Improving the Accuracy of the Cross-Calibrated Multi-Platform (CCMP) Ocean Vector Winds

Round 1

Reviewer 1 Report

The paper is dedicated to accuracy improvement of the cross-calibrated multi-platform (CCMP) ocean winds. The completed CCMP wind analysis agrees better with long-term trend estimates from satellite observations and reanalysis than previous versions. The paper is written well. The results obtained are new, interesting, and valuable for the field. The results are clear, and their discussion is also presented well. Nevertheless, the paper needs some corrections before its publication in the journal.

 

Corrections suggested.

 

1. Please, add e-mail addresses of Lucrezia Ricciardulli and Frank Wentz in the affiliation lines. Also, please, add abbreviations of the authors’ first and last names.

2. Line 53. Please, substitute ‘(CCMP1)’ by ‘(CCMP 1.0)’.

3. Line 75. Please, justify ‘…satellites (TMI, GMI, AMSRE, WindSat and AMSR2)…’. This part of the sentence is not correct as you write about satellites, but you provide a mix of satellite instruments and satellite names, i.e., The Advanced Microwave Scanning Radiometer 2 (AMSR2) instrument launched aboard the Japan Aerospace Exploration Agency (JAXA) Global Change Observation Mission 1st-Water, "SHIZUKU" (GCOM-W1) satellite on May 18, 2012 https://www.earthdata.nasa.gov/learn/find-data/near-real-time/amsr2.

Also, in this connection, please justify Figure 1.

4. Lines 111-112: ‘Additional figures S1-S7 supporting Sections 2 and 4 are shown in the supplementary material.’. It is desirable to organize an Appendix A instead of separate file with a supplementary material.

5. Line 465. Please, substitute ‘(CCMP V2.0)’ by ‘(CCMP 2.0)’.

6. Please, prepare all the references exactly with the journal requirements. Use the journal template and latest published papers as examples.

 

 

So, the paper needs minor revision.

Author Response

1. Please, add e-mail addresses of Lucrezia Ricciardulli and Frank Wentz in the affiliation lines. Also, please, add abbreviations of the authors’ first and last names.
Partly done
We do not understand the “abbreviation” part of this comment.  We don’t see anything about this in the instructions or template
2. Line 53. Please, substitute ‘(CCMP1)’ by ‘(CCMP 1.0)’.
Done
3. Line 75. Please, justify ‘…satellites (TMI, GMI, AMSRE, WindSat and AMSR2)…’. This part of the sentence is not correct as you write about satellites, but you provide a mix of satellite instruments and satellite names, i.e., The Advanced Microwave Scanning Radiometer 2 (AMSR2) instrument launched aboard the Japan Aerospace Exploration Agency (JAXA) Global Change Observation Mission 1st-Water, "SHIZUKU" (GCOM-W1) satellite on May 18, 2012 https://www.earthdata.nasa.gov/learn/find-data/near-real-time/amsr2.
Also, in this connection, please justify Figure 1.
Done by adding a table of satellites, instruments and acronyms used in the paper.
4. Lines 111-112: ‘Additional figures S1-S7 supporting Sections 2 and 4 are shown in the supplementary material.’. It is desirable to organize an Appendix A instead of separate file with a supplementary material.
Done
5. Line 465. Please, substitute ‘(CCMP V2.0)’ by ‘(CCMP 2.0)’.
Done
6. Please, prepare all the references exactly with the journal requirements. Use the journal template and latest published papers as examples.
We have reformatted all references using the MDPI style specification for Zotero.  

Reviewer 2 Report

Overall comments:

As the title suggests, this manuscript is dedicated to improving the well-known CCMP sea surface vector winds version 2.0. The manuscript is well written. The methods are well introduced with some examples. Overall speaking, I recommend this manuscript be published soon. However, it would be better that some minor questions can be addressed.

1>   The CCMP winds are produced at 00:00, 06:00, 12:00, and 18:00 per day. The wind data represent ocean surface winds at one moment or average winds in 6 hours? This needs explanation.

2>   It is not clear that how the authors handled time? Scatterometer or radiometer data are provided at anytime in day.

3>   It would be helpful if the authors could give a chart which showing data processing flow for producing one given time of winds. Explaining how you handled the grids which only have scatterometer data, radiometer, both scatterometer and radiometer, multiple data source, or no satellite data?

4>   It is not clear that how the radiometer winds speed contribute to CCMP winds? If I understand correctly, the u and v wind components are used in the processing. How scatterometer, radiometer, and ERA5 winds are merged?

5>   How the data gap are filled? For example, when producing winds at 12:00, and there is a grid with no satellite data available. Then ERA5 winds at 12:00 will be used as initialization?

 

6>   Can the CCMP3.0 data reveal the diurnal patterns of winds in regional areas?

