Aerosol transport model evaluation of an extreme smoke episode in Southeast Asia

Biomass burning is one of many sources of particulate pollution in Southeast Asia, but its irregular spatial and temporal patterns mean that large episodes can cause acute air quality problems in urban areas. Fires in Sumatra and Borneo during September and October 2006 contributed to 24-h mean PM10 concentrations above 150 mu g m(-3) at multiple locations in Singapore and Malaysia over several days. We use the FLAMBE model of biomass burning emissions and the NAAPS model of aerosol transport and evolution to simulate these events, and compare our simulation results to 24-h average PM10 measurements from 54 stations in Singapore and Malaysia. The model simulation, including the FLAMBE smoke source as well as dust, sulfate, and sea salt aerosol species, was able to explain 50% or more of the variance in 24-h PM10 observations at 29 of 54 sites. Simulation results indicated that biomass burning smoke contributed to nearly all of the extreme PM10 observations during September November 2006, but the exact contribution of smoke was unclear because the model severely underestimated total smoke emissions. Using regression analysis at each site, the bias in the smoke aerosol flux was determined to be a factor of between 2.5 and 10, and an overall factor of 3.5 was estimated. After application of this factor, the simulated smoke aerosol concentration averaged 20% of observed PM10, and 40% of PM10 for days with 24-h average concentrations above 150 mu g m(-3). These results suggest that aerosol transport models can aid analysis of severe pollution events in Southeast Asia, but that improvements are needed in models of biomass burning smoke emissions. (C) 2010 Elsevier Ltd. All rights reserved.