Use of Time Series Analysis to predict the Ozone Concentration in Central Taiwan

• Main Authors: Chang, Yau-Hwa, 張耀華
• Other Authors: Cheng Wan-Li
• Format: Others
• Language: zh-TW
• Published: 1996
• Online Access: http://ndltd.ncl.edu.tw/handle/50048793991566524499

碩士 === 東海大學 === 環境科學系 === 84 === The purpose of this study is to analyze the ozone concentration at six air pollution monitoring stations set by Environment Protection Agency in Taichung area. Time series analysis was used to evalute the hourly ozone concentration. Three time series models, Univariate-ARIMA, ARIMA-Transfer Function and ARIMA-Intervention Model were used to predict the ozone concentration data. Ththat, in predicting ozone concentration data, there were similar results between ARIMA and ARIMA-Trans Function models. Adding NO2 as an explanatory variable wo''nt make the prediction performance better because of the uncertainty of NO2 concentration prediction and the effect of physical, chemical and meteorological effects. ARIMA- Intervntion Model can be used to correct ozone concentration data when special weather type occurs. It has a better prediction result on the following day of special weather type. But after the day, the performance of ARIMA-Intervention Model and Univariate-ARIMA are about the same.The result of three Time Series Models do not have significant difference.It take the least effort and time to establish Univariate-ARIMA Model. Therefore, Univariate-ARIMA Model should be utilized for predicting short term ozone concentration.