A review of aerial survey density estimates of bearded seals (Erignathus barbatus) in the Canadian Arctic highlights important knowledge gaps and research needs

There has been significant sea ice loss in the Arctic as a result of climate change, with measurable impacts on available habitat for ice-obligate marine mammals. Bearded seals (Erignathus barbatus) have a circumpolar distribution and primarily inhabit coastal areas of shallower depths with seasonal pack ice cover. However, there is limited information and no formal estimates on population abundance and densities of bearded seals within Canada. Here, we review historic field reports and publications from aerial surveys conducted from 1974 to 2022 to compile the density and abundance estimates of bearded seals across Canadian waters. Aerial surveys for marine mammals have been flown in several areas across the Canadian Arctic, including the Beaufort Sea (1974–2006), Canadian High Arctic (1979–2022), Baffin Bay-Davis Strait (1979–2014), and the Hudson Complex (1994–2017). After reviewing all published data sources, we report that the density of bearded seals in Canada were highest (and highly variable) in the Beaufort Sea (0.01–8.68 seals*km−2), then in Baffin Bay-Davis Strait (0.004–8.3 seals*km−2), and lowest and less variable inter-annually in the Hudson Complex (0.02 to 0.12 seals*km−2) at the time of these surveys. We also determined that bearded seals are more often found in areas of patchy ice cover (50–75%) with shallow water depths ≤ 500 m. Further, this review identifies regions within the Canadian Arctic and sub-Arctic that require updated aerial survey information. Quantifying the abundance and density estimates of bearded seals in Canada is essential for monitoring population status over time to better understand how this species is responding to environmental variation from anthropogenic activity and climate change.


