Spatial distribution and source apportionment of air pollution in Malaysia through environmetric techniques
This research involves the analyses of secondary air quality data collected at twelve monitoring stations in Malaysia between 2001 and 2009. Several environmetric techniques were applied on this nine-year daily average database. The environmetric techniques incorporated discriminant analysis (DA) t...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English |
| Published: |
2013
|
| Subjects: | |
| Online Access: | http://psasir.upm.edu.my/id/eprint/41233/1/FPAS%202013%208R.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | This research involves the analyses of secondary air quality data collected at twelve monitoring stations in Malaysia between 2001 and 2009. Several environmetric
techniques were applied on this nine-year daily average database. The environmetric techniques incorporated discriminant analysis (DA) to investigate the significant
discriminating air quality variables, hierarchical agglomerative cluster analysis (HACA) to access the spatial air quality patterns and principal component analysis
(PCA) to determine the probable sources of air pollutants. The artificial neural network (ANN) analysis used to determine the air quality model structure as well as
the most significant variable that influenced the Air Pollutant Index (API). A combined receptor model (PCA/MLR) was then used is to assess the source apportionment of the significant variables. The DA computed nine significant
variables to discriminate the five levels of air quality. HACA grouped the twelve air monitoring stations into three different clusters. The PCA results showed that the probable sources of air pollution within the study areas were combustion of fuels in all modes of transportation, offshore oil installation, agriculture operations,
combustion of wood and industrial activities. For the overall air quality spatial assessment, ANN produced the best fit model with high R2 values (0.803 ≤ R2 ≤0.807, p<0.05). It also revealed that more than 80% of the air quality variability is explained by the nine significant variables (CO, O3, PM10, NO2, SO2, temperature,humidity and wind speed). Further, the ANN analysis showed that among the nine significant variables, PM10 was the most important variable that influenced the API value variation. In addition, the combined receptor model (PCA/MLR) showed that
in all three clusters, more than 70% of the API values were influenced by ozone, O3 (secondary gas pollutant) and particulate matter with diameter of less than 10 micrometers, PM10 (non-gas air pollutants). The research verifies that environmetric techniques are highly viable and effective for analyzing large amounts of complex
data to glean vital knowledge about air quality, especially the behavior characteristics of specific air pollutants and air pollution patterns. This knowledge can be employed as decision tools for policy makers in planning for more effective air quality monitoring programs. |
|---|