Precipitation explains GRACE water storage variability over large endorheic basins in the 21st century

Abstract

Introduction: Seasonal—interannual variations in surface water storage revealed by the Gravity Recovery and Climate Experiment (GRACE) satellites have received less attention than storage trends in the literature. We focus on six large endorheic basins and develop variability attribution diagnostics against independent precipitation and evapotranspiration (hereafter P and E) datasets.Methods: We generate a flux-inferred storage (FIS), representing the integral of the component flux anomalies into and out of a region, enabling a comparison between the P and E contributions to GRACE water storage anomalies on seasonal to interannual timescales. Additionally, a monthly budget closure approach is applied, giving self-consistent coupled water and energy exchanges from 2002 to 2020.Results: On seasonal timescales, P and E data show insufficient cancellation, implying over-large seasonal variations in surface storage. In most basins, P drives the seasonal storage cycle with E dampening storage amplitudes, although in the Caspian Basin, seasonal storage is driven by E, with P remaining seasonally constant when integrated over the whole drainage basin. Budget closure mostly adjusts E, which has larger uncertainties, in fitting the GRACE data. On year-to-year and multi-year timescales, there is a strong correlation between P-driven storage and the observed GRACE variability, which ranges between 0.55 and 0.88 across all basins, and this is maintained after budget closure. However, storage changes driven by P alone from GPCP are too large compared to GRACE, with E data from FLUXCOM generally having only very weakly compensating interannual variations. After budget closure, interannual E variability is substantially increased. Closed energy budgets often show interannual amplitudes, partly driven by radiation and partly by water budget variation through shared latent heat losses, although these have not been independently verified.Discussion: Although water flux trends cannot be detected with significance due to the large interannual variability, the strong agreement between multi-annual GRACE storage and precipitation variations, especially over the Caspian basin, lends no support to the suggestion that E changes driven by climate change are responsible for water storage trends seen by GRACE.