CROJFE - Volume 32, Issue 2

32

Issue 2

2011

2

Uvodnik – Editorial

FORMEC – Progress Made and the Way Ahead

Dear colleagues, As a member of the editorial board of the Croatian Journal of Forest Engineering (CROJFE) and the president of the scientific network FORMEC (International Symposium on Forestry Mechanization), it is a great honor and pleasure for me to present the current CROJFE edition. In anticipation of the 44th Symposium »Pushing the boundaries with research and innovation in forest engineering«, we have succeeded for the first time in reviewing selected FORMEC paper contributions and publishing them in a CROJFE issue. This will ensure a significant improvement in the quality of the meeting, as well as benefitting the whole scientific community in forest engineering.

FORMEC was founded in 1966 in the former Czechoslovakia. The basic idea of the meeting was to serve as a bridge between East and West and bring together scientists to discuss current challenges related to the mechanization of forestry. The founding concept of FORMEC as a forest engineering communication platform remains, but some specific changes have been made over the years. The official language of the symposium has changed from German to English in order to increase international participation and to encourage the attendance of young researchers and practitioners. This has had a significant impact on the number of participants as well as the quality of their contributions. In the beginning, participation fluctuated between 30 and 40 researchers.

This year in Graz, Austria more than 200 participants from 38 different countries have been registered. There is already keen interest in the future meetings scheduled for Dubrovnik, Croatia in 2012 and Dresden, Germany in 2013. FORMEC has be-

Croat. j. for. eng. 32(2011)2

come one of the most important scientific meetings in forest engineering. This is recognition of, and at the same time a mandate for, continuous improvement in the future – much in keeping with the adage: »Who stops going forward, stops being excellent«.

Forest engineering research deals with the analysis and design of production systems (from forest to mill) for the material and energy utilization of wood resources. It includes the evaluation of impacts on the environment and humanity. The research concentrates on the direct applicability of results for the forest industry and provides practical solutions to current problems (applied research). The challenge is to direct development of forest engineering methods towards system and network oriented approaches.

FORMEC, as a scientific platform, will contribute to this development. In this context the encouragement and promotion of young scientists will play an important role. At the moment we have a lack of young and highly qualified researchers in forest engineering, and professions can only be developed with enough young talent.

Finally, I would like to thank all those who have helped ensure that the meeting in Austria will succeed and who made it possible to finish the current CROJFE issue in time. We look forward to the valuable contributions that all participants can make at this meeting.

Karl Stampfer

469

Original scientific paper – Izvorni znanstveni rad

Improving Forest Operations Management through Applied Research Mark Brown, Martin Strandgard, Mauricio Acuna, Damian Walsh, Rick Mitchell Abstract – Nacrtak A great challenge of applied research is translating results into industry innovation. Increasingly, forest managers do not have the capacity to interpret research results but prefer to be presented with tools based on the research results that can be readily implemented. The Cooperative Research Centre (CRC) for Forestry, based in Australia, has focused on delivering research results to industry partners in novel ways that can be easily applied in the field.

This paper discusses six approaches taken by the CRC to help transfer applied research results to industry, including basic benchmarking curves for feller-bunchers, a toolbox for operational machine evaluation, a productivity model, a method to predict productivity with existing data, a guide for effective use of onboard computers and an optimised transportation planning tool. For each approach the paper will discuss how these approaches were developed and applied with industry collaboration.

Keywords: innovation, implementation, efficiency, operations management

1. Introduction – Uvod When working in the field of applied forest research, particularly collaboratively with industry, presenting results in a way that industry can readily make use of is as important as ensuring quality and relevant research. A classic example is Skogsforsk, a world leading applied forest research organisation based in Sweden, whose mission is to supply applicable knowledge, services and products to enhance the competitiveness of Sweden’s forest industry.

Skogforsk’s focus on industry collaboration and delivering research results in a usable form for industry has decreased real harvest costs in Sweden over the past 20 years, with tools like FlowOpt and applied new work methods adapted to new technology (Fryk and Radstrom 2011). To provide results in a usable form to drive innovation, it is necessary for the research provider to understand the value their industry can add to research knowledge, issues of change management and the innovation process.

Translation of research results to innovation requires effective communication to industry to allow both the research provider and industry to play active roles in implementation.

Croat. j. for. eng. 32(2011)2

Private industry is one of the most effective and valuable avenues for science to disseminate its knowledge more broadly within society. Industry is motivated to satisfy its customers’ expectations of good value products and services to remain competitive and can be very effective at converting research knowledge to useable products and services (Cribb and Hartomo 2002). Research can also assist with development of more cost-effective delivery of existing products and services.

Implementing research inevitably involves making change within an organisation. Effecting change requires that management has to understand and believe in the projected impact of the change and effectively communicate that to all involved or impacted (Grover et al. 1995). This management-level understanding can be facilitated by having all stakeholders actively engaged through the research process (Dentoni and English 2011). This can be particularly effective for a small focused industry segment with a clear set of influencing variables, but difficult to do effectively for a broader industry group subject to a wider range of influences.

For the broader industry group, the experience of the CRC for Forestry has been engaging stakeholders

471

M. Brown et al.

Improving Forest Operations Management through Applied Research (471–480)

as much as possible in setting the research direction, within the bounds of »good research«; it is valuable to ensure relevance, but needs to be supported by creative delivery of the research results to facilitate uptake. Providing a simple report often locks the results into a fixed set of assumptions that may not be directly applicable to a specific company and limits trust in the results. Adapting the research results into flexible transparent tools for industry can facilitate understanding of not only the actual research results but, more importantly, how the research results apply in their particular case and what the projected impacts will be within their business.

1.1 CRC for Forestry: the context of collaborative research in Australia – Centar za istra`ivanja u {umarstvu: pozadina istra`iva~ke suradnje u Australiji The CRC concept program is funded by the Commonwealth Government of Australia to facilitate a collaborative approach between research providers and industry. Through a competitive process an industry group, representing a significant portion of the national industry, comes together with research providers to develop a research program to address particular issues that are common to, and will help improve the competitiveness of, the Australian industry, and more specifically the collaborating partners. A CRC has defined time and resources to meet a set of objectives that are agreed and managed by all the partners. Where an ongoing need for a sustained research effort is identified in the area covered by the CRC, an ongoing institute can be established based on the foundation created by the CRC and its industry partners.

The CRC for Forestry is a very successful CRC currently nearing the end of its third seven-year round of funding. In an independent third year review of the CRC for Forestry it was found that key outputs have or have the potential to deliver in excess of 80% internal rates of return (IRR) for the partners (Agtrans 2008). The success of the CRC for Forestry is further highlighted in its application for a five year extension submitted this year, where more than 80% of the forest industry in Australia are supporting the bid and industry contribution to the research program is increased. Building and maintaining strong industry and Commonwealth support has required that the CRC’s research is not only world standard, but also relevant and applicable »on the ground«. This has always involved strong and regular interaction with industry and in recent years a greater focus on how the research results are presented to industry, with greater direct interaction

472

and development of tools that fit their management needs and support uptake.

This paper explores six case studies of delivering collaborative research results from the CRC for Forestry through different tools to facilitate their use for innovation within industry. The cases show how the approach of developing a tool, in addition to formally reporting results, has facilitated industry uptake. It is also shown that the tool need not be particularly complex to effectively demonstrate research results and meet industry needs. Where tools are sufficiently mature and have been used in operational settings by industry, and where the industry partner has been willing to share their results, the impact achieved from the use of the research is noted.

2. Cases – Slu~ajevi 2.1 Managing productivity – benchmark production curve for feller-bunchers Krivulje proizvodnosti za »feller buncher« 2.1.1 Background and research – Pozadina istra`ivanja Many studies have shown feller-buncher productivity to be highly dependent on average tree size (e.g. Gingras 1988, Visser and Stampfer 2003). Other important factors include tree spacing (Long et al.

2002), slope (Oliveira et al. 2009) and undergrowth density (Granskog and Anderson 1981). Productivity data (m3/productive machine hour (PMH)) from

Fig. 1 Estimate of trees cut per PMH0 for field evaluation Slika 1. Procjena broja posje~enih stabala pomo}u efektivnoga vremena rada za vrednovanje na terenu Croat. j. for. eng. 32(2011)2

Improving Forest Operations Management through Applied Research (471–480)

seven short-term studies of feller-bunchers clearfelling E. globulus plantations in Western Australia, and from six published overseas studies clearfelling various species, were plotted against average tree size (m3). Only hot saw and shear head feller-buncher results were used. All delays were excluded (PMH0). PMH0 productivity is preferred, as it is easiest to collect in the field by users of the benchmark. A curve fitted to the data from the 13 different field studies gave a good fit (r2adj = 93%). Estimated trees cut per PMH0 was derived from the regression and plotted against average tree size (m3). Lines at ± 10% were added for guidance when using the benchmark curves as shown in Fig. 1. With a strong relationship presented in the curve, armed only with average tree size and a watch, managers and machine owners can quickly determine if the machine is operating at an expected level and take action where appropriate.

2.1.2 Strategy for transfer to industry – Strategija prijenosa znanja industriji The feller-buncher benchmark curves were distributed to harvest contractors and other potential users by: Þ CRC Bulletin – key research findings and the benchmark curves. Distributed to client organisations (Strandgard and Mitchell, 2010), Þ Meetings with client organisations, Þ Article in the Australian Forest Contractors Association newsletter, 23 November 2010 (http:/ /www.afca. asn.au/afca%20log/Log_23_November_ 2010[1][1].pdf), Þ Distributed directly to harvesting contractors using email or in face-to-face meetings.

2.1.3 Implementation by industry and impact Primjena u industriji Contractors and machine operators use the benchmark curves by estimating average tree size from estimates of merchantable volume per hectare and stems per hectare provided by the forest manager. Average tree size is then used to estimate the expected number of trees cut per PMH0 from the benchmark line.

This figure is compared with actual performance to identify and remedy underperformance and reward over-performance. For a typical West Australian E. globulus stand, a feller-buncher operating 10% less productively than expected would cost ~ 40 000 $ more per year to cut the same volume of wood as a machine operating at the benchmark rate. Requests have been received from both contractors and forest managers to have similar curves produced for other common machines.

Croat. j. for. eng. 32(2011)2

M. Brown et al.

2.2 Equipping operations to do effective evaluation – Machine Evaluation Toolbox Alat za procjenu proizvodnosti rada i strojeva u {umarstvu 2.2.1 Background and research – Pozadina istra`ivanja Being able to effectively measure and understand machine performance is critical to having efficient mechanised forest operations, but the need for machine evaluation is infrequent and rarely a core skill for managers. The challenge for time-constrained managers and contractors is maintaining their basic skills and a consistent approach when the need arises for machine evaluation to support important management decisions.

Initially the CRC for Forestry research team created a standard framework for machine evaluation in Australia, based on the recognised international framework from the International Union for Forest Research Organizations (IUFRO) (Bjorheden et al.

2000), to guide the research program and ensure internal consistency. What became immediately clear from the introduction of the framework was that the industry would need assistance to interpret and implement it to maintain the consistency of approach established by the CRC for Forestry. To meet this need a series of workshops were provided to the industry.

Feedback from the workshops was that they provided an excellent introduction, but industry participants typically did not have machine evaluation as a primary focus when they returned to their jobs. When the need for evaluation arose, the workshop knowledge had been lost and needed to be refreshed. Based on this feedback the CRC for Forestry Machine Evaluation Toolbox was created, providing industry with instructions for four common evaluation methods, field forms, basic calculators and report generators for the most common forest machines. The tools are available in a simple software format using intuitive menus to access the appropriate set of directions and tools for a given evaluation on a given machine type.

Evaluations done with the toolbox are consistent with CRC for Forestry methods, making the results comparable with research results and benchmarks.

2.2.2 Strategy for transfer to industry – Strategija prijenosa znanja industriji As part of the CRC’s collaborative approach to industry the toolbox was directly distributed to industry partners for initial review and was updated based on feedback received. As part of a broader industry engagement strategy, lunch-hour workshops on CRC for Forestry research outputs have been offered to industry partners at their locations for their

473

M. Brown et al.

Improving Forest Operations Management through Applied Research (471–480)

staff – the machine evaluation toolbox has been the most requested topic. It has also been directly implemented by contractors who co-operated in CRC for Forestry trials and wanted to expand on the trial work.

2.2.3 Implementation by industry and impact Primjena u industriji Though early in its introduction, the toolbox is seeing use within industry to address questions about productivity as it relates to different operating conditions and machine utilisation. Initial use has been limited to contractors seeking to improve specific knowledge for the management of their operations, contract tendering and negotiations.

2.3 Predicting productivity from harvester stem files – Predvi|anje proizvodnosti iz podataka prikupljenih radom harvestera 2.3.1 Background and research – Pozadina istra`ivanja Creating a harvester productivity model has traditionally been a time-consuming task using stop watches, or video cameras and data loggers, to estimate harvest cycle times. Model accuracy is limited by the funds and time available.

Harvester productivity studies can be heavily influenced by operator performance differences (Ovaskainen et al. 2004). This effect can be reduced by conducting trials of multiple operators and machines or combining study results to create generalised productivity models. Few such models have been developed (e.g. Nurminen et al. 2006, Spinelli et al.

2010), which reflects the significant time and effort required.

Recently the use of StanForD stem file records generated by single-grip harvesters was trialled by the CRC for Forestry to estimate harvest cycle time.

Stem files record the time and date a tree was cut and its log product information. Differences between times recorded in consecutive stem files can be used to estimate cycle time and develop a productivity model. Compared with current methods, using stem files has the potential to rapidly produce productivity models as data are readily available. The drawbacks are that there is no onsite observer able to note unforeseen changes in trial conditions (Nuutinen et al. 2008) and elemental time information is not recorded.

Four trial sites in Pinus radiata plantation clearfell operations across southern Australia were used to compare stem file productivity models with those produced using time and motion studies. The stem file data were filtered to remove cycle times with large delays and trees with broken tops and

474

multiple leaders. At all sites, the models produced by the stem file approach were not statistically significantly different to those produced with time and motion studies, which suggests the stem file approach can be applied more widely when stem files are collected under consistent, known stand and site conditions.

2.3.2 Strategy for transfer to industry – Strategija prijenosa znanja industriji The stem file productivity model tool is still under development. The planned tool will import stem files and apply user-selected filters to remove cycle times with large delays and trees with broken tops and multiple leaders. Models will be generated from the filtered data for typical model forms used for harvester productivity models (linear, logarithmic and power). The tool will be distributed electronically through CRC for Forestry networks, workshops and direct interaction with industry stakeholders.

2.3.3 Implementation by industry and impact Primjena u industriji The tool is intended to be used by harvesting contractors to estimate rates and manage operations, by forest managers to plan harvest schedules and estimate harvest costs and by researchers to evaluate harvester performance. Savings are mainly expected from reduced time spent developing models, which may amount to thousands of dollars per model.

There may be further savings from increased availability of productivity models, but these are currently difficult to quantify.

2.4 Guide for improved operations management with onboard systems – Vodi~ za upravljanje radnim operacijama pomo}u ugra|enoga sustava 2.4.1 Background and research – Pozadina istra`ivanja Experience in Europe and North America has shown that effective use of onboard computing equipment in forest harvesting machines can produce gains of up to 30% in availability, utilisation and productivity (e.g. Jamieson 2004). In Australia, lack of information about onboard computer capabilities and implementation has been identified as a key barrier to their uptake. The only area where there has been widespread uptake is bucking optimisation in P. radiata plantations, which has been driven by the requirements of forest managers.

Results of trials installing and testing onboard computing equipment in Australia have been used to develop a selection and implementation guide.

Croat. j. for. eng. 32(2011)2

Improving Forest Operations Management through Applied Research (471–480)

Users of the guide select their problem or concern from a list and are taken to a description and examples, key points and case studies of the onboard computer(s) best suited to address that concern. Basic steps needed to implement each category of onboard computer including installation; setup and basic information on the use and analysis of the data collected by the computer are also covered in the guide.

An initial study identified a number of onboard computing systems with the potential to improve harvesting machine performance. Three trials were established to test the most promising of these onboard computing systems across a range of Australian forest harvesting systems and forest types (natural forest, P. radiata and E. globulus plantations). An additional trial was established when the opportunity arose to test a pre-release version of FPInnovation’s FPDat (http://fpsuite.ca).

2.4.2 Strategy for transfer to industry – Strategija prijenosa znanja industriji Forest harvesting contractors are the primary intended users of the onboard systems selection and implementation guide, with forest managers as other potential users. The guide was delivered by: Þ a series of industry workshops across southern Australia, Þ distribution of printed copies of the guide, Þ an online, interactive version of the guide (http:/ /www.crcforestry.com.au/publications/downloads/ CRC-Onboard-Computers-WEBLINK-21_031.pdf).

2.4.3 Implementation by industry and impact Primjena u industriji There is widespread potential to use onboard systems in Australia to identify inefficiencies in harvesting systems and to better understand baseline performance to enable more accurate harvest rate setting.

One of the case studies used in the development of the guide identified savings of approximately 100 000 $ per year from the application of onboard computing systems, for an operation that represents less than 0.5% of the annual harvest in Australia, while another case study (unpublished) determined the main delay causes in an in-field chipping operation.

2.5 Accurate operational predictions of productivity and cost – ALPACA Operativno predvi|anje proizvodnosti i tro{kova pomo}u sustava ALPACA 2.5.1 Background and research – Pozadina istra`ivanja While several comprehensive productivity and costing models have been developed for specific reCroat. j. for. eng. 32(2011)2

M. Brown et al.

gions, few are available in the literature and on the Internet. In Canada, FPInnovations developed ProVue (FPInnovations 2005), based on 30 years of field studies, to help contractors pick the best harvesting system for unique forest conditions or to predict operating productivities when they move to new sites. Contractors search for productivity and cost data by specifying the components of a harvesting system and by terrain conditions, area and study period. In South Africa, the Department of Forest and Wood Science (University of Stellenbosch), in conjunction with the Forest Engineering Programme (Institute for Commercial Forestry Research South Africa) developed the South African Harvesting & Transport Systems Costing Model for both large and emerging contractors (University of Stellenbosch and Institute Forestry Research South Africa 2009). In the US, the Forest Operations Research Unit has developed two spreadsheet-based tools: the Machine Rate Calculator, to analyse system balance, production rates and cost, and the General Ground-Based Harvesting Systems Analysis to estimate total system cost based on stand characteristics, felling, skidding, processing, loading, roads and hauling, and operational factors (USDA Forest Service 2008).

The current version of the Australian Logging Productivity and Cost Assessment (ALPACA) tool was developed as part of a collaborative project between the CRC for Forestry and Oregon State University. It is based on more than 200 production studies on felling, yarding, processing, chipping and loading, and it provides two types of analyses: harvesting system forecast and single machine. In addition to data on times for extraction cycles, delays, etc., data were also assembled on felling breakage, average piece size, payloads, etc. for various types of terrain and felling patterns. Harvest systems include cut-to-length at stump or at roadside, in-field chipping with debarking at stump or at roadside, and cable-logging operations. Using the research results at the core of the tool it predicts daily production, rate per tonne or m3, total harvesting costs and number of days to log the harvest block.

2.5.2 Strategy for transfer to industry – Strategija prijenosa znanja industriji Industry partners were provided with a brief workshop on the ALPACA prototype to allow them to trial the software and provide feedback for changes and improvements. The workshop consisted of a general description of the tool, the harvest systems included, calculation of productivity and costs, the data inputs required and several demonstrations and exercises with the participants. As the tool is still un-

475

M. Brown et al.

Improving Forest Operations Management through Applied Research (471–480)

der development, it has also been distributed to logging contractors and operational staff to get their comments and feedback.

2.6 Optimised transport and logistics planning – FastTRUCK – Pobolj{anje prijevoza drva i logisti~koga planiranja sustavom FastTRUCK

not well-suited to address the specific forest transport issues in Australia.

FastTRUCK’s first version was developed for in-field chipping operations. It is a Microsoft Windows®-based system developed in Visual C++® and uses inputs from the existing parameters of the transport operation to generate a range of alternatives to determine the optimal (or near optimal) operating scenario. Optimal truck schedules are created by FastTRUCK using a simulator and a simulated annealing algorithm. The aim of the system is to minimise total transportation costs. It considers travel loaded and unloaded time, stood down time and fixed costs. FastTRUCK reports the optimal number of trucks required for the operation, total transportation cost, total volume of chips hauled to dumpers, average truck utilisation, average truck waiting time and average loaded running percentage (travel loaded/total travel distance) while maintaining a user-defined chipper utilisation. Detailed results by truck are exported to MS Excel® and include total time, total cost, trips to dumpers, waiting time, utilisation, running loaded percentage, arrival times at forests and dumpers, and the optimal schedule for one day.

A second version for more complex transport operations (multiple products, truck configurations, destinations, etc.) has been developed in Java® (Eclipse IDE) and is being tested with an Australian company in P. radiata plantation haulage operations.

2.6.1 Background and research – Pozadina istra`ivanja

2.6.2 Strategy for transfer to industry – Strategija prijenosa znanja industriji

Several optimal transportation models developed internationally and their corresponding solution approaches were investigated before developing and implementing FastTRUCK. In Chile, a computerised system called ASICAM has been in use since 1990. It is a simulation system embedded with a heuristic solver which produces a complete working schedule for one day (about 100 trucks) in only a few minutes (Epstein et al. 2007). ASICAM has led to reductions in transport costs of between 10% and 20% (Weintraub et al. 1996). In Sweden, Skogforsk developed a system called FlowOpt (Forsberg et al. 2007). The system integrates GIS data with a database and uses a heuristic approach based on a tabu search algorithm.

Tests have reported savings of 5% – 20% compared to manual solutions. Other approaches include a Finnish system called EPO that deals with all stages from strategic to operative planning (Linnainmaa et al. 1995) and a solution approach reported in Palmgren et al. (2004). These are all very effective tools, but in many cases their availability was limited to the organisations that developed them or they were

An industry bulletin on FastTRUCK has been published by the CRC for Forestry (Acuna et al. 2010).

This report presented an analysis of the factors that affect the transport efficiency of in-field chipping operations. Program industry partners were also provided with a brief workshop on the FastTRUCK prototype to allow them to trial the software and provide feedback for changes and improvements. The workshop consisted of an introduction to timber transport, a description of truck dispatching and scheduling, presentation of some of the potential quantitative and qualitative benefits associated with the adoption of optimisation tools and a demonstration of FastTRUCK for in-field chipping operations.

2.5.3 Implementation by industry and impact Primjena u industriji Harvest managers and contractors can use ALPACA to predict harvesting costs and productivity of different systems for specific forest conditions.

Inputs are relatively easy to collect and include among others: harvest unit details (e.g. total compartment area), stand details (e.g. merchantable volume), harvest type (e.g. clearfell), species, harvesting system (e.g. cut-to-length) and extraction details.

In addition to overall costs and productivity other outputs include daily production costs by activity (felling, skidding, etc.) and number of days to log the compartment. The tool could eventually be used to plan the harvest, match harvesting systems to forest conditions, assess the impact of operational factors on productivity and cost, as a control tool allowing managers to compare actual performance against predicted performance, and allocate logging crews to harvest units.

476

2.6.3 Implementation by industry and impact Primjena u industriji During 2009 – 2010, the CRC for Forestry worked with an industry partner as it planned a new woodchip export facility in Western Australia. Working with known annual in-field chipping harvest volumes and constraints on truck access times to the Croat. j. for. eng. 32(2011)2

Improving Forest Operations Management through Applied Research (471–480)

port, it was important to determine the required level of infrastructure at the port.

FastTRUCK proved to be a valuable tool in evaluating the required level of receiving capacity at the proposed woodchip terminal in Western Australia.

Based on the FastTRUCK results, a significant capital expenditure at the new facility was avoided as it was determined only one dumper would be required rather than the two initially planned. The CRC work also identified that there were additional savings to be made, with the potential for the planned operation to reduce annual costs by 10% through improved harvest and truck scheduling. Applied to the roughly 1 million tonne annual harvest transported from in-field chipping operations around Australia each year, FastTRUCK has the potential to deliver more than 1.5 million $ in annual savings.

3. Discussion – Rasprava These six cases demonstrate a range of strategies used by the CRC for Forestry to lift research results from the page and develop them into tools that are better suited to support industry to engage in the innovation process. In all cases additional effort was needed to translate research results into practical »field-ready» tools. This effort ranged from simply reformatting and amalgamating data to make it easier for industry to receive (e.g. the feller-buncher benchmark curve), to complex software programming (e.g. FastTRUCK). What is important is that the tools were developed collaboratively with end users, to meet industry needs and expectations. Such ownership by end users increases the likelihood of research results playing an active role in industry innovation. The success of the tools rests on choosing the right approach to meet industry need, and ensuring they are fit for-purpose – e.g. not creating complex solutions where a simple one meets the need.

This is best managed through ongoing communication and collaboration with industry.

While the level of complexity and resources vary, all six of the tools presented in this paper, with the exception of FastTRUCK, were developed and delivered with commonly available computer desktop tools. In addition to making the development of the tools simpler and limiting resources required, the use of common computer software facilitated the transfer to industry.

It is also important to understand that the development of tools to deliver research results to industry in no way replaces proper reporting of research results. In many cases the research results behind the six presented cases have been or will be published or presented at conferences (Strandgard and Mitchell Croat. j. for. eng. 32(2011)2

M. Brown et al.

2010, Acuna et al. 2010, Strandgard 2009, Strandgard, 2010, Acuna and Ghaffariyan 2011). Reports are an important step to properly document the research and expose the results to proper scrutiny and review by the broader forest research community. While the CRC for Forestry experience has been that reports are not an effective tool to transfer results to industry, they remain an effective mechanism to share results within the research community.

The CRC for Forestry is convinced that the approach of working with industry to present research results in industry-ready, usable tools has significantly increased uptake of our results. This is supported by documented usage and positive feedback from our industry partners. For example, the introduction and use of the feller-buncher benchmarking tool lead to specific requests for the CRC for Forestry to develop similar tools for other common, system-limiting machines such as in-field chippers and harvesters. The onboard system selection and implementation guide was delivered to more than 100 industry stakeholders over six workshops and has lead to supplementary requests to the CRC to assist individual companies implement the technology appropriately to deal with specific needs in their operations. This demonstrates that the objective of this research has been achieved – the guide has initiated the uptake of onboard technology. Only a few weeks after its introduction to industry the machine evaluation toolbox has been the most requested topic of CRC-delivered lunch-hour seminars/workshops. Of the 60 industry stakeholders that have received the tool, more than 90% report they will use it in their work.

While very different in their presentation, complexity and application, what the six cases have in common is they are targeted at a key industry interest or need. All of them hold the potential for significant financial impact for the industry. In 2010, 24.8 million m3 of timber was harvested in Australia (ABARES 2011) with the CRC for Forestry’s partners representing about 80% of that harvest. Changes in utilisation or productivity of as little as 2% gained through the use of the feller-buncher benchmark curve, onboard systems, stem file productivity model, ALPACA or the machine evaluation toolbox, to identify and address an issue in an operation will deliver benefits of between 15 to 20 million $ for the CRC for Forestry partners per year. If the level of improvements that have been reported with early uses are translated across the industry, these benefits increase to more than 80 million $ per year. In the case of FastTRUCK, the second version will be released to all sectors of the industry, and this is expected to reduce transport costs by up to 10% through opti-

477

M. Brown et al.

Improving Forest Operations Management through Applied Research (471–480)

mised schedules, providing the potential for more than 25 million $ in annual benefits to CRC for Forestry partners.

Forsberg, M., Frisk, M., Rönnqvist M., 2005: FlowOpt – a decision support tool for strategic and tactical transportation planning in forestry. International Journal of Forest Engineering 16(2): 101–114.

4. Conclusion – Zaklju~ak

FPInnovations, 2005: ProVue – Harvest and costing model for the Canadian Forest Industry. Available at http://www. feric.ca/en/?CFTOKEN=&OBJECTID=825F91A3-A960-F D3B-E1D7FB47BA9645B0&CFID=612117

Applied research must consider implementation and innovation as its end objective to sustain support. The most effective approach to achieve this is to have an ongoing, strong collaborative approach with industry end users, and to deliver research results in a form that is ready for use within existing industry structures, to address real industry needs.

When presented with effective, easy to use tools, based on quality research and developed specifically for industry needs, large financial opportunities are simply the last motivator to drive change and clearly show the value of research.

5. References – Literatura Acuna, M., Ghaffariyan, M. R., Brown, M., 2010: Analysis of Factors that affect the transport efficiency of in-field chipping operations. CRC for Forestry Industry Bulletin #11.

Acuna, M., Ghaffariyan, M. R., 2011: Fast Truck: A Truck Scheduling System to Improve Efficiency in the Australian Forest Industry. In: 34th COFE annual meeting, June 12–15, Quebec City, Canada.

Agrtans Research, 2008: Economic assessment of selected investments of the Cooperative Research Centre for Forestry.CRC for Forestry Governance Reports (http://www.crcforestry.com.au/publications/downloads/Economic-Report-2008.pdf).

Fryk, J., Nordstrom, L., 2011: The impact of Forest Operations Research. In: 4th Forest Engineering Conference »Innovation in Forest Engineering – Adapting to Structural Change». White River, South Africa. April 5–7.

Grover, V., Jeong, S. R., Kettinger, W. J., Teng, J. T. C., 1995: The implementation of business process reengineering.

Journal of Management Information Systems 12(1): 109–144.

Gingras, J. F., 1988: The effect of site and stand factors on feller-buncher performance. Forest Engineering Research Institute of Canada. Technical Report. TR-84, 23 p.

Granskog, J. E., Anderson, W. C., 1981: Dense undergrowth reduces feller-buncher productivity in shortleaf plne plantations. Research Note SO-274, United States Department of Agriculture, Forest Service, Southern Forest Experiment Station, New Orleans, LA, 4 p.

Jamieson, S., 2004: Eliminating the guesswork. Canadian Forest Industries, 20–21 p.

Linnainmaa, S., Savola, J., Jokinen, O., 1995: EPO A knowledge based system for wood procurement management.

Paper from the 7th Annual Conference on Artificial Intelligence, Montreal, 107–113 p.

Long, C., Wang, J., McNeel, J., Baumgras, J., 2002: Production and cost analysis of a feller-buncher in central Appalachian hardwood forest. In: 2002 Council on Forest Engineering (COFE) Conference Proceedings: »A Global Perspective» Auburn, June 16–20.

Australian Bureau of Agricultural and Resource Economics and Science (ABARES), 2011: Australia’s forests at a glance 2011 with data to 2009 – 2010, Commonwealth of Australia, Canberra, Australia.

Nurminen, T., Korpunen, H., Uusitalo, J., 2006: Time consumption analysis of the mechanized cut-to-length harvesting system. Silva Fennica 40(2): 335–363.

Bjorheden, R., Apel, K., Shiba, M., Thompson, M., 2000: Forest work Study Nomenclature, IUFRO World Congress in Malaysia.

Nuutinen, Y., Väätäinen, K., Heinonen, J., Asikainen, A., Röser, D., 2008: The accuracy of manually recorded time study data for harvester operation shown via simulator screen. Silva Fennica 42(1): 63–72.

Cribb, J., Hartomo, T. S., 2002: Sharing knowledge – A Guide to Effective Science Communication. CSIRO publishing, 208 p.

Dentoni, D., English, F., 2011: Factors Influencing Industry Uptake of Public-Private Research on Marketing: The Case of the Australian Seafood Cooperative Research Centre.

21st Annual IFAMA World Forum and Symposium Frankfurt, Germany.

Epstein, R., Rönnqvist, M., Weintraub, A., 2007: Forest transportation. In: Handbook Of Operations Research In Natural Resources. International Series in Operations Research & Management Science 99(3): 391–403.

478

Oliveira Junior, E., Seixas, F., Batista, J., 2009: Feller-buncher productivity in eucalyptus plantation on steep ground terrain. Floresta 39(4): 905–912.

Ovaskainen, H., Uusitalo, J., Väätäinen, K., 2004: Characteristics and significance of a harvester operator’s working technique in thinnings. International Journal of Forest Engineering 15(2): 67–77.

Palmgren, M., Rönnqvist, M., Varbrand, P., 2004: A near-exact method for solving the log-truck scheduling problem. International Transactions of Operations Research 11 (4): 447–464.

Croat. j. for. eng. 32(2011)2

Improving Forest Operations Management through Applied Research (471–480)

M. Brown et al.

Spinelli, R., Hartsough, B. R., Magagnotti, N., 2010: Productivity Standards for Harvesters and Processors in Italy.

Forest Products Journal 60(3): 226–235.

University of Stellenbosch and Institute Forestry Research South Africa, 2010: South African Harvesting & Transport System Costing Model. Handbook Version 1.5, 37 p.

Strandgard, M., 2009: Onboard Systems for Australian Forest Operations. In: The Biennial Conference of the Institute of Foresters of Australia, Caloundra, Australia. September 6–10.

USDA Forest Service, Forest Operations Research, 2008: Machine rate calculator and General Ground-based harvesting system analysis. Available at http://www.srs.fs. usda.gov/forestops/download.htm

Strandgard, M., 2010: User guide for the selection of onboard technology in Australian harvesting operations. In: Precision Forestry Symposium, Stellenbosch University, South Africa.

Visser, R., Stampfer, K., 2003: Tree-length system evaluation of second thinning in loblolly pine plantations. Southern Journal of Applied Forestry 27(1): 77–82.

Strandgard, M., Mitchell, R., 2010: Benchmarking feller-buncher productivity in Western Australia Blue Gum Plantations. CRC for Forestry Industry Bulletin #12

Weintraub, A., Epstein, R., Morales, R., Serón, J., Traverso, P., 1996. A truck scheduling system improves efficiency in the forest industries. Interfaces 26(4): 1–12.

Sa`etak

Pobolj{anje radnih operacija u {umarskoj praksi i industriji pomo}u primijenjenih istra`ivanja Velik izazov za znanstvenike jest prijenos znanja i rezultata primijenjenih istra`ivanja industriji. [tovi{e, {umarski stru~njaci na terenu ~esto nemaju vremena za tuma~enje rezultata brojnih istra`ivanja u znanosti te daju prednost ve} gotovim alatima dobivenim na temelju rezultata istra`ivanja. Centar za istra`ivanja u {umarstvu Australije (CRC) usmjeren je na tuma~enje rezultata znanstvenih istra`ivanja te njihovo preno{enje i prilago|avanje kako bi bili lako primjenjvi za prate}u industriju. U ovom se radu raspravlja o {est studija koje je Centar prilagodio za prate}u industriju.

U prvom slu~aju podaci iz 13 razli~itih podru~ja istra`ivanja poslu`ili su za dobivanje krivulje proizvodnosti (broj posje~enih stabala u efektivnom vremenu rada stroja) za feller-buncher. Na krivulju proizvodnosti dodane su linije odstupanja od ± 10 % (slika 1) koje predstavljaju smjernice pri kori{tenju referentne krivulje. Promatraju}i prosje~ni obujam stabla, {umarski stru~njaci i vlasnici strojeva mogu brzo utvrditi radi li stroj na o~ekivanoj razini proizvodnosti, tj. sije~e li dovoljan broj stabala u efektivnom radnom vremenu.

Na temelju povratnih informacija iz industrije Centar za istra`ivanja u {umarstvu osmislio je tzv. alat za procjenu strojeva u {umarstvu (Machine Evaluation Toolbox) koji omogu}uje ~etiri metode procjene rada za naj~e{}e kori{tene {umske strojeve, kori{tenje terenskih obrazaca, izra~une i ra~unske operacije te sadr`i generator izvje{taja. Takav alat omogu}uje stru~njacima na terenu brzo djelovanje pri ocjenjivanju proizvodnosti sustava rada i strojeva.

Ina~e je za izradu modela proizvodnosti harvestera bilo potrebno mnogo vremena, no kori{tenjem tzv. programa StanForD iz mati~nih datoteka zapisa jednozahvatnoga harvestera procijenjeno je vrijeme turnusa rada stroja. U usporedbi sa sada{njim metodama, koriste}i mati~ne datoteke stroja, uvelike je ubrzano stvaranje u~inkovitih modela proizvodnosti stroja.

Ra~unalna je oprema ugra|ena u pojedine sustave rada te je stvoren vodi~ za upravljanje radnim operacijama.

Korisnici vodi~a odabiru problem ili pote{ko}u na koji su nai{li u sustavu rada pa ih program vodi opisu i mogu}im primjerima za rje{avanje toga problema.

Alat za operativna predvi|anja proizvodnosti i tro{kova (ALPACA) temelji se na vi{e od 200 izra|enih studija proizvodnje te pru`a dvije vrste analize: 1) sustav za procjenu sje~e i 2) sustav za procjenu stroja. Koriste}i rezultate istra`ivanja, alat omogu}uje predvi|anje dnevne proizvodnja, jedini~nih tro{kova i vrijeme potrebno za sje~u drva.

Za stvaranje rasporeda prijevoza drva kori{ten je sustav FastTRUCK. Sustav pru`a stvaranje rasporeda kamionskoga prijevoza kako bi se smanjili ukupni tro{kovi prijevoza drva te iskoristio najbolji mogu}i broj kamiona potrebnih za prijevoz drva u jednom danu.

Tih {est slu~ajeva pokazuje niz strategija kojima se koristi Centar za istra`ivanja u {umarstvu kako bi se rezultati znanstvenih istra`ivanja iskoristili u {umarskoj praksi i industriji. U svim slu~ajevima bio je potreban Croat. j. for. eng. 32(2011)2

479

M. Brown et al.

Improving Forest Operations Management through Applied Research (471–480)

dodatni napor kako bi se rezultati znanstvenih istra`ivanja iskoristili u prakti~ne svrhe. Va`no je naglasiti da su razvijeni alati i programi nastali u suradnji s krajnjim korisnicima da se zadovolje potrebe i o~ekivanja same industrije. Suo~ena s djelotvornim alatima, jednostavnim za kori{tenje, nastalim na temelju kvalitetnih znanstvenih istra`ivanja te posebno razvijenim za potrebe industrije, velike financijske mogu}nosti kori{tenja tih alata postale su jednostavno zadnji u nizu pokreta~ promjena u industriji te je jasno pokazana vrijednost znanstvenih istra`ivanja.

Klju~ne rije~i: inovacije, primjena, u~inkovitost, upravljanje proizvodnjom

Authors’ address – Adresa autorâ: Mr. Mark Brown Harvesting and Operations Program Manager & Industry Engagement Manager e-mail: mwbrown@unimelb.edu.au University of Melbourne CRC for Forestry Department of Forest and Ecosystem Science Adjunct Research Fellow, University of the Sunshine Coast 1 Water St., 3363, Creswick AUSTRALIA Mr. Martin Strandgard Harvesting and operations program e-mail: mnstra@unimelb.edu.au University of Melbourne CRC for Forestry 500 Yarra blvd., Richmond, VIC AUSTRALIA Mauricio Acuna, PhD. Harvesting and operations program e-mail: mauricio.acuna@utas.edu.au University of Tasmania CRC for Forestry Private bag 12, Hobart, TAS AUSTRALIA Mr. Damian Walsh Harvesting and operations program e-mail: damianw@unimelb.edu.au University of Tasmania CRC for Forestry 500 Yarra blvd., Richmond, VIC AUSTRALIA

Received (Primljeno): July 26, 2011 Accepted (Prihva}eno): September 5, 2011

480

Mr. Rick Mitchell Harvesting and operations program e-mail: rick.mitchell@wapres.com.au CRC for Forestry – WAPRES Level 2, 53 Victoria St, Bunbury, WA AUSTRALIA Croat. j. for. eng. 32(2011)2

Original scientific paper – Izvorni znanstveni rad

Using On-Board GPS to Identify Training Needs of Helicopter Pilots Andy Horcher, Rien Visser Abstract – Nacrtak In forest harvesting, helicopter extraction systems are incredibly versatile due to their ability to avoid many of the obstacles that encumber ground based and skyline systems. Helicopter yarding is used for a variety of reasons including site sensitivity, urgency to remove or deliver the product, lack of access, and slope of the terrain. Because of the high cost of helicopter yarding, maximizing productivity is critical. There are many site and stand factors that affect productivity, but pilot experience and skill is also known to be important. Job training of new pilots can be very expensive, including the loss of productivity during the training phase. Basic time studies can be used to show differences in productivity between pilots. This project shows that by using an on-board GPS system to capture elemental time study data, that is geo-referenced, it becomes possible to isolate in detail during what phase of the turn cycle a trainee is not efficient. Using data collected at three different sites in the Pacific Northwest region of the USA, basic productivity curves were developed for each yarding element and indicated that inexperienced pilots produced between 33 and 43 tons less per productive machine hour. For these case studies, the trainee pilot was losing most of his time positioning the helicopter when hooking trees, although reduced acceleration and maximum top velocity was also noted. With detailed feedback, the trainee pilot and/or trainer can focus efforts to improve training effectiveness and reduce productivity loss during the training period.

Keywords: helicopter yarding, on-board GPS system, pilot training, productivity

1. Introduction – Uvod Helicopter yarding is incredibly versatile due to its ability to avoid many of the obstacles that encumber ground based and skyline systems (Stampfer et al. 2002; Conway 1976; Burke 1973). Today this yarding system is used for a variety of reasons including site sensitivity, urgency to remove or deliver the product, lack of access, and slope of the terrain. The use of helicopters in forestry continues to expand.

Where there were only a few firms offering helicopter logging services in the early 1970’s (Conway 1976), today the Helicopter Association International estimates almost 175 forestry or logging companies use helicopter logging as a principal means of yarding timber (Bruce 2003).

Helicopters have various designs and abilities, and the variety of helicopters used is also fairly extensive. The type of helicopter used will influence speed, angle of ascent, and maximum payload (Conway 1976). For example Dunham (2003) lists 14 manufacturers and models used in British Columbia, Croat. j. for. eng. 32(2011)2

Canada. Helicopters are typically rated by payload capacity that ranges from 1134 kg for the Eurocopter Lama to 12727 kg for the Boeing CH234 for machines commonly used in helicopter yarding.

Helicopter yarding is a relatively high cost extraction method (Keegan et al. 1995). While Hartsough et al. (1997) found ground based skidding to account for approximately 20 – 25% of the stump to truck operation costs, helicopter yarding ranged between 65 and 78% of the stump to truck costs (Krag and Evans 2003; Dunham 2003). Currently helicopter operating costs are at least US$ 500 per hour for the smallest machine, up to approximately US$ 4500 per hour for the larger machines.

Optimizing the payload is a key factor in achieving efficient yarding (Burke 1973; Hartsough et al.

1986). The location and layout of the log landing is also a crucial factor. The primary concern is the yarding distance, which generally is the distance from the hook point to the log landing (Burke 1973). Weather not only limits when operations may occur, but it

481

A. Horcher and R. Visser

Using On-Board GPS to Identify Training Needs of Helicopter Pilots (481–488)

also influences helicopter capability during operations. The density of the air impacts both the ability of the helicopter to achieve lift and the horsepower of the engine (Wagtendonk 1996).

Other operation dependent factors that may influence productivity are the pilots themselves. When a helicopter yarding organization employs a pilot new to logging work, they are likely to experience higher costs (Sloan and Tollenaere 1994; Warren 1996; Stampfer et al. 2002). Stampfer et al. (2002) shows that an experienced pilot delivered 59% more volume to the landing than a trainee-pilot did.

With the high cost and wide range of factors affecting helicopter yarding, the application of new technology to improve and evaluate training systems will be very beneficial. New technology allows us to more accurately measure the helicopter yarding process and better predict the production rates in different conditions. Recent forest operation research has used ground based equipment with on-board Geographic Positioning Systems (GPS) to conduct more precise production and site impact analysis (McDonald et al. 2002, McDonald et al. 2000). GPS is also being used more extensively to monitor, manage and optimize harvesting operations (Flisberg et. al 2007). Heinimann and Caminada (1996) recommend using GPS to gather more precise data on helicopter operations.

This study aims to test the application of on-board GPS to aid the helicopter yarding industry by measuring the impact of pilot experience on productivity. Knowing where a trainee-pilot is likely to need the most improvement may assist the industry in selecting optimized training routines.

2. Describing the Helicopter Yarding Process – Izno{enje drva helikopterom The process of helicopter yarding can be broken into yarding cycles, turns, and elements. The basic definition for a cycle is leaving the service landing, flying a number of turns and returning to the service landing. The basic definition for a turn is leaving the log landing and traveling to the location of the payload (outhaul), picking up the payload (hooking), returning to the log landing with that payload (inhaul), and releasing the payload at the log landing (unhooking). Each segment of the turn just described is an element.

Beginning at the service landing, the helicopter will fly to the harvest area and begin yarding logs.

During the hooking element there will often be a person, the hooker, on the ground with pre-choked logs ready to be connected to the hook at the end of the helicopters long line. The pilot locates the hooker

482

and maneuvers the hook near the hooker. Then the hooker slides the chokers into the hook. The pilot then lifts the logs off of the ground and clear of the forest canopy.

The inhaul element begins and the pilot flies toward the log landing. At the landing the pilot sets the logs on the ground in the drop zone and releases the chokers from the hook. With the load released, the pilot clears the log landing and enters the outhaul element and flies back to the woods for another load of logs. The entire process, hook, inhaul, unhook, and outhaul is referred to as a turn. If no problems occur, this continues for 60 to 90 minutes, until the helicopter must be refueled. The pilot must then return to the service landing for fuel. When the helicopter is in the hooking, inhaul, unhooking, or outhaul elements, this is called the yarding cycle. When the helicopter is flying to or from the service landing or being fueled or repaired, this is called the service cycle.

3. Methodology – Metodologija The data used for this paper is part of a larger comprehensive study into measuring helicopter productivity using GPS and GIS analyses (Horcher 2008). It includes over 35 days of helicopter data gathered at nine different sites on three different helicopters.

3.1 Study Sites – Mjesto istra`ivanja At three locations the operation included an inexperienced pilot, who for the purpose of this study is defined as a pilot with less than 100 hours flying in logging operations. The sites were located in the Pacific Northwest, the operations were harvesting conifer trees on rolling terrain, and all operations were studied in the summer. At each site (Table 1), an attempt was made to capture at least 30 turns for a cycle, whereby the inexperienced and experienced pi-

Table 1 Site descriptions for pilot experience data Tablica 1. Mjesto istra`ivanja Site Mjesto istra`ivanja

Paired Cycles Broj radnih turnusa

Yarding Method Na~in prihvata drva

A

1

Grapple Hvatalo

B

4

C

5

Chokers U`e za vezanje Chokers U`e za vezanje

Harvest Type Vrsta sje~e Seed Tree/Salvage Naplodni sijek Sanitarna sje~a Thinning Proreda Thinning Proreda

Croat. j. for. eng. 32(2011)2

Using On-Board GPS to Identify Training Needs of Helicopter Pilots (481–488)

A. Horcher and R. Visser

Table 2 Description of the numerical variables and time components Tablica 2. Opis broj~anih varijabli i vremenskih sastavnica rada Type Tip

Dependant variables Zavisne varijable

Covariables Kovarijable

Factors ^imbenici

Name Ime

Description – Opis

Unit M. jedinica

FlyOut Let neoptere}enoga helikoptera

Time for the helicopter to fly from landing to hook point Vrijeme leta helikoptera od stovari{ta do mjesta utovara

sec

Hook Kop~anje tereta

Time at hook point, which is defined by a radius of 20 meters around the actual hook point Vrijeme provedeno za kop~anje drva, uklju~uju}i podru~je od 20 m promjera oko mjesta kop~anja drva

sec

FlyIn Let optere}enoga helikoptera

Time for the helicopter to fly from hook point to landing Vrijeme leta helikoptera od mjesta utovara do stovari{ta

sec

Landing Slijetanje

Time at landing, which is defined by a radius of 30 meters around the actual landing Vrijeme provedeno na stovari{tu, uklju~uju}i podru~je od 30 m promjera oko stovari{ta

sec

TurnVol Obujam tovara

Sum of the tree volumes extracted in one turn Obujam tovara iznesen unutar jedne sastavnice rada

kg

TreeVol Obujam drva

Average tree volume Prosje~an obujam komada drva

kg

ExtDist Udaljenost izno{enja

3d extraction distance – Udaljenost izno{enja drva

m

ElvChange Promjene u visini leta

Change in Elevation – Promjene u visini leta tijekom izno{enja drva

m

Slope Nagib

Slope between landing and hook point Razlika u nagibu izme|u stovari{ta i mjesta kop~anja drva

%

PilotEx Iskustvo pilota

Pilot experience, as defined by 100 flying hours in logging Iskustvo pilota (100 radnih sati izno{enja drva)

0/1

ChokDrop Ispu{tanje u`adi

Turn that includes picking up and dropping off a bundle of chokers Podizanje i otpu{tanje u`adi za kop~anje

0/1

Vel Brzina

Max velocity (Flyout and FlyIn) Najve}a brzina (optere}enoga i neoptere}enoga helikoptera)

m/sec

Accel Ubrzanje

Max Acceleration (Flyout and FlyIn) Najve}e ubrzanje (optere}enoga i neoptere}enoga helikoptera)

m/sec2

Decell Usporavanje

Max Deceleration (Flyout and FlyIn) Najve}e usporavanje (optere}enoga i neoptere}enoga helikoptera)

m/sec2

Time Vrijeme

lot flew consecutively yarding cycles ensuring consistency in stand and terrain factors. In all cases the inexperienced pilots flew fewer turns per cycle: in part because they took longer per turn, but were also deliberately replaced by an experienced pilot to reduce the productivity impact. During the study of the seed tree/salvage harvest with a grapple (site A) the inexperienced pilot only flew long enough once to obtain one cycle comparison with an experienced pilot. As the data is consistent with the other sites it has been included in this paper.

Helicopter yarding is a very competitive business and a confidentiality agreement required to collect research data prevents us from identifying the exact location of the operation, or specifically linking the make and model of the helicopter to a specific Croat. j. for. eng. 32(2011)2

data set. The two machines used by the pilots in this experiment were the Boing Vertol BV 107 and the Sikorsky S-61A. Both are rated at approximately 4 ton payload.

3.2 Data Collection – Prikupljanje podataka GPS data was collected with a WAAS enabled Trimble Geo XT with EVEREST technology mounted in the helicopter. Location information was gathered at one-second intervals, and downloaded from the GPS unit when possible.

Some basic programs were developed to aid the evaluation of the data, including auto location of landing and hook point within the GPS data using helicopter velocity (Horcher 2008). Using 35 meter

483

A. Horcher and R. Visser

Using On-Board GPS to Identify Training Needs of Helicopter Pilots (481–488)

radii around the landing and the hook point, the time was divided into the four phases that make up a typical turn cycle.

An overview of all of the time-study elements is presented in Table 2.

It is possible to present the GPS data from a helicopter yarder cycle graphically, for example as overlaid on a GIS based contour as shown in Fig. 1.

Processing the GPS data, and combining it with the payload data as recorded by the load sensor, allowed it to be converted to a more conventional time study form, as presented in Table 3. It includes four timed cycle elements (fly out, hook the load, fly in and releasing the logs at the log landing), and a measured turn weight that can be combined with the time elements to calculate productivity on a per turn basis. In addition, the extraction distance is recorded for each turn. The final column shows the pilot experience covariate.

Statistical analysis was carried out using SAS JMP 7.0, including basic mean comparisons as well as linear regressions to build the models. Scheffe’s multiple comparison procedure was conducted on hook and unhook time for the trainee and experienced pilot data. Comparison of means was tested at

Fig. 1 Mapped helicopter yarding data for a full cycle from an on-board GPS unit Slika 1. Kartografski prikaz turnusa rada helikoptera pomo}u sustava GPS

Table 3 A sample of a typical time study format for helicopter yarding Tablica 3. Uzorak podataka iz studija rada i vremena za izno{enje drva helikopterom Covariate Time Elements – Radni zahvati Factor – ^imbenik Cycle Kovarijabla Turnus Un-hook, sec. Tot-Cycle, sec. Payload, t Prod, m3/hr Extr-Dist, m Exper, 0/1 rada Out, sec. Hook, sec. In, sec.

Let, s Kop~anje, s Let, s Otkop~avanje, s Ukupno, s Teret, t Proizvodnost, m3/h Udaljenost izno{enja, m Iskustvo pilota, 0/1 1 36 44 32 25 137 3.0 75 671 0 2 29 69 37 16 151 3.1 70 687 0 3 27 58 36 15 136 3.1 78 667 0 4 28 40 33 13 114 3.3 99 672 0 5 28 37 36 14 115 2.3 68 674 0 Etc. – – – – – – – – – Itd.

Table 4 Inexperienced pilot elemental time as a percent of experienced pilot elemental time Tablica 4. Usporedba vremena rada neiskusnoga i iskusnoga pilota pri radnim zahvatima Site Mjesto rada A B C

484

Choker Drop Ispu{tanje u`adi N/A Nije primjenjivo No – Ne Yes – Da No – Ne Yes – Da

Outhaul Let neoptere}enoga helikoptera

Hook Kop~anje

Inhaul Let optere}enoga helikoptera

Unhook Otkop~avanje

Total Ukupno

156%

368%

175%

117%

181%

128% 189% 129% 176%

209% 217% 200% 189%

147% 147% 179% 179%

147% 155% 204% 176%

164% 194% 172% 180%

Croat. j. for. eng. 32(2011)2

Using On-Board GPS to Identify Training Needs of Helicopter Pilots (481–488)

A. Horcher and R. Visser

Table 5 Mean hook and unhook times excluding choker drop cycles Tablica 5. Srednje vrijeme kop~anja i otkop~avanja drva bez vremena ispu{tanja u`adi Site Mjesto rada

Element Radni zahvat

Pilot Status* Iskustvo pilota

Turns, n Radne sastavnice, n

Mean, sec.

Srednja vrijednost, s

Hook Kop~anje

1

10

83.5

0

38

22.7

Unhook Otkop~avanje

1

8

19.4

0

39

16.5

Hook Kop~anje

1

44

109.0

0

137

51.7

Unhook Otkop~avanje

1

44

30.8

0

139

16.9

Hook Kop~anje

1

54

106.0

0

152

52.8

Unhook Otkop~avanje

1

56

36.0

0

153

17.6

A

B

C

Scheffe Sign., 0.05 Yes Da No Ne Yes Da Yes Ne Yes Da Yes Da

*Pilot Status; 0 = experienced, 1 = inexperienced – Piloti : 0 = bez iskustva, 1 = s iskustvom

the 0.05 level, whereby stepwise model development used a threshold of 0.10 for parameter inclusion.

4. Results – Rezultati The full data set showed highly significant differences in almost all elements when comparing the impact of pilot experience. An example to show how improvement potential can be assessed compares relative times for the total turn, as well as individual elements within a turn. Table 4 provides a summary of all the data, showing the percent of time an inexperienced pilot spends relative to an experienced pilot. 100 percent indicates performance equivalent to an experienced pilot.

For these case studies, the total time column indicates that a trainee pilot takes at least 64% longer to complete a single turn, and this is exacerbated to 80 or even 90% when adding complexity such as requiring chokers to be picked up and dropped off, or when using a grapple. For the individual elements, both the highest and lowest percentages occur within site A: 368% for hook and 117% for unhook. This indicates that grapples may require minimal experience for unhooking, but significant experience for actually grabbing a log. Overall, the greatest room for improvement usually occurs in the hook element.

Table 5 provides a more detailed look at just the Hook and Unhook phases of the yarding cycle. It indicates how readily significant differences can be Croat. j. for. eng. 32(2011)2

found between pilots, even if the inexperienced pilot flies very few turns.

To better understand the reason for the additional time required for a trainee pilot to complete either the inhaul or the outhaul phase of a cycle, it was possible to look at maximum velocity, as well as the acceleration and deceleration, of the helicopter. For example, at site A maximum inhaul velocity is significantly higher with the experienced pilot, at 99 km/hr, than it is with the inexperienced pilot, 59 km/hr. Site C exhibits a similar situation, where the experienced pilot achieves a higher mean maximum outhaul velocity of 139 km/hr compared to the inexperienced pilot’s mean maximum outhaul velocity of 107 km/hr. The mean outhaul acceleration and deceleration rates at site C are more extreme for the experienced pilot as well, with 3.0 m/s2 and –3.0 m/s2, compared to 1.9 m/s2 and –1.8 m/s2 for the trainee pilot.

Regressions for turn time and productivity (tons/hr) all show a strong correlation. In general, turn time had stronger correlations compared to productivity, adjusted R2 = 0.70 to 0.79, and adjusted R2 = 0.50 to 0.64 respectively. The lower multiple coefficient of determination for productivity relative to turn time is expected to be a result of the variability in wood availability and arrangement at each hook point, which is not described with the explanatory variables. From a production perspective, on average in these studies the inexperienced pilots reduce production from 75.9 to 42.8 tons/hr. This equates to a 44% decrease in productivity.

485

A. Horcher and R. Visser

Using On-Board GPS to Identify Training Needs of Helicopter Pilots (481–488)

Pilot experience (PilotEx) and extraction distance (ExtrDist) were significant for all regression models.

Turn volume (TurnVol) was significant in all turn cycle time models, but not used in the productivity models as it is the quotient component of productivity. Specifically, the turn time and productivity regression equations for the grapple operation (Site A) were: SiteA: Turn(sec) = 22.1 + 112.8(PilotEx) + + 0.0281(TurnVol)

(1)

Adjusted R2 = 0.70 SiteA: Prod(tonnes/hr) = 90.0 – 42.8(PilotEx) – – 0.0094(ExtrDisc) (2) Adjusted R2 = 0.50 For sites B and C, the need to pick up and drop off chokers during a production cycles has a significant influence on both turn time and productivity. Specifically; SiteB: Turn(sec) = 51.8 + 123.4(PilotEx) + + 0.0358(ExtrDist) + 96.7(ChokDrop) + + 0.0200(TurnVol)

(3)

Adjusted R2 = 0.75 SiteB: Prod(tonnes/hr) = 90.6 – 33.1(PilotEx) – – 0.0105(ExtrDist) – 28.1(ChokDrop) (4) Adjusted

R2

= 0.64

For site C an autocorrelation component also became significant in the regression equation for turn time. Autocorrelation would indicate that cycles are not independent of each other. In this case it suggests that as the pilot returns to a hook point, the accumulation of prior knowledge of the site helps decrease the overall cycle time. The effect of slope was also picked up in the regression equation for productivity at Site C. SiteC: Turn(sec) = 58.5 + 122.4(PilotEx) + + 0.0407(ExtrDist) + 63.3(ChokDrop) – – 0.58 (Turn of Cycle) + 0.0111(Payload) + + 3.9(Logs)

Integration of new technologies can provide significant opportunities to improve productivity of existing timber harvesting operation. This study has demonstrated the opportunity for using onboard GPS for the benefit of identifying training needs for inexperienced helicopter pilot flying in logging operations. It identified turn time and productivity differences at both the cycle and elemental level. This not only allows a true opportunity cost for operating with trainee pilots to be established, but also allows for targeted training by indicating the specific phases where a trainee pilot is less efficient compared to an experienced pilot. General regression equations that identified key factors that affect productivity are also readily developed. Although some automation of the data interpretation was achieved during this study, opportunities exist for improved data synthesis.

6. References – Literatura Bruce, B., 2003: Helicopter logging’s bumpy ride: using helicopters to thin forests instead of fighting fires. Timber West July/Aug. Burke, D., 1973: Helicopter logging: Advantages and disadvantages must be weighed. J. For. Sept. 71(9): 574–576.

Conway, S., 1976: Logging Practices: Principles of Timber Harvesting Systems. USA: Miller Freeman Publications.

