Physical and sociocultural factors affecting walrus ...

This article was downloaded by: [University of Alaska Fairbanks] On: 02 April 2013, At: 08:10 Publisher: Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Polar Geography Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/tpog20

Physical and sociocultural factors affecting walrus subsistence at three villages in the northern Bering Sea: 1952–2004 a

b

Martin D. Robards , Alexander S. Kitaysky & John J. Burns

c

a

Wildlife Conservation Society, P.O. Box 751110, Fairbanks, AK, 99775, USA b

Institute of Arctic Biology, University of Alaska Fairbanks, P.O. Box 757000, Fairbanks, AK, 99775, USA c

Alaska Department of Fish and Game - Retired, AK, USA Version of record first published: 02 Apr 2013.

To cite this article: Martin D. Robards , Alexander S. Kitaysky & John J. Burns (2013): Physical and sociocultural factors affecting walrus subsistence at three villages in the northern Bering Sea: 1952–2004, Polar Geography, 36:1-2, 65-85 To link to this article: http://dx.doi.org/10.1080/1088937X.2013.765519

PLEASE SCROLL DOWN FOR ARTICLE Full terms and conditions of use: http://www.tandfonline.com/page/terms-andconditions This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. The publisher does not give any warranty express or implied or make any representation that the contents will be complete or accurate or up to date. The accuracy of any instructions, formulae, and drug doses should be independently verified with primary sources. The publisher shall not be liable for any loss, actions, claims, proceedings, demand, or costs or damages whatsoever or howsoever caused arising directly or indirectly in connection with or arising out of the use of this material.

Polar Geography, 2013 Vol. 36, Nos. 12, pp. 6585, http://dx.doi.org/10.1080/1088937X.2013.765519

Physical and sociocultural factors affecting walrus subsistence at three villages in the northern Bering Sea: 1952 2004

MARTIN D. ROBARDSa*, ALEXANDER S. KITAYSKYb and JOHN J. BURNSc a

Downloaded by [University of Alaska Fairbanks] at 08:10 02 April 2013

b

Wildlife Conservation Society, P.O. Box 751110, Fairbanks, AK 99775, USA; Institute of Arctic Biology, University of Alaska Fairbanks, P.O. Box 757000, Fairbanks, AK 99775, USA; c Alaska Department of Fish and Game - Retired, AK, USA Changes in sea-ice conditions have direct bearing on ice-associated species such as Pacific walrus (Odobenus rosmarus divergens), an important species for coastal Alaska Native subsistence. We explore the dynamic relationships among sea ice, walrus, and subsistence hunting between 1952 and 2004 at three northern Bering Sea villages Diomede, Gambell, and Savoonga. We integrate changes in timing, size, and gender distribution of walrus catches under four environmental regimes that alter the extent, duration, and persistence of sea ice. Our results suggest that the physical ice conditions proximal to the three villages affect timing and migration of walrus herds and thus hunting, but village-specific factors, such as the number and demographics of hunters, impart strong inter-community variability in the magnitude of catches. Decadal-scale climatic regimes are correlated with consistent patterns of timing and magnitude for the walrus hunts at Gambell and Savoonga, and at Diomede until 1989. However, a marked reduction in walrus catches at Diomede since 1989 is attributable to several social changes that compound more difficult hunting conditions. Our study highlights the important linkages between geographic location and the sociocultural capacity to hunt (e.g. number of hunters and local rules) when considering the resilience or vulnerability of village subsistence activities in a changing climate.

Introduction While the rapid rate of change in the earth’s cryosphere is increasingly evident (Intergovernmental Panel on Climate Change 2007; Serreze et al. 2007; Stroeve et al. 2007), significant difficulties persist in applying such broad-scale observations to marine ecosystems at the local scale of a wildlife population or a village dependent on that wildlife (Duerden 2004; Gearheard et al. 2006; Krupnik and Ray 2007; Laidler 2006; Ray et al. 2010). Relationships within social-ecological systems will change over space and time as a consequence of climate change (Liu et al. 2007), so specific environmental conditions may only favor a certain humanenvironment relationship at specific locations and times. Consequently, to assess impacts of environmental change on humanresource relationships requires attention to both spatial and temporal scales of analysis, as well as the interactions across scales and disciplines. Geographers have long grappled with the interdisciplinary challenges associated with the scale-specific relationships between people and their environment. *Corresponding author. Email: [email protected] # 2013 Taylor & Francis


66

M.D. Robards et al.

For example, environment and development geographers have demonstrated how livelihoods are spatially and temporally dynamic, and socially heterogeneous (Perrault 2009). Furthermore, the interaction of physical and social processes operating at various temporal and physical scales leads to scale-specific adaptive capacities of people in relation to climate change (Ruddell and Wentz 2009; Vincent 2007). Consequently, the sustainability of specific livelihoods is a function of both the adaptive capacities of specific groups of people in a dynamic environment and the spatial and temporal dynamics of desired resources. In this paper, we explore the marine mammal subsistence livelihoods and adaptive capacities of three northern Bering Sea communities, as evidenced by their hunting success over four climatic regimes. The northern Bering Sea (figure 1) is a highly productive marine ecosystem between Chukotka (Russia) and Alaska (United States) (Grebmeier et al. 2006; Springer et al. 1996). The region is ice-covered in winter, but in spring and summer,

Figure 1. The northern Bering Sea region showing Gambell and Savoonga on St. Lawrence Island, and Diomede. Note: Diomede is approximately 300 km north of St. Lawrence Island; the intervening waters termed the Chirikov Basin. Image: TerraModis 4 June 2002: NASA.


