• English 
  • Spanish 

Fire activity as a function of fire-weather seasonal severity and antecedent climate across scales in southern Europe and Pacific western USA

Journal: Environmental Research Letters
Year: 2015  
Status: Published
In this status since: 7 Nov 2015
PDF file: 2015_Urbieta_ERL.pdf
DOI: 10.1088/1748-9326/10/11/114013
Authors:
Urbieta, I., Zavala, G., , , San-Miguel-Ayanz, J., Camia, A., Keeley, J., Moreno, J.M.

Climate has a strong but complex influence on fire activity that can vary across areas and scales. We analyzed how fire-weather conditions during the main fire season and antecedent water-balance conditions affected inter-annual changes in fire activity in two Mediterranean regions with different management histories. In the southern countries of the European Union (EUMED) all fires in the territory were analyzed covering a wide longitudinal gradient; in the Pacific western coast of the USA (PWUSA) national forest fire data were analyzed representing a longer latitudinal gradient. Total number of fires (≥1 ha), number of large fires (≥100 ha) and area burned were related to seasonal values of the Fire Weather Index (FWI), and to the Standardized Precipitation-Evapotranspiration Index (SPEI) accumulated in the preceding 3 (spring) or 8 (autumn through spring) months. Calculations were made at three spatial aggregations in each area, and models related first-difference (year-to-year change) of fire and FWI/climate variables to minimize autocorrelation. Across areas and scales, an increase in mean FWI conditions during the fire season resulted in greater fire activity for the three fire variables tested. SPEI contributed little to explain fires, with few exceptions. Negative water-balance (drought) in the antecedent 8 months was more important than positive conditions (moist) in spring, both of which contributed positively to fires. The relationships were similar in both regions, albeit weaker in PWUSA, and the R2 of the models generally increased with the area of aggregation. For total number of fires and area burned, the R2 of the models tended to decrease along a gradient of long-term mean seasonal FWI, suggesting that in areas with less amenable conditions for fires (lower FWI), fires were more susceptible to change with climate variability than in areas with higher mean FWI values. This has implications for anticipating fires with increasing seasonal FWI due to global warming. The large variance explained by some of the models indicates that large-scale seasonal forecast could help anticipating fire activity in the investigated areas.