Jonathan R. Potts

Predicting animal movement with deepSSF: a deep learning step selection framework

This paper introduces deepSSF, a deep learning framework for predicting animal movement by learning from environmental and temporal data, showing it can capture complex habitat interactions and temporal dynamics.

Predicting fine-scale distributions and emergent spatiotemporal patterns from temporally dynamic step selection simulations

This study shows that incorporating fine-scale temporal dynamics improves both hourly and seasonal spatial predictions of water buffalo movement and habitat use.

Variations in nonlocal interaction range lead to emergent chase-and-run in heterogeneous populations

The paper shows that varying interaction ranges between chasers and runners in heterogeneous populations leads to emergent chase-and-run dynamics at the population level.