Fire Danger Tool – Forest Preparation and Preparedness

The Fire Danger Tool provides an indication of the Fire Danger Index (FDI) in the region of interest. The FDI is provided through a ML-based approach that embeds both spatial and temporal dimensions and it represents the probability of forest fire occurrence. The ML method takes into account not only the weather variables, like the traditional Canadian FWI, but also the availability of biomass (vegetation) and human factors, as it is very well understood that a large percentage of forest fires are caused by human activity. The ML-based FDI is produced as a probability measure between 0 and 1 which is mapped into six classes, from very low (0) to very high risk (5). In addition to the daily FDI, the UP also provides the empirical Fire Weather Index (FWI) which only takes into account weather factors such as temperature, wind speed, relative humidity and precipitation.

The Fire Danger Tool contributes to Phase A providing information about the conditions that reflect the potential of fire occurrence over an area for a fire to ignite and spread. In addition, it also contributes to Phase B since it provides information where the fire is most probable to ignite and spread. Therefore, it can be used as input for improving early detection of wildfire (e.g., increasing the monitor frequency by the sensors, planning drone missions to monitor areas which are at fire danger indicated by the fire danger index). 

It has several features which serve the preparation and prevention of forest fires. Among these the most important ones are the following:

• Providing the fire danger index for different temporal scales for e.g. now-casting (between 0-3 hours) daily, as well as seasonal,
• The FDI prediction is based on empirical as well as data-driven approaches, fully utilizing the state-of-the-art in ML methods, exploiting both temporal and spatial dimensions across a wide set of input predictors,
• The service is fully integrated in the Silvanus platform and FDI estimates can be visualized in the integrated Silvanus Dashboard and recovered from Silvanus SAL,

At the moment, the Fire Danger Tool covers the Gargano (Italy) pilot site providing the daily FDI estimates at 2 km of spatial resolution. Other pilots and FDI estimates at different temporal resolutions (e.g., hourly and seasonal) will be soon available through the Silvanus platform.

The following picture shows the daily ML-based FDI maps across the Gargano pilot site (Apulia, Italy) for several consecutive days in May 2024. The FDI consists of six classes, ranging from 0 (least fire danger) to 5 (maximum fire danger).