Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Activity measures of free-ranging grizzly bears (Ursus arctos) in the Flathead drainage McCann, Robert Keith 1991-11-22

You don't seem to have a PDF reader installed, try download the pdf

Item Metadata

Download

Media
831-UBC_1991_A6_7 M32.pdf [ 6.01MB ]
Metadata
JSON: 831-1.0098597.json
JSON-LD: 831-1.0098597-ld.json
RDF/XML (Pretty): 831-1.0098597-rdf.xml
RDF/JSON: 831-1.0098597-rdf.json
Turtle: 831-1.0098597-turtle.txt
N-Triples: 831-1.0098597-rdf-ntriples.txt
Original Record: 831-1.0098597-source.json
Full Text
831-1.0098597-fulltext.txt
Citation
831-1.0098597.ris

Full Text

ACTIVITY MEASURES OF FREE-RANGING GRIZZLY BEARS (Ursus arctos) IN THE FLATHEAD DRAINAGE by ROBERT KEITH MCCANN B.Sc, The University of British Columbia, 1985 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in THE FACULTY OF GRADUATE STUDIES (Department of Animal Science) We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA April 1991 © Robert Keith McCann, 1991 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 of AWVAAV^ <X_AtLUC£-The University of British Columbia Vancouver, Canada Date Apcuu a°\ mi DE-6 (2/88) ABSTRACT Between 1984 and 1988, 4756 hours of activity data were collected on 15 different grizzly bears (Ursus arctos) in the Flathead drainage of southeastern British Columbia and adjacent portions of Montana. Data were collected with the aid of portable chart recorders that recorded the output from motion-sensitive radio collars. While many benefits stem from remote sensing of a study animal as intractable as the grizzly, both the method of data collection and the assumptions employed in translating chart recordings into quantitative measures of bear activity may affect conclusions drawn. Major objectives of this study were: 1) to assess the validity of procedures employed to translate continuous chart recordings of signal patterns from motion-sensitive radio collars into quantitative measures of bear activity; 2) to assess whether active and inactive bout lengths were related to sex and age related differences in energetic requirements and seasonal differences in food type; and 3) to document activity budgets and patterns as functions of sex, age, season, and the daily solar cycle. In the absence of concurrent visual observations of grizzly bears and recorded signal patterns, the validity of procedures used to interpret chart recordings was assessed by estimating percent of time active (%TA) under varying definitions of active and inactive bouts, and by comparing %TA to values found by other researchers. Estimates of %TA were iii stable over the range of activity bout definitions examined. Stability resulted from bears spending most of their time in active and inactive bouts > 30 min duration. Estimates of %TA for this study agreed with results on other populations. Over the non-denning portion of the year, grizzly bears were active about 55% of the time. Analyses of bout durations were plagued by a bias against active bouts to be monitored in their entirety, because when active, bears frequently moved out of range of the chart recorder. The distribution of activity over the 24-hour cycle differed from many other studies in that bears in the Flathead were active mostly in daylight hours. A greater use of darkness by bears in the fall, compared to other seasons, may be related to available daylight or to avoidance of hunters. While activity patterns were generally bimodal with activity peaks in morning and evening, the morning activity peak was not strongly tied to sunrise. Activity in the morning generally reached a peak 1 or more hours after sunrise. Seasonal trends in activity budgets conformed to physiological changes in bears necessitated by requirements for denning. Significant individual variation exists in both activity patterns and budgets, and may be related to body size, to frequency dependent foraging strategies, or to differing competitive ability for defendable resources among sex-age classes of bears. 1 iv TABLE OF CONTENTS Page ABSTRACT ii TABLE OF CONTENTS iv LIST OF TABLES vLIST OF FIGURES viii LIST OF APPENDICES xiACKNOWLEDGEMENTS xiiCHAPTER 1: AN OVERVIEW i) GENERAL INTRODUCTION 1 ii) STUDY AREA 4 iii) FIELD PROCEDURES 6 iv) DISTRIBUTION OF DATA 8 V) LITERATURE CITED 12 CHAPTER 2: ASSESSING ACTIVITY FROM CHART RECORDINGS i) INTRODUCTION 5 ii) METHODS 20 Chart Interpretation 2Stability of Activity Estimates 28 Statistical Analyses 30 iii) RESULTS AND DISCUSSION 1 iv) CONCLUSIONS 42 V) LITERATURE CITED 5 CHAPTER 3: BOUT LENGTHS i) INTRODUCTION 48 ii) METHODS 54 Analyses Within Seasons 5V Annual and Among Seasons Analyses 56 iii) RESULTS 57 Annual and Seasonal Trends 59 Analyses Within Seasonsiv) DISCUSSION 65 V) LITERATURE CITED 71 CHAPTER 4: ACTIVITY BUDGETS AND PATTERNS i) INTRODUCTION 74 ii) METHODS 7 Comparisons of Activity Patterns 78 Comparisons of Activity Budgets 79 iii) RESULTS 80 Activity Patterns 8Activity Budgets 5 iv) DISCUSSION 94 Activity Patterns 96 Activity Budgets 103 V) LITERATURE CITED 9 CHAPTER 5: OVERALL CONCLUSIONS 113 APPENDIX 115 vi LIST OF TABLES Page Table 1. Distribution of recorded data (sum of active and inactive bout durations) by population component and month for 1984 through 1988 combined, in decimal hours. Number of bears in each monthly sample are in brackets. .... Table 2 Distribution of recorded data (sum of active and inactive bout durations) by population component and season for 1984 through 1988 combined, in decimal hours. Number of bears in each seasonal sample are in brackets. Table 3. Means, standard errors of the means, and ranges of percent of time active by sex-age class, and overall, calculated from complete 24-hour samples on grizzly bears Table 4. Means, standard errors of the means, and ranges of percent of time active by month, calculated from complete 24-hour samples on grizzly bears 10 11 34 37 Table 5. Percent of time active by population component and season as determined by complete active and inactive bouts (columns under C) and as determined by all available data (columns under A). Samples sizes of complete bouts are in brackets 58 Table 6. Table 7. Table 8. Tests of nonparametric multiple contrasts for durations of complete inactive bouts within seasons. Values are test statistics (value of contrast divided by its standard error) Tests of nonparametric multiple contrasts for durations of complete active bouts within seasons. Values are test statistics (value of contrast divided by its standard error) Mixed model nested-factorial Analysis of Variance of activity patterns for grizzly bears over seasons. SA Class represents those sex-age classes (subadult males, subadult females, adult females) for which there were sufficient data to test 62 63 81 Table 9. Mixed model nested-factorial Analysis of Variance of activity patterns for sex-age vii classes of grizzly bears over early summer and the berry season 84 Table 10. Mixed model nested-factorial Analysis of Variance of activity budgets for grizzly bears over seasons. SA Class represents those sex-age classes (subadult males, subadult females, adult females) for which there were sufficient data to test 88 Table 11. Mixed model nested-factorial Analysis of Variance of activity budgets for sex-age classes of grizzly bears over early summer and the berry season 93 viii LIST OF FIGURES Page Figure 1. Typical pulse mode patterns obtained from bears wearing 2.5-min reset delay collars, a) Active pulse modes interspersed with brief inactive pulse modes (marked by arrows), and b) inactive pulse modes interspersed with active pulse modes of approximately the 2.5-min reset delay period. One inch of chart trace represents about 4 min 23 Figure 2. Examples of assignment of activity bouts (upper continuous line) to pulse mode patterns (lower broken series) obtained from bears wearing 2.5-min reset delay collars. Read from left to right: U = unknown bout; A = active bout (as measured when following bout is unknown); I = inactive bout (note addition of 2.5 min correction); A = active bout (note subtraction of 2.5 min when following bout is inactive); I = inactive bout; A = active bout. One inch of chart trace represents about 4 min 24 Figure 3. Typical pulse mode patterns obtained from bears wearing 5-s reset delay collars, a) Active pulse mode pattern showing frequent switching between active and inactive pulses, and b) alternating active and inactive (solid lower trace) pulse mode patterns. One inch of chart trace represents about 4 min 26 Figure 4. Examples of assignment of activity bouts (upper continuous line) to pulse mode patterns (lower broken series) obtained from bears wearing 5-s reset delay collars. Read from left to right: I = inactive bout; U = unknown bout (note the inclusion of pulse mode patterns too short to classify at either end of the unknown bout); A = active bout (note inactive pulse mode patterns with durations < 2.5 min); I = inactive bout (note active pulse mode patterns with durations < 1 min); A = active bout. One inch of chart trace represents about 4 min 27 Figure 5. a) Percent of total active time as distributed between active bouts of different durations and seasons, and b) percent of total inactive time as distributed between inactive bouts of different durations and seasons 33 ix Figure 6. Percent of time active by month for a) adult males, b)adult females, c) subadult males, and d) subadult females as calculated from all data 36 Figure 7. Annual percent of time active by hour of the day for a) adult males, b) adult females, c) subadult males, and d) subadult females. For hours of the day, 1 = 0000-0100; 2 = 0100-0200 etc 38 Figure 8 Figure 9, Annual percent of time active by diel period for a) adult males, b) adult females, c) subadult males, and d) subadult females 40 Notched box plots of complete a) active, and b) inactive, bout durations pooled across seasons, for subadult females, adult females, subadult males, and adult males. Outside values were omitted to emphasize notches about the medians. For any 2 boxes with notches about the medians that do not overlap, the 2 medians are significantly different at approximately a 95% confidence level 60 Figure 10. Notched box plots of complete a) active, and b) inactive, bout durations pooled across sex-age classes, for spring (SP), early summer (ES), berry season (BS), and fall (FA). Outside values were omitted to emphasize notches about the medians. For any 2 boxes with notches about the medians that do not overlap, the 2 medians are significantly different at approximately a 95% confidence level Figure 11. First order interaction between seasons and diel periods from Analysis of Variance of grizzly bear activity patterns over seasons (Table 7). Plotted values are predicted cell means Figure 12. First order interaction between age classes and diel periods from Analysis of Variance of grizzly bear activity patterns for sex-age classes over early summer and the berry season (Table 8). Plotted values are predicted cell means Figure 13. Second order interaction between seasons, sex classes, and diel periods from Analysis of Variance of grizzly bear activity patterns for sex-age classes over early summer and the berry season (Table 8). Sex x diel period interactions are plotted for a) early summer, 61 83 86 X Figure 14 Figure 15, Figure 16, Figure 17, Figure 18, Figure 19, Figure 20, and b) the berry season, predicted cell means. Plotted values are 87 First order interaction between seasons and quarter day periods from Analysis of Variance of grizzly bear activity budgets over seasons (Table 9). Plotted values are predicted cell means 90 Second order interaction between seasons, SA classes, and quarter day periods from Analysis of Variance of grizzly bear activity budgets over seasons (Table 9). SA class x diel period interactions are plotted for a) spring, b) early summer, c) the berry season, and d) fall. Plotted values are predicted cell means 91 First order interaction between age classes and quarter day periods from Analysis of Variance of grizzly bear activity budgets for sex-age classes over early summer and the berry season (Table 10). Plotted values are predicted cell means 95 Percent of time active by hour of the day in spring for a) subadult females, b) adult females, and c) subadult males. Arrows indicate approximate times of sunrise and sunset. Plots were composed from all spring data collected. For hours of the day, 1 = 0000-0100; 2 = 0100-0200 etc 98 Percent of time active by hour of the day in early summer for a) subadult females, b) adult females, c) subadult males, and d) adult males. Arrows indicate approximate times of sunrise and sunset. Plots were composed from all early summer data collected. For hours of the day, 1 = 0000-0100; 2 = 0100-0200 etc Percent of time active by hour of the day in the berry season for a) subadult females, b) adult females, c) subadult males, and d) adult males. Arrows indicate approximate times of sunrise and sunset. Plots were composed from all berry season data collected. For hours of the day, 1 = 0000-0100; 2 = 0100-0200 etc Percent of time active by hour of the day in fall for a) subadult females, b) adult females, and c) subadult males. Arrows indicate approximate times of sunrise and sunset. Plots were composed from all fall 99 100 xi data collected. For hours of the day, 1 = 0000-0100; 2 = 0100-0200 etc 101 xii LIST OF APPENDICES Page Appendix 1. Data totals (sum of active and inactive bout durations) by bear, year, and season, in decimal hours 115 xiii ACKNOWLEDGEMENTS Many people have made contributions to this project - I can only thank a few explicitly. First, I thank Dr. D. M. Shackleton, my supervisor, for the opportunity to do this study and for his helpful insights. Special thanks goes to Dr. Bruce N. McLellan who first suggested the study topic, provided valuable advice and expertise on all aspects of the research, and collected some of the data used in this project. Bruce, and his wife, Celine Doyon, also made available an excellent field camp. Additional logistical support came from Dr. Chris Servheen of the U.S. Fish and Wildlife Service and Ray DeMarchi of the B.C. Wildlife Branch. Dan Carney arrived in the field at several opportune times and provided timely advice, good stories and an occasional chew. Tom Radandt helped with some of the field work in 1988. Dr. M. Grieg and S. Kita of the U.B.C. Computing Center, and Dr. Ray Peterson all provided statistical advice at various stages of the project. Discussions with F. W. Hovey and F. L. Bunnell contributed to some of the approaches used. 1 CHAPTER 1: AN OVERVIEW GENERAL INTRODUCTION Continuing expansion of human pursuits into the range of the grizzly bear (Ursus arctos) indicates that continued survival of many grizzly populations will depend on the ability of humans and bears to coexist. Studies on grizzly populations in environments that have a high level of human intrusion such as Yellowstone National Park and European game reserves, suggest that bears are capable of modifying their behavior not only to avoid areas of predictable human use but to take advantage of times when disturbance from humans is low. Whether such modification of behavior towards greater activity during the hours of darkness impinges on foraging efficiency and effectively lowers the carrying capacity of the environment is unclear. Currently, there is a need for greater understanding of when and how bears modify their behavior to avoid contact with humans, and how other factors such as sex, age, and time of year affect both the amount and scheduling of activity. Ultimately, the relationship between behavior and population demography needs to be determined. The influence of human predation on a species with low reproductive potential such as the grizzly (Bunnell and Tait 1981) is obvious. Of less certainty is the adaptability of the grizzly to habitat modifications (Knight 1980). The species is generally regarded as highly versatile, as is 2 demonstrated by its extensive historical range, solitary social system, omnivorous food habits, and great mobility (Knight 1980). However, the bears' survival depends on access to a variety of habitats within a given annual cycle, and an even greater variety over a life cycle, to alleviate stochastically occurring failures in food resources. Of equal uncertainty is whether the presence of human activity imposes a cost in terms of reduced reproductive fitness through alteration of behavior patterns This question is a contentious issue for many species due to the difficulty in linking disruption of behavior to demographic consequences (Shank 1979). In general, detrimental demographic effects could arise from 2 broad sources; exclusion from critical habitats, or altered activity budgets resulting in a lowered net energy status. Many studies have demonstrated behavioral responses by wildlife species to human activity (e.g., spatial distributions, Archibald et al. 1987; flight responses, Churchill 1982;) or physiological responses (e.g., cardiac rates, MacArthur et al. 1982), but not demographic consequences. While indices of animal condition and demographic parameters can reveal the significance of both habitat and human related behavioral disturbance effects, estimating these can be costly and time consuming. Twenty-four hour activity budgets are measures of how animals partition their time into various behaviors (Boy and Duncan 1979). As a component of bioenergetics, activity budgets have 3 application in population dynamics, habitat manipulation, assessment of resource development impacts (Hudson and White 1985; Jacobsen and Wiggins 1982) and the development of theoretical time-energy budget models (Jacobsen and Wiggins 1982). Activity patterns denote the temporal distribution of activity over the 24-hour cycle. Quantifying these activity parameters of a species under a variety of environmental conditions may ultimately permit their use as indicators of stress placed on a given population by its environment (Roth 1983). However, evidence of altered activity budgets or patterns are not in themselves evidence of detrimental effects on the population. To be of general utility, activity parameters need to referenced against known population status. Previous research on grizzly bears in the Flathead River drainage has documented short-term, overt behavioral responses to human disturbance, spatial distributions as influenced by roads, and population demographics (McLellan 1989a, 1989b, 1989c; McLellan and Shackleton 1989a, 1989b, 1988a, 1988b). The Flathead study area provides a unique opportunity to link activity parameters to a well documented population. Knowledge of activity patterns for grizzlies may also help reduce the occurrence of bear-human encounters and improve research schedules (Harting 1987). The main focus of this research was to document the temporal distributions (activity patterns) and amount of activity (activity budgets) of grizzly bears