基于CA-Markov模型的向海湿地土地覆被变化动态模拟研究

论文借助地理信息技术，采用CA-Markov模型，以1988、1996、2003年3个相同时相的TM遥感影像作为数据基础，并结合地形图和相应的交通、气候等数据对向海自然保护区未来的土地覆被类型进行动态模拟和预测。在模型建立过程中，通过马尔可夫模型求出转移概率矩阵，确定CA模型转换规则。利用CA-Markov模型对向海自然保护区未来的土地覆被类型转化方向进行了动态模拟，将模拟结果与已有的土地覆被类型对比，得出模拟精度达到74.46%。据此对向海自然保护区2010、2018年土地覆被类型做出预测。预测结果表明：向海自然保护区土地覆被类型整体趋势不容乐观，湿地面积在大量减少，林地和草地呈减少趋势，耕地、居民地、未利用地呈增加趋势，其中未利用地面积增加最多。该方法可以对向海自然保护区的湿地保护和动态监测以及可持续发展提供依据。

Dynamic Simulation of Land Cover Changes in Xianghai Wetlands Based on the CA-Markov Model

Based on the three same phase TM remote sensing images of 1988, 1996 and 2003 and base maps, traffic and climate data, the paper expatiates the dynamic simulation and prediction for future land cover types of Xianghai Nature Reserve,combined with CA-Markov Model by means of geographic information technology. In the model-building process, the rules of CA model of the conversion are determined after the transition probability matrix is obtained through the Markov model. The dynamic simulation of Xianghai Natural Reserve is made to research the future transition direction of land cover types with the CA-Markov model, the simulation accuracy is more than 74.46% compared with the existing land cover types. According to the testing process above, the land cover types of Xianghai Nature Reserve are predicted in 2010 and 2018. The forecast results show that the overall trend of the land cover types is not optimistic in the study area, wetland area is in a significant reduction state while woodland and grassland decrease. Arable land, residential areas and the unused land show an upward trend in which the speed of increase is the fastest for unused land. The methods can provide a basis for wetland protection and dynamic monitoring as well as sustainable development for the Xianghai Nature Reserve.