Introduction

Wolverines (Gulo gulo) occur broadly across the Arctic, known by names including qavvik (Iñupiaq), qavvigaarjuk (Inuit/Inuktitut; Laugrand 2017), qafčik (Yup'ik; Bobaljik 1996), kuekuatsheu (Innu, Armitage 1992), иӈгнeй (Nenets; Li 2021), қeper (Chukchi, Weinstein 2018), ńitlә (Tundra Yukaghir; Nikolaeva 2008), and pocoмaxa (Russian). The region is often considered a stronghold for wolverines due to its relatively restricted human footprint and cold climate (Fisher et al. 2022), and the species occupies an important sociocultural role among the region’s Indigenous Peoples (Armitage 1992; Cardinal 2004; Benson 2014). Despite this, many aspects of the species’ ecology in this biome have received little attention by Western science. The paucity of available information has precluded developing formalized research priorities and evaluating potential conservation concerns. Addressing this gap is particularly important as the Arctic changes rapidly in response to anthropogenic disturbance.

The circumpolar Arctic, which we regard here as the area north of latitudinal treeline (Walker et al. 2005), is a dynamic landscape undergoing rapid change (Hinzman et al. 2005). Anthropogenic warming in recent decades has degraded sea ice, snow, and permafrost at a rapid pace, with cascading consequences for wildlife (Post et al. 2013; Berteaux et al. 2017; Huntington et al. 2020). For example, sea ice loss precludes metapopulation mixing among terrestrial island-dwelling species, such as wolves (Canis lupus; Carmichael et al. 2008) and Arctic foxes (Alopex lagopus; Geffen et al. 2007), while enabling range expansions of temperate marine mammals (Kovacs et al. 2011). Thawing permafrost is depleting waterfowl breeding habitat (Perreault et al. 2017), and shifting snowmelt phenology is altering the availability of insects to breeding migratory birds (Saalfeld et al. 2019). Species typically associated with boreal forest, including moose (Alces alces), beavers (Castor canadensis), and snowshoe hares (Lepus americanus) are shifting northward into the treeless Arctic following shrubification (Jung et al. 2016; Tape et al. 2016a, b, 2018), while the ranges of specialist Arctic species such as Lapland longspurs (Calcarius lapponicus) are contracting (Boelman et al. 2015). Disease transmission, mediated by the abiotic and biotic environments, is changing as well (Ruscio et al. 2015; Keatts et al. 2021). Accompanying and driving these changes are an expanding human footprint, largely associated with resource extraction and facilitated by advancements in over-snow travel technology (Landa et al. 2000; Kumpula et al. 2012; Raynolds et al. 2014). Such human activity can produce both positive and negative impacts to individual wildlife species (e.g., through food subsidies, habitat loss, or increased exposure to hunters; Elmhagen et al. 2017; Johnson et al. 2020).

Many aspects of wolverine ecology, including the species’ response to disturbance, are receiving increasing attention in boreal and montane environments, whereas research on this species in the Arctic lags behind. In a recent review of two decades of wolverine research across the species’ global distribution (Fisher et al. 2022), only 10 of 156 articles focused primarily on the Arctic. Although many fundamental aspects of the species’ ecology are undoubtedly similar across biomes, important differences likely exist that necessitate focused research in this region for effective management and conservation. For example, although documented wolverine reproductive dens in the montane contiguous USA occur exclusively in deep, persistent snowdrifts (Copeland et al. 2010) suggesting likely habitat loss with climate change (McKelvey et al. 2011), this habitat requirement is apparently less strict in boreal forest (Webb et al. 2016; Aronsson and Persson 2017). In the Arctic, fundamental facets of the species’ ecology may differ in accordance with the unique characteristics of this landscape, including distinct prey species, forms of structural habitat, and controls over metapopulation dynamics such as sea ice loss for island populations.

To address the lack of formalized research priorities and potential conservation concerns for wolverines in this region, we synthesize existing knowledge, including both Western research and Indigenous Knowledge (IK), and discuss possible implications of anthropogenic change for wolverines in the Arctic. We contextualize our review by updating the pan-Arctic distribution map of wolverines and discussing possible underlying drivers of Arctic wolverine population dynamics.

Methods

We sought to review all available published literature addressing wolverines in the Arctic. To accomplish this, we employed a snowballing procedure (Wohlin 2014), whereby we searched “wolverine* AND Arctic AND ‘Gulo gulo’” in Google Scholar, sorted in descending order of relevance and used the top 20 most relevant studies addressing wolverines north of latitudinal treeline as our start set. We then worked forward and backward, examining papers both cited by and citing each focal paper until we found no new studies. Additionally, to identify studies excluded using English search terms and to locate gray literature (e.g., reports from management agencies) not indexed by Google Scholar or commonly referenced in the peer-reviewed literature, we contacted regional experts across the circumpolar north. We note that, although this approach yielded some studies not identified during the snowballing procedure, our review likely remains biased toward English studies and peer-reviewed literature. All IK we discuss is derived from published reports, which we note limits the contribution of this body of knowledge to this review.

To quantify research and monitoring efforts regarding wolverines in the Arctic, we classified each source listed in References according to topic and whether or not it was published in the peer-reviewed literature. We excluded supporting sources (i.e., those not dealing directly with wolverines in the Arctic) and those used solely to translate “wolverine” into Indigenous languages (e.g., dictionaries). To quantify how this body of literature has grown and changed over time, we then summarized studies according to date published and topical area.

