Monitoring and predicting land use changes using LCM module
(Case study: Marivan region)
Document Type : Research Paper
Detection and prediction of land use changes are powerful tools in natural resource and ecosystem management. This study aims at monitoring and predicting land use changes using LCM module in the western Marivan region. Landsat images dated 1989, 2000 and 2011 were classified in order to generate digital land use maps. The images were classified into four classes including forest, agriculture, water bodies and built-up area. LCM module in Idrisi GIS software is used to evaluate the land use changes and predict the land uses status in 2011, based on ANN and Markov Chain Analysis. ANN was trained with various spatial variables including distance from roads, distance from residential areas, distance from forest edges, land uses, elevation and aspect. The results indicated that 1234 hectares of the forest area have been reduced during the period of 1989-2011 and the deforestation rate was 0.21 % per year. Moreover the built-up areas have been increased 2.46% (924 ha) in comparison to initial situation. Results indicate very dynamic changes in agricultural areas, as they showed 1066 hectares increase and simultaneously 777 hectares decreasing, so in overall 289 hectares have been increased. The comparison of actual and predicted land use change maps, during the period of 1989-2011, indicates that Kappa coefficient for forest, agricultural and built-up areas were 0.37, 0.50, and 0.48, respectively. Based on the obtained results,PredictingLanduse changes using LCM was weak in this study area. To study the role of other variables such as soil types, forest types and socio-economic information to improve the performance of the model is recommended.