Impact of numerical weather models on gravity field analysis

Abstract

Atmospheric pressure variations are one of the major sources of gravity perturbations. Due to the high variability of the atmospheric masses and the sparse sampling of these by GRACE the signals alias into the observations taken by the satellites. The determination of accurate atmospheric gravity field coefficients (AGC) is indispensable for the elimination of these signals. For the determination of AGC it is state of the art to use high resolution Numerical Weather Prediction (NWP) models which take into account the time-variable three-dimensional distribution of the atmosphericmass. By subtracting the gravity spherical harmonics of a long term atmospheric mean field from the ones of the instantaneous atmosphere, the residual gravity spherical harmonic series is obtained. It describes the deviation of the actual gravity field from the mean gravity field due to atmospheric mass variations. NWP models are not perfect as they can show significant differences to in situ measurements. Further these models evolve and change throughout time, which can lead to changes in the pressure data and therefore in the AGC. In this study several aspects of NWP models are investigated, and the influence they have on the determination of the AGC is discussed.We present a strategy that was developed for dealing with changes in the NWP models, and compare our products to those of the GRACE Atmosphere and Ocean Dealiasing level-1B products and those provided by the Groupe de Recherche de Géodésie Spatiale (GRGS).