A FOOD-BASED HABITAT-SELECTION M O D E L FOR GRIZZLY BEARS IN K L U A N E N A T I O N A L P A R K , Y U K O N by JAMES E D W A R D M c C O R M I C K B.Sc. (Zoology), The University of British Columbia, 1988 THESIS SUBMITTED IN PARTIAL F U L F I L M E N T OF THE REQUIREMENTS FOR THE DEGREE OF M A S T E R OF SCIENCE in THE F A C U L T Y OF G R A D U A T E STUDIES (Centre for Applied Conservation Biology) (Department of Forest Sciences) (Faculty of Forestry) We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH C O L U M B I A April 1999 © James Edward McCormick, 1999 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department The University of British Columbia Vancouver, Canada Date APRIL- a q . DE-6 (2/88) Abstract I examined the relationship between plant food abundance and diet, and habitat selection by grizzly bears (Ursus arctos) in the Alsek River Valley, Kluane National Park (KNP) in 1995 and 1996. I built a simple model that combined how much food was present in each bear habitat type (BHT) with how prevalent that food was in the diet of grizzly bears to produce a habitat food value (HFV) for each BHT. I tested the effectiveness of the model using habitat selection data from radio-collared grizzly bears. I designed this model to make a priori predictions of selection of BHTs by grizzly bears. The model combined the relative food abundances from each BHT with the respective seasonal food values to produce a H F V for each BHT. I calculated the relative abundance of 10 grizzly-bear plant foods within 8 BHTs from 478 food abundance plots. Diet was inferred from an analysis of scats collected in KNP. Four dietary seasons were distinguished based on shifts in plant foods eaten. I calculated a food value by dietary season for each plant food based on relative consumption of that food within that season. BHTs were ranked by H F V within each season and these ranks represented predicted habitat selection by grizzly bears. I tested the utility of this model by comparing actual habitat selection with the predictions of my model. Actual selection of BHTs by grizzly bears was measured from aerial locations (n=365) of radio-collared grizzly bears and then ranked within each dietary season. I compared the ranks of actual habitat selection (grizzly bear telemetry locations) to the ranks of predicted habitat selection (HFVs). HFVs were successful predictors of grizzly bear habitat selection. This simple food-based model may be used by Park managers to minimise human disturbance of grizzly bears in the Alsek valley by restricting human activity in areas of high grizzly bear food value. i i Table of Contents ABSTRACT . i i LIST OF TABLES : iv LIST OF FIGURES v ACKNOWLEDGEMENTS vi CHAPTER 1. INTRODUCTION CONTEXT 1 OBJECTIVES 2 STUDY AREA 3 Kluane National Park 3 Alsek River Valley 3 CHAPTER 2. BUILDING A FOOD-BASED HABITAT-SELECTION MODEL FOR GRIZZLY BEARS INTRODUCTION 6 METHODS 7 Habitat Classification 7 Seasonal Food Values 8 Food Abundance 12 Habitat Food Values 16 RESULTS 16 Seasonal Food Values 16 Food Abundance 18 Habitat Food Values 22 DISCUSSION 24 Habitat Classification 24 Seasonal Food Values 25 Food Abundance 27 Habitat Food Values 28 CHAPTER 3. TESTING THE FOOD-BASED HABITAT-SELECTION MODEL INTRODUCTION 30 METHODS 30 Habitat Selection 30 Measured Habitat Selection Compared to Predicted Habitat Food Values 32 RESULTS 33 Habitat Selection 33 Measured Habitat Selection Compared to Predicted Habitat Food Values 35 DISCUSSION 39 CHAPTER 4. SUMMARY AND MANAGEMENT CONSIDERATIONS 41 LITERATURE CITED 47 iii List of Tables Table 1. Bear habitat types (BHTs) identified for the Alsek River study area and the percentage of the study area they cover. 9 Table 2. Major plant foods found in the study area eaten by grizzly bears, as determined from scats collected from 1992-1994 in the montane zone, Kluane National Park (data summarised from McCann 1997). 10 Table 3. Food value (FV) of plant species determined by standardising product of frequency and volume measurements of <2 week old scats collected in the montane zone. Shading indicates high FVs for a given season. Sample size listed in parentheses, (data from McCann 1997) 19 Table 4. Measures of plant food abundance and availability from plots conducted in bear habitat types (BHTs). Mean canopy cover within each BHT is listed for each plant food. The mean proportion of berry producers and berries per 25-cm x 25-cm quadrat are listed for S. canadensis. Standard error of the mean is in parentheses. N/A = not applicable. N/S = not sampled. The rock/ice BHT was not sampled as I assumed it would have no food value. See Table 5 for BHT abbreviations. 20 Table 5. Relative abundances (RA) of food species in the eight different bear habitat types (BHTs) that I sampled. Shaded areas indicate high RAs. 21 Table 6. Habitat food values (HFVs) calculated for each bear habitat type (BHT) within each dietary season. The BHTs were ranked from 1 (high) to 9 (low) within each dietary season based on their HFV. When two or more BHTs had the same HFV, the average rank was given to each of the tied BHTs. The rock/ice BHT was assumed to have no food value. Shaded areas indicate the BHTs with the highest HFVs for that dietary season. 23 Table 7. Telemetry locations in the study area summarised by dietary season, by gender, and by number of individuals. 33 Table 8. Grizzly bear habitat selection calculated using Manly's alpha (Manly et al. 1972). The BHTs were ranked from 1 (high) to 9 (low) within each dietary season based on habitat selection. When two or more BHTs had the same selection value, the average rank was given to each of the tied BHTs. Shaded areas indicate the three BHTs with the highest selection values for that dietary season. Number of telemetry locations listed in parentheses. 34 Table 9. Correlations between rank of predicted habitat selection based on habitat food value (HFV) and rank of actual habitat selection based on radio-collared grizzly bear locations for all four dietary seasons (critical value: (r s)0 0 5 ( 1 ) 9 = 0.6) 37 Table 10. A summary of habitat food values (HFV) and habitat selection values for all four dietary seasons. 45 Table 11. Relative abundance of grizzly bear plant foods in the nine bear habitat types (BHTs). (***** highest abundance; **** 75-99% of highest abundance; *** 50-74% of highest abundance; ** 25-49% of highest abundance; * 1-25% of highest abundance; - not measured; blank = none observed) 46 iv List of Figures Figure 1. Map of north-western North America showing the location of Kluane National Park. 4 Figure 2. Annual dietary patterns of grizzly bears determined from predominance of food items found in scats collected in the montane zone in Kluane National Park between 1992 and 1994 (n=363). Area under the curve reflects the presence of food item in scats. Frequency and volume of food item residue in scats used to determine relative mean percent volume (see text). I used shifts in food consumption patterns to sort the biweekly periods into 4 dietary seasons. 17 Figure 3. Rank of predicted habitat selection based on habitat food values (HFVs) compared to rank of actual habitat selection based on locations of radio-collared grizzly bears. A l l 4 dietary seasons have been combined on this graph. Ranks range from 1 (high) to 9 (low) for each dietary season. 36 Figure 4. Changes in ranks across the 4 dietary seasons for 2 bear habitat types (BHTs). Rank of predicted habitat selection based on habitat food values (HFVs) compared to rank of actual habitat selection based on locations of grizzly bears: deciduous mixed forest (a) and shrub mosaic (b). Ranks range from 1 (high) to 9 (low) for each dietary season. 38 v Acknowledgements This study was funded and logistical support was provided by Kluane National Park and Reserve. This project is part of the larger Kluane National Park Grizzly Bear Research Project. Funding support was also provided by the Northern Studies Training Program. I received the Bert Hoffmeister Scholarship in Forest Wildlife and the Donald S. McPhee Fellowship. I am very grateful to my supervisor, Dr. Fred Bunnell, for his perspicacity and encouragement. My committee members, Dr. Bruce McLellan and Dr. Michael Pitt, provided insightful advice throughout the course of my project. Dr. David Shackleton suggested improvements for the final draft of my thesis. Drs. Valerie LeMay, Antal Kozak, and Peter Marshall provided invaluable statistical advice. Rob McCann helped conceive this project and provided much insight and direction. He generously provided data he collected and analysed for the Kluane National Park Grizzly Bear Research Project that allowed me to both refine and test my model. This project would not have happened without the excellent support and effort from Park Wardens Ray Breneman and Kevin McLaughlin. Sid Waskiewich went above and beyond expectations as a volunteer field assistant for two summers. His dedication and friendship made fieldwork an especially rewarding experience. For their assistance and for the great working environment that they supplied, I thank the Park Wardens of Kluane National Park: Terry Skjonsberg, Rhonda Markell, Bruce Sunbo, Glen Kubian, Tom Buzzell, Mark Eikland, Craig McKinnon, Lloyd Freese, Duane West, Phillip Frost, Andrew Lawrence, and Rick Staley. Henry Hudson used his ingenuity to construct sampling equipment. Jesse Devost prepared excellent GIS maps for ground-truthing habitat types. Mark Lindberg shared strategies in the early stages of the project. Grant MacHutchon offered his insights on bears and habitat. Debbie Wellwood, Siobhan Jackson, and I combined resources and energies vi for a summer of fieldwork. Doug Makkonen provided safe and efficient helicopter piloting under a variety of conditions. Special thanks to Ray and Louise Breneman whose friendship and generous hospitality helped make my stay in Haines Junction so enjoyable. Scott Harrison generously contributed his time, energy, insight, and friendship from the conception of this project to the final write-up. Bob Mooney and Rachel Holt shared their wisdom and camaraderie while teaching me the finer aspects of pool. Nyree Sharp was always there to supply encouragement and a good joke. The "bear guys", Rob Serrouya and Roger Ramcharita, provided stimulating discussion on the ecology of bears. Tanya Wahbe and Kim Lisgo provided perspective and support. Lara Payne contributed killer graphics and nourishment for the soul. Pierre Vernier supplied advice on Geographical Information System technology. Jackie Johnson, a keystone member of the Centre for Applied Conservation Biology, was always generous with her assistance. I thank my friends at the Centre for their