Modeling of present-day atmosphere and ocean non-tidal de-aliasing errors for future gravity mission simulations

Abstract

A realistically perturbed synthetic de-aliasing model consistent with the updated Earth System Model of the European Space Agency is now available over the period 1995–2006. The dataset contains realizations of (1) errors at large spatial scales assessed individually for periods 10–30, 3–10, and 1–3 days, the S1 atmospheric tide, and sub-diurnal periods; (2) errors at small spatial scales typically not covered by global models of atmosphere and ocean variability; and (3) errors due to physical processes not represented in currently available de-aliasing products. The model is provided in two separate sets of Stokes coefficients to allow for a flexible re-scaling of the overall error level to account for potential future improvements in atmosphere and ocean mass variability models. Error magnitudes for the different frequency bands are derived from a small ensemble of four atmospheric and oceanic models. For the largest spatial scales up to $$\hbox {d/o} = 40$$ and periods longer than 24 h, those error estimates are approximately confirmed from a variance component estimation based on GRACE daily normal equations. Future mission performance simulations based on the updated Earth System Model and the realistically perturbed de-aliasing model indicate that for GRACE-type missions only moderate reductions of de-aliasing errors can be expected from a second satellite pair in a shifted polar orbit. Substantially more accurate global gravity fields are obtained when a second pair of satellites in an moderately inclined orbit is added, which largely stabilizes the global gravity field solutions due to its rotated sampling sensitivity.