Managing for Bears in Forested Environments
Revelstoke BC Canada - October 17-19, 2000

The Columbia Mountains Institute of Applied Ecology (CMI) hosted this three day workshop. On October 17 and 18, three sessions addressed topics related to bear population census, forestry and bears, and resolving bear/human conflicts. On October 19, half day field trips augmented the workshop sessions. Over 280 persons attended all or part of the workshop.

The Columbia Mountains Institute of Applied Ecology would like to thank the following agencies for their financial and in-kind support for this workshop:

• Azimuth Forestry and Mapping Solutions

• Bear Brewing

• Bear in Mind Gifts

• Canadian Mountain Holidays

• Columbia Basin Fish and Wildlife Compensation Program

• Columbia Basin Trust/ Affected Areas and Communities Initiatives

• Columbia Power Corporation

• Downie Bonus Fund

• Ministry of Forests

• Mt. Begbie Brewing

• Parks Canada

• Revelstoke Community Forest Corporation

• Tembec Forest Products

The following summaries were provided by authors who presented papers at the conference. Some presenters did not submit a summary. Please contact the authors directly as necessary.

List of Presentations

• Why Do We Need Better Data on Bears? John Woods, Parks Canada
• Meta-Analysis of DNA Mark-Recapture Projects in British Columbia, John Boulanger, Integrated Biological Research
• Grizzly Bear Abundance and Distribution Survey in the Central Purcell Mountains of Southeast BC, Michael Proctor, University of Calgary
• Grizzly Bear Occurrence Relative to Broad-Scale Factors of Habitat and Human Influence near Golden BC, Clayton Apps, Aspen Wildlife Research
• Foothills Model Forest Grizzly Bear Research Program - DNA Component, Gordon Stenhouse, Alberta Environment, and Garth Mowat, Aurora Wildlife Research
• Genetic and Endocrine Monitoring of Population and Disturbance Parameters in Ursids Using Scat Detection Dogs, Sam Wasser, University of Washington
• DNA Degradation in the Field, Curtis Strobeck, University of Alberta
• Monitoring Population Trends in Glacier National Park, Montana Using Non-invasive Genetic Sampling, David Roon, University of Idaho
• Genotyping Errors in DNA-Based Inventories, David Paetkau, Wildlife Genetics International
• Grizzly Bear Population Estimation and Hunt Management, Guy Woods, BC Ministry of Environment, Lands and Parks
• Data for Decision Making - Environmental Assessments in BC, Matt Austin, BC Ministry of Environment, Lands and Parks

Session Two - Managing Forests for Bears

• Habitat, Foods, Energy Balance of Forest Inhabiting Bears, Bruce McLellan, BC Ministry of Forests
• Grizzly Bear Use of Avalanche Chutes in the Columbia Mountains, Roger Ramcharita, University of BC
• Bears, Berries and Silviculture, Tony Hamilton, BC Ministry of Environment
• Grizzly Bear Habitat Management Guidelines for Avalanche Tracks, Garth Mowat, Aurora Wildlife Research, and Matt Besko, Alberta Environment
• Science and Road Access Management in Montana: All Roads Lead to the Court House, Richard Mace, Montana Department of Fish, Wildlife and Parks
• Access: North Fork of the Flathead Experience, Fred Hovey, BC Ministry of Forests
• A Comparison of Resource Selection Patterns between Mountain and Plateau Grizzly Bears in the Parsnip River, BC, Lana Ciarniello, University of Alberta
• Black Huckleberry Biology and Management, Evelyn Hamilton, BC Ministry of Forests
• Bears in Timber Stands: Damage and Preventative Measures, Dale Nolte, US Dept of Agriculture.