over the 24-4 hour cycle as functions of sex, age class, season, and the daily solar cycle. Additional hypotheses concerned durations of both active and inactive periods. However, it was not possible to correlate recorded signal patterns from motion-sensitive collars with concurrent observations of grizzly bears due to the difficulty of observing free-ranging bears. Therefore, it was necessary to first assess procedures employed to translate chart recordings into measures of bear activity through other means. Specific hypotheses are delineated and appraised in each of the following chapters. STUDY AREA The study area was centered along the North Fork of the Flathead River (114° 851 W; 49° 1' N) which drains the extreme southeast corner of British Columbia, Canada, and parts of northwestern Montana, U. S. A. The grizzly bear population in this area has been under study since 1978 by B. N. McLellan and others of the Flathead Grizzly Project (McLellan 1989a, 1989b, 1989c; McLellan and Shackleton 1989a, 1989b, 1988a, 1988b). Boundaries to the project's study area are determined by the home ranges of grizzly bears captured and collared in a 264 km2 central trapping area. These home ranges define an area of 2820 km2 (McLellan 1989a) spanning the international boundary with Montana. Monitoring of bears for activity data was conducted within a subset study area of approximately 1000 5 km2. This subset was bounded by the Clark Range on the east, the McDonald (Whitefish) Range on the west, with extensions 30 km north and 15 km south of the 49th parallel. The Flathead River flows south through the study area along a 5 to 10 km wide valley characterized by river flats, rolling hills and benches. Valley bottom elevations range from 1350 m above sea level in the north to 1100 m above sea level in the south. Mountain ranges rise up to elevations in excess of 2800 m above sea level on either side and harbor numerous side drainages. Lower elevations of the study area lie within the Montane Spruce Biogeoclimatic Zone (B.C.M.O.F. 1988) characterized by cold winters and moderately short, warm summers. Higher elevations fall within the Engelmann Spruce - Subalpine Fir Zone with long cold winters and short cool summers. The uppermost elevations are in the Alpine Tundra Zone which has a harsh climate. Approximate climatological features of the area include mean daily temperatures of -15 °C in January and < 16 °C in July, an average frost free period < 60 days, and mean annual precipitation of 100 - 150 cm (Farley 1979). Wildfires around the turn of the century, coupled with beetle infestations and logging in the mid-1900s and again in the 1970s and 1980s (Zager et al. 1983) have resulted in successional lodgepole pine (Pinus contorta) dominating lower elevations. Englemann spruce (Picea enqelmannii x gXauca), subalpine fir (Abies lasiocarpa) and white bark pine (Pinus albicualis) are locally abundant, becoming 6 common at higher elevations and in side drainages. Larch (Laryx occidentalism and alpine larch (Laryx lvalli) are scattered throughout the lower and upper elevations, respectively. Many high elevation sites, particularly those bordering on the Flathead Basin, have sparse tree cover and are dominated by huckleberry (Vaccinium globulare). Human impacts in the study area are sharply delineated by 3 political boundaries. The Canadian side has undergone extensive clearcut logging and exploration for oil and gas but has little human settlement. The American side is divided between a rural environment along the west side of the Flathead River and the relatively pristine wilderness of Glacier National Park on the east side of the river. With the exception of the park, big game hunting occurs throughout the study area and includes hunting for grizzlies. FIELD PROCEDURES Grizzly bears were captured by personnel of the Flathead Grizzly Project in Aldrich foot snares or culvert traps and outfitted with either 2.5-min or 5-s reset delay activity collars (Telonics Ltd, Arizona; 164 - 166 MHz). Both collar types utilized a mercury tip-switch that when activated by movement changed from an inactive signal pulse rate of approximately 50 pulses per min to an active signal pulse rate of approximately 75 pulses per min. While the pulse rate changed from inactive to active mode 7 instantaneously, the signal remained in active pulse mode for the time delay period (2.5-min or 5-s) before returning to inactive mode. Subsequent movement by the bear before expiry of the delay period caused the collars to reset to the beginning of the delay period without interruption of the active pulse mode. Signal pulse modes and amplitudes transmitted from the collars of individual bears were recorded using 2 remote data recording systems (Telonics Ltd, Arizona) that each consisted of a TR-2 telemetry receiver, a TDP-2 digital data processor, and a Rustrack dual channel strip chart recorder (Gulton Graphic Instruments Division, New Hampshire) all housed in a water resistant, metal case. Twelve volt car batteries or CF12V5PP rechargeable batteries (Eagle-Picher Industries, Missouri) were used to power the recording systems. The chart recorders were set to run at a chart speed of 16 in (40.6 cm) per h (actual speed varied with level of battery charge and ambient temperature). At this speed even isolated occurrences of individual 5 s periods of active pulse mode were readily identifiable. Data collection was conducted from a number of favorable locations generally on hill tops and benches, and when available from a high elevation fire lookout (2100 m above sea level) overlooking the study area. Recording sessions noted the bear's identification, general weather conditions, date, start and stop time of the recording, and location from which the recording was taken. Whenever 8 possible, the bear was relocated using telemetry. Recorders were checked frequently to ensure that the bear was still in range, and to mark time checks on the charts for calibrating chart speed. Sampling of individual bears was frequently opportunistic and data collection on a subject was continuous until the bear went out of range or an adequate sample (at least 24 h) was obtained. Individual recording sessions varied greatly in length from < 1 h to > 48 h. DISTRIBUTION OF DATA Activity recordings on grizzlies used in this study were collected over 5 consecutive years from 1984 through to 1988. A total of 4757 h of recorded data were collected with the majority obtained in 1987 and 1988 (60.6% and 23.5% of total, respectively). Data were collected on 15 individual bears, but due to the long time span over which data were acquired, several individuals passed from subadults (1 - 4 y of age) to adults (> 5 y old) and (or) changed reproductive status. For analytical purposes, bears that occurred in more than 1 sex-age or reproductive class were treated as independent individuals within each class. Appendix 1 outlines total time of recorded data for individuals by year and season and indicates sex, age class, and reproductive status. Pooled data totals on individuals within sex-age classes ranged from 41.5 hours to 641.9 hours. 9 Dates of data collection range from 13 April to 27 October, and for adult females and subadults of both sexes this approximately covers the period from den emergence to den entrance. Adult males generally emerge from dens earlier in the spring and enter dens later in the fall (McLellan pers. comm.). Additionally, data on adult males were sparse in both April and October (Table 1). Few adult females wearing activity collars were with cubs-of-the-year or yearlings during the study (Table 1). Based on phenological development of major grizzly bear forage species and observed habitat and dietary change of the bears (McLellan pers. comm.), the period from den emergence to den entrance was partitioned into four seasons: 1) spring - den emergence to the average date of green-up on May 9; 2) early summer - May 10 to the average date of berry ripening on July 24; 3) berry season - July 25 to the average date of bears leaving high elevation berry fields on September 19; and 4) fall - September 20 to den entrance. Most analyses were conducted within or among seasons (Table 2). Table 1. Distribution of recorded data (sum of active and inactive bout durations) by population component and month for 1984 through 1988 combined, in a decimal hours. Number of bears in each monthly sample are in brackets . Population Component April Mav June July August Sept Oct All Adult 157.98 470. 88 175.07 296.15 224.17 253.32 126. 05 Females (3) (4) (3) (4) (4) (4) (4) b with COYS 64.52 166. 57 37.45 — — 21.90 --(1) (1) (2) (1) with yearlings — — — 70.42 65.25 57.95 81.93 52 . 33 (1) (1) (1) (2) (2) alone or with 93.47 304. 32 67.20 230.90 166.22 149.48 73. 72 2-year-olds (2) (4) (3) (3) (3) (2) (2) Adult 16.02 149. 88 111.23 163.17 252.08 87.47 51. 82 Males (2) (3) (2) (2) (2) (3) (3) Subadult 38.60 258. 62 159.13 170.50 176.52 231.17 122. 25 Females (2) (3) (4) (2) (4) (4) (3) Subadult 74.68 49. 30 166.48 138.72 168.70 285.98 181. 02 Males (1) (3) (4) (3) (3) (4) (4) T0TALC 287.28 928. 68 611.91 768.54 821.47 857.94 481. 14 (8) (11) (11) (11) (11) (12) (11) Some bears appear in more than 1 population component due to changes in age class and reproductive status over the sample years. b COYS = cubs-of-the-year. c TOTAL = sum of adult females, adult males, subadult females, subadult males. Number of bears represents unique individuals. Table 2. Distribution of recorded data (sum of active and inactive bout durations) by population component and season for 1984 through 1988 combined, in a decimal hours. Number of bears in each seasonal sample are in brackets . Population Early Berry Fall Component Sprinq Summer Season All Adult 221.18 , 826.47 399.18 256.83 Females (3) (4) (5) (4) with COYSb 66.93 201.62 — 21.90 (1) (2) (1) with yearlings — 135.68 93.48 98.77 (1) (1) (2) alone or with 154.25 489.18 305.70 136.17 2-year-olds (2) (4) (4) (2) Adult 52.63 316.80 393.43 68.82 Males (2) (3) (2) (3) Subadult 230.48 361.55 313.53 251.27 females (3) (4) (5) (4) Subadult 90.05 291.62 477.33 205.88 Males (2) (5) (4) (4) TOTAL° 594.34 1796.44 1583.47 782.80 (9) (13) (13) (12) Some bears appear in more than 1 population component due to changes in age class and reproductive status over the sample years. b COYS = cubs-of-the-year. c TOTAL = sum of adult females, adult males, subadult females, subadult males, Number of bears represents unique individuals. 12 LITERATURE CITED Archibald, W. R., R. Ellis, and A. N. Hamilton. 1987. Responses of grizzly bears to logging truck traffic in the Kimsquit River valley, British Columbia. Int. Conf. Bear Res. and Manage. 7:251-257. Boy, V., and P. Duncan. 1979. Time-budgets of Camarque horses. 1. Developmental changes in the time-budgets of foals. Behaviour 71:187-202. British Columbia Ministry of Forests, 1988. Biogeoclimatic zones of British Columbia 1988. British Columbia Ministry of Forests, Victoria, British Columbia. 1:2,000,000 colored map. Bunnell, F. L., and D. E. N. Tait. 1981. Population dynamics of bears - implications. In Dynamics of large mammal populations. Edited by. C. W. Fowler and T. D. Smith. John Wiley & Sons Inc., New York, New York. pp. 75-98. Churchill, B. P. 1982. Winter habitat selection and use of clearcuts by elk in the White River drainage of southeastern British Columbia. M.Sc. Thesis, Univ. of British Columbia, Vancouver, British Columbia. 95 pp. Farley, A. L. 1979. Atlas of British Columbia: people, environment and resource use. UBC Press, Vancouver, British Columbia. 136 pp. Harting, A. L. 1987. Activity patterns. In Grizzly bear compendium. Edited by M. N. LeFranc, M. B. Moss, K. A. Patnode, and W. C. Sugg. U.S. National Wildlife Federation, Washington, D.C. pp. 33-35. Herrero, S. 1972. Introduction to the biology and management of bears. Int. Conf. Bear Res. and Manage. 2:11-18. Herrero, S. 1976. Conflicts between man and grizzly bears in the national parks of North America. Int. Conf. Bear Res. and Manage. 3:121-145. 13 Hudson, R. J., and R. G. White. 1985. Preface. In Bioenergetics of wild herbivores. Edited by. R. J. Hudson and R. G. White. CRC Press, Boca Raton, Florida. 314 pp. Jacobsen, N. K,. and A. D. Wiggins. 1982. Temporal and procedural influences on activity estimated by time-sampling. J. Wildl. Manage. 46:313-324. Knight, R. R. 1980. Biological considerations in the delineation of critical habitat. Int. Conf. Bear Res. and Manage. 4:1-3. MacArthur, R. A., V. Geist, and R. H. Johnston. 1982. Cardiac and behavioral responses of mountain sheep to human disturbance. J. Wildl. Manage. 46:351-358. McLellan, B. N. 1989a. Dynamics of a grizzly bear population during a period of industrial resource extraction. I. Density and age-sex composition. Can. J. Zool. 67:1856-1860. McLellan, B. N. 1989b. Dynamics of a grizzly bear population during a period of industrial resource extraction. II. Mortality rates and causes of death. Can. J. Zool. 67:1861-1864. McLellan, B. N. 1989c. Dynamics of a grizzly bear population during a period of industrial resource extraction. III. Natality and rate of increase. Can. J. Zool. 67:1865-1868. McLellan, B. N., and D. M. Shackleton. 1988a. Grizzly bears and resource-extraction industries: effects of roads on behaviour, habitat use and demography. J. Appl. Ecol. 25:451-460. McLellan, B. N., and D. M. Shackleton. 1988b. A comparison of grizzly bear harvest data from Montana and southeastern British Columbia. Wildl. Soc. Bull. 16:371-375. 14 McLellan, B. N., and D. M. Shackleton. 1989a. Grizzly bears and resource-extraction industries: habitat displacement in response to seismic exploration, timber harvesting and road maintenance. J. Appl. Ecol. 26:371-380. McLellan, B. N., and D. M. Shackleton. 1989b. Immediate reactions of grizzly bears to human activities. Wildl. Soc. Bull. 17:269-274. Roth, H. U. 1983. Diel activity of a remnant population of European brown bears. Int. Conf. Bear Res. and Manage. 5:223-229. Shank, C. C. 1979. Human-related behavioral disturbance to northern large mammals: a bibliography and review. Foothills Pipelines (South Yukon) Ltd., Calgary, Alberta. 253 pp. Zager, P. E., C. Jonkel, and J. Habeck. 1983. Logging and wildfire influence on grizzly bear habitat in northwestern Montana. Int. Conf. Bear Res. and Manage. 5:124-132. 15 CHAPTER 2: ASSESSING ACTIVITY FROM CHART RECORDINGS INTRODUCTION The use of biotelemetry to study activity budgets and patterns of free-ranging animals has been extensive with respect to the number of studies and the number of species. Obvious advantages to remote sensing include 1) reduction or elimination of observer effects on the subject's behavior, 2) freedom from the need for direct observation of the subject which can be compromised due to darkness, dense cover or mobility on the part of the subject, and 3) increased observer safety in the case of potentially dangerous animals. Disadvantages stem from relating the monitored analog to the behavioral state of the animal. Instantaneous changes in the variable monitored (e.g., signal integrity, signal pulse rate) introduce uncertainty (and hence inaccuracies), while procedural biases may favour particular behavioral states (Jacobsen and Wiggins 1982). Such biases may evolve from rules employed to categorize the analog to a behavioral state (Smith 1986; Lindzey and Meslow 1977), from sampling techniques, or from temporal changes in the subjects' behavior (Jacobsen and Wiggens 1982). Coupled with this are potential differences in the sensitivity of individual collars and the inability of most tip-switch collars to detect movement on more than one plane. Previous work on biotelemetric monitoring of grizzly and black bear (Ursus americanus) activity either used the 16 signal integrity of fixed pulse rate collars as an indicator of activity (e.g., Bjarvall and Sandegren 1987; Roth and Huber 1986; Harting 1985; Roth 1983; Lindzey and Meslow 1977; Amstrup and Beecham 1976), employed tip-switch collars that alternate between 2 pulse modes (slow and fast) depending on collar orientation (e.g., Harting 1985; Schleyer 1983; Garshelis et al. 1982; Garshelis and Pelton 1980), or used tip-switch collars with various reset delay periods ranging from 1 to 5 min (Clevenger et al. 1990; Ayres et al. 1986; Smith 1986; Garshelis et al. 1982). In the previous studies, collection of data involved periodically sampling the collars' signal pulse patterns (time-sampling) and employing various rules to determine if the bear should be classified as active or inactive. Both sampling period and intervals between samples varied greatly between studies. Only Smith (1986) collected data via continuous monitoring of radio collar signals (as used here continuous monitoring refers to the recording of all-occurrences of the behaviors of interest), however, it was done manually and was not extensive. In all the above studies, resolution of behavior was coarse-grained and involved only classifying behavior as either active or inactive. Studies utilizing signal integrity generally used qualitative rules to assign activity state. Rules used to determine the state of the bear for tip-switch collars, generally relied on counting the number of pulse mode changes per unit time and 17 classifying as inactive those samples containing less than some minimum number of changes. Assignment of a sample as active or inactive was independent of the previous or following samples' assignment. Jacobsen and Wiggins (1982) have termed this "instantaneous time-sampling" and have demonstrated for white-tailed deer (Odocoileus virqinianus) that it is unbiased in estimating time in activity, but is increasingly inaccurate as intervals between samples are increased, or the sample duration is decreased. With the exception of Clevenger et al. (1990), studies using time delay reset collars did not assign activity state to all samples instantaneously. For some samples, depending on the signal pulse mode monitored, assignment of activity state depended on a future sample meeting some criteria. This protocol was designed to differentiate between a collar set in active mode by brief movement of an inactive bear as opposed to a collar set in active mode by an active bear. Smith (1986) and Garshelis et al. (1982), when unable to instantaneously assign activity state to a sample, resampled the signal pulse mode after a time interval equal to or slightly longer than the collar's time delay reset period. Hence, the future sample required for decision making was not part of the systematic sampling scheme. In contrast, Ayres et al. (1986) sampled only systematically, and if one active sample occurred between two inactive samples, they classified the active sample as inactive on the assumption that it was only a brief comfort movement by the bear. When 18 future samples must be examined prior to assignment of behavioral state it is termed "conditional time-sampling" (Jacobsen and Wiggins 1982). Studies