the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 05 Sep 2022
Validation of Aeolus wind profiles using ground-based lidar and radiosonde observations at La Réunion Island and the Observatoire de Haute Provence
Mathieu Ratynski
Sergey Khaykin
Alain Hauchecorne
Robin Wing
Jean-Pierre Cammas
Yann Hello
Philippe Keckhut
Abstract. European Space Agency’s (ESA) Aeolus satellite mission is the first Doppler wind lidar in space, operating in orbit for more than three years since August 2018 and providing global wind profiling throughout the entire troposphere and the lower stratosphere. The Observatoire de Haute Provence (OHP) in southern France and the Observatoire de Physique de l’Atmosphère à La Réunion (OPAR) are equipped with ground-based Doppler Rayleigh-Mie lidars, which operate on similar principles to the Aeolus lidar, and are among essential instruments within ESA Aeolus Cal/Val program. This study presents the validation results of the L2B Rayleigh-clear HLOS winds from September 2018 to January 2022. The point-by-point validation exercise relies on a series of validation campaigns at both observatories: AboVE (Aeolus Validation Experiment) that were held in September 2019 and June 2021 at OPAR, and in January 2019 and December 2021 at OHP. The campaigns involved time-coordinated lidar acquisitions and radiosonde ascents collocated with the nearest Aeolus overpasses. During AboVE-2, Aeolus was operated in a campaign mode with an extended range bin setting allowing inter-comparisons up to 28.7 km. We show that this setting suffers from larger random error in the uppermost bins, exceeding the estimated error, due to lack of backscatter at high altitudes. To evaluate the long-term evolution in Aeolus wind product quality, twice-daily routine Météo-France radiosondes and regular lidar observations were used at both sites. This study evaluates the long-term evolution of the satellite performance along with punctual collocation analyses. On average, we find a systematic error (bias) of −0.92 ms-1 and −0.79 ms-1 and a random error (scaled MAD) of 6.49 ms-1 and 5.37 ms-1 for lidar and radiosondes, respectively.
-
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
(1762 KB)
-
Supplement
(177 KB)
-
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(1762 KB) -
Supplement
(177 KB) - BibTeX
- EndNote
- Final revised paper
Journal article(s) based on this preprint
Mathieu Ratynski et al.
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2022-822', Anonymous Referee #1, 27 Sep 2022
General comments:
I am very happy to read this article. It clearly outlines the data and methods used, and provides an important new result for the validation of the spaceborn Aeolus lidar.
Specific comments:
I am worried about the presentation of the overall statistics, which are an accumulation of Aeolus data for different baselines (for example in the abstact on lines 25,26).
The Aeolus instrument settings as well as the ground processing has seen several major changes during its mission. These will have an effect on statistical properties like bias and standard deviation/MAD. In addition to the combined statistics I think it would be better to split the results and also present them separately for the different baselines. Also it seems near-real-time and reprocessing results are mixed, i.e. baseline 11 was introduced in near-real-time processing on 8-Oct-2020, so the baseline 11 results before that date must be based on reprocessed Aeolus data. I think it would be better to split this as well, since the reprocessing used different calibration data than the near-real-time processing.line 47: you state that Aeolus covers nearly the whole globe within 7 days.
This is not really the case. With a 7 day repeat cycle of the orbit the instrument observes a specific pattern on the earth and the slant curtain above this pattern, but it certainly does not observe every location on earth.line 453: The figure depicts a very specific pattern of oscillating nature.
This pattern is indeed striking, and I have not seen such a thing before in previous Aeolus publications. I think it is important to try and understand what is happening here. But I think you should not call this "instument induced", since you cannot
yet prove that this indeed is the case. There could also be some bug or unforeseen effect in the ground processing or in the data handling of this paper. So I would suggest to find another name and not use the acronym I2OPs. Please contact the Aeolus DISC team and work with them to try and find what is happening here.Technical corrections:
line 13: Aeolus is now flying for over 4 years, so please correct your statement that it is operating for 3 years.