Dunham, M. T., 2003: Helicopter logging with the Bell 214B: group and single-tree selection in low-volume coastal cedar stands. FERIC, Vancouver, B.C. Advantage Report 4(33): 1–19.

Flisberg, P., Forsberg, M., Ronnqvist, M., 2007: Optimization based planning tools for routing of forwarders at harvest areas. Can. J. For. Res. 37(11): 2153–2163.

Hartsough, B. R., Lambert, M. B., Miles, J. A., 1985: Airship logging: parameters affecting load factors. Trans. of ASAE 28(5): 1363–1366.

Hartsough, B. R., Lambert, M. B., Miles, J. A., 1986: Simulating changes to helicopter logging operations. Trans.

ASAE 28(5): 1228–1231.

(5)

Adjusted R2 = 0.79 SiteC: Prod(tonnes/hr) = 87.4 – 33.4(PilotEx) – – 0.0176(ExtrDist) – 0.21(Slope) – – 14.0(ChokDrop) (6) Adjusted R2 = 0.58 This analysis showed that not only can pilot differences be quickly evaluated using GPS technology, but productivity factors identified. Here onboard GPS shows strong promise in evaluating performance.

486

5. Conclusions – Zaklju~ci

Hartsough, B. R., Drews, E.S., McNeel, J. F., Durston, T. A., Stokes, B. J., 1997: Comparison of mechanized systems for thinning ponderosa pine and mixed conifer stands. For. Prd. J. 47(11/12): 59–68.

Heinimann, H. R., Caminada, L., 1996: Helicopter logging in Switzerland: analysis of selective logging operations.

Paper presented to the Mountain Logging and Pacific Northwest Skyline Symposium, May 12 – 16 in Campbell River, Canada.

Horcher, A., 2008: Improving Helicopter Yarding with On-Board GPS. PhD Dissertation, online at www.scholar.lib. vt.edu/theses. Virginia Tech, Blacksburg, USA Croat. j. for. eng. 32(2011)2

Using On-Board GPS to Identify Training Needs of Helicopter Pilots (481–488)

A. Horcher and R. Visser

Jackson, B. D., Morris, R. A., 1986: Helicopter logging of baldcypress in southern swamps. S. J. App. For. 10(2): 92–95.

McDonald, T. P., Taylor, S. E., Rummer, R. B., 2000: Deriving forest harvesting machine productivity from positional data. ASAE An. Int. Meet. Paper No 005011 ASAE.

Keegan, C. E., Fiedler, C. E., Wichman, D. P., 1995: Costs associated with harvest activities for major harvest systems in Montana. For. Prod. J. 45(7–8): 78–82.

McDonald, T. P., Carter, E. A., Taylor, S. E., 2002: Using the global positioning system to map disturbance patterns of forest harvesting machinery. Can. J. For. Res. 32(2): 310–319.

Krag, R., Evans, C., 2003. Helicopter logging on the Queen Charlotte Islands: productivities and costs of a Sikorsky S-64E Skycrane in clearcuts, patch cuts and single-tree selection cuts. FERIC, Vancouver, B.C. Advantage Report 4(19): 1–40.

Sloan, H., Tollenaere, J., 1994: Technology advances in helilogging; a case study of the KMAX in Appalachian hardwoods. in Proceeding of Advanced technology in forest operations: Applied ecology in action. Council on Forest Engineering. Portland, OR.

Lyons, C. K., McNeel, J., 2004: Partial Retention and helicopter turn volume. For. Prod. J. 54(1): 58–61.

Stampfer, K., Gridling, H., Visser, R., 2002: Anaylses of parameters affecting timber extraction. Int. J. For. Eng. 13(2).

Wagtendonk, W. J., 1996: Principles of Helicopter Flight.

New Castle, WA: Aviation Supplies and Academics.

Sa`etak

Kori{tenje sustava GPS za unaprje|enje obuke pilota helikoptera Izno{enje drva helikopterom pokazalo se kao vi{estruko upotrebljiv na~in primarnoga transporta drva zbog svoje sposobnosti svladavanja mnogih terenskih prepreka s kojima se susre}u po tlu kretni te u`etni sustavi prilikom privla~enja drva. Izno{enje drva helikopterom naj~e{}e se koristi u ekolo{ki osjetljivim sastojinama, pri hitnim isporukama ve}e koli~ine drva, zbog nemogu}nosti pristupa sje~ini (manjak {umskih cesta ili `i~nih linija) ili zbog izrazito nepovoljnoga nagiba terena. Postoje razne vrste helikoptera za izno{enje drva te }e vrsta letjelice utjecati na brzinu, kut uspona i najve}u dopu{tenu nosivost tereta. Zbog visoke cijene izno{enja drva helikopterom najva`nije je posti}i najvi{u mogu}u proizvodnost takva sustava rada. Trenuta~no se tro{kovi rada helikoptera (ovisno o vrsti letjelice) kre}u od najmanje 500 US$ po satu pa sve do 4500 US$ po satu.

Postoje mnogi sastojinski ~imbenici koji utje~u na proizvodnost sustava rada, ali iskustvo i vje{tina pilota helikoptera pokazali su se tako|er vrlo zna~ajnima. Obuka novih pilota neposredno u sje~ini mo`e biti vrlo skupa, pa je studijem rada i vremena lak{e i jeftinije uo~iti razlike izme|u pilota s razli~itim radnim iskustvom. Utvr|ivanjem gdje }e pilot vje`benik (pilot s manje od 100 sati leta izno{enja drva) najvjerojatnije trebati dodatnu obuku pomo}i }e u cjelokupnom sustavu rada te ubrzati vrijeme obuke pilota vje`benika.

Kori{tenjem sustava GPS, uklopljena u vozilo radi snimanja georeferenciranih podataka tijekom studija rada i vremena, mogu}e je u detalje izdvojiti dijelove turnusa rada odnosno uo~iti radne zahvate u kojima pilot vje`benik nije dovoljno u~inkovit.

Podaci kori{teni u ovom istra`ivanju dio su opse`nijega mjerenja proizvodnosti rada helikoptera analizom pomo}u GPS-a i GIS-a. To uklju~uje vi{e od 35 radnih dana rada triju razli~itih vrsta helikoptera i podatke s devet razli~itih mjesta istra`ivanja. Podaci GPS-a prikupljeni su sustavom Trimble Geo XT sa WAAS ispravljanjem (korekcijom) i tehnologijom EVEREST postavljenima u helikopteru. Informacije su prikupljene u vremenskim razmacima od 1 sekunde. U tri sje~ine radio je i neiskusan pilot (pilot vje`benik). Istra`ivanje je provedeno ljeti, u sje~inama ~etinja~a u sjeverozapadnom SAD-u (tablica 1). Kori{teni su helikopteri Boing Vertol BV 107 i Sikorsky S-61A, oba nosivosti do 4 tone.

Zabilje`ene su zna~ajne razlike u gotovo svim radnim zahvatima s obzirom na iskustvo pilota (tablica 4). Pilotu vje`beniku trebalo je najmanje 64 % vi{e vremena za dovr{etak jedne radne sastavnice, {to je dodatno pove}ano na 80 ili ~ak 90 % kada je dodana slo`enost u radnim elemenatima (npr. podizanje ili otpu{tanje u`adi za kop~anje, kori{tenje hvatala). Osnovne su krivulje proizvodnosti iza|ene za svaki od radnih zahvata. Neiskustvo je pilota smanjilo proizvodnost sustava rada od 75,9 do 42,8 t/h, odnosno 44 %. Uo~ene su razlike u vremenu potrebnom pilotu vje`beniku pri letu optere}enoga ili neoptere}enoga helikoptera (najve}a brzina leta, ubrzanje i usporavanje letjelice). Ipak, pilotu vje`beniku najvi{e je vremena bilo potrebno prilikom postavljanje helikoptera u sje~ini za kop~anje tereta.

Croat. j. for. eng. 32(2011)2

487

A. Horcher and R. Visser

Using On-Board GPS to Identify Training Needs of Helicopter Pilots (481–488)

Kori{tenje novih tehnologija mo`e pru`iti zna~ajne mogu}nosti za pobolj{anje proizvodnosti postoje}ih sje~nih sustava. Ovo je istra`ivanje pokazalo mogu}nost kori{tenja sustava GPS uklopljenoga u vozilo za bolje budu}e osposobljavanje neiskusnih pilota helikoptera. Omogu}ena je kvalitetnija obuka novih pilota jer su ustanovljene kriti~ne to~ke rada (tijekom turnusa rada, radnih sastavnica i radnih zahvata) na koje treba obratiti vi{e pa`nje pri obuci pilota.

Klju~ne rije~i: izno{enje drva helikopterom, sustav GPS, obuka pilota, proizvodnost

Authors’ address – Adresa autorâ: Andy Horcher e-mail: ahorcher@fs.fed.us Natural Resource Operations Manager USDA Forest Service Savannah River New Ellenton USA

Received (Primljeno): August 04, 2011 Accepted (Prihva}eno): September 5, 2011

488

Assoc. Prof. Rien Visser, PhD. e-mail: rien.visser@canterbury.ac.nz University of Canterbury College of Engineering Private Bag 4800 Christchurch NEW ZEALAND Croat. j. for. eng. 32(2011)2

Original scientific paper – Izvorni znanstveni rad

Integrating Animal and Mechanical Operations in Protected Areas Natascia Magagnotti, Raffaele Spinelli Abstract – Nacrtak The authors tested two alternative treatments for small wood extraction in protected conservation areas, respectively based on direct skidding by small crawler tractors and integrated horse bunching and crawler tractor skidding. The integration of horse bunching with tractor skidding proved cheaper than direct tractor skidding, and allowed extending the distance range of horse skidding. Integration also offers many additional benefits, as it can improve work safety and system sustainability. The performance of the integrated system can be optimized by paying special attention to team balance and by manipulating extraction distance.

In industrialized countries, the number of horse loggers is so small that they may not contribute large wood volumes to the markets: however, the integration of animal and mechanical power may allow making the most efficient use of the few remaining horse logging operations, and increase their contribution to low-impact, cost-effective wood extraction in protected conservation areas. A similar and converse effect could be obtained in developing countries, where integration would allow making the most efficient use of the few available tractors. Efficient use of draught horses may also help increasing horse logger revenues, thus providing a further motivation to stay in business, and contributing to environmentally compatible economic development.

Keywords: logging, protected areas, draught animals, economics

1. Introduction – Uvod In most industrialized countries, the use of draught animals in forest operations represents a curiosity, rather than a technical necessity. The rapid mechanization of all rural activities has brought animal power to the brink of extinction, despite the staunch resistance of its few loyal supporters. Every now and then, the concurrent publication of some works seems to signal a revival of animal power, but the tide cannot be turned. Until now, development pairs with mechanization, and in our modern economies animal power cannot aspire to much more than a small niche. But at least, that will prevent the loss of a huge cultural heritage, which includes genetic resources as well as specific know-how. If the horse was first domesticated in the Bronze Age (Mallory 1997), we are looking at over 5000 years of R&D, an effort that will dwarf the most ambitious of our current research programs. It would be a huge waste to lose this knowledge, if there was some sensible use for it. Many modern authors believe that there are still several reasons for resorting to animal power, and in particCroat. j. for. eng. 32(2011)2

ular to draught horses. Special opportunities could be offered by the current interest in mobilizing non-industrial private forestry (NIPF) resources. These are often too small for cost-effective mechanized harvesting, whose overall cost-efficiency is heavily affected by the fixed cost of moving the operation to the worksite (Väätäinen et al. 2006). That calls for appropriate system design and equipment selection, but the development of dedicated light-weight, low-cost and fully-mechanized operations (Becker et al.

2006) cannot solve all problems, especially when negotiating rough terrain (De Lasaux et al. 2009). Here draught horses are still somewhat popular, even in such an industrialized country as the United States (Toms et al. 1998). Furthermore, animal power can be deployed with much benefit in protected areas (Magagnotti and Spinelli 2011), where it configures as a low-impact alternative to conventional operations (Bahls 1991). Horses can skid through tight spaces in partial cuts with very little damage to residual trees (Thompson and Sturos 1984). Comparative studies have found that the percent of damaged

489

N. Magagnotti and R. Spinelli

Integrating Animal and Mechanical Operations in Protected Areas (489–499)

trees drops to half (Dietz 1981, Schotz 1985) and damage severity to one third (Fickling et al. 1997) when animals are used instead of tractors. Similar considerations are true for soil impacts (Wang 1997, De Paul and Bailly 2005, Shresta et al. 2008), and the lower soil compaction caused by animal logging is indicated by Fries (1977) as the reason why trees growing near animal skid trails show significantly higher yields, compared to similar trees growing near machine skid trails. For these reasons, animal logging is becoming relatively popular in protected areas (Herold et al. 2009) and in the urban fringe (Egan 1998), to the point that some National parks in Italy routinely prescribe animal logging as the sole log extraction method allowed in their most sensitive zones (Proto, personal communication 2010). However, the design of such operations can be improved, with the purpose of maximizing profits, thus providing a solid financial justification for the deployment of draught animals. Draught horses are particularly effective in those bunching tasks that represent the weak spot of most tractor operations (DePaul et al. 2006), and this consideration was the starting point for this study, whose goals are: 1) to determine the performance and the cost of horse bunching in rough terrain; 2) to relate bunching productivity and cost to the main work conditions, such as distance, log volume and crew size; 3) to measure the productivity and cost benefits gained by tractor skidding as a result of pre-bunching; 4) to gauge the potential for improvement of both animal bunching and tractor skidding, and provide guidelines to optimized joint implementation.

2. Materials and methods – Materijal i metode The study was conducted in the Castelli Romani Regional Park, a protected conservation area of strategic importance for its immediate vicinity to the city of Rome, in Italy. The park includes 9000 ha of forests and agricultural crops, and is richly endowed with cultural and historical heritage (www.parcocastelliromani.it). Sweet chestnut (Castanea sativa L.) is the main forest species in the area, favored by the fertile volcanic substrate. Chestnut stands are normally coppiced every 15 – 20 years, with the purpose of producing poles and fencing assortments, particularly appreciated by the many local vine-growers. In its management plan, the Park maintains traditional economic exploitation, but requires that such activities be conducted with the utmost respect for the natural environment. Hence the need for conducting all logging operations with low-impact harvesting techniques, which explains the survival of animal

490

logging and the interest in optimizing rather than replacing it. The study site is described in Table 1, and is representative for the area and for many chestnut stands, which generally grow on the slopes of extinct volcanoes, colonizing rather steep sites. Trees were felled, delimbed and topped by two-man crews, with a main operator felling and topping with a chainsaw, and a helper directing the fall with a pole and delimbing with an axe. Delimbed stems were then pulled downhill with draught horses, which dropped them along the main tractor trails, forming bunches of 5 – 10 pieces. Given the very steep slope, horse teams would use any pre-existing old tracks for climbing uphill to the loading site. If needed, new narrow tracks could be prepared within a very short time, using a pick and a shovel. Loads were dragged downhill, and the animals were trained to move fast and to step aside whenever stopping, so that the load would not hit their rear legs. As the slope gradient increased, choker chains were left longer, to avoid lifting the log ends, thus increasing friction and mitigating the risk of the load slipping downhill and towards the horse. For this very reason, stems would be dragged from the small end (Fig. 1). That would also minimize the need for any new tracks, as trees could be felled towards the old pre-existing mule paths, placed 20 to 40 m apart. In any case, full length stems were more stable than short logs, and had a limited tendency to slip or roll sideways. Two different horse teams were used for the study, each represented by one horse and its driver, because the rough terrain demanded agility and prevented the use of horse pairs. Both animals belonged to the Italian AITPR breed, which offers both strength and speed, since it was specifically selected in the XIX century for pulling coaches (www.caitpr.it). AITPR horses have similar traits to other heavy breeds normally used for logging, such as the Belgian and the Percheron (Pynn 1991), with which they share part of their gene pool.

Both specimens selected for the study were stallions, aged between 7 and 10 years, and weighing about 700 kg. Three drivers were tested, all comparatively young (25 – 35 years of age) but very experienced with horse logging, which they had practiced for at least 5 years. No attempt was made to normalize individual performances by means of productivity ratings, recognizing that normalization or corrections can introduce new sources of errors and uncontrolled variation in the data material (Gullberg 1995).

The study also included the tractor that collected bunched trees and dragged them to the main landing. This was a very simple operation, consisting of a small 44 kW crawler tractor, weighing about 4 tons.

The crawler was manned by one operator, also young and expert. This operation was studied under two Croat. j. for. eng. 32(2011)2

Integrating Animal and Mechanical Operations in Protected Areas (489–499)

N. Magagnotti and R. Spinelli

Table 1 Description of the test site Tablica 1. Opis mjesta istra`ivanja Municipality – Op}ina Province – Pokrajina Altitude – Nadmorska visina Slope gradient – Nagib terena Trail gradient – Nagib vlake Road density – Otvorenost {uma Species – Vrsta drve}a Management – Uzgojni oblik Treatment – Vrsta sje~e Age – Dob Removal – Sje~na gusto}a Residual density – Preostala stabla Tree DBH – Prsni promjer stabla Stem Height – Visina debla Stem volume – Obujam debla

Fig. 1 Delimbed stems were hooked by the small ends and dragged downhill Slika 1. Okresana su debla vezana u tovar tanjim krajem odignutim od tla i vu~ena nizbrdo

Velletri Rome 650 m 61 % 15 % 32 m3/ha Castanea sativa L. Coppice – [ikara Clearcut – ^ista sje~a 18 years – 18 godina 112 m3/ha 925 trees/ha – 925 stabala/ha 85 trees/ha – 85 stabala/ha 0.13 m 9.2 m 0.121 m3

work modes, and namely: skidding with and without horse bunching. In the latter mode, the tractor driver was assisted by a choker man, and reached the loads in the forest, using existing small tracks or natural pathways. The effect of an assistant was also included in the horse bunching study, where observations were divided in two batches, depending on whether the horse team was or was not assisted by a second operator at the hooking site, who prepared the loads and cleared obstacles (Fig. 2). Since the eventual assistant would serve two horse teams at a

Fig. 2 A second operator assisting the horse driver at the loading site Slika 2. Pomo} drugoga radnika vodi~u konja pri slaganju tovara u sje~ini Croat. j. for. eng. 32(2011)2

491

N. Magagnotti and R. Spinelli

Integrating Animal and Mechanical Operations in Protected Areas (489–499)

Table 2 Costing: assumptions, cost centers and total cost Tablica 2. Tro{kovi: pretpostavke, mjesta tro{ka i ukupni tro{ak Unit – Sredstvo za rad Mode – Na~in rada Investment – Investicija Resale – Preprodaja Service life – Razdoblje odr`avanja Utilization – Iskori{tenost Interest rate – Kamatna stopa Depreciation – Pad vrijednosti Interests – Kamate Insurance – Osiguranje Fodder – Krmno bilje Vet – Veterinar Shoeing – Potkivanje Daily care – Dnevna njega Diesel – Dizelsko gorivo Lube – Motorno ulje Repairs – Popravci Total – Ukupno Crew – Broj radnika Labor – Tro{ak radnika Labor – Tro{ak rada Overheads – Op}i tro{kovi Total rate – Ukupna cijena

Euro Euro Years – Godina h/year h/god. % Euro/year Euro/god.

Euro/year Euro/god.

Euro/year Euro/god.

Euro/year Euro/god.

Euro/year Euro/god.

Euro/year Euro/god.

Euro/year Euro/god.

Euro/year Euro/god.

Euro/year Euro/god.

Euro/year Euro/god.

Euro/h n Euro/h Euro/h Euro/h Euro/h

Horse – Konj Horse – Konj Tractor – Traktor Tractor – Traktor No helper Helper No helper Helper Bez pomo}noga radnika S pomo}nim radnikom Bez pomo}noga radnika S pomo}nim radnikom 3500 3500 45000 45000 700 700 13500 13500 10 10 10 10 1000

1000

1000

1000

4

4

4

4

280

280

3150

3150

90

89.6

1233

1233

179

179.2

1233

1233

3000

3000

–

–

400

400

–

–

500

500

–

–

2738

2738

–

–

–

–

6600

6600

–

–

2442

2442

–

–

2520

2520

7.2 1 15 15 4.4 26.6

7.2 1.5 15 22.5 5.9 35.6

17.2 1 15 15 6.4 38.6

17.2 2 15 30 9.4 56.6

Cost in Euro ( ) as on April 30, 2010. – Tro{kovi izra`eni u eurima prema te~aju 30. travnja 2010. (1 = 1.33 US$)

time, only half of its cost was added to the cost of the single horse team when working in the »assisted« mode.

A time-motion study was carried out to evaluate team productivity and to identify the variables that are most likely to affect it, such as extraction distance and payload size (Bergstrand 1991). Each cycle was stop watched individually, separating productive time from delay time (Bjorheden et al. 1995). Extraction distances were determined with a measuring tape. No correction was made for slope gradient, so that these distances represent the actual paths covered by extraction units. Load size was estimated by

492

measuring the diameter at breast height (DBH) of all trees in each cycle. DBH records were converted into stem wood volume using a single-entry tariff table specifically calculated for the purpose, on the basis of 100 sample trees, distributed along all diameter classes. Sample trees were scaled by measuring DBH, total length and diameter at mid-length. Data from individual cycle observations were analyzed with regression techniques in order to calculate meaningful relationships between productive time consumption and work conditions, such as extraction distance and load size. Indicator variables were used to mark differences between treatments (Olsen et al. 1998).

Croat. j. for. eng. 32(2011)2

Integrating Animal and Mechanical Operations in Protected Areas (489–499)

The tractor rate was calculated with the method described by Miyata (1980), on an estimated annual utilization of 1000 scheduled machine hours (SMH) and a depreciation period of 10 years. The costs of fuel, insurance, repair and service were obtained directly from the operator. The cost of the animals was calculated along similar lines, after making some adjustments to account for the difference between machines and living creatures (Akay 2005). The cost of the horse also includes the labor needed for daily animal care, estimated to 0.5 hours per day. In all cases, labor cost was set to 15 SMH-1 inclusive of indirect salary costs. The calculated operational cost of all teams was increased by 20% to account for overhead costs (Hartsough 2003). Further detail on cost calculation is shown in Table 2, where all costs are first presented as annual costs and then converted into hourly costs, to facilitate comparison. These figures are comparable to those recently presented by Blumenstein (2008) and Schroll (2008), after accounting for the lower labor rates of Italy compared to Germany.

The study material consisted of 185 horse turns and 40 tractor turns, necessary for extracting 436 and 263 stems, respectively (70 and 42 m3). Overall, the time study sessions lasted about 50 hours. The experiment was divided in two parts: one for horse bunching and the other for tractor extraction. In both cases, one complete turn would represent a replicate. Replicates were distributed randomly over the full work area, trying to cover the largest range of bunching and extraction distances.

However, readers must note that data were collected from actual commercial operations, which made it very difficult to obtain the balance of factors normally obtained under more artificial conditions. Despite randomization and extensive replication, our

N. Magagnotti and R. Spinelli

results are only representative of the specific case described here, and should not be extrapolated or generalized without much caution. However, these results well represent the actual world, where different systems are seldom used under identical conditions.

3. Results and discussion – Rezultati s diskusijom Table 3 shows the main results obtained from the test. The average net productivity of horse bunching varied between 1.7 and 2.6 m3 SMH–1, depending extraction distance and the presence of a loading assistant at the stump site. Horse bunching did tractor productivity, which jumped from 2.3 to 5.3 m3 SMH–1, despite the longer extraction distance. The system based on horse bunching allowed a 10% reduction of extraction cost, and a potential for reducing skid trail density to one third of the original value.

However, the data shown in Table 3 represent average values recorded under different extraction distances, and a better comparison can be made only after recalculating productivity as a function of distance, using regression analysis. The results are shown in Fig. 3 for horse bunching and in Fig. 4 for tractor skidding. Both regressions are highly significant, but explain only 35 to 55% of the overall variability. That is the effect of delays time, typically erratic and capable of introducing a significant degree of variability in any regression. Better results could be obtained by excluding delay time from the observations, or by spreading it evenly (Spinelli et al.

2009). However, the authors thought that the inclusion of the original delay time records may better

Table 3 Extraction productivity and cost: summary table Tablica 3. U~inkovitost i tro{kovi privla~enja – zbirna tablica Unit – Sredstvo za rad

Horse – Konj No helper Mode – Na~in rada Bez pomo}noga radnika Distance – Udaljenost m 173 3 Stem size – Obujam debla m 0.15 Load size – Veli~ina tovara Pieces – Komada 2.5 3 Load size – Obujam tovara m 0.33 Total time – Ukupno vrijeme min 13.7 Delay – Prekidi % 26.8 3 –1 Productivity – U~inak m SMH 1.73 Cost – Tro{ak Euro/m3 15.4

Horse – Konj Helper S pomo}nim radnikom 99 0.20 2.2 0.39 10.3 33.5 2.60 13.7

Tractor – Traktor Tractor – Traktor Horse bunching Direct skidding Sakupljanje drva konjima Neposredno privla~enje 445 206 0.10 0.19 8.5 4.6 1.30 0.82 17.4 24.4 7.6 24.7 5.31 2.27 7.27 24.94

SMH = Scheduled Machine Time, i.e. worksite time including all delays – Vrijeme stroja na radnome mjestu koje uklju~uje sve prekide rada Croat. j. for. eng. 32(2011)2

493

N. Magagnotti and R. Spinelli

Integrating Animal and Mechanical Operations in Protected Areas (489–499)

Fig. 5 reports the result of a three-way comparison including: a) tractor skidding without horse bunching, b) tractor skidding after horse bunching, over an average bunching distance of 75 m (which is deducted from skidding distance), c) horse extraction up to the maximum recorded distance of 275 m.

Overall cost was calculated after estimating bunching and skidding productivities with the equations in Fig. 3 and 4. The integration of horse bunching and tractor skidding allows significant economies over direct tractor skidding, with savings ranging between 18 and 65%. Among other things, horse bunching allows accumulating larger tractor loads (1.3 vs. 0.8 m3), which limits the effect of extraction distance: therefore, the savings obtained with horse bunching increase with total extraction distance. If the total extraction distance is shorter than 200 m, then horse skidding is the cheapest alternative.

Notes: the curves were calculated based on the equations attached to Fig. 1 and 2; for the integrated horse + tractor treatment it was assumed that the horse would bunch over the first 75 m and the tractor would skid over the remaining distance; maximum horse extraction distance corresponds to the maximum observed in the study.

Fig. 3 Horse bunching productivity as a function of distance and crew size Slika 3. U~inkovitost sakupljanja drva konjima u ovisnosti o udaljenosti i broju radnika represent the inherent variability of the process, and would not invalidate the equations, which do retain a high statistical significance. At any rate, Table 4 reports the time consumption regressions calculated for each time element: readers can use these equations to check the relationship between time consumption for a specific task and the main influencing factors. The authors calculated bunching and extraction productivity both by using the overall productivity equations shown in Fig. 3 and 4, and by using the time consumption equations in Table 4.

The results were very near, and the authors decided to stick with the original all-inclusive productivity equations, which are simpler to use and guarantee a better representation of the observed cycles. It is worth noticing that the bunching productivity increase offered by detaching an assistant to help with loading can barely repay the additional cost of the assistant: therefore, it is best if the horse driver attaches the loads on his own, without additional help.

494

Fig. 4 Tractor skidding productivity as a function of distance and bunching Slika 4. U~inkovitost privla~enja drva traktorima u ovisnosti o udaljenosti i sakupljanju drva Croat. j. for. eng. 32(2011)2

Integrating Animal and Mechanical Operations in Protected Areas (489–499)

N. Magagnotti and R. Spinelli

Table 4 Basic relationships for the analytical calculation of productivity Tablica 4. Osnovni odnosi za analiti~ko izra~unavanje u~inkovitosti Element – Element

Relationship – Odnos Horse bunching – Sakupljanje drva konjima T = 0.243 + 0.016 D + 0.003 DC * D Empty trip – Hod nenatovarenih konja T = 1.445 + 0.800 NP – 0.834 DA Load – Utovar T = – 0.051 + 0.009 D + 0.001 DC * D Loaded trip – Hod natovarenih konja T = 0.450 + 0.303 NP Unload – Istovar 0.47 if no assistant, 0.68 if assistant Delay Factor – Koeficijent prekida 0,47 bez pomo}noga radnika, 0,68 s pomo}nim radnikom 0.33 m3 if no assistant, 0.39 m3 if assistant Load size – Obujam tovara 0,33 m3 bez pomo}noga radnika, 0,39 m3 s pomo}nim radnikom Tractor skidding – Privla~enje drva traktorom T = 0.135 + 0.010 D – 0.002 DB * D Empty trip – Vo`nja praznoga traktora T = – 4.551 + 18.898 V – 12.413 DB * m3 Load – Utovar T = 0.335 + 0.009 D + 0.833 V – 0.834 DB Loaded trip – Vo`nja natovarenoga traktora T = 1.663 + 1.745 V – 1.134 DB Unload – Istovar 0.39 if no horse bunching, 0.12 if horse bunching Delay Factor – Koeficijent prekida 0,39 bez sakupljanja drva konjima, 0,12 sa sakupljanjem drva konjima 0.82 m3 if no horse bunching, 1.30 m3 if horse bunching Load size – Obujam tovara 0,82 m3 bez sakupljanja drva konjima, 1,30 m3 sa sakupljanjem drva konjima

r2

F

p

0.916 989.5 0.340 46.9 0.799 361.1 0.243 58.6 Unpaired t-test Neparni t-test Unpaired t-test Neparni t-test

<.0001 <.0001 <.0001 <.0001

0.966 524.7 0.416 13.2 0.954 250.3 0.324 8.9 Unpaired t-test Neparni t-test Unpaired t-test Neparni t-test

<.0001 <.0001 <.0001 <.0001

0.049 0.007

0.002 0.001

T = time consumption per turn, min×turn-1 – Utro{ak vremena po turnusu, min × turn-1 D = bunching or extraction distance, m – Udaljenost sakupljanja ili priva~enja drva, m NP = number of pieces in the load – Broj komada u tovaru V = load volume size, m3 – Obujam tovara, m3 DA = dummy assistant: 0 if no assistant, 1 if an assistant is detached – Koeficijent pomo}nika: 0 bez pomo}noga radnika, 1 s pomo}nim radnikom DB = dummy bunch: 0 if no horse bunching, 1 if horse bunching – Koeficijent sakupljanja: 0 bez sakupljanja drva konjima, 1 sa sakupljanjem drva konjima DC = dummy crew: 0 if crew A, 1 if crew B – Koeficijent skupine: 0 za skupinu A, 1 za skupinu B Delay Factor = Delay time/Net work time – Koeficijent prekida = vrijeme prekida/efektivno vrijeme (Spinelli i Visser 2009)

4. Discussion – Rasprava

Fig. 5 Total extraction cost as a function of distance and treatment Slika 5. Ukupni tro{kovi u ovisnosti o udaljenosti privla~enja i sredstvima rada Croat. j. for. eng. 32(2011)2

Skidders work best with pre-assembled loads, which explains the enduring success of traditional operations that integrate feller-bunchers and grapple-skidders. However, steep terrain harvesting does not offer favorable conditions to traditional operations, especially when dealing with small trees. Opening a dense trail network may result unprofitable, and will certainly clash with the environmental restrictions imposed on protected areas. On these sites, managers try to contain both the density and the width of extraction trails, which also explains the choice of a small agricultural tractor rather than a rubber-tired skidder. Here, the mechanical alternatives to horse bunching are winches and mini-skidders. Other authors have already compared horse bunching with winching, demonstrating that horses offer a cheaper alternative to tractor-mounted winches (Harstela and Tervo 1981, Hedman 1988) and independent radio controlled winches (Leek 1976, Leinert 1979). As to mini-skidders, most models cannot ne-

495

N. Magagnotti and R. Spinelli

Integrating Animal and Mechanical Operations in Protected Areas (489–499)

gotiate rough terrain, and even in flat terrain horses offer a cheaper service than mini-skidders (Dekking 1984).

Apparently, draught horses are still the best option for tree bunching on steep terrain. In fact, when the extraction distance does not exceed 200 m, draught horses can perform the whole job at a lower cost than crawler tractors, as already reported by Host and Schlieter (1978) some 30 years ago. The fact that these basic relationships have not changed in 30 years is not surprising, because mechanization has replaced animal power in big wood and easy terrain, but is still struggling to find a viable solution to small wood harvesting in steep terrain, especially when environmental performance is a crucial issue.

In industrialized countries, the number of horse loggers is so small that they may not offer this solution on any significant scale, even if they had it: however, the integration of animal and mechanical power may allow making the most efficient use of the few remaining operations. A similar and converse effect could be obtained in developing countries, where integration would allow making the most efficient use of the few available tractors. Efficient use of draught horses may also help increasing horse logger revenues, thus providing a further motivation to stay in business. That may add a further social and historical dimension to conservation, by fostering a work technique which roots deep into local history and culture.

A careful analysis of the bunching sequence may address the question of improvement potential. Any attempts at increasing horse bunching productivity may not follow the same principles normally applied to mechanized operations. The living machine works on muscle power, which gets depleted very quickly and needs frequent rest stops for energy recovery. It is unlikely that one may increase travel speed and/or decrease loading and unloading time, without experiencing a proportional increase in the frequency and duration of rest stops. This was demonstrated by the study, where detaching an assistant to the loading site resulted in a decrease of loading time from 11.9 to 6.6 min per m3, which was followed by an increase of the delay factor (e.g. the ratio between delay time and net work time, see Spinelli and Visser 2009) from 0.65 to 0.76. Both these differences resulted significant to ANOVA testing, with p-values of 0.001 and 0.031, respectively for loading time and delay factors. Apparently, the best strategy to increase horse bunching productivity is the reduction of animal fatigue, by avoiding unnecessary effort. In flat terrain, this would involve the adoption of suitable devices to reduce drag friction, such as sleds, cones or skidding pans (Hedman 1987). These de-

496

vices may not suit steep terrain operations, where friction is required for load control. Here, unnecessary effort can be avoided by reducing extraction distance to the bare minimum, since rest time per turn is significantly correlated (R2 0.55, F 20.8, p < 0.001) to both bunching distance and load size, according to the following equation: T (min turn–1) = 0.486 + 0.028 * distance (m) + 11.540 load (m3). While reducing load size would also reduce productivity, reducing bunching distance will decrease rest time and increase productivity. In turn, the reduction of bunching distance is obtained by integrating animal and mechanical power, which is the very subject of this paper. A further strategy to increase bunching productivity may consist in using more animals per driver, since draught horses can do most of the work independently, without direct human intervention.

This technique is already very common with pack mules, where it certainly offers good results (Ghaffariyan et al. 2009). Its effect with horse skidding could be checked by further studies, which should also determine if the productivity increase is large enough to justify the cost of the additional horse, which represents a 33% increase over the cost of a single horse team. Finally, it must be stressed that any integrated system involves team interaction and the risk for interaction delays: therefore, optimization must also address the crucial questions of organization and team balance, which may be the reason for the low utilization already observed in some integrated animal-machine operations (Shresta et al.

2005).

5. Conclusions – Zaklju~ci The integration of horse bunching with tractor skidding offers a cost-effective solution to small wood extraction in steep terrain and in protected conservation areas. This solution proves cheaper than direct tractor skidding, and allows extending the distance range of horse extraction. The result is a reduction of harvesting cost and skid trail density, both particularly desirable. Integration also offers many additional benefits, as it can improve work safety and system sustainability. Horse bunching is potentially safer than tractor extraction, as drivers can control their horses with voice commands, keeping at a safe distance whenever required (Snoek 2000). Furthermore, horse logging systems are based on renewable resources for 60% of their inputs, whereas tractor systems rely on renewable resources only for 9% of their inputs (Rydberg and Jansén 2002). The performance of the integrated system can be optimized by paying special attention to team balance and by manipulating extraction distance, so Croat. j. for. eng. 32(2011)2

Integrating Animal and Mechanical Operations in Protected Areas (489–499)

N. Magagnotti and R. Spinelli

that horse bunching may occur over reasonably short stretches. This will allow achieving a higher productivity, while making the most efficient use of the few animal crews still available.

Gullberg, T., 1995: Evaluating operator-machine interactions in comparative time studies. Int. J. For. Eng. 7(1): 51–61.

6. References – Literatura

Hartsough, B., 2003: Economics of harvesting to maintain high structural diversity and resulting damage to residual trees. West. J. Appl. For. 18(2): 133–142.

Akay, A., 2005: Determining cost and productivity of using animals in forest harvesting operations. J. Appl. Scie.

Res. 1(2): 190–195.

Harstela, P., Tervo, L., 1981: Bunching of timber by winches and horse. Folia Forestalia 466, 20 p.

Hedman, L., 1987: Tools and equipment for horse logging.

Small Scale For. 1: 10–17.

Bahls, J., 1991: Horsepower logging. Am. For. 97(1/2): 49–59.

Hedman, L., 1988: Skidding with horse to strip road. Small Scale For. 2: 15–19.

Becker, P., Jensen, J., Meinert, D., 2006: Conventional and Mechanized Logging Compared for Ozark Hardwood Forest Thinning: Productivity, Economics, and Environmental Impact. Nort. J. App. For. 23(4): 264–272.

Herold, P., Jutta, J., Scharnhölz, R., 2009: Arbeitspferde im Naturschutz. BfN-Skripten 256. Bonn, Germany, 139 p.

Bergstrand, K.G., 1991: Planning and analysis of forestry operation studies. Skogsarbeten Bulletin 17, 63 p.

Blumenstein, B., 2008: Arbeitswirtschaftliche ehrebungen beim arbeitspferdeeinsatz als kalkulationsgrundlage der produktionsplanung. Thesis, Faculty of Agricultural Ecology and Economy, University of Kassel, Germany.

Björheden, R., Apel, K., Shiba, M., Thompson, M., 1995: IUFRO Forest work study nomenclature. Swedish University of Agricultural Science, Dept. of Operational Efficiency, Garpenberg, 16 p.

Dekking, J., 1984: Goliat, a small tractor with tracks. IEA/ FE/CPC7 Report, 17 p.

De Lasaux, M., Hartsough, B., Spinelli, R., Magagnotti, N., 2009: Small parcel fuel reduction with a low-investment, high-mobility operation. West. J. Appl. For. 24(4): 205–213.

De Paul, M., Bailly, M., 2005: À propos de la pression exercée par les pneus, chenilles et sabots. Forêt Wallonne 78: 21–33.

De Paul, M., Lombaerde, F., Jourez, B., 2006: Approche économique du cheval en forêt. Forêt Wallonne 81: 15–25.

Ficklin, R., Dwyer, J., Cutter, B., Draper, T., 1997: Residual tree damage during selection cuts using two skidding systems in the Missouri Ozarks. Proceedings of the 11th Central Hardwood For. Conf. USDA Forest Service Gen Tech Rep NC-188: 36–46.

Host, J., Schlieter, J., 1978: Low-cost harvesting systems for intensive utilization in small-stem lodgepole pine stands.

USDA Forest Service Research Paper no. INT-201. 20 p.

Leek, N., 1976: Bunching with Radiotir 740 in thinnings.

Nederlands Bosbouw Tijdschrift 48: 151–157.

Leinert, S., 1979: Einsatz veraltungseingener pferde beim vorrücken von schwachholz. Forsttechnische Informationen 1: 4–6.

Magagnotti, N., Spinelli, R., 2011: Financial and energy cost of low-impact wood extraction in environmentally sensitive areas. Ecological Engineering 37(4): 601–606.

Mallory, J., 1997: Encyclopedia of Indo-European Culture.

Ed. Fitzroy Dearborn, London. ISBN 9781884964985 Miyata, E., 1980: Determining fixed and operating costs of logging equipment. General Technical Report NC-55. Forest Service North Central Forest Experiment Station, St. Paul, MN. 14 p.

Olsen, E., Hossain, M., Miller, M., 1998: Statistical Comparison of Methods Used in Harvesting Work Studies. Oregon State University, Forest Research Laboratory, Corvallis, Oregon. Research Contribution n 23. 31 p.

Pynn, L., 1991: Logging with horse power. Can. Geogr. 3: 31–35.

Rickenbach, M., Steele, T., 2006: Logging firms, nonindustrial private forests, and forest parcelization: evidence of firm specialization and its impact on sustainable timber supply. Can. J. For. Res. 36(1): 186–194.

Dietz, P., 1981: Vermeidung und behandlung von rückenschaden. Allgemeine Forstzeitschrift 12: 263–265.

Rydberg, T., Jansén, J., 2002: Comparison of horse and tractor traction using emergy analysis. Ecol. Eng. 19(1): 13–28.

Egan, A., 1998: Clashing values at the urban fringe: is there a niche for horse logging. North Log Timber Process 32: 16–17.

Shresta, S., Rummer, R., Dubois, M., 2005: Utilization and cost of log production from animal logging operations. Int. J. For. Eng. 16: 167–180.

Fries, J., 1977: Sänker skogstractoren tillväten? Skogen 6: 222–224.

Schotz, J., 1985: Bestandesschäden bei der holzernte: forderungen und wünsche des waldbauhaus. Der Forst und Holzwirt, 14: 375–379.

Ghaffaryian, M., Durston, T., Sobhani, H., Mohadjer, M., 2008: Mule logging in northern forests of Iran: a study of productivity, cost and damage to soil and seedlings. Croat.

J. For. Eng. 29(1): 67–75.

Croat. j. for. eng. 32(2011)2

Schroll, E., 2008: Holzrücken mit pferden – Handbuch für die waldarbeit mit pferden. Starke Pferde Verlag, Lemgo, Germany.

497

N. Magagnotti and R. Spinelli

Integrating Animal and Mechanical Operations in Protected Areas (489–499)

Shresta, S., Lanford, B., Rummer, R., Dubois, M., 2008: Soil disturbances from horse/mule logging operations coupled with machines in the Southern United States. Int. J. For. Eng. 19(1): 17–23.

Toms, C., Wilhoit, J., Dubois, M., Bliss, J., Rummer, B., 1998: Animal logging fills important timber harvesting niche in Alabama. Alabama Agricultural Experiment Station. Highlights of Agricultural Research 45(1): 7–8.

Snoeck, B., 2000: Ces chevaux »qui traînent au bois«. Forêt Wallonne 46: 12–23.

Väätäinen, K., Asikainen, A., Sikanen, L., Ala-Fossi, A., 2006: The cost effect of forest machine relocations on logging costs in Finland. For. Studies 45: 135–141.

Spinelli, R., Visser, R., 2009: Analyzing and estimating delays in wood chipping operations. Biomass Bioenergy 33(3): 429–433.

Spinelli, R., Magagnotti, N., Picchi, G., 2009: Complete tree harvesting as an alternative to mulching in early thinnings. Forest Products Journal 59(6): 79–84.

Wang, L., 1997: Assessment of animal skidding and ground machine skidding under mountain conditions. J. For. Eng. 8(2): 57–64.

Sa`etak

Povezivanje animalnoga i strojnoga rada u za{ti}enim podru~jima U ve}ini je razvijenih zemalja upotreba animalne snage pri izvo|enju {umskih radova vrlo rijedak slu~aj. Tek se povremeno pojavi pojedini znanstveni ~lanak koji zagovara ponovnu upotrebu `ivotinja u {umskom radu.

Istra`ivanje upotrebe `ivotinja pri izvo|enju {umskih radova u kombinaciji sa {umskom mehanizacijom po~iva na pretpostavci da vu~ni konji ostvaruju ve}i u~inak pri sakupljanju drva nego rad skidera (DePaul i dr. 2006).

Ciljevi su istra`ivanja: 1) odrediti u~inak i tro{kove sakupljanja drva na te{kom terenu; 2) povezati u~inak sakupljanja drva s utjecajnim ~imbenicima, kao {to su udaljenost privla~enja, obujam tovara, brojnost grupe; 3) odrediti proizvodnost i smanjenje tro{kova traktora koji privla~i sakupljena drva; 4) odrediti mjere za optimiziranje zajedni~ke proizvodnosti `ivotinjskoga sakupljanja drva i privla~enja drva traktorima.

Istra`ivanja su provedena u regionalnom parku Castelli Romani, koji je za{ti}eno podru~je od strate{ke va`nosti zbog neposredne blizine grada Rima. Park se prostire na 9000 ha {umskoga i poljoprivrednoga zemlji{ta, te je na podru~ju bogatoga kulturnoga i povijesnoga naslje|a. Najpro{irenija je vrsta drve}a pitomi kesten (Castanea sativa L.) koji se svakih 15 do 20 godina sije~e ~istom sje~om zbog proizvodnje stupova i kolja za ograde za potrebe lokalnih vinara. Kako je to za{ti}eno brdovito podru~je, potrebno je pri pridobivanju drva koristiti okoli{no prihvatljive tehnologije te se stoga upotrebljavaju vu~ni konji pri izvo|enju {umskih radova. Mjesto istra`ivanja obja{njeno je u tablici 1.

Prilikom istra`ivanja napravljena je studija rada i vremena kako bi se izra~unala u~inkovitost i odredili utjecajni ~imbenici. Studijom rada i vremena snimane su po dvije ina~ice za privla~enje drva traktorom (s pomo}nim radnikom ili bez njega) te dvije ina~ice za sakupljanje drva (s pomo}nim radnikom ili bez njega). Pomo}ni radnik kod traktora imao je ulogu kop~a{a, a u radu s konjima zadatak mu je bio priprema tovara i uklanjanje zapreka s vlake. Sakupljanje i privla~enje drva konjima obavljalo se niz nagib, s tanjim krajem odignutim od tla zbog pove}anja otpora izvla~enja ~ime se izbjegava udar tovara na konjske noge pri kretanju nizbrdo. U istra`ivanju je kori{ten gusjeni~ni traktor, snage motora 44 kW i mase oko 4 tone. Mjernom je vrpcom mjerena stvarna udaljenost privla~enja. Obujam je tovara odre|ivan pomo}u jednoulaznih tablica, napravljenih samo za ovo istra`ivanje na 100 odabranih stabala ravnomjerno raspore|enih po prsnim promjerima. Prikupljeni su podaci dalje obra|eni statisti~kim metodama kako bi se utvrdili odnosi izme|u efektivnoga vremena rada i radnih uvjeta, kao {to su udaljenost privla~enja i obujam tovara.

Izra~un je tro{kova prikazan u tablici 2. Tro{kovi su najprije odre|ivani na godi{njoj razini, a zatim radi lak{e usporedbe prera~unavani su u tro{kove po satu. Ukupni se tro{kovi rada kre}u od 26,6 /h (konj bez pomo}noga radnika) do 56,6 /h (traktor s pomo}nim radnikom). Tijekom istra`ivanja snimljeno je 185 turnusa sakupljanja s konjima i 40 turnusa privla~enja drva traktorom.

Prosje~na se proizvodnost sakupljanja drva konjima kretala od 1,7 m3/h do 2,6 m3/h, ovisno o udaljenosti privla~enja i prisutnosti pomo}noga radnika (slika 3). Sakupljanje drva pove}alo je proizvodnost traktora (slika 4) s 2,3 m3/h na 5,3 m3/h unato~ pove}anju udaljenosti privla~enja (tablica 3).

498

Croat. j. for. eng. 32(2011)2

Integrating Animal and Mechanical Operations in Protected Areas (489–499)

N. Magagnotti and R. Spinelli

Rezultati regresijske analize na slikama 3 i 4 pokazuju veliku zna~ajnost, ali opisuju samo 35 % do 55 % varijabilnosti, jer je velik utjecaj vremena prekida rada na varijabilnost. U~inkovitost se sakupljanja i privla~enja drva izra~unala temeljem jednad`bi dobivenim regresijskom analizom prikazanim na slikama 3 i 4 te tako|er kori{tenjem jednad`bi utro{ka vremena iz tablice 4. Pri tome su dobivene podjednake vrijednosti, ali zbog velike varijabilnosti dobivene regresijskom analizom u daljnjem prora~unu u~inkovitosti kori{tene su samo jednad`be utro{ka vremena.

Na slici 5 prikazani su rezultati usporedbe triju na~ina privla~enja drva: privla~enje drva traktorom bez prethodnoga sakupljanja drva konjima, privla~enje drva traktorom nakon prethodnoga sakupljanja drva konjima (srednja udaljenost sakupljanja 75 m) te izvla~enje drva konjima do udaljenosti izvla~enja od 275 m. Ukupni su tro{kovi odre|eni nakon izra~una u~inka sakupljanja i privla~enja drva pomo}u jednad`bi iz slika 3 i 4. Povezivanje sakupljanja i privla~enja drva omogu}uje zna~ajne u{tede, izme|u 18 % i 65 %, u usporedbi s neposrednim privla~enjem. Me|u ostalim, sakupljanje drva omogu}uje pove}anje tovara traktora (s 0,8 m3 na 1,8 m3), {to ograni~ava u~inak udaljenosti privla~enja, stoga se u{tede napravljene sakupljanjem drva pove}avaju s ukupnom udaljeno{}u privla~enja drva traktorom. Ako je udaljenost privla~enja manja od 200 m, tada je izvla~enje drva konjima jeftinija ina~ica.

Povezivanje sakupljanja i privla~enja drva isplativo je rje{enje u pridobivanju drva maloga obujma na strmim terenima i u za{ti}enim podru~jima. To se rje{enje pokazalo isplativijim od neposrednoga privla~enja drva. Rezultat je toga smanjenje tro{kova pridobivanja drva, smanjenje gusto}e traktorskih putova i vlaka i pove}anje sigurnosti rada. U~inak objedinjenoga sustava mo`e se pobolj{ati ako se obrati pozornost na broj radnika i na udaljenost privla~enja.

Klju~ne rije~i: privla~enje drva, za{ti}ena podru~ja, vu~ne `ivotinje, ekonomika

Authors address – Adresa autora: Raffaele Spinelli, PhD. e-mail: spinelli@ivalsa.cnr.it CNR – Ivalsa Via Madonna del Piano Pal. F I-50019 Sesto Fiorentino ITALY

Received (Primljeno): June 23, 2010 Accepted (Prihva}eno): April 11, 2011 Croat. j. for. eng. 32(2011)2

Natascia Magagnotti, MSc. e-mail: magagnotti@ivalsa.cnr.it CNR – Ivalsa Via Biasi 75 I-38010 San Michele all’Adige (TN) ITALY

499

Original scientific paper – Izvorni znanstveni rad

Utilization Rates and Cost Factors in Timber Harvesting Based on Long-term Machine Data Franz Holzleitner, Karl Stampfer, Rien Visser Abstract – Nacrtak Operating forest machines is not only expensive but accurate monitoring of economic variables can be very difficult. Detailed machine data capture of economic variables within a forest enterprise can be used to support accurate decision making processes, especially costing for new investments.

The objectives of this study were to analyze economic variables of forest machinery based on long-term recorded data from one of the Austrian federal state forest machinery workshops.

The study used data from the enterprise’s resource planning system over the period 2004 to 2008. In total 28 tower yarders, 19 skidders, 12 harvesters and 18 forwarders where analyzed for annual utilization, repair costs, fuel consumption and lubrication costs.

The average annual utilization of all skidders was approximately 1,150 productive machine hours excluding breaks less than 15 minutes (PMH15) per year. Skidders consumed an average 7.3 L/PMH15 with repair costs of 11.4 /PMH15. For the fully mechanized harvesting system the harvesters achieved 2,040 PMH15/year and the forwarders 2,070 PMH15/year.

The annual utilization of cable yarding systems is between 560 PMH15 and 1,500 PMH15.

Keywords: forest machinery, fuel consumption, machine utilization, repair costs

1. Introduction – Uvod In 2009 the Austrian annual timber harvest was 16.9 Mio m3. Due to difficult mountainous terrain only 16% were felled and processed with harvesters.

Extraction of timber from the stand to forest road side is split with 20% being carried out with cable yarding equipment, 49% with skidders, 26% with forwarders, and 5% with other means (Holzeinschlagsmeldung 2009).

Technical machine limitations, as well as social and environmental compatibility, are main decision criterions that restrict system selection. If multiple systems satisfy these criteria then the most cost-effectiveness will be used. In Austria machine cost calculation for timber harvesting is normally based on the FAO-Scheme, combined and adapted with company related data and conditions (FAO-Forestry Paper 99, 1992).

Machine rate estimation itself depends on the quality of information available for it. Jarack (1965) Croat. j. for. eng. 32(2011)2

defined three categories of estimates depending on the sources of data. High quality estimates are calculations based on (1) long-term costs records followed by (2) knowledgeable sources. Estimates with low accuracy (3) are done with not documented or questionable sources or by use of rule-of-thumb values.

Therefore high quality data should be used to assure accurate costing. Improved data recording and analyzing is an elementary component of a basic business strategy like in timber harvesting.

One of the most important factors influencing machine cost calculation is the annual use and utilization rate of forest machinery. Annual utilization rate is the ratio of productive to scheduled machine hours. Machine utilization is affected by different factors such as technical reliability of the machines, weather and road conditions, logistics, proportion of set-up time, and the workers. Such information can support strategic and operational decision making processes within a company, especially accurate costing for new investments.

501

F. Holzleitner et al.

Utilization Rates and Cost Factors in Timber Harvesting Based on Long-term Machine Data (501–508)

ments, fuel quantities and repair and maintenance costs (Table 1).

Both productive and scheduled time was recorded on a daily basis by the crew. Productive time was defined as all machine operating hours including breaks less than 15 minutes in duration (PMH15). Relocation and set-up times are not included in productive machine hours. Scheduled hours include all normal working hours for worker and machine operating in one shift system, but exclude holidays and sick-days and are therefore limited to 1,650 h/year.

For all calculations concerning the annual utilization only machines working a full 12 month period were taken into account.

A spreadsheet-based database was developed to combine and prepare the data. Based on these database relevant variables of different forest machineries were filtered and analyzed. For cumulative hours of machines which are older than the period for analyzing the data (i.e. pre-2004) the starting point was taken from the ÖBF AG resource planning system of the machine itself. Regarding the costs for this longer time period, the consumer price index from the Federal Institute of Statistics was taken to refer the costs back to 2004. This is comparable to the method used by Brinker et al. (2002) to compare harvest machine costs in the USA.

The Austrian Federal Forestry company (ÖBF AG) represents 15% of forest area with a main part under mountainous conditions. It harvests approximately 1.8 Mio m3 per year. Within its organisation it operates two technical divisions that carry out forest operations such as planning, road constructing and timber harvesting. These divisions also offer their services in private forests.

The aim of this paper is to analyse long term machine information from the ÖBF AG to improve data used for cost calculations with a focus on utilization, repair costs, fuel and lubricant consumption. Data set includes harvesters, skidders, forwarders and tower yarders. Additionally factors available for cost calculation of forest machinery are compared with the recorded data.

2. Material and Methods – Materijal i metode The machine rate is usually divided into ownership costs, operating costs, and labor (FAO-Forestry Paper 99, 1992). Operating costs include maintenance and repair costs, fuel and lubricant costs, tire, track, chain, and cable replacement. Maintenance and repair may include everything from simple maintenance items to the periodic overhaul of engine, transmission, clutch, brakes, and other major equipment components (Bushman et al. 1988).

3. Results – Rezultati

2.1 Data capturing – Prikupljanje podataka

3.1 Annual utilization – Godi{nja iskori{tenost

The study used data from the ÖBF AG resource planning system over the period 2004 to 2008. In total 28 tower yarders, 19 skidders, 12 harvesters and 18 forwarders of different brands and payload capabilities were analyzed for annual utilization, repair costs and fuel consumption. For each machine categorized information was recorded including time ele-

Average annual machine use for harvesters was 2,042 productive machine hours. A maximum annual use of 3,120 PMH15 was recorded when operating multi-shifts in wind-throw in Sweden. The majority of their operations are carried out in Austria, but during this time they also contracted abroad due to a big wind-throw event in Scandinavia. The aver-

Table 1 Overview of machines and models observed during data capturing Tablica 1. Pregled po vrstama strojeva i modelima pra}enih tijekom prikupljanja podataka Attributes – Obilje`ja Number of studied machines, n Broj istra`ivanih strojeva, n Number of models, n Broj modela, n Range of engine power, kW Raspon snage motora, kW Range of cumulative operating time 2004 to 2008, PMH15 Raspon kumulativnoga pogonskoga vremena 2004 –2008, h

502

Harvester Harvester

Machine types – Vrste strojeva Forwarder Skidder Forvarder Skider

Tower yarder Stupna `i~ara

12

18

19

28

4

6

6

7

125–204

82–150

75–150

170–330

143–12,937

72–15,349

300–7,102

91–14,948

Croat. j. for. eng. 32(2011)2

Utilization Rates and Cost Factors in Timber Harvesting Based on Long-term Machine Data (501–508)

F. Holzleitner et al.

Table 2 Descriptive statistics concerning annual utilization of machinery, PMH15/year Tablica 2. Deskriptivna statistika godi{nje iskori{tenosti strojeva, pogonskih sati godi{nje Machine type Vrsta stroja Harvester – Harvester Forwarder – Forvarder Skidder – Skider Tower Yarder – Stupna `i~ara

Mean Arit. sredina 2,042 2,068 1,151 1,083

Min.

Max.

938 787 355 541

3,120 4,254 1,619 1,531

St. Dev. Stan. devijacija 408 549 276 202

5 Perc.

5. percentil 1,509 1,467 456 684

95 Perc.

95. percentil 2,737 2,951 1,562 1,398

Scheduled Machine Hours Godi{nji fond sati 3,300 3,300 1,650 1,650

Table 3 Descriptive statistics concerning repair costs of analyzed machinery, /PMH15 Tablica 3. Deskriptivna statistika tro{kova popravaka pra}enih strojeva, /pogonskom satu Machine type Vrsta stroja Harvester – Harvester Forwarder – Forvarder Skidder – Skider Tower Yarder – Stupna `i~ara

Mean Arit. sredina 20.2 11.2 11.4 28.0

Min.

Max.

8.1 2.4 1.4 5.5

46.0 36.3 43.6 68.7

age scheduled machine hours in Austria for a harvester in single shift is 1,650 PMH15. Within this machine group different harvester types show a considerably higher annual machine use than the average.

The reason is that the Austrian Federal Forestry company uses a special work shift model. This system uses two workers and results in the machines being used seven days a week. Using this work shift model the scheduled machine hours are 3,300. The average machine utilization rate for all harvesters was 62% (Table 2).

Forwarders have the largest average productive hours per year with 2,068 PMH15 among the investigated machines. They work the same shift model as harvesters. Variability in the hours is greater than that of harvesters, as forwarders are also used after motor-manual felling. Forwarders have the same annual scheduled machine hours as harvesters, being 3,300 hours. This results in a slightly higher machine utilization rate of 63%.

The results also show for skidder use a clear seasonal effect with higher use in winter time. Because of this effect productive machine hours ranged from 355 to 1,619, and averaged 1,151 PMH15 per year. The annual scheduled machine hours for a skidder are 1,650 hours. This results in a machine utilization rate of 70%.

Tower yarders show the average productive hours of 1,083 PMH15/year. The difference in annual use is clearly visible. The annual scheduled machine hours for a tower yarder are 1,650 hours. This results in a machine utilization rate of 66%. A double shift sysCroat. j. for. eng. 32(2011)2

St. Dev. Stan. devijacija 9.7 5.9 8.3 13.6

5 Perc.

5. percentil 9.4 4.2 3.9 8.8

95 Perc.

95. percentil 41.3 21.1 31.7 57.3

N 36 55 77 91

tem for cable yarding in steep terrain is not possible as light conditions limit the choker-setter and the faller.

