EGU General Assembly 2020
© Author(s) 2020. This work is distributed under
the Creative Commons Attribution 4.0 License.

Global wave resource classification and application to marine energy deployments

Iain Fairley1, Matthew Lewis2, Bryson Robertson3, Mark Hemer4, Ian Masters1, Jose Horrillo-Caraballo1, Harshinie Karunarathna1, and Dominic Reeve1
Iain Fairley et al.
  • 1Swansea University, College of Engineering, Bay Campus, Swansea, United Kingdom of Great Britain and Northern Ireland (
  • 2School of Ocean Sciences, Bangor University, Menai Bridge LL59 5AB, UK
  • 3College of Engineering, Oregon State University, 101 Covell Hall, Corvallis, OR 97331-2409, USA
  • 4CSIRO, GPO Box 1538, Hobart, TAS 7001, AUS

Understanding and classification of the global wave energy resource is vital to facilitate wave energy converter technology development and global roll-out of this promising renewable energy technology. To date, many wave energy converters have been developed based on Northern European wave climates; these are not representative of wave climates worldwide and may not be the best for large scale energy extraction. Classification of resources will highlight alternative wave resource types that may prove fruitful for deployment of future technologies; equally it will enable existing technology to define regions worthy of site exploration. Therefore k-means clustering is used here to classify the global resource from a data-driven, device agnostic perspective.

Parameters relevant to energy extraction (significant wave height, peak wave period, extreme wave height, spectral and directional properties) were extracted from the ECMWF ERA5 reanalysis dataset and used to split the global resource into 6 classes. Only areas within 3 degrees of land (feasible energy transport to user) were considered. The 6 classes returned by the analysis consisted of: 1) low energy high variability areas in enclosed seas; 2) low energy moderate variability areas in semi-enclosed seas and sheltered ocean coasts; 3) moderate energy areas, largely on eastern oceanic coastlines and influenced by local storm activity; 4) moderate energy areas primarily influenced by long period swell and largely on western oceanic coastlines; 5) higher energy areas, with variable conditions, primarily in the northern hemisphere; 6) highest energy areas, primarily on the tips of continents in the southern hemisphere. Consideration of device power matrices show that existing devices only perform well in classes 5 and 6, despite these areas having limited global coverage, which suggests devices should be developed for lower energy classes.

To refine global roll-out planning for existing devices, based on a request from a wave energy converter developer, a second classification is currently being developed with two additional constraints on the areas tested. These constraints are excluding any areas with a mean wave power of less than 15 kW/m (an often-used value for the lower power limit for commercial viability) and a maintenance constraint whereby wave heights must drop below 3m for a minimum of 48hrs per month. These newer results will be presented at the annual assembly and contrasted with our more device agnostic classification.

How to cite: Fairley, I., Lewis, M., Robertson, B., Hemer, M., Masters, I., Horrillo-Caraballo, J., Karunarathna, H., and Reeve, D.: Global wave resource classification and application to marine energy deployments, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-8135,, 2020

Comments on the presentation

AC: Author Comment | CC: Community Comment | Report abuse

Presentation version 1 – uploaded on 30 Apr 2020
  • AC1: Comment on EGU2020-8135, Matt Lewis, 30 Apr 2020

    Great work! It seems that only a few device designs are needed: an equatorial WEC device, a Europe and North America WEC device and a west-coast and east-coast southern ocean WEC device? Have you any idea on the impact of putting the wrong WEC design into the wrong location?


    • AC2: Reply to AC1, Iain Fairley, 01 May 2020

      Possibly - but one thing that occurs to me is that we haven't disentangled resource type/suitability from resource consistency/level; where the former defines the type of device and the latter the energy yield.

  • CC1: Comment on EGU2020-8135, Wolf-Gerrit Fruh, 06 May 2020

    Before 'disentangling' energy yield from resource, I suppose that this classification is now a valuable resource for WEC designers to design a WEC suitable for a particular class of resource  (ie a design guide help rather than just a selection help from exisiting WECs...

    • AC3: Reply to CC1, Ian Masters, 06 May 2020

      Yes, that is partly the point of fig1, the test sites and the designs so far are aiming for a particular type of resource, this analysis suggests other "clusters" of parameters that may form a specification for a device type. (the "new" device, might be an existing concept, just at a different set of dimensions/power output)

    • AC4: Reply to CC1, Iain Fairley, 06 May 2020

      Yes. As Ian said, that was part fo the motivation for the work. If you look at the Applied Energy paper (), Figure 12 shows characteristic Hs-Tp joint occurence matrices for four of the classes for the global classification. The thought was that these could be test cases for WEC design.

      • CC2: Reply to AC4, Wolf-Gerrit Fruh, 06 May 2020

        Thank you!  I will have a look at the full paper.  Thanks for the link