Abstract: Following the relaxation of coronavirus disease 2019 restrictions and the subsequent full economic recovery, Shandong Province, China experienced a 4.3% rebound in the air quality index in 2023, with both fine particulate matter (PM_2.5) and ozone (O₃) concentrations exhibiting noticeable upward trends. Quantifying the drivers of this rebound is essential for developing targeted air quality management strategies. To this end, the analysis focused on the early spring period (ESP, February to April) and autumn harvest period (AHP, September to October) for PM_2.5 pollution and the photochemical season period (PSP, July to October) for O₃ pollution. We developed an interpretable random forest–Shapley additive explanation (RF–SHAP) framework optimized with a tree-structured Parzen estimator (TPE) to assess the impacts of anthropogenic emissions and meteorological factors. The introduction of the TPE optimization technique enhanced RF model performance across these pollution periods. Anthropogenic emissions played the dominant role in PM_2.5 pollution rebounds, contributing 14.1% during the ESP and 19.0% during the AHP, for example, industrial recovery (9.2% increase in energy consumption) and agricultural waste burning (70.0% increase in crop residue burning incident). In contrast, O₃ pollution was more strongly influenced by meteorological conditions, which contributed a 5.8% increase during the PSP. Critical meteorological drivers included strengthened atmospheric oxidation capacity, reduced total cloud cover, and changes in boundary layer height, although precursor emissions from the transportation and petrochemical industries remained indispensable for O₃ formation. This study provides an important scientific basis for precise air quality management in Shandong Province in the postpandemic period.

基于集成机器学习方法揭示2023年山东省三个时期PM_2.5和O₃污染反弹的驱动因素

[1]AssarehN, BeddowsA, StewartG, et al., 2025. What impact does net zero action on road transport and building heating have on exposure to UK air pollution? Environmental Science & Technology, 59(2):1274-1286.

[2]BelisCA, PernigottiD, PirovanoG, et al., 2020. Evaluation of receptor and chemical transport models for PM₁₀ source apportionment. Atmospheric Environment: X, 5:100053.

[3]ChenYT, ZhuLY, WangSH, et al., 2025. Unraveling the complexities of ozone and PM_2.5 pollution in the Pearl River Delta region: impacts of precursors emissions and meteorological factors, and effective mitigation strategies. Atmospheric Pollution Research, 16(2):102368.

[4]DaiSM, ChenXW, LiangJ, et al., 2023. Response of PM_2.5 pollution to meteorological and anthropogenic emissions changes during COVID-19 lockdown in Hunan Province based on WRF-Chem model. Environmental Pollution, 331:121886.

[5]DaiYQ, CaiXM, ZhongJ, et al., 2022. Chemistry, transport, emission, and shading effects on NO₂ and O_x distributions within urban canyons. Environmental Pollution, 315:120347.

[6]ElbirT, TuygunGT, GündoğduS, et al., 2025. Evaluating machine learning-based PM_2.5 estimation using integrated high-resolution datasets across NDVI levels in an urban-industrialized region. Environmental Pollution, 382:126734.

[7]EsuCO, PyoJ, ChoK, 2024. Machine learning-derived dose-response relationships considering interactions in mixtures: applications to the oxidative potential of particulate matter. Journal of Hazardous Materials, 475:134864.

[8]GrangeSK, CarslawDC, 2019. Using meteorological normalisation to detect interventions in air quality time series. Science of the Total Environment, 653:578-588.

[9]GrangeSK, CarslawDC, LewisAC, et al., 2018. Random forest meteorological normalisation models for Swiss PM₁₀ trend analysis. Atmospheric Chemistry and Physics, 18(9):6223-6239.

[10]HossainMM, RoyK, 2025. The development of classification-based machine-learning models for the toxicity assessment of chemicals associated with plastic packaging. Journal of Hazardous Materials, 484:136702.

[11]HouLL, DaiQL, SongCB, et al., 2022. Revealing drivers of haze pollution by explainable machine learning. Environmental Science & Technology Letters, 9(2):112-119.

[12]HouXS, WangXQ, ChengSY, et al., 2025. Evaluating urban carbon neutrality pathways and co-benefits through an integrated downscaling framework: a case study of the Beijing–Tianjin–Hebei region, China. Environmental Science & Technology, 59(49):26479-26491.

[13]JiaHY, YaoS, TangXD, et al., 2026. Multi-objective machine learning for health-oriented O₃ and PM_2.5 control: integrating VOC photochemical consumption and source apportionment. Journal of Hazardous Materials, 505:141483.

[14]JiangFF, MaJ, 2025. Graph-based machine learning for high-resolution assessment of pedestrian-weighted exposure to air pollution. Resources, Environment and Sustainability, 20:100219.

[15]KimHK, KimE, SongCK, et al., 2025. Comparative evaluation of surface-level PM_2.5 modeling across Northeast Asia: relevance for regulatory and scientific applications. Environmental Pollution, 377:126428.

