广东沿海地区极值风速空间插值方法的对比研究
Study on the Comparison of Spatial Interpolation Methods Used to Estimate Extreme Wind Speed Spatial Distribution Using Irregularly Distributed Station Data in Guangdong Coastal Area
-
摘要: 利用反距离加权法、Cressman法、基于规则样条函数和平板样条函数的径向基函数法、普通Kriging法和泛Kriging法等4大类共6种空间插值方法,对位于广东沿海地区的26个国家级气象站、101个自动气象站和6个海洋站在内的133个站点重现期为50年离地10 m高10 min平均的最大风速进行空间插值研究,并进行交叉验证和空间分布对比。结果表明,Cressman法的绝对误差最小;泛Kriging法的均方根误差最小,相关系数最高;泛Kriging法在空间分布上更贴合实际情况,更适用于广东沿海地区极值风速的空间插值。Abstract: Data from 26 national reference weather stations, 101 regional automatic weather stations and 6 ocean observation stations in Guangdong coastal area are collected. Inverse Distance Weighting scheme, Cressman Interpolation scheme, Radial Basis Function scheme and Kriging method are then empolyed to study the continuous extreme wind speed distribution spatial interpolation experiment of 10 min averaged extreme wind speed at 10 meters above ground with a return peroid of 50 years. Cross-validation shows that the Mean Absolute Error of Cressman Interpolation is least while Universe Kriging interpolation lead to the least Root-Mean-Square Error and the best correlation coefficient. It is indicated that Universe Kriging interpolation is more appropriate for extreme wind speed spatial interpolation in Guangdong coastal area.
下载:
