Anticipating the severity of the fire season in Northern Portugal using statistical models based on meteorological indices of fire danger

Abstract

Climate and weather are major drivers of fire activity in Portugal. The aim of the present study is to assess the role of meteorological factors on the inter-annual variability of burned area, for the period 1980-2011, over a region of Central Portugal. Although occupying only 18% of the territory of Portugal, the chosen area is responsible for 43% of all the burned area in August during the study period. A normal distribution model is fitted to the decimal logarithms of monthly burned area during August over the study area. This model is then improved by introducing as covariates two different measures of prevailing meteorological conditions as derived from Daily Severity Rate (DSR), namely a top-down factor which consists of daily cumulated values of DSR from April up to July and a bottom-up factor defined as the square root of the mean of the squared daily positive deviations of DSR in August from the daily climatology. The two models are used to derive a model of fire severity that allows quantifying the probability of having a severe or weak fire season.