the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 22 Aug 2025
GCM clouds and actual clouds as seen from different space lidars: towards a long-term assessment of cloud representation in GCMs using lidar simulator
Abstract. In Earth's radiative budget, clouds play a central role but their representation in General Circulation Models (GCMs) remains a major source of uncertainty for climate projection. Here, we used spaceborne lidar observations to assess cloud distribution in the IPSL-CM6-LR model using the CFMIP Observation Simulator Package (COSP). We focused on the lidars onboard CALIPSO and AEOLUS satellites during 2006–2023 and 2018–2023. While CALIPSO has been widely used for GCMs evaluation, ALADIN was originally designed for wind profiling. However, studies have demonstrated its potential to retrieve reliable cloud profiles. A new module was developed to simulate AEOLUS observations within COSP-lidar, extending original implementations made for CALIPSO, including wavelength change (532 nm to 355 nm), viewing geometry (35° off-nadir) and specific parameters adjustments related to sensivity and resolution. We compared our simulations to 1-year observations for both instruments. Results show that AEOLUS observations can effectively evaluate clouds in GCMs, as it shows similar cloud fraction biases in IPSL-CM6-LR to those obtained with CALIPSO. Significant underestimations of low (up to 20 %) and high clouds in certain regions (e.g. warm pool) were re-assessed for this model. Sensitivity analyses highlighted the small role of instrument-specific parameters in COSP-lidar: viewing geometry, multiple scattering coefficient and cloud detection threshold (associated with wavelength and sensivity). This work lays the foundation for a consistent multi-decades evaluation of cloud representation using different lidar missions, and supports the integration of EarthCARE/ATLID in COSP-lidar for further model evaluation.
-
Notice on discussion status
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
-
Preprint
(6796 KB)
-
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(6796 KB) - Metadata XML
- BibTeX
- EndNote
- Final revised paper
Journal article(s) based on this preprint
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2025-3723', Anonymous Referee #1, 05 Sep 2025
- AC1: 'Reply on RC1', Marie-Laure Roussel, 26 Nov 2025
-
RC2: 'Comment on egusphere-2025-3723', Anonymous Referee #2, 15 Oct 2025
Review of “GCM clouds and actual clouds as seen from different space lidars: towards a long-term assessment of lloud representation in GCMs using lidar simulator”
by M.-L. Roussel et al.This paper analyses the potential use of the AEOLUS lidar cloud products for model evaluation. The authors modify the COSP/CALIPSO simulator to produce model equivalents of the AEOLUS lidar cloud products, and apply both simulators to the LMDZ model. The evaluation provides consistent results against CALIPSO and AEOLUS, providing confidence in the potential of AEOLUS cloud products for model evaluation. This is first step towards the challenge of merging these datasets with ATLID to provide a long-term record of cloud observations from spaceborne lidars.
The paper is well written and the topic is suitable for publication in GMD, but I believe the authors need to provide more evidence on the reasons why they made some methodological choices before it can be accepted for publication.MAJOR COMMENTS
Definition of cloud fraction in 3 layers. My understanding is that GOCCP low-, mid-, and high-level cloud fraction is based on pressure layers: low-level (P > 680 hPa), middle-level (440 < P < 680 hPa) and highest-level (P < 440 hPa). However, in this study height is used instead of pressure (Lines 120-122, Table 1). It is not clear if a different type of processing has been done for this study. This needs clarification.Section 2.3. I'm confused by this section. CALIPSO and AEOLUS overlap, so why not compare cloud retrievals during the overlapping period? That would avoid the impact of internal variability.
Use of daily-averaged input data for COSP calculations. There is no discussion on the impact of this choice. A better motivation of this choice is needed. It would be interesting to see differences between daily averages calculated from 3-hourly inputs and daily-averaged inputs. Are these differences small enough with respect to other sources of uncertainty considered?
I may be misinterpreting the results, but Figure 7h doesn't seem to match with the results shown in Figure 8. Figure 7h shows less AEOLUS cloud fraction nearly everywhere above between 4000m and 8000m altitude, but Figure 8d shows an excess of AEOLUS mid-level cloud with respect to CALIPSO.
SPECIFIC COMMENTS
L45: that is involved -> that participatesL70: inclinaison. Typo that appears many times in the manuscript. I'd recommend replacing all instances of 'inclination' with 'off-nadir pointing angle' to avoid confusion with the inclination of the satellite orbit.
L122: global median of high cloud cover -> median of global-mean high-cloud cover. There are other instances in the text, please correct.
L192: cloud detection threshold s. Capital S is used before in reference to the detection threshold. Please use consistent notation.
Table 2. Please use Greek letters for 'eta' and the wavelength 'l'.
