Open Access

Since 2002, the Gravity Recovery And Climate Experiment GRACE provides global high resolution observations of the time variable gravity field of the Earth. Besides other fields of application, this information is used to derive spatio-temporal changes of the terrestrial water storage body. Due to the lack of suitable direct observations of large scale water storage changes, a validation of derived GRACE mass variations remains difficult. In this study, an approach is presented that allows the appraisal of the quality of GRACE for assessing the terrestrial water storage variations of continental scale hydrological basins. This is done by comparing the GRACE products to aerologic water budgets that are derived from the vertically integrated atmospheric moisture flux divergence. Aerologic water budgets of global atmospheric reanalyses are explored, covering the products of the European Centre for Medium-Range Weather Forecasts (ECMWF) and the National Centers for Environmental Prediction (NCEP). Furthermore, dynamic downscaling is applied to obtain high-resolution refinements of the coarse global reanalysis fields. For that purpose, the Weather Research and Forecast modeling system (WRF) is chosen. Prior to the evaluation with GRACE, all aerologic products are validated against global gridded observations data sets of precipitation and 2m-temperature. Atmosphere related uncertainty bounds for the terrestrial water storage variation are obtained from different configurations of the downscaling model WRF in combination with the two involved global reanalyses. The resulting atmospheric uncertainty ranges are opposed and compared with uncertainty ranges originating from three different GRACE products of the data centers GeoForschungsZentrum Potsdam (GFZ), Center for Space Research (CSR) and Jet Propulsion Laboratory (JPL). The results show that regional atmospheric downscaling is able to add value to the global reanalyses, depending on the geographical location of the considered catchments. In general, global and regional atmospheric water budgets are in reasonable agreement with GRACE derived terrestrial water storage variations (r=0.6-0.9). However, both atmospheric and satellite based approaches reveal significant uncertainties. In particular, for regions with minimal storage change rates, i.e. desert environments, the limitation of both methods is revealed. The study encompasses six different climatic and hydrographic regions. They comprise the Amazonian Basin, the combined river catchments of Lena and Yenisei, the Central Australian Basin, the Saharan Desert, the Lake Chad depression, and the Niger River catchment.