encouragement and support throughout this project. I thank my parents for their lifetime of supporting and encouraging me in the pursuit of my dreams. vii Chapter 1 Introduction Context Kluane National Park and Reserve (KNP) in the Yukon Territory supports a large population of northern interior grizzly bears (Ursus arctos). The KNP Grizzly Bear Research Project was initiated in 1992 within KNP to study the ecology of grizzly bears prior to proposed improvements in visitor access to the Park (Wielgus et al. 1992). One objective of the study was to 'identify, classify, and map grizzly bear habitats so that human use and development can be restricted to non-sensitive and non-dangerous areas' (Wielgus et al. 1992). The Alsek River Valley is an undeveloped wilderness area within KNP and, with the Kaskawulsh River Valley, is believed to support the highest grizzly bear densities in the Park (Canadian Parks Service 1990). By 1987, the Alsek River had become a destination for commercial and private W h i t e w a t e r rafting trips. In 1987, one group of 11 people went down the river, but in 1994, at least 300 people in 29 different groups made the trip (KNP unpublished data). The rapid rise in popularity of rafting in this area prompted concern about the potential impacts of human activity, particularly camping, on grizzly bears. Human-bear interactions are an important issue in parks and protected areas that support bear populations (Albert and Bowyer 1991; Gunther 1990; Jope 1985; Mace and Waller 1996; McArthur-Jope 1982; McCrory et al. 1986). The backcountry hiking area in the Slims River Valley, KNP, experienced an increase in serious bear incidents from 1981 - 1987 (Leonard et al. 1990). Incidents included defensive reactions by grizzlies to surprise encounters with people and aggressive actions by grizzlies to obtain food from people. No human injuries were reported but 1 Parks staff were required to kill five grizzly bears, relocate five grizzly bears, and close areas for up to 30 days. People and bears must be kept separate to reduce the potential of bear-human interactions. A realistic approach is to understand where grizzly bears are more likely to be and keep people away from those habitats. One method is to construct a model of expected habitat selection based on our knowledge of grizzly bear ecology. Many factors can affect grizzly bear habitat selection including resource requirements, learned behaviour, reproductive strategies, exploration of home ranges, use of cover, slope, aspect, or resource partitioning between sex and age classes (Boudreau 1995; Hamer and Herrero 1983). A simple viable approach is to build a model based on diet and food availability. Many studies have suggested that grizzly bear habitat selection is related to the quality and quantity of available forage (Craighead et al. 1982; Hadden et al. 1985; Hamer and Herrero 1983; Hamilton 1987; Lloyd 1979). By understanding what grizzly bears are eating and having a measure of how much of each food is present in different habitats, insights can be gained into which habitats grizzlies are more likely to frequent. Objectives The objective of my study is to develop and test a food-based habitat-selection model for grizzly bears. This tool will assist Park staff in minimising the impacts of humans on grizzly bears in the Alsek Valley by identifying habitat types with abundant grizzly bear food, that are most likely to receive high use by bears. In Chapter 2 I estimate a seasonal value for each habitat type from food values and food abundances. In Chapter 3 I test the seasonal habitat values against seasonal habitat selection by 2 grizzly bears. In Chapter 41 summarise my results and discuss their implications to the management of visitor activities in the Alsek Valley. Study Area Kluane National Park The 22,015 km 2 Kluane National Park and Reserve is located in the south-west corner of the Yukon Territory (Fig. 1). Together with Tatshenshini-Alsek Wilderness Park, British Colombia, and Wrangell-St. Elias National Park and Reserve and Glacier Bay National Park, Alaska, it makes up the largest internationally protected area in the world (97,000 km2), a UNESCO World Heritage Site. Only 18% of the total Park area is vegetated; the St. Elias Mountains, which include Canada's highest peak and the world's largest nonpolar icefields, comprise most of the Park (Gray 1987). The lower valleys and slopes make up the montane zone. The montane zone is dominated by white spruce (Picea glauca) forests which are interspersed with shrub, herb, marsh, and fen vegetation communities (Douglas 1974). The subalpine zone occurs above the continuous forest. It is dominated by tall shrubs, mainly willow species (Douglas 1974). The lower alpine zone is characterised by a low shrub mosaic and the upper alpine zone is distinguished by dwarfed vascular plants (Douglas 1974). Alsek River Valley The study area extends from Serpentine Creek on the Dezadeash River to Goatherd Mountain on the Alsek River, a length of approximately 57 km. It is comprised of the montane vegetation zone (Douglas 1974) to a maximum distance of 1500 m from the edge of the active 3 Figure 1 . Map of north-western North America showing the location of Kluane National Park. 4 riverbed on each side of the valley. This arbitrary distance was chosen to depict the area of potential influence of rafting campsites on grizzly bears. Rafting groups often put in at Serpentine Creek and raft down past Goatherd Mountain and the Lowell Glacier. The Dezadeash River enters the Park through a gap in the Kluane Mountain Range. It joins the Kaskawulsh River and together they form the south-flowing Alsek River. The massive Lowell Glacier resides across from Goatherd Mountain and this dynamic lump of ice, snow, and debris has a climatic influence on the valley. The Alsek River flows down a broad, U-shaped, glacial valley that is intersected with smaller side canyons and streams. Most of the rivers in the area are glacial fed and highly silt-laden. Summertime flows often are heavy from glacial melt and the river beds are dynamic and braided. 5 Chapter 2 Building a Food-Based Habitat-Selection Model for Grizzly Bears Introduction It is critical that grizzly bears consume enough food during the non-denning period to meet the energy demands of reproduction and growth, and to develop fat reserves for hibernation. Grizzly bears in KNP can spend between 162 and 235 days of the year hibernating (McCann 1996). During denning, grizzly bears lose a large percentage of their autumn weight. Females can lose up to 40% of their fall weight, while males lose about 22% of their fall weight (Kingsley et al. 1983). Two adult male grizzly bears in a simulated denning experiment lost an average of 0.44 kg/day (Watts and Jonkel 1988). Body weight may decrease further after emergence in spring and it is not until mid-summer that weight is regained. Bears are omnivores, and that seems to enable rapid weight gains during the pre-denning period, but the digestive system of bears is not adapted to digest cellulose (Bunnell and Hamilton 1983; Rogers 1976). Bears have made some morphological adaptations that aid with herbivory but retain the relatively short digestive system of a carnivore (Bunnell and Hamilton 1983; Herrero 1978). The digestion of fibrous forage is inefficient (Bunnell and Hamilton 1983) and passage rates of plant material are faster than passage rates of meat.(Pritchard and Robbins 1990). Therefore, bears need to consume easily digestible forage in abundant quantities to enable weight gain prior to denning. The importance of finding high quality food in a short period is supported by studies of grizzly bear activity and habitat use. Activity budgets for grizzly bears show that the majority of time is spent feeding (MacHutchon 1996; Stelmock and Dean 1986). The spatial distribution of major food items is thought to influence grizzly bear habitat selection (Mace 1986). Bears 6 appeared to select more productive habitat (Mattson et al. 1987) and movements of grizzly bears in Yellowstone increased with increasing food ingestion (Blanchard and Knight 1991). Studies of bears and habitat use have shown that distribution of food influences where bears are located (Clark et al. 1994; Hamilton 1987; Kansas and Riddell 1993). In this chapter, I outline how I constructed a simple model to predict habitat selection by grizzly bears. By quantifying diet and plant food abundance, I built a model that reflects the food value of different habitats to grizzly bears. A model enables habitat evaluations that are well documented, repeatable, and quantifiable (Stelfox 1991). Models are useful because they reveal implicit assumptions. They thus provide an opportunity to test these assumptions and to direct further questions. While building the model, one can identify deficiencies in the data and one's understanding of the system being modelled (Stelfox 1991). Several assumptions are made in my model; specifically, an analysis of the frequency and volume of food items in grizzly bear scats reflects the diet of grizzly bears and the cover and availability estimates of plant foods reflect the relative abundance of those plant foods. The objectives of this chapter are to: 1) define and map habitat types based on broad vegetation types; 2) determine a food value for plant food species based on scat analyses; 3) estimate the relative abundance of plant foods in each habitat type; and 4) combine food values and abundance values to estimate habitat food values. Methods Habitat Classification I based my bear habitat type (BHT) map on the existing KNP Geographic Information Systems (GIS) vegetation map that was digitised from polygons drawn on a 1:50,000 airphoto-7 mosaic. There were too many vegetation communities described for the area (Douglas 1974; Douglas 1980) to allow adequate collection of food abundance data for each community or to provide enough grizzly bear locations for an adequate summary of habitat use. Instead, I classified the study area into nine different BHTs that were similar in vegetation structure, geomorphology, or proximity to the Alsek River (Table 1). A limited number of habitats is recommended for tests of habitat selection (Alldredge and Ratti 1986). I classified the vegetation polygons on the GIS map into one of the nine BHTs using 1:25,000 black and white air photos, 10-cm x 15-cm oblique colour photos taken from a helicopter, and ground truthing. Problems, including georeferencing, edge matching of mosaics, discontinuous polygons across mosaics, and inaccurate labelling, were encountered within the KNP GIS vegetation map (McCann 1997). Where I observed large discrepancies between the actual BHT polygon boundaries and the GIS map, I adjusted polygon boundaries on the GIS map. I estimated the correct position of the polygon boundaries on the GIS map using both the air photos and the oblique colour photos. Seasonal Food Values I determined the seasonal value of food (FV) for grizzly bears in KNP using scats collected and analysed as part of the KNP Grizzly Bear Research Project. Methods used to collect scats are presented by McCann (1997). R. McCann (unpublished data) supplied content and volume data from all scats less than two-weeks-old collected in the montane zone (n=363). He summarised the data in biweekly intervals from May 1s t to October 15 t h. From this scat analysis, I identified common plant foods that also occurred in my more restricted study area (Table 2). I defined different dietary seasons for non-denning grizzly bears in KNP by identifying dietary shifts in the scat analysis summarised in biweekly intervals. 