PREVENTION OF BY-CATCU OF SMALL CETACEANS IN PELAGIC TRAWLS BY TECHNICAL MEANS (PROJECT CETASEL)

Transcription

1 .',. This paper not to be cited without prior reference to the author INTERNATIONAL COUNCIL FOR TUE EXPLORATION OF TUE SEA ICES CM 1997/Q:20 Fish Capture Committee PREVENTION OF BY-CATCU OF SMALL CETACEANS IN PELAGIC TRAWLS BY TECHNICAL MEANS (PROJECT CETASEL) by D. de Haan, P. Y. Dremiere, B. Woodward, R. A. KasteIein, M. Amundin and K. Hansen D. de Haan: DLO-Rijksinstituut voor Visserijonderzoek (NL), P. Y. Dremiere: Institut Fran~ais de Recherche pour l' Exploitation de la Mer (F), B. Woodward: Underwater Acoustic Group Loughborough University (UK), R. A. KasteIein: Harderwijk Marine Mammal Park (NL), M. Amundin: KoImärdens Djur and Natur Park (SWE), K. Hansen: Danish Institute for Fisheries Technology and Aquaculture (DK)

2 - I I, Abstract The European Commission (DG XIV) AIR III project CETASEL (CETAcean SELectivity, contract AIR III-CT ) aims at developing technical means acceptable to the fishing industry to reduce the by-catch of small cetaceans in pelagic trawls. The research complements studies of by-catch in pelagic trawls in two other recent European projects. To investigate the causes of by-catch, non-intrusive acoustic detection methods have been developed to monitor how the animals behave when confronted with obstacles. Sea trials have been conducted to study the origin of the cetacean by-catch. A tracking system has been developed to monitor cetacean positions relative to the trawl. An exciuder prototype was modeled in a flume tank and scale models were presented to captive dolphins. A fuh scale excluder was tested in a pelagic trawl to observe the behaviour of cetaceans and the effect of the exciuder on fishing efficiency. Studies of the first enciosure trial showed one of the aversive sounds forced a harbour porpoise to pass through a panel with meshes of 3.60 m. Within the CETASEL time scale the main goal of the project could not be fully reached. The development and the adaptation of the dolphin tracking system to the pelagic fishing operation required more effort than was foreseen. Analysis from other acoustic sensors and observer data can provide new information on depth and distance ofcetaceans relative to the trawl. Key words: acoustic, by-catch, cetacean, deterrents, pelagic trawl, tracking.

3 Table of contents 1. Introduction Methods and individual results Inventory of by-catch data Research on the behaviour of a harbour porpoise (Phocoena phocoena) in an enclosure Design ofa dolphin tracking system Research on the behaviour of cetaceans in the vicinity of a pelagic trawl 4 3. Modification ofgear design The effects of an excluder panel to the behaviour of captive bottlenose dolphins FuII scale exciuder panei Sea trials on the efficiency ofthe separation ofcetaceans 7 5. Discussion 7 6. Acknowledgements 7 7. References 8 8. APPENDIX 9