To update the distribution map of wolverines in the Arctic, we relied on reports arising (1) during our literature review, (2) from targeted searches for specific geographical areas, and (3) from communication with regional experts. We note that defining the Arctic by latitudinal treeline excludes nearly all of the Scandinavian Peninsula; this region is therefore not included in the present review (Fig. 2). We analyzed the updated distribution map in a Geographical Information System to calculate the portion of wolverines’ global distribution in the Arctic and the portion of the species’ Arctic distribution that occurs on islands, which we use to contextualize our discussion of island metapopulation dynamics. For non-Arctic areas of wolverines’ global distribution, we used the distribution reported by Copeland et al. (2010) since this is the most recent effort to collate expert opinion of global wolverine distribution of which we are aware. To visualize the discrepancies between currently reported Arctic distribution and our update, we use the range reported by the International Union for Conservation of Nature (Abramov 2016), since this is typically considered the authority on species’ distributions and it broadly parallels Copeland et al. (2010) for the Arctic.

With the exception of northwestern Alaska and northern Labrador (Schmelzer 2006; Poley et al. 2018), as well as snow track surveys near mines in the Northwest Territories (NWT; Golder Associates Ltd. 2017; De Beers 2018), we are not aware of any systematic effort to survey distribution of wolverines across the Arctic. Therefore, evaluating wolverine presence relies mostly on opportunistic reports or harvest (i.e., hunting or trapping) by local residents, land users, and researchers, as well as a few systematic studies focusing on other aspects of wolverine biology (e.g., remote camera grids to measure population density; Mulders et al. 2007; Awan and Boulanger 2016; Awan et al. 2020). We erred conservatively in designating presence, since wolverines occur naturally at low densities and many remote regions of the Arctic are visited only sporadically by people. Specifically, we designated areas to be within the species’ distribution if reproduction had been documented or if the region was consistently considered occupied by wolverines among local land users or regional experts. We defined an intermediate level (“occasionally observed”) for areas with at least one verifiable report but insufficient observations to support the presence of a reproductive population. For islands, we designated wolverine occupancy on a per-island basis and did not attempt to determine within-island variation (except Greenland). Data and citations for distribution on Arctic islands are presented in Fig. 2 and Online Resource 1. This map, while reflective of the current state of knowledge, is subject to biases associated with obtaining presence records for low-density, wide-ranging species in a vast region with relatively few people. Specifically, we suspect that wolverines may occur in some regions where no records exist and that some regions currently reported to support wolverines may no longer do so.

Bibliometrics

We identified 59 peer-reviewed studies and 32 reports, books, or theses dealing directly with wolverines in the Arctic (Table 1, Fig. 1). Of these, the category “harvest” was most represented (n = 18 across publication types), followed by “parasitism, disease, and toxicology” (n = 15), and occupancy/population density (n = 13, Table 1). Studies published in the peer-reviewed literature generally became more abundant in the late 1990s with a recent peak in 2021. Impacts of human development to wolverines have received particularly little attention (n = 3 across publication types), as have diet (n = 7) and sociocultural role/Indigenous knowledge (n = 6).

Table 1 Number of studies regarding wolverines (Gulo gulo) in the Arctic published between 1943 and 2022
Full size table
Fig. 1
figure 1

Peer-reviewed studies of wolverines (Gulo gulo) by subject area and year of publication in the Arctic

Full size image

Sociocultural role

The few available published surveys regarding the sociocultural role of wolverines among Arctic Indigenous communities unsurprisingly affirm that Arctic Indigenous Peoples demonstrate a refined and complex appreciation of wolverine biology (Nelson 1983; Cardinal 2004; Braund 2010). Indeed, conversations of IK holders in the Canadian High Arctic include much of the information contained in the present review, as well as aspects of wolverine biology not yet documented by Western science (Cardinal 2004). There have, however, been few targeted efforts to collate and publish this knowledge, and much of the information discussed here therefore spans both the Arctic and subarctic. As such, there are both scientific and ethical imperatives for the involvement of Arctic Indigenous Peoples in producing management objectives, biological knowledge, and conservation priorities for wolverines in this region.

Indigenous hunters and trappers in northern Canada predominantly characterize the wolverine as smart and tough, adept at raiding traplines and food caches, and enduring harsh conditions (Laugrand 2017; Bonamy et al. 2020). Among the Inuit, the animal’s consumption of human food and resources position it as an adversary and have made it the subject of hatred (Laugrand 2017), whereas for the Innu of Ungava the wolverine is central to creation stories and occupies an archetypal trickster role in mythology (Armitage 1992). In Arctic Russia, Indigenous Peoples including the Nenets report occasional interactions with wolverines as predators of semi-domestic reindeer (Rangifer tarandus) and raiders of supplies (Forbes 1999; Terekhina et al. 2021). Although they report rarely encountering wolverines, Gwich’in hunters and trappers in boreal and Arctic Canada frequently cross and follow tracks, revealing that wolverines travel impressive distances and disregard topography (Benson 2014). Wolverines appear frequently in legend and oral history among the Gwich’in, often characterized as tricksters or by their intelligence (Benson 2014).

Wolverines are hunted and trapped across the Arctic, predominantly for their fur or to mitigate predation of reindeer. Wolverine fur is valued for its resistance to accumulating frost and is therefore desirable for the ruffs of parka hoods (Hardy 1948). In contrast to more southerly regions, where wolverines are typically trapped, wolverines in the Arctic are commonly hunted and taken opportunistically (Fig. 3; Cardinal 2004; Braund 2010). The prevalence of hunting for wolverines increased with the arrival of the snowmachine; prior to this, take of wolverines in the Arctic was likely considerably lower. Total annual reported take based on carcass collections in Nunavut for the years 2009–2012 ranged from 61 to 124, with most animals taken on the mainland in the Kivalliq and Kitikmeot regions and far fewer in the more northerly Qikiqtaaluk region (Awan and Szor 2012). Reported harvest varies spatially across the Arctic and is broadly concentrated near communities (Braund 2010; Awan and Szor 2012), reflecting the importance of access in determining harvest patterns. In Alaska, reported wolverine harvest for Game Management Units 18, 22, 23, and 26, roughly corresponding to the Arctic portion of the state, is strongly male biased and annual totals ranged from 70 (1994) to 177 (2016) for the period 1987–2018 (Fig. 3; Alaska Department of Fish and Game 1986, 1988a, 1988b, 2013, 1990a; 1990b, 1991, 1993, 1998, 2001, 2007, personal communication). However, these values are widely assumed to underestimate true harvest, since most furs in Arctic communities are used or sold locally and therefore not reported (possibly as many as 90% of the total harvest; Magoun 1985; Cardinal 2004), despite reporting mandates in some regions (e.g., Alaska; Alaska Department of Fish and Game 2021).

