基于“三位一体”空间抽样理论研究及其二联查找表研建

我国环境与经济社会高度分异且快速变化，高效监测、准确估计、及时报告是科学调控的依据，空间抽样将提供关键方法论。抽样效率受被调查对象的地学特征，抽样方法和估计量三位一体所控制，每一维都有许多类型或选择,形成众多组合。一些组合导致最优线性无偏估计(BLUE，Best Linear Unbiased Estimation)，而有些则导致抽样效率损失。由于先验知识、经费、野外条件、观测仪器的局限，实际地学抽样组合与理想BLUE组合的偏差在所难免。针对上述问题，本项目拟开展基于三位一体空间抽样理论研究，系统地揭示三一组合的统计性质、内在关系、以及互相转换条件，据此研建空间抽样查表。空间抽样查找表将以简单的二联表形式揭示地学空间抽样理论体系，为地学抽样实践提供方便快捷的查表：因地制宜地选择和设计优化抽样包；对已有抽样策略的偏性和效率作出判断和优化建议；对已发布的统计报告的可信度作出评估并推荐纠偏方案。

China environment and social-economic conditions are highly heterogeneous and are rapidly changing. Spatial sampling is an essential tool for monitoring and estimating the variations. A high efficient sampling is related to the sampling trinity: property of a target population, way of sampling and method of estimation, each has many options. Some of the numerous combinations (packages) of the triple elements would lead to the BLUE (best linear unbiased estimate), while others cause sampling efficiency loss. Sampling practices are often biased from a BLUE sampling due to the lack of prior knowledge, restrictions in the field work or instruments, and limitations of budget and estimators. In this study, we will exhaustively investigate typical sampling packages and develop a Table of spatial sampling. The Table of spatial sampling can be used as a quick reference for choosing a BLUE sampling package from numerous options; assessing the quality of an existing sampling package and giving an advice for improvement; and judging the reliability of a statistic report.