Conservation status of polar bears (Ursus maritimus) in relation to projected sea-ice declines
Abstract
Loss of Arctic sea ice owing to climate change is the primary threat to polar bears throughout their range. We evaluated the potential response of polar bears to sea-ice declines by (i) calculating generation length (GL) for the species, which determines the timeframe for conservation assessments; (ii) developing a standardized sea-ice metric representing important habitat; and (iii) using statistical models and computer simulation to project changes in the global population under three approaches relating polar bear abundance to sea ice. Mean GL was 11.5 years. Ice-covered days declined in all subpopulation areas during 1979–2014 (median −1.26 days year−1). The estimated probabilities that reductions in the mean global population size of polar bears will be greater than 30%, 50% and 80% over three generations (35–41 years) were 0.71 (range 0.20–0.95), 0.07 (range 0–0.35) and less than 0.01 (range 0–0.02), respectively. According to IUCN Red List reduction thresholds, which provide a common measure of extinction risk across taxa, these results are consistent with listing the species as vulnerable. Our findings support the potential for large declines in polar bear numbers owing to sea-ice loss, and highlight near-term uncertainty in statistical projections as well as the sensitivity of projections to different plausible assumptions.
1. Introduction
Polar bears (Ursus maritimus) depend on sea ice for most aspects of their life history [1]. Anthropogenic climate change is the primary threat to the species because, over the long term, global temperatures will increase and Arctic sea ice will decrease as long as atmospheric greenhouse gas concentrations continue to rise [2,3]. The global population of approximately 26 000 polar bears [4] is divided into 19 subpopulations, which are grouped into four ecoregions reflecting sea-ice dynamics and polar bear life history (figure 1; [5]). The subpopulations currently exhibit variable status relative to climate change [6]. Two have already experienced sea-ice-related demographic declines [7,8]. Others show signs of nutritional stress [9], have been reported as stable or productive [10] or have unknown status owing to deficient data [11].
Methods to forecast the effects of continued sea-ice loss on polar bears have included structured elicitation of expert opinion [12], Bayesian network models evaluating cumulative stressors [5], demographic projections for individual subpopulations [8] and mechanistic models linking vital rates to environmental factors [13]. To date, there has been no global assessment of polar bear abundance data relative to sea ice. We explored future changes in mean global population size (MGPS) for the species, using population projections under three approaches. Approach 1 reflected the hypothesis that environmental carrying capacity (K) is directly proportional to the availability of sea ice. Approaches 2 and 3 estimated relationships between changes in sea ice and observed changes in polar bear abundance. We evaluated projection outcomes, over three polar bear generations, relative to thresholds for threatened categories under criterion A3 of the IUCN Red List of Threatened Species (hereafter Red List; [4,14]). The scientific basis of Red List categories is discussed by Mace et al. [15].
2. Methods
Projection timeframes can incorporate biological differences across species by referencing to generation length (GL, the average age of parents of the current cohort; [14]). We estimated GL as the mean age of adult female polar bears with new cubs based on live-capture data from 11 subpopulations. Females with 1 year-old cubs in year t + 1 were counted as pseudo-observations in year t. Variation in GL was evaluated, using a bootstrap procedure (electronic supplementary material, table S1).
Satellite data of sea-ice concentration were collected between years 1979–2014 to develop an index of K for polar bears (figure 2; [16]). Within each of the 19 subpopulation areas, daily sea-ice area was calculated by summing the product of ice concentration and grid cell area over all 25 × 25 km grid cells with concentration more than 15%. We then determined the midpoint between summer-minimum and winter-maximum ice areas, and calculated the metric ice as the number of days per year that ice area was above the midpoint (i.e. the number of ‘ice-covered’ days). Mean values of ice were projected forward, using linear models, which facilitated projections at the spatial scale of polar bear subpopulations (electronic supplementary material, table S2).
