Land Cover Change Modeling based on Artificial Neural Networks and transmission potential method in LCM (Case Study: Forests Gilan-e Gharb, Kermanshah Province)

In order to land cover change modeling and detect to possibility of predict the future trend of Land Change modeler (LCM) was used. VNIR Data ASTER Sensor of TERRA satellite with spatial resolution of 15m for three periods 2000, 2007 and 2016 from Gilan-e-Gharb forests of Kermanshah province were analyzed. Land cover maps of years 2000, 2007 and 2016 four categories: forest cover, pasture lands, agricultural lands and built-up area areas for each of images were extracted. The results of data analysis in the first period (2000-2007) and the second period (2007-2016) showed the greatest increase in agricultural lands and pasture lands have the greatest decrease area. Based on these changes and by taking eight independent variable, transition potential modeling of 2016 was done using Artificial Neural Network. Then by hard predict model and images were classified of first period (2000- 2007), the land cover map in 2016 using Land Change Modeler was predicted. After evaluating the model, 83.09 and 71.10 overall accuracy was obtained for the first and second periods showed the consistency between prediction map and classified map of year 2016. The land cover maps by entering the second period (2007-2016) to Land Change Modeler the land.