Skip booklet index and go to page content

COSEWIC Assessment and Update Status Report on the Atlantic Walrus in Canada

Designatable Units

Use of the terms “stock” and “population” in the walrus literature has been inconsistent and the basis of delineation (management or genetic) has often not been stated (Stewart 2002). DFO has in the past delineated four stocks for hunt management purposes on the basis of geographical distributions, genetics and lead isotope data. These populations inhabit: 1) South and East Hudson Bay, 2) Northern Hudson Bay-Davis Strait, 3) Foxe Basin, and 4) Baffin Bay (High Arctic) (Figure 4) (Born et al. 1995; Outridge and Stewart 1999; Stewart 2002; Outridge et al. 2003). Born et al. (1995) provide a detailed discussion of the seasonal distribution of walruses within each of these supposed populations.

Figure 3: Place Names Used in Text

Place names used in text.

Figure 4: Approximate Summer and Winter Distributions of the South and East Hudson Bay (A), Northern Hudson Bay-Davis Strait, (B), Foxe Basin (C), and Baffin Bay (High Arctic) (D) Atlantic Walrus Populations in Canadian Waters

Approximate summer and winter distributions of the South and East Hudson Bay (A), Northern Hudson Bay-Davis Strait, (B), Foxe Basin (C), and Baffin Bay (High Arctic) (D) Atlantic walrus populations in Canadian waters.

Question marks (?) indicate uncertainty with respect to distributions and/or movements.

A fifth “Northwest Atlantic” or “maritime” population was once abundant along the Atlantic coast of Canada and in the Gulf of St. Lawrence but has long since been extirpated (Reeves 1978; Richard and Campbell 1988).

These populations may have been contiguous in the past. The degree of genetic exchange between them is uncertain and each may consist of local sub-units that mix little if at all (Outridge et al. 2003). Some recent research has heightened doubts as to the homogeneity of some of these stocks; and in particular has raised concerns that some local hunts may be supported by smaller numbers than formerly believed.

COSEWIC faces, and will continue to face, problems of defining ‘designatable units’ for marine mammals--which may or may not be the same as management stocks--as the techniques available for analysis of population structure become ever more refined. Five populations of Atlantic walrus ranging from Nova Scotia to the high Arctic are recognized for managing hunting based on geographical distributions, genetics and lead isotope data. However, COSEWIC does not feel that the data about population structure is sufficient at this time to recognize and assess these populations as separate designatable units.

Scientific and community knowledge about Atlantic walrus are collected and compiled by management units called stocks or populations. Thus information used to assess the status of walrus in Canada is organized by the five populations recognized and used to manage walrus. However, COSEWIC does not consider these populations to be designatable units.

South and East Hudson Bay Population

The South and East Hudson Bay population is distributed over an area of about 65 000 km2, from the Ottawa Islands south to the Ekwan Point area of western James Bay (Figure 4). Separation from the Northern Hudson Bay–Davis Strait population has been inferred on the basis of geographical distributions, changes in abundance, and lead isotope ratios. There appears to be a gap in the distributions of these populations between Mansel Island and the Ottawa Islands. The apparent decline in abundance of walruses in the South and East Hudson Bay population has not been accompanied by a similar decline in the Coats Island area, which suggests that immigration from Hudson Strait or northern Hudson Bay is limited (Born et al. 1995). Differences in the ratios of lead isotopes (208Pb/207Pb) in the teeth of animals harvested by hunters from Akulivik and Inukjuak also support separation of these populations (Outridge and Stewart 1999; Outridge et al. 2003). These measurements compare the lead accumulated within the walrus’s teeth over its lifetime and indicate that, on average, the animals harvested by these communities inhabit geochemically different habitats over most of their lifetimes. Akulivik traditionally hunts walruses from Nottingham Island and Inukjuak mostly from the Ottawa, King George, or Sleeper archipelagos (Olpinski 1993; Portnoff 1994; Reeves 1995; Brooke 1997).

