||The Accuracy of Mass Estimates Computed From the GRACE Gravity Field
Most scientific applications of GRACE begin by using the GRACE gravity fields to compute changes in mass at the Earth's surface; changes in the global distribution of mass in some cases, changes within specific regions in others. But how good are those mass estimates? Direct validation is hard, because of the difficulty of finding a region large enough for GRACE to resolve, where the mass variability can be independently determined. Instead, we have been estimating mass uncertainties using the GRACE data themselves. The method relies on the plausible but unproven assumptions that the gravity field errors do not include significant annually varying components, and that the errors are not notably correlated with the true gravity signal. The method and the resulting error estimates will be described here.
The fact that plausible error estimates can be constructed is one of the outstanding characteristics of GRACE. It's what makes it possible to use GRACE to assess and validate model predictions, and it provides a level of confidence to use when interpreting features in the GRACE results.