The conservation of arboreal marsupials in the montane ash forests of the central highlands of Victoria, south-eastern Australia. VIII. Landscape analysis of the occurrence of arboreal marsupials

Abstract

We describe the results of landscape analysis of the occurrence of a suite of species of arboreal marsupials in the montane ash forests of the central highlands of Victoria, south-eastern Australia. The species targeted for study were: Yellow-bellied Glider ( Petaurus australis), Mountain Brushtail Possum ( Trichosurus caninus), Greater Glider ( Petauroides volans) and Leadbeater's Possum ( Gymnobelideus leadbeateri). Our analyses involved exploring statistical relationships between the presence/absence of these four species and compositional variables used to characterise the landscape within 20 and 80 ha circles surrounding each of 166, 3 ha, sites surveyed for arboreal marsupials. Many potential statistical models were examined for each of the four species targeted for analysis. With the exception of P. australis, the final model chosen was the null model; i.e. none of the landscape variables were statistically significant as predictors of the presence of animals at a site. In the case of P. australis, three measures of the landscape within the 80 ha circles surrounding our survey sites were found to be statistically significant: (1) the presence of old forest, (2) the steepness of the forest terrain, and, (3) the area of forest on different aspects. There was a significantly higher probability of detecting the species on northerly and westerly aspects, on either steep or flat terrain, and within circles supporting old growth montane ash forest. P. australis has a home range of 40–60 ha which is at least an order of magnitude larger than the other species of arboreal marsupials examined. Thus, the size of the 80 ha circle used to capture data on landscape composition matched the home range size of P. australis more closely than other species of arboreal marsupials examined in this study. This may, in part, explain why it was the only species for which landscape composition variables were found to be significant predictors of occurrence. Our findings tentatively suggest that changes in the spatial coverage of old growth forest that have resulted from past wildfires, post-fire salvage logging and more recent intensive and extensive clearfell logging operations, have created landscape patterns unsuitable for P. australis. The implications of our results for the conservation of the species are discussed. In addition, some difficulties were experienced in the use of landscape-scale data in our statistical analyses, such as identifying attributes of landscape composition that are biologically relevant to the taxa targeted for analysis. We briefly discuss the nature of these problems.