Background Land use and climate changes are altering wildfire regimes globally, necessitating forest fire danger rating indices that can accurately identify spatio-temporal fire danger patterns for effective wildfire risk management. Aims Many existing models primarily rely on weather conditions, often overlooking fuel and providing daily information. This study introduces the latest version of the RISICO model, developed in the early 2000s and used operationally by the Italian Civil Protection system for decades. RISICO explicitly incorporates fuel into its computations, providing hourly outputs. Methods The fuel classification used by RISICO is based on a wildfire susceptibility model developed using machine learning techniques, reducing data requirements. The model’s performance was tested against various fire danger indices, considering the past 16 years of Italian wildfires. Key results Analysis of the RISICO outputs demonstrated improved performance in accurately identifying periods and locations of increased fire danger, more precisely narrowing the areas considered at risk compared with other models from the literature. Conclusions RISICO proves to be an effective tool for operational use in civil protection contexts, thanks to its ability to identify fire danger conditions accurately and with ease of implementation. Implications This study highlights the importance of including fuel characteristics within fire danger models to enhance their discrimination ability and improve wildfire risk management.
Perello et al. (Sun,) studied this question.