MUFFINS
Edito

Edito

Floods are the first weather related disaster worldwide, expected to increase with climate change, and effective flood warning systems are crucially needed to mitigate their devastating effects. Nevertheless, performing accurate flood forecasts in terms of location, magnitude and timing of runoff and flooding, and identifying areas ”at risk”, i.e. prone to trigger fatalities and economic losses, remains a key challenge especially for Mediterranean convective rainfall event. The development of suitable flood forecasting systems, first ones already implemented in France and worldwide, is still faced with serious challenges due to: (1) complex coupled and multi-scale physical processes; (2) lack of data and robustness especially on ungauged (often devastating) watercourses, computational limitations for high resolution accuracy over large domains and quickly evolving hydro-meteorological situations; (3) limited adequation with operational needs and risk-damages modeling.

Therefore, the overarching objective of the MUFFINS project is to propose new accurate and computationally efficient multi-scale flood forecasting approaches enabling transferring information between scales and dealing with multi-scale issues (local sudden runoff, flash floods on ungauged catchments, inundations, risks-damages), better integrating multi-source data (in situ, earth observation, opportunistic) available at the different spatial and temporal scales.
A multidisciplinary consortium, with recognized experience in the relevant scientific and operational fields, will address the key elements required to build such effective flood forecasting methods: 

  • multi-scale forecasting capabilities, from small scales (∼ 100m2) to catchment scales (∼ 5000km2) with anticipation times ranging from 1h or less to ∼ 1d, obtained by coupling short range precipitation forecasts, spatially distributed scalable and regionalized hydrological models and high resolution effective hydraulic models including hydrological processes,
  • innovative methods performed to integrate information from multi-source data and to reach fast computation times, which include model reduction and offline-online strategies combined with statistical, data assimilation, machine learning methods,
  • the impact modeling and expert specifications-evaluations for the forecasting chains, achieved thanks to multi-scale (flash) flood cases, rich datasets, damage models and expertise provided by end users. 

The case studies will be located in the French Mediterranean region, especially in the Var, Bouches du Rhˆone and Alpes-Maritime departments where several recent flood events offer a particularly favorable context in terms of data availability (including damage data) and of applicability of the panel of multiscale approaches developed. These events include the June 2010 and October-November 2018 floods in the Argens watershed, the January 2014 floods in the Gapeau watersheds, the October 2015 and October 2020 floods in the Alpes-Maritimes region. Rich datasets composed of physiographic and hydrometric datasets, post-field surveys and new data (earth observation, video camera, opportunistic) will enable improving models accuracy.
The main expected benefits of the project should be: - New methodologies and tools for next generation of spatially distributed flood forecasting methods with multi-sourced data integration, affordable in real time and transferable to researchers and end users at national, regional and local scales. - Operational demonstrators of hydrological-hydraulic flood forecasting methods affordable in real-time and refined risk-cost modeling. - Upgrade of the VigicrueFlash operationnal flood forecasting chain for ungauged catchments and methods for future infra-departmental flood vigilance, such as runoff sensitivity mapping. - Improved capability of insurance or reinsurance (CCR) companies to take effective action after each event, through anticipating the volume of claims filed.

Modification date: 04 May 2023 | Publication date: 09 June 2022