Assessment of pluri-annual and decadal changes in terrestrial water storage predicted by global hydrological models in comparison with the GRACE satellite gravity mission

Abstract

Abstract. The GRACE (Gravity Recovery And Climate Experiment) satellite gravity mission enables global monitoring of the mass transport within the Earth's system, leading to unprecedented advances in our understanding of the global water cycle in a changing climate. This study focuses on the quantification of changes in terrestrial water storage with respect to the temporal average based on an ensemble of GRACE solutions and two global hydrological models. Significant changes in terrestrial water storage are detected at pluri-annual and decadal timescales in GRACE satellite gravity data that are generally underestimated by global hydrological models though consistent with precipitation. The largest differences (more than 20 cm in equivalent water height) are observed in South America (Amazon, São Francisco and Paraná River basins) and tropical Africa (Congo, Zambezi and Okavango River basins). Smaller but significant (a few centimetres) differences are observed worldwide. While the origin of such differences is unknown, part of it is likely to be climate-related and at least partially due to inaccurate predictions of hydrological models. Pluri-annual to decadal changes in the terrestrial water cycle may indeed be overlooked in global hydrological models due to inaccurate meteorological forcing (e.g. precipitation), unresolved groundwater processes, anthropogenic influences, changing vegetation cover and limited calibration/validation datasets. Significant differences between GRACE satellite measurements and hydrological model predictions have been identified, quantified and characterised in the present study. Efforts must be made to better understand the gap between methods at both pluri-annual and decadal timescales, which challenges the use of global hydrological models for the prediction of the evolution of water resources in changing climate conditions.