3.2 Repair costs – Tro{kovi popravaka Average repair cost for harvesters were 20.2 /PMH15 based on the consumer price index benchmarked back to 2004. Forwarder and skidders had almost the same cost per hour with 11.2 and 11.4 , respectively. Tower yarders are the highest with 28.0 /PMH15 (Table 3). No correlation was found between the amount of repair cost and annual utilization or the summarized utilization per year.

3.3 Fuel consumption and lubricants Potro{nja goriva i maziva Fuel consumption of harvesters ranged from 10.2 to 24.3 L/PMH15 with an average of 15.6 l/PMH15.

The calculations yielded the average fuel consumption for forwarders of 11.1 L/PMH15. Skidders show the lowest consumption rate with 7.3 L/PMH15. They also have lighter engines as shown in Table 1. Tower yarders consume 16.0 L/PMH15. In combination with fuel prices, the fuel cost per hour including cost for lubricants were also analyzed (Table 4).

When analyzing fuel consumption according to engine power, tower yarders show the lowest fuel consumption with 0.06 L per hour and kilowatt, followed by skidders with 0.08 L/kW, PMH15. Forwarders and harvesters are close together with 0.10 and 0.09 L/kW, PMH15 (Table 5).

503

F. Holzleitner et al.

Utilization Rates and Cost Factors in Timber Harvesting Based on Long-term Machine Data (501–508)

Table 4 Descriptive statistics concerning fuel consumption and lubricant costs Tablica 4. Deskriptivna statistika tro{kova goriva i maziva Cost factors Tro{kovni faktori

Machine type Vrsta stroja Fuel consumption, Harvester – Harvester Forwarder – Forvarder L/PMH15 Potro{nja goriva, Skidder – Skider L/pog. satu Tower Yarder – Stupna `i~ara Harvester – Harvester Fuel cost, Forwarder – Forvarder /PMH15 Tro{ak goriva, Skidder – Skider /pog. satu Tower Yarder – Stupna `i~ara Harvester – Harvester Lubricant cost, Forwarder – Forvarder /PMH15 Tro{ak maziva, Skidder – Skider /pog. satu Tower Yarder – Stupna `i~ara Harvester – Harvester Lubricant cost, Forwarder – Forvarder % of fuel cost Tro{ak maziva, Skidder – Skider % od tro{ka goriva Tower Yarder – Stupna `i~ara

Mean Arit. sredina 15.6 11.1 7.3 16.0 13.0 9.2 5.5 12.6 1.6 0.8 0.4 1.7 12.6 7.9 7.2 12.9

Min.

Max.

10.2 1.3 3.6 5.3 8.4 1.1 2.8 3.3 0.3 0.2 0.0 0.2 1.4 2.0 1.0 3.6

24.3 20.5 11.3 24.8 22.7 19.0 10.2 21.9 3.1 2.7 1.0 4.4 30.8 18.2 29.7 32.2

St. Dev. 5 Perc.

Stan. devijac. 5. percentil 3.3 11.3 3.1 7.4 2.1 4.0 4.2 8.9 3.4 9.6 3.1 5.3 1.7 3.0 3.8 6.2 0.8 0.4 0.4 0.3 0.2 0.1 1.0 0.4 7.6 2.0 3.8 3.2 4.3 2.6 6.1 5.1

95 Perc.

95. percentil 23.0 17.4 10.8 23.2 20.3 13.7 8.4 18.6 3.0 1.5 0.8 3.9 28.9 15.5 14.5 25.4

N 36 55 77 91 36 55 77 91 34 53 77 91 34 53 77 91

Table 5 Descriptive statistics concerning fuel consumption per hour dependent on the engine power, L/kW, PMH15 Tablica 5. Deskriptivna statistika potro{nje goriva po pogonskom satu ovisno o snazi motora Machine type Vrsta stroja Harvester – Harvester Forwarder – Forvarder Skidder – Skider Tower Yarder – Stupna `i~ara

Mean Arit. sredina 0.095 0.098 0.075 0.060

Min.

Max.

0.077 0.011 0.031 0.023

0.119 0.146 0.117 0.108

St. Dev. Stan. devijacija 0.012 0.020 0.020 0.017

5 Perc.

5. percentil 0.077 0.077 0.041 0.034

95 Perc.

95. percentil 0.118 0.129 0.108 0.089

N 36 55 77 91

Table 6 Analysis of variances concerning the variables Tablica 6. Analiza varijance odabranih obilje`ja

Adjusted Model – Prilago|eni model Constant – Konstanta Engine power – Snaga motora Machine type – Vrsta stroja Std. Error – Stand. pogre{ka Sum – Zbroj Adjusted Sum of Variation Prilago|ena suma varijabilnosti Sum – Zbroj

Sum of Square Zbroj kvadrata 4,046.514 358.487 448.686 871.332 2,372.846 45,725.424

df Stup. slobode

6,419.360

258

6,134.037

255

Further analysis estimated the rate of fuel consumption per hour depending on the engine power and type of forest machine. The significant covariates are the machine type (MT) and power of en-

504

4 1 1 3 254 259

Mean of Square Varijanca 1,011.628 358.487 448.686 290.444 9.342

F 108.289 38.374 48.029 31.090

Sig. Zna~ajnost <0.001 <0.001 <0.001 <0.001

gines used by different machines. In the next step a model based on the data was developed (Table 6 and Table 7). The model (1) shows an adequate R-Square with 63%.

Croat. j. for. eng. 32(2011)2

Utilization Rates and Cost Factors in Timber Harvesting Based on Long-term Machine Data (501–508)

F. Holzleitner et al.

Table 7 Descriptive statistics concerning engine power of analyzed machine types Tablica 7. Deskriptivna statistika snage motora analiziranih vrsta strojeva Machine type Vrsta stroja Harvester – Harvester Forwarder – Forvarder Skidder – Skider Tower Yarder – Stupna `i~ara All Machines – Svi strojevi

Mean Arit. sredina 162 118 99 271 176

Engine power – Snaga motora, kW Minimum Maximum 5 Perc.

Najmanja vrijednost Najve}a vrijednost 5. percentil 125 204 125 82 150 82 75 150 75 170 330 170 75 330 75

95 Perc.

95. percentil 204 150 150 330 330

With the previous analysis the following model was developed:

4. Discussion – Rasprava

fc = 5.055 + power0.04+CMT

Information regarding machine utilization and repair costs gives forest engineers a useful tool for cost-evaluation in logging operations. This study used long term machine data captured by the ÖBF AG for 4 different machine categories, and each category included multiple machines for a total of 77 machines. The data used were from the period 2004 – 2008.

The average annual use is reported for all machines. The data show that the average utilization rates ranged from a low of 62% for the harvesters up to a high of 70% for skidders. The new working shift model, used by the Austrian Federal Forestry Company for their harvesters and forwarders, shows a clear increase in the annual machine use in comparison to the skidder and tower yarder. According to

(1)

Where: fc fuel consumption, L/PMH15 P engine power, kW CMT coefficient machine type: Tower yarder – CMT = 0 Harvester – CMT = 3.924 Forwarder – CMT = 1.488 Skidder – CMT = –1.744 Based on the developed model, fuel consumption can be estimated for different forest machine types. Fig. 1 shows observed data and estimates. Estimating fuel consumption per hour was just done based on the machine data available.

Fig. 1 Observed fuel consumption and the model based estimates of different forest machinery Slika 1. Snimljena potro{nja goriva i procjene zasnovane na modelu razli~itih {umskih strojeva Croat. j. for. eng. 32(2011)2

505

F. Holzleitner et al.

Utilization Rates and Cost Factors in Timber Harvesting Based on Long-term Machine Data (501–508)

Spinelli et al. (2011), the average annual utilization of harvesters and forwarders with 1,439 and 1,617 hours per year reaches 70% and 78%, respectively, of the value in this study. Spinelli and Visser (2008), based on a large number of separate time studies, determined an average delay for harvesters of 28.9%. Therefore, the utilization rates from that study are almost the same. Repair costs showed a high variability between machine types and with the cumulative operating time.

In Finland the repair and maintenance costs for harvester and forwarder averaged 9.66 /h and 5.06 /h, respectively. The fuel consumption for harvesters and forwarders were about 12.79 L/h and 10.76 L/h. Repair and service costs (classified as variable costs) of a logging contractor with a harvester-forwarder in Finland was studied by Väätäinen et al. (2006) and covered 6.2% of total costs. Maintenance costs defined as fixed costs accounted for 5.3% of total costs.

Pausch (2002) reported fuel consumption of 14.1 L/h for a medium sized harvester compared to 15.4 L/h for the model developed in this study.

Löffler (1991) estimates the average fuel consumption for a forwarder with a medium sized engine of approximately 9.7 L/h compared to 10.9 L/h in this study.

The model for estimating fuel consumption could be improved with further investigation using more detailed data capture. It must also be mentioned that for life cycle analysis, fuel consumption of machinery has to be calculated differently. Self-driven tower yarders already included the fuel necessary for relocating the machinery. Skidders, harvesters and forwarders just show the figures concerning the logging processes without any relocating processes.

Currently data is being recorded manually by the ÖBF AG and their crews. In future data availability and calculation of results could be automated with defined interfaces between costing and enterprise resource planning system. Cost calculation could, therefore, be based on online data from the ERP for new machinery or for harvesting costs. Further work could also include more detailed data capture for investigating parameters influencing the repair costs of forest machines. The results of this paper can also be applied as basic information in life cycle assessment.

506

Acknowledgements – Zahvala The authors would like to thank Erwin Stampfer from the Austrian Federal Forest Agency for providing the data and useful information for this paper.

They also want to thank Mohammad Reza Ghaffariyan for the data preparation.

6. References – Literatura Brinker, R., Kinard, J., Rummer, B., Lanford, B., 2002: Machine rates for selected forest harvesting machines. Circular 296 (Revised). Alabama Agricultural Experiment Station, Auburn University, AL. 32 p.

Bushman, S.P., 1987: Determining labor and equipment costs of logging crews. Department of Forest Engineering, Oregon State University, Corvallis, OR. 123 p.

Holzeinschlagsmeldung, 2009: http://www.lebensministerium.at Jarack, W., 1965. Machine Rate Calculations. American Pulpwood Association, Technical Rel. 77-R-32. 6p.

Löffler, H., 1991: Manuskript zu den Lehrveranstaltungen Forstliche Verfahrenstechnik (Holzernte) für Studierende der Forstwissenschaft. 2. überarb. Auflage. München: Eigenverlag des Lehrstuhls für Forstliche Arbeitswissenschaft und Angewandte Informatik.

Pausch, R., 2002: Ein Systemansatz zur Untersuchung von Zusammenhängen zwischen Waldstruktur, Arbeitsvolumina und Kosten der technischen und biologischen Produktion in Forstrevieren ost- und nordbayrischer Mittelgebirge. Dissertation am Department für Ökologie und Landschaftsmanagement, Lehrstuhl für Forstliche Arbeitswissenschaft und Angewandte Informatik. TU München, 301 p.

Sessions, J., 1992: Cost control in logging and road construction. FAO, Forestry paper: 99, Rome. 121 p.

Spinelli, R., Magagnotti, N., Picchi, G., 2011: Annual use, economic life and residual value of cut-to-length harvesting machines. Journal of Forest Economics, Article in Press.

Spinelli, R., Visser, R., 2008: Analyzing and Estimating Delays in Harvester Operations. International Journal of Forest Engineering 19(1): 35–40.

Väätäinen, K., Asikainen, A., Sikanen, L., Ala-Fossi, A., 2006: The cost effect of forest machine relocations on logging costs in Finland. Forestry studies/Metsanduslikund Uurimused 45: 135–141.

Croat. j. for. eng. 32(2011)2

Utilization Rates and Cost Factors in Timber Harvesting Based on Long-term Machine Data (501–508)

F. Holzleitner et al.

Sa`etak

Iskori{tenost i tro{kovni faktori strojeva pridobivanja drva temeljem dugoro~noga pra}enja Me|u najva`nije ~imbenike koji utje~u na izra~un tro{kova strojnoga rada jest godi{nja uporaba i stopa iskori{tenosti {umskih strojeva. Godi{nja je stopa iskori{tenosti stroja omjer proizvodnoga (pogonskoga) i ukupnoga radnoga vremena. Iskori{tenost je stroja pod utjecajem razli~itih ~imbenika, kao {to su: tehni~ka ispravnost strojeva, vremenski uvjeti, stanje cesta, logistika, priprema rada (npr. monta`a i demonta`a `i~nih linija), radnici.

Takve informacije mogu pomo}i u postupku dono{enja strate{kih i operativnih odluka u tvrtki, a osobito u to~nom utvr|ivanju tro{kova prilikom novih investicija.

Austrijske savezne {ume (ÖBF AG) gospodare s 15 % povr{ine austrijskih {uma, ~iji je glavni dio u planinskim podru~jima te sije~e oko 1,8 mil. m3 godi{nje. Poduze}e je organizirano u dva odjela koji provode {umske zahvate: planiranje radova, izgradnju cesta i pridobivanje drva. Ti odjeli izvode radove i u privatnim {umama.

Cilj je ovoga rada analizirati dugoro~ne podatke o strojevima dobivene od ÖBF-a radi unapre|enja kori{tenja podataka za izra~un tro{kova s naglaskom na godi{nju iskori{tenost strojeva, tro{kove popravaka te potro{nju goriva i maziva. Podaci obuhva}aju harvestere, skidere, forvardere i stupne `i~are. Dodatni podaci o {umskim strojevima dostupni za izra~un tro{kova uspore|eni su s ranijim objavama.

Za istra`ivanje su kori{teni podaci ÖBF-ova sustava planiranja resursa za razdoblje od 2004. do 2008. godine.

Ukupno je 28 stupnih `i~ara, 19 skidera, 12 harvestera i 18 forvardera razli~itih proizvo|a~a i nosivosti tereta analizirano s obzirom na godi{nju iskori{tenost, tro{kove popravaka i potro{nju goriva. Za svaki su stroj snimljeni kategorizirani podaci, koji su obuhva}ali: utro{ke vremena, koli~inu goriva te tro{kove odr`avanja i popravaka.

Proizvodno i ukupno radno vrijeme na dnevnoj osnovi snimali su radnici. Proizvodno (pogonsko) vrijeme odre|eno je kao svi radni sati stroja, uklju~iv{i i prekide rada kra}e od 15 minuta (pogonski sati rada). Premje{tanje strojeva i vrijeme pripreme rada (kod `i~ara monta`a i demonta`a) nisu uklju~eni u proizvodno vrijeme. Ukupno radno vrijeme uklju~uje sve normalne sate rada za radnika i stroj u jednoj smjeni, ali isklju~uje praznike i bolovanja te je stoga ograni~eno na 1650 sati godi{nje. Za sve izra~une povezane s godi{njom iskori{teno{}u uzeti su u obzir samo strojevi koji su radili svih 12 mjeseci u godini.

Baze podataka koje slu`e za pripremu i obradu podataka razvijene su u tabli~nom kalkulatoru. Temeljem tih baza podataka odabrane su i analizirane odgovaraju}e varijable razli~itih {umskih strojeva. Za cjelokupne (kumulativne) sate rada strojeva koji su stariji od razdoblja za koje su analizirani podaci (tj. prije 2004), po~etna je to~ka preuzeta iz ÖBF-ova sustava planiranja resursa za pojedini stroj. Za isto razdoblje podaci su o tro{kovima definirani pomo}u indeksa potro{a~kih cijena dobivenih od Saveznoga zavoda za statistiku.

Prosje~na godi{nja uporaba harvestera iznosi 2042 pogonska sata rada. Najve}a godi{nja uporaba od 3120 pogonskih sati rada zabilje`ena je u vi{esmjenskom radu pri saniranju vjetroizvala u [vedskoj. Ve}inu vremena strojevi su radili u Austriji, ali tijekom promatranoga razdoblja ugovoreni su i radovi u inozemstvu zbog velike koli~ine vjetroizvala u Skandinaviji. Prosje~an fond radnih sati harvestera u Austriji za jednu smjenu iznosi 1650 pogonskih sati rada godi{nje.. Unutar te grupe strojeva razli~iti tipovi harvestera pokazuju zna~ajno vi{u godi{nju uporabu stroja od prosjeka. Razlog je u tome {to Austrijsko savezno {umarsko poduze}e koristi poseban model smjenskoga rada. Taj sustav koristi dva radnika, a rezultat je da strojevi rade sedam dana u tjednu. Primjenom toga modela smjenskoga rada godi{nji fond radnih sati iznosi 3300 sati. Prosje~na je stopa iskori{tenosti harvestera 62 %.

Me|u analiziranim strojevima forvarderi imaju najvi{u prosje~nu ostvarenost od 2068 pogonskih sati rada, pri ~emu rade u istom smjenskom modelu kao i harvesteri. Kod forvardera varijabilnost s obzirom na broj ostvarenih pogonskih sati godi{nje ve}a je nego kod harvestera jer se forvarderi koriste i nakon ru~no-strojne sje~e i izradbe drva. Forvarderi imaju isti fond radnih sati kao i harvesteri (3300 pogonskih sati rada), {to dovodi do malo ve}e iskori{tenosti vozila (63 %).

Rezultati, tako|er, pokazuju jasan sezonski utjecaj pri uporabi skidera s ve}om iskori{teno{}u u zimskom razdoblju. Zbog toga utjecaja pogonski sati rada kre}u se u rasponu od 355 do 1619, a u prosjeku iznose 1151 pogonski sat godi{nje. Fond radnih sati skidera iznosi 1650 sati godi{nje, {to rezultira iskori{teno{}u od 70 %.

Stupne `i~are ostvaruju prosje~no 1083 pogonska sata godi{nje. Razlika u godi{njoj uporabi `i~are jasno je vidljiva. Godi{nji fond radnih sati za stupne `i~are iznosi 1650 sati, {to rezultira stopom iskori{tenosti od 66 %.

Croat. j. for. eng. 32(2011)2

507

F. Holzleitner et al.

Utilization Rates and Cost Factors in Timber Harvesting Based on Long-term Machine Data (501–508)

Sustav dvostrukih smjena za izno{enje drva `i~arom na nagnutom terenu nije mogu} zbog ograni~enja trajanja dnevnoga svjetla za kop~a{a i sjeka~a.

Prosje~ni tro{kovi popravaka za harvester bili su 20,2 po pogonskom satu rada temeljeno na indeksu potro{a~kih cijena iz 2004. godine. Forvarderi i skideri imali su gotovo isti tro{ak po satu, odnosno 11,2 i 11,4 .

Stupne su `i~are imale najvi{i tro{ak od 28,0 po pogonskom satu rada (tablica 3). Nije utvr|ena korelacija izme|u iznosa tro{kova popravaka i godi{nje iskori{tenosti strojeva.

Prosje~na potro{nja goriva harvestera iznosi 15,6 L, a forvardera 11,1 L po pogonskom satu rada. Skideri pokazuju najni`u prosje~nu potro{nju goriva od 7,3 L (u odnosu na druge strojeve imaju i najlak{e motore, tablica 7), a stupne `i~are najvi{u od 16,0 L po pogonskom satu rada. U kombinaciji s cijenom goriva analizirani su i tro{kovi goriva po satu, uklju~uju}i i tro{kove maziva.

Trenuta~no radnici ÖBF-a snimaju podatke ru~no. Dostupnost podataka i izra~un rezultata u budu}nosti mogu biti automatizirani, s definiranim su~eljima izme|u sustava planiranja resursa i tro{kova unutar poduze}a.

Izra~un tro{kova mo`e biti temeljen na online podacima ERP-a za nove strojeve ili za tro{kove pridobivanja drva.

Daljnji rad mo`e tako|er uklju~iti prikupljanje preciznijh podataka za utvr|ivanje parametara koji utje~u na tro{kove popravaka {umskih strojeva. Rezultati ovoga rada mogu se tako|er primijeniti kao temeljna informacija u postupku utvr|ivanja analize `ivotnoga ciklusa {umskih strojeva.

Klju~ne rije~i: {umski strojevi, potro{nja goriva, iskori{tenost stroja, tro{kovi popravaka

Authors’ addresses – Adresa autorâ: Franz Holzleitner, MSc. e-mail: franz.holzleitner@boku.ac.at Assoc. Prof. Karl Stampfer, PhD. e-mail: karl.stampfer@boku.ac.at University of Natural Resources and Applied Life Sciences Vienna Department of Forest and Soil Sciences Institute of Forest Engineering Peter Jordan Strasse 82/3 1190 Vienna AUSTRIA

Received (Primljeno): August 6, 2011 Accepted (Prihva}eno): September 5, 2011

508

Assoc. Prof. Rien Visser, PhD. e-mail: rien.visser@canterbury.ac.nz University of Canterbury College of Engineering Private Bag 4800 Christchurch NEW ZELAND Croat. j. for. eng. 32(2011)2

Original scientific paper – Izvorni znanstveni rad

The Correlation between Long-Term Productivity and Short-Term Performance Ratings of Harvester Operators Thomas Purfürst, Ola Lindroos Abstract – Nacrtak Human operators are key determinants of the performance of most production systems, so individual performance is of intrinsic interest when evaluating current and proposed systems for forest operations. Such evaluations can be useful for diverse purposes, for instance, planning, incentive-setting, control and costing. Hence, various evaluation methods have been developed, all with pros and cons. Here, we compare subjective, short-term ratings of the work-related behavior of 12 harvester operators and their long-term output (harvested volume per unit time), based on observation periods of a few hours and data gathered over two months, respectively. It was found that competent raters can filter the many, interacting behavioral components and translate short-term observations into grades that reflect the operator’s long-term output well (Spearman’s rs > 0.9). Moreover, substantial variations in performance values obtained by both methods were found, probably at least partly attributable to variations in individual performance of both the operators and the raters. We argue that both of the studied methods could be used to adjust population norms (e.g. productivity functions) to the individual’s performance, with sufficient accuracy for normal production purposes (e.g. planning). However, in a scientific context it could be questioned whether the expected uncontrolled variation in operators’ performance is most efficiently minimized by the introduction of uncontrolled variation in rater’s behavior and/or historical data, or if other precautions could be taken to improve the reliability of the data.

Keywords: performance measurement, operator rating, CTL harvester thinning, StanForD, harvester operator, operator influence, human factors

1. Introduction – Uvod The scientific discipline of work science systematically evaluates current and proposed human-machine-environment (HME) systems (Björheden 1991; Wilson 1998), with an intrinsic interest in the performance of systems. However, it is important to realize the ambiguity of the term performance, because it has different meanings depending on context. Interested readers are advised to compare, for instance, organizational-oriented literature (e.g. Tangen 2003; Tangen 2005) and literature oriented towards the focus of this paper; the performance of individuals.

Since human operators are key determinants of the performance of most production systems, individual performance has been intensively investigated by, inter alia, work and organizational psychologists (SonCroat. j. for. eng. 32(2011)2

nestag and Frese 2002). Progress in recent decades has clarified and extended the concept of individual performance, which is now generally considered to be multi-dimensional and dynamic. In fact, methodologies to manage the extreme variations in individual performance that can occur have been proposed (Beheshti and Lollar 2008). Extensive models have also been proposed for exploring the effects of diverse aspects of individual performance. A thorough review of such models is beyond the scope of this paper, but numerous other authors have reviewed them (e.g. Arvey and Murphy 1998; Sonnestag and Frese 2002; Newman et al. 2004; Tubré et al. 2006).

However, to set the context for the study presented here the conceptual framework is roughly outlined below.

509

T. Purfürst and O. Lindroos

The Correlation between Long-Term Productivity and Short-Term Performance ... (509–519)

Two aspects of individual performance are generally differentiated: action and outcome, i.e. the behavior and results of individuals’ actions, respectively (e.g. Arvey and Murphy 1998; Sonnestag and Frese 2002; Newman et al. 2004). Behavior, here, refers to what an individual does in the work situation (e.g. operating a harvester), but not all behavior performed during work is considered within the concept. As with the concept of forest work time differentiation (e.g. Björheden 1991), only behavior that is relevant to the operational goals is considered. Hence, the behavioral aspect of performance is not defined by the action itself, but by judgmental and evaluative processes (Sonnestag and Frese 2002). The outcome aspect refers to work-related results of the individual’s behavior (e.g. harvested volume per unit time). The behavioral and outcome aspects are often related, but outcome depends on external factors in addition to the individual’s behavior, and interactions between behavior and external factors may either enhance or adversely affect individuals’ performance output. The influence of external factors can be readily understood in forestry, because of the strong influences of environmental factors (e.g. tree size and forwarding distance). Moreover, output is known to be highly dynamic (Vöry 1954; Steinlin 1955; Appelroth 1980) and there are substantial differences in output performance between operators; the most productive operators have been found to be at least 114%, 300% and 80% more productive than the least productive operators in manual (Harstela 1975), motor-manual (Reichel 1999) and mechanized work (Purfürst 2009), respectively. However, despite these well-known variations, individual performance is seldom considered in scientific studies. For example, in a review of 53 productivity models for harvester work, Purfürst (2009, p 24) found that 15 only provided information on the experience of the studied operators, three provided some additional information about the operators, and the remaining 35 models did not recognize the operator as an influential factor at all.

The motives for analyzing individual performance depend on whether the objectives are practical or scientific. For a practitioner, individual performance is something that should ideally be enhanced and optimized (Sonnestag and Frese 2002). In contrast, for a forest work scientist, the objective has usually been to normalize the operator influence in order to make valid generalizations for a population of operators. Hence, operator influence has mainly attracted interest as a noise factor. Two approaches for handling inevitable operator effects have been adopted in forest work science (Lindroos 2010). In one, output data are corrected to account for between-operator

510

variations, either objectively or subjectively. Typically, if this is done subjectively, operator behavior (e.g. speed of movements) during the observed task is rated in relation to some kind of norm. Hence, the subjective method is often called performance rating, although it is a very specific application of performance rating, as discussed below. In the other approach, in which output data are not corrected to account for individual behavior, operator blocking is generally applied, i.e. during the observations all operators work with all tested methods/machines (Lindroos 2010).

Subjective correction of output data has a tradition in forestry in continental Europe and Great Britain, but it has been criticized for its subjectivity, i.e. it is biased by the expert’s interpretations even when collecting the data (Mattson Mårn 1953; Steinlin 1955; Kärkkäinen 1975; Samset 1990; Samset 1992), whereas operator-blocking has been criticized for neglecting the well-known variability in human physiological and psychological characteristics (Appelroth 1989; Thompson 1992). A recent contribution that explored the strengths and limitations of operator-blocking (Lindroos 2010) and this work on the uses of performance rating in an updated framework, may (to some extent at least) mediate in the old, and sometimes heated, argument between advocates of the two approaches (see e.g. Sundberg 1988).

In disciplines concerned with the management of human resources, contemporary use of performance assessments is not mainly for minimizing operator influence, but for addressing the underlying factors that contribute to differences in individual performance. Hence, the interest lies in detecting, quantifying and analyzing both behavioral and output aspects of individual performance. In this context, subjective performance ratings by observers play a key role. The strengths and weaknesses of the many set-ups that can be used in performance ratings for these purposes have been scrutinized and discussed at length by work psychologists (e.g. Arvey and Murphy 1998; Sonnestag and Frese 2002; Newman et al. 2004), but surprisingly little attention has been paid to them in forestry work analyses. Some efforts have been made to evaluate and develop methodology for studies concerned with enhancing individual performance (Gellerstedt 2002; Ovaskainen et al.

2004; Ovaskainen 2005; Ovaskainen and Heikkilä 2007), but there has been little corresponding focus on methods for assessing individual performance.

Commonly used methods that enable assessment of (individual) performance in forestry are time studies, performance rating and follow-up studies (analyses of historical output records) (Table 1). In time studies the input-output ratio of an individual is obCroat. j. for. eng. 32(2011)2

The Correlation between Long-Term Productivity and Short-Term Performance ... (509–519) T. Purfürst and O. Lindroos

served, normally during a rather short period of time. Even when there is no intention to correct data prior to analysis, behavior is intrinsically observed during data collection. However, unless it is included in the experimental design (e.g. experiments are specifically designed to compare methods or analyze the work element distribution), behavior is not normally included in analyses except as circumstantial information or for explaining abnormalities in the data. To enable comparisons of individuals’ performance under such settings, there are high requirements for external conditions to be equivalent between time studies, and studies should ideally include repetitions to account for the variation that inevitably occurs, regardless of the care taken to control conditions. Furthermore, including controls and repetitions to exclude as much variation as possible requires substantial time and resources, and time studies are seldom intended to assess individuals (despite the practical potential of such assessments), but rather to predict work performance, often in terms of a productivity norm for a considered cohort of machine operators (cf. standard time for a task).

As previously mentioned, the worker’s behavior is generally either not considered at all, or it is compensated for in the process of synthesizing work study observations into productivity norms. Irrespective of the methodology used, an established norm enables time studies to be conducted in terms of comparing an individual’s observed times with pre-determined norm times for given tasks.

The term performance rating is often used for the time study methodology of establishing productivity norms by subjectively adjusting observed output to account for variations in behavior (see, for instance, Wittering (1973), Bains (1995) or Nieble and Freiwalds (2003) for procedural descriptions). However, here the term refers to the subjective rating of an individual’s performance in relation to other individuals. In an assessment an individual is observed by a rater, normally during a rather short period of time, with the main focus on his/her behavior.

The rater’s assessment should consider external influences, thus decreasing the requirements for equivalent external conditions. Ideally, the rater should also compensate for behavioral variations of an individual. Output is also intrinsically observed during data collection, but is not necessarily a formalized part of the rating (i.e. there is no required measurement of input/output ratios). A limitation of performance rating generally is its strong dependency on the rater’s competence and judgment (i.e. the rater’s performance), since it has long been known (and well documented) that ratings vary between and within raters (e.g. Barnes 1937; Erler 1985; Arvey and Croat. j. for. eng. 32(2011)2

Murphy 1998; Nieble and Freiwalds 2003; Murphy et al. 2004; Roch et al. 2009). However, in work psychology there is optimism regarding the use of subjective rating in assessments of individual performance. The variation in ratings is no longer viewed intrinsically as rater »errors«, but as true variation arising from various sources (Arvey and Murphy 1998). Thus, rating variation can be considered to be a mix of variation in performance by both observed individual and rater. Hence, there is increased recognition that subjective rating does not inevitably introduce rater error or bias, and that rating can often provide valid reflections of individuals’ true performance at low cost (Arvey and Murphy 1998).

Gathering data from observations of normal production activities is the core of follow-up studies.

The methodology applied can range from self-reporting to use of existent records, with the benefit of requiring little resources for gathering data over long periods of time. Generally, long-term data gathering provides more accurate data about normal performance than short-term studies, because infrequent but expected work components are likely to be included. Moreover, the data gathering should not, ideally, interfere with normal work (i.e. there should be no observer effect) and, thus, minimize the well-known effect that individual performance tends to increase when studied (Mayo 1933; Vöry 1954).

However, the level of detail and accuracy of the data acquired are generally lower than when researchers themselves gather data. Computerized automatic data gathering in high-tech forestry machines nowadays offers an attractive alternative for assessing individual performance, since although the data acquired are generally of inferior quality, this is compensated by superior quantity.

Due to the revived acceptance of performance rating, and the emerging potential of automatic gathering of output data, it is of interest to examine the different methodologies of performance assessment in a forest work setting to evaluate their inter-changeability, in order to facilitate the selection of appropriate methodology according to the study objectives. Therefore, here we evaluate the correlations between assessments of individual performance by long-term follow-up studies of output and short-term performance rating of behavior. Based on previous research from other fields, our hypotheses for the study are that: a) the results from the two assessment methods are correlated, and b) the correlation of performance raters’ assessments with follow-up assessments varies between raters.

511

T. Purfürst and O. Lindroos

The Correlation between Long-Term Productivity and Short-Term Performance ... (509–519)

Table 1 Selected forestry work study methods that allow assessment of individual performance, their use of performance components, observational time duration and level of intentionally included subjective elements Tablica 1. Odabrane metode studija rada koje omogu}uju procjenu individualne izvedbe, kori{tene sastavnice, vrijeme trajanja promatranja i razinu hotimi~no uklju~enih subjektivnih elemenata Method Metoda Time study Studij vremena Performance rating Procjena izvedbe Historical data (follow up) Podaci iz pro{losti (dugoro~no pra}enje)

Performance component – Sastavnica izvedbe rada Behavior (action) Output (result) Zahvat (radnja) Output (rezultat)

Duration of observations Trajanje promatranja

Subjective elements Subjektivni elementi

x

X

Short – Kratkoro~no

x – X*

X

x

Short – Kratkoro~no

X

–

X

Long – Dugoro~no

x

Note: The main component observed is indicated by an upper case X, the additional component (if observed) by a lower case x, and lack of observation of either component by –. * Depending on whether or not data are subjectively corrected Bilje{ka: Glavna promatrana sastavnica ozna~ena je velikim slovom X, dodatna sastavnica (ako je promatrana) malim nako{enim slovom x, a nepostojanje promatranja znakom –. * Ovisno o tome jesu li podaci subjektivno korigirani

All of the data were collected in Germany in 2004 – 2006, during first or second thinnings of pre-marked trees in pine-dominated stands in flat terrain with a cut-to-length system, using similarly-sized harvesters (John Deere 1070, Valmet 901 and Ponsse Beaver) to minimize variation due to differences in machinery. All 12 operators had at least two years of relevant work experience at the time of the study, and most had entered their profession via harvester work education at a vocational school. Their age ranged between 20 – 45 years (median 26 years).

ments of harvester operators at the Training Centre for Forest Work and Forestry Technology in Neheim-Hüsten, western Germany. The assessment form included a category of overall performance (i.e. behavior) and 11 subcategories (e.g. harvester head positioning, and speed and carefulness in crane movements). However, in this study only the rating for overall performance was used. For each category, raters graded the operator on a five-level integer scale in which 1 was the best performance and 5 was the worst performance. The scale was constructed so mean performance should correspond to a grade 3.

The raters graded each operator having observed him during 2 – 3 hours of work. Both raters observed the operators under their normal working conditions, but from different locations; rater 1 observed from a distance of ca. 25 – 30 m and rater 2 sat in the cab with the operator, as he normally did during his vocational training. The two raters had no knowledge of the operators’ prior performance and graded independently of each other without any interactions or calibrations of grades or grading. All operators were visited, observed and graded within a period of one week.

2.2 Observers and grading – Procjenitelji i ocjenjivanje

2.3 Historical output data – Izlazni podaci iz pro{losti

Rater 1 was a 28-year-old male work science researcher and had several years of experience of evaluating harvester work. Rater 2 was a 37-year-old male teacher of harvester work in a vocational school and had six years of experience of training and evaluating harvester operators. Raters assessed harvester operators (i.e. individuals) using the form commonly used in vocational training and professional assess-

Output data were collected from normal work through the automatic data recording systems of harvesters. Information on times, dates, harvesting data, operators and software was also collected. The time used for the performance assessment was the productive work time, including interruptions shorter than 15 minutes (PWh15). Data used were stored in defined files according to the StanForD-standard (2007):

2. Material and Methods – Materijal i metode To address the objective and test the hypotheses, two experts each rated 12 harvester operators’ work, then their gradings were compared to the operators’ normalized historical output from the two months preceding the rating observations. Here the study setup is only briefly described, since more details are provided in Purfürst (2009).

2.1 Environmental conditions and individuals Okoli{ni uvjeti i djelatnici

512

Croat. j. for. eng. 32(2011)2

The Correlation between Long-Term Productivity and Short-Term Performance ... (509–519) T. Purfürst and O. Lindroos

Eq. 2 was based on the 12 rated operators’ and an additional 20 operators’ work (in total 32 operators) in thinnings of pine-dominated forest during a data collection period of three years. The function was based on ca. 65 500 work hours distributed among 3 351 stands with mean stem sizes of 0.04 – 0.32 m3 (under bark), and explained a significant (p < 0.001) proportion of the observed variance (R2 = 0.61).

Since the historical data were time-limited, the productivity in individual stands was time-weighted rather than volume-weighted when calculating the mean performance of an operator. This was done using Eq. 3, in which the observed normalized productivity in a given stand was weighted with the stand proportion of the total work time analyzed for the given operator. In Eq. 3, P N is the time-weighted normalized mean productivity of a given operator and ti is the time worked in stand i (PWh15). P N and time-weighted absolute mean productivity values during the two-month period are presented in Table 2.

*.prd Total harvesting production data. *.pri Harvesting production data for each individual log and stem. *.drf Operational monitoring data, covering both work time and repair time data. *.stm Stem data (length and diameter values) To enable indications of long-term performance to be obtained, but avoid including performance trends (e.g. long-term performance changes), the work period compared to the raters’ grade was set to the 60 days preceding the rating observation. The characteristics of the stands where thinning was carried out by operators are summarized in Table 2. To minimize the influence of differences in environmental factors on the data, observed productivity in the stands was normalized (Eq. 1) relative to the mean expected productivity according to:  Pi  PiN =  $  × 100 (%) P

[1]

  n   n P N =  ∑ Pi N × ti  ×  ∑ ti    i=1   i=1

where Pi is the observed productivity in the stand i (m3 under bark /PWh15) and the expected productivity in the stand. was calculated as a function of the stand’s mean stem size (V, m3 under bark) according to Eq. 2 (Purfürst 2009).

P$ = e0.684×ln(V)+3.543 (m3/PWh15)

−1

[3]

(%)

2.4 Statistical analysis – Statisti~ka analiza Due to the categorical features of the rating scale, non-parametric tests were generally used to avoid violating parametric tests assumptions, e.g. of con-

[2]

Table 2 Stand characteristics and time-weighted mean productivity for operators according to data from two months of historical (follow-up) data Tablica 2. Sastojinske zna~ajke i vremenski uprosje~ena proizvodnost za voza~e harvestera naspram povijesnih podataka o dugoro~noj proizvodnosti Operator Voza~

A B C D E F G H I J K L a a

Stands Sastojine

Harvested volume Posje~eni obujam

Work time Vrijeme rada

Stem size Obujam deblovine

N

m3

PWh15

5 11 5 4 8 4 4 5 5 4 3 13

2247 2118 1345 1280 1383 1156 1563 1318 1597 1297 2132 2088

203 136 161 195 181 156 132 142 210 162 213 236

meana[SD] 0.15 (0.02) 0.22 (0.09) 0.16 (0.05) 0.12 (0.07) 0.15 (0.06) 0.09 (0.05) 0.18 (0.13) 0.17 (0.08) 0.10 (0.05) 0.10 (0.05) 0.14 (0.03) 0.13 (0.10)

Productivity, meana[SD] Proizvodnost, arit_sreda[SD] Relative Absolute Apsolutna Relativna, P N m3/PWh15 % 11.1 (1.6) 16.2 (4.6) 8.3 (1.8) 7.0 (2.4) 7.7 (2.1) 7.2 (2.2) 11.5 (5.5) 9.6 (3.1) 7.5 (2.2) 8.1 (2.7) 10.0 (0.8) 9.0 (4.4)

115 (11) 132 (10) 85 (2) 88 (3) 84 (13) 113 (6) 116 (7) 96 (4) 107 (7) 112 (5) 114 (7) 108 (13)

– time-weighted mean and standard deviation (SD, see Eq. 2) – vremenski vagana aritmeti~ka sredina i standardna devijacija (SD, vidi jednad`bu 2)

Croat. j. for. eng. 32(2011)2

513

T. Purfürst and O. Lindroos

The Correlation between Long-Term Productivity and Short-Term Performance ... (509–519)

tinuous and normally distributed values. Relationships were tested with the Spearman correlation test and illustrated with parametric linear regressions.

The significance of differences in ranking between observers was tested with the Wilcoxon signed rank test. In all the non-parametric tests the rank of a given value relative to other values was used instead of the actual value. When assumptions for parametric tests are fulfilled, non-parametric tests can still be used, but they are less good at distinguishing relationships and mean differences. Under such conditions, Wilcoxon and Spearman tests are 5% and 9%, respectively, less efficient than the corresponding parametric tests (paired T-test, and Pearson’s correlation test) (Zar 1996). SPSS 15.0 (SPSS Inc., U.S.A.) was used for all statistical analyses, with the critical significance level set to 5%.

3. Results – Rezultati There were significant negative relationships (p < 0.001) between the long-term relative productivity level and both raters’ separate grading of operators (Fig. 1), indicating a congruency between objective long-term measurements and short-term, subjective ratings.

The level of correlation was generally high (rs > 0.9), but slightly different for the two raters, as illus-

Fig. 1 Relationships between operator performances assessed by normalized long-term, objective productivity data and subjective, short-term rating by the two raters Slika 1. Odnos u~inkovitosti voza~a koju su utvrdila dva procjenitelja normaliziranim dugoro~nim pra}enjem (objektivne proizvodnosti) i kratkoro~nim pra}enjem izvedbe 514

Fig. 2 Relationship between operator’s performance assessed by normalized objective, long-term (two months) productivity data and subjective short-term rating (two raters). Triangles indicate mean values for an operator, and bars indicate standard deviations Slika 2. Odnos u~inkovitosti voza~a utvr|ene normaliziranim dugoro~nim pra}enjem (dva mjeseca) podataka proizvodnosti i subjektivnom kratkoro~nom procjenom izvedbe (dva procjenitelja). Trokuti}i ozna~uju prosje~ne vrijednosti za voza~a, a crtane oznake standardne devijacije trated by the difference in slopes of the linear regression functions in Fig. 1. Combining the two raters’ grading into a mean value for each operator improved the correlation (rs = –0.944, p < 0.001) and the combined regression function had a similar slope to the relationship for rater 2. This was mainly due to the off-the-scale grade 0 that rater 2 gave one operator who was considered especially skillful. With that grade transformed to the scale’s highest grade for skillfulness (1), the relationship between objective, long-term relative productivity and rater 2’s grading remained strong (rs = –0.900, p < 0.001), with a slope approaching that of rater 1 (y = 8.25 – 0.06x). However, this correction did not result in different correlation values for the mean over raters.

When the variation in the long-term, relative productivity data and the variation in rating due to differences between raters were compared, a clear relationship remained between the objective and subjective assessments (Fig. 2). The variation around the mean relationship seemed, however, to be rather large; the SD-based area was ca. 20% wide and two grades high.

There was a significant positive relationship (rs = 0.831, p < 0.001) between the two raters’ grading of operators (Fig. 3). However, rater 1 systematically Croat. j. for. eng. 32(2011)2

The Correlation between Long-Term Productivity and Short-Term Performance ... (509–519) T. Purfürst and O. Lindroos

0.2×(grade of rater 2)) to correspond to the mean grading of rater 1.

Both raters identified that the average performance of the operator sample was above average (population average = 100%, observed median = 110%), because their median grades (rater 1 = 2.3, rater 2 = 2) were lower than the scale center (3) (Fig. 4). However, the raters apparently had different reference values for their distribution of grades, as illustrated by the deviation in the distributions of grades between raters in Fig. 4. Grades given by rater 1 can be transformed to long-term normalized productivity intervals because there is only one overlap in productivity level between grades (Fig. 1). Hence, rater 1 seems to have graded according to: 1 > 114%, 2 = 105 – 114%, 3 = 95 – 104%, 4 = 85 – 94% and 5 < 85%. Overlaps were more frequent for rater 2 and, thus, the rater’s grades could not be readily transformed to long-term, normalized productivity intervals.

Fig. 3 Relationship between the two raters’ grading of harvester operators (N=12; ratings for four operators coincide) Slika 3. Odnos ocjena voza~a harvestera dvaju procjenitelja (N=12, procjene za ~etiri operatora koincidiraju)

4.1 Results – Rezultati

gave higher grades than rater 2 (Wilcoxon Signed Rank test, Z = –2.6, p = 0.008), although the between-grader difference diminished as the grade increased. Based on the regression function in Fig. 3, grades set by rater 2 would have to be increased by (1.2 –

The strong correlation between rater grades and long-term productivity confirms our hypothesis that the results from the two assessment methods are correlated. The results clearly indicate that competent raters can successfully filter the many, interacting behavioral components (e.g. speed and appropriateness

4. Discussion – Rasprava

Fig. 4 Relative distributions of the two raters’ gradings (left panel) and the relative productivity (right panel) of the sample of operators (N = 12). Expected population mean performance is located at grade 3 and 100% relative productivity, respectively Slika 4. Relativne distribucije dvaju procjenitelja (lijevo) i relativna proizvodnost (desno) uzorkovanih voza~a (N = 12). O~ekivana prosje~na vrijednost uzorka smje{tena je kod ocjene 3 i relativne proizvodnosti od 100 % Croat. j. for. eng. 32(2011)2

515

T. Purfürst and O. Lindroos

The Correlation between Long-Term Productivity and Short-Term Performance ... (509–519)

of movements) and translate their rather short-term observations into grades that reflect operators’ long-term output. However, our hypothesis that the correlation between the two assessment methods varies between raters was also confirmed. This result is consistent with the vast body of previous research (e.g.

Barnes 1937; Murphy et al. 2004; Roch et al. 2009) and shows that despite the high concordance between rater grades, grades are not directly interchangeable.

The differences are likely to increase the more graders’ perception of normal work performance vary.

Moreover, the results demonstrate that the variation in assessed individual performance is rather high, irrespective of the method used. This is consistent with expectations of dynamic performance of individuals (e.g. Arvey and Murphy 1998; Sonnestag and Frese 2002). Hence, for a given operator in this study, the observed variation in long-term productivity is most likely due to a combination of variation in the working conditions that is not accounted for by the normalization standard (Eq. 2, i.e. variation in variables other than mean stem size) and variation in individual performance, whereas the variation in an individual’s performance grades is most likely due to performance variation between raters. Hence, there are likely to be two contributory sources of variation in individual performance to the variation in correlation between the methods; one from the operator and one from the rater.

Further, it was interesting to note that both raters clearly deliberately modified their rating scale, according to their perceptions of how to award individuals appropriate grades. Hence, the five-level scale seems to have conflicted with the raters’ internal need for detail to obtain appropriate accuracy. This conflict might have been reduced if the grading alternatives had been increased by, for instance, allowing half-grades (e.g. 2.5). However, the need for accuracy should be balanced against the need for simplicity.

The scale modifications and rating discrepancies indicate that both the construction and use of scales for subjective performance rating require thorough consideration. The use of scales also implies a need for some kind of rating criteria, which are often quite difficult to formulate because knowledge of the most appropriate behavior under the given conditions is required, and in the complex work of operating a harvester in a highly variable forest environment it is not straightforward to identify, or agree upon, the most appropriate behavior (cf. Ovaskainen et al. 2004).

However, the raters obviously drew similar conclusions, although they could have disagreed with each other to some extent if they had provided detailed descriptions of their rating criteria. Such a procedure is consistent with the training and calibration of raters;

516

an obvious thing to do when aiming to ensure that gradings are as similar as possible. Such efforts are likely to decrease the variation within and between raters, but will not eliminate it (e.g. Barnes 1937; Arvey and Murphy 1998).

4.2 Strengths and limitations – Prednosti i ograni~enja To our knowledge, this is the first study to address the correlation between subjective, short-term assessments and objective long-term assessments of individuals in forestry work. Compared to many previous studies on harvester work, the number of observed operators (N = 12) was rather high. However, the number of raters should ideally have been higher. Nevertheless, the results provide interesting and valid indications of correlation between the two assessment methods, although the findings, especially the level of correlation, should not be generalized without proper precautions.

From a philosophical standpoint it can be questioned if any methods are truly objective when considering the inherently present subjective elements in, for instance, selecting and applying methods or in the data analyzes. However, the dichotomy of objectivity-subjectivity used in this study is considered to be justified by the differences in intentional subjective influence in the applied methods and is, thus, a relative definition. The influence of subjective features is limited when applying the established long-term methodology, which is therefore considered the most objective method of the two. In contrast, the application of performance rating is constructed to contain an intrinsic element of subjectivity and is therefore considered the most subjective method of the two.

As mentioned in the introduction, a general limitation of the use of historical data is that quantity is obtained at the expense of quality and control. In this study one of the difficulties (which did not affect the results) was that substantial effort was required to handle variation in software and realization of the StanForD standard between harvester manufacturers. Moreover, despite the information and instructions provided to operators about the follow-up study, there was, as expected, little control over how data were recorded and whether or not operators actually managed their harvester computer in the stipulated manner. However, the data recording for the follow-up study did not differ from their normal recording and reporting procedures, so no new and unfamiliar procedures were introduced. Moreover, the productivity levels observed in the total follow-up material (Eq. 2) was reasonably consistent with those reported in previous studies (e.g. Sirén and Aaltio 2003; Nurminen et al. 2006).

Croat. j. for. eng. 32(2011)2

The Correlation between Long-Term Productivity and Short-Term Performance ... (509–519) T. Purfürst and O. Lindroos

4.3 Practical applications – Prakti~ne primjene When considering the practical applications of the study presented here, it is first necessary to recognize the variation in objectives and available resources for work measurements (e.g. Björheden 1991; Nieble and Freiwalds 2003). For instance, Sanders (1975) pointed out that »Work measurements are carried out for a purpose, which may be planning, incentives, controls, costing or some combination of these, and each purpose has its own precision requirement that must be met.« The studied methods assessed individual performance similarly, and either could be used to acquire relevant data for adapting planning, incentives and costing to an individual level in forestry, i.e. either could be used to adjust population norms to the individual’s performance, with reasonable accuracy for normal production purposes. Moreover, it could also be possible to both assess the need for vocational training and its effects (for control purposes). This could be done rapidly, with little effort, by performance rating, with the proviso that it would not be automatically possible to compare individuals graded by different raters. The analysis of historical data would probably provide a more accurate estimation of performance over time, but would require greater effort, especially if production data were not normally recorded.

In terms of applications in forest work science, one also has to consider both the objective and the need for accuracy. If the purpose of a study is to rate individual performance, either of the two methods can provide relevant information according to our results. However, if the objective is to minimize the operator effect in order to conduct comparative studies or to construct productivity norms, the appropriateness of the two methods is questionable.

If, for instance, it is only possible to compare a machine operated by one individual with another machine operated by another individual, the operator effect has to be handled in order to generalize results. One possible way to do so is to correct output data by dividing acquired values by relative performance levels assessed by either of the two methods.

However, the question is how to ascertain whether such a procedure will decrease and not increase the uncontrolled variation in the data. Hence, before designing such a study, one should consider whether the expected uncontrolled variation in operators’ performance can be most efficiently minimized by introducing uncontrolled variation in a rater’s behavior or in historical data, or if other precautions could be taken to improve the reliability of the data.

Croat. j. for. eng. 32(2011)2

5. References – Literatura Anon., 2007: Standard for forest data and communications – StanForD. Skogsforsk. Uppsala, Sweden. 12 p.

Appelroth, S.-E., 1980: Comparability of work study results.

In Proceedings from the IUFRO Symposium on Stand Establishment Techniques and Technology in Moscow and Riga 3–8 Sept 1979. IUFRO Group 3.02-00, 414–419.

Appelroth, S.-E., 1989: The analysis and interpretation of forest work study results. In Proceedings of a Symposium on the Equipment/Silviculture Interface in Stand Establishment Research and Operations. Jasper, Alberta. Information Report O-X-40, Ontario Region, Forestry Canada, 173–183.

Arvey, R. D., Murphy, K. R., 1998: Performance evaluation in work settings. Annual Review of Psychology 49(1): 141–168.

Bains, A., 1995: Work measurement – the basic principles revisited. Work Study 44(7): 10–14.

Barnes, R., 1937: Motion and Time Study. 1st ed. John Wiley & Sons, Inc. New York. 285 p.

Beheshti, H. M., Lollar, J. G., 2008: Fuzzy logic and performance evaluation: discussion and application. International Journal of Productivity and Performance Management 57(3): 237–246.

Björheden, R., 1991: Basic time concepts for international comparisons of time study reports. International Journal of Forest Engineering 2(2): 33–39.

Erler, J., 1985: Durchschnittlicher Zeitverbrauch oder Zeitbedarf bei Durchschnittsleistung? Zeitschrift für Arbeitswissenschaft 39(3): 166–168.

Gellerstedt, S., 2002: Operation of the single-grip harvester: motor-sensory and cognitive work. International Journal of Forest Engineering 13(2): 35–47.

Harstela, P., 1975: Factors affecting the consumption of working time and the strain on the worker in some forest work methods. A theoretical and empirical analysis.

Communicationes Instituti Forestalis Fenniae 87.2. University of Helsinki, Faculty of Agriculture and Forestry.

Helsinki. 130 p.

Kärkkäinen, M., 1975: Foundations of forest work research.

A critical review. Research notes No. 31. Department of Logging and utilization of forest products, University of Helsinki. Helsinki. 167 p.

Lindroos, O., 2010: Scrutinizing the theory of comparative time studies with operator as a block effect. International Journal of Forest Engineering 21(1): 20–30.

Mattson Mårn, L., 1953: Arbetsstudier – Ett av arbetslärans viktigaste hjälpmedel [Work studies – One of work science’s most important tools]. Skogshögskolan [Royal College of Forestry]. Stockholm. 133 p.

Mayo, E., 1933: The Human Problems of an Industrial Civilization. Macmillan Company. New York.

Murphy, K. R., Cleveland, J. L., Skattebo, A. L., Kinney, T. B., 2004: Raters who pursue different goals give different ratings. Journal of Applied Psychology 89(1): 158–164.

Newman, D. A., Kinney, T. B., Farr, J. L., 2004: Job performance ratings. In Thomas, J. C. (eds) Comprehensive Hand-

517

T. Purfürst and O. Lindroos

The Correlation between Long-Term Productivity and Short-Term Performance ... (509–519)

book of Psychological Assessment, Vol. 4: Industrial/Organizational Assessment. Wiley. New York. p. 956–1008.

Nieble, B., Freiwalds, A., 2003: Methods, Standards and Work Design. 11th ed. McGraw-Hill. Boston, MA. 747 p.

Nurminen, T., Korpunen, H., Uusitalo, J., 2006: Time consumption analysis of the mechanized cut-to-length harvesting system. Silva Fennica 40(2): 335–363.

Ovaskainen, H., Uusitalo, J., Väätäinen, K., 2004: Characteristics and significance of a harvester operators’ working technique in thinning. International Journal of Forest Engineering 15(2): 67–78.

Ovaskainen, H., 2005: Comparison of harvester work in forest and simulator environments. Silva Fennica 39(1): 89–101.

Ovaskainen, H., Heikkilä, M., 2007: Visuospatial cognitive abilities in cut-to-length single-grip timber harvester work.

International Journal of Industrial Ergonomics 37(9–10): 771–780.

Sonnestag, S., Frese, M., 2002: Performance concepts and performance theory. In Sonnestag, S. (eds) Psychological Management of Individual Performance. John Wiley & Sons, Ltd. Chichester, UK. p. 3–25.

Steinlin, H., 1955: Zur Methodik von Feldversuchen im Hauungsbetrieb [Methodology of logging field experiments]. Mitteilungen Band XXXI, heft 2. Schweizerischen Anstalt fur Forstliches Versuchswesen. Zurich. 249–320 p.

Sundberg, U., 1988: The emergence and establishment of forest operations and techniques as a discipline in forest science. Communication 41.8. Norwegian Forest Research Institute. Ås, Norway. 107–137 p.

Tangen, S., 2003: An overview of frequently used performance measures. Work Study 52(7): 347–354.

Tangen, S., 2005: Demystifying productivity and performance. International Journal of Productivity and Performance Management 54(1): 34–46.

Purfürst, F. T., 2009: Der Einfluss des Menchen auf die Leistung von Harvestersystemen [The operator’s influence on harvester productivity]. PhD-thesis. Institut für Forstnutzung und Forsttechnik, Technische Universität Dresden. Dresden, Germany. 307 p. (In German with English summary).

Thompson, M. A., 1992: Observation and analysis of performance in forest work. In Proceedings from the IUFRO International Symposium: Work Study – Measurement and Terminology. Göttingen, Germany 10–12 June 1992.

Institute of Forest Engineering, Georg-August-University of Göttingen. p. 202–219.

Reichel, K., 1999: Relative time studies – An empirical survey based on individual tasks during logging operations.

Allgemeine Forst Und Jagdzeitung 170(8): 143–148.

Tubré, T., Arthur, W., Bennett, W., 2006: General models of job performance: Theory and practice. In Bennett, W. and Woehr, D. (eds) Performance Measurements: Current Perspectives and Future Challenges. Lawrence Erlbaum Associates, Inc. Mahwah, New Jersey. p. 175–204.

Roch, S. G., Paquin, A. R., Littlejohn, T. W., 2009: Do raters agree more on observable items? Human Performance 22(5): 391– 409.

Samset, I., 1990: Some observations on time and performance studies in forestry. Communication 43.5. Norwegian Forest Research Institute. Ås, Norway. 80 p.

Samset, I., 1992: Forest operations as a scientific discipline.

Communication 44.12. Norwegian Forest Research Institute. Ås, Norway. 48 p.

Sanders, N. W., 1975: Precision in work measurements.

Work Study 8: 15–23.

Sirén, M., Aaltio, H., 2003: Productivity and costs of thinning harvesters and harvester-forwarders. International Journal of Forest Engineering 14(1): 39–48.

Wilson, J. R., 1998: A framework and a context for ergonomics methodology. In Wilson, J. R. and Corlett, E. N. (eds) Evaluation of Human Work. Taylor & Francis Ltd. London. p. 1–39.

Wittering, W. O., 1973: Work study in forestry. Forestry Commission Bulletin 47. London, U.K. 100 p.

Vöry, J., 1954: Analysis of the time study materials of some forest jobs. Publication No. 31. The Forest Work Studies Section of the Central Association of Finnish Woodworking Industries, Metsäteho. Helsinki, Finland. 117 p.

Zar, J. H., 1996: Biostatistical Analysis. 3 ed. Prentice-Hall.

Upper Saddle River, N.J. 662 p.

Sa`etak

Odnos dugoro~ne proizvodnosti i kratkoro~ne procjene izvedbe rada voza~a harvestera Radnici su klju~ni i odlu~uju}i ~imbenici provedbe ve}ine proizvodnih sustava pa su njihovi individualni rezultati va`ni pri procjeni sada{njih, ali i predlo`enih sustava {umarskih operacija. Analize radnoga procesa mogu biti korisne za razli~ite namjene, npr. za planiranje, za odre|ivanje stimulacija, za kontrolu ili za odre|ivanje tro{kova proizvodnje.

Zbog toga su s vremenom razvijene razli~ite metode pra}enja rada od kojih sve sadr`e elemente i za njihovu primjenu i protiv njihove primjene. To su tri metode: studij vremena, kratkoro~na procjena izvedbe rada i metoda

518

Croat. j. for. eng. 32(2011)2

The Correlation between Long-Term Productivity and Short-Term Performance ... (509–519) T. Purfürst and O. Lindroos

dugoro~ne proizvodnosti na osnovi vremenskih podataka. Najstarija je metoda metoda studija vremena rada. Metoda pra}enja kratkoro~ne izvedbe subjektivna je metoda na osnovi koje se naj~e{}e donose norme za radove u {umarstvu. Zasniva se na usporedbi razli~itih individualnih podataka o u~inkovitosti i na~inima rada. Pri pra}enju rada procjenitelj subjektivno ocjenjuje djelatnika s naglaskom na njegove na~ine rada, i to u kra}em razdoblju. Negativna je strana te metode velika ovisnost rezultata o kvaliteti i iskustvu procjenitelja. Metoda utvr|ivanja proizvodnosti na osnovi vremenskih podataka zasniva se na prikupljanju i ra{~lambi podataka o stvarno ostvarenom radnom procesu. Op}enito ta metoda nije zahtjevna kao prve dvije kratkoro~ne metode te daje realnije podatke.

