GBP_2007_03_AnnRpt_2006 by fRI Research

FOOTHILLS MODEL FOREST GRIZZLY BEAR RESEARCH PROGRAM 2006 ANNUAL REPORT

Prepared and edited by Gordon B. Stenhouse and Karen Graham March 2007

Disclaimer This report presents preliminary findings from the 2006 research program within the FMF grizzly bear research program. It must be stressed that these data are preliminary in nature and all findings must be interpreted with caution. Opinions presented are those of the authors and collaborating scientists and are subject to revision based on the ongoing findings over the course of this study.

Suggested citation for information within this report: e.g. Larsen T. and Karine Pigeon 2006, The diet of grizzly bears (Ursus arctos) in NorthCentral Alberta, Canada, part of a new food-based model approach to mapping grizzly bear habitat In: Stenhouse, G. and K. Graham (eds). Foothills Model Forest Grizzly Bear Research Program 2006 Annual Report. Hinton, Alberta. 87 pp. This is an interim report not to be cited without the express written consent of the senior author.

ACKNOWLEDGEMENTS A program of this scope and magnitude would not be possible without the dedication, hard work and support of a large number of people. We would like to thank the capture crew members: Bernie Goski, Dave Hobson, Terry Larsen, Rick Booker, Jay Honeyman Saundi Stevens and Codi Schmidt. Thanks also to Fish and Wildlife officers and biologists for all their help during the capture season: Dwayne Matier, Dave Robertson, Byron Schram, Dave Stepnisky, Simon Tatlow, Lewis Watson and Chris Watson. A special thanks to Ken Schmidt. Exemplary flying (and field!) skills were provided by John Saunders of Peregrine Helicopters of Hinton and fixed wing pilot Mike Dupuis of Wildlife Observation Air Services. Thanks also to the veterinarians who assisted with the captures: Marc Cattet, Nigel Caulkett, Erin Geymonat, and Johan Lindsjo. Appreciation is also extended to the vegetation plot crewmembers Terry Larsen, Karine Pigeon, Rebecca Vaughan, Cody Schmidt and Jennifer Cave whose hard work and enthusiasm ensured a successful field season. A huge thank you to Cliff Henderson for his assistance with aircraft support and Alberta Sustainable Resource Development for their help with logistical support and camp accommodations. Also a big word of thanks to Bernie Morin of Canfor Corporation for assisting with field staff accommodations. The Grizzly Bear Research Program Steering Committee and the Board of Directors of the Foothills Model Forest provided valuable support and assistance that allowed the research to proceed in order to address management needs and we thank Don Podlubny and Jim LeLacheur for this. The financial support of our many program partners allowed us to focus our attention on the delivery of the program goals within this multidisciplinary program. Thanks to Julie Duval, Melissa Pattison, and Debbie Mucha for their help in all areas relating to GIS and a special thanks to Jerome Cranston for all his GIS expertise. A word of praise goes out to Lisa Jones and Greg Nelson for keeping up with media needs and the special communication requirements associated with our program. Thank you to Sheri Fraser, Fran Hanington, Denise Lebel, Judy Astelos and especially Angie Larocque for an excellent job in managing the administrative details of this program. Definitely not an easy task! The staff at the Hinton Training Centre provided a great deal of assistance in many ways this year including food and lodging for field crews during their short stays. John Boulanger provided superb statistical advice throughout the year. Dr. David Paetkau at Wildlife Genetics International completed the lab work on all DNA hair samples and Matsonâ&#x20AC;&#x2122;s Lab conducted our tooth aging. This program would not have been possible without the many program sponsors (See Appendix 3).

TABLE OF CONTENTS ACKNOWLEDGEMENTS .............................................................................................. i TABLE OF CONTENTS ................................................................................................. ii LIST OF TABLES ............................................................................................................ v LIST OF FIGURES ........................................................................................................ vii BACKGROUND ............................................................................................................... 1 INTRODUCTION............................................................................................................. 1 CHAPTER 1. GRIZZLY BEAR CAPTURE AND COLLARING - 2006 FIELD SEASON............................................................................................................................. 3 Introduction ..................................................................................................................... 3 Study Area ...................................................................................................................... 3 Methods........................................................................................................................... 4 Summary of Captures ..................................................................................................... 5 Capture Locations ....................................................................................................... 5 Sex and Age Characteristics ....................................................................................... 6 Capture Type ............................................................................................................... 6 Telemetry .................................................................................................................... 7 Status of Captured Grizzly Bears ................................................................................ 8 Capture Related Mortalities ........................................................................................ 8 Black Bears ................................................................................................................. 8 Grizzly bear vs. black bear captures ........................................................................... 9 Other Non-Target Species......................................................................................... 13 CHAPTER 2: REMOTE SENSING AND HABITAT MAP PRODUCTION 2006 . 14 Introduction ................................................................................................................... 14 Methods......................................................................................................................... 14 Study Area ................................................................................................................ 14 Field Sampling .......................................................................................................... 15 Airborne Mission ...................................................................................................... 15 Satellite Image Acquisition and Pre-Processing ....................................................... 16 Map Updates and Change Detection......................................................................... 16 Land Cover................................................................................................................ 18 NDVI Phenology ...................................................................................................... 18 Crown Closure and Species Composition................................................................. 19 Results and Discussion ................................................................................................. 19 Map Updates and Change Detection......................................................................... 19 Land Cover................................................................................................................ 20 NDVI Phenology ...................................................................................................... 22

Crown Closure .......................................................................................................... 23 Species Composition................................................................................................. 24 Summary ....................................................................................................................... 25 CHAPTER 3. SEASONAL GRIZZLY BEAR HABITAT FOR SIX POPULATION UNITS OF ALBERTA, CANADA ................................................................................ 27 Introduction ................................................................................................................... 27 Methods......................................................................................................................... 27 Defining grizzly bear habitat use .............................................................................. 27 Defining female grizzly bear range (occupied versus unoccupied habitat) .............. 28 Defining final grizzly bear habitat and validation .................................................... 29 Results ........................................................................................................................... 29 Grizzly bear habitat use ............................................................................................ 29 Female grizzly bear range (occupied versus unoccupied habitat) ............................ 29 Final grizzly bear habitat map................................................................................... 30 Discussion ..................................................................................................................... 30 CHAPTER 4: THE DIET OF GRIZZLY BEARS (URSUS ARCTOS) IN NORTHCENTRAL ALBERTA, CANADA, PART OF A NEW FOOD-BASED MODEL APPROACH TO MAPPING GRIZZLY BEAR HABITAT ...................................... 43 Abstract ......................................................................................................................... 43 Introduction ................................................................................................................... 43 Methods......................................................................................................................... 44 Study Area ................................................................................................................ 44 Field Collection......................................................................................................... 45 Laboratory Analysis .................................................................................................. 45 Data Analysis ............................................................................................................ 45 Results ........................................................................................................................... 46 Swan Hills ................................................................................................................. 46 Wapiti ........................................................................................................................ 48 Hinton ....................................................................................................................... 48 Chinchaga ................................................................................................................. 48 Simonette .................................................................................................................. 49 Discussion ..................................................................................................................... 49 Sampling Bias and Error ........................................................................................... 49 West-central Alberta ................................................................................................. 49 Swan Hills ................................................................................................................. 50 Wapiti ........................................................................................................................ 51 Seasonal Food Selection ........................................................................................... 51 Appendix ....................................................................................................................... 54 CHAPTER 5: ANIMAL HEALTH ............................................................................... 55 Objectives ..................................................................................................................... 55 Animal Health Update .................................................................................................. 55 Laboratory work........................................................................................................ 55 Grizzly Bear Health Assessment .............................................................................. 56

iii

Chapter 6. KNOWLEDGE TRANSFER AND PRODUCT DELIVERY ................. 58 Objectives ..................................................................................................................... 58 Knowledge Transfer and Product Delivery .................................................................. 58 APPENDIX 1: RECENT PUBLICATIONS/REPORTS FROM THE FMF GRIZZLY BEAR RESEARCH PROGRAM ............................................................... 60 APPENDIX 2: PUBLISHED PAPERS AND THESES RESULTING FROM THE FMFGBP – JANUARY 2006. ........................................................................................ 62 APPENDIX 3: FMF GRIZZLY BEAR PROJECT PARTNERS ............................. 66

LIST OF TABLES Table 1. Number of collars allotted to each capture area. .................................................. 5 Table 2. Grizzly bears captured in each capture area. ........................................................ 5 Table 3. Sex and age of captured grizzly bears. ................................................................. 6 Table 4. Captured female grizzly bears with cubs. ............................................................. 6 Table 5. Grizzly bear capture types. ................................................................................... 7 Table 6. Status of 2006 research grizzly bears as of December 2006. ............................... 8 Table 7. Black bear captures by sex and age classifications............................................... 9 Table 8. Number of grizzly bears vs. black bears captured by year. .................................. 9 Table 9. Confusion Matrix for the 2005 Change Update Layer ....................................... 20 Table 10. Dates used to define seasonal habitat use. ....................................................... 32 Table 11. Environmental variables used to predict the relative probability of seasonal habitat use by grizzly bears in six population units of Alberta, Canada. .................. 32 Table 12. Habitat coefficients ( y of habitat use within home ranges by population unit for the hypophagia period (season 1: 1 May to 15 June). ................................................................................................ 33 Table 13. Habitat coefficients ( bility of habitat use within home ranges by population unit for the early hyperphagia period (season 2: 16 June to 31 July). .................................................................................. 34 Table 14. Habitat coefficients ( relative probability of habitat use within home ranges by population unit for the late hyperphagia period (season 3: 1 August to 15 October)........................................................................... 35 Table 15. Coefficients describing the probability of female grizzly bear occupancy in population units 1 to 6. Model significance (LR 2, df = 8) of 988.5 (p<0.001) and percent deviance explained of 50.8. Natural sub-region categories indicated with avoid had perfect avoidance within the study area and were assumed a probability of 0................................................................................................................................. 36 Table 16. Predictive accuracy of seasonal population unit habitat maps based on spearman rank correlations (rs) of habitat bin and area-adjusted frequency of use per habitat bin. Results reported by training (within sample) and testing (withheld outof-sample) datasets.................................................................................................... 37

Table 17. Summary of grizzly bear scats collected and analysed from north-central Alberta in 2005 and 2006. ......................................................................................... 46 Table 18. The seasonal importance of the percent volume of digestible dry matter from major food items consumed by grizzly bears in the Hinton, Chinchaga, and Grande Cache study areas in Alberta, Canada, 2006............................................................. 49 Table 19. A comparison between the percent volume of digestible dry matter from major food items consumed seasonally by grizzly bears in the Swan Hills, Alberta, Canada, in 2005 and 2006. ........................................................................................ 54 Table 20. The percent volume of digestible dry matter from major food items consumed seasonally by grizzly bears in the Wapiti study area in Alberta, Canada, 2006. ...... 54

LIST OF FIGURES Figure 1. Locations of Capture Areas. ................................................................................ 4 Figure 2. Total snare sites by year. ................................................................................... 10 Figure 3. Snare sites with known installation and removal dates by year. ....................... 10 Figure 4. Number of days snare sites were active ............................................................ 11 Figure 5. Grizzly bear capture rates per day. .................................................................... 11 Figure 6. Total black bear capture rate per day. ............................................................... 11 Figure 7. Total bears capture rate per day......................................................................... 11 Figure 8. Grizzly bear population units per day. .............................................................. 12 Figure 9. Average grizzly bear captures ........................................................................... 12 Figure 10. Average black bear captures per day ............................................................... 13 Figure 11. Average total bears captures per day ............................................................... 13 Figure 12. Location of the Phase 6 study area in Western Alberta, including the phase 6 extension area............................................................................................................ 15 Figure 13. Airborne sensor flight transect. ....................................................................... 16 Figure 14. Temporal resolution of the Phase 5 orthomosaic. ........................................... 17 Figure 15. Classification decision rules with suggested starting values. .......................... 18 Figure 16. Disturbance and regeneration in the Phase 6 study area. ................................ 20 Figure 17. Land cover map for the Phase 6 study area. .................................................... 21 Figure 18. 16-day NDVI phenology composites for the 2005 growing, Phase 6 study area. ................................................................................................................................... 22 Figure 19. Continuous crown closure map. ...................................................................... 23 Figure 20. Continuous species composition map (% conifer). ......................................... 24 Figure 21. Grizzly bear population units of Alberta. Habitat maps were produced for population units 1/2 (Waterton and Livingstone), 3 (Clearwater), 4 (FMF core), 5 (Grande Cache), and 6 (Swan Hills). Only the Alberta North (unit 7) remains unclassified. .............................................................................................................. 38

vii

Figure 22. Edge distance index used for assessing and predicting the attraction of edge habitats by grizzly bears............................................................................................ 39 Figure 23. Model diagram illustrating the process of determining final grizzly bear habitat values for individual population units and seasons (re-parameterized and repeated for each season and population unit combination). .................................... 40 Figure 24. Predicted range of female grizzly bears (occupancy of home range) for six grizzly bear population units of Alberta. This map was used as a scalar to ensure extrapolated habitat values were consistent with occupancy of the area. ................. 41 Figure 25. Predicted grizzly bear habitat rank during late hyperphagia (1 August to 15 October) for six grizzly bear population units of Alberta, Canada. .......................... 42 Figure 26. Map of the study area boundaries in Alberta; primary highways and major cities displayed. ......................................................................................................... 44 Figure 27. Map of the study area boundaries in Alberta relative to natural subregions; primary highways and major cities displayed. .......................................................... 44 Figure 28. Percent volume of important food items based on seasonal changes in grizzly bear diet in the Swan Hills Study Area, Alberta, 2005. ............................................ 47 Figure 29. Percent volume of important food items based on seasonal changes in grizzly bear diet in the Swan Hills Study Area, Alberta, 2006. ............................................ 47 Figure 30. Percent volume of important food items based on seasonal changes in grizzly bear diet in the Wapiti Study Area, Alberta, 2006. ................................................... 48 Figure 31: Map area where landcover mapping has been completed circa 2005 conditions. ................................................................................................................. 59

viii

BACKGROUND The challenge facing land managers is to learn how to ensure the long-term survival of grizzly bears while addressing human and societal demands on the same land base. If we are to sustain both human use activities and grizzly bears, intensive management based on sound biological information is required. In 1999, the Foothills Model Forest (FMF) initiated a co-operative, international, multidisciplinary, six-year grizzly bear research program in west-central Alberta to address these concerns. The primary goal of the research is to assess grizzly bear populations, bear response to human activities, and habitat conditions in order to better understand the requirements of this species and integrate those requirements into the land management decision-making framework. The research addresses important management questions for which data are required. Many of the activities currently underway within this research program are directly linked to the 2005 Provincial Grizzly Bear Recovery Plan (draft). An important outcome of this program is the ongoing development of tools and techniques that address landscape level conservation issues, which is a critical component to the successful management of grizzly bear populations throughout Alberta and North America. The Foothills Model Forest Grizzly Bear Research Project was designed to take a holistic approach to questions concerning grizzly bear response to human activities and landscape conditions. In this regard the project has a series of key program elements that are directly linked to one another. Each program element is the responsibility of research collaborators from a variety of disciplines. Not all these research collaborators are grizzly bear biologists but each researcher brings a unique perspective to the research program objective and makes a significant contribution to the overall team in providing a necessary data component.

