Study on species invasion warning modeling using GIS and data mining

Biological invasion has been one of the most dramatic ecological even in human history that threatens our economy, public health and ecological integrity. GIS and Remote Sensing technology should be integrated with spatial data mining to recognize the patterns of invasive species over space and time and predict the distribution at the large-scale. Presented with the challenge of problems during the prediction modeling including the uncertainty in biodiversity data, the uncertainty in model selection, and the uncertainty in niche cross the geographic space, this paper used information-theoretic approaches based on a set of GIS/RS environment layers to generate two kinds of species invasion warning models: global species invasion warning model (G-SIWM) and local species invasion warning model (L-SIWM) and illustrated the approach through a habitat-suitability analysis of ragweed (Ambrosia artemisiifolia L.).