Predicting locational trend of land use changes using CA-Markov Model (Case study: Kohmare Sorkhi, Fars province)

Land use changes act as a significant factor in the environmental changes and have become a global threat. Monitoring and prediction these changes by satellite images and models can help the planners and managers to make more conscious planning decisions. In this regard, the current research aimed to monitor, model and predict land use changes using CA-Markov model in Kohmare Sorkhi region, Fars province in 2024 for a period of 25 years (1987-2012). To implement the mentioned model, the land use map was first prepared by ETM and TM sensors during three years (1987, 2000, 2012). Then, validation of maps and change detection process were performed. The results of change detection for the first period (1987-2000) and second period (2000-2012) with an accuracy of 83% and Kappa index of 88% have shown the greatest increase in the rangeland area (4224.24 ha) and the greatest decrease in the forest area (3953.75 ha). In the next stage, in order to calibrate the CA-Markov model, land use map for 2012 was predicted; on the other hand, regarding Error Matrix between the modeling land use map and the reference land use map, the Kappa index wad given as 75%. Finally, considering the previous stage, the land use map for the outlook of 2024 was predicted. The final results for 2024 indicated that the forest area would endure the great amount of changes in comparison with 2012. The forests would change into the irrigated agricultural area and rangelands which can be considered in sustainable development planning by decision makers.