一种融合遥感和地面观测资料的雪深空间插值方法

针对积雪观测站点稀少的问题,提出一种考虑海拔影响,能够融合MODIS积雪面积产品和站点观测的雪深空间插值方法,该方法利用去云后MODIS积雪面积产品构建的无积雪“虚拟站点”弥补站点分布不均匀和稀少的不足,利用泛协克里金插值方法考虑海拔对雪深的影响。利用北疆地区50个气象站点的逐日雪深观测资料、逐日MODIS积雪面积产品和AMSR-E被动微波雪水当量和雪深产品,对普通克里金、泛克里金、普通协克里金和泛协克里金插值结果进行了比较研究。研究结果表明:积雪覆盖范围较大时,站点雪深与海拔之间相关系数较大,利用泛协克里金插值结果精度高且稳定;否则利用普通克里金插值精度较高且稳定。通过增加“虚拟站点”,能够提高雪深插值精度,并在一定程度上修正了克里金插值中存在的平滑效应。


A Novel Spatial Interpolation Method for Snow Depth by Integrating Satellite and Ground Observations

For the problem of rare snow observation stations,this paper proposed a new scheme to produce spatial distribution of snow depth on based on different kriging interpolation methods,which can not only consider elevation effects,but also fuse MODIS snow cover products.This scheme uses snow\|free pixels in the cloud\|removed MODIS snow cover image to build virtual stations with zeros snow depth to compensate for the scarcity and uneven distribution of stations.Additionally,the universal co\|kriging interpolation method is used to consider the impact of elevation on snow depth.The daily snow depth observations at 50 meteorological stations in northern Xinjing province are chosen to evaluate the proposed scheme.Four types of kriging methods are also compared such as ordinary kriging,universal kriging,ordinary co\|kriging and universal co-kriging.Results show that universal co\|kriging can achieve the best performance with a larger snow cover area and a bigger value of correlation coefficient between snow depth and elevation.Otherwise,the best performance is achieved by the ordinary kriging.The added virtual stations can improve the accuracy of interpolation and reduce smoothing effect in kriging interpolation.