the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Measurement Report: Observed Increase in Southern Hemisphere Reflected Energy from Clouds During December 2020 and 2021
Abstract. Measured backscattered UV radiances at 388±1.5 nm are converted to Lambert Equivalent Reflectivity (LER) are from EPIC (Earth Polychromatic Imaging Camera) onboard the DSCOVR spacecraft (Deep Space Climate Observatory) orbiting about the Sun-Earth Lagrange-1 (L1) gravitational balance point. The average percent of reflected solar energy in the 388±1.5 nm band is 29.2 % of the global incident solar energy in that band. Maximum reflected 388 nm solar energy RSE, mostly from clouds, occurs during the summer solstice in each hemisphere, December in the Southern Hemisphere SH and June in the Northern Hemisphere NH. The global average RSE (90° S to 90° N) has a maximum in December and a minimum in June showing that the SH cloud reflected energy is greater than that in the NH. Backscattering from land and oceans at 388 nm is small since the average clear-sky reflectivity of the Earth’s surface free of snow and ice is about 0.05. Calculations of RSE based on the 388 nm LER show a 7 % increase during December 2020 in RSE at 40° S to 50° S when the backscattering angle BA was 178.05°, and 6 % at 30° S to 40° S in November 2021 when BA = 177.5° compared to previous years, 2015–2019, with a smaller BA. Comparison of 380 nm RSE at 40° S to 50° S during December 2020 from the low Earth polar-orbiting nadir mapper in the Ozone Mapping and Profiler Suite (OMPS-NM) near 13:30 local solar time suggests that there has been a 5 % increase in SH cloud reflection during December 2020 compared to previous years. This suggests that the observed increase by EPIC is mostly from an increase in cloud cover and not from enhanced backscatter. In the NH RSE values at large EPIC BA (177.5° in June 2020 and 178.2° in June 2021) between 30° N to 60° N show a percent decrease 4.8 % in RSE at 45° N during June 2021 and a 6 % increase during June 2020 at 55° N compared to the previous 4 years. This also suggests that the increase and decrease in RSE are probably related to changes in cloud cover and not backscatter angle effects. Annual integrals of percent reflected solar energy over complete years are almost constant at all latitudes.
This preprint has been withdrawn.
-
Withdrawal notice
This preprint has been withdrawn.
-
Preprint
(7800 KB)
Interactive discussion
Status: closed
-
RC1: 'Comment on acp-2022-481', Anonymous Referee #1, 23 Dec 2022
The authors analyze EPIC ultraviolet reflectance data, focusing on peaks in December 2020 and November 2021, and (I think) attempt to answer the question of whether these peaks are physical (due to clouds) or retrieval artifacts due to changing viewing angles. Based on differences between the Northern and Southern Hemisphere peaks and a comparison with an independent instrument, they conclude the December 2020 and November 2021 peaks are due to increased Southern Hemisphere cloudiness.
As detailed below, I am confused about the motivation for the paper and about whether the authors applied corrections to the EPIC data to account for the changing viewing angles. The paper is also poorly written and edited at present. I cannot recommend publication barring major revisions to the manuscript.
General comments:
A. Overall motivation: From the paper I remain unconvinced that December 2020 and 2021 had interesting anomalies that really need explaining. It would be helpful to show that the differences seen in those months are large compared to interannual variability in general (e.g., as measured by the interannual standard deviation of previous years).
The introduction and concluding paragraph frame the study as primarily motivated by the question of whether angular dependencies are substantially affecting EPIC measurements. I did not pick up on this motivation from the abstract or title of the paper.
The authors also claim they correct the EPIC data for changes in orbit in 2020 and 2021, although they do not describe how this is done. If they are able to correct for the orbit changes, then what added benefit is the analysis shown? Or is the analysis supposed to be a confirmation that the corrections indeed work and the differences seen after correction are due to cloud changes?
B. General sloppiness: Overall I found the paper incredibly hard to follow. I believe a large portion of this could be fixed by better organization and explanation on the part of the authors. The use of equations is particularly galling, with 9 equations stacked in a row with minimal explanation at one point and a tenth equation placed within a figure caption. There are also a large number of typos and grammar mistakes.
Specific comments:
Lines 21-22: Isn’t there an effect of eccentricity here too, as the Earth is closer to the Sun in December?
Lines 88-89: I don’t believe the fact that increased reflection was observed in 2020 and 2021 has been introduced. It might be good to show that here and give a sense of how large the anomalies are relative to other years.
Lines 115-116: What is the correction? As mentioned above, this seems fundamental to the paper’s motivation.
Equations 1-9: The equations could definitely be explained better. Maybe interspersing them within the text instead of just listing them all at once would be an improvement?
Eq. 5: LER is not defined. Should B be Bi?
Line 127: “and” should be “an”?
Figure 1: Why is data from 2019 excluded? Could you please explain the data gap.
Line 153: Isn’t this Figure 4?
Line 176: Could you explain how Figure 3 supports this?
Figures 4-8, 10-11: What is the y-axis showing? Are you reporting the amount of sunlight reflected in a given latitude band divided by the entire amount of sunlight received by the Earth? So the sum of everything should be ~30%?
Figure 5 bottom right panel: What does a negative value for percent of sunlight reflected mean?
Figure 5 legend: The red line should be labeled 2016-2017, not 2017-2017. The blue line should be labeled 2018-2019, not 2018-2919.
Line 176: There is no justification for this being an Appendix with only one sentence and one figure. Why not just bring it into the main text?
Figure 6: Maybe put error bars of 1 standard deviation around the 2015-2018 values to show if the December 2020 values are really that extreme?
Figure 6/Eq. 10: Are you defining an equation inside of a figure caption? Please define the equation within the main text.
Line 202: How does Figure 2 support this?
