A causal analysis of ground-level ozone, meteorological factors, and other air pollutants: an in-depth AI-based study applied to the climate of Craiova City, Romania

Abstract

Abstract Ground-level ozone is a major air pollutant whose concentrations are influenced by meteorological conditions and other air pollutants. Understanding the causal relationships among ozone, meteorological variables, and other air pollutants is important for two reasons: (1) accurate prediction and forecasting and (2) effective air quality management. This study employs a combination of Granger causality testing, cross-correlation analysis, multiple linear regression, deep NARMAX modelling, and structural equation modelling to investigate the causal and time-lagged effects of meteorological and anthropogenic factors on ozone formation and depletion. A five-year dataset of hourly measurements of ozone, other air pollutants, and meteorological parameters is analysed for Craiova, Romania, to identify the dominant drivers of ozone variability. The results provide a basis for developing improved predictive models and offer insights into the delayed effects of air pollutants and weather conditions on urban ground-level ozone concentrations, supporting informed strategies for air quality management.