CRIMAC cruise report 2024

CRIMAC is a centre of research-based innovation funded by the research council of Norway through their program for research-based innovation (SFI). Sustainable, healthy food production and clean energy production for a growing population are important global goals, and CRIMAC will contribute to these by obtaining accurate underwater observations of gas, fish, nekton and other targets. The data will be used in conjunction with CRIMAC data from other surveys to build a reference data set for optical and acoustic target classification. The classification libraries will be used for developing methods and products toward the fishing industry and marine science.

1.1 - Survey objectives

The first task was to mount, test and calibrate the new 18 kHz BB transducer on RV G. O. Sars in addition to the other echosounders frequencies. This task was conducted between November 6^th – 7^th .

The first leg was conducted between November 8^th (Bergen) and 11^th (Bergen) in the Osterfjorden area north of Bergen ( Figure 1 ). The main objective was to test the performance of the new 18 kHz BB transducer on mesopelagic fish, with the objective to discriminate mesopelagic fish from other scatterers and to estimate the length using the resonance frequency. We also tested the Deep Vision system with the krill trawl and used that for validating the targets.

The second leg was conducted between November 11^th (Bergen) and 13^th (Bergen) approximately 25 nmi west of Fitjar ( Figure 2 ). The main objective was to test a suppression sheet to reduce the sediment plume and allow optical system to work in bottom trawl.

1.2 - Timeline

1.3 - Vessel details

The cruise was conducted with RV G. O. Sars ( Figure 4 ) operated by the Institute of Marine Research.

RV G.O. Sars is 77.5 m length overall, has a m aximum speed of 17 knots and a c rew of 15 in addition to accommodation for 30 scientific crew members including instrument technicians. The vessel is equipped with Kongsberg Maritime EK80 scientific broadband echosounders (operating at 18, 38, 70, 120, 200, and 333 kHz centre frequency) and a range of other sensors (sonars, ADCPs). The vessel is equipped to deploy a wide range of additional equipment (e.g. probes, towed vehicles, pelagic and demersal trawls). More information about the vessel can be found online ( https://www.hi.no/resources/brosjyre-g.o.sars.pdf ).

1.4 - Cruise participants

RV G. O. Sars is equipped with six drop-keel mounted echosounders (Simrad EK80) capable of continuous wave (CW)/narrowband or frequency modulated (FM)/broadband pulse generation. These have nominal frequencies at 18, 38, 70, 120, 200, and 333 kHz. A new broadband-capable 18 kHz transducer (Simrad ES18-11 mk2) was installed on ship drop-keel at the beginning of the survey.

The Simrad EC150-3C ADCP / echosounder is also installed on the ship drop-keel and is capable of CW and FM pulse generation both when operated as an ADCP and as a scientific fisheries echosounder of rather narrow beamwidth (2.5 ° ).

TS-probe is a probe equipped with four echosounders (Simrad EK80) capable of continuous wave (CW)/narrowband or frequency modulated (FM)/broadband pulse generation (except 38 kHz). The four transducers are depth-compensated versions (ES-38D, ES70-7CD, ES120-7CD, ES200-7CD). Probe can be deployed to desired depth from the hangar of a stationary ship. Power and communication is achieved via combined fibre-optic and power cable.

2.1 - Echosounder calibration

Ship echosounders were operated with CW and FM acoustic pulses. Settings for these were chosen to fit survey objectives and to avoid undesirable effects such as acoustic “cross-talk” in broadband data. This influenced the choice of the acoustic bandwidth, power, and pulse duration settings ( Table 1 ). Standard CW pulse settings  (Korneliussen et al. , 2008)  were used but with reduced power (this is to match power setting of alternating CW / FM pulses that were used during parts of this survey). The standard IMR FM pulse settings for broadband acoustic backscatter data collection were used.

Ship drop-keel mounted echosounders and TS-Probe echosounders (2024.11.06-07, Sandviksflaket) were calibrated using standard methods  (Demer et al. , 2015)  and metallic spheres of various sizes made of tungsten carbide with 6 percent cobalt binder. The calibration sphere diameter was chosen based on the best fit for the bandwidth in question in terms of the “null” positions in the frequency response of the sphere ( Table 2 and Figure 6 )

Example calibration results are shown in ( Figure 7 ). CTD cast was performed immediately prior to the start of the calibration procedures and echosounder environment updated accordingly.

A second calibration target of a different size was used where needed to ensure calibration data across the entire bandwidth of the chosen acoustic pulse and the two calibration results merged as per EK80 software procedures for it.

Calibration target diameters used: 57.2 mm, 38.1 mm, 35 mm, 25 mm, 22 mm, and 20 mm (henceforth referred to in the format “WC57.2” indicating tungsten carbide sphere of 57.2 mm diameter). Calibration targets are traceable, and laser engraved with an ID number.

The EC150-3C ADCP is mounted on the drop keel along with fisheries echosounders and capable of operation as ADCP and as a split-beam echosounder of a rather narrow beamwidth (about 2.5 ° ) with both narrow- and broad-band acoustic pulses. It was calibrated with WC38.1.

An additional weight (400g shackle) was used to stabilize spheres of smaller size (WC38.1, WC35, WC22, and WC20) when calibrating ship echosounders. It was suspended 6 m below the calibration target by 0.30mm diameter nylon line. WC57.2 was used alone with no additional weights. No additional weights were used for TS-Probe calibration. All spheres had nylon line netting with 2 m long loop to ensure the calibration target is removed in range from the three winch-line suspension rig line and knot echoes that are present just above the calibration target.

Ship EK80 and EC150-3C echosounder calibration conditions and quality were good to excellent. Calibration results text files (*.xml) may benefit from check-up and calibration re-run from acoustic raw data files before these are used to scale fish acoustic frequency response data.

TS-Probe echosounders were calibrated at the same location and date as ship echosounders. Probe was deployed into water so that transducers were at 1.5m depth. Calibration target was suspended 6 m below the probe ( Table 2 , Table 3 ). ES38D transducer was found to be defective as evident in the calibration data.

3.1 - Objectives

The introduction of a broadband 18 kHz transducer is an important task in the CRIMAC project. The frequency band can be used for identifying the resonance frequency of mesopelagic fish.

A broadband 18 kHz transducer was mounted in the instrument keel of RV G. O. Sars and tested during the 2023001016 CRIMAC survey in November 2023, but the transducer was not accepted by IMR due to high noise levels. The objective here is accepting a new modified18-kHz broadband system. The 18 kHz system using the new ES18-11 Mk2 broadband transducer in 2023 performed worse than the original from the 2022 survey ( Figure 8 ), and it was decided to switch back to the original narrowband 18 kHz transducer. Note also that the performance of the 333-kHz channel improved greatly when the original 333-kHz transducer was replaced by a 333-kHz transducer with different grounding (sea-water grounding).

3.2 - Methods

After a modification by Kongsberg a new ES18-11 Mk2 transducer was again mounted in the instrument keel of RV G. O. Sars and tested. The modification was to remove an unused connector (called “loddeører”). The comparison had to be done in narrowband mode (CW) since the old 18-kHz transducer was narrowband.

3.3 - Preliminary results

The echograms of the narrowband and broadband 18-kHz-channels from 2022 and 2024 shows quite similar noise estimates, although the noise from the pulse-compressed channel in this case is slightly lower ( Figure 9 d vs Figure 9 h) . One advantage of broadband 18-kHz-channel (14 – 22 kHz) is the possibility to detect the resonance-peak (c.f. Figure 9 i) that may be used to estimate size of a gas-inclusion such as a swimbladder. Figure 10 shows a resonance peak together with a Deep Vision image. Figure 11 shows no noise-peaks in the 14 – 22 kHz band.

The extensive noise found in an 18-kHz-transducer in 2023 tests was not apparent in 2024. In conclusion the modification made by Kongsberg to the ES18-11 Mk2 transducer prior to the 2024001019 survey turned out to be successful.

