miii FACULTEIT LANDBOUWKUNDIGE EN UNIVERSITEIT TOEGEPASTE BIOLOGISCHE WETENSCHAPPEN GENT

Academiejaar 2003 - 2004

EVALUATION OF BY-CATCH IN THE BELGIAN BROWN SHRIMP (Crangon crangon L.) FISHERY AND OF TECHNICAL MEANS TO REDUCE DISCARDING

EVALUATIE VAN DE BIJVANGST IN DE BELGISCHE VISSERIJ OP GRIJZE GARNAAL (Crangon crangon L.) EN VAN TECHNISCHE MIDDELEN OM TERUGGOOI TE VERMINDEREN

d o o r

ir. Hans POLET

Thesis submitted in fulfillment of the requirements for the degree of Doctor (Ph.D.) in Applied Biological Sciences

Proefschrift voorgedragen tot het bekomen van de graad van Doctor in de Toegepaste Biologische Wetenschappen

op gezag van:

R ecto r: Prof. dr. ir. A. DE LEENHEER

Decaan: Promotor: Prof. dr. ir. H. VAN LANGENHOVE Prof. dr. ir. R. VERSCHOORE

The author and the promotor give the authorization to consult and to copy parts of this work for personal use only. Any other use is limited by the Laws of Copyright, particularly concerning the obligation to mention the source when reproducing parts of this work.

De auteur en de promotor geven de toelating dit proefschrift voor consultatie beschikbaar te stellen en delen ervan te kopiëren voor persoonlijk gebruik. Elk ander gebruik valt onder de beperkingen van het auteursrecht, in het bijzonder met betrekking tot de verplichting uitdrukkelijk de bron te vermelden bij het aanhalen van resultaten uit dit proefschrift.

Gent, 15 december 2003

Promotor: Auteur:

Prof. dr. ir. R. Verschoore ir. Hans Polet

0-1

Table of contents

Table of contents...... 1 Dankwoord...... 3 1 Introduction...... 5 1.1 Fisheries...... 5 1.2 Area of study...... 8 1.3 Organisational and financial framework of this study...... 11 2 Problems associated with fisheries in general and Brown Shrimp fisheries in particular 13 2.1 Overfishing...... 13 2.1.1 Overcapacity ...... 13 2.1.2 Technological development ...... 14 2.2 Science and management...... 15 2.3 Monitoring and control...... 16 2.4 Environmental degradation...... 16 2.5 Discarding...... 17 2.6 The Brown Shrimp fishery...... 18 3 Objectives ...... 21 4 The shrimp trawler fleet ...... 23 4.1 Introduction...... 23 4.2 Materials and methods...... 23 4.3 Results...... 24 4.3.1 The Belgian shrimp trawler fleet ...... 24 4.3.2 The North Sea shrimp trawler fle et ...... 31 4.4 Discussion...... 34 4.5 Conclusions...... 36 5 Discarding by the shrimp trawler fleet ...... 37 5.1 Introduction...... 37 5.2 Materials and methods...... 37 5.2.1 General plan ...... 37 5.2.2 Planning o f the sea trials ...... 38 5.2.3 Sampling and analysis o f the catch ...... 39 5.2.4 Data analysis...... 40 5.3 Results...... 42 5.3.1 National data...... 42 5.3.2 North Sea data ...... 52 5.4 Discussion...... 54 5.5 Conclusions...... 57 6 Biological and economic consequences of discarding in the Brown Shrimp fishery...... 59 6.1 Introduction ...... 59 6.2 Materials and methods ...... 59 6.3 Results ...... 61 6.3.1 The biological and economic consequences o f discarding in the Crangon fisheries 61 6.3.2 The benefits arising from the introduction o f selective fishing gears ...... 62 6.3.3 The effects o f a closed season in the German fishery ...... 65 6.4 Conclusions ...... 65 7 Selectivity of the shrimp beam trawl...... 68 7.1 Introduction ...... 68 7.2 Materials and methods ...... 70 7.2.1 The fishing area, the vessel and the trawl ...... 70 7.2.2 The cod-end and the net covers ...... 72 7.2.3 Data collection and analysis ...... 73 73 Results...... 74 7.3.1 Fishing conditions ...... 74 7.3.2 Cod-end selectivity ...... 77 Table of contents 0-2 7.3.3 Whole trawl selectivity ...... 79 7.4 Discussion...... 84 7.5 Conclusions ...... 86 8 Evaluation of the Nordmore sorting grid as a selectivity improving device ...... 88 8.1 Introduction...... 88 8.2 Materials and methods...... 89 8.2.1 Fishing area, vessels and trawls ...... 89 8.2.2 The cod-end and the outlet covers...... 91 8.2.3 Data collection and analysis...... 92 8.3 Results...... 92 8.3.1 Narrative o f the sea trials...... 92 8.3.2 Cod-end selectivity...... 94 8.3.3 Sorting grid selectivity ...... 96 8.4 Discussion...... 106 8.5 Conclusions...... 109 9 Evaluation of the sieve net as a selectivity improving device ...... I l l 9.1 Introduction...... I l l 9.2 Materials and methods...... 112 9.2.1 Fishing area, vessels and trawls ...... 112 9.2.2 The cod-end and the outlet covers...... 113 9.2.3 Data collection and analysis...... 113 9.3 Results...... 113 9.3.1 Narrative o f the sea trials...... 113 9.3.2 Sieve net selectivity ...... 114 9.4 Discussion...... 122 9.5 Conclusions...... 125 10 Electric fishing ...... 127 10.1 Introduction...... 127 10.2 General arrangements...... 128 10.3 Materials and methods...... 129 10.3.1 Phase 1 - laboratory experiments ...... 129 10.3.2 Phase 2 - sea trials...... 133 10.4 Results...... 135 10.4.1 Phase 1 - laboratory experiments ...... 135 10.4.2 Phase 2 - sea trials...... 142 10.5 Discussion...... 152 10.5.1 Phase 1 - laboratory experiments ...... 152 10.5.2 Phase 2 - sea trials...... 154 10.6 Conclusions ...... 157 11 Comparison of the different trawls and devices tested...... 161 12 General discussion and conclusions...... 165 13 Abstract...... 169 14 Samenvatting ...... 171 15 Executive summary...... 173 16 Uitgebreide samenvatting ...... 180 17 References...... 188 18 Appendix 1 - Glossary of terms and abbreviations...... 199 19 Appendix 2 - Curriculum vitae...... 206 0-3

Dankwoord

De reis was lang maar het land is in zicht ! Wat de mooiste momenten van dit project waren is moeilijk te zeggen. Op zee gaan is altijd een avontuur en ik hou er goede herinneringen aan over. Ook het werk op het lab en de nationale en internationale samenwerking waren interessant en aangenaam. Maar net zoals ik na elke zeereis blij was dat we de haven binnenliepen en de problemen achter de mg waren, ben ik tevreden dat ik met dit dankwoord bijna aan het einde van de rit ben voor dit proefschrift. Toen ik in 1995 aan het eerste project in een lange reeks begon, had ik geen idee hoe lang het zou duren. Het heeft uiteindelijk meer dan acht jaar gekost om een voorlopig punt te kunnen zetten achter dit werk. De vele zeereizen zullen mij het best bijblijven - de sfeer aan boord samen met de collega’s en vissers, een spiegelgladde zee ais het dan toch eens mooi weer werd, storm en eens niet zeeziek zijn of gewoon een net zien bovenkomen zonder averij. Wat ons te wachten stond op zee wisten we nooit maar dat er problemen zouden zijn stond wel vast. Er zijn nog zekerheden in het leven. Gescheurde netten, stormen, zeeziekte, te grote of te kleine vangsten, motorpech enz. hoorden er bij. Een doorgedreven samenwerking van het team aan boord was noodzakelijk om de problemen de baas te kunnen. Net zoals teamwerk centraal stond op zee, is dit proefschrift veeleer de vrucht van een ganse ploeg enthousiastelingen dan van de auteur alleen. Ik ben blij dat ik heb mogen samenwerken met heel wat mensen uit verschillende disciplines en met verscheidene achtergronden en wil iedereen daarvoor bedanken. Eerst en vooral wil ik mijn promotor, professor Verschoore, bedanken. In België is er geen enkele universitaire vakgroep die zich bezighoudt met visserijtechnieken en ik was blij dat hij bereid was om het promotorschap voor deze vreemde eend in de bijt op zich te nemen. Ik herinner mij vooral onze gezamenlijke zeereis waarbij bleek dat die prof toch veel meer wist over het werk op zee dan onze zeemannen verwacht hadden. Het verhaal van de “oude wijven-knoop” gaat nog altijd mee. Ik ben ook dankbaar voor het geduld dat professor Verschoore heeft opgebracht en het begrip voor mijn werksituatie en natuurlijk voor de bijstand, begeleiding en interesse tijdens die acht jaar. Dan is er natuurlijk de ploeg van het Departement Zeevisserij die bergen werk heeft verzet en aan wie ik veel dank verschuldigd ben. Het zogenaamd zeegaand personeel dat mij evenzeer bijstond aan de wal: Eddy Buyvoets, Johan Coenjaerts, Omer De Cock, Louis de Gheest, Fernand Delanghe, Bernard Demaerel, Katrien Ghesquière, Christian Vanden Berghe - een aantal van hen is er slechts tijdelijk geweest omdat jammer genoeg tijdelijke contracten soms niet konden verlengd worden, anderen hebben de tocht van begin tot eind meegemaakt. Het aantal uren op zee is niet meer bij te houden, het aantal vissen dat door onze handen ging is niet te schatten op enkele duizenden nauwkeurig. Samenwerken met jullie liep altijd gesmeerd, ook in moeilijke omstandigheden. Ik stond er telkens van versteld hoe weinig overleg er nodig was; de werkverdeling ging quasi automatisch. Onontbeerlijk bij het praktische werk was de samenwerking met vissers. Schippers Henry Goutsmit, Manfred Van Elslande, Franky Van Troye, Fernand Verleene, Benny Viaene en de bemanningen van de vaartuigen Black Jack (N.64), Benny (0.101), Norman Kim (0.225), Zeesymphonie (0.455), Virtus (0.533), Bisiti (0.700), Z.403 en Asannat (Z.582) wil ik speciaal bedanken voor de hulp bij de proeven op zee. Onze nieuwe vistuigen waren niet altijd ontworpen voor handig gebruik in de praktijk en de bijsturingen van de vissers werden door ons in dank aanvaard. Onze werkmethodes waren ook dikwijls moeilijk te combineren met de routine aan boord maar er was altijd genoeg goede wil om het werk vlot te laten verlopen. Ook reder Willy Versluys wil ik bedanken voor de aanzet die hij gaf bij het project Dankwoord 0-4 elektrisch vissen en voor de hulp en interesse tijdens het project. Veel van de campagnes op zee gingen door aan boord van de “Belgica”. Mijn dank aan de mensen van BMM, in het bijzonder André Pollentier en Johan Bakkers en aan de bemanning van de Belgica die ons dag en nacht bij stond en die ons dikwijls uit de nood heeft geholpen toen het mis ging. Naast het praktische werk is er ook het wetenschappelijk luik van het project waarbij ik heb kunnen rekenen op de steun van vele collega’s. Ik dank hierbij in de eerste plaats Ronald Fonteyne, Hans Hillewaert en Frank Redant die me inspiratie gaven, die altijd klaarstonden ais ik om advies vroeg en die geregeld daadwerkelijk meehielpen met het werk. Daan Delbare was een welkome hulp bij het ontwerp van onze aquariuminfrastructuur. Ik wil ook Wilfried Vyncke en Rudy De Clerck, departementshoofden van het Departement Zeevisserij, bedanken voor de vrijheid die ze me gaven bij de uitvoering van dit doctoraat. Ook de collega’s van het secretariaat hebben hun steentje bijgedragen bij het verwerken van de overmaat aan projectadministratie. Verder wil ik nog professor Alex Van den Bossche van de vakgroep Elektrische Energie, Systemen en Automatisering van de Universiteit Gent bedanken voor het ontwerp en de ontwikkeling van de labo-pulsgenerator. Ook Bart Huyghebaert ben ik dank verschuldigd die ais thesisstudent deze handige generator in elkaar gezet heeft. Het toestel voldeed perfect aan onze wensen. No international projects without colleagues outside the Belgian borders. These projects were the engine of this PhD and I’m grateful to all who cooperated in applying for projects, executing them and bringing them to a good end. I’ll take the risk of making a list hoping I don’t forget anyone: Jean-Claude Brabant and Serge Mortreux (Ifremer, Boulogne), Per- Sand Christensen (DIFRES, Copenhagen), Erdmann Dahm, Thomas Neudecker, Martin Purps and Harald Wienbeck (Bfafi, Hamburg), Norman Graham (IMR, Bergen), Sean Pascoe, Clive Radcliffe and Stefan Riemann (University of Humberside), Andy Revill (CEFAS, Lowestoft) and Bob van Marlen (RIVO, Ijmuiden), who took the initiative for the first project in this series. I also thank Rene Holst (Constat, Hirtshals) for assisting me in statistical matters and Peter Stewart (Marine Laboratory, Aberdeen) for giving advice and helping me to write the better English. Ten slotte zijn er nog de mensen die niet rechtstreeks geholpen hebben aan dit proefschrift maar die ik gewoon graag zie en die door mijn vrije tijd waardevol te maken de lasten van het werk hebben verlicht: mijn vrouwtje Leen die ook meerdere malen de tekst heeft doorgelezen, mijn ouders en schoonouders, mijn familie, vrienden en vriendinnen...

