Popis: |
UIDB/04209/2020 UIDP/04209/2020 The lengthening of the fire season in Portugal has been understudied, although there is evidence of changes in recent decades. In this brief study, we focused on the annual cycles of fire activity and related meteorological fire danger, thus bypassing the problem of subjective definition of what should mark the beginning and end of the fire season. Based on the daily time series of burned areas and occurrences collected in mainland Portugal by state services from 1980 to 2018, we searched for significant changes in the monthly and daily relative in-year distributions. In particular, an exceedance date method was used to determine day-scale trends in the anticipation/prolongation of fire activity in the year. We found an unequivocal diminishing trend in the proportion of annual burned area that occurs in the summer months (July to September), which was offset by an increase in the fraction from January to June. Apparently, the month of March plays an important role in these changes that should be clarified in future analyses. The daily analysis shows a clear shift of the cumulative curves of burned area to the left, which suggests an in-year anticipation of fire activity. Considering the dates of exceedance of the 15% of annual burned area, the found linear trend implies an anticipation rate of 1.34 days per year. Next, we evaluated seasonal changes in fire weather through the annual cycle of cumulative DSR, computed from daily FWI values. The results show, on average, an earlier increase in DSR when comparing the latter second half of the period (2000-2018) with the first one former (1980-1999). This difference develops from April to the end of September. Interestingly, a closer look shows a first increase in the gap as early as March, which disappears in April before steadily increasing again. Finally, we suggested that up-to-date knowledge of the interannual and interdecadal changes in the seasonality of both wildfires and fire weather is a decisive component, but not the only one, of a comprehensive and complex management of the fire season. A good illustration of this complexity is the fact that the annual definition of the fire season window necessarily affects the annual fire cycle by conditioning the ignition calendar. publishersversion published |