INTRODUCTION In 2006 the FMF Grizzly Bear Project began a new phase of our ongoing program. The new work we embarked on this year was made possible by the support of two new major grants. One of these was from Alberta Innovation and Science and the second grant was from the NSERC/CRD program. Both of these grants required the ongoing support of our industry partners who have continued to provide the needed support for our work. Without this combined support from all the program sponsors it would not be possible to meet the ambitious program objectives we have established. This year our research team embarked on the expansion of our previous efforts in two key areas: 1.

The status evaluation of grizzly bear health at both the population and individual level. This work included significant new work in laboratories at two Canadian universities. The spatial delineation of landscape structure and change to match existing grizzly bear movement data from 1999-2006.

The focus of these research areas is to try to gain an understanding of the relationships between grizzly bear health and landscape condition and change.

While this work was underway our research team continued with the ongoing development of important research products and tools to assist program partners in making land use management decisions in grizzly bear habitat in Alberta. Some of these products and tools include:  New remote sensing mapping products and procedures allowing the creation of grizzly bear habitat maps for large landscape areas. This year we have completed new products for an area of approximately 227,000 km2, which covers an area along the eastern slopes of the Rocky Mountains from Grande Prairie south to the Montana border.  New validated resource selection function (RSF) models that build on the remote sensing mapping products and identify important grizzly bear habitats on the landscape. These models provide a map that identifies the spatial distribution of high quality grizzly bear habitat. This year we have prepared new versions of these RSF products that are linked to the grizzly bear population units we have identified through genetic analysis. These new models have used the most current and comprehensive GPS movement data from collared animals (1999-2006).  New models, using graph theory analysis, which identifies grizzly bear movement corridors across the landscape. In 2006 this work has focused on completing the area south of Highway 1 south to the Montana border.  New techniques to monitor and assess grizzly bear health. This year our research team has focused on new approaches to quantify and synthesize existing grizzly bear health data as well as to develop new techniques to identify and measure chronic stress.  New procedures and techniques for the capture and handling of grizzly bears for research and management purposes. These procedures are now being adopted as new leading edge standards for grizzly bear handling in Alberta and other jurisdictions in North America. Our long term-term project objective remains the same: To provide knowledge and planning tools to land and resource managers to ensure the long-term conservation of grizzly bears in Alberta. The status and management of grizzly bears in Alberta remains a management concern and priority within the Department of Sustainable Resource Development and for companies and organizations concerned about sustainable forest and land management. Given the profile of this species, the wise management of grizzly bears is also a serious concern to industry and environmental groups in Alberta. The research results are crucial in supporting sustainable resource development activities in Alberta. These results will allow industry to develop best management practices in grizzly bear habitat and minimize impacts on this important species that is often viewed as an indicator of eco-system health. We believe that Albertan’s expect that we will continue to use the best science and information available to manage the resources we are fortunate to have in this province.

CHAPTER 1. GRIZZLY BEAR CAPTURE AND COLLARING 2006 FIELD SEASON Dave Hobson, Jerome Cranston and Gordon Stenhouse. Foothills Model Forest, Hinton, Alberta.

Introduction The 2006 grizzly bear spring capture session was the 8th conducted by the Foothills Model Forest Grizzly Bear Research Project. The original study area encompassed 10,000 km2 in an area south of Highway 16, between Edson and Jasper in the north and the Brazeau River in the south. Since 2003 the study area has expanded to include all of the grizzly bear range between the Berland River and the Montana border (62% of grizzly bear range in Alberta). In 2005, the study area expanded again to include areas between the Berland and Wapiti Rivers plus the Swan Hills. In 2006, the study area expanded northwards to include the Chinchaga River, The Hotchkiss River and the Meikle River areas. We also included the area south of the Wapiti River and returned to the Swan Hills. Concurrent with the capture program, resource selection function (RSF) maps have and are being developed and refined for this entire area as new and additional data is collected. These maps will serve as a surrogate for habitat suitability maps and will define the landscape in terms of probability of grizzly bear occurrence. Spatial data generated by radio-collared grizzly bears captured during this session will be used to verify the accuracy of the RSF maps generated in 2005, and to develop new RSF maps for this yearâ&#x20AC;&#x2122;s expansion area. Tooth, blood and tissue samples collected from captured grizzly bears will also be used to help understand population dynamics, assist in defining population units through genetic analysis, and track population health. The goal of this years capture session was to deploy 20-25 GPS radio-collars on grizzly bears in the study areas. This report summarizes the results of the 2006 capture session and includes number of bears captured, location of captures, and sex and age ratios. Study Area Three individual capture areas were designated within the 2006 study area (Figure 1). These capture areas were as follows: 1. The Chinchaga capture area (Chinchaga, Hotchkiss and Meikle Rivers). 2. The Wapiti River capture area. 3. The Swan Hills capture area.

The 2006 study area comprised parts of Bear Management Areas (BMAs) 1 (Chinchaga), 2A (Wapiti), 2B (Wapiti and Swan Hills) and 3A (Swan Hills). These BMAs make up approximately 66% of grizzly bear range in Alberta. The population estimate for these BMAs, as of 2002, based on Alberta Fish and Wildlife Division modelling, was 476 grizzly bears. This was 49% of the estimated Alberta population in 2002. Between 1988 and 2002, a total of 1836 grizzly bear licences were issued and 106 grizzly bears were harvested from these 4 BMAs. Hunter success over these years averaged 5.8%. Methods Capture efforts were conducted by 2 helicopter-based crews. Each crew consisted of biologists with experience in grizzly bear capture and a Veterinarian. Crews moved into new capture areas when the allotted collars were deployed (Table 1) or when it was deemed that enough time had been spent in the area. Capture efforts ended on 22 June 2006.

Figure 1. Locations of Capture Areas.

Table 1. Number of collars allotted to each capture area. Capture Area Proposed Collar Distribution Chinchaga 9 Wapiti 6 Swan Hills 6 Bears were immobilized using a drug combination of Telazol and xylazine (XZT). The drugs were administered by rifle/pistol once the bear had been restrained in a snare. Atipamazole was used to reverse the xylazine. Once immobilized, grizzly bears were weighed, and measured (chest girth, zoological length, and straight-line length). Samples were collected (blood, hair, ear plug tissue and tooth). Radio-collar and ear tag transmitters were attached. Vital functions and bloodoxygen levels were monitored during the processing. A wildlife veterinarian was present at most bear captures. Black bears and other non-target species were marked with ear tags and released from the snare after immobilization. Vital conditions were monitored while under anaesthesia but measurements and samples were not collected. After administering the reversal, all bears were monitored until they became mobile. As per the capture permit conditions, all captured bears were checked within 24 hours of capture to ensure that they had recovered from immobilization. Summary of Captures Capture Locations In total, 15 grizzly bears and 38 black bears were captured. Of the grizzly bear captures, 2 were caught in the Chinchaga capture area, 6 in the Wapiti capture area, and 7 in the Swan Hills capture area (Table 2). In some cases (G212, G214, G233, and G234), the captured grizzly bears were too small for radio-collars. These bears were tagged with eartag transmitters only. Table 2. Grizzly bears captured in each capture area. Capture Area Chinchaga Wapiti Swan Hills

Captured Grizzly Bear IDs G240, G241 G228, G233, G234, G235, G236, G237 G200, G204, G205, G211, G212, G213, G214

Sex and Age Characteristics Of the 15 grizzly bears captured, 7 (47%) were adults, 6 (40%) were sub-adults, and 2 (13%) were yearlings. Eleven (73%) were males and 4 (27%) were females (Table 3). Sub-adult males made up the largest component of captured grizzly bears (33%) while adult males, adult females, sub-adult females and yearling males comprised 27%, 20%, 7% and 13% of captured grizzly bears respectively. Table 3. Sex and age of captured grizzly bears. Grizzly Bear IDs G228 G233 G234 G235 G236 G237 G240 G241 G200 G204 G205 G211 G212 G213 G214

Area Wapiti Wapiti Wapiti Wapiti Wapiti Wapiti Chinchaga Chinchaga Swan Hills Swan Hills Swan Hills Swan Hills Swan Hills Swan Hills Swan Hills

Age 10 Sub-adult Sub-adult 7 14 2 3 3 3-4 3 12-13 12 Yearling 15 Yearling

Sex Male Male Male Male Female Male Male Male Male Female Female Female Male Male Male

Of the 3 captured adult females, all were either observed with cubs or were lactating (Table 4). Table 4. Captured female grizzly bears with cubs. Grizzly Bear IDs G205 G211 G236

Cubs Lactating 3 yearling cubs Lactating

Capture Type Capture types were categorized as ground capture with snare and ground capture with culvert trap. Most ground capture sites used snares in 3 different formats: pail sets, cubby sets and/or trail sets. A Culvert trap was used in only 1 ground site. Of the 16 capture events (G241 was captured twice) (Table 5), helicopter captures accounted for 6.25% (1) of captures events and ground captures accounted for 93.75% (15) of capture events. Fourteen ground-capture events involved snares and 1 involved a culvert trap.

Table 5. Grizzly bear capture types. Grizzly Bear IDs G228 G233 G234 G235 G236 G237 G240 G241 G241 2nd capture G200 G204 G205 G211 G212 G213 G214

Capture Type Aerial Snare Snare Snare Snare Snare Snare Culvert Snare Snare Snare Snare Snare Snare Snare Snare

Telemetry Ten radio-collars were deployed. Four captured grizzly bear were not radio-collared due to the bearâ&#x20AC;&#x2122;s small size. Fourteen grizzly bears were tagged with an ear-tag transmitter, some with both a regular, battery powered ear-tag transmitter and an experimental, solar powered ear-tag transmitter. Radio-collars deployed consisted of 2 types, Tellus (new Televit model), and Telonics GPS/Argos. Tellus collars collect locations on the following schedule: April 1 to November 30. 1 location/hour. December 1 to March 31. 1 location/day. The Telonics collar collects GPS coordinates, which are stored onboard. Six locations are transmitted through an Argos satellite to a ground station on a schedule. The information is forwarded via e-mail. The location collection and data transfer schedule is as follows: March 1 to October 31, 2006: 1 location every 2 hours. November 1, 2006 to March 15, 2007: No locations. March 16, 2007 to end of battery life: 1 location every 2 hours. Data transfer via Argos satellite twice every 7 days during period of location collection. Four hours at 1200h UTC and 4 hours at 2200h UTC. All radio-collars were outfitted with a timed release or remote release mechanism with a rot-off as a backup.

Status of Captured Grizzly Bears Table 6 lists status of radio-telemetered bears as of mid-September 2006.

Table 6. Status of 2006 research grizzly bears as of December 2006. Grizzly Bear IDs G200 G204 G205 G211 G212 G213 G214 G228 G233 G234 G235 G236 G237 G240 G241

Fate as of August 2006 Collar still on and collecting data Collar still on and collecting data Collar came off, bear still alive Collar still on and collecting data Still with G211 Collar came off, bear still alive Still with G211 Not relocated since August, still alive in July Not relocated in August, still alive in July Not relocated in August, still alive in July Collar came off 4 June 06; bear alive at the time Collar still on and collecting data Collar still on and collecting data Collar failed, bear not relocated since July Not relocated in August; still alive in July

Capture Related Mortalities A grizzly bear sow was killed when she charged the capture crew in the Wapiti study area. She had two 2-year-old cubs in snares. Enforcement Field Services investigated the incident. Field staff completed and submitted a full incident report on the day of the occurrence. A black bear yearling died under anaesthesia in the Wapiti study area. A post-mortem examination was conducted, but cause of death remains undetermined. Black Bears A total of 38 black bears were captured (18 in the Chinchaga area, 16 in the Wapiti area and 4 in the Swan Hills area; Table 7). Cementum analyses were not conducted on black bears during this study. Adult males constituted the largest number of captured black bears (37%). adult females, subadult females and subadult males comprised 18%, 21% and 13% of captured black bears respectively. Male and female yearlings constituted 5% each of the totals. No cubs of the year were captured. Male/female percentages of captured black bears were 63/37 and adult/subadult/yearling percentages were 55/34/11.

Table 7. Black bear captures by sex and age classifications.

Sex Male Female Total Male Female Total Male Female Total

Chinchaga area Cub of the Year Yearling Sub-adult 0 1 5 0 1 2 0 2 7 Wapiti area 0 1 3 0 1 2 0 2 5 Swan Hills 0 0 0 0 0 1 0 0 1

Adult

Total 7 2 9

13 5 18

5 4 9

9 7 16

2 1 3

2 2 4

Grizzly bear vs. black bear captures The numbers of grizzly bears vs. black bears captured have varied between years and areas (Table 8). Table 8. Number of grizzly bears vs. black bears captured by year. Year 1999 2000 2001 2002 2003 2004 2005 2006

Grizzly Bears 24 25 29 28 28 25 23 15

Black Bears 5 13 10 5 14 25 22 38

Total 29 38 39 32 42 50 45 53

An analysis was conducted to determine and compare capture probabilities for both grizzly and black bears in all study areas between 1999 and 2006. There have been 355 ground-based snare sites set by the Grizzly Bear Research Project (GBRP) since 1999 (Fig. 2). Capture efforts were focused immediately south of Hinton from 1999 to 2003. Efforts moved further south to the U.S. border in 2004 and then north to the Swan Hills and the Grande Prairie area in 2005. In 2006, capture efforts were concentrated in the Swan Hills, Grande Prairie area and the Chinchaga. Due to missing removal dates for sites from 1999 to 2000, only 274 of the 355 sites had valid duration data. Of these 274 sites, 62 had an installation date, a removal date but

lacked information on period closures between installation and removal. This left 212 sites with an accurate record of number of days open (Fig. 3).

Figure 2. Total snare sites by year.

Figure 3. Snare sites with known installation and removal dates by year.

Duration of activity varied widely from 1 to 61 days, with an average duration of 16 days (standard deviation = 11 days). In general, southern sites (2004) were active for less time (10 days avg.) than northern sites (17 and 20 days avg., 2005 and 2006 respectively) (Fig. 4). Capture probabilities for the 274 sites with recorded installation and removal dates were calculated as number of bears (grizzly, black and both) caught at each site divided by the number of days the site was active (Fig. 5, 6 and 7).

Figure 4. Number of days snare sites were active

Figure 5. Grizzly bear capture rates per day.

Figure 6. Total black bear capture rate per day.

Figure 7. Total bears capture rate per day.