8 Table 1. Bear habitat types (BHTs) identified for the Alsek River study area and the percentage of the study area they cover. Bear Habitat Type (BHT) Percentage within Description Study Area Spruce forest (FS) 37.3 >25% canopy cover of Picea glauca, <25% canopy cover of deciduous; slopes of 3% -32%; overstory cover usually greater than 30% Shrub mosaic (SM) 17.3 a broadly defined category with a relatively open canopy cover (range from 0 - 25%); frequently on a slope (average 28%, range 6 - 52%) Alder shrub (SA) 11.0 >75% canopy cover of tall shrub, primarily Alnus spp.; often on a slope of >25%; dense overstory cover >40% Dryas herb (HD) 10.2 Dryas drummondii is the dominant plant and very little other vegetation is usually present; flat ground, usually no overstory cover Herb (HE) 9.2 dominated by herb communities (other than Dryas drummondii); includes bog, fen, and seepage areas; flat river bars to open hillsides; sparse i f any overstory cover Rock/Ice (RI) 6.5 <5 % vegetation; includes rock, gravel, and ice Deciduous/mixed forest (FM) 3.6 >25% canopy cover of deciduous; overstory cover greater than 30% Floodplain shrub (SF) 3.4 only in flat floodplain areas near and influenced by the Alsek River; appears to only be in areas <2 m above water level of river; vegetation ranged from open herb to shrub to deciduous and coniferous tree; low to high overstory cover Grey willow shrub (SW) 1.5 >50 % canopy cover of Salix glauca, <25% canopy cover of spruce Table 2. Major plant foods found in the study area eaten by grizzly bears, as determined from scats collected from 1992-1994 in the montane zone, Kluane National Park (data summarised from McCann 1997). Code Plant Species Common Name Part Eaten A R U U Arctostaphylos uva-ursi bearberry berries ELCO Elaeagnus communtata silverberry berries EMNI Empetrum nigrum crowberry berries SHCA Shepherdia canadensis soapberry berries EQAR Equisetum arvense horsetail foliage C A R E X Car ex spp. sedges foliage GRASS Grass spp. grasses foliage H E A L Hedysarum alpinum alpine hedysarum roots O X C A Oxytropis campestris field locoweed flowers ASTR Astragalus spp. milk vetch roots, leaves 10 I calculated a relative value for each food within each dietary season based on the percent of the diet made up by that food. Diet was determined from the analysis of scats. I started with a measure from the scat analysis that incorporated both the number of times a dietary item appeared in the scats and the volume of the food in the scat: relative percent volume (R%V) (McCann 1997). I have summarised the method used to calculate the R % V from McCann (1997): R % V = (RF)(M% V) (Equation 1) (will sum to 100 for each dietary season) where ^ , . „ number of scats having the food item RF = relative frequency = total number of scats M % V = average percent volume of the food item in all the scats in which it was found Many of the scats analysed contained more than one food item. I calculated R % V for each of the main grizzly bear plant foods found in my study area. I standardised all R%V's among dietary seasons by dividing each R % V by the maximum R % V for that season. This gave me a range of plant food values (FV) from 0 - 1 for each dietary season. F V , therefore, is the estimated measure of relative importance of a food species based on how common it is in the diet. Some results from the scat analyses specified only genus or a choice between two genera. Based on my experience and the experiences of the KNP project researchers, I interpreted the data as follows: 11 1) A l l Equisetum scats were classified as E. arvense. Studies in other areas have noted that this is the species commonly eaten (Hamer and Herrero 1987; Mattson et al. 1991; McLellan and Hovey 1995). 2) A l l Hedysarum scats were classified as H. alpinum. Observations in the field show exclusive digging ofH. alpinum. 3) A l l Car ex and grass scats were classified as graminoids because of the difficulty and inconsistency in identification within scats. 4) A l l Elaeagnus scats were classified as E. communata. This is the only Elaeagnus species in the study area. 5) A l l Oxytropis scats were classified as O. campestris. This is the most common species of Oxytropis found in the area and the species that grizzly bears were observed to be feeding on in the montane zone. I made the following assumptions and decisions about the scat analysis data used to determine seasonal FVs. I assumed that all scats collected throughout the montane zone in KNP would be indicative of scats deposited by grizzly bears within the smaller study area. Many of the scats were collected within or at the edge of the study area. For the food value calculations, I included all the plant food species that composed greater than 5% of the diet in one or more dietary seasons. The one exception was Salix spp. catkins. These were found in some scats deposited in the montane zone but observations supported my assumption that grizzly bears were feeding on Salix spp. predominantly in the sub-alpine zone. Food Abundance Within BHT types, I measured the relative abundance of plants and plant parts that were known grizzly bear foods. Bear food sampling was conducted over two summers: June 5 -12 August 23, 1995, and July 2 -21 , 1996. Data collection did not begin until Shepherdia canadensis berries had started to ripen and finished before the majority of berries started to drop. I pooled data from 1995 and 1996 to increase sample size. I believed that the distribution of plant foods would be similar among years because most of the plant foods were rooted perennials, there was similar moisture levels in the two years, and successional changes occur very slowly in the Alsek Valley. I located plots from a random starting point at 50 m intervals along a random compass bearing. Systematic plot sampling along a randomly selected line is acceptable as long as one looks for possible periodic trends in spatial pattern (Krebs 1989). I sampled ten plots within each BHT polygon; fewer plots were sampled in polygons too small to allow a 500-m transect. Plots consisted of a circle of radius 7.5 m. I measured bear plant foods and other variables within each of these plots using the following techniques. Within the plot, I estimated the percent canopy cover of each bear food species (Daubenmire 1968). For//, alpinum,E. arvense, O. campestris, andS. canadensis, the four main food species, I visually estimated the canopy cover to the nearest whole number. For the remainder of the plant food species, I recorded the canopy cover following Daubenmire (1968) cover classes (0, <1%, 1-5%, 5-25%, 25-50%, 50-75%, 75-95%, 95-100%). Daubenmire cover classes in a 15-m diameter circular plot provide an adequate estimate of stand composition (Douglas 1974). To calculate the average percent canopy cover of the species, I used the midpoints of the cover classes (Douglas 1974). I assumed that canopy cover was proportional to food abundance for most of these species. I tested this assumption for O. campestris and H. alpinum. 13 I used a 1-m x 1-m quadrat to compare canopy cover estimates to food available for O. campestris. I estimated the O. campestris canopy cover within the quadrat and then clipped and counted all the flower heads. I calculated the regression of estimated canopy cover to number of flower heads and found the relationship to be significant (r2 = 0.81, p « 0 . 0 0 5 , n = 12). This suggests that canopy cover of O. campestris is a sufficient indicator of number of flowerheads available [total flower heads = -1.3 + 5.02 canopy cover]. I compared root biomass to above ground biomass for H. alpinum. I collected samples for six weeks in the summer of 1994. I selected different locations in a range of habitat types, slopes, aspects, and moisture levels. I also sampled a range of above ground biomass measurements. At sample points, I clipped all the H. alpinum plants that originated within a 0.5-m x 0.5-m quadrat and removed all the H. alpinum roots up to 15 cm below the ground within the quadrat. I oven dried the samples prior to weighing. I removed two of my observations from the data: the first had been collected in an unusual location for H. alpinum to be growing; the second was twice as large as the next largest observation. I did a curvilinear regression and found a significant result (r2 = 0.62., p « 0 . 0 0 5 , n=28) indicating that foliage biomass is an adequate indicator of below ground biomass fori/ , alpinum [root biomass = 18.017 + (1.205) foliage biomass + (-20.180) 1/foliage biomass]. My findings are similar to those of Vandehey (1991) who determined an r 2 of 0.68 for the same variables ofH. sulphurescens. Although the data were variable, I assumed a positive correlation between canopy cover and above ground biomass because of the difficulty in measuring or estimating above ground biomass. Percent canopy cover is not the only important variable for S. canadensis berries. In important S. canadensis feeding sites, canopy cover varies from dense to scattered (>75% canopy cover to <5% canopy cover; Hamer and Herrero 1987). More consistent, however, is the 14 observation that "bears feed on shrubs bearing abundant fruits" (Hamer 1996). S. canadensis is also a dioecious species so only female plants bear fruit. Therefore, in addition to canopy cover estimates, I also measured the proportion of shrubs bearing fruit and estimated the number of berries per m 2 of fruit-bearing shrub. Berry density was estimated using a double sampling scheme. Pilot studies I conducted in 1993 and 1994 suggested visual estimates of berry numbers within a 25-cm x 25-cm quadrat were good predictors of actual berry counts (r2=0.90 and 0.93, respectively, p«0 .005) . I continued double sampling randomly in 1995 and 1996. I sampled 4 shrubs per plot in 1995 and 16 shrubs per plot in 1996, to obtain a better estimate of berry abundance. I used the resulting regressions to correct berry estimates: 1995: berry count = 1.27 x berry estimate, r 2 = 0.93, p « 0 . 