A Bayesian Approach to Characterizing Habitat Use By, and Impacts of Anthropogenic Features On, Woodland Caribou (Rangifer tarandus caribou) in Northeast British Columbia

Posted on Jun 22, 2015

Author

Steven F. WILSON and Craig A. DEMARS

Correspondence: Steven F. Wilson, EcoLogic Research, 302-99 Chapel Street, Nanaimo, British Columbia, V9R 5H3, Canada. Email: steven.wilson@ecologicresearch.ca

Submitted 18 February 2015 – Accepted 22 June 2015


Abstract

Woodland caribou (Rangifer tarandus caribou) ranges in northeast British Columbia (BC) have been altered by industrial forestry and oil and gas development. In addition to numerical and functional changes in the predator-prey system, these habitat changes might be altering habitat use by caribou. Our objective was to understand impacts of anthropogenic disturbances and ire on the composition of woodland caribou home ranges to inform decisions about where, and to what extent, diferent habitats need to be restored to serve as efective caribou habitat. We used a non-parametric Bayesian analysis to model selection of habitats in their current condition, and then estimated selection in an hypothesized landscape without anthropogenic features and ire to identify potential areas for habitat restoration in the Parker woodland caribou range of northeast BC. Treed bogs and poor fens were selected by woodland caribou among all seasons analyzed (based on comparisons with a random dataset) and areas of high linear feature density were avoided, particularly in late fall and late winter. Unexpectedly, areas closer to early seral/immature forest habitats were selected. he slopes of selection curves were relatively shallow and model precision (i.e., proportion of correct predictions) was <60% for all seasons, suggesting that woodland caribou were not particularly discriminating in their selection of home ranges, at least with respect to the factors analyzed. Anthropogenic features and ires negatively afected 43-75% of the Parker range and selection of these areas by woodland caribou was 8-11% lower than expected compared to the hypothesized landscape without anthropogenic features and ire, depending on season. he models generated by the Bayesian approach were more intuitive than traditional parametric methods and provided a platform for generating predictions regarding the efectiveness of mitigation or restoration activities.

Key Words: Anthropogenic Disturbance, Bayesian Approach, Boreal Forest, Habitat Use, Woodland Caribou

 

View Full-Text    Download PDF

Top