Introduction
Accelerated climate warming in the Arctic over the past several decades has been four times that of the global mean rate of warming (IPCC 2014;Rantanen et al. 2022) Significant reductions in summer ice extent, multiyear ice, and ice thickness in the Arctic (Comiso 2012) have led to longer periods of ice-free habitat and warming sea surface temperatures (Stroeve et al. 2014;Serreze et al. 2019).Arctic marine mammals, especially pinnipeds, are vulnerable to these climate shifts due to their dependence on sea ice for raising young, molting, resting, and feeding (Laidre et al. 2008;Kovacs et al. 2012).To detect geographic variation in changes to Arctic marine mammal subpopulations related to environmental change, data on abundance and density estimates must be collected annually in order to quantify range shifts, changes to population growth, and habitat preference (Laidre et al. 2015;Ferguson et al. 2018).Further, the incorporation of historical records from early survey and monitoring records can be valuable for filling in the gaps for abundance and density estimates of large megafauna (McClenachan et al. 2012).Metrics of density and abundance for marine mammals are key because marine spatial planning requires explicit information on the overlap between species' distributions and stock assessments for management and mitigation measures (Grech et al. 2011).
Bearded seals (Erignathus barbatus) are large pinnipeds (~ 400 kg; Cameron et al. 2010) that have a circumpolar distribution, extending from the Arctic Ocean (85° N) south to Sakhalin Island (45° N) in the Pacific Ocean and to southern Hudson Bay and James Bay, along the Labrador coast, and south to northern Newfoundland (~ 50° N) in the Atlantic (Burns 1981;Smith 1981;Cameron et al. 2010;Kovacs et al. 2018;NAMMCO-North Atlantic Marine Mammal Commission 2023).Within Canada, bearded seals are year-round residents in Arctic waters, including the Hudson Bay region, Eastern Beaufort Sea and Amundsen Gulf, Arctic Archipelago (Barrow Strait, Jones Sound, North Water Polynya), Labrador Sea, and Baffin Bay (Cleator 1996;Cameron et al. 2010).The bearded seal is generally split into two subspecies, the Pacific E. b. nauticus (Pallas 1811) and the Atlantic E. b. barbatus (Erxleben 1777), which are spatially divided in central Canada, but there is still a need for geographic resolution on the distribution of these two subspecies.Generally, bearded seal distribution is described as patchy with relatively low densities (Smith 1981).Although found in both pelagic and coastal environments, bearded seals are primarily benthic feeders and forage on marine invertebrates and fishes in benthic habitats (Burns 1981;Smith 1981;Kienle and Berta 2016).While the bulk of bearded seal diet consists of mollusks, crustaceans, and fish, juveniles eat primarily fish, whereas adults tend to be benthic foragers, primarily consuming mollusks (Finley and Evans 1983;Cameron et al. 2010;Young et al. 2010).Due to their benthic foraging strategies and dependence on ice for pup rearing, bearded seals are associated with shallower areas within the continental shelf (200 m or less), and, depending on the region, occupy both pack and landfast ice in winter and spring and drifting and patchy sea ice during summer (McLaren 1958;Burns 1967Burns , 1981;;Smith 1981;Simpkins et al. 2003;Cameron et al. 2010Cameron et al. , 2018)).Despite the widespread geographic distribution of bearded seals across the Canadian Arctic, regional estimates of their abundance and densities remain limited (NAMMCO-North Atlantic Marine Mammal Commission 2023).
Monitoring animal populations in the Arctic is challenging since species occur across large geographic areas and can have patchy distributions (Smith 1981), often located far from human settlements with little to no access (Wang et al. 2019).A straightforward method to monitor bearded seal population status over time is through quantitative information on the overall population abundance and inter-annual (or periodic) variation in abundance estimates across regions.Periodic standardized aerial surveys are often employed to estimate the abundance of Arctic marine mammals and allow for assessments of population abundance and density over time (Young et al. 2015;Ferguson et al. 2018;Conn et al. 2021).Historically, aerial surveys for Arctic marine mammals in Canadian waters, specifically pinnipeds (Myers and Bowen 1989;Young et al. 2015), have occurred either during spring (~ April-June) or late-summer (July-September).Seals spend more time on the sea ice during spring while molting and sea ice breakup has begun, which allows for estimates of animals across the sea ice.In late-summer, when the Arctic is mostly ice-free, some pinniped species (e.g., harbor seal, Phoca vitulina and Atlantic walrus, Odobenus rosmarus rosmarus) are concentrated at haul out locations, while many other species, including bearded seals, are widely dispersed in marine waters, thus making them difficult to survey during this time.While aerial surveys have provided insights into the distribution and abundance of Arctic marine mammals, information on bearded seal populations across Canada is still extremely data deficient as indicated by the last Committee on the Status of Endangered Wildlife in Canada (COSEWIC) assessment in 2007 (COSEWIC 2007;Cameron et al. 2010).
Despite the widespread distribution of bearded seals, much of the existing information on the species' abundances and trends across circumpolar regions is available only in federal reports (Cameron et al. 2010), unpublished consulting reports (e.g., Koski 1980a, b;Koski andDavis 1979, 1980), and a select few published reviews (Cleator 1996;Scherdin et al. 2022).Further, abundance estimates and an understanding of habitat associations for bearded seals within Canada are even more limited (Cameron et al. 2010).In a recent review by Scherdin et al. (2022), knowledge of the Atlantic populations of bearded seals was synthesized, but there is minimal information reported on specific regions within the Canadian range; instead the review focuses on Greenland and a select few areas near the North Water Polynya in the Canadian High Arctic.Additionally, of the existing publications that report observations of bearded seals in Canada, much of the information is anywhere from 5 to 50-years old and there is a lack of recent estimates for bearded seals (Cameron et al. 2010).Using harvest reports, it was crudely estimated by McLaren (McLaren 1958), and then later synthesized by Cleator (1996), that approximately 186,000-190,000 bearded seals inhabit the regions of Hudson Bay, Baffin Island, Foxe Basin, and Hudson Strait.These rough estimates were based on a combination of harvest data and aerial surveys with much of the data being extrapolations of aerial survey observations.An overview of the existing and more recent knowledge of bearded seal abundance estimates across Canada and a synthesis of previous survey efforts, can provide a more detailed understanding of their population structure in Canadian waters.A synthesis of bearded seal abundance estimates and survey efforts across Canada will inform priorities and direction for future surveys in support of bearded seal assessment and monitoring to aid pinniped conservation and management practices.
This review focuses on available research publications (including journal articles and government and consulting reports) that provide information on bearded seal abundance and density estimates within the Canadian Arctic.Specifically, we reviewed all existing aerial survey efforts that were conducted for marine mammals and detected bearded seals (directly through study objectives or indirectly through incidental observations) from published studies that occurred from 1974 to 2022.By compiling existing information on bearded seal density and abundance in Canada, we can provide a reference point for future conservation assessments and highlight geographic areas that have not been surveyed or where surveys are dated and should be flown again with specific aims for estimating bearded seals.

