Abstract
Predicting the dispersal of an invasive species over a heterogeneous region is a difficult yet important task, since these predictions can inform biosecurity efforts and threatened industries. A spatially-discrete population model of the citrus pest and disease vector Diaphorina citri is developed. The population is stratified by life cycle and sex, allowing phenological processes and second order interactions to be modelled effectively. The model is formalised as a continuous time Markov jump process and simulated using a tau-leaping algorithm. The model introduces several unknown parameters which are related to mating, trapping, density dependence, and dispersal. The dispersal kernel incorporates both natural and human mediated dispersal; the latter uses a gravity model. Three example parameter estimation problems are presented which use both real and simulated data. Response surface optimisation techniques and bootstrap confidence intervals were used to estimate the parameters.