Estimated effects of the vertical structure of atmospheric mass on the time-variable geoid

[1] The GRACE satellite mission is designed to map the Earth's gravity field at a resolution of a few hundred kilometers every 30 days beginning in 2002. At these timescales, much of the change in the gravity field may be attributed to processes involving the redistribution of water on the surface of the Earth. Contributions from continental water storage, the oceans, and the atmosphere will all be present in the time-varying gravity solutions. Isolating the hydrological and oceanographic signals will first require the removal of the atmospheric component of the gravity field estimates provided by GRACE. The vertical distribution of mass in the atmosphere is typically neglected when calculating the atmospheric gravity signal. We examine the accuracy of this approximation, as well as the accuracies of models which determine idealized atmospheric vertical structure from surface values of temperature and pressure. Using isobaric geopotential height data from a global forecast center to characterize the true atmospheric density distribution, we compute an exact atmospheric gravity signal with which to compare the gravity signal of each of these models. In addition, we examine the effects of including the aspherical component of the Earth's shape when calculating the atmospheric component of the gravity field. Because gravity estimates from GRACE will have limited spatial resolution, we average our results over regions of 200 to 500 km. At these length scales, our results show that using models based solely on surface data can introduce errors in the time variable surface mass signal inferred from GRACE as large as a few millimeters equivalent water thickness, with a global RMS of about 1 mm.