国土空间监测网络布局优化方法研究

研究目的:服务于自然资源综合监管以及全国国土空间的多级综合动态监测,以县域为基本空间单元,采用多目标优化求解方法,优选具有国土空间代表性的典型县域监测样点,使得选取的每个样点县均能同时代表某种类型的农业生产空间、城乡建设空间和生态保育空间,形成国土空间监测网络优化布局方案,为构建国土空间监测网络提供参考依据。研究方法:基于2015年土地利用变更调查分县数据,结合自然本底、社会经济统计数据,构建空间采样数据集;采用土地垦殖率、城乡建设强度、生态覆盖度三个指标表征国土空间类型,选择与其相关性较大的自然、社会经济指标作为辅助变量,采用重复二分聚类方法进行国土空间分区;根据监测点对不同的农业生产空间类型、城乡建设空间类型、生态保育空间类型的代表性,以及地理空间分布的均匀性,设定多目标优化函数,采用多路空间模拟退火方法进行多目标优化求解,从全国县域单元中优选出能够同时对不同类型的农业生产、城乡建设、生态保育空间均具有代表性的样点县,形成国土空间监测网络布局备选方案;基于专家知识,充分考虑管理实践对监测样点布设的现实要求,从多套理论方案中优选出国土空间监测网络布局的优化方案。研究结果:基于成本最小原则,优选出34个县域监测样点,形成了国土空间监测网络布局参考方案。研究结论:人类对自然资源的开发能力日益增强,国土空间格局正在发生剧烈演变。为动态获取国土空间变化信息,掌握国土空间演变规律,及时制定国土空间优化调控策略,优化国土空间开发格局,亟需加快构建覆盖全国的国土空间多级综合监测网络。本文面向全国国土空间监测网络建设优选出的34个县域监测样点,能够以最少的样本数量实现对主要国土空间类型的全面覆盖,可为全国国土空间监测网络优化布局提供参考。同时,研究提出的国土空间分区和监测样点优化布局方法可为重点区域国土空间监测网络构建提供借鉴。

Spatial Optimized Distribution Method for China Land Monitoring Network

The purpose of this study is to select typical counties/cities from all the Chinese county-level administrative districts based on the multi-objective spatial optimization sampling methods, which can simultaneously represent the specific types of agricultural production space, urban-rural construction space, ecological reservation space, and then to put forward spatial distribution proposal for the construction of China Land Monitoring Network that serves as the natural resources integrated management and multi-level, comprehensive and dynamic land monitoring across the country. Based on land use data of the year 2015 from national land use surveying, natural and social-economic data, the spatial sampling data set was established; three indicators i.e. reclamation rate, urban and rural construction intensity, ecological space covering rate, were selected to reflect main land function types, and combined with highly relevant natural and social-economic variables to conduct China land use zoning by repeated bisection clustering method; taking representativeness on different territorial spaces and geographical distribution uniformity of the monitoring samples as the optimization objectives, the multi-objective optimization function was built and multi-path spatial simulated annealing was used for multi-objective optimization solution, and the common typical counties for different territorial space types were selected as the alternatives samples; fully considering the land management practical requirements on the monitoring samples distribution, one set of samples were selected by virtue of expert knowledge from alternatives as the reference of China land monitoring network. Eventually, 34 counties were selected as the monitoring samples for construction of China Land Monitoring Network based on the principle of minimum cost. In conclusion, human''s abilities of land resources developments are increasingly enhanced and spatial patterns of territory development are sharply changing. In order to dynamically obtain land system data of different scales, explore land evolution rules, and put forward land optimization and control strategies, it is imperative to construct China Land Monitoring Network from the national, regional and local perspectives. The 34 samples selected cover all the types of above three territorial spaces and can be used as the reference for the construction of China Land Monitoring network. Meanwhile, the methods of land zoning and monitoring samples distribution optimization proposed by this study can provide the reference for regional land monitoring network design and construction.