[Numerical Response Analysis of PM_2.5-O₃ Compound Pollution in Beijing]

The multi-scale variation trend of PM_2.5-O₃ compound pollution events was analyzed based on air quality data, meteorological data, and COVID-19 data in Beijing from 2015 to 2020. For the threshold of compound pollution, a compound pollution index was proposed, and the numerical response trend was evaluated based on the generalized additive model. A distributed lag nonlinear model was introduced to analyze the risk response relationship between compound pollution and influencing factors. The results showed that the events of PM_2.5-O₃ compound pollution in Beijing decreased annually. At the same time, due to the influence of pollutant emissions and meteorological conditions, there were obvious seasonal effects, week effects, holiday effects, and epidemic effects. The composite pollution index had no correlation with rainfall but had a linear positive correlation with O₃ and air temperature and a nonlinear correlation with other explanatory variables. Air pollutants and meteorological conditions had obvious lag effects on the composite pollution index, and the lag effects were mainly concentrated in 1-3 d. PM_2.5, PM₁₀, O₃, SO₂, and air temperature in high-value areas significantly increased the risk of compound pollution. The CO (1-6 mg·m^-3), NO₂ (38-118 μg·m^-3), and relative humidity (54%-87%) in the median section would also increase the risk of compound pollution, as would low wind speed. The compound pollution events showed a trend of multi-day continuous pollution in the numerical response. Compared with PM_2.5 and PM₁₀, compound pollution events were more dependent on O₃, and the compound pollution rate in high-value areas was 30.7%-47.5%. CO and relative humidity had little effect on compound pollution events. The air temperature had the greatest impact, and 84.7% of the composite pollution incidents occurred at 20-30℃.