Austral Summer and Winter Analysis of Upper Tropospheric Wind Speed Trends for Brazil from 1980 to 2022

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

I find the article interesting, where a detailed analysis of wind in Brazil is presented. The layout of the article is correct and the individual chapters/sub-chapters have a logical continuity. The article's content is facilitated by figures, the number of which is proportional to the content and their quality is good. I recommend the article for publication after taking into account the following comments:

1) …The resultant seasonal trends are finally analyzed at each mandatory pressure level through a vertical profile and by constructing a 3D model through geographical information systems (GIS)…. This is currently too general a description. Please indicate which GIS tools were used? What was the criterion for their selection? Would relying on other software have produced different results?

2) I do not find in the article a broader reference of the results in terms of their applicability.

3) The article mentions land cover in several places. Against this background, please give more details about the coverage of the analysed area and their changes during the analysed period against the results obtained.

4) The results should be linked to Atmospheric teleconnections or at least cite the relevant literature in this respect in the Discussion chapter.

5) The Conclusions chapter should be rewritten. Here the most important results of your own research should be included: their strengths and weaknesses. Citing other publications here is inappropriate and should be in the Discussion chapter.

Author Response

Enclosed is a copy of our responses (in bold) to reviewers #2 comments (in italic) that evaluated the last submission of this manuscript submitted to Meteorology.

Response to Reviewer #2:

I find the article interesting, where a detailed analysis of wind in Brazil is presented. The layout of the article is correct and the individual chapters/sub-chapters have a logical continuity. The article's content is facilitated by figures, the number of which is proportional to the content and their quality is good. I recommend the article for publication after taking into account the following comments:

We thank reviewer #2 for taking the time to critically review and provide feedback to improve the manuscript. Based on your feedback, we have made the recommended corrections to providing additional content of how the 3-D mapping was completed, what role land cover change may have on lower level tropospheric winds, why atmospheric teleconnections should be considered as part of the analysis, and finally a conclusion section that summarizes the key findings, pitfalls, and possible future research related to upper-level winds in Latin America.

1) …The resultant seasonal trends are finally analyzed at each mandatory pressure level through a vertical profile and by constructing a 3D model through geographical information systems (GIS)…. This is currently too general a description. Please indicate which GIS tools were used? What was the criterion for their selection? Would relying on other software have produced different results?

We agree with reviewer #2 that more context is required for the construction of 3-D mapping of wind, temperature or geopotential height changes across the study area. In the materials and methods section, we provide a more in-depth analysis of how each reanalysis data was intersected with its grid field resolution and fields that only overlapped the study area were included as part of the analysis.  From that point, each pressure height was assigned or given a value. This was done to “help exaggerate” the distance between each layer when the map was created. To help make the maps visually appealing and readable 4 pressure heights were selected when 3-D maps were created for a given variable.  This rationale was chosen based on the number of hours working with the data and determining what atmosphere pressure heights are the most important to include as part of the manuscript. Two height groupings were chosen and using GIS software, a Local Scene (3-D) tool was used to visualize the pressure heights of the study area.

We decided on implementing a “step” ladder approach in the second version of the manuscript to reduce any clarity issues with the maps. If any issues arise in terms of results, it would be related to the type of GCS or PCS used in the analysis and what tools used to select grid fields based on the study area. The geovisualization is a tool to display the results, if any errors do exist in the analysis, it would be based on how the statistical test were performed. For this study, third-party software (i.e., R) was utilized to calculate any mean values and their statistical trends and their findings were merged into GIS software.

2) I do not find in the article a broader reference of the results in terms of their applicability.

We agree with reviewer #2 that the manuscript is missing a broader impact and applicability statement. Therefore, in the introduction section (L81-84), we include a statement about how wind energy production (renewable source) is becoming an important resource in northeastern/southern Brazil over the last two decades. Furthermore, we included a statement in the final paragraph (L724-726) of the conclusion section to remind the reader about how upper-tropospheric winds may influence future energy production across Brazil during the 21^st century.