That the South and East Hudson Bay population is distinct from the Foxe Basin population is supported by differences in their organochlorine signatures (Muir et al. 1995), in metal concentrations (Wagemann and Stewart 1994), and lead isotope ratios (206Pb/207Pb and 208Pb/207Pb) (Outridge and Stewart 1999). However,high-precision analyses of lead isotope signatures in the walrus teeth, using thermal ionization mass spectrometry (TIMS) (see also Evans et al. 1995), suggest that part of the group formerly harvested in Foxe Basin by hunters from Hall Beach may have moved southward intoeastern Hudson Bay (Stewart et al. 2003), perhaps into the Sleeper Islands.

The relationship between walruses in the Sleeper and Belcher archipelagos of eastern Hudson Bay and those to the south at Cape Henrietta Maria and inside James Bay is unknown.

Northern Hudson Bay–Davis Strait Population

The Northern Hudson Bay–Davis Strait population is distributed over an area of about 385 000 km2, from Arviat on the west coast of Hudson Bay north and east through Hudson Strait to Clyde River on the east coast of Baffin Island (Figure 4) (Richard and Campbell 1988; Born et al. 1995; Stewart 2002). Born et al. (1995) considered animals at Digges and Mansel Islands to be part of this population. The rationale for distinguishing it from the South and East Hudson Bay population is discussed above.

Distinction of the Foxe Basin population from the Northern Hudson Bay–Davis Strait population is based on distance and on differences in growth patterns and lead isotope ratios. Distance may not completely separate these populations but it must limit interchange between them. The gap between them (Figure 4) is putative, since the seasonal distribution in southeastern Foxe Basin is poorly known. There is some north-south movement of walruses in Foxe Basin but no evidence of concerted movement to or from Hudson Strait (Anderson and Garlich-Miller 1994). Walruses winter in both areas, so they are not moving to seek wintering habitat. Animals sampled from Foxe Basin in the 1980s and 1990s were significantly larger than those sampled from northern Hudson Bay in the 1950s (Garlich-Miller and Stewart 1998). This suggests genetic or habitat separation of animals in these areas. The lead isotope ratios (206Pb/207Pb and 208Pb/207Pb) in teeth of walruses harvested by hunters from Akulivik and Coral Harbour are different from those of walruses harvested from northern Foxe Basin (Outridge and Stewart 1999; Outridge et al. 2003). However, lead isotope signatures in the walrus teeth suggest that part of the group harvested by Hall Beach hunters may move southward into northeast Hudson Bay (Outridge et al. 2003; Stewart et al. 2003).

The Northern Hudson Bay-Davis Strait population may consist of separate sub-populations that inhabit northern Hudson Bay, Hudson Strait, and Davis Strait. These distinctions are supported by measurements of lead isotope ratios (206Pb/207Pb and 208Pb/207Pb)--significantly different between Coral Harbour and Akulivik (Outridge et al. 2003)--and by morphometric observations--Inuit have noted differences in body size and tusk length between Nottingham Island (harvested by Akulivik) and Coats Island (harvested by Coral Harbour), and between the Chesterfield Inlet area and the Repulse Bay area (Fleming and Newton 2003).

Separation of the Northern Hudson Bay–Hudson Strait population from the Baffin Bay population has been inferred mostly from information on walrus distribution and movements (Born et al. 1995). There may be a gap in walrus distribution along the east coast of Baffin Island between Clyde River and Pond Inlet (Mansfield 1967). In the 1970s, hunters from Clyde Inlet travelled north to Scott Inlet to hunt walruses (Kemp 1976) and hunters from Pond Inlet travelled south to the Cape Macculloch area (Lands Directorate 1981). The intervening stretch of the east Baffin coast is remote from both communities and seldom visited in summer, and has not recently been surveyed; it is uncertain whether there are walruses there or not. Walruses in northwest Greenland have different patterns and levels of organochlorine contaminants in their blubber from those in southeast Baffin Island (Loks Land), which also suggests these animals belong to different populations (Muir et al. 2000).

The amount of exchange between the Northern Hudson Bay–Davis Strait and Central West Greenland populations is unknown. Dunbar (1955) and Vibe (1967) both suggested that such a connection might exist. It is supported by observations of walruses offshore over deep waters between southeastern Baffin Island and western Greenland (Born et al. 1994) and by a recent movement of a female tagged in West Greenland to southern Baffin I. within one month (R. Dietz, pers. comm.).

