Spatially-correlated extinction in a metapopulation model of Leadbeater's Possum

The importance of considering spatially-correlated extinction in metapopulation viability analyses was investigated using a model of the population dynamics of Gymnobelideus leadbeateri McCoy (Leadbeater's Possum). Fire caused local extinction of G. leadbeateri and induced changes in the suitability of the habitat over a period of decades and centuries. Spatially-correlated fires, in which the correlation between the incidence of fire declines with distance, and uniformly-correlated fires were simulated. The predicted risk of metapopulation extinction increased: (i) as the variance in the number of fires each year increased, (ii) as the mean fire interval decreased, and (iii) as the mean dispersal distance decreased. Incorporating spatial correlation in the incidence of fires between patches had little effect on the results, provided the variance in the number of fires per year remained the same and fires modified habitat quality. The predicted risk of metapopulation extinction was greater for spatially-correlated fires than for uniformly-correlated fires when fires only caused local extinction but did not change habitat suitability. Incorporating spatial correlation in the incidence of fire within patches, which allowed partial burning of patches, reduced the predicted risk of extinction. This effect was only slight when patches were smaller than about 50 ha. The results of our simulations demonstrate the importance of considering correlations in disturbance regimes in metapopulation models, especially if these models are used to assist the design of nature reserves.