France Develops AI-Based Products to Predict Disasters

17th February 2025

1 minute read

France Develops AI-Based Products to Predict Disasters

The French meteorological service is currently working on new artificial intelligence-based products to improve weather disaster forecasting and ensure public safety and property protection.

A report published by the French website Futura Sciences states that artificial intelligence has become the new hope for meteorology. All forecasting organizations are now betting on AI, and the French Meteorological Agency is also investing in this field, according to Bruno Pradel, a data scientist at the AI laboratory of the French meteorological service, who is working on developing future forecasting products.

According to Pradel, one of the main revolutions brought by AI in meteorology is the speed at which forecasts can be calculated. Traditional models require an hour of computation to generate a forecast, whereas AI can achieve the same result in just one minute after a learning phase.

He adds: “This means that weather forecasting models will be able to benefit from continuous updates, which is not currently available.”

Pradel emphasizes that the key objective is to enhance the prediction of extreme weather phenomena that could lead to disasters. This is precisely the goal of the Espresso program (Estimation of Precipitation Using Satellite Observations), which already allows for better anticipation of rare events at a lower cost, according to the report.

The Espresso concept aims to estimate weather events in areas that are traditionally difficult to access, such as oceans, mountains, and overseas territories.

Bruno Pradel explains that this technology enables real-time observations thanks to satellite images updated every 15 minutes.

Regarding hurricanes, this tool represents a significant advancement for meteorologists. It has already been used by the crisis management unit and played a crucial role in decision-making during Cyclones Chidou and Dikilidi in Mayotte. According to the report, forecasters who have used this tool have provided very positive feedback.