Session Three - Living in Bear Country

• Living with Bears in British Columbia, Richard Daloise, BC Ministry of Environment
• People and Bears - How to Prevent Problems, Darcy Lutz, BC Conservation Foundation
• Developing and Delivering Public Education Messages - Debby Robinson
• Whistler's Bear Aware Program, Sylvia Dolson, J.J. Whistler Bear Foundation
• Canmore - Bear-Proofing a Community, Andreas Comeau, Town of Canmore, Alberta
• Moving Bears - Does it Work? John Woods, Parks Canada
• Aversive Conditioning Results on 12 Radio Collared Problem Bears, Hal Morrison, Parks Canada
• Electric Bear Fencing at Landfills, Jeff Marley, Margo Supplies
• Behavior of Grizzly Bears Before and After Landfill Closures in North-Central BC, Mari Wood, Peace-Williston Fish and Wildlife Compensation Program
• Northern Bear Awareness, Tony Boschmann
• The Partners in Life Project - Keeping Bears Wild, Carrie Hunt, Wind River Bear Institute, and Tim Manley, Montana Department of Fish, Wildlife, and Parks

Evening Presentation

• Causes and Avoidance of Bear Attacks, Stephen Herrero, University of Calgary

Columbia Mountains Institute of Applied Ecology
Box 2568 Revelstoke BC Canada V0E 2S0
office@cmiae.org
250-837-9311

Contact information:
John Woods, Parks Canada,
Box 350, Revelstoke, British Columbia, Canada, V0E 2S0
250-837-7527, john_woods@pch.gc.ca

As of the time of the CMI Revelstoke meeting there have been 13 projects in British Columbia which have attempted to use DNA methods with grizzly bear populations. The majority of mark-recapture efforts have focused upon the estimation of population size and density (Woods et al., 1999). The main challenges in obtaining reliable population estimates have been meeting the assumption of population closure, minimizing capture probability variation, and obtaining adequate sample sizes in term of bear recaptures (White et al., 1982). Additional assumptions regarding the genetic analysis of samples are also required for DNA mark-recapture as discussed further in the talk.

Meeting the assumption of population closure is one of the biggest challenges to mark-recapture and most other methods which attempt to estimate population size of large carnivores (Smallwood & Schonewald, 1996). Analysis of data from BC suggests that estimation of population size of bears which frequent the sampling grid and surrounding area (the "superpopulation" of bears as defined by Kendall,(1999)) is possible with this technique if studies are designed appropriately to allow use of robust estimation models. The scaling of superpopulation estimates to density estimates is difficult unless sampling grid areas are topographically closed or radio collared bears are used to index movements across grid boundary areas (Powell et al. 2000). An ad-hoc procedure to gain further inference into the sources of closure violation and provide partially corrected estimates is discussed (Boulanger & McLellan, In review). Results of field and simulation studies suggest that mark-recapture data is too sparse to reliably detect capture probability variation using program CAPTURE statistical tests. However, simulation studies suggest that the heterogeneity estimators in program CAPTURE are reasonably robust to most forms of capture probability variation as long as sample size levels are adequate. Case studies are discussed to further illustrate the importance of sampling design and analysis strategy.

Potential methods for estimation of population trend are being explored for use with bear populations as an alternative to costly population estimation-based methods. In general, estimation of trend and survival is relatively robust to issues such as population closure and capture probability variation, which challenge the estimation of population size. Recent developments in mark-recapture methodology allow further inference into biological hypothesis and population trends from data sets (White & Burnham, 1999). This shift has allowed researchers and managers to gain more ecological insight from their data. Basically, the question "What is the population size?" is being replaced by "What factors influence the trend and survival of this population across time and space?" (Anderson et al. 1995; Cooch & White 2000). Results of simulation studies are given to further illustrate these newer methods.

In conclusion, the results of BC studies show that reliable estimates of superpopulation size are possible when study design objectives are met. The estimation and interpretation of population density is a greater challenge given the wide spread movements of bears and the patchy spatial distribution of bears on sampling grids (Clayton Apps, In prep). Quantitative tools are available to produce estimates of population size, density, and trend however the ultimate quality and reliability of estimates is determined by sound attention to sampling design, field implementation, and genetic analysis.

Anderson, D. R., White, G. C. & Burnham, K. P. (1995). Some specialized risk assessment methodologies for vertebrate populations. Environmental and Ecological Statistics 2: 91-115.

