Coherent Satellite Monitoring of the Water Cycle Over the Amazon. Part 2: Total Water Storage Change and River Discharge Estimation

In companion paper 1, the SAtellite Water Cycle (SAWC) satellite data set integration approach was presented. SAWC accounts for (1) the closure of the water budget at the sub-basin scale by (2) using upstream/downstream dependencies. Here, the SAWC database is used to reconstruct a missing water component. The total water storage change (dS) can be reconstructed prior to the Gravity Recovery and Climate Experiment (GRACE) period. In terms of seasonal and interannual variations, SAWC provides long-term dS estimates that are highly consistent with the GRACE estimate (2002–2015). When compared to the extent of inundation, correlation of anomalies reaches 0.6 over sub-basins presenting large floodplain. A strong relationship (correlation of 0.7) exists between the monthly dS and the El Niño index. The river discharge R can also be estimated using the SAWC approach during the GRACE period. A scheme was developed to estimate R continuously along the river through the indirect interpolation of discharge station measurements by using satellite estimates of dS, precipitation P, and evaporation E. These R estimates along the river were evaluated with 80 independent measurements and had a correlation of 0.86 and a Kling-Gupta efficiency index of 0.71. A comprehensive monthly river discharge estimate was proposed for the first time across the Amazon basin over 13 years (2002–2015) at a 0.25° resolution. The SAWC framework is a pure observation tool for integrating a large range of satellite estimates, potentially facilitating the assimilation of such data into land surface models.