Physical and sociocultural factors affecting walrus subsistence

67

the dual processes of sea-ice melt (thinning) and retreat (reduction in extent) results in the northward passage of ice through the Bering Strait into the Chukchi Sea and Arctic Basin. In conjunction with this annual retreat, populations of marine mammals such as bowhead whales (Balaena mysticetus) and Pacific walruses (Odobenus rosmarus divergens) migrate northwards through the Bering Sea, and have supported the nutritional and cultural needs of Native villages in Alaska and Chukotka for over 2000 years (Krupnik and Ray 2007). The subarctic climate of the northern Bering Sea is characterized by both large interannual variability and decadal-scale changes in two dominant climate patterns: the Pacific decadal oscillation and Arctic oscillation (Stabeno and Hunt 2002). Recent decadal-scale climatic changes have resulted in a general decline in the extent and quality of sea ice in the marginal ice zone of the Bering Sea, more open water, earlier breakup, and later arrival of the fall pack ice (Grebmeier et al. 2006; Huntington 2000; Stabeno et al. 2007; Stroeve et al. 2007). Benson and Trites (2002) suggest that climatic forcing indirectly affects marine mammal species through changes in the distribution and abundance of prey. However, Pacific walruses use ice as a platform for resting, giving birth, and nursing (Burns 1970; Fay 1982). Consequently, ecological associations for walrus are likely to change in concert with alterations in the extent and quality of their sea-ice habitat, and not just their prey. In turn, social-ecological associations between walrus and the people who rely on them for subsistence will be mediated by physical constraints on hunting, the ecology of walrus, and the social capacity to hunt. Coastal villages in the northern Bering Sea are frequently located at ecological and cultural boundaries, where the inherent diversity and connections are thought to have benefited their resilience to environmental and social changes over time (Krupnik 1993; Meek et al. 2008; Turner et al. 2003). Nevertheless, Alaskan hunters are currently reporting broad-scale changes in sea-ice and weather patterns that reduce their ability to safely and efficiently find, access, retrieve, and return walrus to communities (Metcalf and Robards 2008). For example, Patrick Omiak Sr. of Diomede states that ‘the ice is always gone so fast that we are not catching walrus like we used to’; Leonard Apangalook of Gambell states ‘our walrus season is very short now’ (Eskimo Walrus Commission 2003); and Conrad Oozeva of Gambell (in Oozeva et al. 2004) states that ‘because the ice melted too fast, the walruses moved faster and in shorter time than usual.’ In this paper, we address how much shorter, how much faster, and what has been the capacity of communities to adapt to these consequences of climatic regime changes. Adaptation is regarded as critical for enhancing resilience, or reducing the vulnerability of arctic communities to a changing climate (Ford 2009). We focus on adaptation at Diomede (Inaliq), Gambell (Sivuqaq), and Savoonga (Sivungaq), which have been the primary walrus-hunting villages in Alaska over at least the past 60 years. We address three critical factors that affect the magnitude and timing of walrus subsistence hunting over a 53-year period: (1) walrus ecology; (2) sea-ice conditions; and (3) the social and technological setting (Fay 1982; Hughes 1960; Stoker 1983). We recognize the importance of weather, but data of sufficient resolution and consistency to cover the entire period were not available. However, see Kapsch et al. (2010) for a more detailed analysis of a portion of this data (since 1980) in relation to meteorological and local observations of weather. The national boundary with Russia may also constrain the hunting activities of some villages, although practical evidence for this is mixed.


68

M.D. Robards et al.

We hypothesized that climatic regimes will correlate with timing and success of walrus hunting, and that regime shifts will be reflected in parallel changes among villages. Alternatively, if patterns diverge among villages, we hypothesized that differences would reflect village-specific sea-ice, weather, or a village’s inherent capacity to adapt to new conditions. Initial efforts to describe the relationship between climate, sea mammals, and subsistence hunting in the Bering Sea were made by Krupnik and Bogoslovskaya (1999). They found that climatic factors that altered sea-ice and weather dynamics led to profound decadal-scale variations in the relative success of village-based marine mammal hunting. During the relatively warm period of the 1930s and 1940s, communities on the northern coast of Chukotka generally caught greater numbers of marine mammals, while those in southern regions obtained relatively few; in subsequent cooler conditions, this situation was reversed (Krupnik and Bogoslovskaya 1999). Such alternating regimes resulted from the modified connections between climate, sea ice, weather, wildlife, and people. More recently, local-scale impacts to walrus and human communities from a changing arctic environment have been described through sharing of knowledge types, using local observations and broadscale scientific perspectives (Gearheard et al. 2006; Kapsch et al. 2010; Krupnik and Ray 2007; Laidler 2006; Metcalf and Robards 2008). This manuscript continues to build from these efforts.

Interdisciplinary background to the ecological, physical, and social components of subsistence walrus hunting Walrus ecology and sea-ice dynamics The use of sea ice by walrus is balanced between finding ice of suitable thickness to support their weight while out of the water ( 0.6 m) and opportunities for passage while swimming, or ice thin enough for them to break through from beneath to breathe and rest (B0.2 m; Fay 1982). In the late winter and spring, the period of maximum annual ice extent, walrus congregate in areas of unconsolidated ice, usually within 100 km of the southern margin of the pack ice (Burns 1970). Consequently, although some walrus remain relatively far north, in open polynyas and leads throughout the winter, most overwinter south of St. Lawrence Island, only moving north during the spring recession of sea ice (Fay 1982). Sea-ice characteristics in winter result in two major walrus concentrations, one extending from southwest of St. Lawrence Island into the Gulf of Anadyr, and one in northern Bristol Bay (Fay 1982; Krupnik and Ray 2007; Speckman et al. 2011). The northerly spring migration from wintering areas coincides with favorable ice movements during the melt and retreat of ice through the northern Bering Sea (figure 2). Walrus travel north with the ice, but are not restricted to passively riding on it; their principal progress may be accomplished by swimming as far north as ice conditions permit (Burns 1965; Fay 1982). During the 1930s, walrus passed St. Lawrence Island in May and June (Collins 1940; Oozeva et al. 2004). Heinrich (1947) documented gender segregation at Diomede, with females and calves preceding males. Gambell hunters have also reported two phases of the spring migration passing their village, which may correspond to the two different winter aggregations (Eskimo Walrus Commission 2003; Krupnik and Ray 2007). Subsequently, females and calves have generally continued north, remaining with the sea



69

Figure 2. Typical spring ice disintegration and retraction through observation of the maximum extent of the 30% sea-ice concentration along the meridian 168.58 W in 1982. Note: Mean catch dates for the Alaska Native villages of Diomede, Gambell, and Savoonga are shown.

ice throughout summer, whereas most males and some females and calves remain at terrestrial haulouts on the Russian and Alaskan coasts. Despite the predicted reductions or complete loss of summer sea-ice in the Arctic basin, sea ice is expected to continue forming over much of the Bering Sea during winter (ACIA 2005). Consequently each spring, sea ice will continue to recede north through the Bering and Chukchi seas into the Arctic Basin. Presuming winter sea-ice north of the Bering Strait is not so broken as to allow year-round access for large numbers of walrus into the Chukchi Sea, sea ice will continue to first preclude (in winter), and then, as it retreats, allow walrus and other marine mammals to use the productive areas north of the Bering Strait. The extent and quality of winter sea-ice are affected by both weather and climate. The recent climatic regime shifts near 1977, 1989, and 1998, which altered sea-ice extent and concentration, are now well described by oceanographers and marine ecologists (Benson and Trites 2002; Hunt et al. 2002; Overland et al. 2008; Stabeno et al. 2007; Table 1). Prior to 1977, a ‘cold regime’ for the period after 1947 supported increasingly heavy ice conditions in the southeast Bering Sea from at least the 1950s to 1970s (Niebauer 1998; Stabeno et al. 2001, 2007; Walsh and Johnson 1979). After 1977, a ‘warm regime’ dominated until 1989. Ice was reduced in extent and concentration, with shorter residence time in the Bering Sea. Spring breakup was earlier, with maximum ice extent occurring in March instead of April or May (Stabeno and Overland 2001). Between 1989 and 1999, a ‘cool period’ (Stabeno and Hunt 2002) existed, although not to the extent of the early 1970s. Sea ice persisted longer in the southeastern Bering Sea during this period than in the 1980s, and was characterized by an extremely rapid melt-back in April leading to about a one-week earlier retreat through the northern Bering Sea (Stabeno and Hunt 2002). Since 1999, the timing of spring ice retreat has been highly variable (Stabeno and Overland 2001; Stabeno et al. 2007). Sea ice now extends farther south, and has continued to disintegrate and retreat earlier and more rapidly than between 1989 and 1999,

70

M.D. Robards et al.

