Reliability of trans-generational genetic mark-recapture for enumerating Pacific salmon

Abstract

As Pacific salmon (Oncorhynchus spp.) continue to decline across much of their range, it is imperative to further develop reliable tools to quantify trends in population abundance. Estimating the number of spawning adult Pacific salmon (escapement) is a fundamental but challenging objective for fishery managers. Escapement monitoring of Pacific salmon typically requires intensive sampling across the return migration, often in remote locations, and is subject to changing environmental conditions that can reduce data quality. Alternative methods that increase the efficiency, accuracy, and precision of escapement monitoring are desirable, particularly during times of shrinking budgets for fisheries management. With the advent of high-throughput genotyping, biologists can combine molecular pedigree information with life-history data to make robust inferences about the size of wild populations using trans-generational genetic mark-recapture (tGMR). Minimally invasive genetic samples 'mark' adults and 'capture' juveniles, then multi-locus genotypes are used to identify parent-offspring pairs ('recaptures'), provided that these kinship relationships can be identified and that certain mark-recapture assumptions such as an equal probability of capture and a closed population are met. Our project aimed to help optimize the emerging tGMR methodology by comparing tGMR estimates to a traditional mark-recapture project for Chinook salmon (Oncorhynchus tshawytscha) from the Chilkat River in southeast Alaska. We further identified potential biases arising from violations of the equal probability of capture and closed population assumptions using an individual-based simulation model to evaluate the accuracy and precision of tGMR under varying demographic and sampling scenarios. We leveraged empirical values to parameterize simulations exploring the influence of age-specific reproductive success and sampling selectivity on tGMR estimates. Finally, we determined how adult sampling location and timing can impact tGMR estimates by comparing adult samples collected from the mainstem of the Chilkat River in June from samples collected in the upriver tributaries in August. Our assessment of the reliability of tGMR informs this method's potential for accurate and precise monitoring of populations.