the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 04 Sep 2026
Operational Weather Windows for Offshore Wind in the Taiwan Strait: Long-Term Variability, Regional Wind Modes, and Climate Linkages
Abstract. Previous studies have either focused on shorter operational records or on ENSO-related variability over more limited periods, leaving the role of regional wind modes and longer-term climate context less explored. Here, we use more than 60 years of ERA5 reanalysis data to examine weather-window variability relevant to Taiwan's offshore wind development and to identify the proximal regional wind modes and large-scale climate linkages associated with this variability. Compared with winter, summer provides the greatest number of operational weather windows and exhibits relatively stable year-to-year variability, making it the primary season for offshore operational activities near Taiwan. Interannual variability in June–July–August mean weather-window counts is dominated by two regional wind modes across the Taiwan Strait, with secondary contributions from episodic tropical cyclone activity and broader climate linkages involving the Western North Pacific summer monsoon and ENSO-related sea surface temperature signal. Together, the regional wind modes, tropical cyclone activity, and climate-linkage indicators account for more than 60 % of the variance in summer weather-window availability. Notably, during the period corresponding to the onset of Taiwan's offshore wind development (2018–2024), summers exhibited near-maximum accessibility relative to other 7-year windows in the 65-year record, indicating that such favorable conditions should not be assumed to persist. These results demonstrate the value of long-term weather-window analysis for addressing how offshore wind operating conditions are associated with broader climate variability and for supporting risk-informed construction and operations planning.
- Preprint
(2322 KB) - Metadata XML
-
Supplement
(545 KB) - BibTeX
- EndNote
Status: closed
-
RC1: 'Comment on wes-2026-121', Anonymous Referee #1, 18 Sep 2026
The comment was uploaded in the form of a supplement: https://wes.copernicus.org/preprints/wes-2026-121/wes-2026-121-RC1-supplement.pdfCitation: https://doi.org/
10.5194/wes-2026-121-RC1 -
RC2: 'Comment on wes-2026-121', Anonymous Referee #2, 18 Sep 2026
Dear authors, I recommend that this manuscript should be reconsidered after major revisions.
The article lack information on the validity of the data basis, both on the ERA5 data used, but also the weather window criterias. While it may be interesting for a synoptic meteorology audience and specialists in PCA, I think that for the WES journal the article needs to be improved on several aspects listed below.
The suggested major revisions are:
1) ERA5 data validation:
a) The ERA5 100m winds should be compared against measurements, not only CWA-WRF data. If no measurements can be used, the validity of the CWA-WRF data should be demonstrated against the measurements mentioned in the paper.
b) The ERA5 wave timeseries should also be validated, against measurements. The spatial variations in wave conditions across the straight should also be investigated.
c) for a) and b), scatter plots including quantile-quantile and other statistics should be provided.
d) the consistency of ERA5 over the long term, in particular before 1979, should be checked. Include a change-point analysis for wind speed, significant wave height and SST, and conclude on the validity of the chosen long-term period.
2) Weather windows criteria:
a) Wind speeds in Table S1 are provided at 10m. How these should be used for 100m wind need be explained. If 10m need to be used (I guess they need to be used for some vessels), 10m should be validated in the same way as in 1). Note: ERA5's understimation is more severe closer to the surface, this should be accounted for.
b) Table S1 mentioned wave height in m, it should explained whether this is significant wave height or maximum wave height. (Moore 2014) uses maximum wave height in their Table 3 for Cable Laying. If maximum wave height is used, the paper needs to explain how it is computed from the significant wave height.
c) Same as above, for the wind speed: are these 3-s, 1-minute, 10-minute, 1-hour averages? ERA5 provide hourly values but due to grid size the effective time resolutions is longer, this needs to checked, discussed and concluded upon (including correction factors).
c) The paper needs to explain why peak period has not been accounted for. Long swell waves, maybe occuring during summer, can make floating installation vessels operations risky and difficult.3) Include winter conditions
From a practical perspective, summer periods do not seem problematic for offshore operations. Winter periods may be more relevant to the audience of this journal: high production, but also almost no calm periods to fix the turbines when they fail. Understanding how these calm period form would be more relevant, in my opinion. I recommend to include winter conditions, and cut on the discussion with the PC correlations elsewhere in the paper.
