Abstract
Molecular approaches to calculate effective population size estimates (Ne) are increasingly used as an alternative to long-term demographic monitoring of wildlife populations. However, the complex ecology of most long-lived species and the consequent uncertainties in model assumptions means that effective population size estimates are often imprecise. Although methods exist to incorporate age structure into Ne estimations for long-lived species with overlapping generations, they are rarely used owing to the lack of relevant information for most wild populations. Here, we performed a case study on an elusive woodland bat, Myotis bechsteinii, to compare the use of the parentage assignment Ne estimator (EPA) with the more commonly used linkage disequilibrium (LD) Ne estimator in detecting long-term population trends, and assessed the impacts of deploying different overall sample sizes. We used genotypic data from a previously published study, and simulated 48 contrasting demographic scenarios over 150 years using the life history characteristics of this species The LD method strongly outperformed the EPA method. As expected, smaller sample sizes resulted in a reduced ability to detect population trends. Nevertheless, even the smallest sample size tested (n = 30) could detect important changes (60%-80% decline) with the LD method. These results demonstrate that genetic approaches can be an effective way to monitor long-lived species, such as bats, provided that they are undertaken over multiple decades.