the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 08 Jul 2026
Frequent surface–rotor stability decoupling limits the representativeness of surface-based offshore observations
Abstract. Understanding atmospheric stability throughout the wind turbine rotor layer is critical for predicting wake evolution, power production, wake steering performance, and structural loading. Here, we use sonic anemometer, profiling lidar, and thermodynamic profiler observations collected from an instrumented barge deployed during June–September 2024 as part of the Third Wind Forecast Improvement Project (WFIP3) campaign off the U.S. northeast coast to evaluate how well near-surface measurements represent atmospheric stability at wind turbine hub height. Surface-based and rotor-layer stability classifications disagree more than 30 % of the time, with the dominant decoupling regime characterized by unstable surface conditions beneath a stably stratified rotor layer. The frequency of these decoupling events increases through late summer into fall as hub-height stratification strengthens, while individual events become shorter-lived, indicating frequent transitions between coupled and decoupled states. Decoupling events are characterized by slower wind speeds consistent with weaker turbulent mixing through the marine boundary layer. Surface observations often fail to represent the atmospheric stability experienced by offshore wind turbines during summer marine conditions. Consequently, relying solely on near-surface measurements may underestimate the occurrence of stable rotor-layer conditions that influence wake behavior, offshore wind farm power performance, and estimates of structural loads.
Competing interests: At least one of the (co-)authors is a member of the editorial board of Wind Energy Science.
Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.- Preprint
(4927 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
- RC1: 'Comment on wes-2026-115', Anonymous Referee #1, 31 Jul 2026
-
RC2: 'Comment on wes-2026-115', Anonymous Referee #2, 15 Oct 2026
The paper “Frequent surface–rotor stability decoupling limits the representativeness of surface-based offshore observations” by Vasquez-Barros et al., addresses an important question: how well do measurements near the sea surface describe atmospheric stability at offshore turbine heights? The combination of sonic anemometers, lidars, and temperature-profile measurements provides a valuable dataset. The occurrence of unstable surface conditions beneath stable air at hub-height is interesting with potential relevance to offshore wind measurements and modeling.
However, I recommend major revision. The authors should better establish the reliability of the classifications, impact of uncertainty of observations, clarify their use of “decoupling”, and distinguish the observed stability differences from their possible effects on power. Some of the reviewer’s major and minor comments are provided below.
1. Quantify uncertainty in the temperature differences and stability classifications.
The authors apply motion corrections to the barge, which is appropriate. However, they should assess how the measurement and retrieval uncertainties affect the reported decoupling frequency. The analysis should account for the relationship between errors at different heights, rather than assume they are independent, where possible.
For example, the authors could use the TropOE temperature retrieval’s error covariance to estimate uncertainty in the virtual potential temperatures across 120–160 m. Also explain whether the retrieval can resolve this layer, interpolating profiles to a finer grid does not add independent vertical information, and smoothing may weaken or miss real temperature gradients. For the sign-based classification used here, the sign of Ri depends on the sign of this temperature difference. Wind-shear uncertainty affects Ri’s magnitude but generally cannot change its sign. The key question is therefore whether the remaining temperature-difference uncertainty is small enough to identify stable and unstable conditions reliably?
Uncertainty in surface heat-flux also needs attention, particularly when the flux is close to zero. Please show how plausible errors in both measurements affect the overall frequency, monthly pattern, and event durations. The results may prove to be robust, but the current sampling confidence intervals do not establish this.
2. Clarify how “decoupling” relates to boundary-layer mixing.
The paper defines decoupling as disagreement between surface and hub-height stability classifications. In the wider boundary-layer literature, decoupling usually refers to weak turbulent exchange between layers. These concepts are related, but classification disagreement alone does not demonstrate reduced exchange. An unstable surface beneath stable air is consistent with shallow surface mixing, but stable air can still contain shear-driven or intermittent turbulence. Similarly, matching stability classifications do not necessarily demonstrate strong mixing between the layers. Please explain this distinction and describe the main result as “surface–hub stability disagreement,” or clearly qualify “decoupling” as the operational definition used in this study. Stronger claims about turbulence being confined below the rotor should be supported by additional observations or presented as possible explanations. Also, because Ri’s sign follows the temperature difference, sign-based Ri and static-stability classifications are not independent checks.