4 t 1. INTRODUCTION Reducing any adverse environmental impact on fishing operations has become an important issue in fisheries management. This is especially thc casc for the incidental catch of marine mammals. The ASCOBANS agreement, signed in 1994 by European countries, addressed environmental issues rclated to marine mammal populations in thc North Seas and the Baltic. Interactions between marine mammals and fisheries are complex. The debate is often emotive, and anecdotal information suggests that dolphins may be caught in large numbers in pelagic trawls during a single tow. It is also possible that high numbers of strandings in the Northeast Atlantic may be attributable to pelagic trawling [Colet & Mison, 1995]. Thc mackerel pelagic fishery was identified as the most Iikely cause of a mass mortality of common dolphins in thc Southwcst of England in and a four nation E.C. project BIOECO [Morizur et al. 1996] was started to investigate by-catches in thc Northeast Atlantic seasonal pelagic fisheries In 1994 a team consisting of fishing gear technologists, zoologists and a sonar research group started project CETASEL, co-funded by the European Commission, to study possiblc technical solutions to avoid by-catches of small cetaceans in pelagic gear. The study consisted of six main tasks: 1. Tocarryoutan inventoryofpelagic trawl by-catchdata; 2. Research on cetacean sensory systems in enclosures; 3. Design and modification ofpassive cetacean tracking equipment; 4. Research on the bchaviour ofcetaceans in thc vicinity ofa pelagic fishing gear; 5. Modification ofgear design; 6. Seatrails on thc efficiency ofthc separation ofcetaceans. Therc are indications that thc dolphins in thcir natural habitat are awarc of thc prcsencc of towed fishing gear, and probably cxploit it for catching fish. In somc situations, howcvcr, thcy obviously makc mistakcs, probably whilc disturbed duc to unfamiliar circumstances or a reduced use of sensory information. By studying thc underwater behaviour of these animals in thc vicinity of a commercial pelagic trawl, and studying the way dolphins and porpoiscs in captivity react to simulated net confrontations, it may bc possiblc to identify thc sensory capabilitics of thc animals. Thc aim of CETASEL is to guide thc dcvelopment of any nccessary technical adjustments to thc trawl, such as an excluder panel or acoustic deterrents. 2. METHODS AND INDIVIDUAL RESULTS The six main tasks of CETASEL were performed by the following methods: 2.1. Inventory of by-catch data By-catch data have been analysed from previous projects primarily BIO-ECO [Morizur et al. 1996], MAMDIS [Couperus,1996] and from a compilation of non fish by-catches in New Zealand [Baird, 1994 ]. Ncw data havc also been analysed throughout thc duration ofthe projcct. Thc BIO-ECO and MAMDIS projects show that over aperiod of six years Atlantic whitcsided dolphin (LagellorhYc/llls aclltlls) dominate the by-catch (49%) and that by-catch mainly occurrcd Southwest of Ireland in thc early spring. The dissection analysis showed landed Atlantic white-sided dolphins were mainly feeding on mackerel shortly beforc their death. No horse mackerel remains were found. By-~atch reports obtained from thc Dutch 1

5 0" "'." '.1''''!" ',J.:",1'\.. ",.. ' fleet, fishing to the South of Ireland showed a rather irregular pattern and numbers differing from year to year. ". : ( I~ ;,..' - :.. ':"." t.:".!"::... ", '(' Results of the by-catch of common dolphins (Delphinus delphis) in New Zealand fisheries show some striking similarities with by-catches made by European pelagic fisheries. Most of the specimens (one excepted) were by-caught at night. Generally one or two animals were by-caught, but multiple by-c'ritches did occur. The quantity of observed by-catch in the BIOECO project was too lo\v to ailow a confident discrimination of the. factors leading to by-catch of marine mammai. A comparison was carried out on the observed by-catch rates reported by the U.S. govemment in the Northwest Atlantic [\Varing et al., 1990]. These reports inciudc a total of 538 by-catches and target the foilowing as important factors influencing by-catch:.:.. Target species of the fishery; prevalence of maritmals; susceptibility of species to entrapment; towing duration; hauling level in water column; light level; net opening size; haul back speed; gear design Research on the behaviour of a harbour porpoise (Phocoena phocoena) in an enclosure. The research involved two experiments, a net panel study to determine the minimum mesh size a small cetacean would pass through voluntarily, and tests on the effectiveness of different acoustic sources as deterrents. The research was conducted in June 1995 in a sea enciosure at Neeltje Jans using a 3-year old harbour porpoise. The experiments were carried out in a rectangular floating net pen (34 m x 19.5 m; 3.2 m deep at the sides and 4.7 m deep in the centre). Meshes of stretched size 1.8, 3.6,4.8, 7.2, and 10 m were used to simulate the walls of the front part of a pelagic trawl (Figllre 1). The rigging of each net panel divided the basin into two almost equal parts and the net panels were tested in order of increasing mesh size. Each session consisted of a baseline period of 20 minutes, in which the animal's normal reference behaviour was recorded, directly followed by a test period of the same period once the net panel was set. To determine the aversive effects of sounds, nine types of sound were used, which inciuded new sounds as weil as those commonly found around trawls. Those sounds were: Frequency sweeps, tones and clicks transmitted through an array of four hydrophones, and two commercially available single-chip sound sources producing a wide band FM sweep signal. Deterrents proven earlier to reduce by-catch of. larger whales [Lien et al., 1992.] and to have significant effects on juvenile porpoises [Kastelcin et al., 1995]. Thc two devices, initially thought to be identical, were found to have differing effects on the cetaceans. On analysis of the frequency spectrum it was found that this might have been caused by one unit having a distorted sinusoidal waveform with significant harmonic energy above 40 khz, with a particularly strang component at 17.5 khz. The second alarm was 4 db stronger and produced a pure sinusoidal frequency with little evidence ofharmonics but had a lesser effecl. A Tri-tech scanning sonar survey produced various numbers of 325 khz pulses depending on the settings of the sector to be scanned. Clicks were generated by Scanmar distance sensors, which are normally used as apart of a cableless system to obtain information on the geometry of trawls. The master sensor emits a 110kHz signal to the slave, when received by the slave this unit responds with the same frequency 10 db lower. A coded telemetry up-date signal (42 khz) is send every 12 seconds from the master. A video camera, mounted 5 m above the pool, covered the entire. pen to record the positions of surfacings. In addition underwater cameras and hydrophones were used to monitor the porpoise's behaviour around the barriers (Figllre 2)..,.. 2

