Analyzing the relationship between meteorological elements and criteria atmospheric pollutants in Tabriz using statistical modeling

Introduction: The rapid increase in population, growth of urbanization and industrialization in recent years, which is generally associated with an increase in demand and energy consumption, and as a result, an increase in pollutant emission sources, has exacerbated air pollution as one of the biggest current crises of urban societies and consequently health risks and related social inequalities in terms of time and space. On the other hand, meteorological parameters directly affect the amount of pollutants as well as the duration of their presence in the atmosphere, and the present research was conducted in order to investigate this effect and discover the relationships between criteria air pollutants and atmospheric elements.
Materials and Methods: Besides investigating the status of meteorological elements (temperature, precipitation, wind speed, relative humidity, radiation, sunshine hours and cloudiness) and criteria air pollutants (carbon monoxide (CO), nitrogen dioxide (NO2), sulfur dioxide (SO2), ozone (O3) and particulate matters with aerodynamic diameters less than 10 microns and 2.5 microns (PM10 and PM2.5)) in Tabriz city during 2004-2021, the present study has explored the relationships between pollutants and meteorological parameters in monthly and seasonal time scales using Pearson's correlation test at the 95% confidence level and the effect of these elements on the concentration of pollutants using Multiple Linear Regression (MLR) and Generalized Additive Model (GAM) in R 4.3.1 statistical software.
Results and Discussion: Based on the results of Pearson correlation analysis, NO2 and PM2.5, SO2 and PM2.5 pollutants and PM2.5 and PM10 pollutants have shown significant positive correlation in pairs, so it seems that these pollutants have similar emission sources. Also, the results of this research demonstrate that the concentration of criteria air pollutants in Tabriz was affected by weather conditions during the entire statistical period in the monthly and seasonal time scales and NO2 and PM2.5 pollutants have the most negative monthly correlation with the parameters of temperature, wind speed and sunshine hours and the most positive correlation with relative humidity; PM2.5 has the most positive correlation with pressure; CO and SO2 have the most negative correlation with radiation; O3 has a strong positive correlation with temperature, wind speed and sunny hours and the most negative correlation with pressure, relative humidity and cloudiness; and NO2 and PM10 pollutants have the most positive correlation with cloudiness. The results of fitting Multiple Linear Regression (MLR) and Generalized Additive Model (GAM) for each criteria air pollutant in Tabriz city indicate the better performance of GAM in analyzing the relationships between all air pollutants and the set of independent variables except NO2.
Conclusion: The results of this research indicate that the effect of atmospheric elements on the concentration of criteria pollutants in Tabriz city is different depending on the type of pollutant and at different times, and it can be acknowledged that the effect of a specific meteorological parameter on air pollution is uncertain. However, wind speed, radiation, temperature and air pressure are the most important meteorological elements related to the concentration of pollutants in Tabriz city. Also, the consequences of the research suggest that both MLR and GAM can describe the variability of the response variable by a set of predictor variables and explain the linear and non-linear relationships between them. However, considering the non-linear relationship between the concentration of atmospheric pollutants and meteorological elements, GAM is able to justify a higher percentage of changes in all criteria atmospheric pollutants except NO2.