Working Group on Nephrops Surveys (WGNEPS outputs from 2020)

ICES SCIENTIFIC REPORTS RAPPORTS SCIENTIFIQUES DU CIEM ICES INTERNATIONAL COUNCIL FOR THE EXPLORATION OF THE SEA CIEM CONSEIL INTERNATIONAL POUR L’EXPLORATION DE LA MER WORKING GROUP ON NEPHROPS SURVEYS ( WGNEPS; outputs from 2020) VOLUME 03 | ISSUE 36 International Council for the Exploration of the Sea Conseil International pour l’Exploration de la Mer H.C. Andersens Boulevard 44 -46 DK -1553 Copenhagen V De nmark Telephone (+45) 33 38 67 00 Telefax (+45) 33 93 42 15 www.ices.dk info@ ices.dk ISSN number: 2618 -1371 This document has been produced under the auspices of an ICES Expert Group or Committee. The contents therein do not necessarily represent the vie w of the Council. © 2021 International Council for the Exploration of the Sea. This work is licensed under the Creative Commons Attribution 4.0 International License (CC BY 4.0). For citation of datasets or conditions for use of data to be included in other databases, please refer to ICES data policy . ICES Scientific Reports Volume 03 | Issue 36 WORKING GROUP ON NEPHROPS SURVEYS (WGNEPS ; outputs from 2020) Recommended format for purpose of citation: ICES. 2021. Working Group on Nephrops Surveys (WGNEPS ; outputs from 2020) ICES Scientific Reports. 03:36. 114pp. https://doi.org/10.17895/ices.pub.8041 Editors Jennifer Doyle Authors Mikel Aristegui-Ezquibela • Jacopo Aguzzi • Candelaria Burgos • Jennifer Doyle • Niall Fallon Spyros Fifas • Jónas Jónasson • Patrik Jonsson • Mathieu Lundy • Michela Martinelli • Ivan Masmitja Gerald McAllister • Damir Medvešek • Atif Naseer • Charlotte Reeve • Cristina Silva • Julien Simon Jean-Philippe Vacherot • Maria Vigo-Fernandez • Yolanda Vila • Adrian Weetman • Kai Wieland ICES | WGNEPS 2020 | i Contents i Executive summary ....................................................................................................................... 2 ii Expert group information .............................................................................................................. 3 iii Terms of Reference ....................................................................................................................... 4 iv Work Plan Summary ...................................................................................................................... 6 1 Survey coordination (ToR a) .......................................................................................................... 7 2 Technological developments (ToR d) .......................................................................................... 15 2.1 Burrow emergence rhythms of Nephrops norvegicus: UWTV, surveying biases and novel technological scenarios ................................................................................. 15 2.2 Creel fishing and acoustic tracking trials in the No-Take zone off Palamós-Roses (Northwestern Mediterranean Sea) at 350-420 m depth.............................................. 17 2.3 Acoustic tracking of Nephrops norvegicus by networked moored hydrophones in a deep-sea no-take reserve of the North Western Mediterranean Sea. ....................... 22 2.4 Training Neural Networks on Nephrops survey datasets............................................... 26 2.4.1 The dataset .................................................................................................................... 26 2.4.2 The neural network ........................................................................................................ 28 2.4.3 Evaluating the neural network ....................................................................................... 28 2.4.3.a Comparison using the image per image method for the human analysis .................... 28 2.4.3.b Comparison watching the video at half speed ............................................................ 29 2.4.4 2.4.4 Examples of objects identified and confidence threshold number ...................... 31 2.5 Nephrops norvegicus detection and classification from underwater videos using Deep Neural Network. ................................................................................................... 32 2.6 Reducing uncertainty & Assessing Bias in estimates of Nephrops norvegicus population size (working paper). ................................................................................... 41 2.7 A review of FU 30 survey area definition ....................................................................... 48 2.8 High definition reference sets ........................................................................................ 53 3 Miscellaneous .............................................................................................................................. 55 3.1 GitHub update (ToR c) ................................................................................................... 55 3.2 Other ToR’s .................................................................................................................... 55 Annex 1: List of participants.......................................................................................................... 56 Annex 2: Resolutions .................................................................................................................... 58 Annex 3: Survey summaries .......................................................................................................... 59 Ireland ......................................................................................................................................... 59 UK Northern Ireland FU 15 .......................................................................................................... 70 UK Scotland ................................................................................................................................. 73 UK England FU 6 .......................................................................................................................... 80 UK England FU 14 ........................................................................................................................ 84 Denmark and Sweden FU 3-4: Skagerrak and Kattegat ............................................................... 87 Denmark FU 33: Off Horns Rev ................................................................................................... 93 Spain FU 30: Gulf of Cadiz ........................................................................................................... 93 Portugal FU 28-29: southwestern Portugal ................................................................................. 93 France FU 23-24: Bay of Biscay .................................................................................................... 94 Iceland FU 1: Off South Iceland ................................................................................................. 106 Italy and Croatia GSA 17 and 18: Adriatic Sea ........................................................................... 109 Annex 4: List of presentations .................................................................................................... 112 Annex 5: Action list ..................................................................................................................... 114 2 | ICES SCIENTIFIC REPORTS 03:36 | ICES i Executive summary The Working Group on Nephrops Surveys (WGNEPS) is the international coordination group for Nephrops underwater television and trawl surveys within ICES. This report summarizes the national contributions on the results of the surveys conducted in 2020 together with time series covering all survey years, problems encountered, data quality checks and technological improvements as well as the planning for survey activities for 2021. In total, 19 surveys covering 25 functional units (FU’s) in the ICES area and 1 geographical subarea (GSA) in the Adriatic Sea were discussed and further improvements in respect to survey design and data analysis standardization and the use of most recent technology were reviewed. Due to the COVID-19 pandemic there were disruptions across several functional units to: survey operations (FU 30, GSA 17, FU 10, FU 13 – Jura, FU 34, FU 28-29); data processing (FU 23-24) and survey coverage (FU 9). Further results from field studies on behaviour aspects of burrow emergence using bottom cages monitored by an automated camera system and on short-range migration using acoustic tracking and remote operated vehicle (ROV) surveys in marine protected areas have become available and are summarised in this report. Geostatistical investigations to reduce uncertainty estimates showed comparable results to historical trends for one survey area in the North Sea. Other preliminary work to redefine survey area using best available datasets was also discussed. Reference sets compilation and count evaluations using still image annotations and Lin’s concordance correlation coefficient (CCC) quality control were presented. Automatic burrow detection based on deep learning methods applied to a test dataset with annotated burrow counts from a HD camera system from two projects showed promising results. The working group members were encouraged to provide more material with annotated burrow counts for further development of machine learning tools. An underwater television (UWTV) survey manual has been accepted for publication in the ICES Techniques in Marine Environmental Sciences (TIMES) series. The working group is currently developing plans for a Nephrops UWTW database to be established at the ICES data centre. ICES | WGNEPS 2020 | 3 ii Expert group information Expert group name Working Group on Nephrops Surveys (WGNEPS) Expert group cycle Multiannual Year cycle started 2019 Reporting year in cycle 2/3 Chair(s) Jennifer Doyle, Marine Institute, Ireland Meeting venue(s) and dates 17-19 November 2020, Online Meeting (Webex), 26 participants 4 | ICES SCIENTIFIC REPORTS 03:36 | ICES iii Terms of Reference ToR Description Background Science Plan topics addressed Duration Expected Deliverables a To review any changes to design, coverage and equipment for the various Nephrops UWTV and full-scale trawl surveys since 2018 and to update the Series of ICES Survey Protocols (SISP) as required To ensure surveys used by WGCSE, WGBIE and WGNSSK are fit for purpose. 3.1, 3.2 Recurrent annual update Survey summary including and description of alterations to the plan, to relevant assessment-WGs (WGCSE, WGNSSK, WGBIE) and SCICOM. Planning of the upcoming surveys for the survey coordinators and cruise leaders, and update the SISP accordingly if necessary. b Develop an international database for Nephrops UWTV survey data which will hold burrow counts, ground shape files and associated data. There is a need to centralize UWTV data in a single international database. Ensure data is available externally. 3.5 Year 1-3 ICES database c Update R scripts for Nephrops UWTV survey data processing including functions to quality control, analyze and visualize data, and interface the tools with the international database for Nephrops UWTV survey data Improving standarisation of data QC and data processing. Support new developing surveys on data analysis. 3.1 Recurrent annual update Document and R packages for UWTV survey data on github site. d To review video enhancement, video mosaicking, automatic burrow detection and other new technological developments applied in Nephrops UWTV surveys and to update the Series of ICES Survey Protocols (SISP) as required . WGNEPS should periodically review emerging technologies that might improve survey methodologies. 4.1 Recurrent annual update To update the SISP based on conslusions if necessary. Other publications when appropriate. e Review and report on the utility of UWTV and trawl Nephrops surveys as platforms for collecting data for purposes other than Nephrops assessment (e.g. the collection of data for OSPAR and MFSD indicators). Nephrops UWTV surveys have a role in relation to benthic habitat monitoring and the collection of other environmental and ecosystem variables. 1.5 Year 2 Joint workshop/meeting report with users ICES | WGNEPS 2020 | 5 f Analyse existing data from UWTV and trawl Nephrops surveys to evaluate possible factors affecting burrow emergence of Nephrops (e.g. currents and light) Recent behaviour aspects have been investigated in the laboratory. Important to investigate correlation with field data. 1.3 Year 3 Review paper 6 | ICES SCIENTIFIC REPORTS 03:36 | ICES iv Work Plan Summary Year Summary Year 1 All ToRs will be adressed in this year but the the main task in year 1 will be to establish the UWTV database and to provide updated shape files of Nephrops FUs and survey domains (ToR b) Year 2 All ToRs will be adressed in this year. In addition to this focus will be on ToR e in year 2 Year 3 All ToRs will be adressed in this year. Focus in year 3 will be on new technologies and, if appropriate, an update of the SISP (ToR b) as well on the review of field date on factors affecting burrow emergence and occupancy (ToR f) Meeting dates Venue Reporting details Comments (change in Chair, etc.) Year 2019 12-14 November Split, Croatia 1st Interrim report by 6 January to EOSG Election of new chair(s) Year 2020 17-19 November Online meeting (Webex) 2nd Interrim report by 17 December 2020 to EOSG Change of chairs: Outgoing: Kai Wieland and Adrian Weetman Incoming: Jennifer Doyle Year 2021 16-18 November Cadiz, Spain Final report by 1 February 2022 to EOSG ICES | WGNEPS 2020 | 13 Figure. 1.4d Nephrops abundance (with 95 % confidence interval) in FU 22, FU 23-24, FU 30, FU 33 and FU 34. Dashed lines show proxy for MSY reference point Btrigger. 