Table 1. General characteristics of sea ice during four environmental regimes between 1947 and 2004. Regime Regime 1: 19471976a Regime 2: 19771988 Regime 3: 19891997


Regime 4: 19982004b

Conditions ‘‘Cold’’ with heavy ice conditions ‘‘Warm’’ with sea ice reduced in extent, of lower concentration, and shorter residence time in the Bering Sea ‘‘Cool’’ with rapid retreat of sea ice leading to a one-week earlier passage through the northern Bering Sea ‘‘Variable’’ with generally an early and rapid retreat of sea ice

Note: Regime shifts were near 1977, 1989, and 1998. a We describe Regime 1 from 1947 based on Niebauer (1998) who uses 1947 as a start date for his analysis. b Regime 4 did not end in 2004; rather, data reported by Stabeno et al. (2007) end at 2004.

resulting in the northern and eastern Bering Sea becoming ice-free earlier than in the previous decades (Stabeno and Hunt 2002; Stabeno et al. 2007). Nevertheless, retreat of ice over the northern shelf has remained highly variable, persisting longer in 2001 than was common in 19891999 (Stabeno et al. 2007).

Alaska spring subsistence walrus hunting Populations of the two villages on St. Lawrence Island (Gambell and Savoonga) steadily increased over the study period (figure 3B). They now use better firearms (first introduced in the 1800s) and more numerous and faster outboard enginepowered aluminum boats (introduced in the 1970s), which has increased both their hunting range and capacity to catch walrus (Robards and Joly 2008). In contrast, Diomede has not increased in population (figure 3B), has fewer young hunters, and uses fewer boats to accomplish their subsistence needs. Most walrus hunting in the northern Bering Sea occurs in late spring, when sea ice is diminishing, weather conditions are favorable, and walrus herds are migrating north past villages (Fay 1982). The speed of ice melt and retreat affects the ability of hunters to reach, hunt, and retrieve walrus. In heavy ice, walrus may still migrate past villages, but hunters do not have easy or safe access to them. Conversely, thin ice may also be problematic when it is ‘flimsy’ and thus unsuitable for hunting as reported by Leonard Apassingok Sr. who notes that the walrus are going ‘past [Gambell] very quick, even when ice was there’ at the 2008 spring hunter meetings. These comanagement events allow for in-person conversations between federal agency staff and hunters before the hunting season. Weather conditions, and particularly the wind, are ‘decisive’ factors in the ability to hunt on a particular day pushing ice onshore, thus preventing access to walrus; or offshore, thus taking walrus too far away (e.g. Fay 1982; Hughes 1960; Kapsch et al. 2010; Benter, pers. comm.). For example, in 1968, south winds in spring pushed the sea ice far out from Savoonga and Northeast Cape on St. Lawrence Island, with the result that Savoonga’s catch in spring was only 57 walrus, in contrast to 455 for the same period in 1966 (Burns 1969). However, in 1995 wind maintained favorable ice conditions close to Savoonga for longer than normal, extending their access to walrus. Wind may also exacerbate difficult sea-ice conditions, such as in



71

Figure 3. Historical pattern of walrus and human populations. (A) Estimates of the total ( ) and Russian component (') of the Pacific walrus population; data from Fay et al. (1997) and Speckman et al. (2011). Note that differing survey methods preclude reliable assessment of population trends. (B) Human population trends for the villages of Diomede ('), Gambell ( ), and Savoonga (j); data from Alaska Department of Commerce, Community, and Economic Development. Note: Vertical dashed lines represent boundaries between the four environmental regimes described between 1952 and 2004.

the late 1940s when wind consolidated heavy ice, leading to food shortages and ‘distress’ (Hughes 1960). In the 1870s, after commercial whalers decimated the Pacific walrus herds, hunting success even in the core habitat areas was drastically reduced, leading to starvation of villagers (Bockstoce 1986). However, although the absolute size of the walrus population is thought to have varied markedly during the past 50 years (figure 3A), we expect that the proximity of the vast majority of migrating walrus to Gambell, Savoonga, and Diomede will reduce the effects of walrus population size on timing of hunts or magnitude of catches as compared to villages closer to the periphery of the current range.

Methods To address how the success of walrus hunts varied among environmental regimes, we examined three sub-questions: (1) has the timing of the spring hunt changed among villages and between regimes? (2) have the magnitudes of spring walrus

72

M.D. Robards et al.

catches changed among villages and between regimes? and (3) what are the roles of social factors in our results? We delineate our analyses of catch data with the 1977, 1989, and 1998 regime shifts in the Bering Sea described above and in Table 1.


Walrus catch data Total catch numbers for each village were reported in Appendix 1 of Garlich-Miller et al. (2006). We use total numbers for adults from that dataset (19602002) and extend it with information, as available between 1952 and 1959, and 2003 and 2004 from agency reports for those years. U.S. Fish and Wildlife Service and the Eskimo Walrus Commission (EWC) recorded the daily walrus catch records for the spring hunt at Alaskan villages between 1992 and 2005. The USFWS Walrus Harvest Monitoring Program records date, gender, and age-class for every walrus observed during the spring hunt. Monitors meet most returning boats; therefore, relatively few walrus are unreported (Garlich-Miller and Burn 1999). For data prior to 1992, we used data archived by prior agency researchers involved in similar monitoring programs: the late Bud Fay (for 19521975) archived with Brendan Kelly at University of Alaska Southeast, and the private collections of John Burns (19581978), Kae Lourie (19801984), Scott Schliebe (19801989), and the EWC. Sease (1986) compiled information for animals of known age from Fay’s materials, although we use a more comprehensive dataset that includes all walrus known to be older than calves (but not necessarily of known age). Basic methods of reporting have been consistent over time, but we omit calves from our analysis as they are irregularly recorded in hunt monitor reports. Although walrus may be caught at other times of the year, we limit our data to a 3-month period between April 1st and June 30th that encompasses the spring hunt. We recognize reporting inconsistencies and the limitations of hunt monitoring programs, but assume that the proportion of unreported animals is relatively consistent across the spring season. We calculated mean date of the spring walrus hunt for each year and at each village based on the methodology of Sease (1986), where mean date is the sum of the number of walrus caught on each day multiplied by Julian date, divided by number of walrus caught. Differences in timing of male and female walrus at villages were examined using the Student’s t-test. For multiple comparisons of timing, we used analysis of variance (ANOVA). A post hoc Tukey’s test for pair-wise differences was used for significant ANOVA results to isolate environmental regimes and villages that differed in hunt timing. We statistically evaluated our three core questions with analysis of covariance (ANCOVA). The dependent variable (catch) was not normally distributed, so data were log10 transformed to accomplish normality. The ANCOVAs include village and regime as factors and timing as the covariate. Results Size of catches was retrieved from historical data for 49 of the 53 years between 1952 and 2004 for the villages of Diomede, Gambell, and Savoonga (no data were available for 1959, 1978, 1990, or 1991). We recovered daily-resolution catch data for gender-specified adult walrus at Diomede, Gambell, and Savoonga for 34, 39, and 33 years, respectively, and for unsexed adults for 36, 40, and 34 years, respec-