Van harte bedankt,

Hans Fisheries 1-5

1 Introduction

1.1 Fisheries From the mid 20th century, world fisheries and aquaculture production have shown a steady increase (Fig. 1.1A). Between 1950 and 1970, world capture fisheries (exclusive of aquaculture) production increased on average by 6 % per year from 18 million tonnes in 1950 to 56 million tonnes in 1969 (FAO, 2002). During the 1970s and 80s, the average rate of increase declined to 2 % per year, falling to almost zero in 1990. From 1995 to 2001, capture fisheries production remained rather stationary and the world total estimated production for 2000 was 95 million tonnes. Preliminary catch reports for 2001 from major fishing countries indicate that there may be a marked decrease in global capture production, to about 91 million tonnes. The total world fisheries production, including aquaculture, for 1999 was estimated at 137 million tonnes. About one quarter was produced by aquaculture, which is becoming increasingly important, and showed a significant increase during the last decade. During the first two decades after the second world war, the North East Atlantic capture fisheries played an important role in world marine fish landings, contributing roughly 30 % to the total marine catch (Alverson et al., 1998) (Fig. l.lB). Catches peaked in 1976 at 12 million tonnes. Today, this region only contributes about 10 % of the global capture harvest.

140 14

120 12

o 100 10 C o 80 8

60 6

40 4

20 2

0 1950 1960 1970 1980 1990 1950 1960 1970 1980 1990 Year Year □ Aquaculture □ All other capture fisheries ■ North East Atlantic capture fisheries i North East Atlantic capture fisheries

Fig. 1-1 - Trend in total world (A) and North East Atlantic (B) fisheries production (FAO, 2002).

About 3/t of the total fisheries production is used for human consumption and accounts for about 15 % of the animal protein intake of the human population (FAO, 2002). The other Ya is used for fish meal and oil. The average per capita consumption of fish increased from 9 kg per annum in 1960 to 16 kg in 1997. The vast majority of the fish consumed (75 %) consists of finfish. Shellfish supplies 25 % or 4 kg per capita, subdivided into 1.4 kg of crustaceans, 2.2 kg of molluscs and 0.4 kg of cephalopods. Introduction 1-6 Total world capture fisheries production during the last 5 years appears to be rather stable. There are, however, significant regional differences. The Southeast Atlantic, the Southwest Pacific and the Western Central Pacific showed positive trends in landings in recent years (FAO, 2000). All other areas showed minor changes or declines. The production of individual stocks1, though, shows a huge variability. A number of species had a remarkable increase in production, e.g. Patagonian grenadiers (Macruronus magellanicus), up by 285 % between 1997-98. Other species, however, declined severely e.g. anchoveta (Cetengraulis mysticetus), down by 78 % in the same period. The FAO report on the state of world fisheries and aquaculture (FAO, 2002) made an estimate of the degree of exploitation of the major marine fish stocks. An estimated 25% are under-exploited or moderately exploited. About 47% are fully exploited and have no room for further expansion. Another 18% are over-exploited and need remedial action. The remaining 10% of stocks have been depleted or are recovering from depletion. The historical trends of the state of the world fish stocks were studied by Alverson and Larkins (1994a). They noted the increasing number of overfished stocks and the steady decrease of under-exploited stocks. The trend as from 1970 is given in Fig. 1-2. Over two-thirds of the world’s over-exploited, depleted and recovering stocks are reported to be in the Atlantic Ocean and associated seas, where only 25 % of the global marine fish catch is produced - surprisingly a region subject to rather intense fishery monitoring and investigations (Alverson and Dunlop, 1998). On top of this, the magnitude of overfishing is probably seriously underestimated by the official statistics. Myers and Worm (2003) demonstrated that industrialised fisheries reduced fish community biomass by 80% within 15 years after the start of exploitation, which often is before the start of any monitoring time series. So modem management is usually missing a baseline to determine correct exploitation levels. They showed that large predatory fish biomass is only about 10% of pre-industrial levels.

60

50

40

30

20

1970 1980 1990 2000 Year — Fully exploited - - Under-or moderately exploited Overexploited, depleted or recovering

Fig. 1-2- Trend in the degree of exploitation of world fish stocks (Alverson and Larkins, 1994a).

1 A glossary of terms and abbreviations is given in Appendix 1 Fisheries 1-7 Fisheries in the North Sea can be grouped into three types, i.e. pelagic, industrial and demersal fishing. The trends in landings from the North Sea from 1970 onwards, as given by the Advisory Committee for Fisheries Management (ACFM) of the International Council for the Exploration of the Sea (ICES), are shown in Fig. 1-3. The pelagic fishery landings dropped to a minimum in the late 1970s with the collapse of the Herring stock and the closure of the fishery. Between 1985 and 1995 pelagic landings rose to over 1 million tonnes and since then dropped to about half of this amount. Industrial fisheries depend on species with a short life span and with a quite variable reproduction. A specific year class will only have a short effect in time and consequently landings can be quite variable. In recent years, landings have fluctuated between 1 and 1.5 million tonnes. For the demersal fishery, the most important fish species are Cod Gadus ( morhua), Haddock ( Melanogrammus aeglefinus), Whiting (Merlangius merlangus), saithe (Pollachius virens), Sole (Solea solea) and Plaice (Pleuronectes platessa). The landings of these species have steadily declined over the period 1970-2000 from a total of 1500 tonnes to less than 400 tonnes (Fig. 1.4).

■ Demersal □ Pelagic □ Industrial ■ Cod □Haddock □ Whiting nSaithe DSole □ Plaice

Fig. 1-3 - Trends in landings of demersal, pelagic Fig. 1-4 - Trends in landings of the major and industrial fish species in the North Sea demersal species in the North Sea (ICES, (ICES, 2001a). 2001a).

In the past 10 years, the state of the stocks of most round- and flatfish species (demersal fishery) in the North Atlantic Seas has deteriorated (ICES, 2001a). Some of these stocks have reached a historical low during this period. Atlantic Cod used to be a huge resource with high directed fishing effort. In the early nineties, however, the northwest Atlantic Cod stock collapsed and despite severe capture restrictions the stock still remains close to its historical low, with weak representation of all year classes (NAFO, 2001a). In 1996, the International Union for the Conservation of Nature (IUCN) assigned Atlantic Cod to their “vulnerable” threat category with an increased extinction probability (IUCN, 1996). In the North Sea, the Cod stock is also at a historical low level (ICES, 2001a). In the year 2001, the European Commission’s (EC) advisory committee STECF recommended severe catch restrictions and technical measures, like a mesh size increase and the introduction of Introduction 1-8 selective gears, (Anon., 2001a) and the EC installed the “Cod recovery plan”. ACFM also states that the stock of Whiting has shown a continued decline over time and is outside safe biological limits (ICES, 2001a). The stock of Haddock presently profits from a good year class recruiting into the spawning stock, but the exploitation rate is still too high. Plaice is outside safe biological limits and fishing mortality of both Plaice and Sole are high and unsustainable in the long term. The spawning stock biomass of Sole is decreasing. The saithe stock, however, is now considered to be within safe biological limits. For Brown Shrimp (Crangon crangon), the landings are quite variable but the trend is not decreasing (ICES, 2001b) and catches remain high compared to historical data. On several occasions, concern about the situation of fish stocks and the urgent need to take management actions has been expressed. The most important statements on a global scale were the “UN Convention on the Law o f the Sea ” (UNCLOS), the “New York Agreement on Conservation and Management of Straddling Stocks and Highly Migratory Fish ”, the “FAO Code o f Conduct For Responsible Fisheries'” (1995) and the "International Conference on the Sustainable Contribution o f Fisheries to Food Security, Kyoto” (1995). For the North Sea, important recommendations were made in the frame of the North Sea conferences like the “Esjberg Declaration ” (Anon., 1995), the “Bergen Declaration ” (Anon., 2002) and the “Intermediate Ministerial Meeting on the Integration o f Fisheries and Environmental Issues ” (Anon., 1997), the EC “Green Paper” (Anon., 2001b) and the ACFM advice (ICES, 2001a). They all focused on reductions in fishing mortality, reductions of fishing effort and implementation of management measures to reduce the amounts of juveniles caught and discarded. The study, monitoring and management of fish stocks are thus of major concern to governments. In addition to the concerns expressed about individual fish stocks, there is a growing worldwide interest in and concern about ecosystems and the impact that fishing may have on their structure and function (e.g. Haii, 1999, Kaiser and de Groot, 2000, FAO, 2002). A key-study on the environmental impact of fishing in the North Sea (Lindeboom and de Groot, 1998) concluded that bottom trawling generates significant changes in the marine ecosystem and recommended that further research should be carried out and management actions undertaken. This was endorsed by the EC (Anon., 2001c) and ICES (ICES, 2000a). Overall, the international community is well aware of the crisis facing fishery resources and fisheries, and has taken a number of important steps towards improved global management of fisheries. The problem for the 21st century is to implement these successfully and in particular, to translate policy to the level of application by the stakeholders: fishing companies, vessel owners, small-scale fishers, skippers and coastal communities. The global fishery crisis consists, however, of a large number of problems associated with individual fisheries, each with its own characteristics. To improve fishery management overall, solutions are needed for each individual over-exploited fishery. The present study follows this approach and concentrates on one small segment that contributes to the fishery crisis, i.e. the demersal fishery for Brown Shrimps Crangon( crangon) in the North Sea and the associated by-catch and discard issues. Quantification of the discarding in this fishery, its biological and economic consequences, together with the identification of possible technical solutions to reduce the discarding, were the major themes of the study.

1.2 Area of study The area of study for the Belgian experiments was the shrimp fishing grounds on the Belgian Continental Shelf, located in the southern part of the North Sea. Brown shrimp are fished, Area o f study 1-9 however, all over the North Sea, all along the continental coasts and in isolated parts of the UK east coast (Fig. 1-5). The present study focussed on the Belgian area but all Brown Shrimp fishing grounds were covered. The North Sea is a relatively shallow semi-enclosed basin (Fig. 1-5). Its depth ranges from about 30 m on average in the southeast to 200 m in the northwest. Although it is small in global ocean tenns, it is not homogeneous as regards temperature, water type and the nature of the substrate. The high productivity of the North Sea is associated with the relatively shallow depth and the existence of mixing mechanisms transporting nutrients from the nutrient-rich bottom layer to the nutrient-poor upper layers of the water column. The southern part of the North Sea is pennanently mixed, whilst the middle and northern parts become stratified in suimner. The North Sea and adjacent seas constitute one of the world’s major shelf areas and, as such, one of the major fish producing ecosystems in the world. With an annual production in the range of 2.5 million tonnes, the area contributes 4 % of the world’s capture fish production (Anon., 200Id). The commercially important human consumption stocks include Cod, Haddock, Whiting, Saithe, Plaice, Sole, Herring, Mackerel, Nephrops norvegicus , Pandalus borealis, and Brown Shrimp (Crangon crangon). A number of stocks are mainly or exclusively exploited for reduction to fish meal and oil: sand eel, Norway pout, blue Whiting, and sprat. Beyond these stocks, there are landings of a variety of demersal species such as Turbot, anglerfish, gurnards, lemon Sole, rays, and sharks.

60 c

Via

Ive

VHct

Fig. 1-5 - The north western European waters with the ICES sub-areas and the major Brown Shrimp fishing grounds indicated. The North Sea comprises ICES sub-areas IVa, b and c.

The total biomass of all fish in the North Sea is estimated at approximately 10 million tonnes. The total annual landings of fish fluctuate around 2.5 million tonnes, and approximately the same amount of fish is eaten annually by the main predatory fish species: Cod, Haddock, Whiting, saithe and Mackerel (Table 1.1). The amount of fish eaten by birds and marine mammals amounts to approximately 3/4 million tonnes. Many stocks of fish Introduction 1-10 spawn in the North Sea, and for most of them the water circulation distributes the young from the spawning to the nursery grounds. For a number of commercially important species of fish (e.g. Plaice), the spawning grounds tend to be in the northern and western parts of the area, whereas the nursery grounds are in the more shallow parts of the southeastern North Sea. The North Sea ecosystems do not fluctuate strongly in terms of the main species; neither are they entirely stable and, even in the absence of human influence, there would be continuous changes in the abundance of all species. Long-term changes may be caused by gradual changes in climate, or in hydrography. Superimposed on these trends are annual fluctuations caused, for example, by cold winters or warm summers or by short-term changes in current flow. Demersal human consumption fisheries in the North Sea usually target a mixture of roundfish (Cod, Haddock, Whiting), flatfish (Plaice and Sole) with a by-catch of roundfish or crustaceans (Nephrops or shrimps) often with a mixed by-catch. A wide variety of fishing gears is used: demersal trawls (otter and beam trawl, pair trawl), pelagic otter and pair trawls, Danish and Scottish seine, set nets and purse seines. The focus of this study was the Brown Shrimp fishery in the South-Eastern North Sea, an area characterised by the presence of a large number of sandbanks; a turbulent and very dynamic area with strong currents and intense sediment transport. The prevailing sediment is sand. The sandbanks are 1 to 3 km wide and 10 to 60 km long. The height varies from 3 to 40m and the cross-section is usually asymmetrical with a gentle and a steep slope. Several fishing methods are used, but trawling is the most common with the flatfish or shrimp beam trawl as the most important fishing gears. It is an area of intense human activity, with heavy shipping traffic, tourism, fisheries, sand and gravel extraction, dumping of dredge spoils, presence of pipelines etc.

