FACULTY OF BIOSCIENCES, FISHERIES AND ECONOMICS DEPARTMENT OF ARCTIC AND MARINE BIOLOGY
POLAR BEARS (URSUS MARITIMUS) IN THE BARENTS SEA AREA: Population biology and linkages to sea ice change, human disturbance and pollution
Magnus Andersen A dissertation for the degree of Doctor Philosophiae March 2013
Polar bears (Ursus maritimus) in the Barents Sea area: Population biology and linkages to sea ice change, human disturbance and pollution
By
Magnus Andersen
Norwegian Polar Institute
Dr. Philos. Thesis
University of Tromsø Faculty of Biosciences, Fisheries and Economics Department of Arctic and Marine Biology
2013
Table of contents 1. Acknowledgements
3
2. Summary
4
3. List of papers included in the thesis
6
4. Introduction
7
5. Objectives
12
5.1. Overall objective
12
5.2. Specific objectives
12
6. Results and discussion
13
6.1. Population biology and threats
13
6.1.1. Population size and development
13
6.1.2. Habitat use and sea ice reduction
15
6.1.3. Denning distribution and sea ice change
18
6.1.4. Human disturbance
22
6.1.5. Feeding habits
23
6.1.6. Levels of pollutants
25
6.2. Multiple stressors, monitoring and management 7. Concluding remarks
29 34
7.1. Specific conclusions
34
7.2. Overall conclusion
36
8. References
37
2
1. Acknowledgements This thesis is based on data collected by the Norwegian Polar Institute (NPI) during studies mainly funded by NPI, the Norwegian Ministry of the Environment, the Norwegian Research Council and the Governor of Svalbard. The findings presented are the result of the effort of many highly skilled specialists within several scientific fields, and I feel privileged to have worked with you all. The list of authors included in the publications is long, and I thank you all for the cooperation (in alphabetical order): Aksel Bernhoft, Andrew. E. Derocher, Andrei N. Boltunov, Carla Freitas, Christian Lydersen, Edmond Hansen, Elisabeth Lie, Gerald W. Garner (deceased), Jon Aars, Justin P. Gwynn, Janneche U. Skaare, Kit M. Kovacs, Martin Biuw, Mark Dowdall, Mette Skern-Mauritzen, Olga Pavlova, Stanislav E. Belikov, Stein Sandven, Steve T. Buckland, Tiago A. Marques and Øystein Wiig. From this list I want to thank Jon Aars, Andrew E. Derocher and Carla Freitas in particular, for giving me the opportunity to use the papers on which they are primary authors in my thesis. Øystein Wiig, Andrew E. Derocher, Jon Aars, Dag Vongraven and Kit M. Kovacs commented on the thesis. Thank you so much for your contributions! I have been lucky to work with some very enthusiastic polar bear biologists during more than a decade so far. Andrew E. Derocher introduced me to polar bear research and has since been an inspiration to me. Thanks for all the support you have given me during the work with this thesis! Jon Aars took over as polar bear researcher at NPI and continued the tradition of the polar bear program. Apart from a rather chilly swim in Storfjorden some years back, the work we have done together so far has been challenging, interesting and fun. This is a dr. philos. thesis and therefore I have had no formal supervisors. I have had several informal ones, though, and Øystein Wiig is the most important person filling that role. Not only are you a very nice person, you are also very knowledgeable and have the ability to inspire. Thanks for the help you have given me with this thesis! I work with a great group of people at NPI, and you are all part of the reason why I look forward to another day in the office. I have to specifically thank my master degree supervisor, Christian Lydersen, for introducing me to arctic marine mammalogy. Also, thanks for all the interesting discussions over a cup of coffee in the morning; they have all had at least a core of biology in them! Thanks also to Kit M. Kovacs, for being a good boss and friend and for everything you have done in all these years to make my job interesting and fun! Finally I would like to thank my family for all your support and for all the good times we have together! When I was a kid the Andersen family (Sidsel, Knut and little brother Auden) always spent a lot of time outdoors; fishing, hiking and skiing. My interest for nature grew with that and it became a very important part of my life. I thank you for that! Dear Ann Merete, Elise and Sigri: Nothing means more to me than the three of you. Thanks for making my life a busy, fun and exciting journey! Tromsø, March 2013 Magnus Andersen
3
2. Summary Polar bears in the Barents Sea population have been protected from hunting in Russia since 1956 and following the signing of the international Polar Bear Agreement in 1973 in Norway. This thesis seeks to summarise current knowledge on key population biology issues four decades after the Norwegian protection and almost six after the Russian. Further, it discusses threats that have developed in the decades following protection against human harvesting. It concludes with a discussion of the effect of multiple stressors on the population, and some thoughts on future research, monitoring and management. Polar bears in Svalbard and the Barents Sea area have been studied during the last 40 years with the aim of gaining knowledge regarding population biology and to evaluate potential sources of impact on the population from anthropogenic activity and changes to their habitat. The initial threat to polar bears in the region was unquestionably overharvest. Polar bear numbers were reduced quite drastically and hunting was clearly not sustainable. After the harvesting was stopped, the population grew in size to an estimated 2650 (1900-3600) in 2004. We believe that population recovery led to a wider distribution of maternity denning in the Svalbard Archipelago, compared to the period just after the protection of the population in 1973. However, during recent decades, the population has faced challenges from a variety of new anthropogenic impacts. The population has been exposed to a range of pollutants and an increasing level of human presence and activity within their range. Contaminants are bioaccumulated through the trophic levels in the marine food web, culminating in this top predator that consumes primarily ringed, bearded and harp seals. Females with small cubs use the land-fast sea ice for hunting, and are vulnerable to human disturbance. Changes in sea ice conditions also affect polar bears in the region, and reduced access to denning areas on the eastern islands of Svalbard is currently a concern. A decrease in spring land-fast ice close to important denning areas could negatively affect the survival of cubs. Research and monitoring provides advice to management bodies both locally and globally. Information on the presence of toxic compounds in High Arctic systems has resulted in progress in recent decades in having better control of harmful substances and in some cases international bans on their production and use. This has resulted in declining contaminant burdens in polar bears. Unfortunately, new harmful substances are finding their way to the Arctic, 4
while others, such as radionuclides, are stored locally (within Russian Territories) in large quantities, representing potential sources of pollution. The protection of important habitats locally with restrictions on motorized traffic may help reduce negative impacts from human activity on polar bears in the region. The fate of polar bears with regard to climate change is uncertain, but significant negative effects have been documented and these impacts are expected to increase in the coming decades. Relevant research and monitoring of polar bears is essential for future management of the species. The arctic environment should be managed in such a way that the combined effects of stressors on polar bear populations are minimized.
5
3. List of papers included in the thesis Paper 1. Aars J, Marques T A, Buckland ST, Andersen M, Belikov S, Boltunov A, Wiig Ø. 2009. Estimating the Barents Sea polar bear subpopulation size. Marine Mammal Science, 25: 35-52 Paper 2. Andersen M, Derocher AE, Wiig Ø, Aars J. 2008. Movements of two Svalbard polar bears recorded using geographical positioning system satellite transmitters. Polar Biology, 31: 501-507 Paper 3. Freitas C, Kovacs KM, Andersen M, Aars J, Sandven S, Mauritzen M, Pavlova O, Lydersen C. 2012. Importance of fast ice and glacier fronts for female polar bears and their cubs during spring in Svalbard, Norway. Marine Ecology Progress Series, 447: 289-304 Paper 4. Andersen M, Derocher AE, Wiig Ø, Aars J. 2012. Polar bear (Ursus maritimus) maternity den distribution in Svalbard, Norway Polar Biology, 35: 499-508 Paper 5. Derocher AE, Andersen M, Wiig Ø, Aars J, Hansen E, Biuw M. 2011. Sea ice and polar bear den ecology at Hopen Island, Svalbard. Marine Ecology Progress Series, 441: 273-279 Paper 6. Andersen M and Aars J. 2008. Short-term behavioural responses of polar bears to disturbance by snowmobiles. Polar Biology, 31: 501-507 Paper 7. Derocher AE, Wiig Ø, Andersen M. 2002. Diet composition of polar bears in Svalbard and the western Barents Sea. Polar Biology, 25: 448-452 Paper 8. Andersen M, Lie E, Derocher AE, Belikov SE, Bernhoft A, Boltunov AN, Garner GW, Skaare JU, Wiig Ø. 2001. Geographic variation of PCB congeners in polar bears (Ursus maritimus), from Svalbard to the Chukchi Sea. Polar Biology, 24: 231-238 Paper 9. Andersen M, Gwynn JP, Dowdall M, Lydersen C, Kovacs KM. 2006. Radiocaesium in marine mammals from Svalbard, the Barents Sea and the North Greenland Sea. The Science of the Total Environment, 363: 87-94
6
4. Introduction The polar bear (Ursus maritimus) is a large, charismatic mammal that represents both an important mythical symbol and a subsistence resource for local peoples of the Arctic, as well as being a “flagship” species in modern nature conservation. It is currently a highly political species that is iconic in the context of climate change. Polar bear science has a history that involves extensive international cooperation regarding both research and management, which has taken place over a period of about 50 years (Larsen and Stirling 2009). Polar bears are widely distributed across the circumpolar Arctic, including regions of drifting sea ice. The world population size is suggested to be 20,000 – 25,000 animals, contained within 19 sub-populations (Obbard et al. 2010). Polar bears are specialised predators that mainly feed on seals and other marine mammals (Stirling and Archibald 1977; Smith 1980), and because of their close association with the ocean they are characterised as marine mammals. Polar bears not only rely on sea ice to get access to their prey, but also as a substrate facilitating travel between hunting and denning habitats. Polar bears hunt primarily on sea ice, but they utilize land for denning throughout most of their range. Further, sea ice characteristics are important for reproduction in most polar bear populations, because males search the sea ice in spring to locate mates (Molnar et al. 2008). Polar bears are highly mobile and individuals can roam over large areas. However, significant variations in movement behaviours have been documented even within populations with home ranges varying from less than 200 square kilometres to almost 400 000 square kilometres in the Barents Sea region (Mauritzen et al. 2001). In Svalbard some bears move over the entire Barents Sea during their annual seasonal movements while others remain local within the Svalbard Archipelago (Wiig 1995; Mauritzen et al. 2001). Polar bears are generally found in low densities throughout the Arctic, but can also concentrate close to or on land during parts of the year, either during maternity denning in winter (for example Kong Karls Land, Norway) (Larsen 1986) or during summer and autumn when they are stranded until the sea ice freezes (for example Hudson Bay, Canada) (Derocher and Stirling 1990). Odd Lønø started to study the ecology of polar bears in Norway in an organised way in 1964. His early work was summarized in “The polar bear (Ursus maritimus, Phipps) in the Svalbard area”, published in 1970. This publication was the first to describe the population biology of polar bears in the archipelago. The issue of hunting and human impacts on the bears 7
was also thoroughly addressed. Lønø (1970) collected and analysed all available data on polar bear hunting in Svalbard up to a few years before protection of the population was enacted. His work documented the very high take of polar bears in this region from about 1870 to 1970. Between 100 and 900 bears were shot annually in northern Greenland and the Barents Sea region during this period. The hunt was controlled only to a limited degree, and it soon became apparent that the population was in danger of being extirpated if the harvest was allowed to continue (Anon 1965). The same situation was seen in other Arctic regions, and consequently international action to protect polar bears was initiated (Prestrud and Stirling 1994). In the late 1960s and early 1970s polar bears became an animal of political interest, and as more scientific data became available, it became clear that immediate action was needed if polar bear populations throughout the Arctic were going to be conserved (Anon 1965). Initiatives among the polar bear nations, which at that time were Canada, Denmark (now Greenland), Norway, the Soviet Union (now Russia), and the USA, which were facilitated by the IUCN, resulted in the signing of ”The Agreement on the conservation of polar bears” (hereafter called “the Agreement”) in 1973 (http://pbsg.npolar.no/en/agreements/agreement1973.html). Article II in the Agreement states that “each Contracting Party shall take appropriate action to protect the ecosystems of which polar bears are a part, with special attention to habitat components such as denning and feeding sites and migration patterns, and shall manage polar bear populations in accordance with sound conservation practices based on the best available scientific data”. Further, Article VII states that, to achieve this goal: “the Contracting Parties shall conduct national research programmes on polar bears, particularly research relating to the conservation and management of the species. They shall as appropriate co-ordinate such research with research carried out by other Parties, consult with other Parties on the management of migrating polar bear populations, and exchange information on research and management programmes, research results and data on bears taken”. This Agreement has subsequently spawned management actions and monitoring activities with the aim to secure the well-being of the world’s polar bears. During the period between the first meeting among polar bear nations in 1965 and the signing of the Agreement, the scientific “branch” of the negotiating parties established the IUCN/Species Survival Commission (SSC) Polar Bear Specialist Group (PBSG). The work of the PBSG was important in the process leading to the Agreement. It provided the necessary scientific data. The Parties to the Agreement have not met between 1981 and 2009, but the PBSG has 8
managed the Agreement and guided national authorities in their management of polar bears, and since 2009 they have been acting as an independent advisor to the Parties of the Agreement. Polar bears have now been included in The Bern Convention and the Washington Convention (CITES). In Norway, the Svalbard Environmental Protection Act (http://www.regjeringen.no/en/doc/ Laws/Acts/Svalbard-Environmental-Protection-ct.html?id=173945) defines how management of the environment in Svalbard shall be conducted, and several regulations are in place to protect polar bears and their habitat. The Norwegian Ministry of the Environment, who is responsible for nature conservation and management in Norway, has ambitious goals for the management of Svalbard and its wildlife, and the polar bear is a key species. Status of polar bears and population threats are therefore specifically dealt with in the Management Plan for Lofoten and the Barents Sea (Anon. 2010), which was presented to the Norwegian Parliament in March 2006, and revised in 2010. The first IPCC report to mention the consequences of climate change for sea ice cover in the Arctic was the Third Assessment, which was published in 2001 (IPCC 2001). Based on this report the IUCN asked the PBSG for a new evaluation of the international Red List status of polar bears, leading to a classification change from Near Threatened to Vulnerable in 2006. As part of the work with the national evaluation in the USA, under the Endangered Species Act, US Fish and Wildlife Service called for a meeting in 2007. Polar bear authorities from all polar bear nations (Range States) were invited, with an aim to exchange information about polar bear research and management and to discuss status of populations and measures to conserve the species. It was agreed that a meeting of the parties to the Polar Bear Agreement of 1973 should be held biannually. During the meetings in 2009 and 2011 the main goal was to develop a rangewide Action Plan for polar bears, and this work is still ongoing; it will be finalized at the next meeting in Russia in 2013 (www.polarbearmeeting.org). The above mentioned initiatives came as a response to concerns that had been raised about climate change effects on polar bears. Global warming (Comiso 2002; IPCC 2007; Comiso et al. 2008) is believed to represent a threat to polar bear populations throughout their range due to the declining area, connectivity (Sahanatien and Derocher 2012), and suitability of sea ice habitats for bears (Stirling and Derocher 1993; Derocher et al. 2004; Amstrup et al. 2008; Wiig et al. 2008; Durner et al. 2009; Molnar et al. 2010, 2011). The decrease in available habitat for polar bears and their main prey (ringed seals) may lead to reduction in population sizes and possibly to complete loss of some populations (Amstrup et al. 2008, 2010; Durner et al. 2009; Molnar et al. 9
2010). The Barents Sea population has been identified as one of the populations where predicted reductions in sea ice in coming decades are particularly severe (Durner et al. 2009). In 2010, an initiative was taken under the auspices of the Arctic Council working group Conservation of Arctic Flora and Fauna (CAFF) to develop a circumpolar monitoring plan for polar bears. A background paper was presented at the CAFF biennial meeting in February 2011 (Vongraven and Peacock 2011), and a circumpolar monitoring framework has been developed (Vongraven et al. in 2012). This framework identifies several threats and stressors on polar bear populations, identifies recommended monitoring parameters, knowledge gaps and suggestions on how to fill these gaps and improve monitoring. The conclusions are in agreement with threats previously identified for the Barents Sea population, in a plan designed to monitor Svalbard and Jan Mayen (MOSJ - Sander et al. 2005), but argue that a more comprehensive monitoring program is needed on a circumpolar level to coordinate monitoring activities, utilize monitoring capacities in a more efficient manner and facilitate monitoring that feeds into an adaptive management regime. The framework presented by Vongraven et al. (2012) uses the ecoregion classification concept, outlined in Amstrup et al. (2008). Polar bear populations throughout the Arctic are categorised according to the characteristics and predicted changes in the sea ice habitat (divergent, convergent, archipelago and seasonal sea ice). Vongraven et al. (2012) recommend that high intensity monitoring should be conducted in at least 6 of the 19 polar bear populations throughout the Arctic; the Barents Sea is one of the chosen areas. The Barents Sea population is chosen as a representative of a divergent sea ice ecoregion (Amstrup et al. 2008) because baseline data is available, there is a high risk of climate change effects and high pollution levels are well documented. The first polar bear was live-captured and tagged as part of the Norwegian polar bear research program in 1966 (Larsen 1967; 1970), initiating a new era in polar bear research and management in the region. In the years following, a range of population studies were conducted (e.g. Harington 1965; Lentfer 1969; Jonkel 1970; Larsen 1972). In 1975 concern was raised for the first time regarding high levels of pollutants found in polar bear tissues (Bowes and Jonkel 1975). The contaminant issue continues to be a significant threat to polar bear health (Obbard et al. 2010; Sonne 2010) and recent findings of effects on immune responses and metabolism highlights the complexity of this issue (Lie et al. 2004; Braathen et al. 2004; Lie et al. 2005; Villanger et al. 2011). Today, several polar bear monitoring programs take place because polar bears are seen as indicators of the environmental condition of the Arctic and because of 10
international obligations (Monitoring of Svalbard and Jan Mayen (MOSJ) (Sander et al. 2005), Arctic Monitoring and Assessment Program (AMAP) (AMAP 2009), Conservation of Arctic Flora and Fauna (Vongraven and Peacock 2011; Vongraven et al. 2012)). Since the time of protection of polar bears in Svalbard, the Norwegian Polar Institute has been responsible for the polar bear research and monitoring programme in Norway. The main aim of the programme has been to develop relevant knowledge needed by management authorities. As new questions have appeared, the program has adapted to answer these questions, while also maintaining a long-term perspective. The main focus of the NPI programme was initially to study the effect of hunting, but later pollution, anthropogenic development and tourism, and most recently climate change, have been major issues given address. Our understanding of the importance of distributional changes and abundance dynamics in relation to sea ice changes affiliated with climate change are growing with respect to polar bears. It is believed that polar bears worldwide will face significant challenges in the years ahead (e.g. Amstrup et al. 2008, 2010; Durner et al. 2009; Stirling and Derocher 2012). Further, questions of the combined effects of different stressors (e.g. climate change, pollution, harvest, human activity and disturbance) acting simultaneously have been raised (Jenssen 2006; UNEP/AMAP 2011; Dietz et al. 2013), and this issue will undoubtedly be given significant research attention in the future. It is thus more important than ever to study polar bear ecology with the aim of reducing negative human impacts on populations.
11
5. Objectives This thesis presents nine peer-reviewed papers published from 2000 until 2012, stemming from the NPI Polar Bear Research Program. The papers herein are based on various data collected during the period from 1972 to 2011.
5.1. Overall objective The main objective of the thesis is to describe key aspects of polar bear population biology in Svalbard and the broader Barents Sea Region, after hunting stopped in 1973, and to explore potential impacts of new threats such as sea ice change, human disturbance and pollution. The findings are discussed in relation to future management and monitoring of polar bears.
5.2. The specific objectives of the thesis are to: 1. Estimate the population size of polar bears in the Svalbard and Barents Sea area, to evaluate the current status of the population and provide a reference point for future monitoring. (Paper 1). 2. Study activity and habitat use of female polar bears in Svalbard, with the aim to describe movement behaviour, identify critical habitat and evaluate the effects of sea ice reduction on the population (Papers 2, 3 and 5). 3. Describe denning distribution and analyse the effect of sea ice reductions on denning (Papers 4 and 5). 4. Study behavioural responses of polar bears to the main type of motorized traffic (snowmobiles) in Svalbard, in the fast ice habitat (Paper 6). 5. Describe predator prey relationships in the population, through studying the diet of polar bears in Svalbard, and evaluate the numerical and energetic importance of different prey species (Paper 7). 6. Analyse tissues from polar bears from the Svalbard and adjacent populations to determine levels of anthropogenic contaminants (persistent organic and radionuclide pollutants) (Papers 8 and 9).
12
6. Results and discussion 6.1. Population biology and linkages to threats 6.1.1. A population estimate for the Svalbard and Barents Sea polar bears; how has the population developed during the last 40 years? The Barents Sea polar bear population is shared between Norway and Russia (Mauritzen et al. 2002), and has been protected against hunting since 1956 in Russia and 1973 in Norway (Prestrud and Stirling 1994). Larsen (1986) suggested that there were between 3000 and 6700 polar bears (depending on the population borders) in the Barents Sea in the beginning of the 1980s. This was based on data from multiple sources including den counts and spatially restricted non-random aerial surveys, which were extrapolated to larger areas. No study covering the whole area in question was available prior to the survey conducted in 2004 (Aars et al. 2009, hereafter Paper 1). Most population estimates for polar bears have been derived using capture-recapture methods (e.g., DeMaster et al. 1980; Taylor et al. 2005). But, obtaining sufficiently large sample sizes is time consuming and expensive (Wiig and Derocher 1999), but on the other hand the method yields valuable data on individuals for a range of other population ecology studies. Recent statistical developments have made distance sampling one of the most widely used methods for estimating animal abundance in the last decade (Buckland et al. 2004), and is today regarded as being more cost efficient than capture-recapture to achieve high levels of precision (Borchers et al. 2002), in particular for populations occurring at low densities over large areas, such as the Barents Sea polar bear population. Our study concluded that the Barents Sea population had approximately 2,650 (95% CI approximately 1,900–3,600) bears in August 2004. We found significant geographic variability in densities of bears across different types of habitats in the study area. The density of bears on land-fast ice and pack ice in the Russian areas to the east were much higher (> 2 bears/100 km2) than farther west in the Norwegian territories (0.4 - 1 bears/100 km2). The mean density of polar bears across the whole region was however, close to the densities described elsewhere in the Arctic (Taylor and Lee 1995; Evans et al. 2003). Polar bear spatial patterns are known to vary with both season and year. Individual polar bears in the Barents Sea show high seasonal fidelity to specific areas (Mauritzen et al. 2001). Many of the polar bears that are distributed around the islands of Svalbard in spring, are distributed along the ice edge further north-east in the Russian 13
area and around Franz Josef Land in August. During our survey there were three times as many bears in the Russian parts of the northern Barents Sea compared to the Norwegian area. Both the number of maternity dens (Larsen 1986; Andersen et al. 2012, hereafter Paper 4) and the relatively high number of recaptures of bears in the Svalbard area (Derocher 2005) indicate that more polar bears are present in the Svalbard area in spring compared to other times of the year. This is partly explained by the need for pregnant females to find suitable denning habitat on land and raise cubs in a stable ice habitat in spring. Bears may also be attracted by the generally good breeding habitat for ringed (Phoca hispida) and bearded seals (Erignathus barbatus) in Svalbard fjords and the resulting good spring hunting habitat for the polar bears, particularly on the east coast (see Freitas et al. 2012, hereafter Paper 3; Paper 4; Derocher et al. 2002, hereafter Paper 7). Between 1909 and 1970 an average of 320 polar bears were harvested annually in Svalbard and adjacent areas (Lønø 1970). Assuming an even sex ratio in the harvest, the sustainable take of a closed polar bear population under optimal conditions is considered to be 3.2 % (Taylor et al. 1987). This implies that the Barents Sea population should have numbered some 10,000+ polar bears to have sustained the recorded harvest. The harvest obviously was not sustainable, but the calculation still indicates that the historical population size must have been significantly higher than the current size. The large difference between this number and the upper confidence limit (3,600) of our estimate in 2004, after 40 years of protection is noteworthy. Larsen (1986) indicated that the population approximately doubled in size over a decade after protection in 1973, and suggested that there were close to 2,000 bears in the Svalbard area, and 3,000– 6,700 in the area between East Greenland and Franz Josef Land in 1980. The growth rate from then and up to 2004 is unknown. Changes in population age structure suggest that population growth has been positive, but also that the growth rate today is much lower than earlier (Derocher 2005). One possible explanation for the large difference in the estimated size in 2004 and the theoretical historical size (10,000) could be a significant immigration from less hunted neighbouring areas. However, the discrepancy between our recent estimate and the historical harvest levels are so significant that it is not likely that migration alone can explain the difference. We speculate that either the population size today is far from the carrying capacity of the region, or the carrying capacity has changed. Derocher et al. (2003) and Derocher (2005) suggested that the population recovery may have been slow after protection due to high levels of organic pollutants (see for example Andersen et al. 2001, hereafter Paper 8) 14
in polar bears in the area, having a negative effect on survival and reproductive rates (Derocher 2005). The time needed for the population to recover to its carrying capacity could therefore be longer than expected from demographic rates typical for other, less polluted, populations. The carrying capacity in the area may also have decreased during the last few decades and may continue to decrease in the future as a response to sea ice loss (Derocher 2005; Schliebe et al. 2006; Heggberget et al. 2006; Durner et al. 2009).
