Marine renewable developments in Scottish waters: review of benthic ecological surveying
This study reviews different intertidal and seabed ecology survey methods, used to identify baselines for environmental assessments.
3 Review of seabed and intertidal biological survey approaches
3.1 Introduction
3.1.1 Scope of benthic survey activities to be considered
Benthic surveying, in the context of MREDs, includes the collection of geophysical/ geotechnical (remote sensing), biological, physical, and chemical information. This is conducted throughout the lifespan of a development, from pre-development (characterisation) to the monitoring stages (pre/post-construction/operation), and decommissioning.
The data gathered and information generated about the types and status of the seabed communities can inform key site and route selection decisions and determine the suitability of a proposed plan of works for licensing purposes. The data can also help a developer report potential effects with regard to licence conditions.
The tools, technologies, and techniques employed for the purpose of surveying benthic habitats and species are diverse and vary from traditional and widely used to novel and emerging. Traditional methods such as grab sampling typically provide a strong evidence-based methodology that conforms to industry standards for application across diverse marine sectors. Over time, traditional methods may be replaced by those that are new or innovative which may currently be limited in operational and industrial application, but could provide wider benefits (i.e. refined approaches, reduced costs, low environmental impact etc.).
There may be additional factors to consider that influence the survey method selection process, for example survey and data continuity, availability and cost of equipment or site-specific conditions. The chosen tools, technologies and techniques must be effective at acquiring the required benthic data to fulfil the survey objectives, however, they should also be appropriate for use in Scottish waters, the sector, development stage, and local environmental conditions.
Four stages are examined within the following sections of the report:
- Section 3 describes the range of sampling platforms, sample gathering tools, technologies and techniques, the forms of analysis applied to the samples and the types of data applications, as well as uses and forms of presentation that may be applied to the data created.
- Section 4 presents a series of key considerations that need to be taken into account when planning an individual survey or a survey campaign over a number of locations or years.
- Section 5 describes the systematic analysis and comparison carried out on these candidate tools, technologies and techniques which is presented in full as an Evaluation Matrix in Appendix 1.
- Section 6 distils these comprehensive findings into a set of more descriptive pros and cons associated with the various tools, technologies and techniques.
Benthic sampling tools, technologies, and analytical techniques are summarised in Section 3.2 and critically analysed and compared within Appendix 1. Geotechnical data acquisition is included within Section 2 for completeness since it can help inform certain questions from a seabed ecology perspective. However, it is not discussed further within this report as its application relates more specifically to engineering aspects of MREDs.
3.1.2 Structure of seabed surveying tools, techniques and technologies adopted for this study
There are five key elements of benthic data acquisition that are applicable to the MRE sector:
- Existing data: Making full use of existing bathymetric, geological, hydrographic and ecological data to best understand the likely prevailing conditions in an area.
- Geophysical/geotechnical remote sensing: The collection of acoustic data and sub-surface sampling of the seabed.
- Biological: The collection of biological information relating to species, habitats, and biotopes.
- Physical: The collection of sediments to determine physical and morphological characteristics.
- Chemical: The collection of sediment and water samples to determine chemical composition.
For these five elements, the types of samples that can be collected and how they can be analysed and/or reported over a project cycle are presented in Table 3.1.
Table 3.1 Summary of main surveying categories and associated attributes linked to gathering existing data, gathering new data and gathering example/ representative specimens and materials
Category: Biological, chemical, physical
Main application: Gathering existing data
Criteria |
Associated attributes |
|---|---|
What type of material gathered |
|
Tools used |
|
How is the material analysed |
|
How the data is reported |
|
Category: Geophysical
Main application: Gathering geophysical/geotechnical data
Criteria |
Associated attributes |
|
|---|---|---|
What type of material gathered |
|
|
Tools used |
|
|
How is the material analysed |
|
|
How the data is reported |
|
|
Category: Physical
Main application: Gathering visual imagery
Criteria |
Associated attributes |
|
|---|---|---|
What type of material gathered |
|
|
Tools used |
|
|
How is the material analysed |
|
|
How the data is reported |
|
|
Category: Biological
Main application: Gathering visual imagery
Criteria |
Associated attributes |
|
|---|---|---|
What type of material gathered |
|
|
Tools used |
|
|
How is the material analysed |
|
|
How the data is reported |
|
|
Category: Biological
Main application: Biological species/specimen sampling
Criteria |
Associated attributes |
|
|---|---|---|
What type of material gathered |
|
|
Tools used |
|
|
How is the material analysed |
|
|
How the data is reported |
|
|
Category: Chemical
Main application: Sediment, rock, water, material sampling
Criteria |
Associated attributes |
|
|---|---|---|
What type of material gathered |
|
|
Tools used |
|
|
How is the material analysed |
|
|
How the data is reported |
|
|
Category: Geophysical
Main application: Geophysical/geotechnical material sampling
Criteria |
Associated attributes |
|
|---|---|---|
What type of material gathered |
|
|
Tools used |
|
|
How is the material analysed |
|
|
How the data is reported |
|
|
3.2 Review of seabed and intertidal sampling tools, technologies, and techniques
This section provides an overview of benthic sampling tools, technologies, and analytical techniques and their applicability for wind, wave, and tidal energy developments (MREDs).
The first step consists of a review of the various platforms from which sample gathering activities are undertaken. This is followed by a review of the sample gathering approaches that can be used from these platforms. The final part of this section considers how the gathered samples might be analysed to create datasets and briefly considers how these data may be used and presented.
3.2.1 Types of sampling platform that can be used
Direct surveyor observation and specimen/sample collection
Cable landfall locations vary in complexity (from sand to rocky shore) and accessibility (easy access to remote). Walkover surveys are a common and inexpensive method of collecting intertidal data across MRED cable landfalls whereby a survey team is deployed on-foot with a Global Positioning System (GPS) to ground-truth the habitats/biotopes and species throughout a survey area. This can include the collection of samples using hand cores and quadrats. Walkovers are suitable for small scale surveys and rocky shores. A minimum of two persons per team is required for health and safety reasons. Data can be hand-written or logged digitally using in-field mapping software. Walkovers can provide wider site detail and precise biotope mapping and identification whilst causing low disturbance to sensitive animals and habitats.
There are safety implications to consider, as well as assessing if the size of the site and coverage can be achievable during a tidal ebb. Further, GPS accuracy and the battery life of electronic equipment needs to be carefully considered in remote and enclosed areas. Intertidal surveys are typically performed as walkover surveys, however, hovercrafts can be deployed for covering larger distances to collect hand cores and quadrats. Hovercrafts enable safe and efficient sampling over difficult and often vast and inaccessible intertidal terrain such as estuaries and mudflats where the tide can flood quickly. In remote locations, local knowledge may need to be sought to identify how best to approach a site (i.e., by land or sea).
Commercial diving teams using Self-Contained Underwater Breathing Apparatus (SCUBA) can be utilised to collect benthic data in shallow (<30 m) waters; however, SCUBA surveys are not a common approach in operational MREDs in the UK. Benthic surveys within wind developments typically occur deeper than recreational diving depths (>30 m) and alternatives for diving deeper incur greater expense. SCUBA is used more frequently in the wave and tidal MRE sectors to conduct photography surveys and epifaunal sample collection in otherwise difficult to sample areas (e.g., EMEC tidal test sites).
Sampling from crewed surface vessels
Crewed vessels are most often utilised to deploy benthic monitoring equipment in the shallow subtidal, nearshore, and offshore. The vessel provides transportation, accommodation and shelter as well as working platform in terms of storage, space and lifting capacity. The vessel is often selected based on the type of survey required, however the size and capacity of the individual vessel can influence the choice of sampling equipment and the methods employed to obtain the required data. Broadly, vessels can be split into two main categories, large offshore vessels (>12 m), and nearshore coastal vessels (<12 m) which include hard boats, Rigid Hull Inflatable Boats (RHIBs) and hovercraft. For MREDs in Scottish waters a combination of vessels may be required. Nearshore vessels must be powerful and nimble enough to withstand and safely navigate the inherent challenging local conditions at wind, wave and tidal development sites and along cable route approaches to shore, to ensure that sampling is effective.
The equipment available on the vessel is also critical. The specifications of winches, cranes, crew skills, deck space, position keeping capacity, accommodation, endurance, speed and motion control may all be important.
Vessel tethered Remotely Operated Vehicles (ROVs) are deployed to collect data during the characterisation, pre-operational, and operational stages of MREDs in UK waters. They are also used during decommissioning of the oil and gas sector which is likely to be cross comparable in methodology.
Sampling from autonomous craft and vehicles
The development and advancements in marine robotics have revolutionised the marine survey industry. Uncrewed vessels, which include all autonomous and semi-autonomous platforms, are now used widely in the global marine sector. As such, several guidelines have been produced for selecting and using uncrewed vessels for the purpose of benthic monitoring (Noble-James et al., 2018). Uncrewed Aerial Vehicles (UAVs) are also used extensively for surveying MREDs.
Uncrewed aerial vehicles
UAVs (more commonly known as drones) are used extensively by the wind farm sector for surveying landfall and potential landfall locations and for identifying sensitive habitats and species. They are not suitable for subtidal surveys in turbid UK waters but have high applicability for Scotland’s shallow subtidal where water depth is shallow with high clarity (Ahmed et al., 2022; Carpenter et al., 2022; Price et al., 2022).
Uncrewed sea surface vessels
The use of Uncrewed Surface Vehicles (USVs) has recently expanded in the wind farm sector. These small, low energy, remotely piloted vessels can conduct a variety of tasks in close proximity to surface and subsea infrastructure without the additional risks associated with larger vessels and tethered equipment.
Autonomous underwater vehicles
Autonomous Underwater Vehicles (AUVs) are most suitable for the characterisation stage of MREDs for surveying over simple seabed topography, or for conducting operational monitoring across export cables. Vertical subsea/subsurface structures that require a complex survey design can present collision risks. This risk is minimised by using a Hybrid AUV (H-AUV). This is an AUV that can be connected via a tether (like an ROV). These adaptations enable greater manoeuvrability than standard ROVs and more stability than AUVs. Local environmental conditions will dictate the suitability of untethered autonomous vessels as strong currents impact stability, with resulting data gaps from equipment moving off course or poor data outputs caused by increased turbulence.
Autonomous seabed vehicles
The surveying needs and capacity of tethered and remote underwater operations are expanding rapidly. Seabed vehicles (both tethered and autonomous) are widely used in marine industry specifically for cable and pipe laying and deep-sea mining. Conducting benthic activities in situ with vehicles that drive across the seabed reduces many of the inherent risks that can impact the collection of data using more traditional approaches (e.g., due to wind, tide, surface/midwater currents). Therefore, the applicability and suitability of seabed vehicles to the MRE sector has become a recent research and development focus, particularly within the tidal sector, where they are being increasingly used to deploy and install moorings at tidal energy sites. It is therefore likely that in the near future seabed vehicles will develop into key tools with wide applicability to the MRE sector, however further research and development is required to understand their effectiveness and suitability for benthic data collection, including direct and indirect impacts to marine habitats and species.
Remote sensing
Satellite and aerial sensing data can be used to help provide information on seabed and intertidal ecology, particularly for intertidal and nearshore habitat characterisation. Visible spectrum photography can show shoreline features and shallow water distribution of rocks and sediments. Radar/lidar type technology can display shoreline topography as a Digital Terrain Model (DTM) creating a DTM dataset. Offshore gravitational pull and sea level detection techniques can be used to map larger scale bathymetric features.
3.3 Methods of data gathering
3.3.1 Making use of existing data
The sources and availability of existing datasets are not a major focus of this present study. However, there are datasets available which can help scope any planned survey effort effectively and successfully. The types of data that can be accessed relevant to seabed ecology include:
- Historical species records,
- Navigational chart data – depth, seabed features, sediment type,
- Nationally held bathymetry – swath bathymetry at varying resolutions (1 m to 50 m),
- Geological information and geophysical data – sediment classification and distribution, rock types, bedforms, rocky reef structures (e.g., moraines),
- Previously gathered seabed photos and videos,
- Hydrographic information (currents, wave action, mixing zones, water chemistry),
- Fisheries data (e.g., spawning grounds, feeding areas),
- Species and habitat records and extent information including features of conservation interest,
- Data supporting previous project development in an area (baseline and monitoring surveys), and
- Previous research (widespread subject matter and sources).
3.3.2 Bathymetry and geophysical data gathering
Geophysical surveys (also referred to as acoustic approaches) can include a combination of: Multi Beam Echosounder (MBES), Single Beam Echosounder (SBES), Multi Beam Backscatter (MBBS), Side-Scan Sonar (SSS), Sub Bottom Profiling (SBP) and magnetometer. There is a general requirement that geophysical surveys for the purposes of benthic characterisation and temporal monitoring are designed to meet the International Hydrographic Organisation (IHO) ‘Order 1a’ standard as per IHO Special Publication S44 (Ed 6.1.0) (IHO, 2022). Key standards and guidance notes for the acquisition and processing of geophysical data for this purpose include Plets, Dix & Bates (2013) and Judd (2012).
The sensors can be hull mounted and vessel towed as well as mounted onto autonomous or semi-autonomous systems. Acoustic equipment can be deployed from a variety of vessel types, nearshore and offshore. In shallow waters (<10 m) USVs mounted with acoustic sensors are generally deemed to be suitable.
Geophysical approaches are widely used during all stages of a development and across all marine sectors to characterise the seabed, detect sub-surface features including man-made objects and obstructions such as wrecks, ordinance, and debris and geomagnetic anomalies within and around an area of influence. Geophysical data is also commonly used to delineate distinct boundaries between differing sediments and seabed features to provide a broad description of the seabed and assign benthic habitat types, it is usually applied in conjunction with traditional ground-truthing methods (e.g., Drop-Down Camera (DDC) and grab sampling).
Acoustic sonar systems coupled with ground-truthing are capable of detecting and distinguishing between fine scale changes and reflective signatures associated with ecological seabed features, that include biogenic reef (Modiolus modiolus, Mytilus edulis, and Sabellaria spinulosa) (e.g., see Sanderson et al., 2014 for M. modiolus), seagrass beds (Greene et al., 2018; Gumusay et al., 2019; MMT, 2021) and maerl beds (De Esteban et al., 2018; Noble-James et al., 2023).
Modes of acquisition have differing scales of influence on the marine environment depending on the sound source, exposure time, frequency, and depth of use. Therefore, the use of geophysical equipment may require licensing for disturbance of European Protected Species (EPS), particularly cetaceans, and basking sharks.
3.3.3 Seabed imagery
Seabed imagery is extensively employed in marine industries, including all MRED sectors. It is collected for multiple purposes, including ground-truthing and delineating habitat features in geophysical surveys, recording species and habitats (particularly those of conservation importance), monitoring collision risks (wave and tidal devices) and investigating potential subsea hazards prior to grab sampling.
Optical imaging systems (including freshwater chambers for use in turbid conditions) vary in style and design, therefore the aims of the survey and the environmental conditions that dominate the site should dictate the type of system used. Operational guidelines developed for remote monitoring of biota (Hitchin, Turner & Verling, 2015) and interpreting seabed imagery (Turner et al., 2016) are widely used and cited by industry but are now largely superseded by the work of the UK’s Benthic Imagery Action Plan, (NMBAQC, 2023a) or Big Picture Imagery Analysis Working Group for the NE Atlantic Biological Analytical Quality Control (NMBAQC), and resulting quality assurance guidance (NMBAQC, 2023b).
The benefits of seabed imagery systems are that they are relatively inexpensive compared to other methods and can be adapted to a range of conditions. However, systems can be limited by operational depth, on-site subsea conditions (turbidity and currents) and weather.
ROV or vertical drop system video surveillance surveys are often one of the first on-site checks to be made on a prospective site or cable route in the wave and tidal sectors. This may be followed up by a stills camera survey or more systematic video survey once the habitats present have been confirmed.
Ocean underwater imaging sensors and laser profiling
Recent advances in ultra-high resolution digital imaging systems such as ‘CathX’ Ocean underwater imaging sensors and laser profiling are increasingly being used within MREDs. For example, in the wind sector, they can be attached to semi-autonomous equipment and used for conducting array and export cable route inspection surveys. They can be used to measure biological and physical impacts such as biofouling, introduction of Invasive Non-Native Species (INNS), habitat loss (e.g., Annex I/Priority Marine Feature (PMF)), scour, sedimentation, and structural integrity. The benefits are the collection of continuous ultra-high resolution imagery that can be used to produce photogrammetric models. Limitations include the overall expense of hiring equipment coupled with the associated platform requirements (Working Class ROV) and costs of hiring a large vessel. Further, large volumes of data are produced that require regular storage and processing. Recent advancements in affordable high-speed satellite internet connectivity have improved data transfer issues. However, there are additional costs associated with the storage and access of big data that may need consideration.
Baited Remote Underwater Video (BRUV)
Baited Remote Underwater Video (BRUVs) are video cameras deployed and left in situ for a determined period of time (soak time). The cameras are baited to attract species towards the camera’s field of view. This method promotes a holistic view of species assemblages in situ, particularly fish (including commercial and priority species) and other mobile epibiota that are less likely to be recorded using traditional methods due to their behavioural reaction to disturbance.
BRUVs are commonly used as a benthic survey tool in Scotland by Non-Governmental Organisations (NGOs) involved in habitat restoration and species recovery. BRUVs are not widely deployed within MREDs, however the development and application of BRUVs and BRUV technology in MREDs has been driven forward in recent years through collaboration with researchers at Swansea University, specifically for the benthic characterisation and operational stages of wind developments in the UK (Griffin et al., 2016; Jones et al., 2019). Exclusion zones created by MREDs have the potential to create refuge areas for mobile species which could over time promote a “spillover” effect of species into commercial fishing grounds (Halouani et al., 2020). Deploying BRUVs could therefore enable the collection of data that promotes quantification of the spillover effect. Consideration is needed when planning and executing BRUV surveys to ensure sampling is representative to the life-history stages, feeding, and migratory patterns of mobile species. Survey parameters such a soak time and sampling time require standardisation to factor in the survey bias associated with bait use. Bait selection and preparation also requires careful consideration.
Timelapse and trigger imagery
Timelapse and trigger photography and/or video can also be used with stimulus sources such as light, bait, or sound or around structures to examine behaviour over time. This approach works for macro- and megafauna such as crustaceans, echinoderms, and bottom living/demersal fish. These techniques show movement and use of an area over time and can also be used to record more occasional/sporadic mobile faunal behaviour. Their advantage is that they are more energy and memory efficient than continuous monitoring and recording approaches.
Sediment Profile Imaging (SPI)
Novel and emerging methods used less commonly in MREDs include acoustic cameras and Sediment Profile Imaging (SPI). Acoustic cameras such as Adaptive Resolution Imaging Sonar (ARIS) and Gemini models are mounted on ROVs for the primary purpose of object detection/collision avoidance as they can detect hard features in turbid conditions when camera visibility is poor. They can be pivoted towards the seabed to capture hard sediment features such as bedrock, boulders, and biogenic reef. They have proven to be particularly useful in turbid conditions which are a favoured environmental condition of S. spinulosa which form reefs (Annex I biogenic reef) throughout the UK, but mainly the east coast of England and Wales, as well as the Moray to Aberdeenshire coast of Scotland (Pearce & Kimber 2020).
Whilst not widely used within MREDs, acoustic cameras offer an alternative sampling method during operational wind farm seabed remediation surveys and the benthic characterisation stage to confirm the presence and absence of biogenic reef (Griffin et al., 2020). The acoustic signature of the S. spinulosa differs from bedrock and stony structures and whilst not tested, this method could potentially be utilised on other biogenic reef habitats such as M. modiolus, Ostrea edulis, or Lithothamnion sp. to identify whether different biogenic structures present differing acoustic signatures.
SPI is a method for rapid assessment of the quality of seafloor habitats. Identified as a more cost and time efficient method than grab sampling, it works by inserting the flat viewport of a camera into the top 20 cm of seafloor sediments and taking images of the upper sediment structure and sediment/water interface to determine the chemical and biological characteristics of the sediment, including the percentage of oxic and anoxic sediments and presence of bioturbation features. SPI cameras produce both cross-sectional and surface imaging, they consist of a wedge-shaped prism with a Plexiglas faceplate and a back mirror mounted at a 45-degree angle. An internal strobe is used to provide light. The mirror reflects the image of the sediment profile to a digital camera mounted horizontally on top of the prism (NewFields, 2023).
SPI systems can be deployed down to 4000 m and have been widely used for decades to assess the anthropogenic impacts of aquaculture, trawling, sewage effluent and oil spills. They require specialised image analysis and interpretation software in order to measure multiple physical and biological parameters (sediment type/colour, prism penetration depth, grain size, surface boundary roughness, mud clasts, redox potential discontinuity depth, presence of oxygen/methane, presence of benthic organisms, faecal pellets, burrows, surface tubes, feeding mound/voids, infaunal succession stage, organism-sediment index, benthic habitat quality index and INNS).
The benefits of SPI systems include the ability to convey the ecological outputs in an easily understandable format to the end user. Limitations include a smaller field of view compared to other imaging techniques, and images of coarser heterogeneous sediments are less reliable and require more replicates than finer sediments (Smith et al., 2003; Blanpain et al., 2009). Additionally, ‘smearing’ (displacement of oxygen rich sediments to deeper layers) during camera penetration into sediments can result in sediments appearing healthier than they are (Moser et al., 2021). SPI has been trialled in cable routes and wind turbine sites assessments for offshore energy projects in North America to reduce the risks and costs for offshore wind projects. SPI imagery was integrated with plan view imagery and MBES acoustic data (‘forward scouting’) to optimise route selection, provide ground-truthing, and characterise benthic habitats and demonstrated a collaborative approach to cable routing and site characterisation (Carey, Doolittle & Smith, 2019).
3.3.4 Gathering physical seabed samples
Benthic sediments are sampled using a variety of tools to provide data on the physical, biological and chemical condition of the seabed. The type of sampling device selected will depend on the method of deployment (mechanical or by hand), water depth, and the known or predicted composition of the sediment. Sediment samplers come in a variety of sizes depending on how/where they will be deployed and typically collect the top 10 - 20 cm of sediment.
Sediment sampling is widely adopted within wind farm developments for the benthic characterisation and operational monitoring phases and currently adopted less across the wave and tidal industries. The most common and widely used subtidal sediment samplers applied to MREDs are Day Grab, Hamon Grab, Mini-Hamon Grab, Van Veen, Dual Van Veen, and Shipek Grab (Coggan & Birchenough, 2007; Judd, 2012; Bender et al., 2017; Hemery, Mackereth & Tugade, 2022). The main benefit of sediment sampling is the ability to conduct repeat quantitative sampling with high positional accuracy to obtain samples of suitable size for multiple analyses. However, sediment sampling is limited by the on-site subsea conditions (sediment type) and weather (wind/wave action during deployment and recovery).
Sediment cores are often collected during the benthic characterisation stage of wind farm developments to provide an undisturbed sample of sediment. These are commonly collected by mechanical means (e.g., Box Core, Vibrocore), however other methods such as Hand Cores and Push Cores can be collected by SCUBA divers and ROVs (Coggan & Birchenough, 2007; Judd, 2012; Bender, Francisco & Sundberg, 2017; Cochrane, Hemery & Henkel, 2017; Hemery, Mackereth & Tugade, 2022). The limitations of core sampling are the relatively small surface area of sample that can be obtained compared to a grab sampler.
Demersal trawl sampling
Remote sampling of mobile benthic epifauna (mainly fish, shellfish and highly mobile megafaunal species) is traditionally conducted subtidally using benthic trawls to provide quantitative and semi-quantitative estimates of abundance and diversity. There are a diverse variety of trawls used to sample benthic species. The choice of trawl is largely dependent on the nature of the seabed sediments, the environmental considerations of the site, and temporal scale cohesion with local fisheries management.
Recommendations and standards relating to benthic trawl use within MREDs is generally lacking, however, recommendations, considerations, and operating guidelines are available via;
- Mapping European Seabed Habitats (MESH) (Curtis & Coggan, 2006)
- Centre for Environment, Fisheries and Aquaculture Science (Cefas) Regional Seabed Monitoring Programme (RSMP) protocols for sample collection and processing (OneBenthic, 2020)
- Joint Nature Conservation Committee (JNCC) (Noble-James et al., 2018)
Further guidance documents are available in relation to aggregate dredging (Boyd, 2022). The use of benthic trawls poses some inherent impact on the seabed and species (Somerton, 2001; Hiddink et al., 2019; Jac et al., 2022; Noble-James et al., 2023). Additional considerations are required in Scottish waters to minimise bycatch of critically endangered PMFs such as the flapper skate, including its egg cases which are attached to the seabed or encrusting epifauna.
3.4 Methods of data analysis on seabed and intertidal samples
3.4.1 Bathymetry, geophysical and geotechnical analysis
The primary purposes of gathering bathymetry, geophysical and geotechnical data relate to the engineering design process, but this information can also help inform likely seabed habitat type and extent. The key factors that should be discernible from such data include:
- Accurate water depths, seabed slope and roughness/texture
- Defining areas of bedrock, broken rock, cobbles, or areas of sediment
- Defining the depth of sediments
- Identifying any shallow natural gas deposits or seepage/pockmark activity
- Defining the limits of mineralised biogenic reefs
- Defining dynamics of mobile sediment features
Making good use of the resources put into geophysical/geotechnical activities is therefore very useful and important for seabed ecological investigations.
3.4.2 Seabed imagery analysis
Stills and video imagery
Seabed imagery is typically analysed using standardised approaches and recommended guidelines (Jones et al., 2016; Golding et al., 2019; 2021) that align with best practice and Quality Control (QC) procedures (NMBAQC, 2016) and are endorsed by governance frameworks. This includes the use of image annotation techniques which utilise cloud-based annotation platforms such as BioImage, Graphical Labelling and Exploration (BIIGLE) (Langenkämper et al., 2017). Imagery can also be analysed for the purpose of assessing wildlife interactions with equipment operating on the seabed (i.e., subsea turbines) (Hutchison, Secor & Gill, 2020).
BRUVs
BRUV imagery undergoes quantitative analysis using dedicated open-source software such as SeaGIS EventMeasure (SeaGIS, 2023) to conduct fast, efficient analysis of movie sequences. The user can train the software to dynamically analyse recorded events and report on abundance measures such as Mean Relative Abundance (MaxN), cumulative MaxN and mean count (by species and stage). Whilst initial software training can take time, overall, it is a cost-effective method of analysing hours of video footage that requires a non-technically skilled user.
3.4.3 Biological seabed sample analysis
Directly gathered specimens
Samples of seaweed, shoreline and seabed animals including mobile, epifauna and even larger infauna can be gathered by hand, sometimes assisted by hand tools. Such samples can sometimes be inspected, photo recorded, notes and measurements taken and the samples then returned to the environment. In other cases, the specimens may need to be anaesthetised and, if appropriate, preserved in fixative. Then at a suitable time and location they may be photo recorded, measured and then restored or discarded. Where certain chemical analyses or DNA analyses are to be carried out, preservation through freezing or other defined means may be necessary.
Grab and box core sampling for biology
The treatment of a seabed sediment sample for biological analysis is relatively universal but has a number of subsequent analysis pathways that are commonly used. Firstly, the obtained sample may be processed whole or may be sub-sampled using a hand core placed into the intact recovered seabed sample. The resultant sample is then placed onto a suitable mesh size of sieve and seawater used to wash away the finer sediment leaving macrofauna and any stones. The latter may be discarded or retained separately to avoid abrasion of the biological specimens. The biological material will generally be anaesthetised/relaxed and then fixed in preservative for later sorting, identification and workup.
The suite of analytical techniques conducted on benthic grab samples depends on the purpose of sampling and the requirements (for example as part of a development licence). Traditional analytical methods such as infaunal and microbenthic identification and analysis (e.g., abundance and biomass) are widely used by marine industry to provide baseline data on the sediment and species composition. Samples are typically analysed by laboratories that participate in the NE Atlantic Marine Biological Analytical Quality Control (NMBAQC) (note this is not an accreditation) scheme , whereby the methods follow high quality control (QC) standards (Mason, 2022; Worsfold, 2023). The MRE sector has generally adopted these traditional analytical techniques, particularly the wind sector. In the UK the wind sector currently conducts sampling and analysis at the benthic characterisation stage to inform monitoring approaches adopted throughout the operational lifespan of a development and to inform assessment of development impacts, and appraisal of alternative and mitigation options.
eDNA analysis
eDNA sampling is a relatively new method of obtaining presence data on assemblages of species. Analysis of eDNA samples provides presence data on species that may be under-reported if traditional methods are used (e.g., mobile/cryptic/rare/invasive species) or at locations that are logistically difficult to sample (subsea/subsurface marine infrastructure). Samples can be collected from different media (predominantly water/sediment), to record the presence of pelagic, near benthic, and benthic species. For example, Matejusova et al. (2021) demonstrated the use of eDNA analytical techniques from water samples as a cost-effective way to detect the presence of the invasive sea squirt, Didemnum vexillum.
Methods of sampling eDNA to inform on the community composition on marine infrastructure have been trialled and compared by Alexander et al. (2023). Trials included ROV scrapes, water samples, a keel crab (underwater drone) scrapes, plankton tows and PUF tows (cylindrical polyurethane foam attached to a funnel). These methods may be of benefit to MRED benthic monitoring, particularly for describing community composition over time.
Other novel methods of eDNA sample collection have recently been researched and trialled using SCUBA divers and platforms of opportunity (BRUV frames) (Cai et al., 2022; Harper et al., 2023; Neave et al., 2023; Neave, Mariani & Meek, 2023). Water eDNA samples often require extensive sampling and filtration which may vary depending on water quality (turbidity/salinity/organic compounds). Sediment eDNA sampling is relatively cost and time efficient as samples can be collected during the standard suite obtained from grab samples and frozen until required. However, it is logistically more difficult to conduct repeat sediment sampling post-construction, i.e. in close proximity to subsea/ subsurface MRE infrastructure.
The persistence of eDNA molecules in the marine environment is relatively short (Collins et al., 2018), degrading rapidly (48 hours) due to many combined abiotic and biotic factors, therefore producing a snapshot of the species assemblages within the sediment/water column at a given point in time. Further, studies have shown that eDNA degrades faster inshore than offshore (Collins et al., 2018). Multi-collection methods may also be required to accurately represent diversity across benthic and epibenthic surfaces, and depths (Koziol et al., 2018; Antich et al., 2020; West et al., 2022; Alexander et al., 2023), increasing survey and analytical costs. eDNA results are heavily influenced by sampling method, substrate and assay selection (Wegleitner et al., 2015; Koziol et al., 2018; Sakata et al., 2020). Therefore, to be effective, a survey design should be driven by knowledge of the target assemblage, habitat and depth (Alexander et al., 2023). Novel field collection methods have only been trialled under a narrow range of environmental conditions, therefore further research is required into the costs and benefits for different eDNA sampling methods in Scottish waters and where used, repeat sampling should ensure consistent collection method(s) are applied when comparing sites or different time points.
Despite its limitations, eDNA is being widely adopted throughout the marine sector. The aquaculture industry has led the development of eDNA research and industry guidance (Jones et al., 2020; NatureMetrics, 2022; Wort et al., 2022; SEPA, 2023). Research conducted by Elliott et al. (2023) identified the usefulness of eDNA within the MRE sector, having completed a comparison study recently between eDNA and traditional techniques for monitoring around wind farms (Elliott et al., 2023). The research focussed on fish ecology; however initial invertebrate results have proved encouraging (Elliott et al., 2023).
Used alone or in conjunction with visual surveys, eDNA can reduce the need for taxonomic expertise and the costs/logistical limitations associated with traditional methods. At inshore MREDs, eDNA water sampling could be co-located with Water Framework Directive (2000/60/EC) sampling stations (if collected) but will likely incur additional sampling requirements and associated time/monetary costs for sampling and analysis by a dedicated laboratory.
3.4.4 Geophysical and chemical sediment sample analysis
Common geophysical parameters derived from benthic samples include sediment composition via Particle Size Distribution (PSD) analysis. Chemical analysis of samples includes contaminant analysis (heavy metal and hydrocarbon concentrations), Total Organic Carbon (TOC), Redox, other water chemistry measures (pH, temperature), and mineral composition.
It is critical that the handling of samples and sub-samples required for multiple purposes does not undermine the integrity of the planned analyses, particularly where the methods employed for one analysis pathway may be contrary to that required for another pathway.
A detailed description of environmental sedimentology and chemistry procedures is beyond the scope of this present study. Suffice to say that it needs to be clear from the outset of survey planning what is envisaged for the end point analyses and appropriate resources and sample capacity need to be made available to:
- ensure an appropriate chain of custody for each sample type
- ensure sufficient representative samples are obtained
- ensure that sufficient replicates are collected for initial analysis and any follow-up verification
- ensure suitable long-term storage of reference samples if needed
- ensure consistent and compatible analyses through laboratory continuity where critical
3.5 Data interpretation, use and presentation of results
3.5.1 Habitat/biotope classification
The combined results of PSD and macrobenthic analyses plus seabed imagery and acoustic data are collectively used to produce standardised hierarchical seabed classifications (i.e. Broadscale Habitats (BSH), or biotopes). In the UK two main habitat classification systems are used – European Nature Information System (EUNIS) and the Marine Habitat Classification for Britain and Ireland (MHC). EUNIS is a Europe-wide classification system whilst the MHC is a UK-wide classification system. The MHC is cross-transferable to the EUNIS system. Guidance on assigning benthic biotopes using both habitat classification systems has been produced by JNCC (Parry, 2019).
3.5.2 Habitat mapping
Habitat classifications assigned from ground-truthed samples are used alongside geophysical acoustic data to delineate boundaries between habitat types to produce digitised maps and spatial datasets of the seabed. The same method is used to identify potential reef areas as it is evident that different geogenic (stony/bedrock) and biogenic reef structures produce differing acoustic signatures (Brown et al., 2011; McGonigle & Collier, 2014; Jenkins et al., 2018; Griffin et al., 2020). Habitat mapping utilises Geographical Information Systems (GIS) software from developers such as ESRI which ranges from free open-source software (e.g. Quantum Geographical Information Systems (QGIS)/Grass) to expensive programs that require licences to use (e.g. ArcGIS/ArcPro). There is freely available guidance online for utilising these software packages. This method can be subjective and influenced by the experience of a habitat mapping specialist and therefore should be contextualised with the relative level of confidence in the data and outputs. Habitat mapping is a widely used and accepted method of producing digitised maps and sharing spatial data.
3.5.3 Predictive habitat modelling
Predictive habitat modelling is a broadscale statistical approach used to predict the likelihood that an area of unknown habitat is similar/dissimilar to an area of known habitat. It utilises data including acoustic (geophysical) sources, physical, and environmental variables coupled with ground-truthed classifications. There are numerous methods available, and no unified approach is adopted. Popular predictive mapping techniques utilise either a supervised (classes are assigned – Maximum Likelihood Classification, Random Forest) or unsupervised (classes are unassigned – Iso Cluster analysis) approach. A pixel or object-based analysis is then performed on the data layers to assign classes. The EU SeaMap 2021 (Vasquez et al., 2021) is an example of a Europe-wide predictive map used widely by marine industry during survey planning/design and monitoring in order to identify the range of habitats likely to occur within the given area. Local-scale predictive maps may be of benefit throughout MREDs to refine broad-scale predictive maps if sufficient data and coverage of ground-truthed habitat is available. The methods of Boswarva et al. (2018) have been used to conduct predictive mapping outputs for MREDs in England, further guidance is also available and the method in general is widely researched and applied (Miller et al., 2017).
3.5.4 Predictive biotope modelling
The predictive power of biotope scale modelling is inherently lower than BSH modelling as the biological information that informs a biotope is not represented visually within acoustic (geophysical) data utilised to produce predictive maps. Species Distribution Modelling (SDM) and Habitat Suitability Modelling (HSM) (also called Ecological Niche Modelling (ENM)) provide alternative methods of capturing species level information. SDM is a similar concept to predictive habitat modelling, whereby specialised statistical software is used (e.g., R/MatLab/ArcMap/ArcPro) to produce a predicted output of a species current distribution utilising occurrence data. Software packages range from open source and free to requiring expensive licences and expertise. HSM/ENM combines observed species data with environmental factors to predict the potential distribution of a species given its suitable habitat within a realised niche (Brown & Griscom, 2022). This is not a widely adopted method in the MRED sectors, however, JNCC have produced a framework that utilises an ensemble modelling approach to predict suitable habitat for Zostera marina beds, M. modiolus beds and S. spinulosa reefs in the UK (Castle et al., 2022).
3.5.5 Novel analytical techniques
Throughout the UK marine sector, novel approaches for recording species assemblages have recently developed from being largely research-based to being increasingly used for commercial application. In MREDs data from eDNA and BRUVs is increasingly being collected and analysed as an alternative to benthic trawl surveys, particularly within the wind sector.
All of the data gathered and analysed needs to be processed, assessed and interpreted. There are a number of tools that can help with this and increasingly these are likely to draw upon AI approaches to make them more effective.
Artificial intelligence and machine learning
Artificial Intelligence (AI) is the intelligence of machines that perform extensive computational tasks and machine learning is the statistical process in which the machine is trained to automate a task, removing the need for human supervision. However, the translation of data into programming scripts and algorithms remains challenging (Blowers et al., 2021). In the marine sector AI and machine learning are used as novel methods of analysing expansive datasets, particularly seabed imagery and video to automatically detect marine life. There are numerous algorithms created that target object and feature detection, the prominent being Neural Networks, which loosely simulate the human brain (Blowers et al., 2021).
An extensive review of the current methods and applications of use was conducted by the Marine Directorate (MD) (Blowers, Evans & McNally, 2020). The leading platforms include Google Object Detection API, Coral Net, BIIGLE, FishTick and Video and Image Analytics for Marine Environments (VIAME). VIAME is a Convolutional Neural Networks (CNN) detector used to produce object detectors, full frame classifiers, image mosaics, rapid model generation, search image and video, and perform stereo measurements. In Scotland it has been applied as a novel method of conducting shellfish stock assessments (Ovchinnikova et al., 2021).
The major limitation of machine learning is the challenge of training a programme to detect and assign the required features automatically. Studies are limited but have shown that the accuracy of detection is dependent on the feature in question with some features (such as burrows) requiring greater processing and annotation steps than others (such as fish or sea pens) (Blowers et al., 2020). Many of the existing platforms are in varying stages of development and require training and expertise to execute.
Substantial training datasets are required which take time to develop, however once developed could act as a resource to facilitate future development. Portals used for standard image annotation such as BIIGLE and VIAME could act as suitable platforms to develop a standardised training dataset, facilitated by imagery data obtained from strategic sampling regimes.
3D Photogrammetric modelling
Three-Dimensional (3D) photogrammetric modelling is a relatively new analytical tool. Its primary use originally was for monitoring damage to tropical coral reefs caused by climate-induced bleaching. In recent years the technique has been utilised in temperate seas and marine industry primarily for monitoring the structural integrity of subsea infrastructure, including the biomass of marine growth in the operational (e.g., cables) and decommissioning stages (e.g., for oil and gas platforms). The process involves the collection of high-quality digital stills of which there is sufficient overlap between imagery (>60 %) to create a mosaic, and the presence of scaling tools. Specialist processing software is required, most commonly Agisoft which ranges from free to expensive depending on requirements and use. 3D photogrammetry can therefore be a relatively low-cost method of visualising habitats and subsurface structures that has global applications. For example, citizen science groups and NGOs can conduct 3D photogrammetry surveys in shallow waters relatively inexpensively using a GoPro and lighting system attached to a hand-held frame. At the opposite end of the scale, 3D photogrammetry can be conducted in deeper and more challenging waters by utilising ROVs mounted with skids that contain multiple cameras and lighting systems.
3D photogrammetry techniques are widely applicable within the MRE sector, methods are under development for modelling the extent and quality of biogenic reef structures on the seabed and subsea infrastructure. Recent advances in geo-referenced photogrammetry aim to incorporate 3D photogrammetry into a bespoke geospatial data portal to improve visualisation and promote effective decision making by government agencies and stakeholder groups.
3.5.6 Overall assessment of analysis options
The analytical techniques discussed here range from traditional and widely used to novel and infrequently applied. The techniques also vary in their suitability and are likely to be required at different stages of a development’s lifespan rather than selected by sector, spatial scale of a development, or site complexity. The likelihood that a particular analytical technique is applied to a given development stage is summarised in
Table 3.2.
Likelihood is based on the technique’s current use and applicability to each major stage within the MRE sector and in the case of novel methods the likelihood of it being applied in the future;
Likely = The technique is widely applied and applicable to the specified development stage/a novel technique that will likely be applied.
Possible = The technique is identified as applicable to MRE and the specified development stage but it may not be a requirement.
Unlikely = The technique is identified as applicable to the sector but it is unlikely to be used at the specified development stage.
Analytical Techniques |
Benthic Characterisation |
Operational |
Decommissioning |
|---|---|---|---|
PSD |
Likely |
Likely |
Likely |
Macrobenthic |
Likely |
Likely |
Likely |
Physiochemical |
Likely |
Possible |
Likely |
Seabed Imagery |
Likely |
Likely |
Likely |
eDNA |
Likely |
Possible |
Possible |
BRUVs |
Possible |
Possible |
Possible |
3D Photogrammetry |
Unlikely |
Possible |
Possible |
Laser Profiling |
Unlikely |
Possible |
Possible |
BSH Classification |
Likely |
Possible |
Possible |
Biotope Classification |
Likely |
Possible |
Possible |
Conservation Features Mapping |
Likely |
Likely |
Likely |
Habitat/Biotope Mapping |
Likely |
Unlikely |
Likely |
Predictive Habitat Mapping |
Possible |
Unlikely |
Unlikely |
Species Distribution Modelling |
Possible |
Unlikely |
Unlikely |
Contact
Email: ScotMER@gov.scot
There is a problem
Thanks for your feedback