SUMBAWS Report 1 Oct 2002 - 31 Dec 2005

FINAL REPORT SUMMARY Section1: Project identification

NOT CONFIDENTIAL

Title of the project: Sustainable management of interactions between aquaculture and wild salmonid fish Acronym of the project: SUMBAWS Type of contract: Shared cost RTD

Total project cost: (in euro) 2 370 803 €

Contract number

Duration (in months)

Q5RS-2002-00730

39 Months

Commencement date

Period covered by the progress report

1 October 2002

1 October 2002 – 31 December 2005

EU contribution: (in euro)

PROJECT CO-ORDINATOR Name:

Title:

Address:

Neil Hazon

Dr

University of St Andrews Gatty Marine Laboratory East Sands, St Andrews, Fife KY16 8LB UK

Telephone:

Telefax:

E-mail address:

+44 - 1334 - 463451

+44 - 1334 - 4634432

nh1@st-andrews.ac.uk

Key words (5 maximum - Please include specific keywords that best describe the project.). salmonid, parasite, socio-economics, sea lice, aquaculture World wide web address (the project’s www address ) http://www.st-andrews.ac.uk/~sumbaws

1 615 787 €

SUMBAWS

Final Report

LIST OF PARTICIPANTS Participant 1 and Project Co-ordinator: University of St Andrews, USTAN Dr Neil Hazon/ Prof Chris Todd School of Biology Gatty Marine Laboratory University of St Andrews Fife KY16 8LB United Kingdom

Telephone Fax E-mail

+44-1334 463451 +44-1334 463443 nh1@st-andrews.ac.uk

Participant 2: Economic and Social Research Institute, ESRI Prof Brendan Whelan Economic and Social Research Institute 4 Burlington Road Dublin 4 Ireland

Telephone Fax E-mail

+353-1 6671525 +353-1 6686231 brendan.whelan@esri.ie

Participant 3: Central Fisheries Board, CFB Dr Patrick Gargan Central Fisheries Board Mobhi Boreen, Glasnevin Mobhi Road Dublin 9 Ireland

Telephone Fax E-mail

+353-1 8379206 +353-1 8360060 Paddy.gargan@cfb.ie

Participant 4: Norwegian Institute for Nature Research, NINA Dr Bengt Finstad NINA Tungasletta 2 Trondheim N-7485 Norway

Telephone Fax E-mail

+47-73 80 14 00 +47-73 80 14 01 bengt.finstad@nina.no

Participant 5: Norwegian Institute of Fisheries and Aquaculture Ltd, NIFA Dr Pål Arne Bjorn NIFA University Campus, Breivika Tromsø N-9291 Norway

Telephone Fax E-mail

+47-77 66 74 18 +47-77 66 74 10 paal-arne.bjorn@fiskeriforskning.no

Participant 6: Radboud Universiteit Nijmegen, RUN Prof S.E. Wendelaar Bonga Radboud Universitiet Nijmegen Department of Animal Physiology PO Box 9010 Toernooiveld 1 Nijmegen 6500 GL Netherlands

Telephone Fax E-mail

+31-24 3652681 +31-24 3652714 wendelaar@sci.kun.nl

2

SUMBAWS

Final Report

Participant 7: Norwegian College of Fishery Science University of Tromsø, NCFS Dr Roar Kristofferson NCFS, University of Tromsø University Campus Breivika Tromsø N-9037 Norway

Telephone Fax E-mail

+47-77 64 51 11 +47-77 64 60 20 roarkr@nfh.uit.no

LIST OF SUB-CONTRACTORS Alpharma (Sub-contractor to Participant 4, NINA) Dr Bernt Martinsen Alpharma Harbitzalleen 3, P.O. Box 158 Skøyen, N-0212 Oslo Norway

Section 2: Project Final Report

Telephone Fax E-mail

+47-22 52 90 88 +47-22 52 90 90 bernt.martinsen@alpharma.no

NOT CONFIDENTIAL

Research Objectives The main Objectives of this proposal are eight-fold: Objective 1. An assessment of the socio-economic importance of salmon/sea trout game-angling and salmon aquaculture in Scotland, Ireland and Norway, including the interactions of the two industries, and identification of the potential social and economic benefits from improved sea lice control. Objective 2.

Identification of the initial migratory routes of salmon and sea trout, utilizing telemetry of wild post-smolts in a selected Norwegian fjord. Objective 3. Identification of areas within fjords of post-smolt susceptibility to sea lice infestation. This will be based on Objective 2 and involve studies of caged fish.

Objective 4. An analysis of the effects of parasite burden on subsequent migratory behaviour; especially ‘premature return’ of sea trout. Objective 5. Specification, for post-smolt wild sea trout and salmon, of the threshold infestation levels at which physiological stress is significantly elevated and to potentially lethal levels and investigations of the physiological effects of ‘premature return’. Objective 6. A synthesis of the results of Objectives 2-5 to understand the effects of the ‘premature return’ on the biology of sea trout.

Objective 7. To quantify and assess the efficacy of Substance EX/Emamectin (= SLICE®) as protection from sea lice infestation. We will assess the possibilities of a remedial, problem-solving, sustainable approach to local stock management by enhancing the returns of naturally-produced smolts. Objective 8.

Recommendations. Based on all aspects of the project.

Expected deliverables The expected deliverables of the project are (a) a socio-economic analysis of the value of aquaculture and sport fishing and how these industries interact(b) the identification of migratory pathways and risk of sea lice infestation in relation to infestation pressure (c) the development of an infection risk for sea trout and salmon for use in the future location of farm sites in relation to migration pathways, (d) a series of recommendations for setting the threshold burden of sea lice for wild and farmed fish (e) determination of the physiological benefits of premature return (f) assessment of the use of prophylaxis as a remedial treatment for the enhancement of threatened wild stocks (g) a series of recommendations to promote the future sustainable development and increased benefits of both the aquaculture and sports fishing industries.

3

SUMBAWS

Final Report

Project’s actual outcome Telemetry techniques were used to monitor salmonid post-smolt movements in fjoirdic areas with different salmon lice pressures. Manual tracking of wild sea trout post-smolts (Salmo trutta) showed that fish swam in the fjord in all compass directions. Both wild and hatchery-reared sea trout post-smolts had a slower rate of migratory progression than wild and hatchery-reared Atlantic salmon post-smolts and remained largely in the inner part of the fjord system, close to the shore. Atlantic salmon post-smolts stayed longer in the inner part than the outer parts of the fjord system, but in contrast to sea trout, migrated through the whole fjord system into the ocean. The swimming speed of recorded fish did not differ between wild and hatchery-reared Atlantic salmon post-smolts indicating that the behaviour of hatcheryreared post-smolts was representative of wild salmon. Neither salmon lice infection nor pharmaceutical prophylaxis had any effect on survival and migration of Atlantic salmon post-smolts. Field observations indicated the importance of predation pressure with several migrating tagged salmon smolts eaten by gulls or predatory fish. A multivariate model was developed to predict the pattern and risks of salmon lice infestations on wild salmonids and this model can be used to predict the effects of new fish farms on specific areas of fjords and coastal areas.

A protection zone has been established in the Romsdalsfjord system, Norway, in which no new salmonid farms can be established and existing farms must operate under a strict management regime. Wild sea trout sampled outside the protection zone (Karlsøyfjord) had higher prevalence and significantly higher abundances of sea lice than sea trout inside the protection zone (Eresfjord). However, infestation levels observed on wild sea trout inside the protection zone were higher than in completely farm-free fjords.

These results imply that protection zones are effective, but must be re-evaluated particularly in terms of size. Sentinel cages were used as an alternative method of monitoring salmon lice infestation pressure in fjord systems, and to confirm the relationship between extensive fish farming and lice infestation pressure.

Naive hatchery reared Atlantic salmon smolts were placed in both pelagic and littoral areas of 3 fjords, and lice infestation was assessed after 14 days. Eresfjord (no farms) had the lowest infestation risk, Langfjord (2 farms) was intermediate, while Karlsøyfjord (extensive farming) had the highest infestation pressure.

No clear differences between littoral and pelagic areas of the fjord were found.

To determine any species differences in susceptibility to salmon lice infestation, Atlantic salmon, sea trout, and Arctic charr were studied in the Altafjord system. Atlantic salmon smolts were only captured during June/July and the majority were uninfested with salmon lice. Pelagic-feeding sea trout and Arctic charr had surprisingly high infestation levels in June/July. In contrast, littoral-feeding sea trout and Arctic charr had very low infestations during June, which increased in July, and peaked in August. These observations indicate that the direct migration of Atlantic salmon out of coastal areas allows them to avoid lice infestation, in contrast to sea trout and Arctic charr, which feed within the fjords throughout the summer and consequentially are exposed to increased infestation.

The physiological consequences of ‘premature return’ to freshwater (FW) in response to sea lice infestation were examined in laboratory studies by infesting SW-acclimated wild sea trout smolts with L. salmonis and either, 1) maintaining fish in SW or 2) returning fish to FW 19 days post infestation (with respective non-infested controls for both groups). Following FW return, the mean infestation intensity and number of mortalities were significantly reduced compared to fish maintained in SW. Plasma concentrations of chloride and lactate were significantly higher in the SW infested group than in all other groups after 21 dpi. Liver glycogen content was significantly decreased following sea lice infestation and remained reduced in the SW infested group, but there was evidence of recovery following return to FW. Plasma cortisol concentration increased in both infested groups at 14 dpi compared to non–infested controls, but returned to control levels in fish returned to FW. These results suggest that premature return to FW is of considerable physiological benefit to post-smolt sea trout.

The physiological effects of simultaneous abrupt seawater (SW) entry and sea lice infestation were investigated in the laboratory for both wild sea trout smolts and farmed Atlantic salmon. Osmoregulatory, metabolic and stress physiology parameters were assessed in both species alongside observations of general epithelial integrity of skin and gill tissue in the sea trout. Results for sea trout indicated mobile lice caused significant increases in plasma concentrations of chloride, glucose, lactate, cortisol and plasma osmolality and a significant reduction in haematocrit. Deterioration of skin and gill epithelium and increased numbers of chloride cells, mucus cells and proliferating cells were also evident. Salmon results indicated significant decreases in haematocrit and increases in plasma concentrations of chloride, glucose, lactate and cortisol and plasma osmolality, following development of sea lice to the mobile stages. The 4

SUMBAWS

Final Report

novel application of piecewise linear models allowed, for the first time, the derivation of a threshold lice intensity, above which the host fish suffered sublethal physiological stress. This analysis identified a consistent breakpoint across several physiological markers and allowed the objective determination of an overall threshold lice burden of 13 mobile L. salmonis fish-1 for sea trout, and 20 mobile L. salmonis fish-1 for salmon.

In Norway groups of hatchery reared Atlantic salmon smolts were Carlin tagged and subsamples treated with Substance EX. Fish were released at 2 sites, (Ims and Halselva) in 2003 and 2004. Results suggest that EX treated Atlantic salmon had significantly better growth than untreated smolts, which may influence spawning success and effects on future generations. In NW Scotland wild sea trout smolts were PIT-tagged and a subsection treated with Substance EX. Treatments took place at the Manse Loch System (2003-2005) and Shieldaig (2004-2005).Whilst prophylactic treatment had no significant effect on survivorship, length or weight, there were significant effects on condition factor between the treated and control groups. In Ireland (Invermore and Gowla) a subsection of wild sea trout smolts were treated with SLICE® prior to release. Equal numbers of fish from treated and control groups were recorded from the Gowla, whereas a greater number of SLICE®-treated fish returned to the Invermore in 2004. On the Gowla, in 2005, there were no statistical differences between treated and control groups in terms of growth, weight, chalimus abundance, total lice level or days spent at sea. Prophylactic treatment of hatchery-reared salmon smolts revealed a significant difference in the return rate of SLICE®-treated and control salmon smolts in three of four release groupings. These data suggest reduced mortality of smolts in the treated groups and hence protection from sea lice infestation in aquaculture bays.

Lepeophtheirus salmonis and Caligus elongatus on returning wild 1SW Atlantic salmon was monitored annually at Strathy Point, Scotland. From 7 years of data the key findings are: 1. Mean abundance of L. salmonis in any one year shows an apparent inverse relationship to variation in C. elongatus abundance. 2.

Within any one year there are strong positive correlations between the intensity of L. salmonis and the intensity of C. elongatus on any one fish. Thus, individual fish in any one year appear to be equally vulnerable to infestation by either species, but the two species do fluctuate in abundance independently. C. elongatus infections of wild salmon evidently have been seriously underestimated in the past and our concern regarding this lack of knowledge is exacerbated by the fact that C. elongatus is a major natural parasite of cod (Gadus morhua), the next species to comprise a major international aquaculture industry.

Socioeconomic analysis of the overall economic impact of Atlantic salmon to the national economies of Ireland, Norway and Scotland was estimated at €1,320m., of which €1,110m. was in aquaculture (representing 15,000 FTE’s), €200m. in salmon angling (6,400 FTE’s) and €8m. in commercial fishing (460 FTE’s). Cost/Benefit ratios of the economic impact of lice treatments demonstrated that a 20% improvement in smolt runs would lead to benefits exceeding costs in all three countries and a 10% improvement in smolt survival, would lead to benefits exceeding costs in Scotland and Ireland. The socioeconomic an d scientific material has been used to develop recommendations including: (a) a series of guidelines for setting the threshold burden of sea lice for wild and farmed fish, (b) considerations for the future location of farm sites in relation to migration pathways, and (c) the use of prophylaxis as a remedial treatment for the enhancement of threatened wild stocks. Together these will allow the future sustainable development and increased benefits of both the aquaculture and sports fishing industries.

Broad dissemination and use intentions for the expected outputs The major beneficiaries of the project are the aquaculture and tourism industries in peripheral regions of the EU with fragile economies as the outputs from the project will allow the future sustainable development of both industries. The results also benefit government policy and decision makers at both national and EU levels in setting uniform environmentally relevant industry standards. University and government scientists with interests in fundamental aspects of the physiology and ecology of sea lice host interactions, and the increased understanding of the spread of salmon lice between farmed and wild salmonids directly benefit from an increased understanding of these complex scientific issues

5

SUMBAWS

Final Report

FINAL REPORT TABLE OF CONTENTS 1

Scientific outcomes, presented as separate scientific papers for each workpackage Executive Summary Conclusions Abstract Introduction Materials and Methods Discussion Conclusions References Workpackage 1 Workpackage 2 Workpackage 3 Workpackage 4 Workpackage 5 Workpackage 6 Workpackage 7 Workpackage 8

7 36 86 137 151 192 198 223

2

Exploitation and dissemination of results

279

3

Policy related benefits

288

4

Annex

290

6

SUMBAWS

Final Report

Workpackage 1 Lead partner: ESRI Partners involved: ESRI, NINA OBJECTIVES • • •

quantify the socio-economic importance of the aquaculture and game angling sectors to the individual national economies. determine the economic effects (benefits and costs) of mutually acceptable sea lice treatments in the aquaculture sector, including the inter-industry effects. provide an overall cost-benefit analysis of sea lice, including both financial and socio-economic elements.

DELIVERABLES A report entitled “Assessment of the socio-economic value of aquaculture and sport angling for wild salmonids in north-western Europe: implications for treatments for sea lice infestation”. This report will contain a cost-benefit analysis of treatments to control sea lice infestation of farmed salmon and their potential impacts on wild stocks. Specifically, cost-benefit analyses of the interventory prophylactic protection of wild smolts in the small, defined river systems (in Ireland, Scotland and Norway WPs 5-7) also will be made.

EXECUTIVE SUMMARY The basic approaches used included the following.

•

Literature search: Gather all existing documentation from government, industry and academic sources relating to the production and economics of the aquaculture and game angling businesses.

•

Review literature to develop country specific aquaculture models that relate inputs (costs and quantities) to output levels (value and quantities).

•

Review literature on the economic and social benefits of game angling to develop indicators which relate game angling expenditure, dependent employment, and incomes to fishing quality and wild population levels.

•

Incorporate results from WPs 2, 4-7, on the biological effects on wild stocks, and the cost to the aquaculture industry of the sea lice treatments into the economic models to estimate the economic implications for both the aquaculture and game angling sectors.

•

Meet with key government and industry personnel to (a) review the models developed for the aquaculture and game angling sectors, (b) elicit opinions on how the models might be adjusted for improvement, (c) discuss model results with respect to the socio-economic implications of the sea lice treatments, (d) elicit comments on how the analysis might be improved based on the key personnel’s knowledge of the sectors.

The report “Assessment of the socio-economic value of aquaculture and sport angling for wild salmonids in north-western Europe: implications for treatments for sea lice infestation” consists of four chapters which are outlined in the following papers.

Paper 1 is entitled Economic Evaluation of the Salmon-Based Businesses. Building on the review of methodology and results from the literature on economic evaluation of the three salmon-based industries (angling, commercial fishing and fishing farming), this chapter presents a comparative economic analysis across the three countries Ireland, Norway and Scotland.

7

SUMBAWS

Final Report

The literature on economic evaluation of Atlantic salmon was thoroughly reviewed and a decision taken to adopt an “economic impact” approach. This means that certain types of non-consumptive value are not taken into account but the existing literature and data is too limited to provide reliable, comparable estimates of these. The approach adopted did, however, allow us to estimate for the first time the overall economic impact of the Atlantic salmon, across the three main sectors (game angling, commercial netting and salmon farming) in Ireland, Norway and Scotland.

The overall gross value of Atlantic salmon was estimated €1,900m., of which approximately €1,500m. relates to salmon aquaculture, €350m. to salmon angling and €10 m. to commercial netting of salmon. When these gross values are corrected for “import content”, displacement effects and multiplier effects, the overall economic impact of Atlantic salmon to the national economies was estimated at a total of €1,320m., of this €1,110m. in aquaculture, €200m. in salmon angling and €8m. in commercial fishing.

Overall full-time job equivalents were estimated at 22,000 FTE, of which 15,000 were in the farming industry, 6,400 in angling tourism and 460 in commercial fishing.

There are striking differences in the size and nature of the businesses in each country. Norway has a huge farming business and a significant angling tourism sector, which is, however, less developed than in Scotland and Ireland. Scotland has a significant farming business and a well developed angling tourism.

Ireland has a relatively substantial commercial fishery, which has, however, only limited economic impact, and a small farming business compared to Norway and Scotland. Generally, commercial fishing is insignificant in all countries compared to the two other sectors. The angling sector is, in view of the limited attention it receives in economic discussions, of surprising economic importance in all the countries. It is a relatively significant generator of jobs, especially on a “jobs per fish” basis. This is consistent with the predominance of service-type jobs in sectors linked to tourism.

Paper 2, Costs of Sea Lice Treatment on Farms, is based on published data, where this was available, and also on interviews and contacts with industry participants, veterinary experts and specialists from state agencies. It presents detailed estimates of the patterns and costs of sea lice treatments in the three countries.

These costs include medication, labour, oxygen etc. As a way of operationalising these cost data for our models, we derive an illustrative estimate of the cost of an extra or “marginal” treatment on all farms. This figure is input into the cost/benefit analysis described in paper 4 below.

Cost associated with improved/intensified sea lice treatments in farms were calculated for all three countries, and the costs are strongly correlated to the overall production of farmed salmon in each country.

For illustrative purposes, the costs of a marginal, prophylactic treatment was estimated at €11.65m.; €3.28m. and €0.38m. for Norway, Scotland and Ireland respectively. The costs as a percentage of total production value were found to lie in the range 0.46 to 0.36 % for the three countries. The cost for a marginal treatment on a ‘per kg produced fish’ basis was very similar across the countries and lay in a very narrow range, € 0.011 – €0.018. These estimates are on a nationwide basis in each case. It is likely that treatment could be carried out more cost-effectively if it were targeted at areas or river systems where the returns were especially high, such as where very significant runs of wild fish occur and/or where infestation levels are especially high. More precise information is needed on the quantitative relationship between reductions in lice levels and improved runs.

The third paper, Consequences of Stock Collapse and Costs of Preventative Treatment, assembles information on the Irish sea trout collapse and provides unique empirical data on the economic consequences of that collapse arising from reduced angling. Based on the experience of colleagues engaged on WP 7, it also presents estimates of the cost of trapping and treating wild smolts on a preventative basis.

8

SUMBAWS

Final Report

This part of the report focuses on what can be learned from the collapse in sea trout populations in the West of Ireland since the late 1980s. It documents and quantifies the resulting decline in angler numbers and estimates the consequential drop in tourist expenditure in the region. Angler numbers fell by more than 50% and the loss in revenue from angling is estimated at â‚Ź1.4m per annum for the Connemara region (in present day values).

By utilising some of the results and experience gained in Workpackage 7 of the SUMBAWS project, which is also based in the West of Ireland region, cost estimates for the preventative treatment of wild smolts are produced. It is shown that the feasibility of widespread use of preventative treatment is questionable and that, even if feasible, very substantial costs would be involved.

Paper 4, Modelling the Economic Effects of Sea Lice Treatment, develops a computer-based model of the lice treatment process and its probable economic effects in the three countries. This model can be used to simulate the effects of different assumptions about the magnitudes and responsiveness of the variables in question (costs, catches, angling participation rates etc.), to derive estimates of the economic benefits arising and to show how these benefits relate to the costs of treatment. A number of plausible sets of assumptions are examined. A simulation is also carried out of a regional situation where wild rather than farmed fish are treated.

Based on the material in earlier chapters, a computer model is developed of the key stages in the assumed causal chain between enhanced sea lice treatment, through better smolt survival and improved returns to the coast to economic, and social benefits in the form of greater income and employment in the commercial and angling sectors. The main output from the model is a series of Cost/Benefit ratios aimed at summarising the economic impacts from treatments with varying levels of effectiveness. A number of plausible scenarios are simulated to illustrate how these ratios vary with changes in the assumptions. It is found that a 10% improvement in the smolt survival, combined with a moderate response by anglers to the improved runs, would lead to costs of treatment exceeding benefits in Norway, while benefits would exceed costs in Scotland and Ireland. A 20% improvement in smolt runs would lead to costs exceeding benefits in all three countries. In all the cases examined, the benefits would exceed costs in Ireland (due to the small size of the farming sector and the good returns in terms of expenditure and employment to improved angling catches).

Some of the estimates incorporated in the model are speculative. However, the main value of the modelling approach is that it clearly highlights what factors are most important for the Cost/Benefit ratio and how they interact. Thus, it clear that the cost of marginal treatment depends crucially on the size of the fish farming sector. This means, for instance, that the cost in Norway for a nationwide treatment is very high. It would be very important in this case to target and limit the intervention to particular regions or systems where the benefits are expected to be especially large, such as key angling systems or areas where lice infection is especially high.

The models also emphasise what types of benefit are likely to arise and to whom they will accrue. Thus, in all three countries the largest benefits from improved runs are likely to accrue to the angling sector rather than to the commercial fishing operations, due to the relatively small size of the latter. The modelling approach also indicates the key areas on which future economic research should focus. These include improved estimates of angler response, better quantified biological information on the likely response in terms of improved runs to improvements in treatment and to more detailed data to improve the estimates of the costs involved on a national, or perhaps even more importantly, a regional or river system basis.

9

SUMBAWS

Final Report

CONCLUSIONS The present report provides a socio-economic evaluation of the interactions between the various salmonbased businesses (aquaculture, commercial fishing and angling) in three countries Ireland, Norway and Scotland. It compiles up-to-date, comparable figures for the value of the various sectors, examines in detail the costs involved in lice reduction and devises a methodology for estimating and comparing the costs and benefits involved.

It is important to emphasise two dimensions which are excluded from this evaluation. First, it is aimed at assessing the costs and benefits accruing to the participants in the salmon-based industries and to examining how these would change under different sea lice control regimes. It does not attempt to assign responsibility for who should pay or compensate for any negative effects arising. For this reason, it does not discuss in any detail the application of the principles underlying the EU Habitats Directive to salmon. This Directive “aims to anticipate, prevent and attack the causes of significant reduction or loss of biodiversity at the source” (European Commission 1998) and implies that policy should aim to prevent damage to wild stocks occurring on any substantial scale. Its effect is similar to the “polluter pays” principle which requires the person or organisation causing the damage to pay for its alleviation. In deciding on policy, these principles would have to be taken into consideration along with the economic costs and benefits described below.

Secondly, the evaluation is based on a static “as is” situation. It does not take account of the fact that stocks of wild salmon in most countries are today well below their historic, long-run (and hence potential) levels. This means that runs could be substantially and sustainably improved if spawning escapement to freshwater was greater. Thus, if improved sea lice control led to better runs this would ultimately increase future stocks and hence yield benefits over and above the current short-run benefits as estimated in this report.

10

SUMBAWS

Final Report

Paper 1: ECONOMIC EVALUATION OF THE SALMON-BASED BUSINESSES Ø. Aas, I. Unglem, J. Curtis, B. Dervo and B. Whelan (Manuscript) Abstract This paper examines the economic significance of the Atlantic salmon (Salmo salar L.) in three countries: Ireland, Norway and Scotland. These countries are among the most important countries for conservation and use of Atlantic salmon. The paper begins by considering the various approaches which can be used in economic evaluation of fisheries and aquaculture, and opts in this instance to focus on the “economic impact” method. It derives broad-brush estimates for output, employment and overall economic impact for the three main sectors based on Atlantic salmon (recreational angling, commercial fishing and aquaculture) in each country. It goes on to show how these can be used to characterise the manner in which the salmon stocks are utilised in each country and concludes that the economic impact of recreational exploitation, compared to aquaculture, measured either in terms of output, employment or net impact, is more significant than is sometimes thought. Commercial fishing for salmon is insignificant compared to the two other sectors. In the final section, it enumerates a number of considerations which need to be borne in mind when using the data presented to make decisions and plans about the management, conservation and enhancement of salmon stocks, and identifies further research needs.

Introduction The Atlantic salmon (Salmo salar L.) is probably the best known and most valuable freshwater fish species around the North Atlantic coast (Mills, 1991). It is one of the world’s most prized game fish, attractive in fish markets, and, during the last decades, has become a major aquaculture species. In short, the Atlantic salmon makes a valuable contribution to the economies of several countries in the North Atlantic.

Despite a general recognition of the huge social and economic values associated with Atlantic salmon, and the development of a substantial literature on individual regions and countries, few studies have reported and estimated the overall economic value of Atlantic salmon across a number of countries and compared the value of different types of use.

There are probably a number of reasons for this dearth of work. First, as will be shown below, the term “economic value” can be defined in many ways, and therefore the studies use a wide variety of approaches and conceptual frameworks, making accurate comparison very difficult (see, for instance, Radford et al.

2001). There have also been wide variations in the methodology and data used. Furthermore, most studies are focused on elucidating policy decisions in a single country, region or locality.

Conceptual Approaches Overall, the literature on economic evaluation of fisheries utilises two basic frameworks of analysis which differ fundamentally in their objectives, data requirements and policy focus. Pollock et al. (1994) label these the “net value” and “economic impact” approaches. Each of these approaches yields a value expressed in monetary terms, but they differ sharply in what they are valuing and to whom this value accrues.

Net Economic Value This approach attempts to quantify the benefit received by an individual or group from a resource, product, service or experience, having allowed for the cost involved in obtaining it. The benefits can be monetary (e.g. net income of commercial fishermen) or non-monetary (e.g. the satisfaction an angler gets from a trip to a particular fishery). Costs can also include non-monetary factors such as time or inconvenience.

Net economic value1 can include both consumptive and non-consumptive (non-depleting) uses of the resource. The main categories of non-consumptive value are: 1

In the literature there is neither a standard classification of the components of Net Economic Value nor a uniform terminology.

11

SUMBAWS

Final Report

Existence value: the value placed by society on sustaining a resource for its own sake Bequest value: the value placed by society on passing on the stock in healthy condition to posterity Option value: the value placed by non-users on retaining the possibility of using the stock in the future These values have been estimated as very substantial e.g. a UK study of the value of restoring/keeping salmon in the Thames gave a figure of € 18 m. (Spurgeon et al 2001).

The main consumptive use values of a fish resource relate to the net incomes (and producer surplus if any) that commercial fishermen derive from the fishery, along with the satisfaction that anglers derive from using the fishery. The latter can be estimated by assessing their “net willingness to pay” or total consumer surplus. Net economic value is derived by subtracting the “opportunity cost” of the resources used in producing the angling experience. In open access (unpriced) fisheries such as are almost universal in the USA, a variety of survey techniques have been developed to measure anglers’ consumer surplus. These include: • The Travel Cost Method (e.g. Clawson and Knetch 1966) • Hedonic Pricing (e.g. Gillen 1982) • Contingent Valuation (e.g. Pollock et al. 1994 Section 16.5) With private ownership of recreational fishing rights, owners of these rights can charge anglers for access to the fishery. Thus much of the anglers’ consumer surplus that would exist in an open access fishery is converted to an income flow for fishery owners. The market price or capital value of the fishery is capitalisation of the annual income flow (Radford et al 1991). In these circumstances, the net economic (use) value comprises anglers’ consumer surplus and the owners’ income flow.

In short, the net economic value approach attempts to assess the overall value to society of the net benefits arising from a resource, such as a fishery, whether or not these benefits actually give rise to a monetary transaction.

Economic Impact Studies This approach focuses mainly on statistics of turnover, expenditure and related variables. The principal aim is to assess the impact of expenditure on income generation, jobs and overall economic activity in a region. In the case of angling, the key assumption is usually that anglers’ expenditure is bringing under- (or un-) utilised resources into productive use. In assessing the effect of angling expenditure, it is vital to focus on the extent of “displacement”, i.e. to consider what would happen to the expenditure if the resource were to disappear or become inaccessible. Only those expenditures that would disappear if the resource became unavailable can be counted as part of its economic impact.

However, the economic interconnection between angling, commercial fishing and aquaculture is well established. Commercial fishing and recreational fishing compete for the same fish; farmed fish compete in the fish market with netted wild fish; and farming can cause harm to wild stocks, and thereby both to angling and commercial fishing by the spreading of sea lice and by escapees with the potential to have a long-term genetic impact on the stocks (McGinnity et al 2003). To make good, overall management assessments of the whole economic sector related to Atlantic salmon, it is important to take into account the overall economic benefits arising from the presence of Atlantic salmon as a resource. To be able to do this, a better understanding of the economic impact on different countries and different uses of this species is needed.

Given existing literature and data sources, and in light of the ownership structure of salmon fisheries in Europe, it is considered that the economic impact approach is likely to be the most relevant for the present study. This approach is likely to yield values which are conceptually comparable with the estimated incomes and costs in the aquaculture sector.

Objective Ireland, Norway and Scotland are all among the major countries for Atlantic salmon in the eastern part of the North Atlantic (ICES working group on Atlantic salmon report 2004). This paper summarises and updates existing data on the economic value of angling, commercial fishing and farming of Atlantic salmon in the three countries. On this basis, major similarities and differences between the businesses and the three countries are identified and discussed. Identified differences can be used to improve the economic value in different regions and sectors. Lastly, further research needs are identified.

12

SUMBAWS

Final Report

Materials and Methods As indicated above, there are a variety of studies in the different countries based on different years, having different conceptual and methodological approaches and covering different geographic areas.

These were reviewed in order to identify those most suitable for our present purpose in terms of recency, comparability and coverage. In Scotland, a recent study (Radford et al. 2004) assessed the economic impacts of angling. In Ireland, a national study on the wild salmon sector (angling and commercial) was published in 2003 (Indecon 2003). In Norway, no recent national study of angling or commercial fishing impacts was available, so different statistics and numbers were combined to provide the best overall estimate. In order to examine the impacts from commercial fishing and aquaculture, official catch and production statistics were used in all three countries. A lot of the data used, especially for Norway, were found in so-called grey literature (Colette 1990).

We endeavoured to make the information as comparable as possible by choice of reference year (2002 where possible, 2001 otherwise) and translated monetary amounts into Euro at the exchange rate prevailing in mid 2002 (€1 = £ 0.65 and NOK 9.0, as indicated by data from the European Central Bank). Table 1 gives an overview of data sources and the reference years for each country.

Table 1 Data sources (year published and reference year) for information on the economic impact of Atlantic salmon in Ireland, Scotland and Norway.

Number of anglers Number of fishing days Angling expenditures Angling employment Commercial fishing –catches Commercial fishing – gross value Commercial fishing – employment Aquaculture production Aquaculture production – turnover

IRELAND Indecon 2003 2001 Indecon2003 2001 Indecon 2003 2001 Indecon 2003 2001 CFB 2003 2002 Indecon 2003 2001 Indecon 2003 2001 Bord Iascaigh Mhara 2003 2002 Bord Iascaigh Mhara 2003 2002

SCOTLAND Radford et al 2004 2002 Radford et al 2004 2002 Radford et al 2004 2002 Radford et al 2004 2002 Fisheries research Services 2003 2002 Fisheries research Services 2003 2002 Fisheries research Services 2003 2002 Fisheries research Services 2003 2002 Fisheries research Services 2003 2002

NORWAY Dervo & Knutsen 2004/Aas 1997 2002 Fiske & Aas 2001 2001 Fiske & Aas 2001 2002 N.A. Mørkved & Krokan 1997/ Norway official statistics 2004 2002 Mørkved & Krokan 1997/ Norway official statistics 2004 2002

FHL Aquaculture 2004 2002 FHL Aquaculture 2004 2002

Results The results, based on these sources, are set out in Table 2.

Ireland – angling Based on the number of salmon licences issued, Ireland has approximately 33,000 salmon anglers. These anglers consist of one third foreign anglers and two thirds domestic. In total, they catch around 29,000 salmon (78 tonnes). Their combined expenditure is estimated at €80m. and generates employment of about 1,300 full-time job equivalents (FTE).

13

SUMBAWS

Final Report

Ireland - commercial fishing There are two main commercial salmon fisheries in Ireland – drift and draft net fisheries. They catch a total of 227 000 salmon (604 tonnes), with a gross value of about €4.8m. The drift net fishery accounts for approximately 80 – 85 % of the catches. The employment effect is estimated at 301 FTE. Ireland – farming Salmon aquaculture in Ireland totalled 21 400 tonnes in 2002. The gross value of the production was estimated at €78m with a total employment effect of 300 FTE. Scotland – angling There were 33 878 salmon and grilse caught and retained and a further 24 042 caught and released in the Scottish angling fishery in 2002 (Fisheries Research Services, 2003). There are no comprehensive data on the numbers of people who fish for salmon in Scotland, but a rough estimate is considered to be 60 000 anglers. Their overall expenditure is estimated at €113m. The employment effect was estimated at 2 200 FTE. Scotland – commercial fishing There are two main remaining Scottish commercial salmon fisheries. These are the net and coble fishery and the fixed engine fishery. The fixed engine includes stationary bag nets, stake nets and jumper nets. The net and coble are beach seine operations in estuaries and lower parts of rivers. They were reported as catching 24 000 salmon with a gross value of about €0.6m. representing a job effect of 10 – 15 FTE. Scotland – farming The Scottish aquaculture sector produced 159 000 tonnes of salmon in 2002, with a gross value of about €460m. The employment effect was estimated at 5 000 FTE. Norway – angling Norwegian angling-based catches of Atlantic salmon have been around 110 000 fish per year during the last three year period. There is no comprehensive, recent study of the number of salmon anglers, and their expenditures and valuation of salmon and sea-trout fishing in Norway.

Every salmon angler in Norway is supposed to pay a national salmon angling licence fee, but there is evidence that many fish without paying this fee (Aas 1997). The number of anglers paying the fee have, during the last 15 years, varied between 75 000 and 150 000, with approximately 90 000 anglers paying the licence in recent years. When asking a sample of the Norwegian population about their participation in hunting and fishing, far higher numbers of salmon anglers occur (Aas 1997), but this method generally overestimates participation rates. Taking these different figures together, we assess the number of salmon anglers in Norway to be approximately 150 000, of which 35 000 are visiting anglers from abroad (Dervo & Knutsen 2004). Overall expenditures are estimated at €160m. There are no available calculations on jobs generated, but assuming a similar ratio of expenditure/jobs as in Ireland and Scotland, (approximately €55 000 per job), an estimate of about 2 900 FTE can be derived.

Norway – commercial fishing The remaining commercial salmon fisheries in Norway consist of two types of fixed net fisheries; the bag net fishery and the bend net fishery. These fisheries take a total of approximately 125 000 salmon, with a gross value of €3m. No figures are available for jobs generated, but studies show they are small. Table 2 shows an estimated figure of 150 FTE derived by dividing the total catch value by the average catch value per job in Scotland and Ireland (about €20 000 per job).

Norway – farming Norway has an overall salmon aquaculture production of 470 000 tonnes, representing a gross value of 1 billion Euro. The FTE for the business is estimated at 10 000 jobs.

14

SUMBAWS

Final Report

Table 2 Overview of the value of fresh-water game angling, commercial net fishing and aquaculture production of Atlantic salmon in Ireland, Scotland and Norway Ireland Freshwater Angling Salmon Catch* 29 000 Number of anglers (approx) 33 000 of which out-of-state anglers 10 000 Overall expenditure (€m.)

80.0 Employment (full time equivalents) 1 300 Marine/estuarine Commercial fishing Harvested salmon in net fisheries 227 400 Gross value (€m.)

4.9 Employment (FTE) 300 Aquaculture Production (2002, t) 21 Gross value (€m.)

78.0 Employment (FTE) 300

Scotland

Norway

Total

57 900 60 000 28 000 113.0 2 200

110 000 150 000 35 000 160.0 2 900**

196 900 243 000 73 000 353.0 6 400

23 700 0.6 12

125 000 3.0 150***

376 100 9 462

146 460.0 5 000

470 1 000.0 10 000

637 1 538.0 15 300

* Includes both caught and retained and caught and released fish ** Estimate by the authors based on expenditure/job in Ireland and Scotland *** Estimate by the authors based on gross value per job in Ireland and Scotland Economic Impact The studies on angling in Ireland and Scotland (Indecon 2003 and Radford et al. 2004) estimated the overall economic impact of salmon angling by making three adjustments.

•

•

•

First, they allowed for the import content of anglers’ expenditure. If an angler spends on, say, petrol in a region this does not generate activity in the region to the extent of the wholesale cost of the fuel because this had to be imported. Both studies make allowance for this “import leakage”, the Irish study by assuming an overall import content figure of about 40% and the Scottish one by taking each of 14 categories of angler expenditure and examining their impact individually using known import contents for each category.

Next, they make allowance for “displacement” i.e. the extent to which the angler’s expenditure would still occur in the region even if the fishery were to disappear or become inaccessible. It is commonly assumed, for instance, that local anglers will continue to spend in the area (on different products or services) even in the absence of the fishery. In contrast, it is usually assumed that foreign anglers will not visit the region if the fishery is not there. The Indecon study in Ireland assumed a displacement of 85% for Irish anglers while the Scottish study assumed that all the purely local expenditure, and half of the “intra Scotland” trips would represent displacement.

Thirdly, they took account of the “multiplier effects” of spending i.e. the economic impact attributable to rounds of spending subsequent to the first one. The magnitude of this effect was about 1.2 in both studies, but they used different approaches to arrive at it. The Irish study made an overall assumption based on previous results from input/output tables while the Scottish study modelled the inter-sectoral transactions explicitly. When these two adjustments have been made, the resulting estimate represents the economic impact on output, income and employment.

In the case of the commercial and aquaculture sectors, the economic impact is easier to calculate since it is necessary only to take account of import content and the multiplier effect. Little firm information is available on the import content of the commercial netting sector. It is likely to be relatively small and a figure of 20% is assumed. Aquaculture, being more capital intensive and having high feed costs, is likely to have much higher import content and 40% is assumed. These calculations and assumptions allow us to derive the figures for overall economic impact shown in Table 3 and Figure 1. The huge size of Norway’s aquaculture industry is evident, as is the significant contribution made by angling in Scotland.

15

SUMBAWS

Final Report

Table 3 Overall Economic Impact for fresh-water game angling, commercial net fishing and aquaculture in Ireland, Scotland and Norway Ireland

Scotland

Norway*

Total

65 2.9 720

205 8 1,107

â‚Źm.

Freshwater Angling Commercial fishing Aquaculture

15.6 4.7 56

124.6 0.6 331

* The angling figure for Norway is an estimate by the authors assuming foreign anglers in Norway spend 50% more than domestic anglers, and that 85% of local angler's expenditure can be considered "displaced".

800 700 600 500 400 300 200 100 0 Fr es hw at er An gl

Norway* Scotland Ireland

Co m Aq m er ua cia cu l fi ing ltu sh re ing

Figure 1 Economic Impact by Business Sector and Country (Source: Table 1) Discussion The data presented above are derived from a very wide variety of sources of varying methodological and statistical quality. Furthermore, as can be seen from the notes to the tables, assumptions (sometimes quite major ones) were needed to remedy the absence of data for certain cells. The figures should therefore be interpreted as, at best, broad orders of magnitude serving to point up the similarities and differences between the countries and the sectors. This being said, we feel that the estimates for each country and sector are credible and will be of use for further studies in the field.

Despite these deficiencies, the figures illustrate that the Atlantic salmon as a species is of substantial economic significance to countries in the North-East Atlantic. A very crude estimate of gross (unadjusted) value can be obtained by aggregating expenditure on angling, the value of commercial landings and the gross value of aquaculture production, which yields a figure in excess of â‚Ź1.8 billion across the three countries. Total associated employment is estimated to exceed 20 000 FTEs. These figures also contain several striking and interesting differences. First, the values associated with commercial fishing are very small compared to those of angling and fish farming. Only in Ireland is the preponderance of the wild catch devoted to the commercial fishing sector. The proportions of the total wild catch caught by the commercial sector in each country are: Ireland: 88.7% Scotland: 29.0% Norway: 56.0%.

16

SUMBAWS

Final Report

Second, the value of angling, especially when measured in terms of associated employment, is higher than the general impression among businessmen and decision makers. The proportions of total employment in the three sectors accounted for by angling are as follows: Ireland: 68.4% Scotland: 30.5% Norway: 22.2%.

It is clear that the angling sector, with its predominance of service-type jobs, is much more labourintensive than aquaculture. This can be illustrated by relating employment to the size of the angling catch to derive the employment per 100 salmon caught (including caught and released) in the angling sector. The figures for the different countries are: Ireland: 4.5 jobs per 100 fish Scotland: 3.8 jobs per 100 fish Norway: 3.3 jobs per 100 fish.

Table 3 above showed the estimated economic impact of each of the three sectors in each of the countries, when account is taken of the import content of the output of each sector and of the “displacement effect” described above. Commercial fishing accounts for a very low proportion of total economic impact even in Ireland where it generates an estimated impact of only €4.7m., despite the fact that 88% of the total wild catch is devoted to this sector. The international trend is strongly towards a reduction in the commercial catch in favour of increased exploitation by angling (cf Indecon 2003). The economic impact as shown in Table 3 emphasises the current significance of aquaculture whose overall economic impact is estimated to be in excess of €1.1 billion, largely accounted for by the huge Norwegian farming industry.

Key decisions in fishery management relate to how the resources are utilised and how interactions and competition between the sectors are regulated. Gargan et al. (1998) present a modelling methodology for estimating the effect of changing the balance between commercial and angling exploitation. They emphasise the need to focus on the marginal quantum of extra fish and show that the biological and economic outcome of the change in exploitation policy will depend on who will catch these fish (local anglers, visiting anglers, not caught) and the responsiveness of tourist anglers to an improvement in the perceived probability of angling success. They also stress that this model depends on the assumption that the extra fish are over and above the number required to reach the optimum spawning escapement for each river system. If escapement is below the optimum level, increased escapement will benefit not just the angling sector in the season in question, but, since anglers are likely to catch only a minority of the extra fish, the remainder will be available to boost spawning and hence to increase overall stocks in future years. These considerations apply to evaluating the costs and benefits of any salmon management policy, whether it be the choice between angling versus commercial exploitation or balancing the costs and benefits of eliminating negative interactions between aquaculture and angling. In utilising the figures derived above to value improvements in the stocks available to the angling sector, these considerations must be borne carefully in mind.

These crude figures could hopefully also serve as a starting point for further research in the area of economic values and impacts of the Atlantic salmon. As indicated, more precise data are needed to verify these estimates. Even more important, however, is the need to focus new studies on how changes in management, allocation of resources, and conservation measures in each of these interconnected sectors can increase the overall value of Atlantic salmon.

References Aas, Ø. (1997) Characteristics of freshwater anglers fishing without the compulsory national fishing licence in Norway. Fisheries Management and Ecology 4, 77-79.

Bord Iascaigh Mhara (2003) Irish Aquaculture Production 2003 BIM Dublin Central Fisheries Board (2003) Wild Salmon and Sea trout Tagging Scheme. Fisheries Statistics Report 2001-2003 CFB, Dublin Clawson and Knetch (1966) Economics of Outdoor Recreation. Johns Hopkins Press. Baltimore Maryland Collette, B.B. (1990) Problems with grey literature in fishery science. Pp. 27-32 in Hunter, J. (ed.). Writing for fishery journals. American Fisheries Society.

17

SUMBAWS

Final Report

Dervo, B. & Knutsen, S. (2004) Handlingsplan. Næringsutvikling i utmark. (Task plan – development of nature based tourism based on fish and wildlife resources in Norway). Norwegian Commercial Development Fund, Oslo. In Norwegian only. 17 pp.

FHL Aquaculture (2005) Aquaculture in Norway a statistical overview. Last cited 21 June 2005. Link from http://www.fhl.no/category/English/category.php?categoryID=73.

Fisheries Research Services (2003a) Scottish Fish Farms Annual Production Survey 2002. Scottish Executive Edinburgh Fisheries Research Services (2003b) Scottish Salmon and Sea Trout Catches 2002. Scottish Executive Edinburgh Fiske, P. & Aas, Ø. (2001) Laksefiskeboka. (How to manage salmon fisheries). NINA Temahefte 20. In Norwegian only. 100 p. ISBN 82-426-1267-6.

Gargan, P.G., K.F. Whelan and B.J. Whelan (1998) An Illustrative Grilse Survival Model for an Irish Salmon Fishery. In Hickley and Tompkins (eds) Recreational Fisheries – Social, Economic and Mangement Aspects FAO ISBN 0-85238-248-0 Gillen (1982) Economic Value of Salmon Angling: Estimates from Hedonic Functions. Canadian Regional Science Association Meetings, Vancouver Indecon (2003) An Economic / Socio-Economic Evaluation of Wild Salmon in Ireland. Central Fisheries Board, Dublin McGinnity P., P. Prodo, A. Ferguson, R. Hynes, N. O´ Maoiléidigh, N. Baker, D. Cotter, B. O’Hea1, D. Cooke, G. Rogan, J. Taggart and T. Cross (2003) Fitness reduction and potential extinction of wild populations of Atlantic salmon, Salmo salar, as a result of interactions with escaped farm salmon. The Royal Society, London Mills, D. (1991) Ecology and Management of Atlantic salmon. Chapman and Hall. London Mørkved, O.J. & Krokan, P.S. (1997) Inntekts- og kostnadsforhold i det norske sjølaksefisket med faststående redskap. (Revenue- and cost aspects of the Norwegian fixed net fisheries for Atlantic salmon).

Report from The College of Nord-Trøndelag to the Norwegian Sea Ranching Program. In Norwegian only.

61 pp.

Norwegian Official Statistics (2005). Statistics of catches of salmon in fixed net fisheries. Last cited 21 June 2005. Link: http://www.ssb.no/sjofiske/.

Pollock K., C.M. Jones and T.L. Brown (1994) Angler Survey Methods and Their Application in Fisheries Management. American Fisheries Society Special Publication 25, Bethesda, Maryland Radford A.F., A. Hatcher, and D. Whitmarsh (1991) An Economic Evaluation of Salmon Fisheries in Britain. Research Report No. 16 Centre for the Economics and Management of Aquatic Resources, University of Portsmouth Radford A.F., G. Riddington and D. Tingley (2001) Economic Evaluation of Inland Fisheries. Technical Report Environment Agency R& D Technical Report W2-039. Module A Produced by MacAlister Elliott and Partners Radford, A., G. Riddington, J. Anderson and H. Gibson (2004) The Economic Impact of Game and Coarse Angling in Scotland. Scottish Executive Edinburgh Spurgeon J., G. Colarullo, A. Radford, G. Riddington and D. Tingley (2001) Economic Evaluation of Inland Fisheries. Technical Report Environment Agency R& D Technical Report W2-039. Module B Produced by MacAlister Elliott and Partners

18

SUMBAWS

Final Report

Paper 2: COST OF SEA LICE TREATMENT ON FARMS Ø. Aas, I. Unglem, J. Curtis, B. Dervo and B. Whelan (Manuscript) Abstract The aim of this paper is to provide estimates of the costs associated with more effective sea lice treatment in Irish, Scottish and Norwegian salmon farms. To investigate this concept in our models, we have chosen to estimate the costs of one additional, or marginal, prophylactic treatment. It is assumed for illustrative purposes that this additional treatment would reduce the infection level on wild fish significantly. Moreover, the reported cost estimates are compared with the overall value of salmonid production in Ireland, Scotland and Norway. The total estimated costs of one additional or marginal treatment across all salmonid fish farms correlated closely to the production volume in the three countries, with Ireland having by far the lowest cost (€0.38m.) and Norway having the highest (€11.65m.). The total cost estimate for Scotland was €3.28m. The estimated treatment costs as a proportion of the production value (Norway: 0.46%, Scotland: 0.44% and Ireland 0.36%) and the cost per kg produced fish (Norway and Scotland: €0.013 and Ireland: €0.011) were nevertheless similar. The cost estimates presented in this report are most likely maximum estimates of the cost of additional treatment against salmon lice. If an additional treatment were carried out as a mitigative action in a realistic situation it would be possible to reduce the costs since not all farming areas interact with important areas for wild salmonid migration. Furthermore, detailed knowledge of migration patterns and habitat use might enable a pinpointed treatment in cases where the areas overlap and it might be unnecessary to intensify the treatment in areas with a naturally low infestation level. Finally, optimization of the current treatment schedules might also reduce the need of increased treatment.

Introduction There is growing acceptance in the industries and among management agencies in the North Atlantic, as well as a growing number of scientific studies, that supports the theory that the problem of sea-lice infection on wild salmonids is related to salmonid farming. It is believed that salmonid aquaculture affects wild salmonid stocks negatively by drastically increasing the number of hosts for salmon lice in coastal areas where both wild and farmed salmonid smolts meet the marine environment for the first time (Gravil 1996; O’Donoghue et al. 1998; Bjørn et al. 2001; Heuch & Mo 2001; Holst et al. 2003; McKibben & Hay 2004; Penston et al. 2004; Heuch et al. 2005). For instance, Tully and Whelan (1993) and Butler (2002) estimated that > 95% of salmon lice eggs and larvae produced on the west coasts of Ireland and Scotland, respectively, were of farm origin. In addition, infective lice stages may also derive from escaped farmed salmon (Heuch & Mo 2001).

Intensified treatment against salmon lice in fish farms might affect the number of wild fish returning to rivers positively, and as such have effects on the economic returns from tourism fishing based on wild anadromous salmonids. In this context, it is of interest to estimate the economic costs of reducing sea-lice levels in farms to a level that would cause no damage or that will significantly reduce the negative effects on wild stocks.

In the current paper the objective was to estimate costs of intensified treatment. Such data are essential in cost/benefit analyses of the economic gain of precautionary treatments, in terms of increased economic returns from recreational exploitation of wild salmonid stocks. It is, however, complicated to calculate the costs of the treatment required to eliminate completely the negative effects on wild stocks of sea lice originating from aquaculture. Firstly, few studies have looked at the overall costs associated with sea lice treatment in salmon farms. Secondly, there is no knowledge regarding the extent and quality of treatment needed to eliminate the problems imposed on wild stocks by salmonid aquaculture completely. Thirdly, infection rates, treatment pattern and costs of treatments vary significantly between farms, regions and countries. These are variations both on a larger spatial/geographical scale, on a smaller spatial scale and variation over time. In addition, some of the costs related to sea lice are not those related to treatment or lack of treatment, but to those of feeding efficiency and food costs, which are affected, by both sea lice infections themselves, and the stress imposed on the fish from bath treatments.

19

SUMBAWS

Final Report

The problems with sea lice vary between southern and northern farm locations, and between locations depending on freshwater influence. In Norway, sea lice problems have been generally limited in northernmost farming districts, as well as in areas with a good freshwater influence, but larger in southern locations with higher salinity. Local oceanic conditions such as current patterns and freshwater influence also create huge local differences in sea lice problems between locations, sometimes over very small distances. In addition, variation in sea temperatures and freshwater runoff cause temporal variations. High freshwater runoff, very low or very high temperatures all reduce sea lice population growth.

The treatment strategies also vary between countries. Generally, treatment against sea lice has three main measures; oral medicaments, varying bath treatments and use of biological control through placement of wrasse in the pens. Normally, smaller fish are treated with oral medicaments. Wrasse are used for controlling lice levels on smaller fish (mostly in Norway south of Troms county). On larger fish, bath treatments are most common, but oral treatment is widespread on larger fish in Scotland and Ireland. Oral treatment is also becoming more common on larger fish in Norway. Strategies for treatment differ somewhat between the three countries, partly because of differences in traditions, because of different regulations relating to chemotherapeutants, and partly due to differences in how the aquaculture industry experiences the sea lice problem. Nevertheless, it is possible to assess some of the costs of treatments at a large scale across a majority of farms in all three countries, in order to indicate the range for such costs.

The aim of this report was to estimate the costs of a more intensive sea lice treatment effort in the farms, and as an example, we have estimated the costs associated with one additional, prophylactic treatment at an overall national level. This additional, or marginal, treatment is envisaged as acting as a precautionary action to supplement the existing treatment schedule.

We do not have evidence that a treatment regime of this kind would be the most appropriate or effective in practice. However, for illustrative purposes, it seems reasonable to assume that such a treatment would have a significant effect on the overall lice infestation levels in the fish farms and thus probably also on the infestation levels in wild salmonid stocks. Theoretically, this could produce an economical benefit for recreational fisheries based on wild salmonids. However, it is speculative to link one additional treatment to a specific reduction in lice infestation level on farmed fish (for instance zero level of female ovigerous lice) to a given increase in wild fish survival and/or returns. More research is needed to quantify this link more precisely.

Furthermore, the costs of intensified treatment were compared to the overall value of salmonid production in Ireland, Scotland and Norway. The reported cost estimates will subsequently be used for modelling the relationship between costs and benefits under different hypothetical scenarios, which represent qualified judgments of the effects of sea lice on salmonid returns to the fjords and rivers, and the economic effects of this on tourism businesses and landowners (see paper 4 below).

Material and methods Data sources Ireland: We began by estimating the typical pattern and costs of sea lice treatment on salmon farms in Ireland. Data are not available in Ireland for total stocks in each month or for the exact national pattern of harvesting. The approach adopted was, therefore, to set out a simplified picture for 2004 of on-farm stocks, using assumed mortality, growth and biomass levels based on known patterns as described by local experts and industry participants. These data are then combined with information on chemotherapeutant and treatment costs to obtain an approximate estimate of the total national annual costs of treatment. It is emphasized that this approach provides a means for estimating overall national costs. It should not be interpreted as giving an empirically valid description of all aspects of the Irish salmon farming industry. We are grateful to the following who provided advice and data: Dr Hamish Rodger of Atlantic Veterinary Services Galway, Mr R. Flynn of the Irish Salmon Growers Association and Dr D. Jackson (SEARCH partner) of the Marine Institute.

Scotland: Detailed estimates for the overall costs of treatment in the Scottish industry were provided to us by Mr Chris Wallace (SEARCH partner) and his colleagues at Marine Harvest Ltd. As in Ireland, national data on the total monthly biomass in the sea are not available and the approach adopted was to build up estimates from informed assumptions about stocks, mortality, growth and biomass based on known patterns derived from detailed discussions with local experts and industry participants. These data are then combined with information on medicine and treatment costs to obtain the desired figures.

20

SUMBAWS

Final Report

Norway: Monthly data on the total biomass in the sea is available for Norway and this was utilized in our work. The data used for the calculations were obtained from the statistics of the Norwegian Seafood Federation (J. Grøttum, FHL), from the Norwegian Directorate for Fisheries (B. Halsteinsen & K. Johnsen, NDF), from recent literature (Øvretveit 2004, Josefsen 2004) and from aquaculture companies (Skretting & Ewos). The data used for the calculations comes from 2004.

Calculation of costs The calculation methods vary somewhat between the countries. This is mainly a result of varying availability and level of details of information and source data (see above). In general the calculations for Ireland and Scotland were performed in a somewhat simpler way than for Norway. However, when the same method as was used for Ireland and Scotland was applied to the data for Norway the results were similar to those presented in the report.

Costs for Ireland and Scotland The costs for Ireland and Scotland were estimated following a three-step modelling approach. First, the number of smolts needed to produce the actual amount of slaughtered biomass fish from one batch of smolts released into the sea was estimated from the assumed mortality over the period from release to slaughter (Ireland: 7 million smolts, Scotland: 39.5 million smolts). Secondly, the biomass, originating from one smolt batch, which was in the sea per month, was calculated on basis of standard growth and mortality models.

These models were based on country specific data.

In Ireland it was assumed that one batch of smolts was released per year, while it was assumed that two batches per year was put in the sea in Scotland. In addition, the fact that the biomass in the sea at any actual point of time originates from batches put in the sea over several years was taken into account, i.e. the total amount of salmon that is in the sea at a specific date might originate from one or two smolt releases in the same year and in addition from smolts put into the sea in earlier years. This is important since a marginal treatment has to be carried out for the entire biomass in the sea at a given time.

The assumed treatment strategies over a year (2004) also varied between the two countries. For Scotland it was assumed that two Excis (bath), two Salmosan (bath), and two SLICE® (oral) treatments were carried out in one year. For Ireland two SLICE® and one Excis was assumed to be carried out for smolt batch 1 (put into the sea in 2003) and one SLICE® treatment for smolt batch 2 (put into the sea in 2004). The costs for SLICE® treatment were calculated as cost of the required amount of SLICE® feed needed to complete a seven day course for the estimated biomass at the relevant times. The costs of bath treatment were the sum of the costs for chemicals, labour, oxygen and loss of biomass due to mortality and growth reduction following bath treatment. These costs were calculated as described in Table 1 using costs specific to each country. The calculations of increased mortality and growth reduction were based on information from Marine Harvest (C. Wallace, personal communication).

Table 1 Methods for calculation of costs associated with bath treatment Costs Chemicals Labour Oxygen Mortality Growth reduction

Formulas (Biomass / treated biomass per m3) × chemical costs per m3 treated Labour cost per treatment × number of treatments cost per treated tonne of fish Number of fish × 0.004 × average price per kg salmon Biomass × 0.00075 × average price per kg salmon

The cost of one additional treatment was calculated as the average of the assumed costs for all treatments in Scotland since all smolt batches were assumed to be treated simultaneously. In Ireland the cost of one additional treatment was calculated as the average of the treatment costs for smolt batch 1 plus the costs of one oral treatment for smolt batch 2. These costs were calculated in local currencies and the Scottish figure translated into Euros using the appropriate conversion ratio for 2004. The costs of one additional treatment were then expressed as the proportions of the value of the total production in each country and as the treatment cost per kg of fish produced.

21

SUMBAWS

Final Report

Costs for Norway The calculations for Norway are based on actual production data for different regions and months in 2004. These data include reported numbers and sizes of fish in the sea at the end of each month. Thus, the calculation method differs from the calculations for Ireland and Scotland since modelling of the monthly biomass and fish number was not required. In addition, oral treatment was assumed to be the only treatment used for fish ≤ 1 kg, while bath treatment only was applied for fish > 1 kg. As in the other countries local costs of chemicals, oxygen and labour were used in the calculations. Two types of medicines are used for bath treatment in Norway, Alpha Max™ and BetaMax™, and the average price of these was used in the calculations. Apart from this, costs for one single treatment were calculated in the same way as for Ireland and Scotland (Table 1). One difference was that the availability of monthly data on the biomass in the sea enabled calculation of the costs for each month during the year. The average cost of one additional treatment was thus calculated as the average of the monthly costs. Moreover, the treatment cost per kg slaughtered fish was estimated on the basis of the total reported numbers of fish in culture at the end of each month in 2004 for different regions. The total fish number was corrected for an assumed average survival from treatment to slaughter (91%) and the resulting number of fish was multiplied with the mean slaughter weight for each month. Mean weight at slaughter was defined on the basis of data for 2004. This enabled the cost of one additional treatment as a proportion of the value of the production and the cost per kg of fish produced to be estimated as for Scotland and Ireland.

Results Costs for each country The estimated costs for one additional treatment across all farms in Ireland would amount to about €0.38m. in 2004 (Table 2). The cost for Scotland was about €3.28m. (Table 3). Furthermore, the results showed that the total cost for one additional treatment for Norway as a whole was €11.65m. (Table 4).

Table 2 Costs of one additional treatment in Ireland in € m.

Treatment 2004 Chemicals Labour Oxygen Mortality Growth reduction ® SLICE , February, 0.22 Smolt batch 1 ® SLICE , September 0.12 Smolt batch 1 SLICE®, December 0.18 Smolt batch 2 Excis, May 0.09 0.06 0.04 0.03 0.06 Smolt batch 1 Cost for one marginal treatment as the average of batch 1 plus batch 2, €m.

Sum 0.22 0.12 0.18 0.28 0.38

Table 3 Costs of one additional treatment in Scotland in £ m.

Treatment 2004 SLICE®, February SLICE®, May Salmosan, July Excis, August Salmosan, September Excis, October Average costs in £m.

Average costs in €m.

Chemicals 2.19 2.58 0.72 1.61 0.64 1.43

Oxygen/ Labour

Mortality

Growth reduction

0.29 0.30 0.31 0.31

0.13 0.13 0.14 0.14

0.60 0.56 0.53 0.50

Sum 2.19 2.58 1.74 2.60 1.62 2.39 2.19 3.28

22

SUMBAWS

Final Report

Table 4 Costs of one additional treatment in Norway, amounts in m. NOK unless other is specified Treatment 2004 SLICE® costs, average for 12 months, fish ≤ 1 kg Bath costs, average for 12 months, fish > 1 kg Cost in m. NOK Costs in €m.

Chemicals

Labour

Oxygen

Mortality

Growth reduction

14.2 24.7

Sum 14.2

35.7

6.2

5.6

10.8

82.9 97.15 11.65

Comparison across the three countries Table 5 below shows the estimates of the cost of a marginal or extra treatment in each of the countries as derived from the available calculations. As can be seen, the cost of treatment varies with national variation in production and ranges from €11.65m. in Norway to €0.38m. in Ireland. When expressed as a proportion of the value of national output, the cost is highest in Norway and Scotland and lowest in Ireland. The cost of a marginal treatment varies from 0.36 % of the total slaughter value of the year’s production in Ireland to 0.46 % in Norway. The costs per kg produced fish are relatively low and vary from €0.011 in Ireland to €0.013 in Scotland and Norway. Thus, treatment in Ireland can be considered the cheapest on a per unit basis, while the costs for Norway and Scotland are slightly higher than Ireland.

Table 5 Summary of the costs of one additional treatment in each country

Norway Scotland Ireland

Currency Cost of a marginal Cost Total value of (m.) treatment in €m. production €m.

NOK 97.15 11.65 2513 £ 2.19 3.28 752 € 0.38 0.38 106

Costs as % of total production 0.46 0.44 0.36

Costs per kg produced fish € 0.013 0.013 0.011

Discussion Not surprisingly, the total estimated costs of one additional or marginal treatment across all salmonid fish farms correlated closely and positively with the production volume in the three countries, with Ireland having by far the lowest cost and Norway having the highest. The costs as a proportion of the production and as the actual cost per kg of fish produced were nevertheless similar, with the proportional cost being somewhat lower for Ireland than the two other countries. This implies that the two different approaches for modelling the costs of one additional treatment produced comparable results.

The results also showed that the cost of an intensified treatment per kg fish was low in all three countries.

However, since the profit margins in salmonid culture are relatively small, an increased cost in terms of one additional treatment against salmon lice might be significant for the profitability of salmonid aquaculture, especially in periods of low prices. Nevertheless, the stakeholders, according to the “polluter pays” principle, might argue that any sea lice infestation level harming wild fish should be treated, even if this is not “profitable”.

The cost estimates presented in this report are mainly designed to increase understanding of the cost range which would be associated with a more intensive and successful sea lice treatment on salmon farms.

As such, these estimates are likely to be maximum estimates of the costs of a better and more intensive sea lice treatment strategy. If an additional treatment were being carried out as a mitigative action in a realistic situation, there are several strategies by which it would be possible to reduce the cost connected with the treatment. First, the areas where salmonid farming is carried out do not always correspond with the most important areas for the wild Atlantic salmon migration. This is, for example, the case in Scotland where most of the salmon farming is carried out on the west coast, while the most important areas for the wild Atlantic salmon stocks are on the east coast. In such cases, an additional treatment might have only a limited effect on the wild stocks. Secondly, in cases where the farming areas and the important areas for wild stocks overlap the extent of an additional treatment might be reduced if the migration routes or the most used areas for wild salmonids are known. An additional treatment would obviously be of relatively low importance in farms that are located far away from these areas. Thus, an additional treatment might for instance be carried out only in farms placed in particularly vulnerable areas. Thirdly, some farm sites have a naturally low 23

SUMBAWS

Final Report

infestation level and an additional treatment might also have a low effect in such locations. This might, for instance, be the case in northern Norway, where problems with sea lice infestations are considerably less than further south in Norway, and in Ireland and Scotland. Finally, optimization of the present treatment schedules (and thus reductions in the current costs associated with this treatment) might also lead to a reduction in the infestation levels in the farms. It might actually be possible for an optimized treatment schedule to be more effective that an additional treatment. For instance, the efforts made in Norway to synchronize the treatments within regions have proved to be very successful. Moreover, improvements of the treatment methodology might also reduce the infestation levels substantially without increased costs.

The use of wrasse is, for instance, regarded as being a less costly, but still an effective, approach for reducing infestation levels compared to some chemical treatments in appropriate locations.

The cost estimates presented in this report are necessary inputs to the models that are developed in paper 4 below. This explores the relationship between the costs of a better treatment against salmon lice and the possible economic benefits, in terms of improvement of the recreational fishery, for the tourism businesses and the landowners. However, this cost/benefit analyses is greatly hampered by the lack of precise knowledge regarding the degree of improvement of the wild salmonid stock following an increased treatment. Even though it is very likely that an additional treatment against sea lice in the fish farms will have a positive effect on the infestation levels it will be quite speculative to predict how much it will reduce the infestation level and how much it will subsequently improve the wild stocks. Nevertheless, the reported cost estimates are useful in cost/benefit analyses under a range of hypothetical scenarios, which represent qualified judgments of the positive effect on the infestation levels in the farms and the consequent increase in salmonid returns to the fjords and rivers (see paper 4 below).

References Bjørn, P.A., Finstad, B. & Kristoffersen, R. (2001). Salmon lice infection of wild sea trout and Arctic char in marine and freshwaters: the effects of salmon farms. Aquaculture Research. 32, 947-962.

Butler, J.A. (2002). Wild salmonids and sea lice infestations on the west coast of Scotland: sources of infection and implications for the management of marine salmon farms. Pest Manage. Sci. 58, 595-608.

Gravil, H.R. (1996). Studies of the biology and ecology of the free swimming larval stages of Lepeophtheirus salmonis (Krøyer) and Caligus elongatus Nordmann 1832 (Copepoda: Caligidae) pp. 1-299.

PhD thesis, University of Stirling, Stirling, Scotland UK Heuch, P.A. & Mo, T.A. (2001). A model of salmon louse production in Norway: Effects of increasing salmon production and public management measures. Dis. Aquat. Org. 45, 145-152.

Heuch, P.A., Bjørn, P.A., Finstad, B., Holst, J.C., Asplin, L. & Nilsen, F. (2005). A review of the Norwegian national action plan against salmon lice on salmonids: the effect on wild salmonids. Aquaculture 246, 79-92.

Holst, J.C., Jacobsen, P., Nilsen, F., Holm, M., Asplin, L., & Aure, J. (2003). Mortality of seawardmigrating post-smolts of Atlantic salmon due to salmon lice infection in Norwegian salmon stocks. In Salmon at the Edge. (ed. by. D. M.s), pp. 136-137. Oxford: Blackwell Science.

Josefsen, E. (2004). Lakselus til besvær – en lønnsomhetsanalyse av ulike avlusningsmetoder på laks.

Fiskerikandidatoppgave, Norges Fiskerihøgskole, Universitetet i Tromsø, Tromsø, Norway. 87pp.

McKibben, M.A. & Hay, D.W. (2004). Distributions of planktonic sea lice larvae Lepeophtheirus salmonis in Loch Torridon, Western Scotland in relation to salmon farm production cycles. Aquaculture Research.

35, 742-750.

O’Donoghue, G., Costello, M. & Costello, J. (1998). Development of a management strategy for the reduction/elimination of sea lice larvae Lepeophtheirus salmonis parasites of salmon and trout. Marine Resource Series NO. 6, pp. 1-51. The Marine Institute, Dublin, Ireland Penston, M.J., McKibben, M.A., Hay, D.W. & Gilibrand, P. (2004). Observations on open-water densities of sea lice larvae in loch Shieldaig, Western Scotland. Aquaculture Research. 35, 793-805.

Tully, O. & Whelan, K.F. (1993). Production of nauplii of Lepeophtheirus salmonis (Krøyer) (Copepoda: Caligidae) from farmed and wild salmon and its relation to the infestation of wild sea trout (Salmo trutta L.) off the west coast of Ireland in 1991. Fish. Res. 17, 187-200. Øvretveit, S. (2004). Feed conversion ratio in practical Atlantic salmon (Salmo salar L.) farming. Master thesis. Department of Fisheries and Marine Biology, University of Bergen, Bergen, Norway. 90pp.

24

SUMBAWS

Final Report

Paper 3: CONSEQUENCES OF STOCK COLLAPSE AND COST OF PREVENTATIVE TREATMENT Ă˜. Aas, I. Unglem, J. Curtis, B. Dervo and B. Whelan (Manuscript) Abstract This paper focuses on what can be learned for the purposes of the present project from the collapse in sea trout populations in the West of Ireland since the late 1980s. It documents and quantifies the resulting decline in angler numbers and estimates the consequential drop in tourist expenditure in the region. By utilising some of the results and experience gained in Workpackage 7 of the SUMBAWS project, which is also based in the West of Ireland region, cost estimates for the preventative treatment of wild smolts are produced. The paper shows that the feasibility of widespread use of preventative treatment is questionable and that, even if feasible, very substantial costs would be involved.

Introduction A considerable literature now exists on the collapse of sea trout fisheries in the West of Ireland which occurred in the late 1980s and early 1990s and whose effects persist to the present. An important part of this literature has focused on the role of sea lice infestation in this collapse and on the origin of such infestation.

For the purposes of the present project, this literature can illustrate: (a) the economic consequences arising from a stock collapse in a migratory salmonid species which had been popular with anglers, and (b) the cost and feasibility of rehabilitating these fisheries, with a view to either restoring stocks to their pre-collapse levels or sustaining a minimal number of spawners sufficient to maintain the genetic integrity of the stock.

We begin with a brief general account of the sea trout collapse based mainly on Whelan and Poole (1996).

Irish west coast sea trout are relatively slow-growing and long-lived compared to salmon. They remain in freshwater for two to four years before migrating to the sea as smolts and often return three or more times to spawn in freshwater. Over the last 30 or 40 years there has been a slow secular decline in abundance which has been attributed to a variety of factors including poaching, excessive commercial exploitation, field and arterial drainage and subsequent maintenance, fertilisation of the surrounding land, afforestation and erosion due to overgrazing by sheep. A sharper and more serious decline has been documented beginning in 1986 which had by 1989 resulted in a population collapse in most fisheries between Galway Bay in the south to Clew Bay in the north (Whelan 1991, 1992, 1993). The precipitate and enduring nature of the collapse can be illustrated by the comprehensive data from the Burrishoole system which is systematically monitored as an index river system (see Table 1 and Figure 1 from Poole et al. 2004). Broadly similar patterns of escapement prevail in most of the other systems in the affected area (Anon 2002).

Table 1 Mean upstream migration in five year periods for silvered and unsilvered trout, including the proportion of 0+ sea age and the spawning escapement (Poole et al (2004)) Migration Years 1970-1974 1975-1979 1980-1984 1985-1989 1990-1994 1995-1999 2000-2003

Total Silver 2130 2624 1719 978 206 177 98

No. of Finnock 1065 868 740 455 124 111 57

%0+ Sea Age Unsilvered 50 31.6 43.5 42.6 60 74 62 112 63.5 56

%0+ Sea Age

73.4 75.3 71.6

Total Spawning Migration Escapement 2130 1812 2624 2369 1719 1622 978 906 280 279 289 289 155 155

25

SUMBAWS

Final Report

4000

Unsilver

Total Upstream Count

3500

Silver

3000 2500 2000 1500 1000 500 2003

2000

1997

1994

1991

1988

1985

1982

1979

1976

1973

1970

0

Year

Figure 1 Annual numbers of upstream migrating sea trout through the Burrishoole fish traps for 1970-2003, showing silvered and unsilvered migrants separately since 1990 (Poole et al 2004) A number of official and unofficial bodies and groups investigated the collapse including the Sea Trout Action Group (Anon 1992) and the Sea Trout Working Group (Anon 1991-1994). As many as eighteen different hypotheses were advanced and investigated to account for the collapse, including an outbreak of an unknown disease, loss of fodder fish such as sand eels, increased predation by sea fish such as pollack on concentrations of sea trout in the vicinity of salmon cages, acidification etc. Some of these factors appeared likely to have had a gradual long term negative effect on sea trout stocks but most could not account for the sudden very substantial collapse in the late 1980s. Whelan and Poole (1996) conclude that “the only consistent factors, therefore, to emerge from the research carried out to date are the premature return of both smolts and kelts to estuaries in late May and the presence of abnormal numbers of juvenile lice�. Detailed further research on the sea lice (Tully and Whelan 1993; Tully et al 1993a and Tully et al 1993b) showed that: (a) in the region studied some 95% of the total nauplius larval production of Lepeophtheirus salmonis was from farmed salmon; (b) that the morphological and physiological impact of the lice on the sea trout was significant and sufficient to cause mortality; (c) that, in a number of cases, lice infestation levels rose and fell in response to the presence or absence of salmon cages.

On the basis of this evidence, the hypothesis that sea lice infestation attributable to salmon farming played a key role in the sea trout collapse appears plausible.

Materials and Methods The material presented in this paper is derived from two main sources: (a) a study by Fingleton (1990) based on a sample survey of accommodation owners in the Connemara (west Galway/Mayo) region which attempted to quantify the scale of business lost due to the collapse in sea stocks and (b) information based on personal communications from our partner the Irish Central Fisheries Board in connection with WP 7 of the SUMBAWS project.

Results The Economic Effects of the Collapse The sample used by Fingleton (1990) was very small and various assumptions and simplifications were needed in order to arrive at an estimate of the loss; but the results are of interest since we are not aware of any comparable studies in the literature. The figures, though necessarily crude and tentative, illustrate how angler numbers react to a serious decline in stocks and indicate the type of decline in business experienced by the region in question. Fingleton’s principal results are shown in Table 2. He estimates that demand for bednights by anglers has fallen by just over 50% , with a fall of the same magnitude in total expenditure.

26

SUMBAWS

Final Report

Updating the value of the loss to current (2004) values by the Consumer Price Index, the loss of IR£ 721 000 in 1990 would amount to €1.4m. in present day values.

Table 2 Estimated Angler Bednights and Expenditure in the Connemara Region before and after the Sea Trout Collapse No. of hotel/guesthouse angling bednights Origin of anglers Hotel/guesthouse bednights by origin Percentage of anglers using above accommodation type No. of anglers bednights at all accommodation types No. of anglers bednights from all origins at all accommodation types Average expenditure per bednight (IR£) Total expenditure, all origins (IR£)

Prior to 1987 11 778 Irish

Foreign

In 1990 5 498 Irish

Foreign Irish Foreign

34% 66% 4 005 7 773 37.5% 66.1% 10 680 11 759

34% 66% 1 869 3 629 37.5% 66.1% 4 984 5 490

22 439

10 474

38.76

79.74

1 352 000

38.76

Loss 6 280 34% 2 135 37.5% 5 693

66% 4 145 66.1% 6 271

11 964

79.74 38.76 79.74

631 000

721 000

The Cost and Feasibility of Stock Restoration As stated above, this section relies heavily on the work of our partner the Irish Central Fisheries Board, to whom we express our sincere thanks, in connection with WP 7 of the SUMBAWS project. The details of how smolt capture and treatment is organised are given by the CFB in their description of WP 7. Here we concentrate on the cost and feasibility issues in order to reach a preliminary judgement about the practicality and economics of using a chemotherapeutant such as Substance EX in the treatment of wild fish.

A key issue is the definition of the objective. If one wishes to treat all or most of the descending smolts (and kelts), then large fixed traps will be required and these may not be feasible at all in larger systems. If the objective is just to treat a proportion of the run (say a number large enough to maintain the genetic integrity of the stock) then it may be feasible to use devices such as rotary screw traps which would have a lower capital cost. However, the numbers of fish involved would not be sufficient to restore the stock to a level that could provide attractive angling.

Consideration also has to be given to the likely cost in terms of mortality from trapping and treating. In heavy floods it is probable that some of the smolts could get crushed in the traps and all handling of wild fish for treatment purposes engenders risks. Estimates of the likely mortality rate are not available.

Costs will also be affected by whether one uses a food-based treatment like SLICE® or a 30 minutes bath treatment like Substance EX. The former involves retaining the smolts in captivity for longer and ensuring that they habituate to feeding in this environment. This will make the feed-based option much more expensive.

We attempted to make a minimum cost estimate by assuming that a fixed trap was feasible (i.e. that it was a small system) and that a bath treatment such as Substance EX could be used. It was further assumed that the objective is only to treat, so that tagging and re-capture of the returning smolts are not required. As shown in Table 3 below, the total cost, making these minimising assumptions, is estimated at about €27 000 per trap (river) in year 1 and €16 800 per annum thereafter (ignoring depreciation/maintenance costs).

Table 3 Estimated cost per Trap per annum (Assumes: no tagging, small system, bath treatment Capital cost of trap Tanks, pumps, oxygen, nets, medicaments Mileage = 300 per week @ €1 per mile Staff = 1 person per trap @ €800 per week Total Cost in year 1 Annual operating costs thereafter

Cost [€] 10 000 2 500 3 900 10 400 26 800 16 800

If it was desired to treat all the sea trout rivers in Connemara (the region worst affected by the collapse) then the list would be as follows: 27

SUMBAWS

Final Report

Large Systems Costello Screebe Invermore Gowla Ballynahinch Clifden Dawros Erriff Delphi Carrowniskey Bunowen Newport Burrishoole

Smaller systems Furnace Lettermucka Inverbeg Carna Ballinaboy Doohulla Owengarve Belclare Crumlin Culfin Ardbear

Even if it were feasible to operate fixed traps on the larger systems (which is quite doubtful), the total annual cost of treating all these systems would be in the order of €0.64m. (24 systems @ €26.8 K per system in year 1 and €0.40m thereafter). This compares with an estimated total annual expenditure by visiting (i.e. non-local) sea trout anglers in pre-collapse conditions (at present day money values) of €2.8 m. (=IR£1.35m from Table 2 above, increased by the growth in the Consumer Price index between 1987 and 2004, and expressed in €) Discussion The experience of the sea trout collapse in the West of Ireland shows that angling activity, angler numbers and spending by tourist anglers all are highly responsive to a sustained deterioration in the probability of success. The estimated economic effects given in Table 2 are based only on expenditure by non-local anglers, i.e. those who, in the main, stay overnight in the region when angling. They do not include any provision for reduction in expenditure by local anglers or for any increased cost incurred by these anglers by having to switch species or travel further for their fishing. It is, however, likely that these impacts would be less than those caused by the decline in angling by non-locals.

It must also be borne in mind that the approach adopted, as discussed in paper 1 above, is an economic impact approach. Thus, it excludes any estimate of the existence, option or bequest values of the sea trout stocks. No data are available to estimate these types of value but they are undoubtedly positive and could be substantial.

From an economic point of view, two major issues arise in relation to the preventative treatment of wild smolts. First, the practicality of such treatment is in doubt. As described above, the feasibility of treatment depends on the size of the system, the type of trap, the prevailing water conditions and numerous other factors.

Mortalities are likely to occur and could be serious, especially in a heavily depleted stock where each smolt is of critical importance to the survival and maintenance of the genetic integrity of the stock. Secondly, even if trapping for treatment is judged feasible, substantial costs would be involved.

References Anon. (1992) Sea Trout Action Group Report, 1991. Sea Trout News. 3 February 1992.

Anon. (1991-94) Reports of the Sea Trout Working Group. Fisheries Research Centre. Department of the Marine, Dublin Anon. (2002) Report of the Sea Trout Working Group 2002. Fisheries Research Centre. Department of the Marine, Dublin Fingleton, P. (1990) Declining Sea Trout Runs in Western Ireland: Some Tourism Effects. The Sea Trout Action Group Poole, W.R., M. Dillane, E. DeEyto, G. Rogan, P. McGinnity and K. Whelan (2004) Characteristics of the Burrishoole sea trout population: census, marine survival, enhancement and stock recruitment, 19712003. Paper presented to the 1st International Symposium on the Conservation and Management of Sea Trout. Cardiff, July, 2004.

28

SUMBAWS

Final Report

Tully O. and K.F. Whelan (1993) Production of nauplii of Lepeophtheirus salmonis (Kroyer) (Copepoda: Caligidae) from farmed and wild salmon and its relation to the infestation of wild sea trout (Salmo trutta L.) off the west coast of Ireland in 1991. Fisheries Research 17, 187-200 Tully, O., Poole, W. R., and Whelan K. F. (1993a) Infestation Parameters for Lepeophtheirus salmonis (Kroyer) copepoda:caligidae) Parasitic on Sea Trout (Salmo trutta L.) Post Smolts on the West Coast of Ireland during 1990 and 1991. Aquaculture and Fisheries Management. 24, 545-557.

Tully, O., Poole, W. R., Whelan, K. F., and Merigoux, S. (1993b) Parameters and Possible Causes of Epizootics of Lepeophtheirus salmonis (Kroyer) Parasitic on Sea Trout (Salmo trutta L.) on the West Coast of Ireland. Proceedings of the First European Crustacea Conference. Paris, August 31 – September 5, 1992 Whelan, K.F. (1991) Disappearing sea trout - decline or collapse? The Salmon Net 23, 24-31.

Whelan, K.F. (1992) Management of salmon and sea trout stocks. In: Environment and Development in Ireland. Proceedings of a conference held at University College Dublin. The Environmental Institute, University college Dublin. 457-466.

Whelan, K.F. (1993) Historic overview of the sea trout collapse in the west of Ireland. In: Aquaculture in Ireland - towards sustainability. Ed. J. Meldon. Proceedings of a Conference held at Furbo, Co. Galway.

30th April - 1 May. 1993. 51-53. An Taisce, Dublin.

Whelan, K.F. and W.R. Poole (1996) The Sea Trout Collapse 1989-92. in J.D. Reynolds (ed.) The Conservation of Aquatic Systems 101-110 Royal Irish Academy, Dublin

29

SUMBAWS

Final Report

Paper 4: MODELLING THE ECONOMIC EFFECTS OF SEA LICE TREATMENT Ø. Aas, I. Unglem, J. Curtis, B. Dervo and B. Whelan (Manuscript) Abstract This paper develops a computer model of the key stages in the assumed causal chain between enhanced lice treatment, through better smolt survival and improved returns to the coast to economic and social benefits in the form of greater income and employment in the commercial and angling sectors. The model incorporates data from the earlier papers on economic impact and the marginal costs of an extra nationwide treatment in each country. The main output from the model is a series of Cost/Benefit ratios aimed at summarising the economic impacts from treatments with varying levels of effectiveness. A number of plausible scenarios are simulated to illustrate how these ratios vary with changes in the assumptions.

Introduction We now attempt to integrate all of material developed in papers 1 to 3 by developing a simple computer model of the lice treatment process and its economic effects in the three countries, in terms of Costs and Benefits (C/B). The conceptual foundation for the model is based in knowledge, assumptions and hypotheses established in the overall SUMBAWS project, supplemented by general literature on linkages between wild and farmed fish stocks, recreational fisheries, and how they are each valued. The model is built up in stages starting with money spent on sea lice treatment, leading though the hypothesised improvements in smolt survival, likely improved adult returns, improved catches and, from this, increased economic impacts in the wild fish businesses. The strength of the computer model is that it can be used to simulate the effects of different assumptions about the magnitudes of the variables in question and to derive estimates of economic benefits and how these benefits relate to the costs of treatment. Given the uncertainties about the magnitudes of the variables used, the model can only provide indicative answers about the consequences of the assumed changes. However, we believe that it does capture some of the key relationships and that its results will be of use as a guide to policy. A number of plausible sets of assumptions are examined. A simulation is also carried out of a regional situation where wild rather than farmed fish are treated.

The cost data used refer to our illustrative case of a nationwide, marginal treatment. As was emphasised in paper 2 above, this additional, or marginal, treatment is envisaged as acting as a precautionary action to supplement the existing treatment schedule. We do not have evidence that a treatment regime of this kind would be the most appropriate or effective in practice. However, for illustrative purposes, this treatment regime is utilised in order to operationalise our model.

Data and Methods The best way of explaining how the model is derived and how it can be used is to work through an example. Figure 1 and Table 1 show, diagrammatically and arithmetically, a realisation of the model for a particular set of assumptions. The first line in Figure 1 shows the cost, in million euro, of a marginal lice treatment in each country. These figures are derived from Tables 2-4 (Paper 2) above. The following lines show the effects of different factors on smolt survival. These are: • The (smolt) “survival improvement factor” i.e. the percentage change (increase) in the number of smolts surviving past the coast after treatment. (The biological dimensions of the overall SUMBAWS project should shed some light on this).

• The “smolt/adult conversion factor” i.e. a factor allowing for the possibility that the extra surviving smolts have a different survival rate to adult than the pre-treatment fish. In most cases, we have assumed this =1.0 • The “commercial catch improvement factor” i.e. the ratio of the change (increase) in the commercial catch to the change in the number of returning adult fish. A value of 1 indicates that the commercial catch increases proportionately with the increase in the adult returns. If, say, the commercial catch were limited by quota to its pre-treatment level then this factor would be set to zero.

30

SUMBAWS

•

•

Final Report

The “angling catch improvement factor” i.e. the ratio of the change (increase) in the angling catch to the change in the number of returning adult fish. A value of 1 indicates that the angling catch increases proportionately with the increase in the adult returns. If, say, anglers catches rose less than proportionately than to the size of the run because of the timing of spates, then this factor would be less than one.

The “angler response factor”. This factor is the “elasticity of angler expenditure with respect to catch” i.e. the proportionate increase in angling expenditure due to a unit increase in overall angling catch. It summarises the combined effect of a number of processes. The basic rationale is that salmon runs increase, angling success improves, angling activity increases (days fished per angler rise and/or the number of anglers increases) and expenditure rises. Note that there is no necessary proportionality between these processes. Because more fish are present it does not mean that expenditure will rise proportionately.

To illustrate how the model works, let us consider the first column of numbers on Table 1 which refers to Norway. Table 4 in paper 2 above showed that the cost of a marginal treatment in Norway was €11.65m.

The assumed survival improvement factor is 1.1, giving a 10% increase in smolt numbers. The smolt/adult conversion factor is 1 i.e. the extra surviving smolts have the same probability of returning as the pretreatment smolts once they leave the coast. The commercial and angling catch factors are each assumed to be 1 i.e. both angling and commercial catch rise proportionately with adult returns. Angler response is assumed to be 0.5 because of a general assumption of diminishing returns and the likelihood that as catch per angler rises, expenditure per angler does not rise proportionately.

The combined effect of all these factors is applied to the current commercial catch (€3m) and to the current value of angling expenditure (€160m.) giving increased value of €0.30m and €8m respectively. The “total economic benefit” is the sum of these two quantities €8.30m. The Cost/Benefit ratio is then about 1.40, showing that the increase in costs exceeds the increased value of the increase in economic activity.

Note that the C/B ratios are below one for both Scotland and Ireland. In Scotland, costs are about 60% of benefits, while in Ireland the C/B ratio is 0.08, indicating that benefits dramatically exceed costs by a factor of about twelve to one.

Amount spent

Spend on lice treatment

Survival Improvement factor

Smolt/adult conversion factor

Improved smolt survival

Improved adult returns Improved commercial catch

Catch Improvement factors

Improved angling catches Value of Income: Commercial Catch

Angler response factor

Improved angler expenditure

Increased economic benefits

Figure 1 Model for economic effects of sea lice treatment 31

SUMBAWS

Final Report

Table 1 Economic Effects of Improved Salmon Survival from Lice Treatment (Survival Improvement =10%, Angler Response=0.5, All other factors =1) Modelling Factors Survival Improvement Change in smolt/adult conversion Commercial catch improvement Angling catch improvement Angler response

(= ratio of improved smolt survival to pre-treatment survival) (=ratio of improved adult survival to pre-treatment adult survival) (=ratio of improved commercial catch to pre-treatment commercial catch) (=ratio of improved angler catch to pre-treatment angler catch) (= elasticity of angler expenditure with respect to angling catch)

Spend on lice treatment Improved smolt survival Improved adult return Improved commercial catch Improved angling catches Value of income for commercial catches Improved angler expenditure Increased economic benefits C/B Ratio Total current value of commercial catch Total current value of angling expenditure

Units €m. % change % change % change % change €m. €m. €m. €m. €m.

Norway 11.65 10% 10% 10% 10% 0.30 8.00 8.30 1.40 3.0 160.0

Scotland 4.49 10% 10% 10% 10% 0.06 5.65 5.71 0.79 0.6 113.0

Ireland 0.38 10% 10% 10% 10% 0.49 4.00 4.49 0.08 4.9 80.0

Results: Other Scenarios Tables 2-5 illustrate what happens when one changes the assumptions. Table 2 shows the implications of an even lower smolt survival, 5% improvement in smolt survival. In this case, costs exceed benefits in both Norway and Scotland while benefits still exceed costs in Ireland. Table 3 illustrates the effect of a 20% improvement in survival - here; benefits clearly exceed costs in all three countries. Table 4 shows a negative situation: 5% survival combined with poor angling catch improvement and the angler response factor set at 0.5. Here costs exceed benefits in both Norway and Scotland but benefits still exceed costs in Ireland. On reflection, it is clear that these general patterns are influenced by a number of considerations: (a) the large size of the Norwegian fish farming industry and thus the high cost of treatment (b) the high expenditure per fish caught by anglers in Ireland and (c) the low cost of treatment in Ireland (primarily due to the relatively small size of the Irish fish-farming industry). A reasonable inference, especially for Norway, would be that focussing treatment on areas with a particularly good return (plentiful runs of salmon or low costs of treatment) would be a better strategy than a nationwide treatment.

Table 5 shows a realisation of the model relating to the treatment of wild sea trout in the Connemara area as discussed in stage (c) above. Assuming a 10% overall improvement in survival and an angler response of 0.5, it shows a Cost/Benefit ratio of 2.9 i.e. costs exceed benefits by almost three to one. This emphasises the high costs and limited feasibility of trapping and treating wild smolts.

32

SUMBAWS

Final Report

Table 2 Economic Effects of Improved Salmon Survival from Lice Treatment (Survival Improvement =5%, Angler Response=0.5, All other factors=1) Spend on lice treatment Improved smolt survival Improved adult return Improved commercial catch Improved angling catches Value of income for commercial catches Improved angler expenditure Increased economic benefits C/B Ratio Total current value of commercial catch Total current value of angling expenditure

Units €m % change % change % change % change €m. €m. €m. €m. €m.

Norway 11.65 5% 5% 5% 5% 0.15 4.00 4.15 2.81 3.0 160.0

Scotland 4.49 5% 5% 5% 5% 0.03 2.83 2.86 1.57 0.6 113.0

Ireland 0.38 5% 5% 5% 5% 0.25 2.00 2.25 0.17 4.9 80.0

Scotland 4.49 20% 20% 20% 20% 0.12 11.30 11.42 0.39 0.6 113.0

Ireland 0.38 20% 20% 20% 20% 0.98 8.00 8.98 0.04 4.9 80.0

Table 3 Economic Effects of Improved Salmon Survival from Lice Treatment (Survival Improvement =20%, Angler Response=0.5, All other factors=1) Spend on lice treatment Improved smolt survival Improved adult return Improved commercial catch Improved angling catches Value of income for commercial catches Improved angler expenditure Increased economic benefits C/B Ratio Total current value of commercial catch Total current value of angling expenditure

Units €m. % change % change % change % change €m. €m. €m. €m. €m.

Norway 11.65 20% 20% 20% 20% 0.60 16.00 16.60 0.70 3.0 160.0

Table 4 Economic Effects of Improved Salmon Survival from Lice Treatment (Survival Improvement =5%, Angler Response=0.5, Angling Catch Improvement =0.5, All other factors=1) Spend on lice treatment Improved smolt survival Improved adult return Improved commercial catch Improved angling catches Value of income for commercial catches Improved angler expenditure Increased economic benefits C/B Ratio Total current value of commercial catch Total current value of angling expenditure

Units €m. % change % change % change % change €m. €m. €m. €m. €m.

Norway 11.65 5% 5% 5% 3% 0.15 2.00 2.15 5.42 3.0 160.0

Scotland 4.49 5% 5% 5% 3% 0.03 1.41 1.44 3.11 0.6 113.0

Ireland 0.38 5% 5% 5% 3% 0.25 1.00 1.25 0.31 4.9 80.0

33

SUMBAWS

Final Report

Table 5 Economic Effects of Improved Salmon Survival from Lice Treatment (Survival Improvement =10%, Angler Response=0.5, All other factors=1) Spend on lice treatment Improved smolt survival Improved adult return Improved commercial catch Improved angling catches Value of income for commercial catches Improved angler expenditure Increased economic benefits C/B Ratio Total current value of commercial catch Total current value of angling expenditure

Units €m. % change % change % change % change €m. €m. €m. €m. €m.

Wild Sea Trout in Connemara 0.4 10% 10% 10% 10% 0.00 0.14 0.14 2.86 0.0 2.8

Discussion The overall purpose of the modelling exercise is to illustrate how investments in sea lice treatment (either on farmed or wild fish) can lead to improved revenues to the wild fish sector, primarily in the recreational tourism fishing. The model is a simplified version of the assumed chain of causation linking sea lice on salmon farms with the effects on wild salmonids and the businesses dependent on them.

However, central elements in the model remain uncertain and speculative. First, while there is general acceptance that sea lice levels and sea lice treatment strategies in farms or on wild fish have the potential to influence wild fish returns, the current state of knowledge does not permit us to establish a firm “dose/response” relationship between the level or type of treatment and a quantified increase in returns to the coast. Furthermore, the assumption that improved smolt survival due to sea lice treatments will provide improved adult returns is complicated. It is well established in the literature that a range of factors affect salmon growth and survival at sea (such as availability of fodder fish, variations in predator populations etc.) and these factors might at times mask the effect of improved sea lice treatment.

While the assumption that increased adult returns normally increase catches for both the commercial and angling sectors is well established and documented, the assumption that increased catches increase expenditures may be more questionable. Furthermore, the magnitude of the increase in expenditure in response to any given improvement in (expected) catch needs to be estimated. There is, however, some evidence that this response can be sizable in the case of salmon angling. (Bell 1989). It is quite possible that this relationship is non-linear and subject to threshold effects (i.e. the increase needs to be above a certain absolute size to evoke any response) or satiation/saturation effects (e.g. due to good angling spaces/pools being limited even if runs improve). More research is needed to elucidate these issues.

The limitations of the “economic impact” approach on which all the data and modelling is based must also be borne in mind. On the one hand, as explained in paper 1, this approach excludes “non-consumptive” values such as existence, option or bequest values (Pollock et al. 1994). On the other hand, the figures presented in Tables 1 to 5 are based on total angling expenditure and the value of total commercial catch in each country. They do not make allowance for multiplier effects or for displacement effects in the case of angling and operating costs in the case of commercial fishing. If allowance could be made for all these factors, the Cost/Benefit ratios would change, possibly to a substantial extent.

The main value of the modelling approach is that it clearly highlights what factors are most important for the Cost/Benefit ratio and how they interact. Thus, it clear that the cost of marginal treatment depends crucially on the size of the fish farming sector. This means, for instance, that the cost in Norway for a nationwide treatment is very high. It would be very important in this case to target and limit the intervention to particular regions or systems where the benefits are expected to be especially large such as key angling systems or areas where lice infection is especially high. The models also emphasise what types of benefit are likely to arise and to whom they will accrue. Thus, in all three countries the largest benefits from improved runs are likely to the angling sector rather than to the commercial fishing operations, due to the relatively small size of the latter. The modelling approach also indicates the key areas on which future economic research should focus. Improved estimates of angler response were mentioned above. It would also be important to obtain quantified biological information on the likely response in terms of improved runs to 34

SUMBAWS

Final Report

improvements in treatment and to improve our estimates of the costs involved on a national, or perhaps even more importantly, a regional or river system basis.

References Bell, F. W. (1989) The Main Quarry Hypothesis and Salmon Angling. Journal of Marine Resource Economics, pp. 71-82.

Pollock K., C.M. Jones and T.L. Brown (1994) Angler Survey Methods and Their Application in Fisheries Management. American Fisheries Society Special Publication 25, Bethesda, Maryland

35

SUMBAWS

Final Report

Workpackage 2 Lead partner: NINA Participating partners: NIFA, NCFS OBJECTIVES • • •

Identify migratory speed and behaviour of wild and hatchery reared salmon and sea trout smolts through a fjord system. We will use a Norwegian model fjord system (Eresfjord in Møre and Romsdal) as a model system for collecting detailed information using acoustic telemetry.

Analyze migration data in relation to environmental conditions associated with migratory behaviour, such as water temperature, water salinity, water depth and distance from shore and habitat type.

Analyse the effect of salmon lice infestation on behaviour, migratory speed and habitat utilization.

DELIVERABLES • • •

The routes, speed, behaviour and habitat preferences of wild and hatchery-reared Atlantic salmon and sea trout smolts, during their fjordic migration, will be ascertained and linked to coastal environmental conditions.

The migratory differences between the species and between wild and hatchery-reared fish will be identified.

The effect of sea lice (by means of deploying Substance EX or artificially high salmon lice infection) on migration and behaviour patterns of sea trout and Atlantic salmon post-smolts will be identified.

EXECUTIVE SUMMARY The lack of information about migratory routes and geographical distribution of Atlantic salmon (Salmo salar L.) and sea trout (Salmo trutta L.) post-smolts in fjords and near-shore areas is particularly critical, because the heaviest mortality of salmon and sea trout in the sea apparently occurs during the first months in seawater. The increasing salmon lice production in farmed fish has been suggested to contribute to unnaturally high infestations on wild fish. This undesirable situation has been implicated in the collapse of many sea trout and salmon stocks in fjords and coastal areas with heavy fish farming activity. However, the proportion of total mortality due to salmon lice versus predation is not known. Thus, to improve our knowledge of mortality in the first marine stage of salmonid smolts, we developed and used telemetry techniques to monitor fish movements in the marine environment with different salmon lice infestation pressures. This information is used in a developed model for simulating the risk of salmon lice infestation in different areas used by Atlantic salmon and sea trout post-smolts. By using this model we can better evaluate where and when the observed infestation is acceptable or not.

Hatchery-reared Atlantic salmon and wild sea trout post-smolts were tracked during active migratory swimming in the fjord. The observed direction of movement was dependent on the actual movement of the fish and not the water current. Water currents were not systematically used as an orientation cue, either in Atlantic salmon or sea trout post-smolts. The actual movements of sea trout were in all compass directions, with no prevailing or systematic patterns. Atlantic salmon also moved in all compass directions, albeit with the lowest frequency of actual movement towards the fjord (Paper 1). The objective of Paper 6 was for the first time to record actual swimming speeds and directions of wild Atlantic salmon post-smolts in a Norwegian fjord system, and to compare results with previous studies of hatchery-reared post-smolts. The swimming speeds and orientation of the wild post-smolts resembled previous results from hatchery-reared post-smolts.

Rate of progression through the fjord by hatchery-reared Atlantic salmon and wild sea trout smolts differed during their outward migration at 9, 32, 48 and 77 km from the outlet of the home river. Seventeen 36

SUMBAWS

Final Report

salmon (68%) and 8 trout (53%) were recorded after release. Sea trout post-smolts had a slower rate of progression between different receiver sites than did Atlantic salmon post-smolts. Atlantic salmon were recorded 9, 48 and 77 km from the river mouth on average 28, 65 and 83 hours after release, respectively.

Sea trout were recorded 9 km from the release site 438 hours after release. Only four (23%) sea trout were detected in the outer part of the fjord system, whereas the remainder of the tagged sea trout stayed in the inner fjord system near the home river outlet. Atlantic salmon post-smolts stayed longer in the inner part than the outer part of the fjord system but, distinct from sea trout, migrated through the whole fjord system into the ocean (Paper 2). The migration of Atlantic salmon post-smolts was a more direct migration than was the migration of sea trout.

Sublethal effects due to salmon lice infestation, such as reduced growth and swimming performance, must impair fish from reaching off shore feeding areas, finding food and avoiding predators. Thus, we examined the effects of salmon lice infestation and salmon lice protection on survival and rate of progression of hatchery-reared Atlantic salmon and sea trout post-smolts during the first stage immediately after the start of the marine migration. The infested groups were artificially exposed to infective salmon lice larvae in the laboratory and released in the inner part of the fjord to simulate a naturally high infection pressure near the home river outlet. Neither salmon lice infection nor pharmaceutical prophylaxis had any effects on survival and rate of progression of migrating Atlantic salmon post-smolts compared to control fish. Hatchery-reared Atlantic salmon post-smolts spent on average 6.3 days (maximum 15.2 days) in migrating through the 80 km fjord system and, thus entered the ocean when the more pathogenic pre-adult and adult (“mobile”) lice stages of the infestation had developed. The hatchery-reared sea trout post-smolts remained in the inner part of the fjord system to a larger extent than did hatchery-reared Atlantic salmon post-smolts. No effect of salmon lice infestation, or protection, was found in sea trout post-smolts during the first weeks of their fjord migration. First time migrating sea trout post-smolts will to a larger extent than first time migrating Atlantic salmon post-smolt stay in the fjord areas when salmon lice infections reach the more pathogenic “mobile” parasite stages. Contrary to Atlantic salmon, sea trout will therefore possess the practical capability of returning to freshwater when encountering severe salmon lice attacks because these parasites cannot survive in fresh water (Paper 3).

There was no difference in survival between wild and hatchery-reared Atlantic salmon post-smolts from release in the river mouth to passing receiver sites 9 km and 37 km from the release site. The mortality approached 65 % during the first 37 km of the marine migration for both groups. The mortality of these fish is suggested to be attributed to reasons other than salmon lice infestation alone. There was no difference between wild and hatchery-reared salmon either in time from release near the home river outlet to first recording at 9.5 km, or in the rate of progression. At 37 km, the hatchery-reared salmon were recorded sooner after release than the wild salmon, but rate of progression in terms of body lengths per second did not differ. The wild sea trout post-smolts remained longer in the inner part of the fjord system, with much slower rates of progression during the first 9 km. The wild and hatchery-reared salmon post-smolts passed 9 km from the home river mouth in the middle of the fjord as well as closer to the shore; by contrast the wild sea trout post-smolts remained closer to the shore than the middle of the fjord. Combining data for all groups of Atlantic salmon and sea trout post-smolts showed a tendency for more post-smolts to pass on one side of the fjord than on the other. Also 37 km from the release site, the salmon passed equally along the middle of the fjord as the shore, whereas sea trout remained rather closer to shore than the middle. There was no indication that the fish passed more often on one side or other of the fjord, except that there were few recordings on the receiver closest to shore on the northern side of the fjord. Most salmon smolts followed a straight route out of the fjord, although some fish were recorded migrating back and forth between 9 and 37 37

SUMBAWS

Final Report

km from the home river (Paper 4).

The predation on seaward migrating salmonid smolts can be a major regulating factor for local populations of salmon and sea trout. In freshwater, and especially in the transition zone from river to the sea, avian predators have been reported to be of especial importance. During the study of the behavior and migration of salmon smolts, several of the tagged smolts were observed to be eaten by gulls or predatory fish within a few hours of descent. Through visual observations and stomach analyses, predation by saithe (Pollachius virens L.) and cod (Gadus morhua L.) on both Atlantic salmon and anadromous sea trout was recorded. Indeed, the results indicate that some individual saithe may have specialized in preying on smolts at this time of the year. Wild and hatchery-reared smolts were equally valnerable to these predators (Paper 5).

Results from the study investigating the temporal and spatial distribution of Atlantic salmon and sea trout post-smolt in a Norwegian fjord system were used for developing a multivariate model for predicting the pattern of the salmon lice infestations on wild salmonids in different parts of the fjord (Paper 7). The model can be used to evaluate if observed infestation levels in fjord systems are acceptable or can be considered safe: i.e., if the infestation levels are harmful or not for wild salmonids. Furthermore, the model also enables prediction of the risks and levels of smolts being infested by salmon lice at specific sites in a fjord. Hence, the model can be used to predict effects of establishing fish farms at new sites in fjords and coastal areas by identifying areas where the negative effects on wild salmonid populations will be minimal. Based on data from 2002, 2003 and 2004 for Atlantic salmon post-smolts from the River Eira, the model predicted low salmon lice infestation levels and no mortality during the period in fjordic waters. The majority of infestations were predicted to occur in the outer part of the fjord system. When applied to data for sea trout in the same area, the model predicted a lice infestation level that would cause harmful physiological stress for a low proportion of the sea trout utilizing the entire fjord system, whereas smolts that remained in the inner part of the fjord were predicted never to experience harmful infestation levels. However, the infestation simulated on sea trout is likely to be an underestimate because the infestation levels were based on data from the spring. Normally, infestation increases during summer and simulations for sea trout will be repeated with increasing infestation levels until the simulated data reflect the recorded lice counts from gillnet fisheries.

CONCLUSIONS All objectives, deliverables and milestones have been met for WP2. Public management decisions, such as where to place fish farms and the level of acceptable fish farming intensity, need to be based on scientific appraisals of the relations between wild and farmed salmonids. The extensive data generated by the project will be integrated into the developed model to predict the risk of infestation from sea lice. The goal here is therefore to build locally calibrated models that describe louse population dynamics, and to allow prediction of the consequences that these parasites may have on stocks of wild salmon. While models may not be used for the specific prediction of lice burdens, they may for example indicate maximum sustainable levels of parasites and on different host fish species. Knowledge of exactly what levels are detrimental to fish may allow farmers to reduce the amount of medication used, benefiting the environment and the farmer. The end result of this project will be the integration, into the present model describing the interactions between fish and parasites, of the large amounts of data generated. This will enable us to model “safe� sites that would be suitable for the establishment of fish farms, and where the risk of infestation of wild smolts is low, in addition to facilitating the determination of areas of high risk to wild smolts.

38

SUMBAWS

Final Report

Paper 1: MIGRATION SPEEDS AND ORIENTATION OF ATLANTIC SALMON AND SEA TROUT POST-SMOLTS IN A NORWEGIAN FJORD SYSTEM E. B. Thorstad, F. Økland, B. Finstad, R. Sivertsgård, P. A. Bjørn and R. S. McKinley (published in Environmental Biology of Fishes (2004) 71, 305-311) Introduction Anadromous salmonids move from fresh water to the ocean to gain weight, but Atlantic salmon (Salmo salar L.) and sea trout (Salmo trutta L.) differ in their strategies during the sea phase. Atlantic salmon migrate to the open ocean (e.g. Hansen et al., 2003), whereas the sea trout remain in the inner fjord systems (e.g. Jonsson, 1985; Knutsen et al., 2001). The understanding of the early marine phase of the Atlantic salmon and the environmental factors that may influence their behaviour and distribution in the sea is limited (Moore et al., 2000), and even less is known about the sea trout. This lack of information is particularly critical because the heaviest mortality of salmon in the sea apparently takes place during the first months after the smolts leave fresh water (see Hansen et al., 2003).

During the last decades, intensive fish farming in fjord and coastal areas has led to higher concentrations of salmon lice (Lepeophtheirus salmonis Krøyer) (Heuch and Mo, 2001), and salmon lice infestations have been reported to cause significant sub-lethal and lethal effects on wild Atlantic salmon and sea trout (Tully et al., 1999; Bjørn et al., 2001; Heuch and Mo, 2001; Tully and Nolan, 2002). The developing fish farming industry has been charged with reducing the negative effects of sea lice on wild salmonid populations. One strategy is to locate fish farms in areas where the risk of salmon lice infestations is reduced. This requires information on migratory routes of the post-smolts and the amount of time they spend in different fjord areas. However, knowledge about migratory routes, geographical distribution and swimming speeds of postsmolt salmonids in fjord and near shore areas is sparse (Voegeli et al., 1998; Lacroix and Voegeli, 2000; Moore et al., 2000). The aim of the present study was, therefore, to record observed and actual swimming speeds of Atlantic salmon and sea trout post-smolts in a Norwegian fjord system, and to initiate studies on the orientation mechanisms of the post-smolts.

Material and methods Study area The study was carried out in the inner part of the Romsdalsfjord system (Figure 1). The difference between low and high tide is approximately 1.6 m. Important Atlantic salmon and sea trout rivers, such as the River Eira, empty into the system (Figure 1). The River Eira has a mean annual water discharge of 15.5 m3 s-1.

In the Eresfjord in 2002, water temperature and salinity were measured at four stations in the middle of the fjord, from the release site to 10 km from the river mouth (Figure 1). The water temperature at the surface was 8.6-10.3 °C on 14 May and 10.5-14.5 °C on 2 June. At 1 m depth, the water temperature was 9.2-9.9 °C on 14 May and 12.0-12.5 °C on 2 June. The salinity increased along the fjord towards the sea, and was 29.3-30.1 ppt at the surface on 14 May and 24.2-26.4 ppt on 2 June. At 1 m depth, the salinity was 29.3-30.1 ppt on 14 May and 24.2-26.4 ppt on 2 June.

39

SUMBAWS

Final Report

Figure 1 The Romsdalsfjord system in Middle Norway. Site for release of acoustically tagged smolts (•) and stations for temperature and salinity recordings (ĂŒ) are indicated.

Handling, tagging and release of fish Hatchery-reared Atlantic salmon (n = 5) and wild sea trout (n = 4) smolts were tagged with acoustic transmitters (VEMCO Ltd., Canada, V8SC-6L, 9 x 20 mm, weight in air of 3.3 g) at the Statkraft hatchery in Eresfjord. The Atlantic salmon (mean total length 26.3 cm, range 22.6-30.0; mean weight 181 g, range 105-278) were two-year old smolts from the hatchery. A seawater tolerance test (Blackburn and Clarke, 1987) performed on the hatchery-reared salmon on 3 May revealed plasma chloride levels at 130 mM, indicating that they had smolted (Sigholt and Finstad, 1990). The salmon were also ready to actively migrate, as indicated by salmon from the same group that were tagged with coded transmitters and recorded by automatic listening stations 48 km from the release site on average 65 hours later (i.e. minimum migration speed of 0.69 body lengths per second, own unpublished data). The sea trout (mean total length 20.3 cm, range 18.1-24.5, mean weight 62 g, range 43-100) were caught in a trap during downstream migration in the River Eira and transported in a tank with oxygenated water to the hatchery. The fish were allowed to recover from the stress of being captured, and the individuals that seemed most robust were tagged after 0-6 days.

The fish were anaesthetised by a 3-min immersion in an aqueous solution of 2-phenoxy-ethanol (EC No 204-589-7, SIGMA Chemical Co., USA, 0.5 ml/l). A 1.2-1.3 cm incision was made on the ventral surface posterior to the pelvic girdle. The transmitter was inserted through the incision and pushed forward into the body cavity. The incision was closed using two independent silk sutures (4/0 Ethicon). Mean handling time was 2.8 min.

Tagged smolts were transported in plastic bags with water to a cage in the fjord at the mouth of the River Eira (mean time from anaesthetisation to release in the cage was 17 min), where they could recover from anaesthetisation and tagging without being prone to predation for 1-3 days before release. The cage (90 cm diameter x 75 cm) was constructed of knotless nylon net (mesh size 13x13 mm). The fish were released together with 25-30 non-tagged hatchery-reared Atlantic salmon smolts.

40

SUMBAWS

Final Report

Fish tracking and simultaneous current measurements The fish were manually tracked from a boat using a VEMCO VR60 receiver in the period 10 May to 1 June 2003. The fish position was fixed every 10th minute after release using a GPS. Individual fish were followed for up to 14 hours (salmon average 9.3 hours, range 6.6-13.9; trout average 8.6 hours, range 6.113.9).

To correct fish swimming speeds and directions for the water current, a separate boat and crew used a current drift drogue to simultaneously record the speed and direction of the water current. Every 30th minute, the current drift drogue was put into the water at the site where the fish was recorded. After ten minutes, the position of the current drift drogue was recorded using a GPS. The current drift drogue was made of canvas, and had three wings (57 cm wide and 1 m deep) mounted at 120째 angles to each other. The current drift drogue was attached 30 cm below a floater at the surface and, hence, reflected the current at 0.3-1.3 m depth. Atlantic salmon post-smolts usually swim close to the surface, in the upper brackish layer (Fried et al., 1978; LaBar et al., 1978; Holm et al., 2003), which was also assumed to be true for sea trout in an estuary (Moore et al., 1998).

Actual swimming speeds and directions of the fish were calculated by vector analysis based on observed movements of the fish and the direction and speed of the water current. Statistical analyses were performed with SPSS 11.5 and the GPS mapping software OziExplorer 3.90.3a.

Results Swimming speeds Mean observed migration speed was 1.27 bl.s-1 (SD = 0.43, individual means from 0.94 to 2.00) for Atlantic salmon and 0.56 bl.s-1 (SD = 0.23, individual means from 0.33 to 0.88) for sea trout. When corrected for the speed and direction of the water current, the actual swimming speed was on average 1.32 bl s-1 (SD = 0.28, individual means from 1.10 to 1.79) for Atlantic salmon and 0.68 bl s-1 (SD = 0.29, individual means from 0.48 to 1.11) for sea trout. The actual migration speed (bl s-1) was not dependent on time from release (linear regression, r2 = 0.004, p = 0.48).

The fish did not follow a straight route seawards along the fjord (see below). The most seaward position of the Atlantic salmon during the tracking period was on average 4,702 m (range 2,195-10,022 m) from the release site, and of the sea trout 1,687 m (range 113-3,702 m). This corresponds to a mean net rate of seaward movement of 506 m h-1 for Atlantic salmon and 196 m h-1 for sea trout.

Orientation The current direction varied considerably during tracking, however, the highest number of recordings showed the current flowing out of the fjord (Figure 2). Mean current speed was 0.16 m s-1 (range 0-0.74, SD= 0.13).

The highest frequency of observed movements of the Atlantic salmon was out of the fjord (northern and western to northern directions, Figure 2). The actual movement, when corrected for the speed and direction of the water current, revealed the lowest frequencies in towards the fjord (eastern to southern directions, Figure 2). For Atlantic salmon, the direction of neither observed nor actual movement was dependent on the direction of the water current (linear regressions, observed: r2 = 0.021, p = 0.22, actual: r2 = 0.009, p= 0.42).

However, the direction of the observed movement was dependent on the actual movement (linear regression, r2 = 0.78, p < 0.001).

The sea trout had less directed movement patterns than the Atlantic salmon, and no patterns in the direction of observed or actual movement were seen (Figure 2). As for the salmon, neither the direction of observed nor the actual movement was dependent on the direction of the water current (linear regressions, observed: r2 = 0.029, p = 0.24, actual: r2 = 0.005, p = 0.64), but the direction of the observed movement was dependent on the actual movement (linear regression, r2 = 0.50, p < 0.001).

Mean distance to shore was 374 m for the Atlantic salmon (based on individual means, which ranged from 146-468 m) and 125 m for the sea trout (individual means from 26-468 m).

41

SUMBAWS

Final Report

Figure 2 Directions of the water current and observed and actual (i.e. after correcting for the speed and direction of the water current) movements of Atlantic salmon (n = 5) and brown trout (n = 4) post-smolts tagged with acoustic transmitters and manually tracked in a Norwegian fjord. The figures show the frequency of 10-min recordings in the directions north to east (N to E, 0-90°), east to south (E to S, 90-180°, i.e. inwards the fjord), south to west (S to W, 180-270°) and west to north (W to N, 270-360°, i.e. outwards the fjord). In the inner part of the fjord, migration in northern directions, as performed by some salmon, will also be more or less directed seaward. For the salmon, each group of directions was therefore divided into two bars to present the data in more detail (N to E is presented as one bar for 0-45° and one bar for 46-90°).

Stacked bars indicate contribution from each individual.

42

SUMBAWS

Final Report

Discussion The present study showed that both the Atlantic salmon and sea trout post-smolts were not passively drifting with the current, but actively swimming (actual swimming speeds of 1.3 bl s-1 for Atlantic salmon and 0.7 bl s-1 for sea trout). There was no relationship between the direction of observed movement and the direction of the water current, indicating that the post-smolts were moving in random directions in relation to the water current. The strong relationship between the direction of the actual and observed movement of the post-smolts indicates that the observed directions of movement were highly dependent on the actual movement of the fish and not of the water current.

