Observed Surface Wind Field Structure of Severe Tropical Cyclones and Associated Precipitation

Round 1

Reviewer 1 Report

The authors analyzed the wind structure of 28 TCs observed by synthetic aperture radar, and compared them with the ERA5 reanalysis, TCIAA model simulation and IBTrACS data. They found that the SAR data had high reliability. The authors also proposed a new model for tropical cyclone wind based on solely the ERA5 reanalysis data. They also analyzed the relationship among the SST, wind speed and precipitation. Over all, the paper is interesting and worth publication. One suggestion is to polish more the language, currently the language is a bit rough. For example, the first sentence in the abstract is not easy to understand. Line 24, ‘originate’ should be ‘originates’. Etc.

Author Response

Points 1: The authors analyzed the wind structure of 28 TCs observed by synthetic aperture radar, and compared them with the ERA5 reanalysis, TCIAA model simulation and IBTrACS data. They found that the SAR data had high reliability. The authors also proposed a new model for tropical cyclone wind based on solely the ERA5 reanalysis data. They also analyzed the relationship among the SST, wind speed and precipitation. Over all, the paper is interesting and worth publication. One suggestion is to polish more the language, currently the language is a bit rough. For example, the first sentence in the abstract is not easy to understand. Line 24, ‘originate’ should be ‘originates’. Etc.

Response 1: Thanks. We tried our best to improve the language and make some changes to the manuscript. We also asked a native speaker to smooth our paper. We marked the changes by using the ‘Track Changes’ function in MS word. Hope the manuscript is much better now.

Reviewer 2 Report

This paper presents an interesting use of SAR wind speed measurements to help improve tropic cyclone models.  My key issue is the SAR wind speeds.  Co-pol (VV or HH) imagery has long been used for wind speed but are best validated below 20 m/s.  There are  newer cross-pol algorithms (VH or HV) that have problems at low wind speed but are much better at wind speeds > 20 m/s and thus more appropriate for the study of TCs.  Perhaps I have misunderstood the GMFs applied, but I would like the authors to address this issue. Perhaps cross-pol imagery was not available.  I have attached a PDF with some minor grammar suggestions.

Comments for author File: Comments.pdf

Author Response

Points 1: This paper presents an interesting use of SAR wind speed measurements to help improve tropic cyclone models.  My key issue is the SAR wind speeds.  Co-pol (VV or HH) imagery has long been used for wind speed but are best validated below 20 m/s.  There are newer cross-pol algorithms (VH or HV) that have problems at low wind speed but are much better at wind speeds > 20 m/s and thus more appropriate for the study of TCs.  Perhaps I have misunderstood the GMFs applied, but I would like the authors to address this issue. Perhaps cross-pol imagery was not available.  I have attached a PDF with some minor grammar suggestions.

Response 1: We sincerely thanks for your professional review work on our article. About the SAR wind speed, we added more references and provided more details in the revised manuscript:

P2, Line 81-87: Wind speeds from dual-polarized SAR are used in the study. The dual-polarized SAR combined both co-polarized and cross-polarized channel in observation. The previous study showed that the co-polarization (VV or HH) worked well under low to moderate wind speed, but the sensitivity decreased at high wind speeds. The cross-polarization (VH or HV) exhibited improved sensitivity comparing with co-polarization and with the combination of both channels, the dual-polarized SAR have quite high accuracy for wind speeds larger than 25m/s.

Reviewer 3 Report

This manuscript assesses the surface wind fields near CT centers from SAR,  ERA5, and TCIAA wind model. A new statistical wind structure model is set up using ERA5 data based on the assessment. Other discussions are about SAR’s effectiveness in depicting surface wind structure and the correlation between the warm sea surface, TC wind fields, and precipitation. The manuscript is generally well written. It is important to develop improved TC wind model. However, the authors do not discuss how they made many of their design choices, while there are other possible options available. This affected the soundness of this research.

First, Why TCIAA, not some other modes, is included in this study? Please also consider to give a name to the new model developed. Why the authors chose those 44 different TC times when there are a lot more. Why use only 12 TC snapshots for comparisons in Table 1? What is considered to be having few default values? (by the way, the meaning of “default values in SAR data” mentioned in Page 4, line 148, is not explained. It is “missing values”?)

