Performance comparison between CERES-MODIS and OMI in retrieving SSA across diverse aerosol regimes

Devi, Archana; Satheesh, Sreedharan K; Srinivasan, Jayaraman

doi:10.5194/egusphere-2025-4146

Preprints

https://doi.org/10.5194/egusphere-2025-4146

Preprints

01 Oct 2025

| 01 Oct 2025

Status: this preprint is open for discussion and under review for Atmospheric Measurement Techniques (AMT).

Performance comparison between CERES-MODIS and OMI in retrieving SSA across diverse aerosol regimes

Archana Devi, Sreedharan K Satheesh, and Jayaraman Srinivasan

Abstract. A comparative evaluation of aerosol single scattering albedo (SSA) retrieved from CERES-MODIS and OMI satellite instruments is presented across diverse aerosol environments. The analysis focuses on seven key regions: biomass burning areas (Amazon and South African Congo forests), clean and polluted oceanic zones (South Pacific, Arabian Sea, and Bay of Bengal), clean and polluted land areas (North America, Europe, Indo-Gangetic Plain, and Eastern China), and the Sahara Desert. Monthly aerosol type concentrations from MERRA-2 reanalysis data are used to investigate the seasonal variation in aerosol loading and composition. Results show that CERES-MODIS consistently captures SSA variability more effectively than OMI, particularly in regions dominated by absorbing aerosols such as black carbon. In biomass burning regions, CERES-MODIS displays a strong negative correlation between fire count and SSA, unlike OMI, which shows weaker or negligible correlations. Over clean regions, both instruments perform comparably, while over polluted zones and deserts, CERES-MODIS demonstrates greater sensitivity to aerosol type and seasonal trends. The findings highlight the relative strengths and limitations of both algorithms in aerosol monitoring under diverse atmospheric conditions.

Received: 25 Aug 2025 – Discussion started: 01 Oct 2025

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

Archana Devi, Sreedharan K Satheesh, and Jayaraman Srinivasan

Status: open (until 22 Nov 2025)

Post a comment Subscribe to comment alert

RC1: 'Comment on egusphere-2025-4146', Anonymous Referee #1, 11 Nov 2025 reply

Devi et al., “Performance comparison between CERES-MODIS and OMI in retrieving SSA across diverse aerosol regimes”, presents a comparative analysis of SSA retrievals from CERES-MODIS and OMI across a wide range of aerosol conditions. After reading the manuscript, I am not convinced about what new knowledge is provided. The current impression is that these two satellite-based SSA products are compared without solid justification or quantitative support. Given the issues elaborated below and the extensive revisions required, my recommendation is that the manuscript should be rejected in its current form, as it would require substantial additional analysis and restructuring to become acceptable.
The manuscript includes 28 figures in the main text, many showing similar comparisons for different regions. This approach feels overwhelming and does not enhance clarity. I strongly recommend consolidating figures, grouping regions with similar aerosol characteristics, and moving some plots to the supplementary material. A more concise and reader-friendly presentation would help highlight the key findings. It is unclear whether OMAERUV, OMAERO, or both were used, since both were mentioned in the "Data used" section. This ambiguity reflects the general lack of careful writing. Based on context, I assume OMAERUV was used, but this should not be left to the reader to guess.
My major comment has to do with the lack of quantitative insight and comparison against earlier findings. Now, the conclusions that CERES-MODIS performs better than OMI are not convincingly supported. The manuscript would benefit from including AERONET SSA data for validation or at least referencing existing literature that provides relevant quantitative information. There are large amount of useful earlier studies, now Just to give couple of examples. Eck et al. (2010) reported SSA seasonality in Beijing (Figure 6b), which aligns with OMI but not CERES-MODIS. Arola et al. (2015) show SSA seasonality (Figure 2) over the Indo-Gangetic Plain, again better captured by OMI (I guess IGP region is in your Figure 24, although the caption says otherwise). Ansari et al. (2025) providing SSA seasonality based on AERONET over many regions, which should be widely utilized and discussed also against your findings of CERES-MODIS and OMI SSA. Their results suggest that there are many regions, for instance Europe, where CERES-MODIS is not following very clearly the AERONET-based SSA seasonality. Currently, the manuscript does not clarify the size or exact location of regions used, making direct comparisons with literature difficult and weakening the credibility of conclusions.
The manuscript includes SSA comparisons for regions with very low aerosol loading (e.g., South Pacific Ocean, AOD ~0.1). Under such conditions, retrieval sensitivity is poor and uncertainties are large. Comparing SSA products in these cases is not meaningful unless limitations are clearly discussed.
The use of MERRA-2 aerosol composition data is promising but currently superficial. If the authors aim to support CERES-MODIS seasonal SSA variations, they should explicitly discuss expected seasonal patterns of absorbing components (BC, dust, OC). In several cases, CERES-MODIS trends do not align with these expectations (if roughly estimated assuming the scattering AOD is based on sea salt and sulfate and partly from OC divided by total AOD), raising questions about consistency and support that MERRA-2 is argued to provide.
The comparison between SSA and fire counts could be insightful, but the conceptual basis needs clarification. SSA is an extensive property, while fire counts are intensive. Now this part was superficially presented, while the manuscript should explain how these metrics relate and to what extent observed correlations are physically meaningful, why south-Africa and South-America cases differ (so one really cannot extrapolate results from one region to the other), and so on.

