1. As human activities expand globally, there is a growing need to identify and mitigate barriers to animal movements. Fencing is a pervasive human modification of the landscape that can impede the movements of wide-ranging animals. Previous research has largely focused on whether fences block movements altogether, but a more nuanced understanding of animals’ behavioral responses to fences may be critical for examining the ecological consequences and prioritizing conservation interventions.

2. We developed a spatial- and temporal-explicit approach, Barrier Behavior Analysis (BaBA, available as an R package), to examine individual-level behaviors in response to linear barriers. BaBA classifies animal-barrier encounters into six behavior categories: quick cross, average movement, bounce, back-and-forth, trace, and trapped. We applied BaBA to wide-ranging female pronghorn (Antilocapra americana) and mule deer (Odocoileus hemionus) in an area of western Wyoming, USA with > 6,000 km of fencing.

3. We found both species were extensively affected by fences, with nearly 40% of fence encounters altering their normal movements, though pronghorn were more strongly affected than mule deer. On average, an individual pronghorn encountered fences 250 times a year – twice the encounter rate of mule deer. Pronghorn were more likely to bounce away from fences, whereas deer engaged in more back-and-forth, trace, and average movement near fences.

4. We aggregated these behavioral responses to demonstrate how BaBA can be used to examine species-specific fencing permeability and to identify problematic fence segments in order to guide fence modification or removal.

5. Synthesis and applications: Our work provides empirical evidence on how fences affect wildlife movement. Importantly, BaBA can be applied to other linear features (such as roads, railways, and pipelines) and habitat edges, enhancing our ability to understand and mitigate widespread barrier effects to animal movement.

Data files include: 

1) A shapefile (FENCE.shp) shows fence locations in the study area as depicted in Figure 1. 

2) A  CSV file (PRONGHORN.csv) that contains 20 pronghorn-year movement locations as depicted in Figure S1.

3) A CSV file (MULEDEER.csv) that contains 20 mule deer-year movement locations as depicted in Figure S1.

Code used to analyze these datasets can be found on GitHub: https://github.com/wx-ecology/BaBA