Citation: https://doi.org/10.5194/egusphere-2025-3723-RC2 - AC2: 'Reply on RC2', Marie-Laure Roussel, 26 Nov 2025
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2025-3723', Anonymous Referee #1, 05 Sep 2025
General Comment:
This study compares the cloud representation within the LMDZ atmospheric model of the IPSL GCM using COSP simulations of two lidar instruments: CALIOP aboard CALIPSO, and ALADIN aboard AEOLUS. The latter being implemented with the COSP-lidar simulator for the first time here. The study aims to verify that AEOLUS observations are suitable to evaluate GCM cloudiness, as the instrument was originally designed for wind profiling. The overarching goal of the study is to enable a long term evaluation framework of GCM cloudiness using different spaceborn lidars through COSP, particularly in preparation for ATLID aboard EarthCare (which like ALADIN operates at 355nm), as CALIPSO (which has been extensively used for GCM evaluation through COSP retrievals) ceased operation in 2023.
Sensitivity testing of COSP-lidar parameters is performed to determine suitable settings for simulating AEOLUS retrievals, and those set out within previous literature are deemed appropriate. Results show that measurements from COSP-AEOLUS are suitable to assess GCM cloud coverage in a similar manner to those from COSP-CALIPSO. With discrepancies between the simulated retrievals attributable to instrumental differences.
The paper is comprehensive, well written, and lays the foundations for the future evaluation of GCM cloudiness through comparisons to observed clouds with forthcoming satellite products. My main comments ask for additional clarity around certain choices made in the study which aren't clear at present. Please see my specific comments.
Specific Comments:
Line 20: Including references to further works on cloud radiative feedbacks, particularly from older CMIP generations (CMIP3 Bony & Dufresne (2005), CMIP5 Vial et al. (2013)) would help contextualize how big the issue is.
Bony, S., & Dufresne, J.‐L. (2005). Marine boundary layer clouds at the heart of tropical cloud feedback uncertainties in climate models.
Geophysical Research Letters, 32(20), L20806.Vial, J., Dufresne, J. L., & Bony, S. (2013). On the interpretation of inter‐model spread in cmip5 climate sensitivity estimates. Climate Dynamics,
41(11–12), 3339–3362.Line 31: This paragraph is confused by the introduction of an objective in the paper in the middle of introducing COSP. I would suggest rounding out the paragraph on COSP and moving everything following "Extending the COSP-lidar..." to the beginning of the next paragraph.
Line 36: Introducing the overarching goal that this study contributes towards is important, however this message is interspersed throughout the introduction, causing a bit of confusion. The paragraph beginning on line 36 is around the overarching goal this study contributes to. The following paragraph beginning line 44 is around the LDMZ model specific to this study. Then the paragraph beginning line 51 begins with the objective of this study before again reiterating the long term goal. These need to be separate paragraphs for clarity, one being what is happening in this paper, the other being the long term goal this work contributes towards.
Line 82: The description of CALIPSO is excellent. However, the paragraphs beginning lines 82/89 are describing data selection for the experimental procedure which is detailed further in section 2.3. I would prefer this information to be in section 2.3 as the reader doesn't yet know why we are talking about these individual years yet.
Section 2.3: Multiple specific years and dates ranges are used in this section. A lot of effort has been put into making sure these two individual years are comparable, but no readily apparent justification is given as to why were only looking at these individual years. Please see the following specific comments.
Line 114: Why these years specifically? In section 2.1 it is stated that 2008 data from CALIPSO is used to for consistency with simulations, but why not simulate 2020, and compare to CALIPSO 2020 and AEOLUS 2020. I assume due to the lack of AMIP forcing data.
Further to the previous point, why then cant a comparison between the 2008-2018 CALIPSO climatology and 2018-2023 AEOLUS climatology be made, thus avoiding the need to compare individual years and worry too much about inter-annual variability?
Is only 2020 produced by Titus et al. (2025)? This would need to be stated. If so comparisons of AEOLUS 2020 to the CALIPSO 2008-2018 or CALIPSO 2020 could be made.
Line 117: So the effect of interannual variability is only assessed using the CALIPSO product and assumed to be the same for AEOLUS. Based on the results presented further in the manuscript these products are highly comparable and would thus be fine, but would be worth mentioning at this stage. A supplementary figure detailing where 2020 fits within the inter-annual variability of 2008-2018 would be good.
It would also be worth mentioning that 2008 and 2020 are both La Nina years in terms of comparing clouds in the pacific.
Line 194: The multiple scattering coefficient n is what is different between Figure 5 (b) and (d), not the inclination. Do you mean (a) and (b), or, (c) and (d)?
Line 331: the sentence beginning: "We have reassessed the significant underestimation of cloud fractions at low levels in the LMDZ model" should be moved to later in the conclusion, as the focus should be on the AEOLUS-CALIPSO comparison.
