Articles | Volume 18, issue 12
https://doi.org/10.5194/gmd-18-3707-2025
© Author(s) 2025. This work is distributed under
the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
https://doi.org/10.5194/gmd-18-3707-2025
© Author(s) 2025. This work is distributed under
the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Development and technical paper
| 
23 Jun 2025
Development and technical paper |
| 23 Jun 2025
| 23 Jun 2025
Knowledge-inspired fusion strategies for the inference of PM2.5 values with a neural network
Matthieu Dabrowski
CORRESPONDING AUTHOR
Centre de Recherche en Informatique Signal et Automatique de Lille, Univ. Lille, CNRS, Centrale Lille, UMR 9189 CRIStAL, 59000 Lille, France
José Mennesson
IMT Nord Europe, Institut Mines-Télécom, Univ. Lille, Centre for Digital Systems, 59000 Lille, France
Jérôme Riedi
Univ. Lille, CNRS, UMR 8518 Laboratoire d'Optique Atmosphérique (LOA), 59000 Lille, France
Chaabane Djeraba
Centre de Recherche en Informatique Signal et Automatique de Lille, Univ. Lille, CNRS, Centrale Lille, UMR 9189 CRIStAL, 59000 Lille, France
Pierre Nabat
Centre National de Recherches Météorologiques – Météo-France, Groupe de Météorologie Grande Echelle et Climat, 31057 Toulouse, France
Related authors
No articles found.
Bosi Sheng, Massimo A. Bollasina, Alexandre S. Gagnon, Laura J. Wilcox, Thomas P. S. Reynolds, Christopher T. S. Beckett, Haolin Wang, Qingxiang Li, Pierre Nabat, Robert J. Allen, Bjørn H. Samset, Joonas Merikanto, Geeta G. Persad, Toshihiko Takemura, Kostas Tsigaridis, Sharar Ahmadi, Maxwell T. Elling, Knut von Salzen, Daniel M. Westervelt, Naga Oshima, and Tsuyoshi Koshiro
EGUsphere, https://doi.org/10.5194/egusphere-2026-2198, https://doi.org/10.5194/egusphere-2026-2198, 2026
This preprint is open for discussion and under review for Atmospheric Chemistry and Physics (ACP).
Short summary
Short summary
Observational records indicate that austral summer rainfall over southern Africa experienced substantial wetting during the late 20th century, followed by pronounced drying over the past two decades. We show that East Asian sulphate aerosol emissions played a key role in shaping these rainfall trends. These findings have important implications for reducing uncertainties in regional rainfall projections and for understanding their links to ecosystem changes.
Athulya Saiprakash, Martina Krämer, Christian Rolf, Jérôme Riedi, and Odran Sourdeval
EGUsphere, https://doi.org/10.5194/egusphere-2026-2423, https://doi.org/10.5194/egusphere-2026-2423, 2026
This preprint is open for discussion and under review for Atmospheric Chemistry and Physics (ACP).
Short summary
Short summary
This second part of the study tested the new method by comparing its results with satellite and aircraft observations and by examining how sensitive it is to key assumptions. The method agreed well with observations in most cases, although clouds formed over mountains remained harder to capture. It also showed which conditions and processes matter most for reconstructing how cirrus form and evolve.
Athulya Saiprakash, Martina Krämer, Christian Rolf, Patrick Konjari, Jérôme Riedi, and Odran Sourdeval
EGUsphere, https://doi.org/10.5194/egusphere-2026-2416, https://doi.org/10.5194/egusphere-2026-2416, 2026
This preprint is open for discussion and under review for Atmospheric Chemistry and Physics (ACP).
Short summary
Short summary
To better understand how observed high ice clouds, known as cirrus, form and change, satellite images were combined with calculations that reconstruct how the air moved before the clouds formed. Case studies showed that cloud history differs with height, age, and surrounding conditions, and that small temperature changes strongly shape cloud growth. This helps reveal and characterize the mechanisms behind observed cirrus.
Léon Roussel, Marie Dumont, Marion Réveillet, Delphine Six, Marin Kneib, Pierre Nabat, Kévin Fourteau, Diego Monteiro, Simon Gascoin, Emmanuel Thibert, Antoine Rabatel, Jean-Emmanuel Sicart, Mylène Bonnefoy, Luc Piard, Olivier Laarman, Bruno Jourdain, Mathieu Fructus, Matthieu Vernay, and Matthieu Lafaysse
The Cryosphere, 19, 5201–5230, https://doi.org/10.5194/tc-19-5201-2025, https://doi.org/10.5194/tc-19-5201-2025, 2025
Short summary
Short summary
Saharan dust deposits frequently turn alpine glaciers orange. Mineral dust reduces snow albedo and increases snow and glaciers melt rate. Using physical modeling, we quantified the impact of dust on the Argentière Glacier over the period 2019–2022. We found that the contribution of mineral dust to the melt represents between 8 % and 16 % of Argentière Glacier summer melt. At specific locations, the impact of dust over one year can rise to an equivalent of 1.2 m of melted ice.
George Jordan, Florent Malavelle, Jim Haywood, Ying Chen, Ben Johnson, Daniel Partridge, Amy Peace, Eliza Duncan, Duncan Watson-Parris, David Neubauer, Anton Laakso, Martine Michou, and Pierre Nabat
Atmos. Chem. Phys., 25, 13393–13428, https://doi.org/10.5194/acp-25-13393-2025, https://doi.org/10.5194/acp-25-13393-2025, 2025
Short summary
Short summary
The 2014–15 Holuhraun eruption created a vast aerosol plume that acted as a natural experiment to assess how well climate models capture changes in cloud properties due to increased aerosol. We find that climate models represent the observed shift to smaller, more numerous cloud droplets well. However, climate models diverge in their aerosol-induced changes to large-scale cloud properties, particularly cloud liquid water content. Our study shows that Holuhraun had a cooling effect on the Earth.
Thomas Drugé, Pierre Nabat, Martine Michou, and Marc Mallet
Atmos. Chem. Phys., 25, 11651–11671, https://doi.org/10.5194/acp-25-11651-2025, https://doi.org/10.5194/acp-25-11651-2025, 2025
Short summary
Short summary
This atmospheric modelling study focuses on the radiative and climatic impacts of coarse-aerosol scattering in the long-wave range. The impact of this process on several climate variables is studied, showing, in particular, significant changes in long-wave radiation and daily surface minimum temperature. Other changes in atmospheric fields are also simulated. The findings of this study highlight the importance of incorporating aerosol scattering in the long-wave spectrum in climate models.
Martin Cussac, Martine Michou, Pierre Nabat, Béatrice Josse, and Sophie Pelletier
EGUsphere, https://doi.org/10.5194/egusphere-2025-1933, https://doi.org/10.5194/egusphere-2025-1933, 2025
Preprint archived
Short summary
Short summary
This study evaluates three chemistry schemes of varying complexity, one mainly stratospheric and two tropospheric-stratospheric, in the latest version of the climate model ARPEGE-Climat. Stratospheric ozone and water vapour are better represented. Despite issues with carbon monoxide in one scheme and with winter nitrogen species in the other, tropospheric ozone is overall realistically simulated. These modelling evolutions strengthen future research on chemistry-climate interactions.
Gabriel Chesnoiu, Isabelle Chiapello, Nicolas Ferlay, Pierre Nabat, Marc Mallet, and Véronique Riffault
Atmos. Chem. Phys., 25, 1307–1331, https://doi.org/10.5194/acp-25-1307-2025, https://doi.org/10.5194/acp-25-1307-2025, 2025
Short summary
Short summary
The ALADIN regional climate model at 12.5 km resolution allows us to study the evolution of surface solar radiation (SSR) and key associated atmospheric parameters. Over northern France and Benelux, influenced by anthropogenic aerosols and cloudy conditions, regional evaluation of recent hindcast simulations shows satisfying results and high spatial variability. Future SSR evolution by the end of the century for two contrasting CMIP6 scenarios highlights large decreases in SSR for SSP3-7.0.
Marc Mallet, Aurore Voldoire, Fabien Solmon, Pierre Nabat, Thomas Drugé, and Romain Roehrig
Atmos. Chem. Phys., 24, 12509–12535, https://doi.org/10.5194/acp-24-12509-2024, https://doi.org/10.5194/acp-24-12509-2024, 2024
Short summary
Short summary
This study investigates the interactions between smoke aerosols and climate in tropical Africa using a coupled ocean–atmosphere–aerosol climate model. The work shows that smoke plumes have a significant impact by increasing the low-cloud fraction, decreasing the ocean and continental surface temperature and reducing the precipitation of coastal western Africa. It also highlights the role of the ocean temperature response and its feedbacks for the September–November season.
Karlie N. Rees, Timothy J. Garrett, Thomas D. DeWitt, Corey Bois, Steven K. Krueger, and Jérôme C. Riedi
Nonlin. Processes Geophys., 31, 497–513, https://doi.org/10.5194/npg-31-497-2024, https://doi.org/10.5194/npg-31-497-2024, 2024
Short summary
Short summary
The shapes of clouds viewed from space reflect vertical and horizontal motions in the atmosphere. We theorize that, globally, cloud perimeter complexity is related to the dimension of turbulence also governed by horizontal and vertical motions. We find agreement between theory and observations from various satellites and a numerical model and, remarkably, that the theory applies globally using only basic planetary physical parameters from the smallest scales of turbulence to the planetary scale.
Alkiviadis Kalisoras, Aristeidis K. Georgoulias, Dimitris Akritidis, Robert J. Allen, Vaishali Naik, Chaincy Kuo, Sophie Szopa, Pierre Nabat, Dirk Olivié, Twan van Noije, Philippe Le Sager, David Neubauer, Naga Oshima, Jane Mulcahy, Larry W. Horowitz, and Prodromos Zanis
Atmos. Chem. Phys., 24, 7837–7872, https://doi.org/10.5194/acp-24-7837-2024, https://doi.org/10.5194/acp-24-7837-2024, 2024
Short summary
Short summary
Effective radiative forcing (ERF) is a metric for estimating how human activities and natural agents change the energy flow into and out of the Earth’s climate system. We investigate the anthropogenic aerosol ERF, and we estimate the contribution of individual processes to the total ERF using simulations from Earth system models within the Coupled Model Intercomparison Project Phase 6 (CMIP6). Our findings highlight that aerosol–cloud interactions drive ERF variability during the last 150 years.
Sarah Tinorua, Cyrielle Denjean, Pierre Nabat, Véronique Pont, Mathilde Arnaud, Thierry Bourrianne, Maria Dias Alves, and Eric Gardrat
Atmos. Meas. Tech., 17, 3897–3915, https://doi.org/10.5194/amt-17-3897-2024, https://doi.org/10.5194/amt-17-3897-2024, 2024
Short summary
Short summary
The three most widely used techniques for measuring black carbon (BC) have been deployed continuously for 2 years at a French high-altitude research station. Despite a similar temporal variation in the BC load, we found significant biases by up to a factor of 8 between the three instruments. This study raises questions about the relevance of using these instruments for specific background sites, as well as the processing of their data, which can vary according to the atmospheric conditions.
Fangxuan Ren, Jintai Lin, Chenghao Xu, Jamiu A. Adeniran, Jingxu Wang, Randall V. Martin, Aaron van Donkelaar, Melanie S. Hammer, Larry W. Horowitz, Steven T. Turnock, Naga Oshima, Jie Zhang, Susanne Bauer, Kostas Tsigaridis, Øyvind Seland, Pierre Nabat, David Neubauer, Gary Strand, Twan van Noije, Philippe Le Sager, and Toshihiko Takemura
Geosci. Model Dev., 17, 4821–4836, https://doi.org/10.5194/gmd-17-4821-2024, https://doi.org/10.5194/gmd-17-4821-2024, 2024
Short summary
Short summary
We evaluate the performance of 14 CMIP6 ESMs in simulating total PM2.5 and its 5 components over China during 2000–2014. PM2.5 and its components are underestimated in almost all models, except that black carbon (BC) and sulfate are overestimated in two models, respectively. The underestimation is the largest for organic carbon (OC) and the smallest for BC. Models reproduce the observed spatial pattern for OC, sulfate, nitrate and ammonium well, yet the agreement is poorer for BC.
Sarah Tinorua, Cyrielle Denjean, Pierre Nabat, Thierry Bourrianne, Véronique Pont, François Gheusi, and Emmanuel Leclerc
Atmos. Chem. Phys., 24, 1801–1824, https://doi.org/10.5194/acp-24-1801-2024, https://doi.org/10.5194/acp-24-1801-2024, 2024
Short summary
Short summary
At a French high-altitude site, where many complex interactions between black carbon (BC), radiation, clouds and snow impact climate, 2 years of refractive BC (rBC) and aerosol optical and microphysical measurements have been made. We observed strong seasonal rBC properties variations, with an enhanced absorption in summer compared to winter. The combination of rBC emission sources, transport pathways, atmospheric dynamics and chemical processes explains the rBC light absorption seasonality.
Xavier Ceamanos, Bruno Six, Suman Moparthy, Dominique Carrer, Adèle Georgeot, Josef Gasteiger, Jérôme Riedi, Jean-Luc Attié, Alexei Lyapustin, and Iosif Katsev
Atmos. Meas. Tech., 16, 2575–2599, https://doi.org/10.5194/amt-16-2575-2023, https://doi.org/10.5194/amt-16-2575-2023, 2023
Short summary
Short summary
A new algorithm to retrieve the diurnal evolution of aerosol optical depth over land and ocean from geostationary meteorological satellites is proposed and successfully evaluated with reference ground-based and satellite data. The high-temporal-resolution aerosol observations that are obtained from the EUMETSAT Meteosat Second Generation mission are unprecedented and open the door to studies that cannot be conducted with the once-a-day observations available from low-Earth-orbit satellites.
Huazhe Shang, Souichiro Hioki, Guillaume Penide, Céline Cornet, Husi Letu, and Jérôme Riedi
Atmos. Chem. Phys., 23, 2729–2746, https://doi.org/10.5194/acp-23-2729-2023, https://doi.org/10.5194/acp-23-2729-2023, 2023
Short summary
Short summary
We find that cloud profiles can be divided into four prominent patterns, and the frequency of these four patterns is related to intensities of cloud-top entrainment and precipitation. Based on these analyses, we further propose a cloud profile parameterization scheme allowing us to represent these patterns. Our results shed light on how to facilitate the representation of cloud profiles and how to link them to cloud entrainment or precipitating status in future remote-sensing applications.
Thomas Drugé, Pierre Nabat, Marc Mallet, Martine Michou, Samuel Rémy, and Oleg Dubovik
Atmos. Chem. Phys., 22, 12167–12205, https://doi.org/10.5194/acp-22-12167-2022, https://doi.org/10.5194/acp-22-12167-2022, 2022
Short summary
Short summary
This study presents the implementation of brown carbon in the atmospheric component of the CNRM global climate model and particularly in its aerosol scheme TACTIC. Several simulations were carried out with this climate model, over the period 2000–2014, to evaluate the model by comparison with different reference datasets (PARASOL-GRASP, OMI-OMAERUVd, MACv2, FMI_SAT, AERONET) and to analyze the brown carbon radiative and climatic effects.
Samuel Rémy, Zak Kipling, Vincent Huijnen, Johannes Flemming, Pierre Nabat, Martine Michou, Melanie Ades, Richard Engelen, and Vincent-Henri Peuch
Geosci. Model Dev., 15, 4881–4912, https://doi.org/10.5194/gmd-15-4881-2022, https://doi.org/10.5194/gmd-15-4881-2022, 2022
Short summary
Short summary
This article describes a new version of IFS-AER, the tropospheric aerosol scheme used to provide global aerosol products within the Copernicus Atmosphere Monitoring Service (CAMS) cycle. Several components of the model have been updated, such as the dynamical dust and sea salt aerosol emission schemes. New deposition schemes have also been incorporated but are not yet used operationally. This new version of IFS-AER has been evaluated and shown to have a greater skill than previous versions.
Simone Tilmes, Daniele Visioni, Andy Jones, James Haywood, Roland Séférian, Pierre Nabat, Olivier Boucher, Ewa Monica Bednarz, and Ulrike Niemeier
Atmos. Chem. Phys., 22, 4557–4579, https://doi.org/10.5194/acp-22-4557-2022, https://doi.org/10.5194/acp-22-4557-2022, 2022
Short summary
Short summary
This study assesses the impacts of climate interventions, using stratospheric sulfate aerosol and solar dimming on stratospheric ozone, based on three Earth system models with interactive stratospheric chemistry. The climate interventions have been applied to a high emission (baseline) scenario in order to reach global surface temperatures of a medium emission scenario. We find significant increases and decreases in total column ozone, depending on regions and seasons.
Andy Jones, Jim M. Haywood, Adam A. Scaife, Olivier Boucher, Matthew Henry, Ben Kravitz, Thibaut Lurton, Pierre Nabat, Ulrike Niemeier, Roland Séférian, Simone Tilmes, and Daniele Visioni
Atmos. Chem. Phys., 22, 2999–3016, https://doi.org/10.5194/acp-22-2999-2022, https://doi.org/10.5194/acp-22-2999-2022, 2022
Short summary
Short summary
Simulations by six Earth-system models of geoengineering by introducing sulfuric acid aerosols into the tropical stratosphere are compared. A robust impact on the northern wintertime North Atlantic Oscillation is found, exacerbating precipitation reduction over parts of southern Europe. In contrast, the models show no consistency with regard to impacts on the Quasi-Biennial Oscillation, although results do indicate a risk that the oscillation could become locked into a permanent westerly phase.
Sarah J. Doherty, Pablo E. Saide, Paquita Zuidema, Yohei Shinozuka, Gonzalo A. Ferrada, Hamish Gordon, Marc Mallet, Kerry Meyer, David Painemal, Steven G. Howell, Steffen Freitag, Amie Dobracki, James R. Podolske, Sharon P. Burton, Richard A. Ferrare, Calvin Howes, Pierre Nabat, Gregory R. Carmichael, Arlindo da Silva, Kristina Pistone, Ian Chang, Lan Gao, Robert Wood, and Jens Redemann
Atmos. Chem. Phys., 22, 1–46, https://doi.org/10.5194/acp-22-1-2022, https://doi.org/10.5194/acp-22-1-2022, 2022
Short summary
Short summary
Between July and October, biomass burning smoke is advected over the southeastern Atlantic Ocean, leading to climate forcing. Model calculations of forcing by this plume vary significantly in both magnitude and sign. This paper compares aerosol and cloud properties observed during three NASA ORACLES field campaigns to the same in four models. It quantifies modeled biases in properties key to aerosol direct radiative forcing and evaluates how these biases propagate to biases in forcing.
Ramiro Checa-Garcia, Yves Balkanski, Samuel Albani, Tommi Bergman, Ken Carslaw, Anne Cozic, Chris Dearden, Beatrice Marticorena, Martine Michou, Twan van Noije, Pierre Nabat, Fiona M. O'Connor, Dirk Olivié, Joseph M. Prospero, Philippe Le Sager, Michael Schulz, and Catherine Scott
Atmos. Chem. Phys., 21, 10295–10335, https://doi.org/10.5194/acp-21-10295-2021, https://doi.org/10.5194/acp-21-10295-2021, 2021
Short summary
Short summary
Thousands of tons of dust are emitted into the atmosphere every year, producing important impacts on the Earth system. However, current global climate models are not yet able to reproduce dust emissions, transport and depositions with the desirable accuracy. Our study analyses five different Earth system models to report aspects to be improved to reproduce better available observations, increase the consistency between models and therefore decrease the current uncertainties.
Daniele Visioni, Douglas G. MacMartin, Ben Kravitz, Olivier Boucher, Andy Jones, Thibaut Lurton, Michou Martine, Michael J. Mills, Pierre Nabat, Ulrike Niemeier, Roland Séférian, and Simone Tilmes
Atmos. Chem. Phys., 21, 10039–10063, https://doi.org/10.5194/acp-21-10039-2021, https://doi.org/10.5194/acp-21-10039-2021, 2021
Short summary
Short summary
A new set of simulations is used to investigate commonalities, differences and sources of uncertainty when simulating the injection of SO2 in the stratosphere in order to mitigate the effects of climate change (solar geoengineering). The models differ in how they simulate the aerosols and how they spread around the stratosphere, resulting in differences in projected regional impacts. Overall, however, the models agree that aerosols have the potential to mitigate the warming produced by GHGs.
Julien Jouanno, Rachid Benshila, Léo Berline, Antonin Soulié, Marie-Hélène Radenac, Guillaume Morvan, Frédéric Diaz, Julio Sheinbaum, Cristele Chevalier, Thierry Thibaut, Thomas Changeux, Frédéric Menard, Sarah Berthet, Olivier Aumont, Christian Ethé, Pierre Nabat, and Marc Mallet
Geosci. Model Dev., 14, 4069–4086, https://doi.org/10.5194/gmd-14-4069-2021, https://doi.org/10.5194/gmd-14-4069-2021, 2021
Short summary
Short summary
The tropical Atlantic has been facing a massive proliferation of Sargassum since 2011, with severe environmental and socioeconomic impacts. We developed a modeling framework based on the NEMO ocean model, which integrates transport by currents and waves, and physiology of Sargassum with varying internal nutrient quota, and considers stranding at the coast. Results demonstrate the ability of the model to reproduce and forecast the seasonal cycle and large-scale distribution of Sargassum biomass.
Josué Bock, Martine Michou, Pierre Nabat, Manabu Abe, Jane P. Mulcahy, Dirk J. L. Olivié, Jörg Schwinger, Parvadha Suntharalingam, Jerry Tjiputra, Marco van Hulten, Michio Watanabe, Andrew Yool, and Roland Séférian
Biogeosciences, 18, 3823–3860, https://doi.org/10.5194/bg-18-3823-2021, https://doi.org/10.5194/bg-18-3823-2021, 2021
Short summary
Short summary
In this study we analyse surface ocean dimethylsulfide (DMS) concentration and flux to the atmosphere from four CMIP6 Earth system models over the historical and ssp585 simulations.
Our analysis of contemporary (1980–2009) climatologies shows that models better reproduce observations in mid to high latitudes. The models disagree on the sign of the trend of the global DMS flux from 1980 onwards. The models agree on a positive trend of DMS over polar latitudes following sea-ice retreat dynamics.
Cited articles
Bach, S., Binder, A., Montavon, G., Klauschen, F., Müller, K.-R., and Samek, W.: On Pixel-Wise Explanations for Non-Linear Classifier Decisions by Layer-Wise Relevance Propagation, PLOS ONE, 10, e0130140, https://doi.org/10.1371/journal.pone.0130140, 2015. a
Beckh, K., Müller, S., Jakobs, M., Toborek, V., Tan, H., Fischer, R., Welke, P., Houben, S., and von Rueden, L.: Explainable Machine Learning with Prior Knowledge: An Overview, arXiv [preprint], https://doi.org/10.48550/arXiv.2105.10172, 2021. a, b, c
Belle, V. and Papantonis, I.: Principles and Practice of Explainable Machine Learning, Front Big Data, 4, 688969, https://doi.org/10.3389/fdata.2021.688969, 2021. a
Bose, S., Hansel, N., Tonorezos, E. S., Williams, D. L., Bilderback, A., Breysse, P. N., Diette, G. B., and McCormack, M. C.: Indoor Particulate Matter Associated with Systemic Inflammation in COPD, J. Environ. Protect., 6, 566–572, https://doi.org/10.4236/jep.2015.65051, 2015. a
Ceamanos, X., Six, B., and Riedi, J.: Quasi-Global Maps of Daily Aerosol Optical Depth From a Ring of Five Geostationary Meteorological Satellites Using AERUS-GEO, J. Geophys. Res.-Atmos., 126, e2021JD034906, https://doi.org/10.1029/2021JD034906, 2021. a
Chen, D., Gu, X., Guo, H., Cheng, T., Yang, J., Zhan, Y., and Fu, Q.: Spatiotemporally continuous PM2.5 dataset in the Mekong River Basin from 2015 to 2022 using a stacking model, Sci. Total Environ., 914, 169801, https://doi.org/10.1016/j.scitotenv.2023.169801, 2024. a
Chu, Y., Liu, Y., Li, X., Liu, Z., Lu, H., Lu, Y., Mao, Z., Chen, X., Li, N., Ren, M., Liu, F., Tian, L., Zhu, Z., and Xiang, H.: A Review on Predicting Ground PM2.5 Concentration Using Satellite Aerosol Optical Depth, Atmosphere, 7, 129, https://doi.org/10.3390/atmos7100129, 2016. a
Cohen, A. J., Brauer, M., Burnett, R., Anderson, H. R., Frostad, J., Estep, K., Balakrishnan, K., Brunekreef, B., Dandona, L., Dandona, R., Feigin, V., Freedman, G., Hubbell, B., Jobling, A., Kan, H., Knibbs, L., Liu, Y., Martin, R., Morawska, L., Pope, C. A., Shin, H., Straif, K., Shaddick, G., Thomas, M., van Dingenen, R., van Donkelaar, A., Vos, T., Murray, C. J. L., and Forouzanfar, M. H.: 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, 1907–1918, https://doi.org/10.1016/S0140-6736(17)30505-6, 2017. a
Confalonieri, R., Coba, L., Wagner, B., and Besold, T. R.: A historical perspective of explainable Artificial Intelligence, WIREs Data Min. Knowl., 11, e1391, https://doi.org/10.1002/widm.1391, 2021. a
Dabrowski, M.: Knowledge-inspired fusion strategies for the inference of PM2.5 values with a Neural Network – code for experiments (Initial), Zenodo [code], https://doi.org/10.5281/zenodo.13920070, 2024a. a
Dabrowski, M.: Knowledge-inspired fusion strategies for the inference of PM2.5 values with a Neural Network – CAMS data for experiments (Initial), Zenodo [data set], https://doi.org/10.5281/zenodo.13929498, 2024b. a
Dey, S., Purohit, B., Balyan, P., Dixit, K., Bali, K., Kumar, A., Imam, F., Chowdhury, S., Ganguly, D., Gargava, P., and Shukla, V. K.: A Satellite-Based High-Resolution (1-km) Ambient PM2.5 Database for India over Two Decades (2000–2019): Applications for Air Quality Management, Remote Sens., 12, 3872, https://doi.org/10.3390/rs12233872, 2020. a
Di, Q., Amini, H., Shi, L., Kloog, I., Silvern, R., Kelly, J., Sabath, M. B., Choirat, C., Koutrakis, P., Lyapustin, A., Wang, Y., Mickley, L. J., and Schwartz, J.: An ensemble-based model of PM2.5 concentration across the contiguous United States with high spatiotemporal resolution, Environ. Int., 130, 104909, https://doi.org/10.1016/j.envint.2019.104909, 2019. a
Engel-Cox, J. A., Hoff, R. M., and Haymet, A.: Recommendations on the Use of Satellite Remote-Sensing Data for Urban Air Quality, J. Air Waste Manage. Assoc., 54, 1360–1371, https://doi.org/10.1080/10473289.2004.10471005, 2004. a
Geng, G., Zhang, Q., Martin, R. V., van Donkelaar, A., Huo, H., Che, H., Lin, J., and He, K.: Estimating long-term PM2.5 concentrations in China using satellite-based aerosol optical depth and a chemical transport model, Remote Sens. Environ., 166, 262–270, https://doi.org/10.1016/j.rse.2015.05.016, 2015. a
Gilik, A., Ogrenci, A. S., and Ozmen, A.: Air quality prediction using CNN+LSTM-based hybrid deep learning architecture, Environ. Sci. Pollut. Res., 29, 11920–11938, https://doi.org/10.1007/s11356-021-16227-w, 2022. a
Goodfellow, I., Pouget-Abadie, J., Mirza, M., Xu, B., Warde-Farley, D., Ozair, S., Courville, A., and Bengio, Y.: Generative adversarial networks, Commun. ACM, 63, 139–144, https://doi.org/10.1145/3422622, 2020. a, b, c, d
Gratton, Y.: LE KRIGEAGE: LA MÉTHODE OPTIMALE D'INTERPOLATION SPATIALE, Les articles de l'institut d'analyse géographique, p. 5, https://www.researchgate.net/publication/229009107_Le_krigeage_la_methode_optimale_d'interpolation_spatiale (last access: 16 June 2025), 2002. a
Guo, W., Zhang, B., Wei, Q., Guo, Y., Yin, X., Li, F., Wang, L., and Wang, W.: Estimating ground-level PM2.5 concentrations using two-stage model in Beijing-Tianjin-Hebei, China, Atmos. Pollut. Res., 12, 101154, https://doi.org/10.1016/j.apr.2021.101154, 2021. a
Gupta, P. and Christopher, S. A.: Particulate matter air quality assessment using integrated surface, satellite, and meteorological products: Multiple regression approach, J. Geophys. Res.-Atmos., 114, D14205, https://doi.org/10.1029/2008JD011496, 2009. a, b
Ho, T. K.: Random Decision Forests, https://web.archive.org/web/20181105063147/http://ect.bell-labs.com/who/tkh/publications/papers/odt.pdf (last access: 16 June 2025), 2018. a
Hu, X., Waller, L. A., Lyapustin, A., Wang, Y., Al-Hamdan, M. Z., Crosson, W. L., Estes, M. G., Estes, S. M., Quattrochi, D. A., Puttaswamy, S. J., and Liu, Y.: Estimating ground-level PM2.5 concentrations in the Southeastern United States using MAIAC AOD retrievals and a two-stage model, Remote Sens. Environ., 140, 220–232, https://doi.org/10.1016/j.rse.2013.08.032, 2014. a
Irrgang, C., Boers, N., Sonnewald, M., Barnes, E. A., Kadow, C., Staneva, J., and Saynisch-Wagner, J.: Towards neural Earth system modelling by integrating artificial intelligence in Earth system science, Nature Machine Intelligence, 3, 667–674, https://doi.org/10.1038/s42256-021-00374-3, 2021. a
Isola, P., Zhu, J.-Y., Zhou, T., and Efros, A. A.: Image-to-Image Translation with Conditional Adversarial Networks, in: 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), IEEE, Honolulu, HI, 5967–5976, ISBN 978-1-5386-0457-1, https://doi.org/10.1109/CVPR.2017.632, 2017. a
Jin, C., Yuan, Q., Li, T., Wang, Y., and Zhang, L.: An optimized semi-empirical physical approach for satellite-based PM2.5 retrieval: embedding machine learning to simulate complex physical parameters, Geosci. Model Dev., 16, 4137–4154, https://doi.org/10.5194/gmd-16-4137-2023, 2023a. a
Jin, J., Henzing, B., and Segers, A.: How aerosol size matters in aerosol optical depth (AOD) assimilation and the optimization using the Ångström exponent, Atmos. Chem. Phys., 23, 1641–1660, https://doi.org/10.5194/acp-23-1641-2023, 2023b. a
Lary, D. J., Lary, T., and Sattler, B.: Using Machine Learning to Estimate Global PM2.5 for Environmental Health Studies, Environmental Health Insights, 9, 41–52, https://doi.org/10.4137/EHI.S15664, place: United States, 2015. a
Lee, S., Park, S., Lee, M.-I., Kim, G., Im, J., and Song, C.-K.: Air Quality Forecasts Improved by Combining Data Assimilation and Machine Learning With Satellite AOD, Geophys. Res. Lett., 49, e2021GL096066, https://doi.org/10.1029/2021GL096066, 2022. a
Li, J., Zhang, M., Xu, K., Dickerson, J., and Ba, J.: How does a Neural Network's Architecture Impact its Robustness to Noisy Labels?, in: Advances in Neural Information Processing Systems, edited by: Ranzato, M., Beygelzimer, A., Dauphin, Y., Liang, P., and Vaughan, J. W., Curran Associates, Inc., vol. 34, 9788–9803, https://proceedings.neurips.cc/paper/2021/hash/51311013e51adebc3c34d2cc591fefee-Abstract.html (last access: 16 June 2025), 2021. a
Li, T., Shen, H., Zeng, C., Yuan, Q., and Zhang, L.: Point-surface fusion of station measurements and satellite observations for mapping PM2.5 distribution in China: Methods and assessment, Atmos. Environ., 152, 477–489, https://doi.org/10.1016/j.atmosenv.2017.01.004, 2017. a
Lyu, B., Huang, R., Wang, X., Wang, W., and Hu, Y.: Deep-learning spatial principles from deterministic chemical transport models for chemical reanalysis: an application in China for PM2.5, Geosci. Model Dev., 15, 1583–1594, https://doi.org/10.5194/gmd-15-1583-2022, 2022. a
Ma, Z., Liu, R., Liu, Y., and Bi, J.: Effects of air pollution control policies on PM2.5 pollution improvement in China from 2005 to 2017: a satellite-based perspective, Atmos. Chem. Phys., 19, 6861–6877, https://doi.org/10.5194/acp-19-6861-2019, 2019. a
Ma, Z., Dey, S., Christopher, S., Liu, R., Bi, J., Balyan, P., and Liu, Y.: A review of statistical methods used for developing large-scale and long-term PM2.5 models from satellite data, Remote Sens. Environ., 269, 112827, https://doi.org/10.1016/j.rse.2021.112827, 2022. a, b, c
Madrigano Jaime, Kloog Itai, Goldberg Robert, Coull Brent A., Mittleman Murray A., and Schwartz Joel: Long-term Exposure to PM2.5 and Incidence of Acute Myocardial Infarction, Environ. Health Persp., 121, 192–196, https://doi.org/10.1289/ehp.1205284, 2013. a
Mallet, M. and Nabat, P.: CNRM-ALADIN64 – Regional climate simulation over the Euro-Mediterranean region, Aeris [data set], https://doi.org/10.25326/703, 2024. a
Mangai, U. G., Samanta, S., Das, S., and Chowdhury, P. R.: A Survey of Decision Fusion and Feature Fusion Strategies for Pattern Classification, IETE Technical Review, 27, 293–307, https://doi.org/10.4103/0256-4602.64604, 2010. a
Marais, W. J., Holz, R. E., Reid, J. S., and Willett, R. M.: Leveraging spatial textures, through machine learning, to identify aerosols and distinct cloud types from multispectral observations, Atmospheric Measurement Techniques, 13, 5459–5480, https://doi.org/10.5194/amt-13-5459-2020, 2020. a
Martin, R. V., Brauer, M., van Donkelaar, A., Shaddick, G., Narain, U., and Dey, S.: No one knows which city has the highest concentration of fine particulate matter, Atmos. Environ., 3, 100040, https://doi.org/10.1016/j.aeaoa.2019.100040, 2019. a
Matheron, G.: Principles of geostatistics, Econ. Geol., 58, 1246–1266, https://doi.org/10.2113/gsecongeo.58.8.1246, 1963. a, b, c
Mukai, S., Sano, I., Satoh, M., and Holben, B. N.: Aerosol properties and air pollutants over an urban area, Atmos. Res., 82, 643–651, https://doi.org/10.1016/j.atmosres.2006.02.020, 2006. a
Murdoch, W. J., Singh, C., Kumbier, K., Abbasi-Asl, R., and Yu, B.: Definitions, methods, and applications in interpretable machine learning, P. Natl. Acad. Sci. USA, 116, 22071–22080, https://doi.org/10.1073/pnas.1900654116, 2019. a
Neophytou, A. M., Costello, S., Brown, D. M., Picciotto, S., Noth, E. M., Hammond, S. K., Cullen, M. R., and Eisen, E. A.: Marginal Structural Models in Occupational Epidemiology: Application in a Study of Ischemic Heart Disease Incidence and PM2.5 in the US Aluminum Industry, Am. J. Epidemiol., 180, 608–615, https://doi.org/10.1093/aje/kwu175, 2014. a
Nie, Z., Lin, T., Jiang, H., and Kara, L. B.: TopologyGAN: Topology Optimization Using Generative Adversarial Networks Based on Physical Fields Over the Initial Domain, J. Mech. Design, 143, 031715, https://doi.org/10.1115/1.4049533, 2021. a
Park, S., Lee, J., Im, J., Song, C.-K., Choi, M., Kim, J., Lee, S., Park, R., Kim, S.-M., Yoon, J., Lee, D.-W., and Quackenbush, L. J.: Estimation of spatially continuous daytime particulate matter concentrations under all sky conditions through the synergistic use of satellite-based AOD and numerical models, Sci. Total Environ., 713, 136516, https://doi.org/10.1016/j.scitotenv.2020.136516, 2020. a
Raissi, M., Perdikaris, P., and Karniadakis, G.: Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations, J. Comput. Phys., 378, 686–707, https://doi.org/10.1016/j.jcp.2018.10.045, 2019. a
Reid, C. E., Jerrett, M., Petersen, M. L., Pfister, G. G., Morefield, P. E., Tager, I. B., Raffuse, S. M., and Balmes, J. R.: Spatiotemporal Prediction of Fine Particulate Matter During the 2008 Northern California Wildfires Using Machine Learning, Environ. Sci. Technol., 49, 3887–3896, https://doi.org/10.1021/es505846r, 2015. a
Reid, C. E., Considine, E. M., Maestas, M. M., and Li, G.: Daily PM2.5 concentration estimates by county, ZIP code, and census tract in 11 western states 2008–2018, Sci. Data, 8, 112, https://doi.org/10.1038/s41597-021-00891-1, 2021. a
Ronneberger, O., Fischer, P., and Brox, T.: U-Net: Convolutional Networks for Biomedical Image Segmentation, in: Medical Image Computing and Computer-Assisted Intervention – MICCAI 2015, edited by: Navab, N., Hornegger, J., Wells, W. M., and Frangi, A. F., Lecture Notes in Computer Science, Springer International Publishing, Cham, 234–241, ISBN 978-3-319-24574-4, https://doi.org/10.1007/978-3-319-24574-4_28, 2015. a, b
Roscher, R., Bohn, B., Duarte, M. F., and Garcke, J.: Explainable Machine Learning for Scientific Insights and Discoveries, IEEE Access, 8, 42200–42216, https://doi.org/10.1109/ACCESS.2020.2976199, 2020. a, b
Sinha, A., Chen, H., Danu, D., Kirubarajan, T., and Farooq, M.: Estimation and decision fusion: A survey, Neurocomputing, 71, 2650–2656, https://doi.org/10.1016/j.neucom.2007.06.016, 2008. a, b
Son, Y., Álvaro R. Osornio-Vargas, O'Neill, M. S., Hystad, P., Texcalac-Sangrador, J. L., Ohman-Strickland, P., Meng, Q., and Schwander, S.: Land use regression models to assess air pollution exposure in Mexico City using finer spatial and temporal input parameters, Sci. Total Environ., 639, 40–48, https://doi.org/10.1016/j.scitotenv.2018.05.144, 2018. a
Song, W., Li, S., Fang, L., and Lu, T.: Hyperspectral Image Classification With Deep Feature Fusion Network, IEEE T. Geosci. Remote, 56, 3173–3184, https://doi.org/10.1109/TGRS.2018.2794326, 2018. a
Su, Z., Lin, L., Chen, Y., and Hu, H.: Understanding the distribution and drivers of PM2.5 concentrations in the Yangtze River Delta from 2015 to 2020 using Random Forest Regression, Enviro. Monit. Assess., 194, 284, https://doi.org/10.1007/s10661-022-09934-5, 2022. a
Sun, L., Gao, H., Pan, S., and Wang, J.-X.: Surrogate modeling for fluid flows based on physics-constrained deep learning without simulation data, Comput. Meth. Appl. Mec. Eng., 361, 112732, https://doi.org/10.1016/j.cma.2019.112732, 2020. a
Sun, Q.-S., Zeng, S.-G., Liu, Y., Heng, P.-A., and Xia, D.-S.: A new method of feature fusion and its application in image recognition, Pattern Recogn., 38, 2437–2448, https://doi.org/10.1016/j.patcog.2004.12.013, 2005. a
Szegedy, C., Ioffe, S., and Vanhoucke, V.: Inception-v4, Inception-ResNet and the Impact of Residual Connections on Learning, arXiv [preprint], https://doi.org/10.48550/arXiv.1602.07261, 2016. a
Thanasutives, P., Numao, M., and Fukui, K.-I.: Adversarial Multi-task Learning Enhanced Physics-informed Neural Networks for Solving Partial Differential Equations, in: 2021 International Joint Conference on Neural Networks (IJCNN), 1–9, https://doi.org/10.1109/IJCNN52387.2021.9533606, , 2021. a
Unik, M., Sitanggang, I. S., Syaufina, L., and Jaya, I. N. S.: PM2.5 Estimation using Machine Learning Models and Satellite Data: A Literature Review, International Journal of Advanced Computer Science And Applications, 14, 5, https://doi.org/10.14569/IJACSA.2023.0140538, 2023. a, b, c, d
van Donkelaar, A., Martin, R. V., and Park, R. J.: Estimating ground-level PM2.5 using aerosol optical depth determined from satellite remote sensing, J. Geophys. Res.-Atmos., 111, D21201, https://doi.org/10.1029/2005JD006996, 2006. a, b
Wang, J. and Christopher, S. A.: Intercomparison between satellite-derived aerosol optical thickness and PM2.5 mass: Implications for air quality studies, Geophys. Res. Lett., 30, 2095, https://doi.org/10.1029/2003GL018174, 2003. a
Wang, W. and Lu, Y.: Analysis of the Mean Absolute Error (MAE) and the Root Mean Square Error (RMSE) in Assessing Rounding Model, IOP Conf. Ser. Mater. Sci. Eng., 324, 012049, https://doi.org/10.1088/1757-899X/324/1/012049, 2018. a
Wang, Y., Gonzalez-Garcia, A., van de Weijer, J., and Herranz, L.: SDIT: Scalable and Diverse Cross-domain Image Translation, in: Proceedings of the 27th ACM International Conference on Multimedia, MM '19, Association for Computing Machinery, New York, NY, USA, 1267–1276, ISBN 978-1-4503-6889-6, https://doi.org/10.1145/3343031.3351004, 2019. a
Wang, Y., Khan, S., Gonzalez-Garcia, A., van de Weijer, J., and Khan, F. S.: Semi-Supervised Learning for Few-Shot Image-to-Image Translation, in: 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), IEEE, Seattle, WA, USA, 4452–4461, ISBN 978-1-72817-168-5, https://doi.org/10.1109/CVPR42600.2020.00451, 2020. a
Wu, J., Yao, F., Li, W., and Si, M.: VIIRS-based remote sensing estimation of ground-level PM2.5 concentrations in Beijing–Tianjin–Hebei: A spatiotemporal statistical model, Remote Sens. Environ., 184, 316–328, https://doi.org/10.1016/j.rse.2016.07.015, 2016. a
Xiao, Y., Wang, Y., Yuan, Q., He, J., and Zhang, L.: Generating a long-term (2003-−2020) hourly 0.25° global PM2.5 dataset via spatiotemporal downscaling of CAMS with deep learning (DeepCAMS), Sci. Total Environ., 848, 157 747, https://doi.org/10.1016/j.scitotenv.2022.157747, 2022. a
Xin, J., Zhang, Q., Wang, L., Gong, C., Wang, Y., Liu, Z., and Gao, W.: The empirical relationship between the PM2.5 concentration and aerosol optical depth over the background of North China from 2009 to 2011, Atmos. Res., 138, 179–188, https://doi.org/10.1016/j.atmosres.2013.11.001, 2014. a
Yang, L., Treichler, S., Kurth, T., Fischer, K., Barajas-Solano, D., Romero, J., Churavy, V., Tartakovsky, A., Houston, M., Prabhat, M., and Karniadakis, G.: Highly-scalable, Physics-Informed GANs for Learning Solutions of Stochastic PDEs, in: 2019 IEEE/ACM Third Workshop on Deep Learning on Supercomputers (DLS), 1–11, https://doi.org/10.1109/DLS49591.2019.00006, 2019. a
Yang, L., Zhang, D., and Karniadakis, G. E.: Physics-Informed Generative Adversarial Networks for Stochastic Differential Equations, SIAM J. Sci. Comput., 42, A292–A317, https://doi.org/10.1137/18M1225409, 2020. a
Zhang, L., Zhang, L., Mou, X., and Zhang, D.: FSIM: A Feature Similarity Index for Image Quality Assessment, IEEE T. Image Process., 20, 2378–2386, https://doi.org/10.1109/TIP.2011.2109730, 2011. a, b, c
Zhu, J.-Y., Park, T., Isola, P., and Efros, A. A.: Unpaired Image-to-Image Translation Using Cycle-Consistent Adversarial Networks, in: 2017 IEEE International Conference on Computer Vision (ICCV), IEEE, Venice, 2242–2251, ISBN 978-1-5386-1032-9, https://doi.org/10.1109/ICCV.2017.244, 2017a. a
Zhu, J.-Y., Zhang, R., Pathak, D., Darrell, T., Efros, A. A., Wang, O., and Shechtman, E.: Toward Multimodal Image-to-Image Translation, in: Advances in Neural Information Processing Systems, vol. 30, Curran Associates, Inc., https://proceedings.neurips.cc/paper/2017/hash/819f46e52c25763a55cc642422644317-Abstract.html (last access: 16 June 2025), 2017b. a
Short summary
This work focuses on the prediction of aerosol concentration values at the ground level, which are a strong indicator of air quality, using artificial neural networks. A study of different variables and their efficiency as inputs for these models is also proposed and reveals that the best results are obtained when using all of them. Comparison between network architectures and information fusion methods allows for the extraction of knowledge on the most efficient methods in the context of this study.
This work focuses on the prediction of aerosol concentration values at the ground level, which...