3. Check the exclusion of near-neutral conditions.
The manuscript states that near-neutral surface conditions were excluded by requiring (|L|>50) m. As written, this appears incorrect: near-neutral conditions are approached as the absolute value of L becomes large. Please check whether this is an error in the description or implementation. Would recommend repeating the sensitivity test using a justified criterion based on (z/L), which approaches zero under neutral conditions. This correction matters because the paper uses this test to support the robustness of its main result.
4. Account for dependence between observations and possible sampling bias.
Each 20-minute sonic estimate is assigned to two 10-minute Ri observations. These observations are not independent, and atmospheric persistence adds further dependence. Please explain how this is handled when calculating confidence intervals. Resampling consecutive time blocks would be more appropriate than treating every timestamp as independent. Please also report valid sample counts by month and hour. Assess whether missing measurements are associated with particular weather, cloud, wind, or sea conditions. Normalizing by available observations does not remove bias if some conditions are less likely to be sampled.
Event durations may also depend on missing data and small, uncertain sign changes. Test the results using a consistent 20-minute averaging period and explain how gaps and deployment boundaries affect event identification.
5. Separate changes in temperature gradients from changes in wind shear.
Increasing Ri is interpreted as stronger stratification at hub height. However, Ri can also increase because wind shear decreases. Please show these two contributions separately before attributing the monthly changes to stronger stratification.
Similarly, changes in L alone do not establish changes in surface buoyancy forcing, because L also depends on friction velocity. Showing heat flux and friction velocity separately would strengthen the explanation.
Representative vertical profiles would help support the proposed shallow unstable layer beneath stable air. The conclusions should also reflect that the observations cover one location during one summer and early fall.
6. Clarify what is known about the impact on power.
The paper discusses implications for power production and wakes, but does not quantify them. Decoupling during 30.6% of observations does not imply a corresponding energy loss or significant wake impacts, as these are mostly transition events and could have lesser impact on actual energy. The decoupled periods also have lower average wind speeds than coupled-stable periods. A direct power comparison would therefore mix stability effects with wind-speed differences.
A useful first step would be to apply a representative turbine power curve to individual observations and estimate the fraction of potential energy production occurring during decoupled periods. This would show their energy relevance, although it would not measure a loss caused by decoupling.
A further analysis could compare wake-model power predictions using surface-based and hub-height stability information while keeping other inputs consistent. A full power study is not essential for this observational paper, but without one, the authors should state clearly that the size of the power impact remains unknown.
A few smaller points need clarification:
- Distinguish the dominant disagreement category, 30.6%, from total classification disagreement, approximately 34.2%.
- Would recommend stating the layer from 120–160 m as a hub-height layer rather than the full rotor layer.
- Please distinguish mean and median event durations in Figure 12: September’s mean decreases, but its median exceeds those of June and July.
- Check and explain the unusually asymmetric confidence interval reported around the 30.6% estimate.
Citation: https://doi.org/10.5194/wes-2026-115-RC2
Model code and software
BargeStabilityDecoupling V. Vasquez-Barros https://github.com/valeria-vasquez-barros/BargeStabilityDecoupling
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 367 | 147 | 105 | 619 | 129 | 94 |
- HTML: 367
- PDF: 147
- XML: 105
- Total: 619
- BibTeX: 129
- EndNote: 94
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
This study investigates differences in atmospheric stability between surface and rotor layer heights at an offshore site during WFIP3 campaign and how those differences can lead to misrepresentations in lower-level stability interpretations. This is a timely topic for offshore wind energy research with important implications for both campaign and modeling frameworks. As a whole, the overall story and structure of the presented manuscript are clear and easy to follow. However, there are some major scientific points that require further justification, as well as multiple inconsistencies and misspellings in the writing necessary to address before referring to the manuscript for publication. As such, I recommend major revision; please see the specific comments and suggestions below and address accordingly.