Discussion point: Only one GCM was evaluated upon here. How would results be different if a different GCM was evaluated upon?
Technical Corrections:
Title: should the last word of the title be the plural of simulator: ....using lidar simulators
Line 6: It would be good to specify that ALADIN is the instrument aboard AEOLUS
Line 10: Single "
Line 68: "layer height" is this a common variable? I would expect a mention of cloud coverage here as well
Line 181: The inclinaison used instead of inclination.
Line 185: The following section would then be 3.4?
Line 266: Shouldn't this be CALIPSO observations in 2008 and AEOLUS observations in 2020.
Line 293: Shouldn't this be Figure 7 (g)?
Line 321 and furthermore: Figure 8 doesn't have a panel (g) or (h).
Line 323: Table numbers appearing as ??.
Citation: https://doi.org/10.5194/egusphere-2025-3723-RC1 - AC1: 'Reply on RC1', Marie-Laure Roussel, 26 Nov 2025
-
RC2: 'Comment on egusphere-2025-3723', Anonymous Referee #2, 15 Oct 2025
Review of “GCM clouds and actual clouds as seen from different space lidars: towards a long-term assessment of lloud representation in GCMs using lidar simulator”
by M.-L. Roussel et al.This paper analyses the potential use of the AEOLUS lidar cloud products for model evaluation. The authors modify the COSP/CALIPSO simulator to produce model equivalents of the AEOLUS lidar cloud products, and apply both simulators to the LMDZ model. The evaluation provides consistent results against CALIPSO and AEOLUS, providing confidence in the potential of AEOLUS cloud products for model evaluation. This is first step towards the challenge of merging these datasets with ATLID to provide a long-term record of cloud observations from spaceborne lidars.
The paper is well written and the topic is suitable for publication in GMD, but I believe the authors need to provide more evidence on the reasons why they made some methodological choices before it can be accepted for publication.MAJOR COMMENTS
Definition of cloud fraction in 3 layers. My understanding is that GOCCP low-, mid-, and high-level cloud fraction is based on pressure layers: low-level (P > 680 hPa), middle-level (440 < P < 680 hPa) and highest-level (P < 440 hPa). However, in this study height is used instead of pressure (Lines 120-122, Table 1). It is not clear if a different type of processing has been done for this study. This needs clarification.Section 2.3. I'm confused by this section. CALIPSO and AEOLUS overlap, so why not compare cloud retrievals during the overlapping period? That would avoid the impact of internal variability.
Use of daily-averaged input data for COSP calculations. There is no discussion on the impact of this choice. A better motivation of this choice is needed. It would be interesting to see differences between daily averages calculated from 3-hourly inputs and daily-averaged inputs. Are these differences small enough with respect to other sources of uncertainty considered?
I may be misinterpreting the results, but Figure 7h doesn't seem to match with the results shown in Figure 8. Figure 7h shows less AEOLUS cloud fraction nearly everywhere above between 4000m and 8000m altitude, but Figure 8d shows an excess of AEOLUS mid-level cloud with respect to CALIPSO.
SPECIFIC COMMENTS
L45: that is involved -> that participatesL70: inclinaison. Typo that appears many times in the manuscript. I'd recommend replacing all instances of 'inclination' with 'off-nadir pointing angle' to avoid confusion with the inclination of the satellite orbit.
L122: global median of high cloud cover -> median of global-mean high-cloud cover. There are other instances in the text, please correct.
L192: cloud detection threshold s. Capital S is used before in reference to the detection threshold. Please use consistent notation.
Table 2. Please use Greek letters for 'eta' and the wavelength 'l'.
Citation: https://doi.org/10.5194/egusphere-2025-3723-RC2 - AC2: 'Reply on RC2', Marie-Laure Roussel, 26 Nov 2025
Peer review completion
Journal article(s) based on this preprint
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 1,899 | 73 | 27 | 1,999 | 35 | 30 |
- HTML: 1,899
- PDF: 73
- XML: 27
- Total: 1,999
- BibTeX: 35
- EndNote: 30
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Marie-Laure Roussel
Hélène Chepfer
Zacharie Titus
Marine Bonazzola
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(6796 KB) - Metadata XML
General Comment:
This study compares the cloud representation within the LMDZ atmospheric model of the IPSL GCM using COSP simulations of two lidar instruments: CALIOP aboard CALIPSO, and ALADIN aboard AEOLUS. The latter being implemented with the COSP-lidar simulator for the first time here. The study aims to verify that AEOLUS observations are suitable to evaluate GCM cloudiness, as the instrument was originally designed for wind profiling. The overarching goal of the study is to enable a long term evaluation framework of GCM cloudiness using different spaceborn lidars through COSP, particularly in preparation for ATLID aboard EarthCare (which like ALADIN operates at 355nm), as CALIPSO (which has been extensively used for GCM evaluation through COSP retrievals) ceased operation in 2023.
