Analysis of the relationship between the weather index of fire danger and occurrences of rural fires. Case study: centro region of portugal
¹Polytechnic University of Castelo Branco, Faculty of Agriculture, Department of Natural Resources and Sustainable Development, Quinta da Sra. de Miércoles, apartado 119, PT6001-909 Castelo Branco, Portugal
²CERNAS - Research Center for Natural Resources, Environment and Society. Escola Superior Agrária de Coimbra, Bencanta, PT3045-601 Coimbra Portugal
³MED - Mediterranean Institute for Agriculture, Environment and Development & CHANGE - Global Change and Sustainability Institute, Universidade de Évora, Pólo da Mitra, Ap. 94, PT7006-554 Évora, Portugal
⁴Instituto Português do Mar e da Atmosfera, Rua C do Aeroporto, PT1749-077 Lisboa, Portugal
⁵QRural - Qualidade De Vida No Mundo Rural, Av. Pedro Álvares Cabral, n. 12, PT6000-084 Castelo Branco, Portugal
*Correspondence: npedro@ipcb.pt
Abstract:
The aim of this study was to design an approach for establishing a plausible relationship between FWI and the monthly average burned area (ABA) and the average number of ignitions (ANI) supported by geographic information systems (GIS). The application of these results will allow the projection of burned areas in forest fires in the future, making mitigation actions possible. This approach was applied to the region of Central Portugal, and to achieve the aims of the study, the following steps were completed: (1) geoprocessing the spatial data of the daily FWI indices, burned area and number of fire ignitions and (2) developing statistical regression models capable of reproducing the variability in burned area and ignition occurrence series from FWI data during the 2001–2017 period. The predicted equations for the burned area as a function of the FWI presented high coefficients of determination for most of the considered periods, thus allowing the projection, with a high degree of confidence, of the monthly burned area values according to the various future climate scenarios. The prediction of the average number of ignitions from the FWI values class proved to be effective for establishing highly adjusted forecast models for July and August. In the spatial analysis at the district level, the ABA and ANI estimation equations were obtained from the FWI values with determination coefficients above 0.90 for most of the districts. Significant differences were observed between the districts in the number of ignitions analysed.
Key words:
burned area, climate, fire occurrence, FWI system, wildfires