Uncertainties in photochemical modeling of ozone and its precursors in the eastern United States

This study examines the uncertainties associated with episodic modeling of ozone over the eastern United States. Meteorological inputs to the Urban Airshed Model (UAM-V) are derived from simulations of the Regional Atmospheric Modeling System (RAMS) and two versions of the Penn State-NCAR Mesoscale Model (MM5). Measured ozone and precursor concentrations are obtained from the U.S. Environmental Protection Agency (EPA) AIRS database.; The time series of the simulated and observed precursor concentrations are decomposed into short-term (intra-day), diurnal (24 hours), synoptic (2–21 days), and longer-term (baseline) components. It is found that not only the variability associated with the intra-day component is under-represented in the model, but also the observed and modeled precursor concentrations are uncorrelated on this scale. The results show the need for model simulations to cover longer time periods since they dictate the variability of ozone and its precursors.; To examine the sensitivity to differing meteorological inputs, modeled ozone concentrations from RAMS/UAM-V and MM5/UAM-V are compared with each other, and with observations from monitoring sites in the eastern United States. The results reveal that there is a significant amount of variability—around 20% at the 95% level of confidence—in the modeled daily maximum of 1-hr ozone concentration from one modeling system to the other. The model-to-model variability in simulated ozone levels is for most part attributable to unsystematic errors.; UAM-V simulations using meteorological inputs from two versions of MM5 that differ in the prescribed physical processes are also examined. The results reveal that the differences in prescribed physical process within the same meteorological model are also a source of significant uncertainty in ozone modeling.; The directionality for emission controls (i.e., NO_x versus VOC sensitivity) is also evaluated using hypothetical emission reduction scenarios. The results show that improvement in ozone as well as VOC sensitive and NO _x sensitive regimes are influenced by the differences in the meteorological fields. The NO_x limitation threshold of the indicator species such as H₂O₂/HNO₃ are found to exhibit model-to-model and episode-to-episode variability.