14 | ICES SCIENTIFIC REPORTS 03:36 | ICES The conclusions for future work are as follows: • WGNEPS recommends continuing with the use of high definition camera systems and still images with the objective to annotate images so that deep learning algorithms can be developed in future to identify features. • WGNEPS recommends promoting and facilitating when possible on UWTV surveys, staff exchange from national laboratories. • WGNEPS recommends promoting and facilitating when possible on UWTV surveys, staff exchange from other institutes who may use survey data. • WGNEPS recommends that national laboratories invest effort in calculating mean burrow size for specific grounds. The edge effect calculation is based on field of view (FOV) and burrow diameter. Mean burrow diameter can vary a lot over time for most grounds and this could have an impact on the edge effect. This will be added as a term of reference for this working group. • WGNEPS recommends exchange of technical expertise so that new and developing surveys may benefit from others. • WGNEPS agrees that it is mandatory that each station is read by at least two readers in accordance with agreed survey data processes. If there are any deviations to survey data work-up this is to be flagged prior to the time the data are to be used for assessment to the stock co-ordinator and chair of the relevant assessment working group. ICES | WGNEPS 2020 | 15 2 Technological developments (ToR d) 2.1 Burrow emergence rhythms of Nephrops norvegicus: UWTV, surveying biases and novel technological scenarios Aguzzi J., Bahamon N. and O'Malley C., Berry A., Gaughan P., Doyle J., Lordan C., Tuck I.D., Chiarini M., Martinelli M., Marini S., Thomsen L., Flögel S., Albiez J., Torkelsen T., Pfannkuche O., Rune Godo O., Henning W., Lopez-Vasquez V., Zuazo A., Rodriguez E., Valencia J., Calisti M., Stefanni S., Mirimin L., Del Río J.,Francescangeli M., Fahalazed A., Navarro J., Vigo M., Masmitjia I., García J.A., Chumbinho R., Company J.B. The occupancy assumption “one burrow system, one animal” (Sardà and Aguzzi, 2012) raises a number of generic research questions concerning the true occupation of burrows in many Nephrops stocks. The burrow system acts as the centre of a strong territorial rhythmic behaviour (Rice and Chapman, 1971; Farmer, 1975) leading the adults’ lobsters to evict subordinates from burrows in a dominance hierarchy framework (Sbragaglia et al., 2017); indeed, two wild adult lobsters are rarely found in the same shelter (Cobb and Wang, 1985). Other studies showed evidence that no spatial segregation occurs between juveniles and adults (Maynou and Sardà, 1997) achieving the establishment of adult-juvenile complexes (at least 1 adult and 1 juvenile per burrow), which become separated as juveniles grow (Tuck et al., 1994). Moreover, Nehprops burrows systems could also be inhabited by other benthic crustacean species (e.g. Munida sp.) or may remain empty and intact for an unknown period of time after animals’ death (Sardà and Aguzzi, 2012). These factors still create uncertainties about the true numbers of animals occupying burrow systems, representing an important issue when providing a relative or absolute index for determination of Nephrops’ stock status (i.e. Harvest Rate; Sardà and Aguzzi, 2012). For a better tuning of the occupancy assumption “1 burrow system, 1 animal”, an accurate temporal description of burrow emergence rhythmicity should be provided. The diel rhythm of burrow emergence can be subdivided in three different phases (Aguzzi et al., 2003, 2007): full emergence, full retraction and door-keeping (i.e. an intermediate period in which individuals wait at the burrow entrance; Sbragaglia et al., 2015). In Aguzzi et al. (submitted) more than three thousand video transects reporting densities by depth of full emergence and door-keeping animals and burrow systems collected in past decades around Ireland waters, are analysed. All density data were grouped per depth ranges based on both the available ones and the previous knowledges from trawl catch patterns (Aguzzi et al., 2003) as nominal: 15-50, 51-100, 101-160 and 340570 m. A waveform analysis on UWTV survey data were conducted to describe averaged full emergence and door-keeping behavioural rhythms over the 24-h within the established depth range. Such an analysis indicate that Nephrops full emergence varied from nocturnal toward midday hours with increasing depth of sampling, while door-keeping behaviour coincided with full emergence only on the upper shelf (15-50 m depth) and the shelf-break (101-160 m depth). To further improve the analysis GAM models for emergence and door-keeping behaviours by depth range were developed as well. The statistical model result by GAM revealed an overall pattern of full emergence and door-keeping behaviour similar to that found by the previous waveform analysis. The emergence behaviour is predominantly dusk and dawn-oriented above 50 m, bimodal and tending to be diurnal between 50 and 100 m, temporally diffused between 101 and 160 m, and finally fully diurnal between 340 and 570 m. The door-keeping behaviour is 16 | ICES SCIENTIFIC REPORTS 03:36 | ICES only temporally defined above 50 m (being nocturnal) and bimodal with a nocturnal increase between 100-160 m. Finally, estimated densities of visible animals engaged in both emergence and door-keeping behaviours (i.e. all individuals) were compared with burrow system counts and derived density estimates, to provide evidence putative biases to the standard stock assessment assumption that “1 burrow system is occupied and maintained by one animal” (Leocadio et al., 2018). A temporally integrated chart of all waveform and GAM results shows an average of about 1 visible individual per 10 burrows, at most, suggesting that a high proportion of the population remains cryptic even during periods of peak emergence. In last years, the novel technologies have become increasingly common in fish-stock assessment using video imagery from worldwide cabled observatory networks (Aguzzi et al., 2020; Del-Rio et al., 2020). The novel scenarios allow to collect observations on visible Nephrops individuals as well as their burrows through cabled observatory instrumented fields for ecological monitoring of fishery resources (e.g. OBSEA-www.obsea.es; and SmartBay Observatory-https://www.smartbay.ie/). Hence, the next steps for fishery-independent assessment calibration should be focused on new advanced imaging packages used on autonomous robotic platforms (e.g. crawlers, AUVs and stand-alone cameras) to tune the fishery-independent assessment equation “1 burrow-1 animal”. References Aguzzi, J., Sardà, F., Abelló, P., Company, J. B., and Rotllant, G. 2003. Diel and seasonal patterns of Nephrops norvegicus (Decapoda: Nephropidae) catchability in the western Mediterranean. Marine Ecology Progress Series, 258: 201–211. http://www.jstor.org/stable/24867045 (Accessed 11 May 2018). Aguzzi, J., Company, J. B., and Sardà, F. 2007. The activity rhythm of berried and unberried females of Nephrops norvegicus (Decapoda, Nephropidae). Crustaceana, 80: 1121–1134. Aguzzi, J., and Sardà, F. 2008. Biological rhythms in the marine environment: The Norway lobster as a case study. Contributions to Science, 3: 493–500. http://publicacions.iec.cat/repository/pdf/00000050/00000097.pdf. Aguzzi, J., Chatzievangelou, D., Company, J. B., Thomsen, L., Marini, S., Bonofiglio, F., Juanes, F., et al. 2020. The potential of video imagery from worldwide cabled observatory networks to provide information supporting fish-stock and biodiversity assessment. ICES Journal of Marine Science. https://doi.org/10.1093/icesjms/fsaa169. Atkinson, R. J. A., and Eastman, L. B. 2015. Burrow dwelling in Crustacea. The natural history of the Crustacea, 2: 78–117. Cobb, J. S., and Wang, D. 1985. Fisheries Biology of Lobsters and Crayfishes. In: Provenzano A.D. (Ed.) The biology of the Crustacea, 10: 167–247. https://books.google.com/books?hl=it&lr=&id=wU1mOBoer5IC&oi=fnd&pg=PA167&dq=Fishery+biology+of+lobsters+and+crayfish&ots=dXNqAKw61q&sig=qHELI0eDiRZToozZ8p8yA3YzO6Y (Accessed 23 October 2020). Del-Rio, J., Nogueras, M., Toma, D. M., Martínez, E., Artero-Delgado, C., Bghiel, I., Martinez, M., et al. 2020. Obsea: A Decadal Balance for a Cabled Observatory Deployment. IEEE Access, 8: 33163–33177. Farmer, A. S. D. 1975. Synopsis of the biological data on the Norway lobster Nephrops norvegicus (Linnaeus, 1758). FAO Fisheries Synopsis, 112: 1–97. Leocadio, A., Weetman, A. & Wieland, K. 2018. Using UWTV surveys to assess and advise on Nephrops stocks. ICES Cooperative Research Report No. 340. 49. Maynou, F., and Sardà, F. 1997. Nephrops norvegicus population and morphometrical characteristics in relation to substrate heterogeneity. Fisheries Research, 30: 139–149. Elsevier. Rice, A. L., and Chapman, C. J. 1971. Observations on the burrows and burrowing behaviour of two muddwelling decapod crustaceans, Nephrops norvegicus and Goneplax rhomboides. Marine Biology: ICES | WGNEPS 2020 | 17 International Journal on Life in Oceans and Coastal Waters, 10: 330–342. http://link.springer.com/10.1007/BF00368093 (Accessed 8 June 2018). Sardà, F., and Aguzzi, J. 2012. A review of burrow counting as an alternative to other typical methods of assessment of Norway lobster populations. Reviews in Fish Biology and Fisheries, 22: 409–422. http://link.springer.com/10.1007/s11160-011-9242-6 (Accessed 11 May 2018). Sbragaglia, V., Aguzzi, J., Garcıa, J. A., Sarriá, D., Gomariz, S., Costa, C., Menesatti, P., et al. 2013. An automated multi-flume actograph for the study behvioural rhythms of burrowing organisms. J. Exp. Mar. Biol. Ecol., 446: 177–185. Sbragaglia, V., García, J. A., Chiesa, J. J., and Aguzzi, J. 2015. Effect of simulated tidal currents on the burrow emergence rhythms of the Norway lobster (Nephrops norvegicus). Marine Biology, 162: 2007–2016. Springer Berlin Heidelberg. Sbragaglia, V., Leiva, D., Arias, A., Antonio García, J., Aguzzi, J., and Breithaupt, T. 2017. Fighting over burrows: the emergence of dominance hierarchies in the Norway lobster (Nephrops norvegicus). The Journal of Experimental Biology, 220: 4624–4633. http://jeb.biologists.org/lookup/doi/10.1242/jeb.165969. Tuck, I. D., Atkinson, R. J. A., and Chapman, C. J. 1994. The structure and seasonal variability in the spatial distribution of Nephrops norvegicus burrows. Ophelia, 40. 2.2 Creel fishing and acoustic tracking trials in the No-Take zone off Palamós-Roses (Northwestern Mediterranean Sea) at 350-420 m depth. Maria Vigo, Joan Navarro, José A. García, Jacopo Aguzzi, Guiomar Rotllant , Nixon Bahamón and Joan B. Company Marine Protected Areas (MPAs) have proven to be useful tools for conservation (Day et al., 2019), and they can offer many other benefits such as improving commercial fish stocks, including habitat restoration (Kerwath et al., 2013; Langton et al., 2020). In the context of the Spanish research project called RESNEP (CTM2017-82991-C2-1-R, “Marine no-take areas as a tool to recover iconic Mediterranean fisheries in decline: the case of Nephrops norvegicus”), a pilot marine reserve was established in an overfished ground at 350-400 m depth in the NW Mediterranean Sea, where Norway lobster (Nephrops norvegicus) dominated the target species fished by local and regional fisheries (BOE-A-2020-9015). Norway lobster constitutes an iconic fishing resource for European fisheries (Leocádio et al., 2012), whose landings have diminished the last two decades, especially in deep-water overfished benthic Mediterranean ecosystems (García-De-Vinuesa et al., 2020; Piroddi et al., 2020). The main objective of this marine no-take reserve, stablished on 2017, was to recover the population of Norway lobster as well as the recovery of the benthic assemblage and the habitat state. 18 | ICES SCIENTIFIC REPORTS 03:36 | ICES Figure 1. Spatial distribution of Norway lobster catches along the Catalan coast, and the location of the marine no-take reserve (green square) and the control area (yellow square) where the visual transects with ROV have been performed. The blue gradient indicates the accumulated catches of Norway lobster between 2006-2019. In the present communication, we present the preliminary results related to the ecological effects of the marine no-take reserve, after 3 years of implementation using ROV (Remotely Operated Vehicle) visual census, a non-invasive monitoring method. For this purpose, we conducted 24 h visual transects in the marine reserve and in a control area where fishing activity is still undergoing. These visual transects were performed in February 2020 on board R/V Sarmiento de Gamboa (Figure.2) Control MPA ICES | WGNEPS 2020 | 19 Figure 2. The Remotely Operated Vehicle (ROV) Liropus tfhat was used to performed all visual transects (A). The Research Vessel Sarmiento de Gamboa (B). The monitor in which the 24h visual transects were transmitted to annotated all the species that appear (C). During them, we assessed the abundance of Norway lobsters in the marine reserve and in the control area (Figure. 1). We will estimate, when was possible, the size of the individuals observed by calibrating the size with the scale of 10 cm provided by the ROV. Moreover, we annotated and classified the number of Norway lobsters located outside the burrows, or the ones performing door keeping behaviour (Aguzzi et al., 2007) in which we can see only the cephalotorax (Figure. 3). We also counted all the burrows, in presence of Norway lobsters or empty. Figure 3. Behaviours studied in Nephrops norvegicus: 1: outside the burrows; and 2: door keeping behaviour. Norway lobsters’ abundance showed high numbers in the marine reserve than in the control area (Figure. 4). The temporal analysis of ROV census data showed that lobsters were mainly outside their burrows during light hours, as previously confirmed in other studies 1 2 20 | ICES SCIENTIFIC REPORTS 03:36 | ICES (reviewed by Aguzzi & Sardà, 2008; Sardà & Aguzzi, 2012), being absent or at the tunnel entrance (i.e. neither visible as door keeping) at the darkness. Figure 4. Firsts results of diel activity of Norway lobster obtained with visual transects with ROV. Abundance of Norway lobster outside the burrow along the day (A). Abundance of Norway lobster doing door keeping behaviour along the day (B). The total swept area covered by the ROV is the same in both control and MPA areas being approximately 0.02km2. In addition to Norway lobster, our objectives were also to identify all the species that appear in both areas and measure the individuals (Figure. 5). Figure 5. Examples of other species found in the marine reserve and the control area during the visual transects with ROV. Lasers indicate 10 cm. In relation to the habitat state, the results showed that the control area presented alterations in the seabed, such as scraping and ploughing, directly associated with trawling activity. In contrast, the MPA showed a clear recovery of the benthos, evidencing the presence of well-structured burrow systems of Norway lobsters. We are annotating also all the debris that appear in both areas (Figure. 