Tracing carbon sources of southern Beaufort Sea and Chukchi Sea polar bears using stable isotope analyses

Abstract

The Southern Beaufort Sea (SBS) polar bear (Ursus maritimus) subpopulation has declined in response to sea ice loss, while the Chukchi Sea (CS) subpopulation appears stable. The substantial population decline in the SBS subpopulation in recent years is concurrent with increases in the proportion of polar bears coming on shore, and the duration they spend there. Both of these changes have been associated with the loss of access to their primary sea ice habitat, which is mainly used as a platform to hunt seals. The first objective of this study was to determine if the SBS and CS polar bear subpopulations could be distinguished based on stable isotope signatures (delta¹³C and delta¹⁵N values) of bone collagen. The second objective was to examine patterns in SBS polar bear trophic level and terrestrial carbon sources over a 65-year time period, as polar bears have increasingly used coastal environments. We analyzed 112 SBS and CS polar bear bones (predominantly mandibles) from 1954-2019 that had been archived at the University of Alaska Museum of the North, as well as bones from subsistence-harvested polar bears. In addition to delta¹³C and delta¹⁵N, samples from the SBS bears were analyzed for compound-specific stable carbon isotopes of amino acids (delta¹³CAA values). Another 50 bone collagen samples from terrestrial mammals and pinnipeds from northern Alaska were analyzed for delta¹³CAA values to provide a regional comparative dataset. Our study showed a significant difference in bulk delta¹³C (p<0.001) values, but not delta¹⁵N (p=0.654) values between the CS (-13.0‰±0.3‰ and 22.0‰±0.9‰, respectively) and the SBS bears (-14.7‰±1.3‰ and 22.2‰±1.0‰, respectively). We performed a logistic regression analysis (LR) using bulk delta¹³C and delta¹⁵N values of the polar bears to predict their placement into these two subpopulations. Using Icy Cape, AK as the geographical boundary, LR correctly placed polar bears in their respective subpopulations 82% of the time. Overall accuracy of placement changed to 84% when using the current geographical boundary at Utqiaġvik, AK. Bone collagen has a slow turnover rate, providing long-term, potentially life-long stable isotope signatures. Our findings could be used to determine the association of harvested polar bears to Alaska subpopulations, thus aiding in harvest quota management. The LR predicted samples collected from the Wainwright, AK region to be 58% CS and 42% SBS polar bears. This indicates that the area between Wainwright and Icy Cape is a polar bear mixing zone that includes bears from both subpopulations. Over the 65-year study period, two distinct groups of SBS polar bears were identified based on their delta¹³C values of the amino acid proline: a high delta¹³CPro group (1.8‰±2.3‰, n=45) and a low delta¹³CPro group (-15.7‰±1.9‰, n=26). The high proline polar bear group had delta¹³CPro values similar to those of Arctic brown bears (Ursus arctos; 0.4‰±1.6‰), while the low group had delta¹³CPro values similar to ice seals (-15.3‰±1.2‰). Among the available samples, there were more high proline/pelagic bears (n=17) after the 2007 sea ice minimum than high proline/coastal bears (n=12), which is opposite of what we expected. This study provides evidence that two distinct ecotypes in Southern Beaufort Sea polar bears, pelagic and coastal, have existed since at least the 1950's. Overall, our results represent a detailed isotopic view of the Alaskan polar bear subpopulations, demonstrating the possibility of distinguishing and categorizing individuals as either SBS or CS, while also highlighting the existence of two ecotypes in the SBS subpopulation.