Multivariate methods for ground-level ozone modeling

作者：Ozbay Bilge^*; Keskin Gulsen Aydin; Dogruparmak Senay Cetin; Ayberk Savas

来源：Atmospheric Research, 2011, 102(1-2): 57-65.

DOI：10.1016/j.atmosres.2011.06.005

• Ozone
• Statistical analysis
• Multiple regression analysis
• Principle component analysis
• Kocaeli

The aim of this study is to apply multivariate statistical methods in predicting ozone (O-3) concentrations at the ground level of the troposphere as the function of pollution and meteorological parameters. PM10, SO2, NO, NO2, CO, O-3, CH4, NMHC, temperature, rainfall, humidity, pressure, wind direction, wind speed and solar radiation were measured hourly for one year period in order to predict O-3 concentrations of 1 h later. In the study, relationships between O-3 data and other variables were investigated by bivariate correlation analysis. CH4, NMHC, NO2 exhibited considerable negative correlations with O-3 described with the Pearson correlation coefficients of -0.67. -0.55, -0.51, respectively whereas highest positive correlation was noted for temperature with correlation coefficient of 0.60. Multiple regression analysis (MLR) was used for modeling annual and seasonal O-3 concentrations. Adjusted R-2 values were determined as 0.90, 0.85 and 0.92 respectively for annual period, cooling and warming seasons. In order to decrease the number of input variables principle component analysis (PCA) was applied by using annual data. MLR analysis was repeated using four principle components and new adjusted R-2 was calculated as 0.63.

• 出版日期2011-10