[1] We investigated trace gas emissions from a biomass fire near Otavi, northern Namibia, on 13 September 2000 as part of the Southern African Regional Science Initiative 2000 (SAFARI 2000). Observations included fast measurements of carbon monoxide (CO), ozone (O 3 ), acetone (propanone, CH 3 COCH 3 ), and acetonitrile (ethanenitrile or methyl cyanide, CH 3 CN). Additionally, flask samples were taken and analyzed for nonmethane hydrocarbons (NMHCs). Measurements close to the fire were used to quantify the emissions of the different compounds. Several transects were flown through the plume downwind to study the chemical changes in the plume during the first 2 hours after the emission. Fast production of O 3 was observed, with the O 3 /CO molar enhancement ratio ER(O 3 /CO) reaching $0.10 in the plume after 2 hours. Acetone molar enhancement ratios (ER(acetone/CO) were $5 Â 10 À3 in the very young plume but $9 Â 10 À3 after 1-2 hours of aging, pointing to a substantial fast secondary acetone formation. To understand the chemical processes occurring in the plume, we simulated the plume chemistry with a dilution box model. The fast O 3 production is well reproduced, whereas acetone mixing ratios in the aging plume are underestimated by the model. This points to additional yet unidentified acetone precursors. The model suggests an overall molar enhancement ratio of these compounds with respect to CO of 7.5 Â 10 À3 and a reaction rate coefficient of 6 Â 10 À11 cm 3 s À1 for the reaction with HO radicals. The secondary acetone production may cause the net acetone source from biomass burning to be underestimated when only data from observations immediately near the fires are considered. The atmospheric implications of these findings are discussed.

On the other hand, high mixing ratios of sulfur dioxide (up to 1.5 ppb) and aerosol sulfate (up to 3 ppb) indicate the influence of fossil fuel burning. During most flights the contributions from these two sources were well mixed within the same air mass, suggesting that the sources on the ground are also close to each other. This is consistent with the assumption that biomass is mainly burnt as biofuel for domestic use in populated areas, where fossil fuel is also used. The ratios dX/dCO (X:acetone, acetonitrile, sulfur dioxide, potassium, or sulfate) measured during the flights indicate that most of the CO in the continental outflow is due to biomass or biofuel burning, whereas the majority of the aerosols results from fossil fuel burning.