In contrast to the results in the present study, several other studies suggest that water currents are the major transport factor in seaward migration of Atlantic salmon smolts, although most studies also emphasize observations of active directed swimming (Fried et al., 1978; LaBar et al., 1978; Lacroix and McCurdy, 1996; Moore et al., 2000). The actual movements of the fish may have been masked by strong water currents in some of these studies. The active swimming component by the fish may be easier to detect when recording migrations in a fjord system with relatively light currents, as in the present study. The differences between the present study and previous studies may also be a result of the detailed and simultaneous water current recordings made in the present study, as compared to studies with a few current recordings at fixed stations, or corrections based solely upon the direction of the tidal flow. However, another explanation for differences between studies could be that post-smolts are not swimming actively in strong currents.

In Norway, most of the Atlantic salmon post-smolts leave the rivers from May to July (Hvidsten et al., 1998) and migrate to the open ocean (Moore et al., 2000). In contrast, sea trout seem to remain in the inner fjord systems (e.g. Jonsson, 1985; Knutsen et al., 2001), which may be the reason why we recorded what seemed to be a slower and less directed migration closer to shore for the sea trout than for the Atlantic salmon. Even less is known about the marine phase of sea trout than of Atlantic salmon, and we know no other publications of telemetry studies of the marine migration of sea trout, except estuarine studies by Moore and Potter (1994) and Moore et al. (1998). The sea trout post-smolts were passively moving with the ebb tide in the upper sections of the estuaries, but were actively swimming in the lower sections (Moore and Potter, 1994; Moore et al., 1998), as they did in the fjord areas in the present study.

Orientation mechanisms used by migrating salmonid smolts and post-smolts apparently differ among species and habitats (Groot, 1965; LaBar et al., 1978; Thorpe et al., 1981; Quinn and Brannon, 1982; Moser et al., 1991). Landmarks, celestial cues, currents, electric and magnetic fields and olfaction are among suggested orientation and navigation cues for fishes (e.g. Stasko, 1973; Lucas and Baras, 2001). Water currents seemed not to be systematically used as an orientation cue by either Atlantic salmon or sea trout in the present study. However, the sea trout were predominantly recorded relatively close to the shore, hence, the shoreline may be used to orient. Magnetic material has been found in the lateral line of Atlantic salmon, and in the nose of several fish species, which may enable them to use the earth's magnetic field for navigation (Potter and Moore, 1991; Lucas and Baras, 2001). However, the lack of highly directional movements in this study did not indicate precise navigation, although the lowest frequency of movements of Atlantic salmon post-smolts was in towards the fjord, resulting in a net seaward movement.

The high swimming speeds recorded in the present study indicates that Atlantic salmon post-smolts have the potential to pass fjord areas with fish farms and high concentrations of salmon lice during a relatively short period. However, the lack of directional movements will increase their stay in inner fjord areas and increase the vulnerability of being sea lice infested. Based on the results in the present study, it seems difficult to predict migration routes of post-smolts based on currents, but similar studies in outer fjord systems could reveal other patterns. Sea trout post-smolts are clearly more vulnerable to sea-lice infestations in inner fjord areas than Atlantic salmon, with their slower migration speeds and longer stay in these areas.

Wild Atlantic salmon smolts in most Norwegian watersheds are too small to be tagged with the acoustic transmitters that are currently available. Migratory routes of wild and hatchery-reared Atlantic salmon were similar in the studies of Lacroix and McCurdy (1996) and Voegeli et al. (1998). However, we suggest that the approach in the present study should be further explored with wild fish, and extended to a longer time period during their outward migration.

43

SUMBAWS

Final Report

References Bjørn, P. A., Finstad, B. and Kristoffersen, R. (2001). Salmon lice infection of wild sea trout and Arctic char in marine and freshwaters: the effects of salmon farms. Aquaculture Research 32, 947-962.

Blackburn, J. and Clarke, W. C. (1987). Revised procedure for the 24 hour seawater challenge test to measure seawater adaptability of juvenile salmonids. Canadian Technical Report of Fisheries and Aquatic Sciences 1515, 1-39.

Fried, S. M., McCleave, J. D. and LaBar, G. W. (1978). Seaward migration of hatchery-reared Atlantic salmon, Salmo salar, smolts in the Penobscot River estuary, Maine: riverine movements. Journal of the Fisheries Research Board of Canada 35, 76-87.

Groot, C (1965). On the orientation of sockeye salmon (Oncorhynchus nerka) during their seaward migration out of lakes. Behaviour 14 (Suppl.), 1-198.

Hansen, L.P., Holm, M., Holst, J.C. and Jacobsen, J.A. (2003). The Ecology of post-smolts of Atlantic salmon. In Salmon at the edge (ed. D. Mills), pp. 25-39. Blackwell Science.

Heuch, P. A. and Mo, T. A. (2001). A model of salmon louse production in Norway: Effects of increasing salmon production and public management measures. Diseases of Aquatic Organisms 45, 145-152.

Holm, M., Holst, J. C., Hansen, L. P., Jacobsen, J. A., O'Maoiléidigh, N. and Moore, A. (2003).

Migration and distribution of Atlantic salmon post-smolts in the North Sea and North-East Atlantic. In Salmon at the edge (ed. D. Mills), pp. 7-23. Blackwell Science.

Hvidsten, N. A., Heggberget, T. and Jensen, A. J. (1998). Sea water temperature at Atlantic salmon smolt enterance. Nordic Journal of Freshwater Research 74, 79-86.

Jonsson, B. 1985. Life history patterns of freshwater resident and sea-run migrant brown trout in Norway.

Transactions of the American Fisheries Society 114, 182-194.

Knutsen, J. A., Knutsen, H., Gjøsæter, J. and Jonsson, B. (2001). Food of anadromous brown trout at sea. Journal of Fish Biology 59, 533-543.

LaBar, G. W., McCleave, J. D. and Fried, S. M. (1978). Seaward migration of hatchery-reared Atlantic salmon (Salmo salar) smolts in the Penobscot River estuary, Maine: Open water movements. Journal du Conseil International pour l' Exploration de la Mer 38, 257-269.

Lacroix, G. L. and McCurdy, P. (1996). Migratory behaviour of post-smolt Atlantic salmon during initial stages of seaward migration. Journal of Fish Biology 49, 1086-1101.

Lacroix, G. L. and Voegeli, F. A. (2000). Development of automated monitoring systems for ultrasonic transmitters. In Advances in fish telemetry, Proceedings of the third conference on fish telemetry in Europe (eds. A. Moore and I. Russell), pp. 37-50. CEFAS, Lowestoft.

Lucas, M. and Baras, E. (2001). Migration of Freshwater Fishes. Blackwell Science, Oxford. 420 pp.

Moore, A. and Potter, E. C. E. (1994). The movement of wild sea trout, Salmo trutta L., smolts through a river estuary. Fisheries Management and Ecology 1, 1-14.

Moore, A., Lacroix, G. L. and Sturlaugsson, J. (2000). Tracking Atlantic salmon post-smolts in the sea.

In The ocean life of Atlantic salmon - environmental and biological factors influencing survival (ed. D. Mills), pp. 49-64. Fishing News Books, Oxford.

Moore, A., Ives, M., Scott, M. and Bamber, S. (1998). The migratory behaviour of wild sea trout (Salmo trutta L.) smolts in the estuary of the River Conwy, North Wales. Aquaculture 168, 57-68.

Moser, M. L., Olson, A. F. and Quinn, T. P. (1991). Riverine and estuarine migratory behaviour of coho salmon (Oncorhynchus kisutch) smolts. Canadian Journal of Fisheries and Aquatic Sciences 48, 1670-1678.

Potter, T. and Moore, A. (1991). Salmon migration mystery: Scientists magnetic navigation theory. Fish.

News 4030, 6.

Quinn, T. P. and Brannon, E. L. (1982). The use of celestial and magnetic cues by orienting sockeye salmon smolts. Journal of Comparative Physiology A 147, 547-552.

44

SUMBAWS

Final Report

Sigholt, T. and Finstad, B. (1990). Effect of low temperature on sea-water tolerance in Atlantic salmon (Salmo salar) smolts. Aquaculture 84, 167-172.

Stasko A. B., Sutterlin, A. M., Rommel, S. A. and Elson, P. F. (1973). Migration-orientation of Atlantic salmon (Salmo salar L.). In International Atlantic Salmon Symposium Special Publication Series 4 (eds. M. W. Smith and W. M. Carter), pp 119-137.

Thorpe, J. E., Ross, L. G., Struthers, G. and Watts, W. (1981). Tracking Atlantic salmon smolts, Salmo salar L., through Loch Voil, Scotland. Journal of Fish Biology 19, 519-537.

Tully, O. and Nolan, D. (2002). A review of the population biology and host-parasite interactions of the sea louse Lepeophtheirus salmonis (Copepoda: Caligidae). Parasitology 124, 165-182.

Tully, O., Gargan, P., Poole, W. R. and Whelan, K. F. (1999). Spatial and temporal variation in the infestation of sea trout (Salmo trutta L.) by the caligid copepod Lepeophtheirus salmonis (Krøyer) in relation to sources of infection in Ireland. Parasitology 119, 41-51.

Voegeli, F. A., Lacroix, G. L. and Anderson, J. M. (1998). Development of miniature pingers for tracking Atlantic salmon smolts at sea. Hydrobiologia 371/372, 35-46.

45

SUMBAWS

Final Report

Paper 2: MIGRATION OF HATCHERY-REARED ATLANTIC SALMON AND WILD SEA TROUT POST-SMOLTS IN A NORWEGIAN FJORD SYSTEM B. Finstad, F. Økland, E. B. Thorstad, P. A. Bjørn and R. S. McKinley (published in Journal of Fish Biology (2005) 66, 86-96) Introduction Salmonid smolting includes morphological, physiological, biochemical and behavioural changes (Wedemeyer et al., 1980; Hoar, 1988; Boeuf 1993; Finstad and Jonsson, 2001). The smolting process is influenced by photoperiod (Saunders and Henderson, 1970; McCormick and Saunders, 1987; McCormick et al., 1995) and water temperature (Johnston and Saunders, 1981; McCormick et al., 1999, 2000), and the process is partly driven by endogenous rhythms (Eriksson and Lundqvist, 1982). In wild populations, the smolt migration from rivers to the sea will occur during a few weeks in spring, and there will be an annual variation in the timing in each river, as well as variations among rivers at different latitudes (Hvidsten et al., 1998), at least partly as a consequence of variation in temperature (Jonsson and Ruud-Hansen, 1985) and light condition (Metcalfe et al., 1988).

Atlantic salmon (Salmo salar L.) and sea trout (Salmo trutta L.) differ in their strategies during the sea phase. Atlantic salmon migrate to the open ocean (e.g. Hansen et al., 2003), whereas sea trout remain in the inner fjord systems (e.g. Jonsson 1985, Knutsen et al., 2001). However, knowledge about migratory routes and geographical distribution of salmonids in fjords and near shore areas is sparse (Johnstone et al., 1995; Moore et al., 1990, 2000; Lacroix and McCurdy, 1996; Voegeli et al., 1998; Lacroix and Voegeli, 2000) and most studies have tracked salmonid smolts only during the freshwater and estuarine stages of migration (Fried et al., 1978; Moser et al., 1991; Greenstreet, 1992; Moore et al., 1995, 1998a). A method to increase this knowledge is the use of tracking and telemetry techniques to monitor the movements of post-smolts in the marine environment, as described in Moore et al., (2000) and Lacroix and Voegeli (2000).

The objective of this study was therefore to record the rate of progression of Atlantic salmon and sea trout post-smolts in a Norwegian fjord system on the West Coast of Norway by the use of coded miniature acoustic transmitters.

Materials and Methods Study area The study was carried out in the Romsdalsfjord system on the West Coast of Middle Norway, 62º40’N, 8º10’E (Figure 1). Receiver sites 1, 2, 3 and 4 were 9, 32, 48 and 77 km, respectively, from the smolt release site at the estuary of the River Eira. The difference between low and high tide is approximately 1.6 m.

Atlantic salmon and sea trout rivers, such as the River Eira (Figure 1), flow into the system. The mean annual water discharge in the River Eira is 15.5 m3 s-1 (Jensen et al., 2003).

46

SUMBAWS

Final Report

Figure 1 The Romsdalsfjord system in Middle Norway. Sites for receivers are given by open circles (12) and site of release of acoustically tagged smolts is indicated by an asterisk. The different fjord areas described in the paper are divided into areas 1 to 4.

Handling, tagging and release of fish Hatchery-reared Atlantic salmon (n = 25) and wild sea trout (n = 15) smolts were tagged with coded acoustic transmitters (VEMCO Ltd., Canada, model V8SC-6C-R256, 9 × 20 mm, weight in air of 3.0 g, weight in water of 1.7 g, frequency 69 kHz, see Lacroix and Voegeli (2000) and references therein) at the Statkraft hatchery in Eresfjord. Water temperature in the hatchery was 5.5-10.7 ºC. The trout (mean total length 19.9 cm, range 16.0-24.1, mean weight 70 g, range 38-122 g) were caught in a trap during downstream migration in the River Eira, transported to the hatchery in a tank with oxygenated water and tagged after 0-4 days, during 5-20 May 2002. The salmon (mean total length 29.5 cm, range 18.7-49.2; mean weight 270 g, range 64-1264 g) were two-year old smolts from the hatchery, with the exception of three three-year old smolts (843-1264 g), and were tagged during 9-20 May 2002. The ratio of transmitter weight in water to fish weight was from 1.4-4.5% (mean ratio 2.4%) and from 0.1-2.7% (mean ratio 0.6%), for sea trout and Atlantic salmon, respectively. A seawater tolerance test (Blackburn and Clarke, 1987) performed on the hatchery-reared salmon on 3 May revealed plasma chloride levels at 130 mM, indicative that the smolts were ready to be released into seawater (Sigholt and Finstad, 1990).

The fish were anaesthetised by immersion in an aqueous solution of 2-phenoxy-ethanol (EC No 204-5897, SIGMA Chemical Co., USA, 0.7 ml/l, mean time 3.4 ± 1.0 min) and placed ventral side up onto a Vshaped surgical table. Approximately 1.3 cm incision was made on the ventral surface posterior to the pelvic girdle using a scalpel. The transmitter was inserted through the incision and pushed into the body cavity above the pelvic girdle. The incision was closed using two or three independent silk sutures (3/0 Ethicon).

The fish were regularly showered with water during surgery (mean handling time 2.6 ± 1.1 min, mean recovery time 4.4 ± 3.1 min). Prior to each incision, surgical equipment was rinsed in 96% ethanol.

Groups of tagged smolts were transported in plastic bags (30x250 cm), with water and oxygen to a cage (Figure 1 – release site) in the fjord at the mouth of the River Eira (mean time from anaesthetisation to release into the cage, 39 ± 12 min). The cage consisted of two rings in top and bottom connected by a knotless nylon net with mesh size 13 × 13 mm. The distance between the rings was 75 cm and the diameter of each ring was 90 cm, giving a total volume of 0.5 m³. All fish were recovered in the cage for two or three 47

SUMBAWS

Final Report

days prior to release. Between one and eight fish were kept in the same cage and released together with 25-30 non-tagged hatchery-reared salmon smolts, with the exception of five tagged salmon smolts released alone into a shoal of smolts observed at the release site. The cage was attached to a small pontoon harbour, and the smolts were released outside the harbour (Figure 1). Shoals of wild and Carlin-tagged smolts were frequently seen inside and outside the harbour early in the study period. The majority of the fish initiated migration 10-30 minutes after release from the cages. The criterion here was that they were not observed near the cages after this migration.

Recording og tagged fish after release Fish were recorded by 12 individual receivers (VEMCO model VR2) moored at different sites throughout the fjord system (9-77 km from the river mouth, Figure 1), which recorded identification number and time stamp from transmitters as fish travelled through receiver range. Range was typically a radius of 120-260 m for transmitters at 0.5-3.0 m depth, but varied considerably (radius of 45-620 m during range tests) with factors such as wave action, salinity and depth. The depth of the water column (sea depth) at receiver sites were 7 to 35 m and receivers were attached to the mooring 2.5-3.0 m below surface. Data were downloaded on 7 June and 5 September. The outermost VR2 on the northern side was found to be missing on 5 September, probably due to strong currents relocating the mooring to greater depth.

Analyses Analysis of migration speed and residence time in different parts of the fjord system required that the fjord system to be divided into four areas, according to localisation of the automatic receivers at four different sites (Figure 1). Sites 1, 2, 3 and 4 were 9, 32, 49 and 77 km, respectively, from the release site at the estuary of the River Eira. Two receivers were placed at each site, one on each side of the fjord, except site 3, where five receivers were placed across the fjord (Figure 1). Since the range of the receivers at each site did not cover the width of the fjord, thus, all passing smolts would not be detected.

The data were used to calculate the rate of progression through the fjord system (i.e. time from release to that of the first recording at the four different sites, and corresponding migratory speeds), time spent in different parts of the fjord system (e.g. time spent in area 2 is time from the last recording at site 1 to the last recording at site 2) and for migration speeds between the different receiver sites (e.g. migration speed in area 2 is distance between site 1 and 2 divided by the time from the last recording at site 1 to the first recording at site 2). All migration speeds are provided the shortest migration route, since we do not know the exact migration route between receiver sites.

Time spent in the different parts of the fjord system could be calculated only for those smolts recorded both entering and leaving a particular fjord area. The same was true for calculations of swimming speeds.

The sample sizes for these analyses were, therefore, smaller than the total number of smolts recorded.

Results Water temperature and salinity Water temperature and salinity were measured at eight sites in the middle of the fjord distributed 1-60 km from the river mouth 3-4 May, 14 May, 24 May, 2 June and 7 June at 1 m depth. Salinity increased outwards in the fjord system. Water temperature varied between 8.2 and 17.7 ºC (mean 12.2 ± 3.0ºC). Mean water temperature increased during the season, from 8.4 ºC on 3-4 May to 16.7 ºC on 7 June. Salinity varied between 16.7 and 31.1 ppt (mean 25.8 ± 3.8 ppt). Mean salinity decreased during the season, from 29.1 on 3-4 May to 21.8 ppt on 7 June.

48

SUMBAWS

Final Report

Smolt migration Seventeen salmon (68%) and eight trout (53%) post-smolts were recorded by the automatic receivers (Table 1). Mean times, from release to first recording at the receiver site, as well as the corresponding migration speed, showed that salmon smolts had a faster and more orientated migratory pattern than sea trout smolts.

Table 1 Number and proportion of acoustically tagged hatchery-reared Atlantic salmon and wild sea trout smolts recorded at the different automatic receiver sites in the Romsdalsfjord system in 2002. The rate of progression through the fjord system (mean time from release to first recording at the different receiver sites) is also given, as is the corresponding migration speed measured as km h-1 and bl s-1 (body lengths per second).

Atlantic salmon

Sea trout

Mean Number Mean time (h) Mean migration Receiver Number Mean time (h) speed from recorded from release to site recorded from release to migration (%) first recording at release site to (%) first recording speed from receiver site receiver site at receiver site release site to (km h-1/ bl s-1) receiver site (range, SD) (range, SD) (km h-1/ bl s-1) 1 14 (56) 28 (6-92, 23) 0.54/0.41 8 (53) 438 (29-1657, 515) 0.07/0.11 2 7 (28) 67 (36-143, 36) 0.56/0.44 3 (20) 247 (56-448, 196) 0.25/0.33 3 13 (52) 65 (40-166, 32) 0.69/0.69 4 (27) 270 (65-451, 161) 0.25/0.38 4 4 (16) 83 (70-91, 9) 0.83/0.78 0 (0) N/A N/A There was no difference between salmon and trout in proportions of the total release recorded by the automatic receivers after release. However, the proportion recorded in outer parts of the fjord system, i.e. outside area 1, was higher for salmon (n = 15, 60%) than for trout (n = 4, 27%) (Pearson Chi-square test, Ď&#x2021;2 = 4.18, p = 0.041.

Salmon and trout smolts showed no significant difference in the numbers recorded or not recorded with respect to date released (based on recordings by automatic receivers, Mann-Whitney U-test, salmon: U = 0.53, p = 0.41, trout: U = 13.5, p = 0.094). Salmon recorded after release were larger (median 260 g) than those not recorded (median 153 g, Mann-Whitney U-test, U = 30.5, p = 0.027). For trout, there was no difference in body size for those recorded (median 75 g) and not recorded (median 61 g) (Mann-Whitney Utest, U = 19.5, p = 0.34). There was no difference in body weight for smolts recorded at the different receiver sites, for either species (Kruskal-Wallis tests, salmon: Ď&#x2021;2 = 1.81, p = 0.61, trout: Ď&#x2021;2 = 0.99, p= 0.61).

Salmon stayed considerably longer in the inner part of the fjord system (area 1), than in the outer parts, even though this was the shortest fjord area (Table 2). Time spent in area 1 was not dependent on body size (linear regression, weight: r2 = 0.021, p = 0.62, length: r2 = 0.045, p = 0.47).

Table 2 Mean time spent by acoustically tagged hatchery-reared Atlantic salmon smolts in the different parts of the Romsdalsfjord system in 2002 (e.g. time spent in fjord area 2 is time from the last recording at site 1 to the last recording at site 2).

Receiver site 1 2 3 4

Time (h) spent in fjord area mean range SD 107 13-584 192 29 16-46 11 55 4-296 118 38 31-53 12

Length of fjord area (km) n 14 5 6 3

9.4 22.1 16.9 29.0

Migration speed in area 1 was higher for salmon than trout (Mann-Whitney U-test, km h-1: U = 15.0, p=0.002, bl s-1 (body lengths per second): U = 10.0, p = 0.001, Table 2.3), and for salmon, migration speed, measured as km h-1 was dependent on body weight but not body length (linear regressions, weight: r2 = 0.33, p = 0.031, length: r2 = 0.19, p = 0.12), while there was no relationship between migration speed, measured as body lengths s-1, and body size (weight: r2 = 0.003, p = 0.86, length: r2 = 0.017, p = 0.66). No 49

SUMBAWS

Final Report

relationships between migration speed and body size were found for trout (r2 from 0.24 to 0.34, p from 0.13-0.22). Migration speeds appeared to increase throughout the fjord system (Table 3), but sample sizes were too small for statistical comparison.

Table 3 Migration speeds for acoustically tagged hatchery-reared Atlantic salmon and wild sea trout smolts in different parts of the Romsdalsfjord system in 2002, measured both as km h-1 and body lengths per second (bl s-1) (i.e. migration speeds between different receiver sites: for example migration speed in fjord area 2 is distance between receiver site 1 and 2 divided by time from last recording at site 1 to first recording at site 2; for fjord area 1 the release time and site is used together with first recording at receiver site 1).

Fjord area

n

1

14

2

6

3

6

4

3

Atlantic salmon km h mean bl s-1 mean (range, SD) (range, SD) 0.54 0.49 (0.10-1.58, 0.39) (0.06-1.10, 0.28) 0.77 0.91 (0.26-1.37, 0.39) (0.26-1.74. 0.52) 1.57 1.82 (0.29-2.14, 0.67) (0.28-2.71, 0.87) 1.18 1.18 (0.81-1.85, 0.58) (0.78-1.64, 0.43) -1

n 8 2 3 0

Sea trout km h mean bl s-1 mean (range, SD) (range, SD) 0.07 0.11 (0.01-0.32, 0.10) (0.01-0.52, 0.17) 1.33 2.11 (1.20-1.46, 0.18) (1.94-2.27, 0.23) 1.76 2.60 (0.92-2.78, 0.94) (1.48-4.33, 1.52) -1

N/A

N/A

Smolts usually did not migrate back and forth between the receiver sites. The exceptions were one salmon (28.3 cm) recorded at site 1 after it had been recorded at site 2, and one trout (22.5 cm) recorded at site 1 after it had been recorded at site 2 and 3 (release: 23 May, site 2; 2 June, site 3; 2-3 June and 30 July, site 2; 31 July, site 1).

There was no fixed migration route for salmon and trout smolts through the fjord system. At site 1, six salmon and six trout were recorded only by the receiver on the western side of the fjord, two salmon only by the receiver on the eastern side of the fjord, whereas six salmon and two trout were recorded by the receivers on both sides of the fjord (five salmon and one trout were recorded simultaneously by both receivers, and one salmon and one trout were recorded subsequently by the receivers). At site 2, six salmon and one trout were recorded by the receiver on the northern side of the fjord, one salmon and one trout by the receiver on the southern side of the fjord, whereas one trout was recorded by both receivers (subsequent recordings). In addition, different migration routes were recorded past islets in the outer part of the Langfjord (site 3).

Discussion In Norway, most of Atlantic salmon smolts migrate from rivers to the ocean from May to July (Hvidsten et al., 1998), and follow outgoing surface currents through fjords (Fried et al., 1978; Dutil and Coutu, 1988; Levings et al., 1994) and coastal waters to the Norwegian Coastal Current and open ocean (Moore et al., 2000). Sea trout (Salmo trutta L.) exhibit various life history patterns and exploit freshwater habitats and coastal areas (Jonsson, 1985, 1989; Berg and Jonsson, 1990; Rikardsen et al., 2000; Knutsen et al., 2001), however, knowledge about migratory routes and geographical distribution of salmonids in fjords and near shore areas is sparse (Moore et al., 1990, 2000; Lacroix and McCurdy, 1996; Voegeli et al., 1998).

Comparison between Atlantic salmon and sea trout post-smolts in the present investigation showed that the percentage of fish recorded on the receivers in the fjord system was good, representing 68 and 53% of reared Atlantic salmon and wild sea trout, respectively. Salmon migrated through the whole fjord system into the ocean, but stayed for a longer time in the inner part as opposed to the outer part, perhaps because they needed time to acclimate and orientate themselves after release from the cages. Approximately 5 days after the release of salmon in the estuary of the River Eira, all fish had left area 1 and were recorded in area 2. Subsequent migration was more straightforward and orientated toward the ocean.

Calculated migration speeds assumes a straight line migration path between detection sites and should therefore be considered as minimum estimates in this study. Migration speeds were higher for salmon than for trout in fjord area 1 (from release site to receiver site 9 km from release site). These results correlate with those in Table 2.1, where trout had a mean time to first recording at site 1 of 438 h compared to 28 h for salmon. Thorstad et al. (2004) performed manual tracking of post-smolts in this same area (fjord area 1), and mean observed migration speed in ten-minute intervals was 1.27 and 0.56 bl s-1 (body lengths per 50

SUMBAWS

Final Report

second), respectively for Atlantic salmon and sea trout. Except of four sea trout detected at receiver site 3 in the present experiment, sea trout did not appear to have a targeted outwards migration and remained in the inner fjord system where they were released.

The behaviour of Atlantic salmon post-smolts was in accordance with a previous study on smolts tagged by miniature acoustic tags in New Brunswick, Canada, (Lacroix and McCurdy, 1996; Voegeli et al., 1998), which showed that the migratory routes of wild and hatchery-reared fish were similar and that the postsmolts moved rapidly away from coastal waters. In a study by Johnstone et al., (1995), sea trout were individually fitted with external acoustic transmitters and followed by manual tracking. These studies showed that three smolts returned to freshwater within a few hours after release and that the remaining smolts continuously were tracked for periods up to 68 h and that one smolt was followed for over 10 days.

Smolts were found to generally remain in shallow water in the littoral and immediate sub-littoral zone.

Similar tagging studies have been performed on wild sea trout in the estuary of the River Conwy, North Wales (Moore et al., 1998b), but these fish were not followed during the fjord migration. Further, Thorstad et al. (2004), performed a similar study with manual tracking of Atlantic salmon and sea trout. The results showed that the observed direction of movement was dependent on the actual movement of the fish and not the water current. Water currents were not systematically used as an orientation cue either in Atlantic salmon or sea trout, as the actual movements were random compared to the direction of the water current.

The actual movement of sea trout was in all compass directions, with no systematic pattern. The Atlantic salmon also moved in all compass directions, but with the lowest frequency of actual movement towards the fjord. The present study thus revealed that some sea trout had a high migratory speed and were monitored at site 3 (48 km from the river mouth) about 65 h after release. In contrast to salmon migrations, we did not record any sea trout at site 4 which is 77 km from the river mouth.

The significant relationship between size and probability for detection for salmon may depend on different mortality or different behaviour/migratory pattern between different size classes of fish, which varied from 18.7-49.2 cm. The size range for trout varied from 16.0-24.1 cm, and a lack of a relationship between trout size and probability of detection might be due to the smaller size range of tagged fish. Further, salmon of large size have a better swimming ability than smaller fish, but related to body lengths per second (LT s-1) there seems to be no difference.

In the present project, the migratory behaviour of both wild and farmed salmonid postsmolts has been investigated through use of automatic detection stations. . By a combination of automatic- and manual tracking (Thorstad et al., 2004), it will be possible to map the migratory routes of wild and farmed salmonid post-smolts in fjord systems and combine this with hydrodynamic models (e.g. Asplin et al., 1999). The overall goal is to build locally calibrated models, which may be used in coastal zone management such as identifying suitable sites for fish farms that will minimize the interference with wild salmonid populations.

References Asplin, L., Salvanes, A.G.V. and Kristoffersen, J.B. (1999). Non-local wind-driven fjord-coast advection and its potential effect on plankton and fish recruitment. Fisheries Oceanography 8, 255-263.

Berg, O.K. and Jonsson, B. (1990). Growth and survival rates of the anadromous trout, Salmo trutta, from the Vardnes River northern Norway. Environmental Biology of Fishes 29, 145-154.

Blackburn, J. and Clarke, W.C. (1987). Revised procedure for the 24 hour seawater challenge test to measure seawater adaptability of juvenile salmonides. Canadian Technical Report of Fisheries and Aquatic Sciences 1515, 1-39.

Boeuf, G. (1993). Salmonid smolting: a preadaptation to the oceanic environment. In Fish Ecophysiology.

(Rankin, J.C. and Jensen, F.B., eds.), pp. 105-135. London: Chapman and Hall.

Dutil, J.-D. and Coutu, J.-M. (1988). Early marine life of Atlantic salmon, Salmo salar, postsmolts in the northern Gulf of St Lawrence. Fisheries Bulletin 86, 197-212.

Eriksson, L. O. and Lundqvist, H. (1982). Circannual rhytms and photoperiod regulation of growth and smolting in Baltic salmon Salmo salar L. Aquaculture 28, 113-120.

Finstad, B. and Jonsson, N. (2001). Factors influencing the yield of smolt releases in Norway. Nordic Journal of Freshwater Research 75, 37-55.

51

SUMBAWS

Final Report

Fried, S.M., McCleave, J.D. and LaBar, G.W. (1978). Seaward migration of hatchery-reared Atlantic salmon, Salmo salar, smolts in the Penoboscot River estuary, Maine: riverine movements. Journal of the Fisheries Research Board of Canada 35, 76-87.

Greenstreet, S.P.R. (1992). Migration of hatchery reared juvenile Atlantic salmon, Salmo salar L., smolts down a release ladder. 1. Environmental effects on migratory activity. Journal of Fish Biology 40, 655-666.

Hansen, L.P., Holm, M., Holst, J.C. and Jakobsen, J.A. (2003). The ecology of post-smolts of Atlantic salmon. In Salmon at the edge. Edited by D. Mills. Blackwell Science, Oxford. pp. 25-39.

Hoar, W.S. (1988). The physiology of smolting salmonids. In Fish physiology: The physiology of developing fish. Viviparity and posthatching juveniles, Volume XIB (Hoar, W.S. and Randall, D.J., eds.), pp.

275-343. New York, NY: Academic Press.

Hvidsten, N.A., Heggberget, T. and Jensen, A.J. (1998). Sea water temperature at Atlantic salmon smolt enterance. Nordic Journal of Freshwater Research 74, 79-86.

Jensen, A.J., Finstad, B., Hvidsten, N.A., Jensås, J.G., Johnsen, B.O. and Lund, R. (2003).

Fiskeribiologiske undersøkelser i Auravassdraget. Årsrapport 2002 (Fish biology surveys in the Aura watercourse. Annual report 2002). NINA Oppdragsmelding 781, 1-36 (In Norwegian with English summary).

Johnston, L. E. and Saunders, R.L. (1981). Parr smolt transformation of yearling Atlantic salmon, Salmo salar L., at several rearing temperatures. Canadian Journal of Fisheries and Aquatic Sciences 38, 11891198.

Johnstone, A.D.F., Walker, A.F., Urquhart, G.G. and Thorne, A.E. (1995). The movements of sea trout smolts, Salmo trutta L., in a Scottish west coast sea loch determined by acoustic tracking. Scottish Fisheries Research report. 56. S.O.A.E.F.D. Jonsson, B. (1985). Life history patterns of freshwater resident and sea-run migrant brown trout in Norway.

Transactions of the American Fisheries Society 114, 182-194.

Jonsson, B. (1989). Life history and habitat use of Norwegian brown trout (Salmo trutta). Freshwater Biology 21, 71-86.

Jonsson, B. and Ruud-Hansen, J. (1985). Water temperature as the primary influence of timing of seaward migrations of Atlantic salmon (Salmo salar) smolts. Canadian Journal of Fisheries and Aquatic Sciences 42, 593-595.

Knutsen, J.A., Knutsen, H., Gjøsæter, J. and Jonsson, B. (2001). Feeding of anadromous brown trout at sea. Journal of Fish Biology 59, 533-543.

Lacroix, G.L. and McCurdy, P. (1996). Migratory behaviour of post-smolt Atlantic salmon during initial stages of seaward migration. Journal of Fish Biology 49, 1086-1101.

Lacroix, G.L. and Voegeli, F.A. (2000). Development of automated monitoring systems for ultrasonic transmitters. In Advances in fish telemetry. Proceedings of the third conference on fish telemetry in Europe.

(Moore, A. and Russell, I., eds.), pp. 37-50. Lowestoft: CEFAS Levings, C.D., Hvidsten, N.A. and Johnsen, B.O. (1994). Feeding of Atlantic salmon (Salmo salar L.) postsmolts in a fjord in central Norway. Canadian Journal of Zoology 72, 834-839.

McCormick, S. D. and Saunders, R.L. (1987). Preparatory physiological adaptations for marine life of salmonids: osmoregulation, growth and metabolism. American Fisheries Society Symposium 1, 211-229.

McCormick, S. D., Björnsson, B.T., Sheridan, M., Eitertson, C., Carey, J.B. and O’Dea, M. (1995).

Increased daylength stimulates plasma growth hormone and gill Na+, K+-ATPase in Atlantic salmon (Salmo salar). Journal of Comparative Physiology B 165, 245-254.

McCormick, S. D., Cunjack, R.A., Dempson, B.D., M. F. O’Dea, M.F. and Carey, J.B. (1999).

Temperature-related loss of smolt characteristics in Atlantic salmon (Salmo salar) in the wild. Canadian Journal of Fisheries and Aquatic Sciences 56, 1649-1658.

McCormick, S.D., Moriyama, S. and Björnsson, B.T. (2000). Low temperature limits photoperiod control of smolting in Atlantic salmon through endocrine mechanisms. American Journal of Physiology Regulatory Integrative Comparative Physiology 278, R1352-R1361.

52

SUMBAWS

Final Report

Metcalfe, N. B., Thorpe, J.E. and Huntingford, F.A. (1988). Determinants of variation in life-history strategies in Atlantic salmon. - Abstract, 2nd. Internat. Conf. Behav. Ecol. Vancouver, Canada.

Moore, A., Russell, I.C. and Potter, E.C.E. (1990). Preliminary results from the use of a new technique for tracking the estuarine movements of Atlantic salmon, Salmo salar L., smolts. Aquaculture and Fisheries Management 21, 369-371.

Moore, A., Potter, E.C.E., M., Milner, N.J. and Bamber, S. (1995). The migratory behaviour of wild Atlantic salmon (Salmo salar) smolts in the estuary of the River Conwy, North Wales. Canadian Journal of Fisheries and Aquatic Sciences 52, 1923-1935.

Moore, A., Ives, S., Mead, T.A. and Talks, L. (1998a). The migratory behaviour of wild Atlantic salmon (Salmo salar L.) smolts in the River Test and Southampton Water, southern England. Hydrobiologia 372, 295-304.

Moore, A., Ives, M., Scott, M. and Bamber, S. (1998b). The migratory behaviour of wild sea trout (Salmo trutta L.) smolts in the estuary of the River Conwy, North Wales. Aquaculture 168, 57-68.

Moore, A., Lacroix, G.L. and Sturlaugsson, J. (2000). Tracking Atlantic salmon post-smolts in the sea. In The ocean life of Atlantic salmon - environmental and biological factors influencing survival. (Mills, D., ed.), pp. 49-64. Oxford: Fishing News Books.

Moser, M.L. Olson, A.F. and Quinn, T.P. (1991). Riverine and estuarine migratory behaviour of coho salmon (Oncorhynchus kisutch) smolts. Canadian Journal of Fisheries and Aquatic Sciences 48, 1670-1678.

Rikardsen, A.H., Amundsen, P.-A., Bjørn, P.A. and Johansen, M. (2000). Comparison of growth, diet and food consumption of sea-run and lake-dwelling Arctic charr. Journal of Fish Biology 57, 1172-1188.

Saunders, R. L. and Henderson, E.B. (1970). Influence of photoperiod on smolt development and growth of Atlantic salmon (Salmo salar). Journal of the Fisheries Research Board of Canada 27, 1295-1311.

Sigholt, T. and Finstad, B. (1990). Effect of low temperature on sea-water tolerance in Atlantic salmon (Salmo salar) smolts. Aquaculture 84, 167-172.

Thorstad, E.B., Økland, F., Finstad, B., Sivertsgård, R., Bjørn, P.A. and McKinley, R.S. (2004).

Migration speed and orientation of Atlantic salmon and sea trout post-smolts in a Norwegian fjord system.

Environmental Biology of Fishes 71, 305-311.

Voegeli, F.A., Lacroix, G.L. and Anderson, J.M. (1998). Development of miniature pingers for tracking Atlantic salmon smolts at sea. Hydrobiologia 371/372, 35-46.

Wedemeyer, G. A., Saunders, R.L. and Clarke, W.C. (1980). Environmental factors affecting smoltification and early marine survival of anadromous salmonids. Marine Fisheries Review 42, 1-14.

53

SUMBAWS

Final Report

Paper 3: EFFECTS OF SALMON LICE INFECTION AND SALMON LICE PROTECTION ON FJORD MIGRATING ATLANTIC SALMON AND BROWN TROUT POST-SMOLTS R. Sivertsgård, E.B. Thorstad, F. Økland, B. Finstad, P.A. Bjørn, N. Jepsen, T. Nordal and R.S. McKinley (Submitted to Hydrobiologia) Introduction Atlantic salmon, Salmo salar L., and brown trout, Salmo trutta L., smolts leave freshwater in the spring.

Atlantic salmon migrate to the open ocean, whereas brown trout remain feeding in fjords and near coastal areas (e.g. Lyse et al., 1998; Hansen et al., 2003; Klemetsen et al., 2003; Rikardsen, 2004). During the last decades, the marine parasite salmon lice Lepeophtheirus salmonis (Krøyer) has become one of the most serious pathogens on outward migrating post-smolts (Tully and Nolan, 2002; Heuch et al., 2005). Fish farming is believed to contribute significantly to elevated salmon lice levels observed in wild post-smolts (Bjørn et al., 2001; Butler, 2002; Gargan et al., 2003). Major physiological disturbances, such as high levels of plasma cortisol and reduced osmoregulatory ability, occur in both Atlantic salmon and brown trout when the lice develop into mobile stages (reviewed by Tully and Nolan, 2002), and may even kill heavily infected hosts (Finstad et al., 2000; Bjørn et al., 2001). Sublethal effects, such as reduced growth and swimming performance, have also been reported (e.g. Bjørn and Finstad, 1997; Nolan et al., 1999; Wagner et al., 2003). This may altogether impair fish from reaching feeding areas, finding food and avoiding predators, but such behavioural effects of salmon lice infection have never been studied in free-swimming salmonid hosts.

The aim of this study was to examine the effects of salmon lice infection and salmon lice protection on survival and rate of progression of Atlantic salmon and brown trout post-smolts during the first stage of their marine migration. Groups of post-smolts were artificially infected with salmon lice (Bjørn and Finstad, 1998), given a prophylactic pharmaceutical protection against salmon lice, or served as untreated control groups. Post-smolts were followed through a fjord system by using acoustic telemetry, and survival and rate of progression were compared among the groups. Infected fish were simultaneously kept in the hatchery to indicate infection intensity and lice development in free-swimming hosts during the study period.

Materials and methods The study was carried out in the Romsdalsfjord system, Norway (62º40’ N 8º10’ E) (Figure 1). Firstgeneration hatchery-reared offspring of Atlantic salmon and brown trout of the River Eira stock (aged 2+) were tagged with acoustic transmitters (VEMCO Ltd., Canada, model V8SC-6L, 9×20 mm, weight in air 3.3 g, weight in water 2.0 g, life time 70 days) as described in Finstad et al. (2005). Individuals of each species were divided into groups 1) artificially infected with salmon lice, 2) protected against salmon lice infection with a prophylactic pharmaceutical product (Pharmaq, Norway), and 3) untreated controls (Table 1). The prophylactic substance protects the fish against salmon lice infection for up to 16 weeks, have low toxicity for warm-blooded animals and humans, and its function is to prevent synthesis of chitin in salmon lice (Bernt Martinsen, Pharmaq, pers. comm.).

For artificial salmon lice infection, infective salmon lice copepodids from Austevoll Research Station (Boxaspen and Næss, 2002) were added to the infection tank (34 ‰, 7 ºC) in accordance with the methods of Bjørn and Finstad (1998). Each fish in the tank was exposed to approximately 200 newly hatched copepodids in a common garden experiment (Glover et al., 2003) containing both trout and salmon. Control fish were sham infected. The protected group was treated with lice protection in a separate trial, and the treatment period was 30 minutes in oxygenated water > 7 mg/l O2. In this period, 1000 L of water was withheld and 1 g of active substance (10 ml/1000 l) was added to the fish in the tank.

The fish were allowed one day of recovery in tanks, before being transported to a nylon net recovery cage (volume 0.5 m3) at the release site in the fjord (Figure 1), from where they were released 12 hours later at high tide during 10-14 May 2003. A total of 72 Atlantic salmon and 72 brown trout were tagged and released in the inner part of the fjord system near the mouth of the River Eira (Table 1). An equal number of fish from each group were released at the same time, and they were released together with 20-25 nontagged smolts. The recovery cage was placed at salinity ≥ 30 ‰ at two meters depth, because juvenile 54

SUMBAWS

Final Report

stages of salmon lice can be killed at ≤ 20‰ salinity (Johnson and Albright, 1991).

Table 1 Number of Atlantic salmon and brown trout smolts tagged with acoustic transmitters and recorded at different receiver sites in the Romsdalsfjord system in 2003 (The number tagged given in parentheses is the number of fish recorded by the receiver at the release site (see text). Total body length and body mass is given as mean values ± SD) Number tagged

Group

Atlantic salmon 22 (16) Lice infected 25 (16) Lice protected 25 (17) Control Brown trout 22 (17) Lice infected 25 (19) Lice protected 25 (20) Control

Total body length Body mass Mean ± SD (mm) Mean ± SD (g)

Number (proportion) at site 1

Number Number (proportion) (proportion) at site 2 at site 3

252 ± 2.7 254 ± 2.1 254 ± 2.3

147 ± 39 149 ± 31 141 ± 32

13 (59%) 13 (52%) 11 (44%)

10 (45%) 9 (36%) 9 (36%)

4 (18%) 3 (12%) 3 (12%)

235 ± 1.5 239 ± 1.6 236 ± 1.5

133 ± 25 121 ± 20 112 ± 20

7 (31%) 8 (32%) 2 (8%)

5 (23%) 3 (12%) 2 (8%)

2 (9%) 1 (4%) 0 (0%)

Figure 1 The Romsdalsfjord system, Norway, with the release site for acoustically tagged Atlantic salmon and brown trout smolts in 2003. Automatic receiver sites 1-3 are indicated.

Tagged fish were automatically recorded when passing 15 VR2- receivers (VEMCO Ltd.) moored at three sites in the fjord system, 7.1 km (site 1), 29-37 km (site 2) and 64-75 km (site 3) from the release site (Figure 1). The range of the receivers did not cover the whole transect of the fjord (each receiver covered typically a radius of 200-260 m for transmitters at 0.5-3.0 m depth, but range varied with conditions).

Hence, ‘survival rates’ are minimum estimates because some individuals may have passed without being recorded. However, standardised methods allow for comparison among groups. The time of release for individual fish was determined by a receiver placed at the release site. Due to technical problems, some fish were not recorded by this receiver. Results on rate of progression through the fjord system were, therefore, based on a smaller sample size than the total number of tagged fish (Table 1). All receivers were operating until the last week in August.

To control infection intensity and lice development of the tagged fish, 20 Atlantic salmon and 12 brown 55

SUMBAWS

Final Report

trout from the infection tank were retained in the hatchery (laboratory control groups). Between 5 and 10 fish of each species were examined for salmon lice development 15 and 25 days after infection, respectively. The saltwater in the hatchery was supplied from the fjord, and a similar lice development was therefore expected for the laboratory control groups and the fish tracked during the fjord migration, even though the latter groups may have been reinfected during their fjord migration after release and not necessarily be directly comparable to the laboratory control groups. The terms prevalence, intensity, abundance, median, intensity range and mean variance to mean ratio follow Bush et al. (1997).

Salinity was high (> 30 ‰) at all depths in the inner part of the fjord system on 21 May. From the second week of June, the salinity decreased in the surface layer, with salinity < 23 ‰ down to 1 m depth.

The salinity was still approximately 30 ‰ at all depths in the outer part of the system. The water temperature was around 7 ºC in the upper 5 m on 21 May. The surface temperature increased to 10-14 ºC in the second week of June and to 16ºC in mid-July.

Results Survival There was no difference among groups in proportions of fish detected at receiver site 1, 2 and 3, respectively, for neither Atlantic salmon (Pearson Chi-square tests; χ2 = 1.1, p = 0.59; χ2 = 0.58, p = 0.75; χ2= 0.49, p = 0.78) nor brown trout (χ2 = 4.6, p = 0.10; χ2 = 2.2, p = 0.33; χ2 = 2.4, p = 0.30) (Table 1).

Thus, no differences in survival among groups were detected based on proportions of post-smolts recorded at the different receiver sites.

Rate of progression Atlantic salmon post-smolts spent on average 6-8 days in passing the entire fjord system, with the slowest migrating individual spending 15.2 days in the fjord system (Table 2). Brown trout remained to a larger extent within the fjord system, and the last individual was recorded 31 days after release (Tables 1, 2). If including brown trout individuals for which the exact release time was not recorded (see methods), the last individual was recorded in the fjord system 11 weeks after tagging, but only three individuals (one protected and two control fish) were recorded more than five weeks after release.

Table 2 Time from release to last recording at the different receiver sites of acoustically tagged post-smolts in the Romsdalsfjord system in 2003 (Time is given as mean with range and SD in parenthesis. N is number of fish) Site

1 2 3 1 2 3

Salmon lice infected N Time (h) Atlantic salmon 3 11 (8-15, 4) 8 127 (80-365, 96) 4 137 (100-160, 26) Brown trout 5 286 (37-537, 194) 3 390 (79-575, 222) 3 575 (413-736, 162)

Salmon lice protected N Time (h)

Control Time (h)

N

7 8 3

32 (7-56, 17) 92 (65-131, 24) 144 (135-158, 12)

7 6 2

40 (10-101, 29) 67 (34-105, 23) 186 (164-207, 31)

4 2 1

665 (25-2062, 655) 133 (106-160, 39) 636 ( - )

2 2 0

305 (94-516, 299) 643 (632-653, 15)

Rate of progression through the fjord system, from release to last recording measured as body lengths per second, did not differ among the groups in neither Atlantic salmon (Kruskal-Wallis test; χ² = 1.13, p = 0.57) nor brown trout (Kruskal-Wallis test; χ² = 1.22, p = 0.55). Rates of progression (mean ± SD) were for Atlantic salmon 0.41 ± 0.20 bl s-¹ (protected, n = 18), 0.45 ± 0.30 bl s-¹ (control, n = 15) and 0.48 ± 0,20 bl s-¹ (infected, n = 15), and for brown trout 0.17 ± 0.14 bl s-¹ (protected, n = 7), 0.06 ± 0.03 bl s-¹ (control, n = 4) and 0.13 ± 0,15 bl s-¹ (infected, n = 11).

56

SUMBAWS

Final Report

Development of salmon lice in laboratory control groups The infection intensity was high for both Atlantic salmon and brown trout, and 100 % of the fish in the experiment were infected (Table 3). There was no difference in intensity between 15 and 25 days post infection neither in Atlantic salmon nor brown trout (Mann-Whitney U-test, salmon: U = 49.0, p = 0.97, brown trout U = 22.0, p = 0.20, Table 2). After 15 days, the infection was dominated by early chalimus stages and after 25 days, the infection was dominated by mobile stages (Figure 2). No significant differences in intensities were found between the species neither at day 15 nor at day 25 (Mann-Whitney Utests, day 15: U = 12.0, p = 0.1, day 25: U = 22.0, p = 0.20, Table 3) (Figure 2).

100

100

Salmon Day 15

Salmon Day 25 N = 10

N = 10 80

Frequency (%)

Frequency (%)

80

60

40

60

40

20

20

0

0 AD

D A

D A

G G FE

M

F

F P2

M P2

F P1

1

H2

H

G G FE D

M D

F D

M P1

C

C

A

A

A

F P2

M P2

F P1

2

1

M P1

H C

H C

Developmental stage

100

100

Trout

Trout Day 25

Day 15

N=7

N=5 80

Frequency (%)

Frequency (%)

80

60

40

60

40

20

20

0

0 AD

D A

D A

G G FE

M

F

F P2

M P2

F P1

1

H2

H

M P1

C

C

DM

G G FE AD

A

F AD

F P2

M P2

F P1

H2

M P1

C

1 CH

Figure 2 Frequency distribution of developmental stages of salmon lice in laboratory control groups of Atlantic salmon and brown trout post-smolts 15 and 25 days after artificial infection. (CH1 = first and second chalimus stage combined, CH3 = third and fourth chalimus stage combined, P1M = first pre-adult male, P1F = first pre-adult female, P2M = second pre-adult male, P2F = second pre-adult female, ADM = adult male, ADF = adult female, ADFEGG = adult female with egg strings)

57

SUMBAWS

Final Report

Table 3 Recording of salmon lice on laboratory control groups 15 and 25 days after artificial infection.

Days after infection Number of fish Prevalence (%) Intensity (mean ± SD) Inter quartile range (IQR) Median Intensity range Variance to mean ratio (S² χ¯-¹)

Atlantic salmon 15 25 10 10 100 100 43.1 ± 19.1 39.7 ± 6.9 41.8 12.8 41.5 39.5 21-70 31-51 8.9 1.2

Brown trout 15 25 5 7 100 100 24.4 ± 10.2 33.9 ± 7.4 18 16 30.0 34.0 9-33 24-44 4.2 1.6

Discussion Neither artificial salmon lice infection nor the pharmaceutical prophylaxis against salmon lice infection had any effects on survival or rate of progression in fjord migrating Atlantic salmon and brown trout postsmolts. Fast migration through the fjord system (average 6.3 days, maximum 15.2 days), might be the reason for no differences in survival and progression rates between infected and non-infected Atlantic salmon post-smolts. At 15 days post infection, laboratory retained Atlantic salmon and brown trout were only infected with chalimus-larvae, and developmental stages in the released fish were expected to be similar. A possible reinfection of salmon lice in released Atlantic salmon post-smolts was also expected to be dominated by chalimus larvae (Finstad et al., 2000). Heavy chalimus infection has in a series of laboratory experiments mainly been associated with stress responses, whereas more severe alterations in physiological homeostasis and mortality mainly occur when pre-adult and adult stages appear (reviewed by Tully and Nolan, 2002; Heuch et al., 2005). Thus, in the present study, the Atlantic salmon post-smolts had already left the 80 km long Romsdalsfjord system before preadult and adult stages of lice had developed, and started to physiologically compromise the fish more severely (Tully and Nolan, 2002). Accordingly, positive effects of the pharmaceutical prophylaxis against salmon lice infection would not be expected until after the salmon had left the fjord system and entered the open ocean. However, more important, no adverse effects of lice protection were detected in survival or rate of progression on fjord migrating Atlantic salmon post-smolts.

The released brown trout remained to a larger extent than Atlantic salmon in the inner part of the fjord system. This is in accordance with previous telemetry studies (Finstad et al., 2005), as well as other studies of wild Norwegian sea trout populations (Berg and Berg, 1987; Lyse et al., 1998; Knutsen et al., 2001). As in Atlantic salmon, effects of salmon lice infection and protection were expected to increase a few weeks after release. By then, the artificial salmon lice infection would have developed into to the more pathogenic pre-adult and adult stages. By this time, released fish in the infected and control group would also have been susceptible for reinfection by salmon lice copepodids. However, only three individual brown trout were recorded more than five weeks after release. This indicates that either all groups of brown trout suffered a high mortality in the fjord system, or that they stayed in parts of the system not covered by the listening stations. New results also indicate that predation risk can be very high in estuarine and inner fjord areas (Thorstad et al., 2005, Jepsen et al. 2005). Together, this might explain the unexpectedly low recording rate in trout, which usually feed only within a short distance of their native river mouth (Berg and Berg, 1987; Berg and Jonsson, 1990). A few individuals may also have moved to the outer coastal areas, but the lower proportion recorded by the listening stations indicated that they to a lesser extent than salmon migrated through the fjord system, and concur with earlier studies of marine sea trout ecology (see references above). Brown trout post-smolts also tend to move closer to shore than Atlantic salmon postsmolts (Thorstad et al., 2005), such that the likelihood of brown trout being recorded when passing the receivers should be higher than for Atlantic salmon.

In conclusion, the present study indicates that chalimus salmon lice infection have limited effects on Atlantic salmon and brown trout post-smolts fjord survival and migration. Thus, previous laboratory studies indicating only minor effects on post-smolts of chalimus larvae are confirmed under natural field conditions. When the infection had developed into mobile stages, Atlantic salmon post-smolts had most likely left the fjord and near-coastal areas. Unlike brown trout, heavily infected Atlantic salmon post-smolts therefore do not possess the practical capability of returning to freshwater for delousing when encountering severe salmon lice attacks (Bjørn et al., 2001), and might die unnoticed at sea (Holst et al., 2003). Brown trout will to a larger extent stay in the fjord areas when severe infection loads are reached, and can 58

SUMBAWS

Final Report

counteract this by returning to freshwater for delousing. Trout will, on the other hand, be more susceptible to adverse effects of continuous reinfection of salmon lice, because they remain in the fjord and coastal areas with extensive fish farming activity and a high infection pressure (e.g. Tully et al., 1999; Bjørn et al., 2001; Heuch et al., 2005).

References Berg, O. K. and Berg, M. (1987). Migrations of sea trout, Salmo trutta L., from the Vardnes river in northern Norway. Journal of Fish Biology 31, 113-121.

Berg O. K. and Jonsson, B. (1990). Growth and survival rates of the anadromus trout, Salmo trutta, from the Vardnes River northern Norway. Environmental Biology of Fishes 29, 145-154.

Bjørn, P. A. and Finstad, B. (1997). The physiological effects of salmon lice infection on sea trout post smolts. Nordic Journal of Freshwater Research 73, 60-72.

Bjørn, P. A. and Finstad, B. (1998). The development of salmon lice (Lepeophtheirus salmonis) on artificially infected post smolts of sea trout (Salmo trutta). Canadian Journal of Zoology 76, 970-977.

Bjørn, P. A., Finstad, B. and Kristoffersen, R. (2001). Salmon lice infection of wild sea trout and Artic char in marine and freshwaters: the effects of salmon farms. Aquaculture Research 32, 947-962.

Boxaspen, K. and Næss, T. (2000). Development of eggs and planktonic stages of salmon lice (Lepeophtheirus salmonis) at low temperatures. Contributions to Zoology 69, 51-55.

Bush, A. O., Lafferty, K. D., Lotz, J. M. and Shostak, A. (1997). Parasitology meets ecology on its own terms: Margolis et al. revisited. Journal of Parasitology 83, 575-583.

Butler, J. R. A. (2002). Wild salmonids and sea lice infestation on the west coast of Scotland: sources of infection and implications for the managment of marine salmon farms. Pest Managment Science 58, 595608.

Finstad, B., Bjørn, P. A., Grimnes, A. and Hvidsten, N. A. (2000). Laboratory and field investigations of salmon lice [Lepeophtheirus salmonis (Krøyer)] infestation on Atlantic salmon (Salmo salar L.) postsmolts. Aquaculture Research 31, 795-803.

Finstad, B., Økland, F., Thorstad, E. B., Bjørn, P. A. and McKinley, R. S. (2005). Migration of hatchery-reared Atlantic salmon and wild anadromous brown trout post-smolts in a Norwegian fjord system. Journal of Fish Biology 66, 86-96.

Gargan, P. G., Tully, O. and Poole, W. R. (2003). Relationship between sea lice infestation, sea lice production and sea trout survival in Ireland, 1992-2001. In Salmon at the edge (ed. D. Mills), pp. 119-135.

Blackwell Science.

Glover, K. A., Skaala, Ø., Nilsen, F., Olsen, R., Teale, A. J. and Taggart, J. B. (2003). Different susceptibility of anadromous brown trout (Salmo trutta L.) populations to salmon louse (Lepeophtheirus salmonis (Krøyer, 1837)) infection. ICES Journal of Marine Science 60, 1139-1148.

Hansen, L. P., Holm, M., Holst, J. C. and Jacobsen, J. A. (2003). The Ecology of post-smolts of Atlantic salmon. In Salmon at the edge (ed. D. Mills), pp. 25-39. Blackwell Science.

Heuch, P. A., Bjørn, P. A., Finstad, B., Holst, J. C., Asplin, L. and Nilsen, F. (2005). A review of the Norwegian ’National action plan against salmon lice on salmonids’: The effect on wild salmonids.

Aquaculture 246, 79-92.

Holst, J. C., Jacobsen, P., Nilsen, F., Holm, M., Asplin, L. and Aure, J. (2003). Mortality of seawardmigrating post-smolts of Atlantic salmon due to salmon lice infection in Norwegian salmon stocks. In Salmon at the edge (ed. D. Mills), pp. 136-137. Blackwell Science.

Jepsen, N., Holthe, E. and Økland, F. (2005). Saithe and cod as specialised smolt predators. Journal of Fish Biology, submitted.

Johnson, S. C. and Albright, L. J. (1991). Development, growth and survival of Lepeophtheirus salmonis (Copepoda: caligidae) under laboratory conditions. Journal of the Marine Biological Association of the United Kingdom 71, 425-436.

59

SUMBAWS

Final Report

Klemetsen, A., Amundsen, P.-A., Dempson, J. B., Jonsson, B., Jonsson, N., O’Connell, M. F. and Mortensen, E. (2003). Atlantic salmon Salmo salar L., brown trout Salmo trutta L. and Artic charr Salvelinus alpinius (L.): a review of aspects of their life histories. Ecology of Freshwater Fish 12, 1-59.

Knutsen, J. A., Knutsen, H., Gjøsæter, J. and Jonsson, B. (2001). Food of anadromous brown trout at sea. Journal of Fish Biology 59, 533-543.

Lyse, A. A., Stefansson, S. O. and Fernö, A. (1998). Behaviour and diet of sea trout post-smolts in a Norwegian fjord system. Journal of Fish Biology 52, 923-936.

Nolan, D. T., Reilly, P. and Wendelaar Bonga, S. E. (1999). Infection with low numbers of the sea louse Lepeophtheirus salmonis induces stress-related effects in postsmolt Atlantic salmon (Salmo salar).

Canadian Journal of Fisheries and Aquatic Sciences 56, 947-959.

Rikardsen, A. H. (2004). Seasonal occurrence of sea lice Lepeophtheirus salmonis on sea trout in two north Norwegian fjords. Journal of Fish Biology 65, 711-722.

Thorstad, E. B., Økland, F., Finstad, B., Sivertsgård, R., Bjørn, P. A. and McKinley, R. S. (2004).

Migration speeds and orientation of Atlantic salmon and sea trout post-smolts in a Norwegian fjord system.

Environmental Biology of Fishes 71, 305-311.

Thorstad, E. B., Økland, F., Finstad, B., Sivertsgård, R., Plantalech, N., Bjørn, P. A. and McKinley, R. S. (2005). Comparing migratory behaviour and survival of wild and hatchery-reared Atlantic salmon and wild anadromous brown trout postsmolts during the first stages of marine migration. Hydrobiologia, submitted.

Tully, O. and Nolan, D. T. (2002). A review of population biology and host-parasite interactions of the sea lice Lepeophtheirus salmonis (Copepoda: Caligidae). Parasitology 124, 165-182.

Tully, O., Gargan, P., Poole, W. R. and Whelan, K. F. (1999). Spatial and temporal variation in the infestation of sea trout (Salmo trutta L.) by the caligid copepod Lepeophtheirus salmonis (Krøyer) in relation to sources of infection in Ireland. Parasitology 119, 41-51.

Wagner, G. N., McKinley, R. S., Bjørn, P. A. and Finstad, B. (2003). Physiological impact of sea lice on swimming performance of Atlantic salmon. Journal of Fish Biology 62, 1000-1009.

60

SUMBAWS

Final Report

Paper 4: COMPARING MIGRATORY BEHAVIOUR AND SURVIVAL OF WILD AND HATCHERY-REARED ATLANTIC SALMON AND WILD ANADROMOUS BROWN TROUT POST-SMOLTS DURING THE FIRST STAGES OF MARINE MIGRATION E.B. Thorstad, F. Økland, B. Finstad, R. Sivertsgård, N. Plantalech, P.A. Bjørn and R.S. McKinley (Submitted to Hydrobiologia) Introduction At the end of the freshwater stage, Atlantic salmon Salmo salar L. and anadromous brown trout Salmo trutta L. juveniles transform physiologically and morphologically and migrate to sea as smolts (e. g.

Wedemeyer et al., 1980; Hoar, 1988; Finstad and Jonsson, 2001). Atlantic salmon migrate to the open ocean for feeding, whereas the brown trout remain feeding in the fjord and coastal areas (e.g. Jonsson, 1985; Hansen et al., 2003; Klemetsen et al., 2003; Rikardsen, 2004). The understanding of the early marine phase of the Atlantic salmon and the environmental factors that may influence their behaviour and distribution in the sea is limited (Moore et al., 2000), and even less is known about the brown trout. This lack of information is particularly critical because the heaviest mortality of salmon in the sea apparently takes place during the first months after the smolts leave fresh water (Hansen et al., 2003).

The first stages of the marine migration of post-smolts can be monitored using telemetry techniques (Moore et al., 2000; Lacroix and Voegeli, 2000). For Atlantic salmon, the method has been limited by the relatively large size of acoustic transmitters, resulting in most studies investigating behaviour of larger hatchery-reared posts-molts (e.g. LaBar et al., 1978; Lacroix and McCurdy, 1996; Finstad et al., 2005; Thorstad et al., 2004). However, the behaviour of hatchery-reared post-smolts may differ from that of wild post-smolts (e.g. Jonsson et al., 1991), and as pointed out by Lacroix and McCurdy (1996), there is a general need to obtain information on the movements of normally sized wild Atlantic salmon post-smolts.

A limited number of miniature acoustic transmitters were produced for this project (Thelma, Norway), allowing tagging of wild Atlantic salmon posts-molts. The objective was to compare the rate of progression and survival of wild Atlantic salmon post-smolts with hatchery-reared Atlantic salmon and wild brown trout post-smolts in a Norwegian fjord system. It was hypothesised that the rate of progression of Atlantic salmon post-smolts through the fjord system was not dependent on their origin (as suggested by Lacroix and McCurdy, 1996; Voegeli et al., 1998; Lacroix et al., 2004). Furthermore, it was hypothesised that survival did not differ between wild and hatchery-reared Atlantic salmon (as suggested by Hvidsten and Lund, 1988). It was also hypothesised that the sea trout post-smolts to a larger extent than the wild Atlantic salmon post-smolts remained in the inner part of the fjord system (as suggested by Finstad et al., 2005).

Materials and methods Study area The study was carried out in the Romsdalsfjord system on the West Coast of Middle Norway, 62º40’N 8º10’E (Figure 1, described in Finstad et al., 2005).

61

SUMBAWS

Final Report

Figure 1 The Romsdalsfjord system in Middle Norway with the release site for acoustically tagged Atlantic salmon and anadromous brown trout in 2004 and automatic receiver sites 9.5 km (4 receiver sites), 37.0 km (7 receiver sites) and 65.0 km (2 receiver sites) from the release site.

Handling, tagging and release of fish Wild (n = 43) and hatchery-reared (n = 71) Atlantic salmon and wild brown trout (n = 34) smolts were tagged with coded acoustic transmitters (Table 1). The smallest 27 wild salmon were tagged with transmitters from Thelma, Norway, model THS-04 (7 × 19 mm, mass in air of 1.9 g, mass in water of 1.2 g, guaranteed life time 28 days), whereas the remaining smolts were tagged with transmitters from VEMCO Ltd., Canada, model V8SC-6L-R256 (9 × 20 mm, mass in air of 3.3 g, mass in water of 2.0 g, guaranteed life time 70 days). The tag to body mass ratio was larger in wild than in hatchery-reared salmon (MannWhitney U-test, p < 0.001, Table 1). All transmitters were on frequency 69 kHz and had a power output of 139 db re 1uPa@1 m.

The wild Atlantic salmon and sea trout were caught in a trap close to the river mouth during their downstream migration in the River Eira, transported to the Statkraft hatchery in Eresfjord in a tank with oxygenated water and tagged after 0-5 days. The hatchery-reared Atlantic salmon were two-year old smolts from the Statkraft hatchery with wild salmon from the River Eira as parents. The fish were tagged during 6 May - 4 June 2004, using methods described in Finstad et al. (2005). Water temperature in the hatchery was 2.7-7.0 ºC. A seawater tolerance test (Blackburn and Clarke, 1987) performed on the hatchery-reared salmon on 6 May revealed plasma chloride levels at 147 mM, indicative that the smolts were ready to be released into seawater (Sigholt and Finstad, 1990).

Groups of tagged smolts were transported in water filled plastic bags to a cage (volume 0.5 m³) in the fjord at the mouth of the River Eira 0-2 days after tagging. The fish were kept in the cage for 2-21 hours prior to release (Figure 1, Finstad et al., 2005). Shoals of wild and Carlin-tagged Atlantic salmon and brown trout post-smolts were frequently seen at the release site during the study period. All fish were released during high tide in daylight (from 3 h before to 1 h after the highest tide).

62

SUMBAWS

Final Report

Table 1 Atlantic salmon (wild and hatchery-reared) and anadromous brown trout (wild) post-smolts tagged with acoustic transmitters and released at the mouth of the River Eira in the Romsdalsfjord system in 2004.

Group

Number tagged

Wild salmon

43

Hatcheryreared salmon

71

Wild trout

34

Mean body length (LT) mm (range, SE) 152 (136-173, 1.2) 198 (160-245, 2.6) 171 (143-242, 4.6)

Mean body Mean transmitter mass Release dates mass in water to fish mass g (range, SE) ratio (%) (range, SE) 25 6.0 7, 8, 9, 12, 20 and 22 (18-38, 0.66) (4.0-8.0, 0.17) May 70 3.1 7, 8, 9, 12, 20 and 31 (37-124, 2.7) (1.6-5.4, 0.12) May 42 5.6 7, 8, 9, 12, 20, 22, 26 (23-114, 3.8) (1.8-8.7, 0.31) and 31 May, 6 June

Recording of fish after release Fish were recorded when passing 19 individual receiver/data loggers (VEMCO model VR2) moored at sites 9.5, 37.0 and 65.0 km from the fish release site at the estuary of the River Eira (Figure 1). At 9.5 km from the release site, where the fjord is 1.48 km wide, four moorings were evenly distributed across the fjord. Receivers were attached to each mooring on 3 m depth. An additional receiver was attached to two of the moorings on 10 m depth (site 1 and 3, Figure 1). At 37.0 km from the release site, where the fjord is 2.60 km wide, seven moorings were evenly distributed across the fjord. Receivers were attached to each mooring on 5 m depth. An additional receiver was attached to four of the moorings on 10 m depth (site 1, 3, 5 and 7, Figure 1). At 65.0 km from the release site, where the fjord is 2.46 km wide, one receiver was attached on each side of the fjord at 5 m depth. The receiver at site 1 (Figure 1) disappeared before data were downloaded.

The sea depth at receiver moorings was 24 to 288 m. Receiver range for detecting signals from transmitters was typically a radius of 200-260 m for transmitters at 0.5-3.0 m depth, but varied considerably (radius of 45-620 m during range tests) with factors such as wave action and salinity. Thus, based on range tests it is likely that nearly all fish were recorded when they passed receiver sites at 9.5 and 37.0 km, since 185 m was the longest distance from a receiver possible to pass when passing the receiver sites. This is supported by the fact that only one of 43 fish (2 %) that was recorded in the outer areas had not been recorded when passing receiver sites further in (see results). Furthermore, the fish passing were in nearly all cases recorded by several of the receivers and with a high number of recordings per fish, indicating a good coverage by the receivers (unpublished results). The receivers collected data until 24 or 25 July 2004.

Manual tracking was carried out in the area around the mouth of the River Eira on 27 July 2004.

Data analysis Data collected did not meet the assumptions for parametric tests, and therefore, non-parametric tests were employed. Rates of progression in terms of body lengths per second are based on time from release to first recording at the different receiver sites. This must not be regarded as actual swimming speeds, but as minimum speeds, since the fish most likely did not follow the shortest migration routes (Thorstad et al., 2004). For analysis of where in the fjord the fish passed, only recordings by receivers at 3 m depth were used 9.5 km from the release site, and at 5 m depth 37.0 km from the release site, for standardisation.

Results Water temperature and salinity Water temperature and salinity were measured at eight sites in the middle of the fjord 1-60 km from the river mouth on 4, 14 and 30 May and 11 June at 1 m depth. Water temperature varied between 7.5 and 13.8ยบC (mean 10.8). Salinity varied between 7.7 and 31.9 ppt (mean 25.5).

63

SUMBAWS

Final Report

Proportions of fish recorded after release There was no difference in the proportions recorded between wild and hatchery reared salmon neither 9.5 km nor 37.0 km from the release site (Fisher's exact test, 9.5 km: p = 0.85, 37.0 km: p = 1.00, Table 2).

However, the proportion of trout recorded was smaller than the proportion of salmon both places (Fisher's exact test, 9.5 km: p = 0.001, 37.0 km: p = 0.002, Table 2). There were no differences in body length or transmitter/body mass ratio between individuals recorded and those not recorded neither 9.5 nor 37.0 km from the receiver site for any of the groups (wild salmon, hatchery-reared salmon and trout, Mann-Whitney U-tests, p = 0.17-0.92), except that wild salmon recorded 37.0 km from the release site were larger than those not recorded (p = 0.039).

Only one fish recorded 37.0 km from the release site was not recorded when passing the receivers 9.5 km from the release site (a wild salmon). The three hatchery-reared salmon recorded 65.0 km from the release site (Table 2) were all recorded when passing receivers both 9.5 and 37.0 km from the receiver site.

On 27 July, transmitters from 6 wild salmon (14 %), 6 hatchery-reared salmon (8 %) and 7 trout (21 %) were located in the area close to the river mouth by manual tracking. One transmitter that had belonged to a hatchery-reared salmon (body size 171 mm, 43 g) was found in a saithe Pollachius virens L. that was caught by an angler in this area.

Table 2 Number and proportion of Atlantic salmon and anadromous brown trout post-smolts recorded at receiver sites 9.5 km, 37.0 and 65.0 km from the release site in the Romsdalsfjord system in 2004.

Group Wild salmon Hatchery-reared salmon Wild trout

Number (%) recorded 9.5 km from the release site 25 (58 %) 39 (55 %) 8 (24 %)

Number (%) recorded 37.0 km from the release site 15 (35 %) 25 (35 %) 3 (9 %)

Number (%) recorded 65.0 km from the release site 0 3 (4 %) 0

Migration routes and patterns Nearly all fish were recorded by more than one of the receivers when passing 9.5 and 37.0 km from the release site. For each fish, the receiver with the highest number of recordings was used to indicate where in the fjord the fish had passed (Table 3 and 4), although exact distance to shore cannot be determined. At 9.5 km from the release site, the wild and hatchery-reared salmon were recorded in the middle of the fjord as well as closer to the shore, whereas the trout were recorded closer to shore (Table 3). Combining data for all groups showed a tendency that more post-smolts passed on the south-western side of the fjord than on the north-eastern side (Table 3). Also 37.0 km from the release site, the salmon were recorded both in the middle of the fjord and along the shore, whereas the trout stayed closer to shore (Table 4). There was no indication that the fish passed more often on one side of the fjord than on the other side, except that there were few recordings on the receiver closest to shore on the northern side of the fjord (Table 4).

Most of the salmon followed a straight route outward the fjord. The exceptions were one wild salmon and five hatchery-reared salmon (no different proportion between the groups, Fisher's exact test, p = 0.38) recorded 9.5 km from the release site again after they had been recorded 37.0 km from the release site. None of the three trout recorded at 37.0 km were returning later to 9.5 km from the release site.

Table 3 Number of tagged Atlantic salmon and anadromous brown trout post-smolts with the highest number of recordings at the different receiver sites 9.5 km from the release site in the Romsdalfjord system in 2004 (Receiver site numbers correspond to those in Figure 1) Group Wild salmon Hatchery-reared salmon Trout Total

Site 1 7 9 5 21

Site 2 7 17 0 24

Site 3 3 9 0 12

Site 4 8 4 3 15

64

SUMBAWS

Final Report

Table 4 Number of tagged Atlantic salmon and anadromous brown trout post-smolts with the highest number of recordings at the different receiver sites 37.0 km from the release site in the Romsdalfjord system in 2004 (Receiver site numbers correspond to those in Figure 2.6. Two fish had an equal number of recordings at two sites, which is the reason for 'half of a fish' being recorded at some sites) Group Wild salmon Hatchery-reared salmon Trout Total

Site 1 4 2 0 7

Site 2 1 4 1 6

Site 3 3.5 5.5 0 9

Site 4 1 3.5 0 4.5

Site 5 1.5 6 0 7.5

Site 6 4 2 2 8

Site 7 0 1 0 1

Rates of progression There was no difference between wild and hatchery-reared salmon neither in time from release to first recording 9.5 km from the release site (Mann-Whitney U-test, p = 0.11), nor in the rate of progression in terms of body lengths per second (Mann-Whitney U-test, p = 0.43) (Table 5). At 37.0 km from the release site, the hatchery-reared salmon were recorded a shorter time after release than the wild salmon (MannWhitney U-test, p = 0.033), but the rate of progression in terms of body lengths per second did not differ between the groups (Mann-Whitney U-test, p = 0.18) (Table 5). Hatchery-reared salmon post-smolts recorded at 37.0 km from the release site (n = 25) had a faster migration to 9.5 km from the release site than those not recorded at 37.0 km (n = 14) (Mann-Whitney U-test, p = 0.038), indicating a higher loss of the slowest migrating fish. Such a difference was not found for wild salmon post-smolts (n = 14 recorded and 11 not recorded, p = 0.27).

The trout spent longer time from release to first recording 9.5 km from the release site than the salmon (wild and hatchery-reared salmon combined, Mann-Whitney U-test, p < 0.001, Table 3). Also the rate of progression in terms of body lengths per second was slower for the trout than for the salmon (MannWhitney U-test, p < 0.001, Table 5). At 37.0 km from the release site, only three trout were recorded (Table 5).

Table 5 Time and rate of progression for tagged Atlantic salmon and anadromous brown trout post-smolts from release to first recording at receiver sites 9.5 km, 37.0 and 65.0 km from the release site in the Romsdalsfjord system in 2004.

First recordings at 9.5 km First recordings at 37.0 km First recordings at 65.0 km from release site from release site from release site Mean rate of Mean rate of Mean rate of Mean time Mean time Mean time progression progression progression Group [h] [h] [h] [bl s-1] [bl s-1] [bl s-1] (range, SE) (range, SE) (range, SE) (range, SE) (range, SE) (range, SE) 135 0.53 450 0.56 Wild salmon (9-667, 41) (0.03-1.88, 0.10) (35-1493, 133) (0.04-1.89, 0.14) Hatchery80 0.56 168 0.77 154 0.64 reared salmon (7-949, 27) (0.01-1.62, 0.07) (26-670, 40) (0.08-1.85, 0.12) (82-201, 37) (0.40-1.00, 0.18) 350 0.06 1309 0.05 Wild trout (134-646, 70) (0.02-0.12, 0.01) (1199-1528, 109) (0.04-0.06, 0.01)

Discussion This study indicates that results and conclusions from previous studies of behaviour of hatchery-reared Atlantic salmon post-smolts (e.g. LaBar et al., 1978; Lacroix and McCurdy, 1996; Finstad et al., 2005; Thorstad et al., 2004) may be representative for wild Atlantic salmon. The proportions of fish recorded 9.5 and 37.0 km from the river mouth did not differ between wild and hatchery-reared post-smolts. Similarly, time from release to first recording and rate of progression in terms of body lengths per second from release to 9.5 km from the river mouth did not differ. The only difference found was that time from release to first recording 37.0 km from the river mouth was shorter for hatchery-reared than for wild fish. However, this difference was attributed to the larger body size of the hatchery-reared salmon, since the rate of progression in terms of body lengths per second was not different between the two groups. It is therefore concluded that 65

SUMBAWS

Final Report

no behavioural differences between wild and hatchery-reared salmon post-smolts were detected, but that time from release to reaching 37.0 km from the release site was dependent on the body size.

The rates of progression for the salmon (0.53-0.77 body lengths per second) were similar to previously recorded for larger hatchery-reared salmon in the same fjord system (mean body length 29.5 cm, rates of progression 0.54-0.69 body lengths per second, Finstad et al., 2005). Hence, the rates of progression in terms of body lengths per second are consistent between years and for different size groups of post-smolts.

Higher ground speeds have been recorded for Atlantic salmon post-smolts in estuaries (e.g. Moore et al., 1992; 1995; LaBar et al., 1978; Fried et al., 1978) and in Passamaquoddy Bay in Canada (Lacroix and McCurdy, 1996), probably at least partly due to higher water velocities in these localities. Ground speed depends on water velocities and by which methods the rates of progression are recorded, and may therefore be difficult to compare among studies.

Swimming speeds have also been recorded in the same fjord system as the present study for hatcheryreared Atlantic salmon post-smolts by continuous manual tracking, with post-smolts swimming at a mean ground speed of 1.27 body lengths per second over 10-min periods (Thorstad et al., 2004). If these ground speeds are representative, it must be concluded that the salmon post-smolts did not follow the straightest route during the first 9.5 and 37.0 of the sea migration in the present study, but had been actually swimming approximately twice the distance. However, it must be pointed out that the individual variation in the rate of progression, and thereby likely in the migration pattern, was large. Also, some individuals were returning to 9.5 km from the release site after they had been recorded at 37.0 km, indicating that some salmon postsmolts may move around in the fjord system for some time. According to Thorstad et al. (2004), the postsmolts were actively swimming (at a speed of 1.32 body lengths per second when corrected for the speed and direction of the water current) and not only drifting passively with the current.

The trout had a slower rate of progression outward the fjord than the salmon, and a smaller proportion of the brown trout post-smolts was recorded both 9.5 and 37.0 km from the river mouth. Thus, the trout remained a longer period in the inner part of the fjord system than both the wild and hatchery-reared Atlantic salmon, although some individuals also reached the outer fjord areas (37.0 km from the river mouth) within the two month study period. The results are in accordance with a previous study on anadromous brown trout (Finstad et al., 2005). The longer time spent in the inner part of the fjord system, may expose the trout to a higher predation pressure than the Atlantic salmon, if the predation pressure is higher in inner than in outer fjord areas (see below).

Both the wild and hatchery-reared Atlantic salmon post-smolts migrated in the middle of the fjord as well as closer to the shore, with a large individual variation. However, the results indicated that there may be some patterns in fjord migration routes, with more fish being recorded on one side of the fjord than on the other side, but it is not known whether this corresponded with for instance current patterns in the fjord system.

Provided all passing fish were recorded at the receiver sites, the mortality of the salmon post-smolts was 40-45% during the 9.5 first km (the individual not recorded at 9.5 km but at 37.0 is included), and 36-42 % during the next 27.5 km, with a total mortality during the first 37.0 km of the sea migration of 65 %. These proportions must be regarded as maximum mortality compared to non-tagged post-smolts, because a few fish may theoretically have passed receivers without being recorded, and because the mortality may be increased due to handling and tagging (see below). However, the results correspond with reports of high predation rates in other Norwegian estuaries by cod Gadus morhua L. and saithe. In the estuary of the River Orkla, predation of cod was estimated to 20 % (Hvidsten and Lund, 1988), and in a small area in the estuary of the River Surna to 25 % (Hvidsten and Møkkelgjerd, 1987). A large number of saithe with wild untagged post-smolts in their stomach were caught close to the river mouth during the present study, suggesting a high loss of post-smolts due to predation (Jepsen et al., 2005). The 12 salmon transmitters located in this area in late July were probably from fish being eaten by saithe or cod, and the transmitter later being ejected from the stomach of the predator. Thus, the first phase of the marine migration seems to imply a heavy mortality on the salmon post-smolts, as suggested by Hansen et al. (2003).

Hvidsten and Lund (1988) found no difference in predation rates between wild and hatchery-reared salmon smolts and no evidence of a selective predation on the smallest wild and hatchery-reared smolts.

This corresponds with the results from proportions of fish recorded in the present study, except that there was an indication of a higher survival to 37 km from the release site for larger than for smaller wild salmon smolts. Hatchery-reared post-smolts recorded at 37.0 km from the release site had a faster migration to 9.5 km from the release site than those not recorded at 37.0 km, indicating a higher loss among the slowest migrating fish. Thus, the risk of being predated may increase with time spent in the fjord system. An alternative explanation may be that the fish with the lowest migration speeds were the weakest fish, and 66

SUMBAWS

Final Report

therefore more likely to be predated.

Handling and tagging may influence the behaviour of the fish. For Atlantic salmon post-smolt studies in the sea, Moore et al. (2000) recommended tags to be less than 5 % of fish mass to minimise effects on behaviour and survival. Some individuals had larger tags in the present study, and the wild post-smolts had relatively larger transmitters than the hatchery-reared post-smolts. According to Peake et al. (1997), wild salmon smolts may also be generally more affected by tagging than hatchery-reared smolts. The most likely potential negative effects of transmitters in the present study are higher mortality and lower swimming speeds for the smallest fish (e.g. McCleave and Stred, 1975), although transmitters weighing up to 12 % of the body mass did not affect the swimming performance of juvenile rainbow trout (Oncorhynchus mykiss) (Brown et al., 1999). No difference in tag to body mass ratio between fish recorded or not recorded after release was found in the present study, indicating that potential negative effects of the tagging did not affect the conclusions of the study. However, if conclusions were affected it was most likely by overestimating mortality and underestimating swimming speeds for the wild compared to hatchery-reared post-smolts.

References Blackburn, J. and Clarke, W. C. (1987). Revised procedure for the 24 hour seawater challenge test to measure seawater adaptability of juvenile salmonides. Canadian Technical Report of Fisheries and Aquatic Sciences 1515, 1-39.

Brown, R. S., Cooke, S. J., Anderson, W. G. and McKinley, R. S. (1999). Evidence to challenge the "2% rule" for biotelemetry. North American Journal of Fisheries Management 19, 867-871.

Finstad, B. and Jonsson, N. (2001). Factors influencing the yield of smolt releases in Norway. Nordic Journal of Freshwater Research 75, 37-55.

Finstad, B., Økland, F., Thorstad, E. B., Bjørn, P. A. and McKinley, R. S. (2005). Migration of Atlantic salmon and sea trout post-smolts in a Norwegian fjord system. Journal of Fish Biology 66, 86-96.

Fried, S. M., McCleave, J. D. and LaBar, G. W. (1978). Seaward migration of hatchery-reared Atlantic salmon, Salmo salar, smolts in the Penoboscot River estuary, Maine: riverine movements. Journal of the Fisheries Research Board of Canada 35, 76-87.

Hansen, L. P., Holm, M., Holst, J. C. and Jacobsen, J. A. (2003). The Ecology of post-smolts of Atlantic salmon. In Salmon at the edge (ed. D. Mills), pp. 25-39. Blackwell Science.

Hoar, W. S. (1988). The physiology of smolting salmonids. In Fish physiology: The physiology of developing fish. Viviparity and posthatching juveniles, Volume XIB (eds. W. S. Hoar and D. J. Randall), pp.

275-343. Academic Press.

Hvidsten, N. A. and Lund, R. A. (1988). Predation on hatchery-reared and wild smolts of Atlantic salmon, Salmo salar L., in the estuary of River Orkla. Journal of Fish Biology 33, 121-126.

Hvidsten, N. A. and Møkkelgjerd, P. I. (1987). Predation on salmon smolts, Salmo salar L., in the estuary of the River Surna. Journal of Fish Biology 30, 273-280.

Jepsen, N., Holthe, E. and Økland, F. (2005). Saithe and cod as specialised smolt predators. Journal of Fish Biology, submitted.

Jonsson, B. (1985). Life history patterns of freshwater resident and sea-run migrant brown trout in Norway.

Transactions of the American Fisheries Society 114, 182-194.

Jonsson, B., Jonsson, N. and Hansen, L. P. (1991). Differences in life history and migratory behaviour between wild and hatchery-reared Atlantic salmon in nature. Aquaculture 98, 69-78.

Klemetsen, A., Amundsen, P.-A., Dempson, J. B., Jonsson, B., Jonsson, N., O'Connell, M. F. and Mortensen, E. (2003). Atlantic salmon Salmo salar L., brown trout Salmo trutta L. and Arctic charr Salvelinus alpinus (L.): a review of aspects of their life histories. Ecology of Freshwater Fish 12, 1-59.

LaBar, G. W., McCleave, J. D. and Fried, S. M. (1978). Seaward migration of hatchery-reared Atlantic salmon (Salmo salar) smolts in the Penobscot River estuary, Maine: Open water movements. Journal du Conseil International pour l' Exploration de la Mer 38, 257-269.

67

SUMBAWS

Final Report

Lacroix, G. L. and McCurdy, P. (1996). Migratory behaviour of post-smolt Atlantic salmon during initial stages of seaward migration. Journal of Fish Biology 49, 1086-1101.

Lacroix, G. L. and Voegeli, F.A. (2000). Development of automated monitoring systems for ultrasonic transmitters. In Advances in fish telemetry, Proceedings of the third conference on fish telemetry in Europe (eds. A. Moore and I. Russell), pp. 37-50. CEFAS, Lowestoft.

Lacroix, G. L., McCurdy, P. and Knox, D. (2004). Migration of Atlantic salmon post-smolts in relation to habitat use in a coastal system. Transactions of the American Fisheries Society 133, 1455-1471.

McCleave, J. D. and Stred, K. A. (1975). Effect of dummy telemetry transmitters on stamina of Atlantic salmon (Salmo salar) smolts. Journal of the Fisheries Research Board of Canada 32, 559-563.

Moore, A., Lacroix, G. L. and Sturlaugsson, J. (2000). Tracking Atlantic salmon post-smolts in the sea.

In The ocean life of Atlantic salmon - environmental and biological factors influencing survival (D. Mills ed.), pp. 49-64. Fishing News Books.

Moore, A., Potter, E. C. E. and Buckley, A. A. (1992). Estuarine behaviour of migrating Atlantic salmon (Salmo salar L.) smolts. In Wildlife telemetry (eds. I. M. Priedeand S. M. Swift), pp. 390-399. Ellis Horwood Limited.

Moore, A., Potter, E. C. E., Milner, N. J. and Bamber, S. (1995). The migratory behaviour of wild Atlantic salmon (Salmo salar) smolts in the estuary of the River Conwy, North Wales. Canadian Journal of Fisheries and Aquatic Sciences 52, 1923-1935.

Peake, S., McKinley, R. S., Scruton, D. A. and Moccia, R. (1997). Influence of transmitter attachment procedures on swimming performance of wild and hatchery-reared Atlantic salmon smolts. Transactions of the American Fisheries Society 126, 707-714.

Rikardsen, A. H. (2004). Seasonal occurrence of sea lice Lepeophtheirus salmonis on sea trout in two north Norwegian fjords. Journal of Fish Biology 65, 711-722.

Sigholt, T. and Finstad, B. (1990). Effect of low temperature on sea-water tolerance in Atlantic salmon (Salmo salar) smolts. Aquaculture 84, 167-172.

Thorstad, E. B., Økland, F., Finstad, B., Sivertsgård, R., Bjørn, P. A. and McKinley, R. S. (2004).

Migration speed and orientation of Atlantic salmon and sea trout post-smolts in a Norwegian fjord system.

Environmental Biology of Fishes 71, 305-311.

Voegeli, F. A., Lacroix, G. L. and Anderson, J. M. (1998). Development of miniature pingers for tracking Atlantic salmon smolts at sea. Hydrobiologia 371/372, 35-46.

Wedemeyer, G. A., Saunders, R. L. and Clarke, W. C. (1980). Environmental factors affecting smoltification and early marine survival of anadromous salmonids. Marine Fisheries Review 42, 1-14.

68

SUMBAWS

Final Report

Paper 5: SAITHE AND COD AS SPECIALISED SMOLT PREDATORS N. Jepsen, E. Holthe and F. Økland (Submitted Fisheries Management and Ecology) Introduction The predation on seaward migrating salmonid smolts can be a major regulating factor for local populations of salmon and sea trout. High predation rates on smolts in freshwater have been reported from pike (Esox lucius L.), pikeperch (Sander lucioperca L.), burbot (Lota lota L.) and northern pikeminnow (Ptychocheilus oregonensis Richardson) (e.g. Larsson, 1985; Rieman et al., 1991; Jepsen et al., 1998). In freshwater, and especially in the transition zone from river to the sea, avian predators have been reported to be of importance (Reitan et al., 1987; Schreck et al., 2002; Dieperink et al., 2002). Other studies have shown high levels of predation on smolts from marine fishes like saithe, cod and pollack (Gadus pollachius L.) (e.g. Hvidsten and Møkkelgjerd, 1987; Hvidsten and Lund, 1988). Previous studies from Norwegian rivers have estimated the smolt predation based on analyses of stomach content of potential fish predators in the estuaries (Hvidsten and Møkkelgjerd, 1987; Hvidsten and Lund, 1988; Hvidsten and Johnsen, 1993), and the predation from cod and saithe was found to be up to 25 % in a small area in the estuary of River Surna and 20 % by cod alone in the Orkla estuary. Estimating the specific predation/mortality rate of smolts entering saltwater, is difficult due to varying conditions (DeAngelis and Petersen, 2001), and large year-toyear variations must be expected.

During a study of the behaviour and migration of salmon smolts from the River Eira, entering the sea, a number of smolts were tagged with acoustic transmitters and tracked during their first hours at sea (for details see Thorstad et al., 2004). Several of the tagged smolts were eaten by gulls or predatory fish within few hours after descent. Based on visual observations as well as analyses of stomach content, this paper aims to provide additional evidence for the complicated predator-prey relationship that influences survival of salmonid smolts entering the sea and raise the question of predator specialisation.

Materials and methods The River Eira of western Norway (N62o 41` E8o 07`) holds wild populations of Atlantic salmon (Salmo salar L.) and sea trout (Salmo trutta L.), but due to extensive hydropower development, compensatory stockings of hatchery-reared fish have been carried out for decades. The number of stocked hatchery-reared Atlantic salmon and sea trout smolts by far exceeds the number of wild and the total mass of hatchery-reared smolts leaving the river exceeds the mass of wild fish by almost an order of magnitude.

Visual observations of smolt behaviour can provide valuable information (Bakshtanskiy et al., 1980), but is usually not easily carried out. At the mouth of the River Eira, conditions allowed for very detailed observations of not only the behaviour of smolts moving from freshwater into full saltwater, but also of the predators taking advantage of the apparent vulnerability of the smolts. Schools of typically 20 – 50 hatchery-reared and wild smolts were visible, swimming around in the river mouth close to the shore and larger groups could be seen in the local boat harbour.

During the study period (May – June) several hours of boat-tracking of tagged fish were performed every day close to or at the river mouth. The water was clear making it possible to see individual smolts at depths < 2-3 m. Visual observations indicated that when a school of smolts left the river, gulls attacked them from above, especially when they were in shallow water over tidal sand flats reaching out 200 – 300 m from shore, whereas cod and saithe attacked the schools as soon as they moved out over the edge to deeper water.

Typically, cod were charging the smolts from beneath whereas saithe chased the smolts at high speed close to the surface. Often a smolt chased by saithe was seen jumping several times out of the water. This attracted the attention of gulls and most times such chases ended with the smolt being taken by a gull. Due to a significant difference in size (Atlantic salmon smolts: mean 23 vs 79 g) and morphology, it was possible to distinguish between wild and hatchery smolts and in most of the smolt groups observed, both types were present. Fishing with artificial lures along the drop-off during the short periods of predator activity was performed in order to investigate the amount of smolts eaten by predators.

69

SUMBAWS

Final Report

Results Fishing with artificial lures along the drop-off during the short periods of predator activity, produced almost constant strikes and a number of saithe and cod were caught this way. The fish were killed and the stomach screened for smolts. Only larger items were recorded and unidentifiable remnants were not further studied. From 27 May to 2 June, 57 cod and saithe were caught (Table 1) of which, 27 had remains of 1 â&#x20AC;&#x201C; 8 smolts in the stomach.

Table 1 Number and sizes of the captured predators, and their stomach content of smolts. Only the least digested of the smolts could be identified (Atlantic salmon or trout; hatchery-reared or wild).

Number of predators caught Predator size (cm) Total number of smolts found in stomach Species Origin

Saithe 39 59 - 86 72 8 trout, 19 salmon 11 hatchery, 14 wild

Cod 19 49 - 65 17 1 trout, 14 salmon 11 hatchery, 4 wild

In these 27 stomachs a total of 89 smolts were identified, giving a mean of 3.29 smolts per predator.

Forty of the smolts (8 trout, 32 salmon) were still in a shape where it was possible to distinguish wild fish from hatchery-reared (larger and adipose fin-clipped). In total 22 of the eaten smolts were hatchery-reared and 18 were wild. As it appears from Table 1, the results indicate that cod ate less wild than hatchery-reared smolts, whereas the opposite was the case for the saithe. Most of the cod, contained other food items like crab (Carcinus maenas L.), herring (Clupea harengus L.) or sprat (Sprattus sprattus L.), whereas all saithe caught at the river mouth contained smolts or nothing. This could be seen as an example of predators specialising in pursuing one type of prey. Such specialisation has also been observed in reservoirs, where the majority of the female pikeperch and a few female pike seemed to have adjusted their behaviour to hunt smolts during the smolt run (Jepsen et al., 2000). Similarly, Hvidsten and Lund (1988) found that the frequency of smolts in saithe stomachs was 73 % higher than in cod stomachs.

The observations indicated that a substantial part of the predation took place in restricted area during restricted periods. This area was the very drop off, where the depth increases from less than 1 m to > 25 m.

The time with most (visible) predation activity was during low tide at the darkest hours.

Discussion In river Eira, a study to estimate the run of wild salmon smolts, and the sea-survival of hatchery-reared smolts has been going for several years. A smolt trap is mounted in the river to catch migrating smolts.

Using the results from the trap catches, the number of migrating smolt was estimated. During the three latest years the trap has caught approximately ten percent of the smolts in the river. These results are derived from catch-recapture estimates from the trap (Jensen et al., 2002). Thus, the total run of wild smolts from River Eira in 2004 is estimated to be 20,000 Atlantic salmon and 4,500 trout smolts. The total weight of the wild smolts is calculated to be 800 kg. In 2004, 69,800 hatchery-reared Atlantic salmon smolts and 2,900 hatchery-reared trout smolts were released in the river, constituting a total weight of 5,800 kg. Thus, there seems to be a significant food resource available during a short period of time, obviously drawing the attention of predatory fish. Also fish eating birds are attracted and the rate of predation has been reported to be relatively high in the River Eira, where up to 10 % of the tags from released smolts were found on the riverbanks (Reitan et al., 1987). Our observations clearly indicate that large numbers of gulls were actively feeding on the migrating smolts, but it is unknown if the gulls, unlike the saithe, selectively eat hatchery smolts.

Other studies have failed to find a (expected) difference in mortality between hatchery-reared and wild fish (Jepsen et al., 1998; Hvidsten and Lund, 1988) and there is growing evidence that wild and hatcheryreared smolts are equally vulnerable to predation during their transition from a lentic to a lotic environment.

It was surprising that more wild than hatchery-reared smolts were found in the saithe, considering that hatchery-reared smolts outnumber wild smolts by 3:1. The results also indicate that the cod were more capable of catching the larger, less vigorous hatchery-reared smolts.

Calculations based on the gastric evacuation rate of saithe at 9 oC, show that the expected (max) daily ratio for 60 â&#x20AC;&#x201C; 70 cm saithe is about 50 g (from Andersen and Riis-Vestergaard, 2003). This means that if one 70

SUMBAWS

Final Report

assumes 50 days of smolt-run, it would take over 2000 individuals to eat all the smolts, but only 330 individuals to eat all the wild smolts. We have no data on the total number of saithe present in the Eresfjord, but estimations from Hvidsten and Lund (1988) showed a cod-population of nearly 5000 individuals in the Orkla estuary In conclusion, the transition from the freshwater in River Eira to the marine environment in the Eresfjord, constitutes a high risk of predation for both wild and hatchery-reared salmonid smolts. Saithe, specialising in smolt predation seem to play a major role in this predation and large year to year variations in the number of saithe may influence numbers of returning adult salmon.

References Andersen, N. G. and Riis-Vestergaard, J. (2003). The effects of food consumption rate, body size and temperature on net conversion efficiency in saithe and whiting. Journal of Fish Biology 62, 395-412.

Bakshtanskiy, E. L., Nesterov, V. D. and Neklydov, M. N. (1980). The behaviour of Young Atlantic Salmon (Salmo salar), during Downstream Migration. Journal of Icthyology 20, 93-100.

DeAngelis, D. L. and Petersen, J. H. (2001). Importance of the predator´s ecological neighbourhood in modelling predation on migrating prey. Oikos 94, 315 – 325.

Dieperink, C., Bak, B. D., Pedersen, L., Pedersen, M.I. and Pedersen, S. (2002). Predation on Atlantic salmon and sea trout during their first days as postsmolts. Journal of Fish Biology 61, 848-852.

Hvidsten, N. A. and Johnsen, B. O. (1993). Increased recapture rate of adult Atlantic salmon released as smolts into large shoals of wild smolts in the River Orkla, Norway. North American Journal of Fisheries Management 13, 272 – 276.

Hvidsten, N. A. and Lund, R. A. (1988). Predation on hatchery-reared and wild smolts of Atlantic salmon in the estuary of River Orkla, Norway. Journal of Fish Biology 33, 121 – 126.

Hvidsten, N. A. and Møkkelgjerd, P. I. (1987). Predation on salmon ismolts in the estuary of the River Surna, Norway. Journal of Fish Biology 30, 273 – 280.

Jensen, J. A., Finstad, B., Hvidsten, N. A., Jensås, J. G., Johnsen, B. O. and Lund, E. (2002). Fish biology surveys in the Aura watercourse. Annual report 2002. NINA Oppdragsmelding 781. 36pp. (In Norwegian with English summary).

Jepsen, N., Aarestrup, K., Økland, F. and Rasmussen, G. (1998). Survival of radio-tagged Atlantic salmon (Salmo salar L.) and trout (Salmo trutta L.) smolts passing a reservoir during seaward migration.

Hydrobiologica 372, 347-353.

Jepsen, N., Pedersen, S. and Thorstad, E. (2000). Behavioural interactions between prey (trout smolts) and predators (pike and pikeperch) in an impounded river. Regulated Rivers: Research and Management 16, 189-198.

Larsson, P. O. (1985). Predation on migrating smolt as a regulating factor in Baltic salmon, Salmo salar, populations. Journal of Fish Biology 26, 391-397.

Reitan, O., Hvidsten, N. A. and Hansen, L. P. (1987). Bird predation on hatchery reared Atlantic salmon smolts released in the River Eira, Norway. Fauna Norvegica 8, 35– 38.

Rieman, B. E., Beamsderfer, R. C., Vigg, S. and Poe, T. P. (1991). Estimated loss of juvenile salmonids to predation by northern squawfish, walleyes, and smallmouth bass in John Day Reservoir, Columbia River.

Tranactions of the American Fisheries Society 120, 448-458.

Schreck, C. B., Clements, S., Jepsen, D. and Karnowski, M. (2002). Evaluation of migration and survival of juvenile salmonids following transportation. Annual Report 2002, Project TPE-00-1. U.S. Army Corps of Engineers, Walla Walla District, Walla Walla, Washington.

Thorstad, E. B., Økland, F., Finstad, B., Sivertsgård, R., Bjørn, P. A. and McKinley, R. S. (2004).

Migration speeds and orientation of Atlantic salmon and sea trout post-smolts in a Norwegian fjord system.

Environmental Biology of Fishes 71, 305–311.

71

SUMBAWS

Final Report

Paper 6: SWIMMING SPEEDS AND ORIENTATION OF WILD FJORD SYSTEM

ATLANTIC SALMON POST-SMOLTS IN A NORWEGIAN

F. Økland, E.B. Thorstad, B. Finstad, R. Sivertsgård, N. Plantalech, N. Jepsen and R.S. McKinley (Submitted to Fisheries Management and Ecology) Introduction Atlantic salmon usually stay in fresh water for 2-4 years as juveniles, transform into the smolt stage and migrate to the ocean for feeding, before returning to freshwater to spawn 1-3 years later (Klemetsen et al., 2003). The heaviest mortality in the sea seems to take place during the first months of the marine migration (see Hansen et al., 2003). Knowledge about migratory routes, geographical distribution and swimming speeds of post-smolts in near shore areas is sparse (Voegeli et al.,1998; Lacroix and Voegeli, 2000; Moore et al., 2000), even though this is crucial information when addressing predation, salmon lice Lepeophtheirus salmonis infestations, and other important mortality factors (e.g. Thorstad et al., 2005; Sivertsgård et al., 2005).

Several studies of hatchery-reared Atlantic salmon post-smolts suggest that water current is the major transport factor in the seaward migration (Fried et al., 1978; LaBar et al., 1978; Lacroix and McCurdy, 1996; Moore et al., 2000). In contrast, Thorstad et al. (2004) found that the active swimming of the postsmolts was the major transport factor in a fjord system with weak water currents. The difference between studies may be a result of the simultaneous, in situ water current recordings made by Thorstad et al. (2004), as compared to studies with a few current recordings at fixed stations, or corrections based solely upon the direction of the tidal flow. Strong water currents in some of the previous studies may also have been dominating compared to the actual movements of the fish.

The behaviour of hatchery-reared post-smolts may differ from that of wild post-smolts (e.g. Jonsson et al., 1991) and as pointed out by Lacroix and McCurdy (1996) and Thorstad et al. (2004), there is a general need to obtain information on the movements of normally sized wild Atlantic salmon post-smolts. The acoustic transmitters previously available were too large for tagging wild Atlantic salmon smolts, but a limited number of smaller transmitters were produced for this project (Thelma, Norway). The objective of the present study was, therefore, for the first time to record observed and actual swimming speeds and directions of wild Atlantic salmon post-smolts in a Norwegian fjord system, and to compare these with previous results from hatchery-reared post-smolts (Thorstad et al., 2004).

Material and methods Eight wild Atlantic salmon smolts (mean total length 14.7 cm, range 13.0-19.9; mean mass 25 g, range 18-41), were caught in a trap close to the river mouth during their downstream migration in the River Eira, transported to the Statkraft hatchery in Eresfjord in a tank with oxygenated water, and tagged with acoustic transmitters 0-3 days later, using the methods described in Finstad et al. (2005). Seven of the smolts were tagged with transmitters (Thelma, Norway; 18 × 7 mm, mass in air/water of 1.9/1.2 g), and the largest smolt with a different transmitter (VEMCO, Canada; 9 × 20 mm, mass in air/water of 3.3/2.0 g). One to three days after tagging, the fish were transported in water-filled plastic bags to a cage (volume 0.5 m³) in the fjord, where they were kept for 2-3 hours before release. The fish were released in the inner part of the Romsdalsfjord system in south western Norway (Figure 1), at the mouth of the River Eira (mean annual water discharge of 15.5 m3 s-1), during 22 May - 5 June 2004. The fish were released together with 25-30 non-tagged hatchery-reared Atlantic salmon smolts.

The fish were manually tracked from a boat using a VEMCO VR60 or VR28 receiver. The fish position was fixed every 10th minute using a GPS. Individual fish were followed for mean 5.0 hours (range 2.9-6.5).

To correct fish swimming speeds and directions for the water current, a separate boat and crew simultaneously recorded the speed and direction of the water current using a current drift drogue. The drift drogue was put into the water at the site where the fish was recorded. Ten minutes later, the position of the drift drogue was recorded using a GPS. During tracking of individual fish, an average of 8.5 (range 2-19) 10-min recordings with simultaneous fish tracking and current recording were repeated. For fish not starting to move immediately after release, current recordings were not made before the fish started moving. The 72

SUMBAWS

Final Report

number of recordings for individual fish was dependent on boats and crew available, since the study was part of a larger field project.

The current drift drogue was made of canvas, and had three wings (57 cm wide and 1 m deep) mounted at 120째 angles to each other. The drift drogue was attached 30 cm below a floater at the surface and, hence, reflected the current at 0.3-1.3 m depth. Atlantic salmon post-smolts usually swim close to the surface, in the upper brackish layer (Fried et al., 1978; LaBar et al., 1978; Holm et al., 2003).

Actual swimming speeds and directions of the fish were calculated by vector analysis based on observed movements of the fish and the direction and speed of the water current over 10-minute periods. Statistical analyses were performed with SPSS 13.0 and the GPS mapping software OziExplorer 3.90.3a.

Figure 1 The Romsdalsfjord system in Middle Norway, indicating the site where acoustically tagged smolts were released.

Results The current direction varied during tracking, with a slightly higher number of recordings with the current flowing in towards the fjord than in other directions (Figure 2). Mean current speed was low (0.09 m s-1, range 0-0.54, SD = 0.11).

Mean observed migration speed (ground speed) of the post-smolts was 1.22 bl s-1 (based on individual means, range 0.42-1.87, SD = 0.60). When corrected for the speed and direction of the water current, the actual swimming speed was on average 1.17 bl s-1 (based on individual means, range 0.33-1.89, SD = 0.63).

Hence, the major transport factor of the post-smolts was their active swimming.

The highest frequency of both observed and actual movements were out of the fjord (northern to western directions, Figure 2). The actual movement, when corrected for the speed and direction of the water current, revealed the lowest frequencies in towards the fjord (eastern to southern directions, Figure 2). The direction of neither observed nor actual movement was dependent on the direction of the water current (linear regressions, observed: r2 = 0.001, p = 0.78, actual: r2 = 0.004, p = 0.62). However, the direction of the observed movement was dependent on the actual movement (linear regression, r2 = 0.63, p < 0.001).

73

SUMBAWS

Final Report

20

Frequency

15

10

5

0 N to E

E to S S to W W to N

Water current direction 25

Frequency

20 15 10 5 0 N to E

E to S S to W W to N

Observed swimming direction 25

Frequency

20 15 10 5 0 N to E

E to S S to W W to N

Actual swimming direction

Figure 2 Directions of the water current and observed and actual (i.e. after correcting for the speed and direction of the water current) movements of eight Atlantic salmon post-smolts. The figures show the frequency of 10-min recordings in the directions north to east (N to E, 0-90째), east to south (E to S, 90-180째, i.e. inwards the fjord), south to west (S to W, 180-270째) and west to north (W to N, 270-360째, i.e. outwards the fjord). Different colours indicate contribution from each individual.

74

SUMBAWS

Final Report

Discussion Atlantic salmon of hatchery origin may for several reasons show migration patterns differing from wild salmon (e.g. Jonsson et al., 1991; Thorstad et al., 1998; 2005), and it has been questioned whether results from migration studies of hatchery-reared post-smolts are representative of wild fish (e.g. Lacroix and McCurdy, 1996; Thorstad et al., 2004). The results from wild post-smolts in the present study resemble results from a previous study of swimming speeds and directions of hatchery-reared post-smolts in the same fjord system (Thorstad et al., 2004). The wild Atlantic salmon post-smolts in the present study were, similarly to hatchery-reared post-smolts, not passively drifting with the current, but actively swimming (actual swimming speeds of 1.2 bl s-1, as compared to actual swimming speeds of 1.3 bl s-1 for the hatcheryreared fish). In the same way as for hatchery-reared smolts, there was no relationship between the direction of observed movement and the direction of the water current, indicating that the post-smolts were moving in random directions in relation to the water current. The strong relationship between the direction of the actual and observed movement of both wild (present study) and hatchery-reared post-smolts (Thorstad et al., 2004), indicates that the observed directions of movement were highly dependent on the actual movement of the fish and not of the water current.

Orientation mechanisms used by migrating salmonid smolts and post-smolts differ among species and habitats (Groot, 1965; LaBar et al., 1978; Thorpe et al., 1981; Quinn and Brannon, 1982; Moser et al., 1991). The lack of highly directional movements in this study did not indicate precise navigation of the wild Atlanic salmon post-smolts, although the highest frequency of movements was out of the fjord, resulting in a net seaward movement. Studies of wild Atlantic salmon post-smolts using automatic listening stations in the same fjord system, showed migration speeds of 0.53 and 0.56 bl s-1 from release at the river mouth to passing sites 9.5 and 37 km from the river mouth, respectively (Thorstad et al., 2005). The swimming speeds recorded over 10-minute periods in the present study, thus, indicate that the post-smolts had actually been swimming approximately twice the distance from release when reaching the listening stations 9.5 and 37 km from the river mouth.

Both the present study and Thorstad et al., (2005) confirm that fjord migration of wild Atlantic salmon post-smolts resembles the migration of hatchery-reared post-smolts, and that management decisions may be based on studies of hatchery-reared post-smolts, even though information from wild fish should be acquired when possible. Information on swimming speeds, migration efficiency and orientation is essential for models of mortality factors such as sea lice infections and predation in different parts of the fjord systems and near coastal areas. The approach in the present study should be further extended to studies in the outer fjord areas to investigate whether results are consistent during the different stages of the outward migration.

References Finstad, B., Ă&#x2DC;kland, F., Thorstad, E. B., Bjørn, P. A. and McKinley, R. S. (2005). Migration of hatchery-reared Atlantic salmon and wild anadromous brown trout post-smolts in a Norwegian fjord system.

Journal of Fish Biology 66, 86-96.

Fried, S. M., McCleave, J. D. and LaBar, G. W. (1978). Seaward migration of hatchery-reared Atlantic salmon, Salmo salar, smolts in the Penobscot River estuary, Maine: riverine movements. Journal of the Fisheries Research Board of Canada 35, 76-87.

Groot, C (1965). On the orientation of sockeye salmon (Oncorhynchus nerka) during their seaward migration out of lakes. Behaviour 14 (Suppl.), 1-198.

Hansen, L. P., Holm, M., Holst, J. C. and Jacobsen, J. A. (2003). The Ecology of post-smolts of Atlantic salmon. In Salmon at the edge (ed. D. Mills), pp. 25-39. Blackwell Science.

Holm, M., Holst, J. C., Hansen, L. P., Jacobsen, J. A., O'MaoilĂŠidigh, N. and Moore, A. (2003).

Migration and distribution of Atlantic salmon post-smolts in the North Sea and North-East Atlantic. In Salmon at the edge (ed. D. Mills), pp. 7-23. Blackwell Science.

Jonsson, B., Jonsson, N. and Hansen, L. P (1991). Differences in life history and migratory behaviour between wild and hatchery-reared Atlantic salmon in nature. Aquaculture 98, 69-78.

75

SUMBAWS

Final Report

Klemetsen, A., Amundsen, P.-A., Dempson, J. B., Jonsson, B., Jonsson, N., O'Connell, M. F. and Mortensen, E. (2003). Atlantic salmon Salmo salar L., brown trout Salmo trutta L. and Arctic charr Salvelinus alpinus (L.): a review of aspects of their life histories. Ecology of Freshwater Fish 12, 1-59.

LaBar, G. W., McCleave, J. D. and Fried, S. M. (1978). Seaward migration of hatchery-reared Atlantic salmon (Salmo salar) smolts in the Penobscot River estuary, Maine: Open water movements. Journal du Conseil International pour l' Exploration de la Mer 38, 257-269.

Lacroix, G. L. and McCurdy, P. (1996). Migratory behaviour of post-smolt Atlantic salmon during initial stages of seaward migration. Journal of Fish Biology 49, 1086-1101.

Lacroix, G. L. and Voegeli, F. A. (2000). Development of automated monitoring systems for ultrasonic transmitters. In Advances in fish telemetry, Proceedings of the third conference on fish telemetry in Europe (eds. A. Moore and I. Russell), pp. 37-50. CEFAS, Lowestoft.

Moore, A., Lacroix, G. L. and Sturlaugsson, J. (2000). Tracking Atlantic salmon post-smolts in the sea.

In The ocean life of Atlantic salmon - environmental and biological factors influencing survival (ed. D. Mills), pp. 49-64. Fishing News Books, Oxford.

Moser, M. L., Olson, A. F. and Quinn, T. P. (1991). Riverine and estuarine migratory behaviour of coho salmon (Oncorhynchus kisutch) smolts. Canadian Journal of Fisheries and Aquatic Sciences 48, 1670-1678.

Quinn, T. P. and Brannon, E. L. (1982). The use of celestial and magnetic cues by orienting sockeye salmon smolts. Journal of Comparative Physiology A 147, 547-552.

Sivertsgård, R., Thorstad, E. B., Økland, F., Finstad, B., Bjørn, P. A., Jepsen, N., Nordal, T. & McKinley, R. S. (2005). Effects of salmon lice infection and salmon lice protection on fjord migrating Atlantic salmon and brown trout post-smolts. Hydrobiologia, submitted.

Thorpe, J. E., Ross, L. G., Struthers, G. and Watts, W. (1981). Tracking Atlantic salmon smolts, Salmo salar L., through Loch Voil, Scotland. Journal of Fish Biology 19, 519-537.

Thorstad, E. B., Heggberget, T. G. and Økland, F. (1998). Migratory behaviour of adult wild and escaped farmed Atlantic salmon, Salmo salar L., before, during and after spawning in a Norwegian river.

Aquaculture Research 29, 419-428.

Thorstad, E. B., Økland, F., Finstad, B. Sivertsgård, R., Bjørn, P. A. and McKinley, R. S. (2004).

Migration speeds and orientation of Atlantic salmon and sea trout post-smolts in a Norwegian fjord system.

Environmental Biology of Fishes 71: 305-311.

Thorstad, E. B., Økland, F., Finstad, B., Sivertsgård, R., Plantalech, N., Bjørn, P. A. and McKinley, R. S. (2005). Comparing migratory behaviour and survival of wild and hatchery-reared Atlantic salmon and wild anadromous brown trout post-smolts during the first stages of marine migration. Hydrobiologia, submitted.

Voegeli, F. A., Lacroix, G. L. and Anderson, J. M. (1998). Development of miniature pingers for tracking Atlantic salmon smolts at sea. Hydrobiologia 371/372, 35-46.

76

SUMBAWS

Final Report

Paper 7: MODELLING SALMON LICE INFESTATION ON WILD ATLANTIC SALMON AND SEA TROUT IN FJORD SYSTEMS

F. Økland, O. Diserud, R. Sivertsgård, E.B. Thorstad, N. Plantalech, N. Jepsen, I. Uglem, B. Finstad and R.S. McKinley (Manuscript) Introduction The salmon louse (Lepeophtheirus salmonis K.) is commonly found on salmonids in the marine environment, and is a serious pathogen both for farmed and wild salmonids (Finstad et al., 2000; Heuch et al., 2005). Salmon lice infestations represent a considerable economic loss in salmonid farming due to costs associated with treatments, management, reduced growth, impaired welfare and lowered product quality (Johnson et al., 2004). Salmonid aquaculture might also affect wild salmonid stocks negatively by drastically increasing the number of hosts for salmon lice in coastal areas where both wild and farmed salmonid smolts meet the marine environment for the first time (Gravil, 1996; O’Donoghue et al., 1998; Bjørn et al., 2001; Heuch and Mo, 2001; Holst et al., 2003; McKibben and Hay, 2004; Penston et al., 2004; Heuch et al., 2005). Tully and Whelan (1993), and Butler (2002), estimated that more than 95% of salmon lice eggs and larvae produced on the west coasts of Ireland and Scotland were of farm origin. In addition, infective lice stages may derive from escaped farmed salmon (Heuch and Mo, 2001).