73

tively. Daily dates of catches for known-sex adult walrus at Diomede, Gambell, and Savoonga represented 83%, 83%, and 94%, respectively, of the known catches, and for all adult walrus (not differentiating gender) represented 84%, 87%, and 94%, respectively, of all the known catches.


Timing of the spring walrus hunt among villages and between regimes Overall, the mean timing of the hunt for females preceded that of males at Diomede by 5.5 days (paired t-test; p B0.01, N 25 comparison years) and Gambell by 2.3 days (paired t-test; p 0.02, N 34 comparison years). The timing of catches for males and females at Savoonga did not differ (paired t-test; p 0.97, N 26 comparison years). Differences in timing between sexes remained significant for all regime periods at Diomede, but only for the middle two regime periods at Gambell (paired t-test; p B0.05). Gender differences of less than a week were much less than the overall variability within a regime (figure 4); we, therefore, grouped sexes for comparisons among regimes and villages.

Figure 4. Mean catch date for adult male ( ) and female ( ) walrus retrieved by hunters from the villages of Diomede, Gambell, and Savoonga between 1952 and 2004. Note: General trends (lines of best fit) are indicated for male (× × × ×) and female (*) walrus. Vertical dashed lines represent boundaries between the four environmental regimes described between 1952 and 2004. k


74

M.D. Robards et al.

In the colder heavy ice conditions at the end of Regime 1 (Table 1), Savoonga’s walrus catches were almost a month later than those a decade earlier. The catches at Savoonga in the early 2000s (Regime 4) are significantly earlier than in the first two regimes (one-way ANOVA, pB0.05; post hoc Tukey’s test, pB0.05). Most strikingly, Savoonga catches had advanced in timing by nearly 6 weeks in 2004, compared to where they had been in the early 1970s. Timing showed no statistical difference at Diomede or Gambell for all regimes (one-way ANOVA, p 0.27). The northerly migration of walrus means that Gambell and Savoonga usually have the opportunity to catch walrus earlier than Diomede. Accordingly, we found that Gambell and Savoonga catches were both significantly earlier in timing than Diomede in Regime 1 (one-way ANOVA, pB0.01; post hoc Tukey’s test, pB0.01), but did not differ from each other (post hoc Tukey’s test, p 0.09 and 0.78, respectively). All three villages differed significantly from each other in timing during Regime 2 (one-way ANOVA, pB0.01; post hoc Tukey’s test, pB0.01). In the warmer Regime 3, only Diomede differed from Gambell (one-way ANOVA, p0.02; post hoc Tukey’s test, p B0.05). Savoonga’s timing did not significantly differ from that for Diomede or Gambell (Tukey’s test, p0.38, 0.19, respectively). In the most recent regime, Gambell and Savoonga were again both significantly earlier in timing than Diomede (one-way ANOVA, pB0.01; post hoc Tukey’s test, p B0.01). The Gambell hunting season was the most consistently timed of the three villages, with a long-term mean catch date of May 18th (SD 9.6 days, N 13,298 adult walrus). Overall, the long-term mean catch date at Savoonga was 5 days later on May 23rd (SD 13.7 days, N 10,667 adult walrus). Finally, Diomede, although exhibiting significant trends in timing during more recent regimes, was, in the longterm, only a little more variable than Gambell, although about 2 weeks later. The long-term mean catch date was June 4 (SD 10.2 days, N 10,114 adult walrus). Close examination of catch timing suggests strong directional change, with timing getting earlier at Diomede and Savoonga during Regime 3, followed by a rapid delay in timing at the transition to Regime 4 in 1998. Heavy pack ice during the spring of 1998 precluded Savoonga hunters from even launching boats for most of that season, reducing their catches (Garlich-Miller et al. 2006) and delaying timing. However, from 1998 until 2004, there was an increase in hunt success, and a continued directional trend of advanced timing of the spring hunt at all three villages (figure 4). We used the difference between mean catch dates of males and females at St. Lawrence Island and Diomede as a proxy for the speed of the migration across the Chirikov Basin (figure 1). Although there was no significant overall trend, our data suggest a consistently quicker passage by females (of about 3 days) compared to males, once they pass St. Lawrence Island in their approximately 23-week transit to Diomede Island (figure 5). To avoid bias from occasional catches of a few walrus prior to the main spring season hunting, we delineated a window of the spring hunt based on the first and last day on which at least five walrus were returned to a village (Table 2). Diomede generally had the shortest season and Savoonga the longest season of the three villages. The shorter spring hunting season for Diomede during Regimes 3 and 4 were accompanied by only half the days on which more than five walrus were caught, compared to Gambell and Savoonga. Nevertheless, the window of hunting should be treated cautiously because walrus may still be present once local needs are met and hunting effort declines.

75



Figure 5. Duration between mean date of spring walrus hunting at St. Lawrence Island and Diomede for adult male ( ) and female ( ) walrus. Note: General trends (lines of best fit) are indicated for male (. . . .) and female (*) walrus. k

Magnitude of spring walrus catches among villages and between regimes The size of walrus catches in different villages varied markedly among years and regimes, including when controlling for different human populations (figure 6; Table 3). ANCOVA results suggest significant differences between mean magnitude of catch among regimes, but not among villages; models including regime having the strongest effect compared to those including village or timing (Table 4). Catches rose to their maximum during Regime 2 (figure 7). Subsequently, catches declined, although at Diomede the change was greater than at Gambell and Savoonga. During Regime 4, catches of walrus at Gambell and Savoonga both increased, while Diomede diverged from what had been a parallel pattern among villages, catches continuing to decline (figure 7). Within-cell regression (ANCOVA) suggested no linear effect of timing on the magnitude of catch across the four regimes combined (F 1.16, p0.28; Table 4). Table 2. Mean window (days between first and last day when 5 walrus were returned to a village), and mean number of days when 5 walrus were returned during a season to the villages of Diomede, Gambell, and Savoonga. 19521976

Diomede Gambell Savoonga

19771988

19891997

1998

Window

Days 5

Window

Days 5

Window

Days 5

Window

Days 5

21 24 32

10 6 12

27 29 41

11 14 9

17 35 34

6 11 10

22 28 29

5 13 12

Note: Periods represent the four environmental regimes described between 1952 and 2004.We strongly emphasize that correlations in the data should be treated cautiously. Effects of social factors and walrus population size undoubtedly contribute to the observed patterns.