Author Response

The main change we made to the manuscript in response to Reviewer #2 was to expand the description of the CCMP Variational Analysis Method to include details to make it more generally understandable.  However, we do want to rewrite the much more extensive discussion that has already be published in Atlas et al. 2011 (Reference 5).
1>    The CCMP winds are produced at 00:00, 06:00, 12:00, and 18:00 per day. The wind data represent ocean surface winds at one moment or average winds in 6 hours? This needs explanation.
Response:  Each “map” represents an estimated snap shot at the synoptic time.  Added a phrase at line 366 to make this more clear.  
“The resulting gridded map corresponds an estimate of the wind field at a single synoptic time instead of the range of times represented in a satellite dataset.”
The expanded discussion of the CCMP VAM also makes this more clear.
2>    It is not clear that how the authors handled time? Scatterometer or radiometer data are provided at anytime in day.
We added a substantially longer description of the CCMP VAM starting at line 415.  The new description includes a brief description of how we de-weight observations away from the analysis time as well the “First Guess at Appropriate Time” that is intended to handle the time evolution of the wind field using a first-order approximation.
3>    It would be helpful if the authors could give a chart which showing data processing flow for producing one given time of winds. Explaining how you handled the grids which only have scatterometer data, radiometer, both scatterometer and radiometer, multiple data source, or no satellite data?
We hope that the expanded discussion of the CCMP VAM satisfies this suggestion.  In the discussion we added material pertaining to each of the cases above.
4>    It is not clear that how the radiometer winds speed contribute to CCMP winds? If I understand correctly, the u and v wind components are used in the processing. How scatterometer, radiometer, and ERA5 winds are merged?
The radiometer wind speeds are part of the second term in the cost function.  This is clearer in the expanded discussion.
5>    How the data gap are filled? For example, when producing winds at 12:00, and there is a grid with no satellite data available. Then ERA5 winds at 12:00 will be used as initialization?
If there is no nearby data, the adjusted ERA5 winds are the main contributor to the analyzed field.  This is clearer is the expanded discussion.
6>    Can the CCMP3.0 data reveal the diurnal patterns of winds in regional areas?
We have not yet evaluated the accuracy of any diurnal patterns in the CCMP analysis.  Also, because there are only 4 analysis times per day, any semi-diurnal variation is not adequately sampled.  We added a caveat about diurnal fidelity to the discussion at the end of the manuscript.

 

Reviewer 3 Report

In this manuscript, the authors investigated the adjustment of the ERA5 background wind field and radiometer wind data to be used in the CCMP variational analysis for better agreement with satellite results. Experiment result showed the proposed scheme is effective, especially for high wind speeds. I suggest the authors consider the following problems in revision:

 

1.     Line 73, decipher “RSS” at its first location.

2.     Line 108, decipher “NWP” at its first location.

3.     Line 188, the adjustment factors are obtained empirically. Machine learning methods could do better job in reanalysis wind correction, see

1)     Y. Li, et al, An Adversarial Learning Approach to Forecasted Wind Field Correction with An Application to Oil Spill Drift Prediction, Int. J. Appl. Earth Observ. Geoinf., vol. 112, p. 102924, 2022.

2)     A. Gonzalez-Arceo, et al, Calibration of Reanalysis Data against Wind Measurements for Energy Production Estimation of Building Integrated Savonius-Type Wind Turbine. Appl. Sci., 10, 9017, 2020.

You may discuss it and consider it as future work.

4.     Line 271, incomplete sentence.

5.     Line 405, explain how λ’s are determined.

6.     You validated your correction results with other satellite data, how can you ensure the satellite data are precise?

Author Response

1.    Line 73, decipher “RSS” at its first location.
It is already defined on line 60
2.    Line 108, decipher “NWP” at its first location.
It is already defined on line 65
3.     Line 188, the adjustment factors are obtained empirically. Machine learning methods could do better job in reanalysis wind correction, see
1)     Y. Li, et al, An Adversarial Learning Approach to Forecasted Wind Field Correction with An Application to Oil Spill Drift Prediction, Int. J. Appl. Earth Observ. Geoinf., vol. 112, p. 102924, 2022.
2)     A. Gonzalez-Arceo, et al, Calibration of Reanalysis Data against Wind Measurements for Energy Production Estimation of Building Integrated Savonius-Type Wind Turbine. Appl. Sci., 10, 9017, 2020.
You may discuss it and consider it as future work.
This approach looks interesting.  Added a mention of machine learning approaches at the end of section 2.2 (line 179).
3.    Line 271, incomplete sentence.
Fixed
4.    Line 405, explain how λ’s are determined.
We have substantially expanded the description of the CCMP Variational Analysis, including a brief discussion of how the λ’s were chosen.  A more detailed treatment would risk rewriting Atlas et al., 2011, (reference [5]).
5.    You validated your correction results with other satellite data, how can you ensure the satellite data are precise?
The purpose of this study is to describe the method for creating a merged product that is consistent with satellite measurements.  The ASCAT-B measurements that we used as independent validation data for our wind analysis had been validated using wind measurements from moored buoys in references 15, 16 and 24. Moreover, towards the end of the manuscript we stated that “Further validation of this new version of the CCMP winds using independent buoy wind measurements are underway. The results will be reported in subsequent manuscripts”.

Round 2

Reviewer 3 Report

I don’t have other comments.