Natural disasters have a strong impact on West African cities, and climatic and demographic changes tend to increase their severity.

The STEWARd project proposes to study, from a physical and social point of view, high impact meteorological events (intense rains, droughts, heat waves, dust storms and air quality) on several large West African cities with the aim of proposing and developing, with local partners, decision support tools for the operational early warning of risks to populations.

This project has two main objectives :

• Identify high-impact events on the population of two major West African cities, Abidjan (Ivory Coast) and Dakar (Senegal). These cities have high density and vulnerable populations, and were affected by recent hazards. The local collected observations and efficient collaborations with local partners (researchers, public institutions) will support this study. This objective will be achieved through a multidisciplinary approach, by developing two new databases. The first will focus on hazards and will take into account a wide range of characteristics (definition, occurrence, intensities) as well as the environment in which they occur. The other database will be as accurate as possible a census of the impacts on the population (mortality, morbidity). It is thus a question of two crossed approaches, one focused on physics to define and monitor hazards and the second on social sciences to focus on impacts.
• Extraction of robust, meaningful, and understandable early warning indices of weather hazards from probabilistic forecasts of weather hazards and the complex teleconnections between these hazards and their impacts identified in the first phase of the program. This will be achieved through a better understanding of the mechanisms through case studies, statistical models and investigations of the sources of uncertainty (related to the timing of the forecasts, the spatial and temporal scales of the hazards and the type of hazard). Differences between cities will be analyzed in detail to initiate the spatial variability of these hazard forecasts.