Possibilities and limits of prospective GIS land cover modelling-a compared case study: Garrotxes (France) and Alta Alpujarra Granadina (Spain)

This study focuses on the possibilities and the limits of a prospective GIS land cover modelling applied to two case studies (France and Spain). The methodology, based on available GIS tools, consists of using earlier land cover maps and relevant environmental factors (calibration data) to model actual, known land cover to validate the model. The model aggregates Markov chain analysis for time prediction and multi-critera evaluation, multi-objective evaluation and cellular automata to perform spatial contiguity of modelled land cover scores. The first results give an accurate, pixel by pixel prediction rate of approximately 75%. An important issue of this study consists of analysing prediction residues to improve the model.