Changes in NO 3 Radical and Its Nocturnal Chemistry in Shanghai From 2014 to 2021 Revealed by Long‐Term Observation and a Stacking Model: Impact of China's Clean Air Action Plan

Nitrate radical (NO 3 ), an important nocturnal tropospheric oxidant, was observed continuously in the eastern Chinese megalopolis of Shanghai from July 2018 to August 2021using differential optical absorption spectroscopy equipped with LED lamp. To assess changes in NO 3 radical during the implementation of China's Clean Air Action Plan, a stacking model was developed to predict historical NO 3 radical from 2014 to 2018 with good performance. Combining the observed and predicted data sets, the highest concentration of NO 3 exceeded 300 pptv in 2021, while the highest monthly mean of 36.2 ± 30.9 pptv occurred in summer 2017. The impact of meteorology, precursors, and emissions on the evolution of nocturnal NO 3 was quantified using the Kolmogorov‐Zurbenko filter. The rapid increase in annual NO 3 from 2014 to 2017 was caused by a combination of persistently high precursor O 3 concentrations, insufficient SO 2 ‐represented primary pollutant reductions, and meteorological conditions favorable to NO 3 . As China's clean air actions advanced further, significant CO‐represented and SO 2 ‐represented emission reductions overcame the impact of meteorological conditions and significantly lowered the NO 3 levels between 2018 and 2021. These efforts drastically affected nocturnal NO 3 chemistry, reducing NO 3 loss rates by approximately 30% in both spring and summer in 2021 compared to 2014, although the NO 3 levels were comparable between the 2 years. In addition, substantial reductions in CO‐represented and SO 2 ‐represented primary pollutants, as well as effective management of NO 2 and PM 2.5 , resulted in varying decreases in the rates of direct (removal by reactions with VOCs, etc.) and indirect (removal by heterogeneous consumption of N 2 O 5 ) losses, thereby moderating secondary aerosol formation. Nitrate radical (NO 3 ) is an important nocturnal tropospheric oxidant and impacts the formation of secondary pollutants. To assess changes in nocturnal atmospheric chemistry due to the implementation of the clean air actions of Chinese government starting in 2013, a machine learning model was developed to predict NO 3 radical from 2014 to 2018 based on observed data between 2018 and 2021. The long‐term NO 3 radical series shows an overall trend with an increase from 2014 to 2017, a decrease from 2017 to 2019, and an increase from 2019 to 2021. A statistical analysis approach indicates that the significant reduction in primary pollutants overcame the impact of met