Modelling Long Term Forest Fire Risk Using Fire Weather Index Under Climate Change in Turkey

Abstract

Fire weather indices (FWIs) are among the most effective techniques to define real time or long term forest fire risk using meteorological data. In this research, long term forest fire risk of Turkey was modelled using a fire weather index called F index for present (1990 - 2010) and future (2061 - 2080) periods. Dry bulb temperature, relative humidity and maximum wind speed were mapped using 945 meteorological stations in Turkey, with a spatial resolution of 250 m. Long term mean F index values (from 1990 to 2010 and from 2061 to 2080) were calculated for 7 months representing fire seasons from April to October. Average fire occurrence of each month and monthly mean F index values of the forestlands were correlated using Pearson correlation statistic and determination coefficiency (R-2) was 0.82. Additionally, projected annual mean temperature and humidity based on HadGEM2-ES model RCP 4.5 scenario were used to derive future F index. Mean F index values of the forestlands were shown that forest fire risk of Turkey will have an increase of 21.1% in 2070s.