Modeling of urban development using cellular automata techniques and the Markov chain: A case study of Yazd

This study aims at the change of the land usage around the city of Yazd from 1994 to 2013 and the prediction of these changes until 2020. To this end, multi-temporal Landsat satellite images taken in three years of 1999,2006 and 2013 were studied and analyzed. The corresponding classified images were used to detect and characterize changes from a cross-tab. In order to predict the changes during these years and in future until 2020, two different models including CA-Markov and LCM were used. The results of this study showed that during these intended 14 years, the CA-Markov model could present a land use map with a 70% accuracy, and the final map for 2020 displayed the maximum usage of the land for residential purposes at the rate of 51.11%. On the other side Wasteland may decrease around 29%. The accuracy of the LCM model for the prediction of the land use map was estimated to be around 83%, and the final plan for 2020 exhibited a 68% increase of residential land and a 24.8% decrease of wasteland. However, the results demonstrated a better accuracy for the LCM model as compared to CA-Markov, which is the consequence of the use of independent variables as the input parameters in the LCM model. In general, it can be concluded that both models have almost the same accuracy and they are powerful methods for predicting the development of the city of Yazd in the intended period until 2020. Finally, the physical development of Yazd to the suburbs through the west, west-south and south regions is expected. This expansión tends toward the neighboring towns like Hamidia and Shahedieh.