Fig. 2
figure 2

Wolverine (Gulo gulo) distribution in the Arctic. For islands, distribution was assigned to whole islands; no effort was made to distinguish distribution patterns within islands (except Greenland). “Occasional” indicates at least one verifiable record but insufficient observations to support a reproductive population. Latitudinal treeline indicates the southern boundary of the region considered in this review, and the southern boundary of wolverine distribution is derived from wolverine range reported by Copeland et al. (2010). We compared our updated range with that reported by the International Union for Conservation of Nature (IUCN; Abramov 2016); diagonal hatching indicates either that (i) IUCN considers the region occupied, whereas our review found no or only occasional reports or (ii) IUCN considers the region unoccupied, whereas our review found reports of occupation. References for wolverine presence records are in Online Resource 1

Full size image
Fig. 3
figure 3

Annual wolverine (Gulo gulo) harvest in Arctic Alaska between 1987 and 2018 from fur sealing records. Black lines depict boundaries of Game Management Units 18, 22, 23, and 26, which align roughly with latitudinal treeline. Reported harvest underestimates true harvest, possibly by as much at 90% (Magoun 1985; Cardinal 2004; Lee 2016)

Full size image

Distribution

We estimated that 25% of wolverines’ global distribution by area occurs in the Arctic and of this 28% is on islands (i.e., 7% of global distribution is on Arctic islands; Fig. 2). In the mainland Arctic regions of North America and Eurasia, wolverines occur everywhere except northern Labrador and Québec (i.e., the Ungava Peninsula). In this region, the wolverine population was severely reduced or extirpated as a result of hunting/trapping and/or caribou (Rangifer tarandus) declines in the early/mid 1900s (Banfield and Tener 1958; Schmelzer 2006; Committee on the Status of Endangered Wildlife in Canada 2014). The species then reappeared in the region’s harvest records between 1963 and 1979 (Moisan 1996; Committee on the Status of Endangered Wildlife in Canada 2014), but since then only two occurrences of wolverines have been confirmed (both trapped in 2019; Guillaume Szor, personal communication), although other unconfirmed observations (i.e., lacking physical evidence or photographs) have been reported since 1979 in both Québec and Labrador (Moisan 1996; Committee on the Status of Endangered Wildlife in Canada 2014). In the Low Arctic (defined as Bioclimate Subzones D and E; Walker et al. 2005) of Alaska, occupancy is non-uniform and associated with terrain ruggedness and soil drainage (Poley et al. 2018).

On High Arctic islands of the Canadian Archipelago and Russia, the species’ distribution is sporadic. Most islands of the Canadian Arctic are considered occupied by wolverines, although the species’ presence on many islands is supported by only a few published reports, and there are few or no reports on the westernmost Queen Elizabeth Islands (van Zyll de Jong 1975; Committee on the Status of Endangered Wildlife in Canada 2014). Wolverines are reported on Ellesmere Island, but there are few historical reports of the species in Greenland, despite being separated from Ellesmere Island only by the 25–35 km seasonally frozen Kennedy Channel. Of Greenland reports, none are verified with photographs, although the abundance of track observations by skilled naturalists in the Thule region during the 1930s and 1940s provides strong support for at least occasional forays by wolverines to the island (Muus et al. 1981). Additionally, although wolverines have been documented on Baffin Island and the northern islands of Hudson Bay (Southampton, Coats, and Mansel), their presence has not been reported for any islands of the more northerly Foxe Basin (Manning 1943; van Zyll de Jong 1975; Mallory et al. 2001; Committee on the Status of Endangered Wildlife in Canada 2014). Occasional sightings are reported on several of the High Arctic Russian islands, but only Wrangel Island has a verified reproductive population apparently resulting from recent colonization (Vekhov 1999; Kolodeznikov 2013; Starova et al. 2014).

The mechanism underlying the sporadic distribution of wolverines across Arctic islands is currently unknown, although availability of prey and carrion is a likely driver (van Zyll de Jong 1975; Cardinal 2004; Committee on the Status of Endangered Wildlife in Canada 2014). There is generally high accordance between the distribution of wolverines’ primary prey species (caribou and muskoxen, Ovibos moschatus) and wolverines. Caribou and muskoxen are sparse or absent on many of the Canadian islands lacking wolverines (Ellef Ringnes, King Christen, Cornwall, Meighen, and Lougheed Islands; Jenkins et al. 2011). Reindeer are similarly unreported on most islands of the Russian High Arctic that lack wolverines, particularly Franz Joseph Land and Severnaya Zemlya (Mizin et al. 2018), whereas muskoxen are absent from all except Wrangel Island, where they were introduced in 1975 followed by the colonization of wolverines in subsequent decades (Starova et al. 2014; Cuyler et al. 2020). Conversely, caribou are present on all islands throughout the Arctic where wolverines are reported, although in regions with lower caribou densities muskoxen comprise a higher portion of the wolverine’s diet (Awan and Szor 2012).

In turn, these patterns of herbivore presence are likely linked to availability of suitable forage. The absence of muskoxen and caribou on High Arctic islands correlates well with Arctic Bioclimate Subzone A, the region of the Arctic characterized by the lowest net primary productivity, lowest mean summer temperatures, and lowest diversity of vascular plants (Walker et al. 2005). Most of the islands lacking wolverines are either predominantly glaciated or belong to subzones A or B. Collectively, these patterns suggest a northern limit to the bioclimatic conditions suitable for wolverines (Copeland et al. 2010), defined by net primary productivity and resulting carrion/prey availability. This hypothesis warrants formal evaluation using species distribution models (e.g., MaxEnt; Elith et al. 2011) and evaluation against possible competing explanations (e.g., availability of suitable denning habitat or cumulative abundance of prey biomass, including small mammals).

