Comparison of three hybrid models to simulate land use changes: a case study in Qeshm Island, Iran

Land use change simulation is an important issue for its role in predicting future trends and providing implications for sustainable land management. Hybrid models have become a recognized strategy to inform decision-makers, but further attempts are needed to warrant the reliability of their projected results. In view of this, three hybrid models, including the cellular automata-Markov chain-artificial neural network, cellular automata-Markov chain-logistic regression, and Markov chain-artificial neural network, were applied to simulate land use change on the largest island in Iran, Qeshm Island. The Figure of Merit (FOM) was used to measure the modeling accuracy of the simulations, with the FOMs for the three models 6.7, 5.1, and 4.5, respectively. Consequently, the cellular automata-Markov chain-artificial neural network most precisely simulates land use change on Qeshm Island and is, thus, used to simulate land use change until 2026. The simulation shows that the incremental trend of the built-up class will continue in the coming years. Meanwhile, the areas of valuable ecosystems, such as mangroves, tend to decrease. Despite the protection plans for mangroves, these areas require more attention and conservation planning. This study demonstrates a referential example to select the proper land use models for informing planning and management in similar coastal zones.