| 19 Aug 2026
Aerosol variability in the Río de la Plata estuary: Multimodal analysis of four years of AERONET observations
Abstract. We analyze four years (2020–2023) of AERONET observations from Montevideo_FING, an urban-estuarine site at the Río de la Plata, to characterize aerosol optical properties and to assess whether inversion-retrieved particle-volume size distributions (PVSDs) are better represented by a multimodal than by a standard bimodal form. The record has a median aerosol optical depth at 500 nm (AOD500) of 0.067 and weak total-load seasonality, but the fine-coarse balance varies markedly: the fine mode fraction (FMF) increases from 0.42 in summer to 0.72 in winter, and wind-resolved analysis shows stronger fine mode contribution under northerly and westerly flow than under southerly and easterly flow.
In the almucantar inversion subset, a modified Gaussian Mixture Model (mGMM), adapted here for inversion-retrieved 22-bin PVSDs, classifies 68 % of cases as multimodal, indicating that strict bimodality is not the dominant state under inversion-qualified conditions. In a spectral-consistency test based on Mie calculations and the Ångström exponent of AOD, the median signed relative difference decreases from 17 % with the standard bimodal PVSD representation to 3.4 % with the multimodal representation obtained with mGMM.
In a multisite comparison, Montevideo occupies an intermediate position between marine and urban-continental reference ranges, consistent with a mixed coastal-urban setting. TROPOMI NO2 columns are also higher in the overland quadrants, providing independent contextual support for stronger lower-tropospheric anthropogenic influence under overland flow. Overall, the 2020–2023 record shows a clear seasonal and wind-dependent redistribution between fine and coarse aerosol contributions.
### Review of
Agesta, A., Frins, E., O'Neill, N. T., and Holben, B.: Aerosol variability in the Río de la Plata estuary: Multimodal analysis of four years of AERONET observations, EGUsphere [preprint], https://doi.org/10.5194/egusphere-2026-4619, 2026.
### Summary
This study reports time series and seasonal variability of aerosol properties as measured by an Aeronet sunphotometer located in Montevideo, Uruguay. In addition, the study implements a novel aerosol size distribution modal fitting approach allowing it to investigate the presence of multi-aerosol modes in the atmospheric column. The new approach, while not thoroughly validated, appears to work better in reproducing observed angstrom exponents compared to angstrom exponents calculated from the standard bimodal almucantar inversions already available from the Aeronet portal. This is a promising approach that deserves further exploration. In addition, this study reports aerosol optical properties in a region of the world for which there are very few studies.
Overall, this work deserves to be published and made available to the community. However, prior to publication, I have some general and design comments that I would like to be addressed as detailed below.
### Main commentary
One overarching comment is that the title of the paper is a bit misleading: Is this a paper about aerosol variability in the Rio de la Plata region or is it a paper on the application of a new multimodal analysis to Aeronet observations? If trying to address both subjects, I find this paper a bit incomplete.
If it is an aerosol variability study, it should include the Aeronet stations at Buenos Aires and La Plata and analyze a longer time data set. As it is now, I see little value in carrying out such a statistical analysis with only four years of data (se comment on Section 4.1.1 below). Or if the goal is to carry out an application of multimodal fitting to Aeronet observations, it should include other sites. If the goal is to do both, the title and the way this study is presented need some rewriting to provide the appropriate framing and context of this study.
The application of the multi-mode fitting is novel and not widely known in the Aeronet community. It looks like an interesting subject for a dedicated study. Most importantly, this paper is convincing in showing that there is useful information on whether there are multi-modes present. This is an interesting finding, but it remains to be verified with independent observations. As shown, it is just a finding based on a mathematical fit that better fits the observations.
But are there actually more aerosol modes out in the atmosphere? It sounds like a likely possibility, and it would be desirable if the authors expanded on the available evidence from independent measurements. Particularly in the case of smoke, which is probably the aerosol type most common at this site, that can be compared with other major studies that use in-situ data.
Therefore, please clarify what the intention is, and I'll leave it to the editor on whether to make this paper one way or another or leave it as it is. Each topic could perfectly be split into two different papers.
**Other general comment:** When printed on paper, the lines in the figures are too thin and faint, and the fonts are very thin and small. Please make the lines thicker and use larger fonts.
### Detailed comments:
**Figure 1A:** Please enlarge the inset figure showing Uruguay to include the full Rio de La Plata estuary.
**Line 65 to 70:** This sentence is not clear. On the one hand, it mentions dust and organic matter, but the source of these from the north is not obvious, particularly referring to dust. Also, what is the season for it? A season for biomass burning is mentioned, but not for dust and organic matter.
**Section 2.2:** Please clarify what the TM5-MP model is. Also, make clear what version of TROPOMI products are available through Google Earth Engine. I have previously found that GEE does not necessarily report product versioning and that it often does not update to the latest versions.
**Section 3.1:** It should be made clear that the SDA products are measured throughout the day with the direct measurements, whereas the almucantar measurements are only measured early in the morning and late in the afternoon.
**Line 189 - 193:** Explain if the coefficients ai, bi, ci have a physical interpretation. It is reasonable to assume they represent parameters such as the mean, standard deviation, and concentration of the respective mode.
**Section 3.3:** Did you consider the Sellmeier equation? The reason is that this equation is more suitable for wavelengths in the near-infrared, such as 1020nm available in Aeronet. Please elaborate on why the choice of the Cauchy equation over this one: https://en.wikipedia.org/wiki/Sellmeier_equation
**Line 220:** Assuming a constant imaginary index of refraction is an important weakness, particularly during smoke events. This should be emphasized and pointed out throughout the paper when discussing these Mie computations.
**Section 4.1.1:** A general comment on this section. I think that some of the average analyses, such as periodicity and climatological comments, are not very useful due to the length of the dataset. It is just 3.75 years of data, and 2020 and 2021 were the years of the COVID-19 pandemic, and local pollution patterns are probably different than in the other years. Therefore, much of the discussion here about average values is not that important or significant because of the size of the data set.
Then perhaps some of the plots and analyses are not necessary. For example, Figure 3A and Figure 4B are not very informative given the size of the data set. The same applies to the analysis of annual periodicity discussed in lines 275 to 280, Figures 5 and 6. It seems like a not very informative section. I think this analysis could be included if, for example, more years are analyzed from the same site or if the sites in Buenos Aires and La Plata are included to provide a more thorough or comprehensive take on aerosols in the estuary.
However, Figures 3A and 4A are informative, and it will be useful to explore the variability observed in these plots in the context of your physical events. For example, note the differences in the years 2020 and 2021 compared to 2022 and 2023 in the months of June and March. Those are differences worth exploring. Why? Were there more fires in one year than the other? Was one year wetter or did it have more rain than the other?
**Section 4.1.2:** This is an interesting analysis. Although I emphasize the comment I made before, it would be better if additional years were included because, again, this seems to be too short of a time frame.
**Figures 7 and 8:** These are interesting, but for context, I would draw the coastlines given that the locations of Buenos Aires and Montevideo appear to be included, but it's not very easy to place them geographically.
**Figures 6, 7, and 8:** The meaning of the colors is a bit confusing. Is this a frequency of the parameters shown, or is it related to the magnitude of those parameters? It just needs to be clarified what this means in the caption and in the text.
**Lines 306 to 310:** In the discussion for Figure 9B, there is very interesting information in the contrast between the northern and west quadrants compared with the Southern and Eastern quadrants. Please expand or put this into the context of why there are differences if they are geophysical, that is, variability in the aerosol type.
**Figure 11 and related discussion:** I am a bit confused because the image clearly shows a third peak around 0.8 microns, and nothing about it is mentioned in the text. The text refers to two modes being present, so please clarify or discuss this point.
**Section 4.2.2:** This section should be revised or reanalyzed in the context of my previous comment on the short time series used in this analysis.
**Section 4.2.3:** Please clarify the Aeronet ID names for each of the stations used. Also, it is not clear in the analysis or comparison with the other stations whether the same wind quadrants were applied. These are all stations with different wind regimes, and it is not entirely clear if the same quadrants apply to all of them. More details need to be provided on how this analysis was done. Also, note the northern Atlantic sites are influenced by dust, so they should be taken into consideration in the discussion.
**Table 3:** Only coarse modes are reported. I wonder if the fine mode retrieval changes much when a single or a two-coarse mode is retrieved? Please comment on this point or perhaps expand the table to include the fine mode radius for the single and dual coarse mode retrievals.
**Line 415-418:** Please comment if multiple modes are retrieved during smoke events or any aerosol type event. Or clarify if there is no clear correlation between the aerosol type observed and the detection of multi-modes.
**Section 4.3:** This is very interesting. Nice analysis.
**Conclusions:** This is mostly a summary of the fitting performance, and there is very little discussion on the geophysical events at play here. After all, the title of the paper is about aerosol variability, and little is mentioned here. So, for example, it is relevant here to discuss what aerosol types were present and what features they have.
AI disclosure: This reviewer composed and created the text. An LLM was used only to correct typos, word repetitions and format the text. LLM was only used to formatting and cosmetic adjustments of the text. .