Grizzly bear range in Alberta south of Highway 16 has been divided into 4 population units based on genetic analysis of tissue samples taken from dead bears that have been registered by Alberta Fish and Wildlife Division (Proctor and Paetkau 2004). Genetic samples were insufficient to determine population units north of Highway 16. For the purposes of this analysis, grizzly bear range north of Highway 16 was divided into 3 populations roughly based on grizzly bear management areas (BMAs) (Fig. 8). The average daily rate of grizzly bear capture was highest in the Clearwater and Waterton population units and lowest in the Alberta North population unit (Fig. 9).

Figure 8. Grizzly bear population units per day.

Figure 9. Average grizzly bear captures

Average daily rate of black bear capture, however, was highest in the Alberta North population unit and lowest in the Clearwater population unit (Fig. 10). When capture rates for grizzly and black bears are combined, the highest average daily capture rate was in the Alberta North population unit and lowest in the central foothills area (FMF core and Clearwater population units) (Fig. 11). Although this analysis suggests that grizzly bear population densities may be highest in the central and southern foothills area of the province and lowest in northern Alberta and that grizzly bear densities are inversely related to black bear densities, further data and analysis is required to more accurately determine the relationship between capture rates and population density.

Figure 10. Average black bear captures per day

Figure 11. Average total bears captures per day

Other Non-Target Species There were no other non-target species captured.

CHAPTER 2: REMOTE SENSING AND HABITAT MAP PRODUCTION 2006 David Laskin1, Greg McDermid1, Alysha Pape2, Ame Wunderle2, and Jennifer Klassen1. 1

Foothills Facility for Remote Sensing and GIScience, Department of Geography, University of Calgary, Calgary, AB. 2

Environmental Remote Sensing Laboratory, Department of Geography, University of Saskatchewan , Saskatoon, SK.

Introduction 2006 marks the seventh year in which researchers and technicians from the University of Calgary and the University of Saskatchewan have contributed remote sensing map products in support of grizzly bear ecology and conservation efforts within the Foothills Model Forest Grizzly Bear Research Program. McDermid (2007a) described a methodological framework for creating land cover and vegetation information over large areas from remote sensing and digital elevation models (DEMs). The approach generates an information base consisting of four attributes: land cover, crown closure, species composition, and normalized difference vegetation index (NDVI) phenology. The products have been demonstrated to provide both high-quality representation of land and vegetation conditions over large, multi-jurisdictional landscapes, as well as an effective foundation for explaining observed patterns of grizzly bear occurrence (McDermid et al., 2006c). This report documents mapping activities in the Phase 6 area of the Projectâ&#x20AC;&#x2122;s study area. Methods Study Area The phase 6 study area covers approximately 25 million hectares along the western portion of Alberta, and includes an extension area of more than 5 million hectares in the north (Figure 12).

Figure 12. Location of the Phase 6 study area in Western Alberta, including the phase 6 extension area. Field Sampling Field crews from the University of Calgary’s Foothills Facility for Remote Sensing and GIScience conducted ground surveys in the summer of 2006 to support two broad objectives: (i) filling in previous portions of the study area that previously contained sparse ground plots (the Phase 4 expansion area in the south, and Wilmore/Swan Hills regions in the north); and (ii) supporting map extension activities and an airborne mission in the Phase 6 expansion area. We used the area’s extensive road networks to gain access to field sites distributed across the region’s widely-varying land base, selected using a multi-stage sampling design. Our efforts netted a total of 470 new ground plots, including 123 in the Phase 4 extension, 218 in the Phase 5 extension, and 129 in the Phase 6 extension. We used established field methods (McDermid 2005) to acquire observations, measurements, and hemispherical photography designed to characterize land cover, crown closure, and tree species composition (proportion coniferous) across a sample plot roughly analogous to a 30 x 30 meter Landsat pixel. Airborne Mission A combined optical/LiDAR airborne mission was flown in the Phase 6 extension area to support on-going mapping and research activities. We contracted Itres Research to acquire highresolution (1 meter pixel size) optical imagery across 37 spectral bands using the Compact Airborne Spectrographic Imager (CASI). The firm Laser Imaging Technologies was commissioned to deploy a discrete-pulse LiDAR system designed to record ground/forest canopy elevations at a density of roughly two samples per square meter. Both sensors were flown along a spatially/temporally coincident transect throughout the Phase 6 extension (Figure 13). These data worked mutually as a surrogate for ground sampling in areas inaccessible to field crews.

Figure 13. Airborne sensor flight transect. Satellite Image Acquisition and Pre-Processing We used satellite imagery from two different sensor systems in 2006: (i) Thematic Mapper (TM) data from the Landsat 5 satellite and (ii) Moderate Resolution Imaging Spectrometer (MODIS) imagery from the Terra/Aqua satellites. The TM imagery was used to map land cover, crown closure, and species composition, while MODIS data were used to track NDVI, and assist with change detections. The extension portion of the Phase 6 study area was composed of four separate Landsat scenes, west to east: 47/21, 46/21, 45/21, and 44/21. Each image was orthorectified, transformed to topof-atmosphere radiance, and radiometrically normalized to a pre-existing master image (45/22 from the Phase 3 study area) using the linear transformation procedure described in McDermid et al. (2007b). We used the tasseled cap transformation of Crist and Ciccone (1984) to generate the standard orthogonal components: brightness, greenness, and wetness. A DEM from DMTI Spatial was acquired for this study through an academic agreement with the University of Calgary library. Wall-to-wall DEM coverage of the study area was achieved by mosaicking 1:250,000 map sheets in ArcMap. We used morphometric processing to derive additional topographic variables such as slope and angle of incidence to assist in subsequent mapping activities. Map Updates and Change Detection The core orthomosaic â&#x20AC;&#x201C; and subsequent map products â&#x20AC;&#x201C; of the Phase 5 study area were derived from TM images acquired from 2000 to 2003 (Figure 14). We used change detection methods to update these products to circa 2005 standards for the Phase 6 delivery. A total of 18 new Landsat scenes and three new MODIS tiles were acquired in support of this work. We used the Enhanced Wetness Difference Index method of Franklin et al. (2001) on top-of-atmosphere corrected Landsat scenes to reveal change features to 2005. The change layer was imported into

Definiens Professional 5.0 and subject to a series of decision rules (Figure 15) designed to classify four basic disturbance categories: well sites, cut blocks, burns, roads, and mine expansions. In areas where no recent Landsat scenes were available, we used 250-meter MODIS data (NDVI tiles from MOD13Q1) as the basis for change detection. Previous experience (Pape 2006) has shown that these data are only capable of capturing about 60% of the common change elements on the landscape, but extensive Landsat coverage meant that MODIS data were only required over a very small area.

Figure 14. Temporal resolution of the Phase 5 orthomosaic.

Figure 15. Classification decision rules with suggested starting values. Land Cover This yearâ&#x20AC;&#x2122;s land cover mapping methodology was focused on an intuitive decision-rule based classification process designed to reduce the heavy reliance on training samples. As in previous years, we conducted our classification in a piece-wise manner (McDermid et al., 2007b) using work zones defined by natural subregion. We used Definiens Professional 5.0, to create image objects and implement a hierarchical decision-rule-set based on a combination of image interpretation skills, field sample data, and a-priori ground knowledge of the areas in question. NDVI Phenology We used MODIS data from the MOD13Q1 vegetation index series (Huete et al., 1999) to monitor NDVI across the 2005 growing season. A series of 16-day normalized difference vegetation index (NDVI) composites from the Terra satellite were acquired across MODIS tiles h10v03, h10v04 and h11v03 from April 7 to October 31, 2005. We used the MODIS Reprojection Tool to mosaic the required scenes and re-project them to UTM zone 11, NAD 83. In order to facilitate subsequent analysis and modeling activities by program partners, we generated a series of phenology metrics using software written in IDL 6.3. Metrics include Maximum NDVI, Timing of Maximum NDVI, NDVI range, and NDVI Change for each composite period.

Crown Closure and Species Composition Crown closure and species composition were modeled as continuous variables using regression techniques in the SPLUS statistical software package. Ground data for crown closure was collected using vertical 180 degree hemispherical photography and LiDAR measurements. The hemispherical photos were processed with WinScanopy image analysis software (Regent Instruments Inc.), while crown closure measurements were derived from LiDAR data using software written in IDL 6.3 (McLane 2007). Tree species composition (measured as the proportion of conifer trees) was measured using variable-area prism sweeps and high-resolution hyperspectral imagery. We followed the procedures outlined by McDermid et al. (2007b) to map crown closure and species composition in a piecewise fashion across the study area using source models where field data were concentrated and destination models where ground data was sparse. Crown closure was modeled using conventional regression techniques on arcsine-transformed response variables. Species composition was modeled using binomial-family generalized linear models with a logit link designed to preserve the sample proportions of conifer vs. broadleaf (Crawley 2002). We used Akaikeâ&#x20AC;&#x2122;s Information Criterion (AIC) to select minimum adequate models based on spectral (brightness, greenness, and wetness) and topographic (slope, incidence, and elevation) variables following the principle of parsimony. We verified variable selections using F-tests and analyses of variance. The final Phase 6 surface was tested for predictive accuracy using multiple aggregation confusion matrices. Results and Discussion Map Updates and Change Detection Areas of disturbance derived from the change detection are shown in Figure 16. The layer has an overall accuracy of 83.7% (Kappa=0.755), with individual class accuracies ranging from 70-97% (Table 9). The userâ&#x20AC;&#x2122;s accuracy of the No Change class is highly affected by omission errors associated with well sites, and can be attributed to the small sample size of the No Change class. However, our main concern here is with the accuracy of the disturbed classes, and the proportional representation of the Change/No Change classes would produce extremely high accuracies as <10% of the landscape has changed. Accuracy of the Cutblock class is high although some spatial error may exist within polygons due to slight misregistration. Lastly, it is intended that this product be used in conjunction with the road data available in the FMF GIS database.

Figure 16. Disturbance and regeneration in the Phase 6 study area. Table 9. Confusion Matrix for the 2005 Change Update Layer CHANGE Cutblock Wellsite No Change Total Producer Overall Accuracy

Cutblock 89 1 11 101 88.1% 83.7

Wellsite 5 64 22 91 70.3%

No Change 7 1 89 97 91.7% KAPPA STATISTIC

Total 101 66 122 289

User 88.1% 97.0% 73.0%

0.755

Land Cover The composite land cover map over the Phase 6 study area is shown in Figure 17. The overall accuracy of the map varies from region to region, ranging from 75% (Kappa=0.46) the Phase 5 expansion area to 88% (Kappa=0.70) in the Phase 6 expansion area, with the results predictably tied to the number of ground points available in each workzone. Upland treed areas were generally well classified across all regions except phase 5, where many upland treed objects were mis-classified as wetland herb. We attribute this to the thinner broadleaf forests in the area displaying spectral values to other wetland classes. Smaller amounts of confusion between

wetland treed and upland treed classes is evident in phase 6, explained in part by the relatively small sample size of the wet-treed class in that area. We recommend further refinement in the delineation of wetland classes in order to increase the accuracy of future map products.

Figure 17. Land cover map for the Phase 6 study area. Overall, the rule-based classification method produced results similar to that reported by McDermid et al (2007a) over the Phase 3 study area, while significantly decreasing computation and handling time. We regard this new method of land cover mapping as a significant step forward.

NDVI Phenology The NDVI range product reflects the magnitude of seasonal vegetation dynamics, where NDVI change indicates the rates and/or levels of change occurring in vegetated surfaces between composite periods (Figure 18). Other NDVI composite surfaces include Minimum, Maximum, and the timing of maximum NDVI. Reflecting the per-pixel level and time during which maximum/minimum surface vegetative production is reached during the growing season (Pettorelli et al., 2005).

Figure 18. 16-day NDVI phenology composites for the 2005 growing, Phase 6 study area.

Crown Closure The intensified ground sample effort was successful in improving measures of crown closure in areas that were previously under-sampled in Phase 5, while the LiDAR mission helped enable the generation of high-quality map products across the Phase 6 extension in the north. A seamless mosaic of the crown closure model surfaces is the result (Figure 19). Overall accuracy of a four-class configuration of crown closure in the Phase 6 extension was calculated at 78%.

Figure 19. Continuous crown closure map.

Species Composition Once again, the supplementary field sampling in summer 2006 combined with the airborne mission in the north enabled the production of a consistent species composition map (Figure 20). Accuracy of product in a four-class configuration in the Phase 6 extension was calculated at 80%.

Figure 20. Continuous species composition map (% conifer).

Summary Field and laboratory efforts from personnel in the remote sensing team at the Foothills Facility for Remote Sensing and GIScience at the University of Calgary and the Environmental Remote Sensing Laboratory at the University of Saskatchewan have enabled the production of highquality maps of land cover, crown closure, species composition, and NDVI phenology over the Foothills Model Forest Grizzly Bear Research Program’s Phase 6 study area, covering almost 25 million hectares of rugged, multi-jurisdictional terrain in western Alberta. Further efforts to extend these map products north to the Northwest Territories border – the Phase 7 expansion area – are currently underway and will be completed in 2007. Literature Cited Crawley, M.J. (2002) Statistical Computing, and Introduction to Data Analysis Usng S-Plus. Wiley, West Sussex, England. Crist, E. P. and R. C. Cicone, 1984: A physically based transformation of Thematic Mapper data - The TM Tasseled Cap. IEEE Transactions on Geoscience and Remote Sensing, Vol. GE22, pp. 256 - 263. Franklin, S.E., Lavigne, M.B., Moskal, L.M., Wulder, M.A. & McCaffery, T.M. (2001). Interpretation of Forest Harvest Conditions in New Brunswick Using Landsat TM Enhanced Wetness Difference Imagery (EWDI). Canadian Journal of Remote Sensing. 27, 118-128. Fung, T. & LeDrew, E. (1987). Application of principal components analysis to change detection. Photogrammetric Engineering and Remote Sensing. 28, 681-684 Huete, A., Kerola, D., Didan, K., van Leeuwen, W.J.D., Ferreira, L (1999). Terrestrial Biosphere Analysis of SeaWiFS data over the Amazon Region with MODIS and GLI Prototype Vegetation Indices. presented at the IEEE- IGARSS'98, Seattle, WA, July 6-10. McLane, A.J. (2007) Use of profiling airborne LiDAR to estimate canopy closure in northern Alberta: An empirical relations investigation. Unpublished master’s thesis, University of Calgary, Department of Geography. McDermid, G. J., S. E. Franklin, and E. F. LeDrew (2005). Remote Sensing for Large-area, Multi-jurisdictional Habitat Mapping. Unpublished Ph.D. thesis, Department of Geography, University of Waterloo. McDermid, G. J., S. E. Franklin, and E. F. LeDrew (2007a). A multi-attribute approach to mapping vegetation and land cover over large areas in support of wildlife habitat assessment. Remote Sensing of Environment, in review. McDermid, G. J., S. E. Franklin, and E. F. LeDrew (2007b). Radiometric normalization and continuous-variable model extension for the operational mapping of large areas with Landsat imagery. International Journal of Remote Sensing, in review.