76

M.D. Robards et al.

Figure 6. The reported catch of walrus at Diomede, Gambell, and Savoonga. Note: Vertical dashed lines represent boundaries between the four environmental regimes described between 1952 and 2004. No data were available for 1959, 1978, 1990, or 1991. Correlations in the data should be treated cautiously due to socio-political factors that undoubtedly contribute to the observed patterns.

Table 3. Human populations, mean walrus catch, and per-capita catch rate (number of walrus per inhabitant) for the villages Diomede, Gambell, and Savoonga during the four environmental regimes described between 1952 and 2004.a Diomede

Gambell

Savoonga

Regime Population Catch Rate Population Catch Rate Population Catch Rate 1 2 3 4

88 139 178 146

397 660 147 94

4.5 4.7 0.8 0.6

358 445 525 649

223 654 328 476

0.6 1.5 0.6 0.7

304 491 519 643

271 414 230 367

0.9 0.8 0.4 0.6

a We strongly emphasize that correlations in the data should be treated cautiously. Effects of social factors and walrus population size undoubtedly contribute to the observed patterns.

77


Table 4. Results of the ANCOVA on the effects of village and regime on catches of Pacific walrus in the spring hunt at Diomede, Gambell, and Savoonga, with timing of the hunt as the covariate. Source


Regime Village Regime Village Regime Timing Village Regime Timing Village Timing Timing

df

MS

F

p

3,98 6,98 2,98 3,95 11,87 2,96 1,98

0.74 0.56 0.39 0.31 0.11 0.11 0.11

9.88 7.50 5.20 4.60 1.50 1.43 1.43

B0.01 B0.01 0.07 B0.01 0.15 0.24 0.28

However, although the relationship between timing and hunt success was parallel among villages (F 1.43, p0.24), the signs of this relationship were different among regimes (F 4.60, p 5 0.01). We investigated this more closely by plotting regressions of season timing versus catch for the four regimes (figure 8). In Regime 1, cold conditions and heavy ice led to more productive hunting during years when the spring hunt continued later into May and June. In the warmer conditions of Regime 2, with its less challenging ice conditions, catches were generally high and were less dependent on timing within a season. This pattern changed in the cooling conditions of Regime 3. During Regime 3, later hunting contrasted to those in Regime 1 with lower, as opposed to higher, catches. We attribute this to the more rapid ice recession, which limits hunting opportunities during the late season. Finally, in the variable conditions of Regime 4, catches at Diomede and Savoonga, as in Regime 2, were less dependent on timing, whereas Gambell’s catches continued to benefit in extended seasons. We attribute this to the ability of Gambell hunters to hunt in a wide array of conditions, including travel in excess of 160 km on a single foray (Benter, pers. comm.); so later seasons provide opportunities for a greater number of successful days (Table 2).

Figure 7. Mean catch (log10 transformed) of Pacific walrus at Diomede, Gambell, and Savoonga during the four environmental regimes described between 1952 and 2004.


78

M.D. Robards et al.

Figure 8. Relationship between catch (log10 transformed) of Pacific walrus and mean timing of the spring walrus hunt at the villages of Diomede (^, *), Gambell ( , . . . .), and Savoonga (h, ----) during the four environmental regimes described between 1952 and 2004. Note: Points are annual values. k

The role of social factors Interannual variability in the timing and magnitude of the spring walrus hunt at Diomede, Gambell, and Savoonga reflects the dynamic social-ecological system in which hunting takes place. Although the broad patterns we describe here correlate with climatic regimes and known walrus ecology, we strongly emphasize the contribution of social factors. Diomede, in particular, during the latter two regimes has been subject to a suite of societal changes that reduce their capacity to hunt walrus. These include human tragedies, incarceration of some key hunters, shifts in demographics such as an aging population, and emigration away from Diomede of young hunters; and changes in technology through loss of traditional large, oceangoing skin boats because of the loss of woman with skills required to prepare the necessary split walrus skins, which restricted hunting range and acceptable hunting conditions; and economic needs through reduced reliance on ivory carving. Village needs and preferences for specific components of the walrus population bias our data. Gambell hunters preferentially hunt females and calves, whereas Diomede hunters have historically focused on male walrus. Preferential hunting of a specific gender can delay the focus on the other. For example, purposely delayed taking of males at Gambell in 1987 was attributable to an extended period over which females were present. Subsequently, as females and low numbers of males continued moving north, and away from Gambell, the main movement of males was delayed due to late breakup of ice east of St. Lawrence Island. This shows up as an


79


outlier in our data (figure 4) lending credence to this methodology being sensitive to non-climatic factors. There have also been increasing tensions concerning local rules for maintaining the preceding hunting season for bowhead whales until the whales have passed at Diomede and Gambell (EWC 2003). Whaling requires quieter tactics, and has generally utilized skin boats moving under sail until recently, as compared to walrus hunting in motorized aluminum skiffs. Local rules in Gambell divided the whale and walrus hunting seasons at about the last week of April or first of May. In years when walrus appear early, they may not be hunted, or if they are, by fewer hunters until whaling is finished (Sease 1986; USFWS 2005). This factor may have limited the advance in timing for the Gambell walrus hunting season in our data by reducing walrus caught in April.

