Data-driven machine learning quantifies ozone transport in the Hangzhou Bay urban cluster

Abstract

Severe ozone (O₃) pollution has always been a serious problem faced by areas with rapid economic development, and the regional O₃ transport between cities is a major cause of this problem. Therefore, we used a bidirectional long short-term memory (Bi-LSTM) model to quantitatively identify the regional O₃ transport in Hangzhou Bay, China. Combined with the meteorological removal method, we were able to model O₃ concentrations that were not affected by transport. The contribution of regional transport to Shanghai’s O₃ was quantified and validated using two different simulation schemes, which yielded highly consistent results of 18.41 μg/m³ (24% contribution) and 20.52 μg/m³ (27% contribution). According to the model simulation results, we found that approximately 24% of the O₃ pollution in Shanghai originates from other cities in the summer when the O₃ pollution is high. In addition, the regional O₃ transport was mainly concentrated during the high-value weather of O₃ pollution in Shanghai, and transport on non-pollution days was not apparent. Therefore, the regional O₃ transport from other cities is an important source of O₃ pollution in Shanghai. Overall, our study demonstrates the potential of machine-learning models coupled with meteorological removal for quantifying the inter-city influence of atmospheric pollutants.

Graphical abstract

Keywords

Ozone transport / Ozone pollution / Machine learning / Meteorological removal / Bi-LSTM

Highlight

	● The Bi-LSTM model can identify and quantify the regional ozone transport.
	● The meteorological removal method can identify regional ozone transport drivers.
	● Ozone pollution characteristics are related to economics and geography.
	● One-quarter of Shanghai’s June ozone is transported from other cities.
	● Regional ozone transport is evident on heavily polluted days.

References

[1] Blanchard C L , Tanenbaum S , Hidy G M . (2014). Spatial and temporal variability of air pollution in Birmingham, Alabama. Atmospheric Environment, 89: 382–391

[2] Cheng L J , Wang S , Gong Z Y , Li H , Yang Q , Wang Y Y . (2018). Regionalization based on spatial and seasonal variation in ground-level ozone concentrations across China. Journal of Environmental Sciences, 67: 179–190

[3] Cohen A J , Brauer M , Burnett R , Anderson H R , Frostad J , Estep K , Balakrishnan K , Brunekreef B , Dandona L , Dandona R . . (2017). Estimates and 25-year trends of the global burden of disease attributable to ambient air pollution: an analysis of data from the Global Burden of Diseases Study 2015. The Lancet, 389(10082): 1907–1918

[4] Fang X , Zou B , Liu X P , Sternberg T , Zhai L . (2016). Satellite-based ground PM_2.5 estimation using timely structure adaptive modeling. Remote Sensing of Environment, 186: 152–163

[5] Feng R , Zheng H J , Gao H , Zhang A R , Huang C , Zhang J X , Luo K , Fan J R . (2019). Recurrent Neural Network and random forest for analysis and accurate forecast of atmospheric pollutants: a case study in Hangzhou, China. Journal of Cleaner Production, 231: 1005–1015

[6] Foret G , Eremenko M , Cuesta J , Sellitto P , Barré J , Gaubert B , Coman A , Dufour G , Liu X , Joly M . . (2014). Ozone pollution: what can we see from space? A case study. Journal of Geophysical Research: Atmospheres, 119(13): 8476–8499

[7] Gao J H , Zhu B , Xiao H , Kang H Q , Hou X W , Shao P . (2016). A case study of surface ozone source apportionment during a high concentration episode, under frequent shifting wind conditions over the Yangtze River Delta, China. Science of the Total Environment, 544: 853–863

[8] GaoWTieX XXuJ MHuangR JMaoX QZhouG QChangL Y (2017). Long-term trend of O₃ in a Mega City (Shanghai), China: characteristics, causes, and interactions with precursors. Science of the Total Environment, 603–604: 603–604