6 It was found that the porpoise would not pass through net panels with 1.8 m and 3.6 m meshes voluntarily, however the animal passed through the panel with 4.8 m meshes being very hesitant at first, but showing more confidence with time. The animal had no problem passing through the 7.2 m meshes, but surprisingly refused to pass through a p10 m mesh [Kastelein er al, 1997]. The harbour porpoise reactions differed greatly to the various sounds presented. \Vhen sounds were activated, the swimming speed, the surfacing rate and the respiration rate increased slightly and the swimming depth decreased. All these parameters quickly returned to their baseline levels during the recovery period and no habituation to the sounds were observed in any of the test sessions. The sweep sounds produced by the four hydrophone array, in the range of 17.5 khz to';140 khz were found to be the most effective in altering the behaviour ofthe animal in the floating pen. To complete both the net panel and acoustic experiments, a combined experiment was conducted to investigate a sweep sound with a centre frequency of 70 khz could scare the harbour porpoise through a net panel with a relatively small mesh size (3.60 m). After the activation of the sound, the animal passed through within 10 seconds. She did not breath after the sound was switched on until she passed the net panel. Just before passing the net panel the porpoise reduced speed and then accelerated strongly when passing through. She returned to the North side of the basin when the sound was switched off, passing through at slow speed [Kastelein er al, 1997]. ' 2.3. Design of a dolphin tracking system The need to detcct the presence and position of dolphins swimming in and around towed pelagic fishing gear is important if the underlying causes of cetacean mortality are to be understood. The acoustic approach utilises the echolocation 'clicks', generated by the dolphins whilst they are foraging within the net, to indicate their spatial position. Dolphin signals arriving at a sparse array of hydrophones (minimum 4) are detected and, if the spatial configuration ofthe hydrophones is known, the arrival times of these signals can be used to compute the position of the source. Since the data rates required to extract timing information from individual clicks are very high the first major problem concerns the transfer of this data to the ship for processing. This approach suggests that four wideband channels of data might require an uncompressed digital serial transmission clocking at frequencies as high as 40 MHz. This is weil above the capabilities of the netnetsonde coaxial cable, but could be accommodated using an upgraded optical fibre link. The use of aspare optical fibre, available in the ROV umbilical, was initially exploited, because the hydrophone positions could be related to the position ofthe ROV. To study the feasibility of tracing cetaceans the performance of a prototype threedimensional tracking algorithm was investigated using computer simulation techniques [Coggrave, 1994]. After the first trial at sea in March 1995 the methodology of tracking cetaceans had to be re-assessed, because of operationallimitations ofthe ROV as carrier platform of the underwater hydrophone array, such ~s: 1- ROV operation was limited to daylight and sea state 6, and by-catch figures indicate research at night has to considered; 2- the optimum dimensions ofthe hydrophone array exceeded those ofthe ROV frame; 3- the towing ofthe hydrophone array at the backofthe ROV could he a risky, unstable operation;, 4- the self-noise ofthe ROV (manoeuvring using hydraulic power) was deemed too high to allow adequate detection. After the first two sea trials, an alternative attachment was developed, which utilised the head rope and top panel of the actual trawl as the stahle towing position for a new 3

