GRACE时变重力场模型反演地表质量变化的组合滤波算法分析

Combined Filtering Algorithm Analysis of Ground Surface Mass Variation Recovery Using GRACE Time-variable Gravity Model

受测量误差和卫星轨道特性影响，直接应用GRACE卫星时变重力场模型反演地球表层质量变化的结果会呈现严重的南北方向条带状误差。空间滤波或去相关滤波都可在一定程度上削弱条带误差，但不能完全消除，而组合滤波则结合了空间滤波和去相关滤波的优势。实验结果表明：组合滤波对球谐系数高阶次项误差和奇偶项间相关误差都进行了消除，相比较单一滤波方法，对条带误差具有更好的滤波效果。

Affected by measuring errors and characteristics of satellite orbit, the earth＇s surface mass inversion result will con- tain serious north-to-south direction stripe error when directly using GRACE time-variable gravity models. Spatial filtering or de-cor- related filtering can reduce the stripe error to some extent but not completely remove it. Combined filtering combines the advantages of spatial filtering and de-correlated filtering. The experiment shows that it can reduce both high-degree harmonic coefficient errors and correlation errors and achieve a better effect in dealing with stripe error than that of a single filtering method.