Skip to Main Content
Because of its strong ability in capture the nature of complex systems, cellular automata (CA) have been widely used to simulate suburban land use changes and urban pattern evolution. Driving forces impacting suburban land use changes could be derived from the multi-temporal remotely sensed imageries to retrieve the transition rules of geographic CA (geo-CA). With the images of two stages ranging 16 years, a CA model based on logistic regression was accurately calibrated. Under a GIS environment, the suburban land use changes in Minhang district of Shanghai are simulated with this geo-CA model from 1992 to 2008. Confusion (error) matrix and the Kappa coefficient between the simulation results generated by the Geo-CA model and the classified images from remote sensing images were calculated. This research has proved that the Geo-CA models are especially appropriate for modelling land use and cover changes of suburban areas.