Prediction of ground level ozone exceedences and return period using distribution funtion

Abstract

Ozone (03) is considered as one of the crucial air pollutant in atmosphere which affects human health, vegetation and forests. According to DoE, the annual average daily maximum one-hour ozone concentrations had slightly increased in 2014 compared to 2013. Thence, it is crucial to come out with a model that is suitable to predict ground level ozone concentration in order to prevent adverse air pollution effects to be aggravated. Therefore, the aim of this study is to predict ground level ozone exceedences and return period at Cheras, Kuala Lumpur and Tanjung Malim, Perak using distribution function. Data analysis in this study was done by MATLAB. The selected distribution functions to fit 03 concentration data in 2012 are Gamma, Inverse Gaussian/ Wald and Rayleigh. Next, parameter for each distribution was estimated by Maximum Likelihood Estimator (MLE) method. The best fit distribution was determined by the selected performance indicators based on the highest accuracy measures which are close to 1 and the smallest error measures which are close to 0. Results showed that the best distribution that fits the 03 observations was found to be Gamma distribution for Cheras and Tanjung Malim. The probabilities of exceedences were calculated and predicted the return period by the cumulative density function (cdf) obtained from the best-fit distribution. For Cheras, it was predicted to exceed O.lppm for 6.7 days with a return period of once per 55 days. Tanjung Malim was predicted to exceed O.lppm for 3.6 days with a return period of once per 101 days. In Malaysia, modelling using distribution function is still a new approach and it is believe that it will be a good alternative for ground level ozone prediction.