6.1.2. Movements and habitat use by polar bears, and their vulnerability to sea ice change Polar bears depend on sea ice as a platform for hunting ice-associated seals (Stirling and Archibald 1977; Smith 1980; Thiemann et al. 2008; Paper 7). Sea ice is also a platform for mating and travelling to and from terrestrial maternity denning areas (see Wiig et al. 2008; Derocher et al. 2011, hereafter Paper 5). Evidence of declines in polar bear body condition, reproductive success, survival and abundance have been documented in the Canadian Arctic and Beaufort Sea in Alaska, and are thought to be caused by nutritional limitations imposed by declining sea ice (Stirling et al. 1999; Regehr et al. 2007; Regehr et al. 2010; Rode et al. 2010; Stirling and Derocher 2012). It is essential to describe polar bear habitat use and identify especially important habitats to be able to make predictions regarding the future impacts of climate change. It is believed that polar bear habitat in Svalbard and the Barents Sea will be significantly reduced during the coming decades, and it has been suggested that the population will decrease as a consequence (Amstrup et al. 2008, 2010; Durner et al. 2009). The use of satellite telemetry in the study of polar bear movement and distribution was first applied between Svalbard, Norway and Greenland in 1979, when four polar bears were equipped with satellite transmitters (IUCN/SSC 1981; Larsen et al. 1983). One of the latest technological developments for studies of wildlife has been to use the Global Positioning System (GPS) to determine the location of animals (e.g., Johnson et al. 2002; Gau et al. 2004; Frair et al. 2004; Morales et al. 2004), and to use the Argos System to collect these data from the transmitters remotely (e.g., Yasuda and Arai 2005; Parks et al. 2006). Andersen et al. (2008) (hereafter Paper 2) was the first study that deployed GPS collars on polar bears, and also the first to investigate the effectiveness of GPS satellite collars in polar bear studies. The location data generated in this project described activity and movement patterns of 15
individuals in far more detail than previously reported in the polar bear literature (e.g., Messier et al. 1992; Wiig 1995; Amstrup et al. 2000; Mauritzen et al. 2001; Wiig et al. 2003). The data also described in great detail the behaviour of the animals, including data on diurnal activity patterns. Stirling (1974) described the behaviour of polar bears on the sea ice at Devon Island, Canada, through direct observations. He found that polar bears spent 66.6% of their time inactive (sleeping, lying and still-hunting). He related the activity patterns to the haul out behaviour of ringed seals, the main prey of polar bears. We found no diurnal activity pattern during summer when the bears were on sea ice. We did, however, find low activity (relative to other times of the day) late in the day when bears were stranded on land in summer and during long range directional movements northward on the pack ice in late summer. Both in winter and autumn the pattern was opposite, with higher activity late in the day compared to early in the day. Messier et al. (1992) studied the seasonal activity patterns of female polar bears in the Canadian Arctic. They found clear seasonal patterns with peak activity periods in May-July for all females, regardless of reproductive status, and concluded that there was a close link between activity and the behaviour of ringed seals. Females with cubs-of-the-year (COYs) had decreasing activity from June to November, low activity until March, and then increasing activity until June. In Paper 2, the bear that was accompanied by two COYs showed a pattern similar to that decribed by Messier et al. (1992). The movement rates we reported (Paper 2: maximum 4.6 km/h during a 4 hour period) are at the high end of those previously reported for polar bears (between 5.3 and 18.2 km/day)( Garner et al. 1990; Born et al. 1997; Ferguson et al. 2001; Wiig et al. 2003). Movement rates were affected by changes in the number of positions included in calculations of the paths, a pattern also discussed by others (Amstrup et al. 2001; Parks et al. 2006). We showed that low sampling frequency significantly underestimated actual movement rates, and home range estimates that were 30% smaller than those calculated using GPS data with a higher sampling frequency (lower step length). As the number of locations increased (step length decreased), home range estimates moved towards an asymptote; these findings are similar to those of Girard et al. (2002). Thus, in Paper 2, we showed how GPS collars are useful for studies of fine-scale habitat use, movement behaviour, energetics and activity patterns. For large scale studies, such as distribution and annual home range size, conventional Argos positioning collars may be suitable if the position collection frequency is sufficiently high. 16
The GPS collar technology gave us the opportunity to further explore polar bear habitat use on a fine scale (Paper 3). Previous studies have shown that polar bear distribution is significantly affected by sea ice concentration and sea ice type. Polar bears typically select ice concentrations ranging from 25 to 100%, depending on the season and the region (Stirling et al. 1993; Arthur et al. 1996; Ferguson et al. 2000a; Ferguson et al. 2001; Mauritzen et al. 2003; Durner et al. 2009). In the Canadian Arctic, females with COYs select landfast ice with pressure ridges during the spring, while lone adult females and males show strong preferences for ice-edge areas (Stirling et al. 1993). Females with COYs were thought to select landfast ice habitats to feed on ringed seal pups, and also to avoid adult males, that are rare in this habitat; male bears sometimes prey on cubs (Stirling et al. 1993). In Svalbard and the Barents Sea area, female polar bears with COYs also show a year-round tendency to be located on more solid ice than lone adult females (Mauritzen et al. 2003). In Paper 3 we found that female polar bears with COYs predominantly occupied inshore landfast ice areas during spring (April), and within this habitat they spent most time close to glacier fronts. In an aerial-survey study in the Canadian Arctic, Stirling et al. (1993) also reported that females with COYs showed a strong preference for landfast ice during spring. However, in Svalbard they concentrated their time in landfast ice close to glacier fronts while in the Canadian Arctic they selected fast-ice with snow drifts along pressure ridges, which were sometimes located far offshore. These preferred areas, in the respective locations, are linked to ringed seal pupping habitat. Ringed seals give birth during spring inside lairs that are constructed in snow that accumulates in stable sea-ice areas (Smith and Stirling 1975; Kingsley et al. 1985; Furgal et al. 1996). Nutritionally stressed polar bear females with COYs need a predictable food source when emerging from the maternity dens in spring and hence these ringed seal pupping areas are a vital resource. In such areas, the female bears hunt the ringed seal pups and sometimes their mothers (Stirling and McEwan 1975; Pilford et al. 2012; C. Lydersen, personal communication) without having to move long distances. Accordingly, most females with COYs in the present study spent their entire tracking period/spring in the landfast ice habitats, close to known denning areas (Paper 4). Ringed seals occur in high densities in landfast ice areas (Krafft et al. 2007) during April and bearded seals and harp seals (Pagophilus groenlandicus) also occur in the pack ice close to shore around Svalbard during spring (Haug et al. 1994; Isaksen and Wiig 1995) All of these 17
species have been recorded in the diet of polar bears from this area (Lønø 1970; Paper 7). Even if seal density is lower in the pack-ice, bearded and harp seals are larger prey and thus represent a larger energy package for polar bears than ringed seals. It is possible that female polar bears in Svalbard face a trade-off between being in landfast ice areas that provide a safe substrate (habitat), especially for cubs, and where prey items are predictable but small, and being in less stable drift ice where prey items are more unpredictable but also more profitable when they are captured. Paper 3 clearly emphasizes the importance of coastal fast-ice, in particular close to glacier fronts, for polar bear females with young cubs in Svalbard. Reductions in the extent of landfast ice have been observed in recent years in Svalbard (Haarpainter et al. 2001; Gerland and Hall 2006; Gerland et al. 2007; Høyland 2009). Glacier fronts that have contact with the ocean have also retreated in Svalbard in recent years (Blaszczyk et al. 2009). The eventual disappearance of these prey-rich and stable sea-ice habitats close to the preferred denning habitat, where polar bear with COYs concentrate during spring, is likely to alter present distribution and hunting patterns and also reduce the survival of cubs.
6.1.3. Maternity den distribution and the effect of sea ice reduction on denning behaviour The use of maternity dens in snow is a characteristic adaptation in polar bears to the harsh Arctic environment (Blix and Lentfer 1979). Polar bears typically den at low densities throughout the circumpolar Arctic, but concentrated denning areas exist at Wrangel Island, Russia (Belikov 1980), Kong Karls Land, Svalbard, Norway (Larsen 1985), and SW Hudson Bay, Canada (Jonkel et al. 1972). Most maternity dens are located on land, although a small amount of denning does occur in multiyear sea ice off the Alaskan coast (Harington 1968; Lentfer 1975; Amstrup and Gardner 1994; Fischbach et al. 2007). In Hudson Bay, polar bears den in earth dens that are sometimes far inland, up to 80 km from the coast but they move into snow dens as snow accumulates in autumn (Jonkel et al. 1972; Richardson et al. 2005). Denning philopatry among female polar bears has been shown in Hudson Bay (Ramsay and Stirling 1990), in Svalbard (Zeyl et al. 2010) and the Beaufort Sea (Amstrup and Gardner 1994). In Paper 4, we found that most maternity dens in Svalbard are close to the coast (< 10 km). Denning occurs in most parts of the Svalbard Archipelago, but the number of dens seems to 18
be highest in the eastern parts of the Archipelago. The six most important denning areas are: 1) north-western Spitsbergen, 2) southern Spitsbergen, 3) northern parts of Nordaustlandet, 4) Barentsøya and Edgeøya, 5) Kong Karls Land and 6) Hopen. Our data revealed that polar bears captured in the Svalbard or in the central parts of the Barents Sea also den in the Franz Josef Land Archipelago in Russia, as noted by Wiig (1998). Lønø (1970) suggested that denning in Svalbard was restricted to the eastern parts, including Kong Karls Land, Nordaustlandet, and along the northern part of the east coast of Spitsbergen. Larsen (1985) concluded that Kong Karls Land was the main denning area and that 90% of all dens in the Archipelago were on the islands Edgeøya, Barentsøya, Nordaustlandet and Kong Karls Land. The small island Hopen was not considered an important denning site by Lønø (1970) or Larsen (1985), because only a few observations of dens or females with COYs had been made there. Based on these earlier findings (Lønø 1970; Larsen 1985), we believe that denning distribution in Svalbard is currently wider than it was in the decades before protection from hunting in 1973. We suggest that this apparent expansion is a result of reestablishment of denning areas after a long period of harvest. Fidelity to denning areas by female polar bears (Ramsay and Stirling 1990; Zeyl et al. 2010) might have delayed re-establishment associated with the population recovery. Factors determining the distribution of polar bear dens are poorly understood but in Svalbard some areas can only be used for denning if sea ice in autumn reaches them, making them accessible (Paper 5). The linkage between denning and sea ice conditions has also been described by others (Ferguson et al. 2000b). Early snow cover is also necessary in most areas, and the terrain is important for snow accumulation. In the Beaufort Sea, about half of the dens were on drifting pack ice, half on land, and some few on landfast ice (Amstrup and Gardner 1994). There has been no evidence of offshore denning in Svalbard (Lønø 1970; Larsen 1986), and the highly dynamic sea ice conditions in the region may explain why this behavior is not seen. Our study indicates that most denning areas in Svalbard are close to fast ice areas where ringed seals give birth to their pups. This agrees well with the findings in Paper 3, that in April females with COYs use landfast ice areas and single females or females with older cubs use other habitat types more frequently. Human activities can influence polar bear denning distribution (Lentfer and Hensel 1980; Stirling and Andriashek 1992; Amstrup 1993). Svalbard has a long history of polar bear harvest 19
(Larsen 1986). A substantial part of the harvest occurred with the use of set-guns (a baited gun, typically built into a wooden box, which the bear fires when taking the bait) onshore, and females emerging from dens with COYs were particularly vulnerable (Lønø 1970). Set-guns were widely used and were very effective, but also non-selective. Larsen (1985) argued that denning in the Edgeøya region had been heavily affected by more than 70 years of hunting, and that only after 10 years of protection, in the early 1980s, the area was again used frequently for denning. The same could be the case for both Hornsund in the south and the fjords in the north of Spitsbergen, because both of these areas also experienced high hunting pressure during the decades before protection in 1973. Both trappers and station personnel hunted bears on Hopen from the early 1900s onwards (Lønø 1970). During the early to mid-1900s only two dens were recorded on Hopen (Lønø 1970). The reason for the larger number of dens on Hopen during the years 1995 to 2008 (Paper 5) compared to the earlier period when sea ice was likely more suitable for denning, is unknown, but it may be related to the difference in the number of adult females in the population. Between 1945 and 1970, an average of 41 bears per year were harvested at Hopen (total reported harvest on Hopen was 951 bears from 1946 to 1968; Lønø 1970). The population was thought to have been depleted before hunting ended in 1973 (Larsen 1986; Prestrud and Stirling 1994) and protection allowed the population to recover over the next 30 years (Derocher 2005). The larger number of maternity dens we observed may be a result of the re-establishment of Hopen as a denning area as the population increased (Papers 1 and 5). Papers 3 and 4 describe den distribution and female habitat use just after den emergence in spring, respectively. However, sea ice is also a critical habitat for female polar bears in autumn, when they prepare to enter their winter birthing dens (Paper 5). The dates of arrival and departure of sea ice near Hopen has varied substantially over time, reflecting its location near the southern edge of where sea ice exists in the Barents Sea (Shapiro et al. 2003). A trend towards later arrival of sea ice has been observed at Hopen, coinciding with a reduction in sea ice thickness observed over the last four decades (Gerland et al. 2008). The arrival of sea ice at Hopen in autumn shifted from late October to mid-December during the period from 1979-2010. In Paper 5, we show that fewer maternity dens were found on Hopen in years when sea ice arrived later in the autumn. If sea ice formed too late, no dens were found.
20
Further, later arrival of sea ice in autumn was correlated with lower body mass of adult females and their cubs at den emergence in the spring. This relationship suggests that recent environmental conditions have negatively affected female condition. Body mass is an indication of energy stores (Molnár et al. 2009) that are critical for supporting female polar bears during the denning period when energy is required for maternal maintenance, gestation and nursing until cubs can leave the den (Watts and Hansen 1987; Derocher et al. 1993; Molnár et al. 2011). Maternal body mass in spring has been correlated with body mass of cubs and with cub survival (Derocher and Stirling 1996, 1998). Our finding that cub mass was lower when the date of arrival of sea ice was later, suggests that the timing of arrival of pregnant females at den areas may impact reproductive success. After leaving the den, young polar bear cubs are vulnerable to hypothermia if exposed to cold water (Blix and Lentfer 1979; Aars and Plumb 2010). In most years, it was evident that there was sufficient sea ice for females with young cubs to leave Hopen without having to cross open water. However, the suitability of a maternity denning area for raising cubs is determined in part by the timing of sea ice arrival, sea ice departure and sea ice type and stability (see Paper 3). There is reason to believe that the fast ice habitat has deteriorated around Hopen in recent years, an effect of the generally lower sea ice concentration and thickness in the area. The reproductive success of females that manage to den on Hopen could be negatively affected if the sea ice departs earlier in spring in the future. Climate change is the most important conservation concern for polar bears due to the declining area, connectivity (Sahanatien and Derocher 2012), and suitability of sea ice habitats (Stirling and Derocher 1993; Derocher et al. 2004; Amstrup et al. 2008; Wiig et al. 2008; Durner et al. 2009; Molnar et al. 2010, 2011). The loss of one maternity denning area may not be a major cause for concern because females are able to den in other areas. However, the loss of habitat is symptomatic of larger ecosystem changes that cumulatively may threaten the persistence of polar bears (Hunter et al. 2010; Amstrup et al. 2010; Molnar et al. 2010, 2011). Further, the Hopen situation might reflect the situation at other important denning areas in Svalbard (Norwegian Polar Institute, unpublished data). Monitoring maternity denning areas at the margin of the polar bear range will be important to better understand how adult female polar bears, and ultimately the species, will respond as sea ice patterns change.
21
6.1.4. The effect of human disturbance of polar bears in the critical fast ice habitats Although polar bears are not harvested in Norway at present, as they were prior to the signing of the Polar Bear Agreement (Prestrud and Stirling 1994), they are still vulnerable to human presence and impacts (Lunn et al. 2002). Recreational activities (e.g. tourism, camping trips) are the source of most polar bear–human encounters in the Svalbard area. Tourism and the local use of snowmobiles have increased in Svalbard over the last 40 years (Overrein 2002; MOSJ 2012 (http://mosj.npolar.no/en/influence/traffic/indicators/snowmobile)). A large part of the snowmobile driving in Svalbard occurs on landfast sea ice, due to the steep and mountainous terrain. On the ice, polar bears hunt ringed seals (Paper 7), and the stable fast ice habitat is particularly important for females with COYs (Paper 3). The sea ice is also a substrate for movement between hunting habitats and denning areas (Mauritzen 2002; Papers 3, 4 and 5). Increasing anthropogenic activity in many Arctic regions made it important to have a more complete understanding of issues related to human disturbance of wildlife in the region. Studies of disturbance are rarely able to assess effects on survival or reproductive success or other effects at the population level. Population level studies would be extremely demanding both in terms of resources and effort, and we will therefore most likely have to depend on studies of effects on behaviour and physiological responses as indicators. Such studies can be valuable if the biology of the species is well understood and one can make plausible interpretations about how these responses link to demographic processes. Another limitation apparent in many disturbance studies is that the effect measured on an individual has a short duration. Cumulative population level effects are difficult to assess in most wild populations, and particularly in a longlived and highly mobile species such as the polar bear. Andersen and Aars (2008 - hereafter Paper 6) found that polar bears in Svalbard reacted to snowmobile disturbance at relatively long distances, although the variability between individuals was significant. Except for adult males, bears typically had a pronounced response and frequently fled from snowmobiles and continued to flee at long distances (up to 5 km). Females with COYs in particular showed strong reactions to this disturbance source. Polar bears are highly mobile on large temporal and spatial scales, but when considering small scale movement behaviour within a limited period of time (such as within a fjord) polar bears can have restricted movements (Paper 2 and 3). Stirling (1974) described the behaviour of polar bears on the sea ice at Devon Island, Canada, through direct observations. He found that 22
polar bears spent most of their time inactive. In Paper 6, we observed polar bears running at least one km after being disturbed by snowmobiles, and several bears left the ringed seal breathing hole where they were still-hunting when vehicles approached. We believe that repeated disturbance in this important fast ice habitat (Paper 3), leading to running and interrupted hunting could result in increased energetic stress on the animals, during a time when they are rebuilding energy stores that are critical for survival of cubs. Additionally, polar bears are not adapted to running quickly over extended distances, and large individuals in particular overheat quickly if pursued for very long (Øritsland 1970). Paper 6 demonstrates that females with cubs respond most strongly to snow-mobile disturbance, and the added stress experienced by the family group could have negative effects on reproductive success of females and perhaps even survival. Such stress could force polar bears to use sub-optimal habitats and spend more time in the water (polar bears tend to take refuge in water when startled). It could also lead to more frequently interrupted hunting situations or suckling/feeding bouts, which both could affect body condition and growth of both adults and cubs/COYs. Tourism and associated disturbance is a potential stressor that can act on a local spatial scale during short periods of the year. Local planning and regulations could significantly reduce the negative effects of tourism if relevant and sufficient knowledge of polar bear ecology locally is available. In Svalbard regulations on snowmobile traffic in sensitive areas in spring have been implemented, giving the authorities the ability to limit traffic, and reduce disturbance of females with cubs that have just emerged from their dens.
6.1.5. Spring diet of polar bears in the Svalbard and Barents Sea area Both movement patterns and choice of denning locations by polar bears can largely be explained by the accessibility of suitable prey. Polar bears are the most carnivorous of the ursids and are adapted to hunt seals on sea ice (Stirling and Archibald 1977; Smith 1980; Gjertz and Lydersen 1986; Stirling and Øritsland 1995). The diet of polar bears is still poorly understood in large parts of their range; little is known regarding the relative energetic contribution of prey species and the seasonal composition of prey. The only previous study of polar bear diet in Svalbard comes from bears harvested throughout the year near Svalbard, and remains of 52 ringed seals, 10 bearded seals, and 6 harp seals that were found in their stomachs (Lønø 1970). Harp seals were only found during summer 23
(June-August) and most bearded seals (9/10) were found in the same period. These findings were similar to the composition of the 114 samples of known species in Paper 7 (76% ringed seal, 15% bearded seal, and 9% harp seal). Similar to earlier studies, ringed seals are the dominant prey of polar bears numerically. However, on a biomass basis, the results from Lønø (1970) together with our study suggest that the diet of polar bears in Svalbard and the western Barents Sea has a significant contribution from bearded seals, due to their large body size compared to ringed seals (Andersen et al. 1999). In the eastern Barents Sea, a Russian study reported 68% ringed seal, 22% walrus (Odobenus rosmarus) and other miscellaneous items in the diet of polar bears (Parovshchikov 1964), perhaps reflecting further geographic variation in the same population. Most information on polar bear diet from our areas is from spring, but Iversen (2011) reported findings based on scat samples from both spring and summer. Their study showed that polar bears in Svalbard feed on eggs, reindeer (Rangifer tarandus platyrhynchus) and vegetation in addition to seals. Reindeer predation by Svalbard polar bears was also documented by Derocher et al. (2000), and Hedberg et al. (2011) who found fatty acids in polar bear milk that indicated that they had fed on reindeer. Distribution and abundance of marine mammal resources available to polar bears in Svalbard are only partially described. Bearded seals are widely distributed throughout Svalbard and the western Barents Sea (Benjaminsen 1973). The abundance of bearded seals is uncertain but may number a few hundred thousand in the North Atlantic (Burns 1981; Kovacs et al. 2009). The size of the ringed seal population in the Svalbard area is unknown but the global population likely numbers in the millions (Reeves 1998; Kovacs et al. 2009). In Svalbard and the western Barents Sea, ringed seals give birth in both landfast ice (Smith & Lydersen 1991) and in drifting pack ice (Wiig et al. 1999). Paper 3 showed that landfast ice is especially important for female polar bears with COYs, and explained this by the combination of a stable substrate and the high density of ringed seal breeding lairs in this habitat. The Barents Sea harp seal population is approximately 2.2 million animals (Nilssen et al. 2000) and represents a seasonally abundant food source for polar bears. However, harp seals do not usually reach polar bear habitat until April-May and then increase in abundance along the drift ice edge until October when they return south (Haug et al. 1994; Nordøy et al. 1998). Annual home range size of adult female polar bears in Svalbard ranged from 185 to 373,539 km2 (Mauritzen et al. 2001) and dietary differences were postulated to explain the 24
different space use patterns (Mauritzen et al. 2001). In particular, Mauritzen et al. (2001) suggested that near-shore bears relied more on the landfast ice and preyed largely on ringed seals during spring, while pelagic bears preyed more on bearded and harp seals over a longer period. Paper 3 and Paper 7 support this suggestion, because ringed seal kills were most numerous in landfast ice areas in spring and bearded and harp seal kills were mainly found in pack ice areas in summer.
6.1.6. Levels of different pollutants in Svalbard polar bears; consequences of a polluted diet The relevance of feeding in explaining levels of contaminants in wildlife is often discussed (see Paper 8 and Andersen et al. 2006, hereafter Paper 9). McKinney et al (2009) linked diet, sea ice changes and changes in contaminant levels in Hudson Bay polar bears. They documented how the change in diet, as a result of sea ice change and prey availability, increased the levels of several contaminants in bear tissue(s). Thus, prey composition is an important element in understanding the ecotoxicology of polar bears. If climate change alters the distribution and abundance of prey (Stirling and Derocher 1993), documentation of the current predation patterns is essential for understanding how exposure to environmental pollutants might vary as a result of climate change (McKinney et al. 2009). A wide range of manmade environmental pollutants have been transported by air and ocean currents from southern industrialised areas to the Arctic during the last decades, among them organochlorines (OCs)(Oehme 1991; Barrie et al. 1992; De March et al. 1998). These compounds are highly lipophilic and persistent to biological degradation; they accumulate in the marine environment and biomagnify up food chains (Muir et al. 1988; Barrie et al. 1992). Arctic organisms are adapted to dealing with short periods of high production during which lipid energy stores are built, resulting in high dependence on fat at most trophic levels (Barrie et al. 1992). Polar bears have the capacity to metabolize several organic pollutants (Letcher et al. 2000), but the metabolites resulting from this process are believed to have an even more negative effect than the original compounds (Cheek et al. 1999; Marchesini et al. 2008; Gutleb et al. 2010). PCBs were first identified in polar bears in the 1970s (Bowes and Jonkel 1975). Svalbard polar bears have shown PCB levels comparable to those found in ringed seals from the Baltic Sea, where reproductive disorders were reported (Norheim et al. 1992; Olsson et al. 1992; 25
Bernhoft et al. 1997). In polar bears at Svalbard, a possible immunotoxic effect (Bernhoft et al. 2000) and negative association between OCs and retinol and thyroid hormones have been reported (Skaare et al. 2000). Studies indicate that negative effects of organic pollutants on immune response and metabolism exist in polar bears (Lie et al. 2004, 2005; Braathen et al. 2004; Villanger et al. 2011). Contaminants in polar bears have been studied in most parts of the species range (Norstrom et al. 1998). However, limited data from most parts of the Russian Arctic have precluded an understanding of circumpolar PCB patterns. Paper 8 demonstrated regional variation in PCB contamination in polar bear blood between the European, Russian and western North American Arctic regions. We found the PCB levels to be highest in the western part of the Russian Arctic, and that the relative contribution of the low chlorinated congeners increasing while the higher chlorinated congeners decrease from west to east. Further, the study showed that the proportion of the PCB congeners 118 and 156 were higher in the Chukchi Sea compared to Svalbard. These two congeners represent the most acutely toxic congeners in this study. We believe that the variation observed in the study is due to different PCB exposure between the regions. Variation in PCB congener levels and patterns could be explained by regional prey differences. Polar bears are typically considered to be predators of ringed and bearded seals (Stirling and Archibald 1977; Smith 1980), which feeds on sympagic and benthic species (Gjertz and Lydersen 1986; Hjelset et al. 1999). However, in some populations, polar bears feed on pelagic feeding harp seals (Lønø 1970), and benthic feeding white whales (Delphinapterus leucas), narwhals (Monodon monoceros) (Lowry et al. 1987; Smith and Sjare 1990) and walruses (Calvert and Stirling 1990; Ovsyanikov 1995). It is known from satellite tracking of individuals, that polar bears in the Kara and Laptev seas spend considerable amounts of time in multiyear ice (Belikov et al. 1998). This is also the case for parts of the population in Svalbard and Franz Josef Land (Wiig 1995). If the structure of the ice-associated food web in these areas causes greater bioaccumulation of contaminants compared to other areas, this could result in higher levels of PCB in these polar bears. Our findings are based on analyses of blood samples from polar bears captured in five different geographic regions. Our data are homogenous in that only samples from adult females were included. Further, all females were captured in spring. However, there are differences in age, body condition and reproductive status at capture. Females emerging from dens with young 26
cubs are very lean (Derocher and Stirling 1998), while others who might have lost their cubs or have older cubs would have been feeding for a longer period before capture and may have been in better condition. Both long- and short-term differences in feeding history (and thus body condition) presumably influence the concentrations and patterns of organochlorines, and this can be a problem, particularly when considering blood sample analyses (Lydersen et al. 2002). In addition, females with offspring can shed PCBs through milk and this also complicates interpretation of the results of contaminant analyses (Bernhoft et al. 1997; Polischuk 1999; Bytingsvik et al. 2012). Movement behaviour further complicates the issue. For example, Olsen et al. (2003) explained differences in contaminant levels as a result of varying activity seen in small versus large home range sizes in Barents Sea polar bears. As mentioned earlier, geographic variation in feeding habits may not only result in geographic variation in contaminant levels, it may also affect the relative patterns observed in the compounds. Since different species have a varying ability to metabolise contaminants in their food (Wolkers et al. 2004), the path these compounds travel up the food web will determine the pattern seen in the upper trophic levels. Organic pollutants are typically lipophilic and are accumulated in fatty tissues. Other groups of contaminants, for example radionuclides, use other pathways, but nevertheless end up at higher levels in top predators. The tendency for Arctic marine food chains to be dependent on benthic and sea ice associated systems provides an efficient mechanism for biomagnification of contaminants, and in combination with the longevity of marine mammals this results in high uptake rates of radionuclides (e.g. Pentreath et al. 1982; Aarkrog et al. 1997; Brown et al. 1999; Carroll et al. 2002). 137
Cs makes its way into the Arctic marine environment via global fallout from
atmospheric weapon testing, discharges from European reprocessing and power facilities and fallout from the Chernobyl accident in 1986. The monitoring of radioactivity in Arctic marine mammals is important for a number of reasons. Information on current levels of contamination is required for monitoring, for the understanding of impacts and behaviour of radionuclides in arctic ecosystems and in the evaluation of potential consequences of future contamination on specific species. The low current low
137
137
Cs levels observed in the marine mammals studied in Paper 9 reflect the
Cs activity in sea water in the European Arctic, following the reduction in 27
discharges from reprocessing facilities at Sellafield, UK in the mid-1970s. Recently reported 137
Cs activities in sea water from the study area ranged from 2.0 to 3.4 Bq/m3, compared to peak
values of 20 to 45 Bq/m3 for the Svalbard area and Barents Sea in the 1980s (Hallstadius et al. 1982; Kershaw and Baxter 1995; Strand et al. 2002). A large number of potential local sources of radionuclide contamination are known in the region (e.g. nuclear reactor dump sites and radioactive waste, atmospheric nuclear bomb testing sites on Novaya Zemlya). It appears from our data in Paper 9 that these potential sources currently have little impact on marine mammals in the European Arctic. A number of studies have shown that 137Cs biomagnifies through marine food chains (e.g. Calmet et al. 1992; Kasamatsu and Ishikawa 1997; Watson et al. 1999; Heldal et al. 2003), but that this happens mostly at lower trophic levels (Brown et al. 2004). Paper 9 has shown that 137Cs contamination of marine mammals in the European Arctic region is low at present. Comparison of concentration factors suggests that
137
Cs is biomagnified through marine food chains through
to seal species, while the situation with regard to further trophic transfer to polar bears remains unclear. In general, pollution is acting across large temporal and spatial scales, potentially having negative effects on polar bear reproduction and survival in several populations. The ban on PCB usage is an example of how positive results can be achieved, as the decreasing trends of these contaminant in arctic biota show (Wolkers et al. 2008). New compounds are, however, being detected in polar bear tissues, calling for new research and management initiatives. While most sources of organic pollutants are found outside the Arctic, and output is continuous but slow, radioactive contaminant sources are found many places in the Arctic and the potential for acute and significant contamination is present. Nuclear power plants and waste disposal sites represent potential sources of contamination and comprehensive plans for managing and monitoring such sources are needed.