7 hydrophone array. The only remaining method of communication was then via the ship's spare netsonde coaxial cable with a consequent severe reduction in signal bandwidth. The new system was configured from a flat rectangular,.self calibrating array of four hydrophones. This geometry can derive target positions in three dimensions, but suffers from a 'mirror image' ambiguity in the third dimension (depth). The sensing apparatus for the third sea trial comprises two oil-filled hydrophone "streamers" attached to the head rope with sufficient separation between them to form a flat sparse tracking array. To remove the depth ambiguity a fifth hydrophone was mounted out of the plane of these streamers and this was attached to the netsonde'cable (Figllre 3). The new design incorporates an acoustic monitoring system to permit all cetacean vocalisations to be observed for behavioural analysis and to assess the self-noise generated by the fishing gear and vessel. In the new electronic design the timing data, derived from each of the streamer hydrophones, is processed underwater within a small pressure housing, which is also attached to the fishing gear. The timing information extracted from the five sensing positions is encoded on sub-carriers and transmitted to the surface, where a computer system onboard the vessel is used to process the information into a spatial position relative to the array and trawl. The cable bandwidth limitations of this system have restricted the definition of the system such that in certain conditions only vectors can bc obtained to the cetacean Research on the behaviour of cetaceans in the vicinity of a pelagic trawl. Experiments at sea were conducted onboard RV Tridens to examinc the origin of bycatches and to test possible tools to reduce them. Research areas were mainly on the Northeast Atlantic at positions along the continental shelf edge off the Southem coast of Ireland, the Bay of Biscay and the Spanish shelf edge. Fishing experiments were carried out using a 4200 meshes and occasionally a 5600 meshes pelagic trawl. To minimise the by-catch of cetaceans during the operational tests of the dolphin tracking system and to respect the sanctuary of the 200 miles lrish zone, fishing with an open trawl was decided. This also meant operations could be extended without needing to recover the trawl. This trawl truncation was first modelied in a flume tank and showed little change in the shape of the front end of the trawl if three sections wcre left behind the sharks teeth (Le. junction of large and small mesh panels). To support the acoustic analysis of dolphin vocalisations the sightings of dolphins were registered by observers on the ship's top deck. Thc meteorological conditions and ship's data were logged every fivc minutes throughout the trial using the ship's data logger. Acoustic equipment availablc throughout the trials, other than the tracking system, ineluded a hydrophone unit capable of detecting all vocalisations of small cetaceans for connection to the trawl, hydrophone attachments for the ROV and sonobuoys (SSQ-904), all of which could bc recorded on instrumentation recorders (Racal VSTORE and RDAT Sony TCD-D7) and analysed on board using task specific softwarc. To observe the performance of the trawl, sensor systems and dolphin behaviour in elosc range underwater observations were carried out using the remotely operated vehiele (ROV). This system includes a Tri-tech scanning sonar system (ST 325 V), a low light SIT camera (Osprcy OE 1323), an experimental ICCD-camera in a spherc housing and an optional hydrophonc or CCD camera. Thc ROV could be towed by thc ship and manoeuvred around thc trawl using magnus rotors~ Thc tests on all except thc last trial were based on the collection ofdata and testing of new equipment. Acoustic and visual data werc stored and logged with thc aim of further offline analysis into cetacean behaviour. The last trial included thc testing of some possiblc solutions to reduce the by-catch.. 4

