Urban Growth Modeling in Bojnurd by Using Remote Sensing Data (Based on neural network and Markov modeling changes of land)

In the recent decades, population growth, increasing urbanizationand invading to agricultural land have become a serious environmental problem. Change assessment of Land use/land cover (LULC) is receiving considerable attention in the prospective modeling domain. The study's purposes refers to analyze and predicte of LULC change by using Land Change Modeler and the neural network with an integrated Markov model in 2032. MLP neural network was used for generating LULC maps by using Landsat images from 2005 to 2014. The overall accuracy and kappa coefficients of the maps were up to 82%. The accuracy of transition potential modeling showed high accuracy more than 95.2% in all sub-models. According to the results, the most salient increase was in urban areas from 1529.38 ha in 2005 to 1837 ha in 2014. This ascending trend will continue in the future and will increase to 2856.31 ha of the total area by year 2003. In conclusion, the study revealed that such models were useful for recognizing the Spatio-temporal LULC change.