[9] Grange S K , Carslaw D C , Lewis A C , Boleti E , Hueglin C . (2018). Random forest meteorological normalisation models for Swiss PM₁₀ trend analysis. Atmospheric Chemistry and Physics, 18(9): 6223–6239

[10] Guan Y , Xiao Y , Wang Y M , Zhang N N , Chu C J . (2021). Assessing the health impacts attributable to PM_2.5 and ozone pollution in 338 Chinese cities from 2015 to 2020. Environmental Pollution, 287: 117623

[11] (2021). . , 115: 26–34

[12] Hou L L , Dai Q L , Song C B , Liu B W , Guo F Z , Dai T J , Li L X , Liu B S , Bi X H , Zhang Y F , Feng Y C . (2022). Revealing drivers of haze pollution by explainable machine learning. Environmental Science & Technology Letters, 9(2): 112–119

[13] Han Y, Lam J C K, Li V O K, Reiner D (2021). A Bayesian LSTM model to evaluate the effects of air pollution control regulations in Beijing, China. Environmental Science & Policy, 115: 26–34

[14] Hu J , Li Y C , Zhao T L , Liu J , Hu X M , Liu D Y , Jiang Y C , Xu J M , Chang L Y . (2018). An important mechanism of regional O₃ transport for summer smog over the Yangtze River Delta in eastern China. Atmospheric Chemistry and Physics, 18(22): 16239–16251

[15] Jiang X Y , Wiedinmyer C , Carlton A G . (2012). Aerosols from fires: an examination of the effects on ozone photochemistry in the western United States. Environmental Science & Technology, 46(21): 11878–11886

[16] Kang Q , Song X , Xin X Y , Chen B , Chen Y Z , Ye X D , Zhang B Y . (2021). Machine learning-aided causal inference framework for environmental data analysis: a COVID-19 case study. Environmental Science & Technology, 55(19): 13400–13410

[17] Kim H S , Park I , Song C H , Lee K , Yun J W , Kim H K , Jeon M , Lee J , Han K M . (2019). Development of a daily PM₁₀ and PM_2.5 prediction system using a deep long short-term memory neural network model. Atmospheric Chemistry and Physics, 19(20): 12935–12951

[18] Li K , Jacob D J , Liao H , Shen L , Zhang Q , Bates K H . (2018a). Anthropogenic drivers of 2013–2017 trends in summer surface ozone in China. Proceedings of the National Academy of Sciences of the United States of America, 116(2): 422–427

[19] Li S , Wang T J , Huang X , Pu X , Li M M , Chen P L , Yang X Q , Wang M H . (2018b). Impact of East Asian summer monsoon on surface ozone pattern in China. Journal of Geophysical Research: Atmospheres, 123(2): 1401–1411

[20] Liu C , Zhang H R , Cheng Z , Shen J Y , Zhao J H , Wang Y C , Wang S , Cheng Y . (2021a). Emulation of an atmospheric gas-phase chemistry solver through deep learning: case study of Chinese Mainland. Atmospheric Pollution Research, 12(6): 101079

[21] Liu X , Lu D W , Zhang A Q , Liu Q , Jiang G B . (2022). Data-driven machine learning in environmental pollution: gains and problems. Environmental Science & Technology, 56(4): 2124–2133

[22] Liu X F , Guo H , Zeng L W , Lyu X P , Wang Y , Zeren Y Z , Yang J , Zhang L Y , Zhao S Z , Li J , Zhang G . (2021b). Photochemical ozone pollution in five Chinese megacities in summer 2018. Science of the Total Environment, 801: 149603

[23] Liu X L , Huang W M , Gill E W . (2017). Wind direction estimation from rain-contaminated marine radar data using the ensemble empirical mode decomposition method. IEEE Transactions on Geoscience and Remote Sensing, 55(3): 1833–1841

[24] Lu X , Zhang L , Chen Y F , Zhou M , Zheng B , Li K , Liu Y M , Lin J T , Fu T M , Zhang Q . (2019). Exploring 2016–2017 surface ozone pollution over China: source contributions and meteorological influences. Atmospheric Chemistry and Physics, 19(12): 8339–8361

