FIREFLY is a data assimilation system based on the numerical modeling of fire spread and which integrates the observations available on a given event to reduce uncertainties and thus represent the envelope of possible fire behavior scenarios useful for decision-making in crisis situations.
FIREFLY was born out of a collaboration between Cerfacs and the University of Maryland in the 2010s, which led to the development of the concept of data assimilation for fire applications.
FIREFLY continues today as part of the ANR project of the same name, involving more players: CNRM/Météo-France, SPE/Université de Corse, CERTEC/Université Polytechnique de Catalogne, and Worcester Polytechnic Institute, with the ambition of combining recent advances in coupled fire-atmosphere modelling, learning, remote sensing and infrared observations.