6) A B Lepidorhombus boscii Helicolenus dactylopterus Cerianthus sp. Nephrops norvegicus ICES | WGNEPS 2020 | 21 In conclusion, our results suggest that the implementation of marine no-take reserves could be an effective strategy contributing to recover the population of Norway lobster and other demersal species by reducing fishing pressure and promoting restoration of their habitats. References Aguzzi, J., Company, J.B., & Sardà, F. (2007). The activity rhythm of berried and unberried females of Nephrops norvegicus (Decapoda, Nephropidae). Crustaceana 80(9): 1121-1134. https://doi.org/10.1163/156854007782008577 Aguzzi, J., & Sardà, F. (2008). A history of recent advancements on Nephrops norvegicus behavioral and physiological rhythms. Reviews in Fish Biology and Fisheries, 18(2), 235–248. https://doi.org/10.1007/s11160-007-9071-9 Boletín Oficial del Estado. BOE-A-2020-9015. , Pub. L. No. 208, «BOE» núm. 208, de 1 de agosto de 2020, páginas 62842 a 62847 (6 págs.) 61561 (2020). Day, J., Dudley, N., Hockings, M., Holmes, G., Laffoley, D., Stolton, S., … Wenzel, L. (2019). Guidelines for applying the IUCN protected area management categories to marine protected areas Second edition. In Best Practice Protected Area Guidelines Series. Retrieved from www.iucn.org/pa_guidelines García-De-Vinuesa, A., Breen, M., Benoît, H. P., Maynou, F., & Demestre, M. (2020). Seasonal variation in the survival of discarded Nephrops norvegicus in a NW Mediterranean bottom trawl fishery. Fisheries Research, 230(May), 105671. https://doi.org/10.1016/j.fishres.2020.105671 Kerwath, S. E., Winker, H., Götz, A., & Attwood, C. G. (2013). Marine protected area improves yield without disadvantaging fishers. Nature Communications, 4, 1–6. https://doi.org/10.1038/ncomms3347 Langton, R., Stirling, D. A., Boulcott, P., & Wright, P. J. (2020). Are MPAs effective in removing fishing pressure from benthic species and habitats? Biological Conservation, 247(May), 108511. https://doi.org/10.1016/j.biocon.2020.108511 Leocádio, A. M., Whitmarsh, D., & Castro, M. (2012). Comparing trawl and creel fishing for Norway Lobster (Nephrops norvegicus): Biological and economic considerations. PLoS ONE, 7(7). https://doi.org/10.1371/journal.pone.0039567 Piroddi, C., Colloca, F., & Tsikliras, A. C. (2020). The living marine resources in the Mediterranean Sea Large Marine Ecosystem. Environmental Development, (May), 100555. https://doi.org/10.1016/j.envdev.2020.100555 Sardà, F., & Aguzzi, J. (2012). A review of burrow counting as an alternative to other typical methods of assessment of Norway lobster populations. Reviews in Fish Biology and Fisheries, 22(2), 409–422. https://doi.org/10.1007/s11160-011-9242-6 Plastic bottle Trawl mark 22 | ICES SCIENTIFIC REPORTS 03:36 | ICES 2.3 Acoustic tracking of Nephrops norvegicus by networked moored hydrophones in a deep-sea no-take reserve of the North Western Mediterranean Sea. Ivan Masmitja, Spartacus Gomariz, Joaquim del Rio; Universitat Politècnica de Catalunya (UPC), Barcelona, Spain. Joan Navarro, María Vigo, Jacopo Aguzzi, Nixón Bahamón, José Antonio García, Guiomar Rotllant, Joan B. Company; Institut de Ciències del Mar (ICM-CSIC), Barcelona, Spain. Knowing the displacement capacity and mobility patterns of fished marine resources is pivotal to establish effective conservation management strategies in marine ecosystems. Accurate behavioural information of deep-sea fished ecosystems is necessary, but currently scarce, to establish the sizes and adequate locations of marine protected areas within the framework of large international societal programs (e.g. European Community H2020, as part of the Blue Growth economic strategy). A breakthrough in the autonomous capability of mobile platforms to deliver data on animal behaviour beyond traditional fixed platform capabilities (e.g. cabled observatories) is overcoming these limitations. Here, we present useful example of that potential in relation to the implementation of autonomous underwater vehicles (AUVs) and remotely operated vehicles (ROVs) as an aid for acoustic long-baseline localization systems for autonomous tracking of Norway lobster (Nephrops norvegicus), one of the key resources exploited in European waters. We reported the outcomes of that monitoring in combination with seafloor moored acoustic receivers to detect and track the movements of 33 tagged individuals at 400 m depth over more than three months. We identified best procedures to localize both the acoustic receivers and the tagged-lobsters, based on cutting-edge algorithms designed for off-the-self acoustic tags identification. These procedures represent an important step forward for prolonged, in situ monitoring of deep-sea benthic animal behaviour at meter spatial scales. Figure 1. The strategy designed to track Norway lobsters (Nephrops norvegicus) is represented. Four receivers created an acoustic LBL localization system, where each one was in self-recording mode and was not accessed in real time. The tags transmitted periodically an acoustic ping, which was recorded by the static receivers and the underwater vehicles; both systems were used to track the lobsters’ movements. ICES | WGNEPS 2020 | 29 The precision (P), the recall (R) and the F1 measure are calculated after a confidence threshold of 0.6 on at least 2 frames. Table 1: Results on stations 2, 8, 69 (LANGOLFTV19) and 115, 145 (UWTV19) 2.4.3.b Comparison watching the video at half speed Each of the 5 stations have also been analyzed by 4 humans using VLC player and watching the video at half speed. For each object the humans could detect, the video was paused, the data (time, specie, number…) was reported in an excel sheet. This method is more representative of the usual method for analyzing the videos. However, in some case, some objects are not detected by the humans as the sledge is moving too fast or as the objects are in the shadow. In this study, the 4 humans analyzed the 5 stations in order to have the inter-observer variability and to compare the counts of the neural network (Figure. 4) to the counts of the 4 humans. The burrows counts are reported table 2. It appear that the neural network is over counting. This is mainly due because the neural network is counting the burrows and not the complexes while humans count the complexes. Classes TP FP FN P R mAP0,5 F1 all 564 535 408 0,513 0,58 0,385 0,544 nephrops_norvegicus 86 27 12 0,761 0,878 0,652 0,815 pennatulacea 136 70 14 0,66 0,907 0,331 0,764 shrimp 12 2 69 0,857 0,148 0,284 0,252 actiniaria 37 11 32 0,771 0,536 0,656 0,632 munida 96 7 35 0,932 0,733 0,471 0,821 actinopterygii 31 25 11 0,554 0,738 0,421 0,633 burrow 166 391 235 0,298 0,414 0,259 0,347 30 | ICES SCIENTIFIC REPORTS 03:36 | ICES Figure 4: Comparison between humans counts of different species. Black dots = the humans count watching at half speed, red dots = neural network counts, orange triangle = ground truth (human counts image per image) Station Counter 1 Counter 2 Mean Neural Network Stn2_LANGOLF19 36 30 33 31 Stn8_LANGOLF19 100 95 97,5 100 Stn69_LANGOLF19 13 16 14,5 5 Stn13_LANGOLF20 166 178 172 187 Stn145_LANGOLF20 124 130 127 208 Stn167_LANGOLF20 68 74 71 90 Table 2: Burrow counts results from different stations ICES | WGNEPS 2020 | 31 2.4.4 2.4.4 Examples of objects identified and confidence threshold number Figure. 1 and 2 show Nephrops burrows identified and associated confidence threshold number. Figure. 3 shows Nephrops norvegicus and Munida species identified with confidence number. Figure 1. Still image from neural network programme. Figure 2. Still image from neural network programme. 32 | ICES SCIENTIFIC REPORTS 03:36 | ICES Figure 3. Still image from neural network programme. 2.5 Nephrops norvegicus detection and classification from underwater videos using Deep Neural Network. Atif Naseer 1. Introduction Spanish Institute of Oceanography has a research group working on Nephrops norvegicus identification and counting. They are conducting the survey on yearly basis. The survey is conducted through special equipment and underwater camera. A 10-12 minutes video was made on each point of interest and the whole survey has more than 20-30 points of interest yearly. Currently they are counting the holes manually by reviewing the video frame by frame in multiple parallel session and conclude the results on consensus of all members. This exercise cost lot of resources in terms of time, human and cost. There is no system available that can help them in solving their current problem. During the past many years Nephrops are counted manually (counting from TV surveys) from underwater videos which is very tedious and time-consuming task. These species are usually lived under the seabed and leaving behind some pattern of burrows. To identify this specie in underwater, one need to identify these patterns and judge the availability of Nephrops. The Nephrops burrows are very specific in their characteristics. Some of the major characteristics of burrows are: 1. At least one burrow opening is usually distinctly crescentic (half-moon) in shape. Where the angle of view permits sight of the tunnel beyond this opening, the angle of descent is usually shallow. 2. There is often evidence of expelled sediment, usually in a broad delta-like ‘fan’ at the burrow opening, and scrapes and tracks are often apparent. 3. Nephrops may be present (either in or out of burrow). ICES | WGNEPS 2020 | 33 The objective of this research project is to develop a deep learning model to automatically detect, classify and count the Nephrops burrows. To achieve A deep learning based automatic system to detect, classify and count the Nephrops Burrow complexes will be developed. The proposed work is using current state of the art Deep neural networks for objects detection and classification. To improve the detections the models, require some fine tuning and addition of more layers. In this work, the Nephrops surveys from Cadiz and Ireland are analyzed using Faster RCNN deep neural networks. The results show some good true positive detection from Cadiz and Ireland data. 2. Research Methodology The system main objective is to develop an auto detection mechanism to classify and count the Nephrops burrows systems. Following are the main phases that are required to achieve the objective. A. Data Preparation a) Data Collection The data used for experimentation and model training is from Cadiz and Ireland stations. The proposed deep learning model requires homogeneous data for training. The data collected from Cadiz is in the form of High Definition videos from the survey of 2018 and 2019.The duration of each video is 9-11 minutes. Each video is 25 frames per seconds. An individual video consists of 15000 frames on average. The data collected from Ireland is in the form of HD quality images. More than 1000 images were collected from Ireland. Table 1. Shows the raw dataset and its attributes. Table 1: Dataset Attributes Station Year Videos Images Cadiz 2018 100 minutes 150,000 Cadiz 2019 100 minutes 150,000 Ireland 2019 NA 1650 b) Data Cleaning In the initial step all the images from Cadiz and Ireland were studied and removed if the lightening conditions and contrast of images are too bad to recover. Also, the repeated frames from the same video will not be considered in the dataset used for annotations. The available data require preprocessing due to its heterogeneous nature. The quality of videos will be improved by improving the lightening effects, noise mitigation, color compensation and image contrast enhancement. c) Ground truth image annotations The major step to prepare a good dataset is to annotate the Nephrops burrows. The ground truth annotations are the key for model training. To annotate the images, the Visual Object Tagging Tool (VOTT) from Microsoft has been used. VOTT helps in end to end machine learning pipeline. The tool allows to download the annotation in various format like csv, Jason, XML etc. From the Ireland dataset, out of 1650 images, 1133 images annotated and recorded 1699 annotations of Nephrops burrows in these images. From the Cadiz dataset only 266 images annotated and recorded 350 annotations. 34 | ICES SCIENTIFIC REPORTS 03:36 | ICES d) Testing and Validation of annotations Once all the ground truth annotations are recorded, now its time to validate the annotations before preparing the dataset for model training. The annotation validation is only possible from experts of Nephrops. Dr. Yolanda Vila from Cadiz helps in validating the ground truth annotations of Cadiz and Jennifer Doyle from Marine Institute of Ireland validated the ground truth annotations of Ireland. e) Data preparation for Model Testing The last step of this phase is to prepare the dataset for training the model. Table 2. Shows the annotated images of each station from Ireland and Cadiz that will be used in the model training and testing. Only 2018 survey of Cadiz is used in this dataset preparation. Total seven stations are annotated from Cadiz and recorded 266 annotated images. From Ireland survey, seven stations are annotated and recorded 1133 annotated images. Table 2: Dataset Preparation Cadiz Dataset Ireland Dataset Station* Annotations Station Annotations RF01 42 Stn1 141 RF03 75 Stn10 201 RF04 34 Stn11 145 RF05 31 Stn15 179 RF07 13 Stn16 154 RF08 36 Stn26 155 RF09 35 Stn27 158 Total 266 Total 1133 B. Model Training In model training phase, a deep neural model will be trained using the prepare dataset. Following are the steps required for training a model. a) Dataset Format Each annotated image has downloaded in an xml file which contains the information of image name, Class name (Nephrops), and bounding box detail of each annotation in the form of Xmin, Ymin, Xmax, Ymax. These Pascal VOC. b) Dataset Distribution To train a deep neural model, the data should be divided into train, validate and test. Table 3. Shows the distribution of this Cadiz and Ireland dataset. ICES | WGNEPS 2020 | 35 Table 3: Dataset Distribution Cadiz Dataset Ireland Dataset Training Images Validation Images Testing Images Training Images Validation Images Testing Images 200 (75%) 18 (7%) 48 (18%) 619 (55%) 155 (14%) 359 (31%) Total Images = 266 Total Images = 1133 c) Model Training Faster RCNN is an object detection architecture presented by Ross Girshick, Shaoqing Ren, Kaiming He and Jian sun in 2015, and is one of the famous object detection architectures that uses convolution neural networks. We trained more complex and denser model based on Faster RCNN, those are: i. MobileNet v2 ii. Inception v2 iii. Resnet50 iv. Resnet101 d) Combination of Dataset for Training and Testing With these complex models, we used combination of our available dataset from Cadiz and Ireland for training and testing. To train the models following combination of datasets are used. i. Cadiz Dataset ii. Ireland Dataset iii. Hybrid Dataset (Combination of Cadiz and Ireland) Each model is trained with 70k iterations and precision are calculated on every 10k iteration. The tables from 4 to 7 shows the combination of training and testing dataset. Table 4: Dataset for MobileNetV2 36 | ICES SCIENTIFIC REPORTS 03:36 | ICES Table 5: Dataset for InceptionV2 . Table 6: Dataset for ResNet50 ICES | WGNEPS 2020 | 37 Table 7: Dataset for ResNet101 For each model used in our study, we performed certain number of experiments based on the combination of data we used. For every model used, at least nine different combination of dataset are applied. Each model is run with 70k iterations, so every model has in total 63 experiments performed. Table 8. shows the number of experiments performed for all the models. Table 8: Total number of experiments performed. C. Model Testing a) Test Data From Cadiz dataset 48 images are used in the testing of model and 359 images from Ireland dataset is used in the testing. b) Quantitative Analysis The Table 9. shows the performance evaluation of all the models used in the experimentation. The models are trained by Cadiz, Ireland and Hybrid dataset. While tested by all the combination of these dataset. A total of nine combination of experiments performed for each model to measure the performance evaluation in terms of mean Average Precision (mAP). 38 | ICES SCIENTIFIC REPORTS 03:36 | ICES Table 9: Performance evaluation of Models c) Qualitative Analysis Here we compare the visual results of Nephrops burrows detection with all the models. The Inception and ResNet101 performs better in detecting more numbers of True Positive burrows. The figure below shows the detections of all the models. ICES | WGNEPS 2020 | 45 Estimates of mean abundance derived using CGS were all lower than the estimates using the current method, apart from the 2016 estimate (Figure. 5), but the difference was only of borderline significance (F(2, 42) = 1.5, p = 0.09). The resampling-based mean abundance estimates tended to be closer to (but were also generally lower than) standard estimates, and followed the same overall temporal trend. There was no significant difference between the resample-based estimates and those derived using the standard method (F(2, 42) = 1.5, p = 0.42). Geostatistical abundance estimates, and their associated confidence limits were all within the uncertainty bounds of the standard abundance estimates, indicating that any deviations in temporal trend observed in the time-series of CGS estimates (e.g. slight difference between the two methods in trend from 2014 to 2016) were plausible in the context of the standard method, and the current Figure 4. Mean Nephrops burrow density distribution for FU12, calculated across 500 CGS realisations using 2011 UWTV burrow density data (overlaid as a black bubble plot where bubble area is proportional to burrow density, the black x symbols represent zero density observations). Darker red pixels represent areas of higher Nephrops burrow density, and lighter yellow pixels represent areas of low density. Cream coloured pixels represent areas where burrows are absent. 