Restoring farmland for biodiversity : a carnivorous perspective

Habitat loss and degradation are leading drivers of biodiversity loss and species decline worldwide, with a large proportion of this loss due to land conversion for primary production. Revegetation and habitat restoration programs are therefore critical in agricultural regions to maintain, create and connect areas of high biodiversity value and prevent species extirpation. To be effective, such programs must look beyond re-establishing vegetation cover and towards restoring the full suite of ecological processes needed to maintain biodiversity. Often, maintaining or re-establishing fauna populations is critical in the restoration of these processes, and predators are particularly relevant as they regulate and promote biodiversity at lower trophic levels. Restoring predator guilds in modified ecosystems can therefore be an important conservation tool, however the traditional emphasis on vegetation has meant that the field of restoration ecology lacks a coherent framework for fauna restoration. While the restoration of suitable habitat is a clear prerequisite, habitat requirements may vary with species, sex, developmental stage, season and wider landscape context. The overarching goal of this thesis was to inform the restoration of a native mammalian carnivore guild as part of an ongoing habitat restoration program in the Tasmanian Midlands. Despite a long history of land clearing and modification, this region is one of Australia’s 15 biodiversity hotspots. The region still supports populations of three native marsupial carnivores which are extinct or endangered on the Australian mainland, the Tasmanian devil (Sarcophilus harrisii), spotted-tailed quoll (Dasyurus maculatus) and eastern quoll (Dasyurus viverrinus) although at least two of these species have declined in recent decades. Restoration of the region’s native carnivore guild would entail maintaining or restoring healthy populations of these species and controlling the introduced feral cat (Felis catus), which is widespread throughout the region and Tasmania. The first aim of the thesis, addressed in Chapter 1, was to assess the current status of these species within the region and to identify factors driving patterns of species occupancy and abundance. We surveyed 110 camera sites across the Midlands in summer and winter, and analysed the results using hierarchical occupancy modelling to account for imperfect detection. We compared the ability of three of the most commonly applied theoretical landscape ecology frameworks to explain patterns of occupancy and abundance for each species. No single framework was consistently superior, but the continuum framework showed the highest and most consistent predictive power. We suggest that this is because species responses are directly related to continuous patterns of underlying environmental variation rather than to a binary perception of habitat/non-habitat, as assumed by the island biogeography and habitat amount frameworks. The feral cat, Tasmanian devil and spotted-tailed quoll were found throughout the region, whereas the eastern quoll was almost exclusively confined to the rugged and less disturbed edges of the region. Factors such as vegetation structure, topographic complexity and the density of linear edge habitats were most important in explaining patterns of species persistence, highlighting the importance of variation in habitat quality and suggesting that restoration practitioners should place more emphasis on habitat structure, complexity and location than on habitat area alone. In addition, the scale and method of derivation of parameters within all frameworks strongly influenced model performance. Model parameters in habitat restoration planning should therefore reflect the underlying biology and behaviour of the species in question. Having established that variation in habitat quality influences the persistence and abundance of species, Chapter 2 aimed to understand how such variation influences home-range size. The size of an animal’s home-range is crucial in the design of effective species management, whether for species conservation or control. Home-range size can vary up to two orders of magnitude between populations of the same species and is determined by the balance between the individual’s metabolic needs and energetic constraints. We compared home-range sizes and average nightly movements and population densities for populations of the marsupial spotted-tailed quoll and eutherian feral cat at four sites throughout the Midlands. These two mesocarnivores show considerable overlap in body size and diet, and both responded strongly to gradients in habitat productivity across the sites studied. Both species showed a significant decrease in population density with declining productivity, however only spotted-tailed quolls also showed a significant increase in home-range size. We suggest that information on both of these parameters is needed to accurately determine how species respond to variation in habitat productivity, and therefore for effective management of species across their geographic range. Within the Midlands, restoration practitioners will need to take variation in home range into account when designing the scale and location of restoration plantings. Chapter 3 aimed to investigate the implications of restoring native carnivores and controlling the introduced feral cat. Invasive mammalian carnivores have had a disproportionate impact on global biodiversity loss and feral cats, in combination with red foxes (Vulpes vulpes), are a leading cause of decline and extinction of native vertebrate fauna in Australia. In this chapter, we combine multiple theoretical frameworks to contrast the potential predation impact of the feral cat with its closest native counterpart, the spotted-tailed quoll. We found that the predation impact of feral cats is likely to be several orders of magnitude higher than that of spotted-tailed quolls, regardless of the anti-predator response adopted by prey species. Prey are 10 – 100 times more likely to encounter feral cats within the landscape, in part due to much higher population densities, but also a greater intensity of habitat use and broader habitat preferences. These latter characteristics also mean that the costs to prey of adopting anti-predator behaviour against feral cats are likely to be much higher, due to the reliability and ubiquity of feral cat cues. These results suggest that restoring a native carnivore guild and controlling feral cats would reduce overall predation pressure on native prey species, but that competition with introduced prey may increase if native carnivores are not effective in controlling species with high reproductive rates such as rabbits (Oryctolagus cuniculus) and black rats (Rattus rattus). Our results highlight the benefits of a mechanistic, animal-centric understanding of species responses to local environmental gradients and processes when planning restoration programs. We demonstrate that habitat quality is more important than overall habitat area in predicting patterns of species occupancy and abundance, suggesting that restoration programs with a greater focus on habitat structure and complexity are likely to be more successful than those which focus on area alone. We demonstrate that even within a relatively small region like the Midlands, the choice of restoration location will strongly influence its success, with both native and introduced carnivores showing very large variation in home range size and/or population densities in response to environmental productivity. Finally, by demonstrating that the potential predation impact of an introduced predator, the feral cat, is up to several orders of magnitude higher than its closest native counterpart, we confirm that the composition of the carnivore guild can strongly influence top-down forces acting on the local ecosystem. Our results support the contention that managing fauna consumers can be critical in achieving the goal of restoring ecosystem processes, and challenge the historic emphasis on vegetation establishment in restoration ecology