New Estimates of Variations in Water Flux and Storage over Europe Based on Regional (Re)Analyses and Multisensor Observations

International audiencePrecipitation minus evapotranspiration, the net flux of water between the atmosphere and Earth’s surface, links atmospheric and terrestrial water budgets and thus represents an important boundary condition for both climate modeling and hydrological studies. However, the atmospheric–terrestrial flux is poorly constrained by direct observations because of a lack of unbiased measurements. Thus, it is usually reconstructed from atmospheric reanalyses. Via the terrestrial water budget equation, water storage estimates from the Gravity Recovery and Climate Experiment (GRACE) combined with measured river discharge can be used to assess the realism of the atmospheric–terrestrial flux in models. In this contribution, the closure of the terrestrial water budget is assessed over a number of European river basins using the recently reprocessed GRACE release 05 data, together with precipitation and evapotranspiration from the operational analyses of high-resolution, limited-area NWP models [Consortium for Small-Scale Modelling, German version (COSMO-DE) and European version (COSMO-EU)] and the new COSMO 6-km reanalysis (COSMO-REA6) for the European Coordinated Regional Climate Downscaling Experiment (CORDEX) domain. These closures are compared to those obtained with global reanalyses, land surface models, and observation-based datasets. The spatial resolution achieved with the recent GRACE data allows for better evaluation of the water budget in smaller river basins than before and for the identification of biases up to 25 mm month−1 in the different products. Variations of deseasoned and detrended atmospheric–terrestrial flux are found to agree notably well with flux derived from GRACE and discharge data with correlations up to 0.75. Finally, bias-corrected fluxes are derived from various data combinations, and from this, a 20-yr time series of catchment-integrated water storage variations is reconstructed