Abstract
This article presents a multi-temporal uncertainty-based method that incorporates a statistical regression model with a view to establishing the risk (probability) of land cover changes as a function of a set of environmental and socio-economic driving factors. The morphologic, climatic and socio-economic variables were examined using an Artificial Neural Network (ANN) model and the Multi-Layer Perceptron (MLP). Following the analysis, maps indicating the suitability to future changes were generated on the basis of observed transitions. From these maps two possible land use scenarios were built, applying the Markov chain principle. The region of Basilicata, in southern Italy, was selected for the analysis. The results highlight: a) a good inclination to change towards specialised crop systems, provided there is sufficient water supply; b) that some cropping patterns are not suitable for changes, partly because they are found in a context with severe limitations for alternative uses.