Metapopulation Viability of Leadbeater's Possum, Gymnobelideus Leadbeateri, in Fragmented Old‐Growth Forests

Abstract

Risk assessment to determine the probability of persistence of populations has an increasingly important role in the development of conservation and resource use strategies. We used the computer program VORTEX to estimate the viability of populations of Leadbeater's possum, Gymnobelideus leadbeateri, an endangered species of forest‐dependent marsupial inhabiting timber production areas in southeastern Australia. The study simulated population dynamics and genetic variability in metapopulations occupying small numbers of habitat patches of varying size. The impacts of different rates of migration between subpopulations were also examined. Computer simulations with subpopulations of 20 or fewer G. leadbeateri were characterized by very rapid rates of extinction, and most metapopulations typically failed to persist for longer than 50 yr. Increasing either the rate of migration or the number of small subpopulations exacerbated the demographic instability of metapopulations when subpopulations contained <20 individuals and when migration rates were kept within plausible values for dispersal of this species between disjunct habitat patches. This was reflected by lower rates of population growth and depressed probabilities of metapopulation persistence. These effects appeared to be associated with substantial impacts of demographic stochasticity on very small subpopulations together with dispersal of animals into either empty patches or functionally extinct (i.e., single‐sex) subpopulations. There were significant differences between metapopulation dynamics of 40 animals and those comprising 20 or fewer individuals. Increased migration and addition of subpopulations of 40 G. leadbeateri resulted in higher rates of population growth, lower probabilities of extinction, and longer persistence times. Extinctions in these scenarios were also more likely to be reversed through recolonization by dispersing individuals. At the highest rates of migration, subpopulations of 40 G. leadbeateri were essentially panmictic and behaved genetically as a single larger population. Increased numbers of subpopulations and accelerated rates of migration slowed the loss of expected heterozygosity in all scenarios. However, there was a significant (>10%) loss in expected heterozygosity over 100 yr even at the highest rates of migration among five subpopulations of 40 animals. Our analyses predicted that while demographic stability might occur in metapopulations of 200 G. leadbeateri, considerably more individuals than this might be required to avoid a significant decline in genetic variability over 100 yr. Thus, genetic and demographic stability in G. leadbeateri occurred at different metapopulation sizes. Metapopulation structures used in our investigation were hypothetical. However, our results might emulate the dynamics of some populations of arboreal marsupials over the next century within substantial areas of wood production forests in the Central Highlands of Victoria. In many of these areas, there are now only a few and typically very small remaining patches of old‐growth forest that will provide suitable habitat for G. leadbeateri in the long term. Thus, over the next 100 yr, the species might be lost from extensive parts of its present range within montane ash forests that are utilized for timber production. Our study also indicated that there might be metapopulation structures in which the addition of subpopulations and moderate migration could have a negative effect on subpopulation persistence. These findings highlight the importance of understanding the size, number, and isolation of subpopulations that are targeted for management.