Evaluation of prospective modelling methods: fuzzy logics and cellular automaton applied to deforestation in Venezuela

Abstract

Spatial evolutions of anthropized ecosystems and the progressive transformation of spaces through the course of time emerge more and more as a special interest issue in research about the environment. This evolution constitutes one of the major concerns in the domain of environmental space management. The landscape evolution of a regional area and the perspectives for a future state raise particularly important issue. What will the state of the region be in 15, 30 or 50 years? Time can produce transformations over a regional area such as emergence, disappearance or the union of spatial entities. These transformations are called temporal phenomena. We propose two different methods to predict the forestry development for the forthcoming years in the experimental area, which reveals these spatial transformations. The proposed methods are based on fuzzy logic and Cellular Automata (CA). The methods are supported by the analysis of the landscape dynamics of a test site located in a tropical rain forest country: the oriental piedmont of the Andes Mountains in Venezuela. This large area, at the scale of a Spot satellite image, is typical of tropical deforestation in a pioneer front. The presented approaches allow the geographer interested in environmental prospective problems to acquire type cartographical documents showing future conditions of a landscape. The experimental tests have showed promising results.