Performance evaluation of an air quality forecast modeling system for a summer and winter season - Photochemical oxidants and their precursors

The predictions of O-3, precursor species and the key transit species from an Air Quality Forecast Modeling System (AQFMS) are evaluated for July 2001 and January 2004 utilizing an extensive measurement data set from the PMTACS-NY "Supersite" program. The AQFMS, consisting of a chemical transport model coupled with a mesoscale meteorological forecasting model, operated routinely over the course of the six-year study. The domain-wide 8-h average O-3 predictions in the summer season have an average mean normalized bias (MNB) of 8.6%. The AQFMS captured the day-to-day variations of O-3, CO, NOx (NOy) and SO2 at the Queens College (urban) and Pinnacle State Park (rural) sites during both summer and winter. During July 2001, the linear regressions of CO vs. NOx at Queens College and CO vs. NOy at Pinnacle State Park are in reasonable agreement with observations. However, during January 2004 the slopes of the linear regressions are significantly overestimated suggesting more uncertainties with the winter emission inventories. The Ozone Production Efficiency (OPE) is under predicted by 45% at Pinnacle State Park during July 2001 which may be caused by the underestimation of NOz removal and/or the underestimation of OH concentrations. Concentrations of HONO, a key transient species for OH production are significantly under predicted by the AQFMS indicating deficiencies in the chemical mechanism in the AQFMS. The underestimation of OH concentrations is much more significant during January 2004 which suggests larger uncertainties with chemical mechanism for winter conditions. (c) 2008 Elsevier Ltd. All rights reserved.