the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Cloud condensation nuclei concentrations derived from the CAMS reanalysis
Karoline Block
Mahnoosh Haghighatnasab
Daniel G. Partridge
Philip Stier
Johannes Quaas
Abstract. Determining concentrations of cloud condensation nuclei (CCN) is one of the first steps in the chain in analysis of cloud droplet formation, the direct microphysical link between aerosols and cloud droplets, a process key for aerosol-cloud interactions (ACI). However, due to sparse coverage of in-situ measurements and difficulties associated with retrievals from satellites, a global exploration of their magnitude, source, temporal and spatial distribution cannot be easily obtained. Thus, a better representation of CCN is one of the goals for quantifying ACI processes and achieving uncertainty reduced estimates of their associated radiative forcing.
Here, we introduce a new CCN dataset which is derived based on aerosol mass mixing ratios from the latest Copernicus Atmosphere Monitoring Service (CAMS) reanalysis (RA: EAC4) in a diagnostic model that uses CAMSRA aerosol properties and a simplified kappa-Köhler framework suitable for global models. The emitted aerosols in CAMS are not only based on input from emission inventories using aerosol observations, they also have a strong tie to satellite-retrieved aerosol optical depth (AOD) as this is assimilated as a constraining factor in the reanalysis. Furthermore, the reanalysis interpolates for cases of poor or missing retrievals and thus allows for a full spatio-temporal quantification of CCN.
This paper illustrates the temporal and spatial structure of CCN and their abundance in the atmosphere. A brief evaluation with ground based in-situ measurements demonstrates the improvement of the modeled CCN over the sole use of AOD as a proxy for CCN.
The CCN dataset, which is now freely available to users (Block, 2023), features 3-D CCN concentrations of global coverage for various supersaturations and aerosol species covering the years from 2003 to 2021 with daily frequency. This dataset is one of its kind as it offers lots of opportunities to be used for evaluation in models and in ACI studies.
- Preprint
(28645 KB) - Metadata XML
- BibTeX
- EndNote
Karoline Block et al.
Status: final response (author comments only)
-
RC1: 'Comment on essd-2023-172', Anonymous Referee #1, 01 Aug 2023
This manuscript introduces a new, comprehensive atmospheric CCN data set based on aerosol reanalysis, and shortly investigates the performance of this proxy against observations and compared with an earlier proxy. The new CCN product introduced in the paper is highly relevant for aerosol-cloud climate investigations, and is anticipated to have a wide range of applications. This paper is scientifically robust, and very well written and structured. I have only a few, relatively minor, issues that should be fixed before accepting this paper for publication.
The manuscript provides a valuable discussion on the problems associated with CCN estimated from AOD measurements (section 1). Related to this, although the new CCN proxy shows major improvements over the AOD-derived one, it is still far from perfect, as can be seen from Figures 3 and 4 and Table 3. The authors should better acknowledge the remaining uncertainties, rather than claiming a good agreement with measurements (lines 378-379).
The paper would benefit from a bit more detailed discussion (1-2 paragraphs) on what might cause the remaining problems with the new CCN proxy (see my previous comment), and what could be done to improve the proxy further. One clear issue related to this is the assumption of an external aerosol mixture in the aerosol reanalysis product. This poses certainly challenges for estimating CCN concentrations, as many of the simulated components (especially SU, OM and BC) are closer to internal than external mixtures in large parts of the atmosphere.
Saying that the correlation between the CCN proxy and measured CCN concentration almost doubles when using the new CAMS-derived CCN proxy compared with AOD-derived CCN proxy (lines 400, 416 and 450) is not a statistically robust statement. For example, doubling the correlation from 0.49 to 0.98 would be a huge improvement, while doubling it from 0.01 to 0.02 would not help at all in practice. Please reword using statistically relevant measures in expressing the improvement in the new proxy compared with the old one.
Finally, I do not understand what is meant by the last sentence in section 3.2 (lines 350-351).