46 | ICES SCIENTIFIC REPORTS 03:36 | ICES understanding of the stock’s dynamics, given the uncertainty around those estimates. Differences in magnitude aside, the three time series were highly correlated (pairwise Pearson correlation coefficients >0.88), suggesting generally good agreement between the trends observed across methods (Figure. 6). Figure 5. Time series estimates of FU12 Nephrops abundance with 95% CIs using the standard method (blue line and polygon) overlaid with geostatistical estimates of mean total abundance with 95% quantiles (orange points and lines), and resample estimates (black boxplots; the centre line is the mean, the whiskers are at the 95% quantiles of the bootstrap distribution). Figure 6. Pairwise comparisons of mean abundance estimates derived using CGS (Geostatistical abundance estimates), the resampling routine, and the standard method. Each plot panel includes a 1:1 line to aid in comparison of time series. ICES | WGNEPS 2020 | 47 Compared to the standard method, there was a mean reduction in coefficient of variation of 87% using the CGS method (Figure. 7), suggesting that CGS may represent a viable abundance estimation approach for FU12 Nephrops with greatly decreased uncertainty compared to the standard method. Discussion CGS can provide estimates of Nephrops abundance for FU12, which have reduced uncertainty when compared to the standard method, while being of a comparable magnitude and following similar historical trends. As such, CGS may offer a solution to the long-standing issue of highly uncertain abundance estimates for that management area. Ultimately, the outcome of the method relies heavily on the ability to fit a useful variogram model. It is thus useful to fully explore the sensitivity of the variogram model fits to the assumptions applied in the calculation of the empirical variogram (e.g. distance lag). It may be useful to expand this analysis to multiple FUs to assess the performance of the CGS estimation method against the standard method in different scenarios. Given the non-significant difference between the resampling-based method and the standard method, it does not appear that substantial bias has been introduced to the assessment due to the UWTV sample allocation method. Regardless, it would be favourable to correct the minor discrepancies in proportionate sample allocations for future surveys. References ICES 2018. Report of the Working Group on Nephrops Surveys (WGNEPS). 6-8 November. Lorient, France. ICES CM 2018/EOSG:18. 226 pp. MINES ParisTech / ARMINES 2020. RGeostats: The Geostatistical R Package. Version: 11.2.13. Free download from: http://cg.ensmp.fr/rgeostats. Petitgas, P., Woillez, M., Rivoirard, J., Renard, D., and Bez, N. 2017. Handbook of geostatistics in R for fisheries and marine ecology. ICES Cooperative Research Report No. 338. 177 pp. Figure 7. Time series of coefficients of variation for standard (grey line) and geostatistical (orange line) estimates of FU12 Nephrops abundance. 48 | ICES SCIENTIFIC REPORTS 03:36 | ICES Rivoirard, J., Simmonds, J., Foote, K. G., Fernandes, P. G., and Bez, N. 2000. Geostatistics for Estimating Fish Abundance. Blackwell Science, Oxford. 206 pp. Woillez, M., Rivoirard, J., and Fernandes, P. G. 2009. Evaluating the uncertainty of abundance estimates from acoustic surveys using geostatistical simulations. – ICES Journal of Marine Science, 66: 1377–1383. 2.7 A review of FU 30 survey area definition Yolanda Vila and Candelaria Burgos ISUNEPCA UWTV survey is carried out in the Gulf of Cadiz (UF 30) yearly in spring-summer since 2014, although the first survey is considered as exploratory. ISUNEPCA is a multi-disciplinary survey and different specific objectives are established: 1. To obtain estimates of Nephrops burrows densities 2. To confirm the boundaries of the Nephrops area distribution 3. To obtain estimates of macro benthos species and the occurrence of trawl marks and litter on the sea bead 4. To collect oceanographic data by means of a CTD coupled to the sledge 5. To collect sediment samples 6. Seabed morphological and backscatter analysis The design of the survey follows a randomized isometric grid at 4 nm spacing. Since 2016, stations are allocated in the grid in a rhomboidal way. A total of about 65-70 stations are yearly planned covering the Nephrops area distribution established in the last benchmarck (ICES, 2016). Footages have been recorded by a HD camera during the period 2015-2017 while a 4K UHD recording camera is used since 2018 (ICES, 2018). Unfortunately, ISUNEPCA UWTV survey could not be conducted in 2020 due the COVID-19 pandemic. A review of the ISUNEPCA UWTV survey area has been carried out and presented during WGNEPS 2020. The current survey area used to obtain the Nephrops abundance estimate in the Gulf of Cadiz (FU30) was established mainly based on a combination of VMS and logbook data analysis (2011-2012) (ICES, 2016). Additional information as the Nephrops abundance from ARSA IBTS surveys (SP-GCGFS-Q1 and Q4) time series (1994-2014) and bathymetric and morphologic information (Díaz del Río et al., 2014; Vila et al., 2016) was also used. This area corresponds to 3000 Km2 and covers depths ranging between 90 m to 600 m, approximately. However, data compiled and the experience acquired during ISUNEPCA UWTV survey time series suggest that the shallowest limit and the Southern border could be different and, as a consequence, the survey area should be probably smaller than the current area. These facts could directly affect the Nephrops abundance estimate. Besides, visibility at those depths is very poor and the presence of other species with a burrowing behavior generates a high uncertainty in the Nephrops burrows identification. For that reason, the stations located in the shallowest limit of the area have been considered stations with zero Nephrops density in the last three years (ICES, 2017; 2018; 2020). New and more accurate information is available now. One of them is the Andalusian monitoring system , called SLSEPA (“Sistema de Localización y Seguimiento de embarcaciones Pesqueras Andaluzas”), installed in most of fleets in the gulf of Cadiz, that transmit hour and positions (provided by (GPS), course and speed to the control centre every three minutes, (instead the two hours interval of European VMS) allowing for an accurate estimate of the actual fishing activity using a quite simple method not relying on strong assumptions. Additionally, updated data from ARSA IBTS survey time series (1993-2020) and beam trawl information obtained in the ICES | WGNEPS 2020 | 49 ISUNEPCA UWTV survey during 2017-2019 periods, as well as recent habitat, sediment and the seabed morphology information (Lozano et al., 2019; Lozano et al., 2020) could be also very useful in order to redefine the survey area in FU 30. Figure 1 shows the SLSEPA information linked to sales notes analysis in 2019 for the bottom trawl fleet in the Gulf of Cadiz (FU30). Landings data were apportioned to estimated fishing points for mapping the spatial distribution of the catch according to Gerritsen and Lordan (2010). Different filters were applied, as selecting records with speed value less or equal than 5 knots and deleting records located in shallow waters, less than 100 m deep where Nephrops is not targeted. The spatial distribution of the catches was estimated by summing the catch of points within 0.5 nm2 grid cells, a sufficient resolution based on the total size of study area. a) b) Figure 1. Analysis of Andalusian vessel monitoring system (SLSEPA) linked to sales notes from the bottom trawl fleet in 2019. a) Taking into account all vessels; b) Eliminating vessels than have not fish at more than 200 m deep in the same day with catches lower than 5Kg/day. Red polygon represents the current area used in ISUNEPCA UWTV surveys. Most of the points located in 100-200 m stratum correspond to vessels that have also fished at more than 200 m deep the same day. Nevertheless, there are a proportion of vessels than only fish in shallower waters and have not fished in the deeper strata. So, a more detailed analysis of the 100-200 m stratum was carried out. Catches have been analyzed by ranges and it get have verified that in the shallower area in front of Cadiz bay, catches never exceeded 5 Kg/day, while higher catches correspond to vessels having fish also close to the 200 m isobaths. Vessels positions in the 100-200 m stratum with catches lower than 5Kg/day were excluded (Figure 1b). The Nephrops abundance from ARSA IBTS surveys (SP-GCGFS-Q1 and Q4) time series indicates a very few quantities of Nephrops in that stratum (100-200 m), as well as in the Southern border of the current UWTV survey area, with only some exceptions during the time series (1993-2020) (Figure 2a). The results obtained from the beam trawl hauls conducted during ISUNEPCA UWTV surveys in the 2017-2019 period showed presence of burrowing crustaceans as Goneplax rhomboids in the 100-200 m stratum but no individuals of Nephrops were caught in them (Figure 2b). 50 | ICES SCIENTIFIC REPORTS 03:36 | ICES a) b) Figure 2. a) Nephrops abundance from ARSA IBTS surveys time series (1994-2020); b) Beam trawl hauls from ISUNEPCA UWTV surveys (2017-2019). The symbol + corresponds to zero Nephrops. Red polygon represents the current area used in ISUNEPCA UWTV surveys. Different geological and oceanographic processes determine the distribution of a wide of geomorphological features, habitats and species in the Gulf of Cadiz. Channels, diapiric ridges and mud volcanoes can be found in the area (Figure 3) which harboring distinct benthic and demersal associated communities and habitats (Díaz del Río et al., 2014; Rueda et al., 2012). Some of them were taken account to establish the Nephrops distribution area used to ISUNEPCA UWTV survey in 2016 (Vila et al., 2016). However, more detailed seabed morphology information, as well as, new information about sediment and habitats in the Gulf of Cadiz are now available (Lozano et al., 2019; Lozano et al., 2020), which can be very useful for this issue. Figure 3. Main geomorphological seafloor features in the Gulf of Cadiz. Source: INDEMARES/CHICA Project (LIFE07/NAT/E/000732). Figure 4a overlaps the results of SLSEPA analysis and Nephrops abundance from ARSA IBTS surveys, with the surrounding area (green line) and the current ISUNEPCA survey area (dark red line). The geomorphic seafloor features have been taken account only in a rough way up the moment. Nevertheless, a more detailed redefinition of the area will be done in a near future, considering that information, as well as, the sediment composition and habitats results obtained by Lozano Chica 1&2 Anastasya Gazul Albolote Tarsis Pipoca ICES | WGNEPS 2020 | 51 and collaborators in 2019 and 2020. On the other hand, some stations carried out during the ISUNEPCA UWTV survey time series, where Nephrops burrows systems were identified, would stay out of the new area proposed (Figure 4b). This survey is a relatively new, as it started in 2014. The low experience in the identification and quantification of the Nephrops burrows when the time series started could be the explication for the presence of Nephrops in this part of the area. For this reason, a review of the Nephrops density in those stations is needed in order to check them. a) b) Figure 4. Preliminary (green polygon) and current (red polygon) ISUNEPCA UWTV survey area overlapped on the Andalucian vessel monitoring system (SLSEPE) linked to sales notes analysis from the bottom trawl fleet in 2019: a) Nephrops density from ARSA IBTS Survey time series (1993-2020) in green bubbles; b) Nephrops density from ISUNEPCA UWTV survey time series (2015-2019) in blue bubbles. Conclusions and recommendations 1. Results obtained indicate that the ISUNEPCA UWTV survey area should be reduced, mainly in the shallowest and Southern border. 2. The survey area presented in this WG must only be considered as preliminary. 52 | ICES SCIENTIFIC REPORTS 03:36 | ICES 3. The Nephrops density in those stations staying out of the proposed area must be checked. In addition, the more detailed geomorphological seafloor features, sediment and habitat available information, should be taken into account. 4. The WGNEPS recommends finalizing this analysis before WGBIE, where a working document should be presented with the work conducted and the proposed new area for the ISUNEPCA UWTV survey. WGBIE should establish the procedure to follow in order to change the survey area in FU30. References Díaz del Rio, V., Bruque, G., Fernández-Salas, L.M., Rueda, J.L., González, E., López, N., Palomino, D., López, F.J., Farias, F., Sánchez-Leal, R., Vázquez, J.T., Rittierott, C.C., Fernández, A., Marina, P., Luque, V., Oporto, T., Sánchez-Gillamón, O., García, Urra, J., Bárcenas, P., Jiménez, M.P., Sagarminaga, R. and Arcos, J.M., 2014. Volcanes de fango del golfo de Cádiz, Proyecto LIFE + INDEMARES. Ed. Fundación Biodiversidad del Ministerio de Agricultura, Alimentacion y Medio Ambiente. 2014. Gerritsen, H and Lordan, C., 2010. Integrated Vessel Monitoring System (VMS) data with daily catch data from logbooks to explore the spatial distribution of catch and effort at hig resolution. ICES Journal of Marine Science. 68. 10.1093/icesjms/fsq137. ICES, 2016. Report of the Benchmark Workshop on Nephrops stocks (WKNEP). ICES CM: 2016/ACOM: 38 ICES. 2018. Report of the Working Group on Nephrops Surveys (WGNEPS). ICES CM 2018/EOSG:18. 226 pp. ICES. 2020. Working Group on Nephrops Surveys (WGNEPS; outputs from 2019). ICES Scientific Reports. 2:16. 85 pp. http://doi.org/10.17895/ices.pub.5968. Lozano, P., Rueda, J.L., Gallardo-Núñez, M., Farias, C., Urra, J. Vila, Y., López-González, N., Palomino, D., Sánchez-Guillamón, O., Vázquez, J.T. and Fernández-Salas, L.M., 2019. Habitat distribution and associated biota in different geomorphic features within a fluid venting area of the Gulf of Cádiz (South Western Iberian Peninsula, NE Atlantic Ocean). In: Seafloor Geomorphology as Benthic habitat. GeoHAB Atlas of Seafloor Geomorphic Features and Benthic Habitats, chapter 52. 2ª edition. Eds: P. Harris & E. Baker. 10.1016/B978-0-12-8149607.00052-X. Lozano, P., Fernández-Salas, L.M., Hernández-Molina, F., Sánchez-Leal, R.F., Sánchez-Guillamón, O., Palomino, D., Farias, C., Mateo-Ramírez, A., López-González, N., García, M., Vazquez, J.T., Vila, Y. and Rueda, J.L., 2020. Multiprocess interaction shaping geoforms and controlling substrate types and benthic community distribution in the Gulf of Cádiz. Marine Geology. 423. 106139. 10.1016/j.margeo.2020.106139. Rueda, J.L., Díaz del Río, V., Sayago-Gil,M., López-González, N., Fernández-Salas, L.M. and Vázquez, J.T., 2012. Fluid Venting Through the Seabed in the Gulf of Cadiz (SE Atlantic Ocean, Western Iberian Peninsula): Geomorphic Features, Habitats, and Associated Fauna. In: Seafloor Geomorphology as Benthic Habitat GeoHAB Atlas of Seafloor Geomorphic Features and Benthic Habitats. Chapter 61. 1º edition. Eds: P. Harris & E. Baker. 2012 10.1016/B978-0-12-385140-6.00061-X. Vila, Y., Burgos, C., and Soriano, M.M., 2016. Nephrops (FU 30) UWTV Survey on the Gulf of Cadiz Grounds. WD presented to ICES Benchmark on Nephrops stocks (WKNEPS 2016). 