Assimilation of GRACE Follow‐On Inter‐Satellite Laser Ranging Measurements Into Land Surface Models

Abstract

The monthly mean level‐2 (L2) time‐variable gravity as well as level‐3 (L3) Mass Concentration blocks (mascons) data from Gravity Recovery and Climate Experiment (GRACE) and GRACE Follow‐On (GRACE‐FO) missions are frequently used for improving land surface models. The conventional data assimilation approach requires several pre‐processing steps like filtering (to suppress the random and systematic errors), correction (to reduce the leakage effect) and scaling (to un‐do the attenuation effect by filters) before integrating the data with the models, for example, when computing total water storage (TWS) changes from L2 data. Moreover, due to the monthly sampling of L2 and L3 data, the model estimates are updated only once a month. This confines the applications of the approaches to cases where water storage is slowly varying at seasonal or interannual time‐scales. We present a new methodology based on direct assimilation of along‐orbit Line‐of‐sight Gravity Difference (LGD) measurements from the GRACE‐FO laser ranging interferometer (LRI) which overcomes the limitation of the conventional approach. Inter‐satellite ranging data reflect the gravitational changes at satellite altitude caused by the instantaneous TWS changes at the Earth surface. Therefore, they pose information at wide‐ranging time‐scales from hours to multiple years. The proposed method is applied globally to assimilate LRI data into a land surface model. Evaluation against multiple satellite‐based and ground data shows the superiority of our proposed approach. The new approach also offers better performance in capturing high‐frequency water storage variations imposed by sub‐monthly climatic events due to its higher number of data assimilation cycles within a month.