摘要
如何把离散的气象台站资料通过合适的空间内插方法转变成规则的网格数据 ,对于气候变化分析和模拟研究具有重要的意义。作者利用中国区域 1 60个常用台站 1 0年降水观测资料为例 ,分别采用克里格 (Kriging)插值、反距离加权、Delaunay三角剖分线性插值、双谐样条 (BiharmonicSpline)插值和Cressman客观分析等几种常见的空间内插方法 ,较为系统地分析和比较了这几种内插方法插值结果之间的异同 ,对其优缺点和适用范围进行了适当讨论 ,并采用不同的网格分辨率和影响半径对Cressman客观分析方法做了进一步分析比较。结果表明 :在台站分布密集的区域不同空间内插方法之间差异较小 ,在台站分布稀疏的区域则差异较大 ;与更加密集的台站观测资料的比较显示 ,自动调节影响半径的Cressman客观分析方法与双谐样条插值方法误差相对较小。
How to transfer the scattered observed station data into the regular grid data is vital to the research of climatic change and simulation, which needs spatial interpolation method to achieve. The inter-comparison is made among Kriging, Inverse distance weighting, Delaunay triangulation linear interpolation, Biharmonic spline and Cressman objective analysis spatial interpolation based on observed scattered station precipitation data over China. The advantages, disadvantages and application areas of every spatial interpolation method are discussed. The further research of Cressman objective analysis is also made by altering grid resolution and influencing radius. The results show that biases of different methods are very small in the dense data area and obvious in the sparse data area. The relative errors caused by Cressman objective analysis that influencing radius is adjusted according to the density of stations and Biharmonic spline method are less than other methods.
出处
《气候与环境研究》
CSCD
2004年第2期261-277,共17页
Climatic and Environmental Research
基金
国家重点基础研究发展规划项目G1990 4 34 0 0资助
