Chiang Mai Journal of Science

 Air pollution is one of the major environmental problems in Malaysia caused by rapid development and industrialization. For instance, industrial processing and transportation may lead to the production of ground level ozone (O3) precursors, namely oxides of nitrogen (NOX) and volatile organic compounds (VOC). O3 forms when NOX and VOC react in the presence of sunlight. It is estimated that 0.47 million deaths occur worldwide due to high O3 concentration. Hence, ground level ozone (O3) forecasting is of great importance and plays a major role in air pollution preparedness. This study used the vector autoregressive (VAR) model for forecasting O3 concentration in Malaysia. Four locations were selected to represent industrial and background monitoring stations. The particulate matter (PM10), gaseous pollutants and meteorological parameters were considered in the development of VAR models. The Granger Causality test was found that the parameters influence O3 concentration model is PM10 and CO in Pasir Gudang, PM10 and NO2 in Nilai, RH and CO in Jerantut while in Perai only PM10. Akaike Information Criterion (AIC) suggested that VAR1 is the most appropriate model for forecasting O3 concentrations at all these monitoring stations. The accuracy of these models was validated by using three performance errors, which is mean absolute error (MAE), mean absolute percentage error (MAPE) and root mean square error (RMSE). In summary, the appropriate VAR models can be used to forecast O3 concentration and help relevant agencies to develop policies related to O3 pollution.