Northern Prairie Wildlife Research Center
Garshelis (1994) subsequently showed how the Lincoln-Peterson estimates could have been influenced if (1) age class and sex-related differences existed in capture rates and (2) the composition of the population changed over time. According to Garshelis (1994), the Lincoln-Peterson estimates were severely biased and had so much influence on results that trends in bear numbers could have been misinterpreted.
Unlike previous analyses, our J-S estimates for female bears were not affected by sex-related differences in capture rates. We believe perceptions of trends were robust to age-related differences because they were affected only slightly when we restricted our analysis to adult females. Estimates were similar from 1969 through 1970, increased abruptly during 1971-72, remained high through 1974, and declined abruptly in 1975. Thus, our results, like those of Kemp (1976) and Young and Ruff (1982), suggest bear density increased when adult males were removed (Fig. 2). However, neither their results nor ours should be used to calculate actual densities (i.e., bears/km²). The CLSA was small relative to home range sizes of bears; consequently, the effective area sampled was not known with sufficient accuracy for density estimation (White et al. 1982).
Changes in estimated numbers of females could have resulted from (1) parallel changes in the effective area sampled by trapping or (2) the temporary departure of bears marked prior to 1975; however, these alternative explanations seem comparatively unlikely. First, distances between capture locations suggest movements of females may have been more extensive during 1968-71 than during 1972-75 (Young and Ruff 1982). Resulting changes in the effective area trapped would have dampened, rather than produced, the trends in our estimates. Second, the abrupt increase in estimates during 1971-72 persisted even when we made departures effectively permanent by truncating capture histories in 1974. Finally, our results are consistent with subjective perceptions of field staff from the Cold Lake study, who reported a notable increase in the frequency of bear sightings after males were removed.
An abrupt increase during 1971-72 in the number of females using the CLSA nevertheless seems surprising. Typically, adult females monitored for ≥2 years occupy the same home range in successive years (e.g., Alt et al. 1980, Reynolds and Beecham 1980, Powell 1987). Emigration by females is thought to be uncommon for reasons related to reproductive fitness (Rogers 1987). However, shifts in home ranges of adult females are not unprecedented. For example, adult females monitored by Lindzey et al. (1986) on Long Island, Washington, USA, adjusted home ranges and increased their use of the adjacent mainland as bear densities and competition for resources increased. Moreover, the removal of adult males probably subjected bears to abrupt changes in social pressure or competition for resources that were more extreme than typically occur while individual bears are being monitored.
Female bears radiomarked within the CLSA during 1974-77 sometimes made long-range exploratory movements (Pelchat and Ruff 1986, R.L. Ruff unpublished data). Presumably, bears living near the CLSA did so as well and were aware the area comprised primarily aspen and aspen-spruce habitats, a better source of food (Pelchat and Ruff 1986), preferred over muskeg and spruce that dominated the surrounding area (Young and Ruff 1982, Pelchat and Ruff 1986). Thus, habitat differences may have motivated female bears from the surrounding area to make greater use of the CLSA after adult males were removed. Subadult males may have been attracted to potential mates as well as by habitat.
Settling rates were probabilities that marked bears would survive from 1 capture period to the next and continue using the study area. They were not annual survival rates and did not distinguish mortalities from departures. Thus, we were able to estimate annual mortality rates only for bears marked with radios during 1974-77. Results were surprising because the CLSA was closed to sport hunting. Mortality rates were much higher than those of previously studied unhunted populations (Bunnell and Tait 1985, Clark and Smith 1994, Doan-Crider and Hellgren 1996). During 1975-77, mortality rates equaled or exceeded estimates of the maximum sustainable by black bears (22-24%; Bunnell and Tait 1981).
Annual variation in survival rate estimates was substantial, but regulations governing human exploitation of bears did not change during the study. Thus, immigration may have played an important role, even prior to the removal of adult males, in sustaining bear densities within the CLSA. This new finding has important implications, because previous interpretations of this study have been predicated on the belief that human exploitation was negligible. Bear densities may not have been naturally regulated at the outset of the Cold Lake study, and high rates of human-caused mortality may have contributed to the abrupt decline in density that occurred in 1975. Young and Ruff (1982) previously reported a more gradual decline resulting from changes in social structure as males matured on the study area.
