108 Journal of Environmental Informatics ISEIS Journal of Environmental Informatics 19(2) 108-119 (2012) www.iseis.org/jei Generating a Future Land Use Change Scenario with a Modified Population-Coupled Markov Cellular Automata Model S. T. Y. Tong 1,* , Y. Sun 1 , and Y. J. Yang 2 1 Department of Geography, University of Cincinnati, Cincinnati, Ohio 45221, USA 2 National Risk Management Research Laboratory, U.S. Environmental Protection Agency, Cincinnati, Ohio 45268, USA Received 2 April 2011; revised 14 Feburary 2012; accepted 7 April 2012; published online 22 June 2012 ABSTRACT. With the population increasing and land use patterns changing, there will be environmental consequences. To solve these impending problems, information on the future land use pattern is needed. This study attempted to develop an enhanced land use model, capable of predicting future conditions. The traditional Markov model was modified by incorporating a Cellular Automata (CA) and a population variable to depict the neighboring effects and the impacts of population growth on urbanization. The performance of this new model was quantitatively assessed by generating the 2001 land use patterns of the East Fork Little Miami River watershed in southwest Ohio with and without the CA and the population variable and compared with the actual 2001 land use imagery. From the comparison, it was apparent that the land use map generated with the CA and population variable was more accurate. To further ascertain its applicability in a larger watershed, the same procedure was used to model the entire Little Miami River watershed. The validation results demonstrated that the performance of the modified CA-Markov model at both watershed scales was acceptable, and the inclusion of the CA and population variable could markedly improve model predictability. Based on these findings, the 2030 land use scenario for the LMR watershed was postulated. The resultant map showed much urban expansion in the western and southern portions of the basin. This information can be useful to planners and resource managers, enhancing their efforts in generating more sustainable future development strategies. Keywords: Markov, CA-Markov, population growth, land use modeling, urbanization, multi-criteria evaluation 1. Introduction Our world is changing rapidly in land use patterns; never in our history had we witnessed such a rate or magnitude of change. With the economic development, population growth, and in-migration, many places are experiencing expeditious ur- ban expansion and sub-urban sprawl, which can cause signifi- cant environmental consequences, such as changes in surface runoff and water quality (Tong et al., 2011). Moreover, future climate changes may further interact with these physical and socio-economic factors, exacerbating the transformation of lan- dscape and degrading environmental qualities (Lambin et al., 2001). With the anticipation of these changes and the associa- ted environmental problems, it is of paramount importance to be able to postulate the future land use conditions so that we can better plan for sustainable future developments. However, the prediction of future land use is often complicated as there are many intrinsic, inter-dependent, and interrelated socio-eco- nomic and biophysical drivers controlling the process of land use change (Parker et al., 2003). * Corresponding author. Tel.: +1 513 5563435; fax: +1 513 5563370. E-mail address: susanna.tong@uc.edu (S. T. Y. Tong). ISSN: 1726-2135 print/1684-8799 online © 2012 ISEIS All rights reserved. doi:10.3808/jei.201200213 This paper attempted to develop an enhanced Markov- based spatial dynamic modeling procedure to predict the pros- pective land use conditions. The goal was to improve the pre- diction accuracy of the original Markov land use model by coupling it with a Cellular Automata (CA) and the trend of po- pulation growth through Multi-Criteria Evaluation (MCE). The integration of a CA with Markov will take into account the nei- ghboring effects in calculating the transition probabilities (East- man, 2006). Additionally, since population growth is often an important socio-economic factor driving urban growth (Li et al., 2003, Liu et al., 2005), it is hoped that by considering po- pulation density in the study area, this modified Markov Cellu- lar Automata Land Use Change Model (CA-Markov) could ex- plicitly simulate the tendency of urbanization and suburban sp- rawl. Through the validation process, the efficacy of this enhan- ced population-coupled CA-Markov land use model in simula- ting urban expansion and in making realistic predictions of the future land use conditions was explored. 2. Methodology 2.1. Study Areas This research used the East Fork Little Miami River (EFLMR) watershed, a sub-watershed of the Little Miami Ri- ver (LMR) in southwest Ohio, as a pilot study to first develop an enhanced land use model. After the model was developed