Advances in mark-recapture models with identification error

Speaker: Richard Vale ( University of Canterbury, Christchurch, New Zealand and Inland Revenue NZ)

Abstract

Capture-recapture is an important technique for estimating animal populations. Sometimes animals have to be identified from incomplete genetic material or blurry photographs, which can make it difficult to tell whether or not a particular animal has been captured before. Recently there has been some interest in taking account of identification errors in capture-recapture studies. We describe the model M_{t, alpha} for capture-recpature with identification errors and explain a new way to calculate the likelihood function, which is much faster than previous methods. I will describe our attempts to apply the model to real and simulated data. This is joint work with Rachel Fewster, Emma Carroll and Nathalie Patenaude.