28
6.2.The multiple stressor issue and perspectives on future monitoring and management A stressor has been defined as a variable (biotic or abiotic) that adversely affects individual physiology or population performance (Barrett et al. 1976; Vinebrooke et al 2004). Stressors can be both natural and anthropogenic, and they often interact to produce a combined impact, which can be synergistic, additive or antagonistic towards the organism. Stressors are synergistic when their combined effect is larger than the effect of their individual effects added together (additive) and antagonistic describes when the combined effect is smaller (Vinebrook et al. 2004). In this thesis I have shown how sea ice changes, disturbance and pollution are threats to polar bears that act on different spatial and temporal scales. Consequently, they are linked to different aspects of polar bear population biology as stressors. The multiple stressor perspective is currently receiving progressively more attention from management authorities and the scientific community (for example Vinebrooke et al. 2004; Jenssen 2006; Obbard et al. 2010; Vongraven and Peacock 2011; UNEP/AMAP 2011; Vongraven et al. 2012; Dietz et al. 2013). A recent AMAP report (AMAP 2011) concluded that “the complex processes involving transport pathways,
intercompartmental
distribution,
bioaccumulation,
and
transformation
of
anthropogenic contaminants will be affected by the recently observed climate change in the Arctic environment”. They describe how the behaviour of pollutants can change with regard to both abiotic and biotic properties as the climate changes. Further, they noted that climate change is expected to result in increased human development and increased contaminant discharge in the Arctic. A number of knowledge gaps with regard to the combined effects of climate change and pollution were identified, and it is clear that further research is needed. Jenssen (2006) pointed out that pollutants with endocrine-disruptive properties are the second most serious anthropogenic threat in the Arctic, after climate change, and that the combination of these two stressors may be a “worst-case combination for Arctic marine mammals and birds”. Dietz et al. (2013) stated that different contaminants (for example various POPs and heavy metals) can act together making it difficult to determine the effect of individual compound in free-ranging animals. Further, they noted that confounding factors such as age, sex, reproductive status, body condition, and presence of diseases or other stressors further complicates analyses. They suggested nevertheless that increasing trends of mercury (Hg) in 29
polar bears from northeast Canada and East Greenland might represent a health risk to the most susceptible animals when stress from climate change, shifts in pathogen organisms, decreased access to food and other contaminations are simultaneously taken into account (Dietz et al. 2013). It has been documented that contaminants have negative effects on thyroid hormones, sex steroid homeostasis and the immune system of marine mammals, including polar bears (Haave et al. 2003; Olsen et al. 2003; Oskham et al; 2003, 2004, Braathen et al. 2004; Lie et al. 2004, 2005; Letcher et al. 2010). Both sea ice changes and human disturbance, although acting on quite different scales, may reduce access to food or increase energy expenditure through less effective hunting (including less prey) or longer walking or swimming distances. In combination with a generally weaker state of health, the net effect of the stressors might affect survival significantly, especially for very young or old animals. The effect of a longer ice-free period in Hudson Bay lowered survival in these two age groups (Regehr et al. 2007). The Hudson Bay population is not considered to have problems with pollution, and disturbance from humans is limited to the period when bears are fasting on land. However, there is reason to believe that there will be an increase in disturbance and human-bear conflicts as climate change progresses, both in this population and elsewhere in the Arctic. Lowered survival as a result of sea ice changes was also found in polar bears in the Southern Beaufort Sea (Regehr et al. 2010). For Svalbard, analyses of the effect of sea ice reduction on survival have not yet been conducted. However, the contamination issue has been suggested as a possible explanation for the slow population recovery after the heavy harvest stopped in 1973 (Derocher 2005). In addition, polar bear habitat in this region is projected to be significantly reduced in the decades to come. The reduction of available habitat will probably lead to decreasing population size (Amstrup et al. 2008, 2010, Durner et al. 2009). Sea ice in the Svalbard region is characterized by active drift ice combined with stable landfast ice. Even if ringed seal breeding has been documented in the drift ice, the landfast ice is thought to be essential to normal pup production levels (Smith and Lydersen 1991). During recent years, pup production in Svalbard has been very low in several important ringed seal breeding fjords (Kovacs et al. 2011). It is reasonable to believe that polar bear females with small cubs might be particularly vulnerable to the loss of this potential prey, because these females are the most nutritionally stressed group of bears at this time of the year (Watts and Hansen 1987; Atkinson and Ramsay 1995). Since fat-soluble contaminants are released into blood circulation when 30
stored fat is metabolized, the mothers experience high levels of contaminants carried to vital organs during the first months of lactation, which causes transfer of pollutants to the foetuses in utero and to the cubs through the milk (Polischuk et al. 2002). It is reasonable to believe that the combined effect of pollution and sea ice reductions in the Barents Sea population acts in an additive, or synergistic, negative fashion (Derocher 2005; Jenssen 2006). Further research and monitoring is needed to increase our understanding of the effects of these stressors on reproduction and survival in this population. Recently, a paradigm shift has been suggested in the field of biological monitoring, with the introduction of the term adaptive monitoring (Lindenmayer and Likens 2009). The key feature of adaptive monitoring is that a monitoring process should evolve as an iterative process as new knowledge emerges or as new questions arise. The need for a conceptual model and the importance of choosing the relevant monitoring parameters is also highlighted by Lindenmayer and Likens (2009). Monitoring of polar bear populations is a perfect example of why adaptative monitoring is so important because knowledge about important sources of impacts on this species have changed through time. The need for better monitoring programs for polar bears that seek to understand processes related to the population level effects of a range of stressors has recently been identified (Vongraven and Peacock 2011; Vongraven et al. 2012). The idea of an adaptive monitoring program also includes the concept of adaptive management (Lindenmayer and Likens 2009; Vongraven et al. 2012). When changes occur in a population, management regimes should change accordingly. In this context it is important to realise that population changes might be rapid as thresholds are crossed, and that plans for how to deal with such changes should be made early (Andrew E. Derocher personal communication). Management regimes and strategies might be challenged as new threats appear. One must ask - what is the ultimate aim of polar bear circumpolar management? Is maximising harvest the goal or is it conservation of the species. Aims are obviously different among different management jurisdictions. Currently some stakeholders believe that polar bears are not in danger of extinction and that proper resource management is sufficient to ensure the survival of populations (see Wiig 2005; Vongraven 2009). In other words, it has been argued that more powerful conservation measures, such as lowered quotas or total protection of animals and their habitats are not needed at this time (see Vongraven et al. 2012 for discussion). This may currently 31
be true for some polar bear populations, but for others the status is far more uncertain. Some populations are declining (Hunter et al. 2010; Regehr et al. 2007, 2010), for others data is unavailable (Obbard et al. 2010). Therefore a more careful conservation approach should be taken. The Barents Sea population is currently one of few populations that is totally protected and some critical habitat has varying degrees of protection, but this is restricted to areas inside Norwegian territorial waters. In this population, the precautionary principle has been used as a conservation tool. Overharvesting and poaching were the main conservation concerns when the work on the Polar Bear Agreement was initiated in the 1960s (Anon 1965; Prestrud and Stirling 1994). From 1870 to 1970, several polar bear populations were overharvested, which lead to the implementation of quota systems in some populations and total protection in others (Prestrud and Stirling 1994). The on-going and future large-scale habitat losses in combination with other threats such as pollution and human development are much more serious challenges, which require a broad management approach. It is generally agreed that the rapid climate change seen during the last century is caused mainly by human activity (IPCC 2007), but whether changes in climate are reversible is still under discussion. Amstrup et al. (2010) argued that a significant reduction in the emission of CO2 to the atmosphere could reduce the rate of sea ice loss, and consequently increase polar bear population persistence. It has been suggested that strict conservation measures must be initiated to secure the survival of the species (Vongraven and Peacock 2011; Vongraven et al. 2012), and that this can only be made possible through a coordinated effort from all Range States. Similar more general advice has been given to the international community with regard to conservation of mammals globally (Rondinini et al. 2011). Rondinini et al. (2011) identified key elements for a successful large scale conservation strategy to include an institution with recognised authority, clear goals and objectives, relevant species data, a priority list and well developed indicators. The cooperation around the Convention on Biological Diversity (http://www.cbd.int) was proposed as a possible starting point for a new global initiative for the conservation of mammals. Rondinini et al. (2011) further suggested that an expanded version of the International Union for the Conservation of Nature (IUCN) Red List (http://www.iucnredlist.org/) would be a suitable future tool. Wilson et al. (2011) pointed out that one of the main challenges in mammal conservation is prioritizing what to focus on, since “we cannot do everything, everywhere, all the time”. 32
However, Rondinini et al. (2011) stressed that a global mammal conservation strategy is urgent, and although there are still significant knowledge gaps, the work cannot be delayed. The recent initiatives taken internationally to improve knowledge and monitoring of polar bear populations throughout the range of the species (e.g. Anon 2009; Vongraven et al. 2012) are in line with these views. Work on the international Action Plan for polar bears is ongoing, and a framework for improved research and monitoring has been developed. It is the responsibility of researchers, management authorities and other stakeholders to cooperate in gaining the necessary knowledge to provide sound management advice and conservation action. It is important for management strategies to be based on the best available scientific knowledge to ensure best practice with respect to preserving the global population of polar bears.
33
7. Concluding remarks 7.1. Specific conclusions 1. The population size of polar bears in the Svalbard and Barents Sea area We estimated that the Barents Sea polar bear population contained approximately 2,650 (95% CI approximately 1,900–3,600) bears in August 2004. The estimated size of the Barents Sea polar bear population is much smaller than the minimum size that must have been present prior to the high hunting pressure from 1870 to 1970. This indicates that, historically, the population was much larger and that the population acted as a sink for animals from other areas.
2. Activity and habitat use of female polar bears in Svalbard We found both monthly and diurnal patterns in fine scale polar bear movements, with maximum movement rates above 4 km/h during 4 hour periods. Simulations showed that a commonly used sampling regime of one location every 6th day would have significantly underestimated the movement rates and the home range sizes compared to our estimates, thus GPS collars with high accuracy and high sampling frequency are essential for fine scale analyses of habitat use. Using this technology, we found that space use patterns differed according to reproductive state; females with COYs had smaller home ranges and used fast-ice areas close to glacier fronts more frequently than lone females. The eventual disappearance of these important habitats might become critical for the survival of polar bear cubs in Svalbard and other regions with similar habitat characteristics.
3. Denning distribution and the effect of sea ice variability on denning behaviour The highest number of dens was recorded on the islands in the eastern and northern parts of the Svalbard Archipelago with fewer dens found further west on the island Spitsbergen. The majority of dens (62%) in Svalbard were located on land within ca. 1 km of the shore. Our observations of den distribution indicates that denning is now more widespread in the archipelago compared to 40-50 years ago and reflects a reestablishment of denning areas following decades of protection. The arrival of sea ice at Hopen Island in autumn shifted from late October to midDecember during the period 1979 to 2010. Fewer maternity dens were found on the island in years when sea ice arrived later in the autumn. Later arrival of sea ice in the autumn was 34
correlated with lower body mass of adult females and their cubs at emergence. Timing of arrival and departure of sea ice is already affecting the denning ecology of polar bears at the southern extent of their range in Svalbard.
4. The effect of disturbance on polar bears in important landfast ice habitats Females with cubs and single medium sized bears tended to show more intense responses to motorized vehicle disturbance in landfast ice habitats than adult males or lone adult females. On average, bears were alerted to snowmobiles at 1,164 m, and showed locomotive responses at 843 m. There was a statistically significant difference in reaction distance between sex and age classes. The response intensity was affected by wind direction. Female polar bears with COYs may be at greater risk via disturbance, since they react at greater distances with amplified responses.
5. Diet of polar bears in Svalbard and the Barents Sea Prey composition was dominated numerically by ringed seals (63%), followed by bearded (13%), harp seals (8%) and unknown species (16%). When known prey were converted to biomass, the total diet composition was dominated by bearded seals (55%), followed by ringed seals (30%) and harp seals (15%), which indicated that bearded seals are an important dietary item for polar bears in the western Barents Sea. Different patterns of space use by different bears may result in geographic variation of diet within the same population.
6. Levels of anthropogenic contaminants (persistent organic and radionuclide pollutants) Our results indicate that polar bears from Franz Josef Land and the Kara Sea have the highest PCB levels in the Arctic. Decreasing trends were seen eastwards and westwards from this region. Of the congeners investigated in the present study, the lower chlorinated PCBs increase and the high chlorinated compounds decrease from Svalbard eastward to the Chukchi Sea. Different pollution sources, compound transport patterns and regional prey differences could explain variation in PCB congener levels and patterns seen in polar bears. The results of our radionuclide study indicated low specific activities of
137
Cs in Arctic marine mammals in the
Barents and Greenland Seas. Concentration factors (CF) of 137Cs from seawater were determined 35
for polar bears, ringed, bearded, harp and hooded seals. Mean CF values were higher than those reported for fish and benthic organisms in the literature, suggesting bioaccumulation of
137
Cs in
the marine ecosystem.
7.2. Overall conclusion The biological traits that make polar bears well adapted to the Arctic environment are problematic in the context of encounters with human activity, pollution and significant changes to their sea ice habitat. This thesis has described how polar bears in Svalbard have been negatively affected by human activity in the last century, but that the threats have changed through time. There are clear linkages between population biology and current anthropogenic threats, and it is reason to believe that the combination of several stressors have significant negative effects on polar bears. It seems clear, however, that the processes involved and the population level effects are not well understood. An international Action Plan for polar bears is under construction and a comprehensive monitoring program that aims to understand the consequences of multiple stressors, has been recommended by an international expert group. Norway is obliged to manage the Norwegian population based on the best available scientific data, as stated by Article VII of the Agreement, and thus should follow the advice given by the group. Improving future management of the species requires relevant research and monitoring through increased scientific effort in the Barents Sea population. The arctic environment as a whole should be managed in such a way that the combined effects of stressors on populations, including polar bears, are minimized.
36
8. References Aarkrog A, Baxter MS, Bettencourt AO, Bojanowski R, Bologa A, Charmasson S, Cunha I, Delfanti R, Duran E, Holm E, Jeffree R, Livingston H, Mahapanyawong H, Nies S, Osvath I, Pingyu L, Povinec P, Sanchez A, Smith J, Swift D. 1997. A comparison of doses from 137Cs and 210Po in marine food: a major international study. Journal of Environmental Radioactivity, 34: 69-90 Aars J, Marques TA, Buckland ST, Andersen M, Belikov S, Boltunov A, Wiig Ø. 2009. Estimating the Barents Sea polar bear subpopulation size. Marine Mammal Science, 25: 35-52 Aars J, Plumb A. 2010. Polar bear cubs may reduce chilling from icy water by sitting on mother's back. Polar Biology, 33: 557-559 AMAP. 2009. Arctic Pollution 2009. Arctic Monitoring and Assessment Programme, Oslo. Norway, 83pp AMAP. 2011. Combined Effects of Selected Pollutants and Climate Change in the Arctic Environment. By: Kallenborn R, Borgå K, Christensen JH, Dowdall M, Evenset A, Odland JØ, Ruus A, Aspmo Pfaffhuber K, Pawlak J, Reiersen L-O. Arctic Monitoring and Assessment Programme (AMAP), Oslo, Norway, 108 pp. Amstrup SC. 1993. Human disturbance of denning polar bears in Alaska. Arctic, 46: 246-250 Amstrup SC, DeWeaver ET, Douglas DC, Marcot BG, Durner GM, Bitz CM, Bailey DA. 2010. Greenhouse gas mitigation can reduce sea-ice loss and increase polar bear persistence. Nature, 468: 955-958 Amstrup SC, Durner GM, McDonald TL, Mulcahy DM, Garner GW. 2001. Comparing the movement patterns of satellite tracked male and female polar bears. Canadian Journal of Zoology, 79: 2147-2158 Amstrup SC, Durner GM, Stirling I, Lunn NN, Messier F. 2000. Movements and distribution of polar bears in the Beaufort Sea. Canadian Journal of Zoology, 78:948-966 Amstrup SC, Gardner C. 1994. Polar bear maternity denning in the Beaufort Sea. Journal of Wildlife Management, 58: 1-10 Amstrup SC, Marcot BG, Douglas DC. 2008. A Bayesian network approach to forecasting the 21st century worldwide status of polar bears. In: Arctic sea ice decline: observations,
37
projections, mechanisms, and implications. DeWeaver ET, Bitz CM, Tremblay LB (eds) American Geophysical Union, Washington, DC, pp 213-268 Andersen M, Derocher AE, Wiig Ø, Aars, J. 2012. Polar bear (Ursus maritimus) maternity den distribution in Svalbard, Norway. Polar Biology, 35: 499–508 Andersen M, Derocher AE, Wiig Ø, Aars J. 2008. Movements of two Svalbard polar bears recorded using geographical positioning system satellite transmitters. Polar Biology, 31:905-911 Andersen M, Hjelset AM, Gjertz I, Lydersen C, Gulliksen B. 1999. Growth, age at sexual maturity and condition in bearded seals (Erignathus barbatus) from Svalbard, Norway. Polar Biology, 21: 179-186 Andersen M, Aars J. 2008. Short-term behavioural response of polar bears to disturbance by snowmobiles. Polar Biology, 31: 501-507 Andersen M, Gwynn JP, Dowdall M, Lydersen C, Kovacs KM. 2006. Radiocaesium in marine mammals from Svalbard, the Barents Sea and the North Greenland Sea. The Science of the Total Environment, 363: 87-94 Andersen M, Lie E, Derocher AE, Belikov SE, Bernhoft A, Boltunov AN, Garner GW, Skaare JU, Wiig Ø. 2001. Geographic variation of PCB congeners in polar bears (Ursus maritimus), from Svalbard to the Chukchi Sea. Polar Biology, 24: 231-238 Anon. 1965. Proceedings of the first international scientific meeting on the polar bear. Fairbanks, Alaska, USA, 6-10 September 1965, 72 pp. Anon. 2009. Meeting of the parties to the 1973 Agreement on the Conservation of Polar Bears Tromsø, Norway, 17 – 19 March 2009. Outcome of meeting. 7 pp. Anon. 2010. Meld. St. 10 (2010–2011) Melding til Stortinget. Oppdatering av forvaltningsplanen for det marine miljø i Barentshavet og havområdene utenfor Lofoten ( In Norwegian). 143 pp. Arthur SM, Manly BFJ, McDonald LL, Garner GW.1996. Assessing habitat selection when availability changes. Ecology 77: 215-227 Atkinson SN, Ramsay MA. 1995. The effects of prolonged fasting on the body-composition and reproductive success of female polar bears (Ursus maritimus). Functional Ecology, 9: 559-567 Barrett GW, Vandyne GM, Odum EP. 1976. Stress Ecology. Bioscience, 26: 192-194 38