[25] Pak U , Ma J , Ryu U , Ryom K , Juhyok U , Pak K , Pak C . (2020). Deep learning-based PM_2.5 prediction considering the spatiotemporal correlations: a case study of Beijing, China. Science of the Total Environment, 699: 133561

[26] Qi L , Tian Z G , Jiang N , Zheng F Y , Zhao Y C , Geng Y S , Duan X L . (2023). Collaborative control of fine particles and ozone required in China for health benefit. Frontiers of Environmental Science & Engineering, 17(8): 92

[27] Qu Z W , Li H T , Li Z H , Zhong T . (2022). Short-term traffic flow forecasting method with M-B-LSTM hybrid network. IEEE Transactions on Intelligent Transportation Systems, 23(1): 225–235

[28] Reichstein M , Camps-Valls G , Stevens B , Jung M , Denzler J , Carvalhais N . (2019). Deep learning and process understanding for data-driven Earth system science. Nature, 566(7743): 195–204

[29] Roberts D R , Bahn V , Ciuti S , Boyce M S , Elith J , Guillera-Arroita G , Hauenstein S , Lahoz-Monfort J J , Schröeder B , Thuiller W . . (2017). Cross-validation strategies for data with temporal, spatial, hierarchical, or phylogenetic structure. Ecography, 40(8): 913–929

[30] Siami-NaminiSTavakoliNNaminA S (2019). The performance of LSTM and BiLSTM in forecasting time series. In: Proceedings of 2019 IEEE International Conference on Big Data (Big Data). Los Angeles: IEEE, 3285–3292

[31] Sicard P , Agathokleous E , De Marco A , Paoletti E , Calatayud V . (2021). Urban population exposure to air pollution in Europe over the last decades. Environmental Sciences Europe, 33(1): 28

[32] Song C B , Wu L , Xie Y C , He J J , Chen X , Wang T , Lin Y C , Jin T S , Wang A X , Liu Y . . (2017). Air pollution in China: status and spatiotemporal variations. Environmental Pollution, 227: 334–347

[33] Suciu L G , Griffin R J , Masiello C A . (2017). Regional background O₃ and NO_x in the Houston-Galveston-Brazoria (TX) region: a decadal-scale perspective. Atmospheric Chemistry and Physics, 17(11): 6565–6581

[34] Tao Q , Liu F , Li Y , Sidorov D . (2019). Air pollution forecasting using a deep learning model based on 1D convnets and bidirectional GRU. IEEE Access, 7: 76690–76698

[35] Targino A C , Harrison R M , Krecl P , Glantz P , de Lima C H , Beddows D . (2019). Surface ozone climatology of South Eastern Brazil and the impact of biomass burning events. Journal of Environmental Management, 252: 109645

[36] Tong X N , You L H , Zhang J J , He Y L , Gin K Y H . (2022). Advancing prediction of emerging contaminants in a tropical reservoir with general water quality indicators based on a hybrid process and data-driven approach. Journal of Hazardous Materials, 430: 128492

[37] Vu T V , Shi Z B , Cheng J , Zhang Q , He K B , Wang S X , Harrison R M . (2019). Assessing the impact of clean air action on air quality trends in Beijing using a machine learning technique. Atmospheric Chemistry and Physics, 19(17): 11303–11314

[38] Wang B , Yuan Q Q , Yang Q Q , Zhu L Y , Li T W , Zhang L P . (2021). Estimate hourly PM_2.5 concentrations from Himawari-8 TOA reflectance directly using geo-intelligent long short-term memory network. Environmental Pollution, 271: 116327

[39] Wang T , Xue L K , Brimblecombe P , Lam Y F , Li L , Zhang L . (2017). Ozone pollution in China: a review of concentrations, meteorological influences, chemical precursors, and effects. Science of the Total Environment, 575: 1582–1596

