Polar bear (Ursus maritimus) COSEWIC assessment and status report: chapter 7

7. Population Size and Trends

This section updates summaries presented in the previous COSEWIC report on the status of polar bears (COSEWIC 2002). It largely follows from submissions by participants of the Working Meeting of the IUCN/SSC Polar Bear Specialist Group (2006), including the authors of this report for text related to Section 7.1, Table 6, and summaries of Nunavut subpopulations. We also include recently collected ATK and information presented at the Federal/Provincial/Territorial Polar Bear Technical Committees of 2006 and 2007 (PBTC 2006, 2007).

The world’s population of polar bears is estimated to number some 20,000–25,000 animals, of which approximately 15,500 are in Canada or in subpopulations shared with Canada (Table 6). Because several subpopulations are shared internationally and each is managed independently, it is difficult to meaningfully discuss dynamics of a single “Canadian population” of polar bears (Figure 2, Section 3.3). Table 6 summarizes for each subpopulation current estimates of abundance, past trends in subpopulations, current human-caused mortality, and estimated current trajectories (annualized finite rates of increase, l) and projections of trend where recent data on abundance and rates of survival and reproduction allow. Also presented are the zero-harvest estimates of lfor each subpopulation (i.e., estimated growth rate should harvest immediately cease). Projections of trend into the future (population viability analysis [PVA]) and current estimates of lare based on simulation models using age- and sex-structured, stochastic population models presented for subpopulations using information on abundance, reproduction (Table 3), estimated harvest over the past 5 years, and estimated survival rates (Tables 4 and 5). Sections 7.2–7.14 present detailed information on the recent history and status of each Canadian subpopulation, including literature sources for data presented in Tables 3–6.

7.1 Status Table

Table 6 presents subpopulation sizes and uncertainty in estimates as 95% confidence intervals (CI). These estimates are based on scientific research using mark-recapture analysis, except where foot noted. The years in which data were collected are presented to provide an indication of the current reliability of subpopulation estimates. Where >1 estimate of abundance was available, the 2 latest estimates are presented. Note, however, that a difference in abundance from a previous to a current estimate may not necessarily entail any real increase or decrease in abundance, but rather a correction from a past under- or over-estimate of abundance. The past trend of subpopulations as suggested by the relevant citations of abundance are indicated as decline, stable, or increase.

For most subpopulations, harvesting of polar bears is regulated; hunting is by far the major cause of mortality for polar bears. In most jurisdictions, the total number of bears killed by humans in pursuit of sport and subsistence hunting, accident, and in defence of life or property are documented. Table 6 presents the 5-year means of known human-caused mortalities (removals) for each subpopulation (2002–2007).

For most subpopulations, recent quantitative estimates of abundance and parameters of natural survival, reproduction, and human kill are available to determine the current finite rate of increase (l±1 standard error [SE]) of the population (e.g., over the next 5 years). The computation oflis an application of birth and death rates (and their uncertainty), carried out using the simulation model RISKMAN v. 1.9005 (Taylor et al. 2003). Values of l > 1.0 indicate population growth and l < 1 indicate population decline (l= 1.0 suggests stability). Stated values are estimates of the current trend of a given subpopulation, which in some cases updates trend information obtainable from past and current estimates of abundance. These values are our best interpretation of the current, short-term trajectory of a subpopulation. For comparison, also presented for each subpopulation is the estimated lunder harvest moratorium. This value conveys the importance of harvest on determining current population trends, and instances of where climate change may have impacted survival rates to such an extent that subpopulations decline even without harvesting. We caution that estimates of lcannot be averaged across subpopulations (i.e., to develop a population-wide rate of increase), as subpopulations are managed and modelled as unique demographic units, and assume no rescue effects or metapopulation dynamics (Hanski and Gilpin 1997).

We also used RISKMAN (Taylor et al. 2003) to estimate likelihoods of future subpopulation declines over 3 generations of polar bears (36 years; Section 5.1). Recent publications using the RISKMAN PVA model include Dobey et al. (2005), McLoughlin et al. (2005), Wear et al. (2005), Clark and Eastridge (2006), and Howe et al. (2007).

RISKMAN is designed to incorporate uncertainty into population simulations at several levels, including sampling error in initial subpopulation size, variance about vital rates due to sample size and annual environmental variation (survival, reproduction, sex ratio), and demographic stochasticity. RISKMAN uses Monte Carlo techniques to generate a distribution of results, and then uses this distribution to estimate population size at a future time, projected mean annual finite rate of population increase over the period involved, and proportion of runs that result in population decline set at a predetermined level by the user. We adopted the latter to estimate the probability of persistence.

Our approach to variance in simulations was to partition total variance of vital rates by a ratio of 3:1 sampling to environmental variation (Taylor et al. 2008a,b,c). We did this because variances for reproductive parameters often did not lend themselves to directly separating the sampling component of variance from environmental variance. Simulation results suggested that sampling error had a greater impact on population viability compared to environmental error, and so we weighted sampling error more heavily in the simulations to produce more conservative results. Differences based on weighting sampling vs. environmental error were minor, however (see Taylor et al. 2008a,b,c).

Individual runs could recover from “depletion”, but not from extinction. Simulations also assumed no immigration or emigration among subpopulations. Because estimates of likelihood of decline were calculated without consideration of metapopulation dynamics (in particular rescue effects between subpopulations), overall results are more likely to indicate decline than is the case.

