Coupled hydrological-hydrodynamic and data assimilation of the entire river network of the Maroni basin using SWOT river products and other EO missions

Abstract

The SWOT mission (NASA, CNES, UK-SA, CSA) launched in December 2022 provides observations of surface of inland water bodies at unprecedented resolution and accuracy. Here we focus on the Level 2 river products (heights and widths at 200m scale) and we assess their usability in creating coupled hydrological-hydrodynamic simulations of large scale basin where in-situ data are sparse. First we use an automated toolchain that generates (i) the mesh and processed input data for the hydrological models SMASH [1] or MGB [2], (ii) the coupling points between the hydrological and the hydrodynamic models, (iii) the mesh for the hydrodynamic 1D model (DassFlow-1D [3]) using either SWOT Level2 river observations of water heights and widths or other EO missions (ICESat-2, Copernicus Sentinels). Then we conduct experiments of data-assimilation of conventionnal altimetry missions (ICESat-2, Sentinel 3), in-situ level heights and SWOT Level 2 river heights in order to correct the unobserved quantities (channel bathymetry and friction coefficient) and the inflow discharges using advanced techniques taking into account correlated effects of control variables and simulated water surface properties. The accuracy obtained using this method is assessed by comparing with the sparse existing in-situ data and in terms of physical consistency of simulated flow signatures with some EO data selected for validation. This methodology and the inference capabilities are illustrated on the Maroni basin (French Guyana) which is the first application of variational data assimilation over a multi-branch river network at basin scale. A large parameter vector composed of spatially distributed friction coefficient and channel bathymetry plus inflow/lateral hydrographs are successfully estimated at various spatio-temporal resolution given data cocktails of varying spatio-temporal densities and informative content.   [1] SMASH (Spatially distributed Modelling and ASsimilation for Hydrology) - https://smash.recover.inrae.fr/ [2] https://www.ufrgs.br/lsh/mgb/what-is-mgb-iph/ [3] https://mathhydronum.insa-toulouse.fr/codes_presentation/pres_dassflow/