Preprints
https://doi.org/10.5194/egusphere-2026-4322
© Author(s) 2026. This work is distributed under
the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
https://doi.org/10.5194/egusphere-2026-4322
© Author(s) 2026. This work is distributed under
the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
17 Aug 2026
| 17 Aug 2026
Status: this preprint is open for discussion and under review for Atmospheric Chemistry and Physics (ACP).
A multi-model analysis of the transatlantic dust transport, deposition, and optical properties in the AeroCom Phase III
Abstract. Transatlantic dust from North Africa is the largest intercontinental aerosol flux on Earth, affecting radiative forcing, cloud processes, and biogeochemical cycling, yet the inter-model spread of dust transport remains large in previous AeroCom experiments. With the progress in 17 AeroCom Phase III (P3) models, this study evaluates the transatlantic dust cycle for 2010 against AERONET inversions, satellite products (MODIS, CALIOP, MISR, IASI, POLDER, and MODIS-TIR), DustCOMM, CAMS, MERRA-2, in-situ records, and GPCP precipitation across subregions from North Africa to the Caribbean. We found that AeroCom P3 narrows mid-visible dust optical depth (DOD) diversity to 30–45 % across the transport domain relative to AeroCom P2 (38–59 %), yet this convergence masks compensatory biases, as models with lower (or higher) mass loading may have higher (or lower) mass extinction efficiency (MEE), yielding similar DOD despite a large spread of mass representations. Compared with observational estimates, model biases intensify with transport distance, with median DOD underestimated by ~50 % and ~70 % over mid- and long-range regions, and loss frequencies exceed satellite estimates by a factor of 2–3. Such biases remain unresolved even with super-coarse extensions, indicating that constraining mid-visible DOD alone is inadequate for the dust cycle. The excess removal implicates deficiencies in gravitational settling, wet scavenging, and vertical transport parameterizations. This systematic underestimation could propagate into dust radiative forcing, cloud-dust interactions, and nutrient delivery to Atlantic and Amazonian ecosystems. Progress requires size-resolved benchmarking (DOD at 550 nm and 10 µm with size-segregated mass), non-spherical settling treatments, and mechanisms sustaining coarse particles aloft.
How to cite. Zheng, J., Yu, H., Rajapakse, C., Chin, M., Schulz, M., Balkanski, Y., Bauer, S. E., Bian, H., Colarco, P., Kim, D., Checa-Garcia, R., Derimian, Y., Dubovik, O., Ginoux, P., Kipling, Z., Kok, J. F., Kokkola, H., Le Sager, P., Li, L., Lund, M. T., Myhre, G., Matsui, H., Olivie, D. J. L., Remy, S., Seland, O., Song, Q., Takemura, T., Tan, Q., Tsigaridis, K., Noije, T. V., Winker, D., Xie, B., Zhang, H., Zhang, X., and Zhang, Z.: A multi-model analysis of the transatlantic dust transport, deposition, and optical properties in the AeroCom Phase III, EGUsphere [preprint], https://doi.org/10.5194/egusphere-2026-4322, 2026.
Competing interests: At least one of the (co-)authors is a member of the editorial board of Atmospheric Chemistry and Physics.
Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.Download & links
- Preprint
(10185 KB) - Metadata XML
-
Supplement
(12608 KB) - BibTeX
- EndNote
Status: open (extended)
Comment types: AC – author | RC – referee | CC – community | EC – editor | CEC – chief editor
| : Report abuse
- RC1: 'Such an important topic requires a more in-depth presentation', Anonymous Referee #1, 28 Sep 2026 reply
Viewed
Total article views: 452 (including HTML, PDF, and XML)
Cumulative views and downloads
(calculated since 17 Aug 2026)
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 251 | 125 | 76 | 452 | 72 | 95 | 87 |
- HTML: 251
- PDF: 125
- XML: 76
- Total: 452
- Supplement: 72
- BibTeX: 95
- EndNote: 87
Viewed (geographical distribution)
Total article views: 366 (including HTML, PDF, and XML)
Thereof 366 with geography defined
and 0 with unknown origin.
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
1
Latest update: 16 Oct 2026
GESTAR-II, University of Maryland at Baltimore County, Baltimore, Maryland, USA
Earth Sciences Division, NASA Goddard Space Flight Center, Greenbelt, Maryland, USA
Hongbin Yu
Earth Sciences Division, NASA Goddard Space Flight Center, Greenbelt, Maryland, USA
Chamara Rajapakse
Earth Sciences Division, NASA Goddard Space Flight Center, Greenbelt, Maryland, USA
Science Systems and Application Inc., Lanham, Maryland, USA
Mian Chin
GESTAR-II, University of Maryland at Baltimore County, Baltimore, Maryland, USA
Earth Sciences Division, NASA Goddard Space Flight Center, Greenbelt, Maryland, USA
Michael Schulz
Norwegian Meteorological Institute, Oslo, Norway
Yves Balkanski
Laboratoire des Sciences du Climat et de l’Environnement, LSCE/IPSL, CEA-CNRS-UVSQ, UPSaclay, Gif-sur-Yvette, France
Susanne E. Bauer
NASA Goddard Institute for Space Studies New York City, New York, USA
Huisheng Bian
GESTAR-II, University of Maryland at Baltimore County, Baltimore, Maryland, USA
Earth Sciences Division, NASA Goddard Space Flight Center, Greenbelt, Maryland, USA
Peter Colarco
Earth Sciences Division, NASA Goddard Space Flight Center, Greenbelt, Maryland, USA
Dongchul Kim
GESTAR-II, University of Maryland at Baltimore County, Baltimore, Maryland, USA
Earth Sciences Division, NASA Goddard Space Flight Center, Greenbelt, Maryland, USA
Ramiro Checa-Garcia
Laboratoire des Sciences du Climat et de l’Environnement, LSCE/IPSL, CEA-CNRS-UVSQ, UPSaclay, Gif-sur-Yvette, France
Yevgeny Derimian
Laboratoire d'optique Atmosphèrique (LOA), Universite Lille, Villeneuve d ascq, France
Oleg Dubovik
Laboratoire d'optique Atmosphèrique (LOA), Universite Lille, Villeneuve d ascq, France
Paul Ginoux
NOAA Geophysical Fluid Dynamics Laboratory, Princeton, New Jersey, USA
Zak Kipling
European Centre for Medium-Range Weather Forecasts, Reading, UK
Jasper F. Kok
University of California Los Angeles, Los Angeles, California, USA
Harri Kokkola
Department of Technical Physics, University of Eastern Finland, Kuopio, Finland
Atmospheric Research Centre of Eastern Finland, Finnish Meteorological Institute, Kuopio, Finland
Philippe Le Sager
Royal Netherlands Meteorological Institute, De Bilt, the Netherlands
Department of Atmospheric and Oceanic Sciences & Institute of Atmospheric Sciences, Fudan University, Shanghai, China
Marianne Tronstad Lund
CICERO Center for International Climate Research, Oslo, Norway
Gunnar Myhre
CICERO Center for International Climate Research, Oslo, Norway
Hitoshi Matsui
Institute for Space–Earth Environmental Research, Nagoya University, Nagoya, Japan
Dirk J. L. Olivie
Norwegian Meteorological Institute, Oslo, Norway
Samuel Remy
HYGEOS, Lille, France
Oyvind Seland
Norwegian Meteorological Institute, Oslo, Norway
Qianqian Song
GESTAR-II, University of Maryland at Baltimore County, Baltimore, Maryland, USA
Toshihiko Takemura
Research Institute for Applied Mechanics, Kyushu University, 6-1 Kasuga-koen, Kasuga, Fukuoka, Japan
Qian Tan
Bay Area Environmental Research Institute, Moffett Field, California, USA
NASA Ames Research Center, Moffett Field, California, USA
Kostas Tsigaridis
NASA Goddard Institute for Space Studies New York City, New York, USA
Center for Climate Systems Research, Columbia University, New York City, NY, USA
Twan van Noije
Atmospheric Research Centre of Eastern Finland, Finnish Meteorological Institute, Kuopio, Finland
David Winker
NASA Langley Research Center, Hampton, Virginia, USA
Bin Xie
Chinese Academy of Meteorological Sciences, Beijing, China
Hua Zhang
Chinese Academy of Meteorological Sciences, Beijing, China
Xindan Zhang
Department of Atmospheric and Oceanic Sciences & Institute of Atmospheric Sciences, Fudan University, Shanghai, China
Zhibo Zhang
GESTAR-II, University of Maryland at Baltimore County, Baltimore, Maryland, USA
Short summary
Winds carry vast amounts of Saharan dust across the Atlantic Ocean, influencing climate, air quality, and ocean ecosystems. Within the Aerosol Comparisons between Observations and Models (AeroCom) project, we evaluated 17 numerical models against observations. Models deposit dust too quickly, so too little reaches the Caribbean. Better simulations require emitting larger particles, treating their non-spherical shapes, and capturing processes that keep dust aloft during transport.
Winds carry vast amounts of Saharan dust across the Atlantic Ocean, influencing climate, air...
The multi-model analysis of transatlantic dust transport by Zheng et al examines the performance of Aerocom III models against a number of reference datasets: satellite observations, reanalysis and inverse modelling products, and in-situ records. The main conclusion is that Aerocom III narrows the variability ("diversity") compared to Aerocom II. However, the authors warn us of compensatory biases as when new particle sizes are introduced, this reflects to an addapted mass extinction efficiency. Moreover, the authors notice an increased particle loss in the models downwind of the source regions, compared to observations (underestimate by 50-70%). The biases remain unexplained despite extended representations able to include super-coarse dust.
This topic is quite significant and of high interest, because it touches on one of the most debated questions around atmospheric dust: "the mysterious long-range transport of giant mineral dust particles" (https://doi.org/10.1126/sciadv.aau2768). I feel however that the paper is not yet mature and is at times too technical in tone and not completely connecting the dots between its own statements. The paper is a big collection of figures (14 in the main text and 13 in the supplement) that are not always described in full in the text. Some methodological points are not convincing. All in all I do not feel that the paper is ready for a publication in ACP and I feel that it could benefit from a radical review in the way the narrative is built and in the selection of materials displayed. I encourage a thourough critical revision of the material presented and the re-writing of the paper in the optics of an improved narrative (not mere technical corrections).
MAJOR POINTS:
1) The main conclusion of the paper is that there is an improvement (reduced diversity) in Aerocom III against Aerocom II (see abstract). However there is not a single figure in the paper displaying this comparison and the text discussing this is limited.
2) The authors note that a more accurate microphysical representation (inclusion of super-coarse particles) does not improve the distance-dependent bias (end of page 11). On the other hand, is it not expected that adding supercoarse particles would cause particles to sediment more quickly, if a process countering gravitational settling is not also added? The way the narrative is construed suggests that this is a surprise to the authors, whereas it is expected.
3) Some of the results presented are strikingly similar to those of another paper that is in open discussion at the same time in ACP (Ratcliffe et al, https://doi.org/10.5194/egusphere-2025-5143). I would encourage the authors to discuss the similarities and differences between the two papers.
4) The vertical distribution of dust in models in the Eastern Atlantic had been studied in O'Sullivan et al (https://doi.org/10.5194/acp-20-12955-2020), which noticed dust removal to be too fast and ascribed issues to compensations in the model between the fine and the coarse mode. The authors went further to propose an augmentation of the set of aerosol observations available on a global scale for constraining models, with a better focus on the vertical distribution and on the particle size distribution. I feel that such results, albeit six years old, could tie well with some of the points raised, and I would encourage the authors into a more general discussion of modelling limitations in the context of the existing literature. For example, at lines 700-710 the authors notice how the use of AOD as a model constrain can alter the model behaviour, which was commented in that paper.
5) How have the Aerocom III models been constrained? Is there any data assimilation for example, or have any parameters been tuned to match the AERONET AOD? Please add a column in table 1 detailing how models were constrained to representing the observations. Understanding this is key to some of the discussions in the paper itself, e.g. around different quantities compensating when a model is modified (e.g. DOD vs mass loading).
6) Section 3.2.2. The remote sensing retrievals from GRASP, AERONET, POLDER and MODIS will generally be stated as optical products (AOD) and to compute mass loadings a conversion is needed. Like for a model, wee need to assume a MEE. If the models are unable to constrain this quantity, then it should be clarified that also the remote sensing methods may have uncertainties in the same respect. The comparisons of mass loadings could therefore be affected by the same issue that they want to resolve. I would therefore be much more careful here!
7) Following up on this, the authors advocate at line 785 an improvement in remote sensing retrievals. The issue here is that remote sensing only offers an indirect estimate of the particle size, which is needed to compute the MEE. The issue may not be in the retrievals but in the information content of optical observations. A nice study on the MEE is in the paper by Papetta et al (https://doi.org/10.5194/acp-26-2055-2026), and I suggest that you incorporate the concerns of the remote sensing community around this issue within your paper's narrative.
MINOR POINTS:
8) Table 1: only 3 models in group 2 (ECMWF-CY45, GISS-MATRIX and GISS-OMA) have a representation for particles larger than 30 micron. It is therefore an oversimplification that group 2 represents an uper limit of 40 um (line 146). I suggest to improve the description of this group.
9) Why is 2010 chosen as the year of the simulations. Is it linked to data availability? Or is this year particular for any reason?
10) Line 165: transatlantic domain average. Define the "transatlantic domain" since the only 4 domains defined are the ones in figure 1.
11) Line 167: regridding to a 1x1 degree resolution. Given that most of the models have a resolution of 2-3 degrees (see table 1), is there not a risk of oversampling here? Please explain the consequences (including statistical ones) of this regridding.
12) Line 170: the exclusion of group 3 models is unclear. From what are they excluded exactly? They are listed in table S1, for example. At line 180, we read that means and medians are calculated "from the available models", so are group 3 also included? Similarly, in figure 1a are group 3 models included? The caption does not specify. More clarity is generally needed around this point. There are many other instances in the paper where this question may arise.
13) Surface dust concentrations: why are Barbados and Miami selected as locations for this comparison? Please discuss also if the selection of the locations can introduce a bias in the conclusions. In particular, why are there no locations in the Eastern side of the Atlantic? and is Miami not a bit too far North to be representative for average transatlantic dust path?
14) AERONET sites (section 3.2): the analysis oscillates between using 5 sites (figure 3) and 19 (line 271). The reason for selecting these sites should be made explicit. Also, indicate (even qualitatively) how far each site is from the sources, so as to be able to visualise statements such as the one at line 261. Also note that figure 3 shows that the AVERAGE model DOD is progressively underestimated with transport distance, but that this is not true for individual model representations (some models are well above the observations): please clarify this point better.
15) I do not understand why MEE cannot be meaningfully averaged due to the dependence on the PSD. This statement could be revisited (line 338).
16) Line 444: why only 10 models?
17) Line 458: is the MBL not typically at 0-1 km altiude? why do you place it at 1-3 km altitude?
18) Line 731: the mention of particle shape comes here a bit as a suprise. Perhaps the authors should discuss the internal representation of dust in the models in section 2, instead.
19) Figures in the supplement are only briefly called in the paper without any discussion of their content (see e.g. lines 212, 299, 432, etc.). I would therefore say that these figures are not needed in the paper and could be removed altogether.
20) The supplementary material is not always made clear. I struggle understanding the descriptions S1-S5: the text not being well written and full of undefined mathematical symbols.