Barrie LA, Gregor D, Hargrave B, Lake R, Sherer R, Tracey B, Bidleman T. 1992. Arctic contaminants: sources, occurrence and pathways. Science of the Total Environment, 122: 1-74 Belikov SE. 1980. Distribution and structure of dens of female polar bears in Wrangel Island. International Conference on Bear Research and Management, 4:117 Belikov SE, Garner GW, Wiig Ø, Boltunov AN, Gorbunov YA. 1998. Polar bears of the Zevernaya Zemlya Archipelago of the Russian Arctic. Ursus, 10: 33-40 Benjaminsen T. 1973. Age determination and the growth and age distribution from cementum growth layers of bearded seals at Svalbard. Fiskeridirektoratets Skrifter, Serie Havundersøkelser, 16: 159-170 Bernhoft A, Skaare JU, Wiig Ø, Derocher AE, Larsen HJS. 2000. Possible immunotoxic effects of organochlorines in polar bears (Ursus maritimus) at Svalbard. Journal of Toxicology and Environmental Health, 59: 561-574 Bernhoft A, Wiig Ø, Skaare JU. 1997. Organochlorines in polar bears (Ursus maritimus) at Svalbard. Environmental Pollution, 95:159-175 Blaszczyk M, Jania JA, Hagen JO. 2009. Tidewater glaciers of Svalbard: Recent changes and estimates of calving fluxes. Polish Polar Research, 30: 85-142 Blix AS, Lentfer JW. 1979. Modes of thermal protection in polar bear cubs - at birth and on emergence from the den. American Journal of Physiology, 236: 67-74 Borchers DL, Buckland ST, Zucchini W. 2002. Estimating Animal Abundance. Closed populations. Springer-Verlag, London Born EW, Wiig Ø, Thomassen J. 1997. Seasonal and annual movements of radio-collared polar bears (Ursus maritimus) in northeast Greenland. Journal of Marine Systems, 10: 67-77 Bowes GW, Jonkel CJ. 1975. Presence and distribution of polychlorinated biphenyls (PCB) in arctic and subarctic marine food chains. Journal of the Fisheries Research Board of Canada, 32: 281-286 Braathen M, Derocher AE, Wiig Ø, Sørmo EG, Lie E, Skaare JU, Jenssen BM. 2004. Relationships between PCBs and thyroid hormones and retinol in female and male polar bears. Environmental Health Perspectives, 112: 826-833
39
Brown J, Børretzen P, Dowdall M, Sazykina T, Kryshev I. 2004. The derivation of transfer parameters in the assessment of radiological impacts on the arctic marine biota. Arctic, 57: 279 – 289 Brown JE, Kolstad AK, Brungot AL, Lind B, Rudjord AL, Strand P, Føyn L. 1999. Levels of 99Tc in seawater and biota samples from Norwegian coastal waters and adjacent seas. Marine Pollution Bulletin, 38: 560-571 Buckland ST, Anderson DR, Burnham KP, Laake JL, Borchers DL, Thomas L, (eds). 2004. Advanced Distance Sampling. Oxford University Press, Oxford. Burns JJ. 1981. Bearded seal Erignathus barbatus Erxleben, 1777. In: Ridgway SH, Harrison RJ (eds) Handbook of marine mammals, vol. 2: Seals. Academic Press, London, pp 145170 Bytingsvik J, Lie E, Aars J, Derocher AE, Wiig Ø, Jenssen BM. 2012. PCBs and OH-PCBs in polar bear mother-cub pairs: A comparative study based on plasma levels in 1998 and 2008. Science of the Total Environment, 417: 117-128 Calmet D, Woodhead D, Andre JM. 1992. 210Pb, 137Cs, and 40K in three species of porpoises caught on the eastern tropical Pacific Ocean. Journal of Environmental Radioactivity, 15: 153-169 Calvert W and Stirling I. 1990. Interactions between polar bears and overwintering walruses in the Central Canadian High Arctic. International Conference on Bear Research and Management, 8: 351-356 Carroll J, Wolkers H, Andersen M, Rissanen K. 2002. Bioaccumulation of radiocaesium in Arctic seals. Marine Pollution Buletin, 44: 1366-1371 Cheek, AO, Kow, K, Chen, J, McLachlan, JA. 1999. Potential mechanisms of thyroid disruption in humans: interaction of organochlorine compounds with thyroid receptor, transthyretin, and thyroid-binding globulin. Environmental Health Perspectives, 107: 273 Comiso JC. 2002. A rapidly declining perennial sea ice cover in the arctic. Geophysical Research Letter 29: L1956 Comiso JC, Parkinson CL, Gersten R, Stock L. 2008. Accelerated decline in the Arctic sea ice cover. Geophysical Research Letter, 35: L01703
40
De March BGE, de Wit CA, Muir CG, Braune BM, Gregor DJ, Norstrom RJ, Olsson M, Skaare JU, Stange K. 1998. Persistent Organic Pollutants. In: AMAP Assessment Report: Arctic Pollution Issues, Wilson SJ, Murray, JL, Huntington HP (eds), Oslo, Norway, pp 183-371 DeMaster DP, Kingsley MCP, Stirling I. 1980. A multiple mark and recapture estimate applied to polar bears. Canadian Journal of Zoology, 58: 633-638 Derocher AE. 2005. Population ecology of polar bears at Svalbard, Norway. Population Ecology, 47: 267–275 Derocher AE, Andriashek D, Arnould JPY. 1993. Aspects of milk composition and lactation in polar bears. Canadian Journal of Zoology, 71: 561-567 Derocher AE, Lunn NJ, Stirling I. 2004. Polar bears in a warming climate. Integrative and Comparative Biology, 44: 163-176 Derocher AE, Stirling I. 1996. Aspects of survival in juvenile polar bears. Canadian Journal of Zoology, 74: 1246-1252 Derocher AE, Stirling I. 1998. Maternal investment and factors affecting offspring size in polar bears (Ursus maritimus). Journal of Zoology, 245: 253-260 Derocher AE, Stirling I. 1990. Observations of aggregating behaviour in adult male polar bears (Ursus maritimus). Canadian Journal of Zoology, 68: 1390-1394 Derocher AE, Wiig Ø, Andersen M. 2002. Diet composition of polar bears in Svalbard and the western Barents Sea. Polar Biology, 25: 448-452 Derocher AE, Wiig Ø, Bangjord G. 2000. Predation of Svalbard reindeer by polar bears. Polar Biology, 23: 675-678 Derocher AE, Andersen M, Wiig Ø, Aars J, Hansen E, Biuw M. 2011. Sea ice and polar bear den ecology at Hopen Island, Svalbard. Marine Ecology Progress Series 441: 273-279 Derocher AE, Wolkers H, Colborn T, Schlabach M, Larsen TS, Wiig Ø. 2003. Contaminants in Svalbard polar bear samples archived since 1967 and possible population level effects. The Science of the Total Environment, 301: 163-174 Dietz R, Basu N, Braune B, O’Hara T, Scheuhammer T, Sonne C, Andersen M, Andreasen C, Andriashek D, Asmund G, Aubail A, Baagøe H, Born EW, Chan HM, Derocher AE, Grandjean P, Knott K, Kirkegaard M, Lunn N, Messier F, Obbard M, Olsen MT, Peacock E, Renzoni A, Rigét FF, Skaare JU, Stern G, Stirling I, Taylor M, Wiig Ø, 41
Wilson S, Aars J. 2013 What are the Toxicological Effects of Mercury in Arctic Biota? The Science of the Total Environment, 443: 775-790 Durner GM, Douglas DC, Nielson RM, Amstrup SC, Mcdonald TL, Stirling I, Mauritzen M, Born EW, Wiig Ø, DeWeaver E, Serreze MC, Belikov SE, Holland MM, Maslanik J, Aars J, Bailey DA, Derocher AE. 2009. Predicting 21st century polar bear habitat distribution from global climate models. Ecological Monographs, 79: 25-58 Evans TJ, Fischbach A, Schliebe S, Manly B, Kalxdorff S, York G. 2003. Polar bear aerial surveys in the eastern Chukchi Sea: A pilot study. Arctic, 56: 359-366 Ferguson S, Taylor M, Born E, Rosing-Asvid A, Messier F. 2001. Activity and movement patterns of polar bears inhabiting consolidated versus active pack ice. Arctic, 54: 49-54 Ferguson SH, Taylor MK, Messier F. 2000a. Influence of sea ice dynamics on habitat selection by polar bears. Ecology, 81: 761-772 Ferguson S, Taylor M, Rosing-Asvid A, Born E, Messier F. 2000b. Relationships between denning of polar bears and conditions of sea ice. Journal of Mammalogy, 81: 1118-1127 Fischbach AS, Amstrup SC, Douglas DC. 2007. Landward and eastward shift of Alaskan polar bear denning associated with recent sea ice changes. Polar Biology, 30: 1395-1405 Frair JL, Nielsen SE, Merrill EH, Lele SR, Boyce MS, Munro RHM, Stenhouse GB, Beyer HL. 2004. Removing GPS collar bias in habitat selection studies. Journal of Applied Ecology, 41: 201-212 Freitas C, Kovacs KM, Andersen M, Aars J, Sandven S, Mauritzen M, Pavlova O, Lydersen C. 2012. Importance of fast ice and glacier fronts for female polar bears and their cubs during spring in Svalbard, Norway. Marine Ecology Progress Series, 447: 289-304 Furgal CM, Innes S, Kovacs KM. 1996. Characteristics of ringed seal, Phoca hispida, subnivean structures and breeding habitat and their effects on predation. Canadian Journal of Zoology, 74: 858-874 Garner GW, Knick ST, Douglas DC. 1990. Seasonal movements of adult female polar bears in the Bering and Chukchi Seas. International Conference on Bear Research and Management, 8: 219-226 Gau RJ, Mulders R, Ciarniello LJ, Heard DC, Chetkiewicz CLB, Boyce M, Munro R, Stenhouse G, Chruszcz B, Gibeau ML, Milakovic B, Parker KL. 2004. Uncontrolled field
42
performance of Televilt GPS-simplex (Tm) collars on grizzly bears in western and northern Canada. Wildlife Society Bulletin, 32: 693-701 Gerland S, Renner AHH, Godtliebsen F, Divine D, Løyning TB. 2008. Decrease of sea ice thickness at Hopen, Barents Sea, during 1966-2007. Geophysical Research Letter, 35: L06501 Gerland S, Hall R. 2006. Variability of fast-ice thickness in Spitsbergen fjords. Annals of Glaciology 44: 231-239 Gerland S, Gascard JC, Ivanov B, Nielsen CP, Nilsen F, Pavlova O, Leu E, Tverberg V, Barrault S. 2007. Fast ice evolution in Kongsfjorden compared with other Svalbard fjords. 8th Ny-Ålesund seminar, Cambridge, UK , October 2007, 3 pp Girard I, Ouellet J-P, Courtois R, Dussault C, Breton L. 2002. Effects of sampling effort based on GPS telemetry on home range size estimations. Journal of Wildlife Management, 66: 1290-1300 Gjertz I, Lydersen C. 1986. Polar bear predation on ringed seals in the fast-ice of Hornsund, Svalbard. Polar Research, 4: 65-68 Gutleb AC, Cenijn P, van Velzen M, Lie E, Ropstad E, Skaare JU, Malmberg T, Bergman A, Gabrielsen GW, Legler J. 2010. In Vitro Assay Shows That PCB Metabolites Completely Saturate Thyroid Hormone Transport Capacity in Blood of Wild Polar Bears (Ursus maritimus). Environmental Science & Technology, 44: 3149-3154 Haarpaintner J, Haugan PM, Gascard JC. 2001. Interannual variability of the Storfjorden (Svalbard) ice cover and ice production observed by ers-2 sar. In: Jeffries MO, Eicken H (eds) Annals of Glaciology, International Glaciological Society, Cambridge, 33: 430436 Haave M, Ropstad E, Derocher AE, Lie E, Dahl E, Wiig Ø, Skaare JU, Jenssen BM. 2003. Polychlorinated biphenyls and reproductive hormones in female polar bears at Svalbard. Environmental Health Perspectives, 111: 431-436 Hallstadius L, Holm E, Persson B, Aarkrog A, Nilsson K. 1982. 137Cs in the Svalbard area. In: Society for Radiological Protection. Proceedings: Third international symposium, Inverness, Scotland, 1982: Radiological protection – advances in theory and practice. Reading: Society for Radiological Protection, 2: 500-505 Harington CR. 1965. The life and status of the polar bear. Oryx 8: 169-176 43
Harington CR. 1968. Denning habits of the polar bear (Ursus maritimus Phipps). Canadian Wildlife Service Report Series, 5: 1-30 Haug T, Nilssen KT, Øien N, Potelov V. 1994. Seasonal distribution of harp seals (Phoca groenlandica) in the Barents Sea. Polar Research, 13: 163-172 Hedberg GE, Derocher AE, Andersen M, Rogers QR, DePeters EJ, Lonnerdal B, Mazzaro L, Chesney RW, Hollis B. 2011. Milk composition in free-ranging polar bears (Ursus maritimus) from Svalbard, Norway as a model for captive rearing milk formula. Zoo Biology, 30: 550–565 Heggberget T, Bjørge A, Swenson JE, Syvertsen PO, Wiig Ø, Øien N. 2006. Pattedyr Mammalia. Pages 365-372 in J.A. Kålås, Å. Viken and T. Bakken, eds. Norsk Rødliste 2006 – 2006 Norwegian Red List. Artsdatabanken, Norway Heldal HE, Føyn L, Varskog P. 2003. Bioaccumulation of Cs-137 in pelagic food webs in the Norwegian and Barents Seas. Journal of Environmental Radioactivity, 65: 177-185 Hjelset AM, Andersen M, Gjertz I, Lydersen C, Gulliksen B. 1999. Feeding habits of bearded seals (Erignathus barbatus) from the Svalbard area, Norway. Polar Biology, 21: 186193 Høyland KV. 2009. Ice thickness, growth and salinity in Van Mijenfjorden, Svalbard, Norway. Polar Research, 28: 339-352 Hunter CM, Caswell H, Runge MC, Regehr EV, Amstrup SC, Stirling I. 2010. Climate change threatens polar bear populations: a stochastic demographic analysis. Ecology, 91: 28832897 IPCC. 2007. Climate change 2007: Synthesis Report. Intergovernmental Panel on Climate Change, Geneva, Switzerland Isaksen K, Wiig Ø. 1995. Seasonal distribution of harbour seals, bearded seals, white whales and polar bears in the Barents Sea. Norsk Polarinstitutt Meddelelser, 136: 47-59 IUCN/SSC. 1981. Proceedings of the 8th Working Meeting of the IUCN/SSC Polar Bear Specialist Group, January 1981, p 67 Iversen M. 2011. The diet of polar bears (Ursus maritimus) from Svalbard, Norway, inferred from scat analysis. Master Thesis, University of Tromsø, Norway.
44
Jenssen BM. 2006. Endocrine-disrupting chemicals and climate change: A worst-case combination for arctic marine mammals and birds? Environmental Health Perspectives, 114: 76-80 Johnson CJ, Heard DC, Parker KL. 2002. Expectations and realities of GPS animal location collars: results of three years in the field. Wildlife Biology, 8: 153-159 Jonkel CJ. 1970. Some comments on polar bear management. Biological Conservation, 2: 115119 Jonkel CJ, Kolenosky GB, Robertson R, Russell RH. 1972. Further notes on the polar denning habits. International Conference on Bear Research and Management, 2: 142-158 Kasamatsu F, Ishikawa Y. 1997. Natural variation of radionuclide 137Cs concentrations in marine organisms with special reference to the effect of food habits and trophic level. Marine Ecology Progress Series, 160: 109-120 Kershaw PJ, Baxter A. 1995. The transfer of reprocessing wastes from north-west Europe to the Arctic. Deep Sea Research II, 42: 1413-1448 Kingsley MCS, Stirling I, Calvert W. 1985. The distribution and abundance of seals in the Canadian High Arctic, 1980-82. Canadian Journal of Fisheries and Aquatic Science, 42: 1189-1210 Kovacs KM, Haug T and Lydersen C. 2009. Marine mammals in the Barents Sea. Pp. 453-496 In: Sakshaug E, Johnsen G and Kovacs K (eds.) Ecosystem Barents Sea. Tapir Academic Press, Trondheim, Norway. Kovacs KM, Lydersen C, Overland JE, Moore SE. 2011. Impacts of changing sea-ice conditions on Arctic marine mammals. Marine Biodiversity, 41: 181-194 Krafft BA, Kovacs KM, Lydersen C. 2007. Distribution of sex and age groups of ringed seals, Pusa hispida, in the fast-ice breeding habitat of Kongsfjorden, Svalbard. Marine Ecology Progress Series, 335: 199-206 Larsen T. 1967. The trapping and study of polar bears, Spitsbergen, 1966. Polar Record, 13: 589593 Larsen T. 1970. Polar bear Thalarctos maritimus research in Spitsbergen. Oryx 10: 368-372 Larsen T. 1972. Air and ship census of polar bears in Svalbard (Spitsbergen). Journal of Wildlife Management, 36: 562-570
45
Larsen T. 1985. Polar bear denning and cub production in Svalbard, Norway. Journal of Wildlife Management, 49: 320-326 Larsen T. 1986. Population biology of the polar bear (Ursus maritimus) in the Svalbard area. Norsk Polarinstitutt Skrifter, 184: 1-55 Larsen T, Jonkel C, Vibe C. 1983. Satellite radio-tracking of polar bears between Svalbard and Greenland. International Conference on Bear Research and Management, 5: 230-237 Larsen TS, Stirling I. 2009. The agreement on the conservation of polar bears – its history and future. Norsk Polarinstitutt Rapportserie nr. 127, 16pp Lentfer JW. 1969. Polar bear tagging in Alaska, 1968. Polar Record 14: 459-462,453 Lentfer JW. 1975. Polar bear denning on drifting sea ice. Journal of Mammalogy, 56: 716-718 Lentfer JW, Hensel RJ. 1980. Alaskan polar bear denning. In Bears - their biology and management, Edited by C.J. Martinka and K.L. McArthur. Washington DC, pp 101-108 Letcher RJ, Bustnes JO, Dietz R, Jenssen BM, Jorgensen EH, Sonne C, Verreault J, Vijayan MM, Gabrielsen GW. 2010. Exposure and effects assessment of persistent organohalogen contaminants in Arctic wildlife and fish. Science of the Total Environment, 408: 29953043 Letcher RJ, Klasson-Wehler E, Bergman Å, Paasivirta J. 2000. Methyl sulfone and hydroxylated metabolites of polychlorinated biphenyls, in: Passivirts, J. (Ed.), The Handbook of Environmental Chemistry. Springer-Verlag, Berlin Heidelberg, pp. 315-359 Lie E, Larsen HJS, Larsen S, Johnsen GM, Derocher AE, Lunn NJ, Norstrom RJ, Wiig Ø, Skaare JU. 2004. Does high organochlorine (OC) exposure impair the resistance to infection in polar bears (Ursus maritimus)? Part I: Effect of OCs on the humoral immunity. Journal of Toxicology and Environmental Health-Part A-Current Issues. 67: 555-582 Lie E, Larsen HJS, Larsen S, Johansen GM, Derocher AE, Lunn NJ, Norstrom RJ, Wiig Ø, Skaare JU. 2005. Does high organochlorine (OC) exposure impair the resistance to infection in polar bears (Ursus maritimus)? Part II: Possible effect of OCs on mitogenand
antigen-induced
lymphocyte
proliferation.
Journal
of
Toxicology
and
Environmental Health-Part A-Current Issues 68: 457-484 Lindenmayer DB, Likens GE. 2009. Adaptive monitoring: a new paradigm for long-term research and monitoring. Trends in Ecology & Evolution, 24: 482-486
46
Lønø O. 1970. The polar bear (Ursus maritimus Phipps) in the Svalbard area. Norsk Polarinstitutt Skrifter, 149: 1-115 Lowry LF, Burns JJ, Nelson RR. 1987. Polar bear, Ursus maritimus, predation on belugas, Delphinapterus leucas, in the Bering and Chukchi seas. Canadian Field-Naturalist, 101: 141-146 Lunn NJ, Schliebe S, Born EW. 2002. Polar Bears: Proceedings of the 13th Working Meeting of the IUCN/SSC Polar Bear Specialist Group, Nuuk, Greenland. IUCN, Gland Switzerland and Cambridge, UK. Vii + 153 pp Lydersen C, Wolkers H, Severinsen T, Kleivane L, Nordøy ES, Skaare JU. 2002. Blood is a poor substrate for monitoring pollution burdens in phocid seals. Science of the Total Environment, 292: 193-203 Marchesini, GR, Meimaridou, A, Haasnoot, W, Meulenberg, E, Albertus, F, Mizuguchi, M, Takeuchi, M, Irth, H, Murk, AJ. 2008. Biosensor discovery of thyroxine transport disrupting chemicals. Toxicology and Applied Pharmacology, 232: 150-160 Mauritzen M. 2002. Patterns and processes in female polar bear space-use. Ph.D. Thesis. University of Oslo, Norway Mauritzen M, Belikov SE, Boltunov AN, Derocher AE, Hansen E, Ims RA, Wiig Ø, Yoccoz N. 2003. Functional responses in polar bear habitat selection. Oikos, 100: 112-124 Mauritzen M, Derocher AE, Wiig Ø. 2001. Space-use strategies of female polar bears in a dynamic sea ice habitat. Canadian Journal of Zoology, 79: 1704-1713 Mauritzen M, Derocher AE, Wiig Ø, Belikov SE, Boltunov A, Hansen E, Garner GW. 2002. Using satellite telemetry to define spatial population structure in polar bears in the Norwegian and western Russian Arctic. Journal of Applied Ecology, 39: 79-90 McKinney MA, Peacock E, Letcher RJ. 2009. Sea ice-associated diet change increases the level of chlorinated and bromiated contaminants in polar bears. Environmental Science & Technology, 43: 4334-4339 Messier F, Taylor M, Ramsay M. 1992. Seasonal activity patterns of female polar bears (Ursus maritimus) in the Canadian Arctic revealed by satellite telemetry. Journal of Zoology, 226: 219-229 Molnár PK, Derocher AE, Klanjscek T, Lewis MA. 2011. Predicting climate change impacts on polar bear litter size. Nature Communications, 2: 186 47
Molnár PK, Derocher AE, Thiemann GW, Lewis MA. 2010. Predicting survival, reproduction and abundance of polar bears under climate change. Biological Conservation, 143: 1612-1622 Molnár PK, Klanjscek T, Derocher AE, Obbard ME, Lewis MA. 2009. A body composition model to estimate mammalian energy stores and metabolic rates from body mass and body length, with application to polar bears. Journal of Experimental Biology, 212: 2313-2323 Molnar PK, A. E. Derocher, Lewis MA, Taylor MK. 2008. Modelling the Mating System of Polar Bears: a Mechanistic Approach to the Allee Effect. Proceedings of the Royal Society B-Biological Sciences, 275: 217-226 Morales JM, Haydon DT, Frair J, Holsiner KE, Fryxell JM. 2004. Extracting more out of relocation data: building movement models as mixtures of random walks. Ecology, 85: 2436-2445 Muir DCG, Norstrom RJ, Simon M. 1988. Organochlorine contaminants in the Arctic marine food chains: accumulation of specific polychlorinated biphenyls and chlordane-related compounds. Environmental Science and Technology, 22: 1071-1079 Nilssen KT, Pedersen OP, Folkow LP, Haug T. 2000. Food consumption estimates of Barents Sea harp seals. NAMMCO Scientific Publications, 2: 9-27 Nordøy ES, Folkow LP, Potelov V, Prichtchemikhine V, Blix AS. 1998. Migratory patterns and dive behaviour of Barents Sea harp seals. The World Marine Mammal Science Conference, Monaco, 20-24 January 1998, pp 97 Norheim G, Skaare JU, Wiig Ø. 1992. Some heavy metals, essential elements, and chlorinated hydrocarbons in polar bear (Ursus maritimus) at Svalbard. Environmental Pollution, 77: 51-57 Norstrom RJ, Belikov SE, Born EW, Garner GW, Malone B, Olpinski S, Ramsay MA, Schliebe S, Stirling I, Stishov MS, Taylor MK, Wiig Ø. 1998. Chlorinated hydrocarbon contaminants in polar bears from eastern Russia, North America, Greenland, and Svalbard: Biomonitoring of Arctic pollution. Archives of Environmental Contamination and Toxicology, 35: 354-367 Obbard ME, Thiemann GW, Peacock E, DeBruyn TD. 2010. Polar bears: Proceedings of the 15th working meeting of the IUCN Polar Bear Specialist Group. Occasional Paper of the 48
IUCN Species Survival Commission No. 43. IUCN, Gland, Switzerland and Cambridge, UK Oehme M. 1991. Dispersion and transport paths of toxic persistent organochlorines to Arctic – levels and consequences. Science of the Total Environment, 106: 43-53 Olsen GH, Mauritzen M, Derocher AE, Sørmo EG, Skaare JU, Wiig Ø, Jenssen BM. 2003. Space-use strategy is an important determinant of PCB concentrations in female polar bears in the Barents Sea. Environmental Science & Technology, 37: 4919-4924 Olsson M, Andersson Ö, Bergman Å, Blomkvist G, Frank A, Rappe C. 1992. Contaminants and diseases in seals from Swedish Waters. Ambio, 21: 561-562 Oskam IC, Ropstad,E, Dahl E, Lie E, Derocher AE, Wiig Ø, Larsen S, Wiger R, Skaare JU. 2003. Organochlorines affect the major androgenic hormone, testosterone, in male polar bears (Ursus maritimus) at Svalbard. Journal of Toxicology and Environmental HealthPart A, 66: 2119-2139 Oskam IC, Ropstad E, Lie E, Derocher AE, Wiig Ø, Dahl E, Larsen S, Skaare JU. 2004. Organochlorines affect the steroid hormone cortisol in free-ranging polar bears (Ursus maritimus) at Svalbard, Norway. Journal of Toxicology and Environmental Health-Part A, 67: 959-977 Overrein Ø. 2002. Virkninger av motorferdsel på fauna og vegetasjon: Kunnskapsstatus med relevans for Svalbard (In Norwegian). Norsk Polarinstitutt Rapportserie nr. 119. 28pp Ovsyanikov N. 1995. Polar bear predation upon walrus on Wrangel Island. Bulletin MOIP Section Biology, 100: 3-15 Parks EK, Derocher AE, Lunn NJ. 2006. Seasonal and annual movement patterns of polar bears on the sea ice of Hudson Bay. Canadian Journal of Zoology, 84: 1281-1294 Parovshchikov VY. 1964. A study on the population of polar bear, Ursus (Thalarctos) maritimus Phipps, of Franz Joseph Land. Acta Societatis Zoologicae Bohemoslovenicae, 28: 167177 Pentreath RJ, Jefferies DF, Talbot JW, Lovett MB, Harvey BR. 1982. Transuranic cycling behaviour in marine environment. IAEA-TECDOC-265. Vienna: International Atomic Energy Agency, 1982, p 121-128.