4.1 - Objective

The objective was to assess whether the new broad-banded 18 kHz transducer could be used to separate size groups of mesopelagic fish and test whether we could identify mesopelagic fish.

4.2 - Method

The Osterfjorden area was chosen for the experiment, were known distributions of mesopelagic fish are present.

For each replication of the experiment, we did one outward, and one inwards acoustic transect combined with a trawl haul from the inner part of the fjord and outwards. The trawl was kept at three or four depths, for approximately 10-15 minutes. The depth intervals were chosen based on the distribution in the echosounder. The trawl was equipped with the Deep Vision optical system. The images from the deep vision system were stored¹ in a separate folder for each trawl haul. A machine learning model trained to detect and enumerate mesopelagic fish was used (Westergerling et al., in prep), and the predictions were stored in an xml file for each trawl haul². The LSSS system was used to annotate the layers from the echosounder that correspond to the trawling layers. The data was exported from the LSSS system for further analysis³. The TS probe stations that were originally planned were not performed due to a water leakage in the plug that connected the probe with the winch. Four CTD stations were conducted, one in the inner part, one in the outer part and two in the middle ( Figure 2 ).

4.3 - Preliminary results

A total of 10 repetitions of the protocol were performed, with both an inbound and outbound acoustic transect ( Table 4 ), an additional “detailed” transects with high ping rate covering the upper part of the water column, and a trawl hauls corresponding to the layers observed in the echosounder data ( Table 5 ).

An example (ST5) of the DV image ( Figure 12 ), and the distribution from the trawled layers combined with the frequency response from the acoustics ( Figure 13 , Figure 14 , Figure 15 ) are provided for illustration. The CTD profiles show consistent hydrographic conditions throughout the experiment ( Figure 16 ).

5.1 - Objective

A challenge with using cameras in demersal trawls is poor visibility due to sediment swirled up by the trawl ground gear. DTU Aqua in Denmark have developed a method where a sediment suppression sheet is attached behind the ground gear for improved image clarity  (Sokolova et al. , 2022) . In this cruise we tested this method on the IMR Campelen 1800 demersal survey trawl. The main objective was to optimise a setup of Campelen trawl with camera and sediment suppression sheet.  

Research Questions: 

1. How does video quality change with the inclusion of the sediment suppression sheet (SSS)?  

1. What is the impact of the SSS on trawl geometry and flow? 

1. Does the SSS impact on trawl selectivity? 

1. Where is the optimal position of in-trawl camera? 

5.2 - Method

5.2.1 - Experimental design

The experiment was carried out about 25 nm west of Fitjar ( Figure 17 ). Bottom depth along the transects was between 282m to 287m. Seabed substrate was mostly mud with a small area of sandy mud. Weather conditions were good with a light to moderate breeze from W - NW. The Campelen 1800 sampling trawl ( Figure 18 ) was used in the experiment. A sediment suppression sheet was prepared by Seilmaker Hansen in Bergen before the cruise based on the Danish design and adapted to the Campelen trawl.  Three different setups with an 8 m, 5 m and no sediment suppression sheet were carried out in a block design where the order of the setups within blocks were altered ( Table 6 ). 30-minute-long trawl hauls were made in parallel lines with about 0.5 nm between the lines ( Figure 17 ). Eight trawl lines were used. Survey design was a result of limitations by bottom type and effort required to map sea floor before trawling and reducing risk of damaging the trawl.    

5.2.2 - Rigging the sediment suppression sheet

The sediment suppression sheet (PVC) was attached in the trawl bottom panel ( Figure 18 ). The front part of the SSS was attached to the ground gear and into the bottom panel netting using plastic rings and elastic rope ( Figure 19 ). Two different sized SSS were used, 8m and 5m ( Figure 20 ).   