Citation: https://doi.org/10.5194/essd-2023-172-RC1 -
RC2: 'Comment on essd-2023-172', Anonymous Referee #2, 01 Sep 2023
Review of Manuscript
Title: Cloud Condensation Nuclei Concentrations Derived from the CAMS Reanalysis
Author(s): Block et al.
MS No.: essd-2023-172
MS Type: Data Description Paper
This study employs the CAMS reanalysis aerosol mass mixing ratios from various aerosol species to calculate CCN number concentrations using estimated parameters of the aerosol particle size distribution. The results are provided globally over a 19-year period, and various spatial distributions, time series, and a brief comparison with observations are presented. In my opinion, this manuscript is well-written, and the analyses within it are presented clearly. This will be a great paper and the community will benefit substantially from this study. Therefore, I recommend publication after addressing the comments below.
General Comments:
- Your methodology is explained clearly. It would be great to elaborate on the exclusion of dust particles from the CCN analysis.
- While your study excludes dust from CCN calculations, it's worth noting that many studies consider dust to be an important contributor to CCN concentrations. For instance, Che et al. (2022) employed UKESM to investigate CCN sources, and they included dust in their analysis.
- They highlighted that even though dust particles are generally insoluble, wettable dust particles with larger diameters can function as CCN. Moreover, small dust particles can accumulate soluble materials through internal mixing during transport, significantly enhancing their activation potential (Bègue et al., 2015; Dusek et al., 2006; Gibson et al., 2007; Hatch et al., 2008).
- This significance of Saharan dust as CCN at a 0.2% supersaturation has also been emphasized in studies by Weinzierl et al. (2017) (Figure 12) and Haarig et al. (2019) (Figure 3). - While you've provided an explanation about the treatment of dust in CAMS in Section two, it is excluded from the results in Section three. I know you provided a brief explanation about dust as CCN towards the end of Section 3.3, however you could elaborate on your justification for excluding dust particles, ideally accompanied by supporting citations. Additionally, I recommend relocating this discussion to Section two, potentially integrating it at the end of 2.1 or the beginning of 2.2.
- It would be good to compare your results with previous studies and highlight the similarities and differences. In particular, Sections 3.1 and 3.2 (maps, time-series, etc.) do not provide any observations. I've added some examples in the specific comment part, but I encourage you to add any other helpful references with explanations.
- Given the significant influence of CCN on low clouds, have you examined the figures for the lower troposphere or surface, rather than the whole troposphere? I really don’t want to impose additional analysis, but if you already have those figures, it would be ideal to include them in the supplementary.
- Figures are very informative. Many figures are based on 2007 averages. Why not use the 19-year averages? If you plotted figures with the full climatology, you could add show them in the paper, or add a few of them to the supplementary.
- Most figures and tables are based on one supersaturation (0.2% or 4%). Can you add a brief explanation on 1) why 0.2% is preferrable and 2) how different the results are when using other supersaturation values.
- The abstract could benefit from a more specific description of the methods and results. I've provided a few suggestions in the specific comments part.
- If you haven't done so already, please provide a description of limitations on your dataset website. For instance, mention the possible biased values over polar regions, as CAMS is not constrained in polar regions (as you mentioned in L413); or the AOD assimilation to CAMS calculation of mass mixing ratio is less reliable over bright surfaces or for cloudy conditions.
- The Conclusions part could benefit from summarizing the limitations of your study.
- Following up on the previous comment, the limitations in the polar region should be listed in a paragraph in the conclusions and possibly in the abstract. You can mention a note in the abstract, for example, "The derived CCN based on CAMS is more reliable over the tropics and mid-latitudes." Also, the issue with the polar region needs to be mentioned earlier when describing Figure 1.
- Since CAMS outputs are provided every 3 hours, can you comment on 3-hourly CCN? Do you expect less agreement with observations? I understand that 3-hourly data can take up so much space. Therefore, if you plan to continue this work in the future, it might be helpful to consider a daily maximum dataset in addition to daily mean data.
- Equations and derivations are described clearly. I encourage the authors to check again and add units for all variables (The units are missing for some variables). This will help better understand the variables and will make a huge difference for scientists who want to implement the derivations into their codes/models.
