The effects of different forestry practices on two native rodent species, the swamp rat (Rattus lutreolus) and the long-tailed mouse (Pseudomys higginsi)

Forest fragmentation, modification and loss can have a range of negative impacts on wildlife, including reduced foraging opportunities, increased competition for resources, loss of habitat connectivity and restricted dispersal, and increased predation risk due to removal of habitat cover. Harvesting practices such as clearfelling (clearcutting) in native forests typically remove all standing mature forest elements, resulting in large tracts of land with little vegetation cover and altered biodiversity. An alternative practice, aggregated retention, was developed with the objectives to 'lifeboat' species and processes, retain and enhance structural complexity and improve connectivity within the landscape, by retaining patches of unlogged forest within 'islands' and surrounding 'edges' in the harvested matrix. Although this practice has been successful in retaining biodiversity and mature forest species for some taxa, there have been relatively few studies on small ground mammals, particularly in the Southern Hemisphere, and very little attention has been given to landscape connectivity. The main aim of this thesis was to determine the effects of different forestry practices (clearfelling, unlogged native forest and aggregated retention treatments) in wet eucalypt forest in Tasmania on two native rodent species: the swamp rat (Rattus lutreolus), a cover-dependent species, and the long-tailed mouse (Pseudomys higginsi), a habitat generalist. The first part of this project involved a field investigation of rodent abundances, demographics and habitat use (Chapter 2). The distinctly different responses of the two species to each practice, particularly within aggregated retention, then prompted investigations into physiological responses of both species (Chapter 3), and the genetic (Chapter 4) and behavioural (Chapter 5) implications of forestry practices on the cover-dependent swamp rat. A major field study examining rodent abundances (Chapter 2) showed that the cover-dependent swamp rat declined with increasing disturbance among the three treatments, with abundance highest in unlogged forest, lowest in clearfelled and intermediate in aggregated retention. These responses to disturbance were also seen in the different habitat types created within aggregated retention sites, with lowest abundances in the harvested matrix, highest in the forested edges and intermediate in the forested islands. There was also a significant positive relationship between swamp rat abundance and lower strata vegetation cover in harvested areas. In contrast, the abundance of the long-tailed mouse was not significantly different among treatments nor within the different habitat types in aggregated retention sites and there were no clear relationships with vegetation cover. The abundance results indicated that swamp rats were highly sensitive to harvesting while long-tailed mice were resilient and able to persist in harvested areas. Interestingly, the physiological data (blood profiles and body condition, Chapter 3) did not reflect this result, with no indication of stress responses nor differences in general condition in swamp rats, while long-tailed mice showed poorer body condition in clearfelled sites compared to unlogged sites. Long-tailed mice may only inhabit harvested areas out of necessity rather than showing resilience to disturbed habitats. Swamp rats were rarely found in harvested areas and may be minimising physiological impacts by preferentially residing in forested areas. Alternatively, populations may be experiencing elevated physiological stress in both harvested and unlogged sites due to fragmentation of the latter by minor roads. Habitat fragmentation can impede movement of animals between suitable habitat, restricting dispersal and gene flow, and resulting in population differentiation. Analyses of swamp rat genetic samples (Chapter 4) from aggregated retention and unlogged sites revealed no evidence of inbreeding, but increased relatedness in aggregated retention island patches, which is most likely due to restricted dispersal across the 'hostile' harvested matrix. Surprisingly, analyses also revealed that swamp rats do not easily move across unpaved, narrow (< 10 m) and seldom-used roads. While harvesting may result in immediate and large-scale changes to suitable habitat, roads may pose a longer-term hindrance to dispersal. Swamp rats prefer dense vegetation cover (Chapter 2), although the importance of ground-level structural cover and overhead visual cover was not clear from the field trial. Therefore, captive behavioural trials (Chapter 5) were run to test habitat cover preferences by swamp rats using ground-level structural cover and 1 m high overhead (visual) cover in low risk (dark) and high risk (light) conditions. There were no clear preferences for different densities of structural or visual cover. However, the walls of the experimental arena (essentially a type of structural cover, perhaps analogous to large logs) were preferred over the centre area of the arena, regardless of cover density or risk conditions for both structural and visual cover types. This thesis highlights the importance of using multiple disciplines (ecology, physiology, genetics, and behaviour) to investigate anthropogenic disturbances on wildlife. Despite persistence within the harvested matrix, long-tailed mice showed decreased body condition, which may have longer-term health and reproductive consequences. Additionally, while swamp rat populations appear to be thriving in unlogged forests, population differentiation is occurring due to the presence of unpaved, narrow, and seldom-used roads acting as dispersal barriers. It also confirmed that the practice of aggregated retention as an alternative to clearfelling is beneficial for small ground-dwelling mammals for the objective of life-boating, but may not be providing landscape connectivity as there are some restrictions for dispersal of cover-dependent species, at least over the short-term.