Simulations are presented using rates of survival and reproduction that were estimated within the past 10 years, and therefore assume effects of climate change leading to today’s climatic conditions only. RISKMAN does not incorporate effects of directional environmental or habitat change on demographic parameters. Thus, regardless of length of simulation, results are relevant for near-term status assessment only.

We modelled PVA as the proportion of 2,500 simulation runs after 36 years resulting in a decline that is ≥30% from initial population size. This value is currently used by COSEWIC for distinguishing between Threatened and Special Concern status of designatable units (Appendix E3, COSEWIC Operations and Procedures Manual, draft dated April, 2005); however, as stated above, due to unknown effects of directional climate change on survival and recruitment, results should be used to interpret current and short-term likelihoods of decline only. Further, each modelled subpopulation is in itself not a Designatable Unit (Section 3.2). Note that persistence probabilities converge on a single percentage without error after running a large number of simulations (>1,000), and so no error estimate is applicable to likelihood of decline.

Required parameter and standard error inputs included: annual natural and/or total survival rates (stratified by age and sex as supported by the data; Tables 4 and 5), stratified rates of harvest if not using total survival rates (we used 5-year means; Table 6), data on reproduction (age of first reproduction, age-specific litter production rates for females available to have cubs [i.e., females with no cubs and females with 2-year-olds], and litter size; Table 3), and abundance (Table 6). We only used total survival rates (Table 4) if estimates were very recent and thus incorporated current harvest levels (i.e., those comprising the 5-year mean [2002–2007] for a subpopulation; Table 6). If harvest data had changed since collection of survival data, we modelled populations based on estimated natural survival (Table 5) from which the current age and sex-stratified hunting mortality was deducted.

The standing age distribution was female-biased for all subpopulations, likely due to harvesting of males. Because we wished to err on the side of caution, for all simulations we used the stable age distribution expected for the subpopulation at the anticipated annual removal rate as the initial age/sex distribution (i.e., by initializing the subpopulation at the stable age distribution, more conservative outcomes were produced compared to runs initialized at the current standing age distribution).

7.2 Southern Beaufort Sea

The subpopulation of polar bears inhabiting the Southern Beaufort Sea is shared between Canada and the United States (Alaska). On the Canadian side of the border, the historical harvest of bears has been relatively light. The subpopulation experienced an increase in hunting activity in the late 1950s due to an increase in fur prices (Usher 1976); however, by the mid-1970s polar bears were only killed opportunistically during hunts for other species by Aklavik and Inuvik hunters (Usher 1976). Hunters of Tuktoyaktuk recall people from their community also hunting polar bears during this time (Frank Pokiak, Chair, Inuvialuit Game Council, letter to COSEWIC Terrestrial Mammals Specialist Subcommittee, January 19, 2007). The Cape Bathurst area was reported to be an important area for hunting polar bears (Usher 1976).

During the early 1980s, radio-collared individuals were tracked from the Canadian portion of the Southern Beaufort Sea into the eastern Chukchi Sea of Alaska (Amstrup et al. 1986; Amstrup and DeMaster 1988). Telemetry data combined with re-sightings of tagged individuals suggested that bears of the Southern Beaufort Sea comprised a single population with an eastern boundary between Paulatuk and Ballie Island, NWT, Canada, and a western boundary near Icy Cape, Alaska (Amstrup et al. 1986; Amstrup and DeMaster 1988; Stirling et al. 1988). Recognition that bears were shared by Canada and Alaska prompted the Polar Bear Management Agreement for the Southern Beaufort Sea (the Agreement). The Agreement, between the Inupiat hunters of Alaska and the Inuvialuit hunters of Canada, was ratified by both parties in 1988. The Agreement included provisions to protect bears in dens and females with cubs, and stated that the annual sustainable harvest from the Southern Beaufort Sea would be shared between the 2 jurisdictions. Harvest levels were to be reviewed annually in light of the best scientific information available (Treseder and Carpenter 1989; Nageak et al. 1994). Brower et al. (2002) evaluated the effectiveness of the Agreement after the first 10 years and concluded that, overall, it had been successful in ensuring that the total harvest and the harvest of adult females remained within what were thought to be sustainable limits.

Amstrup et al. (1986) estimated the size of the Southern Beaufort Sea subpopulation to be approximately 1,800 bears in 1983, with a minimum and maximum of 1,300 and 2,500 bears, respectively. Research incorporating mark and recapture and radio-telemetry has continued on a nearly annual basis through to the present time. Capture-recapture models applied to data collected from 2001 to 2006 suggest that there were 1,526 (95% CI 1,211–1,841) polar bears in the region in 2006 (Regehr et al. 2006).

Rates of survival and recruitment have recently been developed for bears of the Southern Beaufort Sea (Regehr et al. 2006; 2007b; PBTC 2007; Tables 3–5). Given the current combined U.S.-Canadian harvest of bears in the Southern Beaufort Sea, the population is likely declining at the present time (Table 6). Importantly, should harvest be decreased to zero, there is a high likelihood that the population would continue to decline (Table 6).

7.3 Northern Beaufort Sea

Studies of polar bears in the Northern Beaufort Sea have used telemetry and mark and recapture programs at regular intervals since the early 1970s (Stirling et al. 1975, 1988; DeMaster et al. 1980; Lunn et al. 1995). Results suggested that there were separate subpopulations in the Northern and Southern Beaufort Sea areas and not a single subpopulation as was initially thought (Stirling et al. 1988; Amstrup 1995; Taylor and Lee 1995; Bethke et al. 1996). An abundance estimate of 1,200 polar bears in the late 1980s (Stirling et al. 1988) was believed to be unbiased, but is now dated. Stirling et al. (2007) updated both the previous and current estimates of abundance in the Northern Beaufort Sea, and conclude current stability in the population. Recent analyses, using data from satellite tracking of female polar bears and new spatial modelling techniques, indicate the boundary between the Northern Beaufort Sea and the Southern Beaufort Sea subpopulations may need to be adjusted, probably by expanding the area occupied by bears from Northern Beaufort Sea and reducing that of Southern Beaufort Sea (Amstrup et al. 2004).

