A stochastic metapopulation model with variability in patch size and position (Q1909292)

We describe and analyse a new stochastic presence-absence metapopulation model which allows for differences in size and quality of patches and incorporates the explicit spatial distribution of patches. The principal issues we address are: (i) estimating extinction probabilities and (ii) calculating the probability distribution function of the state of the system, given that the metapopulation is not extinct. We relax the assumption that all patches are identical. We focus our attention on presence-absence models, formulating a model with a discrete, finite state space because this allows us to avoid the mathematical and computational complications which arise from stochastic models which track the number of individuals in each patch. We apply our model to malleefowl, Leipoa ocellata, inhabiting fragments of habitat in the Bakara region, southern Australia. We consider an existing system of eight patches, where the patches range in size from 100 to 2700 hectares. For this metapopulation, we calculate extinction probabilities and the quasi-stationary distribution. We also examine the effect of patch removal on the metapopulation extinction probability. To illustrate the influence of variation in patch size, we analyse this same system assuming that all patches have the same area, but maintaining the same total area of habitat. We also explore the effect of explicit spatial structure, by analysing both of these systems with the additional assumption that all patches are equally accessible.