24-28 October 2016, Cádiz (Spain). ICES | WGNEPS 2020 | 53 2.8 High definition reference sets Mikel Aristegui, Marine Institute, Ireland Since 2019, Irish UWTV surveys have been recorded in high definition camera. The digital format of the new footage allows remote analysis of the images in laptops and do not need any more CRT monitors and DVD players. This became a key feature in 2020, since COVID-19 restrictions did not allow the footage to be counted as usual onboard the Celtic Voyager. However, UWTV reference sets used by the Marine Institute (Ireland) until 2019 were recorded in DVDs using the previous UWTV standard definition camera. This means that counters would have not been able to be trained remotely before counting 2020 survey footage. Therefore, prior to the 2020 UWTV season, the Marine Institute decided to renew all their reference sets (FU16, FU17, FU2021 and FU22) using high definition footage from 2019 surveys. In order to undertake such an important job, the Marine Institute followed the reference set compilation recommendations from WKNEPS (ICES 2018). The detailed procedure carried out for every Functional Unit’s reference set is detailed below: A) Selection of stations: 1. Take all the UWTV 2019 survey stations. 2. Split stations in High, Moderate and Low densities (tertiles). 3. Sort each density group by Lin’s CCC obtained by the 2019 pair of counters. 4. By default: Choose the three highest Lin’s CCC stations from each density group. But ensuring there is a variety of features among stations, such as:  Presence/absence of trawl marks.  Presence/absence of sea-pens.  Different ground types.  Nephrops in and out.  Some stations with low Lin’s CCC. 5. End up with 9 stations for the reference set. B) Generate reference counts: 1. Two experts involved: one expert running the annotation app (Aristegui 2020) and sharing the screen remotely with the second one. 2. Open the station with the Annotation app in SIC_matching mode, which shows the annotations made by the 2 reviewers who counted the station back in 2019 (Figure 2.7.1). 3. The two experts review together every single annotation made by those 2 reviewers, and confirm or reject each annotation. Generating the reference counts was a time consuming task for the two scientists involved. The lowest density stations were reviewed in around 15 minutes, but more than one hour was needed for the highest density stations. The plan was to split each FU in three work sessions, aiming to review three stations in each session. However, a total of 14 sessions were needed for generating the four reference sets. The final output of the full process was four HD reference sets (FU 16, FU 17, FU 20-21 and FU 22), each of them containing nine UWTV stations of eight minutes. The new reference sets do not only contain the count of burrows per minute (as standard definition sets used to contain), but also each burrow’s annotation in the footage. Afterwards, training versions of the reference sets were created, including only annotations of the first two minutes of each station (or alternatively 54 | ICES SCIENTIFIC REPORTS 03:36 | ICES for very low density stations, annotations of the first few burrows of the station). The annotated reference sets are a highly valuable tool and will be key in future burrow identification training. Figure 2.7.1. Marine Institute’s annotation app example (Aristegui 2020). Left: UWTV high definition still image with a Nephrops burrow annotated by the two counters in 2019 (yellow and red circles). Top-right: summary of the station with a coloured map showing all the annotations from the two counters and a coloured table with the number and percentage of their matches. Bottomright: clickable list of every single annotation, which allows instant visualization of each of them. References Aristegui, M. (2020) Image annotation R Shiny app. Marine Institute. http://doi.org/d24n ICES. 2018. Report of the Workshop on Nephrops Burrow Counting (WKNEPS). 2-5 October. Aberdeen, UK. ICES CM 2018/EOSG:25. 44 pp. ICES | WGNEPS 2020 | 61 Beam Trawling Operations. Due to time constraints in 2020 beam trawl fishing operations were not carried out on the Aran Nephrops grounds (FU 17) and the Smalls Nephrops grounds (FU 22). Other Benthic fauna distributions. Monitoring the occurrence and frequency of other sea-pens observed on Nephrops grounds is important but is dependent on national resources. An OSPAR special request to record sea pens species (Virgularia mirabilis, Funiculina quadrangularis and Pennatula phosphorea) using a key devised to categorise the density (ICES, 2011) exists. In 2020 presence/absence of these three species was recorded in FU 16, 17, 19, 20-21 and 22. Figure 4 shows the 2020 stations on the Porcupine Nephrops grounds where the aforementioned sea-pen species were identified and noted as present or absent. The deep water sea-pen Kophobelemnon stelliferum has been observed during the UWTV survey on the Porcupine Banks (FU 16) Nephrops ground. It is an easy species to identify from the image data due to its specific shape and colour. Seapen presence/absence data from the FU 16 Porcupine UWTV survey was provided as part of a 2020 datacall for new information on Vulnerable Marine Ecosystems (VME) in the North Atlantic for the Joint ICES/NAFO Working Group on Deep-water Ecology (ICES, 2020). Table 1. Nephrops UWTV survey datasets currently available on the Marine Institute Data Catalogue. Nephrops UWTV Survey Dataset Marine Data Catalogue Link FU 22 https://tinyurl.com/yxo6ltnh FU 20-21 combined https://tinyurl.com/y3yfgzq9 FU 16 https://tinyurl.com/y2s6pbgx Table 2. 2020 UWTV mean adjusted density, abundance estimate, CV (relative standard error) and Lin’s Concordance Correlation Coefficient (CCC) threshold by Functional Unit. UWTV Survey Mean density adjusted (burrow/m²) Final Abundance Estimate (millions of individuals) CV (Relative standard error) Lin’s Concordance Correlation Coefficient Threshold to screen survey Counts FU 16 0.17 1264 4% 0.6 FU 17 Aran Grounds only 0.29 359 4% 0.6 FU 19 0.16 320 15% 0.5 FU 20-21 combined 0.102 1020 5% 0.5 FU 22 0.27 750 8% 0.6 62 | ICES SCIENTIFIC REPORTS 03:36 | ICES Figure 1. Time series of the total number of UWTV stations carried out by Ireland in each Functional Unit. Stations in FU 14 and FU 15 are carried out in collaboration with AFBI in UK-NI and CEFAS UK E&W. Figure 2. Screenshot of the Image annotation R shiny app used to annotate UWTV footage during the 2020 surveys. Blue circle denotes annotated burrow system. ICES | WGNEPS 2020 | 63 Figure 3. 2020 Mean adjusted density estimates (burrow/m²) by station for Nephrops grounds in ICES Subarea 7. Figure 4. FU16 grounds: 2020 stations where Virgularia mirabilis (VAM), Funiculina quadrangularis (FAQ), Pennatula phosphorea (PNP) and Kophobelemnon stelliferum (KOP) were identified and noted as present or absent. Closed circles indicated presence and open circles denotes TV stations with no sea-pen observations. 64 | ICES SCIENTIFIC REPORTS 03:36 | ICES UWTV Survey FU16: Porcupine Banks. The 2020 survey was multi-disciplinary in nature collecting UWTV and other ecosystem data. In total 65 UWTV stations were successfully completed in a randomised 6 nautical mile isometric grid covering the full spatial extent of the stock. The mean burrow density observed in 2020, adjusted for edge effect, was 0.17 burrows/m². The final krigged abundance estimate was 1264 million burrows with a CV of 4% and an estimated stock area of 7,130 km2. The 2020 abundance estimate was 25% higher than in 2019. Using the 2020 estimate of abundance and updated stock data implies catches between 2653 and 3290 tonnes in 2021 that correspond to the F ranges in the EU multiannual plan for Western Waters (assuming that all catch is landed). Four species of seapen; Virgularia mirabilis, Funiculina quadrangularis, Pennatula phosphorea and the deepwater seapen Kophobelemnon stelliferum were observed during the survey. Trawl marks were also observed on 22% of the stations surveyed. Further details on this survey available at: http://hdl.handle.net/10793/1655 Figure 5. FU 16 Porcupine Bank: Violin and box plot of adjusted burrow density distributions by year for the available time series 2012 to 2020. No UWTV survey in 2015. The blue line indicates the mean density over time. The horizontal black line represents the median, white box is the inter quartile range, the black vertical line is the range and the black dots are outliers. UWTV Survey FU17: Aran grounds, Galway Bay and Slyne Head Nephrops grounds. In 2020 the nineteenth annual underwater television on the Aran, Galway Bay and Slyne head Nephrops grounds, ICES assessment area; Functional Unit 17 was successfully carried out. The survey was multi-disciplinary in nature collecting UWTV and other ecosystem data. In 2020 a total of 44 UWTV stations were successfully completed, 34 on the Aran Grounds, 5 on Galway ICES | WGNEPS 2020 | 65 Bay and 5 on Slyne Head patches. The mean burrow density observed in 2020, adjusted for edge effect, was medium at 0.29 burrows/m². The final krigged burrow abundance estimate for the Aran Grounds was 359 million burrows with a CV (Coefficient of Variance; relative standard error) of 4%. The final abundance estimate for Galway Bay was 27 million and for Slyne Head was 7 million, with CVs of 13% and 4% respectively. The total abundance estimates have fluctuated considerably over the time series. The 2020 combined abundance estimate (394 million burrows) is 20% lower than in 2019, and it is below the MSY Btrigger reference point (540 million burrows). Using the 2020 estimate of abundance and updated stock data implies catches between 443 and 508 tonnes in 2021 that correspond to the F ranges in the EU multi annual plan for Western Waters, assuming that discard rates and fishery selection patterns do not change from the average of 2017–2019. Virgularia mirabilis was the only sea-pen species observed on the UWTV footage. Trawl marks were present at 7% of the Aran stations surveyed. Further details on this survey available at: http://hdl.handle.net/10793/1656 Figure 6. FU17 Aran grounds: Violin and box plot of adjusted burrow density distributions by year from 2002-2020. The blue line indicates the mean density over time. The horizontal black line represents the median, white box is the inter quartile range, the black vertical line is the range and the black dots are outliers. 66 | ICES SCIENTIFIC REPORTS 03:36 | ICES Figure 7. FU17 Galway Bay and Slyne Head: Violin and box plot of adjusted burrow density distributions by year from 2002-2020. The blue line indicates the mean density over time. The horizontal black line represents the median, white box is the inter quartile range, the black vertical line is the range and the black dots are outliers. UWTV Survey FU19. South and South west coast of Ireland. The survey was multi-disciplinary in nature collecting UWTV other ecosystem data. In 2020 a total 42 UWTV stations were successfully completed. The mean density estimates varied considerably across the different patches. The 2020 raised abundance estimate was a 20% decrease from the 2019 estimate and at 320 million burrows is below the MSY Btrigger reference point (430 million). Using the 2020 estimate of abundance and updated stock data implies catch in 2021 that correspond to the F ranges in the EU multi annual plan for Western Waters are between 531 and 595 tonnes (assuming that discard rates and fishery selection patterns do not change from the average of 2017–2019). Two species of sea pen were observed; Virgularia mirabilis and Pennatula phosphorea which have been observed on previous surveys of FU19. Trawl marks were observed at 26% of the stations surveyed. Further details on this survey available at: http://hdl.handle.net/10793/1654 ICES | WGNEPS 2020 | 67 Figure 8. FU19 grounds: Violin and box plots of adjusted burrow density distributions by year for 2006-2020 for each ground. The blue line indicates the mean density over time. The horizontal black line represents the median, white box is the inter quartile range, the black vertical line is the range and the black dots are outliers. No TV survey from 2007 – 2010. UWTV Survey FU20-21: Labadie, Jones and Cockburn Banks. The 2020 survey achieved full coverage of the stock area for the seventh successive time. Area of this ground is calculated at 10 014 km² which is the largest Nephrops ground in ICES area 7 (ICES, 2014). The 2020 survey was multi-disciplinary in nature collecting UWTV and other ecosystem data. A total of 97 UWTV stations were completed at 6nm intervals over a randomised isometric grid design. The mean burrow density was 0.102 burrows/m2 compared with 0.06 burrows/m2 in 2019. The 2020 geostatistical abundance estimate was 1020 million, a 65% increase on the abundance from 2019, with a CV of 5%, which is well below the upper limit of 20% recommended by SGNEPS 2012. Low to medium densities were observed throughout the ground. Using the 2020 estimate of abundance and updated stock data implies catch in 2021 that correspond to the F ranges in the EU multi annual plan for Western Waters are between 1682 and 1710 68 | ICES SCIENTIFIC REPORTS 03:36 | ICES tonnes (assuming that discard rates and fishery selection patterns do not change from the average of 2017–2019). One species of sea-pen (Virgularia mirabilis) were recorded as present at the stations surveyed. Trawl marks were observed at 36% of the stations surveyed. Further details on this survey available at: http://hdl.handle.net/10793/1657 Figure 9. FU20-21 grounds: Violin and box plot of adjusted burrow density distributions by year from 2013-2020. The blue line indicates the mean density over time. The horizontal blacks line represents medians, white boxes the inter quartile ranges, the black vertical lines are the range and the black dots are outliers. UWTV Survey FU22: The Smalls. The 2020 survey was multi-disciplinary in nature collecting UWTV and other ecosystem data. A total of 40 UWTV stations were surveyed successfully (high quality image data), carried out over an isometric grid at 4.5nmi or 8.3km intervals. The precision, with a CV of 8%, was well below the upper limit of 20% recommended by SGNEPS (ICES, 2012). The 2020 abundance estimate was 33% lower than in 2019 and at 750 million is below the MSY Btrigger reference point (990 million). Using the 2020 estimate of abundance and updated stock data implies catch in 2021 that correspond to the F ranges in the EU multi annual plan for Western Waters are between 1238 and 1560 tonnes (assuming that discard rates and fishery selection patterns do not change from the average of 2017–2019). One species of sea pens was recorded as present at the stations surveyed: Virgularia mirabilis. Trawl marks were observed at 48% of the stations surveyed. Further details on this survey available at: http://hdl.handle.net/10793/1658 ICES | WGNEPS 2020 | 69 Figure 10. FU22 Smalls grounds: Violin and box plot of adjusted burrow density distributions by year from 2006-2020. The blue line indicates the mean density over time. The horizontal black lines represent medians, white boxes the inter quartile ranges, the black vertical lines the range and the black dots are outliers. References Aristegui, M. 2020. Image annotation R Shiny app. Marine Institute. http://doi.org/d24n ICES. 2011. Report of the ICES Advisory Committee 2011. ICES Advice.2011. Book 1: Introduction, Overviews and Special Requests. Protocols for assessing the status of sea-pen and burrowing megafauna communities, section 1.5.5.3 ICES. 2014. Report of the Benchmark Workshop on Celtic Sea stocks (WKCELT), 3–7 February 2014, ICES Headquarters, Copenhagen, Denmark. ICES CM 2014\ACOM:42. 194 pp. ICES. 2020. ICES/NAFO Joint Working Group on Deep-water Ecology (WGDEC). ICES Scientific Reports. 