Impacts of Climate Change on the Offshore Wind Industry in Metropolitan France: Insights from the 2C NOW Project

Kervella, Youen; Chevallier, Tessa; Oueslati, Boutheina; Raillard, Nicolas; Yates, Marissa; Pimoult, Matéo; Poppeschi, Coline; Patra, Anindita; Luxcey, Neil; Guinot, Florent; Dubus, Laurent

doi:10.5194/wes-2025-266

Preprints

https://doi.org/10.5194/wes-2025-266

Preprints

15 Dec 2025

| 15 Dec 2025

Status: this preprint is currently under review for the journal WES.

Impacts of Climate Change on the Offshore Wind Industry in Metropolitan France: Insights from the 2C NOW Project

Youen Kervella, Tessa Chevallier, Boutheina Oueslati, Nicolas Raillard, Marissa Yates, Matéo Pimoult, Coline Poppeschi, Anindita Patra, Neil Luxcey, Florent Guinot, and Laurent Dubus

Abstract. The offshore wind energy sector in France is poised for significant growth, with ambitious targets to install 18 GW of offshore capacity by 2035 and 45 GW by 2050. This expansion is crucial for achieving France’s goal of generating 20 % of its electricity from offshore wind by 2050. The 2C NOW project, led by France Energies Marines, investigates the impact of climate change on metocean conditions along France’s three maritime fronts (two representative points for each seafront). This is critical in the context of sector development, which aims to support long-term energy transition goals and the sustainability of wind farms over several decades. Using global climate models from the CMIP6 project (IPCC), a statistical downscaling of the datasets is performed using the CDF-t method (Michelangeli et al., 2009) based on the best available numerical reanalyses (selected using extensive in-situ data). The mean conditions show a general decrease in wind speed and wave height (multi-model average) for all SSP scenarios and for all seafronts in continental France. This downward trend is more significant for the long-term future period (horizon 2100), on the Atlantic and Mediterranean coasts. These average trends are nevertheless accompanied by strong model uncertainties. Regarding extreme conditions, an increase in extreme values of significant wave height is observed for the future climate scenarios, while there is less consensus on the wind speed. Water levels show significant increasing trends, regardless of the seafront or the conditions concerned.

Received: 01 Dec 2025 – Discussion started: 15 Dec 2025

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.

Download & links

Youen Kervella, Tessa Chevallier, Boutheina Oueslati, Nicolas Raillard, Marissa Yates, Matéo Pimoult, Coline Poppeschi, Anindita Patra, Neil Luxcey, Florent Guinot, and Laurent Dubus

Status: final response (author comments only)

RC1:
'Comment on wes-2025-266', Anonymous Referee #1, 20 Feb 2026
Review of “Impacts of Climate Change on the Offshore Wind Industry in Metropolitan France: insights from the 2C NOW project”
This study uses CMIP6 data and a statistical downscaling approach to assess the climate change impact on wind and wave parameters at a number of French coastal sites. The subject is relevant for offshore wind development in France. However there are quite a few points that need to be addressed or substantially revised before the study be considered publication.
I won’t go into all details in this round. I’ll put focus on the methodology, as it is the soul of the results.
I think the title is too broad with the use of “offshore wind industry”.

This study has quite some references for climate change assessment, but very little on the method for calculating extreme values.

The method of calculating extreme values in this study: periodical maximum value from each month is used before applying a distribution function for fitting – is it the Gumbel that is used here? The reviewer doubts its validity of choosing monthly maximum values. When applying a fitting function such as the Gumbel, it is required that the extreme event samples come from the same/similar mechanism. The extremes in summer and in winter in the studied areas are of different weather systems. One may argue that what I said is only a theory, and the samples the authors collect do perform a good distribution. Then the reviewer would encourage the authors to show such distribution with assessment of uncertainties. It would also help if you use blocks larger than 1 month, say 3 months, 6 months and 12 months, to see the sensitivity. Compare with the measurements please. If the readers are not convinced that this calculation is accurate, we are not motivated to read to the end of the paper to see the impact.

The model data have different temporal and spatial resolutions. It is easier to accept the analysis of climate change to a qualitative degree, if the historical data and the future data are of the same temporal and spatial resolution. To move from there to the assessment of impact, we need the absolute magnitude to be more accurate. Then it becomes relevant to ask how the model data used here with a temporal resolution of 3 hours and 1 hours, compare with the IEC standard, which is often referring to a temporal resolution of 10 min.