Sensitivity testing of COSP-lidar parameters is performed to determine suitable settings for simulating AEOLUS retrievals, and those set out within previous literature are deemed appropriate. Results show that measurements from COSP-AEOLUS are suitable to assess GCM cloud coverage in a similar manner to those from COSP-CALIPSO. With discrepancies between the simulated retrievals attributable to instrumental differences.
The paper is comprehensive, well written, and lays the foundations for the future evaluation of GCM cloudiness through comparisons to observed clouds with forthcoming satellite products. My main comments ask for additional clarity around certain choices made in the study which aren't clear at present. Please see my specific comments.
Specific Comments:
Line 20: Including references to further works on cloud radiative feedbacks, particularly from older CMIP generations (CMIP3 Bony & Dufresne (2005), CMIP5 Vial et al. (2013)) would help contextualize how big the issue is.
Bony, S., & Dufresne, J.‐L. (2005). Marine boundary layer clouds at the heart of tropical cloud feedback uncertainties in climate models.
Geophysical Research Letters, 32(20), L20806.
Vial, J., Dufresne, J. L., & Bony, S. (2013). On the interpretation of inter‐model spread in cmip5 climate sensitivity estimates. Climate Dynamics,
41(11–12), 3339–3362.
Line 31: This paragraph is confused by the introduction of an objective in the paper in the middle of introducing COSP. I would suggest rounding out the paragraph on COSP and moving everything following "Extending the COSP-lidar..." to the beginning of the next paragraph.
Line 36: Introducing the overarching goal that this study contributes towards is important, however this message is interspersed throughout the introduction, causing a bit of confusion. The paragraph beginning on line 36 is around the overarching goal this study contributes to. The following paragraph beginning line 44 is around the LDMZ model specific to this study. Then the paragraph beginning line 51 begins with the objective of this study before again reiterating the long term goal. These need to be separate paragraphs for clarity, one being what is happening in this paper, the other being the long term goal this work contributes towards.
Line 82: The description of CALIPSO is excellent. However, the paragraphs beginning lines 82/89 are describing data selection for the experimental procedure which is detailed further in section 2.3. I would prefer this information to be in section 2.3 as the reader doesn't yet know why we are talking about these individual years yet.
Section 2.3: Multiple specific years and dates ranges are used in this section. A lot of effort has been put into making sure these two individual years are comparable, but no readily apparent justification is given as to why were only looking at these individual years. Please see the following specific comments.
Line 114: Why these years specifically? In section 2.1 it is stated that 2008 data from CALIPSO is used to for consistency with simulations, but why not simulate 2020, and compare to CALIPSO 2020 and AEOLUS 2020. I assume due to the lack of AMIP forcing data.
Further to the previous point, why then cant a comparison between the 2008-2018 CALIPSO climatology and 2018-2023 AEOLUS climatology be made, thus avoiding the need to compare individual years and worry too much about inter-annual variability?
Is only 2020 produced by Titus et al. (2025)? This would need to be stated. If so comparisons of AEOLUS 2020 to the CALIPSO 2008-2018 or CALIPSO 2020 could be made.
Line 117: So the effect of interannual variability is only assessed using the CALIPSO product and assumed to be the same for AEOLUS. Based on the results presented further in the manuscript these products are highly comparable and would thus be fine, but would be worth mentioning at this stage. A supplementary figure detailing where 2020 fits within the inter-annual variability of 2008-2018 would be good.
It would also be worth mentioning that 2008 and 2020 are both La Nina years in terms of comparing clouds in the pacific.
Line 194: The multiple scattering coefficient n is what is different between Figure 5 (b) and (d), not the inclination. Do you mean (a) and (b), or, (c) and (d)?
Line 331: the sentence beginning: "We have reassessed the significant underestimation of cloud fractions at low levels in the LMDZ model" should be moved to later in the conclusion, as the focus should be on the AEOLUS-CALIPSO comparison.
Discussion point: Only one GCM was evaluated upon here. How would results be different if a different GCM was evaluated upon?
Technical Corrections:
Title: should the last word of the title be the plural of simulator: ....using lidar simulators
Line 6: It would be good to specify that ALADIN is the instrument aboard AEOLUS
Line 10: Single "
Line 68: "layer height" is this a common variable? I would expect a mention of cloud coverage here as well
Line 181: The inclinaison used instead of inclination.
Line 185: The following section would then be 3.4?
Line 266: Shouldn't this be CALIPSO observations in 2008 and AEOLUS observations in 2020.
Line 293: Shouldn't this be Figure 7 (g)?
Line 321 and furthermore: Figure 8 doesn't have a panel (g) or (h).
Line 323: Table numbers appearing as ??.