Land use Change Modeling Using Artificial Neural Network and Markov Chain(Case Study: Minab Esteqlal Dam Watershed)

One of the methods used by planners and managers to manage the land use change is its modeling. The aim of this research is to forecast the land use changes in Minab Esteqlal Dam Watershed using Multilayer Perceptron Network and Markov chain in 1409 horizon. The images of Landsat 5 TM (1995), Landsat 7 ETM+ (2003), and Landsat 8 OLI (2016) were used to prepare the land use maps. For this purpose, Maximum Likelihood Algorithm was used in three mentioned time periods. Transition potential modeling was done using Perceptron Neural Network and some of the static and dynamic variables and Markov chain was used to forecast the changes of land use in future. The parameters of GEOMOD method and Kappa statistics were used to evaluate the accuracy of the forecasting. Results of evaluating the calibration periods using GEOMOD methods and parameters of N (n), N (m), H (m), M (m), K (m), P (m), and P (p) and Kappa statistics indicated that the calibration period of 1995 to 2016 had the highest accuracy to forecast the land use in 2030. Results of the land use change in calibration period indicated that among six land use categories of forest, rangeland, agriculture, residential areas, bare lands, and water resources, the highest increase was related to the agriculture land use with an area of 627.05 Km2 and the highest decrease was related to the rangeland land use with an area of 580.35 Km2. Rangeland degradation has been done for developing the agricultural lands and residential areas. Also, results of land use change modeling for 2030 indicated that, agricultural areas has increased by 101128 Km2 and it has increased from 7.36% to 16.9% in the study time period. The area of rangelands decreased by 1000Km2 and reached from 56.8% to 47.4%. If the current trend continues until 1409, decreasing the area of rangelands and their conversion to agricultural lands will occur to increase land use productivity.