U ovom se radu uspore|uju zadnje dvije metode pra}enja rada 12 voza~a harvestera: metoda kratkoro~ne procjene izvedbe rada i njihova dugoro~na proizvodnost (output, obujam posje~enoga drva u razdoblju). Istra`ivanje je provedeno u Njema~koj od 2004. do 2006. godine. Kratkoro~na je procjena rada temeljena na opa`anjima svakih nekoliko sati. Procjenitelj broj 1 bio je 28-godi{nji znanstvenik iz podru~ja prou~avanja rada sa vi{egodi{njim iskustvom pra}enja rada harvestera, dok je procjenitelj 2 bio instruktor za rad harvestera i imao je {estogodi{nje iskustvo u obuci voza~a. Manual za pra}enje imao je ocjenu ukupne radne aktivnosti i jo{ 11 potkategorija ocjene rada voza~a. Dugoro~na je procjena obavljena prihvatom i ra{~lambom podataka iz automatiziranoga ra~unalnoga sustava harvestera uz podatke o utjecajnim ~imbenicima rada. Vrijeme kori{teno u analizi bilo je proizvodno vrijeme rada s prekidima do 15 minuta (PMH15). Podaci su pohranjeni u datotekama u standardu StanForD. Postavljene su dvije hipoteze istra`ivanja: 1) rezultati su usporedbe dviju metoda procjene povezani i 2) zna~ajno su razli~ite procjene izme|u dvaju procjenitelja u odnosu na dugoro~no ostvarenu proizvodnost.

U rezultatima je potvr|ena hipoteza o sna`noj povezanosti dviju metoda procjene. Spoznalo se da utjecajni ~imbenici mogu filtrirati mnoge interaktivne sastavnice radnoga procesa i prevesti kratkoro~na opa`anja u ocjene koje odra`avaju dugoro~nu uspje{nost voza~a (Spearman’s rs > 0,9). Osim toga, utvr|ena su zna~ajna odstupanja u ostvarenim rezultatima rada primjenom obiju metoda, koje se najvjerojatnije mogu djelomi~no pripisati subjektivnoj izvedbi i kod voza~a i kod procjenitelja. Dokazano je da obje prou~avane metode mogu biti primijenjene za utvr|ivanje normi (funkcije proizvodnosti) prema izvedbi pojedinoga voza~a, s dovoljnom to~nosti za normalnu proizvodnju (planiranje rada). Me|utim, u znanstvenom je kontekstu mogu}e ispitati je li predvi|ena nekontrolirana varijacija pri radu voza~a naju~inkovitije minimizirana zbog uvo|enja nekontrolirane varijacije u djelovanju procjenitelja i/ili povijesnih (iskustvenih) podataka, ili se moraju poduzeti druge mjere radi pobolj{anja pouzdanosti ulaznih podataka.

Klju~ne rije~i: mjerenje u~inkovitosti, ocjena operatera, CTL proreda harvesterom, StanForD, voza~ harvestera, utjecaj voza~a, ljudski ~imbenici

Authors’ addresses – Adrese autorâ: Thomas Purfürst, PhD. e-mail: thomas.purfuerst@forst.tu-dresden.de Institute of Forest Utilization and Forest Technology Technische Universität Dresden D-01737 Tharandt GERMANY

Received (Primljeno): September 1, 2010 Accepted (Prihva}eno): January 17, 2011 Croat. j. for. eng. 32(2011)2

Assist. Prof. Ola Lindroos, PhD.* e-mail: ola.lindroos@slu.se Department of Forest Resource Management, Swedish University of Agricultural Sciences, SE-901 83 Umeå SWEDEN * Corresponding author – Glavni autor

519

Original scientific paper – Izvorni znanstveni rad

Bunching with a Self-levelling Feller-Buncher on Steep Terrain for Efficient Yarder Extraction Mauricio Acuna, Justin Skinnell, Tony Evanson, Rick Mitchell Abstract – Nacrtak A research trial was conducted in Victoria, Australia, to evaluate a self-levelling feller-buncher on steep terrain and its potential to improve the overall productivity of steep terrain cable logging. The production study was conducted for a mechanized harvesting system using a Valmet 445 EXL self-levelling tracked feller-buncher and a Madill 124 swing yarder while operating in a clear fell plantation. This study quantified the equipment productivity of steep slope harvesting in a 33 year-old Pinus radiata D. Don (radiata pine) plantation.

Mechanized felling was an integral part of this operation, although there were areas of motor-manually felled trees due to terrain and stream restrictions. Thus the difference in productivity of the yarder for bunched and unbunched trees was quantified.

For an average piece size of 0.8 m3, a productivity of 138 m3/PMH was predicted for the feller-buncher. Bunching substantially improved the productivity of the swing yarder. Mean volume per cycle for the swing yarder was 1.9 m3 for bunched trees versus 1.3 m3 for unbunched trees. For a yarding distance range between 150 and 240 metres, bunching increased the productivity by 25%. These results show the potential of self-levelling feller-bunchers in cable logging operations and suggest that research into mechanised felling be directed towards acquiring more information on the performance of steep terrain feller-bunchers in larger trees sizes, and under other slope and soil conditions in Australia.

Keywords: self-levelling feller-buncher, swing yarder, bunched trees, harvesting productivity

1. Introduction – Uvod Worldwide there is a trend towards increased mechanization of forest harvesting operations. Advantages of mechanized felling include: increased production rate compared to manual felling; providing the opportunity to bunch stems for higher extraction productivity; improved value recovery through reduced stump height and tree breakage; and reducing operator exposure to physical harm (Murphy 2003, Visser 2008, Evanson and Amishev 2010).

Logging contractors have been recently using self-levelling feller-bunchers for steep slope harvesting in cable logging/yarder operations in parts of Australia and New Zealand. Purpose-built level-swing tracked feller-bunchers have been available for more than 30 years and have been used both in Australia and New Zealand clearfell harvesting operations for at least the last 15 years (Evanson 2010). A self-levelCroat. j. for. eng. 32(2011)2

ling feller-buncher increases the payload in comparison to a conventional feller-buncher as, in the latter case, the superstructure tends to swing downhill under the force of gravity with a resultant reduced lifting capacity. Also, tilting the cab too steeply makes it very uncomfortable for the operator (MacDonald 1999). Bunching harvesters not only improve efficiency compared with manual felling, they influence the following cable yarder productivity by concentrating the logs into bunches.

Bunching is not a new concept; its effect on yarder productivity was first investigated in the seventies in the USA (Kellogg 1976). It has been extensively used for improving extraction in plantations and natural forests (Spinelli and Hartsough 2000) and thinning and clearfelling (Bergström et al. 2010). Technology developed in recent years has made it possible to harvest on terrain well over 35 – 45% with the use of feller bunchers and harvesters (Carson et al.

521

M. Acuna et al.

Bunching with a Self-levelling Feller-Buncher on Steep Terrain for Efficient Yarder Extraction (521–531)

1985, Kirk and Kellog 1990, Visser and Stampfer 1998).

Stampfer and Steinmuller (2001) studied a tracked harvester Valmet 911 named »Snake« (whose four single wheels were replaced with trapezoidal tracked undercarriages) on slopes between 22% and 56%.

In comparison to a thinning operation, an 11% increase in productivity was obtained in a clearfell operation for a slope of 36% and tree volume of 0.6 m3.

In a commercial thinning operation with a Syncrofalke yarder in Austria, Heinimann et al. (1998) reported increases in productivity of 25% for a yarder when trees were felled and logs bunched with a Skogsjan 687 harvester. In New Zealand, Amishev and Evanson (2010) investigated the extraction phase of the system that used an excavator log-loader to bunch stems and present them to the grapple yarder.

The use of excavator bunching/presenting resulted in a significantly larger haul size to be extracted than grapple yarding using a spotter (3.2 versus 2.4 trees/ cycle), which accounted for a 33% increase or an estimated 17 m3/PMH extra production.

Based on these good experiences and the interest in mechanized felling and bunching, especially for cable extraction, a research trial was conducted to explore the potential of a mechanized felling/bunching system that could be utilised more extensively in Australia. The aim of this study was to evaluate a self-levelling feller-buncher on steep terrain and its potential to improve the overall productivity of steep terrain cable logging.

2. Materials and methods – Materijal i metode 2.1 Study site and layout – Podru~je istra`ivanja The study site was located near Yarram, on the South Gippsland coast of Victoria, Australia (latitude/longitude: 38°30'45''S/146°33'54''E). The stand was a 33-year-old radiata pine plantation of approximately 1065 trees/ha with no notable understory.

The dry, sedimentary-based soils enabled good traction in the steep terrain. The principle objective of this clearfelling operation was to produce a mixture of sawlog and pulp material. This site had never been thinned or pruned. A 0.58 ha plot containing 618 trees was laid out for observation of the feller-buncher; a pre-treatment description of the harvest plot is given in Table 1. The swing yarder was observed in an adjacent area (of approximately 0.6 ha and separated from the feller-buncher plot by about 50 metres) over a period of two days performing normal operations. A description of the felling of the trees and the layout of the area is presented in Fig. 1.

522

Operational harvest scheduling and equipment allocation made it not possible to conduct the feller-buncher and the swing yarder time study on exactly the same plot. However, both areas were consistently felled and yarded by the same operators and their work methods were identical on both sides. In addition, a visual inspection was also conducted to make sure that the bunches were similar in both areas. Although a detailed inventory was not carried out in the yarding area, plot data collected with the Atlas cruiser inventory system (ATLAS Technology 2010) was provided for the study area to confirm that tree size and the diameter distribution was similar in the feller-buncher and swing yarder area.

The feller-buncher machine with a self-levelling cab was responsible for felling and bunching all trees except those that the machine was unable to fell in a nearby creek due to environmental constraints.

The creek area was demarcated with pegs and tapes to clearly identify the remaining trees that were motor-manually felled and consequently not bunched.

It was not possible to layout two parallel corridors, one with pre-bunched trees and other with no pre-bunched trees, due to the high costs that repre-

Table 1 Pre-treatment description of the harvest plot Tablica 1. Osnovni podaci o mjestu sje~e Plot Attribute Value or range Svojstva istra`ivane plohe Raspon podataka Mean DBH, cm – Srednji prsni promjer, cm 31.5 DBH range, cm – Raspon prsnih promjera stabla, cm 12 to 47 3 3 Mean tree size, m – Prosje~ni obujam stabla, m 0.8 Tree size range, m3 – Raspon obujma stabala, m3 0.14 to 1.89 Mean basal area, m²/ha – Srednja temeljnica, m²/ha 82.6 Ground slope range, % – Nagib terena, % 32 to 47

Fig. 1 Layout of the study area Slika 1. Prikaz podru~ja istra`ivanja Croat. j. for. eng. 32(2011)2

Bunching with a Self-levelling Feller-Buncher on Steep Terrain for Efficient Yarder Extraction (521–531)

Fig. 2 Feller-Buncher »Valmet« 445EXL Slika 2. Gusjeni~no sje~no vozilo Valmet 445 EXL

Fig. 3 Swing Yarder »Madill« 124 Slika 3. [umska `i~ara s hvatalom Madill 124 sented for the contractor to motor-manually fell trees that were located out of creek areas. Although this issue does not affect the productivity comparison between bunched and unbunched trees, it could eventually limit the scope of the results obtained in the study.

2.2 Harvesting system and work method Sustav pridobivanja drva i radne metode The harvesting system comprised a Valmet 445 EXL tracked, swing-to-tree type feller-buncher, a Madill 124 swing yarder and grapple, a Komatsu PC 300 with a Waratah 622 processing head, a Hitachi 280LC excavator loader, a tail hold excavator and a bulldozer. For the purpose of this study only the feller-buncher and the swing yarder were time studied. The feller-buncher was equipped with a Valmet 233 fixed felling head (chain saw) and a self-levelling Croat. j. for. eng. 32(2011)2

M. Acuna et al.

cab up to 27 degrees (Fig. 2). The Madill 124 swing yarder (57.6 tons and 450 HP) was equipped with an 18.3 meters yarding boom, paired with a mobile tailspar (30 ton excavator) (Fig. 3).

Harvesting with the feller-buncher was carried out in parallel extraction tracks that were 15 metres apart. The observed operating method was for the machine to work a felling swath directly up the slope (moving at right-angles to the contour), laying bunches at right-angles to the line of movement. Most of the time, trees were cut when moving uphill, and then slewed to the right (the best visibility for the operator). Trees felled tended to be in the uphill semicircle (from about 270 to 70 degrees) of the machine’s working radius. The operator was able to fell only one tree at a time because of the characteristics of the felling head and the size (DBH and height) of the trees being handled.

The swing yarder was used to haul the trees to a central landing where they were processed into logs, sorted and decked. The yarder was paired with a mobile tailspar, which was a key element in the functionality of this yarding system. To maintain productive cycles, mobility at the back end of a cable operation was equally important. For that purpose, a 30 ton excavator with raised swivelling fairleads was used. The mobile tailspar (excavator) was operated by a man when road changes were needed. This person (»spotter«) also gave radio instructions to the operator (due to lack of sight from the cab) during the yarding phase.

At the top of the corridor, there was a log chute formed in front/beneath the swing yarder where the trees could be stacked until the processor could grab them and begin processing each stem. Once processed, the log loader would sort the logs into their respective place in the log-deck, ready for loading onto trucks. The feller-buncher and hand-faller worked several days ahead of the extraction crew to avoid machine conflicts and ensure there was wood on the ground at all times for extraction.

2.3 Data Collection – Prikupljanje podataka Before data collection began, tree diameters at breast height (DBH) (1.3 m) were measured and marked using a colour coding system. Eight different colour codes were used, in 5 cm classes from 12 cm (± 2.5 cm) to 42 cm (± 2.5 cm). Ground slope was measured at several points, averaging 36% with 47% maximum slope. Over the period of two days, the operation of the feller-buncher was filmed with a video camcorder. As each tree was felled, the colour was recorded in order to identify the felling times by diameter classes later when evaluating the film. Tree size (m3) for individuals was determined from vol-

523

M. Acuna et al.

Bunching with a Self-levelling Feller-Buncher on Steep Terrain for Efficient Yarder Extraction (521–531)

ume equations and coefficients provided by the Atlas cruiser inventory system.

The swing yarder was filmed yarding both bunched and unbunched trees with the number of pieces per cycle being recorded throughout the filming.

Pre-harvest inventory data and some tree measurements were used to calculate the average tree size for the bunched and the unbunched tree areas. Maximum yarding distance was 310 metres, with an average yarding distance of approximately 155 metres (range 25 – 300 metres) for the bunched trees and 195 metres for the unbunched trees (range 150 – 240 metres).

The detailed time study was conducted in the office by reviewing field operations recorded by the camcorder. The software Timer ProTM (Applied Computer Services Inc. 2007) with a PDA (DellTM Axim x51) and a spreadsheet, were used for recording equipment cycle times. Cycle times of the machines were divided into work elements that were considered typical of the harvesting process of each machine. In addition, variables believed to have an impact on the productivity of each piece of equipment were recorded together with the work elements. For

the feller-buncher, this included tree size while for the swing yarder this included number of pieces per cycle, yarding distance and a dummy variable describing if the load was bunched or unbunched.

For two travel cycles, a GPS travel recorder was attached to the inside cab window and also on the felling head itself to record the machine’s travel. The GPS receiver placement on the relatively protected part of the felling head produced improved data in comparison to the receiver attached to the inside cab window, which experienced poor positional data.

2.4 Data analysis – Obrada podataka Data collected with the detailed time study were used to determine the productivity of the feller-buncher and swing yarder. The statistical analysis consisted of simple (feller-buncher) and multiple linear (swing yarder) regression models for predicting cycle times per tree and productivity. In the swing yarder model the dummy variable »Unbunched« took a value of 1 for unbunched trees and 0 for bunched trees. Models were checked against regression assumptions and evaluated with the multiple R-squared, the standard error of the residuals, and

Table 2 Summary of feller-buncher’s time study Tablica 2. Studij rada i vremena sje~noga vozila Work element – Radne sastavnice Move to tree, or Re-position Premje{tanje vozila ka stablu ili izmje{tanje vozila Swing-to-fell – Postavljanje sje~ne glave za ru{enje Cut – Sje~a Swing-to-bunch Postavljanje sje~ne glave za sakupljanje stabala Second cut, or Cut stump Drugi rez ili rez na panju Fell and bunch dead – Sje~a i skupljanje su{aca Adjust bunch – Uhrpavanje Travel – Kretanje Total – Ukupno

No. of Observations Broj mjerenja

Mean time per cycle, sec.

Srednja vremena po radnom ciklusu, s

% of cycle time Postotni udio u turnusu rada, %

305

2.5

12.0

618 618

6.1 3.5

29.2 16.7

618

6.6

31.5

33

0.3

1.4

19 25 5 618

0.4 0.4 1.1 20.9

1.9 1.9 5.2 100.0

Move-to-tree, Re-position: Machine moving uphill in a straight line between successive tree felling and bunching activities, or machine movement laterally, adjusting the move-to-tree line Premje{tanje vozila ka stablu ili izmje{tanje vozila: Vozilo se kre}e ka stablu uz nagib ili po slojnicama izme|u radnih sastavnica ru{enja ili sakupljanja stabala Swing-to-fell: Machine slewing and extending the boom to position the felling head to fell a tree Postavljanje sje~ne glave: Namje{tanje hidrauli~ne dizalice sa sje~nom glavom u najbolji mogu}i polo`aj za sje~u Cut : Saw operation to fell the tree – Sje~a: sje~a stabala Swing-to-bunch: Slewing the felled tree and lower to the ground or onto a bunch Postavljanje sje~ne glave za sakupljanje stabala: Spu{tanje ustavljenoga stabla na tlo ili slaganje u skupinu oborenih stabala Second cut, Cut stump: A second extension of the saw to sever a tree not felled after the first cut, or a cut to lower the height of a stump Drugi rez ili rez na panju: Dugi rez sje~nom glavom ili spu{tanje sje~ne glave ni`e na panju stabla Fell and bunch dead trees: Slewing, cutting and bunching or disposing of a dead tree – Sje~a i skupljanje su{aca: Sje~a, sakupljanje ili uklanjanje su{aca Adjust bunch: Move trees in a bunch to reduce spread of the butts – Slaganje slo`aja: Slaganje stabala u slo`aju radi smanjenja veli~ine slo`aja Travel: Machine movement (downhill) from the end of a felling swath to the start of the next – Kretanje: Kretanje vozila niz nagib ka novoj sje~noj liniji

524

Croat. j. for. eng. 32(2011)2

Bunching with a Self-levelling Feller-Buncher on Steep Terrain for Efficient Yarder Extraction (521–531)

M. Acuna et al.

the F-statistic. The statistically significant difference of the cycle time models for the swing yarder (null hypothesis that coefficient associated with the dummy variable is equal to cero) was determined through an Extra-sum-of-squares F-test. Also, t-tests were conducted to determine the effect of diameter on felling time as well as the effect of DBH on other work elements. All the tests presented in the paper were conducted at the p = 0.05 level of significance. Productivity is reported in delay-free productive machine hours (PMH) following standard methodologies used in harvesting (Nurminen et al. 2006, Acuna and Kellogg 2008).

3. Results – Rezultati 3.1 Time and motion study – Studij vremena i pokreta A total of 618 trees (158 bunching cycles) were timed for the feller-buncher. The time per tree associated with each work element in a full cycle is presented in Table 2. Swing-to-bunch and swing-to-fell were the most time consuming work elements, accounting for 31.5% and 29.5% of the total cycle time, respectively. Average bunch size was 4.2 trees ranging from 2 to 6 trees. On average there were 1.7 moves/bunch, 2.4 trees cut between each move element, and 10.1 seconds/bunch for move-to-tree and reposition elements.

Statistically significant differences of tree diameter class on cut time are presented in Table 3. Several of the mean cut times for the breast height diameter classes were significantly different, indicating a relationship between tree diameter and cut time (Fig. 4).

Also, for the individual 5 cm diameter classes there was no significant difference between swing loaded

Table 3 Average cut times for diameter classes Tablica 3. Prosje~no vrijeme sje~e po debljinskim razredima Tree diameter class, cm Mean cut time, sec. Significant difference* Debljinski razred, cm Srednje vrijeme sje~e, s Signifikantna razlika* 12 1.6 a 17 1.7 a 22 2.0 a 27 2.3 b 32 3.1 c 37 3.8 d 42 4.4 e 47+ 6.5 f * Values with the same letter are not significantly different at p > 0.005 * Vrijednosti iste oznake nemaju signifikantnu razliku za p > 0,005 Croat. j. for. eng. 32(2011)2

Fig. 4 Effect of DBH on cut time Slika 4. Utjecaj prsnoga promjera stabla na vrijeme sje~e and bunch time (swing-to-bunch) and tree diameter.

However, statistically significant differences between all classes were observed when combining data into larger classes (17 – 27 cm, 32 – 37 cm, 42 – 45 cm).

Mean swing-to-bunch times were 5.6, 6.3, and 7.3 seconds for each class, respectively. As expected, larger trees required more time to swing-to-bunch.

Travel time per tree to return to start of the felling swath averaged 1.1 seconds/cycle (5 observations of 100, 177, 143, 174 and 176 seconds). Total machine movement time (including move-to-tree, re-position and travel elements) averaged 3.6 seconds/cycle or 17.2% of total cycle time. A downhill travel speed of 0.61 m/s (2.2 km/hr) was obtained from the data collected with the GPS. Average move-to-tree speed uphill, during felling and bunching was estimated at 0.47 m/s (1.7 km/hr).

A total of 184 haul cycles were collected during the swing yarder’s time study. From this total, 142 cycles were completed from bunched trees and 42 cycles were completed from unbunched trees. The average number of pieces per cycle was 2.3 for the bunched trees and 1.5 for the unbunched trees, with an average volume per cycle of 1.9 m3 (average tree size = 0.81 m3) and 1.3 m3 (average tree size = 0.87 m3), respectively.

On a per cycle basis, drop/hook and outhaul times were 11.9% and 11.8% longer when yarding unbunched trees (Table 4). The bunched trees made a larger target for the operator to hit when dropping the grapple. Also, concentration of the trees in fewer

525

M. Acuna et al.

Bunching with a Self-levelling Feller-Buncher on Steep Terrain for Efficient Yarder Extraction (521–531)

Table 4 Summary of swing yarder’s time study Tablica 4. Studij rada i vremena {umske `i~are Time (min) per cycle (bunched trees) Vrijeme po turnusu rada (slo`ena stabla), min

Time (min) per cycle (unbunched trees) Vrijeme po turnusu rada (razasuta stabla), min

Swing-to-outhaul – Odmicanje

0.31 [14.9%]

0.30 [14.0%]

Outhaul – Pomicanje praznoga hvatala

0.34 [16.7%]

0.38 [17.6%]

Inhaul – Pomicanje punoga hvatala

0.81 [39.7%]

0.82 [38.0%]

Drop/hook – Spu{tanje hvatala i utovar

0.59 [28.7%]

0.66 [30.4%]

Total – Ukupno

2.05 [100.0%]

2.18 [100.0%]

Work element – Radne sastavnice

Swing-to-outhaul: Yarder swing after dropping a load at the landing chute and is ready to start a new outhaul Odmicanje: Odmicanje hvatala {umske `i~are nakon istovara tereta na pomo}nom stovari{tu Outhaul: Grapple movement downhill (empty) until it is lowered down to get a load Pomicanje praznoga hvatala: Pomicanje praznoga hvatala niz nagib do trenutka spu{tanja hvatala zbog utovara Inhaul: Grapple movement uphill with a load of logs until the load is dropped at the landing chute Pomicanje punoga hvatala: Pomicanje hvatala s teretom do trenutka ispu{tanja stabala na pomo}nom stovari{tu Drop/hook: Grapple descending towards the ground until the logs have been secured and the grapple starts moving up towards the landing Spu{tanje hvatala i utovar: Spu{tanje hvatala prema tlu, utovar oborenih stabala do trenutka kretanja punoga hvatala

locations improved overall visibility. Drop and hooking time ranged from 0.21 to 2.64 minutes/cycle for the unbunched trees and from 0.09 to 3.78 minutes/ cycle for the bunched trees. The outhaul time difference is explained by the shorter yarding distance of the bunched trees in comparison with the unbunched trees (155 versus 195 metres). As expected, and considering all the cycles, outhaul time was statistically different at different distance ranges (25 – 100, 100 – 200, 200 – 300 metres). On average, outhaul time increased at a rate of 1.2 seconds for every 10 metres.

No substantial differences between bunched and unbunched trees were revealed in the cycle time for the elements swing-to-haul and inhaul. Swing-to-haul is independent of the yarding distance or turn size. In the case of inhaul, the effect of a shorter yarding distance for the bunched tree system was offset by the greater number of pieces and payload per cycle. On average, for the same yarding distance (between 150 and 240 metres), the inhaul time was reduced by 0.40 minutes when yarding unbunched trees, which resulted from the fewer pieces per cycle and the lower payload hauled to the landing. As depicted in Fig. 5, one or two pieces were hauled in more than 90% of the unbunched tree cycles. This contrasts with the bunched tree cycles where there was an even distribution of one, two, and three pieces (which accounted for about 85% of the cycles), with four and five pieces being hauled in the remaining 15% of the cycles. As in the case of outhaul time, inhaul time was statistically different at different distance ranges (25 – 100, 100 – 200, 200 – 300 metres). On average, inhaul time increased at a rate of 3 seconds for every 10 metres.

526

3.2 Cycle time and productivity models Vrijeme ciklusa rada i modeli proizvodnosti Linear regression models developed to determine the effect of tree size on the feller-buncher’s cycle time per tree, and the effect of bunching, number of pieces per cycle and yarding distance on the yarder’s cycle time per turn are presented in Table 5. Both models met the regression assumptions and all the variables were statistically significant. In the yarder model, the null hypothesis that the coefficient associated with the dummy variable is equal to zero was rejected through the Extra-sum-of squares F-test, indicating a significant difference between the models with and without the dummy variable.

Based on the results obtained with the models, piece size explains 32% of the feller-buncher’s cycle time variance. The number of cycles per PMH drops by 19.1% (from 196.1 to 158.7) when tree size increases from 0.1 to 1.3 m3 (Fig. 6). For an average tree size of 0.8 m3, the model predicts 172.4 cycles (trees)/PMH. In the case of the swing yarder, 62% of the cycle time per turn variance is explained by yarding distance, number of trees per cycle and bunching system. For a yarding distance ranging between 150 and 240 metres, and 2.3 and 1.5 pieces/cycle for the bunched and unbunched tree systems, the cycle time predicted with the model is 0.37 minutes shorter for the unbunched trees than for the bunched trees. On average, this represents 3.2 extra cycles/PMH for the unbunched tree system.

Productivity curves were obtained from the cycle time models developed for the feller-buncher and the swing yarder. Fig. 6 shows the feller-buncher’s productivity curve for a range of tree sizes. For a tree Croat. j. for. eng. 32(2011)2

Bunching with a Self-levelling Feller-Buncher on Steep Terrain for Efficient Yarder Extraction (521–531)

M. Acuna et al.

Table 6 compares the productivity between bunched and unbunched trees for a yarding distance of 180 metres. On average, the cycle time for the bunched trees is 17% longer than for the unbunched trees (2.8 minutes versus 2.4 minutes) with the corresponding fewer number of cycles per PMH (21.5 versus 24.7).

However, the longer average time per cycle for the bunched trees is offset by a 33% increase in the number of pieces per PMH, which results in a 24% increase in the volume yarded per PMH (40.1 m3 for bunched trees versus 32.3 m3 for unbunched trees).

The surface chart in Fig. 8 shows the combined effect of yarding distance and number of pieces per cycle on the swing yarder’s productivity when bunch-

Fig. 5 Proportion of pieces per cycle for bunched and unbunched trees Slika 5. Udio komada tereta u radnom ciklusu za slo`ena i razasuta stabla size range between 0.1 and 1.3 m3, the productivity increases from 19.6 to 206.3 m3/PMH. For an average tree size of 0.8 m3, the model predicts a productivity of 138.0 m3/PMH. Fig. 7 shows the swing yarder’s productivity curve for bunched and unbunched trees using 2.3 and 1.5 pieces/cycle, respectively. For the yarding distance range from where the bunched trees were collected (150 to 240 metres), productivity boosts by 25% (7.5 m3/PMH) when bunched trees are yarded.

The increased productivity of the bunched tree system is proportionally larger as yarding distance increases. Thus, for a yarding distance of 150 and 240 metres, productivity for the bunched trees is 22% and 27% higher than for the unbunched trees.

Fig. 6 Feller-buncher’s productivity curve and cycles per PMH for a range of tree sizes Slika 6. Krivulja proizvodnosti sje~noga vozila

Table 5 Cycle time models for feller-buncher and swing yarder Tablica 5. Modeli radnih ciklusa za sje~no vozilo i {umsku `i~aru Feller-buncher – Sje~no vozilo Cycle time, min/cycle = 0.30 + 0.06 x Tree size, m3 – Vrijeme radnoga ciklusa, min/turnusu = 0.30 + 0.06 x obujam stabla, m3 Residual standard error: 0.042 on 616 df – Preostala standardna pogre{ka: 0,042 za 616 stupnjeva slobode Multiple r2 = 0.32, 618 observations – Multipla regresija r2 = 0,3 za 618 mjerenja F-statistic: 170.1 on 1 and 616 df; p-value = 0 – F-statistika: 170,1 za 1 i 616 stupnjeva slobode; p-vrijednost = 0 Swing yarder – [umska `i~ara Cycle time, min/cycle = 1.11 + 0.01 x Distance, m – 0.05 x N, pieces/cycle – 0.41 x Unbunched (0/1) Vrijeme radnoga ciklusa, min/turnusu = 1,11 + 0,01 x Udaljenost izno{enja (m) – 0,05 x N, komada/turnusu – 0,41 x razasuta stabla (0/1) Residual standard error: 0.504 on 180 df – Preostala standardna pogre{ka: 0,504 za 180 stupnjeva slobode Multiple r2 = 0.62, 184 observations – Multipla regresija r2 = 0,62 za 184 mjerenja F-statistic: 97.48 on 3 and 180 df; p-value = 0 – F-statistika: 97,48 za 3 i 180 stupnjeva slobode; p-vrijednost = 0

Croat. j. for. eng. 32(2011)2

527

M. Acuna et al.

Bunching with a Self-levelling Feller-Buncher on Steep Terrain for Efficient Yarder Extraction (521–531)

Table 6 Productivity measurements for bunched and unbunched trees (yarding distance = 180 metres) Tablica 6. Mjerenja proizvodnosti za sakupljena i razasuta stabla pri udaljenosti privla~enja od 180 m Attribute – Zna~ajke Average time per cycle, min Prosje~no vrijeme trunusa, min Cycles/PMH Turnusi/satu strojnoga rada Average no. pieces/PMH Prosje~an broj komada po satu strojnoga rada Average volume/PMH*, m3 – *Prosje~ni obujam komada, m3 po satu strojnoga rada*

Bunched trees Skupljena stabla

Unbunched trees Razasuta stabla

2.8

2.4

21.5

24.7

49.5

37.1

40.1

32.3

* Based on an average piece size of 0.81 m3 for the bunched trees and 0.87 m3 for the unbunched trees *Prosje~ni obujam komada od 0,81 m3 za sakupljena stabla i 0,87 m3 za razasuta stabla

Fig. 7 Swing yarder’s productivity curve for bunched and unbunched trees Slika 7. Krivulja proizvodnosti {umske `i~are za (ne)sakupljena stabla

for yarding distance of 300 metres (from 12 m3/PMH to 63 m3/PMH). The figure also shows that in proportion, yarding distance has a slightly greater effect on productivity when more pieces per cycle are hauled. Thus, when yarding distance increases from 30 to 300 metres, there is a 3-fold increase in productivity for cycles where one piece is yarded (from 12 m3/PMH to 36 m3/PMH), and a 3.3-fold increase in productivity for cycles where five pieces are yarded (from 63 m3/PMH to 209 m3/PMH).

4. Discussion – Rasprava

Fig. 8 Effect of yarding distance and number of pieces per cycle on the swing yarder’s productivity (bunched trees) Slika 8. Utjecaj udaljenosti izno{enja drva i broja komada po radnom turnusu na proizvodnost {umske `i~are (za sakupljena stabla) ed trees are yarded. For an average yarding distance of 150 metres, there is a five-fold increase in productivity (from 19 m3/PMH to 103 m3/PMH), when the number of pieces per cycle increases from 1 to 5. The effect of pieces per cycle is slightly bigger with shorter yarding distances. Thus, when tree size increases from one to five, productivity boosts about 5.86 times for a yarding distance of 30 metres (from 36 m3/PMH to 210 m3/PMH) and about 5.25 times

528

The aim of this study was to evaluate a self-levelling feller-buncher on steep terrain and its potential to improve the overall productivity of steep terrain cable logging.

Previous studies have identified tree size as a major issue with tracked feller-buncher performance (Acuna and Kellogg 2008). Both tree size (mass) and DBH affect cutting time, and the ability to swing and bunch or drop the tree. Previous studies in medium to large tree size clearfelling operations in Australia and New Zealand have compared productivity rates of self-levelling feller-bunchers. In a recent study, in atypical New Zealand conditions of high stocking (736 stems/ha) and small tree sizes (1.0 m3), a Valmet 445 EXL equipped with a Satco 630 felling head achieved a productivity of 100 trees/PMH. Slopes travelled averaged 19.4% and move time in this stocking comprised 16% of total cycle time (Evanson 2008).

This current study confirmed that both felling (cut-time) and bunching (swing-to-bunch time) were significantly affected by DBH and tree size. In our study, maximum tree size of around 50 cm DBH did Croat. j. for. eng. 32(2011)2

Bunching with a Self-levelling Feller-Buncher on Steep Terrain for Efficient Yarder Extraction (521–531)

not appear to present any problems for the machine and larger trees were felled and bunched using the same methods as average sized trees (31.5 cm DBH).

We are aware that the sole inclusion of tree size, although statistically significant, affected the predictive capability of the feller-buncher’s cycle time and productivity models, with the corresponding low R-squared values. Some researchers (e.g. Pan et al.

2008) have developed models that include additional independent variables such as »move to tree distance« and »move to bunch distance«, which have resulted in more accurate cycle time models and high R-squared values. However, these variables are time consuming and difficult to collect in the field, and the use of these models are limited for operational staff when felling takes place in different harvest and forest conditions.

Feller-buncher performance is also affected by stocking. In our study, the high stocking of 1000 stems per hectare enabled a high ratio of trees to be felled per move-to-tree element (average 2.4). Move-to-tree time was also affected by the required bunch size. The average bunch size was four trees (varying from two to six trees depending on tree size) to try to match the grapple capacity so that by each haul the grapple could extract a complete bunch for maximum efficiency.

For the swing yarder, total cycle time per turn increased by 12% when yarding bunched trees, mainly due to the longer inhaul time involved when yarding more volume per cycle. However, on a per cycle basis, drop/hook for bunched trees was 11% shorter than for unbunched trees. The easily visible bunches provided a larger and easier target for the yarder operator to engage the grapple which reduced drop/ hook times for the bunched trees. It is clear though that some time was used in making bunch sizes that were not suited to the grapple capacity. Although the feller-buncher was able to produce bunches with an average of 4.3 trees, only 2.3 trees/cycle were hauled to the landing. This is mainly explained by the holding capacity of the grapple and the necessity for the operator to maintain yarder productivity without spending excessive time hooking logs.

Despite the fact that bunching could eventually affect the feller-buncher productivity and that bunch sizes were not suited to the grapple capacity, the most noticeable difference observed between bunched and unbunched trees was the greater number of pieces per cycle in bunched stems (2.3 versus 1.5), which on average resulted in a 25% increase in productivity for the swing yarder. The results are very similar to some reported in previous studies. In New Zealand, different breakout methods resulted in different number of trees hauled per cycle (Evanson Croat. j. for. eng. 32(2011)2

M. Acuna et al.

and Amishev 2010). For an average tree size of 0.85 m3, 2.4 trees/cycle were hauled when the trees were grappled using a spotter and 1.5 trees/cycle (mainly unbunched) were hauled when trees were grappled by the yarder operator only. The use of excavator bunching/presenting resulted in a 33% increase (17 m3/ PMH) extra production.

Both yarding distance and number of pieces per cycle showed to have an important impact on the yarding productivity, especially at shorter distances and when more pieces per cycle were yarded. Yarding productivity was slightly more sensitive to the number of pieces per cycle than to yarding distance.

Although cycle times increased by 15% when bunched trees were yarded (more pieces per cycle), this effect was offset by the additional volume per cycle, which in turn resulted in a higher productivity per PMH. These results are consistent with other studies found in the literature. In Canada, Peterson (1987) reported a 57% increase in the number of pieces yarded per PMH when bunched trees were yarded (average piece size = 0.75 m3). For a yarding distance of 150 metres, cycle time for bunched trees was 5% longer than for unbunched trees, mainly explained by a 20% increase in inhaul time.

5. Conclusions – Zaklju~ci Results of this study indicate that in good conditions (relatively small clearfell tree size – average 0.8 m3 – and dry, sedimentary-based soils that enabled good traction on slopes of 36 to 47%) a high production rate can be achieved by a tracked self-levelling feller-buncher.

Mechanical felling and bunching operations are particularly advantageous if working in smaller tree sizes because extraction efficiency can be improved through bunching for optimal yarding sizes. For an average piece size of 0.8 m3, a productivity of 138 m3/PMH was predicted for the feller-buncher.

Bunching improved substantially the productivity of the swing yarder. The mean volume per cycle for the swing yarder was 1.9 m3 for the bunched trees versus 1.3 m3 for the unbunched trees. For a yarding distance range between 150 and 240 metres, bunching increased the productivity of the swing yarder by 25%.

Despite the limitations of our study (one stand with specific terrain and forest conditions, study layout, productivity model for the feller-buncher based solely on tree size, no time study of felling without bunching), these results show the potential of self-levelling feller-bunchers in cable logging operations and suggest that research into mechanized felling be directed towards acquiring more information on the

529

M. Acuna et al.

Bunching with a Self-levelling Feller-Buncher on Steep Terrain for Efficient Yarder Extraction (521–531)

performance of steep terrain feller-bunchers in larger pieces size, and under other slope and soil conditions in Australia.

Acknowledgements – Zahvala The authors thank the following people and institutions for their support in carrying out this research project: Hancock Victoria Plantations, Victoria, Australia ANC Forestry, Victoria, Australia

6. References – Literatura Acuna, M., Kellogg, L., 2008: Evaluation of alternative cut-to-length harvesting technology for native forest thinning in Australia. International Journal of Forest Engineering 20(2): 19–27.

Amishev, D., Evanson, T., 2010: Innovative methods for steep terrain harvesting. In: FORMEC 2010 conference: »Forest Engineering: Meeting the needs of the society and the environment«. July 11–14, Padova, Italy.

Applied Computer Services Inc., 2010: Timer pro. Professional version. Englewood, CO. USA. ATLAS Technology, 2010: ATLAS cruiser. Rotorua, New Zealand. ().

Bergström, D., Bergsten, U., Nordfjell T., 2010: Comparison of boom-corridor thinning and thinning from below harvesting methods in young dense Scots pine stands.

Silva Fennica 44(4): 669–679.

Carson, B., Mann, C. N., Schiess, P., 1985: An Evaluation of Cable Yarding Bunched Trees on Steep Slopes. In: Proceedings of the Council on Forest Engineering: Forest Operations in politically and environmentally sensitive areas.

August 18–22, Tahoe City, California, USA: 95–102.

Evanson, T., 2008: Valmet 445/Satco 630. FFR Draft Report, Unpublished. Future Forests Research Ltd., Rotorua, New Zealand.

Evanson, T., 2010: Valmet 445 EXL Self-levelling feller-buncher. FFR Report 3(8). Forest Future Research Ltd., Rotorua, New Zealand.

Evanson, T., Amishev, D., 2010: Productivity impacts of bunching for yarder extraction. In: FORMEC 2010 conference: »Forest Engineering: Meeting the needs of the society and the environment«. July 11–14, Padova, Italy.

Heinimann, H. R., Visser, R. J., Stampfer, K., 1998: Harvester-cable yarder system evaluation on slopes – a Central European study in thinning operations. In: Proceedings of the Council of Forest Engineering Conference: »Harvesting logistics: from woods to markets«, Portland, OR, 20–23 July 1998, 41–46 p.

Kellogg, L. D., 1976: A case study of bunching and Swinging. A Thinning System for Young Forests. M.F. Paper, Oregon State University, Corvallis, USA. 88 p.

Kirk, R. J., Kellogg, L. D., 1990: Mechanized Felling on a Cable Yarding Operation. In: Proceedings of the 13th Annual Meeting of the Council on Forest Engineering: Managing Forestry Operations in a Changing Environment. August 12–16, 1990, Outer Banks, North Carolina, USA: 168–174.

MacDonald, A. J., 1999: Harvesting systems and equipment in British Columbia. FERIC Handbook, ISSN 0701–8355, No. HB-12. 197 p.

Murphy, G. E., 2003: Mechanisation and value recovery: worldwide experiences. In: Proceedings of the Wood for Africa 2002 Conference, July 2002, Pietermaritzburg, South Africa.

Nurminen, T., Korpunen, H., Uusitalo, J., 2006: Time consumption analysis of the mechanized cut-to-length harvesting system. Silva Fennica 40(2): 335–363.

Pan, F., Han, H. S., Johnson, L., Elliot, W., 2008: Production and cost of harvesting, processing, and transporting small-diameter (£ 5 inches) trees for energy. Forest Products Journal 58(5):47–53.

Peterson, J. T., 1987: Effect of falling techniques on grapple yarding second-growth timber. FERIC Technical Note TN-107.

Spinelli, R., Hartsough, B., 2000: Trials with a self-leveling CTL harvester in a naturally regenerated mixed-conifer stand. Rivista Di Ingegneria Agraria 2: 82–88.

Stampfer, K., Steinmuller, T., 2001: A New Approach To Derive A Productivity Model for the Harvester »Valmet 911 Snake«. In: Proceedings of the International Mountain Logging and 11th Pacific Northwest Skyline Symposium, 254–262. Seattle, WA, 10–12 December.

Visser, R., 2008: Is there a slope limit to mechanised felling on steep terrain? In NZ logger P14. Allied Publications Ltd. New Zealand.

Visser, R., Stampfer, K., 1998. Cable Extraction of Harvester-Felled Thinnings: An Austrian Case Study. Journal of Forest Engineering 9(1): 39–46.

Sa`etak

Utjecaj sakupljanja stabala feler-ban~erom na strmom terenu na u~inkovito izno{enje drva `i~arom Istra`ivanje rada sje~noga vozila na strmim terenima, njegove mogu}nosti i pobolj{anja u proizvodnosti u pridobivanju drva u`etnim sustavima na strmim terenima provedeno je u saveznoj dr`avi Viktorija u Australiji.

Obavljena je strojna ~ista sje~a i izno{enje drva u 33-godi{njoj planta`i bora (Pinus radiata D. Don) pomo}u

530

Croat. j. for. eng. 32(2011)2

Bunching with a Self-levelling Feller-Buncher on Steep Terrain for Efficient Yarder Extraction (521–531)

M. Acuna et al.

sje~noga vozila Valmet 445 EXL i {umske `i~are Madill 124. Strojna je sje~a okosnica ovoga istra`ivanja, iako treba napomenuti da je zbog terenskih prilika na pojedinim dijelovima sastojine obavljena i ru~no-strojna sje~a motornom pilom lan~anicom. Razlika u proizvodnosti rada {umske `i~are promatrana je usporedbom prethodno sakupljenih i oborenih stabala sje~nim vozilom i onih posje~enih motornom pilom i razasutih po {umskom bespu}u.

Rezultati ovoga istra`ivanja pokazuju da je u povoljnim sastojinskim uvjetima (relativno mala veli~ina stabala – prosje~noga obujma 0,8 m3, te na ocjeditim sedimentnim tlima) omogu}ena dobra kretnost vozila na nagibima od 36 do 47 % s visokom proizvodnosti. Na temelju regresijskoga modela izra~unata je proizvodnost sje~noga vozila od 138 m3/h za prosje~ni obujam stabla od 0,8 m3.

Na proizvodnost {umske `i~are uvelike je utjecalo da li su stabla bila prethodno sakupljena ili su bila razasuta po {umskom bespu}u nakon sje~e. Sakupljanje oborenih stabala omogu}ilo je skra}ivanje turnusa rada {umske `i~are jer je spu{tanje hvatala i prihvat tereta bilo kra}e za 11 %. Lako vidljive grupe (oborenih pa sakupljenih) stabala omogu}ile su operateru {umske `i~are lak{e i to~nije usmjeravanje hvatala, pa je i vrijeme prihvata tereta bilo kra}e. Iako je sje~nim vozilom Valmet 445 EXL mogu}e skupiti u prosjeku 4,3 stabla po grupi (slo`aju), samo su 2,3 stabla iznesena na pomo}no stovari{te u jednom turnusu. To je uglavnom zbog veli~ine hvatala same {umske `i~are i nastojanja radnika za smanjenjem gubitka vremena pri prihvatu tereta.

Unato~ ~injenici da bi sakupljanje oborenih stabala na kraju moglo utjecati na proizvodnost samoga sje~noga vozila te iako veli~ine sakupljenih stabala nisu bile prikladne za hvatalo `i~are, zamjetljiva razlika izme|u sakupljenih i razasutih stabala jest ve}i broj iznesenih komada po radnom turnusu (2,3 komada/turnusu kod skupljenih stabala, nasuprot 1,5 komada/turnusu kod stabala razasutih po {umskom bespu}u), s prosje~nim obujmom iznesenoga drva od 1,9 m3/turnusu za skupljena stabla, u odnosu na 1,3 m3/turnusu za razasuta stabla. U prosjeku je za udaljenosti izno{enja drva od 150 do 240 metara sakupljanje oborenih stabala pove}alo proizvodnost {umske `i~are s hvatalom za 25 %.

Unato~ ograni~enjima ovoga istra`ivanja (istra`ivanje je provedeno u jednoj sastojini, model proizvodnosti sje~noga vozila Valmet 445 EXL temelji se isklju~ivo na prosje~noj veli~ini oborenih stabala, nije izra|ena studija rada i vremena od dviju sastavnica, tj. posebno sje~a stabala pa onda sakupljanje oborenih stabala, ovi rezultati pokazuju mogu}nosti kori{tenja (sje~a te sakupljanje) gusjeni~nih sje~nih vozila uz izno{enje drva {umskim `i~arama na strmim terenima u Australiji. Potrebna su daljnja istra`ivanja na razli~itima terenima, nagibima te tipovima tla, ali i pri obaranju stabala ve}ih dimenzija.

Klju~ne rije~i: gusjeni~no sje~no vozilo, {umska `i~ara, sakupljanje stabala, proizvodnost

Authors’ address – Adresa autorâ: Mauricio Acuna, PhD. Harvesting and operations program e-mail: mauricio.acuna@utas.edu.au University of Tasmania CRC for Forestry Private bag 12, Hobart, TAS AUSTRALIA Mr. Justin Skinnell, Undergraduate student e-mail: skinnelj@onid.orst.edu Oregon State University 204 Peavy Hall, Corvallis, OR 97331 USA Mr. Tony Evanson, Senior Researcher e-mail: Tony.Evanson@scionresearch.com SCION (New Zealand Forest Research Institute) 49 Sala Street, Rotorua NEW ZEALAND

Received (Primljeno): July 26, 2011 Accepted (Prihva}eno): September 5, 2011 Croat. j. for. eng. 32(2011)2

Mr. Rick Mitchell, Research Technician e-mail: rick.mitchell@wapres.com.au CRC for Forestry – WAPRES Level 2, 53 Victoria St, Bunbury, WA AUSTRALIA

531

Original scientific paper – Izvorni znanstveni rad

Efficiency of John Deere 1470D ECOIII Harvester in Poplar Plantations Milorad Danilovi}, Ivan Toma{evi}, Dragan Ga~i} Abstract – Nacrtak This article presents the results of researching John Deere 1470 D ECO III harvester in clear cuttings of Populus×euramericana 'I-214' poplar trees in lowland regions. Four different methods of the harvester work were analyzed from the aspect of its movement direction and the number of trees that were cut from one standing point. Apart from that, the effect of forks on the harvester productivity was analyzed. A study of work and time was carried out in the research. Duration of the working operations was measured on the chronometer, by the time flowing method. The method of work had the most significant effect on the moving and the positioning time of the harvester, while the stem forking greatly affected the stem processing time. The differences in the time length of various felling operation phases affected the productivity that the harvester achieved in different methods of work. The average productivity ranges from 30.3 to 34.7 m3/h, depending on the method of work. The harvester achieves the highest productivity when it moves backwards between two rows and cuts a stem in the row to the right and then in the row to the left, observed from the moving direction, while on its way back, it drives forward between the rows and cuts a stem from the row on the left first and then from the row on the right, observed from the moving direction. When this method is applied, the harvester productivity amounts to 34.7 m3/h. Its productivity is significantly affected by stem dimensions, i.e. the increasing volume of timber wood decreases the time needed for the processing of one unit product.

Keywords: harvester, poplar plantations, tree characteristics, method of work, productivity, costs

1. Introduction – Uvod Poplar is a tree species that covers 1.9% of the total forest area in Serbia. It grows both in natural stands and on artificially established plantations in lowland areas.

According to their production goal, the most common plantations are long rotation plantations from 20 to 25 years of age, intended for the production of saw logs for mechanical wood processing. Plantations with combined production goals, which are thinned between the age of 8 and 12, are significantly less frequent.

The processes of felling and processing assortments in long rotation plantations are characterized by larger trees, thicker branches and bigger forks in comparison with the processes of felling and processing logs in short rotation plantations. One of the aims of this research is to determine the effects of these characteristics of poplar trees from long rotation plantations on the harvester productivity.

Croat. j. for. eng. 32(2011)2

Up until 2008 chain saws were solely used in the felling and processing operations in mature poplar plantations in Serbia, with the application of the group selection system. Since 2008, apart from chain saws, John Deere 1470 D ECO III harvester has been used for the purposes of felling and processing logs in mature poplar plantations.

The harvester has been introduced in the regular cutting operations of poplar trees in order to achieve better mechanization of felling and processing operations, to reduce the number of workers with occupational diseases, to overcome labor shortage and to achieve better production outputs.

Daily production outputs of the workers who are employed in felling and processing logs with a chainsaw in poplar plantations mature for felling are high, i.e. unit costs of felling and processing are considerably lower than the costs of cutting and processing the logs in the final felling of hard broadleaves.

533

M. Danilovi} et al.

Efficiency of John Deere 1470D ECOIII Harvester in Poplar Plantations (533–548)

Processing logs into long cellulose wood and cordwood in poplar plantations mature for felling on the territory of Serbia has become a regular practice, and it has a considerable positive effect on the achieved outputs.

In such working conditions, the use of the harvester is profitable only when high outputs are produced.

The efficiency of the harvester in felling and processing operations depends on a great number of factors, the most important being terrain conditions and stem characteristics.

Site slope significantly affects the productivity of the harvester and it is often a limiting factor in its use. With the development of a driving system that enables harvesters to operate on steep slopes, the effect of this factor is considerably smaller, which has been proved by the results of recent researches (Stampfer 1999, Stampfer and Steinmüeller 2001).

The work of the harvester in different working conditions has been studied by a great number of authors (Brunberg 1991, Tufs and Brinker 1993, Kellogg and Bettinger 1994, Tufts 1997, Stampfer 1999, Glöde 1999, Stampfer and Steinmüeller 2001, Krpan et al.

2001, Krpan and Por{insky 2002, Eliasson 1998, Wester and Eliasson 2003, Kärhä et al. 2004, Nurminen et al. 2006). The results of these researches show that the harvester productivity is highly variable and it depends on a great number of factors.

However, all the researchers stress the stem volume as the most significant factor. The productivity of the harvester increases with the growth of stem volume.

The productivity of the harvester also varies with the felling intensity or the number of stems intended for clear-cutting (Eliasson et al. 1999).

Apart from that, skilled and experienced operators can increase the productivity of the harvester to a great extent (Sirén 1998).

The use of the harvester in regular poplar fellings is seemingly a simple solution, since here we deal with clear-cuttings on flat terrains, a homogeneous stand from the aspect of tree dimensions and arrangement and lower wood hardness. However, even in such conditions there are a number of factors that affect the productivity of the machine.

Poplar plantations are characterized by low bearing capacity of the terrain and large stems with distinctive external characteristics (excessive branching, sweep, forks and others).

The effect of stem characteristics on the harvester productivity is more significant in cutting and processing broadleaved tree species due to a higher rate of branching in the crown region, sweep of the lower

534

or the most valuable part of the stem and the incidence of forking.

Sweep of the lower or the most valuable part of the stem is particularly characteristic of the plantations that are exposed to flood waters. This part of the stem is the most valuable and according to the previous research, the assortments that are in qualitative marking for cross-cutting obtained from this part of the stem are logs for sliced and peeled veneer (Danilovi} 2006).

Damage arising from processing curved tree stems leads to more intensive drying of peripheral parts of the assortments, which has a negative effect on the quality of peeling in the mechanical wood processing.

Trees that are regularly felled often form forks, which can affect the time required for processing a unit product. The frequency of forking, as well as the height at which the forks appear, depend on the measures taken to correct the tops of young plants or prune thicker branches off young poplar standing trees.

In the early years of plantation development, it is necessary to form the top of the trees by removing thicker branches in order to avoid subsequent stem forking. Occurrence of forks cannot be completely eliminated, but it can be minimized by proper and timely pruning of thicker lateral branches off the young standing poplar trees.

Moreover, Populus euramericana trees have an excess of moisture in the central part of the stem, which can greatly contribute to the occurrence of cracks in the lower, most valuable part of the stem during the felling operations, but even more during the vegetation. Cracks are most frequently developed when cutting large leaning trees.

The productivity of the harvester is also affected by the method of work used in the process of felling and processing wood assortments. Great progress has been made in the development of the methods and systems of work (Makkonen 1991, Glöde and Sikström 2001, Suadicani and Fjeld 2001, Andersson and Eliasson 2004, Nurminen et al. 2006, Jirou{ek et al. 2007).

The productivity of the harvester-forwarder joint cutting system has been more frequently dealt with in the studies of American researchers than in the studies of European researchers who have mostly studied these systems separately (Nurminen et al.

2006).

There are some other factors influencing the harvester productivity, amongst which is the manner in which the harvester moves through the plantation Croat. j. for. eng. 32(2011)2

Efficiency of John Deere 1470D ECOIII Harvester in Poplar Plantations (533–548)

during the harvest, as well as the number of trees cut from one standing point.

The aim of this article is to study the productivity of John Deere 1470D ECO III harvester when four different methods of work are used in the course of cutting poplar trees in the plantations with long rotation periods and to assess the effects of forks on the harvester productivity.

The initial hypotheses are: Þ There are significant differences between the investigated methods and Þ Forks have a considerable effect on the time needed for stem processing.

2. Material and Methods – Materijal i metode The research was conducted in mature plantations of Populus×euramericana 'I-214' on the territory of Ravni Srem in Serbia. They cover an area of 6.89 ha.

Harvesting was carried out by John Deere 1470 D ECO III harvester equipped with a harvester head H 480 (Fig. 1). The harvester operator had two years of previous experience in operating the harvester, and many years of experience in operating a forwarder.

Recording was carried out in winter conditions in the period from December 2009 to February 2010.

The weather conditions were favorable, with no precipitation. Before felling and processing wood assortments, larger undergrowth had been removed in order to facilitate the movement of the harvester through the cutting site. Main characteristics of the poplar plantation and terrain are shown in Table 1.

While felling and processing logs, the harvester moved along the longer side of the felling site, which

Fig. 1 John Deere 1470D ECOIII Harvester Slika 1. Harvester John Deere 1470D ECOIII Croat. j. for. eng. 32(2011)2

M. Danilovi} et al.

Table 1 Main characteristics of the plantation Tablica 1. Osnovne zna~ajke planta`e Age of the plantation, years Dob planta`e, godina Planting spacing, m Razmak sadnje, m No.of trees per hectare at the time of establishment Broj stabala po hektaru u vrijeme osnivanja nasada No.of trees per hectare at the end of rotation Broj stabala po hektaru na kraju ophodnje Medium DBH, cm Prosje~ni prsni promjer, cm Mean tree height, m Prosje~na visina stabla, m Total volume of assortments, m3 Ukupni obujam sortimenata, m3 Average net volume of assortments per stem m3/stem Prosje~ni obujam sortimenata po stablu, m3/stablu Slope, ° Nagib terena, ° Cutting method Na~in sje~e Soil type Tip tla

25 6×6 278 192 40.4 31.7 2056.9 1.55 » 0° clear cutting ~ista sje~a mollic gleysol moli~ni glej

is 750 m long. This was the only solution because of the direction of stem inclination, caused by prevailing winds characteristic in this part of Serbia.

Four different methods of harvester work in the operations of felling and processing logs were recorded on one part of the compartment area (Fig. 2).

The investigated methods differed depending on the harvester driving direction, its positioning-to-cut and the number of stems cut from one standing point. The following methods of the harvester work were investigated: Þ Method 1 – The harvester moves in reverse and cuts trees from the row to the left, observed from the moving direction. It cuts and processes one stem from one standing point. When it finishes cutting in one row, it moves to the beginning of the next row and starts harvesting in the same manner. The general direction of stem inclination is opposite to the direction of the harvester movement. Þ Method 2 – The harvester moves in reverse and cuts trees from the row to the left, observed from the moving direction. It cuts and processes assortments from one stem per standing point. The direction of stem inclination is opposite to the direction of the harvester movement. It starts cutting trees in the next row from the opposite side of the cutting site, by driving forward. The gen-

535

M. Danilovi} et al.

Efficiency of John Deere 1470D ECOIII Harvester in Poplar Plantations (533–548)

eral direction of stem inclination is the same as the direction of the harvester movement. From one standing point, it cuts and processes one tree on its right hand side observed from the moving direction of the harvester. Felling and processing operations in the following rows are conducted in the same manner as in the previous two rows. Þ Method 3 – The harvester moves backwards between two rows, cuts a stem in the row to the right, and then in the row to the left, observed

from the moving direction. It cuts and processes two stems from one standing point. The general direction of stem inclination is opposite to the direction of harvester movement. It starts harvesting stems in the next rows without changing the side and in the same manner as in the previous two rows. Þ Method 4 – The harvester moves backwards between two rows, cuts a stem in the row to the right, and then in the row to the left, observed

Fig. 2 Methods of harvester work Slika 2. Metode rada harvestera 536

Croat. j. for. eng. 32(2011)2

Efficiency of John Deere 1470D ECOIII Harvester in Poplar Plantations (533–548)

M. Danilovi} et al.

Table 2 The values obtained for the measured elements Tablica 2. Opis vrijednosti mjerenih elemenata Measured elements – Mjereni elementi Number of analyzed stems Broj analiziranih stabala Mean diameter at breast height, cm Prosje~ni promjer na prsnoj visini, cm Mean diameter of the stems without forks, cm Prosje~ni promjer stabala bez ra{alja, cm Mean diameter of forked stems, cm Prosje~ni promjer ra{ljavih stabala, cm Volume of assortments processed from the stems with the mean cut diameter, m3 Obujam sortimenata izra|enih od prosje~noga sje~noga stabla, m3 Volume of assortments processed from the mean stems without forks, m3 Obujam sortimenata izra|enih od srednjega stabla bez ra{alja, m3 Volume of assortments processed from the mean forked stems, m3 Obujam sortimenata izra|enih od srednjega ra{ljavoga stabla, m3 The average number of assortments per stem Prosje~an broj sortimenata izra|enih od stabla Percentage of forked stems, % Udjel ra{ljavih stabala, %

from the moving direction. It cuts two stems from one standing point. The general direction of stem inclination is opposite to the direction of the harvester movement. It starts cutting trees in the next two rows from the opposite side of the felling site, by driving forward, in the direction of stem inclination. First, it cuts a stem in the row to the left, and then in the row to the right, observed from the moving direction. It cuts two stems from one standing point.

