The effect of low-level thin arctic clouds on shortwave irradiance: evaluation of estimates from spaceborne passive imagery with aircraft observations

Chen, Hong; Schmidt, Sebastian; King, Michael D.; Wind, Galina; Bucholtz, Anthony; Reid, Elizabeth A.; Segal-Rozenhaimer, Michal; Smith, William L.; Taylor, Patrick C.; Kato, Seiji; Pilewskie, Peter

doi:https://doi.org/10.5194/amt-14-2673-2021

Articles | Volume 14, issue 4

https://doi.org/10.5194/amt-14-2673-2021

© Author(s) 2021. This work is distributed under
the Creative Commons Attribution 4.0 License.

https://doi.org/10.5194/amt-14-2673-2021

© Author(s) 2021. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 14, issue 4

Research article

|

07 Apr 2021

Research article |

| 07 Apr 2021

The effect of low-level thin arctic clouds on shortwave irradiance: evaluation of estimates from spaceborne passive imagery with aircraft observations

Hong Chen, Sebastian Schmidt, Michael D. King, Galina Wind, Anthony Bucholtz, Elizabeth A. Reid, Michal Segal-Rozenhaimer, William L. Smith, Patrick C. Taylor, Seiji Kato, and Peter Pilewskie

Download

Final revised paper (published on 07 Apr 2021)
Supplement to the final revised paper
Preprint (discussion started on 28 Oct 2019)

Interactive discussion

Status: closed

AC: Author comment | RC: Referee comment | SC: Short comment | EC: Editor comment

- Printer-friendly version

- Supplement

RC1: 'Review of Chen et al.', Anonymous Referee #1, 26 Nov 2019
- AC1: 'Response to Reviewer #1', Hong Chen, 20 Apr 2020
- AC4: 'Response to Reviewer #1 (annotated changes of manuscript)', Hong Chen, 20 Apr 2020
RC2: 'Review of Chen et al.', Anonymous Referee #2, 12 Dec 2019
- AC2: 'Response to Reviewer #2', Hong Chen, 20 Apr 2020
- AC5: 'Response to Reviewer #2 (annotated changes of manuscript)', Hong Chen, 20 Apr 2020
RC3: 'Review of Chen et al.', Anonymous Referee #3, 23 Dec 2019
- AC3: 'Response to Reviewer #3', Hong Chen, 20 Apr 2020
- AC6: 'Response to Reviewer #3 (annotated changes of manuscript)', Hong Chen, 20 Apr 2020

Peer-review completion

AR: Author's response | RR: Referee report | ED: Editor decision

AR by Hong Chen on behalf of the Authors (19 May 2020) Author's response Manuscript

ED: Referee Nomination & Report Request started (25 May 2020) by Manfred Wendisch

RR by Anonymous Referee #2 (08 Jun 2020)

Suggestions for revision or reasons for rejection

The authors put a lot of work into improving the manuscript! It is now more clear, what the analysis of the airborne measurements aims for and the results are presented more clearly. I also acknowledge the detailed uncertainty analysis. However, understanding now, what is done in the study, I see one major issue which needs to be addressed more carefully.

Based on the reviews, the authors now mention the difference between albedo in cloud-free (blue-sky) and cloudy (white sky) conditions. It is acknowledged, that the use of measurements in cloud-free conditions might induce some bias in the albedo.
On page 14 line 5 of the revised manuscript the authors conclude, that the blue-sky and white-sky albedo do not differ in their measurements. I think, this conclusion is made from a wrong comparison. Figure 12 shows ratios of modelled vs. measured upwelling irradiance for tau=0. This means, that the measurements at tau=0 again represent cloud-free conditions, which is similar to the albedo assumed in the simulations. As far as I understand, here you can not make any conclusion on the effect of blue-sky vs. white-sky albedo.

However, I actually see no need to apply a wrong albedo in the simulations. You have airborne measurements below clouds. Here you can derive an albedo for overcast conditions (white-sky) and implement this into the simulations. This would also avoid the complex derivation of the cloud-free surface albedo from parametrizing the camera observations. As carefully analysed by the authors, all observations show no cloud gaps, which makes the assumption of a white-sky albedo the first choice. As the differences between both albedo are not minor, I strongly suggest to consider this approach even when the parametrization of the cloud-free albedo might become redundant.
Or you put everything into a different context and ask: What surface albedo would be available from a MODIS albedo product when aiming for calculating cloud radiative effects? In that case, I would show the difference between assuming parametrized blue-sky albedo and directly measured white-sky albedo.

Minor comments:

P2, L17: add "effect on the downwelling (upwelling) irradiance". just to avoid misunderstandings with CRF.

P10, L35: slightly misleading: the "0911-above-cloud" case does not account for changes along flight track.

P11, L6: Just an idea: You could test the effect of assuming a constant SF vs. variable SF, when you do the same for the "below-cloud" case (use a fixed SF) and compare to the simulations when changes of surface albedo are considered.

P11, 15: Still the potential effects of the cirrus are not sufficiently discussed or quantified to my point of view. Your comparison may just coincidently match! You compare upward irradiance which for MODIS may assume a higher optical thickness than present (MODIS optical thickness will include both cirrus and low level cloud). At the same time, the measured downward irradiance in flight altitude is reduced but not the simulated downward Irradiance (no cirrus in the RTM). Both would add up for higher upwelling irradiance simulated with MODIS compared to the observations. This needs to be discussed! Describe how the cirrus will affect the irradiances and may be a short simulation will estimate the impact in W m-2.

