Forecasting the Danger of the Forest Fire Season in North-West Patagonia, Argentina

Abstract

An average of 1300 forest fires are officiallyreported in Argentina annually, triggeredeither by natural or intentional causes. Oncean outbreak has started, topographic andvegetation (fuel) conditions and the weathersituation determine the fire behavior. Thegeneral evolution of fire danger can be monitoredthrough the Fire Weather Index (FWI),part of the Canadian Forest Fire Danger RatingSystem (CFFDRS). This indicator was successfullyimplemented in Argentina since thebeginning of 2000. It is based on temperature,relative humidity, wind intensity, and precipitation.Among its advantages, it allows evaluatingthe possibility of fire ignition, intensity,and spreading. The objective of this work is toachieve a characterization of the danger of thedifferent forest fire seasons in the northernArgentinean Patagonian mountain range,based on Bariloche station data, using theFWI and find a method to forecast the danger.For this purpose, the index was related toatmospheric circulation and sea surface temperaturepatterns that dominate the regionbefore and during the fire season, using thecorrelation method. It allowed defining predictorsthat were used to generate somestatistical models, to predict the dangerousnessof the future seasons. Sea surface temperatures,anomalies of wind, and the AntarcticOscillation were the best predictors for thispurpose. Some derived statistical models arevery useful tools to improve the decisionmakingregarding the coming fire seasons inthis particular area. The incorporation ofnon-linear techniques seems to improve theefficiency of linear regression models.