Cooch, E. & White, G. C. (2000). Analysis of encounter data from marked animal populations: Program MARK: A gentle introduction, Cornell University (Available online at: canuck.dnr.cornell.edu/mark/)

Kendall, W. L. (1999). Robustness of closed capture-recapture methods to violations of the closure assumption. Ecology, 80: 2517-2525.

Powell, L. A., Conroy, M., Hines, J. E. & Krementz, D. G. (2000). Simultaneous use of mark-recapture and radiotelemetry to estimate survival, movement, and capture rates. J. Wildlife Manage. 64, 302-313.

Smallwood, K. S. & Schonewald, C. (1996). Scaling population density and spatial pattern for terrestrial carnivores. Oecologia 105: 329-335.

White, G. C., Anderson, D. R., Burnham, K. P. & Otis, D. L. (1982). Capture-recapture and removal methods for sampling closed populations. Los Alamos National Laboratory. Available online at: . http://www.cnr.colostate.edu/class_info/fw663/White1982/WhiteList.html

White, G. C. & Burnham, K. P. (1999). Program MARK: Survival estimation from populations of marked animals. Bird Study Supplement 46: 120-138. http://www.cnr.colostate.edu/~gwhite/mark/mark.htm)

Woods, J. G., Paetkau, D., Lewis, D., McLellan, B. L., Proctor, M. & Strobeck, C. (1999). Genetic tagging free ranging black and brown bears. Wildlife Society Bulletin 27: 616-627

See www.ecological.bc.ca/refs.htm for more DNA MR papers.

Contact Information:
John Boulanger, Integrated Ecological Research
924 Innes St. Nelson, BC V1L 5T2,
250-352-2605, boulange@ecological.bc.ca, www.ecological.bc.ca

Contact Information:
Michael Proctor, University of Calgary
250-353-7349, mproctor@netidea.com

Conservation of wide-ranging species requires consideration of habitat and population distribution at scales that extend from geographic range to micro-sites. Although the scale of regional populations is often considered in planning decisions, the necessary information on habitat potential and population distribution are typically lacking. For grizzly bears, hair-snag sampling and DNA analysis techniques, developed for population estimation, hold promise for understanding and modeling population distribution and abundance at broad spatial scales.

I describe a case study in relating broad-scale factors of habitat and human influence to grizzly bear distribution and abundance in the West Slopes Study Area, near Golden, British Columbia. Grizzly bear occurrence was sampled using hair-snag methods in 3 different sampling grids during June and July, 1996 to 1998. DNA analysis confirmed 244 independent grizzly bear visits to 168 station and sample session combinations, while 845 station/sessions received no confirmed visits. I analyzed grizzly bear occurrence relative to 24 variables of habitat and human influence derived from existing digital data sources. Within a GIS, each variable was "aggregated" at each of 3 spatial scales, reflecting mean attribute composition within circular landscapes of successively larger radii. For some variables, associations with grizzly bear occurrence depended on the spatial scale considered. However, grizzly bear occurrences were generally associated with broad landscapes of high elevation and complex terrain of predominantly southern aspect, with higher composition of alpine, avalanche chutes, burns, and other open habitats, while vegetation productivity was relatively low. Although grizzly bears were positively associated with old forest composition at broader scales, occurrences were positively related to only open or non-forested habitats at the finest scale. Negative associations with human influence were evident across scales.

Variables that were at least marginally related to grizzly bear occurrence were entered into a multivariate analysis to define a minimum combination that best described grizzly bear distribution within the sampling area. The resulting model was highly significant, correctly classifying 77% of the sample dataset. I applied the grizzly bear occurrence model to the greater West Slopes Study Area using algebraic GIS modeling. The GIS-based model exhibited significant predictive power in an independent validation analysis against landscapes "occupied" by 49 radio-collared grizzly bears. I then used the model to spatially interpolate a population estimate (J. Boulanger, unpublished) derived for the combined multi-year sampling area. This illustrated predicted grizzly bear density and distribution over the greater study area, from which an extrapolated population estimate was possible. I discuss important limitations of this and other spatial modeling approaches, and provide suggestions for future sampling designs.