- There are many occurrences of the word "concentration," but often it is not mentioned whether it is about number concentration or mass concentration. You can add a note at the beginning of the paper clarifying that you mean number concentration unless stated otherwise (If that is true).
- Are spatial averages weighted by the area of each grid cell? This can be mentioned in the paper.
Specific Comments:
- L1 and 16: Perhaps change "concentrations" to "number concentration."
- L13-15: The results of your study need to be mentioned more specifically. Instead of the first sentence, maybe mention that the derived CCN dataset captures the general trend and spatial and temporal distribution of total CCN and CCN from different aerosol species. For the second sentence, it would be good to add the correlation values, as they have been improved.
- L22-26: I encourage you to add references for each sentence.
- L39-40: Although AOD and AI do not inform about the aerosol species, the angstrom exponent is related to aerosol size and therefore implies aerosol type. It would be good to mention this, along with relevant references, e.g., Kapustin et al., (2006) and Shinozuka et al. (2015) (Figure 3).
- L40: I think the use of "this study" should be reserved for instances where you refer to your current manuscript. Please mention the exact reference to avoid confusion.
- L53-68: Is this about the surface or vertical distribution of aerosol properties?
- L61: Are the "various techniques" here in-situ measurements, ground-based remote sensing, or satellite remote sensing?
- L78-80: Please mention if this is 3D or just at the surface.
- L85-87: Perhaps a rephrasing is needed, as it is not clear if GCMs and observations are compared, or if different observational methods are compared.
- Paragraph starting at L99: Please add the temporal and spatial resolutions of the CAMS model. Additionally, does CAMS assimilate meteorological data in addition to satellite observations?
- L115-120: Please add references.
- L153-156: Can you be more specific about the cloud properties in CAMS? Does CAMS have a cloud microphysics model coupled to the aerosol model, or does it receive cloud parameters from another model/reanalysis?
- L165-167: MODIS has an AOD product called deep-blue (DB) that covers over bright surfaces where dark-target (DT) AOD is missing. It seems that CAMS model does not use DB because it was not available when assimilation of DT was implemented in CAMS. It might be good to mention a possible solution to DT limitations and how much of a difference it might make. Please refer to Garrigues et al. (2022) and relevant references in that study.
- L173: Remove the "tau" symbol after "AOD" and the "lambda" symbol after "wavelength." Those symbols haven't been used anywhere else in the manuscript.
- L192-194: It is first mentioned that CCN is averaged daily, but then it is said that the data is computed at 00 UTC.
- Paragraph starting at L191: Instead of “CCN”, it seems more proper to use “aerosol number concentration” for the first paragraph of Section 2.2.
- Paragraph starting at L201: To make a clear transition to the new topic, you can add something like this at the beginning of the paragraph: "After deriving the aerosol number concentration of aerosol species from the CAMS reanalysis, CCNs are calculated diagnostically…" (If I understand correctly)
- L209: Please define the geometric standard deviation in one sentence.
- L210: I don't think it is accurate to say Equation 2 is derived by dividing density by mass. It is an assumption for the particle size distribution.
- L234: What do "w" and "a" stand for in the subscript of tension parameter? Are they "water" and "air"?
- L240-242:
- I encourage you to rephrase this part. What I understand from this is that although supersaturation in warm clouds typically ranges between 0.1% and 1.5% (Spracklen et al., 2011), you selected values near the lower limit of s (Pohlker et al., 2023). Therefore, s << 1 is valid. Is that correct?
- Also, how is it justified to select the lower limit of s for warm clouds?
- Additionally, you mentioned in L195 that s is selected from 0.1 to 1% in your dataset. Most of these values do not satisfy (s << 1). An explanation would be helpful.
- Equations 7 and 8: It seems that these are based on radius, whereas equations 3-6 are based on diameter. Is adding Equations 7 and 8 necessary? There is a simple relationship between diameter and radius that can be applied at the very end.