Table 1.1: Recent landings from the North Sea and adjacent areas - 1991-1999 (Anon., 2001d) Stock Average, Stock Average, 1,0001 1,0001 1991-1999 1991-1999 Sand eel, North Sea 851 Whiting, North Sea and Eastern 87 Channel Herring, North Sea and Skagerrak 509 Sand eel, Skagerrak 37 (autumn spawners) Mackerel, North Sea and Skagerrak 323 Sprat, Skagerrak 29 Sprat, North Sea 190 Sole, North Sea 27 Norway pout, North Sea and Skagerrak 178 Brown shrimp {Crangon crangon) 23 Haddock, North Sea and Skagerrak 139 Shrimp {Pandalus), North Sea and 15 Skagerrak Cod, North Sea, Skagerrak and Eastern 120 Anglerfish, North Sea 14 Channel Saithe, North Sea, Skagerrak and West of 115 Plaice, Skagerrak and Kattegat 10 Scotland Plaice, North Sea 102 Plaice, English Channel East 6 Blue Whiting, North Sea and Skagerrak 95 Hake, North Sea and West of 6 Scotland Horse Mackerel, North Sea and Skagerrak 91 Sole, English Channel East 4 Organisational and financial framework o f this study 1-11 1.3 Organisational and financial framework of this study The author of this report has been working in the “Departement voor Zeevisserij” (Centrum voor Landbouwkundig Onderzoek - Gent) in Oostende, Belgium, during the course of this study. The work was done within the “Technical Fisheries Research” Section and supported by expertise from the “Biology” Section of the same institute. Most of the tasks were carried out within a framework of a series of projects. The author of this PhD thesis was one of the initiators of these projects and was actively engaged in the research. The following projects were carried out and are an integral part of this PhD thesis:

• BIO 94/C 144/04 "Research into Crangon fisheries unerring effect" (RESCUE) Aim: To make an inventory of fishing gear and vessel characteristics of the Brown Shrimp trawler fleet and quantify the discards in this fishery. Period: 01/03/95 -31/05/97 Participants: Departement voor Zeevisserij, Oostende, Belgium; Rijksinstituut voor Visserij Onderzoek, IJmuiden, Netherlands; DIFMAR, Charlottenlund, Denmark; University of Humberside, Grimsby, UK; Institut für Seefischerei, Bundesforschungsanstalt fuer Fischerei, Hamburg, Germany; IFREMER, Boulogne sur mer, France. Author’s task: national coordinator, field work, data analysis, reporting; carried out in co­ operation with the Biology section of the Sea Fisheries Department (SfD), Oostende.

• ECO - 97/SE/025 “Economic Consequences of Discarding in the Crangon Fisheries” (ECODISC) Aim: To determine the biological and economic consequences of discarding in the Brown Shrimp fisheries Period: 01/03/98 - 28/02/99 Participants: Departement voor Zeevisserij, Oostende, Belgium; DIFMAR, Charlottenlund, Denmark; University of Humberside, Grimsby, UK; Institut für Seefischerei, Bundesforschungsanstalt fuer Fischerei, Hamburg, Germany; Cemare, University of Portsmouth, Portsmouth, UK; Cefas, Lowestoft, UK; ARBEE Computer Consultants, Grimsby, UK. Author’s task: data supply, evaluate model; carried out in co-operation with the Biology section of SfD.

• BIO 98/012 “Reduction of discards in Crangon trawls”(DISCRAN) Aim: To determine the selective properties of sorting grids and sieve nets for the Brown Shrimp fishery. Period: 01/03/99 - 28/02/01 Participants: Departement voor Zeevisserij, Oostende, Belgium; Rijksinstituut voor Visserij Onderzoek, IJmuiden, Netherlands; University of Newcastle, Newcastle, UK; Institut fur Seefischerei, Bundesforschungsanstalt fuer Fischerei, Hamburg, Germany. Author’s task: national coordinator, field work, data analysis, reporting.

• 5BW/EOGFL 30B-A.4.1 "Ontwikkeling van een milieuvriendelijke visserijmethode voor de garnaalvisserij, gebaseerd op stimulering door elektrische pulsen " Aim: To develop a selective shrimp trawl based on alternative stimulation, i.e. electric pulses. Introduction 1-12 Period: 20/11/97 - 31/09/01 Participants: Departement voor Zeevisserij, Oostende, Belgium; Redersentrale, Oostende, Belgium; Brevisco bvba, Bredene, Belgium. Author’s task: coordinator, field work, data analysis, reporting. Overfishing 2-13

2 Problems associated with fisheries in general and Brown Shrimp fisheries in particular

2.1 Overfishing At his inaugural speech at the London Fisheries Exhibition in 1882, biologist Thomas Huxley declared ‘7 believe that it may be affirmed with confidence that, in relation to our present modes of fishing, a number of the most important sea fisheries, such as the cod fishery, the herring fishery, and the mackerel fishery, are inexhaustible. And I base this conviction on two grounds, first, that the multitude of these fishes is so inconceivably great that the number we catch is relatively insignificant; and, secondly, that the magnitude o f the destructive agencies at work upon them is so prodigious, that the destruction effected by the fisherman cannot sensibly increase the death-rate ” (Huxley, 1884). Quite recently, Hudson (1996) noted the “widespread and ingrained belief that some fishes are inherently resilient to extinction and will not disappear, however hard they are fished’'. Hutchings (2001) mentioned the “contemporary perceptions of the ability of marine fishes to recover from population decline”. This belief in the vastness of ocean resources and the ability to recover has driven the huge expansion of the world fisheries and has led to a high pressure on exploited stocks. Overfishing is, however, not new and historical extinctions of marine animals due to human activity have been documented and are quite likely severely underestimated (Carlton et al., 1999). Jackson et al. (2001a and b) demonstrated that primitive communities can have a serious impact on coastal ecosystems and that humans transformed marine ecosystems long before modem ecological investigations began. Historical abundances of a number of consumer species were very high in comparison to recent observations, and many top predators and grazers have disappeared. The authors also noted that extinction caused by overfishing precedes all other pervasive human disturbance to coastal ecosystems. The impact of present day fishing is considerable. Pauly et al. (1995) calculated that for fisheries on continental shelves and in upwelling areas between 24 and 35% of the primary production is needed to sustain the reported catches. They concluded that there are broad limits to the carrying capacity of natural aquatic ecosystems and expressed their concern for sustainability of the fisheries and biodiversity. A quite pessimistic view was presented by Jackson (2001) saying “no wild Atlantic coastal fishery is sustainable at anything close to present levels of exploitation”. The arguments supporting this statement were 1) the continuing human population growth driving a constantly increasing demand, 2) development of mechanised fishing technologies severely damaging the environment, 3) cheap and rapid transportation, making even the most distant populations vulnerable and 4) management, consistently failing in conserving depleted stocks. The North Sea demersal fishery is beset with problems and the causes of fish stock declines, observed in the North Sea, are indeed manifold.

2.1.1 Overcapacity Overfishing is often characterised as “too many fishermen chasing too few fish” but the enormous harvesting capacity due to an industrial approach to fishing has become one of the main causes. Fish stocks in the North Sea are a common resource and the production of individual vessels is interdependent. Skills and equipment are key elements in the profitability of individual vessels. In an open access fishery, this often leads to the so-called “race to fish” and Problems associated with fisheries in general and Brown Shrimp fisheries in particular 2-14 excessive growth of harvesting capacity (Lindebo, 1999). In the North Sea fishery, the so- called “ratchet effect” (Ludwig et al., 1993) has boosted capacity growth. According to this mechanism, stable periods with sufficient resources encourage investment in vessels and processing capacity. Science often is unable to predict future overfishing with enough certainty to convince management to disinvest. As stocks decline, the fishing industry continues to expand and appeals are made to the government for help to overcome the difficult period; substantial investments and many jobs are at stake. The governmental response typically is to provide direct or indirect subsidies and no disinvestment occurs. The long-term-outcome, as observed in the North Sea in the 1980s and 90s, is a heavily subsidised industry that over harvests the resource. In a World Bank publication (Milazzo, 1998), it is calculated that the global fisheries subsidies are between 17 and 25 % of the industry’s revenues. In this publication, the EU was indicated as one of the largest subsidisers. The EC explicitly stated in its green paper in the year 2001:The “ current fleet is much too large ” (Anon., 2001b). As far back as 1983, however, a first “multiannual guidance programme” (MAGP I - 1983-86) was set up by the European Community including a system of financial aid for the removal of vessels from national fishing fleets. MAGP’s have been the key element of structural policy and management of fishing effort on a consistent, longer-term basis. MAGP I was followed by MAGP II (1987-1991), MAGP III (1992-1996), MAGP IV (1997-2002) and MAGP V (2001-2006) with more ambitious objectives for overall reductions. These programmes have, however, often been criticised for failing to control adequately the real growth in the catching capacity due to technical advances in fishing methods. There is also the apparent contradiction in a policy that has simultaneously provided financial aid for both increasing and decreasing fleet capacity (Hatcher, 1999). The EC stated in its green paper “The M AGP’s were set by the Council at levels that were not ambitious enough to address the problem of excess capacity effectively and have often not been enforced. Subsidies for construction/modernisation and running costs may have aggravated the current situation” (Anon., 2001b).

2.1.2 Technological development Fishing capacity is not only determined by the number of vessels and engine power but also by the technological efficiency of the fishing vessels and their gear. New fishing methods, increase of the size and improvement of the efficiency of existing fishing gears, mainly after 1945, increased the size of catches, allowed fishing on new fishing grounds (previously inaccessible) and allowed fishing for new species (Safina, 1995). In his review, Valdemarsen (2001) indicated the following developments as the main contributions to technological advance in capture fisheries: developments in diesel engines, hydraulics, electronics and refrigeration; the introduction of synthetic fibres; new fishing methods and net designs; new types of groundgear; acoustic fish detection, navigation aids and trawl monitoring instruments. A study on e.g. the northern Australian prawn fishery indicated that the introduction of GPS navigation aids increased the fishing power by 12 % (Robins et al., 1998). This is also illustrated by Wilen (1990), citing the example of the northeast Pacific halibut fishery. In the face of powerful and persistent increases in fleet strength and technical efficiency over the years, the season was shortened from nine months in the 1930s to about five days by the mid 1980s which was by then all it could stand. In addition, the fish processors reported 50% of the fish not iced and 30% not gutted, leading to waste, such was the haste with which the fleet worked in the short time available. This clearly demonstrates the problem of over investment. Science and management 2-15 Quantitative data on the increase in fishing power due to technological advance are rare, but technological development has undoubtedly counteracted efforts to reduce fishing capacity by decommissioning schemes in the North Sea.

