the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 15 Jul 2025
Mapping 532 nm Lidar Ratios for CALIPSO-Classified Marine Aerosols using MODIS AOD Constrained Retrievals and GOCART Model Simulations
Abstract. The NASA-CNES Cloud-Aerosol Lidar and Infrared Pathfinder Satellite Observations (CALIPSO) mission provided a spaceborne global record of atmospheric aerosol and cloud profiles from June 2006 to June 2023. As an elastic backscatter lidar, the CALIPSO Cloud-Aerosol Lidar with Orthogonal Polarization (CALIOP) typically required an assumption of the aerosol lidar ratio (extinction-to-backscatter ratio; Sa) to retrieve aerosol extinction and column-integrated aerosol optical depth (AOD). In all previous versions of its data products, the CALIPSO extinction algorithms first determine the aerosol types then assign one Sa value globally for each aerosol type (e.g., 23 sr for marine at 532 nm). One of the major changes for the final CALIPSO data products release (Version 5, or V5) is the implementation of regional and seasonal Sa tables for CALIOP-classified “marine” aerosols. In this study, we describe the process of creating the tables using 12 years (June 2006–August 2018) of Aqua Moderate Resolution Imaging Spectroradiometer (MODIS) total column AODs to constrain collocated CALIOP backscatter profiles in a Fernald inversion scheme and infer Sa (at 532 nm), focusing solely on the CALIOP “marine” aerosol type. The Goddard Chemistry Aerosol Radiation and Transport (GOCART) global aerosol model is used to estimate sea salt volume fraction (SSVF) that are collocated with the constrained Sa retrievals. Patterns of smaller SSVF (< 65 %) and larger constrained Sa (> 40 sr) are found near land masses, while larger SSVF (> 95 %) and smaller constrained Sa (< 30 sr) are generally observed in the remote oceans. The inverse empirical relationship found between modeled SSVF and constrained Sa over global oceans yields values of ~21 sr for SSVF of 100 % (i.e., “pure” marine) and ~58 sr for SSVF of 0 % (i.e., the absence of marine aerosol). This relationship is applied to develop regional and seasonal hybrid (retrieval and model-assisted) climatological Sa maps for CALIOP-classified marine aerosols; i.e., when MODIS-constrained results are not available, the model-assisted values are used. These hybrid Sa maps are subsequently used to retrieve new CALIPSO Level 2 (L2) aerosol extinction profiles and column AODs in the V5 release. For a 4-month (January, April, July, and October 2015) analysis, the V5 L2 CALIPSO AODs compared better to CALIPSO Ocean Derived Column Optical Depth (ODCOD) than the CALIPSO Version 4.51 (V4.51) standard AODs in several regions, most notably the Bay of Bengal/Arabian Sea, where smoke/pollution typically mixes with marine aerosols. Also, the V5 CALIPSO AODs likely provide a lower AOD bias and root-mean-square-error than V4.51 AODs relative to coastal and island Aerosol Robotic Network (AERONET) AODs, as found in a validation study using data from June 2006 through October 2022. The technique described in this study contributes to CALIPSO’s final V5 data products release and provides critical Sa information for future spaceborne elastic backscatter lidars.
- Preprint
(10315 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2025-2832', Anonymous Referee #2, 05 Aug 2025
“Mapping 532 nm Lidar Ratios for CALIPSO-Classified Marine Aerosols using MODIS AOD Constrained Retrievals and GOCART Model Simulations” by Toth et al. documents the updates to the lidar ratio selection methodology for marine aerosols in v5. The updated method uses MODIS AODs and GOCART modeled aerosols to create seasonally and spatially varying maps from which to select their marine aerosol lidar ratios. They find that these updates provide AODs that better align with those calculated through the ODCOD than the previous version (v4.51) and also better agree with those measured by AERONET sites in coastal and island locations. The paper provides valuable documentation of the updated CALIOP data product, which, despite CALIPSO’s retirement in 2023, still provides a valuable long-term dataset for cloud/aerosol research. This update to marine aerosol lidar ratios represents a significant advancement through addressing regional and seasonal variability that was previously unaccounted for in the previous fixed value lidar ratio assignment. The paper is generally well organized and written; however, this reviewer found it to be a bit on the long side. I would recommend publication after some minor revisions.