McDermid, G. J., R. J. Hall, A. Sanchez, S. E. Franklin, G. B. Stenhouse, T. Kobliuk, and E. F. LeDrew (2007c). Remote sensing and forest inventory for wildlife habitat assessment. Journal of Wildlife Management, in review. Pape, A. (2006). Multiple Spatial Resolution Change Detection for Environmental Management. Unpublished MSc Thesis. Faculty of Social Science, Department of Geography, University of Saskatchewan. Pettorelli, N., Olav Vik, J., Mysterud, A., Gaillard, J.M., Tucker, C.J., and Stenseth, N.C. (2005). Using satellite-derived NDVI to assess ecological responses to environmental change. Trens in Ecology and Evolution. 20(9): 503-510

CHAPTER 3. SEASONAL GRIZZLY BEAR HABITAT FOR SIX POPULATION UNITS OF ALBERTA, CANADA Dr. Scott Nielsen, Conservation Analytics Inc; 1817 Tufford Way, NW; Edmonton, Alberta

Introduction Critical to management and conservation of remaining grizzly bear populations in Alberta are spatial depictions of habitat. Given a consistent and objective habitat product, resource stakeholders can communicate potential impacts of future land-use actions on grizzly bear populations and suggest alternatives that minimize habitat degradation. To date, there have been a number of models or maps representing grizzly bear habitat (Nielsen et al. 2002, 2003, 2004a; Stevens 2002; Theberge 2002, Nielsen 2005). Although these maps have proven useful in local resource and grizzly bear management, such products have been limited in spatial extent, often being available for only a portion of a population unit. Moreover, existing products are often scaled differently making comparisons or integration difficult. Here I present a consistent set of seasonal grizzly bear habitat products representing the relative probability of habitat use for six of seven grizzly bear population units in Alberta. This represents an area of 132,076 km2 (55% of currently defined grizzly bear range in Alberta; Figure 21). I describe the model approach, discuss habitat predictions, and assess model accuracy. Methods Defining grizzly bear habitat use For each of three seasons (Table 10) and six population units of interest (Figure 21), relative probability of habitat use was estimated from the distribution of grizzly bear animal locations relative to 16 environmental variables that represented landcover type, forest canopy characteristics, soil wetness, edge distance metrics for selected landcover classes and distance to streams (Table 11). Landcover and forest canopy variables were based on remote sensing classifications provided by Dr. Greg McDermid (University of Calgary). Soil wetness (compound topographic index) was generated from a 30m DEM using an ArcInfo AML. Using the landcover classification and a stream network, a number of edge variables were produced that relate the distance of any pixel to the edge of a forest or stream. Edge variables were chosen to highlight the attraction or potential habitat of forest edges and streamside riparian areas, which usually contain a higher diversity and abundance of food resources. Although buffers could be used to define edges, the choice of distance thresholds used is not immediately apparent and likely varies among different landcover classes. Use of straight line distances, on the other hand, assumes equal influence for any particular change in distance regardless of whether it is near to an edge or at some large distance that becomes meaningless. To emphasize greater influence of edge habitat and little to no relevance at larger distances, an exponential transformation of straight line distance was used to scale values from 0 (at edge) to 1 (at some large distance from an edge). Specifically, the transformation took the general form, y 1 e



d

500



where d represents the distance in meters to an edge and y the resulting distance index for forest edge or streamside habitat. A 500 meter constant was chosen based on the performance of preliminary analyses that tested 50, 100, 250, 500, and 1000 meters values respectively. Using this transformation and the chosen 500 meter parameter, 95% of the change in the index occurred at 3-times the constant (i.e., 500 meters  3), or a value of 0.95 at 1,500 meters, while approximately 50% of the index (i.e., 0.5) was at 350 meters (Figure 22). This ensured that large distances from edges were effectively removed from relevance, while distances near edges had the greatest impact. Resource selection functions (Manly et al. 2002) were estimated for each grizzly bear and season combination with a meta-analysis used to subsequently estimate population level coefficients for each population unit. Due to small sample sizes (both in the number of collared animals and GPS relocations), the Waterton and Livingstone units were combined into a single regional habitat model. Where sample size permitted, models where estimated from female grizzly bear locations (e.g., Clearwater and FMF core) to emphasize the most sensitive sex-age class for population persistence. The other three model areas (e.g., Livingstone/Waterton, Grande Cache, and Swan Hills) considered both male and female grizzly bears for final population-level habitat. All population units, however, were scaled according to predicted probability of female grizzly bear occupancy to emphasize habitat for female bears. Defining female grizzly bear range (occupied versus unoccupied habitat) Although habitat was defined as areas of differential habitat use within animal home ranges, a female bear range scalar was used to account for potential differences in range (occupancy) of female bears compared to more wide-ranging male animals. Minimum convex polygon (MCP) female home ranges were estimated for each known female grizzly bear in the six population units of interest (including the East Slopes Grizzly Bear Project). A series of systematic locations were generated at a spacing of 10 kilometers to approximate the average scale of female home ranges and classified as either female ‘occupied’ (inside home range) or ‘unoccupied’ range. I assumed all locations within the Rocky Mountain parks to be occupied even without known female home ranges, since most studies focused on the front ranges and female grizzly bear distribution can largely be assumed to be occupied to the west of the front ranges in the protected mountain parks to the British Columbia border. Logistic regression was used to classify occupied and unoccupied female bear range using a suite of potential landscape and/or human variables. Although a number of variables were significant in univariate assessments, only natural sub-region and proportion of agriculture (based on a classification provided by Greg McDermid) within a 10 kilometer radius window were used, as these variables described occupancy well, with natural sub-region acting as a surrogate for a number of other environmental factors. Lower foothills were used as the reference category (indicator contrast) for comparisons of other natural sub-regions. Therefore positive natural sub-region coefficients would indicate greater likelihood of female bear occupancy than the lower foothills, while negative coefficients would indicate a lower likelihood of female bear occupancy than lower foothills. The probability of female bear occupancy was estimated for each pixel in population units 1 to 6 using coefficients estimated from the final logistic regression model and a GIS.

Defining final grizzly bear habitat and validation Final habitat maps were generated in a GIS through customization of an ArcToolbox Tool with individual models for each season and population unit (Figure 23) developed in Model Builder (ESRI, Model Builder 9.1). Habitat was first estimated for the entire population unit (i.e., extrapolated) based on habitat selection within home ranges (section 2.1) and three mask variables used to define non-habitat conditions. Mask variables included a water and ice pixels, agricultural fields, and non-vegetated sites in the alpine natural sub-region. The mask of nonvegetated sites in alpine areas was used to remove high elevation rock. To account for differences in the likelihood of female occupancy across population units (except Swan Hills), the habitat model was re-scaled so that values were on a similar scale to the female occupancy model (low of 0 to a maximum of 1) and combined with the female occupancy model using the product of the two maps. Maps were subsequently reclassified into 11 bins using 0 for mask pixels (water, ice, alpine barren, and agricultural fields) and 1 to 10 as ordinal habitat bins based on the distribution of final habitat values inside home ranges (using quantile classification). Using defined classification breakpoints in home ranges, estimates of habitat (0 to 10 bins) were generated for the entire population unit. These methods were repeated for each season and population unit (Swan Hills was binned without the female occupancy model). Each map was evaluated for its predictive accuracy based on predicted habitat values for animal locations used to generate habitat models (i.e., training data) and animal locations withheld for independent validation (testing data). A spearman rank correlation of habitat bin number (0 to 10) and percent use of each bin (adjusted for the aerial extent of bin in home ranges) was used to gauge the predictive accuracy of seasonal population unit habitat maps (Boyce et al. 2002). Results Grizzly bear habitat use Seasonal grizzly bear habitat use was defined by landcover types, forest canopy characteristics, soil wetness, and distance to forest edge or streamside habitat (Tables 12-14). Generally, open herbaceous and open forest habitats were selected. When forest use occurred, it tended to be along forest edges or streamside habitats. Interpretation of individual landcover categories based on coefficients was complicated by other model factors, particularly the distance to edge variables for those same landcover types. Final habitat values for individual landcover types were therefore highly variable and dependent on within stand characteristics (i.e., forest cover, forest species composition, and soil wetness) and landscape metrics (distance to forest edges and streams). Female grizzly bear range (occupied versus unoccupied habitat) Female grizzly bears were less likely to establish home ranges (occupancy) in the eastern fringes of population units where boreal, parkland, or grassland natural sub-regions occurred and particularly where the green zone transitioned into the agricultural white zone (Table 15; Figure 24). Compared to lower foothills, the upper foothills, montane, sub-alpine, and alpine natural sub-regions had significantly higher probabilities of female occupancy. The central mixedwood had significantly lower probabilities of occurrence, while the foothills fescue and foothills parkland were not significantly different from that of the lower foothill natural sub-region (Table 15). No female home ranges were observed in dry mixedwood, mixed grassland, central parkland, and Peace River parkland natural sub-regions. These areas were therefore assumed to

have a probability of female occupancy of 0. Mapped probabilities were consistent with the distribution of known female home ranges. Final grizzly bear habitat map Final habitat maps, ranked from 0 (considered non-habitat) to 10 (highest habitat use), reflected the product of both within home range habitat use and predicted female grizzly bear occupancy (Figure 25). Predictive assessments of maps suggested good predictive accuracy of maps (Table 16), although there did appear to be consistent over-use of masked habitats (bin 0) for most seasons and population units. An informal investigation of these animal locations found many of the locations to be in barren landcover types of the mountains (i.e., rocky mountainsides). With the presence of herbaceous landcover types adjacent to these barren, rocky sites and the presence of steep terrain interfering with satellite number and strength, GPS positional error seems to be a logical cause of misclassification. Future analyses might benefit from compositional analyses of landcover types and predictive accuracy of final habitat maps using circular windows of 150m to 300m in diameter to account for possible inaccuracies of animal locations. Regardless of possible positional inaccuracies, the hypophagia season generally had the poorest predictive accuracy, while the hyperphagia seasons generally had good to excellent predictive accuracy (Table 16). In some areas, such as the Livingstone and Waterton population unit, predictive accuracies should be considered an under-estimate of female grizzly bear habitat, since male bears were used to generate and test habitat models and in some instances occurred beyond the expected range of female bears. In contrast, predictions of habitat for the periphery of the Swan Hills may be over-estimated, since the female occupancy scalar was not used for this population unit. Future calibration of this unit should be considered for inter-population unit comparisons and evaluations of habitat value for areas of the population unit that likely have low female occupancy rates as depicted in figure 24. Discussion These results provide for the first time a consistent and objective definition of grizzly bear habitat across multiple population units of Alberta. Although this represents an important step for management and conservation of grizzly bears, further considerations of mortality risk (Nielsen et al., 2004b, 2006), food resources (Nielsen et al. 2003), and animal density should be given. Moreover, large-scale modeling and mapping of grizzly bear habitat requires trade-offs in detail. Undoubtedly, more specific definitions of grizzly bear habitat can be estimated for smaller regions of population units by using more detailed geo-spatial datasets and limiting the extent of environmental gradients assessed. For example, not all spatial datasets are available for Alberta grizzly bear range. Forest stand history, which has been shown to be important for predicting bear habitat use (Nielsen et al. 2004a), has been difficult to compile and map consistently across Forest Management Agreements. Specific to environmental gradients, it has become apparent that grizzly bear habitat use for specific landcover types can vary substantially among grizzly bear population units or even within larger population units as a function of elevation gradients, natural sub-regions, or differences in learned behaviors of animals. Despite limitations, evaluation of grizzly bear habitat maps suggests good to excellent predictive accuracy. Future habitat modeling should consider models having flexible habitat coefficients based on landscape context and resource gradients, while also integrating estimates of mortality risk predicting habitat states or safe harbor and attractive sink habitats as described in Nielsen et al. (2006).

References Boyce, M.S., Vernier, P.R., Nielsen, S.E., Schmiegelow, F.K.A., 2002. Evaluating resource selection functions. Ecological Modelling 157, 281-300. Manly, B.F.J., McDonald, L.L., Thomas, D.L., McDonald, T.L., Erickson, W.P., 2002. Resource selection by animals: Statistical design and analysis for field studies. Second edition. Kluwer Academic Publishers, Dordrecht, the Netherlands. Nielsen, S.E., 2005. Habitat ecology, conservation, and projected population viability of grizzly bears (Ursus arctos L.) in west-central Alberta, Canada. PhD Dissertation, University of Alberta, Edmonton, Alberta, Canada. Nielsen, S.E., Boyce, M.S., Stenhouse, G.B., Munro, R.H.M., 2002. Modeling grizzly bear habitats in the Yellowhead Ecosystem of Alberta: Taking autocorrelation seriously. Ursus 13, 45−56. Nielsen, S.E., Boyce, M.S., Stenhouse, G.B., Munro, R.H.M., 2003. Development and testing of phenologically driven grizzly bear habitat models. Ecoscience 10, 1−10. Nielsen, S.E., Boyce, M.S., Stenhouse, G.B., 2004a. Grizzly bears and forestry I: selection of clearcuts by grizzly bears in west-central Alberta, Canada. Forest Ecology and Management 199, 51–65. Nielsen, S.E., Herrero, S., Boyce, M.S., Benn, B., Mace, R.D., Gibeau, M.L., Jevons, S., 2004b. Modelling the spatial distribution of human-caused grizzly bear mortalities in the Central Rockies Ecosystem of Canada. Biological Conservation 120, 101-113. Nielsen, S.E., Boyce, M.S., Stenhouse, G.B., 2006. A habitat-based framework for grizzly bear conservation in Alberta. Biological Conservation, 130, 217-229. Stevens, S., 2002. Landsat TM-based greenness as a surrogate for grizzly bear habitat quality in the Central Rockies Ecosystem. M.Sc. thesis, University of Calgary, Calgary, Alberta, Canada. Theberge, J.C., 2002. Scale-dependent selection of resource characteristics and landscape pattern by female grizzly bears in the eastern slopes of the Canadian Rocky Mountains. Ph.D. dissertation, University of Calgary, Calgary, Alberta, Canada.

Table 10. Dates used to define seasonal habitat use. Season number Season name

Dates (period of activity)

hypophagia

1 May to 15 June

early hyperphagia

16 June to 31 July

late hyperphagia

1 August to 15 October

Table 11. Environmental variables used to predict the relative probability of seasonal habitat use by grizzly bears in six population units of Alberta, Canada. Variable Landcoverâ&#x20AC; upland-treed wet-treed regenerating forest shrub wet-herb upland-herb non-vegetated Forest canopy crown closure in treed (wet or upland) sites crown closure in regenerating forest species composition in upland treed sites Terrain-based soil wetness compound topographic index (150m average) Distance to edge (transformed) distance to opening in upland-treed distance to opening in wet-treed distance to forest edge in upland-herb distance to forest edge in regenerating forest distance to forest edge in non-vegetated Distance to stream (transformed) distance to stream edge

Data range

Abbreviation

0 or 1 0 or 1 0 or 1 0 or 1 0 or 1 0 or 1 0 or 1

uptree wettree regencut shrub wetherb upherb nonveg

1 to 100 1 to 100 0 to 100

cc_treed cc_regen sc_uptree

3.4 to 24.0

cti

0 to 1 0 to 1 0 to 1 0 to 1

edge_uptree edge_wettree edge_upherb edge_regen edge_nonveg

0 to 1

dist_stream

â&#x20AC; upland treed used as the reference category in indicator contrasts of landcover types.

