Implementation and performance evaluation of an air quality forecast modeling system (AQFMS) for Northeastern United States of America

An Air Quality Forecast Modeling System (AQFMS) has been developed and operated to provide real-time 24-hr air quality forecasts for ozone, PM2.5 and their precursors for the Northeastern US. The goal of this work is to assess the reliability of the AQFMS in predicting gas phase species, free radicals and particulate matter, and to study the uncertainties therein.; The domain wide 8-hr average O3 predictions in the summer have a mean normalized bias error of 14.5% with the AQFMS generally under predicts CO and NOx and over predicts SO2 for both summer and winter. During July 2001, the CO vs. NOx correlation at Queens College (QC) and CO vs. NOy correlation at Pinnacle State Park (PSP) are in reasonable agreement with observations while the Ozone Production Efficiency (OPE) is under predicted at PSP. During January 2004 the feature of O3 removal by titration are successfully captured by the AQFMS. The higher predicted CO vs. NOx ratio at QC suggests more uncertainties with winter emissions.; For PM2.5 predictions, the AQFMS has better performance in predicting sulfate in summer than in winter. The larger portion of sulfate production from heterogeneous reactions may have added more uncertainties in winter sulfate predictions. The predictions for organic carbon particulate matter are biased low domain wide, with significant spatial variability. The analysis of prediction data show that at QC the oxidation of toluene and xylene contribute 67% and 86% of the total secondary organic carbon (SOC) production for summer and winter respectively, with 13% of the summer SOC derived from the oxidation of biogenic VOCs.; In addition, this research has revealed that the introduction of HONO heterogeneous formation and HONO emission improves both the prediction of HONO, and HOx (OH and HO2) significantly. The predicted OH concentrations increased by 25% during summer and by a factor of 2.5 during winter. The predictions for ozone concentrations, OPE and production of secondary PM2.5 have also improved accordingly.