面向城市建成区扩展模拟的CA-Markov模型改进

Improvement of CA-Markov Model for Extended Modeling of Urban Built-up Area

对城市建成区扩展的预测是防止城市蔓延的重要管理依据。目前,元胞自动机-马尔可夫链模型,已成为城市建成区扩展预测的重要方法。该模型对指标权重的赋值方法较为敏感,以往的单一指标赋值法,影响了城市建成区扩展预测的精度和可信度。为此,本研究提出整合传统权重赋值法的AHP和逻辑回归模型改进CA-Markov模型。研究选择云南省大理市为案例,对2020、2030年的城市建成区扩展进行模拟和预测,最后进行精度验证。研究结果表明:(1)Kappa指数可达到96.8%,预测结果有较好的一致性。(2)大理市的城市建成区扩展均表现为继续向外扩展,以东南、西北方向和两片建成区之间为主要扩展方向。研究提供了组合权重赋值法改进CA-Markov模型,这将为规划者在未来规划中提供强有力的支持。

The prediction of the expansion of urban built-up areas is an important management basis for preventing the spread of cities. At present,the CA-Markov model has become an important method for the expansion of urban built-up areas. The model is more sensitive to the method of index weight assigmnent.In the past,the single index assignment method ＂affected the accuracy and credibility of the extended prediction of urban buih-up areas.To this end,this paper proposes to improve the CA-Markov model by integrating AHP and logistic regression models based on the traditional weighted assignment method.The paper selects Dali city,Yunnan province as a case study to simulate and forecast the expansion of urban built-up areas in 2020 and 2030,and finally verifies the aceuracy.The resuhs show that ：（1）Kappa index can reach 96.8%, the prediction results have a good consisteney.（2）The expansion of urban built-up areas in Dali City continues to expand outward,with the southeast and northwest directions as the main expansion direction between the two built-up areas. Research provides a combination of weight assignment method to improve the CA-Markov model, which will provide planners with strong support in the future planning.