Second, this study used “quadratic fitting curve” to get the new model. Curve fitting, and other fitting methods, such as"machine learning" regression, all approximate data with functions. "machine learning” may not have same efficiency and accuracy of the general curve fitting algorithms, but the authors still need to discuss why they chose this method over others without the need to compare the results from several options .

In addition, why only use the average TC surface wind field structure of 14 TC times in the comparison between SAR observation, ERA5, TCIAA model and developed model (Figure 7), while forecasting individual TCs matters more their average. What if the new model look better in the “average” plot, but have large difference in some of the TC times. How did the authors select the 3 “TC” Time used in Fig. 2?

The authors are suggested to use graphics such as bar charts instead of listing numbers in Table 1. For example, the authors could use 3 bar charts for RMW, Vmax, and R17, respectively. In each bar chars, use bars for  SAR, ERA5 and TCIAA model of all selected CTs. People can tell the difference between SAR, ERA5 and TCIAA quickly and clearly by just just looking at the heights of bars.

Here are some minor issues:

A comma may be needed in line 13, page 1, after “observations” and before “the fifth generation ECMWF reanalysis for the global climate and weather (ERA5)”.  In Figure 9, it is better to put labels of “SAR”， “new model”, “ERA5”, “GPM” inside the subplots, like figure 2 and 3. 

Author Response

We sincerely thanks for your professional review work on our article. We upload the response as a PDF file and please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

The only systemic comment I have is that I worry about the correlation between precipitation and SAR winds.  The rain could affect the surface roughness and hence the retrieved winds.

Some modest suggested changes:

On line 46, "strongly ocean basins dependent" should be "strongly ocean-basin dependent"

On line 52, "the momentum equations as assuming a constant boundary layer depth" should read "the momentum equations and assuming a constant boundary layer depth"

On line 64, SST should be defined as sea surface temperature in its first use.

On line 71, "and it been less affected by clouds" should read "and being less affected by clouds".

One line 75, insert a space before "[27]"

On line 83, "SAR combined both co-polarized and cross-polarized channel in observation."  should read "SAR combines both co-polarized and cross-polarized channel to make a wind speed observation."

On line 85, inset a space before  "[28, 29]"

On line 88, inset a space before  "[27]"

On line 88, "100m" should read "100 m"

On line 90, "in the Eastern Pacific and files with missing values which led to incomplete TC coverage" should read "in the Eastern Pacific. Files with missing values which led to incomplete TC coverage"

On line 93, does "2-3" minutes refer to geographical extent and not time? This should be made clearer.

On line 101, "34kt (about 17m/s)" should read "34 kt (about 17 m/s)" I will not note this further in the review, numbers and units should be linked with a non-breaking space. 

On line 104, "was" should be replaced with "were" since the term "data" is plural.

On line 109, "u" and "v" be italicized since they represent variable.

On line 141, "speed, RMW and a fitting parameter, α, ? is the surface motion vector of the TC and r" should read "speed, RMW and a fitting parameter, α. ?_b is the surface motion vector of the TC and r"  "r" should be italicized.

On line 147, "when" should be "where".

On line 154, "belong" should be "belonging".

On line 160, "figure 2 and figure 3" should "Figure 2 and Figure 3"

On line 183, "(Figures 2&3)" should read "(Figures 2 and 3)"

On line 194, decide whether "Figure" should be spelled out or abbreviated.

On line 208, "Considering that SAR is not available" should read "Considering that SAR data are not available"

Lines 213 and 215, "a" and "b" should be italicized.  Remember that in the manuscript, variables should be italicized.

On the paragraph starting with line paragraph 244, is it possible that high rate rate cm affect the surface roughness and hence the SAR backscatter?

Author Response

Responds to Reviewer 2

Thank you for your constructive comments on my manuscript. We have carefully considered the suggestion and make some changes.

1) In the study, we didn’t consider the influence of surface roughness. So, we add the discussion and the reference in the final part at line 298.

Line 298-301: The relationship between wind and precipitation involves complicated dynamic and thermodynamic processes. In addition, the rain perturbations can be resulted from changes of sea surface roughness by impinging drops and influence the wind vector retrieval [42]. As a result, the connection between wind speed and precipitation needs to be further explored.

2) The grammatical and writing mistakes have been corrected. We marked the changes by using the ‘Track Changes’ function in MS word. Hope the manuscript is much better now.