Fisheries Research 123–124 (2012) 16–20
Contents lists available at SciVerse ScienceDirect
Fisheries Research journal homepage: www.elsevier.com/locate/fishres
Development and test of a remotely operated Minisampler for discrete trawl sampling Niels Madsen a,∗ , Kurt E. Hansen b , Rikke Petri Frandsen a , Ludvig Ahm Krag a a b
DTU Aqua, National Institute of Aquatic Resources, North Sea Science Park, DK-9850 Hirtshals, Denmark SINTEF Fisheries and Aquaculture, North Sea Science Park DK-9850 Hirtshals, Denmark
a r t i c l e Keywords: Trawl Sampling Selectivity Survey Minisampler
i n f o
a b s t r a c t Systems that take discrete samples of the catch while trawling are widely used in fisheries investigations. We have developed the Minisampler, a relatively small system to be used particularly on small vessels and small trawls. The device is simple, has low weight and is reasonably cheap to make. The Minisampler consists of a tele-command deck unit, a dunking transducer, a stainless steel frame and one or two acoustical trigger-units. Each of them releases a metal bar that opens/closes a collecting bag. Thus is possible to take two or three remotely controlled discrete samples during a tow. We tested the system successfully in a flume tank and during sea trials from a commercial trawler. The Minisampler is a multipurpose tool and provides opportunity for new experimental designs to improve scientific work using demersal or pelagic trawls in marine or freshwater environments. © 2011 Elsevier B.V. All rights reserved.
1. Introduction Various systems have been developed to take discrete samples of fish or plankton in the water column. One such system is the MultiSampler system, which was particularly designed to sample from large pelagic trawls (Engås et al., 1997) and is widely used to support hydroacoustic fish stock surveys (Ona, 2003; Røstad et al., 2006; Pedersen et al., 2009), as well as macroplankton surveys (Korneliussen et al., 2009; Wiebe et al., 2010). The MultiSampler is 1.3 m wide and 2.5 m long and weighs ∼300 kg in air. It is built for use on research vessels, where the working space on deck is generally good. Recently the MultiSampler has been used in relation to studies of codend selectivity in commercial demersal trawls (Madsen et al., 2008a; Grimaldo et al., 2009). These experiments provided important knowledge about the fish escape, which was found to be substantial during the haul-back process. This escape increases the probability of unaccounted fishing mortality (Madsen et al., 2008a; Grimaldo et al., 2009) and it appears to be important to assess when testing selective devices. However, it is very difficult or impossible to handle a MultiSampler on a small vessel, including commercial vessels where the working space is limited. Furthermore, different test codends are often compared by fishing them simultaneously (e.g., in a twin trawl setup) and under these conditions, the ease of handling the sampling units is crucial.
∗ Corresponding author. Tel.: +45 3396 3200; fax: +45 3396 3260. E-mail address:
[email protected] (N. Madsen). 0165-7836/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.fishres.2011.11.016
Here we describe the development of a new and simple device that can be attached to a trawl to divide the catch into two or three fractions. The small size of the device makes it possible to use from smaller vessels. The concept may lead to new experimental designs that will support scientific work and provide valuable information about the fishing process. The system can be used for multiple purposes ranging from marine and freshwater surveys with demersal and pelagic trawls to trawl selectivity experiments conducted from commercial vessels. 2. Materials and methods We have constructed two different Minisampler devices: a dualsampler and a triple-sampler (Figs. 1–3). The dual-sampler consists of a frame with an attached collecting bag that can be released by an acoustic release unit triggered from a deck unit on a vessel. The triple-sampler has two collecting bags that can be released by two different acoustic release units. The stainless steel frame of the dual-sampler weighs ∼10 kg in air and the triple-sampler ∼14 kg. The main difference between the dual-sampler and the triplesampler is the height, 92 cm and 127 cm, respectively (Figs. 1 and 2). The main frame is made of 20-mm tubes with a wall thickness of 2 mm. Distance pieces of 10 mm attach a 10 mm tube along the outside of the main frame. This tube is used to fasten the device to the codend or cover (Fig. 3). The net of the collecting bag is attached to an upper bar and to metal rings that run on vertical sliding tubes (25 mm) located inside the main frame and to the release bar, which also runs on the sliding tubes. The acoustic releasers are attached by ropes to the upper panel of the codend or cover well in front of the
N. Madsen et al. / Fisheries Research 123–124 (2012) 16–20
17
Fig. 1. Drawing of the triple-sampler frame. Dimensions are indicated in mm and Ø indicates diameter. The height of the dual-sampler is indicated. The letters indicates: (a) net attachment bar; (b) sliding tube; (c) main frame; (d) frame for net attachment; (e) lock; and (f) sliding ring for net attachment.
frame. They are positioned with the receiver transducer pointing towards the vessel and the release latch towards the frame. It may be possible to attach the acoustic releasers in a fixed holder directly onto the frame although this may make the system more sensitive to handling. One rope from the releaser to the middle or two ropes to the sides of the release bar keeps the release bar at the top of the frame until the release mechanism is activated. When the loop rope is released, two kites made of halved float spheres (Fig. 3) and the 3.8 kg weight cause the release bar to open the collection bag and simultaneously prevent access to the former collecting bag. The kites have a 20 cm and 27 cm diameter for collecting bags 2 and 3, respectively. A 3 cm diameter hole in the centre of the kite reduces wobbling movements. A net skirt (Fig. 3) is attached to netting from about 0.5 m in front of the frame descending down to the frame. This ensures that fish cannot pass through the small opening between the release bar and the main frame into the collecting bag before it is released. In addition the triple-sampler has a net skirt in the lower panel preventing the fish from entering the area under the lock (Fig. 1) and subsequently entering collecting bag 2 when collecting bag 3 is released. In the final version of the triple-sampler we have attached a vertical bar to the main frame for attachment of the skirt netting. The frame is mounted in a 90◦ vertical position. If it was sloping backwards from top to bottom the release bar would be dragged down more easily assisted by the water flow. But it would also have to be larger to keep the effective circumference of the netting the
Fig. 2. Picture of the dual-sampler installed in netting with the acoustic releaser on top of the netting. The white frame in the front is for attachment to wires when tested in the flume tank. No net skirts are mounted.
same and hence more difficult to handle. Simple locks at each side of the device in the lower part of the frame (Fig. 1) prevent the release bar from moving back up and reopening the codend or cover. Locks are placed 18 cm above the end of the sliding tube on the triplesampler and 5 cm above on the dual-sampler. The difference is due to the use of 10 net rings for each collecting bag and an extra release bar. The acoustic release system (model OCEANO 500) is a product available from IXSEA (www.ixsea.com). It weighs 6.5 kg in air and 2 kg in water. The depth range is 400 m. The deck equipment consists of a tele-command unit (TT701) and a dunking transducer unit (OIPET801-30). We made the first sampling collecting bag (bag 1) not releasable to reduce the size of the sampler by not using space on the sliding tubes. The first sampling bag can be opened if a flushing period is needed before starting sampling. The length of the collecting bags
18
N. Madsen et al. / Fisheries Research 123–124 (2012) 16–20
Fig. 4. The triple-sampler in the flume tank. The letters indicates: (a) acoustic releasers; (b) cod-line, inner bags; (c) Minisampler; and (d) kites.
The sides of the inner bags and of the outer bag are increased in the triple-sampler to fit the dimensions of the frame. 3. Results
Fig. 3. The dual-sampler and triple-sampler and examples on use. The numbers indicates: (1) collecting bag 1; (2) collecting bag 2; and (3) released collecting bag 2 (dual-sampler) or bag 3 (triple-sampler).
