Integrated Population Modeling Provides the First Empirical Estimates of Vital Rates and Abundance for Polar Bears in the Chukchi Sea

Eric V. Regehr,Karyn D. Rode,Ryan R. Wilson,Michelle St. Martin,Nathan J. Hostetter,Sarah J. Converse

doi:10.1038/s41598-018-34824-7

Abstract

Large carnivores are imperiled globally, and characteristics making them vulnerable to extinction (e.g., low densities and expansive ranges) also make it difficult to estimate demographic parameters needed for management. Here we develop an integrated population model to analyze capture-recapture, radiotelemetry, and count data for the Chukchi Sea subpopulation of polar bears (Ursus maritimus), 2008–2016. Our model addressed several challenges in capture-recapture studies for polar bears by including a multievent structure reflecting location and life history states, while accommodating state uncertainty. Female breeding probability was 0.83 (95% credible interval [CRI] = 0.71–0.90), with litter sizes of 2.18 (95% CRI = 1.71–2.82) for age-zero and 1.61 (95% CRI = 1.46–1.80) for age-one cubs. Total adult survival was 0.90 (95% CRI = 0.86–0.92) for females and 0.89 (95% CRI = 0.83–0.93) for males. Spring on-ice densities west of Alaska were 0.0030 bears/km2 (95% CRI = 0.0016–0.0060), similar to 1980s-era density estimates although methodological differences complicate comparison. Abundance of the Chukchi Sea subpopulation, derived by extrapolating density from the study area using a spatially-explicit habitat metric, was 2,937 bears (95% CRI = 1,552–5,944). Our findings are consistent with other lines of evidence suggesting the Chukchi Sea subpopulation has been productive in recent years, although it is uncertain how long this will continue given sea-ice loss due to climate change.

Highlights

ObjectivesOur objectives were to (i) estimate abundance for the Chukchi Sea (CS) subpopulation using data collected in a geographically-limited study area; (ii) estimate vital rates by concurrently analyzing multiple types of data, including parameters that would be unobservable or confounded if data types were analyzed separately; (iii) reduce bias in demographic parameters by modeling the movement of animals in to and out of the study area; and (iv) develop a flexible modeling framework that addresses key analytical challenges for CS polar bears, and can be adapted to other subpopulations or species
Most captured females age ≥4 years were fitted with satellite collars (2008–2010; Telonics, Mesa, AZ, USA) or Global Positioning System (GPS) collars (2011–2016; Telonics)
A total of 166 unique males and 135 unique females were physically captured and released in 2008–2011, 2013, and 2015–2016. These numbers do not reflect captures of C0s and C1s, which were not included as individuals in the capture-recapture model

Summary

Objectives

Our objectives were to (i) estimate abundance for the CS subpopulation using data collected in a geographically-limited study area; (ii) estimate vital rates by concurrently analyzing multiple types of data, including parameters that would be unobservable or confounded if data types were analyzed separately; (iii) reduce bias in demographic parameters by modeling the movement of animals in to and out of the study area; and (iv) develop a flexible modeling framework that addresses key analytical challenges for CS polar bears, and can be adapted to other subpopulations or species

Methods

Results

Discussion

Conclusion