The average values of the measured elements for the analyzed stems, as well as the percentage of forked stems for the investigated methods, are presented in Table 2.

The research is based on the study of work and time. The time length of the working operations was measured on the chronometer, by the flowing method, with one second accuracy.

The following working operations were recorded: Þ Moving: it starts when the harvester wheels start moving from one standing point and ends when they stop in the next standing point Þ Positioning-to cut: it starts when the harvester wheels stop moving and ends when the harvester head grips the stem Þ Felling: it starts when the chain starts moving and ends when the tree falls onto the ground Þ Processing: it starts when the tree stem starts moving through the harvester head and ends when the harvester wheels start moving.

Croat. j. for. eng. 32(2011)2

Method 1 Metoda 1

Method 2 Metoda 2

Method 3 Metoda 3

Method 4 Metoda 4

350

310

356

311

40.5

40.3

41.0

40.2

39.7

39.5

40.2

38.7

46.0

44.0

47.3

47.7

1.54

1.53

1.58

1.55

1.49

1.49

1.52

1.46

1.93

1.78

2.03

1.90

4.1

4.3

4.4

4.4

12.3

14.4

13.5

13.8

All the hang-ups in the felling and processing operations of the harvester work were recorded.

Apart from that, records were made of all trees that deviated from the desired felling direction, by stating the rate and cause of deviation.

Furthermore, a separate record was made of all forked stems, i.e. the stems that could not be processed without previous rotating the harvester head when cutting the limbs of the forks.

When the harvester moved backwards, wood assortment were processed on the spot that was on its right side, observed from the moving direction, at the distance of 2.5 m, and when it moved frontally it was on the same spot to the left.

The logs were positioned in such a way that they ensured free movement of the forwarder through the cutting site and easy transport of logs.

The most significant factors in processing logs into assortments of certain quality were minimal dimensions prescribed by the national standards for the quality of roundwood (SRPS) and the stem characteristics (knots and sweep). It has been research-proved that these two factors are crucial in the qualitative marking stems for cross-cutting (Danilovi} 2006). The logs were cross-cut through knots since in this way they can be regarded as an allowance and eliminated as a wood defect. At the same time, cracking of the face of the processed assortments was avoided.

537

M. Danilovi} et al.

Efficiency of John Deere 1470D ECOIII Harvester in Poplar Plantations (533â&#x20AC;&#x201C;548)

Before the harvester started the work, the stems had been marked for cross-cutting and stem diameter at breast height had been measured. Afterwards, the diameter at stump height and the diameter at the thinner end of the log were measured as well. They were measured with the accuracy of up to a millimeter. The volume of the processed assortments was calculated by Huber formula with an accuracy of two decimal places.

The height of the trees was determined by using the height curve based on the sample of 234 trees.

Two criteria were used in the classification of stems for the purposes of the analyses within each investigated method. They are: diameter at breast height and forks.

The significance and the effects of these factors on the differences between the harvester working elements in cutting and processing wood assortments were determined by two-factor analysis of variance in the statistical program Statgraphics Plus. Regression analysis was applied to make a relationship between the variable.

3. Results â&#x20AC;&#x201C; Rezultati In the course of the research, a total 1327 stems were felled and processed. The mean cut diameter at stump height is 46.3 cm and at breast height 40.4 cm.

Out of the total number of cut and processed stems, forked stems account for 13.6%. Forked stems yield averagely 6.1 assortments, while the stems without forks yield 4.3 assortments with the average

Fig. 3 Structure of the effective harvester working time Slika 3. Struktura efektivnoga vremena rada harvestera 538

length of 5.1 m. The diameter at the thinner end of the log averages 20.1 cm.

Fig. 3 shows the structure of the effective working time for each method studied.

When the first working method was used, the moving time made 15.7% of the harvester total effective working time and amounted to 0.36 min/stem or 0.23 min/m3. Out of the total amount of moving time, 60.2% was spent on moving from one stem to another, while the rest of the time was spent on moving from one row to another, or on the return of the harvester from the end of the row in which it had performed the cutting operations to the beginning of the next row. The average driving speed of the harvester, when it moved from one row to another, was 59.9 m/min and 41.1 m/min when it moved from one tree to another.

When the second method was used, the moving time made 9.9% of the harvester total effective working time and amounted to 0.22 min/stem or 0.16 min/m3. With the the third method, the harvester spent 8.5% of its total effective working time on moving, i.e. 0.18 min/stem or 0.11 min/m3. Out of the total amount of moving time, 64.7% was spent on moving from one tree to another, while the rest of the time was spent on moving from one row to another.

The percentage of transition time was the lowest when the fourth method was used and it accounted for 5.8% of the harvester total effective working time and amounted to 0.12 min/stem or 0.08 min/m3.

The results of the two-factor analysis of variance show that there are statistically significant differences in the average moving time values between the investigated methods (F-203.1, p-0.000).

The size of the stem and the degree of its forking do not have a significant effect on the differences in the transition time (p>0.05).

There are statistically significant differences in the average moving time between the first and the other three methods of work (Table 3). The effects of other factors such as undergrowth, soil moisture or topography do not differ between the studied areas.

The average positioning time ranged from 12.1 to 15.9% of the effective working time and was from 0.26 to 0.37 min/stem or from 0.25 to 0.18 min/m3, depending on the method used. It was the longest when the first method was used (0.37 min/stem or 0.25 min/m3) and the shortest when the fourth method was used (0.26 min/stem or 0.17 min/m3).

The respective values for the second method were 0.34 min/stem or 0.22 min/m3 and 0.28 min/stem or 0.18 min/m3 for the third (Table 5).

The investigated methods of work showed statistically significant differences (F-11.65, p-0.000) in the Croat. j. for. eng. 32(2011)2

Efficiency of John Deere 1470D ECOIII Harvester in Poplar Plantations (533–548)

M. Danilovi} et al.

Table 3 Average time of working elements for different working methods (Duncan multiple range test) Tablica 3. Prosje~ni utro{ci vremena radnih sastavnica pri razli~itim metodama rada (Duncan multiple range test) Working elements Radne sastavnice

Method 1 Method 2 Method 3 Method 4 Covariate Metoda 1 Metoda 2 Metoda 3 Metoda 4 Nezavisna varijabla Average time consumption, min/stem – Prosje~ni utro{ci vremena, min/stablo Moving – Premje{tanje 0.36 a 0.22 b 0.18 b 0.12 b – Positioning-to-cut – Zauzimanje polo`aja 0.37 a 0.34 a 0.28 b 0. 26 b – Felling – Ru{enje 0.121 a 0.124 a 0.136 b 0.138 b – Processing – Izradba 1.48 a 1.53 a 1.57 a 1.58 a Stem volume – Obujam stabla, p<0.05 a, b – time consumptions of working elements without differences between working methods – utro{ci vremena radnih sastavnica bez zna~ajnih razlika me|u metodama rada

average positioning time values. Stem forking contributed greatly to the differences in the harvester positioning time. This means that the harvester operator spent a considerably greater amount of time on positioning when felling forked stems.

The average positioning time did not differ between the first and the second, as well as between the third and fourth method of harvester work (Table 3), which proved that the driving direction of the harvester in the cutting and processing operations did not affect the positioning time, while the number of trees cut from one standing point had a significant effect on the positioning time and differentiated the investigated methods into two homogeneous groups.

Felling time ranged from 5.2 to 6.7% of the effective working time, depending on the method of work, i.e. from 0.12 min/stem or 0.09 min/m3 in the first method to 14 min/stem or 0.1 min/m3 in the fourth method. The investigated methods of work showed statistically significant differences (F-8.41, p-0.000) in the average felling time values. The average felling time did not differ between the first and second, as well as between the third and fourth method of harvester work (Table 3). Furthermore, forking and stem volume contributed to the differences in the cutting and felling time (p<0.05).

Cutting and felling time increases with the growth of stem volume due to larger crowns or less space available for free stem felling. There is a linear relationship between the variables (Fig. 4).

Minimum stem processing time was 0.22 and maximum 5.91 min/stem. Processing time took up the greatest percentage of the total effective time and amounted to 70% (Fig. 3) The average processing time did not differ significantly between the investigated methods of work (F-1.02, p-0.381). In the average it ranged from 0.91 to 1.0 min/m3, depending on the investigated method of work.

Croat. j. for. eng. 32(2011)2

Stem forking has a significant effect on the stem processing unit time (F-173.9, p-0.000). The average time needed for processing a stem without forks amounted to 0.89 min/m3 and 1.35 min/m3 for stems with formed forks.

Processing time increases with the growth of stem volume. The relationship between the stem volume and stem processing time is expressed by regression equations (Table 4).

The model of the function that was selected after testing several different models shows a moderately strong relationship between the variables. The relationship between the stem size and processing time is moderately strong. In other words, the relationship between these two variables accounts for less than half of the variations (Table 4).

The relationship between stem volume and the time spent on cutting and felling is linear (Tab. 3).

Fig. 4 Dependence of felling time on stem volume Slika 4. Ovisnost vremena ru{enja o obujmu stabla 539

M. Danilovi} et al.

Efficiency of John Deere 1470D ECOIII Harvester in Poplar Plantations (533–548)

Table 4. Functions and function parameters Tablica 4. Funkcije i parametri funkcija Working elements Radne sastavnice Felling – Ru{enje

Processing – Izradba

Function Funkcija Method 1 and 2 – Metoda 1 i 2 Method 3 and 4 – Metoda 3 i 4

 min  tf = a + b × V    tree 

Stems without forks – Stabla bez ra{alja Forked stems – Ra{ljava stabla All stems – Sva stabla

Fig. 5 Dependence of processing time on stem volume Slika 5. Ovisnost vremena izradbe drva o obujmu stabla Non-productive time accounted for 34.91% of the total time of harvester work. Out of the total non-productive time, 18.4% was spent on justified and unjustified interruptions.

Justified interruptions included breaks, consultations, harvester maintenance, breakfast, odd jobs and technical failures. Out of the total non-productive time, 24.2% was taken by preliminary work time, while the average time used for breaks amounted to 19.6% of the total interruption time length. Lunch

 min  t pr = a × V b    tree 

Parameters – Parametri a b 0.083 0.023

0.061

0.076

0.035

0.101

0.764 1.164 0.781

1.147 1.129 1.259

0.415 0.492 0.449

R2

breaks amounted to 18.4%. The interruptions which occurred when the chain fell off when felling larger stems with the diameter at stump height above 55 cm averaged 16.2% of the total interruption time.

Other interruptions included harvester maintenance, consultations and odd jobs.

Extra working time of the harvester, which is derived by excluding all unnecessary and unjustified interruptions from the non-productive working time, amounts to 28.5%. Felling and processing time per unit product varies with the method of work and the differences are primarily caused by the positioning and moving time (Table 5).

The timber volume of a stem with a mean cut diameter is 1.55 m3. The effective time needed for processing a stem with a mean cut diameter is from 1.35 to 1.54 min/m3, depending on the working method.

Based on the results of the analysis of variance, it can be concluded that there are statistically significant differences in this time between the used methods of harvester work (F-2.82, p-0.038), as well as between the stems with and without forks (F-98.2, p-0.000).

There are also statistically significant differences between the fourth and the other three working methods.

The total effective time spent on stem felling and processing can be calculated by the following mathematical equation (Eq. 1): 4  4  min  ∑ a ⋅ tm + ∑ b ⋅ tp + t + tpr   3  j j f   i i j =1  j =1  m 

te =

1 V

aj =

0, i ≠ j 0, i ≠ j , bj = 1, i = j 1, i = j

(1)

Table 5 Harvester moving and positioning time depending on the method of work Tablica 5. Vrijeme premje{tanja i zauzimanja polo`aja harvestera za primijenjene metode rada Working elements Radne sastavnice Moving tm, min/stem – Premje{tanje tm, min/stablo Positioning-to-cut tp, min/stem – Zauzimanje polo`aja tp, min/stablo

540

Method 1 Metoda 1 0.36 0.37

Method 2 Metoda 2 0.22 0.34

Method 3 Metoda 3 0.18 0.28

Method 4 Metoda 4 0.12 0.26

Croat. j. for. eng. 32(2011)2

Efficiency of John Deere 1470D ECOIII Harvester in Poplar Plantations (533–548)

M. Danilovi} et al.

Fig. 6 Harvester productivity vs. work method Slika 6. Ovisnost proizvodnosti harvestera o primijenjenoj metodi rada

Fig. 7 Influence of stem forkness on harvester productivity Slika 7. Utjecaj ra{ljavosti stabala na proizvodnost harvestera

where: tm – moving time (Table 5) tp – positioning time (Table 5) – felling time (function in Table 4) tf tpr – processing time (function in Table 4) V – stem volume The average productivity ranges from 39.0 to 44.4 m3/h, depending on the method of work. The time needed for processing a stem with a mean cut diameter depending on the method of work, provided that the extra time is included, ranges from 1.73 to 1.98 min/m3, which means that the harvester achieves the average productivity of 30.3 to 34.7 m3/h.

The highest productivity is achieved when the harvester moves between two rows and cuts two stems from one standing point (Method 4), and the lowest when it moves backward and cuts stems from one row, i.e. one stem from one standing point (Method

1). The average productivity of the harvester, when the second method is used, amounts to 32.3 m3/h, and 34.0 m3/h when the third method is used.

The productivity of the harvester increases with the growth of timber volume. The relationship between the productivity and the stem diameter is presented in Fig. 6 The productivity of the harvester in felling and processing forked stems is lower than the productivity achieved in felling stems without forks (Fig. 7).

The relationship between the stem volume and productivity is expressed by regression equations (Table 6).

Direct costs of John Deere 1470D ECOIII harvester work are calculated on the basis of 1600 operating hours and they amount to 121.3 EUR/h. Cost calculation for the harvester work was made on the basis of input data (Table 7).

Table 6 Statistic elements of regression models Tablica 6. Statisti~ki elementi regresijskih modela Method – Metode Method 1 – Metoda 1 Method 2 – Metoda 2 Method 3 – Metoda 3 Method 4 – Metoda 4 Stems without fork – Stabla bez ra{alja Forked stems – Ra{ljava stabla All stems – Sva stabla

Croat. j. for. eng. 32(2011)2

Function – Funkcija

P=

1  m3  b h a+   v

a 0.0244 0.0242 0.0238 0.0236 0.022 0.030 0.024

b 0.013 0.01 0.008 0.005 0.009 0.012 0.008

ta 7.1 9.4 12.8 16.1 23.9 10.5 28.4

tb 2.8 2.8 2.9 2.5 7.45 2.6 4.8

R2 0.221 0.274 0.263 0.213 0.481 0.340 0.311

541

M. Danilovi} et al.

Efficiency of John Deere 1470D ECOIII Harvester in Poplar Plantations (533â&#x20AC;&#x201C;548)

Table 7 Machine cost data Tablica 7. Podaci o tro{kovima stroja Purchase price, EUR Nabavna cijena, EUR Engine output power, kW Snaga motora, kW Machine life, years Radni vijek stroja, godine Salvage value, % of purchase price Preostala vrijednost, % od nabavne cijene Machine utilization rate,% Stupanj iskoristivosti stroja, % Repair and maintenance cost, % of depreciation Tro{kovi popravaka i odr`avanja, % od amortizacije Interest rate, % of average yearly investment Kamatna stopa, % prosje~ne vrijednosti investicije Insurance and tax rate, % of average yearly investment Osiguranje i takse, % prosje~ne vrijednosti investicije Fuel consumption rate, L/h Utro{ak goriva, L/h Fuel cost, EUR/L Cijena goriva, EUR/L Oil and lubrication consumption rate, L/h Utro{ak ulja i maziva, L/h Oil and lubrication cost, EUR/L Cijena ulja i maziva, EUR/L Operatorâ&#x20AC;&#x2122;s wage, EUR/h Zarada voza~a, EUR/h

442900 180 5 10 80 50 5.5 5 14.8 1.1 0.62 2.5 5.7

Fig. 9 Influence of stem forkness on unit costs Slika 9. Utjecaj ra{ljavosti stabala na jedini~ne tro{kove The Unit costs of work are calculated on the basis of direct costs of work and the productivity, achieved by the harvester in different working methods of felling and popalr plantation, vary with the method of work. The costs are the lowest when the Method 4 is used and amount to 3.5 EUR/m3, and the highest with the the first method, amounting to 4.0 EUR/m3.

Costs per unit product decrease with the growth of timber volume, due to the increase in the productivity achieved by the harvester (Fig. 8).

Furthermore, the average unit costs of felling forked stems or the stems whose processing requires rotating of the harvester head are higher than the average costs of processing a unit product from stems with the same dimensions but without forks.

On the basis of these results it can be concluded that Method 4 is economically most convenient for mature poplar plantations.

4. Discussion â&#x20AC;&#x201C; Rasprava

Fig. 8 Harvester unit costs depending on methods of work Slika 8. Jedini~ni tro{kovi harvestera u ovisnosti o primijenjenoj metodi rada 542

At the end of the production cycle, the productivity of John Deere 1470D ECOIII harvester, employed in the clear-cutting of poplar trees varied from 30.3 to 34.7 m3/h, depending on the method of work.

The achieved outputs are lower in comparison with the outputs achieved by the harvester in clear-cutting of mature conifer stands (e.g. spruce) (Eliasson et al. 1999) and higher than the outputs Croat. j. for. eng. 32(2011)2

Efficiency of John Deere 1470D ECOIII Harvester in Poplar Plantations (533–548)

achieved in hard broadleaved stands (beech) mature for felling.

Poplar trees have less dense and consequently less hard wood than hard broadleaves. It makes their wood easier for processing. Felling and processing stems with a lower rate of branching, less noticeable sweeps and without major forks can be carried out without any significant delays. The damage to the processeed assortments is minimal, especially during the vegetative resting period.

Large trees (above 55 cm in diameter at stump height) are not always possible to fell in the desired direction. The percentage of such trees was 4.1% of the total number of felled trees. Deviation from the desired felling direction was often greater than the distance between the trees in a row, which often caused damage to the adjacent standing trees, or the trees that had already been felled. This happened more often when felling and cutting forked trees because their individual inclination was significantly different from the general inclination direction of the trees in the stand. Minor deviations occurred when felling thinner stems. They often got stuck in the forks of adjacent standing trees and required more time for felling. Apart from that, it caused a significant breakage of standing trees.

When felling forked trees with larger diameters at stump height, it took more time to position the harvester in order not to threaten the operator safety.

In this case the distance between the harvester and the tree was approximately 2.5 m, and when felling and cutting thinner stems the distance was about 4 m.

It follows that when the harvester moves between two rows and cuts stems from both rows, the harvester operator has to perform an additional positioning before felling a tree, which means that it is not always possible to fell two trees from one standing point without moving the harvester.

Operating the harvester is easier when the first method of work is used because the operator can focus his attention on the trees in one row, which causes less mental fatigue in comparison with the working methods that imply frontal movement of the harvester between the rows. Operating the harvester does not involve only physical activities. The harvester operator responsibilities are to follow a series of variable conditions and to respond to a specific situation in the best possible manner (Nonaka et al. 1995). Furthermore, we are dealing with voluminous work, which must be qualitatively divided into several different phases if we want to achieve maximum financial effect.

The time needed to drive the harvester from one row to another, when the first method is used, averCroat. j. for. eng. 32(2011)2

M. Danilovi} et al.

ages 39.8% of the total moving time, which affects the total time needed for felling and processing a unit product, i.e. the achieved productivity of the harvester. However, this time should be seen as the time for a mental break of the harvester operators, who certainly appreciate it. However, we cannot assess the extent of its positive effect on the mental fatigue of the workers on the basis of this research. The moving time does not depend on the size of the trees, but on the planting density and the number of trees cut from one standing point (Klunder and Stokes 1994, Eliasson et al. 1999, Krpan and Por{insky 2002, Nurminen et al. 2006).

When the harvester moves backwards and cuts trees to the right observed from the moving direction, there is much more free space available for felling and the visibility is better because there is not a standing tree in front of the tree that is being felled.

In this case, the angle at which the harvester cuts the tree in relation to its moving axis usually ranges from 0 to 25° (Fig. 2). Since there are no standing trees in the area from 0 to 10°, it is easier to perform the felling operations. The situation is the same when the third method is used, except that the felling area on the right side observed from the moving direction is free for felling from 0 to 20°. When the second and the fourth methods are used, this space is occupied by standing trees that make it more difficult to fell trees in the desired direction, especially if they are forked.

Felling trees with large diameters and trees that deviate from the general felling direction increases the amount of time needed for felling. In this case the operator drives the harvester to a position in which he can use the harvester grip to push the tree in the desired felling direction. The safety of the operator is assessed for each individual operation. The direction of the felled tree affects the time needed for stem processing because of additional manipulation with the stems in order to position the assortments in such a way to allow free movement of the forwarder.

Processing time does not vary significantly with the working method of the harvester, but there are statistically significant differences in the average time needed to process stems with and without forks. Longer time is needed to process forked stems because the harvester head has to be rotated 180° to cut off the limbs of the forks.

The harvester operator has to rotate the harvester head, then to cut the stem just below the fork in order to achieve maximum wood assortment utilization, and finally he needs some time to lift the rest of the forked stem off the ground and cut it up.

Since the forks have a significant effect on the stem processing time (Fig. 6), in the early years of

543

M. Danilovi} et al.

Efficiency of John Deere 1470D ECOIII Harvester in Poplar Plantations (533â&#x20AC;&#x201C;548)

plantation development, it is necessary to remove thick branches that lead to the formation of large forks. This is especially true when it comes to clones of vigorous growth.

In comparison with mountainous regions, where both stem and terrain characteristics are of great importance, terrain characteristics are not regarded crucial for the use of harvester in lowland areas.

Terrain characteristics do not have a direct effect on the harvester working performances, because it can approach each tree and take the most suitable felling and processing position. The indirect effects of the terrain characteristics can be observed through the impact of flood waters on the occurrence of multiple sweep in the lower or the most valuable part of the stem, which subsequently affects the stem processing time on the one hand and the extent of damage caused by the harvester head on the other hand.

The damage was analyzed on the sample of 74 stems or 286 logs. The results show that cracks appeared on the face of the logs in 7.56% of all logs, while mechanical injuries in the form of dents accounted for 16.8%. The average depth of damage was 3.2 cm, the maximum 12 cm.

The first log had no mechanical injuries, the second had injuries in 19.8% of all cases, while the damage on the third and the fourth log occurred in 77.4% of all cases. Mechanical injuries that occurred in the upper part of the tree stem did not have a significant effect on the quality of the processed assortments since this part of the stem rarely yields best quality assortments.

Stem characteristics greatly affect the harvester productivity, but not to such extent to call for the application of the combined method, as is the case with the final fellings of hard broadleaves.

It follows that from the technical point of view, the use of harvester in felling and processing soft broadleaves is more convenient than its use in felling and processing hard broadleaves, because stem characteristics produce less significant effects on the harvester productivity, primarily due to lower wood density.

Results of this kind have also been reached by other authors who have studied the efficiency of the harvester in the stands of different tree species. They agree that the use of harvester in natural stands of hard broadleaves is less effective and more costly than its use in conifer cultures and plantations of soft broadleaves (Krpan et al. 2004).

There are certain other factors that affect the harvester productivity. They are: operator skills, terrain conditions, felling intensity, felling time, etc. (Mak-

544

konen 1991, Eliasson 1998, Stampfer 1999, Nurminen et al. 2006). These factors were not important for this research because the harvester was operated by only one worker on the flat terrain in winter working conditions.

This research has proved just like the studies of other authors that stem size affects the productivity of the harvester (Brunberg et al. 1989, McNeel and Rutherford 1994, Kellogg and Bettinger 1994, Lageson 1997, Glode 1999).

The initial hypotheses that the method of harvester work and stem forking have significant effects on the harvester productivity in the mature poplar plantations have been confirmed.

5. Conclusions â&#x20AC;&#x201C; Zaklju~ci The average productivity of John Deere 1470 D ECO III harvester, used in the regular cuttings of poplar trees in Populus Ă&#x2014; euramericana 'I-214' plantations varied from 30.3 to 34.7 m3/h, depending on the method of work.

The harvester is the most productive when it moves backwards between two rows and cuts two stems from one standing point, first a stem in the row to the right and then a stem in the row to the left, observed from the moving direction. It starts felling trees in the next two rows from the opposite side of the felling site. The same method of felling is used except that the harvester now moves frontally.

Moving and positioning time has the greatest effect on the average productivity of the harvester and varies with the method of work, while the stem processing time does not vary with the used methods and does not affect its productivity.

The average time needed to process forked stems amounts to 1.35 min/m3 and it is considerably longer than the average time needed to process stems without forks, which amounts to 0.89 min/m3, because there is no need to rotate the harvester head when processing them.

The effects of the forks on the harvester productivity can be reduced by removing thicker lateral branches off the young poplar trees.

Felling and processing wood assortments in mature poplar plantations does not require the employment of the combined method since the harvester can perform the felling and processing of wood assortments without any major delays. Therefore, it can be concluded that from a technical point of view John Deere 1470D ECOIII harvester is an efficient working instrument of clear-felling in poplar plantations.

Croat. j. for. eng. 32(2011)2

Efficiency of John Deere 1470D ECOIII Harvester in Poplar Plantations (533–548)

6. References – Literatura Andersson, J., Eliasson, L., 2004: Effects of three harvesting work methods on Harwarder productivity in final felling.

Silva Fennica 38(2): 195–202.

Brunberg, T., Thelin A., Westerling S., 1989: Basic data for productivity standards for single-grip harvesters in thinning. The Forest Operations Institute of Sweden. Report 3: 21.

Danilovi}, M., 2006: Study of quality factors in poplar Populus × euramericana 'I-214' and Populus × euramericana 'Ostia' plantations from the aspect of the application of national and European quality standards. PhD Thesis, Forestry Faculty of Belgrade University, 204 p.

M. Danilovi} et al.

Kellogg, L. D., Bettinger, P., 1994: Thinning productivity and cost for mechanized cut-to-length system in the Northwest Pacific coast region of the USA. International Journal of Forest Engineering 5(2): 43–52.

Klunder, R. A., Stokes, B. J., 1994: Productivity and cost of three harvesting methods. Southern Journal of Applied Forestry 18(4):168–174.

Lageson, H., 1997: Effects of thinning type on the harvester productivity and on the residual stand. International Journal of Forest Engineering 8(2): 7–14.

Makkonen, I., 1991: Silver Streak single-grip harvester in Nova Scotia. Forestry Engineering Research Institute of Canada, FERRIC, Pointe Claire, Que. Field Note TR–94: 18.

Eliasson, L. Bengtsson, J., Cedergren, J., Lageson, H, 1999: Comparison of single-grip harvester productivity in clearand shelterwood cutting International Journal of Forest Engineering 10(1): 43–48.

McNeel, J. F., Rutherford, D., 1994: Modelling harvester-forwarder system performance in a selection harvest. International Journal of Forest Engineering 6(1): 7–14.

Eliasson, L., 1998: Analyses of single-grip harvester productivity. Doctoral thesis. Swedish University of Agricultural Sciences. Department of Operational Efficiency. Umeå. 24 p.

Nakagawa, M., Hamatsu, J., Saitou, T., Ishi|a, H., 2007: Effect of tree size on productivity and time required for work elements in selective thinning by a harvester. International Journal of Forest Engineering 18(2): 24–28.

Eliasson, L., 2000: Effects of establishment and thinning of shelter-woods on harvester performance. International Journal of Forest Engineering 11(1): 21–27.

Glöde, D., 1999: Single and double grip harvesters productive measurements in final cutting of Shelterwood. International Journal of Forest Engineering 10(2):63–74.

Glöde, D., Sikström, U., 2001: Two felling methods in final cutting of shelterwood, single-grip harvester productivity and damage to the regeneration. Silva Fennica 35(1): 71–83.

Hånell, B., Nordfjell, T., Eliasson, L., 2000: Productivity and costing of shelterwood harvesting. Scandinavian Journal of Forest Research 15(5): 561–569.

Hartsough, B. R., Cooper. D. J., 1999: Cut-to-length harvesting of short rotation Eucalyptus. Forest Products Journal 49(10): 69–75.

Jirou{ek, R., Klva~ R., Skoupý, A., 2007: Productivity and costs of the mechanised cut-to-length wood harvesting system in clear-felling operations. Journal of Forest Science 53(10): 476–482.

Kärhä, K., Rönkkö, E., Gumse S.I. 2004. Productivity and cutting costs of thinning harvesters. International Journal of Forest Engineering 15(2): 43–56.

Krpan, A. P. B., Por{insky, T., 2001: Harvester Timberjack 1070 in Croatia. [umarski list 125(11–12): 619–624.

Krpan, A. P. B., Por{insky, T., 2002: Proizvodnost harvestera Timberjack 1070 pri proredi kulture obi~nog bora. [um. list 126(11–12): 551–561.

Krpan, A. P. B., Por{insky, T., Stanki}, I., 2004: Efficiency of Mechanical Felling and Processing in Soft and Hardwood broadleaved stands – Part 3: Efficiency of harvester in natural thinning stands of hardwood broadleaf species). [umarski list 128(9–10): 495–508.

Croat. j. for. eng. 32(2011)2

Nonaka, I., Takeuchi, H., 1995: The knowledge-Creating Company. Oxford University Press. 1995. New York.

Puttock, D., Spinelli, R., Hartsough, B. R., 2005: Operational trials of cut-to-length harvesting of poplar in a mixed wood stand. International Journal of Forest Engineering 16(1): 39–49.

Nurminen, T., Korpunen, H., Uusitalo, J., 2006: Time consumption analysis of the mechanized cut-to-length harvesting system, Silva Fennica 40(2): 335–363.

Ovaskainen, H., Uusitalo, J., Väätäinen, K., 2004: Characteristics and Significance of Harvester Operators’ Working Technique in Thinnings. International Journal of Forest Engineering 15(2): 67–77.

Sirén, M., 1998: One grip harvester operation, its silvicultural result and possibilities to predict tree damage.

PhD Thesis, Finnish Forest Research Institute, Research Papers 694, 179 p.

Stampfer, K., Steinmüeller, T., 2001: A new approach to derive a productivity model for the harvester »Valmet 911 Snake«, International mountain logging and 11th Pacific Northwest Skyline Symposium, Institute of forest and mountain risk engineering, University of Agricultural science Vienna, Austria.

Stampfer, K., 1999: Influence of terrain conditions and thinnings regimes on productivity of a trackbased steep slope harvester. In: Proceedings of the International Mountain Logging and Tenth Pacific Nortwest Symposium, March 28 – April 1, 1999, Cornvallis, Oregon 78–87 p.

Suadicani, K., Fjeld, D., 2001: Single-tree and group selection in montane Norway spruce stands: Factors influencing operational efficiency. Scandinavian Journal of Forest Research 16(1): 79–87.

545

M. Danilovi} et al.

Efficiency of John Deere 1470D ECOIII Harvester in Poplar Plantations (533–548)

Tufs, R. A., Brinker, R. W., 1993: Walmet’s woodstar series harvesting system: A case stady. Southern Journal of Applied Forestry 17(2): 69–74.

Wester, F., Eliasson, L., 2003: Productivity in final felling and thinning for a combined harvester-forwarder (Harwarder).

International Journal of Forest Engineering 14(2): 45–51.

Tufts, R. A., 1997: Productivity and cost of the Ponsse 15-series, cut-to-length harvesting system in southern pine plantations. Forest Products Journal 47(10): 39–46.

Sa`etak

Djelotvornost harvestera John Deere 1470D ECOIII u topolovim planta`ama U nastojanju za ve}om razinom mehaniziranosti pridobivanja drva, u {umarstvu u Srbiji od 2008. godine za sje~u i izradbu drva u topolovim planta`ama duge ophodnje koristi se harvester John Deere 1470D ECOIII. Budu}i je cilj zamjena radnika sjeka~a harvesterima na mjestima na kojima postoje uvjeti za mehaniziranu sje~u i izradbu drva, najvi{e zbog pomanjkanja kvalificirane radne snage. Primjena malih harvestera u proredama mekih lista~a logi~no je rje{enje, ali je njihov udjel u ukupnoj povr{ini planta`a u Srbiji vrlo malen. Poslednjih petnaest godina uglavnom se osnivaju planta`e simetri~noga rasporeda sadnje namijenjene proizvodnji obloga drva za mehani~ku preradbu.

Imaju}i na umu ~injenicu da sjeka~i pri sje~i i izradbi drva u topolovim planta`ama zrelim za sje~u ostvaruju velike u~inke, me|u ostalim i zbog toga {to je posljednjih godina sve vi{e zastupljena izradba vi{emetarskoga prostornoga drva, primjena harvestera promatrana s ekonomskoga aspekta mo`e biti isplativa samo pri njihovim visokim u~incima.

Primjena harvestera u redovitim sje~ama topole naizgled je jednostavno rje{enje s obzirom na to da je rije~ o ~istoj sje~i na ravnim terenima, homogenoj strukturi sastojine s aspekta dimenzija i rasporeda stabala te o drvu male tvrdo}e. U svim, pa i u ovim uvjetima rada postoji velik broj ~imbenika koji utje~u na proizvodnost sredstava za rad, me|u kojima su najzna~ajniji terenski uvjeti te zna~ajke stabala.

Topolove planta`e duge ophodnje nalaze se na terenima ograni~ene nosivosti s visokom razinom podzemnih voda te s ~estim poplavama koje dovode do izra`ene zakrivljenosti donjega dijela stabla. Osim toga stabla su velikih dimenzija s izra`enim vanjskim zna~ajkama (granatost, zakrivljenost, ra{ljavost i dr.).

Utjecaj zna~ajki stabala na djelotvornost harvestera ogleda se u prekidima rada, osobito pri kresanju ja~ih grana, o{te}enju donjega najvrednijega dijela debla stabla zbog vi{estruke zakrivljenosti, kao i zastoja pri trupljenju ra{lji.

Pojava ra{ljavosti stabala lista~a ~esta je pojava, me|utim udjel se tih stabala mo`e znatno smanjiti uzgojnim mjerama, kao {to su korekcija vrha mladih biljaka te kresanje grana s mladih dube}ih topolovih stabala.

Tako|er euroameri~ke topole imaju izra`enu vla`nost srca debla, {to zna~ajno utje~e na raspucavanje donjega najvrednijega dijela stabla pri njegovu ru{enju. Utjecaj na proizvodnost harvestera imaju i ostali ~imbenici, me|u kojima su i metode rada.

Cilj je ovoga rada da se istra`i djelotvornost harvestera John Deere 1470D ECOIII u topolovim planta`ama duge ophodnje primjenom ~etiriju razli~itih metoda rada te da se ocijeni zna~enje, odnosno utjecaj ra{ljavosti stabala na djelotvornost strojne sje~e i izradbe drva harvesterom.

Istra`ivane su se metode razlikovale u ovisnosti o smjeru kretanja harvestera, o njegovu polo`aju u odnosu na red koji sije~e te broju stabala koja sije~e s jednoga stajali{ta.

Istra`ivanje je provedeno u planta`i topole (Populus × euramericana 'I-214') na podru~ju ravnoga Srijema u Srbiji, na povr{ini od 6,89 ha u razdoblju od prosinca 2009. do velja~e 2010. godine. Primjerne povr{ine na kojima je obavljeno snimanje razli~itih metoda rada postavljene su duljom stranicom sje~ine, a radni su uvjeti bili isti za sve ~etiri istra`ivane metode rada. Srednji je promjer stabala na panju iznosio 46,3 cm, a na prsnoj visini 40,4 cm.

Pri istra`ivanju je primijenjen studij rada i vremena, a trajanje radnih operacija snimano je proto~nom metodom kronometrije.

Vrijeme rada harvestera podijeljeno je na ove radne sastavnice: premje{tanje vozila, zauzimanje polo`aja, ru{enje stabla te izradba drva.

546

Croat. j. for. eng. 32(2011)2

Efficiency of John Deere 1470D ECOIII Harvester in Poplar Plantations (533–548)

M. Danilovi} et al.

Posebno su evidentirana ra{ljava stabla, odnosno stabla koja nije bilo mogu}e izraditi bez prethodnoga zakretanja harvesterske glave pri trupljenju krakova ra{lji. Udjel tih stabala na svim ~etirima primjernim povr{inama iznosi 13,6 %.

Pri izradbi sortimenata presudne su bile minimalne dimenzije propisane odredbama nacionalnih normi za oblo drvo i zna~ajke debla stabla (kvrge i zakrivljenost). Po`eljno trupljenje drva preko kvrga zbog smanjena raspucavanja izra|ene oblovine bilo je posebno va`no s obzirom na mogu}i utjecaj na utro{ak vremena izradbe drva.

Stabla za analizu u okviru svake od istra`ivanih metoda razvrstana su po dva kriterija. Prvi je kriterij bio promjer na prsnoj visini stabla, a drugi ra{ljavost stabla.

Razlike izme|u srednjih vrijednosti zna~ajki istra`ivanih metoda rada te izme|u srednjih vrijednosti zna~ajki ra{ljavih i stabala bez ra{alja istra`ene su dvostrukom analizom varijance. Zavisne su varijable bili utro{ci vremena sastavnica rada harvestera, a faktori metoda rada te ra{ljavost stabla. Osim toga uklju~ena je i kovarijàbla za elemente gdje je postojala potreba, odnosno gdje su ispunjeni uvjeti za njezino uklju~ivanje.

Udio vremena premje{tanja vozila u efektivnom vremenu rada zna~ajno se razlikuje u ovisnosti o primijenjenoj metodi rada te iznosi od 5,8 do 15,7 %; najve}i je u prvoj, a najmanji u ~etvrtoj metodi rada. Na razlike je najvi{e utjecalo vrijeme premje{tanja harvestera s kraja jednoga na po~etak drugoga reda.

Smjer kretanja harvestera pri sje~i i izradbi drva nema zna~ajan utjecaj na vrijeme zauzimanja polo`aja, dok broj stabala koje harvester sije~e s jednoga stajali{ta zna~ajano utje~e na vrijeme zauzimanja polo`aja i diferencira istra`ivane metode u dvije homogene skupine. Utro{ci vremena iznose od 12,1 % do 15,9 % efektivnoga vremena rada, odnosno 0,25 m3 (prva metoda rada) do 0,18 min/m3 (~etvrta metoda rada).

Ra{ljavost je stabla, tako|er, zna~ajno utjecala na vrijeme zauzimanja polo`aja harvestera, odnosno voza~ je harvestera utro{io vi{e vremena na zauzimanje polo`aja pri ru{enju ra{ljavih stabala zbog uklje{tenja stabla koje se ru{i u ra{lje preostalih dube}ih stabala ili zbog pojave raspucavanja pridanka pri sje~i stabla koje se ru{i. Udio vremena ru{enja stabala u efektivnom vremenu po istra`ivanim metodama rada iznosi od 5,2 do 6,7 %, pri ~emu su razlike vezane uz prostor bez dube}ih stabala. Primjera radi, kada se harvester kre}e unazad i kada se sije~e stablo s desne strane promatrano u smjeru kretanja, oslobo|en je ve}i prostor i preglednost tijekom ru{enja jer se ispred stabla koje harvester ru{i ne nalazi dube}e stablo. U ovom slu~aju kut pod kojim harvester ru{i stabla u odnosu na os kretanja vozila je 0 do 25° (slika 2), s tim {to se u prostoru od 0 do 10° ne nalaze dube}a stabala, pa je lak{e ru{enje. Takva je situacija prisutna i u tre}oj metodi rada, s tim {to je prostor za ru{enje stabla s desne strane promatran sa smjerom kretanja slobodan do 20°. Pri primjeni druge i ~etvrte metode rada u tom se prostoru nalaze dube}a stabla koja ote`avaju ru{enje stabala u `eljenom smjeru, a posebno ra{ljavih. Promjer stabla nema zna~ajan utjecaj na utro{ak vremena ru{enja, iako stabla promjera na panju iznad 55 cm nije bilo uvijek mogu}e sru{iti u `eljenom smjeru. U tim je slu~ajevima zauzimanje polo`aja harvestera trajalo dulje, ponajprije zbog sigurnosti pri radu.

Pri tome se harvester kre}e izme|u redova i sije~e stabla iz oba reda, a voza~ harvestera dodatno zauzima polo`aj prije ru{enja stabla, odnosno nije u mogu}nosti da u svim slu~ajevima sru{i dva stabla s jednoga stajali{ta bez pomicanja vozila.

Vrijeme izradbe drva stabla iznosi u prosjeku oko 70 % efektivnoga vremena rada i ne razlikuje se zna~ajno me|u istra`ivanim metodama te, ovisno o istra`ivanim metodama rada, iznosi od 0,91 do 1,0 min/m3. Veza je izme|u obujma stabla i utro{ka vremena izradbe drva povezana.

Ra{ljavost stabla zna~ajno utje~e na vrijeme izradbe drva, odnosno izme|u prosje~noga vremena izradbe ra{ljavih stabala i stabala bez ra{alja postoje zna~ajne statisti~ke razlike. Prosje~no vrijeme izradbe stabala bez ra{alja iznosi 0,89 min/m3, a ra{ljavih stabala 1,35 min/m3. Na ve}e utro{ke vremena izradbe ra{ljavih stabala utjecalo je zakretanje harvesterske glave pri trupljenju krakova ra{lji.

Prosje~ni u~inci u ovisnosti o metodi rada harvestera kre}u se od 30,3 do 34,7 m3/h uz dodatno vrijeme od 28,5 %.

Najve}i je u~inak harvester ostvario kada se kretao izme|u dvaju redova i kada je sjekao dva stabla s jednoga stajali{ta (metoda 4), a najmanje kada se kretao unazad i kada je sjekao stabla iz jednoga reda, odnosno jedno stablo po stajali{tu (metoda 1). Prosje~an u~inak harvestera pri primjeni metode 2 iznosi 32,3 m3/h, a pri metodi 3 prosje~an je u~inak 34,0 m3/h.

Izravni tro{kovi rada harvestera John Deere 1470D ECOIII izra~unati na na osnovi 1600 pogonskih sati i iznose 121,3 EUR/h.

Jedini~ni su tro{kovi najmanji pri primjeni metode rada 4 te iznose 3,5 EUR/m3, a najve}i pri metodi rada 1 i iznose 4,0 EUR/m3. Na osnovi ovih istra`ivanja izlazi da je najve}a djelotvornost harvestera kada se primjenjuje metoda rada 4.

Croat. j. for. eng. 32(2011)2

547

M. Danilovi} et al.

Efficiency of John Deere 1470D ECOIII Harvester in Poplar Plantations (533–548)

Zna~ajke topolovih stabala u planta`ama duge ophodnje imaju velik utjecaj na djelotvornost rada harvestera, ali ne toliko da uvjetuju primjenu kombinirane ru~no-strojne sje~e i izradbe drva motornim pilama lan~anicama sa strojnom sje~om harvesterom, kao {to je to ~esto slu~aj u sastojinama tvrdih lista~a.

Klju~ne rije~i: harvester, topolove planta`e, zna~ajke stabla, metode rada, proizvodnost, tro{kovi

Authors’ addresses – Adresa autorâ: Asst. Prof. Milorad Danilovi}, PhD. e-mail: milorad.danilovic@sfb.rs Asst. Prof. Dragan Ga~i}, PhD. e-mail:dragan.gacic@sfb.rs Forestry Faculty of Belgrade University Kneza Vi{eslava 1 11030 Belgrad SERBIA

Received (Primljeno): May 20, 2010 Accepted (Prihva}eno): July 6, 2011

548

Ivan Toma{evi}, BSc. ivan.tomasevic@sezampro.rs »Vojvodina{ume« Public enterprise Preradovi}eva 2 21000 Novi Sad SERBIA Croat. j. for. eng. 32(2011)2

Original scientific paper – Izvorni znanstveni rad

Analysis of Accidents During Cable Yarding Operations in Austria 1998–2008 Petros A. Tsioras, Christian Rottensteiner, Karl Stampfer Abstract – Nacrtak This paper deals with accidents in the period 1998 – 2008 reported during harvesting operations in ÖBf AG, Austria’s largest forest enterprise, with a focus on cable yarding. In total 1888 accidents were recorded with 8.7% of these associated with cable yarding activities.

The overall accident rate amounted to 36 incidents per million cubic meter extracted by cable yarder. There was a clear spike in accidents between 2 and 3 pm. Most accidents occurred on Wednesdays and Mondays (26 and 25%, respectively). The four months of March, June September and November accounted for 45% of all accidents. The majority of accidents (63.2%) were caused by broken spar and anchor trees, bouncing cables and falling objects.

Frequently injured body parts were the extremities such as hand and feet (64%) and the head and neck area (15.2%). Contusions (37.8%) are the most common kind of injury followed by bone fractures (12.8%), sprain or strains (11.6%) and punctures or lacerations (10.4%). An average cable yarding accident required 17.9 days for recovery, compared to the 25.6 days for manual extraction and 29.7 days for extraction by means of a tractor or skidder. Accident severity varied among body parts: eye injuries resulted in three lost work days, while injured extremities required 20 days for recovery. During wood extraction, some stems with branches may be unhooked and fall down. In this case, the accidents caused are the most severe needing 27 days for recovery.

Keywords: cable yarding, forest operations, safety, accident statistics, incident rate, Austria

1. Introduction – Uvod Sustainable forest management sets new levels of standards on forest operations. In this context, forest operations should incorporate technically feasible, economically viable, environmentally sound and institutionally feasible solutions (Heinimann 2000).

Due to technological advancements a wide variety of forest machinery is currently available. However, the optimization of wood harvesting systems implies in depth knowledge of all their elements and factors that could possibly affect their efficiency.

The steepness of the terrain is a limiting factor with regard to the use of machinery in timber extraction. Cable yarding represents an environmentally friendly option of wood extraction, especially in steep terrain compared to other existing technologies (Holzwieser 1998, Visser and Stampfer 1998).

The fact that most forestlands of Austria are located in the mountainous regions of this country, where approximately 62% of the forest area has slope greater than 30%, and 24.7% of the area is located on Croat. j. for. eng. 32(2011)2

slopes greater than 60% (BFW 2011), has led to extensive use of cable yarding systems. Use of mobile cable-yarders is mainly carried out in areas exceeding 40% in slope (Gschwantner 2009). The importance of cable yarding systems for Austrian forestry is underlined by the fact that every fifth cubic meter of timber is extracted by cable yarding systems (BMLFUW 2009).

A crucial dimension of the social aspects of forest work is its impact on the safety and health of the workforce (Heinimann 2000). Forestry professions belong to the most dangerous jobs in all fields of production (Peters 1991, Poschen 1993, Mitchell et al.

2001, Bentley et al. 2005, Poto~nik et al. 2009, Lindroos and Burström 2010). Timber harvesting, with or without machinery, is difficult, especially on steep slopes and is connected with high risk of accidents. The increased accident frequency in forest operations has triggered many studies with regard to accident analysis in forestry. Most of them focused on chainsaw accidents (Mc Farlane 1977, Doyle and Conroy 1989,

549

P. A. Tsioras et al.

Analysis of Accidents During Cable Yarding Operations in Austria 1998–2008 (549–560)

Wang et al. 2003, Bentley et al. 2005, Montorselli et al.

2010) and fatal accidents in forest operations (Jarl 1980, Peters 1991, Rodriguez-Acosta and Loomis 1997, Mitchell et al. 2001, Thelin 2002). A general evaluation of the severity of accidents in forest operations in Slovenia was done by Poto~nik, et al. (2009), while Suchomel and Balanová (2009) analyzed the impact of weather conditions on injury during harvesting operations. However, no study so far has examined the characteristics of cable yarding accidents.

The objective of this study was to increase our knowledge about accidents during cable yarding operations through the analysis of the ÖBf accident archive in the period 1998 – 2008. The analysis of these data is discussed and proposals for the promotion of safety and health during work are made.

2. Material and Methods – Materijal i metode The evaluation is based on data of the Austrian State Forest Enterprise ÖBf AG. ÖBf AG is responsible for the management of 14.8% of the total forest area of Austria (Bundesamt für Wald 2011) and, among other activities, harvests an annual timber volume of more than two million m3 (Österreichische Bundesforste 2011). According to the current legislation in Austria, all accidents resulting in at least three days of lost work have to be reported (ASVG 2011). However, the company has been systematically keeping record of all accidents of its employees since 1981, including all minor accidents which resulted in less than three lost work days.

During the data capture period, tower yarders were mainly used in combination with processors.

The most common mode of extraction was the whole-tree-method. The trees were felled and their top was removed by chainsaw. The trees, after being extracted uphill, were processed and bucked on the forest road and later transported to the sawmill. All forest workers had been offered protective equipment (helmet with ear and eye protection, cut-proof trousers, boots and gloves) as well as training courses by the company.

In the context of this study, the ÖBf archives for the period 1998 – 2008 were examined for cable yarding accidents. All accidents that occurred: a) due to a fall from trees or cable towers during the setting-up, dismantling, repair and maintenance of cable yarders or b) during wood extraction by means of cable yarding technology, were identified as cable yarding accidents and included for further analyses.

The final version of the dataset included details on five fields: personal data of the injured worker (year of birth etc.), temporal data (month, day and

550

time of accident), cause of accident, injured part of the body and type of injury. Six causes of accident were identified: falling down from a tree or a cable tower, work with a cable yarder (during the setting up, dismantling, repair or maintenance process), falling objects (e.g. support spar), falling loaded tree(s) or loading carriage, during loading or unloading the yarder carriage and stab wound caused by cable.

Injured body parts were categorized into seven groups: head and neck, chest and back, abdomen and pelvis, arms and hands, legs, feet and multiple injured body parts. In the case of multiple injuries no more detailed information was available. With regard to the type of injury, four major groups were identified: by mechanical force (e.g. lacerations, sprains, broken bones, eye injuries, etc), by natural forces (e.g. lightning, sunburns), by chemical action or poisoning and others (e.g. insect bites).

Production data of the company enabled the estimation of commonly used incidence rates such as the number of accidents per million production units, in our case cubic meters, as well as per million work hours for every year. The average annual number of work hours per full time worker was estimated to be 1808, excluding vacation days and national holidays.

It should be noted that production data for cable yarding were available only for the period 2001 – 2008. The recorded lost work hours per accident were used for the classification of accident severity.

To obtain the number of lost work days an eight-hour workday was assumed.

Statistical analysis has been conducted with the help of the statistical package SPSS 17. Statistical differences have been checked with the help of chi-square tests, with Yates correction for continuity when necessary (Bremmer et al. 2000). The test for significance level was set to 5%.

3. Results and discussion – Rezultati i rasprava 3.1 All types of accidents – Svi tipovi nesre}a During the period 1998 – 2008 a total of 1888 accidents occurred within all fields of production of the enterprise (Table 1). The large majority of accidents (1769 – 93.7%) affected forest workers of ÖBf while 119 (6.3%) administrative employees. Also, 64 commuting accidents have been recorded, 56 of which occurred on the way to or back from work and eight accidents occurred during work activities.

The overall incident rate for the study period for all types of accidents was 92.8 accidents per million work hours with its lowest value (70.8) in 2005 and Croat. j. for. eng. 32(2011)2

Analysis of Accidents During Cable Yarding Operations in Austria 1998–2008 (549–560)

P. A. Tsioras et al.

Table 1 Accidents and accident rate during the study period Tablica 1. Nesre}e i stopa nesre}a u istra`ivanom razdoblju Year Godina 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 Total Ukupno

Accidents Nesre}e Non fatal Bez smrtne posljedice 242 243 168 169 163 206 121 102 138 164 161

Fatal Smrtne 2 3 0 0 0 0 2 0 1 3 0

1877

11

Type of injured employees Commuting accidents Vrsta ozlije|enih zaposlenika Nesre}e pri putovanju Forest workers Admin. personnel On the way to work During work [umarski radnici Admin. osoblje Na putu za posao Za vrijeme posla 232 12 8 0 231 15 3 3 161 7 4 0 155 14 9 0 152 11 6 2 200 6 2 1 120 3 4 0 100 2 5 0 128 11 7 1 146 21 4 1 144 17 4 0 1769

its highest (119.5) in 2003. Accordingly, the overall incident rate per million production units was estimated to be 77.6 accidents per million cubic meters, ranging from 49 (2005) to 118 (1999) (Table 1).

A total of 11 fatal accidents have been recorded during the study period, resulting in a rate of 0.49 fatal accidents per million cubic meters or 0.6 fatal accidents per million work hours.

119

56

8

Accident rate per million work hours Stopa nesre}a na milijun radnih sati 107.3 112.3 91.8 89.0 96.2 119.5 79.6 70.8 87.9 91.3 75.2 x = 92.8

3.2 Cable yarding accidents – Nesre}e pri izvla~enju drva `i~arom 3.2.1 Demographic distribution – Demografska raspodjela A total of 164 accidents occurred during cable yarding operations which amounts to 8.7% of all accidents (Table 2). The years with the highest number

Table 2 Trend of overall accidents and accidents during cable yarding operations Tablica 2. Kretanje ukupnih nesre}a i nesre}a pri izno{enju drva `i~arom Year Godina 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 Average Prosjek Total Ukupno

Employees Zaposlenici 1593 1398 1324 1259 1202 1188 1103 1100 1108 1146 1174

Accidents during cable yarding, n Nesre}e pri izno{enju `i~arom, n 23 22 10 9 9 25 10 7 21 15 13

1236

14.9

8.7

13.9

13.595

164

–

–

Croat. j. for. eng. 32(2011)2

Percentage of accidents during cable yarding, % Percentage of overall injured employees, % Udio nesre}a pri izno{enju `i~arom, % Udio ukupno ozlije|enih radnika, % 9.4 15.3 8.9 17.6 6.0 12.7 5.3 13.4 5.5 13.6 12.1 17.3 8.1 11.2 6.9 9.3 15.1 12.5 9.0 14.6 8.1 13.7

551

P. A. Tsioras et al.

Analysis of Accidents During Cable Yarding Operations in Austria 1998–2008 (549–560)

The increased accident frequency in 2003 might be attributed to the large areas of windthrown forest areas in this year (Österreichische Bundesforste 2004). Under pressure to recover large quantities of windthrown wood, more workers and forest machines were recruited by ÖBf. This fact, which has changed the composition of the cable yarding working teams (Eiwegger 2009) along with the increased difficulty of processing and extracting wind thrown trees (Odenthal-Kahabka 2005, Sonnleitner and Seebacher 2003) are possibly related to the increased accident frequency for this year.

Skidding accidents represent 14.1% of the total number of accidents for the study period, with cable yarding responsible for 8.7% of them. This percentage is lower compared to other studies that analysed skidding accidents in Slovenia (24%) (Poto~nik et al.

2009), New Zealand (22%) (Gaskin and Parker 1993) or Sweden (20%) (Engsås 1995) but higher than 4.6% reported for cable yarding accidents in Switzerland (Wettman 2005). Such differences can be attributed to a number of factors, such as different ownership, equipment and technology status (Poto~nik et al.

2009).

All injured employees were male and the average age at the time of the injury was 38.8 years (SD±10.04).

The youngest of them was 18 and the oldest 59 years old. More than half of them (51.2% – 84 forest workers) were between 30 and 45 years old (Fig. 1). After the age of 44 the number of accidents is declining.

This finding is similar to Bentley et al. (2005) who

Fig. 1 Distribution of accidents in age classes Slika 1. Raspodjela nesre}a po dobnim razredima of accidents during cable yarding were 2003 (25 accidents) and 1998 (23 accidents). The highest proportion of cable yarding accidents can be found in 2006 with 15.1% (21 accidents). The minimum number of cable yarding accidents was recorded in 2005, when only seven people (6.9%) were injured, while the lowest proportion of cable yarding accidents was found for the year 2001 with 5.3% (nine accidents).

Table 3 Cable yarding accidents per extracted million cubic meters for the period 2001 – 2008 Tablica 3. Nesre}e pri izvla~enju drva `i~arom na milijun kubnih metara za razdoblje od 2001. do 2008.

Year Godina

Extracted volume Privu~eni obujam

2001 2002 2003 2004 2005 2006 2007 2008 Average Prosjek Total Ukupno

25 1239 23 6921 24 6959 29 0342 29 3799 35 5099 36 2573 32 3768

552

Number of accidents – Broj nesre}a Accidents leading to three or more lost days All accidents Nesre}e s tri ili vi{e dana Sve nesre}e bolovanja 8 9 8 9 20 25 5 10 5 7 17 21 14 15 7 13

Accidents per million m3 – Nesre}e na milijun m3 Accidents leading to three or more lost days All accidents Nesre}e s tri ili vi{e dana Sve nesre}e bolovanja 31.8 35.8 33.8 38.0 81.0 101.2 17.2 34.4 17.0 23.8 47.9 59.1 38.6 41.4 21.6 40.2

29 5088

10.5

13.6

35.6

46.2

236 0700

84

109

–

–

Croat. j. for. eng. 32(2011)2

Analysis of Accidents During Cable Yarding Operations in Austria 1998–2008 (549–560)

P. A. Tsioras et al.

Fig. 2 Breakdown per accident time Slika 2. Raspodjela po vremenu nesre}e also found the age group of 35 – 44 years to be more prone to accidents, even compared to workers in their first months in employment. However, in our study and for the years 2005 – 2008, the comparison between the age distribution of forest workers and cable yarding accident frequency per age group showed no significant differences (c2 = 9.333, df = 7, p = 0.23), suggesting that the accident frequency reflects the workforce age distribution.

The trend of decline of injuries as workers’ seniority increased is evident in other studies (Wang et al.

2003). This fact could be attributed to the increased expertise of the older forest workers. This result is also consistent with other studies (Driscoll et al. 1999) that analysed fatalities in forest work. However, the fact that accidents do happen at increased frequency for experienced workers might imply the need for training throughout the worker’s career, not only at the beginning of employment (Wang et al. 2003).

Incident rates for cable yarding were estimated only for the period 2001 – 2008 (Table 3). During this period, an average of 295,000 m3/year were extracted by means of cable yarding technology, resulting in an incident rate of 46.2 accidents per million cubic meters or 35.6 accidents per million cubic meter if minor accidents are excluded.

3.2.2 Temporal analysis – Vremenska analiza The accidents are not uniformly distributed across different hours (c2= 38.195, df = 10, p < 0.0000) and the majority them occurred between 10:00 – 12:00 and 14:00 – 15:00 (Fig. 2). The risk of injury seems to Croat. j. for. eng. 32(2011)2

be increasing during the morning hours until lunch break. The peak between 10 am and 12 pm has been reported in other studies (Bentley et al. 2005, Wettmann 2005, Fischer 1991, Stadlmann 1991) and could probably be related to the circadian rhythms of the human body. However, in our study, most accidents (18.9%) occurred after lunch in the time interval 14:00 – 15:00. The second peak between 2 and 3 pm could be attributed to the »lunch effect« (Camino-López et al. 2011), according to which food consumption might be related to increased accident frequency in the hours following the lunch. After this time point, accident risk seems to be decreasing till the end of the shift.

The distribution of accidents is uneven across the weekdays (c2 = 46.805, df = 5, p < 0.0000). The highest incident rate was reported on Wednesday (26.8% or 44 accidents) and the next highest on Monday (25% or 41 accidents) compared to other studies that place Monday and Tuesday as the most dangerous days of the week (Wettmann 2005, Jacke 1989, Fischer 1991).