should depend on expected catch rates. We made the inner collecting bags (bags 1 and 2, Fig. 3) 4 m long for expected catches up to about 1–1.5 tonnes, whereas the outer collecting bag (bag 1, Fig. 3) is 8 m long and expected to be able to handle catches up to about 1.5–2 tonnes. The double length of the outer collecting bag makes it easier to handle and empty the outer collecting bag first. If larger fish (>∼50 cm) with good swimming performance are targeted, one should consider increasing the length of collecting bag 2 in the triple-sampler to avoid the risk that fish swimming forward in front of the frame end up in collecting bag 3. We have put a 3 m long zipper on the outer collecting bag making it possible to handle the inner collecting bags and pull them out through the outer collecting bag and empty them this way if necessary. The nominal mesh size (inside mesh opening) in the whole net was 40 mm. The outer collecting bag is made by joining two equalsized sections by a selvedge in the side (Fig. 4), which makes it simple to join he unit directly to a codend cover or a trawl that normally is build in two sections. The outer collecting bag can also be built as a four panel section which fits the rectangular form of the frame and is likely to be more stable when towed. The inner collecting bags (bags 2 and 3) are made in 4 sections to fit the form of the frame optimally. The side sections have thinner twine (1.2 mm) than all other net sections (1.8 mm) because this facilitates a fast effective opening and because less wear is expected on these sides.
The dual-sampler was tested and adjusted in the Hirtshals flume tank (Denmark) in 2007 and the triple-sampler in 2010 (Fig. 4). Both samplers were initially tested without kites connected to the release bars. Kites however improved the speed and ensured an efficient closing which took about 3 s. Two kites were attached to each side of the release bar. In the triple-sampler the ropes of the kites were guided by rings fixed to the net in order to avoid entangling of the kites. This will also keep the kites in distance to the sea bed. We have not experienced that the kites came in contact with the sea bed when using the dual-sampler on a codend cover on a demersal trawl (Fig. 3) close to the sea bed. This could nevertheless be a potential problem that should be considered. Floats neutralise the weight of the Minisampler in water. In the flume tank and during the sea trials we used six floats (1.6 kg buoyancy each) for the dual-sampler and six floats (2.7 kg buoyancy each) for the triple-sampler when tested in the flume tank. Even at very low speed (<0.5 knots) there is sufficient tension on the release ropes to keep the release bar in position at the top of the frame. The water flow was measured in the middle inside the triplesampler 70 cm in front of the frame having a simulated catch of about 100 kg (plastic bags filled with water, Table 1). The reduction in the water flow in this area is very limited and is about the same for all three sampling periods. The drag of the whole system was also measured. At a speed of 1.8 knots (max allowed speed in the flume tank) the drag was a little lower in the first sampling period (bag 1) with limited difference between the two following sampling periods. The largest difference in the drag between sampling periods is, however, very low (∼1%) compared to the drag of a smaller commercial trawl used in demersal marine fisheries. If the speed is increased to 2.5 knots the drag is expected to increase to about 230 kg. Table 1 Measurements of the water flow inside (n = 250) the triple-sampler and the drag (n = 1000) of the triple-sampler (±standard deviation).
Collecting bag 1 Collecting bag 2 Collecting bag 3
Free flow
Flow in sampler
Drag
1.85 1.85 1.85
1.80 ± 0.09. 1.83 ± 0.07 1.86 ± 0.07
112 ± 3.9 127 ± 3.7 131 ± 4.6
N. Madsen et al. / Fisheries Research 123–124 (2012) 16–20
19
Table 2 Retention length (L) of several marine species at 75%/95%/99% retention probability for six different mesh sizes estimated from codend selectivity experiments. All estimates based on assuming a constant selection factor (L50/mesh size) and a constant selection ratio (L75 − L25/L50). Mesh size 10 mm 20 mm 30 mm 40 mm 50 mm 60 mm
Herringa
Codb
Plaiceb
Norway lobsterb
Shrimpc
5.1/6.2/7.2 10.2/12.4/14.3 15.3/18.6/21.5 20.4/24.8/28.6 25.5/31.0/35.8 30.6/37.2/42.9
2.9/3.5/4.1 5.7/7.0/8.1 8.6/10.5/12.2 11.5/14.0/16.2 14.3/17.5/20.3 17.2/21.0/24.3
2.5/2.7/2.9 5.0/5.5/5.9 7.5/8.2/8.8 10.0/10.9/11.7 12.5/13.7/14.6 15.0/16.4/17.6
3.6/4.7/5.7 7.2/9.4/11.4 10.8/14.1/17.1 14.4/18.9/22.8 18.0/23.6/28.4 21.6/28.3/34.1
3.8/4.8/5.6 7.6/9.5/11.2 11.4/14.3/16.8 15.2/19.0/22.4 19.0/23.8/27.9 22.8/28.5/33.5
Mesh size refers to measurements with ICES 4 kg gauge (Fonteyne et al., 2007). Length in cm for fish and mm carapace length for Norway lobster and shrimp. a Suuronen and Millar (1992). b Frandsen et al. (2009). c Lehmann et al. (1993).