[16]LeeBG, JeongKH, KimHE, et al., 2025. Machine learning models for predicting indoor airborne fungal concentrations in public facilities utilizing environmental variables. Environmental Pollution, 368:125684.

[17]LiJ, LiuXY, WangH, et al., 2024. Prediction and interpretation of photocatalytic NO removal on g-C₃N₄-based catalysts using machine learning. Chinese Chemical Letters, 35(2):108596.

[18]LiK, SunRH, 2024. Understanding the driving mechanisms of site contamination in China through a data-driven approach. Environmental Pollution, 342:123105.

[19]LiK, JacobDJ, LiaoH, et al., 2019. A two-pollutant strategy for improving ozone and particulate air quality in China. Nature Geoscience, 12(11):906-910.

[20]LiSY, WangSX, WuQR, et al., 2023. Emission trends of air pollutants and CO₂ in China from 2005 to 2021. Earth System Science Data, 15(6):2279-2294.

[21]LiuHW, YueFG, XieZQ, 2022. Quantify the role of anthropogenic emission and meteorology on air pollution using machine learning approach: a case study of PM_2.5 during the COVID-19 outbreak in Hubei Province, China. Environmental Pollution, 300:118932.

[22]LiuS, WangG, KongFH, et al., 2025a. Chemical composition, multiple sources, and health risks of PM_2.5: a case study in Linyi, China’s plate and logistics capital. Environmental Pollution, 365:125343.

[23]LiuS, WangG, ZhaoN, et al., 2025b. Ozone pollution in a typical region of southeastern Shandong Province, China: spatio-temporal distribution, sensitivity regimes, and sources. Journal of Environmental Chemical Engineering, 13(4):117389.

[24]LiuS, WangG, KongFH, et al., 2025c. PM_2.5 pollution characteristics, drivers, and regional transport during different pollution levels in Linyi, China: an integrated PMF-ML-SHAP framework and transport models. Journal of Hazardous Materials, 494:138534.

[25]LiuYM, WangT, 2020. Worsening urban ozone pollution in China from 2013 to 2017-Part 1: the complex and varying roles of meteorology. Atmospheric Chemistry and Physics, 20(11):6305-6321.

[26]LundbergSM, LeeSI, 2017. A unified approach to interpreting model predictions. Proceedings of the 31st International Conference on Neural Information Processing Systems, p.4768-4777.

[27]LundbergSM, ErionG, ChenH, et al., 2020. From local explanations to global understanding with explainable AI for trees. Nature Machine Intelligence, 2(1):56-67.

[28]MadukpeVN, UgoalaCB, NnadiN, et al., 2025. A topological approach in analyzing the shifts in air pollutants’ dynamics pre- and post-COVID-19 lockdown era. Environmental Monitoring and Assessment, 197(7):719.

[29]MangalathuS, HwangSH, JeonJS, 2020. Failure mode and effects analysis of RC members based on machine-learning-based Shapley additive explanations (SHAP) approach. Engineering Structures, 219:110927.

[30]MEE (Ministry of Ecology and Environment of the People’s Republic of China), 2018. Technical Specifications for Operation and Quality Control of Ambient Air Quality Continuous Automated Monitoring System for SO₂, NO2, O₃ and CO, HJ 818-2018. National Standards of the People’s Republic of China(in Chinese).

[31]MEE (Ministry of Ecology and Environment of the People’s Republic of China), 2024. Ministry of Ecology and Environment Releases National Ambient Air Quality Status for December 2023 and January–December 2023. MEE (in Chinese). https://www.mee.gov.cn/ywdt/xwfb/202401/t20240125_1064784.shtml

[32]MericoE, GrassoFM, CesariD, et al., 2020. Characterisation of atmospheric pollution near an industrial site with a biogas production and combustion plant in southern Italy. Science of the Total Environment, 717:137220.

[33]MOT (Ministry of Transport of the People’s Republic of China), 2024. 2023 National Port Cargo and Container Throughput. MOT (in Chinese). https://xxgk.mot.gov.cn/2020/jigou/zhghs/202404/t20240416_4128399.html

[34]NizamaniMM, ZhangHL, BolanN, et al., 2024. Understanding the drivers of PM_2.5 concentrations in Chinese cities: a comprehensive study of anthropogenic and environmental factors. Environmental Pollution, 361:124783.

[35]PendergrassDC, JacobDJ, OakYJ, et al., 2025. Wintertime trends of fine particulate matter (PM_2.5) in South Korea, 2012–2022: response of nitrate and organic components to decreasing NO_x emissions. Geophysical Research Letters, 52(19):e2025GL116091.

[36]ShenY, ZhangLP, FangX, et al., 2019. Spatiotemporal patterns of recent PM_2.5 concentrations over typical urban agglomerations in China. Science of the Total Environment, 655:13-26.