In Norway, most Atlantic salmon (Salmo salar L.) migrate as smolts from the river to the open ocean where they stay for one to three years before returning to the rivers as adults. The period salmon smolts reside near coastal fish farms, i.e. in areas with elevated salmon lice infestation pressure, is thus restricted to a short period during spring. On the other hand, sea trout smolts (Salmo trutta L.) tend to stay in the coastal areas during their entire period in the sea, and might thus be more exposed to salmon lice infestations than wild salmon. This is exemplified by the fact that the sea trout populations on the west coast of Ireland collapsed already in 1989 after just a few years of salmonid farming in the actual areas (Gargan et al., 2003; Heuch et al., 2005).

Estimation of the variation in risks and levels of salmon lice infestation in fjords and coastal areas is greatly hampered by the lack of detailed information about the spatial and temporal distribution of wild salmonids in these areas. In the present paper we have employed results from a project aimed at studying temporal and spatial distribution of salmonid smolts in Eresfjord on the west coast of Norway to develop a universal model for estimation of risks and levels of salmon lice infestation in fjord systems. The objective of this project was to collect detailed information about swim speed, rate of smolt progression and migratory route, and also to measure salmon lice infestation levels in different parts along the migratory routes throughout the fjord system. The designed model can be used to evaluate whether or not observed infestation levels in other fjord systems are acceptable or safe, i.e. if the infestation levels are likely to be harmful or not for wild salmonids. Furthermore, the model also enables prediction of the risks and levels of smolts being infested by salmon lice at specific sites in a fjord. Hence, the model can be used to predict effects of the establishment of fish farms on new sites in fjords and coastal areas by identifying areas where the negative effects on wild salmonid populations will be minimal. In the future, such information will be crucial for an optimal exploitation of the coastal areas for aquaculture purposes and for conserving the wild salmonid populations.

Material and methods The model The fundamental idea behind the model is that it should be relevant for different fjord or coastal areas.

Thus, identification of variables that can be expected to be relatively constant between fjords and different parts of fjords was a prerequisite. Basically, there are two main factors deciding the risk and level of salmon lice infestation on a post-smolt; the infestation pressure and the time period a smolt is exposed. Over time, our aim is to develop a universal model that can be used in all fjord systems by adjusting variables based on the geographical size and shape, farming activity and information about the local wild salmonid stocks.

The model is based on a variable number of sub-models that can be used to simulate the number of salmon lice a smolt will accumulate within a specific section along its migratory route throughout a fjord.

77

SUMBAWS

Final Report

Each sub-model will simulate the risk and level of salmon lice infestation in one fjord section. The sections should represent fjord areas with relatively homogeneous geographical or physical conditions. The model for lice infestation (I) is then: I ( t ) = IP × (1 − PTF ) ×

FL SS × OE

where IP is the infestation pressure in the fjord section, (1-PTF) is the proportion of time the smolt stays in water with more than 20 ppt salinity, FL is the length of the fjord section, SS the mean swim speed and OE is the orientation efficiency coefficient in the actual section. The product SS × OE equals the rate of progression, so if this information is available it can be used directly. Total infestation (TI) throughout the whole fjord will then be the sum of the infestations from all sub-models.

Infestation pressure (IP) The infestation pressure is recorded by holding smolts in sentinel cages followed by a determination of the number of salmon lice on the fish after a two week period. By distributing cages along the expected migratory route salmon lice infestation levels can be estimated for sections of the fjord system. In 2004, the sentinel cages were moored both near shore and in the offshore pelagic areas since the observed migratory route for Atlantic salmon post-smolts is mainly in the pelagic part of the fjords, whereas sea trout postsmolts predominantly stay close to the shore (Thorstad et al., 2004; Finstad et al., 2005; Sivertsgård et al., 2005; Thorstad et al., 2005).

Proportion of time in water with less than 20 ppt salinity (PTF) If salmon lice copepodids (the infective salmon lice stages) are evenly distributed in the water column the swim depth of the fish should not influence risk and level of salmon lice infestation. In all Norwegian fjords there is an influx of fresh water from rivers, especially during the snow melting in the spring. The influx of freshwater creates a brackish stratum on the top of the water column, which can cover large fjord areas and sometimes be several meters thick (Unpublished data from the Romsdalsfjord system). Infective salmon lice stages have been recorded to avoid water less than 20 ppt in salinity (Heuch et al., 2005), probably because salmon lice have osmoregulatory problems in brackish water (Johnson and Albright, 1991). Thus, this brackish water layer can be regarded as a refuge for salmonid post-smolts. Furthermore, the infective stages of salmon lice are reported to aggregate at steep salinity gradients such as the transition zone between high salinity sea water and the brackish water stratum (Heuch et al., 1995). Hence, swim depth and water salinity might be important for the risk and level of salmon lice infestation, in particular in the first phase after the smolts enters the sea, since the brackish water layer is most pronounced in the estuaries.

At present, data on swim depth and salinity preferences in migrating salmonid smolts are lacking. A study aimed at collecting such data has, however, been initiated but data on swim depth and salinity preferences will not be available before 2006. Until such data are available the model simulates a worst case scenario and assumes that the post-smolts continuously stay in sea water above 20 ppt.

Fjord length (FL) FL is the length of the fjord section in one sub-model. The number of sub-models that is needed to simulate the salmon lice infestation along a migratory route or inside a fjord area will vary. For small areas where the variables included in the model are expected to be relatively homogeneous one sub-model for one section might give good simulations. In other systems infestation pressure might vary and as may the ability for the fish to navigate. Navigation in long and narrow systems is likely to be easier than finding the way through large basins with several alternative routes. Thus, the geographical shape of the system can influence the rate of progression and would also be important for salmon lice infestation.

Swim speed (SS) The swim speed of the fish was recorded by manual tracking of individual smolts tagged with acoustic transmitters (Thorstad et al., 2004, unpublished data). The location of the fish was recorded every 10 minutes. A current drift drogue was placed near the position of the tracked fish and the direction and speed of the water current were recorded for the period that the tagged fish was tracked. By subtracting current 78

SUMBAWS

Final Report

direction and speed from the fish movement, we obtained estimates of the absolute swim speed and direction of the fish.

The swim speed of the smolt increased with increasing body length (Thorstad et al., 2005). Most wild salmon post-smolts, and some wild sea trout post-smolts, are too small to be tagged with acoustic transmitters. Thus, the size of the fish examined in the project in Eresfjord is larger than the average size of wild post-smolt in the fjord system (Thorstad et al., 2004; Finstad et al., 2005; Sivertsgård et al., 2005b; Thorstad et al. 2005; Jensen et al., 2005). This is corrected in the model by expressing the swim speed as body lengths per second (bl s-1). The model can thereby be adapted to any body size and better simulate the movement of the wild fish in the system, even if they are smaller than the fish used to define some the parameters in the model.

Orientation efficiency (OE) How Atlantic salmon smolts are able to find the direction out from the river mouth to the open ocean is not known (Thorstad et al., 2004). Several hypotheses are suggested, but in our model we use a simple approach, adjusting the swim speed by an “orientation efficiency coefficient”. The coefficient is 1 if the smolt swims the shortest way from the river mouth throughout the fjord, and 0 if the smolt swims in circles and has no net “out fjord” movement.

By using automatic data logging stations positioned along the fjord it is possible to estimate migration time through different fjord sections. Hence, such data enables estimation of the rate of progression without the need of measuring individual swim speed and orientation efficiency, as was done in the Eiresfjord by manual tracking of individual tagged fish. However, measurement of the actual swim speed of post-smolts in different fjords and manual estimation of the orientation efficiency coefficient in different parts of fjord systems implies a higher predictive power.

Different simulation scenarios Accumulated infestation of salmon lice on Atlantic salmon post-smolt migrating from the River Eira through the Romsdalsfjord system Three sub-models were developed to estimate the number of salmon lice wild Atlantic salmon post-smolt from the River Eira will accumulate during fjord migration (Figure 1). By identifying the risk of infestation in different parts of the fjord system, we hope to illuminate the required preventive actions along the migratory route, and identify those parts most sensitive to changes in infestation levels. We will then be able to predict if salmon post-smolts already carry lethal infestations in some parts of the fjord system or when it leaves the Romsdalsfjord system and enters the open ocean. The total number of accumulated salmon lice in the Romsdalsfjord system (TI (t)) is based on accumulated numbers of salmon lice in the Eresfjord I1 (t1 ), Langfjord I2(t2) and Romsdalsfjord I3(t3), respectively, i.e.: TI (t) = I1 (t1) + I2(t2)+ I3(t3) where I1 (t1 ), I2(t2) and I3(t3) are predicted by the model; (1)

I i ( ti ) = IPi × (1 − PTFi ) ×

FLi , i = 1, 2, 3 SSi × OEi

where IPi is the average infestation pressure in different pelagic area, (1-PTFi) is the proportion of time in water with more than 20 ppt salinity in the pelagic area, FLi is the length of the fjord section, SSi the mean swim speed, OEi is the orientation efficiency coefficient in the fjord section, and ti is the time used in the fjord section.

79

SUMBAWS

Final Report

Figure 1 Location of the Romsdalsfjord system on the west coast of Norway. Along the migratory route of Atlantic salmon post-smolts from the River Eira (arrow) three geographical areas were identified and individual sub-models developed to estimate the salmon lice infestation from each area separately. Area 1 is Eresfjord, area 2 is Langfjord and area 3 is Romsdalsfjord.

The parameter estimates for 2004 for the different sub-models are given in Table 1. In the calculations we use the average body length 12.8 cm. Note that the scale of the units varies, e.g. time is given both in days and seconds. This implies that the scales have to be standardized before the analyses.

Table 1 The parameter estimates for 2004 used to simulate salmon lice infestation on Atlantic salmon postsmolts migrating through the Romsdalsfjord system from River Eira. (Sections 1 to 3 represent the geographical area along the migration route given in Figure 1. IP is the infestation pressure of salmon lice recorded in the sector, PTF the proportion of time the fish is in brackish water, FL the length of the section the sub-model is developed for, SS is the swim speed and OE the orientation efficiency coefficient used to correct the swim speed to the actual rate of progression).

Section 1 2 3

IP [n days-1] 0.10 0.27 1.49

PTF 0 0 0

FL [km] 9.5 27.5 28

SS [bl s-1] 1.32 1.32 1.32

OE 0.25 1.03 0.60

The average periods a salmon smolt will stay in each of the fjord sections are 113, 53 and 156 hours respectively. The numbers of lice they are expected to collect in each section are 0.03, 0.04 and 0.69 respectively, giving an expected total number of salmon lice TI of 0.76 in the Romsdalsfjord in 2004.

The variation between individual rates of progression for salmon smolts can be estimated by simulating 10 000 fish through the model, and assigning each an OE, or a rate of progression (RoP) drawn randomly from our original set of observations. The number of lice that each simulated smolt can collect in each section will then be Poisson distributed with rate Ii(ti). The results from the simulations are presented in Figure 2. Top left panel gives the TI-values, with a blue point on the base line indicating the expected value 0.76 and two red points indicating the 95 % confidence interval. The simulation shows that even if the average number of lice is below 1, some smolt may pick up as many as 7 lice. The three other panels give the simulated distribution for the three fjord sections. For all simulations, the main contribution of lice is seen in fjord in section 3, i.e. Romsdalsfjord, even if a few smolts spend much time in the inner parts of the fjord and also accumulate lice in sections 1 and 2.

80

SUMBAWS

Final Report

In order to illustrate the effect of varying body size, simulations were also run for a “minimum-smolt” with body length 10 cm and for a “maximum-smolt” with length 17 cm. The relation between length and expected number of lice is inversely proportional, so a decrease in body length will result in more lice due to a reduced rate of progression. A salmon smolt with length 10 cm is expected to accumulate 0.98 lice, whereas a smolt with length 17 cm on average accumulates 0.58 lice.

Figure 2 Simulated number of salmon lice on Atlantic salmon post-smolt from the River Eira for the whole fjord route (TI), and for each fjord section separately. The model is based on data from 2004, and with average smolt length 12.8 cm. The X-axis shows the number of accumulated lice.

When mapping the salmon lice infestation pressure in 2002, lice were detected on smolts only in Romsdalsfjord and, even there, the infestation pressure was low. The model gave low total lice infestation levels, and only 2.3 % of the salmon post-smolts were infected, and then only with 1 lice each.

Estimating salmon lice infestation in sea trout populations utilizing one section in the Romsdalsfjord system Most of the wild sea trout post-smolts in Norway are found near the river mouth, and they seldom migrate more than 80 km from their native watercourse (Thorstad et al., 2004; Finstad et al., 2005; Sivertgård et al., 2005b; Thorstad et al., 2005). Thus, the simplest assumption is to develop one model estimating the lice infestation in the section near the river mouth for the period post-smolts are in the sea.

The model of total accumulated number of salmon lice in sea trout from Eira I1(t1 ), Visdalselva I2(t2 ) and Røa I3(t3 ) is based on a simplified version of the equation above; (2)

I i ( ti ) = IPi × (1 − PTFi ) × ti , i = 1, 2, 3

where IPi is the infestation pressure in the littoral area in the fjord area, (1-PTFi) is the proportion of time in water with more than 20 ppt salinity in the littoral area, and ti is the time used in the fjord area.

81

SUMBAWS

Final Report

Accumulated infestation of salmon lice on sea trout post-smolts from the River Eira utilizing large littoral habitats in the Romsdalsfjord system However, some wild sea trout post-smolts are found to migrate as much as 80 km from the home river outlet in the Romsdalsfjord system (Thorstad et al., 2004; Finstad et al., 2005; Sivertgård et al., 2005b; Thorstad et al., 2005). The migratory route for most of these sea trout was in littoral areas rather than in the pelagic areas. Thus, different sub-models are necessary to calculate the accumulated infestation whithin different parts of the littoral route. Three sub-models were developed to estimate the average total accumulated number of salmon lice for a sea trout population with three differing migration strategies in littoral areas in the Romsdalsfjord system. The first strategy is to spend all the time in the Eresfjord section, with total infestation TI1(t1). The second strategy is to migrate through Eresfjord with “low-directional” swimming (feeding behaviour involves the fish spending most of the time in Langfjord), and then a more resolute swimming back through Eresfjord. The third strategy is the one for the long-distance swimmers which migrate outwards through Eresfjord and Langfjord, and spend most of their time in the outer parts of the fjord before migrating back to the River Eira. The average TI of the sea trout population will then be a weighted sum over the three strategies, where the weights correspond to the proportions of the populations that choose the respective strategies.

The total infestation for the proportion of the population that follows strategy 1, i.e. spending all the time in Eresfjord, can then be calculated directly by: TI1 ( ttot ) = IP1 × (1 − PTF1 ) × ttot

where ttot is the time the sea trout spend in the fjord.

The total infestation for the strategy 2 sea trout becomes: TI 2 ( ttot ) = I1,out + I 2 + I1,in

where I1,out is the infestation the fish experiences swimming out through Section 1, I1,in the infestation on the way back to the river, and I2 the accumulated lice infestation in Section 2 for the remaining time. We will here use rates of progression (RoP) instead of the product SSxOE. The equations, when we assume different RoPs outwards and inwards, are: I1,out = IP1 × (1 − PTF1 ) × t1,out , t1,out =

I1,in = IP1 × (1 − PTF1 ) × t1,in , t1,in =

FL1 RoP1,out

FL1 RoP1,in

I 2 = IP2 × (1 − PTF2 ) × t2 , t2 = ttot − t1,ut − t1,inn

The total infestation for the strategy 3 sea trout becomes: TI 3 ( ttot ) = I1,out + I 2,out + I 3 + I 2,in + I1,in

where I1,in and I1,out are the same as for the Strategy 2 sea trout, I2,in and I2,out similarly for the migration through section 2, and I3 the accumulated lice infestation in Section 3 for the remaining time: I 3 = IP3 × (1 − PTF3 ) × t3 , t3 = ttot − t1,ut − t2,ut − t2,inn − t1,inn

The parameter estimates are from 2002, 2003 and 2004 (Table 3). The average time spent in the fjord over these years, ttot, was 63 days. In the calculations we apply an average body length for the sea trout at 16.5 cm. Note that the FL-values are different from those given in Table 2, due to differing locations for the automatic listening stations.

82

SUMBAWS

Final Report

Table 2 The parameters used to estimate the salmon lice infestation on sea trout post-smolts from River Eira. (Section 1 to 3 is given illustrated in Figure 1, IP is salmon lice infection pressure, PTF is the proportion of time the smolt spend in freshwater, FL is length of the section, RoP out the rate of progression out to the next section and RoP in the migratory speed recorded by Atlantic salmon smolt out the fjord, used to indicate that the return migration might be fast) Section 1 2 3

IP [n days-1] 0.05 0.31 1.87

PTF 0 0 0

FL [km] 9.4 39 29

RoPout [bl s-1] 0.08 0.46 -

RoPin [bl s-1] 0.14 0.88 -

The total infestations for the three strategies are 0.23, 1.16 and 5.73 lice, respectively. In 2003 we followed 56 sea trout, of which 45 fish chose strategy 1, 3 fish strategy 2 and the remaining 8 fish swam all the way out to Section 3 (strategy 3) and spent 41 of 63 days in the outer parts of the fjord. The weighted total infestation of the River Eira population is then 1.06.

More interesting than the average value of the population is perhaps the proportion of the population that accumulates more than a certain number of lice. The population can be separated into a proportion with no effect, a proportion with sub lethal stress effects and a proportion with levels resulting in mortality. This can be illustrated by performing simulations as for those presented for Atlantic salmon. The RoP-values are drawn randomly from the original values (Bootstrap-sampling) and the number of lice on individual sea trout is assumed to be Poisson-distributed with rates I. The results are presented in Figure 3. We see that all sea trout that choose strategies 1 or 2 experienced safe levels of lice infestation, but some of the trout that swam all the way out the fjord (strategy 3) accumulated as many as 20 lice.

Figure 3 Simulated number of salmon lice on sea trout post-smolts from the River Eira with three different strategies. TI1 are lice contributions from Eresfjord only, strategy 1, TI2 is the lice contribution from sea trout post-smolts utilizing both Eresfjord and Langfjord (strategy 2) and TI3 is the contribution from the proportion of the population utilising all three fjord sections (strategy 3). The model is based on data from 2003 and 2004, and with average smolt length 16.5 cm. The X-axis shows the number of accumulated lice.

83

SUMBAWS

Final Report

Discussion By a consistacally tagging Atlantic salmon and sea trout, the Eresfjord project has provided data on movement and distribution of post-smolts in fjords. Manual tracking of single smolts has given detailed information of swim speed and route, while geographical distributions and migratory speeds over large areas have been recorded by use of automatic listening stations. Together with recorded salmon lice infestation levels from the same fjord system, this information made it possible to design a model simulating the salmon lice infestation levels on both sea trout and salmon post-smolts. The model can be used to predict the effects of establishment of fish farms at new sites in fjords and coastal areas by identifying areas where the negative effects on wild salmonid populations will be low. Such information will be crucial for an optimal exploitation of the coastal areas for aquaculture purposes and for conserving the wild salmonid populations.

A salmon lice infestation of 0.75 chalimus larvae per gram fish weight will kill a 40 g Atlantic salmon post-smolt (Finstad et al., 2000). Using the same factor on wild Atlantic salmon smolts sizes from midNorway, which normally weigh from 15 to 20 g, this indicates that as few as 11.3 chalimus larvae can have detrimental effects and lead to mortality (Finstad et al., 2000). Based on this assumption, the results from the simulations for Atlantic salmon smolts carried out in the present study indicate that the levels of salmon lice infestation and also the mortality should be low in 2002. The catch statistics from returning wild Atlantic salmon which left the river in 2002, show that the catch of grilse (one-sea-winter salmon) in 2003 was higher than previous years, whereas the catch of two-sea-winter salmon in 2004 was somewhat reduced (Peder Fiske, NINA, pers. comm.). The total catches of adult salmon originating from the 2002 smolt run were thus slightly higher than an average of the previous years, indicating a minor increase in the total marine survival compared with previous years. The results from the catch statistic from 2003 and 2004 do not therefore contradict the predictions from the model.

Sea trout seem to be more tolerant of salmon lice infestations than salmon. Bjørn et al. (2001) suggest that an infestation of 1.6 chalimus larvae per gram fish weight can lead to mortality for sea trout smaller than 150 g, but physiological stress effects have been recorded at infestation levels as low as 0.35 chalimus larvae per gram fish weight (Wells et al., submitted). Sea trout from the River Eira are 16.5 cm on average, and 35 g in body weight. Physiological stress should thus occur at infestations with 12 or more chalimus larvae, whereas more than 56 lice will lead to mortality. The simulations predicted low lice infestation levels for sea trout with migratory strategies 1 and 2, i.e. fish that spend all their time in the Eresfjord and Langfjord section. However, for sea trout post-smolts displaying migratory strategy 3, (i.e. fish that utilize the outer part of the fjord system) a minor proportion was predicted to accumulate a number of sea lice that would cause a harmful physiological stress level. Hence, on the basis of the simulations, the sea trout population in the River Eira should not experience any major problem from salmon lice infestations in the period from 2002 to 2004.

The infestation levels entered into the model are based on data collected in spring and early summer when the Atlantic salmon is migrating through. Normally, the infestation level increases during summer and the model should therefore underestimate the infestation level on the sea trout. To better describe the infestation on sea trout during the summer, the model will be developed to simulate an increasing infestation level during the summer. This will not influence the simulated salmon lice infestation on Atlantic salmon, but will increase the infestation on the sea trout until the model gives similar results to those reported from the gillnet fisheries in the fjord system (Sivertsgård et al., 2005). The data from the sea trout fisheries will be available soon.

References Bjørn, P.A., Finstad, B. and Kristoffersen, R. (2001). Salmon lice infection of wild sea trout and Artic char in marine and freshwater: the effects of salmon farms. Aquaculture Research 32, 947-962.

Butler, J.A. (2002). Wild salmonids and sea lice infestations on the west coast of Scotland: sources of infection and implications for the management of marine salmon farms. Pest Management Sciences 58, 595608.

Finstad, B., Bjørn, P.A. Grimnes, A. & Hvidsten, N.A. (2000). Laboratory and field investegations of salmon lice (Lepeoptheirus salmonis Krøyer) infestation on Atlantic salmon (Salmo salar L.) post smolts.

Aquaculture Research 31, 795-803.

Gargan, P.G., Tully, O and Poole, W.R. (2003). Relationship between sea lice infestations, sea lice productions and sea trout survival in Ireland, 1992-2001. In Salmon at the Edge (Mills eds.) Blackwell Science, Oxford; 119-135.

84

SUMBAWS

Final Report

Gravil, H.R. (1996). Studies of the biology and ecology of the free swimming larval stages of Lepeophtheirus salmonis (Krøyer) and Caligus elongatus Nordmann 1832 (Copepoda: Caligidae) pp. 1-299.

PhD thesis, University of Stirling, Stirling, Scotland UK Grimnes, A., Birkeland, K., Jakobsen, P.J. and Finstad, B. (1996). Lakselus – nasjonal og internasjonal kunnskapsstatus. NINA Fagrapport 18, 1-20 (In Norwegian with English summary).

Heuch, P.A. and Mo, T.A. (2001). A model of salmon louse production in Norway: Effects of increasing salmon production and public management measures. Diseases of Aquatic Organisms 45, 145-152.

Heuch, P.A., Bjørn. P.A., Finstad, Holst, J.C., Asplin, L. and Nilsen, F. (2005). A review of the Norwegian ”National Action Plan Against Salmon Lice on Salmonids”: The effect on wild salmonids.

Aquaculture 246, 79-92.

Heuch, P.A., Parsons, A. and Boxaspen, K. (1995). Diel vertical migration: A possible host-finding mechanism in salmon louse (Lepeophtheirus salmonis) copepodids? Canadian Journal of Fisheries and Aquatic Sciences 52, 681-689.

Holst, J.C., Jacobsen, P., Nilsen, F., Holm, M., Asplin, L. and Aure, J. (2003). Mortality of seawardmigrating post-smolts of Atlantic salmon due to salmon lice infection in Norwegian salmon stocks. In Salmon at the Edge. (ed. by. D. Mills), pp. 136-137. Oxford: Blackwell Science.

Jensen, A.J., Finstad, B., Hvidsten, N.A., Jensås, J.G., Johnsen, B.O., Lund, E. and Holthe, E. (2005).

Fiskebiologiske undersøkelser i Auravassdraget. Årsrapport 2004. NINA Rapport 16, 1-52 (In Norwegian with English summary).

Johnson, S., Treasurer, J.W., Bravo, S., Nagasawa, K. and Kabata, Z. (2004). A review of the impact of parasitic copepods on marine aquaculture. Zoological Studies 43, 229-243.

Johnson, S.C. and Albright, L.J. (1991). Development, growth, and survival of Lepeophtheirus salmonis (Copepoda: Caligidae) under laboratory conditions. Journal of the Marine Biological Association U.K. 71, 425-436.

McKibben, M.A. and Hay, D.W. (2004). Distributions of planktonic sea lice larvae Lepeophtheirus salmonis in Loch Torridon, Western Scotland in relation to salmon farm production cycles. Aquaculture Research 35, 742-750.

O’Donoghue, G., Costello, M. and Costello, J. (1998). Development of a management strategy for the reduction/elimination of sea lice larvae Lepeophtheirus salmonis parasites of salmon and trout. Marine Resource Series NO. 6, pp. 1-51. The Marine Institute, Dublin, Ireland Penston, M.J., McKibben, M.A., Hay, D.W. and Gilibrand, P. (2004). Observations on open-water densities of sea lice larvae in loch Shieldaig, Western Scotland. Aquaculture Research 35, 793-805.

Sivertsgård, R., Bjørn, P.A., Kristoffersen, R and Finstad, B. (2005a). Fish farming, salmon lice, Lepeophtheirus salmonis Krøyer, and protective zones; effects on wild sea trout, Salmo trutta Linne, populations. In prep.

Sivertsgård, R., Thorstad., E.B., Økland, F., Finstad, B., Bjørn, P.A., Jepsen, N., Nordal., T. and McKinley, R.S. (2005b). Effects of salmon lice infection and salmon lice protection on fjord migrating Atlantic salmon and brown trout post-smolts. Submitted to Hydrobiologia.

Thorstad, E.B., Økland, F., Finstad, B., Sivertsgård, R., Plantalech, N., Bjørn, P.A. & McKinley, R.S. (2005). Comparing migratory behaviour and survival of wild and hatchery-reared Atlantic salmon and wild anadromous brown trout postsmolts during the first stages of marine migration. Submitted to Hydrobiologia.

Thorstad, E.B., Økland, F., Finstad, B., Sivertsgård, R., Bjørn, P.A., McKinley, R.S. 2004. Migration speeds and orientation of Atlantic salmon and sea trout post-smolts in a Norwegian fjord system.

Environmental Biology of Fishes 71, 305-311.

Tully, O. and Whelan, K.F. (1993). Production of nauplii of Lepeophtheirus salmonis (Krøyer) (Copepoda: Caligidae) from farmed and wild salmon and its relation to the infestation of wild sea trout (Salmo trutta L.) off the west coast of Ireland in 1991. Fisheries Research 17, 187-200.

Wells, A., Grierson, C.E., Russon, I., Reinardy, H., Middlemiss, C., Bjørn, P.A., Finstad, B., Wendelaar, S.E., Mackenzie, M., Todd, C.D. and Hazon, N. (2006). The physiological effects of simultaneous abrupt seawater entry and sea lice infestation of wild sea trout smolts. Canadian Jounrnal of Fishery Sciences (submitted) 85

SUMBAWS

Final Report

Workpackage 3 Lead Partner: NIFA Participating Partners: NINA, NCFS OBJECTIVES • • • •

Identify stages of the fjord migration where post-smolts are particularly susceptible to sea lice infestation.

Assess possible differences in the susceptibility to infestation of Atlantic salmon and sea trout postsmolts.

Gather data on the salmon lice infestation pressures in fjords of differing fish farming activity.

Gather data on the natural infestation pressures on wild Atlantic salmon and sea trout post-smolts from a fjord system of moderate infestation pressure.

DELIVERABLES • • • •

Data on stages/locations of fjord migration where fish are most susceptible to salmon lice infestation.

Quantitative data on differences in the susceptibility to infestation of Atlantic salmon and sea trout Natural infestation pressure and the risk of infestation in fjordic/coastal systems with different farming activity.

Quantitative data on the natural infestation of wild salmonids, and assessments of their consequences.

EXECUTIVE SUMMARY To gather data about differences in salmon lice infestation between species, fjord areas and fjord systems (objective 1, 2 and 3), we depended heavily on results from naive salmonids caged in sentinel cages. Large field studies with sentinel cages both in the Romsdalsfjord system (62°40’N, 8°10’E) in middle Norway, and in the Altafjord system (70°05’N, 22°55’E) in northern Norway were therefore initiated. Heavy mortality in caged fish required us to terminate the cage study after the first trial in Romsdalsfjord in 2003.

As a remedial action, we started an intensive campaign of gill-netting for sea trout in the protection farmfree inner area (Eresfjord), as well as the intensively farmed outer area (Karlsøyfjord) of the Romsdalsfjord system both in 2003 and 2004.

In the first paper (paper 1), salmon lice infestation levels in wild sea trout were investigated inside (Eresfjord) and outside (Karlsøyfjord) a salmon farming protection zone in the Romsdalsfjord system, Norway. No new salmonid farms are allowed to be established within the protection zone and those already established have to operate under a stricter regime. Wild sea trout sampled outside the protection zone (Karlsøyfjord) had higher prevalence and significantly higher abundances than sea trout inside the protection zone (Eresfjord). The infection pattern followed the same pattern. Low infestation level was found in spring followed by a summer peak at levels expected to have negative consequences for wild sea trout both inside and outside the protection zone. Our results therefore indicate that wild sea trout will benefit from a farm protection zone in the fjord. However, the infestation level observed on wild sea trout inside the protection zone is higher than in other completely farm-free fjords in Norway. This implies that the zones are too small to have the necessary effect, and this is discussed in relation to the establishment of numerous salmon farming protection zones in Norway. This paper to a large degree answers the first (Identify stages of the fjord migration where post-smolts are particularly susceptible to sea lice infestation) and the third objective (Gather data on the salmon lice infestation pressures in fjords of differing fish farming activity) in the original research plan. The most important results are that the study has shown

86

SUMBAWS

Final Report

differences in infection pressure between the intensively farmed outer area of the Romsdalsfjord system compared to the farm-free inner part of the fjord.

In addition, a new sentinel cage study with improved infrastructure and methodology was initiated in the second year in the Romsdalsfjord system (paper 2). Differences in infestation risks of salmon lice were studied in three connected fjord sections of the Romsdalsfjord system. Eresfjord, in the inner part, lacks salmon farming activity. Langfjord, in the middle part, has two fish farms, whereas Karlsøyfjord is subject to extensive salmon farming activity. Naive Atlantic salmon post-smolts were placed in duplicated sentinel cages in both pelagic and littoral areas of all three fjords. After a fortnight, salmon lice infestation was analysed. The results showed that there were significant differences in infestation pressure between the three fjords. Eresfjord (no farms) had the lowest infestation risks, intermediate risks were found in Langfjord (two farms), whereas Karlsøyfjord (extensive farming activity) had the highest infestation risk. No clear differences between the littoral (inshore) and pelagic (offshore) areas of the fjord were found. Results showed that sentinel cages can be used as an alternative method to monitor salmon lice infestation pressure in fjord systems. Abundances on caged fish are, however, much lower than found in free-swimming wild fish, and this experimental protocol therefore needs to be calibrated. The most important result was that differences in infection pressure between the intensively farmed outer area of the Romsdalsfjord system compared to the farm-free inner part of the fjord were confirmed.

An even larger sentinel cage study with both naive sea trout and Atlantic salmon smolts was initiated in the Altafjord system in 2003 and continued in 2004. Duplicate littoral and pelagic cages containing both species were deployed from the outfall of the Alta river to the fjord mouth areas in three repeated trials in both years. This study was designed to assess possible differences in susceptibility between Atlantic salmon and sea trout post-smolts to sea lice infestation (objective 2). The aims were also to investigate stages where post-smolts are particularly susceptible to infestations (objective 1) as well as gather data on the salmon lice infestation pressure in fjords of differing fish farming activity (objective 3). Although every effort was taken to improve survival of caged fish, mortality was heavy (65%), and further analyses of the results was not given priority. As a remedial action, field data on the natural infestation pressures on wild Atlantic salmon and sea trout post-smolts from fjord systems of moderate infestation pressure (objective four) were strengthened.

Data on differences in salmon lice infection on sympatric populations of fjord-migrating Atlantic salmon post-smolts, sea trout, and Arctic char was studied in three fjords in northern Norway (paper 3). Atlantic salmon post-smolts were captured in the fjords only during late June and early July, and probably left the fjords after this period. Results show that none of these fish were infected with salmon lice. In contrast, sea trout and Arctic char had similar infection patterns during their sampling periods, with very low prevalence and mean infection intensity during June, slight increases in July, and peaking in August. These observations indicate that Atlantic salmon smolts seemingly may have a mismatch between time of peak lice infection fish and their post-smolt fjord migration in northern fjords. In contrast, sea trout and Arctic char feed within the fjords throughout the summer and consequentially have a higher risk of harmful infections in years with suitable environmental conditions for salmon lice development (paper 3). The most important result is therefore that significant differences in risk of salmon lice infection are indicated between pelagic-migrating Atlantic salmon post-smolts and littoral-feeding sea trout and Arctic char.

The fourth paper is a continuum of paper three and, through a two year sampling effort with wild salmonids in the Altafjord system, addresses possible differences in risk and consequences in salmon lice infection between the species and between different habitats (paper 4). Results show that most of the Atlantic salmon post-smolts were uninfected with salmon lice, but smolts captured in the outer fjord area in 87

SUMBAWS

Final Report

2004 carried a few lice. This is the first time that lice have been found on wild Atlantic salmon post-smolts in northern Norway. By contrast, pelagic-feeding sea trout and Arctic char had surprisingly high infection levels during June and July, which suggests that their pelagic patterns of movement expose them to high risks of salmon lice infection. Very low prevalence and mean infection intensity was noted during early summer, with much higher levels in summer and autumn; these data were in contrast marked by those found in littoral-feeding sea trout and Arctic char. These results concur with earlier observations that Atlantic salmon post-smolts have a mismatch between the time of peak lice infection and their post-smolt fjord migration in northern fjords. In contrast, sea trout and Arctic char feed within the fjords throughout the summer and consequentially have higher risk of harmful lice infections. However, risks of infection may also differ between pelagic-feeding sea trout and Arctic char on the one hand, and littoral-feeding fish on the other, as well as between species and age classes. The most important result is that significant differences in risk of salmon lice infection are confirmed between fjord-migrating Atlantic salmon postsmolts and fjord-feeding sea trout and Arctic char.

CONCLUSIONS Although the sentinel cage study was not completely successful, the remedial action taken through intensive gill net fishing and trawling for live fish greatly improved the results. All objectives and deliverables and the expected overall results in the original research plan have therefore been met, but some additional analyses are still needed. Important results are the identification of the migratory stages where salmonids are most susceptible to salmon lice infestation (papers 1 and 2), quantification of the risks and consequences of infestation in wild salmonids in fjord systems with different fish farming activity (papers 1, 2 and 3), and identification of infection risk between the salmonid species (papers 3 and 4). From a management perspective, the most significant results are that, 1) the correlation between higher risks of salmon lice infestation and intensively fish farming activity is confirmed, 2) protection zones must be of a considerable size to give wild salmonids sufficient protection against salmon lice infestation, and 3) risk of infection differs between species, feeding behaviour and age classes. Locally integrated salmon lice management models must, therefore, be initiated and based on regional (e.g. southern vs. northern) and local (fjord) ecological and environmental knowledge to ensure that the management measures applied are both appropriate and sufficient to achieve the objective of protecting wild salmonids stocks.

88

SUMBAWS

Final Report

Paper 1: FISH FARMING, SALMON LICE, LEPEOPHTHEIRUS SALMONIS (KRØYER), AND PROTECTION ZONES; EFFECTS ON WILD SEA TROUT, SALMO TRUTTA LINNE, POPULATIONS R. Sivertsgård, P. A. Bjørn, B. Finstad, R. Kristoffersen, T. Nordal and T. Øverland (Manuscript) Abstract Salmon lice (Lepeophtheirus salmonis K.) infestation levels in wild sea trout (Salmo trutta L.) were investigated inside (Eresfjord) and outside (Karlsøyfjord) a salmon farming protection zone in the Romsdalsfjord system, Norway. No new salmonid farms can be established within the protection zone and the ones already established have to operate under a stricter operational regime. Wild sea trout sampled outside the protection zone (Karlsøyfjord) had a higher prevalence and a significantly higher abundance of salmon lice than sea trout inside the protection zone (Eresfjord). The infection pattern in both zones followed a similar pattern; low levels of infestation in spring followed by a summer peak at levels expected to have negative consequences for wild sea trout. Our results indicate that wild sea trout will benefit from a farm protection zone compared to the infestation of fish captured outside this zone, even in the same fjord system. However, the infestation level observed in wild sea trout inside the protection zone is higher than in completely farm free fjords in Norway. This implies that the zone in the present study is too small to have the necessary effect, and this is discussed in the light of the establishment of numerous salmon farming protection zones in Norway.

Introduction Many sea trout stocks in Norway, Ireland and Scotland have experienced annual salmon lice epidemics during the latest decade (Tully et al., 1993; Anon, 1999; Bjørn et al., 2001; Todd et al., 2000; Butler & Watt, 2003; Gargan et al., 2003; Kålås and Urdal, 2004; Bjørn et al., 2005; Heuch et al., 2005). This is probably a consequence of increased salmon farming activity in fjords and coastal areas (Anon, 1999; Todd et al., 2000; Bjørn et al., 2001; Butler and Watt, 2003; Gargan et al., 2003; Rikardsen, 2004; Heuch et al., 2005).

Within a salmon farm, or within a salmon farming area, the concentration of potential hosts is much larger than in areas without farms. The increased salmon lice production in farmed fish has been suggested to contribute to the high infestation levels in wild fish (Todd et al., 2000; Bjørn et al., 2001; Butler and Watt, 2003; Gargan et al., 2003; Heuch et al., 2005; Krkošek et al., 2005). Due to the year around stock of farmed salmon in coastal waters and fjords, high lice production can also occur during late winter and spring time when wild hosts are scarce (Heuch and Mo, 2001; Heuch et al., 2005). This can cause negative consequences for a first time migrating salmonid smolt, if they become heavily infested, since physiological disturbances due to salmon lice infestation are especially important in smaller fish (Grimnes and Jakobsen, 1996; Bjørn and Finstad, 1997; Finstad et al., 2000; Tully and Nolan, 2002; Wagner et al., 2004, 2005; Wells et al., sub.). Annual salmon lice epidemics have therefore been blamed for the collapse of many sea trout and Atlantic salmon (Salmo salar L.) stocks in fjords and coastal areas with heavy fish farming activity (Anon, 1999; Bjørn et al., 2001; Butler and Watt, 2003; Gargan et al., 2003; Heuch et al., 2005).

The Norwegian Authorities have taken several measures to reduce the infestation pressure in fjords with intensive salmon farming activity (Anon, 1999; Heuch et al., 2005). Of particular importantance is the establishment of salmon farming protection zones in a number of “Norwegian Salmon Fjords” to protect wild Atlantic salmon and wild sea trout from different harmful interactions with cultured salmonids (Anon, 1999). The main argument for the initiative, besides reducing the numbers of escaped farmed salmon in these fjords, is to prevent the spread of salmon lice infestation from farmed to wild salmonids (Anon, 2004; Heuch et al., 2005). Within a protection zone, no new fish farms are allowed to be established. The fish farms already established inside the zone have to work under a stricter operational regime, including compulsory delousing if the salmon lice infestation level in farmed fish reaches a certain threshold (Anon, 1999; Anon, 2004; Heuch et al., 2005). Therefore, wild sea trout inside a protected zone are expected to have a lower infestation level than fish outside a protection zone, even within the same fjord system.

The present study was initiated to investigate possible differences in salmon lice infestation level in sea trout inside (Eresfjord) and outside (Karlsøyfjord) a protection zone of the Romsdalsfjord system. The

89

SUMBAWS

Final Report

results are discussed in the light of the establishment of numerous small Norwegian salmon farming protection zones as a measure against salmon lice infestation Methods Study area Eresfjord and Karlsøyfjord are two fjord areas in the Romsdalsfjord system (62°40’N8°10’E) in Møre & Romsdal county, mid Norway (Figure 1). The Romsdalsfjord system has extensive salmon farming activity, but most of the farms are concentrated in the outer part of the fjord, with only a few farms in the inner part.

The inner part is also incorporated in a protection zone, and Eresfjord is completely devoid of salmon farms.

Many rivers with anadromous populations of sea trout and Atlantic salmon enter both the outer and inner part of the Romsdalsfjord system. Since most sea trout at sea in Norway feed close to their native river (Jensen, 1968; Berg and Berg, 1987; Finstad et al., 2005), the different sea trout populations in the Romsdalsfjord system, are probably also exposed to different salmon lice infestation pressure. Two different gill-netting localities were therefore selected in the present study: Eresfjord inside, and Karlsøyfjord outside, the protection zone in the Romsdalsfjord system. A comprehensive two-year sampling program was initiated, which included regular gill-netting as well as environmental sampling (salinity and temperature) at the two study localities.

Figure 1 Map of the Romsdalsfjord system. Large squares shows the location of the 2003 and 2004 sea trout gill-net area in the inner (Eresfjord) and outer (Karlsøyfjord) area of the Romsdalsfjord system. Circles show the location of the fish farms in operation in 2003 and 2004. Triangles show fish farms only in operation in 2004, while small squares show fish farms only in operation in 2003. The hatched area of the fjords shows the area of the fjord with a salmon farming protection zone.

Sampling and analyses Wild sea trout smolts were regularly sampled by floating gill-nets from May to July/August in 2003 and 2004 after the methods described in Bjørn et al. (2001). Most of the salmonid post-smolts descend the rivers during May in this fjord system (Jensen et al., 2005), and the sampling regime therefore covered most of the seawater phase. Sampled fish were gently cut out off the nets and placed in individual plastic bags and frozen for subsequent examination. In the laboratory, captured fish were thawed and salmon lice were examined according to Bjørn et al. (2001). The fork length (LF) of wild fish was measured to the nearest mm and fish were weighed to the nearest 0.1 g. Temperature and salinity were measured on each sampling occasion at 50 cm intervals to a depth of 12 m.

Statistical tests and handling of data The ecological terms, prevalence, abundance and intensity were used according to Bush et al. (1997).

The degree of aggregation of parasites is measured by the variance (S2) to mean ( x ) ratio (S2 x -¹), which appears to be the most reliable when comparing samples with different prevalence or means (Scott, 1987).

90

SUMBAWS

Final Report

We performed statistical analyses using SPSS 12.0. GPS position of every fish farm in the whole fjord system was taken from the Norwegian Fishery Directorate. The GPS position of the area in the fjord for gillnet fishing was based on the GPS position in the middle of the test fishing area. We used these GPS positions in OziExplorer 3.90 GPS Mapping Software to find the shortest distance between farms and the area of the fjord where we collected the wild sea trout.

To evaluate the relationship between the salmon lice abundance on wild fish sampled in the protected (Eresfjord) and unprotected (Karlsøyfjord) area, and the potentially infestation pressure from the salmon farms, we collected infestation data and the total number of fish in approximately 90% of all farms in operation in 2003 and 2004 in the Romsdalsfjord system.

We are unable to use data from all the fishing farms in operation as we require permission from farmers.

We have a complete data set from fish farms inside the protection zone. All data were drawn from the Norwegian Food Safety Authority. The infestation data are based on louse counts from farmers or veterinarians on a random sample of 20 fish from each farm every second week during the investigation period. Thus, infestation data for every month in each fish farm used in this present study is a mean of two lice countings. This compulsory louse monitoring level follows the procedures from the Norwegian Food Safety Authority (www.mattilsynet.no).

Results Temperature and salinity Table 1 Mean (± SD) salinity and temperature in the investigation period, measured every 50 cm from surface down to 12 m on each day of gill net fishing 2003 and 2004.

Eresfjord Karlsøyfjord

2003 Salinity (‰) Temperature (ºC) (mean ± SD) (mean ± SD) 25.6 ± 7.1 11.8 ± 3.1 28.0 ± 1.2 14.6 ± 1.5

2004 Salinity (‰) Temperature (ºC) (mean ± SD) (mean ± SD) 18.8 ± 6.4 15.1 ± 3.8 28.8 ± 1.6 14.4 ± 3.6

Salmon lice infestation in the protection zone (Eresfjord) In 2003, a low percentage (31%) of all sea trout smolts captured carried salmon lice. Abundance (mean ± SD) and intensity (mean ± SD) was 7.9 ± 25.1 and 25.9 ± 41.4, respectively. The salmon lice infestation increased generally from spring towards summer (Table 2). Low abundance was found in May which increased towards June and peaked in July (Mann-Whitney U-test; p < 0.050). The prevalence followed a similar pattern as the abundance. The infestation in May consisted of 14% larvae, 43% preadult and 43% adult lice (Figure 2). There was a re-infestation of larvae (91% of total infestation) in June, followed by a higher degree of adult stages in July. The infestation in July consisted of fewer larvae (52% larvae in total infection) but more preadult (32% preadult of total infection) and fewer adult (16% adult of total infection) stages than in June. In 2004, 52% of the fish captured were infested with salmon lice. Total abundance was 6.0 ± 13.8 and intensity was 11.7 ± 17.4 (see Table 2). As in 2003, low abundance was found in late spring (May) which increased towards similar levels in June and July (Mann-Whitney U-test; p = 0.002), and was reduced in August (Mann-Whitney U-test; p = 0.050). In May 2004, the infestation consisted of mostly larvae (70%) and a few preadult and adult lice. A high degree of larvae (94%) was also found in June. In July, more preadult (32%) and adult stages (7%) were found. In August, the sea trout were predominantly infested with adult and larval stages.

91

SUMBAWS

Final Report

Table 2 Infestation parameters for fish caught in different weeks in Eresfjord 2003 and 2004.

Year

2003

Week of sampling Number of fish Mass (g) mean ± SD LF (mm) mean ± SD Prevalence (%) Abundance mean ± SD Intensity mean ± SD median IQR Range Variance to mean ratio

21 48

24 43

28 27

21 43

24 25

28 13

32 18

155 ± 137

97 ± 69

168 ± 153

267 ± 304

383 ± 276

196 ± 163

427 ± 413

25 ± 6

21 ± 4

24 ± 6

27 ± 7

27 ± 7

25 ± 6

30 ± 9

4

37

67

28

64

77

72

0.2 ± 0.7

8.9 ± 3.0

20.5 ± 35.5

1.2 ± 2.5

12.2 ± 23.5

10.7 ± 13.4

5.7 ± 6.4

3.5 ± 0.7 3.5 1 3.0 – 4.0

23.9 ± 45.7 9.5 19 1.0 – 186.0

30.2 ± 40.1 8.5 45.3 1.0 – 143.0

4.3 ± 3.1 3.5 6.8 1.0 – 9.0

19.0 ± 27.2 5.5 34.2 1.0 – 96.0

13.9 ± 13.8 10 19.8 1.0 – 44.0

7.9 ± 6.3 7 7.5 2.0 – 24.0

0.1

87.1

53.2

2.2

39.1

13.8

5

a)

b) 100

100

80

80

Frequency (%)

Frequency (%)

2004

60

40

60

40

20

20

0

0 21

24

28

Week

21

24

28

32

Week

Figure 2 Frequency distribution of different developmental stages of salmon lice on infected fish sampled in Eresfjord in different weeks in 2003 (a) and 2004 (b). Larvae (black): all chalimus stages combined; Preadult (white): preadult male and female; Adult (grey): adult male and female.

Salmon lice infestation outside the protection zone A high proportion (88%) of the total sea trout captured outside the protection zone in 2003 was infested with salmon lice (see Table 3). These fish had a moderately high abundance (mean ± SD) 28.8 ± 42.5 and intensity (mean ± SD) 32.8 ± 42.6. Abundance was low in May, peaked in June (Mann-Whitney U-test; p<0.0001) and decreased in July (Mann-Whitney U-test; p < 0.029). In May, the fish were mainly infested with larval (46%) and adult (36%) stages. The infection peak in June showed that there was a second infestation episode inside Karlsøyfjord (90% larvae, 7% preadult and 3% adult). The high degree of larvae (73%) continued in July, but preadult and adult lice were also observed (Figure 3). 83% of the fish were infested in 2004. Total abundance was 15.9 ± 25.0 and intensity was 19.0 ± 26.2 (see Table 3). Abundance increased from May to June (Mann-Whitney U-test; p = 0.050) and July (Mann-Whitney U-test; p = 0.526), but was reduced in August (Mann-Whitney U-test; p = 0.038). The fish caught in May were mostly infected with larva, but also some preadult and adult stages (Figure 3). High numbers of larval stages were also observed in June and July, together with preadult and adult lice. In August, fewer larvae but more preadult and adult lice were detected.

92

SUMBAWS

Final Report

Table 3 Infestation parameters for fish cached in different weeks in Karlsøyfjord 2003 and 2004.

Year

2003

Week of sampling Number of fish Mass (g) mean ± SD LF (mm) mean ± SD Prevalence (%) Abundance mean ± SD Intensity mean ± SD median IQR Range Variance to mean ratio

21 15

24 19

28 8

21 6

24 21

28 9

32 7

425 ± 325

364 ± 194

453 ± 368

490 ± 664

363 ± 189

176 ± 133

322 ± 428

33 ± 7

32 ± 6

32 ± 10

31 ± 11

40 ± 39

38 ± 43

29 ± 8

73

100

88

83

86

89

71

3.3 ± 5.7

52.8 ± 53.1

19.6 ± 15.5

5.7 ± 4.4

18.2 ± 20.1

27.9 ± 42.9

2.4 ± 2.6

4.6 ± 6.2 3 2 1.0 – 23.0

52.8 ± 53.1 38 37 1.0 – 130.0

22.4 ± 14.4 18 30 4.0 – 40.0

6.8 ± 3.8 7 7.5 2.0 – 11.0

21.3 ± 20.2 17 14 1.0 – 93.0

31.4 ± 44.5 10.5 58 3.0 – 123.0

3.4 ± 2.5 4 4.5 1.0 – 7.0

8.5

53.4

9.2

2.2

19.1

63.2

1.9

a)

b) 100

100

80

80

Frequency (%)

Frequency (%)

2004

60

40

20

60

40

20

0

0 21

24

28

Week

21

24

28

32

Week

Figure 3 Frequency distribution of different developmental stages of salmon lice on infected fish sampled in Karlsøyfjorden in different weeks in 2003 (a) and 2004 (b). Larvae (black): all chalimus stages combined; Preadult (white): preadult male and female; Adult (grey): adult male and female.

Comparison of infestation between the zones The total catch in Eresfjord (2003 and 2004) was 217 fish, and in Karlsøyfjord 85 fish (see Tables 2 and 3). More than twice as many fish were infested outside (Karlsøyfjord: 85% infested) than inside (Eresfjord: 40% infected) the protection zone. In 2003, total abundance was also higher (Mann-Whitney U-test; p<0.0001) outside than inside the protection zone (Figure 4). In 2004, the percentage of infested fish was approximately 1.7 times higher in Karlsøyfjord (outside) than in Eresfjord (inside the protected zone). Also, in terms of abundance (Mann-Whitney U-test; p<0.0001) sea trout outside the protected zone had a significantly higher salmon lice infestation than sea trout inside the protection zone (Figure 4).

93

SUMBAWS

Final Report

100

Abundance

80

60

40

20

0 2003

2004

Year

Figure 4 Box-and-whiskers plot showing the abundance (mean number of lice per sampled fish), of all fish sampled in Karlsøyfjord (white squares) and Eresfjord (gray squares) in 2003 and 2004. The lower and the upper hinges give the 25th and 75th percentile.

Salmon lice infestation in fish farms The total number of farmed salmonids in 90% of all fish farms in operation in the Romsdalsfjord system from May to August was higher in 2003 (range 5 090 995 to 5 802 051) than 2004 (range 3 319 621 to 4 576 022) (Mann-Whitney; p = 0.004) (Figure 5). Farmed fish had a significantly higher infestation with sexually mature female sea lice, in comparison with 2004 (Mann-Whitney; p = 0.042) (Figure 5). There was no difference in the abundance of gravid female lice in fish farms between the protection zone and the nonprotection zone in 2003 (Mann-Whitney; p = 0.228) and 2004 (Mann-Whitney; p = 0.357). a) 1,0

Abundance of pregnant female

Total number of farmed fish

5000000

4000000

3000000

2000000

0,8

0,6

0,4

0,2

1000000 0,0 3

4

5

6

Month

7

8

3

4

5

6

7

8

Month

94

SUMBAWS

Final Report

3000000

Abundance of pregnant female

Total number of farmed fish

b)

2500000

2000000

1500000

0,8

0,6

0,4

0,2

0,0 3

4

5

6

7

8

3

Month

4

5

6

7

8

Month

Figure 5 Total numbers of farmed salmonids and abundance of pregnant female salmon lice in farmed salmonids in the Romsdalsfjord system. Data is from March (3) to August (8) in 2003 (a) and 2004 (b).

Dotted line (- - -) is inside the protection zone and straight line (__) is outside the protection zone.

Discussion Salmon lice infestation inside the protection zone (Eresfjord) Moderate numbers of sea trout captured in the protection zone carried salmon lice in both years. Sea trout in Eresfjord had moderate salmon lice infestation during both study years and the infestation level changed seasonally. The lowest prevalence and abundance were observed in spring each year, while the infestation level peaked during late summer (July) and leveled off in early autumn (August 2004). This seasonal pattern is in accordance with most other studies from northern Norway, except that a second peak has often been observed in late autumn (August and September) (Bjørn et al., 2001; Bjørn and Finstad, 2002; Rikardsen, 2004). In a farm-free fjord in southern Norway, Schram et al. (1998) studied seasonal occurrence of salmon lice and Caligus elongatus on wild sea trout. The infestations of both parasites increased in spring, peaked in July and August, and dropped during winter time. In another farm-free fjord in south-eastern Norway, Mo and Heuch (1998) observed higher occurrence of salmon lice in wild sea trout in October compared to August (1992) and in September compared to May (1993). In contrast to the infestation level in Eresfjord (inside protection zone), a much lower intensity has generally been found in farm-free fjords both in northern (Bjørn et al., 2001; Rikardsen, 2004) and southern parts of Norway (Mo and Heuch, 1998; Schram et al., 1998). The results from the present study therefore indicate that establishment of small protection zones, as in the Romsdalsfjord system, may be beneficial for sea trout but that the protection against salmon lice infestation is limited compared to a completely farm-free fjord.

Salmon lice infestation outside the protection zone (Karlsøyfjord) Most of the sea trout captured outside the protection zone were infested with salmon lice in both years, and the sea trout carried a relatively high infestation. The infestation was generally low in May of both years, increased in June and July and was reduced in August (2004). The higher abundance of salmon lice on wild sea trout outside the protection zone is likely to be due to higher densities of farmed hosts (see Figure 5) and the likely associated higher infestation pressure in this area (Heuch and Mo, 2001; Heuch et al., 2005). However, the infestation level in the present study was lower compared to other studies in intensively fish farming areas in Norway (Birkeland, 1996; Bjørn et al., 2001; Kålås and Urdal, 2004; Bjørn et al., 2005). In some of those studies, the infestation levels were very high both in fish captured in freshwater and saltwater (Birkeland, 1996; Bjørn et al., 2001). Infestations at such levels may kill immature fish in seawater if the fish are unable to terminate the sea phase and return prematurely to freshwater to restore physiological homeostasis (Grimnes and Jakobsen, 1996; Finstad et al., 2000; Bjørn et al., 2001; Wagner et al., 2005; Wells et al., sub.) and to delouse (Finstad et al., 1995). In the present study, premature 95

SUMBAWS

Final Report

return to freshwater was not investigated. However, infestation levels expected to have negative consequences for sea trout were frequently found in fish in Karlsøyfjord (Wagner et al., 2004, 2005; Wells et al., sub.). This, together with the concurrent reduction in abundance and variance to mean ratio in the final samples in 2003 and 2004, implies that infestation pressure in Karlsøyfjord was at sufficient levels to cause physiological disturbance and premature return to freshwater. One reason for the lower abundance in Karlsøyfjord compared to other studies in farming areas (e.g. Bjørn et al., 2001), may be that most fish farms in the Romsdalsfjord system are located in outer coastal and fjord areas and quite distant from the sampling locality in Karlsøyfjord (see Figure 1). In Ireland, Gargan et al. (2003) examined the relationship between infestations on wild sea trout with distance to salmon farms over a 10-year period. Sea trout sites less than 20 km from fish farms had the highest mean abundance of both total salmon lice and larval stages.

At sites less than 30 km away from farms, infestation level was lower, and at sites more than 30 km away, they reported very low infestation levels. A model (Gargan et al., 2003) suggests that at distances more than 25 km away from farms, infestations will not exceed 32 lice · fish-1. Although not directly comparable, this corresponds well with results in the present study, since all except one sampling period in Karlsøyfjord had levels below that number. The shortest mean distance from the middle of the gill-net area used in Karlsøyfjord to all fish farms in operation in the Romsdalsfjord system both years, was approximately 26 km.

Comparison of infestation between the zones Very high salmon lice infestation levels have frequently been observed in fjords with intensive fish farming activity compared to fjords without farms (Mo and Heuch, 1998; Schram et al., 1998; Todd et al., 2000; Bjørn et al., 2001; Bjørn and Finstad, 2002; Gargan et al., 2003; Butler and Watt, 2003; Rikardsen, 2004; Heuch et al., 2005; Bjørn et al., 2005; Kålås and Urdal, 2004; Krkošek et al., 2005). Our present data add additional evidence to this. Salmon lice infestation on wild sea trout caught both inside (Eresfjord) and outside (Karlsøyfjord) a protection zone had infestation levels comparable to levels that have only been reported on wild trout from farmed fjords (Birkeland, 1996; Bjørn et al., 2001; Bjørn et al., 2005; Gargan et al., 2003; Kålås and Urdal, 2004). However, the significantly higher abundance outside the protection zone, confirms that there is a higher infestation pressure from infective salmon lice in this part of the fjord system.

These results may indicate a gradually declining infestation pressure from the outer to the inner fjord area of the Romsdalsfjord system.

Information about salmon lice infestation of sea trout inside different areas in the same fjord system is generally limited, but new results on fjord migrating Pacific salmon Oncorhynchus spp. smolts indicate that infestation pressure may be spatially concentrated (Krkošek et al., 2005). Results from the present study indicate that sea trout smolts may also experience different salmon lice infestation pressure in different fjord areas, especially during spring and summer, and that this early infestation pressure is highest outside the protection zone. From May to July the infestation pressure was largest outside the protection zone, followed by a reduced infestation pressure inside Karlsøyfjord (outside the protection zone) along with an increasing infestation inside the protection zone (Eresfjord). In August 2004 there was no significant difference in abundance between sea trout from Eresfjord and Karlsøyfjord. Most of the first migrating smolts in this fjord system, migrate to sea in May (Jensen et al., 2005), and these fish are especially vulnerable to lice infestation the first few weeks after migration (Finstad et al., 2000; Bjørn et al., 2001; Tully and Nolan, 2002). Consequently, a protection against lice in spring and early summer is very important to post-smolt survival and fitness. Therefore, the protection zone partly seems to fulfil a protection role against salmon lice infestation, but the relatively high infestation level inside the protected area also implies that the zone is too small to achieve the long term goal of “reducing the harmful effects of lice on wild salmonids to a minimum” (Heuch et al., 2005).

However, to draw conclusions about the relationships between fish farms, protection zones and infestation levels on wild fish, knowledge about the fish migration patterns as well as dispersal of infective lice inside the fjord system is a prerequisite. We assume that sea trout smolts sampled in Eresfjord and Karlsøyfjord have spent all, or at least most, of their time in the sampling areas. Wild sea trout smolts in Norway are generally found within a few km off their native river mouth, and seldom migrate more than 80 km from their water-course (Jensen, 1968; Berg and Berg, 1987; Finstad et al., 2005; Sivertsgård et al., in press; Thorstad et al., in press). Due to the large distance between the two sampling sites, it is reasonable to believe that most of the sea trout were captured close to their area of origin, and that the infestation level represents the true infestation level at the sampling point. In many circumstances, the highest infestation levels occur at rivers nearest fish farms (Butler and Watt, 2003; Gargan et al., 2003), but lice can also be dispersed over long distances depending on the physical fjord conditions (Asplin et al., in prep). A salmon 96

SUMBAWS

Final Report

farm can be considered as a point source of infection (Krkošek et al., 2005). The waterborne transmission of lice from this source to wild fish will be strongly dependent on both seawater circulation (Costello et al., 1996; Rikardsen, 2004; Heuch et al., 2005), and lice movement, survival and longevity (Johnson and Albright, 1991; Heuch et al., 1995; Boxaspen and Næss, 2000; Vikeså, 2000; Stien et al., 2005). L. Asplin (Institute of Marine Research, Bergen Norway, unpublished data), using model simulation experiments, has shown that infective copepodids may have a passive drift transport from zero to hundreds of kilometres under normal Norwegian coastal current conditions and temperature. Therefore, almost all of the fish farms in the Romsdalsfjord system could potentially disperse infective salmon lice stages throughout most of the system, given the appropriate environmental conditions. The protection zone in the Romsdalsfjord system seems to have positive effects against salmon lice infestation in wild sea trout. This protection is, however, limited compared to a fjord system completely devoid of fish farms (Mo and Heuch, 1998; Schram et al., 1998; Rikardsen, 2004). Establishment of small protection zones inside a fjord area or fjord system can therefore not be the only consensus tool against salmon lice. Regional delousing of entire fjord systems (Heuch et al., 2005), compulsory reporting of lice counts (Anon, 1999), and lower legal limits for the maximum mean number of lice per fish, are probably also necessary in intensively farmed areas (Heuch et al., 2005).

The establishment of a protection zone will also raise the question about which wild species or wild populations we want to protect. Since Atlantic salmon and sea trout smolts differ in their strategies during the marine sea phase (Klemetsen et al., 2004; Finstad et al., 2005) a protection zone like the one in the Romsdalsfjord system will give different protection to different species and populations. Atlantic salmon smolts migrate out of the fjord to the open ocean (Klemetsen et al., 2003; Finstad et al., 2005; Sivertsgård et al., in press; Thorstad et al., in press), while sea trout remain in the fjord system near their home river outlet (Jensen, 1968; Berg and Berg, 1987; Finstad et al., 2005; Sivertsgård et al., in press, Thorstad et al., in press). A protection zone that increases the number of fish farms in the outer part of the fjord system would probably also increase the infestation pressure in the same area. Wild sea trout inside the protection zone will, to a certain degree, benefit from this. The situation can be the opposite for wild Atlantic salmon smolts which have to migrate through intensively farmed areas in order to reach the open ocean.

Knowledge about the direct effects of protection zones is lacking and establishment of the present zones was mostly based on “the precautionary principle”. The topics are of special interest in Norway, since the authorities have chosen area protection as the main measure to protect wild Atlantic salmon at sea (Anon, 1999; Anon, 2004). Inside “National salmon fjords”, fish farms are prohibited or restricted. This will probably increase the number of fish farms and thereby the salmon lice infection pressure in the outer parts of the fjords. Therefore, the time that Atlantic salmon smolts spend inside areas with lower infestation pressure must be balanced by the time spent in areas with higher infestation pressure. In the Romsdalsfjord system, wild Atlantic salmon smolts will probably take approximately 10 – 12 days (Thorstad et al. in press) to move from the protected zone to outer coastal areas. It is therefore questionable if the zones are large enough to give wild salmonids sufficient protection against salmon lice.

Acknowledgements We are pleased to acknowledge the help of numerous helpful people throughout this study. We are grateful to the staff at the Statkraft hatchery in Eresfjord, Bjørg Anne Vike, Petter Sira and Torbjørn Utigard, for extensive help and co-operation during the project. Trond Haukebø at the County Governor in Møre & Romsdal for all support and Ivar Ellingsgård with family at Bolsøya for all their help during the project period. Thanks to all farmers in the Romsdalsfjord system that have give us the permission to use the data and to the Norwegian Food Safety Authority in Molde for all their help. The study was financed by the European Commission contract no: Q5RS-2002-00730, the Norwegian Institute for Nature Research and Statkraft SF, Norway.

References Anonymous. (1999). Til laks åt alle kan ingen gjera? Om årsaker til nedgangen i de norske villlaksbestandene of forslag til strategier og tiltak for å bedre situasjonen, Statens forvaltningstjeneste.

Norges offentlige utredninger. NOU 1999. 9, 1-296. (In Norwegian with English summery).

Anonymous. (2004). Nasjonale laksevassdrag og laksefjorder. Ferdigstilling av ordningen. (Downloadable: www.dirnat.no/archive/attachments/01/61/NLVNL053.pdf) 97

SUMBAWS

Final Report

Berg, O.K. & Berg, M. (1987). Migrations of sea trout, Salmo trutta L., from the Vardnes river in northern Norway. Journal of Fish Biology 31, 113-121.

Birkeland, K. (1996). Salmon lice, Lepeophtheirus salmonis Krøyer, infestations and implications for anadromous brown trout, Salmo trutta L. Dr. scient thesis, University of Bergen.

Bjørn, P.A. & Finstad, B. (1997). The Physiological Effects of Salmon Lice Infection on Sea Trout Post Smolts. Nordic Journal of Freswater Research 73, 60-72.

Bjørn, P.A. & Finstad, B. (2002). Salmon lice, Lepeophtheirus salmonis (Krøyer), infestation in sympatric poulations of Artic char, Salvelinus alpinus (L.) and sea trout, Salmo trutta (L.), in areas near and distant from farms. ICES Journal of Marine Science 59, 131-139.

Bjørn, P.A., B, Finstad & Kristoffersen, R. (2001). Salmon lice infection of wild sea trout and Artic char in marine and freshwaters: the effects of salmon farms. Aquaculture Research 32, 947-962.

Bjørn, P.A., B, Finstad & Kristoffersen, R. (2005). Registration of salmon lice on Atlantic salmon, sea trout and Artic charr in 2004. NINA Rapport 60, 1-26 (In Norwegian with English summary).

Boxaspen, K. & Næss, T. (2000). Development of eggs and planktonic stages of salmon lice (Lepeophtheirus salmonis) at low temperatures. Contributions to Zoology. 69, 51-55.

Bush, A. O., Lafferty, K. D., Lotz, J. M. & Shostak, A. W. (1997). Parasitology meets ecology on its own terms: Margolis et al. revised. Journal of Parasitology 83, 575-583.

Butler, J.R.A. & Watt, J. (2003). Assessing and Managing the impacts of Marine Salmon Farms on Wild Atlantic Salmon in Western Scotland: Identifying Priority Rivers for Conservations. In Salmon at the Edge (Mills eds.). Blackwell Science, Oxford, 93-118.

Castelloe, M., Castelloe., J., Roche, N. (1996). Planctonic dispersion of larval salmon-lice, Lepeophtheirus salmonis, associated with cultured salmon, Salmo salar, in Western Ireland. Journal of the Marine Biological Assosiation of the United Kingdom 76, 141-149.

Finstad, B., Bjørn, P.A. & Nilsen, S.T. (1995). Survival of salmon lice, Lepeophtheirus salmonis Krøyer, on Artic charr, Salvelinus alpinus (L.), in fresh water. Aquaculture Research 26, 791-795.

Finstad, B., Bjørn, P.A. Grimnes, A. & Hvitsten, N.A. (2000). Laboratory and field investegations of salmon lice (Lepeoptheirus salmonis Krøyer) infestation on Atlantic salmon (Salmo salar L.) post smolts.

Aquaculture Research 32, 947-962.

Finstad, B., Økland, F., Thorstad, E. B., Bjørn, P. A. & McKinley, R. S. (2005). Migration of hatcheryreared Atlantic salmon and wild anadromous brown trout postsmolts in a Norwegian fjord system. Journal of Fish Biology 65, 1-11.

Gargan, P.G., Tully, O & Poole, W.R. (2003). Relationship between sea lice infestations, sea lice productions and sea trout survival in Ireland, 1992-2001. In Salmon at the Edge (Mills eds.) Blackwell Science, Oxford, 119-135.

Grimnes, A. & Jakobsen, P. (1996). The physiological effects of salmon lice (Lepeophtheirus salmonis Krøyer) infection on post smolt of Atlantic salmon (Salmo salar). Journal of Fish Biology 48, 1179-1194.

Heuch, P. A. & Mo, T. A. (2001). A model of salmon louse production in Norway: Effects of increasing salmon production and public management measures. Diseases of Aquatic Organisms 45, 145–152.

Heuch, P.A., Parsons, A. & Boxaspen, K. (1995). Diel vertical migration: A possible host-finding mechanism in salmon louse (Lepeophtheirus salmonis) copepodids? Canadian Journal of Fisheries and Aquaculture 52, 681-689.

Heuch, P.A., Bjørn. P.A., Finstad, Holst, J.C., Asplin, L. & Nilsen, F. (2005). A review of the Norwegian ”National Action Plan Against Salmon Lice on Salmonids”: The effect on wild salmonids.

Aquaculture 246, 79-92.

Jensen, K.W. (1968). Sea trout (Salmo trutta L.) of the river Istra, western Norway. Report to the Institute of Freshwater Research Drottningholm 48, 187-213.

Jensen, A.J., Finstad, B., Hvidsten, N.A., Jensås, J.G., Johnsen, B.O., Lund, E. & Holthe, E. (2005).

Fiskebiologiske undersøkelser i Auravassdraget. Årsrapport 2004. - NINA Rapport 16, 1-52.

98

SUMBAWS

Final Report

Johnson, S.C., Albright, L.J. (1991). Development, growth, and survival of Lepeophtheirus salmonis (Copepoda: Caligidae) under laboratory conditions. Journal of the Marine Biological Assosiation of the United Kingdom 71, 425-436.

Klemetsen, A., Amundsen, P-A., Dempson, J.B., Johnson, B., Johnson, N., O’Connell, M.F. & Mortesen, E. (2003). Atlantic salmon Salmo salar L., brown trout Salmo trutta L. and Artic charr Salvelinus alpinus (L.): a review of aspects of their life histories. Ecology of Freshwater Fish 12, 1-59.

Krkošek, M., Lewis, M.A. & Volpe, J.P. (2005). Transmission dynamics of parasitic sea lice from farm to wild salmon. Proceedings of the Royal Society B 272, 689-696.

Kålås, S. & Urdal, K. (2004). Overvaking av lakselusinfeksjonar på tilbakevandra sjøaure i Rogaland, Hordaland og Sogn & Fjordane sommaren 2004. Rådgivende biologer AS 761, 1-40. (In Norwegian with English summary).

Mo, T.A. & Heuch, P.A. (1998). Occurrence of Lepeophtheirus salmonis (Copepoda: Caligidae) on sea trout (Salmo trutta) in the inner Oslo Fjord, south-eastern Norway. ICES Journal of Marine Science 55, 176180.

Rikardsen, A.H. (2004). Seasonal occurrence of sea lice Lepeophtheirus salmonis on sea trout in two north Norwegian fjords. Journal of Fish Biology 65, 711-722.

Schram, T.A., Knutsen, J.A., Heuch, P.A. & Mo, T.A. (1998). Seasonal occurrence of Lepeophtheirus salmonis and Caligus elongatus (Copepoda: Caligidae) on sea trout (Salmo trutta), off southern Norway.

ICES Journal of Marine Science. 55, 163-175.

Scott, M. E. (1987). Temporal changes in aggregation: a laboratory study. Parasitology 94, 583-595.

Sivertsgård, R., Thorstad., E.B., Økland, F., Finstad, B., Bjørn, P.A., Jepsen, N., Nordal., T. & McKinley., S.R. (2005) Effects of salmon lice infection and salmon lice protection on fjord migrating Atlantic salmon and brown trout post-smolts. Hydrobiologia. In press.

Stien, A., Bjørn, P.A., Heuch, P.A. & Elson, D.A. (2005). Population dynamics of salmon lice Lepeophtheirus salmonis on Atlantic salmon and sea trout. Marine Ecology Progress Series 290, 263-275.

Todd, C.D., Walker, A.M., Hoyle, J.E., Northcott, S.J., Walker, F.W. & Ritchie, M.G. (2000).

Infestations of wild adult Atlantic salmon (Salmo salar L.) by the ectoparasitic copepod sea louse Lepeophtheirus salmonis Krøyer: prevalence, intensity and the spatial distribution of males and females on the fish. Hydrobiologia 429, 181-196.

Thorstad, E.B., Økland, F., Finstad, B., Sivertsgård, R., Plantalech, N., Bjørn, P.A. & McKinley, R.S. (2005) Comparing migratory behaviour and survival of wild and hatchery-reared Atlantic salmon and wild anadromous brown trout postsmolts during the first stages of marine migration. In press to Hydrobiologia.

Thorstad, E.B., Økland, F., Finstad, B., Sivertsgård, R., Bjørn, P.A., McKinley, R.S. (2004). Migration speeds and orientation of Atlantic salmon and sea trout post-smolts in a Norwegian fjord system.

Environmental Biology of Fishes 71, 305-311.

Tully, O & Nolan, D.T. (2002). A review of the population biology and host-parasite interactions of the sea louse Lepeophtheirus salmonis (Copepoda: Caligidae). Parasitology 124, 165-182.

Tully, O., Poole, W.R. & Whelan, K.F. (1993). Infestation parameters of Lepeophtheirus salmonis (Krøyer) (Copepoda: Caligidae) parasitic on sea trout Salmo trutta L., off the west coast of Irland during 1990 and 1991). Aquaculture and Fisheries Management 24, 545-555.

Vikeså, V. (2000). Påvirkning av saltholdighet og temperatur på tidlige livsstadier av lakselus, Lepeophtheirus salmonis. Mc thesis, University of Bergen, Norway. 1-69. (In Norwegian).

Wagner, G., McKinley, R.S., Bjørn, P.A., & Finstad, B. (2004). Short-term freshwater exposure benefits sea lice-infected Atlantic salmon. Journal of Fish Biology 64, 1593-1604.

Wells, A., Grierson, C.E., MacKenzie, M., Russon, I.J., Reinardy, H., Middlemiss, C., Bjørn, P.A., Finstad, B., Wendelaar Bonga, S.E., Todd, C.D. & Hazon, N. (2006) The physiological effects of simultaneous abrupt seawater entry and sea lice (Lepeoptheirus salmonis Krøyer) infestation of wild sea trout (Salmo trutta L.) smolts. Canadian Journal of Fisheries and Aquatic Sciences. submitted 99

SUMBAWS

Final Report

Paper2: SPATIAL INFESTATION PRESSURE OF SALMON LICE, LEPEOPHTHEIRUS SALMONIS (KRØYER), IN A NORWEGIAN FJORD SYSTEM: CAN SENTINEL CAGES REVEAL HIGH RISK AREAS? R. Sivertsgård, P. A. Bjørn, B. Finstad, F. Økland and R. Kristoffersen (Manuscript) Abstract Differences in infestation risks of salmon lice (Lepeophtheirus salmonis K.) for Atlantic salmon (Salmo salar L.) and sea trout (Salmo trutta L.) post-smolt were studied in three connected fjord sections in a Norwegian fjord system. The inner half of this fjord system is protected from salmon farming through the establishment of a salmon farming protection zone, while the outer half of the fjord system is outside this zone. This protection zone is designed to give wild salmonids protection from farmed salmon lice. No new fish farms can be established within this protection zone and any farms that were already established must operate under a stricter operation regime. Eresfjord in the inner part of this fjord system is free from salmon farming activity and the nearest farms are in Langfjord, while Karlsøyfjord, in the outer area of this fjord system and outside the protection zone, contains a large number of salmon farms. The large number of salmon farms in the outer reaches of the Romsdalsfjord system causes an increased risk of infestation.

Therefore, the infestation risks throughout the fjord system, as well as between littoral and pelagic areas, were investigated. Atlantic salmon post-smolts were placed in sentinel cages and after a fortnight, the fish were sampled and salmon lice infestation examined. The results showed that there were significant differences in infestation pressure between the three fjord sections. Eresfjord (inner part of the fjord system) had the lowest infestation risks, Karlsøyfjord (outer part of the fjord system) had the highest infestation risks, while the risk of infestation in Langfjord (middle part of the fjord system) was between that of Eresfjord and Karlsøyfjord. No clear difference between littoral and pelagic areas of the fjords were found, although significant differences between pelagic and littoral cages occasionally were observed both in Langfjord and Karlsøyfjord. Results show that sentinel cages can be used as an alternative method of monitoring the infestation pressure in a fjord system. The advantage of using sentinel cages is the capacity to measure the infestation pressure in discrete isolated areas. However, the infestation pressure measured by use of sentinel cages tends to underestimate that found in wild free swimming fish. Thus, the method must be calibrated if used as alternative to gill-net fishing or trawling for free swimming salmonids.

Introduction During the last decades, the marine parasite salmon lice Lepeophtheirus salmonis (Krøyer) has become one of the most serious pathogens on outward migrating post-smolts (Tully and Nolan, 2002; Heuch et al., 2005). This parasite may impair fish from reaching feeding areas and breeding grounds, finding food, avoiding predators and reducing the spawning success. Intensive fish farming in fjords and coastal areas has led to higher concentrations of infective salmon lice (MacKinnon, 1997; Bjørn et al. 2001, Heuch and Mo, 2001). The increasing salmon lice production in fish farms has been suggested to contribute to high infestation pressure on wild fish (Todd et al., 2000; Butler and Watt, 2003; Gargan et al., 2003). Salmon lice infestation historically has mostly been recorded and compared between separate fjords (Todd et al., 2000; Tully et al., 1993; Bjørn et al., 2001; Butler and Watt, 2003; Gargan et al., 2003; Heuch et al., 2005) rather than in different sections of the same fjord. However, different sea trout populations in the Romsdalsfjord system in Norway are exposed to different infestation pressure (Sivertsgård et al., 2005b sub.). Fish sampled in littoral areas in the fjord section with many fish farms showed significantly higher infestation, compared with fish from the farm-free section. However, physical and biological conditions in fjords are constantly changing and the infestation pressure may alter both in time and space (L.Asplin, Institute of Marine Research, Bergen, Norway, pers. comm.) Thus, fish can be exposed to different infestation pressures throughout the fjord system due to changes in infestation pressure existing between different parts of the fjord system. In this case the infestation pressure will be different for different species or populations if they utilise different parts of the fjord.

Wild sea trout post-smolts in Norway frequently stay in their home fjord and often near the home river outfall during their entire period in the sea (Jensen, 1968; Lyse et al., 1998; Finstad et al., 2005; Sivertsgård et al., 2005a sub.; Thorstad et al., 2005 sub.), though some larger fish conduct long migrations (Berg and 100

SUMBAWS

Final Report

Jonsson, 1990; Finstad et al., 2005; Sivertsgård et al., 2005a sub.; Thorstad et al., 2005 sub.). Post-smolts prefer areas closer to the shore line (Lyse et al., 1998; Thorstad et al., 2004; Thorstad et al., 2005 sub.) with scattered distribution in the pelagic areas (Rikardsen and Amundsen, 2005) and often remain at the surface, no deeper than 3-5 m (Lyse et al., 1998; Rikardsen et al., unpubl.data). On the other hand, Atlantic salmon (Salmo salar L.) post-smolts in Norway migrate from the river to the open ocean in the middle as well as closer to the shore of the fjord before returning to the home river as adults (Thorstad et al., 2004; 2005 sub.).

The time period the salmon smolts migrate in fjords is restricted to a short period after they have left the home river as smolts (Finstad et al., 2005; Sivertsgård et al., 2005a sub.; Thorstad et al., 2004; 2005 sub.).

Thus, a higher infestation pressure in the middle of the fjord is likely to affect Atlantic salmon post-smolts more than sea trout post-smolts and vice versa.

The main objective of this study was to investigate differences in infestation pressure between inner and outer fjord areas, and between littoral and pelagic areas of a Norwegian fjord system, using sentinel cages as a method of monitoring natural infestation pressure.

Material and methods Study area Eresfjord, Langfjord and Karlsøyfjord are three connected fjord areas in the Romsdalsfjord system (62°40’N8°10’E) in Møre & Romsdal county, mid Norway (Figure 1). The Romsdalsfjord system has extensive salmon farming activity, but most of the fish farms are localized in outer part of the fjord system, with only a few farms in the inner part. These farms drastically increase the number of salmon lice hosts in the outer parts of the Romsdalsfjord system. The inner half of the Romsdalsfjord system is incorporated in a protection zone (Anon, 2004). Thus, salmon lice infestation pressure may vary longitudinally from outer to inner areas of the fjord system (Sivertsgård et al., 2005b in prep.), and may also vary between littoral and pelagic areas in the fjord. The difference between low and high tide is approximately 1.6 m (Finstad et al., 2005).

Figure 1 Map of the Romsdalsfjord system. Squares show the location of littoral and pelagic sentinel cages in the inner (Eresfjord), middle (Langfjord) and outer (Karlsøyfjord) areas of the Romsdalsfjord system.

Circles show the location of the fish farms in operation in 2004. The hatched area of the fjords shows the extent of the salmon farming protection zone.

Design and deployment of sentinel cages A total of 12 sentinel cages were deployed in duplicate (less than 20 m from each other) in both littoral and pelagic areas in Eresfjord, Langfjord and Karlsøyfjord (Figure 1). The cage rig consists of a 25 kg anchor in littoral areas and a 100 kg (0.6 × 0.6 m) iron plate in pelagic areas. A 5 m, 20 kg steel chain is 101

SUMBAWS

Final Report

attached to the anchor and a 14 mm nylon rope connects the end of the chain to a floating buoy at the surface. The cage is circular and 0.8 m wide and 0.9 m high. The lower and upper ends are held open by a circular polyethylene tube (diameter 1.5 cm). The cage is covered by 1 × 1 cm knotless mesh with an opening on the top of the cage (Figure 2). The top of the cage is connected to the anchor rope 1 m under the buoy through the upper and lower polyethylene ring. The rope used to close the upper opening of the cage is also connected directly to the floating buoy. This, together with a 1.0 kg weight connected to the centre bottom of the cage, ensures that the cage remains at its maximal volume. To ensure that the cage is positioned vertically in the water, a further 5.0 kg weight is connected to the anchor rope 2 m below the cage.

Salinity and temperature were measured every 50 cm from the surface down to 5 m as close as possible to the sentinel cages in Eresfjord, Langfjord and Karlsøyfjord.

Figure 2 Sentinel smolt cage with Atlantic salmon post-smolts

Sampling and analyses 30, two-year old hatchery-reared Atlantic salmon post-smolts from the Statkraft hatchery in Eresfjord with wild salmon parents from the River Eira were placed in each sentinel cage. A seawater tolerance test (Blackburn and Clarke, 1987) performed on the hatchery-reared salmon post-smolts on 6 May revealed plasma chloride levels at 147 mM, indicating that the fish had smolted and were ready to be released into seawater (Sigholt and Finstad, 1990). The fish were starved for 24 h before transport to reduce stress in fish due to handling and transport (Wedemeyer, 1997). The salmon smolts were transported in plastic bags (500 × 900 mm) filled with oxygenated salt water (34‰). The bags were cooled (5 – 7 ºC) and covered so the fish were transported in darkness to reduce the transport stress (Wedemeyer, 1997) from the hatchery to the sentinel cages. The first trial was performed in weeks 19 – 21 (early May) and the second trial was performed during week 22 – 24 (May/June) in 2004. After 14 days in the sentinel cages, the fish were carefully removed from the cages, placed in individual plastic bags and frozen for subsequent analyses. In the laboratory, fish were thawed and salmon lice were examined according to Bjørn et al. (2001). Fork length (LF) was measured to the nearest mm and the fish were weighed to nearest 0.1 g.

Statistical tests and handling of data The ecological terms used in this study; prevalence, abundance, intensity, median and abundance range follow Bush et al. (1997). Statistical analyses were performed with SPSS 12.0 and the GPS mapping software OziExplorer 3.90. 3a. The GPS position of every fish farm in whole fjord system was taken from the Norwegian Fishery Directorate.

102

SUMBAWS

Final Report

Results Salinity and temperature Table 1 Mean (±SD) salinity and temperature measured every 50 cm from the surface down to 5 m near the sentinel cages in Eresfjord, Langfjord and Karlsøyfjord in 2004

Eresfjord Langfjord Karlsøyfjord

Littoral cages Salinity (‰) Temperature (ºC) (mean ± SD) (mean ± SD) 22.9 ± 3.7 11.0 ± 1.2 23.0 ± 0.0 11.4 ± 0.0 29.5 ± 0.4 10.7 ± 0.3

Pelagic cages Salinity (‰) Temperature (ºC) (mean ± SD) (mean ± SD) 18.8 ± 5.0 11.8 ± 7.3 26.9 ± 3.1 10.1 ± 1.0 28.4 ± 1.3 13.2 ± 2.8

Infestation of salmon lice in the fjord section without salmon farms Prevalence and abundance were low both in littoral and pelagic areas of Eresfjord in the first deployment of Atlantic salmon post-smolts in the sentinel cages (Table 2). There were no significant differences in abundance between the two littoral cages (Mann-Whitney U-test, p = 0.44). Prevalence was 11.5% and 5.0%, respectively; mean intensity was 1.0 in both cages. Low prevalence (15.4 and 15.0, respectively) and intensity (1.0 in both cages) was also found in the two pelagic cages. No significant differences were found between the two pelagic cages (Mann-Whitney, p = 0.97).

Extremely low infestations were found in the second deployment in Eresfjord. The fish were uninfested in three out of four cages, and the last (pelagic cage) had a prevalence of 10% and mean intensity of 1 louse · fish-1 (Table 3). There was no difference in abundance between pelagic and littoral cages for the first (Mann-Whitney U-test; p = 0.34) (Figure 3) and the second deployments (Mann-Whitney U-test; p = 0.36) (Figure 4).

Infestation of salmon lice in the fjord section with low farming activity Prevalence in the two littoral cages was 23.8% and 40.0% (Table 4); mean intensity was 1.0 and 1.5, respectively. No significant differences in abundance were found between the two littoral cages (MannWhitney U-test; p = 0.08) and the two pelagic cages (Mann-Whitney U-test; p = 0.20). One of the pelagic cages had a prevalence of 4.8%, while 45.8 % of the fish were infested in the other cage. Mean intensity was 1.0 and 1.25 in the two pelagic cages. Low infestation was also found in the two littoral sentinel cages in the second deployment. Prevalence was 14.3% and 20.0% and mean intensity was 1 in both cages, and no significant difference in abundance was found between the two littoral cages (Mann-Whitney U-test, p = 0.80). Prevalence in the pelagic cage (one cage was lost during the trial) was 42.9%, mean intensity was 1.3 (Table 5). Abundance was highest in the littoral cages compared with the pelagic cages in the first deployment in Langfjord (Mann-Whitney U-test; p = 0.03) (Figure 3), but this was opposite in the second deployment (Mann-Whitney U-test; p = 0.03) (Figure 4).

Table 2 Infestation parameters from the first deployment of Atlantic salmon post-smolt in sentinel cages in Eresfjord.

Eresfjord Cage Week of sampling Number of fish Mass (g) mean ± SD Length (mm) mean ± SD Prevalence (%) Abundance mean ± SD Intensity mean ± SD median IQR Range

Littoral location 1 2 21 21 26 20 57.32 ± 16.18 58.90 ± 18.71 17.37 ± 1.68 17.56 ± 1.58 11.53 5 0.12 ± 0.32 0.05 ± 0.05 1.00 ± 0.0 1 1 1.00 – 3.00

1.00 ± 0.0 1 1 1.00 – 1.00

Pelagic location 3 4 21 21 26 20 62.84 ± 29.77 54.80 ± 22.39 17.5 ± 2.69 16.86 ± 2.24 15.38 15 0.15 ± 0.37 0.15 ± 0.37 1.00 ± 0.0 1 1 1.00 – 1.00

1.00 ± 0.0 1 1 1.00 – 1.00 103

SUMBAWS

Final Report

Table 3 Infestation parameters from the second deployment of Atlantic salmon post-smolt in sentinel cages in Eresfjord.

Eresfjord Cage Week of sampling Number of fish Mass (g) mean ± SD Length (mm) mean ± SD Prevalence (%) Abundance mean ± SD Intensity mean ± SD median IQR Range

Littoral location 1 2 24 24 10 18 44.40 ± 16.18 44.44 ± 13.04 16.38 ± 2.03 16.37 ± 1.61 0 0

Pelagic location 3 4 24 24 22 10 57.09 ± 27.01 54.80 ± 19.69 17.31 ± 2.69 17.45 ± 2.08 0 10 0.10 ± 0.32 1.00 ± 0.0 1 1 1.00 – 1.00

Table 4 Infestation parameters from the first depoyment of Atlantic salmon post-smolt in sentinel cages in Langfjord 2004.

Langfjord Cage Week of sampling Number of fish Mass (g) mean ± SD Length (mm) mean ± SD Prevalence (%) Abundance mean ± SD Intensity mean ± SD median IQR Range

Littoral location 5 6 21 21 26 25 52.8 ± 15.6 72.64 ± 33.63 17.00 ± 1.43 18.51 ± 2.67 23.81 40 0.19 ± 0.42 0.60 ± 0.91 1.00 ± 0.0 1 1 1.00 – 1.00

1.50 ± 0.85 1 1.25 1.00 – 3.00

Pelagic location 7 8 21 21 21 24 55.71 ± 14.13 60.5 ± 23.01 17.22 ± 1.37 17.37 ± 1.89 4.76 45.83 0.05 ± 0.22 0.21 ± 0.50 1.00 ± 0.0 1 1 1.00 – 1.00

1.25 ± 0.50 1 0.75 1.00 – 2.00

Table 5 Infestation parameters from the second deployment of Atlantic salmon post-smolt in sentinel cages in Langfjord 2004.

Langfjord Cage Week of sampling Number of fish Mass (g) mean ± SD Length (mm) mean ± SD Prevalence (%) Abundance mean ± SD Intensity mean ± SD median IQR Range

Littoral location 5 6 24 24 14 20 50.14 ± 31.92 44.50 ± 14.03 16.58 ± 2.58 16.29 ± 1.77 14.28 20 0.14 ± 0.36 0.20 ± 0.41 1.00 ± 0.0 1 1 1.00 – 1.00

1.00 ± 0.0 1 1 1.00 – 1.00

Pelagic location 7 8 24 24 21 0 52.57 ± 13.07 17.54 ± 1.52 42.85 0.57 ± 0.81 1.33 ± 0.71 1 0.5 1.00 – 3.00 104

SUMBAWS

Final Report

Infestation of salmon lice in the fjord section exposed to high farming activity Prevalence in the first deployment in Karlsøyfjord was high: 88% and 83%, respectively, of the fish in the two littoral cages were infested. Mean intensity was 2.1 and 1.4, and there was no difference in abundance between the littoral cages (Mann-Whitney U-test, p = 0.07). 84 % of the fish in the pelagic cage were infested (one cage has been opened by intruders and the fish released), with a mean intensity of 2.6 (Table 6). Relatively high infestations were also found in the second deployment in Karlsøyfjord.

Prevalence was 80.0% and 87.0%, mean intensity was 2.9 and 2.6 in the two littoral cages, respectively, and there were no significant differences between the two littoral cages (Mann-Whitney U-test, p = 0.80) (Table 7). 35.7% and 68.2 % of the fish in the pelagic cages was infested with salmon lice. Mean intensity was 1.0 and 1.9 and the abundance was different between the pelagic cages (Mann-Whitney U-test, p = 0.01) (Table 7). There was no difference in abundance between the littoral and pelagic cages in the first deployment (Mann-Whitney U-test; p = 0.06) (Figure 3). In the second deployment, abundance was highest in littoral cages compared with pelagic cages (Mann-Whitney U-test; p = 0.001) (Figure 4).

Table 6 Infestation parameters from the first deployment of Atlantic salmon post-smolt in sentinel cages in Karlsøyfjord.

Karlsøyfjord Cage Week of sampling Number of fish Mass (g) mean ± SD Length (mm) mean ± SD Prevalence (%) Abundance mean ± SD Intensity mean ± SD median IQR Range

Littoral location 9 10 21 21 25 18 65.44 ± 18.32 50.77 ± 19.98 18.02 ± 1.72 16.80 ± 2.25 88 83.33 1.84 ± 1.31 1.17 ± 0.71 2.09 ± 1.19 1 2 1.00 – 6.00

Pelagic location 11 12 21 21 0 25 52.57 ± 13.07 17.46 ± 2.25 84 2.0 ± 2.16

1.40 ± 0.51 1 1 1.00 – 2.00

2.62 ± 2.1 1 2 1.00 – 10.00

Table 7 Infestation parameters from the second deployment of Atlantic salmon post-smolt in sentinel cages in Karlsøyfjord.

Karlsøyfjord Cage Week of sampling Number of fish Mass (g) mean ± SD Length (mm) mean ± SD Prevalence (%) Abundance mean ± SD Intensity mean ± SD median IQR Range

Littoral location 9 10 24 24 10 23 46.20 ± 14.49 67.00 ± 30.6 16.42 ± 1.97 17.33 ± 2.46 80 86.95 2.30 ± 2.21 2.26 ± 1.60 2.88 ± 2.10 2 3.25 1.00 – 7.00

2.60 ± 1.43 2.5 2.75 1.00 – 5.00

Pelagic location 11 12 24 24 14 22 43.57 ± 15.19 50.54 ± 13.51 16.06 ± 1.93 17.03 ± 1.55 35.71 68.18 0.13 ± 0.49 1.32 ± 1.17 1.00 ± 0.00 1 0 1.0 – 1.0

1.93 ± 0.88 2 1 1.00 – 3.00

105

SUMBAWS

Final Report

Comparisons of infestation in littoral and pelagic areas in the three fjord sections Significant differences in lice abundance between the littoral cages were found between Eresfjord, Langfjord and Karlsøyfjord (Kruskal-Wallis χ2- test, p < 0.0001) in the first deployment. Fish in Karlsøyfjord had the heaviest infestation, followed by Langfjord and finally Eresfjord where fish had the lowest infestation. Abundance in littoral areas was approximately 1.6 lice · fish-1 in Karlsøyfjord, while abundance in Langfjord was 0.39 and for Eresfjord it was 0.09. There was also a significant difference in abundance between the pelagic cages in Eresfjord, Langfjord and Karlsøyfjord (Kruskal-Wallis χ2- test, p < 0.0001). Karlsøyfjord had the highest abundance in the pelagic areas of the three fjord sections. Abundance in the pelagic areas was approximately 1.9 lice · fish-1 in Karlsøyfjord, while abundance in Langfjord and Eresfjord was 0.1.

Significant differences in abundance were also found between the littoral areas in Eresfjord, Langfjord and Karlsøyfjord (Kruskal-Wallis χ2- test, p < 0.0001) in the second deployment. Abundance in littoral areas was highest in Karlsøyfjord, then Langfjord and lowest in Eresfjord. Littoral abundance in Karlsøyfjord was 2.3, Langfjord 0.2 and Eresfjord 0.0 lice · fish-1. Differences in abundance between the three pelagic areas were found (Kruskal-Wallis χ2- test, p < 0.0001); Karlsøyfjord had the highest abundance, then Langfjord and lowest Eresfjord. Abundance in the pelagic areas was approximately 0.9 lice · fish-1 in Karlsøyfjord, 0.6 in Langfjord and 0.3 in Eresfjord.

2,5 P = 0.06

Abundance

Abundance

2,0

1,5

1,0

P = 0.03 0,5 P = 0.34

0,0 Eresfjord

Langfjord

Karlsøyfjord

Figure 3 Abundance (mean ± SD) of salmon lice in the farm-free Eresfjord, the low farmed Langfjord and the high farmed Karlsøyfjord in the first deployment. Black bars represent salmon lice abundance in littoral deployed sentinel cages, and grey bars represent salmon lice abundance in pelagic deployed sentinel cages.

P-values (Mann-Whitney U-test) are given for comparison between abundance in littoral and pelagic cages.

3,0

P = 0.001

2,5

Abundance

Abundance

2,0

1,5 P = 0.03 1,0

0,5

P = 0.36

0,0 Eresfjord

Langfjord

Karlsøyfjord

Figure 4 Abundance (mean ± SD) of salmon lice in the farm-free Eresfjord, the low farmed Langfjord and the high farmed Karlsøyfjord in the second deployment. Black bars represent salmon lice abundance in littorally deployed sentinel cages, and grey bars represent lice abundance in pelagic deployed sentinel cages.

P-values (Mann-Whitney U-test) are given for comparison between abundance in littoral and pelagic cages.

106

SUMBAWS

Final Report

Discussion The present study indicates that the risk of salmon lice infestation differs between the farm-free Eresfjord, the intermediately farmed Langfjord and the intensively farmed Karlsøyfjord. The observed differences may be explained through the different farming level in the fjords. The low, but persistent, infestation level in the farm protection zone of Langfjord and Eresfjord, implies that this zone reduces infestation pressure.

Annual salmon lice epidemics have been regarded by Norwegian authorities as a major threat to the long term survival of wild salmonid populations in areas with intensive fish farming activity (Anon, 2004). As a precautionary approach, it was therefore decided to establish “National Salmon Fjords” in a number of fjords with important salmon rivers (Anon, 2004). In these fjords, no new salmon farms were allowed to be established, and those already established were required to operate under a stricter operational regime.

However, possible effects of “National Salmon Fjords” on e.g. the risks of salmon lice infestation in wild salmonids have not been investigated.

Recent results on wild sea trout (Sivertsgård et al., 2005b in prep.) in the Romsdalsfjord system indicate that the protection zone may be beneficial to wild sea trout populations in Eresfjord. Results from both 2003 and 2004 show that salmon lice infestation pressure is significantly higher outside (intensively farmed Karlsøyfjord), than inside the protection zone of Eresfjord (Sivertsgård et al., 1995b in prep.). However, perhaps surprisingly, high infestations of wild sea trout were also found in the protection zone of Eresfjord compared to levels previously observed in completely farm-free fjord systems in Norway (Heuch and Mo, 1998; Bjørn and Finstad, 2002; Rikardsen, 2004; Heuch et al., 2005). This difference can be explained both through horizontal dispersal of infective lice stages (Heuch et al., 2005), but also by sea trout undertaking long seawater migrations (Berg and Jonsson, 1990) and having spent parts of their seawater journey in the higher risks areas of the Romsdalsfjord system.

The present study of caged fish in different areas of the Romsdalsfjord system confirms our previous results (Sivertsgård et al. 2005b in prep.). The infestation risk was highest in the intensively farmed Karlsøyfjord, gradually reduced in the intermediately farmed Langfjord and lowest in the farm free Eresfjord. This implies that a salmon farming protection zone reduces lice infestation pressure, and is in agreement with Sivertsgård et. al. (in prep). However, no clear difference between littoral and pelagic areas of the fjord were found, although significant differences between pelagic and littoral cages occasionally were observed both in Langfjord and Karlsøyfjord. The dispersal of infective lice stages in natural fjord systems is a complicated process and little is still known about the details of lice dispersal patterns. Infective stages are expected to aggregate in steep salinity gradients (Heuch et al., 1995), avoid salinity < 20 ‰ (Heuch et al., 1995), and aggregate in areas with turbulent water, but they can also be transported over very long distances (L.Asplin, Institute of Marine Research, Bergen, Norway, pers. comm.). Thorstad et al. (2005 sub.) used telemetry to compare the migratory routes of sea trout and Atlantic salmon post-smolts. Sea trout tend to migrate in near shore waters close to the littoral zone on both sides of the fjord. In contrast, Atlantic salmon post-smolts tend to migrate in more pelagic areas and more distant from shoreline (Thorstad et al., 2005 sub.). These findings are also in agreement with earlier studies (Hvidsten et al., 1992; Moore et al.

1995; Holm et al., 1998; Voegeli et al., 1998; Moore et al., 2000; Thorstad et al., 2004; Finstad et al., 2005; Rikardsen et al., unpubl. data). Since farms are usually located quite close to the littoral zone, it has been argued that infective stages of salmon lice may be concentrated in these areas thereby rendering sea trout at higher infestation risks. Results from the present study do not confirm this. On the contrary, lice infestation stages seem to be distributed both in the pelagic and littoral areas of the fjord, and imply that both Atlantic salmon post-smolts and sea trout are at risk of infestation.

Gill-netting of sea trout in the same time period and locality as caged Atlantic salmon post-smolts showed that wild fish are infected to a significantly higher degree than caged fish. Wild sea trout sampled in Eresfjord at the end of the first sentinel cage deployment (week 21), had a prevalence of 28% and intensity of 4.3 lice · fish-1 (Sivertsgård et al., 2005b in prep.). Prevalence was 64 % and intensity was 19.0 at the end of the second cage deployment (week 24). Similar results were also found in Karlsøyfjord. Prevalence and intensity in wild sea trout at the end of first and second cage deployment was 83% and 6.8 and 86% and 21.3, respectively (Sivertsgård et al., 2005b in prep.). In both cases, sentinel cages detect the differences in infestation pressure between areas, although they seem to underestimate the infestation level acquired by wild fish during the same period. However, sentinel cages can be used as an alternative method of investigating risk and consequences of lice infestation provided that infestation levels are better calibrated against levels in wild fish.

107

SUMBAWS

Final Report

Acknowledgements We are pleased to acknowledge the help of numerous people throughout this study. We are grateful to the staff at the Statkraft hatchery in Eresfjord, Bjørg Anne Vike, Petter Sira and Torbjørn Utigard, for extensive help and co-operation during the project. As well we want to thank Tore Øverland and Stig Sandring for their help during the field work. The study was financed by the European Commission contract no: Q5RS2002-00730, the Norwegian Institute for Nature Research and Statkraft SF, Norway.

References Anonymous (2004) Nasjonale laksevassdrag og laksefjorder. Ferdigstilling av ordningen. (Downloadable: www.dirnat.no/archive/attachments/01/61/NLVNL053.pdf) Berg, O.K. & Jonsson, B. (1990) Growth and survival rates of the anadromus trout, Salmo trutta, from the Vardnes River northern Norway. Environmental Biology of Fishes 29, 145-154.

Bjørn, P.A. & Finstad, B. (2002) Salmon lice, Lepeophtheirus salmonis (Krøyer), infestation in sympatric poulations of Artic char, Salvelinus alpinus (L.) and sea trout, Salmo trutta (L.), in areas near and distant from farms. ICES Journal of Marine Science 59, 131-139.

Bjørn, P.A., Finstad, B. & Kristoffersen, R. (2001) Salmon lice infection of wild sea trout and Artic char in marine and freshwaters: the effects of salmon farms. Aquaculture Research 32, 947-962.

Blackburn, J. & Clarke, W.C. (1987) Revised procedure for the 24 hour seawater challenge test to measure seawater adaptability of juvenile salmonides. Canadian Technical Report of Fisheries and Aquatic Sciences 1515, 1-39.

Bush, A. O., Lafferty, K. D., Lotz, J. M. & Shostak, A. W. (1997) Parasitology meets ecology on its own terms: Margolis et al. revised. Journal of Parasitology 83, 575-583.

Butler, J.R.A. & Watt, J. (2003) Assessing and Managing the impacts of Marine Salmon Farms on Wild Atlantic Salmon in Western Scotland: Identifying Priority Rivers for Conservations. In Salmon at the Edge (Mills eds.). Blackwell Science, Oxford, 93-118.

Finstad, B., Økland, F., Thorstad, E. B., Bjørn, P. A. & McKinley, R. S. (2005) Migration of hatcheryreared Atlantic salmon and wild anadromous brown trout postsmolts in a Norwegian fjord system. Journal of Fish Biology 65, 1-11.

Gargan, P.G., Tully, O & Poole, W.R. (2003) Relationship between sea lice infestations, sea lice productions and sea trout survival in Ireland, 1992-2001. In Salmon at the Edge (Mills eds.) Blackwell Science, Oxford, 119-135.

Heuch, P. A. & Mo, T. A. (2001) A model of salmon louse production in Norway: Effects of increasing salmon production and public management measures. Diseases of Aquatic Organisms 45, 145–152.

Heuch, P.A., Parsons, A. & Boxaspen, K. (1995) Diel vertical migration: A possible host-finding mechanism in salmon louse (Lepeophtheirus salmonis) copepodids? Canadian Journal of Fisheries and Aquaculture 52, 681-689.

Heuch, P.A., Bjørn. P.A., Finstad, Holst, J.C., Asplin, L. & Nilsen, F. (2005) A review of the Norwegian ”National Action Plan Against Salmon Lice on Salmonids”: The effect on wild salmonids. Aquaculture 246, 79-92.

Holm, M., Axelsen, B.E., Sturlaugsson, J., Hvidsten, N.A., Ikonen, E. & Johnsen, B.O. (1998).

Behaviour of acoustically tagged post-smolts in the Trondheim fjord – Influence of hydrografical and metrological conditions on migration. International Council for the Exploration of the sea CM 1998/N:17.

Hvidsten, N.A., Johnsen, B.O. & Levings, C.D. (1992) Atferd og ernæring hos utvandrende laksesmolt i Trondheimsfjorden (Behaviour and feeding in postsmolts from the Trondheimsfjord). NINA Oppdragsmelding 164. Report in Norwegian with English summary. Norwegian Institute for Nature Research, Tungasletta 2, N-7485 Trondheim, Norway, 1-14 (in Norwegian).

Jensen, K.W. (1968) Sea trout (Salmo trutta L.) of the river Istra, western Norway. Report to the Institute of Freshwater Research Drottningholm 48, 187-213.

108

SUMBAWS

Final Report

Lyse, A.A., Stefansson, S.O. & Fern, Ö. A. (1998) Behaviour and diet of sea trout post-smolts in a Norwegian fjord system. Journal of Fish Biology 52, 923-936.

MacKinnon, B.M. (1997) Sea lice: a review. World Aquaculture September 1997, 5-10.

Mo, T.A. & Heuch, P.A. (1998) Occurrence of Lepeophthheirus salmonis (Copepoda: Caligidae) on sea trout (Salmo trutta) in the inner Oslo Fjord, south-easern Norway. ICES Journal of Marine Science 55, 176180.

Moore, A., Potter, E.C.E., Milner, N.J. & Bamber, S. (1995) The migratory behaviour of wild Atlantic salmon smolts in the estuary of the River Conwy, North Wales. Canadian Journal of Fisheries and Aquatic Sciences 52, 1923-1935.

Moore, A., Lacroix, G.L. & Sturlaugsson, J. (2000) Tracking Atlantic salmon post-smolts in the sea. In The ocean life of Atlantic salmon - environmental and biological factors influencing survival (Mills, D. Ed.). pp. 49-64. Oxford: Fishing News Books.

Planateck et al in prep.

Rikardsen, A.H. (2004) Seasonal occurrence of sea lice Lepeophtheirus salmonis on sea trout in two north Norwegian fjords. Journal of Fish Biology 65, 711-722.

Rikardsen, A.H. & Amundsen, P.A. (2005) Pelagic marine feeding behaviour of Arctic charr Salvelinus alpinus and sea trout Salmo trutta. Journal of Fish Biology 66, 1163-1166.

Sigholt, T. & Finstad, B. (1990) Effects of low temperature on sea-water tolerance in Atlantic salmon (Salmo salar) smolts. Aquaculture 84, 167-172.

Sivertsgård, R., Thorstad., E.B., Økland, F., Finstad, B., Bjørn, P.A., Jepsen, N., Nordal., T. & McKinley., S.R. (2005a) Effects of salmon lice infection and salmon lice protection on fjord migrating Atlantic salmon and brown trout post-smolts. Submitted to Hydrobiologia.

Sivertsgård, R., Bjørn., P.A., Finstad, B., Kristoffersen, R. & Øverland, T. (2005b) Fish farming, salmon lice, Lepeophtheirus salmonis Krøyer, and protective zones; effects on wild sea trout, Salmo trutta Linne, populations. In prep.

Thorstad, EB., Økland, F., Finstad, B., Sivertsgård, R., Bjørn, P.A., McKinley, R.S. (2004) Migration speeds and orientation of Atlantic salmon and sea trout post-smolts in a Norwegian fjord system.

Environmental Biology of Fishes 71, 305-311.

Thorstad, E.B., Økland, F., Finstad, B., Sivertsgård, R., Plantalech, N., Bjørn, P.A. & McKinley, R.S. (2005) Comparing migratory behaviour and survival of wild and hatchery-reared Atlantic salmon and wild anadromous brown trout postsmolts during the first stages of marine migration. Submitted to Hydrobiologia.

Todd, C.D., Walker, A.M., Hoyle, J.E., Northcott, S.J., Walker, F.W. & Ritchie, M.G. (2000) Infestations of wild adult Atlantic salmon (Salmo salar L.) by the ectoparasitic copepod sea louse Lepeophtheirus salmonis Krøyer: prevalence, intensity and the spatial distribution of males and females on the fish. Hydrobiologia 429, 181-196.

Tully, O., Poole, W.R. & Whelan, K.R. (1993) Infestation parameters for Lepeophtheirus salmonis (Krøyer) (Copepoda: Caligidae) parasitic on sea trout, Salmo trutta L., off the west coast of Ireland during 1990 and 1991. Aquaculture and Fisheries Management 24, 545-555.

Tully, O. & Nolan, D. T. (2002) A review of the population biology and host-parasite interactions of the sea louse Lepeophtheirus salmonis (Copepoda: Caligidae). Parasitology 124, 165-182.

Voegeli, F.A., Lacroix, G.L. & Anderson, J.M. (1998) Development of miniature pingers for tracking Atlantic salmon smolts at sea. Hydrobiologia 371/372, 35-46.

Wedemeyer, G.A. (1997) Effects of rearing conditions on the health and physiological quality of fish in intensive culture. In Fish stress and health in Aquaculture. (Iwama, G.K. ed.), pp 35-71. Society for Experimental Biology Seminar Series 62 Cambridge University Press.

109

SUMBAWS

Final Report

Paper 3: DIFFERENCES IN RISKS AND CONSEQUENCES OF SALMON LICE, LEPEOPHTHEIRUS SALMONIS (KRØYER) INFECTION ON SYMPATRIC POPULATIONS OF ATLANTIC SALMON, BROWN TROUT AND ARCTIC CHAR WITHIN NORTHERN FJORDS P. A. Bjørn, B. Finstad, R. Kristoffersen, A.H. Rikardsen and R. S. McKinley (Submitted to ICES Journal of Marine Sciences) Abstract Differences in salmon lice (Lepeophtheirus salmonis) infection on sympatric populations of fjord migrating Atlantic salmon post-smolts (Salmo salar), brown trout (Salmo trutta) (sea trout) and Arctic char (Salvelinus alpinus) was studied in three fjords in northern Norway with fish farming activity during June – August 2000. Atlantic salmon post-smolts were only captured in the fjords during late June and early July and probably left the fjords after this period. None of these fish were infected with salmon lice. In contrast, brown trout and Arctic char had similar infection patterns during their sampling periods, with very low prevalence and mean infection intensity during June (0-21% and 0-6 lice · fish-1, respectively), slightly increasing in July (8-70% and 6-12 lice · fish-1, respectively), and peaking in August (80-88% and 19-27 lice · fish-1, respectively). The chalimus stages dominated during June and July with a few pre-adult and adult stages observed in July, while all stages were found frequently during August. These observations indicate that Atlantic salmon seemingly may have a mismatch between time of lice infection and their postsmolt fjord migration in northern fjords. In contrast, brown trout and Arctic char feed within the fjords throughout the summer and consequentially have higher risk of harmful infections in years with suitable environmental conditions for salmon lice development, especially in fish farming areas. Arctic char usually spend the shortest time at sea of these three species, and in many occasions the salmon lice may probably not have time to develop to the adult stages on this species.

Introduction The salmonid species in the Northern Hemisphere, Atlantic salmon (Salmo salar L.), brown trout (Salmo trutta L.) (sea trout) and Arctic char (Salvelinus alpinus L.), have an anadromous life history pattern utilising both fresh- and seawater habitats (Jonsson, 1985; Finstad & Heggberget, 1993, 1995; Rikardsen et al., 2000; 2004b; Klemetsen et al., 2003). The migratory life-history pattern includes both major changes in fish physiology and ecology (Hoar, 1988; Boeuf, 1993; Høgåsen, 1998), exposing the fish to a large variation of environmental and biological challenges. One of these challenges is the exposure to the ectoparasitic copepod salmon lice (Lepeophtherius salmonis Krøyer) (Kabata, 1974) in the marine environment (Bakke & Harris, 1998).

Historically salmon lice have been observed in low numbers on wild salmonids, and few adverse effects on the host reported (Boxshall, 1974; Johannessen, 1975; Pemberton, 1976; Wootten et al., 1982). However, since the late 1980s, there have been heavy infestations of salmon lice on anadromous brown trout (sea trout) along the coast of Norway (Birkeland, 1996; Bjørn et al., 2001b), Ireland (Tully et al., 1993a,b; Gargan et al., 2003) and Scotland (Todd et al., 1997; Butler, 2002). Infested trout, mainly post-smolts, have often been reported to be in poor physical condition and some with severely damaged caudal and dorsal fins, and have been observed returning to rivers and estuaries shortly after they have entered the sea (Tully et al., 1993a,b; Birkeland 1996). It has been suggested that the increased infestation rate of salmon lice on brown trout is a result of high lice levels on farmed salmonids in these areas (Tully & Whelan, 1993; Birkeland, 1996; Bjørn et al., 2001b; Butler, 2002; Tully & Nolan, 2002; Gargan et al., 2003). Furthermore, Norwegian investigations have indicated that lice larvae infections occur on fjord-migrating Atlantic salmon smolt (Finstad et al., 2000; Heuch et al., 2005) and Arctic char in areas with salmon farms (Bjørn et al., 2001b; Bjørn & Finstad, 2002).

The risks and consequences of salmon lice infection may, however, vary between the species (Bjørn & Finstad, 2002). This will depend both on the encounter rate as well as the susceptibility to infections, as well as their different life-history (Klemetsen et al., 2003). Given the frequently high numbers of gravid salmon lice carried by the large numbers of cultured fish throughout the year, it is likely that the development of a aquaculture industry has lead to changes in the natural host-parasite relationship, and made possible the production of large amounts of infective dispersal lice stages in addition to the natural production of lice on wild salmonids (Tully & Whelan, 1993; Bjørn et al., 2001b; Heuch & Mo, 2001a,b; Rikardsen, 2004; Heuch 110

SUMBAWS

Final Report

et al., 2005; Krkošek et al., 2005 ). As plankton, these larvae will drift and be dispersed over long distances, but apparently concentrate near the surface by day (Heuch et al., 1995), and probably also near pycnoclines in stratified waters (Heuch, 1995). The density of infective salmon lice stages are, therefore, likely to be greatest in inshore coastal areas and fjords that are subject to constrained tidal flushing. These locations are exploited by feeding and migrating post-smolts and facilitate increased encounter rates between the parasite and the host.

Their different migrating behaviour at sea may also have strong implications on the risk of salmon lice infection. In northern Norway, smolts of anadromous fish migrate to sea for the first time usually during a 23 weeks peak between late May and early July (Rikardsen et al., 1997; 2004a; Klemetsen et al., 2003; Tuff Carlsen et al. 2004). The timing of this period varies between species and populations, but the final decision when to migrate is determined by environmental factors like water temperature, light and water discharge during spring, resulting in annual variation in the peak of the migration for each species (Tuff Carlsen et al., 2004). At the same latitude, Arctic char and brown trout normally spend 1-2 summer months each year at sea before returning to freshwater, where the trout usually return some weeks later than the char (Jonsson, 1985; Rikardsen, 2000). In contrast, Atlantic salmon spends 1-3 years at sea before returning to spawn in freshwater (Klemetsen et al., 2003) and is assumed to move quickly throughout the fjord to the feeding areas in the open sea (Thorstad et al., 2004; Finstad et al., 2005). These species are therefore subject to different environmental challenges and sources of mortality at sea, including differing susceptibility to salmon lice infection.

Due to severe methodological difficulties in capturing Atlantic salmon post-smolts at sea (Holst & MacDonald, 2000; Rikardsen et al., 2004a), only one previous international refereed paper reports salmon lice infection levels in fjord migrating post-smolts (Finstad et al., 2000). In contrast there exist more infection data on Arctic char and especially brown trout (Boxshall, 1974; Tingley et al., 1997; Mo & Heuch, 1998; Bjørn et al., 2001b; Bjørn & Finstad, 2002; Heuch et al., 2002; Rikardsen, 2004), but to the best of our knowledge no study has reported data from sympatric populations of all these three species.

The purpose of this study was therefore to investigate the risk of salmon lice infection in sympatric populations of fjord migrating Atlantic salmon, brown trout and Arctic char in areas with fish farming activity in northern Norway. Possible differences in susceptibility between the species, ecological consequences of the infection and the relationship with the fish farming activity have been discussed.