Discussion Variation in season of village walrus catches Despite profound climate changes in the northern Bering Sea, seasonality of walrus catches by gender and season have remained remarkably consistent at Gambell (figure 4), suggesting a consistent timing of walrus migration past that village since the 1930s (Collins 1940; Heinrich 1947). Nevertheless, local rules impeding the advance of walrus hunting earlier into the spring may have contributed to this pattern in recent years (only Regime 4). Strong currents through the Anadyr Strait (to the west of St. Lawrence Island) create and maintain open leads around Gambell, providing hunters with greater accessibility than at other villages, with easy access from both west- and north-facing beaches, resulting in a variety of conditions in which ice does not block access to walrus. Farther north, Diomede in the central Bering Strait also has strong currents and winds that open leads, even in dense ice, which allows limited but productive hunting. However, Diomede hunters frequently report ice conditions that preclude access from the village (e.g. blocking launch points), even when walrus are present. Variability of ice conditions is likely to be greater in the narrow Bering Strait due to the substantial constriction on ice movements, and helps explain the rapid changes in timing at Diomede during the most recent two regimes (figure 4). For example, the timing of the 1998 regime shift coincided with a 23-week change in timing of the Diomede hunt, affirming local hunter concerns about the great variability of walrushunting conditions. Hunting at Savoonga was the most variable in timing. Savoonga is situated on a generally north-facing beach and is susceptible to the frequent north winds that compact ice along the northern shore of St. Lawrence Island, precluding hunting directly from that village (because boats cannot be launched). Timing of walrus hunting at Savoonga was nearly 6 weeks earlier in 2004 as compared to the early 1970s at the end of the ‘cold regime’ (the only significant change in timing of hunt that we found for a village by regime; figure 4). In recent years, deterioration of sea ice allowed walrus to migrate closer to Savoonga between 1998 and 2004 and hunting has occurred earlier. Walrus are thought to remain as far north as they can during spring, based on seaice conditions (Burns 1970). Consequently, thinner ice conditions in recent years would be expected to allow walrus to move further north in late winter and early


80

M.D. Robards et al.

spring than in prior colder regimes. However, timing of the hunt at Gambell was not significantly altered between 1952 and 2004 (although local rules may hide some advance in timing). In contrast, the significantly earlier season at Savoonga suggests that the passage of walrus northward past St. Lawrence Island is now less restricted through Shpanberg Strait (to the east of St. Lawrence Island) due to earlier amelioration of sea-ice conditions in spring (Fay 1982). Although our results were not statistically significant, they suggest an increasingly rapid passage of males through the Chirikov Basin, supporting our contention of quicker passage north past Savoonga (figure 5). Reduced ice concentrations in the eastern Bering Sea (Stabeno et al. 2007) further support the more rapid passage of males through this area. Nevertheless, once past St. Lawrence Island, the female passage ahead of males in areas of strong current (i.e. Anadyr Strait, western and central Bering Strait) has remained, at least until 2004, as a consistent signature of the Pacific walrus spring migration.

Variation in size of village walrus catches The size of walrus catches has been reported elsewhere, although usually with respect to biological removals, rather than the village’s capacity to hunt (Fay et al. 1997; Garlich-Miller et al. 2006). Overall, catches at all three villages peaked in the middle of Regime 2. This period not only coincided with warming climatic conditions and reduced extent and quality of sea ice, but also with growing size of village populations (figure 3B) and introduction of more and faster aluminum boats on St. Lawrence Island. Also, during this period, the walrus population is thought to have attained its maximal recent numbers (figure 3A). High per-capita catch rates at Diomede during Regime 2 suggest a combination of access and capacity, as well as desire to catch greater numbers of walrus. The last two regimes supported significantly lower catches at Diomede, most likely a consequence of social changes. Catches of walrus are an emergent property of the complex system that makes up the humanwalrus subsistence relationship. Although interannual variability in catch is high, the long-term per-capita stability of the catch (except during the 1980s) is noteworthy (Table 3). Apart from during Regime 2, when catches were particularly high at Diomede, and to a lesser extent at Gambell, the per-capita catch rate has been relatively consistent at between 0.6 and 0.9 walrus per person per year. We suggest two environmental and one social hypothesis to explain per-capita stability. First, in environmental terms, hunting conditions may have deteriorated in parallel with increased capacity (through technology and number of hunting crews) to hunt; second, the walrus population may have declined enough to constrain success of hunting crews, thus requiring more hunters to sustain a village’s percapita return of walrus (catches support this for Gambell and Savoonga); third, in social terms, informal institutions may have restrained hunting despite their increased capacity to find and catch animals. Hunters have raised less concern about a reduced walrus population than about constraints on their ability to hunt under difficult conditions (Eskimo Walrus Commission 2003); however, core communities will still experience large numbers of walrus under a reduced population scenario due to the entire population passing close by.


81


Implications for communities A change in the seasonality or success of walrus hunting presents the communities with at least two challenges. First, the timing of walrus migrations alters the window of the hunt, and for places like Gambell, changes in timing of one subsistence hunt can conflict with other hunting activities such as whaling. Second, the size of catches caters to the economic and sustenance needs of communities, and shortfalls require substitution of other nutritional resources or sources of local income. Here, as elsewhere in the Arctic, subsistence-based communities have continued to adapt to the challenges presented by changing climatic conditions, including timing and size of marine mammal catches (Ford et al. 2008; Krupnik 1993), and new social conditions (AHDR 2004). Of particular note is the continued importance of the walrus hunt to communities. Adaptability of communities is facilitated by local knowledge, continued learning, strong social networks, flexibility in resource use, and institutional support (Ford et al. 2008; Metcalf and Robards 2008). Our results suggest that Gambell and Savoonga have been resilient in their contemporary ability to catch walrus (maintaining their per-capita catch rate). Recognizing that human populations have increased, community resilience is fostered through the increased capacity to hunt and efficiency at accomplishing goals (e.g. number of hunting crews, speed, and navigation). Gambell appears the best located of the three villages for continued successful hunting of walrus; strong winds from several directions, together with strong currents, provide more consistent hunter access to migrating animals in most years. The dangers associated with these winds and currents are under some circumstances mediated by opportunities to launch boats to either the west or north of the island (due to the villages’ location at the northwest tip of St. Lawrence Island; figure 1). Diomede hunters report a shorter season in recent years and less favorable ice conditions for hunting (EWC 2003). In conjunction, Diomede has a much reduced capacity to hunt based on number of boat crews compared to Gambell and Savoonga, and in the recent regimes have only had half as many successful walrus hunting days (those days where 5 walrus are returned). Savoonga, in contrast, has maintained a relatively consistent level of catch, despite profound changes in timing of their season. Savoonga is also hauling boats to the south side of the island now with snowmachines and hunting some walrus there in conjunction with whaling, although hunt monitors do not record the success of these new practices. A quicker walrus migration conspires with weather and economics, exacerbating local concerns about the future implications to economy, lifestyle, and cultural traditions. Historically, hunters utilized larger and slower boats and rarely travelled more than 40 km from villages. The more numerous and faster motor-powered boats now being used on St. Lawrence Island travel greater distances to find walrus when seasons get shorter, but increase the economic costs. Increased human populations, shorter hunting seasons, and the potential for reductions in the walrus population also create a scenario of greater competition among hunters and pressure on fewer animals. Such acknowledged pressures, and the increasing potential of regulation under the Marine Mammal Protection Act or Endangered Species Act, are being responded to by the communities of Gambell and Savoogna through their recently reestablished historical hunting ordinances, which include trip-hunting limits for walrus. Such moves reflect continued village adaptation within the prevailing environmental and social milieu.