There are three interesting exceptions to the above patterns. Svalbard, northern Greenland, and the Ungava Peninsula all contain ungulates, sufficient in Greenland and Ungava to sustain populations of wolves at least periodically (Parker and Luttich 1986; Bergerud et al. 2008; Marquard-Petersen 2009). In Ungava, concurrent crashes of caribou, wolves, and wolverines in the early/mid 1900s, followed by the reappearance of wolverines in the trapping record between 1963 and 1979 associated with an increase in caribou, support the role of caribou in influencing wolverine distribution (Banfield and Tener 1958; Bergerud et al. 2008). Nonetheless, caribou rebounded to over a million individuals in Québec by the 1990s, accompanied by an increase in wolf populations (Courtois et al. 2003), but wolverines apparently did not capitalize on this surge of prey and carrion (Fortin et al. 2005). Ungava caribou populations subsequently experienced dramatic declines between 2000 and 2020 (Gagnon et al. 2019), which may be a contributing factor to wolverines’ continued failure to recolonize the region (Fortin et al. 2005). An additional possible contributing factor is the absence of Arctic ground squirrels (Urocitellus parryii), which have been identified as an important food source in some instances where/when ungulates are less prevalent (Magoun 1987). This persistent absence of wolverines on the Ungava Peninsula presents a ripe opportunity to study factors controlling wolverine recolonization dynamics and the species’ distribution. Similarly, the absence of wolverines in northern Greenland despite the presence of ungulates and carrion-providing wolves, as well as its relative proximity to Ellesmere Island, presents a puzzle that deserves further attention. The absence of wolverines in Svalbard despite the presence of Svalbard Reindeer is also intriguing, perhaps explained simply by its greater distance from potential mainland source populations.

Population structure

Genetic structure of wolverines in the Arctic has been studied in the Low Arctic region of North America. There, populations generally exhibit little nuclear genetic structure (Kyle and Strobeck 2001, 2002), but higher mitochondrial genetic structure (Wilson et al. 2000; Chappell et al. 2004; Tomasik and Cook 2005; Zigouris et al. 2013; Krejsa et al. 2021). This dynamic likely reflects the species’ male-biased dispersal (Vangen et al. 2001; Aronsson and Persson 2018), resulting in lower gene flow among maternally inherited markers. However, the spatial scale of such structuring remains uncertain; Wilson et al. (2000) detected mitochondrial genetic structuring among populations separated by < 100 km in Nunavut, whereas Dalerum et al. (2007) did not find genetic support for male-biased dispersal in the Western Brooks Range of Alaska. Mitochondrial markers also indicate that wolverines in Nunavut may be genetically distinct from those in Alaska and NWT (Tomasik and Cook 2005; Zigouris et al. 2013). Wolverines in Russia are broadly distinct from wolverines in Arctic North America, although specific location information about Russian samples (i.e., Arctic versus non-Arctic, Beringian versus non-Beringian) is generally not reported (Kyle and Strobeck 2002; Zigouris et al. 2013; Rico et al. 2015). We are aware of no study to evaluate genetic structure of wolverines within Russia.

Although 7% of global wolverine distribution occurs on Arctic islands, we are aware of no study to evaluate metapopulation dynamics (i.e., connectivity or genetic structure) among island-dwelling populations or the potential role of sea ice or its loss as a barrier to gene flow. Wolverines are occasionally observed traveling across sea ice (up to seven kilometers from land in some instances; Glass, unpublished data, Starova et al. 2014), and use of sea ice for dispersal is implied from the species’ presence on Arctic islands (e.g., St. Lawrence and Wrangel Islands, 70 km and 140 km from potential mainland source populations, respectively), but dispersal rates to, from, and between islands are unknown.

Population density

Broadly, IK holders report higher wolverine densities in the Low Arctic, decreasing across a northward gradient, with the lowest densities on islands of the High Arctic in Canada (Cardinal 2004). IK holders and genetic evidence suggest an increase in wolverine populations across much of Arctic Canada during recent decades (Tomasik and Cook 2005; Committee on the Status of Endangered Wildlife in Canada 2014), possibly rebounding from historical carnivore poisoning programs (Cardinal 2004). The only quantitative population density estimates come from the North American Low Arctic, and these estimates are inconsistent. In the Low Arctic regions of Napaktulik and Aberdeen Lakes, Nunavut, and Toolik Lake, Alaska, spatial capture–recapture analyses of bait station visits yielded population densities of approximately 2–4 individuals/1000 km2 (Awan and Boulanger 2016; Awan et al. 2020, Glass unpublished data), whereas estimates based on VHF telemetry (Utukok River, Alaska) and closed population models of bait station visits (Daring Lake, NWT) produced estimates of approximately 17–21 individuals/1000 km2 (Magoun 1985; Mulders et al. 2007). The habitat in these regions is superficially similar; all are within roughly 200 km of the latitudinal treeline, fall within the migration paths of substantial caribou herds, and have similar assemblages of potential prey species. Although some of the difference is likely attributable to bias associated with the analytical approaches (Mowat et al. 2020), this may not explain the entire discrepancy, since the territory sizes of radio-collared individuals studied contemporarily are consistent with the densities, i.e., much smaller in the Utukok region than the Toolik region (Magoun 1985, Glass unpublished data). The disparate estimates preclude a generalizable assessment of population density in Low Arctic tundra and highlight the need for additional studies, ideally accounting for spatial and temporal fluctuations in carrion availability, alternative food sources, and wolverine mortality, which can affect density and territorial dynamics of at least reproductive female wolverines (Aronsson and Persson 2018).

