SPATIAL POINT PROCESS ANALYSIS OF AT SEA SEAL ENCOUNTERS

Abstract

We consider a data set comprising encounters between gray seals off Sable Island in the autumn of 2009. We apply Spatial Point Pattern data analysis approaches to this data set and demonstrate the effectiveness of these techniques. Specifically we propose using point processes for clustered data; the Inhomogeneous Poisson process, and the subset of cluster processes called Cox processes. We utilize maximum likelihood methods to produce a model for the mean intensity of the seal encounters which includes spatial covariates. We then consider the impact of inhomogeneity and review minimum contrast methods for fitting the Cox process models, specifically the Thomas and Matern cluster processes. We demonstrate that these methods allow us to gain valuable insight into the nature and extent of grey seal encounters at sea. Finally we validate both our conclusions and the approach taken using the most recent data available, that of autumn 2010.