2.2 Science and management The scientific advice supporting management, aims at a sustainable fishery and plays an important role in modem fisheries. Modem fisheries management depends heavily on scientific assessment of stocks. The advice, however, holds a high degree of uncertainty. This uncertainty can be caused by e.g. sampling errors, inadequate data collection systems, misreporting and the lack of complete understanding of the population and community dynamics (Francis and Shotton, 1997). The absence of risk assessment in scientific advice, highlighting this uncertainty, has often led to poor management decisions (Francis and Shotton, 1997, Caddy and Cochrane, 2001). One example is the case of the Canadian northern Cod stock where poor scientific advice played an important role in the collapse because, according to Hutchings (1996), the effects of fishing on the exploited Cod populations were constantly and consistently underestimated (i.e. poorly understood). During the last decade, a lot of effort has been put into improving the quality of the advice and breakthroughs in modelling complex ecosystems, together with the growing use of GIS technology look promising (Caddy and Cochrane, 2001). The EC, however, expressed its fear (Anon., 2001) that fisheries scientists in the EU Member States are too involved with the year-to-year routine of providing advice on TACs to allow time for innovative thought on alternative management measures. The Common Fisheries Policy (CFP) of the European Union, established in 1983, can be seen as one of the most advanced economic and political integrative processes that developed in Europe after World War II and has been confronted with major challenges ever since. In theory, the CFP had good potential but has not achieved its goals, despite some positive results like containing conflicts at sea, the provision of some degree of stability in the fisheries sector and, despite the high fishing pressure, avoiding the total collapse of stocks as some areas of the world have occasionally witnessed. The main flaw is that the policy has not delivered long-term sustainable exploitation of fisheries resources. Botsford and Castilla (1997) argued that continuous socio-political pressure to increase total allowable catch, along with scientific uncertainty, has kept fishery managers from sustaining the resource. The criticism of the CFP was summarized by Bailly and Collet (1999) as follows: the national political agendas have had too high an influence on the policy choices; fishermen’s participation in the decision-making process has been too low, which led to low compliance; regulations have been too technocratic rather than pragmatic; there was a lack of enforcement, monitoring and means for control and there were contradictions in the policy objectives. The EC itself stated that (Anon., 2001b) it has not been possible to take advantage of all tools available in the legislation. There was limited progress in adopting multi-annual guidance programmes and fishing effort management yielded poor results. Difficulties with TACs were caused by the Councils fixing them, in some cases, at levels higher than indicated in the scientific advice, and by illegal or black landings etc. It was also admitted that the new technical measures regulation in force since early 2000 can only partially remedy current problems. Authorized mesh sizes remain too small for effective protection of juveniles, inspection of mesh sizes remains difficult, regulations are too complex and the use of selective fishing techniques has not reached its full potential. Problems associated with fisheries in general and Brown Shrimp fisheries in particular 2-16 2.3 Monitoring and control Management measures have to be implemented and controlled. Monitoring and control activities to enforce the CFP are widely seen as insufficient and discriminatory (Anon., 2001b). The absence of harmonization of sanctions and the limited powers of Community inspectors are major obstacles. Neither has the Commission been granted the human resources or the powers to enable an effective organization of monitoring and control.

2.4 Environmental degradation It has been well demonstrated that fisheries impact individual commercial fish stocks (FAO, 2002; ICES, 2001a). Fisheries may, however, impact the whole marine ecosystem. By removing biomass, fisheries change the food base for other biota, change predation pressure, change the population structure of species and may even remove certain species from the ecosystem. Discarding creates an additional food source for scavengers and mechanical effects of trawling may alter the habitat and benthic communities. The impact of fishing on the species composition of marine ecosystems has been revealed in the trends in species composition of landings, referred to by Pauly et al. (1998) as“fishing down marine food webs”. He demonstrated that landings from global fisheries have shifted in the last 45 years from large piscivorous fishes towards smaller invertebrates and planktivorous fishes, especially in the Northern Hemisphere. The authors suggest that this may imply major changes in the structure of marine food webs and may lead to widespread fisheries collapses. Garcia and Newton (1997) concluded that despite technological advances and the increase in world fleet size, the landing rate of high value fish species decreased by 25 % between 1970 and ’89. Caddy and Garibaldi (2000) support this but found clear regional differences in the causes of this phenomenon because other human induced changes in the marine ecosystem can affect populations. They stated, however, that for the Northern Atlantic, “fishing down the marine food webs” seems to be a candidate mechanism for explaining these catch trends. Two Canadian studies (Pauly et al., 2001; Bundy, 2001) provided a local confirmation of this trend. For the Canadian east coast, it was concluded that higher trophic level fish like Cod had greatly disappeared from the landings while that of invertebrates like the snow crab (Chionoecetes opilio) had greatly increased. A similar trend has been observed in the North Sea, where landings of the pelagic and industrial fisheries, which exploit species lower in the food web, have been able to maintain their catch levels while landings of demersal species have shown a steady decline over the past 30 years (Fig. 1-3). The size structure of exploited fish populations changes with increasing exploitation towards a lower abundance of larger individuals (ICES, 1999). Studies of historical trawl catch data from the North Sea also indicated large changes in population densities for several species (ICES, 1999). Apart from commercial fish species, non-commercial fish and invertebrates can also be affected (Haii, 1999; Kaiser and de Groot, 2000). Rijnsdorp et al., 1996, found that of the 19 species recorded, 18 were found to have decreased between the beginning of the 20th century and the 1990s, for several down to 1% or less of the original densities. The susceptibility of populations to fisheries impact is assumed to be associated with some life history characteristics including large ultimate body size, slow growth rates, high age at maturity and low fecundity. This is confirmed by the IMPACT II study (Lindeboom and de Groot, 1998) that found significant changes in community structure in the North Sea due to fishing activities, with a dominance of opportunistic short-lived species and a decrease of long-living sessile organisms. This was confirmed by the study on long-term trends in the North Sea by Clark and Frid (2001) who also found that non-commercial fish species increased in density, while commercial species decreased. Discarding 2-17 Apart from the mortality caused by catching marine organisms, the fishery may also have physical effects on the seafloor and the organisms that live in and on it. Fishing gear, especially trawls, causes physical damage to animals or indirect damage through frequent perturbation of the sediment (Lindeboom and de Groot, 1998, Kaiser and de Groot, 1999). In that way, extra food becomes available to scavengers, which may alter the interactions between species. Fisheries may also affect seabirds and marine mammals. By increasing the availability of food at the sea surface when fish are discarded or by reducing the abundance of fish populations, which are important food items, seabird populations may be affected (Camphuysen et al., 1995, ICES, 2000b). For marine mammals, incidental capture in fishing operations is the most obvious hazard but fisheries may also affect these animals by depleting or changing their food source (ICES 2001c). The degree to which fisheries are responsible for observed ecosystem changes is very hard to determine (Haii, 1999). Other influences like pollution, eutrophication, climate change and introduction of alien species have effects on marine ecosystems and the fishing industry needs a better environmental policy. A number of studies have indeed proved that climate and nutrient input have had a major influence on long-term changes in the North Sea ecosystems (Clark and Frid, 2001; Kröncke, 1995; Thatje and Gerdes, 1997). The environmental impact of the North Sea Brown Shrimp fisheries has not yet been studied.

2.5 Discarding The discarding at-sea of harvested fish and the associated mortalities is one aspect of fisheries impact on the marine environment. It has been recognized as an inherent problem in the management of world fisheries. Before the 1980s, quantitative data on discards were rare and scattered but in 1983 a first global assessment of by-catch and discard levels was made by Sada (1983). In the report, Sada estimated a minimum world discard of fish and shellfish of more than 6 million tonnes. It was admitted that only minimal or partial data were available and that this number was an underestimate, but the report was seen as an eye- opener for the discard problem. The next global discards assessment was made by Alverson et al. in 1994. The authors provided a provisional estimate of 27 million tonnes, roughly one- third of the world capture fisheries production. Shrimp trawl fisheries (especially tropical) were found to generate more discards than any other fishery type and accounted for over one-third of the global total. The Northeast Atlantic fisheries produced almost 15 % of the global discards. The weight-based discard ratio (ratio of discarded to total catch weight) for the world’s major fisheries, ranged from a maximum of 84% for shrimp and prawn fisheries to almost zero for the Mackerel fishery. This ratio for North Sea shrimp trawling was estimated at 59%. Although these figures may look impressive, large quantities of discards do not necessarily lead to significant biological or ecological impacts. Conversely, to presume effects are minimal because discard quantities are low may also be misleading. More detailed studies that link the discard data with population or ecosystem characteristics are necessary to improve the understanding of fishery impact. Still, the biological or ecological impact of discarding is of major concern to fisheries managers (e.g. Anon., 2002). The discard issue is also one of waste; the millions of tons of protein dumped in the ocean and the waste of animal lives is often condemned on moral grounds. Discarding may also have economic and social consequences (Alverson et al., 1994). Discarding practices by a certain fleet may affect the stock of its target species or the target species of other fleets, and may cause conflicts and generate additional costs without Problems associated with fisheries in general and Brown Shrimp fisheries in particular 2-18 affecting the revenues. Discarding also raises management and enforcement costs and gives fishermen a bad public image. The absence or poor quality of discard data introduces uncertainty into stock assessment, although the new National Data Gathering Program of the EC aims to improve these data significantly.

2.6 The Brown Shrimp fishery Although the Brown Shrimp stock does not show signs of decline and over exploitation, over capacity and heavy subsidising has been recognised as a problem in the North Sea inshore fishery. A study by De Wilde (1999) showed that this was particularly the case for the inshore beam trawler fleet in the North Sea, of which the shrimp trawler fleet is one component. Moreover, these boats were not only bigger than the recorded tonnage figures, they were also more powerful than their nominal main engine power of 221 kW. Most of these boats had main engines installed de-rated to 300 hp, but had continuous outputs of 450 hp up to 520 hp. Consequently, the capacity of this fleet is larger than is reflected by the figures used by scientists and managers. Fishing effort is the main measure used to evaluate the activity of the North Sea shrimp trawler fleet. The standard definition of fishing effort is “fishing capacity x fishing time”. Fishing capacity is a function of vessel size, engine power, gear type and crew skill. This definition cannot always be used due to lack of data. Temming (1992) presented different measures of effort each confirming an overall increase in fishing effort in spite of decreasing vessel numbers in the German fleet. An excessive growth of the fishing capacity has, however, been counteracted during the last decade by an industry driven agreement on catch and effort restrictions aiming at a stabilisation of prices. Also the overall landed catches of Brown Shrimp remained at a high level in the second half of the 1990s (ICES, 2000c) indicating that the stock was not declining. However, the effect of this fishery, carried out in vulnerable coastal areas and estuaries, on the stocks of other species, remained unclear. The technological development of the Brown Shrimp fleet has not been recorded and no systematic inventory of vessels, equipment and fishing gears was available. Therefore, the impact of technology on the fishing capacity of this fleet is difficult to appraise. It is, however, a matter of course that the shrimp trawler fleet would have followed, at least to some extent, technological developments in navigation, catch processing and fishing gear resulting in a higher capacity per unit of engine power. Some indication of the impact of technology is given by Neudecker (1999) who calculated that the trawled area by the German shrimp trawler fleet in 1956, with 630 vessels, was smaller compared to that in 1996 with only 247 vessels. Temming (1992) presented a list of technological improvements introduced since the early 1960s: mechanical sorting sieves, hydraulic winches with four drums, light and durable synthetic nets, separator panels, more powerful engines, kort nozzles, hull and propeller design, larger vessels, radar, echo sounders, navigation equipment, automatic steering, radio facilities, modem cookers, refrigerated fish holds and conveyor belts. According to Boddeke (1989), crews could at that time process three tons of consumption shrimps, compared to only one ton in 1960. For Brown Shrimp, no stock assessment is carried out and the number of scientists working on this species is quite low, despite the fact that Brown Shrimp was the fourth most valuable species landed in the North Sea in 1999 (ICES, 2000c). The Netherlands (with the second most important Brown Shrimp fleet) have, for example, not been delegating a scientist to the ICES Crangon Working Group for a number of years. Basic information, like fishing effort, is very difficult to assess because each country uses different measures and the growth and life cycle of the species is not yet clearly understood. As a consequence, science can give little guidance to the management of the fishery. The problems associated with monitoring The Brown Shrimp fishery 2-19 and control activities to enforce the CFP in general (see Section 2.3) also apply to the North Sea shrimp fisheries. The environmental impact of the North Sea Brown Shrimp fisheries has not been studied in detail as has been done for the flatfish beam trawl fishery in the North Sea (Lindeboom and de Groot, 1998). Some degree of similarity can be expected due to the similarities in the fishing method. So seafloor disturbance, direct and indirect mortality to target and non-target fish and invertebrate species are likely to occur. Long term changes in the ecosystem are also likely to be a consequence of intense shrimp fishing. Some examples exist of the environmental effects of shrimp trawling. Riesen & Reise (1982) reported that extensive sub- tidal Sabellaria spinulosa reefs were lost in the Waddensee between 1924 and 1982; they reported that local shrimp fishermen claimed to have deliberately destroyed them with "heavy gear" as they were in the way of the shrimp trawling. Reise & Schubert (1987) reported similar losses and attributed them to the same causes. Shrimp trawling still occurs in these areas and the S. spinulosa have not reappeared, but have effectively been replaced by mussel Mytilus edulis communities and assemblages of sand dwelling amphipods (Reise & Schubert, 1987). Berghahn and Vorberg (1997) found that the shrimp trawl penetrates up to 2cm in the bottom and that trawling has a negative effect on the abundance of a number of benthic animal species. They state, though, that this fishery has no major impact on the benthic ecosystem in the highly dynamic and rapidly changing Wadden Sea but discarding practices may bring about a population shift. Concern is expressed about the trend of increasing fishing effort with bigger vessels, more frequent trawling and heavier gear. The Brown Shrimp fishery is carried out in coastal zones and estuaries with small meshed nets. The discarding practices associated with it have been regarded as a problem for many years. Already in the early 1930s, the high mortality of the fish discards and the possible biological and economic consequences of the destruction of large quantities of young fish in the Belgian, Dutch and German shrimp fishery were addressed (Wulff and Bückmann, 1935, Gilson, 1935). Klausing and Lefevere (1961) pointed at the possible consequences of landing large amounts of undersized shrimps and other by-catch (juvenile fish, crabs etc.) for industrial processing instead of returning the by-catch to the sea. The by-catch issue appeared in the literature between 1950 and ‘90 (e.g. Gilis, 1951, Mistakidis, 1958, Meyer and Thiews, 1965, Kelle, 1976/77), each time pointing at high amounts of discards in the Brown Shrimp fishery. The absence of a sound knowledge of population parameters of commercial fish species, however, made impact assessments on stocks problematic and the real consequences of these practices could not be estimated with confidence. Despite early investigations into selective shrimp trawls (e.g. Kurc and Faure, 1965, Van Middelem and Cleeren, 1967, Brabant, 1973), implementation of selective devices has been rare in the North Sea. In the late 1980s and 1990s, however, the pressure on the shrimp fisheries increased considerably. At the second North Sea Conference (Anon., 1987), no attention was given to fisheries, but at the third North Sea Conference (Anon., 1990), a recommendation was made to draw attention to the problem of fisheries impact on the North Sea ecosystem and to give high priority to the protection of the Waddensee, an area intensely fished by the shrimp trawler fleet. At the fourth North Sea Conference (Anon., 1995), the Ministers agreed to request the competent management authorities to take actions to investigate and reduce the impact of fisheries and discarding. Coastal areas and estuaries were seen as extremely vulnerable and a ban on fisheries in large parts of the Waddensee was demanded (Hagena, 1991, Will, 1993). The problem of discarding in the shrimp fishery also turned up in the media and public opinion slowly turned against the shrimp fishermen (e.g. Berghahn, 1992). The discussion, however, was difficult since no sufficiently reliable data were available covering the whole fishery. There was no discussion on the issue that large amounts of by­ Problems associated with fisheries in general and Brown Shrimp fisheries in particular 2-20 catch were returned to the sea, but the seriousness of the problem and the possible consequences for the fish stocks and the ecosystem were minimised by the fishermen and other fisheries were blamed (Hagena, 1991, Carstensen, 1992). Environmentalists claimed the opposite. The need for data on this issue and a solution for the discard problem was the starting point of the present study. The Brown Shrimp fishery 3-21

