Density, movements and hunting of feral cats in relation to fire and grazing in northern Australia

Vegetation structure plays a pivotal role in predator-prey interactions, because it influences the effectiveness with which predators are able to detect, pursue or ambush their prey, and determines the opportunities for prey to avoid being exposed to predators. Therefore, changes in vegetation structure may alter the impacts of predation on populations of prey species. In recent decades, populations of small mammals have collapsed across much of the savannah biome of northern Australia. The causes of this decline are not known, but predation by feral cats, intensification of fire regimes, and intensified grazing by introduced herbivores have all been implicated as possible causes by correlative and circumstantial evidence. However, we do not understand the mechanisms by which these factors, alone or in combination, affect population size in small mammals. Without this mechanistic understanding it is difficult to prescribe management interventions that will protect small mammals from further declines and allow recovery. I investigated interactions between predation by feral cats and vegetation structure in a study area in the central Kimberley region of northwestern Australia, to test whether one effect of fire and grazing regimes is to amplify the impacts on small mammals of predation by feral cats. I addressed three specific objectives. First, I measured population density of cats differed in areas with contrasting fire and grazing regimes. The second objective was to test how the movements of individual cats were influenced by vegetation and habitat structure, and especially to show whether cats preferentially hunted in burnt or grazed areas. The third objective was to derive direct measures of kill rates of feral cats, to show how these were influenced by habitat characteristics. To determine cat density, I deployed six arrays of infrared cameras across the study area during 2012 and 2013. The arrays were allocated evenly to areas with and without large introduced herbivores, and ranged across areas with differing fire histories. Each array consisted of 15 or more cameras deployed for one month. All cats 'captured' on camera were individually identified, and density was estimated using spatially-explicit mark-recapture analyses. Density at all six arrays was generally similar (mean 0.18 cats per km2, with SE of 0.08). There was no significant difference in density between grazed and un-grazed arrays, despite a five-fold difference in small mammal abundance. These results demonstrate that cats occur at low density in the Kimberley compared to other parts of Australia, and that population density does not increase when small-mammal abundance increases in ungrazed areas. To investigate movement behaviour and fine-scale habitat selection by cats, I deployed GPS collars on 32 cats in landscapes with contrasting fire and grazing treatments, and used discrete-choice modelling in relation to detailed habitat measures to reveal the movement decisions that underlie habitat selection by individual cats. Cats selected areas with more open grass cover, heavily grazed areas, and areas close to standing water. Mild fire scars were avoided, but cats strongly selected for areas recently burnt by intense fires in habitats with high abundance of small mammals. Not only was this the strongest influence on movement decisions made by cats within their home ranges, but cats also predictably journeyed up to 15 km to temporarily exploit areas recently burnt at high intensity outside the borders of their home range. I deployed collar-borne video cameras on cats to record hunting events and to measure the effects of habitat on hunting success. In this way I obtained 98 hours of cat activity, during which 101 hunting events were observed. The success or failure of cats in killing prey at a hunting event was strongly affected by habitat. For hunting events in which prey were in a grass tussock or a rocky refuge, the kill rate was four times lower (18 %) than in open areas (76 %). That is, cats enjoyed higher hunting success in the habitat types that they preferred to enter. This research demonstrates how a relatively sparse population of cats can have major impacts on prey populations if disturbance regimes are not managed appropriately. In particular, cats respond well to the conditions created by intense fire and grazing, increase their hunting in such areas, and accomplish higher hunting success. Vegetation structure is pivotal to creating ideal landscapes for predators to hunt, or conversely, to providing refuges for prey to avoid predation. Intense fires are likely to become more prevalent with climate change, and this could increase the impacts of many species of small to medium sized predators. A useful broad principle to reduce the impacts of this predation could be to increase the spatial and temporal cover of ground vegetation wherever possible.