49
Pilfold NW, Derocher AE, Stirling I, Richardson E, Andriashek D. 2012. Age and sex composition of seals killed by polar bears in the Eastern Beaufort Sea. PLoS ONE 7(7): e41429. doi:10.1371/journal.pone.0041429 Polischuk SC, Norstrom RJ, Ramsay MA. 2002. Body burdens and tissue concentrations of organochlorines in polar bears (Ursus maritimus) vary during seasonal fasts. Environmental Pollution, 118: 29-39 Polischuk SC. 1999. Organochlorine dynamics in free-ranging polar bears and their cubs. PhD Thesis, University of Saskatchewan, Saskatoon, Canada, pp 188 Prestrud P, Stirling I. 1994. The international Polar Bear Agreement and the current status of polar bear conservation. Aquatic Mammals, 20: 1-12 Ramsay M, Stirling I. 1990. Fidelity of female polar bears to winter den sites. Journal of Mammalogy, 71: 233-236 Reeves RR. 1998. Distribution, abundance and biology of ringed seals (Phoca hispida): an overview. NAMMCO Scientific Publications, 1: 9-45 Regehr EV, Hunter CM, Caswell H, Amstrup SC, Stirling I. 2010. Survival and breeding of polar bears in the Southern Beaufort Sea in relation to sea ice. Journal of Animal Ecology, 79: 117-127 Regehr EV, Lunn NJ, Amstrup SC, Stirling L. 2007. Effects of earlier sea ice breakup on survival and population size of polar bears in Western Hudson Bay. Journal of Wildlife Management, 71: 2673-2683 Richardson E, Stirling I, Hik DS. 2005. Polar bear (Ursus maritimus) maternity denning habitat in western Hudson Bay: a bottom-up approach to resource selection functions. Canadian Journal of Zoology, 83: 860-870 Rode KD, Amstrup SC, Regehr EV. 2010. Reduced body size and cub recruitment in polar bears associated with sea ice decline. Ecological Applications, 20: 768-782 Rondinini C, Rodrigues ASL, Boitani L. 2011. The key elements of a comprehensive global mammal conservation strategy Introduction. Philosophical Transactions of the Royal Society B-Biological Sciences, 355: 2591-2597 Sahanatien V, Derocher AE. 2012. Monitoring sea ice habitat fragmentation for polar bear conservation. Animal Conservation, 15: 397-406
50
Sander G, Hanssen-Bauer I, Bjørge A, Prestrud P. 2005. Miljøovervåking av Svalbard og Jan Mayen – MOSJ. Rapportserie nr. 123, Norsk Polarinstitutt, Tromsø, Norge, 70 pp. Schliebe S, Wiig Ø, Derocher AE, Lunn N. 2006. Ursus maritimus. In: IUCN 2006. 2006 IUCN Red List of Threatened Species. . Shapiro I, Colony R, Vinje T. 2003. April sea ice extent in the Barents Sea, 1850-2001. Polar Research, 22: 5-10 Skaare JU, Bernhoft A, Derocher A, Gabrielsen GW, Goksøyr A, Henriksen E, Larsen HJ, Lie E, Wiig Ø. 2000. Organochlorines in top predators at Svalbard – occurrence, levels and effects. Toxicological Letters, 112-113: 103-109 Smith TG. 1980. Polar bear predation of ringed and bearded seals in the land-fast sea ice habitat. Canadian Journal of Zoology, 58: 2201-2209 Smith TG, Lydersen C. 1991. Availability of suitable land-fast ice and predation as factors limiting ringed seal populations, Phoca hispida, in Svalbard. Polar Research, 10: 585594 Smith TG, Sjare B. 1990. Predation of belugas and narwhals by polar bears in nearshore areas of the Canadian High Arctic. Arctic, 43: 99-102 Smith TG, Stirling I. 1975. The breeding habitat of the ringed seal (Phoca hispida). The birth lair and associated structures. Canadian Journal of Zoology, 53: 1297-1305 Sonne C. 2010. Health effects from long-range transported contaminants in Arctic top predators: An integrated review based on studies of polar bears and relevant model species. Environmental International, 36: 461-491 Stirling I. 1974. Midsummer observations on behavior of wild polar bears (Ursus maritimus). Canadian Journal of Zoology, 52: 1191-1198 Stirling I, Andriashek D. 1992. Terrestrial maternity denning of polar bears in the eastern Beaufort Sea area. Arctic, 45: 363-366 Stirling I, Andriashek D, Calvert W. 1993. Habitat preferences of polar bears in the Western Canadian Arctic in late winter and spring. Polar Record, 29: 13-24 Stirling I, Archibald WR (1977) Aspects of predation of seals by polar bears. Journal of the Fisheries Research Board of Canada, 34:1126-1129 Stirling I, Derocher AE. 1993. Possible impacts of climatic warming on polar bears. Arctic, 46: 240-245 51
Stirling I, Lunn NJ, Iacozza J. 1999. Long-term trends in the population ecology of polar bears in Western Hudson Bay in relation to climatic change. Arctic, 52: 294-306 Stirling I, McEwan EH. 1975. Caloric value of whole ringed seals (Phoca hispida) in relation to polar bear (Ursus maritimus) ecology and hunting behavior. Canadian Journal of Zoology, 53: 1021-1027 Stirling I, Øritsland NA. 1995. Relationships between estimates of ringed seal (Phoca hispida) and polar bear (Ursus maritimus) populations in the Canadian Arctic. Canadian Journal of Fisheries and Aquatic Sciences, 52: 2594-2612 Stirling I, Derocher AE. 2012. Effects of climate warming on polar bears: a review of the evidence. Global Change Biology, 18: 2694–2706 Strand P, Howard BJ, Aarkrog A, Balonov M, Tsaturov Y, Bewers JM, Salo A, Sickel M, Bergman R, Rissanen K. 2002. Radioactive contamination in the Arctic - sources, dose assessment and potential risks. Journal of Environmental Radioactivity, 60: 5-21 Taylor MK, Laake J, McLoughlin PD, Born EW, Cluff HD, Ferguson SH, Rosing-Asvid A, Schweinsburg R, Messier F. 2005. Demography and viability of a hunted population of polar bears. Arctic, 58: 203-214 Taylor MK, Lee LJ. 1995. Distribution and abundance of Canadian polar bear populations: A management perspective. Arctic, 48: 147-154 Taylor MK, Demaster DP, Bunnell FL, Schweinsburg RE. 1987. Modelling the sustainable harvest of female polar bears. Journal of Wildlife Management, 51: 811-820 Thiemann GW, Iverson SJ, Stirling I. 2008. Polar bear diets and arctic marine food webs: Insights from fatty acid analysis. Ecological Monographs, 78: 591-613 UNEP/AMAP. 2011. Climate Change and POPs: Predicting the Impacts. Report of the UNEP/AMAP Expert Group. Secretariat of the Stockholm Convention, Geneva. p. 62. Sonne C, Dietz R. 2011. Exposure to mixtures of organohalogen contaminants and associative interactions with thyroid hormones in East Greenland polar bears (Ursus maritimus). Environmental International, 37: 694-708 Vinebrooke RD, Cottingham KL, Norberg J, Scheffer M, Dodson SI, Maberly SC, Sommer U. 2004. Impacts of multiple stressors on biodiversity and ecosystem functioning: the role of species co-tolerance. Oikos, 104: 451-457 Vongraven D. 2009. Guest editorial: the ballyhoo over polar bears. Polar Research, 28: 323-326 52
Vongraven D, Peacock E. 2011. Development of a pan-Arctic monitoring plan for polar bears: background
paper.
Circumpolar
Biodiversity
Monitoring
Programme,
CAFF
Monitoring Series Report No. 1. CAFF International Secretariat, Akureyri, Iceland Vongraven D, Aars J, Amstrup S, Atkinson SN, Belikov SE, Born EW, DeBruyn TD, Derocher AE, Durner G, Gill M, Lunn N, Obbard ME, Omelak J, Ovsyanikov N, Peacock E, Richardson E, Sahanatien V, Stirling I, Wiig Ø. 2012. A circumpolar monitoring framework for polar bears. Ursus Monograph Series Number 5, 2012. Watson WS, Sumner DJ, Baker JR, Kennedy S, Reid R, Robinson I. 1999. Radionuclides in seals and porpoises in the coastal waters around the UK. Science of the Total Environment, 234: 1-13 Watts PD, Hansen SE. 1987. Cyclic starvation as a reproductive strategy in the polar bear. Symposia of the Zoological Society of London, 57: 305-318 Wiig Ø. 1998. Survival and reproductive rates for polar bears at Svalbard. Ursus, 10: 25-32 Wiig Ø, Aars J, Born EW. 2008. Effects of climate change on polar bears. Science Progress, 91: 151-173 Wiig Ø, Born EW, Pedersen LT. 2003. Movements of female polar bears (Ursus maritimus) in the East Greenland pack ice. Polar Biology, 26: 509-516 Wiig Ø, Derocher AE, Belikov SE. 1999. Ringed seal (Phoca hispida) breeding in the drifting pack ice of the Barents Sea. Marine Mammal Science, 15: 595-598 Wiig Ø. 1995. Distribution of polar bears (Ursus maritimus) in the Svalbard area. Journal of Zoology, 237: 515-529 Wiig Ø. 2005. Are polar bears threatened? Science, 309:1814-1815 Wiig Ø, Derocher A. 1999. Application of aerial survey methods to polar bears in the Barents Sea. Pages 27-36 in Garner et al., eds. Marine Mammal Survey and Assessment Methods. Balkema, Rotterdam. Wilson KA, Evans MC, Di Marco M, Green DC, Boitani L, Possingham H P, Chiozza F, Rondinini C. 2011. Prioritizing conservation investments for mammal species globally. Philosophical Transactions of the Royal Society B-Biological Sciences, 366: 2670-2680 Wolkers H, Van Bavel B, Derocher AE, Wiig Ø, Kovacs KM, Lydersen L, Lindstrøm G. 2004. Congener-specific accumulation and food chain transfer of polybrominated diphenyl
53
ethers in two Arctic food chains. Environmental Science and Technology, 38: 16671674 Wolkers H, Krafft BA, van Bavel B, Helgason LB, Lydersen C, Kovacs K M. 2008. Biomarker responses and decreasing contaminant levels in ringed seals (Pusa hispida) from Svalbard, Norway. Journal of Toxicology and Environmental Health, 71: 1009-1018 Yasuda T, Arai N. 2005. Fine-scale tracking of marine turtles using GPS-Argos PTTs. Zoological Science, 22: 547-553 Zeyl E, Ehrich D, Aars J, Bachmann L, Wiig Ø. 2010. Denning-area fidelity and mitochondrial DNA diversity of female polar bears (Ursus maritimus) in the Barents Sea. Canadian Journal of Zoology, 88: 1139-1148 Øritsland NA. 1970. Temperature regulation of the polar bear (Thalarctos maritimus). Journal of Comparable Biochemistry and Physiology Series A, 37: 225-233
54
Paper 1
Photo: Magnus Andersen
MARINE MAMMAL SCIENCE, 25(1): 35–52 ( January 2009) C 2008 by the Society for Marine Mammalogy DOI: 10.1111/j.1748-7692.2008.00228.x
Estimating the Barents Sea polar bear subpopulation size J. AARS Norwegian Polar Institute, N-9296 Tromsø, Norway E-mail:
[email protected]
T. A. MARQUES S. T. BUCKLAND Centre for Research into Ecological and Environmental Modelling, University of St Andrews, St Andrews KY16 9LZ, Scotland
M. ANDERSEN Norwegian Polar Institute, N-9296 Tromsø, Norway
S. BELIKOV A. BOLTUNOV All-Russian Research Institute for Nature Protection, 113628 Moscow, Russia
Ø. WIIG Natural History Museum, University of Oslo, N-0318 Oslo, Norway
ABSTRACT A large-scale survey was conducted in August 2004 to estimate the size of the Barents Sea polar bear subpopulation. We combined helicopter line transect distance sampling (DS) surveys in most of the survey area with total counts in small areas not suitable for DS. Due to weather constraints we failed to survey some of the areas originally planned to be covered by DS. For those, abundance was estimated using a ratio estimator, in which the auxiliary variable was the number of satellite telemetry fixes (in previous years). We estimated that the Barents Sea subpopulation had approximately 2,650 (95% CI approximately 1,900–3,600) bears. Given current intense interest in polar bear management due to the potentially disastrous effects of climate change, it is surprising that many subpopulation sizes are still unknown. We show here that line transect sampling is a promising method for addressing the need for abundance estimates. Key words: helicopter surveys, distance sampling, line transects, polar bear, Barents Sea, subpopulation size.
35
36
MARINE MAMMAL SCIENCE, VOL. 25, NO. 1, 2009
Polar bears were heavily hunted during the first half of the 20th century all over the Arctic. In the early 1960s scientists and managers from the five polar bear nations (Canada, Denmark (Greenland), Norway, Soviet Union, and U.S.A.) started to work on an international plan for conservation of polar bears. The outcome of this work was the international “Agreement on the Conservation of Polar Bears”, signed in 1973 (see Prestrud and Stirling 1994). One central statement in the agreement is that “Each Contracting Party. . . shall manage polar bear populations in accordance with sound conservation practices based on the best available scientific data.” Monitoring change in population size is the most direct method for monitoring population status. The importance of using scientifically derived population estimates in the management of polar bears has recently been stressed (Wiig 2005, Aars et al. 2006). The Barents Sea polar bear subpopulation, shared between Norway and Russia, has not been harvested since 1956 in Russia and since 1973 in Norway (Prestrud and Stirling 1994). Persistent pollutants (Bernhoft et al. 1997, Norstrom et al. 1998, Andersen et al. 2001), climate change (Derocher 2005), and oil development (Downing and Reed 1996, Isaksen et al. 1998) are possible threats to the subpopulation. No meaningful estimate of the size of the subpopulation based on any study that covers the whole area in question has been available until now (Wiig and Derocher 1999). In the early 1980s, Larsen (1986) suggested that the Barents Sea polar bear subpopulation size was between 3,000 and 6,700 (dependent on subpopulation definition). This was based on data from multiple sources including den counts and spatially restricted, non-random air surveys, and extrapolation to larger areas. With recent statistical developments, distance sampling (DS) is one of the most widely used methods for estimating animal abundance (Buckland et al. 2001), and is today considered to be more cost efficient than capture–recapture to achieve a given precision (Borchers et al. 2002), in particular for populations occurring at low densities over large areas. Wiig and Derocher (1999) advocated a line transect study for the Barents Sea area to estimate the subpopulation size, due to the large area and the lack of logistical bases needed for capture–recapture-based estimates. Pilot studies on polar bears in Norway and the United States have examined the applicability of aerial surveys over sea ice, employing both strip transects and line transects. However, density estimates have not been extended to subpopulation estimates due to small sample sizes, restricted coverage, and methodological uncertainties (Wiig and Bakken 1990, Wiig 1995, Manly et al. 1996, Wiig and Derocher 1999, Evans et al. 2003). Wiig and Derocher (1999) reviewed aerial survey attempts in the Barents Sea, comparing them with mark–recapture techniques, and concluded that line transect surveys would be the most economical and effective means to estimate and monitor the polar bear subpopulation size. In the present paper we report on a large-scale line transect survey of polar bears that was conducted in the Barents Sea in August 2004.
MATERIALS AND METHODS Study Area The aerial survey was conducted between 26 July and 1 September 2004, in the Barents Sea area (Fig. 1). The Barents Sea polar bear subpopulation has earlier been defined as the animals occupying the area between longitudes of 10◦ E and
AARS ET AL.: ESTIMATING BARENTS SEA POLAR BEARS
37
60◦ E, and latitudes of 72◦ N and 83◦ N including the Svalbard (Norway) and Franz Josef Land (FJL, Russia) archipelagos (Wiig and Derocher 1999), but we chose to include both the whole of FJL, and areas extending slightly west of 10◦ E and east of 60◦ E allowing parallel survey lines extending north from the ice edge (Fig. 1). The latitudinal boundaries vary seasonally with the sea ice extent, and in August bears are not found south of about 76◦ 30 . In 2004 there was less sea ice around the islands than in most recent years, and the sea ice edge was in most places farther north with a very distinct edge, and very stationary during the period of survey. Movements of bears are lowest at the time of year the survey was conducted (Mauritzen et al. 2003), and they are not expected to be directional as in spring or late autumn when bears move between hunting and denning areas, or earlier in summer when sea ice distribution changes considerably.
Survey Design Before the survey was conducted, we used capture and satellite tracking data, as well as historical knowledge, to plan the survey design. Bears could be encountered either on or around islands (land, glaciers, or on land-fast sea ice) or in the pack ice in the Arctic Basin (hereafter termed pack ice or PI). The Svalbard areas were considered to be either: a 0 areas where bears were unlikely to be encountered in late summer, or a p areas where there could be polar bears present. Areas in category a 0 were excluded from the survey (bear density assumed 0). For areas where polar bears could be present, a further sub-division was made: a pc small areas not well suited for line transect sampling, with expected high densities of polar bears and a pd areas well suited for line transect sampling, mostly with presumed intermediate densities. Areas a pc were either small islands or relatively narrow strips of land between fjords and steep cliffs, and were free of snow. It was therefore possible to conduct complete counts (with variance assumed to be zero) of these areas. Areas a pd , together with areas of sea ice and all the islands of FJL, were surveyed using line transect sampling. The DS survey design for areas a pd was created using the survey design engine in Distance (Thomas et al. 2006). Line spacing was 3 km for all islands and sea ice areas. This spacing was chosen according to flight hours and survey time available. In different subareas line orientation was determined to minimize transects along shore areas, placing lines perpendicular to any potential density gradients. The line transect coordinates were imported into Arcview (www.ESRI.com), and from there uploaded into a GPS unit that was used in the helicopter during the survey. When transects were flown over land and glaciers, they were continued over the sea wherever there was ice (fast ice or drift ice) present. Around Svalbard, there was virtually no sea ice left. Only 41 km of transect line was flown to cover what was left, and no bears were observed. Thus we ignored these transects in the analyses (bear density assumed 0). Also some very open drift ice (Fig. 1) was not covered due to safety considerations. While most of the land areas were surveyed according to the planned 3-km line spacing pattern, the line spacing of glaciers on FJL was often changed. Because fog often prevented flying across glaciers, every second or third line were flown (Fig. 1), resulting in 6- or 9-km line spacing, respectively. In Svalbard, the coverage on glaciers was even more sporadic than on FJL (Fig. 1), because of almost continuous dense fog over the glaciers during the survey period.
38
MARINE MAMMAL SCIENCE, VOL. 25, NO. 1, 2009
Figure 1. Study area. Line transects are marked with solid lines. The seven strata used in distance sampling analyses were: (1) Glacier and (2) Land in Norway (area G), (3) Glacier, (4) Land and (5) Sea ice in Russia (area H), and (6) Pack Ice in Norway (area A) and in (7) Russia (areas B + C). Dotted lines show planned survey lines not covered or extensions of these lines north to 336 km from the ice edge, in areas were telemetry fixes from collared bears were used to estimate bear densities (areas D + E + F).
For the PI area, 89 transect lines 9 km apart were laid roughly northward from the ice edge, with the most western and eastern lines running from 9◦ 39.4 E, 81◦ 26.0 N to 6◦ 13.2 E, 82◦ 30.3 N and 60◦ 28.4 E, 81◦ 54.5 N to 63◦ 48.6 E, 82◦ 41.0 N, respectively (see Fig. 1). These lines extended 185 km north from the ice edge. Due to time and fog-induced safety constraints, 17 lines were not covered, and most surveyed lines were shorter than 185 km (Fig. 1). For analysis we considered seven geographic strata, based on the interaction of administrative regions (Russia and Norway) and habitats (Pack Ice, Land, Glacier, and Sea Ice, the latter around islands, only in Russia) (Fig. 1).
AARS ET AL.: ESTIMATING BARENTS SEA POLAR BEARS
39
Field Methods Two single-engine helicopters (Eurocopter AS350 Ecureuil) were used for flying total counts and line transects. One helicopter was stationed at Longyearbyen, Svalbard, covering the Svalbard archipelago from 26 to 31 July, and 14 to 26 August. The second helicopter was stationed onboard the RV Lance (from 1 to 31 August) operating along the ice edge in the Barents Sea and in the FJL archipelago area. The helicopters operated with four observers including the pilot. The two observers in the front seats focused on the transect line in front of the helicopter, while the two rear observers focused to each side of the helicopter farther away from the transect line, but with some overlap with the front observers’ search areas. The helicopters generally flew at 200 ft (61 m) above ground and at 100 knots (185 km/h). Due to weather conditions, other combinations of altitude and speed were used for one transect at the PI, and at a few transects above glaciers at FJL. The speed and altitude for flying were chosen after a pilot study (Aars et al., unpublished data) to ensure that the probability of observing a bear on the transect line, g(0), was maximized. We tested different altitudes and varying speed and found that a speed of 185 km/h and altitude of 60 m was optimal. The predetermined transect lines were flown using the GPS unit to which the survey design had been previously uploaded. This GPS unit also recorded a detailed track of the actual flight path. Details can be found in Marques et al. (2006). The helicopter that operated from the research vessel in the Russian area and along the PI had two complete survey teams including two pilots, to allow transects to be flown almost continuously in periods of good weather. When single transect lines covered different habitat types (land, glacier, fast ice, drift ice), each stretch of one habitat type was categorized as a section. For analysis, all sections of the same habitat type along a single line were then pooled (e.g., a flown line with 10 km ice + 10 km land + 10 km ice was considered as two transects, one with 20 km ice and one with 10 km land). For each bear observation, habitat structure was recorded as a covariate and scored as: 1 = relatively flat surface with only minor or no structure of the size that could make it difficult to spot a polar bear; 2 = habitat with some structure that could make it harder to spot a bear in the area (e.g., some screwed sea ice or crevasses on glaciers); 3 = major structures present that make it considerably more difficult to detect polar bears (e.g., heavily screwed sea ice). Data Analysis We used DS as described by Buckland et al. (2001, 2004) to estimate bear abundance. For populations that potentially occur in well-defined clusters, like polar bears, the perpendicular sighting distances to the center of detected clusters allow the modeling of a detection function, g(y), which represents the probability of detecting a cluster, given that it is at distance y from the transect line (Buckland et al. 2001). Parametric models are assumed for this detection function, and their parameters estimated via maximum likelihood, with the possibility to include adjustment terms for fit improvement. The probability of detecting a cluster in the covered areas, P, is the mean value of the detection function with respect to the available distances. While conventional methods use only the perpendicular distance y for detection function modeling, multiple covariate distance sampling (MCDS) extends
40
MARINE MAMMAL SCIENCE, VOL. 25, NO. 1, 2009
the methods to include additional covariates z to help model the detection function. The scale parameter of the detection function becomes a function of these additional covariates (Marques and Buckland 2003). If MCDS is used, the probability of detection associated with cluster i, Pa(z i ) (i = 1, 2 , . . . , n, where n is the total number of detected clusters) is conditional on the corresponding covariate values z i . An estimate of cluster abundance in the covered area (corresponding to the covered strips of width 2w, i.e., extending a distance w either side of the lines) of size a is then obtained by (e.g., Marques and Buckland 2003) ˆ cs = N
n
1
i=1
Pˆ a (z i )
.
To estimate animal abundance, this estimate can be multiplied by an estimate of mean cluster size, using the regression method (Buckland et al. 2001:73–75) to correct for size bias. Given an appropriate random sample of lines (see Strindberg et al. 2004 for design details), one can extrapolate the results obtained in the covered area, based on the design properties, to the wider survey region. In our case, purely design-based extrapolation was not possible for the entire PI because it was not possible to survey all the transect lines, so that an additional procedure was required (see below). We define the surveyed area as the area within which a systematic sample of surveyed transects was conducted, and for which we can estimate abundance using line transect sampling alone. Other areas within the study region, to the north of covered strips or within the gap in survey coverage (areas D, E, and F, Fig. 1) are referred to as the unsurveyed area. The conventional methods used in our survey rely on a set of assumptions (Buckland et al. 2001), the most important being: (1) a large number of transects are randomly allocated in the study area independently of the distribution of the population of interest; (2) all animals on the line are detected with certainty (g(0) = 1); (3) animal movement is slow with respect to observer movement; (4) distances are measured without error. An evaluation of the extent to which each of these was fulfilled is provided in the Discussion. Preliminaries Both the tracks and the waypoint files from the transects flown were downloaded to a computer and analyzed using the program GPS Map Explorer version 2.34 (http://home.tiscali.no/gpsii/). Perpendicular distances to bear clusters required for DS analysis were obtained by calculating the shortest distance from the line flown, as recorded by GPS, to the waypoint representing the location of the bear (or bear cluster) (Marques et al. 2006). Extensive plotting of detected distances, number of detections, encounter rates, and cluster sizes was done during the survey, allowing the removal of virtually all typos and recording errors. Analysis in Distance 5.0 The line transect data were analyzed using program Distance 5.0. (Thomas et al. 2006). A preliminary analysis of all the data with half-normal and hazard-rate
AARS ET AL.: ESTIMATING BARENTS SEA POLAR BEARS
41
detection functions showed a well-behaved detection function decreasing with distance and no evidence of any major problems. Preliminary analyses showed that truncation of 5% of the largest distances (w = 1,068 m) was adequate to preclude fitting spurious bumps in the tail of the detection function, and that was the truncation used in subsequent analysis. A number of candidate covariates were available for modeling the detection function: habitat, structure (around the position the bear was observed), cluster size, and helicopter altitude. Habitat was a factor covariate with three levels, structure was considered both as a factor with three levels and as a continuous covariate (with values 1, 2, or 3), and the remaining two were continuous covariates. Their inclusion was assessed by using minimum Akaike’s Information Criterion (AIC), and the overall fit was evaluated using the standard chi-square, the Kolmogorov–Smirnov, and the Cram´er-von Mises tests available in Distance. A priori, it was thought that altitude could reflect overall flying conditions (because under ideal conditions the altitude would be 61 m). The perception based on the fieldwork was that cluster size would not be relevant, because clusters tend to act as single cues. Both habitat and structure were believed to be potentially useful for modeling the detection function. As well as testing the inclusion of cluster size as a covariate, we used size bias regression methods, namely by regressing the log of the detection function on the log of cluster size (e.g., Buckland et al. 2001:73–75), to evaluate and correct for the eventual presence of size bias. For abundance estimation, we considered a common detection function model across strata, with density estimated by stratum conditional on the observed covariate values.
Correction for Areas not Surveyed Despite the carefully planned survey design, it proved impossible to survey 17 lines in the Russian region, leaving a gap in the coverage of the PI region between Svalbard and FJL (see area F, Fig. 1). This was due to bad weather and related safety considerations arising from flying a one-motor helicopter in a remote area. Additionally, the average length of the lines flown in this area was 125 km (range 43–232 km, Fig. 1) instead of the 185 km planned. Rather than assuming that density in the surveyed areas could be extrapolated to these areas, we opted for a ratio estimator. Considering boxes as described below, we used the auxiliary variable number of telemetry fixes (from adult females, see below) within each box, to predict the number of bears that would have been detected had the lines been flown. The 89 lines from the study design, including the 17 that were not covered by line transects, were considered to be the centerlines of rectangles or “boxes,” each 9 km wide (4.5 km either side of the line) and 336 km long (north from the ice edge). Each box was divided into two, the southern portion corresponding to the surveyed centerline, and the northern portion extending the box from where survey effort stopped to the boundary 336 km north of the ice edge (the maximum distance recorded by telemetry fixes). The 17 lines not covered by the survey in the Russian sector were divided into two; a southern part was 92 km long (the average length of the Russian transects flown) and a northern part was 244 km long (= 336–92 km). Thus, the size of the study region north of the ice edge was 269,136 km2 (89 × 9 km × 336 km), of which 80,839 km2 (30%) was surveyed by line transect sampling, with abundance for the remaining 188,297 km2 (70%) being estimated using the ratio estimator, as described below.
