Estimating Population Size from Dung-based DNA Capture�Recapture Data

Abstract

Non-invasive samples of animal dung are being collected for use with DNA-based capture–recapture analysis to estimate the abundance of the species. A typical sampling strategy involves repeated sampling of transects or grids to search for dung. Searches performed on different transects or on different occasions are often considered the capture sessions for capture–recapture analysis. It is possible that multiple samples of dung from the same individual are found within an occasion. Most current capture–recapture models do not account for this additional data. We present a generalization of the geographically and demographically closed population capture–recapture models that use the numbers of dung found for each individual during each sampling occasion to help estimate heterogeneity in capture probability. We demonstrate that in the face of heterogeneity caused by unequal amounts of dung available per individual, the estimator developed here out performs standard capture–recapture abundance estimation. The estimator allows the benefit of using all data collected while not adding additional costs to the study. We present an example using dung-based capture–recapture data from African elephants to demonstrate the method.