engleski

Application of multivariate methods in an investigation of the effect of NO2, SO2, CO, PM10 and meteorological factors on ozone concentrations in an urban area

Presents an investigation of the importance of meteorological and air pollutants' variables in predicting ozone concentrations by using linear regression, principal component analysis, and principal component regression method. O3, NO2, CO, SO2 and PM10 concentrations determined in urban area in summer period are presented for the first time. The study focuses on the evaluation of the impact of temperature (T), relative humidity (RH), wind speed (WS), wind direction (WD), NO2, SO2, CO and PM10 concentrations on ozone variability. The principal component regression method showed that RH, T, WS, the wind vector component that explains air mass movement on the axis east to west, NO2, CO and SO2 were responsible for most variations in ozone concentrations (R2≈0.82). Any remaining variability could be attributed to other causes i.e parameters that were not monitored in this study. Results showed that the use of principal components as inputs improved multiple regression models prediction by reducing their complexity and eliminating data multicollinearity.

atmospheric pollutants; meteorological factors; principal component regression

nije evidentirano