In this investigation we have only used a nominal 40 mm mesh size. The mesh size in the collecting bags can be chosen according to the desired retention of the targeted species. We have made a theoretical estimate of the retention lengths of five rather different marine species to indicate the expected 75%, 95% and 99% retention length for mesh sizes ranging from 10 to 60 mm (Table 2). The selectivity of the net will, however, depend on the degree of mesh opening that is influenced by the net design, the net material (stiffness, twine thickness, etc.) and the catch drag from the collecting bags, so Table 2 should only be considered as a general guide line. We tested the system during demersal trawl selectivity experiments in the Kattegat and Skagerrak (Madsen et al., 2008b) from a relatively small (20 m) commercial fishing vessel which had no ramp. Two dual-samplers were placed on codend covers (Madsen et al., 2001), attached to collect escaping fish from the trawl codends (Fig. 3, dual-sampler example) used simultaneously in a twin-trawl system. The systems functioned well and could be handled relatively easily. The collecting bag was released immediately before haul-back to sample fish escaping from the codend during haul-back. Out of 32 releases, only 2 were unsuccessful (i.e., the acoustical unit was not triggered). In those cases, no fish were caught in the collecting bag as should be expected. Catches in collecting bag 1 was up to 742 kg and catches in collecting bag 2 was 606 kg. The transducer was lowered directly into the water from the side of the vessel. In some cases it might be necessary to a planer board or a bar to drag it out from the vessel to avoid propeller noise and ensuring the direction of the acoustic signal.
4. Discussion Our system weighs considerably less than 10% of a MultiSampler. Consequently, it is much easier to handle, and it can be used from small vessels, including commercial vessels where working space is limited. In addition the price is modest (to our knowledge, <10% of the cost of a MultiSampler system) and the system is simple to operate opening wider perspectives for use as a standard device for various monitoring tasks including use in freshwater trawl surveys. A disadvantage compared to the MultiSampler system is that there is no signal to report whether the sampler was successfully released and locked. However, based on observations in the flume tank, we adjusted the system to ensure that the release bar operated efficiently and as swiftly as possible. Towing speed will normally be higher (>2 knots) than the speed in the flume tank and the drag from the kites will then be increased. If needed, release systems that report whether the acoustic release was successfully activated are available, although at a higher price. It might also be possible to add a function that reports or observes if the bar was successfully locked. The successful use of two Minisamplers fishing simultaneously in a twin trawl rig opens new possibilities for conducting comparative selectivity experiments that will provide very important
additional information. This experimental setup might change the interpretation of the results that in the end can change management decisions because not only quantities but also quality of the escape can be assessed. The reason is that escape very late in the fishing operation might induce additional unaccounted fishing mortality due to fish suffering from decompression and other damages (Madsen et al., 2008a; Grimaldo et al., 2009). Two new sea trials with the dual-sampler are scheduled in 2010. The performance of the triple-sampler has only been assessed in the flume tank but it is seems to function well. It is intended to use for selectivity sea trials and for survey sampling in the future. Routine acoustic surveys to estimate stock size of commercially important pelagic species do not use depth-stratified sampling as a standard routine although this information can be important to support the identification of species and size distributions in