Materials and Methods Study area Three study sites in northern Norway were selected in this study: the Altafjord in Finnmark County (Figure 1B) and Malangsfjord and Løksebotten in the Troms County (Figure 1A). The Altafjord had extensively fish farming activity in 2000, including 17 large salmon farms, and the fjord was classified to be relatively intensively farmed (Bjørn et al., 2001a). The inner part of the Malangsfjord system was without fish farms during the study period, and only three fish farms were situated in the outer part of the fjord system. The fjord was therefore classified to have low exposure to fish farming (Bjørn et al., 2001a). The farming activity south of Malangsfjord, especially the close area to Løksebotten, was however intensively farmed (Bjørn et al., 2001a). Large populations of both Atlantic salmon, sea trout and Arctic char are present inside the Altafjord system and a sampling program for all three species was established in 2000 (Figure 1B). Atlantic salmon post-smolts were also captured in the Malangsfjord system, while brown trout and Arctic char from this county were captured in Løksebotten, 70 km south of the Malangsfjord system (Figure 1A). See Rikardsen et al. (2004a) for further information of the Altafjord and Malangsfjord systems.

Sampling procedures and analyses Post-smolts of Atlantic salmon was captured by a newly developed pelagic trawl, the FISH-lift (Holst & MacDonald, 2000), while most of the Arctic char and brown trout where captured by gillnetting in the littoral zone as described in Bjørn et al. (2001b). The pelagic trawl has proven to be very efficient in capturing fjord migrating post-smolts, and fish were sampled in two periods during 2000: week 24 and 26 in Malangsfjord and week 25 and 27 in the Altafjord (Rikardsen et al., 2004a). The fjords were divided into three areas, and several trawl hauls were performed in pelagic areas of the fjords and along the assumed routes of migrating post-smolts (Figure 1A and 1B). Floating gill nets were set 45-90° to the shore both day and night to capture brown trout and Arctic char in their littoral feeding areas in June (week 25), July (week 28) and August (week 32) of the Altafjord (Figure 1B) and in Løksebotten fjord system (Figure 1A). The nets were 25 m long, 2 m deep, and with mesh sizes ranging from 19.5 to 35 mm (Bjørn et al. 2001b; Rikardsen, 2004).

111

SUMBAWS

Final Report

All fish were gently removed from the trawl and the gill-nets, and the fish were immediately placed into individually tagged plastic bags. Weight and total lengths were measured and the fish were examined for lice under a stereoscope as presented in Bjørn & Finstad (1998). Ecological terms recommended by Bush et al. (1997) were used: Prevalence is defined as the percentage of infested fish, abundance is the mean number of parasites per fish caught, median intensity is the median number of parasites per infested fish of a sample, and mean intensity is the average number of parasites per infested fish. Owning to lack of normal distribution and the highly aggregated distributions among the samples, non-parametric Mann-Whitney or Kruskal-Wallis tests were chosen for analyses of statistical differences. Furthermore, the variance (S2) to mean ratio (abundance) was also used to describe the degree of aggregation in the different samples. In all tests, a probability level of p ≤ 0.05 was considered significant.

Figure 1 The geographic location of the study in Finnmark and Troms country, northern Norway. Finnmark and Troms country is showed together with the field locality in Malangen/Løksebotten (A) (69°30' N – 18°20' E / 68°55' N – 17°41' E) and Altaford (B) (70°03' N – 23°05' E) where the fishing was conducted.

Atlantic salmon post-smolt trawling was performed in pelagic areas (grey areas) of different zones (hatched lines) in both the Malangsfjord (A) and the Altafjord system (B). Sea trout and Arctic char test fishing localities in the Malangsfjord/Løksebotten- and Altafjord system are shown by filled circles.

Fish Material In Altafjord, a total of 47 brown trout, 36 Arctic char and 153 Atlantic salmon post-smolts were captured with average weights of 240, 360 and 22 gram, respectively (Table 1). In Løksebotten and in Malangen the total catch was 19 brown trout, 30 Arctic char and 93 Atlantic salmon post-smolts with average weights of 170, 260 and 22 gram, respectively (Table 1). All brown trout and Arctic char were captured in the littoral zone, while all Atlantic salmon post-smolts were captured in the pelagic zone of the fjords. The largest fish captured were the few brown trout caught in June in the Altafjord. Most of these fish were maturing fish, while most of the other brown trout and Arctic char captured during the sampling periods were immature 112

SUMBAWS

Final Report

fish on their first- (post-smolt) or second sea migration. All the Atlantic salmon post-smolts were immature.

In general, Arctic char were captured most frequently in June and July, while most brown trout were captured in July and August. The Atlantic salmon were captured during early July in Malangsfjord (inner-, middle- and outer zone) and Altafjord (inner- and middle zone) while some fish were also captured in late June in Malangsfjord (inner zone).

Table 1 Fish sampled in the Altafjord (Atlantic salmon, brown trout and Arctic char), in the Malangsfjord (Atlantic salmon) and at Løksebotten (brown trout and Arctic char) localities. The mean weight ± standard deviation and the number (n) of Atlantic salmon, brown trout and Arctic char in each sampling week are presented.

Week 25/26 (June) 29 (July) 32/33 (August)

Species Sea trout Arctic char Atlantic salmon Sea trout Arctic char Atlantic salmon Sea trout Arctic char Atlantic salmon

Altafjord 626.3 ± 645.0 (6) 338.6 ± 150 (14) (0) 202.3 ± 213.7 (22) 407 ± 449.7 (15) 22.4 ± 4.6 (153) 161.7 ± 122.3 (19) 296.7 ± 260.2 (7) -

Malangsfjord/Løksebotten 61.0 ± 16.9 (2) 220.9 ± 206.2 (14) 23.1 ± 5.2 (10) 245.3 ± 342.57 (7) 289.3 ± 116.3 (16) 22.1 ± 6.7 (83) 134.2 ± 74.4 (10) (0) -

Results Sea trout and Arctic char There were no significant differences in infection intensity between sea trout and Arctic char from the same sampling occasions and locations in June and July (Mann-Whitney U-test; p > 0.05). The infection parameters on the two species were therefore pooled to present an overview of the salmon lice infection among hosts feeding in the littoral zone with time.

Both the prevalence and infection intensity differed significantly between the different sampling periods for the combined material of sea trout and Arctic char (Kruskal-Wallis; p < 0.05), and showed much of the same general pattern at both localities (Table 2). In June, all sea trout and Arctic char in the Altafjord were uninfected with salmon lice, while only 21 % of the fish in Løksebotten were infected and had low infections. In July, only 8% of the fish in Altafjord were infected with mean intensities of 12 lice. In contrast, the infection in Løksebotten had increased more in June with a prevalence of 70%, although the mean intensity was similar as in June (Kruskal-Wallis; p > 0.05). However, in August, both the prevalence and intensity in both fjords had increased significantly (Kruskal-Wallis; p < 0.05). In Løksebotten, 80 % of the fish were infected and had a mean intensity of 27 lice, while 88% of the fish in Altafjord were infected with a mean intensity of 19 lice. Maximum values in Løksebotten and Altafjord was 59 and 78, respectively.

Table 2 Infection intensity on pooled groups of sea trout and Arctic char from the Altafjord and Løksebotten in June (weeks 25 and 26), July (week 29) and August (weeks 32 and 33) 2000 (SW= sea water, n = total number of fish captured; Prev= Percentage (%) of infected fish of the total number of fish; mean= mean numbers of lice on infected fish only (intensity); SD = standard deviation; IQR= interquartile range; min and max = minimum and maximum number of lice; s2/ x = variance to mean ratio) SAMPLING WEEK (MONTH)

Habitat

n

Prev.

26 (June) 29 (July) 32 (August)

SW SW SW

20 37 26

0 8.1 88.4

25 (June) 29 (July) 33 (August)

SW SW SW

24 33 10

mean

SD

median

IQR

min

max

s2/ x

14.0

22

4 1

27 78

14.0 18.3

5.0 4.0 24.5

8.5 6 46

1 1 4

13 25 59

3.4 5.7 18.7

Altafjord 12.0 13.0 18.9 18.6 Løksebotten 20.8 6.4 4.6 69.7 5.9 5.9 80.0 26.5 22.3

113

SUMBAWS

Final Report

In general, the chalimus stages dominated during in June and July, but with a few pre-adult and adult stages observed in July (especially in Løksebotten), while all stages were found more frequently during the peak in August (Figure 2). In Altafjord in July, low numbers of early (I and II) and late chalimus (III and IV) stages dominated but only a few adult males were found on the fish in this month (Figure 2a). In August, higher numbers of especially early and late chalimus stages were found, but there had also been an aggregation of older lice stages on the fish in this fjord. The lice population on fish from Løksebotten was dominated by a few chalimus stages in June (Figure 2b). The chalimus infection had slightly increased in July, and older stages also aggregated on the fish in this month. In August, a new infection of lice larvae had occurred in Løksebotten but also a significant aggregation of older lice on the fish.

A) Altafjord

B) Løksebotten Week 26 N=0 n = 20

,5

CH1

CH3 P1M

P1F

P2M

n = 24

,6 ,4

,2

CH1 CH3 P1M P1F P2M P2F ADM ADF

P2F ADM ADF

Week 29 N = 36 n = 37

,8 ,6 ,4 ,2

Week 29

1,0

Mean number of salmon lice

1,0

Mean number of salmon lice

N = 32 ,8

0,0

0,0

0,0

N = 36 n = 37

,8 ,6 ,4 ,2 0,0

Week 32

10

N = 434 n = 26

5

0 CH1 CH3 P1M P1F P2M

P2F ADM ADF

Developmental stages

CH1 CH3 P1M P1F P2M P2F ADM ADF

Mean number of salmon lice

CH1 CH3 P1M P1F P2M P2F ADM ADF

Mean number of salmon lice

Week 25

1,0

Mean number of salmon lice

Mean number of salmon lice

1,0

Week 32

10

N = 434 n = 26

5

0 CH1 CH3 P1M P1F P2M

P2F ADM ADF

Developmental stages

Figure 2 Distribution of developmental stages (%) of salmon lice on pooled groups of sea trout and Arctic char captured in saltwater in the Altafjord (A) and the Malangen/Løksebotten (B) system. The fish was sampled in June (week 25/26), July (week 29) and August (week 32/33). (N is the total number of lice on the fish and n is the number of fish sampled. Developmental stages are designated as follows: CH1, First and second chalimus stage combined; CH3, Third and fourth chalimus stage combined; P1M, First preadult male; P1F, first preadult female; P2M, second preadult male; P2F, Second preadult female; ADM, Adult male; ADF, Adult female.)

114

SUMBAWS

Final Report

The relative intensity (number of lice · g-1 fish weight) in fish from Løksebotten did also increase significant during the sampling period (Kruskal-Wallis; p < 0.05) (Figure 3). In June and July very low relative intensities were found, and the levels peaked in August at median levels close to 0.2 lice · g-1 fish weight. In the Altafjord system, approximately similar median relative intensities (0.15 lice · g-1 fish weight) were found in July and August, but relatively few individuals were infected early in the season in contrast to the majority of the population in August. Maximum value in Altafjord was 0.7 lice · g-1 fish weight. In Løksebotten, approximately 25 % of all infected post-smolts carried relative intensities between 0.15 and 0.7 lice · g-1 fish weight in August.

Figure 3 Box-and-whiskers plot showing the relative intensity of lice (number of lice · g-1 fish weight) in the smallest fish (< 200 g) in a) Altafjord in July (week 29; n = 2) and August (week 32; n = 17) and b) Løksebotten in June (week 25; n = 1), July (week 29; n = 8) and August (week 33; n = 8). Horizontal lines indicate medians. The lower and upper hinges give the 25th and 75th percentile. Outliers (o) are presented, and the whiskers give the largest and smallest observed values that are not outliers. The horizontal line gives the relative intensity expected to cause minor osmoregulatory disturbances of the fish (Wagner et al., 2003, 2004).

Atlantic salmon No salmon lice were found on pelagic captured post-smolts of Atlantic salmon from the Alta- and Malangsfjord at any of the sampling periods in late June and early July, which contradict the results described from sea trout and Arctic char captured in the littoral zone (Table 2). This includes the few fish captured in the inner zone of Malangsfjord in late June, and in the inner, middle and outer zone of the fjord in July (week 27). The same pattern was found for post-smolts in the Altafjord system captured in the inner zone of Altafjord in June and in the middle in July (no fish were captured in the outer zone in this fjord).

Discussion The present study indicate that the risks of salmon lice infection may differ between sympatric populations of Atlantic salmon post-smolts, sea trout and Arctic char in north Norwegian fjords, which may be explained by generic differences in their marine life-history, including migratory timing, behavioural pattern and duration of fjord residency.

In the Altafjord, both sea trout and Arctic char captured in the littoral zone and pelagic captured Atlantic salmon post-smolts were uninfected with salmon lice in late June. The same pattern was found on Atlantic salmon post-smolts from Malangsfjord in late June, while only a few lice were found on littoral feeding Arctic char and sea trout from Løksebotten during the same period. In July, more than two third of the sea trout and Arctic char in Løksebotten were infected, while less than 10% of the Arctic char and sea trout in Altafjorden were infected with a few salmon lice. No Atlantic salmon post-smolts were infected in any of the fjords during the same period, and the post-smolts of this species had probably left the fjord during the last part of July. However, in the middle of August (week 32/33) the lice infection on sea trout and Arctic char had increased severely, and showed an epidemic tendency through the concurrent rise in prevalence, intensity and variance to mean ratio in both systems. This pattern of infection is to a large extent in consistence with earlier studies in northern areas where the infection pressure have been found to usually increase in August after relatively low levels in June and July (Bjørn et al., 2001a; Bjørn & Finstad, 2002).

115

SUMBAWS

Final Report

There are, however, also variations between years and increased infection pressure has also been observed in early July both in Altafjorden (Bjørn & Finstad, 2002) and at different localities in Troms County (Bjørn et al., 2000).

Although the sampling periods for Atlantic salmon post-smolts were not completely comparable with Arctic char and sea trout due to small differences in sampling time and space, the results indicate that a difference in risks of infection may exist between pelagic migratory Atlantic salmon post-smolts compared to the littoral fjord feeding sea trout and Arctic char in northern fjords. The infection pattern in sea trout and Arctic char showed a remarkable similarity between the localities: very low infections were found in June, while the prevalence increased in July and both prevalence and mean intensity peaked in August at levels significantly higher than assumed historical levels (Boxshall, 1974; Pemberton, 1976) and in areas without fish farms in recent years (Tingley et al., 1997; Schram et al., 1998; Bjørn et al., 2001b; Bjørn & Finstad, 2002; Rikardsen, 2004).

The relationship between the growth of the fish farming industry along the coast, and co-occurring salmon lice attacks on both farmed and wild salmonids has intensively been discussed in recent time (reviewed by Pike & Wadsworth, 1999; Tully & Nolan, 2002; Heuch et al., 2005). Direct evidence of lice transfer from farmed to wild hosts has, however, not been generated. There is, however, substantial evidence indicating that infective copepodids from salmon lice on farmed fish have a role in generating the observed epidemics on wild salmonid fish in farming areas (Tully & Whelan, 1993; Tully et al., 1993a,b; Heuch & Mo, 2001a,b; Bjørn et al., 2001a,b; Bjørn & Finstad, 2002; Gargan et al., 2003; Krkošek et al., 2005 ).

Historical infection levels on sea trout (Boxshall, 1974), and infection levels in areas without fish farming, are generally occurring at a relative high prevalence but low intensities (Tingley et al., 1997; Schram et al., 1998; Bjørn et al., 2001b; Bjørn & Finstad, 2002, Rikardsen, 2004). The variability between years also seems to be low in areas with no fish farm activity (Tingley et al., 1997; Schram et al., 1998), and probably represent a stable long term situation with few adult lice, low transmission rates, and no adverse effect on the fish (Tingley et al., 1997).

Within Atlantic salmon farms, or within farming areas, large numbers of hosts are continuously present, facilitating a build up of large number of reproducing female lice and a continuous re-infection (Heuch & Mo 2001a,b). As the potential for larvae production is substantially increased under marine cage culture conditions (Tully & Whelan, 1993; Heuch & Mo 2001a,b), years with optimal condition for lice reproduction and dispersal, may potentially result in salmon lice epidemics in wild salmonids (Bjørn et al., 2001a, Stien et al., 2005). At least in sea trout populations, these epidemics are characterised by high infection pressure leading to physiological damaging, or even lethal lice infections levels, premature return to freshwater of the most heavily infected fish, as well as indices of direct parasite induced mortality of heavily infected fish (Bjørn et al., 2001b). The risks of infection from free swimming salmon lice copepodids derived from cultured fish to wild salmonids will depend on e.g. the number and dispersal of lice from fish farms, the behaviour, survival and longevity of infective copepodids (Stien et al., 2005), and the feeding or migratory areas of wild salmonids in relation to farms (Thorstad et al., 2004; Finstad et al., 2005).

The risks of salmon lice infection may, therefore, also differ between the salmonid species as indicated in the results of the present study, and may also differ from the infection risk in more southern areas. Results from Bjørn & Finstad (1998) imply, although not directly tested, that salmon lice have a quite similar development and growth rate on the congeneric sea trout, Arctic char and Atlantic salmon compared with the Pacific salmon species Oncorhynchus spp. (Johnson & Albright, 1992; Johnson, 1993). Similar results have also been indicated from field studies where both sea trout and Arctic char have become heavily infected with salmon lice when feeding in fjords and coastal areas of intensive fish farming activity (Bjørn et al., 2001b; Bjørn & Finstad, 2002). Fjord migrating post-smolts of Atlantic salmon has been found with low salmon lice infection in fjord almost without fish farming activity (Finstad et al., 2000), but dramatically high infection rates have also been found in posts-smolts descending intensively farmed areas of western Norway (Holst & Jakobsen, 1999; Heuch et al., 2005).

However, the marine migratory behaviour of Atlantic salmon, Arctic char and sea trout diverge on several important aspects, although the knowledge of the details of this is still limited. Most of the information gathered to date suggest that post-smolts of Atlantic salmon move relatively quickly through estuary and the fjord close to the surface (Hvidsten et al., 1992; Moore et al., 1995; Holm et al., 1998; Voegeli et al., 1998; Moore et al., 2000; Thorstad et al., 2004; Finstad et al., 2005), although this also may vary between populations and years (Hutchinson & Mills, 2000; Rikardsen et al., 2004a; Knudsen et al., 2005). In contrast, sea trout and Arctic char usually feed in littoral areas close to their native river throughout the summer and autumn (Berg & Jonsson, 1990; Finstad & Heggberget, 1993; Lyse et al., 1998; 116

SUMBAWS

Final Report

Rikardsen et al., 2000; Rikardsen & Amundsen, 2005). Tagging experiments by use of data logger tags on trout and char have shown that the fish spend more than 90% of the time at the surface no deeper than 3 meter in the Altafjord (Rikardsen et al., in press). Sea trout and Arctic char therefore seem to belong to a “near shore, surface orientated quild of fishes” as previously suggested by Grønvik and Klemetsen (1987), although these fish may also display pelagic feeding in open water within fjords at some occasions (Rikardsen & Amundsen, 2005). Fish farms are usually located quite close to the littoral zone. Infective dispersal stages of salmon lice may therefore be more concentrated in nearshore areas and fjords and lochs with a turbulent current pattern and often distinct thermoclines and haloclines, which seem to be preferred areas both for sea trout and Arctic char (Lyse et al., 1998; Bjørn et al., 2001b; Rikardsen et al., 2000; Rikardsen, 2004) as well as for salmon lice copepodids (Heuch et al., 1995; McKibben & Hay, 2004; Penston et al., 2004).

There are also differences in the timing of the migration between the species; often large veterans of Arctic char descend prior to veteran sea trout, followed by smolts of Atlantic salmon, Arctic char and sea trout (Tuff Carlsen et al., 2004). Similarly, there are also differences in the sea water residency of Arctic char and sea trout within fjords. Arctic char usually ascend earlier than sea trout in late summer or autumn, where large veterans predominant among the early ascenders while post-smolts return later (Berg & Jonsson, 1990; Rikardsen et al., 1997). Furthermore, some sea trout are also shown to stay in saltwater during late autumn and winter within some northern fjords (Rikardsen, 2004). In addition to possible immunological differences between the salmonid species (Dawson et al., 1997), these behavioural differences may lead to different risks of infection. Furthermore, the sea water temperature is usually also much lower in the north than in the southern fjords of Norway (Rikardsen et al., 2004a). As a result of this, the peak in infection pressure often occur on trout and char in August - October at northern latitudes (Bjørn & Finstad, 2002; Rikardsen, 2004; this study) compared to June - August at more southern latitudes (Schram et al., 1998; Heuch et al., 2005). Usually, the Atlantic salmon post-smolts have left the northern fjord (Rikardsen et al., 2004a) when the infection pressure is expected to be highest. Atlantic salmon post-smolts may therefore usually experience “mismatch” conditions in northern fjords between the peak in infection risk and the normal migration period. This may also be the case for veteran migrants of sea trout, and especially Arctic char veterans who often ascend the rivers in July or early August (Berg & Jonsson, 1990; Rikardsen et al., 1997). Immature Arctic char and sea trout often stay longer in seawater (Berg & Jonsson, 1990), and some sea trout may even spend the whole autumn and winter at sea (Rikardsen, 2004). Especially immature sea trout may therefore often get rather high salmon lice infections, and results from this study, as well as previous results from the Altafjord system (Bjørn et al., 2001a; Bjørn & Finstad, 2002; Bjørn et al., 2002), indicate that relative infection intensities reaches levels that may impair fish physiology (Bjørn & Finstad, 1997; Nolan et al., 1999; Bjørn et al., 2001b; Wagner et al., 2003, 2004; Wells et al., submitted).

The August infection levels found on the most heavily infected small sea trout from both the Altafjord and Løksebotten, may therefore have consequences for the population on the longer terms (Wagner et al. 2003).

In contrast, Arctic char may in general spend to short time at sea (average 30-50 days, Finstad & Heggberget, 1993, 1995; Rikardsen, 2000) for the salmon lice to develop to the more harmful mobile preadult and adult stages, although severe and harmful infections levels has also been observed on this species as well at some years (Bjørn et al., pers.obs.). However, knowledge of the general early marine ecology of salmonids is still limited. For example, recent results from the Altafjord indicate prolonged feeding periods of Atlantic salmon post-smolts on energy rich fish larva within the fjord system (Rikardsen et al., 2004a; Knudsen et al., 2005). If that is the case, risks of salmon lice infection may be severely increased also in northern areas. This and other aspects with the marine migration behaviour of the species should be addressed more carefully in order to reduce the risk of salmon lice infection from the increasing fish farming industry in at the northern latitudes.

Acknowledgements Financial support for this study was given by the Norwegian Research Council (The Wild Atlantic Salmon Program no. 149796/720), the Norwegian Directorate for Nature Management, and the European Commission Contract no: Q5RS-2002-00730.

117

SUMBAWS

Final Report

References Bakke, T.A. & Harris, P.D. (1998). Diseases and parasites in wild Atlantic salmon (Salmo salar) populations. Canadian Journal of Fisheries and Aquatic Sciences 55, 247-266.

Berg, O.K. & Jonsson, B. (1990). Growth and survival rates of the anadromus trout, Salmo trutta, from the Vardnes River northern Norway. Environmental Biology of Fishes 29, 145-154.

Birkeland, K. (1996). Salmon lice, Lepeophtheirus salmonis Krøyer, infestations and implications for anadromous brown trout, Salmo trutta L. Dr. scient thesis, University of Bergen.

Bjørn P. A. & Finstad B. (1997). The physiological effects of salmon lice infection on sea trout post smolts. Nordic Journal of Freshwater Research 73, 60-72.

Bjørn P.A. & Finstad B. (1998). The development of salmon lice (Lepeophtheirus salmonis) on artificially infected sea trout (Salmo trutta) post smolts. Canadian Journal of Zoology 76, 970-977.

Bjørn, P.A. & Finstad, B. (2002). Salmon lice, Lepeophtheirus salmonis (Krøyer), infestation in sympatric populations of Arctic char, Salvelinus alpinus (L.), and sea trout, Salmo trutta (L.), in areas near and distant from salmon farms. ICES Journal of Marine Sciences 59, 131-139.

Bjørn P.A., Kristoffersen R. & Finstad B. (2000). Lakselus på vill sjøørret og sjørøye i Troms sommeren 1999 (Salmon lice on wild sea trout and Arctic charr in Troms – summer 1999). – Rapport til Fiskehelse og Miljøgruppa i Troms, Fiskeridirektoratet, region Troms, 1-34 (in Norwegian).

Bjørn P. A., Kristoffersen R. & Finstad B. (2001a). Registreringer av lakselus på laks, sjøørret og sjørøye i 2000. (Registrations of salmon lice on Atlantic salmon, sea trout and Arctic charr in 2000). NINA Oppdragsmelding 698. Report in Norwegian with English summary. Norwegian Institute for Nature Research, Trondheim, Norway. 40 pp.

Bjørn, P. A., Finstad, B. & Kristoffersen, R. (2001b). Salmon lice infections of wild sea trout and Arctic char in marine and freshwaters: the effects of salmon farms. Aquaculture Research 31. 795-803.

Bjørn, P. A., Finstad, B. & Kristoffersen R. (2002). Registreringer av lakselus på laks, sjøørret og sjørøye 1999. (Registrations of salmon lice on Atlantic salmon, sea trout and Arctic charr in 2001). NINA Oppdragsmelding 737. Report in Norwegian with English summary. Norwegian Institute for Nature Research, Trondheim, Norway. 33 pp.

Boeuf, G. (1993). Salmonid smolting: a preadaptation to the oceanic environment. In Fish Ecophysiology.

(Rankin, J.C. & Jensen, F.B., eds.), pp. 105-135. London: Chapman & Hall.

Boxshall, G. A. (1974). Infection with parasitic copepods in north sea marine fishes. Journal of Marine Biology Association U.K. 54, 355-372.

Bush, A. O., Lafferty, K. D., Lotz, J. M. & Shostak, A. W. (1997). Parasitology meets ecology on its own terms: Margolis et al. revised. Journal of Parasitology 83, 575-583.

Butler, J. R. A. (2002). Wild salmonids and sea louse infestations on the west coast of Scotland: sources of infection and implications for the management of marine salmon farms. Pest Management Science 58, 595608.

Dawson, L.H.J., Pike, A.W., Houlihan, D.F. & McVicar, A.H. (1997). Comparison of the susceptibility of sea trout (Salmo trutta L.) and Atlantic salmon (Salmo salar L.) to sea lice (Lepeophtheirus salmonis (Krøyer, 1837)) infections. ICES Journal of Marine Science 54, 1129-1139.

Finstad, B. & Heggberget, T. G. (1993). Migration, growth and survival of wild and hatchery-reared anadromous Arctic charr (Salvelinus alpinus) in Finnmark, northern Norway. Journal of Fish Biology 43, 303-312.

Finstad, B. & Heggberget, T. G. (1995). Seawater tolerance, migration, growth and recapture rates of wild and hatchery-reared Arctic charr (Salvelinus alpinus (L.). Nordic Journal of Freshwater Research 71, 229236.

Finstad, B., Grimnes, A., Bjørn, P.A. & Hvidsten, N.A. (2000). Laboratory and field investigations of salmon lice [Lepeophtheirus salmonis (Krøyer)] infestation on Atlantic salmon (Salmo salar L.) postsmolts.

Aquaculture Research 31, 795-803.

118

SUMBAWS

Final Report

Finstad, B., Økland, F., Thorstad, E.B., Bjørn, P.A. & McKinley, R.S. (2005). Migration of hatcheryreared Atlantic salmon and wild sea trout post-smolts in a Norwegian fjord system. Journal of Fish Biology, in press.

Gargan, P.G., Tully, O., and Poole, W.R. (2003) Relationship between sea lice infestation, sea lice production and sea trout survival in Ireland, 1992-2001. In Salmon at the edge. (Mills, D., ed.), pp. 119-135.

Oxford,: Blackwell Science.

Grønvik, S. & Klemetsen, A. (1987). Marine food and diet overlap of co-occurring Arctic charr Salvelinus alpinus (L.), brown trout Salmo trutta L. and Atlantic salmon S. salar L. off Senja, N. Norway. Polar Biology 7, 73-177.

Heuch, P.A. (1995). Experimental evidence for aggregation of salmon louse copepodids, Lepeophtheirus salmonis, in step salinity gradients. Journal of the Marine Biology Association of the United Kingdom 75, 927–939.

Heuch, P. A. & Mo, T. A. (2001a). A model of salmon louse production in Norway: Effects of increasing salmon production and public management measures. Diseases of Aquatic Organisms 45, 145–152.

Heuch, P. A. & Mo, T. A. (2001b). Sluttrapport for NFR-prosjektet 114.109/122 Spredning av lakselus fra oppdrettsanlegg. (Dispersion of salmon lice from seafarms). Report in Norwegian. National Veterinary Institute, Oslo, Norway.

Heuch, P. A. Parsons, A. & Boxaspen, K. (1995). Diel vertical migration: a possible host-finding mechanism in salmon louse (Lepeophtheirus salmonis) copepodids? Canadian Journal of Fisheries and Aquatic Sciences 52, 681–689.

Heuch, P. A., Knutsen, J. A., Knutsen, H. & Schram. T. (2002). Salinity and temperature effects on sea lice over-wintering on sea trout (Salmo trutta) in coastal areas of the Skagerrak. Journal of the Marine Biology Association of the United Kingdom 82, 887-892.

Heuch, P. A., Bjørn, P. A., Finstad, B., Holst, J. C. H., Asplin, L. & Nilsen, F. (2005). A review of the Norwegian “National Action Plan Against Salmon Lice on Salmonids”: The effects on wild salmonids.

Aquaculture 246, 79-92.

Hoar, W.S. (1988). The physiology of smolting salmonids. In Fish Physiology (Hoar, W.S. & Randall, D.J., eds.), Vol. 11B, pp. 275-343. New York: Academic Press.

Holm, M., Axelsen, B.E., Sturlaugsson, J., Hvidsten, N.A., Ikonen, E. & Johnsen, B.O. (1998).

Behaviour of acoustically tagged post-smolts in the Trondheim fjord – Influence of hydrografical and metrological conditions on migration. International Council for the Exploration of the sea CM 1998/N:17.

Holst, J.C. & Jakobsen, P.J. (1999). Lakselus knekker vestlandslaksen (Salmon lice kills the salmon in Western Norway). Norsk Fiskeoppdrett 16: 38–39 (in Norwegian).

Holst, J. C. & McDonald, A. (2000). FISH-LIFT: a device for sampling live fish with trawls. Fisheries Research 48, 87-91.

Hutchinson, P. & Mills, D. (2000). Executive summary. In The Ocean Life of Atlantic salmon.

Environmental and Biological factors Influencing Survival (Mills, D., ed.), pp. 7-18. Oxford: Blackwell Science.

Hvidsten, N.A., Johnsen, B.O. & Levings, C.D. (1992). Atferd og ernæring hos utvandrende laksesmolt i Trondheimsfjorden (Behaviour and feeding in postsmolts from the Trondheimsfjord). NINA Oppdragsmelding 164. Report. In Norwegian with English summary. Norwegian Institute for Nature Research, Tungasletta 2, N-7485 Trondheim, Norway, 1-14 (in Norwegian).

Høgåsen, H. R. (1998). Physiological changes associated with the diadromous migration of salmonids.

Canadian Special Publications for Fisheries and Aquatic Science 127, 1-128.

Johannessen, A. (1975). Lakselus, Lepeophtheirus salmonis Krøyer (Copepoda, Caligidae). Frittlevende stadier, vekst og infeksjon på laks (Salmon salar L.) fra oppdrettsanlegg og kommersielle fangster i vestnorske farvann 1973-1974. Cand. Real. thesis, University of Bergen, University Library Bergen, Norway (in Norwegian).

Johnson S. C. (1993). A comparison of development and growth rates of Lepeophtheirus salmonis (Copepoda: Caligidae) on naïve Atlantic (Salmo salar) and Chinook (Oncorhynchus tshawytscha) salmon.

119

SUMBAWS

Final Report

In: Pathogens of wild and farmed fish, sea lice pp. 68-80. Ed. By G. A. Boxhall and Defaye. Ellis Horwood.

London.

Johnson S.C. & Albright, L.J. (1992). Comparative susceptibility and histopathology of the response of naive Atlantic, chinook and coho salmon to experimental infection with Lepeophtheirus salmonis (Copepoda: Caligidae). Diseases of Aquatic Organisms 14, 179-193.

Jonsson, B. (1985). Life history patterns of freshwater resident and sea-run migrant brown trout in Norway.

Transactions of the American Fisheries Society 114, 182-194.

Kabata, Z. (1974). Moth and mode of feeding of Caligidae (Copepoda), parasites of fishes, as determined by light and scanning electron microscopy. Journal of the Fisheries Research Board of Canada 31, 15831588.

Klemetsen, A., Amundsen, P.-A., Dempson, B., Jonsson, B., Jonsson, N., O’Connell, M.F. & Mortensen, E. (2003). Atlantic salmon Salmo salar L., brown trout Salmo trutta L. and Arctic charr Salvelinus alpinus (L.): a review of aspects of their life histories. Ecology of Freshwater Fish 12, 1-59.

Knudsen, R, Rikardsen, A.H., Dempson, J.B., Bjørn, P.A., Finstad, B., Amundsen, P.-A. & Holm, M. (2005). Intestinal trophic transmitted parasites as bio-indicators of prolonged fjord-feeding of Atlantic salmon (Salmo salar) post-smolts in Northern Norway. Journal of Fish Biology 66, 1-15.

Krkošek, M., Lewis, M. A. & Volpe, J. P. (2005). Transmission dynamics of parasitic sea lice from farm to wild salmon. Proceeding of the Royal Society B 272, 689-696.

Lyse, A. A., Stefansson, S. O. & Fernö, A. (1998). Behaviour and diet of sea trout post-smolt in a Norwegian fjord system. Journal of Fish Biology 52, 923-936.

McKibben, M. A. & Hay, D. (2004). Distribution of planktonic sea lice larvae Lepeophtheirus salmonis in the inter-tidal zone in Loch Torridon, Western Scotland in relation to salmon farm production cycles.

Aquaculture Research 35, 742-750.

Mo ,T.A. & Heuch, P.A. (1998). Occurrence of Lepeophtheirus salmonis (Copepoda: Caligidae) on sea trout (Salmo trutta) in the inner Oslo Fjord, south-eastern Norway. ICES Journal of Marine Science 55, 176180.

Moore, A., Potter, E.C.E., Milner, N.J. & Bamber, S. (1995). The migratory behaviour of wild Atlantic salmon smolts in the estuary of the River Conwy, North Wales. Canadian Journal of Fisheries and Aquatic Sciences 52, 1923-1935.

Moore, A., Lacroix, G.L. & Sturlaugsson, J. (2000). Tracking Atlantic salmon post-smolts in the sea. In The ocean life of Atlantic salmon - environmental and biological factors influencing survival (Mills, D. Ed.). pp. 49-64. Oxford: Fishing News Books.

Nolan, D.T., Reilly, P., and Wendelaar Bonga, S.E. (1999). Infection with low numbers of the sea louse Lepeophtheirus salmonis induces stress-related effects in postsmolt Atlantic salmon (Salmo salar).

Canadian Journal of Fisheries and Aquatic Sciences 56, 947–959.

Pemberton, R. (1976). Sea trout in the North Argyll sea lochs: population, distribution, and movements.

Journal of Fish Biology 9, 157-179.

Penston, M. J., McKibben, M. A., Hay, D. W. & Gillibrand, P. A. (2004). Observation on open-water densities of sea lice larvae in Loch Shieldaig, Western Scotland. Aquaculture Research 35, 793-805.

Pike A.W. & Wadsworth S.L. (1999). Sealice on Salmonids: Their Biology and Control. Advances in Parasitology 44, 233–337.

Rikardsen, A.H. (2000). Effects of Floy and soft VIalpha tags on growth and survival of juvenile Arctic char. North American Journal of Fisheries Management 20, 719-728.

Rikardsen, A.H. (2004). Seasonal occurrence of salmon lice Lepeophtheirus salmonis on sea trout Salmo trutta in North Norwegian fjords without fish farms. Journal of Fish Biology 65, 711-722.

Rikardsen, A. H. & Amundsen, P.-A. (2005). Pelagic marine feeding behaviour of Arctic charr Salvelinus alpinus and sea trout Salmo trutta. Journal of Fish Biology 66, 1163-1166.

Rikardsen, A.H., Svenning, M.-A. & Klemetsen, A. (1997). The relationships between anadromy, sex ratio and parr growth of Arctic charr in a lake in North Norway. Journal of Fish Biology 51, 447-461.

120

SUMBAWS

Final Report

Rikardsen, A.H., Amundsen, P.-A., Bjørn, P.A. & Johansen, M. (2000). Comparison of growth, diet and food consumption of sea-run and lake-dwelling Arctic charr. Journal of Fish Biology 57, 1172-1188.

Rikardsen, A.H., Haugland, M., Bjørn, P.A. Finstad, B., Knudsen, R., Dempson, B., Holm, M., Holst, J.C. & Hvidsten, N.A. (2004a). Geographical differences in early marine feeding of Atlantic salmon postsmolt in Norwegian fjords. Journal of Fish Biology 64, 1655-1679.

Rikardsen, A.H., Thorpe, J.E. & Dempson, B. (2004b). Modelling the life-history variation of anadromous and resident Arctic charr Salvelinus alpinus (L.). Ecology of Freshwater Fish 13, 305-311.

Schram T.A., Knutsen J.A., Heuch P.A. & Mo T.A. (1998). Seasonal occurrence of Lepeophtheirus salmonis and Caligus elongatus (Copepoda: Caligidae) on sea trout (Salmo trutta), off southern Norway.

ICES Journal of Marine Science 55, 163-175.

Stien, A., Bjørn, P. A., Heuch, P. A. & Elston, D. A. (2005). Population dynamics of salmon lice Lepeeophtheirus salmonis on Atlantic salmon and sea trout. Marine Ecology Progress Series 290, 263-275.

Thorstad, E.B., Økland, F., Finstad, B., Sivertsgård, R., Bjørn, P.A. & McKinley, R.S. (2004) Migration speeds and orientation of Atlantic salmon and sea trout post-smolts in a Norwegian fjord system.

Environmental Biology of Fishes 71, 305-311.

Tingley, G. A., Ives, M. J. & Russel, I. C. (1997). The occurrence of lice on sea trout (Salmo trutta L.) captured in the sea off the East Anglian coast of England. ICES Journal of Marine Science 54, 1120-1128.

Todd, C. D., Walker, A. M., Wolff, K., Northcott, S. J., Walker, A. F., Ritchie, M. G., Hoskins, R., Abbott, R. J. & Hazon, N. (1997). Genetic differentiation of populations of the copepod sea louse Lepeophtheirus salmonis (Krøyer) ectoparasitic on wild and farmed salmonids around the coasts of Scotland: Evidence from RAPD markers. Journal of Experimental Marine Biology and Ecology 210, 251274.

Tuff Carlsen, K., Berg, O.K., Finstad, B. & Heggberget, T.G. (2004). Diel periodicity and environmental influence on the smolt migration of Arctic charr, Salvelinus alpinus, Atlantic salmon, Salmo salar , and brown trout, Salmo trutta, in northern Norway. Environmental Biology of Fishes 70(4), 403-413.

Tully, O. & Whelan, K. F. (1993). Production of nauplii of Lepeophtheirus salmonis (Krøyer) (Copepoda: Caligidae) from farmed and wild salmon and its relation to the infestation of wild sea trout (Salmo trutta L.) off the west coast of Ireland in 1991. Fisheries Research 17, 187-200.

Tully, O., Poole, W.R. & Whelan, K.R. (1993a). Infestation parameters for Lepeophtheirus salmonis (Krøyer) (Copepoda: Caligidae) parasitic on sea trout, Salmo trutta L., off the west coast of Ireland during 1990 and 1991. Aquaculture and Fisheries Management 24, 545-555.

Tully, O., Poole, W.R., Whelan, K.R. & Merigoux, S. (1993b). Parameters and possible causes of epizootics of Lepeophtheirus salmonis (Krøyer) infesting sea trout (Salmo trutta L.) off the west coast of Irland. In Pathogens of wild and farmed fish: sea lice. Ed. by (G.A. Boxshall, G.A. & Defaye, D., eds), pp 202-213. Ellis Horwood. London.

Tully, O. & Nolan, D. T. (2002). A review of the population biology and host-parasite interactions of the sea louse Lepeophtheirus salmonis (Copepoda: Caligidae). Parasitology 124, 165-182.

Voegeli, F.A., Lacroix, G.L. & Anderson, J.M. (1998). Development of miniature pingers for tracking Atlantic salmon smolts at sea. Hydrobiologia 371/372, 35-46.

Wagner, G. N., McKinley, R.S., Bjørn, P.A. & Finstad, B. (2003). Physiological impact of sea lice on the swimming performance of Atlantic salmon. Journal of Fish Biology 62, 1000-1009.

Wagner, G., McKinley, R.S., Bjørn, P.A., & Finstad, B. (2004). Short-term freshwater exposure benefits sea lice-infected Atlantic salmon. Journal of Fish Biology 64, 1593-1604.

Wells, A., Grierson, C.E., MacKenzie, M., Russon, I.J., Reinardy, H., Middlemiss, C., Bjørn, P.A., Finstad, B., Wendelaar Bonga, S.E., Todd, C.D. & Hazon, N. (2006) The physiological effects of simultaneous abrupt seawater entry and sea lice (Lepeoptheirus salmonis Krøyer) infestation of wild sea trout (Salmo trutta L.) smolts. Canadian Journal of Fisheries and Aquatic Sciences. submitted Wootten, R., Smith, J. W. & Needham, E. A. (1982). Aspects of the biology of the parasitic copepods Lepeophtheirus salmonis and Caligus elongates on farmed salmonids, and their treatment. Proceedings of the Royal Society of Edinburgh 81B, 185-197.

121

SUMBAWS

Final Report

Paper 4: RISKS OF SALMON LICE, LEPEOPHTHEIRUS SALMONIS POPULATIONS OF ATLANTIC SALMON, SEA TROUT AND HYPOTHESIS

(KRØYER) INFECTION ON SYMPATRIC ARCTIC CHAR: THE MATCH-MISMATCH

P. A. Bjørn, R. Sivertsgård, B. Finstad and R. Kristoffersen (Manuscript) Abstract Differences in salmon lice (Lepeophtheirus salmonis) infection on sympatric populations of fjord migrating Atlantic salmon post-smolts (Salmo salar), sea trout (Salmo trutta) (sea trout) and Arctic char (Salvelinus alpinus) were studied in the intensively fish-farmed Altafjord system in northern Norway.

Atlantic salmon post-smolts were captured only in the pelagic areas of the fjord during late June and early July. After this time they probably left the fjord. Most of these fish were uninfected with salmon lice, but smolts captured in outer fjord area in 2004 carried a small number of lice. This is the first time lice have been found on wild Atlantic salmon post-smolts in northern Norway. In contrast, pelagic feeding sea trout and Arctic char had surprisingly high infection levels during June and July, and this suggests that their pelagic behavioural pattern exposes them to high risks of salmon lice infection. Very low prevalence and mean infection intensity during June, and much higher levels in summer and autumn, were in contrast found in littoral feeding sea trout and Arctic char. These results agree with earlier observations that Atlantic salmon post-smolts may avoid sea lice infestation due to a mismatch between the peak of lice infection risk and their post-smolt fjord migration in northern fjords. In contrast, sea trout and Arctic char feed within the fjords throughout the summer and consequentially have higher risk of harmful infections with sea lice.

However, risks of infection may also differ between pelagic feeding sea trout and Arctic char, and littoral feeding fish, as well as between species and age classes. Integrated salmon lice management models must therefore be based on local ecological and environmental knowledge to ensure that the appropriate measures are taken at a local level.

Introduction During the last decade, the salmon farming industry has been introduced into pristine marine areas of the Northern Hemisphere. In these areas, some of the most important rivers for Atlantic salmon (Salmo salar L.) are situated, e.g. the Tana river in the Tanafjord system and the Alta river in the Altafjord system. These areas are also vital for wild stocks of anadromous sea trout (Salmo trutta L.) (sea trout) and anadromous Arctic char (Salvelinus alpinus L.) (Jonsson, 1985; Finstad & Heggberget, 1993, 1995; Rikardsen et al., 2000; 2004b; Klemetsen et al., 2003).

In the Altafjord system and close surroundings, approximately 8.2 million farmed Atlantic salmon were present in 2003 and 2004 (Heuch et al. 2005), and some of these farms are situated only a few kilometres from the mouth of the River Alta. The wild Atlantic salmon in the Alta river are one of the world’s most important stocks of Atlantic salmon, and the Alta river is famous for its large 3 SW fish. The river also contains important populations of sea trout and Arctic char (Thorstad et al. 2004). In addition, several other smaller river systems in the fjord contain populations of Atlantic salmon, sea trout and Arctic char (Finstad and Heggberget, 1993, 1995).

In northern Norway, smolts of anadromous fish migrate to sea for the first time during a 2 – 3 week peak between late May and early July (Rikardsen et al., 1997; 2004a; Klemetsen et al., 2003; Tuff Carlsen et al.

2004). Atlantic salmon spend 1 – 3 years at sea before returning to spawn in freshwater (Klemetsen et al., 2003), and are assumed to move quickly throughout the fjord (Thorstad et al., 2004; Finstad et al., 2005), but new results also indicate that Atlantic salmon post-smolts can feed for prolonged periods especially in the productive fjords of northern Norway (Rikardsen et al. 2004a, Knudsen et al. 2005) before they migrate to the feeding areas in the open sea (Hansen et al. 2003). At the same latitude, Arctic char and sea trout normally spend 1-2 summer months each year at sea before returning to freshwater, and trout usually return several weeks later than the char (Jonsson, 1985; Rikardsen, 2000). Therefore, the life-history pattern of salmonid species in the Northern Hemisphere depends to a large degree on the productivity of the marine fjord environment (Gross 1987). There is general concern that increased marine salmonid aquaculture and wild salmonids will not interact in a sustainable way (Bakke and Harris 1998, Gross 1998, Johnson et al.

122

SUMBAWS

Final Report

2004). Special concern has been given to the exposure of wild salmonids to salmon lice epidemics in areas with intensive fish farming aquaculture.

Since the late 1980s, there have been heavy infestations of salmon lice on anadromous sea trout (sea trout) along the coast of Norway (Birkeland, 1996; Bjørn et al., 2001), Ireland (Tully et al., 1993a,b; Gargan et al., 2003) and Scotland (Todd et al., 1997; Butler, 2002). Similar results have also recently been reported in Atlantic salmon post-smolts migrating through intensively farmed fjords in western Norway (Holst et al.

2003), and Pacific salmon smolts migrating through intensively farmed Canadian fjords (Krkošek et al.

2005). It has been suggested that the increased infestation rate of salmon lice on sea trout and salmon is a result of high lice levels on farmed salmonids in these areas (Tully & Whelan, 1993; Birkeland, 1996; Bjørn et al., 2001; Bjørn & Finstad, 2002; Butler, 2002; Tully & Nolan, 2002; Heuch et al. 2005).

The risks and consequences of salmon lice infection may, however, vary between the species (Bjørn & Finstad, 2002). This will depend both on the encounter rate, the susceptibility to infection (Dawson et al.

1997), and the different life-history of the species (Klemetsen et al., 2003). Due to the apparently quite equal susceptibility to infection in the congeneric Atlantic salmon, sea trout and Arctic char (Bjørn and Finstad 1998, Bjørn et al. 2001, Bjørn and Finstad 2002), different migrating behaviours at sea are believed to have strong implications on the specific risks of salmon lice infection (Bjørn et al. in prep).

Due to severe methodological difficulties in capturing Atlantic salmon post-smolts at sea (Holst & MacDonald, 2000; Rikardsen et al., 2004a), only one previous international refereed paper reports salmon lice infection levels in fjord-migrating post-smolts (Finstad et al., 2000). Observations in intensively farmed western fjords of Norway indicate that infection and its consequences can be considerable (Holst et al. 2003; Heuch et al. 2005). In contrast there is considerably more infection data on Arctic char and sea trout (Boxshall, 1974; Tingley et al., 1997; Mo & Heuch, 1998; Bjørn et al., 2001; Bjørn & Finstad, 2002; Heuch et al., 2002; Rikardsen, 2004). A new study (Bjørn et al. in prep) of differences in risks and consequences of salmon lice infection on sympatric populations of Atlantic salmon, sea trout and Arctic char within northern fjords, indicates that there is a general mismatch between the timing of Atlantic salmon fjord migration and the risk of salmon lice infection. On, the other hand, sea trout and Arctic char seem to be at a higher risk of infestation in northern fjords by remaining in inshore waters (Bjørn et al. in prep).

The purpose of this study was therefore to extend the work of Bjørn et al. (in prep), and further investigate the risk of salmon lice infection in sympatric populations of fjord migrating Atlantic salmon, sea trout and Arctic char in an area with fish farming activity in northern Norway. A comprehensive two year sampling program of wild salmonid post-smolts was therefore initiated in the intensively farmed Altafjord system.

Materials and Methods Study area Altafjord in Finnmark County was chosen as study area (Figure 1). The Altafjord system (70°05’N, 22°55’E) is a broad and open sill fjord with several smaller side branches, and has a maximum width of 14 km and is 30 km long. The outer part of the fjord branches off into three 20 – 50 km long inlets; Stjernsund, Rognsund and Vargsund. The large River Alta is an important freshwater source in the head of the fjord. Altafjord had extensive fish farming activity in 2003 and 2004. Several fish farms operated in the vicinity of the sampling localities, and in total 8.2 million farmed salmon were present in the Altafjord system and surroundings in 2003 and 2004. Some of these farms are situated in the inner part of the fjord system, close to the sampling areas. Large populations of Atlantic salmon, sea trout and Arctic char are present in the Altafjord system and a sampling program for all three species was established in 2003 and 2004. See Rikardsen et al. (2004a) for further information of the Altafjord system.

Sampling procedures and analyses Post-smolts of Atlantic salmon were captured by a newly developed pelagic trawl, the FISH-lift (Holst & MacDonald, 2000). The Arctic char and sea trout where both captured by gill-netting in the littoral zone as described in Bjørn et al. (2001), and by the FISH-lift trawl in pelagic areas of the fjord. The pelagic trawl has proven to be very efficient in capturing fjord-migrating post-smolts of Atlantic salmon, and captures also veteran migrants of sea trout and Arctic char. Fish were sampled in four periods during 2003 and 2004: weeks 25, 26, 28 and 30 in 2003 and weeks 25, 27, 29 and 30 in 2004. Several trawl hauls were performed in pelagic areas of the fjords and along the assumed routes of migrating post-smolts (Figure 1) in each sampling week.

123

SUMBAWS

Final Report

Talte anlegg 2003

Figure 1 The geographic location of the study in the Altafjord system in Northern Norway in 2003 and 2004. Grey zones show the pelagic trawling areas along the assumed migratory routs of Atlantic salmon post-smolts, and circles show the littoral gillnet-fishing locations. Crosses in the map to the bottom right, show the number and location of fish farms in the Altafjord system and surroundings.

Floating gill nets were also set 45 – 90° to the shore at both day and night to capture sea trout and Arctic char in their littoral feeding areas in weeks 26, 29 and 32 in 2003, and weeks 27, 30, 33 and 36 in 2004 (Figure 1). The nets were 25 m long, 2 m deep, and with mesh sizes ranging from 19.5 to 35 mm (Bjørn et al. 2001; Rikardsen, 2004).

All fish were gently removed from the trawl and the gill-nets, and the fish were immediately placed into individually tagged plastic bags. Ecological terms recommended by Bush et al. (1997) were used: prevalence is defined as the percentage of infested fish, abundance is the mean number of parasites per fish caught, median intensity is the median number of parasites per infested fish of a sample, and mean intensity is the average number of parasites per infested fish. Owning to lack of normal distribution and the highly aggregated distributions among the samples, non-parametric Mann-Whitney or Kruskal-Wallis tests were chosen for analyses of statistical differences. Furthermore, the variance (S2) to mean (abundance) ratio was also used to describe the degree of aggregation in the different samples. In all tests, a probability level of p ≤ 0.05 was considered significant.

Fish Sampling In Altafjord, a total of 67 sea trout, 9 Arctic char and 210 Atlantic salmon post-smolts were captured in the pelagic areas of the fjord in 2003. In the littoral areas of Altafjord, 31 sea trout and 26 Arctic char were captured in 2003. In 2004, a total of 220 Atlantic salmon post-smolts, 29 sea trout and 2 Arctic char were captured by trawling in pelagic areas of the fjord. In littoral areas, 106 sea trout and 29 Arctic char were captured. Sea trout in particular, but also a few Arctic char, were captured both in the littoral and pelagic 124

SUMBAWS

Final Report

zone, while all Atlantic salmon post-smolts were captured in the pelagic zone of the fjord. The largest sea trout and Arctic char were generally captured in June and July in the Altafjord system. Most of these fish were maturing fish, whereas most of the other sea trout and Arctic char captured during the later (August and September) sampling periods were immature fish on their first (post-smolt) or second sea migration. All the Atlantic salmon post-smolts were immature. In general, Arctic char were captured most frequently in June and July, while most sea trout were captured in July, August and September. The Atlantic salmon were captured during late June and early July in the inner, middle and outer zone of the Altafjord system. Most of the fish sampled in 2003 and 2004 were examined for lice under a stereoscope according to Bjørn & Finstad (1998). The average weight and number of examined fish from the current months are presented in Table 1.

Table 1 Fish sampled in the Altafjord (Atlantic salmon, sea trout and Arctic char). The mean weight ± standard deviation and the number (n) in each sampling month are presented.

Month

Species

Sea trout June Arctic char Atlantic salmon Sea trout July Arctic char Atlantic salmon Sea trout August Arctic char Atlantic salmon Sea trout September Arctic char Atlantic salmon

Littoral zone 2003 2004 295 ± 166 (12) 292 ± 46 (12) 238 ± 83 (12) 227± 127 (14) 822 ± 393 (7)

694 ± 422 (10) 365 ± 220 (13) 385 ± 461 (49) 130 ± 74 (14)

Pelagic zone 2003 419 ± 250 (33) 409 ± 226 (4) 24 ± 5 (12) 211 ± 147 (34) 335 ± 297 (4) 28 ± 6 (68)

2004 297 ± 194 (19) 25 ± 9 (101) 126 ± 35 (6) 27 ± 6 (64)

412 ± 440 (29) 268 ± 78 (2) 253 ± 206 (18)

Results Salmon lice infection in the Altafjord system in 2003 Pelagic zone No salmon lice were found on pelagic captured post-smolts of Atlantic salmon from the Altafjord system at any of the sampling periods in late June and early July. This includes fish captured in the inner, the middle and the outer zone of the Altafjord system.

In contrast, pelagic sea trout were already infected with sea lice in late June, but their abundance was low. In early July (week 28), abundance had increased significantly (Kruskal-Wallis, p < 0.05), median abundance was close to 10 lice, and approximately 25% of the fish captured carried more than 30 lice (Figure 2). In late July (week 30), abundance was slightly reduced, and fewer fish carried high infections although a few fish with 30 – 40 lice were found. The lice structure was dominated by lice larvae in June (week 24 and 26), while older lice gradually also were found later in the season (Figure 3).

125

SUMBAWS

Final Report

100

Pelagic sea trout in Altafjorden in 2003

90

Abundance

80 70 60 50 40 30 20 10 0 24

26

28

30

Week

Figure 2 Abundance of salmon lice in pelagic captured sea trout in the Altafjord system in 2003. Number of fish captured was 16, 17, 22 and 12 in week 24, 26, 28 and 30, respectively. The fish were captured in zone 1 and 2 in the Altafjord system (see Figure 1 for details) Sea trout 100

Stages

Frequency (%)

80

ADF ADM

60 P2F P2M 40 P1F P1M

20

CH3 0

CH1 24

26

28

30

Week number

Figure 3 Developmental stages of salmon lice in pelagic captured sea trout in the Altafjord system in 2003.

A few Arctic char was also captured in pelagic areas of Altafjord in 2003. In late June abundance was low.

Abundance had significantly increased in early July (Kruskal-Wallis, p < 0.05), and some of the char carried high numbers of lice (Figure 4).

80

Pelagic charr in Altafjorden in 2003

Abundance

60

40

20

0 26

28

30

week

Figure 4 Abundance of salmon lice in pelagic captured Arctic char in the Altafjord system in 2003. Number of fish captured was 4, 4 and 1 in week 26, 28 and 30, respectively. The fish were captured in zone 1 and 2 in the Altafjord system (see Figure 1 for details) 126

SUMBAWS

Final Report

Only adult lice were found on the few char captured in late June (week 26). Lice larvae dominated in early July, but older stages were also found (Figure 5).

Arctic char 100

Stages

Frequency (%)

80

ADF ADM

60 P2F P2M 40 P1F P1M

20

CH3 0

CH1 26

28

30

Week number

Figure 5 Developmental stages of salmon lice in pelagic captured Arctic char in the Altafjord system in 2003.

There were no significant differences in infection intensity between sea trout and Arctic char from the same sampling occasions and locations in June and July (Mann-Whitney U-test; p > 0.05), and the infection followed generally the same pattern.

Littoral zone The abundance differed significantly between the different sampling periods of sea trout and Arctic char in the littoral zone (Kruskal-Wallis; p < 0.05). In June, all littoral feeding sea trout and Arctic char in the Altafjord system were almost uninfected with salmon lice (Figure 6). In July (week 29), abundance had increased significantly (Post-Hoc test; p < 0.05), both in sea trout and in Arctic char (Mann-Whitney; U-test; p > 0.05). However, in August, abundance had decreased slightly in sea trout, and most of the fish carried low numbers of lice. No Arctic char were captured in the last sampling round in the littoral zone.

40

Littoral sea trout/charr in Altafjorden 2003

Abundance

30

20

10

0

25

29

33

Week

Figure 6 Box-and-whiskers plot showing abundance of salmon lice in littoral captured sea trout and Arctic char in the Altafjord system in 2003. Number of fish captured was 12, 12, and 7 sea trout, and 12, 14 and 0 arctic char in week 25, 29 and 33, respectively. The fish were captured in zone 1 and 2 in the Altafjord system (see Figure 1 for details) 127

SUMBAWS

Final Report

In general, chalimus stages dominated during in June and July, but there were also a relatively large number of pre-adult and adult stages in July (especially in sea trout), whereas only early chalimus larvae were found in the few infected sea trout in June (Figure 7). In Altafjord in July, low numbers of early (I and II) and late chalimus (III and IV) stages dominated, but only a few adult males were found on the fish in this month. In August, a new infection of lice larvae had occurred in sea trout, but also an aggregation of adult female and male lice occurred on the sea trout.

Artic char

Sea trout 100

120

Stages

Stages

100

80

80

ADM P2F

60 P2M P1F

40

Frequency (%)

Number of Sea lice

ADF

ADF ADM

60 P2F P2M 40 P1F

P1M

P1M

20

20

CH3

CH3 CH

0 25

29

CH1

0 25

Week number

29

33

Week number

Figure 7 Developmental stages of salmon lice in littoral captured sea trout in the Altafjord system in 2003. (see Figure 1 for details) Salmon lice infection in the Altafjord system in 2004 Pelagic zone No salmon lice were found on the pelagic captured post-smolts of Atlantic salmon from the Altafjord system in late June (week 27). In early July (week 29), a few fish were infected with low numbers of salmon lice. This was also the case later in July (week 30), although only a few fish were captured (Figure 8). With one exception (chalimus 3), only early (chalimus 1 and 2) chalimus stages were found on the fish (results not shown).

3,0

Atlantic salmon post smolt Altafjorden 2004

Abundance

2,5

2,0

1,5

1,0

0,5

0,0 27

29

30

Week

Figure 8 Box-and-whiskers plot showing abundance of salmon lice in pelagic captured Atlantic salmon post-smolt in the Altafjord system in 2004. Number of fish is 101, 60, and 4 in week 27, 29 and 30, respectively. No fish were caught in week 25. The fish were captured in both zone 1, 2 and 3 in the Altafjord system (see Figure 1 for details) 128

SUMBAWS

Final Report

Pelagic captured sea trout were infected with small numbers of salmon lice in late June (week 27).

Median abundance was also low in week 30 but a few of the sea trout carried more than 40 lice (Figure 9).

50

Pelagic sea trout in Altafjorden 2004

Abundance

40

30

20

10

0 27

30

Week

Figure 9 Box-and-whiskers plot showing abundance of salmon lice in pelagic captured sea trout in the Altafjord system in 2004. Number of fish is 20 and 6 in week 27 and 29, respectively. No fish were caught in week 29. The fish were captured in both zone 1 and 2 in the Altafjord system (see Figure 1 for details).

The lice population structure was dominated by larvae in week 27, but in addition to lice larvae more preadult and adult lice were observed in week 30 (Figure 10).

CH1%

100

CH3% P1M% P1F% P2M% P2F%

80

ADM%

Frequency (%)

ADF%

60

40

20

0 27

30

Week

Figure 10 Developmental stages of salmon lice in pelagic captured sea trout in the Altafjord system in 2004. (see Figure 1 for details) Littoral zone The abundance differed significantly between the different sampling periods of sea trout and Arctic char in the littoral zone (Kruskal-Wallis; p < 0.05). In late June (week 27), all littoral feeding sea trout and Arctic char in the Altafjord system were uninfected with salmon lice (Figure 11). Much higher abundance was found in both sea trout and Arctic char in mid July (week 30). Median abundance in sea trout was, however, only a few lice, but approximately 25 % of the fish had close to 10 lice or above. The same structure was found in Arctic char, but general infection in July tended to be slightly lower. Abundance in sea trout increased gradually during August and September, median levels above 5 lice, and more than 25 % of the population carried more than 10 lice. Only two Arctic char were captured in August and only one in September, and these were also infected with a few lice.

129

SUMBAWS

Final Report

Littoral sea trout/charr in Altafjorden in 2004

30

25

Abundance

20

15

10

5

0

27

30

33

36

Week

Figure 11 Box-and-whiskers plot showing abundance of salmon lice in littoral captured sea trout in the Altafjord system in 2004. Number of fish captured was 12, 12, and 7 sea trout, and 12, 14 and 0 arctic char in week 25, 29 and 33, respectively. The fish were captured in zone 1 and 2 in the Altafjord system (see Figure 1 for details) In general, chalimus stages dominated on both sea trout and Arctic char in mid July (week 30). Chalimus stages were also found on sea trout in August along with both preadult and adult lice, and the same situation was observed in September (Figure 12) Arctic char Sea trout CH1%

100

CH1%

100

CH3%

CH3%

P1M%

P1M%

P1F%

P1F% P2m%

P2m% P2F%

60

40

P2F%

80

ADM%

ADM%

ADF%

ADF%

Frequency (%)

Frequency (%)

80

60

40

20

20

0

0 30

36

Week

30

33

36

Week

Figure 12 Developmental stages of salmon lice in pelagic captured sea trout in the Altafjord system in 2004. (see Figure 1 for details)

Discussion The present study confirms that the risks of salmon lice infection may differ between sympatric populations of Atlantic salmon post-smolts, sea trout and Arctic char in the same north Norwegian fjord system. This may be explained by generic differences in their marine life-history, including migratory timing, behavioural pattern and duration of fjord residency in relation to the typical pattern of lice infection in northern fjords.

In the Altafjord system, captured Atlantic salmon post-smolts were uninfected with lice in 2003. The same general pattern appeared in 2004, but a few post-smolts were infected with small numbers of lice in the very outer area of the fjord in 2004 (Pül Arne Bjørn, Fiskeriforskning, Norway, pers. obs.). This is the first time that outwardly-migrating Atlantic salmon post-smolts have been found infected with salmon lice in northern Norwegian fjord systems, despite extensive sampling programs which have been carried out since 2000 (Rikardsen et al 2004). Of particular interest in the 2004 sampling program was the success in following the fish all the way from the river outfall to the fjord mouth. While all post-smolts in the inner 130

SUMBAWS

Final Report

area of the fjord were uninfected, almost half of the fish carried a few lice by the time they were captured at the fjord mouth (Pål Arne Bjørn, Fiskeriforskning, Norway, pers. obs.). The inner part of the Altafjord system only harbours a few farms, while the farming activity is more intense in outer fjord areas. These observations are therefore in accordance with those of Krkošek et al. 2005, and indicate that the risks of salmon lice infection increases as the smolts pass through intensively farmed areas and towards the open sea.

However, abundances of salmon lice on Atlantic salmon post-smolts were much lower than on pelagicfeeding sea trout, and the few pelagic-feeding Arctic char, captured in the same areas as those through which the salmon migrate. In contrast to Atlantic salmon post-smolts, which are believed to transit Norwegian fjord systems quite rapidly (Rikardsen et al. 2004), sea trout and Arctic char remain actively feeding on herring fry and sand lance in pelagic fjord areas (Rikardsen and Amundsen, 2005). This behavioural pattern may render pelagic sea trout and char to especially high infection risks, both due to their very active feeding behavioural pattern and habitat choice which may contain high abundances of prey species but also present a high abundance of infective lice stages. Both zooplankton and pelagic fish tend to aggregate in fjord areas with upwelling and a turbulent current pattern. This may also be the preferred areas for infective salmon lice stages (Heuch, 1995), and may explain the surprisingly high infection found among some of the pelagic feeding sea trout and char whilst Atlantic salmon post-smolts captured concurrently carried only low salmon lice infections.

In contrast, both sea trout and Arctic char captured in the littoral zone were uninfected with salmon lice in late June. By July, most of the sea trout and Arctic char in Altafjord were infected with salmon lice, and mostly the same pattern with little difference between the species was found in both 2003 and 2004.

However, while lice infection in littoral feeding sea trout and Arctic char levelled off in 2003, both prevalence and intensity increased among sea trout in August and September 2004. In contrast, only few littoral feeding Arctic char were captured in August and September, and these also tended to carry fewer lice. Atlantic salmon post-smolts were infected with small numbers of lice in July, but the post-smolts of this species had probably left the fjord during the latter part of July when infection pressure apparently increased. This pattern of infection is, to a large extent, consistent with earlier studies in northern areas where the infection pressure has been found to increase usually in August after relatively low levels in June and July (Bjørn et al., 2001; Bjørn & Finstad, 2002). There are, however, also between year variations and increased infection pressure has also been observed in early July, both in Altafjord (Bjørn & Finstad, 2002) and at different localities in Troms County (Bjørn et al., 2000).

These results indicate that a difference in risks of infection may exist between pelagic migratory Atlantic salmon post-smolts compared to pelagic and littoral fjord feeding sea trout and Arctic char in northern fjords. The infection pattern in sea trout and Arctic char showed remarkable similarity between years: very low infections were found in June, while the prevalence increased in July and both prevalence and mean intensity peaked in August and September at levels significantly higher than assumed historical levels (Boxshall, 1974; Pemberton, 1976) and in areas without fish farms in recent years (Tingley et al., 1997; Schram et al., 1998; Bjørn et al., 2001; Bjørn & Finstad, 2002; Rikardsen, 2004).

Recently the relationship between the expansion of the fish farming industry along the coast, and cooccurring salmon lice infestation on both farmed and wild salmonids, has been discussed intensively (see e.g. Pike & Wadsworth, 1999). Direct evidence of lice transfer from farmed to wild hosts has so far not been generated. There is, however, substantial evidence indicating that infective copepodids from salmon lice on farmed fish have a role in generating the observed epidemics on wild salmonid fish in farming areas (Tully & Whelan, 1993; Tully et al., 1993ab; Heuch & Mo, 2001; Bjørn et al., 2001; Bjørn & Finstad, 2002).

Within Atlantic salmon farms, or within farming areas, large numbers of hosts are continuously present, facilitating a build up of large numbers of gravid female lice and a continuous re-infection (Heuch & Mo, 2001). Years with optimal conditions for lice reproduction, (optimal temperature and salinity) may potentially result in salmon lice epidemics in wild salmonids (Bjørn et al., 2001). At least in sea trout populations, these epidemics are characterised by high infection pressures leading to physiologically damaging, or even lethal lice infections levels, premature return to freshwater of the most heavily infected fish, as well as indices of direct parasite induced mortality of heavily infected fish (Bjørn et al., 2001). The risks of infection of wild salmonids from free swimming salmon lice copepodids derived from cultured fish will depend on e.g. the number and dispersal of lice from fish farms, the behaviour and longevity of infective copepodids, and the feeding or migratory areas occupied by wild salmonids in relation to farms.

The risks of salmon lice infection may, therefore, differ between salmonid species, behavioural choices and age classes, as indicated in the results of the present study, and may also differ from the infection risk in more southern Norwegian areas. Results from Bjørn & Finstad (1998) imply, although not directly tested, 131

SUMBAWS

Final Report

that salmon lice have similar development and growth rates on sea trout, Arctic char and Atlantic salmon compared with the Pacific salmon species Oncorhynchus spp. (Johnson & Albright, 1992; Johnson, 1993).

Similar results have also been obtained from field studies where both sea trout and Arctic char have become heavily infected with salmon lice when feeding in fjords and coastal areas of intensive fish farming activity (Bjørn et al., 2001; Bjørn & Finstad, 2002). Fjord migrating post-smolts of Atlantic salmon have been found with low salmon lice infection in fjords almost lacking fish farming activity (Finstad et al., 2000), but dramatically high infection rates have also been found in posts-smolts descending intensively farmed areas of western Norway (Holst & Jakobsen, 1999).

However, the marine migratory behaviour of Atlantic salmon, Arctic char and sea trout diverge in several important aspects, although detailed knowledge is still limited. Most of the information gathered to date suggests that post-smolts of Atlantic salmon move relatively quickly through the estuary and the fjord close to the surface (Hvidsten et al., 1992; Moore et al., 1995; Holm et al., 1998; Voegeli et al., 1998; Moore et al., 2000; Thorstad et al., 2004; Finstad et al., 2005), although this may vary between populations and years (Hutchinson & Mills, 2000; Rikardsen et al., 2004a; Knudsen et al., 2005). In contrast, sea trout and Arctic char usually feed in littoral areas close to their native river throughout the summer and autumn (Berg & Jonsson, 1990; Finstad & Heggberget, 1993; Lyse et al., 1998; Rikardsen et al., 2000; Rikardsen & Amundsen, 2005). Sea trout and Arctic char furthermore seem to belong to a “near shore, surface orientated quild of fishes” as previously suggested by Grønvik and Klemetsen (1987), although these fish may also display pelagic feeding in open water within fjords at some occasions (Rikardsen & Amundsen, 2005).

Fjord-feeding sea trout and Arctic char may therefore more actively seek distinct thermoclines and haloclines or other areas with a turbulent current pattern, where both fish prey and infective salmon lice copepodids seem to be aggregated (Heuch et al., 1995).

There are also differences in the timing of the migration between the species; often large migrant veteran Arctic char descend prior to veteran sea trout, followed by smolts of Atlantic salmon, Arctic char and sea trout (Tuff Carlsen et al., 2004). Similarly, there are also differences in the sea water residency of Arctic char and sea trout within fjords. Arctic char usually ascend earlier than sea trout in late summer or autumn.

The large veterans predominate among the early ascenders, while first-migrant post-smolts return later (Berg & Jonsson, 1990; Rikardsen et al., 1997). In addition to possible immunological differences between the salmonid species (Dawson et al., 1997), these behavioural differences may lead to different risks of infection at sea. Furthermore, the sea water temperature is usually much lower in the north than in the southern fjords Norway (Rikardsen et al., 2004a). As a result of this, the peak in infection pressure often occurs on trout and chars in August - October at northern latitudes (Bjørn & Finstad, 2002; Rikardsen, 2004; this study), compared to early summer at more southern latitudes (Schram et al., 1998). Usually, the Atlantic salmon post-smolts have left the northern fjords (Rikardsen et al., 2004a) before the infection pressure is expected to be highest. Atlantic salmon post-smolts may therefore usually experience “mismatch” conditions, in northern fjords, between the peak in infection risk and the normal migration period. This may also be the case for veteran migrant sea trout, and especially Arctic char veterans which often ascend the rivers in July or early August (Berg & Jonsson, 1990; Rikardsen et al., 1997). Immature first-migrant sea trout often stay longer in seawater (Berg & Jonsson, 1990), and veteran sea trout seem also to have a more pelagic feeding pattern that Arctic char.

Sea trout may, therefore, often be subject to rather high salmon lice infections, and results from this study, as well as previous results from the Altafjord system (Bjørn et al., 2001; Bjørn & Finstad, 2002), indicate that relative infection intensities reach levels that may impair fish physiology (Bjørn & Finstad, 1997; Nolan et al., 1999; Bjørn et al., 2001; Wagner et al., 2003, 2004). In contrast, Arctic char may in general spend too short a time at sea (average 30-50 days, Finstad & Heggberget, 1993, 1995; Rikardsen, 2000) for the salmon lice to develop to the more harmful mobile pre-adult and adult stages, and may also behaviourally to a lesser extent seek feeding areas which also have high infection risks. Results from the present study therefore imply that sea trout may be rendered at especially high risks of salmon lice infection through their behavioural patterns at sea. Atlantic salmon post-smolts in northern Norwegian fjords seem to have low risks of infection, while Arctic char seemingly may be in an intermediate position. Integrated salmon lice management models must therefore be based on regional and local (fjord) ecological and environmental information to ensure that the measures taken are both appropriate and sufficient to protect wild salmonid stocks.

132

SUMBAWS

Final Report

Acknowledgements Financial support for this study was given by the Norwegian Research Council (The Wild Atlantic Salmon Program no. 149796/720), the Norwegian Directorate for Nature Management and the European Commission contract no: Q5RS-2002-00730.

References Bakke, T.A. & Harris, P.D. (1998). Diseases and parasites in wild Atlantic salmon (Salmo salar) populations. Canadian Journal of Fisheries and Aquatic Sciences 55, 247-266.

Berg, O.K. & Jonsson, B. (1990) Growth and survival rates of the anadromus trout, Salmo trutta, from the Vardnes River northern Norway. Environmental Biology of Fishes 29, 145-154 Birkeland, K. (1996). Salmon lice, Lepeophtheirus salmonis Krøyer, infestations and implications for anadromous brown trout, Salmo trutta L. Dr. scient thesis, University of Bergen.

Bjørn, P. A. and Finstad, B. (1997). The physiological effects of salmon lice infection on sea trout post smolts. Nordic Journal of Freshwater Research 73, 60-72.

Bjørn P.A. & Finstad B. (1998). The development of salmon lice (Lepeophtheirus salmonis) on artificially infected sea trout (Salmo trutta) post smolts. Canadian Journal of Zoology 76, 970-977.

Bjørn, P.A. & Finstad, B. (2002). Salmon lice, Lepeophtheirus salmonis (Krøyer), infestation in sympatric populations of Arctic char, Salvelinus alpinus (L.), and sea trout, Salmo trutta (L.), in areas near and distant from salmon farms. ICES Journal of Marine Sciences 59, 131-139.

Bjørn P.A., Kristoffersen R. & Finstad B. (2000). Lakselus på vill sjøørret og sjørøye i Troms sommeren 1999 (Salmon lice on wild sea trout and Arctic char in Troms – summer 1999). – Rapport til Fiskehelse og Miljøgruppa i Troms, Fiskeridirektoratet, region Troms, 1-34 (in Norwegian).

Bjørn, P. A., Finstad, B. & Kristoffersen, R. (2001). Salmon lice infections of wild sea trout and Arctic char in marine and freshwaters: the effects of salmon farms. Aquaculture Research 31, 795-803.

Boxshall, G. A. (1974). Infection with parasitic copepods in north sea marine fishes. Journal of Marine Biology Association U.K. 54, 355-372.

Bush, A. O., Lafferty, K. D., Lotz, J. M. & Shostak, A. W. (1997). Parasitology meets ecology on its own terms: Margolis et al. revised. Journal of Parasitology 83, 575-583.

Butler, J. R. A. (2002). Wild salmonids and sea louse infestations on the west coast of Scotland: sources of infection and implications for the management of marine salmon farms. Pest Management Science 58, 595608.

Dawson, L.H.J., Pike, A.W., Houlihan, D.F. & McVicar, A.H. (1997). Comparison of the susceptibility of sea trout (Salmo trutta L.) and Atlantic salmon (Salmo salar L.) to sea lice (Lepeophtheirus salmonis (Krøyer, 1837)) infections. ICES Journal of Marine Science 54, 1129-1139.

Finstad, B. & Heggberget, T. G. (1993). Migration, growth and survival of wild and hatchery-reared anadromous Arctic charr (Salvelinus alpinus) in Finnmark, northern Norway. Journal of Fish Biology 43, 303-312.

Finstad, B. & Heggberget, T. G. (1995). Seawater tolerance, migration, growth and recapture rates of wild and hatchery-reared Arctic charr (Salvelinus alpinus (L.). Nordic Journal of Freshwater Research 71, 229236.

Finstad, B., Grimnes, A., Bjørn, P.A. & Hvidsten, N.A. (2000). Laboratory and field investigations of salmon lice [Lepeophtheirus salmonis (Krøyer)] infestation on Atlantic salmon (Salmo salar L.) postsmolts.

Aquaculture Research 31, 795-803.

Finstad, B., Økland, F., Thorstad, E.B., Bjørn, P.A. & McKinley, R.S. (2005). Migration of hatcheryreared Atlantic salmon and wild sea trout post-smolts in a Norwegian fjord system. Journal of Fish Biology, in press.

133

SUMBAWS

Final Report

Gargan, P.G., Tully, O., and Poole, W.R. 2003. Relationship between sea lice infestation, sea lice production and sea trout survival in Ireland, 1992-2001. In Salmon at the edge. (Mills, D., ed.), pp. 119-135.

Oxford,: Blackwell Science.

Grønvik, S. & Klemetsen, A. (1987). Marine food and diet overlap of co-occurring Arctic charr Salvelinus alpinus (L.), brown trout Salmo trutta L. and Atlantic salmon S. salar L. off Senja, N. Norway. Polar Biology 7, 73-177.

Gross, M. R. (1987). Evolution of diadromy in fishes. American Fisheries Society Symposium 1, 14-25.

Gross, M. R. (1998). One species with two biologies: Atlantic salmon (Salmon salar) in the wild and in aquaculture. Canadian Journal of Fisheries and Aquatic Sciences 55 (Suppl. 1.) 131-144.

Heuch, P. A. & Mo, T. A. (2001). A model of salmon louse production in Norway: Effects of increasing salmon production and public management measures. Diseases of Aquatic Organisms 45, 145–152.

Heuch, P.A., Parsons, A. & Boxaspen, K. (1995). Diel vertical migration: A possible host-finding mechanism in salmon louse (Lepeophtheirus salmonis) copepodids? Canadian Journal of Fisheries and Aquaculture 52, 681-689.

Heuch, P. A., Knutsen, J. A., Knutsen, H. & Schram. T. (2002). Salinity and temperature effects on sea lice over-wintering on sea trout (Salmo trutta) in coastal areas of the Skagerrak. Journal of the Marine Biology Association of the United Kingdom 82, 887-892.

Hansen, L P., Holm, M., Holst, J. C. and Jakobsen, J. A. (2003). The Ecology of Post-Smolts of Atlantic salmon (eds. D. Mills). In Salmon at the edge. Blackwell Science, Oxford. UK. Heuch, P. A., Bjørn P. A., Finstad, B., Holst, J. C., Asplin, L. & Nilsen, F. (2005) Relationship between salmon lice on wild and farmed salmonids: A review of population dynamics, management measures and effects on wild salmonid fish stocks in Norway. Aquaculture 246, 79-92.

Holm, M., Axelsen, B.E., Sturlaugsson, J., Hvidsten, N.A., Ikonen, E. & Johnsen, B.O. (1998).

Behaviour of acoustically tagged post-smolts in the Trondheim fjord – Influence of hydrografical and metrological conditions on migration. International Council for the Exploration of the sea CM 1998/N:17.

Holst, J.C. & Jakobsen, P.J. (1999). Lakselus knekker vestlandslaksen (Salmon lice kills the salmon in Western Norway). Norsk Fiskeoppdrett 16: 38–39 (in Norwegian).

Holst, J. C. & McDonald, A. (2000). FISH-LIFT: a device for sampling live fish with trawls. Fisheries Research 48, 87-91.

Holst, J. C., Jakobsen, P., Nilsen, F., Holm, M., Asplin, L. and Aure, J. (2003). Mortality of SeawardMigrating Post Smolts of Atlantic Salmon Due to Salmon Lice Infections in Norwegian Salmon Stocks. (eds. D. Mills). In Salmon at the edge. Blackwell Science, Oxford. UK. Hutchinson, P. & Mills, D. (2000). Executive summary. In The Ocean Life of Atlantic salmon.

Environmental and Biological factors Influencing Survival (Mills, D., ed.), pp. 7-18. Oxford: Blackwell Science.

Hvidsten, N.A., Johnsen, B.O. & Levings, C.D. (1992) Atferd og ernæring hos utvandrende laksesmolt i Trondheimsfjorden (Behaviour and feeding in postsmolts from the Trondheimsfjord). NINA Oppdragsmelding 164. Report in Norwegian with English summary. Norwegian Institute for Nature Research, Tungasletta 2, N-7485 Trondheim, Norway, 1-14 (in Norwegian).

Jonsson, B. (1985). Life history patterns of freshwater resident and sea-run migrant brown trout in Norway.

Transactions of the American Fisheries Society 114, 182-194.

Johnson S.C. (1993). A comparison of development and growth rates of Lepeophtheirus salmonis (Copepoda: Caligidae) on naïve Atlantic (Salmo salar) and Chinook (Oncorhynchus tshawytscha) salmon.

In: Pathogens of wild and farmed fish, sea lice. pp. 68-80. Ed. By G. A. Boxhall and Defaye. Ellis Horwood. London.

Johnson, S.C., Albright, L.J. (1991). Development, growth, and survival of Lepeophtheirus salmonis (Copepoda: Caligidae) under laboratory conditions. Journal of the Marine Biological Assosiation of the United Kingdom 71, 425-436.

134

SUMBAWS

Final Report

Johnson, S.C., Treasurer, J. W., Bravo, S., Nagasawa, K. and Kabata, Z. (2004). A review of the impact of parasitic copepods on marine aquaculture. Zoological studies 43 (2), 229-243.

Klemetsen, A., Amundsen, P.-A., Dempson, B., Jonsson, B., Jonsson, N., O’Connell, M.F. & Mortensen, E. (2003). Atlantic salmon Salmo salar L., brown trout Salmo trutta L. and Arctic charr Salvelinus alpinus (L.): a review of aspects of their life histories. Ecology of Freshwater Fish 12, 1-59.

Knudsen, R, Rikardsen, A.H., Dempson, J.B., Bjørn, P.A., Finstad, B., Amundsen, P.-A. & Holm, M. (2005). Intestinal trophic transmitted parasites as bio-indicators of prolonged fjord-feeding of Atlantic salmon (Salmo salar) post-smolts in Northern Norway. Journal of Fish Biology 65, 000-000, in press.

Krkošek, M., Lewis, M.A. & Volpe, J.P. (2005). Transmission dynamics of parasitic sea lice from farm to wild salmon. Proceedings of the Royal Society 272, 689-696.

Lyse, A.A., Stefansson, S.O. & Fern, Ö. A. (1998). Behavior and diet of sea trout post-smolts in a Norwegian fjord system. Journal of Fish Biology 52, 923-936.

Mo ,T.A. & Heuch, P.A. (1998). Occurrence of Lepeophtheirus salmonis (Copepoda: Caligidae) on sea trout (Salmo trutta) in the inner Oslo Fjord, south-eastern Norway. ICES Journal of Marine Science 55, 176180.

Moore, A., Potter, E.C.E., Milner, N.J. & Bamber, S. (1995) The migratory behaviour of wild Atlantic salmon smolts in the estuary of the River Conwy, North Wales. Canadian Journal of Fisheries and Aquatic Sciences 52, 1923-1935.

Moore, A., Lacroix, G.L. & Sturlaugsson, J. (2000) Tracking Atlantic salmon post-smolts in the sea. In The ocean life of Atlantic salmon - environmental and biological factors influencing survival (Mills, D. Ed.). pp. 49-64. Oxford: Fishing News Books.

Nolan, D.T., Reilly, P., and Wendelaar Bonga, S.E. (1999). Infection with low numbers of the sea louse Lepeophtheirus salmonis induces stress-related effects in postsmolt Atlantic salmon (Salmo salar).

Canadian Journal of Fisheries and Aquatic Sciences 56, 947–959.

Pemberton, R. (1976). Sea trout in the North Argyll sea lochs: population, distribution, and movements.

Journal of Fish Biology 9, 157-179.

Pike A.W. & Wadsworth S.L. (1999). Sealice on Salmonids: Their Biology and Control. Advances in Parasitology 44, 233–337.

Rikardsen, A.H. (2000). Effects of Floy and soft VIalpha tags on growth and survival of juvenile Arctic char. North American Journal of Fisheries Management 20, 719-728.

Rikardsen, A.H. (2004). Seasonal occurrence of salmon lice Lepeophtheirus salmonis on sea trout Salmo trutta in North Norwegian fjords without fish farms. Journal of Fish Biology 65, 711-722.

Rikardsen, A.H. & Amundsen, P.A. (2005) Pelagic marine feeding behaviour of Arctic charr Salvelinus alpinus and sea trout Salmo trutta. Journal of Fish Biology 66, 1163-1166.

Rikardsen, A.H., Svenning, M.-A. & Klemetsen, A. (1997). The relationships between anadromy, sex ratio and parr growth of Arctic charr in a lake in North Norway. Journal of Fish Biology 51, 447-461.

Rikardsen, A.H., Amundsen, P.-A., Bjørn, P.A. & Johansen, M. (2000). Comparison of growth, diet and food consumption of sea-run and lake-dwelling Arctic charr. Journal of Fish Biology 57, 1172-1188.

Rikardsen, A.H., Haugland, M., Bjørn, P.A. Finstad, B., Knudsen, R., Dempson, B., Holm, M., Holst, J.C. & Hvidsten, N.A. (2004a). Geographical differences in early marine feeding of Atlantic salmon postsmolt in Norwegian fjords. Journal of Fish Biology 64, 1655-1679.

Rikardsen, A.H., Thorpe, J.E. & Dempson, B. (2004b). Modelling the life-history variation of anadromous and resident Arctic charr Salvelinus alpinus (L.). Ecology of Freshwater Fish 13, 305-311.

Schram, T.A., Knutsen, J.A., Heuch, P.A. & Mo, T.A. (1998). Seasonal occurrence of Lepeophtheirus salmonis and Caligus elongatus (Copepoda: Caligidae) on sea trout (Salmo trutta), off southern Norway.

ICES Journal of Marine Science. 55, 163-175.

135

SUMBAWS

Final Report

Thorstad, E.B., Økland, F., Finstad, B., Sivertsgård, R., Bjørn, P.A. & McKinley, R.S. (2004).

Migration speeds and orientation of Atlantic salmon and sea trout post-smolts in a Norwegian fjord system.

Environmental Biology of Fishes 71, 305-311.

Tingley, G. A., Ives, M. J. & Russel, I. C. (1997). The occurrence of lice on sea trout (Salmo trutta L.) captured in the sea off the East Anglian coast of England. ICES Journal of Marine Science 54, 1120-1128.

Todd, C. D., Walker, A. M., Wolff, K., Northcott, S. J., Walker, A. F., Ritchie, M. G., Hoskins, R., Abbott, R. J. & Hazon, N. (1997). Genetic differentiation of populations of the copepod sea louse Lepeophtheirus salmonis (Krøyer) ectoparasitic on wild and farmed salmonids around the coasts of Scotland: Evidence from RAPD markers. Journal of Experimental Marine Biology and Ecology 210, 251274.

Tuff Carlsen, K., Berg, O.K., Finstad, B. & Heggberget, T.G. (2004). Diel periodicity and environmental influence on the smolt migration of Arctic charr, Salvelinus alpinus, Atlantic salmon, Salmo salar , and brown trout, Salmo trutta, in northern Norway. Environmental Biology of Fishes 70, 403-413.

Tully, O. & Whelan, K. F. (1993). Production of nauplii of Lepeophtheirus salmonis (Krøyer) (Copepoda: Caligidae) from farmed and wild salmon and its relation to the infestation of wild sea trout (Salmo trutta L.) off the west coast of Ireland in 1991. Fisheries Research 17, 187-200.

Tully, O., Poole, W.R. & Whelan, K.R. (1993a). Infestation parameters for Lepeophtheirus salmonis (Krøyer) (Copepoda: Caligidae) parasitic on sea trout, Salmo trutta L., off the west coast of Ireland during 1990 and 1991. Aquaculture and Fisheries Management 24, 545-555.

Tully, O., Poole, W.R., Whelan, K.R. & Merigoux, S. (1993b). Parameters and possible causes of epizootics of Lepeophtheirus salmonis (Krøyer) infesting sea trout (Salmo trutta L.) off the west coast of Irland. In Pathogens of wild and farmed fish: sea lice. Ed. by (G.A. Boxshall, G.A. & Defaye, D., eds), pp 202-213. Ellis Horwood. London.

Tully, O. & Nolan, D. T. (2002). A review of the population biology and host-parasite interactions of the sea louse Lepeophtheirus salmonis (Copepoda: Caligidae). Parasitology 124, 165-182.

Voegeli, F.A., Lacroix, G.L. & Anderson, J.M. (1998) Development of miniature pingers for tracking Atlantic salmon smolts at sea. Hydrobiologia 371/372, 35-46.

Wagner, G. N., McKinley, R. S., Bjørn, P. A. and Finstad, B. (2003). Physiological impact of sea lice on swimming performance of Atlantic salmon. Journal of Fish Biology 62, 1000-1009.

Wagner, G., McKinley, R.S., Bjørn, P.A., & Finstad, B. (2004). Short-term freshwater exposure benefits sea lice-infected Atlantic salmon. Journal of Fish Biology 64, 1593-1604.

136

SUMBAWS

Final Report

Workpackage 4 Lead Partner: NINA Participating Partners: USTAN, CFB, NIFA, KUN OBJECTIVES To assess the behavioural and physiological consequences of heavily-infested sea trout post-smolts returning prematurely to freshwater infected with (1) copepodids/chalimus or (2) the older ‘mobile’ stages of salmon lice.

DELIVERABLES • •

Assess the physiological effect of salmon lice infestation in salt, brackish and freshwater.

Assess the importance and consequences of “premature return” to freshwater as a behavioural response to promote survival of heavily salmon lice-infested sea trout post-smolts.

EXECUTIVE SUMMARY Coincident with the first reports of stock crashes of sea trout were reports of ‘premature return’ of post smolt sea trout to freshwater (FW), within a few weeks of their first migration to sea. These fish typically bore heavy infestations of largely juvenile (chalimus) stages of lice, perhaps suggesting that host fish carrying mobile sea lice had succumbed to infestation. Available data on the causes, importance and consequences of premature return to freshwater for sea trout post smolts are limited. As a marine parasite, Lepeophtheirus salmonis are readily lost from the host fish on migration to brackish or freshwater for extended periods. The phenomenon of ‘premature’ migratory return to freshwater has been interpreted as an adaptive response by the host to osmoregulatory dysfunction. The overall aim of this workpackage was to assess and quantify the effect of sea lice infestation on osmoregulatory, metabolic and stress physiology before and after return to freshwater.

The physiological consequences of premature return to freshwater were examined in the laboratory by infesting seawater (SW)-acclimated wild sea trout smolts with infective copepodids of L. salmonis. The fish were split into 4 experimental groups, two of which were held in SW and two of which were maintained in SW until 19 days post infection (dpi), at which point they were returned to freshwater (FW). In both salinities, there was an infested group and an uninfested control group. At 19 dpi, L. salmonis had reached the mobile preadult stages. Fish were sampled at 0 dpi (prior to infestation with L. salmonis) and then at 7, 14, 21, 28 and 35 dpi, in order to sample infested fish with sea lice at all stages throughout the life cycle.

Following transfer to FW, the mean infestation intensity was significantly reduced, suggesting that the majority of sea lice were unable to survive transfer to FW for periods of more than a few days. In addition, there was a markedly higher mortality of infested fish maintained in SW when compared with infested fish which were returned to FW at 19 dpi. These results suggest that premature return to FW may be of benefit to sea trout post-smolts in terms of lice infestation intensity and survivorship.

Plasma chloride, an osmoregulatory marker, was significantly higher in the SW infested group than in all other groups following return to FW at 19 dpi. Disturbance of water balance and ion homeostasis is one of the most characteristic aspects of stress in fish, and such stress responses may occur due to structural changes in the gills. Our results therefore suggest that premature return to FW alleviated the marked osmoregulatory disturbances observed in infested fish in SW, again indicating that this behaviour may be of physiological benefit to sea trout. It should be noted however, that there may be further consequences to the fish, such as secondary fungal infection, which was noted on a number of previously lice infested fish following return to FW. Over a prolonged period of time these infections may prove deleterious to the fish.

137

SUMBAWS

Final Report

Significant effects on the metabolic markers, plasma lactate and liver glycogen were also observed.

Elevated plasma glucose and plasma lactate levels have been reported following both acute and chronic stress. No significant differences were observed in plasma glucose between groups, but plasma lactate concentration in the SW infested group was elevated above the controls at 21 dpi, and above all other groups at 28 and 35 dpi. Liver glycogen content increased rapidly in all groups at the start of the experiment, but by 21 dpi liver glycogen content was significantly higher in both control groups that in the infested groups.

Liver glycogen remained depressed in the SW infested group for the remainder of the experiment, but there was evidence of recovery in the infested group that were returned prematurely to FW by 35 dpi. However, it is important to note that in this experiment, fish which were returned to FW had the same access to food as fish remaining in SW. However, in the wild, food availability is considerably reduced in comparison with the SW environment, perhaps further compromising liver glycogen stores. However, our results suggest that premature return to FW may be of physiological benefit to sea trout.

An elevation in plasma cortisol is the most widely used indicator of stress in fish. There was a significant increase in plasma cortisol in both infested groups which was apparent at 14 dpi. However, by the end of the experiment only the SW infested group had a plasma cortisol concentration that was significantly above the control levels, again indicating that premature return to FW may be of physiological benefit to sea trout.

To our knowledge this is the first attempt to measure the effects of premature return in laboratory studies, using a series of osmoregulatory, metabolic and stress measures. It would therefore, appear that despite possible consequences of secondary fungal infection, premature return to FW confers considerable physiological benefits on sea trout infested with L. salmonis.

CONCLUSIONS It was decided to focus on the physiological consequences of sea trout post-smolts returning to FW, infested with mobile lice stages, rather than copepodids/chalimus stages as outlined in the objectives for this workpackage. It was clear from our prior laboratory results in WP 5 that sustained physiological effects were generally not apparent until the sea lice had reached the late chalimus or mobile stages, and a similar effect was observed in the laboratory studies in WP 4. The results from the feeding study were difficult to interpret, due to the surprisingly low sea lice intensity on the experimentally infested fish. However, we successfully measured the physiological effects of premature return of sea lice infested sea trout post smolts in a controlled laboratory environment. Coupled with the use of appropriate statistical modelling techniques we have, for the first time, provided evidence of the physiological benefits of premature return to FW for sea trout post-smolts. Our original objectives, deliverables and milestones have therefore been met.

138

SUMBAWS

Final Report

Paper 1: THE PHYSIOLOGICAL CONSEQUENCES OF ‘PREMATURE RETURN’ TO FRESHWATER FOR WILD SEA TROUT (SALMO TRUTTA L.) POST-SMOLTS INFESTED WITH SEA LICE (LEPEOPHTHEIRUS SALMONIS KRØYER) A. Wells, C. E. Grierson, L. Marshall, M. MacKenzie, I. Russon, H. Reinardy, R. Sivertsgård, P. A. Bjørn, B. Finstad, S. E. Wendelaar Bonga, C. D. Todd and N. Hazon.

(Manuscript) Abstract Infestation of juvenile salmonids with ectoparasitic caligid copepod sea lice can induce osmoregulatory dysfunction and stress of the host fish. Caligids are marine parasites which do not survive in freshwater.

Here we investigated the physiological consequences of ‘premature migratory return’ for wild sea trout (Salmo trutta L.) post-smolts infested with Lepeophtheirus salmonis (Krøyer). Osmoregulatory, metabolic and stress markers were analysed to assess the potential beneficial consequences for the host of return to freshwater at 19 days post-challenge with copepodid larvae. Mean sea lice infestation intensity was significantly reduced following transfer to freshwater. Mortality rates of infested fish maintained in seawater were markedly higher than for fish that were returned to freshwater. Significant sea lice effects, which were consistent across all physiological markers, were apparent once L. salmonis developed to the mobile stages but some effects were observed when the lice were at the (sessile) late chalimus stages. Plasma chloride, lactate and cortisol concentration were significantly higher than control values, and liver glycogen concentration was significantly reduced in sea lice infested fish in seawater. Following return to freshwater, these physiological measures returned to control levels, whereas significant differences remained for fish held in seawater throughout the experiment. Premature migratory return to freshwater of sea trout infested with L. salmonis confers significant physiological benefits as shown by a range of osmoregulatory, metabolic and stress markers.

Introduction In northwest Europe, sea trout (Salmo trutta L.) are the anadromous form of the freshwater brown trout.

Unlike Atlantic salmon (Salmo salar L.), which undertake extensive oceanic migrations (Hansen 2000), sea trout remain in nearshore waters throughout their marine residence and many return to freshwater to overwinter as non-reproductive juveniles (Klemetsen et al. 2003). In recent years there has been considerable concern expressed over the catastrophic declines of certain sea trout stocks in Ireland, Scotland and Norway (Anonymous 1993, 1995; Northcott & Walker 1996). These stock declines appear to be attributable primarily to increased marine mortality (Pike & Wadsworth 1999), and there is now strong circumstantial evidence that sea lice may be a major contributory factor to marine mortality rates (Butler 2002; McKibben & Hay 2004; Penston et al. 2004).

Coincident with the first reports of stock crashes were reports of ‘premature return’ of post-smolt sea trout to freshwater, within a few weeks of their first migration to sea. These fish typically bore heavy infestations of largely juvenile (chalimus) stages of lice (MacKenzie et al. 1998; Tully et al. 1993a; Tully & Whelan 1993). Despite their large sample sizes and high abundances of juvenile sea lice stages on infested fish, Tully et al. (1993a) failed to capture juvenile sea trout bearing adult sea lice, and concluded that any such host fish had succumbed to infestation.

The copepod sea lice, Lepeophtheirus salmonis (Krøyer) and Caligus elongatus (Nordmann) are the primary ectoparasites of farmed and wild salmonids (Pike & Wadsworth 1999). High abundances of sea lice induce stress and osmoregulatory imbalance of the host fish and these effects are especially intense when the parasites develop to the mobile preadult and adult stages (Bjørn & Finstad 1997; Dawson et al. 1998; Grimnes & Jakobsen 1996). Sea lice feed on the host epidermis, and feeding activity by mobile sea lice can result in increased skin damage (Bjørn & Finstad 1998; Pike 1989), which can be fatal to hatchery-reared sea trout and salmon (Bjørn & Finstad 1997; Dawson et al. 1998; Grimnes & Jakobsen 1996). Infestation of sea trout smolts with large numbers of the sessile chalimus larval stages leads to increased plasma levels of the stress hormone cortisol (Bjørn and Finstad 1997). Cortisol causes a range of effects, including hydromineral imbalance, changes in intermediary metabolism, reduction in the immune response and reduced capacity to tolerate subsequent or additional stressors (Wendelaar Bonga 1997).

139

SUMBAWS

Final Report

Available data on the causes, importance and consequences of premature return to freshwater for sea trout post-smolts are limited. As a marine parasite, L. salmonis is readily lost from the host fish on migration to brackish or freshwater for extended periods (McLean et al. 1990). However, there are reports of clear host species differences, and L. salmonis infesting Arctic char (Salvelinus alpinus L.) can survive for up to 3 weeks in freshwater (Finstad et al. 1995). The phenomenon of premature migratory return to freshwater has been interpreted as an adaptive response by the host to osmoregulatory dysfunction (Birkeland 1996; Birkeland & Jakobsen 1997; Bjørn et al. 2001; Tully et al. 1993a; Tully et al. 1993b). Although wild Atlantic salmon post-smolts do not naturally show premature return, swimming performance of sea liceinfested salmon post-smolts improved upon transfer to freshwater, providing further evidence that a link exists between the physiological effects of sea lice infestation and premature return (Wagner et al. 2004).

Despite early return to freshwater allowing recovery from sea lice infestation, this benefit must be balanced against long-term health consequences, including minor osmoregulatory disturbances and increased mortality arising from sea lice-induced tissue damage (Birkeland 1996; Bjørn et al. 2001). Such lesions may, for example, render the fish susceptible to fungal or bacterial infection in freshwater, and even if premature returning fish survive, their growth opportunities and future fecundity may be compromised by an abbreviated migration to sea. Our primary objective was to assess, under controlled laboratory conditions, the physiological consequences of premature return to freshwater for infested wild sea trout post-smolts.

Materials and Methods Wild sea trout smolts in freshwater (FW) were trapped from the Hals watercourse, Altafjørd, Northern Norway (70°05’N, 22°55’E), between June 24-28, 2004 and transferred to tanks in the adjacent Talvik Research Station. Smolts were randomly allocated to experimental groups in duplicate tanks and maintained in FW under natural photoperiod (24 h light) until transfer to seawater (SW) (33‰) on 29 June. There were 4 experimental groups: Control (SW) – non-infested fish maintained throughout in SW; Infested (SW) – sea lice-infested fish maintained throughout in SW; Control (SW - FW) – non-infested fish maintained in SW until 19 days post infestation (dpi) then transferred to FW; Infested (SW - FW) – sea lice-infested fish maintained in SW until 19 dpi then transferred to FW. Because only six 800 litre tanks were available, two 150 litre tanks were also used. In order to ensure that all sea lice infestations occurred in identical tanks, the two 150 litre tanks were used as control tanks, but in both control groups one 150 litre tank and one 800 litre tank were used as duplicate pairs. 800 litre tanks were allocated to the remaining experimental groups on a random basis. Fish were fed to satiation on a combined diet of chopped squid and commercial fish food pellets, which were mixed with cod roe in order to encourage feeding.

Ovigerous female sea lice (L. salmonis) were collected and cultured at The Marine Institute, Bergen.

Infective copepodids (2-4 d old) were transported by air in flasks of cooled SW and immediately used to infest the fish. The total number of larvae was estimated by extrapolating from the number of larvae in five 1 ml samples drawn from the freshly mixed, pooled contents of three flasks (total volume 3200 ml).

At infestation the volume of water in the experimental tanks was reduced to 100 litres and the water supply turned off. The water was continuously aerated during the infestation, allowing circulation to be maintained in the tanks. A calculated number of copepodids was introduced with the intention of achieving the desired infestation level of ~50 lice · fish-1: approximately 100 lice · fish-1 were therefore added to each experimental tank. The remaining control tanks were treated in an identical manner, but no copepodid larvae were added.

The infestation was allowed to progress for 4 h in total darkness, following which all tanks were refilled, and the continuous light regime reinstated.

Control samples of fish were taken from all tanks on the day prior to sea lice infestation. All tanks were then sampled at 7, 14, 21 and 28 dpi. Fish were individually netted from the experimental tanks, taking care to minimise disturbance to the remaining fish, and anaesthetised in 2-phenoxyethanol (2.5 ml l-1) using individual containers for each fish. The anaesthetic was passed through a 40 µm filter in order to collect and classify any lice which became dislodged during sampling. Five fish from each tank (10 fish for each lice intensity) were removed at each sampling date and blood and tissues were collected within 15 min. Blood samples (0.5 ml) were taken from the caudal blood sinus, added to 50 µl 15% ammonium EDTA and processed immediately (see below). The fish were then opened along the mid-ventral line, the livers removed and freeze clamped in liquid nitrogen. The frozen livers were placed in pre-weighed vials and stored at -20ºC until analysis. Fish were placed in individual plastic bags and stored at 4ºC or -20ºC for later examination for sea lice. All dissection was performed on a piece of paper towelling and this was also bagged and subsequently examined for dislodged lice. The fork length and weight of all fish were recorded and expressed as the composite condition factor, Fulton’s K (weight in g/(length in cm)3 ×100). Ratio condition factors are sensitive to isometric variation and departures from the cube relationship, but the size 140

SUMBAWS

Final Report

range of experimental fish was narrow and Fulton’s K has been shown to have a high correlation with whole fish lipid content (Neff & Cargnelli 2004; Sutton et al. 2000). A measure of condition which is more independent of length than Fulton’s K is the relative mass index, WR (WR = W/WS), whereby W is the observed mass and WS the standard mass of individual fish calculated from a specific mass-length equation (Blackwell et al. 2000; Brenden et al. 2003). We currently do not have access to a mass-length relationship for Norwegian sea trout, but there was a significant correlation (r2 = 0.993) between the present K values and WR, as calculated from a length-weight relationship for Scottish sea trout (A.F. Walker, unpubl. data).

Blood plasma was separated by centrifugation at 13,000 rpm for 1 min, and frozen at -80ºC until analysis. Samples were transferred on dry ice to University of St Andrews, Scotland and Radboud University, Nijmegen, Netherlands, for analysis. Plasma chloride was measured by colorimetric titration (Chloride analyser 925, Corning, UK). Plasma glucose concentrations were determined in duplicate using a commercially available kit (Randox laboratories Ltd, Co Antrim, UK – ref: GL 2623), and absorbance was read at 490 nm. Plasma lactate concentrations were determined in duplicate using commercially available lactate reagent (Randox laboratories Ltd, Co Antrim, UK – ref: LC 2389) and read at 540 nm. Both of the latter assays were modified for use in 96-well plates and absorbance was read on a plate reader (Biotech Instruments Inc, USA). Liver glycogen was assessed following hydrolysation by amyloglucosidase using the method of Keppler and Decker (1974). Plasma cortisol was measured by RIA, as described by van Anholt et al (2003).

The body surface, fins and gills were separated analytically into a total of 11 zones as previously described (Bjørn & Finstad 1998) (Figure 1). These zones on all infested fish were examined for sea lice, and the frequencies of each developmental stage were recorded. Sea lice at chalimus stages 1 and 2 and chalimus stages 3 and 4 were pooled into two groups, termed C1 and C3 respectively. Older (post larval) lice stages were classified and their sex determined according to Johnson and Albright (1991) and Schram (1993).

Figure 1 The 11 zones on the body of the infested sea trout smolts examined for sea lice (Bjørn & Finstad 1998). A, head and operculum; B, anterior dorsal area of the body above the lateral line from the head to the posterior end of the dorsal fin; C, posterior dorsal area of the body above the lateral line from the posterior end of the dorsal fin to the caudal fin; D, anterior ventral area of the body below the lateral line from the head to the posterior end of the dorsal fin; E, posterior dorsal area of the body below the lateral line from the posterior end of the dorsal fin to the caudal fin; F, dorsal fin; G, pectoral fin; H, ventral fin; I, anal fin; J, caudal fin; K, gills. The anaesthetic solutions, plastic storage bags, and paper towelling on which dissection took place were also examined and are shown as zone L. Data Analysis A maximum of 3 fish were removed from the data during analysis due to high influence points, as determined by Cook’s distances (Fox 1997). The number of sea lice, fish length, weight, and condition factor, treatment group and dpi were considered as candidate variables in each model, in addition to interaction terms involving sea lice, dpi and group. The data also were checked for significant sampling order and tank effects. Length and weight were too highly correlated to include simultaneously in the models; these had variance inflation factors of 19.80 and 21.34 respectively, and since length resulted in better-fitting models overall, this was chosen instead of weight.

141

SUMBAWS

Final Report

It was suspected that nonlinear relationships between the covariates and response variables existed, and thus Generalized Additive Models (GAMs) were used to model nonlinearities. Specifically, cubic regression splines were fitted inside a linear model framework and the necessity of these nonlinearities (as compared with their linear alternatives) was assessed using F-tests. Interaction effects involving time, sea lice and group were also considered in each model because the treatments given to each group and the effects of sea lice on the fish were closely related to time.

For models in which sea lice intensity was retained as a variable, it was important to obtain groupspecific measures of average sea lice numbers across time. For example, it was clear during the experiment that freshwater re-entry resulted in sea lice detachment, and fish sampled after re-entry had substantially fewer sea lice than were originally attached. This led to the group-specific modelling of average sea lice numbers across time using a generalized nonlinear quasipoisson model. The tank-specific sea lice loadings resulting from this model, at each time point, were then used to generate model predictions for each physiological measure to help ensure that a realistic number of sea lice at each time point was used for each model. Mean values of model covariates were used to generate these predictions.

A tank identifier was included in the model to account for any tank-specific effects in sea lice number because averages were significantly different across the two infested SW tanks and the two infested SW-FW tanks. Sea lice predictions based on this model were then averaged across the tanks at each time point, for both the infested SW and infested SW-FW groups; this resulted in slightly higher sea lice numbers for fish in the infested SW-FW group (Table 3). These results were used to predict levels of liver glycogen and plasma chloride, glucose and lactate through time, as well as plasma cortisol at 14 dpi (Figures 2-5).

For each response variable, predictions were obtained for each time point based on the experimental group and the relevant predicted number of sea lice for that group at that time. Any other covariates included in the model were fixed at their mean across dpi, except for plasma cortisol which was considered at each time point. In this analysis, non-overlapping intervals are considered to represent significant differences between underlying means in each time-specific group; i.e. intervals seen to overlap, even in part, are not considered to represent significant differences between group means. It should also be noted that these intervals are not adjusted for pair-wise comparisons.

Models for each physiological marker were selected using AIC scores (Akaike 1973), and p-values resulting from F-tests and likelihood ratio tests (Fox 1997). Typically, a preliminary model was chosen using the AIC score which was followed by the omission of non-significant covariates at the 5% level; minor variations to this procedure are mentioned individually in the results section. All models were initially fitted assuming normally distributed, independent errors with constant error variance; these assumptions were all assessed visually and supplemented by tests of significance. Specifically, Breusch-Pagan, DurbinWatson and Shapiro-Wilks tests were used for model residuals (Breusch & Pagan 1979; Fox 1997; Royston 1982). Any models found to have non-constant error variance, but which exhibited symmetry, were fitted using a generalised least squares framework, allowing a more appropriate mean-variance relationship.

Models with response data bounded by zero were modelled using a generalized linear model with Gamma errors, as were models exhibiting non-normal errors and non-constant error variance.

Results Development of L. salmonis The progression of L. salmonis development is summarised in Table 1. L. salmonis numbers were obtained for sampled fish at each experimental timepoint (7, 14, 21, 28 & 35 dpi). The sporadic fish mortalities also were examined for L. salmonis infestation and this allowed observations between sampling days and the recording of the first appearance of mobile L. salmonis stages on the fish. Mobile, preadult L. salmonis were observed on fish from 17 dpi onwards. Prevalence of L. salmonis was 100% on each sampling occasion, with the exception of fish that were returned to FW at 19 dpi. Fish in these groups were found to have no sea lice at 28 and 35 dpi and in all cases the infestation intensity was greatly reduced in comparison with the infested groups which remained in SW. It is, however, assumed that fish in these experimental groups were similarly infested earlier in the experiment.

142

SUMBAWS

Final Report

Table 1 Stage and frequency of L. salmonis at sampling. Infested fish (n = 20) were sampled at day 3, 7, 14, 21 & 28 dpi. Stages are designated as follows: C1, first and second chalimus stages combined; C3, third and fourth chalimus stages combined; PAM1, first preadult male; PAF1, first preadult female; PAM2, second preadult male; PAF2, second preadult female; AM, adult male; AF, adult female; GF, gravid female. dpi 7 14 21 28 35

Cop 9%

C1 91% 1%

C3 99% 2%

Stage and Frequency PAM1 PAM2 PAF1 PAF2

11%

35% 1%

50% 4%

3% 27%

AM

AF

GF

50% 50%

18% 19%

30%

The zonal distribution of L. salmonis on the body surface of sampled fish is shown in Table 2. At 7 and 14 dpi, when almost all of the L. salmonis were at the chalimus stages (see Table 1), the majority of lice were found attached to the fins. Once L. salmonis attained the mobile stage (21-35 dpi), the majority were found on the body surface, but a substantial proportion was also found dislodged in the anaesthetic solution.

Group-specific models of average sea lice numbers across time which were used to generate model predictions for each physiological marker are shown in Table 3. It is clear that the numbers of lice 路 fish-1 were somewhat higher than the original target loading of ~50 lice 路 fish-1, particularly at 14 dpi.

Table 2 Frequency distribution (%) of L. salmonis on the body surface of sampled fish. % of total number of L. salmonis observed on each body zone of the experimental fish at each time point. See Figure 1 for an explanation of the different body zones. The stage of lice at each sampling point is shown in Table 1. dpi 7 14 21 28 35

A 3.3 5.2 11.8 9.7 18.7

Body Surface B C D 0.7 3.6 2.8 1.5 6.2 4.6 17.5 20.4 6.0 14.7 24.4 1.3 17.5 21.6 10.5

E 2.5 4.6 12.5 9.7 13.5

F 27.0 28.1 1.7 0.8 0.0

G 14.7 16.4 0.5 0.0 0.0

Fins H 11.3 7.8 0.3 0.0 0.0

I 7.3 5.4 0.2 0.0 0.0

J 13.9 15.7 1.5 0.0 0.0

Gills K 5.7 3.2 1.5 0.0 0.0

Anaesthetic L 7.1 1.3 26.3 39.5 18.1

Table 3 Predictions of the mean intensity of L. salmonis in the two infested groups.

Group

SW Infested

SW-FW Infested

dpi 7 14 21 28 35 7 14 21 28 35

Mean L. salmonis intensity 51.55 100.43 76.89 64.93 46.53 70.52 121.71 62.97 11.09 1.65

Mortalities There was a single fish mortality in the SW control group and 6 mortalities in the SW-FW control group.

There were 13 mortalities in the SW-FW infested group, 9 of which occurred prior to return to FW at 19 dpi.

There were 46 mortalities in the SW infested group and mortalities increased rapidly once L. salmonis reached the mobile stages (17 dpi). It was noted that from 15 dpi onwards, infested fish showed a marked decrease in feeding. Whilst the feeding activity of the fish in the infested SW-FW tanks recovered somewhat 143

SUMBAWS

Final Report

following return to FW at 19 dpi, food intake in both groups remained lower than the control groups throughout the remainder of the experiment.

Modelling A three-way interaction between sea lice numbers, dpi and the infested SW-FW group was retained for the AIC model for plasma chloride. Due to the difficulties of interpreting a high order interaction, the BIC (which incurs a larger penalty for each parameter) was used for comparison. In this case a three-way interaction and two of its three associated two-way interactions were replaced by an interaction between the infested SW group and time. Because the reduction in predictive power for the BIC model was only marginal (0.503 compared to 0.475), and the prediction plots were virtually identical, the BIC model was chosen in the interests of parsimony. No significant differences were observed between groups at 0, 7 or 14 dpi for plasma chloride concentration (Figure 2). After 19 dpi, when the lice had developed to the mobile stages (Table 1) plasma chloride in the SW lice-infested group was significantly elevated above all over groups (21-35 dpi), whereas no significant difference was observed between the SW-FW infested and control groups.

Plasma Chloride (mmol.l-1)

200 Control (SW) Control (SW - FW) Infested (SW) Infested (SW - FW)

150

100

50

0 0

7

14

21

28

35

Days Post Infection Figure 2 Predicted mean values with 95% confidence intervals for plasma chloride. Control (SW) – solid black line and ○, Control (SW-FW) – broken line and ●, Infested (SW) – broken line and ∆, Infested (SWFW) – dotted line and ▲. The vertical grey line represents the point at which two groups were returned to FW (19 dpi).Data are slightly offset for clarity.

A similar pattern of significance to that observed for plasma chloride also was observed for plasma lactate (Figure 3). However, at 14 dpi, plasma lactate was significantly elevated above control values in the SW-FW infested group. At 21 dpi, plasma lactate was significantly higher in both infested groups when compared to the controls. At 28 and 35 dpi plasma lactate in the SW lice-infested group was significantly elevated above both the SW control group and the SW-FW infested group (21 dpi), whereas no significant difference was observed between the SW-FW infested and control groups. There were no significant differences in plasma glucose concentration between groups at any timepoint, although there was some variation in glucose concentration over time with the highest values in all groups recorded at 21 dpi (Table 4).

144

SUMBAWS

Final Report

Plasma Lactate (mmol.l-1)

4 Control (SW) Control (SW - FW) Infested (SW) Infested (SW - FW)

3

2

1

0 0

7

14

21

28

35

Days Post Infection Figure 3 Predicted mean values with 95% confidence intervals for plasma lactate. Details as in Figure 2.