line 32: wind profiling crucial => wind profiling is crucial
line 49: the first ever Doppler-Rayleigh Wind Lidar =>the first ever Doppler-Rayleigh-Mie Wind Lidar
line 135: classified using particle backscatter coefficient
The classification method was changed to use SNR threshold on 31-Oct-2019 with the start of baseline 7 for the Mie channel, and on 8-Oct-2020 with the start if baseline 11 for the Rayleigh channel.
line 159/160: the end of the mission's extended life in November 2022. ==>the end of the mission's extended life in spring 2023.
line 168: the difference between vLOS and HLOS becomes negligible
No this is not true. If w is small, than the sine term in equation (1) becomes negligible, but the cosine term stll remains. Therefore there still is a difference by a factor of cas(Psi) between vLOS and vHLOS.
line 181/182: multiple RBC settings are activated at the same time.
No this is not true. Each channel has just one RBS at any given time. But the RBS can be changed multiple times per orbit.line 192: The downsampling begins with an averaging of the reference measurements between the middle points of the reference bins
This phrasing is confusing and maybe I misunderstood.
Each Aeolus wind result has just one middle point, so there is no in between. So I think the correct way is to take the reference measurements between the top and bottom edge of the Aeolus measurement bin, and average these results, before comparing to the Aeolus result. That way no interpolation at all is needed.line 311: to average every profile => to average every Aeolus profile
line 345: shown in Fig. 5a => shown in Fig. 4a
line 387: One reason Sun et al. (2014) raised
One important contribution for orbital phase biases is the telescope temperature effect explained by Weiler et al., 2021.
I think you should mention this as well here.line 439: At the same time, the radiosonde drifts along
You could mention here that not only the distance between Aeolus and radiosonde changes with time, but also the time difference between the two systems changes with time and therefore also with altitude.line 512: Once can thus conclude => One can thus conclude
line 530/531: the end of the extended mission lifetime in November 2022, =>the end of the extended mission lifetime in spring 2023,
line 580/581: there are fewer particles at higher altitude levels. =>there are fewer molecules at higher altitude levels.
line 603: higher by and average => higher by an average
Citation: https://doi.org/10.5194/egusphere-2022-822-RC1 - AC1: 'Reply on RC1', Mathieu Ratynski, 23 Dec 2022
- AC2: 'Reply on RC1', Mathieu Ratynski, 23 Dec 2022
-
RC2: 'Comment on egusphere-2022-822', Anonymous Referee #2, 09 Nov 2022
In my opinion this is an excellent paper summarizing the comparison of an extensive data set of radiosonde and surface-based lidar observations with Aeolus wind estimates. The paper is quite comprehensive in its analysis of the data set. I especially like that the authors have compared their results with results from other groups carrying out similar investigations at different locations and using different instruments. This work is quite important for evaluating Aeolus performance and assessing the potential utility of the Aeolus observations.
Most of my comments are minor and editorial in nature, and I leave it to the authors and editor to decide on whether or not to include them in a revised manuscript.
Specific comments
Line 32: It seems that there should be an “is” inserted after “wind profiling”.
Line 56: Perhaps I missed it, but it’s a bit unclear whether the work described in the paper is investigating total bias or residual bias. As noted, bias correction schemes were implemented to the Aeolus data at different times in the mission. Although a fully bias-corrected data set is being developed for Aeolus, this analysis seems to be using data that may have included intermittent changes to the bias. A sentence or 2 to perhaps discuss bias correction in the Aeolus data set and which data are being used for comparison could be helpful to the reader.
Line 79: Is LIOvent an acronym? If so, please define.
Line 93: Does the LIOvent lidar measure only Rayleigh winds?
Line 230, Figure 1 caption: I think that the caption could use a bit more explanation. Although I think that the straight colored trajectories represent lidar measurements from Aeolus and the ground-based lidar, this isn’t described in the caption so it required some scrutiny on my part to discern what they represent. The same is true for the Figure 2 caption.
Line 244: Looks like a word is missing after “however”.
Line 300 (Figure caption): It would be useful for the caption to differentiate between the lighter and darker shading and c and relate them (presumably) to the orbit.
