Mathematical Modeling of Gas-Phase Organic Air Pollutants

Volatile organic compounds (VOCs) play a significant role in the production of ozone in urban atmospheres. In addition, VOCs are of concern because some of them are toxic, and because the atmospheric oxidation of directly emitted VOCs can form condensable products which contribute to airborne particulate matter concentrations. In this study, a general model that relates pollutant emissions to ambient VOC concentrations is described. Model performance is evaluated both for ozone and VOCs for the August 27-29, 1987 period in the Los Angeles area using data from the Southern California Air Quality Study (SCAQS). Improved chemical composition profiles for major VOC emission sources are presented, and use of these profiles results in significant changes to previous emissions estimates for many individual VCs. Reconciliation of emission data with speciated ambient VOC concentration data from the Los Angeles area indicates that there is much more unburned gasoline in the atmosphere than the emission inventory suggests. Three photochemical airshed models are presented that predict the ambient concentrations of VOCs. The first model predicts concentrations of lumped VOC classes (e.g. lumped canes, monoalkyl benzenes). Satisfactory model performance is obtained only after the VOC emission estimates prepared by the government are scaled up to match emission rates measured in the Van Nuys tunnel. An enhanced airshed model with 53 individual VOCs represented explicitly also is described and tested. The best performance is obtained for aromatic hydrocarbons which are predicted to within ±20% by the model in most cases; concentrations of most other species are predicted to better than ±50%. Finally, a model for gas-phase toxic organic air pollutants is described and tested for species including aromatics, aldehydes, ketones, and 1,3-butadiene among others. Significant contributions to total ambient concentrations from atmospheric photochemical formation are found for formaldehyde, acetaldehyde, propionaldehyde, acrolein and methyl ethyl ketone. Therefore, control programs for some toxic air pollutants must consider photochemical formation pathways in addition to direct emissions. The novel aspects of this study include the analysis and improvement of speciated VOC emission estimates, and the development and testing of airs hed models for lumped and individual VOCs using data from SCAQS