Hunting of polar bears of the Northern Beaufort Sea has historically focused on the Amundsen Gulf (Usher 1976; Farquharson 1976), although the western coast and associated sea ice of Banks Island are also important for Inuit hunters (Usher 1976). Little ATK on the status of polar bears in this area has been recorded, but what is available supports this assertion. In a 2001 interview for the Paulatuuq Oral History project, an elder hunter suggested that the population in the area had been stable over the past 30 years (Parks Canada 2004).

Trend for this subpopulation is believed to be stable by Stirling et al. (2007), with abundance currently at 980 bears (95% CI: 825–1,135). Reported survival rates for some age groups (e.g., yearlings) have high variability and are unreasonably low (e.g., by half compared to cubs-of-the-year, and lower than ever reported in the literature; Table 4), precluding simulations based solely on North Beaufort data (Table 6). Lower than expected survival rates are likely due to unmodelled heterogeneity in the capture data.

7.4 Viscount Melville Sound

Only in the past 30 years have polar bears of the Viscount Melville Sound experienced regular hunting pressure. Farquharson (1976) noted that by the mid-1970s, hunters from the Holman area had expanded their traditional hunting range to kill polar bears along the western and northern coasts of Victoria Island to Glenelg Bay. At the same time, Inuit from Cambridge Bay began travelling by land or air to reach northern Victoria Island to hunt polar bears. In response to increased interest in hunting bears of the Viscount Melville Sound, the Government of the Northwest Territories established quotas. When quotas were originally allocated in the 1970s, the size and productivity of the Viscount Melville Sound subpopulation was overestimated. Polar bear density is lower in Viscount Melville Sound compared to other regions because of large expanses of multi-year ice and low densities of ringed seals (Kingsley et al. 1985). The consequence of overestimating abundance when initially setting quotas was substantial over-harvest of bears in the region during the 1980s and early 1990s (e.g., 1985–1990 mean of 19.6 bears/year; Taylor et al. 2002).

A 5-year moratorium on hunting was enacted in 1994/1995. Hunting resumed in 1999/2000 with an annual quota of 4 bears. In 2004/2005 the annual quota was increased to 7 bears/year (Northwest Territories 4, Nunavut 3) to accommodate hunters on both sides of the new territorial border. Polar bear numbers in the Viscount Melville Sound should be increasing with this increase in quotas (Table 6); however, the subpopulation remains at historically low levels and the abundance estimate is now becoming dated.

A 5-year study of movements and size of the Viscount Melville Sound subpopulation of polar bears using satellite telemetry and mark and recapture sampling was completed in 1992 (Messier et al. 1992, 1994; Taylor et al. 2002). Current boundaries are based on observed movements of females with satellite radio-collars and movements of bears tagged inside and outside of the study area (Bethke et al. 1996; Taylor et al. 2001). The published 1996 abundance estimate of 215 bears (SE = 58) in Taylor et al. (2002) was based on the 1993 estimate plus 3 years of simulated population growth. Polar bears in the Viscount Melville Sound are likely to benefit from a warming climate (at least over the short-term), which may increase the abundance and accessibility of seals by reducing amounts of multi-year ice.

7.5 Norwegian Bay

The polar bear subpopulation of Norwegian Bay is bounded by multi-year ice to the west, islands to the north, east, and west, and polynyas to the south (Taylor et al. 2001; Taylor et al. 2008b). Based on data from mark-recapture studies and satellite radio-tracking of adult females, it appears that most bears concentrate along coastal tide cracks and ridges in the northern, eastern, and southern regions of Norwegian Bay (Taylor et al. 2001). The preponderance of multi-year ice through most of the central and western areas contributes to low densities of ringed seals (Kingsley et al. 1985) and, consequently, low polar bear density. Grise Fiord hunters reported high concentrations of polar bears in Norwegian Bay during the early 1970s (Riewe 1976); however, based on unpublished data, the current (1993–97) estimate for this subpopulation is 190 bears (SE = 48.1; Taylor et al. 2008b). Estimates of survival rates (Tables 4 and 5) for Norwegian Bay are derived from pooled Lancaster Sound and Norwegian Bay data because these 2 subpopulations are adjacent and because the number of bears captured in Norwegian Bay was too small for reliable survival estimates (Taylor et al. 2008b). Risk of decline (Table 6) is high for this subpopulation because of a relatively low rate of reproduction (Table 3) and low abundance; however, polar bears in Norwegian Bay are likely to benefit from a warming climate (at least over the short term) which may increase abundance of and accessibility to seals. The harvest quota for the Norwegian Bay subpopulation was reduced to 4 bears (3M:1F) in 1996 and remains at this level today. The population is probably stable at the current time but there is substantial incertitude about its trend.

7.6 Lancaster Sound

The central and eastern portion of the Lancaster Sound subpopulation is characterized by high productivity and thus high densities of ringed seals and polar bears (Schweinsburg et al. 1982; Kingsley et al. 1985; Welch et al. 1992). Inuit hunters of Resolute, Grise Fiord, and Arctic Bay have all historically hunted polar bears in Lancaster Sound (Brody 1976; Riewe 1976). The western third of this region (eastern Viscount Melville Sound) is dominated by multi-year ice and apparently low biological productivity, leading to low densities of ringed seals (Kingsley et al. 1985). In the spring and summer, densities of polar bears in the western third of the area are low; however, as break-up occurs, polar bears move west to summer on the multi-year pack.