Line 319: I find the paragraphs beginning at line 305 and line 320 to be unclear. Does paragraph 305 refer to the radiosonde comparisons while 320 refers to the lidar comparisons? Also the text beginning on line 305 refers to figures 3a, 3c, (wind measurement differences), while the text beginning on line 320 refers figures 3c(wind measurement differences) , 3d (data count), but seem to be doing an equivalent comparison. These two paragraphs seem to require some clarification or correction.
Lines 345 and 347: I think the authors mean “4a” instead of “5a” here.
Citation: https://doi.org/10.5194/egusphere-2022-822-RC2 - AC3: 'Reply on RC2', Mathieu Ratynski, 23 Dec 2022
-
RC3: 'Comment on egusphere-2022-822', Anonymous Referee #3, 09 Dec 2022
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2022/egusphere-2022-822/egusphere-2022-822-RC3-supplement.pdf
- AC4: 'Reply on RC3', Mathieu Ratynski, 23 Dec 2022
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2022-822', Anonymous Referee #1, 27 Sep 2022
General comments:
I am very happy to read this article. It clearly outlines the data and methods used, and provides an important new result for the validation of the spaceborn Aeolus lidar.
Specific comments:
I am worried about the presentation of the overall statistics, which are an accumulation of Aeolus data for different baselines (for example in the abstact on lines 25,26).
The Aeolus instrument settings as well as the ground processing has seen several major changes during its mission. These will have an effect on statistical properties like bias and standard deviation/MAD. In addition to the combined statistics I think it would be better to split the results and also present them separately for the different baselines. Also it seems near-real-time and reprocessing results are mixed, i.e. baseline 11 was introduced in near-real-time processing on 8-Oct-2020, so the baseline 11 results before that date must be based on reprocessed Aeolus data. I think it would be better to split this as well, since the reprocessing used different calibration data than the near-real-time processing.line 47: you state that Aeolus covers nearly the whole globe within 7 days.
This is not really the case. With a 7 day repeat cycle of the orbit the instrument observes a specific pattern on the earth and the slant curtain above this pattern, but it certainly does not observe every location on earth.line 453: The figure depicts a very specific pattern of oscillating nature.
This pattern is indeed striking, and I have not seen such a thing before in previous Aeolus publications. I think it is important to try and understand what is happening here. But I think you should not call this "instument induced", since you cannot
yet prove that this indeed is the case. There could also be some bug or unforeseen effect in the ground processing or in the data handling of this paper. So I would suggest to find another name and not use the acronym I2OPs. Please contact the Aeolus DISC team and work with them to try and find what is happening here.Technical corrections:
line 13: Aeolus is now flying for over 4 years, so please correct your statement that it is operating for 3 years.
line 32: wind profiling crucial => wind profiling is crucial
line 49: the first ever Doppler-Rayleigh Wind Lidar =>the first ever Doppler-Rayleigh-Mie Wind Lidar
line 135: classified using particle backscatter coefficient
The classification method was changed to use SNR threshold on 31-Oct-2019 with the start of baseline 7 for the Mie channel, and on 8-Oct-2020 with the start if baseline 11 for the Rayleigh channel.
line 159/160: the end of the mission's extended life in November 2022. ==>the end of the mission's extended life in spring 2023.
line 168: the difference between vLOS and HLOS becomes negligible
No this is not true. If w is small, than the sine term in equation (1) becomes negligible, but the cosine term stll remains. Therefore there still is a difference by a factor of cas(Psi) between vLOS and vHLOS.
line 181/182: multiple RBC settings are activated at the same time.
No this is not true. Each channel has just one RBS at any given time. But the RBS can be changed multiple times per orbit.line 192: The downsampling begins with an averaging of the reference measurements between the middle points of the reference bins
This phrasing is confusing and maybe I misunderstood.
Each Aeolus wind result has just one middle point, so there is no in between. So I think the correct way is to take the reference measurements between the top and bottom edge of the Aeolus measurement bin, and average these results, before comparing to the Aeolus result. That way no interpolation at all is needed.line 311: to average every profile => to average every Aeolus profile
line 345: shown in Fig. 5a => shown in Fig. 4a
line 387: One reason Sun et al. (2014) raised
One important contribution for orbital phase biases is the telescope temperature effect explained by Weiler et al., 2021.