0 0 5 , n = 47; 1996: berry count = 1.23 x berry estimate, r 2 = 0.92, p « 0 . 0 0 5 , n = 69; The importance of Elaeagnus communata and Astragalus spp. as bear foods in KNP was not realised before the 1995 sampling season. The abundance of these foods was only measured in BHT polygons sampled in 1996. I estimated the abundance of plant foods from average canopy cover values. I calculated the mean canopy cover for each food species in each BHT. I also averaged berry production and berry estimates for S. canadensis berries. Averages were determined from all plots sampled in the BHT. Relative abundances (RAs) are an estimate of how much food is present in each BHT relative to all the other BHTs. I calculated them by standardising the abundance averages for each plant food for each BHT. To do this, I divided all the averages by the maximum BHT 15 average for each food species. This produced RAs for each plant food species ranging from zero to one. Habitat Food Values Habitat food values (HFVs) are an estimate of how much food is present in each BHT and how prevalent that food is in the diet of grizzly bears. I calculated seasonal HFVs by multiplying the seasonal FVs by their respective RAs for each BHT and summing the products (Equation 2). H F V j k = i t i (FV* x MV) (Equation 2) *=i j=\ /=i where F V = seasonal food value R A = relative abundance / = bear plant foods (n=9) j = bear habitat types (BHTs) (n=9) k = dietary seasons (n=4) .HFVs were ranked from 1 - 9 for each dietary season. I assumed the rock/ice BHT had no food present and therefore an H F V of zero. When two or more BHTs had the same HFV, the average rank was given to each of the tied BHTs. Results Seasonal Food Values I identified four dietary seasons from the scat analyses (Fig. 2). Spring diets (May 1 -June 14) were dominated by E. arvense, Arctostaphylos uva-ursi berries, and H. alpinum roots; 16 3 < o E £ 3 3 a. 03 > _, c g o .S C C 1 ) B o _ t o "4 * U PP O pt) 00 •g 5 O -cj <*3 jxi CJ o d 5 CJ u ° 12 N C o -c s o e 3 CJ J 3 [i , " C O B, 5 3 •o g S 8 CJ .13 3 o u o S o a> w .S -au ° t2 w o -a 3 3 a; 3 o •a 5 <8 a 3 y p—i XI o . C J O G 1 ) 3 a 3 a I * 5-1 o 3 CJ •a § CJ ON f i S) CJ i i o > 3 u I—H cj a< 3 CJ E g CJ CJ s & CS U •3 2 _ C3 S d 3 0 CJ 2 § cn 15 w w o o 3 •5 E •£ 1 ) CJ 17 early summer foods (June 15 - July 14) consisted predominantly of E. arvense and O. campestris flowers; S. canadensis berries and H alpinum roots were the main diet in late summer (July 15 -August 31); and autumn foods (September 1 - October 14) consisted mainly of H alpinum roots and various species of berries. Four food plants species were dominant diet items (Table 3). Grizzly bears ate E. arvense in May, June, and July, and this plant had the highest F V for both the spring and early summer seasons. Grizzlies dug H. alpinum roots in all seasons of bear activity, but most frequently in September and May; these roots had the highest F V in autumn and second highest in late summer. Grizzly bears ate O. campestris flowers primarily in June, with reduced consumption in July and August; these flowers were the second most important food in early summer. Shepherdia canadensis berries were consumed by grizzlies from the latter half of July to mid September; the berries had the highest F V in the late summer season. Food Abundance I conducted a total of 57 bear food transects, comprised of 478 plots, over the two summers. I found considerable variation in plant food abundance both within and among BHTs. The results are summarised in Table 4 and the standardised RAs in Table 5. Hedysarum alpinum occurred predominantly in the floodplain shrub BHT. Equisetum arvense was found predominantly in moist sites within grey willow shrub and the two forest BHTs. Oxytropis campestris occurred in open patches within floodplain shrub, herb, and Dryas herb BHTs. Shrub mosaic was the best BHT for S. canadensis berries followed by grey willow shrub. Deciduous/mixed forest had the highest canopy cover of S. canadensis but a low proportion of berry producing plants and low berry production on the plants that did produce. Other BHTs 18 Table 3. Food value (FV) of plant species determined by standardising product of frequency and volume measurements of <2 week old scats collected in the montane zone. Shading indicates high FVs for a given season. Sample size listed in parentheses, (data from McCann 1997) Spring Early Summer Late Summer Autumn Dietary Item (n=45) (n=105) (n=157) (n=56) Arctostaphylos uva-ursi 0.364 0.045 0.015 0.027 Astragalus spp. 0.097 0.054 0.021 0.000 Elaeagnus communata 0.000 0.000 0.000 0.191 Empetrum nigrum 0.000 0.001 0.073 0.242 Equisetum arvense. 1.000 1.000 0.224 0.025 Graminoids 0.127 0.080 0.061 0.016 Hedysarum alpinum 0.349 0.078 0.533 1.000 Oxytropis campestris 0.216 0.825 0.054 0.006 Shepherdia canadensis 0.009 0.039 1.000 0.388 19 Table 4. Measures o f plant food abundance and availability f rom plots conducted in bear habitat types (BHTs). Mean canopy cover within each B H T is listed for each plant food. The mean proportion o f berry producers and berries per 25-cm x 25-cm quadrat are listed for S. canadensis. Standard error o f the mean is in parentheses. N/A = not applicable. N/S = not sampled. The rock/ice B H T was not sampled as I assumed it would have no food value. See Table 5 for B H T abbreviations. Bear Habitat Type (BHT) FS S M SA H D HE F M SF s w no. transects 13 13 3 5 14 3 4 2 no. plots 98 81 28 50 135 30 36 20 Arctostaphylos uva-ursi 2.1 5.3 0.6 0.1 2.2 14.3 3.9 6.8 (0.6) (0.9) (0.5) (0.1) (0.6) (4.2) (1.3) (1.6) Astragalus spp. 0.8 0.0 N/S 0.0 0.0 4.1 0.0 0.9 (0.5) (0.0) (0.0) (0.0) (1.9) (0.0) (0.4) Car ex spp. 0.6 0.8 1.1 0.0 7.5 0.1 0.0 2.0 (0.3) (0-4) (0.7) (0.0) (1.7) (0.1) (0.0) (1.9) Elaeagnus communtata 0.0 0.0 N/S 0.0 0.0 0.0 1.5 0.0 (0.0) (0.0) (0.0) (0.0) (0.0) (1.5) (0.0) Empetrum nigrum 0.9 0.0 0.0 0.0 0.0 0.0 0.0 0.2 (0.3) (0.0) (0.0) (0.0) (0.0) (0.0) (0.0) (0.2) Equisetum arvense 6.0 1.9 0.8 0.0 0.3 8.6 2.6 11.3 (1.4) (0.9) (0.3) (0.0) (0.1) (2.5) (1.5) (4.3) Grass 2.0 2.4 5.5 5.2 13.7 3.0 2.2 5.8 (0.5) (0.5) (1.3) (1.2) (1.4) (1.8) (0.8) (1.6) Hedysarum alpinum 0.4 0.4 0.0 0.0 0.4 0.0 17.2 0.2 (0.3) (0.2) (0.0) (0.0) (0.3) ( 0 0 ) (3.1) (0.2) Oxytropis campestris 0.0 0.8 0.0 1.7 3.1 0.7 3.2 0.9 (0.0) (0.2) (0.0) (0.3) (0.3) (0-3) (0.6) (0-6) Shepherdia canadensis 10.7 23.5 1.3 3.0 7.0 26.9 4.8 13.2 (1.4) (2.0) (0.9) (0-9) (1.2) (4.9) (1.7) (3.1) proportion o f shrubs 0.26 0.41 0.29 0.44 0.38 0.21 0.32 0.48 wi th berries (0.03) (0.02) (0.04) (0.06) (0.04) (0.05) (0.07) (0.06) number o f berries per 21.0 32.1 10.2 32.4 31.3 9.9 28.0 33.1 25-cm x 25-cm quadrat (3.0) (3.5) (3.8) (8.2) (4.7) (2.0) (10.3) (4.9) 20 Table 5. Relative abundances (RA) of food species in the eight different bear habitat types (BHTs) that I sampled. Shaded areas indicate high RAs. BHT A R U U ASTR C A R E X ELCO EMNI EQAR GRASS HEAL OXCA SHCA X3 FS 0.15 0.20 0.07 0.00 1.00 0.54 0.15 0.02 0.00 0.19 S M 0.37 0.00 0.11 0.00 0.00 0.17 0.17 0.02 0.26 1.00 SA 0.04 - 0.14 - 0.02 0.07 0.40 0.00 0.00 0.01 HD 0.00 0.00 0.00 0.00 0.00 0.00 0.38 0.00 0.52 0.14 HE 0.15 0.00 1.00 0.00 0.00 0.02 1.00 0.03 0.98 0.27 F M 1.00 1.00 0.01 0.00 0.00 0.76 0.22 0.00 0.22 0.18 SF 0.27 0.00 0.00 1.00 0.00 0.23 0.16 1.00 1.00 0.14 SW 0.47 0.22 0.27 0.00 0.17 1.00 0.42 0.01 0.27 0.67 FS = Spruce forest SM = Shrub mosaic SA = Alder shrub HD = Dryas herb HE = Herb RI = Rock/Ice FM = Deciduous/mixed forest SF = Floodplain shrub SW = Grey willow shrub SHCA X 3 = S. canadensis including proportion of shrubs producing and berries per quadrat 21 with high overstory cover (alder shrub and spruce forest) also had low numbers of berries on producing bushes. Arctostaphylos uva-ursi was most common in the deciduous/mixed forest BHT with abundant pockets also in the grey willow shrub and the hillsides of shrub mosaic BHTs. Empetrum nigrum and E. communtata were not common in the study area. Habitat Food Values HFVs for each of the four dietary seasons are presented in Table 6. Each H F V was ranked within the dietary season; tied values were given an average of the ranks over which they ranged. Floodplain shrub and grey willow shrub were consistently in the top three HFVs for each dietary season. By contrast, alder shrub and Dryas herb were consistently low food value BHTs. Deciduous/mixed forest had a high H F V in the spring and early summer but decreased in late summer and autumn. Shrub mosaic had the opposite trend, starting with a low H F V in the first two dietary seasons but a higher HFV for the last two dietary seasons based on a high density of S. canadensis berries. 22 Table 6. Habitat food values (HFVs) calculated for each bear habitat type (BHT) within each dietary season. The BHTs were ranked from 1 (high) to 9 (low) within each dietary season based on their HFV. When two or more BHTs had the same HFV, the average rank was given to each of the tied BHTs. The rock/ice BHT was assumed to have no food value. Shaded areas indicate the BHTs with the highest HFVs for that dietary season. Spring Early Summer Late Summer Autumn BHT H F V rank H F V rank H F V rank H F V rank Spruce forest (FS) 0.6 4 0.6 5 0.4 5 0.4 3 Shrub mosaic (SM) 0.4 5.5 0.4 6.5 1.1 1 0.4 3 Alder shrub (SA) 0.1 7.5 0.1 8 0.0 8 0.0 8 Dryas herb (HD) 0.1 7.5 0.4 6.5 0.2 7 0.1 6.5 Herb (HE) 0.4 5.5 0.9 4 0.4 5 0.2 5 Rock/Ice (RI) (not sampled) 0 9 0 9 0 9 0 9 Deciduous/mixed forest (FM) 1.3 1.5 1.1 3 0.4 5 0.1 6.5 Floodplain shrub (SF) 0.9 3 1.2 2 0.8 3 1.3 1 Grey willow shrub (SW) 1.3 1.5 1.3 1 1.0 2 0.4 3 23 Discussion I used food characteristics of different habitat types to rank their importance to grizzly bears. Food is a suitable variable to use because of its obvious importance to grizzly bears and it can be quantified. I combined diet derived from scat analysis and food abundances assessed from plots to estimate a habitat food value for grizzly bears. This model can help Park managers rank habitat value and possibly predict grizzly bear habitat use. Habitat Classification There is no single, natural unit of vegetation classification (Whittaker 1978). I classified the study area into BHTs based on the amount of food that bears might find in each type. I used predominant species and the vegetation layer (tree, shrub, or herb) as the distinguishing features for most BHTs. The floodplain shrub BHT was one exception to this classification scheme because its dominant layer varied from herb to shrub to tree, but