The introduction of the “statistical downscaling” method is way too brief here. Please elaborate and show some examples what it did: from the original coarse resolution to the downscaled one. Please list the added values from this approach for the analyzed parameters. Currently, some of this is packed in Figure 1. Please improve the presentation here: what is in each subplot? What are on the x- and y-axes? What are we supposed to see to understand the added values?

The authors introduced measurement data. How has each of them be used?

The overall writing of the paper needs substantial improvement. Please make sure each variable in an equation is explained. For instance ξ in Eq. 3. Also a sentence like “More details can be found in (Raillard et al in prep)” is difficult for the readers to take.
Citation: https://doi.org/10.5194/wes-2025-266-RC1
- AC1:
  'Reply on RC1', Youen Kervella, 02 Apr 2026
  Dear reviewer,
  
  We would like to thank you for your thorough reading and for detailing all the technical points to be addressed in the manuscript, both in terms of form and content. Below are the responses to all these points:
  
  I think the title is too broad with the use of “offshore wind industry”.
  
  We thank the reviewer for this comment. To be more accurate, we changed the title to: “Climate Change Impacts on Offshore Wind Power and Components Design Along the French Coasts: Insights from the 2C NOW Project”.
  
  This study has quite some references for climate change assessment, but very little on the method for calculating extreme values.
  
  Thank you for your feedback and for your vigilance on the lack of references linked to the method for calculating extreme values. In addition to the link to this article submitted in the meantime (https://arxiv.org/pdf/2603.08101v1), also from the 2C NOW project, we have added references in the article on these methods (Coles, 2001), particularly applied to the ocean (Meucci et al., 2023; Young et al., 2011) and the offshore sector (Goda, 2010; IEC, 2019) or accounting for climate change (Ewans and Jonathan, 2023; Meucci et al., 2023).
  
  The method of calculating extreme values in this study: periodical maximum value from each month is used before applying a distribution function for fitting – is it the Gumbel that is used here? The reviewer doubts its validity of choosing monthly maximum values. When applying a fitting function such as the Gumbel, it is required that the extreme event samples come from the same/similar mechanism. The extremes in summer and in winter in the studied areas are of different weather systems. One may argue that what I said is only a theory, and the samples the authors collect do perform a good distribution. Then the reviewer would encourage the authors to show such distribution with assessment of uncertainties. It would also help if you use blocks larger than 1 month, say 3 months, 6 months and 12 months, to see the sensitivity. Compare with the measurements please. If the readers are not convinced that this calculation is accurate, we are not motivated to read to the end of the paper to see the impact.
  
  We thank the reviewer for this important remark. The text has been amended to reflect some changes and hopefully to make it clearer. We used a non-stationary GEV approach, based on monthly maxima. The parameters are estimated from the data, so the Gumbel distribution is covered when the parameter $\xi \to 0$, but our model is broader and can cover bounded to heavy tails. We agree that the extremes in the different seasons come from different distributions, that is the reason why the model is made non-stationary, by allowing its parameters to vary smoothly with time. We also added a reference to the paper submitted in the meantime (as mentioned above) where we extend further this model and provide more details about the results. We believe that this discussion goes beyond the scope of the present paper and hope the additions made here are sufficient.
  
  The model data have different temporal and spatial resolutions. It is easier to accept the analysis of climate change to a qualitative degree, if the historical data and the future data are of the same temporal and spatial resolution. To move from there to the assessment of impact, we need the absolute magnitude to be more accurate. Then it becomes relevant to ask how the model data used here with a temporal resolution of 3 hours and 1 hours, compare with the IEC standard, which is often referring to a temporal resolution of 10 min.
  
  We thank the reviewer for pointing out this difference. Regarding model temporal resolution, we imagine that you are referring to climate models versus reanalyses? Indeed, these two datasets can have different input temporal resolutions. Therefore, in this case, we duplicated the climate model values (without interpolation) to apply the CDF-t method, which acts as a bias correction (this is the commonly used approach). For the IEC standard, we believe that the 10-minute resolution is more applicable to mean winds measured in-situ or to high-resolution regional models such as WRF or AROME (https://hal.science/hal-05380788/). Currently, climate data do not offer such a high resolution, but an hourly time resolution is sufficient for our wind applications, whether for energy production (impact models like PyWake use 1-hour inputs) or for design calculations, which are based on wind/wave environments.
  
