Akademik Veri Yönetim Sistemi
JOURNAL OF FORESTRY RESEARCH, cilt.30, sa.2, ss.577-587, 2019 (SCI-Expanded)
Fuel moisture content is an important variable for forest fires because it affects fuel ignition and fire behavior. In order to accurately predict fuel ignition potential, fuel moisture content must be assessed by evaluating fire spread, fireline intensity and fuel consumption. Our objective here is to model moisture content of surface fuels in normally stocked Calabrian pine (Pinus brutia Ten.) stands in relation to weather conditions, namely temperature, relative humidity, and wind speed in the Mugla province of Turkey. All surface fuels were categorized according to diameter classes and fuel types. Six fuel categories were defined: these were 0-0.3, 0.3-0.6, and 0.6-1cm diameter classes, and cone, surface litter, and duff. Plastic containers 15x20cm in size with 1x1mm mesh size were used. Samples were taken from 09:00 to 19:00h and weighed every 2h with 0.01g precision for 10days in August. At the end of the study, samples were taken to the laboratory, oven-dried at 105 degrees C for 24h and weighed to obtain fuel-moisture contents. Weather measurements were taken from a fully automated weather station set up at the study site prior to the study. Correlation and regression analyses were carried out and models were developed to predict fuel moisture contents for desorption and adsorption phase for each fuel type categories. Practical fuel moisture prediction models were developed for dry period. Models were developed that performed well with reasonable accuracy, explaining up to 92 and 95.6% of the variability in fuel-moisture contents for desorption and adsorption phases, respectively. Validation of the models were conducted using an independent data set and known fuel moisture prediction models. The predictive power of the models was satisfactory with mean absolute error values being 1.48 and 1.02 for desorption and adsorption as compared to the 2.05 and 1.60 values for the Van Wagner's hourly litter moisture content prediction model. Results obtained in this study will be invaluable for fire management planning and modeling.