Boyce, M. S., and L. L. McDonald. 1999. Relating populations to habitats using resource selection functions. TREE 14:268-272.

Klopatek, J. M., and R. H. Gardner. 1999. Landscape ecological analysis: issues and applications. Springer-Verlag, New York, NY.

Manley, B. F. J., L. L. McDonald, and D. L. Thomas. 1993. Resource selection by animals: statistical design and analysis for field studies. Chapman and Hall, New York, NY.

Morrison, M. L., B. G. Marcot, and R. W. Mannan. 1998 Wildlife-habitat relationships: concepts and applications. University of Wisconsin Press, Madison, Wisconsin.

Woods, J. G., D. Paetkau, D. Lewis, B. N. McLellan, M. Proctor, and C. Strobeck. 1999. Genetic tagging of free-ranging black and brown bears. Wildlife Society Bulletin 27:616-627.

Contact Information:
Clayton Apps, Aspen Wildlife Research
403-270-8663, aspen@cadvision.com

Given the poor capture success at rub pad sites and during the third session, we combined the live capture and hair capture datasets to estimate population size in order to increase the sample size of the mark-recapture dataset. There were 71 captures of 48 different bears in the combined dataset; we created a fourth capture session using bears live captured before hair capture began. The population estimate was 107 using Chao's Model Mth. This estimate is likely to be biased high because the assumption of geographic closure was unlikely to have been met (White et al. 1982). We used the method of Kenward et al. (1981) to correct for closure which essentially weights the point estimate by average residency. We had sufficient locations (mean n=174, range 29-338) to estimate residency for 12 bears. The final population estimate for this study area was 100 bears (95% CI 66-185) which generates a density estimate of 18.7 bears / 1000 km2. This is a moderate density when compared to other areas in the Northern Rockies. This result should be viewed cautiously in light of issues relating to capture probability and study area boundary closure. Further analyses of available data sets from this project are currently being analyzed to provide more insights on these topics.

We cannot recommend the use of carpet tack rub pads for grizzly bears, we did achieve somewhat better success with black bears. We only tried one bait on rub pads in this study and perhaps other more effective baits exist. The use of 3 sample sessions can save considerable field cost but was not well advised in our case given the large variation in capture heterogeneity in our data. Live capture may affect subsequent capture probability causing heterogeneity in capture probabilities. Designs using 2 or 3 sessions with our current study design appears to carry significant risk and may only be advisable when previous experience has shown that there is little risk of behavior or heterogeneity in capture probabilities. We suggest future workers attempt to put at least 10 guard hairs (15 or 20 may be better) and over 30 under fur in samples to ensure genotyping success.

In addition to the DNA hair sampling conducted in 1999 our research program conducted a co-operative effort with Sam Wasser to investigate the use of trained dogs to locate bear scat during the same sampling periods when the hair capture sessions were underway. The focus of this work was to investigate other DNA collection methods. Sam Wasser will present the results of this work later in this session.

Kenward, R. E., V. Marcström, and M. Karlbom. 1981. Goshawk winter ecology in Swedish pheasant habitats. Journal of Wildlife Management 45:397-408.

Stenhouse, G.B. and R. Munro. 2000. Foothills Model Forest Grizzly Bear Research Project 1999 annual report. Foothills Model Forest, Hinton, Alberta. 107 pp.

White, G. C., D. R. Andersen, K. P. Burnham, and D. L. Otis. 1982. Capture-recapture and removal methods for sampling closed populations. Los Alamos Nat. Laboratory LA-8787-NERP.

Woods, J. G., D. Paetkau, D. Lewis, B. N. McLellan, M. Proctor, and C. Strobeck. 1998. Genetic tagging free ranging black and brown bears. Wildl. Society Bull. 27:616-627.

