Global sensitivity analysis of ozone production and O3-NOx-VOC limitation based on field data

Global sensitivity analysis was performed to study the effects of model uncertainties on the predictions of urban ozone production and its limitation by NOx or VOCs. Uncertainties were assigned for hundreds of model parameters including measurements used to constrain the model as well as kinetic rate coefficients and product yields of chemical reactions. Monte Carlo simulations were run using a zero-dimensional box model with 76 representative base cases of different initial conditions extracted from the measurements of a field campaign conducted in Houston, Texas. The results showed that relative uncertainty (+/- 1 sigma) of ozone production exhibits a pattern of higher uncertainty at morning rush hour (about 30-40%) and lower uncertainty in the afternoon (about 20-30%). The model uncertainty comes mostly from uncertainties in chemical schemes while the uncertainties in measurements are less influential. The most important model parameters are generally associated with the amounts of acetaldehyde, the photolysis of HONO and HCHO(-> HO2), and the reactions of OH with NO2, HO2 with NO, and xylenes with OH. The uncertainties in these parameters could also shift the ozone-precursor relation between the NOx-sensitive and VOC-sensitive regimes. The greater values of the NO amount and the reaction rates of NO2 + OH, NO + HO2 and ISOP (isoprene peroxy radicals) + NO increase VOC-sensitivity, while greater values of aldehydes amounts and kinetic rate coefficients for reactions of OH with aldehydes and xylenes and of ISOP with HO2 increase NOx-sensitivity of ozone under the studied conditions. (C) 2012 Elsevier Ltd. All rights reserved.