Table 4 Predicted mean values with 95% confidence intervals for plasma glucose concentration. dpi 0 7 14 21 28 35

SW Control 3.50 ± 0.33 3.92 ± 0.42 5.69 ± 0.50 6.58 ± 0.63 5.20 ± 0.57 4.31 ± 0.48

SW Infested 3.50 ± 0.33 3.50 ± 0.35 4.50 ± 0.70 6.40 ± 0.73 3.39 ± 1.01 3.87 ± 1.06

SW-FW Control 3.50 ± 0.33 3.92 ± 0.42 5.69 ± 0.50 6.58 ± 0.63 5.20 ± 0.57 4.31 ± 0.48

SW-FW Infested 3.48 ± 0.43 4.10 ± 0.48 6.04 ± 1.03 7.66 ± 0.87 4.95 ± 0.60 4.29 ± 0.56

Liver glycogen concentration was extremely low at 0 dpi in all groups, but had increased significantly for all groups by 7 dpi (Figure 4). There were no significant differences between any groups at 0 or 7 dpi. At 21 and 28 dpi, when the sea lice were at the mobile stages (Table 1), liver glycogen was significantly lower in both infested groups when compared with their respective controls. However, at 28 dpi liver glycogen in the SW-infested group was significantly lower than in the SW-FW infested group. At 35 dpi, a significant difference between the SW control and SW infested groups remained, but there was no significant difference between the SW-FW control and SW-FW infested groups at this time point.

Given the importance of plasma cortisol as a stress indicator, this marker was examined over time in further detail. Specifically, a separate model with potentially different covariates was fitted for each time point (Figure 5). No significant differences were observed for plasma cortisol until 14 dpi, when both infested groups were significantly elevated above their respective controls. At 21 dpi, plasma cortisol in the SW-infected group was significantly elevated above the mean cortisol concentration for all other fish. Data for fish sampled at 28 and 35 dpi were pooled due to small sample sizes; only 7 fish were sampled from the infested SW group at 28 dpi and just 6 from this group at 35 dpi. At this point mean cortisol concentration was significantly higher in the SW-infested group than mean cortisol concentration for all other fish.

145

SUMBAWS

Final Report

Liver Glycogen (mg.g-1 wet weight tissue)

100 Control (SW) Control (SW - FW) Infested (SW) Infested (SW - FW)

80

60

40

20

0 0

7

14

21

28

35

Days Post Infection Figure 4 Predicted mean values with 95% confidence intervals for liver glycogen. Details as in Figure 2.

300 Control (SW) Control (SW - FW) Infested (SW) Infested (SW - FW)

Plasma Cortisol (ng.ml-1)

250

200

150

100

50

0 0

7

14

21

28

35

Days Post Infection Figure 5 Predicted mean values with 95% confidence intervals for plasma cortisol. Details as in Figure 2.

146

SUMBAWS

Final Report

Discussion We were able to successfully acclimate wild sea trout smolts to tank conditions in a relatively short period of time. This was achieved by careful optimisation of flow rate, coupled with an intensive period of hand-feeding, particularly at the start of the experiment, and fish in all experimental tanks were feeding well within 1 wk of capture. However, from 15 dpi onwards, infested fish in both groups showed a marked decrease in feeding. Whilst the feeding activity of the fish in the infested SW-FW tanks recovered somewhat following return to FW, food intake in both groups remained lower than the control groups throughout the remainder of the experiment.

Sea lice development time, infestation success and the frequency distribution of sea lice on the body surface were similar to those previously described for laboratory infestation of hatchery-reared sea trout in northern Norway (Bjørn & Finstad 1998). A small number of copepodid and chalimus larvae were found in the gills but no adult or preadult lice were found there.

Although infestation intensities were rather higher than targeted, it should be noted that infestation intensity was only quantified after sampling had occurred and could not be altered once the infestation had started. The four infested tanks received equivalent numbers of sea lice according to the same protocol, but variability in infestation levels are an inherent problem with this type of experiment. All fish received a sea lice infestation which was considerably higher than the identified sublethal threshold infestation intensity of 13 L. salmonis ¡ fish-1, above which significant physiological effects on the host fish were to be expected (Scottish sea trout smolts; mean weight, 37 g) (Wells et al. 2006), and infestation intensity achieved allowed the assessment of the effects of premature return. Indeed, there were a number of mortalities arising from sea lice infestation, particularly in the infested SW group. As expected, following transfer back to FW, the mean infestation intensity was significantly reduced, showing that the majority of sea lice were unable to survive transfer to FW for periods of more than a few days. This in itself suggests that premature return to FW may be of benefit to heavily infested sea trout post-smolts, as also indicated by the markedly higher mortality of infested fish maintained in SW again indicating a survivorship benefit of premature return to FW. It should be noted however, that there may be further consequences of premature return over longer time periods which could not be accounted for by the present experimental design. In particular, secondary fungal or microbial infection was noted on a number of previously sea lice-infested fish following return to FW. Whilst these infections seemingly were not severe, and had no major physiological consequences over the experimental period, it is possible that over a more prolonged period such (freshwater) infections may prove deleterious to individual fish.

Mean plasma chloride levels remained stable throughout the experimental period in the control SW group. No significant differences were observed between the control and infested groups until L. salmonis had reached the mobile stages (21 dpi), in agreement with our previous study on wild sea trout smolts in Scotland (Wells et al. 2006). Plasma chloride continued to increase with time for infested fish in SW but was not significantly different from control values in the infested SW-FW group. Increases in plasma chloride may have resulted from the increased feeding activity of late chalimus and mobile preadult/adult lice and their associated skin damage (Bjørn & Finstad 1998; Pike 1989). Disturbance of water balance and ion homeostasis is one of the most characteristic aspects of stress in fish, due mainly to the intimate relationship between the body fluids in the gills and the ambient water (Wendelaar Bonga 1997). The disturbance of hydromineral balance, which is mediated by cortisol, may occur due to marked structural changes in the gills which are themselves characteristic of stress. These results therefore suggest that premature return to FW can alleviate the marked osmoregulatory disturbances observed in infested fish in SW, and that this behaviour may be of physiological benefit to post-smolt sea trout.

Hyperglycemia (elevation of plasma glucose) and elevated plasma lactate levels have been reported following both acute and chronic stress, and are seen as useful stress indicators in fish (Barton 2000; Farbridge & Leatherland 1992; Iversen et al. 1998; Wendelaar Bonga 1997). Anaerobic conditions caused by stress results in the breakdown of muscle glycogen to lactate, with some of this lactate being released into the bloodstream (Wood 1991). However, it is important to note that liver glycogen, plasma glucose and plasma lactate, are influenced also by the metabolic status of the animal. In the present study, the experimental animals were intercepted during their natural migration to sea following a non-feeding period of several months over winter. Following this period of natural fasting, experimental fish were fed to satiation and hence increases in metabolic measures are to be expected, and are not necessarily linked to stress.

Plasma lactate concentration in the SW infested group was elevated above the controls by 21 dpi, and above all other groups by 28 dpi. Increases in plasma lactate concentration for infested fish may reflect an increased energy demand to the animal, resulting in anaerobic breakdown of muscle glycogen to lactate, 147

SUMBAWS

Final Report

although it was noted that there was no associated rise in plasma glucose concentration in infested groups.

Plasma lactate also was elevated above control values at 14 and 21 dpi in the SW-FW infested group, but returned to control values by 28 dpi. The recovery of plasma lactate concentration to control values following return to FW (after 19 dpi) may itself suggest a benefit of premature return to FW for infested fish.

The enforced period of natural fasting immediately prior to the experiment is reflected in the initial liver glycogen levels at the start of the experiment (Figure 4), which were extremely low (2.2-3.4 mg.g-1 wet weight tissue). Liver glycogen content increased rapidly in all groups, probably as a result of experimental feeding, but stabilised in the controls by 14 dpi. At 21 dpi, once the sea lice had developed to the mobile preadult stages, liver glycogen content was significantly reduced in the two infested groups. Liver glycogen content remained depressed in the SW infested group, but following return to FW there was evidence of some recovery in the SW-FW infested group; by 35 dpi there was no significant difference in liver glycogen content between the SW-FW infested group and the respective control group.

Reductions in liver glycogen content of infested fish may reflect an increased energy demand to the host animal, either by increased activity in terms of flashing behaviour and leaping, or due to increased metabolic demands, for example in maintaining osmoregulatory integrity following physical damage. However, the decrease in liver glycogen observed in the present study is contrary to the current consensus for fish liver â&#x20AC;&#x201C; that cortisol (and therefore stress) significantly enhances the rate of gluconeogenesis (Mommsen et al.

1999). It is possible therefore that the observed decrease in liver glycogen is not a direct result of cortisolinduced stress, but rather is an indirect consequence of the reduction in feeding observed for both groups of infested fish, once the sea lice developed to the preadult and adult stages. An additional effect of cortisol is however, the suppression of appetite (Mommsen et al. 1999). Moreover, fish which were returned to FW had the same access to food as they previously had in SW and liver glycogen content in previously infested fish which were returned to FW appeared to recover to some extent. In contrast, in the wild it is likely that surviving, prematurely returning sea trout would in addition have to face reduced food availability in the freshwater environment, perhaps compromising liver glycogen stores further.

An elevation in plasma cortisol is the most widely used indicator of stress in fish (Wendelaar Bonga 1997). Cortisol may cause structural changes in the gills, such as necrosis of the branchial epithelial cells or epithelial lifting, which may have pronounced effects on the major functions of the gill, often leading to disturbance of water and ion homeostasis (Wendelaar Bonga 1997). There was a significant increase in plasma cortisol in both infested groups which was apparent at 14 dpi. This differed from our previous study of wild sea trout smolts in Scotland, whereby significant increases in plasma cortisol were not observed until L. salmonis had moulted to the mobile stages (Wells et al. 2006). By the end of the present experiment only the SW infested group had a plasma cortisol concentration significantly above the control levels. This elevation in plasma cortisol is consistent with the well-characterised stress response (Wendelaar Bonga 1997), and the lack of a chronic increase in plasma cortisol in the SW-FW group also suggests that premature return to FW is a benefit to sea lice infested fish.

To our knowledge, this is the first attempt to measure the effects of premature return in laboratory studies, using a comprehensive suite of osmoregulatory, metabolic and stress measures. Return to FW resulted in the loss of L. salmonis from the host body surface within 1 wk of transfer, and mortality in this group was considerably lower than in infested fish. Significant increases in plasma chloride, lactate and cortisol and a significant decrease in liver glycogen content all were apparent in infested fish which were maintained in SW, but these physiological effects were alleviated in fish which returned prematurely to FW. It is important to note that heavily sea lice-infested wild fish, particularly those individuals with significant lesions resulting from sea lice feeding would be prone to bacterial/fungal infection following their return to FW. However, the overall results of the present study indicate that there may be clear physiological benefits of premature return to FW for sea lice-infested sea trout post-smolts.

References Akaike, H. 1973. Information theory and an extension of the maximum likelihood principle In International Symposium on Information Theory. Edited by B. N. Petrov, and F. Csaki. Akademia Kiado, Budapest.

Anonymous. 1993. Report of the Sea Trout Working Group 1993. Page 127. Department of the Marine, Abbotstown.

Anonymous. 1995. Report of the Sea Trout Working Group 1994.254 pp.

148

SUMBAWS

Final Report

Barton, B. A. 2000. Salmonid fishes differ in their cortisol and glucose responses to handling and transport stress. N. Am. J. Aqualcult. 62:12-18.

Birkeland, K. 1996. Consequences of premature return by sea trout (Salmo trutta) infested with the salmon louse (Lepeophtheirus salmonis Kroyer): Migration, growth, and mortality. Can. J. Fish. Aquat. Sci. 53:2808-2813.

Birkeland, K., and Jakobsen, P. J. 1997. Salmon lice, Lepeophtheirus salmonis, infestation as a causal agent of premature return to rivers and estuaries by sea trout, Salmo trutta, juveniles. Environ. Biol. Fishes 49:129-137.

Bjørn, P. A., and Finstad, B. 1997. The physiological effects of salmon lice infection on sea trout post smolts. Nordic J. Freshw. Res. 73:60-72.

Bjørn, P. A., and Finstad, B. 1998. The development of salmon lice (Lepeophtheirus salmonis) on artificially infected post smolts of sea trout (Salmo trutta). Can. J. Zool.-Rev. Can. Zool. 76:970-977.

Bjørn, P. A., Finstad, B., and Kristoffersen, R. 2001. Salmon lice infection of wild sea trout and Arctic char in marine and freshwaters: the effects of salmon farms. Aquac. Res. 32:947-962.

Blackwell, B. G., Brown, M. L., and Willis, D. W. 2000. Relative weight (Wr) status and current use in fisheries assessment and management. Rev Fish Sci 8:1-44.

Brenden, T. O., Murphy, B. R., and Birch, J. B. 2003. Statistical properties of the relative weight (Wr) Index and and alternative procedure for testing Wr differences between groups. North Am. J. Fish Manage.

23:1136-1151.

Breusch, T. S., and Pagan, A. R. 1979. A simple test for heteroscedasticity and random coefficient variation. Econometrica 47:1287-1294.

Butler, J. R. A. 2002. Wild salmonids and sea louse infestations on the west coast of Scotland: sources of infection and implications for the management of marine salmon farms. Pest Manag. Sci. 58:595-608.

Dawson, L. H. J., Pike, A. W., Houlihan, D. F., and McVicar, A. H. 1998. Effects of salmon lice Lepeophtheirus salmonis on sea trout Salmo trutta at different times after seawater transfer. Dis. Aquat. Org. 33:179-186.

Farbridge, K. J., and Leatherland, J. F. 1992. Plasma growth hormone levels in fed and fasted rainbow trout (Oncorhynchus mykiss) are decreased following handling stress. Fish Physiol. Biochem. 10:67-73.

Finstad, B., Bjorn, P. A., and Nilsen, S. T. 1995. Survival of salmon lice, Lepeophtheirus salmonis Kroyer, on Artic charr, Salvelinus alpinus (L.), in fresh water. Aquac. Res. 26:791-795.

Fox, J. 1997. Applied regression, linear models and related methods. Sage Publications, Thousand Oaks, California.

Grimnes, A., and Jakobsen, P. J. 1996. The physiological effects of salmon lice infection on post- smolt of Atlantic salmon. J. Fish Biol. 48:1179-1194.

Hansen, L. P. J. J. J. 2000. Distribution and migration of Atlantic salmon, Salmo salar L., in the sea In The ocean life of Atlantic salmon. Environmental and biological factors influencing survival. Edited by D. Mills. Fishing News Books, Oxford.

Iversen, M., Finstad, B., and Nilssen, K. J. 1998. Recovery from loading and transport stress in Atlantic salmon (Salmo salar L.) smolts. Aquaculture 168:387-394.

Johnson, S. C., and Albright, L. J. 1991. The Developmental Stages of Lepeophtheirus Salmonis (Kroyer, 1837) (Copepoda, Caligidae). Can. J. Zool.-Rev. Can. Zool. 69:929-950.

Keppler, D., and Decker, K. 1974. Glycogen: determination with amyloglucosidase In Methods of Enzymatic analysis. Edited by H. U. Bergmeyer. Academic Press.

Klemetsen, A., Amundsen, P. A., Dempson, J. B., Jonsson, B., Jonsson, N., O'Connell, M. F., and Mortensen, E. 2003. Atlantic salmon Salmo salar L., brown trout Salmo trutta L. and Arctic charr Salvelinus alpinus (L.): a review of aspects of their life histories. Ecol. Freshw. Fish 12:1-59.

MacKenzie, K., Longshaw, M., Begg, G. S., and McVicar, A. H. 1998. Sea lice (Copepoda : Caligidae) on wild sea trout (Salmo trutta L.) in Scotland. ICES J. Mar. Sci. 55:151-162.

149

SUMBAWS

Final Report

McKibben, M. A., and Hay, D. W. 2004. Distributions of planktonic sea lice larvae Lepeophtheirus salmonis in the inter-tidal zone in Loch Torridon, Western Scotland in relation to salmon farm production cycles. Aquac. Res. 35:742-750.

McLean, P. H., Smith, G. W., and Wilson, M. J. 1990. Residence Time of the Sea Louse, Lepeophtheirus salmonis K, on Atlantic Salmon, Salmo salar L, after Immersion in Fresh-Water. J. Fish Biol. 37:311-314.

Mommsen, T. P., Vijayan, M. M., and Moon, T. W. 1999. Cortisol in teleosts: dynamics, mechanisms of action, and metabolic regulation. Rev. Fish. Biol. Fish. 9:211-268.

Neff, B. D., and Cargnelli, L. M. 2004. Relationships between condition factors, parasite load and paternity in bluegill sunfish, Lepomis macrochirus. Environ. Biol. Fishes 71:297-304.

Northcott, S. J., and Walker, A. F. 1996. Farming salmon, saving sea trout; a cool look at a hot issue In Aquaculture and sea lochs. Edited by K. Black. Scottish Association for Marine Science, Oban.

Penston, M. J., McKibben, M. A., Hay, D. W., and Gillibrand, P. A. 2004. Observations on open-water densities of sea lice larvae in Loch Shieldaig, Western Scotland. Aquac. Res. 35:793-805.

Pike, A. W. 1989. Sea Lice - Major Pathogens of Farmed Atlantic Salmon. Parasitol. Today 5:291-297.

Pike, A. W., and Wadsworth, S. L. 1999. Sealice on salmonids: their biology and control. Advances in Parasitology 44:233-337.

Royston, P. 1982. An extension of Shapiro and Wilk's W Test for normality to large samples. Applied Statistics 31:115-124.

Schram, T. A. 1993. Supplimentary descriptions of the developmental stages of Lepeophtheirus salmonis (Kroyer, 1837) (Copepoda: Caligidae) In Pathogens of Wild and Farmed Fish: Sea Lice. Edited by G. A. Boxshall, and D. Defaye. Ellis Horwood, Chichester.

Sutton, S. G., Bult, T. P., and Haedrich, R. L. 2000. Relationships among fat weight, body weight, water weight, and condition factors in wild Atlantic salmon parr. Trans. Am. Fish. Soc. 129:527-538.

Tully, O., Poole, W. R., and Whelan, K. F. 1993a. Infestation parameters for Lepeophtheirus salmonis (Kroyer) (Copepoda: Caligidae) parasitic on sea trout, Salmo trutta L., off the west coast of Ireland during 1990 and 1991. Aquaculture and Fisheries Management 24:545-555.

Tully, O., Poole, W. R., Whelan, K. F., and Merigoux, S. 1993b. Parameters and possible causes of Lepeophtheirus salmonis (Krøyer) infesting sea trout (Salmo trutta L.) off the west coast of Ireland In Pathogens of wild and farmed fish. Sea lice. Edited by G. A. Boxshall, and D. Defaye. Ellis Horwood, Chichester.

Tully, O., and Whelan, K. F. 1993. Production of Nauplii of Lepeophtheirus salmonis (Kroyer) (Copepoda, Caligidae) from Farmed and Wild Salmon and Its Relation to the Infestation of Wild Sea Trout (Salmo trutta L) Off the West-Coast of Ireland in 1991. Fish Res. 17:187-200. van Anholt, R. D., Spanings, T., Koven, W., and Wendelaar Bonga, S. E. 2003. Effects of acetylsalicylic acid treatment on thyroid hormones, prolactins, and the stress response of tilapia (Oreochromis mossambicus). Am. J. Physiol.-Regul. Integr. Comp. Physiol. 285:R1098-R1106.

Wagner, G. N., McKinley, R. S., Bjorn, P. A., and Finstad, B. 2004. Short-term freshwater exposure benefits sea lice-infected Atlantic salmon. J. Fish Biol. 64:1593-1604.

Wells, A., Grierson, C. E., MacKenzie, M., Russon, I. J., Reinardy, H., Middlemiss, C., Bjørn, P. A., Finstad, B., Wendelaar-Bonga, S. E., Todd, C. D., and Hazon, N. 2006. The physiological effects of simultaneous abrupt seawater entry and sea lice (Lepeophtheirus salmonis Krøyer) infestation of wild sea trout (Salmo trutta L.) smolts. Can. J. Fish. Aquat. Sci. Submitted Wendelaar Bonga, S. E. 1997. The stress response in fish. Physiol. Rev. 77:591-625.

Wood, C. M. 1991. Acid-base and ion balance, metabolism, and their interactions, after exhaustive excercise in fish. J. Exp. Biol. 160:285-308.

150

SUMBAWS

Final Report

Workpackage 5 Lead Partner: USTAN Participating Partners: USTAN, NINA, NIFA, KUN OBJECTIVES • • • • •

Determine the effect of increasing sea lice infestation on osmoregulatory physiology including drinking rate, branchial sodium flux, Na+-K+ATPase activity in osmoregulatory epithelia and plasma ion composition.

Assess the effect of increasing sea lice infestation on the plasma levels of key hormones controlling osmoregulation.

Determine the degree of stress levels in fish with increasing infestation of sea lice.

Assess the effects of sea lice infestation on the number and distribution of branchial chloride cells.

Establish the minimum level of infestation that significantly affect normal physiology and before physical damage is apparent on the fish.

DELIVERABLES • • • • •

Determining the effect of increasing sea lice infestation on key osmoregulatory parameters.

Effects of lice infestation on hormonal control mechanisms for salt/water balance in wild sea trout smolts.

Assessment of stress caused by increasing sea lice infestation as measured by established plasma markers.

Epithelial cell type and distribution and cell turnover in uninfested fish and after sea lice infestation.

Elucidation of the interaction of stress caused by sea lice infestation and the onset of osmoregulatory failure.

EXECUTIVE SUMMARY The overall aims of this workpackage were to assess and quantify the effect of sea lice infestation on osmoregulatory physiology, stress physiology and to determine the minimum levels of infestation that significantly affect normal physiology. In order to achieve these aims we performed a series of laboratory based studies, in which a suite of parameters involved in stress and osmoregulatory physiology were assessed. In addition, we employed for the first time a novel application of piecewise linear models in order to determine threshold lice levels above which the host suffers significant physiological stress.

Previous experimental reports of the physiological effects of Lepeophtheirus salmonis infestation on sea trout are restricted to seawater (SW) acclimated, hatchery-reared fish. These fish are much larger than wild sea trout smolts and the fins (primary sites for first attachment of juvenile sea lice) often are stunted, raising questions about the validity of extrapolating experimental outcomes to wild fish. In addition, acclimation to SW itself is also a stressful process and simultaneous challenge from ectoparasite infestation during this period will constitute an environmentally valid significant additional stressor. In Paper 1, the physiological effects of simultaneous abrupt seawater entry and sea lice infestation were investigated in wild sea trout smolts. In a series of laboratory experiments designed to mimic environmentally realistic conditions, we investigated the time course of physiological responses arising from varying intensities of infestation with the sea louse, L. salmonis.

Analysis of osmoregulatory, metabolic and stress markers allowed the determination of the louseinduced onset of stress. Significant lice effects, consistent across all measured parameters, were not apparent until L. salmonis developed to the ‘mobile’ preadult and adult stages. Preadult L. salmonis caused significant increases in plasma chloride, osmolality, glucose, lactate, cortisol and a significant reduction in haematocrit.

Piecewise linear statistical approaches allowed the determination of abrupt changes in these physiological markers, attributable to the intensity of L. salmonis infestation on individual fish, and identification of an 151

SUMBAWS

Final Report

overall sublethal threshold lice burden. 13 L. salmonis · fish-1 was a consistent breakpoint across several physiological measures. Identification of this sublethal threshold burden is an essential prerequisite in formulating effective wild fisheries management and stock conservation policy, and these results will provide a sound basis for continued policy development. We would recommend that the conservative and precautionary approach would be for wild fishery managers to adopt a critical level of 13 mobile L. salmonis · fish-1 for juvenile sea trout primarily in their first ‘sea’ year. This level would provide a clear indication of the proportion of sea trout within a population which are subject to physiological stress and potential death from sea lice infestation.

The epithelial response of the skin and gills of sea trout smolts arising from varying intensities of infestation with L. salmonis were investigated in Paper 2, in a parallel study to Paper 1. Specifically, changes in cell type, distribution and morphology as well as changes in general epithelial integrity were investigated on samples taken from the same fish used in Paper 1. The fish epithelium forms the first barrier between the external and internal environment and demonstrates characteristic changes in fish exposed to stressors. Disruption of the gill has deleterious effects on the fish because the gills are the primary site for ion regulation, gaseous exchange and excretion of nitrogenous waste. The majority of research in this area has therefore been focused on the gill, but recent studies have suggested that the assessment of integrity of the skin is also an important area for study.

Changes in the epithelia of fish take longer to develop than do hormonal changes and are therefore a reliable indicator of prolonged chronic stress. In this study, differences were not observed until 14 days post infection (dpi) coinciding with the first appearance of mobile lice and the associated increased feeding activity of lice at this stage (Paper 1). No differences were observed between control fish or fish infested with low lice intensity (~10 lice · fish-1) when skin and gill samples were analysed using scanning electron microscopy (SEM). However, at medium lice intensities (~30 lice · fish-1), above the predicted threshold level determined in Paper 1, deterioration of the skin and gill epithelium was clearly evident from 14 dpi onwards. In addition, increases in the number of chloride cells and proliferating cells (suggesting an increased cell turnover) as well as increases in the number of migrating mucus cells and chloride cells were observed in the medium lice intensity group from 14 dpi onwards. These data are consistent with the results of Paper 1, with significant lice-induced stress at medium levels of infestation once the lice reach the mobile stages.

Previous workpackages have highlighted the clear behavioural, ecological and demographic differences between Atlantic salmon and sea trout, and these differences highlight the pitfalls of extrapolating research from one species to the other. In order to avoid these problems we conducted a further series of laboratory experiments designed to determine a threshold level of sea lice infestation for farmed Atlantic salmon (Paper 3). This work was undertaken following discussion with our SEARCH clustering partners and representatives of the aquaculture industry. Similarly to the results presented in Paper 1, for sea trout smolts, significant decreases in salmon haematocrit and increases in plasma concentrations of chloride, glucose lactate and cortisol and plasma osmolality were observed once the sea lice had reached the mobile stages.

Piecewise linear statistical approaches again allowed the determination of abrupt changes in these physiological markers, attributable to the intensity of L. salmonis infestation on individual fish, and identification of overall threshold lice burden 20 mobile L. salmonis · fish-1 was a consistent breakpoint across several physiological measures indicating the onset of physiological stress above this lice intensity.

This important information will be made available to the aquaculture industry.

152

SUMBAWS

Final Report

CONCLUSIONS Due to the logistics of completing the complex and time-consuming laboratory experiments conducted in WP 5, and following a series of preliminary experiments, we decided to reprioritise the sampling regime in order to maximise the information gathered during the available time. Therefore, branchial sodium fluxes were not measured and drinking rate was measured in a separate series of experiments, as detailed in the Year 3 report. During our preliminary investigations it became clear that we would not be able to collect sufficient blood from wild sea trout smolts to allow the measurement of all the key hormones involved in osmoregulation. We therefore focused on cortisol and MSH as hormonal markers. Despite these minor changes to the initial experimental workplan we have met all the objectives and deliverables from the original research contract, and in addition, we have conducted a further series of laboratory experiments on farmed salmon smolts. Perhaps most significantly, we have, for the first time objectively identified a consistent threshold lice burden of 13 L. salmonis 路 fish-1 for wild sea trout smolts. Identification of this threshold is an essential prerequisite in formulating effective wild fisheries management and stock conservation policy and these results will provide a sound basis for continued policy development.

153

SUMBAWS

Final Report

Paper 1: THE PHYSIOLOGICAL EFFECTS OF SIMULTANEOUS ABRUPT SEAWATER ENTRY AND SEA LICE (LEPEOPHTHEIRUS SALMONIS KRØYER) INFESTATION OF WILD SEA TROUT (SALMO TRUTTA L.) SMOLTS A. Wells, C. E. Grierson, M. MacKenzie, I. J. Russon, H. Reinardy, C. Middlemiss, P. A. Bjørn, B. Finstad, S. E. Wendelaar Bonga, C. D. Todd and N. Hazon.

(Submitted to Canadian Journal of Fisheries and Aquatic Sciences) Abstract For wild sea trout (Salmo trutta L.) smolts, the physiological consequences of abrupt transfer to seawater and simultaneous challenge with copepodid larvae of the sea louse, Lepeophtheirus salmonis Krøyer were investigated in the laboratory. Analysis of osmoregulatory, metabolic and stress markers allowed the derivation of a sublethal threshold burden of L. salmonis, above which the host suffers significant physiological stress. Significant lice effects, consistent across all measured markers, were not apparent until L. salmonis developed to the ‘mobile’ preadult and adult stages. Preadult L. salmonis caused significant increases in plasma chloride, osmolality, glucose, lactate, cortisol and a significant reduction in haematocrit.

Piecewise linear statistical approaches allowed the determination of abrupt changes in these physiological markers, attributable to the intensity of L. salmonis infestation on individual fish, and identification of overall threshold lice burdens. 13 L. salmonis · fish-1 (weight range 19-70 g) was a consistent breakpoint across several physiological measures. This information will provide a valuable, objectively derived tool to aid in the formulation of effective wild fisheries management policy concerning S. trutta conservation.

Introduction Sea trout (Salmo trutta L.) have an anadromous life history pattern and are exposed to a variety of environmental and biological challenges, including parasitic infection. Sea trout numbers in NW Europe have fallen since the 1950s (Anonymous 2004), but during the late 1980s and early 1990s the first reports of collapses in certain wild sea trout stocks, and the possible association of these collapses with sea lice (Lepeophtheirus salmonis, Krøyer), arose in Ireland (Tully et al. 1993), Scotland (Northcott and Walker 1996), and Norway (Birkeland 1996). Unlike salmon, which undertake extensive oceanic migrations (Hansen and Jacobsen 2000), sea trout remain in nearshore waters throughout their marine residence (Klemetsen et al. 2003) and stock declines appear to be attributable primarily to increased marine mortality (Anonymous 1993, 1995; Northcott and Walker 1996). Caligid sea lice infestations can debilitate or kill the host fish (Pike and Wadsworth 1999), and there is very strong circumstantial evidence that sea lice derived from aquaculture pens might be a major contributory factor to marine mortality rates (Butler 2002; McKibben and Hay 2004; Penston et al. 2004).

Copepod sea lice, L. salmonis and Caligus elongatus (Nordmann), are the primary marine ectoparasites of farmed and wild salmonids in the North Atlantic. Sea lice infestation of the host fish is characterised by stress and osmoregulatory imbalance, which occur once the parasites moult to the ‘mobile’ preadult and adult stages (Bjørn and Finstad 1997; Dawson et al. 1998; Finstad et al. 2000; Grimnes and Jakobsen 1996).

Feeding activity by preadult/adult sea lice can result in increased skin damage (Bjørn and Finstad 1998; Nolan et al. 1999; Pike and Wadsworth 1999), which can be fatal to hatchery-reared sea trout and salmon (Bjørn and Finstad 1997; Dawson et al. 1998; Grimnes and Jakobsen 1996). Infestation of sea trout smolts with large numbers of the sessile chalimus larval stages leads to increased plasma levels of the stress hormone cortisol (Bjørn and Finstad 1997). Cortisol causes a range of effects, including hydromineral imbalance, changes in intermediary metabolism, reduction in the immune response and reduced capacity to tolerate subsequent or additional stressors (Wendelaar Bonga 1997).

Experimental studies on Atlantic salmon and Arctic char juveniles indicate that 30 to 50 attached chalimus larvae may cause the death of a 40 g post-smolt (Grimnes et al. 1996). In the case of sea trout, experimental data indicate that ~50 preadult and adult lice may cause the death of a 60 g post-smolt (Bjørn and Finstad 1997). However, from the perspective of conservation and management of wild and farmed salmonid stocks, the objective establishment of a threshold infestation intensity, above which there are significant but sublethal physiological effects on the host, is more important than the determination of lethal lice levels. Here we report a threshold infestation intensity for wild sea trout infested by L. salmonis, which

154

SUMBAWS

Final Report

is derived from appropriate statistical models, and which is consistent across a range of physiological measures.

Previous experimental reports of the physiological effects of L. salmonis infestation on sea trout are restricted to seawater (SW) acclimated, hatchery-reared fish. These fish are much larger than wild sea trout smolts, and the fins (primary sites for first attachment of the infective copepodid stage of L. salmonis (Finstad et al. 1994)) often are stunted, raising questions about the validity of extrapolating experimental outcomes to wild fish. In addition, acclimation to seawater alone is a stressful process and simultaneous challenge from ectoparasite infestation during this period will constitute a significant additional stressor.

Here, in laboratory experiments designed to mimic environmentally realistic circumstances, we investigated the time-course of physiological responses, and identify sublethal threshold ectoparasite intensities for wild smolts transferred abruptly to SW and simultaneously challenged with L. salmonis.

Materials and Methods Wild sea trout smolts (mean weight 37.3g; range 19.3-69.9 g) were trapped and netted from the River North Esk, East Scotland, from 18-24 May 2004 and transported (90 min) to the Gatty Marine Laboratory in a smolt tank containing constantly oxygenated freshwater (FW). On arrival, smolts were randomly split into duplicated experimental groups and held in brackish water (468 ± 9 mOsm.kg-1) in eight, 300 l aquarium tanks under natural photoperiod (18 h light: 6 h dark). Smolts were maintained in this manner for 7-13 d prior to experimentation. All fish were fed on chopped squid (Ammodytes Company Ltd) at a ration of ~2% body weight.d-1. The sea trout smolts adapted to tank conditions in a relatively short period of time following an intensive period of hand-feeding, particularly at the start of the experiment, to the extent that fish in all experimental tanks were feeding well on chopped squid within 1 week of capture. We continued to feed chopped squid throughout the experiment because the fish refused commercial pellet even when mixed with squid. Eighteen hours prior to the L. salmonis infestation, the flow of FW was stopped and the SW flow was increased accordingly. The salinity of the tanks reached full SW (972 ± 3 mOsm kg-1) within 3 h and the mean aquarium temperature was 13.6°C. Ovigerous female L. salmonis were collected from Atlantic salmon at a commercial West Scotland fish farm and transported to the Gatty Marine Laboratory in cooled SW. Egg strings were removed from the lice and evenly distributed between six, 2 l conical flasks containing filtered SW (10°C). Flasks were continuously aerated to ensure that egg strings remained in constant motion. Nauplii began hatching within 24 h and were collected on a 40 µm filter and transferred to a 20 l plastic aquarium containing filtered, UVsterilised, through-flowing SW. Larvae developed through the two naupliar moults to the infective copepodid stage. Small numbers of larvae were sampled daily and microscopically examined to monitor the rate of larval development. On the day of infestation, 2 to 4 d old copepodid larvae were concentrated in 1 l of SW by filtering through a 40 µm filter. The total number of copepodids was estimated from five replicate subsample counts (mean 22 lice ml-1: range 16-24).

At infestation, the volume of water in the experimental tanks was drained to one third (100 l) and the water supply turned off. The water was continuously aerated during the infestation to maintain a circulating current in the tanks. Copepodids were added in appropriate numbers to achieve the desired infestation level.

Our aim was to achieve three duplicated infestation levels: low (~10 lice · fish-1), medium (~30 lice · fish-1) and high (~50 lice · fish-1). The duplicate control tanks (zero lice) were treated in an identical manner, but no copepodid larvae were added. Infestation progressed for 2 h in total darkness, following which the experimental tanks were refilled and the 18 h light: 6 h dark photoperiod regime was reinstated.

Infested and control fish were sampled at 3, 7, 14, 21 and 28 days post infestation (dpi). Fish were individually netted from the experimental tanks, taking care to minimise disturbance to the remaining fish, and killed by overdose of 2-phenoxyethanol (2.5 ml l-1) using individual containers for each fish. On removal from the anaesthetic the spinal cord of each individual fish was severed using a scalpel blade. The anaesthetic was subsequently passed through a 40 µm filter in order to collect any dislodged L. salmonis.

Five fish from each duplicate tank (10 fish for each lice intensity) were removed at each sampling date and blood and tissues were collected within 15 min. Blood samples (0.5 ml) were taken from the caudal blood sinus immediately following sacrifice, added to 50 µl 15% ammonium EDTA and processed immediately (see below). The fish was then opened along the mid ventral line and the liver was removed and freeze clamped in liquid nitrogen to prevent glycogenolysis. The frozen liver was placed in a pre-weighed vial and stored at -80ºC until analysis. Fish carcasses were placed in individual plastic bags and stored at 4ºC or 20ºC for subsequent microscopic determination of L. salmonis numbers. All dissection was performed on a piece of paper towelling and this also was bagged and subsequently examined for dislodged L. salmonis.

The fork length and weight of all fish were recorded and expressed as the composite condition factor, 155

SUMBAWS

Final Report

Fulton’s K (weight in g/(length in cm)3 × 100). Ratio condition factors are sensitive to isometric variation and departures from the cube relationship, but the size range of experimental fish was narrow and Fulton’s K has been shown to have a high correlation with whole fish lipid content (Neff and Cargnelli 2004; Sutton et al. 2000).

Blood haematocrit was determined on fresh blood using a Compur microspin microhaematocrit centrifuge. Plasma was separated by centrifugation at 13,000 rpm for 1 min, and frozen at -80ºC until analysis. Plasma chloride was measured by colorimetric titration (Chloride analyser 925, Corning, UK).

Plasma osmolality was measured by freezing point depression (Roebling Osmometer, Camlab, UK). Plasma glucose concentration was determined in duplicate using a commercially available kit (Randox laboratories Ltd, Co Antrim, UK – ref: GL 2623), and absorbance was read at 490 nm. Plasma lactate concentration was determined in duplicate using commercially available lactate reagent (Randox laboratories Ltd, Co Antrim, UK – ref: LC 2389) and read at 540 nm. Both of the latter assays were modified for use in 96-well plates and absorbance was read on a plate reader (Biotech Instruments Inc, USA). Liver glycogen was assessed following hydrolysation by amyloglucosidase using the method of Keppler and Decker (1974). Plasma cortisol was measured by RIA, as described by van Anholt et al. (2003).

The body surface, fins and gills were separated analytically into a total of 11 zones as previously described (Bjørn and Finstad 1998) (Figure 1). These zones on all infected fish were examined for L. salmonis, and the frequencies of each developmental stage were recorded. L. salmonis at the sessile first/second chalimus stages and third/fourth chalimus stages were pooled into two groups, termed C1 and C3 respectively. Older (post-larval) mobile L. salmonis stages were categorised and their sex determined according to Johnson and Albright (1991) and Schram (1993).

Figure 1 The 11 zones on the body of the infected sea trout smolts examined for the presence of L. salmonis (Bjørn and Finstad 1998). A, head and operculum; B, anterior dorsal area of the body above the lateral line from the head to the posterior end of the dorsal fin; C, posterior dorsal area of the body above the lateral line from the posterior end of the dorsal fin to the caudal fin; D, anterior ventral area of the body below the lateral line from the head to the posterior end of the dorsal fin; E, posterior dorsal area of the body below the lateral line from the posterior end of the dorsal fin to the caudal fin; F, dorsal fin; G, pectoral fin; H, ventral fin; I, anal fin; J, caudal fin; K, gills. The anaesthetic solutions, plastic storage bags, and paper towelling on which dissection took place were also examined and are shown as zone L. Data Analysis Eight fish had lice values above 100 (maximum 232) and therefore were excluded from the analyses to avoid substantial gaps in the lice covariate range. The number of lice, fish length, fish weight, the condition of each fish and dpi were considered as candidate variables in each statistical model. Length and weight were too highly correlated to allow both to be considered for model selection; their variance inflation factors (Fox and Monette 1992) were 26.57 and 25.38 respectively. Consequently, only length was considered for selection since this variable returned slightly better fit results overall.

In order to incorporate the effects of lice, length, condition factor and time, the potentially nonlinear relationships between each variable and the physiological measure of interest were fitted using Generalized Additive Models (GAMs). GAMs were fitted to the data pooled across time to allow the detection of any significant effects due to lice, fish length, condition and time. In order to detect any significant tank or 156

SUMBAWS

Final Report

duplicate tank effects, tank and tank duplicate variables were also included in each model. The tank loading (i.e. the lice loading targeted for the various tanks) was also considered for selection in each model with a significant lice relationship to ensure that tank loading was no longer significant once the number of lice per fish was included. No tank or tank duplicate effects were found to be significant and so all data were pooled across tank loadings to obtain data across the lice range. GAM analysis was also carried out for each response at each time point to allow assessment of the way in which lice numbers and the other covariates change with time. This approach was considered appropriate because L. salmonis developed through their entire life cycle as the experiment progressed and distinct life cycle stages of L. salmonis were observed at each experimental time point (Table 1). This decision to model each stress indicator at each time point was supported also by the data; significant lice-time interactions were detected for all stress measures when the data were pooled across time.

In order to determine abrupt changes (putative thresholds) in the lice relationship for each physiological measure, a piecewise linear approach was used. This approach assumes linearity between breakpoints and is commonly used to identify changes in functional form and/or threshold values (Bernatchez and Wilson 1998; Molinari et al. 2001; Shea and Vecchione 2002), but this is the first application of the method to physiological studies of stress.

Observed lice values were used as candidate breakpoints in each model and the number and location of the breakpoints for the piecewise linear models were objectively identified using the AIC statistic (Akaike 1973). After the first breakpoint was identified and fixed in the model, the remaining set of unique lice values were considered for the second breakpoint. This forwards selection procedure was repeated until the AIC score no longer improved. The model was constrained so that breakpoints were separated by at least three unique lice values to ensure each linear component fitted between breakpoints was supported by data, in keeping with the recommendations in Wold (1974). This breakpoint selection process also was carried out while including fish length and condition factor in each model. In practice, GAM analysis at each time point revealed that nonlinear relationships were justified only for the lice covariate and so condition factor and length were included as linear terms in each model.

To ensure the complexity of a piecewise linear approach was suitable, the fit of each model (as judged by the AIC) was compared with more and less flexible alternatives. Specifically, each piecewise linear model was compared with the more restrictive standard linear term for lice, and two more flexible models which allow more flexible curves between the chosen breakpoints. B-splines were used to fit these models in R (R Development Core Team 2004), using a specified number of (internal) breakpoints. The AIC statistic formed the basis for these comparisons and all models were fitted assuming a normal error distribution.

Normality, independence and constant error variance were checked using Shapiro-Wilk, Durbin-Watson and Breusch-Pagan tests on the residuals (Breusch and Pagan 1979; Fox 1997; Royston 1982) and models found to exhibit non-constant error variance were fitted inside a generalized least squares framework. This framework naturally accommodates an increasing mean-variance relationship which is often encountered with biological data.

Results Development of L. salmonis L. salmonis development is summarised in Table 1. L. salmonis numbers were assessed on sampled fish at each experimental timepoint (3, 7, 14, 21 & 28 dpi). However, as mortalities also were counted for L. salmonis, we were able to observe development on intervening days and also record the first appearance of the various L. salmonis stages on the fish. Mobile, preadult L. salmonis were observed on fish from 10 dpi onwards. Prevalence of L. salmonis was 100% on each sampling occasion, with the exception of the low lice intensity group at 28 dpi. In that group, no L. salmonis were found on two fish, but it is assumed that these fish had been infected earlier in the experiment. Only copepodids and early chalimus stages (C1) were found at 3 dpi. By 7 dpi, 67% of the L. salmonis had developed to the late chalimus stages (C3). One week later, the majority of male L. salmonis were at the second preadult stage, whereas the majority of female L. salmonis had reached only the first preadult stage. At day 21, most male L. salmonis had reached the adult stage whereas females were either at the second preadult stage or the adult stage. By day 28 all L. salmonis were adults, 62% of which were gravid females.

157

SUMBAWS

Final Report

Table 1 Stage and frequency of major groups of L. salmonis at sampling. Infected fish (n = 30 at day 3, 7 and 14, n = 20 at day 21 and n = 10 at day 28) were sampled at day 3, 7, 14, 21 & 28 dpi. Stages are designated as follows: C1, first and second chalimus stages combined; C3, third and fourth chalimus stages combined; PAM1, first preadult male; PAF1, first preadult female; PAM2, second preadult male; PAF2, second preadult female; AM, adult male; AF, adult female; GF, gravid female. dpi

Cop 73%

3 7 14 21 28

C1 27% 33%

C3 67% 1%

Stage and Frequency PAM1 PAM2 PAF1 PAF2

6%

44%

49%

<1% 25%

AM

AF

48% 38%

27%

GF

62%

The frequency distribution of L. salmonis on the body surface of sampled fish is shown in Table 2. At 3 dpi, when 73% of the L. salmonis were at the copepodid stage (see Table 1), the majority (74%) of L. salmonis were found detached in the anaesthetic solution (zone L). At 7 dpi, the majority of the L. salmonis were observed attached to the fins, with 34% on the dorsal fin (zone F). Only 1% of the L. salmonis were detached from the host, and hence collected from the anaesthetic solution. Once L. salmonis attained the mobile stage (14-21 dpi), the majority were on the body surface, but a substantial proportion was also found dislodged in the anaesthetic solution. At 28 dpi, adult L. salmonis were observed only on the posterior zones C and E. Table 2 Frequency distribution (%) of L. salmonis on the body surface of sampled fish. Percent of total number of L. salmonis observed on each body zone of the experimental fish at each time point. See Figure 1 for the different body zones. The developmental stage of lice at each sampling point is shown in Table 1.

dpi 3 7 14 21 28

A 0.3 2.8 15.9 22.7 0

Body Surface B C D 0.1 0.7 1.0 1.2 4.4 1.2 15.7 18.8 9.5 19.9 20.5 5.9 0 42.9 0

E 0.7 2.4 3.4 1.7 23.8

F 6.7 34.3 0.6 0 0

G 5.2 19.9 0.3 0 0

Fins H 3.2 9.7 0.2 0 0

I 3.4 7.7 0.1 0 0

J 3.3 13.6 0.7 0 0

Gills K 1.1 1.4 0.1 0 0

Anaesthetic L 74.3 1.4 34.7 29.3 33.3

Mortalities There was a single fish mortality in each of the control groups and the low lice infestation group. There were 54 mortalities in the medium and high lice intensity groups and mortalities increased rapidly once the L. salmonis reached the mobile stages (10 dpi). For reasons of animal welfare, any remaining fish in the high lice intensity and medium lice intensity tanks were sacrificed at 14 and 21 dpi respectively. It was noticed that from 10 dpi onwards, infected fish in the medium lice intensity and high lice intensity groups showed a marked decrease in feeding, and this continued for the remainder of the experiment.

Size and condition factor The exponent of the log length-log weight relationship was 2.90 (standard error = 0.002), which was a significant departure (p < 0.05) from an exponent of 3. However, Fultonâ&#x20AC;&#x2122;s condition factor showed no significant length dependence (K = 0.939 â&#x20AC;&#x201C; 0.00058.length, r = 0.127, p = 0.099), due to the restricted size range of the experimental fish (124-198 mm; 19.3-69.9 g).

Generalized Additive Models Predicted mean values with 95% confidence intervals based on the GAMs fitted at each time point are shown in Figures 2-4. Specifically, predictions and associated intervals were generated for fish of average length and of average condition at each time point using the fitted GAMs and lice levels of 0, 10, 30 and 50 158

SUMBAWS

Final Report

L. salmonis 路 fish-1. Predictions for 50 lice 路 fish-1 at 21 dpi and 10, 30 and 50 lice 路 fish-1 at 28 dpi were excluded from Figures 2-4 to avoid extrapolation; the maximum levels observed at 21dpi and 28 dpi were 44 lice and 5 lice 路 fish-1 respectively.

No significant lice effects were observed at 3 dpi or 7 dpi for haematocrit or plasma chloride (Figure 2), and, whereas the lice relationship was not statistically significant at the 5% level for plasma osmolality at 3 dpi, there was some evidence of a lice effect at 7 dpi (p = 0.06; Figure 2). A significant lice effect was apparent at 14 dpi and 21 dpi for plasma chloride and plasma osmolality. Despite there being no significant lice effects at 21 dpi for haematocrit, there was weak evidence (p = 0.08) of a lice effect at 14 dpi.

There was no evidence of a lice effect at 3 dpi or 28 dpi for plasma glucose, plasma lactate or liver glycogen (Figure 3); however, a lice effect was significant at 14 dpi for these three measures. Whilst there were no lice effects at 7 dpi for liver glycogen there was a significant elevation in liver glycogen concentration across all experimental groups. A significant lice effect was exhibited for plasma glucose at 7 dpi and plasma lactate at 21 dpi, whereas only weak evidence (p = 0.08) of a lice effect was evident at 21 dpi for plasma glucose. Plasma cortisol (Figure 4) exhibited a significant lice effect at 3, 14 and 21 dpi, but lice effects were not statistically significant at the 5% level at 7 or 28 dpi.

Piecewise linear modelling Standard linear terms were preferred for lice for the plasma osmolality and haematocrit responses at 14 dpi, whereas a breakpoint at as few as 13 L. salmonis per fish was identified for plasma cortisol at this time point (Figure 5). The breakpoints identified by the AIC were notably higher for plasma lactate (34, 43, 67) and glucose (68) at 14 dpi (Figures 5 & 6); however breakpoints for plasma glucose (6, 13) and plasma lactate (13, 24) were substantially reduced at 21dpi (Figures 5 & 6 and Table 3).

The AIC statistic suggested a standard linear lice term was preferable for the plasma chloride and plasma cortisol models at 21 dpi, indicating that increasing L. salmonis levels resulted in a constant increase in these physiological measures. Whereas this suggests that the extra parameters required to fit piecewise linear terms for plasma chloride and plasma cortisol were not justified at 21 dpi, these measures did also identify breakpoints at 13 L. salmonis per fish when piecewise linear terms for lice were fitted (Table 3).

Table 3 Predicted breakpoints at 21 dpi. Breakpoints identified at 21 dpi using piecewise linear models. In the interests of parsimony the AIC statistic was used to discriminate between a standard linear term and a piecewise linear term for lice. Where a linear model was preferred (under the AIC) the breakpoints chosen using a piecewise approach are shown in parentheses.

Physiological Measure Plasma glucose Plasma lactate Plasma osmolality Plasma chloride Plasma cortisol Haematocrit

Breakpoints at 21 dpi 6,13 13,24 12 Linear (6,13,24) Linear (8,24) Linear (6,13,24)

159

SUMBAWS

Final Report

40

Haematocrit (%)

38 36 34 32 30 28

Plasma Chloride (mmol.l-1)

26 0

7

14

21

28

0

7

14

21

28

150 138 126 114 102

Plasma Osmolality (mOsm.kg-1)

90

470 Control -1 10 lice.fish -1 30 lice.fish -1 50 lice.fish

456 442 428 414 400 0

7

14

21

28

Days Post Infection (dpi) Figure 2 Predicted mean values with 95% confidence intervals for haematocrit (top) plasma chloride (middle) and plasma osmolality (bottom) at four fixed lice loadings (Control (no lice) – solid black line and ○, 10 L. salmonis · fish-1 – broken line and ●, 30 L. salmonis · fish-1 – broken line and ∆, 50 L. salmonis · fish-1 – dotted line and ▲). These predictions were obtained by fixing other covariates (length and condition factor) at their means for each time point. Data are slightly offset for clarity.

160

Plasma Glucose (mmol.l-1)

SUMBAWS

Final Report

4.0

3.5

3.0

2.5

2.0 0

7

14

21

28

0

7

14

21

28

Plasma Lactate (mmol.l-1)

3.0 2.5 2.0 1.5 1.0 0.5

Liver Glycogen (mg.g-1 wet wt tissue)

0.0

40 Control -1 10 lice.fish -1 30 lice.fish -1 50 lice.fish

30 20 10 0 0

7

14

21

28

Days Post Infection (dpi) Figure 3 Predicted mean values with 95% confidence intervals for plasma glucose (top) plasma lactate (middle) and liver glycogen (bottom) at four fixed lice loadings. Details as in Figure 2.

161

SUMBAWS

Final Report

300 Control -1 10 lice.fish -1 30 lice.fish -1 50 lice.fish

Plasma Cortisol (ng.ml-1)

250

200

150

100

50

0 0

7

14

21

28

Days Post Infection (dpi)

Figure 4 Predicted mean values with 95% confidence intervals for plasma cortisol at four fixed lice loadings. Details as in Figure 2.

Figure 5 Results from the piecewise linear models; where a standard linear term for lice was preferred (under the AIC statistic) a linear term is shown (e.g. Plots A and D). A. plasma osmolality at 14 dpi; B. plasma osmolality at 21 dpi; plasma cortisol at 14 dpi; D. plasma cortisol at 21 dpi. Breakpoint values are shown in Table 3.

162

SUMBAWS

Final Report

Figure 6 Results from the piecewise linear models. A. plasma lactate at 14 dpi; B. plasma lactate at 21 dpi; plasma glucose at 14 dpi; D. plasma glucose at 21 dpi. Breakpoint values are shown in Table 3.

Discussion This is the first investigation of the physiological effects of L. salmonis infestation on wild sea trout smolts. Previous physiological studies of the impact of L. salmonis infection of salmon (S. salar) smolts have concerned parasitic challenge of fully SW-acclimated juvenile host fish, some days or weeks after their experimental transfer to SW. Such data are of limited value in ascertaining the overall impacts of ectoparasitic infection of wild fish because L. salmonis commence infestation of the host fish immediately on entry to sea water. The present study is the first attempt to model a sublethal threshold number of L. salmonis · fish-1, above which significant physiological stress occurs. The setting of an objectively derived minimum ‘target sea lice burden’ known to have negligible physiological effects on the host had earlier been identified as being central to any strategy for monitoring L. salmonis on wild and farmed fish (Sea lice monitoring protocols, Unpublished report of UK Coordinator of Fisheries Research and Development (UKCFRD), Pitlochry, October 1999).

Previous studies have demonstrated that the developmental time course of L. salmonis on sea trout (Bjørn and Finstad 1998) corresponds to reports for Atlantic salmon (Grimnes and Jakobsen 1996; Johnson and Albright 1991). Male L. salmonis on salmon develop faster than females, taking approximately 29-40 and 38-52 d respectively to reach the adult stage at 10°C. In the present study, L. salmonis development was somewhat quicker than those described for hatchery-reared sea trout (Bjørn and Finstad 1998), reflecting the higher ambient water temperature in our aquarium (13.6°C). However, infestation success and the spatial distribution of L. salmonis over the body surface were similar to those previously described by Bjørn and Finstad (1998). A small proportion of copepodid and chalimus larvae were found on the gills, but no adult or preadult L. salmonis were observed there. A large proportion of the total number of copepodids and mobile L. salmonis also were recovered from the anaesthetic solutions, emphasising the importance of careful collection of detached and dislodged sea lice in this type of study.

163

SUMBAWS

Final Report

Plasma chloride & osmolarity Mean plasma chloride levels remained relatively stable throughout the experimental period in the control group. No difference was observed between control and infected fish when the L. salmonis were at the late chalimus stage (7 dpi), in contrast to previous studies on hatchery-reared sea trout (Bjørn and Finstad 1997), but in agreement with experiments on Atlantic salmon (Grimnes et al. 1996; Grimnes and Jakobsen 1996).

However, a significant parasite-induced increase in plasma chloride was apparent once L. salmonis reached the mobile preadult stages (14 & 21 dpi). This may be due in part to the increased feeding activity by mobile preadult L. salmonis and associated skin damage (Bjørn and Finstad 1998; Pike and Wadsworth 1999), leading to a reduction in osmoregulatory ability. In addition, disturbance both of water and ion homeostasis is a common aspect of stress in fish (Wendelaar Bonga 1997). The disturbance of hydromineral balance may occur due to marked structural changes in the gills which can result from a range of stressors (Wendelaar Bonga 1997). A similar effect was observed for plasma osmolality but for this physiological measure there was weak evidence for a lice effect at 7 dpi, when the parasites were still at the attached chalimus stage (Table 1).

Plasma cortisol An elevation in plasma cortisol is the most widely used indicator of stress in fish (Wendelaar Bonga 1997). Typically, plasma cortisol levels rise rapidly a few minutes after exposure to an acute stressor, and the return to normal levels may take one or more hours. When the stressor is chronic, cortisol levels may remain elevated, although well below peak levels (Wendelaar Bonga 1997). Cortisol has a broad activity spectrum in fish, including the regulation of hydromineral balance, energy metabolism and a reduction in the immune response (Wendelaar Bonga 1997). In addition, cortisol may be responsible for structural changes in the gills, such as necrosis of the branchial epithelial cells or epithelial lifting. These may have pronounced effects on the major functions of the gill, often leading to disturbance of water and ion homeostasis (Wendelaar Bonga 1997). In the present study, an initial lice effect on cortisol was observed as early as 3 dpi, but by 7 dpi no lice effect was apparent. The lice effect at 3 dpi may indicate an initial acute stress arising from the process of copepodid attachment. During the first 2-3 dpi fish in the medium and high lice intensity tanks often showed distinct ‘flashing’ behaviour or jumped clear of the water surface. By 7 dpi this behaviour had effectively stopped and was not observed again until the L. salmonis developed to the mobile stages. The significant lice effect observed at 14 and 21 dpi may indicate chronic stress arising from increased feeding activity by mobile, preadult L. salmonis. These significantly elevated plasma cortisol concentrations are consistent with the well characterised response to stress (Wendelaar Bonga 1997), and a similar response was observed in previous studies on hatchery-reared sea trout (Bjørn and Finstad 1997).

Metabolic indicators Both acute and chronic stress typically are associated with increased metabolic rate, because response to a stressor is energy-demanding (Mommsen et al. 1999). To satisfy the increased energy demand, fish mobilise substrates to fuel cellular processes. There have been many reports of an elevation of plasma glucose (hyperglycemia), and a decrease in liver glycogen (glycogenolysis) during stress (Wendelaar Bonga 1997). Hyperglycemia represents a secondary stress response and also is useful as a stress indicator in fish (Barton 2000). In addition to hyperglycemia, elevated plasma lactate levels have been reported following both acute and chronic stress (Farbridge and Leatherland 1992; Iversen et al. 1998). Anaerobic conditions caused by exhaustive exercise result in the breakdown of muscle glycogen to lactate, with some of this lactate being released into the bloodstream (Wood 1991). However, it is important to note that liver glycogen, plasma glucose and plasma lactate all are influenced by the metabolic status and feeding history of the animal.

Plasma glucose and lactate There was a significant lice-induced elevation of both plasma glucose and plasma lactate concentrations at 14 and 21 dpi, perhaps reflecting an increased energy demand to the animal due to the stress of L. salmonis infestation. This increased energy demand may have resulted in anaerobic breakdown of muscle glycogen to lactate, perhaps connected to the ‘flashing’ behaviour observed at this time. These extra energy demands may explain the significant lice effect observed for plasma glucose at 7 dpi.

The unavoidable short delay in freeze-clamping the liver samples (following prior removal of blood samples) may have resulted in a slight underestimation of liver glycogen content, due to initiation of liver 164

SUMBAWS

Final Report

glycogen breakdown. However, liver sampling was completed within 3 min of sacrifice, in common with previous salmonid studies (Plisetskaya et al. 1994), and there was an equivalent delay in sampling for fish in all experimental groups. Liver glycogen was low at 3 dpi, perhaps reflecting reduced feeding activity in FW. In general there was an increase in liver glycogen at 7 dpi. It is most likely that this increase was attributable to SW transfer, one week earlier. In the present study liver glycogen concentration returned to levels approaching basal in all groups and remained low for the remainder of the experiment.

Identification of threshold parasite intensities The piecewise linear approach permitted the objective identification of abrupt changes, or threshold intensities, which occurred as a result of the effects of increasing numbers of L. salmonis for each physiological measure. At 21 dpi a strikingly consistent breakpoint of 12-13 L. salmonis · fish-1 was identified across the majority of physiological measures. Earlier studies of a restricted range of measures (plasma ions and cortisol) indicated that 50 mobile L. salmonis · fish-1 may cause the death of a small sea trout post-smolt (60 g; c.f. our mean of 37 g) (Bjørn and Finstad 1997). Similarly, as few as 30 adult L. salmonis were found to cause significant osmoregulatory imbalance in plasma ion concentrations (Bjørn and Finstad 1997). This reported breakpoint/threshold of L. salmonis infestation intensity is markedly below those previously reported, because this experimental approach is environmentally more relevant in reflecting the dual challenge of infestation of wild sea trout smolts and physiological acclimation to SW. Identification of the 12-13 L. salmonis · fish-1 threshold is an essential prerequisite in formulating effective wild fisheries management and stock conservation policy, and these results should provide a sound basis for continued policy development (Heuch et al. 2005). However, difficulties remain in translating physiological predictors into a useable tool in the management of both farmed and wild populations. For example, the age and size of first migrant juvenile sea trout are known to vary considerably throughout the species geographic range (Klemetsen et al. 2003). However, in the absence of an objective means of scaling lice loadings on individual fish of differing size, we would recommend that the conservative and precautionary approach would be for wild fishery managers to adopt a critical level of 13 mobile L. salmonis · fish-1 for juvenile sea trout primarily in their first ‘sea’ year. This level would provide a clear indication of the proportion of sea trout within a population which are subject to physiological stress and potential death from sea lice infestation.

Acknowledgements This study was carried out with the support of European Commission DG Fisheries Contract No. Q5RS2002-00730 “Sustainable Management of Interactions Between Aquaculture and Wild Salmonid Fish (SUMBAWS)”. Fish were provided with the permission of the Esk District Salmon Fishery Board and were collected from the Kinnaber trap with the help of Fisheries Research Services personnel. We thank A.F. Walker for comments on the manuscript.

References Akaike, H. 1973. Information theory and an extension of the maximum likelihood principle In International Symposium on Information Theory. Edited by B. N. Petrov, and F. Csaki. Akademia Kiado, Budapest.

Anonymous. 1993. Report of the Sea Trout Working Group 1993. Department of the Marine, Abbotstown.

Anonymous. 1995. Report of the Sea Trout Working Group 1994. Department of the Marine, Abbotstown.

Anonymous. 2004. Statistical Bulletin. Scottish salmon and sea trout catches, 2003. Report No. Fis/2004/1.

Fisheries Research Services, Aberdeen.

Barton, B.A. 2000. Salmonid fishes differ in their cortisol and glucose responses to handling and transport stress. N. Am. J. Aquacult. 62: 12-18.

Bernatchez, L., and Wilson, C.C. 1998. Comparative phylogeography of Nearctic and Palearctic fishes.

Mol. Ecol. 7: 431-452.

165

SUMBAWS

Final Report

Birkeland, K. 1996. Consequences of premature return by sea trout (Salmo trutta) infested with the salmon louse (Lepeophtheirus salmonis Krøyer): migration, growth, and mortality. Can. J. Fish. Aquat. Sci. 53: 2808-2813.

Bjørn, P.A., and Finstad, B. 1997. The physiological effects of salmon lice infection on sea trout post smolts. Nordic J. Freshw. Res. 73: 60-72.

Bjørn, P.A., and Finstad, B. 1998. The development of salmon lice (Lepeophtheirus salmonis) on artificially infected post smolts of sea trout (Salmo trutta). Can. J. Zool. 76: 970-977.

Breusch, T.S., and Pagan, A.R. 1979. A simple test for heteroscedasticity and random coefficient variation. Econometrica 47: 1287-1294.

Butler, J.R.A. 2002. Wild salmonids and sea louse infestations on the west coast of Scotland: sources of infection and implications for the management of marine salmon farms. Pest Manag. Sci. 58: 595-608.

Dawson, L.H.J., Pike, A.W., Houlihan, D.F., and McVicar, A.H. 1998. Effects of salmon lice Lepeophtheirus salmonis on sea trout Salmo trutta at different times after seawater transfer. Dis. Aquat. Org. 33: 179-186.

Farbridge, K.J., and Leatherland, J.F. 1992. Plasma growth hormone levels in fed and fasted rainbow trout (Oncorhynchus mykiss) are decreased following handling stress. Fish Physiol. Biochem. 10: 67-73.

Finstad, B., Bjorn, P.A., Grimnes, A., and Hvidsten, N.A. 2000. Laboratory and field investigations of salmon lice Lepeophtheirus salmonis (Kroyer) infestation on Atlantic salmon (Salmo salar L.) post-smolts.

Aquac. Res. 31: 795-803.

Finstad, B., Johnsen, B.O., and Hvidsten, N.A. 1994. Prevalence and mean intensity of salmon lice, Lepeophtheirus salmonis Krøyer, infection on wild Atlantic salmon, Salmo salar L., postsmolts. Aquac.

Fish. Man. 25: 761-764.

Fox, J. 1997. Applied regression, linear models and related methods. Sage Publications, Thousand Oaks, California.

Fox, J., and Monette, G. 1992. Generalized collinearity diagnostics. J. Am. Stat. Ass. 87: 178-183.

Grimnes, A., Finstad, B., and Bjorn, P.A. 1996. Okologiske og fysiologiske konsekvenser av lus på laksefisk i fjordsystem. NINA Oppdragsmelding 381: 1-38.

Grimnes, A., and Jakobsen, P.J. 1996. The physiological effects of salmon lice infection on post-smolt of Atlantic salmon. J. Fish Biol. 48: 1179-1194.

Hansen, L.P., and Jacobsen, J.J. 2000. Distribution and migration of Atlantic salmon, Salmo salar L., in the sea. In The ocean life of Atlantic salmon. Environmental and biological factors influencing survival.

Edited by D. Mills. Fishing News Books, Oxford.

Heuch, P.A., Bjorn, P.A., Finstad, B., Holst, J.C., Asplin, L., and Nilsen, F. 2005. A review of the Norwegian 'National Action Plan Against Salmon Lice on Salmonids': the effect on wild salmonids.

Aquaculture 246: 79-92.

Iversen, M., Finstad, B., and Nilssen, K.J. 1998. Recovery from loading and transport stress in Atlantic salmon (Salmo salar L.) smolts. Aquaculture 168: 387-394.

Johnson, S.C., and Albright, L.J. 1991. The developmental stages of Lepeophtheirus salmonis (Krøyer, 1837) (Copepoda, Caligidae). Can. J. Zool. 69: 929-950.

Keppler, D., and Decker, K. 1974. Glycogen: determination with amyloglucosidase. In Methods of Enzymatic analysis. Edited by H.U. Bergmeyer. Academic Press.

Klemetsen, A., Amundsen, P.A., Dempson, J.B., Jonsson, B., Jonsson, N., O'Connell, M.F., and Mortensen, E. 2003. Atlantic salmon Salmo salar L., brown trout Salmo trutta L. and Arctic charr Salvelinus alpinus (L.): a review of aspects of their life histories. Ecol. Freshw. Fish 12: 1-59.

166

SUMBAWS

Final Report

McKibben, M.A., and Hay, D.W. 2004. Distributions of planktonic sea lice larvae Lepeophtheirus salmonis in the inter-tidal zone in Loch Torridon, Western Scotland in relation to salmon farm production cycles. Aquac. Res. 35: 742-750.

Molinari, N., Daures, J.-P., and Durand, J.-F. 2001. Regression splines for threshold selection in survival data analysis. Stat. Med. 20: 237-247.

Mommsen, T.P., Vijayan, M.M., and Moon, T.W. 1999. Cortisol in teleosts: dynamics, mechanisms of action, and metabolic regulation. Rev. Fish. Biol. Fish. 9: 211-268.

Neff, B.D., and Cargnelli, L.M. 2004. Relationships between condition factors, parasite load and paternity in bluegill sunfish, Lepomis macrochirus. Environ. Biol. Fishes 71: 297-304.

Nolan, D.T., Reilly, P., and Wendelaar Bonga, S.E. 1999. Infection with low numbers of the sea louse Lepeophtheirus salmonis induces stress-related effects in postsmolt Atlantic salmon (Salmo salar). Can. J. Fish. Aquat. Sci. 56: 947-959.

Northcott, S.J., and Walker, A.F. 1996. Farming salmon, saving sea trout; a cool look at a hot issue. In Aquaculture and sea lochs. Edited by K. Black. Scottish Association for Marine Science, Oban.

Penston, M.J., McKibben, M.A., Hay, D.W., and Gillibrand, P.A. 2004. Observations on open-water densities of sea lice larvae in Loch Shieldaig, Western Scotland. Aquac. Res. 35: 793-805.

Pike, A.W., and Wadsworth, S.L. 1999. Sealice on salmonids: their biology and control. Adv. Parasitol.

44: 233-337.

Plisetskaya, E.M., Moon, T.W., Larsen, D.A., Foster, G.D., and Dickhoff, W.W. 1994. Liver glycogen, enzyme activities, and pancreatic hormones in juvenile Atlantic salmon (Salmo salar) during their 1st summer in seawater. Can. J. Fish. Aquat. Sci. 51: 567-576.

R Development Core Team 2004. R: A language and environment for statistical computing. Vienna, R Foundation for Statistical Computing.

Royston, P. 1982. An extension of Shapiro and Wilk's W Test for normality to large samples. Applied Statistics 31: 115-124.

Schram, T. A. 1993. Supplementary descriptions of the developmental stages of Lepeophtheirus salmonis (Krøyer, 1837) (Copepoda: Caligidae). In Pathogens of wild and farmed fish. Sea lice. Edited by G. A. Boxshall, and D. Defaye. Ellis Horwood, Chichester.

Shea, E.K., and Vecchione, M. 2002. Quantification of ontogenetic discontinuities in three species of oegopsid squids using model II piecewise linear regression. Mar. Biol. 140: 971-979.

Sutton, S.G., Bult, T.P., and Haedrich, R.L. 2000. Relationships among fat weight, body weight, water weight, and condition factors in wild Atlantic salmon parr. Trans. Am. Fish. Soc. 129: 527-538.

Tully, O., Poole, W.R., Whelan, K.F., and Merigoux, S. 1993. Parameters and possible causes of Lepeophtheirus salmonis (Krøyer) infesting sea trout (Salmo trutta L.) off the west coast of Ireland. In Pathogens of wild and farmed fish. Sea lice. Edited by G. A. Boxshall, and D. Defaye. Ellis Horwood, Chichester. van Anholt, R.D., Spanings, T., Koven, W., and Wendelaar Bonga, S.E. 2003. Effects of acetylsalicylic acid treatment on thyroid hormones, prolactins, and the stress response of tilapia (Oreochromis mossambicus). Am. J. Physiol. 285: R1098-R1106.

Wendelaar Bonga, S.E. 1997. The stress response in fish. Physiol. Rev. 77: 591-625.

Wold, S. 1974. Spline functions in data analysis. Technometrics 16:1-11.

Wood, C.M. 1991. Acid-base and ion balance, metabolism, and their interactions, after exhaustive exercise in fish. J. Exp. Biol. 160: 285-308.

167

SUMBAWS

Final Report

Paper 2: ASPECTS OF THE STRESS RESPONSE OF SEA TROUT SMOLTS, SALMO TRUTTA (L.), TO VARYING SALMON LICE, LEPEOPHTHEIRUS SALMONIS (KRﾃ郎ER), INTENSITIES WHEN ENTERING SEAWATER: THE EPITHELIAL RESPONSE. I.J. Russon, A. Wells, C.E. Grierson, C.D. Todd, N. Hazon and S.E. Wendelaar Bonga.

(Manuscript) Introduction The integrated stress response of fish comprises three levels. The primary response involves the activation of brain centres, causing the release of catecholamines and corticosteroids. The secondary response involves the actions of these hormones on the blood and tissue of the fish e.g. disturbance of hydromineral balance, oxygen uptake, and mobilisation of energy. Finally, tertiary responses extend to the level of the whole organism and population, e.g. inhibition of growth, reproduction, the immune response, and a reduced capacity to tolerate additional stressors (Wendelaar Bonga, 1997).

The skin epithelium of fish forms the first barrier between the external and internal environment and, when influenced by a number of endocrine factors, demonstrates characteristic changes in individuals exposed to stressors. One of the best ways to evaluate the effects of environmental and external stressors in fish is to quantify changes in the epithelial tissues, such as skin and branchial epithelia, that interface with the external environment (Iger et al, 1995).

Disruption of the gill has deleterious effects on the fish because the gills are the main site of ion regulation, gaseous exchange, and excretion of nitrogenous waste. This has led the majority of research to focus on the gills but, more recently, the response of teleost skin to stressors has also been investigated (e.g.

Nolan, 2000a). Epithelial changes following parasite infestation are mediated by cortisol and/or catecholamines circulating in the blood (e.g. increased permeability of gills to water and ions; Wendelaar Bonga, 1997), and as a result these effects also occur in areas not affected directly by the attached parasites (Nolan, 1999). This is evidence of the indirect effect of parasites on the teleost integument.

Teleost fish must maintain a constant balance between their internal electrolytes and the electrolyte levels of the external medium. Branchial chloride cells have a primary role in ion regulation (Perry, 1998), and in marine teleosts they actively secrete Na+ and Cl- ions in order to counteract the effects of the hyperosmotic conditions of seawater (Daborn et al, 2001). Chloride cells have been observed to increase in number and size on transfer of fish from freshwater to seawater (e.g. Foskett and Hubbard, 1981; Pisam et al, 1987; Pisam and Rambourg, 1991). In addition to SW transfer, cortisol is known to cause an increase in the synthesis of branchial Na+/K+-ATPase in the chloride cell (Metz et al, 2003) and the development of new chloride cells in salmonids (Richman and Zaugg, 1987).

Fish mucus has many roles including gaseous exchange, ionic and osmotic regulation, reproduction, excretion, disease resistance, feeding and nest building (Shephard, 1994). The relationship between stressor and mucus cell numbers appears to be highly complex, the response being specific to both the stressor and fish species. Initial hyperactivity of the mucus cells, followed by a decrease in cell number, is a common stress response (Burkhardt-Holm et al, 1997). No changes in mucus cell number were observed in carp (Cyprinus carpio) epidermis following cortisol administration at low concentrations and subsequent low intensity infestation with Argulus japonicus (van der Salm et al, 2000). Reduced numbers of epidermal mucus cells have been reported for Atlantic salmon (Salmo salar) exposed to low infestation intensities of sea lice, Lephtheirus salmonis (Nolan et al, 1999). However, an increase in the number of branchial mucus cells has been observed in Rainbow trout (Oncorhynchus mykiss) during infection with bacterial gill disease (Ferguson et al, 1992).

Perry (1998) states that lamellar chloride cell proliferation is beneficial to ionic regulation. The rate of cell turnover can be assessed through observations of levels of cell proliferation (growth), and cell death through apoptosis (programmed cell death) and/or necrosis (un-programmed cell death). Increased cell turnover is associated with stress (Wendelaar Bonga, 1997). An increase in proliferation of mucus and branchial chloride cells will increase the ability of the fish to withstand the effects of stress. Cell proliferation is a fundamental process in adaptation to disease (Boulton and Hodgson, 1995) where dying cells are replaced, and newly-proliferated cells are likely to be more efficient than mature cells.

The following work focused mainly on the secondary stress response of sea trout (Salmo trutta) to the sea louse L. salmonis, with an emphasis on changes in head skin and branchial epithelia. Specific changes in 168

SUMBAWS

Final Report

cell type, distribution and morphology were investigated, together with changes in the general epithelial integrity.

Materials and Methods: Collection of fish, infestation with sea lice, tank conditions and sampling regimes are detailed elsewhere (Wells et al., 2006; Paper 1 in Final Report, WP5). However, only the control, low (~10 lice · fish-1) and medium (~30 lice · fish-1) infestation groups were analysed in the present study. Furthermore, no samples were analysed from 28 days post infection (dpi) due to the mortalities and low lice levels as detailed in Wells et al. (2006).

Sampling procedures Skin biopsies from the anterior part of the head (head skin) and gill sections from the second gill arch of the left side were prefixed for electron microscopy (EM) in a 3% gluteraldehyde in 0.1M cacodylate buffer solution for 15 minutes. For postfixation, samples were immediately transferred into a 2% osmium tetroxide in 0.1M cacodylate buffer solution for 1 hour. Samples were then rinsed in demineralised water (minimum of five times over 30 minutes) and finally placed in 70% ethanol for storage until processing. Further head skin biopsies and the entire second gill arch from the right side of the fish were fixed for light microscopy (LM). These samples were placed in Bouin’s fixative for 24 hours, rinsed in tap water to remove excess picric acid, and then placed in 70% ethanol for storage.

Sample processing Light Microscopy (LM) Bouin fixed samples (n = 6 per group and sample point) were dehydrated through an ethanol series, infiltrated with paraffin, and cut in 5µm sections. Head skin samples were cut in longitudinal sections, with gill samples cut at right angles to the plane of the secondary lamellae.

MUCUS CELLS Head skin and gill samples were stained with Periodic Acid-Schiff’s (PAS) stain to identify total mucus cell numbers. This involves initial oxidation of samples with 0.5% Periodic Acid, and subsequent submersion in Schiff’s stain. Acidophilic glycoprotein-containing mucus cells were stained using Alcian blue II (AB.2) at pH 2.6. AB.2 stains both carboxylated and sulphated mucins dark blue (AB.2 positive (+ve) cells), whereas neutral glycoprotein-containing mucus cells (AB.2 negative (-ve)) have no, or very pale blue, staining. Cells were quantified using the method described by Nolan et al (1999). Total cell numbers were counted using a Carl Zeiss light microscope in 10 randomly selected 300 µm sections. The mean number of mucus cells per mm length of epidermis was calculated for the head skin and the mucus cells per mm length of lamellae were calculated for the branchial epithelium. The same quantification method was used for all subsequent histological and immunohistochemical procedures. Images were taken using a Leica DN-RBE fluorescent microscope, fitted with a Leica DC 500 digital camera. AB.2+ve and AB.2-ve cells were distinguished in the head skin. Furthermore a distinction between migrating and nonmigrating mucus cells in the branchial epithelia was also made, with non-migrating cells defined as being present at the base of the gill lamellae and migratory cells occurring along the length of the gill lamellae.

CHLORIDE CELLS Branchial chloride cells were stained immunohistochemically using an antibody for Na+/K+ATPase, as described by Metz et al (2003). Both total cell and migratory/non-migratory chloride cell numbers were counted.

PROLIFERATING CELLS A monoclonal antibody against proliferating cell nuclear antigen (PCNA) was used to visualise proliferating cells, as described by Van der Salm et al (2000). Total numbers of PCNA-positive cells were counted as described above.

Scanning Electron Microscopy (SEM) SEM analysis (JEOL SEM 6330 scanning electron microscope) was used to observe the structural integrity of skin and branchial epithelia. Gluteraldehyde – osmium fixed head skin and gill samples, stored 169

SUMBAWS

Final Report

in 70% ethanol, were dehydrated through an ethanol series, and chemically dehydrated using a 100% ethanol:HMDS (1,1,1,3,3,3-hexamethyldisilazane) series (2:1, 1:1, 1:2, 100% HMDS × 2; each for 1-2 minutes). Dried samples were then mounted on aluminium blocks and gold sputtered under an argon atmosphere using a Balzers Union gold sputter coater.

Statistics All cell count data obtained from light microscopy were tested for homogeneity of variance using Levene’s statistic. Non-homogenous data were log transformed to attempt to obtain homogeneity of variance, but only the data for total branchial mucus cells stained with PAS, and the ratio of AB.2 +ve/–ve epithelial cells showed homogeneity of variance. Log transformed total branchial PAS positive cells, total (+ve and –ve) AB.2 skin cells, log transformed AB.2 +ve/-ve skin mucus cells, and PCNA positive branchial epithelium cells all were tested using ANOVA and differences between treatment groups identified using Bonferroni’s multiple range test. Pearson’s correlation analysis was used to examine the relationship between time, infestation intensity and the cell counts. All other cell count data were analysed using non-parametric ANOVA, with differences between treatment groups being identified using Dunnett’s C test. Spearman rank correlation analysis was used to examine the relationship between time, infestation intensity and the cell counts.

Results Mucus cells AB.2 total mucus cell counts (AB.2 +ve and –ve) in the skin epithelia show no lice-induced effect until day 14, when a significant (p < 0.05) reduction in cell number was evident for the medium infestation group (Figure 1). At 21 dpi, the medium infestation mucus cell numbers had returned to levels comparable to the control and low infestation group values at all previous time points (Figure 1). When the ratio of AB.2 +ve/AB.2 -ve skin epithelial mucus cells was examined (Figures 2 and 3) there was a significant increase in ratio for the medium infestation group at 3 and 14 dpi (p < 0.05). At day 21 the ratio of AB.2+ve/AB.2-ve cells of the medium infested group had returned to control levels. No information for AB.2 branchial mucus cells is shown due to extremely low numbers being present.

There were no significant differences in total PAS +ve branchial mucus cells between treatment groups (Figure 4). However, there was a significant reduction in PAS +ve cells within all groups between 3 dpi and 21 dpi (p < 0.01). The ratio of mucus cells migrating along the gill lamellae compared to the non-migratory interlamellar mucus cells varied with infestation intensity (Figure 5), with evidence of a lice-induced increase in the ratio of migratory/non-migratory PAS +ve cells at 3 and 14 dpi (p < 0.05). There was a reduction in the ratio of migratory/non-migratory mucus cells between 3 and 21 dpi (p < 0.01). Examples of migratory and non-migratory mucus cells are shown in Figure 6 (14 dpi, control and medium infestation group). In the skin epithelium there was a significant reduction in PAS +ve mucus cells in the medium infestation group at 14 dpi when compared to the control and low infestation groups at 14 dpi, and in comparison with all other time points for the same infestation intensity group (Figure 7).

70

mucous cells per mm

60 50

Control 0 lpf

40

lpf Low10infestation 30 lpf

Medium infestation

30

*

20 10 0 3

7

14

21

Days Post Infection

Figure 1 Total mucus cells (AB.2+ve and –ve) per mm length epidermis present in the skin epithelia of sea trout infested with differing sea lice intensities (p < 0.05) (bars = S.E.) Asterisk shows significant difference from the control group at that time point.

170

SUMBAWS

Final Report

AB.2 +ve/AB.2 -ve mucous ratio

7

*

6 5 4

0 lpf Control 10infestation lpf Low

3

Medium 30 lpfinfestation

*

2 1 0 3

7

14

21

Days Post Infection

Figure 2 Ratio AB.2 +ve/ AB.2 –ve mucus cells present in the skin epithelia of sea trout infested with differing sea lice intensities (p < 0.05). Asterisk shows significant difference from the control group at that time point.

100µm

a)

b)

Figure 3 AB.2 stained skin epithelia from a) control fish 14 dpi, b) a fish infested with 26 lice (medium infestation intensity) 14dpi. Arrow = AB.2 +ve mucus cell, broken arrow = AB.2 –ve mucus cell.

PAS +ve cells per mm

20 15 0 LPF

Control

10

LPF Low10 infestation Medium infestation 30 LPF

5 0 3

7

14

21

Days post infection

Figure 4 Total PAS +ve mucus cells (migratory/non-migratory) per mm length filamental epithelium in the branchial epithelium of fish infested with differing intensities of sea lice (bars = S.E.)

171

SUMBAWS

Final Report

Ration migratory/nonmagratory cells

0.6

*

*

0.5 0.4

0 LPF Control

0.3

Low10 infestation LPF Medium infestation 30 LPF

0.2 0.1 0 3

7

14

21

Days post infection

Figure 5 Ratio migratory/non-migratory branchial PAS +ve mucus cells present on sea trout infested with differing sea lice intensities (p < 0.05). Asterisk shows significant difference from the control group at that time point.

100Âľm

a)

b)

Figure 6 PAS stained gill epithelia from a) control fish 14 dpi, b) a fish infested with 26 lice (medium infestation intensity) 14 dpi. Arrowhead = migratory PAS +ve mucus cell.

90 80

Mean cells per mm

70 60 0 lpf Control

50

10 lpf Low infestation 40

Medium 30 lpf infestation

30

*

20 10 0 3

7

14

21

Days Post Infection

Figure 7 PAS +ve mucus cells per mm length epidermis in the skin epithelia of sea trout infested with differing sea lice intensities (p < 0.05) (bars = S.E.). Asterisk shows significant difference from the control group at that time point.

172

SUMBAWS

Final Report

Number of cells

Chloride cells There was a significant lice-induced increase in total chloride cells at 3 and 14 dpi (p < 0.05) (Figure 8).

There also was a lice-induced increase in the ratio between migratory and non-migratory chloride cells (p < 0.05) which was apparent at 3, 7 and 14 dpi in the medium infestation group and at 7 and 14 dpi in the low infestation group (Figure 9). There was no difference between the control and low infestation groups at 3 and 21 dpi. An example of the increase in migratory chloride cells along the gill lamellae between a control fish and a fish infested with 26 lice 路 fish-1 at 14 dpi is shown in Figure 10.

180 160 140 120 100 80 60 40 20 0

*

** 0 LPF Control Low 10 infestation LPF Medium infestation

30 LPF

3

7

14

21

Days Post Infection

migratory/non-migratory cell ratio

Figure 8 Total chloride cells (migratory and non-migratory) per mm length filamental epithelium in the branchial epithelium of fish infested with differing intensities of sea lice (bars = S.E.). Asterisk shows significant difference from the control group at that time point.

0.25

*

0.2

*

0.15

*

*

0 LPF Control 10 infestation LPF Low

*

0.1

Medium infestation 30 LPF

0.05 0 3

7

14

21

Days Post Infection

Figure 9 Ratio migratory/non-migratory branchial chloride cells present on sea trout infested with differing sea lice intensities. Asterisk shows significant difference from the control group at that time point.

173

SUMBAWS

Final Report

a)

b)

Figure 10 Na+/K+ ATPase stained chloride cells in the gill epithelia from a) control fish 14 dpi, b) a fish infested with 26 lice (medium infestation intensity) 14dpi. Arrow = Non-migratory chloride cell, arrowhead = migratory chloride cell.

Proliferating cells At 14 dpi the medium infestation group had a significantly higher number of proliferating branchial cells than did the control and low infestation groups (p < 0.05) (Figures 11 and 12). The number of proliferating branchial cells in the medium infestation group also was significantly higher at 14 dpi than at all other time points (Figure 11). There was a significant lice-induced increase in proliferating cells of the skin epithelium which was apparent in the medium infestation group at 3, 7, and 14 dpi (p < 0.01) (Figures 13 and 14). *

100

Mean cells per m m

90

80 0 lpf Control 70

10infestation lpf Low Medium 30 lpf infestation

60

50

40 3

7

14

21

Day Pos t Infe ction

Figure 11 Proliferating cells per mm length filamental epithelium in the branchial epithelium of fish infested with differing intensities of sea lice (p < 0.05) (bars = S.E.). Asterisk shows significant difference from the control group at that time point.

a)

b)

Figure 12 Proliferating (PCNA +ve) cells in the gill epithelia from a) control fish 14 dpi, b) a fish infested with 26 lice (medium infestation intensity) 14dpi.

174

SUMBAWS

Final Report

Mean cells per mm length of epithelium

120

100

*

80

Control 0 lpf

*

60

10infestation lpf Low

*

30 lpf infestation Medium

40

20

0 3

7

14

21

Days Post Infection

Figure 13 Proliferating cells per mm length epithelium in the skin epithelia of sea trout infested with differing numbers of sea lice (p < 0.01) (bars = S.E.). Asterisk shows significant difference from the control group at that time point.

a)

b)

Figure 14 Proliferating (PCNA +ve) cells in the skin epithelia from a) control fish 14 dpi, b) a fish infested with 26 lice (medium infestation intensity) 14dpi.

SEM analyses Figures 15â&#x20AC;&#x201C;17 are SEM images of skin epithelium in control, low and medium infestation groups, respectively, at 14 dpi. The integrity of the skin of control and low infestation groups are similar, with no disruption to normal epithelial integrity. Concentric microridges at the apical membranes of the pavement cells and tight connections between these cells can be seen. Disruptions in epithelial integrity in the medium infestation intensity group can be seen in Figure 17. The loss of the concentric microridge structure (*), the presence of intercellular spaces (arrowheads), and a greater number of pores of mucus cells (arrows) are all seen in the medium infestation group.

Figures 18 â&#x20AC;&#x201C; 20 are SEM images of the branchial epithelium in control, low and medium infestation groups, respectively, at 14 dpi. As with the skin epithelia, the gill filaments of the low infestation group were in a condition similar to the control group. The gills from the medium infestation group show folding of the gill lamellae.

175

SUMBAWS

Final Report

Figure 15 Skin Epithelia from a control fish 14 dpi. Broken arrow = intercellular connections.

Figure 16 Skin Epithelia from a fish infested with 9 lice (low infestation intensity) 14 dpi.

Figure 17 Skin Epithelia from a fish infested with 26 lice (medium infestation intensity) 14 dpi. * = deterioration of pavement cell apical microridge structure. Arrow = mucus pores. Arrowhead = intercellular spaces.

176

SUMBAWS

Final Report

F

Figure 18 Branchial epithelia from a control fish 14dpi. F = Gill filament. Arrows = secondary gill lamellae.

Figure 19 Branchial epithelia from a fish infested with 9 lice (low infestation intensity) 14 dpi.

Figure 20 Branchial epithelia from a fish infested with 26 lice (medium infestation intensity) 14 dpi.

Arrowhead = folding of gill lamellae.

177

SUMBAWS

Final Report

Discussion It has been well documented that the variety of changes in the external epithelia of fish induced by stress, occur mainly due to the effects of stress corticosteroids and catecholamines (e.g., Iger et al, 1995; Nolan et al, 2000a and 2000b; van der Salm et al, 2002). Changes in the epithelia of fish take more time to develop or disappear than do changes in hormone levels, therefore epithelial changes are a useful indicator of prolonged chronic stress. The increased effect of an additional stressor on the physiology of already chronically stressed fish has been reported by e.g., Ruane et al (2002). Carp (Cyprinus carpio) under high stocking densities showed an increased stress response to net confinement in comparison to carp maintained at lower stocking densities.

In many of the cell types observed in this study, e.g. total skin mucus cells with both AB.2 and PAS, and proliferating cells, differences were not noted until 14 dpi of the experiment. This coincides with the development of the larval lice to the mobile pre-adult stage, and the associated increase in lice feeding activity. This will have led to chronic stress following the initial acute stress of acclimation to seawater and simultaneous infestation with larval sea lice earlier in the experiment. In the present study, in some instances, an effect on the cells was seen at 3 dpi, reflecting this early acute stress on the fish.

The integument of fish is protected with a chemically and functionally complex mucus coat, which is discharged by specialised mucus cells in the epidermis (Shephard, 1994). An increase in mucus production is one of the anti-parasitic defence mechanisms recognisable in fish (Buchmann, 1999) and in the present study changes in mucus cell numbers were observed in both the skin and gills. However, these varied depending upon epithelium type. In the skin epithelia, the greatest change was observed at 14 dpi within the medium (10 lice · fish-1) infestation intensity group, probably due to increased feeding activity of the mobile lice. There appeared to be no lice-induced changes in branchial mucus cell numbers. However, a liceinduced increase in the number of migrating PAS stained branchial mucus cells was observed both at 3 and 14 dpi. Ferguson et al (1992) observed that the greatest increase in mucus-producing cells occurred around the tips of the lamellae in rainbow trout infected with bacterial gill disease. Cells migrating along the lamellae were larger and less constricted than those present at the base of the lamellae, which allows for a greater mucus content. In addition, mucus cells on the lamellae are closer to the surface of the epithelia, thus making it easier to discharge the mucus. AB.2 stains acidic glycoprotein containing mucus cells (Berntssen et al, 1997). It is interesting to note that a reduction in the number of skin mucus cells coincided with an increase in the AB.2 +ve compared to AB.2 –ve cells (indicative of a relative increase in acid glycoproteincontaining skin mucus cells as observed by Berntssen et al (1997) in Atlantic salmon (S. salar) and Ferguson et al (1992) in rainbow trout (O. mykiss)), even though young cells always contain neutral (i.e.

AB.2 –ve) glycoproteins. As cells age the glycoprotein content becomes more acid (i.e. AB.2 +ve).

Chloride cells play an important role in fish ionic regulation because they are involved in the active exchange of ions (Perry, 1997). Under stress conditions there is an increased permeability of the gills to water and ions mediated by elevated plasma catecholamine concentration (Wendelaar Bonga, 1997).

Therefore, an increased level of active ion transport (uptake or secretion) is necessary, regardless of the external salinity, and this explains the increase in the number of chloride cells that is often observed in stressed fish. In the present study a significant lice-associated increase in chloride cell numbers was apparent at 3 and 14 dpi. In addition, a lice-induced increase in the proportion of chloride cells migrating to the lamellae from the interlamellar area was observed, similar to the observations of Nolan et al (1999) in postsmolt Atlantic salmon (Salmo salar). Chloride cells on the lamellae appear to be larger, and have an increased surface area in contact with the water, than those in the interlamellar regions. This probably allows for an increased and more efficient level of ion exchange. It was also noted that the size of the chloride cells appeared to be greater, possibly due to less constriction than when the cells are located in the interlamellar area. Metz et al (2003) previously observed this phenomenon in common carp (Cyprinus carpio) following temperature induced stress. This is an important area for future work and further analysis of the collected samples will be performed using morphometric analyses in order to determine quantitatively if there is a lice-induced increase in the size of chloride cells.

The lice-induced increase in proliferating cells could explain why there often were few changes in actual cell numbers. Apoptotic or necrotic cells may have been replaced by differentiating cells at a similar rate as they disappeared. Increased cell turnover, particularly in chloride cells is indicative of stress, and can be inferred from an increase in the number of apoptotic, necrotic, and newly proliferated immature chloride cells (Wendelaar Bonga, 1997). An increase in cell turnover will lead to an increase in new, and therefore more efficient, chloride or mucus cells. It was originally intended to quantify apoptosis using the TUNEL procedure, but this proved to be impossible, due to lack of reproducible results. However, SEM analysis indicated qualitatively that there was an increase in apoptosis.

178

SUMBAWS

Final Report

It was observed in all cases with the SEM images, and in many of the cell counts by LM, that the control and low infestation groups were similar, with no observable stress-related changes. However, stressor effects were apparent in the medium infestation group which is best seen in the SEM images where deterioration of the skin and gill epithelium where evident. Epithelia of fish in the low infestation group remained similar to those of the control fish. This suggests that low levels of infestation did not evoke a clear stress response in sea trout smolts.

Wells et al. (2006), using the same fish as sampled in the present study, deduced a sublethal threshold level of 13 lice · fish-1, above which the lice have significant physiological effects. Analyses of skin and branchial epithelium do not allow an accurate prediction of a threshold level of lice, although a significant effect is apparent above the predicted threshold of 13 lice · fish-1. It can be concluded that the data concerning the epithelial parameters show that sea trout smolts demonstrate no sign of stress when infested with a low lice intensity (10 lice · fish-1), but there does appear to be a significant lice-induced effect of stress at medium levels of infestation (30 lice · fish-1).

These lice effects were not observed until the lice developed to mobile stages, at which point a breakdown of skin and gill epithelial integrity and changes in cell number, size and distribution were evident. These observations are therefore consistent with the 13 lice · fish-1 threshold reported by Wells et al. (2006).

References Berntssen, M.H.G., Kroglund, F., Rosseland, B.O., and Wendelaar Bonga, S.E., 1997. Responses of skin mucous cells to aluminium exposure at low pH in Atlantic salmon (Salmo salar) smolts. Canadian Journal of Fisheries and Aquatic Science 54, pp. 1039-1045.

Boulton, R.A., and Hodgson, J.F., 1995. Assessing cell proliferation: a methodological review. Clinical Science 88, pp. 119-130.

Buchmann, K., 1999. Immune mechanisms in fish skin against monogeneans – a model. Folia Parasitologica 46, pp. 1-9.

Burkhardt-Holm, P., Escher, M., and Meier, W., 1997. Waste-water management plants effluents cause cellular alterations in the skin of brown trout. Journal of Fish Biology 50, pp. 744-758.

Daborn, K., Cozzi, R.R.F., and Marshall, W.S., 2001. Dynamics of pavement cell-chloride cell interactions during abrupt salinity change in Fundulus heteroclitus. The Journal of Experimental Biology 204, pp. 1889-1899.

Ferguson, H.W., Morrison, D, Ostland, V.E., Lumsden, J., and Byrne, P., 1992. Responses of mucusproducing cells in gill disease of rainbow trout (Oncorhynchus mykiss). Journal of Comparative Pathology 106, pp. 255-265.

Foskett, J.K., and Hubbard, G.M., 1981. Hormonal control of chloride secretion by teleost opercular membrane. Annals of the New York Academy of Science 372, pp. 643.

Iger, Y., Balm, P.H.M., Jenner, H.A., and Wendelaar Bonga, S.A., 1995. Cortisol induces stress-related changes in the skin of rainbow trout (Oncorhynchus mykiss). General Comparative Endocrinology 97, pp.

188-198.

Metz, J.R., van den Burg, E.H., Wendelaar Bonga, S.E., and Flik, G., 2003. Regulation of branchial Na+/K+-ATPase in common carp Cyprinus carpio L. acclimated to different temperatures. The Journal of Experimental Biology 206, pp. 2273-2280.

Nolan, D.T., Reilly, P., and Wendelaar Bonga, S.E., 1999. Infection with low numbers of the sea louse Lepeophtheirus salmonis (Kroyer) induces stress-related effects in post-smolt Atlantic salmon (Salmo salar L.). Canadian Journal of Fisheries and Aquatic Science 56, pp. 947-959.

Nolan, D.T., 2000a. Skin responses of fish to stressors. PhD thesis, Katholieke Universiteit Nijmegen.

Nolan, D.T., 2000b. Juvenile Lepeophtheirus salmonis (Krøyer) affect the skin and gills of rainbow trout Oncorhynchus mykiss (Walbaum) and the host response to a handling procedure. Aquaculture Research 31, pp. 823-833.

179

SUMBAWS

Final Report

O’Byrne-Ring, N., Dowling, K., Cotters, D., Whelan, K., and MacEvilly, U., 2003. Changes in mucus cell numbers in the epidermis of the Atlantic salmon at the onset of smoltification. Journal of Fish Biology 63, pp. 1625-1630.

Perry, S.F., 1997. THE CHLORIDE CELL: Structure and function in the gills of freshwater fishes. Annual Review of Physiology 59, pp. 325-347.

Perry, S.F., 1998. Relationships between branchial chloride cells and gas transfer in freshwater fish.

Comparative Biochemistry and Physiology – Part A: Molecular and Integrative Physiology 1, pp. 9-16.

Pisam, M., Caroff, A., and Rambourg, A., 1987. Two types of chloride cells in the gill epithelium of a freshwater-adapted euryhaline fish: Lebistes reticulatus; their modifications during adaptation to saltwater.

The American Journal of Anatomy 179, pp. 40-50.

Pisam, M., and Rambourg, A., 1991. Mitochondria-rich cells in the gill epithelium of teleost fishes: An ultrastructural approach. International Review of Cytology 130, pp. 191-232.

Richman, N.H., Zaugg, W.S., 1987. Effects of cortisol and growth hormone on osmoregulation in pre- and desmoltified Coho salmon (Oncorhynchus kisutch). General and Comparative Endocrinology 65, pp. 189198.

Ruane, N.M., Carballo, E.C., and Komen, J., 2002. Increased stocking density influences the acute physiological stress response of common carp Cyprinus carpio (L.). Aquaculture Research 33, pp. 777-784.

Shephard, K.L., 1994. Functions for fish mucous. Reviews in Fish Biology and Fisheries 4, pp. 401-429.

Van der Salm, A.L., Nolan, D.T., Spanings, F.A.T., and Wendelaar Bonga, S.E., 2000. Effects of infection with the ectoparasite Argulus japonicus (Thiele) and administration of cortisol on cellular proliferation and apoptosis in the epidermis of common carp, Cyprinus carpio L., skin. Journal of Fish Diseases 23, pp. 173-184.

Van der Salm, A.L., Nolan, D.T., and Wendelaar Bonga, S.E., 2002. In vitro evidence that cortisol directly modulates stress-related responses in the skin epidermis of the rainbow trout (Oncorhynchus mykiss Walbaum). Fish Physiology and Biochemistry 27, pp. 9-18.

Wendelaar Bonga, S.E., 1997. The stress response in fish. Physiological Reviews 77, pp. 591-625.

Wells, A., Grierson, C.E., MacKenzie, M., Russon, I.J., Reinardy, H., Middlemiss, C., Bjørn, P.A., Finstad, B., Wendelaar Bonga, S.E., Todd, C.D. & Hazon, N. (2006) The physiological effects of simultaneous abrupt seawater entry and sea lice (Lepeophtheirus salmonis Krøyer) infestation of wild sea trout (Salmo trutta L.) smolts. Canadian Journal of Fisheries and Aquatic Sciences. submitted

180

SUMBAWS

Final Report

Paper 3: THE PHYSIOLOGICAL EFFECTS OF SIMULATANEOUS ABRUPT SEAWATER ENTRY AND SEA LICE (LEPEOPHTHEIRUS SALMONIS KRØYER) INFESTATION OF FARMED ATLANTIC SALMON (SALMO SALAR) SMOLTS

A. Wells, C. E. Grierson, M. MacKenzie, H. Reinardy, K. Zischka, C. D. Todd and N. Hazon (Manuscript) Abstract For farmed Atlantic salmon (Salmo salar L.) smolts (weight range 52-159g), the physiological consequences of abrupt transfer to seawater and simultaneous challenge with copepodid larvae of the sea louse, Lepeophtheirus salmonis Krøyer, were investigated in the laboratory. Analysis of osmoregulatory, metabolic and stress measures allowed the derivation of a sublethal threshold burden of L. salmonis, above which the host suffers significant physiological stress. Significant sea lice effects, consistent across all measured measures, were not apparent until L. salmonis developed to the ‘mobile’ preadult and adult stages.

Preadult L. salmonis caused significant increases in plasma chloride, osmolality, glucose, lactate and cortisol and a significant reduction in haematocrit. Piecewise linear statistical approaches allowed the determination of abrupt changes in these physiological measures, attributable to the intensity of L. salmonis infestation on individual fish, and the identification of overall threshold lice burdens. 20 L. salmonis · fish-1 was a consistent breakpoint across several physiological measures. This information will provide a valuable, objectively derived tool to aid in the formulation of management policy for Atlantic salmon smolts in the first year of aquaculture production.

Introduction Caligid copepod sea lice, Lepeophtheirus salmonis (Krøyer) and Caligus elongatus (Nordmann), are the primary marine ectoparasites of farmed and wild salmonids in the North Atlantic. The most recent estimate of the full economic cost (fish mortalities, harvest downgrades, chemotherapeutants and labour) of caligids to the global salmon aquaculture industry is ~US$100m annually (Johnson et al. 2004). Heavy sea lice infestation of the host fish is characterised by stress and osmoregulatory imbalance, which occur once the parasites moult from the sessile chalimus IV to the ‘mobile’ preadult and adult stages (Bjørn & Finstad 1997; Dawson et al. 1998; Finstad et al. 2000; Grimnes & Jakobsen 1996). Feeding activity by preadult/adult sea lice can result in increased skin damage (Bjørn & Finstad 1998; Nolan et al. 1999; Pike & Wadsworth 1999), which itself can be fatal to hatchery-reared sea trout and salmon (Bjørn & Finstad 1997; Dawson et al. 1998; Grimnes & Jakobsen 1996). Infestation of salmon smolts with large numbers of the chalimus larval stages leads to increased plasma levels of the stress hormone cortisol (Finstad et al. 2000).

Cortisol induces a range of physiological effects, including hydromineral imbalance, changes in intermediary metabolism, reduction of the immune response and reduced capacity to tolerate subsequent or additional stressors (Wendelaar Bonga 1997).

We recently identified a threshold sea louse intensity for wild sea trout (Salmo trutta L.) smolts of 13 L. salmonis · fish-1, above which significant negative physiological effects were observed (Wells et al. 2006).

There is, to date, no information on the physiological effects of early stages of L. salmonis on farmed salmon smolts following their transfer to seawater (SW). The potential effects of stress arising from infestation with L. salmonis presents key management and fish welfare/husbandry issues and sublethal detrimental effects could include reduced growth rate and immunocompetence (Wendelaar Bonga 1997).

Either or both of the latter will result in a reduction of fish quality at harvest. In this study, using laboratory experiments designed to mimic environmentally realistic circumstances, we investigated the time-course of physiological responses and identify sublethal threshold ectoparasite intensities for farmed Atlantic salmon smolts transferred abruptly to SW and simultaneously challenged with L. salmonis.

Materials and Methods Farmed salmon smolts (mean weight 97.4 g; range 52-159 g) were provided by Marine Harvest (Scotland) Ltd. and transported to the Gatty Marine Laboratory in a smolt tank containing constantly oxygenated fresh water (FW). On arrival, smolts were randomly split into duplicate experimental groups and 181

SUMBAWS

Final Report

held in brackish water (468 ± 9 mOsm.kg-1) in eight, 300 l aquarium tanks under natural photoperiod (18 h light: 6 h dark). Smolts were maintained in this manner for 7-13 d prior to experimentation. All fish were fed commercial fish pellets. Eighteen hours prior to L. salmonis infestation, the flow of FW was stopped and the SW flow was increased accordingly. The salinity of the tanks reached full SW (972 ± 3 mOsm.kg-1) within 3 h and the mean aquarium temperature was 13.6°C. Ovigerous female L. salmonis were collected from wild salmon caught in commercial nets at Montrose, East Scotland. Egg strings were removed from the sea lice and evenly distributed between six, 2 l conical flasks containing filtered SW (10°C). Flasks were continuously aerated to ensure that egg strings remained in constant motion. Nauplii began hatching within 24 h and were collected on a 40 µm filter and transferred to a 20 l plastic aquarium tank containing filtered, UV-sterilised, through-flowing SW. Larvae developed through the two naupliar moults to the infective copepodid stage. Small numbers of larvae were sampled daily and microscopically examined to monitor the rate of larval development. On the day of infestation, 2 to 4 d old copepodid larvae were concentrated in 1 l of SW by filtering through a 40 µm filter. The total number of copepodids was estimated from five replicate subsample counts.

At infestation, the volume of water in the experimental tanks was drained to one third (100 l) and the water supply turned off. The water was continuously aerated during the infestation to maintain a circulating current in the tanks. Copepodids were added in appropriate numbers to achieve the target infestation levels of ~10, ~15 and ~25 sea lice · fish-1 in duplicate tanks. These target intensities were chosen to include infestation intensities that were not likely to cause mortality, but which would definitely embrace a sublethal threshold level of infection. Duplicate control tanks (zero sea lice) were treated in an identical manner, but no copepodid larvae were added. Infestation progressed for 2 h in total darkness, following which the experimental tanks were refilled and the 18 h light: 6 h dark photoperiod regime was reinstated.

Infested and control fish were sampled at 7, 14, 21 and 28 days post infestation (dpi). Fish were netted individually from the experimental tanks, taking care to minimise disturbance to the remaining fish, and killed by overdose of 2-phenoxyethanol (2.5 ml l-1) using individual containers for each fish. On removal from the anaesthetic the spinal cord of each fish was severed using a scalpel blade. The anaesthetic was subsequently passed through a 40 µm filter in order to collect any dislodged L. salmonis. Five fish from each duplicate tank (10 fish for each sea louse intensity) were removed at each sampling date and blood and tissues were collected within 15 min. Blood samples (0.5 ml) were taken from the caudal blood sinus immediately following sacrifice and processed immediately (see below). The fish was then opened along the mid ventral line and the liver was removed and freeze clamped in liquid nitrogen to prevent glycogenolysis.

The frozen liver was placed in a pre-weighed vial and stored at -80ºC until analysis. Fish carcasses were placed in individual plastic bags and stored at 4ºC or -20ºC for subsequent microscopic determination of L. salmonis numbers. All dissection was performed on paper towelling and this also was bagged and subsequently examined for dislodged L. salmonis. The fork length and weight of all fish were recorded and expressed as the composite condition factor, Fulton’s K (weight in g/(length in cm)3 × 100). Ratio condition factors are sensitive to isometric variation and departures from the cube relationship, but the size range of experimental fish was narrow and Fulton’s K has been shown to have a high correlation with whole fish lipid content (Neff & Cargnelli 2004; Sutton et al. 2000).

Blood haematocrit was determined on fresh blood using a Compur microspin microhaematocrit centrifuge. Plasma was separated by centrifugation at 13,000 rpm for 1 min, and frozen at -80ºC until analysis. Plasma chloride was measured by colorimetric titration (Chloride analyser 925, Corning, UK).

Plasma osmolality was measured by freezing point depression (Roebling Osmometer, Camlab, UK). Plasma glucose concentration was measured in duplicate using a commercially available kit (Randox laboratories Ltd, Co Antrim, UK – ref: GL 2623), and absorbance was read at 490 nm. Plasma lactate concentration was measured in duplicate using commercially available lactate reagent (Randox laboratories Ltd, Co Antrim, UK – ref: LC 2389) and read at 540 nm. Both of the latter assays were modified for use in 96-well plates and absorbance was read on a plate reader (Biotech Instruments Inc, USA). Liver glycogen was assessed following hydrolysation by amyloglucosidase using the method of Keppler and Decker (1974). Plasma cortisol was measured by RIA, as described by van Anholt et al. (2003).

The body surface, fins and gills were separated analytically into a total of 11 zones as previously described (Bjørn & Finstad 1998) (Figure 1). These zones for all infected fish were examined for L. salmonis, and the frequencies of each developmental stage were recorded. L. salmonis at the sessile first/second chalimus stages and third/fourth chalimus stages were pooled into two groups, termed C1 and C3 respectively. Older (post-larval) mobile L. salmonis stages were categorised and their sex determined according to Johnson and Albright (1991) and Schram (1993). Statistical analysis was performed according to the methods described in Wells et al. (2006).

182

SUMBAWS

Final Report

Figure 1 The 11 zones on the body of the infected sea trout smolts examined for the presence of L. salmonis (Bjørn and Finstad 1998). A, head and operculum; B, anterior dorsal area of the body above the lateral line from the head to the posterior end of the dorsal fin; C, posterior dorsal area of the body above the lateral line from the posterior end of the dorsal fin to the caudal fin; D, anterior ventral area of the body below the lateral line from the head to the posterior end of the dorsal fin; E, posterior dorsal area of the body below the lateral line from the posterior end of the dorsal fin to the caudal fin; F, dorsal fin; G, pectoral fin; H, ventral fin; I, anal fin; J, caudal fin; K, gills. The anaesthetic solutions, plastic storage bags, and paper towelling on which dissection took place were also examined and are shown as zone L. Results Development of L. salmonis Prevalence was 100% on each sampling occasion and the proportion of sea lice at each developmental stage is summarised in Table 1. At 7 dpi 40% of sea lice were at the early chalimus stages whereas 60% had developed to the late chalimus stages. One week later, the majority of males had moulted to the second preadult stage, whereas all females were at the first preadult stage. At day 21, all males had attained adulthood though females were split between the second preadult stage and adults. By day 28 all lice were adults, with 35% gravid females. Because sea lice numbers also were assessed on mortalities, we were able to observe the first appearance of each of the sea lice stages on the fish. Mobile, preadult lice were observed on fish from 11 dpi onwards.

Table 1 Stage and frequency of major groups of sea lice at sampling of salmon. (Infected fish (n = 30 at 7, 14 & 21 dpi; n = 20 at 28 dpi) were sampled at day 7, 14, 21, 28 dpi. Stages are designated as follows: C1, first and second chalimus stages combined; C3, third and fourth chalimus stages combined; PAM1, first preadult male; PAF1, first preadult female; PAM2, second preadult male; PAF2, second preadult female; AM, adult male; AF, adult female; GF, gravid female.) dpi 7 14 21 28

Cop 40%

C1 60%

C3

Stage and Frequency PAM1 PAM2 PAF1 PAF2 6%

42%

AM

AF

GF

38% 52%

36% 13%

35%

51% 25%

The frequency distribution of sea lice on the body surface of sampled fish is shown in Table 2. At 7 dpi, when the majority of lice were still at the copepodid or early chalimus stages (Table 1), the majority of the sea lice were observed on the fins, with 17.5% attached to the dorsal fin (zone F; Figure 1). Only 1% of the sea lice were detached from the host (in the anaesthetic; zone L). Once L. salmonis attained the mobile stage (14-21 dpi), the majority of sea lice were found on the body surface, but a significant proportion were also found in the anaesthetic solution. At 28 dpi, when all the sea lice were adults, the majority of sea lice still attached to the fish were found on the head.

183

SUMBAWS

Final Report

Table 2 Frequency distribution (%) of sea lice on the body surface of sampled salmon. (n is the total number of lice in each group at that sample point. See Figure 1 for an explanation of the different body zones. The stage of lice at each sampling point is shown in Table 1.) dpi

A 7.1 44.3 40.4 54.1

7 14 21 28

Body Surface B C D 4.1 2.1 4.6 16.7 12.8 2.2 27.6 14.6 0.4 7.1 17.1 0

E 2.4 0.6 0.4 0.6

F 17.5 0.1 0 0

Fins G H 19.9 6.9 0.1 0 0 0 0 0

I 4.0 0 0 0

J 7.2 0.1 0 0

Gills K 23.3 0 0 0

Loose L 1.0 23.1 16.7 21.2

Mean lice intensity for each group is shown in Table 3. At 7 and 14 dpi the mean lice intensities were somewhat higher than the target infestation levels of ~10, ~30 and ~50 sea lice · fish-1.

Table 3 Mean number of lice per infestation group. Data are anti-logged mean lice numbers. dpi 7 14 21 28

Low 30.1 ± 0.1 21.8 ± 0.2 10.6 ± 0.1 2.3 ± 0.2

Lice Group Medium 40.1 ± 0.1 33.2 ± 0.1 17.2 ± 0.2 5.7 ± 0.2

High 72.1 ± 0.1 49.1 ± 0.2 24.5 ± 0.1

Mortalities There were no mortalities in the control group and only 2 mortalities occurred in the low sea lice intensity group. Eight mortalities occurred in the intermediate intensity group with 26 in the high intensity group. All mortalities occurred after L. salmonis had developed to the mobile stages (11 dpi). For reasons of animal welfare, any remaining fish in the high intensity group were sacrificed, and the experiment truncated, following sampling at 21 dpi.

Physiological Measures Mean values for the physiological measures investigated are shown in Figures 2-4. No significant sea lice effects were observed at 7 dpi for either haematocrit or plasma chloride (Figure 2), but there was a significant sea lice effect at 7 dpi for plasma osmolality (p = 0.004). Significant sea lice effects were observed for all three osmoregulatory measures at 14, 21 and 28 dpi.

There was no evidence of a sea louse effect for plasma glucose or for liver glycogen at 7 dpi (Figure 3), but a significant effect was observed for plasma lactate at this time point (p = 0.039). Significant sea lice effects were observed for plasma glucose at 14, 21 and 28 dpi and for plasma lactate at 21 dpi. For plasma cortisol at 14, 21 and 28 dpi (Figure 4) there were significant sea louse effects, attributable largely to the high intensity group at 14 and 21 dpi.

184

SUMBAWS

Final Report

50

Haematocrit (%)

45 40 35 30 25 20

Plasma Chloride (mmol/l)

200 180 160 140 120

Plasma Osmolality (mOsm/kg)

100

450 400 350 300

Control Infected (Low) Infected (Medium) Infected (High)

250 200 0

7

14

21

28

35

Days Post Infection (dpi) Figure 2 Mean values ± SEM for haematocrit (top) plasma chloride (middle) and plasma osmolality (bottom) at four fixed lice loadings (Control (no lice) – solid black line and ○, 10 L. salmonis · fish-1 – broken line and ●, 30 L. salmonis · fish-1 – broken line and ∆, 50 L. salmonis · fish-1 – dotted line and ▲).

Data are slightly offset for clarity.

185

SUMBAWS

Final Report

Plasma Glucose (mmol/l)

10 8 6 4 2 0

Plasma Lactate (mmol/l)

5 4 3 2 1

Liver Glycogen (mg/g wet wt tissue)

0

50 Control Infected (Low) Infected (Medium) Infected (High)

40 30 20 10 0 0

7

14

21

28

35

Days Post Infection (dpi)

Figure 3 Mean values Âą SEM for plasma glucose (top) plasma lactate (middle) and liver glycogen (bottom) at four fixed lice loadings. Details as in Figure 2.

186

SUMBAWS

Final Report

250 Control (0 lice) Infected (low) Infected (medium) Infected (high)

Plasma Cortisol (ng/ml)

200

150

100

50

0 0

7

14

21

28

35

Days Post Infection (dpi)

Figure 4 Mean values ± SEM for plasma cortisol at four fixed lice loadings. Details as in Figure 2.

Piecewise linear modelling Standard linear terms were preferred for the effects of sea lice abundance on plasma chloride and plasma glucose at 14 dpi. On some occasions no breakpoint could be predicted because no covariates justified their inclusion in the model. Breakpoints identified by the AIC are shown in Table 4. A low breakpoint of 14 L. salmonis · fish-1 at 14 dpi and 13 L. salmonis · fish-1 at 21 dpi was predicted for haematocrit. Breakpoints for plasma osmolality and plasma cortisol were somewhat higher at 14 dpi (43 and 39 sea lice respectively), but at 21 dpi consistent breakpoints between 20 and 24 L. salmonis · fish-1 were identified. Breakpoints for 5 of the 7 physiological measures were predicted at 20 or 21 L. salmonis · fish-1. Plots of the predicted breakpoints are shown in Figure 5. Following the predicted breakpoints, it is relevant to note that there was a decrease in haematocrit and an increase for plasma osmolality, chloride, glucose and lactate, as might be expected from the experimental outcomes in Figures 2-5.

Table 4. Predicted breakpoints for all physiological markers.

Response Haematocrit Chloride Osmolality Glucose Lactate Glycogen Cortisol

14 dpi 14 Linear 43 Linear 39

21 dpi 13 21 21 20 24 20 20

Breakpoints identified using piecewise linear models. In the interests of parsimony the AIC statistic was used to discriminate between a standard linear term and a piecewise linear term for lice. Where ‘Linear’ is included in the table standard linear terms were preferred for lice.