its proximity to the Dezadeash and Alsek rivers seemed to provide good growing conditions for H. alpinum. The shrub mosaic BHT was the designated category for many of the vegetation communities that were too finely-scaled to fit anywhere else. I considered other factors when defining BHTs. I classified spruce forest as one BHT despite the large proportion of the study area covered by spruce forest (>37%) and the high variability in the plant communities beneath the canopy. Although my approach follows other habitat classification systems for bears that emphasise non-forested vegetation (Despain 1986), the primary reason for not further stratifying spruce forest was that most of this BHT was mapped into only six large polygons. Remapping these areas at a finer scale was impossible within the context of this project. The grey willow shrub BHT made up the smallest proportion 24 of the study area (1.5%). It was included because of its suspected high bear use. The Dryas herb BHT was included as a distinct BHT, as opposed to being grouped within the herb BHT, because I suspected it would have lower amount of bear plant foods and it covered a significant proportion of the study area (10.2%). Seasonal Food Values Like all animals, bears forage to acquire energy for growth and reproduction. However, researchers do not yet understand the effects of micronutrients, toxins, handling time, and learned behaviour on bear food habits (Hamer and Herrero 1983; Robbins 1993) and cannot predict an optimal diet for bears. I inferred features of grizzly bear diets and foraging seasons from the collection and analyses of scats. That approach is subject to four types of error (McLellan and Hovey 1995). In brief, there is unequal probability of collecting each scat, black bear scats may be confused with grizzly bear scats, each scat is given equal weight in the analysis although individual scat volumes vary, and lastly, the proportion of different foods eaten is not necessarily reflected in the proportion of faecal residue (McLellan and Hovey 1995). To address the latter problem, correction factors have been calculated that adjust the volume of food remnants in a scat to reflect how much of that food was eaten (Hewitt and Robbins 1996). Such correction factors had not yet been incorporated into the summarised scat analysis data that I used to calculate food values (FVs). The uncorrected scat analysis data I used may have over-estimated the proportion of E. arvense, O. campestris, H. alpinum, Astragalus, and graminoids, slightly over-estimated the proportion of A. uva-ursi, E. nigrum, and E. communtata, and under-estimated the proportion of S. canadensis and any meat in the diet (Hewitt and Robbins 1996). Meat was not considered in my model. There were two main faunal species that potentially could have been eaten: moose (Alces alces) and ground squirrels (Spermophilus spp.). Moose are not 25 common in the study area and grizzly bears appear to dig for ground squirrels predominantly in alpine areas. Review of the seasonal rankings of BHTs suggest that corrections for variable digestibility would have changed rankings little. I examined diet studies conducted on other populations of grizzly bears, and the literature for nutritional studies of the food plants in question. I limit my discussion of grizzly bear diets in KNP to major foods where digestible energy and protein are the likely factors influencing why bears shift from one food to another. The diets of grizzly bears in KNP were similar to those of other grizzly bears in North America (Hamer and Herrero 1983; MacHutchon 1996; McLellan and Hovey 1995) and the seasonal patterns of energy and protein availability suggest that the scat analysis is consistent with expected foraging patterns. Bears may eat H. alpinum roots and over-wintered berries in early spring, prior to green-up (very few scats have been collected from this period in KNP) (Fig. 2). Hedysarum roots are higher in protein and energy and lower in fibre before new growth begins (Hamer and Herrero 1983; McLellan and Hovey 1995). The feeding on over-wintered berries may be associated with the sugar content in these mealy fruits (Hamer and Herrero 1983). The bears switched to the more accessible green vegetation when it became available (Fig. 2). Equisetum arvense plants have a higher protein content than Hedysarum roots (Hamer and Herrero 1983; McLellan and Hovey 1995) and often occur in dense patches which makes it easy for grizzly bears to consume large quantities. Flowers of O. campestris and other species were eaten when they became available (Fig. 2). Oxytropis campestris often grows in large clumps which again makes for easier grazing. When berries were ripe they quickly became the main food source (Fig. 2). Shepherdia canadensis berries are higher in energy, in fact they are higher than any other plant food in KNP, and this energy is probably important for conversion to fat for winter hibernation (Hamer and Herrero 1983; McLellan and Hovey 1995). 26 These berries are usually an abundant food source in KNP. As the main berry season ended, roots and later ripening species of berries came to dominate the diet (Fig. 2). The roots of Hedysarum began to increase in protein and energy, and decrease in fibre after the plants had set seeds (Hamer and Herrero 1983; McLellan and Hovey 1995). Food Abundance Comparison of food abundance values from this study with values in others is not possible because of geographical differences in vegetation types; therefore, I can draw only general conclusions. The BHTs where S. canadensis was most productive (berries/m2 of shrub) were those with the lowest overstory cover. Hamer (1996) found that fruits/m2 decreased rapidly with increasing forest canopy closure when canopy closure was >45%. He found a strong relationship, with 70% of the variation in S. canadensis fruit abundance explained by forest canopy cover. The abundance and productivity of other berry producing plants was also negatively correlated with increasing overstory cover (Noyce and Coy 1990). I observed edges of forested BHTs were productive sites for S. canadensis. Quantity of food is not the only variable influencing locations chosen by grizzlies to excavate roots. At sites where grizzly bears dig for roots, the most important factor influencing site selection is ease of digging (Edge et al. 1990; Holcroft and Herrero 1984; Mattson 1997). Plant abundance did not appear to influence selection of a site for excavation in two of the studies (Holcroft and Herrero 1984; Mattson 1997). However, Edge et al. (1990) found that mean canopy cover at used sites was significantly greater than at unused sites. I originally chose 'digability' of the soil and canopy cover as appropriate measures of H. alpinum availability. I measured digability using a 'clawometer', modelled after that used by Holcroft and Herrero (1984), in food abundance plots containing H. alpinum. This measurement represented an index 27 of the accessibility ofH alpinum roots to grizzly bears. When I summarised the plant food abundance results, I found that H. alpinum occurred predominantly in the floodplain shrub BHT (Table 4) where I measured the digging to be relatively easy. Therefore, to simplify calculations, I did not include digability when determining relative abundance of H. alpinum. To be useful to wildlife management practitioners, methods for measuring plant foods should not only be accurate and precise, but also be easy and quick to conduct. Because of the variability within BHTs, I sacrificed some accuracy for speed to enable more estimates to be made. I chose visual canopy cover estimates of plant foods as representative of the amount of bear food present. Observers estimating canopy cover may provide different assessments on different occasions, however, increasing experience leads to consistency and the ability "to give a fairly reliable comparison between different communities" (Grieg-Smith 1964). In my study, all observations were made by one observer to reduce the variability from comparing observations made by more than one observer. Canopy cover and digability estimates supported the floodplain shrub BHT as the best habitat for H. alpinum. Dense patches of E. arvense occur in moist pockets, most frequently in the spruce and deciduous/mixed forests and the grey willow shrub BHTs. The open areas of herb and shrub floodplain BHTs provided the most abundant canopy cover of O. campestris and grizzly bears browsed the flower heads in these areas. The highest berry densities of S. canadensis were on the open hillsides of shrub mosaic, grey willow shrub, and herb BHTs. Habitat Food Values Other studies have attempted to build food-based habitat models for bears (Clark et al. 1994; Kansas and Riddell 1993; Mace 1986; Noyce and Coy 1990). Most follow similar approaches: determine food items and their relative importance from scat analyses and the 28 literature, estimate bear plant food abundance within habitat types from canopy cover values, and calculate a habitat importance combining the two. My model differs from these models in three ways: 1)1 quantified fruit production for the major berry species (S. canadensis) using double sampling; 2) I estimated canopy cover values of the four main species to the nearest percent as opposed to putting them in a class, and 3) I standardised FVs and plant food canopy cover values. My model predicts the seasonal importance of a BHT to grizzly bears. Because there is variation in the amount of plant food within BHTs, grizzly bears might use any BHT that contains localised concentrations of seasonal food. The density and size of seasonal food patches within BHTs is reflected in the random sampling and calculation of an H F V for each BHT. Within a dietary season, the rank of a BHT reflects its estimated H F V relative to the H F V of other BHTs. The H F V rank estimates the attractiveness of a BHT to grizzly bears relative to the other BHTs for specific dietary seasons. I compare the predicted attractiveness with actual use in Chapter 3. 