Delineating Arctic marine areas for bearded seals
Within the Arctic, several key Large Ocean Marine Areas (LOMAs) have been identified by the Circumpolar Biodiversity Monitoring Program (CBMP) and the Arctic Council, through international government collaborators including Canada, United States, Russia, Greenland, Denmark, Iceland, and Norway (Gill et al. 2011).As outlined by the 2008 CBMP-Marine Plan, LOMAs were identified as having diverse taxonomic data available (ranging from low to high trophic level species), being a biological hotspot, being important for traditional foods, occurring on physical oceanographic boundaries (i.e., ice edges, distinct currents), and in areas that may change profoundly from anthropogenic impact (Gill et al. 2011).We subdivided and slightly modified the identified LOMAs in the Canadian Arctic, that are within the bearded seal distribution for this review; (1) Eastern Beaufort Sea and Amundsen Gulf, (2) High Arctic Archipelago, (3) Hudson Complex, and (4) Davis Strait-Baffin Bay (Fig. 1).

Eastern Beaufort Sea and Amundsen Gulf
The Eastern Beaufort Sea area encompasses the Canadian extent of the Beaufort Sea, including the Mackenzie Delta and Tuktoyaktuk Peninsula, while Amundsen Gulf includes Prince Albert Sound, waters to the north and east around Banks Island, and other waters east of the Eastern Beaufort Sea (Fig. 1).This region also includes the marine waters around Banks Island, Sverdrup Islands, and Parry Islands as far east as Viscount Melville Sound (Gill et al. 2011).The Eastern Beaufort Sea LOMA is characterized most notably by the Beaufort Continental Shelf, along with a short icefree season and high primary productivity with high rates of sediment and freshwater input in spring and summer (Cobb et al. 2008).In this region, a polynya coincides with freshwater input and the Beaufort Gyre (Gill et al. 2011).The Eastern Beaufort Sea generally has relatively shallow waters (mean depth ~200 m) with considerable amounts of pack ice in the northern region and seasonal ice in the south (Cobb et al. 2008;Gill et al. 2011).In the Eastern Beaufort Sea and Amundsen Gulf, landfast ice forms in late-September and eventually converges with drifting ice along the coast (Cobb et al. 2008).

High Arctic Archipelago
The Canadian High Arctic Archipelago is a marine complex that extends to Victoria Island and Melville Sound in the west to Jones Sound and Barrow Strait in the east (Gill et al. 2011;Fig. 1).Other regions in this LOMA include the marine waters around Somerset Island, northwestern Baffin Island near the western edge of Lancaster Sound, Norwegian Bay, Cornwallis Island, Prince Phillip Basin, Byam and Austin channels, and Prince of Wales Island (Fig. 1).The High Arctic Archipelago is characterized by perennial sea ice at high latitudes and seasonal sea ice and relatively high primary productivity, particularly at lower latitudes of the archipelago (Gill et al. 2011).

Davis Strait-Baffin Bay
The marine region of Davis Strait-Baffin Bay extends from the northern edge of the Labrador Sea all the way to the northern edge of Baffin Bay (Fig. 1).The northern end of Baffin Bay includes the marine waters surrounding Ellesmere Island, Devon Island, and Baffin Island (Fig. 1).Additionally at the north of Baffin Bay near Smith Sound there is the North Water Polynya (NOW), which contributes to high productivity in that region (Tremblay et al. 2002;Gill et al. 2011).Within the Canadian region of the Davis Strait-Baffin Bay LOMA there is also high primary productivity in Lancaster Sound, Prince Regent Inlet, and Admiralty Inlet and along the Baffin Current on the eastern side of Baffin Island (Gill et al. 2011).The Davis-Baffin LOMA is mostly covered by sea ice in winter (some exceptions in eastern Davis Strait) and then is primarily ice-free in summer with prolonged ice cover in western areas (Tremblay et al. 2002;Gill et al. 2011).

Hudson Complex
The Hudson Complex is a relatively enclosed system that extends westward and north to Fury and Hecla Strait, eastward to Hudson Strait and south to James Bay (Fig. 1).Generally, the Hudson Complex has depths around 200 m or less, but this depth increases in Foxe Channel and Hudson Strait (Gill et al. 2011).The Hudson Complex is characterized by high primary productivity as a result of strong tidal mixing and two polynyas in the northwestern Hudson Bay and Foxe Basin (Ferland et al. 2011).The Hudson Complex is covered seasonally by ice with little to no multi-year ice (Ferland et al. 2011;Gill et al. 2011).In general, this LOMA includes subregions of the Hudson Strait, Hudson Bay, James Bay, Ungava Bay, Foxe Basin, and the marine area around Southampton Island (Fig. 1).

Regional abundance estimates
Surveying for bearded seals is challenging due to their solitary behavior, patchy distribution, and low density across their range, which has led to extrapolated population estimates that are uncertain and outdated for much of the Canadian distribution of bearded seals.Further, bearded seals have not often been a focal species for Arctic research, and, therefore, minimal funding and few surveys have been dedicated to this species, which limits assessments of historical and current population abundance and density estimates across their Canadian range.However, regional population abundances and densities of bearded seals have been estimated across several distinct areas within Canadian waters for certain time periods.

Eastern Beaufort Sea and Amundsen Gulf
Within the Eastern Beaufort Sea, systematic strip-transect aerial surveys on seals hauled out on ice during spring were conducted from 1974 to 1979 (Stirling et al. 1975(Stirling et al. , 1977(Stirling et al. , 1982; Online Resource 1) and then again from 1981 to 1984 with surveys that also included Amundsen Gulf (Kingsley and Lunn 1983;Kingsley 1985; Table 1; Online Resource 1).These surveys were primarily designed to estimate ringed seal (Pusa hispida) abundances but also noted bearded seal sightings.Strip-transect surveys in the Eastern Beaufort Sea area revealed maximum densities ranging from 0.01 to 8.68 seals*km −2 of ice (Stirling et al. 1975(Stirling et al. , 1977(Stirling et al. , 1982;;Kingsley and Lunn 1983;Kingsley 1985; Fig. 1).Ice conditions at the time of the surveys in the Eastern Beaufort Sea and Amundsen Gulf varied substantially across years in the 1970s and 1980s, which could have led to the large range in densities estimates across aerial surveys (Stirling et al. 1975(Stirling et al. , 1977(Stirling et al. , 1982;;Kingsley and Lunn 1983;Kingsley 1985).Based on the aerial surveys conducted by Stirling et al. (1975Stirling et al. ( , 1977Stirling et al. ( , 1982)), bearded seal abundance estimates ranged from 1197 (1975) to 3019 (1978; Table 1).Indices calculated by Kingsley andLunn (1982, 1983)

High Arctic Archipelago
In the Canadian High Arctic Archipelago aerial surveys were flown from 1978 to 1982 primarily in May-July to quantify the distribution of Atlantic walruses, ringed seals, and bearded seals (Fig. 1; Online Resource 1).Unfortunately information on survey effort or individuals observed is insufficient to estimate densities of bearded seals in the High Arctic Archipelago (Table 2).No abundance estimates were calculated for these surveys due to the low number of bearded seal observations.Kingsley et al. (1982a, b)

Davis Strait-Baffin Bay
Strip-transect surveys of bearded seals in the Davis Strait-Baffin Bay area reported maximum densities ranging from  3; Fig. 1).From aerial surveys conducted from May-July of 1979 along the eastern edge of Baffin Bay, Koski (1980a) estimated around 8,462 bearded seals based on 369 individual observations (Table 3; Online Resource 1).Koski (1980b) also estimated 115 bearded seals from July to September along the eastern edge of Baffin Bay (Table 3).Following this, Koski and Davis (1980) reported 112 individuals during marine mammal surveys in 1980, and calculated uncorrected estimates of 7400-9500 bearded seals (Table 3).Finley and Renaud (1980) 3).It is important to note that all the Davis-Baffin surveys were flown in spring (April-June), except for the 1980 surveys by Koski and Davis (1980) which were flown in July-October.

