Approach to land cover change modelling using cellular automata

Nowadays one of the most effective and widely used ways for the land cover change modelling is an application of cellular automata (CA). One of the key factors in influencing on the effectiveness of the CA application is defining the CA transition rules in each specific case. The paper proposes a compositional probabilistic approach to defining CA transition rules with more complex nature compared to the traditional approach to the CA land cover change modelling. The preliminary investigation results show that proposed approach allows performing the modelling more effectively and its results could be more accurate compared to ones obtained by existing models of land cover change. The accuracy estimation of the proposed approach is conducted using model time series thematic maps and truth thematic maps obtained from the time series satellite images RESURS-O1 for the Uymon steppe area (Altay region, Russia) on 1998, 1999, 2000.