  The introduction of the “statistical downscaling” method is way too brief here. Please elaborate and show some examples what it did: from the original coarse resolution to the downscaled one. Please list the added values from this approach for the analyzed parameters. Currently, some of this is packed in Figure 1. Please improve the presentation here: what is in each subplot? What are on the x- and y-axes? What are we supposed to see to understand the added values?
  
  Thanks to the reviewer for pointing out this lack of clarity on the "statistical downscaling" method, which we have renamed "bias correction" to be clearer. We have detailed in the text the explanatory paragraph of the CDF-t (Michelangeli et al., 2009) as well as its advantages compared to a classical quantile-quantile method. We have also commented better on Figure 1 as requested, insisting on the correction of the full distribution of the variable (in this case significant wave height) for each climate model, as well as the seasonal cycles of these climate models which closely follow that of the reference (reanalysis), and the annual means of the climate models which exhibit less variability around this reference.
  
  The authors introduced measurement data. How has each of them be used?
  
  We thank the reviewer for this remark. In the current paper, in-situ measurements were used to determine the best reanalyses for each parameter (wind speed, significant wave height, and water level) and for each French coastline. Only long-term series (over 10 years) were selected, verified (QC), and then used as references to establish the scores of each reanalysis.
  
  The overall writing of the paper needs substantial improvement. Please make sure each variable in an equation is explained. For instance ξ in Eq. 3. Also a sentence like “More details can be found in (Raillard et al in prep)” is difficult for the readers to take.
  
  We thank the reviewer for his complete proofreading and for noting these errors. The definitions of the different variables in each equation have been added. A preprint of Raillard et al. has been deposited on arxiv (https://arxiv.org/pdf/2603.08101v1) pending the editor's decision on this article. We will do our best to improve the overall quality of the article's writing.
  
  Thank you again for your time.
  
  Citation: https://doi.org/10.5194/wes-2025-266-AC1

RC2: 'Comment on wes-2025-266', Anonymous Referee #2, 09 Mar 2026

AC2: 'Reply on RC2', Youen Kervella, 02 Apr 2026

Dear reviewer,

We would like to thank you for your thorough reading and for detailing all the technical points to be addressed in the manuscript, both in terms of form and content. Below are the responses to all these points:

Major comments

L72: One must be very careful with this. Are these averaged values of U and V, or the average of the wind speed? Note that these may differ substantially and can affect your conclusions. The 3-hourly instantaneous zonal and meridional components are available from the ESGF repositories.

We thank the reviewer for this remark. Indeed, the data we used were average wind speeds (sfcWind in ESGF repositories). We will mention it in the paper. In the present study, we only use daily data of average wind speed and we plan to use 3 hourly data in future work.

The extrapolation of the winds to 150 m height is not well explained. Please elaborate.

Thanks to the reviewer for this comment. The vertical extrapolation from 10 meters to 150 meters was performed using a classical power law. Rather than using a constant value for the exponent α, for example α =0.12 as recommended by DNV (DNV-RP-C205n 2021) for open sea with waves, we used a time-of-day and month-dependent value for α, calculated from the 10 m and 100 m wind speed of the CERRA reanalysis. This method is inspired by recent work by the Copernicus Climate Change Service for Energy (C3SEnergy) which computed a gridded α matrix on a 0.25°x0.25° grid (24 (hours) x 12 (months)), from 10 years of ERA5 hourly data of 10 m and 100 m WS (for more detail, see the Confluence web page of the product).

I believe what you refer to as ”downscaling” is merely a bias correction based on quartiles. Is that correct? The downscaling section warrants further elaboration.

The reviewer is absolutely correct, as it is indeed a bias correction for the entire distribution. We have corrected the description in the paper and the section has been renamed "Bias correction".

Table 5. It would be useful to show a map of the sites. Also, are the areas homogeneous? Are you sure all these points are over water in all GCMs? If part of the GCM grip point is over land, other factors could be introduced

Thank you for this entirely justified comment. We have checked the wind GCM and all models and scenarios share the same coordinates, each site has their own extracted coordinates. All extracted coordinates are over water. Below is the summary table, showing the distance between the representative points and the nearest grid point of the models:

Site	Lon	LAT	CMIP6 LON	CMIP6 LAT	DISTANCE (KM)
EChan	0.2	50.1	0.227338	49.892002	4.5
WCAN	-2.75	49.2	-2.536347	48.853733	4.7
NATL	-3.85	47.45	-3.561458	47.330273	4.7
SATL	-1.9	45.6	-1.923157	45.833633	4.4
WMED	3.3	42.65	3.270830	42.813477	3.6
EMED	4.75	43.05	4.605993	43.080486	3.5

In the Limitation section, the possible changes in shear are also not included in this study, and it should be mentioned. Other factors that can introduce uncertainty should also be considered.