42
MARINE MAMMAL SCIENCE, VOL. 25, NO. 1, 2009
The telemetry data are well described in Mauritzen et al. (2001, 2002). Satellite telemetry fixes have high position accuracy compared to the scale on which polar bears operate. The telemetry accuracy was estimated to be less than 1,400 m for 92% of the positions (Mauritzen et al. 2001). We used only positions from the period when the ice edge is furthest north (July–September) from 44 different bears tracked between 1989 and 1999. Only positions north of 79◦ N and at least 20 km from land were used (169 data points). If more than one position was available for a given bear, only positions a minimum of 6 d apart were used to reduce dependence between bear locations. The average distance between 93 pairs of fixes 6 d apart in time was 57 km (SD = 28 km), and thus relatively large compared to the distance between lines (9 km). Using standard ANOVA (P > 0.1), there was no significant yearly difference in the distribution of bears relative to the distance from the ice edge, despite the latitudinal position of the ice edge varying substantially among years. There was also no significant difference in the distance from the ice edge when comparing data for July plus September with data for August. Each point was allocated to the appropriate box. As a box was defined in terms of distance from the ice edge, a fix of say 20 km north of the ice edge in 1989 would be placed 20 km north of the ice edge in our survey, even though the ice edge was in a different absolute position. The number of bears that would have been detected if transects had been flown in the areas not surveyed, nˆ u n s , was estimated using the following ratio estimator: nˆ u n s = r
178
Xi ,
i=73
where
72
r =
i=1 72
Yi Xi
i=1
and Xi is the number of telemetry fixes within box i and Yi the number of bear groups detected within box i during the survey. Note i = 1, 2 , . . . , 72 (Fig. 1, area A + B + C) represents boxes associated with transect lines surveyed while i = 73, 74 , . . . , 178 (Fig. 1, area D + E + F) represents boxes corresponding to transect line segments not surveyed. ˆ u n s ) is then calculated The estimated number of animals for the unsurveyed areas ( N by ˆ ˆ u n s = E (s )nˆ u n s , N Pˆ d Pc |s where the relevant quantities were obtained in the DS component of the survey, i.e., Eˆ (s ) is an estimate of mean cluster size, obtained using the regression method (Buckland 2001:73–75) to allow for size bias, Pˆ d is the estimated probability of detection of animals that are within the truncation distance w (this is obtained as the average estimated probability of detection within each stratum, conditional on the observed covariates in that stratum) and Pc |s is the probability that a bear is within a distance w from the survey lines, given that it is within the surveyed areas (= 2w/box width).
AARS ET AL.: ESTIMATING BARENTS SEA POLAR BEARS
43
Total Estimate and Variance Estimation The total estimated number of bears was obtained as the sum of: (1) the estimated numbers in the surveyed area, (2) the estimated numbers in the unsurveyed area, and (3) the numbers obtained by total counts. Although the final population estimate is made up of three components (total count, surveyed areas, and unsurveyed areas), the first of these does not contribute to the variance, as by definition a total count has no variance. To obtain estimates of variance for the estimated abundances we used a nonparametric bootstrap (999 bootstrap resamples). Variance estimates were obtained by resampling lines within strata, as described in Buckland et al. (2001). In the case of the PI strata, the 336 km long transects were resampled, along with all the information contained in them (survey data and telemetry fixes). To incorporate the variance involved in estimation for both surveyed and unsurveyed areas, the bootstrap procedure had to be implemented outside Distance, using software R (version 2.0.1, R Development Core Team 2004), with Distance called from R; in this way, for each bootstrap resample, an estimate of all random quantities involved both in the DS component and the ratio estimator component were obtained. As both detections and fixes are resampled together within the bootstrap, no assumption of independence is made between the estimate for the unsurveyed and surveyed areas when calculating the variance of the whole subpopulation. Confidence intervals were obtained by the percentile method (Buckland et al. 2001). By bootstrapping all the information within the boxes (around each line), we ensure comparability across the two components of estimation. As shown by Davison and Hinkley (1997), it is theoretically superior not to bootstrap at lower levels when data are hierarchical, hence we do not bootstrap fixes within boxes in addition to boxes. In this way, we also avoid having to assume that fixes are independent, replacing this by the weaker assumption that data for different boxes are independent. RESULTS Total Counts A total of 31 polar bears were observed during total counts on the island of Spitsbergen. An additional 27 bears were observed on different small- and mediumsized islands in Svalbard. On Kvitøya, in the northeast of the Svalbard area, 32 bears were observed. In addition to these observations from areas that we allocated to a total count survey, we added six bears that were observed from RV Lance or from the helicopter in the Russian area, in areas of loose drift ice not surveyed by line transects. Thus total counts add up to 96 bears. Line Transects We flew 20,975 km of transect lines distributed in seven strata (Table 1). These were considered to be 1,018 independent transect units for variance estimation. A total of 189 polar bear clusters were detected. These clusters consisted of 139 lone bears and 50 adult females with 1–3 (average 1.48) juveniles. The mean cluster size was 1.39 and the total number of bears observed was 263. In Table 2 we present summary statistics for the candidate models considered for the detection function, with combinations of the covariates available (models
44
MARINE MAMMAL SCIENCE, VOL. 25, NO. 1, 2009
Table 1. Estimates of polar bear density and abundance within areas of seven different strata covered by line transect surveys, based on analyses in the program Distance. Nor = Norwegian area, Rus = Russian area, G = Glacier, L = Land, SI = Sea Ice around FJL, PI = Pack Ice, obs = number of distance observations used to fit the curves after 5% truncation of ˆ = estimated number of bears in ˆ = estimated bear density (per 100 km2 ), N observations, D surveyed areas. Strata
Km lines
Obs
ˆ D
ˆ 95% CI D
ˆ N
ˆ 95% CI N
Km2
Nor G Nor L Nor PI Rus G Rus L Rus PI Rus SI Total
1,055 2,771 6,903 1,757 888 4,706 2,895 20,975
3 25 16 16 38 39 43 180
0.5 1.0 0.4 1.6 5.7 2.1 2.3 1.1
(0.0; 1.2) (0.6; 1.5) (0.2; 0.6) (0.8; 3.1) (3.3; 8.8) (1.3; 3.0) (1.4; 3.4) (0.8; 1.4)
67 84 212 226 137 483 185 1,394
(0, 160) (49, 121) (106, 318) (106, 424) (79, 212) (305, 678) (113, 271) (1,060, 1,743)
13,271 8,082 57,877 13,714 2,421 22,962 8,009 126,336
shown only if AIC < 2). This table provides further reassurance on the quality of the data. Despite using different key functions (half-normal and hazard rate) with several combinations of covariates, the global density estimates, and corresponding associated measures of precision, were remarkably close. Additionally, none of the three absolute goodness-of-fit measures considered indicated that any of the top models gave a poor fit. AIC did not indicate that cluster size affected detectability (AIC of 5.35 and 8.44 for the half-normal and hazard rate models, respectively). This was in accordance with field perception. We were rarely able to distinguish between a family group Table 2. Details of the MCDS models considered for the detection function, with combinations of the available covariates. Available covariates were habitat (Hab), structure as factor (Str) or as a continuous covariate (nfStr), cluster size (CS), and altitude (Alt) used to describe the model. HN corresponds to half-normal and HR to hazard rate. Columns are Akaike Information Criterion (AIC) and the difference between lowest AIC and model AIC (AIC), number of parameters in the model (par), global density estimate pooled across strata (D), the corresponding 95% confidence interval (95% CI), and coefficient of variation (CV), as well as three goodness of fit (GOF) measures, a chi-square test (ChS), a Kolmogorov-Smirnov (KS), and a Cram´er-von Mises test (CvM).
Name HN + nfStr HN + nfStr + Hab HN + Str HR + nfStr HN + Str + Hab HR + nfStr + Hab
Par AIC
AIC
D
95% CI
D CV
GOF ChS
GOF GOF KS CvM
2 4
0.00 0.09
2,417.75 0.013 (0.01, 0.017) 0.129 0.145 0.518 2,417.84 0.014 (0.011, 0.018) 0.130 0.087 0.568
0.5 0.7
3 3 5
0.10 0.43 1.09
2,417.86 0.014 (0.01, 0.017) 0.129 0.115 0.472 2,418.19 0.014 (0.011, 0.018) 0.130 0.092 0.635 2,418.85 0.014 (0.011, 0.018) 0.130 0.064 0.451
0.6 0.9 0.7
5
1.15
2,418.91 0.014 (0.011, 0.018) 0.131 0.051 0.474
0.9
AARS ET AL.: ESTIMATING BARENTS SEA POLAR BEARS
0 200 400 600 800 1000 1200 1400 >1400
Rus PI
Rus SI
0 200 400 600 800 1000 1200 1400 >1400
Rus L
0 200 400 600 800 1000 1200 1400 >1400
0 200 400 600 800 1000 1200 1400 >1400
Rus G
Nor PI
0 200 400 600 800 1000 1200 1400 >1400
Nor L
0 200 400 600 800 1000 1200 1400 >1400
0 200 400 600 800 1000 1200 1400 >1400
Nor G
45
Figure 2. Histograms of all 189 observations grouped into 200-m distance classes from survey lines. Detections are divided by habitat (glacier = G, land = L, pack ice = PI, and sea ice around islands = SI) and by region (Nor = Norwegian Arctic, Rus = Russian Arctic). The histogram to the upper right shows all 189 observations pooled. Observations truncated in the distance analyses are also shown.
and a lone animal before we were very close, and hence at the spatial scale we were operating, bear clusters seemed to offer a single cue independently of their size. Similarly, altitude was not important for detection function modeling based on AIC, and hence was subsequently ignored in the analysis. Overall, conditional on the covariates, the half-normal key function was usually more parsimonious than the hazard rate. Although the model with lowest AIC included only structure as a covariate, we opted to use the next best model (AIC = 0.1), which included the habitat covariate, for further inference. This was justified because we were interested in estimation over different strata with considerably different habitats, and from the field there was a clear perception that the detection process was different in different habitat types. Figure 2 shows the distribution of observed detection distances in different habitats and areas. Similar trends across strata made it possible to fit the half-normal detection function to all strata, with habitat and structure as covariates. Based on this model, we estimated bear abundance (and corresponding variances) for the seven strata (Table 1). In total 1,394 (95% CI: 1,060–1,743) bears were estimated to be in the areas surveyed by line transects. Bear densities were much higher in Russian than Norwegian areas, and higher on land than in other habitat types (Table 1). Pack Ice Areas not Surveyed The ratio estimator was r = 0.875 based on 56 observations and 64 telemetry fixes within the PI areas covered by line transects (Fig. 3). Based on line transect data, the size bias regression estimate of mean cluster size was 1.389, probability of detection within the covered strip of half-width 1,068 m was 0.472 (Russia) and
46
Fixes N
5
MARINE MAMMAL SCIENCE, VOL. 25, NO. 1, 2009
5
0
North
5
Obs S
0
Fixes S
South
Norway 0
20
Russia
40
60
80
Transect number
Figure 3. The number of observations by line transect flown (Obs S, in square A + B + C), the corresponding number of telemetry fixes allocated to boxes associated with the same transects (Fixes S, in square A + B + C), and the number of telemetry fixes in unsurveyed areas, extended from where transects terminated to 336 km north of the ice edge (D + E) and in the gap not surveyed (F). The area codes A–F correspond to those in Figure 1.
0.442 (Norway), and the proportion of the surveyed area covered was 1,068/4,500. Based on this, the number of bears in the unsurveyed areas of the PI was estimated to be 1,154 (95% CI: 659—1,845, see Table 3). The wide confidence interval obtained was partly due to the considerable spatial variation in numbers of observations and fixes (Fig. 3).
The Total Estimate Adding the estimated numbers of bears detected by total counts, those predicted to be within line transect survey areas and those predicted to be in unsurveyed areas of the PI, the total is 2,644 bears (95% CI: 1,899–3,592).
AARS ET AL.: ESTIMATING BARENTS SEA POLAR BEARS
47
Table 3. The estimated number of bears in areas of the pack ice not surveyed by line ˆ together with confidence intervals based on a ratio transects (area D + E + F, Fig. 1) N, estimator. Exp # groups = the number of bear groups expected to have been detected if the areas had been surveyed, based on the detection functions for the PI line transect strata.
Nor Uns Rus Uns Nor + Rus Uns
Exp # groups
ˆ N
95% CI for n
17.5 74.4 91.9
232 922 1,154
133; 377 527; 1,480 659; 1,845
DISCUSSION Population Estimate We estimated the Barents Sea area to host between approximately 1,900 and 3,600 polar bears, combining total counts (in small, easily surveyed areas), estimates from line transects (over large areas), and a ratio estimator (where planned line transects were not implemented due to safety and weather constraints). The ratio estimator approach used to estimate abundance in areas not surveyed was the best of a number of non-ideal alternatives, and the potential problems are discussed at length in a separate section below. However, it is unlikely that any such survey, covering such a large area of quickly changing weather conditions, in a necessarily narrow time window, will avoid similar difficult choices. We have attempted to represent the considerable uncertainty in our estimate fairly, resulting in a rather wide confidence interval. It is assumed that the Arctic has 19 relatively discrete subpopulations with a total of about 20,000–25,000 polar bears (Aars et al. 2006). Thus, the Barents Sea subpopulation contains a considerable fraction of the world’s population. Larsen (1986) suggested there were close to 2,000 bears in the Svalbard area, and 3,000– 6,700 in the area between East Greenland and FJL in 1980. Uncertainties around his and our estimate preclude a direct comparison. Derocher (2005) assumed that the subpopulation had increased in size until recently. Evaluation of Methodological Assumptions Total count—The relatively low number of bears observed during the total counts had little influence on the total estimate. The failure to detect bears during these counts is therefore likely to have negligible effect on the total abundance estimate. A few bears were also missed in areas of very open drift ice that we did not cover by helicopter. These areas were of minor extent, and we were fortunate to encounter a rather distinct ice edge that moved little during the period of the survey. Larger movements of the habitat during the survey would certainly have complicated the study. Six bears were seen from RV Lance in the FJL region, in areas that were not covered by line transects. This was in areas with very open water not suited for search by a one-engine helicopter, and where we beforehand had decided densities of bears would be too low to justify surveying. These observations suggest that a considerable number of bears might have been outside the areas we covered, and hence missed. In the absence of contemporaneous maps showing distribution of areas having very loose drift ice, it is not possible to provide a data-based number for individuals in
48
MARINE MAMMAL SCIENCE, VOL. 25, NO. 1, 2009
these areas. As an educated guess we believe that this might account for 100–200 bears. Distance sampling—The large number of randomly allocated line transects guarantees that the design was adequate for estimating abundance within the surveyed area. Avoidance movement before detection would lead to a biased estimate, but is likely to generate negligible bias due to the slow speed of the bears relative to the helicopter. Another possible problem would be if movement away from a surveyed line generated a temporarily higher density around neighboring survey lines before the helicopter reached these areas. All experience we have with polar bear behavior after disturbance is that they only react for a short period of time when the helicopter approaches. Also, in most cases, the helicopter traveling at 185 km/h would reach the next line before a polar bear, averaging a maximum of 4 km/h (Andersen et al. 2008), even when fueling or switching observational teams, and hence this is unlikely to be a significant issue. The adequacy of the measurement procedure was evaluated in Marques et al. (2006), and it is reasonable to say that, at the scale we were working, measurements were virtually error free. Therefore, it is unlikely that any of these assumptions might have had an impact on the final estimates. The most important issue regarding the reliability of the DS population estimate is g(0). If, say 10% of bears on the line passed undetected, then we expect our estimate to be only 90% of true abundance (Buckland et al. 2001). Our own perception was that bears were unlikely to be missed on land (gray or brown background), on most areas of glaciers, and on flat, newly formed sea ice. We think that along transects covering these habitats, g(0) was either 1 or very close to 1. However, in some sections of transects in the PI, with heavily packed ice, it is possible that polar bears close to the line could be missed. Heavily packed ice was most commonly found close to open water, at the ice edge. Two polar bear observations out of 56 in the PI area were in structure 3 habitat, and this habitat accounted for a small proportion of the total area covered. Even if g(0) was considerably lower than 1 in structure 3 habitat, we think that g(0) was close to 1 when averaged over all transects. However, in accordance with recommendations from, e.g., Borchers et al. (2006), estimation of g(0) as an integral part of the study should be a priority for future similar surveys. An eventual g(0) < 1, coupled with some underestimation on the areas covered with total counts and areas assumed to have no bears, means that, if anything, our estimate might underestimate the subpopulation size. Nonetheless, it is our belief that provided, as expected, g(0) was not considerably lower than 1, it is unlikely that these factors were enough to seriously affect the quality of our estimate. Area not surveyed—The most challenging problem resulted from the large number of transects not surveyed in the Russian part of the PI and the restricted coverage in areas more than 90 km north of the ice edge. Fixes of tagged females showed a tighter longitudinal spread than did observations made in the line transect sampling, which were relatively uniform in distribution from east to west (Fig. 3). This is presumably linked to non-random selection of animals for tagging (tagging effort was concentrated in eastern Svalbard). Because there was no survey effort in the north, we cannot assess whether there might be a similar mismatch between north and south; we must assume that the tagged bears have a representative distribution in late summer with respect to latitude, for the ratio estimator to be approximately unbiased. The bears can walk more than 50 km in a day (Amstrup 2003). Bears tagged in Svalbard can have home range areas of many thousand square kilometers, and use the sea ice of the PI as far as the most easterly areas surveyed (Mauritzen et al. 2002). Thus the geographic distances per se are unlikely to limit where they
AARS ET AL.: ESTIMATING BARENTS SEA POLAR BEARS
49
can be found along the PI. Nonetheless it is plausible to believe that bears from FJL are more commonly found at the areas farther to the east, and also that areas furthest west had more bears from northern Svalbard rather than from eastern Svalbard. Furthermore, some bears seen furthest west and east are likely to be from neighboring subpopulations, East Greenland to the west and Kara Sea to the east. This would explain the mismatch of distributions of telemetry fixes and observations from east to west (Fig. 3). This mismatch in longitudinal distributions raises the concern about whether the ratio estimator will give a biased estimate for the southern area (Rus Uns Fig. 1 and Fig. 3, area F) not surveyed. The number of fixes (22) in this area predicts an average 1.3 detections per line segments had they been covered. This is similar to what we had in the Russian sector to the west (1.5) and east (1.3). Thus we have no indication of which direction a bias for this area would be directed, and it seems unlikely that the bias is substantial given densities recorded were similar in neighboring areas. It is thus of greater concern whether the latitudinal distribution of fixes (converted to distance north of the ice edge) from several years match with how the bears were distributed. A possible source of bias would be if polar bears were found closer to the ice edge in years where the ice edge was located far north. Due to the high mobility of the polar bears, it is not likely that such a relative switch in positions relative to the edge would be caused by an increased distance from the islands per se. However, polar bears prefer areas of more shallow water with higher productivity (Ferguson et al. 2000). Areas further north from the ice edge have deeper water. The ice edge in 2004 was further north than average, and thus it could be that the ratio estimator using distributions from earlier years overestimated the number of bears in the uncovered areas to the north. However, the failure to reveal any heterogeneity in distances from the ice edge in fixes among years, despite a large year-to-year variation in the position of the ice edge, means that such a bias might be small. Another indication of this is the fact that 65 of the telemetry fixes used were from 1999, a year with a very similar ice edge location to 2004. The average distance of fixes from the ice edge was 57 nm in 1999, only slightly shorter than that for all the 169 fixes used in the analyses (61 nm). Other potential sources of bias are the possibility that males and females or animals at different age classes have different distributions. Little is known about this, because telemetry data are almost exclusively available for adult females. To avoid all the possible sources of bias, it would obviously have been preferable to have a representative sample of collared bears with respect to both area and status, and from the year of the survey. In our case, given that such data are not available, it might be possible to improve the estimate during coming years by adding new information from telemetry fix data if polar bears in the future are fitted with satellite transmitters in areas from where we currently have low coverage (northern part of Svalbard and Russian areas). Comparison with other studies—In our study, the average density of polar bears was 1.1/100 km2 in the areas surveyed by line transects (125,000 km2 ). In comparison, Taylor and Lee (1995) estimated an average density of 0.4 (range: 0.1–1.0) bears per 100 km2 on the PI in the Canadian Arctic in April. In the Chukchi Sea, Evans et al. (2003) estimated an average of 0.7 bears/100 km2 in August. Average densities of polar bears across different types of habitat in the Barents Sea are close to these estimates. However, we found a profound geographic variability. Densities on fast ice and PI in the Russian area were much higher (>2 bears/100 km2 ) than farther west in the Norwegian area. Such spatial patterns however will vary a lot with both seasons and years. Polar bears in the Barents Sea show high seasonal fidelity (Mauritzen et al. 2002), and many of the polar bears that are present around the islands of Svalbard in
50
MARINE MAMMAL SCIENCE, VOL. 25, NO. 1, 2009
the spring, will be hunting along the ice edge, particularly further northeast in the Russian area, and around FJL in August. During our survey there were three times as many bears in the Russian area as in the Norwegian area. Both the number of maternity dens (Larsen 1986) and the relatively high number of recaptures of bears in the Svalbard area (Derocher 2005) indicate that many more polar bears are present in the Svalbard area in spring. The fact that a large proportion of the subpopulation seems to migrate regularly across territorial borders emphasizes the need for a joint management between the two nations. Conclusions With the challenge of a warmer Arctic climate and significant habitat loss, it is a serious concern that several of the 19 existing subpopulations of polar bears still do not have reliable population size estimates (Aars et al. 2006). This is the first largescale line transect study aimed at estimating the size of a polar bear subpopulation. Difficulties and limitations were clearly demonstrated, particularly connected to icing conditions and fog preventing helicopter flying, which forced us to use a ratio estimator comparing telemetry data from earlier years to line transect observations to estimate bear densities in areas not surveyed. In future studies, time should be allocated to ensure a good coverage by transects independent of weather conditions, and ideally, helicopters capable of flying further north of the ice edge should be used. In face of the increasing demand for reasonable estimates of polar bear subpopulation sizes in the near future, we argue that line transect surveys are the best estimation method for such widely distributed populations. ACKNOWLEDGMENTS A large number of people were involved in preparing this study. We thank D. Vongraven, E. Born, A. E. Derocher, K. Kovacs, M. Ekker, A. Studenetski, and H. Goodvin for their contributions. S. Christensen-Dalsgaard, E. Johansen, T. Severinsen, and V. Bakken helped with the field survey. M. Mauritzen provided help with satellite telemetry data files. Thanks also to the crew on RV Lance, the pilots and mechanics from AIRLIFT, and The Governor of Svalbard. The Norwegian Ministry of the Environment funded the survey. Three anonymous reviewers and Associate Editor W. Testa all contributed to considerably improve the manuscript.
LITERATURE CITED AARS, J., N. J. LUNN AND A. E. DEROCHER, EDS. 2006. Polar bears. Proceedings of the 14th working meeting of the IUCN/SSC Polar Bear Specialist Group, Seattle, WA, 20–24 June 2005. IUCN, Gland, Switzerland and Cambridge, U.K. ˚ Viken, and T. Bakken, AMSTRUP, S. C. 2003. Polar bears. Pages 365–372 in J. A. K˚al˚as, A. eds. Wild mammals of North America. Biology, management and conservation. The Johns Hopkins University Press, Baltimore, MD. ANDERSEN, M., E. LIE, A. E. DEROCHER, S. E. BELIKOV, A. BERNHOFT, A. N. BOLTUNOV, G. W. GARNER, J. U. SKAARE AND Ø. WIIG. 2001. Geographic variation of PCB congeners in polar bears (Ursus maritimus) from Svalbard east to the Chukchi Sea. Polar Biology 24:231–238. ANDERSEN, M., A. E. DEROCHER, Ø. WIIG AND J. AARS. 2008. Recording detailed movements of Svalbard polar bears using geographical positioning system satellite transmitter. Polar Biology. In press.
AARS ET AL.: ESTIMATING BARENTS SEA POLAR BEARS
51
BERNHOFT, A., Ø. WIIG AND J. U. SKAARE. 1997. Organochlorines in polar bears (Ursus maritimus) at Svalbard. Environmental Pollution 95:159–175. BORCHERS, D. L., S. T. BUCKLAND AND W. ZUCCHINI. 2002. Estimating animal abundance. Closed populations. Springer-Verlag, London, U.K. BORCHERS, D. L., J. L. LAAKE, C. SOUTHWELL AND C. G. M. PAXTON. 2006. Accommodating unmodeled heterogeneity in double-observer distance sampling surveys. Biometrics 62:372–378. BUCKLAND, S. T., D. R. ANDERSON, K. P. BURNHAM, J. L. LAAKE, D. L. BORCHERS AND L. THOMAS. 2001. Introduction to distance sampling. Oxford University Press, Oxford, U.K. BUCKLAND, S. T., D. R. ANDERSON, K. P. BURNHAM, J. L. LAAKE, D. L. BORCHERS AND L. THOMAS, EDS. 2004. Advanced distance sampling. Oxford University Press, Oxford, U.K. DAVISON, A. C., AND D. V. HINKLEY. 1997. Bootstrap methods and their application. Cambridge University Press, New York, NY. DEROCHER, A. E. 2005. Population ecology of polar bears at Svalbard, Norway. Population Ecology 47:267–275. DOWNING, K., AND M. REED. 1996. Object-oriented migration modelling for biological impact assessment. Ecological Modelling 93:203–219. EVANS, T. J., A. FISCHBACH, S. SCHLIEBE, B. MANLY, S. KALXDORFF AND G. YORK. 2003. Polar bear aerial surveys in the eastern Chukchi Sea: A pilot study. Arctic 56:359–366. FERGUSON, S. H., M. K. TAYLOR AND F. MESSIER. 2000. Influence of sea ice dynamics on habitat selection by polar bears. Ecology 81:761–772. ISAKSEN, K., V. BAKKEN AND Ø. WIIG. 1998. Potential effects on seabirds and marine mammals of petroleum activity in the northern Barents Sea. Norsk Polarinstitutt Meddelelser 154. 66 pp. LARSEN, T. 1986. Population biology of the polar bear (Ursus maritimus) in the Svalbard area. Norsk Polarinstitutt Skrifter 184. 55 pp. MANLY, B. F. J., L. L. MCDONALD AND G. W. GARNER. 1996. Maximum likelihood estimation for the double-count method with independent observers. Journal of Agricultural, Biological, and Environmental Statistics 1:170–189. MARQUES, F. F. C., AND S. T. BUCKLAND. 2003. Incorporating covariates into standard line transect analyses. Biometrics 59:924–935. MARQUES, T. A., M. ANDERSEN, S. CHRISTENSEN-DALSGAARD, S. BELIKOV, A. BOLTUNOV, Ø. WIIG, S. T. BUCKLAND AND J. AARS. 2006. Comparing distance estimation methods in a helicopter line transect survey. Wildlife Society Bulletin 34:759–763. MAURITZEN, M., A. E. DEROCHER AND Ø. WIIG. 2001. Space-use strategies of female polar bears in a dynamic sea ice habitat. Canadian Journal of Zoology 79:1704–1713. MAURITZEN, M., A. E. DEROCHER, Ø. WIIG, S. E. BELIKOV, A.N. BOLTUNOV, E. HANSEN AND G. W. GARNER. 2002. Using satellite telemetry to define spatial population structure in polar bears in the Norwegian and western Russian Arctic. Journal of Applied Ecology 39:79–90. MAURITZEN, M., A. E. DEROCHER, O. PAVLOVA AND Ø. WIIG. 2003. Female polar bears, Ursus maritimus, on the Barents Sea drift ice: Walking the treadmill. Animal Behaviour 65:107–113. NORSTROM, R. J., S. E. BELIKOV, E. W. BORN, G. W. GARNER, B. MALONE, S. OLPINSKI, M. A. RAMSAY, S. SCHLIEBE, I. STIRLING, M. S. STISHOV, M. K. TAYLOR AND Ø. WIIG. 1998. Chlorinated hydrocarbon contaminants in polar bears from eastern Russia, North America, Greenland, and Svalbard: Biomonitoring of arctic pollution. Archives of Environmental Contamination and Toxicology 35:354–367. PRESTRUD, P., AND I. STIRLING. 1994. The International Polar Bear Agreement and the current status of polar bear conservation. Aquatic Mammals 20:1–12. R DEVELOPMENT CORE TEAM. 2004. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria.