Citation: https://doi.org/10.5194/acp-2022-481-RC1 -
AC1: 'Reply on RC1', Jay Herman, 17 Jan 2023
I have made many changes, marked in green, in the text as shown in the supplement file both from referee comments and from re-reading the paper myself. I have also revised the figures in response to referee comments.
General comments:
- Overall motivation: From the paper I remain unconvinced that December 2020 and 2021 had interesting anomalies that really need explaining. It would be helpful to show that the differences seen in those months are large compared to interannual variability in general (e.g., as measured by the interannual standard deviation of previous years).
I have included error bars based on the standard deviations of previous years.
The introduction and concluding paragraph frame the study as primarily motivated by the question of whether angular dependencies are substantially affecting EPIC measurements. I did not pick up on this motivation from the abstract or title of the paper.
The authors also claim they correct the EPIC data for changes in orbit in 2020 and 2021, although they do not describe how this is done. If they are able to correct for the orbit changes, then what added benefit is the analysis shown? Or is the analysis supposed to be a confirmation that the corrections indeed work and the differences seen after correction are due to cloud changes?
After corrections for the change in viewable area when the viewing angles are large (near 180 degrees), the differences are still present.
- General sloppiness: Overall I found the paper incredibly hard to follow. I believe a large portion of this could be fixed by better organization and explanation on the part of the authors. The use of equations is particularly galling, with 9 equations stacked in a row with minimal explanation at one point and a tenth equation placed within a figure caption. There are also a large number of typos and grammar mistakes.
The equations are now separated by text
Hopefully, I have improved this aspect.
Specific comments:
Lines 21-22: Isn’t there an effect of eccentricity here too, as the Earth is closer to the Sun in December? Yes, the earth-sun distance is included
Lines 88-89: I don’t believe the fact that increased reflection was observed in 2020 and 2021 has been introduced. It might be good to show that here and give a sense of how large the anomalies are relative to other years. Later figures show that explicitly
Lines 115-116: What is the correction? As mentioned above, this seems fundamental to the paper’s motivation.
This is followed by a correction for the increased viewable area caused by the satellite’s orbit getting close to the Earth-Sun line in 2020-2021.
Added to the text
Equations 1-9: The equations could definitely be explained better. Maybe interspersing them within the text instead of just listing them all at once would be an improvement?
Done
Eq. 5: LER is not defined. Should B be Bi? Fixed
Line 127: “and” should be “an”? Fixed
Figure 1: Why is data from 2019 excluded? Could you please explain the data gap.
The satellite pointing system (gyroscope + momentum wheels) failed in 2019 and were replaced by the star tracker, external propulsion jets, and new software, which took almost a year to develop, test, and deploy. See the caption in Fig. 1
Line 153: Isn’t this Figure 4? Fixed
Line 176: Could you explain how Figure 3 supports this?
It should have been (Fig. 4 and its magnified version in Appendix Fig. A1)
Figures 4-8, 10-11: What is the y-axis showing? Are you reporting the amount of sunlight reflected in a given latitude band divided by the entire amount of sunlight received by the Earth? So the sum of everything should be ~30%?
Looking at Fig. 1 the global percent reflected of incident energy varies from 24% to 38% depending on the month of the year. The global reflected energy is dominated by the SH summer (December-January). In the following figures (4-8) and (10-11) the y-axis is the percent of total energy on the illuminated Earth disk.
The annual average is PSE = 29.2% of the incident solar energy including the surface contribution (Lines 118 to 120)
The 29.2% was stated in the original text, but the above has been added.
Figure 5 bottom right panel: What does a negative value for percent of sunlight reflected mean?
The original data are not negative (grey circles). The negative values are an artifact of the averaging technique I used, Loess(325pts). When I use Adjacent Averaging AA(325 pts) the values are not negative. The 60S-70S figure now uses AA.
Figure 5 legend: The red line should be labeled 2016-2017, not 2017-2017. The blue line should be labeled 2018-2019, not 2018-2919. Fixed. I apologize for the unnoticed typos.
Line 176: There is no justification for this being an Appendix with only one sentence and one figure. Why not just bring it into the main text?
I could put the figure in the main text, but the Appendix figure is the same as the one in the main text but with the time scale magnified. It has no additional information except the summer small SEV effect is easier to see.
Figure 6: Maybe put error bars of 1 standard deviation around the 2015-2018 values to show if the December 2020 values are really that extreme?
Error bars have been added
Figure 6/Eq. 10: Are you defining an equation inside of a figure caption? Please define the equation within the main text.
The equation was not in the Figure Caption. Some text has been added in front of Eq 10.
Line 202: How does Figure 2 support this?
Figure 2 does not support this. However, the explanation of the minimum over Africa supports Figure 2. Reflectivity at 388 nm is mostly from clouds with the background land having low reflectivity (less than 4% except for a small desert in Libya). Africa usually has almost no cloud cover per km2 compared to other regions, both land and oceans. Therefore, the 388 nm reflectivity is a minimum over Africa. In addition, the files are time stamped and can be matched with visible wavelength images such as the one shown below. Northern Africa is almost cloud-free. A sentence and reference have been added to this effect.
Africa has little cloud cover per km2 compared to other regions (e.g., https://epic.gsfc.nasa.gov/?date=2022-06-09) and the land surfaces are dark (LER < 0.05) at 388 nm (Herman et al., 2018b).
Herman, Jay, Guoyong Wen, Alexander Marshak, Karin Blank, Liang Huang, Alexander Cede, Nader Abuhassan, Matthew Kowalewski, Reduction in Earth Reflected Radiance during the Eclipse of 21 August 2017, Atmos. Meas. Tech., 11, 4373–4388, https://doi.org/10.5194/amt-11-4373-2018. 2018b.