We thank the reviewer for this very relevant comment. We added it to the Limitations section. We will soon be studying (in a new R&D project) other atmospheric parameters like shear and turbulence, using higher spatial resolution datasets.

Minor comments and corrections

L7: CMIP6 and IPCC and two terms not directly linked. Yes, CMIP6 was done to serve as a basis for the 6th report. Please rewrite.

Thank you for pointing out this error. We have corrected the sentence in the paper: “Using climate data from the CMIP6 project, which served as the basis for the latest IPCC report (AR6), ..”

L16: “Climate change is expected to profoundly reshape the environmental conditions which govern offshore wind energy production.” I am not sure this sentence is true; at least it is exaggerated. Can you provide a reference?

Thank you to the reviewer for pointing out this sentence, which is indeed ambiguous. We have removed the sentence from the paper.

L26: Again, can you provide a reference to the statement “... leading to significant losses in energy production”?

A reference to this assertion is given previously: Rapella et al. (2023). The authors of this article also cite another reference that supports this view: Cutululis N A, Litong-Palima M and Sørensen P E 2012 Offshore Wind Power Production in Critical Weather Conditions (Copenhagen: European Wind Energy Association (EWEA)) (https://backend.orbit.dtu.dk/ws/portalfiles/portal/7894079/Offshore_Wind_Power_Production.pdf).

L28: “due to changes in atmospheric circulation and land-sea thermal contrasts — a phenomenon known as global terrestrial stilling” I thought the stilling was mostly associated with changes in surface roughness. Please check that the statement is correct

We thank the reviewer for pointing out this error. Indeed, global terrestrial stilling is due to large-scale changes in atmospheric circulation, but primarily to increased surface roughness over the continents. We therefore modify the wording in the manuscript.

Table 3. I don’t think this is the actual resolution. These models have dlat x dlon grid spacing. Please check this statement.

Thanks to the reviewer for pointing out this point. We checked and the horizontal resolution (rounded to 10km) is the square root of the surface area of the Earth divided by the number of grid points for the atmosphere. This information is described in the table AII.5 of the AR6 IPCC report (IPCC, 2021: Climate Change 2021: The Physical Science Basis. Contribution of Working Group I to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change [Masson-Delmotte et al.]. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA, 2391 pp. doi:10.1017/9781009157896).

We added this point to the manuscript for clarity.

Figure 1. Please label the different panels on the plot (a), (b), etc. Also, what is the y-axis on each plot? It is not clear in the top row of figures.

Thank you for pointing out this technical detail. The figure has been updated in the manuscript.

Figure 4. This is not the same period as in the previous figure. It would be good to be consistent.

Indeed, the reviewer is right. The climatic period used for the extreme analyses is 30 years, whereas it is 20 years for the other calculations. This is described in line 132 of the manuscript.

L208–209: “Intra- and inter-farm wake effects are not considered in this study.” Please explain if this could be important.

In the manuscript, the idea is to translate the evolution of wind speed into the evolution of wind power output. We therefore use a single turbine and its power curve for conversion. In reality, offshore wind farms are composed of several turbines, and overall, this leads to a reduction in power due to the wake effect. The wake effect refers to the trail of slower, more turbulent air produced behind operating wind turbines or wind farms. By impacting downstream wind quality, wakes can reduce the energy yield of neighboring turbines and farms by several percent, while increasing mechanical wear and tear on the turbines located within that flow (https://www.science.org/doi/10.1126/science.aau2027 and https://iopscience.iop.org/article/10.1088/1742-6596/2265/2/022008). We added these references to the manuscript.

L225: Please use SI units. These are required by WES.

Thank you for pointing out this error. The references to m/s have been replaced with m.s⁻¹

Table 11: The turbine capacity should be listed here, not the height.

We thank the reviewer for this comment. The legend for Table 11 has been modified to show the turbine type and its capacity.

Figure 6: The legend at the top is too small to understand what the different colours mean. It will be better to replace these with colour bars.