3) The article mentions land cover in several places. Against this background, please give more details about the coverage of the analyzed area and their changes during the analyzed period against the results obtained.

We agree with reviewer #2. We included a more thorough background in the discussion section of the manuscript which talks about the rate or trend of winds between the lower and upper troposphere and how past research found similar outcome.  We also agree that additional research needs to be completed to provide a clearer understanding of how near-surface winds are influencing conditions within the planetary boundary layer.

4) The results should be linked to Atmospheric teleconnections or at least cite the relevant literature in this respect in the Discussion chapter.

We agree with reviewer #2. In the discussion, we include a conversation about how other synoptic (SAMS and SACZ) and planetary circulations (ENSO) are attributing to the year-to-year weather but also short and long-term climate of Brazil.

5) The Conclusions chapter should be rewritten. Here the most important results of your own research should be included: their strengths and weaknesses. Citing other publications here is inappropriate and should be in the Discussion chapter.

We rewrote the conclusion section of the manuscript and removed any necessary references. Now, the conclusion chapter highlights the key/major findings, any pitfalls and/or shortcomings and finally what possible future research should be considered to advance this topic area forward.

Reviewer 2 Report

Comments and Suggestions for Authors

I’m happy to see the analysis on upper tropospheric wind speed in South America, which is highly needed for the wind community. In my mind, the comparison between the IGRA and reanalysis should be further highlighted. The current MS has a substantial shortage in the comparison and validation parts; how much are the differences between the IGRA and reanalysis? The language and framework should also be further polished. I would recommend a major revision.

 

Major comments

1. Sometimes, the references are over-cited, e.g., L35 [2-12], L41 [22-32]. It’s not common, please be concise and cite the key references for the statements. I can find some inappropriate descriptions like:

1) L57 what is the meaning of North America (China)?

2) L67 850 and 700 hPa are not near-surface.

3) L71 wind on which level?

Above all, I suggest the author carefully revise the Introduction section.

2. Since the South American monsoon has strong effects on modulating the local atmospheric circulation. I suggest the author perform some analysis on long-term monsoon changes by using the widely used monsoon index (such as the definition in 2008 Wang Global monsoon dominant mode of annual variation in the tropics). By checking the variations of monsoon time series and monsoon circulation, it would be helpful to understand the current phenomena.
3. It would also be interesting to show the time series of area-averaged tropospheric wind speed of six datasets, as well as the BoH and STJ indices.

 

Minor comment：

1. Since the paper already has the entire data of Brazil's wind speed. Why not also include an analysis of spring and autumn? Yet, this is an optional choice.
2. L240: What is “southern and southern Brazil”?
3. L288: “3.3.3. D Wind Trends” to “3.3. 3-D Wind Trends”
4. L337: Do you mean the trends?
5. Figure 5: The wind vectors are too small.
6. Figures 4, 6, 7, and 9: Please add an ensemble mean plot in each figure.
7. Figure 8: Why does the CFSR show very different time series from others? As you already said that the MERRA2 and CFSR share the same GLADS to assimilate wind speed (L 485).

Author Response

Enclosed is a copy of our responses (in bold) to reviewers #3 comments (in italic) that evaluated the last submission of this manuscript submitted to Meteorology.

Response to Reviewer #3:

I’m happy to see the analysis on upper tropospheric wind speed in South America, which is highly needed for the wind community. In my mind, the comparison between the IGRA and reanalysis should be further highlighted. The current MS has a substantial shortage in the comparison and validation parts; how much are the differences between the IGRA and reanalysis? The language and framework should also be further polished. I would recommend a major revision.

We thank reviewer #3 for providing feedback and critique to improve the manuscript. As recommended by the reviewer, we reviewed and removed unnecessary references in the introduction, revised the wind vectors (arrows) for BoH and address any other concerns with the manuscript within the given time frame.

