5.4 Assessment of an integrated fire rating index for Sardinia, Italy

Wednesday, 24 October 2007: 11:00 AM
The Turrets (Atlantic Oakes Resort)
Costantino Sirca, Univ. of Sassari, Sassari, Italy; and D. Spano and P. Duce

Ichnusa Fire Index (IFI) is an integrated fire danger index developed for Sardinia, Italy. IFI includes a Drought Code (DC) describing water status of plants, a Meteo Code (MC) related to turbulence and weather conditions, a Fuel Code (FC) that accounts for structure and moisture of vegetation, and a Topological Code (TC) describing slope and aspect of the site. IFI values were calculated using the software IFI ver. 1.3. In this paper, preliminary results on the IFI performance are reported and compared with some common used fire danger indexes (e.g., FWI, Port and IMPI). The accuracy of IFI in predicting days with or without fire events was assessed using several statistics (e.g., Mahalanobis distance and the Score 1, 2 and 3). The IFI performance was assessed for the years 2000-2006 in 57 areas corresponding to weather stations of Agrometeorological Sardinian Service - SAR. Fire events were included in the statistical analysis. In addition, IFI values were calculated for 2005 and 2006 summer periods using weather forecast from a limited area model (BOLAM) with a 5 km x 5 km grid resolution developed by the SAR. The analysis included 3422 fire events that burned 14962 ha. IFI showed the highest Mahalanobis distance test, indicating better results when compared with FWI, Port and IMPI. Also, Scores 1, 2 and 3 confirmed the good performance of IFI in comparison with the other indexes. A good relationship of IFI forecast with both fire event number and burned surfaces indicates the promising capability of IFI to give realistic forecast of fire events in Sardinia, Italy. In addition, weather forecasts over a limited size cell grid can give useful information to Regional Forest Services for improving forest fire management.
- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner