Research on Spatio-temporal Simulation Evolution of Rural Settlements in Beijing

• 1、Land Science & Technique School, China University of Geosciences, Beijing 100083
• 2、Land Science & Technique School，China University of Geosciences，Beijing 100083

Abstract：The spatio-temporal evolution pattern of the rural settlements is the core to solve the problems of extensive utilization for the rural residential land. Aiming at comparing CLUE-S model and Markov-CA model(hereinafter referred to as the two models), this article taking Beijing as the research area, selected 14 land-use driving factors with 220 m×220m as grid unit and utilized the two models to simulate and predict the land-use change of Beijing in 2000 and 2005. Kappa index and accuracy were used to compared and analyzed the predicted results.The main conclusions are as follows: CLUE-S model whose core-part is the Logistic Regression Model, has obvious advantage to simulate and predict relatively scattered rural settlements. Markov-CA model which considers the neighborhood, has comparative advantage to simulate and predict urban land that has the characteristic of neighbor diffusion. And of all the land uses, only urban land has this characteristic. This is the main reason for that CLUE-S model's predicting effect is better than Markov-CA model in 2000 and 2005.

Keywords： Land Resources Management Beijing rural settlements driving theory CLUE-S model Markov-CA model