Correction of Inconsistencies in ECMWF’s Operational Analysis Data During De-Aliasing of GRACE Gravity Models

philip kotler;

The main objective of the Gravity Recovery and Climate Experiment (GRACE) Atmospheric and Oceanic De-Aliasing Level-1B product (AOD1B) is the removal of high-frequency nontidal mass variations due to sub-monthly mass transport in the atmosphere and oceans. Application of AOD1B shall avoid aliasing of these high-frequency signals into monthly gravity models derived from modern gravity missions and shall help to derive consistent orbit solutions for altimetry and Satellite Laser Ranging missions. The AOD1B 6-h series of spherical harmonic coefficients up to degree and order 100 are routinely generated at the German Research Centre for Geoscience and distributed to the GRACE Science Data System and the user community. Inputs for this product are acquired from numerical weather prediction models which are regularly revised and consequently not stable in time. The latest AOD1B release 5 (RL05) is based, as all other releases, on input from ECMWF and does not resolve this problem of discontinuities present in the surface pressure and surface geopotential input data. This might contaminate the gravity field variations derived from atmospheric mass variations. In this paper we present a method to overcome this problem during future AOD1B product generation, aswell as two new Level-2 products (GAEand GAF) that, over land, fix a posteriori the two jumps present in the already distributed Level-2 RL05 monthly gravity models which were based on AOD1B RL05. The impact of the proposed correction on the variations and long-term trend of the total mass of the atmosphere and on the ice mass balance over Antarctica and over Greenland is also illustrated. We found that the GAE/GAF-corrected trend of the global atmospheric mass over the GRACE mission lifetime significantly decreased from −0.05 to −0.02 mm yr−1 in terms of geoid height. A considerable effect (33 per cent) was also found in the quadratic term of ice mass loss over Antarctica which results in an acceleration of 3.2 Gt yr−1 yr−1 smaller than without applying this correction.