Major Points:
- The study provides a valuable update to the assignment of marine aerosol lidar ratios; however, the approach raises fundamental questions about CALIOP’s aerosol typing framework. The manuscript would benefit from directly addressing how these new spatially/seasonally varying marine lidar ratios relate to the existing aerosol typing framework, particularly the distinction between marine and dusty marine. Is differentiating between dusty marine and marine needed or useful anymore with these new methods? Many of the regions where the largest differences in lidar ratios occur, such as the Bay of Bengal, are regions where dusty marine and other aerosol mixtures are common. Here the study assigned V5 marine lidar ratios exceeding the V4.51 dusty marine value of 37 sr in some regions. This convergence between the new variable marine lidar ratios and the dusty marine values raises two questions for me: 1) may some of these aerosol layers currently classified as marine in these regions be misclassified dusty marine? 2) Does this new method render the discernment between marine and dusty marine somewhat obsolete? Connecting lidar ratios to modeled sea salt volume fractions suggests that this approach could be beneficially extended to other marine-influenced aerosol classes, which perhaps is covered by the tables/maps noted at L142, but not shown in this paper.
- The exclusion of modeled dust aerosols from the SSVF calculations (L407-409) warrants reconsideration or further justification. This study already focuses specifically on CALIOP-identified marine layers, which have already passed the CALIOP typing algorithm’s criteria (depolarization or otherwise) for classification as marine vs. dust (or other types). These marine-classified layers would still contain some dust at concentrations below levels that would trigger classification as dusty marine or something else. By omitting dust from the SSVF denominator, SSVFs would be inflated, especially in transitional, dustier, regions.
- The manuscript is a bit lengthy and could be strengthened through some editing, especially in the introduction. The extensive literature review from L148 to 222 largely duplicates the information already presented and effectively summarized in Table 2. Streamlining the literature review and highlighting only the most significant studies would benefit readers.
Minor Points:
L164: is the MPL at 532 or 523nm?
L330: This sentence read a bit weird. Consider: “Note that this approach produces a negligible proportion of negative Sa values (less than 0.05%), and our methodology minimizes the influence of these outliers by using median values when creating the Sa maps (Sections 3 and 4).” or something similar. The phrase “our use of medians” sounded off to me…
L347: What are typical stratospheric AOD values in the SAPP? It strikes me that removing stratospheric AOD from the column would result in a fairly small correction outside of volcanic/pyrocumulonimbus events.
Figure 13: The regional boxes encompass a lot of land. I would recommend being more explicit that the analysis only includes at the oceanic parts of the domain.
L763: State why ODCOD is expected to be greater than v4.51
L816: Consider adding that models parameterize sea salt emissions by wind speed.
Citation: https://doi.org/10.5194/egusphere-2025-2832-RC1 -
CC1: 'Reply on RC1', Gregory L. Schuster, 19 Aug 2025
Firstly, I would like to mention that I no longer work for NASA (since January 2025) and therefore my comments are merely those of a citizen scientist and do not represent a consensus view of the co-authors. I have not met with the co-authors to discuss this work since January, 2025.
Interesting comment about the dusty marine, though… When we started this project, we knew that we did not have time to alter the classification algorithm shown in Figure 1 (which is identical to the Version 4 selection process shown in Kim et al. 2018). So the algorithm discrimination between Marine and Dusty Marine remains unchanged. However, it is clear from the elevated constrained lidar ratios in Figures 3 & 6 that much of the world’s oceans are not purely marine (or ’Clean Marine’) when the algorithm selects the Marine aerosol type; otherwise, we should see lidar ratios of 23 +/-5 sr everywhere (per Kim et al. 2018, Table 2). We kept the wording ‘Marine’ because we expect at least some of the aerosols in these plumes to be sourced from surface water.
We considered renaming the Marine aerosols as another type for certain regions and seasons depending upon the lidar ratio climatology, but we ultimately left that for the users to decide. For instance, one might set an appropriate lidar ratio threshold and relabel the CALIPSO ‘Marine’ aerosols as ‘polluted marine’ for some of the grid boxes (e.g., off the India coasts in the non-summer months, per Figure 6).
Regarding Dusty Marine, I presented lidar ratio maps of Dusty Marine at several conferences in 2023. Those constrained lidar ratio maps are very similar to the constrained lidar ratio maps for Marine aerosols shown in Figure 3; globally, the mean difference between the Marine and Dusty Marine maps is 0.5 sr and the standard error is 0.04 sr. The largest differences occurred at tropical/mid-latitudes over the Atlantic where dust plays an important role — the mean difference there is 1.8 sr with a standard error of 0.12 sr. We originally intended to use the Marine maps as a baseline for Dusty Marine (because the Dusty Marine data was too sparse to build robust seasonal maps) and include Dusty Marine in this paper, but that concept was nixed by higher powers.
