The whale shark is the largest fish living in our oceans but little is

known about its ecology and natural history. Tagging studies allow us to learn more about whale shark movements and habitat use. This knowledge is essential for planning the management and conservation of whale sharks to ensure their continued coexistence with humans.

This thesis is concerned with modelling the movements of whale sharks using stochastic models estimated from tracks obtained by tagging studies. In particular, it focusses on using stochastic differential equations being driven by potential functions. Approximations are presented that reduce the task of estimating the potential function to regression problems. I present a method for obtaining smooth potential functions and explicitly modelling measurement error.

The primary scientific questions addressed by the thesis are: Where do whale sharks go, both in terms of geographical location and habitat? Do whale sharks interact with each other? Remote sensing data on sea surface temperature, chlorophyll concentration and sea surface currents are readily available. These time varying covariates are incorporated into the potential function model allowing investigation of their influence on the whale shark's movements.