187

SUMBAWS

Final Report

Figure 5 Results from the piecewise linear models.

188

SUMBAWS

Final Report

Discussion The present study is the first to model a sublethal sea lice threshold burden for Atlantic salmon smolts, above which significant physiological stress occurs. This information, whilst not directly relevant to wild fish due to the considerable size difference between farmed and wild smolts when they first enter SW, is a valuable addition to the detrimental levels of infestation predicted for wild sea trout smolts (Wells et al.

2006) and may also be useful to the aquaculture industry as a management tool.

Development of L. salmonis on Atlantic salmon smolts was similar to that previously described for wild sea trout smolts (Wells et al. 2006), reflecting the similar ambient water temperatures in both studies.

Infestation success also closely approximately previously described experiments (Bjørn & Finstad 1998; Wells et al. 2006). However, in the present study a high proportion (23%) of copepodid and chalimus larvae were found on the gills, which was not the case for sea trout. Furthermore, the majority of adult L. salmonis were observed on the head of salmon smolts in the present study, whereas they tended to be located on the posterior body zones in the sea trout.

The mortalities in the high intensity group most likely occurred due to the higher than expected sea lice intensities which occurred at 7 and 14 dpi. It should be noted that variability in infestation levels are an inherent problem with type of experiment, but at 21 dpi, mean lice intensity was extremely close to the target infestation levels.

Osmoregulatory Measures From 14 dpi onwards when L. salmonis were at the mobile stages there was a significant decrease in haematocrit for all infested groups. A similar effect has been observed in previous studies on Atlantic salmon (Grimnes & Jakobsen 1996) and sea trout (Bjørn & Finstad 1997). In those studies, decreased haematocrit was attributed to osmotic shrinkage of red blood cells, resulting from increased plasma ions, coupled with anaemia due to leakage of blood components from sea lice damaged areas of the body surface.

Lesions which resulted from the feeding activity of mobile lice were observed in the present study, particularly for the high sea lice intensity group. However, haematocrit also was significantly reduced in fish from the intermediate and low intensity groups, for which no sea lice-induced lesions were noted.

The patterns observed for plasma chloride and plasma osmolality were similar, with significant sea lice effects observed at 14, 21 and 28 dpi. In addition, a further sea louse effect was observed for plasma osmolality at 7 dpi, presumably due to the extremely small variation in this study. There were significant increases in plasma chloride and osmolality in the high lice intensity group which were obvious by 14 dpi, once L. salmonis had reached the mobile preadult stages. This response perhaps was attributable to the increased feeding activity of the lice causing a loss of skin integrity (Bjørn & Finstad 1998; Pike & Wadsworth 1999). A similar effect of the mobile L. salmonis stages was observed for sea trout smolts (Wells et al. 2006). Disturbance of water and ion homeostasis is mediated by cortisol and is a common aspect of stress in fish which, in addition to lice-induced skin damage, may arise from structural changes in the gills. Additionally, cortisol may mediate structural changes in the gills, such as necrosis of branchial epithelial cells or epithelial lifting, often leading to disturbance of water and ion homeostasis (Wendelaar Bonga 1997).

Plasma cortisol An elevation in plasma cortisol is the most widely used measure or indicator of stress in fish, and cortisol concentration may remain elevated during chronic stress (Wendelaar Bonga 1997). Cortisol has a broad range of effects in fish including the regulation of hydromineral balance and energy balance and it is involved in reducing the immune response (Wendelaar Bonga 1997). Sea lice effects on plasma cortisol were observed from 14 dpi onwards, reflecting chronic stress arising from the movement and increased feeding activity of mobile stages of L. salmonis, and these effects are consistent with the well characterised stress response (Wendelaar Bonga 1997). A similar cortisol response had been described in studies on hatchery-reared (Bjørn & Finstad 1997) and wild (Wells et al. 2006) sea trout smolts. However, in the present study there was no significant increase in plasma cortisol when the sea lice were at the chalimus stage (7 dpi) and this was in marked contrast to previous studies of salmon smolts (Finstad et al. 2000).

Metabolic Measures An increase in metabolic rate often is associated with stress and in order to satisfy increased energy demands, fish mobilise substrates to fuel cellular processes (Mommsen et al. 1999). Hyperglycemia (elevation of plasma glucose) and decreases in liver glycogen have been reported during stress (Wendelaar 189

SUMBAWS

Final Report

Bonga 1997) and both hyperglycemia (Barton 2000) and elevated plasma lactate concentration (Farbridge & Leatherland 1992; Iversen et al. 1998) have been reported in response to both acute and chronic stress.

Lactate, resulting from the breakdown of muscle glycogen, may be released into the bloodstream due to aerobic conditions following exhaustive exercise (Wood 1991).

Whilst there were no obvious pattern in relevant metabolic measures related to increasing sea lice intensity, plasma glucose and lactate were significantly elevated in the high sea lice intensity group.

Following the development of L. salmonis to the mobile stages (at 11 dpi) fish in the high lice intensity tanks in particular showed distinct ‘flashing’ behaviour, and even jumped clear of the water surface. The increases in plasma glucose and lactate could perhaps reflect an increased energy demand to the animals arising from this behaviour and due to the stress of L. salmonis infection.

Identification of threshold parasite intensities A piecewise linear approach was employed to enable the identification of abrupt changes in the response of the various measures to increasing sea lice intensities. Such breakpoints will provide evidence of threshold parasite intensities which induce marked changes in physiological responses indicative of stress.

This same approach was applied to experimental data for sea trout smolts and 13 sea lice · fish-1 was a consistent analytical outcome (Wells et al. 2006). In the present study, breakpoints for salmon smolts were identified at14 dpi, but consistently we recorded breakpoints of 20-21 L. salmonis · fish-1 across most of the physiological measures at 21 dpi.

The large size of the salmon smolts used in the present study, which were reared for commercial ongrowing in salmon cages, makes extrapolation and comparison of threshold values to wild fish problematic.

Wild salmon smolts typically enter SW at ~30g and sea lice-mediated responses are likely to be allometrically related to body size. However, information on the number of lice which are likely to cause sublethal physiological stress to farmed Atlantic salmon smolts is lacking, and such stress is almost certain to exert detrimental effects on growth, subsequent immunocompetence and ultimately value at harvest. The results of the present study suggest that L. salmonis intensities exceeding 20 mobile lice · fish-1 would be expected to have significant deleterious physiological consequences on the fish. Whilst it is very unlikely that farmed post-smolts in their first year at sea would be infested to such high intensities in well managed farms, this objectively derived threshold intensity does provide the industry with a clear target for fish welfare and husbandry.

References Barton, B. A. 2000. Salmonid fishes differ in their cortisol and glucose responses to handling and transport stress. N. Am. J. Aqualcult. 62:12-18.

Bjørn, P. A., and Finstad, B. 1997. The physiological effects of salmon lice infection on sea trout post smolts. Nordic J. Freshw. Res. 73:60-72.

Bjørn, P. A., and Finstad, B. 1998. The development of salmon lice (Lepeophtheirus salmonis) on artificially infected post smolts of sea trout (Salmo trutta). Can. J. Zool.-Rev. Can. Zool. 76:970-977.

Dawson, L. H. J., Pike, A. W., Houlihan, D. F., and McVicar, A. H. 1998. Effects of salmon lice Lepeophtheirus salmonis on sea trout Salmo trutta at different times after seawater transfer. Dis. Aquat. Org. 33:179-186.

Farbridge, K. J., and Leatherland, J. F. 1992. Plasma growth hormone levels in fed and fasted rainbow trout (Oncorhynchus mykiss) are decreased following handling stress. Fish Physiol. Biochem. 10:67-73.

Finstad, B., Bjorn, P. A., Grimnes, A., and Hvidsten, N. A. 2000. Laboratory and field investigations of salmon lice Lepeophtheirus salmonis (Kroyer) infestation on Atlantic salmon (Salmo salar L.) post-smolts.

Aquac. Res. 31:795-803.

Grimnes, A., and Jakobsen, P. J. 1996. The physiological effects of salmon lice infection on post- smolt of Atlantic salmon. J. Fish Biol. 48:1179-1194.

Iversen, M., Finstad, B., and Nilssen, K. J. 1998. Recovery from loading and transport stress in Atlantic salmon (Salmo salar L.) smolts. Aquaculture 168:387-394.

Johnson, S. C., and Albright, L. J. 1991. The Developmental Stages of Lepeophtheirus Salmonis (Kroyer, 1837) (Copepoda, Caligidae). Can. J. Zool.-Rev. Can. Zool. 69:929-950.

190

SUMBAWS

Final Report

Johnson, S. C., Treasurer, J. W., Bravo, S., Nagasawa, K., and Kabata, Z. 2004. A review of the impact of parasitic copepods on marine aquaculture. Zool. Stud. 43:229-243.

Keppler, D., and Decker, K. 1974. Glycogen: determination with amyloglucosidase In Methods of Enzymatic analysis. Edited by H. U. Bergmeyer. Academic Press.

Mommsen, T. P., Vijayan, M. M., and Moon, T. W. 1999. Cortisol in teleosts: dynamics, mechanisms of action, and metabolic regulation. Rev. Fish. Biol. Fish. 9:211-268.

Neff, B. D., and Cargnelli, L. M. 2004. Relationships between condition factors, parasite load and paternity in bluegill sunfish, Lepomis macrochirus. Environ. Biol. Fishes 71:297-304.

Nolan, D. T., Reilly, P., and Wendelaar Bonga, S. E. 1999. Infection with low numbers of the sea louse Lepeophtheirus salmonis induces stress-related effects in postsmolt Atlantic salmon (Salmo salar). Can. J. Fish. Aquat. Sci. 56:947-959.

Pike, A. W., and Wadsworth, S. L. 1999. Sealice on salmonids: their biology and control. Advances in Parasitology 44:233-337.

Schram, T. A. 1993. Supplimentary descriptions of the developmental stages of Lepeophtheirus salmonis (Kroyer, 1837) (Copepoda: Caligidae) In Pathogens of Wild and Farmed Fish: Sea Lice. Edited by G. A. Boxshall, and D. Defaye. Ellis Horwood, Chichester.

Sutton, S. G., Bult, T. P., and Haedrich, R. L. 2000. Relationships among fat weight, body weight, water weight, and condition factors in wild Atlantic salmon parr. Trans. Am. Fish. Soc. 129:527-538. van Anholt, R. D., Spanings, T., Koven, W., and Wendelaar Bonga, S. E. 2003. Effects of acetylsalicylic acid treatment on thyroid hormones, prolactins, and the stress response of tilapia (Oreochromis mossambicus). Am. J. Physiol.-Regul. Integr. Comp. Physiol. 285:R1098-R1106.

Wells, A., Grierson, C. E., MacKenzie, M., Russon, I. J., Reinardy, H., Middlemiss, C., Bjørn, P. A., Finstad, B., Wendelaar-Bonga, S. E., Todd, C. D., and Hazon, N. 2006. The physiological effects of simultaneous abrupt seawater entry and sea lice (Lepeophtheirus salmonis Krøyer) infestation of wild sea trout (Salmo trutta L.) smolts. Can. J. Fish. Aquat. Sci. submitted Wendelaar Bonga, S. E. 1997. The stress response in fish. Physiol. Rev. 77:591-625.

Wood, C. M. 1991. Acid-base and ion balance, metabolism, and their interactions, after exhaustive excercise in fish. J. Exp. Biol. 160:285-308.

191

SUMBAWS

Final Report

Workpackage 6 Lead Partner: NINA Participating Partners: USTAN, NIFA, NCFS OBJECTIVES To study the effects of salmon lice infestation on survival, growth and life history pattern in hatcheryreared populations of Atlantic salmon.

• Established protocols on how to protect salmonids from salmon lice infection by treatment with Substance EX. • Established research stations on both river systems with access to specialist staffing, equipment and infrastructure.

• Established fish traps on the rivers conferring complete census control of the populations.

DELIVERABLES • • •

Evaluation of the effects of salmon lice infestations on survival, growth and life history strategies in two Norwegian populations of hatchery-reared Atlantic salmon.

Substance EX as a partial remedial action to enhance survival of Atlantic salmon post-smolts.

These data will be complementary to those obtained in WP 7 for a Scottish system.

The efficacy of hatchery-reared fish as a remedial action to regenerate wild salmonid stocks.

EXECUTIVE SUMMARY Since 1996, individually Carlin-tagged wild and hatchery-reared Atlantic salmon and sea trout smolts have been treated with Substance EX (which is the subject of a pending patent and is not related to any previously applied sea lice treatments). Preliminary results from these field-experiments were positive and have indicated that salmon lice in Norway may interfere with both the survival and growth of fish at sea.

These investigations have, however, long been hampered by low tag recapture rates, primarily because they have been undertaken in commercially exploited systems. In the present project we were able, for the first time, to include experiments on the only two Norwegian watercourses with permanent fish traps. This offered complete control and census of the populations and an evaluation of the effectiveness of a laboratory-based treatment protocol under field conditions. In the first and second years we released hatchery-reared, “protected” (Substance EX treated) and unprotected control groups of 2000 Atlantic salmon smolts from the two fish traps on river systems of contrasting farming activity, and subsequently examined sea lice infestation pressure (Ims, Høgsfjord; Talvik, Altafjord). The fish were individually Carlin-tagged, and the traps provided total control over ascending/returning salmonids. Salmon remain at sea for at least one year following tagging, and some for 2+ years, hence the requirement for repeated and longer term trapping of fish.

During this experiment, protocols for how to protect salmonids from salmon lice infection by treatment with substance EX were developed in collaboration with Pharmaq (former Alpharma) and with USTAN and CFB (WP7). Research stations at both Norwegian river systems provided access to specialist staffing, equipment, and infrastructure and the complete control of fish populations resulted in considerable success.

Results from these field-experiments have indicated that salmon lice in Norway may interfere with both the survival and especially the growth of fish at sea. We still need to continue to monitor recaptures until summer/autumn 2006 before we can summarize all results (releases from 2004). Data will be split into time of recaptures, where fish were recaptured and then related to fish growth and survival. For the Imsa and Talvik releases the results of the present experiment so far show that whilst Substance EX-treated Atlantic salmon smolts were significantly longer and heavier on return than untreated smolts, there was no 192

SUMBAWS

Final Report

significant increase in condition. Increases in size and condition may have an influence on the spawning success of the fish and therefore affect future generations. However, further recaptures are needed to draw the final conclusions. The plan is to publish all the data relating to Substance EX in 2006 or 2007, either separately or as a summary, together with partners from USTAN (Scotland) and CFB (Ireland). The compilation of the data concerning the efficacy of Substance EX and SLICE速 in providing smolts with protection from sea lice infestation will be all the more robust in embracing multiple years and different stocks in Norway, Scotland and Ireland.

CONCLUSION The description of the work from the original grant proposal has been successful. For the Norwegian (Imsa and Talvik) releases in the present experiment, Substance EX-treated Atlantic salmon had a significantly better growth than that shown by untreated smolts. However, further recaptures in 2006 are needed before the final conclusions can be drawn. In years with high salmon lice infection pressure in fjord systems, the use of this protection may increase survival in salmonid post-smolts during their stay at sea and may also offer a tool to regenerate small identifiable wild salmonid stocks in areas where populations have been compromised partly due to salmon lice infestations. The economics and practicalities of such releases and inventory management are discussed further in WP8 Final Report.

193

SUMBAWS

Final Report

Paper 1: THE EFFECTS OF PROTECTION TO SALMON LICE INFESTATION ON SURVIVAL, GROWTH AND LIFE HISTORY PATTERN IN HATCHERY-REARED POPULATIONS OF ATLANTIC SALMON B. Finstad, P.A. Bjørn, C.D. Todd, N. Hazon, A. Wells, P.G. Gargan and B. Martinsen (Manuscript) Introduction Since 1996, individually Carlin-tagged wild and hatchery-reared Atlantic salmon and sea trout smolts have been treated with Substance EX. Preliminary results from these field-experiments are positive and have indicated that salmon lice in Norway may interfere with both the survival and growth of fish at sea. These investigations have, however, been hampered by low tag recapture rates, primarily because they have been performed in commercially exploited systems. This project will allow for the first time experiments on the only two Norwegian watercourses with permanent fish traps and which offer complete control and census of the populations.

In the first and second years we will release hatchery-reared, protected and unprotected groups of 2000 Atlantic salmon from the two fish traps on river systems of contrasting farming activity and thence lice infestation pressure (Ims, Høgsfjord; Talvik, Altafjord). The fish will be individually Carlin-tagged, and the trap will provide total control over ascending/returning salmonids. Salmon will be at sea for at least one year following tagging, hence the requirement for repeated and longer term trapping of fish.

Materials and Methods Two groups Atlantic salmon smolts were Carlin-tagged and released below the fish trap in May at NINA’s Research Station at Ims in Rogaland in 2003 and 2004 (Figures 1a and 2). One group (n = 2000) was treated with Substance EX while the other (control) group (n = 3000) was unprotected. The same treatments (group EX-treated (n = 2029) and control group (n = 2215)) were performed at the Talvik Research Station in Finnmark (Figures 1b and 2). Fish were released below the fish in the trap in the River Halselva by the end of June. The prophylactic Substance EX (Pharmaq, Norway), protects the fish against salmon lice infection for up to 16 weeks, have low toxicity for warm blooded animals and humans, and its function is to prevent synthesis of chitin in salmon lice (Bernt Martinsen, Pharmaq, pers. comm.).

The aim here was to quantify differences in sea growth and recapture rates (=marine survivorship) between the two treatments. Prior to release the fish were given a standardized seawater challenge test to show that they were osmoregulatorily ready for release. According to Sigholt & Finstad (1990), smolt quality was good (plasma chloride levels at release <160 mM) at both locations. Recaptures from the smolts released in the River Imsa and in the River Halselva in 2003 and 2004 are given in the Tables 1-8 below.

a)

b)

Figure 1 a) River Imsa fish trap, (b) River Halselva (Talvik) fish trap

194

SUMBAWS

Final Report

Figure 2 Releasing sites at Ims (River Imsa) and Talvik (River Halselva).

Table 1 Recaptures of returning adults (1SW) Atlantic salmon in the river Imsa, other rivers and in the sea for salmon smolts released in the River Imsa in 2003 (Chi square test: No significant difference. χ²= 0.025, df=1, p = 0.87. Only one 2SW Atlantic salmon was recaptured in 2004 (750 mm, 3400 g)) Year

Treatment

2003 2003 Total

EX Control

Numbers released 2000 3000 5000

River Imsa 33 44 77

Recaptures Other rivers Sea 2 7 10 7 12 14

Total 42 61 103

% 2.1 2.03 2.06

Table 2 Length (mm) and weight (g) of recaptured returning adults (1SW) Atlantic salmon from the releases in the River Imsa in 2003 (L=length in mm; Wt= weight in g; WR=relative mass index. Length at release refers only to those fish recaptured. *There were significant differences in length at release (t-test; p < 0.0001), length at recapture (t-test; p = 0.0001) and weight at recapture (t-test; p = 0.0008). Only one 2SWAtlantic salmon was recaptured in 2004 (750 mm, 3400 g)) Year 2003 2003 Total

Treatment EX Control

No. recaptured 42 61 103

L (release) 171.3* ± 2.5 151.1 ± 1.5

L (recapture) 588.8* ± 67.2 539.3 ± 57.2

Wt (recapture) 1720.0* ± 615.6 1371.8 ± 526.7

WR 0.71 ± 0.02 0.73 ± 0.02

Table 3 Recaptures of returning adults (1SW) Atlantic salmon in the river Imsa, other rivers and in the sea for smolts released in the River Imsa in 2004 (Chi-square test: Too few fish for testing.)

Year

Treatment

2004 2004 Total

EX Control

Numbers released 2000 3000 5000

River Imsa 2 2 4

Recaptures Other river Sea 0 0 0 2 0 2

Total 2 4 6

% 0.07 0.2

195

SUMBAWS

Final Report

Table 4 Length (mm) and weight (g) in recaptured returning adults (1SW) Atlantic salmon from the releases in the River Imsa in 2004 (L=length in mm; Wt= weight in g; WR=relative mass index. Length at release refers only to those fish recaptured. *There was a significant difference in weight at recapture (t-test; p = 0.023)) Year 2004 2004 Total

Treatment EX Control

No. recaptured 2 4 6

L (release) 142.5 ± 8.5 155.0 ± 9.0

L (recapture) 513.0 ± 45.3 538.3 ± 7.2

Wt (recapture) 1022.5* ± 98.3 1441.0 ± 183.6

WR 0.69 ± 0.10 0.77 ± 0.01

Table 5 Recaptures of returning adults (1SW) Atlantic salmon in the river Halselva, other rivers and in the sea for Atlantic salmon smolts released in the River Halselva in 2003 (Chi-square test: No significant differences. χ²=0.0078, df=1, p = 0.92) Year

Treatment

2003 2003 Total

EX Control

Numbers released 2029 2215 4244

Recaptures Other rivers Sea 0 3 0 5 0 8

River Halselva 4 3 7

Total 7 8 15

% 0.35 0.36 0.35

Table 6 Length (mm) and weight (g) in recaptured returning adults (1SW) Atlantic salmon from the releases in the River Halselva in 2003 (L=length in mm; Wt= weight in g; WR=relative mass index. Length at release refers only to those fish recaptured. *There was a significant difference in length at recapture (t-test; p = 0.0174) and weight at recapture (t-test; p = 0.0352)) Year 2003 2003 Total

Treatment EX Control

No. recaptured 7 8 15

L (release) 190.3 ± 2.5 182.9 ± 3.1

L (recapture) 609.4* ± 35.8 557.9 ± 37.2

Wt (recapture) 2304.0* ± 515.0 1816.5 ± 341.7

WR 0.86 ± 0.04 0.91 ± 0.04

Table 7 Recaptures of returning adults (2SW) Atlantic salmon in the river Halselva, other rivers and in the sea for Atlantic salmon smolts released in the River Halselva in 2003 (Chi-square test: No significant differences. χ²=0.340, df=1, p = 0.56) Year

Treatment

2003 2003 Total

EX Control

Numbers released 2029 2215 4244

River Halselva 0 2 2

Recaptures Other rivers Sea 0 3 0 3 0 6

Total 3 5 8

% 0.15 0.23 0.19

Table 8 Length (mm) and weight (g) in recaptured returning adults (2SW) Atlantic salmon from the releases in the River Halselva in 2003 (L=length in mm; Wt= weight in g; WR=relative mass index. Length at release refers only to those fish recaptured. *There was a significant difference in length at release (t-test; p = 0.032)) Year 2003 2003 Total

Treatment EX Control

No. recaptured 3 5 8

L (release) 192.7* ± 4.7 182.2 ± 1.2

L (recapture) 883.3 ± 65.1 719.0 ± 175.1

Wt (recapture) 8733.3 ± 1662.3 4677.6 ± 3599.0

WR 0.93 ± 0.03 0.74 ± 0.10

196

SUMBAWS

Final Report

As seen from the data we still need to monitor recaptures until 2006 before we can summarize all results, including the return of multi sea-winter (MSW) fish. Both experimental releases showed no significant difference in recapture rates for the two groups (Tables 1, 3, 5 and 7). However, for the Imsa and Halselva releases, EX-treated Atlantic salmon had significantly better growth than did untreated smolts (Tables 2 and 6). Further recaptures are needed to provide a clear answer to this, but the data strongly indicate that growth and survival are enhanced for treated smolts. Even if numbers of treated fish show no significant increase, the improved sea growth of treated fish will have important consequences for individual fecundity and egg deposition.

Discussion Recapture results from these field-experiments are positive and have indicated that salmon lice in Norway may interfere with both the survival and growth of fish at sea (Finstad and Birkeland, 1997; Finstad and Jonsson, 2001; present results). These investigations have, however, long been hampered by low tag recapture rates, primarily because they have been undertaken in commercially exploited systems. In the present project we were able, for the first time, to include experiments on the only two Norwegian watercourses with permanent fish traps and which offered complete control and census of the populations and an evaluation of the effectiveness of a laboratory-based treatment protocol under field conditions.

We still need to monitor recaptures until 2006 before we can summarize all results (releases from 2004).

Data will be split into time of recaptures, where fish were recaptured and then related to fish growth and survival. For the Imsa and Talvik releases the results of the present experiment so far show that whilst Substance EX-treated Atlantic salmon smolts were significantly longer and heavier on return than untreated smolts, there was no significant increase in condition. In addition, for the Imsa releases in 2003, there were sufficient returns to the fish trap to allow sex-specific differences in subgroups of control and treated fish to be examined. Returning summer 1 SW males tend to be longer and heavier than females, but no difference in WR is expected between the sexes and within years. Information on sex could only be collected from fish which returned to the trap, because this information was not collected from fish caught by rod and line or in the commercial fishery. At Imsa, treated male fish were significantly longer (p = 0.0001) and heavier (p = 0.0001) than their respective controls, but showed no significant difference in WR. Treated female fish were not significantly longer or heavier than control female fish; however treated female fish actually returned at a significantly lower condition (p = 0.0089). Increases in size and condition may have an influence on the spawning success of the fish and therefore affect future generations. However, further recaptures are needed to draw the final conclusions.

References Bjørn, P.A., Finstad, B. and Kristoffersen, R. (2005). Registreringer av lakselus på laks, sjøørret og sjørøye i 2004 (Registrations of salmon lice on Atlantic salmon, sea trout and Arctic charr in 2004. NINA Rapport 60, 1-26.

Carlin, B. (1955). Tagging of salmon smolts in the River Lagan. Report from the Institute for Freshwater Research, Drottningholm 36, 57-74.

Finstad, B. and Birkeland. (1997). Salmon lice infestations in orally treated and non-treated sea trout. ICES CM 1997/MR: 4, 1 pp.

Finstad, B. and Jonsson, N. (2001). Factors influencing the yield of smolt releases in Norway. Nordic Journal of Freshwater Research 75, 37-55.

Heuch, P.A., Bjørn, P.A., Finstad, B., Holst, J.C., Asplin, L. and Nilsen, F. (2005). A review of the Norwegian ‘National Action Plan Against Salmon Lice on Salmonids’: The effect on wild salmonids.

Aquaculture 246, 79-92.

Jensen, A. (ed.). 2004. Geografisk variasjon og utviklingstrekk i norske laksebestander (Geographical variation and population trends in Norwegian Atlantic salmon). NINA Fagrapport 80, 1-79.

Sigholt, T. and Finstad, B. (1990). Effect of low temperature on sea-water tolerance in Atlantic salmon (Salmo salar) smolts. Aquaculture 84, 167-172.

197

SUMBAWS

Final Report

Workpackage 7 Lead Partner: USTAN Participating Partners: CFB, NINA, NIFA OBJECTIVES • Assessment of the efficacy of Substance EX in protecting wild smolts from initial infestations of sea lice.

• Use of existing traps for intercepting (i) all downstream-migrating sea trout and salmon smolts, and (ii) all upstream migrating juvenile/adult sea trout and adult salmon on the Manse Lochs system in Scotland and Gowla and Inver river systems in Ireland. Complete census, and PIT-tagging of samples of the salmon and sea trout smolt runs in each downstream migratory season (April/May).

• Treatment of subsamples of tagged fish (both species) with Substance EX (prophylaxis against sea lice infestation).

• Trap recapture of any “premature returning” (lice infested) juvenile sea trout within the first few days or weeks of their migration to seawater; trap recapture of naturally-returning immature ‘finnock’ in the summer/autumn prior to their overwintering in freshwater.

• Recapture of adult salmon, returning after one sea-winter, in the summer following their downstream migration. First returns are expected in July/August 2004 following tagging in April/May 2003.

DELIVERABLES At all sites detailed quantification and statistical analysis of: • full censuses, seasonal timing and density of the natural smolt runs – both species, all years.

• adult salmon and mature sea trout runs in the Irish and Scottish systems. These fish are the natural broodstock and the basis of any rod fishery; detailed numerical data will be required to set permit process for the developing Scottish fishery and the levels of sustainable fishing effort.

• the extent of “premature return” of juvenile sea trout. Such post-smolts inevitably are heavily infested with juvenile sea lice and return to freshwater within a few weeks or days of first migration. Few of these survive but the natural return of immature finnock, to overwinter in freshwater, will also be quantifiable.

• growth and successful spawning history of adult sea trout and salmon (year 2 onwards), following first tagging. Sea trout, but not salmon, often return repeatedly to spawn.

• the efficacy of Substance EX in protecting wild smolts of both species from sea lice infestation in their first 12-16 weeks at sea. Comparisons will be made of the sea growth (weight increment) and survivorship of treated versus control tagged smolts.

EXECUTIVE SUMMARY Preliminary studies in Norway have demonstrated that releases of Atlantic salmon and sea trout smolts, which were protected from sea lice infestation by prophylactic treatment, resulted in higher recaptures (0.90%) compared to unprotected fish (0.03%) in areas with intense fish farming activity. The aim of the present study was to assess the efficacy of prophylactic treatment (Substance EX; bath treatment for 30 minutes) in wild sea trout smolts intercepted from small, clearly defined Scottish loch systems (Paper 1), and to assess the efficacy SLICE® for sea trout and salmon smolts in Ireland (Paper 2).

Full censuses of the sea trout smolt runs were completed for the Manse Loch System (2003-2005) and the Shieldaig System (2004-2005) in Scotland (Paper 1). In 2003, 649 sea trout smolts were trapped in the Manse Loch System, of which 590 were PIT tagged and included in the trial. However, in 2004, only 224 sea trout smolts were trapped and PIT tagged and in 2005 only 8 fish were trapped. At Shieldaig, 837 fish were included in the trial in 2004, and 1171 were included in 2005. For much of 2003, which was a particularly dry year, there was insufficient water flowing through the Manse trap to allow the fish to move either downstream or upstream. This almost certainly was a factor in the low return of fish in 2003, because fish were regularly seen in the pool below the trap. In addition, at times of high water flow is was also 198

SUMBAWS

Final Report

evident that fish were able to pass over the top of the trap in either direction. In all three study years it was probable that numerous downstream-migrating smolts and returning fish were not trapped for this reason.

Indeed, several overwintered finnock, which were tagged in 2003 but not recorded on their upstream migration, were recorded in the downstream trap in 2004.

For these reasons, the study was extended to Shieldaig in 2004 and 2005, and despite a disappointing migratory return at both sites in 2004, we had encouraging and similar results at the Manse Loch System in 2003 and at Shieldaig in 2005. Whilst prophylactic treatment had no significant effect on sea trout survivorship, length or weight there were significant effects on condition factor between the treated and control groups. At the Manse Loch system, this result should be treated with caution, because it was apparent only when Fulton’s K was used as an index of condition factor. Fulton’s K is length-dependent when the isometry of length/weight departs from an exponent of 3. WR, or relative condition, is a more reliable length-independent measure. At Shieldaig in 2005, significant effects were observed when both Fulton’s K and WR were used. This suggests that EX-treated fish were subject to lesser constraints on their growth at sea and this may have significant impacts on their subsequent survivorship, growth to adulthood and reproduction. This is of particular importance to stock conservation because of the allometric relationship between female size and fecundity. Survivorship of post-smolt salmonids tends to be sizedependent, so any agent that enhances growth rate ought also to reflect enhanced survivorship.

Paper 2 outlines work carried out by the CFB on the efficacy of SLICE® and its use as a remedial tool for sustainable stock enhancement of wild salmonids on the Gowla and Invermore systems. Wild sea trout smolts were trapped during their downstream migration, batch tagged, and then evenly distributed between control and experimental tanks. Fish were fed on fish pellets initially and when seen to be actively feeding, SLICE® pellets were introduced to the experimental tank. Fish were treated for approximately seven days and then released below the traps to continue their migration. In addition, tagged salmon smolts were transferred to lake cages on both systems. Fish were fed normally for six weeks before SLICE® pellets were introduced to the experimental fish as above. The salmon smolts were then released close to the sea.

Highest lice infestation pressure in Bertraghboy bay (Gowla System) (2004) coincided with highest sea lice infestation on sea trout, a high rate of premature return, poor finnock marine growth and low marine finnock survival. High lice infestation pressure in Kilkieran bay (Invermore) in both years also coincided with high lice infestation on sea trout, premature return and poor marine growth. In the absence of farmed salmon in Bertraghboy bay in 2005, a very low lice infestation on sea trout, absence of premature returns, highest recorded marine growth and an increase in marine survival was observed. In 2005 sufficient fish returned from control and experimental groups on the Gowla system to allow statistical analysis. Returns revealed no statistical difference between SLICE® and control groups in terms of growth, weight, chalimus abundance, total lice level or days spent at sea.

Results for salmon smolts from the present study reveal a significant difference in the return rate of SLICE® treated and control fish in three of four release groupings, and suggest increased (early) mortality of smolts from sea lice infestation in aquaculture bays. The greater weight of adult salmon returning after SLICE®-treatment could be expected from enhanced growth of fish protected from high sea lice infestation early in their migration. The findings of the present study demonstrate the potential impact of marine salmon farms on salmon post-smolt marine mortality in very short coastal bays in the West of Ireland.

199

SUMBAWS

Final Report

CONCLUSIONS There were several problems which occurred during this workpackage. There were clear problems with the fish trap at the Manse Loch System. In 2003, it became apparent that the trap would reliably function only between well-defined water levels. For much of 2003, which was a particularly dry year, there was insufficient water flowing through the trap to allow the fish to move either downstream or upstream. In addition, at times of high water flow it was also evident that fish were able to pass over the top of the trap in either direction. In all three study years it was certain that a number of smolts and returning fish were not trapped for this reason. It was also decided in year one that, due to the small number of wild salmon smolts in the Manse Loch System, salmon would not be included in the trial in Scotland. However, many of these problems were highlighted after the first year of study and were ameliorated thereafter by the extension of the study to the Shieldaig system. It should be noted that Shieldaig is an exclusively sea trout system and no salmon were therefore included in Scotland. In Ireland, permission for the experimental use of Substance EX was not granted, resulting in our assessing SLICE速 as an alternative treatment. Despite these problems, all objectives and deliverables from the original research contract have been met for this workpackage.

200

SUMBAWS

Final Report

Paper 1: EFFICACY OF SUBSTANCE EX AS A REMEDIAL TOOL FOR THE REGENERATION OF DEPLETED STOCKS OF WILD SALMONIDS BY NATURALLY-SPAWNED SMOLTS IN TWO SCOTTISH SYSTEMS A. Wells, D. Hay, M. McKibben, N. Hazon and C. Todd.

(Manuscript) Introduction Sea trout (Salmo trutta L.) numbers in NW Europe have fallen since the 1950s (Anonymous, 2004), but during the late 1980s and early 1990s the first reports of collapses in certain wild sea trout stocks, and the possible association of these collapses with sea lice (Lepeophtheirus salmonis, Krøyer), arose in Ireland (Tully et al., 1993b), Scotland (Northcott and Walker, 1996), and Norway (Birkeland, 1996). Sea trout remain in nearshore waters throughout their marine residence (Klemetsen et al., 2003) and stock declines appear to be attributable primarily to increased marine mortality (Anonymous, 1993; Anonymous, 1995; Northcott and Walker, 1996). Caligid sea lice infestations can debilitate or kill the host fish (Pike and Wadsworth, 1999), and there is strong circumstantial evidence that sea lice might be a major contributory factor to marine mortality rates (Butler, 2002; McKibben and Hay, 2004).

Copepod sea lice, L. salmonis and Caligus elongatus (Nordmann), are the primary marine ectoparasites of farmed and wild salmonids. Sea lice infestation of the host fish is characterised by stress and osmoregulatory imbalance, which occur once the sea lice moult to the mobile preadult and adult stages (Bjørn and Finstad, 1997; Dawson et al., 1998; Finstad et al., 2000; Grimnes and Jakobsen, 1996). Feeding activity by preadult/adult sea lice can result in increased skin damage (Bjørn and Finstad, 1998; Nolan et al., 1999; Pike and Wadsworth, 1999), which can be fatal to hatchery-reared sea trout and salmon (Bjørn and Finstad, 1997; Dawson et al., 1998; Grimnes and Jakobsen, 1996).

Coincident with the first reports of stock crashes were reports of ‘premature return’ of post-smolt sea trout to fresh water, within a few weeks of their first migration to sea. These fish typically bore heavy infestations of largely juvenile (chalimus) stages of lice (MacKenzie et al., 1998; Tully et al., 1993a; Tully and Whelan, 1993). Despite their large sample sizes and high abundances of juvenile sea lice stages on infested fish, Tully et al. (1993a) failed to capture juvenile sea trout bearing adult sea lice, and concluded that any such host fish had succumbed to infestation.Preliminary studies in Norway have demonstrated that releases of Atlantic salmon and sea trout smolts, which were protected from sea lice infestation by prophylactic treatment, resulted in higher recaptures (0.90%) compared to unprotected fish (0.03%) in areas with heavy fish farming activity (Finstad and Jonsson, 2001). The aim of the present study was to assess the efficacy of prophylactic treatment in wild sea trout smolts intercepted from small, clearly defined Scottish loch systems.

Materials and Methods Study sites The Manse Loch System, Assynt, Sutherland: This system comprises three freshwater lochs, each approximately 1 km in length and interconnected by a small stream, the lowest of which empties into a fjordic sea loch. Salmon and sea trout migrate to the highest of these three lochs and spawn in feeder streams. The Manse Loch System has a total catchment area of 1570.25 ha., and a total area accessible to fish of 39.4 ha. (Shona Marshall, pers. com.). The outfall into the estuarine head of the sea loch is narrow (8m), and at this point an existing concrete dam, with two gabion basket weirs immediately below, has been modified to allow the installation of an upstream/downstream trap for salmonids with a view to allowing a total census of upstream and downstream migrants as close as possible to saltwater. The trap has been operated by the West Sutherland Fisheries Trust and volunteers from the Assynt Crofters Trust since 1999.

Shieldaig, Wester Ross: The Shieldaig River in North West Scotland, which flows into Loch Shieldaig, is enclosed by an adult and juvenile fish trap situated at the mouth of the river. The river has been stocked each year with 25,000 adipose-clipped trout fry.

201

SUMBAWS

Final Report

Tagging and treatment of fish The trial was conducted from 2003-2005 in Assynt and 2004-2005 at Shieldaig. The downstream traps of both systems were checked for fish twice a day throughout April and May. Downstream migrants were removed from the trap and anaesthetised in batches of 10-20 fish in 20 litres of 0.1% phenoxyethanol.

Salmon smolts in Assynt were measured and then immediately released following recovery from anaesthesia. Sea trout smolts at both sites were PIT tagged, using the method of Gries & Letcher (2002). In summary, following anaesthesia, a small incision (approximately 2 mm) was made with a scalpel just posterior to the pectoral fins, along the mid-ventral line. The PIT tag was then gently pushed posteriorly into the body cavity using the experimenter’s fingers. All PIT tagged fish were fin clipped, by removal of the adipose fin. The weight and length of the smolts was recorded and scale samples were taken from approximately 10% of fish in order to estimate the age at migration. The sea trout were allowed to recover in 75 litre tanks before being transferred to duplicate treatment/control tanks. Half of the smolts were treated with Substance EX (Alpharma), the other half forming a control group. The control group was subject to identical handling and tagging as the treated group but was ‘treated’ with water only. All tanks were continuously supplied with air and, following a 30 minute bath treatment, the fish were removed to a keep net (Assynt) or a separate section of the downstream trap (Shieldaig) and allowed to recover overnight. The following day a sub-sample of the fish was checked for the presence of a PIT tag using a hand held tag reader in order to estimate tag loss, and then all fish were released. At Shieldaig, a maximum of 320 fish were treated per day to remain within the discharge consent for Substance EX. Surplus fish on these days were released untreated.

Returning fish The upstream trap was checked once daily in April and May and then twice daily from June – October.

All fish were lightly anaesthetised, measured and weighed prior to release and all sea trout were checked for the presence of a PIT tag using a hand held tag reader. The presence of sea lice was assessed by placing the fish into white trays containing anaesthetic and counting the lice. Sea lice were then recorded according to Johnson and Albright (1991) and Schram (1993). Weight and length were recorded and host condition factor was calculated for all fish sampled for parasite analysis, using 2 of the standard measures (Blackwell et al.

2000, Brenden et al. 2003; Marshall et al. 2004; Neff & Cargnelli 2004). Fulton’s index (K = 100(Weight in g)/(Length in cm)3) assumes growth to be isometric and scaled to the cube of the length. Any departure from an exponent of 3 will, therefore, result in a length-dependence of individual fish CF. A measure more independent of length is the relative mass index, WR (WR = W/WS), whereby W is the observed mass and WS the standard mass of individual fish calculated from a specific mass-length equation (Blackwell et al. 2000, Brenden et al. 2003).

Results Assynt In 2003, 649 sea trout smolts were trapped, of which 590 were PIT tagged. Of these fish, 288 were treated with Substance EX and 289 were untreated controls. In 2004, 224 sea trout smolts were trapped, all of which were PIT tagged. Of these fish, 109 were treated with Substance EX and 115 were untreated controls. In addition 14 overwintered finnock, which were PIT tagged in 2003, were also trapped returning to sea. It should be noted however that 10 of these fish were not recorded in the upstream trap during the sea trout return in 2003, demonstrating that upstream-migrating fish are able to bypass the upstream trap. In 2005, only 8 sea trout smolts were trapped, all of which were PIT-tagged but none were treated. The mean length, weight, condition factor and relative mass of sea trout smolts are shown in Table 1.

Table 1 Length, weight and condition factor of sea trout smolts (Mean ± SEM) Year 2003 2004 2005

Control EX-Treated Control EX-Treated Control EX-Treated

Mean Length 182.5 ± 1.08 183.0 ± 1.04 189.4 ± 1.77 190.1 ± 1.89 206.9 ± 6.5 -

Mean Weight 59.8 ± 1.05 60.1 ± 1.16 67.7 ± 1.83 69.0 ± 1.95 88.3 ± 6.4 -

K 0.97 ± 0.02 0.97 ± 0.02 0.97 ± 0.01 0.98 ± 0.01 0.99 ± 0.02 -

WR 0.85 ± 0.02 0.85 ± 0.01 0.85 ± 0.00 0.86 ± 0.00 0.88 ± 0.01 202

SUMBAWS

Final Report

PIT tagged sea trout returns for 2003 are summarised in Table 2. In total 136 sea trout returned to the system of which 41 were PIT tagged. 95 returning sea trout bore no detectable tags are we therefore assume that these fish originated from a different system. For the PIT tagged fish there were no significant differences between the treated and untreated groups in terms of their survivorship, length or weight on return to fresh water. There also were no significant differences in prevalence or intensity of sea lice.

However, there was a significant decrease in condition factor for the control group which was not apparent in the EX treated group. Importantly, the decrease in condition was not apparent when measured using WR. In 2004, only 2 sea trout were recovered in the upstream trap (one from each group), and in 2005 no returning fish were recorded.

Table 2 Summary of PIT tagged sea trout (finnock) returns to the Manse Loch System. (Values for length, weight and condition factor are Mean ± SEM. Values for length, weight and condition factor as smolts refer only to those fish which survived to return as finnock. *p < 0.05 represents a significant difference between condition factor as a smolt and its condition factor on return (Paired t-test)).

2003 Number (Survivorship) Mean Length Mean Weight K WR

Control Smolt

186 ± 3.9 mm 67.3 ± 6.6 g 1.00 ± 0.05 0.85 ± 0.04

Return 22/289 (7.6%) 288.4 ± 7.2 mm 203.2 ± 11.0 g 0.87 ± 0.04* 0.81 ± 0.03

EX-Treated Smolt Return 19/288 (6.6%) 178.9 ± 3.7 mm 272.1 ± 3.6 mm 56.8 ± 3.7 g 192.4 ± 6.3 g 0.97 ± 0.02 0.95 ± 0.02 0.86 ± 0.02 0.87 ± 0.02

Shieldaig In 2004, 1135 sea trout smolts were trapped, all of which were PIT tagged. Of these fish, 420 were treated with Substance EX and 417 were untreated controls. In 2005, 1610 sea trout smolts were trapped and PIT tagged. Of these fish, 581 were treated with Substance EX and 590 were untreated controls. The mean length, weight and condition factor of sea trout smolts are shown in Table 3.

Table 3 Length, weight and condition factor of sea trout smolts (Mean ± SEM) Year 2004 2005

Control EX-Treated Control EX-Treated

Mean Length 176 ± 0.73 178 ± 0.73 175 ± 0.75 175 ± 0.65

Mean Weight 53 ± 0.68 54 ± 0.69 49 ± 0.63 49 ± 0.59

K 0.95 ± 0.004 0.95 ± 0.003 0.90 ± 0.004 0.90 ± 0.004

WR 0.83 ± 0.003 0.83 ± 0.003 0.78 ± 0.004 0.79 ± 0.003

PIT tagged sea trout returns for both years are summarised in Table 4. In 2004 only 4 PIT tagged finnock from the Substance EX trial returned to the Shieldaig system, all of which were from the treated group. In 2005, 23 finnock from the EX trial returned, of which 9 were controls and 14 were treated with Substance EX. There were no significant differences between the treated and untreated groups in terms of their survivorship, length or weight on return to fresh water. However, there was a significant increase in condition factor (both K and WR) in returning fish from the EX treated group which was not apparent in the control group.

203

SUMBAWS

Final Report

Table 4 Summary of sea trout (finnock) returns to the Shieldaig Trap (Values for length, weight and condition factor are Mean ± SEM. Values for length, weight and condition factor as smolts refer only to those fish which returned as finnock. *p < 0.05 represents a significant difference between condition factor as a smolt and condition factor on return (Paired t-test)).

Control 2004 Number (Survivorship) Mean Length Mean Weight K WR 2005 Number (Survivorship) Mean Length Mean Weight K WR

EX Treated

Smolt

Return

-

0/417 (0%) -

174.6 ± 2.7 mm 48 ± 2.6 g 0.89 ± 0.02 0.78 ± 0.01

9/590 (1.5%) 252.3 ± 4.5 mm 147 ± 8.9 g 0.91 ± 0.03 0.83 ± 0.02

Smolt

Return

193 ± 26 mm 72 ± 29 g 0.96 ± 0.04 0.83 ± 0.02

4/420 (0.71%) 276 ± 24 mm 244 ± 77 g 1.13 ± 0.08 0.97 ± 0.07

175.1 ± 1.5 mm 47.7 ± 1.8 g 0.88 ± 0.02 0.77 ± 0.02

14/581 (2.4%) 255.2 ± 2.9 mm 165.3 ± 7.0 g 0.99 ± 0.02* 0.90 ± 0.02*

Discussion The overall returns of sea trout at both sites in Scotland have been disappointing. In addition, the number of sea trout smolts migrating from the Manse Loch System showed a marked decline over the three years of the present study. The reasons for this decline remain unclear, but it should be noted that the system was stocked with 50,000 sea trout fry for 4 years between 1996 and 2000. Given that we observed that 50% of smolts migrating from this system were 3+ fish it is possible that the majority of fish migrating from the system during this study were stocked fish, and that the decline in smolt numbers was due to the cessation of stocking in 2000.

In addition to the foregoing there were clearly some problems with the trap at the Manse Loch System. In 2003, it became apparent that the trap would function only between well-defined water levels. For much of 2003, which was a particularly dry year, there was insufficient water flowing through the trap to allow the fish to move either downstream or upstream. This was almost certainly a contributory factor to the low return of fish in 2003 because fish were regularly seen in the pool below the trap. In addition, at times of high water flow is was also evident that fish were able to pass over the top of the trap in either direction. In all three study years it was likely that a number of smolts and returning fish were not trapped for this reason.

Indeed, several overwintered finnock which were tagged in 2003, but not recorded on their upstream migration, were recorded in the downstream trap in 2004.

It was largely for these reasons that the study was extended to Shieldaig in 2004 and 2005 and, despite a disappointing return at both sites in 2004; we had encouraging and similar results at the Manse Loch System in 2003 and at Shieldaig in 2005. Whilst prophylactic treatment had no significant effect on survivorship, length or weight there were significant effects on condition factor between the treated and control groups. At the Manse Loch system, this result should be treated with caution, because it was only apparent when Fulton’s K was used as an index of condition factor, but in Shieldaig in 2005 significant effects were observed when both Fulton’s K and WR were used. This suggests that EX-treated fish were subject to lesser constraints on their growth at sea and this may have significant impacts on their subsequent growth to adulthood and reproduction. This is of particular importance to stock conservation because of the allometric relationship between female size and fecundity. Similar results have also been observed in Norway in previous studies (Finstad and Jonsson, 2001) and as demonstrated in WP6.

Concern must, however, be raised regarding the extremely poor survivorship and returns of sea trout to the Shieldaig system, irrespective of their treatment with Substance EX. Because sea trout post-smolts undergo only a comparatively short migration to sea during the summer months it would be expected that a large proportion would survive to return and over-winter. The causes of the very low survivorship in Loch Torridon are not apparent but should remain the focus of continued research. These results also demonstrate the intense vulnerability of large and small river stocks of sea trout in Scotland.

204

SUMBAWS

Final Report

References Anonymous. (1993). Report of the Sea Trout Working Group 1993, pp. 127. Abbotstown: Department of the Marine.

Anonymous. (1995). Report of the Sea Trout Working Group 1994. 254 pp.

Anonymous. (2004). Statistical Bulletin. Scottish salmon and sea trout catches, 2003. Report No. Fis/2004/1, pp. 29. Aberdeen: Fisheries Research Services.

Birkeland, K. (1996). Consequences of premature return by sea trout (Salmo trutta) infested with the salmon louse (Lepeophtheirus salmonis Kroyer): Migration, growth, and mortality. Canadian Journal of Fisheries and Aquatic Sciences 53, 2808-2813.

Bjørn, P. A. and Finstad, B. (1997). The physiological effects of salmon lice infection on sea trout post smolts. Nordic J. Freshw. Res. 73, 60-72.

Bjørn, P. A. and Finstad, B. (1998). The development of salmon lice (Lepeophtheirus salmonis) on artificially infected post smolts of sea trout (Salmo trutta). Can. J. Zool. 76, 970-977.

Blackwell B.G., Brown M.L. and Willis D.W. (2000) Relative weight (Wr) status and current use in fisheries assessment and management. Rev Fish Sci 8:1-44.

Brenden T.O., Murphy B.R. and Birch J.B. (2003) Statistical properties of the relative weight (Wr) index and an alternative procedure for testing Wr differences between groups. N Am J Fish Manage 23:1136–1151.

Butler, J. R. A. (2002). Wild salmonids and sea louse infestations on the west coast of Scotland: sources of infection and implications for the management of marine salmon farms. Pest Management Science 58, 595608.

Dawson, L. H. J., Pike, A. W., Houlihan, D. F. and McVicar, A. H. (1998). Effects of salmon lice Lepeophtheirus salmonis on sea trout Salmo trutta at different times after seawater transfer. Diseases of Aquatic Organisms 33, 179-186.

Finstad, B., Bjorn, P. A., Grimnes, A. and Hvidsten, N. A. (2000). Laboratory and field investigations of salmon lice Lepeophtheirus salmonis (Kroyer) infestation on Atlantic salmon (Salmo salar L.) post-smolts.

Aquaculture Research 31, 795-803.

Finstad, B. and Jonsson, N. (2001). Factors Influencing the Yield of Smolt Releases in Norway. Nordic J. Freshw. Res. 75, 37-55.

Gries, G. and Letcher, B. H. (2002). Tag retention and survival of age-0 Atlantic salmon following surgical implantation with passive integrated transponder tags. North American Journal of Fisheries Management 22, 219-222.

Grimnes, A. and Jakobsen, P. J. (1996). The physiological effects of salmon lice infection on post- smolt of Atlantic salmon. Journal of Fish Biology 48, 1179-1194.

Johnson, S. C. and Albright, L. J. (1991). The Developmental Stages of Lepeophtheirus Salmonis (Kroyer, 1837) (Copepoda, Caligidae). Canadian Journal of Zoology-Revue Canadienne De Zoologie 69, 929-950.

Klemetsen, A., Amundsen, P. A., Dempson, J. B., Jonsson, B., Jonsson, N., O'Connell, M. F. and Mortensen, E. (2003). Atlantic salmon Salmo salar L., brown trout Salmo trutta L. and Arctic charr Salvelinus alpinus (L.): a review of aspects of their life histories. Ecology of Freshwater Fish 12, 1-59.

MacKenzie, K., Longshaw, M., Begg, G. S. and McVicar, A. H. (1998). Sea lice (Copepoda : Caligidae) on wild sea trout (Salmo trutta L.) in Scotland. Ices Journal of Marine Science 55, 151-162.

Marshall C.T., Needle C.L., Yaragina N.A., Ajiad A.M. and Gusev E. (2004) Deriving condition indices from standard fisheries databases and evaluating their sensitivity to variation in stored energy reserves. Can J Fish Aquat Sci 61:1900–1917.

McKibben, M. A. and Hay, D. W. (2004). Distributions of planktonic sea lice larvae Lepeophtheirus salmonis in the inter-tidal zone in Loch Torridon, Western Scotland in relation to salmon farm production cycles. Aquaculture Research 35, 742-750.

205

SUMBAWS

Final Report

Neff B.D. and Cargnelli L.M. (2004). Relationships between condition factors, parasite load and paternity in bluegill sunfish, Lepomis macrochirus. Env Biol Fishes 71:297â&#x20AC;&#x201C;304.

Nolan, D. T., Reilly, P. and Wendelaar Bonga, S. E. (1999). Infection with low numbers of the sea louse Lepeophtheirus salmonis induces stress-related effects in postsmolt Atlantic salmon (Salmo salar).

Canadian Journal of Fisheries and Aquatic Sciences 56, 947-959.

Northcott, S. J. and Walker, A. F. (1996). Farming salmon, saving sea trout; a cool look at a hot issue. In Aquaculture and sea lochs, (ed. K. Black), pp. 72-81. Oban: Scottish Association for Marine Science.

Pike, A. W. and Wadsworth, S. L. (1999). Sealice on salmonids: their biology and control. Adv. Parasitol.

44, 233-337.

Schram, T. A. (1993). Supplimentary descriptions of the developmental stages of Lepeophtheirus salmonis (Kroyer, 1837) (Copepoda: Caligidae). In Pathogens of Wild and Farmed Fish: Sea Lice, eds. G. A. Boxshall and D. Defaye), pp. 30-47. Chichester: Ellis Horwood.

Tully, O., Poole, W. R. and Whelan, K. F. (1993a). Infestation parameters for Lepeophtheirus salmonis (Kroyer) (Copepoda: Caligidae) parasitic on sea trout, Salmo trutta L., off the west coast of Ireland during 1990 and 1991. Aquaculture and Fisheries Management 24, 545-555.

Tully, O., Poole, W. R., Whelan, K. F. and Merigoux, S. (1993b). Parameters and possible causes of Lepeophtheirus salmonis (Krøyer) infesting sea trout (Salmo trutta L.) off the west coast of Ireland. In Pathogens of wild and farmed fish. Sea lice, eds. G. A. Boxshall and D. Defaye), pp. 202-213. Chichester: Ellis Horwood.

Tully, O. and Whelan, K. F. (1993). Production of Nauplii of Lepeophtheirus salmonis (Kroyer) (Copepoda, Caligidae) from Farmed and Wild Salmon and Its Relation to the Infestation of Wild Sea Trout (Salmo trutta L) Off the West-Coast of Ireland in 1991. Fisheries Research 17, 187-200.

206

SUMBAWS

Final Report

Paper 2 EFFICACY OF SUBSTANCE EX/SLICE® AS A REMEDIAL TOOL FOR THE REGENERATION OF DEPLETED STOCKS OF WILD SALMONIDS BY NATURALLY-SPAWNED SMOLTS P. Gargan, N. Mc Donnell and G. Forde (Manuscript) Introduction Sea trout stocks began to decline in West Coast Irish rivers in 1987 and 1988. This was followed by a total collapse in spawning stocks and rod catches in 1989 and 1990. The Connemara district rod catch fell from an average of 9570 sea trout from 1974-1988 to only 240 sea trout in 1990 (Gargan et al., 2002).

Following this collapse sea trout were observed returning prematurely to freshwater with heavy sea lice infestation. There was also a change in the sea trout population structure (Anon, 1994a).

Since the early 1980s, there has been an increasing trend in the production of finfish and shellfish at aquaculture sites in the west of Ireland. The production of farmed salmon (Salmo salar) increased from 21t in 1980 to 14000t in 1996 (McMahon, 2000). Salmon farms are a potential source of high numbers of sea lice and studies have been carried out on the possible relationship between the occurrence of sea lice on wild sea trout and lice levels on Atlantic salmon farms in Ireland (Tully et al., 1993, Gargan et al., 2002). In 1992 and 1993, the Sea Trout Working Group found a statistically significant relationship between lice infestation on sea trout and distance to the nearest fish farm. However the relationship was weaker in 1994, which coincided with reduced sea lice levels on fish farms during this period (Anon, 1994). Gargan et al (2002) also found a statistical relationship between lice infestation on sea trout and distance to the nearest salmon farm over a ten year period.

The sea louse, Lepeophtheirus salmonis (Krøyer) is a parasite of salmonid fishes. Sea lice lodge on fish skin, resulting in two main physiological changes to their host. The primary response is an increase in stress (Poole et al., 2000) and a compromised immune response (Wootten et al., 1982; Wagner & Kinley, 2004).

The second response is osmoregulatory failure (Grimnes & Jakobsen, 1996). Other effects include blood loss, skin and fin erosion, (Wootten et al., 1982; Wagner & Kinley, 2004), reduced growth and reproductive failure (Bjorn & Finstad, 1997) and at heavy lice levels fish may die at sea (Bjorn et al., 2001). The ability of sea trout and salmon to survive depends on the severity of the infection. Work carried out by Finstad et al. (2000) found that 11 lice per fish kill juvenile Atlantic salmon weighing 15g or less. However even at high infestation levels the damage is minimal until sea lice reach the pre-adult and adult stage (Bowers et al., 2000). Lice survival is reduced at lower salinities and they generally do not survive more than a few days once they enter freshwater, although Wootten et al. (1982) did report survival up to 21 days after entering freshwater.

The Owengowla (Gowla) and Invermore are two river systems situated in Connemara in the west of Ireland. These locations were chosen as study sites because of the existing upstream and downstream trapping facilities that allow the quantification of downstream and upstream movement of salmonids. Both systems have been studied in detail over the past decade and extensive baseline data were already available.

The Owengowla system enters the sea at Bertraghboy Bay and the Invermore system enters the sea in Kilkieran Bay. Salmon farming activities began in Kilkieran Bay and Bertraghboy Bay in 1982 and 1985 respectively, with production increasing rapidly in the late 1980’s. A sea trout stock collapse was experienced on both river systems in 1989. In 1990 sea trout were observed returning prematurely in May to both rivers with high juvenile lice levels.

In the present study the effectiveness of a prophylactic lice treatment, SLICE®, on sea trout and salmon survival was investigated. SLICE® is an in-feed treatment which has been successfully used for the control of Lepeophtheirus salmonis and Caligus elongatus, two species of sea lice affecting farmed salmonids.

Emamectin benzoate is the active component of SLICE®. When sea lice feed on tissues of treated fish, emamectin is taken up into the tissues of the louse. Emamectin then binds to ion channels of nerve cells and disrupts transmission of nerve impulses, which results in paralysis and death of the parasite. Emamectin benzoate is excreted slowly by fish and metabolized to inactive compounds, resulting in long-lasting protection. The experimental programme was designed to determine marine survival, marine growth and length of time that sea trout and salmon spent at sea before returning to freshwater.

207

SUMBAWS

Final Report

Material and Methods The location of the two river systems is shown in Figure 1.

Figure 1 Location map of the river Gowla and Invermore Sampling Regime for Sea Trout The number of sea trout in ascending and descending traps was monitored daily and more frequently during flood periods from 2003 - 2005. Fish length (cm) and a sample of fish scales were taken for later analysis.

Sea trout were removed from traps and measured, placed in anaesthetic (2-phenoxy-ethanol) before a visible implant (VI) elastomer tag was inserted in the clear membrane above the eye. Fish then were either retained for use in the SLICE® experiment or released downstream of the trap.

Sea trout were not treated with SLICE® during spring 2003 because final permission for the study had not then been granted. Large numbers of sea trout smolts migrated in the Gowla during the treatment period (2004-2005), and three separate treatments were used each year. Lower numbers of fish migrated on the Invermore and only one treatment of SLICE® was carried out in 2004 and 2005.

Once tagged, fish were evenly distributed between a control and an experimental tank (SLICE® treated).

In the control tank sea trout were fed on normal pellets. In the experimental tank, fish were fed on pellets initially and when seen to be actively feeding, SLICE® pellets were introduced for a period of seven days prior to fish being released below the traps.

The upstream traps were monitored on a daily basis on both systems from February 1st to September 30th each year to record returning sea trout. Sea trout caught in the trap were removed and the length, and number of chalimus and post-chalimus sea lice were recorded for each fish. Fish returning between February 1st and June 1st were classified as premature returns. Data also were recorded on whether the fish were VI tagged or fin-clipped. Sea trout scales and weight were taken from a sample of returning fish to assess age and marine growth.

A Mann-Whitney U-test was used to test the differences between lengths of smolts in the downstream traps between sites. Comparisons were made between sea lice levels on the control fish and those treated with SLICE® in the experimental tank.

208

SUMBAWS

Final Report

Sampling Regime for Salmon In December/January 2003-2004 (thereafter 2004 release), 20 000 salmon smolts were micro-tagged with four separate batch codes and the fish were adipose fin-clipped. The experiment was repeated in December 2004/January 2005 (thereafter 2005 release).

The microtagged fish were transferred to lake cages on the Invermore and Gowla for imprinting, to improve homing as adults. Salmon smolts were divided into four groups, two of which were placed in cages on the Invermore and two on the Gowla. At each site there was a control and experimental group. The fish were fed normally for about six weeks and then SLICE® was introduced to the diet of fish in the experimental cage for a period of seven days prior to release. Both groups of salmon smolts were released close to the sea. Samples of 50 fish were taken from each group to test for tag loss and to record weight and condition. The number of salmon smolts in control and experimental groups released in 2003 and 2004 are shown in Tables 13 and 15.

The majority of recaptures of salmon smolts released in 2004 were taken the following year (2005) in the offshore commercial salmon drift net fishery, operating during June and July. Upstream traps were monitored for the return of smolts released in spring 2004 and returned as 1SW grisle in 2005. The Marine Institute operates a micro-tag recovery programme from the commercial fishery and recovers microtags from adipose fin-clipped salmon. A core of flesh was removed from the head region, dissected under a microscope, and the micro-tag recovered and read by staff from the Marine Institute. The return rate of SLICE®-treated and untreated fish, both from the commercial fishery and from upstream traps was determined.

A chi-square test was used to test if there was a statistical difference in survival between the SLICE® and control groups from both fisheries. A Mann-Whitney U-test was used to determine the difference in salmon condition factor, mean weight and mean length for salmon which returned in 2004. Data on length and weight of returned salmon in 2005 were not available at the time of writing this report.

Estimation of Sea Lice Infestation Pressure An estimate of sea lice infestation pressure from farmed salmon was calculated for Bertraghboy bay and Kilkieran bay in 2003 and 2004. This estimate was derived by taking the average sea lice (L.salmonis) data on ovigerous and total lice in April recorded on farmed salmon (Source: Irish Marine Institute) and multiplying these data by the estimated number of grower salmon (excluding smolts) in each bay in each year.

Results Freshwater Descending Sea Trout Smolts Production of sea trout smolt numbers has increased substantially on the Gowla during the study period, with 4 231 sea trout smolts passing through the downstream trap in 2005 (Table 1). However, there has been a dramatic decline in the numbers of sea trout smolts migrating on the Invermore with the lowest number of sea trout smolts (334) recorded in 2004. A flood event in March 2003 may have resulted in significant numbers of sea trout smolts descending the Gowla without being encountered in the downstream trap. A total of 3 965 sea trout were tagged in both systems over the study period,(Table 1). In total, 1 197 sea trout were tagged on the Invermore and 2 768 were tagged on the Gowla.

Table 1 Number of fish tagged in control and experimental (SLICE®) groups (* possibly underestimated due to flood event) Fishery Invermore

Gowla

Year 2003 2004 2005 2003 2004 2005

Total Smolt Run 1391 334 465 143* 1302 4231

Number of sea trout tagged 764 194 239 84 1093 1591

Number SLICE® 0 96 115 0 569 850

Number controls 0 98 124 0 524 741

% sea trout tagged 54.9 58.1 51.4 58.7 83.9 37.6

209

SUMBAWS

Final Report

In both 2004 and 2005, treatment of sea trout with SLICE速 took place three times during the smolt run.

There was a smaller smolt run from the Invermore fishery and only the one treatment of SLICE速 was used each year.

Smolt Length There was a statistical difference in smolt length between all years (p < 0.05) on the Gowla system and between 2003 versus 2004-2005 on the Invermore system (p < 0.05). There was no significant difference in smolt length between the Gowla and the Invermore in 2003 (Mann-Whitney: p = 0.293, z = -1.052, n = 894). In 2004 smolts were smaller than in 2003 on both systems and fish in the Gowla were smaller than those in the Invermore (Mann-Whitney: p = 0.000, z = -5.853, n = 1539) (Figure 2). In 2005, smolts were of similar length in both systems (Mann-Whitney: p = 0.146, z = -1.455, n = 2093).

Gowla 2003 (n=82)

250

Inver 2003 (n=812) Number

200 150 100 50 24-24.9

23-23.9

22-22.9

21-21.9

20-20.9

19-19.9

18-18.9

17-17.9

16-16.9

15-15.9

14-14.9

13-13.9

0

Length (cm)

Gowla 2004 (n=1211)

350

Inver 2004 (n=328)

300 Number

250 200 150 100 50 24-24.9

23-23.9

22-22.9

21-21.9

20-20.9

19-19.9

18-18.9

17-17.9

16-16.9

15-15.9

14-14.9

13-13.9

0

Gowla 2005 (n=1857)

500 450 400 350 300 250 200 150 100 50 0

24-24.9

23-23.9

22-22.9

21-21.9

20-20.9

19-19.9

18-18.9

17-17.9

16-16.9

15-15.9

14-14.9

Inver 2005(n=236)

13-13.9

Number

Length (cm)

Length (cm )

Figure 2 Length frequencies of sea trout smolts on the Invermore and the Gowla in 2003, 2004 and 2005.

210

SUMBAWS

Final Report

Smolt Age Scale samples were taken from a subsample of sea trout smolts for age analysis and these results are summarised in Tables 2 (Gowla) and 3 (Invermore).

Table 2 Percentage of sea trout smolts occurring within each age group on the Gowla Year

Total number (n)

2003 2004 2005

17 54 422

1+ 0 0 <1

Age category (%) 2+ 3+ 41 35 24 56 15 71

4+ 24 2 13

Table 3 Percentage of sea trout smolts occurring within each age group on the Invermore Year

Total number (n)

2003 2004 2005

185 101 86

1+ 0 0 1

Age category (%) 2+ 3+ 21 74 28 67 17 62

4+ 5 5 20

Returning sea trout Although a large number of sea trout smolts was tagged in 2004 and 2005, few were recorded in upstream traps (Table 4). Equal numbers of fish from both groups were recorded from the Gowla, whereas a greater number of SLICE®-treated fish returned to the Invermore in 2004. No tagged sea trout returned to the Invermore in 2005.

Table 4 Return rate of tagged sea trout in control and experimental groups Fishery Gowla Invermore

Year 2004 2005 2004 2005

Number tagged 1093 1591 194 239

Number of returns 5 (0.45%) 23 (1.45%) 7 (3.6%) 0 (0%)

SLICE® 3 11 5 0

Control 2 12 2 0

Sea Trout Length Frequency The length of sea trout captured in upstream traps was recorded over the 2003 to 2005 period (Figure 3).

There was a statistical difference in sea trout length between the Gowla and Invermore systems in 2003 with larger sea trout (26 cm) recorded on the Gowla system (Mann-Whitney U-test : p < 0.001, z = -3.898, n = 203). The mean sea trout length on the Gowla was 25.2 cm ± 0.22 (median: 26 cm) whereas on the Invermore the mean length was only 22.9 cm ± 0.52 (median: 22.5 cm).

A Mann-Whitney test revealed no significant difference in length between sea trout recorded in upstream traps on either fishery in 2004 (Mann-Whitney: p = 0.4, z = -0.841, n = 65) or in 2005 (Mann-Whitney: p = 0.950, z = -0.063, n = 178). Median length of sea trout on the Invermore in 2004 was 21.6 cm and slightly smaller at 20 cm on the Gowla system. The largest median sea trout size on the Gowla system during the three year period was recorded in 2005 at 27 cm, comparable to that from the Invermore from a very small sample size (median: 27.0, n = 10).

Length of sea trout in the upstream trap on the Gowla varied during the three year period (2003: 26 cm, 2004: 20 cm, 2005: 27 cm) (p < 0.05). There was no statistical difference in length on the Invermore between 2003 and 2004.

211

SUMBAWS

Final Report

35 Invermore 2003( n=72)

30

Gowla 2003( n=154)

Number

25 20 15 10 5

41-41.9

39-39.9

37-37.9

35-35.9

33-33.9

31-31.9

29-29.9

27-27.9

25-25.9

23-23.9

21-21.9

19-19.9

17-17.9

15-15.9

0

9 8 7 6 5 4 3 2 1 0

Invermore 2004( n=36)

41-41.9

39-39.9

37-37.9

35-35.9

33-33.9

31-31.9

29-29.9

27-27.9

25-25.9

23-23.9

21-21.9

19-19.9

17-17.9

Gowla 2004( n=41)

15-15.9

Number

Length (cm)

45 40 35 30 25 20 15 10 5 0

Invermore 2005( n=10)

41-41.9

39-39.9

37-37.9

35-35.9

33-33.9

31-31.9

29-29.9

27-27.9

25-25.9

23-23.9

21-21.9

19-19.9

17-17.9

Gowla 2005( n=175)

15-15.9

Number

Length (cm)

Length (cm)

Figure 3 Length frequency of returning sea trout on the Invermore and Gowla in 2003, 2004 and 2005.

Sea trout weight was available from a subsample of returning sea trout (Table 5). Table 5 includes fish from outwith the SLICE速 trial in order to provide a comparison between systems. While the median condition of sea trout was higher on the Gowla in 2005 than 2004, this was not statistically different (weight: z = -0.772, p = 0.440, n = 109; Condition factor: z = 1.023, p = 0.306, n = 109). There was no statistical difference in weight (z = -1.831, p = 0.067, n = 21) or condition factor (z = -0.141, p = 0.888, n = 21) between sites in 2004.

212

SUMBAWS

Final Report

Table 5 Weight and condition of finnock

Gowla 2004 (n = 10) Gowla 2005 (n = 99) Invermore 2004 (n = 11)

Mean wt (g)

Median wt (g)

Mean K

Median K

Mean WR

Median WR

191 ± 26

223

0.989 ± 0.1

0.96

0.725 ± 0.07

0.71

222 ± 4

220

1.1 ± 0.2

1.1

0.799 ± 0.01

0.80

130 ± 18

112

0.998 ± 0.1

1.05

0.737 ± 0.08

0.78

Age of Upstream Migrants The majority of sea trout returning on both systems were immature finnock (Table 6) returning to freshwater in the same year as their first migration to sea. A smaller number of sea trout spent one full year at sea before returning as 1SW maiden spawners. The highest percentage return of finnock was on the Gowla in 2003 (99%). The highest percentage of 1SW maidens (27%) returned to the Gowla in 2004.

Table 6 Number (and proportion) of sea trout present within each category Age Group Finnock Total 1-sea-wintermaiden Total

2.0+ 3.0+ 4.0+ Finnock 2.1 3.1 4.1 1 SW maiden

2003 33 (25%) 87 (66%) 10 (8%) 99% 1 (<1%) 1 (<1%) 1%

Gowla 2004 3 (20%) 8 (53%) 73% 3 (20%) 1 (7%) 27%

2005 8 (9%) 79 (84%) 4 (4%) 97% 3 (3%) 3%

Invermore 2003 2004 9 (19%) 2 (13%) 34 (73%) 10 (67%) 1 (2%) 2 (13%) 94% 93% 2 (4%) 1 (2%) 1 (7%) 6%

7%

Sea lice infestation data Growth, chalimus abundance, total lice abundance and days at sea were compared between control and SLICE®-treated groups using a Mann-Whitney U-test. Growth and weight values were divided by the number of days at sea and multiplied by 30 to allow comparison between fish. As there were few returning sea trout present in each of the groups in 2004 no statistical analysis was undertaken. In 2005 sufficient fish returned from control and experimental groups on the Gowla group to allow statistical analysis. There were no statistical differences between SLICE®-treated and control groups in terms of growth, weight, chalimus abundance, total lice level or days spent at sea (Table 7).

Table 7 Comparison of median growth, chalimus abundance, total lice and days at sea between control and experimental groups from the Gowla in 2005 (Note: One fish was removed from analysis as it was tagged in 2004 and returned as a 1-sea winter maiden) Parameter Growth (cm) Weight (g) Chalimus (no.)

Total lice (no.)

Days at sea

SLICE® Median Mean 8.3 8.1 ± 0.8 151 152 ± 17 0 3.7 ± 2 6 6.1 ± 2 71 90 ± 2

Median 8.1 141.5 0 7 82

Control Mean 8.6 ± 0.7 141.6 ± 15 3.4 ± 1.4 9.4 ± 3 99 ± 11

Mann-Whitney U-test z = 0.755, n = 22, p = 0.450 z = 0.572, n = 19, p = 0.568 z = 0.571, n = 22, p = 0.568 z = 0.715, n = 22, p = 0.474 z = 0.164, n = 22, p = 0.869

Sea lice levels on returning sea trout from both the Gowla and Invermore are shown in Tables 8 and 9.

No statistical differences were observed, which is perhaps unsurprising, given the small sample size. Median condition factors of the SLICE®-treated and control groups in 2005 were 1.02 and 0.9820 respectively.

These value were not statistically significantly different (Mann-Whitney: z = -0.735, p = 0.462, n = 19).

213

SUMBAWS

Final Report

Table 8 Mean sea lice infestation data on sea trout from SLICE速-treated and control groups in 2004 (both sites) Treatment Control SLICE速

Sea lice infestation SD Median IQR 15.0 0 22.5 12.14 0 17.3

Mean 7.5 6.5

n 4 8

Days at sea Mean Max 26 57 30 88

Max 30 29

Table 9 Mean sea lice infestation data on sea trout from SLICE速-treated and control groups in 2005 (both sites) Treatment Control SLICE速

Sea lice infestation SD Median IQR 10.8 7 22 7.4 6 9

Mean 9.4 6.1

n 11 11

Days at sea Mean Max 99.2 170 90.5 166

Max 30 22

The highest sea lice levels during this study were recorded in 2004 on sea trout returning to the Gowla system. Levels were highest in May and early June and returned to lower levels towards the end of the summer. Highest levels on the Invermore were experienced in late July (Figures 4 and 5).

4

Mean Chalimus

3.5

Mean total lice

3

LN(x+1)

2.5 2 1.5 1 0.5 Late Sept 05

Late Aug 05

Early Sept 05

Early Aug 05

Late July 05

Early July 05

Late June 05

Late May 05

Early June 05

Early May 05

Late July 04

Early Aug 04

Early July 04

Late June 04

Late May 04

Early June 04

Early May 04

Late July 03

Early Aug 03

Early July 03

Late June 03

Early June 03

0

Figure 4 Mean number of Chalimus and total lice on sea trout on the Gowla in 2003, 2004 and 2005.

4

Mean Chalimus

3.5

Mean total lice

3

LN(x+1)

2.5 2 1.5 1 0.5

Late Sept 05

Early Sept 05

Late Aug 05

Early Aug 05

Late July 05

Early July 05

Late June 05

Early June 05

Late May 05

Early May 05

Early Aug 04

Late July 04

Early July 04

Late June

Early June 04

Late May 04

Early May 04

Late July 03

Early July 03

Late June 03

Early June 03

0

Figure 5 Mean number of Chalimus and total lice on sea trout on the Invermore in 2003, 2004 and 2005.

214

SUMBAWS

Final Report

Sea lice infestation data on sea trout were tested to determine whether there were differences between the two bays between years, and secondly whether lice infestation between the two sites was different in each year. Data for the Invermore represents lice infestation data from sea trout in the upstream trap only whereas data for the Gowla combines trap data and data from netting surveys in Bertraghboy bay. Sea lice abundance data includes all fish recorded. Intensity data only includes fish with sea lice and all zero values are removed from the analysis. Sea lice summary statistics are shown in Table 10.

Table 10 Sea lice summary statistics Year n Mean SD Median IQR Abundance Gowla (all fish, including those with zero lice) 2003 156 5.5 13.2 0 4.8 2004 54 35.1 41.3 20.5 64.8 2005 180 9.7 11 6 16 Abundance Invermore (all fish, including those with zero lice) 2003 73 16.8 24.1 3 26 2004 38 9.8 16.8 0 15.5 2005 11 9.5 12.6 3 18 Intensity Gowla (infected fish only) 2003 53 16.2 18.5 9 18 2004 39 49.4 40.7 38 57 2005 124 14 10.6 10.5 15.25 Intensity Invermore (infected fish only) 2003 37 33.2 24.5 26 34.5 2004 15 24.9 18.5 20 26 2005 6 17.5 12.3 15 24.25

Max 81 191 53 121 70 34 81 191 53 121 70 34

The data could not be normalised and therefore non parametric testing was carried out. Sea lice levels were statistically different on the Gowla between all years (p < 0.05). Levels were highest in 2004 with a median value of 20.5 lice per fish. There was little variation in lice levels on the Invermore between years (p > 0.05). Lice levels were also compared between sites each year. In 2003, sea lice levels were statistically higher on the Invermore than the Gowla (Mann-Whitney: z = -3.557, p < 0.001, n = 229). This pattern was reversed in 2004 when statistically higher sea lice levels were experienced on the Gowla (Mann-Whitney: z = -3.473, p = 0.001, n = 92). In 2005 median sea lice levels were not statistically different between the Gowla and the Invermore (Mann-Whitney: z = -0.791, p = 0.429, n = 169).

Because ANOVA is robust to departures from the assumptions to normality and homogeneity of variances (Underwood, 1997), a two-way ANOVA was undertaken to compare sites and years. Both year and site were significant (Table 11).

Table 11 Two-way ANOVA analysis to test site and year variance in sea lice infestation.

Source Year Site Error

df 2 1 275

SS 25.2 4.979 210.254

MS 12.606 4.979 0.765

F 16.488 6.513

p <0.001 0.011

Premature return Upstream traps were monitored from the 1st of February each year to monitor for the presence of premature return of sea trout smolts. Fish were not observed in upstream traps before June 1st on the Gowla in 2003 and 2005 or on the Invermore in 2005 (Table 12). In 2004, almost half of the returns to the Gowla were premature-returning fish which entered the upstream trap before June 1st. Premature return was also observed on the Invermore system in 2004, with 34% of sea trout returning before June 1st. The number of sea trout returning to the Invermore has steadily declined over the last three years with only 11 sea trout returning in 2005 (Table 12). This represented a survival from smolt of 2.4%. In 2004, survival to return was 11.4%. Variation in return rates may be expected with such low smolt runs. A long dry spell in July and 215

SUMBAWS

Final Report

August 2005 prevented sea trout from returning to rivers along the west coast. On the Gowla survival to return was 3.2% in 2004 and 4.1% in 2005. There was a significant difference in mean lice abundance between smolts returning before and after June 1st on the Gowla (t-test; p < 0.0001) but no significant difference on the Invermore.

Table 12 Sea trout premature return to freshwater & finnock marine survival (*possibly underestimated due to flood event, # Mean lice abundances (antilogged))

2003 143* 155 0 0

Smolt run Total return % total return Premature return % premature return Mean lice abundance before June 1st.# Mean lice abundance after June 1st.#

Gowla 2004 1302 42 3.2 19 47.6 19.3 1.1

2005 4231 175 4.1 0 0

Invermore 2003 2004 2005 1391 334 465 75 38 11 5.4 11.4 2.4 3 13 0 4 35.1 0 0.9 3.8

Salmon smolt release programme and adult salmon returns The number of 1SW salmon (grilse) recovered in 2004 from the 2003 smolt release is shown in Table 13.

As only a proportion of the salmon catch is examined for micro-tagged fish each year out of the total catch in the Western (Galway/Limerick) district, the numbers of recovered tags are raised taking account of the total grilse catch to derive an estimate of marine survival to the coast. Estimates of non-catch fishing mortality (drop out of nets, predation) are also included in this estimate.

Table 13 Recoveries of adult salmon in 2004 from 2003 smolt releases Fishery

Group

Gowla A Gowla B Invermore A Invermore B

SLICE® Control SLICE® Control

Number released 4955 4822 4589 4594

Unraised recovery (2004) 35 3 17 9

Raised total (2004) 218 20 83 36

Survival to coast (%) 4.4 0.4 1.8 0.8

There was a statistical difference between salmon survival from the SLICE®-treated and control groups on both the Gowla (chi-square test: p < 0.001) and the Invermore (chi-square test: p < 0.001) in 2004, with more fish returning from the SLICE®-treated group on both systems.

Data on salmon weight were available only for the Invermore. Statistical analysis (Mann-Whitney Utest) was carried out on condition factor, length and weight for the control and experimental groups. There was no statistical difference in condition factor (WR) (p = 0.171, U = 5.0) or length (p = 0.606, z = -0.516, n = 14) between SLICE®-treated and control groups on the Invermore; however, the weight of the returning salmon from the SLICE®-treated group was significantly higher (p = 0.042, z = -2.032, n = 10) (Table 14).

Table 14 Mean length (cm), weight and condition factor of salmon

Invermore Invermore

Group SLICE® Control

Mean length (cm) 62.98 (n = 8) 60.2 (n = 6)

Mean weight (kg) 2.62 (n = 6) 2.05 (n = 4)

K 1.137 (n = 6) 0.962 (n = 4)

WR 1.13 (n = 6) 0.97 (n = 4)

Returning fish from the releases in 2004 were encountered in the commercial fishery from late May 2005 and in both upstream traps (Table 15). A total of 46 salmon returned from the SLICE®-treated group, in comparison to 22 from the control group on the Gowla in 2005. A chi square test revealed that there was a statistically significant increase in salmon survival in the SLICE®-treated group compared to the control (p = 0.037). However there was no difference in survival on the Invermore system (p = 0.431).

216

SUMBAWS

Final Report

Table 15 Recoveries of adult salmon in 2005 from 2004 smolt releases Fishery

Group

Gowla A Gowla B Invermore A Invermore B

SLICE速 Control SLICE速 Control

Number released 4982 5034 4972 4994

Number recovered in trap 10 4 1 0

Number recovered in commercial fishery 36 18 29 22

Unraised recovery 2004 46 22 30 22

Sea Lice Infestation Pressure Data on sea lice levels on salmon farms (Table 16) were combined with farm stock numbers (Table 17) to derive an estimate of sea lice infestation pressure in Bertraghboy and Kilkieran bays during spring 2003 and 2004. Numbers of farmed salmon in their second year of production were multiplied by mean ovigerous and total lice (all mobile stages) numbers in April each year to estimate sea lice infestation pressure from farm sources. This related to the period of release of groups of salmon smolts from the Gowla and Invermore rivers (Figure 6).

Table 16 Sea lice monitoring results from salmon farms (Source: Marine Institute)

26/02/2003 13/03/2003 28/03/2003 10/04/2003 25/04/2003 2003 February March April May

Ovigerous 0 0 0.08 0.03 0 Mean Ovigerous 0.91 0.58 0.17 0.46

Bertraghboy Bay (Gowla) Total 0 11/02/2004 0.02 19/03/2004 3.32 31/03/2004 1.66 14/04/04 0.35 23/04/04 Kilkieran Bay (Invermore) Mean 2004 Mobile 1.93 February 1.89 March 2.56 April 2.3 May

Ovigerous 0.35 1.29 1.93 1.65 1.39 Mean Ovigerous 1.07 1.3 0.46 1.5

Total Mobile 2.72 7.97 7.94 6.28 5.25 Mean Mobile 4.42 7.97 2.68 7.78

Table 17 Estimated numbers of farmed salmon present in spring (Source: Western Regional Fisheries board)

April 2003 April 2004 April 2005

Bertraghboy Bay (Gowla) 800 000 289 000 0

Kilkieran Bay (Invermore) 783 000 260 000 300 000

217

SUMBAWS

Final Report

3000000

2500000

2003 2004

Estimated Number of Lice

2000000

1500000

1000000

500000

0

-500000

Bertraghboy Ovigerous

Kilkieran Ovigerous

Bertraghboy Total

Kilkieran Total

-1000000

Figure 6 Estimated sea lice infestation pressure in Bertraghboy and Kilkieran bays in 2003 and 2004

Discussion L. salmonis is a parasite of salmonids and originates from wild salmon, wild sea trout and farmed salmon (Tully & Whelan, 1993a). During the late 1980â&#x20AC;&#x2122;s, many sea trout fisheries in the west of Ireland experienced a dramatic decline in sea trout catches (Whelan & Poole, 1996, Gargan, 2000). The history of the sea trout stock collapse in western Ireland, and subsequent events, have been well documented (Poole et al., 1996; Gargan, 2000). In 1989, when sea trout stocks collapsed in Western fisheries, sea trout were observed in the lower pools of the Delphi Fishery in Connemara in late May with heavy infestations of juvenile sea lice (Lepeophtheirus salmonis). Sampling of rivers began in 1990 to determine if this phenomenon was widespread and sea trout post-smolts and some sea trout kelts were recorded in all sampled rivers carrying infestations of predominantly juvenile sea lice, indicating recent transmission (Tully et al., 1993). This has been linked with the development of salmon farming in the mid-west zone at that time (Gargan et al., 2002).

The collapse in sea trout populations was characterized by a number of events. Smolts and kelts returned to freshwater prematurely, a proportion was heavily infected with sea lice, some fish were emaciated and a reduction in spawning stock was recorded (Anon, 1991). A correlation was found between L. salmonis and the collapse of sea trout stocks in Ireland (Whelan and Poole, 1996). The increase in sea lice levels has been linked to an increase in fish farming activities in western Ireland (Gargan et al., 2002). A study by Bjorn et al. (2001) also found that salmon lice infection on wild sea trout and Artic char in Norway was significantly higher close to fish farming activity than in unfarmed areas.

In the present study, field studies were undertaken over three years to examine sea trout population dynamics in two rivers in the west of Ireland entering bays with salmon aquaculture. SLICEÂŽ, the commercial in-feed treatment against sea lice, was used to treat wild sea trout and hatchery-reared salmon to determine its efficacy in protecting salmon smolts from sea lice infestation and determine its effects on their marine survival. Because the returns of sea trout from both groups in the SLICEÂŽ experiment were low, the performance of all returning sea trout was examined with reference to the sea lice infestation pressure in both bays over the three year period and general trends are discussed.

Local estuarine topography, migration routes of fish, tidal movements and other factors affecting lice distribution are critical in determining the degree of lice infestation likely to be encountered by salmon and sea trout. However, salmon farm stock numbers and recorded lice levels on farmed fish may provide an index of 218

SUMBAWS

Final Report

lice infestation pressure in both bays (Figure 6) and the performance of returning sea trout can be assessed with regard to this sea lice infestation pressure. Highest lice infestation pressure in Bertraghboy bay (2004) coincided with highest sea lice infestation on sea trout, a high rate of premature return, poor finnock marine growth and low marine finnock survival. High lice infestation pressure in Kilkieran bay in both years also coincided with high lice infestation on sea trout, premature return and poor marine growth.

Sea trout post-smolts have been observed ascending Irish estuaries and rivers only weeks after smolt migration (Tully et al., 1993a, 1993b), a phenomenon termed ‘premature return’. This behavior has been interpreted as a response by sea trout to reduce stress caused by an infection and to enhance survival (Bjorn et al., 2001). Many post-smolts exhibiting this behavioral response were heavily infested with sea lice. Sea lice infestation affects osmoregulation but return to estuarine water may enable the fish to retain its osmotic balance and hence improve survival (Birkeland, 1996, Wagner et al., 2004). Poor marine growth is likely to be the result of the short duration of post-smolt marine feeding but could also be the energetic cost of the louse infestation (Birkeland, 1996). Reduced growth resulting from lice infestation may reduce resources available for egg production and may reduce fecundity and reproductive success (Birkeland, 1996).

In the absence of farmed salmon in Bertraghboy bay in 2005, a very low lice infestation on sea trout, absence of premature returns, highest recorded marine growth and an increase in marine survival was observed. In 2005 sufficient fish returned from control and experimental groups on the Gowla group to allow statistical analysis. Returns revealed no statistical difference between SLICE®-treated and control groups in terms of growth, weight, chalimus abundance, total lice level or days spent at sea; results which might be expected in the absence of farmed salmon in the bay in 2005. These results are consistent with those of previous studies (Birkeland 1996; Birkeland & Jakobsen 1997; Bjorn et al, 2001; Butler & Watt 2002, Gargan et al. 2002) indicating the negative effects of lice infestation from marine salmon farms on sea trout populations.

Release of SLICE®-treated and control groups of hatchery-reared salmon smolts into aquaculture bays allows assessment of the efficacy of an in-feed sea lice treatment to be tested. While hatchery salmon smolts are larger than wild salmon smolts, they are likely to be broadly representative and treated and control groups and should reflect conditions in the marine environment. Bjorn & Finstad (1997) have shown a correlation between smolt size, sea lice infestation and osmoregulatory disturbances, indicating that smaller fish are more likely to suffer physiological compromise at a lower level of sea lice infestation. These data suggest that any differences in survival between groups of treated and untreated hatchery smolts are likely to be manifested to a greater degree with wild smolts in the same environment.

Results from the present study reveal a significant difference in the return rate of SLICE®-treated and control salmon smolts in three of four release groupings; these data suggest reduced mortality of smolts in the treated groups and hence perhaps protection from sea lice infestation in aquaculture bays. Similar results were found by Skilbrei & Wennevik (2005) using salmon smolts released in June in the River Dale, western Norway.

Previous studies in Ireland (Tully and Whelan, 1993, Gargan et al., 2002), Scotland (Butler 2002) and Norway (Heuch and Mo, 2001) have indicated that in spring, the majority of nauplii arise from ovigerous lice infesting farmed salmon. Tully et al. (1999) have demonstrated that the presence of salmon farms significantly increased the level of sea lice infestation on sea trout post-smolts. Similar findings have been reported from Norway (Grimnes et al., 2000) and Scotland (Mackenzie et al., 1998, Butler, 2002). Given the presence of a significant source of sea lice infestation from marine salmon farms in both bays in the present study, increased mortality of salmon smolts can be expected because previous studies have found that the feeding activity of sea lice causes mechanical damage such as skin and fin erosion (Dawson et al., 1998; Bjørn and Finstad, 1997), osmotic stress and death (Grimnes and Jakobsen, 1996; Nolan et al. 1999; Finstad et al., 2000).

Research has indicated that lice infestation of 30 or more pre-adult sea lice can affect osmoregulation of salmon post-smolts (Grimnes & Jakobsen, 1996), and as few as 11 lice or more on salmon smolts can cause mortality (Finstad et al., 2000). Both Anderson & Gordon (1982) and Bjørn et al. (2001) noted excessive mortality of the most heavy infested fish, suggesting that these fish die at sea. In areas with epizootics, lice have been implicated in the mortality of 32-47% of all migrating sea trout smolts (Bjørn et al. 2001) and 4886% of wild salmon smolts (Holst and Jakobsen, 1998).

The greater returning weight exhibited by the SLICE®-treated adult salmon in Invermore could be expected from enhanced growth of fish protected of high sea lice infestation (Johnson et al. 2004). Skilbrei & Wennevik (2005) found a similar difference in weight in SLICE®-treated hatchery salmon smolts released in Norway.

Gargan et al. (2002) demonstrated a statistical relationship between lice infestation on sea trout and distance to the nearest salmon farm. A similar trend has been recorded in Scottish (Butler & Watt 2002; Mackenzie et al. 1998) and Norwegian studies (Anon 1997; Birkeland and Jakobsen 1997; Bjorn et al.

219

SUMBAWS

Final Report

2001). Tully et al. (1999) suggest that there is substantial retention of larvae within the bay in which they are produced by selective tidal stream transport, a suggestion supported by the findings of Gargan et al.

(2002). Most records available suggest that salmon post-smolts move relatively quickly into the ocean, swimming as fast as 2 body lengths s-1 (Lacroix & McCurdy, 1996). A migration route distance of 12-15 km from release point to the open ocean can be assumed for both bays in the present study, indicating that salmon post-smolts could reach the open ocean, outside the influence of marine salmon farms, within 1-2 days, a period much shorter than that for migration through elongated and large fjordic systems. Thus the findings of the present study support the view of substantial retention of larvae within bays and demonstrate the potential impact of marine salmon farms on salmon post-smolt marine mortality even in very short coastal bays.

References Anderson, R.M. & Gordon, D.M. (1982). Processes influencing the distribution of parasite numbers within host populations with special emphasis on parasite influenced host mortalities. Parasitiology 85, 373-398.

Anon (1991) Report of the Sea Trout Working Group. Fisheries Research Centre, Department of the Marine, Dublin, 40pp.

Anon (1994) Report of the Sea Trout Working Group, 1993, Department of the Marine, Dublin.

Anon (1994a) Report of the Sea trout task force, Department of the Marine, Dublin.

Birkeland, K. (1996) Consequences of premature return by sea trout, Salmo trutta, infested with the salmon lice Lepeophtheirus salmonis Kroyer: migration, growth and mortality. Canadian Journal of Fisheries and Aquatic Science 53, 2808-2813.

Birkeland, K. and Jakobsen, P.J. (1997) Salmon lice, Lepeophtheirus salmonis, infestation as a causal agent of premature return to rivers and estuaries by sea trout, Salmo trutta, juveniles. Environmental Biology of Fisheries 49, 129-137.

Bjorn and Finstad (1997) The Physiological effects of salmon lice infection on sea trout post smolts.

Nordic Journal of Freshwater research 73, 60-72.

Bjorn, P.A., Finstad, B. and Kristoffersen, R. (2001) Salmon lice infection of wild sea trout and Arctic char in marine and freshwater: the effects of salmon farms. Aquaculture Research 32, 947-962.

Bowers, J.M., Mustafa, A., Speare, D.J., Conboy, G.A., Brimacombe, M., Sims, D.E. and Burka, J.F. (2000) The physiological response of Atlantic salmon, Salmo salar L., to a single experimental challenge with sea lice, Lepeophtheirus salmonis. Journal of fish Diseases, 23, 165-172.

Boxshall, G. A. (1976). The host specificity of Lepeophtheirus pectoralis (Muller, 1776) (Copepoda: Caligidae). Journal of Fish Biology 8, 255-264.

Butler, J.R.A. & Watt, J. (2002) Assessing and managing the impacts of marine salmon farms on wild Atlantic salmon in western Scotland: identifying priority rivers for conservation. In: Salmon on the Edge (ed. D. Mills), pp. 93-118. Blackwell Science, Oxford, UK. Dawson, L.H.J., Pike, A.W., Houlihan, D.F. & McVicar, A.H. (1998). Effects of salmon lice Lepeophtheirus salmonis on sea trout Salmo trutta at different times after seawater transfer. Diseases of Aquatic Organisms 33; 179-186.

Finstad, B., Bjorn, P.A., Grimnes, A. and Hvidsten, N.A. (2000) Laboratory and field investigations of salmon lice (Lepeophtheirus salmonis Kroyer) infestations on Atlantic salmon (Salmo salar L) post-smolts.

Aquaculture Research 31, 795-803.

Gargan, P.G. (2000) The impact of the salmon louse (Lepeophtheirus salmonis) on wild salmonids in Europe and recommendations for effective management of sea lice on marine salmon farms. In : Aquaculture and the protection of wild salmon (P.Gallaugher, P. and C. Orr eds.) Workshop Proceedings, July 2000, p37-46. Simon Fraser University, Vancouver, British Columbia, Canada.

Gargan, P.G., Tully, O. and Poole, W.R. (2002) The relationship between sea lice infestation, sea lice production and sea trout survival in Ireland, 1992-2001. In: Salmon on the Edge (ed. D. Mills), pp. 119-135.

Blackwell Science, Oxford, UK. 220

SUMBAWS

Final Report

Grimnes, A. and Jakobsen, P.J. (1996) The physiological effects of salmon lice infection on post-smolt of Atlantic salmon. Journal of Fish Biology, 48, 1179-1194.

Grimnes, A., Finstad, B., and Bjorn, P.A. (2000). Registrations of salmon lice on Atlantic salmon, sea trout and charr in 1999. NINA Oppdragsmelding 634, 1-34. (In Norwegian with English Abstract).

Heuch, P.A., and Mo, T.A. (2001). A model of salmon louse production in Norway: effects of increasing salmon production and public management measures. Diseases of Aquatic Organisms 45, 145-152.

Holst, J.C., and Jakobsen, P.J. (1998). Dodelighet hos utvandrende postsmolt av laks som folge av lakselusinfeksjon. Fiskets Gang 8, 13-15. (In Norwegian).

Johnson, S. C., Treasurer, J. W., Bravo, S., Nagasawa, K., Kabata, Z. (2004). A Review of the impact of parasitic copepods on marine aquaculture. Zoological Studies 43, 229-243.

Lacroix, G.L. & McCurdy, P. (1996) Migratory behaviour of post-smolt Atlantic salmon during initial stages of sea word migration. Journal of Fish Biology, 49, 1086-101.

Mackenzie, K., Longshaw, M., Begg, G.S., and McVicar, A.H. (1998). Sea lice (Copepoda: Caligidae) on wild sea trout (Salmo trutta L.) in Scotland. ICES Journal of Marine Science 55, 151-162.

McMahon, T. (2000) Regulation and monitoring of marine aquaculture in Ireland. Journal of Applied Ichthyology, 16, 177-181.

Nolan, D. E. T., Reilly. P. & Wendelaar Bonga S.E. (1999). Infection with low numbers of the sea louse Lepeophtheirus salmonis (Kroyer) induces stress-related effects in post-smolt Atlantic salmon (Salmo salar L.). Canadian Journal of Fisheries and Aquatic Sciences 56, 1-14.

Poole, W. R., Whelan, K.F., Dillane, M.G., Cooke, D.J. and Matthews M.(1996) The performance of sea trout, Salmo trutta L., stocks from the Burrishoole system Western Ireland, 1970-1994. Fisheries Management and Ecology 3, 73-92.

Poole W.R., Nolan, D.T. and Tully, O. (2000) Modelling the effects of capture and sea lice (Lepeophtheirus salmonis (Krøyer)) infestation on the cortisol stress response in trout. Aquaculture Research, 31, 835-841.

Skilbrei, O.T and Wennevik, V. (2005) Survival and growth of sea-ranched Atlantic salmon treated against salmon lice prior to release. ICES Journal of Marine Science, In Press.

Tully, O. and Whelan, K.F. (1993) Production of nauplii of Lepeophtheirus salmonis (Kroyer) (Copepoda: Caligidae) from farmed and wild salmon and its relation to the infestation of wild sea trout (Salmo trutta L.) off the west coast of Ireland in 1991. Fisheries Research 17, 187-200.

Tully, O., Poole, W.R., Whelan, K.F. and Merigoux, S. (1993) Parameters and possible causes of epizootics of Lepeophtheirus salmonis (Krøyer) infesting sea trout (Salmo trutta) off the west coast of Ireland. In: Pathogens of Wild and Farmed Fish: Sea lice (ed. By G.A. Boxshall & D. Defaye). pp. 30. Ellis Horwood, Chichester, U.K. Tully, O., Poole, W.R., Whelan, K.F. (1993) Infestation parameters for Lepeophtheirus salmonis (Kroyer) parasitic on sea trout (Salmo trutta L.) postsmolts off the west coast of Ireland during 1990 and 1991.

Aquaculture and Fisheries Management 24, 545-557.

Tully, O., Poole, W.R., Whelan, K.F. and Merigoux, S.(1993a) Parameters and possible causes of Lepeophtheirus salmonis (Kroyer) infesting sea trout (Salmo trutta L.) off the west coast of Ireland. In: Pathogens of Wild and Farmed Fish: Sea Lice (ed. Boxshall, G.A. and Defaye, D.), pp.202-213. Ellis Horwood, Chichester.

Tully, O and Poole, R. (1999) Parameters and impacts of sea lice infestation of sea trout along a salinity gradient in clew Bay, Ireland. Report to the Atlantic Salmon Trust, Pitlochry and the Salmon Research Agency, Newport, Ireland.

Underwood, A, J. (1997) Experiments in Ecology: their logical design and interpretation using analysis of variance. Cambridge University Press. Cambridge, U.K. Wagner, G.N., McKinley, R.S., Bjorn, P.A. and Finstad, B. (2004) Short-term freshwater exposure benefits sea lice-infected Atlantic salmon. Journal of Fish Biology, 64, 1593-1604.

221

SUMBAWS

Final Report

Went, A.E.J. (1949) Sea trout of the Owengowla (Gowla River). Scientific Proceedings of the Royal Dublin Society 25, 55-64.

Went, A.E.J. (1962) Irish sea trout, a review of investigations to date. Scientific Proceedings of the Royal Dublin Society 1A(10), 265-296.

Whelan, K.F. and Poole, W.R. (1996) The sea trout collapse, 1989-92. In: The conservation of aquatic systems (J.D. Reynolds, ed.). Proceedings of seminar held on 18-19th February, 1993. Royal Irish Academy, Dublin, pp 101-110.

Wootten, R., Smith, J.W. and Needham, E.A. (1982) Aspects of the biology of the parasitic copepods Lepeophtheirus salmonis and Caligus elongates on farmed salmonids, and their treatment. Proceedings of the Royal Society of Edinburgh 81B, 185-197.

222

SUMBAWS

Final Report

Workpackage 8 Lead partner: USTAN Participating Partners: ESRI, CFB, NINA, NIFA, KUN, NCFS OBJECTIVES • • • •

Integration of experimental data to identify lethal and sublethal threshold levels of lice infestation for Atlantic salmon and sea trout smolts. Inputs from WPs 5-7.

Assessment of likely vulnerability of migrating smolts to lice infestation with respect to the location of aquaculture cages. Inputs from WPs 2,3,4.

Socio-economic analyses of the costs and benefits of achieving identified lice loadings on farmed salmon. Inputs from WPs 1,4,5,6 and integration of data with the Norwegian NFR Villaksprogram mathematical model.

Appraisal of the effectiveness of Substance EX as a remedial prophylactic treatment in protecting wild smolts from initial infestations by sea lice. Inputs from WP 7, with relevant supportive data from WPs 1,2,3,4.

DELIVERABLES • •

•

Annual monitoring of infestations on wild adult salmon and identification of the levels of sea lice infestations which cause lethal stress and significant, but sublethal, stress to smolts of both salmon and sea trout.

Modelling the cost-benefit analysis in attaining/maintaining given levels of infestations on farmed stocks. Recommendations for farmed stocks in isolation (maximizing fish health) and for farmed stocks with respect to further reductions in levels of infestation that will minimize putative consequential impacts on wild stocks.

Scientific and socio-economic appraisal of the viability and desirability of interventory prophylactic bath-treatment of wild salmonid smolts as a sustainable strategy for enhancing small, definable wild stocks and fisheries.

223

SUMBAWS

Final Report

GENERAL RECOMMENDATIONS 1.

Farming of Atlantic salmon (Salmo salar L.) commenced in the 1970s and worldwide production has increased steadily to >1.1 million tonnes, with the majority in the Northern Hemisphere and largely in the NE Atlantic. Norway remains the single largest producer nation in the Northern Hemisphere.

2.

Lepeophtheirus salmonis (Krøyer) is a specialist ectoparasite of salmonids in their marine phase and infests all species of wild salmonids in both the North Pacific and North Atlantic oceans, and quickly became a major pest at aquaculture sites during the 1970s. Furthermore, farmed and wild Atlantic salmon in the North Atlantic also are impacted by the host generalist Caligus elongatus (Nordmann), which has been recorded from >80 species of elasmobranch and teleost host species. In the North Pacific ocean, there are 14 species of caligid copepods (two Caligus spp. and 12 Lepeophtheirus spp.) parasitizing many different species of marine fish. However, only L. salmonis and Caligus clemensi (Parker & Margolis) are thought to pose a potential threat to both farmed and wild salmon, although Lepeophtheirus cuneifer Kabata is common on farmed salmon in some regions.

3.

L. salmonis and C. elongatus together are the most economically important metazoan pathogens to S. salar aquaculture in the North Atlantic, and the most recent estimates of the worldwide costs to the industry of caligids range up to 100 million US $ p.a. For the Scottish industry alone, the total cost is at least £25 million p.a.

4.

In NW Europe there has been a long-standing controversy regarding the extent to which cultured salmon are a significant source of sea louse infestations of both wild sea trout (Salmo trutta L.) and Salmo salar populations.

5.

L. salmonis can cause stress, pathological damage and death of the host fish, and the small post-smolt stage is perhaps the most vulnerable. Concerns about the possible detrimental impacts of caligids on populations of wild salmonids first arose from observations in Ireland (and thence Scotland and Norway) of stock collapses of S. trutta in the late 1980s/early 1990s, and apparently associated epizootics of L. salmonis, particularly on juvenile hosts. Heavily infested fish often showed “premature” migratory return to freshwater, were in poor condition and bore heavy infestations of the sessile larval chalimus stages of the parasite. More recently, epizootics on wild, juvenile pink and chum salmon in British Columbia have raised similar concerns for Pacific species.

6.

An early resolution of the controversy surrounding the caligid infestation interactions between farmed and wild salmonids would have been readily achievable if it were possible to unequivocally determine, by direct means, the source of infective copepodids of L. salmonis parasitizing wild host fish.

7.

Despite the intractability of physically tracking individuals of the planktonic, infective, larval sea lice stages from initial hatching to ultimate settlement on a host fish, the evidence is now persuasive that farm-origin caligids do impact wild salmonids, and almost certainly have at least contributed to recent wild salmonid stock declines proximate to salmon farms in areas of Norway, the British Isles, and W Canada. That evidence (e.g. modeling of sea lice population dynamics, plankton surveys of incidence and seasonality of larval sea lice in relation to farming activity, and population genetic studies), although very strongly indicative, is nonetheless indirect and primarily circumstantial or correlative. As the deliverables from SUMBAWS are published in the primary literature this evidence will be considerably strengthened. It is well-established that sea lice can exert marked detrimental impacts on 224

SUMBAWS

Final Report

salmonids, and few would argue that sea lice from salmon farms do not infect wild fish. But the quantitative extent of those interactions, and the direct implication of intensive aquaculture as a contributory causal factor in certain salmonid stock declines, do remain contentious issues.

8.

Our overall economic impact analyses for Ireland, Scotland and Norway showed that, in terms of cash value and job numbers, commercial fishing for wild salmon is of minor importance to all three countries. Freshwater game angling is especially important in Norway and Scotland, but the size differentials of the aquaculture and game angling industries in these two countries are such that the cash economic value of aquaculture is approximately 2.7 times more important in Scotland, and 11 times more important in Norway.

9.

Game angling is relatively labour intensive and the estimates range from 3.3 jobs per 100 fish caught in Norway, to 4.5 jobs per 100 fish caught in Scotland. Ireland was intermediate at 3.8 jobs per 100 fish.

Especially in rural and remote areas, game angling is therefore a significant source of employment, and access to healthy and productive wild salmonid stocks arguably remains a very important facet of the cultural heritage and quality of life of many individuals.

10. Concerning the sea lice impacts of farm salmon on wild stocks of salmonids, the data are not yet available to allow modelling of the quantitative contribution of farmed salmon larval production to infestations of wild salmonids. In some respects an ideal would be that sea lice levels on farms are maintained at zero ovigerous females year-round (but see also paragraph 31, below). By that means, it could be assured that farms are not subject to self-reinfestation, that separate farms do not cross-infest, and that wild salmonids are protected from sea lice infestations of farm origin. The economic (chemicals and labour) and environmental (pollutant discharge) costs of achieving zero ovigerous lice year-round on farmed salmon are, however, almost certainly unacceptable at the present stocking levels and with the current two-year production cycle. Co-ordinated treatment strategies within fjords and sealochs are, however, undoubtedly effective in reducing infestations and may have significant beneficial economic effects. It is recommended that the more appropriate objective be adopted of achieving zero ovigerous sea lice on salmon farms during the spring-autumn period, including both the initial migration to sea of sea trout and salmon smolts, and the return to freshwater of juvenile sea trout. The timing of this critical period will be reflected by the migratory seasonality of wild fish (salmon, sea trout, char) and will differ in Ireland and Scotland, as well as latitudinally throughout southern to northern Norway. In Ireland, for example, this period may extend from February to October, whereas in northern Norway it could be confined to May to September.

11. In specifically adopting a precautionary perspective, we considered it expedient to estimate by modelling the economic costs of a single further, marginal sea louse treatment beyond that typically undertaken by the industry. The assumption was that such a strategically-deployed additional treatment would have beneficial consequences to wild salmonid stocks in reducing infestation pressures on fish outwith the farm environment. Co-ordination and synchronisation of treatments within fjords, sealochs and bays might elevate the economic benefit to farms to a greater extent than presently estimated, but for a two-year production cycle, the analysis showed that an additional treatment might raise production costs of fish to slaughter by the order of 0.4% (Ireland) to 0.6% (Scotland). In Scotland especially there may be critical and additional environmental and regulatory considerations, because farms often are deploying all of their licenced discharge consent for chemotherapeutants during a typical production cycle.

225

SUMBAWS

Final Report

12. Cost-benefit analyses were undertaken to allow a fundamental evaluation of the principle of intercepting and treating wild smolts to protect them from sea lice infestation during the early weeks of their marine residence. The analysis included the costs of constructing and operating fish traps in small river systems in relation to the perceived benefits of increased adult returns. The analysis was costed for western Ireland and the unequivocal outcome was that the costs were unacceptably high in relation to perceived benefit, and in terms of the local cash expenditure sustained by game anglers. The recommendation from the socio-economic modelling alone is that the adoption of an interventory prophylactic treatment of smolts in wild fisheries against sea lice infestation is not an appropriate management or enhancement strategy (but see also Paragraph 24).

13. Acoustic telemetry of migrating smolts was efficiently and successfully achieved with VEMCO V8SCGL tags and VEMCO VR2 automatic data logging units. The equipment all performed to specification and expectation. It is recommended that this system is very well suited to the purposes of acoustically logging tagged fish in semi-enclosed fjords. Manual tracking provided a great deal of useful and detailed data for individual fish, but the acquisition of such data is extremely labourintensive. Automated logging systems detect and record tagged fish within prescribed operational detection limits, but where the coastal topography is appropriately narrow this technology is very efficient in recording the passage and thereby the migration speeds of large numbers of acoustically tagged fish. Coastal coverage can be improved by mooring the logging units in deeper waters and such has been shown to be possible to 300 m depth with this equipment.

14. Telemetry studies of tagged salmon smolts in Norwegian fjords confirmed that they migrate seawards in the surface few metres and that Atlantic salmon and sea trout smolts/post-smolts occupy or migrate through different areas of the fjord. Net catches confirmed that the exploitation of nearshore (“littoral”) habitats by sea trout juveniles differs from salmon smolts, which tend to migrate along offshore (“pelagic”) corridors of the fjord. Post-smolts of either species are vulnerable to sea lice infestation because lice larvae tend to aggregate at the brackish layer interface where the smolts commonly are located. Smolts of either species also are vulnerable to predation from wild gadids (especially cod and saithe) as well as sea gulls immediately following descent from their natal river. Given these various sources of mortality on first entering seawater, the survivorship of tagged smolts during migration through the fjords was often poor. Based on data on lice infestation pressures (sentinel cages), on observations of the migratory pathways, migratory speeds and residence times of salmon and sea trout smolts in definable regions of the experimental fjords, a model was developed to predict likely infestation and survivorship of fish during the early stages of their marine migration.

15. Infestation results for one fjord system (the Karlsøy/Romsdals/Lang/Eresfjord system) indicated that the farm-free “protection” zone was effective in reducing impacts of sea lice on resident sea trout and migrating salmon smolts. Infestation intensities on sea trout in the farm-free zone were, however, still high and perhaps outwith the levels that we would recommend as being critical. It is recommended that additional farm protection zones are evaluated in other fjord systems, as a precautionary measure to protect wild salmonids whilst also permitting the continuance of the aquaculture industry.

16. Previous studies reported that infestation of juvenile salmonids with sea lice can induce osmoregulatory dysfunction and stress of the host fish, but those studies were restricted in their quantification of stress markers or physiological correlates thereof. Here, we investigated a comprehensive suite of 226

SUMBAWS

Final Report

osmoregulatory, metabolic and stress markers for salmon and sea trout post-smolts. Because caligid copepods are marine parasites, which do not survive in freshwater, the possibility has been raised that premature migratory return of infested post-smolts to freshwater might be physiologically beneficial to juvenile sea trout. Specifically we assessed the consequences for the host of return to freshwater at 19 days post-challenge with copepodid larvae.

17. Significant sea lice effects, which were consistent across all physiological markers, were apparent once L. salmonis developed to the mobile stages; but some significant effects were observed when the lice were at the younger, (sessile) late chalimus stages. Plasma chloride, lactate and cortisol concentrations all were significantly elevated by infestation, and liver glycogen concentration was significantly reduced for infested fish held throughout in seawater. Following experimental transfer of fish back to freshwater, these physiological measures returned to control levels, whereas significant differences remained for those held throughout in seawater. From these data, premature migratory return to freshwater of sea trout infested with L. salmonis does confer significant physiological benefits.

18. A range of analytical modelling procedures was applied to physiological stress data obtained for wild sea trout and for farmed salmon post-smolts, with the overall objective of quantifying the lice loadings that trigger, or induce, sublethal stress of the host fish. It is recommended that the particular stress markers utilized here (plasma chloride, glucose, cortisol and total osmolality and liver glycogen levels), and the statistical modelling techniques applied to the data, both are effective in determining sublethal stress of fish species for which husbandry and welfare conditions are a major concern. For wild sea trout smolts (mean weight 37 g) a piecewise modelling approach allowed the objective identification of a threshold level of 13 mobile lice ¡ fish-1. From extensive monitoring of sea lice infestations of wild sea trout captured from sites in Ireland, >50 km distant from salmon farms, we find that ~9% carry a total of â&#x2030;Ľ13 sea lice (chalimus and mobile stages combined) (Figure 1).

1000

Number of sea trout

900

Total sea lice (chalimi + mobile stages)

800 700 600 500

Fish exceeding the recommended threshold of 13 sea lice¡fish-1

400 300 200 100 0 0

1-12

13-19

20-29

30-39

40-49

50-59

60-69

70-79

80-89 90-155

Number of sea lice. fish -1

Figure 1 Infestation intensities (range 1-155) of caligid ectoparasites on 1208 Irish sea trout sampled >50 km from farms. 9.4% carried 13 or more total sea lice (chalimi + mobile stages).

227

SUMBAWS

Final Report

19. On this basis, it is recommended that a level of 10% or fewer of wild sea trout in any given population in Ireland bearing total infestations of ≥13 lice · fish-1 should be adopted as indicative of a satisfactory or acceptable lice loading. Within any given sea trout stock, frequencies of heavilyinfested juvenile sea trout (i.e. those ≥13 lice · fish-1) >10% should perhaps be considered a cause for concern. A figure of >10% might, for example, be utilized by regulatory authorities as a trigger level requiring obligatory interventory treatment of sea lice on local salmon farms. It is essential to note, however, that it can be difficult in Scotland and Ireland to capture wild sea trout in the marine environment, although this problem is not so acute in Norway. The avoidance of bias in the sampling of these fish is an essential pre-requisite to evaluations as to whether or not a given stock or population is threatened. In Scotland, for example, sea trout are seldom readily or reliably caught in numbers away from river estuaries and/or brackish water influences. They are most easily captured in brackish estuaries or at river outfalls because at least components of the population have a tendency to aggregate there. In brackish water, the mobile preadult/adult stages of L. salmonis (and especially the adults of C. elongatus) have a marked tendency to detach from the host fish. Fish monitored in brackish waters (or close to river outfalls) may, therefore, show consistently lower infestations of mobile sea lice than the prevailing levels of local infection. Gill-netted fish also can lose parasites (and scales) in the nets.

However, the sessile chalimus stages are not so readily lost and hence the recommendation would be that the total count (of chalimi and adults) should apply in setting this threshold infestation intensity. The critical proportion of fish exceeding the ≥13 lice · fish-1 threshold may vary among geographic localities. Evaluations of Scottish and Norwegian data for sea trout captured >50 km from salmon farms are necessary. For example, our initial assessments of Norwegian data show that juvenile (finnock) sea trout in some farm-free areas do typically show >10% bearing ≥13 lice · fish-1. Also it is certain that a single yardstick will not be applicable throughout Norway. It is very likely that sea trout in the southern, central and northern coastlines of Norway will show consistent and geographically predictable differences in background lice infestation levels. Thus, whilst we would recommend 13 as the threshold infestation intensity, the acceptable proportion of fish exceeding that threshold should be evaluated separately for the five areas (Ireland, Scotland, southern/central/northern Norway).

