Decomposition of pollution contributors to urban ozone levels concerning regional and local scales

Research output: Journal Publications and Reviews (RGC: 21, 22, 62)21_Publication in refereed journalpeer-review

20 Scopus Citations
View graph of relations



Original languageEnglish
Pages (from-to)97-103
Journal / PublicationBuilding and Environment
Issue number1
Publication statusPublished - Mar 2012


An investigation of historical database of environmental pollutants (RSP, NO, NO2, NOx, Ozone, SO2, NO/NO2 and NO2/NOx) and meteorological parameters at selected air monitoring stations in Hong Kong, is reported in this paper. Multiple regression and principal component analysis are employed to predict ozone levels from data of pollutants and meteorological variables. The results indicated that the performance of logarithmic transformation of ozone values is better than original data; based on this transformation, the variable NO2/NOx explained 75% of variation of the ozone level at Sha Tin and 83% at Kwun Tong. The ozone level increased exponentially with increase of NO2/NOx but the critical point was judged to be 0.5. In addition, due to close relationships between these studied variables, influence of meteorological variables on ozone concentration was confirmed being incorporated in pollutant variables and could not be segregated in the regression model. The results showed that meteorological variables account for 31% and 34% of ozone variation at Sha Tin and Kwun Tong, respectively. The cloud amount was observed as the most important factor at both stations. Data of Sha Tin and Kwun Tong were compared to ascertain the effect of geographical location on ozone variation. It seems that ozone level depended more heavily on variations in pollutants and meteorological variables. It can be deduced that Kwun Tong located near Victoria harbor and surrounded by hilly terrain of Kowloon and Hong Kong Island is in a disadvantageous position in terms of ozone dispersion, which results in relatively high ozone concentration. © 2011 Elsevier Ltd.

Research Area(s)

  • Air pollutant, Multicollinearity, Multiple regression analysis, Ozone, Principal component analysis