Deep learning for air pollutant forecasting: opportunities, challenges, and future directions

Abstract

Deep learning methods are increasingly employed to forecast air quality from an ever-increasing stream of data and algorithms. However, the efficacy of current approaches may be questionable when evaluated not solely in terms of greater forecasting fidelity, but also concerning the decision-making process in pollution early warning. Here, rather than amending classical machine learning algorithms, we argue that now is the time to push the frontiers of air pollutant forecasting beyond state-of-the-art approaches. This can be achieved through near real-time assimilation of multi-scale observations for laying the foundation of training data, enhanced attribution methods for impending heavy pollution, diagnostics for forecasting uncertainty, and advanced climate-chemistry emulators for improving seasonal forecasting. To harness this potential, it is essential to address several key challenges in deep learning methods, particularly generalization ability in extreme events, physics-informed interpretable approaches, and the mitigation technology of cumulative errors in multi-process coupled systems. This interdisciplinary endeavor will remain a central pursuit in the quest to anticipate and manage environmental change.

Graphical abstract

Keywords

Deep learning / Air pollution forecasting / Data assimilation / Seasonal forecasting

Highlight

	● Real-time data assimilation is essential to build reliable and actionable forecasting tools.
	● Advances in model architectures help to overcome the limitations of deep learning-based air quality forecasting.
	● Cross-sphere coupling hold promise for enabling data-driven seasonal air pollutant forecasting.

References

[1] AbnarSZuidemaW H (2020). Quantifying attention flow in transformers. In: Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics. Stroudsburg: Association for Computational Linguistics

[2] Agapiou A . (2017). Remote sensing heritage in a petabyte-scale: satellite data and heritage earth engine© applications. International Journal of Digital Earth, 10(1): 85–102

[3] Appel K W , Bash J O , Fahey K M , Foley K M , Gilliam R C , Hogrefe C , Hutzell W T , Kang D W , Mathur R , Murphy B N . . (2021). The Community Multiscale Air Quality (CMAQ) model versions 5.3 and 5.3.1: system updates and evaluation. Geoscientific Model Development, 14(5): 2867–2897

[4] Bai K X , Li K , Shao L Q , Li X R , Liu C S , Li Z Q , Ma M L , Han D , Sun Y B , Zheng Z . . (2024). LGHAP v2: a global gap-free aerosol optical depth and PM_2.5 concentration dataset since 2000 derived via big Earth data analytics. Earth System Science Data, 16(5): 2425–2448

[5] Bai L B , Du S H , Zhang X Y , Wang H Y , Liu B , Ouyang S . (2022). Domain adaptation for remote sensing image semantic segmentation: an integrated approach of contrastive learning and adversarial learning. IEEE Transactions on Geoscience and Remote Sensing, 60: 5628313

[6] Baklanov A , Brunner D , Carmichael G , Flemming J , Freitas S , Gauss M , Hov Ø , Mathur R , Schlünzen K H , Seigneur C . . (2017). Key issues for seamless integrated chemistry–meteorology modeling. Bulletin of the American Meteorological Society, 98(11): 2285–2292

[7] Baklanov A , Schlünzen K , Suppan P , Baldasano J , Brunner D , Aksoyoglu S , Carmichael G , Douros J , Flemming J , Forkel R . . (2014). Online coupled regional meteorology chemistry models in Europe: current status and prospects. Atmospheric Chemistry and Physics, 14(1): 317–398

[8] Baró R , Palacios-Peña L , Baklanov A , Balzarini A , Brunner D , Forkel R , Hirtl M , Honzak L , Pérez J L , Pirovano G . . (2017). Regional effects of atmospheric aerosols on temperature: an evaluation of an ensemble of online coupled models. Atmospheric Chemistry and Physics, 17(15): 9677–9696

[9] Beucler T , Pritchard M , Rasp S , Ott J , Baldi P , Gentine P . (2021). Enforcing analytic constraints in neural networks emulating physical systems. Physical Review Letters, 126(9): 098302