Diet and food habits

Most studies of wolverine diet in the Arctic have taken place near latitudinal treeline and suggest that wolverines rely heavily on ungulates for food, consistent with conspecifics in boreal and montane areas. Caribou and reindeer (hereafter “caribou”) account for the majority of ungulates’ contribution to wolverine diet in the Arctic, although muskoxen and moose are also well represented in some regions and individuals. Specifically, muskoxen may be particularly important in regions of the High Arctic where or when caribou are scarce (Awan and Szor 2012), while moose occur more frequently in the diet of wolverines near treeline and in regions with tall shrubs (Dorendorf et al. 2018). Wolverine consumption of ungulates is generally highest during winter, a time when alternate food sources are scarce and ungulates are vulnerable to predation and better preserved as carcasses (Magoun 1987). The species’ reliance on caribou extends even to regions without overwintering caribou, where wolverines consume bones, hide, and remains cached or otherwise left from seasons of caribou presence (Magoun 1987). Although wolverines have been observed to hunt and kill caribou on tundra (Magoun et al. 2018), most ungulates are likely consumed as carrion (Magoun 1987). This point is supported by interannual trends in wintertime consumption of caribou, which is influenced by caribou mortality rate and not by annual fluctuations in caribou abundance (Dalerum et al. 2009). Since overwinter food availability is important to wolverine reproduction (Persson 2005), these findings collectively suggest that the abundance of caribou, muskoxen, and moose carcasses in the Arctic is an important factor in regulating wolverine populations.

In addition to ungulates, wolverines supplement wintertime diet with a variety of species, including Arctic ground squirrels, Arvicoline rodents, ptarmigan (Lagopus spp.), beavers, snowshoe hares, seals (Phoca spp.), and cached geese and goose eggs (Makridin 1964; Magoun 1987; Lee and Niptanatiak 1993; Dorendorf et al. 2018; L’Herault 2018). Of these, Arctic ground squirrels and Arvicoline rodents have been found to occur in wolverine diet at frequencies comparable to ungulates in certain conditions (Magoun 1987; Dorendorf et al. 2018). Arctic ground squirrels consumed during winter are mostly killed and cached the preceding summer, rather than excavated during hibernation from burrows, although at least one instance of the latter has been observed (Magoun 1987). In coastal areas, wolverines consume marine mammals, particularly seals, in accordance with their proximity to the sea (Rausch and Pearson 1972; L’Herault 2018; Robards, unpublished data), and wolverines’ occasional forays onto sea ice are presumably associated with hunting or scavenging seals at least in some cases, as evidenced by spatiotemporal overlap with seal pupping (Glass, unpublished data, Starova et al. 2014).

During spring and summer, wolverines transition to a heavier reliance on hunting, particularly Arctic ground squirrels and Arvicoline rodents, and scavenge or hunt vulnerable caribou and muskoxen calves (Magoun 1987; Starova et al. 2014). As snow melts, wolverines on tussock tundra hunt Arvicolines that are apparently forced to the tussock tops as the inter-tussock space floods with snowmelt unable to permeate the still-frozen ground (Dorendorf et al. 2018). Wolverines have also been observed removing and caching eggs from waterfowl, shorebird, and raptor nests (Samelius et al. 2002; Hoover and Dickson 2007; Chris Latty and Devin Johnson personal communication), and hunting adult geese (Chen spp.; Samelius et al. 2002) and eiders (Somateria mollissima; Hoover and Dickson 2007). The presence of mud-encased waterfowl carcasses and eggshells in wintertime caches made by wolverines or foxes suggests that these food sources can help wolverines survive the relatively food-scarce winter (Magoun 1987, Glass unpublished data).

Parasitism, disease, and toxicology

Most studies of Arctic wolverine health have focused on parasitism. The species’ large territories, generalist diet, and position in regional food webs likely predispose wolverines to high overall parasitism rates, which generally exceed 80% in the Arctic and make the species a good candidate as a sentinel of landscape-level parasite spread (Watson et al. 2020; Sharma et al. 2021). Historical or longitudinal data of wolverine parasitism remain sparse for evaluating such trends, but information is improving. Documented parasites include several taxa of both helminths and protozoa, with > 80% taxa-specific prevalence of Trichinella spp., Taenia twitchelli, Sarcocystis spp., and Baylisascaris devosi and 40–60% prevalence of the protozoan Toxoplasma gondii in some, but not all, regions of the North American Arctic (Rausch 1959; Addison and Boles 1978; Reichard et al. 2008b, a; Dubey et al. 2010; Sharma et al. 2019; Watson et al. 2020). Wolverine is the preferred definitive host for T. twitchelli, which likely uses porcupine (Erethizon dorsatum) as its intermediate host and therefore may be limited to the Low Arctic (Rausch 1959). The recently discovered Trichinella chanchalensis has only been documented in wolverines of Arctic and boreal Canada, although previous methods may not have been able to distinguish it from T. nativa in other species and regions (Sharma et al. 2020). We are not aware of any ectoparasite reports among wolverines in the Arctic, although ear canker mite (Otodectes cynotis) was reported in a wolverine in northern boreal forest of Alaska (Wilson and Zarnke 1985).

Scavenging may increase wolverines’ susceptibility to viral and bacterial diseases, which can be transferred from other scavengers or arise from the scavenged animal’s necrobiome (Dalerum et al. 2005; Watson 2020). The few studies to catalogue potentially pathogenic viruses and bacteria in Arctic wolverines, however, generally indicate low incidence of both. Among 64 wolverines harvested in Alaska’s Brooks Range during 1998–2002, four had antibodies for canine distemper virus, one had antibodies for canine parvovirus type 2, and none had antibodies for canine adenovirus (Dalerum et al. 2005). A single apparently lethal rabies infection has been documented in Arctic Alaska, the only case of rabies in wolverine of which we are aware (Alaska Department of Fish and Game 2012). In Nunavut, pathogenic bacteria accounted for only 0.7% of wolverine gut microbiota, possibly attributable to low exposure mediated by caching behavior, defenses provided by the gut microbiome, or high stomach acidity (Inman et al. 2012; Watson 2020).