2:62. 188 pp. https://doi.org/10.17895/ices.pub.7503 Lin, L. I-K. 1989. A Concordance Correlation Coefficient to Evaluate Reproducibility. Biometrics, 45(1), 255-268. doi:10.2307/2532051 70 | ICES SCIENTIFIC REPORTS 03:36 | ICES UK Northern Ireland FU 15 (Mathieu Lundy) Functional Unit FU 15 Area name Western Irish Sea Survey design Random grid Previous surveys 2003-2019 Country (ies) UK & Ireland Vessel name (s) R/V Corystes Survey code (s) CO3120 Dates (start/end) 6th – 12th Aug 2020 Number scientific staff 5 Staff exchanges N/A Number of stations (planned/completed/used in analysis) 100/99/99 Deviations from the survey plan (e.g. coverage/weather related problems, technical problems, potential biases, etc.) No deviations. Ship position used for distance over ground as in 2019 Distance over ground source used Ship Average field of view (cm) Analogue cam: 68 cm Adjusted mean density 0.82 Adjusted abundance, CV 4872 million, CV=2.91% Overall footage quality (poor, medium, good) Medium Reference footage for survey area generated Yes Quality control of station counts (Lin’s CCC or consensus count) State Lin’s CCC threshold Lin’s CCC threshold 0.5 Other survey activities (CTD, Trawl, sediment samples, sediment profile images, % stations with trawl marks recorded, presence/absence sea-pen distribution etc.) CTD Beam trawl hauls Nephrops otter trawls Data storage, level of analysis and dissemination (by data type) Nephrops burrow counts 9706 Nephrops burrows counted, storage: DVD up to 2020, level of analysis: kriged estimates as for last year dissemination: WGCSE CTD 99 Trawl 48 Sediment 0 Other 0 ICES | WGNEPS 2020 | 77 • including these two new areas in the MRV Scotia survey significantly increased the mileage and associated steaming time, reducing UWTV deployment time. Considering these added challenges it was not possible to conduct all the standard activities during the survey, and priorities were identified and modifications applied: • UWTV activity in the six main functional units was the main priority of the survey; • no sediment samples were taken; • no trawling was undertaken; • the Noup, Devils Hole and the Sound of Jura were not surveyed; • the planned number of stations in each area were reduced proportionally in relation to previous years, depending on various factors (except in the South Minch where variability in the area traditionally remained high). Following modified COVID inductions, training and a review of amended risk assessments aboard the vessel, the first area to be surveyed was the Firth of Forth, where station numbers were reduced significantly as the area has a well-established, steady fishery and homogenous grounds. Fladen was then surveyed, and although the abundance data has been relatively stable over time, there has been a slight downturn over the last three years. In addition, due to the size of the grounds any major reduction in station numbers would have a disproportionally affect on the analysis, and so the number of stations were reduced the least at Fladen. The survey then continued down west side of the North Minch and into the South Minch. With the North Minch survey area based on one strata (that of VMS data) this allowed a slightly larger reduction in stations than other areas; whereas the South Minch traditionally showed high variance due to a wide range of benthic strata over a large geographical area and therefore planned station numbers remained unchanged. In both Minches, a number of COMPASS moorings were recovered and replacement arrays deployed. These moorings are part of a long term, Interreg project involving five institutions which aims to build cross-border capacity for effective monitoring and management of Marine Protected Areas (MPAs). The moorings associated with this survey were laid on the seabed with various acoustic devices attached to enable the monitoring of passing fauna by recording and counting the number of vocal interactions to establish the frequency and variety of cetaceans visiting the west coast of Scotland. The vessel then proceeded into the Clyde, and although a relatively small area to survey, the benthic variability resulted in only a marginal reduction in the number of stations in this functional unit. Due to the reduced number of days available on this survey, the ability to adapt the survey plan to ensure surveying the Clyde was conducted during the weekend (when trawler activity is not permitted) was not possible. Therefore the survey was conducted during the week resulting in poorer visibility which was reflected in the QC plots. On completing all the scheduled TV stations the vessel recovered a marine passive acoustic monitoring mooring (MarPAMM) from alongside the MPA to the south of the Isle of Arran, which was also providing data for the COMPASS project (Figure 3). The remaining stations in the South and then North Minch were completed on the return leg of the journey. However due poor weather the expected time available to survey the Moray Firth was reduced, impacting on the achievable number of stations even further. As a result less than half the number of the stations that are normally surveyed during the August/September MRV Alba-na-Mara survey were conducted. In addition, although randomly generated, many of these 78 | ICES SCIENTIFIC REPORTS 03:36 | ICES stations appeared to be located near to the edge of the known muddy habitat. This increased the variability between counters and introducing a third counter on this occasion did not resolve the situation. Due to the limited number of reviewers aboard and the reduced time available for reviewing whilst at sea, some first stage counts and a number of third counts had to be completed following the survey. This had a significant impact on the provision of data ahead of the annual assessment working groups, and although all required deadlines were met, this situation highlighted the need to ensure time was made available for critical staff post-survey if this atypical scenario was ever repeated. All video footage, both at sea and onshore, was reviewed in accordance with WGNEPS guidance, with quality control being carried out on all data using Lin’s CCC, with third counts applied where thresholds were not met (see Table 1 below). Reference sets for the three remaining areas ((Firth of Forth (FU 8), Moray Firth (FU9) and Clyde (FU 13) using 2018 footage was collated, completing the revised reference sets for MSS. However due to staff and time constrictions during the survey the footage remained unassessed. All survey data were uploaded to the bespoke MSS UWTV database. Conclusions/recommendations/aspirations: • To further encourage and promote national and international staff exchange. • To continue to promote the UWTV surveys to being open to alternative, but appropriate and collaborative, use of staff experience and ship’s time to improve cost and time efficiencies, widen the survey remit and increase staffs’ skill base. • To increase the number of MSS staff suitably trained to assist in UWTV surveys. • To submit tenders to replace the failing motion compensated sea going balances. • To submit tenders for the provision of a copper/fibre optic hybrid cable and associated high definition camera. • To prepare and present at WGNEPS 2021 updated analysis of Nephrops morphometric and maturity related data, gathered from UWTV surveys. • To prepare and present at WGNEPS 2021, updated analysis of Nephrops weight/length data, gathered from UWTV surveys. • To ensure sufficient staff aboard surveys to carry out all analysis in a timely fashion; or if footage has to be reviewed post-survey, to prioritise this work, ensuring sufficient time is allocated to achieve this task as soon as possible. • To continue collaborating with the Joint Nature Conservancy Council (JNCC) in analysing UWTV footage for associated studies; and continue to contribute to the UK marine image collation, processing, storage, annotation and promotion work shops (The Big Picture) chaired by JNCC. ICES | WGNEPS 2020 | 79 Table 1 Summary of Nephrops burrow abundance related activities carried out within the six survey areas during the MRV Scotia cruise in June 2020. Survey design: RS – S, random stratified based on sediment; RS – E, random stratified based on VMS effort; Fixed, survey stations are fixed due to the challenging topography and/or a legacy component. Figure 3. Map illustrating the location of the UWTV stations and COMPASS mooring recoveries/deployments that were conducted within the six survey areas during the MRV Scotia cruise during June 2020. -8 -6 -4 -2 0 2 56 57 58 59 Scotia UWTV Nephrops Surv Lon Lat Completed Survey Activities, c Survey ActivitySurvey Activity Completed TV Stations Moorings Area Number of TV sledge deployments Number of fishing trawls Number of sediment samples Linn’s CCC threshold Lin’s CCC pass rate Survey design type Firth of Forth 34 0 0 0.5 61.7 RS –S Fladen 61 0 0 0.7 49.0 RS –S North Minch 33 0 0 0.5 75.8 RS – E & F South Minch 45 0 0 0.5 55.5 RS -S Clyde 34 0 0 0.5 47.0 RS -S Moray Firth 34 0 0 0.5 54.1 RS –S Totals 231 0 0 80 | ICES SCIENTIFIC REPORTS 03:36 | ICES UK England FU 6 (Charlotte Reeve) Functional Unit FU 6 Area name Farn Deeps Survey design Fixed Grid Previous surveys 1997-2019 (Except 1999 & 2000) Country (ies) England Vessel name (s) RV Cefas Endeavour Survey code (s) CEND0920 Dates (start/end) 29 th June - 10 th July 2020 Number scientific staff 10 Staff exchanges 0 Number of stations (planned/completed/used in analysis) 110/110/110 Deviations from the survey plan (e.g. coverage/weather related problems, technical problems, potential biases, etc.) Inclement weather from 4 th to 5 th July slowed operations and reduced visibility at 10 stations. Distance over ground source used Ships positioning/transponder Average field of view (cm) Width of view 82.5 (distance between lasers) Adjusted mean density 0.35 Adjusted abundance, CV 1102 million Overall footage quality (poor, medium, good) 76% Good, 21% Moderate, 3% Poor Reference footage for survey area generated No. Footage from 2018 used. Quality control of station counts (Lin’s CCC or consensus count) State Lin’s CCC threshold Lin’s CCC, threshold 0.5 Other survey activities (CTD, Trawl, sediment samples, sediment profile images, % stations with trawl marks recorded, presence/absence sea-pen distribution etc.) CTD twice daily. ESM2 logger attached recording turbidity reading, depth, salinity, and oxygen levels. Data storage, level of analysis and dissemination (by data type) Nephrops burrow counts Storage: MP4 files Level of Analysis: Krigged Dissemination : ICES Advice, WGNSSK CTD Twice daily Trawl None Sediment None Other None ICES | WGNEPS 2020 | 81 Table 1: UWTV Summary FU 6. Year Number of Stations (used in the analysis) Abundance adjusted estimate (millions of burrows) CV on Burrow estimate % 1997 87 1500 4.3 1998 91 1090 4.2 1999 - - - 2000 - - - 2001 180 1685 2.0 2002 37 1048 5.5 2003 73 1085 4.2 2004 76 1377 3.7 2005 105 1657 4.6 2006 105 1244 4.7 2007 105 858 1.4 2008 95 987 2.0 2009 76 682 2.8 2010 95 785 1.4 2011 97 878 1.0 2012 97 758 0.9 2013 110 706 1.3 2014 110 755 0.9 2015 110 568 1.3 2016 110 697 1.2 2017 110 909 1.4 2018 109 950 1.2 2019 91 1163 1.2 2020 110 1102 1.1 82 | ICES SCIENTIFIC REPORTS 03:36 | ICES Figure 1a: FU 6 Map of density by station for each year. ICES | WGNEPS 2020 | 83 Figure 1b: FU 6 Map of density by station for each year. 84 | ICES SCIENTIFIC REPORTS 03:36 | ICES UK England FU 14 (Charlotte Reeve) Functional Unit FU 14 Area name Eastern Irish Sea Survey design Fixed Grid Previous surveys 2007-2019 Country (ies) England & N.Ireland Vessel name (s) RV Corystes Survey code (s) CO3120 Dates (start/end) 6 th – 8 th July 2020 Number scientific staff See FU 15 report Staff exchanges 0 Number of stations (planned/completed/used in analysis) 46/43/43 Deviations from the survey plan (e.g. coverage/weather related problems, technical problems, potential biases, etc.) See FU15 report. No Cefas staff exchanged on survey due to Covid-19, Cefas participated in station counting. Distance over ground source used Ship positioning Average field of view (cm) 68cm Adjusted mean density 0.46 Adjusted abundance, CV 496 million (CV: 8.6%) Overall footage quality (poor, medium, good) No footage quality recorded Reference footage for survey area generated No Quality control of station counts (Lin’s CCC or consensus count) State Lin’s CCC threshold Lin’s CCC threshold 0.5 Other survey activities (CTD, Trawl, sediment samples, sediment profile images, % stations with trawl marks recorded, presence/absence sea-pen distribution etc.) CTD Presence/absence ancillary data collected Data storage, level of analysis and dissemination (by data type) Nephrops burrow counts Storage: DVD Level of Analysis: Krigged Dissemination: ICES Advice, WGCSE CTD Every station Trawl None Sediment None Other None ICES | WGNEPS 2020 | 85 Table 1: FU 14 UWTV Summary. Year Number of Stations (used in the analysis) Abundance adjusted estimate (millions of burrows) CV on Burrow estimate % 2007 - - - 2008 32 407 - 2009 32 350 - 2010 26 422 - 2011 26 449 11.8 2012 26 693 7.8 2013 31 487 9.1 2014 34 449 10.7 2015 42 590 7.9 2016 48 429 12.6 2017 45 579 7.8 2018 46 513 12.6 2019 46 399 9.3 2020 43 496 8.6 86 | ICES SCIENTIFIC REPORTS 03:36 | ICES Figure 1: FU 14 Map of density by station for each year. ICES | WGNEPS 2020 | 93 Denmark FU 33: Off Horns Rev (Kai Wieland) Bi-annual survey. No survey planned in 2020. Next survey scheduled for 3 – 12 May 2021. See ICES. 2020. Working Group on Nephrops Surveys (WGNEPS; outputs from 2019). ICES Scientific Reports. 2:16. 85pp. http://doi.org/10.17895/ices.pub.5968 for results of the previous surveys. Spain FU 30: Gulf of Cadiz (Yolanda Vila, Candelaria Burgos) Due to COVID-19 pandemic the UWTV survey was not carried out in 2020. See ICES. 2020. Working Group on Nephrops Surveys (WGNEPS; outputs from 2019). ICES Scientific Reports. 2:16. 85pp. http://doi.org/10.17895/ices.pub.5968 for results of the previous surveys. Portugal FU 28-29: southwestern Portugal (Cristina Silva) Due to COVID-19 pandemic and vessel issues the trawl survey was not carried out in 2020. 94 | ICES SCIENTIFIC REPORTS 03:36 | ICES France FU 23-24: Bay of Biscay (Spyros Fifas, Jean-Philippe Vacherot) Historical context The UWTV survey named "LANGOLF-TV" has been conducted since 2014 aiming to demonstrate the technical feasibility of such a survey in the local context and to identify the necessary competences and equipment for its sustainability. During the first two years, 2014 and 2015, video sampling was associated to a trawl one for the purpose of providing Nephrops LFDs by sex and estimating the proportion of other burrowing crustaceans (mainly Munida) which can induce bias in the burrows counting. The surface involving in Nephrops is precisely delimited owing two information: (1) on the sedimentary structure of the seabed already taken into account during the former LANGOLF trawl survey on years 2006-2013 (5 spatial strata; Figure. 1); (2) on the systematic grid of video tracks combined with VMS data for the fishery (Figure. 2; data source: National Fisheries Direction; compilation: Ifremer). Sampling of landings and discards (onboard and at auction) has provided yearly dataset since 1987 and mainly since 2003 owing to the monitoring of the European DCF plan (Table 1; Figure. 