Inference from single occasion capture experiments using genetic markers

© 2016 Dr. Chathurika Kumari Hettiarachchi HettiarachchigeAccurate estimation of the size of animal populations is an important task in ecological science. However, it is still a challenge in the case of estimation for many endangered species. Classical capture-recapture models have not been able to address this issue to complete satisfaction. Estimating the population size using genetic data also has sometimes led to estimates that differ markedly from each other and also from classical capture-recapture estimates. We use genetic markers to estimate the number of mothers in a population using a single occasion capture experiment. Our work is based on the idea that the degree of biological relatedness between individuals in a sample contributes to the estimation of the size of a population, especially when the population is closed, following Skaug [33]. We mainly construct our estimators using method of moments. First we show that it is feasible to estimate the population size using mother-daughter pairs in the sample, identified using genetic data, observed from a population consisting of two generations. Then we incorporate other relationships into the estimators and extend our method to populations consisting of more than two generations. We show that our estimators are statistically consistent and estimate their standard errors. We also consider the maximum likelihood method in this setting and discover problems that preclude its use. However, we show the method of moments produces reliable estimates. Finally, we study the impact of the misclassification of relationships on the proposed estimators. We illustrate our methods on small subsets of data collected at several sites at Mt Hotham, Victoria, Australia between 2010-2012