Simulation and prediction of land use change in Dongguan of China based on ANN cellular automata - Markov chain model

Abstract

Simulation and prediction of land use cover change (LUCC) in cities can effectively identify trends in land conversion and support the decision-making process regarding future urban development. To identify urban LUCC and the effects of its inherent climatic, geographical, and socioeconomic drivers, this study developed a hybrid method for predicting future land type area based on a Markov chain model and simulating the spatial distribution of land use based on an artificial neural network, the cellular automata approach, and the Land Expansion Analysis Strategy. The proposed method was shown effective in the following aspects: (1) identification of land use variation, (2) simulation of future land use change distribution, and (3) analysis of the contribution of different driving factors to LUCC. The approach was applied to the city of Dongguan. The land use patterns changed in a balanced manner, according to the movement distances of gravity center during 2010–2020. Compared with the land use pattern in 2020, the area of ecological land was projected to diminish by 0.67 %, 1.42 %, and 2.07 % in 2025, 2030, and 2035, respectively. Based on the simulation and prediction of the land-use areas in Dongguan, the area of woodland would reduce by 7.42 %, and the growth rate of building land would be slow. It was recommended that a conservation district should be established, especially in the central mountainous region of Dongguan.

Keywords

• Land use cover change
• Land use patterns
• ANN-CA
• Driving factors
• Land use prediction
• Dongguan