Combining genetics with age/length data to estimate temporal changes in year-class strength of source populations contributing to mixtures

Abstract

We developed an approach for estimating changes in relative year-class strength of source populations contributing to mixtures by incorporating ages or lengths of mixture individuals in genetic stock identification models. The approach is intended for long-lived fishes with high pre-recruitment and low post-recruitment mortality rates for which consistent temporal changes in recruitment can be assumed. Age- and collection-year specific contributions of sources to the mixture are modeled as linear functions of two source-specific parameters: an intercept representing relative recruitment for the first modeled year class and a slope representing how relative recruitment changes annually. Based on simulations, the estimation approach performed reasonably well under diverse conditions, including varying numbers of sources, levels of genetic divergence among sources, degrees of change in year-class strength, durations and frequencies of sampling from mixtures, age ranges of individuals from mixtures, and sample sizes from mixtures. The estimation approach was also robust to aging error and uncertainty in length-age relationships. We applied the approach to genetic and length data for lake sturgeon Acipenser fulvescens from Green Bay, Lake Michigan, collected from 2002 to 2008, with sources corresponding to five Lake Michigan tributaries. Two of the Lake Michigan sources showed declining recruitment levels, whereas the other three sources showed increasing recruitment levels. We believe our proposed approach for indexing changes in year-class strength shows promise as a tool for identifying sources potentially at risk due to declining recruitment and for relating changes in recruitment to biotic or abiotic factors.