• Adaptive management;
  • frailty;
  • growth rate;
  • harvest;
  • waterfowl


We examined individual heterogeneity in survival and recruitment of female Pacific black brant (Branta bernicla nigricans) using frailty models adapted to a capture–mark–recapture context. Our main objectives were (1) to quantify levels of heterogeneity and examine factors affecting heterogeneity, and (2) model the effects of individual heterogeneity on harvest dynamics through matrix models. We used 24 years of data on brant marked and recaptured at the Tutakoke River colony, AK. Multievent models were fit as hidden Markov chain using program E-SURGE with an adequate overdispersion coefficient. Annual survival of individuals marked as goslings was heterogeneous among individuals and year specific with about 0.23 difference in survival between “high” (0.73)- and “low” (0.50)-quality individuals at average survival probability. Adult survival (0.85 ± 0.004) was homogeneous and higher than survival of both groups of juveniles. The annual recruitment probability was heterogeneous for brant >1-year-old; 0.56 (±0.21) and 0.31 (±0.03) for high- and low-quality individuals, respectively. Assuming equal clutch sizes for high- and low-quality individuals and that 80% of offspring were in the same quality class as the breeding female resulted in reproductive values about twice as high for high-quality individuals than low-quality individual for a given class of individuals producing differential contributions to population growth among groups. Differences in reproductive values greatly increased when we assumed high-quality individuals had larger clutch sizes. When we assumed that 50% of offspring were in the same quality class as their mothers and clutches were equal, differences in reproductive values between quality classes were greatly reduced or eliminated (breeders [BRs]). We considered several harvest scenarios using the assumption that 80% of offspring were in the same quality class as their mothers. The amount of compensation for harvest mortality declined as the proportion of high-quality individuals in the harvest increased, as differences in clutch sizes between groups decreased and as the proportion of BRs in the harvest increased. Synthesis and applications. Harvest at the same proportional level of the overall population can result in variable responses in population growth rate when heterogeneity is present in a population. λ was <1.0 under every scenario when harvest rates were >10%, and heterogeneity caused as much as +2% difference in growth rates at the highest levels of proportional harvest for low-quality individuals and the greatest differences in qualities between classes of individuals, a critical difference for a population with λ near 1.0 such as the brant. We observed less response in overall survival in the presence of heterogeneity because we did not observe heterogeneity in the annual survival of BRs. This analysis provides a comprehensive view of overall compensation at the population level and also constitutes the first example of a survival-recruitment model with heterogeneity. Individual heterogeneity should be more explicitly considered in harvest management of vertebrates.