P13, L2: You adjusted SF to match the cloud-free measurements and used 1640 nm wavelength. Using this wavelength, which is very sensitive to snow properties (snow grain size), and may not result in a correct broadband albedo. You mentioned, that at 860 nm the agreement was worse. So ice fraction might differ from you SF estimate. Therefore, I recommend to name this adjusted SF an "effective SF".

P14, L23: change into "approach that did help" ... later you show answers to your questions.

P15, L27-30: Here and in the intruduction the uncertainty of available and need for new cloud retrieval over snow surfaces is discussed. Some work into this direction has been done in recent years which should be mentioned in such a discussion:

Ehrlich, A., Bierwirth, E., Istomina, L., and Wendisch, M.: Combined retrieval of Arctic liquid water cloud and surface snow properties using airborne spectral solar remote sensing, Atmos. Meas. Tech., 10, 3215–3230, https://doi.org/10.5194/amt-10-3215-2017, 2017.

Rolland, P., and Liou, K. N. ( 2001), Surface variability effects on the remote sensing of thin cirrus optical and microphysical properties, J. Geophys. Res., 106( D19), 22965– 22977, doi:10.1029/2001JD900160.

Figure 7: To present a consistent data analysis, think about adding also histograms to Figure 7 (upward irradiance above cloud) similar to the comparison of the below-cloud case.

P31, There is something wrong with the figure. No number and caption.

Hide

RR by Anonymous Referee #1 (13 Jun 2020)

Suggestions for revision or reasons for rejection

General comments

This paper is a substantially revised version of a manuscript already reviewed in November 2019. The points raised by this reviewer were considered adequately in the revised version. In particular an uncertainty analysis was included that was missing in the first version.

The paper should be published after minor corrections. Please see specific comments below.

Specific comments

Page 2, line 17: “The radiative effect of clouds that were detected…was -40 W m-2 (-39 W m-2)….”. I think this statement is a leftover from the first version. These results are not discussed in the manuscript. Instead the effect of the undetected clouds should be quantified here.

Page 9, line 31: The flight altitude of 240 m does not fit to that given in Fig. 2 (134 m in accordance with the video information).

Page 11, line 25: I think there is an inconsistency regarding the FOV definition. With a 45° FOV you wouldn’t cover half of the hemispheric irradiance (isotropic), you need 90° FOV and your FOV diameter is 14 km when you fly at 7 km altitude. I assume that this was actually applied because the horizontal error bars in Fig. 7 seem to have the correct ±7 km size.

Page 12, line 14: “… except for the time period before 22:22:48”. This was statement was already noted in the first version. Do you mean 22:21:48? In any case the x-scales in Figs. 8 (a) and (b) are possibly wrong. Was there a 1 min break in the time series (22:23 is missing)? If so that should be indicated by an empty period or at least a vertical line.

Page 13, line 11: “Of all pixels along the flight leg with a MODIS-COD below the detection threshold of 0.5 (i.e., “clear”), 27% (highlighted in green) are actually cloudy where MODIS cloud detection algorithm identified as clear-sky.” This statement is unclear and somehow in contradiction with the Abstract information: “27% of clouds remained undetected”. So what was calculated? The fraction of undetected cloud / total cloud periods or the fraction of undetected cloud / total clear (MODIS) periods? Please compare with the statement in the first review and clarify.

Page 15, lines 14 and 16. There is little quantitative information the reader finds in the Conclusions. So the numbers you give should be accurate and have a clear meaning. The 40 Wm-2 given for the above cloud case was stated as 30 Wm-2 within the text (page 11, line 34). It refers to time periods where you think thin clouds were present. For the below cloud case the 45 Wm-2 is a mean of the whole measurement period during which clouds were occasionally undetected (Fig. 8 (c)). So the numbers are not comparable, as implied in the text.

Page 18, line 17 (and Fig. A3): Please check for consistency with the FOV definition discussed above.

Fig. 1, Fig. 2 (caption), Fig. 3 (caption) and Fig. 8: Please use consistent notations of longitudes and latitudes.

Fig. 2: “The diameter of the field of view was about 380 m” Please check again. Is this the diameter or the radius of the field of view?

Fig. 3 (a) and (b): Typo “Temperature”

Fig. 6: “…cloud optical thinness of…”?

Hide

ED: Reconsider after major revisions (15 Jun 2020) by Manfred Wendisch

AR by Hong Chen on behalf of the Authors (31 Aug 2020) Author's response Manuscript

ED: Referee Nomination & Report Request started (11 Sep 2020) by Manfred Wendisch

RR by Anonymous Referee #2 (13 Sep 2020)

ED: Publish as is (14 Sep 2020) by Manfred Wendisch

AR by Hong Chen on behalf of the Authors (14 Sep 2020) Manuscript

Post-review adjustments

AA: Author's adjustment | EA: Editor approval

AA by Hong Chen on behalf of the Authors (01 Apr 2021) Author's adjustment Manuscript

EA: Adjustments approved (01 Apr 2021) by Manfred Wendisch

Short summary

In this paper, we accessed the shortwave irradiance derived from MODIS cloud optical properties by using aircraft measurements. We developed a data aggregation technique to parameterize spectral surface albedo by snow fraction in the Arctic. We found that undetected clouds have the most significant impact on the imagery-derived irradiance. This study suggests that passive imagery cloud detection could be improved through a multi-pixel approach that would make it more dependable in the Arctic.