Modelling land use dynamics in Luxembourg cross border region . The use of Cellular Automata and decision Tree Learning Model

Human activities (urbanisation, industry) and natural hazard effects (floods, drought, etc.) shape land use and land cover features over space and time. Stakeholders need planning tools to develop land policy strategies (e.g. zoning regulations) and to improve various aspects in urban planning. Therefore in the last decades, they relied on scientist to provide them with efficient decision support tools. In this paper, a decision support system is presented that aims at helping stakeholders in forecasting future changes in land use, which will result in better planning. This system consists of an integrated model combining Cellular Automata (CA) and decision tree learning. The later is used to define the transition rules of the former. The results from applying this model on Luxembourg and its cross-border regions show that the integrated model performed very well in predicting the change in land use and in detecting its patterns. This model also proved to be a useful for finding out why and where changes in land use occur?