Prediction of tropospheric ozone concentration using artificial neural networks at traffic and background urban locations in Novi Sad, Serbia

In this paper, we described generation and performances of feedforward neural network models that could be used for a day ahead predictions of the daily maximum 1-h ozone concentration (1hO₃) and 8-h average ozone concentration (8hO₃) at one traffic and one background station in the urban area of Novi Sad, Serbia. The six meteorological variables for the day preceding the forecast and forecast day, ozone concentrations in the day preceding the forecast, the number of the day of the year, and the number of the weekday for which ozone prediction was performed were utilized as inputs. The three-layer perceptron neural network models with the best performance were chosen by testing with different numbers of neurons in the hidden layer and different activation functions. The mean bias error, mean absolute error, root mean squared error, correlation coefficient, and index of agreement or Willmott's Index for the validation data for 1hO₃ forecasting were 0.005 μg m^-3, 12.149 μg m^-3, 15.926 μg m^-3, 0.988, and 0.950, respectively, for the traffic station (Dnevnik), and - 0.565 μg m^-3, 10.101 μg m^-3, 12.962 μg m^-3, 0.911, and 0.953, respectively, for the background station (Liman). For 8hO₃ forecasting, statistical indicators were - 1.126 μg m^-3, 10.614 μg m^-3, 12.962 μg m^-3, 0.910, and 0.948 respectively for the station Dnevnik and - 0.001 μg m^-3, 8.574 μg m^-3, 10.741 μg m^-3, 0.936, and 0.966, respectively, for the station Liman. According to the Kolmogorov-Smirnov test, there is no significant difference between measured and predicted data. Models showed a good performance in forecasting days with the high values over a certain threshold.