Table 12. Habitat coefficients ( population unit for the hypophagia period (season 1: 1 May to 15 June).

Variable Landcover wettree regencut shrub wetherb upherb nonveg Forest canopy cc_treed cc_regen sc_uptree Soil wetness cti Distance to edge_uptree edge_wettree edge_upherb edge_regen edge_nonveg dist_stream

Livingstone/Waterton SE 

Clearwater SE 

FMF core SE 

Grande Cache SE 

Swan Hills SE 

 -0.091 -0.265 -0.275 0.513 0.503 0.342

0.298 0.211 0.163 0.408 0.246 0.248

-1.468 -1.528 -1.622 -5.159 1.224 -3.955

0.151 0.230 0.175 0.203 0.225 0.197

-1.734 3.426 -1.575 -3.072 0.013 -1.674

0.315 0.491 0.318 0.409 0.298 0.326

N.A.

-1.945 -0.651

5.600 0.347

-4.233 4.11

3.109 7.115

N.A.

-0.133 -1.648

0.357 0.484

4.539 0.221

7.128 2.855

-0.009 0.035 -0.007

0.005 0.048 0.002

-0.038 0.008 -0.137 0.100 0.057 0.0734

-0.008 0.001 -0.001 0.002 0.001 0.0012

-0.015 0.002 -0.002 0.002 -0.005 0.0011

-0.008 0.004 -0.05 0.006 -0.006 0.0014

0.053

0.028

0.060

0.046

-0.006

0.011

0.005

0.011

0.071

0.025

-0.087

0.315

-2.371

0.654

N.A.

-1.277 -0.273 -2.784 -0.042

0.636 0.1 0.517 0.294

-0.772 -0.619 -0.973 0.083 -0.746 -0.445

0.091 0.27 0.435 0.386 0.263 0.071

0.211 0.099

0.896 0.998 0.661 0.220

0.112 0.451 0.319 0.523 0.238 0.083

-2.067 0.618

-1.566 -1.463 -1.304 0.619

-1.95 0.509 -0.246 -0.768 -4.130 -0.572

-5.814

1.202

-0.379

0.154

Table 13. Habitat coefficients ( population unit for the early hyperphagia period (season 2: 16 June to 31 July).

Livingstone/Waterton SE 

Clearwater SE 

FMF core SE 

Grande Cache SE 

Swan Hills SE 



0.304 0.457

-1.616 -2.266 -3.148 -8.017 -2.523 -2.283

1.324 1.265 1.146 2.887 1.266 1.164

-0.111 -0.214 -0.53 -0.333 0.497 -0.015

0.266 0.188 0.153 0.372 0.220 0.209

-1.135 -2.188 -2.745 -5.762 0.713 0.823

0.125 0.197 0.142 0.193 0.191 0.213

-1.125 -1.441 -2.946 -3.418 -2.373 -3.757

0.179 0.302 0.218 0.292 0.224 0.246

-0.016 0.064 -0.013

0.004 0.021 0.002

-0.008 0.010 -0.024

0.004 0.017 0.010

-0.012 -0.010 -0.001

0.001 0.002 0.001

-0.020 -0.003 -0.014

0.002 0.002 0.001

-0.027 -0.017 -0.014

0.003 0.004 0.001

0.090

0.025

0.09

0.026

-0.029

0.011

0.088

0.009

0.104

0.016

0.345

0.285

-1.942

0.25

N.A.

-0.550 -4.085 -3.847 -0.476

0.428 2.906 0.366 0.180

-1.629 -1.687 -0.286 1.035 -2.809 -1.261

0.077 0.262 0.620 0.277 0.429 0.063

0.127 0.364

0.715 0.988 0.866 0.207

0.099 0.439 0.309 0.463 0.236 0.076

-1.060 -0.181

-2.301 -1.042 -3.178 -0.357

-1.347 -0.008 -0.581 -1.308 -4.104 -0.703

-0.267

0.269

N.A.

-5.792 -1.289

2.540 0.285

N.A.

-1.240 -1.382

-2.168

0.106

Table 14. Habitat coefficients ( population unit for the late hyperphagia period (season 3: 1 August to 15 October).

Livingstone/Waterton SE 

Clearwater SE 

FMF core SE 

Grande Cache SE 

Swan Hills SE 



0.273 0.385

-4.347 -2.061 -2.782 -2.139 -3.619 -3.297

0.822 1.485 1.079 0.884 1.158 1.123

-0.025 -0.182 -0.421 -0.487 0.398 0.054

0.328 0.203 0.162 0.401 0.227 0.219

-0.148 -2.374 -3.342 -4.305 -1.864 -3.468

0.113 0.201 0.151 0.184 0.207 0.220

-0.177 -0.396 -1.316 -1.601 -1.357 -1.576

-0.009 -0.004 -0.015

0.004 0.015 0.002

-0.014 0.012 -0.027

0.003 0.019 0.011

-0.011 -0.010 -0.001

0.001 0.002 0.001

-0.029 -0.018 -0.01

0.002 0.002 0.001

-0.034 0.003 -0.023 0.003 0.0002 0.001

-0.008

0.025

0.043

0.024

-0.055

0.011

0.208

0.009

-0.015

0.017

-0.875

0.273

-3.359

0.259

N.A.

0.837 -3.838 -2.707 -0.009

0.349 0.803 0.308 0.160

-1.5 -1.473 -0.957 -1.439 -1.690 -0.997

0.079 0.226 0.569 0.351 0.419 0.06

0.132 0.386

0.669 0.985 0.842 0.188

0.104 0.547 0.306 0.471 0.235 0.079

-1.594 -1.638

-3.518 2.385 -4.005 -1.061

-1.407 -0.008 -0.547 -1.315 -4.171 -0.717

-1.024

0.195

N.A.

-5.544 -1.064

2.006 0.302

N.A.

-0.916 1.127

0.181 0.278 0.222 0.271 0.232 0.226

-0.823

0.093

Table 15. Coefficients describing the probability of female grizzly bear occupancy in population units 1 to 6. Model significance (LR ď Ł2, df = 8) of 988.5 (p<0.001) and percent deviance explained of 50.8. Natural sub-region categories indicated with avoid had perfect avoidance within the study area and were assumed a probability of 0. 95% Confidence Interval Lower Upper

Variable Natural sub-region Boreal Forest Central Mixedwood Dry Mixedwood

ď ˘

S.E.

-1.679 avoid

0.386 -

<0.001 -

-2.436 -

-0.922 -

Grassland Foothills Fescue Mixed Grassland

-0.226 avoid

0.673 -

0.737 -

-1.545 -

1.093 -

Foothills Upper Foothills Lower Foothills

1.536

0.191 <0.001 1.163 1.910 [ natural sub-region indicator category ]

Parkland Central Parkland Foothills Parkland Peace River Parkland

avoid -0.099 avoid

0.672 -

0.882 -

-1.417 -

1.218 -

Rocky Mountains Alpine Montane Sub-alpine

5.208 1.685 4.399

0.722 0.266 0.402

<0.001 <0.001 <0.001

3.793 1.164 3.611

6.623 2.206 5.188

Agriculture proportion within 10 km

-7.982

1.718

<0.001

-11.350

-4.614

Constant

-0.865

0.121

<0.001

-1.103

-0.627

Table 16. Predictive accuracy of seasonal population unit habitat maps based on spearman rank correlations (rs) of habitat bin and area-adjusted frequency of use per habitat bin. Results reported by training (within sample) and testing (withheld out-of-sample) datasets. Training data

Testing data

A. Season 1 (hypophagia) 1,2 Livingstone/Waterton 3 Clearwater 4 FMF core 5 Grande Cache 6 Swan Hills Mean

rs 0.200 0.620 0.627 0.718 0.964 0.626

p 0.555 0.042 0.039 0.013 0.0001

rs 0.582 0.508 0.591 0.782 0.936 0.680

p 0.060 0.111 0.056 0.005 0.0001

B. Season 2 (early hyperphagia) 1,2 Livingstone/Waterton 3 Clearwater 4 FMF core 5 Grande Cache 6 Swan Hills Mean

0.864 0.873 0.746 0.836 0.991 0.862

0.001 0.001 0.009 0.001 0.0001

0.746 0.822 0.782 0.836 0.982 0.834

0.009 0.0003 0.005 0.001 0.0001

C. Season 3 (late hyperphagia) 1,2 Livingstone/Waterton 3 Clearwater 4 FMF core 5 Grande Cache 6 Swan Hills Mean

0.627 0.900 0.746 0.855 0.982 0.822

0.039 0.0002 0.009 0.001 0.0001

0.655 0.918 0.736 0.846 0.991 0.829

0.029 0.0001 0.010 0.001 0.0001

6 5

4 Population unit: Alberta North

Clearwater FMF Core

Grande Cache Livingstone Swan Hills Waterton

Figure 21. Grizzly bear population units of Alberta. Habitat maps were produced for population units 1/2 (Waterton and Livingstone), 3 (Clearwater), 4 (FMF core), 5 (Grande Cache), and 6 (Swan Hills). Only the Alberta North (unit 7) remains unclassified.

Edge distance index

0.8 0.6

0.4 0.2

2400

2200

2000

1800

1600

1400

1200

1000

800

600

400

200

Distance (meters) to edge

Figure 22. Edge distance index used for assessing and predicting the attraction of edge habitats by grizzly bears.

Figure 23. Model diagram illustrating the process of determining final grizzly bear habitat values for individual population units and seasons (re-parameterized and repeated for each season and population unit combination).

Figure 24. Predicted range of female grizzly bears (occupancy of home range) for six grizzly bear population units of Alberta. This map was used as a scalar to ensure extrapolated habitat values were consistent with occupancy of the area.

Figure 25. Predicted grizzly bear habitat rank during late hyperphagia (1 August to 15 October) for six grizzly bear population units of Alberta, Canada.

CHAPTER 4: THE DIET OF GRIZZLY BEARS (URSUS ARCTOS) IN NORTH-CENTRAL ALBERTA, CANADA, PART OF A NEW FOODBASED MODEL APPROACH TO MAPPING GRIZZLY BEAR HABITAT Terrence Larsen and Karine Pigeon, Foothills Model Forest, Hinton, Alberta

Abstract The diet of the grizzly bear (Ursus arctos) has been well documented in the mountainous and forested regions of the Rocky Mountains. However, little is known about food selection of grizzly bears in Northern portions of the province. As part of a new foodbased model approach to mapping grizzly bear habitat in Alberta, we analysed grizzly bear scats (n=209) collected from GPS locations of collared bears from five study areas during field site investigations in 2005 and 2006. Although grizzly bear diets differed both spatially and temporally across study areas, similarities were detected in the timing and types of foods consumed when compared to a previous scat analysis from 2001-2003. In all study areas, succulent vegetation such as graminoids, horsetails, cow parsnip, and clover formed an important part of a grizzly bearâ&#x20AC;&#x2122;s diet. The frequency of use between these foods likely reflected seasonal changes in plant phenology and occurrence on the landscape. Major dietary differences observed among study areas were most commonly attributed to the quantity and type of mammalian prey, insects, and fruit consumed. Annual and regional variation in used food types is likely the result of differences in regional climatic conditions, landscape attributes and configuration of natural and anthropogenic disturbances, and the behaviours and abilities of individual bears. Introduction In Alberta, there is increasing pressure on the landscape from forest harvesting, oil and gas development, and human recreation that may have negative effects on the provincial grizzly bear population. Understanding current landscape conditions relative to grizzly bear food availability, abundance, and distribution are important factors that may limit population growth through reproductive fitness and survival (Iverson et al. 2001; Kasbohm et al. 1995; MacHutchon and Wellwood 2003; McLellan and Hovey 1995). Understanding how seasonal changes in food habits influence grizzly bears habitat selection is an important aspect to consider for land-use planning (MacHutchon 1989). The diet of grizzly bears, corresponding to seasonal changes in food use, has been well documented within the mountainous and forested regions of the Rocky Mountains (Hamer et al. 1991; McLellan and Hovey 1995; Munro and Price 2005; Raine and Riddell 1991; Servheen 1985). Between April and October 2001-2003 the Foothills Model Forest Grizzly Bear Research Program (FMFGBRP) investigated food habits of grizzly bears within the mountains and foothills of west-central Alberta (Hinton; Figure 26). However, little is known about food selection of grizzly bears in northern regions of Alberta, including the Swan Hills, Wapiti, and Chinchaga (Figure 26). Although habitats are similar among these areas, geographical differences (natural subregions) likely affect bear feeding strategies (Hamer et al. 1991; Figure 27).

The objective of this project is to better understand the regional ecology of grizzly bears, particularly in areas where diet information is not readily available. Findings included in this report will be used to further develop and validate a Resource Selection Function (RSF) food-based model in areas where diet information is available. The food model incorporates the percent volume of food items that occur seasonally in grizzly bear diets. From these values, the model predicts corresponding landscape values, both spatially and temporally, based on current landscape conditions and seasonal changes in food distribution. Methods Study Area The 2006 study area includes portions of the original core study area of the FMFGBRP from 1999-2003 (Hinton; Figure 26). The additional expansion areas form phase 5,6, and 7 of the remote sensing mapping efforts and include Wapiti, Swan Hills, Simonette, and Chinchaga (Figure 26)

Figure 26. Map of the study area boundaries in Alberta; primary highways and major cities displayed.

Figure 27. Map of the study area boundaries in Alberta relative to natural subregions; primary highways and major cities displayed.

