THE CONSERVATION OF ARBOREAL MARSUPIALS IN THE MONTANE ASH FORESTS OFTHE CENTRAL HIGHLANDS OF VICTORIA, SOUTH-EASTERN AUSTRALIA .6. THE PERFORMANCE OF STATISTICAL-MODELS OF THE NEST TREE AND HABITAT REQUIREMENTS OF ARBOREAL MARSUPIALS APPLIED TO NEW SURVEY DATA

Statistical models are widely used as a tool to provide a concise desc ription of data by identifying and quantifying relationships between k ey explanatory variables. The adequacy of the fit of such models to th e data can be gauged by various diagnostic procedures. However, the pe rformance of statistical models, when applied to new datasets, can yie ld valuable insights about the various assumptions that underpin such models and may identify potentially important new explanatory variable s. In this study we describe the results of a field survey that was un dertaken specifically to assess the performance of previously develope d statistical models of the nest tree and habitat requirements of arbo real marsupials inhabiting the montane ash forests of the central high lands of Victoria, south-eastern Australia, The study involved compari ng legit and Poisson regression models developed from surveys at 152 s ites completed between 1983 and 1989 with a new dataset comprised of 5 5 sites that was gathered in 1992-93. Our analyses demonstrated that m ost of the models developed in earlier studies performed reasonably we ll when they were applied to the recently collected data. Thus, for th e two datasets, there were similarities in the statistical relationshi ps between characteristics of trees with hollows and their occupancy b y various species of arboreal marsupials. Similar results were recorde d from assessments of the performance of the habitat models. Notably, for some species such as Leadbeater's possum, Gymnobelideus leadbeater i McCoy, there was greater congruence between the results of the two s urveys than recorded in similar comparisons with data collected front retained linear strips (wildlife corridors). This finding further emph asizes previous observations that the species appeals to be disadvanta ged in narrow areas of linear habitat.