Abstract
Principal component analysis (PCA) and a two-stage clustering were used to classify the long-term meteorological conditions over Thailand into six synoptic meteorological patterns. The procedure was applied to daily 0700 LST meteorological data for eight years (1992-1999) during months with high ozone in BMR (November-May). The level and spatial distribution of the highest 1-h ozone in Bangkok associated with each pattern were determined.
Two approaches were used in this study. The first approach, using meteorological parameters from a single Bangkok Metropolitan weather station, could not produce significant differences in spatial ozone distribution patterns among the synoptic categories.
The second approach used regional meteorological data from nine weather stations including 4 in Thailand showed a better distinction in spatial distribution. Stepwise multiple linear regression models were developed based on 1998-1999 hourly ozone data to predict the highest 1-h ozone in Bangkok.
|