[37]ShiZB, SongCB, LiuBW, et al., 2021. Abrupt but smaller than expected changes in surface air quality attributable to COVID-19 lockdowns. Science Advances, 7(3):eabd6696.

[38]SongCB, LiuBW, ChengK, et al., 2023. Attribution of air quality benefits to clean winter heating policies in China: combining machine learning with causal inference. Environmental Science & Technology, 57(46):17707-17717.

[39]SongQ, HuangL, ZhangY, et al., 2025. Driving factors of PM_2.5 pollution rebound in north China plain in early 2023. Environmental Science & Technology Letters, 12(3):305-312.

[40]SPBS (Shandong Provincial Bureau of Statistics), 2024. 2024 Shandong Statistical Yearbook. China Statistical Publishing House, Beijing, China(in Chinese).

[41]SPDEE (Shandong Provincial Department of Ecology and Environment), 2024. 2023 Bulletin on the Ecological Environment of Shandong Province. SPDEE (in Chinese). http://xxgk.sdein.gov.cn/wryhjjgxxgk/zlkz/zkgb/202406/P020240606454202370719.pdf

[42]SPPG (Shandong Provincial People’s Government), 2024. Air Quality Status and Dust Fall Monitoring results for January to December 2023 in the Province. SPPG (in Chinese). http://www.shandong.gov.cn/art/2024/1/29/art_305267_10347461.html?xxgkhide=1

[43]WangG, QianYX, KongFH, et al., 2023. Non-methane hydrocarbon characteristics and their ozone and secondary organic aerosol formation potentials and sources in the plate and logistics capital of China. Atmospheric Pollution Research, 14(9):101873.

[44]WangG, YangXN, LiuS, et al., 2025. Unveiling the drivers of PM_2.5 pollution in an industrial inland city in China during heating seasons (2021–2024) by an integrated machine learning method. Journal of Hazardous Materials, 500:140495.

[45]WenZ, MaX, XuW, et al., 2024. Combined short-term and long-term emission controls improve air quality sustainably in China. Nature Communications, 15(1):5169.

[46]WuK, ZhuSP, Mac KinnonM, et al., 2023. Unexpected deterioration of O₃ pollution in the South Coast Air Basin of California: the role of meteorology and emissions. Environmental Pollution, 330:121728.

[47]XiaoQY, ZhengYX, GengGN, et al., 2021. Separating emission and meteorological contributions to long-term PM_2.5 trends over eastern China during 2000–2018. Atmospheric Chemistry and Physics, 21(12):9475-9496.

[48]XuQC, WangSX, JiangJK, et al., 2019. Nitrate dominates the chemical composition of PM_2.5 during haze event in Beijing, China. Science of the Total Environment, 689:1293-1303.

[49]YaoS, JiaHY, FanFJ, et al., 2025a. Assessing ozone formation impact through SHAP interaction redistribution analysis: a novel framework for evaluating VOC photochemical loss and source interactions. Environmental Impact Assessment Review, 115:108017.

[50]YaoS, FanFJ, JiaHY, et al., 2025b. Quantitative impacts of VOC sources on atmospheric oxidation capacity and O₃ formation from a megacity in China. Atmospheric Environment, 344:121033.

[51]Yuval, LeviY, BrodayDM, 2024. Revealing causality in the associations between meteorological variables and air pollutant concentrations. Environmental Pollution, 345:123526.

[52]ZhangHY, GuoWT, WangJG, et al., 2025. Scenario-dependent O₃ driver identification and formation responses to VOCs/NO_x emission reduction pathways from integrated and sectoral sources in the Beijing–Tianjin–Hebei region, China. Journal of Hazardous Materials, 496:139468.

[53]ZhangKR, ChenQL, HongYW, et al., 2025. Elucidating contributions of meteorology and emissions to O₃ variations in coastal city of China during 2019–2022: insights from VOCs sources. Environmental Pollution, 366:125491.

[54]ZhangQ, ZhengYX, TongD, et al., 2019. Drivers of improved PM_2.5 air quality in China from 2013 to 2017. Proceedings of the National Academy of Sciences of the United States of America, 116(49):24463-24469.

[55]ZhangW, XuQW, GaoLJ, et al., 2025. ITSC fault diagnosis for PMSM by using adaptive filtering and tree-structured Parzen estimator optimized-automated random forest. Electrical Engineering, 107(4):4711-4725.

[56]ZhangZL, WangSY, WangP, et al., 2024. Regional transport dominates air pollution events in all seasons in Beijing in 2020. Atmospheric Environment, 323:120395.

[57]ZhaoN, ZhangHY, WangG, 2025. Revealing the nonlinear responses of PM_2.5 and O₃ to VOC and NO_x emissions from various sources in Shandong, China. Journal of Hazardous Materials, 489:137655.

Similar articles