4) Paper clarity
a) Explain in simpler terms what are the blue bars in Figure 3, with an example.
b) Sections 2.4, 2.,5, 2.6 are very dense and refer to a vast amount of data science study which I had difficulty to see the relevance of. Revisit which one of these checks are actually needed, and if/when they are, include a figure.
c) Simplify the text and discussion.
Citation: https://doi.org/10.5194/wes-2026-121-RC2 -
RC3: 'Comment on wes-2026-121', Anonymous Referee #3, 14 Oct 2026
Review of wes-2026-121
Operational Weather Windows for Offshore Wind in the Taiwan Strait: Long-Term Variability, Regional Wind Modes, and Climate Linkages
The manuscript relates the occurrence of operational weather windows for offshore wind farms in the Taiwan Strait to regional and global climate modes. Weather windows are defined as ERA5 wind below a simple threshold of 12 m/s at a single grid point, which represents an area of 4° × 4 °. The weather windows are further correlated to large-scale climate indices.
The manuscript is challenging to read, which obscures the study's objective. If the only criterion for identifying weather windows is that the 100-meter wind speed must be below 12 m/s, it is essential to analyse the ERA5 data to determine when and how frequently this condition occurs. While there are long-term correlations between large-scale climate patterns, such as El Niño and the Southern Oscillation Index (SOI), the paper does not demonstrate that forecasting these indices is more effective than predicting the ERA5 wind data. Additionally, the article fails to provide adequate validation of the ERA5 100-meter winds against observations, nor does it specify the criteria for the weather windows.
The introduction of Empirical Orthogonal Function (EOF) analysis is unclear and appears irrelevant to the overall analysis. Additionally, the study simplifies the problem by focusing on a single season. While the findings may interest researchers in synoptic meteorology and large-scale climate dynamics, the paper ultimately fails to demonstrate the practical significance of these concepts for wind energy applications. Therefore, I must recommend that it be rejected for publication in WES.
Some further comments regarding the manuscript:
- The abstract's first sentence is introduced all of a sudden without explaining the paper's theme.
- Many necessary figures and one table are shown in the supplementary material.
- Data such as the CWA-WRF is introduced but not used in the study.
- Abbreviations are multiply defined.
- Problems with ordering and definitions. Make sure terms are defined before being used. E.g. weather-windows counts in L71 are only defined later on in section 2.2
Citation: https://doi.org/10.5194/wes-2026-121-RC3
Status: closed
-
RC1: 'Comment on wes-2026-121', Anonymous Referee #1, 18 Sep 2026
The comment was uploaded in the form of a supplement: https://wes.copernicus.org/preprints/wes-2026-121/wes-2026-121-RC1-supplement.pdf
-
RC2: 'Comment on wes-2026-121', Anonymous Referee #2, 18 Sep 2026
Dear authors, I recommend that this manuscript should be reconsidered after major revisions.
The article lack information on the validity of the data basis, both on the ERA5 data used, but also the weather window criterias. While it may be interesting for a synoptic meteorology audience and specialists in PCA, I think that for the WES journal the article needs to be improved on several aspects listed below.
The suggested major revisions are:
1) ERA5 data validation:
a) The ERA5 100m winds should be compared against measurements, not only CWA-WRF data. If no measurements can be used, the validity of the CWA-WRF data should be demonstrated against the measurements mentioned in the paper.
b) The ERA5 wave timeseries should also be validated, against measurements. The spatial variations in wave conditions across the straight should also be investigated.
c) for a) and b), scatter plots including quantile-quantile and other statistics should be provided.
d) the consistency of ERA5 over the long term, in particular before 1979, should be checked. Include a change-point analysis for wind speed, significant wave height and SST, and conclude on the validity of the chosen long-term period.
