A closer look at the Selkie GIS T-E Tool

A closer look at the Selkie GIS T-E Tool

If you are a newcomer to the SELKIE project you might have gleaned from our website that the project has been working towards delivering a series of open-source tools to catalyse the development to commercialisation of marine energy devices, with a focus towards tidal and wave energy converters.

Our Geographic Information System Tecno-Economic calculator (GIS T-E) offers a wealth of vital information to any potential developer looking for a suitable site to deploy a marine energy device, providing up to date projections, metrics and calculations with the click of a button. 

Below are some screen captures of some of the layers the tool provides with a description to provide further interpretation.

Annual Wave Energy Production

Fig 1. Annual Wave Energy Prodction (MWh)

Wave Data

A 20 year hindcast of significant wave height and wave period data for the study area was required. Such data was freely available via the Copernicus Marine Service (CMS) website. There were two similar products within this website which met the Selkie wave data requirements. Both were used for comparison of accuracy.

The ‘NWS Ocean Waves Reanalysis Product’ gives a hindcast for the European Northwest Shelf area beginning from 1980 – 01 – 01 to Present. The underlying wave model is WaveWatch III, which covers the extents of 16°W – 13°W; 62.75°N – 46°N with a spatial resolution of 0.017° x 0.017°, or ~ 1.5 to 3 km. The ‘Atlantic -Iberian Biscay Irish – Ocean Wave Reanalysis Product’ is based on the MFWAM model developed by Meteo-France (MF). It is fed by the ERA 5 reanalysis wind data from ECMWF, covers the extents 19°W – 5°W; 56°N – 26°N and has a spatial resolution of 0.05° x 0.05°, or ~ 3 to 5 km.

The variables downloaded were ‘Spectral significant wave height (Hm0)’, ‘Spectral moments (-1,0) wave period (Tm-10)’ and ‘Wave period at spectral peak / peak period (Tp)’. The time series applied was 2000-01-01 00:00:00 to 2019-12-30 23:00:00 for downloading each. As the website has a download limit of 1024MB, the data was downloaded as 90 separate geographically subset NetCDF files, each covering a different segment of the study area.

The AEP layers were precalculated in MATLAB by applying the power curve or power matrix to the historical climatic data (20 years for wind and wave and one year for tidal currents). For power matrices, a script containing a bespoke loop to find all pairs of significant wave height (m) and wave period (s) corresponding to a given sea state, and then converting that sea state to the corresponding power output (MWh) based on the power matrix of the device. In both scripts, the ‘meshgrid’ MATLAB function was used to repeat these conversions for each and every geographic grid square, the size of which varied depending on the spatial resolution of the netCDF climatic data. The ‘pcolor’ MATLAB function was used to create pseudocolor plot of the geographical distribution of the resulting power output (MWh) values and the ‘geotiffwrite’ function converted these plots to GIS compatible rasters which could be further manipulated and symbolised in ArcGIS Pro before sharing to the cloud for accessing via the web tool. The AEP layer shown in the map is for a Generic WEC Power Matrix found in Robertson et al. (2016).


Fig 2. (above) Bathymetry of SELKIE study area (m)
Bathymetry zoom giving depth contours (m) off SW Irish coast

Bathymetry data

Covering the European sea regions (36°W – 43°E; 90°N – 15°N), the harmonised EMODnet Digital Terrain Model (DTM) was downloaded in raster format to the geographic extent of the Selkie study area. The 2016 DTM release was the latest available version at the time of download which has a spatial resolution of approx. 115m2. The product has been generated from selected bathymetric survey data sets, composite DTMs and satellite derived bathymetry products, whilst any data gaps are filled using integration of GEBCO Digital Bathymetry. It is freely available to view and download via EMODnet’s Viewing and Downloading service.  The DTM raster data then was processed in ArcGIS Pro in order to make it fit to share as a web-layer for integration to the GIS-TE tool. This involved masking out areas >300m deep (beyond the edge of the continental shelf at which point MRE infrastructure is unlikely to be deployed), converting to polygon, and symbolising appropriately for user legibility. A depth contour layer was also developed using the ArcGIS Pro ‘Contour’ geoprocessing tool (selecting a 10m contour interval) to make it clearly visible what the depth range is in a given area of interest. The data from these layers feeds the Site Selection Aid and appears in the Techno-Economic Calculator under ‘Site Information’.

Marine Traffic and Fishing Density

Fig 4. (above) AIS Shipping Density (hrs/km²/month)
Fig 5. AIS Fishing Density (hrs/km²/month)

AIS Data

The EMODnet Human Activities web-page facilitates access to geographic marine data on human activities performed in EU waters, one theme of which is vessel density. The vessel density data is based on the AIS (Automatic Identification System) which is used by ships to regularly report their geographic position. The data is open access and can be viewed on an interactive map of Europe. The spatial resolution of the data is 1km by 1km, with each grid square expressing the vessel density in hours per km2 per month. It can be separated by vessel type and can be downloaded in the form of GeoToiff rasters. For the Selkie GIS-TE tool, a three year hindcast of this data was downloaded for cargo and tanker vessels together, to represent general shipping, and then for fishing vessels separately, due to the known importance of considering the fishing industry when siting ORE infrastructure. At the time of download, the yearly average values were available for 2017, 2018 and 2019 and were downloaded separately as individual geotiff files. These were then uploaded to ArcGIS Pro and run through a purpose-built model for the necessary processing steps required to meaningfully display the data, both for all traffic (Fig. a) and for fishing traffic exclusively (Fig. b). This involved:

  1. Creating an individual raster from the three separate rasters (2017, 2018 and 2019) depicting the average values over the three-year time series (‘Raster Calculator’ tool)
  2. Converting the data to integer values remove meaningless decimal values (‘Int’ tool)
  3. Clipping the data to the extent on the Selkie study area (‘Clip Raster’ tool)
  4. Converting the raster data to polygon in order to make it web compatible (‘Raster to Polygon’ tool)
  5. Removing land areas from any parts of the dataset which erroneously straddled over coastal boundaries of countries within the study area (‘Erase’ tool)
Figure a. (above) The model developed in ArcGIS Pro to process the AIS data for all ships.
Figure b. The model developed in ArcGIS Pro to process the AIS data for fishing vessels.

The resulting layers were then symbolised in ArcGIS Pro to show the top 1% and 1.5% busiest areas (in the Irish and UK EEZ) for shipping and the top 10% and 20% busiest areas (in the Irish and UK EEZ) in terms of fishing vessel activity. These thresholds were chosen as per an interview study of offshore development zones. In that study, it was agreed by the majority of respondents that including the TSS as a hard constraint is more important than including the AIS data as ships need to strictly adhere to them, unlike the AIS data which simply indicates where they have been over a given period. The majority of respondents also agreed that the top 20% busiest areas should be regarded as a constraint to ORE development, but that this should only be guidance as thorough and delicate consultation with the fishing communities is required on a project-by-project basis. The TSS layers were symbolised with a red outline to indicate their extent.


Robertson, B., Hiles, C., Luczko, E. and Buckham, B., 2016. Quantifying wave power and wave energy converter array production potential. International Journal of Marine Energy, 14, pp.143-160.