Assimilation of transformed water surface elevation to improve river discharge estimation in a continental-scale river

[摘要] Quantifying continental-scale river discharge isessential for understanding the terrestrial water cycle, but it is susceptible toerrors caused by a lack of observations and the limitations of hydrodynamicmodeling. Data assimilation (DA) methods are increasingly used to estimateriver discharge in combination with emerging river-related remote sensingproducts (e.g., water surface elevation (WSE), water surface slope, riverwidth, and flood extent). However, directly comparing simulated WSE tosatellite altimetry data remains challenging (e.g., because of large biasesbetween simulations and observations or uncertainties in parameters), andlarge errors can be introduced when satellite observations are assimilatedinto hydrodynamic models. In this study we performed direct, anomaly, andnormalized value assimilation experiments to investigate the capacity of DAto improve river discharge within the current limitations of hydrodynamicmodeling. We performed hydrological DA using a physically based empiricallocalization method applied to the Amazon basin. We used satellite altimetrydata from ENVISAT, Jason 1, and Jason 2. Direct DA was the baselineassimilation method and was subject to errors due to biases in the simulatedWSE. To overcome these errors, we used anomaly DA as an alternative todirect DA. We found that the modeled and observed WSE distributions differedconsiderably (e.g., differences in amplitude, seasonal flow variation, and askewed distribution due to limitations of the hydrodynamic models).Therefore, normalized value DA was performed to improve dischargeestimation. River discharge estimates were improved at 24 %, 38 %, and62 % of stream gauges in the direct, anomaly, and normalized valueassimilations relative to simulations without DA. Normalized valueassimilation performed best for estimating river discharge given the currentlimitations of hydrodynamic models. Most gauges within the river reachescovered by satellite observations accurately estimated river discharge, withthe Nash–Sutcliffe efficiency (NSE)  > 0.6 . The amplitudes of WSE variationwere improved in the normalized DA experiment. Furthermore, in the Amazonbasin, normalized assimilation (median NSE  =0.50 ) improved river dischargeestimation compared to open-loop simulation with the global hydrodynamicmodel (median NSE  =0.42 ). River discharge estimation using direct DAmethods was improved by 7 % with calibration of river bathymetry based onNSE. The direct DA approach outperformed the other DA approaches whenrunoff was considerably biased, but anomaly DA performed best when the riverbathymetry was erroneous. The uncertainties in hydrodynamic modeling (e.g.,river bottom elevation, river width, simplified floodplain dynamics, and therectangular cross-section assumption) should be improved to fully realizethe advantages of river discharge DA through the assimilation of satellitealtimetry. This study contributes to the development of a global riverdischarge reanalysis product that is consistent spatially and temporally.

[学科分类] 妇产科学