Eastern brown snakes (Pseudonaja textilis) are the most common snakes encountered in suburban Canberra, with hundreds translocated out of conflict situations each year due to fear of snakebite. As rates of both urbanization and human-snake conflict increase, understanding the impacts of management strategies like translocation are essential. Previous research has shown that translocating snakes out of their home range can negatively influence their movement, reproduction, and survival. However, despite their critical role in human-snake conflict and their wide occurrence across Australia, brown snakes remain grossly understudied. Our research aims to answer key questions about brown snake spatial ecology and to determine the impacts of translocation. To address this, we compared the survival and movement behaviour of 11 translocated and 9 resident snakes implanted with VHF radio-transmitters. We calculated metrics like distance per move/per day, displacement, and area used (utilization distribution) to test for differences between translocated and resident snakes. Additionally, because consistent individual differences in behaviour (personality) can affect animal spatial movements and/or survival, we sought to identify the relationship between snake personality and movement behaviour in the field. We conducted standardised behavioural tests with 4-5 trials per individual to quantify variation in among-individual behaviour (personality) and within-individual behaviour (plasticity). The next stage of our research will quantify the degree to which personality/plasticity explains post-translocation movement and survival. By understanding sources of variation in venomous snake behaviour and individual responses to translocation, we can improve strategies for managing human-snake conflict along the urban edge.

Koalas in Southeast Queensland face increasing pressures from urban expansion, habitat fragmentation, and environmental change. These factors can significantly affect koala ecology including the natural dispersal behaviours of juvenile koalas within the landscape. This research investigates fine scale dispersal patterns of juvenile koalas from remnant, translocated, and receiving populations to assess how landscape modification and translocation practices affect movement ecology. Using GPS tracking data, detailed tree and site-level habitat variables, and environmental data (e.g., weather, vegetation structure), this study analyses dispersal distances, tree use, and movement strategies of juvenile koalas. Comparative spatial and statistical analyses will explore how factors such as gender, habitat composition, and connectivity influence dispersal success. Ultimately, the project aims to identify key barriers and facilitators to successful dispersal in urbanised environments, with findings intended to inform more effective conservation planning and translocation protocols. By advancing our understanding of juvenile koala movement in human-altered landscapes, this research will provide evidence-based strategies to enhance population resilience and ensure effective, informed actions for koala conservation.

In early 2020, bushfires on Kangaroo Island, South Australia, destroyed approximately 84% of the known koala habitat, displacing an estimated 5,000 individuals into fragmented native vegetation and unburnt Eucalyptus globulus (Tasmanian Blue Gum) plantations. These plantations initially served as critical refuges, but ongoing harvesting is now removing key food and shelter resources. With koala densities reaching up to five individuals per hectare, the risk of displacement and mortality has increased, particularly where surrounding native vegetation remains poorly regenerated or already overpopulated.
This study applies movement ecology to assess how koalas respond behaviourally to plantation harvesting in a post-fire environment. Using GNSS tracking collars, spatial analysis, and home range modelling, we evaluated movement patterns and tested the effectiveness of current forestry protocols in reducing disturbance impacts. Prior to harvest, males and females had significantly different home range sizes (mean = 4.4 ha and 1.2 ha respectively; p = 0.02). After harvesting began, home ranges increased significantly for both sexes (p = 0.001), with no remaining difference between them.
Movement across recently harvested areas was constrained by felled tree debris, with some individuals remaining in the same unharvested tree or moving less than 100 m until the area was cleared. Observational data revealed up to five koalas sharing single unharvested trees three months after harvest.
By analysing koala movement and displacement under the combined pressures of bushfire, habitat fragmentation, and industrial forestry, this research provides critical insight into the behavioural responses of koalas in plantation landscapes. The findings offer evidence to inform more effective forestry management practices and improve conservation strategies for koalas as they face increasing environmental stress.

Predictions of animal movement are vital for understanding and managing wild populations. However, the fine-scale, complex decision-making of animals can pose challenges for the accurate prediction of trajectories, and typically require high quality data that is representative of multiple individuals and environmental configurations. Animal movement data can be expensive and logistically challenging to collect, but there has been significant progress in other fields of leveraging knowledge gained from one task to improve performance on related tasks, referred to as transfer learning. We tested this for predicting animal movement by training a deepSSF model on GPS-tracking data from two similar species; water buffalo (Bubalus bubalis) and feral cattle (Bos indicus), across multiple regions in northern Australia. We designated a target site and assessed which combination of pre-training and fine-tuning led to the best model performance. In our case, pre-training allowed the model to learn fundamental principles of animal movement, as well as extract relevant environmental features, such as rivers, water bodies and the configuration of habitat patches, which was fine-tuned to the region of interest, improving overall performance. We discuss promising possibilities of a general model that has been pre-trained on many more regions, species and ecosystems that can be fine-tuned to few data.