29 Chapter 3 Testing the Food-Based Habitat-Selection Model Introduction In Chapter 2,1 used scat analyses and measurements of plant food abundance to estimate seasonal habitat food values (HFVs) for nine different bear habitat types (BHTs). In this chapter, I test the effectiveness of this model using an evaluation of grizzly bear habitat selection. Habitat selection refers to the process of choosing habitats; it differs from preference which is the likelihood of a habitat being chosen i f it is offered on an equal basis with other habitats (Johnson 1980). Use, defined as the number of locations in a certain habitat, will be selective i f a habitat is used disproportionately more than its availability. I evaluate habitat selection with a use-availability study design. The method compares the number of radio-telemetry locations in each BHT to the relative area of each BHT. The whole study area is considered available to each radio-collared grizzly bear. To test my model, I compare the measured grizzly bear habitat selection to the habitat selection predicted by HFVs. The objectives of this chapter are to: 1) estimate seasonal habitat selection of BHTs by grizzly bears from locations of radio-collared grizzly bears, and 2) compare predicted food-based selection (HFVs) of BHTs to measured selection of BHTs. Methods Habitat Selection I determined habitat use from aerial locations of radio-collared grizzly bears within the study area. The KNP Grizzly Bear Research Project provided telemetry location data for collared grizzly bears collected from 1989-1996 (McCann, unpublished data). Ninety-five 30 percent of the locations were collected from 1992-1996. Data were collected using methods outlined in the study design (Wielgus et al. 1992). Radio-collared bears were monitored from the time that they left the den in the spring until they entered the den in the fall. Observers flew telemetry flights about once per week using fixed-wing aircraft. Additional aerial telemetry was conducted opportunistically during helicopter flights. Observers recorded locations of radio-collared grizzly bears on a 1:50,000 topographic map. Habitat or vegetation type, as determined from the air, was usually recorded with each location. Aerial telemetry was used because limited road access in KNP prohibited ground telemetry. The advantage of aerial telemetry was that the observer saw the collared grizzly bear on 59% of recorded locations and, therefore, could improve the accuracy of map locations. The disadvantage of aerial telemetry was that it restricted data collection to daylight hours resulting in 66%o of all locations being recorded between 11:00 and 14:59 (McCann 1997). The estimated error for telemetry locations on this project was a radius of 200 m (R. McCann, pers. comm.). Site investigations of telemetry locations were impossible because of limited access. To determine which BHTs grizzly bears used, I took the Universal Trans Mercator co-ordinates of collared grizzly bear locations and re-plotted them on 1:50,000 topographic maps. I placed the location onto a 1:25,000 black and white air photo using measurements made on the maps. I used observers' notes to assist in an accurate placement. From the airphoto, I recorded the BHT that the collared bear was in when located. I summarised the number of locations recorded in each BHT for each dietary season. I assigned most of the telemetry locations to a BHT without difficulty. I had difficulty assigning locations to a BHT for 5% of cases because of the lack of standardised vegetation types and because of the transition zones between vegetation types around the edges of BHTs. 31 The preferred method of analysing habitat use from radio telemetry data is to use the individual animal as the sampling unit (Aebischer et al. 1993). This was not a feasible option for this study because most of the individual collared bears had low numbers of locations per dietary season within the study area. No collared bear had 10 or more locations in the spring season, 2 collared bears had 10 locations in the early summer season, 5 collared bears had 10 or more locations in the late summer season, and 4 collared bears had 10 or more locations in the autumn season. The accuracy of estimated habitat use for individual bears would be very low (Aebischer et al. 1993). The low number of locations per individual per dietary season (no individual bear had greater than 17% of the total locations in any dietary season) reduced the possibility that one or a few bears would bias pooled habitat selection measures. Although habitat selection can vary among years (Schooley 1994), I pooled the data across years because there was no indication, such as a berry crop failure, to suggest that bears would be behaving differently among years. I used telemetry locations as the sampling unit, and I measured habitat selection using Manly's alpha, a standardised selection index (Chesson 1978; Manly et al. 1993; Manly et al. 1972). If all alpha are equal (0.111 for 9 habitats), then habitat use is random and no selection occurred. I calculated this selection index for each BHT in each dietary season using the program PREFER (Krebs 1989). I determined the area of each BHT, which represented the availability of that BHT, from the GIS map. Measured Habitat Selection Compared to Predicted Habitat Food Values To test the food-based habitat-selection model, I compared the HFVs against grizzly bear habitat selection. I compared the ranks of HFV against the ranks of habitat selection using Spearman's rank correlation (Zar 1984) to test if there was a correlation between food 32 abundance, modified by food values, and grizzly bear use. I used a one-tailed test because I was testing for a positive correlation. Results Habitat Selection I used 365 collared grizzly bear locations to measure habitat selection (Table 7). The number of grizzly bear locations varied among dietary seasons and among BHTs. The 6 weeks of the late summer season had 164 collared grizzly bear locations recorded, while only 31 locations were recorded for the spring season. The spruce forest BHT consistently had the highest number of locations, and the rock/ice BHT had the fewest number of grizzly bear locations recorded in it. Table 7. Telemetry locations in the study area summarised by dietary season, by gender, and by number of individuals. Spring Early summer Late summer Autumn Totals no. female locations 17 40 95 61 213 no. male locations 14 20 69 49 152 total no. locations 31 60 164 110 365 no. female grizzly bears 5 8 17 11 17 no. male grizzly bears 6 9 17 13 20 total no. individuals 11 17 34 24 37 There were similarities in habitat selection for BHTs across the four dietary seasons (Table 8). BHTs that had the highest selection — floodplain shrub and grey willow shrub — were selected disproportionately in all four dietary seasons. Similarly, BHTs that had low selection -alder shrub, Dryas herb, and rock/ice ~ were ranked low in all dietary seasons. There were two 33 Table 8. Grizzly bear habitat selection calculated using Manly's alpha (Manly et al. 1972). The BHTs were ranked from 1 (high) to 9 (low) within each dietary season based on habitat selection. When two or more BHTs had the same selection value, the average rank was given to each of the tied BHTs. Shaded areas indicate the three BHTs with the highest selection values for that dietary season. Number of telemetry locations listed in parentheses. Bear Habitat Type % study area Spring (n=31) Early summer (n=60) Late summer (n=164) Autumn (n=110) Manly's alpha rank Manly's alpha rank Manly's alpha rank Manly's alpha rank Spruce forest (FS) 37.6 0.10 4 0.09 4 0.06 5.5 0.07 5 Shrub mosaic (SM) 17.3 0.03 6 0.07 5 0.12 3 0.11 3 Alder shrub (SA) 11.0 0 8 0.02 7 0.03 8 0.02 8.5 Dryas herb (HD) 10.2 0 8 0 3.5 0.04 7 0.05 6.5 Herb (HE) 9.0 0.05 5 0.05 6 0.09 4 0.10 4 Rock/Ice (RI) 6.5 0 8 0 3.5 0.01 9 0.05 6.5 Deciduous/mixed forest (FM) 3.6 0.13 3 0.14 3 0.06 5.5 0.02 8.5 Floodplain shrub (SF) 3.4 0.54 1 0.22 2 0.31 1 0.42 1 Grey willow shrub (SW) 1.5 0.15 2 0.41 1 0.28 2 0.15 2 34 interesting exceptions to this pattern. Deciduous/mixed forest was ranked third highest in selection for both spring and early summer but dropped to the lowest rank for selection in the autumn season. Shrub mosaic was ranked sixth highest in selection in the spring but rose to third highest for late summer and autumn. Measured Habitat Selection Compared to Predicted Habitat Food Values Rank testing between HFVs of BHTs and collared grizzly bear locations resulted in significant correlations. Generally, BHTs ranked high in HFV were also ranked high in habitat selection, while those ranked low in HFV were correspondingly low in habitat selection (Fig. 3). A Spearman rank correlation indicated significant correlations between modelled HFV ranks and actual habitat selection ranks for each of the four dietary seasons (Table 9). I also looked at BHTs that had significant changes in position, defined as increasing or decreasing three or more ranks, across the four dietary seasons. Both deciduous/mixed forest and shrub mosaic BHTs showed this change. A plot of HFV vs habitat selection ranks suggests that as HFV increased, selection for that BHT also increased, and as HFV declined, selection for that BHT did also (Fig. 4) . 35 Rank of Predicted Habitat Selection Based on Habitat Food Values (HFVs) Spring • Early Summer O Late Summer x Autumn A O O • • x • O x x A • O A • x 3 Rank of Actual Habitat Selection Based on Locations of Radio-collared Grizzly Bears Figure 3. Rank of predicted habitat selection based on habitat food values (HFVs) compared to rank of actual habitat selection based on locations of radio-collared grizzly bears. A l l 4 dietary seasons have been combined on this graph. Ranks range from 1 (high) to 9 (low) for each dietary season. 36 Table 9. Correlations between rank of predicted habitat selection based on habitat food value (HFV) and rank of actual habitat selection based on radio-collared grizzly bear locations for all four dietary seasons (critical value: (r s) 0.05(i),9= 0.6) Dietary season Significance Spring 0.93 p< 0.001 Early Summer 0.89 p < 0.0025 Late Summer 0.92 p< 0.001 Autumn 0.86 p < 0.005 37 a) 9 T 8 + Rank of Predicted Habitat Selection Based on Habitat Food Values (HFVs) 7 6 + 5 4 + 3 2 1 A - + -1 b) 9 T 8-Rank of Predicted Habitat Selection Based on Habitat Food Values (HFVs) 2 4-1 -I 1 A 1 1 1 1 1 1 1 2 3 4 5 6 7 8 9 Actual Rank of Selection Based on Radio-collared Grizzly Bear Locations Figure 4. Changes in ranks across the 4 dietary seasons for 2 bear habitat types (BHTs). Rank of predicted habitat selection based on habitat food values (HFVs) compared to rank of actual habitat selection based on locations of grizzly bears: deciduous mixed forest (a) and shrub mosaic (b). Ranks range from 1 (high) to 9 (low) for each dietary season. 