Comparison of fire-produced gases from wind tunnel and small field experimental burns

Abstract

Composition of pyrolysis gases for wildland fuels is often determined using ground samples heated in non-oxidising environments. Results are applied to wildland fires where fuels change spatially and temporally, resulting in variable fire behaviour with variable heating. Though historically used, applicability of traditional pyrolysis results to the wildland fire setting is unknown. Pyrolytic and flaming combustion gases measured in wind tunnel fires and prescribed burns were compared using compositional data techniques. CO2 was dominant in both. Other dominant gases included CO, H2 and CH4. Relative amounts of CO, CO2 and CH4 were similar between fire phases (pyrolysis, flaming combustion); relatively more H2 was observed in pyrolysis samples. All gas log-ratios with CO2 in pyrolysis samples were larger than in flaming combustion samples. Presence of live plants significantly affected gas composition. A logistic regression model correctly classified 76% of the wind tunnel samples as pyrolysis or flaming combustion based on gas composition. The model predicted 60% of the field samples originated from pyrolysis. Fire location (wind tunnel, field) and fire phase affected gas composition. The compositional approach enabled analysis and modelling of gas compositions, producing results consistent with the basic characteristics of the data.