Understanding the response of a woodland-dependent mammal to habitat loss and fragmentation in the Midlands bioregion, Tasmania

Habitat loss and fragmentation are the leading cause of global biodiversity decline, with agricultural expansion being the dominant driver. The socio-economic importance of agriculture, together with growing human population and resource demand, is expected to result in a further one billion hectares of land being converted for agriculture and crop land by the year 2050. Species are increasingly being restricted to smaller remnants of their original range, in habitats that have undergone significant ecological change. It is increasingly acknowledged that protecting habitat is not sufficient to protect biodiversity, and instead ecological restoration will become more important for biodiversity persistence. For ecological restoration to be successful, a grounded and detailed understanding of species responses to fragmentation is needed. Over the last 240 years, the Australian mainland has experienced the greatest loss of mammalian diversity and abundance of any comparable land area in the world. This is largely due to the introduction of invasive predators, loss of habitat and changing fire regimes. The mammalian species most at risk fall within the 'critical weight range' (CWR) of 35-5500g (Woinarski, Burbidge & Harrison 2014; Woinarski, Burbidge & Harrison 2015b). Many CWR species are threatened or near extinction. The island state of Tasmania is believed to have secure populations of CWR mammals, mainly because of the absence of the red fox in Tasmania. However, CWR mammals in Tasmania also face pressures from habitat conversion and other disturbances, especially in the agriculturally dominated Midlands bioregion. Little is known regarding the distribution and response of CWR mammalian species to such stressors. Current management efforts are in place to restore native vegetation in the Midlands to promote connectivity for wildlife. This project focuses on a critical weight range mammal, the eastern bettong (Bettongia gaimardi), to provide essential knowledge on the fundamental ecology of the species needed to plan management and restoration of habitat. The first aim of this study was to determine the distribution of the eastern bettong across the Midlands bioregion and identify habitat attributes influencing the probability of presence. A landscape-scale camera survey was carried out using 62 sites, repeated in summer and winter. Using occupancy modelling, I modelled habitat and landscape features to determine what variables predicted the presence of eastern bettongs. My analysis was designed to evaluate the predictions of alternative hypotheses on the effects of habitat loss on the persistence of species: the Island Biogeography (or metapopulation) hypothesis suggests that the occurrence of species will increase with patch size, and decreased patch isolation; the Habitat Amount hypothesis suggests that occurrence will increase with the amount of habitat available in a landscape (at a scale relevant to the movement capacity of individual animals), regardless of its patchiness. Occupancy was best explained by habitat amount within a 1km buffer rather than by patch size and isolation, together with the quality of the available habitat (quality was indicated by the density number of regenerating stems of canopy trees). These results highlight the value of small patches in fragmented landscapes for species such as the eastern bettong, and the significance of the quality of those habitat remnants. The second chapter focuses on providing a mechanistic understanding of how the eastern bettong responds to fragmentation through variation in home ranges in fragmented landscapes. I used GPS tracking of 24 individuals across three different fragmented sites to describe variation in home ranges, and I modelled the effects of home-range size on the quality, amount and fragmentation of habitat within radii of 750 m and 250m from the mean centre of activity of individual home ranges. I also estimated population density using spatially-explicit capturerecapture analyses at the three sites, and for comparison an additional protected area with a large tract of continuous habitat. My results showed that habitat quality, the amount of woodland and population density were the most important determinants of home-range size, while there was no effect of fragmentation. Habitat quality was the strongest determinant: home ranges were smaller in areas of higher habitat quality. On the other hand, bettongs increased their home-range size with higher density and greater amount of woodland. These results suggest that fragmentation does not limit home-range size of bettongs, who can compensate for fragmentation by increasing their ranges to incorporate more patches provided there is enough woodland accessible to them, but that quality is crucial for habitat use and therefore persistence. However, population density was lower in fragmented than in intact, contiguous woodland. Furthermore, large variations in bettong density across sites suggests habitat area is important for bettong persistence, and knowledge of baseline density is important for conservation and management. In the last chapter, I provide a finer scale understanding of bettong habitat use and selection by investigating the movement pathways of 24 GPS tracked individuals across three sites. In the first study to use state-space modelling (Hidden Markov Modelling) on a small terrestrial mammal, I identified three distinct behaviour states based on movement patterns: denning, foraging and fast travelling. Transitions between behaviours were associated with the density of vegetation and the percent of woodland cover. Bettongs foraged in woodland and denned in areas with greater vegetation density but lower woodland cover. Across sites, bettong movements differed according to the availability of high-quality habitat. Moreover, movement of individual bettongs was not hindered by fragmentation as they readily crossed gaps, but their gap-crossing potential was improved by the presence of low vegetation and or stands of trees. My results suggest that the eastern bettong is able to persist in the Midlands bioregion with the current extent of fragmentation and patch degradation. However, there is still cause for concern over the future of the eastern bettong in this region. Estimates of population density suggest bettongs are sensitive to habitat quality decline associated with the further risk of habitat loss and fragmentation. To improve the population status of bettongs in the Midlands, restoration efforts should focus on improving the quality of woodland remnants and adding to the total amount or area of woodland habitat. Where habitats are degraded, restoration by planting vegetation to increase the biomass of fine roots (for example: Lomandra longifolia, Eucalyptus sp, Acacia sp.) can promote the growth of ectomycorrhizal fungi, which are the major food resource for bettongs and therefore increase the chance of persistence of bettong populations. Ensuring there is a high density of low-cover vegetation will provide nesting material for animals, while also providing areas that could be used by animals when crossing gaps between isolated woodland patches. Finally, many of the sites within which bettongs are found are under covenant protection giving these sites protection from disturbances such as grazing and fire. Using covenant sites as focal points for restoration and promoting their connectivity, could be an avenue of preserving and increasing the amount of high-quality woodland, which would benefit bettong populations.