Major Comments:

1. Sometimes, the references are over-cited, e.g., L35 [2-12], L41 [22-32]. It’s not common, please be concise and cite the key references for the statements.

We agree with reviewer #3.  The authors revisited and removed any unnecessary references/citations from the introduction. We retained the most recent references and citations that particularly focus on Latin America.

I can find some inappropriate descriptions like:

• L57 what is the meaning of North America (China)?

We removed the parenthesis of positive (negative) wind trends for North America/western Europe (China) to reduce confusion. The sentence now discusses the wind speed trends related to North America and western Europe only.

• L67 850 and 700 hPa are not near-surface.

We agree with reviewer #3. The authors corrected the error and changed it from “near-surface” to “lower tropospheric wind trends (850 and 700 hPa)….”

• L71 wind on which level?

The authors corrected the sentence to read as followed: “McVicar et al. (2012) documented that near-surface wind trend climatology literature has been emphasized to North America, Europe, and Asia at the expense of Central and South America over the last decade”…

Above all, I suggest the author carefully revise the Introduction section.

2. Since the South American monsoon has strong effects on modulating the local atmospheric circulation. I suggest the author perform some analysis on long-term monsoon changes by using the widely used monsoon index (such as the definition in 2008 Wang Global monsoon dominant mode of annual variation in the tropics). By checking the variations of monsoon time series and monsoon circulation, it would be helpful to understand the current phenomena.

To construct a monsoon index [based on the definition of Wang and Ding (2008)] related to South America for 5 reanalysis products is not feasible given the short window to make corrections to the manuscript. We agree it would provide more clarity of how the South American monsoon may attribute to upper-level wind speed trends across the study area.  The authors include more literature related to the impact of the monsoon in the discussion section of the manuscript.  In the conclusion section of the paper, the authors also include a statement about how future research should examine how upper tropospheric wind trends may be interconnected to South American monsoon.

3. It would also be interesting to show the time series of area-averaged tropospheric wind speed of six datasets, as well as the BoH and STJ indices.

We agree with reviewer #3 that conducting an area-average tropospheric wind speed would enhance the manuscript. But with the given time frame allocated to make revisions, it would require additional time to calculate, map, and write the outcomes from such an analysis. We were able to complete an average tropospheric wind speed trend analysis for radiosonde dataset based on a 10-year period/cycle (2001-2010 and 2011-2020). If required by reviewer #3, we would perform such an analysis for wind speeds, BoH, and STJ.

Minor Comments:

1. Since the paper already has the entire data of Brazil's wind speed. Why not also include an analysis of spring and autumn? Yet, this is an optional choice.

In the original manuscript, a full analysis was completed for both annual and seasonal (DJF, MAM, JJA, and SON) wind trends.  However, to keep the scope of the manuscript at a reasonable length and appropriate number of figures and tables (a total of 11), austral spring and fall were excluded from the analysis. The author would like to complete a second study examining the “transition” seasons of upper-level wind speed trends since most studies only investigate wind speed trends at an annual or warm/cool season perspective.

2. L240: What is “southern and southern Brazil”?

The authors have corrected the statement to read as “southeastern and southern Brazil”.

3. L288: “3.3.3. D Wind Trends” to “3.3. 3-D Wind Trends”

The authors have corrected the subtitle to read as “3.3. 3-D Wind Trends”.

4. L337: Do you mean the trends?

The authors have added the word of “trend” as part of the sentence, to read as “It is anticipated when the BoH weakens, upper-level wind speed trends across central-west and southeastern (southern) Brazil decrease (increase).”

5. Figure 5: The wind vectors are too small.

The wind vector color has been changed from gray to black to better visualize the direction. The vectors were small because several of the reanalysis consist of higher resolution. Therefore, if the vectors were to be enhanced in size the graphics would be difficult to interpret and view. The authors were able to reduce the number of vectors from each reanalysis and increase the size of the wind vectors.  Please inform the authors if the color or size requires additional adjustment.

