From Global Composition to Regional Air Quality: The Influence of Near-Real-Time Satellite Observations on Chemical Forecasts during the 2006 INTEX-B and TEXAQS Campaigns

The global component of the Realtime Air Quality Modeling System (RAQMS global) was used to provide 1- to 5-day atmospheric chemical composition forecasts for aircraft flight planning during the 2006 INTEX-B (Intercontinental Transport Experiment phase B) campaign and for satellite/insitu data synthesis during the 2006 TEXAQS (TEXas Air Quality Study) campaign. A unique aspect of these predictions is the infusion of near-real-time satellite data to improve the chemical forecasts, including assimilation of ozone observations from OMI and the use of MODIS fire counts to generate biomass burning emissions. These data infusions introduce additional observational constraints into the chemical forecast. However, there are large uncertainties inherent in inferring chemical emissions from fire counts. We assess the fidelity of the 2006 RAQMS global composition forecasts through comparisons with remote and insitu measurements including satellite, ozonesonde, aircraft and surface measurements. We focus on North American air quality during the intensive measurement periods of INTEX-B and TEXAQS. We then analyze locations and times of model forecast successes and failures to identify dominant sources of model composition forecast errors, including meteorological forecast uncertainty, emissions uncertainty, and uncertainties associated with the photochemical calculations. In particular, because biomass burning emissions are episodically quite large and vary on timescales from daily to interannual, their accuracy and uncertainty directly impact the accuracy and uncertainty of atmospheric composition forecasts.