IRIS – Rapid response fire spread forecasting system: Development, calibration and evaluation

Abstract

This work presents the development of a rapid response forecasting system for the prediction of wildfire spread. Named IRIS, the forecasting system was primarily designed and developed to support the operational fire suppression activities of the Greek Fire Service. It employs a coupled atmosphere-fire modeling system, comprised of a state-of-the-art numerical weather prediction model and an advanced fire spread model, accounting for the two-way interactions between fire and weather. For the implementation of IRIS, a prototype fuel models’ geospatial dataset was constructed for Greece, exploiting high-resolution, open-source pan-European vegetation and land use datasets of the Copernicus Land Monitoring Service. Calibration of the fire spread component of IRIS was carried out by retrospective forecasting of 8 wildfires that took place in Greece during the 2016–2017 fire seasons, while the evaluation of the calibrated forecasting system was conducted by forecasting 4 major wildfires of the 2018 fire season. Our overarching aim was to assess the capacity of IRIS for providing accurate real-time wildfire spread simulations. Results clearly highlight that reasonably accurate fire spread predictions can be obtained by proper calibration of the fire spread component. IRIS will be operationally deployed in Greece during the 2019 fire season, providing 6 h and 24 h wildfire spread forecasts, when necessary, in about 15 min and 60 min, respectively. To our best knowledge, IRIS is one of the few operational fire spread forecasting systems based on coupled atmosphere-fire modeling.