Broad-scale habitat fragmentation is a visible result of human land-use throughout the world, often resulting in deleterious ecological outcomes. Animals inhabiting fragmented landscapes need to access different habitats to acquire resources such as food and shelter which can pose risks if human disturbances are present. To mitigate these risks, animals may change their patterns of resource selection on a temporal scale.
Phillip Island, located in south-east Australia, is a highly human-modified and fragmented landscape that contains patches of native vegetation amongst a matrix of agricultural farmland and urban developments. The island supports an abundant population of swamp wallabies (Wallabia bicolor), but little is known about how they use a modified landscape and select resources within it.
We developed inexpensive custom-made wildlife trackers to gather fine-scale GPS data of 48 swamp wallabies and examined the circadian variation in resource selection within habitats of low (i.e. remnant vegetation) to high (i.e. roads) human disturbances. We compared real wallaby trajectories to simulate random walks and applied generalised linear mixed models to analyse the data on a temporal scale. Firstly, we investigated which habitats are selected by wallabies and secondly, we determined whether the distance to different habitat types changes, depending on the level of disturbance. Finally, we created habitat suitability maps to indetify aras of use and avoidance. Information generated by the analysis will inform and improve the conservation of swamp wallabies in a fragmented and human-distrubed landscape by implementing effective management actions and more generally enhance our understanding of the challenges faced by fauna in changing environments worldwide.