Weed Invasion Susceptibility Prediction (WISP) Model for Use with Geographic Information Systems

Abstract

The Weed Invasion Susceptibility Prediction (WISP) model was developed as an extension of the ArcView Geographic Information System to predict potential risk of invasion by individual weed species in rangelands. Existence potential was determined by comparing growth requirements of each weed species with respect to nine site characteristics obtained from geographic data layers: distance from water and disturbance sources, elevation, annual precipitation, soil texture and pH, aspect, slope, and land cover. Disturbance is important for predicting weed infestations; an innovative aspect of the WISP model is that we use data layers for transportation and energy development as an estimate of disturbance. Data from weed surveys conducted at the Jack Morrow Hills Study Area in southwestern Wyoming, USA, were used to test WISP model predictions for the occurrence of Hyoscyamus niger (black henbane), Cardaria draba (hoary cress), and Lepidium latifolium (perennial pepperweed). Data acquired from the U.S. Department of Interior, Bureau of Land Management were used to test model predictions for the occurrence of Euphorbia esula (leafy spurge) and Centaurea maculosa (spotted knapweed) for four counties in northwestern Wyoming, USA. The model accuracy, based on false-negative errors where a species was present but was predicted not to occur, averaged 89% for the five weed species. The WISP model allows land managers to predict more easily and accurately the potential for weed invasions in order to prioritize areas for detailed monitoring.