Formation of Ozone and Growth of Aerosols in Young Smoke Plumes from Biomass Burning
The combustion of biomass is a major source of atmospheric trace gases and aerosols. Regional- and global-scale models of atmospheric chemistry and climate take estimates for these emissions and arbitrarily “mix” them into grid boxes with horizontal scales of 10-200 km. This procedure ignores the complex non-linear chemical and physical transformations that take place in the highly concentrated environment of the young smoke plumes. In addition, the observations of the smoke plume from the Timbavati savannah fire [Hobbs et al., 2003] show much higher concentrations of ozone and secondary aerosol matter (nitrate, sulfate, and organic carbon [OC]) in the smoke plume than are predicted by current atmospheric chemistry models. To address these issues, we developed a new model of the gas- and aerosol-phase chemistry of biomass burning smoke plumes called ASP (Aerosol Simulation Program). Here we use ASP to simulate the gas-phase chemistry and particle dynamics of young biomass burning smoke plumes and to estimate the errors introduced by the artificial mixing of biomass burning emissions into large-scale grid boxes. This work is the first known attempt to simultaneously simulate the dynamics, gas-phase chemistry, aerosol-phase chemistry, and radiative transfer in a young biomass burning smoke plume.
We simulated smoke plumes from three fires using ASP combined with a Lagrangian parcel model. We found that our model explained the formation of ozone in the Otavi and Alaska plumes fairly well but that our initial model simulation of the Timbavati smoke plume underestimated the formation of ozone and secondary aerosol matter. The initial model simulation for Timbavati appears to be missing a source of OH. Heterogeneous reactions of NO2 and SO2 could explain the high concentrations of OH and the rapid formation of ozone, nitrate and sulfate in the smoke plume if the uptake coefficients on smoke aerosols are large [O(10-3) and O(10-4), respectively]. Uncharacterized organic species in the smoke plume were likely responsible for the rapid formation of aerosol OC. The changes in the aerosol size distribution in our model simulations were dominated by plume dilution and condensational growth, with coagulation and nucleation having only a minor effect.
We used ASP and a 3D Eulerian model to simulate the Timbavati smoke plume. We ran two test cases. In the reference chemistry case, the uncharacterized organic species were assumed to be unreactive and heterogeneous chemistry was not included. In the expanded chemistry case, the uncharacterized organic compounds were included, as were heterogeneous reactions of NO2 and SO2 with uptake coefficients of 10-3 and 2 x 10-4, respectively. The 3D Eulerian model matched the observed plume injection height, but required a large minimum horizontal diffusion coefficient to match the observed horizontal dispersion of the plume. Smoke aerosols reduced the modeled photolysis rates within and beneath the plume by 10%-20%. The expanded chemistry case provided a better match with observations of ozone, OH, and secondary aerosol matter than the reference chemistry case, but still underestimated the observed concentrations. We find that direct measurements of OH in the young smoke plumes would be the best way to determine if heterogeneous production of HONO from NO2 is taking place, and that these measurements should be a priority for future field campaigns.
Using ASP within an Eulerian box model to evaluate the errors that can be caused by the automatic dilution of biomass burning emissions into global model grid boxes, we found that even if the chemical models for smoke plume chemistry are improved, the automatic dilution of smoke plume emissions in global models could result in large errors in predicted concentrations of O3 , NOx and aerosol species downwind of biomass burning sources. The thesis discusses several potential approaches that could reduce these errors, such as the use of higher resolution grids over regions of intense biomass burning, the use of a plume-in-grid model, or the use of a computationally-efficient parameterization of a 3D Eulerian plume chemistry model.