6. Figures 4, 6, 7, and 9: Please add an ensemble mean plot in each figure.

 

The authors require further clarification about creating an ensemble mean plot for Figures 4,6,7, and 9. Would a linear trend time series or spatial plot be necessary for each figure.  If so, additional time would be required to build and add the ensemble plots to the 4 figures.

7. Figure 8: Why does the CFSR show very different time series from others? As you already said that the MERRA2 and CFSR share the same GLADS to assimilate wind speed (L 485).

This is a great question posed by reviewer #3, we examined several references but we still have not be able to determine what factors or parameterizations are influencing the result differences between MERRA-2 and CFSR. To correctly determine this discrepancy between the two reanalyses will require further investigation by the authors.

Reviewer 3 Report

Comments and Suggestions for Authors

Attached

Comments for author File: Comments.pdf

Author Response

Enclosed is a copy of our responses (in bold) to reviewers #4 comments (in italic) that evaluated the last submission of this manuscript submitted to Meteorology.

Response to Reviewer #4:

The article demonstrates the spatial and vertical variation of wind speed at 7 pressure levels over Brazil between 1980 and 2022 (robust data base) using radiosonde data and several reanalyses. The article robustly shows wind speed variability using radiosonde and reanalysis data. Below I include some suggestions to help the article become a reference in the field.

We thank reviewer #4 for reviewing and providing feedback related to the submitted manuscript. We reconstructed the 3-D maps as a “step” ladder approach and we believe it will provide a easier way to view and interpret the results for any reader. We also provided additional metadata data collected for each radiosonde location used in the study and included further background on possible mechanisms (ENSO, SAMS, and SACZ) contributing to the wind speed trends across Brazil and finally added a future research statement about examining future wind speed trends and how it may influence wind energy production across specific regions of Brazil.

Errors to be corrected:

Page 1, line 12:

“...lowest wind speed trend changes trends...”

Suggested correction: remove the second “trends” → “lowest wind speed trend changes”.

We adjusted the language to read to as “lowest wind speed trend changes” as recommended by the reviewer.

Page 3, line 100:

“Figure. 1 shows the location...”

Suggested correction: remove the period after “Figure” → “Figure 1 shows...”.

We removed the period from L100 to describe Figure 1.

Page 20, line 591:

“Allen and Sherwood (2008) shows that...”

Suggested correction: “Allen and Sherwood (2008) show that...”.

We changed the wording from “shows” to “show”.

Page 21, line 609:

“south (north) of the SJT...”

Suggested correction: use “STJ” (Subtropical Jet) instead of “SJT”.

We corrected the acronym error from “SJT” to “STJ”.

1. Introduction

The authors present well the problem addressed, emphasizing the scarcity of (only one) studies analyzing upper-level wind speed variability over Latin America.

Suggestion: include papers that compare radiosonde-measured wind speed and reanalysis data.

We agree with the reviewer about including and citing past literature that have examined radiosonde and reanalysis products. At L74 we include a statement about how reanalysis and radiosonde data has been used to describe upper-level wind changes across the globe.  By adding this sentence, it allows the following sentence to better connect with recent research conducted in Argentina and how upper-level wind research is still limited in Latin America.

2. Materials and Methods

The authors used IGRAv2 data at 7 pressure levels and two daily synoptic times. The authors used the criterion of having more than half (12 years) of the years available for each site. However, it is necessary to state the percentage of data for each site separated by launch time (00Z or 12Z) (include in Figure 1 or in a supplementary table). This is important to assess whether the wind speed at each site is biased toward one radiosonde launch time.

The authors constructed a table that displays the name of the radiosonde site, ICAO id, total count of observations used in the study for austral summer and winter, and frequency/percentage related to radiosonde launch time (0 and 12 UTC).  The table was embedded within Figure 1 and the caption was updated to reflect that change.

