The Role of the Atmosphere for Satellite Gravity Field Missions

Traditionally, in satellite derived static global gravity field models, the total mass of the Earth is considered to be the sum of the solid Earth mass and the water masses in land (hydrology), oceans, ice (continental ice and sea ice) and atmosphere. The total mass is reflected in the GM value (gravity constant times Earth mass), which usually is a parameter of geodetic reference systems. As the whole system Earth is mass conserving in principle GM should be constant over time. Any time dependency of GM could only be caused by wrong estimates for GM (or the corresponding zero degree coefficient of the spherical harmonic series). For analysis of the time variable gravity field the situation is more complex, because masses are fluctuating at various temporal and spatial scales and are exchanged between components of the Earth system. Disregarding mass variations inside the solid Earth (e.g. by earthquakes, mantle convection, etc.) the water cycle represents the major source of mass variations, of which the atmosphere plays the most prominent role (by pressure field variations, by forcing ocean circulation, by precipitation and evaporation). In order to take into account such temporal variations for gravity field analysis the sampling pattern of the satellite mission and its sensors in addition play a crucial role. This means that during gravity field recovery the known part of mass variations with respect to the mean value have to be modelled independently and corrected for in order to avoid aliasing of temporal signals into the resulting fields. In contrast, unknown time variable mass variations have to be estimated. Both approaches are applied nowadays in the GRACE data analysis simultaneously. The paper introduces the problem of atmospheric mass variability for gravity field determination with satellite data. In the main part of the paper estimates for the total mass of the atmosphere and its variability in time as well as an analysis of the impact of the vertical structure and of potential errors in the atmospheric fields are provided. Results indicate, that the total atmospheric mass during the last 20 years slightly increased, what could be a hint to global warming. Results of error analysis also show, that for reaching the ultimate performance with GRACE data analysis, the vertical structure of the atmosphere as well as potential errors in the atmospheric fields are crucial.