Polar bear (Ursus maritimus) COSEWIC assessment and status report: chapter 3

3. Distribution

3.1 Global Range

Polar bears are circumpolar in the northern hemisphere. Initially, polar bears were believed to occur as a single, homogeneous population that ranged throughout the circumpolar Arctic, with animals carried passively on sea ice by the predominant currents (Amstrup 2003). However, recent studies based on satellite telemetry and mark-recapture demonstrate that, although some movements may be exceptionally large (Messier et al. 2001), polar bears do not wander as nomads throughout the Arctic, but rather show seasonal fidelity to local areas (Taylor and Lee 1995; Bethke et al. 1996; Taylor et al. 2001). Today, the circumpolar distribution of the polar bear is usually divided into 19–20 “subpopulations” of which 13 (excluding bears of the Arctic Basin, which are at low density and thought to generally support transient bears) range into or are entirely contained within Canada (Figures 1, figure2, figure3; Taylor et al. 2001).

3.2 Canadian Range

In North America, the area of occupancy (Figure 3) of the polar bear extends from the southern edge of the permanent multi-year pack ice of the Arctic Ocean (Arctic Basin) to include sea ice and coastal areas of Greenland, the Canadian Arctic Archipelago, east to the Labrador coast, south to James Bay, and west to the Bering Sea. A few polar bears regularly appear as far south as the island of Newfoundland. Bears have occasionally been noted in the Gulf of St. Lawrence in years when thick pack ice drifts farther south than normal. These observations are included in the species’ extent of occurrence (Figure 3). The present area of occupancy of the polar bear in Canada appears to correspond to the species’ historic area of occupancy after the end of the last major glacial retreat (10,000 years ago), although some authors suggest that polar bears were more common in southern Labrador and Newfoundland in previous centuries than they are today (Smith et al. 1975; Stirling and Kiliaan 1980). Smith et al. (1975) communicates that early explorers commonly encountered polar bears in Newfoundland, and bears were sometimes observed in high densities in southern Labrador (e.g., a congregation of 32 bears observed by Cartwright in the 1770s feeding on salmon at the mouth of the Eagle River). Historic reductions in the abundance of polar bears in southern Labrador and Newfoundland may have been a consequence of several hundred years of human habitation and associated hunting, or a response to climate warming after climax of the cooling events commonly referred to as the “Little Ice Age,” particularly the substantial glacial (and sea ice) advances in North America of 1711–1724 and 1835–1849 (Lamb 1977, p. 453).

3.3 Designatable Units

Delineation of subpopulations in Canada (Figure 2) has been largely based on hierarchical cluster analysis of movements of radio-collared females (Bethke et al. 1996; Taylor et al. 2001; Amstrup et al. 2004). Within most subpopulations, population dynamics appear to be determined from internal rates of birth and death (mainly from harvest, which is set at different levels according to subpopulation), rather than emigration/immigration, although annual rates of exchange between adjacent subpopulations range from 0.4–8.9% (Taylor et al. 2001). Genetic distances between sampled individuals from subpopulations based on FST (a correlation of allele frequencies between populations [Weir and Cockerham 1984]) suggest the possibility of 4 population clusters among identified subpopulations (Table 1 and Paetkau et al. 1999); however, misclassification rates in assignment tests among clusters and subpopulations (Table 2) do not support definitive boundaries in terms of genetic isolation across the range of the polar bear (Paetkau et al. 1999). Values of pairwise FST are low: among all Canadian subpopulations the highest observed difference in pairwise FST by Paetkau et al.(1999) was 0.091 (Southern Beaufort Sea vs. Foxe Basin), with a Canada-wide mean pairwise FST of 0.039 (Table 1). The latter is relatively small in comparison to mean pairwise FST data from other species of carnivores in North America, including populations of non-hybridizing gray wolves, Canis lupus, and coyotes, Canis latrans(0.168 and 0.107, respectively; Roy et al. 1994); grizzly bears in the central Rocky Mountains (0.096; Proctor et al. 2005), and wolverines, Gulo gulo (0.067; Kyle and Strobeck 2001). When comparing individuals of pairs of widely separated subpopulations of black (Ursus americanus), brown, and polar bears, polar bears exhibited the lowest values of intraspecific genetic distance, DLR (Figure 5 of Paetkau et al. 1997). Among adjacent potential subdivisions of polar bears identified in Paetkau et al.(1999; Table 1), FST differences of included subpopulations ranged from 0.024 to 0.061. The data of Paetkauet al.(1999) strongly support the hypothesis of a polar bear population that--despite the presence of regional differences in dynamics and environmental conditions--maintains considerable genetic interchange among subpopulations, with a gradation in genetic relatedness across the range. No localized adaptations have led to the genetic isolation of any subpopulation: identified units are not evolutionarily significant (Paetkau et al.1999).

