Simulation of land use change in Mashhad Metropolitan Area: Using Multi-Layer Perceptron and Markov chine model for ۲۰۰۰-۲۰۳۰

The purpose of this study is to simulate land cover and land use(LCLU) change and identify the driving factors affecting its changes in Mashhad Metropolitan Area(MMA) during ۲۰۰۰ to ۲۰۳۰. Multilayer perceptron (MLP) and Markov chain model (MLP-Markov) were used in TerrSet and GIS software. To this end, Landsat ۷ and ۸ satellite images for ۲۰۰۰, ۲۰۱۰ and ۲۰۱۹ were used. Six driving factors were used to identify the variables affecting the simulation of ۲۰۳۰ LCLU change. With maximum likelihood probability equation land uses were classified into ۹ categories. According to the results of the study, during the period ۲۰۰۰ to ۲۰۱۹, net positive changes of LCLU change in MMA was ۷۶۸۵.۴۱ha (۳۱.۳۷%) for residential, industrial, commercial and construction uses (RICUs). The remaining six land uses -۹.۳۷% decreased, including: Mass and light vegetation (-۱.۲۵%), uncultivated land (-۱۱.۲۲%), Barren land (sand and soil) (-۵.۴۰%) and mountainous land (-۳۰.۴۸%). To confirm the prediction of land use changes until ۲۰۳۰, the results of using the overall coefficient of Kappa (۰.۹۱۰۶) and ROC (۰.۵۵۵) were acceptable. The result of using Cramer's V statistics (۰.۹۳۶۱) to measure the effect of driving variables on land use forecast was approved. Predicting LCLU change for the ۲۰۱۹-۲۰۳۰ periods using MLP-Markov showed that RICUs will increase by ۱۷۷۹.۴۴ ha (۵.۵۳%). While the area of remaining six LCLU change will be reduced to -۲.۴۰%. In general, the distance from river, DEM, and industrial centers have played an important role in changing the sub-models of land uses. The spatial distribution of land use changes from ۲۰۰۰ to ۲۰۱۹ represents two types of patterns. Panned distribution in the form of city Master Plans on barren lands on the west side and mountain slopes on the south side of the city. Irregular and unplanned distribution on agricultural and barren lands along roads in the north, northwest and east of the city. Findings are useful for planning authorities in MMA where is highly affected by population growth and LCLU change

Keywords:

Driving variables , Mashhad , MLP- Markov chain , Projection