References
Ansari, K., & Ramachandran, S. (2025). Global insights on absorption characteristics of aerosols. Science of The Total Environment, 959, 178178. https://doi.org/10.1016/j.scitotenv.2024.178178
Arola, A., Schuster, G. L., Pitkänen, M. R. A., et al. (2015). Direct radiative effect by brown carbon over the Indo-Gangetic Plain. Atmospheric Chemistry and Physics, 15, 12731–12740. https://doi.org/10.5194/acp-15-12731-2015
Eck, T. F., et al. (2010). Climatological aspects of aerosol optical properties in Beijing based on AERONET measurements. Journal of Geophysical Research, 115(D00K30).

Reply

Citation: https://doi.org/10.5194/egusphere-2025-4146-RC1
RC2:
'Comment on egusphere-2025-4146', Anonymous Referee #2, 12 Nov 2025 reply
“Performance comparison between CERES-MODIS and OMI in retrieving SSA across diverse aerosol regimes” by Devi et. al. compares the performance of CERES-MODIS and OMI aerosol single scattering albedo SSA algorithms in different aerosol regimes, like biomass burning regions, deserts and polluted and not polluted ocean and land.
The manuscript offers interesting comparisons between the SSA derive from two algorithms, but substantial revisions and additional analysis are required before the paper is suitable for publication:
The study compares SSA at two different wavelengths 550nm for CERES-MODIS and 388nm for OMI, the expected differences of the SSA should be explained and taken into account when comparing those.
A discussion section highlighting also weaknesses of the own algorithms and the critical question of the comparability of the SSA at the different wavelengths should be added.
The authors promote their own algorithms by comparing their results to the one derived from a weaker instrument with lower information content and at another wavelength. A comparison with stronger instruments like POLDER would offer the option of a more critical reflection of the own results. Another option would be a comparison with SSA from AERONET to have more reliable values to compare with.
In multiple cases, the results shown in the figures are inconsistent with the corresponding descriptions in the text. It looks like the text includes what the authors want to see in the plots not what is in there.
Concreate points in the manuscript that need further investigations are listed below:
I count six regions as said in the introduction not seven as in the abstract. Please correct this. I would also recommend to add a map showing the different study areas as an overview to make it easier to keep the overall picture.

OMI is the instrument name not the algorithm retrieving the SSA, so the wording should be changed. The same question for CERES-MODIS, this is the combination of instruments is it also the algorithm name?

Reanalysis data from MERRA-2 is used to derived the AOD per aerosol type in the bar charts. Please explain how the aerosol type AOD is derived from the MERRA-2 data and cite the explicate dataset used in the study. Also, it is not clear from what timeframe the plotted data is from: climatology or simply one year or something else?

The same for the OMI and CERES-MODIS data also here it is not clear from what time the data is from and if it is from the same time.

The source of the datasets especially the OMI SSA should be cited and listed in the Data availability section. Is the CERES-MODIS SSA data public available? This should be mentioned in the Data availability section.

I would recommend to explain the methodology of the CERES-MODIS SSA algorithms clearer and add how the uncertainties given in Pg5 ln1-7 are derived and answer why it requires sufficient variations in AOD values in the region (last sentence of section 2.1

You write that the OMI SSA data is validated against airborne, ground based, and satellite observations (Pg6 ln13-15). Do these validations show the same limitations you see in your comparison?

