This study developed a new semi-statistical model based on a Lagrangian approach. The overall effects on the observed pollutant levels at a receptor site were divided into two groups, one including the effects of emissions from various upwind sources and the other including all other effects (including the overall effects of atmospheric dilution. chemical transformation, and wet and dry depositions). The former effects were directly accounted for by a new parameter, an emission factor, defined as the accumulated emission uptake along the air trajectory toward the analyzed receptor site. All other effects were represented by a pollutant transfer coefficient. Meteorological parameters, excluding wind direction, were suggested to simulate this coefficient. The model was used to simulate variations in daily average NOx concentrations at a receptor site in south Taiwan during 1995-1999. Four meteorological factors, temperature, humidity, wind speed and pressure, were used to simulate the pollutant transfer coefficient. The full model successfully explained 61% of the analyzed concentration variations. The emission factor was the single most important factor in the model. When this factor was omitted, the determination coefficient of the model decreased from 61% to 48%. However, the pollutant transfer coefficient still dominated the analyzed variations of concentration.
Atmospheric environment,vol. 37, no.15, pp. 2051-2059