on the effects of conditional time-sampling by Jacobsen and Wiggins (1982) may not be directly applicable to studies utilizing reset delay collars and different conditional time-sampling rules. However, caution is warranted as a strong tendency for conditional time-sampling to underestimate time in activity has been demonstrated (Jacobsen and Wiggins 1982). Continuous monitoring of activity offers a more accurate account of time expenditures and sequencing among activities than does sampling at intervals (Jacobsen and Wiggins 1982). Following Altmann (1974), continuous monitoring results from the collection of focal animal samples and involves recording all-occurrences of the behaviors of interest performed by specific (focal) individuals over a sample period of known length. Such data are suited to studies of behavioral sequences, amount and percent of time spent in various behaviors, and durations of behavioral bouts (Altmann 1974). However, the collection of continuous data tends to be limited by the effort required to collect the data (Jacobsen and Wiggins 1982). In coarse-grained studies of activity (active vs inactive), diverse behaviors with diverse energetic costs are pooled into the active state (much less variation is expected among those behaviors pooled into inactivity). Aside from compromising time-energy budgets, a problem 19 arises when working with continuous data in terms of minimum bout lengths that can be recognized. For instance, energetically expensive activity of short duration should warrant special recognition as should short bouts of rest interspersed within vigorous activity. However, few activity studies on any species have employed continuous monitoring. Sorokin and Berger (1939), studying time-budgets of humans recorded every activity that lasted > 5 min. Cederlund (1981) in work with roe deer (Capreolus  capreolus) also did not recognize bouts < 5 min but rather included such short bouts with adjacent activity categories. Cederlund and Lemnell (1980) presented results of continuous activity monitoring on roe deer and mountain hares (Lepus  timidus). with activity bouts > 2 min recognized as distinct units. These latter 2 studies also distinguished 3 different types of active behavior. Smith (1986) studied hibernating black bears and recognized all-occurrences of active behavior, with brief occurrences of < 1 min duration recorded as 1 min bouts. Because uncertainty exists in interpreting the monitored analog, as shorter and shorter activity bouts are recognized it is expected that the error rate in activity classification will increase. But, errors in classification must be random if the mean estimates of activity are to be unbiased. Alternatively, systematic errors in classification will bias the means but will not affect the 20 results of significance testing, provided that the magnitude and direction of errors are comparable across samples. Specific objectives of this chapter were to: 1) assess the stability of chart interpretations with varying definitions of minimum bout duration; 2) describe general trends in activity budgets and patterns; and 3) assess the validity of procedures used to interpret continuous chart recorded samples of bear activity, through qualitative comparisons with previous research on activity of bears. METHODS Chart Interpretation Field procedures were as described in Chapter 1. Activity collars used in this study were either of 5-s or 2.5-min reset delay types (82% and 18% of recorded data, respectively). Interpretation of chart recordings had to be consistent between the 2 collar types if data were to be pooled. For the 2.5-min reset delay collars, the minimum detectable inactive duration had a lower bound of 2.5 min while active periods meant that the bear was triggering the tip-switch at least once every 2.5 min. Data from the 5-s reset delay collars were interpreted in a manner consistent with these restrictions. There was little opportunity to directly observe collared bears (bears wearing 2.5-min reset delay collars were never observed). When collared bears could be observed and simultaneously recorded or listened to with a receiver, 21 it was apparent that bears wearing 5-s reset delay collars could be active for almost 1 min without tripping the tip-switch from inactive mode. Therefore, signal pulse patterns had differential reliability of the information that they contained - active pulse modes indicated movement but short durations of inactive pulse modes did not preclude movement by the bear. Based on the limitations of the 2.5-min reset delay collars and the observations of bears wearing 5-s reset delay collars, minimum durations for recognition of active and inactive bouts were chosen to be 1 min and 2.5 min, respectively. Active bouts were measured to a shorter duration to account for the differential in information reliability. Interpretation of charts consisted of 2 steps. Pulse mode patterns were first categorized into those consistent with active and inactive states based on mutually exclusive, quantitative rules. Categorized patterns were then assigned as active or inactive bouts if the pattern persisted for at least the respective minimum duration (1 or 2.5 min) required for recognition. Categorized pulse mode patterns too short to be recognized as bouts were combined with adjacent recognizable bouts. Durations > 5 min, where the bears* state could not be determined, were recorded as unknown bouts, shorter durations were interpolated across if they occurred between like bouts, or split evenly if they occurred between unlike bouts. 22 Because there was variation in the chart speeds, due to the level of battery charge and (or) ambient temperature, the chart distance that represented a given duration varied within and between charts. Average chart speed was calculated between time checks on the charts and calipers were set to the distances that represented the durations used to differentiate pulse mode patterns and bouts. Pulse mode patterns differed greatly between the 2 collar types and necessitated different rules for categorizing patterns into the 2 states. Specific rules for each collar type were as follows: a) 2.5-min Reset Delay Collars The 2.5-min delay period greatly constrained the frequency with which the pulse modes could alternate. Movement which tripped the tip-switch resulted in a minimum active mode of 2.5 min. However, the delay period did not apply to inactive pulse modes and the collar could return to inactive mode for very brief periods before been tripped by movement again (Fig. 1). Because of the reset delay period, points of change in pulse mode from active to inactive were biased indicators of when the bear ceased movement. Therefore, a correction of 2.5 min was added to the beginning of any inactive pulse mode and subtracted from the end of the preceding pulse mode (Fig. 2). Because of the required 2.5 min correction, any occurrence of inactive pulse mode was recognized as an inactive bout. Active pulse Figure 1. Typical pulse mode patterns obtained from bears wearing 2.5-min reset delay collars, a) Active pulse modes interspersed with brief inactive pulse modes (marked by arrows), and b) inactive pulse modes interspersed with active pulse modes of approximately the 2.5-min reset delay period. One inch of chart trace represents about 4 min. I Figure 2. Examples of assignment of activity bouts (upper continuous line) to pulse mode patterns (lower broken series) obtained from bears wearing 2.5-min reset delay collars. Read from left to right: U = unknown bout; A = active bout (as measured when following bout is unknown); I = inactive bout (note addition of 2.5 min correction); A = active bout (note subtraction of 2.5 min when following bout is inactive); I = inactive bout; A = active bout. One inch of chart trace represents about 4 min. 25 modes that preceded inactive bouts were recognized as an active bout when the duration between the first active pulse mode and the reset delay correction was > 1 min (Fig. 2). When active pulse modes preceded a bout classified as unknown, the last movement by the bear to trigger the tip-switch was assumed to be at the last active pulse mode. An active bout was recognized if the active pulse mode duration was > 3.5 min thus ensuring that the subject minimally was active for > 1 min (Fig. 2). If the pulse mode duration was < 3.5 min, the active pulse modes were pooled with the unknown bout. b) 5-s Reset Delay Collars Pulse mode patterns were much more variable for 5-s reset delay collars and active bears generally had a pattern characterized by frequent switching back and forth between active and inactive pulses (Fig. 3). In some instances, active bears did demonstrate a steady active pulse. Pulse patterns consistent with the active state were where active pulse modes occurred within 2.5 min of each other. An active bout was recognized if a pulse pattern consistent with the active state had a duration > 1 min, as measured from the first active pulse to the last active pulse (Fig. 4). Durations of inactive pulse modes > 2.5 min, as measured from the first to the last inactive pulse, were recognized as inactive bouts (Fig 4). ~~r~ —1 Figure 3. Typical pulse mode patterns obtained from bears wearing 5-s reset delay collars, a) Active pulse mode pattern showing frequent switching between active and inactive pulses, and b) alternating active and inactive (solid lower trace) pulse mode patterns. One inch of chart trace represents about 4 min. Figure 4. Examples of assignment of activity bouts (upper continuous line) to pulse mode patterns (lower broken series) obtained from bears wearing 5-s reset delay collars. Read from left to right: I = inactive bout; U = unknown bout (note the inclusion of pulse mode patterns too short to classify at either end of the unknown bout); A = active bout (note inactive pulse mode patterns with durations <2.5 min); I = inactive bout (note duration <1 min of active pulse mode patterns within inactive bout); A = active bout. One inch of chart trace represents about 4 min. 28 For each recorded sample on a bear, the length on the chart of each bout (including unknown bouts) was measured to the nearest 1.0 mm and entered sequentially into a computer file along with the chart speed and bout category. Data files were subsequently collated with information on sex, age, reproductive status, collar type, and date. Starting times of recordings and the starting and stopping times of each bout were calculated to the nearest 1.0 min. Time of day was in Mountain Daylight Time. Total durations spent in active and inactive bouts were determined by multiplying each bout's measured length by the chart speed to determine duration of time, and summing by bout category. Percent of time active (%TA) was determined as total time in the active bout category divided by the sum of total time in both active and inactive categories. All continuous 24-hour samples of bear activity that were 98% complete (i.e., < 2% of sample coded as unknown bouts) were identified. Stability of Activity Estimates Chart interpretation rules were designed primarily to permit a detailed and repeatable characterization of the chart pulse patterns. Short active and inactive bouts may not represent true activity states on the part of the animal. Short active bouts occurring within long inactive bouts may only reflect brief head movements or comfort movements of a resting bear. Short inactive bouts interspersed between long active bouts may be due to collar 29 insensitivity to movement in certain planes or to the animal being in a vigilant state. To assess the impact of potential errors in activity classification on resultant estimates of %TA, I wrote a computer program that permitted me to increase the minimum durations required for recognition of bouts. The durations of all 3 bout types (active, inactive, unknown) could be varied. Lower bounds to durations were determined by those used to interpret the charts. Decision rules were applied by the program's algorithm to the data in hierarchical fashion in successive passes through the data. In the first pass, the duration of each bout was checked against the new respective minimum duration and coded as a success if it met the minimum duration. Bouts failing to meet the new minimum duration were coded as a failure and could occur either singly or in groups between successes. In the second pass, single failures were pooled with or split between adjacent successful bouts. The third pass analyzed the groups of failures as to the type of activity they best represented in relation to bounding successful bouts, and pooled, split or assigned an activity classification as required. All 5-s reset delay collar data were analyzed under 5 different selected sets of definitions with respect to minimum bout durations required for recognition: set #1 -active bouts = 1 min, inactive bouts = 2.5 min, missing bouts = 5 min (this set was identical to the chart 30 interpretation rules); set #2 - active bouts = 2.5 min, inactive bouts = 2.5 min, missing bouts = 5.0 min; set #3 -active bouts, inactive bouts, and missing bouts, all = 5.0 min; set #4 - active bouts, inactive bouts, and missing bouts, all = 7.5 min; set #5 - active bouts, inactive bouts, and missing bouts, all = 10.0 min. Sets of definitions were selected arbitrarily but were designed to encompass a range of bout definitions to permit exploration of the impact of varying definitions on %TA. Statistical Analyses The effect of the different sets of bout definition rules on %TA was investigated within each season using fixed model repeated measures analyses of variance (Hicks 1982, p. 239), where pooled data for each sex-age class were the subjects on which repeated measures (the 5 sets of rules) were taken. Complete 24-hour recordings were used to test differences in %TA between sex, age and individuals in a mixed model nested-factorial analysis of variance (Hicks 1982, p. 233), and differences between months and individuals in a mixed model repeated measures analysis of variance. Analyses based on complete 24-hour recordings were unbalanced with missing cells. Data were arcsine square root transformed (Sokal and Rohlf 1981, p. 427) prior to analyses. Analyses were performed on UBC GENLIN (Grieg and Bjerring 1978). Differences between means were 31 investigated with Newman-Keuls range tests (Hicks 1982, p. 51). The alpha level for all tests was 0.05. RESULTS AND DISCUSSION From a practical perspective, %TA tended to be resistant to the different sets of definitions tested. Subadult males in the spring demonstrated the largest range in %TA (54.3% - 62.3%) between definition sets #2 and #5, respectively. Adult males in the fall (34.5% - 37.2%) and spring (30.8% - 33.2%) showed the next largest ranges between sets #5 and #1, respectively. All 3 of these "subjects" were characterized by small amounts of data. Analyses of variance showed significant differences in %TA among the sets of definitions in both the berry season and fall, with %TA for definition set #1 being significantly less than %TA for all other sets of definitions in both seasons. However, the largest absolute difference in %TA among the sets of definitions was small in every season (2.28% for spring, 0.75% for early summer, 0.85% for berry season, 1.38% for fall). Absolute differences were related to the total hours of activity data in each season - the more hours of activity data, the less was the largest absolute difference - suggesting an averaging-out effect. Large changes in %TA were precluded by the distributions of active and inactive time across bouts of different durations. Under the rules used for chart interpretation, the most frequently occurring bout durations were those < 5 32 min, but most of the bears' time was spent in bouts > 30 min (Fig. 5a and 5b). Consequently, never more than about 15% of the data were at risk of been reinterpreted by the program. Without independent estimates of activity (e.g., concurrent visual observations of grizzly bears and recorded signals) there were no criteria for selection of one set of definitions over another. However, given the small differences between estimates of %TA, the original chart interpretations were deemed acceptable and used for all subsequent analyses. Overall %TA for all data combined was 55.4%. Data were not evenly distributed over the diel cycle but were biased to daytime hours. Applying equal importance to each hour of the diel cycle resulted in an unweighted overall %TA of 53.7%. Unweighted estimates for sex-age classes were 47.3% (adult males), 54.3% (adult females), 55.3% (subadult males), and 55.1% (subadult females). A total of 44 complete 24-hour samples, representing 22% of the entire data set, were obtained and used to test effects of sex, age, and individual bears. Only individual bears were significant. Estimates of %TA calculated from complete 24-hour samples were similar to unweighted estimates calculated from all data combined (Table 3). Ranges of %TA for complete 24-hour samples were similar to those reported by Hechtel (1985; 15% - 74%). All sex-age classes showed a definite trend for %TA (weighted by hour of the day) to increase to a maximum LU F 5 < I UL o UJ u QC LU CL 100-1 80-80-70 60-60-40-£ 30-20-•m LU LU > < z e LL o LU u DC LU 0. 