In section 2.3 you write that the MODIS Fire count data has assigned confidence levels, allowing users to apply thresholds depending on their application. What are you using in this work? Maybe add a sentence explaining why fire count is a suitable proxy for aerosols from biomass burning.

I recommend to rename subsection 2.4 to: “MERRA-2 reanalysis data” to be consistent with the other subsections in section 2 and add what part of the MERRA-2 data you used in the study.

I would recommend to shorten the results description by putting the example regions for one big region together, showing in the plots of both and describing what is common and where the differences are not repeating the description of the same behavior. I the two biomass burning regions together, the two polluted ocean regions and so on.

Could you please explain where one point in the figures 5 to 8 is coming from, are those fire counts summed up over a month in the hole region and the SSA is a mean over the time?

When I compare Fig. 3 and 4 to the numbers and months in the text Pg13 ln8 (nearly 80000 in August in the text and below 700000 and peak in Juli in Fig. 4) and Pg 14 ln2 also 800000 in August instead of 700000 at the peak in Juli from Fig. 4 and nearly 200000 in the text and peak values below 150000 in Fig. 3. Please correct to be consistent between plots and text.

Pg 20 ln 12 correct the reference to the Figure – should be Figure 3 not 5.10.

In Pg21 ln2 its 60-70% but looking at Fig. 9 I see 80% so I would correct 60-80%.

Please explain the meaning of the white dot in the boxplots (for example Fig. 10).

10 and the belonging description do not fit together. You write about lower SSA values from June to November but the boxplot shows significant lower values in DJF than in JJA, please correct this.

Pg 24 ln 7 how can you say anything about particular months (July and August) from a plot showing only seasons?

Pg 24 ln11-13 can you really say anything about the aerosol loading from the SSA plots? And I would specify that you can see a better sensitivity to biomass burning aerosols.

In Pg 25 ln8 you write about an increase in sulphate and dust but Fig. 13 shows a decrease in those.

Pg 26 ln 13 “austral” should be Australian?

Pg 26 ln 12-15 contradicts with Fig. 14: the plot shows the highest values in JJA and the lowest in the seasons listed in the text to be the highest. Please correct this.

Pg 27 ln 5 I would say that only the CERES-MODIS SSA from Fig. 14 fits to the bar graph in Fig. 13.

Pg 27 ln 13-16 could this “underestimation” in the OMI data be caused by the different wavelength used? What does the OMI SSA validation says in the respect of the potential error?

Pg 28 ln 3-9 stays in conflict with the Fig 15: dust is the dominant component in the summer months (May to August); and due to the wind direction, it does not come from Thar Desert but from Arabian deserts and Danakil-Somalia. The sulphate, OC and BC have increased levels in the “not summer” months – mainly in November, December, March and April which make sense because of the winter monsoon bringing air from India.

Pg 29 ln 6ff: in March OC is higher than Sea Salt and in May the Dust AOD is around the same size as sulphate and Sea Salt so saying that the peak is driven by Sea Salt and sulphate is strange. Please be concreate which months are meant when talking about pre-monsoon period. The increase in OC and BC can be observed from November till March in the Fig. 16. The lower OC, BC, Dust and Sulphate values could also be caused by the wind direction in the monsoon season (June to September) coming mainly from the ocean.

Pg 31 ln 14 specify that theses variations in the SSA are observed in CERES-MODIS algorithm.

PG 32 ln 2-3 the SSA boxplots do not give information about the aerosol loading so please correct the sentence.

Pg 32 ln 6-10: This is only observed in CERES-MODIS so please be concreate.

Pg 34 ln 5 the peak is from April to September, please correct.

Pg 34 ln 6-8 can you please explain the correlation between the warmer months and the increased industrial activity and vehicular emissions.

Pg 34 ln 8-10 for example Prospero et al 2021 identifies the dust peak season in June and Juli – how would you explain the difference compare to Fig 19?

Pg 35 ln 7 from Fig 20 its March to August.

Pg 35 ln 7-8 explain the correlation of increased industrial emissions in the warmer month.

Pg 35 ln 8: dust AOD over Europe is higher in spring than in summer so that should be spring here.

Pg 35 ln 23-25: dust AOD has the peak in MAMJ in both regions following Fig. 19 and 20

Comparing Fig. 21 and Fig. 22 it looks like there is more variation in the SSA in Europe than in North America, which is not consistent to a comparison of Fig. 19 and Fig. 20. Can you please explain this.