Hudson Complex
The

Sea ice cover
Information relating bearded seal abundances to habitat features are generally lacking across Canada, but some published studies have reported associations with specific environmental and habitat characteristics.Bearded seals often inhabit areas with pack ice, which they use for resting, pupping, and moulting (Fay 1974) and they have the capability of maintaining breathing holes in sea ice using their strong claws on their front flippers (Stirling and Smith 1977;Smith 1981).Cameron et al. (2010) concluded that bearded seals tend to occur more often in areas with access to patchy drift ice (range: 70-90% ice cover), either from glaciers or stratified annual sea ice (Koski 1980b;Simpkins et al. 2003;Scherdin et al. 2022).Across all four LOMAs bearded seals were mainly found in areas of patchy, medium to high (50-75%) sea ice concentration.In the Eastern Beaufort Sea, although habitat preference and occupancy models were not examined, bearded seals were observed to be associated with ice concentrations of 53-97% and in areas where large and small ice floes converge (Stirling et al. 1975(Stirling et al. , 1977(Stirling et al. , 1982;;Kingsley andLunn 1983, 1985).For the Eastern Beaufort Sea and Amundsen Gulf, sea ice cover depends on the specific area since pack ice and landfast ice can vary across the Amundsen Gulf, thus affecting bearded seal occupancy (Smith 1981).Bearded seals occurred in higher densities within the Davis-Baffin area during spring (April-May) and in areas with relatively high sea ice coverage (range: 25-75%; Koski 1980a, b;Koski and Davis 1980;Heide-Jørgensen et al. 2013, 2016).Similarly, in the Canadian High Arctic Archipelago, bearded seals were typically observed in areas with a sea ice coverage of 50-75% (range: 25-100%; Kingsley et al. 1982aKingsley et al. , b, 1985;;Kingsley and Lunn 1982).Many existing studies in the High Arctic Archipelago either did not analyze or describe habitat associations or did not have enough bearded seals observations to make statistical comparisons (Finley and Renaud 1980;Young et al. 2019;Carlyle et al. 2021).In the Hudson Complex, bearded seals were associated with relatively higher amounts of sea ice cover with observations being highest at ~ 75-85% sea ice cover (range: 40-99%; Lunn et al. 1997;Chambellant et al. 2012;Elliot et al. 2013), but some of these surveys (e.g., Elliot et al. 2013) were conducted earlier in the year, which likely explains the higher ice coverage reported.It is important to note that bearded seals may not always be able to select environmental features (i.e., sea ice cover, depth, benthic habitat) independently from one another, especially given that such features vary within and between years.

Water depth
Bearded seals in Canadian waters prefer coastal areas with relatively shallow water (< 150-200 m water depth;Cameron et al. 2010).Bearded seals in the Eastern Beaufort Sea area were often associated with water depths of less than 75 m, with most observations ranging between 25 and 75 m (Stirling et al. 1975(Stirling et al. , 1977(Stirling et al. , 1982;;Smith 1981;Kingsley andLunn 1983, 1985).In contrast, the Davis-Baffin area had much deeper water depths where bearded seals were detected, which is likely an artifact of the regions deeper waters (> 200 m), particularly in the NOW.Further, in the eastern High Arctic (e.g., NOW and Davis-Baffin area), bearded seals move into the offshore pack ice in deep water and return to shallower coastal waters as the ice recedes (Koski 1980a, b).Bearded seals in the Davis-Baffin area were observed in waters 400 m or less (Koski 1980a;Heide-Jørgensen et al. 2013, 2016), although Koski and Davis (1980a) suggested that < 200 m is preferred bearded seal habitat in this area during spring, and probably in winter.
In the High Arctic Archipelago, bearded seals were most often seen around depths of less than 75 m (Kingsley et al. 1982a(Kingsley et al. , b, 1985)), but in some cases they occurred in depths of 200-500 m (Finley and Renaud 1980)

