Predicting Land Use and Land Cover Changes on Sand Dunes Expansion Using CA-Markov Model (Case Study: Southeastern Coastal Desert of Iran)

Due to the intense and unprincipled changes in land use, it is necessary to detect land cover/use change over time using satellite imageries for sustainable land management. This study aims to monitor land use and land cover changes, model and predict their status in the next 15 years in Jagin plain, Jask coastal desert. In this study, Landsat satellite images of 1987, 2003 and 2018 were used to detect changes. After applying geometric and atmospheric corrections on the images, land use maps  of each year at 11 land use /land cover classes including farmlands, aquaculture, erg, halophyte, mangrove, habitable zones, nebka, vegetation affected by water flow vegetation (rangeland), bare lands, were prepared using hybrid classification method. The Markov chain forecasting model was used to validate the model in the first stage, and was used in the second stage to predict land cover/land change in 2033. Comparison of the simulated map of 2018 and the real map with the Kappa coefficient of 83% confirmed the capability of the Cellular Automata Markov Chain in predicting of land cover/land changes. According to the results, in 2032, agriculture lands in the region will have an upward trend and grow by 74% compared to 2018. Habitable zones, marine access network and aquaculture lands will also increase during the study period. In the study area, vegetation and riverbed vegetation areas show a downward trend and will mainly converted to agricultural lands. Sand dunes, which are among the main threats to people and mangrove forests, will be extended by 2032, threatening the southern part of Sorglam village and the eastern part of the mangrove forest.