Gravity Field Modeling on the Basis of GRACE Range-Rate Combinations

A new functional model is proposed for gravity field modeling on the basis of KBR data from the GRACE satellite mission. This functional model explicitly connects a linear combination of gravitational potential gradients with a linear combination of range-rate measurements at several successive epochs. The system of observation equations is solved in the least-squares sense by means of the pre-conditioned conjugate gradient method. Noise in range-rate combinations is strongly dependent on frequency, so that a proper frequency-dependent data weighting is a must. The new approach allows a high numerical efficiency to be reached. Both simulated and real GRACE data have been considered. In particular, we found that the resulting gravity field model is rather sensitive to errors in the satellite orbits. A preliminary gravity field model we obtained from a 101 day set of GRACE data has a comparable accuracy with the GGM01S model derived by CSR.