General Comments and Suggestions:
Technical Comments:
Line 61: Since “Section does not appear at the beginning of the sentence, it should be abbreviated as “Sect.” per the journal submission guide.
Line 78: used lidars: Please change “used lidars” to “employed lidars”; it took me a moment to realize you weren’t saying used, as in ‘not new’.
Line 88: “The barge was deployed on tation for a total of 85 days, over the 120 day period, as it was recovered and towed to shore for 2 large wave events events (Table 1)”. Station? Also, the word ‘events’ is repeated.
For the same sentence and the one that follows as, “During the deployed periods, its 16 instruments document the MABL in the region, collecting wind and thermodynamic data, as well as precipitation profiles, and wildlife information (birds and bats),” the tenses switch from past (“recovered”) to present (“document”). Make sure to fix this so it is consistent here and anywhere else this occurs in the text.
Table 2: a border at the bottom of the table would make this table cleaner.
Chapter 2: There are a lot of subsections for Chapter 2 that I think can be organized to be a little easier to digest. Something like this:
Line 109: “Tropospheric Remotely Observed Profiling via Optimal Estimation,…” This needs a reference/references!
Line 118: There should be a comma after altitude, not a period.
Line 124: ‘…relative to water surface assuming a 1.52 m barge draft)”… I think this is good information on the barge height to introduce earlier when first discussing the barge at the beginning of the Chapter, as it was a question I originally had at that section.
Figure 3: the caption needs a period at the end. As an aside, Figure 3 strikes me more as belonging in the Appendix, though that’s personal preference.
Paragraph 145: There are inconsistencies in the use of the full word “minute” and just “min.” Please pick one here and elsewhere – I personally vouch for the full word.
Line 157: Figure 2 shouldn’t be mentioned after Figure 3. I suggest switching them and moving both what is now Figure 2 and Table 1 closer to 2.7 where they are referenced. There are also places where you mention the Figure + panel, such as Figure 4a, before introducing Figure 4 itself, which needs to be avoided.
Figure 4: is there a specific reason you have so many intervals for the wind classifications wind rose? I suggest reducing them to 5 or 6 to better see the differences between heights.
Line 174: In this sentence, make the w in “We” lowercase and add commas before and after the parentheses.
Line 225: In this checklist, please add a space before each / to maintain consistency with the rest of the text. There are also inconsistencies in the section title capitalization throughout (e.g., 3.2 Coupled and Decoupled Stability has all non- text capitalized, but in 3.3 ‘Event identification’, Identification isn’t). According to the journal rules ,
Line 243: There needs to be some sort of introductory sentence into the results. Add a sentence or two about how you begin with assessing meteorological conditions, starting with wind speed shown in Figure 4 wind roses. Otherwise, jumping straight into describing 4a comes out of nowhere for the reader.
Figure 4 caption: Add the word ‘panel’ here before (a) and (b) and any other reference to figure letters.
Line 256: Since you already define MABL, you can just replace atmospheric boundary layer in ‘the resulting destabilization of the lowest portion of the atmospheric boundary layer.’ with MABL. Same with the Obukhov length and Bulk Richardson naming, you already named them L and Ri so just use those.
Line 272: Add comma after 8 and capitalize B in bulk to remain consistent (or, as mentioned above, just use the abbreviations).
Line 286: Remove extra parenthesis
Figure 9-11: the heatmap values are quite small-- I think that can just be solved by increasing the figure size, but also consider removing the decimal places in Figure 10 since they aren’t super necessary. On a semi-related note, three figures would probably be better combined since they’re somewhat connected, which would also fix the abruptness of jumping between them.
Line 301: Take out one of the ‘of decoupling’s.
A stylistic comment throughout: there are quite a few occurrences of short paragraphs. While a few of these might help improve readability, others would arguably serve better as a single paragraph. Some examples of the latter are the last two paragraphs of 3.1, the first two of 4.3, the paragraph at 305 and the single sentence after it.