Trends and prediction of air pollutants in Pasir Gudang Industrial Area, Johor, Malaysia

Abstract

The trends and prediction of the air quality of Pasir Gudang industrial area in Johor are discussed and presented in this thesis. An attempt was also made to study the pollutants concentrations recorded by the Larkin monitoring station. However, studies on the trends, meteorological influences and the predictions of atmospheric pollution were given a greater emphasis for the Pasir Gudang industrial area. The statistical analysis based on a simple correlation coefficient and regression analysis showed that although there is a relationship between each pollutant i.e ozone (O3), particulate Matter with diameter of 10 micrometers or less (PM10), nitrogen dioxide (NO2), sulfur dioxide (SO2) and carbon monoxide (CO) concentrations and a combination of meteorological parameters such as wind speed, temperature, humidity rate and solar radiation in Pasir Gudang with the correlation coefficient (r) of 0.64, 0.42, 0.71, 0.55 and 0.49, respectively, the inclusion of the previous day’s pollutants concentrations significantly presents better prediction models with the correlation coefficient of 0.73, 0.68, 0.83, 0.68 and 0.67, respectively. Subsequently, the prediction of PM10 based on its previous day’s concentrations through artificial neural network resulted in a much better model prediction with the value of r=0.69 and 0.70 compared to the statistical model with the value of r=0.64. The spatial variation of SO2, NO2 and PM10 emitted from various industrial sources in Pasir Gudang were also predicted using American Meteorological Society/Environmental Protection Agency Regulatory Model (AERMOD) air dispersion model. The Weather Research and Forecasting (WRF) model was applied to simulate the required meteorological variables for the selected date i.e 2-16 July, 2010. The WRF output values i.e. temperature, wind speed and wind direction were compared with the onsite measured data in Pasir Gudang, Senai, KLIA and Kluang stations. The results showed the accuracy of WRF model performance in simulating temperature and wind speed with the Mean Absolute Error (MAE) and Root Mean Square Error (RMSE) value of less than 2.8 and 3.5, respectively, while it has some difficulties in simulating the wind direction near a coastal area. The maximum ground level concentration of pollutants i.e SO2, NO2 and PM10 simulated through AERMOD coupled with WRF in Pasir Gudang industrial area was 36.2, 59.8 and 5.4 ug/m3, respectively, which were within the Malaysia ambient air quality guidelines over the receptor grid. The evaluation of AERMOD through the quantile-quantile (Q-Q) plots showed that most of the predicted and observed pair points are lying close to the one-to-one line. Besides, the sensitivity of AERMOD model to its input parameters i.e stack characteristics and meteorological variables showed that the model is more sensitive to stack gas temperature and stack height as well as wind speed.