82

M.D. Robards et al.

The advantages of living on cultural and ecological boundaries such as in the Bering Strait region support the notion that this provides a ‘greater capacity for flexibility’ and increases social-ecological resilience (Turner et al. 2003, p. 439). All three villages have adapted to extreme variability over the past millennia, but our results suggest that the village-specific responses to the recent climatic regime shifts have differed. As the transition from winter to summer has become more ‘liquid,’ Gambell seems the least vulnerable of the three villages to changes affecting walrus hunting. However, Savoonga may benefit from the new suite of environmental and technological conditions that leave it less vulnerable to the consequences of heavy ice, and with greater numbers of suitable hunting days (Kapsch et al. 2010). Nevertheless, Savoonga’s walrus hunters may need to travel further across the ocean or over land to other launching points to catch a sufficient number of walrus in a shorter season. Diomede may be the most vulnerable to environmental change due to social changes, a much lower village population, fewer hunting crews, and being subjected to a more rapid passage of walrus and greater variability of ice conditions. The continued existence of communities dependent on walrus for millennia, and a consistent catch-per-capita over the recent decades is indicative of an inherent adaptive capacity among hunters. Looking forward, the ability of communities to adapt and remain resilient in the face of environmental change will continue to be mediated by local weather conditions and the specific social milieu in which the acquisition of subsistence resources takes place. However, much greater attention is needed to explain the specific localized social factors that link both the ability and desire of hunters (individually and collectively) to acquire walrus under a range of local environmental conditions. Acknowledgements National Park Service’s Shared Beringian Heritage Program provided financial support. We acknowledge the help and hospitality of the communities of Bering Strait, and the Eskimo Walrus Commission, particularly Vera Metcalf, Chris Perkins, and Charles Brower. Igor Krupnik (Smithsonian Institution) and Henry Huntington provided impetus to this project through their invaluable advice. Bradley Benter, Brendan Kelly, Joel Garlich-Miller, Kae Lourie, Lori Quakenbush, and Scott Schliebe provided help accessing unpublished data. Amy Lovecraft, Peter Schweitzer, and Terry Chapin III (University of Alaska, Fairbanks) provided comments on earlier drafts.

References ACIA (ARCTIC CLIMATE IMPACT ASSESSMENT), 2005. Cambridge, UK: Cambridge University Press, 1042pp. AHDR (ARCTIC HUMAN DEVELOPMENT REPORT), 2004. Akureyri, Iceland: Stefansson Arctic Institute. BENSON, A.J., and TRITES, A.W., 2002, Ecological effects of regime shifts in the Bering Sea and eastern North Pacific Ocean. Fish and Fisheries, 3, pp. 95113. BOCKSTOCE, J.R., 1986, Whales, Ice and Men (Seattle, WA: University of Washington), 400pp. BURNS, J.J., 1965, The walrus in Alaska: its ecology and management. Federal Aid in Wildlife Restoration Report (Juneau: Alaska Department of Fish and Game), 48pp.



83

BURNS, J.J., 1969, Marine mammal report. Volume X, Annual Project Segment Report, Federal Aid in Wildlife Restoration (Juneau: Alaska Department of Fish and Game), 17pp. BURNS, J.J., 1970, Remarks on the distribution and natural history of pagophilic pinnipeds in the Bering and Chukchi Seas. Journal of Mammalogy, 51(3), pp. 445454. COLLINS, G., 1940, Habits of the Pacific walrus (Odobenus divergens). Journal of Mammalogy, 21(2), pp. 138144. DUERDEN, F., 2004, Translating climate change impacts at the community level. Arctic, 57(2), pp. 204212. ESKIMO WALRUS COMMISSION, 2003, Pacific walrus. Conserving our culture through traditional management. Report of the Oral History Project under Cooperative Agreement #701813J506, Section 119 (Nome, AK: Eskimo Walrus Commission, Kawerak, Inc.). FAY, F.H., 1982, Ecology and biology of the Pacific walrus, Odobenus rosmarus divergens Illiger. United States Department of the Interior, Fish and Wildlife Service, North American Fauna, number 74. FAY, F.H., EBERHARDT, L.L., KELLY, B.P., BURNS, J.J., and QUAKENBUSH, L.T., 1997, Status of the Pacific walrus population, 19501989. Marine Mammal Science, 13, pp. 537 565. FORD, J.D., 2009, Dangerous climate change and the importance of adaptation for the artic’s Inuit population. Environmental Research Letters, 4, pp. 19. FORD, J.D., SMIT, B., WANDEL, J., ALLURUT, M., SHAPPA, K., ITTUSARJUAT, H., and QRUNNUT, K., 2008, Climate change in the Arctic: current and future vulnerability in two Inuit communities in Canada. The Geographical Journal, 174(1), pp. 4562. GARLICH-MILLER, J.L., and BURN, D., 1999, Estimating the harvest of Pacific walrus, Odobenus rosmarus divergens, in Alaska. Fishery Bulletin, 97(4), pp. 10431046. GARLICH-MILLER, J.L., QUAKENBUSH, L.T., and BROMAGHIN, J.F., 2006, Trends in age structure and productivity of Pacific walruses harvested in the Bering Strait region of Alaska, 19522002. Marine Mammal Science, 22(4), pp. 880896. GEARHEARD, S., MATUMEAK, W., ANGUTIKJUAQ, I., MASLANIK, J., HUNTINGTON, H.P., LEAVITT, J., KAGAK, D.M., TIGULLARAQ, G., and BARRY, R.G., 2006, ‘‘It’s not that simple’’: a collaborative comparison of sea-ice environments, their uses, observed changes, and adaptations in Barrow, Alaska, USA, and Clyde River, Nunavut, Canada. Ambio, 35(4), pp. 203211. GREBMEIER, J.M., OVERLAND, J.E., MOORE, S.E., FARLEY, E.V., CARMACK, E.C., COOPER, L.W., FREY, K.E., HELLE, J.H., MCLAUGHLIN, F.A., and MCNUTT, S.L., 2006, A major ecosystem shift in the northern Bering Sea. Science, 311, pp. 14611464. HEINRICH, A., 1947, Their friend, the walrus. General Superintendent, Alaska Native Service. HUGHES, C.C., 1960, An Eskimo Village in the Modern World (Ithaca, NY: Cornell University Press). HUNT, G.L., STABENO, P., WALTERS, G., SINCLAIR, E., BRODEURE, R.D., NAPP, J.M., and BOND, N.A., 2002, Climate change and control of the southeastern Bering Sea pelagic ecosystem. Deep-Sea Research II, 49, pp. 58215853. HUNTINGTON, H.P., 2000, Using traditional ecological knowledge in science: methods and applications. Ecological Applications, 10(5), pp. 12701274. INTERGOVERNMENTAL PANEL ON CLIMATE CHANGE. 2007. Fourth assessment report. http:// www.ipcc.ch. Accessed 31 December 2011. KAPSCH, M.-L., EICKEN, H., and ROBARDS, M.D., 2010, Sea ice distribution and ice use by indigenous walrus hunters on St. Lawrence Island, Alaska. In SIKU: Knowing Our Ice: Documenting Inuit Sea-Ice Knowledge and Use, I. Krupnik, C. Aporta, S. Gearheard, G.J. Laidler, and L.E. Holm (Eds.), pp. 115144 (New York, NY: Springer). KRUPNIK, I., 1993, Arctic Adaptations: Native Whalers and Reindeer Herders of Northern Eurasia (Hanover, CT: University Press of New England), 375pp.