Contact Information:
Garth Mowat, Fish and Wildlife Division, Timberland Consultants
Box 171, Nelson, B.C, V1L 5P9,
250-359-7606, gmowat@telus.net

Gordon Stenhouse, Alberta Environment, Hinton, AB

Curtis Strobeck, Dept. of Biological Sciences, University of Alberta
Edmonton, Alberta, T6G 2E9

Robin Munro and Kelly Stalker, Foothills Model Forest, Hinton, AB

Unbiased, cost effective collection methods are clearly needed for concurrently estimating the number, distribution and degree of disturbance of multiple species at risk over large remote areas. Our project aimed to validate and implement such methods, combining noninvasive fecal DNA and hormone technology with highly trained detection dogs used to locate scat from target species. Scat sampling with detection dogs has the potential to be relatively free of collection biases that have plagued many of the more traditional monitoring techniques because it utilizes high drive dogs, rewarded with play for locating samples independent of the target species' gender or behavior.

Four K-9 teams collected grizzly and black bear scat samples over a 5,350 km2 area of the Yellowhead ecosystem along the eastern slopes of Alberta Canada. Forty percent of the study area is within Jasper National Park, whereas the remaining 60% is in a multi use study area to the north, exposed to a variety of human disturbances. Scat samples were geo-referenced upon collection using a hand-held GPS unit and plotted on a Geographic Information System (GIS) that also maps disturbances in the study area. Hair sampling (G. Mowat and K. Stalker) and radio-collaring (G. Stenhouse and R. Munro) were conducted concurrently, and compared as part of the assessment of our scat methods.

Both hair snag and scat collection methods obtained comparable numbers of samples (~400 samples each), despite hair snags occurring in over 20% more of the study area (i.e., 64 versus 40 grids). Multiple scat samples collected in the same grid and session also appeared more likely to represent multiple individuals compared to hair snag collections. Visual comparisons suggested close correspondence between hair and fecal sample distributions of both black and grizzly bears. These distributions also appeared similar to telemetry based distributions of the radio collared grizzly bears from the study population. However, hair snags collected 0.47 black bear samples per grizzly bear sample whereas detection dogs collected 2.40 black bear samples per grizzly bear sample.

The stress data demonstrated gender and species differences, with stress levels being higher among grizzly versus black bears, and among males versus females within each species. We are still investigating the relationship between these stress levels and levels of human activities using GIS tools. While more data are needed, cross-sectional fecal cortisol measures across the landscape are beginning to suggest that stress levels for grizzly and black bears may correspond positively to Ursid densities and human disturbances.

The primary problem faced in this study was a lower microsatellite DNA amplification success rate than anticipated. Indeed, the ultimate success of this method hinges on enhancing microsatellite amplification success beyond that obtained in this study. We believe that these problems are surmountable and discuss ways to circumvent them in the future. Once resolved, coupling scat detection dog methods with DNA and endocrine techniques should offer an overall monitoring approach that is ideally suited to address a wide variety of Ursid conservation and management concerns.

Contact Information:
Samuel K. Wasser, University of Washington School of Medicine
Box 354693, Seattle, WA 98195,
206-534-0670, wassers@u.washington.edu

Gordon Stenhouse, Alberta Environment
Hinton, AB,
780-865-8331, gordon_stenhouse@gov.ab.ca

Contact Information:
Curtis Strobeck, University of Alberta
780-492-3515, curtis.strobeck@ualberta.ca

Contact Information:
David Roon, University of Idaho
208-885-5005, roon8505@uidaho.edu

Craighead, Paetkau, Reynolds, Vyse & Strobeck (1995) J. Heredity 86:255.

Gagneux, Boesch & Woodruff (1997) Molecular Ecology 6:861.

Paetkau, Waits, Clarkson et al. (1998) Conservation Biology 12:418.

Taberlet, Griffin, Goossens et al. (1996) Nucleic Acids Res. 24:3189.

Woods, Paetkau, Lewis, et al. (1999) Wildlife Society Bull. 27:616.

