Estimating Global River Bathymetry by Assimilating Synthetic SWOT Measurements

The next-generation satellite altimetry mission, Surface Water and Ocean Topography (SWOT) intends to provide simultaneous mapping of inundation area and water surface elevation (WSE) of inland waters (i.e., river, lakes, wetlands, and reservoirs), using a Ka-band radar interferometer. The 2-dimensional dynamic WSE maps by SWOT can be used to measure the storage and extent changes of terrestrial waters and also to estimate river discharge.

The river channel cross-section will not be fully determined from SWOT as it cannot measure the bathymetry under the water surface. Instead, the SWOT directly measures the changes in water depth and the cross-sectional area above the lowest measured WSE. To improve discharge estimation from SWOT measurements, it is necessary to estimate true channel cross-section. Data assimilation can be used to estimate unobserved variables from the satellite measurements.

We conduct an observing system simulation experiment (OSSE) using CaMa-Flood global river model to assimilate WSE and river bathymetry using simulated SWOT observations. The assimilation scheme consists of Local Ensemble Transform Kalman Filter (LETKF) in combination with state-parameter estimation scheme. The effectiveness of assimilation was mainly evaluated with AI, which describes how close the assimilated value gets to the true value.

The OSSE suggested that SWOT observations have the potential to improve global-scale river discharge estimation by correcting river bathymetry; however, further studies are required to apply the developed data assimilation framework to real SWOT observations in the future. In this study, we only considered one critical uncertainty: river bathymetry errors but our future work will consider other uncertainties in the hydrodynamic model, such as the Manning’s roughness coefficient. 

We believe our global frame work can be used for future real time data assimilation frame works to assimilate the model variables.