We thank the reviewer for this comment. The figure's legend has been enlarged and new colors have been introduced.

Larsen et al (2023) on L410 was rejected by the journal and should not be cited.

Thank you to the reviewer for pointing out this error. The reference has been replaced with the following: https://www.frontiersin.org/journals/energy-research/articles/10.3389/fenrg.2024.1404791/full

Writing

The overall quality of the paper’s writing needs significant enhancement to meet the required academic standards.

Thank you for this comment; many passages of the manuscript have been rewritten in a more academic style.

Length

The manuscript length is adequate.

Figures and tables

Many figures and tables require revision as outlined above. Overall, figure captions should be revised to ensure they contain all required information.

We thank the reviewer for their vigilance. The figure captions in the document have been revised.

Title

I am not sure whether non-French readers are familiar with the meaning of the term “Metropolitan” in the title. Perhaps the term “European” France is more familiar.

Thank you to the reviewer for this comment; to avoid any confusion, the title has been changed to: “Climate Change Impacts on Offshore Wind Power and Components Design Along the French Coasts: Insights from the 2C NOW Project”.

Abstract and Plain text summary

The abstract his basically fine, but too much text is devoted to describing the project and objectives, and very little to the method and the study’s conclusions. I suggest reorganising it when the rest of the manuscript is revised.

The authors agree with the reviewer and propose a new abstract where the results are more detailed:

“The offshore wind sector in France is rapidly expanding, making it essential to anticipate how climate change may alter future metocean conditions. This study assesses projected changes in wind, wave and water-level conditions along the French Atlantic, Channel and Mediterranean coasts using CMIP6 climate models, which served as the basis for the latest IPCC report (IPCC, 2022). These models have been bias-corrected with the CDF-t method, whose reference analysis has been validated against in-situ observations.

Downscaled CMIP6 projections indicate a general decrease in mean wind speed and significant wave height across all scenarios, reaching –1.5 % to –6 % by late century depending on location, nevertheless accompanied by strong model uncertainties. Mean sea level is projected to rise by +15 to +17 cm by 2100. In contrast, extreme significant wave heights increase markedly (up to +15 %), while extreme wind changes show weaker consensus. Storm surge extremes also intensify slightly. These evolving conditions translate into moderate reductions in annual energy production for a 15 MW turbine (–1 % to –8 %) but a slight decrease in fatigue loads (–1 % to –2 %). Morphodynamic modeling of a representative landfall site shows increased shoreline retreat and enhanced sensitivity to extreme events.

Overall, results highlight the combined influence of mean trends and changing extremes on offshore wind performance, design conditions, and coastal infrastructure. They emphasize the need to account for model uncertainty and to integrate climate‑resilient approaches into future offshore wind planning.”

Thank you again for your time.

Citation: https://doi.org/10.5194/wes-2025-266-AC2

Youen Kervella, Tessa Chevallier, Boutheina Oueslati, Nicolas Raillard, Marissa Yates, Matéo Pimoult, Coline Poppeschi, Anindita Patra, Neil Luxcey, Florent Guinot, and Laurent Dubus

Viewed

Total article views: 688 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	BibTeX	EndNote
397	242	49	688	38	39

HTML: 397
PDF: 242
XML: 49
Total: 688
BibTeX: 38
EndNote: 39

Views and downloads (calculated since 15 Dec 2025)

Month	HTML	PDF	XML	Total
Dec 2025	115	63	7	185
Jan 2026	107	56	4	167
Feb 2026	87	68	14	169
Mar 2026	65	43	17	125
Apr 2026	23	12	7	42

Cumulative views and downloads (calculated since 15 Dec 2025)

Month	HTML	PDF	XML	Total
Dec 2025	115	63	7	185
Jan 2026	107	56	4	167
Feb 2026	87	68	14	169
Mar 2026	65	43	17	125
Apr 2026	23	12	7	42

Viewed (geographical distribution)

Total article views: 663 (including HTML, PDF, and XML) Thereof 663 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 11 Apr 2026

Short summary

France aims for major offshore wind growth. We examined future climate impacts on wind, waves, and water levels. Results suggest that mean winds and waves may weaken, but extreme waves and sea levels will increase. These trends are nevertheless accompanied by strong model uncertainties. These findings are necessary for designing durable offshore wind farms in France and ensuring reliable energy production for decades to come.


Total:	0
HTML:	0
PDF:	0
XML:	0