Assessment of desertification vulnerability using soft computing methods

Abstract

In this work Artificial Neural Networks and Genetic Programming are applied in order to assess the desertification status, a kind of land degradation, of an area, from meteorological and land use data. The approach has been tested in the Sannio (central Italy) region. Both the used soft computing methods show low error rates, and the Genetic Programming offers the advantage of an explicit representation of the factors that favour or delay the desertification. This methodology allows preventive actions to face the upcoming desertification.