[10] Bi J Z , Knowland K E , Keller C A , Liu Y . (2022). Combining machine learning and numerical simulation for high-resolution PM_2.5 concentration forecast. Environmental Science & Technology, 56(3): 1544–1556

[11] Bi K F , Xie L X , Zhang H H , Chen X , Gu X T , Tian Q . (2023). Accurate medium-range global weather forecasting with 3D neural networks. Nature, 619(7970): 533–538

[12] Bodnar C , Bruinsma W P , Lucic A , Stanley M , Allen A , Brandstetter J , Garvan P , Riechert M , Weyn J A , Dong H Y . . (2025). A foundation model for the Earth system. Nature, 641(8065): 1180–1187

[13] BonevBKurthTHundtCPathakJBaustMKashinathKAnandkumarA (2023). Spherical Fourier neural operators: learning stable dynamics on the sphere. In: Proceedings of the 40th International Conference on Machine Learning. Honolulu: JMLR.org, 117

[14] Bucsela E J , Krotkov N A , Celarier E A , Lamsal L N , Swartz W H , Bhartia P K , Boersma K F , Veefkind J P , Gleason J F , Pickering K E . (2013). A new stratospheric and tropospheric NO₂ retrieval algorithm for nadir-viewing satellite instruments: applications to OMI. Atmospheric Measurement Techniques, 6(10): 2607–2626

[15] Buizza R . (2019). Introduction to the special issue on “25 Years of Ensemble Forecasting”. Quarterly Journal of the Royal Meteorological Society, 145(S1): 1–11

[16] Byun D , Schere K L . (2006). Review of the governing equations, computational algorithms, and other components of the models-3 community multiscale air quality (CMAQ) modeling system. Applied Mechanics Reviews, 59(2): 51–77

[17] Caldwell P M , Mametjanov A , Tang Q , Van Roekel L P , Golaz J C , Lin W Y , Bader D C , Keen N D , Feng Y , Jacob R . . (2019). The DOE E3SM coupled model version 1: description and results at high resolution. Journal of Advances in Modeling Earth Systems, 11(12): 4095–4146

[18] Chen J X , Zhu S Q , Wang P , Zheng Z H , Shi S , Li X Y , Xu C , Yu K X , Chen R J , Kan H D . . (2024a). Predicting particulate matter, nitrogen dioxide, and ozone across great britain with high spatiotemporal resolution based on random forest models. Science of the Total Environment, 926: 171831

[19] Chen K , Han T , Ling F H , Gong J C , Bai L , Wang X Y , Luo J J , Fei B , Zhang W L , Chen X . . (2025). The operational medium-range deterministic weather forecasting can be extended beyond a 10-day lead time. Communications Earth & Environment, 6(1): 518

[20] Chen K H , Li G B , Li H W , Wang Y Q , Wang W Z , Liu Q Y , Wang H C . (2024b). Quantifying uncertainty: air quality forecasting based on dynamic spatial-temporal denoising diffusion proba-bilistic model. Environmental Research, 249: 118438

[21] Chen L , Zhong X H , Li H , Wu J , Lu B , Chen D L , Xie S P , Wu L B , Chao Q C , Lin C S . . (2024c). A machine learning model that outperforms conventional global subseasonal forecast models. Nature Communications, 15(1): 6425

[22] Chen W Y , Lu X C , Yuan D H , Chen Y A , Li Z N , Huang Y Q , Fung T , Sun H C , Fung J C H . (2023). Global PM_2.5 prediction and associated mortality to 2100 under different climate change scenarios. Environmental Science & Technology, 57(27): 10039–10052

[23] Christopher S A , Gupta P . (2010). Satellite remote sensing of particulate matter air quality: the cloud-cover problem. Journal of the Air & Waste Management Association, 60(5): 596–602

[24] 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

[25] Collins W J , Lamarque J F , Schulz M , Boucher O , Eyring V , Hegglin M I , Maycock A , Myhre G , Prather M , Shindell D . . (2017). AerChemMIP: quantifying the effects of chemistry and aerosols in CMIP6. Geoscientific Model Development, 10(2): 585–607

[26] Copernicus Sentinel-5P (processed by ESA) (2020). TROPOMI Level 2 Formaldehyde Total Column products. Version 02. Paris: European Space Agency

[27] Di Q , Kloog I , Koutrakis P , Lyapustin A , Wang Y J , Schwartz J . (2016). Assessing PM_2.5 exposures with high spatiotemporal resolution across the continental United States. Environmental Science & Technology, 50(9): 4712–4721

[28] Domeisen D I V , White C J , Afargan-Gerstman H , Muñoz Á G , Janiga M A , Vitart F , Wulff C O , Antoine S , Ardilouze C , Batté L . . (2022). Advances in the subseasonal prediction of extreme events: relevant case studies across the globe. Bulletin of the American Meteorological Society, 103(6): E1473–E1501

[29] DosovitskiyABeyerLKolesnikovAWeissenbornDZhaiX HUnterthinerTDehghaniMMindererMHeigoldGGellyS, . (2021). An image is worth 16 × 16 words: transformers for image recognition at scale. In: Proceedings of the 9th International Conference on Learning Representations. Virtual Only Conference, Vienna, Austria

[30] Ehrendorfer M . (1997). Predicting the uncertainty of numerical weather forecasts: a review. Meteorologische Zeitschrift, 6(4): 147–183

[31] FanXWangZ HLinY TZhangYXiangYLiH (2025). MVAR: MultiVariate Auto Regressive Air Pollutants Forecasting Model. Shanghai: Shanghai Academy of Artificial Intelligence for Science, arxiv.org/abs/2507.12023

[32] FranchiGLaurentOLeguéryMBursucAPilzerAYaoA (2024). Make me a BNN: a simple strategy for estimating Bayesian uncertainty from pre-trained models. In: Proceedings of 2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Seattle: IEEE, 12194–12204

[33] GalYGhahramaniZ (2016). Dropout as a Bayesian approximation: representing model uncertainty in deep learning. In: Proceedings of the 33rd International Conference on Machine Learning. New York: JMLR.org, 1050–1059

[34] Geng G N , Xiao Q Y , Liu S G , Liu X D , Cheng J , Zheng Y X , Xue T , Tong D , Zheng B , Peng Y R . . (2021). Tracking air pollution in China: near real-time PM_2.5 retrievals from multisource data fusion. Environmental Science & Technology, 55(17): 12106–12115

[35] GriffithsP TWilcoxL JAllenR JNaikVO’ConnorF MPratherM JArchibaldA TBrownFDeushiMCollinsW, . (2024). The role of AerChemMIP in advancing climate and air quality research. Atmospheric Chemistry and Physics, 25, 8289–8328

[36] Guastavino S , Piana M , Tizzi M , Cassola F , Iengo A , Sacchetti D , Solazzo E , Benvenuto F . (2022). Prediction of severe thunder-storm events with ensemble deep learning and radar data. Scientific Reports, 12(1): 20049

[37] Hammer M S , Van Donkelaar A , Bindle L , Sayer A M , Lee J , Hsu N C , Levy R C , Sawyer V , Garay M J , Kalashnikova O V . . (2023). Assessment of the impact of discontinuity in satellite instruments and retrievals on global PM_2.5 estimates. Remote Sensing of Environment, 294: 113624

[38] HeK MChenX LXieS NLiY HDollárPGirshickR (2022a). Masked autoencoders are scalable vision learners. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. New Orleans: IEEE, 15979–15988

[39] He T L , Jones D B A , Miyazaki K , Bowman K W , Jiang Z , Chen X K , Li R , Zhang Y X , Li K N . (2022b). Inverse modelling of Chinese NO_x emissions using deep learning: integrating in situ observations with a satellite-based chemical reanalysis. Atmospheric Chemistry and Physics, 22(21): 14059–14074

[40] HettigeK HJiJ HXiangS LLongCCongGWangJ Y (2024). AirPhyNet: harnessing physics-guided neural networks for air quality prediction. In: Proceedings of the 12th International Conference on Learning Representations. Vienna: ICLR

[41] Hirsch E , Koren I . (2021). Record-breaking aerosol levels explained by smoke injection into the stratosphere. Science, 371(6535): 1269–1274

[42] HoJJainAAbbeelP (2020). Denoising diffusion probabilistic models. In: Proceedings of the 34th International Conference on Neural Information Processing Systems. Vancouver: Curran Associates Inc., 574

[43] Hong C P , Zhang Q , Zhang Y , Davis S J , Tong D , Zheng Y X , Liu Z , Guan D B , He K B , Schellnhuber H J . (2019). Impacts of climate change on future air quality and human health in China. Proceedings of the National Academy of Sciences of the United States of America, 116(35): 17193–17200

[44] Hu M Y , Lu X C , Chen Y A , Li Z N , Wang Y Y , Fung J C H . (2025). AirQFormer: improving regional air quality forecast with a hybrid deep learning model. Sustainable Cities and Society, 119: 106113

[45] Huang C H , Hu J L , Xue T , Xu H , Wang M . (2021a). High-resolution spatiotemporal modeling for ambient PM_2.5 exposure assessment in China from 2013 to 2019. Environmental Science & Technology, 55(3): 2152–2162

[46] Huang C Y , Hu T T , Duan Y S , Li Q Y , Chen N , Wang Q , Zhou M G , Rao P H . (2022). Effect of urban morphology on air pollution distribution in high-density urban blocks based on mobile monitoring and machine learning. Building and Environment, 219: 109173

[47] Huang L , Liu S , Yang Z Y , Xing J , Zhang J , Bian J , Li S W , Sahu S K , Wang S X , Liu T Y . (2021b). Exploring deep learning for air pollutant emission estimation. Geoscientific Model Development, 14(7): 4641–4654

[48] Jairi I , Ben-Othman S , Canivet L , Zgaya-Biau H . (2024). Enhancing air pollution prediction: a neural transfer learning approach across different air pollutants. Environmental Technology & Innovation, 36: 103793

[49] Jiang F , Zheng Z H , Coe H , Healy R M , Poulain L , Gros V , Zhang H , Li W J , Liu D T , West M . . (2025). Integrating simulations and observations: a foundation model for estimating the aerosol mixing state index. ACS ES&T Air, 2(5): 877–890

[50] KarrasTAittalaMLaineSAilaT (2022). Elucidating the design space of diffusion-based generative models. In: Proceedings of the 36th International Conference on Neural Information Processing Systems. New Orleans: Curran Associates Inc., 1926

[51] Kelp M M , Jacob D J , Kutz J N , Marshall J D , Tessum C W . (2020). Toward stable, general machine-learned models of the atmospheric chemical system. Journal of Geophysical Research: Atmospheres, 125(23): e2020JD032759

[52] Kelp M M , Jacob D J , Lin H P , Sulprizio M P . (2022). An online-learned neural network chemical solver for stable long-term global simulations of atmospheric chemistry. Journal of Advances in Modeling Earth Systems, 14(6): e2021MS002926

[53] Kochkov D , Yuval J , Langmore I , Norgaard P , Smith J , Mooers G , Klöwer M , Lottes J , Rasp S , Düben P . . (2024). Neural general circulation models for weather and climate. Nature, 632(8027): 1060–1066

[54] Lam R , Sanchez-Gonzalez A , Willson M , Wirnsberger P , Fortunato M , Alet F , Ravuri S , Ewalds T , Eaton-Rosen Z , Hu W H . . (2023). Learning skillful medium-range global weather forecasting. Science, 382(6677): 1416–1421

[55] Lamarque J F , Emmons L K , Hess P G , Kinnison D E , Tilmes S , Vitt F , Heald C L , Holland E A , Lauritzen P H , Neu J . . (2012). CAM-chem: description and evaluation of interactive atmospheric chemistry in the community earth system model. Geoscientific Model Development, 5(2): 369–411

[56] Larkin A , Anenberg S , Goldberg D L , Mohegh A , Brauer M , Hystad P . (2023). A global spatial-temporal land use regression model for nitrogen dioxide air pollution. Frontiers in Environmental Science, 11: 1125979

[57] Li J W , Han Z W , Surapipith V , Fan W X , Thongboonchoo N , Wu J , Li J , Tao J , Wu Y F , Macatangay R . . (2022a). Direct and indirect effects and feedbacks of biomass burning aerosols over mainland southeast Asia and south China in springtime. Science of the Total Environment, 842: 156949

[58] Li J Y , Cai Y X , Li Q , Kou M Y , Zhang T X . (2024a). A review of remote sensing image segmentation by deep learning methods. International Journal of Digital Earth, 17(1): 2328827

[59] Li M Y , Yang Q Q , Yuan Q Q , Zhu L Y . (2022b). Estimation of high spatial resolution ground-level ozone concentrations based on Landsat 8 TIR bands with deep forest model. Chemosphere, 301: 134817

[60] Li S W , Xing J . (2025). Enhancing 72-hour air quality forecasting with an observation-driven deep learning chemistry transport model. Environment International, 202: 109689

[61] Li Z Y , Han W , Zhang Y , Fu Q F , Li J X , Qin L Z , Dong R Y , Sun H , Deng Y , Yang L J . (2024b). Learning spatiotemporal dynamics with a pretrained generative model. Nature Machine Intelligence, 6(12): 1566–1579

[62] Libonati R , Geirinhas J L , Silva P S , Russo A , Rodrigues J A , Belém L B C , Nogueira J , Roque F O , DaCamara C C , Nunes A M B . . (2022). Assessing the role of compound drought and heatwave events on unprecedented 2020 wildfires in the Pantanal. Environmental Research Letters, 17(1): 015005

[63] Lin Y . (2024). Progressive neural network for multi-horizon time series forecasting. Information Sciences, 661: 120112

[64] Liu H , Xu Y N , Chen C . (2019). Improved pollution forecasting hybrid algorithms based on the ensemble method. Applied Mathematical Modelling, 73: 473–486

[65] Liu Z , Lin Y T , Cao Y , Hu H , Wei Y X , Zhang Z , Lin S , Guo B N . (2021a). Swin transformer: hierarchical vision transformer usingshifted windows. In: Proceedings of 2021 IEEE/CVF International Conference on Computer Vision. Montreal: IEEE, , 9992–1000

[66] Liu N , Xiong A Y , Zhang Q , Liu Y J , Zhan Y J , Liu Y M . (2021b). Development of basic dataset of severe convective weather for artificial intelligence training. Journal of Applied Meteorological Science, 32(5): 530–541

[67] Liu W Q , Xing C Z . (2024). Needs and challenges of optical atmospheric monitoring on the background of carbon neutrality in China. Frontiers of Environmental Science & Engineering, 18(6): 73

[68] Liu X , Zhu Y J , Xue L , Desai A R , Wang H K . (2022). Cluster-enhanced ensemble learning for mapping global monthly surface ozone from 2003 to 2019. Geophysical Research Letters, 49(11): e2022GL097947

[69] LiuZLinY TCaoYHuHWeiY XZhangZLinSGuoB N (2021). Swin transformer: hierarchical vision transformer using shifted windows. In: Proceedings of 2021 IEEE/CVF International Conference on Computer Vision. Montreal: IEEE, 9992–10002

[70] Lu M Y , Lao T F , Yu M Z , Zhang Y D , Zheng J Q , Li Y C . (2021). PM_2.5 concentration forecasting over the central area of the Yangtze River delta based on deep learning considering the spatial diffusion process. Remote Sensing, 13(23): 4834

[71] Lu X , Zhang S J , Xing J , Wang Y J , Chen W H , Ding D , Wu Y , Wang S X , Duan L , Hao J M . (2020). Progress of air pollution control in China and its challenges and opportunities in the ecological civilization era. Engineering, 6(12): 1423–1431

[72] LundbergS MLeeS I (2017). A unified approach to interpreting model predictions. In: Proceedings of the 31st International Conference on Neural Information Processing Systems. Long Beach: Curran Associates Inc., 4768–4777

[73] LyapustinAWangY (2022). MODIS/Terra+Aqua Land Aerosol Optical Depth Daily L2G Global 1km SIN Grid. Washington, DC: NASA EOSDIS Land Processes DAAC

[74] LyuBHuangRWangX LWangW GHuY T (2024). FastCTM (v1.0): atmospheric chemical transport modelling with a principle-informed neural network for air quality simulations. Geoscientific Model Development Discussions, doi: 10.5194/gmd-2024-1982024

[75] Ma J , Li Z , Cheng J C P , Ding Y X , Lin C Q , Xu Z R . (2020). Air quality prediction at new stations using spatially transferred bi-directional long short-term memory network. Science of the Total Environment, 705: 135771

[76] MariottiARutiP MRixenM (2018). Progress in subseasonal to seasonal prediction through a joint weather and climate community effort. npj Climate and Atmospheric Science, 1(1): 4

[77] MechosoC R (2020). Interacting Climates of Ocean Basins: Observations, Mechanisms, Predictability, and Impacts. Cambridge: Cambridge University Press

[78] Miyazaki K , Eskes H J , Sudo K . (2012). Global NO_x emission estimates derived from an assimilation of OMI tropospheric NO₂ columns. Atmospheric Chemistry and Physics, 12(5): 2263–2288

[79] Mo X Y , Li H , Zhang L . (2022). Design a regional and multistep air quality forecast model based on deep learning and domain knowledge. Frontiers in Earth Science, 10: 995843

[80] MolnarC (2025). Interpretable Machine Learning: a Guide for Making Black Box Models Explainable. 3rd ed. ISBN: 978-3-911578-03-5

[81] Montavon G , Samek W , Müller K R . (2018). Methods for interpreting and understanding deep neural networks. Digital Signal Processing, 73: 1–15

[82] Price I , Sanchez-Gonzalez A , Alet F , Andersson T R , El-Kadi A , Masters D , Ewalds T , Stott J , Mohamed S , Battaglia P . . (2025). Probabilistic weather forecasting with machine learning. Nature, 637(8044): 84–90

[83] RanNXiaoPWangYShiWLinJ XMengQAllmendingerR (2025). HR-Extreme: a high-resolution dataset for extreme weather forecasting. In: Proceedings of the 13th International Conference on Learning Representations. Singapore: ICLR

[84] RaviRKumariN SGeethikaP S SRaoK VRaoM S (2024). Air pollution forecasting using deep learning algorithms: a review. In: Lin F M, Patel A, Kesswani N, Sambana B, eds. Accelerating Discoveries in Data Science and Artificial Intelligence I. Cham: Springer, 511–517

[85] 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

[86] Ren X , Mi Z Y , Cai T , Nolte C G , Georgopoulos P G . (2022). Flexible Bayesian ensemble machine learning framework for predicting local ozone concentrations. Environmental Science & Technology, 56(7): 3871–3883

[87] Requia W J , Di Q , Silvern R , Kelly J T , Koutrakis P , Mickley L J , Sulprizio M P , Amini H , Shi L H , Schwartz J . (2020). An ensemble learning approach for estimating high spatiotemporal resolution of ground-level ozone in the contiguous United States. Environmental Science & Technology, 54(18): 11037–11047

[88] SchreckJ SGagneD JBeckerCChapmanW EElmoreKFanDGantosGKimEKimparaDMartinT, . (2024). Evidential deep learning: enhancing predictive uncertainty estimation for Earth system science applications. Artificial Intelligence for the Earth Systems, 3(4): 230093

[89] Selvaraju R R , Cogswell M , Das A , Vedantam R , Parikh D , Batra D . (2020). Grad-CAM: visual explanations from deep networks via gradient-based localization. International Journal of Computer Vision, 128(2): 336–359

[90] Sha Y K , Sobash R A , Gagne D J . (2024). Generative ensemble deep learning severe weather prediction from a deterministic convection-allowing model. Artificial Intelligence for the Earth Systems, 3(2): e230094

[91] Shen L , Mickley L J . (2017). Seasonal prediction of US summertime ozone using statistical analysis of large scale climate patterns. Proceedings of the National Academy of Sciences of the United States of America, 114(10): 2491–2496

[92] Shen S Y , Li C , van Donkelaar A , Jacobs N , Wang C G , Martin R V . (2024). Enhancing global estimation of fine particulate matter concentrations by including geophysical a priori information in deep learning. ACS ES&T Air, 1(5): 332–345

[93] Song Z H , Chen B , Zhang P , Guan X D , Wang X , Ge J M , Hu X Q , Zhang X Y , Wang Y X . (2022). High temporal and spatial resolution PM_2.5 dataset acquisition and pollution assessment based on FY-4A TOAR data and deep forest model in China. Atmospheric Research, 274: 106199

[94] Tao C L , Jia M , Wang G Q , Zhang Y Q , Zhang Q Z , Wang X F , Wang Q , Wang W X . (2024a). Time-sensitive prediction of NO₂ concentration in China using an ensemble machine learning model from multi-source data. Journal of Environmental Sciences, 137: 30–40

[95] Tao C L , Peng Y B , Zhang Q Z , Zhang Y Q , Gong B , Wang Q , Wang W X . (2024b). Diagnosing ozone-NO_x-VOC-aerosol sensitivity and uncovering causes of urban-nonurban discre-pancies in Shandong, China, using transformer-based estima-tions. Atmospheric Chemistry and Physics, 24(7): 4177–4192

[96] Thongthammachart T , Araki S , Shimadera H , Matsuo T , Kondo A . (2022). Incorporating Light Gradient Boosting Machine to land use regression model for estimating NO₂ and PM_2.5 levels in Kansai region, Japan. Environmental Modelling & Software, 155: 105447

[97] VaswaniAShazeerNParmarNUszkoreitJJonesLGomezA NKaiserŁPolosukhinI (2017). Attention is all you need. In: Proceedings of the 31st International Conference on Neural Information Processing Systems. Long Beach: Curran Associates Inc., 6000–6010

[98] WangC GPritchardM SBrenowitzNCohenYBonevBKurthTDurranDPathakJ (2024). Coupled Ocean-Atmosphere Dynamics in a Machine Learning Earth System Model. Princeton: Princeton University

[99] Wang Y , Puķīte J , Wagner T , Donner S , Beirle S , Hilboll A , Vrekoussis M , Richter A , Apituley A , Piters A . . (2018). Vertical profiles of tropospheric ozone from MAX-DOAS measurements during the CINDI-2 campaign. Part 1-development of a new retrieval algorithm. Journal of Geophysical Research: Atmospheres, 123(18): 10637–10670

[100] Wang Z X , Li J , Wu L , Zhu M M , Zhang Y J , Ye Z L , Wang Z F . (2022). Deep learning-based gas-phase chemical kinetics kernel emulator: application in a global air quality simulation case. Frontiers in Environmental Science, 10: 955980

[101] WatanachaturapornPAroraMVarshneyP (2006). Sub-Pixel Land Cover Classification Using Support Vector Machines. In: ASPRS 2006 Annual Conference. May 1–5, Reno, Nevada

[102] Wei J , Li Z Q , Lyapustin A , Wang J , Dubovik O , Schwartz J , Sun L , Li C , Liu S , Zhu T . (2023). First close insight into global daily gapless 1 km PM_2.5 pollution, variability, and health impact. Nature Communications, 14(1): 8349

[103] 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

[104] WhittakerTDiLuca A (2025). Pushing the Limits of Extreme Weather: Constructing Extreme Heatwave Storylines with Differentiable Climate Models. Montréal: Université du Québec à Montréal, arXiv:2506.10660

[105] Xie Y Y , Lin M Y , Decharme B , Delire C , Horowitz L W , Lawrence D M , Li F , Séférian R . (2022). Tripling of western US particulate pollution from wildfires in a warming climate. Proceedings of the National Academy of Sciences of the United States of America, 119(14): e2111372119

