Multiple linear regression (MLR) models for PM2.5 concentration forecasting in Greater Klang Valley

Prolonged exposure to fine particulate matter (PM2.5), which is less than 2.5 micrometers in diameter, has been linked to significant health risks, including increased mortality from cardiopulmonary diseases and lung cancer. Developing effective local-level forecasting models is crucial, as these models allow authorities and the public to anticipate and respond to unhealthy air quality conditions. This study aims to develop Multiple Linear Regression (MLR) models to estimate PM2.5 concentrations diurnally across nine different locations in the Greater Klang Valley (GKV) and to examine the influence of meteorological factors, namely temperature and humidity. Model validation using ground-based data yielded R-squared values ranging from 0.3558 to 0.7188 and RMSE values from 0.093 to 1.7687, indicating that temperature and humidity significantly affect PM2.5 levels. Additionally, a correlation analysis between Landsat 8 satellite data and ground-based PM2.5 measurements revealed that Band 8 had the highest correlation coefficient of 0.574, with an R-squared value of 0.329 and a highly significant p-value of 0.005. Spatial distribution map of PM2.5. temperature and humidity have been generated and it showed that PM2.5 levels were lower in northern GKV from July to December 2022 but increased across central and southern regions in early 2023, especially between February and May. These patterns indicate seasonal variation and suggest that higher temperatures and lower humidity may influence pollutant concentration.