3 Objectives

The main objectives of this study were to quantify the biological and economic consequences of discarding in the Belgian Brown Shrimp (Crangon crangon) fishery and to evaluate possible technical alterations to the shrimp beam trawl to reduce discarding in this fishery. The study was carried out in five consecutive steps: • the collection of data on the characteristics of the Belgian shrimp trawler fleet, • the collection of data on the discarding practices in the Belgian Brown Shrimp fishery, • estimation of the biological and economic significance of the discarding, • study of the selectivity of the shrimp beam trawl and • evaluation of discard reducing technical alterations to the shrimp beam trawl as possible management tools. Since detailed knowledge of the structure, characteristics and exploitation patterns of the shrimp fishery in the North Sea, in a uniform data format, was lacking, a thorough inventory of the fleet, outlined in Chapter 4, was the first step in the study. These data were indispensable for the further course of the study, i.e. 1) consistent data on fishing grounds, fishing effort and seabed surface area fished were necessary for the extrapolation of the discard data collected in the discard sampling programme; 2) knowledge on the degree of variability in vessels and fishing gears was necessary to choose representative vessels for the sampling programme; 3) this knowledge would give an idea on the reliability of the extrapolations; 4) information on the fishing gears was very useful for the further technical research on the fishing gear. North Sea wide data on the discarding practices in the Brown Shrimp fishery were lacking and the collection of these data, following a standardised methodology, was one of the main objectives of the present study. The data on numbers of commercial fish species and Brown Shrimps discarded are presented inChapter 5. At the start of this project, the impact of discarding by the Brown Shrimp fishery was a hot issue, but discussions were not backed up by sound data. The collection of these data was therefore the obvious next step in the project. This information was needed for the appraisal of the biological and economic consequences of discarding in the Brown Shrimp trawler fleet. Taken at face value, data on juvenile commercial fish discarded in the Crangon fisheries can be impressive, but the impact of the amounts of fish returned to the sea remained largely unknown. In order to put these numbers in their proper perspective, the biological and economic significance of the discarding had to be studied. To be useful for scientists and managers, this was done in terms of the potential impact upon recruitment and spawning stock biomass of the fish species concerned, and hence on the potential lost landings and their associated market value. This information was also valuable to steer further technical gear research. The analysis was done by means of a purpose-built model. The results are given in Chapter 6. This analysis drew upon the joint expertise of fish and shrimp biologists, gear technologists and fisheries economists in order that the multitude of biological, technical and economic factors be taken into account. The task of the author has been to provide the basic data in a suitable format, to comment on the newly designed model and to give advice on the fishing gear aspects of the exercise. Biologists and economists, however, carried out the modelling work and consequently, the model is not elaborated in detail in this report. Because of its importance for further work, however, a short description of the model design and a summary of the results are presented. Objectives 3-22 A detailed study of the selectivity of the Brown Shrimp beam trawl was considered an essential preparatory step in the search for technical solutions to improve the selectivity and reduce the discards. In addition to the selectivity of the cod-end, which is generally seen as the most important part for selection, the selectivity of the other parts of the trawl was also studied. The selective properties of the shrimp beam trawl are described in Chapter 7. Several management options exist to reduce discards in a fishery: reduction of fishing effort, closed areas, closed seasons etc. In this study, however, the objective was to try and evaluate discard reducing technical alterations to the shrimp beam trawl as a possible management tool. Three options were selected as having potential: the selective sorting grid, the selective sieve net and electric fishing. The selective sorting grid was chosen as an option because it had been successfully implemented in the northern shrimp{Pandalus) fishery. The sieve net was already known in the Brown Shrimp fishery, to avoid problems with large by-catch like jellyfish but the selective properties for fish had not been studied. The fact that it was already being used on a voluntary basis made the sieve net an obvious alternative. Electric fishing was not an obvious choice but experience in the North Sea in the 1980s and present day commercial application in the Chinese Penaeid shrimp fishery pointed at some potential. The three options, each quite different in design and functioning, were presumed to be sufficient to find a solution to reduce discarding and were expected to be feasible within the financial and time constraints of the study. The results are given in Chapters 8,9 and 10. The present study has focused on the Belgian shrimp fishery and the author has had a major role in the application for the projects, planning of the trials, field work, data analysis and reporting throughout the course of projects. Where appropriate, the work has been done with a wider perspective. In view of the cross-border nature of the discard issue, all discard work has been carried out in an international framework. This was a necessity since shrimp trawler fleets are not only active in their own territorial waters and the consequences of their discarding habits affect target species of other fleets in other countries’ waters. Many commercial fish species migrate over long distances and discarding in one area may affect the stock in other areas. Also stock assessment is an international activity and the involvement of stock assessment expertise in this project was needed. Consequently, international co-operation in this study was not only logical but essential for the soundness, wider applicability and acceptance of the results and conclusions of the research. Some parts of the fishing gear work, i.e. trawl selectivity and electric fishing, were carried out at a national level. In view of the imminent inclusion in the technical measures of EU-legislation, however, the sorting grid and sieve net research was carried out with international co­ operation. Introduction 4-23

4 The shrimp trawler fleet

4.1 Introduction In order to obtain detailed information on the structure, characteristics and exploitation patterns of the shrimp fishery in the North Sea, a thorough inventory was carried out. The aim of the inventory was to support the discard sampling programme. Consistent data on fishing grounds, fishing effort and surface area fished are necessary for the sampling programme to extrapolate the discard sampling data to fleet level. Knowledge about the degree of variability in vessels and fishing gears is necessary to choose representative vessels for the sampling programme and to estimate the reliability of the extrapolations. This information is also useful for further technical research. The inventory was compiled in co-operation with Denmark, France, Germany, the Netherlands and the UK in the second half of 1995. The Belgian inventory was entirely carried out by the author and the staff of DvZ. The project was funded by the European Commission under the EU-study contract no. 94/044 - RESCUE - Module 1. The Belgian data are presented in detail in this report and extended with more recent data. A summary for the whole North Sea shrimp trawler fleet is added.

4.2 Materials and methods As a first step in the inventory, the Belgian shrimp trawlers had to be identified. Shrimp trawlers cannot as such be extracted from the Belgian fisheries statistics as they belong to the broader group “coastal trawlers”. These are multi-purpose trawlers that can switch between shrimp and flatfish beam trawling, otter and pair trawling, depending on the season. A number of these vessels do not carry out shrimp fishing at all. The official shrimp vessel list, set up according to EU Council Regulation 3554/90, on the other hand, is too narrow as it only contains vessels for which the shrimp directed horns fishing is higher than 50 %. For the purpose of this project it was necessary to identify all shrimp trawlers, even if only one shrimp directed trip was carried out per year, since all shrimp fishing had to be covered in the discard sampling programme. Therefore vessel selection was based on the landings data in the national database. Because only vessels with an engine power <221 kW are allowed to fish for shrimp (EU Council Regulation 55/87), the extraction of landings data was restricted to those vessels. Data on the landings in 1994 and the first half of 1995 by all Belgian vessels with an engine power of 221 kW or less were extracted from the national fishery database. All vessels that landed shrimps were identified and as a double check, these vessels were compared with the vessels in the yearly updated coastal fishery vessels lists set up according to Commission regulations 55/87 and 3554/90. This new list of shrimp trawlers formed the basis for the inventory. Before the start of the inventory, prior consultation with the fishing community was established through informal meetings with individual skippers in order to communicate the purpose of the project and to get feedback on the suggested methodology. A standard inquiry form was then compiled to be used by all partners in the project. It contained detailed questions on the vessel, the fishing gear, equipment etc. A series of port visits was made and skippers were interviewed on board their vessels. The inquiry forms were filled out based on the answers of the skippers, measurements of the fishing gears and observation of the processing equipment. The shrimp trawler fleet 4-24 All data were entered in a project database for further processing. The data were collected on a confidential basis, which was a prerequisite to ensure the good co-operation of the fishermen. Consequently, all data have been presented in a way not to be traceable to individual vessels or skippers. By definition, fishing effort is fishing capacity x fishing time. It is often expressed, however, in different units, e.g. days at sea, hours fished, hp-days etc depending on the data available. In this report, different units are used, dependent on the data source, since conversion to one standard usually is impossible.

4.3 Results The national fishery database, held by the Ministry of Agriculture, was consulted in 1995 and provided data on fishing effort, landings and vessels. This information was stored in an Excel spreadsheet for further use in the discard sampling programme. In the period March - September 1995, regular port visits were carried out for the skipper interviews. The co­ operation by the fishermen was very good for 3/4 of the vessels. The other skippers were quite reluctant to provide all detailed information for the questionnaire and only gave some basic information. It was felt that coverage of 3/i of the shrimp trawler fleet was quite adequate and that the good co-operation of the fishermen was most likely thanks to: • The Sea Fisheries Department having a long established working relationship with the shrimp fishermen, • the Sea Fisheries Department not being connected with fisheries legislation enforcement and • the assured anonymity of the survey gaining many skipper’s confidence.

4.3.1 The Belgian shrimp trawler fleet

4.3.1.1 Size and features Estimating the number of shrimp trawlers was difficult since the national fisheries statistics were incomplete for a number of vessels. A careful analysis of the data, however, together with information from the inquiry led to an estimate of 51 vessels targeting shrimps in 1995. About 25 to 30 of these vessels could be classified as genuine shrimp trawlers having shrimp as their main target species for at least 8 months a year. About 10 vessels could be classified as seasonal shrimp trawlers, targeting shrimp from August to November. The other vessels had a rather low shrimp directed fishing effort without a clear pattern and were classified as a-typical shrimp trawlers. In 2001 the number of shrimp trawlers had dropped to 37, with 16 genuine, 14 seasonal and 7 a-typical vessels. The shrimp fishery is a coastal and estuarine fishery often carried out close to the homeport (Fig. 4-1). The fishing grounds are usually sandy and free of large obstacles like stones and boulders. The homeports of the shrimp trawler fleet are Nieuwpoort, Oostende, Zeebrugge and some small ports on the river Schelde (Bouchoute, Kieldrecht, Rupelmonde). In 1995, Oostende and Zeebrugge contained the largest fleet with 19 and 18 vessels respectively, 5 vessels had Nieuwpoort as their homeport and the Schelde fleet consisted of 9 vessels. Most of the skippers usually land their catches in their homeport. There was, however, an increasing trend to land in a foreign harbour, especially when there was an important price difference. In 2001 over one third of the catches were sold in foreign harbours. Fishermen indicated Breskens or Vlissingen in the Netherlands as the most visited harbours. The fact Results 4-25 that by 2001, a number of Belgian shrimp trawlers were in the hands of Dutch companies, also contributed to this trend. The main shrimp fishing grounds for the Belgian shrimp trawler fleet are indicated as highlighted areas in Fig. 4-1. Most shrimp fishing occurred within the 6 mile zone, on the “Stroombank”, “Oostendebank”, “Wenduinebank”, “Wandelaar” and “Vlakte van de Raan”. In wintertime, however. Brown Shrimps migrate to deeper waters and then some of the fishing effort was located further from the coast. The “Westpit” and “Steenbank” were indicated by the fishermen as most visited.

North Seri

© Stroombank ©Oostendebank ©Wenduinebank ©W andelaar Belgium ©Vlakte van de Raan © W estpit ©Steenbank

Fig. 4-1 - The southeastern part of the North Sea with the shrimp fishing grounds highlighted and the banks most visited by Belgian fishermen indicated.

Seasonal fishing effort, seasonal shrimp landings, yearly fishing effort and landings of the Belgian shrimp trawler fleet are given in Fig. 4-2. A clear pattern can be distinguished in the seasonal effort and landings (Fig. 4-2 A and B) with a peak in summer and autumn for the years 1995, 1999 and 2000. In the years 1980 and 1990, a considerable amount of fishing effort was also deployed in spring. The main shrimp season, as is obvious from the landings data, falls in the months of August, September and October. A peak in total yearly fishing effort > 12000 kW hours was observed in 1975 (Fig. 4-2 C). By the end of the 1970s this had dropped below 8000 and varied between 4000 and 8000 ever since. The overall trend of the fishing effort and especially of the landings is decreasing. The yearly fishing effort of the Belgian shrimp trawlers (Fig. 4-3 E and F - expressed as hours fished per year) varied considerably from one vessel to another and ranges from less than 500 hrs to over 2000 hrs. Effort data, expressed as hours fished, are available on vessel level and will be used for the calculations in the discard sampling prograimne. The shrimp trawler fleet 4-26 Most of the Belgian shrimp trawlers are multi-purpose vessels that carry out seasonal fisheries based on catch opportunities for shrimp. Sole and Cod. In spring, when Sole is abundant on the Belgian coast (spawning season), most vessels will target this species. In that case the gear used is a flatfish beam trawl. The exact period and time span is dependent on yearly changes in conditions. In wintertime and if catch opportunities for Cod are high on the Belgian coast, this species is targeted with the otter trawl or pair trawl. Since these catch opportunities do not occur every year, it is impossible to predict the activities of the shrimp fishing fleet. Moreover, the duration of the different fishing seasons will vary considerably from one vessel to another.

A/ Seasonal fishing effort

^ 1250 .c i= 1000 w ------1990 = 750 .c g 500 1999 t 250 ë m 0

Month

B/ Seasonal landings

1980 O)« 150 c 1990 — • — 1995 1999 — • — 2000

Month

C /Yearly fishing effort and landings 16000

Effort 1000 -2 o Landings

Year

Fig. 4-2 - Fishing effort (expressed as kW hours fished) and shrimp landings of the Belgian shrimp trawler fleet (data source: ICES 2001b). Results 4-27 The average haul duration in the shrimp fishery is 1.5 hours. In wintertime, when the catch volume is low, the tow duration can be up to 3 hours. If catch volumes are excessively high or when the meshes get clogged with hydroids. the tow duration is decreased, sometimes to less than 30 minutes. In Belgium, the shrimp fishery is usually carried out during night time. The vessels mainly stay at sea for only 12 hours. In wintertime one trip can take up to 36 hours. The catchability of shrimp depends on the light intensity close to the seafloor: catchability diminishes with increasing light intensity. Consequently, day-fishing will only be successful if the visibility in the water is poor and fishermen will select fishing grounds according to this characteristic. With an average speed of 2.75 knots and an average beam length of 7.65 m a typical fishing vessel of this fleet will fish an area of 0.08 km2 in 1 hour. Knowing that about 15 % of the fishing time is used for hauling and setting the gear, this vessel will, in practice, fish a surface of 0.07 km2 in one fishing hour.

4.3.1.2 The vessel Shrimp trawlers are rather small vessels (length over all (LOA) < 22m. keel depth 2.3m) with an engine power <221 kW built to operate in an inshore fishery (Fig. 4-3).

A/ Numbers of shrimp trawlers by engine B/ Numbers of shrimp trawlers by beam power class (1995) width class (1995) 25 25 20 20 tri 15 -Q2 15 I 10 10 z 5 5 0 <147 [147-184[ [184-213[ [213-221] 7 m 8 m 9 m Engine power class (kW) Beam width (m)

C l Numbers of shrimp trawlers by age class DI Numbers of shrimp trawlers by length (1995) class (1995) 25 25 20 20 tri w d) cu XJ 15 n 15 b 10 b 10 z Z 5 5

1961 -19701971 -19801981 -1990 >1990 <16 [16-18[ [18-20[ [20-22[ >=22 Age class Length over all (m)

El Numbers of shrimp trawlers by fishing F/ Numbers of shrimp trawlers by fishing effort class -1995 effort class - 2001 25 i ------25 20 20 w tn no3 15 2 15 I 10 I 10 z Z 5 5 0 1 0 <500 [500-1000[ [1000-1500[1500-2000[ =2000 <500 [500-1000[ [1000-1500[[1500-2000[ =2000 Fishing effort (hours fished) Fishing effort (hours fished)

Fig. 4-3 - Data on the Belgian shrimp trawler fleet. The shrimp trawler fleet 4-28 As for tonnage, two units of measurement are used in the national statistics: gross registered tonnage (GRT) and gross tonnage (GT). GRT is an older unit. GT is a unit agreed upon in the Convention of Geneva in 1969 and all recent measurements are in this unit. Consequently, averages were calculated for two groups of vessels, i.e. 36 GRT and 61 GT. Most of the vessels have a steel hull and only a minority of wooden vessels were still active. For all vessels the power supply, for purposes other than propulsion, was taken from the main engine. 70 % had a secondary engine only used in case of a main engine breakdown. These secondary engines had an engine power between 4 and 15 kW. The average propeller diameter was 1.3 m, in 80 % of the cases provided with a Kort nozzle. None of the vessels, though, had controllable pitch. The fleet is quite old; many vessels date from before 1970 and between 1990 and 1995 only one new vessel was built. After 1995, however, there was a tendency for newer vessels, the so-called Eurocutters, to enter the shrimp fishery. These vessels were built to fish within the 12-mile zone with dimensions and engine power close to the upper limit allowed on these fishing grounds. At the time of the inventory, for the Eurocutters, shrimp trawling was in an experimental phase while other types of fishery were also being explored. By 1995, GPS (Global Positioning System) had become the standard for navigational purposes and replaced the less accurate Decca system. In the shrimper fleet it was used on board 90 % of the vessels. All vessels were equipped with radar, autopilot and an echosounder. Three quarters of the vessels used a track plotter for navigation and storage of fish tracks.

4.3.1.3 Deck machinery and catch processing The catch handling equipment on board the shrimp trawlers was quite uniform, and consisted of a container to collect the catch, a rotating shrimp riddle (Fig. 4-4), a washing drum, a shrimp cooker and a cooling tray. Aboard most of the vessels, catch handling follows the same procedure (Fig. 4-5). After hauling the fishing gears, the cod-end catch is collected in a container, where a continuous water flow leads it on to a conveyer belt. This belt leads to the rotating shrimp riddle where three catch fractions are separated: (a) large shrimp with small fish and invertebrates, (b) small shrimp with small fish and invertebrates and (c) by-catch and trash. These three fractions are collected in baskets. Fish to be landed are picked out manually, in the container and/or at the end of the rotating shrimp riddle. Unwanted by-catch and small shrimp are usually thrown overboard manually (surface disposal). Large shrimp are poured into the boiler. After boiling, the commercial shrimp fraction, which often contains small fish, is manually scooped and moved into a washing drum, which speeds up the cooling process and washes out the small fish. Shrimps are then collected in cooling racks, which are placed outboard of the vessel. After cooling, the shrimps are stored in baskets or boxes, on deck, until landed. If the trip takes longer than one night or if the air temperature is high, the shrimps are often stored on ice in the fish hold. Aboard the modem Eurocutters, catch handling is more automated. The unwanted by-catch is discarded through a tube which leads straight from the rotating shrimp riddle to a sub­ surface opening in the hull. The transfer of large shrimp from the rotating shrimp riddle to the boiler is, in a few cases, done by means of a tube. Scooping the shrimp out of the boiler and transfer to the cooling device is sometimes automated. The shrimp are stored on ice in plastic bags or boxes in a refrigerated fish hold. Results 4-29

1: wheelhouse 2: catch receptacle 3: converyor belt ' 4: rotating shrimp riddle 5: rotating cooling and cleaning drum 6: shrimp boiler 7: shaking riddle 8: cooling trays 9: discards: manual surface disposal

Fig. 4-4 - The rotating shrimp riddle. Fig. 4-5 - Standard deck lay-out aboard a shrimp trawler.

4.3.1.4 The shrimp beam trawl The beam trawl (Fig. 4-6) is a demersal fishing gear used to target flatfish and shrimp. The net is held open horizontally by means of a steel beam, supported at both sides by the beam trawl shoes. The construction of the net is quite simple and consists of a top and a lower panel and a cod- end where the catch accumulates. The top panel is attached to the headline, which is rigged to the beam trawl shoes. The lower panel is attached to the bobbin rope, which assures bottom contact. A fishing vessel equipped for beam trawling tows two gears simultaneously, one at each side, by means of two outrigger beams.

groundrope

Fig. 4-6 The shrimp beam trawl.

The data on the dimensions of the gear are summarised in Fig. 4-7. Average dimensions are given in ïmn and minima and maxima are given in brackets. The width of the beam trawls used in the Belgian shrimp fishery varied within the narrow range of 7 to 8 m, with one exception of 9 m. The weights of the shoe and beam were 200 kg and 260 kg respectively. Tickler chains are never used when targeting shrimp. The total weight of the gear was around 1 ton. The length of the headline was usually 20 cm shorter than the beam. The diameter of the headline was 16 ïmn for the traditional shrimpers and 25 ïmn for the Eurocutters. Headline material was polyamide (PA), polyethylene (PE) or a mixture of PE and steel (Atlas). The groundrope length varied between 8 and 11 m, depending on the length of the beam. Its diameter was 14 or 16 ïmn and the material was mixed PE with steel wire. The warp/depth The shrimp trawler fleet 4-30 ratio was 2 to 3 depending on the water depth and sea floor conditions. On softer silt grounds, this ratio will be smaller compared to the harder sandy grounds. All vessels used single warps with a diameter between 16 and 20 ïmn.

.390 (300 - 500)

140 (1 20 - 1 50) 200

A 270 (200 - 500) No, of bobbins: shrimpers: 30 cr 32 1 000 (700 - 1220) eurocutters: 37 Material bobbins: modified rubber

Fig. 4-7 - Dimensions of the shrimp beam trawl (all length measurements in mm).

The net-design was standard for practically all vessels. Most fishermen constructed the net themselves and only two skippers used a net provided by the netting industry. Fig. 4-8 shows a netplan of a typical shrimp net.

UPPER PANEL LOWER PANEL

headline: 7.80 m groundrope: 9.80 m mesh mesh size size

5 28 135 1N2B\V AN \

Fig. 4-8 - A typical shrimp net used in the Belgian shrimp fishery. Results 4-31 The standard netting material was single braided PA. The reduction in mesh size from the front to the aft part of the net and the cutting rates were similar on all vessels. The cod-end was 200 meshes long and 200 on the circumference and made of 22 mm single braided PA. At the time of the study, the minimum legal mesh opening was 20 mm. All nets were equipped with a large mesh cover (lifting bag), 50 meshes long and 50 meshes on the circumference, used for protection of the cod-end. One quarter of the vessels used a sieve net, made of 50 to 70 mm PE, for a short period of the year. This selective device was used when large quantities of jellyfish appeared in the catches.