3). The 2016’s WKNEP benchmark validated the UWTV survey and the assessment combining burrows counting and the SCA model for this stock. The change of the stock status from category 3 to 1 implies annual advice instead of the biennial one applied previously. ICES | WGNEPS 2020 | 95 Figure 1. Spatial stratification of the Bay of Biscay according to sedimentary criteria as considered from the first UWTV survey onwards (2014) and sampling design 2020 before COVID-19 crisis (left) and finally retained (right). Figure 2. UWTV stations on a systematic grid and VMS data for retained catches of Nephrops (example of the year 2016; source: National Fisheries Direction; compilation: SIH Ifremer). 96 | ICES SCIENTIFIC REPORTS 03:36 | ICES Figure 3. LFDs (size in carapace length, mm) for landings and discards by sex. Example of dataset 2019. Sampling Protocol In accordance with other routinely UWTV surveyed stocks, the sampling protocol applied since 2014 has been a systematic one advantaged by wider spatialised explorations on collected data. A distance of 4.7 nautical miles was retained similarly to the FU22 Smalls Ground. From 2016 onwards the survey duration has been longer than previously: 14 effective working days were planned (instead of 10). Thus, it has been allowed to cover for the first time the area contained in the outline of the Central Mud Bank no belonging to any sedimentary stratum: this area known as not trawled due to rough sea bottom concentrate moderate fishing effort targeting Nephrops (16164 km² were covered by sampling instead of 11676 km² of the historical five sedimentary strata). In the 2018's UWTV survey, an additional area of ≈2200 km² was investigated with 31 validated stations added to the 184 ones contained in the 2016's benchmarked area of 16164 km². In 2019 a supplementary area of ≈930 km² was sampled with 7 validated stations whereas the standard benchmarked area contained 145 ones. In 2020, due to the COVID-19 pandemic, the survey initially scheduled at late April/early May was strongly compromised, before being rescheduled from 22nd July to 4th August, with only two Irish scientists experienced in this type of mission in order to respect the obligatory social distancing on board (31 m vessel: "Celtic Voyager"; Irish company P&O). It was decided to reduce the sampling plan to 130 stations allowing to obtain statistically acceptable precision level of estimates and to make all video interpretations by Ifremer agents in lab after the end of the survey. The basis of the 2020's plan was the 2018's survey because its coverage was more complete than in 2019. Among the 2018's 184 validated stations contained in the Central Mud Bank benchmarked outline, 10 corresponding to zero burrows counted in 2018 as well as in 2019 were erased. The choice of 130 stations was ended by a random process eliminating 44 stations among the 174 remaining ones. Owing to favourable meteorological conditions, the initial goal was exceeded and 134 validated stations were finally sampled. ICES | WGNEPS 2020 | 97 Table 1. Nephrops in the Bay of Biscay (VIIIab). Above: Landed and discarded weights since the DCF routinely conducted sampling onboard. Below: Discards and landings in numbers (103 individuals) obtained by sampling onboard and at auction. Only years with sampling onboard are presented. Landings (1) Total Discards Catches Year FU 23-24 (2) FU 23 FU 24 Unallocated (MA N) (3) Total VIIIa,b used by WG FU 23-24 Total VIIIa,b VIIIa VIIIb VIIIa,b VIIIa,b 2003 1 3564 322 49 3886 1977 5863 2004 na 3223 348 5 3571 1932 5503 2005 na 3619 372 na 3991 2698 6689 2006 na 3026 420 na 3447 4544 7990 2007 na 2881 292 na 3176 2411 5587 2008 na 2774 256 na 3030 2123 5154 2009 na 2816 212 na 2987 1833 4820 2010 na 3153 245 na 3398 1275 4673 2011 na 3240 319 na 3559 1263 4822 2012 na 2290 230 na 2520 1012 3532 2013 na 2195 185 na 2380 1521 3900 2014 na 2699 108 na 2807 1326 4133 2015 na 3425 144 na 3569 1822 5391 2016 na 3873 217 na 4091 2531 6622 2017 na 3283 129 na 3412 2387 5799 2018 na 2038 86 na 2125 1571 3696 2019 na 2065 89 na 2154 634 2789 (1) WG estimates (2) landings from VIIIa and VIIIb aggregated until 1974 (3) outside FU 23-24 Italic font: revised value between WGBIE 2019 and 2020 (from 1627 t to 1571 t) 98 | ICES SCIENTIFIC REPORTS 03:36 | ICES Year Discards Landings % discarding 1987 268 244 288 974 48 1991 151 634 217 338 41 1998 150 995 161 549 48 2003 201 841 152 485 57 2004 222 089 139 753 61 2005 315 346 166 165 65 2006 487 288 127 942 79 2007 214 788 117 273 65 2008 198 031 115 274 63 2009 174 480 123 504 59 2010 113 530 138 120 45 2011 121 603 108 011 53 2012 117 935 101 424 54 2013 154 914 114 853 57 2014 117 930 121 594 49 2015 156 400 138 921 53 2016 200 973 161 371 55 2017 200 600 143 502 58 2018 151 926 83 463 65 2019 59 102 96 919 38 In 2020, LANGOLF-TV was carried out on 10 actual days (July 22nd-31st). The equipment (sledge, computing hardware, screens, recorders) were provided by the "Marine Institute". The sledge is based on the Scottish material (2.5 m*2.7 m*2.5 m; weight=80 kg); its speed is around 20 m/min. As for 2019's survey, the new HD system CathX was adopted this year. The reduction in the number of stations was based on the 2018 campaign (239 stations also including the area outside the benchmarked edge of the Central Mud Bank; 184 stations in the stock validated area and 55 elsewhere) as follows:  10 stations to zero burrows in 2018 and 2019  7 rocky stations in 2018  5 stations not validated in 2018  12 stations intentionally abandoned in 2018 on sandy areas with no appearance of burrows in previous years ICES | WGNEPS 2020 | 99  31 stations outside the Benchmark 2016 framework  44 stations removed by random draw and including all strata Acquiring images on the sea bottom requires a preliminary use of multi-beam sounder aiming to determine the nature of the sediment and to avoid technical problems due to rough ground. The recording starts when the sledge reaches the adequate speed (∼0.8 knots), the contact with the sediment is conform. Recording lasts 10 min even with no Nephrops burrows on the track; 7 min minimum are necessary for the validation of the footage. Up to 2019's survey, the provisional absence of reference footage in the Bay of Biscay implied the use of other support coming from grounds with similar conditions (density of burrows) to the Bay of Biscay: the Smalls grounds (FU22, Celtic Sea, UWTV surveyed since 2006) was chosen. A validation by the test CCC (Figure. 4) allows to decide on the conformity or not of each reader. The delay of the survey in 2020 and the impossibility to read footage onboard induced lack of time between the end of the survey and the deadline for stock assessment and advice. There was additionally unavailability of sufficient experienced Ifremer agents having the readers agreement because of many other oceanographic surveys. As consequence of that, the recordings were read by only one person (8 minutes counted per station, 7 taken into account for processing) apart from 10 common stations. Accordingly to recommendations of the WGNEPS, all readings will be doubled before the next year's survey. Method More details can be found in Cochran (1977), Frontier (1983). The stratified sampling plan allows to calculate a ratio estimator (noted Y) of two variables, the numbers of burrows by video track and the surface of the track: With: h= stratum [h=1,…,ns]; i= station by stratum h [i=1, …, nh]; Sh= total surface of the stratum h; sjh= surface for the station i, stratum h; xih= total number of burrows by station i in the stratum h (by adding the total recorded and validated minutes by station averaged according to the number of observers usually equal to 2)1 The variance of Y, noted V[Y], is given by: ()() ],[... 2][.][.. [.][][ ] 2 2 1 11 ihih h h ihih h h nh i ih h ns h h ns h sxCov S Y nhxVnhsV S Y nh s S YVYV −+           == ∑ ∑∑ = == with V[xih], V[sih] and Cov[xih,sih] variances and covariance of xih and sih. 1 The stratified estimator was also investigated under a sub-sampling plan (primary unit: station; secondary unit: observer*minute). It was proved that including the 2nd level increases the total variance only by 1.8-2.2% for years 20142018 (but ≈5.5% in 2019 and ≈8.6% in 2020); thus, the stratified plan is further developed on only one sampling level. 100 | ICES SCIENTIFIC REPORTS 03:36 | ICES Figure 4. Conformity test CCC. 2020’s results. Raising 1. Raising to the five historical sedimentary strata (from the former trawl survey 2006-2013). The whole area of the five historical strata was covered in 2014 although only 2/3 of the total number of stations were carried out in 2015. In the period 2016-2020, 100% of the Central Mud Bank was sampled (respectively 160, 94, 148, 116 and 117 validated stations). The 2017’s lower sampling level is explained by the coverage of a wide area exceeding the actual Central Mud Bank of the Bay of Biscay whereas the additional sampling effort outside the edge in 2018 affected the sampling level in the 2016's benchmarked area in a lesser degree. In 2019, the sampling coverage was also impacted by the weather conditions. Table 2 shows results of raising of burrow densities (/m²)² associated to their CVs by stratum for years 2014-2020. Results for 2020 show a steep decrease by -24% compared to 2019 (+18% between 2017 and 2018, +6% between 2018 and 2019). Table 2. Total number of burrows (106), densities/m² and CVs by spatial stratum and for the whole area. Years 20142020 (values not corrected by the cumulative bias factor). nb/m² total burrows CV (%) %burrows nb/m² total burrows CV (%) %burrows nb/m² total burrows CV (%) %burrows 0.442 5164.53 5.82 0.386 4501.89 8.25 0.386 4505.52 7.86 CB 0.317 802.68 15.68 15.54% 0.151 383.85 25.66 8.53% 0.258 654.41 19.84 14.52% CL 0.171 196.72 28.30 3.81% 0.306 352.28 18.57 7.83% 0.237 272.72 20.87 6.05% LI 0.354 1651.31 8.69 31.97% 0.320 1492.89 16.38 33.16% 0.283 1319.12 13.86 29.28% VS 1.656 1048.72 11.05 20.31% 0.875 553.75 30.48 12.30% 0.839 531.18 17.92 11.79% VV 0.544 1465.10 13.19 28.37% 0.639 1719.13 10.99 38.19% 0.642 1728.09 14.52 38.35% 2014 (156 stations) 2015 (96 stations) 2016 (160 stations) ICES | WGNEPS 2020 | 101 2. Raising including the rough sea bottom. From 2016 supplementary area assumed to not be trawled as occupied by rough ground was also covered (Table 3). This additional stratum concentrating a moderate fishing pressure level as illustrated by VMS data were included in the five strata considered since the former trawl survey 2006-2013. Table 3. Total number of burrows (106), densities/m² and CVs by spatial stratum and for the whole area. Years 20162020 after including rough sea bottom contained in the outline of the Central Mud Bank (16164 km² instead of 11676 km² for the five sedimentary strata sensu stricto). As for the other raising options, the number of burrows seems to have steeply declined between 2016 and 2017 (-19%) then increased by +12% and +9% respectively in 2018 and 2019. In 2020, a reduction of –17% was observed. Anyway, for any year the two more compact muddy strata (VS and VV) corresponding to less than 20% of the overall surface concentrate around 40-45% of the total number of burrows. Correction factors. Edge effect: The edge effect calculated on 2014’s data are represented by a corrective coefficient of 1.15 and it is associated to a low uncertainty (relative precision≈11%). This value is still used for 2016-2020’s data. The adoption of the HD system since 2019 suggests the necessity to update this coefficient. nb/m² total burrows CV (%) %burrows nb/m² total burrows CV (%) %burrows nb/m² total burrows CV (%) %burrows 0.303 3534.20 9.85 0.357 4172.82 8.44 0.378 4413.87 8.59 CB 0.152 384.49 20.10 10.88% 0.259 656.93 19.56 15.74% 0.259 436.35 25.39 9.89% CL 0.262 302.03 14.76 8.55% 0.517 595.61 23.64 14.27% 0.517 464.82 43.28 10.53% LI 0.210 978.48 14.75 27.69% 0.228 1064.10 13.27 25.50% 0.228 1363.72 14.34 30.90% VS 1.147 726.44 27.94 20.55% 0.841 532.43 23.30 12.76% 0.841 370.94 21.46 8.40% VV 0.425 1142.76 19.82 32.33% 0.492 1323.75 17.30 31.72% 0.492 1778.04 12.12 40.28% 2019 (116 stations) 2018 (148 stations) 2017 (94 stations) nb/m² total burrows CV (%) %burrows % surf 0.286 3343.31 10.18 CB 0.072 182.34 24.46 5.45% 21.72% CL 0.229 263.73 44.46 7.89% 9.87% LI 0.195 911.55 18.76 27.26% 39.94% VS 0.903 571.69 20.14 17.10% 5.42% VV 0.525 1414.01 16.96 42.29% 23.05% 2020 (117 stations) nb/m² total burrows CV (%) %burrows nb/m² total burrows CV (%) %burrows nb/m² total burrows CV (%) %burrows 0.320 5167.67 7.84 0.259 4181.95 9.87 0.291 4696.84 8.30 CB 0.258 654.41 19.84 12.66% 0.152 384.49 20.10 9.19% 0.259 656.93 19.56 13.99% CL 0.237 272.72 20.87 5.28% 0.262 302.03 14.76 7.22% 0.517 595.61 23.64 12.68% LI 0.283 1319.12 13.86 25.53% 0.210 978.48 14.75 23.40% 0.228 1064.10 13.27 22.66% VS 0.839 531.18 17.92 10.28% 1.147 726.44 27.94 17.37% 0.841 532.43 23.30 11.34% VV 0.642 1728.09 14.52 33.44% 0.425 1142.76 19.82 27.33% 0.492 1323.75 17.30 28.18% RO 0.148 662.15 29.61 12.81% 0.144 647.75 34.23 15.49% 0.117 524.02 31.79 11.16% 2016 (196 stations) 2018 (184 stations) 2017 (124 stations) nb/m² total burrows CV (%) %burrows nb/m² total burrows CV (%) %burrows % surf 0.316 5100.64 8.34 0.263 4247.08 12.74 -16.73% CB 0.172 436.35 25.39 8.55% 0.072 182.34 24.46 4.29% -58.21% 15.69% CL 0.403 464.82 43.28 9.11% 0.229 263.73 44.46 6.21% -43.26% 7.13% LI 0.292 1363.72 14.34 26.74% 0.195 911.55 18.76 21.46% -33.16% 28.85% VS 0.586 370.94 21.46 7.27% 0.903 571.69 20.14 13.46% 54.12% 3.92% VV 0.661 1778.04 12.12 34.86% 0.525 1414.01 16.96 33.29% -20.47% 16.65% RO 0.153 686.77 28.17 13.46% 0.201 903.76 46.57 21.28% 31.60% 27.76% 2020 (134 stations) 2019 (145 stations) 102 | ICES SCIENTIFIC REPORTS 03:36 | ICES Detection: a very good visibility characterized footage during the four UWTV years (e.g. in 2014, 946 minutes of reading on 1095, i.e. 86%, have very high quality of image) and a correction factor of 0.94 is retained. Species identification: The coexistence between Norway lobsters (Nephrops norvegicus) and squat lobsters (Munida sp.) and a certain capacity of the second species to colonise Nephrops burrows affect the correction factor of the "species identification". The interaction Nephrops and Munida is not relevant to many other Nephrops stocks already routinely video surveyed either because of the depth (Iberic stocks, bank of Porcupine) or due to the latitude as Munida is more southerly spread than Nephrops in the NW Atlantic waters. Video on years 2014-2020 allows to investigate the basic differences of dial activities for both species: Nephrops is active during a more restrictive time interval within a day whereas the activity of Munida is more widely spread on 24 h. The intuitively expected case of Nephrops activity around dawn and dusk was observed on data collected in September 2014, May 2016 and May 2017, although 2015’s data presented a different profile (see WGBIE 2017) and 2018's data showed no relevant pattern to be fitted. Munida showed wider profile of emergence with two close study cases of minimized activity near dawn and dusk (September 2014, May 2017); at the opposite, 2016's and 2018's observations do not correspond to the same scheme whereas 2015's data are not relevant. The last two years reveal similar pattern for both crustaceans modelled according to Gauss curves (Figure. 