| This study is based on the photochemical air quality modeling system of the Laboratory for Atmospheric Research, Washington State University. The model system consists of MM5 (a prognostic meteorological model), CALMET (a diagnostic meteorological model), and CALGRID (an eulerian photochemical grid model). The research of this thesis is focused on applying process analysis, sensitivity analysis and uncertainty analysis to investigate details of ozone formation and transport in the Puget Sound region. The efforts have been aimed at performing photochemical air quality modeling on a comprehensive level and improving the photochemical air quality model system.; The results of process analysis show that formation of peak ozone near Enumclaw area was due to ozone chemical production and accumulation on a continuous basis from the early morning through the mid-afternoon until the air mass reached the receptor. NO titration led to near-zero nighttime ozone levels in the urban core areas, while dry deposition tended to deplete surface ozone in the rural and mountain areas during the nighttime. Sensitivity analysis indicates domain peak ozone levels always had a linear relationship with VOC emissions. The peak ozone was more sensitive to VOC than to NOx emissions. The effective strategy for controlling ozone is to reduce VOC emissions of both anthropogenic area and mobile sources. Uncertainty analysis indicated the uncertainty of simulated domain peak ozone due to emission inventory was higher than that due to wind field. The uncertainty of initial and boundary concentrations is also an important factor for the uncertainty of domain peak ozone levels. These results suggest the urgent need to improve emission inventory and the selection of initial and boundary concentrations. |
