Until now, few studies have used the mainstreaming models to simulate the land use changes in the cities of rapid urbanizing regions. Therefore, we aimed to develop a methodology to simulate the land use changes in rapid urbanizing regions that could reveal the land use change trend in the cities of the regions. Taking the urban areas of Wuhan, a typical rapid urbanizing region in China, as the study area, this study built a Markov chain-artificial neural network (ANN)-cellular automaton (CA) coupled model. The model used land use classification spatial data with a spatial resolution of 5 m in 2010 and 2020, obtained by remote sensing image interpretation, and data on natural and socio-economic driving forces for land use change simulation. Using the coupled model, the land use patterns of Wuhan urban areas in 2020 were simulated, which were validated in comparison with the actual land use data in 2020. Finally, the model was used to simulate the land uses in the study area in 2030. The model validation indicates that the land use change simulation has a high accuracy of 90.7% and a high kappa coefficient of 0.87. The simulated land uses of the urban areas of Wuhan show that artificial surfaces will continue to expand, with an area increase of approximately 7% from 2020 to 2030. Moreover, the area of urban green spaces will also increase by approximately 7%, while that of water bodies, grassland, cropland, and forests will decrease by 12.6%, 13.6%, 34.9%, and 1.3%, respectively, from 2020 to 2030. This study provides a method of simulating the land use changes in the cities of rapid urbanizing regions and helps to reveal the patterns and driving mechanisms of land use change in Wuhan urban areas.
Keywords: Markov–ANN–CA model; driving force; dynamic land use change; remote sensing image interpretation; simulation-based prediction.