- Figure 1: What is the reason for averaging over 2007? Why not use the full climatology? Also, it looks like high-elevation regions show lower CCN. Is this because the CCN is integrated over a column with smaller height for these regions? If so, it would be nice to see maps of surface CCN and address this artefact. You can include these in the supplementary.
- L270: Perhaps mention that CCN load is the vertical integration of CCN over all levels up to 10 km.
- L273-274: One sentence to describe the reason for this would be great.
- L276-277: I am not sure if this can be seen in the vertical profiles. Maybe mentioning the exact latitude and height can help. I don’t see it in maps either.
- L280-281: Is this due to small CCNs that correspond to larger supersaturation? See Figure 7 in Che et al. (2017).
- L282: Is this CCN decoupling or just very low CCN concentration over Antarctica? It might be good to briefly define the West Wind Drift.
- Figure 2b: When you average over all latitudes for each longitude, it is possible that some high values of N.H. summer cancel low values of S.H. winter. This might lead to less distinction among seasons on this panel.
- L286-287: If you have checked figures for MO and SU values, please mention "figure not shown" after MO and SU.
- L290: It seems to me that there is a seasonal cycle, but overall, the CCN concentration is low over the Pacific. Also, did you mean 140W?
- L296-297: Can this be confirmed in your data by looking at time series of different components?
- L312: The IMO 2020 regulation limits sulfur in ship's fuel. Is that implemented in the CAMS model, and if so, is it detectable in the SU component of CCN?
- L315: Do you mean OM? OC has not been mentioned before.
- Figure 4: Again, why not show 19-year averages?
- L335: Please add "North Hemispheric" before "high latitudes."
- L345-346: Please refer to Figure 2a.
- L350-351: This sentence doesn’t seem correct and can be deleted or rephrased.
- L353 and L371-374: Is there a reference for ARM datasets and their preparation and quality control?
- L378-379: Please add that on average CAMS overestimates CCN for three stations (MAG, PGH, PVC) and this is stronger for PGH with the largest CCN.
- Figure 7: Related to previous comment, is there an explanation for CAMS CCN overestimation or does it depend on the sites? Have you seen evidence that CAMS produces underestimated CCN?
- L379: Add "as shown by dashed lines" after "within a factor of 10".
- L381: The spread of data seems larger for MAG because it is on the lower end of the graph on a log-log scale. In fact, the spread of data for PGH spans more than 5000 cm-3 (at least for CAMS CCN), whereas the MAG data spread is ~ 2000 cm-3.
- L384: Some changes are Figure 7b reveals almost balanced contributions from SU and OM for SGP and PVC, OM dominance for PGH, and SU dominance for GRW and MAG.
- L387: PGH doesn't seem to be an exception. Perhaps, delete "except for PGH."
- Table 3 Caption:
- Perhaps refer to the text where you define observational measurement uncertainties and select the value of 40%. Is this uncertainty arbitrary? Depending on this value, the CAMS CCN can be within or outside the range of observational uncertainty.
- Can you add a brief description of how to interpret the 'bias' and NRMSE? For example, a value of 1 for 'bias factor' represents the best agreement.
- Perhaps the use of “bias factor” or “bias ratio” is more accurate?
- What does 'Q' in Q50, etc., refer to? If it stands for quartile, it should be Q1, Q2, Q3, and Q4, which correspond to P25, P50, P75, and P100 (P refers to percentile). - L390-395: This highlights a possible source of bias in reanalysis and GCMs.
- This bias might not be evident from total AOD (Figure 20, first row, third column in Inness et al., 2019), but the substantial positive difference in OM AOD is clearly evident in the third row, third column.
- This effect is more pronounced over China but can also be observed to a lesser extent over Europe and the Eastern USA (Figure 20 in Inness et al., 2019).
- It would greatly reinforce your findings to include brief comparisons with other studies that compare models and observations for these specific regions (assuming such studies exist). - Table 4: Including correlation analysis is helpful. A few questions:
- What is the number of data points for each site? I noticed that the total value in Table 4 is approximately half that in Table 3. Does this suggest that some sites (e.g., MAG) have a very low number of data points? Adding the number of data points as a new column to Table 4 would be informative.