100-1 80-80-Bout Durations ZZ] <5 min ES3 >5 min,<10 min ED >10 min <15 min KB >15 min,<30 min >30 min SPRING EARLY BERRY SUMMER SEASON FALL SPRING EARLY BERRY SUMMER SEASON FALL Figure 5. a) Percent of total active time as distributed between active bouts of different durations and seasons, and b) percent of total inactive time as distributed between inactive bouts of different durations and seasons. oo oo 34 Table 3. Means, standard errors of the means, and ranges of percent of time active by sex-age class, and overall, calculated from complete 24-hour samples on grizzly bears. Sex-acre class X SE n Rancre Adult males 48.2 7.0 5 26.4 - 65.3 Adult females 50.0 5.4 15 11.7 - 82.3 Subadult males 53.2 4.1 10 23.6 - 67.7 Subadult females 55.6 4.4 14 17.8 - 73.3 Overall 52.3 2.6 44 11.7 - 82.3 35 around July and August followed by a marked decline through September and October (Fig. 6). Levels of %TA were similarly low in April with the exception of subadult females, however, recorded data totals were small for both subadult females and adult males in this month (see Table 1) • Using complete 24-hour samples to test months (October was excluded due to only 1 observation), individuals, and their interaction, resulted in significant month and individual main effects. A range test on the predicted means for months resulted in only April being significantly less than August. Complete samples obtained in the months of April, May and June may be biased towards inactivity (Table 4). Bears exhibit more frequent movements in these months (McLellan and Shackleton 1989) and frequently went out of range of the recorders. Consequently, complete samples may overrepresent the behavior of bears that periodically restricted their activity to localized areas. Annual activity patterns by hour of the day were strongly bimodal for all sex-age classes (Fig. 7), and showed consistent trends for low nocturnal levels of %TA and a diurnal minimum in %TA around 1400 hours. Shifts in peak levels of %TA between 0700 - 0900 hours and between 2000 -2100 hours were apparent between the different sex-age classes. Times of sunrise and sunset vary throughout the year and data on different sex-age classes were not equally distributed throughout the months of data collection. To PERCENT OF TIME ACTIVE PERCENT OF TIME ACTIVE H. (D o» o & n> c 3 i- rt rt o 0) t—' rt 0) H-W 9 - (D P> 0> 3 O Oi rt H-PJ < — (D 0, o C 3 M rt rt M) Hi (D O S H 01 l_< p) (0 ^-01 (1) 01 e M O rt 9) t—1 9 O 0> C M I—1 (D 0) 01 rt-(0 n o. O J= 3 M rt 0) M M) M (0 3 0) H-1 rt o> dt m o PERCENT OF TIME ACTIVE PERCENT OF TIME ACTIVE 37 Table 4. Means, standard errors of the means, and ranges of percent of time active by month, calculated from complete 24-hour samples on grizzly bears. Month X SE n Rancre April 30.7 15.3 2 15.5 - 46.0 May 47.4 3.4 9 33.1 - 62.1 June 34.5 8.7 4 17.7 - 54.4 July 69.0 5.2 5 55.7 - 82.3 August 62.3 3.7 11 26.4 - 71.8 September 50.1 4.9 12 11.7 - 73.3 October 42.8 _ _ _ 1 90-2 F LL O 80 70 eo B0-| 40 30 20 ioH 0 ll Ii ••l 1 ii •i llm Ik II •iii 1 •111 1 2 3 4 6 6 7 8 9 10 11 12 13 14 15 16 17 IB 192021222324 TIME OF DAY LU 2 F 90 80 70 80 80 40 30 M H 10 0 b J ri L i Pi I J i •ill ii I ii iiiii 1 2 3 4 6 S li I 111 III lilii iisllii it •••I i ii in Hi 7 S 9 10 11 12 13 H 18 16 17 18 1920 21222324 TIME OF DAY 100 90 H 80 70H 80 60 40-30-20 10 0 11 * J III 1111 1 ! ii 2 3 4 6 8 7 8 9 K> 11 12 13 14 16 18 17 18 19 20 21 22 23 24 TIME OF DAY 100-90 < LU 80 5 o J- 40 LU O 30 20-| 10 0 d ll lilii -J I »1 ill Jllllllllllfciiii •l • Iii I— Ml 1 2 3 4 8 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 TIME OF DAY Figure 7. Annual percent of time active by hour of the day for a) adult males, b) adult females, c) subadult males, and d) subadult females. For hours of the day, 1 • 0000-0100; 2 = 0100-0200 etc. U) 00 39 standardize the data with respect to solar events, four diel periods were defined. Morning and evening (crepuscular) periods were each 3 h long and extended from 1.5 h before, to 1.5 h after, sunrise and sunset respectively. The intervening durations between these crepuscular periods were assigned as diurnal and nocturnal periods (Fig. 8). In spite of the strong bimodal patterns (Fig. 7), only adult males showed a strong crepuscular pattern (Fig. 8) while the remaining sex-age classes showed small and inconsistent differences in %TA between diurnal and crepuscular periods. Activity peaks for adults males tended to be narrow (Fig. 7) while the other sex-age classes displayed much wider peaks that extended into the diurnal period. Among other studies, average activity measures are over varying portions of the spring, summer and fall. However, many estimates are similar to those reported here. Roth (1983) found European brown bears to be active 47% of the time as calculated from all data. For 7 data sets with adequate data grouped by bear and month, he calculated a mean active level of 53%. Roth and Huber (1986) reported average active levels of European brown bears of 61% and 49% for a 2-year-old female and 4-year-old male, respectively. Combined means of fully active behavior and associated stationary active behavior for grizzlies on the east front of the Rockies (Aune and Stivers 1983), resulted in an estimate of 11.9 h of activity per 24-hour cycle (49.6%). Aune and Kasworm (1989) concluded that grizzlies on the east vQ C CD CO PERCENT OF TIME ACTIVE PERCENT OF TIME ACTIVE 3 3 PJ fl> 01 - T3 fl> 0 ^ — o CD cn 3 £ ft cr ca o & M> c M rt rt H-3 CD P> n> o 01 ft - H-<J PJ (D 3 & er & — o. 01 (D cr pj *rJ CL CD H" H-rt O 0< H) CD Hi o 3 PJ M (D PJ 01 ^ P> O. 3 PJ H-1 CD 01 PJ PERCENT OF TIME ACTIVE PERCENT OF TIME ACTIVE 41 front of the Rockies were involved in 10 - 15 h of vigorous activity over a 24-hour period (41.6 - 62.5%). Schleyer (1983) found grizzlies in Yellowstone to exhibit a yearly average active level of about 8 - 12 h per day (33.3 -50.0%). Harting (1985) also worked in Yellowstone and did not present an average activity estimate, however, the probability of bears been active averaged across the months of his study was 0.61 (calculated from data of Fig. 10, Harting 1985). From direct observational studies of grizzlies in northern Alaska, annual active estimates were 77.5% (calculated from data of Table 1, Phillips 1987), 64.0% (Gebhard 1982), and 40% (calculated from data of Hechtel 1985). Most observations made by Hechtel (1985) were obtained prior to mid-June. For a female grizzly observed from spring to fall Hechtel (1985) reported an activity level of 55%. The general pattern of variation in activity over months reported here, is similar to that found by Harting (1985) for Yellowstone grizzlies where active levels increased from May through to a peak in July, and then declined through to September. High daily active levels in July and August coincide with the occurrence of hyperphagic behavior by grizzly bears (Watts and Jonkel 1988). Garshelis and Pelton (1980) found a similar pattern for black bears in the Great Smoky Mountains. Studies on grizzly and black bears support a general trend of low daily 42 active levels in post- and pre-denning months (e.g., Schleyer 1983; Amstrup and Beecham 1976). Bimodal activity patterns by hour of the day or diel period are commonly reported in the literature for both grizzlies (Aune and Kasworm 1989; Harting 1985; Schleyer 1983; Roth 1983) and black bears (Ayres et al. 1986; Garshelis and Pelton 1980). The main difference among studies (and among bears within studies) tends to be whether bears are relatively more active during the nocturnal hours separating the bimodal peaks or during the diurnal hours. Such variation in activity patterns may be a product of several factors including foraging behavior (Ayres et al. 1986; Harting 1985) and level of human disturbance (Aune and Kasworm 1989; Roth 1983). ^ CONCLUSIONS The consequences of potential errors in activity classification appear to be small based on the assumption that the longer the minimum duration required for recognition the more likely the bout is to be classified correctly. Under the bout definitions examined, durations of active and inactive pulse patterns less than the minimums recognized, accounted for a small percent of the bears' time. The influence of a bias to active bouts in the chart interpretation rules introduced by recognition of bouts down to > 1 min also appears minimal. Short active bouts occur with the greatest frequency in the data but their influence 43 on estimates of %TA is minimal because the bouts are weighted in direct proportion to their duration. Systematic errors in classification using time-sampling, likely would have a greater effect because each sample is weighted equally. Ideally, statistical comparisons of activity budgets should be based on complete 24-hour samples of activity to capture complete diel cycles and to standardize the duration of time that a sample encompasses. However, such samples are rare in the data and are poorly distributed across months and individuals. Analyses based on complete 24-hour samples were unbalanced and suffered from missing cells. All alpha levels, as well as assumptions of no correlations between means for range tests based on studentized ranges, may be compromised (Grieg and Bjerring 1978). In a study such as this one, determination of what constitutes an observation is not straightforward. Time-sampling of the charts and retaining each sample as an observation would have resulted in a sample size of several thousands. However, these are unlikely to be independent. On the other hand, the use of 24-hour samples restricts statistical tests to a small portion of the total data base. Analyses of %TA within and between seasons are presented later and rely on a compromise between the 2 above noted extremes in observations. General activity measures presented here are well within the range of results from other studies. Comparisons 44 with other studies and exploration of the sensitivity of %TA to alternative definitions of bouts do not replace visual verification of pulse mode patterns and actual behavior, but do strengthen the assumption that bear activity has been adequately quantified by the methods I used. 45 LITERATURE CITED Altmann, J. 1974. Observational study of behavior: sampling methods. Behaviour 49:227-265. Amstrup, S. C., and J. Beecham. 1976. Activity patterns of radio-collared black bears in Idaho. J. Wildl. Manage. 40:340-348. Aune, K., and T. Stivers. 1983. Rocky Mountain front grizzly bear monitoring and investigation. Mont. Dep. Fish, Wildl. and Parks. Helena, Montana. 180 pp. Aune, K., and W. Kasworm. 1989. Final report: east front grizzly bear study. Mont. Dep. Fish, Wildl. and Parks. Helena, Montana. 332 pp. Ayers, L. E., L. S. Chow, and D. M. Graber. 1986. Black bear activity patterns and human induced modifications in Sequoia National Park. Int. Conf. Bear Res. and Manage. 6:151-1654. Bjarvall, A., and F. Sandegren. 1987. Early experiences with the first radio-marked brown bears in Sweden. Int. Conf. Bear Res. and Manage. 7:9-12. Cederlund, G. 1981. Daily and seasonal activity pattern of roe deer in a boreal habitat. Viltrevy 11:315-353. Cederlund, G., and P. A. Lemnell. 1980. Activity recording of radio-tagged animals. Biotelemetry Patient Monitg. 7:206-214. Clevenger, A. P., F. J. Purroy, and M. R. Pelton. 1990. Movement and activity patterns of a European brown bear in the Cantabrian Mountains, Spain. Int. Conf. Bear. Res. and Manage. 8:205-211. Garshelis, D. L., and M. R. Pelton. 1980. Activity of black bears in the Great Smoky Mountains National Park. J. Mammal. 61:8-19. 46 Garshelis, D. L., H. B. Quigley, C. R. Villarrubia, and M. R. Pelton. 1982. Assessment of telemetric motion sensors for studies of activity. Can. J. Zool. 60:1800-1805. Gebhard, J. G. 1982. Annual activities and behavior of a grizzly bear (Ursus arctos) family in northern Alaska. M.Sc. Thesis. Univ. Alaska, Fairbanks, Alaska. 218 pp. Grieg, M., and J. Bjerring. 1978. UBC GENLIN - a general least squares analysis of variance program. Computing Centre. Univ. of British Columbia, Vancouver, British Columbia. 48 pp. Harting, A. L. 1985. Relationships between activity patterns and foraging strategies of Yellowstone grizzly bears. M.Sc. Thesis, Mont. State Univ., Bozeman, Montana. 103 pp. Hechtel, J. L. 1985. Activity and food habits of barren-ground grizzly bears in Arctic Alaska. M.Sc. Thesis. Univ. Montana, Missoula, Montana. 74 pp. Hicks, C. R. 1982. Fundamental concepts in the design of experiments. 3rd edn. Holt, Rinehart and Winston, New York, New York. 425 pp. Jacobsen, N. K., and A. D. Wiggins. 1982. Temporal and procedural influences on activity estimated by time-sampling. J. Wildl. Manage. 46:313-324. Lindzey, F. G., and E. C. Meslow. 1977. Home range and habitat use by black bears in southwestern Washington. J. Wildl. Manage. 41:413-425. McLellan, B. N., and D. M. Shackleton. 1989. Grizzly bears and resource-extraction industries: habitat displacement in response to seismic exploration, timber harvesting and road maintenance. J. Appl. Ecol. 26:371-380. Phillips, M. K. 1987. Behavior and habitat use of grizzly bears in northeastern Alaska. Int. Conf. Bear Res. and Manage. 7:159-167. 47 Roth, H. U. 1983. Diel activity of a remnant population of European brown bears. Int. Conf. Bear Res. and Manage. 5:223-229. Roth, H. U., and D. Huber. 1986. Diel activity of brown bears in Plitvice Lakes National Park, Yugoslavia. Int. Conf. Bear Res. and Manage. 6:177-181. Schleyer, B. 0. 1983. Activity patterns of grizzly bears in the Yellowstone ecosystem and their reproductive behavior, predation and use of carrion. M.Sc. Thesis, Mont. State Univ., Bozeman, Montana. 13 0 pp. Sokal, R. R., and F. J. Rohlf. 1981. Biometry. 2nd edn. W. H. Freeman and Co. New York, New York. 859 pp. Smith, T. R. 1986. Activity and behavior of denned black bears in the lower Mississippi River valley. Int. Conf. Bear Res. and Manage. 6:137-143. Sorokin, P. A., and C. Q. Berger. 1939. Time-budgets of human behavior. Harvard University Press, Cambridge, Massachusetts. 204 pp. Watts, P. D., and C. Jonkel. 1988. Energetic cost of winter dormancy in grizzly bear. J. Wildl. Manage. 52:654-656. 48 CHAPTER 3: BOUT LENGTHS INTRODUCTION Time-energy budgets determine an organism's energetic status. Coarse-grained activity budgets which resolve only active and inactive behavioral states lack information concerning the energetic costs or gains of time spent in either behavior. Despite this limitation, broad expectations of how grizzly bears manage their time may be derived from theoretical considerations, and from empirical observations gained from interspecific studies. In this chapter, the emphasis is on the lengths of both active and inactive bouts. Although total time active will be dealt with explicitly in a later chapter, it is also an important determinant of bout lengths because changes in total time active translate into changes in active bout durations, in frequencies (and hence, changes in inactive bout lengths), or in both. Theoretical treatments of time allocation are rare (Bunnell and Gillingham 1985; Herbers 1981). However, energy acquisition is a fundamental component of time-budget theory (Herbers 1981). The total time bears spend foraging (searching and consuming food) likely depends on individual specific factors (weight, sex, age, reproductive status) that determine energetic requirements and constraints, and on environmental factors that limit the individual's options for resolving energy problems (see review in Bunnell and 49 Gillingham 1985). Foraging bout lengths may be viewed as a function of gut size, ingestion and digestion rates (Bunnell and Gillingham 1985). Generally, most of the active time that a mammal spends is devoted to foraging (Bunnell and Harestad 1989), while most nonforaging time is spent resting (Herbers 1981). Consequently, total time active should relate broadly to factors determining total time spent foraging, and both active and inactive bout lengths should relate to factors determining foraging bout lengths and foraging frequency. The importance of body weight is generally viewed from an interspecific perspective. As body weight increases, several physical relationships between an organism and its environment change, but at different rates (Demment 1983). Therefore, animals of different weights are required to respond to their environment in different ways (Demment 1983). Among species, one significance of being heavier is that requirements for total daily energetic intake increase but at a decreasing rate (Demment 1983) because energy requirements scale to metabolic weight (Kleiber 1975, p. 202). Gut capacity, however, increases near linearly with body weight (Demment 1983). These general relationships with body weight apply within species as well, although the actual rates may be somewhat different (Feldman and McMahon 1983; Demment 1983; Huesner 1982). The functional consequence of this relationship between gut size and metabolic requirement is that smaller individuals must 50 exhibit higher rates of passage than larger individuals to meet their energy requirements (Demment 1983). Within grizzly bears, weight is covariate with both sex and age. Adults are larger than subadults, males are larger than females. Based on absolute energetic requirements dictated by weight, and assuming an identical forage type across different classes of bears, adult males may be expected to spend more total time foraging than other sex-age classes. However, adult males may, for example, reduce their total foraging time a) by being less selective in forage components consumed (a strategy available to them due to the relationship between gut capacity and metabolic requirement), b) by expropriating the most profitable foraging patches for themselves (e.g., Egbert and Stokes 1976; Stonorov and Stokes 1972), or c) by foraging more intensively (Bunnell and Gillingham 1985). The other sex-age classes have additional energetic requirements to those determined by weight alone. Adult females must contend with the increased energetic and nutritive demands of gestation and lactation (Lloyd et al. 1978). Subadults have increased energetic and nutritive requirements for growth (Lloyd et al. 1978), and are potentially hampered by foraging inefficiencies due to inexperience (Bunnell and Gillingham 1985) or to their smaller size. Gebhard (1982) noted that of 2 grizzly yearlings in a family group that he observed throughout their active year, the smaller female yearling was less efficient than was her male sibling, and both 51 yearlings were less efficient than their mother when foraging for important but physically difficult to extract foods such as roots and ground squirrels. Predictions of total time active, therefore, must incorporate sex, age and reproductive status differences in energy requirements, foraging strategies and abilities. An individual bear's partitioning of total active time into a sequence of active bouts that alternate with inactive bouts, will be broadly a function of the rate at which the individual fills and empties its gut. Ingestion rates will depend on search time, on components of handling time such as pursuing and subduing prey, on bite rate, on bite size (the amount of food that can be taken in each bite as determined by forage characteristics), and also on physical mouth size which sets an absolute upper limit on bite size. Under the assumption that mouth size scales to body size in a fashion similar to gut capacity, different sized bears feeding on the same forage should be able to fill their gut at the same rate, provided that bite size is not limiting. However, for some food types such as berries that are distributed singly or in small clusters, bite size is essentially independent of the size of the bear, so small bears should fill their gut faster, unless larger bears can compensate by increasing their bite rate. For bears of different sizes on the same forage type, the significance of being smaller will be expressed in a requirement for more foraging bouts due to higher passage 52 rates needed to meet their metabolic requirements. The covariate nature of size and age indicates that smaller bears will have additional energetic requirements for growth that should intensify the differential in required foraging bouts. Smaller, younger bears may offset an increase in the number of foraging bouts required, by selectively foraging on higher quality components of the forage (a strategy available to smaller animals with absolutely lower energetic requirements), perhaps at the cost of longer foraging bouts. A similar requirement for more or longer foraging bouts per unit time is expected when adult females incur reproductive costs, especially lactation. The above arguments have been formulated under the restriction of identical forage types across sex-age and reproductive classes of bears. Within seasons this may be generally true. However, unlike many species (e.g., herbivores) that persistently consume similar forage species, bears exhibit dramatic shifts in food habits throughout the annual cycle (Bunnell and Hamilton 1983; Pearson 1975). Variation in food types is expected to influence both total time active and the partitioning of time into active and inactive bouts through searching and handling times. The potential for different sex-age or reproductive classes to increase or decrease forage selectivity is also likely constrained differentially among seasons. 