Fire in African savanna: testing the impact of incomplete combustion on pyrogenic emissions estimates

• Published in: Ecological applications 2005-06, Vol.15 (3), p.1074-1082
• Main Authors: Smith, A.M.S, Wooster, M.J, Drake, N.A, Dipotso, F.M, Perry, G.L.W
• Format: Article
• Language: English
• Subjects: ash; Ashes; Biomass burning; biomass burning and global climate change; Botswana; carbon; Combustion; combustion in natural fires; Fire ecology; Forest fires; Fuel combustion; gas emissions; global budgets; global emission budgets; Grassland fires; loss on ignition (LOI); mathematical models; N and C emissions; Nitrogen; particulate emissions; savannas; Wildfire seasons; wildfires; Wildland firefighting
• ISSN: 1051-0761; 1939-5582
• DOI: 10.1890/03-5256

African savannah fires are key sources of trace gases and aerosols, yet their true magnitude remains relatively poorly constrained. Here we present a detailed investigation of the amount of unburned material remaining in the ash produced by such fires, and examine whether this quantity needs to be considered when calculating atmospheric emissions. Emissions estimates for individual fires are usually based upon a calculation of the amount of fuel biomass combusted, usually estimated via a "simple subtraction" of the pre- and post-fire fuel loads. However, certain studies have concluded that this approach leads to gross overestimation of emissions, by perhaps 100%, due to the fact that a proportion of unburned fuel normally remains in the "charred" ash and that the simple subtraction method fails to take this fully into account. Whilst this latter point is true, we show that the mathematical relationship used previously to calculate the implications of this for emissions calculation is flawed. We demonstrate a new first-principles derivation of the quantity of mass combusted, taking into account any fuel remaining in the ash via measurement of ash mass loss on ignition (LOI). Using data from dry-season experimental fires in Botswana, we compared estimates of biomass combusted calculated via our LOI-based method to those obtained via the simple subtraction approach. We found a mean difference of < 10%, far less than the factor of 2 difference obtained when using the previous erroneous formula. Similar results are obtained for carbon and nitrogen emissions, and the findings are further supported by reassessment of previous data from early-to-mid-dry season savanna fires in Tanzania. These results indicate that the simple subtraction method overestimates emissions by far less than previously suggested, and that the method is likely to provide sufficient accuracy in most situations where emissions calculations are required. For fires where very substantial amounts of unburned material are expected to remain within the ash, and where an LOI-based approach to emissions calculation may therefore be particularly warranted, it is important that the correct formula derived herein is used, since errors inherent in the previous formula increase with increasing ash LOI.