Mark-recapture data and data on movements of adult females fitted with satellite radio-collars have been collected for bears of Lancaster Sound (Taylor et al. 2001, 2008b). The current abundance estimate of 2,541 bears (SE = 391) is based on an analysis of mark-recapture data current to 1997 (Taylor et al. 2008b). This estimate is considerably larger than the 1979 estimate of 1,031 ± 236 bears (mean ± 95% CI) published by Schweinsburg et al. (1982); however, given the substantial differences in study area boundaries between Schweinsburg et al. (1982) and Taylor et al. (2008b), it is difficult to compare estimates. Schweinsburg et al. (1982) focused on a much smaller area that extended into northern Baffin Bay, compared to the bounds used by Taylor et al. (2008b) and presented in Figure 2. Recent survival rates of polar bears Lancaster Sound are presented by Taylor et al. (2008b). Note that Taylor et al. (2008b) pooled the similar survival rates of Lancaster Sound with those of Norwegian Bay to minimize sampling errors (Tables 4 and 5).

7.7 M'Clintock Channel

The current boundaries for the M’Clintock Channel subpopulation are based on recoveries of tagged bears and movements of adult females with satellite radio-collars in adjacent areas (Taylor and Lee 1995; Taylor et al. 2001). These boundaries appear to be a consequence of large islands to the east and west, the mainland to the south, and the multi-year ice in Viscount Melville Sound to the north. A 6-year mark-recapture study in the mid-1970s covered most of this area (Furnell and Schweinsburg 1984). An estimate of 900 bears was derived from data collected within the boundaries proposed for the M’Clintock Channel subpopulation from results of Furnell and Schweinsburg (1984), who identified an abundance estimate of 1,100 animals for an area that overlapped both M’Clintock Channel and the Gulf of Boothia. After the study was published, local hunters suggested the estimate of 900 animals may have been too high; hence, the Federal/Provincial/Territorial Polar Bear Technical Committee accepted a recommendation to reduce the abundance estimate to 700 bears (M.K. Taylor, Department of Environment, Government of Nunavut), which would have been current to 1978 (i.e., the last year of Furnell and Schweinsburg’s sampling program). No accurate confidence intervals are available for this estimate.

Following completion of a mark-recapture inventory in spring of 2000, the subpopulation was estimated to number only 284 bears (SE = 59.3; Taylor et al. 2006a). The legal harvest (averaging 34.0 bears/year from 1979–1999) for M’Clintock Channel was clearly unsustainable. The Government of Nunavut implemented a moratorium on hunting for the 2001/2002 and 2002/2003 hunting seasons. The current annual quota for M'Clintock Channel is 3 bears and the population is now likely growing (Table 6); however, the subpopulation remains at risk due to low abundance.

Scientific data which suggests low abundance of polar bears in M’Clintock Channel due to over-harvest is supported by ATK. Recently, hunters of Gjoa Haven reported that the number of bears near their community has declined over the past 30 years (Keith et al. 2005). Other areas where decreased numbers of polar bears have been reported include the Royal Geographical Society Islands, Pasley Bay, northern King William Island, Gateshead Island, Larsen Sound, and the M’Clintock Channel itself (Atatahak and Banci 2001). Inuit suggest that polar bears are no longer present in the Queen Maud Gulf area (Keith et al. 2005). Inuit hunters also report a decline in the number of adult male bears in M’Clintock Channel but that large males can be found further to the north (Atatahak and Banci 2001; Keith et al. 2005). This finding is consistent with what one could expect from a relatively heavy, male-biased hunt. Northern M’Clintock Channel is the recent focus of the polar bear hunt out of Gjoa Haven (Keith et al. 2005).

In addition to unsustainable harvesting, recent changes in habitat and disturbance by humans have been identified by Inuit as potential reasons for the reduced abundance of bears in M’Clintock Channel (Keith et al. 2005). One noted habitat change has been the recent absence of multi-year ice and icebergs, although this may offer improved habitat for ringed seals and so may not necessarily be detrimental to polar bears. Human disturbances such as the construction of DEW (Distant Early Warning) line sites, construction of Inuksuit, and noise from aircraft and snowmobiles are also thought to have contributed to the low density of bears around Gjoa Haven (Keith et al. 2005).

7.8 Gulf of Boothia

Boundaries of the subpopulation of polar bears inhabiting the Gulf of Boothia were largely based on movements of tagged bears (Taylor and Lee 1995), movements of collared females in the Gulf of Boothia and adjacent areas (Taylor et al. 2001), and information from Inuit hunters about how local conditions influence the movements of polar bears. Distinction between the Gulf of Boothia and M'Clintock Channel subpopulations of polar bears has, however, recently been questioned by Inuit hunters (Keith et al. 2005), and new genetic analyses (Saunders 2005) suggest considerable interchange between the subpopulations.

