地统计方法学研究进展

地统计方法学已经成为空间预测与不确定性分析的关键性工具。本研究从文献计量学和方法学演变过程两个角度开展了1967~2005年的地统计方法学研究综述研究。首先从宏观角度分析其发展趋势、应用情况与模式,然后总结其演化规律、适宜性和选择原则。研究表明:地统计方法学的演化规律表现为稳态向非稳态演变、单变量向多变量(含二次信息)演变、参数与非参数方法相互补充、线性向非线性方法演变和空间静态向时空动态演变;此外,未来研究发展方向集中在半变异函数估计新方法、不确定性地统计学、时空地统计学与多点地统计学、机理模型与地统计学耦合研究和基于地统计学模拟的不确定性决策等。

State-of-art on geostatistical methodology

Geostatistical methodology becomes an important tool for spatial prediction and uncertainty analysis.In terms of scientometric analysis and methodology development,advanced research for geostatistical methodology from 1967 to 2005 has been proposed in this study.First,its developing trend,application and patterns were identified;its development mode,suitability and choosing principles were then summarized.Development mode could be summarized as five transformations from stationary to non-stationary,single-variable to multivariate,parametric to non-parametric,linear to non-linear and spatial static to spatiotemporal dynamic geostatistics.Finally,the future researches were preliminarily discussed in this field,which is of great significance to geostatistical methodology and applications in the future.It includes:1)developing new methods for variogram estimation to reduce the analytical complexity,2)refining uncertain geostatistics to reflect hybrid uncertainties(input/output dataset,model structure,model parameters),3)exploiting spatiotemporal/multi-point geostatistics and its algorithm/software to simulate complex realizations,4)developing hybrid approach of process-mechanism and geostatistics to comprehensively uncover potential process-mechanism,in spite of the difficulties within the process simulations,and 5)expanding geostatistical-based decision-making under uncertainty,which supports the risk-based decision-making in social work,environmental pollution,agricultural production,public health,etc.