38 Discussion My study selected and measured habitat variables prior to assessing habitat use and made testable predictions. Plant food abundance adjusted for dietary pattern was a successful predictor of habitat selection by grizzly bears in the Alsek River Valley. I found a significant correlation between predicted H F V ranks and actual habitat selection ranks despite the low number of locations available to estimate habitat selection in the spring and early summer season. The overall similarity of selection ranks for each BHT among the four dietary seasons (Table 8) suggested that grizzly bears could be selecting habitats based on factors that did not change seasonally. However, for shrub mosaic and deciduous/mixed forest, the two BHTs that fluctuated in ranks among dietary seasons, a notable difference in selection rank between dietary seasons was also reflected by a similar change in H F V rank (Fig. 4). That is, i f a BHT's H F V increased or decreased, its use by grizzly bears correspondingly increased or decreased. This pattern further supports the predictive ability of this model. Other studies have modelled grizzly bear habitat selection using a food-based system. The scale of habitat classification affects the predictive success. In a grizzly bear habitat model developed for Banff, Jasper, Yoho, and Kootenay National Parks, six of seven months tested had significant correlations between food availability and grizzly bear habitat use with 174 ecosite units (Kansas and Riddell 1993). When ecosites were grouped into 61 functional units, only four of six months were significant (Kansas and Riddell 1993). In coastal B C , the habitat use of two grizzly bear females was not significantly correlated at the scale of 110 habitat types but when similar types were grouped into 14 'Bear Habitat Units', habitat use was correlated with food quantity and quality (Hamilton 1987). There is no apparent pattern in success between large scale habitat units and small scale habitat units. 39 The success of my model can be linked to several factors. Grizzly bears in KNP did not have a predictable source of animal protein, such as salmon or ungulates, so their diet was tightly tied to plant foods (c/Hamilton 1987; Lloyd 1979; Mattson et al. 1991; McLellan and Hovey 1995). The number of food plant species consistently used by grizzly bears was smaller than in other studies modelling habitat use (c/Clark et al. 1994; Kansas and Riddell 1993). Particularly noteworthy was the fact that only one major berry species (S. canadensis) and three minor species (E. nigrum, A. uva-ursi, and E. communtata) were used by grizzlies in KNP. Other studies involved the use of more than 10 different berry species that occur in many different habitat types (e.g. Hamilton 1987). The layout of my study area also probably contributed to the success of my model. The area was relatively small (about 200 km2) and could be realistically classified into nine habitat types. It was restricted to one valley, so plant food distribution was likely less variable within BHTs than if other valleys had been included. Food is an appropriate variable to predict grizzly bear habitat selection. Grizzly bears need to consume large quantities of high quality food in a relatively short period of time. Their movements are less constricted by security concerns than other animals because such threats are limited to conspecifics and humans. Lastly, the quantifiable nature of food makes it ideal for modelling. 40 Chapter 4 Summary and Management Considerations I examined the relationship between plant food abundance and diet, and grizzly bear habitat selection along a section of the Alsek River Valley, Yukon. My objectives were to: 1) estimate grizzly bear plant food abundance in different habitat types; 2) calculate plant food values from an analysis of collected scats; 3) derive a habitat food value as a product of plant food abundance and plant food value; 4) measure habitat selection by collared grizzly bears; and 5) determine whether grizzly bears selected habitats according to plant food abundance. I estimated the abundance of 10 plant foods in representatives of the 8 vegetated bear habitat types (BHTs). I measured a total of 478 plots along 57 transects during the summer of 1995 and 1996. I averaged and then standardised the food abundance estimates for each food in each BHT. To calculate plant food values, I summarised an analysis of scats collected in the montane zone of KNP (McCann 1997). Equisetum arvense was the dominant item in spring scats with H. alpinum roots and over-wintered A. wva-ursi berries making up much of the remaining diet. Equisetum arvense foliage and O. campestris flowers were favoured foods in the early summer. Late summer was the berry season and S. canadensis berries were the most common food consumed with H. alpinum roots the second most common food item. In the autumn season, H. alpinum roots were eaten most often and the berries of S. canadensis, E. nigrum , and E. communtata were eaten less frequently. Food values (FVs) were determined for each plant food species for each of the four dietary seasons by standardising the product of frequency and volume measurements from the scats. 41 For each dietary season, I calculated a H F V for each BHT from the sum of products of food abundance and food value for each plant food species. I ranked the HFVs within dietary season as a predictor of grizzly bear habitat selection. Grey willow shrub, mixed/deciduous forest, and floodplain shrub BHTs had the highest HFVs in the spring and early summer seasons. Shrub mosaic, grey willow shrub, and floodplain shrub BHTs had the greatest quantity of seasonal food in late summer. The floodplain shrub BHT had the highest amount of seasonal food in autumn. I grouped 365 collared grizzly bear locations by BHT and dietary season, and measured habitat selection using Manly's alpha (Chesson 1978; Manly et al. 1993; Manly et al. 1972). I ranked habitat selection within each dietary season to permit testing of model predictions. Floodplain shrub and grey willow shrub BHTs were the top two selected habitats in all four dietary seasons. Deciduous/mixed forest and shrub mosaic BHTs were the third highest selected habitats for the spring and early summer seasons and the late summer and autumn seasons, respectively. I tested the prediction of habitat selection, rank of HFV, against the measure of habitat selection, rank of habitat selection. The food-based habitat-selection model successfully predicted grizzly bear habitat selection. Grizzly bears selected BHTs of highest HFV. Grizzly bears require productive habitat to provide enough food for hibernation, growth, and reproduction (Bunnell and Tait 1981). The predictable presence of people in important food areas can impact bears. Concerns include displacing bears into less productive habitats, causing strong reactions that can be energy depleting and disruptive to bears, and encouraging aggressive responses of bears to humans (McLellan and Shackleton 1989). 42 Grizzly bears may become negatively conditioned to humans. Grizzly bears in the Jewel Basin Hiking Area, Montana, were significantly farther than expected from lakes with campsites and trails (Mace and Waller 1996). In Pelican Valley, Yellowstone National Park, occupancy significantly reduced the number of bear sightings within 400 m of backcountry campsites; observers recorded approximately 67% fewer bear sightings within 400 m of the campsites when they were occupied than when they were unoccupied (Gunther 1990). Grizzly bears may be forced into lower quality habitats if campsites are set up in or near habitats with high food abundance. People in backcountry areas can cause avoidance responses in grizzly bears that encounter them. In the Flathead River drainage, British Columbia, grizzlies always fled from people on foot in areas of low human use (n=10); in seven instances the bears ran more than a kilometre or out of the immediate drainage (McLellan and Shackleton 1989). In 18 observations of fleeing bears in Pelican Valley, the bears ran an average of at least 422 m before stopping in forest cover (Gunther 1990). The cost of these responses include disturbance of activity patterns and to a lesser extent, increased energetic outputs. Putting people in bear habitat obviously increases the chance of encountering a grizzly bear. In some cases, bears may respond aggressively towards humans by charging or attacking. Bears that feel threatened, like a surprised bear or a female with cubs, are more likely to react aggressively towards people. Even if bears do not respond aggressively, they may exhibit a strong tendency to investigate people. To minimise displacement of grizzly bears and the potential bear-human interactions, rafting campsites should be restricted to areas of lower quality bear habitat. They should be located at least 500 m from high quality habitat. From this project, we can infer habitat quality 43 from food abundance. A complete summary of H F V and habitat selection of BHTs is presented in Table 1 0 . The relative abundance of different plant foods within BHTs is presented in Table 1 1 . Given the apparent predictive capacity of the model, I can offer some clear guidelines to managing 'haul out' locations for rafters. During the prime rafting season from June through August, Dryas herb BHTs would provide little food incentive for grizzly bears and relatively flat areas for camp set-up. Conversely, the floodplain shrub BHT should be avoided by people whenever possible — it has high H F V and grizzly bear selection during all four dietary seasons. 44 Table 10. A summary of habitat food values (HFV) and habitat selection values for all four dietary seasons. Spring Early Summer Late Summer Autumn BHT H F V select- H F V select- H F V select- H F V select-ion ion ion ion Spruce forest (FS) med med med med med low med low Shrub mosaic (SM) med low med low high high med med Alder shrub (SA) low low low low low low low low Dryas herb (HD) low low med low low low low low Herb (HE) med low high low med med low med Rock/Ice (RI) (not sampled) low low low low low low low low Deciduous/mixed forest (FM) high high high high med low low low Floodplain shrub (SF) high high high high high high high high Grey willow shrub (SW) high high high high high high med high 45 Table 11. Relative abundance of grizzly bear plant foods in the nine bear habitat types (BHTs). (***** highest abundance; **** 75-99% of highest abundance; *** 50-74% of highest abundance; ** 25-49% of highest abundance; * 1-25% of highest abundance; - not measured; blank = none observed) B E A R FOOD Bear Habitat Type (BHT) A R U U ASTR C A R E X E L C O E M N I E Q A R GRASS H E A L O X C A SHCA Spruce forest (FS) * * * **** ** * * * Shrub mosaic (SM) #* * $ + ajc ***** Alder shrub (SA) * _ * * * ** * Dryas herb (HD) * * * * * * Herb (HE) * ***** s(e + % $ $ s|e jfc ** Rock/Ice (RI) (not sampled) Deciduous/mixed forest (FM) ***** ***** * $ $ $ $ * * * Floodplain shrub (SF) ***** $ $ $ $ $ $ $ $ $ $ $ * Grey willow shrub (SW) * * * * * * ** * ** *** A R U U = A. uva-ursi ASTR = Astragalus spp. C A R E X = Car ex spp. E L C O = E. communtata EMNI = E. nigrum E Q A R = E. arvense GRASS = grasses H E A L = H. alpinum O X C A = 0. campestris SHCA = S. canadensis 46 Literature cited Aebischer, N . J., P. A . Robertson, and R. E. Kenward. 