Ozone precursor levels and responses to emissions reductions: Analysis of regional oxidant model results

An analysis of results from the Regional Oxidant Modeling for Northeast Transport (ROMNET) study ( U.S. EPA, 1991, EPA-450/4-91-002a) has investigated the chemical conditions under which air quality was predicted to improve with reductions in ROG and/or NO ξ emissions, or with changes in the composition of ROG emissions. The ROMNET simulations used emissions projected to the year 2005, with meteorological conditions from July 1988. Predicted concentrations of PAN, HNO 3, H 2O 2 and HCHO are shown along with O 3 for the 2005 base case, allowing limited comparisons to be made with field observations and results from other modeling studies. Predicted secondary pollutant concentrations indicate an unusual degree of photochemical activity over much of the model domain, directionally consistent with the extreme nature of the July 1988 episode. Reducing NO ξ emissions was predicted to reduce O 3 in grid cells in which reactive nitrogen (NO y ) concentrations were below about 25 ppb, but to be counterproductive for some cells with higher NO y . The New York City area where NO ξ control was predicted to be counterproductive was characterized by very high NO ξ to NO y ratios. Ozone was relatively insensitive to ROG controls in grid cells with NO y concentrations below 5–10 ppb. Comparison of unweighted ROG concentrations with concentrations weighted by HO rate constants (i.e. reactivity) showed that the latter varied less across locations. Predicted spatial gradients of NO y were generally sharper than those of reactivity-weighted ROG, supporting a dominant role for variations in NO y in controlling the sensitivity of ozone to its precursors. Reductions in reactivity-weighted ROG achieved with composition changes were similar to reductions achieved with ROG emissions cuts, explaining the similar response of ozone to these two control strategies.