Visible wavelengths showing almost cloud-free Noarthern Africa
-
AC1: 'Reply on RC1', Jay Herman, 17 Jan 2023
-
RC2: 'Comment on acp-2022-481', Anonymous Referee #2, 08 Jan 2023
Dear authors,
Wishes for a happy New Year,In this study by Herman et al. EPIC backscattered UV radiances at 388nm analyzed and they concluded that the observed increases in reflected solar radiation were attributed to increase in cloud cover and not to the decrease of the small satellite viewing angle from the Sun-Earth line (SEV). This argument is also supported by analyzing data from an independent platform (OMPS-NM). Changes in cloudiness is of great importance for climate change and for solar power generation, so a paper giving evidence of such changes it would be interesting. But, regarding the present manuscript there are many things to improved. Apart from the fact that the manuscript is hard to follow and needs improvements, there are parts in the analysis that are unclear.
I recommend this paper for publication. I think it will be improved if the authors can address the following major comments:
1) Make the scientific questions clear. The goals of the study that are stated in the introduction are not related with the final conclusions. An apparent increased in back reflection observed by EPIC was anticipated in 2020 and 2021 as it is stated in the introduction and a correction for the changing orbit in 2020-2021 has been applied as it is stated in the last sentence of the introduction and in section 2. So, the performance of the correction is tested compare to previous years? And the changes of the corrected values of 2020-2021 compared to previous years are explained by changes in cloud cover?
2) Try to clarify and put the scientific questions and the related conclusions into perspective and discuss the importance of the results for the community. Studies dealing with changes in cloudiness are of great importance for climate and for the solar energy sector too.
3) From my point of view, the main dataset was EPIC measurements, and the OMPS-NM were used additionally for supporting the EPIC results and only for SH. So, make also clear the scientific question being addressed using the OMPS-NM data.
4) The manuscript is hard to follow and essential parts of the analysis are unclear due to this. I suggest the following changes:
- use different sections for data and methods
- in methods, it was hard to follow the equations presented. I suggest to explain step by step the equations, by explaining every symbol (not all symbols were explained e.g LER or Gi) when they first appear. A figure illustrating the problem geometry it would be also helpful.
2) Specific comments
Lines 24: My opinion is that it is missing from the abstract a link explain why enhanced backscatter effects are anticipated for backscatter angles ~178 for 2020 and 2021 and this “problem” first appears in line 31. The relevant discussion is provided by the authors in the introduction (lines 45-51), but I think that it would be helpful to better understand the concluded remarks if a relevant sentence included in the abstract too.
Lines 52-59: If I understand correctly the goal of the study wasn’t to compare (line 55) EPIC and OMPS-NM data, but to use OMPS-NM data to support EPIC results. Please clarify better and also in this paragraph only 388nm are given, while OMPS data correspond to different wavelengths.
Lines 82-83: In previous paragraph (lines 78-80) it is stated that the observed UV reflectivity comes mostly from clouds. Please, try not to repeat and be consistent with previous statements.
Lines 115-116: I strongly recommend here to explain in more detail the correction applied and move equation given in lines 167-168 and 174 here.
Lines 117-119: I miss here the connection to the EPIC measurements. The EPIC measured backscatter UV radiances at 388nm are converted to LER as mentioned in the abstract, but nowhere up to now in the main text this description is given and for the first time appeared in eq (5) with a different symbol LER without explanation.
Lines 123-129: I cannot see an obvious increase in PSE in Fig 1. Also, is there any comment for the gap for the last ~6months of 2019?
Lines 125-126: see 3rd specific comment
Lines 151-157: This description is for Fig 4 and not 3
Lines 158-164: This paragraph fits better after the description of fig 4
Line 174: RA or AC the correction? And 2019 and not 2018?
Line 180: Is Fig4 smoothed like Fig 1B?
Line 234: There is no blue curve at fig 11.
3) Technical corrections
Line 19: RSE corresponds to reflected energy for different wavelengths, so please change the sentence accordingly.
Line 78: TOA acronym is introduced here, please put it inside parenthesis and in general check acronym throughout the manuscript.
Line 85: OMPS-NM instead of OMPS
Line 86: BA first appears in main text, please explain symbol
Line 99: OMPS-NM instead of NPP-NM
Figure 1: The y-axis doesn’t have PSE label
Figure 2: Please correct y-axis, P(t), Δθ
Figure 3: Remove the second “The ratio” from the beginning of the legend. Additionally, it is 2019 and not 2019? The same for fig 9.
Figure 4: please replace “energy in the PSE” with energy PSE in the and correct y-axis, the same in fig 8.
Figure 5: Please correct y-axis, red line is for 2016-2017?
Figure 6: Please use the same color for lines and legends, here make Dec 2020 the same color with line. Additionally, is it 2018 or 2019?
Figure 7: Please correct y-axis.
Table 2: 2018 or 2019? Also Latitude in degrees.
Line249: in Fig 10 instead of 11.
Line 259: OMPS-NM instead of OMI-NM
Citation: https://doi.org/10.5194/acp-2022-481-RC2 -
AC2: 'Reply on RC2', Jay Herman, 17 Jan 2023
- Make the scientific questions clear. The goals of the study that are stated in the introduction are not related with the final conclusions. An apparent increased in back reflection observed by EPIC was anticipated in 2020 and 2021 as it is stated in the introduction and a correction for the changing orbit in 2020-2021 has been applied as it is stated in the last sentence of the introduction and in section 2. So, the performance of the correction is tested compare to previous years? And the changes of the corrected values of 2020-2021 compared to previous years are explained by changes in cloud cover?
As now more clearly stated in the paper, the differences in cloud reflected energy still appear after the correction for changing observing area is 2020 and 2021 compared to previous years. The change is associated with small SEV angles.
- Try to clarify and put the scientific questions and the related conclusions into perspective and discuss the importance of the results for the community. Studies dealing with changes in cloudiness are of great importance for climate and for the solar energy sector too.
The key here is that the annual integral of cloud reflected energy at 388 nm was almost constant as stated in the paper even though there was a summer increase at Southern latitudes.
- From my point of view, the main dataset was EPIC measurements, and the OMPS-NM were used additionally for supporting the EPIC results and only for SH. So, make also clear the scientific question being addressed using the OMPS-NM data.
I have further addressed this problem
4) The manuscript is hard to follow and essential parts of the analysis are unclear due to this. I suggest the following changes:
- use different sections for data and methods.
Hopefully, the new version has fixed this
- in methods, it was hard to follow the equations presented. I suggest to explain step by step the equations, by explaining every symbol (not all symbols were explained e.g LER or Gi) when they first appear. A figure illustrating the problem geometry it would be also helpful.
The figure for this is pretty standard, namely the area of a spherical cap. I just took the difference between two spherical caps, which gives the area of a latitude band.
2) Specific comments
Lines 24: My opinion is that it is missing from the abstract a link explain why enhanced backscatter effects are anticipated for backscatter angles ~178 for 2020 and 2021 and this “problem” first appears in line 31. The relevant discussion is provided by the authors in the introduction (lines 45-51), but I think that it would be helpful to better understand the concluded remarks if a relevant sentence included in the abstract too.
I have added text to the abstract. “The non-repeating tilted Lissajous orbit about L1 has time varying backscatter angles BA in the Earth’s atmosphere that varied between 176O to 174O during 2015 to 2019 and increased to BA near 178O in 2020 and 2021. Previous studies have shown that backscattering from the Earth’s surface in the visible and near infrared wavelengths increases at high backscatter angles.”
Lines 52-59: If I understand correctly the goal of the study wasn’t to compare (line 55) EPIC and OMPS-NM data, but to use OMPS-NM data to support EPIC results. Please clarify better and also in this paragraph only 388nm are given, while OMPS data correspond to different wavelengths.
Reference to the different wavelength for OMPS-NM has been added and a sentence added, “This paper uses the 380±0.55 nm LER data from the downward looking spatial mapper OMPS-NM to validate the observations from EPIC.”
Lines 82-83: In previous paragraph (lines 78-80) it is stated that the observed UV reflectivity comes mostly from clouds. Please, try not to repeat and be consistent with previous statements.
Fixed
Lines 115-116: I strongly recommend here to explain in more detail the correction applied and move equation given in lines 167-168 and 174 here. Fixed
Lines 117-119: I miss here the connection to the EPIC measurements. The EPIC measured backscatter UV radiances at 388nm are converted to LER as mentioned in the abstract, but nowhere up to now in the main text this description is given and for the first time appeared in eq (5) with a different symbol LER without explanation.
Typo: It should all be LER
The following has been added in the introduction: “Measured backscattered UV radiances at 388±1.5 nm are converted to Lambert Equivalent Reflectivity (LER) as a function of latitude, longitude, and time (Herman et al. 2018a).”
Lines 123-129: I cannot see an obvious increase in PSE in Fig 1. Also, is there any comment for the gap for the last ~6months of 2019?
The increase in the global average is small compared to that in the SH because the NH decrease cancels a portion of the SH increase.
The global percent reflected energy PSE appears to slightly increase during December 2020 (SEV = 1.95O) (Table 1) compared to the average from previous years (2015-2019) when the SEV angles are larger (Fig.1). The effect during the summer is larger at southern latitudes and decreased during summer (June) at northern latitudes. Sections 2.1 (SH) and 2.2 (NH) will show which latitude bands contribute to the change in reflected energy during 2020 – 2021 when the SEV angles are smallest.
Yes. Failure of the gyroscopes in June 2019 that were replaced with a star tracker and gas-jet propulsion. Normal operation was resumed in February 2020. The explanation is now in the text.
Lines 125-126: see 3rd specific comment
Lines 151-157: This description is for Fig 4 and not 3
Fixed
Lines 158-164: This paragraph fits better after the description of fig 4
Moved
Line 174: RA or AC the correction? And 2019 and not 2018? RA. AC has been removed
RA = Area in 2020/<Area 2015-2019>, where <…> denotes average
(10)
The correction is RA the ratio of the area viewed when SEV is small compared to when SEV was larger 4 – 6 degrees.
Line 180: Is Fig4 smoothed like Fig 1B? Not smoothed to that extent. It is a weekly average as described in the text and now specified in the caption
Line 234: There is no blue curve at fig 11.
In Fig.5 the monthly averages are June to June starting in June 2015 and ending in June 2019. The blue curve is June 2018 to June 2019. The data then starts again in June 2020 to June 2021. In Figure 11, they are December 2015 to December 2018. The blue curve would have been December 2018 to December 2019 but the period July 2019 to February 2020 is missing data. Fig 11 data starts again in December 2020 through December 2021.
I changed to figure caption to read “Fig. 11 Monthly average of NH annual time series of percent reflected energy at 388±1.5 nm for 4 years, 2015 – 2021. The period 2018-2019 is not shown since the data are missing starting on 28 June 2019 to February 2020.”
3) Technical corrections
Line 19: RSE corresponds to reflected energy for different wavelengths, so please change the sentence accordingly. For this paper, RSE applies only to 388±1.5 nm
Line 78: TOA acronym is introduced here, please put it inside parenthesis and in general check acronym throughout the manuscript. (TOA) done
Line 85: OMPS-NM instead of OMPS Fixed
Line 86: BA first appears in main text, please explain symbol Done
Line 99: OMPS-NM instead of NPP-NM Fixed
Figure 1: The y-axis doesn’t have PSE label Fixed
Figure 2: Please correct y-axis, P(t), Δθ Should be PSE (Percent)
Figure 3: Remove the second “The ratio” from the beginning of the legend. Additionally, it is 2019 and not 2019? The same for fig 9. Fixed and now references Eq. 10.
Figure 4: please replace “energy in the PSE” with energy PSE in the and correct y-axis, the same in fig 8. Fixed
Figure 5: Please correct y-axis, red line is for 2016-2017? Fixed
Figure 6: Please use the same color for lines and legends, here make Dec 2020 the same color with line. Additionally, is it 2018 or 2019? Fixed, and it is 2018, since December 2019 is missing
Figure 7: Please correct y-axis. Fixed
Table 2: 2018 or 2019? Also, Latitude in degrees. <2015 – 2018> degrees fixed
Line249: in Fig 10 instead of 11. Fixed
Line 259: OMPS-NM instead of OMI-NM Fixed
-
AC2: 'Reply on RC2', Jay Herman, 17 Jan 2023
Interactive discussion
Status: closed
-
RC1: 'Comment on acp-2022-481', Anonymous Referee #1, 23 Dec 2022
The authors analyze EPIC ultraviolet reflectance data, focusing on peaks in December 2020 and November 2021, and (I think) attempt to answer the question of whether these peaks are physical (due to clouds) or retrieval artifacts due to changing viewing angles. Based on differences between the Northern and Southern Hemisphere peaks and a comparison with an independent instrument, they conclude the December 2020 and November 2021 peaks are due to increased Southern Hemisphere cloudiness.
As detailed below, I am confused about the motivation for the paper and about whether the authors applied corrections to the EPIC data to account for the changing viewing angles. The paper is also poorly written and edited at present. I cannot recommend publication barring major revisions to the manuscript.
General comments:
A. Overall motivation: From the paper I remain unconvinced that December 2020 and 2021 had interesting anomalies that really need explaining. It would be helpful to show that the differences seen in those months are large compared to interannual variability in general (e.g., as measured by the interannual standard deviation of previous years).
The introduction and concluding paragraph frame the study as primarily motivated by the question of whether angular dependencies are substantially affecting EPIC measurements. I did not pick up on this motivation from the abstract or title of the paper.
The authors also claim they correct the EPIC data for changes in orbit in 2020 and 2021, although they do not describe how this is done. If they are able to correct for the orbit changes, then what added benefit is the analysis shown? Or is the analysis supposed to be a confirmation that the corrections indeed work and the differences seen after correction are due to cloud changes?
B. General sloppiness: Overall I found the paper incredibly hard to follow. I believe a large portion of this could be fixed by better organization and explanation on the part of the authors. The use of equations is particularly galling, with 9 equations stacked in a row with minimal explanation at one point and a tenth equation placed within a figure caption. There are also a large number of typos and grammar mistakes.
Specific comments:
Lines 21-22: Isn’t there an effect of eccentricity here too, as the Earth is closer to the Sun in December?
Lines 88-89: I don’t believe the fact that increased reflection was observed in 2020 and 2021 has been introduced. It might be good to show that here and give a sense of how large the anomalies are relative to other years.
Lines 115-116: What is the correction? As mentioned above, this seems fundamental to the paper’s motivation.
Equations 1-9: The equations could definitely be explained better. Maybe interspersing them within the text instead of just listing them all at once would be an improvement?
Eq. 5: LER is not defined. Should B be Bi?
Line 127: “and” should be “an”?
Figure 1: Why is data from 2019 excluded? Could you please explain the data gap.
Line 153: Isn’t this Figure 4?
Line 176: Could you explain how Figure 3 supports this?
Figures 4-8, 10-11: What is the y-axis showing? Are you reporting the amount of sunlight reflected in a given latitude band divided by the entire amount of sunlight received by the Earth? So the sum of everything should be ~30%?
Figure 5 bottom right panel: What does a negative value for percent of sunlight reflected mean?
Figure 5 legend: The red line should be labeled 2016-2017, not 2017-2017. The blue line should be labeled 2018-2019, not 2018-2919.
Line 176: There is no justification for this being an Appendix with only one sentence and one figure. Why not just bring it into the main text?
Figure 6: Maybe put error bars of 1 standard deviation around the 2015-2018 values to show if the December 2020 values are really that extreme?
Figure 6/Eq. 10: Are you defining an equation inside of a figure caption? Please define the equation within the main text.
Line 202: How does Figure 2 support this?
Citation: https://doi.org/10.5194/acp-2022-481-RC1 -
AC1: 'Reply on RC1', Jay Herman, 17 Jan 2023
I have made many changes, marked in green, in the text as shown in the supplement file both from referee comments and from re-reading the paper myself. I have also revised the figures in response to referee comments.
General comments:
- Overall motivation: From the paper I remain unconvinced that December 2020 and 2021 had interesting anomalies that really need explaining. It would be helpful to show that the differences seen in those months are large compared to interannual variability in general (e.g., as measured by the interannual standard deviation of previous years).
I have included error bars based on the standard deviations of previous years.
The introduction and concluding paragraph frame the study as primarily motivated by the question of whether angular dependencies are substantially affecting EPIC measurements. I did not pick up on this motivation from the abstract or title of the paper.
The authors also claim they correct the EPIC data for changes in orbit in 2020 and 2021, although they do not describe how this is done. If they are able to correct for the orbit changes, then what added benefit is the analysis shown? Or is the analysis supposed to be a confirmation that the corrections indeed work and the differences seen after correction are due to cloud changes?
After corrections for the change in viewable area when the viewing angles are large (near 180 degrees), the differences are still present.
- General sloppiness: Overall I found the paper incredibly hard to follow. I believe a large portion of this could be fixed by better organization and explanation on the part of the authors. The use of equations is particularly galling, with 9 equations stacked in a row with minimal explanation at one point and a tenth equation placed within a figure caption. There are also a large number of typos and grammar mistakes.