Figure 3. The extent of occurrence (bold polygon, 8.7 × 106 km²) and area of occupancy (shaded region with dotted outline, 5.6 × 106km²) of polar bears in Canada (areas estimated using Lambert Equal-Area Azimuthal [North Pole] projection in ArcGIS, v. 9.1 [ESRI, 380 New York Street, Redlands, CA]). Note that these areas include all freshwater, ocean, and land in the computation of area. The extent of occurrence as defined by COSEWIC (COSEWIC's Assessment Process and Criteria, reviewed and approved by COSEWIC in April 2006) is: “the area included in a polygon without concave angles that encompasses the geographic distribution of all known populations of a species (up to and including the international border).” The area of occupancy is defined as the: “area within 'extent of occurrence' that is occupied by a taxon, excluding cases of vagrancy.”

Figure 3. The extent of occurrence (bold polygon, 8.7 × 106 km2) and area of occupancy (shaded region with dotted outline, 5.6 × 106 km2) of polar bears in Canada (areas estimated using Lambert Equal-Area Azimuthal [North Pole] projection in ArcGIS, v. 9.1 [ESRI, 380 New York Street, Redlands, CA]).

Table 1. Genetic distances among Canadian and worldwide subpopulations of polar bears ( see legend of Figure 2 for abbreviations for Canadian subpopulations [identified in bold typeface]) presented in Paetkau et al. (1999): F ST (× 100) below diagonal, D LR above. Highlighted rectangles identify distances within 4 suggested population clusters by Paetkau et al. (1999). F ST is a correlation of allele frequencies between populations (Weir and Cockerham 1984) and D LR is the mean genotype log likelihood ratio across individuals from the 2 subpopulations (Paetkau et al. 1997). Source: table reproduced from Paetkau et al. (1999) and © 1999 Blackwell Science Ltd.
Location WH FB DS BB KB LS GB MC VM NW NB SB CS FN SV EG
WH
-
0.5
1.4
3.9
3.6
3.9
3.5
5.2
4.4
4.3
5.4
5.9
6.6 5.6 5.0 5.9
FB
0.9
-
1.1
3.3
3.3
3.9
4.1
4.7
4.7
4.9
6.3
6.4
7.8  5.5  5.3  5.8 
DS
2.1
1.4
-
1.6
2.1
2.6
2.4
2.7
3.0
3.4
3.7
4.6
4.9 3.1 3.2 3.8
BB
5.2
4.4
2.6
-
0.1
0.9
1.3
1.5
1.7
2.5
3.0
3.5
3.9 2.8 2.7 2.7
KB
5.6
4.8
3.6
0.3
-
1.1
1.5
1.7
1.7
2.2
2.7
3.3
4.0 3.0 2.3 2.2
LS
5.4
5.2
3.9
1.0
1.1
-
0.7
0.8
0.8
1.7
2.5
3.4
4.5 3.0 2.8 2.6
GB
4.8
5.1
3.6
1.4
2.1
1.0
-
1.1
1.3
2.6
2.5
2.8
3.8 2.8 3.0 3.1
MC
5.7
5.2
3.2
0.9
1.3
0.5
1.1
-
0.9
3.7
2.5
3.7
3.9 2.6 2.4 2.7
VM
6.1
6.3
4.3
1.9
2.4
1.7
1.9
0.8
-
2.2
1.6
2.3
3.6 2.5 2.6 2.5
NW
6.8
7.5
5.8
3.5
2.4
2.4
3.7
3.9
3.5
-
3.4
4.3
5.4 4.5 4.4 4.1
NB
8.1
8.9
6.5
3.8
3.8
3.9
3.3
2.7
2.6
5.2
-
0.5
0.9 1.3 1.0 1.3
SB
8.6
9.1
7.4
4.6
4.9
5.8
4.3
4.2
3.7
6.3
0.5
-
0.9 1.4 1.8 2.1
CS 9.5 10.8 8.2 5.3 6.0 6.8 5.1 5.0 5.5 7.9 0.8 1.0 - 1.2 1.1 1.9
FN 8.5 8.4 5.3 3.9 4.6 5.1 4.4 3.3 3.7 6.8 1.4 1.8 1.7 - 0.0 0.4
SV 7.6 7.9 5.3 3.1 3.2 4.1 3.9 2.5 3.5 5.9 1.0 2.2 1.3 0.2 - 0.4
EG 8.7 8.8 6.5 3.4 3.2 4.2 4.2 3.0 3.8 5.8 1.4 2.6 2.1 0.8 0.3 -

