Abstract
Abstract. Terrestrial water storage (TWS) composed of surface waters, soil moisture, groundwater and snow where appropriate, is a key element of global and continental water cycle. Since 2002, the Gravity Recovery and Climate Experiment (GRACE) space gravimetry mission provides a new tool to measure large-scale TWS variations. However, for the past few decades, direct estimate of TWS variability is accessible from hydrological modeling only. Here we propose a novel approach that combines GRACE-based TWS spatial patterns with multi-decadal-long in situ river level records, to reconstruct past 2-D TWS over a river basin. Results are presented for the Amazon Basin for the period 1980–2008, focusing on the interannual time scale. Results are compared with past TWS estimated by the global hydrological model ISBA-TRIP. Correlations between reconstructed past interannual TWS variability and known climate forcing modes over the region (e.g., El Niño-Southern Oscillation and Pacific Decadal Oscillation) are also estimated. This method offers new perspective for improving our knowledge of past interannual TWS in world river basins where natural climate variability (as opposed to direct anthropogenic forcing) drives TWS variations.
Highlights
Terrestrial water storage is an important component of the global and continental water cycle
The present study has established for the first time direct, observation-based, estimate of Terrestrial water storage (TWS) spatiotemporal variability over the Amazon Basin for the past ∼3 decades
It is only based on TWS observations from Gravity Recovery and Climate Experiment (GRACE) and in situ water levels
Summary
Terrestrial water storage (hereafter noted as TWS) is an important component of the global and continental water cycle. The method developed in the present study combines spatial information on TWS from GRACE (over 2003–2008) with multi-decade-long (over 1980–2008) but sparse river level time series based on in situ gauges. In the Amazon Basin, previous studies (e.g., Xavier et al, 2010; Vaz de Almeida, 2009) have shown that at seasonal and interannual time scales, river water level fluctuations can locally be correlated to TWS (as observed by GRACE). Such a correlation suggests that, at these time scales, TWS (including underground waters) and surface waters co-vary in a similar way. The final products are gridded (i.e., 2-D) time series of past TWS
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