The peak on Monday could be attributed to the change to job tasks after the weekend rest (Jacke 1989), and the high number of accidents on Wednesday to fatigue. After Wednesday, accident frequency is declining, reaching its lowest value on Friday with 9.8% or 16 accidents. The second half of the week shows a decrease in incidents which might be attributed to higher motivation due to the upcoming weekend.

The highest number of accidents (19) was reported for October, followed by March, June and November with 18 accidents each (Fig. 3). No statis-

553

P. A. Tsioras et al.

Analysis of Accidents During Cable Yarding Operations in Austria 1998–2008 (549–560)

Fig. 3 Distribution of accidents per month Slika 3. Raspodjela nesre}a po mjesecima tical differences have been found (c2 = 11.555, df = 11, p = 0.398). Finally, April was the month with the lowest incident frequency of eight.

3.2.3 Cause of accident – Uzrok nesre}e The prevailing activities at the time of the accident were the hooking and unhooking of loads using a choker cable (43%) and the setting-up, dismantling, maintenance or repair (33%), respectively. In the majority of incidents, the forest workers were struck by or striked against an object (76%). The rest injury initiating events included vehicle rollovers (18%), and slips, trips and falls (6%).

According to the analyzed dataset, broken support and anchor trees, bouncing cables and falling objects or tree stems caused almost two third of all accidents during cable yarding operations. This category of accidents is very common in forest operations (Driscoll et al. 1999, Wang et al. 2003, Bentley et al. 2005).

According to Peters (1991), such accidents are often due to violations of safe work practices and poor working technique and need proper investigation.

Unfortunately, limited conclusions can be drawn with regard to the working activity during an accident and the cause of accident due to the lack of more detailed data. Our inability to investigate the underlying reasons stems from the accident recording system used by the company, which successfully fulfills the requirements set by AUVA, however inhibits further analysis of accidents. This drawback could be addressed by the recording of supplemental information on the accident report sheet (Eiwegger 2009).

554

3.2.4 Type of injuries – Oblik ozljeda The most frequent types of injuries were contusions (37.8%), a finding similar to other studies (Poto~nik et al. 2009, KWF 2011) (Table 4). Bone fractures (12.8%), sprains and strains (11.6%) and punctures or lacerations (10.4%) were the rest most common types of injuries. It should be noted that 19 accidents resulted in multiple types of injuries, but unfortunately, no more information on them is available.

Extremities (legs and arms) were the most affected body parts (64% or 105 accidents). This result is close to 66% reported for Slovenia (Poto~nik et al.

2009) and 64% for Germany (KWF 2011) but higher than 51% for the Jilin Province in China (Wang et al.

2003) or 50% for Louisiana (Lefort et al. 2003). Head and neck injuries (15.2%), chest and back injuries (9.8%) and injured abdomen or pelvis (3.0%) followed. Extremities were the most frequently affected body parts with regard to accidents. As a more elaborate analysis reveals (Fig. 4), finger and thumbs represent the extremity parts most frequently affected, which is quite expected due to the nature of cable work. Eyes follow (7.3%), indicating not continuous use of protective equipment but this percentage is lower compared to 16.5% reported by Poto~nik et al. (2009). Head-neck injuries accounted for 15.2% of all accidents, lower than 28% reported

by Wang et al. (2003). Finally, accidents that affected multiple body regions accounted for 7.9% of all incidents.

Croat. j. for. eng. 32(2011)2

Analysis of Accidents During Cable Yarding Operations in Austria 1998–2008 (549–560)

P. A. Tsioras et al.

Table 4 Frequency related to the kind of injuries Tablica 4. U~estalost nesre}a prema vrsti ozljeda Kind of injury – Vrsta ozljede Contusion – Nagnje~enje Bone fracture – Prijelom kosti Sprain, strain trauma or luxation – Uganu}e, istegnu}e ili i{~a{enje Puncture or laceration – Otvorene ili razderane rane Foreign body (e.g. in the eye) – Strano tijelo (npr. u oku) Ligament rupture and meniscus injury – Kidanje ligamenata i ozljede meniska Myorrhexis – Mioreksa, raskidanje mi{i}a Herniated disc and spine defect – O{te}enje diska i ozljede kralje{nice Sinew injury – Ozljede mi{i}a, tetiva Separation of extremities – Odvajanje ekstremiteta Concussion – Potres Comminuted fracture – Usitnjeni lomovi Vascular injury – Ozljede `ila Miscellaneous – Razno Missing information – Bez podataka Total – Ukupno

Accidents, n – Ozljede, n 62 21 19 17 9 5 4 2 1 1 1 1 1 19 1 164

Percentage, % – Udio, % 37.8 12.8 11.6 10.4 5.5 3.0 2.4 1.2 0.6 0.6 0.6 0.6 0.6 11.6 0.6 100

Fig. 4 Injured body regions Slika 4. Ozlije|eni dijelovi tijela Croat. j. for. eng. 32(2011)2

555

P. A. Tsioras et al.

Analysis of Accidents During Cable Yarding Operations in Austria 1998–2008 (549–560)

Fig. 5 Frequency and lost work days per cause of accident Slika 5. U~estalost nesre}a i izgubljeni dani prema uzroku nesre}e 3.2.5 Accident severity – Te`ina nesre}e An average number of 17.9 (or 21.4 lost work days excluding all minor injuries) days were lost per cable yarding accident compared to 25.6 for manual extraction and 29.7 lost days for extraction by means of tractor or skidder, respectively. This is the first indication that cable yarding, despite higher accident

frequency and difficulties caused by steeper terrain, results in less severe accidents. These accident severity values are lower than 32.5 lost work days reported for skidding accidents by (Poto~nik et al.

2009) but surpass the 11 lost work days reported by Bentley et al. (2005). Finally, during the study period only one fatal accident has been recorded. Therefore,

Fig. 6 Frequency and accident severity per injured body part Slika 6. U~estalost i te`ina nesre}a prema ozlije|enim dijelovima tijela 556

Croat. j. for. eng. 32(2011)2

Analysis of Accidents During Cable Yarding Operations in Austria 1998–2008 (549–560)

the incident rate of fatal accidents is very low (0.42 accidents per million cubic meters).

Injuries caused as a result of unhooking loaded stems during extraction demanded the most days (26.6) for recovery (Fig. 5). Breakdown of support and anchor trees, bouncing cables and falling objects are the most frequent reasons that lead to an accident, and resulted in an average of 14.7 lost work days. The least severe incidents were caused by broken cable cords (9.6 lost work days).

Multiple injuries were the most severe resulting in 35 lost work days (Fig. 6). Injuries in the leg and chest-back areas follow with 20.7 and 18.7 lost work days, respectively. On the contrary, eye injuries resulted in an average of only three lost days, while head and neck injuries resulted in an average of eight lost work days.

4. Conclusions and summary – Zaklju~ak i sa`etak High accident frequency in forest operations is a well known fact and has initiated many studies all over the world. Most of these studies have focused on work with chainsaw or report the general accident levels in one country for a specific period of time. Exceptions include works of Väyrynen (1982), who analyzed accidents during the maintenance of heavy forest machinery, as well as of Lindroos et al.

(2008), Lindroos and Burström (2010), who focused on the accidents of special groups of forest workers.

The originality of the current study is its focus on accidents during cable yarding operations. In this context, the current study is focused on cable yarding accidents for a period of eleven years in a large forest enterprise. The choice of ÖBf is justified by the amount of existing data, which enabled the calculation of incident rates, as well as better understanding of the working environment.

The comparison of the presented results with the results of other studies is limited mainly due to the lack of reported incidence rates, such as the number of accidents per million production units or million work hours. According to the analysis of data, cable yarding seems to be an extraction technology with increased frequency but reduced accident severity compared to other skidding methods used by ÖBf. However, in order to reach safer conclusions, more similar studies must be conducted with focus on cable yarding as well as other special aspects of forestry work or equipment.

In a large number of published studies, the extent and importance of minor accidents is unknown as these accidents are not recorded. This is often due to the standards set by the National Insurance or Labor Croat. j. for. eng. 32(2011)2

P. A. Tsioras et al.

Inspectorates (Poto~nik et al. 2009 ASVG 2011). In order to address this problem, the current study includes all cable yarding accidents including the minor ones and provides incident rates for both cases.

This approach enables better understanding of the prevailing situation. Still, there might be a possibility that a limited number of cable yarding accidents have not been included in the present study, as a result of missing explanatory information (e.g. some commuting accidents during work might have taken place with cable yarding machinery on their way to wood extraction sites).

The social dimension of accidents should not be underestimated, and it is expressed in low professional prestige and even an important reason for forest workers to change their profession (Lewark and Härle 1991, Tsioras 2011). This fact underlines the need for more specialized studies on accident analysis during forest operations. The reduction of both accident frequency and severity will only be possible through concerted actions: training of forest workers and use of protection equipment can maximize their effects on prevention strategies, if they are properly assisted by well organized accident recording systems. In times of declining numbers of forest workers (Jacob et al. 1994, Axelsson 1998, Gröger and Lewark 2002, BUS/BUWAL 2003, Salminen et al.

1999, Tsioras 2010), forest worker safety should remain a top priority of national forest policies worldwide.

Acknowledgements – Zahvala The authors wish to thank ÖBf AG, and especially Mr. Stefan Trzesniowski, for the kind provision of accident data. The Editor and the two anonymous reviewers are also thanked for their thoughtful comments and recommendations.

5. References – Literatura ASVG, 2011: Allgemeines Sozialversicherungsgesetz. http://www.jusline.at/363._Unfallmeldung_ASVG.html > (Accessed 3 February 2011).

Axelsson, S.Å., 1998: The mechanization of logging operations in Sweden and its effect on occupational safety and health. International Journal of Forest Engineering 9(2): 25–31.

Bentley, T., Parker, R., Ashby, L., 2005: Understanding felling safety in the New Zealand forest industry. Applied Ergonomics 36(2): 165–175.

BFW 2011. Personal communication with the Head Unit of Inventory Design, Evaluation and Remote Sensing DI Richard Büchsenmeister.

557

P. A. Tsioras et al.

Analysis of Accidents During Cable Yarding Operations in Austria 1998–2008 (549–560)

BMLFUW, 2009: Holzeinschlagsmeldung über das Kalenderjahr 2008 (Timber harvesting report 2008). Wien, Bundesministerium für Land – und Forstwirtschaft, Umwelt und Wasserwirtschaft, 112 p.

Bremmer, J. M., Morgan, B. J. T., Jollife, L. T., Jones, B., North, P. M., Bassett, E. E., 2000: Statistics: Problems and Solutions – Second Edition. World Scientific Publishing.

Bundesamt für Wald, 2011: <> (Accessed 18.3.2011).

BUS/BUWAL, 2003: Wald und Holz – Jahrbuch. Bundesamt fur Statistik/BUWAL, Eidg. Forstdirektion.

Camino-López, M. A., Fontaneda, I., Gonzalez-Alcantara, O. J., Ritzel, D. O., 2011: The special severity of occupational accidents in the afternoon: »The lunch effect«. Accident Analysis and Prevention (Article in press).

Doyle, C. G., Conroy, R. M., 1989: Prevention of timber felling and chainsaw-related accidents in the Republic of Ireland. Accident Analyisis and Prevention 21(6): 529–534.

Driscoll, T., Healey, S., Hendrie, L., Mandryk, J., Mitchell, R., 1999: Work-related traumatic fatalities involving timber activities in Australia, 1989–1992. Epidemiology Unit, National Occupational Health and Safety Commission, Canberra, Australia.

Eiwegger, A., 2009: Arbeitssicherheit bei der Seilrückung – Auswertung von Arbeitsunfällen (Work safety during cable yarding – Evaluation of accidents). Diplomarbeit am Institut für Forsttechnik (Diploma thesis at the Institute of Forest Engineering). Universität für Bodenkultur, Wien, 57 p.

Engsås, J., 1995: Accidents in the private woodlot sector.

Small Scale Forestry 23(1): 15–22.

Fischer, M., 1991: Unfallursachenforschung als Grundlage für die Unfallverhütung (Cause analysis of accidents as basis for accident prevention). Diplomarbeit am Institut für Forsttechnik (Diploma thesis at the Institute of Forest Engineering). Universität für Bodenkultur, Wien, 38 p.

Gaskin, J. E., Parker, R. J., 1993: Accidents in forestry and logging operations in New Zealand. Unasylva 44(172): 19–24.

Gröger, V., Lewark, S., 2002: Der arbeitende Mensch im Wald – eine standige Herausforderung fur die arbeitswissenschaft. Schriftenreihe der Bundesanstalt fur Arbeitsschutz und Arbeitsmedizin Dortmund/Berlin/Dresden, 174 p.

Gschwantner, T., 2009: Technische und ökonomische Rahmenbedingungen der modellierten Holzernte (Technical and economic framework of modelling harvest operations).

BFW-Praxisinformation. 18/2009: 10–12.

Heinimann, H. R. 2000: Forest Operations under Mountainous Conditions. In: Forests in Sustainable Mountain Development – a State of Knowledge Report for 2000 (Price, M. F., Butt, N., eds). Vol. IUFRO Research Series No. 5: p. 224 – 230. CABI Publishing, Wallingford UK. Holzwieser, O., 1998: Die Mechanisierung der Holzernte und des forstlichen Bringungswesens von 1955 bis 1998 (Mechanization in harvesting and extracting operations from 1955 to 1998). In: Beiträge zur Walderschließung und

558

forstlichen Arbeitswissenschaft (Articles on opening-up and ergonomics in forestry) (Stampfer, K., ed.), Universität für Bodenkultur, Wien (8): 37–48.

Jacke, H., 1989: Die Unfallhäufigkeit im Wochenverlauf (Accident frequency along the week days). Allgemeine Forst Zeitschrift 48(6): 1276–1281.

Jacob, J., Huber, M., Wirz, R., Huerle, P. R., Lewark, S., 1994: Warum wechseln waldarbeiter ihren Beruf? Allgemeine Forst und Jagdzeitung 165: 1–6.

Jarl, T., 1980: Fatal occupational accidents in Sweden: Personal safety equipment and rescuing. Accident Analysis & Prevention 12(3): 185–188.

KWF, 2011: Unfallstatistik 1999–2008 (Accident statistics 1999–2008). <http://www.kwf-online.org/unfallstatistik.html> (Accessed 21 February 2011).

Lefort, A. J. Jr., De Hoop, C. F., Pine, J. C., Marx, B. D., 2003: Characteristics of Injuries in the Logging Industry of Louisiana, USA: 1986 to 1998. International Journal of Forest Engineering 14(2): 75–89.

Lewark, S., Härle, P. R., 1991: Das Ausscheiden von Waldarbeitern der Staatsforstverwaltung 1983 bis 1990. Frühinvalidität und Berufswechsel am Beispiel des Bezirks Darmstadt (Departure of forest workers from the state forest administration from 1983 to 1990. Early disability and career changes on the example of the district of Darmstadt). Allg Forstz für Waldwirtschaft und Umweltvorsorge 46: 470–474.

Lindroos, O., Aspman, E. W., Lidestav, G., Neely, G., 2008: Accidents in family forestry’s firewood production. Accident Analysis and Prevention 40(3): 877–886.

Lindroos, O., Burström, L., 2010: Accident rates and types among self-employed private forest owners. Accident Analysis and Prevention 42(6): 1729–1735.

McFarlane, I., 1977: Severe Chainsaw Injuries in RotoruaTaupo Area. Journal of Bone and Joint Surgery (British volume) 59(4): 515–515.

Mitchell, R., Driscoll, T., Healey, S., Mandryk, J., Hendrie, L., Hull, B., 2001: Fatal injuries in forestry and logging work in Australia, 1989 to 1992. Journal of Occupational Health and Safety – Australia NZ 17(6): 567–577.

Montorselli, N. B., Lombardini, C., Magagnotti, N., Marchi, E., Neri, F., Picchi, G., Spinelli, R., 2010: Relating safety, productivity and company type for motor-manual logging operations in the Italian Alps. Accident Analysis and Prevention 42(6): 2013–2017.

Odenthal-Kahabka, J., 2005: Handreichung Sturmschadensbewältigung. Hrsg. Landesforstverwaltung Baden-Württemberg und Landesforsten Rheinland-Pfalz. Österreichische Bundesforste, 2011: Material for the OEBf website < http://www.oebf.at/> (Accesed 25 January 2011). Österreichische Bundesforste, 2004: Nachhaltigkeitsbericht zum Geschäftsjahr 2004 (Sustainability report for the year 2004).

Croat. j. for. eng. 32(2011)2

Analysis of Accidents During Cable Yarding Operations in Austria 1998–2008 (549–560)

Peters, P. A., 1991: Chainsaw felling fatal accidents. Transactions of the American Society of Agricultural Engineers 34(6): 2600–2608.

Poschen, P., 1993: Forestry, a safe and healthy profession?

Unasylva 172: 3–12.

Poto~nik, I., Pentek, T., Poje, A. 2009: Severity Analysis of Accidents in Forest Operations. Croatian Journal of Forest Engineering 30(2): 171–184.

Rodriguez-Acosta, R. L., Loomis, D. P., 1997: Fatal occupational injuries in the forestry and logging industry in North Carolina, 1977–1991. International Journal of Occupational and Environmental Health 3(4): 259–265.

Salminen, S., Klen, T., Ojanen, K., 1999: Risk taking and accident frequency among Finnish forest workers. Safety Science 33(3): 143–153.

Sonnleitner, G., Seebacher, D., 2003. Unfallverhütung bei der Aufarbeitung von Sturmschadholz (Accident protection during processing of windthrown wood). BFW Praxis Information 1: 6–8.

Stadlmann, H., 1991: Motorsägen – Unfallanalyse 1990/91 (Chainsaws – Analysis of accidents 1990/91). SVB, Wien 16 p.

P. A. Tsioras et al.

Suchomel, J., Belanová, K., 2009: Influence of selected meteorological phenomena on work injury frequency in timber harvesting process. Croatian Journal of Forest Engineering 30(2): 185–191.

Thelin, A., 2002: Fatal accidents in Swedish farming and forestry, 1988–1997. Safety Science 40(6): 501–517.

Tsioras, P. A., 2010: Perspectives of the forest workers in Greece. iforest, 3: 118–123.

Tsioras, P. A., 2011: Status and job satisfaction of Greek forest workers. Small-scale Forestry (Article in Press).

Väyrynen, S., 1982: Occupational accidents in the maintenance of heavy forest machinery. Journal of Occupational Accidents 4(2–4): 175–175.

Visser, R., Stampfer, K., 1998: Cable extraction of harvester felled thinnings: An Austrian case study. Journal of Forest Engineering 9(1): 39–46.

Wang, J., Bell, J. L., Grushesky, S. T., 2003: Logging injuries in Jilin Province of the People’s Republic of China. Journal of Safety Research 34(3): 273–279.

Wettmann, O., 2005: Berufsunfaelle in Forstbetrieben 2003 (Occupational accidents in forest enterprises for the year 2003). 26p. SUVA, Luzern.

Sa`etak

Analiza nesre}a pri izno{enju drva `i~arom u Austriji u razdoblju od 1998. do 2008.

U radu se prikazuju nesre}e na radu nastale pri privla~enju drva u austrijskom najve}em {umskom poduze}u ÖBf AG u razdoblju od 1998. do 2008. godine. Posebno se pritom analiziraju nesre}e nastale pri izno{enju drva `i~arom. U promatranom je razdoblju zabilje`eno ukupno 1888 nesre}a, od ~ega je 8,7 % povezano s aktivnostima na izno{enju drva `i~arom. Svi zaposlenici ozlije|eni u analiziranim nesre}ama bili su mu{karci u prosjeku od 38,8 godina u vrijeme ozljede (SD ± 10,04). Najmla|i me|u njima imao je 18, a najstariji 59 godina. Vi{e od polovice njih (51,2 % – 84 {umarska radnika) imalo je izme|u 30 i 45 godina.

Godina s najve}im brojem nesre}a pri izno{enju drva `i~arom bila je 2003. godina, {to se mo`e povezati s nevremenom i opse`nim {umskim podru~jima s velikim vjetroizvalama u toj godini. Zbog potrebe izno{enja velikih koli~ina drva iz takvih podru~ja u {umskom poduze}u ÖBf AG te je godine na pridobivanju drva anga`irano mnogo vi{e radnika i {umskih strojeva. S obzirom na vrijeme pojavljivanja nesre}a utvr|en je jasan vrh u broju nesre}a koje se doga|aju izme|u 2 i 3 sata popodne. Najvi{e se nesre}a dogodilo srijedom i ponedjeljkom (26 % odnosno 25 %). U ~etiri mjeseca, tj. u o`ujku, lipnju, rujnu i studenom, dogodilo se 45 % ukupnoga broja nesre}a.

Najve}i je dio nesre}a (63,2 %) uzrokovan polomljenim sidrenim i potpornim stablima `i~are, zatim odskakivanjem ~eli~noga u`eta `i~are i padom razli~itih objekata.

U~estalo ozlije|eni dijelovi tijela bili su ekstremiteti, odnosno dijelovi ruku i nogu kao {to su {ake i stopala (64 %) te podru~je glave i vrata (15,2 %). Razli~ite kontuzije i nagnje~enja bili su naj~e{}i oblik ozljeda, a za njima slijede prijelomi kostiju (12,8 %), uganu}a i istegnu}a (11,6 %) te razli~iti oblici otvorenih rana, posjekotina i razderotina (10,4 %). Te`ina nesre}a iskazana brojem izgubljenih radnih dana zbog nesre}e razlikovala se prema dijelovima tijela koji su ozlije|eni: ozljede oka imale su za posljedicu tri izgubljena radna dana, dok su ozljede ekstremiteta tra`ile 20 dana bolovanja za oporavak. Najte`e nesre}e koje su zahtijevale 27 dana oporavka zabilje`ene su u slu~ajevima pridobivanja drva u kojima bi se izno{eno deblo pri privla~enju `i~arom otkva~ilo i palo.

Stopa nesre}a pri izvla~enju drva `i~arom u radu je utvr|ena za razdoblje od 2001. do 2008. godine. U tom je razdoblju prosje~no `i~arom privu~eno 295 000 m3/godina. Stopa nesre}a utvr|ena na osnovi toga obujma iznosi Croat. j. for. eng. 32(2011)2

559

P. A. Tsioras et al.

Analysis of Accidents During Cable Yarding Operations in Austria 1998–2008 (549–560)

46,2 nesre}e na milijun kubnih metara drva ili 35,6 nesre}a na milijun kubnih metara, ako se isklju~e manje i lak{e nesre}e. Prosje~an broj izgubljenih radnih dana zbog nesre}a na izno{enju drva `i~arom iznosi 17,9 dana (ili 21,4 dana ako se isklju~e sve manje ozljede). U usporedbi s tim broj izgubljenih radnih dana zbog ozljeda pri ru~nom privla~enju drva ili privla~enju drva traktorima i skiderima iznosi 25,6 odnosno 29,7 dana. Takav je odnos prvi znak da se pri izno{enju drva `i~arom, usprkos ve}oj u~estalosti nesre}a i te`ini postupaka zbog strmoga terena, ipak doga|aju manje ozbiljne nesre}e u odnosu na one pri ru~nom privla~enju i privla~enju traktorima i skiderima.

Mogu}nost je usporedbe prikazanih rezultata s rezultatima drugih istra`ivanja ograni~ena uglavnom zbog nedostatka izvje{taja o stopama nesre}a, npr. o broju nesre}a na milijun proizvodnih jedinica ili milijun radnih sati.

Na osnovi analize prikupljenih podataka mo`e se zaklju~iti da izno{enje drva `i~arom rezultira pove}anom u~estalo{}u, ali i smanjenom te`inom nesre}a u usporedbi s ostalim metodama privla~enja drva koje se primjenjuju u ÖBf AG. Me|utim, radi potvrde takvih zaklju~aka i postizanja sigurnijih rezultata potrebno je provesti vi{e sli~nih istra`ivanja s te`i{tem na izno{enje drva `i~arom i na druge posebnosti {umskoga rada i {umarske opreme.

U velikom broju provedenih istra`ivanja i objavljenih radova opseg i va`nost manjih nesre}e ostaje nepoznat s obzirom na to da se takve nesre}e ne evidentiraju. To je ~esto posljedica razli~itih standarda koje postavljaju dr`avna socijalna i zdravstvena osiguranja i inspekcije rada. S namjerom da se pa`nja posveti i tomu problemu, u ovom su radu obuhva}ene sve nesre}e na izno{enju drva `i~arom, do onih najmanjih, te se daju njihove stope za oba slu~aja. Smatra se da takav pristup omogu}uje bolje razumijevanje postoje}e situacije.

Socijalna dimenzija koju imaju nesre}e na radu ne smije biti podcijenjena, {to je izra`eno u niskom profesionalnom ugledu, pa i u va`nom razlogu za promjenu zanimanja {umskih radnika. Ta ~injenica nagla{ava potrebu za specijaliziranim istra`ivanjima nesre}a pri {umskim radovima. Smanjenje u~estalosti i te`ine nesre}a pritom je mogu}e samo konkretnim akcijama: osposobljavanje i izobrazba {umskih radnika te kori{tenje za{titne opreme, {to mo`e pove}ati njihov utjecaj na razne preventivne strategije, ako su one potpomognute dobro organiziranim sustavima evidencije nesre}a. U doba smanjivanja broja {umskih radnika sigurnost i za{tita zdravlja {umskih radnika trebala bi ostati prioritet nacionalnih {umarskih politika.

Klju~ne rije~i: izno{enje drva `i~arom, {umski radovi, sigurnost, statistika nesre}a na radu, stopa nesre}a, Austrija

Authors’ address – Adresa autorâ: Petros A. Tsioras, PhD. e-mail: ptsioras@for.auth.gr Lab. of Forest Utilization Aristotle University (POB 227) GR-541 24 Thessaloniki GREECE

Received (Primljeno): July 27, 2011 Accepted (Prihva}eno): September 5, 2011

560

Christian Rottensteiner, MSc. e-mail: christian.rottensteiner@bocu.ac.at Assoc. Prof. Karl Stampfer, PhD. e-mail: karl.stampfer@bocu.ac.at University of Natural Resources and Applied Life Sciences Vienna Department of Forest and Soil Sciences Institute of Forest Engineering Peter Jordan Straße 82 1190 Wien AUSTRIA Croat. j. for. eng. 32(2011)2

Original scientific paper – Izvorni znanstveni rad

Workload Benefits of Using a Synthetic Rope Strawline in Cable Yarder Rigging in Norway Giovanna Ottaviani, Bruce Talbot, Morten Nitteberg, Karl Stampfer Abstract – Nacrtak This study examined the difference in workload brought about by exchanging a 3.5 mm steel rope with a 4.0 mm synthetic fiber rope when dragging a strawline up a 300 m corridor in setting up a new cable-yarding line. Physiological workload was monitored through heart rate measurement, while the physical forces acting on the subject (rope mass and friction) were quantified using a dynamometer attached to a belt. While there was a substantial difference in force between rope types at full extent (140 N vs. 40 N), the result was less significant when seen against the total work required in moving the subjects own body mass up the slope. The direction of the resultant force vector appears to play an important role in the way that strain is experienced. It was discovered that 300 m was the maximum hauling distance for a single person using this rigging method with a steel wire strawline, whereas for the synthetic rope, the same tensile force would only be reached at 1200 m. This alone has important implications for labor saving amongst small cable logging teams.

Keywords: cable logging, workload, heart rate, synthetic rope

1. Introduction – Uvod In Norway, topography and surface conditions are often challenging for ground based harvesting systems. Much of the mature timber on steep slopes is found in the coastal belt. Here, high growth rates in plantation forests (Picea spp.) established in the 1950s through 1970s have resulted in volumes in excess of 600 m3 ha–1 which need to be harvested in coming years. Amongst forest stands mature for harvesting, around 28% (1 213 000 ha) is on slopes exceeding a gradient of 33% (classified as steep slopes), of which 43% (524 000 ha) exceed slopes of 50% (Larsson and Hylen 2007). The challenge of increasing the proportion of the annual cut from steep slopes can partly be met by increasing cable yarding activity. Given a current production of under 30 000 m3 per year per unit, there are opportunities for tens of new cable yarding crews to enter the sector over the next 20 – 30 years.

However, the shortage of labor willing to work in cable yarding operations is considered an obstacle to further expansion. The present situation is dependent on migratory labor that has no initial skills in mounCroat. j. for. eng. 32(2011)2

tain logging, and is at high risk of being lost to other sectors after becoming established in the Norwegian labor market. The retention of such workers is a priority in the industry. Ergonomic benefits reducing work strain, danger or discomfort of working with cable yarders would likely benefit the industry through improved recruitment, productivity and retention.

Depending on the actual function being carried out, forestry workers in steep terrain work continuously at or over the endurance limit (Stampfer 1997).

One infrequent but strenuous activity is rigging the yarder into a new corridor. For downhill yarding, which is most common in Norway, a strawline is manually pulled up the corridor (c. 3 – 400 m) connected through an end block and used to winch successively heavier lines up. Walking directly up the slope is one of the more strenuous activities associated with cable logging and classified as moderately heavy to heavy work (Vik 1992, Kirk and Sullman 2001, Stampfer et al. 2010).

The use of synthetic ropes (ultra-high molecular weight polyethylene) has become popular in a range of logging applications in North America (Pilkerton et al. 2004, Hartter and Garland 2006, Garland et al.

561

G. Ottaviani et al. Workload Benefits of Using a Synthetic Rope Strawline in Cable Yarder Rigging in Norway (561–569)

2003), and has also been adapted and tested in specific applications in Europe. Talbot (2007) found that in winching large hardwood logs, a 14 mm synthetic rope increased the working radius for a single operator and tractor setup from 15 m to 80 m as against using a 12 mm steel rope, while Stampfer et. al (2010a) showed that synthetic guylines reduced the workload from 2 to 1 person, and still reduced the heart rate for the single person by 30%. Depending on the application and dimensions required, synthetic ropes can offer a mass saving of 70 – 85% as against conventional steel cables of equivalent load capacity, while abrasion resistant covers and the absence of elastic energy ensures their technical integrity (Kirth et al.

2007, Stampfer et al. 2010b). The purchase price of synthetic ropes is 3 to 4 times that of steel ropes, while durability is application dependent and not well documented in the literature.

The goal of this study was to quantify the work strain saving that could be achieved in switching from a 3.5 mm steel wire strawline to a lightweight synthetic rope.

2. Materials and methods – Materijal i metode 2.1 Corridor – Ispitna trasa A 300 m (59% slope) corridor on an existing harvesting site in central Norway was divided into 12 successive segments of 25 m. Each segment profile was measured separately using a Vertex IV hypsometer and rangefinder. There was a plateau in the profile at 75 m – 100 m (40%) and 100 – 120 m (20%), otherwise the slope was relatively consistent though increasing with increasing distance – i.e. concave form (Fig. 1). The profile data was used in determining the altitude (above starting point) of each of the segments. Walking conditions on the slope were generally firm with only limited interference from harvesting slash or loose rocks.

2.2 Treatments and measurements – Postupci i mjerenja The experiment aimed to quantify the workload saving of switching from the steel strawline to the synthetic strawline. This was done by monitoring the force that the subject was exposed to while walking up the slope, as well as monitoring the heart rate.

Three treatments were defined; (i) pulling out the 3.5 mm steel wire strawline weighing 39 g m–1 (STEEL, W) (ii) pulling out the 4.0 mm synthetic strawline with a braided cover weighing 11g m–1 (SYNTH, X) and (iii) walking the profile with no load (ZERO, Z).

The ZERO treatment was designed to be the control.

562

Fig. 1 Profile of the 300 m corridor (slope length) Slika 1. Profil ispitne trase u duljini 300 m (duljina nagiba) Table 1 Study layout Tablica 1. Postavka istra`ivanja Replication Block Blok ponavljanja

Subject Radnik

Sequence and treatment code (rep-subject-treatment) Slijed i {ifra postupka (postupak ponavljanja radnika) 1 2 3

A

1AX

1AZ

1AW

B

1BZ

1BX

1BW

A

2AW

2AZ

2AX

B

2BZ

2BX

2BW

A

3AW

3AX

3AZ

B

3BZ

3BX

3BW

1

2

3

Time of day Razdoblje dana Morning Jutro Morning Jutro Afternoon Popodne Afternoon Popodne Morning Jutro Morning Jutro

The trial was designed in a randomized block with each subject drawing a random sequence of the 3 treatments within each of the 3 replication blocks (Table 1). Subjects were alternated and had approximately 1 hour rest between treatments. The 3 replication blocks were carried out in the morning, the afternoon, and the subsequent morning.

The tensile force that the subject pulled (constituting the mass of the rope and friction on the ground Croat. j. for. eng. 32(2011)2

Workload Benefits of Using a Synthetic Rope Strawline in Cable Yarder Rigging in Norway (561–569)

G. Ottaviani et al.

Table 2 Details of the two subjects included in the study Tablica 2. Detaljni opis dvaju radnika u istra`ivanju Subject Radnik A B

Age Dob year godina 26 37

Height Visina

Mass Masa

cm

kg

181 186

88 110

Initial heart rate HRr Po~etni puls srca HRr bpm otkucaji/min 93 (13) 75 (8)

Body mass index Indeks tjelesne mase – 26,9 31,8

Fig. 2 Illustration of the belt, dynamometer (A) and steel strawline being pulled out (B) Slika 2. Prikaz remena, dinamometra (A) te izvla~enja ~eli~noga pomo}noga u`eta (B) and in the winch drum) was measured continuously using a 3.5 kN AEP dynamometer fitted to a belt and equipped with a wireless transmitter (Fig. 2). The tensile force (N) combined with the distance (m) and height increment (m) was used to calculate the amount of work done (J) and the rate of work (W) in pulling the rope between segment endpoints and for the entire profile. The calculation of the rate of work done included lifting the subject’s own body mass through 143 vertical meters. The rate of work (W) was calculated by m (kg) s.(m2) t (s–3), where m is vertical meters and s is time in seconds.

2.3 Physiological measurements – Fiziolo{ka mjerenja Two male subjects (A and B) from a 3 man logging team participated in the study. Both had been working in steep terrain for more than 2 years, were considered well experienced in yarder rigging, and accustomed to hard physical work despite their relatively high BMI. Heart rate was measured to assess the level of physical stress of each treatment using a Polar RS400 pulse monitor with continuous data logging and storage. Heart rate measurement inCroat. j. for. eng. 32(2011)2

cluded the downhill return leg, where recovery time was also monitored. Resting heart rate (HRr) was not recorded as the no-load (zero) treatment was to be used as the benchmark for comparison. However, HRr was estimated from the initial heart rate readings (first 5 s) from each treatment (Table 2). Percentage of heart rate reserve (%HRR), a measure of exercise intensity, was calculated from the mean working heart rate (HRw), the proxy resting heart rate (HRr) and the maximum heart rate (HRmax) for each subject as: %HRR = (HRy-HRr)·100/(HRmax-HRr). HRmax was approximated from the rule stating HRmax = 220-subject age.

Walking speed was monitored, but not regulated, for each of the 12 segments in the profile using SIWORK 3 time study software running on an Allegro data logger.

3. Results – Rezultati 3.1 Force and work done – Sila i izvr{eni rad The force (N) that the subjects were exposed to increased linearly for both treatments, and was made up of a mass and a friction component. At the full

563

G. Ottaviani et al. Workload Benefits of Using a Synthetic Rope Strawline in Cable Yarder Rigging in Norway (561â&#x20AC;&#x201C;569)

Fig. 3a The tensile force exerted on the subject for the steel wire and synthetic rope, respectively. The resistance in the drum of the steel cable was higher than for the synthetic rope at 0 m despite the linear regression indicating otherwise Slika 3a. Primijenjena sila radnika za izvla~enje ~eli~noga i sinteti~koga u`eta. Otpor bubnja kod ~eli~noga u`eta ve}i je od otpora kod sinteti~koga u`eta, iako linearna regresija pokazuje obrnuto

Fig. 3b The cumulative work done in pulling out the steel and synthetic ropes, net of body mass Slika 3b. Kumulativne vrijednosti obavljenoga rada pri izvla~enju ~eli~noga i sinteti~koga u`eta bez mase tijela

Fig. 4 Mean rate of work (W) by 25 m segment for subject A (left) and subject B (right) for each treatment Slika 4. Srednje vrijednosti snage rada po segmentima trase od 25 m za radnika A (lijevo) i radnika B (desno) 564

Croat. j. for. eng. 32(2011)2

Workload Benefits of Using a Synthetic Rope Strawline in Cable Yarder Rigging in Norway (561–569)

G. Ottaviani et al.

Table 3 Summary data on subject’s heart rate response to treatments Tablica 3. Zbirni podaci o pulsu srca radnika ovisno o postupku

Subject – Radnik Steel rope – ^eli~no u`e, W Synthetic rope – Sinteti~ko u`e, X Zero – Bez optere}enja, Z

Weighted mean working heart rate Vrednovani srednji radni puls srca HRw bpm – otkucaji/min A B 170 162 165 157 166 148

300 m, the subjects were exposed to a mean force of approximately 140 N with the steel wire and 40 N with the synthetic rope (a). On average, subjects were exposed to higher force from the steel wire for 80 seconds longer than the synthetic rope. Fig. 3a shows smoothed forces, averaged over the 25m intervals and does therefore not include all force spikes seen in the continuous measurement. The cumulative work done (due to treatments only) was 21 kJ for the steel wire and 8 kJ for the synthetic rope (Fig. 3b).

In considering the rate at which work was done, both the treatment force and the subject’s own body mass play a role. The lighter subject (A) worked at a mean rate of 236 W, with a minimum of 100 W and a maximum of 402 W for treatment zero (Fig. 4a). The heavier subject (B) had a mean output of 277 W with a minimum of 93 W and a maximum of 472 W (b).

For both subjects and all treatments, a clear decrease in work output is seen in segments 4 and 5 (75 – 100 m and 100 – 125 m), which correlates to the plateau in the terrain profile.

The working pace was not strictly regulated, and from the heart rate measurements, subject A worked at an intensity of over 70% HRR for each treatment, though for differing lengths of time (Table 3). The higher working pace resulted in subject A rapidly ascending to 140 bpm and following the same trajectory for all 3 treatments (Fig. 5). Subject B, who worked at a slower pace, showed more distinct heart rate differences between treatments. Heart rate recovered more rapidly and uniformly for subject B than subject A, although the route of the descent was not monitored.

4. Discussion – Rasprava Despite the 25 m intervals set up in order to regulate the working pace, this was not done carefully enough during the trial and subjects had a tendency to work toward maximum capacity, irrespective of the treatment they were subjected to. This blurred the Croat. j. for. eng. 32(2011)2

Duration Trajanje t s A 589 529 497

B 612 548 542

Percentage heart rate reserve Postotak pri~uve pulsa srca % HRR % A B 76 80 71 75 72 67

expected result in heart rate differences and therewith the inference of the study. Due to inter-person variability, the rule of HRmax=220-age appears not to be specific enough for robust scientific analysis, and HRmax ought to be determined in laboratory conditions (Robergs and Landwehr 2002). HRr was estimated using the initial heart rate as a proxy and is therefore not the true resting heart rate. In this study, subject A appeared to have an unnaturally high initial heart rate for his age despite the extra care taken to ensure total rest before each treatment. Also, his recovery rate was slower and more erratic than for subject B. The dynamometer tests showed a clear difference in force required to overcome the mass and drag of the respective ropes (100 N at 300 m). The work done in drawing out the steel wire was 21 kJ as against approximately 8 kJ for the synthetic rope. In an attempt to standardize the evaluation of work by compensating for different walking speeds, we calculated the rate at which it was done (W) for the treatment only, and for the treatment including the subject’s own mass. Considering the ergonomic theory that a manual worker can sustain approximately 100 W throughout a working day, with shorter spikes of up to 400 W (Witney 1988), our data seemed to fit well. The mean power output of roughly 250 W during that time equates to strenuous work and this is corroborated by the heart rate data. A similar study cites a categorization of exertion as being 'Very heavy work' for a heart rate of 131 – 150, and 'Extremely heavy work' for a heart rate of 151 – >170 for 20 – 30 year olds (Pilkerton et al. 2004). In this study, subjects worked in latter zone for much of the time. In reality, a logging crew member would likely adjust the intensity of the task to his own capacity – a tendency known as constant strain behavior – especially as this is an infrequent task carried out only a few times per week and the speed of performance would have only a limited effect on productivity. Future studies ought to be carefully planned to ensure that working pace is

565

G. Ottaviani et al. Workload Benefits of Using a Synthetic Rope Strawline in Cable Yarder Rigging in Norway (561â&#x20AC;&#x201C;569)

Fig. 5. Heart rate monitoring of subject A (upper) and subject B (lower), respectively Slika 5. Izmjereni puls srca za radnika A (gore) i radnika B (dolje) kept constant if heart rate is to be used as an effective proxy for physiological workload.

Irrespective of the rate of work, we found that when pulling the steel wire, the load on the subject towards the end of the corridor (140 N) was the maximum load that a single person could pull on the given slope, and that completion of the final segment was exceedingly strenuous. This implies that

566

300 m is the maximum operational rigging length for a single person using the steel wire. Following the force gradient exerted by the synthetic rope in , such a barrier (140 N) would only be reached at 1200 m, which is three to four times normal yarding distance.

Although the mass of the fully extended steel wire was only 11.7 kg, it is the resultant force vector (on a 59% slope) that contributes substantially to the effort Croat. j. for. eng. 32(2011)2

Workload Benefits of Using a Synthetic Rope Strawline in Cable Yarder Rigging in Norway (561–569)

required to overcome it. The resultant force is not transferred downwards through the muscular-skeletal structure, but backwards and downwards, and therefore requires extra compensatory effort by the person. It should also be pointed out that in this trial, no rigging equipment (end block, guy-lines, tensioners, etc.) were carried. These would normally load the subject with an additional 15 kg or more, making them more sensitive to differences in the forces they are subjected to.

Rolling resistance in the winch drum was not isolated in the analysis. Assuming 800 m of line on each drum, the initial mass of steel wire on the drum would be 31.2 kg, and for the synthetic rope only 9.5 kg. The mass of the drum itself and the resistance would add to this. The forces measured in the first segment (25 m) could be used as a proxy measure as the rope mass on the ground is relatively insignificant. The difference here is 17 N for the steel wire and 10.8 N for the synthetic rope on average. Differences in drum rolling resistance would decrease with distance.

Using a wireless transmitter on the dynamometer with a receiving computer introduced some difficulties as the full 300m range was not covered and the person had to be followed in the terrain. A smaller datalogger attached to the dynamometer and carried by the subject would likely have provided a more effective and robust solution.

Some yarder operators use a different approach to laying out the strawline. A backpack mounted drum allows the person to walk up the slope along a more natural path, connect through the end block and walk directly down the corridor. However, this method transfers the full mass of a 6 – 800 m strawline to the subject from the beginning – a shift to synthetic rope would imply a large saving in mass to be carried.

Acknowledgements – Zahvala The authors wish to acknowledge the financial support from the Norwegian Research Council (grant No. 186912) and the Norwegian Forestry Development Fund (Skogtiltaksfondet) as well as assistance from colleagues Eirik Nordhagen, Helmer Belbo and Daniel Kindernay. We also express our thanks to forest entrepreneur Frivik A/S, whose logging team and equipment participated in the study.

G. Ottaviani et al.

5. References – Literatura Garland, J. J., Sessions, J., Pilkerton, S. J., Hartter, J., 2003: Synthetic Rope to Replace Wire Rope in Mountain Logging Operations. Austro2003: High Tech Operations for Mountainous Terrain Schlaegl, Austria.

Hartter, J., Garland, J. J., 2006: Synthetic Rope End Connections for Use in Timber Harvesting. International Journal of Forest Engineering 17(1): 39–51.

Kirk, P. M., Sullman, M. J. M., 2001: Heart rate strain in cable hauler choker setters in New Zealand logging operations. Applied Ergonomics 32(4): 389–398.

Kirth, R., Schiemer, S., Nemestothy, N., Sperrer, S., 2007: Further Developments of Synthetic Ropes for Logging Applications in Forestry. Austro2007/FORMEC’07: Meeting the Needs of Tomorrow’s Forests – New Developments in Forest Engineering. Vienna and Heiligenkreuz, Austria, FORMEC. Larsson, J. Y., Hylen, G., 2007: Statistics of Forest Conditions and Forest Resources in Norway. Viten fra Skog og landskap. Ås, Norwegian Forest and Landscape Institute.

Pilkerton, S. J., Garland, J. J., Hartter, J., 2004: Applications of synthetic rope for improved ergonomic, economic and environmental performance in mountainous logging. 2004 International Mountain Logging Conference. Vancouver, BC, Canada.

Robergs, R., Landwehr, R., 2002: The Surprising History of the »HRmax=220-age« Equation. Journal of Exercise Physiology 5(2): 1–10.

Stampfer, K., 1997: Stress and Strain Effects of Forest Work in Steep Terrain. In Heinimann, H.R. & Sessions, J. (Eds.) Forest Operations in Himalayan Forests with Special Consideration of Ergonomic and Socio-Economic Problems.

Thimphu, Bhutan, IUFRO 3.06.00 Stampfer, K., Karpf, F., Visser, R., 2010a: Efficiency and Ergonomic Advantages of Synthetic Rope for Guying Cable Yarders. In Belbo, H. (Ed.) Forest Operations Research in the Nordic Baltic Region. Honne, Norway. Norwegian Forest and Landscape Institute.

Stampfer, K., Leitner, T., Visser, R., 2010b: Efficiency and Ergonomic Benefits of Using Radio Controlled Chokers in Cable Yarding. Croatian Journal of Forest Engineering 31(1): 1–9.

Talbot, B., 2007: Reduced Impact Logging in Certified Hardwood Stands. In Gingras, J-F. (Ed.) 3rd World Forest Engineering Conference. Mont-Tremblant, Quebec, Canada, FP-Innovations.

Vik, T., 1992: Work Science research at NISK (In Norwegian with English summary). Norwegian Institute of Forest Research – 75 years (Skogforsk).

Witney, B., 1988: Choosing and using farm machines, London, Longman. 412 pp.

Croat. j. for. eng. 32(2011)2

567

G. Ottaviani et al. Workload Benefits of Using a Synthetic Rope Strawline in Cable Yarder Rigging in Norway (561â&#x20AC;&#x201C;569)

Sa`etak

Olak{anje radnoga optere}enja primjenom sinteti~koga pomo}noga u`eta za postavljanje trase `i~are u Norve{koj Cilj je rada odrediti razliku u radnom optere}enju koja nastaje zamjenom ~eli~noga u`eta promjera 3,5 mm sinteti~kim u`etom ve}ega promjera 4,0 mm prilikom izvla~enja pomo}noga u`eta za postavljanje nove `i~ne linije na udaljenosti od 300 m.

U Norve{koj su reljefne zna~ajke terena ~esto nepovoljne za pridobivanje drva s kretnim sustavima po tlu.

Ve}ina se zrelih sastojina nalazi na strmim terenima u obalnom pojasu. Velik prirast u {umskim planta`ama smreke (Picea sp.), koje su osnivane od 1950-ih do 1970-ih, rezultirao je visokom drvnom zalihom, preko 600 m3/h, koju treba posje}i idu}ih godina. Me|u sastojinama koje dolaze na red za sje~u oko 28 % (1 213 000 ha) nalazi se na nagibu ve}em od 33 % (ozna~en kao strmi teren), od ~ega se 43 % (524 000 ha) nalazi na nagibu ve}em od 50 % (Larsson i Hylen 2007). Pove}enje godi{njega sje~ivoga etata mo`e se donekle ostvariti pove}anjem upotrebe {umskih `i~ara. S godi{njom proizvodno{}u `i~are od 30 000 m3 drva u idu}ih 20 do 30 godina otvara se mogu}nost upotrebe na desetke novih {umskih `i~ara.

Ipak, smanjenje broja zainteresiranih za rad na {umskim `i~arama ograni~avaju}i je ~imbenik. Trenuta~na je situacija ovisna o sezonskoj radnoj snazi koja nema dovoljno iskustva u pridobivanju drva na strmim terenima te ~esto odlazi na druge poslove koji se otvaraju na norve{kom tr`i{tu rada. Zadr`avanje je takve radne snage prioritet u {umarstvu. Pobolj{anje ergonomskih uvjeta, smanjenje radnoga optere}enja, opasnosti i neudobnosti rada sa {umskim `i~arama vjerojatno bi pove}alo zapo{ljavanje, proizvodnost i zadr`avanje {umskih radnika.

Jedan od rje|ih, ali napornijih poslova jest postavljanje {umske `i~are na novu trasu. U Norve{koj se naj~e{}e {umskim `i~arama iznose drva niz nagib. Pri tome se pomo}no u`e, koje se koristi za podizanje te`ih u`adi pri postavljanu `i~are, mora ru~no vu}i uz nagib (300 do 400 m). Kretanje okomito na slojnice terena jedan je od najnapornijih aktivnosti u radu sa `i~arom. Upotreba sinteti~ke u`adi (polietilenska u`ad vrlo visoke molekularne te`ine) za pridobivanje drva postaje sve ~e{}a u SAD-u (Pilkerton i dr. 2004, Hartter i Garland 2006, Garland i dr.

2003) te je tako|er ispitivana na pojedinim radovima u Europi. Ovisno o vrsti rada za koji je namijenjena i dimenziji u`eta, sinteti~ka u`ad ima 70 â&#x20AC;&#x201C; 85 % manju masu od konvencionalne ~eli~ne u`adi jednake nosivosti, dok otpornost na o{te}enja i elasti~nost jam~i tehni~ku ispravnost (Kirth i dr. 2007, Stampfer i dr. 2010b). Nabavna cijena sinteti~koga u`eta ve}a je 3 do 4 puta nego nabavna cijena ~eli~noga u`eta, dok je izdr`ljivost ovisna o vrsti posla za koji se u`e koristi i nedovoljno je opisana u literaturi.

Istra`ivanje je provedeno u sredi{njoj Norve{koj na ve} postoje}oj sje~ini. Trasa `i~are duljine 300 m (nagib 59 %) podijeljena je na 12 uzastopnih dijelova duljine 25 m. Profil svakoga dijela zasebno je izmjeren pomo} visinomjera i daljinomjera Vertex IV. Nagib je na svim dijelovima podjednak, raste s pove}anjem udaljenosti, konkavnoga je oblika. Podaci su profila kori{teni za odre|ivanje visine (od po~etne to~ke) svakoga pojedinoga dijela. Dvojica od trojice radnika (A i B) na `i~ari sudjelovala su u istra`ivanju. Obojica imaju dvogodi{nje iskustvo u radu na strmim terenima i smatraju se dovoljno iskusnim u postavljanju `i~are, te su naviknuti na te`ak fizi~ki posao unato~ svojemu visokomu indeksu tjelesne mase (BMI). Za svaki im je radni zahvat, pomo}u ure|aja Polar RS400, mjeren broj otkucaja srca kako bi se odredio fizi~ki napor kojemu su izlo`eni.

Smanjenje radnoga optere}enja, prilikom zamjene ~eli~noga u`eta sinteti~kim, odre|ivano je mjerenjem sile kojom su radnici bili izlo`eni tijekom hoda uz nagib i mjerenjem otkucaja srca. Fizi~ko optere}enje mjereno je u tri postupka: 1) izvla~enje ~eli~noga u`eta promjera 3,5 mm, te`ine 39 g/m (STEEL, W); 2) izvla~enje sinteti~koga u`eta promjera 4,0 mm, te`ine 11 g/m (SYNTH, X) i 3) hod uz nagib bez izvla~enja u`eta (ZERO, Z). ZERO na~in rada postavljen je kao kontrolni uzorak.

Sila izvla~enja u`eta (~ije su sastavnice masa u`eta, trenje izme|u tla i u`eta te trenje na bubnju vitla) mjerena je pomo}u dinamometra 3,5 kN AEP, opremljenoga be`i~nim oda{ilja~em, koji je bio pri~vr{}en na remen {umskoga radnika. Kombinacijom sile izvla~enja u`eta (N), udaljenosti (m) i pove}anja visine (m) izra~unata je koli~ina utro{ene energije (J) i izvr{ena snaga rada (W) pri izvla~enju u`eta izme|u krajnjih to~aka pojedinoga dijela i du` cijeloga profila. Pri ra~unanju izvr{ene snage rada u izra~un je uklju~ena i masa samoga radnika dok je radnik prelazio 143 metra visinske razlike. Izvr{ena snaga rada izra~unata je kao m (kg) s.(m2) t (s-3), gdje je visinska razlika u metrima, a vrijeme u sekundama.

568

Croat. j. for. eng. 32(2011)2

Workload Benefits of Using a Synthetic Rope Strawline in Cable Yarder Rigging in Norway (561–569)

G. Ottaviani et al.

Sila (N), koja se sastojala od mase u`eta i sastavnica trenja, kojom je radnik bio izlo`en u oba slu~aja se linearno pove}avala. Na maksimalnoj udaljenosti od 300 m radnik je bio optere}en silom srednje vrijednosti 140 N prilikom izvla~enja ~eli~noga u`eta, a 40 N prilikom izvla~enja sinteti~koga u`eta. Radnik je bio izlo`en ve}emu optere}enju, u prosjeku 80 sekundi, prilikom izvla~enja ~eli~noga u`eta. Slika 3(a) prikazuje prosje~ne vrijednosti sile na dijelu od 25 m te zbog toga ne prikazuje najve}a optere}enja koja su se pojavljivala tijekom mjerenja. Utro{ak energije bio je 21 kJ pri radu s ~eli~nim u`etom i 8 kJ pri radu sa sinteti~kim u`etom (slika 3b). Uzimaju}i u obzir snagu rada, u oba slu~aja i sila i radnikova tjelesna masa imaju utjecaj na optere}enje. Radnik (A) s manjom tjelesnom masom radio je srednjom vrijednosti snage 236 W (najmanja vrijednost 100 W i najve}a 402 W) prilikom rada bez optere}enja (slika 4a). Radnik (B) s ve}om tjelesnom masom radio je sa srednjim vrijednostima snage od 277 W (najmanja vrijednost 93 W i najve}a 427 W). Kod oba radnika u svim slu~ajevima, u segmentima 4 i 5 (75 – 100 m i 100 – 125 m udaljenosti od po~etne to~ke), vidljivo je smanjenje snage izvr{enoga rada, {to je povezano sa zaravni u profilu terena.

Mjerenja sile pokazuju o~itu razliku potrebnu za svladavanje mase i izvla~enja u`eta (100 N na 300 m).

Utro{ak je energije prilikom izvla~enja ~eli~noga u`eta 21 kJ nasuprot 8 kJ prilikom izvla~enja sinteti~koga u`eta.

U poku{aju standardiziranja procjene obavljenoga rada kompenzacijom za razli~ite brzine kretanja izra~unali smo utro{enu snagu za istra`ivane slu~ajeve uklju~uju}i i tjelesnu masu radnika. Srednja vrijednost utro{ene snage od 250 W s vremenom se izjedna~ava, {to potvr|uju i otkucaji srca. Bez obzira na intenzitet posla, zaklju~ujemo da je prilikom izvla~enja ~eli~noga u`eta optere}enje radnika (od 140 N) pri kraju trase `i~are bilo najve}e optere}enje koje jedna osoba mo`e podnijeti na takvu nagibu. Zavr{etak je zadnjega dijela u profilu bio osobit napor, {to pokazuje da je 300 m maksimalna operativna udaljenost prilikom koje jedna osoba postavlja `i~aru koriste}i ~eli~no u`e.

Provedeno je istra`ivanje va`no zbog uo~avanja ergonomskih koristi pri kori{tenju lak{e sinteti~ke pomo}ne u`adi u usporedbi s uobi~ajenom ~eli~nom u`adi. Budu}a bi se istra`ivanja trebala usmjeriti na smanjenje radnoga optere}enja i ergonomska pobolj{anja, {to bi zna~ajno pridonijelo pove}anju pridobivanja drva na strmim terenima u Norve{koj.

Klju~ne rije~i: izno{enje drva `i~arama, radno optere}enje, puls srca, sinteti~ko u`e

Authors’ addresses – Adrese autorâ: Giovanna Ottaviani, PhD. student* e-mail: gio@skogoglandskap.no Bruce Talbot, PhD. e-mail: bta@skogoglandskap.no Morten Nitteberg, Scientific Engineer e-mail: nim@skogoglandskap.no Norwegian Institute for Forest and Landscape Section for Forest Technology and Economics Høgskoleveien 8, 4132 Ås NORWAY

Received (Primljeno): July 26, 2011 Accepted (Prihva}eno): September 5, 2011 Croat. j. for. eng. 32(2011)2

Assoc. Prof. Karl Stampfer, PhD. karl.stampfer@boku.ac.at University of Natural Resources and Applied Life Sciences Vienna Institute of Forest Engineering Peter Jordan Strasse 82 1190 Vienna AUSTRIA * Corresponding author – Glavni autor

569

Original scientific paper – Izvorni znanstveni rad

Discussion on Economic and Energy Balances of Forest Biomass Utilization for Small-Scale Power Generation in Kanuma, Tochigi Prefecture, Japan Kazuhiro Aruga, Ayami Murakami, Chikara Nakahata, Reiko Yamaguchi, Takuyuki Yoshioka Abstract – Nacrtak In this study, the economic and energy balances of forest biomass utilization for small-scale power generation are discussed, considering the spatial distribution of the forest biomass resources using the geographic information system (GIS) in the Kanuma area of Tochigi Prefecture, Japan. First, the optimum scales of two power-generation plants are discussed. For a direct combustion power-generation plant operating at an optimum scale of 5 MW generation capacity, the electricity cost would be 23.7 yen/kWh. For a small-scale gasification power plant operating at an optimal scale of 2.4 MW generation capacity, the electricity cost would be 12.8 yen/kWh. As the average electricity price in Japan is 22.2 yen/kWh, the electricity generated from the small-scale gasification power-generation plant could be economical. The energy balance and CO2 emissions from the energy utilization of forest biomass resources were analyzed using the life cycle inventory (LCI) method. For both types of power generation, the ratio of energy output to input was calculated to be about 20, indicating that the system examined in this study could be feasible as an energy production system. The CO2 emission from the direct combustion power generation with a generation capacity of 5 MW was 754.9 tCO2/year, while the CO2 emission of the small-scale gasification power plant with a generation capacity of 2.4 MW was 381.9 tCO2/year. However, the reductions in the amount of CO2 emission that would result from replacing fossil fuel were 15,707 tCO2/year and 6,275 tCO2/year, respectively.

1 = 114 yen on June 27, 2011.

Keywords: economy balance, energy balance, CO2 balance, forest biomass resources, small-scale power generation, GIS, LCI

1. Introduction – Uvod Forests play an important role in the carbon balance of the planet by drawing carbon from the atmosphere and producing wood, a renewable resource that stores the removed carbon. Therefore, forests need to be continuously and properly managed and the wood utilized at all levels – from building materials, furniture, board and paper, to chemical products and fuel. Woody biomass can be categorized into forest residues (referred to as forest biomass resources in this study), sawmill residues, and conCroat. j. for. eng. 32(2011)2

struction waste timber. The recent years have witnessed a steady increase in the introduction of wood-fired boilers and generators and the production of wood pellets in Japan. However, a large amount of woody biomass, particularly forest residues, still remains unused (Forestry Agency 2009).

