GARP modeling of natural and human factors affecting the potential distribution of the invasives Schismus arabicus and Brassica tournefortii in 'El Pinacate y Gran Desierto de Altar' Biosphere Reserve

摘要

Invasive plant encroachments threaten biodiversity and ecosystem viability worldwide. Effects of invasives on native Sonoran Desert ecosystems are, for example, a growing concern among ecologists and land managers. Potential distributions of the exotic Brassica tournefortii and Schismus arabicus were modeled for 'El Pinacate y Gran Desierto de Altar' Biosphere Reserve (PBR) in Sonora Mexico using genetic algorithms within a geographic information system (GIS) database. Spatial models were created using known presence points for B. tournefortii and S. arabicus, and the merged and separated influence of natural and human-related variables, respectively. Best performing models were summed to produce probability models for each species and suite of variables. Models derived from human related variables showed higher significance and accuracy values for both invasives. Combined probability models were reclassified to isolate areas of high-predicted presence, which served to identify individual variables that favor most the potential occurrence of invasives. Variables showing greatest predictive skill were isolated and recombined to produce models of merged conditions of high probability presence. Internal road networks were strongest predictors of presence, revealing the potential importance of humans as vectors of invasiveness in the PBR. (C) 2007 Elsevier B.V. All rights reserved.