Short-term prediction of atmospheric concentrations of ground-level ozone in Karaj using artificial neural network

Air pollution is a challenging issue in some of the large cities in developing countries. Air quality monitoring and interpretation of data are two important factors for air quality management in urban areas. Several methods exist to analyze air quality. Among them, we applied the dynamic neural network (TDNN) and Radial Basis Function (RBF) methods to predict the concentrations of groundlevel ozone in Karaj City in Iran. Input data included humidity, hour temperature, wind speed, wind direction, PM2.5, PM10 and benzene, which were monitored in 2014. The coefficient of determination between the observed and predicted data was 0.955 and 0.999 for the TDNN and RBF, respectively. The Index of Agreement (IA) between the observed and predicted data was 0.921 for TDNN and 0.9998 for RBF. Both methods determined reliable results. However, the RBF neural network performance had better results than the TDNN neural network. The sensitivity analysis related to the TDNN neural network indicated that the PM2.5 had the greatest and benzene had the minimum effect on prediction of groundlevel ozone concentration in comparison with other parameters in the study area.