4.3.2 The North Sea shrimp trawler fleet At the time of the inventory, the North Sea shrimp trawler fleet consisted of a total of 643 shrimp trawlers, i.e. 51 Belgian, 22 Danish, 225 Dutch, 247 German and 98 UK vessels. In France, 133 vessels targeted shrimp but only 3 of them were active in the North Sea. The main shrimp fishing grounds in the North Sea are indicated in Fig. 1-5. The total North Sea Brown Shrimp catch in 2000 amounted to 25,894 t and represented a value of approximately 80 million Euro (ICES, 2001b). In terms of value of landings, Brown Shrimp is in the top five of North Sea species. The wholesale market for landed Brown Shrimp, generally bought directly from fishermen, is dominated by a single buyer, based in the Netherlands. Starting in 1992 the fishermen of Denmark, Germany and the Netherlands agreed on effort limitation (ICES, 2000c) by means of a restriction of the number of fishing days per week. The main purpose of this regulation was to avoid over saturation of the markets and to guarantee stabilisation of the prices. It was not clear whether this measure had effectively reduced the fishing effort, because fishing most likely intensified during the open periods. Moreover, in 1997 the fishermen, together with the main processors of shrimp, agreed to restrict the total catch per week and per boat, if the shrimp supply did not match with the market demands and processing capacities. This led to weekly limits per boat in the range of It (Dec. 1999) to 6t (Oct. 1998). Without doubt the introduction of weekly quotas has supported price increases for the fishermen, especially in 1999 when the average price in Germany was about 3.2 Euro/kg, compared to a normal level of about 2.3 Euro/kg in 1998. The agreement stopped in 2000. A clear seasonal pattern in shrimp landings can be distinguished in each country (Fig. 4-9A). Except for Denmark, the main shrimp season falls in late summer and autumn. Unlike Belgium, shrimp landings in spring are also quite important in Germany, France, the Netherlands and the UK. In Denmark, the main shrimp season is spring. The yearly landings since 1975 are given in Fig. 4-9B and C. The landings for Denmark, Germany, the Netherlands and the UK have stayed rather stable in that period or even have a slight tendency to increase. For Belgium and France, on the other hand, landings have steadily decreased. Fishing effort for the national shrimp fleets is expressed in different units for each country and a common representation is hardly possible. The ICES Working Group on Crangon Fisheries and Life History (ICES, 2000c), however, made an attempt to standardise the data and came to the following conclusions: total effort was fairly stable with a mean of 10.6 million hp-days in the period 1973-81. During the period 1986-96 the total effort increased from 10.9 to 15.2 million hp-days. The corresponding time series of LPUE (Landings Per Unit of Effort) revealed overall stable LPUEs for the Netherlands and Denmark and decreasing trends for Belgium (down to 30% of the initial values) and to a lesser extent for Germany, even though Belgian LPUE was very similar to the German and Dutch levels in the early seventies. The shrimp trawler fleet 4-32

A/ Seasonal landings (average 1995 - 2000) 2500 160 140 2000 Denmark 120 Germany 1500 - - -A- - Netherlands

c Z 1000 ---- A—- Belgium .A — -A- - - -Q- - France •G- - -■Q. — -UK

jan feb mrt apr mei jun jul aug sep okt nov dec Month

B/ Yearly landings for Denmark, Germany and the Netherlands 15000

10000 — • — Denmark .. -o- - - Germany — • — Netherl. 5 5000

1975 1980 1985 1990 1995 2000 Year

C l Yearly landings for Belgium, France and the UK

2000

— * — Belgium 'O. ■ ■-ç> - - France 5 1000 —•—UK

'-•G-

1975 1980 1985 1990 1995 2000 Year

Fie. 4-9 - Seasonal and yearly landings for the North Sea (+ France) shrimp trawler fleets (data source: ICES 2001b).

Some basic data on the shrimp trawler fleets are given in Fig. 4-10. The Dutch and the Gennan fleets are by far the most important in the Brown Shrimp fishery, representing almost 3/i of the whole North Sea shrimper fleet. The Gennan shrimp trawler fleet is mainly active in the Wadden Sea. With an average age of 24 years, the Gennan fleet is rather old and renewal of the fleet is at a low level. Most of the vessels have a LOA between 14 and 20 m and an engine power between 150 and 221 kW. The width of the beam is usually 8 or 9 m. The Dutch fleet operates from the Wadden Sea, along the Dutch coast, up to the Belgian coast. The vessels’ average LOA and engine power is 20.8 m and 183 kW respectively. As for width of the beam, 8 and 9 m are the most coimnon sizes. The Dutch fleet has been far more modernized compared to the other Crangon fleets but no data were available to calculate an average age. Results 4-33

Al Number of shrimp trawlers by B/ Number of shrimp trawlers by beam engine power class (1995) width class (1995) 350 200

□ France B Netherl. □ Netherl. ¡2 200 B Belgium □ Belgium □ Denmark □ Denmark ? 150 □ Germany □ Germany ■ UK ■ UK

<150 [150- [180- [210- 23456789 10 180[ 210[ 221 [ Beam width (m) Engine power class (kW)

Cl Average towing speed (1995) D/ Number of shrimp trawlers by 250 length class (1995) 140

200

□ France B France 150 B Netherl. □ Netherl. ■Q H Belgium □ Belgium □ Denmark = 60 □ Denmark Z 100 □ Germany □ Germany ■ UK ■ UK

IT 'SoOOÍNsrCDOOOÍNsrCDfS “ OOÓCN^-CDOÓÓCN^-

Towing speed (knots) Length over all (m)

Fig. 4-10 - Data on the North Sea (+ France) shrimp trawler fleet.

The Danish shrimp trawlers fish along the Danish west coast with vessels of an average age of 27 years and an average engine power of 200 kW. Most of the vessels have a LOA between 14 and 18 m. The width of the beam is 9 or 10 m. Typical for this fishery is the enforced use of a selective sieve net. In the UK, the east coast shrimp fishery is mainly carried out in the area of The Wash and the Humber estuary. Most of the vessels have been built in the 1980s, a time with a remarkable effort increase in that fishery. The average LOA and engine power is 11 m and 106 kW. The width of the beam is usually below 7 m. Although twin beam trawling for shrimp has become standard, in the UK 15 % of the vessels The shrimp trawler fleet 4-34 still operate a single beam trawl. The French fleet mainly consisted of small vessels (average LOA 10 m) with low engine power (mostly < 150 kW) and a low towing speed (Fig. 4-10). In contrast to the other fleets, the most common fishing gear for targeting shrimps was the otter trawl. The majority of the vessels used a selective device in the trawl similar to the sieve net, called “châlut asselin”. Belgium is the only country to report by-landings of white fish of any significance.

4.4 Discussion Good co-operation with the fisheries sector was very important for drawing up the inventory of the fleet as most of the information was obtained from the skipper interviews. For the discards sampling programme, which had to be carried out aboard commercial vessels, co­ operation was also essential. The way fishermen are approached in a project with a very sensitive subject like discarding is critical for the co-operation of the sector. The Dutch partner in the project experienced the consequences of a wrong approach. The rumour spread amongst the fishing community that a shrimp fishery discards project was going on before the institute informed the fishermen. Later, the project was presented in a large-scale meeting between scientists and fishermen and a video “Juvenile Fish Killers” was shown. Skippers were not contacted personally for the subsequent interview. Instead, the inquiry form was sent by post to be cost effective. The fishermen disliked the content of the project and the way it was presented and refused further co-operation. Consequently, little information came from the skipper interviews and the discard sampling programme had to be cancelled. The Dutch partner produced no data. Long term contacts, respect and an open mind towards the problems in the sector are key factors for good collaboration. The Belgian shrimp trawler fleet is quite heterogeneous in terms of seasonality of the shrimp fishery. Three groups emerged, i.e. the genuine shrimp trawlers, the seasonal shrimp trawlers and the a-typical shrimp trawlers. Although the landings statistics indicated that most of the coastal fishing fleet targeted shrimp for a period of the year, only a limited number of vessels spent all their time on shrimp. Many skippers also fished for other species using other trawls on a seasonal basis, using the opportunities of high abundance of certain species in certain periods of the year. This was also the case for many genuine shrimp trawlers. The results from the interviews showed that the a-typical shrimp trawlers used shrimp as an escape route. Their skippers chose to target shrimp when quotas for their main target species were exhausted or because their vessels were too small to follow the high abundance of fish to distant fishing grounds. Independent of the type of shrimp trawler, however, the main shrimp season (August - October) attracted almost all vessels to this fishery. In the period December to June, shrimp directed fishing effort is very low. Before the 1990s, though, the period April to June was quite important for the shrimp fishery and fishing effort was not considerably lower compared to the real shrimp season. The reason why this pattern has changed is unclear. The fisheries statistics for 2001 showed that between 1995 and 2001 the shrimp trawler fleet shrank from 51 vessels to 37. Most striking was the drop in the number of genuine shrimp trawlers, from more than 25 to 16 vessels. In 2001 the genuine shrimp trawlers had almost completely disappeared from the Zeebrugge fleet. Eurocutters, carrying out a seasonal shrimp fishery, had replaced them. These vessels are equipped with a refrigerated fish hold and stay at sea for several days. They land their catches for the national and international market. Apparently, this fleet has followed the modernisation trend as observed in the Dutch so called “large fishery” fleet (van Marlen et al., 1997b). This fleet consists of modem vessels operating on the in- and offshore shrimp grounds and making sea trips of more than one day. The fact that a number of these vessels are Dutch “flag ships” (owned by Dutch Discussion 4-35 companies but fishing with a Belgian licence) probably contributed to or initiated this trend. Oostende and Nieuwpoort, on the other hand, preserved a more traditional shrimp fishery. Older vessels mostly go at sea for one night only and sell their catches at the local fish market. The general design of the shrimp fishing gear was quite similar for all Belgian shrimp trawlers. The typical features of a shrimp beam trawl, i.e. the steel construction to assure the opening of the net, the bobbin rope, the rather short length of the groundrope and the design of the net were almost identical for all vessels. Also the characteristics of the vessels varied in a rather narrow range, especially if compared with other fleet sectors like the flatfish beam trawlers (Lindeboom and de Groot, 1998). Consequently, based on the inventory, no sub­ fleets could be detected in the Belgian shrimp trawler fleet. This was an important conclusion for the selection of vessels to cooperate in the discard sampling programme (see Section 5). Instead of selecting vessels based on vessel and gear characteristics to cover all vessel types in the fleet, more attention could be given to “expected co-operation of the fishermen” which proved to be extremely important for the success of the programme. The information collected from the existing databases and from the interviews allowed to provide all necessary information for the subsequent discard-sampling programme (see Section 5). The choice of the vessels, fishing gears, fishing grounds, timing of the sampling and the calculation of surfaces fished could now be based on sound data. The information on the fishing gear was also deemed useful for the selectivity experiments later in the project (see Section 7-10). All Belgian shrimp trawlers are equipped with a rotating shrimp riddle, allowing quick processing of the catch. Compared to the old shaking sieve, survival of discards improves with this machinery (Berghahn and Vorberg, 1997, Boddeke, 1989). Aboard most of the vessels, however, the discards are collected in baskets before returning them to the sea. The time between hauling the catch and disposal of discards can be considerable. Especially when large catches have to be handled, this can be more than 30 minutes for some part of the discards, resulting in a higher mortality. The negative effect on the survival of discards of a long handling time on board was observed by Neilson et al. (1989), Van Beek et al. (1990) and Thorsteinsson (1995). Discard mortality on modem shrimp trawlers with immediate sub­ surface disposal of discards will thus be lower. The landings statistics demonstrate that the North Sea Brown Shrimp fishery is in good condition. Overall catch levels have risen during the last decade and are at a historically high level and overall LPUEs have remained stable. In Belgium (and France - ICES, 2001b), however, catch levels are historically low and have gradually declined since the 1970s. The fact that the coastal fishing fleet has diminished in that period and fishing effort decreased, has certainly contributed to this decline. The LPUEs, though, have also shown a gradual decline (ICES, 2000c); 0.125 kg/hp hour in 1973-82 compared to 0.075 in the 1990s. This observation may point at a more fundamental trend of decreasing densities of Brown Shrimp in the Southern North Sea. Possibly, the reduction of nursery areas in the Scheldt estuary due to reclamation of land may have reduced the reproductive potential of the species. Since 1950, the wetlands in this area have reduced by roughly 50 km2 (Redant, F., pers. comm.). The Belgian coast is near the southern boundary of the area where Brown Shrimp is available in exploitable quantities. Observations by fishermen support the idea of a northward movement of the southern limit of the population. Older fishermen claim that 30 to 40 years ago, high densities of Brown Shrimps occurred in the western part of the Belgian continental shelf and good catches could be obtained there (pers. comm, with fishermen). Nowadays, no vessels target shrimps in that area since few shrimps are to be caught. It may therefore be The shrimp trawler fleet 4-36 possible that the southern boundary of the Brown Shrimp population is moving northward, thus reducing the fishing grounds for Belgian shrimp fishermen. This observation is, however, highly speculative since no research has been done to support it. Some support, though, is found in research on the relation between abundance and climate change for other species. O’Brien et al. (2000) state that the stock of North Sea Cod is threatened by a decline in the production of young Cod that has paralleled warming of the North Sea over the past ten years. Other studies have come to a similar conclusion (Cook et al., 1997; Attrill and Power, 2002). Dippner (1997) found a relation between climate variability and recruitment of Cod, saithe and Whiting in the North Sea. Klyashtorin (1998) found good correlations between stock dynamics of the main commercial fish species and climate cycles.

