Forecasting Ozone Concentrations Using Box-Jenkins ARIMA Modeling in Malaysia

Abstract

Time series analysis and forecasting has become a major tool in many applications in air pollution and environmental management fields. Among the most effective approaches for analyzing time series data is the model introduced by Box and Jenkins. In this study, we used Box-Jenkins methodology to build Autoregressive Integrated Moving Average (ARIMA) model on the average of monthly ozone data taken from three monitoring stations in Klang Valley for the period 2000 to 2010 with a total of 132 readings. Result shows that ARIMA (1,0,0)(0,1,1)12 model was successfully applied to predict the long term trend of ozone concentrations in Klang Valley. The model performance has been evaluated on the basis of certain commonly used statistical measures. The overall model performance is found to be quite satisfactory as indicated by the values of Root Mean Squared Error, Mean Absolute Percentage Error and Normalized Bayesian Information Criteria. The finding of a statistically significant upward trend of future ozone concentrations is a concern for human health in Klang Valley since over the last decade, ozone appears as one of the main pollutant of concern in Malaysia.