In order to utilize forest biomass resources for bio-energy, it is crucial to determine the relationship between the annual available amounts and the procurement (harvesting and transporting) costs of forest biomass resources. Ranta (2005) and Yoshioka and Sakai (2005) carried out detailed analyses of the po-

571

K. Aruga et al.

Discussion on Economic and Energy Balances of Forest Biomass Utilization ... (571â&#x20AC;&#x201C;586)

tential supply of forest biomass using a geographic information system (GIS). Nord-Larsen and Talbot (2004) and Aruga et al. (2006a) discussed the longterm feasibility of timber and forest biomass resources by predicting future forest resources using growth models while optimizing the allocation of fuelwood using linear programming or random search. Yagi and Nakata (2006), Aruga et al. (2006b), and Panichelli and Gnansounou (2008) discussed the scales and locations of bio-energy facilities based on the relationship between the annual available amounts and the procurement costs of forest biomass resources.

In addition to the economic balances discussed in these studies, many other studies have discussed energy balances (Faaij et al. 1997, Forsberg 2000). However, these studies have focused on large-scale power plants whose net power outputs ranged from tens to hundreds of megawatts for the energy conversion of biomass. Therefore, Yoshioka et al. (2005) conducted analyses of energy and CO2 balances for a small-scale energy-conversion system that used forest biomass resources from conventional Japanese

forestry as fuel. However, the study did not consider the spatial distribution of the forest biomass resources. Therefore, the present study discusses the economy and energy balances of forest biomass utilization for small-scale power generation considering the spatial distribution of forest biomass resources using GIS in the Kanuma Area of Tochigi Prefecture, Japan.

2. Materials and Methods â&#x20AC;&#x201C; Materijal i metode Forest biomass resources can be categorized into logging residues, thinned wood, and broad leaved trees (Yoshioka and Sakai 2005). Forest resources, the slope of the land, and public and forest road layers of GIS were obtained from the Tochigi Prefectural Government in order to estimate the harvesting volumes and costs of timber and forest biomass resources. Future forest resources in each stand were predicted using the system yield table, Local Yield Table Construction System (LYCS, Shiraishi 1985). Then, the

Fig.1 Study site Slika 1. Podru~je istra`ivanja 572

Croat. j. for. eng. 32(2011)2

Discussion on Economic and Energy Balances of Forest Biomass Utilization ... (571â&#x20AC;&#x201C;586)

stand harvesting schedules were planned by balancing harvesting volumes of timber and forest biomass resources using random search while minimizing procurement costs. First, the optimum scales of a direct combustion power plant and a small-scale gasification power plant were discussed with the viewpoint of economic balances. Then, the energy balance and the CO2 emission from the energy utilizations of the forest biomass resources were analyzed using the life cycle inventory (LCI) for the power-generation plants with the optimum generation capacity determined by the analyses of the economical balances.

K. Aruga et al.

2.1 Study site and Data â&#x20AC;&#x201C; Podru~je istra`ivanja i podaci The study site was the Kanuma area, consisting of Kanuma city and Nishikata town (Fig. 1).

This area encompasses 52,000 hectares, of which about 65% is forested (Fig. 2). Most of forests are man-made forests (79%) planted with conifers; Japanese cedar (Sugi) and Japanese cypress (Hinoki) account for 54% and 23% of the trees, respectively.

Most of the conifers are about 45 â&#x20AC;&#x201C; 50 years old. The Kanuma area was one of the famous forestry areas in Tochigi Prefecture. However, the forestry industry

Fig.2 Tree species in the study area Slika 2. Vrste drve}a na podru~ju istra`ivanja Croat. j. for. eng. 32(2011)2

573

K. Aruga et al.

Discussion on Economic and Energy Balances of Forest Biomass Utilization ... (571–586)

Table 1 Operation patterns in sub-compartments to be felled Tablica 1. Primjeri postupaka u pododjelima za sje~u Forest – [uma

Age, year Dob, godina

31–60 Man-made and coniferous Umjetno podignute i ~etinja~e Over 61 Iznad 61

Naturally regenerated and broad-leaved Prirodno pomla|ene i lista~e 1 1

Over 31 Iznad 31

Operation pattern – Primjer postupaka [Forest biomass resources: Thinned trees] Thinning is carried out with a 20 – 50% thinning rate based on stand situations, and whole trees are used as energy sources1 [Resursi {umske biomase: stabla iz proreda] Prorede se izvode intenzitetom od 20 do 50 % na osnovi stanja sastojine i cijela se stabla koriste kao izvor energije1 [Forest biomass resources: Logging residues] Clearcutting is carried out on a 60-year cycle. Trees are limbed and bucked, logs are harvested, and tops and branches are used as energy sources [Resursi {umske biomase: ostaci pridobivanja drva] ^iste se sje~e provode u 60-godi{njim ciklusima. Stabla se kre{u i izra|uju, trupci se odvajaju, a vrhovi i grane se koriste kao izvori energije [Forest biomass resources: Broad-leaved forests] Clearcutting is carried out on a 30-year cycle, and whole trees are used as energy sources [Resursi {umske biomase: sastojine lista~a] ^iste se sje~e provode u 30-godi{njim ciklusima i cijela se stabla koriste kao izvor energije

In this study, we assume that all the cut material from thinning operations can be used as an energy source, considering the actual Japanese market value U ovom je radu pretpostavljeno da se sav materijal iz proreda mo`e koristiti kao izvor energije, uzimaju}i u obzir stvarne vrijednosti na japanskom tr`i{tu

has been in decline for a long time. Therefore, bio-energy is attracting a great deal of attention since energy utilization of forest biomass resources is expected to contribute to the revitalization of the forestry industry as well as the maintenance of the appropriate ecological, economic, and social functions of man-made forests. The site-index ranks the order of the production capacity of the stands into three classes: the smaller the number, the larger is the production capacity. Site-index 1 is 53%, site-index 2 is 43%, and site-index 3 is 4%. As for the operation-site inclination, most of the forests are in relatively steep terrain, sloping 30 degrees or more. The density of the road network in the Kanuma area is 18 m/ha.

Forest registration data (stand age, tree species, and site indexes) and GIS data (information on roads and sub-compartment layers) from the Tochigi Prefectural Government were used in this study, as well as 50-m-grid digital elevation models (DEM) from the Geographical Survey Institute. The data were converted into 50-m-grid raster data for consistency with the DEM data. Using these materials and the GIS, the annual available amounts of timber and forest biomass resources were estimated based on sub-compartments. The DEM was used to estimate the slope of each sub-compartment and to judge the skidding/yarding direction (uphill or downhill) for cost estimations. The analysis was conducted on the basis of sub-compartments, which are common operational units in Japan.

574

2.2 Procurement costs – Tro{kovi pridobivanja The harvesting and transporting systems for forest biomass resources were classified into two types (Fig. 3) depending on the resources used (logging residues or the whole tree). Table 1 lists the operation patterns of the sub-compartments to be felled. Logging residues are considered to be a by-product in conventional forestry. Therefore, the system boundary of logging residues begins with comminuting the logging residues at the landing of the logging site by a mobile chipper (Fig. 3).

In this study, it is assumed that tractors (cable skidders), swing yarders (backhoe with winch drums), tower yarders (mobile yarders), and yarders are used for the skidding/yarding process. Tractors can be used on slopes below 11 degrees for uphill travel and 19 degrees for downhill travel. Swing yarders and tower yarders can be used within 100 m and 300 m yarding distances, respectively. In this study, one type of machine from the four types mentioned here is assumed to be selected for each stand so that skidding/yarding costs are minimized within the topographic condition of each stand.

Table 2 shows the equations for estimating the efficiency, and Table 3 shows the equations for estimating the procurement costs of timber and forest biomass resources. In both sets of equations, the variables are the slope q [degree], average stem volume Vn [m3/stem], harvesting volumes per ha V [m3/ha], number of trees harvested per ha NF [stem/ha], skidCroat. j. for. eng. 32(2011)2

Discussion on Economic and Energy Balances of Forest Biomass Utilization ... (571–586)

K. Aruga et al.

Fig.3 Classification of systems according to wood material harvested Slika 3. Klasifikacija sustava prema pridobivanomu drvnomu materijalu Table 2 Machine specification Tablica 2. Zna~ajke strojeva Machine – Stroj

Chainsaw Motorna pila Tractor Traktor Swing yarder Okretna `i~ara Tower yarder Stupna `i~ara Yarder @i~ara Processor Procesor Chipper Ivera~ Truck Kamion

Efficiency, m3/hour Efikasnost, m3/sat

Remarks Napomena

Mass, kg Masa, kg

Lifetime, hour @ivotni vijek, sat

Conifer ^etinja~e

6

2,700

219Vn N F + 3, 000

2.8

6

2,700

2.0

2.8

6,000

6,480

5,440/LY

4.3

3,200

4,200

1,080/(2LY+80)

13.0

7,425

5,400

4,860/(2LY+243)

3.0

3,000

6,300

12LY–0.21

2.8

6,770

6,480

151Vn

2.0

7,802

5,000

13

28.0

7,960

5,500

247,422/LT

8.2

7,960

5,500

199,500/LT

8.2

Broadleaf Lista~e Whole tree Cijelo stablo Whole tree Cijelo stablo Whole tree Cijelo stablo Whole tree Cijelo stablo Whole tree Cijelo stablo Solid Kruto Log Trupac Chip Iver

Croat. j. for. eng. 32(2011)2

21600 , Vn N F

Fuel consumption, l/hour Potro{nja goriva, l/sat

575

K. Aruga et al.

Discussion on Economic and Energy Balances of Forest Biomass Utilization ... (571–586)

Table 3 Procurement cost, yen/m3 Tablica 3. Tro{kovi pridobivanja, jen/m3 Machine Stroj Chainsaw Motorna pila Tractor Traktor Swing yarder Okretna `i~ara Tower yarder Stupna `i~ara Yarder @i~ara Processor Procesor Chipper Ivera~ Truck Kamion

Remarks Napomena Conifer ^etinja~e Broadleaf Lista~e Whole tree Cijelo stablo Whole tree Cijelo stablo Whole tree Cijelo stablo Whole tree Cijelo stablo Whole tree Cijelo stablo Solid Kruto Log Trupac Chip Iver

Cost – Tro{ak 179 + 2,453/Vn N F 1,472 0.97LY + 27,510e0.12q/V + 1,771 22LY + 3,913,500/LYV + 919 5.4LY + 5,870,250/LYV + 747 991LY0.21 + 5,071,896/LYV + 161,236/V + 196 532/Vn 1,093 0.027LT + 778 0.033LT + 778

ding/yarding distance LY [m], and transporting distance LT [m]. In addition to the direct costs of labor, machinery, and fuel, the indirect costs of labor (55% of the direct cost of labor), machine moving costs (50,000 yen/each), and overhead costs (20% of the total direct costs) are also considered.

Finally, the following items on topography are processed by the GIS software. The average angle of inclination of each sub-compartment is calculated.

To determine the skidding/yarding distance of each sub-compartment, the distance between the center of gravity of the sub-compartment and the nearest road from the sub-compartment is calculated. A landing is to be arranged at a point on the nearest road from the sub-compartment. The skidding/yarding direction (uphill or downhill) is judged by comparing the altitude between the center of gravity and the landing. The transporting distance from the landing to the energy-conversion plant is calculated by the shortest path algorithm using the Dijkstra method (Dijkstra 1959). The energy-conversion plant is assumed to be located on the log market. By applying the topographical data for each sub-compartment to the equations listed in Table 3, the procurement costs of timber and forest biomass resources can be calculated.

2.3 Available amounts of timber and forest biomass resources – Raspolo`ive koli~ine drva i {umske biomase The available amounts of timber and forest biomass resources can be estimated from the stem volume recorded in the forest register and the coefficients listed in Table 4. Consequently, by applying Tables 1 and 4 to the forest register, the available amounts of timber and forest biomass resources for the first year can be estimated. To estimate the future available amounts of timber and forest biomass resources, the system yield table, LYCS (Shiraishi 1985)

Table 4 Methods for calculating the amount of forest biomass resources Tablica 4. Metode za izra~un koli~ine resursa za {umsku biomasu

1 1

Biomass resources Resursi biomase

Equation (s.v.: Stem volume) Jednad`ba (s.v.: obujam debla)

Logging residues1 [umski ostatak1

Amount (tDM) – Koli~ina (tST) = s.v. × 15/92 × 0.40

Thinned trees Stabla iz proreda

Amount (tDM) – Koli~ina (tST) = s.v. × t.r. × 100/92 × 0.40

Broad-leaved forests Sastojine lista~a

Amount (tDM) – Koli~ina (tST) = s.v. × 100/80 × 0.56

Notes Zabilje{ke 15/92: Ratio of top and branches’ volume to stem volume 0.40: Density of a coniferous tree 15/92: Odnos obujma ovr{aka i grana prema obujmu debla 0,40: Gusto}a ~etinja~a t.r.: Thinning rate, 20 – 50/100 100/92: Ratio of the whole tree’s volume to stem volume 0.40: Density of a coniferous tree t.r.: Intenzitet prorede, 20 – 50/100 100/92: Odnos obujma cijeloga stabla prema obujmu debla 0,40: Gusto}a ~etinja~a 100/80: Ratio of the whole tree’s volume to stem volume 0.56: Density of a broad-leaved tree 100/80: Odnos obujma cijeloga stabla prema obujmu debla 0,56: Gusto}a lista~a

The method for calculating the cut volume of logs in clearcutting is as follows: Volume of logs (m3) = s.v. × 85/92 (85/92: Ratio of logs’ volume to stem volume) Metoda za izra~un obujma trupaca u ~istim sje~ama je sljede}a: obujam trupaca (m3) = s.v. × 85/92 (85/92: odnos obujma trupaca prema obujmu debla)

576

Croat. j. for. eng. 32(2011)2

Discussion on Economic and Energy Balances of Forest Biomass Utilization ... (571–586)

is applied to the forest register. The time interval is five years. In order to allow the steady operation of the energy-conversion plant, the forest biomass resources should be provided to the plant on a continuous basis. In this study, stand harvesting schedules were planned for sixty years by balancing harvesting volumes of timber and forest biomass resources using random search while minimizing procurement costs (Aruga et al. 2006a).

2.4 Energy-conversion plants – Energetska postrojenja Two types of energy-conversion are considered in this study. One is direct combustion and the other is small-scale gasification. Small-scale gasification is a technology currently under development. Table 5 shows the basic specifications for the power-generation plants. The net power output and steam flow change in relation to the plant scale.

Yagi and Nakata (2007) reported that the energy-conversion efficiency E [%] of direct combustion and small-scale gasification could be expressed by the following equations as a function of the generation capacity G [kW]: Direct combustion power-generation: E = 5.35 × LN(G) – 24.59

(1)

K. Aruga et al.

Small-scale gasification power-generation: E = 3.14 × LN(G) + 5.10

(2)

The generation capacity of small-scale gasification was assumed to be within 3,000 kW. Fig. 4 shows the relationship between the generation capacity and the demand for fuels. The costs of the power-generation plant construction C [yen] are assumed to be proportional to the power of 0.7 of the net power output of the plant, C µ G0.7 because the quantity of materials required for a power-generation plant is reported to be proportional to the power of 0.7 of the net power output of the plant (Tahara et al. 1998). Half of the initial costs are assumed to be subsidized. Some of the generated power and steam is assumed to be utilized in the plant, and the surplus power and steam are assumed to be sold.

2.5 Energy balance – Energetska bilanca The energy input into the system consists of the equipment and operation energies over the entire lifetime of the plant (Yoshioka et al. 2005). Equipment energy is defined as the energy used for manufacturing the equipment that constitutes the system, that is, the forestry machinery and the power-generation plant in this study, and is composed of the »material«, »production«, »transport« and »construction« energies. On the other hand, operation energy

Table 5 Power-generation plant specifications Tablica 5. Zna~ajke energetskoga postrojenja Direct combustion Izravno izgaranje

Small-scale gasification Malo plinoficiranje

Fuel consumption, tDM/day – Potro{nja goriva, tST/dan

202

6.8

Operating rate, % – Stupanj kori{tenja, %

72

87

2,663

208

24

6

Item – Zna~ajka

Net power output, kW – Proizvodnja energije, neto, kW Steam flow, t/h – Protok pare, t/h Initial cost, million yen – Po~etni tro{kovi, milijun jena

1,464

68

Subsidy ratio, % – Iznos potpora, %

50

50

Repair and maintenance coefficient, % – Koeficijent popravaka i odr`avanja, %

5

5

Operators, n – Rukovatelji, n

7

1

Labor cost, yen/year/person – Tro{ak radne snage, jen/godina/osoba

7,000,000

7,000,000

Land tenancy cost, yen/year – Tro{ak zakupa zemlji{ta, jen/godina

8,400,000

0

Overhead rate to initial cost, % – Op}i tro{kovi prema po~etnomu tro{ku, %

5

5

Surplus power ratio, % – Stupanj vi{ka energije, %

79

75

Surplus steam ratio, % – Stupanj vi{ka pare, %

83

100

Power selling price, yen/kWh – Prodajna cijena energije, jen/kWh

8

8

Steam selling price, yen/kg – Prodajna cijena pare, jen/kg

0.5

0.5

Life time, years – @ivotni vijek, godine

30

30

Croat. j. for. eng. 32(2011)2

577

K. Aruga et al.

Discussion on Economic and Energy Balances of Forest Biomass Utilization ... (571â&#x20AC;&#x201C;586)

Fig.4 Generation capacity and fuel consumption Slika 4. Proizvodni kapacitet i potro{nja goriva is defined as the energy necessary for operating the system and is composed of the fuel consumption of the forestry machinery and the Âťrepair and maintenanceÂŤ energy of the power-generation plant.

The material energies are calculated using the weight of each kind of material used and the energy density of each material. Based on the analysis by Hondo et al. (2000), all the parts of each forestry machine are assumed to be made of steel in this study.

The weight of the required material for each machine is calculated from the mass, the lifetime of each machine, the productivity of each machine, and the annual amount of forest biomass resources required for the plant (Yoshioka et al. 2005). The required materials for a forest biomass power-generation plant are calculated with reference to a 1,000 MW coal-fired power-generation plant (Uchiyama and Yamamoto 1991). The energy density of steel is 4,709 MJ/t (500 kWh) electricity and 20,930 MJ/t coal. The energy density of aluminum is 164,826 MJ/t (17,500 kWh) electricity and 46,047 MJ/t oil. The energy density of concrete is 184 MJ/t (20 kWh) electricity, 435 MJ/t oil, and 255 MJ/t coal. According to Uchiyama and Yamamoto (1991), the sum of the production, transport, and construction energies is assumed to be equivalent to 20% of the total material energy.

The quantity of required fuel is calculated from the fuel consumption of each machine, the productivity of each machine, and the annual amount of forest biomass resources required for the plant. Gasoline (petrol) is used as fuel for chainsaws, and diesel fuel is used for all other machines. The energy densities of

578

gasoline and diesel fuel are 34.6 MJ/l and 38.2 MJ/l, respectively. The repair and maintenance energy of a power-generation plant is assumed to be equivalent to 5% of the equipment energy over the lifetime of the plant provided that the repair and maintenance of the plant is performed every year so that all parts of the plant may be updated in 20 years (Uchiyama and Yamamoto 1991).

The goal of this study is to investigate the following two environmental load profiles of the defined biomass procurement and bio-energy supply chain.

First, the energy balance factor (EBF) is the ratio of energy output to input, which is used to examine whether the system is feasible as an energy production system. Second, the energy payback time (EPT) is the index that accounts, by energy production, for the number of years required to recover the total energy input into the system over its entire lifetime.

The forest biomass power generation is compared to fossil and renewable resources from the perspectives of EBF and EPT. The basic theoretical equations for the two environmental load profiles defined in this study (EBF and EPT) are based on the rule of Uchiyama and Yamamoto (1991).

Furthermore, the CO2 emissions from all the processes of the system are examined. These emissions are calculated from the energy input into each process and the CO2 emission per unit energy of each energy resource. The CO2 emissions from the electricity, coal, gasoline, and diesel fuel per unit energy are 392.33 kgCO2/MWhe, 90.61 kgCO2/GJ, 67.10 kgCO2/GJ, and 68.70 kgCO2/GJ, respectively (Uchiyama and Yamamoto 1992, Ministry of the Environment 2005).

Fig.5 Harvesting volumes and procurement costs Slika 5. Koli~ina i tro{kovi pridobivanja goriva Croat. j. for. eng. 32(2011)2

Discussion on Economic and Energy Balances of Forest Biomass Utilization ... (571â&#x20AC;&#x201C;586)

3. Results and discussion â&#x20AC;&#x201C; Rezultati i rasprava 3.1 Economic balance â&#x20AC;&#x201C; Ekonomska bilanca The maximum available amount of forest biomass resources is 28,872 tDM/year, which is enough to meet the fuel requirement of a 5 MW direct combustion power plant (Fig. 4 and 5).

K. Aruga et al.

According to the harvesting volumes of the forest biomass resources, the logging residues can be harvested based on the stand harvesting schedule. If the forest biomass resources are insufficient, broad-leaved forests and thinned trees can be harvested to meet the required volumes (Fig. 6).

A comparison of the procurement costs of the forest biomass resources revealed that logging residues were the cheapest at 9,271 yen/tDM; this was fol-

Fig.6 Harvesting volumes Slika 6. Koli~ina priodobivanja drva

Fig.7 Procurement costs Slika 7. Tro{kovi pridobivanja Croat. j. for. eng. 32(2011)2

579

K. Aruga et al.

Discussion on Economic and Energy Balances of Forest Biomass Utilization ... (571â&#x20AC;&#x201C;586)

Fig. 8 Generation capacity and cost of a direct combustion power plant Slika 8. Proizvodni kapacitet i tro{kovi energetskoga postrojenja s izravnim izgaranjem

Fig. 9 Generation capacity and cost of a small-scale gasification power plant Slika 9. Proizvodni kapacitet i tro{kovi maloga energetskoga postrojenja s plinoficiranjem lowed by broad-leaved forests at 12,995 yen/tDM; thinned trees were the most costly, 15,325 yen/tDM (Fig. 7). As the target volume of the forest biomass resources decreases, the procurement costs of the forest biomass resources also decreases due to the reduction in the harvesting volumes of thinned trees and broad-leaved trees that have high procurement costs (Fig. 5).

580

Next, the optimum scale of the power-generation plant was discussed using data from Fig. 4 and 5, applied to a direct combustion power plant and a small-scale gasification power plant. With regard to direct combustion power generation, the optimum scale of the power-generation plant was a generation capacity of 5 MW and an energy-conversion efficiency of 21% (Fig. 8).

Croat. j. for. eng. 32(2011)2

Discussion on Economic and Energy Balances of Forest Biomass Utilization ... (571â&#x20AC;&#x201C;586)

Fig. 10 Generation capacity and economic balance Slika 10. Proizvodni kapacitet i ekonomska bilanca The component of the cost of electricity referred to fuel procurement was 12.2 yen/kWh for a 5 MW generation capacity. Furthermore, other costs were reduced with an increase in generation capacity.

Therefore, the total cost of electricity was the smallest for a 5 MW generation capacity: 23.7 yen/kWh. On the other hand, the minimum fuel cost of a small-scale gasification power plant was 6.8 yen/kWh for 1.5 MW generation capacity, which was within the

K. Aruga et al.

estimations in this study; however, a lower cost can be obtained if the generation capacity of the plant is increased. Moreover, other costs can be reduced with an increase in generation capacity. For 2.4 MW generation capacity, the minimum total cost was 12.8 /kWh (Fig. 9).

Since this estimated cost was lower than the average electricity price in Japan, 22.2 yen/kWh, the electricity generated from the small-scale gasification power-generation plant can be used in houses; hence, it is important to develop small-scale gasification technology. On the other hand, this estimated cost was higher than the average price of electricity sold to grids in Japan, 8 yen/kWh; therefore, selling the electricity to grids was not a viable option even for small-scale gasification. If the steam could be sold to houses at a price of 0.5 yen/kg, the economic balance of small-scale gasification would be positive, while the economic balance of direct combustion would remain negative (Fig. 10). In this case, half of the initial costs that are assumed to be borrowed for constructing the 2.4 MW small-scale gasification power-generation plant would be paid back over 3.2 years.

3.2 Energy balance â&#x20AC;&#x201C; Energetska bilanca Energy input increases with an increase in generation capacity. For direct combustion of 5 MW generation capacity, the harvesting equipment and operation energy inputs are 442 GJ/year and 7,159 GJ/year, respectively, while plant equipment and operation

Fig. 11 Generation capacity and energy input of a direct combustion power plant Slika 11. Proizvodni kapacitet i potro{nja energije energetskoga postrojenja s izravnim izgaranjem Croat. j. for. eng. 32(2011)2

581

K. Aruga et al.

Discussion on Economic and Energy Balances of Forest Biomass Utilization ... (571â&#x20AC;&#x201C;586)

Fig. 12 Generation capacity and energy input of a small-scale gasification power plant Slika 12. Proizvodni kapacitet i potro{nja energije maloga energetskoga postrojenja s plinoficiranjem

Fig. 13 Generation capacity and energy balance factor Slika 13. Proizvodni kapacitet i pokazatelj energetske bilance energy inputs are 2,636 GJ/year and 3,954 GJ/year, respectively (Fig. 11).

On the other hand, for the small-scale gasification of 2.4 MW generation capacity, the harvesting equipment and operation energy inputs are 127 GJ/year and 2,301 GJ/year, respectively, while plant equipment and operation energy inputs are 1,577 GJ/year and 2,365 GJ/year, respectively (Fig. 12).

582

The energy balance factors of direct combustion with 5 MW generation capacity and small-scale gasification with 2.4 MW generation capacity are 20.3 and 21.8, respectively (Fig. 13).

The energy balance factors of 1,000 MW generation capacity, a large-scale power generation system with coal and oil, are 17.2 and 20.8, respectively (Uchiyama and Yamamoto 1991). The energy payback times of plants using direct combustion with 5 MW generation capacity and small-scale gasification with 2.4 MW generation capacity are 0.33 and 0.38 years, respectively (Fig. 14). The energy payback times of wind power generation and solar power generation are 1.99 and 10.00 years, respectively (Uchiyama and Yamamoto 1991). Therefore, forest biomass power-generation is relatively superior to other renewable energy resources from the perspective of energy payback time while being similar to fossil energy resources from the perspective of the energy balance factor.

Regarding the energy input, the CO2 emission increases as the generation capacity increases (Fig. 15).

For direct combustion of 5 MW generation capacity, the harvesting equipment and operation CO2 emission are 36.1 tCO2/year and 202.5 tCO2/year, respectively, while the plant equipment and operation CO2 emission are 206.5 tCO2/year and 309.8 tCO2/year, respectively. On the other hand, for small-scale gasification of 2.4 MW generation capacity, the harvesting equipment and operation CO2 emission are 10.4 tCO2/year and 62.7 tCO2/year, respectively, while Croat. j. for. eng. 32(2011)2

Discussion on Economic and Energy Balances of Forest Biomass Utilization ... (571–586)

K. Aruga et al.

spectively. These reductions are much larger than the CO2 emissions from forest biomass power generation; thus, they can contribute to achieving the goals of the Kyoto Protocol in the first period of commitment, starting in 2008, when Japan’s goal was to reduce its greenhouse gas emissions by 6% of the amount recorded in 1990. Yoshioka et al. (2005) assumed that Japan has the potential of forest biomass resources to supply 100 direct combustion power-generation plants with 3 MW generation capacity.

The CO2 emission and reduction from a direct combustion plant with 3 MW generation capacity are 507 tCO2/year and 10,471 tCO2/year, respectively.

Therefore, 100 power-generation plants can reduce CO2 emissions by 996,400 tCO2/year. This figure is 1.3% of the 74,000,000 tCO2/year that is the amount of greenhouse gas emission that Japan must reduce.

Fig. 14 Generation capacity and energy payback time Slika 14. Proizvodni kapacitet i vrijeme povrata energije

Fig. 15 Generation capacity and CO2 emission Slika 15. Proizvodni kapacitet i emisija CO2 plant equipment and operation CO2 emission are 123.5 tCO2/year and 185.3 tCO2/year, respectively.

However, using surplus electricity and steam reduces CO2 emissions from fossil energy resources (Aruga et al. 2007). The CO2 reduction from direct combustion with 5 MW generation capacity and small-scale gasification with 2.4 MW generation capacity are 15,707 tCO2/year and 6,275 tCO2/year, reCroat. j. for. eng. 32(2011)2

4. Conclusions – Zaklju~ci In this study, the economic and energy balances of forest biomass utilization for small-scale power generation were discussed considering the spatial distribution of forest biomass resources using GIS in the Kanuma Area of Tochigi Prefecture, Japan. First, the optimum scales of a direct combustion power plant and a small-scale gasification power plant (technology currently under development) were discussed.

It is important to develop this small-scale gasification technology because its electricity cost, 11.5 yen/kWh, with the optimum scale of a small-scale gasification power plant, was lower than the average electricity price in Japan, 22.2 yen/kWh. Next, the energy balance and CO2 emissions from the energy utilization of forest biomass resources were analyzed using LCI. The results show that the system examined in this study is feasible as an energy production system.

Although the spatial distribution of forest biomass resources was considered in this study, only one power-generation plant was discussed. The introduction of wood-fired boilers and generators and the production of wood pellets have been increasing steadily in recent years in Japan. Multiple forest biomass utilization facilities can be located in a single area. Therefore, multiple facilities should be considered, and the scales and locations of these facilities should be optimized in a future study.

Acknowledgments – Zahvala We are grateful to the Tochigi Prefecture Government for providing the data. We also thank the anonymous reviewers for their constructive comments.

This study was supported by Nissei Zaidan.

583

K. Aruga et al.

Discussion on Economic and Energy Balances of Forest Biomass Utilization ... (571–586)

5. References – Literatura Aruga, K., Yoshioka, T., Sakurai, R., 2006a: Long-term feasibility of timber and forest biomass resources at an intermediate and mountainous area-Balance of harvesting volumes using random search. J. Jpn. For. Eng. Soc. 21(1): 49–59.

Aruga, K., Tasaka, T., Yoshioka, T., Sakurai, R., Kobayashi.

H., 2006b: Long-term feasibility of timber and forest biomass resources at an intermediate and mountainous area (2) – Examining the optimum scale of an energy plant. J. Jpn. For. Eng. Soc. 21(3): 185–192.

Aruga, K., Saito, M., Tasaka, T., Yoshioka, T., 2007: Long-term feasibility of timber and forest biomass resources at an intermediate and mountainous area (3) – Discussion on energy balance. J. Jpn. For. Eng. Soc. 21(4): 251–256.

Dijkstra, E. W., 1959: A note on two problems in connection with graphs. Numerische Mathematik 1(1): 269–271.

Faaij, A., Ree, R. van, Waldheim, L., Olsson, E., Oudhuis, A., Wijk, A. van, Daey-Ouwens, C., Turkenburg, W., 1997: Gasification of biomass wastes and residues for electricity production. Biomass and Bioenergy 12(6): 387–407.

Forestry Agency, The Ministry of Agriculture, Forestry and Fisheries of Japan, 2009: Annual Report on Trends in Forest and Forestry. Fiscal Year 2008 (summary), Tokyo, 31 p.

Forsberg, G., 2000: Biomass energy transport: Analysis of bioenergy transport chains using life cycle inventory method. Biomass and Bioenergy 19(1): 17–30.

Hondo, H., Uchiyama, Y., Moriizumi, Y., 2000: Evaluation of power generation technologies based on life cycle CO2 emissions: Re-estimation using the latest data and effects of the difference of conditions. Socioeconomic Research Center Report No. Y99009, Central Research Institute of Electric Power Industry, Tokyo.

Ministry of the Environment, 2005: Guidelines for the method of calculating GHG emissions from organizations.

Ministry of the Environment, Tokyo, 83 p.

Panichelli, L., Gnansounou, E., 2008: GIS-based approach for defining bioenergy facilities location: A case study in Northern Spain based on marginal delivery costs and resources competition between facilities. Biomass and Bioenergy 32(4): 289–300.

Ranta, T., 2005: Logging residues from regeneration fellings for biofuel production-A GIS-based availability analysis in Finland. Biomass and Bioenergy 28(2): 17–182.

Shiraishi, N., 1985: Study on the growth prediction of even-aged stands. Bulletin of Tokyo University Forests 75: 199–256.

Tahara, K., Kojima, T., Inaba, A., Ogi, T., Yokoyama, S., 1998: Reduction in CO2 emission by biomass power generation with sustainable afforestation: Evaluation by LCA. J. Jpn. Inst. Energy 77: 403–409.

Uchiyama, Y., Yamamoto, H., 1991: Energy analysis on power generation plants. Economic Research Center Report No. Y90015, Central Research Institute of Electric Power Industry, Tokyo.

Uchiyama, Y., Yamamoto, H., 1992: Greenhouse effect analysis of power generation plants. Economic Research Center Report No. Y91005, Central Research Institute of Electric Power Industry, Tokyo.

Yagi, K., Nakata, T., 2007: Economic analysis on small-scale forest biomass gasification considering regional resource distribution and technical characteristics. J. Jpn. Inst. Energy 86(2): 109–118.

Yoshioka, T., Sakai, H., 2005: Amount and availability of forest biomass as an energy resource in a mountain region in Japan: A GIS-based analysis. Croatian J. For. Eng. 26(2): 59–70.

Yoshioka, T., Aruga, K., Nitami, T., Kobayashi, H., Sakai, H., 2005: Energy and carbon dioxide (CO2) balance of logging residues as alternative energy resources: System analysis based on the method of a life cycle inventory (LCI) analysis. Journal of Forest Research 10(2): 125–134.

Nord-Larsen, T., Talbot, B., 2004: Assessment of forest-fuel resources in Denmark: Technical and economic availability. Biomass and Bioenergy 27(2): 97–109.

584

Croat. j. for. eng. 32(2011)2

Discussion on Economic and Energy Balances of Forest Biomass Utilization ... (571â&#x20AC;&#x201C;586)

K. Aruga et al.

Sa`etak

Rasprava o ekonomskoj i energetskoj bilanci kori{tenja {umske biomase za malu proizvodnju energije u Kanumi, prefekturi Tochigi, u Japanu [ume imaju zna~ajnu ulogu u odr`avanju ravnote`e ugljika na Zemlji povla~enjem ugljika iz atmosfere i proizvodnjom drva, obnovljivoga resursa koji pohranjuje odstranjeni ugljik. Zbog toga {umama treba kontinuirano i pravilno gospodariti, a drvo treba biti kori{teno na svim razinama â&#x20AC;&#x201C; od gra|evnoga materijala, namje{taja, plo~a i papira do kemijskih proizvoda i goriva. Drvna se biomasa pritom mo`e kategorizirati na {umske ostatke (u ovom se istra`ivanju spominju kao resursi {umske biomase), ostatke pilana i gra|evno otpadno drvo. Posljednjih je godina u Japanu zabilje`en siguran porast u uvo|enju kotlovnica i generatora u kojima se koristi drvo te u proizvodnji drvnih peleta. Ipak, velik dio drvne biomase, osobito {umskih ostataka, i dalje ostaje neiskori{ten. Radi kori{tenja izvora {umske biomase za proizvodnju bioenergije klju~no je odrediti odnos izme|u godi{nje raspolo`ivih koli~ina i tro{kova pridobivanja (prikupljanja i transporta) {umske biomase. Mnogobrojna su istra`ivanja uz ekonomska pitanja koja se raspravljaju u ovom radu analizirala i energetska pitanja kori{tenja biomase. Me|utim, ve}ina je tih istra`ivanja usmjerena na velika energetska postrojenja ~ija se neto proizvodnja energije kre}e u rasponu od nekoliko desetaka do vi{e stotina megavata energije. Neki su autori analizirali sustave maloga opsega koji kao gorivo za pretvaranje energije upotrebljavaju {umsku biomasu iz konvencionalnoga japanskoga {umarstva, ali pritom nisu uzimali u obzir prostorni raspored raspolo`ivih resursa {umske biomase.

U ovom se radu razmatraju ekonomske i energetske bilance kori{tenja {umske biomase u proizvodnji energije maloga opsega, uzimaju}i u obzir prostornu razdiobu resursa {umske biomase i primjenu geografskoga informacijskoga sustava (GIS) na podru~ju Kanuma, u prefekturi Tochigi, u Japanu. Podru~je Kanuma, koje se kao istra`iva~ki poligon sastoji od gradova Kanuma i Nishikata, obuhva}a 52 000 hektara, od ~ega je oko 65 % po{umljeno. Ve}ina je {umskih sastojina umjetno podignuta (79 %), prevladavaju ~etinja~e i glavnina je sastojina stara oko 45 â&#x20AC;&#x201C; 50 godina. U radu su kori{teni inventurni podaci ({umski resursi, dob sastojina, vrste drve}a, indeksi stani{ta) i podaci GIS-a (informacije o javnim i {umskim cestama, nagib terena i dr.) koji su dobiveni od Uprave prefekture Tochigi. Na osnovi toga materijala i podataka, uz pomo} GIS-a, u radu su procijenjene godi{nje dostupne koli~ine drva i {umske biomase te tro{kovi njihova pridobivanja na razini pododjela koji su uobi~ajena operativna jedinica u japanskom {umarstvu.

Istra`ivanjem su obuhva}ene dvije vrste elektrana, odnosno dva tipa pretvaranja i proizvodnje energije. Jedno je energetsko postrojenje s izravnim izgaranjem, a drugo je malo energetsko postrojenje s plinoficiranjem. Postrojenje s plinoficiranjem predstavlja tehnologiju koja je trenuta~no u razvoju. U radu su prvo na osnovi ekonomske bilance odre|eni optimalni kapaciteti promatranih postrojenja za proizvodnju energije, a zatim su analizirane energetska bilanca i emisija CO2 u proizvodnji energije kori{tenjem resursa {umske biomase. Pritom je upotrijebljena metoda inventure `ivotnoga ciklusa (LCI). Dodatno su uspore|ene proizvodnja energije iz {umske biomase s energijom iz fosilnih goriva i drugih obnovljivih izvora.

Rezultati istra`ivanja pokazuju da bi u slu~aju energetskoga postrojenja s izravnim izgaranjem, kojemu je optimalni proizvodni kapacitet 5 MW, cijena struje iznosila 23,7 jena/kWh (1 = 114 jena, 27. lipnja 2011). Za malo energetsko postrojenje s plinoficiranjem i optimalnim kapacitetom proizvodnje od 2,4 MW cijena bi struje iznosila 12,8 jena/kWh. S obzirom na to da je prosje~na cijena struje u Japanu 22,2 jena/kWh, proizvodnja struje u malom energetskom postrojenju s plinoficiranjem mogla bi biti ekonomi~na. [to se ti~e energetske bilanca i emisije CO2, za oba dva tipa proizvodnje energije utvr|eno je da odnos izme|u outputa i inputa energije iznosi oko 20, {to pokazuje da elektrane analizirane u radu mogu biti ostvarivi sustavi za proizvodnju energije. Emisija CO2 kod energetskoga postrojenja s izravnim izgaranjem proizvodnoga kapaciteta 5 MW iznosila je 754,9 tCO2/godina, dok je kod maloga postrojenja s plinoficiranjem i proizvodnim kapacitetom 2,4 MW iznosila 381,9 tCO2/godina. Ipak, smanjenje koli~ine emisije CO2 koje bi se postiglo zamjenjivanjem fosilnih goriva iznosi 15 707 tCO2/godina, odnosno 6275 tCO2/godina. Tako|er je utvr|eno da je u usporedbi s fosilnim gorivima i drugim obnovljivim izvorima energije proizvodnja energije iz {umske biomase relativno superiorna s obzirom na vrijeme povrata energije

Croat. j. for. eng. 32(2011)2

585

K. Aruga et al.

Discussion on Economic and Energy Balances of Forest Biomass Utilization ... (571–586)

(indeks koji ra~una broj godina potrebnih za dobivanje energije ukupno unesene u sustav) i sli~na po faktoru energetske bilance (odnos izme|u izlaza i ulaza energije). Rezultati rada pokazuju da su prikazani sustavi izvedivi i da mogu biti vrijedna rje{enja u proizvodnji energije.

Klju~ne rije~i: ekonomska bilanca, energetska bilanca, bilanca CO2, resursi {umske biomase, mala proizvodnja energije, GIS, LCI

Authors’ address – Adresa autorâ: Assoc. Prof. Kazuhiro Aruga, PhD. e-mail: aruga@cc.utsunomiya-u.ac.jp Ayami Murakami, MSc. e-mail: ayamy_mumu@yahoo.co.jp Chikara Nakahata, BSc. e-mail: c.nakahata0927@gmail.com Reiko Yamaguchi, BSc. e-mail: r.yamaguchi0210@gmail.com Utsunomiya University Faculty of Agriculture Department of Forest Science 350 Mine Utsunomiya 321-8505 JAPAN

Received (Primljeno): April 05, 2011 Accepted (Prihva}eno): June 21, 2011

586

Asst. Prof. Takuyuki Yoshioka, PhD. e-mail: yoshioka.takuyuki@nihon-u.ac.jp Nihon University College of Bioresource Sciences Department of Forest Science and Resources 1866 Kameino Fujisawa 252-0880 JAPAN Croat. j. for. eng. 32(2011)2

Original scientific paper – Izvorni znanstveni rad

Comparing Terrain and Roadside Chipping in Mediterranean Pine Salvage Cuts Enrico Marchi, Natascia Magagnotti, Lisa Berretti, Francesco Neri, Raffaele Spinelli Abstract – Nacrtak In central Italy, the increasing demand for fuel chips and the epidemic spread of maritime pine bast scale have favored the development of large-scale industrial logging operations. After years of extensive commercial trials, local operators have developed their own industrial harvesting systems, through a wise mix of Scandinavian and North American methods. The result is original and effective, and allows keeping harvesting cost below 20 gt–1. The study compared terrain chipping with roadside chipping, as applied to the coastal pine stands of Tuscany. Under the conditions of our study, roadside chipping was over four times more productive than terrain chipping, and it allowed reducing harvesting cost by one third (12.3 vs. 18.3 gt–1). Despite the intense use of diesel, total fossil energy inputs accounted for less than 3% of the potential energy in the wood chips. Terrain chipping and roadside chipping yielded 36 and 47 times the energy they used, respectively. The coexistence of the two systems was most interesting. The harvesting systems described in the study perform best in clear-cuts, but they can also work in partial cuts, including thinning operations. They are actually used in thinnings in the same Regional Park of San Rossore, although their productivity is lower than in clear-cuts.

Keywords: biomass, logistics, productivity, economics, salvage cuts, Mediterranean

1. Introduction – Uvod Many studies forecast a significant increase in the use of energy biomass over the coming years (Berndes et al. 2003), and the forest industry is already exploring this new growing market. Biomass harvesting does add to the complexity of forestry, but it also offers a significant opportunity to increase efficiency, raise value recovery and reduce logging and management costs (Björheden 2000). The recovery of forest biomass generally requires some form of processing – chipping or bundling – aimed at increasing the density and the homogeneity of the feedstock (Spinelli and Magagnotti 2009). Under ideal access conditions, the biomass can be chipped in the stand, and chips rather than trees can be extracted to the roadside landing (Kalaja 1984): among other things, direct delivery of chip loads to the roadside reduces the landing space requirements, and makes this system most suited to the situations where the forest infrastructure is poor or fragmented (Kofman 1993).

However, terrain chipping requires dedicated equipment, often smaller and less productive compared to Croat. j. for. eng. 32(2011)2

truck or trailer mounted chippers, working at the roadside. Furthermore, chipping at the roadside allows dumping the chips directly into the trucks, which avoids the additional cost of loading onto transportation vehicles. These benefits of roadside chipping may offset the higher cost of extracting bulky uncomminuted residues. On the other hand, both systems are faced with interaction delays, which occur between the chipper and the chip extraction fleet in the case of terrain chipping, and between the chipper and the transportation fleet in the case of roadside chipping (Spinelli and Visser 2009).

When choosing between these two options, one is confronted with two conceptual problems: point of comminution and economy of scale (Björheden 2008). These problems are exacerbated in industrial operations, due to the massive material flow, the high operating costs and the reduced operational flexibility. Industrial logging operations are increasingly popular in Central Italy, favored by a large biomass demand and the urgent need for extensive salvage logging, consequent to the epidemic spread of

587

Enrico Marchi et al.

Comparing Terrain and Roadside Chipping in Mediterranean Pine Salvage Cuts (587–598)

maritime pine bast scale (Matsucoccus feytaudi). This is a specific pest of Maritime pine that is endemic in the Iberian Peninsula and southwestern France, and invasive in southeastern France, Italy and Corsica, where it is causing large scale forest damage (Kerdelhué and Decroocq 2006). Despite the rapid evolution of pest management techniques, clearcutting of infected forests is still the most common control measure applied to most cases, and especially in central Italy (Brockerhoff et al. 2006). Clear-cutting is also applied to overmature umbrella pine plantations, which are very common along the Tuscan coastline (Barbero et al. 1998). These stands were planted about a century ago and are now old, weak and increasingly vulnerable to sea winds (Cantiani e Scotti 1988). The mainstream silvicultural prescription is again clear-cutting, followed by replanting or by re-naturalization if the quality of the hardwood understory is good (Zerbe 2002) In turn, large clear-cuts have favored the introduction of heavy industrial machinery, which performs best under such conditions. This has resulted in a steady reduction of supply costs, and a parallel search for operational optimization. Large and expensive operations are especially vulnerable to poor planning, and their managers are especially keen on finding the best deployment strategy. After all, harvesting and transportation cost can represent approximately 70% of the total biomass cost (Panichelli and Gnansounou 2008), and today this cost represents one of the most important barriers to the increased use of biomass (Rentizelas et al. 2009). In Tuscany, salvage harvesting operations have been optimized to a point that the Tuscan enterprises can afford ferrying their biomass across the Mediterranean sea to Sardinia, for co-firing in a large power station. This would not be possible unless harvesting incurred a very low cost. Until now, most operators adopted terrain chipping with heavy industrial units, and that seemed to be the key to their enduring success. However, one operator has recently introduced an even more powerful truck-mounted chipper for chipping at the roadside, turning upside down the mainstream operational philosophy of the region. That is a radical innovation, and the large size of the new machine has raised many questions.

The international literature can offer little support, due to the very large size of these machines, and to their deployment in the middle of the Mediterranean basin, where few people are credited with much experience of large scale mechanized harvesting operations, except perhaps for the Tuscan entrepreneurs (Spinelli et al. 2009a).

The goal of this study is to provide a quantitative comparison of terrain and roadside chipping opera-

588

tions, conducted with very powerful chippers under the conditions of industrial clear-cutting in a Mediterranean environment. Such comparison spans over the whole operation – from standing tree to chips loaded on the trucks – and includes both technical and economical aspects, as represented by productivity, workplace time allocation and harvesting cost per unit product. Sensitivity analysis is used to refine the comparison, by gauging the effect of varying conditions on harvesting cost, for both systems.

2. Material and Methods – Materijal i metode The trials were conducted near Pisa, Italy, inside the Regional Park of San Rossore, which encloses a surface of about 3,000 hectares (ha) and is in large part covered by pine plantations (Spinelli et al. 2009b).

The trial took place during commercial harvesting of two different woodlots, both in flat even ground.

Terrain chipping was applied to a 120 years old umbrella pine (Pinus pinea L.) plantation, over-mature and declining. In contrast, roadside chipping was applied to an 80 years old maritime pine (Pinus pinaster L.) plantation, with severe pine bast damage. Both stands were clear-cut, using 27-ton JD 759J swing-to-tree tracked feller-bunchers. These machines worked well ahead of the chippers, felling, separating the basal logs and bunching tops, large branches and small trees in big piles. The feller-bunchers also broke large branches, in order to facilitate forwarding and chipping. Hence, field conditions for chipping were quite similar in both stands, despite the different age and species. The actual concentration

Fig. 1 Terrain chipping operation: forwarder-mounted chipper and chip shuttle Slika 1. Iveranje u sastojini: forvarder opremljen ivera~em i vozilo za transport ivera Croat. j. for. eng. 32(2011)2

Comparing Terrain and Roadside Chipping in Mediterranean Pine Salvage Cuts (587â&#x20AC;&#x201C;598)

of energy biomass was not measured, but was estimated to 80% of the total harvest, or 250 green tons (gt) haâ&#x20AC;&#x201C;1, based on earlier studies of the same sites and operations (Spinelli et al. 2002). This estimate accounted for both operations, since the amount of residue left on site appeared very similar. Besides, the quoted earlier studies covered both stand types and found little difference between them, at least in terms of quantity and size of residue biomass.

The terrain chipping operation was very simple, with a forwarder-mounted chipper blowing chips directly into chip shuttles (Suadicani 2003). Compared to the classic Danish example, the Italian version used a bigger chipper and simpler shuttles (Fig. 1). The chipper was an Erjo 12/90 model, powered by a 430 kW independent engine and mounted on a John Deere JD1410 D forwarder. The chip shuttles consisted of two 128 kW Valtra T161 farm tractors, each pulling a large silage trailer, with a capacity of 22 and 30 m3. The terrain chipper did not carry its own built-in container, and blew chips directly into the tractor trailers. Once full, the trailers were driven to a large paved landing about 2 km away, and their content was dumped on the ground for subsequent loading on trucks. The chipping operation was manned by two operators only, one on the chipper and the other taking turns on the tractors, alternatively parking the empty trailer under the chipper and driving the full one to the landing. If the chipper had to move to a new stack, the chipper driver would first drive the chipper to its new work station, then dismount, move the shuttle and finally climb back onto the chipper cab to resume chipping.

Chips would be loaded into open-top chip vans with a 20-ton Liebherr 904 excavator, equipped with a clam bucket and a high-raise cab. This system was introduced to the region 10 years ago and is now the most common, with about 7 operations running the same way. Local operators know its potential and limits, and have refined its application over the years.

The roadside chipping operation was more complex, involving more units and a significantly higher investment. Three 14-ton JD 1410 forwarders were used to move uncomminuted energy biomass to a roadside landing, about 150 m from the centre of the woodlot. Here the biomass was chipped with an Erjo 15/120 drum chipper, powered by two 515 kW engines, for a 1030 kW total delivered power (Fig. 2).

The chipper was mounted on a semitrailer and relocated using a truck tractor. It was fed by a modified 26-ton Liebherr 924 excavator, with a log grapple and a high-raise cab. The chips were blown directly into open top chip vans, so that separate re-loading was not necessary. A chip shuttle of the type describCroat. j. for. eng. 32(2011)2

Enrico Marchi et al.

Fig. 2 Roadside chipping operation: the heavy chipper filling a chip van Slika 2. Iveranje na stovari{tu: te{ki ivera~ puni kamion iverom ed above was parked by the chipper and used as a surge bin, if trucks were delayed (Blair 1998). Once full, the chip shuttle would be driven to the landing used for the terrain chipping operation, whence the chips could be reloaded on trucks using the clam-bucket. This operation was manned by four operators, three on the forwarders and one on the excavator. The latter would also operate the chipper, using a remote control.

Both operations were owned and managed by the same main contractor, so that company policy and manager skills would not change between treatments. Both systems worked hot-deck, with limited buffers between the chipping and the extraction units (Han et al. 2004). Both chippers were equipped with the standard bar screen, designed for producing industrial chips.

All operators included in the study were experienced professionals, who knew their job and equipment, and had at least 5 years of experience with the type of machine they were using. The only exception was the operator of the roadside chipper, whose machine had been commissioned one year earlier. However, he was a very experienced chipper operator, who had run industrial chippers for over 20 years.

In order to determine productivity and workplace time distribution, we carried out a typical time and motion study (Bergstrand 1991). The study focused on the chippers, considered as the pivotal element of the chipping operation. Chip shuttles, forwarders and the excavator-base loader were considered as auxiliary to the chippers, and their hourly cost was simply added to the hourly cost of the chippers they served. If machine unbalance affected the chipper, this would be reflected by the presence of chipping delays. If that affected the auxiliary units,

589

Enrico Marchi et al.

Comparing Terrain and Roadside Chipping in Mediterranean Pine Salvage Cuts (587–598)

sented by the delay factor (DF), i.e. the ratio of delay time to net work time. Contrary to the incidence of delays over total time, a DF has no internal correlation and is easier to generalize.

The output was estimated by measuring the volume of all chip shuttles and chip vans, and by taking all chip vans to the certified weighbridge installed at the Park gate and used by the forest owner to quantify the sale. The bulk density figure obtained from the chip vans was then applied to the chip shuttles, converting all output into weight figures. Twenty 1-kg chip samples were randomly collected from each test and taken to the laboratory: half of the samples were used for determining moisture content according to the European standard CEN/TS 14774-2, and half for determining particle size distribution according to the European Standard CEN/TS 15149-1. Although

the study would not detect it, but these units were already accounted for their full cost, regardless of whether they worked full time or not. We also conducted a separate study at the landing in order to determine the productivity and the cost of loading. In all cases, a full chip load (chip shuttle or chip van) was considered as a full work cycle.

Each cycle was stop watched individually, using a conventional 3-watch time-study board (Picchio et al. 2009). Productive time was separated from delay time (Björheden et al. 1995), but all delays were included in the study, and not just the delays below a certain duration threshold, because such practice may misrepresent the incidence of downtime (Spinelli and Visser 2008). However, delays generated by the study itself were separated and removed from the data sets. The incidence of delays was repre-

Table 1 Costing assumptions and machine cost, excluding labor Tablica 1. Pretpostavke tro{kova i tro{ak stroja, ne uklju~uju}i tro{ak radnika Machine Radni stroj Purchase price Nabavna cijena Economic life Radni vijek Resale value Preprodajna vrijednost Interest rate Kamatna stopa Fuel consumption Potro{nja goriva Depreciation Amortizacija Annual use Godi{nja iskori{tenost Total fixed cost Ukupni fiksni tro{kovi Fuel & lubricant Tro{ak goriva i maziva Repair & maintenance Tro{kovi popravaka i odr`avanja Total variable cost Ukupni varijabilni tro{ak Overhead (20%) Op}i tro{kovi Total cost Ukupni tro{kovi

Type Vrsta Model Model

Terrain chipper Chip Shuttle Forwarder Ivera~ u sastojini Vozilo za dopremu ivera Forvarder

Roadside chipper Ivera~ na stovari{tu

26-t loader 20-t loader 26-tonski utovariva~ 20-tonski utovariva~

Erjo 12/90

Valtra 161

JD 1410D

Erjo 15/120

Liebherr 924

Liebherr 904

500000

120000

320000

1000000

140000

110000

Years Godine % new % novog

10

10

10

10

10

10

25

33

25

25

25

33

%

5

5

5

5

5

5

l SMH–1

35

10

20

110

15

15

year–1 god–1

37500

8040

24000

75000

10500

7370

SMH

1500

1500

1500

1500

1500

1500

SMH–1

38.3

8.7

24.5

76.5

10.7

12.0

SMH–1

50.1

14.3

28.6

157.3

21.5

21.5

SMH–1

12.5

1.9

5.6

17.5

2.5

2.6

SMH–1

62.6

16.2

34.2

174.8

24.0

24.1

SMH–1

20.2

5.0

11.7

50.3

6.9

7.2

SMH–1

121.1

29.9

70.4

301.6

41.6

43.3

Note: costs in Euro ( ), as on July 26th, 2011. 1 = 1.44 US Dollars; SMH = Scheduled Machine Hours, including delays (e.g. workplace time) Napomena: cijene su u eurima ( ) za 26. srpnja 2011. 1 = 1,44 ameri~ka dolara; SMH = planirani sati rada stroja, uklju~uju prekide rada (vrijeme na radnome mjestu)

590

Croat. j. for. eng. 32(2011)2

Comparing Terrain and Roadside Chipping in Mediterranean Pine Salvage Cuts (587–598)

Enrico Marchi et al.

Table 2 Total operation cost, including labor Tablica 2. Ukupni tro{kovi rada stroja s uklju~enim tro{kom radnika Operation Radni zahvat n Terrain chipper Ivera~ u sastojini Chip shuttle Vozilo za dopremu ivera Forwarder Forvarder Roadside chipper Ivera~ na stovari{tu 26–t loader 26–tonski utovariva~ 20–t loader 20–tonski utovariva~ Crew Radna skupina Total Ukupno

Terrain Chipping Iveranje u sastojini SMH–1

Roadside chipping Iveranje na stovari{tu n SMH–1

n

Loading Utovar SMH–1

1

121.1

0

0.0

0

0.0

2

59.8

1

29.9

0

0.0

0

0.0

3

211.2

0

0.0

0

0.0

1

301.6

0

0.0

0

0.0

1

41.6

0

0.0

0

0.0

0

0.0

1

43.3

2

40.0

4

80.0

1

20.0

220.9

664.3

63.3

Note: SMH = Scheduled Machine Hours, including delays (e.g. workplace time) Napomena: SMH = planirani sati rada stroja, uklju~uju prekide rada (vrijeme na radnome mjestu)

this analysis separated the classic six size classes (< 3, 3 – 16, 16 – 45, 45 – 63, 63 – 100 and > 100 mm), we grouped classes into three main functional categories to make interpretation easier. These were: oversize (> 63 mm), accepts (3 – 63 mm) and fine (< 3 mm) particles.

Machine costs were calculated with the method described by Miyata (1980), over costing assumptions provided by the contractor himself. Fuel consumption was determined by recording the quantities of diesel added to the tanks during the trials, as well as the tank levels at the beginning and at the end of the trials. The calculated operational cost was increased by 20% in order to include relocation and administration costs, the former already capable of representing up to 10% of the total machine cost (Väätäinen et al. 2006). Further detail on cost calculation is shown in Table 1. Table 2 reports the total cost of each operation, and includes labor cost, estimated to 20 hour–1 inclusive of all taxes and benefits.

The actual harvesting cost was calculated as the sum of felling, extraction, chipping and loading. As to chipping and extraction, the unit cost was obtained by dividing the system costs reported in Table 2 by the respective system productivities. The loading cost was obtained by dividing loader cost by loader productivity. The resulting figure was added entirely to the cost of the terrain chipping system, where all chips had to be re-loaded on chip vans.

With roadside chipping, we calculated the percent of Croat. j. for. eng. 32(2011)2

chips blown in the surge bins, and used this figure to pro-rate the cost of loading for the roadside system.

The cost of felling and bunching was obtained from the contractor, and resulted to be 2.2 gt–1. This figure excludes the separation of butt logs, as this cost is fully charged to the round wood product. Felling and bunching cost was added equally to both treatments, in order to calculate the total harvesting cost, from standing tree to chips on the chip van.

Both direct and indirect fossil energy consumption were estimated, with the exclusion of manual work. Direct energy inputs were estimated by multiplying the measured diesel consumption by its energy content of 37 MJ l–1 (Bailey et al. 2003), and then inflating this value by 1.2 in order to account for the additional fossil energy consumed in the production and transportation of diesel fuel (Pellizzi 1992). Indirect energy inputs incurred during machine manufacturing, repair and maintenance were estimated as 30% of direct energy consumption (Mikkola and Ahokas 2010). The heating value of conifer chips was assumed to be equal to 20 MJ dry kg–1 (Spinelli et al.

2011) The Mann-Whitney non-parametric comparison test was used to check the statistical significance of differences between treatments (SAS 1999). Non-parametric statistics were adopted, since data distribution deviated from normality.

The study on terrain chipping lasted 29.3 workplace hours, during which the terrain chipper pro-

591

Enrico Marchi et al.

Comparing Terrain and Roadside Chipping in Mediterranean Pine Salvage Cuts (587–598)

duced 57 loads, or 446 gt. The study on roadside chipping covered 22.5 workplace hours, during which the chipper produced 51 loads, or 1281 gt. Finally, the study on loading lasted 11.1 workplace hours, during which the loader filled 17 chip vans, for a total of 476 gt.