3. Results

3.1

Suggestion: A decadal analysis (1980–1990, 1991–2000, etc.) could be included to assess whether trends are stable or if there are periods of reversal. Do winds show acceleration or deceleration in certain periods (e.g., post-2000)?

We collected 23 years of radiosonde data since most sites (airports) only began collecting in the early 2000s in Brazil. Therefore, we could only complete a decade wind speed trend analysis at 2001–2010 and 2011–2020 perspectives. This decadal wind information is included as part of Figure 2. We believe it provides a better understanding of how wind speeds can change not only over an entire study period (2000–2022) but also at a decadal perspective as well. A paragraph was included as part of the analysis to explain the key findings of the decade analysis.

3.2

Because both reanalysis and radiosonde data are used, a direct comparative analysis is missing indicating, if possible, which of the five reanalyses best reproduces the radiosonde-observed patterns.

We agree and the authors included two paragraphs in the discussion section of the manuscript to compare similarities and differences between radiosonde and reanalysis products and if any of the 5 climate products reproduce the findings of balloon launch data.

3.3.3. D Wind Trends (rename to 3.2.3)

The subtitle and heading were renamed to the recommendations of Reviewer #3.

The 250 hPa level in Figures 4, 6, 7, and 9 is hard to visualize.

Suggestion: redraw these figures presenting all levels with a layout similar to the 850 hPa panel for better visibility of trends.

The figures were redrawn to represent them as a “step” ladder but still in a 3-D visualization. We hope the new layout is easier to read and interpret for the readers.  If the study area was smaller (regional) it would be simpler to construct and visualize the multi-level approach from the original version of the manuscript.

The authors could include a quantitative correlation between wind and temperature / geopotential height trends to determine statistical relationships among the variables.

We agree with reviewer #4.  First, a spearman’s rank correlation test was performed between wind speed and temperature trends. We were able to determine that the latest generations of climate products (ERA5, MERRA-2, and CFSR) show a statistically significant correlation between the two variables in the upper tropospheric (between 400 and 200 hPa) while the older generation of reanalysis exhibits a low or minimal correlation. This statistical result is explained in L432-437 of the revised manuscript.

For the second part of the analysis on geopotential height (Figure 8), we conducted a similar correlation test.  We found that wind speed and geopotential trends are negatively correlated with statistical significance for all reanalysis products except for NCEP/DOE R2.  A sentence was included in the results section of manuscript (L447-449).

4. Discussion

Although briefly mentioned in the Introduction, the authors did not include any analysis of how variability modes such as El Niño/La Niña may have contributed to the observed trend patterns, even though they modulate the subtropical jet and the BoH.

Suggestion: include a discussion.

We agree with reviewer #4 that the variability of teleconnections should be addressed as part of the discussion. We include a summary of how BoH is interconnected with broader synoptic (SAMS and SACZ) and planetary circulation (ENSO) and how the dry and wet episodes can influence the wind trend patterns determined for this study.

The discussion mentions the BoH (Bolivian High) and STJ but could address other mechanisms such as changes in the South Atlantic Convergence Zone (SACZ/ZCAS) or South Atlantic teleconnections. The discussion could examine whether there are signals of acceleration or regime shifts (not just linear trends).

We incorporated this recommendation as part of the discussion as well.  SACZ has been identified has an important feature that can influence weather (wind) conditions across southern and southeastern Brazil during summer. Past research found that SACZ is migrating poleward and therefore wind speeds across the S/NE Brazil are influenced by this climatological shift (signals).  Future research should examine this relationship more closely as described in the conclusion section of the manuscript.

5. Conclusions

Include future recommendations derived from the conclusions.

Suggestions: future projections (CMIP6) and analyses of impacts on wind speed trends under the influence of variability modes such as ENSO, PDO, etc.

We agree with reviewer #4. A final paragraph was written to highlight how future projections may impact wind speed changes related to synoptic and planetary circulation patterns and the possible effect it may have on wind energy production seen along coastal (NE and S) Brazil over the past two decades.

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

I have no questions.