- Additionally, could you consider showing the statistical significance (or p-value) for each correlation? There are look-up tables that provide p-values based on correlation and the number of data points (or degree of freedom). Incorporating this information would enhance the analysis, as the correlation value can be influenced by the number of data points. - L400: Perhaps rephrase as 'The correlation coefficient for all data points increases from 0.37 to 0.71...
- L401: Considering the previous point, it might not be appropriate to directly compare changes in correlation among different sites, as they have varying numbers of data points. Additionally, the change from 0.07 to 0.18 doesn't necessarily represent a substantial improvement due to the overall weak correlation. Nevertheless, the following discussion based on previous studies is insightful and worth retaining.
- L415-418: You can omit the first two sentences and place the last one at the end of the first sentence in L401.
- L420: You can provide a more accurate statement by saying, 'Since the reanalysis aerosol mass mixing ratios are generated through the assimilation of satellite AOD...
- L443: Change to something like 'The contribution of black carbon to CCN is significantly smaller than that of sulfate and organic matter…
- L444: You may consider deleting this line or providing a more specific and detailed explanation.
- L472: The latter part of this sentence is difficult to understand.
Technical Corrections:
- Ensure that acronyms and abbreviations are defined in their first occurrence in the manuscript. Some examples:
L8: Replace “RA” with “(CAMSRA)”. Also, “EAC4” is not clear to me.
L24: Define AERONET
L45: Define ECHAM-HAM
L50: Define POLDER
L59: Define MAC
L67: Change “Caliop” to “the Cloud-Aerosol Lidar with Orthogonal Polarization (CALIOP)”
L85: GCM should be defined here.
L104-105: Define LMD-Z, GEMS, MACC.
L143-144: Define MEGAN2.1 and MERRA-2
L201: Define HadGEM3-UKCA
Appendix B: Define “CN” and “NC”.
- Grammar and Typos:
L1-2: Consider rephrasing to correct the grammar, such as adding "and" before "a process."
L56: Change "related with" to "related to."
L91: Change "Deviations to" to "Deviations from."
L162: Change "on board of the Aqua" to "on board the Aqua."
L262: Change "radii is" to "radii are."
L278-279: Change "stay" and "decrease" to "stays" and "decreases."
L296: Change "where" to "were."
L299: Change "eruption on Iceland" to "eruption in Iceland."
Figure 3 caption: Change "latitudional" to "latitudinal" and "timeseries" to "time-series."
L331: Change "wind speed are" to "wind speeds are."
Figure 8 caption: Remove the extra "the" in "the the ARM."
L519: Change "is" to "are" in "A demonstration and a possible extension of the formula is given."
L527: Change "Thus study" to "This study."
L535: Change "office of science" to "Office of Science."
- Use hyphens when two words function together as an adjective before a noun. Some examples:
L5: Change “uncertainty reduced estimates” to “uncertainty-reduced estimates”.
L14: Change “ground based in-situ measurements” to “ground-based in-situ measurements”.
L38: Change “satellite retrieved AOD” to “satellite-retrieved AOD”.
L58: Change “CCN related aerosol” to “CCN-related aerosol”.
L81 and similar occurrences: Change “CCN relevant” to “CCN-relevant” throughout the manuscript.
L134: Change “latitude dependent e-folding time scale” to “latitude-dependent e-folding time scale”.
L191 and 419: Change “19 year long global CCN” to “19-year-long global CCN”
L353 and similar occurrences: Change “CAMS derived” to “CAMS-derived” and “quality assured” to “quality-assured” throughout the manuscript.
L359: Change “pyramid like” to “pyramid-like”.
L425: Change “cloud free” to “cloud-free”.
L493: Change “quality checked” to “quality-checked”.
- Be sure to be consistent about past tense and present tense throughout the manuscript. Example:
L375: Perhaps, change “ensured” to “ensure” in “We use daily means of quality-checked CCN measurements and ensured that”
- Ensure consistent spelling throughout the manuscript. Decide between British and American English spellings and stick to one style. Examples:
L125: Change “aging” to “ageing”.
