Statistical modeling of meteorological conditions and air pollution over Athens, Greece

Abstract

The importance of meteorological parameters and topography in determining the air pollution levels in a specific area is well established. The aim of this study is to model the association between air pollution levels and meteorological parameters for a single site in Athens, Greece. The statistical analysis is based on the Multiple Linear Regression (MLR) models for simulating the relationship amongst primary and secondary pollutants (CO, NO, NO2, O3 and SO2) and air temperature, wind speed, relative humidity and atmospheric boundary layer (ABL) depth. The meteorological variables are used as explanatory variables for training different statistical models for each pollutant. The analysis is performed for a twenty-year period (1990-2009) at 00Z and 12Z. Special emphasis is given to the most accurate representation of the ABL depth by using three different methods (i.e. Holzworth, virtual Richardson number and potential temperature gradient). The modeling results indicate the superior performance in the case where the ABL depth was calculated by the virtual Richardson number method. The results indicate the importance of meteorology in air quality along with the significance of other factors that increase air pollution variability in urban environments.