Our predecessors described a population that comprised bears using the CLSA (Kemp 1972, 1976; Young and Ruff 1982). Thus, population growth rates were determined by 4 processes: reproduction, mortality, immigration, and emigration. Density-dependence was indicated if any of these processes were influenced by abundance. Unfortunately, this common perspective (e.g., Caughley 1977, Shenk et al. 1998) poses 2 difficulties for interpretation when small "populations" are arbitrarily spatially delimited from larger groups of individuals (Harrison and Cappucino 1995). First, the occurrence of immigration, emigration, and changes in density are scale-dependent (Fig 3; Wells and Richmond 1995). Second, dynamics of population numbers over time are not distinguished from movements of individuals (Allen and Hoekstra 1992), although the distinction is of critical importance for bear management.
|Fig. 3. Demographic concepts are ambiguous when small study populations are arbitrarily delimited. Removing individuals from the circle causes the actual density to decline in the square, but immigration maintains the actual density in the circle. Relative local densities are increased within and reduced outside the circle.|
To resolve semantic difficulties, we believe bear biologists should reserve population status for disjunct (Wells and Richmond 1995) or, minimally, much larger groupings (e.g., local populations; Goodwin and Fahrig 1998) of bears than was studied at Cold Lake. When populations are defined this way, the critical processes of population dynamics are birth and death (Allen and Hoekstra 1992); movements of individuals are no less important, but are said to influence local densities and metapopulation dynamics.
Regardless of how they are defined, populations cannot increase without limit and seldom dwindle to extinction. Regulation is perhaps best defined as the tendency of populations to fluctuate, over the long term, about stationary mean densities (Turchin 1995). Limiting factors influence these means (Sinclair 1991). Direct density dependence is an inverse relation between per capita population growth rates and present or past population densities (Murdoch and Walde 1989, Turchin 1995). Direct density-dependence is a necessary (but not sufficient) condition for population regulation (Murdoch and Walde 1989).
Like most studies of black bears, the Cold Lake experiment was not controlled. Thus, the removal of adult males was confounded with other influences on bears of the CLSA. However, the timing, magnitude, and duration of changes in numbers of bears using the CLSA strongly suggest that the presence of adult males influenced local bear densities. In keeping with their use of population, our predecessors interpreted this effect as evidence of density dependence. However, we note that (1) effects were documented only for a localized area; (2) adult males depressed densities of bears in other sex and age classes; and (3) interactions with conspecifics are 1 of many factors that may operate on local densities in a relative sense, across a wide range of actual densities. Thus, in keeping with our synthesis of terminology, we conclude that adult males limited relative local densities of female and subadult male bears. Our results do not permit inferences about density dependence or population regulation, as we defined these terms. We could not estimate population growth rates, test for changes in survival or reproduction, or estimate population densities across an area large enough to encompass effects of the experimental treatment.
At Cold Lake, adult males were removed from a restricted area during a short period. The removal of bears from broad geographic areas over long time spans is more typical of hunted black bear populations. In such cases, hunter harvests of adult males are unlikely to cause sudden changes in the dispersion of subadult and female bears. Further, the response observed at Cold Lake may have occurred only because the CLSA was a population sink comprising aspen and spruce habitats, which bears preferred over surrounding muskeg. Removing males from less desirable habitats might not have a similar effect.
Density-dependent population regulation is a topic of special interest to bear biologists interested in maximizing sustainable harvest rates (McClellan 1993, 1994; Derocher and Taylor 1994; Garshelis 1994; Taylor 1994), but efforts to conclusively demonstrate density dependence within black bear populations may be doomed to failure. Formidable logistical barriers limit the accuracy and precision of density estimates. Unambiguous detection of density dependence is difficult (Pollard and Lakhani 1987), even with error free estimates, and some tests designed to accommodate sampling error are invalid (Shenk et al. 1998). Moreover, most black bear populations are studied at moderate densities, where changes in density probably have little effect on reproduction or survival (Fowler 1981a,b; Strong 1986). Finally, even at high densities, regulatory agents may reduce survival or suppress reproduction at irregular intervals, so that a "vague" (Strong 1986) relationship between density and demographic performance of a population should be expected.
Similarly, density dependence is difficult to detect by comparing the demographic performance of different populations that vary in density. Garshelis (1994) noted some obstacles, including the difficulty of measuring relevant parameters with sufficient accuracy, complications that arise from methodological differences among studies, and factors other than density that influence vital rates. More important, however, is the fact that carrying capacity varies among populations. Thus, density alone is likely to be a poor predictor of demographic performance, even if compensatory responses are strong and population density can be estimated accurately.
In large mammals, strong density-dependent responses usually occur only when populations near carrying capacity are reduced (Fowler 1981a,b). Probably few managed bear populations are near carrying capacity (Taylor 1994). Thus, much current interest in density dependence and population regulation might more profitably be directed toward factors that limit growth rates of populations at low densities. The management significance of such factors may be much greater. Relief from factors that limit growth rates can buffer low populations from overharvest and increase sustainable yields, whether or not density-dependent responses occur (Gasaway et al. 1992, Van Ballenberghe and Ballard 1994). Although we could not assess effects of adult males on population growth rates, their influence on relative local densities should encourage further study of this issue, especially with respect to mortality rates and causes of death during juvenile dispersal.