3. Results – Rezultati Table 3 shows how the loading and felling costs were calculated. As an average, the loader took about 20 minutes to fill a chip van: 10 more minutes were needed to cover the load and tie the tarpaulin, so that the average terminal time of a chip van was 30 minutes, excluding delays. Chip van delays were not quantified with this study, but loader delays were, and they added about 38% to net time consumption. About three quarters of the loader delays consisted in waiting idle for a new chip van to show up. While all the chips produced by the terrain chipping operation were dumped on the ground and had to be loaded, only 10% of those produced by the roadside operations had to be dumped on the ground, because directed to the surge bin, which could not be dumped directly into a truck. The remaining 90% was blown directly into the chip vans and required no separate loading. Hence the additional loading cost for the roadside chipping operation was 10% of the full loading cost, i.e. 0.15 gt–1.

Table 4 shows the productivity of the two chipping operations, the total harvesting cost and its breakdown by process step. Under the conditions of our study, roadside chipping was over four times more productive than terrain chipping, and it al-

lowed reducing harvesting cost by one third. Once up and running, the powerful roadside chipper could process over 100 green tons of chips per hour, filling a chip van in less than 20 minutes. Productivity remained exceptional even after including all accessory work time and delays, the latter adding over 30% to net time consumption (chipping + accessory work time). The graphs in Fig. 3 show the higher incidence of delays for the roadside chipping operation. A large proportion of the roadside chipper delays is related to the higher maintenance needs. In contrast, terrain chipping is less affected by delay time. Here, moving between stacks and repeated parking of the chip shuttles are the main elements limiting productivity. The incidence of waiting times is small with both systems. Given the recent introduction of the new roadside chipping system, further improvements may be expected as a consequence of technological learning (Junginger et al.

2005).

As expected, chipping and extraction were the most expensive process steps, accounting for 80% of the total harvesting cost. Felling and bunching represented between 12 and 17% of the total harvesting cost. Loading had a minor yet significant impact.

The detailed time study allowed checking the cost effect of placing a surge bin by the roadside chipper (Table 5). Without a surge bin, extraction, chipping and loading cost would have increased by 3.2%, i.e. 0.33 gt–1.

Both operations used large amounts of fuel, with the roadside chipper burning over 100 l of diesel per hour. Fuel cost represented 31% and 36% of the total harvesting cost for terrain chipping and roadside

Table 3 Calculating the additional cost of loading and felling Tablica 3. Obra~un dodatnih tro{kova utovara i sje~e Operation Radni zahvat Loading net productivity Neto proizvodnost utovara Loading gross productivity Bruto proizvodnost utovara Delay Factor for loading Udio op}ih vremena pri utovaru Loading cost Tro{ak utovara % mass loaded Masa ivera ispu{tena na tlo, pa utovarena, % Cost of loading Dodatni tro{ak utovara Cost of felling Tro{ak sje~e

Terrain chipping Iveranje u sastojini

Roadside chipping Iveranje na stovari{tu

gt PMH–1

59.2

59.2

gt SMH–1

43.0

43.0

%

37.7

37.7

gt–1

1.5

1.5

%

100

10

gt–1

1.5

0.1

gt–1

2.2

2.2

Notes: PMH = Productive Machine Hours, excluding delays; SMH = Scheduled Machine Hours, including delays (e.g. workplace time) Napomena: PMH = pogonski sati rada stroja bez prekida rada; SMH = planirani sati rada stroja s prekidima rada

592

Croat. j. for. eng. 32(2011)2

Comparing Terrain and Roadside Chipping in Mediterranean Pine Salvage Cuts (587–598)

Enrico Marchi et al.

Table 4 Chipper productivity and total harvesting cost Tablica 4. Proizvodnost ivera~a i ukupni tro{kovi pridobivanja ivera Operation Radni zahvat Observations Broj opa`anja Load size Veli~ina tovara Moisture content Udio vlage Chipper productivity Proizvodnost ivera~a Net operation productivity Neto proizvodnost radnoga zahvata Gross operation productivity Bruto proizvodnost radnoga zahvata Delay factor for chipping Udio op}ih vremena pri iveranju Unit production cost Jedini~ni tro{ak proizvodnje Felling cost Tro{ak sje~e Chipping & Extraction cost Tro{kovi iveranja i privla~enja drva Loading cost Tro{ak utovara

Terrain chipping Iveranje u sastojini

Roadside chipping Iveranje na stovari{tu

N

57

51

gt t (svje`ega ivera)

7.8

25.1

< 0.0001

%

49.3

48.8

0.6831

gt chip only hour bt ivera h–1

23.3

111.1

< 0.0001

gt PMH–1

16.7

90.9

< 0.0001

gt SMH–1

15.2

66.7

< 0.0001

%

10.7

32.4

0.5333

gt–1

18.3

12.3

< 0.0001

12.0

17.8

79.9

81.0

8.1

1.2

p

–1

% of total cost % udio u ukupnim tro{kovima % of total cost % udio u ukupnim tro{kovima % of total cost % udio u ukupnim tro{kovima

Notes: PMH = Productive Machine Hours, excluding delays; SMH = Scheduled Machine Hours, including delays (e.g. workplace time); p = significance of differences between the average values for terrain and roadside chipping as resulting from the Mann-Whitney non parametric test, conducted at the 5% level.

Napomena: PMH = pogonski sati rada bez prekida rada; SMH = PPS planirani sati rada stroja s prekidima rada; p (signifikantna razlika) = zna~ajnost razlike izme|u prosje~nih vrijednosti iveranja u sastojini i na stovari{tu odre|ivana je Mann-Whitneyevim neparametarskim testom provedenim na razini od 5 %.

Fig. 3 Breakdown of chipper workplace time for terrain chipping (left) and roadside chipping (right) Slika 3. Podjela radnih zahvata ivera~a u sastojini (slika lijevo) i na pomo}nom stovari{tu (slika desno) chipping, respectively. This justified the concern for the effect of diesel price on harvesting cost, given the ever increasing fuel prices. Fig. 4 shows the results of a sensitivity analysis, tying total harvesting cost to fuel price. If fuel price increased over 50% and Croat. j. for. eng. 32(2011)2

passed from the base 1.3 l–1 assumption to a maximum 2 l–1, harvesting cost would increase by 22 and 25% for terrain chipping and roadside chipping, respectively. Total harvesting cost would still be limited and below 23 gt–1, in the worst case.

593

Enrico Marchi et al.

Comparing Terrain and Roadside Chipping in Mediterranean Pine Salvage Cuts (587–598)

Fig. 4 Sensitivity of harvesting cost to diesel fuel price Slika 4. Osjetljivost tro{kova pridobivanja drva s obzirom na kretanje cijene dizelskoga goriva Table 5 Impact of a surge bin on roadside chipping cost Tablica 5. Utjecaj kori{tenja lomilice na tro{kove iveranja na pomo}nom stovari{tu Surge bin Lomilica Felling & extraction rate Cijena sje~e i privla~enja Felling & extraction productivity Proizvodnost sje~e i privla~enja Felling & extraction cost Tro{kovi sje~e i privla~enja Loading cost Tro{ak utovara Total cost Ukupni tro{kovi Cost increment Uve}anje tro{kova Cost increment Uve}anje tro{kova

Yes Da

No Ne

SMH–1

664

634

gt SMH–1

66.7

60.8

gt–1

9.96

10.43

gt–1

0.15

0

gt–1

10.11

10.43

gt–1

–

0.33

%

–

+3.2

The intense use of fossil fuel remains a concern with respect to energy efficiency, even if harvesting cost is kept within reasonable bounds. Hence the energy balance drawn in Table 6. Fossil energy inputs are 554 and 423 MJ odt–1 for terrain and roadside chipping, respectively. About half of the total fossil energy inputs derive from chipping, whereas felling, bunching extraction and loading represent the other half. In any case, total fossil energy inputs account for less than 3% of the potential energy in the wood chips. Terrain chipping and roadside chipping yield 36 and 47 times the energy they use, respectively.

594

Fig. 5 Particle size distribution (%) of chips obtained with the two systems (n = 20) Slika 5. Postotni udio veli~ine ~estica ivera dobivenih na oba na~ina (n = 20) Fig. 5 shows the particle size distribution of chips produced with the two machines. The percent incidence of accepts and fines is significantly different between the two machines, according to the MannCroat. j. for. eng. 32(2011)2

Comparing Terrain and Roadside Chipping in Mediterranean Pine Salvage Cuts (587–598)

Table 6 Energy balance for terrain and roadside chipping Tablica 6. Energijska potro{nja pri iveranju u sastojini i na pomo}nom stovari{tu

Inputs Ulazni podaci Felling Sje~a Chipping & extraction Iveranje i privla~enje Loading Utovar Total Ukupno Total output Ukupni u~inak Output/Input U~inak/ulaz Input/Output % Ulaz/u~inak %

Terrain chipping Iveranje u sastojini MJ h–1 MJ odt–1

Roadside chipping Iveranje na stovari{tu MJ h–1 MJ odt–1

959

54

959

54

3516

456

12468

365

959

44

959

4

554

423

20000

20000

36.1

47.3

2.8

2.1

Note: odt = oven-dry ton Napomena: suha tvar u tonama

-Whitney non-parametric comparison test, conducted at the 5% level (accepts, p = 0.0015; fines, p = 0.0008).

Differences in the proportion of oversize particles are also large, but not statistically significant (p = 0.1208).

The smaller terrain chipper offers a superior product, with very little fines and oversize particles.

4. Discussion and Conclusions Rasprava sa zaklju~cima After years of extensive commercial trials, Tuscan operators have developed their own industrial harvesting systems, through a wise mix of Scandinavian and North American methods. The result is original, effective and much different from the Central European family of biomass harvesting techniques, popular in the Alpine regions (Stampfer and Kanzian 2006). In Tuscany, the extensive use of disc-saw feller-bunchers is borrowed from North American operations, and is justified by the large proportion of low quality biomass obtained from pine clearcuts.

On the other hand, Nordic forwarders are given preference over conventional skidders, in an attempt to reduce product contamination. The technology mix can be steered more towards the Scandinavian or the North American prototypes, depending on operational conditions.

The terrain chipping operation analyzed in this study mirrors the typical Danish system, based on a terrain chipper and a chip shuttle (Talbot and SuadiCroat. j. for. eng. 32(2011)2

Enrico Marchi et al.

cani 2005). The main adaptations consist in the adoption of a larger chipper and in the replacement of the expensive forwarder-based chip shuttle with cheaper tractor-based equivalents. The lower investment cost of the latter solution allows using two chip shuttles instead of one, thus building more buffer capacity while avoiding the extra cost of a built-in chip container on the forwarder. In turn, renouncing the integral chip container is a pre-requisite to the adoption of a bigger chipper, as the forwarder could not support both a bigger chipper and a container. The reason for using a bigger chipper is based on the type of cut: the traditional Danish system was designed for thinnings (Brenøe and Kofman 1990), whereas the Italian version is sized on clear-cuts. Of course, terrain chipping is especially dependent on favorable terrain conditions, and in particular moderate slopes and good soil bearing capacity.

The roadside chipping operation reflects North American operational philosophy, which favors processing at the landing, in an effort to achieve better scale economy. For this reason, roadside chipping has become increasingly popular also in Europe, including the Nordic Countries (Tahvanainen and Anttila 2011). Both the investment and the productivity estimated for the Tuscan operation are in line with those reported in similar North American studies (Adebayo et al. 2007, Mitchell and Gallagher 2007).

The coexistence of these two systems is most interesting. On one hand it demonstrates a good availability of woodlots, as well as a significant investment capacity. On the other it hints at a very skilful management, which can discriminate with good accuracy between different work conditions and deploy the best system for each given case. Experience and good managerial skills are also demonstrated by the use of a surge bin, which generates but a marginal benefit. Hence, the operation manager showed the capacity of fine-tuning his operation, taking the right decision even when the difference was not self-evident. (Marchi et al. 2005).

Given the right conditions, roadside chipping allowed a further reduction of harvesting cost, compared to terrain chipping. In fact, the two systems were not compared under exactly the same conditions, especially for what concerned extraction distance. This was much longer for terrain chipping.

However, a longer extraction distance might be considered as the inherent characteristic of terrain chipping, as applied in Tuscany. The use of cheaper silage trailers implies dumping on the ground and reloading onto the transportation vehicles. Hence the need for accumulating enough chips at a single landing to contain loader relocation frequency. At the same time, there is a keen interest in looking for old

595

Enrico Marchi et al.

Comparing Terrain and Roadside Chipping in Mediterranean Pine Salvage Cuts (587–598)

farm yards with a concrete floor pad, so as to reduce chip losses and chip contamination during reloading. Extending chip forwarding distance is an effective way to find an appropriate landing pad. The decision to dump on the ground is not irrational. Tuscan operators have a good knowledge of both roll-on containers and high-dumping chip bins, which could avoid the need for dumping on the ground.

However, container trucks have a high tare weight and are not the ideal way to transport chips over medium to long distances (Talbot and Suadicani 2006).

After the initial pioneering trials, today very few Italian loggers use roll-on containers to transport chips (Spinelli et al. 2007).

Filling the chip vans with a loader has been the favorite system until the introduction of the new roadside chipper. This is because a good loader can fill a chip van in 20 minutes, whereas even the largest chippers available until now would still take about 40 minutes to fill a chip van (Spinelli and Hartsough 2001). So far, separate loading has allowed a substantial reduction of truck idle time and a proportional increase of trucking capacity, for the same fleet. However, the new roadside chipper can also fill a chip van in 20 minutes, filling the gap with separate loading and making it redundant.

The harvesting systems described in this study perform best in clear-cuts, but they can also work in partial cuts, including thinning operations. They are actually used in thinnings in the same Regional Park of San Rossore, although their productivity is lower than in clear-cuts. Much depends on proper operation planning and on a tree selection pattern allowing for efficient machine traffic. This is one more reason to use forwarders rather than skidders, as the former have better maneuverability inside the stand, once loaded with tree sections.

The energy consumption for the studied operation is about 2 times higher than reported for mechanized round wood operations (i.e. 82 MJ gt–1, Athanassiadis 2000). That depends on the high energy input required by chipping. Once the comparison is made with data from other chipping operations the match is quite good (cfr. 210 – 440 MJ gt–1, Gingerich and Hendrickson 1993). The energy balance estimated in this study is also corroborated by previous studies, reporting an input-output ratio in the range of 2% (Timmons and Viteri-Mejia 2010).

Interpreting the differences in particle size distribution is made uncertain by the processing of different species. Although umbrella pine and maritime pine have similar general form and wood characteristics, they are not the same. What is sure, is that the two chippers used different knife layouts. The terrain chipper adopted a classic two-knife design,

596

with knives running the whole length of the drum.

In contrast, the roadside chipper used a four-knife design, with each knife covering half the drum length. The number of full cuts per revolution was the same, but the distribution of impacts was spread more evenly on the larger roadside machine.

5. References – Literatura Adebayo, A., Han, H., Johnson, L., 2007: Productivity and cost of cut-to-length and whole-tree harvesting in a mixed-conifer stand. Forest Products Journal 57: 59–69.

Athanassiadis, D., 2000: Energy consumption and exhaust emissions in mechanized timber harvesting operations in Sweden. The Science of the Total Environment 255: 135–143.

Bailey, A., Basford, W., Penlington, N., Park, J., Keatinge, J., Rehman, T., Tranter, R., Yates, C., 2003: A comparison of energy use in conventional and integrated arable farming in the UK. Agricultural Ecosystems and Environments 97: 241–253.

Barbero, M., Loiser, R., Quezel, P., Richardson, D., Romane, F., 1998: Pines of the Mediterranean basin. In: Ecology and Biogeography of Pinus. (Richardson, D., ed.).Cambridge University Press, Cambridge, UK. 153–170.

Bergstrand, K. G., 1991: Planning and analysis of forestry operation studies. Skogsarbeten Bulletin n. 17, 63 pp.

Berndes, G., Hoogwijk, M., Van Den Broek, R., 2003: The contribution of biomass in the future global energy supply: a review of 17 studies. Biomass and Bioenergy: 1–28.

Björheden, R., Apel, K., Shiba, M., Thompson, M. A., 1995: IUFRO Forest work study nomenclature. Swedish University of Agricultural Science, Dept. of Operational Efficiency, Garpenberg. 16 p.

Björheden, R., 2000: Integrating production of timber and energy – a comprehensive view. New Zealand Journal of Forestry Science 30: 67–78.

Björheden, R., 2008: Optimal point of comminution in the biomass supply chain. In: Suadicani, K., Talbot, B. – The Nordic Baltic Conference on Forest Operations – Copenhagen September 23–25, 2008. Forest & Landscape Working Papers No. 30–2008, 92 pp. Forest & Landscape Denmark, Hørsholm.

Blair, C., 1998: Using a chip storage bin to improve in-woods chipper efficiency and reduce chip van cycle times. FERIC technical note TN–274, 8 pp.

Brenøe, P., Kofman, P. D., 1990: Harvesting early thinnings for energy. Biomass and Bioenergy 22: 159–169.

Brockerhoff, E., Liebhold, A., Jactel, H., 2006: The ecology of forest insect invasions and advances in their management. Canadian Journal of Forest Research 36: 263–268.

Cantiani, M. G., Scotti, R., 1988: Le fustaie coetanee di pino domestico del litorale tirrenico: studi sulla dinamica di accrescimento in funzione di alcune ipotesi selvicolturali alternative. (Even-aged umbrella pine stands on the Tyrrenian coast: studies on growth dynamics as a function of alCroat. j. for. eng. 32(2011)2

Comparing Terrain and Roadside Chipping in Mediterranean Pine Salvage Cuts (587–598)

ternative silvicultural prescriptions). Annali dell’Istituto Sperimentale per l’Assestamento Forestale e per l’Alpicoltura. Vol. XI: 1–54. Trento.

Gingerich, J., Hendrickson, O., 1993: The theory of energy return on investment: A case study of whole tree chipping for biomass in Prince Edward Island. The Forestry Chronicle 69: 300–306.

Han, H., Lee, H., Johnson, L., 2004: Economic feasibility of an integrated harvesting system for small diameter trees in southwest Idaho. Forest Products Journal 54: 21–27.

Junginger, M., Faaij, A., Björheden, R., Turkenburg, W., 2006: Technological learning and cost reductions in wood fuel supply chains in Sweden. Biomass and Bioenergy 29: 399–418.

Enrico Marchi et al.

mass supply chain. Renewable and Sustainable Energy Reviews 13: 887–894.

SAS Institute Inc., 1999: StatView Reference. SAS Publishing, Cary, NC. ISBN-1-58025-162-5. pp. 84–93.

Spinelli, R., Hartsough, B., 2001: A survey of Italian chipping operations. Biomass and Bioenergy 21: 433–444.

Spinelli, R., Visser, R., 2008: Analyzing and estimating delays in harvester operations. International Journal of Forest Engineering 19: 35–40.

Spinelli, R., Mancini, L., Nati, C., Fabbri, P., 2002: Utilizzazione e recupero degli schianti da vento nelle pinete litoranee (Salvaging windblown pine stands in coastal areas). L’Italia Forestale e Montana 57: 481–498.

Kalaja, H., 1984: The example of terrain chipping system in first commercial thinning. Folia Forestalia 583. Helsinki.

Spinelli, R., Nati, C., Magagnotti, N., 2007: Recovering logging residue: experience from the Italian Eastern Alps.

Croatian Journal of Forest Engineering 28: 1–9.

Kerdelhué, C., Decroocq, S., 2006: Characterization of eight new microsatellite loci in the invading maritime pine bast scale Matsucoccus feytaudi (Hemiptera: Coccoidea: Margarodidae). Molecular Ecology Notes 6: 1168–1170.

Spinelli, R., Magagnotti, N., 2009: Logging residue bundling at the roadside in mountain operations. Scandinavian Journal of Forest Research 24: 173–181.

Kofman, P., 1993: Flishugning. Dokumentation af nuværende systemer (Chipping: documentation of innovative systems). Maskinrapport 12, Miljøministeriet, Skov- og Naturstyrelsen, Copenhagen. 39 pp (Danish, with English summary).

Marchi, E., Pesare, A., Spinelli, R., 2005: La cippatura in campo: modelli organizzativi con cippatrice semovente su base forwarder (Terrain chipping: operational systems based on forwarder-mounted terrain chippers). Sherwood – Foreste e Alberi Oggi 108: 1–6.

Mikkola, H., Ahokas, J., 2010: Indirect energy input of agricultural machinery in bioenergy production. Renewable Energy 35: 23–28.

Mitchell, D., Gallagher, T., 2007: Chipping whole trees for fuel chips: a production study. Southern Journal of Applied Forestry 31: 176–180.

Miyata, E. S., 1980: Determining fixed and operating costs of logging equipment. General Technical Report NC-55.

Forest Service North Central Forest Experiment Station, St. Paul, MN. 14 pp.

Spinelli, R., Visser, R., 2009: Analyzing and estimating delays in wood chipping operations. Biomass and Bioenergy 33: 429–433.

Spinelli, R., Magagnotti, N., Nati, C., 2009a: Options for the mechanized processing of hardwood trees in Mediterranean forests. International Journal of Forest Engineering 20: 39–44.

Spinelli, R., Magagnotti, N., Picchi, G., 2009b: Complete tree harvesting as an alternative to mulching in early thinnings. Forest Products Journal 59: 79–84.

Spinelli, R., Nati, C., Sozzi, L., Magagnotti, N., Picchi, G., 2011: Physical characterization of commercial woodchips on the Italian energy market. Fuel 90: 2198–2202.

Stampfer, K., Kanzian, C., 2006: Current state and development possibilities of wood chip supply chains in Austria.

Croatian Journal of Forest Engineering 27: 135–145.

Suadicani, K., 2003: Production of fuel chips in a 50-year old Norway spruce stand. Biomass Bioenergy 25: 35–43.

Tahvanainen, T., Anttila, P., 2011: Supply chain cost analysis of long-distance transportation of energy wood in Finland. Biomass Bioenergy 35: 3360–3375.

Panichelli, L., Gnansounou, E., 2008: GIS-based approach for defining bioenergy facilities location: a case study in Northern Spain based on marginal delivery costs and resources competition between facilities. Biomass and Bioenergy 32: 289–300.

Talbot, B., Suadicani, K., 2005: Analysis of two simulated in-field chipping and extraction systems in spruce thinnings. Biosystems Engineering 91: 283–292.

Pellizzi, G., 1992: Use of energy and labour in Italian agriculture. Journal of Agricultural Engineering Research 52: 111–119.

Timmons, D., Viteri-Mejia, C., 2010: Biomass energy from wood chips: Diesel fuel dependence? Biomass and Bioenergy 34: 1419–1425.

Picchio, R., Maesano, M., Savelli, S., Marchi, E., 2009: Productivity and energy balance in the conversion into high forest system of a Quercus cerris L. coppice in Central Italy; Croatian Journal of Forest Engineering 1: 15–26.

Väätäinen, K., Asikainen, A., Sikanen, L., Ala-Fossi, A., 2006: The cost effect of forest machine relocations on logging costs in Finland. Forestry Studies 45: 135–141.

Rentizelas, A., Tolis, A., Tatsiopoulos, I., 2008: Logistics issues of biomass: the storage problem and the multi-bio-

Croat. j. for. eng. 32(2011)2

Talbot, B., Suadicani, K., 2006: Road transport of forest chips: containers vs. bulk trailers. Forestry Studies 45: 11–22.

Zerbe, S., 2002: Restoration of natural broad-leaved woodland in Central Europe on sites with coniferous forest plantations. Forest Ecology and Management 167: 27–42.

597

Enrico Marchi et al.

Comparing Terrain and Roadside Chipping in Mediterranean Pine Salvage Cuts (587–598)

Sa`etak

Usporedba iveranja u sastojini i na pomo}nom stovari{tu pri sanitarnoj sje~i primorskoga bora U sredi{njoj je Italiji sve ve}a potra`nja za drvenim iverom, a zbog ubrzanoga {irenja primorskoga bora (Pinus pinaster) uvelike se razvijaju radni zahvati u pridobivanju drva. Nakon godina opse`nih komercijalnih poku{aja lokalni su privatnici razvili vlastiti sustav pridobivanja drva kombinacijom skandinavskih i sjevernoameri~kih metoda.

Rezultat je u~inkovit te omogu}uje odr`avanje tro{kova pridobivanja drva ispod 20 po bruto toni. U istra`ivanju je uspore|ena metoda iveranja u sastojini i iveranje na pomo}nom stovari{tu, a primijenjena je u sastojinama primorskoga bora u Toskani. U uvjetima ovoga istra`ivanja iveranje na pomo}nom stovari{tu postiglo je ve}u proizvodnost za vi{e od ~etiri puta od iveranja u sastojini. To je omogu}ilo smanjenje tro{kova pridobivanja drva za tre}inu (12,3 /bt u odnosu na 18,3 /bt). Jednom kada je ivera~ postavljen na pomo}nom stovari{tu, njime se mo`e preraditi vi{e od 100 tona svje`e mase ivera po satu, pune}i kamion iverom za manje od 20 minuta.

Prizvodnost je ostala vrlo visoka i s uklju~enim op}im vremenima rada, iako to ~ini vi{e od 30 % utro{ka vremena (iveranje + op}a vremena). Kao {to se i o~ekivalo, iveranje i privla~enje bili su najskuplji radni zahvati, ~ine}i 80 % od ukupne cijene pridobivanja drva. Na sje~u i uhrpavanje otpalo je od 12 do 17 % ukupne cijene pridobivanja drva. Utovar je imao manji, ali ipak zna~ajan utjecaj. Detaljna izra|ena studija rada i vremena omogu}ila je provjeru u~inka tro{kova u odnosu na postavljanje lomilice uz ivera~ na stovari{tu. Bez lomilice tro{kovi privla~enja, iveranja i utovara pove}ali bi se za 3,2 %, odnosno 0,33 /bt. U oba radna zahvata kori{tene su znatne koli~ine goriva, pri ~emu ivera~ na pomo}nom stovari{tu tro{i i vi{e od 100 l dizela po satu. Tro{ak goriva ~ini 31 % odnosno 36 % od ukupnih tro{kova pridobivanja drva za iveranje u sastojini odnosno na pomo}nom stovari{tu. U slu~aju pove}anja cijene goriva preko 50 % te uz polaznu pretpostavku od 1,3 /l do 2 /l, tro{kovi }e se pridobivanja drva pove}ati za 22 % za iveranje u sastojini te 25 % za iveranje na pomo}nom stovari{tu. U najnepovoljnijem slu~aju ukupni }e tro{kovi pridobivanja drva biti i dalje ispod 23 /bt. Unato~ kori{tenju dizelskoga goriva izra~unato je kako ukupni unos fosilnih goriva (energije) iznosi manje od 3 % potencijalne energije ivera. Iveranje u sastojini daje 36 puta vi{e energije, dok iveranje na pomo}nom stovari{tu 47 puta vi{e energije nego {to se energije utro{i na same radne zahvate. Navedene su metode pridobivanja drva ostvarile najbolje rezultate u ~istoj sje~i, ali se mogu primijeniti i u proredama (kao {to se i primjenjuju u regionalnom parku San Rossore), ali }e ipak proizvodnost sustava rada u proredama biti manja nego u ~istoj sje~i.

Klju~ne rije~i: biomasa, logistika, proizvodnost, ekonomija, sanitarne sje~e, Sredozemlje

Received (Primljeno): August 8, 2011 Accepted (Prihva}eno): September 5, 2011

598

Authors’ addresses – Adrese autorâ: Enrico Marchi e-mail: enrico.marchi@unifi.it Lisa Berretti e-mail: lisa.berretti@libero.it Francesco Neri e-mail: francesco.neri@unifi.it DEISTAF University of Florence Via S. Bonaventura 13 I-50145 Firenze ITALY Natascia Magagnotti e-mail: magagnotti@ivalsa.cnr.it Raffaele Spinelli e-mail: spinelli@ivalsa.cnr.it CNR IVALSA Via Madonna del Piano 10 I-50019 Sesto Fiorentino ITALY Croat. j. for. eng. 32(2011)2

Review article – Pregledni rad

Forest Fire Risk Mapping by Kernel Density Estimation Nazan Kuter, Firdes Yenilmez, Semih Kuter Abstract – Nacrtak When evaluating wildland fires, well prepared forest fire risk maps are regarded as one of the most valuable tools for forest managers, and during the production stage of these maps, association between historical fire data and other factors, such as topographic, anthropogenic and climatic, are often required. One of the most encountered problems in forest fire risk analyses is the fact that historical fire data, the dependent variable, are generally in point format, whereas other factors, the independent variables, are often expressed in areal units and available in raster format. Kernel density estimation is a widely preferred method for converting historical fire data into a continuous surface. In this study, kernel density estimate of forest fire events in the Middle East Technical University (METU) campus, in Ankara, Turkey, between 1993 and 2009, were obtained by using different bandwidth choices. Kernel density maps with regard to seasons and years were also produced and the final result was expressed as mean density value in each polygon of the study area. Actions that should be taken in high-risk areas were given on the basis of the results obtained.

Keywords: wildfire, Kernel density estimation, Kernel bandwidth, METU campus, mean random distance

1. Introduction – Uvod Forests have various crucial functions such as maintaining the balance of climate, conserving soil, water, and biodiversity; however, they have been threatened by a number of destructive factors, like droughts, insect infestations, diseases, encroachment for agricultural applications, and unplanned settlements. However, among these factors, fire events are the most important disturbance factor in forests.

Forest fires in Boreal and Mediterranean landscapes of Europe have always coexisted with human activities and been considered as one of the main disturbance agents, due to the abundance of coniferous trees and highly flammable ground vegetation, especially during hot and dry summer seasons, which favor the occurrence of fire ignition and propagation. Every year, millions of hectares of forested land are burned, which causes a wide variety of effects such as atmospheric emissions, soil erosion, biodiversity loss and drainage alterations. The use of fire for a number of activities such as grazing, agriculture and hunting has significantly modified fire regimes, primarily in the Mediterranean region. As a result of the recent increase in population density and extenCroat. j. for. eng. 32(2011)2

sive use of natural and forest regions for recreation, the number of human-caused fires has increased (Barbosa et al. 2009, Larjavaara et al. 2005, Morgan et al. 2001, Chuvieco et al. 2005).

Being located in Mediterranean climate zone, forest fires are also one of the most important factors that threaten forests in Turkey. The period from June to October is observed as high-season for forest fire events, and especially from July to August with very high temperatures, very low humidity and effective wind. In Turkey, 2 135 forest fires occurred in 2008, devastating a total area of 29 749 ha, 23 577 ha of which were forest land (JRC 2009).

An increase in the trend of occurrence of extreme natural hazards and disasters directly or indirectly related to wildland fires has been observed and fatal accidents resulted in human casualties have occurred in these events for the past several years as in the United States (2003), Canada (2002 – 2003), Greece (2000), Australia (2002 – 2003), and more recently the large fires took place in Iberian Peninsula and France (2003 – 2006). When viewed within this frame, to design and implement operational projects in order to successfully face forest fires for prevention, forecast

599

N. Kuter et al.

Forest Fire Risk Mapping by Kernel Density Estimation (599â&#x20AC;&#x201C;610)

and suppression are important priorities of fire fighting organizations (Balatsos et al. 2007).

Due to the fact that objective tools are required in pre-fire planning in order to monitor when and where a fire is more likely to occur, or when it will have more negative effects, the most important and critical part of fire prevention is the evaluation of fire risk (Chuvieco et al. 2010). Two crucial aspects of fire management determine the factors influencing the occurrence of fire and understanding the dynamic behavior of fire, and therefore a well-prepared fire risk map is needed in order to evaluate forest fire problems and make decisions on solution methods.

Wildfire risk maps would help the managers in planning the main roads, subsidiary roads, inspection paths, etc. and may lead to a reliable communication and transport system to efficiently fight small and large forest fires. Moreover, fire risk mapping definitely constitutes strategic operational advantage to develop a proper decision support system, in which the necessary actions can be taken according to their spatial and temporal priorities inside the high risk zones. Based on the risk zoning in forest fire maps, more efficient allocation of available resources can also be achieved for more effective fire prevention and suppression (Jaiswal et al. 2002, Balatsos et al.

2007).

In order to have a realistic fire risk estimate, the history of fire events in terms of occurrence, frequency and spatial distribution must be available to fire managers. World and/or national fire atlases are very useful sources of information, where the fire event is recorded by means of geographical position and other details. Especially for small fires, data related to perimeters, size and severity are mostly not included in the fire atlases, and even countries significantly affected by forest fires do not have proper data on fire incidence. Besides, a fire event is generally registered by x and y coordinates, totally loosing its surface nature and its spreading behavior. The assumption of considering each fire event as a single point process with a spot nature is currently valid in the Mediterranean fire regimes, where small dimension of fire events and broad pixel resolution of sensors do not allow a surface recording but only an event point process recognition, leading to serious positional inaccuracies since the exact location of ignition points are usually not known (Amatulli et al.

2005, Koutsias et al. 2004). Therefore, a reliable method is necessary to convert ignition point dataset into a continuous surface representation of the fire ignition density as a source of data to create fire occurrence maps and fire risk maps. Another advantage of the use of a continuous density is the possibility to integrate this information with other types of area

600

data, especially in the framework of fire risk assessment (Amatulli et al. 2005, Amatulli et al. 2007, De la Riva et al. 2004).

There are several interpolation techniques to convert data from point observations to continuous fields.

Although most techniques necessitate a variable to be estimated as a function of location, when dealing with positional uncertainties of historical fire data and converting them into a continuous surface, Kernel Density Estimation (KDE), a nonparametric method, has been widely used in many studies recently.

The study by Chuvieco et al. (2005) presented the results of the fire danger component of the Spread project, which was a European funded project for assessing fire risk conditions at several spatial scales. A GIS database covering all the EUMed countries (Portugal, Spain, Italy, Greece and Southern France) was developed within the frame of this project. In the study, KDE approach was used in order to produce continuous fire occurrence density surfaces from the location of ignition points for the representation of fire danger associated to human factors. According to the results of the study, the final product shows promising potential for helping fire managers in simulating different danger scenarios, as well as for obtaining a single evaluation of fire danger conditions for the whole EUMed area.

The spatial distribution patterns of the historical forest fires in DaXingAn mountainous area in the northeast of China were studied by Liu et al. (2010).

In the study, Ripleyâ&#x20AC;&#x2122;s K function and KDE method were combined to reveal the spatial distribution patterns of human-caused and lightning-caused forest fires. The analysis showed that this kind of spatial distribution information is very useful for forest managers to predict and manage forest fires.

In the study of Kuter and Kuter (2010), a wildfire risk map was generated for Bodrum Forest Sub-district in MuĂ°la city, located on the south-west coast in the Mediterranean region of Turkey. Different layers with associated risk ratings and weights regarding anthropogenic, topographic and vegetation factors were combined through GIS overlay analysis, and then using the historical fire data between 1999 and 2009, a continuous density map of fire events in the study area was obtained by KDE method.

KDE approach was also used in the study of Kalabokidis et al. (2007) for analyzing the wildfire dynamics of Greece, which is in the Mediterranean ecosystem. In their paper, they focused on spatial distribution of long-term fire patterns versus physical and anthropogenic elements of the environment that determine wildfire dynamics in Greece. Within a GIS environment, a spatial database, in which logistic regression and correspondence analysis were carried Croat. j. for. eng. 32(2011)2

Forest Fire Risk Mapping by Kernel Density Estimation (599–610)

out, was developed and managed. Cartographic fire data were statistically correlated with the basic physical and human geography factors in order to estimate the degree of their influence at landscape scale.

Since the original data of wildland fire ignition points, control points and livestock activities were provided as point observations, KDE method was applied in the study to transform these point observations into continuous density surfaces.

The main framework of this study is to reveal the spatiotemporal patterns of forest fire events in METU campus, which has suffered from high fire incident rate, by using KDE method. As a first step, the necessary spatial database was built by combining the historical fire data (between 1993 and 2009) and digital stand map of the study area in ArcInfo environment.

Two different bandwidth selection methods were applied; mean random distance and subjective choice. After determining the final bandwidth value, KDE maps with respect to seasons and years were produced. The final density map was converted into one that shows the mean density value in each compartment of the study area.

2. Materials and Methods – Materijal i metode 2.1 Study Area – Podru~je istra`ivanja METU campus is located at the south-west of Ankara, the capital of Turkey, and it covers an area of 40.52 km2, 30.4 km2 of which is forested, and in 1995 it was declared as the natural conservation area by the Republic of Turkey, Ministry of Culture (Fig. 1).

The campus serves about 25000 people including students, academic and administrative staff. The campus area is also important since it hosts Lake Eymir, one of the rarest natural recreational areas in the vicinity of Ankara. The dominant tree species in the campus are Pinus nigra, Pinus syilvestris, Cedrus libani, Quercus sp. and Populus sp. (METU APFF 2009). Due to high anthropogenic pressure and fire-prone tree species, forested land in the campus is a high fire risk area and as a result it was chosen as the study area.

N. Kuter et al.

Fig. 1 Location of study area Slika 1. Polo`aj podru~ja istra`ivanja positions of fire ignition points were not available, the number of ignition points in each compartment was acquired from the historical fire data. Then, the historical fire data were combined with the digital stand map in ArcInfo environment, and the number of fire ignition points in each compartment was assigned to that compartment as the attribute value.

As a next step, a random distribution of 80 ignition points was obtained in the study area by considering the mean distance to the nearest neighbor by ArcInfo.

The mean distance to the nearest neighbor is also known as »inhibition distance« and defines the distance between a point and its nearest neighbor that would be expected if the individual point observations were randomly distributed, and given by (Clark and Evans 1954): r$E =

k 2 r

(1)

2.2 Data – Podaci Stand map of the study area in digital format (in ED50 datum, UTM zone 36N projection) and forest fire data of METU campus between 1993 and 2009 were obtained from the General Secretariat of METU. According to the historical fire data, there were 80 fire ignition points in the study area in the mentioned period.

The stand map consists of 211 polygons, also called »compartments«, and even though the exact Croat. j. for. eng. 32(2011)2

Where: r density of the observed distribution expressed as the number of events per unit of area, k number of sectors in a circle of an infinite radius surrounding the point from which measurements of distance are taken, and generally a single sector is sufficient for the description and comparison of spatial relations in most natural populations.

601

N. Kuter et al.

Forest Fire Risk Mapping by Kernel Density Estimation (599–610)

2.3 KDE Calculations – Izra~un KDE (procjene Kernelove gusto}e) KDE for multivariate case is mathematically defined by (Amatulli et al. 2007, De la Riva et al. 2004, Koutsias et al. 2004, Silverman 1998): 1 n 1 (2) f$( x) = d ∑ K ( x − Xi ) nh i = 1 h Where: n number of point observations, h bandwidth, K kernel function, x a vector of coordinates that represent the location where the function is estimated, Xi vectors of coordinates that represent each point observation, d number of dimensions in space.

A kernel can be selected from a variety of functions, for example normal distribution, triangular function, quartic function, Epanechnikov, etc (De la Riva et al. 2004). In this study, Epanechnikov kernel, which is default in ArcInfo, is used and it is defined by (Silverman 1998):  1 −1 C ( d + 2)(1 − x T x) if x T x < 1 Ke(x) =  2 d (3) otherwise 0 The bandwidth directly influences the smoothness of the density function. If the bandwidth is narrow, a finer mesh density is obtained; otherwise, a smoother distribution density is produced with larger bandwidths, resulting in less variability between areas (Amatulli et al. 2007).

There are several methods to find the appropriate size of the bandwidth. One of them is the selection of the bandwidth subjectively by eye. This method is based on looking at several density estimates over a range of bandwidths and selecting the density that is the most suitable in some sense. This method begins

with the choice of a large bandwidth and then the amount of smoothing is decreased until fluctuations that are more »random« than »structural« start to appear (Wand and Jones 1995). Subjective bandwidth choices of 250 m, 500 m, 750 m, 1000 m, 1250 m, 1500 m, and 2000 m were applied in KDE of 80 ignition points.

Another method to select the kernel bandwidth is the mean random distance (RDmean) calculations, and it can be based on either local or global approach. In the former, mean polygon size and mean number of ignition points per polygon is taken into account, whereas total size of the study area and total number of ignition points is considered in the latter. RDmean is defined as (De la Riva et al. 2004): RDmean =

1 2

A N

(4)

Where: A mean polygon size, N mean number of ignition points falling inside the polygons (for local approach).

The double of RDmean was recommended as the bandwidth value in previous studies (De la Riva et al. 2004, Koutsias et al. 2004). The parameters used in bandwidth calculations and bandwidth values are given in Table 1. KDE maps of 80 fire ignition points were generated using both subjective and RDmean approaches. Firstly, different bandwidths (250 m, 500 m, 750 m, 1000 m, 1250 m, 1500 m, 2000 m) were applied (Fig. 2) in order to select the bandwidth subjectively by eye. While there were too much smoothing in higher bandwidths, lower bandwidths produced spikier kernel density maps. Therefore, 750 m was selected as the most suitable bandwidth in the study area for the subjective method. Secondly, for the mean random distance approach, the bandwidth was calculated by using local and global mean random distance calculations and both approaches yield-

Table 1 Parameters related to bandwidth calculations Tablica 1. Parametri povezani s izra~unavanjem {irine Kernelova pojasa Total size of the study area, A – Ukupna povr{ina istra`ivanoga podru~ja, A Total number of polygons – Ukupan broj poligona Total number of ignition points, N – Ukupan broj inicijalnih to~aka zapaljenja po`ara, N Mean polygon size – Srednja (prosje~na) veli~ina poligona Mean number of ignition points per polygon – Srednji (prosje~ni) broj inicijalnih to~aka zapaljenja po`ara po poligonu Local RDmean, x2 – Lokalna srednja slu~ajna udaljenost, x2 Global RDmean, x2 – Globalna srednja slu~ajna udaljenost, x2 Subjective bandwidth choices – {irina Kernelova pojasa odre|ena subjektivnom metodom

602

40.52 km2 211 80 192 045 m2 0.379 712 m 712 m 250 m, 500 m, 750 m, 1000 m, 1250 m, 1500 m and 2000 m

Croat. j. for. eng. 32(2011)2

Forest Fire Risk Mapping by Kernel Density Estimation (599â&#x20AC;&#x201C;610)

N. Kuter et al.

Fig. 2 KDE maps of campus fires for different bandwidths: a) 250 m, b) 750 m, c) 1250 m, d) 2000 m Slika 2. KDE zemljovid po`ara u kampusu za razli~ite {irine Kernelova pojasa: a) 250 m, b) 750 m, c) 1250 m i d) 2000 m ed the same bandwidth value: 712 m, which was also in compliance with the bandwidth calculated subjectively by eye. The final bandwidth value for the study area was decided to be 731 m by taking the arithmetic mean of the bandwidths found in both Croat. j. for. eng. 32(2011)2

methods. Using the final bandwidth value, KDE maps with respect to seasons and years were obtained. As a final step, the map showing the mean density value in each compartment of the study area was generated.

603

N. Kuter et al.

Forest Fire Risk Mapping by Kernel Density Estimation (599â&#x20AC;&#x201C;610)

3. Results â&#x20AC;&#x201C; Rezultati According to the seasonal KDE maps, fire events show different cluster patterns both spatially and temporally. As expected, the major portion of forest

fires took place in summer season with a percentage of 61% out of 80 ignition points (Fig. 3). This can be attributed to the increase in recreational activities around Lake Eymir and at the north-east part of the campus adjacent to residential areas. It is also notice-

Fig. 3 KDE maps according to seasons: a) fall, b) winter, c) spring, d) summer Slika 3. KDE zemljovid po`ara u razli~ito doba godine a) jesen, b) zima, c) prolje}e i d) ljeto 604

Croat. j. for. eng. 32(2011)2

Forest Fire Risk Mapping by Kernel Density Estimation (599â&#x20AC;&#x201C;610)

N. Kuter et al.

Fig. 4 Satellite image of the study area Slika 4. Satelitski prikaz istra`ivanoga podru~ja Croat. j. for. eng. 32(2011)2

605

N. Kuter et al.

Forest Fire Risk Mapping by Kernel Density Estimation (599–610)

Fig. 5 Fire ignition points with respect to years Slika 5. Pojave po`ara prema godinama able that fire events are highly concentrated around residential areas. Whereas the main campus area is located at the north, the majority of residential areas are along the eastern border of the study area, and Lake Eymir is at the south-east part (Fig. 4).

When KDE maps generated for periods between 1993 – 1996, 1997 – 2000, 2001 – 2004 and 2005 – 2009 were inspected, a gradual increase in ignition points was observed after 2000, reaching its peak between 2005 – 2009 (42.5% of 80 ignition points) (Fig. 5).

In the first period, fire events were dense in east and south-east part of the study area, and during the second period (1997 – 2000) a decrease in the intensity and a shift toward central zone were observed.

Between 2001 and 2004, the rate of fire events started to increase and had a similar spatial pattern as in the first period. Finally, in the last period (2005 to 2009), a significant spread in fire events throughout the whole study area, including the main campus area at the north, was observed (Fig. 6).

4. Concluding Remarks – Zaklju~na razmatranja In order to analyze forest fire risk and to take effective counteractions, well prepared maps indicating the association between fire and topographic, anthropogenic and climatic factors are highly demanded by fire management planners. One major drawback of using historical fire data in this kind of

606

analysis is the fact that it is in point format and also contains positional inaccuracies, whereas anthropogenic, topographic and climatic conditions are mostly expressed as continuous areal maps. KDE is a nonparametric method to obtain continuous surfaces from point observations and has been widely and effectively used for converting historical fire data into forest fire density maps. In this study KDE method was used to map the historical fire data in METU campus in order to analyze the forest fire events.

When preparing KDE maps for forest fires, mean random distance method for the calculation of bandwidth was preferred and applied by several authors such as by De la Riva et al. (2004) in Central Spanish Pre-Pyrenees and the East central Iberian range of Spain, by Koutsias et al. (2004) when estimating kernel densities of wildland fires in Halkidiki peninsula, Greece, and by Kuter and Kuter (2010) for predicting forest fire densities in Bodrum Forest Sub-district located at south-west Mediterranean coast of Turkey.

Alternative approach is the selection of bandwidth subjectively by eye, and as stated by Wand and Jones (1995) it is based on looking at several density estimates over a range of bandwidths and selecting the density that fits the study area best. This approach was also applied in the study of De la Riva et al. (2004).

The calculation of bandwidth by mean random distance method is solely based on mean polygon size and mean number of fire ignition points per polygon; however, subjective approach requires detailed a priori knowledge and expertise about the study area, which is not always possible for the analyst. These two methods were successfully utilized in this study and yielded nearly the same bandwidth values.

Based on interviews with authorities from General Secretariat of METU, it can be concluded that neither a geospatial database of fire events in the campus nor a fire risk map has been prepared, therefore as a first step, such database should be immediately built up and maintained thoroughly, and the geospatial database compiled in this study can be used as such a base. Integration of the fire density map (Fig. 7) generated via KDE with topographic and anthropogenic factors in order to obtain the forest fire risk map of the campus area should certainly be considered as a further stage of this study, and according to the results obtained from the map, fire towers and/or autonomous early warning system at critical zones in the campus area should be installed.

Especially for summer seasons, security measures should be increased around these zones and necesCroat. j. for. eng. 32(2011)2

Forest Fire Risk Mapping by Kernel Density Estimation (599–610)

sary actions should be taken in order to enhance public awareness, since a great percentage of observed fire events are human caused.

As indicated in this study, it should prove to be helpful to managers, as this type of forest fire density

N. Kuter et al.

mapping by effective use of KDE together with GIS geospatial analysis tools would enable fire departments to take appropriate precautions and build up a necessary fire-fighting infrastructure for the areas more prone to fire damage.

Fig. 6 KDE maps according to years: a) 1993 – 1996, b) 1997 – 2000, c) 2001 – 2004, d) 2005 – 2009 Slika 6. KDE zemljovid po godinama: a) 1993 – 1996, b) 1997 – 2000, c) 2001 – 2004, d) 2005 – 2009 Croat. j. for. eng. 32(2011)2

607

N. Kuter et al.

Forest Fire Risk Mapping by Kernel Density Estimation (599–610)

Barbosa, P., Camia, A., Kucera, J., Liberta, G., Palumbo, I., Ayans, J., Schmuck, G., 2009: Assessment of forest fire impacts and emissions in the European Union based on the European forest fire information system. Wildland Fires and Air Pollution, Developments in Environmental Science 8, Eds. A. Bytnerowicz, M. Arbaugh, A. Riebau and C. Andersen. Elsevier B. V., Amsterdam, The Netherlands: 197–208 p.

Chuvieco, E., Camia, A., Bianchini, G., Margaleff, T., Koutsias, N., Martinez, J., 2005: Using remote sensing and GIS for global assessment of fire danger. Proceeding of the 22nd International Cartographic Conference, Global Congresos, Coruna.

Chuvieco, E., Aguado, I., Yebra, M., Nieto, H., Salas, J., Martin, M. P., Vilar, L., Martinez, J., Martin, S., Ibarra, P., De la Riva, J., Baeza, J., Rodriguez, F., Molina, J. R., Herrera, M. A., Zamora R., 2010: Development of a framework for fire risk assessment using remote sensing and geographic information system technologies. Ecological Modelling 221(1): 46–58.

Clark, P. J., Evans, F. C., 1954: Distance to nearest neighbor as a measure of spatial relationships in populations. Ecology 35(4): 445–453.

Fig. 5 Fire ignition points with resFig. 7 Mean fire densities for compartments (KDE bandwidth = 731 m) Slika 7. Srednja gusto}a po`ara po odsjecima (KDE = 731 m)pect to years

Acknowledgments – Zahvala The authors would like to express their gratitude to the General Secretariat of METU for their data sharing and kind supports in this study.

5. References – Literatura Amatulli, G., Perez-Cabello, F., De la Riva, J., 2007: Mapping lightning/human-caused wildfires occurrence under ignition point location uncertainty. Ecological Modelling 200(3–4): 321–333.

Amatulli, G., Rodrigues, M. J., Lovreglio, R., 2005: Mapping forest fire occurrence at national level – assessing fire density by means of the adaptive kernel density technique. Proceedings of the 5th International Workshop on Remote Sensing and GIS Applications to Forest Fire Management: Fire Effects Assessment, Eds. De la Riva, J., Pérez-Cabello, F., Chuvieco, E. Universidad de Zaragoza: 51–55 p.

Balatsos, P., Kalabokidis, K., Koutsias, N., 2007: Fire risk zoning at national level in Greece: Methodological approach and outcome. 4th International Wildland Fire Conference, Seville, Spain, 10 p.

608

De la Riva, J., Perez-Cabello, F., Lana-Renault, N., Koutsias, N., 2004: Mapping wildfire occurrence at regional scale. Remote Sensing of Environment 92(3): 363–369.

Jaiswal, R. K., Mukherjee, S., Raju, K. D., Saxena, R., 2002: Forest fire risk zone mapping from satellite imagery and GIS. International Journal of Applied Earth Observation and Geoinformation 4(1): 1–10.

JRC, 2009: JRC scientific and technical reports, forest fires in Europe 2008, Report No: 9, European Commission, Joint Research Centre, Institute for Environment and Sustainability, Italy. http://effis.jrc.ec.europa.eu/download/forest-fires-in-europe-2008.pdf (Accessed 15 January 2010).

Kalabokidis, K. D., Koutsias, N., Konstantinidis, P., Vasilakos, C., 2007: Multivariate analysis of landscape wildfire dynamics in a Mediterranean ecosystem of Greece Area 39(3): 392–402.

Koutsias, N., Kalabokidis, K. D., Allgöwer, B., 2004: Fire occurrence patterns at landscape level: beyond positional accuracy of ignition points with kernel density estimation methods. Natural Resource Modeling 17(4): 359–375.

Kuter, N., Kuter, S., 2010: Wildfire risk mapping by GIS: A case study in Bodrum Forest Sub-District. The 2nd International Geography Symposium Mediterranean Environment, GEOMED 2010, Kemer, Antalya, Turkey, 196 p.

Larjavaara, M., Kuuluvainen, T., Rita, H., 2005: Spatial distribution of lightning-ignited forest fires in Finland. Forest Ecology and Management 208(1–3): 177–188.

Liu, W., Wang, S., Zhou, Y., Wang, L., Zhang, S., 2010: Spatial distribution patterns analysis of historical forest fires in DaXingAn mountains of China. International Conference on Computer Application and System Modeling (ICCASM 2010), 634–638 p.

Croat. j. for. eng. 32(2011)2

Forest Fire Risk Mapping by Kernel Density Estimation (599â&#x20AC;&#x201C;610)

N. Kuter et al.

METU APFF, 2009: Middle East Technical University Action Plan against Forest Fires. Ankara, Turkey.

Silverman, B. W., 1998: Density estimation for statistics and data analysis. Chapman & Hall/CRC, USA, 175 p.

Morgan, P., Hardy, C. C., Swetnam, T. W., Rollins, M. G., Long, D. G., 2001: Mapping fire regimes across time and space: Understanding coarse and fine-scale fire patterns.

International Journal of Wildland Fire 10(4): 329â&#x20AC;&#x201C;342.

Wand, M. P., Jones, M. C., 1995: Kernel smoothing. Chapman & Hall, UK, 212 p.

Sa`etak

Kartiranje opasnosti pojave {umskih po`ara primjenom metode KDE (Kernelova procjena gusto}e) Pri procjeni je opasnosti pojave {umskih po`ara na odre|enom {umskom podru~ju vrlo va`no posjedovati kvalitetan zemljovid rizika nastanka {umskih po`ara. Takav se zemljovid, u pravilu, temelji na {to potpunijoj bazi podataka o {umskim po`arima na odre|enom podru~ju u pro{losti te naj~e{}e topografskim, klimatskim, antropogenim i ostalim podacima. Pri tome nailazimo na problem da su povijesni podaci o {umskim po`arima (ovisna varijabla) uglavnom u linijskom (to~kastom) formatu, dok su svi ostali podaci (neovisne varijable) prikazani u rasterskom (povr{inskom) formatu. Stoga se za pretvorbu povijesnih podataka o {umskim po`arima u prikaz na kontinuiranim povr{inama vrlo ~esto primjenjuje metoda KDE. Istra`ivanje je provedeno na podru~ju kampusa Srednjoisto~noga tehni~koga sveu~ili{ta u Ankari u Turskoj (METU) (slika 1 i slika 4). Kampus ima povr{inu od 40,52 km2, od ~ega je 30,4 km2 {umskoga podru~ja 1995. godine progla{eno za{ti}enim podru~jem prirode. U kampusu `ivi i radi oko 25 000 ljudi (studenti, profesori i administrativno osoblje), a u njegovu se sastavu nalazi i jezero Eymir, jedno od najvrednijih rekreacijskih podru~ja u blizini Ankare. Me|u drve}em prevladavaju crni bor, bijeli bor, libanonski cedar, vrste hrasta (Quercus sp.) i topole (Populus sp.).

Podaci o sastojini i stani{tu su, kao i podaci o {umskim po`arima na istra`ivanom podru~ju u razdoblju 1993 â&#x20AC;&#x201C; 2009 godine, preuzeti od Uprave METU-a. Navedeni su podaci za svaki od 211 poligona (odsjeka) bili u digitalnom obliku. Na istra`ivanom je podru~ju, u navedenom razdoblju, zabilje`eno 80 {umskih po`ara. Inicijalne to~ke zapaljenja po`ara nisu bile definirane svojim to~nim polo`ajem, ve} su bile povezane uz konkretan odsjek.

Pri izra~unu KDE je izme|u razli~itih funkcija (distribucija) kori{tena Epanechnikova metoda (formula 3).

Jedan od va`nih parametara matemati~ke definicije KDE je Kernelova {irina pojasa (formula 2). Postoje razli~ite metode odre|ivanja najpovoljnije {irine Kernelova pojasa. U ovom su istra`ivanju kori{tene tzv. subjektivna metoda odabira i temeljem nje je odre|ena {irina pojasa od 250, 500, 750, 1000, 1250, 1500 i 2000 m te metoda srednje slu~ajne udaljenosti (formula 4) primijenjena na lokalnoj i na globalnoj razini.

Subjektivnom je metodom kao najpovoljnija {irina Kernelova pojasa odabrana ona od 750 m, dok je metoda srednje slu~ajne udaljenosti (koja se temelji samo na srednjoj povr{ini poligona odnosno odsjeka i na srednjoj vrijednosti broja po`ara po poligonu odnosno odsjeku), i na lokalnoj i na globalnoj razini, rezultirala {irinom Kernelova pojasa od 712 m. U kona~nici je izra~unata aritmeti~ka sredina {irina pojasa odre|enih objema metodama i ona iznosi 731 m. Svi su parametri povezani s izra~unavanjem {irine Kernelova pojasa prikazani u tablici 1.

Na bazi kona~ne {irine Kernelova pojasa izra|eni su zemljovidi KDE u ovisnosti o godini (razdoblju) i godi{njem dobu (sezoni) nastanka {umskih po`ara (slika 3 i slika 6). Zavr{ni je korak bio generiranje zemljovida koji prikazuje srednje vrijednosti gusto}e {umskih po`ara u svakom odsjeku (slika 7).

Zemljovid {umskih po`ara KDE u razli~ito doba godine prikazuje sezonsko pojavljivanje {umskih po`ara i prostorno i vremenski. Glavnina je {umskih po`ara (61 %), prema o~ekivanju, sezonska, vezana uz ljeto, a prostorno za okolicu jezera Eymir (rekreacijske aktivnosti) te uz sjeveroisto~ni dio kampusa koji grani~i s gu{}e naseljenim podru~jima. Promatrano po vi{egodi{njim razdobljima najvi{e je {umskih po`ara bilo izme|u 2005. i 2009. godine (42,5 %), a najmanje u razdoblju od 1993. do 1997. godine (slika 5).

Tijekom ovoga istra`ivanja pripremljene su baze prostornih podataka koje, u suradnji s upravom METU-a, treba doraditi, dopuniti i nastaviti odr`avati te koristiti kao osnovu u protupo`arnoj za{titi kampusa. Zemljovid gusto}e {umskih po`ara po odsjecima, u daljnjim istra`ivanjima, treba dopuniti topografskim i antropogenim ~imbenicima radi kreiranja zemljovida ugro`enosti podru~ja kampusa {umskim po`arima. Takav bi zemljovid bio

Croat. j. for. eng. 32(2011)2

609

N. Kuter et al.

Forest Fire Risk Mapping by Kernel Density Estimation (599–610)

dobra podloga za izlu~ivanje najugro`enijih podru~ja, planiranje {umske protupo`arne infrastrukture (prometne i druge), odabir optimalnih mjesta za postavljanje protupo`arnih osmatra~nica i/ili razli~itih sustava brzoga otkrivanja {umskih po`ara i njihove dojave. U spomenutim bi podru~jima, poglavito tijekom ljeta, trebalo intenzivirati protupo`arnu preventivu te poraditi na edukaciji ljudi jer obi~no ljudi uzrokuju velik broj {umskih po`ara.

Authors’ address – Adresa autorâ: Asst. Prof. Nazan Kuter, PhD. e-mail: nkuter@karatekin.edu.tr Çankýrý Karatekin University Department of Landscape Architecture 18200 Çankýrý TURKEY Research Assistant Firdes Yenilmez, MSc. e-mail: firdes@metu.edu.tr Middle East Technical University Department of Environmental Engineering 06531 Ankara TURKEY

Received (Primljeno): July 06, 2010 Accepted (Prihva}eno): March 18, 2011

610

Instructor Semih Kuter, MSc. e-mail: semihkuter@karatekin.edu.tr Çankýrý Karatekin University Department of Forest Engineering 18200 Çankýrý TURKEY Croat. j. for. eng. 32(2011)2

ISSN 1845-5719

9 771845 571000