 

Table 2. Results of the population assignment test of Paetkau et al.(1999) based on microsatellite analysis of individuals sampled from the world’s subpopulations of polar bears (abbreviations defined in the legend of Figure 2 for Canadian subpopulations [identified in bold typeface]). Each row contains the samples from a study area (‘‘subpopulation’’) and the columns indicate the subpopulations to which these samples were ‘‘assigned’’ (in which their genotypes had the highest likelihood of occurring). For example, of polar bears captured in Davis Strait (DS), 13 animals were correctly assigned to that subpopulation unit by genetic analysis, whereas 17 individuals captured in Davis Strait were assigned to alternate subpopulations. Highlighted rectangles present the 4 suggested population units by Paetkau et al.(1999), as in Table 1. Source: table reproduced from Paetkau et al.(1999) and © 1999 Blackwell Science Ltd.
Location WH FB DS BB KB LS GB MC VM NW NB SB CS FN SV EG Total
WH
21
6
5
  1     33
FB
9
16
3
1
1
        30
DS
1
5
13
4
2
2
1
1
  1     30
BB
3
13
7
2
1
2
1 1   1 31
KB
1
9
9
2
2
4
1
1
      1 30
LS
2
1
2
10
2
5
4
3
  1     30
GB
1
1
4
2
3
10
3
3
2
 
1
        30
MC
2
1
2
7
3
        15
VM
1
2
1
4
2
1
14
1
4
        30
NW
1
1
2
1
3
2
2
17
  1     30
NB
1
1
1
1
11
6
6 1 1 1 30
SB
2
2
6
11
3 5   1 30
CS                     1 7 15 4 1 2 30
FN 1     1       1 2   3 3 1 8 9 3 32
SV           1 1   2   1   1 5 11 9 31
EG         2       2 1 2 1 3 5 4 11 31

Interchange among identified subpopulations is further suggested by recent Aboriginal Traditional Knowledge (ATK, Keith 2005; Nirlungayuk 2008) and results of independent genetic analyses (Crompton 2004, Saunders 2005). For example, using 9 microsatellite loci Saunders (2005, p. 39) found no support for any genetic discontinuities between M’Clintock Channel and Gulf of Boothia. Crompton (2004) additionally concluded that the subpopulation boundaries of Western Hudson Bay, Foxe Basin, and Davis Strait (Figure 2) were not supported by her study; rather, Crompton suggested that polar bears of the Hudson Bay region are structured into at least three breeding groups in the southern portion of Hudson Bay and what appears to be one larger admixed population to the north. Interchange of individuals between the subpopulations of the Southern Beaufort Sea and Northern Beaufort Sea (as currently defined; Figure 2) is also known to be considerable (Amstrup et al. 2004).

At a larger scale, Thiemann et al. (In press) recently argued that the population of polar bears in Canada should be divided by COSEWIC into 5 Designatable Units for conservation purposes. The argument for 5 units by Thiemann et al. (In press) is based partly on the results of Paetkau (1999) but also on differences in diet and ice conditions among subpopulations. The proposed units are identical to the 4 subpopulation clusters of Paetkau et al. (1999) as indicated in Table 1, with the exception that Davis Strait is separated from all other subpopulations because of differences in ecological conditions (principally higher abundances of harp seals in the region). Amstrup et al. (2007), using models of seasonal patterns of ice distribution and movements, similarly proposed regional ecoregions for polar bears for modelling purposes. The ecoregions of Amstrup et al. (2007) differ in several respects from the Designatable Units proposed by Thiemann et al. (In press), with boundaries appearing between the Northern and Southern Beaufort Sea subpopulations but not Norwegian Bay and those of the Canadian Arctic Archipelago, and pooling subpopulations with near complete loss of sea ice in summer (Baffin Bay, Davis Strait, and subpopulations of Hudson Bay are grouped together in Amstrup et al. [2007]).

Although useful for describing local trends in population growth, demographic parameters, behaviours, and managing polar bears, identified subpopulations cannot be considered Designatable Units as per COSEWIC guidelines (Appendix F5, COSEWIC Operations and Procedures Manual, last reviewed and approved by COSEWIC in 2005). Further, despite the arguments of Thiemann et al. (In press), the presence of larger-scale regional subdivisions in the polar bear population do not warrant more than a single Designatable Unit in Canada. However, this does not mean that conservation threats to polar bears are uniform across the range of the species. Because of this, and given historical management of the species on a subpopulation-by-subpopulation basis using bounds presented in Figure 2 (including different harvest rates), trends in the Canadian polar bear population are generally discussed in this report according to identified subpopulations (see Section 7).

Page details

Date modified: