CEFAS25-139 Request for Quotation for the analysis of seabed imagery

Tender description

CEFAS25-139 Request for Quotation for the analysis of seabed imagery (still images and video) collected at Beachy Head East (2023) and Cape Bank (2025) MCZs. Cefas and Natural England (NE) have been working collaboratively to deliver monitoring surveys within Marine Protected Areas (MPAs) as part of the ongoing MPA programme. These surveys aim to support Natural England’s monitoring objectives and provide robust evidence to inform conservation measures. Two recent surveys were undertaken at designated Marine Conservation Zones (MCZs): Beachy Head East MCZ in November 2023 and Cape Bank MCZ in September 2025. Both sites are important for understanding subtidal habitats and associated species, and the surveys were designed to gather high-quality data to meet specific ecological objectives. Beachy Head East MCZ is an inshore site located along the East Sussex coast near Eastbourne, within the Eastern Channel region. Covering an area of approximately 195 km², the survey focused on: • Mapping the extent and distribution of subtidal sediment, rock, and biogenic reef habitats. • Collecting quantitative information on species composition across sedimentary and rocky biotopes. • Identifying the presence of S. spinulosa formations and, where possible, providing data to inform Reefiness assessments. • Recording any presence of non-native species. Cape Bank MCZ, is an inshore site situated in the Western Channel and Celtic Sea region to the west of Land’s End, spans roughly 474 km². The survey objectives included: • Acquiring high-quality video and image data to verify the extent and distribution of subtidal moderate energy circalittoral rock features. • Determining the main characterising species within these habitats to enable temporal comparisons with previous datasets. 2. Requirements: The Supplier is required to undertake the analysis of seabed imagery (video and still images) collected at Beachy Head East and Cape Bank MCZs. This analysis will include the production of imagery-specific Geographic Information System (GIS) deliverables. The imagery for analysis is detailed in Table 1, and maps of the stations sampled are shown in Figures 1 and 2. More detailed metadata will be provided to the awarded Supplier following Contract award. All video observations were captured using a drop-down camera (DDC) system. 2.1. General project specifications: The Supplier will: • Undertake the analysis of seabed imagery (video and still images) as outlined in this document and its appendices. • Use BIIGLE to annotate video and still images in accordance with methodology described in Appendix A. Cefas and NE will provide guidance on use of BIIGLE for MPA projects during the start-up meeting, if needed. o Record the majority of information in BIIGLE first and subsequently used to populate the provided proformas. Where explicitly agreed by Cefas/NE, certain data may be entered directly into the proformas. o Develop and maintain substrate and taxonomic image reference collections in BIIGLE. • Produce all specified deliverables (see section 4), including GIS outputs and a comprehensive analysis report. o The GIS outputs will be provided as an ESRI ArcMap GIS geodatabase (“.gdb”) using the sample information provided (see also section 2.2.3). The database will contain one A still image feature-class (labelled “_STILL”) and two video specific feature-classes (labelled “_VIDSOL” and “_VIDTR”). o The comprehensive analysis report will include a summary of the results, the description of the methodology, and some details on the quality Assurance (QA) process and any remediation undertaken (see also section 4.1) • Ensure that still image and video references used in analysis outputs are identical to those used in the naming of the original media, to enable future reconciliation between data and media. If identical naming is not possible, a suitable alternative should be sought with Cefas (e.g. a lookup table, or closely similar naming) . Any deviation from this methodology must be approved in writing by Project Manager. 2.2. Analysis Methodologies The Supplier must perform all analyses in accordance with the specifications set out below and in Appendix A. Any deviation must be approved in writing by the Project Manager prior to implementation. 2.2.1. Video Data (All Transects) Video will be analysed in BIIGLE using the label trees shown in Table 2. A high-level review will be conducted as described in Turner et al (2016). Tier selection for each tree is specified in Table 2 and further clarified in Appendix A. Each video will be analysed and labelled using provided BIIGLE label trees (unless specified otherwise). The following steps outline the tasks required to generate the necessary information: 1. Where possible, Video will be segmented into areas of continuous seabed broadscale habitat (BSH) type greater than or equal to 5 m along transect distance. The segment labels should be used to delineate these segments and labels from other trees should be attached to each segment using the “add label” tool in BIIGLE. Alternative options (e.g. using time stamp, can be considered. 2. The EUNIS 2007/12 classification system will be used, and a new segment should be started if the habitat classification changes. • EUNIS 2007/12 classification system. Initially this label tree should be used to assign BSH types, that is up to level 3 of the hierarchy (e.g. Sublittoral sands and muddy sands). 3. Each segment will be assigned an Image quality using labels from two label trees. Further analysis of video segments will be dependent on the image quality score. For example, if a segment is given a score of zero, no further analysis should be carried out for that segment: • Image quality, a summary of these labels is shown in Table 3. • NMBAQC image quality, a summary of these scores is shown in Table 4 but described in more detail in section 2.1 of Turner et al (2016). 4. Use the Habitat Features Of Conservation Importance (FOCI) label tree to identify any habitat segments containing the FOCI habitats listed in NE & JNCC, 2010 . 5. Identify evidence of anthropogenic impacts on the seabed: • Use the litter label tree to record the presence of litter using the categories listed in Annex 5.1 of the Joint Research Centres Guidance on Monitoring of Marine Litter in European Seas . • Use the Anthropogenic label tree to annotate trawl marks or anthropogenic impacts other than litter. The tree may be extended with new labels if required in agreement with Cefas (extensions must be documented). 6. Use the habitat label tree to assign biotopes, up to level 6 of the EUNIS 2007/12 classification system and in accordance with Parry (2019) . 7. A reference collection of representative images must be provided for each discreet BSH identified. 2.2.2. Analysis of Stills Imagery Still images must be analysed in BIIGLE using the label trees shown in Table 5. Annotations will be assigned in a two-tier process. Tier 1 annotations are referred to as “image labels” are associated with image metadata (BSH, quality etc) whilst “Tier 2” annotation are faunal annotations applied over observations of biota (See table 5 for further information on the Two-Tier system). Optionally tier 1 labels may be added directly into the proforma instead of being applied to still images, where this is done it must be consistent (i.e. there must not be a mixture of BIIGLE annotations and proforma entries for image quality) and agreed with Cefas/NE. More details on the still annotation types and how they must be applied are provided in Appendix A. When analysing still images the Supplier must use the provided BIIGLE label trees to analyse and label each image. The steps below describe the task required to produce the necessary information: 1. Extract and record NMBAQC image quality. A summary of these scores is shown in Table 2 but described in more detail in section 2.1 of Turner et al 2016. 2. Identify and enumerate epifauna using the Biota label tree. For each still image being analysed, identify and quantify all : • Solitary epifaunal species present (to the lowest taxonomic level for which identification can be guaranteed) and bioturbation traces using counts (achieved by using the point annotation tool within BIIGLE) • Colonial and/or encrusting and erect epifaunal species present as far as possible (to the lowest taxonomic or morpho-type level for which identification can be guaranteed) using percentage cover (achieved by using the polygon, magic wand, or brush tools within BIIGLE). • Presence of Non-native Invasive species (NIS) should be annotated (these individuals will be annotated in Tier 1 and Tier 2 (List provided in Annex B). 3. Use the Faunal turf label tree to record an approximate percentage cover of faunal turf. 4. Use laser points label tree to annotate lasers and calculate image area in BIIGLE (if laser have been used in the survey) 5. Identify evidence of anthropogenic impacts on the seabed • Use the litter label tree to record the presence of litter using the categories listed in Annex 5.1 of the Joint Research Centres Guidance on Monitoring of Marine Litter in European Seas. • Use the Anthropogenic label tree to annotate trawl marks or anthropogenic impacts other than litter. This will not be a complete label tree and new labels may need to be added to the label tree. Document any extensions. 6. Use the substrate label tree to record the substrate types present to the nearest 5%. 7. Any comments from the analyst which are considered of importance (e.g. presence of burrows) can be recorded in the proforma spreadsheet 8. A reference collection of representative images must be provided for each Biota label and discreet broad-scale habitat identified. A subset of each biota label will be annotated using a bounding box achieved using the ‘draw a rectangle’ tool up to a minimum of 10 where labels exist to be used for future work and to populate a machine learning image library. 2.2.3. GIS Processing The production of three separate Arc Map feature-classes within an ESRI geodatabase will be required. The task will centre on taking “sample data” from the still and video proformas provided (e.g. time, position, depth etc.) and combining these with outputs from the imagery analyses (“BSH”, biotope, substrate percentages etc.) and incorporating these into template feature-classes provided. A “File Naming Convention” document, example geodatabase and template geodatabase will be provided. The Supplier must: 1. Complete imagery analyses. 2. Complete “MPAGIS_Deliverables_Master.xlsx spreadsheet, with “Points” and “Line” tabs filled in (combination of proforma “sample data” and analysed outputs from imagery analysis from both the stills and the video datasets). 3. Create point shapefile from “Point” tab of the deliverable spreadsheet. This will be used to load data into the “_STILL” point feature-class and the “_VIDSOL” feature-class. 4. Create line shapefile from the “Line” tab of the deliverables spreadsheet. This will be used to load data into the “_VIDTR” feature-class in the geodatabase. 5. Load data from newly created shapefiles into template feature-classes within the template geodatabase 6. Complete feature-class and geodatabase metadata using metadata editor in ArcMap 10 or ArcPro • Metadata to be compliant with MEDIN Discovery Metadata Standard v3.1.2 and EMODNET DEF v4 . The provided “File Naming Convention” document details the specific formats required of each attribute field. • The example geodatabase provided will serve as a template for metadata completion. 3. Quality Assurance (QA) 3.1. Internal QA process for video analysis A minimum of 10% of video transects analysed (randomly selected by the Supplier) are to be subject to an internal QA process. Suppliers must demonstrate their internal Quality Control (QC) system, which must include routine checks of data and analysis. QC of data and analysis is where data are re-checked and compared against the original and compared to an expected standard/limit of error/deviation from the original. It is also where electronic records (in spreadsheets or databases) are checked against the data on the original field/laboratory datasheets. This measures the repeatability of the method of analysis and data entry, and highlights factors such as inter-worker variability. QC must also include details of remedial action in cases where expected standards/limits are not met. Internal QC must be undertaken by a staff member other than the staff member who conducted the original analysis. 3.2. Internal QA process for image analysis A minimum of 10% of still images analysed (randomly selected by the Supplier) are to be subject to an internal QA process. Suppliers must demonstrate their internal Quality Control (QC) system, which must include routine checks of data and analysis. QC of data and analysis is where data are re-checked and compared against the original and compared to an expected standard/limit of error/deviation from the original. It is also where electronic records (in spreadsheets or databases) are checked against the data on the original field/laboratory datasheets. This measures the repeatability of the method of analysis and data entry, and highlights factors such as inter-worker variability. QC must also include details of remedial action in cases where expected standards/limits are not met. Internal QC must be undertaken by a staff member other than the staff member who conducted the original analysis 4. Deliverables 4.1. Outputs The Supplier will provide analysis results in the MEDIN compliant MS Excel proforma spreadsheets which will be provided. The expected outputs/deliverables of the Contract are listed in Table 6. Please note the Final Report must be no longer than 10,000 words including all tables and appendices. The report must be provided electronically via email as a Microsoft Word document, and will include the following sections: • Methodology • Results • Details of all QA work undertaken, and any remedial action deemed necessary. 4.2. Data Standards: • All sample data (video/still photography analyses) must be entered into Marine Recorder Online and any import spreadsheets to be delivered with the final reports. Natural England will arrange system logins to Marine Recorder Online for the awarded Supplier to use. Guidance ‘Marine Recorder Guidance for Contractors’ will also be provided.. • All GIS datasets need to be provided in ESRI ArcGIS format compatible with ArcMap 10.7 and have attached metadata • All GIS files containing habitat data for each individual survey need to be produced to the EMODnet Seabed Habitats data exchange format, to the most detailed EUNIS (2007 Revised 2012) habitat level possible. • If not included in the GIS data layers listed above all sampling locations, vessels tracks, and links to data obtained should also be included as a single GI layer. • Accompanying metadata for the data set must meet the MEDIN metadata discovery standard.