I think you should mention this as well here.line 439: At the same time, the radiosonde drifts along
You could mention here that not only the distance between Aeolus and radiosonde changes with time, but also the time difference between the two systems changes with time and therefore also with altitude.line 512: Once can thus conclude => One can thus conclude
line 530/531: the end of the extended mission lifetime in November 2022, =>the end of the extended mission lifetime in spring 2023,
line 580/581: there are fewer particles at higher altitude levels. =>there are fewer molecules at higher altitude levels.
line 603: higher by and average => higher by an average
Citation: https://doi.org/10.5194/egusphere-2022-822-RC1 - AC1: 'Reply on RC1', Mathieu Ratynski, 23 Dec 2022
- AC2: 'Reply on RC1', Mathieu Ratynski, 23 Dec 2022
-
RC2: 'Comment on egusphere-2022-822', Anonymous Referee #2, 09 Nov 2022
In my opinion this is an excellent paper summarizing the comparison of an extensive data set of radiosonde and surface-based lidar observations with Aeolus wind estimates. The paper is quite comprehensive in its analysis of the data set. I especially like that the authors have compared their results with results from other groups carrying out similar investigations at different locations and using different instruments. This work is quite important for evaluating Aeolus performance and assessing the potential utility of the Aeolus observations.
Most of my comments are minor and editorial in nature, and I leave it to the authors and editor to decide on whether or not to include them in a revised manuscript.
Specific comments
Line 32: It seems that there should be an “is” inserted after “wind profiling”.
Line 56: Perhaps I missed it, but it’s a bit unclear whether the work described in the paper is investigating total bias or residual bias. As noted, bias correction schemes were implemented to the Aeolus data at different times in the mission. Although a fully bias-corrected data set is being developed for Aeolus, this analysis seems to be using data that may have included intermittent changes to the bias. A sentence or 2 to perhaps discuss bias correction in the Aeolus data set and which data are being used for comparison could be helpful to the reader.
Line 79: Is LIOvent an acronym? If so, please define.
Line 93: Does the LIOvent lidar measure only Rayleigh winds?
Line 230, Figure 1 caption: I think that the caption could use a bit more explanation. Although I think that the straight colored trajectories represent lidar measurements from Aeolus and the ground-based lidar, this isn’t described in the caption so it required some scrutiny on my part to discern what they represent. The same is true for the Figure 2 caption.
Line 244: Looks like a word is missing after “however”.
Line 300 (Figure caption): It would be useful for the caption to differentiate between the lighter and darker shading and c and relate them (presumably) to the orbit.
Line 319: I find the paragraphs beginning at line 305 and line 320 to be unclear. Does paragraph 305 refer to the radiosonde comparisons while 320 refers to the lidar comparisons? Also the text beginning on line 305 refers to figures 3a, 3c, (wind measurement differences), while the text beginning on line 320 refers figures 3c(wind measurement differences) , 3d (data count), but seem to be doing an equivalent comparison. These two paragraphs seem to require some clarification or correction.
Lines 345 and 347: I think the authors mean “4a” instead of “5a” here.
Citation: https://doi.org/10.5194/egusphere-2022-822-RC2 - AC3: 'Reply on RC2', Mathieu Ratynski, 23 Dec 2022
-
RC3: 'Comment on egusphere-2022-822', Anonymous Referee #3, 09 Dec 2022
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2022/egusphere-2022-822/egusphere-2022-822-RC3-supplement.pdf
- AC4: 'Reply on RC3', Mathieu Ratynski, 23 Dec 2022
Peer review completion
Journal article(s) based on this preprint
Mathieu Ratynski et al.
Mathieu Ratynski et al.
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 457 | 221 | 22 | 700 | 43 | 8 | 5 |
- HTML: 457
- PDF: 221
- XML: 22
- Total: 700
- Supplement: 43
- BibTeX: 8
- EndNote: 5
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(1762 KB) - Metadata XML
-
Supplement
(177 KB) - BibTeX
- EndNote
- Final revised paper