Foxe Basin Population

Walruses are widely distributed in the relatively shallow waters of northern Foxe Basin, an area of about 50 000 km2, where they live year-round (Figure 4) (Mansfield 1959; Loughrey 1959; Crowe 1969; Beaubier 1970; Brody 1976; Orr et al. 1986). Genetic analyses support the hypothesis that animals harvested from this area belong to a different population from those in the Resolute Bay–Bathurst Island area (Buchanan et al. 1998; de March et al. 2002).  This suggests that animals may not pass through Fury and Hecla Strait, a movement that both scientists and Inuit have considered to be unlikely (Loughrey 1959; Mansfield 1959; Davis et al. 1980; Garlich-Miller cited in Stewart 2002). Walruses are found on occasion in the Gulf of Boothia south to Pelly (Brice-Bennett 1976) and Committee Bays and to Crown Prince Frederik Island (70°02'N, 86°50'W) (Loughrey 1959; Anders 1966). These animals are believed to come from the Baffin Bay population to the north, either as strays (Loughrey 1959) or when ice does not break up in Barrow Strait (Riewe 1976). The reasons for distinguishing the Foxe Basin population from the Northern Hudson Bay–Hudson Strait population are discussed above.

The existence of sub-populations within Foxe Basin is not so far supported by genetic analyses, which did not find significant differences between the walruses harvested by Igloolik and Hall Beach (de March et al. 2002). However, differences in lead isotope ratios (206Pb/207Pb) in the teeth of walruses harvested by the two communities suggest quite strongly that their hunters are taking animals from different local sub-populations (Outridge et al. 2003; Stewart et al. 2003), or at least from groups with distinct feeding areas. Inuk elders recognize two groups of walruses in northern Foxe Basin on the basis of differences in size, colour, and distribution (DFO 2000).

Baffin Bay (High Arctic) Population

The Baffin Bay population is distributed over an area of about 150 000 km2 that extends west to Bathurst Island and north to Kane Basin (Figure 4) and northwest Greenland. It seems to be separated geographically from the other Canadian walrus populations to the south, as discussed earlier, but is shared by Canada and Greenland, as its distribution along the western coast of Greenland extends southward to about Disko Island (Reeves 1978; Richard and Campbell 1988; Born et al. 1995). Genetic studies have found significant genetic separation between animals from this population and those of the central West Greenland population to the south (Andersen and Born 2000). They suggest that these populations may be connected by some migration of males but that female migration is very restricted. Walruses do move between Canadian and Greenland waters (e.g. Degerbøl and Freuchen 1935; Vibe 1950; Born et al. 1995), but the number of animals involved and the extent of the movements are unknown. There are extralimital reports of walruses from this population at Prince Patrick and Melville islands and near Taloyoak (Spence Bay) (Harrington 1966).

Sub-structure within this regional population has been the subject of study. Lead isotope ratios (206Pb/207Pb and 208Pb/207Pb) in the teeth of walruses from Resolute were different from those sampled near Grise Fiord and Thule, Greenland (Outridge et al. 2003). The lead isotope signatures in the walrus teeth suggest that hunters from Grise Fiord and Thule kill walruses from the same sub-population, as well as from different sub-populations. About 80% of the Grise Fiord and 20% of the Thule animals had similar isotopic signatures.

Early genetic analyses did not find a significant separation between animals in the Resolute Bay/Bathurst Island area and those in the Grise Fiord area (de March et al. 2002), but more recent--and still preliminary--analyses of microsatellites in nuclear DNA have shown some evidence, still not clear, of some differences between groups in eastern Jones Sound, western Jones Sound and west of Devon Island (R. Stewart, pers. comm.). Some of these distinctions have received limited support from tagging studies, which have not shown evidence of movements between western and eastern Jones Sound; tags have not, however, stayed on for more than 3 months.