Field Collection In 2005 and 2006, 294 scats were collected during field investigations of grizzly bear GPS collar locations within selected seasons based on study area boundaries (Figure 26; Table 17). We assumed that all scat samples found within a 20x20m area of the corresponding GPS collar location belonged to the collared bear only if the apparent age matched the bearuse date. Scats found outside the designated grid were collected based on a combination of the following criteria: a grizzly bear bed was present and contained hair, the scat was less than 50m from the plot, the relative age of the scat fit the date of the GPS location, the scat size and composition was consistent with that of a grizzly bear. In most cases scat was collected at grizzly bear bed sites and could be confirmed with the presence of grizzly hair. Scats were individually bagged, labelled, and stored in a freezer prior to analysis. Laboratory Analysis Based on the study area boundaries, one scat per bear-use date was selected for analysis unless an obvious difference could be confirmed from visual observation. We followed this protocol to minimize bias from over sampling any particular day within seasons. All samples (n=209) were thawed, re-hydrated, and washed through stacked 2mm and 212 ď m sieves. In an attempt to avoid underestimating the amount of meat within any given scat, the percent volume was recorded during the washing process. All macroscopic content was placed in paper bags prior to analysis. A 35 x 25 cm grid, separated by percent classes (50, 25, 12.5, and 6.25), was superimposed in a large metal pan to consistently estimate the percent volume of each food item per scat. Vegetative food items found in scats were positively identified using a reference collection of forbs, grasses, fruits, and seeds. Compound and/or dissecting microscopes were used to distinguish between Equisetum spp. bracts and the flower heads of Taraxacum officinale and Trifolium spp. Microscopes were also used to discern between the seeds of fruit. Because of difficulties distinguishing sedges from grasses in some cases, we combined them into a single category, graminoids. Similarly, blueberry species were also combined into a single category (Vaccinium spp.), unless positively identified. Mammalian food items such as hair and bone were identified from reference samples and available literature (Deedrick and 2004; Teerink 2004). In general, hair was identified based on physical appearance (form, colour, diameter, length) and medulla characteristics using a compound microscope at 10-40x. Microtine prey was identified using hair medulla characteristics and dentition patterns. Data Analysis Because of geographic differences among sampled grizzly bears, we analysed scats based on year and study area boundaries delineated by the habitat-mapping project. To maintain consistency with the previous scat analysis of Munro and Price (2005), bi-monthly periods were used to assess the seasonal importance of food items. For each season and corresponding year, the percent value of food items were averaged from the total number of scats found within that season. Munro and Price (2005) applied correction factors (CF; Hewitt and Robbins 1996) to adjust raw percent values to account for the differential digestibility of certain food items and associated bias within scats. Because CFâ&#x20AC;&#x2122;s were designed to convert millilitres of fecal residue into grams of dry digestible matter it was inappropriate to use them in our analysis (Hewitt and Robbins 1996).

Table 17. Summary of grizzly bear scats collected and analysed from north-central Alberta in 2005 and 2006. Bears

Area

1 M, 2 F 2F 1M 1 M, 1 F 1F 1F

Swan Hills Swan Hills Chinchaga Wapiti Simonette Hinton

Scat Collected Scat Analysed 55 132 6 73 14 14

38 80 4 63 14 10

Timeframe July 1 - September 17, 2005 April 21 - September 10, 2006 June 11 - June 19, 2006 June 3 - September 12, 2006 September 20 - October 8 2006 April 28 - May 15, 2006

Results Grizzly bear diets differed both spatially and temporal across study areas. Grizzly bears consumed both mammalian and vegetative food items on a seasonal basis, which likely corresponds to the availability and abundance of foods across the landscape. Because of the high diversity of species consumed by grizzly bears we classified food items into 8 major categories to examine the overall trends of food consumption. Several species within classes have been identified as high value grizzly bear foods and are consistently found in other geographic locations. Swan Hills In 2005 (Figure 28), green vegetation, dominated by cow parsnip, accounted for the majority of food intake in early July (Heracleum lanatum). From late July to early September, the use of green vegetation decreased with a corresponding increase in animal matter, insects, and fruit. Beaver (Castor canadensis) was the main species from late July to late August, and moose during early August and early September. Buffaloberry (Shepherdia canadensis) consumption peaked in late July and tapered off as blueberry species (Vaccinium myrtlloides x caespitosum x vitis-idaea) became increasingly important. Yellow jackets (Vespula spp.) were used consistently from late July to early September, peaking in late August.

Figure 28. Percent volume of important food items based on seasonal changes in grizzly bear diet in the Swan Hills Study Area, Alberta, 2005.

In 2006 (Figure 29), green vegetation dominated the collection period from late April to late July with graminoids peaking in late May, clover (Trifolium spp.) in early June, and cow parsnip in early July. Meat consumption was noticeable in late June and again in late August. Over-wintered ligonberries (Vaccinium vitis-idaea) were important in late April and buffaloberry formed a minor portion of the diet in early August. Beginning in late July, blueberries were consumed in abundance until the end of the study period. Ants were only present during late June.

Figure 29. Percent volume of important food items based on seasonal changes in grizzly bear diet in the Swan Hills Study Area, Alberta, 2006.

Because data-collection periods varied between 2005 and 2006, no comparison is available previous to early July. However, a seasonal comparison between early July and early September suggests that there is some degree of temporal variation. Roots (Hedysarum spp.) were used in 2005 but not in 2006; insects and animal matter varied among years, along with the type and quantity of individual food items. The use of forbs and fruit corresponded seasonally but the proportion used was different. Wapiti In early June, animal matter composed largely of neonate moose and beaver made up a significant portion of the diet (Figure 30). A shift to green vegetation, dominated by horsetails (Equisetum spp.) and cow parsnip, occurred in late June and lasted until early July. Fruits were consumed from early July until late August with buffaloberry peaking in late July. As fruit consumption decreased in early August, cow parsnip became prominent once again with a combination of ants and yellow jackets. In late August and early September, roots were used extensively along with moderate levels of neonate moose, ants, and yellow jackets.

Figure 30. Percent volume of important food items based on seasonal changes in grizzly bear diet in the Wapiti Study Area, Alberta, 2006. Hinton Roots and animal matter were the primary foods during early May. Animal matter consisted largely of bighorn sheep (Ovis canadensis; 30%) and microtine (Microtus spp.; 4%) prey (Table 18). Chinchaga In early June, animal matter was used extensively and included beaver (49 %) and neonate moose (24 %). Green vegetation such as graminoids and horsetails were used to a lesser extent (Table 18).

Simonette In late September, clover was the primary food item; roots and blueberries were consumed to a lesser degree. A switch to roots and blueberries occurred in early October with clover as a minor food item. Table 18. The seasonal importance of the percent volume of digestible dry matter from major food items consumed by grizzly bears in the Hinton, Chinchaga, and Grande Cache study areas in Alberta, Canada, 2006. Area

Season HorsetailGraminoids Forbs Roots Insects Animal Matter Fruit Cones/Seeds

Hinton Early May Chinchaga Early Jun Simonette Late Sept Early Oct

16 -

9 11 -

0 0 64 10

55 0 18 41

34 73 2

1 18 47

Discussion Sampling Bias and Error Caution should be emphasized when using scat analysis data for habitat mapping purposes or associated management applications. Aside from the obvious biases of an individualâ&#x20AC;&#x2122;s ability to accurately identify food items, other factors are also important when considering volumetric measurements of food items. The age at which a scat is collected relative to the date of deposition may influence the amount and type of items that can be positively identified because of decomposition. The most common and notable bias is the differential digestibility of certain food items related to the amount of fecal residue that can be identified (MacHutchon and Wellwood 2003; McLellan and Hovey 1995; Raine and Kansas 1990). Because roots and graminoids are poorly digested their importance may be over represented; conversely foods that are highly digested such as fruits and animal matter may be under represented (MacHutchon and Wellwood 2003; McLellan and Hovey 1995). Other biases may include the inability to find scats in certain habitats because of ground cover, or the inadvertent collection of a black bear scat (McLellan and Hovey 1995). Relative to this study, it is important to understand that this diet information is based on a small sample of collared grizzly bears within study areas. Due to logistic constraints including the number of field personnel as well as the ability to access GPS locations makes data collection challenging. Thus, our data may not reflect the importance of specific bear foods at the population level. However, it is likely that important foods identified by this study are widely available and used by bears across the landscape. We believe that the observed trends display the importance of most major food items. West-central Alberta Based on geographic location and similar habitat types, we expected our results to be somewhat similar to those of Munro and Price (2005) from west-central Alberta. From scat analysis data between 2001-2003, Munro and Price (2005) identified four major shifts in the diet of grizzly bears. These correspond to seasonal availability of important foods: â&#x20AC;&#x2DC;Early springâ&#x20AC;&#x2122; (den emergence to mid-June), grizzly bears dig hedysarum spp. roots and also

consume neonate moose. Green up or ‘late spring’ (from mid-June to the end of July) corresponded to a shift in green vegetation such as horsetails, graminoids, and forbs. From early August to mid-September, fruit was used signifying the ‘summer’ or fruit season. Bears then switched back to hedysarum spp. post fruit season until den emergence. Although kill site information from 2001-2003 displayed major trends in the use of ungulates, because most animal matter was not analysed other mammalian prey may not have been accounted for. Similarities in grizzly bear food habits have been observed in other regions of North America (Hamer et al 1991; McLellan and Hovey 1995; Raine and Riddell 1991; Servheen 1985) Swan Hills In 2006, horsetails and graminoids were used extensively in late April, prior to green-up. Bears likely target these abundant species because handling time is low and intake is high (MacHutchon and Wellwood 2003). In 2005 and 2006, Hedysarum spp. plants were rarely observed and likely occur in low densities, leaving bears with little else to choose from. Because herbaceous foods are difficult to digest, they must be consumed in large volumes to acquire the necessary energy (Mealey 1975). Our data also corresponds to the findings of Munro and Price (2005) who found that over-wintered ligonberries were used during this season. While succulent vegetation provides the necessary protein bears require for growth, over wintered fruit rich in carbohydrates would be important for energy assimilation (Noyce et al 1997). In both 2005 and 2006, bears consumed cow parsnip and clover during green up, similarly to the west-central Alberta findings. This shift in diet likely corresponds to an increase in digestible protein and energy available from newly growing forbs (McLellan and Hovey 1995). Throughout the summer, there is a gradual reduction in green vegetation corresponding to an overall decrease in protein content in these foods (MacHutchon and Wellwood 2003). Fruits were also available from late July to early September and formed the majority of the diet. However, contrary to west-central Alberta, blueberries were used almost exclusively rather than buffaloberry. During field investigations, buffaloberry was rarely found in the Swan Hills and is likely the reason bears selected blueberries. Summer fruit availability and abundance allows bears to acquire the necessary fat reserves for winter denning (Hamer 1996; Hilderbrandt et al 1999; MacHutchon and Wellwood 2003). In some bear populations, fruit productivity may limit bear numbers (Hamer 1996). The use of ants in late June likely corresponds to the phenologic changes in ant colonies (Noyce et al 1997). Noyce et al (1997) found a positive relationship between ant consumption by black bears and the increase in ant pupae and size from north-central Minnesota. In 2005, animal matter and insects formed a major portion of the diet during the ‘summer’ or fruit season contrary to the 2006 and west-central Alberta findings. Beaver was the principal mammalian food item with moose also being important. Little is known of the mechanisms related to how or when bears hunt ungulates or other mammalian prey. It has been well documented that neonate moose are pursued in the spring, post calving, by both grizzly and black bears and is likely a result of their reduced mobility (Hamer and Herrero 1991; McLellan and Hovey 1995; Mowat and Heard 2006). Meat protein is considered the highest quality bear food and is important during the spring to increase body mass and also in the fall, when it is converted to fat for winter denning (McLellan and Hovey 1995; Mowat and Heard 2006). The use of terrestrial meat sources varies considerably among individuals and

geographic area, and is reflected in these data (Mowat and Heard 2006). Roots were consumed during the expected time period (early September) but were only used in 2005. These major differences may simply be attributed to yearly variation and/or particular home ranges of study animals. In 2005, scats were collected from a bear that had a significant portion of its home range overlapping a large burn. Pengelly and Hamer (2006) identified an increase in digging density of Hedysarum alpinum from grizzly bears in areas of Banff National Park, Alberta, where shrubland fires occurred. Wapiti Overall food items selected by grizzly bears within the Wapiti study area were similar to those of west-central Alberta. However, temporal differences and similarities were evident in some of the major food items. Although only part of the pre green up information was available, it coincided with the findings of Munro and Price (2005). Animal matter comprised of neonate moose and beaver were important while horsetails and graminoids were used less often. Moose were again used in early September similarly to west-central Alberta and the Swan Hills. Although earlier pre green up information is not available, it is likely that root digging would have formed a major part of the pre green up period since they were used extensively in early September. While crude protein levels are lower in Hedysarum spp. during the late summer fibre content is also lower, resulting in a higher proportion of digestible crude protein available (MacHutchon and Wellwood 2003). The fruit season began much earlier than anticipated and also ended earlier than expected. However, contrary to the Swan Hills, buffaloberry was used almost exclusively, which is consistent with west-central Alberta. The occurrence of yellow jackets in grizzly bear scat has not been well documented in other North American studies. However, Ciarniello et al (2003) report that grizzly bears used wasps at low intensities during the summer (July 15 September 20) from use site investigations conducted in central interior British Columbia. It is hypothesized that similar to the ant phenology theory, grizzly bears selected yellow jackets based on the developmental stage of nests. In the Hinton study area, food items consumed were similar to the study conducted in 20012003. Roots were the most prevalent item followed by ungulate matter. However, the amount and type of ungulate matter was greater and different then past findings and was most likely the result of abundant bighorn sheep from the reclaimed coal mine. In Chinchaga, the use of neonate moose and green vegetation coincided with the findings of west-central Alberta during the early June season. In the Simonette area, considerable differences were evident. The consumption of clover during late September can be attributed to the newly established vegetation from a recently developed pipeline within the grizzly bears home range. Munro and Price (2005) indicate that in areas of high industrial activity, species such as clover are more abundant then natural levels. In early October, ligonberries and roots were consumed, likely reflecting regional food abundance. Seasonal Food Selection A combination of factors may be attributed to the selection of foods by grizzly bears. The yearly and seasonal variability of food distribution, abundance, and quality, as well as the ability for bears to locate these foods within their home range and associated habitat may influence selection. Availability and handling time has been deemed important within an

omnivores diet largely because of the trade-off between food quality and volume consumed (McLellan and Hovey 1995). However, high quality foods are most likely consumed proportionately to their abundance and low quality foods consumed at high volumes may reflect low food diversity. Because some high quality foods are not as readily available compared to low quality foods, balancing the costs-benefits is an important factor for grizzly bears (McLellan and Hovey 1995). Acknowledgements The FMFGBRP would like to thank Jennifer Cave, Karen Graham, Terrence Larsen, Robin Munro, Karine Pigeon, Cody Schmidt, Rebecca Vaughan, and Chelsey Whenham for their keen attitudes, dedication, and persistence in the collection of pertinent field data. We would also like to thank Edson Fish and Wildlife and the Hinton Training Center for providing the necessary lab space to make grizzly bear scat analysis more enjoyable. A special thanks goes to Robin Munro and Scott Nielsen for their input regarding the analysis of grizzly bear scat. Literature Cited Ciarniello, L.M., D. Seip, and D. Heard. 2003. Parsnip Grizzly Bear Population and Habitat Project. Final Report. Accessed online < http://web.unbc.ca/parsnipgrizzly/progress.htm>. Deedrick, D.W., and S.L. Koch. 2004. Microscopy of Hair Part II: A Practical Guide and Manual for Animal Hairs. Accessed online <http://www.fbi.gov/hq/lab/fsc/backissu/july2004/research/2004_03_research02.htm >. Hamer, D. 1996. Buffaloberry [Shepherdia Canadensis (L.) Nutt.] fruit production in firesuccessional bear feeding sites. Journal of Range Management 49(6):520-529. Hamer, D. and S. Herrero. 1991. Elk, Cervus elaphus, Calves as Food for Grizzly Bears, Ursus arctos, in Banff National Park, Alberta. Canadian Field Naturalist 105(1):101103. Hewitt, D.G., and C.T. Robbins. 1996. Estimating grizzly bear food habits from fecal analysis. Wildlife Society Bulletin. 24(3):547-550. Hildebrandt,G.V., Schwartz,C.C., Robins, C. T., Jacoby,M. E.,Hanley, T. A., Arthur, S. M. & Servheen, C. (1999). The importance of meat, particularly salmon, to body size, population productivity, and conservation of North American brown bears. Can. J. Zool. 77: 132â&#x20AC;&#x201C;138. Iverson, S. J., Smith, L. K., and MacDonald, J. (2001) Changes in diet of free-ranging black bears in years of contrasting food availability revealed through milk fatty acids. Can. J. Zool. 79: 2268-2279. Kasbohm, J.W., Vaughan, M.H., Kraus, J.G. 1998. Black bear home range dynamics and movement patterns during a gypsy moth infestation. Ursus. 10:259-267.

