Thesis (M.Sc.)--Chulalongkorn University, 2013

Multiple linear regression models were constructed to characterize ground-level O3 metrics in Bangkok Metropolis Region where meteorological parameters are different from other studies in cold cities. SAS® 9.2 software analyzed 2.9-million hourly data during 1997 – 2011 including O3, NO2 and meteorological variables such as temperature (T), rainfall (RF), relative humidity (RH), pressure (P), solar radiation (SR), wind speed (WS) and wind direction (WD). These data were classified into 3 seasons that were summer, rainy and winter. O3 had negatively correlated with RH and RF and positively correlated with SR and previous day O3 (O3(d-1)). Regression results showed that the lnO3(d-1) was a main positive predictor and RH is the strongest negative predictor following by a positive SR predictor. These results reveal that high SR and O3(d-1) with low RH caused an increase of ground-level O3 Multicollinearity between predictors was tested and the results showed that there was no multicollinearity. For validation analysis, the lnO3 daily maximum and daytime average in summer show the highest R2 values at 0.573 and 0.568 respectively. This work investigated the effects of Bangkok tropical climate parameters influencing O3 metrics. We tested for seasonal difference of daily O3 and meteorological parameters among 3 seasons and investigated O3 levels in meteorologically extreme days vs. meteorologically normal days by season. Our results showed that hourly O3 and meteorological parameters were concurrently peak at the same time, 13:00-14:00 h. ANOVA mean comparisons of 2 ozone variables and 3 extreme meteorological variables (maximum T and SR and minimum RH) were statistically different for all 3 seasons (p-value <0.001). This indicated that seasonal variation of tropical wet BMR significantly controlled over daily O3. T-test comparisons showed that both daily O3 average and daily maximum were higher in meteorologically extreme days than in meteorologically normal days in most comparison pairs regardless of meteorological parameter type and season (p-value <0.001). Large differences between O3 means of extreme days vs. normal days were found in RH effect investigation in all seasons especially for daily O3 maximum due purely to the strong effect of low RH in promoting O3 level regardless of season. Large differences between O3 means (average and maximum) were most pronounced in winter especially with extremely low RH and extremely high SR but not with extremely high temperature. We observed that season-specific extreme meteorological conditions in BMR tropical wet area could enhance O3 production and accumulation.