Nova Scotia–Newfoundland–Gulf of St Lawrence (Maritime) Population

The south-western Gulf, and the Scotian shelf, have extensive shallow waters with flat sandy bottoms, and a numerous stock of walruses originally used land haul-out sites in the St Lawrence system. Sites mentioned were Seven Islands in the north-western Gulf, Miscou Island in the western Gulf, and several sites on and around the Magdalen Islands in the central Gulf. The original distribution in the St Lawrence extended as far up-river as Rivière-Ouelle. Sable Island off Nova Scotia was also frequented by considerable numbers of walruses, as was Ramea Island off Newfoundland (Born et al. 1995).

These aggregations were heavily harvested, especially in the 17th and 18th centuries, and by the end of the 18th had been extirpated. There have been occasional sightings in recent decades (Kingsley 1998; Camus 2003; Richer 2003), but these have not been considered indications of a re-establishment of this population.

Changes in Seasonal Distribution

There has been a general shift in walrus distribution away from communities to areas that are less accessible (Born et al. 1995). This is not a new phenomenon. It is related to changes in technology that have enabled hunters to range further from home (Brody 1976). It began with the introduction of whaleboats in the 1920s, which extended hunting ranges and enabled open-water hunting; accelerated with the introduction of motorized technology ca. 1940-60; and continues as the range and speed of boats increase (see also Crowe 1969; Beaubier 1970; Orr et al. 1986). The extent to which distributional changes reflect declines as opposed to shifts in the walrus populations is unknown but, until increases in other parts of the range have been documented, it is prudent to assume numbers have been reduced (DFO 2000).

Inuit around Hudson Bay have linked the disappearance of walruses from traditional hunting areas variously to natural shifts in the species’ distribution, to poor and wasteful hunting techniques, and to low harvest rates (Fleming and Newton 2003). In the past, unregulated hunting from motorboats disturbed animals at uglit (haul-outs) in the Belcher and Sleeper islands, and along the west coast of Hudson Bay--possibly with high mortality. Walrus remains were sometimes discarded at the uglit and, together with sinking losses, tainted both the uglit and feeding grounds causing herds to leave the area. Inuit recognize the sensitivity of walruses to habitat disturbance and to the mortality of other walruses in their traditional knowledge.

"When I was growing up, I remember, my father and the others used to say never try to kill a walrus where you think it will sink right into the feeding areas, or never cut up the walrus where they usually bask or rest. The elders used to say never to leave the guts near the islands where they bask. If you do that the walrus will move away from there." (Zach Novalinga, Sanikiluaq).

Some eastern Hudson Bay Inuit have, however, suggested that declines in the number of walruses may stem from harvest rates that are too low to maintain the reproductive rate (Fleming and Newton 2003).

South and East Hudson Bay Population. Walruses were once common in the archipelagos of south and east Hudson Bay (Flaherty 1918; Twomey and Herrick 1942; May 1942; Manning 1946; Freeman 1964; Schwartz 1976; Born et al. 1995). They have not been common in the Belchers since about the late 1950s, and unsuccessful hunting expeditions (see Olpinski 1993; Portnoff 1994) suggest use of uglit in the Ottawa Islands may have declined in the 1990s. In 1993, walruses re-occupied some uglit in the North Belcher Islands and Inuit believed the population might be recovering (Z. Novalinga, Sanikiluaq Environmental Committee and Peter Kattuk, Mayor of Sanikiluaq, pers. comm. 1993). However, DFO (2000) reports that walruses continue to be uncommon in the Belchers. To hunt walrus in fall, Inuit from the area generally travel to the Sleeper Islands.

The walrus distribution in James Bay is much reduced. Along the east coast in the Wemindji area, at Wiipichuutukuwiih, walrus were once numerous and posed a hazard to paddlers (Fleming and Newton 2003). Cree hunted them in the Wemindji-Waskaganish area until at least 1934. Geographical names suggest that they also hauled out at “Walrus Point” and “Pte. du Morse” near Chisasibi but none have been seen there recently. In the west, walruses occurred south to Attawapiskat. They were seen in the early 1960s on ice floes between Lakitusaki River (Lake River) and Bear Island (Johnston 1961), and in the 1970s at Ekwan Point (Fleming and Newton 2003). Residents of Attawapiskat saw fewer walruses in their area in the early 1990s but reported that they were present on the mainland between Akimiski Island and Ekwan River after spring breakup.