53 Predicting total time active, through integration of metabolic requirements determined by weight, age, and reproductive status, with the influences of forage type, foraging strategy, physical ability, and the capacity of the gut in relation to metabolic requirements is a complex task. Many other factors (e.g., available daylight, mating activities, weather) likely exert important influences. Determining a priori how individuals partition total time active into discrete packages and intersperse these bouts with discrete packages of inactivity is perhaps impossible. My intent was not to generate a set of testable predictions through deduction, but rather to provide a theoretical framework from which expectations of differences in bout durations could be generated. Analyses between sex-age classes throughout the year may indicate general differences in how bears of different sexes and ages respond to their environment. If food type influenced bout lengths, this may be revealed through comparisons conducted among seasons. Under the assumption that within any season all bears foraged on similar foods, within season differences in bout lengths should reflect how sex-age or reproductive classes packaged foraging and resting to resolve their energetic requirements and constraints. If different food types influenced this packaging differently for specific classes of bears, this may be broadly indicated by a lack of consistent trends across seasons. 54 My specific objectives were to: 1) assess whether sex-age classes of bears differed in inactive and active bout durations on an annual basis; 2) assess whether inactive and active bout durations pooled for all bears varied as a function of season; and 3) test within seasons for differences in inactive and active bout durations among bears of different sex, age and reproductive status. METHODS Analyses of activity bout durations were conducted on chart recorded activity data of grizzly bears. Active bouts were defined as pulse mode patterns consistent with the active state that had a minimum duration of 1 min. Inactive bouts were defined as patterns consistent with the inactive state that had a minimum duration of 2.5 min. Only complete bouts of each type were used in the analyses. Active and inactive bouts in progress at the beginning or end of a recording session, or those interrupted by bouts assigned as unknown (due to missing data) were excluded. Separate, but identical analyses were conducted for active bouts and inactive bouts. Analyses Within Seasons For analyses conducted on each of the 4 seasonal data sets, linear models of sources of variation were derived under two approaches: 1) a nested-factorial model with sex and age classes as crossed main effects and individual bears 55 nested within sex-age class combinations, and 2) a fully hierarchical model with the highest level (population component) composed of 6 treatments (females with cubs, females with yearlings, females without offspring, adult males, subadult males, and subadult females), with individual bears nested within the treatments. Exploratory analyses of the data indicated that the distributions of both active and inactive bouts were skewed to the right and leptokurtotic (gl and g2 both positive; Sokal and Rohlf 1981, p. 117). Transformations to normality were attempted by using powers < 1 (Velleman and Hoaglin 1981, p. 48) and assessed with Lilliefors•s test (P = 0.05; Wilkinson 1989, p. 604). No suitable transformations were found and nonparametric alternatives were utilized. However, nonparametric techniques are generally awkward when applied to complex models (Conover 1971, p. 274) and may not lend themselves to such analyses. Nonparametric analogs to one-way analysis of variance (Kruskal-Wallis tests; Conover 1971, p. 256) were used to test each bout type within season, with individual bears as levels of the main effect of interest. To extend the set of conclusions drawn from Kruskal-Wallis tests, nonparametric multiple comparisons (Gibbons 1985, p. 181; Dunn 1964) which permit simultaneous statistical inferences were used. The advantage of this procedure was that individuals representing a level of an effect (e.g., a sex or age class) could be pooled and 56 compared to other pooled individuals representing another level of the effect. Contrasts were developed to test effects in accordance with the structure of the data outlined in the above models. Adjustment of the alpha level for each test of a contrast, in accordance with the number of contrasts tested within each seasonal data set, encouraged parsimony in the number of tests performed. Therefore, a subset of contrasts that permitted exploration of potential differences as outlined in the introduction was developed. An experimentwise error rate of 0.10 was maintained for each seasonal analysis. Annual and Among Seasons Analyses Comparisons of active and inactive bout lengths among sex-age classes on an annual basis were performed by pooling all data within each sex-age class by bout type. Differences among seasons were explored by pooling all data within each season by bout type. As with analyses conducted within seasons, the distributions of bout lengths were skewed to the right and leptokurtotic. Analyses did not extend beyond the exploratory data analysis procedure of notched box plots (Wilkinson 1988, p. 206; McGill et al. 1978) which provided comparisons of heuristic value. Box plots provide visual summaries of data with the ends of the box representing the quartiles and the line bisecting the box representing the median. Notched box plots enable comparisons between 2 sample medians at approximately 95% 57 confidence (McGill et al. 1978). However, the error rate is not adjusted for multiple comparisons (Velleman and Hoaglin 1981, p. 74). The practical consequence of this is that any 2 box plots with overlapping confidence intervals will never be significantly different if pairwise error rates are adjusted in order to maintain the experimentwise error rate at 0.05. However, 2 box plots with non-overlapping confidence intervals may not be different if such adjustments are made. RESULTS The use of complete bouts in these analyses probably biased results towards bouts of shorter duration. Longer bouts were more likely to be interrupted by missing data due to signal interference, to frequency drift, or to the bear moving behind obstacles. An expectation that a lower percentage of active bouts than inactive bouts would be complete due to the bears1 movements was only weakly expressed (77% of all active bouts were complete versus 79% of all inactive bouts). However, complete inactive bouts accounted for 70.3% of the total recorded inactive time, while complete active bouts accounted for only 54.1% of the total recorded active time. Percent of time active calculated from complete bout data tended to be higher than that calculated from all available data in the spring but lower in the remaining 3 seasons (Table 5). Table 5. Percent of time active by population component and season as determined by complete active and inactive bouts (columns under C) and as determined by all available data (columns under A). Samples sizes of complete bouts are in brackets. Population Early Berry Component Spring Summer Season Fall C A C A C A C A All Adult Females 51.0 (455) 47.7 47.6 (1639) 56. 6 54.1 (544) 65. 3 39.5 (555) 47.1 b with COYS 38.5 (119) 29.1 42.9 (475) 41. 0 — — 12.6 (8) 86.8 with yearlings — — 64.3 (96) 81. 5 67.3 (66) 76. 4 54.0 (188) 59.6 alone or with 2-year-olds 55.1 (336) 55.8 48.0 (1068) 56. 1 52.3 (478) 61. 9 33.5 (359) 31.6 Adult Males 49.2 (130) 33.2 45.9 (971) 45. 3 46.9 (496) 54. 9 42.3 (171) 37.2 Subadult females 52.4 (666) 53 .8 50.8 (846) 54. 1 54.2 (318) 73. 8 46.8 (542) 47.3 Subadult Males 60.5 (243) 51.9 51.7 (363) 68. 6 53.1 (716) 61. 0 32.5 (335) 35.2 All Bears 52.9 (1494) 49.4 48.5 (3819) 56. 0 52.1 (2074) 63. 1 40.8 (1603) 43.2 Some bears appear in more than 1 population component due to changes in age class and reproductive status over the sample years. COYS = cubs-of-the-year. 59 Annual and Seasonal Trends On an annual basis, sex-age classes did not differ with respect to median active (Fig. 9a), or median inactive (Fig. 9b) bout durations as indicated by overlapping notches on boxplots. For all bears pooled by season, median active bout durations showed a trend (Fig. 10a) of declining from spring through to the berry season and increasing in fall, with median durations in the berry season significantly less than in spring. Inactive bout durations displayed several significant changes (Fig. 10b) seasonally, although significance may not be stable to adjustments of pairwise alpha levels for multiple comparisons. Regardless, median durations of inactive bouts showed an opposite trend to active bouts; increasing from spring through to the berry season and declining in fall. Analyses Within Seasons With the exception of active bouts in fall, Kruskal-Wallis tests of both active and inactive bout lengths were significant (P < 0.05) in each season, indicating that at least one individual differed from the others. Contrasts were generated for each Kruskal-Wallis test (Tables 6 and 7) including fall active bouts (for the purpose of detecting consistent trends across seasons). Many more significant differences were found among contrasts testing inactive bouts than among active bouts. CO CO < scf-a •a III a < i x iu CO AC?-10 20 30 40 50 60 ACTIVE BOUT DURATION (min) CO CO < LU a < i x LU CO b s9-A9-SCT- •a ACf-—I 1 1 1 1 1 10 20 30 40 50 60 INACTIVE BOUT DURATION (min) Figure 9. Notched box plots of complete a) active, and b) inactive, bout durations pooled across seasons, for subadult females, adult females, subadult males, and adult males. Outside values were omitted to emphasize notches about the medians. For any 2 boxes with notches about the medians that do not overlap, the 2 medians are significantly different at approximately a 95% confidence level. I FA z BS O CO 3 ES CO {I SP 0 10 20 30 40 50 60 ACTIVE BOUT DURATION (min) z o CO < LU CO FA -BS ES b SP •{I 0 10 20 30 40 INACTIVE BOUT DURATION (min) i 1— 50 60 Figure 10. Notched box plots of complete a) active, and b) inactive, bout durations pooled across sex-age classes, for spring (SP), early summer (ES), berry season (BS), and fall (FA). Outside values were omitted to emphasize notches about the medians. For any 2 boxes with notches about the medians that do not overlap, the 2 medians are significantly different at approximately a 95% confidence level. Table 6. Tests of nonparametric multiple contrasts for durations of complete inactive bouts within seasons. Values are test statistics (value of the contrast divided by its standard error). Contrast Spring Early Summer Berry Season Fall 1) adult male - subadults a 3.12 -4.10a 0.26 -1.58 2) adult female - subadults a 3.25 -1.92 0.84 -0.31 3) adult M - adult F 1.10 -2.56a -0.48 -1.32 4) subadult M - subadult F a -3.51 1.76 -0.81 -0.21 5) adult F alone - adult F & coys 0.76 0.59 — — 6) adult F alone - adult F & yrls — 1.55 1.19 2 .10 7) adult F & coys - adult F & yrls — 1.17 — — 8) adult M - adult female alone 0.86 a -2.77 -0.76 -1.92 denotes significance at an experimentwise error rate = 0.10. Note: M = males, F = females, coys = cubs-of-the-year, yrls = yearlings. Table 7. Tests of nonparametric multiple contrasts for durations of complete active bouts within seasons. Values are test statistics (value of the contrast divided by its standard error). Contrast Spring Early Summer Berry Season Fall 1) adult M - subadults -1.57 2.05 0.96 -1.00 2) adult F - subadults 1.12 , a 3.05 -0.09 0.11 3) adult M - adult F -2.14 -0.67 0.76 -1.02 4) subadult M - subadult F a 2.41 -0.002 -0.76 -1.03 5) adult F alone - adult F & coys 0.41 -0.53 — 6) adult F alone - adult F & yrls — -1.02 -0.02 -0.48 7) adult F & coys - adult F & yrls — -0.72 — 8) adult M - adult F alone -2.17 -0.28 0.74 -0.87 a denotes significance at an experimentwise error rate = 0.10. Note: M = males, F = females, coys = cubs-of-the-year, yrls = yearlings. 64 a) Effects of Age Age effects were examined by testing pooled adult males and pooled adult females each against all subadults pooled together (contrasts 1 and 2 of Tables 6 and 7). In spring, inactive bouts were significantly longer for both male and female adults than subadults, while active bout lengths were not significantly different. In early summer, both male and female adults tended to have shorter inactive bouts than subadults but only the contrast testing adult males against subadults was significant. Active bouts tended to be longer for adults of either sex than for subadults but only the contrast testing adult females against subadults was significant. Contrasts testing adults against subadults were not significant for inactive or active bouts in the berry season or in fall. b) Effects of Sex The effects of sex were tested within adult and subadult age classes (contrasts 3 and 4 of Tables 6 and 7). Adult males and adult females differed significantly only in their early summer inactive bout lengths which were shorter for adult males. Within subadults, males had significantly shorter inactive bouts than females and significantly longer active bouts in spring. There were no other significant differences found. 65 c) Effects of Reproductive Status Not all contrasts involving adult females of different reproductive status could be tested in each season (contrasts 5, 6, and 7 of Tables 6 and 7). Contrasts testing differences between adult females of differing status were never significant. A general trend, however, was for females without offspring to have longer inactive bouts and shorter active bouts than females with cubs or yearlings. Results for females without offspring tested against adult males (contrast 8 of Tables 6 and 7) paralleled results for all adult females tested against adult males. However, most data for adult females were from those without offspring. DISCUSSION On an annual basis, sex-age classes did not differ significantly in median durations of either inactive or active bouts. However, contrasts testing sex-age classes within seasons revealed that some significant differences did exist in spring and early summer. When viewed annually, some within season differences among sex-age classes apparently cancelled each other out or else were masked by data pooling. Seasonal analyses indicated marked differences in inactive bout durations among seasons for all sex-age classes pooled. For active bout durations, only spring and the berry season were significantly different. The trend 66 from spring through to the berry season for active bout durations to decrease, was contrary to expectations derived from estimates of percent of time active for all available data by season (Table 5). An apparent bias for complete bouts to seriously underrepresent active bouts of long duration was most pronounced for the berry season and for early summer as indicated by the magnitude of the discrepancies between the values of %TA calculated from all available data for all bears (63.1% and 56.0% for the berry season and early summer, respectively) and from complete bouts (52.1% and 48.5% for the berry season and early summer, respectively). These discrepancies paralleled trends from spring through to the berry season for an increase in %TA (Table 5) and a greater percentage of total active time spent in bouts > 30 min duration. This was calculated from all available data and showed the percent of total active time in bouts > 30 min was 68.9% in spring, 78.9% in early summer, 89.3% in the berry season, and 67.6% in fall. The trend for median active bout durations to decrease from spring through to the berry season appeared to be a consequence of a bias against long active bouts due to periodic signal loss. This resulted from frequency drift or bears moving behind obstacles and caused many long active bouts to be classified as incomplete. In an observational study of a grizzly family in northern Alaska, Gebhard (1982) noted that grizzlies displayed longer active and inactive bouts in July and August than in the preceding months. From 67 September to denning he observed long active bouts interspersed with short inactive bouts, however, observations were hampered by increasing darkness. While seasons define shifts in foraging activities by bears, significant variation in food types within seasons may still exist among years or among individuals. Hechtel's (1985) data indicated that the quantities of various seasonal food types utilized by grizzly bears in arctic Alaska varied among years. However, annual variation cannot be separated from individual variation. McLellan (1989) found significant within-season differences in the amount of use different habitats received by sex, age and reproductive classes of grizzly bears in the Flathead. Indirectly, these differences in habitat use suggest differences in food types and (or) abundance. Differences in food types have been shown to affect bout lengths. Schleyer (1983) noted that grizzlies feeding on carcasses were significantly less active and exhibited more sporadic activity patterns than bears utilizing other foods. While differences were insignificant, Phillips (1987) also found grizzlies feeding on carcasses to have longer and more frequent rest periods (indicating shorter intervening active periods) than other bears. The availability of carrion is seasonal and greatest in the spring due to winter kill of ungulates, and again in the fall due to gut piles and crippled animals from hunting. 