As part of the implementation of the CREWS (Climate Risk & Early Warning Systems) West Africa project, WMO and INRAE have signed an agreement to provide expertise to enable INRAE to propose an analysis of existing early warning systems and recommendations for a methodological framework for forecasting flash floods in West Africa.

For INRAE, this involved providing expertise and coordination work on the analysis of existing methodologies worldwide and in the region; participating in exchanges with regional centres and other specialists and researchers involved in flood issues in the region; co-organising and conducting interviews with local services identified by the WMO; to propose to the WMO a methodology for the operational forecasting of flash floods in the region, including recommendations concerning possible approaches to be implemented in West Africa allowing for the development of knowledge in the region, as well as a consultative plan for monitoring the development and implementation; to contribute to the presentation and dissemination of the results; and to participate in meetings with the person in charge at the WMO. The INRAE expert report focused on the analysis of warning systems dedicated to flash floods.

As part of a wider assessment of early warning systems for floods in West Africa, it joins the analyses and reports produced under the coordination of the IRD (regional analysis, also in collaboration with the Université Gaston Berger, in St Louis, Senegal) and SEPIA Conseils (analysis of forecasting systems for floods in urban areas).

The aim of the assessment carried out by INRAE was to provide analytical elements that could contribute to the design of an effective methodology for the operational forecasting of flash floods in West Africa, to be developed and implemented in close collaboration with the relevant regional or national institutions. The study highlighted the challenges involved in developing a relevant and reliable product or service, adapted to the fine scale of the hydrometeorological processes that cause flash floods.

Opportunities to build on existing proven systems/platforms were discussed, and general steps for an action plan were proposed, based on the three key pillars of flood forecasting (data, models and forecasters).