This study develops specific models for estimating and forecasting fire risk based on a combination of time series analysis and remote sensing data. The risk indices used in this procedure are the FPINDWI and the FPINDVI, which differ in terms of the vegetation index used in their calculation, NDVI or NDWI. The FPI (Fire Potential Index) pools meteorological data with information from remote sensing images. Time series analysis has unearthed dynamic patterns in fire behaviour during the study period 2000-2009. From the time series for the period 2000-2008, specific forecast models were developed for the two indices by «fuel type-bioclimatic region». The results show a good fit between the original FPINDWI data and the forecasts for 2009. The study also shows that the FPINDWI’s risk forecasting accuracy is better than the FPINDVI’s, especially for the ecosystems of the north of the study region.
|Translated title of the contribution||Pre-empting fire risk|
|Number of pages||15|
|Journal||Seguridad y medio ambiente|
|Publication status||Published - 2012|