Limitations and future recommendations
Due to the limited frequency, geographic coverage, and design (studies not focused on bearded seals) of aerial surveys conducted in Canadian waters over time, there are several limitations and gaps in our knowledge of the species in Canadian waters.A major limitation of estimating abundances and distributions of bearded seals over time in Canada is that Arctic seal aerial surveys are seldom performed and there are numerous logistical difficulties of covering such expansive area across the Canadian Arctic.Further, bearded seals are solitary by nature and occur at low densities across their circumpolar distribution, making it a difficult species to survey and estimate abundance.Bearded seals in the Canadian Arctic have a patchy and widespread distribution (Burns 1981) and are not typically the focal species for aerial surveys in spring, which contributes to their data deficiency and skewed estimates of absolute abundance across their Canadian range.In addition, incomplete survey coverage could also overestimate density and abundance if surveys cover an area with a high concentration of seals, previous surveys for bearded seals have not covered the entire known geographic range of the species, which is undetermined and likely shifting.
For the areas that have been surveyed, they have not been surveyed recently or frequently enough to support an examination of bearded seal trends over time.Several aerial surveys in recent years (2017-2021) found few to no observations of bearded seals in western Hudson Bay (n = 13), the marine waters around Alert, Nunavut (n = 0; Ferguson and Yurkowski-unpublished data), Archer Fjord (n = 4; Carlyle et al. 2021), Norwegian Bay (n = 5;Florko et al. 2023), and Eclipse Sound (n = 0; Ferguson & Yurkowski-unpublished data), but the high-Arctic surveys were designed to explore for multiple species of marine mammals across a largely understudied area.Most often, abundance estimates, or even anecdotal observations of bearded seals are collected during aerial surveys where other marine megafauna are the focal species such as bowhead whales (Balaena mysticetus), Atlantic walruses, narwhals (Monodon monoceros), belugas (Delphinapterus leucas), and primarily ringed seals, which are more common and in Canada are 20 times more abundant than bearded seals (Koski 1980a;Koski and Davis 1980;Stirling et al. 1975Stirling et al. , 1977Stirling et al. , 1982;;Kingsley andLunn 1983, 1985).It is important to note that in Alaska, U.S.A, there are about twice as many ringed seals as there are bearded seals and thus bearded seals in Alaska are considered relatively more important and are targeted by research surveys and subsistence harvesters more often than bearded seals within the Canadian range (Bengston et al. 2005).Alaskan waters are ideal bearded seal habitat with high benthic productivity and extensive areas with water depths < 200 m, which likely explains the higher densities of bearded seals in Alaska relative to the Canadian High Arctic (Simpkins et al. 2003).Due to this nature of focal species of interest and design of aerial surveys in Canadian waters, some prime bearded seal habitat areas are not being adequately covered during these surveys, or in some cases, are not surveyed at all.
As revealed by this review, most information on bearded seal abundance estimates originated from surveys conducted over several core regions in the Canadian Arctic (Cameron et al. 2010;Gill et al. 2011) and during the same season.However, there are gaps within the Canadian range of bearded seals that have either been barely surveyed since the 1970s and 1980s or have not been surveyed at all, making it impossible to detect changes in abundance and distribution, including documenting potential changes associated with a changing Arctic (e.g., range expansions, contractions, and shifts).For example, few surveys have collected data on bearded seals in the Arctic Basin (Gill et al. 2011), southern edge of the Labrador Sea and along the northern Newfoundland coast, or within the Northwestern Passages.Additionally, while studies in recent years (Carlyle et al. 2021) have detected few bearded seals in the marine area around Ellesmere Island, there must be more effort in place to survey that area of the High Arctic as it is expected to be the last refugia for Arctic marine mammals with global warming (Moore and Huntington 2008).Without proper coverage of the Canadian range of bearded seals, or at least focused efforts within parts of the range that are accessible, it will be a constant challenge to quantify population trends within Canada.
The high interannual variation in seasonal ice cover during surveys, which can influence the probability of bearded seal observations, is another source of data uncertainty.Some surveys across the Canadian distribution were conducted in just one year when there was high sea ice cover (Koski 1980a, b;Kingsley and Lunn 1982;Elliot et al. 2013;Heide-Jørgensen et al. 2013;Carlyle et al. 2021) and, thus, it is possible that bearded seals were underestimated in those years due to limited preferred habitat (i.e., broken, floe ice cover) available to them in the surveyed area.Given that bearded seals appear to respond substantially to interannual variation in sea ice cover (Cameron et al. 2010), potentially due to their foraging behaviors, prey availability, and resource productivity (Burns 1981;Simpkins et al. 2003), it is important for surveys in a given region to span multiple years.Without knowledge of interannual variation in bearded seal abundance or distribution, it is not possible to establish a reference point abundance estimate for certain parts of their range or examine abundance trends over time in response to changes in available habitat.
Further, little has been documented across different regions about when peak haul-out occurs during the moult for bearded seals, which could affect availability of seals to be seen from an aircraft (London et al. 2022).The period when bearded seals haul-out on the ice to moult and thus become available for counting from aircraft, is more protracted and less well-defined than the haul-out period of the ringed seal (Smith 1981).Due to this difference, the date of the seasonal peak of bearded seals hauling out is difficult to establish.Additionally, for bearded seals, the moulting period is longer (119 ± 2 days) than that of ringed seals (28 ± 6 days) indicating that the time-frame for detecting moulting individuals is more variable for bearded seals (Thometz et al. 2021).Smith (1981) suggested that a potential optimum time to count bearded seals would be during the pupping season (late-April and May).Male bearded seals are usually not seen during this time, but counts of females with newborn pups can be used to provide production indices for regional surveys (Smith 1981;Cameron et al. 2010).This time-frame, however, for possible peak bearded seals is at least one full month earlier than aerial surveys for ringed seals, which likely has led to underestimates of bearded seal abundance.If annual surveys are not feasible due to financial or logistical constraints, surveys should attempt to prioritize years or areas where there may be a higher abundance of preferred bearded seal habitat (70-90% sea ice cover, bathymetric depth < 200 m).
In addition to differences in ice condition and the timing of haul-out behaviors, several other factors associated with aerial survey efforts for seals can lead to sources of data uncertainty and biased results.Variation in weather conditions during the time of surveys can influence the likelihood that seals are resting on the ice versus swimming in the water (Kingsley et al. 1985) Additionally, poor weather may influence the detection rates of seals that are resting on the ice if cloud cover inhibits survey efforts.Survey altitude is also known to vary across the aerial survey efforts conducted throughout the Canadian Arctic, which can have significant effects on detection rates of pinnipeds.Lastly, the experience of observers may also lead to biases in results if there are individual differences in the ability to distinguish between marine mammal species from aerial surveys (Bröker et al. 2019).One approach to address factors that can affect survey detection and variation across different survey protocols is to employ estimations of availability and detection probability within calculations of absolute density and abundance (London et al. 2022).If future surveys for pinnipeds, specifically bearded seals, account for processes such as incomplete detection (i.e., perception bias or double-sampling) or availability, then estimates of seal abundances can be directly comparable to each despite difference in survey protocol (e.g., instrument-based or human observed) or conditions (e.g., sea ice cover, weather, altitude).Alternatively, other modes of estimating seal abundances can be used to expedite surveys and obtain high-resolution data.Photographic surveys are becoming increasingly common as a way of avoiding biases from different aerial observers (Bröker et al. 2019;Young et al. 2019).Future bearded seal surveys could implement machine learning to detect individuals in drone imagery (Dujon et al. 2021) or FLIR (forward-looking infrared) devices can be mounted on aircraft and used to count seals hauled out by detecting the heat emitted from their body (Kingsley et al. 1990).Population abundance of pinnipeds can also be estimated using genetic methods collected from hunters in the Canadian Arctic (Conn et al. 2020).
While not included in this review, hunting and harvest data from Indigenous communities can also inform abundance and density estimates for Arctic pinnipeds (McLaren 1958).As reported by Cameron et al. (2010), between 1976and 2003 in the Northwest Territories, Nunavut, Nunavik, and Labrador and Newfoundland, communities harvested between 2,375 to 3,800 bearded seals annually.Indigenous and Local Knowledge can identify and rank habitat by incorporating harvest counts and other local information on seals into population models and habitat suitability models.Further, harvest data and Indigenous Knowledge can highlight the optimal location and timing for aerial surveys which would enable more comprehensive and accurate population abundance and density estimates of bearded seals.Historically, bearded seals were considered a valuable species for Indigenous subsistence practices, but in the last several decades, the demand for bearded seal meat and pelts has declined in communities across the Canadian Arctic (COSE-WIC 2007) as well as in Greenland (Cameron et al. 2010).Due to this trend, and the preference for ringed seals for traditional and economic subsistence, harvest data on bearded seals is limited.In addition, most seal harvests come from Nunavut and the Northwest Territories (Smith and Taylor 1977;Smith 1981).Bearded seals are not commonly harvested by people in northern Manitoba or Ontario and thus there are low harvest numbers from these areas, which are less informative for estimating population distribution and abundance (Smith and Taylor 1977;Smith 1981).Increased effort to collect bearded seal harvest data could provide an alternative way to improve estimates of regional bearded seal abundance.

