Land Use /Cover Change Monitoring and Prediction Using Markov Chain (Case Study: The Abbas Plain)

Remote sensing is a key technology for assessing expansion and rate of land cover changes that awareness of these changes as the basic information has a special importance for various programs.
In this study, land use changes were examined over the past 24 years, and the feasibility of predicting it in the future was evaluated by using the Markov chain model of the Abbas Plain. Landsat TM, ETM, and OLI satellite images for the years 1968, 2003 and 2013, respectively; along with topographic and vegetation maps of the study region were used in this research. The images for three periods were classified into five land-use classes of rangeland, agricultural land (irrigated and rain-fed)), residential land, riverbed and barren and hilly land. According to the results, agricultural land is the most dynamic land-use class in the study area and its area has followed an upward trend during the period 1968 – 2003, so that 4337 ha (7.12%) has been added to this land-use class during this period. The trend of rangeland use change has had a descending trend during the period 1968 – 2003, so that has caused its area to be decreased by 3.19% (6573.6 ha) during this period. The results obtained from Markov chain analysis in the period 1968-2003, for model calibration; the maps for the years 1968 and 2003, and its matrix for predicating land use changes of the year 2023 indicate the Kappa coefficient equal to 80 percent. Based on the obtained results, in the year 2023, 49.1 and 10.1 percent of the study region are comprised of agricultural land and rangeland, respectively.