84

M.D. Robards et al.

KRUPNIK, I., and BOGOSLOVSKAYA, L., 1999, Old records, new stories: ecosystem variability and subsistence hunting pressure in the Bering Strait Area. Arctic Research of the United States, 13, pp. 1524. KRUPNIK, I., and RAY, G.C., 2007, Pacific walruses, indigenous hunters, and climate change: bridging scientific and indigenous knowledge. Deep-Sea Research II, 54, pp. 2946 2957. LAIDLER, G.J., 2006, Inuit and scientific perspectives on the relationship between sea-ice and climate change: the ideal complement? Climatic Change, 78, pp. 407444. LIU, J., DIETZ, T., CARPENTER, S.R., FOLKE, C., ALBERTI, M., REDMAN, C.L., SCHNEIDER, S.H., OSTROM, E., PELL, A.N., LUBCHENCO, J., TAYLOR, W.W., OUYANG, Z., DEADMAN, P., KRATZ, T., and PROVENCHER, W., 2007, Coupled human and natural systems. Proceedings of the National Academy of Sciences, 36(8), pp. 639649. MEEK, C.L., LOVECRAFT, A.L., ROBARDS, M.D., and KOFINAS, G.P., 2008, Building resilience through interlocal relations: case studies of walrus and polar bear management in the Bering Strait. Marine Policy, 32(6), pp. 10801089. METCALF, V., and ROBARDS, M.D., 2008, Sustaining a healthy human-walrus relationship in a dynamic environment: challenges for comanagement. Ecological Applications, 18(2), pp. S148S156. NIEBAUER, H.J., 1998, Variability in Bering Sea ice cover as affected by a regime shift in the North Pacific in the period 19471996. Journal of Geophysical Research, 103(C12), pp. 2771727737. OOZEVA, C., NOONGWOOK, C., NOONGWOOK, G., ALOWA, C., and KRUPNIK, I., 2004. Watching Ice and Weather Our Way. 208pp (Washington, DC: Arctic Studies Center, Smithsonian Institution). OVERLAND, J., RODIONOV, S., MINOBE, S., and BOND, N., 2008, North Pacific regime shifts: definitions, issues and recent transitions. Progress in Oceanography, 77, pp. 92102. PERRAULT, T., 2009, Environment and development. In A Companion to Environmental Geography, N. Castree, D. Demeritt, D. Liverman, and B. Rhoads (Eds.), pp. 442460 (Chichester, U.K.: Wiley-Blackwell). RAY, G.C., OVERLAND, J.E., and HUFFORD, G.L., 2010, Seascape as an organizing principle for evaluating walrus and seal sea-ice habitat in Beringia. Geophysical Research Letters, 37, pp. L20504, 6pp. ROBARDS, M.D., and JOLY, J.L., 2008, Interpretation of ‘wasteful manner’ within the Marine Mammal Protection Act and its role in management of the Pacific walrus. Ocean and Coastal Law Journal, 13(2), pp. 171232. RUDDELL, D., and WENTZ, E.A., 2009, Multi-tasking: scale in geography. Geography Compass, 3/2, pp. 681697. SEASE, J.L., 1986, Historical status and population dynamics of the Pacific walrus. Masters Thesis. University of Alaska, Fairbanks. SERREZE, M.C., HOLLAND, M.M., and STROEVE, J., 2007, Perspective on the Arctic’s shrinking sea-ice cover. Science, 315, pp. 15331536. SPECKMAN, S.G., CHERNOOK, V., BURN, D.M., UDEVITZ, M.S., KOCHNEV, A.A., VASILEV, A., and JAY, C.V., 2011, Results and evaluation of a survey to estimate Pacific walrus population size, 2006. Marine Mammal Science, 27(3), pp. 514553. SPRINGER, A.M., MCROY, C.P., and FLINT, M.V., 1996, The Bering Sea Green Belt: shelf-edge processes and ecosystem production. Fisheries Oceanography, 5, pp. 205223. STABENO, P.J., BOND, N.A., KACHEL, N.B., SALO, S.A., and SCHUMACHER, J.D., 2001, On the temporal variability of the physical environment over the south-eastern Bering Sea. Fisheries Oceanography, 10(1), pp. 8198. STABENO, P.J., BOND, N.A., and SALO, S.A., 2007, On the recent warming of the southeastern Bering Sea shelf. Deep-Sea Research II, 54, pp. 25992618. STABENO, P.J., and HUNT, JR, G.L., 2002, Overview of the inner front and southeast Bering Sea carrying capacity programs. Deep-Sea Research II, 49, pp. 61576168.



85

STABENO, P.J., and OVERLAND, J.E., 2001, Bering Sea shifts toward and earlier spring transition. Eos, Transactions American Geophysical Union, 82(29), p. 317. STOKER, S.W., 1983, Subsistence harvest estimates and faunal resource potential at whaling villages in northwestern Alaska. Prepared for Alaska Consultants Inc. In: Alaska Consultants Inc. with Stephen Braund & Associates. 1984. Subsistence study of Alaska Eskimo whaling villages. Prepared for the U.S. Department of the Interior. Appendix A. STROEVE, J., HOLLAND, M.M., MEIER, W., SCAMBOS, T., and SERREZE, M., 2007, Arctic sea ice decline: faster than forecast. Geophysical Research Letters, 34, p. L09501. TURNER, N.J., DAVIDSON-HUNT, I.J., and O’FAHERTY, M., 2003, Living on the edge: ecological and cultural edges as sources of diversity and resilience. Human Ecology, 31(3), pp. 439461. USFWS (UNITED STATES FISH AND WILDLIFE SERVICE), 2005, Field Report: 2005 Spring Walrus Harvest at Gambell, Alaska (Anchorage, AK: USDOI, FWS). VINCENT, K., 2007, Uncertainty in adaptive capacity and the importance of scale. Global Environmental Change, 17, pp. 1224. WALSH, J.E., and JOHNSON, C.M., 1979, An analysis of Arctic sea ice fluctuations, 195377. Journal of Physical Oceanography, 9(3), pp. 580591.