Hunting in the Gulf of Boothia increased from historic levels through the 1970s (Brice-Bennett 1976); however, unlike the situation in Viscount Melville Sound and M’Clintock Channel, the original quota established by the Government of the Northwest Territories in the Gulf of Boothia was likely less than the maximum sustainable yield. Local hunters reported that the subpopulation increased during the 1980s after results of Furnell and Schweinsburg (1984) suggested abundance at around 300 bears (considering that portion of the Gulf of Boothia included in their study area). Based on Inuit knowledge, recognition of past sampling deficiencies, and an increased understanding of polar bear densities in other areas, the interim subpopulation estimate in the 1990s for the Gulf of Boothia was 900 bears (M.K. Taylor, Department of Environment, Government of Nunavut). Following completion of a mark-recapture inventory in spring of 2000, the subpopulation was estimated to number 1,528 bears (SE = 285; Taylor et al. 2008c). Recruitment and survival rates (Tables 3, 4 and 5) were estimated to be relatively high. The subpopulation is considered to be growing (Table 6), and in 2005 harvest quotas were increased by the Government of Nunavut to 74 bears/year. Hunting success rates are reported to be high, although this does not convey information about status (personal communication of Kotierk [2005]).

7.9 Foxe Basin

Based on 12 years of mark-recapture studies, tracking of female bears with conventional radios, and satellite tracking of adult females in Western and Southern Hudson Bay, the Foxe Basin subpopulation is thought to comprise a demographic unit in Foxe Basin, northern Hudson Bay, and the western end of Hudson Strait (Taylor and Lee 1995). During the ice-free season, polar bears concentrate on Southampton Island and along the Wager Bay coast; however, significant numbers of bears also occur on the islands and coastal regions throughout the Foxe Basin area. Crête et al. (1991) found relatively few bears of the Foxe Basin population along the Quebec shore during the ice-free season. A total abundance estimate of 2,119 (SE = 349) was made in 1996 (M.K. Taylor, Department of Environment, Government of Nunavut, unpubl. data) from a mark-recapture analysis based on tetracycline biomarkers (Taylor and Lee 1994; M.K. Taylor, Department of Environment, Government of Nunavut, unpubl. data). The marking effort was conducted during the ice-free season, and distributed throughout the entire area. The abundance estimate is believed to have been accurate, but is now dated. Simulation studies suggest that harvest quotas prior to 1996 reduced the subpopulation from approximately 3,000 in the early 1970s to 2,100 bears in 1996. Harvest levels were reduced in 1996 to permit recovery of this subpopulation, provided that harvest in Quebec did not increase.

Recent ATK suggests that the subpopulation of Foxe Basin has increased since 1996 (McDonald et al. 1997). For example, at Southampton it has become not unusual for hunters to fill their quota in a matter of days (McDonald et al. 1997).However, ATK from the Ivujivik area indicates a decrease in polar bear numbers. One hypothesis proposed to explain this observation is that ocean currents in the region are now weaker, allowing bears to become distributed more evenly on the ice during mid-winter rather than congregating at the mouth of Hudson Strait (McDonald et al. 1997).After consultations with native communities, Nunavut increased the harvest quota in 2004 to a level consistent with a subpopulation size of 2,300 bears (109 bears/year). Co-management discussions with Quebec are ongoing.

Effects of climate change on the Foxe Basin subpopulation of polar bears have not been evaluated scientifically. Because Foxe Basin is immediately north of Western Hudson Bay and has experienced earlier timing of break-up of sea ice in similar fashion as the rest of Hudson Bay, future analyses might detect negative impacts on polar bears.

7. 10 Western Hudson Bay

The distribution, abundance, and boundaries of the Western Hudson Bay subpopulation of polar bears have been studied since the late 1960s (e.g., Stirling et al. 1977; Derocher and Stirling 1990, 1992, 1995a,b; Taylor and Lee 1995; Lunn et al. 1997, 2006). Between 60–80% of adults have been marked at any given time and there are extensive records from mark-recapture studies and the return of tags from bears killed by Inuit hunters, and from the ongoing and long-term Polar Bear Alert Program of the Government of Manitoba. This population appears to be geographically segregated during the open-water season, although it mixes with those of Southern Hudson Bay and Foxe Basin on the Hudson Bay sea ice during the winter and spring (Stirling et al. 1977; Derocher and Stirling 1990; Stirling and Derocher 1993; Taylor and Lee 1995).

ATK of Inuit elders concerning polar bears in Western Hudson Bay has recently been summarized by Nirlungayuk (2008). Nirlungayuk (2008) suggests that polar bear abundance in the areas of Western Hudson Bay are today considerably higher than in the historic past (50+ years ago), and that this may have been the result of supplemental feeding by garbage (around Churchill). For example, prior to the increase in the population of humans in Churchill in the 1940s, polar bears were best hunted nearer Wager Bay, Southampton Island, and Coates Island; after polar bear hunting regulations came in if people wanted to be guaranteed a polar bear, they would travel down to south of Arviat. Observing polar bear dens by boat in Western Hudson Bay was once a rare event; now “lots of bears are there.” Concurrently but contrary to the scientific re-assessment of abundance (below), Inuit along the western coast of Hudson Bay recently reported seeing greater numbers of polar bears, which they interpreted as evidence of an increasing population (McDonald et al. 1997; Dowsley and Taylor 2006b). Polar bears have been reported as numerous at Chesterfield Inlet in September and have been increasing in that area since 1988. Bears have been present for several years near Arviat, from September to December, but have recently increased in number according to ATK, especially in September.

