Francesca di Giuseppi (ECMWF) - From Fire Danger to Fire Activity Forecast
Francesca Di Giuseppe discusses ECMWF fire prediction systems and the advantages of SPARKY, a new data-driven system for predicting fire activity and burned area.
Traditional fire danger indices estimate environmental conditions conducive to fire but do not directly predict when and where fires will occur. This presentation introduces a shift toward fire activity forecasting, using data-driven approaches that integrate weather, fuel conditions, and ignition sources. By leveraging machine learning and Earth observation data, these systems provide probabilistic forecasts of actual fire occurrence and spread. The approach improves early warning capabilities, reduces false alarms in fuel-limited regions, and offers actionable information for fire management, supporting more effective preparedness, response, and long-term risk mitigation strategies.