Wolverines’ relatively high trophic level and at least occasional consumption of marine mammals may also increase the species’ risk of accumulating toxic levels of anthropogenic contaminants. Total mercury concentrations in wolverines harvested near coastal Arctic Canadian communities, however, were well below expected toxicity level (Hoekstra et al. 2003a), although mercury toxicity has not been directly evaluated for this species. Polychlorinated biphenyl concentrations among the same sample of wolverines were higher than other terrestrial mammals, possibly due to consumption of marine mammals, but still below the expected level of reproductive impairment (Hoekstra et al. 2003b).

The implications of these parasites, viruses, bacteria, and toxins for wolverine and human health remain unclear. The single rabies case is the only Arctic wolverine mortality directly attributable to disease of which we are aware, and sublethal effects of this and other potential pathogens and toxins remain poorly understood. Most parasites found in wolverines use the species as a definitive host, which likely reduces health consequences (Watson et al. 2020), and although several bacterial taxa known to cause disease in other species are found in wolverine, bacterial disease induction in wolverines is so far undocumented (Watson 2020). Toxoplasmosis and Trichinellosis in humans, typically arising from undercooked parasitized meat consumption, are increasing concerns in Arctic communities relying on subsistence hunting (Keatts et al. 2021). Wolverines, which are rarely consumed, are unlikely vectors for these diseases, but handling and skinning animals has been identified as a risk factor in certain contexts and therefore care is warranted when handling live or dead wolverines (McDonald et al. 1990; Keatts et al. 2021; Sharma et al. 2021). Keatts et al. (2021) note the potential for coronavirus transmission between humans and wolverines.

Habitat associations and den-site selection

Arctic Canadian IK holders report that wolverine presence is predominantly determined by food availability, but that wolverines also associate broadly with rugged and rocky habitats, as well as higher elevations (Cardinal 2004). This trend aligns with Western science studies from the winter and spring in northern Alaska tundra, which have found that wolverine habitat selection is driven by terrain ruggedness, soil drainage, streams/rivers, and snow properties (Poley et al. 2018; Glass et al. 2021b). At the occupancy level of selection (i.e., the placement of home ranges), wolverines select more rugged terrain and areas with better drained soils, possibly showing a preference for Arctic ground squirrel habitat (Poley et al. 2018). In Arctic Alaska, this results in higher occupancy in the mountains and foothills of the Brooks Range than the more northerly and less rugged coastal plain (Poley et al. 2018). Within home ranges in Low Arctic Alaska during winter and spring, wolverines select strongly for streams and rivers, where they are observed hunting snowshoe hare and ptarmigan dwelling among the tall shrubs, as well as rugged terrain and deep, dense snow (Glass et al. 2021b).

Snow is used by wolverines across the species’ global range for reproductive den structures and food caching (Copeland et al. 2010; Inman et al. 2012) and its importance may be particularly high in the Arctic where alternative structural habitat is limited (Magoun et al. 2017). Elsewhere in the species’ global distribution, wolverines access subnivean structure including upturned rootwards, abandoned beaver lodges, and boulder complexes for reproductive dens (Magoun and Copeland 1998; Dawson et al. 2010; Scrafford and Boyce 2015; Jokinen et al. 2019). In Arctic Alaska and Russia, wolverines excavate reproductive dens in snow, and in many cases these dens are excavated solely in snowdrifts (Glass et al. In Press; Serebryakov 1983; Magoun and Copeland 1998), although IK holders also report the use of rocks and boulders for dens (Cardinal 2004). Reproductive dens are commonly found in snowdrifts that form in small drainages and along the cut banks associated with lake edges (Glass et al. In Press). Excavations of such dens reveal complex networks of tunnels and chambers, with total tunnel lengths approaching 55 m (Magoun 1985). Wolverine reproductive dens and resting sites have also been documented in subterranean ice caves formed by eroding permafrost (Glass et al. 2021a).

In addition to reproductive dens, wolverines excavate burrows in snow for resting sites and to cache food (Glass et al. In press; Lee and Niptanatiak 1993, 1996). Approximately half of wintertime wolverine resting sites are in burrows, and wolverines rest in these burrows when it is cold and there is little solar radiation, transitioning to the snow surface during periods of higher air temperature (Glass et al. 2021c).

Response to industrial development

The Arctic contains large mineral and petroleum deposits, the extraction of which requires constructing and maintaining transportation networks, buildings, waste processing facilities, and other supporting infrastructure. Russia produces the majority of Arctic petroleum and mineral resources, with petroleum production concentrated in the Yamal and Gydan Peninsulas and gold, iron, apatite, copper, nickel, and cobalt mines distributed across mainland Arctic Russia (Haley et al. 2011; Peters et al. 2011; Mathonniere 2019). In the North American Arctic, petroleum production is concentrated in the Prudhoe Bay region of Alaska’s North Slope, whereas mineral extraction includes several active diamond and gold mines in Arctic Canada, an iron mine on northern Baffin Island, and a zinc mine in Alaska’s Brooks Range (Haley et al. 2011). Trends in extractive infrastructure and activity can be difficult to document and predict, particularly in Arctic Russia where inconsistent reporting standards and methodologies have precluded longitudinal evaluation (Haley et al. 2011). Nonetheless, continued expansion of these industries is indicated by the initiation and support of resource development for economic opportunities by a suite of local communities, as well as Arctic states’ policies signaling continued expansion (Laruelle 2020) and expanding footprints and human presence in most regions with adequate data (Raynolds et al. 2014; Kröger 2019).

