Trend analysis and change point detection of air pollution index in Malaysia

Abstract

Parametric methods are commonly used to conduct the trend analysis of air pollution. These methods require certain statistical assumptions, such as stationarity and normality of the data. However, such assumptions are usually not applicable to trends in Air pollution index (API). In addition, the change points in the time series have not been taken into consideration by most of the analysis of API. Therefore, this study presents a comprehensive investigation of the trend analysis and change point detection of the mean and maximum of API series in Malaysia. The hourly, daily, weekly, monthly, seasonal, and annual API data series were considered in the analysis. The fner time intervals were used to detect any signifcant increasing or decreasing trends of the API series for Malaysia. The API data were collected from 37 air monitoring stations in Peninsular Malaysia. The nonparametric tests, including Mann–Kendall test, Pettitt test, and innovative trend analysis were used to examine the contribution presented herein. Various aspects of API data were studied, including upward trends, downward trends, and change points. Several signifcant monotonic trends and changing points in some of the API measuring stations were found from the Mann–Kendall test results. Signifcant increasing trends of the monthly and seasonal mean, as well as maximum API, were found in the years 2013 and 2014 for some stations. In addition, the magnitudes of the increasing trends in maximum API are larger than the mean API. The detection points captured by the Pettitt test are possibly related to the El-Nino events. In general, the results of the study provide comprehensive information on air quality trends and their atmospheric aspects, which can help in developing strategies to address air quality problems and provide meaningful opportunities to mitigate air pollution problems in Peninsular Malaysia.