Monitoring the changes of land cover of Kermanshah with Markov CA model

Different plans for sustainable urban development requires evaluation and accurate prediction of the scales of land use changes in urban areas. By reviewing these changes through satellite images, predicting them by modeling, and choosing effective managerial policies, we can step ahead in better town planning. The end of this investigation is to review, simulate, predict Kermanshah city expanding through combination of automatic cellular model, Markov chain and logistic regression. To achieve these purposes, firstly the land use maps of Kermanshah were provided by multi-time images of Landsat satellite of 1985, 2000, and 2013. As the second step, validation of these maps and manifestation of the changes was done. Manifestation of changes of the first period (1985-2000) and the second period (2000-2013) showed that urban areas has been expanded by 6541 hectares. According to these changes and selection of seven independent variables influencing the development of the city, modeling of the potential of land use change for the 2013 was done by regression logistic method. For the calibration of Markov chain CA model, land use map of 2013 was predicted and the error between the modeled map and the original land use map of 2013, show a kappa coefficient of 0.73. Then, using the Markov chain method, the estimate of possibility of changes and spatial distribution for 2026 was predicted. Finally, land use maps of 2026 was predicted through Markov chain CA model. Also, its results shows that a large amount of the lands of mountain areas, plant and rock cover