Meteorological impacts on surface ozone: A case study based on Kolmogorov–Zurbenko filtering and multiple linear regression

Abstract

Ozone variation, excluding meteorological effects, is very important to assess the effects of air pollution control policies. In this study, the Kolmogorov-Zurbenko (KZ) filter method and multiple linear stepwise regression are combined to study the impact of meteorological parameters on ozone concentration over the past 5 years (2016–2020) in a petrochemical industrial city in northern China. Monte Carlo simulations were used to evaluate the reliability for the potential quasi quantitative prediction of the baseline component. The average level of the city and the details of five stations in the city were studied. The results show that the short-term, seasonal, and long-term component variances of maximum daily running average 8 h (MDA8) ozone in Zibo city (City) decomposed by the KZ filter account for 32.06%, 61.67% and 1.15% of the total variance, for a specific station, the values were 32.37%–34.90%, 56.64%–62.00%, and .35%–3.14%, respectively. The average long-term component increase rate is 3.19 μg m−3 yr−1 on average for the city, while it is 1.52–5.95 μg m−3 yr−1 for a specific station. The overall meteorological impact was not stable and fluctuated between −2.60 μg m−3 and +3.77 μg m−3. This difference in trends between the city and specific stations implied that the O3 precursor’s mitigation strategy should be more precise to improve its practical effects.