[40] Wang T Y , Zhao B , Liou K N , Gu Y , Jiang Z , Song K , Su H , Jerrett M , Zhu Y F . (2019). Mortality burdens in California due to air pollution attributable to local and nonlocal emissions. Environment International, 133: 105232

[41] Watson G L , Telesca D , Reid C E , Pfister G G , Jerrett M . (2019). Machine learning models accurately predict ozone exposure during wildfire events. Environmental Pollution, 254: 112792

[42] Wei J , Liu S , Li Z Q , Liu C , Qin K , Liu X , Pinker R T , Dickerson R R , Lin J T , Boersma K F . . (2022). Ground-Level NO₂ surveillance from space across China for high resolution using interpretable spatiotemporally weighted artificial intelligence. Environmental Science & Technology, 56(14): 9988–9998

[43] Xue J , Wang F T , Zhang K , Zhai H H , Jin D , Duan Y S , Yaluk E , Wang Y J , Huang L , Li Y W . . (2023). Elucidate long-term changes of ozone in Shanghai based on an integrated machine learning method. Frontiers of Environmental Science & Engineering, 17(11): 138

[44] Xue L K , Wang T , Gao J , Ding A J , Zhou X H , Blake D R , Wang X F , Saunders S M , Fan S J , Zuo H C . . (2014). Ground-level ozone in four Chinese cities: precursors, regional transport and heterogeneous processes. Atmospheric Chemistry and Physics, 14(23): 13175–13188

[45] Zhang K , Zhou L , Fu Q Y , Yan L , Bian Q G , Wang D F , Xiu G L . (2019). Vertical distribution of ozone over Shanghai during late spring: a balloon-borne observation. Atmospheric Environment, 208: 48–60

[46] Zhang Y B , Yu S C , Chen X , Li Z , Li M Y , Song Z , Liu W P , Li P F , Zhang X Y , Lichtfouse E . . (2022). Local production, downward and regional transport aggravated surface ozone pollution during the historical orange-alert large-scale ozone episode in eastern China. Environmental Chemistry Letters, 20(3): 1577–1588

[47] Zhao D D , Xin J Y , Wang W F , Jia D J , Wang Z F , Xiao H , Liu C , Zhou J , Tong L , Ma Y J . . (2022). Effects of the sea-land breeze on coastal ozone pollution in the Yangtze River Delta, China. Science of the Total Environment, 807: 150306

[48] Zhao H , Chen K Y , Liu Z , Zhang Y X , Shao T , Zhang H L . (2021). Coordinated control of PM_2.5 and O₃ is urgently needed in China after implementation of the “Air pollution prevention and control action plan”. Chemosphere, 270: 129441

[49] Zheng D Y , Huang X J , Guo Y H . (2022). Spatiotemporal variation of ozone pollution and health effects in China. Environmental Science and Pollution Research, 29(38): 57808–57822

[50] Zhong S F , Zhang K , Bagheri M , Burken J G , Gu A , Li B K , Ma X M , Marrone B L , Ren Z J , Schrier J . . (2021). Machine learning: new ideas and tools in environmental science and engineering. Environmental Science & Technology, 55(19): 12741–12754

[51] Zhu Q Y , Bi J Z , Liu X , Li S S , Wang W H , Zhao Y , Liu Y . (2022). Satellite-based long-term spatiotemporal patterns of surface ozone concentrations in China: 2005–2019. Environmental Health Perspectives, 130(2): 027004

[52] Zhu Y , Liu J , Wang T J , Zhuang B L , Han H , Wang H M , Chang Y , Ding K . (2017). The impacts of meteorology on the seasonal and interannual variabilities of ozone transport from North America to East Asia. Journal of Geophysical Research: Atmospheres, 122(20): 10612–10636

RIGHTS & PERMISSIONS

The Author(s) 2025. This article is published with open access at link.springer.com and journal.hep.com.cn