Mowat, G., D.C. Heard. 2006. Major components of grizzly bear diets across North America. Can. J. Zool. 84:473-489. MacHutchon, G.A. 1989. Spring and Summer Food Habits of Black Bears in the Pelly River Valley, Yukon. Northwest Science. 63(3)116-118. MacHutchon, G.A., and D.W.Wellwood. 2003. Grizzly bear food habits in the northern Yukon, Canada. Ursus. 14(2):225-235. McLellan, B.N., and F.W. Hovey. 1995. The diet of grizzly bears in the Flathead River drainage of southeastern British Columbia. Can. J. of Zool. 73:704-712. Mealey, S. P. 1975. The natural food habits of free-ranging grizzly bears in Yellowstone National Park, 1973-1974. M.Sc. thesis, Montana State University, Bozeman, MT. 158pp. Munro, R.H.M, and M.H.H. Price. 2005. The Diet of Grizzly Bears, Ursus arctos, in WestCentral Alberta, Canada. Final Report. Accesses online < http://www.fmf.ca/GB/GB_report7.pdf> Noyce, K.V., P.B. Kannowski, and M.R. Riggs. Black bears as ant-eaters: seasonal associations between bear myrmecophagy and ant ecology in north-central Minnesota. Can. J. Zool. 75:1671-1686. Pengelly, I. and D. Hamer. 2006. Grizzly bear use of pink hedysarum roots following shrubland fire in Banff National Park, Alberta. Ursus 17(2):124-131. Raine, R.M. and R.N. Riddell. 1990. Grizzly bear research in Yoho and Kootenay National Parks, 1988-1990. Final report. Prep. for Canadian Parks service, Western Region, Calgary, AB. Prep. by Beak Associates Consulting Ltd., Calgary, AB. Various pagination. Servheen, C. 1985. Grizzly Bear Food Habits, Movements, and Habitat Selection in the Mission Mountains, Montana. J. Wildl. Manage. 47(4):1026-1035. Terrink, B.J. 1991. Atlas and Identification Key to Hair of West European Mammals. Cambridge University Press. Cambridge. Terrink, B.J. 1991. Atlas and Identification Key to Hair of West European Mammals. Cambridge University Press. Cambridge.

Appendix Table 19. A comparison between the percent volume of digestible dry matter from major food items consumed seasonally by grizzly bears in the Swan Hills, Alberta, Canada, in 2005 and 2006. Season Late Apr Early May Late May Early Jun Lat Jun Early Jul Late Jul Early Aug Late Aug Early Sept

05/06 n 4 3 9 13 9

05/06 05/06 Horsetail Graminoids

2 4 5 3 10 14 8 12 13 9

5 8 -

54 36 9 21 14 4 1 2 -

28 24 11 3 2

37 39 85 7 22 3 2 6 4 5

05/06 Forbs 65 26 39 39 40

3 25 3 70 43 88 65 30 35 25

05/06 Roots

05/06 05/06 Insects Animal Matter

1 4 4 8 -

1 1 8 1 1 -

27 41 29 27

2 10 1 1 9 2

05/06 Fruit 2 12 4 20 25

05/06 Cones/Seeds

6 2 1 1 30 64 51 67

2 -

Table 20. The percent volume of digestible dry matter from major food items consumed seasonally by grizzly bears in the Wapiti study area in Alberta, Canada, 2006. Season

Early Jun Late Jun Early Jul Late Jul Early Aug Late Aug Early Sept

8 9 6 12 12 6 10

HorsetailGraminoids Forbs 11 32 28 5 2 4 0

6 9 2 5 7 6 1

3 46 36 17 37 22 2

Roots Insects Animal Matter Fruit Cones/Seeds 0 0 0 0 0 46 48

4 3 1 2 19 17 24

66 7 0 8 0 0 20

9 1 31 60 35 5 3

1 2 2 2 1 0 3

2 -

CHAPTER 5: ANIMAL HEALTH Dr. David Janz, Western College of Veterinary Medicine, University Of Saskatchewan. Dr. Marc Cattet, Canadian Cooperative Wildlife Health Centre, Saskatoon, Saskatchewan. Dr. Matt Viijan, Department of Biology, University of Waterloo, Waterloo, Ontario. Gordon Stenhouse, Foothills Model Forest, Hinton, Alberta.

Objectives Within this program element the team will strive to develop a sensitive proteomics technique for detecting long-term physiological stress in grizzly bears based on analysis of expression profiles of multiple stress-activated proteins found in many body tissues. We will also develop Animal Health profiles for individual bears and attempt to determine relationships between long-term physiological stress and other measures of health (longevity, growth, reproduction, immunity, and activity) in grizzly bears. A team of scientists from the University of Saskatchewan and the University of Waterloo are working in a collaborative yet independent manner to attempt to develop new laboratory techniques to measure and quickly quantify stress in grizzly bears. In addition to the aforementioned laboratory work it is important to collect samples from grizzly bears in landscapes with different levels of human use and landscape structure in order to determine relationships between measures of health and landscape condition. Gordon Stenhouse leads a field capture program from April-July to gather these samples and other necessary data from grizzly bears along the eastern slopes. Animal Health Update Laboratory work The development of long-term stress biomarkers is occurring along two complementary paths. One path, led by Matt Vijayan (U. of Waterloo) is the development and validation of blood serum-based indicators of long-term stress. The other path, led by David Janz (U. of Saskatchewan), is the development of a sensitive protein array for detecting long-term physiological stress. At the U. of Waterloo, we are evaluating the usefulness of corticosteroid-binding globulin (CBG) as a serum-based indicator of long-term stress in grizzly bears. In 2006 we purified CBG from bear serum using affinity chromatography and the purified protein was confirmed as CBG by SDS-PAGE and immunodetection (western blotting) using antibodies specific to human CBG. The purified protein was injected into rabbits to generate antiserum to bear CBG. We are currently purifying the antibodies using chromatography. We have also analysed other proven stress markers, including plasma cortisol and heat shock protein 60 (hsp60) concentrations in grizzly bear serum using commercially available radioimmunoassay and sandwich ELISA kits, respectively. We are in the process of measuring serum hsp70 levels using a commercially available ELISA kit. The work is progressing as planned and we have provided laboratory results from current samples in March 2007 as scheduled.

At the U. of Saskatchewan, we are developing a sensitive antibody-based protein array that specifically targets 30-40 proteins that exhibit altered expression (increased or decreased) during long-term physiological stress. As indicated in our grant submission, the major milestone in Year 1 was to evaluate commercially available antibodies against stress proteins for their cross-reactivity with bear proteins. To date we have tested, using Western blotting, a total of 168 antibodies. We have confirmed bear protein cross-reactivity to 13 of these antibodies. In addition to this approach, we have added two complementary approaches to identify stress proteins of interest: (1) two-dimensional gel electrophoresis with protein identification using mass spectrometry, and (2) commercially available antibody-based protein arrays. We will continue using all of these approaches to develop our list of 30-40 antibodies for use in the array. Our next step will be to print antibodies onto slides and optimize/validate conditions for the protein array. We anticipate this to occur in March-April 2007. Grizzly Bear Health Assessment We developed health profiles for all grizzly bears captured in 2006 using the collective results of several procedures, including (i) physical examination, (ii) physical measurements, (iii) physiological measurements, (iv) the extraction of a premolar tooth for age determination, and (v) the collection of blood for laboratory analyses, i.e., haematology, serum biochemistry, and assays for immunological and reproductive function. When combined with health information collected for the Foothills Model Forest Grizzly Bear Research Program during 1999-2005, we have health profiles available for 145 individual grizzly bears, with some bears captured more than once, bringing the total number of captures to 261. In 2007, we will expand the health profiles of each bear by adding information regarding their movement rates (an index of activity) from the GPS location data, and stress levels from the laboratory work. After assembling data all available health data from grizzly bears for the period 1999-2006, we began to explore the relationships between health, stress, and landscape variables. This work was started in September 2006 and the statistical analyses were based on small sample sizes representing only a subset of the available information. However, the focus was not toward the results per se, but instead toward developing the appropriate methodologies to merge and analyze this complex data set. This preliminary exploration allowed us to identify and address potential obstacles (e.g., repeated measures, confounding variables, data redundancy), and to plan an efficient strategy to conduct data mergers and analyses in year 2 of our program. We have also developed new approaches to quantify and consolidate the large number of health variable measures into a â&#x20AC;&#x153;health profileâ&#x20AC;? in four key categories. This new system will be used to compare individual bears in different population units and specific bears where data was collected in different years. Our intent is to now prepare these new health profile measures for each bear that we have handled since 1999 and then proceed with looking at these measures in relation to landscape conditions and change. This work will be ongoing in 2007 and represents the second year of our Innovation and Science research grant. Overall the research team feels we have made important and significant progress in the animal health component of this research project and major analysis will be underway in

2007 as new laboratory analysis are completed and annual landscape conditions metrics are assembled. At this time all components of the grizzly bear health research effort are on schedule and budget.

Chapter 6. KNOWLEDGE TRANSFER AND PRODUCT DELIVERY Gordon Stenhouse, Foothills Model Forest, Hinton, Alberta.

Objectives The scope and scale of delivering these new products and information to the numerous government departments and agencies is a key step within this program. It is our intention to not only provide digital files (maps, etc.) for each department but to also develop new GIS applications to enhance decision making and to provide the necessary training to enable new users to understand and maximize the use of these products in their work environments. Our research team feels that this program element is key to maximizing the benefits of this leading edge research undertaking. Knowledge Transfer and Product Delivery Over the course of the past year (2006) our research team has worked closely with program partners (e.g. Petro-Canada and others) and Alberta SRD in using the new maps and models that this project has developed. These products are now being used for the evaluation of land use management activities in grizzly bear habitat in Alberta. Recently our program team has been called upon to assist SRD and Sundrie Forest Products (and other FMA holders) in evaluating â&#x20AC;&#x153;emergency harvest plans for mountain pine beetleâ&#x20AC;?. We have also developed and distributed (January 2006) maps, models, and GIS applications to aid in using these products to all program partners. SRD staff has full access to these products through the forest management branch in Edmonton. Our research team has also now entered into an agreement (with the support of our Innovation and Science Program Partners) with ENFORM for the development and delivery of the planned training program. This new partnership will allow program partners from government, the forestry and oil and gas sectors to take a common training program thus allowing standardization of information and experience within the training environment. A training program is scheduled with representatives of our program partner groups (Forestry sector, Oil and Gas sector, Provincial Government, and private consultants) to provide directed input on the training course. This stakeholder input is important to ensure that the needs of program partners will be met. Agreements are now in place between ENFORM and the FMF is regards to roles and responsibilities concerning the development and delivery of this training program. Our research scientists are now preparing course materials for review by the ENFORM education and training staff. Our goal is to have this training program ready for delivery to stakeholder groups in the fall of 2007. The program products from our 2006 program have been completed and packaged for delivery and are included in a DVD being sent to program partners along with this annual report. Partners will receive: 1. A new remote sensing based landcover map that encompasses an area south of Grande Prairie to the Montana border (see figure 31). This map represents 24 landsat scenes and is updated to circa 2005 landscape conditions. It also has accompanying data layers for; crown closure, LAI (3 seasons), leading species and NDVI. 2. A new resource selection function (RSF) map coverage for the land area mapped in #1 above. This new RSF layer is an improvement over previous products with the

addition of new GPS bears location data, improved landcover mapping and clear delineation of population unit boundaries. 3. A detailed PowerPoint presentation providing information on the overall program and recommendations on the process and procedures to follow to conduct assessments of land use activities on grizzly bears and their habitats.

Figure 31: Map area where landcover mapping has been completed circa 2005 conditions. Work is continuing on the completion of the following products that are expected to be delivered to program partners by June of 2007; mortality risk layer, graph theory movement corridor layer for the area south of Highway 1 and revised GIS applications to conform to the newest RSF map products.

APPENDIX 1: RECENT PUBLICATIONS/REPORTS FROM THE FMF GRIZZLY BEAR RESEARCH PROGRAM

REFEREED JOURNAL ARTICLES (PUBLISHED/ACCEPTED) Linke, J., and S.E. Franklin. 2007. Interpretation of landscape structure gradients based on satellite image classification of land cover. Canadian Journal of Remote Sensing, in press. Wunderle, A., S.E. Franklin, and X. Guo. 2007. Regenerating boreal forest structure estimation using SPOT-5 pan-sharpened imagery. International Journal of Remote Sensing, in press. Wulder, M.A., S.E. Franklin, J.C. White, J.Linke and S. Magnussen. 2007. An accuracy assessment framework for large-are land cover classification products derived from mediumresolution satellite data. International Journal of Remote Sensing: in press. REFEREED JOURNAL ARTICLES (SUBMITTED) Cattet, M., J. Boulanger, G. Stenhouse, R. Powell, and M. Reynolds. 2007. Long-term effects of capture and handling in grizzly bears and black bears: implications for ecological studies, conservation programs, and wildlife management. Ecological Applications, submitted. Klassen, J.N., G.J. McDermid, and M. Hall-Beyer. 2007. Remote sensing of phenology: strategies for noise reduction in NDVI time series. Journal of Applied Remote Sensing, submitted. McDermid, G.J., R.J. Hall, G.A. Sanchez-Azofeifa, S.E. Franklin, G.B. Stenhouse, T. Kobliuk, and E.F. LeDrew. 2007. Remote sensing and forest inventory for wildlife habitat assessment. Journal of Wildlife Management, submitted. McDermid, G.J., S.E. Franklin, and E.F. LeDrew. 2007. A multi-attribute approach to mapping vegetation and land cover over large areas in support of wildlife habitat mapping. Remote Sensing of Environment, submitted. McDermid, G.J., S.E. Franklin, and E.F. LeDrew. 2007. Radiometric normalization and continuous-variable model extension for the operational mapping of vegetation and land cover over large areas. International Journal of Remote Sensing, submitted. Pape, A.D. and S.E. Franklin. 2007. MODIS-based change detection for grizzly bear habitat mapping in Alberta. Photogrammetric Engineering and Remote Sensing, submitted.

