A Study Of Simulation And Forecast Of Urban Expansion Based On CA-MARKOV Model

The city we living is a complex system, the change of the city and its surrounding land are not only affected by the situation of vegetation cover and other natural conditions, but also by economic, population, transportation network and other social conditions. Xining will enter into the phase of rapid urbanization in the next ten years, as the economy developing, the population of xining will expand, and the scale of urban and the infrastructure matching the urban itself will follow the growth of the economy, the direct consequence is that the conflict between farmland protection, environmental protection and urban development will burst in the future. Today, we promote the sustainable development of society, not only do we want to meet our demand, but also want to construct a resource-saving and environment-friendly society. So how to solve the contradiction between the urban development and the environment protection so as to keep Xining developing harmoniously would be a major problem.Cellular automata model is a dynamic evolution model with discrete space and time, it is powerful in spatial modeling and parallel computing, so it can take advantage of workstation to simulate the transformation process of complex spatial system. we will use a model based on cellular automata to forecast the development of Xining in the future. Markov Chain model is widely used in forecasting a system which is changing smoothly, in this study, the Markov Chain would provide transition rules for the cellular automata.The research area in this study include four distinct of Xining——chengdong district, chengxi district, chengzhong district and chengbei district. The data is two ETM+ images shot by Landsat in 1996 and 2004. And the resolution of the two images is 28.5 meters. Firstly, we derive the image of land use from the two ETM+ image based on decision tree classification, and the land-use types are artificial land-use, farmland, forest, grassland and water. Then we make images which response to controlling factors, and these images will affect the simulation of the land cover change of Xining. Finally, we predict the land cover of Xining in 2012 based on the Markov model and the CA-Markov model. And we will also compare and analyze the characteristics of the two models and the change of Xining between 1996 and 2004.Cellular automata model is used in the simulation of urban growth recently, while the Markov Chain model is widely used in the dynamic forecast of complex systems, through analyzing the differences between the two models, we find the concept of Markov Chain is not good, because there are so many small polygon on the image, which is different from the real. Using the two models we can predict the land use of Xining in the future. Markov Chain will provide the cellular automata model transformation rules, and the cellular automata will provide Markov Chain spatial concept.In this paper, we can supply some useful advices for the current urban planning by analyzing the development of Xining.