The dangers posed by polar bears in the region are a concern to Inuit, and what is thought by some to be an artificially high number of bears is viewed as a problem (Nirlungayuk 2008). Encounters in the region have increased through the 1970s and 1980s; since the 1980s the Arviat community has been giving warnings to hunters that persons should not go out alone for fear of polar bears. Nunavut Tunngavik Incorporated (NTI) recently collaborated with 5 experienced hunters from communities in Western Hudson Bay to complete a series of interviews and a workshop (NTI 2005). While the final analysis and report are not yet complete, the ATK data indicate an increasing number of bears in the Arviat area since the 1970s, and around Whale Cove and Rankin Inlet since the 1980s. This has also been noted by Inuit of Chesterfield Inlet. In Arviat, the recent increase has been noted in all seasons except winter, while Inuit of other areas report an increase in all seasons. In the Chesterfield Inlet area, groups (gatherings) of polar bears have been observed recently, something that was apparently rare in the past.

Over the past 30 years, the condition of adults and the proportion of independent yearlings caught during the open-water season have declined significantly in Western Hudson Bay (Derocher and Stirling 1992, 1995b; Stirling and Lunn 1997; Stirling et al. 1999; N. Lunn and I. Stirling, unpubl. data presented in IUCN/SSC Polar Bear Specialist Group 2006). Over the same period, the average date of break-up of the sea ice has advanced by 3 weeks (Stirling et al. 1999, 2004; Ferguson et al. 2005), probably due to increasing spring air temperatures (Section 6.1). Stirling et al. (1999) documented that the earlier the timing of break-up, the poorer the condition of adult females. Inuit are intimate with the changing ice conditions in Western Hudson Bay (Nirlungayuk 2008).

The number of polar bears in Western Hudson Bay was most recently assessed scientifically by Regehr et al. (2007a). Regehr et al. (2007a) show that abundance has declined from 1,194 (95% CI = 1,020–1,368) to 935 (95% CI = 794, 1,076) between 1987 and 2004, a reduction of approximately 22%. Progressive declines in the condition and survival of cubs, subadults, and bears 20 years of age and older likely initiated decline in the size of the subpopulation. It is believed that once the subpopulation began to decline, the existing harvest was no longer sustainable so that its additive contribution to the reduction in the size of the subpopulation accelerated between 1988 and 2004. The harvest sex ratio of 2M:1F in Western Hudson Bay has resulted in a sex ratio that is 58% female and 42% male (Derocher et al. 1997).

In summer 2007, the Government of Nunavut conducted a mark-recapture survey of bears from Churchill to Chesterfield Inlet to determine whether or not there were large numbers of bears along the Kivalliq coast during the summer as suggested by ATK (Peacock and Taylor 2007). The survey included those areas identified by ATK as being areas where polar bears were becoming more common. A total of 25 bears were captured during the 3-day survey. The proportion of marked individuals in the capture sample (p = 0.46, SE = 0.11) was lower but not statistically different from the proportion of marked animals in the Canadian Wildlife Service (CWS) capture sample (p = 0.59, SE = 0.01). Statistical power was low for this analysis. Results suggest that actual numbers of bears in Western Hudson Bay and annual survival rates could thus be slightly but not significantly higher (due to unmodelled heterogeneity) than estimated by Regehr et al. (2007a). Peacock and Taylor (2007) recommend that in future years, CWS capture teams work north to Arviat to capture polar bears in the entire area where polar bears summer, but do not contest the conclusions of Regehr et al. (2007a).

Climate change in connection with over-harvest is the major threat to the Western Hudson Bay subpopulation. The population is believed to be declining at a substantial rate (Table 6), and the quota for hunting polar bears in Western Hudson bay is proposed to be reduced to 8 animals in 2008–2009.

7.11 Southern Hudson Bay

Inuit hunting the Southern Hudson Bay subpopulation of polar bears reported an increase in the number of bears that have historically occurred in the area (McDonald et al. 1997). The offshore islands of eastern Hudson Bay apparently had no bears 50 years ago, and the species was rare around Inukjuak, only appearing “recently” (McDonald et al. 1997). Similarly, in Sanikiluaq, it was rare to kill a polar bear in the 1960s but now the community’s annual quota is filled in approximately 3 weeks, with increased observations of bears coming into the community (personal communication of Arragutainaq [2006]). In 1986, Crête et al. (1991) found relatively high numbers of bears near Twin Island in James Bay during the ice-free season. Cree in western James Bay report increased aggressiveness among bears and an increase in litter size (McDonald et al. 1997). Communities along the Hudson Bay and James Bay coasts in Ontario report an increase in bear encounters and property damage caused by polar bears (personal communications of Carpenter [2006]; Solomon [2006]; Kapashesit [2006]). In the past 5 years, polar bears have also been observed to travel more frequently during the open water season all the way to the Moosonee area of southern James Bay (approximately 1 sighting per year). Previously, bears were observed around Moosonee roughly once in 5 or 6 years (personal communications of Kapashesit [2006]; Solomon [2006]). Explanations offered for observations of higher numbers of bears include potential immigration of bears in response to increased ringed seals in the region, an extended ice floe in the area, and hunting quotas below the maximum sustainable yield.

Boundaries of the Southern Hudson Bay subpopulation of polar bears are currently based on data from movements of marked bears of all sexes and telemetry studies of females (Jonkel et al. 1976; Kolenosky and Prevett 1983; Kolenosky et al. 1992; Taylor and Lee 1995). Crompton (2004) suggests that the current boundaries that define the Southern Hudson Bay subpopulation may need to be revisited, as she observed at least three breeding groups in the southern portion of Hudson Bay (including James Bay).