[106] Xing J , Li S , Zheng S , Liu C , Wang X , Huang L , Song G , He Y , Wang S , Sahu S K . . (2022). Rapid inference of nitrogen oxide emissions based on a top-down method with a physically informed variational autoencoder. Environmental Science & Technology, 56(14): 9903–9914

[107] Xing J , Zheng S X , Ding D , Kelly J T , Wang S X , Li S W , Qin T , Ma M Y , Dong Z X , Jang C . . (2020). Deep learning for prediction of the air quality response to emission changes. Environmental Science & Technology, 54(14): 8589–8600

[108] Xiong Y , Yang Q Q , Gao Y , Li K , Yang Y , Lin G X , Lu X , Wang Z L , Zhang H L , Gao M . (2025). Modeling the formation of aerosols and their interactions with weather and climate: critical review and future perspectives. Frontiers of Environmental Science & Engineering, 19(11): 143

[109] Xu J Z , Zhang H R , Cheng Z , Liu J Y , Xu Y Y , Wang Y C . (2022). Approximating three-dimensional (3-D) transport of atmospheric pollutants via deep learning. Earth and Space Science, 9(7): e2022EA002338

[110] Zhan J L , Liu Y C , Ma W , Zhang X , Wang X Z , Bi F , Zhang Y J , Wu Z H , Li H . (2022). Ozone formation sensitivity study using machine learning coupled with the reactivity of volatile organic compound species. Atmospheric Measurement Techniques, 15(5): 1511–1520

[111] Zhang A X , Fu T M , Feng X , Guo J F , Liu C F , Chen J K , Mo J J , Zhang X , Wang X L , Wu W L . . (2023). Deep learning-based ensemble forecasts and predictability assessments for surface ozone pollution. Geophysical Research Letters, 50(8): e2022GL102611

[112] Zhang B R . (2025). Comparative investigation of machine learning and deep learning approaches for air quality prediction. ITM Web of Conferences, 73: 02002

[113] Zhang C X , Niu X H , Wu H Y , Ding Z P , Chan K L , Kim J , Wagner T , Liu C . (2025). Unleashing the potential of geostationary satellite observations in air quality forecasting through artificial intelligence techniques. Atmospheric Chemistry and Physics, 25(2): 759–770

[114] Zhang Z , Zhang S . (2023). Modeling air quality PM_2.5 forecasting using deep sparse attention-based transformer networks. International Journal of Environmental Science and Technology, 20(12): 13535–13550

[115] Zhao N , Zhang Y Q , Xue L K . (2025). Nonlinear relationship between air pollution and precursor emissions in Qingdao, eastern China. Frontiers of Environmental Science & Engineering, 19(1): 9

[116] Zheng Z H , Fiore A M , Westervelt D M , Milly G P , Goldsmith J , Karambelas A , Curci G , Randles C A , Paiva A R , Wang C . . (2023). Automated machine learning to evaluate the information content of tropospheric trace gas columns for fine particle estimates over India: a modeling testbed. Journal of Advances in Modeling Earth Systems, 15(3): e2022MS003099

[117] Zhong Q R , Schutgens N , Veraverbeke S , van der Werf G R . (2024). Increasing aerosol emissions from boreal biomass burning exacerbate arctic warming. Nature Climate Change, 14(12): 1275–1281

[118] Zhou Y L , Chang F J , Chang L C , Kao I F , Wang Y S . (2019). Explore a deep learning multi-output neural network for regional multi-step-ahead air quality forecasts. Journal of Cleaner Production, 209: 134–145

[119] 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

[120] Zhu S Q , Ma J L , Wang S Y , Sun S D , Wang P , Zhang H L . (2023). Shifts of formation regimes and increases of atmospheric oxidation led to ozone increase in North China Plain and Yangtze River Delta from 2016 to 2019. Journal of Geophysical Research: Atmospheres, 128(13): e2022JD038373

RIGHTS & PERMISSIONS

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