Conclusion
Given rapid changes in sea ice and ocean conditions in the Canadian Arctic due to climate change, as well as increases in other anthropogenic impacts, it is even more important to revisit baselines and continue to collect new information to monitor population status, distribution, density estimates, and habitat use of bearded seals through surveys, but also other approaches to understand their biology, ecology, and demographics.In order to effectively manage bearded seals and understand stock structure across Canada, key data on regional population abundances are necessary.The 2016 IUCN assessment recommended that bearded seals continue to be monitored so that more data can become available (Kovacs 2016), but there has been little urgency to study this species since their widespread distribution, solitary nature, and flexible habitat requirements make it difficult to determine their conservation status.While bearded seal ecology and population abundances have been summarized in reports (Cleator 1996;Cameron et al. 2010;Scherdin et al. 2022), there is still no single recent or comprehensive abundance estimate for bearded seals across their Canadian distribution.This lack of information stems from scattered monitoring efforts and large regional gaps in bearded seal abundance and distribution.By highlighting all of the existing knowledge of bearded seal abundance estimates from aerial surveys, we hope this review provides a focused perspective on existing bearded seal population estimates in Canada, including the identification of which geographic areas have not yet been surveyed historically or recently.Additionally, this study aims to identify what information is needed to determine regional estimates and a Canada range-wide population estimate for bearded seals, a goal that cannot be achieved with existing data.Without continued monitoring and a better understanding of bearded seal abundance and distribution, it is nearly impossible to recognize to what degree bearded seals may be vulnerable to accelerated environmental change or human activities.

Fig. 1
Fig. 1 Approximate year-round distribution (a subset of their circumpolar range) of bearded seals (Erignathus barbatus; dark gray) in North American waters including Canada and areas adjacent to Alaska, Iceland, and Greenland.Black outlines represent the four classified Large Ocean Marine Areas (LOMAs) of a Eastern Beaufort Harwood et al. (2007)astern Beaufort Sea and Amundsen Gulf (inclusive of Prince Albert Sound and the waters around Banks Island) indicated that from 1981 to 1984 bearded seal abundance ranged from450 (1984)to 1722 (1983; Table 1).Harwood et al. (2007)also used aerial surveys from 2003 to 2006 to evaluate potential impacts of offshore development on seals in a small (< 1000 km 2 ) area north of the Mackenzie River estuary in the Beaufort Sea during spring.Harwood et al. (2007)observed a total of seven observations of bearded seals during these surveys, but did not generate bearded seal density or abundance estimates given that they were not the focus of this assessment nor were the surveys in prime bearded seal habitat (Table1).Overall, bearded seal abundances varied substantially across years in the Eastern Beaufort area and Amundsen Gulf, likely in part due to variation in weather conditions, ice cover across different time periods during the time of surveys, which ranged from 53 to 90%, and slight differences in survey coverage across years.

Table 1
Eastern Beaufort Sea bearded seal (Erignathus barbatus) survey efforts and abundance estimates from aerial surveys conducted from 1974 to 2006 in western Canada Values with "*" are means calculated from different survey strata for a single year when single-year values were not reported.Values with "^" were inferred from publications based on trends observed in tables and figures, but not explicitly stated in text.Values with "c" account for detection and availability bias during aerial surveys.Values with "i" indicate estimates that account for ice area

Table 2
High Arctic Archipelago bearded seal (Erignathus barbatus) survey efforts and abundance estimates from aerial surveys conducted from 1978 to 2019 in the Canadian High Arctic Values with "^" were inferred from publications based on trends observed in tables and figures, but not explicitly stated in text (2012; Complex has the most comprehensive amount of information on bearded seal abundance estimates in recent decades.All but one year(2012; Elliot et al. 2013)of aerial surveys conducted in the Hudson complex were flown in May-June, when ice was partially broken up.
(Elliott et al. 2013hile no population abundances were estimated, surveys flown in Hudson Strait in March-April 2012 recorded 20 bearded seals(Elliott et al. 2013; Online Resource 1).Similar to the Eastern Beaufort LOMA, there was considerable annual variation in bearded seal abundance estimates in the Hudson Complex, potentially due to differences in resource availability and ice conditions between years.

Table 3
Baffin Bay-Davis Strait bearded seal (Erignathus barbatus) survey efforts and abundance estimates from aerial surveys conducted from 1979 to 2010 in eastern CanadaValues with "*" are means calculated from different survey strata for a single year when single-year values were not reported.Values with "^" were inferred from publications based on trends observed in tables and figures, but not explicitly stated in text.Values with "c" account for detection and availability bias during aerial surveys

Table 4
Hudson Complex bearded seal (Erignathus barbatus) survey efforts and abundance estimates from aerial surveys conducted from 1978 to 2019 in the Hudson Bay and Hudson Strait area of Canada Values with "*" are means calculated from different survey strata for a single year when single-year values were not reported.Values with "^" were inferred from publications based on trends observed in tables and figures, but not explicitly stated in text . Surveys conducted by Elliot et al. (2013) recorded twenty bearded seals over waters less than 200 m deep.Otherwise, no published reports from the Hudson Complex explicitly report water depths related to bearded seal abundance.