Current and projected industrialization highlight the importance of understanding the relationship between wolverines and human infrastructure in an Arctic tundra environment. However, the only systematic attempts to do so, by evaluating habitat selection and movement responses, have been unsuccessful due to low sample sizes or insufficient overlap between monitored wolverines and disturbance features (Johnson et al. 2005; Glass et al. 2021b). Incidental observations suggest possible mechanisms underlying the species’ response, such as apparent short-term attraction to mines in Arctic Canada (Golder Associates Ltd. 2017; De Beers Canada Inc. 2018), lethal control following attraction to grease traps (Gebauer et al. 2014), and possible home range delineation along a highway corridor by six wolverines in Arctic Alaska (Glass, unpublished data). These observations highlight the need for further study of population-level impacts of industrial development to wolverines.

Research and Conservation Priorities

The Arctic is changing rapidly. For wolverines, these changes introduce novel prey species and pathogens, may reduce connectivity among island populations, alter habitat availability, and increase interactions with humans through new infrastructure, transportation corridors, and hunting/trapping opportunities. Some of these changes, such as increased moose abundance, may benefit wolverines, while others will likely incur negative consequences. Although recent studies have dramatically improved our understanding of wolverines in the Arctic, the individual findings of each represent brief periods and typically small spatial scales. As change accelerates, so too does the need to gather longitudinal information about population processes and drivers thereof (Fisher et al. 2022). These research and monitoring efforts should be developed by, or in collaboration with, local Indigenous communities to improve their efficacy, applicability, and equity (Brook et al. 2009). Here, we identify several research priorities focusing on understanding wolverines’ susceptibility and response to ongoing change in the Arctic (summarized in Table 2).

Table 2 Knowledge gaps and areas of potential conservation concern for wolverines (Gulo gulo) in the Arctic
Full size table

Effective population monitoring

Conservation-driven management of wolverines in the Arctic requires an accurate understanding of spatiotemporal trends and drivers of occupancy, population abundance, and demographics. Since obtaining such information can be logistically challenging, we advocate partnering with existing wildlife monitoring efforts (e.g., aerial surveys for caribou and muskoxen) and/or local hunters and trappers to design and implement these studies. For example, existing aerial surveys can be adapted to include wolverine track observations if observers are skilled at identifying wolverine tracks, snow conditions are amenable, and survey protocol such as height above ground and aircraft speed are compatible, potentially yielding occupancy and population density estimates (Golden et al. 2007; Poley et al. 2018). Alternatively, baited camera sites or hair snares operated by local hunters and trappers can simultaneously strengthen collaborative relationships and gather longitudinal data regarding reproduction, population density, and occupancy, as has been successful elsewhere (Webb et al. 2016). In concert with these efforts, we recommend developing community-based reporting initiatives to more accurately document harvest, including demographic information, such as sex and age, as are currently implemented in Arctic Canada (Lee 1994, 2016; Awan and Szor 2012; Kukka and Jung 2016). Such information can supplement population abundance estimates, enabling study of harvest influences on population dynamics (Mowat et al. 2020).

Improved understanding of population dynamics will also facilitate a mechanistic study of the limits of wolverine distribution in the Arctic, including the potential role of bioclimatic conditions and ungulate abundance. Parsing the drivers underlying distribution will aid in predicting and understanding the species’ response to climate warming, including the potential to expand into currently unoccupied regions, as bioclimatic conditions change.

Understanding and forecasting impacts of climate change

We expect that impacts of climate change to wolverines in the Arctic will primarily be mediated by reduced snow and sea ice availability, changing prey species assemblage, and shifts in disease dynamics. We advocate prioritizing study of these changes’ mechanistic links to wolverine fitness and population dynamics.

Net primary productivity is broadly increasing in the Arctic, and this trend extends to the regions currently unoccupied by wolverines (Yu et al. 2017), suggesting future conditions more favorable for wolverine prey species and the possibility of wolverines’ northward range expansion. Arctic shrubification is driving a northward range expansion of moose, beaver, and snowshoe hare, all species identified by modern studies to be consumed by wolverines in Arctic Alaska despite their arrival to the region only decades ago (Tape et al. 2016a, b, 2018). The availability of these species to wolverines will likely continue to increase under climate change (Zhou et al. 2020). Conversely, wintertime ground-icing events, which are increasingly common and severe with climate change, are dampening the peaks of Arvicoline population cycles and driving mass mortality events in caribou and muskoxen (Aars and Ims 2002; Kohler and Aanes 2004; Ims et al. 2008; Jenkins et al. 2011). Earlier springs and later autumns are altering the migration phenology of species including caribou and geese (Lameris et al. 2018; Mallory et al. 2020), and sea ice degradation is reducing wolverines’ access to marine resources (Post et al. 2013). Although wolverines have a broad diet, the unprecedented magnitude and synchrony of these environmental changes have the potential to impact wolverine fitness, particularly where ungulates such as caribou are negatively impacted (Mallory and Boyce 2018). Cumulatively, these changes highlight the need for continued study of wolverines’ diet throughout the Arctic, with a focus on mechanistic links between diet composition and wolverine fitness.

Disease vectors are strongly influenced by environmental conditions, and climate change may facilitate range expansion of temperate vectors into the Arctic and modify transmission pathways of existing Arctic vectors (Bradley et al. 2005; Keatts et al. 2021). For instance, warmer and longer summers can increase developmental rates of some parasites and may ultimately result in higher prevalence and individual parasite load (Bradley et al. 2005). Information regarding pathogen and parasite disease potential to wolverines in the Arctic is insufficient to predict the possible impacts of such changes to the species. We therefore recommend continued monitoring of wolverine health using community-based initiatives in collaboration with hunters and trappers (Oakley et al. 2016; Sharma et al. 2019; Watson et al. 2020) and study of physiological response to potential disease vectors. The connection between human and animal health in the Arctic and subarctic is increasingly recognized under the “One Health” framework (Ruscio et al. 2015), yielding successful longitudinal monitoring efforts of wildlife health across diverse taxa (e.g., caribou, muskoxen, and ringed seal (Phoca hispida); Brook et al. 2009; Tomaselli et al. 2018; Harwood et al. 2020); we advocate development of similar long-term efforts for wolverines.