We used population projections to evaluate changes in MGPS between the years 2015 and (2015 + 3 × GL) based on three approaches relating ice to estimates of subpopulation abundance (N; electronic supplementary material, table S3). Approach 1 assumed a one-to-one proportional relationship between ice and N. Approaches 2 and 3 estimated linear relationships between ice and proportional changes in N, and used these relationships to predict future values of N as a function of projected ice. Approach 2 estimated a global ice–N relationship based on a maximum of two estimates of N per subpopulation, separated by at least 10 years, which were available for seven subpopulations. Approach 3 estimated a separate ice–N relationship for each polar bear ecoregion using a dataset that was similar to approach 2 but included longer time series of N available for four subpopulations. All approaches assumed that changes in N were mediated primarily through changes in K or density-independent habitat effects, and that the ratio N/K was stable relative to other factors [17]. These assumptions were established on the basis that polar bears depend fundamentally on sea ice, that sea-ice changes represent the main source of habitat modification for the species [5], and that other potential stressors are either secondary (e.g. contaminants; [5]) or have been managed (e.g. harvest; [6]) for most subpopulations in recent decades. Projected subpopulation-specific changes in N were scaled to changes in MGPS, using the most recent estimate of N for each subpopulation. Based on 62 500 stochastic projections, we calculated the most likely change in MGPS over three generations. In addition, the probabilities of exceeding 0%, 30%, 50% and 80% reduction thresholds were generated following Red List guidelines [14]. We performed computations in R [18], using the package ‘arm’ [19] to simulate uncertainty in model coefficients. Data and projection methods are described fully in the electronic supplementary material.
3. Results and discussion
The mean subpopulation-specific estimate of GL was 11.5 years (approx. 5th and 95th percentiles = 9.8 and 13.6, respectively) based on 3374 observed reproductive events (electronic supplementary material, table S1). Projections were performed using GL = 11.5 and 13.6 years to reflect variation in GL and approximate natural GL. We did not apply the lower fifth percentile, because harvest likely shortened several empirical estimates of GL [14]. The metric ice declined at a significance level of 0.05 in all 19 subpopulation areas during 1979–2014 (figure 2, median −1.26 days year−1 [95% CI = −3.37 to −0.71]; electronic supplementary material, table S2).
We simulated per cent change in MGPS for six scenarios representing two values of GL and three approaches relating ice and N (table 1). Using GL = 11.5 years, the most likely values for per cent change in MGPS over three generations were −30%, −4% and −43% for approaches 1, 2 and 3, respectively. Across scenarios, the estimated median probabilities of reductions greater than 30%, 50% and 80% in MGPS were 0.71 (range 0.20–0.95), 0.07 (range 0–0.35) and less than 0.01 (range 0–0.02), respectively.
approach for projectionsa | duration of three polar bear generations (years) | per cent change in mean global population size
|
probability of decline
|
|||||
---|---|---|---|---|---|---|---|---|
median | lower 95%CI | upper 95%CI | ≥0% | ≥30% | ≥50% | ≥80% | ||
1 | 35 | −30 | −35 | −25 | 1.00 | 0.56 | 0.00 | 0.00 |
1 | 41 | −34 | −40 | −29 | 1.00 | 0.95 | 0.00 | 0.00 |
2 | 35 | −4 | −62 | 50 | 0.55 | 0.20 | 0.06 | 0.00 |
2 | 41 | −4 | −68 | 56 | 0.55 | 0.24 | 0.08 | 0.00 |
3 | 35 | −43 | −76 | −20 | 1.00 | 0.86 | 0.30 | 0.01 |
3 | 41 | −45 | −79 | −21 | 1.00 | 0.88 | 0.35 | 0.02 |
Our analyses highlight the potential for large reductions in MGPS as climate change and sea-ice loss continue [20] over the next three polar bear generations. Approach 1 was based only on projected changes in habitat, a common method when population data are lacking [14]. Approach 2 estimated a global ice–N relationship that was near 0 and not statistically significant (estimated slope coefficient [β] < 0.001, s.e. = 0.005; electronic supplementary material, table S4). This finding reflects variability in current subpopulation status, uncertainty in estimates of N and the lack of empirical evidence for sea-ice mediated changes in global abundance over recent decades [6]. Approach 3 estimated a separate ice–N relationship for each ecoregion. Relationships were positive at a significance level of 0.01 for the seasonal (β = 0.013, s.e. = 0.002) and divergent ecoregions (β = 0.032, s.e. = 0.009), reflecting observed correlations between declining sea ice and declining abundance (electronic supplementary material, table S4). Relationships were not significant for the convergent (β = −0.008, s.e. = 0.009) and archipelago ecoregions (β = −0.029, s.e. = 0.030). Although approach 3 reflected regional variability in sea-ice dynamics and polar bear ecology, it was strongly influenced by several well-studied subpopulations and did not reflect finer-scale variation. For example, within the divergent ecoregion, multiple estimates of N were available for the declining Southern Beaufort sea subpopulation [7], but not for the Chukchi sea subpopulation, which inhabits a more biologically productive region and has exhibited high recruitment despite sea-ice loss [9].
