A Bayesian approach to evaluating habitat for woodland caribou in north-central British Columbia

Abstract

Woodland caribou (Rangifer tarandus caribou (Gmelin)) populations are in decline throughout much of their range. With increasingly rapid industrial, recreational, residential, and agricultural development of caribou habitat, tools are required to make clear, knowledgeable, and explainable management decisions to support effective conservation of caribou and their range. We developed a series of Bayesian belief networks to evaluate conservation policy scenarios applied to caribou seasonal range recovery areas. We demonstrate the utility of the networks to articulate ecological understanding among stakeholders, to clarify and explicitly depict threats to seasonal range. We also show how simulated forecasts of spatially explicit seasonal range can be compared with landscape potential with range under assumed conditions of natural disturbance. These tools have provided opportunities to operationally define and measure conditions for recovery of caribou in north-central British Columbia.