Wolverines’ occupancy of Arctic islands may make the species susceptible to reduced inter-island dispersal opportunities resulting from sea ice decline (Post et al. 2013). The abrupt loss of sea ice that began in the mid twentieth century is likely reducing gene flow among island populations of this vagile species, a dynamic supported by genetics studies in arctic foxes, wolves, and caribou (Geffen et al. 2007; Carmichael et al. 2008; Jenkins et al. 2016). This rapid change likely increases the susceptibility of wolverines and their prey to stochastic environmental events and genetic drift. Islands that have been historically occupied by wolverines may therefore become unsustainable as sea ice reductions disconnect them from other populations of wolverines and their prey species. We suggest community-based campaigns using tissues from harvested animals or bait stations to monitor gene flow among island populations and possible impacts of changing sea ice conditions, building upon existing efforts that already include southern portions of the Canadian Archipelago (Awan and Szor 2012; Lee 2016).

Finally, spring snowmelt is advancing rapidly in the Arctic, with the duration of the snow-covered period contracting by as much as nearly 9 days per decade in some regions (Callaghan et al. 2011). This advancing melt is likely exposing vulnerable wolverine kits to predators and inclement weather at a younger age, but the capacity of wolverines to respond by shifting reproductive phenology or structural requirements for den sites is unknown. Moreover, advancing snowmelt and warmer air temperatures may reduce the efficacy of springtime food caches upon which wolverines rely to fuel lactation (Inman et al. 2012). The cumulative effects of these snow-mediated changes have received considerable attention in contiguous USA (e.g., McKelvey et al. 2011) and warrant further research in the Arctic.

Understanding and forecasting impacts of industrial development

We suggest prioritizing research regarding development impacts into three possible disturbance pathways. First, wintertime over-snow travel, which is common in the Arctic to reduce tundra compaction, could negatively impact wolverines occupying subnivean reproductive dens and resting burrows, either through compaction or auditory/visual disturbance. Exploratory surveys for mineral or petroleum deposits, which typically deploy high-energy seismic pulses into tundra, can have footprints covering 10% of the landscape (Raynolds et al. 2020). Mitigating the impacts of over-snow travel and/or associated seismic pulses has received considerable attention for polar bears (Ursus maritimus) and ringed seals, which have similar reproductive phenology and denning requirements (Kelly et al. 2010; Wilson and Durner 2020; Owen et al. 2021). Observations of wolverine den abandonment following interaction with humans indicate that this species may be susceptible to these disturbances (Glass et al. In press). However, population-level response of wolverines to den-site disturbance has so far not been evaluated, and we are not aware of any existing measures to mitigate impacts of over-snow travel.

Second, increased ice-breaking ship traffic may reduce connectivity among island populations of wolverines in the Arctic, exacerbating impacts of climate-induced sea ice decline. Such shipping traffic inhibits caribou migration (Dumond et al. 2013), and springtime ice breakers may be of particular concern for wolverines given the temporal overlap with juvenile dispersal (Vangen et al. 2001). We recommend study of wolverine dispersal dynamics across sea ice to better understand and mitigate any impacts from this form of disturbance.

Finally, industrial development may increase mortality by facilitating harvest in new regions or prompting lethal control at development sites. Roads associated with mines and oilfields provide hunting and trapping access to regions where harvest is otherwise more difficult and development-related capital flowing into Arctic communities increases harvest efficiency through modernization of equipment, particularly snowmachines (Gebauer et al. 2014; Fauchald et al. 2017). We suggest community-based monitoring initiatives to address these potential concerns and ensure sustainability of landscape-scale harvest. Mortality arising from lethal control at mine sites or oilfields can be proactively mitigated using appropriate deterrence procedures consistent with those used for other species, such as bear and fox (e.g., secure waste storage and processing containers).

Ensuring harvest sustainability

The absence of information regarding population trends, demographics, or vital rates for most Arctic regions, coupled with unreported harvest due to use of fur within and among communities, complicates efforts to evaluate sustainability of current harvest levels. Most regions report higher harvest of males than females (Awan and Szor 2012; Alaska Department of Fish and Game 2013; Lee 2016) consistent with expectations for a species exhibiting male-biased dispersal. This provides some support for the sustainability of current harvest levels, since persistence of species with low reproductive output requires low mortality of adult females (Kukka et al. 2017). The precise relationship between demographics of hunted/trapped wolverines and harvest sustainability remains unknown, however (Kukka et al. 2017), and additional information is required to evaluate impacts of harvest to wolverine populations in the Arctic. We recommend coupled monitoring of wolverine population parameters and harvest rates, as discussed in “Effective population monitoring,” to address this knowledge gap, with the ultimate goal of harvest management policies that are based in a firm understanding of wolverine population dynamics.

As technological and industrial development continue to facilitate harvest farther from population centers, it may become increasingly important to proactively establish refugia. Harvest is currently concentrated near population centers, and in some cases harvest rates are likely unsustainable at a local level but enabled by immigration from surrounding de facto refugia as is seen in subarctic regions (Kukka et al. 2022). Construction of road networks and development of more efficient and faster over-snow vehicles may enable access to these regions and necessitate codifying refugia (Lee 2016).

Conclusion

The Arctic is often considered a stronghold for wolverines, given the species’ prevalence and the region’s low human population density and cold climate (Fisher et al. 2022). Nonetheless, much remains unknown about wolverines in this region, including the potential impacts of modern anthropogenic change to the species. Climate change in particular offers numerous potential pathways to negatively impact the species, as does the expansion of extraction-based industries and, in many regions, underestimated harvest. Ensuring long-term persistence of wolverines in the Arctic requires considerable additional research and monitoring to resolve mechanisms driving abundance and demographics, and mitigation of growing threats as the Arctic continues to change.