Landscape and restoration genetics of contrasting Tasmanian marsupials

Australia's mammal fauna has suffered a dramatic decline over the last 200 years, with 30 species becoming extinct since European colonisation and many more threatened. The island of Tasmania, 250 km south-east of continental Australia, has historically acted as a refuge for several mammals that have become either extinct or threatened on the mainland. Anthropogenic pressures including habitat loss and fragmentation and climate change are among the major threats to many mammal species, and a clear understanding of their impacts is required for effective conservation. Genetic research can contribute valuable data to inform conservation efforts at range-wide and finer spatial scales. The goal of my PhD was to support conservation efforts for two marsupials, the eastern bettong (Bettongia gaimardi) and spotted-tailed quoll (Dasyurus maculatus), in Tasmania through genetic research at multiple spatial scales. For each species, I investigated patterns and causes of genetic variation across their range in Tasmania, and then identified potential factors affecting genetic connectivity within smaller-scale landscapes. In particular, by investigating landscape factors affecting gene flow, I aimed to provide data to inform an ongoing habitat restoration program in the Midlands region of Tasmania. The Midlands is an important part of both species' ranges that has experienced extensive habitat loss and fragmentation due to agricultural intensification in the last 200 years. To facilitate this aim, in Chapter 2, I reviewed tools from the fields of population and landscape genetics that can be used to assess connectivity and identified novel ways to apply these to the planning and evaluation of ecological restoration projects. The eastern bettong is a medium-sized marsupial with a climatically restricted, patchy distribution in Tasmania, being limited to the relatively dry and warm eastern half of the island. As a woodland-dependent species, it is potentially highly susceptible to the effects of loss and fragmentation of woodland habitat, which has been extensive in the Midlands and other agricultural regions. Thus, my goals were first to test the effect of climatic factors on genetic structure across the species' range, and then to investigate how loss of connectivity of woodland patches affects localised gene flow, and hence genetic structure at the landscape scale. In Chapter 3, I investigated the effects of short-term climate variability (i.e. weather) on genetic diversity in bettongs. I quantified population genetic structure and diversity by surveying ~2900 putatively neutral single nucleotide polymorphism (SNP) markers and the mitochondrial control region in 294 bettongs from 18 localities range-wide. I then assessed whether genetic diversity in this species could be predicted by a weather-based species distribution model (SDM), which is an emerging technique that uses short-term weather variables, rather than multi-decadal climatic averages, to create dynamic models of weather suitability over time. I found substantial genetic structure and variation in genetic diversity between populations. Weather stability was a strong predictor of population genetic diversity in eastern bettongs, which is likely due its influence on patterns of local extinction and recolonisation. This is the first evidence that weather-based SDMs could be a novel and important tool for predicting the impacts of current and future climate change on genetic diversity in weather-sensitive species. In Chapter 4, I tested the effects of habitat fragmentation on gene flow among bettong populations in the Midlands. I quantified genetic differentiation among 9 populations based on SNP genotypes, and measured key fragmentation variables within least-cost transects between each population. I then used maximum likelihood population-effects (MLPE) models to compare the effects of different fragmentation variables on genetic structure. I found strong effects of habitat fragmentation on genetic connectivity between populations. In particular, connectivity was affected by least-cost path length, maximum distance of open habitat crossed, and the proportion of the transect comprised of closed habitat. Amount of habitat edge and size of patches did not strongly predict connectivity, suggesting that it is the amount of closed habitat and distances between patches, not patch size or edge effects, that particularly influence connectivity in bettongs. In contrast to the eastern bettong, the spotted-tailed quoll is a generalist and wide-ranging carnivore, found at low densities across Tasmania. It has substantial dispersal capabilities across a range of habitat types, and thus range-wide and smaller-scale genetic patterns are likely to be affected primarily by the intrinsic dispersal capabilities of the species, and potentially by major historical or contemporary barriers to movement. For this species, my goals were first to quantify the extent of gene flow Tasmania-wide and identify any major barriers, and then to investigate whether climatic variables, habitat loss or other anthropogenic pressures act as barriers to gene flow at a smaller landscape scale. In Chapter 5, I investigated broad-scale historical and contemporary genetic structure in 502 spotted-tailed quolls sampled from across the species' range in Tasmania, using 13 microsatellite loci and the mitochondrial DNA control region. A Bayesian clustering analysis and spatial principal components analysis identified a gradual genetic transition from western to eastern Tasmania, consistent with an isolation-by-distance pattern. I did not find evidence of either contemporary or phylogeographic barriers to gene flow. I found significant positive spatial autocorrelation over distances up to 116 km for females and 133 km for males, confirming that contemporary gene flow in this species is substantial, relative to the size of its Tasmanian range. In Chapter 6, I assessed the effects of bioclimatic variables, anthropogenic land use change and geographic distance on gene flow in spotted-tailed quolls at smaller scales, in three contrasting 15 km x 45 km landscapes in Tasmania. I parameterised resistance surfaces reflecting hypotheses on the effects of landscape variables on gene flow, and tested the relationship between genetic distances and resistance distances in an MLPE framework. I did not find consistent effects of any variable in all three landscapes, and in two of the landscapes, there were only weak relationships between genetic structure and environmental variables. In the third landscape, the northern Midlands, which is more marginal for this species, environmental variables had stronger effects, although these may have been amplified by sampling limitations. My research provides guidance for the conservation of these contrasting species at range-wide and finer landscape scales. For the spotted-tailed quoll, with strong inherent dispersal abilities and gene flow occurring over large scales, climatic variables and landscape change do not currently appear to be creating barriers to gene flow. Thus, for this species, conservation efforts should focus on ensuring that sufficient habitat and resources are available to support the largest population densities possible given land use, which may be at carrying capacity in intact habitat and at lower but sustainable densities in fragmented agricultural landscapes. Conservation measures would include appropriate land conservation across Tasmania, and restoration and expansion of habitat in the Midlands. For the eastern bettong, a species with more restricted dispersal and specialist habitat requirements, targeted conservation actions are needed to protect genetic diversity and connectivity. At the range-wide scale, weather-based SDMs can be used to identify and prioritise areas of potentially high genetic diversity for conservation of bettongs and other weather-sensitive species. Within the Midlands, restoration practitioners should focus particularly on increasing the amount of habitat in the landscape and reducing large gaps between patches to improve connectivity, but my results suggest that the size and shape of these patches are less important. These findings may also be useful for guiding conservation and restoration efforts for other Tasmanian animals with similar dispersal capabilities and habitat requirements to bettongs and quolls.