Air pollution estimation using traffic volume data and primary weather data: case study Mashhad

Rapid urbanization and population growth has resulted in increased traffic congestion and consequently air pollution in most major cities, in particular, in the developing countries. Knowledge on the amount of different air pollutants and their spatial and temporal concentrations is of great importance for decision makers on health, environment and air quality estimation in different scales. Mashhad, as a metropolitan, due to its specific religious, socio-cultural and geographical role in the region is declared as one of the most polluted cities of the country.
Given that there is a direct relationship between traffic volume data and air pollutants (PM2.5, CO and ), this study attempts to estimate the amount of each pollutant based on traffic volume and some primary weather data. We used empirical models proposed in the literature, such as Baker model and AERMOD, as well as linear regression and nonlinear neural network methods to explore the correlation between traffic volume and air pollutants over a period of six months in the city of Mashhad.
The results showed low correlation coefficients between traffic volume and air pollutants in all models, indicating that such models may not be suitable to further estimate air pollutants using only traffic volume and primary weather data. Correlation coefficients were lowest for the pollutant PM2.5 over the time period of the study. Sensitivity analysis demonstrated that vehicle average velocity is by far the most influential variable in the empirical models used.