MODELING THE IMPACTS OF WILDFIRE ON THE VIABILITY OF METAPOPULATIONS OF THE ENDANGERED AUSTRALIAN SPECIES OF ARBOREAL MARSUPIAL, LEADBEATERPOSSUM

Catastrophic events such as intensive wildfires have a major effect on the dynamics of some wildlife populations. In this investigation, the computer package ALEX (Analysis of the Likelihood of EXtinction), was used to simulate the impacts of wildfires on the persistence of metap opulations of the endangered species Leadbeater's possum (Gymnobelideu s leadbeateri) which is restricted to the montane ash forests of the C entral Highlands of Victoria. A range of scenarios was examined. First , the response of G. leadbeateri to fires in hypothetical patches of o ld growth forest of varying size was modelled. Metapopulation dynamics were then modelled in four existing forest management areas: the O'Sh annassy Water Catchment and the Steavenson, Ada and Murrindindi Forest Blocks using GIS-derived forest inventory data on complex spatial arr angements of potentially suitable old growth habitat patches. The impa cts of different fire frequencies and the proportion of forest area th at was burnt in the Steavenson Forest Block and the O'Shannassy Water Catchment were examined. Finally, the combined impacts of both wildfir es and post-fire salvage logging operations on the persistence of popu lations of G. leadbeateri were assessed. Our analyses indicated that, even in the absence of wildfires, populations of G. leadbeateri are ve ry susceptible to extinction within single isolated habitat patches of 20 ha or less. The probability of persistence approached 100% in patc hes of 250 ha. The incorporation of the effects of wildfire was predic ted to have a major negative impact on isolated populations of G. lead beateri. In these cases, the probability of population extinction rema ined above 60%, even when a single patch of 1200 ha of old growth fore st was modelled. In the absence of wildfires, there was a low probabil ity of extinction of G. leadbeateri in the O'Shannassy Water Catchment where very large patches of old growth forest presently exist. The ri sk of extinction of the species was significantly higher in the Murrin dindi and Ada Forest Blocks where there are lower total areas of, and significantly smaller, suitable habitat patches. Wildfires resulted in an increase in the predicted probability of metapopulation extinction in the four areas that were targeted for study. An investigation of t he Steavenson Forest Block and the O'Shannassy Water Catchment reveale d that the predicted values for the probability of extinction were sen sitive to inter-relationships between the frequency of fires and the p roportion of habitat patches that were burnt during a given fire event . The probability of extinction of G. leadbeateri was predicted to be lowest when there were frequent fires that burnt only relatively small areas of a given forest block. Conversely, the results of our analyse s suggested that populations of the species are vulnerable to infreque nt but intensive conflagrations that burnt a large proportion of the f orest. The results of the suite of analyses completed in this study su ggest that four management strategies will be important for the long-t erm conservation of G. leadbeateri. (1) Attempts to suppress wildfires should be maintained as even the largest remaining areas of old growt h forest may be susceptible to being burnt by repeated widespread wild fires that could result in localised and/or global extinction of the s pecies. (2) Because the probability of population persistence is great est in those areas that support more extensive stands of old growth fo rest, it will be important to grow relatively large patches of existin g regrowth forest (over 50 ha) through to ecological maturity. This wi ll be particularly important in some wood production forest blocks whe re there are only limited areas of old growth forest. (3) Patches that are set aside for the conservation of G. leadbeateri should be spatia lly separated to minimise the risk that all of the reserved areas in a region are destroyed in a single major fire event. (4) Salvage loggin g operations should be excluded from stands of old growth forest and r eserved areas that are burnt in a wildfire. This is because such activ ities can have a major negative impact of the development of suitable habitat for G. leadbeateri.