5.2.3 - Trawl and sediment plume monitoring

Trawl geometry was monitored with standard Scanmar trawl sensors including a trawl eye in the top panel above the ground gear and door sensors. The behaviour of the sediment suppression sheet was monitored with GoPro and Dark Vision cameras attached to various places, e.g. top panel looking down, side panel phasing the ground gear. The sediment plume and visibility were monitored with Dark Vison camera attached ahead of the tunnel and DV in the tunnel ahead of the codend.  

Current in the water column was logged with the vessel ADCP. 

The height and acoustic density of the sediment plume behind the ground gear was also measured with a Wideband Autonomous Transceiver (WBAT) with a 200 kHz split beam transducer (ES200-7C) (Kongsberg Discovery AS) ( Table 7 ). The WBAT was mounted in the centre of the top panel, forwardmost in first 42 mm panel (3.9 m behind ground gear). Pointing directly down at the sediment suppression sheet.  

5.2.4 - Catch data

Catches were landed on deck and sorted by species. Total weight of each species was recorded and all individuals length measured. Where catches were large, random subsamples of rabbit fish ( Chimaera monstrosa , havmus), greater argentine ( Argentina silus , vassild) and blue whiting ( Micromesistius poutassou , komule) were taken, weighed and scaled to total species weight. For station 521 the entire catch was subsampled on deck, with approximately 60% sorted, weighed and measured.  

Catches were combined within experimental trawl designs and comparisons made of total catch weight, with significance assessed through one-way ANOVA. Relative frequency of species groups between trawl designs were also compared, with a focus on flatfish compared to roundfish and changes in length distribution of species. Significance assessed through inclusion of length and species terms in GAM.   

5.3 - Preliminary results

5.3.1 - Acoustic density of sediment plume

There is no immediate clear signal on the WBAT data between the treatments, but when the sheet is mounted there are some strong scattering spikes and possibly a slightly lower band above the bottom data ( Figure 21 ). The plume is believed to be approximately 0.5 m at the location of the transducer.

5.3.2 - Image clarity

Qualitatively, the average height of the sediment plume was lower with increasing SSS length ( Figure 22 ; Figure 23 ), and also dissipates within the image frame in parts of tows using an 8m SSS, which is rarely observed in tows without the SSS. 

5.3.3 - Catch composition

Mean catch weight was highest for 0m SSS at 118 kg, decreasing to 111 kg with the 5m and continuing to decline to 97 kg with the 8m SSS ( Figure 24 ). ANOVA indicates non-significant difference F(2, 10) = 0.36, p = 0.706. No clear pattern in catch weights observed between transect groupings.  

By weight, for all hauls blue whiting and rabbit fish were the largest proportion of the catch. Other species observed in high proportions by weight across all gear treatments were greater argentine, beaked redfish ( Sebastes mentella, snabeluer) and European hake ( Merluccius merluccius , lysing) ( Figure 24 ). Similarly, in terms of numbers of individuals the catch predominantly consisted of blue whiting, Norway pout ( Trisopterus esmarkii , øyepål), silvery pout ( Gadiculus argenteus , sølvtorsk) and greater argentine ( Figure 25 ). The length distribution of the catch was broadly consistent across each of the three gears ( Figure 25 ), with mean lengths of 28 cm for both 0m and 5m groups, declining to 23 cm for the 8m SSS.  

5.3.4 - Conclusion

The preliminary results indicate that the sediment suppression sheet effectively reduced the sediment plume and improved visibility in the trawl cameras. There were no clear indications of effects on trawl geometry or catch efficiency from the sediment suppression sheet. However, a more detailed data analyses is required before any conclusions can be made. Furthermore, the data set may be too limited to conclude on the catch comparisons.

The data is organized in accordance with the IMR data organization procedure. In this section the placement of each data set is described as well as a short description of each individual data set. The headings are equal to the folders in the data structure.

The data from RV G. O. Sars is safely stored at IMRs secure data storage system under \cruise_data\2024\ S2024001019_PGOSARS_4174. All data placements below refers to this directory as the top level directory.