2) Weather windows criteria:
a) Wind speeds in Table S1 are provided at 10m. How these should be used for 100m wind need be explained. If 10m need to be used (I guess they need to be used for some vessels), 10m should be validated in the same way as in 1). Note: ERA5's understimation is more severe closer to the surface, this should be accounted for.
b) Table S1 mentioned wave height in m, it should explained whether this is significant wave height or maximum wave height. (Moore 2014) uses maximum wave height in their Table 3 for Cable Laying. If maximum wave height is used, the paper needs to explain how it is computed from the significant wave height.
c) Same as above, for the wind speed: are these 3-s, 1-minute, 10-minute, 1-hour averages? ERA5 provide hourly values but due to grid size the effective time resolutions is longer, this needs to checked, discussed and concluded upon (including correction factors).
c) The paper needs to explain why peak period has not been accounted for. Long swell waves, maybe occuring during summer, can make floating installation vessels operations risky and difficult.3) Include winter conditions
From a practical perspective, summer periods do not seem problematic for offshore operations. Winter periods may be more relevant to the audience of this journal: high production, but also almost no calm periods to fix the turbines when they fail. Understanding how these calm period form would be more relevant, in my opinion. I recommend to include winter conditions, and cut on the discussion with the PC correlations elsewhere in the paper.
4) Paper clarity
a) Explain in simpler terms what are the blue bars in Figure 3, with an example.
b) Sections 2.4, 2.,5, 2.6 are very dense and refer to a vast amount of data science study which I had difficulty to see the relevance of. Revisit which one of these checks are actually needed, and if/when they are, include a figure.
c) Simplify the text and discussion.
Citation: https://doi.org/10.5194/wes-2026-121-RC2 -
RC3: 'Comment on wes-2026-121', Anonymous Referee #3, 14 Oct 2026
Review of wes-2026-121
Operational Weather Windows for Offshore Wind in the Taiwan Strait: Long-Term Variability, Regional Wind Modes, and Climate Linkages
The manuscript relates the occurrence of operational weather windows for offshore wind farms in the Taiwan Strait to regional and global climate modes. Weather windows are defined as ERA5 wind below a simple threshold of 12 m/s at a single grid point, which represents an area of 4° × 4 °. The weather windows are further correlated to large-scale climate indices.
The manuscript is challenging to read, which obscures the study's objective. If the only criterion for identifying weather windows is that the 100-meter wind speed must be below 12 m/s, it is essential to analyse the ERA5 data to determine when and how frequently this condition occurs. While there are long-term correlations between large-scale climate patterns, such as El Niño and the Southern Oscillation Index (SOI), the paper does not demonstrate that forecasting these indices is more effective than predicting the ERA5 wind data. Additionally, the article fails to provide adequate validation of the ERA5 100-meter winds against observations, nor does it specify the criteria for the weather windows.
The introduction of Empirical Orthogonal Function (EOF) analysis is unclear and appears irrelevant to the overall analysis. Additionally, the study simplifies the problem by focusing on a single season. While the findings may interest researchers in synoptic meteorology and large-scale climate dynamics, the paper ultimately fails to demonstrate the practical significance of these concepts for wind energy applications. Therefore, I must recommend that it be rejected for publication in WES.
Some further comments regarding the manuscript:
- The abstract's first sentence is introduced all of a sudden without explaining the paper's theme.
- Many necessary figures and one table are shown in the supplementary material.
- Data such as the CWA-WRF is introduced but not used in the study.
- Abbreviations are multiply defined.
- Problems with ordering and definitions. Make sure terms are defined before being used. E.g. weather-windows counts in L71 are only defined later on in section 2.2
Citation: https://doi.org/10.5194/wes-2026-121-RC3
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 312 | 116 | 86 | 514 | 64 | 96 | 91 |
- HTML: 312
- PDF: 116
- XML: 86
- Total: 514
- Supplement: 64
- BibTeX: 96
- EndNote: 91
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