The bottom line is that the constrained lidar ratio maps of Figures 3, 6, and 10 indicate that CALIPSO Marine aerosols are frequently mixed with anthropogenic pollution, biomass burning, or other aerosol types in various regions and seasons. The same is true for Dusty Marine. Users should keep this in mind when they are using specific aerosol types in their analyses. For example, if one wants to target marine or dusty marine aerosols, they might consider excluding certain coastal regions from their dataset because they are likely mixed with other aerosol types as well. That is my interpretation of the data, anyways.
Citation: https://doi.org/10.5194/egusphere-2025-2832-CC1
-
RC2: 'Comment on egusphere-2025-2832', Anonymous Referee #1, 05 Aug 2025
The paper by Toth et al. describes the creation of regional and seasonal lidar ratio tables for CALIPSO’s final data release (V5). The study focuses on marine aerosol and for the creation of the lidar ratio tables passive and active remote sensing measurements were used, in particular MODIS AOD constrained CALIOP backscatter profiles in addition to GOCART model simulations of sea salt volume fraction (SSVF). This is a great step beyond the single globally-constant lidar ratio value that was used in earlier versions.
In addition to the well described methodology, differences between V4.51 and the new V5 CALIPSO extinction and AOD were also presented together with a preliminary validation using the CALIPSO ODCOD algorithm for seven study regions by utilizing four months from 2015. Improvements due to the new seasonal maps are demonstrated in a case study, focusing on one of the aforementioned regions. Validation using AERONET retrievals was also performed and the results suggest that the lidar ratio tables for marine aerosol improve the AOD retrievals.
Overall, the paper is of good quality, structured and well-written and should be published after only a few minor revisions (listed below).
Line 84: It would be worth mentioning a few of the campaigns used, in addition to a number of studies that revealed a bias induced by the marine aerosol type in the CALIPSO data.
Line 86-89: Similarly, as above, one should acknowledge previous studies that highlighted the necessity for the introduction of the dusty marine type.
Section 2.1: While I really appreciate the fact that the authors provide the filenames of the products that they’ve used, I find it slightly interruptive for reading. The data availability section contains the filename already. Same applies for sect. 4.3 and the names of the variables used.
Line 322: Please cite the available software properly.
Lines 328-330: Why is the lidar ratio allowed to vary to physically meaningless values, i.e., -50 sr? Aren’t the retrievals stable for a range 0 to 150 sr? Could the authors also provide references for the sensitivity studies that they mention?
Line 340: Please explain the quality flags selected for the MODIS data. A reader might not know what a Land Ocean Quality Flag value greater or equal to 1 means.
Lines 361-368: It is not clear why the additional filtering step was used. Please explain and also summarize in 1-2 sentences the main points of Li et al., 2022 regarding the SNR/classification confidence relationship.
Section 4.1: Some details regarding the GEOS GOCART model could have been included earlier in the methods section.
Lines 488-489: Please rephrase, is “replacing” indeed the right word?
Line 496: “Eq.” is missing before the parenthesis.
Lines 526 and 535: How was the minimum of 50 points selected?
Line 553: The authors could discuss the meteorological conditions leading to the seasonal aerosol transport.
Lines 600-606: Could you please elaborate more on the additional procedures, especially on the second one? The minimum lidar ratio of 15 sr is justified from the field measurements. What drove you into implementing the outlier replacement procedure? Where there many outliers and could you please include a statement regarding that?
Lines 618-619: The reported maximum lidar ratios are clearly influenced by non-marine aerosol and they should be discussed together with the SSVF.
Lines 626-629 and Fig. 11: A lot of pixels flagged as “Retrieval” in Fig. 12 are accompanied by the maximum assigned uncertainty of 22% (e.g., South Atlantic and Pacific during SON). Could you please provide a statement with the typical range of the uncertainty (before the assignment of 22%) for the pixels with the assigned 22% uncertainty?
Lines 690-695: How and why were the study regions defined as such? It should be stated that by this selection regions with e.g., modelled-only lidar ratios, high SSVF-low lidar ratios etc. were covered.
Lines 746-747: For clarity, please point out again that these results correspond to 2015.
Citation: https://doi.org/10.5194/egusphere-2025-2832-RC2
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 406 | 32 | 20 | 458 | 15 | 14 |
- HTML: 406
- PDF: 32
- XML: 20
- Total: 458
- BibTeX: 15
- EndNote: 14
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1