摘要
对比分析不同空间插值模型对GPS大气可降水量的插值精度,并对插值误差影响因素进行研究。利用反距离权重法、全局多项式法、局部多项式法、径向基函数法、普通克里金法5种常用空间插值方法对SuomiNet网GPS PWV数据进行插值,应用交叉验证法对比分析各方法的精度,得出5种插值模型均达到一定的插值精度。其中,普通克里金法和径向基函数法插值精度最高,全局多项式法和反距离权重法精度最低,而局部多项式法插值精度中等。对插值误差影响因素的分析表明,GPS站点密度越高、水汽值越低、水汽值变化程度越小,空间插值精度越高。研究成果可为大气水汽含量空间分布特征和规律研究,以及空间插值方法选取提供借鉴。
Analyzed and compared interpolation precision of different spatial interpolation models for GPS Precipitable Water Vapor （PWV） , and researched on the error factors of the interpolation models. The cross validation method was applied to analyze and compare interpolation precision of Inverse Dis- tance Weighting （ IDW ）, Global Polynomial Interpolation （ GPI）, Local Polynomial Interpolation （LPI） , Radial Basis Functions （RBF） and Ordinary Kriging （OK） for SuomiNet GPS PWV interpolation. Re- sults showed that the five models all reached relatively high interpolation precision, among which the pre- cision of OK and RBF were the highest, LPI lower, and GPI and IDW the lowest. Error factors analysis showed that the interpolation precision was higher when the GPS stations density was higher, the average value and the volatility of PWV were lower. The results of this study may provide useful guidance for re- search on spatial distribution characteristic and law of PWV and choice for PWV interpolation models.
出处
《气象与环境科学》
2013年第1期1-6,共6页
Meteorological and Environmental Sciences
基金
民用航天"十二五"预先研究项目
国土资源部地学空间信息技术重点实验室开放基金(KLGSIT2013-15)资助
关键词
大气可降水量
空间插值方法
插值误差
交叉验证法
GPS
Precipitable Water Vapor
spatial interpolation method
interpolation error
cross vail-dation method
GPS