68 Simple conclusions concerning within-season sex and age effects on active and inactive bout durations are elusive for at least 2 reasons. First, effects tested in several contrasts were primarily or wholly composed of data from 1 individual (e.g., almost all data on subadult males in spring was from 1 bear). Consequently, individual variation confounds some interpretations. Second, complete active bouts are biased, although the influence of this bias on the analyses within seasons is difficult to assess. Available evidence suggests that for spring and early summer, the effects of age and sex on active and inactive bout lengths are confounded. Significant differences in inactive bout durations between sex-age classes in spring are difficult to explain simply in terms of interactions between food types and body size relations. Significantly longer inactive bouts by adult bears may indicate that relationships between gut capacity and metabolic requirements result in subadults foraging more frequently. Size related differences in access to carrion may also account for the longer inactive bouts by adults. However, similar differences did not occur in the fall in spite of the 2 seasons being similar in terms of primary forage items (Hedvsarum sulphuresence roots and carrion; McLellan 1989). Significant differences in spring bout lengths between subadult males and subadult females may reflect sex related size differences between males and females. Subadults males 69 likely do not have the same options available for reducing foraging times as do adult males (e.g., expropriating profitable foraging sites). Additionally, after leaving their dens in the spring, grizzly bears undergo a 2 - 3 week period of hypophagia (Nelson et al. 1983) before resuming normal ingestion rates. Some differences in the spring may be due to sampling bears in different physiological states. Differences between adults and subadults tended to reverse themselves in early summer as compared to spring. Both reduced carrion availability or the influence of mating activity could be invoked as a potential explanations. Mating activity occurs almost wholly within the early summer season and involves predominantly adults. Some observational accounts of courting bears (Phillips 1987; Stelmock and Dean 1986; Hechtel 1985) indicate that both total time active and the distribution of time into activity bouts are modified during courtship. Significantly shorter inactive bouts of adult males compared to adult females may reflect increased movements by males as they search for reproductively available females. A failure for increased movements by adult males to be apparent as significantly longer active bouts, could be due to their moving out of range of the recording unit. The berry season represents the period with the least variance in primary food types consumed and with the most restricted movements by bears. A lack of significant differences between sex-age classes of bears in this season 70 may be due to the relationship between rates of ingestion and excretion. While ingestion of berries is a slow process, transit times through the gut are short with many berries showing little indication of having been digested. Cessation of foraging bouts may not be due to gut fill. The hyperphagic behavior of bears in the berry season and the marked weight gains achieved, also indicate that bears are not merely responding to immediate metabolic requirements. Bout lengths are therefore uncoupled from predictions based on such requirements and from simple rules such as 'feed until full' (Bunnell and Gillingham 1985). 71 LITERATURE CITED Bunnell, F. L., and M. P. Gillingham. 1985. Foraging behavior: dynamics of dining out. In Bioenergetics of wild herbivores. Edited by R. j. Hudson and R. G. White. CRC Press, Boca Raton, Florida, pp. 53-79. Bunnell, F. L., and T. Hamilton. 1983. Forage digestibility and fitness in grizzly bears. Int. Conf. Bear Res. and Manage. 5:179-185. Bunnell F. L., and A. S. Harestad. 1989. Activity budgets and body weight in mammals: how sloppy can mammals be? Curr. Mammal. 2:245-305. Conover, W. J. 1971. Practical nonparametric statistics. John Wiley and Sons, Inc., New York, New York. 462 pp. Demment, M. W. 1983. Feeding ecology and the evolution of body size of baboons. Afr. J. Ecol. 21:219-233. Dunn, O. J. 1964. Multiple contrasts using rank sums. Technometrics 6:241-252. Egbert, A. L., and A. W. Stokes. 1976. The social behavior of brown bears on an Alaskan salmon stream. Int. Conf. Bear Res. and Manage. 3:41-56. Feldman, H. A., and T. A. McMahon. 1983. The 3/4 mass exponent for metabolism is not a statistical artifact. Resp. Physiol. 52:149-163. Gebhard, J. G. 1982. Annual activities and behavior of a grizzly bear (Ursus arctos) family in northern Alaska. M.Sc. Thesis. Univ. Alaska, Fairbanks, Alaska. 218 pp. Gibbons, J. D. 1985. Nonparametric methods for quantitative analysis. 2nd edn. American Sciences Press, Inc., Columbus, Ohio. 481 pp. Hechtel, J. L. 1985. Activity and food habits of barren-ground grizzly bears in Arctic Alaska. M.Sc. Thesis. Univ. Montana, Missoula, Montana. 74 pp. Herbers, J. M. 1981. Time resources and laziness in animals Oecologia. 49:252-262. Heusner, A. A. 1982. Energy metabolism and body size. I. Is the 0.75 mass exponent of Kleiber's equation a statistical artifact? Resp. Physiol. 48:1-12. Kleiber, M. 1975. The fire of life: an introduction to animal energetics. Krieger Pub. Co., Huntington, New York. 453 pp. Lloyd, L. E., B. E. McDonald, and E. W. Crampton. 1978. Fundamentals of nutrition. 2nd edn. W. H. Freeman and Company, San Francisco, California. 466 pp. McLellan, B. N. 1989. Effects of resource extraction industries on behavior and population dynamics of grizzly bears in the Flathead drainage, British Columbia and Montana. Ph.D. Thesis, Univ. of British Columbia, Vancouver, British Columbia. 115 pp. McGill, R., J. W. Tukey, and W. A. Larsen. 1978. Variations of box plots. The American Statistician 32:12-16. Nelson, R. A., G. E. Folk, E. W. Pfeiffer, J. J. Craighead, C. J. Jonkel, and D. L. Steiger. 1983. Behavior, biochemistry, and hibernation in black, grizzly, and polar bears. Int. Conf. Bear Res and Manage. 5:284-290 Pearson, A. M. 1975. The northern interior grizzly bear (Ursus arctos L). Can. Wildl. Serv. Rep. Ser. No. 34. 86 pp. Phillips, M. K. 1987. Behavior and habitat use of grizzly bears in northeastern Alaska. Int. Conf. Bear Res. and Manage. 7:159-167. 73 Schleyer, B. 0. 1983. Activity patterns of grizzly bears in the Yellowstone ecosystem and their reproductive behavior, predation and use of carrion. M.Sc. Thesis, Mont. State Univ., Bozeman, Montana. 130 pp. Sokal, R. R., and F. J. Rohlf. 1981. Biometry. 2nd edn. W. H. Freeman and Company, New York, New York. 859 pp. Stelmock, J. J., and F. C. Dean. 1986. Brown bear activity and habitat use, Denali National Park - 1980. Int. Conf. Bear Res. and Manage. 6:155-168. Stonorov D., and A. W. Stokes. 1972. Social behavior of the Alaska brown bear. Int. Conf. Bear Res. and Manage. 2:232-242. Velleman, P. F., and D. C. Hoaglin. 1981. Applications, basics, and computing of exploratory data analysis. Duxbury Press, Boston, Massachusetts. 354 pp. Wilkinson, L. 1989. Systat: the system for statistics. Systat, Inc., Evanston, Illinois. 822 pp. Wilkinson, L. 1988. Sygraph. Systat, Inc., Evanston, Illinois. 980 pp. 74 CHAPTER 4: ACTIVITY BUDGETS AND PATTERNS INTRODUCTION Major reductions in the grizzly bears' range over the last 150 years, and the restriction of current populations to relatively large wilderness areas, has indicated that grizzlies may require seclusion from humans in order to survive (Knight 1980). However, most occupied grizzly habitat lies within areas managed under integrated land management philosophies which pursue increased development, or else within parks that are under increasing recreational use by humans. Recent studies in the Flathead River drainage of southeastern British Columbia and adjacent areas of Montana have documented overt behavioral responses by grizzlies to human activities (McLellan and Shackleton 1989a, 1989b), and habitat loss due to avoidance of habitats close to roads built to support logging and petroleum exploration (McLellan and Shackleton 1988). However, negative demographic responses of the grizzly population to these human intrusions have not been demonstrated (McLellan 1989; McLellan and Shackleton 1988). The implication therefore, is that grizzlies can coexist with some level of human activity if the latter is properly controlled. This is supported by the fact that all known grizzly mortalities were due to human predation (McLellan and Shackleton 1988). A prima facie aphorism that grizzlies can not adapt to humans is slowly being replaced by the realization that the 75 problem with adaptability primarily lies with people, not with bears. In areas of high human use, mammals avoid harassment from humans by switching their activity to locations and times when human activity is minimal (Geist 1971). Spatial avoidance of humans by grizzly bears has been reported for populations other than the Flathead (e.g., Archibald et al. 1987; Mattson et al. 1987; Harding and Nagy 1980). However, Servheen (1981) noted that darkness may provide cover for bears and McLellan and Shackleton (1988) found that bears did use habitats near roads significantly more often during the night than during the day. Several recent studies of activity budgets and patterns of grizzly populations in North America and Europe have indicated that increased nocturnal activity by bears may be an adaptation to avoid human activity (Clevenger et al. 1990; Aune and Kasworm 1989; Bjarvall and Sandegren 1987; Roth and Huber 1986; Roth 1983) and may reduce the extent of habitat loss due to avoidance experienced by populations coexisting with humans (McLellan and Shackleton 1988). Activity budget data have been frequently suggested as a means of studying relationships between the environment (including human induced disturbances) and populations, classes of individuals (e.g., sex, age, reproductive or social classes), and individuals themselves (Roth 1983; Jacobsen and Wiggins 1982; Boy and Duncan 1979; Eberhardt 1977). Roth (1983) suggested that quantifying activity 76 patterns of a species under varying environmental conditions (including human intrusion) may permit some activity parameters to be used as indicators of stress placed on a particular population by its environment. However, to be of use, baseline information on activity parameters must account for variation due to factors such as sex, age, and reproductive status, and be indexed to habitat productivity, to the intensity and type of human intrusions, and to the population's demographic status. Frequently, such vital information is missing or is known only qualitatively. Previous studies of grizzly bear activity patterns and budgets have frequently been hampered by small numbers of individual bears, and relatively small sums of total time monitored. This has generally prevented exploration of sex and age effects, and detracted from examinations of seasonal effects due to confounding. Detailed, quantitative information on how much time bears allocate to activity (as opposed to inactivity) may reveal constraints placed on them by their environment, and allow identification of critical periods or essential environmental factors in their annual cycle. Comparisons among different sex and age classes can reveal critical periods in the life cycle of the species. This chapter addresses time allocation in free-ranging grizzly bears with respect to total time spent in the active state, and the distribution of active time over the 24-hour cycle. Specific objectives were to: 1) compare activity budgets among sex and age classes of grizzly bears both 77 within and among seasons; 2) compare distributions of activity over the diel cycle among sex and age classes of grizzly bears both within and among seasons; and 3) to qualitatively compare diel activity patterns and activity budgets for the Flathead grizzly population with other documented populations. METHODS Field procedures were as described in Chapter 1. Chart analysis followed the rules presented in Chapter 2. Individual recording sessions on bears varied greatly in length. Ideally, statistical comparisons of activity budgets and patterns over the diel cycle should be based on complete 24-hour samples of activity so as to encompass daylight and darkness and standardize the duration of time that a sample encompasses. However, complete 24-hour samples were limited in number (44 samples, or only 22% of the data base) and were poorly distributed across individuals and seasons. Time-sampling the charts could have resulted in several thousand "observations". While previous studies utilizing time-sampling of radio collar signals frequently have used individual time-samples as independent observations in analytical treatments of the data, the assumption of independence is not likely warranted. As a compromise between these 2 extremes in the concept of an observation, the 24-hour cycle was stratified into 4 mutually exclusive time periods. However, this did 78 not exclude some interdependencies existing within the data. For each time period sampled for each bear, percent of time active (%TA) was determined as total time in the active bout category divided by the sum of total time in both active and inactive categories. Many time periods were not sampled completely depending on when the recording was initiated or if the signal was lost. Any time period for which < 75% of the period's duration was recorded was rejected from analyses. Comparisons of Activity Patterns Activity patterns were examined with respect to the daily cycle of daylight and darkness. Stratification of the 24-hour cycle into 4 time periods (hereafter referred to as "diel periods") was based on sunrise and sunset. Morning and evening diel periods were each 3 h long and extended from 1.5 h before, to 1.5 h after, sunrise and sunset, respectively. Diurnal and nocturnal diel periods varied in length throughout the year. Comparisons of activity patterns (percent of time active in each diel period) among sex-age classes and seasons were conducted with a mixed model nested-factorial analyses of variance (Hicks 1982, p. 233). For both spring and fall, there were insufficient data to test adult males. The analysis tested the remaining 3 sex-age classes (adult females, subadult females, subadult males) as treatments of a main effect which was factorial with diel periods and seasons. Individual bears (3 adult 79 females, 3 subadult females, 2 subadult males) were nested within their respective sex-age class. For early summer and the berry season, data were sufficient to test all 4 sex-age classes. A mixed model nested-factorial analysis of variance tested seasons, sex, age, and diel period as crossed factors with individual bears (2 adult males, 3 adult females, 3 subadult males, 2 subadult females), nested within their respective sex-age class. Comparisons of Activity Budgets Activity budgets could not be assessed from the above analyses since the 4 time periods used were not of equal length and mean %TA for main effects did not reflect actual duration of time active. Conversion of %TA for each diel period observation to actual time active, or weighting by the proportion of a 24-hour cycle that the period represented, would have meant that different periods had different potential ranges over which the values of observations could vary. As an alternative, 4 new time periods ("quarter days"), each 6 h in duration (0500 - 1100, 1100 - 1700, 1700 - 2300, 2300 - 0500; MDT) were defined. Analyses were conducted as described above with diel periods replaced by a quarter day effect. All analyses of activity budgets and patterns had disproportionate numbers of individual bears within sex-age 80 classes and unequal numbers of observations on individuals. Analyses comparing activity patterns and budgets among all 4 seasons also suffered from missing cells (notably, subadult males in spring were represented by only 1 individual). Under these conditions the Type 1 error rate may be different than that intended, and can be either more or less conservative. Some effects within analyses also failed to meet the assumption of homogeneity of variance. Results, therefore, must be interpreted with caution. Analyses were performed on UBC GENLIN (Grieg and Bjerring 1978) with arcsine square root transformed values (Sokal and Rohlf 1981, p. 427) . Differences between means were investigated with Newman-Keuls range tests (Hicks 1982, p. 51). The alpha level for all tests was 0.05. All reported activity levels are back-transformed predicted means of %TA ± SEM rather than observed means. Due to the large number and complexity of some interactions, interpretations of interactions were made with respect to trends only. RESULTS Activity Patterns For activity patterns, the effects of interest in the analyses of variance are the diel period main effects and interactions including diel period. Comparisons over all 4 seasons (Table 8) indicated significant differences in bear activity levels among diel periods. However, first order interactions involving diel periods were significant for 81 Table 8. Mixed model nested-factorial Analysis of Variance of activity patterns for grizzly bears over seasons. SA Class represents those sex-age classes (subadult males, subadult females, adult females) for which there were sufficient data to test. Source of variation d.f. M.S. F Season 3 1. 228 12. 018 SA Class 2 0. 021 0. 024 Diel Period 3 10. 000 57. 228 Bear(SA Class) 5 0. 860 11. 474 Season x SA Class 6 0. 250 2. 451 Season x Diel Period 9 0. 248 3. 496 SA Class x Diel Period 6 0. 172 0. 982 Season x Bear(SA Class) 13 0. 102 1. 363 Season x SA Class x Diel Period 18 0. 087 1. 221 Diel Period x Bear(SA Class) 15 0. 175 2. 332 Season x Diel Period x Bear(SA Class) 34 0. 071 0. 947 Prob. <0.001* 0.976 <0.001c <0.001€ 0.083£ 0.004C 0.471c 0.176C 0.299 0.004 0.556 Residual Total 309 423 0.075 a Tested against Season x Bear(SA Class) Tested against Bear(SA Class) ^ Tested against Diel Period x Bear(SA Class) Tested against Season x Diel Period x Bear(SA Class) e Tested against Residual 82 both seasons and individual bears (Table 8). Pooled over all other factors, activity levels for the 4 diel periods indicated low nocturnal activity (16.0 ± 2.8%), similarly high activity levels for the morning and diurnal periods (68.7 ± 2.8% and 69.4 ± 3.1%, respectively), and a peak in activity in the evening period (80.3 ± 2.3%). Regardless of season, the lowest activity levels were for bears in nocturnal periods (Fig. 11). Within any season activity levels were the highest in the evening period (Fig. 11), however, among seasons this was not true. Large variation in activity levels was evident among seasons in the morning period. Within any individual, nocturnal periods pooled over seasons were always the least active. For 7 of the 8 individuals, the nocturnal activity levels were the lowest values overall (range: 6.8 ± 9.2% to 28.1 ± 12.3%). Four bears (1 subadult female, 1 subadult male, 2 adult females) displayed a bimodal activity pattern with activity peaks occurring in the morning and evening periods, the remaining 4 bears (1 adult female, 2 subadult females, 1 subadult male) exhibited an increasing activity level from the morning through to the evening period. Comparisons between early summer and the berry season, with sex and age factorial, also resulted in significant differences among diel periods (Table 9). However, a first order interaction between diel period and age, and a second order interaction involving season, sex and diel period were significant (Table 9). The pattern of activity across diel I morning diurnal evening nocturnal DIEL PERIOD Figure 11. First order interaction between seasons and diel periods from Analysis of Variance of grizzly bear activity patterns over seasons (Table 7). Plotted values are predicted cell means. CO 84 Table 9. Mixed model nested-factorial Analysis of Variance of activity patterns for sex-age classes of grizzly bears over early summer and the berry season. Source of variation d.f. M.S. F Season 1 0. 908 2. 548 Sex 1 0. 082 0. 126 Age Diel Period 1 0. ,130 0. 120 3 10. ,758 88. 970 Bear(Sex x Age) 6 0. ,654 8. 122 Season x Sex 1 0. ,394 1. 105 Season x Age 1 0. ,002 0. 005 Sex x Age 1 0. ,150 0. 229 Sex x Diel Period 3 0. .342 2. 830 Age x Diel Period 3 0. .645 5. 333 Season x Bear(Sex x Age) 6 0. .356 4. 429 Season x Sex x Age 1 0. .321 0. 900 Season x Diel Period 3 0. .032 0. 478 Sex x Age x Diel Period 3 0, .142 1. 176 Season x Sex x Diel Period 3 0. .237 3. 503 Season x Age x Diel Period 3 0, .021 0. 317 Season x Sex x Age x Diel Period 3 0. .014 0. 205 Diel Period x Bear(Sex x Age) 18 0, .121 1. 503 Season x Diel Period x Bear(Sex x Age) 17 0. .068 0. 842 Residual 313 0, .080 Prob. 0.162* 0.7357 0.671 <0.001C <0.001e 0.334a °'946K 0.649 0.068C 0.008C <0.001e 0.380* 0.702 0.347C 0.038d 0.813d 0.891d 0.0876 0.6436 Total 391 * Tested against Season x Bear(Sex x Age Class) Tested against Bear(Sex x Age Class) ^ Tested against Diel Period x Bear(Sex x Age Class) Tested against Season x Diel Period x Bear(Sex x Age) e Tested against Residual 85 periods pooled over all other factors indicated a bimodal distribution with activity levels of 19 ± 2.9%, and 71.6 ± 3.5% in the nocturnal and diurnal periods respectively, while activity levels in the morning and evening periods were 80.0 ± 2.8% and 84.7 ± 2.6% respectively. The age x diel period interaction (Fig. 12) indicated that subadults were more active in all diel periods except for the nocturnal period when they were less active than adults. Adults displayed a bimodal activity pattern across diel periods while subadults exhibited little difference in activity between the morning and diurnal periods and an activity peak in the evening period. Across seasons, sex classes showed little difference in activity in the nocturnal period (Fig. 13a and b). Respectively, sex classes tended to have higher activity levels over the remaining diel periods in the berry season than they did in the early summer. In both seasons, females exhibited a peak in activity in the evening period, males tended to be slightly more active in the morning than in the evening period. In early summer, male and female activity levels contrasted sharply between morning and diurnal periods. Activity Budgets Comparisons of activity budgets over all 4 seasons (Table 10) indicated significant differences in the main effects of seasons, quarter day periods, and individual bears. No differences in activity were apparent for the 100 Age Class subadults adults morning diurnal evening nocturnal DIEL PERIOD Figure 12. First order interaction between age classes and diel periods from Analysis of Variance of grizzly bear activity patterns for sex-age classes over early summer and the berry season (Table 8.) Plotted values are predicted cell means. Figure 13. Second order interaction between seasons, sex classes, and diel periods from Analysis of Variance of grizzly bear activity patterns for sex-age classes over early summer and the berry season (Table 8) . Sex x diel period interactions are plotted for a) early summer, and b) the berry season. Plotted values are predicted cell means. 88 Table 10. Mixed model nested-factorial Analysis of Variance of activity budgets for grizzly bears over seasons. SA Class represents those sex-age classes (subadult males, subadult females, adult females) for which there were sufficient data to test. Source of variation d.f. M.S. F Prob. Season 3 1. 021 20 .143 <0.001a SA Class 2 0. 008 0 .007 0.993 Quarter Day 3 8. 460 43 .202 <0.001C Bear(SA Class) 5 1. 131 16 .239 <o.ooie Season x SA Class 6 0. 127 2 .500 0.083a Season x Quarter Day 9 0. 282 4 .306 <0.001d SA Class x Quarter Day 6 0. 159 0 .812 0.577C Season x Bear(SA Class) 12 0. 051 0 .728 0.7246 Season x SA Class x J Quarter Day 18 0. 140 2 .137 0.027 Quarter Day x Bear(SA Class) 15 0. 196 2 .811 <0.001e Season x Quarter Day x Bear(SA Class) 35 0. 065 0 .939 0.5716 Residual 316 0. 070 Total 430 a Tested against Season x Bear(SA Class) Tested against Bear(SA Class) ^ Tested against Quarter Day x Bear(SA Class) Tested against Season x Quarter Day x Bear(SA Class) e Tested against Residual 89 sex-age class effect. However, all main effects were involved in significant interactions and could not be interpreted reliably. Seasonal activity levels were 41.0 ± 3.5% in the spring, 58.7 ± 2.1% in the early summer, 64.4 ± 2.5% in the berry season and 44.1 ± 3.2% in the fall. A first order interaction between seasons and quarter day periods, and a second order interaction between seasons, sex-age classes and quarter day periods were significant (Table 10). In spring and fall, activity levels pooled over sex-age classes were generally lower across quarter day periods than in early summer and the berry season (Fig. 14). The most notable exception was between 2300 - 0500 hours during which bears in the fall were more active than they were in other seasons. Bears were also more active across quarter day periods in the berry season than they were in the early summer, except between 2300 - 0500 hours (Fig. 14). The second order interaction was complex (Fig. 15a -d). With respect to seasonal activity budgets, sex-age classes tended to be less active across quarter day periods in the spring than in early summer or the berry season except between 1100 - 1700 hours. During this quarter day period, spring and early summer activity levels overlapped considerably, while only the low activity level of adult females in the berry season overlapped with spring activity levels. In the fall, bears were less active in the quarter day periods encompassing 0500 - 1100 and 1700 - 2300 hours, IOO-I 90-0500-1100 1100-1700 1700-2300 2300-0500 QUARTER DAY PERIOD Figure 14. First order interaction between seasons and quarter day periods from Analysis of Variance of grizzly bear activity budgets over seasons (Table 9). Plotted values are predicted cell means. vo o u. O 0600-1)00 1100-1700 fTOO-2300 23OO-0GOO QUARTER DAY PERIOD no tO M 70 to 60 40 30 20-10-0 0600-1100 1100-1700 1700-2300 2300-0600 OUARTER DAY PERIOD HX) (0 80 70 80 SO 40-30-20 10 b wo to BO 70 to 60 40 30-20-10 0 0600-1100 1100-1700 1700-2300 2300-0600 OUARTER DAY PERIOD 0600-1100 1100-1700 1700-2300 2300-0600 OUARTER DAY PERIOD Sex-Age Class subadult males subadult females adult females Figure 15. Second order interaction between seasons, SA classes, and quarter day periods from Analysis of Variance of grizzly bear activity budgets over seasons (Table 9). SA class x diel period interactions are plotted for a) spring, b) early summer, c) the berry season, and d) fall. Plotted values are predicted cell means. 92 but tended to be more active between 2300 - 0500 hours, than they were in early summer or the berry season. Between 1100 - 1700 hours, the activity level for subadult females in the fall exceeded early summer activity levels, and also exceeded the activity level for adult females in the berry season. Spring and fall activity levels overlapped considerably with the exception of the quarter day period ranging from 2300 - 0500 hours for which bears were consistently more active in the fall. Activity levels for sex-age classes in early summer and the berry season overlapped in all quarter day periods with subadults in early summer displaying much higher activity levels between 1100 - 1700 hours than in the berry season. Relationships between the sex-age classes varied among seasons with no consistent trend for one sex-age class to exhibit higher activity levels. Pooled over seasons and quarter day periods, activity levels for each sex-age class were similar (for subadult males 55.7 ± 2.8%, subadult females 56.0 ± 2.4%, adult females 54.6 ± 1.9%). However, pooled over seasons and quarter day periods, individuals varied greatly in activity (range: 34.1 ± 5.4% to 71.0 ± 3.4%). Comparisons between early summer and the berry season, with sex and age factorial, resulted in only the main effects of quarter day periods and individual bears been significant (Table 11). However, both these main effects, as well as age and season, appeared in significant first order interactions (Table 11). Only the main effect of sex 93 Table 11. Mixed model nested-factorial Analysis of Variance of activity budgets for sex-age classes of grizzly bears over early summer and the berry season. Source of variation d.f. M.S. F Season 1 0. 469 1. 822 Sex 1 0. 454 0. 767 Age 1 0. 638 1. 078 Quarter Day 3 8. 997 109. 790 Bear(Sex x Age) 6 0. 591 7. 593 Season x Sex 1 <0. 001 <0. 001 Season x Age 1 0. 006 0. 024 Sex x Age 1 0. 083 0. 141 Sex x Quarter Day 3 0. 192 2. 344 Age x Quarter Day 3 0. 822 10. 027 Season x Bear(Sex x Age) 6 0. 258 3. 307 Season x Sex x Age 1 0. 396 1. 537 Season x Quarter Day 3 0. 088 1. 584 Sex x Age x Quarter Day 3 0. 155 1. 895 Season x Sex x Quarter Day 3 0. 118 2. 127 Season x Age x Quarter Day 3 0. 106 1. 902 Season x Sex x Age x Quarter Day 3 0. 026 0. 470 Quarter Day x Bear(Sex x Age) 18 0. 082 1. 052 Season x Quarter Day x Bear(Sex x Age) 18 0. 056 0. 714 Residual 312 0. 078 Prob. 0.226a 0.4155* 0.339 <0.001C <o.ooie 0.981a 0.883a 0.720 0.107C <0.001C 0.0046 0.261a 0.228 0.167C 0.132d 0.166d 0.707d 0.401e 0.7976 Total 391 a Tested against Season x Bear(Sex x Age Class) Tested against Bear(Sex x Age Class) ^ Tested against Quarter Day x Bear(Sex x Age Class) Tested against Season x Quarter Day x Bear(Sex x Age) e Tested against Residual 94 could be interpreted and indicated no significant difference between males (56.1 ± 2.2%) and females (63.1 ± 2.0%). The first order interaction between age classes and quarter day periods suggested that subadults had higher activity levels than adults in all quarter day periods except between 2300 - 0500 (Fig. 16). Pooled over quarter day periods and seasons, adults were active 56.6 ± 1.8% while subadults were active 64.9 ± 2.3%. The first order interaction between seasons and individuals revealed that 4 individuals (1 adult female, 1 adult male, 2 subadult females) exhibited a distinct increase (range of absolute differences (RAD): 6.8 - 36.1%) in activity from early summer to the berry season, 3 individuals (1 adult female, 2 subadult males) demonstrated little difference (RAD: 0.6 - 2.3%) in activity, and 3 individuals (1 adult female, 1 adult male, 1 subadult male) showed a distinct decrease in activity (RAD: 5.3 - 16.3%). Seasonal activity levels were 56.5 ± 2.0% for early summer and 63.8 ± 2.2% for the berry season. DISCUSSION Within a species, an individual's total time active and the distribution of activity over the diel cycle are determined by individual specific (e.g., weight, age, sex, physiological condition) and environmental (e.g., predation, human activity, thermal stress, food type and abundance, available daylight) factors (Bunnell and Gillingham 1985). I Figure 16. First order interaction between age classes and quarter day periods from Analysis of Variance of grizzly bear activity budgets for sex-age classes over early summer and the berry season (Table 10). Plotted values are predicted cell means. 96 Individual factors assessed in this study were restricted to sex and to broad age classes, although weight is covariate with both these factors. Environmental factors explicitly dealt with were restricted to seasons (and hence, food type and abundance) and the daily solar cycle. Thermal stress and human activity tend to be covariate with seasons. Changes from the predominate spring and fall foods (Hedysarum sulphuresence and carrion) to grasses, forbs and berries, were accompanied by a change in thermal regime. Similarly, use of unroaded high elevation berry fields resulted in a switch to berries being accompanied by a change in the level of human disturbance. Activity Patterns Most research supports a general description of grizzly bears as nocturnal or crepuscular, with activity peaks in the mornings and evenings, and often higher activity levels during the night than during the day (Clevenger et al. 1990; Gunther 1990; Phillips 1987; Roth and Huber 1987; Roth and Huber 1986; Aune and Stivers 1985; Harting 1985; Aune et al. 1984; Roth 1983; Schleyer 1983). Substantial variation in activity patterns has been reported among and within individuals (Clevenger et al. 1990; Aune and Kasworm 1989; Roth and Huber 1986; Harting 1985; Hechtel 1985; Roth 1983; Schleyer 1983; Sizemore 1980), and among seasons, where seasonal information is available (Harting 1985; Schleyer 1983; Gebhard 1982; Sizemore 1980). Some authors have 97 related observed diurnal activity to reduced levels of human intrusion or to use of habitats that have high cover value (Clevenger et al. 1990; Aune and Kasworm 1989; Bjarvall and Sandegren 1987). Many have suggested that observed nocturnal or crepuscular activity is a behavioral adaptation to avoid humans. Unlike most other studies of grizzly bear activity patterns, bears in the Flathead were primarily active during daylight hours. While activity levels expressed in diel periods frequently did not suggest strong bimodality, plots of hourly activity levels utilizing all data (Figs. 17 - 20) generally did display a bimodal activity pattern with peaks in the morning and evening. Morning activity peaks, however, were often not centered about sunrise but rather occurred 1 or more h after sunrise. Evening activity peaks showed a closer relationship to sunset. Reported influences of seasons on diel activity patterns of bears are ambiguous. Schleyer (1983) and Harting (1985) observed Yellowstone grizzlies to be more diurnal in spring and fall than in summer. Sizemore (1980) found grizzlies in the southern Flathead area to be more nocturnal in the spring-summer season, but found no differences in diurnal and nocturnal activity in the summer-fall season. Some studies on European brown bears (Clevenger et al. 1990; Roth and Huber 1986) have reported diurnal activity by individual bears in fall and winter, but individuals did not show consistent patterns across years. KK) n 4 6 8 7 8 9 10 11 12 13 HIS 16 17 18 19 20 2122 23 24 ^ TTME OF DAY ^ 1 2 3 4 B 6 7 8 9 10 11 12 13 14 1B 16 17 18 19 20 21 22 23 24 A. TIME OF DAY A. 100-90-80-G TO" < LU 60 2 fe... LU O 30 E 20 10-0 tm ii II IP 11 I Illl 1 2 3 4 6 8 • i iii 1 iliPIII III! Mill II 7 8 9 10 11 12 13 14 IB 18 17 18 19 20 2122 23 24 TIME 6F DAY ^ Figure 17. Percent of time active by hour of the day in spring for a) subadult females, b) adult females, and c) subadult males. Arrows indicate approximate times of sunrise and sunset. Plots were composed from all spring data collected. For hours of the day, 1 = 0000-0100; 2 = 0100-0200 etc. vo 00 1 2 3 4 6 8 7 8 9 to 11 12 13 14 18 18 17 18 t9 20 21 22 23 24 ^ TIME OF DAY ^ TOO 80-80-R 70-Uj 80 E -o g 40 O 30 20 10 0 I I 1 II I il ill ill II1 II 1 III ilii 11 nil •ft I 11 il in hi in 1 2 3 4 B 8 7 8 9 10 11 T2 13 14 16 18 17 18 19 20 2122 23 24 ^ TIME OF DAY ^ Figure 18. Percent of time active by X females, b) adult females, indicate approximate times from all early summer data 2 = 0100-0200 etc. 100 90 -| LU 80 < Ul 80 2 u. O 60 40 30-20-10 0 1 mt m J i IH 1 HI §§11 ii 1 llllll! W ill I I II 1 2 3 4 6 8 7 8 9 10 11 12 13 14 18 16 17 18 19 20 21 22 23 24 ^ TIME OF 0AY ^ TOO 90 < Ul 80 5 u. O 60 H 40 30-20-10 0 d II il ^ ^ ^ ^ ^ in pin •VI 11 Hi _ II ii HI iiiiiiill 2 3 4 6 8 7 8 9 10 11 12 13 14 15 16 17 18 19 20 2122 23 24 ^ TIME OF DAY ^ Lour of the day in early summer for a) subadult c) subadult males, and d) adult males. Arrows of sunrise and sunset. Plots were composed collected. For hours of the day, 1 = 0000-0100; 100-90-UJ B0-F 70-< LU 80-2 E -o fc- *0 S 20-10-n ' i 1 1 1 1 l III 0 -J-J 100-. 90-U, BO S' ro-< LU 60-5 E -o y 30-E 20-10- I o-U r'T*'v'i"f"r"Y"r'i t" 1 23468769 10 j 11 12 13 14 16 18 17 1 ME OF DAY M 3 192 I I 0 212 I Hi 1 22324 ll 1 2 3 4 5 8 7 8 9 10 11 12 13 14 16 18 17 18 19 20 21 22 23 24 A TIME OF DAY & Figure 19. Percent of time active by hour of subadult females, b) adult female Arrows indicate approximate times composed from all berry season da 1 = 0000-0100; 2 = 0100-0200 etc. 100 90-Uj 80-> < LU 60 2 o |_ 40 S 20 10 0 ill II iiiiiiiiiiiiii lllllllllllllllll .1 iiilll IP iiiiiniiiiini Y- Y. V. '////, V. Y. II -•ii Ml 1 2 3 4 5 6 7 8 9 10 11 12 13 14 IS 16 17 18 19 20 21 22 23 24 A TIME OF DAY A 100 80 Uj 8° > 0 70 < UJ 60-| o |- 40 LU O 30-S 20-10-IP ii||i| I •••••••••1 i^ WPP mil ill 1 2 3 4 5 8 7 8 9 10 11 12 13 14 16 18 17 18 19 20 21 22 23 24 A TIME OF DAY A the day in the berry season for a) , c) subadult males, and d) adult males. of sunrise and sunset. Plots were a collected. For hours of the day, too 80-70-»0 O 20- 11 m IIII il II Jill" in •in ii pEL III HI II II II in II i 2 3 4 6 6 7 8 9 10 11 12 13 14 15 18 T7 18 19 20 21222324 A TIME OF DAY A 100-i 2 3 4 5 8 7 8 9 10 11 12 1314 15 18 17 18 18 20 21222324 TIME OF DAY A T LU 100 80 G 70 < Ul 80 o 8 60-40-30 20-10 0 L •••lllipillMi il i Hi 11 J 2 3 4 5 8 7 8 9 10 11 12 13 14 16 16 17 18 19 20 21222324 A TIME OF DAY A Figure 20. Percent of time active by hour of the day in fall for a) subadult females, b) adult females, and c) subadult males. Arrows indicate approximate times of sunrise and sunset. Plots were composed from all fall data collected. For hours of the day, 1 = 0000-0100; 2 = 0100-0200 etc. 102 In spite of spring and fall being similar in primary foods utilized by bears, I found grizzlies in the Flathead to exhibit the lowest nocturnal activity levels in the spring and the highest nocturnal activity in the fall. Available daylight was shorter in the fall and may contribute to this result. The availability of carrion in the fall is primarily a function of sport hunting (gut piles from hunter kills and crippled animals) while spring carrion is from winter kills and winter weakened animals. Throughout most of the year, human intrusion is predictable and centered along the network of roads, while in the fall, bears have a much higher probability of encountering hunters away from roads. The greater activity by bears during darkness in the fall may be an avoidance reaction to hunter activity, but may also relate to bears searching for remains of hunter kills at night when hunting activity is low. McLellan and Shackleton (1988) found that grizzly bears used areas close (0 - 250 m) to roads significantly less than expected in the Flathead and that this represented an 8.7% loss of available habitat. They also found that bears used areas near roads more often at night than during the day. Habitats close to roads (e.g., riparian areas) tended to be of high value to bears in spring, early summer, and fall. This study, however, indicates that bears have not made a significant shift in their activity patterns to enable them to exploit habitats near roads and suggests that 103 adequate amounts of habitat further away from roads are available to support current population levels. Several authors have suggested that grizzlies are more active during daylight hours in environments with low human intrusion (Aune and Stivers 1989; Bjarvall and Sandegren 1987; Roth and Huber 1986; Roth 1983). Strong evidence indicating high use of daylight hours as normal for grizzlies has generally been lacking. Results of this study support a preference for diurnal activity in an environment where human intrusion is highly localized around roads and hence, is predictable in space. Preferences for daylight may be related to increased opportunities for selective foraging (Bunnell and Gillingham 1985), or to increased searching efficiency. Currently, it is unclear if grizzly populations restricted to nocturnal foraging exploit resources less effectively than they would with diurnal foraging patterns. Activity Budgets Comparing seasonal activity levels found in this study with other research on grizzly bear activity budgets is difficult due to differing definitions of seasons across studies. These differences, in part, reflect geographical variation in food resources, plant phenological development, and duration of the denning period. For the Yellowstone grizzly population, Schleyer (1983) found activity levels to be the lowest (approx. 