Contact Information:
David Paetkau, Molecular Artificer, Wildlife Genetics International
4100 EDC Building, 8308-114 St., Edmonton, AB T6G 2V2
780-491-6114 paetkau@telusplanet.net

Grizzly bear densities in BC have been classified into five categories ranging from 10 km2 bear to 1000 km2 per bear. Each BEC subzone / ecosection type in the province has been assigned a bear capability classification of 1 to 5. These classifications are used to calculate the population capability of an area to support grizzly bear. However, a wide variety of activities have changed the capacity of habitats to support grizzly bears. To take these into account four variables are used, each with three modification factors. The primary impacts are permanent habitat loss (i.e. cities), habitat alteration (i.e. logging), habitat degradation (i.e. roads) and population impacts (i.e. historical hunting). The first three of these primary impacts is rated according to the quality of the habitat lost, altered or degraded, the quantity of the habitat change, and the relative value of the habitat changed. The product of these ratings and the capability population estimate provides current population estimates.

Known population areas are used to calibrate the estimates of change. Once known areas match in the model we extrapolate to similar adjacent areas that do not have very much data available. This extrapolation provides us with MU population estimates, which are then summed to give Grizzly Bear Population Unit, regional and provincial population estimates. In the Kootenays the net result is a population estimate of between 2100 and 3000 grizzly bears at the present time.

Hunting management requires population estimates at the MU level, safe rates of harvest, estimates of non-hunting losses and unknown losses, and harvest data for resident hunters, non-resident hunters and first nations. Kootenay rates of harvest have historically been set at 0 or 3.8 percent annually, but are now more flexible with a range of 0 to 6 percent annually, depending on population goals and trends. Estimates of unknown losses and historical rates of known non-hunting losses are deducted from available harvest first. The remainder is allocated between residence groups, although in the Kootenay there is virtually no first nations harvest.

Resident hunters are allocated 70 percent of the regional harvest and the provincial Limited Entry Hunt (LEH) allocation system is use to provide hunting opportunities. Non-resident hunting opportunities are provided through a quota system, with licensed guide outfitters receiving 30 percent of the available harvest. Harvest balances are sought over a three-year period in an effort to smooth the flow of hunting opportunity and reduce the impact of year to year variations in total harvest. At the end of each three-year period every effort is made to match the harvest achieved with the harvest goal by adjusting hunting opportunities. Grizzly Bear Population Units are also use to better match the harvest with the area ranged by a grizzly bear population. Most Management Units are too small to encompass a GBPU. Kootenay Region total known grizzly kill has averaged 58 animals annually between 1976 and 1999, with an increasing trend peaking in 1996.

Contact Information:
Guy Woods, BC Environment
250-354-6341, guy.woods@gems4.gov.bc.ca

In many cases it may be appropriate to stipulate the most precautionary supposition to be correct (e.g. that an area is important grizzly bear habitat or that there is a high density of grizzly bears in the surrounding area). Those involved can then proceed to considering mitigation to determine how critical additional information (e.g. habitat inventory or a population inventory) is to the decisions being made. Regulators involved in assessing the potential impacts of a project should consider clearly establishing what they believe the general types of potential impacts that may result from the project are and what mitigation techniques might be used to address these impacts as early as possible in the process. Ideally such a strategic overview should allow for assumptions to be easily identified so that in the event that the proponent, or any other person or group, wishes to challenge one or more of the assumptions it will be readily apparent what information could be collected to confirm or refute each of them.

The focus of the environmental assessment process should be on the end result: the decision on whether a project will proceed including any mitigation measures to be implemented if it is approved. If decisions over data collection are considered in terms of how they will specifically contribute towards this end, the result will be a more focused process that provides better information for decision-making. This approach should also help to avoid expending time and effort on the collection of data that is not central to the decision-making process and that may be better focused on the development of effective mitigation measures.

Contact Information:
Matt Austin, BC Environment
250-387-9799, matt.austin@gems7.gov.bc.ca

[Continue to Session 2]

 Columbia Mountains Institute of Applied Ecology
Box 2568 Revelstoke, B.C. VOE 2SO
Tel: 250-837-9311 Fax: 250-837-9311
E-mail: office@cmiae.org
[Home Page]
  
© 1998 - 2008 Columbia Mountains Institute
 Website designed & maintained by Solutions Interactive - Contact Webmaster