Results of Obbard et al. (2007) suggest that contrary to results coming from Western Hudson Bay (Section 6.1, 7.10)--although not suggesting a large increase consistent with ATK (above)--there has been no observable decline in abundance of polar bears in Southern Hudson Bay since the 1980s. A recent analysis of data presented inKolenosky et al. (1992) using new mark-recapture software estimated abundance of polar bears in Southern Hudson Bay as 641 (95% CI: 401–881) in 1986 and 681 (95% CI: 401–961) in 2005. These estimates are lower than previously stated for the Southern Hudson Bay (e.g., 1,000 bears), and are likely an underestimate because of lack of complete coverage of the population (e.g., areas in James Bay). Stirling et al. (2004), in their recent analysis of coastal survey data, also suggested that the abundance of polar bears in Southern Hudson Bay has remained unchanged in recent years.

Stirling et al. (1999) contend that climate-related reductions in sea ice appear to have resulted in declines in body condition and in reproduction in the adjacent Western Hudson Bay subpopulation of polar bears (Sections 6.1 and 7.10). A similar pattern of decline in body condition was documented for the Southern Hudson Bay subpopulation when comparing bears captured in 1984–1986 with those captured in 2000–2004 (Obbard et al. 2006, 2007; PBTC 2006). Lucassie Arragutainaq of Sanikiluaq reported in August, 2006 (Arragutainaq 2006) that, although the animals look healthy and have nice fur, the fat of polar bears in the area no longer has the same consistency as in years previous. Although overall abundance in Southern Hudson Bay appears to have been stable since the 1980s, it is unknown to what extent changes in body condition might impact demographic parameters and thus abundance in the future.

7.12 Kane Basin

Based on movements of adult females equipped with satellite radio-collars and recaptures of tagged animals, the boundaries of the Kane Basin subpopulation include the North Water Polynya (to the south), and Greenland and Ellesmere Island to the west, north, and east (Taylor et al. 2001). Polar bears in Kane Basin do not differ genetically from those in Baffin Bay (Paetkau et al. 1999; Tables 1 and 2). Prior to 1997, this subpopulation was essentially unharvested in Canadian territory because of its distance from Grise Fiord, the closest Canadian community, and because conditions for travel in the region are typically difficult. However, bears from this subpopulation have occasionally been harvested by hunters from Grise Fiord (since 1997) and harvest continues on the Greenland side of Kane Basin. In some years, Greenland hunters also harvest polar bears in western Kane Basin and Smith Sound (Rosing-Asvid and Born 1990).

Few polar bears were encountered along the Greenland coast between 1994 and 1997, possibly because of harvest pressure by Greenland hunters. The current and only estimate of the Kane Basin subpopulation is 164 bears (SE = 35; Taylor et al. 2008a). The best estimate of the Greenland kill is 10 bears/year during 1999–2003 (Born 2005; Born and Sonne 2005). However, the actual number being taken by Greenland hunters is uncertain (Rosing-Asvid 2002; Born and Sonne 2005) and needs to be validated. The Canadian quota for this subpopulation is 5 bears/year. The annual combined Canadian and Greenlandic take of 10–15 bears from this subpopulation is unsustainable (Table 6). Although the habitat appears suitable for polar bears on both the Greenland and Canadian sides of Kane Basin, the density of bears on the Greenland side is much lower than on the Canadian side.

Co-management discussions regarding the hunting of polar bears have been ongoing between Greenland and Canada. Greenland enacted a quota system on January 1, 2006 (West Greenland harvest is not to exceed 100 bears/year, PBTC 2006); however, because Kane Basin, Baffin Bay (Section 7.13), and Davis Strait (Section 7.14) are treated as a single unit for management purposes by Greenland, it is unclear whether reductions in the harvest of bears in Kane Basin will result from the establishment of this quota. The mean kill of polar bears in Kane Basin has been 10 bears/year for hunters of Greenland in recent years, and <1 for hunters of Nunavut (PBTC 2006).

7.13 Baffin Bay

Based on movements of adult females equipped with satellite radio-collars and recaptures of tagged animals, the area in which the Baffin Bay subpopulation occurs is bounded by the North Water Polynya to the north, Greenland to the east, and Baffin Island to the west (Taylor and Lee 1995; Taylor et al. 2001). A relatively distinct southern boundary at Cape Dyer (Baffin Island) is evident from the movements of tagged bears (Stirling et al. 1980) and recent movement data from polar bears monitored by satellite telemetry (Taylor et al. 2001). A study of micro-satellite variation did not reveal any genetic differences between polar bears in Baffin Bay and Kane Basin, although bears of Baffin Bay differed significantly from those of Davis Strait and Lancaster Sound (Paetkau et al. 1999; Tables 1 and 2). An initial subpopulation estimate of 300–600 bears was made by the Government of the Northwest Territories from mark-recapture data collected in spring of 1984–1989. However, recent work has since shown that an unknown proportion of the subpopulation was typically offshore during the spring and, therefore, unavailable for capture. A second study (1993–1997) was carried out annually during the months of September and October, when all polar bears were on shore in summer retreat areas on Bylot and Baffin islands (Taylor et al. 2005). Taylor et al. (2005) estimated the number of polar bears in Baffin Bay (1998 estimate) at 2,074 bears (SE = 266).

The Baffin Bay subpopulation of polar bears is shared with Greenland, which until January 2006, did not limit the number of bears killed in a year. Based on mark-recapture sampling, Taylor et al. (2005) estimated the Greenland annual removal at 18–35 bears for the period 1993–1997. However, Born (2002) reported that the estimated Greenland average annual catch of polar bears from Baffin Bay was 73 bears/year over the period 1993–1998. Greenland documents that the average kill by Greenland hunters in Baffin Bay for the period 2002–2007 was 147 bears/year (range: 75–206 bears/year; PBTC 2008).

