Regional Spatiotemporal Variability of Multi-Pollutant Ambient Air Quality Across Major Regions of Malaysia
DOI:
https://doi.org/10.11113/jagst.v6n2.131Keywords:
Air quality, K-means clustering, Spatial–temporal variability, Criteria pollutantsAbstract
Rapid urbanisation and intensified anthropogenic activities have contributed to increasing spatial–temporal variability in ambient air pollutant concentrations across Malaysia. This study applied pollutant-specific K-means clustering to compare the spatial and temporal concentration patterns of PM₁₀, PM₂.₅, CO, and NO₂ across Malaysia. Daily- and monthly-averaged data from the Department of Environment (DOE) Continuous Air Quality Monitoring (CAQM) network from 2018 to 2021 were analysed following data preprocessing procedures including mean imputation, Min–Max normalization, and Z-score outlier removal. Descriptive statistical analysis revealed significant regional disparities, with the Centre region recording the highest mean concentrations of CO (0.76 ppm), PM10 (30.02 µg/m³), and PM2.5 (21.73 µg/m³), while the lowest levels were observed in the Borneo region. Following data verification and unit harmonisation, nitrogen dioxide (NO₂) showed regional variability, with the Centre region recording the highest mean concentration (0.0126 ppm). Clustering analysis identified distinct pollution regimes, with high-concentration clusters predominantly associated with urbanised zones. Daily PM₁₀ concentrations in the Centre region exceeded the WHO 24-hour guideline during extreme cluster events, reaching values above 100 µg/m³. Monthly clustering further demonstrated seasonal influences, with elevated pollutant concentrations during the Southwest Monsoon associated with regional biomass burning and transboundary haze episodes. These findings indicate that severe pollution events in Malaysia are episodic rather than persistent, yet capable of generating substantial short-term exposure risks in densely populated metropolitan regions. The integration of unsupervised machine learning enables improved identification of hidden pollution regimes and supports the development of region-specific air quality management strategies aligned with Sustainable Development Goal 11.