The equations are now separated by text
Hopefully, I have improved this aspect.
Specific comments:
Lines 21-22: Isn’t there an effect of eccentricity here too, as the Earth is closer to the Sun in December? Yes, the earth-sun distance is included
Lines 88-89: I don’t believe the fact that increased reflection was observed in 2020 and 2021 has been introduced. It might be good to show that here and give a sense of how large the anomalies are relative to other years. Later figures show that explicitly
Lines 115-116: What is the correction? As mentioned above, this seems fundamental to the paper’s motivation.
This is followed by a correction for the increased viewable area caused by the satellite’s orbit getting close to the Earth-Sun line in 2020-2021.
Added to the text
Equations 1-9: The equations could definitely be explained better. Maybe interspersing them within the text instead of just listing them all at once would be an improvement?
Done
Eq. 5: LER is not defined. Should B be Bi? Fixed
Line 127: “and” should be “an”? Fixed
Figure 1: Why is data from 2019 excluded? Could you please explain the data gap.
The satellite pointing system (gyroscope + momentum wheels) failed in 2019 and were replaced by the star tracker, external propulsion jets, and new software, which took almost a year to develop, test, and deploy. See the caption in Fig. 1
Line 153: Isn’t this Figure 4? Fixed
Line 176: Could you explain how Figure 3 supports this?
It should have been (Fig. 4 and its magnified version in Appendix Fig. A1)
Figures 4-8, 10-11: What is the y-axis showing? Are you reporting the amount of sunlight reflected in a given latitude band divided by the entire amount of sunlight received by the Earth? So the sum of everything should be ~30%?
Looking at Fig. 1 the global percent reflected of incident energy varies from 24% to 38% depending on the month of the year. The global reflected energy is dominated by the SH summer (December-January). In the following figures (4-8) and (10-11) the y-axis is the percent of total energy on the illuminated Earth disk.
The annual average is PSE = 29.2% of the incident solar energy including the surface contribution (Lines 118 to 120)
The 29.2% was stated in the original text, but the above has been added.
Figure 5 bottom right panel: What does a negative value for percent of sunlight reflected mean?
The original data are not negative (grey circles). The negative values are an artifact of the averaging technique I used, Loess(325pts). When I use Adjacent Averaging AA(325 pts) the values are not negative. The 60S-70S figure now uses AA.
Figure 5 legend: The red line should be labeled 2016-2017, not 2017-2017. The blue line should be labeled 2018-2019, not 2018-2919. Fixed. I apologize for the unnoticed typos.
Line 176: There is no justification for this being an Appendix with only one sentence and one figure. Why not just bring it into the main text?
I could put the figure in the main text, but the Appendix figure is the same as the one in the main text but with the time scale magnified. It has no additional information except the summer small SEV effect is easier to see.
Figure 6: Maybe put error bars of 1 standard deviation around the 2015-2018 values to show if the December 2020 values are really that extreme?
Error bars have been added
Figure 6/Eq. 10: Are you defining an equation inside of a figure caption? Please define the equation within the main text.
The equation was not in the Figure Caption. Some text has been added in front of Eq 10.
Line 202: How does Figure 2 support this?
Figure 2 does not support this. However, the explanation of the minimum over Africa supports Figure 2. Reflectivity at 388 nm is mostly from clouds with the background land having low reflectivity (less than 4% except for a small desert in Libya). Africa usually has almost no cloud cover per km2 compared to other regions, both land and oceans. Therefore, the 388 nm reflectivity is a minimum over Africa. In addition, the files are time stamped and can be matched with visible wavelength images such as the one shown below. Northern Africa is almost cloud-free. A sentence and reference have been added to this effect.
Africa has little cloud cover per km2 compared to other regions (e.g., https://epic.gsfc.nasa.gov/?date=2022-06-09) and the land surfaces are dark (LER < 0.05) at 388 nm (Herman et al., 2018b).
Herman, Jay, Guoyong Wen, Alexander Marshak, Karin Blank, Liang Huang, Alexander Cede, Nader Abuhassan, Matthew Kowalewski, Reduction in Earth Reflected Radiance during the Eclipse of 21 August 2017, Atmos. Meas. Tech., 11, 4373–4388, https://doi.org/10.5194/amt-11-4373-2018. 2018b.
Visible wavelengths showing almost cloud-free Noarthern Africa
-
AC1: 'Reply on RC1', Jay Herman, 17 Jan 2023
-
RC2: 'Comment on acp-2022-481', Anonymous Referee #2, 08 Jan 2023
Dear authors,
Wishes for a happy New Year,In this study by Herman et al. EPIC backscattered UV radiances at 388nm analyzed and they concluded that the observed increases in reflected solar radiation were attributed to increase in cloud cover and not to the decrease of the small satellite viewing angle from the Sun-Earth line (SEV). This argument is also supported by analyzing data from an independent platform (OMPS-NM). Changes in cloudiness is of great importance for climate change and for solar power generation, so a paper giving evidence of such changes it would be interesting. But, regarding the present manuscript there are many things to improved. Apart from the fact that the manuscript is hard to follow and needs improvements, there are parts in the analysis that are unclear.
I recommend this paper for publication. I think it will be improved if the authors can address the following major comments:
1) Make the scientific questions clear. The goals of the study that are stated in the introduction are not related with the final conclusions. An apparent increased in back reflection observed by EPIC was anticipated in 2020 and 2021 as it is stated in the introduction and a correction for the changing orbit in 2020-2021 has been applied as it is stated in the last sentence of the introduction and in section 2. So, the performance of the correction is tested compare to previous years? And the changes of the corrected values of 2020-2021 compared to previous years are explained by changes in cloud cover?
2) Try to clarify and put the scientific questions and the related conclusions into perspective and discuss the importance of the results for the community. Studies dealing with changes in cloudiness are of great importance for climate and for the solar energy sector too.
