Simulation and Analysis of Animal Movement Paths Using Numerus Model Builder

We address the question of how best to fit animal movement paths, represented by point relocation time series, to a novel stochastic walk model-referred to as M-cubed-in a way that captures movement patterns at several different spatio-temporal scales. We test our approach on simulated data obtained from a high-frequency, multi-mode model constructed using the Numerus Model Builder platform. The advantage of using simulated over empirical data is that we know the processes generating movement. Fitting M-cubed to data requires that we extract movement modes and phases at various scales ranging from subhourly to daily. After fitting the M-cubed model, we evaluate how well it captures movement patterns and find that it performs exceptionally well at minute to hourly scales but needs to be extended to capture daily scale patterns driven by particular landscape features.