CONFERENCE PRESENTATION/POSTER (SPECIFY) Carlson, R.I., M.R.L. Cattet, G.B. Stenhouse and D.M. Janz. Development of a protein microarray to detect long-term stress in grizzly bears. Poster presentation, The Wildlife Society annual meeting, Anchorage, AK, Sept 23-27, 2006. Carlson, R.I., M.R.L. Cattet, G.B. Stenhouse and D.M. Janz. Monitoring long-term stress in grizzly bears using protein microarray technology. Poster presentation, World Microarray Congress, Vancouver, BC, March 24-25, 2006. Hamilton, J., M. Obbard, M. Cattet, G. Stenhouse, and M.M. Vijayan. 2006. Cortisol-binding globulin as a marker of chronic stress in bears. Oral presentation, The University of Waterloo Graduate Symposium, Waterloo, ON, April 3, 2006. Hamilton, J., M. Obbard, M. Cattet, G. Stenhouse, and M.M. Vijayan. 2006. Is cortisolbinding globulin a biomarker of chronic stress in bears? Poster presentation, Canadian Society of Zoologists Annual Meeting, Edmonton, AB, May 4-6, 2006. McDermid, G.J., S.E. Franklin, G.B. Stenhouse. 2006. Remote sensing and forest inventory for grizzly bear habitat mapping and resource selection analysis. Oral presentation, American Society of Photogrammetry and Remote Sensing Annual Conference, Reno, NV, April 2, 2006. Stenhouse, G.B. Grizzly bears in the wild: grizzly bear science and research in Alberta. Invited presentation, Smithsonion Institution, Washington DC, August 3, 2006. Stenhouse, G.B. Recognizing the needs of grizzlies: new knowledge and tools for land management decisions in Alberta. Invited presentation, Petroleum Technology Alliance of Canada Ecological Research Forum and Resource Access Technology Workshop, Kananaskis, AB, October 10-11, 2006. OTHER:

TECHNICAL REPORTS, NON-REFEREED ARTICLES ETC. (SPECIFY)

Pape, A.D. 2006. Multiple spatial resolution image change detection for environmental management applications. M.Sc. thesis, Department of Geography, University of Saskatchewan, 97p. Wunderle, A.L. 2006. Sensitivity of multi-resolution satellite sensor imagery to regenerating forest age and site preparation for wildlife habitat analysis. M.Sc. thesis, Department of Geography, University of Saskatchewan, 89p.

APPENDIX 2: PUBLISHED PAPERS AND THESES RESULTING FROM THE FMFGBP â&#x20AC;&#x201C; March 2007. Boulanger, J., M. Proctor, S. Himmer, G. Stenhouse, D. Paetkau, J. Cranston. 2006. An empirical test of DNA mark-recapture sampling strategies for grizzly bears. Ursus 17:149-158. Boulanger, J., G. Stenhouse, R. Munro. 2004. Sources of heterogeneity bias when DNA mark-recapture sampling methods are applied to grizzly bear (Ursus arctos) populations. Journal of Mammalogy 85:618-624. Cattet, M.R., A. Bourque, B.T. Elkin, K.D. Powley, D.B. Dahlstrom, N.A. Caulkett. 2006. Evaluation of the potential for injury with remote drug-delivery systems. Wildlife Society Bulletin 34:741-749. Cattet, M.R.L., K. Christison, N.A. Caulkett and G.B. Stenhouse. 2003. Physiologic responses of grizzly bears to different methods of capture. Journal of Wildlife Diseases 39(3):649-654. Cattet, M.R.L., N.A. Caulkett, and G.B. Stenhouse. 2003. Anesthesia of grizzly bears using xylazine-zolazepam-tiletamine or zolazepam-tiletamine. Ursus 14(1):88-93. Cattet, M.R.L., N.A. Caulkett, M.E. Obbard and G.B. Stenhouse. 2002. A body-condition index for ursids. Canadian Journal of Zoology 80:1156-1161. Frair, J.L., S.E. Nielsen, E.H. Merrill, S. Lele, M.S. Boyce, R.H.M. Munro, G.B. Stenhouse, and H.L Beyer. 2004. Removing GPS-collar bias in habitat-selection studies. Journal of Applied Ecology 41, 201-212. Franklin, S. E., P. K. Montgomery, and G. B. Stenhouse. 2005. Interpretation of land cover using aerial photography and satellite imagery in the Foothills Model Forest of Alberta. Canadian Journal of Remote Sensing 31:304-313. Franklin, S.E., D.R. Peddle, J.A. Dechka, G.B. Stenhouse. 2002. Evidential reasoning with Landsat TM, DEM and GIS data for landcover classification in support of grizzly bear habitat mapping. International Journal of Remote Sensing 23(21):4633-4652. Franklin, S.E., M.B. Lavigne, M.A. Wulder, and G.B. Stenhouse. 2002. Change detection and landscape structure mapping using remote sensing. The Forestry Chronicle 78(5):618-625. Franklin, S.E., M.J. Hansen, G.B. Stenhouse. 2002. Quantifying landscape structure with vegetation inventory maps and remote sensing. The Forestry Chronicle 78(6):866875.

Franklin, S.E., G.B. Stenhouse, M.J. Hansen, C.C. Popplewell, J.A. Dechka, D.R. Peddle. 2001. An integrated decision tree approach (IDTA) to mapping land cover using satellite remote sensing in support of grizzly bear habitat analysis in the Alberta Yellowhead ecosystem. Canadian Journal of Remote Sensing 27(6):579-592. Gau, R.J., R. Mulders, L.M. Ciarniello, D.C. Heard, C.B. Chetkiewicz, M. Boyce, R. Munro, G. Stenhouse, B. Chruszcz, M.L. Gibeau, B. Milakovic, K. Parker. 2004. Uncontrolled field performance of Televilt GPS-Simplex collars on grizzly bears in western and northern Canada. Wildlife Society Bulletin 32:693-701. Huettmann, F., S.E. Franklin, G.B. Stenhouse. 2005. Predictive spatial modelling of landscape change in the Foothills Model Forest. Forestry Chronicle 81:525-537. Hunter, A., N. El-Sheimy, G. Stenhouse. 2005. GPS/Camera Collar Captures Bear Doings. http://www.gpsworld.com/gpsworld/article/articleDetail.jsp?id=146689 (this will only be here until September 2006 or so). Linke, J. 2003. Using Landsat TM and IRS imagery to Detect Seismic Cutlines: Assessing their Effects on Landscape Structure and on Grizzly Bear (Ursus arctos) Landscape Use in Alberta. MSc. Thesis. Department of Geography, University of Calgary, Calgary, Alberta, Canada. Linke, J., S.E. Franklin, F. Huettmann and G.B. Stenhouse. 2005. Seismic cutlines, changing landscape metrics and grizzly bear landscape use in Alberta. Landscape Ecology 20:811-826. McDermid, G. J., S.E. Franklin and E.F. LeDrew. 2005. Remote sensing for large-area habitat mapping. Progress in Physical Geography 29:449-474. McDermid, G.J. 2005. Remote Sensing for Large-Area, Mult-Jurisdictional Habitat Mapping. PhD. Thesis. Department of Geography, University of Waterloo, Waterloo, Ontario, Canada. Mowat, G., D.C. Heard, D.R. Seip, K.G. Poole, G.Stenhouse, D.W. Paetkau. 2005. Grizzly Ursus arctos and black bear U. americanus densities in the interior mountains of North America. Wildlife Biology 11: 31-48. Munro, R.H.M, S.E. Nielsen, M.H. Price, G.B. Stenhouse, M.S. Boyce. 2006. Seasonal and diel patterns of grizzly bear diet and activity in west-central Alberta. Journal of Mammology 87:1112-1121. Nielsen, S.E., G.B. Stenhouse, M.S. Boyce. 2006. A habitat-based framework for grizzly bear conservation in Alberta. Biological Conservation 130:217-229. Nielsen, S.E. 2004. Habitat ecology, Conservation, and Projected Population Viability of Grizzly Bears (Ursus arctos L.) in West-Central Alberta, Canada. PhD Thesis.

Department of Biological Sciences, University of Alberta, Edmonton, Alberta, Canada. Nielsen, S.E., M.S. Boyce, and G.B. Stenhouse. 2004. Grizzly bears and forestry I: selection of clearcuts by grizzly bears in west-central Alberta. Canada.Forest Ecology and Management 199:51â&#x20AC;&#x201C;65. Nielsen, S.E., R.H.M. Munro, E. Bainbridge, M.S. Boyce, and G.B. Stenhouse. 2004. Grizzly bears and forestry II: distribution of grizzly bear foods in clearcuts of westcentral Alberta, Canada. Forest Ecology and Management 199:67â&#x20AC;&#x201C;82. Nielson, S.E., M.S. Boyce, G.B. Stenhouse, R.H.M. Munro. 2003. Development and testing of phenologically driven grizzly bear habitat models. Ecoscience 10(1):1-10. Nielson, S.E., M.S. Boyce, G.B. Stenhouse, and R.H.M. Munro. 2002. Modeling grizzly bear habitats in the Yellowhead ecosystem of Alberta: taking autocorrelation seriously. Ursus 13:45-56. Pape, A. 2006. Multiple Spatial Resolution Image Change Detection for Environmental Management Applications. MSc. Thesis. Department of Geography, University of Saskatchewan, Saskatoon, Saskatchewan, Canada. Pereverzoff, J.L. 2003. Development of a Rapid Assessment Technique to Identify Human Disturbance Features in a Forested Landscape. MSc. Thesis, Department of Geography, University of Calgary, Calgary, Alberta, Canada. Popplewell, C., S.E. Franklin, G.B. Stenhouse, and M. Hall-Beyer. 2003. Using landscape structure to classify grizzly bear density in Alberta Yellowhead Ecosystem bear management units. Ursus 14: 27-34. Popplewell, C. 2001. Habitat Structure and Fragmentation of Grizzly Bear Management Units and Home Ranges in the Alberta Yellowhead Ecosystem. MSc. Thesis. Department of Biological Sciences, University of Calgary, Calgary, Alberta, Canada. Ritson-Bennett, R.A. 2003. Assessing the Effects of a Heli-portable 3D Seismic Survey on Grizzly Bear (Ursus arctos horribilis) Distribution. MSc. Thesis, Department of Geography, University of Calgary, Calgary, Alberta, Canada. Roever, C.L. 2006. Grizzly Bear (Ursus arctos L.) Selection of Roaded Habitats in a Multiuse Landscape. M.Sc. Thesis, University of Alberta, Edmonton, Alberta, Canada. Schwab, B.L. 2003. Graph Theoretic Methods for Examining Landscape Connectivity and Spatial Movement Patterns: Application to the FMF Grizzly Bear Research. MSc Thesis. Department of Geography, University of Calgary, Calgary AB.

Stenhouse, G.B., J. Boulanger, J. Lee, K. Graham, J. Duval, J. Cranston. 2004. Grizzly bear associations along the eastern slopes of Alberta. Ursus 16:31-40. Stenhouse, G.B. and K.Graham. (Eds.). 2005. Foothills Model Forest Grizzly Bear Research Program 1999-2003 Final Report. 289 pp. Stenhouse, G.B., R. Munro and K.Graham. (Eds.). 2003. Foothills Model Forest Grizzly Bear Research Program 2002 Annual Report. 162 pp. Stenhouse, G.B., R. Munro. (Eds.). 2002. Foothills Model Forest Grizzly Bear Research Program 2001 Annual Report. 126 pp. Stenhouse, G. and R. Munro. (Eds.). 2001. Foothills Model Forest Grizzly Bear Research Program 2000 Annual Report. 87 pp. Stenhouse, G. and R. Munro. (Eds.). 2000. Foothills Model Forest Grizzly Bear Research Program 1999 Annual Report. 98 pp. Wasser, S.K., B. Davenport, E.R. Ramage, K.E. Hunt, M. Parker, C. Clarke, and G.B. Stenhouse. 2004. Scat detection dogs in wildlife research and management: Application to grizzly and black bears in the Yellowhead Ecosystem, Alberta, Canada. Canadian Journal of Zoology 82:475-492. Wulder, M. A., and S. E. Franklin, eds., 2003, Remote Sensing of Forest Environments: Concepts and Case Studies, Kluwer Academic Publishers, Boston, MA, 519p. Wunderle, A.L. 2006. Sensitivity of multi-resolution satellite sensor imagery to regenerating forest age and site preparation for wildlife habitat analysis. M.Sc. thesis, Department of Geography, University of Saskatchewan, 89p.

APPENDIX 3: FMF GRIZZLY BEAR PROJECT PARTNERS •Ainsworth Lumber •Alberta Conservation Association •Alberta SRD •Alberta Fish and Game •Alberta Newsprint •Alberta Advanced Education and Technology (Innovation and Science) •Anadarko •Anderson Exploration Ltd. •Anderson Resources Ltd. •AVID Canada •BP Canada Energy Company •Banff National Park •BC Oil & Gas Commission •Buchanan Lumber-Tolko •Burlington Resources Ltd. •Canada Centre for Remote Sensing •Canadian Hunter •Canadian Wildlife Service •Canfor •Cardinal River Operations •Canadian Forest Service •Conoco Phillips Ltd. •Conservation Biology Institute •Devon Canada Corp. •DMI •Elk Valley Coal •Enbridge Inc. •EnCana Corp. •Environment Canada –HSP •Foothills Model Forest •Fording Coal •FRIAA •GeoAnalytic Ltd. •Gregg River Resources •Husky Energy •Jasper National Park •Komex International Ltd. •Lehigh Inland Cement •Luscar Ltd.-Coal Valley •Manning Forestry Research •Millar Western Ltd. •Mountain Equipment Co-op •Nexen

•Natural Resources Service •Northrock Resources Ltd. •NSERC •Petro Canada •Peyto Exploration •Precision Drilling Ltd. •PTAC (CAPP) •Rocky Mountain Elk Foundation •Shell Canada •Spray Lake Sawmills •Suncor Energy •Sundance Forest Industries •Talisman Energy Ltd. •Telemetry Solutions •Trans Canada Pipelines •University of Alberta •University of Calgary •University of Lethbridge •University of Saskatchewan •University of Washington •Veritas •West Fraser Hinton Wood Products Blue Ridge Lumber Sundre Forest Products Slave Lake Pulp •Western College of Veterinary Medicine •Weyerhaeuser Ltd. •World Wildlife Fund