8 Five sea trials were conducted during the project: 1- March 1995 trial 2- NovemberlDecember 1995 trial 3- AprillMay 1996 trial 4- October 1996 trial 5- April 1997 trial (3 wks) (2 wks) (6 wks) (3 wks) (3 wks) The equipment to monitor dolphin behaviour was tested during the first two trials. On the third trial, the modified tracking system utilising the ship's netsonde cable was tested. The survey was carried out to the Southwest of Ireland mainly along the edge ofthe continental shelf at depths between 120 and 300 m and occasionally at greater depths (\Vhittard canyon). During this cruise many dolphin observations were registered (81 observations, 79 % in the April period). The animals stayed at a distance from the ship and trawl. No underwater video recordings of cetacean behaviour near the trawl were obtained. Acoustic data ofdolphins vocalisations in the vicinity ofthe trawl were obtained using the stand-alone hydrophone unit.. The operation of the tracking system was hampered due to a variety of events, such as deployment, alignment and calibration of the sensitive underwater signal processing electronics. Several options were tried to improve the stability of the hydrophone array, but without success. As the commercial practice of discarding unwanted fish may be a factor in attracting dolphins into the vicinity of a trawl, a test with discarded mackerel was carried out during fishing. The distribution pattern of the discarded fish, which was observed using video cameras on the ROV showed that dead mackerel sink quite slowly. At a distance of700 m from the ship and a towing speed of 4.5 knots the fish appeared to be dispersed between a depth of m. ROV observations showed the experimental open trawl to be more stable than model tests predicted. Trawl profiles at the junctions ofthe three sections were measured using the ROV's Tri-Tech scanning sonar. The fourth trial was carried out in the Bay of Biscay, mainly along the French and Spanish shelf edge, where 56 cetacean encounters were registered, with individuals being sighted. On this trial the operation of the tracking system was further improved. The performance of the hydrophone array was further improved by modifications of the trawl support. Tracking data were recorded on Racal VSTORE tapes to be analysed ashore. At the end ofthe cruise a possible cause ofby-catch was registered on one of the night experiments. While towing the truncated trawl a group of bottlenose dolphins (Tllrsiops trullcatlls, occasionally 50 individuals) were observed to be foraging for small fish in the lights of the ship for several hours, apparently attracted to the lights. When only a single light was in operation the animals started foraging in this lighted area. When extinguished, the behaviour of the dolphins was observed by moonlight. The lights were off for three to four minutes and in this time the number of surfacings reduced to virtually zero. Some dolphins were observed heading for the bow. After a time the flood light was re-lit and within 30 seconds ten bottlenose dolphins started hunting again in the illuminated area. This experiment was repeated twice with the same result. This predatorprey interaction was earlier observed between common dolphins and squid (Illex illecebroslls) [Major, 1986]. During this event, the dolphins' echo-iocation dicks as weil as lower frequency vocalisations were recorded using the tracking system. Range measurements from these signals showed that the animals were dose to the hydrophone array. 5