L180: Change “parameterized” to “parameterised”.
L280 and similar occurrences: Change “analyzed” to “analysed”.
References:
- Che, H. C., Zhang, X. Y., Zhang, L., Wang, Y. Q., Zhang, Y. M., Shen, X. J., ... & Zhong, J. T. (2017). Prediction of size-resolved number concentration of cloud condensation nuclei and long-term measurements of their activation characteristics. Scientific reports, 7(1), 5819. https://doi.org/10.1038/s41598-017-05998-3
- Che, H., Stier, P., Watson-Parris, D., Gordon, H., & Deaconu, L. (2022). Source attribution of cloud condensation nuclei and their impact on stratocumulus clouds and radiation in the south-eastern Atlantic. Atmospheric Chemistry and Physics, 22(16), 10789-10807. https://doi.org/10.5194/acp-22-10789-2022
- Garrigues, S., Chimot, J., Ades, M., Inness, A., Flemming, J., Kipling, Z., ... & Agusti-Panareda, A. (2022). Monitoring multiple satellite aerosol optical depth (AOD) products within the Copernicus Atmosphere Monitoring Service (CAMS) data assimilation system. Atmospheric Chemistry and Physics, 22(22), 14657-14692. https://doi.org/10.5194/acp-22-14657-2022
- Haarig, M., Walser, A., Ansmann, A., Dollner, M., Althausen, D., Sauer, D., Farrell, D., and Weinzierl, B. (2019). Profiles of cloud condensation nuclei, dust mass concentration, and ice-nucleating-particle-relevant aerosol properties in the saharan air layer over barbados from polarization lidar and airborne in situ measurements. Atmospheric Chemistry and Physics, 19(22):13773–13788. https://doi.org/10.5194/acp-19-13773-2019
- Inness, , Ades, M., Agustí-Panareda, A., Barré, J., Benedictow, A., Blechschmidt, … & Suttie, M. (2019). The CAMS reanalysis of atmospheric composition, Atmos. Chem. Phys., 19, 3515–3556, https://doi.org/10.5194/acp-19-3515-2019 .
- Kapustin, V. N., Clarke, A. D., Shinozuka, Y., Howell, S., Brekhovskikh, V., Nakajima, T., & Higurashi, A. (2006). On the determination of a cloud condensation nuclei from satellite: Challenges and possibilities. Journal of Geophysical Research: Atmospheres, 111(D4). https://doi.org/10.1029/2004JD005527
- Shinozuka, Y., Clarke, A. D., Nenes, A., Jefferson, A., Wood, R., McNaughton, C. S., ... & Yoon, Y. J. (2015). The relationship between cloud condensation nuclei (CCN) concentration and light extinction of dried particles: indications of underlying aerosol processes and implications for satellite-based CCN estimates. Atmospheric Chemistry and Physics, 15(13), 7585-7604. https://doi.org/10.5194/acp-15-7585-2015
- Weinzierl, B., Ansmann, A., Prospero, J. M., Althausen, D., Benker,, Chouza, F., … & Walser, A. (2017). The saharan aerosol long-range transport and aerosol–cloud-interaction experiment: Overview and selected highlights. Bulletin of the American Meteorological Society, 98(7):1427 – 1451. https://doi.org/10.1175/BAMS-D-15-00142.1
Citation: https://doi.org/10.5194/essd-2023-172-RC2 - Your methodology is explained clearly. It would be great to elaborate on the exclusion of dust particles from the CCN analysis.
Karoline Block et al.
Data sets
Cloud condensation nuclei (CCN) numbers derived from CAMS reanalysis EAC4 (Version 1) Karoline Block https://doi.org/10.26050/WDCC/QUAERERE_CCNCAMS_v1
Karoline Block et al.
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 313 | 109 | 14 | 436 | 10 | 13 |
- HTML: 313
- PDF: 109
- XML: 14
- Total: 436
- BibTeX: 10
- EndNote: 13
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1