METAPOPULATION VIABILITY OF ARBOREAL MARSUPIALS IN FRAGMENTED OLD-GROWTH FORESTS - COMPARISON AMONG SPECIES

Computer simulation analyses using a generic package for population vi ability analysis (PVA) were used to examine the dynamics of a limited number of metapopulation structures of the mountain brushtail possum, Trichosurus caninus, and the greater glider, Petauroides volans. These arboreal marsupials inhabit tall eucalypt forests in southeastern Aus tralia. As in a parallel study of Leadbeater's possum, Gymnobelideus l eadbeateri, the results indicated that there are some metapopulation s tructures in which increased movement between habitat patches may have a detrimental effect on demographic stability and, in turn, reduce po pulation persistence. These negative impacts were most pronounced in m odel metapopulations composed of an ensemble of very small subpopulati ons of 4, 10, or 20 animals. The patterns were accentuated with increa sed inter-patch movement as well as the addition of further small subp opulations. These counter-intuitive findings appear to result when ani mals from already unstable and declining sub-populations disperse into unoccupied patches of habitat that act as a type of population sink. These trends in population demography were reversed in scenarios in wh ich larger subpopulations of 40 animals were modeled. In such cases th e impacts of movement and additional habitat patches on subpopulation and metapopulation dynamics were positive. In contrast to the results for various demographic parameters, increased movement and added subpo pulations had a beneficial effect on genetic variability among all the types of metapopulation structures that were examined. Results for th e two species were compared to each other and also with investigations of Leadbeater's possum, Gymnobelideus leadbeateri. These comparisons confirmed general trends but also revealed potential quantitative diff erences in the behavior of metapopulations of T. caninus, P. volans, a nd G. leadbeateri. Many of these differences were attributed to variat ions in key life history parameters such as birth and death rates, but trends in metapopulation viability did not correlate simply with intr insic rates of population growth. The three species differed in the si ze of metapopulations required for demographic and genetic stability, and the rank order of their susceptibility to demographic and genetic stochasticity may not be consistent. Different species, even those wit hin the same guild, may vary in vulnerability to disturbance and envir onmental perturbation. Our analyses also emphasized the need to unders tand the structure and composition of metapopulations in the developme nt of strategies for conservation of wildlife at the landscape and reg ional scales.