9 3. MODIFICATION OF GEAR DESIGN 3.1. The effects of an excluder panel to the behaviour of captive bottlenose dolphins.. Model tests ofa trawl excluder device were conductedin November 1996 and February 1997 at the Kolmardens Djur and Natur park in Sweden. The experiments were carried out in the indoor main basin of the dolphinarium. The kidney-shaped basin has a sunace area of 800 m2.. Thc max. width at thc centre is "18 m ; the max. Icngth'45 m. The waterdepth was 4 m. Thc barrier panels werc positioned across the centre of the basin. Three female bottlenose dolphins were available' for the experiments. Each session consisted of a baseline - and a test period, each 20 minutes long. During thc baseline session thc animal's normal behaviour was registered as a reference to the test period. A chain of aluminium net panels was used to herd the animals to the \Vest side of the pool, wherc they werc kept separated during thc deployment of thc net panels. A test session started by pulling the separation panel asidc providing an escape opening between the pool wall and the separation panel. Tests werc caitied out under two different light conditions, bright light (44 lux) and low light (004 lux), which was the threshold level ofthe underwater camera. Rope panels (Figllre 4) were tested with differing ropc spacings and with and without passive acoustic reflectors [Mayo & Goodson, 19~2, Goodson & Mayo, 1995]. Panels with large diamond-shaped meshes were also tested simulating thc front panels of thc trawl. Finally combinations werc tested to simulate the boundary betwecn thc cxeluder and thc top or lower panel of thc trawl. The junction between thc two baitiers was midwater. To monitor the dolphin's vocalisations two hydrophones (Sonar products HS70) were positioned on the opposite side of the net panel in the predicted echolocation path of thc animal. A third hydrophone was placed in the entrance passage to the basin to collect vocalisations of the animals not ineluded in the test, so as not to confuse the analysis. Signals from thc hydrophones were recorded on a high speed Racal VSTORE recorder. Visual information of the animal's behaviour was derived from an underwater low-light camera (Osprey SIT 1323), and two CCD cameras mounted approximately 9 m above the basin, each covering complementary parts of the basin. Thc signal outputs from thc hydrophones were connected to the audio inputs of the video recorders. The preliminary results ofthe experiment showed that all net panels were passed through and no combination was found to bc a completc baitier. So far it secms thc animals did not usc their echo-location sonar to navigate through thc barriers. Even in low-light conditions their visual sense could be strong enough to rely on. The high numberofpassings might have been initiated by the release ofthe animals at the start of thc tests. As soon as thc gap between thc pool wall and thc net frame was widc cnough (approx. 45 cm), thc animals accelerated through it, passing thc barrier with small visible reactions. Experiments with both combination panels showed the animals often escaped through thc rope panel instead of taking thc path through thc largc meshes. Thc animals were more reluctant when the panel was inverted so that they had to escape downwards. Also the animal's character did affect the result. One of the animals was a natural leader and passed through the exeluder readily, using her echo-location sonar. In other experiments, two animals werc involved and onc followed thc other passively Full scale excluder panel Although model tests with bottlenose dolphins show the preference for a downwards escape, this arrangement was deemed impossible while fishing elose to thc bottom.. 6