Our projections (table 1) are broadly consistent with expert opinion [12] and Bayesian network model forecasts [5], although methodological differences preclude direct comparison (see electronic supplementary material). Following the Red List guidelines for risk tolerance ([14]; §3.2.3), the high probability of reductions more than 30% in MGPS, and low probability of reductions more than 50%, were consistent with a categorization of vulnerable (i.e. facing a high risk of extinction in the wild; [4]). Our use of statistical models required estimating few parameters, consistent with sparse data available for Arctic marine mammals [21], and propagated the effects of assumptions on model outcomes in a transparent manner. Future global population assessments could explore the use of hierarchical models [22], integrate data from multiple sources [23], model population processes (e.g. density-dependent interactions between harvest and habitat loss; [17]), consider cumulative effects on polar bear health [24] or consider nonlinear or spatial responses [25].
Data accessibility
The datasets supporting this article have been uploaded as part of the electronic supplementary material.
Authors' contributions
Ø.W. was the IUCN Red List Authority Coordinator for polar bears. All authors contributed to the conception, design, analysis and interpretation of this work. Data were provided by E.V.R., K.L.L., M.O., N.J.L. and T.C.A. H.S. and K.L.L. led ice analyses. E.V.R. led GL and population analyses. E.V.R. and K.L.L. drafted the manuscript, and all authors revised it for accuracy and intellectual content. All authors agree to be held accountable for the content and approve the final version of the manuscript.
Competing interests
We have no competing interests.
Funding
Authors' institutions provided in-kind support. K.L.L. and H.S. were supported by National Aeronautics and Space Administration grant np/ NNX13AN28G and NNX11A063G.
Acknowledgements
E. Holmes and J. Reynolds provided statistical advice. C. Rondinini reviewed an earlier version of this work. J. Aars, E. Born and M. Dyck provided data to estimate GL. Any use of trade, firm or product names is for descriptive purposes only and does not reflect endorsement by the US government. The findings and conclusions are those of the authors and do not necessarily represent the views of the US Fish and Wildlife Service.
References
- 1
Amstrup SC . 2003 Polar Bear (Ursus maritimus). In Mammals of North America: biology, management, and conservation (edsFeldhamer GA, Thompson BC, Chapman JA ), pp. 587–610, 2nd edn. Baltimore, MD: John Hopkins University Press. Google Scholar - 2
Schneider SH . 1989 The greenhouse effect – science and policy. Science 243, 771–781. (doi:10.1126/science.243.4892.771) Crossref, PubMed, Web of Science, Google Scholar - 3
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. (doi:10.1038/nature09653) Crossref, PubMed, Web of Science, Google Scholar - 4
Wiig Ø, Amstrup S, Atwood T, Laidre K, Lunn N, Obbard M, Regehr E, Thiemann G. 2015. Ursus maritimus. The IUCN Red List of threatened species 2015: e.T22823A14871490. (doi:10.2305/IUCN.UK.2015-4.RLTS.T22823A14871490.en) (accessed 06 January 2016). Google Scholar - 5
Atwood TC, Marcot BG, Douglas DC, Amstrup SC, Rode KD, Durner GM, Bromaghin JF . 2016. Forecasting the relative influence of environmental and anthropogenic stressors on polar bears. Ecosphere 7, e01370. (doi:10.1002/ecs2.1370) Crossref, Web of Science, Google Scholar - 6
Obbard ME, Thiemann GW, Peacock E, DeBruyn TD . 2010 Polar bears: Proc. the 15th Working Meeting of the IUCN/SSC Polar Bear Specialist Group, Copenhagen, Denmark, 29 June–3 July, 2009, p. vii + 235. Gland, Switzerland and Cambridge, UK: IUCN. Google Scholar - 7
Bromaghin JF 2015 Polar bear population dynamics in the southern Beaufort sea during a period of sea ice decline. Ecol. Appl. 25, 634–651. (doi:10.1890/14-1129.1) Crossref, PubMed, Web of Science, Google Scholar - 8
Lunn NJ, Servanty S, Regehr EV, Converse SJ, Richardson E, Stirling I . 2016. Demography of an apex predator at the edge of its range: impacts of changing sea ice on polar bears in Hudson Bay. Ecol. Appl. 26, 1302–1320. (doi:10.1890/15-1256) Crossref, PubMed, Web of Science, Google Scholar - 9
