Modeling Biomass Burning Emissions for comparison with MOPITT retrievals: Boreal Forest Case Studies

The Measurement of Pollutants in the Troposphere (MOPITT) instrument on board the Terra satellite, launched in 1999, uses gas correlation spectrometry to measure the atmospheric concentration of carbon monoxide and methane, incorporating a high horizontal and temporal resolution and also extracting information about the vertical distribution of CO through the tropospheric column. Several experiments have demonstrated the sensitivity of the MOPITT retrieval to biomass burning events that produce large fluxes of CO to the atmosphere. A quantitative assessment of the MOPITT signal must be undertaken to determine the potential of these data for testing sophisticated hypotheses generated by emissions models. This assessment, however, requires a numerical simulation of a CO emission event at appropriate spatial and temporal scales. We have evaluated two case studies from fire events in the North American boreal forest in the summer of 2000, in northern Saskatchewan and interior Alaska. Both events were large complexes of fires, 50,000 ha in Saskatchewan and over 300,000 ha in Alaska, mostly in areas of limited suppression. In both cases, the greater part of the area burned was covered by the fire during just a few days, though the fires burned for several weeks. We evaluated ground-based data from fire management agencies together with satellite products to explore the complexities of modeling emissions events at a temporal resolution appropriate for comparison with the MOPITT data. We used all available data to construct several spatially and temporally resolved emissions simulations for these fire events, representing the range of uncertainty from the data sources, as well as theoretical uncertainties in the emissions model used. These emissions simulations were then propagated into the atmosphere using very simple atmospheric models, allowing qualitative comparison with the MOPITT data set. The results obtained from this experiment will contribute improved understanding of input data requirements and modeling methods for subsequent efforts at modeling atmospheric impacts of biomass burning emissions.