ESTIMATION OF WINTER SEASON SULPHUR DIOXIDE CONCENTRATIONS WITH AN ARTIFICIAL NEURAL NETWORK MODEL

Abstract

An understanding of pollution sources and emissions,and their interactions with terrain and the atmosphere is the most importantstep in developing appropriate air pollution management plans and actionstrategies. In this study, relationship between sulphur dioxide (SO2)concentration and meteorological parameters such as wind direction, wind speed,temperature, air pressure, precipitation, sunshine amount, sunshine duration,and relative humidity is modeled by using winter season data. Since therelation between SO2 concentrations and many meteorologicalparameters is complex, an artificial neural network (ANN) model is developed topredict the SO2 levels. Wind direction is modeled as the combinationof two variables, which enables to appropriately define the wind direction. TheANN model exhibited an R–squared value of 0.85