Multi-scenario simulation of land use and land cover based on shared socioeconomic pathways: The case of coastal special economic zones in China.

Urban land-use change simulations without considering the sustainable planning policies, especially in special economic park highly concerned by planners, might lack the reliability and availability. Thus, this study proposes a novel planning support systems integrating the Cellular Automata Markov chain model and Shared Socioeconomic Pathways (CA-Markov-SSPs) for predicting the changing of land use and land cover (LULC) at the local and system level by using a novel machine learning-driven, multi-source spatial data modelling framework. Using multi-source satellite data of coastal special economic zones from 2000 to 2020 as a sample, calibration validation based on the kappa indicates a highest average reliability above 0.96 from 2015 to 2020, and the cultivated land and built-up land classes of LULC is the most significant changes in 2030 by using the transition matrix of probabilities, the other classes except water bodies continue to increase. And the non-sustainable development scenario can be prevented by the multiple level collaboration of socio-economic factors. This research intended to help decision makers to confine irrational urban expansion and achieve the sustainable development.