Abstract

Wildfire management is closely linked to robust forecasts of changes in wildfire risk related to meteorological conditions. This link can be bridged either through fire weather indices or through statistical techniques that directly relate atmospheric patterns to wildfire activity. In the present work the COST-733 classification schemes are applied in order to link wildfires in Greece with synoptic circulation patterns. The analysis reveals that the majority of wildfire events can be explained by a small number of specific synoptic circulations, hence reflecting the synoptic climatology of wildfires. All 8 classification schemes used, prove that the most fire-dangerous conditions in Greece are characterized by a combination of high atmospheric pressure systems located N to NW of Greece, coupled with lower pressures located over the very Eastern part of the Mediterranean, an atmospheric pressure pattern closely linked to the local Etesian winds over the Aegean Sea. During these events, the atmospheric pressure has been reported to be anomalously high, while anomalously low 500hPa geopotential heights and negative total water column anomalies were also observed. Among the various classification schemes used, the 2 Principal Component Analysis-based classifications, namely the PCT and the PXE, as well as the Leader Algorithm classification LND proved to be the best options, in terms of being capable to isolate the vast amount of fire events in a small number of classes with increased frequency of occurrence. It is estimated that these 3 schemes, in combination with medium-range to seasonal climate forecasts, could be used by wildfire risk managers to provide increased wildfire prediction accuracy.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call