Modeling vital rates and age-sex structure of Pacific Arctic phocids: influence on aerial survey correction factors

Abstract

To estimate abundance, surveys of marine mammals often rely on samples of satellite-tagged individuals to correct counts for the proportion of animals that are unavailable to be detected. However, naïve application of this correction relies on the key assumption that availability of the tagged sample resembles that of the population. Here, we show how matrix population models can be used to estimate stable age- and stage-proportions, and how these can be used to adjust aerial survey correction factors so that they represent population-level availability. We illustrate this procedure using data from ice-associated seals in the Bering, Chukchi, and Beaufort Seas. After compiling life history data for bearded (Erignathus barbatus), ribbon (Histriophoca fasciata), ringed (Pusa hispida), and spotted seals (Phoca largha), we find that correction factors ignoring age-sex composition can positively bias spotted seal abundance by an average of 13% and negatively bias ribbon seal abundance by an average of 5%. Note that we did not examine potential bias for bearded or ringed seals due to low sample sizes; as such, we urge caution in interpretation of abundance estimates for these species.