Grizzly Bear Simulation Code Examples

This software release contains the core code and example data for generating animal movement simulations using step selection functions (SSFs). The approach is based on simulation methods used in a series of grizzly bear studies by Sells et al. (2022–2024). In general, an SSF is fit to animal movement data and associated environmental covariates, and the fitted model is used to simulate new movements.
Simulations are presented in two forms:
Undirected simulations: The animal begins at a specified location but has no predetermined endpoint. Movements are governed solely by the SSF.
Directed simulations: The animal begins at a starting location and is given a destination. These simulations use the randomized shortest path (RSP) algorithm (Panzacchi et al. 2016), in conjunction with the SSF, to simulate a path between start and end points.
Directed simulations introduce additional assumptions (e.g., that an animal knows where it is going), which may not be realistic for certain behavioral contexts such as dispersal. Nearly all simulations in the Sells et al. papers used the undirected approach, with the exception of the 2023 connectivity paper, which used both.
For more background and links to manuscripts, visit:
https://www.umt.edu/coop-unit/sellslab/sells-research/grizzly-bears.php