The 2004 estimate of <1,600 bears is based on population simulations (similar to the most recent estimates of abundance in Viscount Melville Sound and M’Clintock Channel; Sections 7.4 and 7.7) that employed the pooled Canadian and Greenland harvest records since 1998 (Table 6; PBTC 2006, 2007). Greenland adopted a quota system effective January 1, 2006, which should see the elimination of extremely high harvests like that of 2002/2003 (206 polar bears). However, assuming that 75­–85 bears taken per year in Baffin Bay will comprise the 100-bear West Greenland quota (historical Greenlander harvest of bears in Kane Basin and Davis Strait has been 5–25 and <5 bears, respectively [PBTC 2008]; last year’s Greenlander take in Baffin Bay under the new quota was 75 bears), the current Nunavut-Greenland harvest will equal approximately 185 bears/year. Simulations suggest that this level of hunting will continue to deplete the subpopulation (Table 6, Taylor et al. 2005).

Contrary to scientific estimates and similar to the situation in Western Hudson Bay (Section 7.10), Inuit have reported higher abundances of polar bears in Baffin Bay in recent years. ATK from 3 Baffin Bay communities (Pond Inlet, Clyde River and Qikiqtarjuaq) indicates that hunters and residents have been seeing more polar bears on the land and around communities in the past few years compared to 10–15 years ago (Dowsley 2005). Significantly more people in the 2 northern communities experienced this increase compared to people in Qikiqtarjuaq (Dowsley 2005). Bear encounters have increased, especially in Pond Inlet and Clyde River, and safety concerns have grown, as well as concerns about damaged property (Dowsley and Taylor 2006a). In response to community suggestions that polar bears increased in abundance in recent years, the Government of Nunavut increased its quota in Baffin Bay from 64 to 105 bears in December of 2004.

Despite the above, the best available scientific information suggests the Baffin Bay subpopulation is substantially over-harvested. The discrepancy between ATK and scientific data regarding the trajectory of the Baffin Bay subpopulation of polar bears is a matter of concern. Local observations of increased abundance may again be due to higher levels of bear activity in response to increased time spent on-shore by polar bears in response to climate warming in the region (Sections 6.1 and 6.2). Movements inland during summer have apparently increased in places in recent years. For example, Inuit have reported that during the open-water season bears can be found much farther into Eclipse sound, up the fiords and inlets where they did not previously occur (Dowsley 2005). Further, all 3 Baffin Bay communities have reported climate change impacts on the sea ice, such as less shore-fast ice, fewer icebergs and thinner ice, which some people (5/12 people who discussed the idea) thought might contribute to changes in polar bear distribution (Dowsley 2005; Dowsley and Taylor 2006a).

7.14 Davis Strait

Based on movements of tagged animals and, more recently, of adult females with satellite radio-collars, the Davis Strait subpopulation is comprised of bears from the Labrador Sea, eastern Hudson Strait, Davis Strait south of Cape Dyer, and along the eastern edge of the Davis Strait-southern Baffin Bay pack ice (Taylor et al. 2001). When bears occur in the latter area they are subject to hunting by Greenlanders (Stirling and Kiliaan 1980; Stirling et al. 1980; Taylor and Lee 1995; Taylor et al. 2001).

The initial subpopulation estimate of 900 bears for Davis Strait (Stirling et al. 1980) was based on a subjective correction from a mark-recapture estimate of 726 bears, which was felt to be too low. Densities of bears were substantially higher in eastern Davis Strait than in the Foxe Basin subpopulation in the survey of the Quebec coast by Crête et al. (1991). In 1993, the Federal/Provincial/ Territorial Polar Bear Technical Committee increased the estimate to 1,400 bears to account for bias in sampling created by the inability of researchers to survey the extensive area of offshore pack ice (M.K. Taylor, Department of Environment, Government of Nunavut). A population inventory by the Governments of Nunavut and Newfoundland and Labrador commenced in 2005. Following 2 years of mark-recapture sampling, a sex/age stratified Lincoln/Peterson estimate of 2,100 bears was developed (Peacock et al. 2006), which is presented in Table 6. No confidence limits are available for this estimate. The population inventory that was begun in summer of 2005 will conclude in summer 2007. Rates of survival and recruitment and final abundance estimates with confidence intervals will be available sometime in 2008.

Within Canada, bears of Davis Strait have traditionally been harvested by Inuit from Nunavut, Quebec, and Labrador (e.g., Kemp 1976; Val 1976; Brice-Bennett 1977; Brazil and Goudie 2006). The combined harvest by these jurisdictions averaged 58.6 over the past 5 years (Table 6). The Greenlander take for this subpopulation remains relatively low (1–11 bears/year in the past 5 years; PBTC 2007). Co-management discussions between Greenland and Canada are continuing, and Greenland has indicated its quota for Davis Strait will be 2 bears/year.

Qualitative observations from elders with considerable knowledge of polar bears in Nain indicated that abundance in Davis Strait was higher now than in the past (Nunatsiavut Government 2006); however, these elders also report that polar bear distribution has changed from primarily coastal and offshore areas to now also include the inland portions of bays (which freeze first), so that bears occur farther inland than previously. How climate change may be impacting polar bears in Davis Strait, and whether Inuit suggestions of higher abundances may contradict scientific estimates of trends in abundance as in Western Hudson Bay and the adjacent Baffin Bay, is unknown. However, like all areas where there is complete loss of sea ice in summer (Figure 5), increased the length of the open-water season due to climate warming is a concern for bears of Davis Strait.

Page details

Date modified: