the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Characteristics of desert dust optical and physical properties from AERONET: spectral absorption, size distributions and spectral AOD
Abstract. The AERONET multi-decade global database is examined to investigate long-term averages of retrieved physical and optical properties and spectral aerosol optical depth (AOD) of significant events at sites in major global dust source and receptor regions. After examples are shown of retrieved particle size and properties such as absorption, attention is paid to average spectral single scattering albedo (SSA) with comparisons to the laboratory measurements of dust from soil samples by Di Biagio et al. (2019). Overall, the AERONET retrieved and laboratory SSA agree very well for specific dust source regions in the Sahara and Sahel, with some regions showing annual cycles at 440 nm, an indicator of changes in iron oxide content. The strongest dust absorption was measured at an Australia desert site, and the least absorption was in dust transported to the Caribbean from the Sahara/Sahel regions. Also notable is regional and seasonal variability in the prevalence of ~0.3 to 1 μm radius dust, especially in the Sahel and Sahara. Such middle mode dust can often account for ~40 % of AOD at 440 nm. These ‘middle mode’ sized particles are a probable significant contributor for the observed relatively large spectral drop-off in AOD in the wavelengths beyond the visible, at 1020 nm and especially 1640 nm, in most dust events where Angstrom Exponent is low. Increased understanding of absorption, size distributions and AOD spectra of dust aerosol may help in improving remote sensing retrievals plus in more accurately characterizing the radiative effects of desert dust.
- Preprint
(16934 KB) - Metadata XML
- BibTeX
- EndNote
Status: open (extended)
- RC1: 'Comment on egusphere-2026-5531', Anonymous Referee #1, 28 Oct 2026 reply
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 602 | 338 | 73 | 1,013 | 79 | 82 |
- HTML: 602
- PDF: 338
- XML: 73
- Total: 1,013
- BibTeX: 79
- EndNote: 82
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
General impression
The manuscript deals with the latest status of the AERONET retrieval capabilities and an update of the retrieval products with focus on desert dust. The authors state: The AERONET dataset is now large enough to revisit key aspects of retrieval dust properties, and their relative modalities in regard to location and seasonality.
A long, deep, and dense study designed and written by a team of professional scientists is now available. The first author is one of the AERONET pioneers: Thomas Eck! ACP can be congratulated that the authors submitted this milestone-like work to this journal.
Here is my list of questions and remarks, all in all pointing to minor revisions.
Detailed comments:
Line 220: Short names (including field campaign names) need to be explained at their first appearance, in this case INDOEX.
Figure 1 is nice! However, you consider many more stations! I would recommend to present a Table (station name, longitude, latitude, altitude) in which all these stations are listed.
Lines 234-262: The discussion should include also ‘critical’ field studies, SAMUM 1, SAMUM 2, SALTRACE articles. Check Weinzierl et al. papers, maybe start with the BAMS 2017 article, and check the references for the Weinzierl SAMUM articles in the BAMS contribution. Furthermore, check the papers of Claire Ryder et al. (e.g., Fennec 2011campaign and other activities). Check Papetta et al. (ACP 2026) as well. In all these studies, it was shown that the fraction of airborne dust particles with diameters larger than 30 µm can be significant and the AERONET retrieval algorithm may then have severe problems, associated with large uncertainties.
Ryder et al., Coarse and giant particles are ubiquitous in Saharan dust export regions and are radiatively significant over the Sahara, Atmospheric Chemistry and Physics, doi: 10.5194/acp-19-15353-2019 (there are further ACP papers 2013, 2015)
Ryder et al., Coarse mode mineral dust size distributions, composition and optical properties from AER-D aircraft measurements over the Tropical Eastern Atlantic. Atmospheric Chemistry and Physics, doi: 10.5194/acp-18-17225-2018
Papetta et al., Volume-to-extinction ratio: an important property of dust, Atmos. Chem. Phys., 26, 2055–2082, doi: 10.5194/acp-26-2055-2026
Weinzierl et al., The Saharan Aerosol Long-Range Transport and Aerosol–Cloud-Interaction Experiment: Overview and Selected Highlights, Bull. Amer. Meteor. Soc., 98, 1427–1451, 2017, doi: 10.1175/BAMS-D-15-00142.1
Line 256: Is DABEX somewhere explained?
Line 273: If only cases with AOD(440) from 0.7 to 3.5 are considered, does that not introduce a bias? …towards strong dust outbreaks?
Lines 288-294: Here again, a ‘more robust’ discussion is needed. AERONET is powerful enough to even discuss shortcomings as they simply exist. Do not make the problem smaller as it is. I do not tell new stuff: One has to be careful, when AERONET observations are performed in or close to desert dust source regions. The giant particle fraction can be considerable and the related uncertainties in the AERONET retrieval products can be large.
Figure 2: A good overview! However, if AOD(440) is 3.5 then the fraction of dust particles with diameters exceeding 30 µm is probably significant. The ‘real-world’ size distributions in Figure 19 (and also in the Weinzierl SAMUM papers and in the Ryder Fennec papers) look rather different for the coarse mode size range than the ‘artificial’ AERONET size distribution in Fig 2a. My question is: These giant dust particles (larger than 15 µm in radius) produce an optical effect. How does the inversion algorithm handle these contributions to the observed AOD spectrum and sky radiances?