Pg 38 ln 17-18: As dust is shown to be the component with the highest AOD values in the Indo-Gangetic Plain in Fig. 23 it should be mentioned here when describing aerosols in this region.

Caption Fig 24 refers to Eastern China but the plot title says Indo Gangetic Plain, please correct.

Pg 40 ln 10 please add the months of the monsoon period mentioned and can you please explain why cloud coverage can generate low SSA values?

Pg 40 ln 8-12 and Fig 24 here OMI SSA shows the expected behavior. This should be mentioned as one of the regions where this algorithm performs better and discussed in the discussion session why this is the case.

Pg 41 ln 4-5: Sulphate AOD of 0.2 (June) to 0.5 (January) is high but not constant.

Pg 41 ln 8-9: Dust AOD is relatively high from March to June and much higher than in October, please correct this.

In Fig. 26 I see nearly constant behavior of the SSA over the year with little lower value in MAM, which also corresponds to Fig 25 but not to the text in Pg 41 ln 12ff. Please correct this.

Pg 42 ln 12-13 Please give a reference for the size of the Sahara

Pg 43 ln 9-10: highest AOD values from March till October and peaking in May in Fig 27, please correct in the text to be consistent with the figure.

Pg 45 ln 6-8: Fig 27 shows a lower percentile of Dust AOD compared to the rest of the year in JJA and a higher percentile of Sulphate AOD – how can you explain the lower SSA values in Fig. 28 in this respect. The expected behavior would be like in the OMI SSA data in Fig. 28.

Pg 45 ln 11-18: The boxplots in Fig. 28 show contradicting behavior for OMI and CERES-MODIS not similar as said in the text. Please correct this.

Pg 45 ln 21: in the list of regions the ones over land are missing, please add.

Pg 46 ln 20 no comparison to ground measurements are shown so where do you get this information from that CERES-MODIS SSA is closer to them than OMI SSA.

Pg 46 ln 24-25 Please specify how cloud cover affects the SSA retrieval – I could expect that no retrieval is possible in the presence of clouds.

Pg 47 ln 7-10: The peak AOD values are from March to August, please correct. And please correct the SSA behavior over the Sahara as mentioned before.

Pg 47 ln 11-13. Be precise in describing with SSA dataset has which advantages and disadvantages.

Pg 47 ln 20: Please add the other used datasets in the data availability section.

One Reference is listed twice, please correct: Torres, Omar, Aapo Tanskanen, Ben Veihelmann, Changwoo Ahn, Remco Braak, Pawan K. Bhartia, Pepijn Veefkind, and Pieternel Levelt. "Aerosols and surface UV products from Ozone Monitoring Instrument observations: An overview." Journal of Geophysical Research: Atmospheres 112, no. D24 (2007).

Reply
Citation: https://doi.org/10.5194/egusphere-2025-4146-RC2
RC3: 'Comment on egusphere-2025-4146', Anonymous Referee #3, 12 Nov 2025 reply

The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2025/egusphere-2025-4146/egusphere-2025-4146-RC3-supplement.pdf
Reply

Citation: https://doi.org/10.5194/egusphere-2025-4146-RC3

Archana Devi, Sreedharan K Satheesh, and Jayaraman Srinivasan

Viewed

Total article views: 174 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	BibTeX	EndNote
130	36	8	174	6	7

HTML: 130
PDF: 36
XML: 8
Total: 174
BibTeX: 6
EndNote: 7

Views and downloads (calculated since 01 Oct 2025)

Month	HTML	PDF	XML	Total
Oct 2025	108	24	8	140
Nov 2025	22	12	0	34

Cumulative views and downloads (calculated since 01 Oct 2025)

Month	HTML	PDF	XML	Total
Oct 2025	108	24	8	140
Nov 2025	22	12	0	34

Viewed (geographical distribution)

Total article views: 170 (including HTML, PDF, and XML) Thereof 170 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 13 Nov 2025

Short summary

This study compares aerosol single scattering albedo data from two algorithms, CERES-MODIS and OMI, across different regions like forests, oceans, land, and deserts. It finds that CERES-MODIS tracks aerosol absorption more accurately, especially in areas with smoke and pollution. In clean regions, both algorithms perform similarly. The study helps scientists understand which satellite gives better data in different conditions, supporting improved climate and air quality research.


Total:	0
HTML:	0
PDF:	0
XML:	0