2.4 - 4.8 hours of activity per 24-104 hour period, or 10 - 20% TA) in post- and pre-denning months (March, and September and October). Activity levels were high in April, June, July, and August (approx. 8.0 - 12.0 hours of activity per 24-hour period, or 33 - 50% TA). Grizzlies were found to be the most active in May, but data were not thought to be representative. Harting (1985) also studied the activity patterns of Yellowstone grizzlies and also found activity levels to be high in the summer months (probability of activity: 0.64 - 0.71) and lower in September (0.52). In contrast to Schleyer's (1983) results, bears were least active in May (0.52). In the southern Flathead region, Sizemore (1980) reported grizzly bear activity levels to be significantly lower in the spring-summer period (den emergence to July 31) than in the summer-fall period (August 1 to den entry). For northern Alaskan grizzly populations, Gebhard (1982) found the activity level of a family unit to be 58% during spring, early summer and late summer, 71% in early fall, and 83% in late fall. Hechtel (1985), however, did not find a seasonal trend for increased activity in the fall by this same family unit in the following year. Information on monthly or seasonal activity levels of European brown bears is limited. Roth (1983) found brown bears in northern Italy to be active 45 - 60% of the time in summer and fall. While data did not permit separation of seasonal trends from individual variability, Roth (1983) noted that the data weakly suggested high activity levels in 105 summer and somewhat lower activity levels in fall. A subadult female brown bear in Yugoslavia spent about 60% of the time in activity from July to October dropping to about 40% of the time in November and December (Roth and Huber 1986). From a larger sample of brown bears in Yugoslavia, Roth and Huber (1987) found activity levels to be highest in May (64%) and September (58%) with high activity levels maintained during the intervening months. Data for an adult male brown bear in Spain (Clevenger et al. 1990) showed activity levels to be lowest (31%) in the post-denning period (den emergence to May 15), highest (43%) in the breeding season (May 16 to August 31), and slightly lower (39%) in fall/winter (September 1 to den entry). Most data thus support a seasonal trend in grizzly bear activity budgets, with post- and pre-denning periods being the least active. This trend is in accordance with the bears' annual physiological phases proposed by Nelson et al. (1983). The post-denning hypophagic period is characterized by a persistence in the biochemical changes associated with denning for up to 3 weeks after den emergence. Bears then enter a period of normal activity, followed by hyperphagic activity. The timing of hyperphagic behavior can be expected to vary between geographic areas relative to the availability of abundant, high quality food resources. Assessing the extent of the hyperphagic period from activity budget data may be difficult. The low activity levels of grizzlies in fall observed by Schleyer (1983), were 106 attributed to them feeding on ungulates. Bears were found to be significantly less active when utilizing carcasses than when using other foods (Schleyer 1983), however, ingestion rates by bears are high when feeding on carcasses and it is unclear if there was a reduction in energetic intake associated with the lower activity levels. In this study, expected trends in seasonal activity levels were for spring and fall to be the least active seasons due to post-denning hypophagic behavior, and the high use of carrion in both seasons. Higher activity levels were expected in early summer as bears entered the normal activity phase. The berry season represented a period of abundant, high quality food and resultant hyperphagic behavior coupled with long handling times was expected to increase activity levels. Expected trends in activity levels over seasons were met for analyses both over all 4 seasons, and over early summer and the berry season. Higher activity in early summer and the berry season, as compared to spring and fall, appeared to be a consistent trend over sex-age classes and individuals. An increase in activity levels from early summer to the berry season was not consistent over sex-age classes or individuals. At the level of sex-age classes, only subadult females and adult males showed the predicted increase in activity. However, the interaction between individuals and seasons in the analysis over early summer and the berry season, indicated that while individual subadult females responded similarly, 107 adult males did not. Male #25's activity level increased from 36.9% in early summer to 73.0% in the berry season, while male #65's activity level decreased from 41.9% in early summer to 25.6% in the berry season. This contrast in activity levels was accompanied by individual differences in foraging strategies, as male #65 utilized riparian river bottoms in the berry season rather than high elevation berry fields like other bears. This suggests that size or frequency dependent differences in foraging strategies exist and as a consequence, the importance of the berry resource to individuals is not relatively the same among sex-age classes. Since subadult females are the most social and physically subordinate class of bears, and are likely poor competitors for food resources, loss of a large, abundant, temporal resource that is not defendable by dominant individuals, may have the greatest effect on subadult females. Pooled over early summer and the berry season, females tended to be more active than males, and subadults tended to be more active than adults. In their review of mammalian activity budgets, Bunnell and Harestad (1989) found a general tendency across species for adult males to be less active than adult females, even when reproductively active males and pregnant or lactating females were excluded from the analyses. Size-related differences in selectivity, and in the case of large herbivores, larger bite size permitting more rapid ingestion rates, were suggested as possible 108 contributing factors to this result. In the case of bears, size-related differences in access to high quality food resources may be more important in explaining both differences between males and females, and between adults and subadults. Increased energetic requirements for growth, coupled with potential foraging inefficiencies (Bunnell and Gillingham 1985) by subadults may also be important. Interestingly, among superconcentrate (e.g., mast) feeders, Bunnell and Harestad (1989) concluded that time advantages accruing to males due to their larger size would be lost if they could not secure profitable patches for themselves. The high activity level of bear #25 in the berry season may be consistent with this observation if bears foraging in the large berry fields of the Flathead Valley have equal access to the same distribution of food patch quality. While results for sex and age classes over the early summer and the berry season met expectations, it appeared to be primarily dependent on the influence of low activity levels by adult males in the early summer and on the high activity level of subadult females in the berry season. Activity levels pooled over all 4 seasons showed no trends attributable to sex-age class effects. Over their entire active portion of the year, adult females and subadults of both sexes were active about the same amount of time. 109 LITERATURE CITED Archibald, W. R., R. Ellis, and A. N. Hamilton. 1987. Responses of grizzly bears to logging truck traffic in the Kimsquit River valley, British Columbia. Int. Conf. Bear Res. and Manage. 7:251-257. Aune, K., T. Stivers, and M. Madel. 1984. Rocky Mountain front grizzly bear monitoring and investigation. Mont. Dep. Fish, Wildl. and Parks. Helena, Montana. 239 pp. Aune, K., and T. Stivers. 1985. Ecological studies of the grizzly bear in the Pine Butte Preserve. Mont. Dep. Fish, Wildl. and Parks. Helena, Montana. 154 pp. Aune, K., and W. Kasworm. 1989. Final report: east front grizzly bear study. Mont. Dep. Fish, Wildl. and Parks. Helena, Montana. 332 pp. Bjarvall, A., and F. Sandegren. 1987. Early experiences with the first radio-marked brown bears in Sweden. Int. Conf. Bear Res. and Manage. 7:9-12. Boy, V., and P. Duncan. 1979. Time-budgets of Camargue horses. 1. Developmental changes in the time-budgets of foals. Behaviour 71:187-202. Bunnell, F. L., and M. P. Gillingham. 1985. Foraging behavior: dynamics of dining out. In Bioenergetics of wild herbivores. Edited by. R. J. Hudson and R. G. White. CRC Press, Boca Raton, Florida, pp. 53-79. Bunnell F. L., and A. S. Harestad. 1989. Activity budgets and body weight in mammals: how sloppy can mammals be? Curr. Mammal. 2:245-305. Clevenger, A. P., F. J. Purroy, and M. R. Pelton. 1990. Movement and activity patterns of a European brown bear in the Cantabrian Mountains, Spain. Int. Conf. Bear. Res. and Manage. 8:205-211. Eberhardt, L. L. 1977. "Optimal" management policies for marine mammals. Wildl. Soc. Bull. 5:162-169. 110 Gebhard, J. G. 1982. Annual activities and behavior of a grizzly bear (Ursus arctos) family in northern Alaska. M.Sc. Thesis. Univ. Alaska, Fairbanks, Alaska. 218 pp. Geist, V. 1971. Bighorn sheep biology. Wildl. Soc. News. 136:61. Grieg, M., and J. Bjerring. 1978. UBC GENLIN - a general least squares analysis of variance program. Computing Centre. Univ. of British Columbia, Vancouver, British Columbia. 48 pp. Gunther, K. A. 1990. Visitor impact on grizzly bear activity in Pelican Valley, Yellowstone National Park. Int. Conf. Bear Res. and Manage. 8:73-78. Harding, L., and J. A. Nagy. 1980. Responses of grizzly bears to hydrocarbon exploration on Richards Island, Northwest Territories, Canada. Int. Conf. Bear Res. and Manage. 4:277-280. Harting, A. L. 1985. Relationships between activity patterns and foraging strategies of Yellowstone grizzly bears. M.Sc. Thesis, Mont. State Univ., Bozeman, Montana. 103 pp. Hechtel, J. L. 1985. Activity and food habits of barren-ground grizzly bears in Arctic Alaska. M.Sc. Thesis. Univ. Montana, Missoula, Montana. 74 pp. Hicks, C. R. 1982. Fundamental concepts in the design of experiments. 3rd edn. Holt, Rinehart and Winston, New York, New York. 425 pp. Jacobsen, N. K., and A. D. Wiggins. 1982. Temporal and procedural influences on activity estimated by time-sampling. J. Wildl. Manage. 46:313-324. Knight, R. R. 1980. Biological considerations in the delineation of critical habitat. Int. Conf. Bear Res. and Manage. 4:1-3. Ill 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. Int. Conf. Bear Res. and Manage. 7:259-273. McLellan, B. N. 1989. Dynamics of a grizzly bear population during a period of industrial resource extraction. III. Natality and rate of increase. Can. J. Zool. 67:1865-1868. McLellan, B. N., and D. M. Shackleton. 1988. Grizzly bears and resource-extraction industries: effects of roads on behaviour, habitat use and demography. J. Appl. Ecol. 25:451-460. McLellan, B. N., and D. M. Shackleton. 1989a. Grizzly bears and resource-extraction industries: habitat displacement in response to seismic exploration, timber harvesting and road maintenance. J. Appl. Ecol. 26:371-380. McLellan, B. N., and D. M. Shackleton. 1989b. Immediate reactions of grizzly bears to human activities. Wildl. Soc. Bull. 17:269-274. Nelson, R. A., G. E. Folk, E. w. Pfeiffer, J. J. Craighead, C. J. Jonkel, and D. L. Steiger. 1983. Behavior, biochemistry, and hibernation in black, grizzly, and polar bears. Int. Conf. Bear Res and Manage. 5:284-290. Phillips, M. K. 1987. Behavior and habitat use of grizzly bears in northeastern Alaska. Int. Conf. Bear Res. and Manage. 7:159-167. Roth, H. U. 1983. Diel activity of a remnant population of European brown bears. Int. Conf. Bear Res. and Manage. 5:223-229. Roth, H. U., and D. Huber. 1986. Diel activity of brown bears in Plitvice Lakes National Park, Yugoslavia. Int. Conf. Bear Res. and Manage. 6:177-181. 112 Roth, H. U., and D. Huber. 1987. Patterns of amount of activity of brown bears in Yugoslavia. Abstr. from Third Congr. of the Croatian Biol. Soc, Mali Losinj, Yugoslavia. Schleyer, B. 0. 1983. Activity patterns of grizzly bears in the Yellowstone ecosystem and their reproductive behavior, predation and use of carrion. M.Sc. Thesis, Mont. State Univ., Bozeman, Montana. 130 pp. Servheen, C. 1981. Grizzly bear ecology and management in the Mission Mountains, Montana. Ph.D. Thesis, Univ. of Montana, Missoula, Montana. 139 pp. Sizemore, D. L. 1980. Foraging strategies of the grizzly bear as related to its ecological energetics. M.Sc. Thesis, Univ. of Montana, Missoula, Montana. 67 pp. Sokal, R. R., and F. J. Rohlf. 1981. Biometry. 2nd edn. W. H. Freeman and Co., New York, New York. 859 pp. CHAPTER 5: OVERALL CONCLUSIONS Human activities can affect grizzly bears in many ways. The potential negative consequences of human activities on grizzlies can be classified broadly into direct and indirect impacts. Direct impacts tend to be immediate, extreme, and result in bear mortalities from hunting, poaching, problem bear control, and accidents along roadways. Indirect impacts are associated with environmental modifications that ultimately express themselves as changes in the behavior, in the energy status of the population, or in both. Destruction or alteration of important habitats are frequent consequences of human settlement, agriculture, and resource extraction. Disturbance from human intrusion may result in costly flight responses by bears, result in bears avoiding habitats close to areas of human activity, or cause bears to alter their daily rhythms of activity and inactivity. If severe enough, indirect impacts may lead to declines in population size and productivity. The ultimate yardstick by which direct and indirect impacts are measured is population demography. However, determining a population's status is time consuming, expensive, and difficult. Frequently, population status is unknown prior to opening an area for resource development, and populations can undergo large declines before it is demonstratively apparent that declines are occurring. Recovery is slow, if not impossible. Conservation of species like the grizzly, which have low reproductive 114 potential, will depend on our ability to assess the occurrence and probable extent of population declines early, and to modify human activities if necessary. Presently, activity parameters are of limited utility in assessing indirect impacts of human activities on bears. This study indicates that bears prefer to use daylight, and apparently will continue to do so in areas where human activity is predictable and localized. The dense cover of the low elevation portions of the Flathead also likely had a mitigating effect. However, nocturnal activity by bears may permit the exploitation of habitats unavailable during daylight and hence, reduce some indirect impacts of human activities. Modifications in behavior of a population are not in themselves indications of reduced energetic status or population decline. The similarity in activity budgets between grizzlies in the Flathead and other populations that follow a nocturnal activity schedule would suggest little difference in the ability of bears to exploit resources in daylight or darkness. However, such comparisons are confounded by differences in habitat quality and predominate food types, and by a lack of vital information (e.g. age of first reproduction, litter size, interbirth intervals, population density) that integrates the animal with its environment. Appendix 1. Data totals (sum of active and inactive bout durations) by bear, year, and season, in decimal hours. Bear Age class, sex, and No. reproductive status soring early summer berrv season 1984 46 subadult male 47 subadult male 48 subadult female 37.60 45.98 67.03 60.72 1985 25 adult male — — 142.67 3 6 subadult female — 15.10 37.00 38 subadult female 46 subadult male 6.45 19.88 47 subadult male — 4.52 48 subadult female 80.82 49.99 8.72 1986 18 36 38 46 48 adult female, yearlings adult female, C0YSa subadult female adult male subadult female 4.13 43.87 41.48 21.90 17.95 10.42 24.25 Appendix 1. Continued. Bear Age class, sex, and No. reproductive status spring 1987 25 adult male 36 adult female, yearlings 38 adult female, alone 45 adult female, alone 46 adult male 48 adult female, alone 58 adult female, alone 63 subadult male 64 subadult male 65 adult male 67 subadult female 68 subadult female 69 subadult male 1988 25 adult male b 27.40 36 adult female, alone 118.93 38 adult female, COYS 66.92 48 adult female, alone 35.363 subadult male 83.60 65 adult male 25.23 67 subadult female 53.88 68 subadult female 95.77 aCOYS = cubs-of-the-year. accompanied by 2-year-olds in spring. early summer berry season fall 60.65 135.67 151.32 38.93 125.75 31.38 106.68 121.72 146.15 90.85 44.80 38.82 150.47 93.45 140.65 52.43 93.37 19.23 201.53 167.25 100.28 100.83 123.12 62.57 53.87 57.28 85.55 50. 62 44.57 56.43 4.55 46.57 84 .73 68.03 52.60 197.48 128.15 3.02 98. 60 62.20 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0098597/manifest

Comment

Related Items