Particulate Matter Forecasting Using Hybrid Autoencoders: The Role of Meteorological Data

Jarosław Bernacki

Abstract


Accurate forecasting of particulate matter (PM) concentrations is crucial for effective air quality management and public health protection. In this study, we propose two deep learning-based forecasting models: a convolutional-recurrent autoencoder and a hierarchical autoencoder. Both models are trained and evaluated using historical data on PM2.5 and PM10 concentrations from multiple monitoring stations in Poland. To assess the influence of meteorological conditions on prediction accuracy, the models are tested in two variants: one using historical PM concentrations and meteorological features such as temperature, wind speed, wind direction, and air pressure, and another that uses only historical PM data. The results clearly show that the inclusion of weather data significantly improves forecasting performance, with lower MAE, and MSE values observed across all test sites. The models trained with meteorological inputs consistently outperform their counterparts trained on PM data alone. The results are also compared with the baseline model. These findings highlight the importance of environmental context in air pollution forecasting and demonstrate the potential of autoencoder-based approaches for this task.


Keywords


air pollution; forecasting; deep models; particulate matter

Full Text:

PDF

References


Borah, J., Nadzir, M.S.M., Cayetano, M.G., Majumdar, S., Ghayvat, H., Srivastava, G. (2024). AiCareAir: Hybrid-ensemble Internet of Things sensing unit model for air pollutant control. IEEE Sensors Journal, 99. https://doi.org/10.1109/JSEN.2024.3397735

Cichowicz, R., Wielgosiński, G., Fetter, W. (2020). Effect of wind speed on the level of particulate matter PM10 concentration in atmospheric air during winter season in vicinity of large combustion plant. Journal of Atmospheric Chemistry, 77, 35–48. https://doi.org/10.1007/s10874-020-09401-w

Czernecki, B., Marosz, M., Jędruszkiewicz, J. (2021). Assessment of machine learning algorithms in short-term forecasting of PM10 and PM2.5 concentrations in selected Polish agglomerations. Aerosol and Air Quality Research, 21, 200586. https://doi.org/10.4209/aaqr.200586

Du, P., Wang, J., Yang, W., Niu, T. (2022). A novel hybrid fine particulate matter (PM2.5) forecasting and its further application system: Case studies in China. Journal of Forecasting, 41, 64–85. https://doi.org/10.1002/for.2785

Gryech, I., Asaad, C., Ghogho, M., Kobbane, A. (2024). Applications of machine learning & Internet of Things for outdoor air pollution monitoring and prediction: A systematic literature review. Engineering Applications of Artificial Intelligence, 137, 109182. https://doi.org/10.1016/j.engappai.2024.109182

Harishkumar, K., Yogesh, K., Gad, I., Doreswamy, N. (2020). Forecasting air pollution particulate matter (PM2.5) using machine learning regression models. Procedia Computer Science, 171, 20572066. https://doi.org/10.1016/j.procs.2020.04.221

Jasiński, R., Galant-Gołębiowska, M., Nowak, M., Ginter, M., Kurzawska, P., Kurtyka, K., Maciejewska, M. (2021). Case study of pollution with particulate matter in selected locations of Polish cities. Energies, 14(9), 2529. https://doi.org/10.3390/en14092529

Kouziokas, G.N. (2020). SVM kernel based on particle swarm optimized vector and Bayesian optimized SVM in atmospheric particulate matter forecasting. Applied Soft Computing, 93, 106410. https://doi.org/10.1016/j.asoc.2020.106410

Kowalski, P., Sapała, K., Warchałowski, W. (2020). PM10 forecasting through applying convolution neural network techniques. International Journal of Environmental Impacts, 3(1), 31–43. https://doi.org/10.2495/EI-V3-N1-31-43

Kryza, M., Werner, M., Dore, A.-J. (2019). Application of degree-day factors for residential emission estimate and air quality forecasting. International Journal of Environment and Pollution, 65(4), 325–336. https://doi.org/10.1504/IJEP.2019.103748

Kujawska, J., Kulisz, M., Oleszczuk, P., Cel, W. (2022). Machine learning methods to forecast the concentration of PM10 in Lublin, Poland. Energies, 15(17), 6428. https://doi.org/10.3390/en15176428

Li, T., Hua, M., Wu, X. (2020). A hybrid CNN-LSTM model for forecasting particulate matter (PM2.5). IEEE Access, 8, 2693326940. https://doi.org/10.1109/ACCESS.2020.2971348

Mauricio-Alvarez, L.-E., Aceves-Fernandez, M.-A., Pedraza-Ortega, J.-C., Ramos-Arreguin, J.-M. (2024). Evaluation of a transformer-based model for the temporal forecast of coarse particulate matter (PMCO) concentrations. Earth Science Informatics, 17, 3095–3110. https://doi.org/10.1007/s12145-024-01330-6

Nidzgorska-Lencewicz, J. (2018). Application of artificial neural networks in the prediction of PM10 levels in the winter months: A case study in the Tricity agglomeration, Poland. Atmosphere, 9(6), 203. https://doi.org/10.3390/atmos9060203

Penkała, M., Rogula-Kozłowska, W., Ogrodnik, P., Bihałowicz, J.S., Iwanicka, N. (2023). Exploring the relationship between particulate matter emission and the construction material of road surface: Case study of highways and motorways in Poland. Materials, 16, 1200. https://doi.org/10.3390/ma16031200

Polichetti, G., Cocco, S., Spinali, A., Trimarco, V., Nunziata, A. (2009). Effects of particulate matter (PM10, PM2.5 and PM1) on the cardiovascular system. Toxicology, 261, 1–8. https://doi.org/10.1016/j.tox.2009.04.035

Połednik, B. (2022). Emissions of air pollution in industrial and rural region in Poland and health impacts. Journal of Ecological Engineering, 23, 250258. https://doi.org/10.12911/22998993/151986

Ramentol, E., Grimm, S., Stinzendorfer, M., &Wagner, A. (2023). Short-term air pollution forecasting using embeddings in neural networks. Atmosphere, 14, 298. https://doi.org/10.3390/atmos14020298

Rogula-Kozłowska, W., Klejnowski, K., Rogula-Kopiec, P., Ośródka, L., Krajny, E., Błaszczak, B., Mathews, B. (2014). Spatial and seasonal variability of the mass concentration and chemical composition of PM2.5 in Poland. Air Quality, Atmosphere & Earth, 7, 41–58. https://doi.org/10.1007/s11869-013-0222-y

Sharma, E., Deo, R.C., Prasad, R., Parisi, A.V., Raj, N. (2020). Deep air quality forecasts: Suspended particulate matter modeling with convolutional neural and long short-term memory networks. IEEE Access, 8, 209503–209516. https://doi.org/10.1109/ACCESS.2020.3039002

Sowka, I., Chlebowska-Styś, A., Pachurka, Ł., Rogula-Kozłowska, W., Mathews, B. (2019). Analysis of particulate matter concentration variability and origin in selected urban areas in Poland. Sustainability, 11, 5735. https://doi.org/10.3390/su11205735

Swetha, G., Datla, R., Vishnu, C., Mohan, C.K. (2024). M2-APNet: A multimodal deep learning network to predict major air pollutants from temporal satellite images. Journal of Applied Remote Sensing, 18, 012005–012005. https://doi.org/10.1117/1.JRS.18.012005

Tariq, S., Loy-Benitez, J., Yoo, C. (2023). Enhancing the sustainable management of fine particulate matter-related health risks at subway stations through sequential forecast and gated probabilistic transformer. Building and Environment, 244, 110780. https://doi.org/10.1016/j.buildenv.2023.110780

Tong, W., Limperis, J., Hamza-Lup, F., Hu, Y., Li, L. (2024). Robust transformer-based model for spatiotemporal PM2.5 prediction in California. Earth Science Informatics, 17, 315–328. https://doi.org/10.1007/s12145-023-01138-w

Tran, D., Nguyen, H., Tran, B., La Vecchia, C., Luu, H.N., Nguyen, T. (2021). Fast and precise single-cell data analysis using a hierarchical autoencoder. Nature Communications, 12, 1029. https://doi.org/10.1038/s41467-021-21312-2

Qin, S., Liu, F., Wang, J., Sun, B. (2014). Analysis and forecasting of the particulate matter (PM) concentration levels over four major cities of China using hybrid models. Atmospheric Environment, 98, 665–675. https://doi.org/10.1016/j.atmosenv.2014.09.046

Vovk, T., Kryza, M., Werner, M. (2024). Using random forest to improve EMEP4PL model estimates of daily PM2.5 in Poland. Atmospheric Environment, 332, 120615. https://doi.org/10.1016/j.atmosenv.2024.120615

Won, W.-S., Oh, R., Lee, W., Ku, S., Su, P.-C., Yoon, Y.-J. (2021). Hygroscopic properties of particulate matter and effects of their interactions with weather on visibility. Scientific Reports, 11, 16401. https://doi.org/10.1038/s41598-021-95834-6

Zeng, Q., Wang, L., Zhu, S., Gao, Y., Qiu, X., Chen, L. (2023). Long-term PM2.5 concentrations forecasting using CEEMDAN and deep transformer neural network. Atmospheric Pollution Research, 14(9), 101839. https://doi.org/10.1016/j.apr.2023.101839

Zheng, Z., Zhang, Z. (2023). A temporal convolutional recurrent autoencoder based framework for compressing time series data. Applied Soft Computing, 147, 110797. https://doi.org/10.1016/j.asoc.2023.110797

https://meteostat.net/en

https://www.gov.pl/web/gios-en




DOI: http://dx.doi.org/10.17951/b.2025.80.0.265-283
Date of publication: 2025-12-05 15:39:24
Date of submission: 2025-07-23 14:55:50


Statistics


Total abstract view - 0
Downloads (from 2020-06-17) - PDF - 0

Indicators



Refbacks

  • There are currently no refbacks.


Copyright (c) 2025

Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 International License.