Air quality pattern assessment in Peninsular Malaysia using statistical approach / Nur Syuhada Aini Ahmad Lokman Bukhari

The causes of air quality problems detected are the high increase in the number of vehicles, the rapidly growing urbanization and industrialization that indirectly poses a threat to the environment and public health. The primary objective of this research was to analyze air quality pattern in Peninsular Malaysia utilizing cluster analysis. The sub-objectives of this study were to classify the air quality pattern using hierarchical cluster analysis, to differentiate between clusters using discriminant analysis and also to identify sources of air pollution using principal component analysis. The data was provided by the Department of Environment (DOE) of Malaysia for the years 2013 until 2016 (4 years). Basically, parameters of air pollution index (API) namely nitrogen dioxide (NO2), carbon monoxide (CO), particulate matter, ozone (O3), sulfur dioxide (SO2) and particulate matter less than10 microns (PM10) were involved in this study. Therefore, statistical approach which is cluster analysis (CA) was used in this study. It cluster three smaller classes compared to 48 air monitoring stations which has the same characteristics. Hierarchical cluster analysis was choosen to study the pattern of air pollution index(API) based on a yearly and monthly basis. It was shown that 2015 has more standout and different pattern. To distinguish between three classes, Discriminant Analysis (DA) has been used where it uses standard, backward and forward stepwise technique. This study also used Wilks’ lambda test to measure how well each variables separates into groups. Principal Component Analysis (PCA) was used to identify significant pollutant parameters which caused the air quality problem in the peninsular area. The study found that CO and PM10 are the major pollutants that have contributed to degrade air quality in the Peninsular Malaysia based on the three classes due to combustion process from vehicles and industries. As conclusion, analyzing huge data sets become better understanding on air quality and also more clearly identify significant air pollutant parameters when using statistical approach.