How much does animal behavior matter in corridor planning?

When relying on a focal species or suite of species for corridor planning, we know that accurately identifying functional corridors depends on our understanding of those species’ dispersal requirements. But do our typical estimates of resource selection actually represent dispersal?

Currently, the most widely used technique to estimate landscape resistance is to use resource selection functions (RSFs), which focus corridor planning on areas at fine spatial scales where dispersal is predicted to occur. Because dispersal events are often difficult to detect in the field, resource selection measured during directed movement states may act as a proxy to infer functional connectivity in lieu of direct dispersal data. However, RSF-derived predictions of landscape connectivity need to be validated with long-distance movement data to assess the validity of this inference.

We tested the efficacy of behavior-specific predictions of connectivity with long-distance dispersal movements of African wild dogs (Lycaon pictus), an endangered species highly sensitive to habitat fragmentation. We used activity data to classify the movement data we collected into three behavioral states: traveling, resting, and hunting. By pairing behavioral state data with GPS data, we could measure resource selection specific to each behavioral state. We were specifically interested in the differences between resource selection measured during the traveling state (‘Movement model’) versus resource selection measured without consideration of behavioral state (‘Combined model’). We found very different estimates of landscape resistance between the ‘Combined model’ (left) and ‘Movement model’ (right).


We then created least-cost corridor predictions from the two RSF-derived resistance models. When we tested these predictions against observed long-distance dispersal paths, we found that the ‘Movement model’ corridors protected 87% of the dispersal movements, whereas the ‘Combined model’ corridors only protected 33% of dispersal movements. These results suggest that behavior-specific resource selection is an important consideration in corridor planning.

We also conducted a systematic literature review of RSF-based connectivity studies to evaluate the extent to which these efforts have incorporated directed-movement behavior and to identify best practices for considering movement behavior for future connectivity studies. Our review helped us create several recommendations to target directed-movement behavior for connectivity analyses:

  • During data collection, collect data from dispersers or during a known dispersal season.
  • For GPS or telemetry data, infer directed-movement behavior by analyzing movement rate, step length and turn angle distributions, or activity loggers.
  • For sign data, exclude locations with sign of resource-use behavior like nesting or feeding.
  • For camera trap data, infer movement rates from spatial capture-recapture analysis.


Abrahms, B., Sawyer, S. C., Jordan, N. R., McNutt, J. W., Wilson, A. M. and Brashares, J. S. 2016. Does wildlife resource selection accurately inform corridor conservation? Journal of Applied Ecology. doi:10.1111/1365-2664.12714.

2016-10-14T10:10:24+00:00 June 16th, 2016|

About the Author:

Briana Abrahms
Briana Abrahms is a spatial wildlife ecologist with emphasis on mitigating the impacts of environmental change on wildlife communities. She is a postdoctoral researcher in the department of Environmental Science, Policy and Management at the University of California, Berkeley. Her research combines GPS tracking, behavioral observation, remote sensing, and simulation modeling to develop novel methods for understanding and restoring connectivity for wildlife movement.