10 Thercfore a provisional cxciudcr dcvicc was dcsigncd with thc assumption that dolphins ean bc guidcd upwards and cseape through enlarged mcshes of the top panel. As ciaustrophobia eould by a eause of thc drowning of dolphins onee thcy are inside the trawl, thc exciudcr was adaptcd to the front part ofthc trawl. A 1/25 seale modcl was tcstcd in a flume tank to assess its influcnee on thc trawl shape and water flow. Thc fuh sealc model eonsistcd of 29 ropes (10 mm nylon) attaehcd to thc bottom pancl dircet1y behind thc footropc to thc top panel (Figllre 5). This arrangement resulted in a rope distance of approx. 2.5 m. Each of the seven cxciuder gates in the top pancl was madc out of four meshes, resulting "in a diamond-shaped opening of approximately 25 m (1) x 8 m (w). ' 4. SEA TRIALS ON THE EFFICIENCY OF THE SEPARATION OF CETACEANS In the final CETASEL sea trial, thc bulbous bow of the ship was equipped with hydrophoncs such that signals provcn to give avcrsive cffccts on harbour porpoiscs in earlier tcsts could bc transmittcd. Thc 4200 mcshcs trawl was riggcd with thc cxciuder pancl and was monitorcd with thc ROV. Thc trawl was also monitored for fishing cfficicncy. During thc final trial 34 cctaccan encountcrs wcrc rcgistercd. Thc outcomc of this trial was hampered by weathcr eonditions, which forccd the operation to inshore waters South of Brittany. Dolphins stayed at somc distance from the vessel and did not approach thc bow or trawl, which left the bow acoustic experiment untried. Observations of the cxciuder dcvice showed somc tcchnical irrcgularitics, which wcrc improvcd during thc cxperiments. Somc influcncc on thc shape of the trawl was noticeable. \Vcather conditions hampered furthcr improvements or measurements on the influence of the excludcr on the eatch. 5. DISCUSSION Analysis of available by-catch data has shown that the highest by-catch risk for thc eetacean species studied in the Northeast Atlantic occurs in the early spring when trawlcrs arc fishing among the mackerel shoals. Somc eetaceans are more susceptible at half light conditions, possibly while hauling. American studies have shown that the trawl gear, fishing and light eonditions ah effect by-catch. There is, as yet, not enough evidence to assurne that eetaceans swim in and out of the side'of the trawl, although enciosure tests tend to show that this may occur in the first stages of the trawl in front of the shark teeth. Analysis of data from the acoustic systems has not yet reachcd a stage where this ean be confirmed in the wild; however, ranges to thc cetaceans show that there is a strong interest in the trawl. Tests on thc exciudcr panel are not complete enough to show any significant effect cither for or against the idea although modelling and sea tests have shown the idea to be technically possible. Enclosure tests are still being analysed but do not seem to confirm behaviour in the wild to such barriers. 6. ACKNOWLEDGEMENTS The authors express their gratitude to the Directorate General of the European Commission DGXIV far co-financing this project. They also are indebted to the captain & crcw ofrv Tridens. All other contributors arc thanked, espccially principally involved in CETASEL: A. D. Goodson, P. Connelly, Underwater Acoustic Group Loughborough University; B. van Marlen, K. Bakker, J. B. van Duyn, RIVO-DLO; N.C. Hoogendoom, Harderwijk Marine Mammal Park; C. Blomquist, Kolmarderis Djur and Natur Park.. 7

11 .. ~ ~ t '. t: '.",'.. 7. REFERENCES r. Colet, A., Mison, V Analyse ofcetacean strandings on the French coasl. Final report BIOECO/93/017, EC contract DG XIV-C-l. Morizur, Y., Tregenza, N., Heessen, H., Berrow, S., Pouvreau, S By-catch and discarding in pelagic trawl fisheries. Final report BIOECO/93/017, EC contract DG XIV-C-l.. Couperus, A. S By-catch of marine mammals and discrirds inpelagic trawl fisheries (MAMDIS). Interim report C003/96. Study in support of the Common Fisheries Policy 94/018. Baird, S. J Fisheries Fishery Assessment \Vorking Group on Nonfish Species and Fisheries Interactions. Draft manuscripl. \Varing, G. T., Gerior, P., Payne, M. P., Parry, B. L. and Nicolas, J. R Incidcntal take of marine mammals in foreign fishery activities off the Northeast United States Fishery Bulletin U.S., 88: Lien, J., Barney, \V., Todd, S. and Seton, R Effects of adding sounds to cod traps on the probability of collisions by Humback whales. In: Marine Mammal Sensory Systems (Eds. Thomas, J. A., Kasteiein, R. A. and Supin, A. Ya.), Plenum Press, New York, Kasteiein, R. A., A. D. Lien, J. and de Haan, D The effects of acoustic alarms on harbour porpoise (PllOcoena phocoena) behaviour. In: Harbour porpoises, laboratory studies to reduce by-catch (Eds. Nachtigall, P. E., Lien, J., Au, W. W. L. and Read, A. J.). De Spil Publishers \Voerdcn, The Nctherlands, Kastclein, R. A., de Haan, D., Goodson, A. D., Staal, C. and Vaughan, N The effects of various sounds on a harbour porpoise (Phocoena pllocoena). In: The biology of thc harbour porpoisc. (Eds. Read, A. J., \Viepkema, P. R., Nachtigall, P.E.). De Spil Publishcrs \Voerden, The Netherlands, Kasteiein, R. A., de Haan, D., Staal, C., Goodson, A. D The response of a harbour porpoise (Phocoena pllocoena) to nets with various mesh sizes, with and without deterring sound. In: The biology of the harbour porpoise. (Eds. Read, A. J., \Viepkema, P. R., Nachtigall, P. E.). De Spil Publishers \Voerden, The Netherlands, Coggrave, C. R. (1994). Performance analysis of 3D underwater tracking system using computer simulation techniques. Coggrave, C. R., Goodson, A. D., Lepper, P. A., Woodward, B An experimental technique for tracking dolphins in the vicinity of a trawl nel. Presented at the Silverjubilee Natinal Symposium on Acoustics, December 1995, Delhi, India. Connely P., \Voodward, B., Goodson, A. D., 'Lepper, P. A., Newborough D Remote sensing methods for cetacean interactions with pelagic trawl fishing gear. Proc. of 11 th Annual Conf. ofeuropean Cetacean Society, Stralsund, March Major, P. F. Notes on a prcdator-prey interaction between common dolphins (Delphinlls delplzis) and short-finncd squid (Illex illecebroslls) in Lydonia Submarine Canyon, Western North Atlantic Ocean. Journal of Mammalogy Vol. 67 No.4: , Mayo, R. H., Goodson, A. D Interaction behaviour between wild dolphin and a moored barrier. European Cetacean Society Conference, San Remo. Goodson, A. D., Mayo, R. H Interactions between free-ranging dolphins, Tllrsiops trll1lcatlls, and passive acoustic gill-net dcterrent devices. In: Sensory Systems of Aquatic Mammals (Eds. Kastelcin, R. A., Thomas, J. A. and Nachtigall, P. E.). De Spil Publishers, \Voerdcn. 8

