Testing Algorithms

One of the novelties announced from the SWOT mission is the river discharge estimation. However, this will not be a direct measurement, but rather a computation made from the different data retrieved by the satellite, plus possibly some ancillary data.

Numerous hydraulic models are able to compute discharges from water heights, width and slope. The novelty in this research area is the ability of these models to deal with unknowns (or sparsely known) river properties, namely the bathymetry and the bed roughness.

The benchmark conducted on 16 rivers with simplified hydraulic models (PEPSI challenge with spatially averaged 1D models) showed that further investigations are required to improve the compromise between robustness and accuracy of a discharge estimate from SWOT data. Particularly the ICUBE - IMFT - IMT research team has proven that the equifinality (non-uniqueness of the identification) between the bathymetry and the bed roughness has to be adequately addressed. An experiment using a single in-situ bathymetry measurement for the rivers of the benchmark gave promising discharge identifications (see Figure, top-left). The conclusions of the benchmark also form a good basis for the future studies in this topic. As such, ongoing studies are conducted to improve these simplified models and the use of complementary datasets.

SWOT data may also represent a wealth of hydraulic information for full 1D hydraulic models (Saint-Venant) that can be used for a majority of river reaches, and 2D models that may be needed for larger rivers, or more complex hydraulic features such as confluences, floodplains or braided rivers.