52
MARINE MAMMAL SCIENCE, VOL. 25, NO. 1, 2009
STRINDBERG, S., S. T. BUCKLAND AND L. THOMAS. 2004. Design of distance sampling surveys and geographic information systems. Pages 190–228 in S. T. Buckland, D. R. Anderson, K. P. Burnham, J. L. Laake, D. L. Borchers and L. Thomas, eds. Advanced distance sampling. Oxford University Press, Oxford, U.K. TAYLOR, M. K., AND L. J. LEE. 1995. Distribution and abundance of Canadian polar bear populations: A management perspective. Arctic 48:147–154. THOMAS, L., J. L. LAAKE, S. STRINDBERG, F. F. C. MARQUES, S. T. BUCKLAND, D. L. BORCHERS, D. R. ANDERSON, K. P. BURNHAM, S. L. HEDLEY, J. H. POLLARD, J. R. B. BISHOP AND T. A. MARQUES. 2006. Distance 5.0. Release 1. Research Unit for Wildlife Population Assessment, University of St. Andrews, U.K. http://www.ruwpa.st-and.ac.uk/distance. WIIG, Ø. 1995. Distribution of polar bears (Ursus maritimus) in the Svalbard area. Journal of Zoology 237:515–529. WIIG, Ø. 2005. Are polar bears threatened? Science 309:1814–1815. WIIG, Ø., AND V. BAKKEN. 1990. Aerial strip surveys of polar bears in the Barents Sea. Polar Research 8:309–311. WIIG, Ø., AND A. DEROCHER. 1999. Application of aerial survey methods to polar bears in the Barents Sea. Pages 27–36 in G. W. Garner, S. C. Amstrup, J. L. Laake, B. F. J. Manly, L. L. McDonald and D. G. Robertson, eds. Marine mammal survey and assessment methods. Balkema, Rotterdam, The Netherlands. Received: 17 October 2007 Accepted: 30 May 2008
Paper 2
Photo: Magnus Andersen
Polar Biol (2008) 31:905–911 DOI 10.1007/s00300-008-0428-x
ORIGINAL PAPER
Movements of two Svalbard polar bears recorded using geographical positioning system satellite transmitters Magnus Andersen · Andrew E. Derocher · Øystein Wiig · Jon Aars
Received: 20 June 2007 / Revised: 29 January 2008 / Accepted: 30 January 2008 / Published online: 21 February 2008 © Springer-Verlag 2008
Abstract Until recently, studies on polar bear (Ursus maritimus) movements and space use have used data collected by satellite telemetry collars that provided positions infrequently (typically weekly) and with low precision (by Doppler Shift method). A new generation of transmitters incorporated into collars use the Global Positioning System (GPS) to provide highly accurate positions, and have the ability to provide many positions per day. We used data from two GPS collars Wtted to female polar bears, that attempted to collect six positions per day (4-h apart) for 546 days (from April 2000 to September 2001) and 413 days (from April 2000 to May 2001) to estimate how estimated speed of movement and home range size increase with increasing number of data points. Using all the positions, we estimated that the bears moved a minimum of 14.3 and 15.8 km per day on average. The fractal dimension (D) of the movement pathways for the two bears were D = 1.28 and 1.31, respectively, indicating low tortousity of the movements. Their minimum estimated annual home range areas were 20,794 and 112,183 km2. Simulations showed that a commonly used sampling regime of one location every 6th day would have signiWcantly underestimated the movement rates and the home range sizes compared to our estimates. We also used the high accuracy of GPS positions M. Andersen (&) · A. E. Derocher · J. Aars Norwegian Polar Institute, 9296 Tromsø, Norway e-mail:
[email protected] Present Address: A. E. Derocher Department of Biological Sciences, University of Alberta, Edmonton, Canada, T6G 2E9 Ø. Wiig Natural History Museum, University of Oslo, PO Box 1172, Blindern, 0318 Oslo, Norway
to look at distances moved within 4-h periods. Maximum movement rate during a 4-h period for the two bears was 4.21 and 4.58 km/h, respectively. Variation in median values by month was signiWcant (0.01 km/h in November for N23476 to 1.48 km/h in December for N7955). Diurnal variation was observed to diVer between deWned periods. Keywords Polar bear · Satellite telemetry · GPS-collar · Movements · Activity · Svalbard
Introduction Animal movement and space use reXect their ecology, since their movements relate to how they acquire resources such as food and dens or nests, how they interact intraspeciWcally, and how they respond to their physical environment (Bell 1991; Ims 1995; Mauritzen 2002). Therefore the study of movement and space-use in a species can give fundamental knowledge of the biology of the animal. The use of satellite telemetry in the study of polar bear (Ursus maritimus) movement and distribution was Wrst applied between Svalbard, Norway and Greenland in 1979 when four polar bears were equipped with satellite transmitters (IUCN/SSC 1981; Larsen et al. 1983). These early transmitters were large and heavy (5 kg) and were built into a backpack attached to the animal with a harness. The designs quickly improved and the second generation transmitters were smaller and built into collars around the animal’s neck (Harris et al. 1990). Collar locations were calculated by measuring the Doppler Shift on the frequency of the satellite transmitter. The Wrst satellite system used the NIMBUS 6, while later transmitters have been sending signals to the Argos System (Harris et al. 1990). One of the latest technological developments for studies of wildlife
123
906
has been to use the Global Positioning System (GPS) to determine the location of the animal (e.g., Johnson et al. 2002; Gau et al. 2004; Frair et al. 2004; Morales et al. 2004), and the Argos System to collect these data from the transmitters (e.g., Yasuda and Arai 2005; Parks et al. 2006). The aim of the current study was to investigate the eVectiveness of GPS satellite collars in polar bear studies, describe the activity and movement patterns of individuals, and assess the increase in information gained by using a GPS sampling regime.
Materials and methods As a pilot project three newly developed GPS collars were deployed on female polar bears in Svalbard. Two of the collars started transmitting as expected, whereas one, for unknown reasons, failed to initialize. One GPS collars was Wtted to a 14 year old female polar bear (N23476) with two cubs-of-the-year, one on a 9 year old female (N7955) with a year old cub. Hereafter, we will use the ID codes of the two bears when discussing the GPS collar. The polar bears were captured on sea ice at Spitsbergen, Svalbard, Norway (74–81°N, 15–45°E) in April 2000 as a part of a research program on the ecology of the Barents Sea subpopulation. The bears were caught by remote injection of a dart (Palmer Cap-Chur Equipment, Douglasville, Georgia, USA) containing the drug Zoletil® (Virbac, Carros, France) Wred from a helicopter (Stirling et al., 1989). Animal handling methods were approved by the National Animal Research Authority (Norwegian Animal Health Authority, Oslo, Norway). The GPS transmitters were programmed to collect and send six positions daily with an expected lifetime of 15 months. The satellite telemetry system deployed were the Gen 2 GPS Collar (model TGW-100, Telonics Inc., Mesa, Arizona, USA), and the data collected were transmitted via the Argos System using the Data Collection Service oVered by CLS (Collecte Localisation Satelites, France). Data were processed using the Argos Data Converter -T03 software (ADC-T03, version 2.0) designed by Telonics Inc, and only locations assigned the quality indicator (Wx status) “Good” was included in the analyses. Diurnal and monthly activity was described by box plots (S-PLUS® 7.0 for Windows). We used box plot notches to assess statistical diVerences between periods. When notches of two plots did not overlap this was considered as strong evidence that the two medians did diVer (P · 0.05) (Chambers et al. 1983). Diurnal patterns are only reported for movements between positions recorded 4 h apart, and analysed by season, habitat and movement behaviour. Positions were compared to sea ice information obtained from manual inspection of daily updated ice maps (www.met.no) to conWrm the use of sea ice versus use of land by the bears.
123
Polar Biol (2008) 31:905–911
Estimation of distances moved was based on all locations collected, while estimations of movement rates were based on location intervals 100 m. The nature of typical snow conditions was shown to eVectively dampen both vibration and noise levels. These results are in agreement with the Wndings of Amstrup (1993), who concluded that some denning female polar bears in Alaska tolerated high levels of activity in the vicinity of their dens; however, Amstrup (1993) also noted instances of potential den abandonment. The current discussion illustrates varying responds to snowmobile traYc and other motorized vehicles and types of disturbance by a number of species. The role of habituation in relation to disturbance is only brieXy discussed in most studies, mainly due to the lack of relevant data (e.g. McLaren and Green 1985). This is also the case in the current study. During the pilot project in 2003; however, one observation illustrates that polar bears may become habituated to snow mobile traYc and human presence. In April and May 2003, a female polar bear with a one year old cub stayed within a small area (5 £ 10 km) near Tempelfjorden (17°20⬘E, 78°20⬘N) on the west coast of Spitsbergen, Svalbard, for at least 3 weeks during the peak snow mobile season. She spent most of her time along the 3 km glacier front at the base of the fjord. This area is one of the most frequently visited sites in Svalbard by snowmobile tourists during this time of the year. During the peak season as many as 100–200 snowmobiles make daily visits, mainly as part of organized trips from Longyearbyen. This female was observed nursing her cub with approximately 50 snowmobiles present within 100 m, while vehicles entered and left the area. Only when one snowmobile approached as close as 25 m did she respond by moving closer to the glacier front and hide among broken glacier ice. Remains of both successful and unsuccessful seal hunting were found along the glacier front and it was obvious that the area was good hunting habitat, which might be why the female chose to stay despite the daily disturbance and sporadic harassment. Polar bears are extremely mobile animals that can move several thousand kilometres each year, and thus can easily move out of an area if suYciently disturbed. Many polar bears have large home ranges (up to 373, 539 km2, Mauritzen et al. 2001), and these individuals will have no problem Wnding areas where disturbance is low. Other bears; however, have been shown to have very small home ranges (185 km2, Mauritzen et al. 2001), and these may be more vulnerable if the area experiences heavy disturbance.
123
506
However, studies referred to earlier (Blix and Lentfer 1992; Amstrup 1993 and Dyck and Baydack 2004) have shown that polar bears are fairly tolerant to some types of disturbance. On a large scale, polar bears are very mobile, but when considering small scale movement behaviour (such as within a fjord) polar bears have been shown to have restricted movements. Stirling (1974) described the behaviour of polar bears on the sea ice at Devon Island, Canada, through direct observations. He found that polar bears spent 66.6% of their time inactive (sleeping, lying, still-hunting), and 25% walking. Only a few instances of short bursts of running were observed by Stirling (1974), all in association with meetings between unrelated polar bears. In our study, we observed polar bears running at least 1 km after being disturbed by snowmobiles, and several bears left the ringed seal breathing hole where they were still-hunting when vehicles approached. We believe that repeated disturbance leading to running and interrupted hunting could result in increased energetic stress on the animals. Further, in extreme situations, overheating from running could lead to death of the animal. Polar bears are not adapted to high pace movement over extended distances, and large individuals in particular overheat quickly if pursued over time (Øritsland 1970). The current study has shown that females with cubs respond most strongly to disturbance, and the added stress experienced by the cubs could have negative eVects on their survival. Such stress could involve more swimming in open water (polar bears tend to take refuge in water when startled) or more frequently interrupted suckling/feeding bouts, both could aVect body condition and growth. Wildlife disturbance studies rarely are able to assess eVects on survival, reproductive success or other eVects on the population level. This is a problematic shortcoming, since it is the most important piece of information needed for authorities to evaluate the net eVect to bears and the population, and to make sound management decisions. With increasing tourism in many Arctic regions it is important to have a more complete understanding of this issue. Such studies would however be extremely demanding, regarding both resources and eVort, and we will therefore unfortunately depend on studies of eVects on behaviour and maybe also on physiological responses. This can however be valuable as long as the biology of the species is well understood and thus one can make plausible interpretations about how these responses may link to demographic processes. The current project has shown that polar bears, especially females with small cubs, have a potential to get heavily disturbed by snowmobiles because they react at long distances. Their reactions are often profound, even though the source disturbance was relatively mild. However, long-
123
Polar Biol (2008) 31:501–507
term behavioral responses to snowmobile disturbance and their signiWcance to the welfare of populations are still unknown. Acknowledgments This project was funded by The Governor of Svalbard and the Norwegian Polar Institute. We would like to thank Petter Wabakken and Øystein Overrein for their assistance in the Weld and for constructive discussions, and Audun Stien for helping in statistical analyses. We would also like to thank the anonymous reviewers for comments and suggestions that greatly improved the manuscript.
References Aars J, Lunn NJ, Derocher AE (2006) Polar bears: proceedings of the 14th working meeting of the IUCN/SSC Polar Bear Specialist Group, 20–24 June 2005, Seattle. IUCN, Gland, Switzerland and Cambridge, UK Amstrup SC (1993) Human disturbance of denning polar bears in Alaska. Arctic 46:246–250 Babb TA, Bliss LS (1974) EVects of physical disturbance on arctic vegetation in The Queen Elisabeth’s Islands. J Appl Ecol 11:549–562 Blix AS, Lentfer JW (1992) Noise and vibration levels in artiWcial polar bear dens as revealed to selected petroleum exploration and developmental activities. Arctic 45:20–24 Born EW, Riget FF, Dietz R, Andriachek D (1999) Escape responses of hauled out ringed seals (Phoca hispida) to aircraft disturbance. Polar Biol 21:171–178 Calef GW, DeBock EA, Lortie GM (1976) The reaction of Barren Ground Caribou to aircraft. Arctic 29:201–212 Chapin III FS, Shaver GR (1981) Changes in soil properties and vegetation following disturbance of Alaskan arctic tundra. J Appl Ecol 18:605–617 Colman JE, Jacobsen BW, Reimers E (2001) Summer response distances of Svalbard reindeer (Rangifer tarandus platyrhynchus) to provocation by humans on foot. Wildl Biol 7:275–283 Creel S, Fox JE, Hardy A, Sands J, Garrott B, Peterson RO (2002) Snowmobile activity and glucocorticoid stress responses in wolves and elk. Conserv Biol 16:809–814 Derocher AE, Wiig Ø, Andersen M (2002) Diet composition of Polar Bears in Svalbard and the Western Barents sea. Polar Biol 25:448–452 Dunnet GM (1977) Observations on the eVects of low-Xying aircraft at seabird colonies on the coast of Aberdeenshire, Scotland. Biol Conserv 12:55–64 Dyck MG, Baydack RK (2004) Vigilance behaviour of polar bears (Ursus maritimus) in the context of wildlife-viewing activities at Churchill, Manitoba, Canada. Biol Conserv 116:343–350 Eckstein RG, O’Brien TF, Rongstad OJ, Bollinger JG (1979) Snowmobile eVects on movements of White-tailed Deer: a case study. Environ Conserv 6:45–51 Eid PM, Prestrud P, Eide NE (2001) Menneskelig forstyrrelse av fjellrev, rein og sel. Forrapport til Sysselmannen på Svalbard (In Norwegian) 6 pp Freddy DJ, Bronaugh WM, Fowler MC (1986) Responses of mule deer to disturbance by persons afoot and snowmoliles. Wildl Soc Bull 14:63–68 Gabrielsen GW, Smith EN (1995) Physiological responses of wildlife to disturbance. In: Knight RL, Gutzwiller KJ (eds) Wildlife and recreationists. Coexistence through management and research. Island, Washington, DC 95–107 Kelly PB, Burns JJ, Quakenbush LT (1988) Responses of ringed seals (Phoca hispida) to noise disturbance. In: Port ocean engineering under arctic conditions. Vol. II. Sackinger WM, JeVries MO (eds)
Polar Biol (2008) 31:501–507 Symposium on noise and marine mammals, August 1987. University of Alaska, Fairbanks. p 27–39 Lunn NJ, Schliebe S, Born EW (2002) Polar Bears: proceedings of the 13th working meeting of the IUCN/SSC Polar Bear Specialist Group, Nuuk, Greenland. IUCN, Gland Switzerland and Cambridge. Vii + 153 pp MacArthur RA, Johnston RH, Geist V (1979) Factors inXuencing heart rate in free-ranging bighorn sheep: a physiological approach to the study of wildlife harassment. Can J Zool 57:2010– 2021 Mauritzen M (2002) Patterns and processes in female polar bear spaceuse. Ph.D. Thesis, University of Oslo, Norway Mauritzen M, Derocher AE, Wiig Ø (2001) Female polar bear space use strategies in a dynamic sea ice habitat. Can J Zool 79:1704– 1713
507 McLaren MA, Green JE (1985) The reactions of muskoxen to snowmobile harassment. Arctic 38:188–193 Overrein Ø (2002) Virkninger av motorferdsel på fauna og vegetasjon: Kunnskapsstatus med relevans for Svalbard (In Norwegian). Norsk Polarinstitutt Rapportserie 119. 28pp Prestrud P, Stirling I (1994) The international polar bear agreement and the current status of polar bear conservation. Aquat Mamm 20:113–124 Stirling I (1974) Midsummer observations on behavior of wild polar bears (Ursus-maritimus). Can J Zool 52:1191–1198 Tyler NJC (1991) Short-term behavioural responses of Svalbard reindeer (Rangifer tarandus platyrhyncus) to direct provocation by a snowmobile. Biol Conserv 56:179–194 Øritsland NA (1970) Temperature regulation of the polar bear (Thalarctos maritimus). J Comp Biochem Physiol A 37:225–233
123
Paper 7
Photo: Magnus Andersen
Polar Biol (2002) 25: 448–452 DOI 10.1007/s00300-002-0364-0
O R I GI N A L P A P E R
Andrew E. Derocher Æ Øystein Wiig Æ Magnus Andersen
Diet composition of polar bears in Svalbard and the western Barents Sea
Accepted: 21 January 2002 / Published online: 20 March 2002 Ó Springer-Verlag 2002
Abstract We estimated both the numerical and biomass composition of the prey of polar bears (Ursus maritimus) from 135 opportunistic observations of kills in Svalbard and the western Barents Sea collected from March to October 1984–2001. By number, the prey composition was dominated by ringed seals (Phoca hispida) (63%), followed by bearded seals (Erignathus barbatus) (13%), harp seals (P. groenlandica) (8%) and unknown species (16%). However, when known prey were converted to biomass, the composition was dominated by bearded seals (55%), followed by ringed seals (30%) and harp seals (15%). Results indicated that bearded seals are an important dietary item for polar bears in the western Barents Sea. We believe that different patterns of space use by different bears may result in geographic variation of diet within the same population.
relative energetic contribution of prey species and the seasonal composition of prey. An earlier study of polar bear diet in Svalbard also suggested that ringed and bearded seals were the main prey (Lønø 1970). Uncertainty in diet composition arises because polar bears also consume a variety of other species, including walrus (Odobenus rosmarus) (Calvert and Stirling 1990), white whales (Delphinapterus leucas) and narwhal (Monodon monoceros) (Lowry et al. 1987; Smith and Sjare 1990), harp seals (P. groenlandica) (Lønø 1970), seabirds (Stempniewicz 1993) and carrion (Christiansen 1981). However, most studies concur that seals are the main prey. In this paper, we document polar bear predation on ringed, bearded and harp seals in Svalbard and the western Barents Sea from observations of seal kill sites. We also discuss the relative importance of these species in the diet of polar bears.
Introduction Materials and methods The most carnivorous of the Ursidae, polar bears (Ursus maritimus) are thought to prey largely on ringed seals (Phoca hispida) and, to a lesser extent, on bearded seals (Erignathus barbatus) (Stirling and Archibald 1977; Smith 1980; Gjertz and Lydersen 1986; Stirling and Øritsland 1995). However, the diet of polar bears is still poorly understood, with limited information about the
A.E. Derocher (&) Æ M. Andersen Norwegian Polar Institute, 9296, Tromsø, Norway E-mail:
[email protected] Tel.: +47-77-750524 Fax: +47-77-750501 Ø. Wiig Zoological Museum, University of Oslo, P.O. Box 1172 Blindern, 0318, Oslo, Norway
Sampling occurred from 1984 to 2001, from March to October, on an opportunistic basis. The total sampling area is approximately contained by the outermost locations of seal kills (Fig. 1). Polar bears in the sampling area represented all age, sex and reproductive classes in the population, and we believe the seal kills were representative of those taken by the population. Seal kills on the sea ice are easy to identify due to the presence of blood, scavengers (glaucous gulls, Larus hyperboreus, ivory gulls, Pagophila eburnea, arctic fox, Alopex lagopus), or bears at the kill. We were able to exclude arctic fox predation based on presence or absence of tracks. Most (n=113) seal kill sites were located while tracking polar bears by helicopter, during research on polar bear ecology. Observations collected during August (1999) were obtained during an aerial survey to estimate polar bear abundance. Some information on kills was collected from other researchers working in the area. We included 22 samples noted in earlier studies of seal predation to increase geographic coverage. We included 6 ringed seal kills from Gjertz and Lydersen (1986), 1 ringed seal kill from Lydersen and Gjertz (1986) and 15 ringed seal kills from Wiig et al. (1999). Limited information from kills was available from most observations. Identification of species and age class of the kill was based on assessment from the air, ship or by examination of
449
Results
Fig. 1. Distribution of ringed seal, bearded seal, harp seal and unknown seal species killed by polar bears in Svalbard and the western Barents Sea in 1984–2001. The study area is approximately delineated by the outermost kill sites remains. It was sometimes not possible to identify the prey species or age as only bone fragments or blood remained. During the spring pupping season, kills at digs in rough pressure-ridged ice with little blood in the area were classified as ringed seal pups. Adult bearded and harp seals were obvious from their size. Because teeth and claws were available from only a few samples (less than ten animals), ages from these were not determined. We estimated prey biomass in the diet of polar bears using a mean mass as an approximation of the size for each seal species. We excluded all prey of unknown species and age class for biomass estimation. The mass of adult seals was obtained from growth curves using the mean asymptotic sizes of females and males. Adult mass of ringed seals was set at 57 kg (Lydersen and Gjertz 1987), 273 kg for bearded seals (Andersen et al. 1999) and 131 kg for harp seals (Innes et al. 1981). We pooled juvenile and adult seals and used adult size in calculations to treat the three species in a similar manner, but recognised that this may be a source of error. We averaged the birth and weaning masses to estimate pup body mass, from published studies. For ringed seals, we used 11 kg, based on a birth mass of 4.55 kg and a weaning mass of 18 kg (Lydersen et al. 1992). Similarly, we used 62 kg for bearded seal pups, which was based on a birth mass of 33 kg (Burns 1981) and a weaning mass of 92 kg (growth of 3.3 kg/day over 18 days) (Lydersen et al. 1994).
Table 1. Number of seal kills found opportunistically in Svalbard and the western Barents Sea from 1984 to 2001 and attributed to polar bear predation, in relation to month
Month
Information was obtained on 135 seal kills (Table 1) with 120 (89%) collected in 1995–2001. Numerically, the majority of the kills were ringed seals (63%), followed by bearded seals (13%), harp seals (8%) and 16% of unknown species. The numerical kill composition of known species (n=114) was 75% ringed seal, 16% bearded seal and 9% harp seal. Most (14/21) of the unknown prey species were from August, when sampling was conducted during an aerial survey and checking kills on the ground was not possible. Of the ringed seal kills, pups composed 72% (54/75) of those classified to age group. For bearded seals, 31% (5/16) were pups. The prey composition by biomass (n=98) was estimated to be composed of 50% adult bearded seal, 20% adult ringed seal, 10% ringed seal pup, 15% harp seal and 5% bearded seal pup. The seasonal distribution of predation was difficult to assess due to non-representative sampling through the months. However, ringed seal predation appeared to dominate in spring during the pupping season, and predation on bearded seals was more evenly distributed through the sampling period. Observation of harp seal predation was restricted to June. Twenty-four kills were observed during August, with eight in multiyear ice. While many of these kills were not identified to species, bearded seals were abundant in the area and two of four adult bearded seal kills in August were in multiyear ice. Insufficient data were available to determine the spatial patterns of the kills but it appeared that predation events occurred throughout the study area (Fig. 1).
