Progress Toward Evaluation of a Four Revolution Correction Proceedure for GRACE Type Missions

Abstract:

At the GRACE Science Team meeting in October, 2013, the possibility of trying an alternate proceedure for correcting for spurious accelerations and other sources of low frequency noise in the data from GRACE type missions was discussed. The usual proceedure is to apply anti-aliasing corrections and then to fit corrections based on once/rev differential acceleration coefficients and a few other parameters to each one revolution arc of data. However, the uncertainties in the anti-aliasing corrections over land areas and at high latitudes over the oceans can be fairly large. Thus an alternate proceedure called Ocean Calibration is being evaluated. It relies mainly on using the satellite separation results over low latitude ocean areas in determining an empirical correction function. The data arcs used are for periods when the satellites cross the equatorial Pacific on 4 successive revolutions. The data when the satellites are between -30 and +30 deg lat. over the Pacific plus one short arc each over the Atlantic and Indian Oceans are fit. In addition, 5 crossings of the S. Pole and 4 crossings of the N. Pole are used. But 2 parameters corresponding to the mean separations at the poles are included in the fitting parameters, so that only the variations in the geopotential between crossings of the poles are assumed to be small.The simulations are based on the energy conservation approximation, where each satellite speeds up and slows down as the geopotential varies. For a bump in the geopotential, this will be seen at slightly different times for the two satellites, so there will be a bump in the separation between the satellites. Initially the simulations have been based on a model for the uncertainties in the geopotential variations over the equatorial Pacific based on the amplitude of the variations in mass in the ECCO-JPL ocean model, and a model for low frequency noise in the satellite separation. With these models and fitting 16 parameters to the satellite separation, the residual contribution to the geopotential height uncertainty is about 3 mm. However, it is hoped that studies based on real GRACE data can be carried out soon.