20. For farmed Atlantic salmon post-smolts the stress markers and piecewise modelling approach indicated a rather higher threshold of ~20 lice · fish-1 resulting in sublethal stress. These experimental post-smolts were considerably larger than wild salmon smolts (which typically are 12-15 cm long and ~20-30 g in wet weight when they first enter seawater), but their mean weight (97 g) is typical of the sizes of smolts that the aquaculture industry first put to sea. Such a level of infestation should require immediate interventory treatment, both from the perspective of fish welfare/husbandry, and as a means of ensuring optimal growth and quality of fish. It should, however, be emphasized that such levels of sea lice infestation of farmed post-smolt salmon would be extremely unusual in the first year of the two-year production cycle and will, in all probability, never pertain on a well-managed farm installation. On this basis, it is recommended to the aquaculture industry that a mean infestation of 20 lice.post-smolt-1 be applied as a threshold level, and that this should be indicative of sublethal stress to cultured juvenile salmon.

21. Parasite infestations of hosts generally do not show a normal distribution of variation among individual hosts. Typically, parasite populations show “over-dispersion”, or “aggregation” on certain individual hosts. That is, some, many or most hosts are parasite-free, but a small number of hosts carry exceptionally heavy infestations. Even when prevalence is 100% (as, for example, has been shown here 228

SUMBAWS

Final Report

consistently for L. salmonis infesting wild adult salmon), the parasites typically are over-dispersed amongst individual host fish. In some instances farm infestations can be less over-dispersed, but if fish within a cage are randomly sampled (including runts, and perhaps malformed and poor condition fish) it is highly likely that over-dispersion will pertain. From a statistical viewpoint it is quite inappropriate to calculate the arithmetic mean and standard deviation of infestation intensities if the data are not normally distributed. Transformations of the data are therefore almost inevitably required prior to the calculation of the normal mean and error terms. A log transformation usually will stabilize the variance and render the error terms normal. The back-transformed mean will, however, always be lower than the arithmetic mean in such cases. It is accordingly recommended that infestation data are inspected for departure from normality and that means of transformed infestation intensities are utilised in monitoring lice infestations on farmed and wild salmonids. This will be essential in establishing whether or not given populations show infestation levels in excess of the above, objectivelyderived threshold intensities. Power analysis also should be undertaken to ascertain the minimum levels of replication (of fish) that are necessary to provide an appropriately reliable measure of mean infestation intensity.

22. Prophylactic treatment of downstream-migrant smolts, to provide protection from early infestation by sea lice, was variously undertaken in Ireland (SLICE®; hatchery-reared salmon, wild sea trout), Scotland (Substance EX; wild sea trout) and Norway (Substance EX; hatchery-reared salmon). At the time of completing the Final Report not all fish tags have been returned and decoded. From the results to date it is, however, apparent that treatment of both species with either the unlicenced experimental Substance EX, or the licenced emamectin benzoate (SLICE®), gave varied results. The returns of tagged sea trout in Scotland were consistently very poor and their low survivorship at sea – irrespective of whether they received protective treatment for sea lice infestation – is itself a cause for great concern. Marine mortality of juvenile and adult salmonids is attributable to numerous factors (including predation by fish, birds and mammals, diseases, parasites, targeted commercial and by-catch fisheries and game anglers) and sea lice alone are not the only important influence. Those sea trout that were recaptured in Scotland did show significantly higher condition factor compared with untreated controls, but these results are tentative because of the small sample sizes.

23. In Ireland there were no significant differences between treatment and control groups of sea trout. For salmon, there were significant increases in survivorship and return of SLICE®-treated smolts, but the data are inadequate to assess these for differences in condition factor because most recaptures were from the commercial net fishery and, as a result, lengths and weights of individual fish are not available.

24. In Norway, Substance EX-treated hatchery-reared salmon smolts released in two river systems in 2003 and 2004 showed consistent results. Overall, there was no significant difference in survivorship or condition factor of returning 1SW salmon, but males were of significantly higher condition. Treated salmon did, however, show significantly increased length and weight over control fish. Multiple releases of large groups of Substance EX-treated and untreated sea trout in Norway since 1996 (prior to SUMBAWS) have shown a variable pattern of response to prophylactic treatment against sea lice infestation. In years of high sea lice infestation pressure there has been significant improvement in survivorship and return of Substance EX-treated fish. In years of relatively low sea lice infestation pressure there were no differences in survivorship, but Substance EX-treated fish did return with 229

SUMBAWS

Final Report

significantly higher condition factor. Female fish of higher condition factor should show increased fecundity and this could be viewed as beneficial in elevating female egg deposition. However, the socio-economic analyses reported (paragraph 12, above) showed that the costs of installing and running fish traps, and of treating fish (especially with SLICE®), outweighed the benefits in terms of numbers of returning fish. One of the major drawbacks of utilizing SLICE® in this way is the need to retain fish in freshwater cages and to feed them SLICE®-treated pellets for at least a week prior to release.

Substance EX provides the opportunity of retaining fish only for 30 minutes in a bath treatment.

Substance EX was, therefore, expected to be the more appropriate treatment in this regard, but again, the benefits were outweighed by the costs. Previous and other studies in Norway have indicated a significant benefit of Substance EX treatment on the survivorship of sea trout post-smolts. Its use in conserving severely threatened stocks in small rivers or those on the verge of extinction would be justified, but from a purely financial perspective the costs of constructing and operating traps will always be considerable. Netting (and treatment) of downstream-migrating smolts is an alternative approach in small river systems; but the unpredicatibility of smolt run timing, and that smolts often migrate at night and in high water conditions, all would engender uncertainties of success and incur considerable labour costs. Accordingly, we would make two distinct recommendations on the management use of Substance EX, dependent on the size of the river system, presence/cost of installing permanent interceptory trapping facilities, and extinction status of the stock. First, it is impossible to place a financial cost on conserving a stock that might otherwise go extinct; for such cases, Substance EX would appear to be a valuable management tool. Second, for the enhancement of stocks not threatened with extinction, the interception and treatment of downstream-migrating smolts is not likely to be financially cost-effective.

25. Annual monitoring of sea lice (L. salmonis, C. elongatus) infestations over the past 8 years at a commercial fishery on the north coast of Scotland has allowed the determination of the ranges of intensities of infestation for both species infecting adult wild salmon. L. salmonis showed 100% prevalence and typically high mean infestation levels (18-32 preadult/adult lice · fish-1). C. elongatus occurred at 92-100% prevalence and typically was of significantly lower intensity (3-24 adult lice · fish-1) than L. salmonis. Detailed analyses of the two parasite species showed that infestation abundances are not driven by host condition factor, and that although both show an apparent inverse temporal pattern of abundance over the time-series there is no evidence to suggest a competitive or agonistic association between the two on individual host fish. Within any one year, individual hosts that carry a high abundance of the one species tend to carry a high abundance of the other. Variation in infestation of adult salmon appears to be attributable to certain individual fish being (by chance) more vulnerable to, or at greater risk from, infestation by sea lice. It has to be emphasized that these caligids are extremely unusual in typically showing 100% prevalence, or infection success, of all host fish (see paragraph 21, above). The effectiveness of the infective copepodid larva in locating wild host fish in the open Atlantic ocean is quite exceptional. It is therefore not unexpected that juvenile sea trout and salmon are extremely vulnerable to sea lice infestation whilst migrating through, or resident in, enclosed inshore waters.

26. Mean intensities for L. salmonis monitored on adult 1SW salmon in Scotland equalled or exceeded the threshold infestation of 20 lice · fish-1 that was deduced (paragraph 20, above) to cause sublethal stress to post-smolt Atlantic salmon. But the sampled wild 1SW salmon typically were 50-75 cm in length and 1.3-4.5 kg in weight. Although the scaling effects of intensity of L. salmonis on stress for host 230

SUMBAWS

Final Report

salmon ranging in size from post-smolts to adults cannot be stated with confidence, it is clear that postsmolts are especially vulnerable to caligid infestations and that larger adult fish can withstand markedly higher intensities. No data are available on the physiological impacts of L. salmonis on wild adult salmon but studies of pre-adult hatchery salmon show compromised swimming performance, and L. salmonis infestation can exert detrimental effects on subsequent reproduction of adult female Arctic char (Salvelinus alpinus L.).

27. The time-series monitoring data provide the best available for assessing progressive changes in sea lice infestations of adult wild salmon as, for example, salmon aquaculture production increases. But it should be emphasized that these wild fisheries can, by definition, sample only those hosts that survive to adulthood and migratory return. Fish that die as juveniles in inshore or offshore waters (from whatever cause) can never be sampled. Thus, even if coastal infestations of migrating smolts increase or decrease in intensity it is unlikely that this will be reflected by the mean intensities of parasites on return migrant adult salmon some 12-14 months after their first migration to sea.

28. From the monitoring studies, there was no specific indication that the surviving adult 1SW fish were detrimentally affected by sea lice (but see paragraph 26), although the typical epidermal damage and bleeding from the feeding activity of sea lice was commonly noted. Analyses of annual variations in condition factor of returning adult 1SW salmon from a single, identifiable river stock on the east coast of Scotland over the period 1993-2005 showed a marked pattern of recent and significant declines in growth success. Moreover, an identical pattern of variation in condition factor was also found for the (unknown) mixed British Isles stocks which were sampled on the north coast of Scotland. These declines in condition factor were not associated with sea lice infestations, but undoubtedly are of major consequence in being reflected throughout components of the ICES â&#x20AC;&#x153;southernâ&#x20AC;? European stock and imply radical shifts in the feeding quality and growth success of salmon at sea. Sea Surface Temperature (SST) has risen steadily throughout the North Atlantic over the period 1993-2005 and very significant negative correlations were found between annual variations in NE Atlantic SST anomalies and condition factor of whole year classes. These results have profound implications for management of wild salmon because, in contrast to the accepted paradigm of warmer SSTs being conducive to salmon survival and growth, they are strongly indicative of an intensely negative impact of recent ocean surface warming on salmon growth. Survivorship of salmon at sea is believed to be strongly growth-dependent; years of high growth generally correlate with high survivorship of that year class. The present observations show that average fecundity of 1SW salmon will have fallen by ~1522% over the period 1997-2005 and this alone will have important implications for the conservation of threatened stocks, especially in small river systems dominated by 1SW spawning adults.

29. It is recommended that time-series monitoring returning Atlantic salmon from commercial interceptory bag net fisheries, and single-stock in-river fisheries, should be continued. One problem here is that fishery managers and conservationists understandably and justifiably place a high priority on the closure of interceptory fisheries because they exploit multiple stocks. It must, however, be acknowledged that in the absence of these fisheries there will be no opportunity of obtaining comprehensive - and scientifically invaluable - information on those stocks. Were it not for our access to extensive and detailed time-series data for both single and multiple stocks in Scotland we would not have been able to infer the significant detrimental impacts of recent ocean warming on wild salmon growth. Monitoring of fish condition factor (and thereby estimates of mean levels of egg deposition) 231

SUMBAWS

Final Report

also will provide invaluable additional information to salmon fishery managers needing to determine minimum spawning escapements for individual rivers or stocks over a wide geographic range.

Knowledge of the numbers of returning salmon, alone, is not sufficient to calculate or estimate egg deposition and hence predict spawning escapements.

30. Because populations of North Atlantic L. salmonis have been shown by DNA studies to be genetically homogeneous and demographically open it will be impossible to eradicate this as a pest species on salmon farms. A very real concern, therefore, must lie in the paucity of the range of chemotherapeutants available to the aquaculture industry for sea lice control, and the commercial expense and lead time for developing new active compounds. Presently, there is a widespread and very heavy reliance by the industry on the one major compound, emamectin benzoate (SLICEÂŽ, ScheringPlough).

31. There are also consequences of farm-wild and wild-farm infestation interactions (which result in population genetic homogeneity) and the possible evolution of resistance of L. salmonis to chemical treatments. One of the tenets of pest management and control is that, in the absence of pesticides, resistant genotypes generally are selected against, or resistant individuals are less â&#x20AC;&#x2DC;fitâ&#x20AC;&#x2122;. Given the demographic openness of L. salmonis populations on farms, it is important to understand how resistance might develop on farms because resistance genes ought to be selected against on wild fish and outwith the farm environment. Nonetheless, in view of the possibility of resistance developing, the challenge to the aquaculture industry is to adopt co-ordinated husbandry practices (which themselves probably are economically beneficial) and to deploy the few currently available and licenced treatments at minimum effective dosages and perhaps to engage in rotation of treatment use. It is recommended that the enactment of precautionary strategies to protect wild salmonid stocks (by reducing the levels of ovigerous sea lice on farmed fish) must be balanced judgements, taking account of the desirability of minimizing sea lice levels whilst not compromising the efficacy of the few active licenced compounds available to the industry.

32. The past focus of research into copepod sea lice biology in the North Atlantic has necessarily been on the salmon louse, L. salmonis, but the importance of the host-generalist, C. elongatus, to both farmed and wild salmon and other teleost fish in the North Atlantic, should not be under-estimated. Atlantic cod (Gadus morhua L.) farming presently is expanding rapidly in the NE Atlantic. It is recommended that the wealth of experience and knowledge that has been garnered in relation to L. salmonis and salmonids should be readily transferable to the possibly detrimental impacts on wild gadid and salmonid stocks that C. elongatus will inevitably exert as cod farming develops. Furthermore, G. morhua also is host to a wide array of other (non-caligid) parasitic copepod species, and many of these are associated also with yet other, commercially and ecologically important gadid host species.

For example, Lernaeocera branchialis (Family Penellidae) is a serious economic pest species on wild cod, whiting and haddock, and Clavella adunca (Family Lernaeopodidae) also is a common parasite of these same fish. As cod farming develops it is highly likely that these, and many other species of parasites, may well cause problems for farmed and wild gadids alike. Neither C. adunca nor L. branchialis will impact salmonids, but the potential for interaction between farmed cod and wild gadid species is clear. It is recommended that pending and future national and EU legislation should take account of possible future developments in aquaculture, including new cultured fish species, and the full potential range of their metazoan parasites.

232

SUMBAWS

Final Report

Paper 1 POLYANDRY IN THE ECTOPARASITIC COPEPOD LEPEOPHTHEIRUS SALMONIS DESPITE COMPLEX PRECOPULATORY AND POSTCOPULATORY MATE-GUARDING C. D. Todd, R. J. Stevenson, H. Reinardy and M. G. Ritchie (published in Marine Ecology Progress Series (2005) 303, 225-234) Abstract Lepeophtheirus salmonis (Krøyer) is an economically important pest on cultured salmonids in the North Atlantic, and has been implicated in declines of some wild salmonid populations. Males inseminate newlymoulted adult females following the cementing of a pair of spermatophores to the female’s genital complex.

Females can produce multiple pairs of eggstrings over a period of months, but this species is reported to be monogamous as a result of blockage of the female’s copulatory ducts by the spermatophore tubules. On wild and farmed Atlantic salmon Salmo salar L., respectively, 88 and 78% adult females sea lice bore the typical pair of spermatophores, while 11 and 19% lacked spermatophores. A very few individuals (1% wild, 3% farmed) bore 3 or 4 spermatophores, showing that apparently successful multiple mating is possible.

Multiple paternity was confirmed by dual-locus microsatellite typing of offspring for 3 of 7 females carrying 4 spermatophores, but also for 2 of 3 females carrying a single pair of spermatophores. Probably most females on wild fish lose their initial spermatophores and are polygamous during their extended ovigerous lifetime, although effective blockage of the copulatory ducts by the first male almost certainly assures single paternity of the first few pairs of eggstrings. The total level of polyandry or sperm competition faced by males may be relatively low. The ecological implications of multiple paternity are discussed within the context of integrated pest management as a strategy of ameliorating L. salmonis infestations impacting both wild and farmed salmonids.

Introduction The specialist ectoparasitic copepod Lepeophtheirus salmonis (Krøyer) infests all species of wild salmonids in the North Pacific (Nagasawa 2001) and occurs with 100% prevalence on wild adult salmon Salmo salar L. in the North Atlantic (Jacobsen & Gaard 1997, Todd et al. 2000). This ectoparasite species, and to a lesser extent the generalist Caligus elongatus Nordmann, remains a major problem to the Atlantic salmon aquaculture industry. At sufficient intensities L. salmonis can debilitate or kill the host fish, and the circumstantial evidence that sea lice are at least a major contributory factor to the recent crashes of wild sea trout Salmo trutta L. stocks in parts of Ireland, Scotland and Norway is now considerable (e.g. Butler 2002, Todd et al. 2004). There is good evidence that L. salmonis may be detrimental to wild Atlantic salmon smolts migrating through certain Norwegian fjords (Finstad et al. 2000). Within the broader framework of integrated pest management (IPM), control of sea lice in Scottish aquaculture has focused largely on chemotherapeutants (e.g. organophosphates, pyrethroids, avermectins, hydrogen peroxide) in concert with periodic fallowing of farm sites, single year-class culture within sea lochs and co-ordinated sea lice treatments amongst some farms in voluntary area management agreements (Grant 2002, Lindsay & Rae 2003).

Despite these measures, the most recent global estimate of the total economic cost of sea lice to the aquaculture industry exceeds 100 million US $ per annum (Johnson et al. 2004), and for the Scottish industry alone is at least £25 million per annum (Rae 2002). Effective pest-species control and management demands detailed knowledge of the parasites’s life history and population biology, and Lepeophtheirus salmonis has been intensely studied (e.g. Boxshall & Defaye 1993, Pike & Wadsworth 1999, Tully & Nolan 2002), especially in the farm environment (e.g. Revie et al. 2002). The typical pattern for L. salmonis populations infesting wild Atlantic salmon is one of adult (~90%) and female (~70%) predominance (Jacobsen & Gaard 1997, Todd et al. 2000), because of an extended life span of ovigerous females. Adult female caligids store sperm in a seminal receptacle (Huys & Boxshall 1991), have internal fertilisation, and produce multiple pairs of uniseriate eggstrings over their reproductive life. The female retains each eggstring pair and the hatching stage is a free-swimming nauplius larva. After 2 larval moults during an obligatory planktonic phase of 2 to 9 d (Johnson & Albright 1991), the infective planktonic copepodid stage is attained and colonization of the host fish results in the first of 4 sessile chalimus stages. The fourth 233

SUMBAWS

Final Report

chalimus moults to the first of 2 mobile preadult stages, and both genders undergo a final moult to the mature adult. Adult male caligids actively locate and attach to Preadult II females for hours or days, and await the definitive adult female moult prior to copulation (Boxshall 1990, Ritchie et al. 1996a). Such precopulatory ‘proximate mate-guarding’ (e.g. Simmons & Siva-Jothy 1998) appears to be a defensive behaviour by the male. At copulation, the male cements a pair of spermatophores onto the posterior ventral surface of the female’s genital complex and remains clasped postcopula for some hours after spermatophore deposition (Ritchie et al. 1996a). Males take up to 4 d to synthesise spermatophores following copulation (Hull et al. 1998). Tubules extend from the attached spermatophores and cross the midline to enter the female copulatory pores on opposite sides (see Huys & Boxshall 1991, p. 274, for L. pectoralis). Emptying of the spermatophore occurs within hours and it is reported that the combination of inner secretions of the spermatophores, the spermatophore tubules themselves, and the male’s cement all effectively block the female copulatory ducts and prevent subsequent mating (Ritchie et al. 1996b). Simmons & Siva-Jothy (1998) term such behaviour ‘remote post-copulatory mate-guarding’. Ritchie et al. (1996a) deduced that Lepeophtheirus salmonis is monogamous, and Pike & Wadsworth (1999) concluded that although singlymated female L. salmonis could produce multiple pairs of fertile eggstrings, there was no information to date about female L. salmonis showing multiple mating. Given the importance of a detailed understanding of the mating system for any commercially or ecologically important pest species, and in view of the prevalence of multiple mating throughout the animal kingdom (e.g. Birkhead & Møller 1998, Neff & Pitcher 2002) and probable benefits of multiple mating to females (Arnqvist & Nilsson 2000, Tregenza & Wedell 2000), we sought to empirically assess whether or not L. salmonis is truly monogamous. Ritchie et al. (1996b) reported that non-virgin female L. salmonis (from farmed salmon) without spermatophores are very rarely found, and also that gravid females were never observed with an empty seminal receptacle. There is no observational evidence, but it is likely that female L. salmonis are incapable of removing the blocking empty spermatophores and cement, and will therefore presumably remain effectively monogamous for at least as long as these remain on the genital complex. The typical and maximum life spans of female Lepeophtheirus salmonis either on farmed or wild Atlantic salmon remain unknown, but almost certainly extend to several months (Pike & Wadsworth 1999, Tully & Nolan 2002), during which time perhaps up to 11 pairs of eggstrings may be produced (Heuch et al. 2000). Adult female L. pectoralis (parasitising plaice Pleuronectes platessa) can survive for up to 10 mo (Boxshall 1974). Anstensrud (1990) showed that spermatophores of L. pectoralis remained cemented to the female for a mean of 26 d (range 2 to 43 d), and that females can be copulated more than once, but only after the spermatophores have been lost. The deduction was that male L. pectoralis can only temporarily seal the female copulatory ducts as a result of spermatophore deposition. Female L. salmonis may differ markedly from L. pectoralis in perhaps retaining their spermatophores throughout reproductive life and remaining monogamous; but assuming that spermatophore retention/loss is similar in L. salmonis and L. pectoralis, monogamy of L. salmonis evidently still is possible if (1) the female copulatory ducts remain permanently blocked following first insemination, (2) all potential eggstrings can be fertilised by sperm stored in the seminal receptacle from the one mating, and (3) if old adult females show an increasing tendency to have lost the spermatophores. Because the last is unlikely to apply (Ritchie et al. 1996a, see last paragraph) this implies that multiple mating of L. salmonis may occur commonly, despite laboratory observations of a strong mating preference of males for virgin adult females (Hull et al. 1998). There has been a perception in the Scottish industry that aquaculture sites are liable to self-reinfestation by L. salmonis (e.g. Bron et al. 1993, Costello 1993, see also Heuch et al.

2003). If this were indeed a significant feature of the infection dynamics of L. salmonis on farms then confirmation of obligate female monogamy would have potentially important implications for the possible inclusion of interference of male fertility within a wider IPM strategy. Our objectives were 2-fold. First, to compare spermatophore occurrence on adult female Lepeophtheirus salmonis from both wild and farmed Atlantic salmon, and second, to ascertain paternity of embryo clutches for individual females from wild hosts. The first few pairs of eggstrings clearly will be fertilised by the initial male but, given the extended life span of adult female L. salmonis, and the observations of spermatophore loss in L. pectoralis, it was considered likely that L. salmonis females can lose their spermatophores and be multiply mated during their ovigerous lifetime. However, whether any such subsequent matings result in successful fertilisation by sperm of a second male (P2 fertilisation, e.g. Simmons & Siva-Jothy 1998) remains unclear. Because of their commonly high levels of polymorphism, single-locus DNA microsatellites have long been the method of choice in parentage analyses (e.g. Strassmann et al. 1996). Herein we applied dual-locus genotyping of samples of embryos from single eggstrings for 10 female L. salmonis, using specifically developed microsatellites (Todd et al. 2004).

234

SUMBAWS

Final Report

Materials and methods Spermatophore numbers and frequency Ethanol preserved adult female Lepeophtheirus salmonis from wild and farmed Atlantic salmon were examined for spermatophores. Bulk samples (not separated for each host fish) were obtained from 2 Scottish salmon farms. One farm sample (June 1998; Farm 3 of Todd et al. 2004) was from the mainland coastline of Wester Ross, and the other from the Western Isles (September 1997; Farm 5 of Todd et al. 2004). Complete collections of L. salmonis from individual wild Atlantic salmon were obtained from the commercial coastal nets at Strathy Point, north Scotland (58°06’N, 04°00’W) in July 1999, 2002, 2003 and 2004. Female Lepeophtheirus salmonis from wild salmon were distinguished as ‘young’ and ‘old’ adults, according to the degree of expansion of the genital complex (see Ritchie et al. 1996a,b), which proceeds in newly moulted adult females irrespective of copulation. The Preadult I and II female stages were distinguished for the Strathy Point 2002 to 2004 samples, but in 1999 only total preadult females were recorded. All adult females were scored as bearing 0 to 4 spermatophores. Females carrying the typical single pair of spermatophores (see Huys & Boxshall 1991, p. 279) are referred to here as ‘paired spermatophore’ individuals. Individuals bearing an additional 1 or 2 spermatophores generally had these cemented along the outer edge of the initial pair of spermatophores; these we refer to as ‘multiple spermatophore’ females.

Microsatellite typing of females and their families Multiple spermatophore females clearly had been multiply mated, and 7 such females were specifically chosen for microsatellite genotyping. Our primary requirements were to assess whether or not (1) specifically chosen females that had unequivocally mated multiply, and (2) randomly chosen females that carried the typical pair of spermatophores, were monogamous or producing polyandrous clutches of offspring. In 3 instances of multiply mated females (Females 4, 5, 6), we randomly sampled from the same respective fish another ovigerous, paired-spermatophore female (Females 1, 2, 3; Table 1). For microsatellite PCR, a variable number of offspring embryos (5 to 10) from both the proximal (attached) and distal (free) ends of 1 of the 2 available eggstrings were isolated for each of 10 females. Of the 7 multiple spermatophore females, 4 (Nos. 4, 5, 6, 9) had broken eggstrings: for these the individual embryos were isolated from both the proximal and the broken end, but were known to not have been maximally separated within the original eggstring. For Females 1–3 (2 spermatophores) and 7, 8 and 10 (multiple spermatophores), the sampled offspring were contiguous and equally balanced from the 2 ends of the intact eggstring (Table 1) to provide offspring genotypes for the first and last fertilised embryos within those eggstrings. PCR was undertaken for genomic DNA of individual embryos from other females in a preliminary study, with extraction by means of either a NucleoSpin (Abgene) kit or by the ‘single fly’ preparation (Gloor & Engels 1992). The latter methodology consistently resulted in the clearer PCR products and was subsequently adopted throughout. Female parent tissue for PCR genotyping was obtained by excising an appropriately small piece of the anterior cephalothorax and also by removal of a maxilliped.

These were extracted with the same protocol as the individual embryos and the single limb was found to consistently provide the clearer PCR products. PCR for all families included primers for 2 highly polymorphic Lepeophtheirus salmonis microsatellites (see Todd et al. 2004 for primer sequences and details of PCR amplifications). LsalSTA3 (GenBank Accession No. AY509256) and LsalSTA5 (AY509258), respectively, had shown 38 and 43 alleles in a previous analysis of population structuring of L. salmonis from throughout the North Atlantic. PCR products were visualised by PAGE silver-staining and alleles were sized against 10 bp ladders and against a previously genotyped individual as a positive control.

Statistical analyses Spermatophore frequencies for the separate years and for ‘wild’ and ‘farmed’ Lepeophtheirus salmonis were analysed for heterogeneity by the G-test. For families with apparently single paternity, statistical power was assessed by means of the Monte Carlo simulation model of Neff & Pitcher (2002). The probability of detecting multiple paternity within a clutch (Neff & Pitcher’s PrDM) is a function of the numbers of (1) loci used, (2) alleles in the population, (3) offspring genotyped, and (4) effective males. Moreover, males may have differential fertilisation success. Neff & Pitcher’s programme is dimension-limited to 30 alleles per locus. For present purposes we deleted alleles with the lowest frequencies from our original Atlantic frequency distributions for the 2 loci (Todd et al. 2004), and thereby reduced population diversity from 38 to 26 alleles (for LsalSTA3) and 43 to 27 alleles (for LsalSTA5), respectively. Population allele frequencies were recalculated for the 2 truncated distributions and included in the model. Each simulation included the 235

SUMBAWS

Final Report

known maternal genotype for the 2 loci and 2 (genetically unknown) males were presumed for each set of offspring analysed; the probability of detecting dual paternity was computed assuming both equal (0.5:0.5) and highly skewed (0.9:0.1) fertilisation success of 2 males. The tabulated probabilities are the averages for 10 runs of the model.

Table 1 Lepeophtheirus salmonis. The 10 females and single eggstrings used for paternity analysis. All were sampled from returning wild adult Atlantic salmon Salmo salar netted at Strathy Point, North Scotland. For host fish, year of sampling, fish identifier and host sea-age (1SW, 2SW = 1/2 sea-winter[s]) are given. For female parasites, condition of sampled eggstring, egg complement (intact eggstrings only; na = not available), number of spermatophores borne and numbers of offspring genotyped (positions within eggstring) are shown Female sea louse 1 2 3 4 5 6 7 8 9 10

Year/ Fish no./ Sea-age 2001/ 02/ 2SW 2001/ 39/ 1SW 2001/ 28/ 1SW 2001/ 02/ 2SW 2001/ 39/ 1SW 2001/ 28/ 1SW 1999/ 33/ 1SW 2001/ 31/ 1SW 2002/ 59/ 1SW 2002/ 63/ 1SW

Eggstring Intact Intact Intact Broken Broken Broken Intact Intact Broken Intact

Total egg no.

202 445 528 na na na 261 429 na 325

No. of spermatophores 2 2 2 4 4 4 4 4 4 4

No. offspring typed (position) 10 (5 proximal + 5 distal) 10 (5 proximal + 5 distal) 10 (5 proximal + 5 distal) 10 10 10 16 (8 proximal + 8 distal) 20 (10 proximal + 10 distal) 20 20 (10 proximal + 10 distal)

Results Spermatophore numbers and frequency Most adult female Lepeophtheirus salmonis from wild and farmed host fish bore a pair of spermatophores (Table 2), cemented in the typical posterioventral position on the genital complex. It is generally impossible to visually assess whether females carrying the typical paired spermatophores were (1) singly mated, or (2) had possibly lost their initial spermatophores plus cement—perhaps some weeks after copulation (see Anstensrud 1990 for L. pectoralis)—and been multiply copulated. Some individuals evidently had lost one of the pair, as indicated by traces of remaining male cement. Few females lacked spermatophores, but many of these also showed traces of cement, indicating that they had been inseminated at least once. A very low percentage of females from both wild and farmed salmon did, however, carry multiple (3 or 4) spermatophores, and almost certainly had been mated by at least 2 separate males. There was significant heterogeneity amongst the 4 yr for wild salmon (Table 2; G = 15.09, 6 df, p = 0.020): 2002 to 2004 did not differ significantly from one another, but 1999 differed significantly from all 3, due to fewer than expected individuals lacking spermatophores in 1999. There was no significant heterogeneity between the 2 farm samples (G = 3.04, 2 df, p = 0.219), although the sample sizes are small. Comparison of the pooled frequencies for parasites sampled from wild versus farmed hosts showed significant heterogeneity (G = 15.56, 2 df, p < 0.001), with farm lice showing more than expected females lacking spermatophores or carrying multiple spermatophores. Multiple spermatophore individuals almost invariably were old adult females (fully expanded genital complex), although 1 young adult female (small genital complex) was recorded in this category for 2002. Distinguishing the wild adult females into ‘young’ and ‘old’ (Table 3) showed that whilst approximately onethird of young adult females remained virgin, only 8% of old females were lacking spermatophores. The latter had not, however, all recently undergone the definitive moult from the Preadult II stage because, for example, 13% of the 87 old adult females lacking spermatophores in 2003 bore both cement and copulation marks from the male appendages on their genital complex.Together, these features strongly indicate that these ‘old’ females had been previously mated but had lost their spermatophores.

236

SUMBAWS

Final Report

Table 2 Lepeophtheirus salmonis. Adult females from wild and farmed Atlantic salmon Salmo salar, showing % (frequency) bearing 0 to 4 spermatophores Spermatophores 0 1 or 2 (‘paired’) 3 or 4 (‘multiple’)

1999 (n=42) 8.0 (60) 90.5 (677) 1.5 (11)

Strathy Point (wild fish) 2002 2003 2004 (n =53) (n =64) (n =62) 13.0 11.8 10.3 (81) (107) (112) 86.4 87.2 89.1 (538) (788) (971) 0.6 1.0 0.6 (4) (9) (6)

Pooled (n =221) 10.7 (360) 88.4 (2974) 0.9 (30)

Farm (fish) 5 3 Pooled (n =14) (n=11) (n =25) 22.4 12.0 19.3 (26) (6) (32) 74.1 86.0 77.7 (86) (43) (129) 3.4 2.0 3.0 (4) (1) (5)

Table 3 Lepeophtheirus salmonis. Adult females from wild Atlantic salmon Salmo salar netted at Straithy Point over 4 yr period, showing % (frequency) of ‘young’ and ‘old’ lacking (without) or bearing (with) spermatophores Year 199 2002 2003 2004 Pooled

Percent young (n) without with 25.6 (11) 74.4 (32) 38.9 (35) 61.1 (55) 30.3 (20) 69.7 (46) 24.3 (43) 75.7 (134) 29.0 (109) 71.0 (267)

Percent old (n) without with 7.0 (49) 93.0 (656) 8.6 (46) 91.4 (487) 10.4 (87) 89.6 (751) 7.6 (69) 92.4 (843) 8.4 (251) 91.6 (2737)

Microsatellite typing and analysis of paternity For LsalSTA3 and LsalSTA5, respectively, 15 and 18 different alleles were detected amongst the 10 females and their families (Table 4): 3 other alleles at LsalSTA5 were recorded as maternal. Multiple paternity was confirmed among the families for 2 of the 3 females bearing the typical pair of spermatophores.

The test of paternity for the third family in this category (Female 3) was, however, weak due to the lack of data for LsalSTA5 (failed PCR). The single inferred male for Family 8 was homozygous at LsalSTA3 and shared an allele at that locus with Female 8. For Females 1 and 2 (Table 4) single paternity was implied for 1 of the 2 loci, and yet dual paternity was confirmed by the alternative locus, showing that the males again shared alleles at 1 locus despite the potentially high allele numbers. Neff & Pitcher’s PrDM for Female 3 was high (>90%), but only when assuming equal mating success of 2 putative fathers. For the strongly skewed simulation, the probability of detecting multiple paternity for these 10 offspring fell to 52%. For Females 8 and 10, confidence in single paternity was consistently high for both simulations because of the larger families screened. The relative frequencies of the male alleles (both loci) for each family of offspring are shown in Fig. 1. In all 5 cases of multiple paternity—2 of 3 paired spermatophore females (Nos. 1, 2) and 3 of 7 multiple spermatophore (Nos. 4, 7, 9) individuals—all offspring could be genetically explained by just males. Although the family samples are small it is important to note that, for the multiply mated females, the least successful paternal allele always had a very low frequency (~0.08 for both loci; Fig. 1). If these females had mated only twice (no families showed >4 paternal alleles), and males had equal fertilisation success, this frequency would be expected to approximate 0.25. Hence, male fertilization success probably is not equal in this species, but there are numerous processes (variation in sperm depletion or quality, insemination rates, sperm competition, cryptic female choice) which could be responsible for this pattern.

237

SUMBAWS

Final Report

Table 4 Lepeophtheirus salmonis. Family samples for 10 females genotyped at each locus and for both loci.

PrDM (Neff & Pitcher 2002) calculated for families for which single paternity was apparent: values are average probabilities (10 model simulations) of detecting dual paternity, assuming males with either equal (0.5:0.5) or strongly biased (0.9:0.1) mating success. nd: no data (failed PCR). No. typed (no. n-m alleles): no of offspring typed (no. of non-maternal alleles detected); PrDM: probability of detecting multiple paternity (modelled male paternity ratio) No. typed (no. n-m alleles)

Female no.

LsalSA3

LsalSA5

Both loci

1 2 3 4 5 6 7 8 9 10 Total

6 (2) 8 (3) 10 (2) 8 (3) 9 (2) 10 (2) 8 (3) 18 (2) 16 (4) 20 (2) 113 (15)

10 (3) 9 (2) nd 9 (4) 9 (2) 10 (2) 15 (3) 18 (2) 16 (3) 20 (2) 116 (18)

6 8 nd 8 9 10 8 18 16 20

No. of spermatophores 2 2 2 4 4 4 4 4 4 4

PrDM

Inferred paternity

(0.5:0.5)

(0.9:0.1)

Dual Dual Single Dual Single Single Dual Single Dual Single

0.906 0.987 0.991 0.999 0.999

0.524 0.589 0.628 0.832 0.862

Figure 1 Lepeophtheirus salmonis. Proportional frequencies of male alleles for LsalSTA3 (left-hand bar) and LsalSTA5 (right-hand bar) for offspring families of Females 1 to 10. Shading shows number and proportions of male alleles for each family; similar shading does not imply identical alleles across families; 1 or 2 non-maternal alleles = single paternity, S (Females 3, 5, 6, 8, 10); 3 or 4 non-maternal alleles = dual paternity, D (Females 1, 2, 4, 7, 9). Females 1 to 3 bore single pair of spermatophores; Females 4 to 10 bore 4 spermatophores. Data for LsalSTA5 for Family 3 missing due to failed PCR

238

SUMBAWS

Final Report

Discussion Multiple mating and paternity If the cemented spermatophores of Lepeophtheirus salmonis are successful in preventing subsequent male insemination, the simplest explanation of the paternity data is as follows: (1) Of the 7 multiply mated females, 4 (Females 5, 6, 8, 10) still bore their original spermatophores and a subsequent pair. Their copulatory pores remained blocked by the original pair, and despite having been multiply mated they were effectively still monogamous. (2) For the remaining 3 multiply mated females (Females 4, 7, 9), the spermatophores borne were the second and third pairs, and dual paternity of offspring was attributable to the first (lost) and second pairs of spermatophores. The second pair did, however, still block the copulatory pores from the most recently acquired third pair. (3) Of those bearing the typical single pair of spermatophores, Female 3 had been mated only once, whereas Females 1 and 2 bore a replacement second pair, having lost the initial pair of spermatophores and were producing offspring which had been successfully fertilised by the second male (P2 fertilisation). The higher percentage of adult females from farmed hosts carrying no spermatophores (19%; Table 2) was unexpected, given observations of the often very marked predominance of males on farmed salmon (e.g. Bron et al. 1993) in contrast to female predominance on wild salmon (e.g. Todd et al. 2000), and the presumed reduced life span of parasites on farmed fishes. Despite 29% of young adult females on wild salmon being unmated (Table 3), it is very probable that within the 8 d (at 12°C; Ritchie et al. 1996a) required for full expansion of the genital complex and the transformation of an adult female from ‘young’ to ‘old’, all females on wild fish will become inseminated. The likelihood is, therefore, that old adult females lacking spermatophores are seldom, if ever, virgin, but will be previously inseminated individuals that have lost both the spermatophores and their cement, and may well already be polyandrous. Numerical predominance of females over males appears widespread in populations of free-living calanoid copepods (e.g. Cuoc et al. 1997, Bathélémy et al. 1998).

Because the mobile preadult and adult stages of Lepeophtheirus salmonis almost certainly do not transfer between wild host fish (but see Ritchie 1997 for hosts in farm pens), the individual fish is the effective demographic ‘unit’ for these ectoparasites. It clearly is possible that on a given fish there may be insufficient adult males to pair with all the reproductively ‘available’ female complement (Preadult II females, plus unfertilized young adult females and ‘old’ adult females lacking spermatophores): this could therefore lead to incidences of old virgin adult females. Every wild fish sampled at Strathy Point in 2002 to 2004 bore 1 or more adult male sea lice (range 1 to 25), and always adult males on a given fish numerically exceeded Preadult II females (data not shown). Similarly, the number of adult males almost always exceeded the complement of females reproductively available to those males on a given fish (Fig. 2). Predictions of male:male competition, and of female-mating frequency, in both freeliving and parasitic copepods may, therefore, be erroneous if such expectations are based on population proportions of total males and females.

For L. salmonis specifically, the general surfeit of adult males over available females on individual fish, and the consistent occurrence of a large proportion of unfertilised young adult females (Table 3), shows that successful fertilization of adult females does not always occur as soon as they undergo the final moult from the Preadult II stage. The rare occurrence of multiple spermatophores (Table 2) does, however, confirm that adult females already bearing spermatophores may very occasionally be multiply mated. Taken together, these data, and the low frequency of old adult females lacking spermatophores (Tables 2 & 3), all strongly indicate that multiple mating is a common feature in L. salmonis, despite clear laboratory behavioural preferences by adult males for the younger female stages (Hull et al. 1998).

Further data for both wild and farmed host fish are required, but if the observed two-thirds of paired spermatophore females (Nos. 1 and 2) showing dual paternity is representative; it is likely that most females on wild and farmed salmon will be polyandrous. Probably only those females that fail to survive beyond loss of the initial spermatophores and their cement will be effectively monogamous. In the farm environment, therefore, routine chemotherapy and its consequential restriction of the adult life span might result in effective monogamy of females. Spermatophores on a female are visible evidence of the outcome of precopulatory competition amongst various males on a host fish; but the finding of multiple paternity of offspring does not, in itself, necessarily reflect postcopulatory male competition, sperm competition, or even cryptic female choice (e.g. Eberhard 1998). Only detailed manipulative studies of the copulation and fertilization history of individual females, and of changes in individual male availability through their ovigerous lifetimes, will reveal the nature and extent of such interactions. It will, for example, be of especial importance to ascertain the extent of initial sperm depletion prior to loss of the first pair of spermatophores in relation to possible sperm displacement and competition, male precedence, and the mean and variance of second male paternity (P2 fertilisation, e.g. Simmons & Siva-Jothy 1998, Simmons 2001). The temporal and 239

SUMBAWS

Final Report

energetic investment by male caligids in proximate mate-guarding of the Preadult II female, and the subsequently inseminated young adult female, is relatively small—a matter of hours, or perhaps days (Boxshall 1990, Ritchie et al. 1996a). Conversely, the durability of the attached spermatophores and cement provides the individual male with extended, cheap and effective remote postcopulatory mate-guarding (cf.

Hull et al. 1998), perhaps for some weeks, during which the (initially monogamous) female may produce several pairs of eggstrings. Insect mating plugs do generally represent effective shortterm barriers to reinsemination (Simmons 2001), but for the majority of insects (which have a single female genital opening), long-term barriers might also prevent the female from ovipositing and thereby be detrimental to both male and female reproductive success. In sharp contrast to most insects, however, female caligid copepods have separate paired copulatory ducts and gonopores through which the fertilised eggstrings are extruded.

Effective first-male blockage of the female does not, therefore, compromise either first male or female fitness in Lepeophtheirus salmonis. Nonetheless, loss of the blockage does allow a degree of polygamy.

Figure 2 Lepeophtheirus salmonis on wild Atlantic salmon Salmo salar for period 2002 to 2004 inclusive (n = 179 fish; Table 2). Scatterplot of ‘available’ females (= Preadult II females + ‘young’ adult females [0 spermatophores] + ‘old’ adult females [0 spermatophores]) on adult males per host fish. Fitted line indicates equality of numbers of males and ‘available’ females. Abscissa values for Years 2002 and 2004 offset for clarity Monogamy and pest control IPM approaches to pest control are deliberately multifaceted and, amongst other benefits, aim to reduce reliance on insecticides and the evolution of treatment resistance by the target organism. The sterile insect technique (SIT) typically involves the rearing and sterilization of males by irradiation before their subsequent release to the environment (e.g. Curtis 1985); SIT has proved to be a successful means of controlling certain terrestrial pest species (e.g. Ferguson et al. 2005), especially when applied within a broader IPM strategy (e.g. Dargie 2000, Twohey et al. 2003). An effective contribution to pest control by sterile male release (SMR) in the aquatic environment is, however, restricted to attempts to control sea lampreys Petromyzon marinus parasitising lake trout and lake whitefish in the Laurentian Great Lakes (e.g.

Twohey et al. 2003). Migrating male lampreys are trapped, chemically sterilised by individual injection and then released. An attractive element of the SMR approach is its potentially minimal wider environmental effect (Siefkes et al. 2003). Of the licensed sea louse chemotherapeutants available to the Scottish industry, only 2 (cypermethrin, emamectin—both of which nonspecifically interfere with the neural membrane function) are widely used. Teflubenzuron, a chitin synthesis inhibitor, which is accordingly ineffective against the (non-moulting) adult stage, is in use only intermittently (Grant 2002). It is therefore understandable that the industry has concerns over possible self-reinfestation of farm sites and development of resistance to the few compounds available to them. Whilst the direct experimental evidence of resistance of sea lice to chemical treatments (e.g. dichlorvos, Jones et al. 1992; deltamethrin, Sevatdal et al. 2005) is equivocal, there is a perception within the industry that this can be a problem. An IPM strategy that ameliorates larval sea lice production, and that minimises dependence on agents to which resistance may evolve, could be advantageous to the aquaculture industry. Monogamy of a pest species renders it especially vulnerable to SMR, but it can have an adverse consequence in genetic control if already-mated females 240

SUMBAWS

Final Report

immigrate from without the treatment area (Curtis 1985), or if females begin to avoid sterile males (Ferguson et al. 2005). Unlike many insect pests, which may be highly vagile, adult Lepeophtheirus salmonis are effectively sedentary (and thereby manageable and readily treatable) on their host fish in captivity, because it is only the planktonic larval stages that emigrate and immigrate. For polyandrous caligid copepods, SMR, or deployment of male sterilant treatments, might yet offer a plausible addition to current sea lice IPM by the aquaculture industry if extremely high sterile:wild ratios of males could be attained on fish within the treatment area. It will never be feasible to rear (on live fish) and release the required numbers of manipulated males for SMR approaches; and, because of their poor abilities to reattach to host fish (Pike & Wadsworth 1999), released adult males almost certainly will not establish themselves on farms in suitable numbers. Because populations of L. salmonis on salmon farms are demographically open (Todd et al. 2004), it will never be possible to eradicate this pest at aquaculture sites. However, it may be possible to develop a specific, biologically or photochemically degradable male sterilant. The wider environmental impact of such an agent could perhaps be reduced by its targeted application as an in-feed additive, but application would have to be regular in order to continually sterilise males as they develop on the fish from the immigrating infective copepodid larval stages. Our previous study of Atlantic and Pacific Lepeophtheirus salmonis (Todd et al. 2004) revealed that levels of gene flow and cross-infection amongst the different host species and between wild and farmed salmonids are sufficiently high throughout the North Atlantic to prevent significant population genetic differentiation on an ocean-wide scale. Planktonic larvae of L. salmonis exported by currents from a given farm site may well infect both adjacent farms and local populations of wild salmonids, in addition to perhaps ultimately being re-imported and re-infesting the natal farm site. High levels of gene flow (= larval colonisation) between wild and farmed salmonids, together with the considerable numerical imbalance between wild and cultured salmonids in the coastal waters of Scotland (e.g. Butler 2002), lead to the conclusion that the bulk of the larval flux of L. salmonis is probably from farmed to wild hosts. Notwithstanding the logistic and practical difficulties in achieving SMR in the aquatic environment, the incorporation of male sterilisation of L. salmonis within a commercially beneficial and environmentally benign IPM strategy—including co-ordinated area management agreements amongst farm sites—may therefore have significant mutual benefits for the welfare of both wild and farmed salmonid stocks. The effectiveness of the contribution to control by sterilising male parasites as they develop on the host fish would also be enhanced if chemotherapeutants could be used to ensure that ovigerous females generally did not survive beyond loss of their initial spermatophores. From an ecological perspective, only experimental manipulation and long-term maintenance of individual copepods on isolated host fish will allow a full understanding of the possible roles of sperm depletion, competition, displacement or stratification, and patterns of male sperm precedence or cryptic female choice in this economically and environmentally important pest species. The key finding—that polyandry occurs and probably is widespread amongst female Lepeophtheirus salmonis—is important to the modelling of the dynamics and the management of caligid infestations on cultured salmonids (e.g. Revie et al. 2002), and requires quantification in the farm environment. Polyandrous mating of adult females also has implications for the evolution and possible spread of resistance to chemotherapeutant treatments on an ocean-wide scale.

Furthermore, the fact that polyandry is frequent despite post-copulatory mate-guarding in this species should urge caution regarding inferences of single or multiple matings (e.g. Cuoc et al. 1997, Bathélémy et al.

1998) in other free-living copepods based on anatomical observations alone.

Acknowledgements We thank A. E. Magurran, G. A. Boxshall and A. M. Walker for comments on the manuscript and J. A. Graves for molecular advice and assistance. G. B. Shimmield and the Scottish Association for Marine Science, Oban, kindly provided a SAMS Bursary to C.D.T. and R.J.S. We are grateful to S. Paterson (T. R. Paterson & Sons, Salmon Fishing Station, Strathy Point) for access to fish from his netting operation.

References Anstensrud M (1990) Mating strategies of two parasitic copepods [(Lernaeocera branchialis (L.) (Penellidae) and Lepeophtheirus pectoralis (Müller) (Caligidae)] on flounder: polygamy, sex-specific age at maturity and sex ratio. J Exp Mar Biol Ecol 136:141–158 Arnqvist G, Nilsson T (2000) The evolution of polyandry: multiple mating and female fitness in insects.

Anim Behav 60:145–164 241

SUMBAWS

Final Report

Bathélémy RM, Cuoc C, Defaye D, Brunet M, Mazza J (1998) Female genital structures in several families of Centropagoidea (Copepoda: Calanoida). Phil Trans R Soc Lond B 353:721–736 Birkhead TR, Møller AP (1998) Sperm competition, sexual selection and different routes to fitness. In: Birkhead TR, Møller AP (eds) Sperm competition and sexual selection. Academic Press, London, p 757– 781 Boxshall GA (1974) The population dynamics of Lepeophtheirus pectoralis (Müller): seasonal variations in abundance and age structure. Parasitology 69:361–371 Boxshall GA (1990) Precopulatory mate guarding in copepods. Bijdr Dierkd 60:209–213 Boxshall GA, Defaye D (eds) (1993) Pathogens of wild and farmed fish: Sea lice. Ellis Horwood, Chichester Bron JE, Sommerville C, Wootten R, Rae GH (1993) Influence of treatment with dichlorvos on the epidemiology of Lepeophtheirus salmonis (Krøyer, 1837) and Caligus elongates Nordmann, 1832 on Scottish salmon farms. In: Boxshall GA, Defaye D (eds) Pathogens of wild and farmed fish: Sea lice. Ellis Horwood, Chichester, p 263–274 Butler JRA (2002) Wild salmonids and sea louse infestations on the west coast of Scotland: sources of infection and implications for the management of marine salmon farms. Pest Manag Sci 58:595–608 Costello MJ (1993) Review of methods to control sea lice (Caligidae: Crustacea) infestations on salmon (Salmo salar) farms. In: Boxshall GA, Defaye D (eds) Pathogens of wild and farmed fish. Sea lice. Ellis Horwood, Chichester, p 219–252 Cuoc C, Defaye D, Brunet M, Notonier R, Mazza J (1997) Female genital structures of Metridinidae (Copepoda: Calanoida). Mar Biol 129:651–665 Curtis CF (1985) Genetic control of insect pests: growth industry or lead balloon? Biol J Linn Soc 26:359– 374 Dargie J (2000) Nuclear technologies and food security: fields of progress. IAEA Bull 42:23–32 Eberhard WG (1998) Female roles in sperm competition. In: Birkhead TR, Møller AP (eds) Sperm competition and sexual selection. Academic Press, London, p 91–116 Ferguson HM, John B, Ng’habi K, Knols BGJ (2005) Redressing the sex imbalance in knowledge of vector biology. Trends Ecol Evol 20:202–209 Finstad B, Bjørn PA, Grimnes A, Hvidsten NA (2000) Laboratory and field investigations of salmon lice [Lepeophtheirus salmonis (Krøyer)] infestation on Atlantic salmon (Salmo salar L.) post-smolts. Aquac Res 31:795–803 Gloor G, Engels W (1992) Single-fly DNA preparations for PCR. Drosophila Inf Serv 71:148–149 Grant AN (2002) Medicines for sea lice. Pest Manag Sci 58: 521–527 Heuch PA, Nordhagen JR, Schram TA (2000) Egg production in the salmon louse [Lepeophtheirus salmonis (Krøyer)] in relation to origin and water temperature. Aquac Res 31: 805–814 Heuch PA, Revie CW, Gettinby G (2003) A comparison of epidemiological patterns of salmon lice, Lepeophtheirus salmonis, infections on farmed Atlantic salmon, Salmo salar L., in Norway and Scotland. J Fish Dis 26:539–551 Hull MQ, Pike AW, Mordue AJ, Rae GH (1998) Patterns of pair formation and mating in an ectoparasitic caligid copepod Lepeophtheirus salmonis (Krøyer 1837): implications for its sensory and mating biology.

Phil Trans R Soc Lond B 353:753–764 Huys R, Boxshall GA (1991) Copepod evolution. Ray Society, London Jacobsen JA, Gaard E (1997) Open-ocean infestation by salmon lice (Lepeophtheirus salmonis): comparison of wild and escaped farmed Atlantic salmon (Salmo salar L.). ICES J Mar Res 54:1113–1119 Johnson SC, Albright LJ (1991) The developmental stages of Lepeophtheirus salmonis (Krøyer, 1837) (Copepoda, Caligidae). Can J Zool 69:929–950 Johnson SC, Treasurer JW, Bravo S, Nagasawa K, Kabata Z (2004) A review of the impact of parasitic copepods on marine aquaculture. Zool Stud 43:229–243 242

SUMBAWS

Final Report

Jones MW, Sommerville C, Wootten R (1992) Reduced sensitivity of the salmon louse, Lepeophtheirus salmonis, to the organophosphate dichlorvos. J Fish Dis 15:197–202 Lindsay Lord, Rae G (2003) Delivering the solutions—the salmon farmer’s point of view. In: Mills D (ed) Salmon at the edge. Blackwell Science, Oxford, p 159–171 Nagasawa K (2001) Annual changes in the population size of the salmon louse Lepeophtheirus salmonis (Copepoda: Caligidae) on high-seas Pacific salmon (Oncorhynchus spp.), and relationship to host abundance. Hydrobiologia 453/454:411–416 Neff BD, Pitcher TE (2002) Assessing the statistical power of genetic analyses to detect multiple mating in fishes. J Fish Biol 61:739–750 Pike AW, Wadsworth SL (1999) Sealice on salmonids: their biology and control. Adv Parasitol 44:233– 337 Rae GH (2002) Sea louse control in Scotland, past and present. Pest Manag Sci 58:515–520 Revie CW, Gettinby G, Treasurer JW, Rae GH, Clark N (2002) Temporal, environmental and management factors influencing the epidemiological patterns of sea lice (Lepeophtheirus salmonis) infestations on farmed Atlantic salmon (Salmo salar) in Scotland. Pest Manag Sci 58:576–584 Ritchie G (1997) The host transfer ability of Lepeophtheirus salmonis (Copepoda: Caligidae) from farmed Atlantic salmon, Salmo salar L. J Fish Dis 20:153–157 Ritchie G, Mordue AJ, Pike AW, Rae GH (1996a) Observations on mating and reproductive behaviour of Lepeophtheirus salmonis, Krøyer (Copepoda: Caligidae). J Exp Mar Biol Ecol 201:285–298 Ritchie G, Mordue AJ, Pike AW, Rae GH (1996b) Morphology and ultrastructure of the reproductive system of Lepeophtheirus salmonis (Krøyer, 1837) (Copepoda: Caligidae). J Crustac Biol 16:330–346 Sevatdal S, Copley L, Wallace C, Jackson D, Horsberg TE (2005) Monitoring of the sensitivity of sea lice (Lepeophtheirus salmonis) to pyrethroids in Norway, Ireland and Scotland using bioassays and probit modelling. Aquaculture 244:19–27 Siefkes MJ, Bergstedt RA, Twohey MB, Li W (2003) Chemosterilization of male sea lampreys (Petromyzon marinus) does not affect sex pheromone release. Can J Fish Aquat Sci 60:23–31 Simmons LW (2001) Sperm competition and its evolutionary consequences in the insects. Princeton University Press, Princeton, NJ Simmons LW, Siva-Jothy MT (1998) Sperm competition in insects: mechanisms and the potential for selection. In: Birkhead TR, Møller AP (eds) Sperm competition and sexual selection. Academic Press, London, p 341–434 Strassmann JE, Solís CR, Peters JM, Queller DC (1996) Strategies for finding and using highly polymorphic DNA microsatellite loci for studies of genetic relatedness and pedigrees. In: Ferraris JD, Palumbi SR (eds) Molecular zoology. Advances, strategies, and protocols. Wiley-Liss, New York, p 163– 180 Todd CD, Walker AM, Hoyle JE, Northcott SJ, Walker AF, Ritchie MG (2000) Infestations of wild adult Atlantic salmon (Salmo salar L.) by the ectoparasitic copepod sea louse Lepeophtheirus salmonis Krøyer: prevalence, intensity and the spatial distribution of males and females on the host fish.

Hydrobiologia 429:181–196 Todd CD, Walker AM, Ritchie MG, Graves JA, Walker AF (2004) Population genetic differentiation of sea lice (Lepeophtheirus salmonis Krøyer) parasitic on Atlantic and Pacific salmonids: analyses of microsatellite DNA variation amongst wild and farmed hosts. Can J Fish Aquat Sci 61:1176–1190 Tregenza T, Wedell N (2000) Genetic compatibility, mate choice and patterns of parentage: invited review.

Mol Ecol 9:1013–1027 Tully O, Nolan DT (2002) A review of the population biology and host–parasite interactions of the sea louse Lepeophtheirus salmonis (Copepoda: Caligidae). Parasitology 124: 165–182 Twohey MB, Hanson LH, Heinrich JW, Seelye JG, Bergstedt RA, McDonald RB, Christie GC (2003) History and development of the sterile male release technique in sea lamprey management. J Gt Lakes Res 29:410–423 243

SUMBAWS

Final Report

Paper 2: ECTOPARASITIC SEA LICE (LEPEOPHTHEIRUS SALMONIS KRØYER, CALIGUS ELONGATUS NORDMANN) INFESTATIONS OF WILD ADULT ONE SEA-WINTER ATLANTIC SALMON SALMO SALAR L. RETURNING TO SCOTLAND, 1998-2005 C.D. Todd, B.D.M. Whyte, J.C. MacLean, A.M. Walker (Submitted to Marine Ecology Progress Series) Abstract Caligid ectoparasitic copepods are major pathological pests on cultured Atlantic salmon (Salmo salar L.) and their population biology has been well studied in the farm environment. The ecology of caligid infestations of wild salmon is, by contrast, rather poorly understood. We monitored return migrant one seawinter wild Atlantic salmon in Scotland annually between 1998 and 2005 for infestations of 2 caligids, Lepeophtheirus salmonis Krøyer (8 y) and Caligus elongatus Nordmann (7 y). L. salmonis was 100% prevalent in all years, whereas C. elongatus ranged from 90–100%. Abundances fluctuated markedly between years and L. salmonis antilogged mean abundance (min.-max. 17.4–31.0) was significantly greater than for C. elongatus (min.-max. 2.9–23.8) in all except one year. A positive association in abundance of the 2 species, for individual fish within any one year, indicates weak or absent competitive effects on abundances for individual hosts. Individual fish within any one year appeared similarly vulnerable to infestation by either species, though Taylor’s power regression showed clear differences in density-related patterns of overdispersion amongst hosts for the 2 species. Host condition factor (expressed either as K or WR) showed significant variation among years. Parasite species abundances were not, however, determined by host condition factor, and poor condition fish were no more likely to carry high infestations than were high condition fish.

Introduction Throughout the North Atlantic and North Pacific, the specialist copepod Lepeophtheirus salmonis Krøyer is a natural ectoparasite of salmonid fishes in their marine phase (Pike & Wadsworth 1999, Nagasawa 2001). L. salmonis, and the more host-generalist, Caligus elongatus Nordmann – which is recorded from >80 species of marine fish (Kabata 1979) – are the most economically important pathogens to the Atlantic salmon (Salmo salar L.) aquaculture industry (Johnson et al. 2004). Expansion of Atlantic salmon farming has been rapid since its inception ~35 years ago, and global annual production presently exceeds 1.1 million t (ICES 2004). As a consequence, much research effort has focused on the life cycle of L. salmonis on salmon farms and on methods of therapeutic control (Pike & Wadsworth 1999, Tully & Nolan 2002).

Although considerable progress has been made in modeling the dynamics of L. salmonis populations on farms (e.g. Revie et al. 2002), comparatively little is known of the ecological interactions/implications with its wild host associations (e.g. Jacobsen & Gaard 1997, Todd et al. 2000, 2004, Bjørn & Finstad 2002). Even less is known of C. elongatus, which in the NE Atlantic has been variously recorded as scarce to abundant on wild sea trout (Salmo trutta L.) (e.g. Bristow & Berland 1991, Berland 1993, MacKenzie et al. 1998, Schram et al. 1998), Atlantic salmon (Jacobsen & Gaard 1997) and Arctic char (Salvelinus alpinus (L.), Bjørn et al. 2001).

Despite their differences in adult size, both species have similarly short generation times of 50–90 d (Tully 1989) and can cause pathological damage and host death, with the small post-smolt stage being the most vulnerable (e.g. Pike & Wadsworth 1999, Finstad et al. 2000). The quantitative extent to which salmon farms contribute to wild fish infestations is a matter of some debate, but the evidence is now considerable that farm-origin caligids are at least a contributory factor to recent stock declines of wild salmonids proximate to salmon farms in areas of Norway, the British Isles, and W Canada (Bjørn & Finstad 2002, Butler 2002, McKibben & Hay 2004, Penston et al. 2004, Todd et al. 2004, Krkošek et al. 2005).

In the present study, we monitored L. salmonis (1998-2004) and C. elongatus (1999-2004) infestations of return migrant one sea-winter (1SW) Atlantic salmon in Scotland against the backdrop of declining marine survival and abundances of wild salmon over at least three decades (Anon. 2004). Our primary objectives were to quantify variations in the annual abundance and the age-structure of the two parasite species on returning 1SW adults, and to ascertain whether year-to-year variations in infestation might correlate with host fish parameters, including individual condition factor (CF). Returning 1SW salmon re-enter freshwater 244

SUMBAWS

Final Report

in Scotland between May and November following their one winter at sea, and they tend to increase in length and weight as the summer netting/sampling season progresses (Todd et al. 2006). Host CF can be calculated for individual fish using a range of standard protocols based on mass-length relationships (see Blackwell et al. 2000, Brenden et al. 2003, Marshall et al. 2004, Neff & Cargnelli 2004). Analysis of within and between-year variation in CF for 13 year classes of mixed-stock 1SW salmon, netted on multiple dates during June-August each year at the same location as the present study (1993-2005, n = 4916; Todd et al.

2006), showed the unexpected result that in 12 y CF did not vary significantly within a year. In that study, CF was expressed as the predicted weight of a standard fish of 60 cm length on each sampling date, and for only 1 y was there a significant decrease in CF during the season. CF therefore essentially is ‘set’ for entire year classes of 1SW return migrants and showed significant variation between years. Given that low intensity caligid ectoparasite infestations can result in physiological stress, or even death, of the host fish (Wells et al. 2006), the clear possibility was that between-year variations in CF of wild hosts might be determined by, or correlated with, ectoparasite abundances.

Materials & Methods Field site and sample data All data were for wild salmon trapped (but still free-swimming) in bag nets set within ~100 m of the shoreline at Strathy Point, N Scotland (58º06'N 04º00'W). Salmon caught at Strathy Point possibly are at their first migratory landfall and are destined not only for the rivers of both W and E Scotland, but also for Ireland (Shearer 1986). For management purposes, ICES (International Council for the Exploration of the Sea) distinguishes European stocks of Atlantic salmon into four entities: viz. 1SW and multi sea-winter (MSW) ‘northern’ (Russia – Scandinavia) European, and 1SW and MSW ‘southern’ (British Isles – Spain) European components (http://www.ices.dk/reports/ACFM/2005/WGNAS/wgnas05.pdf). Salmon taken at the interceptory fishery at Strathy Point undoubtedly are of mixed stocks and are representative of a significant element of the ICES southern European component (ICES 2004). Both Lepeophtheirus salmonis and especially Caligus elongatus have a propensity to detach from the host fish on encountering hyposaline water (see, e.g., Todd et al. 2000 for data from the river Tay estuary, E Scotland). Because of (i) the coastal distance of Strathy Point from large river estuaries, (ii) the method of trapping, and (iii) the minimal postcapture handling of the fish, salmon taken there offer the most realistic abundance data for these 2 parasites on adult wild salmon in the British Isles. Additional unique features are that, prior to their interception, these salmon have not been subject to recent fishery pressure (which may, for example, be size selective), and will not recently have encountered salmon farms.

The traps were inspected twice daily, the fish removed individually by hand, and killed with a single blow to the head. St Andrews and Fisheries Research Services (FRS), Montrose, personnel collected 2 independent sets of fish data over the period 1998–2005. St Andrews personnel were permitted to sample salmon for parasites on a single occasion in each season between 1998 and 2005 (n = 403). Fork length (to 0.1 cm) and weight (to 0.05 kg) were recorded and scale samples removed from the standard area (Anon.

1985) for subsequent ageing of all fish. MSW fish, farm escapees, and any wild fish that had previously spawned would be subject to different infestation dynamics from the 1SW fish which comprised the bulk of the sample in each year. Therefore, these fish (n = 38) were identified by scale reading and excluded from the data. The annual parasite data generally were obtained on 2 consecutive days, during weeks 25–30. In 1998, however, separate day visits were made in June (wk 26) and July (wk 30).

Since 1993 FRS have routinely monitored fork length, weight and sea age of salmon captured at Strathy Point. The multiple repeat FRS samplings throughout each netting season thus provide a comprehensive time series of fish data (but no parasite information) within and among years. FRS sample days within years ranged from 8 (1998) to 20 (2003), with the earliest and latest dates of their monitoring being Julian days 168 (wk 24, 2003) and 232 (wk 33, 2002 and 2003), respectively. A total of 4584 fish was sampled by FRS between 1993 and 2004: the length/weight data for these fish provided the specific length-mass equation used for calculating condition factor (CF) of the fish sampled independently (1998–2005) by St Andrews personnel for parasites. The FRS 1998-2004 data set (n = 2145 fish), which was contemporaneous with the St Andrews salmon/parasite data, was reduced to 2110 by the exclusion of 11 d on which <5 fish were sampled.

Over the course of the 10 wk June–August netting season, returning salmon tend to increase in length and weight (Todd et al. 2006), and both host fish size and CF were considered likely to be important independent variables for the analysis of within and between-year variation in parasite abundances. That the parasite data could be obtained only for single annual visits therefore necessitated an assessment of how 245

SUMBAWS

Final Report

representative are single samples of, for example, host CF for each year. The FRS data allowed within- and between-year analysis of fish parameters, including CF. During sampling for parasites, fish were inspected by eye and all caligids removed and preserved for later confirmation of developmental stage and sex. For Lepeophtheirus salmonis, the pre-adult I, pre-adult II and adult stages of both sexes were categorized; there is no pre-adult stage in Caligus elongatus, which moults directly from the sessile chalimus IV to the mobile adult. Although stage C III and C IV chalimi can be visually assessed, reliable counts of all the attached post-larval chalimus stages can be obtained only by microscopic examination of fish. Time did not permit such a detailed examination because this is a commercial operation, with the fish being destined for human consumption and requiring icing, processing and onward transport. Nonetheless, time did permit counts of visible (mostly C III, C IV) chalimi (not attributed to species) in 4 y (1999, 2001, 2002, 2004). Typically, and in common with previous reports (e.g.

Jacobsen & Gaard 1997), chalimi were very rare and, where present, usually attached to the dorsal, caudal or paired fins. Data for chalimi were not included in the present statistical analyses, which are confined to ‘mobile’ (pre-adult I, pre-adult II, adult) L. salmonis and (adult) C. elongatus only.

Fish and parasite data analyses Data for both parasite species were obtained for 1999–2005 inclusive: Caligus elongatus was present in 1998, but time constraints precluded their enumeration. Separate statistical analyses were undertaken of the variation in the proportion of female parasites (both species), and of pre-adults (stages I and II pooled, Lepeophtheirus salmonis) over the period 1999–2005. Host CF was calculated for all fish sampled for parasite analysis (except 1998, because weights are lacking for the 18 fish sampled in June 1998), using 2 of the standard measures (Blackwell et al. 2000, Brenden et al. 2003; Marshall et al. 2004; Neff & Cargnelli 2004). Fulton’s index (K = 100(Weight in g/(Length in cm)3)) assumes growth to be isometric and scaled to the cube of the length. Any departure from an exponent of 3 will, therefore, result in a length-dependence of individual fish CF. A measure more independent of length is the relative mass index, WR (WR = W/WS), whereby W is the observed mass and WS the standard mass of individual fish calculated from a specific mass-length equation (Blackwell et al. 2000, Brenden et al. 2003).

The specific mass-length relationship for Salmo salar was determined from the FRS 1993–2004 data (n = 4584 fish). Regression of y (log Weight, g) on x (log Length, cm) was highly significant (y = -2.17 + 3.14.x; r =0.934, p < 0.001), and from this the geometric mean regression (y = -2.57 + 3.36.x; see Ricker 1973) was derived for calculating the standard mass (WS) of individual fish, and thence WR. Body surface area of all individual fish (except 18 from June 1998) was estimated by applying the S. salar regression model of Jacobsen & Gaard (1997) whereby, for a given weight in g, Surface area = 9.5864 x 10log(Weight) x 0.629 .

All parasite count data for replicate fish were log(x+1) transformed, and all proportion data for females and pre-adults, within years and for the separate species, were arcsine transformed prior to analysis. Prior to ANOVA, homogeneity of variances was assessed by Levene’s test and inspection of standardized residuals.

In some instances variances remained heterogeneous following transformation, but ANOVA is robust to departures from the assumption of equality of variances in cases where samples are large and the number of treatments >6 (Underwood 1997). Where they occurred, significant outcomes for Levene’s test are reported.

Variation in parasite species abundances among years was analyzed by one-way ANOVA, followed by Tukey post hoc tests. For more detailed analysis of variation in the 2 species’ abundances, of the proportions of pre-adult Lepeophtheirus salmonis, and of the proportionality of females among individual fish and within each species, stepwise multiple regression analysis was applied to the 1999–2005 data (i.e. excluding 1998 when Caligus elongatus were not enumerated). Separate stepwise analyses were undertaken by including either K or WR as the measure of fish CF. Ten stepwise analyses concerned the separate parasite species, with 2 for both species pooled (Total lice). Both forward and backward stepwise regression was undertaken for each analysis, and in all cases both procedures gave an identical outcome.

In addition to assessing the possibility of inter-specific competition between Lepeophtheirus salmonis and Caligus elongatus, we investigated whether or not fish size might influence parasite abundances, that parasite loadings might drive host CF, or that fish in poor condition are perhaps disproportionately vulnerable to parasite infestation. Furthermore, the degree of spatial aggregation that parasite species display amongst individual hosts can have important consequences for inter- and intra-specific interactions on individual hosts and is conveniently expressed by least squares linear regression of log variance on log mean abundance (Taylor 1961). Taylor’s ‘power law’ typically results in species-specific slopes to such regressions.

246

SUMBAWS

Final Report

Results Choice of measure of condition factor (CF) The 2 CF measures – K and WR – were highly correlated (FRS data, 1993–2004; r = 0.954, n = 4584, p < 0.001) and K showed the expected positive dependence on fish length (r = 0.110, n = 4584, p < 0.001). By contrast, WR showed a significant negative correlation with length for 1993–2004 overall (r = -0.192, n = 4584, p < 0.001). Thus, despite calculation of WR with an exponent of 3.36 to preclude the mathematical problem of length-dependence, WR does reveal a tendency for longer salmon to have weighed less than predicted by the overall geometric mean mass-length relationship, and thereby to show poorer condition.

Variation in host fish size and CF As expected, host length and weight of the FRS sampled fish increased during each 10 wk season. For example, correlation analyses for daily mean fish length and Julian day of sampling showed significant outcomes for every year over the period 1998–2004 (FRS data, p varying from 0.033 (1999, n = 115) to < 0.001 (1998, n = 409; 2000, n = 309)). By contrast, mean CF – as measured by K – showed no significant correlation with Julian day for any of the 7 y. WR showed consistently negative correlation coefficients for mean CF against Julian day of sampling for all 7 y, but none were significant. Thus, the single annual St Andrews samples can be considered representative of CF (here preferably expressed as the length-corrected index, WR) for the 1SW cohort of salmon for each of the years sampled.

For those fish sampled by St Andrews, log length varied significantly among years (ANOVA 1998– 2005, F7,402 = 4.54, p < 0.001), as did log weight (ANOVA 1999–2005, F6,360 = 3.09, p = 0.006), K (ANOVA 1999–2005, F6,360 = 47.34, p < 0.001 (Levene’s test = 2.532, p = 0.021)) and WR (ANOVA 1999– 2005, F6,360 = 49.50, p < 0.001 (Levene’s test = 2.951, p = 0.008)). There was an overall decline in host weight and WR throughout the study period, excluding 1998 (for which weights were missing for 18 fish), but no obvious trend in host length (Table 1, Figure 1).

Table 1 Salmo salar. Minimum to maximum and means of sampled fish lengths (log cm), weights (log kg) and condition factors (K, exponent 3; WR exponent 3.36) for fish sampled for Lepeophtheirus salmonis and Caligus elongatus. Weights were available for only 24 fish in 1998 (*Weights were lacking for 18 fish sampled in June 1998; n = 24 for weights and CF measures) Fork length

Weight

K

WR

Year

No. of fish

Min.-max.

Mean

Min.-max.

Mean

Min.-max.

Mean

Min.-max.

Mean

1998

42

48.5-69.8

61.6

1.2-3.8*

2.38*

0.84-1.24*

1.01*

0.71-1.02

0.85*

1999

39

52.3-66.4

58.9

1.4-3.6

2.43

0.91-1.4

1.19

0.75-1.17

1.00

2000

41

52.0-70.5

60.2

1.6-4.4

2.54

1.02-1.35

1.16

0.85-1.11

0.97

2001

43

51.4-71.2

61.0

1.6-4.0

2.53

0.98-1.3

1.12

0.82-1.08

0.93

2002

53

49.4-71.0

60.3

1.6-4.15

2.57

0.99-1.47

1.17

0.83-1.23

0.98

2003

64

54.0-66.0

60.5

1.4-3.4

2.27

0.68-1.36

1.03

0.57-1.13

0.86

2004

62

54.0-74.0

63.4

1.4-4.6

2.46

0.82-1.18

0.97

0.68-0.97

0.80

2005

59

52.3-74.4

60.5

1.27-4.1

2.23

0.88-1.4

1.02

0.74-1.18

0.85

247

SUMBAWS

Final Report

Figure 1 Salmo salar. Annual variation in host salmon condition factor (K, WR), 1999–2005, for fish sampled by St Andrews (n = 361). Values are means ± 95% c.i.

Annual abundances, Lepeophtheirus salmonis and Caligus elongates Lepeophtheirus salmonis occurred at 100% prevalence throughout all 8 y. Caligus elongatus prevalence in 4 of the 7 y also was 100%, but otherwise varied between 90.2% and 97.4%. L. salmonis antilogged mean abundance varied between 17.4 and 31.0 and there were significant between-year variations (ANOVA F7,402 = 7.95, p < 0.001 (Levene’s test = 2.256, p = 0.029)), with 2 statistically homogeneous groupings (Figure 2). For C. elongatus, abundances typically were low but 2002 was exceptionally high, resulting in annual means varying from 2.9 (2000) to 23.8 (2002). One-way ANOVA for C. elongatus log abundance also was highly significant for year (F6,360 = 37.04, p < 0.001), and there were 5 homogeneous abundance groupings (Figure 2).

An inverse relationship between the mean annual abundances of the 2 species at the population level (Figure 2) superficially might be indicative of inter-specific competition. The loadings of the 2 species on individual fish did, however, reveal an overall significant positive correlation (r = 0.169, n = 361, p < 0.001; Figure 3), which was consistent across 5 of the 7 y (analyses not shown). The exceptions were 2000, when abundances of Caligus elongatus were extremely low and thus the correlation test weak, and 2005 when abundances of both species were moderate to high. In general, within any one year, individual fish that are vulnerable to infestation by one species tend also to be at risk from the other, but it is likely the 2 species’ annual abundance patterns vary independently. One-way ANOVA for total sea lice abundances (log((L.s. + C.e.)+1)) on individual fish (1999–2005), showed a significant year effect (F6,360 = 2.37, p = 0.029 (Levene’s test = 3.650, p = 0.002)). It was expedient, therefore, to ascertain whether or not variation in putatively important host fish parameters might explain this variation, or the abundances of either species at the population level.

Figure 2 Lepeophtheirus salmonis and Caligus elongatus. Annual variations in mean [log(x+1) transformed] abundance ± 95% c.i. Prevalence was 100% for both species, except for C. elongatus in 1999 (97.4%), 2000 (90.2%) and 2004 (96.8%). Post-hoc Tukey groupings, following ANOVA (L. salmonis: F6,343 = 6.10, p < 0.001; C. elongatus: F5,301 = 38.76, p < 0.001), are shown in lower case letters.

248

SUMBAWS

Final Report

Figure 3 Lepeophtheirus salmonis and Caligus elongatus. Co-occurrence of the two species on individual host fish, 1999â&#x20AC;&#x201C;2005. For 5 of the 7 y there was a significant power regression. The fitted regression (log y = 0.628 + 0.254.log x, r = 0.169, p = 0.001) is for the pooled data.

Lepeophtheirus salmonis and host/Caligus parameters Least squares linear regression showed a significant positive correlation between log abundance of Lepeophtheirus salmonis and log host fish length (n = 403, r = 0.245, p < 0.001). The bulk of this variation amongst hosts could, however, be explained by the greater body surface area of the longer fish (log L. salmonis = -1.292 + 0.867.log Surface area; n = 385, r = 0.216, p < 0.001). It is important to note, however, that L. salmonis tends to aggregate in restricted, preferred areas on the host fish body surface (Todd et al. 2000). L. salmonis abundance showed very significant and positive partial correlations only with the abundance of Caligus elongatus and with host fish length (Table 2). Within any one year, longer fish tended to bear greater infestations of either species and fish with high L. salmonis infestations also tended to carry higher intensity infestations of C. elongatus. The proportion of female L. salmonis on individual fish showed strong positive partial correlations with the abundance both of L. salmonis and of C. elongatus (Table 2). A similar outcome was shown for the proportion of pre-adult L. salmonis, except that the partial correlation with C. elongatus was negative. For both of these L. salmonis dependents, therefore, parasite abundance alone appears to be the primary determining factor. Higher proportions of females arise as a result of accumulation of the longer-lived gender, and greater proportions of pre-adult L. salmonis are characteristic of recent infestation.

Caligus elongatus and host/Lepeophtheirus parameters As for Lepeophtheirus salmonis, Caligus elongatus abundance showed an overall significant positive correlation with host fish length (log C. elongatus = -2.264 + 1.824.log fish length; n = 360, r = 0.151, p = 0.004), but this was significant for only 3 (2001, 2003, 2005) of the 7 y. The 2 stepwise analyses for C. elongatus abundance (including either K or WR as the measure of CF) showed closely similar outcomes, and a clear difference to the analytical outcomes for L. salmonis abundance. For C. elongatus abundance, Year, CF, and L. salmonis abundance all proved to be (positive) significant independents (Table 2). The proportion of female C. elongatus on individual fish was consistently high (mean annual min.-max. 0.835 (2002) â&#x20AC;&#x201C; 0.926 (1999)), and much higher than for L. salmonis (mean annual min.-max. 0.598 (2001) â&#x20AC;&#x201C; 0.724 (1998)). The consistency of this C. elongatus parameter amongst individual fish resulted in no significant correlation with any of the independents investigated (Table 2).

249

SUMBAWS

Final Report

Table 8.1 Lepeophtheirus salmonis (L.s.) and Caligus elongatus (C.e.). Summary of stepwise multiple regression for infestation parameters of individual host salmon (n = 361; 1999-2005), for each species and the two summed (Total lice). Parasite abundances, host fish length and host surface area were log transformed. The respective regression coefficients, partial correlation coefficients (in parentheses) and associated multiple regression correlation coefficients (r), and probabilities (p) are shown. Independents excluded by stepwise analysis are denoted by ‘–’; ns = not significant.

Lepeophtheirus salmonis Total

Caligus elongatus Total Total Proportion female (K) (WR)

Proportion female

Proportion male

6.998 (<0.001) 4.753 (0.001)

13.657 (<0.001) -4.867 (0.001)

0.238 (0.002)

–

–

–

–

–

0.382 (0.030)

Total lice

Independent variables Total L.s. (log) Total C.e. (log) Host length (log) Condition factor (K) Condition factor (WR) Host surface area (log)

0.087 (0.010) 2.185 (<0.001) –

0.242 (0.002)

– –

–

–

2.200 (<0.001)

–

–

–

–

–

–

0.528 (0.022) –

–

–

Year

–

–

–

r p

0.319 <0.001

0.303 <0.001

0.321 <0.001

–

–

0.055 (<0.001) 0.300 <0.001

0.057 (<0.001) 0.293 <0.001

ns

Total lice and host parameters Within any one year, the wide range of abundances on individual fish (L. salmonis, 1–114; C. elongatus, 0–107; both species pooled, 4–152) shows that salmon generally do not bear a maximal loading of either species, or the two combined. Stepwise regression confirmed that host length was the only significant independent for total parasite burden (Total lice, Table 2). The finding for individual fish that is, perhaps, of especial importance is that reduced CF (either K or WR) was not associated with, or driven by, total parasite loading although this independent was (positively) significant for C. elongatus (see above). Years of poorer condition fish (i.e. 2003–05) were not characterized by exceptionally high total parasite loadings, and, overall, poor condition fish in any one year showed no greater tendency to carry greater (or lesser) burdens of sea lice than did high condition fish.

The presence of attached chalimi confirms recent infection of the fish. Though sampling for (sessile) chalimi was less rigorous than for the larger, mobile stages, chalimus stages generally were rare (Table 3).

Two fish in 2002 carried 10 and 40 chalimi each, indicating strong patterns of overdispersion (aggregation) of recent colonization by parasites amongst individual host fish. The patterns of aggregation for the 2 species (excluding chalimi) amongst the sampled host fish in the various years are summarized by the power regressions in Figure 4. Although the regression coefficients indicate a convergence of the species’ patterns of spatial dispersion amongst host fish at extremely high mean population densities (~68 lice of each species), over the observed density ranges Caligus elongatus showed consistently greater aggregation, and especially so at low densities. The infection dynamics, demographics and dispersion of these parasite species among individual hosts are, therefore, fundamentally distinct.

250

SUMBAWS

Final Report

Table 3 Lepeophtheirus salmonis and Caligus elongatus. Incidence of sessile chalimus stages on 1SW salmon in four years (no data for 1998, 2000, 2003, 2005). Chalimi were enumerated, but not identified to species, and were excluded from all analyses Numbers of chalimi 0 1 2 3 4 10 40 Total fish

1999 38 0 1 0 0 0 0 39

Number of fish 2001 2002 35 40 4 8 3 3 1 0 0 0 0 1 0 1 43 53

2004 54 4 1 2 1 0 0 60

Figure 4 Lepeophtheirus salmonis and Caligus elongatus. Taylorâ&#x20AC;&#x2122;s power law (regression, log variance on log mean abundance) fits for all 8 y (L. salmonis; log y = -1.47 + 2.64.log x, r = 0.833, p = 0.010) and 7 y (C. elongatus; log y = 0.42 + 1.53.log x, r = 0.774, p = 0.041), respectively. The data for C. elongatus included zero abundances.

Discussion Initially, Caligus elongatus was the only major pest caligid to the Atlantic salmon farming industry in E Canada (Piasecki & MacKinnon 1995), and Hogans & Trudeau (1989) reported C. elongatus at prevalences up to 100%, compared to Lepeophtheirus salmonis at a maximum only of 8%. In recent years, however, L. salmonis has become the primary pest species, just as it is for the Atlantic salmon farming industry both in Europe (e.g. Pike & Wadsworth 1999) and on the Pacific coast of Canada (e.g. Morton et al. 2004).

Nonetheless, analyses of potential inter-specific interactions for ectoparasites on individual host fish are presently lacking. Detailed analyses have been undertaken of the spatial disposition of L. salmonis within host fish, and of the spatial segregation of its two genders (Todd et al. 2000), but the lack of comparable data for C. elongatus within individual hosts presently precludes detailed analysis of inter-specific interactions. Irrespective of this, Bakke & Harris (1998) concluded that these two species were amongst the pathogens most likely to detrimentally impact both wild and farmed Canadian salmonid stocks in the future.

Certainly, the importance of C. elongatus to the Scottish industry has long been recognized, but only recently have population modeling studies, founded on extensive commercial databases, begun to lead to a better understanding of the strong seasonal dynamics of C. elongatus in the farm environment (e.g.

McKenzie et al. 2004).

Our annual time-series of mean abundances for the 2 species on wild salmon did show a generally inverse relationship (Fig. 2), but this pattern cannot be attributed to inter-specific competition: in 5 of 7 years, individual fish that bore higher intensity infestations of Lepeophtheirus salmonis also tended to carry 251

SUMBAWS

Final Report

a higher intensity of Caligus elongatus. Assessments of competitive interactions between these 2 species are, however, additionally problematic, because caligid ectoparasites typically are aggregated in restricted areas of the host body surface and yet the mobile pre-adult and adult stages can readily move over the host body surface. C. elongatus is more mobile than is L. salmonis, and the males and females of L. salmonis are spatially segregated on individual host fish. Females predominate along the posterior dorsal and ventral midlines, and males on the anterior dorsal midline as well as the head (Todd et al. 2000). Rohde (1993) suggested that microhabitat restrictions or specialization could enhance ectoparasite mating encounters, especially for species which typically are sessile and of low population density. The combination of their high to absolute prevalence, and the motility of the pre-adult/adult stages of L. salmonis and adults of C. elongatus, ought to confer high frequencies of male-female encounters. In contrast to Rohdeâ&#x20AC;&#x2122;s suggestion, sexual segregation of L. salmonis may actually be determined by intra-specific agonistic interactions, with the (large) females aggressively maintaining themselves in the most preferred body regions. Both sexes of C. elongatus are small and may similarly be displaced from these areas by female L. salmonis, but there is no evidence from the present data of L. salmonis abundance on individual fish exerting a negative effect on C. elongatus abundance.

Open ocean infestation, and interactions with host CF Chalimus stages are rare on Atlantic salmon captured on their oceanic feeding grounds (e.g. Jacobsen & Gaard 1997). Returning 2SW salmon do, however, typically bear significantly greater mean abundances of mobile L. salmonis than do 1SW fish (Todd et al. 2000) and, as a generalization, overdispersion amongst individual hosts is characteristic of host-parasite associations (e.g. Shaw & Dobson 1995, and Fig. 4). The very low numbers, but high degree of overdispersion, of chalimi (Table 3) suggest that recent coastal infestations by sea lice originating from other wild fish, or from farmed stocks in sea pens, are of relatively minor importance to adult 1SW salmon returning to coastal waters. Wild post-smolt Salmo salar can acquire chalimi within days of first entering seawater (Finstad et al. 1994, 2000); but the adult lice burdens of return migrant 1SW and 2SW salmon, captured some 12+ or 24+ months later, are most unlikely to be survivors of initial infestation (Todd et al. 2000). Post-settlement development from the infective, planktonic copepodid larva to the final sessile (chalimus C IV) stage can take up to 270 degree days post infection (ddpi) for Caligus elongatus (Piasecki & MacKinnon 1995) or 350 ddpi for Lepeophtheirus salmonis (Johnson & Albright 1991). At oceanic spring/early summer temperatures, our fish carrying preadult L. salmonis, or unidentified chalimi, will have been most recently infected 4â&#x20AC;&#x201C;6 wk before capture.

Continued re-infestation in the open ocean therefore is a persistent feature of Lepeophtheirus salmonis on wild salmon (Jacobsen & Gaard 1997; Todd et al. 2000). In the absence of specific identification of attached chalimi, and because C. elongatus lacks a pre-adult stage, it is not possible to extend this conclusion to that species with certainty. Nonetheless, all caligid chalimus stages are sessile, and thereby not so subject to under-estimation as a result of dislodgement due to capture method or post-capture handling; their very scarcity points to C. elongatus also colonising fish persistently and over extended periods.

Host CF is not a significant determinant of abundance of mobile Lepeophtheirus salmonis or total lice for wild salmon (Table 3), and though the fish with 40 chalimi was the individual of lowest CF (WR = 0.843) amongst the 53 sampled that year, the CF (0.987) of the fish with the next highest chalimus density (10) was close the 2002 WR mean (0.999). For Caligus elongatus abundance, CF was shown by stepwise regression to have a positive partial correlation; that is, fish of higher CF tended to carry higher intensities of C. elongatus. The positive association between the 2 species on individual fish for 5 of the 7 y (Fig. 3), and the outcomes of stepwise multiple regression (Table 2), all show that individual fish are similarly vulnerable to infestation by the 2 species. Within any one year, fish with a large burden of L. salmonis tended also to carry a relatively high infestation of C. elongatus, but among years the population abundances of the 2 species probably fluctuate independently.

Irrespective of the evidently highly adapted reproductive strategy of Lepeophtheirus salmonis in colonizing wild Salmo salar with 100% success every year, there clearly must be a sharp dichotomy between the processes and risks of initial infestation of post-smolt S. salar (and both juvenile and adult Salmo trutta) in coastal waters as opposed to the open ocean. Available inshore data indicate that the planktonic larval copepod stages undergo vertical migrations, but typically they are close to the surface and tend to be caught in numbers only adjacent to salmon farms or at river mouths (e.g. Tully & Nolan 2002, McKibben & Hay 2004; Penston et al. 2004). The apparent importance of hyposaline fronts to the local concentration of infective copepodids in coastal waters (McKibben & Hay 2004) can have no relevance to infestation dynamics in the open ocean. It is quite likely that oceanic infestation arises simply as a result of 252

SUMBAWS

Final Report

the larvae maintaining the appropriate depth near the surface or at thermal fronts and the salmon encountering the parasite by chance whilst foraging.

A fundamental constraint of any study of parasite abundance and demography on wild hosts is that, almost by definition, the observer can sample only surviving hosts and this constraint is especially acute for migratory wild salmonids. For sea trout (Salmo trutta), MacKenzie et al. (1998) reported a negative relationship between C I–II intensity and host CF (WR), but also a positive relationship between CF and C III–IV intensity. This they deduced to arise from the differential periods of fish in seawater, but equally these contrasts could be attributable to early mortality of heavily-infested individual fish which were no longer available for capture. Despite their large sample sizes, Tully et al. (1993) failed to capture sea trout juveniles bearing adult Lepeophtheirus salmonis and they concluded that any such host fish probably had died as a consequence of the infestation.

There is no clear evidence to suggest that poor condition Salmo trutta are especially susceptible to infestation by sea lice copepodids, or that poor individual host CF is caused by caligid infestation (Mo & Heuch 1998; Schram et al. 1998; Murray 2002); the latter would, however, seem intuitive at least for heavy but sub-lethal loadings (see Wells et al. 2006). Although no inferences can be drawn for post-smolt early mortalities, for Salmo salar 1SW adults the present data give no indication that ectoparasitic infestation has any detrimental effect on condition of fish that have survived to return to home waters. Notwithstanding the positive partial correlation for CF with Caligus elongatus (Table 2), the indications are that the significant among-year variation in CF reported here (Fig. 1) for the 8 cohorts of 1SW Atlantic salmon is not driven by parasite abundance, but is most probably explained by fluctuations in host feeding and growth opportunities, as has been reported for Pacific salmon species (e.g. Hinch et al. 1995).

Demography of Lepeophtheirus salmonis and Caligus elongatus, and risks of infestation Female and adult predominance is characteristic of Lepeophtheirus salmonis infestations of wild salmon (e.g. Johnson et al. 1996, Jacobsen & Gaard 1997). For individual fish, higher abundance L. salmonis infestations correlated positively with a greater proportion of females (r = 0.166, n = 403, p = 0.001; annual mean min.-max. 0.598 [2001] – 0.724 [1998]), and a greater proportion of pre-adult stages (r = 0.305, n = 403, p < 0.001; annual mean min.-max. 0.033 [1999] – 0.191 [2000]). Higher intensity infestations could be attributable either to elevated rates of recent infestation (high proportion of pre-adults), and/or elevated rates of past infestation (accumulation of the larger and longer-lived gender). Infestation risk almost certainly is not continuous, and high(er) overall past rates of infestation may be temporally and/or spatially heterogeneous as individual fish transit through relatively high(er) and low(er) risk areas (e.g. Shaw & Dobson 1995). A convergence in this regard is apparent with the chalimus data (Table 3), and the 2 species’ patterns of aggregation (Fig. 4): certain individual fish clearly are at either higher risk of infestation, or, by chance, encounter high risk areas of infestation more frequently than do most other fish.

Demographic comparisons between Lepeophtheirus salmonis and Caligus elongatus infestations of wild adult salmon are problematic, if only because of the lack of field data on the oceanic distribution of the infective larval stage, there being no C. elongatus pre-adult stage (precluding the ageing of infestations, and their recent history for individual fish), and the far broader host species range of C. elongatus. Multiple host species might be expected to result in C. elongatus larval production being more widespread and continuous throughout the North Atlantic, and for its infestation probability to be spatially and temporally more homogeneous. Although Taylor’s power regression (Fig. 4) showed the expected significant difference in regression coefficient for the two species, this analysis actually affirms the opposite – a tendency for greater heterogeneity of infestation risk and aggregation at low abundance for C. elongatus. Quantitatively, the infestation risk that C. elongatus presents to wild salmon is distinct from L. salmonis, and the difference may well be explained by inter-specific contrasts in fecundity and successful production (and host encounter) of infective copepodids. A conservative estimate of typical lifetime fecundity for L. salmonis females on wild salmon is several thousand (e.g. Heuch et al. 2000), but for C. elongatus this will be markedly lower, at perhaps only a few hundred even if several pairs of eggstrings are produced in a female’s lifetime (Pike & Wadsworth 1999).

Female Lepeophtheirus salmonis are much larger, slower-developing and longer-lived than males, and the levels of female predominance reported here (annual min.-max. 0.598–0.724) appear typical of L. salmonis infestations of wild Salmo salar (Jacobsen & Gaard 1997; see also Johnson et al. 1996 for Oncorhynchus nerka). For Caligus elongatus, however, despite the two sexes being the same size, female predominance is even greater (annual min.-max. 0.835 to 0.948), indicating a marked differential longevity of males and females on wild fish. Piasecki & MacKinnon (1995) found that male C. elongatus (infesting experimental char, Salvelinus alpinus) die soon after copulation. None of the independents here analyzed by 253

SUMBAWS

Final Report

stepwise regression (Table 2) correlated significantly with variation in the proportion of female C. elongatus amongst individual fish; but if male post-copulatory death is typical of C. elongatus on wild salmon, then this alone could explain the extreme female bias. C. elongatus adults are, however, very much more active across the host fish surface than L. salmonis, and their vagility apparently leads to their commonly detaching from the host fish (e.g. review by Pike & Wadsworth 1999). The possibility therefore remains that males might be more likely to detach from host fish and at least temporarily become planktonic. Certainly, adult C. elongatus commonly can be caught in surface plankton net tows in coastal waters (Todd et al. 1996).

The focus of research into the impacts of sea lice on salmon aquaculture has necessarily remained primarily on Lepeophtheirus salmonis (e.g. Pike & Wadsworth 1999), although Caligus elongatus can – especially in the summer months – be problematic to cultured fish (e.g. Revie et al. 2002). Our data show that C. elongatus infestation of wild salmon can in some years (2002; Fig. 2) equal L. salmonis and that C. elongatus has been under-estimated in its importance on wild fish, either because of the method of fish capture or the interception of host fish in brackish waters. We found consistently high infestations of L. salmonis and that intensities show no statistical association with host CF. Furthermore, the apparent pattern of a decline in CF of returning adult 1SW salmon in recent years appears not to be related to L. salmonis or C. elongatus infestation.

Acknowledgments Funding was variously provided by the Natural Environment Research Council, UK (Grant No. GR3/11122), a Scottish Association of Marine Science Bursary, and the European Commission (SUMBAWS, Contract No. Q5RS-2002-00730), to all of whom we are grateful. For fieldwork assistance, we thank N Hastings, R Stevenson, N Walker and H Reinardy. We thank also M Pawson and G Copp for valuable comments on earlier drafts. S Paterson kindly provided invaluable access to fish, without which this work would not have been possible.

References Anonymous (1985) Atlantic salmon scale reading. Report of the Atlantic salmon scale reading workshop.

ICES, Copenhagen Anonymous (2004) Scottish salmon and sea trout catches, 2003. Statistical Bulletin, No. Fis/2004/1.

Fisheries Research Services, Aberdeen, Scotland Bakke TA, Harris PD (1998) Diseases and parasites in wild Atlantic salmon (Salmo salar) populations. Can J Fish Aquat Sci 55(Suppl 1):247–266 Berland B (1993) Salmon lice on wild salmon (Salmo salar L.) in western Norway. In: Boxshall GA, Defaye D (eds) Pathogens of wild and farmed fish. Sea lice. Ellis Horwood, Chichester, p 179–187 Bjørn PA, Finstad B (1997) The physiological effects of salmon lice infection on sea trout post smolts.

Nord J Freshw Res 73:60–72 Bjørn PA, Finstad B (2002) Salmon lice, Lepeophtheirus salmonis (Krøyer), infestation in sympatric populations of Arctic char, Salvelinus alpinus (L.), and sea trout, Salmo trutta (L.), in areas near and distant from salmon farms. ICES J Mar Sci 59:1–9 Bjørn PA, Finstad B, Kristoffersen R (2001) Salmon lice infection of wild sea trout and Arctic char in marine and freshwaters: the effects of salmon farms. Aquac Res 32:947–962 Blackwell BG, Brown ML, Willis DW (2000) Relative weight (Wr) status and current use in fisheries assessment and management. Rev Fish Sci 8:1-44 Brenden TO, Murphy BR, Birch JB (2003) Statistical properties of the relative weight (Wr) index and an alternative procedure for testing Wr differences between groups. N Am J Fish Manage 23:1136–1151 Bristow GA, Berland B (1991) A report on some metazoan parasites of wild marine salmon (Salmo salar L.) from the west coast of Norway with comments on their interactions with farmed salmon. Aquaculture 98:311–318 Butler JRA (2002) Wild salmonids and sea louse infestations on the west coast of Scotland: sources of infection and implications for the management of marine salmon farms. Pest Manage Sci 58:595–608

254

SUMBAWS

Final Report

Finstad B, Johnsen BO, Hvidsten NA (1994) Prevalence and mean intensity of salmon lice, Lepeophtheirus salmonis Krøyer, infection on wild Atlantic salmon, Salmo salar L., postsmolts. Aquac Fish Manage 25:761–764 Finstad B, Bjørn PA, Grimnes A, Hvidsten NA (2000) Laboratory and field investigations of salmon lice [Lepeophtheirus salmonis (Krøyer)] infestation on Atlantic salmon (Salmo salar L.) post-smolts. Aquac Res 31:795–803 Heuch PA, Nordhagen JR, Schram TA (2000) Egg production in the salmon louse [Lepeophtheirus salmonis (Krøyer)] in relation to origin and water temperature. Aquac Res 31:805–814 Hinch SG, Healey MC, Diewert RE, Henderson MA (1995) Climate change and ocean energetics of Fraser River sockeye (Oncorhynchus nerka). In: Beamish, RJ (ed) Climate change and northern fish populations. Can Spec Publ Can Fish Aquat Sci p. 439–445 Hogans WE, Trudeau DJ (1989) Preliminary studies on the biology of sea lice, Caligus elongatus, Caligus curtus and Lepeophtheirus salmonis (Copepoda: Caligoida) parasitic on cage-cultured salmonids in the lower Bay of Fundy. Can Tech Rep Fish Aq Sci, No. 1715 ICES (2004) Report of the working group on North Atlantic salmon. ICES CM 2004/ACFM:20, Ref. I. Advisory Committee on Fishery Management. ICES, Copenhagen Jacobsen JA, Gaard E (1997) Open-ocean infestation by salmon lice (Lepeophtheirus salmonis): comparison of wild and escaped farmed Atlantic salmon (Salmo salar L.). ICES J Mar Res 54:1113–1119 Johnson SC, Albright LJ (1991) The developmental stages of Lepeophtheirus salmonis (Krøyer, 1837) (Copepoda, Caligidae). Can J Zool 69:929–950 Johnson SC, Blaylock RB, Elphick J, Hyatt KD (1996) Disease induced by the sea louse (Lepeophtheirus salmonis) (Copepoda: Caligidae) in wild sockeye salmon (Oncorhynchus nerka) stocks of Alberni Inlet, British Columbia. Can J Fish Aquat Sci 53:2888–2897 Johnson SC, Treasurer JW, Bravo S, Nagasawa K, Kabata Z (2004) A review of the impact of parasitic copepods on marine aquaculture. Zool Stud 43:229–243 Kabata Z (1979) Parasitic Copepoda of British fishes. Monograph No. 152. Ray Society, London Krkošek M, Lewis MA, Volpe JP (2005) Transmission dynamics of parasitic sea lice from farm to wild salmon. Proc Roy Soc Lond B 272:689-696 MacKenzie K, Longshaw M, Begg GS, McVicar AH (1998) Sea lice (Copepoda: Caligidae) on wild sea trout (Salmo trutta L.) in Scotland. ICES J Mar Sci 55:151–162 Marshall CT, Needle CL, Yaragina NA, Ajiad AM, Gusev E (2004) Deriving condition indices from standard fisheries databases and evaluating their sensitivity to variation in stored energy reserves. Can J Fish Aquat Sci 61:1900–1917 McKenzie E, Gettinby G, McCart K, Revie CM (2004) Time-series models of sea lice Caligus elongatus (Nordmann) abundance on Atlantic salmon Salmo salar L. in Loch Sunart, Scotland. Aquac Res 35:764– 772 McKibben MA, Hay DW (2004) Distributions of planktonic lice larvae Lepeophtheirus salmonis in the inter-tidal zone in Loch Torridon, western Scotland in relation to salmon farm production cycles. Aquac Res 35:742–750 Mo TA, Heuch PA (1998) Occurrence of Lepeophtheirus salmonis (Copepoda: Caligidae) on sea trout (Salmo trutta) in the inner Oslo Fjord, south-eastern Norway. ICES J Mar Sci 55:176–180 Morton A, Routledge R, Peet C, Ladwig A (2004) Sea lice (Lepeophtheirus salmonis) infection rates on juvenile pink (Oncorhynchus gorbuscha) and chum (Oncorhynchus keta) salmon in the nearshore marine environment of British Columbia, Canada. Can J Fish Aquat Sci 61:147–157 Murray AG (2002) Using observed load distributions with a simple model to analyse the epidemiology of sea lice (Lepeophtheirus salmonis) on sea trout (Salmo trutta). Pest Manage Sci 58:585–594 Nagasawa K (2001) Annual changes in the population size of the salmon louse Lepeophtheirus salmonis (Copepoda: Caligidae) on high-seas Pacific salmon (Oncorhynchus spp.), and relationship to host abundance. Hydrobiologia 453/454:411–416 255

SUMBAWS

Final Report

Neff BD, Cargnelli LM (2004) Relationships between condition factors, parasite load and paternity in bluegill sunfish, Lepomis macrochirus. Env Biol Fishes 71:297–304 Penston MJ, McKibben MA, Hay DW, Gillibrand PA (2004) Observations on open-water densities of sea lice larvae in Loch Shieldaig, western Scotland. Aquac Res 35:793–805 Piasecki W, MacKinnon BM (1995) Life cycle of a sea louse, Caligus elongatus von Nordmann, 1832 (Copepoda, Siphonostomatoida, Caligidae). Can J Zool 73:74–82 Pike AW, Wadsworth SL (1999) Sealice on salmonids: their biology and control. Adv Parasitol 44:233– 337 Revie CW, Gettinby G, Treasurer JW, Rae GH (2002) The epidemiology of the sea lice, Caligus elongatus Nordmann, in marine aquaculture of Atlantic salmon, Salmo salar L., in Scotland. J Fish Dis 25:391–399 Ricker WE (1973) Linear regressions in fishery research. J Fish Res Board Can 30:409-434 Rohde K (1993) Ecology of marine parasites. CAB International, Wallingford Schram TA, Knutsen JA, Heuch PA, Mo TA (1998) Seasonal occurrence of Lepeophtheirus salmonis and Caligus elongatus (Copepoda: Caligidae) on sea trout (Salmo trutta), off southern Norway. ICES J Mar Sci 55:163–175 Shaw DJ, Dobson AP (1995) Patterns of macroparasite abundance and aggregation in wildlife populations: a quantitative review. Parasitology 111:S111–S133 Shearer WM (1986) The exploitation of Atlantic salmon in Scottish home water fisheries in 1952-83. In: Jenkins, D Shearer, WM (eds) The status of the Atlantic salmon in Scotland. Institute of Terrestrial Ecology, Abbots Ripton, p 7–49 Taylor LR (1961) Aggregation, variance and the mean. Nature 189:732–735 Todd CD, Laverack MS, Boxshall GA (1996) Coastal marine zooplankton: a practical manual for students. Second edition. Cambridge University Press, Cambridge Todd CD, Walker AM, Hoyle JE, Northcott SJ, Walker AF, Ritchie MG (2000) Infestations of wild adult Atlantic salmon (Salmo salar L.) by the ectoparasitic copepod sea louse Lepeophtheirus salmonis Krøyer: prevalence, intensity and the spatial distribution of males and females on the host fish.

Hydrobiologia 429:181–196 Todd CD, Walker AM, Ritchie MG, Graves JA, Walker AF (2004) Population genetic differentiation of sea lice (Lepeophtheirus salmonis Krøyer) parasitic on Atlantic and Pacific salmonids: analyses of microsatellite DNA variation amongst wild and farmed hosts. Can J Fish Aquat Sci 61:1176–1190 Todd CD, Hughes S, Walker AM, MacLean JC, Whyte BDM (2006) Negative impact of recent oceanic warming on salmon in the NE Atlantic. Manuscript under review.

Tully O (1989) The succession of generations and growth of the caligid copepods Caligus elongatus and Lepeophtheirus salmonis parasitizing farmed Atlantic salmon smolts (Salmo salar L.). J Mar Biol Assn UK 69:279–287 Tully O, Nolan DT (2002) A review of the population biology and host–parasite interactions of the sea louse Lepeophtheirus salmonis (Copepoda: Caligidae). Parasitology 124:S165–S182 Tully O, Poole WR, Whelan KF, Merigoux S (1993) Parameters and possible causes of epizootics of Lepeophtheirus salmonis (Krøyer) infesting sea trout (Salmo trutta L.) off the west coast of Ireland. In: Boxshall GA, Defaye D (eds) Pathogens of wild and farmed fish. Sea lice. Ellis Horwood, Chichester, p 202–213 Underwood AJ (1997) Experiments in ecology: their logical design and interpretation using analysis of variance. Cambridge University Press, Cambridge Wells AW, Grierson CE, Russon IJ, Reinardy H, Middlemiss C, Bjorn PA, Finstad B, Wendelaar Bonga SE, MacKenzie M, Todd CD, Hazon N (2006) The physiological effects of simultaneous abrupt seawater entry and sea lice infestation of wild sea trout (Salmo trutta L.) smolts. Manuscript under review

256

SUMBAWS

Final Report

Paper 3: PARASITE INTERACTIONS BETWEEN WILD AND FARMED SALMONIDS: THE APPLICATION OF GENETIC TECHNIQUES TO COPEPOD SEA LICE (LEPEOPHTHEIRUS SALMONIS (KRﾃ郎ER), CALIGUS ELONGATUS NORDMANN) INFESTING ATLANTIC SALMON (SALMO SALAR L.) AND SEA TROUT (SALMO TRUTTA L.) C. D. Todd (Submitted to Journal of Plankton Research) Abstract The caligid copepods Lepeophtheirus salmonis (Krﾃｸyer) and Caligus elongatus Nordmann are major pathogens of farm and wild salmonids throughout the North Atlantic. Since the early 1990s there has been considerable controversy regarding the extent to which infective larvae of the salmonid specialist L. salmonis originate from aquaculture cage sites and impact wild stocks of Atlantic salmon (Salmo salar L.) and sea trout (Salmo trutta L.). In view of the impracticality of tracking individual planktonic larvae from hatching to final host colonization, reliance has to be placed on indirect analytical and observation methods.

The various non-genetic and genetic experimental approaches taken in addressing this issue are reviewed.

Microsatellite DNA analyses have shown that L. salmonis comprises a single panmictic population throughout the North Atlantic; the levels of gene flow between wild and farmed hosts are sufficiently high as to prevent population genetic differentiation by random drift. However, because of the lack of significant population differentiation, no estimates of the levels of gene flow from farm to wild or vice versa are possible. Another area of concern regarding L. salmonis and the aquaculture industry is the possibility of the evolution of resistance to chemotherapeutant compounds. Reduced efficacy has been reported for numerous active compounds deployed by the industry and identification of a point mutation in domain II of the paratype voltage-sensitive sodium channel gene might be indicative of knockdown resistance to pyrethroid treatments. Whilst the focus of experimental research in the North Atlantic has been on L. salmonis, it is of concern that the more host generalist C. elongatus might become an especially severe pathogen to both salmonid and gadid host populations as the emerging cod (Gadus morhua L.) aquaculture industry develops.

Introduction Copepods comprise an extremely species-rich and diverse crustacean subclass with at least 10,000 named species (Huys and Boxshall, 1991). Their ecological importance is perhaps attributable largely to the key roles of planktonic cyclopoids and calanoids as primary consumers in pelagic freshwater and marine systems respectively; but parasitic copepods do exert important ecological and economic impacts on a wide array of invertebrate and fish species (Rohde, 1993). For example, caligids of the Order Siphonostomatoida are especially important ectoparasites of wild and farmed salmonids alike (Pike and Wadsworth, 1999; Johnson et al., 2004). Farming of Atlantic salmon (Salmo salar L.) commenced in the 1970s and worldwide production has increased steadily to >1.1 million tonnes (ICES, 2004). Lepeophtheirus salmonis (Krﾃｸyer) is a specialist ectoparasite of salmonids in their marine phase and infests all species of wild salmonid in both the North Pacific and North Atlantic (Pike and Wadsworth, 1999, Nagasawa, 2001), and quickly became a major pest at aquaculture sites. Furthermore, farmed and wild Atlantic salmon in the North Atlantic also are impacted by the host generalist Caligus elongatus Nordmann (Wootten et al., 1982; Schram, 1998), which has been recorded from >80 species of elasmobranch and teleost hosts (Kabata, 1979). These two species together are the most economically important metazoan pathogens to S. salar aquaculture in the North Atlantic, and estimates of the worldwide costs to the industry of caligids range up to US $100 million p.a.

(Johnson et al., 2004). For the Scottish industry alone, the total cost is at least ﾂ｣25 million p.a. (Rae 2002).

In NW Europe there has been a long-standing controversy regarding the extent to which cultured salmon are a significant source of sea louse infestations of both wild sea trout (Salmo trutta) and S. salar populations (McVicar et al., 1993; Butler, 2002; Todd et al., 2004). Various experimental and analytical approaches have been applied in attempts to derive an objective means of ascribing the infestation source(s) of sea lice impacting wild fish, and these are briefly reviewed here. Irrespective of that debate, a clear understanding of the population dynamics, demography, and sources of infestation of sea lice impacting farmed salmonids has obvious commercial and management significance regarding their interventory control. L. salmonis can cause stress, pathological damage and death of the host fish (Nolan et al., 1999; Wells et al., 2006), and the small post-smolt stage is perhaps the most vulnerable (Finstad et al., 2000). Notwithstanding their impacts 257

SUMBAWS

Final Report

on welfare of farmed salmon, concerns about the possible detrimental impacts of caligids on populations of wild salmonids first arose from observations in Ireland of stock collapses of S. trutta in the late 1980s/early 1990s, and apparently associated epizootics of L. salmonis, particularly on juvenile hosts (Whelan, 1993; Tully et al., 1993). Heavily infested fish often showed “premature” migratory return to freshwater, were in poor condition and bore heavy infestations the sessile larval chalimus stages of the parasite. The evidence is now persuasive that farm-origin caligids do impact wild salmonids, and almost certainly have at least contributed to recent wild salmonid stock declines proximate to salmon farms in areas of Norway, the British Isles, and W Canada (Bjørn and Finstad, 2002, Butler, 2002, McKibben and Hay, 2004, Penston et al., 2004, Todd et al., 2004, Krkošek et al., 2005). That evidence is, nonetheless, indirect and primarily circumstantial or correlative, due to the intractability of physically tracking individuals of the planktonic, larval sea lice stages from initial hatching to ultimate settlement on a host fish.

Here, I focus on studies of L. salmonis and C. elongatus in the North Atlantic, because it is there that the aquaculture industry has been established the longest and the impacts of these sea lice on wild salmonids are the most long-standing and acute. I provide a brief overview of the range of analytical, non-genetic and genetic methods that have been applied in attempts to confirm either a ‘farm’ or ‘wild’ provenance for individual parasites established on particular host fish. I argue that the most appropriate analytical approach to this generic issue involves the deployment of various molecular genetic techniques and my focus, therefore, also is on recent advances in the application of genetic techniques to other general and more specific taxonomic, systematic and ecological problems concerning sea lice.

Non-genetic techniques to discriminate populations of sea lice Phenotypic variation. Fertilization of L. salmonis and C. elongatus is internal and the extruded eggstrings are retained by the female until the hatching, planktonic nauplius I is attained. The nauplius II moults to the infective planktonic copepodid stage and this attaches to the host fish and moults to the first of four sessile chalimus stages. Planktonic development of all these non-feeding larval stages is temperaturedependent and may require 2-9 d (Johnson and Albright, 1991). The chalimus IV gives rise to the first of two mobile pre-adult stages, prior to the definitive moult to the adult male or female. Nordhagen et al. (2000) noted that adult L. salmonis sampled from wild salmon were significantly larger than those sampled from cultured salmon, but that larvae from these two sources developed in the laboratory to similar sizes.

Moreover, larvae developed at 9˚C gave rise to larger adults than did those developed at 12˚C. Previously, Sharp et al. (1994) had reported apparent geographic variations in size of L. salmonis and C. elongatus infesting S. trutta from the east and west coasts of Scotland, with larger parasites found on the east coast. Their analysis did not, however, take account of the size of the host fish. The same extends to an earlier investigation of size variation in L. salmonis infesting a cage population of S. salar in Ireland (Tully, 1989). In the latter study, smolts were placed at sea in May and sampled monthly up to January the following year. Lengths of males and female L. salmonis both were minimal in summer/autumn (JulySeptember), when temperatures were highest, and parasite size increased progressively from September to January as temperature fell. It is likely that temperature does exert a general influence on adult sizes, but host size and species do also have predictable effects on adult size of L. salmonis infesting wild S. salar and S. trutta (A.M. Walker and C.D. Todd, unpubl.): L. salmonis on S. trutta always are smaller than those from S. salar, and larger host fish of either species bear larger adult L. salmonis. Moreover, L. salmonis adults sampled from S. salar always are larger than those sampled from S. trutta of comparable size. As affirmed

by Nordhagen et al. (2000), it is very clear that size of adult L. salmonis is highly plastic and size variation is of no utility in distinguishing a farm or wild origin of the initial infective copepodid (Sharp et al., 1994).

Shinn et al. (2000a) applied conventional nebulisation ICPMS (Inductively Coupled Plasma Mass Spectrometry) to analyze adult female L. salmonis from four salmon farms and from three wild commercial net fisheries in Scotland. Data for as few as 16 elements allowed sea lice from the seven sites to be distinguished by discriminant analysis, but the general applicability of this technique in attributing a farm or wild provenance of individual parasites remains unclear. The elemental signature of the parasite will, for example, depend upon both host-related and extrinsic seasonal and environmental physicochemical factors.

An alternative chemical approach was to analyze tissue pigment complements. The pink coloration of the flesh of wild salmon is largely attributable to the natural carotenoid pigment astaxanthin. In order to achieve a similar coloration of cultured salmon flesh, the feed pellets are treated with synthetic pigments, such as canthaxanthin. Pigment complements of ectoparasitic sea lice potentially provide an analytical tool for distinguishing sea lice ectoparasitic on farmed and wild salmon (Noack et al., 1997). Although they could not unequivocally identify canthaxanthin by HPLC, Noack et al. (1997) reported that canthaxanthin-like peaks were distinguished from astaxanthin and that the ratios of the two components were shown to differ 258

SUMBAWS

Final Report

between adult sea lice taken from wild and farm fish. But such a tool is of no value in determining the provenance or origin of the infective larvae impacting a given wild (or farm) fish. Even assuming that the tissues of nauplius larvae released by farm lice initially carry a disproportionately high level of canthaxanthin, this â&#x20AC;&#x153;farmâ&#x20AC;? pigment marker will be lost as larval development and growth of the sessile and mobile stages proceeds on a host fish. The same argument extends to the use of stable isotopes in discriminating populations of sea lice.

Consumers (predators, parasites) generally show enrichment of the heavier (less labile) isotopes over those in their diet (McCutchan et al., 2003). Enrichment of consumer 13C arises from differential respiratory and excretory loss of assimilated 12C in respiratory CO2, although lipids tend to be depleted in 13C relative to carbohydrates and proteins. Similarly, excreted nitrogen typically is depleted in 15N relative to assimilation.