3) From my point of view, the main dataset was EPIC measurements, and the OMPS-NM were used additionally for supporting the EPIC results and only for SH. So, make also clear the scientific question being addressed using the OMPS-NM data.
4) The manuscript is hard to follow and essential parts of the analysis are unclear due to this. I suggest the following changes:
- use different sections for data and methods
- in methods, it was hard to follow the equations presented. I suggest to explain step by step the equations, by explaining every symbol (not all symbols were explained e.g LER or Gi) when they first appear. A figure illustrating the problem geometry it would be also helpful.
2) Specific comments
Lines 24: My opinion is that it is missing from the abstract a link explain why enhanced backscatter effects are anticipated for backscatter angles ~178 for 2020 and 2021 and this “problem” first appears in line 31. The relevant discussion is provided by the authors in the introduction (lines 45-51), but I think that it would be helpful to better understand the concluded remarks if a relevant sentence included in the abstract too.
Lines 52-59: If I understand correctly the goal of the study wasn’t to compare (line 55) EPIC and OMPS-NM data, but to use OMPS-NM data to support EPIC results. Please clarify better and also in this paragraph only 388nm are given, while OMPS data correspond to different wavelengths.
Lines 82-83: In previous paragraph (lines 78-80) it is stated that the observed UV reflectivity comes mostly from clouds. Please, try not to repeat and be consistent with previous statements.
Lines 115-116: I strongly recommend here to explain in more detail the correction applied and move equation given in lines 167-168 and 174 here.
Lines 117-119: I miss here the connection to the EPIC measurements. The EPIC measured backscatter UV radiances at 388nm are converted to LER as mentioned in the abstract, but nowhere up to now in the main text this description is given and for the first time appeared in eq (5) with a different symbol LER without explanation.
Lines 123-129: I cannot see an obvious increase in PSE in Fig 1. Also, is there any comment for the gap for the last ~6months of 2019?
Lines 125-126: see 3rd specific comment
Lines 151-157: This description is for Fig 4 and not 3
Lines 158-164: This paragraph fits better after the description of fig 4
Line 174: RA or AC the correction? And 2019 and not 2018?
Line 180: Is Fig4 smoothed like Fig 1B?
Line 234: There is no blue curve at fig 11.
3) Technical corrections
Line 19: RSE corresponds to reflected energy for different wavelengths, so please change the sentence accordingly.
Line 78: TOA acronym is introduced here, please put it inside parenthesis and in general check acronym throughout the manuscript.
Line 85: OMPS-NM instead of OMPS
Line 86: BA first appears in main text, please explain symbol
Line 99: OMPS-NM instead of NPP-NM
Figure 1: The y-axis doesn’t have PSE label
Figure 2: Please correct y-axis, P(t), Δθ
Figure 3: Remove the second “The ratio” from the beginning of the legend. Additionally, it is 2019 and not 2019? The same for fig 9.
Figure 4: please replace “energy in the PSE” with energy PSE in the and correct y-axis, the same in fig 8.
Figure 5: Please correct y-axis, red line is for 2016-2017?
Figure 6: Please use the same color for lines and legends, here make Dec 2020 the same color with line. Additionally, is it 2018 or 2019?
Figure 7: Please correct y-axis.
Table 2: 2018 or 2019? Also Latitude in degrees.
Line249: in Fig 10 instead of 11.
Line 259: OMPS-NM instead of OMI-NM
Citation: https://doi.org/10.5194/acp-2022-481-RC2 -
AC2: 'Reply on RC2', Jay Herman, 17 Jan 2023
- Make the scientific questions clear. The goals of the study that are stated in the introduction are not related with the final conclusions. An apparent increased in back reflection observed by EPIC was anticipated in 2020 and 2021 as it is stated in the introduction and a correction for the changing orbit in 2020-2021 has been applied as it is stated in the last sentence of the introduction and in section 2. So, the performance of the correction is tested compare to previous years? And the changes of the corrected values of 2020-2021 compared to previous years are explained by changes in cloud cover?
As now more clearly stated in the paper, the differences in cloud reflected energy still appear after the correction for changing observing area is 2020 and 2021 compared to previous years. The change is associated with small SEV angles.
- Try to clarify and put the scientific questions and the related conclusions into perspective and discuss the importance of the results for the community. Studies dealing with changes in cloudiness are of great importance for climate and for the solar energy sector too.
The key here is that the annual integral of cloud reflected energy at 388 nm was almost constant as stated in the paper even though there was a summer increase at Southern latitudes.
- From my point of view, the main dataset was EPIC measurements, and the OMPS-NM were used additionally for supporting the EPIC results and only for SH. So, make also clear the scientific question being addressed using the OMPS-NM data.
I have further addressed this problem
4) The manuscript is hard to follow and essential parts of the analysis are unclear due to this. I suggest the following changes:
- use different sections for data and methods.
Hopefully, the new version has fixed this
- in methods, it was hard to follow the equations presented. I suggest to explain step by step the equations, by explaining every symbol (not all symbols were explained e.g LER or Gi) when they first appear. A figure illustrating the problem geometry it would be also helpful.
The figure for this is pretty standard, namely the area of a spherical cap. I just took the difference between two spherical caps, which gives the area of a latitude band.
2) Specific comments
Lines 24: My opinion is that it is missing from the abstract a link explain why enhanced backscatter effects are anticipated for backscatter angles ~178 for 2020 and 2021 and this “problem” first appears in line 31. The relevant discussion is provided by the authors in the introduction (lines 45-51), but I think that it would be helpful to better understand the concluded remarks if a relevant sentence included in the abstract too.
I have added text to the abstract. “The non-repeating tilted Lissajous orbit about L1 has time varying backscatter angles BA in the Earth’s atmosphere that varied between 176O to 174O during 2015 to 2019 and increased to BA near 178O in 2020 and 2021. Previous studies have shown that backscattering from the Earth’s surface in the visible and near infrared wavelengths increases at high backscatter angles.”
