APPLICATION OF GIS AND REMOTE SENSING FOR PREDICTING LAND-USE CHANGE IN THE FRENCH JURA MOUNTAINS WITH THE LCM MODEL: THE IMPACT OF VARIABLES ON THE DISTURBANCE MODEL

摘要

This research studies the land-use change for the Ain watershed in the Jura Mountains, in the East of France. Using the satellite images LANDSAT in 1975, 1992, 2000 and 2010, the land-use data for the four corresponding years was generated in the first place, and the diachronic land-use analysis from 1975 to 2010 was next carried out. In order to generate the four land-use maps, the satellite images were classified using the Maximum Likelihood supervised classification. The land-use map in 2010 is validated based on the observations from ground visits in 2011. Other land-use maps are validated by the ground truth data established from the auxiliary information layers taken in the area. Based on the trend in the past evolution of land-use, a prediction of future land-use is generated, using the Land Change Modeler (LCM). Next, the multi-temporal analysis of land-use change is carried out and the variables affecting the LCM are evaluated using Multi-Layer Perception and Markov transition probabilities. Two pairs of maps (1975, 1992) and (1992, 2000) are used to generate the predictive maps.