Patterns and drivers of variation in Weddell seal survival rates, and population implications of unobserved individual heterogeneity in vital rates

Abstract

Variation in vital rates can result from many different factors, have consequences at the individual level, and may alter the structure of populations with implications for population dynamics. Here, different forms of variation in vital rates are investigated and the consequences to population dynamics are assessed using long-term longitudinal data from a population of Weddell seals (Leptonychotes weddellii) in Erebus Bay, Antarctica. I assessed how large unobserved individual heterogeneity in the probability of reproduction for female Weddell seals influenced population dynamics using an integral projection model populated with vital rate estimates from this population. I found that changes to the distribution of unobserved individual heterogeneity in reproduction led to small changes in the population growth rate and posit that these results are likely due to the canalization of life history traits. Given that survival is predicted to be an important driver of population dynamics, I evaluated sources of variation in survival for two different periods of life. I assessed patterns of age-specific survival for males and females to assess possible actuarial senescence patterns and to compare patterns between the sexes. To estimate survival rates, I used Cormack-Jolly-Seber models in the Bayesian framework allowing for unobserved heterogeneity with finite mixtures and applied flexible basis splines to model age-specific patterns. I found that males not only exhibited lower survival rates from age three years old to the end of life, but they also exhibited stronger senescence after the age of 12 years old than did females. Results for survival senescence were then used to assess support for different hypotheses for the evolution of sex-differences in survival senescence. Finally, I built upon previous research by evaluating the relationship of different variables to the probability that an individual survives early life and returns to the study area between the ages of two and six years old. I found that pup weaning mass was the most important to a pup's probability of being recaptured, but that sex differences may also exist. The results from this dissertation highlight the importance of life-history strategies to understanding demographic and population processes and provide avenues for future research.