Regularization and analysis of GRACE mass anomaly time series by a minimization of month-to-month year-to-year double differences

A methodology has been developed to estimate the accuracy of GRACE monthly solutions as functions of space or time without using independent geophysical models. An application of the methodology to several commonly-used solution time-series reveals that the ITSG-Grace2016 solutions show the highest accuracy among those considered. Furthermore, it is found that the accuracy of background models exploited to produce GRACE RL05 solutions was likely insufficient to describe adequately mass re-distribution in the oceans at both short (< 1 month) and long (>1 month) time scales. Insufficiently accurate modelling of ocean signals may reduce the accuracy of mass anomaly estimates not only over oceans, but also in the coastal areas of continents.