6 and 7). The observed active individuals fluctuated a lot: for Nephrops in the range 235-1369 (minimum in 2019, maximum in 2016) and for Munida in the range 151-2653 (minimum in 2018, maximum in 2014). It is noticeable that Munida was systematically represented by higher numbers apart from the three last years' surveys. Combining those results on footage and trawling experimental catches (for years 2014 and 2015) on both species allow to propose species identification coefficient of 1.05, 1.10 or 1.15. The third value was retained by 2016’s WKNEP benchmark for the stock. The combination of the correction factors above provides a cumulative bias coefficient of 1.24. The advice 2021 for the stock was performed on the basis of the 2020’s UWTV survey results corrected by the cumulative bias coefficient combined with the harvest rate for the year 2019 (LFDs and mean weights for landings and discards, discard survival rate fixed at 50% since the WKNephrops 2019 which revised the historical value of 30%) (Table 4). Table 4. Catch option table for the FU23-24 Nephrops including information from the 2020’s UWTV survey. Variable Value Notes Stock abundance (2021) 3425.061 Number of individuals (millions); UWTV Survey 2020 Mean weight in projected landings 23.82 Average 2017–2019; in grammes Mean weight in projected discards 10.99 Average 2017–2019; in grammes Projected discards 53.6 Average 2017–2019; percentage by number Discard survival * 50 Percentage by number Dead projected discards 37.4 Average 2017–2019; percentage by number ICES | WGNEPS 2020 | 109 Italy and Croatia GSA 17 and 18: Adriatic Sea Martinelli M., Medvešek D., Chiarini M., Angelini S., Belardinelli A., Caccamo G., Cacciamani R., Calì F., Canduci G., Croci C., Domenichetti F., Giuliani G., Grilli F., Guicciardi S., Penna P., Scarpini P., Santojanni A., Zacchetti L., Cvitanić R., Isajlovic I., Vrgoc N., Milone N., Arneri E. The Pomo (or Jabuka) Pits area is one of the main fishing ground for Norway Lobster Nephrops norvegicus and European hake Merluccius merluccius within the GFCM Geographical Sub Areas 17 (Northern and Central Adriatic Sea) and it is shared by the Italian and the Croatian fleets (Russo et al. 2018). Furthermore, this represents a well-known nursery area for M. merluccius (Angelini et al. 2016) and hosts a distinct population of N. norvegicus, characterized by smallsized mature individuals (Froglia and Gramitto 1982; Vrgoć et al. 2004; Colella et al. 2018, Angelini et al. 2020). Due to a decline in landing of both species for the Adriatic Sea (FAO-GFCM 2019), since 2015 the Italian and the Croatian governments implemented some protection measures in that area. Eventually in 2018, the GFCM established a Fishery Restricted Area (FRA; GFCM 2017). Although not covered by DCF, following early trials (Froglia et al. 1997; Morello et al. 2007) a spring UWTV survey is carried out since 2009 (except for 2011 and 2018) in the Pomo Pits by CNR-IRBIM of Ancona (formerly part of CNR-ISMAR), jointly with IOF of Split and under the auspices of the FAO – AdriaMed regional project (Scientific Cooperation to Support Responsible Fisheries in the Adriatic Sea; Martinelli et al. 2013). The latest equipment improvements are dated 2016 (Martinelli et al. 2016). The collected footage is usually analyzed later in the institute labs by a team composed by Italian and Croatian scientists, complying as much as possible with ICES standards (ICES 2017, ICES 2019) and applying some specific thresholds (e.g. on speed and turbidity) settled for the Adriatic footage (Martinelli et al. 2016; Martinelli et al. 2017a). Aiming to produce an index of abundance to use as tuning for a length-based integrated stock assessment model (CASAL; Bull et al. 2005), the Adriatic team is constantly working to address the uncertainties still linked to the application of this method within the study area; therefore in 2019 a complete revision of the time series 2012-2017 was carried out and presented at WGNEPS 2019 (ICES 2020). Unfortunately, due to the COVID-19 pandemic, the survey was not conducted in spring 2020. Usually during the UWTV Adriatic surveys, along with CTD casts, trawl hauls are also carried out by means of an experimental net, at sunrise and sunset, in order to obtain demographic and biological data on N. norvegicus and other species relevant to the area (Martinelli et al. 2017a). Since 2015, an additional autumn trawl survey (using the same net and CTD) is carried out in the western side of the Pomo Pits area; the latter is planned in the framework of an agreement between the Italian Ministry of Agriculture and Forestry (MIPAAF) and CNR-IRBIM and aims to evaluate the effects of the management measures enforced in the area (Figure xx; Martinelli et al. 2017b). Indeed, in 2018, these Adriatic surveys were included in the monitoring plan adopted by the Scientific Advisory Committee on Fisheries (GFCM-SAC) to monitor the Pomo FRA effectiveness. The obtained catch per unit of effort (CPUEs) datasets were analysed in order to statistically detect possible effects of the Pomo FRA implementation on the main target species, in terms of biomass and distribution (Martinelli et al. 2019). The preliminary results were presented to the AdriaMed Working Group on Demersal Fisheries Resources (18 May 2020) and reported to MIPAAF (Martinelli et al. 2020), before upcoming submission to GFCM and publication. Furthermore, aiming to use these CPUEs time series as input for stock assessment models, standardization exercises through generalized additive models (GAM) are in progress. 110 | ICES SCIENTIFIC REPORTS 03:36 | ICES Figure 1. Pomo (Jabuka) Pits area within GSA 17 with indication of bathymetry (EMODNET bathymetry in meters), location of the UWTV stations (points) carried out during the spring surveys and FRA zones (zone A closed to any professional fishing activity, zones B and C subject to fisheries limitations; GFCM 2017). References Angelini S., Hillary R., Morello E.B., Plagányi É.E., Martinelli M., Manfredi C., Isajlović I., Santojanni A. 2016. An Ecosystem Model of Intermediate Complexity to test management options for fisheries: A case study. Ecological Modelling 319: 218-232. Angelini S., Martinelli M., Santojanni A., Colella S. 2020. Biological evidence of the presence of different subpopulations of Norway lobster (Nephrops norvegicus) in the Adriatic Sea (Central Mediterranean Sea). Fisheries Research 221: 105365. Bull, B., Francis, R., Dunn, A., McKenzie, A., Gilbert, D., Smith, M., Bian, R., et al. 2005. CASAL (C++ algorithmic stock assessment laboratory): CASAL User Manual v2. Colella S., Angelini S., Martinelli M., Santojanni A. 2018. Observations on the reproductive biology of Norway lobster from two different areas of the Adriatic Sea. ISSN 1123-4245 Biologia Marina Mediterranea 25 (1):241-242. FAO-GFCM. Fishery and Aquaculture Statistics. GFCM capture production 1970-2017 (FishstatJ). In: FAO Fisheries and Aquaculture Department [online]. Rome. Updated 2019. www.fao.org/fishery/statistics/software/fishstatj/en Froglia, C., Gramitto, M.E., 1982. Alcuni aspetti biologici e gestionali della pesca a strascico sui fondi a scampi dell’Adriatico Centrale. In: Atti del Convegno delle Unità Operative afferenti ai sottoprogetti Risorse Biologiche e Inquinamento Marino, Roma, 10–11 Novembre 1981. Consiglio Nazionale delle Ricerche, Rome, pp. 295–309. Froglia, C., R. J. Atkinson, I. Tuck and E. Arneri. 1997. Underwater television survey, a tool to estimate Nephrops stock biomass on the Adriatic trawling grounds. In: Tisucu Godina Prvoga Spomena Ribarstva u Hrvata (ed. B. Finka), pp. 657–667. Hrvatska Akademija Znanosti I Umjetnosti, Zagreb. ICES | WGNEPS 2020 | 111 ICES. 2017. Report of the Workshop on Nephrops burrow counting. WKNEPS 2016 Report 9-11 November 2016. Reykjavík, Iceland. ICES CM 2016/SSGIEOM:34. 62 pp. ICES. 2019. Report of the Working Group on Nephrops Surveys (WGNEPS). 6-8 November. Lorient, France. ICES CM 2018/EOSG:18. 226 pp. ICES. 2020. Working Group on Nephrops Surveys (WGNEPS; outputs from 2019). ICES Scientific Reports. 2:16. 85 pp. http://doi.org/10.17895/ices.pub.5968 GFCM 2017. Recommendation GFCM/41/2017/3 on the establishment of a fisheries restricted area in the Jabuka/Pomo Pit in the Adriatic Sea. Martinelli M., Belardinelli A., Guicciardi S., Penna P., Domenichetti F., Croci C., Angelini S., Medvesek D., Scarpini P., Micucci D., Giuliani G., Grilli F., Isajlović I., Vrgoč N., Santojanni A. 2016. SP2_LI1_WP1_UO05_D01 - Rapporto della campagna 2015 (ex SP2_WP1_AZ3_UO05_D03 - Report 3° UWTV Survey – RITMARE) - RITMARE La Ricerca ITaliana per il MARE. Martinelli, M., Morello, E. B., Isajlović, I., Belardinelli, A., Lucchetti, A., Santojanni, A., Atkinson, J. A., Vrgoč, N., Arneri, E. 2013. Towed underwater television towards the quantification of Norway lobster, squat lobsters and sea pens in the Adriatic Sea. Acta Adriatica 54(1): 3 – 12. Martinelli M., Belardinelli A., Guicciardi S., Penna P., Domenichetti F., Croci C., Angelini S., Medvesek D., Froglia C., Scarpini P., Micucci D., Isajlović I., Vrgoč N., Santojanni A. 2017a. Report of the Underwater Television survey (UWTV) activities in 2016 in Central Adriatic Sea. Document presented at the 18th Meeting of the AdriaMed Coordination Committee (Tirana, Albania, 16-17 February 2017). FAO AdriaMed: CC/18/info 12. Martinelli M., Morello E.B., Angelini S., Froglia C., Belardinelli A., Domenichetti F., Croci C., Micucci D., Scarpini P., Santojanni A. 2017b. Parte 2: Fermo biologico area di Pomo - Convenzione tra MIPAAF e CNR-ISMAR Ancona per aggiornamento dei piani di gestione delle specie demersali delle GSA: 9 10, 11, 15, 16, 17, 18, 19, fermo biologico nell'area di Pomo, valutazione della pesca dei bivalvi nella fascia costiera compresa nelle 0,3 miglia nautiche e misure gestionali ZTB - CUP J52I15003990001. Martinelli M., Angelini S., Belardinelli A., Chiarini M., Croci C., Domenichetti F., Guicciardi S., Scarpini P., Santojanni A., Zacchetti L. 2019. Report finale Modulo 6. Monitoraggio Fosse di Pomo periodo 20172018 (esteso primavera 2019) Convenzione tra MIPAAFT e CNR-ISMAR Ancona per uno studio propedeutico al rinnovo dell’affidamento della gestione della pesca dei molluschi bivalvi ai Consorzi di Gestione – CUP J53C17000540001. Martinelli M., Angelini S., Belardinelli A., Caccamo G., Cacciamani R., Calì F., Canduci G., Chiarini M., Croci C., Domenichetti F., Giuliani G., Grilli F., Guicciardi S., Penna P., Scarpini P., Santojanni A., Zacchetti L. Accordo tra MIPAAF e CNR-IRBIM Ancona in merito alla proposta progettuale relativa alle attività di monitoraggio periodico delle fosse di Pomo e all’attuazione di misure che, nel rispetto dei piani di gestione, comportino il mantenimento delle condizioni ambientali idonee alla vita e all’accrescimento dei molluschi bivalvi, ponendo in essere misure supplementari tese a proteggere le diverse fasi del ciclo biologico delle specie interessate (CUP J41F19000080001) - Parte Monitoraggio Fosse di Pomo periodo 2019·2020. Secondo interim report - Luglio 2020. Morello, E.B., C. Froglia, and R. J. A. Atkinson. 2007. Underwater television as a fishery-inde-pendent method for stock assessment of Norway lobster (Nephrops norvegicus) in the central Adriatic Sea (Italy). ICES J. Mar. Sci. 64: 1116–1123. Russo T., Elisabetta B Morello E.B., Parisi A., Scarcella G., Angelini S., Labanchi L., Martinelli M., D'Andrea L., Santojanni A., Arneri E., Cataudella S. 2018. A model combining landings and VMS data to estimate landings by fishing ground and harbor. Fisheries Research 199: 218–230. Vrgoć, N., E. Arneri, S. Jukić Peladić, S. Krstulović Šifner, P. Mannini, B. Marčeta, K. Osmani. 2004. Review of current knowledge on shared demersal stocks of the Adriatic Sea. AdriaMed Technical Documents, 12. 91 pp. 112 | ICES SCIENTIFIC REPORTS 03:36 | ICES Annex 4: List of presentations (in order of appearance) Kai Wieland, Patrik Jonsson, Mats Ulmestrand, Sven Koppetsch, Annegrete Dreyer-Hansen, Maria Jarnum, Gert Holst, Ronny Sørensen, Baldvin Thorvaldsson, Anders Wernbo and Filip Bohlin: Nephrops UWTV survey in the Skagerrak and Kattegat (FU 3&4) in 2020. 11 pp. Jennifer Doyle and Mikel Aristegui et al. : 2020 Update on Marine Institute Ireland Nephrops UWTV surveys. 24 pp. Candelaria Burgos and Yolanda Vila: IEO Developments on the UWTV survey in the Gulf of Cadiz (FU 30) 2020. 18 pp. Adrian Weetman: Marine Scotland Science 2020 UWTV surveys summary. 18 pp. Jónas Páll Jónasson, Julian Burgos, Georg Haney, Arnþór Kristjánsson, Anna Ragnheiður Grétarsdóttir, Arnar Björnsson, Auður Bjarnadóttir & Hjalti Karlsson: Development of UWTV survey in Icelandic waters. 13 pp. Martinelli M., Medvešek D., Angelini S., Belardinelli A., Caccamo G., Cacciamani R., Calì F., Canduci G., Chiarini M., Croci C., Domenichetti F., Giuliani G., Grilli F., Guicciardi S., Penna P., Scarpini P., Santojanni A., Zacchetti L., Cvitanić R., Isajlovic I., Vrgoc N., Milone N., Arneri E.: Adriatic UWTV surveys and Pomo monitoring activity. 16 pp. Mathieu Lundy: AFBI Western Irish Sea Nephrops Grounds (FU15) 2020 UWTV Survey and Trawl survey. 19 pp. Charlotte Reeve: CEFAS Survey results and assessment summary for FU 6 and FU14. 10 pp. Spyros Fifas and Jean-Philippe Vacherot: Ifremer FU23-24 Nephrops Analysis of UWTV Survey 2020 results and overview of stock status and technical operations. 19 + 14 pp. Julien Simon: Application of Neural Networks using Langolf Dataset. 28 pp. Atif Naseer: Nephrops norwegicus detection and classification from underwater videos using deep neural network. 57 pp. Jacopo Aguzzi and Bahamon Nixon et al. : Burrow emergence rhythms of Nephrops norvegicus, UWTV, surveying biases and novel technological scenarios, EMSO-Link Transnational Access (TNA) project SMARTLOBSTER 26 pp. Ivan Masmitja, S. Gomariz, J. del Rio (UPC), J. Navarro, J. Aguzzi, M. Vigo, N. Bahamón, J. A. García, J. B. Company (ICM-CSIC): Acoustic tracking by networked moored hydrophones in Nephrops no take zones Deep NW Mediterranean. 29 pp. Maria Vigo, Joan Navarro, José A. García, Jacopo Aguzzi, Guiomar Rotllan, Nixon Bahamón and Joan B. Company: ROV as a non-invasive tool for the assessment of an overexploited protected area in the Northwestern Mediterranean Sea. 19 pp. Niall Fallon: Nephrops norvegicus abundance estimates bias and uncertainly FU 11 and FU 12. 23 pp. ICES | WGNEPS 2020 | 113 Mikel Aristegui and Jennifer Doyle: MI High Definition reference set compilation 2020. 6 pp. Mikel Aristegui: ToR c – Github WGNEPS. 3 pp. Adrian Weetman: ToR b – Developing international database status and update and Nephrops FU and survey areas shapefiles. 1 pp. 114 | ICES SCIENTIFIC REPORTS 03:36 | ICES Annex 5: Action list Action Addressed to Action latest before 1 Provide outstanding parts of the WG report All WG members At latest 17/122020 2 Review and comment on completed draft report All WG members At latest 15/12021 3 Conduct efforts to obtain burrow system size measurements All WG members 01/11/2021 4 Have meeting with ICES database centre One member per institute At latest 18/122020 5 Follow up on meeting with ICES database centre on the UWTV database One member per institute 17/11/2021 6 Check FU’s shapefiles and provide feedback to Rui Catarino at ICES All WG member asap 7 Contact end user for UWTV datasets feedback Patrik, Ade 01/11/2021 8 Submit final version of SISP to TIMES committee and resolution Jennifer asap 9 Update/Upload R scripts for UWTV survey data analysis and quality control on github All WG members Ongoing 10 Develop reference sets for other FU’s and report to WGNEPS National Institutes Ongoing 11 Hold meeting with researchers to decide on best annotation tools to develop training sets One member per institute /Jennifer asap 12 Full review of 2020 FU23-24 UWTV survey data so that at least 2 readers counts per station in line with UWTV TIMES publication (in progress). Ifremer Before WGBIE 2021