<p>Quantifying regional water and energy fluxes much more accurately from observations is essential for improving climate and earth system models, and their ability to simulate future change. This study uses satellite observations to produce monthly flux estimates for each component of the terrestrial water and energy budget over selected large river basins from 2002 to 2013. Prior to optimisation the water budget residuals vary between 1.5 % and 35 % of precipitation by basin, and the imbalance between the net radiation and the corresponding turbulent heat fluxes ranges between ± 10 Wm<sup>−2</sup> in the long term average. In order to further assess these imbalances, a flux-inferred surface storage (FIS) is used for both water and energy, based on integrating the flux observations. This exposes mismatches in seasonal water storage as well as important interannual variability.</p> <p>Our optimisation ensures the flux estimates are consistent with total water storage changes from GRACE on short (monthly) and longer timescales, while also balancing a coupled long term energy budget, by using a sequential approach. All the flux adjustments made during the optimisation are small and within uncertainty estimates using a χ<sup>2</sup> test, and interannual variability from observations is retained. The optimisation also reduces formal uncertainties on individual flux components. When compared with results from previous literature in basins such as the Mississippi, Congo and Huang He River, the FIS metrics show the better agreement with GRACE variability and trends in each case</p>