Line 305: A ‘middle’ mode is introduced. And I started again to ask myself: Is that an ‘artificial’ mode? You have (a) this 15µm size limit issue, (b) the dust particles are assumed to be spheroidal, but they are not spheroidal in shape, they are simply irregularly shaped (the spheroidal shape assumption causes large uncertainties in lidar parameters, lidar ratio, depol ratio, see Haarig et al. 2022), (c) the exact chemical composition of the dust particles is not known (and thus the spectral complex refractive index characteristics), (d) the dust particles may even have hygroscopic material on their surfaces leading potentially to water uptake effects (all this influences the refractive index properties), and (e) there are probably also instrumental problems (e.g., channel saturation effects) when AOD observations (for cases with AOD exceeding 2.0) are performed at low sun elevation angles so that the line-of-sight AOD may exceed 5-10. There are probably further sensitive retrieval-related uncertainty sources…. One needs to mention all this, is my feeling! The present set of results is excellent, no doubt, but one needs to discuss the numerous uncertainty sources and their potential impact as well.
Haarig et al., First triple-wavelength lidar observations of depolarization and extinction-to-backscatter ratios of Saharan dust, Atmos. Chem. Phys., 22, 355–369, doi: 10.5194/acp-22-355-2022
Line 325: I found several times (not only in line 325) …2.5 mm instead of 2.5 µm.
Line 370: Now, many new stations are introduced that are not included in Figure 1! In Figure 3, four new stations are introduced. We need a Table with a full list of all stations used in this manuscript.
Figure 3 (and also in other figures): It is not easy to distinguish the different observations although we have different colors, dotted, dashed, and solid lines, and even symbols… Any idea how to improve this?
Line 417: 440 nm SSA comparison, Barbados vs Canary Island. There are so many reasons for differences….. first of all, dust source regions are different. I would not discuss small differences in such large detail.
Line 460: Three modes in the size distribution… Many airborne in situ observations just show ONE mode, i.e., a pronounced coarse mode with a left wing (small particles). That’s it! This left wing explains the FMF. Even in the case of pure dust outbreaks there is always a FMF with values of 25-35%. Mamouri and Ansmann (ACP, 2014) already used this feature to interpret lidar observations of dust depolarization ratios.
Mamouri, R. E. and Ansmann, A.: Fine and coarse dust separation with polarization lidar, Atmos. Meas. Tech., 7, 3717–3735, https://doi.org/10.5194/amt-7-3717-2014, 2014.
Figure 5: The message obviously is: pure dust scenarios are observed when AE < 0.24 ? This is at least found for Tamanrasset.
Line 491-493: Even if the city center of Dakar would be 100 to 150 km away from the AERONET station there is always a constant, non-negligible pollution contribution to AOD in western central Africa. And my opinion is supported by Figure 5.
Figure 6: Why is this figure shown?
Line 524: Can an AOD of 7.1 be accurately retrieved? The dynamical signal range of the AERONET channels must be able cover several orders of magnitude, especially when the observations were taken at low elevation angles.
I was expecting that the presentation of products starts with the size distributions, followed by a discussion of complex refractive indices, and after that a discussion on SSA. Ok, You prefer to start with the SSA, because it is the ‘most attracting’ product, right? Maybe, provide a short motivating paragraph, why you start with SSA, followed by refractive index, etc..
The paper is very long, do we need Figure 13?
In Sect. 3.2.4 volume size distributions are discussed! I gave my comments to the retrieval and the uncertainty sources already above.
Figure 17 shows size distributions. The middle mode is highlighted and visible even in the Tamanrassed size distribution. I notuiced that the coarse-mode part of the size distribution for Tamanrasset shows the lowest coarse-mode values for radii larger than about 5µm. Why is that? An artifact? And the two further modes (fine mode, middle mode) are trustworthy?
Line 818-820: Snapshot-like airborne in situ observation for given specific height levels are compared with column-integrated retrieval products. This has never been a convincing approach to me. Especially not if the column is filled with a varying pollution-dust mixture in the boundary layer and, in addition, with desert dust after long-range transport in the free troposphere.
Figure 18 a,b,c: All shown coarse-mode shapes are so similar, and ‘nature’ is usually rather variable. It is obvious that these so similar coarse-mode shapes do not just reflect reality.
Lines 833 – 839: Middle mode discussion. For me, this discussion is hypothetically, and not based on ‘observed’ facts in view of all the assumption and constraints applied in the retrieval of the size distribution. Ok, maybe I am wrong. You think, the middle mode is a clear fact?
Figure 19 is in line with all the work of Weinzierl et al. and Ryder et al. I appreciate that such a figure is given, but is presented quite late. The message of Fig. 19 is: The results in Figs. 17, 18, 20, and 21 must be handled with care.
In the summary and conclusions section, 7 key messages (major findings) are listed, the points 1-4 deal with SSA, and the points 5-7 mainly deal with the middle mode.
Also some open problems could be mentioned and AERONET future plans, if there are any.
Is there any plan to check (within field studies) to what extent dust particles with radius larger than 15 µm contribute to the coarse mode fraction (number concentration, mass concentration). One could organize field studies, e.g., in Tamanrasset and Dakar, with a cloud radar at the AERONET photometer station. Cloud radars are sensitive to giant particles with radius of 20-100 µm…. and may allow to quantify/estimate the contribution of giant particles to the to dust (volume) mass concentration.