Work Package 2
MUFFINS
Methodology
Work Package 2

Work Package 2: Reducing uncertainties with new data and new methods

This second WP will: 1) advance the exploitation of new data (earth observation, video camera, opportunistic) and purely ML methods for improved short range nowcasting of rainfall fields based on meteorological models and radar data ; 2) apply state-of-the-art inverse methods (data assimilation) to high resolution/accurate hydro-hydraulics coupled models (from Task 1.3); 3) develop new methods to offer numerical tools with rapid computational times during crisis. This WP is therefore organized in three very complementary tasks from the point of view of the methods implemented and the common cases studies (WP 3) and databases. As a consequence, WP2 will result in uncertainty reductions on the quantities simulated with integrated models from WP1 (states/fluxes, water depth-velocities) and WP3 (risks-damages).

Task 2.1: Datasets preparation, statistical analysis, and inferences by ML

Partners involved: all project partners.

Task 2.2: Multi-source data assimilation for high resolution hydrological-hydraulic models of Task 1.3

The goal of this task is to enhance synergies between multi-source data and hydrological-hydraulic models to obtain accurate high resolution numerical representations of multi-scale floods-inundations. The following scientific directions will be tackled: (i) integration of heterogeneous multi-source data in a coupled hydrological-hydraulic high resolution models, (ii) inference of hydrological-hydraulic model parameters from multi-source data, (iii) for downstream forecasts via the assimilation of upstream observations, (iv) information feedbacks from downstream observations to upstream hydrological components within large domains.

Partners involved: INRAE, IMT

Task 2.3: New methods: hybrid ML - Data Assimilation with offline-online (real time) strategy

 

Partners involved: IMT, INRAE.

See also

Modification date: 15 June 2023 | Publication date: 26 April 2023 | By: Pierre-André Garambois