A git hub repository has been set up to store the code for analysing the data: https://github.com/CRIMAC-WP4-Machine-learning/CRIMAC-meso-18

A “readme” document containing information about positions of sensors and cameras in leg 2 is stored at: S2024001019_PGOSARS_4174\CRUISE_DOCUMENTS\OTHER_DOCUMENTS 

6.1 - ACOUSTIC DATA

The shipboard EK80 echosounder channels were calibrated prior to data acquisition and the subsequent files contain the updated calibration settings. The calibration files are located under \ACOUSTIC\EK80\EK80_CALIBRATION and S2024001019_PGOSARS_4174\PHYSICS\ADCP\EC150_KHZ for the echosounders and the ADCP, respectively.

Ship-borne unprocessed EK80 data from FF G.O. was stored in accordance with the IMR data storage structure under \ACOUSTIC\EK80\EK80_RAWDATA. There were several different strategies for collecting data, so that the content of the EK80 data files are not all the same. During leg 1 the echo sounders operated using the IMR standard settings for BB, and for leg II they were operated using standard settings for CW.

The overall organizing of data from the survey is stored in the survey file localized at \ACOUSTIC\LSSS\LSSS_FILES. A survey-file keeps track of how the directories are organized, e.g. which Work-files to use, which KORONA-files to use and which preprocessing setup to use.

Sv(f) files are exported from LSSS and placed under \ACOUSTIC\LSSS\EXPORT\BroadbandSv. Each of the data-files represents an inbound or outbound transect according to the filename in Table 4 , where the different trawled layers are coded as {MESO1, MESO2, MESO3, MESO4} as an acoustic category within the files, c.f. the table.

6.2 - BIOLOGY

6.2.1 - Trawl sensors

S2024001019_PGOSARS_4174\BIOLOGY\TRAWL_SENSORS\SCANMAR 

• RBR S2024001019_PGOSARS_4174\BIOLOGY\TRAWL_SENSORS\RBR 

• StarOddi S2024001019_PGOSARS_4174\BIOLOGY\TRAWL_SENSORS\STAR_ODDI 

6.2.2 - Deep Vision data

The deep vision data is stored under \BIOLOGY\CATCH_MEASUREMENTS\DEEP_VISION.

(leg 2: only raw images, not geometrically or colour corrected)

6.2.3 - DarkVision

S2024001019_PGOSARS_4174\BIOLOGY\CATCH_MEASUREMENTS\OTHER_MULTIMEDIA\DARK_VISION 

6.2.4 - Catch data

(preliminary) S2024001019_PGOSARS_4174\BIOLOGY\CATCH_MEASUREMENTS\OTHER_MULTIMEDIA\CATCH_PRELIMINARY 

6.3 - OBSERVATON_PLATFORMS

6.3.1 - WBAT

S2024001019_PGOSARS_4174\OBSERVATION_PLATFORMS\WBAT 

The data from the wbat is timestamped and the different treatments can be found by looking up the time interval in Table 6 . The files from the wbat have the file extension .ra_. These are converted to .raw files using the wbat mission planner. To be able to read the files into LSSS the files were reprocessed using the replay function in the EK80 software. The reprocessed files are denoted with a prefix “REPROCESSED-”. These can be imported into LSSS. 

 Demer, D. A., Berger, L., Bernasconi, M., Bethke, E., Boswell, K., Chu, D., Domokos, R., et al. 2015. Calibration of acoustic instruments. ICES Cooperative Research Report No. 326: 133 pp. http://dx.doi.org/10.25607/OBP-185.

Korneliussen, R. J., Diner, N., Ona, E., Berger, L., and Fernandes, P. G. 2008. Proposals for the collection of multifrequency acoustic data. ICES Journal of Marine Science, 65: 982–994.

Sokolova, M., O’Neill, F. G., Savina, E., and Krag, L. A. 2022. Test and development of a sediment suppressing system for catch monitoring in demersal trawls. Fisheries Research, 251: 106323.