Walruses are still present along the Ontario coast of Hudson Bay near Cape Henrietta Maria. In 1993, local Cree reported that there were “lots” of walruses in the area and that they had been seen in July near Peawanuck (Fleming and Newton 2003).

Northern Hudson Bay-Davis Strait Population. The main changes in seasonal distribution of this population occurred in the early to mid-1900s. They include abandonment of uglit along the west coast of Hudson Bay north to Chesterfield Inlet, on Digges Island in northeast Hudson Bay, near the head of Cumberland Sound, and on the Gyrfalcon Islands (59°05'N, 68°57'W) in southern Ungava Bay (Born et al. 1995).

In western Hudson Bay, walruses were rare at Churchill, but were increasingly numerous moving northward where the coastline offered more suitable uglit. Six animals were seen off the coast near Cape Churchill in October 1954 (Johnson in Loughrey 1959). Walruses are uncommon near Whale Cove but were numerous at islands near the community from 1942 to 1945 (Fleming and Newton 2003). They have abandoned various uglit in western Hudson Bay but hauled out in small numbers in summer at Bibby Island (61°53'N, 93°05'W), Term Point (62°08'N, 92°28'W), “Little Walrus Island” in Mistake Bay, Sentry Island (61°10'N, 93°51'W), Wag Island (63°23'N, 90°38'W), Marble Island (62°41'N, 91°08'W), and Fairway Island (63°15'N, 90°33'W) as recently as the 1950s (Low 1906; Degerbøl and Freuchen 1935; Loughrey 1959; Reeves 1978; Born et al. 1995; DFO 2000; Fleming and Newton 2003) (Figure 4). Small groups of walruses are sometimes seen at the floe edge south to Whale Cove (Gamble 1988; Fleming and Newton 2003). Inuit report that walruses were more numerous in the Chesterfield Inlet area in the early 1990s than in the past (Fleming and Newton 2003).

Bell (1884: 33DD) found numerous walruses at Digges Island in the 1880s, but they are apparently rare in the region now (Born et al. 1995). They were common at the head of Cumberland Sound in the 1800s (Kumlien 1879) and were killed in the Kingnait Fiord area in the early 1900s (Anders et al. 1967), but now are uncommon in these areas. Walruses were once common in summer at the Gyrfalcon Islands in southern Ungava Bay (Dunbar 1955), and on islands in Deception Bay, near Sugluk (Loughrey 1959).

In the early 1900s the largest ugli on the east coast of Baffin Island may have been at Padlei, just south of Padloping Island (Mansfield 1958; Reeves 1978). Many walruses were harvested from the area, which is still used but apparently by fewer animals.

Foxe Basin Population. The summer distribution of walruses in Foxe Basin has changed over the past 50 years (Anders 1966; Crowe 1969; Beaubier 1970; Brody 1976; Orr et al. 1986). Until the 1940s or 1950s, walruses were abundant at uglit along the east coast of Melville Peninsula. The establishment of a Hudson’s Bay Company post at Igloolik in 1939 and a DEW Line Station at Hall Beach in 1955–1956 congregated the Inuit of Foxe Basin in this area. This increased local hunting pressure on the walruses and disturbance by boat traffic, and they are now less common. Hunters from Igloolik and Hall Beach do not believe that the number of walruses in northern Foxe Basin has changed over the past 25 years (DFO 2000). They now travel further offshore to hunt in the summer, but attribute this change to a reduction in the ice cover. Three islands in western Foxe Basin continue to have large concentrations of animals in the fall, and three others that had been used for some time are now being used again, but uglit along the east coast of Melville Peninsula have not been reoccupied.

Baffin Bay (High Arctic) Population. The main change observed in seasonal distribution of this population has been in the Avanersuaq (Thule) area of west Greenland, where walruses were once abundant in summer but are now absent (Vibe 1950; Born et al. 1995).

Nova Scotia–Newfoundland–Gulf of St Lawrence (Maritime) Population. Exploitation of these herds continued for over a hundred years spanning the 17th and 18th centuries, but the records available do not give clues about changes in distribution as a result of this heavy harvesting.