1993. Compositional analysis of habitat use from animal radio-tracking data. Ecology 74: 1313-1325. Albert, T. M . , and R. T. Bowyer. 1991. Factors related to grizzly bear-human interactions in Denali National Park. Wildlife Society Bulletin 19: 339-349. Alldredge, J. R., and J. T. Ratti. 1986. Comparison of some statistical techniques for analysis of resource selection. Journal of Wildlife Management 50: 157-165. Blanchard, B. M . , and R. R. Knight. 1991. Movements of Yellowstone grizzly bears. Biological Conservation 58: 41-67. Boudreau, T. A . 1995. The role of topography in habitat selection by grizzly bears in the northcentral Alaska Range. M.Sc. Thesis. University of Alaska, Fairbanks. 86pp. Bunnell, F. L. , and T. Hamilton. 1983. Forage digestibility and fitness in grizzly bears. International Conference on Bear Research and Management 5: 179-185. Bunnell, F. L. , and D. E. N . Tait. 1981. Population dynamics of bears - implications. Pages 75-98 in T. D. Smith and C. Fowler, eds. Dynamics of large mammal populations. John Wiley and Sons, New York N . Y . Canadian Parks Service. 1990. Kluane National Park Reserve Management Plan. 52pp. Chesson, J. 1978. Measuring preference in selective predation. Ecology 59: 211-215. Clark, J. D., D. L . Clapp, K. G. Smith, and B. Ederington. 1994. Black bear habitat use in relation to food availability in the interior highlands of Arkansas. International Conference on Bear Research and Management 9(1): 309-318. Craighead, J. J., J. S. Sumner, and G. B. Scaggs. 1982. A definitive system for analysis of grizzly bear habitat and other wilderness resources. Wildlife-Wildlands Institute Monograph No. 1, University of Montana, Missoula, Montana. Daubenmire, R. 1968. Plant communities: a textbook of plant synecology. Harper & Row, New York. Despain, D. G. 1986. Habitat type and cover type as a base for grizzly bear habitat mapping and evaluation. Pages 230-233 in G. P. Contreras and K. E. Evans, eds. Proceedings - Grizzly Bear Habitat Symposium. USDA General Technical Report INT-207. Intermountain Research Station, Ogden, UT. Douglas, G. W. 1974. Montane zone vegetation of the Alsek River region, southwestern Yukon. Canadian Journal of Botany 52: 2505-2532. Douglas, G. W. 1980. Biophysical inventory studies of Kluane National Park: Vegetation. Douglas Ecological Consultants Ltd., Victoria, B.C. 47 Edge, W. D., C. L. Marcum, and S. L. Olson-Edge. 1990. Distribution and grizzly bear, Ursus arctos, use of yellow sweetvetch, Hedysarum sulphurescens, in northwestern Montana and southeastern British Columbia. Canadian Field-Naturalist 104: 435-438. Gray, B., J. 1987. Introduction. Pages 1-1 to 1-13. Kluane National Park Resource Description and Analysis. Natural Resource Conservation Section, Environment Canada, Parks, Prairie and Northern Region, Winnipeg. Grieg-Smith, P. 1964. Quantitative plant ecology. Univ. Calif. Press, Berkeley, Cal. Gunther, K. 1990. Visitor impact on grizzly bear activity in Pelican Valley, Yellowstone National Park. International Conference on Bear Research and Management 8: 73-78. Hadden, H. A. , W. J. Hann, and C. J. Jonkel. 1985. An ecological taxonomy for evaluating grizzly bear habitat in the Whitefish Range of Montana. Pages 67-77 in G. P. Contreras and K. E. Evans, eds. Proceedings - Grizzly Bear Habitat Symposium. USDA General Technical Report LNT-207. Intermountain Research Station, Ogden, UT. Hamer, D. 1996. Buffaloberry [Shepherdia canadensis (L.) Nutt.] fruit production in fire-successional bear feeding sites. Journal of Range Management 49: 520-529. Hamer, D., and S. Herrero. 1983. Ecological studies of the grizzly bear in Banff National Park: final report. University of Calgary, Calgary, Alberta. 303pp. Hamer, D., and S. Herrero. 1987. Grizzly bear food and habitat in the front ranges of Banff National Park, Alberta. International Conference on Bear Research and Management 1: 199-213. Hamilton, A. N . 1987. Classification of coastal grizzly bear habitat for forestry interpretations and the role of food in habitat use by coastal grizzly bears. M.Sc. Thesis. University of British Columbia, Vancouver, B.C. 86pp. Herrero, S. 1978. A comparison of some features of the ecology, evolution and behavior of black and grizzly/brown bears. Carnivore 1:7-17. Hewitt, D. G., and C. T. Robbins. 1996. Estimating grizzly bear food habits from fecal analysis. Wildlife Society Bulletin 24: 547-550. Holcroft, A . C , and S. Herrero. 1984. Grizzly bear digging sites for Hedysarum sulphurescens roots in southwestern Alberta. Canadian Journal of Zoology 62: 2571-2575. Johnson, D. H. 1980. The comparison of usage and availability measurements for evaluating resource preference. Ecology 61: 65-71. Jope, K. L. 1985. Implications of grizzly bear habituation to hikers. Wildlife Society Bulletin 13: 32-37. Kansas, J. L. , and R. N . Riddell. 1993. Grizzly bear habitat model for the four contiguous mountain national parks: 2nd iteration. Canadian Parks Service Report, Calgary, Alberta. 109pp. 48 Kingsley, M . C. S., J. A . Nagy, and R. H. Russell. 1983. Patterns of weight gain and loss for grizzly bears in northern Canada. International Conference on Bear Research and Management5: 174-178. Krebs, C. J. 1989. Ecological methodology. HarperCollins, New York. Leonard, R. D., R. Breneman, and R. Frey. 1990. A case history of grizzly bear management in the Slims River area, Kluane National Park Reserve, Yukon. International Conference on Bear Research and Management 8: 113-123. Lloyd, K. A . 1979. Aspects of the ecology of black and grizzly bears in coastal British Columbia. M.Sc. Thesis. University of British Columbia, Vancouver, B.C. 151pp. Mace, R. D. 1986. Analysis of grizzly bear habitat in the Bob Marshall Wilderness, Montana. Pages 136-149 in G. P. Contreras and K. E. Evans, eds. Proceedings - Grizzly Bear Habitat Symposium. USD A General Technical Report INT-207. Intermountain Research Station, Ogden, UT. Mace, R. D., and J. S. Waller. 1996. Grizzly bear distribution and conflicts in Jewel Basin Hiking Area, Swan Mountains, Montana. Wildlife Society Bulletin 24: 461-467. MacHutchon, A . G. 1996. Grizzly bear habitat use study, Ivvavik National Park, Yukon. Western Arctic District, Parks Canada, Inuvik, NWT. Manly, B. F. J., L . L . McDonald, and D. L . Thomas. 1993. Resource selection by animals. Chapman & Hall, New York. Manly, B. F. J., P. Miller, and L. M . Cook. 1972. Analysis of a selective predation experiment. American Naturalist 106: 719-736. Mattson, D. J. 1997. Selection of microsites by grizzly bears to excavate biscuit roots. Journal of Mammology 78: 228-238. Mattson, D. J., B. M . Blanchard, and R. R. Knight. 1991. Food habits of Yellowstone grizzly bears, 1977-1987. Canadian Journal of 'Zoology 69: 1619-1629. Mattson, D. J., R. R. Knight, and B. M . Blanchard. 1987. The effects of developments and primary roads on grizzly bear habitat use in Yellowstone National Park, Wyoming. International Conference on Bear Research and Management 7: 259-273. McArthur-Jope, K. L . 1982. Interactions between grizzly bears and hikers in Glacier National Park, Monatana. Cooperative Park Studies Unit, Oregon State University, Corvallis, Oregon. 61pp. McCann, R. K. 1996. Kluane National Park Grizzly Bear Research Project: Year End Report 1994. Centre for Applied Conservation Biology, University of British Columbia, Vancouver, B.C. 83pp. McCann, R. K. 1997. Kluane National Park Grizzly Bear Research Project: Year End Report 1996. Centre for Applied Conservation Biology, University of British Columbia, Vancouver, B.C. 76pp. 49 McCrory, W., S. Herrero, and P. Whitfield. 1986. Using grizzly bear habitat information to reduce human-grizzly bear conflicts in Kokanee Glacier and Valhalla Provincial Parks, BC. Pages 24-30 in G. P. Contreras and K. E. Evans, eds. Proceedings - Grizzly Bear Habitat Symposium. USD A General Technical Report INT-207. Intermountain Research Station, Ogden, UT. McLellan, B. N . , and F. W. Hovey. 1995. The diet of grizzly bears in the Flathead drainage of southeastern British Columbia. Canadian Journal of Zoology 73: 704-712. McLellan, B. N . , and D. M . Shackleton. 1989. Immediate reactions of grizzly bears to human activities. Wildlife Society Bulletin 17: 269-274. Noyce, K. V. , and P. L. Coy. 1990. Abundance and productivity of bear food species in different forest types of northcentral Minnesota. International Conference on Bear Research and Management8: 169-181. Pritchard, G. T., and C. T. Robbins. 1990. Digestive and metabolic efficiencies of black and grizzly bears. Canadian Journal of Zoology 68: 1645-1651. Robbins, C. T. 1993. Wildlife feeding and nutrition. Academic Press Inc, San Diego, California. Rogers, L. 1976. Effects of mast and berry crop failures on survival, growth, and reproductive success of black bears. Transactions of the 41st North American Wildlife and Natural Resources Conference. Wildlife Management Institute, Washington, D.C. 431-438pp. Schooley, R. L . 1994. Annual variation in habitat selection: Patterns concealed by pooled data. Journal of Wildlife Management 58: 367-374. Stelfox, H. A . 1991. Assessing land/wildlife relationships via the ecological land evaluation. Pages 29-46 in H. A . Stelfox, G. R. Ironside, and J. L. Kansas, eds. Guidelines for the integration of wildlife and habitat evaluations with ecological land survey. Wildlife Working Group of the Canada Committee on Ecological Land Classification. Wildlife Habitat Canada and Canadian Wildlife Service (Environment Canada), Ottawa, Ontario. Stelmock, J. J., and F. C. Dean. 1986. Brown bear activity and habitat use, Denali National Park -1980. International Conference on Bear Research and Management 6: 155-167. Vandehey, A. M . 1991. Vegetative productivity of grizzly bear habitat components in southeastern British Columbia. M.Sc. Thesis. University of Wisconsin, Oshkosh. 105pp. Watts, P. D., and C. Jonkel. 1988. Energetic costs of winter dormancy in grizzly bear. Journal of Wildlife Management 52: 654-656. Whittaker, R. H. 1978. Approaches to classifying vegetation. Pages 1-31 in R. H. Whittaker, ed. Classification of plant communities. Dr. W. Junk b.v., Boston. Wielgus, R., R. McCann, and F. L. Bunnell. 1992. Study design for Kluane National Park Reserve grizzly bear research program. Centre for Applied Conservation Biology, University of British Columbia, Vancouver, B.C. 52pp. Zar, J. H . 1984. Biostatistical analysis. Prentice Hall, Englewood Cliffs, New Jersey. 50