Obbard ME, Cattet MRL, Howe EJ, Middel KR, Newton EJ, Kolenosky GB, Abraham KF, Greenwood CJ . 2016. Trends in body condition in polar bears (Ursus maritimus) from the Southern Hudson Bay subpopulation in relation to changes in sea ice. Arctic Sci. 2, 15–32. (doi:10.1139/as-2015-0027) Crossref, Web of Science, Google Scholar - 10
Rode KD, Regehr EV, Douglas DC, Durner G, Derocher AE, Thiemann GW, Budge SM . 2014 Variation in the response of an Arctic top predator experiencing habitat loss: feeding and reproductive ecology of two polar bear populations. Glob. Change Biol. 20, 76–88. (doi:10.1111/gcb.12339) Crossref, PubMed, Web of Science, Google Scholar - 11
Polar Bear Specialist Group of the International Union for the Conservation of Nature. 2015 Summary of polar bear population status per 2014. pbsg.npolar.no/en/status/status-table.html (accessed 01 July 2015). Google Scholar
- 12
O'Neill SJ, Osborn TJ, Hulme M, Lorenzoni I, Watkinson AR . 2008 Using expert knowledge to assess uncertainties in future polar bear populations under climate change. J. Appl. Ecol. 45, 1649–1659. (doi:10.1111/j.1365-2664.2008.01552.x) Crossref, Web of Science, Google Scholar - 13
Molnar PK, Derocher AE, Thiemann GW, Lewis MA . 2010 Predicting survival, reproduction and abundance of polar bears under climate change. Biol. Conserv. 143, 1612–1622. (doi:10.1016/j.biocon.2010.04.004) Crossref, Web of Science, Google Scholar - 14
IUCN Standards and Petitions Subcommittee. 2016. Guidelines for using the IUCN Red List categories and criteria. Version 12. Prepared by the Standards and Petitions Subcommittee. See http://www.iucnredlist.org/documents/RedListGuidelines.pdf (accessed 03 October 2016). Google Scholar
- 15
Mace GM, Collar NJ, Gaston KJ, Hilton-Taylor C, Akçakaya HR, Leader-Williams N, Milner-Gulland EJ, Stuart SN . 2008 Quantification of extinction risk: IUCN's system for classifying threatened species. Conserv. Biol. 22, 1424–1442. (doi:10.1111/j.1523-1739.2008.01044.x) Crossref, PubMed, Web of Science, Google Scholar - 16
Stern HS, Laidre KL . 2016. Sea-ice indicators of polar bear habitat. Cryosphere 10, 2027–2041. (doi:10.5194/tc-10-2027-2016) Crossref, Web of Science, Google Scholar - 17
Regehr EV, Wilson RR, Rode KD, Runge MC . 2015 Resilience and risk—a demographic model to inform conservation planning for polar bears. U.S. Geological Survey Open-File Report 2015-1029, 56. Google Scholar - 18
R Core Team. 2015. R: a language and environment for statistical computing. Vienna, Austria. See http://www.R-project.org/ (accessed 05 May 2015). Google Scholar
- 19
Gelman A, Su Y-S. 2015. arm: Data Analysis Using Regression and Multilevel/Hierarchical Models. R package version 1.8-5. See http://CRAN.R-project.org/package=arm. Google Scholar - 20
Kirtman B et al. 2013 Near-term climate change: projections and predictability. In Climate change 2013: the physical science basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change (eds TF Stocker, D Qin, G-K Plattner, M Tignor, SK Allen, J Boschung, A Nauels, Y Xia, V Bex, PM. Midgley), pp. 953–1028. Cambridge, UK and New York, NY: Cambridge University Press. Google Scholar
- 21
Laidre K 2015. Arctic marine mammal population status, sea ice habitat loss, and conservation recommendations for the 21st century. Conserv. Biol. 29, 724–737. (doi:10.1111/cobi.12474) Crossref, PubMed, Web of Science, Google Scholar - 22
Hobbs NT, Hooten MB . 2015 Bayesian models: a statistical primer for ecologists. Princeton, NJ: Princeton University Press. Crossref, Google Scholar - 23
Rhodes JR, Chooi Fei N, de Villiers DL, Preece HJ, McAlpine CA, Possingham HP . 2011 Using integrated population modelling to quantify the implications of multiple threatening processes for a rapidly declining population. Biol. Conserv. 144, 1081–1088. (doi:10.1016/j.biocon.2010.12.027) Crossref, Web of Science, Google Scholar - 24
Patyk KA 2015. Establishing a definition of polar bear (Ursus maritimus) health: a guide to research and management activities. Sci. Total Environ. 514, 371–378. (doi:10.1016/j.scitotenv.2015.02.007) Crossref, PubMed, Web of Science, Google Scholar - 25
Derocher AE 2013 Rapid ecosystem change and polar bear conservation. Conserv. Lett. 6, 368–375. (doi:10.1111/conl.12009) Crossref, Web of Science, Google Scholar