12 ~ '. 8. APPENDIX Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Net panels and specifications Floating pen and equipment Attachment of the hydrophone array to the top panel of the trawl Aselection of excluder models tested in Kolmärden An experimental full scale excluder design 9

13 .,. Figure 1 Netpanels and specifications ~.<:-.,.. A ;.. net panel 1.80 ;.. >c >0 mesh size mesh ratio twine type twine diameter mxo.54m - 18:12 - Anzalori mm net panel 3.60 mesh size m x 1.08 m B mesh ratio -5.5:4 twine type - Anzalon twine diameter -5.5mm net panel 4.80 mesh size m x 1.44 m C mesh ratio -4.5:3 twine type - Anzalon twine diameter -5.5 mm net panel 7.20 D mesh size m x 2.16 m mesh ratio - 3:1.5 twine type - Anzalon twine diameter -7.5 mm e E 1 mesh size mx3 m mesh ratio - 2:1.5 twinetype - Dyneema SK 60 twine diameter -10 mm bottom seetion mesh size m x3 m mesh ratio - 2:1.5 twine type - Anzalon twine diameter -4mm F mesh size m x 0.24 m twine type - Anzalon twine diameter -3.5mm 10

14 t Figure 2 Floating pen and equipment North ljl [D Transducers [D [D net panel ~nderwater camera underwater camera )I::: Main pen C1 Audio band hydrophone / ~ ~ Feeding pen Veterinary pen II I I Ir \/ Aerial camera f----i2m 11

15 I Figure 3 Attachment 01 the hydrophone array to the top panel 01 the trawl netsonde cable headline 12

16 ,. Figure 4 A selection of excluder models tested in Kolmärden Excluder panel type I 3 I i Combination panel type 3 Profile panel type watersurface animal's I _d_i_re_ct_io_n e./xz] 2 pool tloor Inverted panel type watersurface 9 Sonar reflector type P20 Twine7 mm - TwinelOmm 2 Dimensions in meters animal's direction 2 pool tloor 13

17 .,. Figure 5 An experimental full scale excluder design escape route......\\ ~ 14