Sensitivity of ozone to precursor emissions in urban Beijing with a Monte Carlo scheme

In order to understand the formation mechanisms of high surface ozone and identify the main contributor sources in Beijing, this study investigates the sensitivity of surface ozone to NO, NO2 and nine types of NMVOC emissions during a photochemical smog episode. Monte Carlo sensitivity analysis scheme with fifty simulations is established based on the Nested Air Quality Prediction Model System (NAQPMS). At every simulation, each of the eleven precursor emissions is perturbed with a distinct set of perturbations. The sensitivities of ozone to emissions are identified by multiple linear regressions. The stability of sensitivity results is validated by two experiments with standard deviations of log-normal perturbations set as 30% and 50% respectively. The sensitivity results suggest that the current high surface ozone is strongly stimulated by NMVOC emissions. Among NMVOC emissions, formaldehyde, ethylene and olefins emissions present the greatest impacts on ozone. On the other hand, NOx emissions have a strong inhibitory effect on ozone formation, even after 50% NOx emission reduction. This indicates that the current ozone formation in Beijing is under NOx-saturated conditions. A transition of ozone formation is observed from NOx-saturated to NOx-limited sensitivity behavior with a 75% reduction of NOx emissions. This study gives the implication that abatement of the four NMVOC types mentioned above could be efficient on reducing the high levels of surface ozone in central urban Beijing, while inadequate abatement in NOx emissions probably induces reverse effects. (C) 2010 Elsevier Ltd. All rights reserved.