Client: Ocean Winds
Location: Moray Firth, Scotland
The Caledonia OWF is located along the eastern edge of the Moray Firth, within the United Kingdom Continental Shelf (UKCS), approximately 40 km south-east of Wick, Scotland. The proposed wind farm covers an area of approximately 429 km², with its western boundary adjacent to the existing Moray East OWF (Figure 1). The area surveyed by Kraken Robotics was the fixed foundation site, excluding the floating foundation site further to the south.
Previous geophysical surveys had confirmed that a significant number of boulders were expected across the site. Earlier seabed features interpretation, using side scan sonar (SSS) and other geophysical datasets, interpreted boulder fields as polygon areas, with objects 1.0 m or larger within those boulder fields identified as individual boulders. The following seabed sediment types were expected within the Caledonia OWF area, as presented in Figure 2.
The aim of the trial was to compare previously acquired SSS data with higher-resolution SAS data, assessing the difference in data resolution and the level of detail available for seabed feature interpretation.
The SAS data was acquired at a resolution of 3 cm x 3 cm, using a range of 150 m on both the port and starboard channels. The interpretation focused on seabed objects measuring 0.3 m or greater in any one dimension, while larger seabed features such as bedforms, wrecks, oil and gas wells, and changes in seabed sediment type were also included.
Due to the resolution of the SAS data, numerous boulders and items of debris were resolved across the survey area. It was determined that interpreting boulder fields as polygons would not make full use of the dataset. Instead, Kraken Robotics interpreted individual boulders measuring 0.3 m or greater in any one dimension, providing a more detailed understanding of seabed conditions and additional information to support the ground model.
Given the significant number of boulders expected, manual interpretation by multiple geophysicists was not considered practical within the required timeframe. Kraken Robotics therefore used an AI-assisted interpretation workflow to support the identification of seabed objects across the dataset. Hidrocibalae were contracted to provide their AI solution for the interpretation of objects measuring 0.3 m or greater.
Figure 2 – Highlighting the three predominant seabed sediments within the areas surveyed with SAS in the Caledonia OWF. The Glaciomarine sediments contain a significant number of boulders and cobbles. All seabed sediment description were provided by client supplied information.
Figure 3 - SAS data example of a known wreck surrounded by items of debris and possible boulders.
Figure 4 - Sample of the boulder density map with a 10 m grid.
Across the 19 km² SAS survey area, approximately 78,000 boulders were interpreted. The 3 cm x 3 cm SAS imagery enabled individual boulders measuring 0.3 m or greater to be identified, providing a more detailed seabed features dataset than previous polygon-based boulder field mapping.
The SAS imagery clearly resolved boulders, cobbles, debris, wreck-related features, and changes in seabed sediment type. This provided the end user with a clearer understanding of seabed conditions that may affect WTG locations, offshore substation platforms, and inter-array cable routes.
Kraken Robotics delivered GIS-compatible interpretation files and a boulder density map to support faster review of potential installation constraints. AI-assisted interpretation reduced the requirement for extensive manual boulder picking while still delivering detailed object-level information.
