Feasibility of Full Saint-Venant River Routing at Coarse Resolution in a Large-Scale Hydrological Model

Résumé fourni par la source

Large-scale river routing models often simplify the Saint-Venant equations to limit computational cost. As routing grids are refined, these simplifications may become less appropriate, especially in low-slope reaches where inertia and pressure-gradient effects can influence flood-wave propagation and water levels. This study asks whether the full Saint-Venant equations can be used in a large-scale river routing model (RRM) at coarse spatial resolution, and whether this feasibility depends on the numerical formulation used to solve them. Two full Saint-Venant formulations are implemented in the CTRIP RRM and compared with the operational kinematic routing formulation. The first follows hydraulic practice and solves the coupled continuity and momentum equations with a Preissmann scheme. The second follows a routing-oriented strategy and solves a water-storage balance followed by the full momentum equation using a Crank-Nicolson discretization and a Gauss-Seidel iterative solver. The schemes are evaluated using idealized river test cases and a 20-year simulation over the Adour catchment in southwestern France at 1/12° resolution. Results show that full Saint-Venant routing is feasible under RRM constraints, but that feasibility depends on the numerical formulation. The coupled hydraulic formulation based on a Preissmann scheme is accurate in idealized cases but does not complete the realistic 20-year simulation in its current implementation, due to numerical failures under low-water and steep-slope conditions. The uncoupled full-momentum formulation completes the long simulation when using a 60s routing time step. Compared with kinematic routing, it produces more dynamically consistent downstream water levels, while differences in daily discharge remain limited at the basin scale.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.