PhD in geosciences – Development of a multi-agent modelling framework for the study of interactions between hydrometeorological forecasts and emergency response operation

Intense rainfall events, which are quite frequent in the French Mediterranean region, regularly cause flash floods in small watercourses. These floods prove particularly dangerous and damaging due to their high intensity and rapid onset. Gustave Eiffel University has been working for about fifteen years on the development of short-term forecasting systems designed to anticipate these phenomena by a few hours. Better anticipation should allow civil protection and emergency authorities to have the time needed to warn at-risk populations, provide them with assistance, ensure their safety, and potentially deploy protective measures. However, due to the very rapid evolution of the intense rainfall that causes these floods, forecasts still show significant uncertainties, even for very short forecast horizons (of less than 6 hours). This PhD project aims to study how these forecast uncertainties can affect the effectiveness of emergency response operations.

The objective is to develop a multi-agent model of emergency response operations during flash floods, in order to study how the use of (uncertain) hydrometeorological forecasts would influence these operations. The multi-agent modelling work will build upon a pre-existing model, which will first need to be improved in terms of realism, then extended to a broader case study, making it possible to examine a series of successive flash flood events. The intended study area is the Gard department, which offers many advantages with regard to the thesis objectives: first, this territory is prone to frequent flash flood events; it has been studied by a large amount of research works in the field of hydrometeorological forecasting; and finally, this territory provides access to a very rich dataset on emergency interventions during flood episodes.