Multiple factors including natural and human-induced ones lead to land cover change in the landscape. Therefore, identifying the pattern of land cover change can help inform land-use management and prevent associated issues which can affect the natural resources of the landscape. The aim of this study is to assess land cover change in the Qeshm Island in southern Iran by combining the resulting outputs of multiple modeling methods, cellular automata (CA), Markov chains, and artificial neural networks (ANN) based on land cover maps for the years 1996, 2006, and 2016 that have been extracted from satellite imagery (Landsat 5, 7, and 8). In order to evaluate the accuracy of modeling, the Kappa coefficient was calculated to be 0.8. Then, land cover changes for 2025 were predicted by a hybrid model (CA-Markov-ANN). The results indicate that the classes of built-up areas, vegetation, and mangrove forests have changed more significantly from 1996 to 2016 compared with other classes. Land cover maps generated in this study showed that built-up areas have grown significantly in recent decades due to the region's growing population and development of ports, commercial, and industrial areas. Due to the climate change, the land area covering vegetation has decreased dramatically. The size of the mangrove forests has increased over the time period of the study (1996-2025). The findings of this study can inform land-use planning decisions by providing them with a comprehensive overview of land cover conditions in the future.
Keywords: Artificial neural networks; Cellular automata; Land management; Markov chains; Qeshm Island.