However, isotopic trophic shift for N can vary amongst species according to whether the primary nitrogenous excretory product is ammonia, urea or uric acid, and depending upon nutritional status (Vanderklift and Ponsard, 2003). Trophic step enrichment is, moreover, sensitive to rates of consumer feeding, post-absorption excretion, defaecation and the extent of isotopic discrimination. For example, Olive et al. (2003) have shown that isotopic ratios can change within a matter of days of starvation, or of switching the consumer to a new food source. Butterworth et al. (2004) applied C and N stable isotope analysis to L. salmonis sampled from farmed S. salar from the Pacific and Atlantic coasts of Canada, and from wild Pacific coho salmon (Oncorhynchus kisutch). Their objective was to validate the stable isotope technique as a means of discriminating between adult parasites from wild and farmed fish. Notwithstanding the foregoing limitations of the stable isotope technique, it is clear that this methodology will never be applicable to studies of the provenance, or infection origin, of sea lice populations infesting (wild or cultured) host fish. As for pigment analyses, even if the stable isotope signatures between adult sea lice (and their larvae) on farmed and wild hosts differ significantly, these differences will be lost as the larva grows and develops to the feeding postlarval and adult stages on the host fish.

Genetic studies of caligid sea lice Given our inability to directly track individual larvae from hatching to host encounter and colonization, genetic approaches would appear to offer the most effective indirect means of ascribing a wild or farm source of sea lice infecting given fish. Low levels of gene flow (here, larval cross-infection) between two populations can result in those populations becoming differentiated as a result of random genetic drift.

Conversely, selection at farm sites, perhaps as a result of exposure of sea lice to chemotherapeutants, might also lead to population divergence. If there is sufficient genetic differentiation of populations, the rates of gene flow can be estimated (Wright 1969) and individuals can be objectively allocated to populations by any one of several assignment methods (Hansen et al., 2001). High levels of gene flow will, however, maintain genetic homogeneity of populations and this will preclude both assignment of individuals and the quantification of gene flow amongst farm and wild populations.

Mitochondrial sequence analysis. Because it provides a large character base, and owing to its high rate of evolution compared to nuclear DNA (Brown et al., 1979), analysis of mtDNA often has been successfully used in generating tractable hypotheses and robust predictions of the phylogeny of species. The molecular approach offers indispensable tools for taxonomists and ecologists alike, especially in distinguishing closelyrelated, cryptic or sibling species. The mtDNA molecule is double-stranded, circular and typically encodes 37 genes (two ribosomal RNA, 22 transfer RNA and 13 protein-coding genes); whilst the number and identity of the genes comprising mtDNA is highly conserved, their order is not, at least in invertebrates. The non-coding control region (Displacement loop) is often used in taxonomic and phylogenetic studies, but analytical outcomes can be markedly influenced by the kinds of markers (e.g. RFLPs) investigated, and when there are high levels of DNA polymorphism (Grant et al., 1998).

At present there are 31 complete crustacean mtDNA sequences available (Genbank), of which 16 are for malacostracans and only two, Tigriopus japonicus (Machida et al., 2002) and L. salmonis (Tjensvoll et al., 2005), are for copepods. The mtDNA sequence of T. japonicus is 14,628 bp long, but none of the other reported gene orders are comparable to T. japonicus. Also, unlike most metazoa, T. japonicus is exceptional in having all the genes on the one strand. Roehrdantz et al. (2002), for example, investigated arthropod phylogeny by analysing gross gene rearrangements of mtDNA. Their data included two malacostracans and a single isopod and pointed to the crustaceans as being a sister lineage to insects; whether the inclusion of the unique gene arrangements for L. salmonis, or for T. japonicus, will further reinforce this pattern is unclear.

Other unusual features of the T. japonicus mtDNA genome include the size reductions in the RNA genes, resulting in this being one of the smallest mtDNA genomes in the arthropod lineage.

259

SUMBAWS

Final Report

The mitochondrial cytochrome oxidase I (mtCOI) gene evolves particularly rapidly and sequence comparisons of this gene have often been used in distinguishing closely related taxa. For example, Bucklin et al. (2003) found that analysis of a 639 bp region of mtCOI allowed the reliable distinction of 34 calanoid copepod species embracing ten genera and two families. With specific reference to the ectoparasitic caligid C. elongatus, Øines and Heuch (2005) exploited the variability of mtCOI in separating two distinct haplotypes. Whether or not these are separate species remains to be confirmed, but morphologically they are virtually indistinguishable and the two are intermixed on various host fish species. Certainly, from both ecological and economic standpoints, because C. elongatus is abundant on both wild (Todd et al., 2006) and farmed salmonids (McKenzie et al., 2004), it will be essential to ascertain if these different haplotypes/species are associated with particular host fish species.

Shinn et al. (2000b) investigated partial sequences of the 18S rRNA gene and the ribosomal internal transcribed spacer (ITS-1) region obtained from mtDNA for individual L. salmonis. Their objective was to assess the utility of mtDNA sequence data in perhaps distinguishing parasites from wild and farmed Atlantic salmon. They sampled L. salmonis from eight localities (three salmon farms and five wild commercial net fisheries) in Scotland. Eight individual L. salmonis (two each from two farms and two wild sites) were sequenced for a 900 nucleotide fragment of the 18S rRNA gene. That gene, however, is highly conserved and showed no sequence variation. Thirteen individual L. salmonis, again from two farms and two wild sites, were sequenced for a 454 nucleotide fragment of ITS-1. Overall there was 92% similarity of sequences, with variation concentrated in one hypervariable region. They reported greater sequence similarity for ITS-1 within rather than between farm and wild lice, and proposed this technique to provide a means of assigning lice to either farm or wild populations.

Furthermore, on the basis of this result they suggested that the causal process may be attributed to restricted gene flow within lice populations at S. salar cage sites. But this is counter-intuitive, given the obligatory 2-9 d duration of the free-swimming planktonic larval stages, during which period the nauplius I and II stages are incapable of colonizing host fish. Planktonic export (and hence gene flow from farm to farm and between farm and wild host populations) of nauplius I larvae from cage sites is inevitable: there is no possibility of developing nauplii being retained within cages for the required time period. Re-importation of fully-developed copepodid larvae, and hence ultimate “self-reinfestation” of cage sites in semi-enclosed fjords is, however, clearly possible; but the hydrodynamic regime at any farm site will ensure initial dispersal of planktonic larval stages away from the source host population and thereby the potential for infestation of other farms and/or wild salmonids. The scale of realised larval dispersal in L. salmonis and C. elongatus, and thence gene flow and levels of genetic differentiation of populations, is therefore dependent both upon passive planktonic dispersal of the parasite larval stages and the migratory ranges of the various wild host species. Whereas S. trutta generally remain in coastal waters and close to their natal river, wild S. salar may migrate thousands of km across the North Atlantic (Klemetsen et al., 2003).

Allozyme (protein) variation. In explicitly addressing the issue of scale of population genetic differentiation of L. salmonis, earlier studies of allozyme variation of adult sea lice sampled from farm and wild populations of Atlantic salmon had reported rather conflicting results. Todd et al. (1997) screened two polymorphic allozyme loci (Fum, Got-2) for 403 adult L. salmonis sampled from three W Scotland farms and from wild S. salar and wild S. trutta in two river estuaries in E Scotland. Analysis of Wright’s Fstatistics confirmed no significant genetic differentiation between east and west coasts (geographic range ~770 km), or between farmed versus wild host fish. Neither was there significant genetic differentiation of male and female L. salmonis. By contrast, Isdal et al. (1997) screened four allozyme loci (Est, Pgi, Idh, Pgm) for L. salmonis sampled at six sites spanning a north-south geographic range in Norway of ~1100 km. Their data showed some evidence of differentiation into two (‘northern’ and ‘southern’) geographic groupings, but differentiation was attributable to allele frequencies at only the one locus.

Although allozyme electrophoresis offers a reliable, well-established and relatively cheap means of obtaining genetic data for individuals, and thence measures of population homogeneity/heterogeneity, the overall levels of polymorphism of allozyme loci typically are very low in contrast to those that can be revealed by modern molecular DNA techniques, especially since the advent of PCR methods. An early application of PCR technology to population analysis was the RAPD (randomly amplified polymorphic DNA) technique, which uses anonymous PCR primers. Todd et al. (1997) screened samples of L. salmonis from host fish taken around the west, north and east coasts of Scotland with six primers. Hosts included wild and farmed S. salar, farmed rainbow trout (Oncorhynchus mykiss Walbaum) and wild S. trutta. Analysis of molecular variance (AMOVA) indicated significant differentiation between ‘farm’ and ‘wild’ L. salmonis, and amongst the various farms sampled. Moreover, certain RAPD fragments appeared to be putative 260

SUMBAWS

Final Report

â&#x20AC;&#x2DC;markersâ&#x20AC;&#x2122; exclusive to L. salmonis sampled from farm fish, and two bands for one primer were apparently sex-linked and diagnostic of female L. salmonis.

The identification of genetic markers that might permit unequivocal assignment of individual sea lice to, for example, a farm or wild origin would be of value to the salmon farming industry and wild fishery managers alike: but the weaknesses of the RAPD technique are considerable (Grosberg et al., 1996). The observer has no specific information on the fragments of DNA being amplified, and amplifications appear to be very dependent upon the precise PCR conditions. Moreover, RAPDs are dominant markers and the allelic basis of fragment variation cannot be inferred with confidence. A subsequent study (Dixon et al., 2004) of L. salmonis from around Scotland, using the same RAPD primers, indicated no systematic geographic patterns of differentiation or distinction between L. salmonis of farmed or wild origin. RAPD markers did not, therefore, provide a sufficiently informative or reliable analytical tool to address questions relating to the genetic structuring of copepod sea lice populations, or of farm versus wild provenance of parasites infesting given host fish. Although time-consuming and expensive to develop anew for a given species, microsatellite (VNTR) DNA loci currently offer the best opportunity of addressing such questions by providing unequivocal and highly polymorphic allelic data for neutral loci. For highly structured and differentiated populations, perhaps showing private alleles, microsatellite data offer the possibility of allowing assignment of individuals to populations (or, in the present context, a farm or wild provenance).

Microsatellite DNA variation. Microsatellites are rapidly-evolving, selectively neutral loci in noncoding regions of DNA that include short tandem nucleotide repeats. The quantification of allelic variation in microsatellite repeat length for multiple loci provides a high resolution database for the analysis of population structuring and individual assignment (Strassman et al., 1996). For L. salmonis, Genbank presently includes ten unpublished and nine published microsatellite sequences (Nolan et al., 2000; Todd et al., 2004). No loci are available for other caligid copepods. In their preliminary analysis of small samples of L. salmonis from Ireland, Scotland and Norway, Nolan et al. (2000) confirmed the potential for microsatellites to allow the detection of population differentiation in caligids.

Todd et al. (2004) screened 1007 Atlantic (n = 973) and Pacific (n = 34) L. salmonis for six microsatellite loci. Their Atlantic samples centred on three host species across multiple farm and wild sites around Scotland (n = 856 L. salmonis), but included material from wild sea trout in N Norway (n = 58) and farmed Atlantic salmon in E Canada (n = 59). The small outgroup Pacific sample was taken from farmed S. salar in British Columbia. The Norwegian and eastern Canadian samples were analytically especially important in the geographic context because the infestations carried by those fish would have been acquired locally; by contrast, the remaining wild salmon sampled from Scotland will have ranged migratorily to both the Norwegian Sea and the coasts of Greenland. There they will have intermixed and cross-infected with fish of Russian, European, Icelandic and North American origin (Jacobsen and Gaard, 1997; Todd et al., 2000).

There were no analytical problems from null alleles and, for the North Atlantic samples, F-statistic analyses showed no significant differentiation amongst host species or between farm and wild L. salmonis.

There was no evidence of isolation by geographic distance over the sampled Atlantic range, and this conformed to expectation. The levels of gene flow (larval cross-infection) between farm and wild salmonids, and between the two primary wild species in the North Atlantic (S. salar, S. trutta), are sufficiently high as to prevent genetic divergence of populations over a 6000 km range. As expected, however, there was highly significant differentiation between North Atlantic and North Pacific L. salmonis, presumably attributable to genetic drift arising from the isolation of North Pacific and North Atlantic L. salmonis.

Although these analyses of microsatellite differentiation provide no direct evidence of infection of wild fish by farmed fish, they do conclusively demonstrate that farm sea lice populations are not genetically isolated and that L. salmonis throughout the North Atlantic comprises a single, panmictic population. Given the numerical imbalance between the numbers of wild and farmed salmonids in the North Atlantic, and the fact that cultured salmon are present in coastal waters year-round, the likelihood is that the bulk of the infective interaction concerning wild S. salar and S. trutta is from farm to wild, rather than wild to farm (Butler, 2002; Todd et al., 2004). It has to be emphasized that microsatellites are neutral non-coding loci â&#x20AC;&#x201C; they will not reveal selection for mutations that may, for example, confer resistance to chemotherapeutants.

Paternity analysis. In view of its economic and ecological importance, the life cycle, population dynamics and epidemiology of L. salmonis have been intensively investigated (Pike and Wadsworth, 1999; Revie et al., 2002, 2003, 2005) and several studies had indicated that female L. salmonis are monogamous (Hull, 1998; Ritchie et al., 1996a,b; Pike and Wadsworth, 1999). Adult males actively seek out and clasp onto a pre-adult II female and the male remains attached in a characteristic mate-guarding position (Hull, 1998). Once the female has undergone the definitive adult moult the male transfers and cements a pair of 261

SUMBAWS

Final Report

spermatophores on her genital complex. Tubules extend from the spermatophores and enter the genital pores to allow transfer of the spermatozoa and their internal storage in the receptaculum seminis. The combination of active (pre-copulatory) and remote (post-copulatory) male mate-guarding by the cementing of durable, persistent spermatophores, the tubules of which occlude the genital pores on the female’s genital complex, all pointed to effective monogamy of L. salmonis (Pike and Wadsworth, 1999).

Observations of females from wild hosts evidently having lost their spermatophores, and rare occurrences of female L. salmonis bearing two pairs of cemented spermatophores, raised questions as to whether L. salmonis might show multiple paternity of their offspring, or if the presence of spermatophores did indeed effectively block further (P2) fertilization by a subsequent male (Todd et al., 2005). Whilst the analysis of microsatellite variation has tended to focus on studies of population structure, differentiation and of gene flow, microsatellites do remain the tool of choice for parentage analysis (Fleischer, 1996; Strassman et al., 1996). Parentage of offspring clutches for 10 females sampled from wild S. salar was therefore determined by means of dual-locus microsatellite typing of the parent female and individual embryos within the eggstrings (Todd et al., 2005). Seven of the females were specifically chosen for analysis because they carried multiple spermatophores and evidently had been multiply mated. Three other females, carrying the typical pair of spermatophores, were also chosen randomly from the same fish as three of the multiply-mated female L. salmonis. Four of the seven multiply-mated females showed single paternity and confirmed the effectiveness of the first (P1) male’s spermatophores in preventing entry and storage of other spermatozoa.

The remaining three multiply-mated females and two of the three singly-mated females did, however, show dual paternity of their offspring. The cementing of spermatophores may therefore hinder copulatory success of P2 males in the short term, but the likelihood is that only females that do not survive beyond the production of one or a few pairs of eggstrings will be monogamous. Indeed, if the 2/3 females with the typical pair of spermatophores showing dual paternity are representative, the likelihood is that most females on wild S. salar will be polyandrous during their ovigerous lifetime of several months.

Nonetheless, because of regular, interventory, chemotherapeutant treatment of sea lice infestations of cultured salmon, there must be a strong likelihood of farm females not surviving sufficiently long to display polyandry. Accordingly, seven adult female L. salmonis bearing the typical pair of spermatophores were randomly sampled from harvested salmon at a W Scotland farm in March 2005. The adult females and 6-20 embryos from one of the pair of eggstrings for each female were genotyped for one of the microsatellite loci (LsalSTA5) used in the previous paternity study. The offspring of two of the seven farm females (Numbers 3 and 7, Fig. 1) showed dual paternity. In order to confidently determine the proportion of either wild or farm females that show multiple paternity, screening of larger numbers of females for two or more loci would be necessary. But qualitatively, the important outcome is that both farm and wild L. salmonis do show polyandry, despite pre- and post-copulatory mate-guarding, and clearly not all females are monogamous. As discussed by Todd et al. (2005), this finding has implications for the continued development of sea lice management and control strategies in the aquaculture industry.

Figure 1 Lepeophtheirus salmonis. Frequencies of paternal microsatellite (LsalSTA5) alleles within seven families of offspring. The numbers of embryos typed for each family are given above the bars. Single paternity – Families 1, 2, 4-6; Dual paternity – Families 3 and 7.

262

SUMBAWS

Final Report

Molecular studies of chemical resistance. Many insect populations have developed genetic resistance to pyrethroid insecticides (Zhao et al., 2000) and to organosphosphates and carbamates (Walsh et al., 2001); similar responses by sea lice on salmon farms were therefore to be expected. Pyrethroid resistance involves genetic mutations of neuronal sodium channel proteins (Soderlund and Knipple, 2003), whereas organophosphate and carbamate resistance is conferred by mutations of acetylcholinesterase (AChE) genes (Walsh et al., 2001). The organophosphates dichlorvos and azamethiphos were widely used as louse chemotherapeutants from early in the development of salmon aquaculture and soon there were reports of reduced sensitivity of L. salmonis to dichlorvos in Scotland (Jones et al., 1992). Similarly, Fallang et al.

(2004) reported reduced sensitivity of L. salmonis in Norway and Canada to azamethiphos. Biochemical assays showed differential sensitivities of two AChE enzymes – one was rapidly inactivated, and the other slowly inactivated – but the genetic basis of this remains unclear.

“Knockdown resistance” (kdr) differs from resistance by metabolic detoxification in arising from a reduction in the sensitivity of the nervous system to pyrethroids (Soderlund and Knipple, 2003). The pyrethroids deltamethrin and cypermethrin have been widely used by the aquaculture industry in the control of sea lice, and anecdotal reports of their reduced efficacy have ranged from Ireland and Scotland to Norway (Sevatdal et al., 2005). Mutations of the para-type voltage-sensitive sodium channel gene appear to specifically confer pyrethroid resistance and Fallang et al. (2005) have sequenced a 318 bp fragment of that gene in L. salmonis. Four of the six previously reported kdr primary resistance mutations in arthropods are located in the second of four specific regions of the protein (Soderlund and Knipple, 2003), and the mutation reported by Fallang et al. (2005) also was located in domain II. Although Fallang et al. (2005) have presently only correlated the occurrence of their mutation with pyrethroid resistance, its location in domain II does strongly suggest a causal relationship.

Conclusion An early resolution of the controversy surrounding the caligid infestation interactions between farmed and wild salmonids would have been readily achievable if it were possible to unequivocally determine, by direct observation, the source of infective copepodids of L. salmonis parasitizing wild host fish. With the rare exception of (visibly) large and very short-lived marine planktonic larvae – such as the tadpole larvae of ascidians, which can be followed from release to settlement by SCUBA divers (Davis and Butler, 1989) – it is impractical to physically track the planktonic stages of marine invertebrate larvae in the water column (Todd, 1998). The challenge has therefore been one of developing indirect methods to confirm the connection between larval sea lice production from aquaculture sites and infestations of free-ranging wild salmonids. The most persuasive empirical field data are from Loch Torridon, W Scotland.

Plankton surveys of Loch Shieldaig, within Loch Torridon, showed considerable spatial and temporal heterogeneity of the distribution of caligid nauplius and copepodid larvae (McKibben and Hay, 2004; Penston et al., 2004). Larvae could be captured consistently only close to the outfalls of local rivers and in such locations wild juvenile S. trutta are therefore especially vulnerable to infestation. S. salar are cultured on a two-year production cycle and L. salmonis infestations are especially problematic on farms during the second year (Revie et al., 2005). Over a five-year period, McKibben and Hay (2004) recorded planktonic larvae in their samples only when there were gravid females present on local farms, and always during the second year of the production cycle.

Despite the range of experimental and analytical approaches applied, it is clear that population genetic techniques are the only plausible means of addressing general and broad scale questions relating to the management of wild-farm interactions and copepod infestations of wild salmonids. Molecular DNA techniques have confirmed that farm populations of L. salmonis are not genetically isolated, and the foregoing plankton data should urge wild and farmed fishery managers to be resolute in adopting a precautionary approach to this problem. The very real concern, however, must lie in the paucity of the range of chemotherapeutants available to the aquaculture industry, and the commercial expense and lead time for developing new active compounds. Presently, there is a widespread and very heavy reliance by the industry on the one major compound, emamectin benzoate (SLICE™, Schering-Plough).

Populations of North Atlantic L. salmonis have been shown to be genetically and demographically open.

These features alone dictate that eradication of L. salmonis as a pest species on salmon farms is impossible.

But there are consequences also for the possible evolution of resistance of L. salmonis to chemical treatments. One of the tenets of pest management and control is that in the absence of pesticides resistant genotypes are selected against, or resistant individuals are less ‘fit’ (Devonshire et al., 1998). Given the demographic openness of L. salmonis populations on farms, and high levels of gene flow, it is important to understand how resistance might develop on farms because resistance genes ought to be selected against on 263

SUMBAWS

Final Report

wild fish and outwith the farm environment. Nonetheless, in view of the possibility of resistance developing, the challenge to the aquaculture industry is to deploy the few currently available and licenced treatments at minimum dosages and perhaps to engage in rotation of treatment use.

The fish husbandry and economic problems caused by the monogenean skin fluke, Benedenia seriolae, to the emerging kingfish aquaculture industry in Australia (Chambers and Ernst, 2005) are timely reminders that metazoan parasites other than copepods can present major environmental, economic and animal welfare challenges to intensive marine aquaculture industries. The past focus of research into copepod sea lice biology in the North Atlantic has necessarily been on the salmon louse, L. salmonis, but the importance of the host-generalist, C. elongatus, to both farmed and wild salmon and other teleost fish in the North Atlantic, should not be under-estimated. For example, for wild 0-group cod (Gadus morhua) and haddock (Melanogrammus aeglefinus) of only 13-63 mm length, Nielsen et al. (1987) recorded intensities of C. elongatus at up to 35 and 15 chalimi respectively per fish in the NW Atlantic. On the basis of empirical physiological data, on the stress impacts of L. salmonis on 19-70 g post-smolt S. trutta (Wells et al., 2006), such intensities of infestation of wild gadids will almost certainly be lethal. In the farm environment, high intensity C. elongatus infestations, and consequentially severe head lesions, were reported for juvenile farmed halibut (Hippoglossus hippoglossus) early in the development of what was then a new aquaculture species (Bergh et al., 2001). Cod farming presently is expanding rapidly in the NE Atlantic. The wealth of experience and knowledge that has been garnered in relation to L. salmonis should be readily transferable to the possibly detrimental impacts on wild gadid and salmonid stocks that C. elongatus may well exert as cod farming develops.

Acknowledgements This work was partly funded by the European Commission (SUMBAWS, Contract No. Q5RS-2002-00730).

For assistance with the paternity analyses I am grateful to N. Gagne and M. Feilen. J. Graves provided insightful comments on the manuscript.

References Bergh, Ø., Nilsen, F. and Samuelson, O. B. (2001) Diseases, prophylaxis and treatment of the Atlantic halibut Hippoglossus hippoglossus: a review. Dis. Aqua. Org., 48, 57–74.

Bjørn, P. A. and Finstad, B. (2002) Salmon lice, Lepeophtheirus salmonis (Krøyer), infestation in sympatric populations of Arctic char, Salvelinus alpinus (L.), and sea trout, Salmo trutta (L.), in areas near and distant from salmon farms. ICES J. Mar. Sci., 59, 1–9.

Brown, W. M., George, M. and Wilson, A. C. (1979) Rapid evolution of animal mitochondial DNA. Proc.

Natl. Acad. Sci. USA, 76, 1967–1971.

Bucklin, A., Frost, B. W., Bradford-Grieve, J., Allen, L. D. and Copley, N. J. (2003) Molecular systematic and phylogenetic assessment of 34 calanoid copepod species of the Calanidae and Clausocalanidae. Mar. Biol., 142, 333–343.

Butler, J. R. A. (2002) Wild salmonids and sea louse infestations on the west coast of Scotland: sources of infection and implications for the management of marine salmon farms. Pest Manage. Sci., 58, 595–608.

Butterworth, K. G., Li, W. and McKinley, R. S. (2004) Carbon and nitrogen stable isotopes: a tool to differentiate between Lepeophtheirus salmonis and different salmonid host species? Aquaculture, 241, 529– 538.

Chambers, C. B. and Ernst, I. (2005) Dispersal of the skin fluke Benedenia seriolae (Monogenea: Capsalidae) by tidal currents and implications for sea-cage farming of Seriola spp. Aquaculture, 250, 60–69.

Davis, A. R and Butler, A. J. (1989) Direct observations of larval dispersal in the colonial ascidian Podoclavella moluccensis Sluiter: evidence for closed populations. J. Exp. Mar. Biol. Ecol., 127: 189–203.

Devonshire, A. L., Field, L. M., Foster, S. P., Moores, G. D., Williamson, M. S. and Blackman, R. L. (1998) The evolution of insecticide resistance in the peach-potato aphid, Myzus persicae. Phil. Trans. R. Soc. Lond. B, 353, 1677–1684.

264

SUMBAWS

Final Report

Dixon, B. A., Shinn, A. P. and Sommerville, C. (2004) Genetic characterization of populations of the ectoparasitic caligid, Lepeophtheirus salmonis (Krøyer 1837) using randomly amplified polymorphic DNA. Aquacult. Res., 35, 730–741.

Fallang, A., Ramsay, J. M., Sevatdal, S., Burka, J. F., Jewess, P., Hammell, K. L. and Horsberg, T. E. (2004) Evidence for occurrence of an organophosphate-resistant type of acetylcholinesterase in strains of sea lice (Lepeophtheirus salmonis Krøyer). Pest Manage. Sci., 60, 1163–1170.

Fallang, A., Denholm, I., Horsberg, T. E. and Williamson, M. S. (2005) Novel point mutation in the sodium channel gene of pyrethroid-resistant sea lice Lepeophtheirus salmonis (Crustacea : Copepoda). Dis. Aqua. Org., 65, 129–136.

Finstad, B., Bjørn P. A., Grimnes A. and Hvidsten, N. A. (2000) Laboratory and field investigations of salmon lice [Lepeophtheirus salmonis (Krøyer)] infestation on Atlantic salmon (Salmo salar L.) post-smolts.

Aquac. Res., 31, 795–803.

Fleischer, R. C. (1996) Application of molecular methods to the assessment of genetic mating systems in vertebrates. In Ferraris, J. D. and Palumbi, S. R. (eds), Molecular zoology. Advances, strategies, and protocols. Wiley-Liss, New York, pp. 133–161.

Grant, W. S., Clark A. -M. and Bowen, B. W. (1998) Why restriction fragment length polymorphism analysis of mitochondrial DNA failed to resolve sardine (Sardinops) biogeography: insights from mitochondrial DNA cytochrome b sequences. Can. J. Fish. Aquat. Sci., 55, 2539–2547.

Grosberg, R. K., Levitan, D. R. and Cameron, B. A. (1996) Characterization of genetic structure and genealogies using RAPD-PCR markers: a random primer for the novice and nervous. In Ferraris, J. D. and Palumbi, S. R. (eds), Molecular zoology. Advances, strategies, and protocols. Wiley-Liss, New York, pp.

67–100.

Hansen, M. M., Kenchington, E. and Nielsen, E. E. (2001) Assigning individual fish to populations using microsatellite DNA markers. Fish and Fish., 2, 93–112.

Hull, M. Q., Pike, A. W., Mordue, A. J. and Rae, G. H. (1998) Patterns of pair formation and mating in an ectoparasitic caligid copepod Lepeophtheirus salmonis (Krøyer 1837): implications for its sensory and mating biology. Phil. Trans. R. Soc. Lond. B, 353, 753–764.

Huys, R. and Boxshall, G. A. (1991) Copepod evolution. Ray Society, London.

ICES (2004) Report of the Working Group on North Atlantic Salmon. ICES CM 2004/ACFM:20, Ref. I Advisory Committee on Fishery Management. ICES, Copenhagen, pp. 1–286.

Isdal, E., Nylund A. and Nævdal, G. (1997) Genetic differences among salmon lice (Lepeophtheirus salmonis) from six Norwegian coastal sites: evidence from allozymes. Bull. Eur. Ass. Fish Pathol., 17, 17– 22.

Jacobsen, J. A. and Gaard, E. (1997) Open-ocean infestation by salmon lice (Lepeophtheirus salmonis): comparison of wild and escaped farmed Atlantic salmon (Salmo salar L.). ICES J. Mar. Sci., 54, 1113–1119.

Johnson, S. C. and Albright, L. J. (1991) Development, growth, and survival of Lepeophtheirus salmonis (Copepoda, Caligidae) under laboratory conditions. J. Mar. Biol. Assn, UK, 71, 425–436.

Johnson, S. C., Treasurer, J. W., Bravo, S., Nagasawa, K. and Kabata, Z. (2004) A review of the impact of parasitic copepods on marine aquaculture. Zool. Stud., 43, 229–243.

Jones, M. W., Sommerville, C. and Wootten, R. (1992) Reduced sensitivity of the salmon louse, Lepeophtheirus salmonis, to the organophosphate dichlorvos. J Fish Dis., 15, 197–202.

Kabata, Z. (1979) Parasitic Copepoda of British fishes. Ray Society, London.

Klemetsen, A., Amundsen, P. -A., Dempson, J. B., Jonsson, B., Jonsson, N., O’Connell, M. F. and Mortensen, E. (2003) Atlantic salmon Salmo salar L., brown trout Salmo trutta L., and Arctic charr Salvelinus alpinus (L.): a review of aspects of their life histories. Ecol. Freshw. Fish., 12, 1–59.

Krkošek, M., Lewis, M. A. and Volpe, J. P. (2005) Transmission dynamics of parasitic sea lice from farm to wild salmon. Proc. R. Soc. Lond. Ser. B., 272, 689–696.

265

SUMBAWS

Final Report

Machida, R. J., Miya, M. U., Nishida, M. and Nishida, S. (2002) Complete mitochondrial DNA sequence of Tigriopus japonicus (Crustacea : Copepoda). Mar. Biotech., 4, 406–417.

McCutchan, J. H., Lewis, W. M., Kendall, C. and McGrath, C. C (2003) Variation in trophic shift for stable isotope ratios of carbon, nitrogen, and sulfur. Oikos, 102, 378–390.

McKenzie, E., Gettinby, G., McCart, K. and Revie, C. M. (2004) Time-series models of sea lice Caligus elongatus (Nordmann) abundance on Atlantic salmon Salmo salar L. in Loch Sunart, Scotland. Aquac. Res., 35, 764–772.

McKibben, M. A. and Hay, D. (2004) Distributions of planktonic lice larvae Lepeophtheirus salmonis in the inter-tidal zone in Loch Torridon, Western Scotland in relation to salmon farm production cycles. Aquac.

Res., 35, 742–750.

McVicar, A. H., Sharp, L. A., Walker, A. F. and Pike, A. W. (1993) Diseases of wild sea trout in Scotland in relation to fish population decline. Fish. Res., 17, 175–185.

Nagasawa, K. (2001) Annual changes in the population size of the salmon louse Lepeophtheirus salmonis (Copepoda: Caligidae) on high-seas Pacific Salmon (Oncorhynchus spp.), and relationship to host abundance. Hydrobiologia, 453/454, 411–416.

Nielsen, J. D., Perry, R. I., Scott, J. S. and Valerio, P. (1987) Interactions of caligid ectoparasites and juvenile gadids on Georges Bank. Mar. Ecol. Prog. Ser., 39, 221–232.

Noack, P. T., Laird, L. M. and Priede, I.G. (1997) Carotenoids of sea lice (Lepeophtheirus salmonis) as potential indicators of host Atlantic salmon (Salmo salar L.) origin. ICES J. Mar. Sci., 54, 1140–1143.

Nolan, D. T., Reilly, P. and Wendelaar Bonga, S. E. (1999) Infection with low numbers of the sea louse Lepeophtheirus salmonis induces stress-related effects in postsmolt Atlantic salmon (Salmo salar). Can. J. Fish. Aquat. Sci., 56, 947–959.

Nolan, D. V., Martin, S. A. M., Kelly, Y., Glennon, K., Palmer, R., Smith, T., McCormack, G. P. and Powell, R. (2000) Development of microsatellite PCR typing methodology for the sea louse, Lepeophtheirus salmonis (Krøyer). Aquac. Res., 31, 815–822.

Nordhagen, J. R., Heuch, P. A. and Schram, T.A. (2000) Size as indicator of origin of salmon lice Lepeophtheirus salmonis (Copepoda: Caligidae). Contrib. Zool., 69, 99–108. Øines, Ø and Heuch, P. A. (2005) Identification of sea louse species of the genus Caligus using mtDNA. J. Mar. Biol. Assn. UK, 85, 73–79.

Olive, P. J. W., Pinnegar, J. K., Polunin, N. V. C., Richards, G. and Welch, R. (2003) Isotope trophic step fractionation: a dynamic equilibrium model. J. Anim. Ecol., 72, 608–617.

Penston, M. J., McKibben, M. A., Hay, D. W. and Gillibrand, P. A. (2004) Observations on open-water densities of sea lice larvae in Loch Shieldaig, Western Scotland. Aquac. Res., 35, 793–805.

Pike, A. W. and Wadsworth S. L. (1999) Sealice on salmonids: their biology and control. Adv. Parasitol., 44, 233–337.

Rae, G. H. (2002) Sea louse control in Scotland, past and present. Pest Manage. Sci., 58, 515–520.

Revie, C. W., Gettinby, G., Treasurer, J. W., Rae, G. H. and Clark, N. (2002) Temporal, environmental and management factors influencing the epidemiological patterns of sea lice (Lepeophtheirus salmonis) infestations on farmed Atlantic salmon (Salmo salar) in Scotland. Pest Manage. Sci., 58, 576–584.

Revie, C. W., Gettinby, G., Treasurer, J. W. and Wallace, C. (2003) Identifying epidemiological factors affecting sea lice Lepeophtheirus salmonis abundance on Scottish salmon farms using general linear models.

Dis. Aqua. Org., 57, 85–95.

Revie, C. W., Robbins, C., Gettinby, G., Kelly, L. and Treasurer, J. W. (2005) A mathematical model of the growth of sea lice, Lepeophtheirus salmonis, populations on farmed Atlantic salmon, Salmo salar L., in Scotland and its use in the assessment of treatment strategies. J. Fish. Dis., 28, 603–613.

Ritchie, G., Mordue, A. J., Pike, A. W. and Rae, G.H. (1996a) Observations on mating and reproductive behaviour of Lepeophtheirus salmonis, Krøyer (Copepoda: Caligidae). J. Exp. Mar. Biol. Ecol., 210, 285– 298.

266

SUMBAWS

Final Report

Ritchie, G., Mordue, A. J., Pike, A. W. and Rae, G. H. (1996b) Morphology and ultrastructure of the reproductive system of Lepeophtheirus salmonis (Krøyer, 1837) (Copepoda: Caligidae). J. Crust. Biol., 16, 330–346.

Roehrdantz, R. L., Degrugillier, M. E. and Black, C. W. (2002) Novel rearrangements of arthropod mitochondrial DNA detected with long-PCR: applications to arthropod phylogeny and diversity. Mol. Biol.

Evol., 19, 841–849.

Rohde, K. (1993) Ecology of marine parasites. CAB International, Wallingford.

Schram, T. A., Knutsen, J. A., Heuch, P. A. and Mo, T. A. (1998) Seasonal occurrence of Lepeophtheirus salmonis and Caligus elongatus (Copepoda: Caligidae) on sea trout (Salmo trutta), off southern Norway.

ICES J. Mar. Sci., 55, 163–175.

Sevatdal, S., Copley, L., Wallace, C., Jackson, D. and Horsberg, T. E. (2005) Monitoring of the sensitivity of sea lice (Lepeophtheirus salmonis) to pyrethroids in Norway, Ireland and Scotland using bioassays and probit modelling. Aquaculture, 244, 19-27.

Sharp, L., Pike, A. W. and McVicar, A. H. (1994) Parameters of infection and morphometric analysis of sea lice from sea trout (Salmo trutta L.) in Scottish waters. In Pike, A. W. and Lewis, J. W. (eds), Parasitic diseases of fish. Samara, Dyfed, Wales, pp. 151–170.

Shinn, A. P., Bron, J. E., Gray, D. J. and Sommerville, C. (2000a) Elemental analysis of Scottish populations of the ectoparasitic copepod Lepeophtheirus salmonis. Contrib. Zool., 69, 79–87.

Shinn, A. P., Banks B. A., Tange N., Bron J. E., Sommerville C., Aoki T. and Wootten, R. (2000b) Utility of 18S rDNA and ITS sequences as population markers for Lepeophtheirus salmonis (Copepoda : Caligidae) parasitising Atlantic salmon (Salmo salar) in Scotland. Contrib. Zool., 69, 89–98.

Soderlund, D. M. and Knipple, D. C. (2003) The molecular biology of knockdown resistance to pyrethroid insecticides. Insect Biochem. Mol. Biol., 33, 563–577.

Strassman, J. E., Solís, C. R., Peters, J. M. and Queller, D. C. (1996) Strategies for finding highly polymorphic DNA microsatellite loci for studies of genetic relatedness and pedigrees. In Ferraris, J. D. and Palumbi, S. R. (eds), Molecular zoology. Advances, strategies, and protocols. Wiley-Liss, New York, pp.

163–180.

Tjensvoll, K., Hodneland, K., Nilsen, F. and Nylund, A. (2005) Genetic characterization of the mitochondrial DNA from Lepeophtheirus salmonis (Crustacea: Copepoda). A new gene organization revealed. Gene, 353, 218–230.

Todd, C. D., Walker, A. M., Wolff, K., Northcott, S. J., Walker, A. F., Ritchie, M. G., Hoskins, R., Abbott, R. J. and Hazon, N. (1997) Genetic differentiation of populations of the copepod sea louse Lepeophtheirus salmonis (Krøyer) ectoparasitic on wild and farmed salmonids around the coasts of Scotland: evidence from RAPD markers. J. Exp. Mar. Biol. Ecol., 210, 251–274.

Todd, C.D. (1998) Larval supply and recruitment of benthic invertebrates: do larvae always disperse as much as we believe? Hydrobiologia, 375/376, 1–21.

Todd, C. D., Walker, A. M., Hoyle, J. E., Northcott, S. J., Walker, A. F and Ritchie, M. G. (2000) Infestations of wild adult Atlantic salmon (Salmo salar L.) by the ectoparasitic copepod sea louse Lepeophtheirus salmonis Krøyer: prevalence, intensity and the spatial distribution of males and females on the host fish. Hydrobiologia, 429, 181–196.

Todd, C. D., Walker, A. M., Ritchie, M.G., Graves, J. A. and Walker, A. F. (2004) Population genetic differentiation of sea lice (Lepeophtheirus salmonis) parasitic on Atlantic and Pacific salmonids: analyses of microsatellite DNA variation among wild and farmed hosts. Can. J. Fish. Aquat. Sci., 61, 1176–1190.

Todd, C. D., Stevenson, R. J., Reinardy, H. and Ritchie, M.G. (2005) Polyandry in the ectoparasitic copepod Lepeophtheirus salmonis despite complex precopulatory and postcopulatory mate-guarding. Mar. Ecol. Prog. Ser., 303, 225–234.

Todd, C. D., Whyte, B. D. M., MacLean J.C. and Walker, A.M. (2006) Ectoparasitic sea lice (Lepeophtheirus salmonis Krøyer, Caligus elongatus Nordmann) infestations of wild adult one sea-winter Atlantic salmon Salmo salar L. returning to Scotland, 1998–2005. Under review.

267

SUMBAWS

Final Report

Tully, O. (1989) The succession of generations and growth of the caligid copepods Caligus elongatus and Lepeophtheirus salmonis parasitizing farmed Atlantic salmon smolts (Salmo salar L). J. Mar. Biol. Assn.

UK, 69, 279–287.

Tully, O., Poole, W. R., Whelan, K. F. and Merigoux, S. (1993) Parameters and possible causes of Lepeophtheirus salmonis (Krøyer) infesting sea trout (Salmo trutta L.) off the west coast of Ireland. In Boxshall, G. A. and Defaye, D. (eds), Pathogens of wild and farmed fish. Sea lice. Ellis Horwood, Chichester, pp. 202–213.

Vanderklift, M. A. and Ponsard, S. (2003) Sources of variation in consumer-diet δ15N enrichment: a metaanalysis. Oecologia, 136, 169–182.

Walsh, S. B., Dolden, T. A., Moores, G. D., Kristensen, M., Lewis, T., Devonshire A. L. and Williamson, M. S. (2001) Identification and characterization of mutations in housefly (Musca domestica) acetylcholinesterase involved in insecticide resistance. Biochem. Journ., 359, 175–181.

Wells, A., Grierson, C. E., MacKenzie, M., Russon, I. J., Reinardy, H., Middlemiss, C., Bjørn, P. A., Finstad, B., Wendelaar Bonga, S. E., Todd, C. D. and Hazon, N. (2006) The physiological effects of simultaneous abrupt seawater entry and sea lice (Lepeoptheirus salmonis Krøyer) infestation of wild sea trout (Salmo trutta L.) smolts. Under review.

Whelan, K. F. (1993) Decline of sea trout in the west of Ireland: an indication of forthcoming marine problems for salmon? In Mills, D. (ed.), Salmon in the sea and new enhancement strategies. Fishing News Books, Oxford, pp. 171–183.

Wootten, R., Smith, J. W. and Needham, E. A. (1982) Aspects of the biology of the parasitic copepods Lepeophtheirus salmonis and Caligus elongatus on farmed salmonids, and their treatment. Proc. R. Soc. Edin. B, 81, 185–197.

Wright, S. (1969) Evolution and the genetics of populations. The theory of gene frequencies. Vol. 2.

University of Chicago, Chicago.

Zhao, Y., Park, Y. and Adams, M. E. (2000) Functional and evolutionary consequences of pyrethroid resistance mutations in S6 transmembrane segments of a voltage-gated sodium channel. Biochem. Biophys.

Res. Commun., 278, 516–521.

268

SUMBAWS

Final Report

Paper 4: NEGATIVE IMPACTS OF OCEAN SURFACE WARMING ON GROWTH OF ATLANTIC SALMON IN THE NE ATLANTIC C. D. Todd, T. Marshall, S. Hughes, M. E. Lonergan, B.D.M. Whyte, A. M. Walker and J. C. MacLean (Manuscript) This manuscipt is a collaborative effort with colleagues at Fisheries Research Services (Scotland) and Centre for Environment and Aquaculture Science (England). The present manuscript is incomplete because we chose to wait for the availability of the Strathy Point 2005 data, and only in February 2006 have the River North Esk data been obtained. Some of the analyses have yet to be rerun in order to include the 2005 data, but the patterns of variation in condition factor between Strathy Point 1993-2005 and River North Esk 19932005 are virtually identical. It is clear that the overall outcomes of the present analyses (up to 2004) will be reinforced and not changed to any substantive degree. Condition factor in 2005 remained extremely low, precisely as we had predicted from knowing that SST anomalies were high throughout 2004-5. The amended script will be submitted for publication by middle of March 2006.

These data and analyses remain confidential for the present time. The River North Esk data, in particular, are held as commercial-in-confidence and the numbers of fish in the samples have been removed from the report. Because the netsmen make available most of their catch for scientific monitoring there is a very close concordance between the sample sizes and the actual catch.

Abstract Abundances and growth-mediated survival of North American Atlantic salmon stocks have previously been positively correlated with the geographic area of suitable winter thermal habitat in the NW Atlantic Ocean. However, this paradigm does not extend to the NE Atlantic and European stocks, for which warmer springtime (May) coastal temperatures have variously been correlated either with enhanced individual growth and year class survival, or reduced juvenile growth. 1992-2005 has been a period of steady surface warming throughout most of the North Atlantic. We analysed condition factor, as the predicted weight of a standard 60 cm fish, for the corresponding 13 year classes of southern European one sea-winter adult salmon returning each summer to Scotland between 1993 and 2005. For mixed stocks and for a single, defined E Scotland river stock we show significant annual fluctuations and an overall recent decrease of ~14 and 22% respectively in individual marine growth success. The two time-series of annual variations in predicted weights were very highly correlated between the mixed and single stocks, and showed significant negative correlations with sea surface temperature anomalies in the NE Atlantic, especially in midwinter (January).

These results indicate that ocean climate-linked growth conditions for salmon in the NE Atlantic have deteriorated markedly as recent surface warming has progressed.

Introduction In recent years, Atlantic salmon (Salmo salar L.) abundances have fallen sharply to historical lows (1, 2) and these changes usually have been linked to a long-term decline in marine survivorship (3, 3a).

Assessments of the health and productivity of Atlantic salmon stocks are based largely on numerical estimates of oceanic abundance, adult return rates and spawning escapements of the one sea-winter (1SW) and multi sea-winter (MSW) population components (1). 1SW fish return to spawn after only 12-18 months at sea, whereas MSW fish remain at sea for 2-4 winters. Linkages of North American and European stock abundances to ocean climate fluctuations on both sides of the North Atlantic ocean (3-8, 8a) have led to concerns that these declines might be due, at least in part, to global climate change (2). In parallel with the numerical declines in Atlantic salmon stocks, there also have been marked changes in the age-structure of return migrant populations. Over the past 30 years the proportion of the Scottish catch reported as MSW fish has declined with a corresponding increase of 1SW fish. (9). Similar changes in age structure have been reported specifically for the River Imsa, Norway (9b, 9c). In recognition of the distinctiveness of these maturity groupings in the management of salmon, the International Council for the Exploration of the Sea (ICES) divides European stocks into their 1SW and MSW components and, according to their river of natal 269

SUMBAWS

Final Report

origin, further distinguishes these into ‘northern’ (Russia-Scandinavia-Iceland) and ‘southern’ (British IslesFrance-Spain) components for the purposes of developing pre-fishery abundance estimates and catch advice.

All four groupings are presently considered by ICES to be outside precautionary conservation limits (1).

Results and Discussion Atlantic salmon are opportunist predators of invertebrate zooplankton and nektonic fish at the ocean surface (11). Growth of juvenile parr in freshwater is temperature-dependent (11a) and, because oceanic temperature influences growth and weight-at-age in numerous NW Atlantic pelagic and demersal fish species (12), it is likely that salmon marine survival, individual growth condition, and the determination of their age at maturity (5, 9c), all will to an extent be density-independent (12a) and influenced directly (physiologically) and/or indirectly (trophically) by ocean climate (8, 8a, 11a, 12b). Because Atlantic salmon migrate to specific ocean areas of a restricted temperature range (3, 5-8), and survivorship appears to be strongly growth-dependent (8), indirect temperature-driven effects on prey zooplankton and nekton assemblages (and hence fish growth) might be especially important. Over the period 1992-2005 (Figure 1) sea surface temperature (SST) throughout most of the North Atlantic, including the foraging areas occupied

by Atlantic salmon both in the NW (Labrador Sea, Greenland) and NE Atlantic (Norwegian Sea), have risen steadily (14, 15).

Figure 1 Changes in mean SST in the North Atlantic, 1992-2004 (NOAA-CIRES Climate Diagnostics Center http://www.cdc.noaa.gov). The monthly averaged data for 2441 1º latitude/longitude boxes are derived from a combination of satellite and in situ measurements (13). The overall trends are derived from a simple linear fit to the time-series data. In red areas the annual mean SST has increased since 1992, in blue areas SST has decreased.

Atlantic salmon smolts weigh only ~20-30g on entering seawater and it is probable that juvenile mortality early in the marine phase is typically high (6, 8). Evidence for ocean-atmosphere climate linkages to salmon abundance was first derived for North American stocks (2, 3, 7), whereby the North Atlantic Oscillation (NAO; 16) was found to be in negative phase with the area of conducive (4–8 ºC) winter oceanic habitat in the NW Atlantic (3-5). However in recent years, the patterns of sea level pressure in the North Atlantic have been such that the NAO Index has not been a useful measure of climactic conditions (14b).

The winter thermal habitat area concept that pertains for salmon to the NW Atlantic has not been demonstrable for European stocks occupying the NE Atlantic but, over the period 1965-1993, for two European river populations migrating to the Norwegian Sea, the springtime (May) North Sea and coastal 270

SUMBAWS

Final Report

Norwegian SST did correlate positively with early marine survivorship of juveniles (4, 5, 8), and their subsequent return rates as 1SW adults (6). Nonetheless, analyses of North American stock abundances (3), estimated from NW Atlantic commercial landings over the period 1910-1997, have also shown negative correlations with winter and spring SSTs, respectively, in the Gulf of Maine and Gulf of St Lawrence. The emerging paradigm for Atlantic salmon at sea has therefore been one of growth-mediated survival showing positive correlations with warmer SSTs. Nonetheless, recent analyses of the first 3-4 months of marine growth of post-smolts that returned to Scotland as 2SW spawning adults between 1964 and 1993 do indicate a negative effect of higher temperatures (12b).

Here we focus not on numerical stock abundances, but on condition factor (expressed as the predicted weight of a standard 60 cm fish, 17) – as a measure of ‘quality’, or overall marine growth success – of return migrant southern European 1SW adult salmon caught in NE Atlantic home waters. Because MSW salmon may remain at sea for up to 4 winters, the separate cohorts, or year classes, of 1SW fish offer the best opportunity for assessing integrated environmental effects on growth throughout the marine phase. We obtained data for 13 consecutive cohorts of 1SW adults (1992-3 to 2004-5 inclusive) from two commercial fisheries. The summer coastal trap nets at Strathy Point, N Scotland (58º36'N 04º00'W; n = 4915 fish) are a mixed-stock (“interceptory”) fishery; most salmon caught here are in their home waters, they have completed their marine growth, and are destined for numerous rivers throughout both the west and east coasts of Scotland, with some returning to Ireland (19). Scotland and Ireland together dominate the ICES southern European stock component, and marine growth data from Strathy Point are arguably representative of multiple stocks within that ICES grouping. Furthermore, fish taken at this station have not been subject to prior fishery pressure, which may, for example, be size selective. Catches were sampled repeatedly at intervals throughout each summer netting season (weeks 23–33). Thus, for example, between June and August 2005 we repeatedly sampled the fully-grown 1SW adults derived from the spring 2004 emigrant smolt run. Comparative data over the same time period for a single, identifiable river stock were obtained from the commercial in-river net fishery for the River North Esk, E Scotland (56º46'N 02º26'W; 19a). North Esk data were available for more days, and throughout a much longer season in each year. The analysed annual North Esk data were therefore confined to the period embracing the median start and end dates (Julian days 173 and 225) of the Strathy Point samples.

Within years, as each season progressed, we found the expected significant increases in fish length and weight (ANCOVA, covariate Julian day, P <0.001; 18) for both the single and multiple stocks, but there also was significant between-year variation in both length and weight (P <0.001) and year x Julian day interaction (P <0.001, 18). Condition can be calculated for individual fish by several standard measures, based on length/weight relationships (18a). Most condition factor measures are length-dependent (18b), and to circumvent this problem we here express condition for each sample date within each year as the predicted weight of a standard 60 cm fish from the length/weight relationship of individuals captured that day. Within 12 of the 13 years at Strathy Point the standard predicted weight showed no significant correlation with Julian day of repeat sampling. For 1994, the predicted weight showed a significant decrease over time (r = 0.801, n = 8, p = 0.017) and in 2005 this rose and then fell as the season progressed. The same pattern of no significant change in condition factor within years was noted for River North Esk. Overall, and for any one year class, condition factor therefore essentially is ‘set’ for the entire cohort of return migrant 1SW salmon, irrespective of their river of origin; between-year variation throughout the time-series exceeded within-year variation.

The lack of significant within-year change permitted the use of a single annual value of condition factor in subsequent correlation analyses. The median sample date across all years at Strathy Point was Julian day 200. To provide the measure, for both data sets we derived for day 200 of each year the weight of the standard 60 cm fish using the 13 LOIS lines of best fit to predicted weight at the standard length for each sampling occasion within every year. This expression of condition shows that temporal pattern of growth success of both the single stock, and of multiple mixed stocks of southern European 1SW Atlantic salmon, was identical (p = 0.93) and has declined sharply since the peak in 1997 (Fig. 2).

271

SUMBAWS

Final Report 2

y = -5.8883x + 23524x - 2E+07

2750 2700 2650 2600 2550

Series1 Poly. (Series1)

2500 2450 2400 2350 2300 2250 1992 1994 1996 1998 2000 2002 2004 2006

De-trended Predicted Weight for a standard 60 cm fish, Day 200 2 1.5 1 0.5 0 -0.5 -1

2005

2004

2003

2002

2001

2000

1999

1998

1997

1996

1995

1994

1993

-1.5 -2

Figure 2 Salmo salar. Annual changes in mean condition factor, expressed as the predicted weight ±95% c.i. of a standard 60 cm fish on Julian day 200 (July), for 13 consecutive year classes of returning 1SW southern European adults. A 2nd order polynomial provided the best fit (y = -5.8883x2 + 23524x – 2E+07, r = 0.763, P <0.01) and was used to de-trend the data. The de-trended data were then normalized prior to correlation analysis with the de-trended monthly SST anomalies over the period May 1992-June 2005 (Figure 4).

The ocean migrations of southern European 1SW salmon are known to be confined to the NE Atlantic, and their winter feeding area centres on the Norwegian Sea to the north of Faeroe (6, 7, 8, 10, 20) but south of the Barents Sea polar front. May is the peak month of first entry of British Isles smolts to seawater (20) and they rapidly move northeast towards the Norwegian coast (20), entering the offshore areas of the Norwegian Sea by midsummer (10). By late July a few southern European post-smolts are known to already occupy areas of the northern Norwegian Sea beyond 70oN (10), where they may remain, or subsequently move to the south, prior to commencing their migratory return the following spring/summer. Figure 3 shows an illustrative example of the SST time-series, over the period 1992-2005, for a 1°x1° lat./long. grid box within the Norwegian Sea (Figure 4A; 21). The strong decadal pattern in SST anomalies probably is linked to the NAO (15, 16), but it is also clear that over the analytical period SST has been steadily rising. We sought correlations between variation in the predicted weight of a standard 60 cm fish on day 200 for each of the 13 mixed-stock 1SW cohorts sampled at Strathy Point and the SST time-series for every 1ºx1º grid box throughout the NE Atlantic (30ºW–30ºE, 47ºN–80ºN).

272

SUMBAWS

Final Report

3 2 1 0 -1 -2

1985

1990

1995

2000

2005

Figure 3 Sea surface temperature (SST) in the Norwegian Sea. The red line shows the monthly mean SST anomaly (°C) for an illustrative 1º latitude/longitude grid box (67.5ºN 4.5ºW) from the NOAA Optimal Interpolation SSTv2 dataset. The anomalies were derived by removing the average seasonal values using a climatology derived for the period 1971-2000. Fitting of a 4th order polynomial (black) to the observations permitted de-trending of the monthly anomalies (blue) correlation analysis with annual predicted salmon weight (Figure 4).

Figure 4, B1-14 shows the incidence and geographic extent of significant correlations, over the period 1993-2005, between the de-trended monthly SST anomaly time-series (15) for 1ºx1º grid boxes and the detrended annual time-series of predicted weights of standard fish on Day 200 (Figure 2B). The majority of grid boxes showed no significant correlation. Where significant, correlations were predominantly negative. .

Especially intense and extensive negative correlations occurred both in summer (June-August, adjacent to the British Isles) and midwinter (January); the latter are perhaps especially relevant in coinciding geographically with the known winter feeding areas (Figure 4A) of southern European 1SW salmon (11).

This outcome is in marked contrast with previous studies of NE Atlantic salmon stocks over the earlier period 1965-1993 (5, 6, 8), and a more recent study for the River Imsa in Norway (9c); namely that warmer spring (May) temperatures generally were conducive to salmon growth, and thereby survival and stock abundances. Our data for predicted weight showed no significant correlation with the NAO Index for the period of measurement (r = 0.31; p = 0.32), but do indicate that the recent warming of sub-arctic North Atlantic SST over the past decade (Figs. 1, 3) has been increasingly detrimental to salmon growth (Figure 2), and especially so in the Norwegian Sea when temperatures are at their seasonal minimum (Figure 4 B9).

Figure 4 A: Schematic summary (green) of the post-smolt/adult migratory habitat, and the Norwegian Sea oceanic feeding grounds, exploited by southern European 1SW Atlantic salmon and which originate in Scotland and Ireland (21). Figure 4 B1-14: Correlation analyses (Pearson’s coefficient) for the de-trended annual predicted weight for a standard 60 cm fish (Day 200, July) (1993-2005; Figure 2) with the de-trended monthly sea surface temperature anomaly data (1992-2005) for all NOAA OISSTV2 NE Atlantic 1º grid boxes. Monthly correlation analyses range from May of the previous year (B1, -14 mo), when each annual cohort of smolts will have first entered sea water, to June (B14, -1 mo) of the summer of capture. Significant negative correlations are shown in blue/light blue and positive correlations in red/orange. Only those grid boxes showing significant correlations (critical values: r = 0.553, p = 0.05; r = 0.684, p = 0.01) are illustrated.

273

SUMBAWS

Final Report

Condition factor, as expressed by the annual standard weight for both single and multiple southern European stocks, shows significant year-to-year fluctuations and yet is ‘set’ for whole 1SW year classes, irrespective of stock sub-structuring according to natal river (22). This implies a widespread and subtle, but ecologically profound, growth response of entire cohorts of Atlantic salmon to prevailing ocean climate conditions. Salmon cease feeding on their migratory return, and the effects of the observed annual variation in predicted weight on female fecundity will have important management implications, especially for the many relatively small river systems in, for example, Scotland (23) and Ireland with spawning populations dominated by 1SW adults. Male 1SW summer return migrants are typically longer and heavier than females (18) and egg production is a major investment of female salmon, which typically are semelparous (24): the ovaries alone comprise ~30% of the female’s total energy reserves at spawning (25) and represent about half of the energy expended in maturation, upstream migration and reproduction combined (26, 26a). For the standard 60 cm female 1SW fishat Strathy Point, our observed annual extremes of mean size (Figure 2) represent a 14% weight difference of 2330 g (2004) versus 2717 g (1997), which will be reflected by a difference of ~12% in egg production. For River North Esk, the 2004 cohort showed a reduction in growth of 22% compared to1997. Notwithstanding density-dependent effects on juvenile survival in freshwater, such variation in initial egg deposition will have important implications for population stability and growth for small or threatened river stocks.

Commensurate with present stock abundances being at unprecedented lows, the recent sharp decline in predicted weight for multiple stocks of the southern European 1SW maturity grouping, and the negative correlations with NE Atlantic SST anomalies, provide strong evidence of a recent, marked and continuing deterioration in ocean climate and growth opportunities for wild Atlantic salmon. Whilst acknowledging that numerical abundance data will remain the focus of stock monitoring and management strategies, we would argue that analysis of condition factor, or quality of pre-spawning adults, provides a valuable additional empirical tool in assessing chronic, but ecologically pervasive, effects of ocean climate variation on the performance and status of wild Atlantic salmon stocks. Such marked between-year differences in fecundity should, perhaps, be considered when deriving spawning escapement conservation limits for given stocks.

References 1. “Report of the Working Group on North Atlantic Salmon (WGNAS), 5-14 April 2005, Nuuk, Greenland” (ICES CM 2005/ACFM:17, ICES, Copenhagen, Denmark, 2005). Available at http://www.ices.dk/reports/ACFM/2005/WGNAS/directory.asp 2. K. D. Friedland, D. G. Reddin, J. R. McMenemy, K. F. Drinkwater, Can. J. Fish. Aquat. Sci. 60, 563 (2003).

3. K. D. Friedland, D. G. Reddin, M. Castonguay, ICES J. Mar. Sci. 60, 343 (2003).

3a. B. Jonsson, N. Jonsson, Can. J. Fish. Aquat. Sci. 61, 2369 (2004).

4. K. D. Friedland, D. G. Reddin, J. F. Kocik, ICES J. Mar. Sci. 50, 481 (1993).

5. K. D. Friedland, Can. J. Fish. Aquat. Sci. 55(Suppl. 1), 119 (1998).

6. K. D. Friedland, L. P. Hansen, D. A. Dunkley, Fish. Oceanogr. 7, 22 (1998).

7. D. G. Reddin, W. M. Shearer, Am. Fish. Soc. Symp. 1, 262 (1987).

8. K. D. Friedland, L. P. Hansen, D. A. Dunkley, J. C. MacLean, ICES J. Mar. Sci. 57, 419 (2000).

8a. G. Beaugrand, P. C. Reid, Global Change Biol. 9, 801 (2003).

9. “Statistical Bulletin. Scottish Salmon and Sea Trout Catches, 2003” (Fis/2004/1, Fisheries Research Services, Scotland, 2004). Available at http://www.marlab.ac.uk/FRS.Web/Delivery/Information_Resources/information_resources_view_documen t.aspx?contentid=1375.

9b. N. Jonsson, B. Jonsson, L. P. Hansen, J. Appl. Ecol. 40, 900 (2003).

9c. N. Jonsson, B. Jonsson, J. Fish Biol. 64, 241 (2004).

10. L. P. Hansen, J. A. Jacobsen, ICES J. Mar. Sci., 60, 110 (2003).

11. J. A. Jacobsen, L. P. Hansen, ICES J. Mar. Sci., 58, 916 (2001).

11a. B. Jonsson, T. Forseth, A. J. Jensen, T. F. Næsje, Funct. Ecol. 15, 701 (2001).

274

SUMBAWS

Final Report

12. K. F. Drinkwater, in The Ocean Life of Atlantic Salmon. Environmental and Biological Factors Influencing Survival D. Mills, Ed. (Fishing News Books, Oxford, England, 2000), pp. 116–136.

12a. N. Jonsson, B. Jonsson, L. P. Hansen, J. Anim. Ecol. 67, 751 (1998).

12b. K. D. Friedland, G. Chaput, J. C. MacLean, ICES J. Mar. Sci. 62, 1338 (2005).

13. R. W. Reynolds, N. A. Rayner, T. M. Smith, D. C. Stokes, W. Wang, J. Clim. 15, 1609 (2002).

14. J. T. Houghton et al., Eds., Climate Change 2001: the Scientific Basis. Contribution of Working Group 1 to the Third Assessment Report of the Intergovernmental Panel on Climate Change. (Cambridge Univ.

Press, New York, 2001).

14b. S. L. Hughes and A. Lavín, Eds., The Annual ICES Ocean Climate Status Summary 2004/2005. ICES Cooperative Research Report. No xx, xxpp.

15. Materials and methods are available as supporting material on Science Online.

16. R. R. Dickson, W. R. Turrell, in The Ocean Life of Atlantic Salmon. Environmental and Biological Factors Influencing Survival, D. Mills, Ed. (Fishing News Books, Oxford, England, 2000), pp. 65–74.

18. Results tables S1-S6 are available as supporting material on Science online.

18a. C. T. Marshall, C. L. Needle, N. A. Yaragina, A. M. Ajiad, E. Gusev, Can. J. Fish. Aquat. Sci. 61, 1900-1917 (2004).

18b. B. G. Blackwell, M. L. Brown, D. W. Willis, Rev. Fish. Sci. 8, 1 (2000).

18c. Supporting material, Table S1.

19. W. M. Shearer, in The Status of the Atlantic Salmon in Scotland, D. Jenkins, & W. M. Shearer, Eds. (Institute of Terrestrial Ecology, Abbots Ripton, England, 1986), pp. 7-49.

19a. The catch data from River North Esk are commercial-in-confidence. Sample sizes cannot be reported, but were consistently large and exceeded 334 in every year 20. M. Holm, J. C. Holst, L. P. Hansen, ICES J. Mar. Sci. 57, 955-964 (2000).

21. The delineation of the Norwegian Sea oceanic feeding grounds exploited by southern European 1SW Atlantic (green fill, Figure 4A) was compiled from data in refs. 6, 8 and 10. The northernmost extent probably is set by the Barents Sea polar front, to the east of Bear Is. (74.5ºN 20ºE), which is indicated in Figure 1 by the intersection of contrasting patterns of cooling/warming.

22. A. F. Youngson et al., Fish. Res. 62, 193 (2003).

23. A. Youngson, D. Hay, The Lives of Salmon (Swan Hill, Shrewsbury, England, 1996).

24. B. Jonsson, N. Jonsson, Can. J. Fish. Aquat. Sci. 61, 2369 (2004).

25. B. Jonsson, N. Jonsson, I. A. Fleming, Funct. Ecol. 10, 89 (1996).

26. I. A. Fleming, Can. J. Fish. Aquat. Sci. 55(Suppl. 1), 59 (1998).

26a. N. Jonsson, B. Jonsson, Can. J. Fish. Aquat. Sci. 60, 506 (2003).

27. We thank L. P. Hansen, D. Hay, P. Hutchinson, B. Jonsson, M. Pawson, R. G. J. Shelton, W. R. Turrell and A. F. Youngson for their insightful comments and discussion of the manuscript. Partial support of this work derives from the European Commission (SUMBAWS, Contract No. Q5RS-2002-00730).

Supporting Materials and Methods Sea surface temperature (SST) and condition factor (CF) data processing. Since the late 19th century, global average ocean surface temperatures have increased by ~0.6ºC per century (0.06ºC per decade; 14). Monthly SST data for the North Atlantic were extracted from the NOAA Optimum Interpolation SSTV2 dataset (NOAA-CIRES Climate Diagnostics Center, http://www.cdc.noaa.gov). The monthly averaged data are generated on a 1º latitude/longitude grid from a combination of satellite and in situ measurements (13). The trend in North Atlantic SST during June 1992 to July 2005 was calculated by fitting a linear regression to the SST time-series at each grid box. The average per annum temperature 275

SUMBAWS

Final Report

changes (Figure 1) were obtained from the slope of the linear fit. Over the period 1992-2005, SSTs have warmed at a rate of ~1-2ºC per decade, which is consistent with the accelerated warming reported by the IPCC (14), and the OISSTV2 monthly data reveal that recent North Atlantic warming has been strongest in the summer months (data not shown). In addition to a general rising trend, SST in the North Atlantic also fluctuates with a multi-decadal period. This fluctuation in SST has been in a rising phase since the mid 1980s and warming in many regions of the Atlantic Ocean appears to have accelerated since the mid 1990s (14; Figure 3).

North Atlantic SSTs show a strong seasonal cycle, which exceeds inter-annual variability. Typically, temperatures range seasonally by ~5ºC in the central North Atlantic, to ~10ºC in coastal regions, and even >20ºC in certain areas such as the southern North Sea and the Baltic Sea. Prior to further analysis, the seasonal cycle and was removed from the OISSTV2 data, using a climatology calculated for the base period 1971-2000 (available from SLH; S.Hughes@marlab.ac.uk), to convert the data to monthly anomalies. The resultant anomaly dataset required further de-trending to preclude or reduce autocorrelation due to the strong multi-decadal variability (e.g. Figure 3). The latter appears to be driven by patterns of sea level atmospheric pressure and can be explained, to some extent, by the North Atlantic Oscillation (16). The long-term trends were removed from the anomaly data by fitting a 4th order polynomial for each grid box and subtracting this curve from the data (e.g. Figure 3). Following removal of the long-term trends, the integral time lag (a measure of how far apart data points need to be before they can be classed as independent) reduced from 12 months to 6 months or less throughout the North Atlantic, and ensured that data for a single month of any year can be treated as an independent variable. Year to year changes in the time-series can, therefore, be investigated without introducing autocorrelation effects. To allow correlation analysis between the detrended monthly SST anomaly and the annual predicted weight time-series, a 2nd order polynomial was chosen as the optimal model (based on the AIC test) to remove the downward trend in weight over the observational period.

Statistical analyses. Prior to ANCOVA the dependents (log fork length, log weight) were tested for homogeneity of variances (Levene’s test) and normality (Kolmogorov-Smirnov test, D). At Strathy Point, log fork length (Levene’s test statistic: 1.432, p = 0.143; D = 0.010, P >0.150; n = 4915) was non significant, but weight (test statistic: 1.882, p = 0.032; D = 0.018, P <0.010) was significant. At river North Esk, log fork (Levene’s test statistic: 1.550, p = 0.107 ; but D = 0.020, P <0.010; n = 9743) was non significant, but for log weight (Levene’s test statistic: 4.524, P <0.001; D = 0.010, p = 0.023; n = 9743).

Heterogeneity of variances and departures from normality are not, however, problematic to the assumptions of ANOVA for very large data sets (27), and ANCOVA was applied to the two dependents at each sample site with Julian day [= sample date within Year] as the covariate, and Year as the factor. Sequential models were applied, fitting Year first followed by Julian day, and using the sequential sums of squares for tests.

Predicted weights of a standard 60 cm fish on each sampling date at each site (Strathy Point; River North Esk) were determined by GAM modeling of the weight/length relationship on each day. LOIS plots of best fit were applied to the predicted weight data across time, and within each year. The median sample date for salmon captured throughout the 10-11 week season at Strathy Point in each year typically fell in July (Julian day 200; Supporting Table S1). Correlation analysis showed no significant change in predicted weight of the standard fish for 11 of the 13 years. In one year the predicted weight decreased significantly, and in 2005 this rose, and then fell as the season progressed. The predicted weight on Day 200 was determined from the LOIS plots for each site and this provided the single value of condition factor for each year. July was selected to designate the annual value of predicted weight for SST correlation analyses. Pearson correlations were calculated between the de-trended 13-year Strathy Point salmon weight time-series and the de-trended SST anomaly time-series for every NE Atlantic grid box for each of the 14 months (1992-2005). Salmon sampled at River North Esk were consistently shorter and lighter than those sampled at Strathy Point. But since there was a very high correlation (p = 0.93) between the Strathy Point and River North Esk salmon weight time-series the SST correlation analyses are presented only for the multiple stock data from Strathy Point. The time lags of the 14 correlation analyses therefore ranged from -1 mo (June, the month prior to capture for any one year) to -13 and -14 mo (respectively, June and May of the year previous to capture, when the fish would have first migrated to sea). Only grid boxes showing significant correlations (P <0.05) between the SST and weight time-series are illustrated in Figure 4 B1-14.

276

SUMBAWS

Final Report

Supporting Tables Table S1A Salmo salar. Summary of sample dates, fish numbers, mean lengths and weights for all years at Strathy Point. Mean lengths and weights are back-transformed logarithmic values. The median day 200 was chosen as the standard annual date for SST and weight time-series correlation analyses in Figure 4B1-14.

Year

Dates of first/last sample

Median Julian day

No. of fish samples

1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005

30 Jun/ 11 Aug 14 Jun/ 27 Jul 21 Jun/ 16 Aug 25 Jun/ 21 Aug 17 Jun/ 6 Aug 23 Jun/ 5Aug 22 Jun/ 18 Aug 27 Jun/ 16 Aug 19 Jun/ 15 Aug 25 Jun/ 20 Aug 17 Jun/ 20 Aug 29 Jun/ 4 Aug 22 Jun/ 16 Aug

208 187 200 211 197 196 201 207 209 204 200 198 197

410 384 664 689 292 409 120 316 185 224 597 294 331

annual mean length (cm) weight (kg) 60.9 2.595 60.3 2.690 61.0 2.793 60.7 2.680 61.0 2.842 58.3 2.279 59.6 2.533 62.4 2.928 60.7 2.560 60.8 2.617 60.3 2.510 58.9 2.225 59.8 2.256

Table S1B Salmo salar. Summary of sample dates, mean lengths and weights for all years at River North Esk. Mean weights are back-transformed logarithmic values. Day 200 was chosen as the standard annual date for condition factor (predicted weight of the standard 60 cm fish) time-series for correlation analyses with Strathy Point.

Year

Dates of first/last sample

Median Julian day

1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005

22 Jun/ 16 Aug 20 Jun/ 15 Aug 20Jun/ 18 Aug 21 Jun/ 13 Aug 23 Jun/ 12 Aug 23 Jun/ 13 Aug 21 Jun/ 14 Aug 21 Jun/ 14 Aug 21 Jun/ 13 Aug 21 Jun/ 13 Aug 23 Jun/ 12 Aug 23 Jun/ 17 Aug 21 Jun/ 15 Aug

200 196 195 198 203 196 199 199 199 197 203 195 196

No. of fish samples

annual mean length (cm) weight (kg) 58.7 2.21 59.2 2.34 58.5 2.32 58.7 2.36 58.3 2.34 56.6 1.99 56.8 2.02 58.7 2.28 58.9 2.23 56.2 1.94 58.1 2.07 57.4 1.92 57.6 1.94

277

SUMBAWS

Final Report

Table S3 Salmo salar, Strathy Point. Sequential ANCOVA for fork length (cm); covariate, Julian Day; factor, Year.

Source Yeas Julian day Year*Julian day Error Total

df 12 1 12 4889 4914

Seq. SS 3869.92 10211.10 610.13 94662.40 109353.55

Seq. MS 322.49 10211.10 50.84 19.36

F 16.66 527.37 2.63

P <0.001 <0.001 0.002

Table S4 Salmo salar, Strathy Point. Sequential ANCOVA for log body weight (kg); covariate, Julian Day; factor, Year.

Source Yeas Julian day Year*Julian day Error Total

df 12 1 12 4889 4914

Seq. SS 6.21625 4.05897 0.44538 54.68140 65.40201

Seq. MS 0.51802 4.05897 0.03711 0.01118

F 46.32 527.3362.917 2.633.32

P <0.001 <0.001 <0.001

Table S5 Salmo salar, River North Esk. Sequential ANCOVA for log length (cm); covariate, Julian Day; factor, Year.

Source Yeas Julian day Year*Julian day Error Total

df 11 1 11

Seq. SS 576519 1.208574 0.073167 10.037694 11.895955

Seq. MS 0.0524 1.208574 0.006652 0.001033

F 50.75 442.04 6.44

P <0.001 <0.001 <0.001

Table S6 Salmo salar, River North Esk. Sequential ANCOVA for log body weight (kg); covariate, Julian Day; factor, Year.

Source Yeas Julian day Year*Julian day Error Total

df 11 1 11

Seq. SS 11.65891 10.03456 0.95194 117.30429 139.94969

Seq. MS 1.05990 10.03456 0.08654 0.1207

F 87.42 214.56 7.17

P <0.001 <0.001 <0.001

Supporting material references 27. Underwood, A.J. Experiments in ecology. (Cambridge, Cambridge University Press, 1997)

278

SUMBAWS

2

Final Report

EXPLOITATION AND DISSEMINATION OF RESULTS

1. CONFERENCES ATTENDED (BUT NO PRESENTATIONS) RUN 6th International Symposium on Fish Immunology, Turku, Finland, 24-29 May 2004 2. CONFERENCES ATTENDED (PRESENTATION TYPE: POSTER, TALK, PLENARY TALK) USTAN International Sea Trout Symposium, Cardiff, 6-8 July 2004 Poster presentation (Wells, A., Grierson, C.E., Todd, C.D. & Hazon, N.) “Assessment of a prophylactic substance in the protection of wild sea trout smolts from subsequent infection with sea lice” 9th International Conference on Copepoda, Tunisia, 11-15 July 2005 Plenary talk (Todd, C.) “Molecular genetic approaches to applied and ecological problems concerning the sea louse, Lepeophtheirus salmonis (Krøyer) infesting salmonid fishes” (Manuscript submitted to Journal of Plankton Research) 12th European Association of Fish Pathologists (EAFP) International Conference on Fish and Shellfish Pathology, Copenhagen, Denmark, 11- 16 September 2005 Talk (Russon, I.J., Wells, A., Grierson, C.E., Todd, C.D, Hazon, N. & Wendelaar Bonga, S.E.) “Aspects of the stress response of sea trout, Salmo trutta trutta (L.), to varying salmon lice, Lepeophtheirus salmonis (Krøyer), intensities when entering seawater” ICES/NASCO Symposium. “Interactions between aquaculture and wild stocks of Atlantic salmon and other diadromous fish species: science and management, challenges and solutions”. Bergen, Norway, 18-21 October 2005 Talk (Wells, A., Grierson, C.E., Russon, I., Wendelaar Bonga, S.E., Bjørn, P.A., Finstad, B., Todd, C. & Hazon, N.) “Stress and osmoregulatory dysfunction in laboratory controlled sea lice infestation of sea trout” CFB ICES/NASCO Symposium. “Interactions between aquaculture and wild stocks of Atlantic salmon and other diadromous fish species: science and management, challenges and solutions”. Bergen, Norway, 18-21 October 2005 Abstract (Finstad, B., Bjørn, P.A., Holst, J.C., Heuch, P.A., Kristoffersen, R., Todd, C.D., Hazon, N., Gargan, P., Tully, O. & McKinley, R.S.) “Impacts of salmon lice on wild salmonids” NINA Sea lice and open science forum, Vancouver, Canada, 22-24 February 2003 Abstracts (McKinley, R.S., Finstad, B., Økland, F., Thorstad, E.B. & Bjørn, P.A.) “Smolt migration and salmon lice infection on Atlantic salmon and sea trout in a Norwegian fjord system” Aquaculture 2004, Honolulu, Hawaii, USA, 1-5 March 2004 Abstract (Finstad, B., Økland, F., Thorstad, E.B., Bjørn, P.A., Sivertsgård, R. & McKinley, R.S.) “The salmon lice case: the need for local models in integrated coastal zone management” 7th International Smoltification Workshop, Iwate, Japan, 24-29 July 2005 Talk (Finstad, B., Kroglund, F., Stefanson, S.O., Kristensen, T.K., Rosseland, B.O., Teien, H.C., Salbu, B.)“The effect of suboptimal water quality and salmon lice infestation on Atlantic salmon post-smolts”

279

SUMBAWS

Final Report

ICES/NASCO Symposium. “Interactions between aquaculture and wild stocks of Atlantic salmon and other diadromous fish species: science and management, challenges and solutions”. Bergen, Norway, 18-21 October 2005 Abstract (Finstad, B., Bjørn, P.A., Holst, J.C., Heuch, P.A., Kristoffersen, R., Todd, C.D., Hazon, N., Gargan, P., Tully, O. & McKinley, R.S.) “Impacts of salmon lice on wild salmonids” Abstract (Wells, A., Grierson, C.E., Russon, I., Wendelaar Bonga, S.E., Bjørn, P.A., Finstad, B., Todd, C.D. & Hazon, N.) “Stress and osmoregulatory dysfunction in laboratory controlled sea lice infestation of sea trout” AquaNet II Conference, Delta Beauséjour, Moncton, New Brunswick, Canada, 14-17 September 2002 Poster (Økland, F., Thorstad, E, B., Finstad, B. & McKinley, R.S.) ”Estimating sea lice infestation on Atlantic salmon and sea trout smolts in a Norwegian fjord” Kultiveringsmøtet 2003, Quality Hotel Vøringsfoss, Norway, 4-6 March 2003 Talk (Thorstad, E.B.) ”Smoltvandringer og lakselusinfeksjoner hos laks og sjøaure i et norsk fjordsystem” 5th Conference on Fish Telemetry held in Europe, Ustica, Palermo, Italy, 9-13 June 2003 Talk (Thorstad, E.B., Økland, F., Finstad, B. & McKinley, R.S.) “Sea lice infestation and migration of Atlantic salmon and sea trout smolts through a Norwegian fjord system” Telemetry Workshop: Positional and physiological telemetry in aquatic research: perspectives from Scandinavia and North America, Umeå, Sweden, 6-10 December 2004 Invited talk (Økland, F., Finstad, B., Thorstad, E.B., Bjørn, P.A., Sivertsgård, R., Plantalech, N. & McKinley, R.S.) “Migration, salmon lice infestation and survival of Atlantic salmon and sea trout postsmolts in selected Norwegian fjord systems” 6th Conference on Fish Telemetry held in Europe, Sesimbra, Portugal, 5-11 June 2005 Talk (Plantalech, N., Økland, F., Thorstad, E.B., Finstad, B., Fiske, P., Sivertsgård, R., Bjørn, P.A. & McKinley, R.S.) “Survival and swimming performance related to morphology of Atlantic salmon (Salmo salar) post-smolts migrating through a Norwegian fjord” Talk (Thorstad, E.B., Finstad, B., Økland, F., Bjørn, P.A., Sivertsgård, R. & McKinley, R.S.) “Is it possible for the aquaculture industry and wild fish interests to coexist? An overview from telemetry and salmon lice studies in Norway” Talk (Økland, F., Thorstad, E.B., Dieserud, O., Plantalech, N., Sivertsgård, R., Jepsen, N., Bjørn, P.A., Finstad, B., Butterworth, K., Cubitt, F. & McKinley, R.S.) “Migratory speed and route of Atlantic salmon smolts through fjords - modelling sea lice infestation” Nordic Workshop for PhD students on Anadromous Salmonid Research NoWPAS, Agdenes, Norway, 7-9 April 2005 Invited talk (Thorstad, E.B., Finstad, B., Økland, F., Sivertsgård, R., Bjørn, P.A. & McKinley, R.S.) “Migration, sea lice infestation and survival of Atlantic salmon and sea trout post-smolts in a Norwegian fjord system” American Fisheries Society 135th Annual meeting, Anchorage, Alaska, 11-15 September 2005 Invited talk (Thorstad, E.B., Økland, F., Finstad, B., Bjørn, P.A., Sivertsgård, R. & McKinley, R.S.) “Fjord migration of wild and hatchery-reared Atlantic salmon and brown trout post-smolts and the risk of sea lice infestation” Forskningsmøte om Vossolaksen, Voss, Norway, 9 December 2005 Talk (Thorstad, E.B.) ”Smoltutvandring-hva har vi lært i Romsdalsfjorden og Hardangerfjorden?”

280

SUMBAWS

Final Report

Workshop “Wild and Farmed Salmon – Working Together” Organized by NASCO/North Atlantic Salmon Farming Industry Liaison Group, NTNU, Gløshaugen, Trondheim, Norway, 9 August 2005 Talk (Finstad, B.) “Area Management Initiatives in Norway” NIFA Sea lice and open science forum, Vancouver, Canada, 22-24 February 2003 Abstract (McKinley, R.S., Finstad, B., Økland, F., Thorstad, E.B. & Bjørn, P.A.) “Smolt migration and salmon lice infection on Atlantic salmon and sea trout in a Norwegian fjord system” Aquaculture 2004, Honolulu, Hawaii, USA, 1-5 March 2004 Abstract (Finstad, B., Økland, F., Thorstad, E.B., Bjørn, P.A., Sivertsgård, R. & McKinley, R.S.) “The salmon lice case: the need for local models in integrated coastal zone management” ICES/NASCO Symposium. “Interactions between aquaculture and wild stocks of Atlantic salmon and other diadromous fish species: science and management, challenges and solutions”. Bergen, Norway, 18-21 October 2005 Abstract (Finstad, B., Bjørn, P.A., Holst, J.C., Heuch, P.A., Kristoffersen, R., Todd, C., Hazon, N., Gargan, P., Tully, O. & McKinley, R.S.) “Impacts of salmon lice on wild salmonids” Talk (Wells, A., Grierson, C.E., Russon, I., Wendelaar Bonga, S., Bjørn, P.A., Finstad, B., Todd, C.D. & Hazon, N.) “Stress and osmoregulatory dysfunction in laboratory controlled sea lice infestation of sea trout” AquaNet II Conference, Delta Beauséjour, Moncton, New Brunswick, Canada, 14-17 September 2002 Talk (Bjørn, P. A.) “The physiological and ecological effects of salmon lice infection on anadromous salmonids” Miniseminar i forbindelse med Forskningsdagene Fiskeriforskning, Tromsø, Norway, 2002 Talk (Bjørn, P.A.) ”Har Nasjonal Handlingsplan mot lus vært en fiasko?”

Foredrag for Tromsø Naturvernforbund, Norges Fiskerihøgskole, Tromsø, Norway, 2002 Talk (Bjørn, P.A.) ”Hvordan skal vi ta vare på vill laksefisk og samtidig skape vekstvillkår for ei sunn og god havbruksnæring” Norwegian-Canadian Fishery and Aquaculture Conference, Ottawa, Canada, 2003 Talk (Bjørn, P.A.) “Protecting the environment while securing sustainable growth of aquaculture: The salmon lice case” Foredrag referansegruppa for NFR prosjektet ”Salmon lice as population regulating factor in Norwegian salmon”. Alta rådhus, Alta, Norway, 2003 Talk (Bjørn, P.A. & Finstad, B.) ”Registrering av lakselus på vill laksefisk i Altafjorden i 2003” Stortingets Energi og Miljøkommite, Surnadal, Norway, 2003 Talk (Bjørn, P.A.) ”Statusrapport om forholdet mellom lakselus på oppdrettslaks og på vill sjøørret, sjørøye og laks i Norge” Forskningsrådets kystsonekommite,. Universitetet i Tromsø, Tromsø, Norway, 31 March 2003 Talk (Bjørn, P.A.) ”Hvordan skal vi greie å ta vare på miljøet og samtidig skape vekstvillkår for ei sunn og god oppdrettsnæring?”

Fiskeri- og havbruksnæringens forskningsfond. Fiskeriforskning, Tromsø, Norway, 2003 Talk (Bjørn, P.A.) ”Hvordan skal vi greie å ta vare på miljøet og samtidig skape vekstvillkår for ei sunn og god oppdrettsnæring?”

281

SUMBAWS

Final Report

Forskningsdagene under arrangementet "Bærekraftig utvikling av havbruksnæringen i MidtTroms". Rubbestad havbruksområde, Vangsvik, Norway, 2004 Talk (Dale, T. & Bjørn, P. A.) “Nasjonale laksefjorder. Konflikter mellom oppdrett og miljø” Telemetry Workshop: Positional and physiological telemetry in aquatic research: perspectives from Scandinavia and North America. Umeå, Sweden, 6-10 December 2004 Invited talk (Økland, F., Finstad, B., Thorstad, E.B., Bjørn, P.A., Sivertsgård, R., Plantalech, N. & McKinley, R.S.) “Migration, salmon lice infestation and survival of Atlantic salmon and sea trout postsmolts in selected Norwegian fjord systems” Fiskehelse- og miljøgruppene i Nordland, Troms og Finnmark. Bodø, Norway, 2005 Talk (Dale, T, Bjørn, P. A., Koren, C, Brørs, B. & Slagstad, D.) ”Risiko for lakselussmitte på vill og oppdretta laksefisk” Nordnorsk Havbrukslag. Clarion Bryggen Hotell. Tromsø, Norway, 2005 Talk (Dale, T, Bjørn, P. A., Koren, C, Brørs, B. & Slagstad, D.) ”Risiko for lakselussmitte på vill og oppdretta laksefisk” 6th Conference on Fish Telemetry held in Europe, Sesimbra, Portugal, 5-11 June 2005 Talk (Plantalech, N., Økland, F., Thorstad, E.B., Finstad, B., Fiske, P., Sivertsgård, R., Bjørn, P.A. & McKinley, R.S.) “Survival and swimming performance related to morphology of Atlantic salmon (Salmo salar) post-smolts migrating through a Norwegian fjord” Talk (Thorstad, E.B., Finstad, B., Økland, F., Bjørn, P.A., Sivertsgård, R. & McKinley, R.S.) “Is it possible for the aquaculture industry and wild fish interests to coexist? An overview from telemetry and salmon lice studies in Norway” Talk (Økland, F., Thorstad, E.B., Dieserud, O., Plantalech, N., Sivertsgård, R., Jepsen, N., Bjørn, P.A., Finstad, B., Butterworth, K., Cubitt, F. & McKinley, R.S.) “Migratory speed and route of Atlantic salmon smolts through fjords - modelling sea lice infestation” Nordic Workshop for PhD students on Anadromous Salmonid Research NoWPAS, Agdenes, Norway, 7-9 April 2005 Invited talk (Thorstad, E.B., Finstad, B., Økland, F., Sivertsgård, R., Bjørn, P.A. & McKinley, R.S.) “Migration, sea lice infestation and survival of Atlantic salmon and sea trout post-smolts in a Norwegian fjord system” American Fisheries Society 135th Annual meeting, Anchorage, Alaska, 11-15 September 2005 Invited talk (Thorstad, E.B., Økland, F., Finstad, B., Bjørn, P.A., Sivertsgård, R. & McKinley, R.S.) “Fjord migration of wild and hatchery-reared Atlantic salmon and brown trout post-smolts and the risk of sea lice infestation” RUN 11th European Association of Fish Pathologists (EAFP) International Conference on Diseases of Fish and Shellfish, Malta, 21-26 September 2003 Poster presentation (Walker, P.D., Haond, C., Russon, I.J. & Wendelaar Bonga, S.E.) “The effect of temperature and age of parasite on the off-host survival of Argulus japonicus Thiele (Crustacea: Branchiura)” 12th European Association of Fish Pathologists (EAFP) International Conference on Fish and Shellfish Pathology, Copenhagen, Denmark, 11-16 September 2005 Talk (Russon, I.J., Wells, A., Grierson, C.E., Todd, C.D., Hazon, N. & Wendelaar Bonga, S.E.) “Aspects of the stress response of sea trout, Salmo trutta trutta (L.), to varying salmon lice, Lepeophtheirus salmonis (Krøyer), intensities when entering seawater”

282

SUMBAWS

Final Report

3. PUBLISHED PAPERS USTAN Todd, C.D., Stevenson, R.J., Reinardy, H. & Ritchie, M.G. (2005) Polyandry in the ectoparasitic copepod Lepeophtheirus salmonis despite complex precopulatory and postcopulatory mate-guarding.

Marine Ecology Progress Series 303, 225-234 NINA Thorstad, E.B., Økland, F., Finstad, B., Sivertsgård, R., Bjørn, P.A. & McKinley, R.S. (2004). Migration speeds and orientation of Atlantic salmon and sea trout post-smolts in a Norwegian fjord system.

Environmental Biology of Fishes 71, 305-311.

Finstad, B., Økland, F., Thorstad, E.B., Bjørn, P.A. & McKinley, R.S. (2005). Migration of hatcheryreared Atlantic salmon and wild sea trout post-smolts in a Norwegian fjord system. Journal of Fish Biology 66, 86-96.

NIFA Thorstad, E.B., Økland, F., Finstad, B., Sivertsgård, R., Bjørn, P.A. & McKinley, R.S. (2004) Migration speeds and orientation of Atlantic salmon and sea trout post-smolts in a Norwegian fjord system.

Environmental Biology of Fishes 71, 305-311.

Finstad, B., Økland, F., Thorstad, E.B., Bjørn, P.A. & McKinley, R.S. (2005). Migration of hatcheryreared Atlantic salmon and wild sea trout post-smolts in a Norwegian fjord system. Journal of Fish Biology 66, 86-96.

NCFS Thorstad, E.B., Økland, F., Finstad, B., Sivertsgård, R., Bjørn, P.A. & McKinley, R.S. (2004). Migration speeds and orientation of Atlantic salmon and sea trout post-smolts in a Norwegian fjord system.

Environmental Biology of Fishes 71, 305-311.

4. ACCEPTED MANUSCRIPTS NINA Sivertsgård, R., Thorstad, E.B., Økland, F., Finstad, B., Bjørn, P.A., Jepsen, N., Nordal, T. & McKinley, R.S. (2006). Effects of salmon lice infection and salmon lice protection on fjord migration Atlantic salmon and brown trout post-smolts. Hydrobiologia, in press.

Thorstad, E.B., Økland, F., Finstad, B., Sivertsgård, R., Plantalech, N., Bjørn, P.A. & McKinley, R.S. (2006). Comparing migratory behaviour of wild and hatchery-reared Atlantic salmon and wild anadromous brown trout postsmolts during the first stages of marine migration. Hydrobiologia, in press.

NIFA Sivertsgård, R., Thorstad, E.B., Økland, F., Finstad, B., Bjørn, P.A., Jepsen, N., Nordal, T. & McKinley, R.S. (2006). Effects of salmon lice infection and salmon lice protection on fjord migration Atlantic salmon and brown trout post-smolts. Hydrobiologia, in press.

Thorstad, E.B., Økland, F., Finstad, B., Sivertsgård, R., Plantalech, N., Bjørn, P.A. & McKinley, R.S. (2006). Comparing migratory behaviour of wild and hatchery-reared Atlantic salmon and wild anadromous brown trout postsmolts during the first stages of marine migration. Hydrobiologia, in press.

NCFS Sivertsgård, R., Thorstad, E.B., Økland, F., Finstad, B., Bjørn, P.A., Jepsen, N., Nordal, T. & McKinley, R.S. (2006). Effects of salmon lice infection and salmon lice protection on fjord migration Atlantic salmon and brown trout post-smolts. Hydrobiologia, in press.

283

SUMBAWS

Final Report

5. SUBMITTED MANUSCRIPTS ESRI Whelan, B., Aas, Ø., Curtis, J. & Dervo, B. (2006) The Economic impact of Atlantic salmon angling, commercial fishing and farming: a comparison between Ireland, Scotland and Norway. Submitted to Fisheries Management and Ecology USTAN Wells, A., Grierson,C.E., MacKenzie, M., Russon, I.J., Reinardy, H., Middlemiss, C., Bjørn, P.A., Finstad, B., Wendelaar Bonga, S.E., Todd, C.D. & Hazon, N. (2006) The physiological effects of simultaneous abrupt seawater entry and sea lice (Lepeophtheirus salmonis Krøyer) infestation of wild sea trout (Salmo trutta L.) smolts. Submitted to Canadian Journal of Fisheries and Aquatic Sciences Todd,C.D., Whyte, B.D.M., MacLean, J.C. & Walker, A.M. (2006) Ectoparasitic sea lice (Lepeophtheirus salmonis Krøyer, Caligus elongatus Nordmann) infestations of wild adult one seawinter Atlantic salmon Salmo salar L. returning to Scotland, 1998-2005. Submitted to Marine Ecology Progress Series Todd, C.D. (2006) Parasite interactions between wild and farmed salmonids: the application of genetic techniques to copepod sea lice (Lepeophtheirus salmonis (Krøyer), Caligus elongatus Nordmann) infesting Atlantic salmon (Salmo salar L.) and sea trout (Salmo trutta L.). Submitted to Journal of Plankton Research NINA Jepsen, N., Holthe, E. & Økland, F. (2006) Do saithe and cod “specialise” in smolt predation? Submitted to Fisheries Management and Ecology. Økland, F., Thorstad, E.B., Finstad, B., Sivertsgård, R., Plantalech, N., Jepsen, N. & McKinley, R.S. (2006). Swimming speeds and orientation of wild Atlantic salmon post-smolts in a Norwegian fjord system. Submitted to Fisheries Management and Ecology.

Whelan, B., Aas, Ø., Curtis, J. & Dervo, B. (2006) The Economic impact of Atlantic salmon angling, commercial fishing and farming: a comparison between Ireland, Scotland and Norway. Submitted to Fisheries Management and Ecology NIFA Bjørn, P. A., Finstad, B., Kristoffersen, R., Rikardsen, A. H. & McKinley, R. S. (2006) Differences in risks and consequences of salmon lice, Lepeophtheirus salmonis (Krøyer) infection on sympatric populations of Atlantic salmon, brown trout and Arctic charr within northern fjords. Submitted to ICES Journal of Marine Science.

RUN Wells, A., Grierson, C.E., MacKenzie, M., Russon, I.J., Reinardy, H., Middlemiss, C., Bjørn, P.A., Finstad, B., Wendelaar Bonga, S.E., Todd, C.D. & Hazon, N. (2006) The physiological effects of simultaneous abrupt seawater entry and sea lice (Lepeophtheirus salmonis Krøyer) infestation of wild sea trout (Salmo trutta L.) smolts. Submitted to Canadian Journal of Fisheries and Aquatic Sciences NCFS Bjørn, P.A., Finstad, B., Kristoffersen, R., Rikardsen, A.H. & McKinley, R.S. (2006) Differences in risks and consequences of salmon lice, Lepeophtheirus salmonis (Krøyer) infection on sympatric populations of Atlantic salmon, brown trout and Arctic charr within northern fjords. Submitted to ICES Journal of Marine Science.

284

SUMBAWS

Final Report

6. GREY LITERATURE REPORTS ESRI Whelan, B., Aas, Ø., Uglem, I., Curtis, J. & Dervo, B. (2006) Assessment of the socio-economic value of aquaculture and sport angling for wild salmonids in northwestern Europe. Implications for treatments for sea lice infestation. NINA Report 126.

NINA Finstad, B., Økland, F., Thorstad, E.B., Bjørn, P.A., Sivertsgård, R. & McKinley R.S. (2004) I smoltens fotspor. MRinfo, medlemsinformasjon for Norges jeger og fiskerforbund – Møre og Romsdal: 6-7.

Finstad, B., Økland, F., Thorstad, E.B., Diserud, O., Bjørn, P.A., Sivertsgård, R., Kristofferersen, R. & McKinley, R.S. (2005) I smoltens kjølvann. NINA Temahefte 31: 49-54.

Whelan, B., Aas, Ø., Uglem, I., Curtis, J. & Dervo, B. (2006) Assessment of the socio-economic value of aquaculture and sport angling for wild salmonids in northwestern Europe. Implications for treatments for sea lice infestation. NINA Report 126.

NIFA Finstad, B., Økland, F., Thorstad, E.B., Bjørn, P.A., Sivertsgård, R. & McKinley R.S. (2004) I smoltens fotspor. MRinfo, medlemsinformasjon for Norges jeger og fiskerforbund – Møre og Romsdal: 6-7.

Bjørn, P.A., Finstad, B. & Kristoffersen, R. (2004). Registreringer av lakselus på laks, sjøørret og sjørøye i 2003. NINA Oppdragsmelding 853.

Bjørn, P.A., Finstad, B. & Kristoffersen, R. (2005). Registreringer av lakselus på laks, sjøørret og sjørøye i 2004. NINA Rapport 60.

Finstad, B., Økland, F., Thorstad, E.B., Diserud, O., Bjørn, P.A., Sivertsgård, R., Kristofferersen, R. & McKinley, R.S. (2005) I smoltens kjølvann. NINA Temahefte 31: 49-54.

Holst, J. C., Heuch, P. A., Bjørn, P. A., Stien. A., Finstad, B. & Asplin. L. (2005) Sea Lice as a population regulating factor in Norwegian salmon: status, effects of measures taken and future management. Sluttrapport Norges Forskningsråd (prosjektnummer: 149791/S40).

RUN Russon, I.J., Walker, P.D. & Wendelaar Bonga, S.E. (2003) The effect of temperature and parasite development stage on the off-host survival of Argulus japonicus and Argulus foliaceus (Crustacea: Branchiura). 11th International Conference of the EAFP. Book of abstracts Walker, P.D., Haond, C., Russon, I.J. & Wendelaar Bonga, S.E. (2003) Evidence of blood feeding by the crustacean ectoparasite of fish, Argulus japonicus Thiele (Crustacea: Branchiura). 11th International Conference of the EAFP. Book of abstracts Walker, P.D., Schuurmans Stekhoven, F., Duijf, F., Russon, I.J. & Wendelaar Bonga, S.E. (2004) Are certain fish immune to the ectoparasitic crustacean Argilus japonicus Thiele (Crustacea:Branchiura)?

6th International Symposium on Fish Immunology. Book of abstracts Russon, I.J., Wells, A., Grierson, C.E., Todd, C.D., Hazon, N. & Wendelaar Bonga, S.E. (2005) Aspects of the stress response of sea trout, Salmo trutta trutta (L.), to varying salmon lice, Lepeophtheirus salmonis (Krøyer), intensities when entering seawater. 12th International Conference of the EAFP. Book of abstracts NCFS Finstad, B., Økland, F., Thorstad, E.B., Bjørn, P.A., Sivertsgård, R. & McKinley R.S. (2004) I smoltens fotspor. MRinfo, medlemsinformasjon for Norges jeger og fiskerforbund – Møre og Romsdal: 6-7.

285

SUMBAWS

Final Report

Bjørn, P.A., Finstad, B. & Kristoffersen, R. (2004). Registreringer av lakselus på laks, sjøørret og sjørøye i 2003. NINA Oppdragsmelding 853.

Bjørn, P.A., Finstad, B. & Kristoffersen, R. (2005). Registreringer av lakselus på laks, sjøørret og sjørøye i 2004. NINA Rapport 60.

Finstad, B., Økland, F., Thorstad, E., Bjørn, P.A., Sivertsgård, R., Kristoffersen, R. & McKinley, R.S. (2005) I smoltens kjølvann. In: Svenning, M.A. & Jonsson, B. (eds.) 2005. Kystøkologi: økosystemprosesser og menneskelig aktivitet. NINA Temahefte 31: 48 - 53 (In Norwegian) 7. WEBSITE/VIDEOS NINA Smolt telemetry video USTAN SUMBAWS website (www.st-andrews.ac.uk/~SUMBAWS) 8. WRITTEN ADVICE TO GOVERNMENTAL AND NON-GOVERNMENTAL COMMITTEES ALL PARTNERS

February 2006. Formal written response by SUMBAWS to the Scottish Executive, concerning the current draft Aquaculture & Fisheries Bill going through the Scottish Parliament (see Annex) USTAN 4 October 2004 Briefing of D. Wyman (Scottish Executive Rural Affairs Department) on progress to date on SUMBAWS programme and the outcomes and developments from the Utrecht Open Technical Workshop (below).

CFB December 2005. Advice to the Central Fisheries Board on the outcome of the Irish involvement in SUMBAWS and the possible implications for out migrating salmon smolts in aquaculture bays.

9. OPEN TECHNICAL WORKSHOPS ALL PARTNERS Utrecht Technical Workshop, Utrecht, Netherlands, 18 September 2004 “Population genetic structuring and evidence of chemotherapeutant resistance in sea lice: integrating the management of pest control with environmental impacts and national policy.”

RUN Wageningen Technical Workshop, Wageningen University, Netherlands, 21-25 April 2003 “Fish immunology: lectures and training in practical techniques” 10. OTHER ACTIVITIES ALL PARTNERS

Annual SUMBAWS/SEARCH Project Meeting, 25-26 October 2003, Vancouver, Canada Annual SUMBAWS/SEARCH Project Meeting, 16-18 September 2004, Utrecht, Netherlands Final SUMBAWS/SEARCH Project Meeting, 1-2 September 2005, Dublin, Ireland USTAN ISLM (Integrated Sea Lice Management) group meetings 01. July 2005 25. November 2005 286

SUMBAWS

Final Report

RUN PARITY Meeting, Berlin, Germany 19-23 February 2005 Presentation of SUMBAW-results to the PARITY project members (including Iain Russon) Final PARITY Meeting in Copenhagen, Denmark, 11 September 2005 Presentation of SUMBAWs results to the PARITY project members (including Iain Russon)

287

SUMBAWS

3

POLICY RELATED BENEFITS

3.1

European Dimension of the problem

Final Report

The salmon louse, Lepeophtheirus salmonis, an ectoparasitic crustacean which is common on salmonids in the marine environment, represents one of the most serious pathogens of both sea-farmed and wild Atlantic salmon. Salmon lice epidemics on farmed fish have chronologically followed the pattern of development of the salmon aquaculture industry in Norway, Scotland and Ireland. Similar epizootics have occurred subsequently on wild salmonids, especially sea trout, in all countries developing a major salmon aquaculture industry. Since 1970 Norway has built up the world’s largest salmon farming industry. It is a major employer and a wealth creator in often remote, coastal regions and is one of the country’s largest exporters.

Production has grown from a few thousand tonnes in 1980 to >400,000 t in 1999. Also Scotland and Ireland have experienced similar growth in salmon production and their production in 1999 was ~120,000 and 20,000 tonnes, respectively. Future perspectives show an increased growth in the aquaculture industry with reference to the continued over-exploitation of wild fisheries. The breakdown in traditional fisheries at sea, and the decline in the wild salmonid stocks in Norway, Scotland and Ireland, have led to a need for farmed production to make up the increasing EU consumer demand for fish as a healthy nutritious food source.

During the course of the SUMBAWS project the European dimension of this problem has been confirmed and emphasised by the results obtained. For example, microsatellite analyses of L. salmonis from samples in E. Canada, Scotland and N. Norway have shown no genetic differentiation of the parasite amongst three host species (Atlantic salmon, sea trout, Arctic char) or between samples from farmed and wild fish. The levels of gene flow/ larval cross infection between farm and wild salmonids and between the wild host species in the North Atlantic are sufficiently high as to prevent genetic divergence over a 6000 km range.

Essentially L. salmonis in the N. Atlantic comprises of a single panmictic population and therefore current and future research requires an equivalent international approach.

3.2

The European added value of the consortium

A unique specialist team has been created consisting of both university and government scientists from four European countries and the SUMBAWS research project would have been impossible to achieve at any single, national level. An increased understanding of the spread of salmon lice between farmed and wild fish is fundamental to developing management tools to increase the efficiency of the aquaculture industry and to minimize possible detrimental environmental impacts. The results generated from the consortium have produced a series of recommendations that will substantially deliver this aim. It is envisaged that the SUMBAWS consortium will form the nucleus of a new research proposal to be submitted under the imminent FP7 research programme. This new enlarged consortium will include partners from the EU SEARCH programme resulting directly from the clustering activities that were carried out during the SUMBAWS project.

3.3

Contribution to EU policies

The results and recommendations from SUMBAWS include objectively determined sea lice levels that initiate the stress response in both wild sea trout and farmed Atlantic salmon, establishing the migration routes of wild sea trout and salmon smolts and assessment of the relative sea lice infestation risks for different salmonid species. This information has been used to develop models for predicting the sea lice infestation pressure in the wild that can be used for both • •

optimal location of new aquaculture sites and in determining protective zones for wild salmonid populations within coastal waters.

Article 6 of the treaty establishing the European Community contains the principle that “environmental protection requirements must be integrated into the definition and implementation of Community policies and activities referred to in Article 3 in particular with a view to promoting sustainable development”.

Therefore this project has provided results and recommendations that directly reaffirm this principle, specifically in relation to the Common Fisheries Policy. Furthermore, it is anticipated that the SUMBAWS project will make a significant input to the EU meeting resulting from the joint SUMBAWS/SEARCH Workshop held in Utrecht in 2004. At this stage the date of this meeting is still to be confirmed but it is envisaged that the results and recommendations from the SUMBAWS project will be used as direct input to

288

SUMBAWS

Final Report

future policy and decision making in furthering the development of best practice in the sustainable management of sea lice in aquaculture and within an international context.

In socio-economic terms, both aquaculture and salmonid sport fishing are the major employers in peripheral regions of the EU where there are few, if any, alternative forms of employment. Peripheral regions located on the far north western extremities of the EU are remote both in terms of other parts of Europe and, at a local level, from other communities within these regions. These communities are fragile not only because of their sparse population, but also they are subject to incipient population drift towards cities and therefore are threatened by depopulation. Article 130a of the Treaty on European Union gives the European Community the objective of strengthening the economic and social cohesion of Europe. The SUMBAWS project has developed, for the first time, a socio-economic model that can be used to assess the interactions between the aquaculture and salmonid sport fishing industries. This is the basis for developing objectively derived policies relevant to rural development in peripheral EU regions.

289

SUMBAWS

4

Final Report

ANNEX

WRITTEN RESPONSES OF THE SUMBAWS CONSORTIUM TO THE PUBLIC CONSULTATION PERTAINING TO THE DRAFT AQUACULTURE AND FISHERIES BILL (https://www.scotland.gov.uk/consultations/environment/afb.asp) Submitted to the Scottish Executive, 22 February 2006 The document includes only those questions that were answered.

Q1. Do you agree that FRS/SEERAD would make the most suitable Regulator?

Within Scotland FRS has provided scientific and technical advice to government regarding the fish farming industry since its inception. FRS clearly is the obvious agency to whom the Executive should refer regarding regulation of the industry. However, as a funded agency of SEERAD there are questions regarding the extent to which FRS, as Regulator, can be perceived by the aquaculture industry, or the wider public, as being independent of SEERAD and hence the Scottish Executive.

Q2. Should there be an enabling power to permit the option of charging in the future?

It would appear logical that the possibility of charging should be enabled.

Q3. Should the Bill make provision for parasites in general, or restrict itself only to sea lice?

It would appear scientifically expedient that the Bill should not be restricted to sea lice alone, but should embrace all parasites. As other aquaculture industries (e.g. intensive cod farming) develop it is quite possible that other species of parasites might become problematic to wild and/or farmed fish species. The colloquial term “sea lice” generally is used to refer to Lepeophtheirus salmonis (the salmon louse) and Caligus elongatus. These are copepods of the Order Siphonostomatoida, of the Family Caligidae. Many scientists therefore refer colloquially to “caligid sea lice”. This usage specifically confines the term to caligids, but there are many species of copepods in that family which are parasitic on marine fishes. As a further illustrative example of the necessity for scientific conciseness in categorizing or referring to parasites, we would refer you to Lernaeocera branchialis (Family Penellidae) and Clavella adunca (Family Lernaeopodidae). L. branchialis is a serious economic pest species on wild gadid fish of the cod family (notably cod, whiting and haddock). Clavella adunca is also a common parasite of these same fish. Neither Lernaeocera nor Clavella would be referred to as “sea lice”, but both are copepods of the Order Siphonostomatoida. As cod farming develops in Scotland it is highly likely that these, and many other species of parasites, may well cause problems for farmed and wild fish alike. Neither C. adunca nor L. branchialis will impact salmonids, but the potential for interaction between farmed cod and wild gadid species is clear. The legislation presumably has to take account of possible future developments in aquaculture, including new cultured fish species, and can overcome the potential for taxonomic confusion and non-caligid copepods becoming pests, by the inclusion of all metazoan parasites under the Bill.

Regarding the zero ovigerous lice objective, it should be emphasized that it is precisely this, an objective.

Scottish Quality Salmon’s Area Management Agreement Concordat states that “The objective of continuously achieving zero ovigerous lice on farmed stocks, especially during the period prior to and during wild smolt migrations, has been agreed.” This is not, therefore, a requirement or an obligation.

Strictly speaking it probably is impossible to legislate for precisely zero ovigerous sea lice on a farm. For example, one ovigerous louse in a sample of 30 fish from a farm cage is not zero. The Scottish Executive could beneficially assess the Norwegian regulations and the mechanisms by which they are successfully enforced. In Norway, at temperatures >4ºC the numbers of lice at each farm have to be counted every second week and reported to ‘Mattilsynet’ – an overseeing government agency – by the 7th of the month following. Over the period December 1 to July 1 the numbers of lice shall not exceed 0.5 females or >5 females and mobile stages. On exceeding this limit the farmer has to delouse. (In the Troms and Finnmark districts of northern Norway this period extends from November 1 to July 1). From July 1 to December 1 the numbers of lice shall not exceed 2 females or >10 females and mobile stages. Again, on exceeding this limit the farmer has to delouse. (In Troms and Finnmark this period is from July 1 to November 1). If these regulations not are followed ‘Mattilsynet’ can impose financial penalties on the fish farmer, within certain 290

SUMBAWS

Final Report

guidelines. This fine increases daily and starts from the day the fine was applied until the farmer has started delousing.

In Ireland there is a trigger level range for treatment set at 0.3-0.5 ovigerous females per fish in springtime, coincident with the wild smolt migration. If these levels are breached the farmer must treat, but having lice levels in excess of the trigger is not of itself an offence; it is failure to treat that is an offence.

One especially severe criticism of setting a uniform, nation-wide trigger level for louse treatment based on averages per fish is that this can never take account of the numbers of cultured fish held within a given sea loch. If the ovigerous average per fish is maintained, and yet numbers of farmed fish or production tonnage increases, the total number of lice larvae produced within that sea loch will rise. Precautionary protection of wild salmonids is perhaps more likely to be achieved by setting an allowable larval lice production total for a sea loch. Such an approach also would circumvent any legal dispute as to what constitutes â&#x20AC;&#x153;zero ovigerous sea liceâ&#x20AC;?. The mean number of ovigerous lice per fish that becomes the trigger level for treatment would, therefore, vary among sea lochs and depend upon the tonnage consents and/or numbers of fish placed at sea in cages. Although not as simple to apply as a single, standard value the precedent could be cited of SEPA setting production tonnage and discharge consents on a site by site basis.

From a precautionary perspective it would appear essential for a zero ovigerous lice objective to be realised in Scotland, and some companies are especially resolute in their attempts to minimize their potential impacts. However, even for a company or individual farm wishing to achieve this desirable objective there may be real economic or regulatory reasons why this is not possible. It may be, for example, that a siteâ&#x20AC;&#x2122;s discharge consent has already been taken up in treatments so that further treatments cannot be deployed.

There may be an arguable case for less stringent authorisation procedures to enable a wider range of chemical treatments to be licenced for use on fish farms, and/or for discharge consents to be temporarily increased to permit companies to achieve zero ovigerous lice. It also has to be emphasized that wild salmonids in Scotland are threatened not only during the wild smolt migration period in springtime (AprilJune). For example, juvenile sea trout on the west coast of Scotland may spend their entire period of marine residence of several months in sea lochs and coastal waters before returning to over-winter in freshwater.

Unless ovigerous sea lice are controlled throughout the period from late winter (February) to early winter (October) there will remain the potential for serious detrimental effects of sea lice larvae of farm origin impacting wild salmonid stocks.

Q4. Should the Regulator have both advisory and enforcement functions?

We agree that the Regulator should have both advisory and enforcement functions. The present system of oversight and enforcement in Norway seems to be effective in that veterinarians (Mattilsynet) monitor sea lice levels and have the power to enforce treatment if farmers fail to comply with treatments as indicated by the nationally agreed trigger levels of sea lice.

Q5. What powers should the Regulator have as regards the inspection of data and investigation of potential parasite problems on farms?

The obligations and powers of the Regulator as set out in paragraphs 15-17 appear to be appropriate.

Questions would, however, have to be resolved concerning monitoring data. At present these are treated by the industry as being commercial-in-confidence (cf. Norway). Data may well be presented in confidence to local Tripartite Working Groups but are not public domain. It is highly likely that some members of the public might wish to access farm lice records under freedom of information legislation.

Q6. Should the Regulator have the power to direct treatment?

Whilst it is appropriate that the Regulator should have the power to decide upon and direct treatment of farm fish (for the welfare of farm fish, and as a precaution to protect wild stocks), clear difficulties may arise with SEPA if the farm has already expended its discharge consent for treatments.

291

SUMBAWS

Final Report

Q7. Should the Regulator have the power to arrange treatment through a third party contractor where a direction to treat has not been complied with?

Yes. When animal welfare is compromised there may be insufficient time to resolve the problem and the ability of the Regulator to enforce a decision rapidly would be expedient and allow the necessary urgency of remedial action.

Q8. Should the Regulator have the powers to direct treatment for notifiable diseases?

Yes Q9. Should escapes proposals apply to all the farm types outlined above?

Yes Q10. Should shellfish farms and restocking hatcheries be exempted?

Yes Q11. Do you agree that the Regulator should have powers to inspect and direct with respect to preventing escapes?

Yes Q12. Do you agree that the Regulator should have a role in improving containment to prevent fish escapes?

Yes Q13. Should the Executive introduce a strict liability offence for escapes from fish farms?

Yes Q14. What elements should be addressed in containment plans?

Q15. Do you have any views on the above proposals?

As extensively discussed at the recent NASCO Symposium in Bergen (“Interactions between aquaculture and wild stocks of Atlantic salmon and other diadromous fish species: Science and Management, Challenges and Solutions” Bergen, Norway 18-21 October 2005) escapes of farmed salmon are widely perceived as being a major threat to wild salmonids. One of the difficulties discussed was how one defines an ‘escape’.

Trickle losses and minor incidents might not fall within a legal definition of ‘escape’ and yet can be numerically important over a period of time. How will this be accounted in the legislation?

Q16. Do you agree that the Regulator should have powers to investigate escapes and suspected escapes from fish farms whatever the source of the information?

Yes Q 17. What data, in addition to the production survey data, do you believe is appropriate to submit?

Q18. Do you agree that financial assistance be given to fish farm operators to relocate where there is a clear environmental benefit in doing so?

The Crown Estate presently leases farm sites to companies. It would be relatively straightforward to request relocation to another (unused) site already leased by the same company but presumably farms will 292

SUMBAWS

Final Report

not be permitted to relocate to unlicenced sites. Individual farm sites also carry production tonnage and discharge consents which may not be the same as the site to be closed. Site fallowing rotations might not be conducive to a relocation request. Relocation can therefore only be undertaken within the present legal framework with close co-operative action by the numerous disparate bodies with a regulatory interest.

Financial assistance with relocation would appear to be justifiable.

Q19. Do you agree that the Scottish Ministers should have powers to close fish farms where there is a clear public interest to do so and where owners are not in a position to relocate?

Yes Q20. Under what circumstances might it be useful for the Scottish Ministers to have discretionary powers to pay compensation to fish farm operators?

It would appear appropriate for Scottish Ministers to approve compensation but this should be extended only to those companies that can be demonstrated to comply with all of the accepted industry standards and practices.

Q21. Do you agree with the need to regulate live fish movements out of, and between, marine farm management areas?

Yes Q22. Do you agree there is no general need to restrict live fish movements between freshwater fish farms?

Yes Q23. Is the proposed power to bring in a national standstill provision in case of a novel disease appropriate?

Yes Q24. Should the Regulator be empowered to licence the transfer of fish by wellboats in Scotland?

Yes

293