Understanding interannual variations of biomass burning from Peninsular Southeast Asia, part I: Model evaluation and analysis of systematic bias

The Weather Research and Forecasting model and the Community Multiscale Air Quality model (WRF/CMAQ) modeling system was applied over Peninsular Southeast Asia (PSEA) and East Asia (EA) for 5 consecutive years from 2006 to 2010 in March and April to understand the PSEA biomass burning for its interannual variations, transport pathway, and impacts at local and downwind areas. As many of the modeling applications over PSEA or EA were usually evaluated with limited regional or local observations focusing on short simulation periods, in this work we incorporated ground surface measurements from multiple networks covering different sub-regions and satellite retrievals to comprehensively examine the performance of WRF/CMAQ and probe into the possible uncertainties of the modeling system. We found increasing simulation discrepancy for CO, NO2, and SO2 from 2006 to 2010 at south part of PSEA (<= 17 degrees N) due to outmoded anthropogenic emission in INTEX-B, while local surface observations and CO2 emission data from World Bank suggested substantial growth of anthropogenic emission over PSEA during the 5 years. The spatial allocation of emission based on population distribution was also found to introduce large uncertainty with overestimation at populated urban area and underestimation at industry area. Over north PSEA (>17 degrees N) CMAQ systematically overestimated CO, surface NO2, troposheric column NO2 by around 6%-20%, 8-15%, and 40%-50% respectively, indicating positive bias within the biomass burning emission due to overestimated emission factor as suggested by OMI retrievals. At EA, despite moderate overestimations for surface NO2 and SO2 by 20%-30% and moderate underestimation for AOD by 30%-50%, no significant temporal trend was found. We found CMAQ underestimated PM10 concentrations at north and northeast EA by 50%-60% due to impact of dust storm, yet the dust plume rise scheme within the model was unable to reproduce it. Our results suggested that an urgent research effort is needed for updating the anthropogenic emission of PSEA countries, and the dust emission module within CMAQ need further improvement for applications over EA. (C) 2015 Elsevier Ltd. All rights reserved.