OOS 25-9
Predicting the impact of fire on a vulnerable multi-species community with a dynamic vegetation model

Wednesday, August 13, 2014: 10:50 AM
307, Sacramento Convention Center
Erin E. Conlisk, Energy and Resources Group, University of California, Berkeley, Berkeley, CA
Alexandra D. Syphard, Conservation Biology Institute, La Mesa, CA
Janet Franklin, School of Geographical Sciences and Urban Planning, Arizona State University, Tempe, AZ
Helen M. Regan, Biology, University of California Riverside, Riverside, CA
Background/Question/Methods: Models that synthesize data from disparate sources to make management decisions for multiple species are important conservation tools.  In this paper, we describe a simulation model that ranks a range of fire management strategies from the viewpoint of protecting endangered coastal southern California wildlife.  The model involved interactions between fire, four animal species (Stephens' kangaroo rat, Pacific pocket mouse, coastal cactus wren, and California gnatcatcher), and five vegetation types (grass, coastal sage scrub, obligate seeding and resprouting chaparral, resprouting-only chaparral, and woodlands).  Based on historic fire in the region, we predicted spatially-explicit fire frequencies and ignition probabilities.  With these predictions we simulated the location and extent of fires annually.  Given a site’s simulated fire history, vegetation type conversion was simulated using fire frequency-dependent vegetation transition data from 1933-2003.  Fire occurrence in a location altered habitat suitability directly for each of the endangered species and indirectly by changing the vegetative community.  For some species, such as the endangered Stephens' kangaroo rat and Pacific pocket mouse, fairly frequent fire is required to maintain open habitat.  For other species, such as the coastal cactus wren and endangered California gnatcatcher, frequent fire destroys the mature coastal sage scrub vegetation on which these species depend.  The model included a management component representing fire control decisions, allowing us to rank fire control policies. 

Results/Conclusions .  Over the 50-year modeled time horizon, we found that Stephens’ kangaroo rat and Pacific Pocket Mouse populations were highly variable and prone to extirpation at the site, despite prescribed burns designed to boost their viability.  California gnatcatchers and cactus wren were rarely extirpated in the model and showed much less variability in model outcomes across simulation runs.  The primary finding was that reducing wildfire in coastal sage scrub was the single most beneficial strategy for all four representative animal species.