Wildfires are highly disruptive events for the continuity of the electricity system, as they can directly affect electrical infrastructures, or they need transmission overhead lines to be deactivated for allowing firefighters to deal with the flames. They are also events that are extremely hard to forecast, as they mainly need human actions (or negligence) for ignition.

Fire weather indices compute wildfire danger from meteorological data, thus allowing stakeholders to take preventive measures where there is a high fire spread risk. In this paper, the Canadian Fire Weather Index (FWI) has been used to assess daily wildfire danger over the Italian territory for the past decades, using as input the high-resolution reanalysis MERIDA HRES OI, developed by RSE S.p.A. MERIDA HRES OI is a gridded meteorological dataset at hourly resolution covering Italy, with a 4km spatial resolution.

It is a reconstruction of the past meteorological conditions using a numerical weather model and integrating experimental data through Optimal Interpolation. The FWI dataset obtained from MERIDA HRES has been tested against the EFFIS Burned Area Dataset to assess its accuracy in representing past fire-prone weather, in which wildfires ignited and spread. This analysis shows that the dataset correctly classifies more than 80% of the wildfire events in the “high” danger class or above.

The study on the FWI reanalysis dataset also allowed for the development of an FWI-based wildfire danger forecasting system, currently being tested in an operational environment. A test case study is discussed to showcase the accuracy at different lead times of the FWI forecasting system prototype.