Cellular automata model as an intuitive approach to simulate complex land-use changes: an evaluation of two multi-state land-use models in the Yellow River Delta

• Published in: Stochastic environmental research and risk assessment 2013-05, Vol.27 (4), p.899-907
• Main Authors: Ding, Wen-Juan, Wang, Ren-Qing, Wu, Da-Qian, Liu, Jian
• Format: Article
• Language: English
• Subjects: Aquatic Pollution; Chemistry and Earth Sciences; Computational Intelligence; Computer Science; Earth and Environmental Science; Earth Sciences; Environment; Environmental impact; Land use; Math. Appl. in Environmental Science; Neural networks; Original Paper; Physics; Probability Theory and Stochastic Processes; Regression analysis; Rivers; Statistics for Engineering; Stochastic models; Waste Water Technology; Water Management; Water Pollution Control
• ISSN: 1436-3240; 1436-3259
• DOI: 10.1007/s00477-012-0624-7

Land-use changes are generally recognized as multi-scale complex systems with processes and driving factors operating at different scales. Traditional linear approaches could not adequately acquire the nonlinear features in complex land-use changes. A multi-state artificial neural network based cellular automata (MANNCA) model and a multi-state autologistic regression based cellular automata (MALRCA) model were developed to simulate complex land-use changes in the Yellow River Delta during the period of 1992–2005. Relatively good conformity between simulated and actual land-use patterns indicated that the two models were able to simulate land-use dynamics effectively and generate realistic land-use patterns. The MANNCA model obtained higher fuzzy kappa values over MALRCA model at all the three simulation periods, which indicated that artificial neural networks could more effectively capture the complex relationships between land-use changes and a large set of spatial variables. Although the MALRCA model does have some advantages, the proposed MANNCA model represents a more effective approach to simulate the complex and nonlinear land-use evolutionary process.