FloodDAM-DT Flood Extent

Flood events simulated by 2D local-scale and high-fidelity hydrodynamic model with TELEMAC-2D (www.opentelemac.org), developed over Garonne Marmandaise (France) and Ohio Cannelton-Newburgh (US). Given a forcing inflow discharge at the upstream point, the TELEMAC-2D model simulates the water level (denoted by H [m]) and velocity (denoted by U and V [m.s−1] for the two horizontal components), based on which flood risk can be assessed.

For each study site, a number of different experiments were realized. An Open-loop (OL) experiment represents a model free run without data assimilation, whereas data assimilation (DA) experiment takes into account flood observations (in-situ and/or remote sensing data) to reduce uncertainties related to the model parameters and inputs, in order to provide reanalysis simulation by an Ensemble Kalman Filter. Remote sensing data is derived from the FloodDAM-DT FloodML mapping collection. For more information on the DA, please consult Nguyen et al, AGU Water Resources Research, 2022 and Nguyen et al, IEEE TGRS, 2024. Forecasts of water depths and velocities have been produced by using observed discharge for reanalysis and then keep a persistent discharge value for the forecast (hence called FCTQ). For more information on the forecasts, please consult Nguyen et al, IEEE JSTARS (under review). "What If" scenarios represent OL simulations forced by a hydrological discharge time-series modelled by RAPID (Routing Application for Parallel computatIon of Discharge) when multiplying the input rainfall by a factor of 1, 2 and 3. For more information on the chaining of RAPID and TELEMAC-2D, please consult Nguyen et al, IGARSS, 2023.

License available here.