Resumen
Habitat occupancy models were developed for 10 vertebrate species that we expected would demonstrate a gradient of response to extensive losses of lodgepole pine (Pinus contorta) and other linked habitat alterations resulting from the mountain pine beetle (Dendroctonus ponderosae) infestation and gradual changes in regional climate. A process-based Bayesian Belief Network approach was used to develop interlinked species models focussed at two levels of land management: (1) the forest stand level including changes in forest overstorey and understorey species composition, within-stand structures, canopy closure, and amounts of standing and fallen deadwood; and (2) the landscape level including changes in size of habitat patches, seral stage composition, and proximity to roads. We also considered indirect influences of broad ecological changes including alteration of some key species interactions (e.g., displacementfrom preferred habitat and [or] increased risk of mortality). We used results of this modelling to provide preliminary predictions of species occupancy in a large area of British Columbia designated by the Nature Conservancy of Canada as their Central Interior ecoregion. This work demonstrates an approach to building species occupancy models capable of representing the effects of large-scale disturbances on habitat supply at both the stand and landscape levels of habitat management. The resultant occupancy maps are also useful when integrated into various strategic planning initiatives including species recovery, silvicultural investments, and long-term conservation planning.