Discussion Estimating the diet of any free-ranging animal is a difficult undertaking and prone to inaccuracies. Given the inaccessibility of polar bear habitat, it is extremely problematic to obtain an overview of their diet, particularly in remote areas, during summer when the ice is melting, and during the winter dark period. The methods used in our study were opportunistic and classification into age classes was approximate. Using a mean pup mass and asymptotic mass for adults may result in biases of estimating intake if, for example, bearded seals
Ringed seals Ad/Juv
March 1 April 9 May 7 June – August 3 October 1 Total 21
Bearded seals
Harp seals
Pup
Unk
Ad/Juv
Pup
Unk
Ad/Juv
Unk
2 35 17 – – – 54
– 6 1 – 3 – 10
– 3 2 2 4 – 11
– 1 4 – – – 5
– – 1 – – 1 2
– – – 7 – – 7
– – – 4 – – 4
Unknown
– 4 2 – 14 1 21
450
killed were well below the asymptote. However, we feel that the methods used are a reasonable representation and any method would result in some level of bias. We could not assess possible sampling bias of the different prey. For example, it is possible that a fresh kill of a large bearded seal may be more easily detected than a smaller ringed seal killed some days earlier. Therefore, interpretation of our data must proceed with caution. Marine-mammal resources available to polar bears in the study area are poorly understood and quantitative estimates are unavailable for most potential prey. Bearded seals are widely distributed throughout Svalbard and the western Barents Sea (Benjaminsen 1973), and their distribution overlaps substantially with that of polar bears. Bearded seals are largely benthic feeders and can dive to depths up to 400 m (Burns 1981; Gjertz et al. 2000), so that most of the Barents Sea, which is less than 300 m deep, may provide feeding habitat. The abundance of bearded seals is uncertain but may number in the 300,000 range in the North Atlantic (Burns 1981). The ringed seal population size in the Svalbard area is unknown but the global population likely numbers in the millions (Reeves 1998). In Svalbard and the western Barents Sea, there is ringed seal reproduction in both land-fast ice (Smith and Lydersen 1991) and drifting pack-ice (Wiig et al. 1999). During studies of ringed seal breeding habitat in Svalbard, a discrepancy was noted in the production of ringed seals and the number of ringed seals required to support the polar bear population in the area (Smith and Lydersen 1991). Smith and Lydersen (1991) suggested that pack-ice production of ringed seals may be an important contribution to the population. Our results further confirm the findings of Wiig et al. (1999) that ringed seals breed in the drifting pack ice of the Barents Sea, particularly northwest of Hopen Island. Ringed seals are available to all bears in the study population. The Barents Sea harp seal population is approximately 2.2 million animals (Nilssen et al. 2000) and represents a potentially abundant food source for polar bears. However, harp seals do not reach polar bear habitat until April/May and then increase in abundance along the drift-ice edge until October, when they return south (Haug et al. 1994; Nordøy et al. 1998). Harp seals are available to polar bears particularly in early summer, when both species select open ice (10–60% cover) (Haug et al. 1994). However, some harp seals are pelagic and this portion of the population is unavailable to polar bears (Nordøy et al. 1998). Further research is needed to quantify the importance of harp seals to this polar bear population. The only previous study of polar bear diet in the study area comes from bears harvested throughout the year near Svalbard; 52 ringed seals, 10 bearded seals and 6 harp seals were found in stomachs (Lønø 1970). Harp seals were only found during June/August and most bearded seals (9/10) were found in the same period. The prey composition from this study was 76% ringed seal,
15% bearded seal and 9% harp seal, and was very similar to the composition of the 114 samples of known species in our study. Numerically, similar to earlier studies, ringed seals are the dominant prey of polar bears. However, on a biomass basis, the results from Lønø (1970), together with ours, suggest that the diet of polar bears in Svalbard and the western Barents Sea has a significant contribution from bearded seals. However, in the eastern Barents Sea, a Russian study reported 68% ringed seal, 22% walrus and miscellaneous other items for the diet of polar bears (Parovshchikov 1964) and may reflect further geographic variation in the same population. Studies of fatty-acid profiles of polar bears suggest that geographic variation in diet may be large (Iverson et al. 1999). Polar bears in Svalbard and the western Barents Sea area are part of a common population that extends as far east as Franz Josef Land (Mauritzen et al. 2002). Polar bears living in the study area have two different space use patterns: one group lives near shore and has small annual ranges whereas the other lives offshore and has larger ranges (Mauritzen et al. 2001). Annual range size of adult females ranged from 185 to 373,539 km2 and dietary differences were postulated to explain the different space use patterns (Mauritzen et al. 2001). In particular, Mauritzen et al. (2001) suggested that nearshore bears relied more on the land-fast ice and preyed largely on ringed seals during spring while pelagic bears preyed more on bearded and harp seals over a longer period. Our results support this hypothesis given that most of the kills observed in June and August were in multiyear pack-ice where the pelagic bears tend to summer. Sampling of polar bear kills is not easily accomplished and most other studies have been conducted over a relatively brief period during the spring when ringed seals pup (Stirling and Archibald 1977; Smith 1980). Further, these studies have been restricted to the Canadian Archipelago where stable ice creates good ringed seal breeding habitat (Hammill and Smith 1989, 1991; Furgal et al. 1996). The results from our study suggest that polar bears in the Barents Sea, which summer in the multiyear ice, may feed on seals all year. Polar bears are opportunistic and other prey species are available to them in the study area. Hooded seals (Cystophora cristata) range northward to the ice edge in summer and overlap with polar bears (Gjertz 1991). We found no instances of walrus predation but, as the walrus population in Svalbard recovers from overharvest (Gjertz and Wiig 1995), it is possible that walrus predation may increase. In addition, harbour seals (Phoca vitulina) are found in the Svalbard Archipelago (Gjertz et al. 2001) and, during the ice-free period in August 2001, we observed polar bears attempting aquatic stalks of hauled-out harbour seals in Van Keulen Fjord on the west coast of Svalbard. We did not observe successful predation but did observe several polar bear tracks in the bottom sediments near shore and saw two bears swimming toward hauled-out seals.
451
Polar bears are also opportunistic scavengers. In summer 2001, polar bears were observed feeding on both a white whale carcass and a sperm whale (Physeter macrocephalus) carcass in northern Svalbard (J.O. Scheie, personal communication). In these 2 observations, up to 14 and 17 bears, respectively, were observed on the carcasses, suggesting that scavenging is important for many individuals. Further, observations of predation and scavenging of reindeer (Rangifer tarandus platyrhynchus) (Derocher et al. 2000) attest to the diversity of diet. Prey composition is an important element for understanding the ecotoxicology of polar bears. In Svalbard, there was speculation that harp seals were responsible for the high levels of polychlorinated biphenyls in polar bears (Kleivane et al. 2000; Gabrielsen and Henriksen 2001). However, our results make it clear that more detailed study of polar bear diet is required to understand trophic transfer of pollutants. In particular, mother-offspring transfer of pollutants can result in nursing young having higher pollution loads than the mother (Tanabe and Tatsukawa 1991; Polischuk et al. 1995; Beckmen et al. 1999). Given the large number of ringed and bearded seal pups consumed by polar bears, it is important that the pollution load of seal pups be studied. However, careful quantification of the season-, sex- and age-specific diet of polar bears is required before trophic-level transfer of pollution can be understood. Similarly, if climate change alters the distribution and abundance of prey (Stirling and Derocher 1993), better documentation of current predation patterns is essential for understanding the effects of climate change on polar bears. In summary, similar to other areas, the diet of polar bears in Svalbard and the western Barents Sea is dominated by ringed seals on a numerical basis, but bearded seals make a significant contribution to the diet when biomass is considered. Harp seals likely play an important, but lesser, role in the diet of bears living in more pelagic habitats, but only during the summer months. Acknowledgements This research was funded by the Norwegian Polar Institute and the Norwegian Research Council. We are grateful for reports of seal kills from others working in the Svalbard area, particularly Georg Bangjord, Wayne Lynch and Christian Lydersen. Jon Ove Scheie, Environmental Officer, Governor of Svalbard, kindly provided the reports of polar bear feeding on white and sperm whales.
References Andersen M, Hjelset AM, Gjertz I, Lydersen C, Gulliksen B (1999) Growth, age at sexual maturity and condition in bearded seals (Erignathus barbatus) from Svalbard, Norway. Polar Biol 21:179–186 Beckmen KB, Ylitalo GM, Towell RG, Krahn MM, O’Hara TM, Blake JE (1999) Factors affecting organochlorine contaminant concentrations in milk and blood of northern fur seal (Callorhinus ursinus) dams and pups from St. George Island, Alaska. Sci Total Environ 231:183–200
Benjaminsen T (1973) Age determination and the growth and age distribution from cementum growth layers of bearded seals at Svalbard. Fiskeridir Skr Ser Havunders 16:159–170 Burns JJ (1981) Bearded seal Erignathus barbatus Erxleben, 1777. In: Ridgway SH, Harrison RJ (eds) Handbook of marine mammals, vol. 2. Seals. Academic Press, London, pp 145–170 Calvert W, Stirling I (1990) Interactions between polar bears and overwintering walruses in the Central Canadian High Arctic. Int Conf Bear Biol Manage 8:351–356 Christiansen BO (1981) Isbjørntreff sydvest for Kvitøya pa˚ Spitsbergen. Fauna 34:129–130 Derocher AE, Wiig Ø, Bangjord G (2000) Predation of Svalbard reindeer by polar bears. Polar Biol 23:675–678 Furgal CM, Innes S, Kovacs KM (1996) Characteristics of ringed seal, Phoca hispida, subnivean structures and breeding habitat and their effects on predation. Can J Zool 74:858–874 Gabrielsen GW, Henriksen EO (2001) Persistent organic pollutants in Arctic animals in the Barents Sea area and at Svalbard: levels and effects. Mem Natl Inst Polar Res 54:349–364 Gjertz I (1991) Distribution of hooded seals in Svalbard waters. Fauna Norv Ser A 12:19–24 Gjertz I, Lydersen C (1986) Polar bear predation on ringed seals in the fast-ice of Hornsund, Svalbard. Polar Res 4:65–68 Gjertz I, Wiig Ø (1995) The number of walruses (Odobenus rosmarus) in Svalbard in summer. Polar Biol 15:527–530 Gjertz I, Kovacs KM, Lydersen C, Wiig Ø (2000) Movements and diving of bearded seal (Erignathus barbatus) mothers and pups during lactation and post-weaning. Polar Biol 23:559–566 Gjertz I, Lydersen C, Wiig Ø (2001) Distribution and diving of harbour seals (Phoca vitulina) in Svalbard. Polar Biol 24:209– 214 Hammill MO, Smith TG (1989) Factors affecting the distribution and abundance of ringed seal structures in Barrow Strait, Northwest Territories. Can J Zool 67:2212–2219 Hammill MO, Smith TG (1991) The role of predation in the ecology of the ringed seal in Barrow Strait, Northwest Territories, Canada. Mar Mammal Sci 7:123–135 Haug T, Nilssen KT, Øien N, Potelov V (1994) Seasonal distribution of harp seals (Phoca hispida) in the Barents Sea. Polar Res 13:163–172 Innes S, Stewart REA, Lavigne DM (1981) Growth in northwest Atlantic harp seals Phoca groenlandica. J Zool Lond 194:11–24 Iverson SJ, Stirling I, Lang S (1999) Using blubber fatty acids for ecological insight: the example of foraging behavior of polar bears. 13th Biennial Conference on the Biology of Marine Mammals Kleivane L, Severinsen T, Skaare JU (2000) Biological transport and mammal to mammal transfer for organochlorines in Arctic fauna. Mar Environ Res 49:343–357 Lønø O (1970) The polar bear (Ursus maritimus Phipps) in the Svalbard area. Nor Polarinst Skr 149:1–115 Lowry LF, Burns JJ, Nelson RR (1987) Polar bear, Ursus maritimus, predation on belugas, Delphinapterus leucas, in the Bering and Chukchi seas. Can Field Nat 101:141–146 Lydersen C, Gjertz I (1986) Studies of the ringed seal (Phoca hispida Schreber 1775) in its breeding habitat in Kongsfjorden, Svalbard. Polar Res 4:57–63 Lydersen C, Gjertz I (1987) Population parameters of ringed seals (Phoca hispida Schreber, 1775) in the Svalbard area. Can J Zool 65:1021–1027 Lydersen C, Hammill MO, Ryg MS (1992) Water flux and mass gain during lactation in free-living ringed seal (Phoca hispida) pups. J Zool Lond 228:361–369 Lydersen C, Hammill MO, Kovacs KM (1994) Diving activity in nursing bearded seal (Erignathus barbatus) pups. Can J Zool 72:96–103 Mauritzen M, Derocher AE, Wiig Ø (2001) Space-use strategies of female polar bears in a dynamic sea ice habitat. Can J Zool 79:1704–1713 Mauritzen M, Derocher AE, Wiig Ø, Belikov SE, Boltunov AN, Hansen E, Garner GW (2002) Using satellite telemetry to
452 define spatial population structure in polar bears in the Norwegian and western Russian Arctic. J Appl Ecol 39:79–90 Nilssen KT, Pedersen OP, Folkow LP, Haug T (2000) Food consumption estimates of Barents Sea harp seals. NAMMCO Sci Publ 2:9–27 Nordøy ES, Folkow LP, Potelov V, Prichtchemikhine V, Blix AS (1998) Migratory patterns and dive behaviour of Barents Sea harp seals. The World Marine Mammal Science Conference, Monaco, 20–24 January 1998 Parovshchikov VY (1964) A study on the population of polar bear, Ursus (Thalarctos) maritimus Phipps, of Franz Joseph Land. Acta Soc Zool Bohemoslov 28:167–177 Polischuk SC, Letcher RJ, Norstrom RJ, Ramsay MA (1995) Preliminary results of fasting on the kinetics of organochlorides in polar bears (Ursus maritimus). Sci Total Environ 160/ 161:465–472 Reeves RR (1998) Distribution, abundance and biology of ringed seals (Phoca hispida): an overview. NAMMCO Sci Publ 1:9–45 Smith TG (1980) Polar bear predation of ringed and bearded seals in the land-fast sea ice habitat. Can J Zool 58:2201–2209 Smith TG, Lydersen C (1991) Availability of suitable land-fast ice and predation as factors limiting ringed seal populations, Phoca hispida, in Svalbard. Polar Res 10:585–594
Smith TG, Sjare B (1990) Predation of belugas and narwhals by polar bears in nearshore areas of the Canadian High Arctic. Arctic 43:99–102 Stempniewicz L (1993) The polar bear Ursus maritimus feeding in a seabird colony in Frans Josef Land. Polar Res 12:33–36 Stirling I, Archibald WR (1977) Aspects of predation of seals by polar bears. J Fish Res Board Can 34:1126–1129 Stirling I, Derocher AE (1993) Possible impacts of climatic warming on polar bears. Arctic 46:240–245 Stirling I, Øritsland NA (1995) Relationships between estimates of ringed seal (Phoca hispida) and polar bear (Ursus maritimus) populations in the Canadian Arctic. Can J Fish Aquat Sci 52:2594–2612 Tanabe S, Tatsukawa R (1991) Persistent organochlorines in marine mammals. In: Jones KC (ed) Organic contaminants in the environment – environmental pathways and effects. Elsevier Applied Science, Barking Wiig Ø, Derocher AE, Belikov SE (1999) Ringed seal (Phoca hispida) breeding in the drifting pack ice of the Barents Sea. Mar Mammal Sci 15:595–598
Paper 8
Photo: Magnus Andersen
Polar Biol (2001) 24: 231±238 DOI 10.1007/s003000000201
O R I GI N A L P A P E R
M. Andersen á E. Lie á A. E. Derocher á S. E. Belikov A. Bernhoft á A. N. Boltunov á G. W. Garner J. U. Skaare á é. Wiig
Geographic variation of PCB congeners in polar bears (Ursus maritimus) from Svalbard east to the Chukchi Sea Accepted: 2 October 2000 / Published online: 14 December 2000 Ó Springer-Verlag 2000
Abstract We present data on geographic variation in polychlorinated biphenyl (PCB) congeners in adult female polar bears (Ursus maritimus) from Svalbard eastward to the Chukchi Sea. Blood samples from 90 free-living polar bears were collected in 1987±1995. Six PCB congeners, penta to octa chlorinated (PCB-99, -118, -153, -156, -180, -194), were selected for this study. Dierences between areas were found in PCB levels and congener patterns. Bears from Franz Josef Land (11,194 ng/g lipid weight) and the Kara Sea (9,412 ng/g lw) had similar SPCB levels and were higher than all other populations (Svalbard 5,043 ng/g lw, East Siberian Sea 3,564 ng/g lw, Chukchi Sea 2,465 ng/g lw). Svalbard PCB levels were higher than those from the Chukchi Sea. Our results, combined with earlier ®ndings, indicate that polar bears from Franz Josef Land and the Kara Sea have the highest PCB levels in the Arctic. Decreasing trends were seen eastwards and westwards from this region. Of the congeners investigated in the present study, the lower chlorinated PCBs are increasing and the M. Andersen (&) á A. E. Derocher Norwegian Polar Institute, 9296 Tromsù, Norway e-mail:
[email protected] Tel.: +47-77750534 Fax: +47-77750501 E. Lie á J. U. Skaare á A. Bernhoft National Veterinary Institute, P.O. Box 8156 Dep., 0033 Oslo, Norway S. E. Belikov á A. N. Boltunov All-Russian Research Institute for Nature Conservation, Znamenskoye-Sadki, Moscow 113628, Russia G. W. Garner Alaska Science Center, U.S. Geological Survey, Biological Resources Division, 1011 E. Tudor Road, Anchorage, Alaska 99503, USA J. U. Skaare The Norwegian School of Veterinary Science, P.O. Box 8146 Dep., 0033 Oslo, Norway é. Wiig Zoological Museum, University of Oslo, Sars gt. 1, 0562 Oslo, Norway
high chlorinated PCBs are decreasing from Svalbard eastward to the Chukchi Sea. Dierent pollution sources, compound transport patterns and regional prey dierences could explain the variation in PCB congener levels and patterns between regions.
Introduction During the last decades, a wide range of man-made environmental pollutants have been transported by air and ocean currents from southern industrialised areas to the Arctic (Oehme 1991; Barrie et al. 1992; De March et al. 1998). One major contamination group is the organochlorine (OC) compounds, which are highly lipophilic and resistant to biological degradation. They accumulate in the marine environment and biomagnify up the food chain (Muir et al. 1988; Barrie et al. 1992). The Arctic ecosystem is particularly sensitive to impacts of pollution because it is characterised by low annual productivity, few species and short food chains (Dunbar 1986), creating few opportunities for biodegradation (De March et al. 1998). Further, Arctic organisms are adapted to dealing with short periods of high production during which lipid energy stores are built, resulting in high dependence on fat at most trophic levels (Barrie et al. 1992). Polychlorinated biphenyl (PCB) compounds are one OC group that has been associated with detrimental eects on animal physiology (Peterson et al. 1993; Safe 1994; Skaare et al. 2000). PCBs are chemically stable compounds used in industrial products such as transformers, capacitors and hydraulic oils. Commercial use of PCBs started in 1929, and open use of these compounds is currently banned in all circumpolar countries, except in Russia, where a limited use and manufacture still exist. In other Arctic regions, a signi®cant amount is still found in closed systems (De March et al. 1998). PCBs were ®rst identi®ed in polar bears in the 1970s (Bowes and Jonkel 1975). Svalbard area polar bears
232
have PCB levels comparable to those found in ringed seals (Phoca hispida) from the Baltic, where reproductive disorders were reported (Norheim et al. 1992; Olsson et al. 1992; Bernhoft et al. 1997). In polar bears at Svalbard, a possible immunotoxic eect (Bernhoft et al. 2000) and negative association between OCs and retinol and thyroid hormones have been reported (Skaare et al. 2000). Contamination in polar bears has been studied in most parts of its range. However, limited data from most parts of the Russian Arctic have precluded understanding the circumpolar PCB levels. A study of OCs in polar bears from eastern Russia, North America, Greenland and Svalbard showed geographic trends in levels of several compounds (Norstrom et al. 1998). Norstrom et al. (1998) concluded that SPCB concentrations were higher in bears from the Arctic Ocean, East Greenland and Svalbard compared to the other regions in their study. They also found that concentrations of other chlorinated hydrocarbon contaminants were higher in the eastern parts than the western parts of their study area. Lechter et al. (1995) found a decreasing trend in the ratio PCB-99:PCB-180 from west to east in the western Arctic hemisphere, which demonstrated an increasing proportion of higher chlorinated PCB congeners in the east. Polar bears are distributed throughout the Arctic region, with a world population size of 22,000±27,000 animals in 19 sub-populations (IUCN/SSC Polar Bear Fig. 1 Capture locations (black circles) of the 90 adult female polar bears included in this study. Locations mentioned in the text are shown on the map
Specialist Group 1998). The polar bear is the principal predator in the Arctic marine ecosystem, mainly feeding on ringed seals, but also on bearded seals (Erignathus barbatus) (Stirling and Archibald 1977; Smith 1980). Because of its key role in the Arctic ecosystem and its wide distribution, polar bears are a good indicator in the study and monitoring of environmental contamination (Norstrom and Muir 1994; De March et al. 1998; Norstrom et al. 1998). Bears from dierent populations can provide a circumpolar perspective on contamination status. The objective of our study was to investigate the levels and geographic pattern of PCBs in polar bears from the European, Russian and parts of the American Arctic, and to discuss possible explanations for the variation.
Materials and methods Blood sampling from live captured adult female polar bears was conducted from late March to mid-May in 1987±1995 in the Svalbard area (78°N, 20°E) eastward to the Chukchi Sea (70°N, 170°W) (Fig. 1, Table 1). Polar bears (³1 year of age) were captured by remote injection of a drug-®lled dart (Palmer Cap-Chur Equipment, Douglasville, Ga.) ®red from a helicopter (Stirling et al. 1989). The drug Zoletil was administered in a solution of 200 mg/ ml at a dosage of 5±10 mg/kg of body mass. A vestigial premolar tooth was extracted from all bears for age determination, using the methods of Calvert and Ramsay (1998). Blood samples were
233 Table 1 Geographic regions where adult female polar bears were captured, years in which the bears were captured, number of bears from each region and ages of captured bears (mean and minimum/maximum ages)
Region
n
Sampling years
Mean age (years)
Age range
Svalbard Franz Josef Land Kara Sea Siberian Sea Chukchi Sea
32 17 12 8 21
1991±1994 1995 1991±1994 1992±1993 1987±1992
9 12 13 11 10
5±14 5±22 5±28 6±23 6±18
collected from the femoral vein into evacuated containers. From the bears captured in the Russian Arctic, whole-blood samples were collected and stored cool until frozen (±20 °C). Samples from Svalbard were stored cool until centrifuged within 8 h of collection, after which plasma was pipetted o and stored at )20 °C until analysis. The Norwegian Experimental Animal Committee approved all capture and handling methods used when capturing polar bears in Norway and similar protocols were used in Russia and Alaska. PCB analyses were performed at the Environmental Toxicology Laboratory at the Norwegian School of Veterinary Science. Plasma and blood cell samples from Svalbard polar bears were analysed in 1993 and 1994. For details on methods used and quality assurance and control for these samples, see Bernhoft et al. (1997). In 1996 and 1997, PCBs in whole-blood samples from polar bears captured in Franz Josef Land, Kara, East Siberian and Chukchi Seas were quanti®ed using the same methods but with some modi®cations. Samples of whole blood (ca. 8 g) were weighed, added internal standards (PCB-29, -112, -207), and extracted with cyclohexane and acetone. Extractable fat % was determined gravimetrically using conical, graduated centrifuge tubes and a gentle stream of nitrogen (N2). Thereafter, the extracted fat was redissolved and washed with ultra-pure sulphuric acid to remove the fat. The extracts were transferred to conical, graduated centrifuge tubes and were evaporated to a 0.5-ml ®nal volume under a gentle stream of N2. The extracts were automatically injected (Fisons Autosampler AS 800) on Carlo Erba, HRGC 53 Mega Series (Carlo Erba Instrumentation, Milan, Italy) gas chromatograph, equipped with a split/splitless injector and an electron capture Ni63 detector (ECD). Two columns of dierent polarity and selectivity were used to obtain the desired chromatographic separation (SPB-5 and SPB1701, 60 m, 0.25 mm ID, 0.25 lm ®lm layer, Supleco, Bellefonte, Pa.), both connected to a 1-m deactivated pre-column. Hydrogen was used as the carrier gas (linear velocity 40 cm/s) and 5% Ar/ CH4 was the make-up gas at a ¯ow of 30 ml min±1. The splitless time was 2 min and the injector temperature was 270 °C. The temperature programme was as follows: 90 °C (hold 2 min); 90± 180 °C (25 °C min±1); 180 °C (hold 2 min); 180±220 °C (1.5 °C min±1); 220 °C (hold 2 min); 220±275 °C (3 °C min±1); 275 °C (hold 15 min). Total run time was 70 min. The detector temperature was 310 °C (detector-base 270 °C). Six PCB congeners were quanti®ed in all samples (IUPAC nos.): 99, 118, 153, 156, 180, 194. In 30 samples from dierent regions, 28 additional PCB congeners were determined (IUPAC nos. 28, 31, 47, 52, 56, 66, 74, 87, 101, 105, 110, 114, 128, 136, 137, 138, 141, 149, 151, 157, 170, 183, 187, 189, 196, 199, 206, 209). In these 30 samples, the 6 congeners analysed in all samples constituted an average of 78% of the 28 PCB congeners, with a coecient of variation (CV) of 3.3%. The selection of the six PCB congeners was based on high correlation between concentrations of the different congeners found in polar bears (Bernhoft et al. 1997) and on the congener structure. The analytical quality of the laboratory has been approved in several intercalibration tests containing components and matrices relevant to our work: four steps of the ICES/IOC/OSPARCOM test (Anonymous 1995; De Boer et al. 1996), tests organised by WHO/UNEP in 1982, 1992 and 1996, IUPAC inter-calibration test in 1989±1993 and an intercomparison series organised by the National Reference Laboratories in the EU starting in 1995. The laboratory is accredited for these analyses according to the requirements of NS-EN45001 (1989) and ISO/IEC Guide 25 (1990).
Standard procedures were used to ensure adequate quality assurance and control, and the results were within the laboratory's accredited requirements. Certi®ed international reference materials (CRM 349 and 350, ICES cod liver oil and mackerel oil) were analysed regularly with results within the given ranges. Quanti®cation was carried out within the linear range of the detector. Three to four level calibration curves were used for the calculations. Blank samples were included in each series to test for interference. Detection limits for individual compounds were determined as 3 times the noise level. Detection limits for individual PCB congeners were between 0.002 and 0.010 ng g±1 wet weight. Quanti®cation was performed using PCB-29, -112 and -207 as internal standards in each sample. Recoveries and CV of individual PCB congeners in spiked sheep blood varied from 79 to 115% and 3.7 to 14.2, respectively. Reproducibility was continuously tested by analysing the laboratory's own reference sample (seal blubber). The results were within the given ranges, with CV of 8.2%. To compare PCB levels in polar bears from Svalbard with bears captured further east, PCB levels in plasma and blood cells were transformed to whole-blood levels. Similar levels and good correlation (r=0.59, P
