中国湿地分布的时空变化模拟研究

Wetlands distribution and its spatio-temporal changes are extremely important to global changes, biodiversity, environment and health. The impact factors those influence the spatial and temporal distribution of China wetlands are much more complicated, which have various impacts on wetland changes along with space and time. This project is designed to study the model that was employed to simulate and predict China wetlands spatial distribution. So the project plans are as follows: firstly, the random forest method will be employed to quantitatively differentiate and analyze the aforesaid impact factors to determine their contribution to China wetland changes. Secondly, based on the previous researches, the project will also plan to continue the research of regionalization of china wetland changes by SOFM method. .Thirdly, wetlands potential occurrence model was redesigned by reference of related researches at home and abroad, and then used to predict wetlands potential distribution across China, which was validated according to existed wetland map in 1978 and other maps..After that, China wetland changes model, which is based on CA model and logistic model, was studied elaboratively and employed to predict the China wetland spatial distribution from 1978 to 2008. At the same time, multi-resources satellite images are interpreted to produce China wetland classification maps, with the aid of China potential wetlands distribution map and other auxiliary data. These wetland classification maps will be employed to validate the corresponding predicted wetland distribution, which were produced by CA model and logistic model, respectively. And finally these two models will be analyzed in view of model efficiency, precision of prediction result. These models studied in this project will imply an innovative approach to map China wetlands efficiently in future and is meaningful for wetlands management.

湿地分布及其时空变化，对于全球变化、生物多样性、环境与健康等都具有重要意义。影响中国湿地分布及时空变化的因素复杂、且随时空不同而对湿地具有不同的影响程度；项目拟利用Random forest方法对这些因素进行定量辨析，给出这些因素对湿地影响的贡献度；并基于已有研究成果，利用SOFM方法进行中国湿地变化分区的研究；同时从湿地发生学角度，研究构建中国湿地的潜在分布模型；在此基础上，以现有中国湿地数据（1978-2008年）为基础，分别基于CA模型和logistic模型研究建立中国湿地分布的时空变化模拟模型，模拟中国湿地空间格局的演变；基于多源遥感数据分类和相关调查等辅助地学数据，对模拟模型进行验证，并从运行效率和预测精度等方面对模型进行对比分析。模型方法不仅将为全国大范围湿地制图探索新的途径，也为满足中国湿地管理的需要提供科学支持。

湿地对于全球变化、生物多样性保护以及环境与健康都具有重要意义。但湿地的分布和变化受多种因素影响，且随时空不同而具有不同的影响程度。项目基于中国多期湿地数据（1978-2008），利用地理加权回归分析方法，研究了湿地变化的驱动力因素（自然和社会经济因素等）随时空变化的分布特征。综合遥感和地理信息系统技术，建立了应用于大尺度（国家和全球）湿地分布模拟的模型方法，模拟了中国和全球湿地的分布，在此基础上对全球湿地的减少状况做出了新的判断。针对湿地的动态性特征和单时相遥感数据在提取湿地方面的局限，发展了基于时间序列遥感数据的湿地遥感分类和制图方法，对中国2000年以来的湿地进行了分类和制图，提高了湿地类型的分类精度。项目研究成果不仅为提高大尺度湿地遥感制图精度奠定基础；而且为全球变化研究、生物多样性保护以及环境与健康研究提供了有效的科学数据和结果，为国家湿地保护政策的决策提供了有效支持。