You are here:
Presentation--HABITAT DISTRIBUTION MODELS FOR 37 VERTEBRATE SPECIES IN THE MOJAVE DESERT ECOREGION OF NEVADA, ARIZONA, AND UTAH
Citation:
Boykin, K., D. F. BRADFORD, AND W. G. KEPNER. Presentation--HABITAT DISTRIBUTION MODELS FOR 37 VERTEBRATE SPECIES IN THE MOJAVE DESERT ECOREGION OF NEVADA, ARIZONA, AND UTAH. Presented at 2008 Annual Conference, Miami, FL, November 08 - 12, 2008.
Impact/Purpose:
Healthy Communities and Ecosystems - by providing ecosystem research and methods development at multiple scales, utilizing landscape ecology and associated techniques
Description:
Thirty-seven terrestrial vertebrate species in the Clark County Multi-Species Habitat Conservation Plan (MSHCP) were previously modeled through the Southwest Regional Gap Analysis Project (SWReGAP), using a deductive approach. To increase the applicability of such habitat models to the MSHCP, we revised these 37 deductive models specific to the Mojave Desert Ecoregion using additional information and finer scale datasets not available for the original SWReGAP models. We explored an inductive modeling approach using locality records for four species: desert iguana (Dipsosaurus dorsalis), common chuckwalla (Sauromalus ater), phainopepla (Phainopepla nitens), and desert kangaroo rat (Dipodomys deserti). Differences in extent of habitat predicted between the original and revised deductive models ranged dramatically, with the revised models for the majority of species predicting greater habitat extent than the original. A “gap analysis” was conducted by determining the extent of habitat predicted by the revised models within each SWReGAP conservation status category and within each MSHCP land management category. The results were similar for the two categorization schemes. For most species, the fraction of species’ habitat in the most protected SWReGAP conservation categories was higher for the Mojave Desert Ecoregion and Clark County than for species’ habitat throughout the 5-state SWReGAP region. For the four species addressed by inductive modeling, locality records were obtained from a number of sources. Inductive modeling was an iterative process using locations and SWReGAP datasets and more localized datasets specific to the Mojave Desert Ecoregion. Area under the Curve (AUC) values ranged from 0.750 to 0.951 for the four species with omission error rates ranging from 11% to 67%. Although the habitat distributions developed by the inductive models were within the same footprint as predicted by the deductive models, the inductive habitat models for three of the four species were more refined and differed considerably between the coarser deductive models.