Lines 52-59: If I understand correctly the goal of the study wasn’t to compare (line 55) EPIC and OMPS-NM data, but to use OMPS-NM data to support EPIC results. Please clarify better and also in this paragraph only 388nm are given, while OMPS data correspond to different wavelengths.
Reference to the different wavelength for OMPS-NM has been added and a sentence added, “This paper uses the 380±0.55 nm LER data from the downward looking spatial mapper OMPS-NM to validate the observations from EPIC.”
Lines 82-83: In previous paragraph (lines 78-80) it is stated that the observed UV reflectivity comes mostly from clouds. Please, try not to repeat and be consistent with previous statements.
Fixed
Lines 115-116: I strongly recommend here to explain in more detail the correction applied and move equation given in lines 167-168 and 174 here. Fixed
Lines 117-119: I miss here the connection to the EPIC measurements. The EPIC measured backscatter UV radiances at 388nm are converted to LER as mentioned in the abstract, but nowhere up to now in the main text this description is given and for the first time appeared in eq (5) with a different symbol LER without explanation.
Typo: It should all be LER
The following has been added in the introduction: “Measured backscattered UV radiances at 388±1.5 nm are converted to Lambert Equivalent Reflectivity (LER) as a function of latitude, longitude, and time (Herman et al. 2018a).”
Lines 123-129: I cannot see an obvious increase in PSE in Fig 1. Also, is there any comment for the gap for the last ~6months of 2019?
The increase in the global average is small compared to that in the SH because the NH decrease cancels a portion of the SH increase.
The global percent reflected energy PSE appears to slightly increase during December 2020 (SEV = 1.95O) (Table 1) compared to the average from previous years (2015-2019) when the SEV angles are larger (Fig.1). The effect during the summer is larger at southern latitudes and decreased during summer (June) at northern latitudes. Sections 2.1 (SH) and 2.2 (NH) will show which latitude bands contribute to the change in reflected energy during 2020 – 2021 when the SEV angles are smallest.
Yes. Failure of the gyroscopes in June 2019 that were replaced with a star tracker and gas-jet propulsion. Normal operation was resumed in February 2020. The explanation is now in the text.
Lines 125-126: see 3rd specific comment
Lines 151-157: This description is for Fig 4 and not 3
Fixed
Lines 158-164: This paragraph fits better after the description of fig 4
Moved
Line 174: RA or AC the correction? And 2019 and not 2018? RA. AC has been removed
RA = Area in 2020/<Area 2015-2019>, where <…> denotes average
(10)
The correction is RA the ratio of the area viewed when SEV is small compared to when SEV was larger 4 – 6 degrees.
Line 180: Is Fig4 smoothed like Fig 1B? Not smoothed to that extent. It is a weekly average as described in the text and now specified in the caption
Line 234: There is no blue curve at fig 11.
In Fig.5 the monthly averages are June to June starting in June 2015 and ending in June 2019. The blue curve is June 2018 to June 2019. The data then starts again in June 2020 to June 2021. In Figure 11, they are December 2015 to December 2018. The blue curve would have been December 2018 to December 2019 but the period July 2019 to February 2020 is missing data. Fig 11 data starts again in December 2020 through December 2021.
I changed to figure caption to read “Fig. 11 Monthly average of NH annual time series of percent reflected energy at 388±1.5 nm for 4 years, 2015 – 2021. The period 2018-2019 is not shown since the data are missing starting on 28 June 2019 to February 2020.”
3) Technical corrections
Line 19: RSE corresponds to reflected energy for different wavelengths, so please change the sentence accordingly. For this paper, RSE applies only to 388±1.5 nm
Line 78: TOA acronym is introduced here, please put it inside parenthesis and in general check acronym throughout the manuscript. (TOA) done
Line 85: OMPS-NM instead of OMPS Fixed
Line 86: BA first appears in main text, please explain symbol Done
Line 99: OMPS-NM instead of NPP-NM Fixed
Figure 1: The y-axis doesn’t have PSE label Fixed
Figure 2: Please correct y-axis, P(t), Δθ Should be PSE (Percent)
Figure 3: Remove the second “The ratio” from the beginning of the legend. Additionally, it is 2019 and not 2019? The same for fig 9. Fixed and now references Eq. 10.
Figure 4: please replace “energy in the PSE” with energy PSE in the and correct y-axis, the same in fig 8. Fixed
Figure 5: Please correct y-axis, red line is for 2016-2017? Fixed
Figure 6: Please use the same color for lines and legends, here make Dec 2020 the same color with line. Additionally, is it 2018 or 2019? Fixed, and it is 2018, since December 2019 is missing
Figure 7: Please correct y-axis. Fixed
Table 2: 2018 or 2019? Also, Latitude in degrees. <2015 – 2018> degrees fixed
Line249: in Fig 10 instead of 11. Fixed
Line 259: OMPS-NM instead of OMI-NM Fixed
-
AC2: 'Reply on RC2', Jay Herman, 17 Jan 2023
Data sets
Measurements of Cloud Reflected Energy Jay Herman and Liang Huang https://avdc.gsfc.nasa.gov/pub/DSCOVR/JayHerman/Global_Refl/
Viewed
HTML | XML | Total | BibTeX | EndNote | |
---|---|---|---|---|---|
821 | 165 | 44 | 1,030 | 35 | 32 |
- HTML: 821
- PDF: 165
- XML: 44
- Total: 1,030
- BibTeX: 35
- EndNote: 32
Viewed (geographical distribution)
Country | # | Views | % |
---|
Total: | 0 |
HTML: | 0 |
PDF: | 0 |
XML: | 0 |
- 1