specific layers. The possibility of stratified sampling using demersal survey trawls could be particularly relevant when monitoring the consequences of anthropogenic constructions like wind mills, oil rigs, pipe lines, bridges and artificial reefs where there often is a demand for the use of smaller trawls with high manoeuvrability. The Minisamplers designs as presented here should be considered as a general guideline because they can be modified in several ways for specific purposes. For example mesh size and length of the collecting bags can be adjusted to meet specific criteria for working conditions and desired catches. It can also be considered to build the frame using a synthetic plastic material like nylon that has been used for sorting grids (Madsen and Hansen, 2001) for many years. This would decrease the weight further and could provide some flexibility in the frame making it less vulnerable to damage when the net is wound on the net drum. Acknowledgements Thanks are particularly due to Mogens Andersen from SINTEF who assisted in all testing of the device and to colleagues for comments and discussions, particularly Bo Lundgren. This work was conducted as part of the SELTRA project, which was carried out with the financial support of the European Union and Danish Ministry of Food, Agriculture, and Fisheries. We appreciate the comments from the referees and editor. References Engås, A., Skeide, R., West, C.W., 1997. The “MultiSampler”: a system for remotely opening and closing multiple codends on a sampling trawl. Fish. Res. 29, 295–298. Fonteyne, R., Buglioni, G., Leonori, I., O’Neill, F.G., Fryer, R.J., 2007. Laboratory and field trials of OMEGA, a new objective mesh gauge. Fish. Res. 85, 197–201. Frandsen, R.P., Holst, R., Madsen, N., 2009. Evaluation of three levels of selective devices relevant to management of the Danish Kattegat–Skagerrak Nephrops fishery. Fish. Res. 97, 243–252. Grimaldo, E., Larsen, R.B., Sistiaga, M., Madsen, N., Breen, M., 2009. Selectivity and escape percentages during three phases of the towing process for codends fitted with different selection systems. Fish. Res. 95, 198–205.
20
N. Madsen et al. / Fisheries Research 123–124 (2012) 16–20
Korneliussen, R.J., Heggelund, Y., Eliassen, I.K., Øye, O.K., Knutsen, T., Dalen, J., 2009. Combining multibeam-sonar and multifrequency-echosounder data: examples of the analysis and imaging of large euphausiid schools. ICES J. Mar. Sci. 66, 991–997. Lehmann, K., Valdemarsen, J.W., Riget, F., 1993. Selectivity in shrimp trawl codends tested in a fishery in Greenland. ICES Mar. Sci. Symp. 196, 80–85. Madsen, N., Hansen, K.E., 2001. Danish experiments with a grid system tested in the North Sea shrimp fishery. Fish. Res. 52, 203–216. Madsen, N., Hansen, K.E., Moth-Poulsen, T., 2001. The kite cover: a new concept for covered codend selectivity studies. Fish. Res. 49, 219–226. Madsen, N., Skeide, R., Breen, M., Krag, L.A., Huse, I., Soldal, A.V., 2008a. Selectivity in a trawl codend during haul-back operation – an overlooked phenomenon. Fish. Res. 91, 168–174.
Madsen, N., Frandsen, R.P., Krag, L.A., Herrmann, B., Holst, R., Lundgren, B., 2008b. Development of selective trawls to Danish fisheries – SELTRA. Report for the Danish Ministry of Food. Agric. Fish., Available at: www.orbit.dtu.dk (in Danish). Ona, E., 2003. An expanded target-strength relationship for herring. ICES J. Mar. Sci. 60, 493–499. Pedersen, G., Handegard, N.O., Ona, E., 2009. Lateral-aspect, target-strength measurements of in situ herring (Clupea harengus). ICES J. Mar. Sci. 66, 1191–1196. Røstad, A., Kaartvedt, S., Klevjer, T.A., Melle, W., 2006. Fish are attracted to vessels. ICES J. Mar. Sci. 63, 1431–1437. Suuronen, P., Millar, R., 1992. Size selectivity of diamond and square mesh codends in pelagic herring trawls: only small herring will notice the difference. Can. J. Fish. Aquat. 49, 2104–2117. Wiebe, P.H., Chu, D., Kaartvedt, S., Hundt, A., Melle, W., Ona, E., Batta-Lona, P., 2010. The acoustic properties of Salpa thompsoni. ICES J. Mar. Sci. 67, 583–593.