Thesis (M.Eng.) - Asian Institute of Technology, 2002

A thesis submitted in partial fulfillment of the requirements for the degree of Master of Engineering. School of Environment, Resources and Development

This study aimed to develop synoptic climatological model to predict photochemical smog pollution potential in Bangkok Metropolitan Region (BMR), which accounts for both regional and local meteorological conditions affecting pollutant dispersion. A composite procedure utilizing principal component analysis (PCA) followed by a twostage clustering was used to identify objectively distinct synoptic meteorological patterns. The procedure was applied for 9-year data (1992-2000) during the high ozone months (November-May) at 0700 LST. The typical ozone pollutant levels and spatial distribution in each of identified synoptic patterns were then determined for 3 years of available ozone monitoring ( 1998-2000). Two approaches were used in this study. The first approach using 12 meteorological parameters from the Bangkok Metropolitan weather station, which exhibited six distinct synoptic categories, but could not produce significant differences in ozone levels and spatial distribution between the synoptic categories. The second approach used regional meteorological data form 13 weather stations including Bangkok showed six distinct synoptic categories with a better distinction in ozone levels and spatial distribution over BMR. Stepwise multiple linear regression was used to develop models to estimate daily maximum 1-h average ozone concentration. The results of the multiple linear regression analysis for the period of 1998-2000 revealed reasonable agreements between the predicted and observed ozone concentration for all six synoptic categories. The coefficients of determination (R2 ) which indicate the proportion of the total variance of the maximum hourly ozone concentrations explained by the models, are 73%, 68%, 85%, 61 %, 83% and 98% for synoptic cluster one, two, three, four, five and six, respectively. Corresponding root mean square (RMSEs) are 15.8, 20.4, 11.5, 19.7, 12.8 and 4.0 ppb, and relative errors are 19%, 27%, 14%, 26%, 17% and 4%, respectively. A procedure was developed to produce warning signal of unfavorable meteorological conditions with potential high ozone concentrations during the day based on the 0700 LST meteorological conditions for BMR. There was not clear dependence of ozone levels on the days of a week. However, Monday, Tuesday and Friday showed a slightly higher frequency of occurrence in the clusters with high ozone, and Saturday and Sunday in the cluster with low ozone.