4.5 Conclusions Crangon crangon is caught in the shallow coastal waters and estuaries in the North Sea. This crustacean is caught by trawlers, generally rigged for twin beaming. The trawl is a simple beam trawl with a metal frame, assuring the horizontal and vertical net opening, and a net made of polyamide netting. The design of the net is quite similar over the whole shrimp trawler fleet. The cod-end mesh size usually is 22 mm. Due to the small mesh size necessary to catch the shrimps and the location of the fishing grounds, there is a significant by-catch of juvenile finfish and invertebrates. The shrimp catches are sorted on deck using mechanical riddles and subsequently cooked onboard. The unwanted by-catch is discarded overboard and may suffer from high mortality. Belgium was the only country to report by-landings of white fish of any significance. The Belgian shrimp fishery is a typical seasonal fishery with peak landings and effort in the period August-October, although part of the fleet target shrimps all year round. The shrimp trawler fleet is quite old and the vessels have quite similar characteristics. In 1995, there were respectively 51 and almost 650 vessels engaged in the Belgian and North SeaCrangon fisheries. About 75% of the fleet was based in Germany and the Netherlands. The North Sea Brown Shrimp fishery is in a good condition although the statistics indicate declining catch and effort levels for the Belgian fleet. The LPUE’s have shown to be quite stable for the Netherlands and Denmark and decreasing in Belgium (down to 30% of the initial values) and to a lesser extent also decreasing for Germany, even though Belgian LPUE’s were very similar to the German and Dutch levels in the early seventies. The information collected on fishing effort and landings, vessels and operational characteristics of the fleet, fishing gear and fishing grounds were a sound basis for the further activities in the project such as the discard sampling programme (see Section 5) and the selectivity experiments (see Section 7-10). Introduction 5-37

5 Discarding by the shrimp trawler fleet

5.1 Introduction In order to fill the gap in knowledge on discarding practices in the North Sea Brown Shrimp fishery, a number of institutes in the major Brown Shrimp fishing nations agreed to identity the magnitude of the problem in a cooperative discard sampling programme. In the early nineties, much debate was going on in the scientific world and in the media about the significance of the discarding problem in this fishery, although little basic data were available. There was also strong pressure to enforce the use of selective fishing gear or to close certain areas for fishing, driven by concerns that were not based on sound scientific data. A North Sea wide quantitative measurement of discard rates was thus required. It was deemed logical to do this before seeking technical methods to reduce discarding. The aim of the discard sampling programme was to estimate the order of magnitude of the numbers of juvenile fish and undersized Brown Shrimp discarded by the North Sea shrimper fleets. In the EU-funded project RESCUE - Module 2, an extensive sampling programme was set up with co-operation of the following institutes: • Bundesforschungsanstalt für Fischerei, Hamburg, Germany, • Departement voor Zeevisserij, Belgium, • DIFMAR, Charlottenlund, Denmark, • IFREMER, Boulogne sur mer, France, • Rijksinstituut voor Visserij Onderzoek, IJmuiden, Netherlands, • University of Humberside, Grimsby, UK. The French partner estimated the discard problem in the French shrimp fishery to be insignificant compared to the effort needed to organise the sea trips and did not cooperate in the sampling programme. The Dutch partner was only allowed on board Dutch shrimp trawlers for a limited number of sea trips (see section 4.4) and consequently, too few data were delivered. The results section first addresses the Belgian situation and then gives a summary of the results for the whole of the North Sea. The Belgian part of the programme was entirely carried out by the author and the staff of DvZ.

5.2 Materials and methods

5.2.1 General plan The sampling scheme was programmed to cover a period of one year. The start of the scheme, the first quarter of 1996, was determined by the starting date of the project. Since both the effort deployed in this fishery and the magnitude of the landings are quite dependent on the season, sampling was more concentrated in the main shrimp season, i.e. the third and fourth quarter in the case of Belgium. Each participating institute selected a number of shrimp trawlers representative of the fishery. Scientific crews boarded the vessels according to a sampling schedule for in situ sampling and measurement of the catches. The data collected were stored in a database and extrapolated to cover the whole fleet on a quarterly basis. The sampling procedure, the analysis of the catches and the data and the presentation of results were, as far as possible, standardised for all partners in the project. This facilitated comparability of the data and enhanced the clarity of the report. The standardisation was Discarding by the shrimp trawler fleet 5-38 especially necessary since each nation’s data were to be used in a biological and economic modelling exercise (see Section 6).

5.2.2 Planning of the sea trials The fleet inventory (see Section 4) demonstrated that shrimp fishing was mainly carried out by traditional shrimp trawlers. More modem vessels, the so-called Eurocutters, on the other hand, were exploring the possibilities of shrimp trawling at the time of the inventory and entered the fishery irregularly. It was difficult to predict whether these vessels would become a constant factor in the shrimp fishery. Consequently, no such vessels were selected for the sampling programme. Since the characteristics of the Belgian shrimp trawlers and their gears were quite homogeneous, no sub-fleets had to be defined prior to vessel selection for discard sampling. This selection could therefore be done on the basis of expected co-operation of the skipper and the crew, which was considered essential for the success of the project. The degree to which fishermen were expected to cooperate was estimated by observation of the attitude towards the project during the skipper interviews (see Section 4) and by previous experience. The vessels selected for the sampling scheme are given in Table 5-1. The sampling was exclusively on commercial vessels, carrying out a normal commercial fishery. The scientific crew did not influence the timing of the hauls, the choice of fishing grounds and the routine on board.

Table 5-1: Shrimp trawlers selected for the Belgian discard sampling programme Vessel Engine power (kW) LOA (m) Year built Beam length (m) N 64 150 19.6 1989 8 0 101 190 16.8 1967 8 0 225 190 20.6 1957 7 0 455 184 18.7 1964 8 0 700 182 16.8 1968 8 Z 403 221 21 1961 8 Z 582 221 21 1961 8

Before the start of the sea trials, the project team decided upon a sampling schedule for each partner. After a budget cut during the project contract negotiations with the EC, it was clear that finances would be the limiting factor for the sea trials. The vessel charter cost and the labour cost for embarking crew were a heavy burden on the restricted budget. The project team was aware that catch composition can be very variable. According to the experience of each partner, it was clear that the absence or presence of a certain fish species in the catch depended on the season, the time of the day, tide, fishing ground etc. and catch composition can vary strongly from one haul to the next. It was anticipated that the results would show rather wide confidence bands and the partners were urged to obtain the maximum sampling intensity with the available means. The following schedule was set up for the Belgian partner: • Quarter 1: 3 sea trips • Quarter 2: 3 sea trips • Quarter 3: 6 sea trips • Quarter 4: 6 sea trips Materials and methods 5-39 Prior to the start of the sampling, the project partners agreed upon a mandatory list of species to be measured, in both the landings and the discards. Apart from Brown Shrimp, this list comprised the following species: • Bib and Poor Cod - Trisopterus spp. (not differentiated - in the text referred to as bib) • Cod - Gadus morhua • Whiting - Merlangius merlangus • Gurnards - Trigla spp. (not differentiated) • Turbot - Psetta maxima • Brill - Scophthalmus rhombus • Flounder - Platichtys flesus • Plaice - Pleuronectes platessa • Dab - Limanda limanda • Sole - Solea solea • Solenette - Buglossidium luteum The following data were recorded for each fishing haul: vessel identification, date, fishing ground and depth, time of shooting and hauling the gear, towing speed, wind direction, wind speed and sea state.

5.2.3 Sampling and analysis of the catch After each fishing haul, the cod-end catches were emptied into the catch container on deck and processed through a rotating shrimp riddle (Fig. 4-4) to separate Brown Shrimp from the large by-catch and sub divide the Brown Shrimps into commercial and non-commercial sizes. The number of animals in the catch was usually very high and sub-sampling was inevitable. All Belgian shrimp trawlers were equipped with a rotating shrimp riddle. This gave the opportunity to take samples from the different fractions of the sorted catch instead of taking one sample from the unsorted catch. The advantage was that sorting time was reduced and that the number of animals of a certain species in a sample was less dependent on the total catch composition, which improved the accuracy of the estimates. After sorting, the catch was split into three fractions: • the main by-catch containing the larger fishes, crabs, starfish, debris, etc., • the commercial shrimps containing also 0- and 1-group flatfish together with occasional juvenile roundfish and • the non-commercial shrimps also with the smallest flatfish. The volume of each catch fraction was recorded and samples were taken. The fish in each sample were sorted out and measured immediately after the haul. All fish were measured for total length (TL) to the cm below, from the tip of the nose to the tip of the stretched tail fin. This method was preferred over standard length (i.e. from the tip of the nose to the fork of the tail fin), to be in accordance with the method used to measure fish for their compliance with minimum landing size regulations. A sample of 1.5 1 of Brown Shrimps was taken from the commercial and from the non­ commercial shrimp fractions for later analysis in the laboratory. There, the total length of the shrimps was measured to the mm below from the tip of the scaphocerite to the distal margin of the fans on the stretched uropods. Following the results of a theoretical study on the effect of sample sizes on the estimation of selection parameters for shrimp trawl cod-ends (Polet and Redant, 1999), it was decided to measure at least 250 shrimps per catch fraction. Discarding by the shrimp trawler fleet 5-40 For the sizes of the samples, some general guidelines were agreed upon. On board, however, this was usually left to the experience of the scientific team and was dependent on the abundance of the different fish species in the samples. If the standard sample did not contain sufficient numbers of fish to produce acceptable length frequency distributions, one or more extra samples were analysed, albeit only for those species where the first sample did not suffice. It was considered more worthwhile to direct most of the effort to the commonest and/or commercially most important species (such as Cod, Whiting, Plaice, Dab and Sole), rather than to spend several hours sorting baskets in an attempt to obtain a few extra measurements for the rarest species (such as Turbot and Brill).

5.2.4 Data analysis To get a first idea of the amounts of discards in the shrimp fishery, the volumes of the three catch fractions, i.e. commercial shrimps, discard shrimps and main by-catch, were analysed. The mean percentage (+ 95% confidence limits) of each fraction in the total catch was calculated over all hauls. The discard ratio was calculated as the ratio of the discarded catch to the total catch. Depending on the normality of the data, the arithmetic mean and corresponding confidence limits or the median and the upper and lower quartiles were used as descriptive statistics. The standard procedure to raise the original numbers-at-length in the samples to quarterly and yearly totals of the numbers of fish and shrimp caught, landed and discarded by the national shrimp trawler fleet, can be summarised as follows: • The numbers-at-length in the samples were first raised to numbers-at-length in the three catch fractions using the ratios between the volumes of the catch fractions and the volumes of the corresponding samples as raising factors. These numbers were then summed across catch fractions, to obtain the total numbers-at-length caught, landed and discarded for each haul separately. • For each fishing haul, the swept area (total surface fished) was calculated as the product of the haul duration, the average towing speed and the beam length and expressed in units of 10,000 m2. The catch data, as absolute numbers-at-length caught per haul, were converted to numbers caught per 10,000 m2. These length frequency data were grouped for all hauls carried out during a certain quarter of the year or the whole year and averaged accordingly. This gave quarterly and yearly figures for the numbers caught, landed and discarded per unit of swept area. • Finally, the numbers per unit of swept area were raised to the national fleet level, by means of quarterly and yearly effort data for shrimp directed fishing (i.e. the product of the total number of hours fished * average towing speed * average beam length expressed in 10,000 m2 swept). The frequency distributions of the numbers of fish or shrimps caught over a series of hauls (by quarter or over the whole year) were checked for normality by visual inspection, testing for skewness and kurtosis and testing with the Kolmogorov-Smimov and Shapiro-Wilk tests. Since the numbers of fish and shrimp caught over a series of hauls (by quarter or the whole year) often showed a highly skewed (asymmetric) or bi-modal frequency distribution (see “Results” section), the arithmetical mean and its confidence limits were unreliable. The outliers were considered valid observation. The presence of outliers is inherent to the nature of fishery and the often patchy distribution of fish on the fishing grounds. Since a technique was needed to obtain confidence limits for the data without knowledge on the underlying distribution, it was decided to use an alternative. If one is sceptical to the theoretical Materials and methods 5-41 distribution, empirical-based re-sampling can be a good alternative. Since this technique was beyond the expertise of the author, advice was sought at the statistical department of the University of Humberside (partner in the project) and fisheries statisticians Rob Fryer (Marine Laboratory, Aberdeen, UK) and Cari O’Brien (CEFAS, Lowestoft, UK) for the methodology. It was decided to bootstrap the datasets. In bootstrapping, the data are used as an approximation to the population density function and the data are re-sampled with replacement from the observed sample to create an empirical distribution for the statistic under consideration. The project team was provided with a software package (written in Delphi) calculating the upper and lower 5% percentiles and median for the arithmetic mean from a number of 1,000 repetitions. This technique was applied to the sets of data of discard numbers / 10,000m2 fished, by species, age group, fleet and quarter. In view of the model work to calculate the actual impact of the shrimp fishery on the recruitment of commercial round- and flatfish species, to be carried out in a later stage in the project, the data were also given as numbers-at-age caught. Therefore, the length frequency data were converted to age groups by deconvolution, i.e. superposition of normal distributions that represent the different age classes. Since the numbers of older fish in the catches were usually very small, it was decided to distinguish three age classes, viz. the 0- group, the 1-group and the 2+-group. Deconvolution was done on a quarterly basis for the time of the year when fish growth is slow or stagnant (winter and spring) and on a monthly basis when growth is fastest (summer and autumn). The split into age classes was done only for the commercially important species for which reliable stock assessment parameters were available, i.e. Whiting, Cod, Plaice, Sole and Dab. For the species having a MLS (Minimum Landing Size) in the North Sea, i.e. Cod, Whiting, Plaice and Sole, the catch data were also split into numbers > MLS and < MLS. For Dab and Flounder, species that are of commercial value to the shrimp fishery, no MLS is set by legislation. Based on the lengths usually landed by the fishermen, however, a “reality MLS” of 23 cm was attributed to these species. This was done for clarity in the presentation of the data and to indicate approximately what part of the catch is landed. At the time of the sampling programme, the MLS for Whiting was 23 cm. All further calculations were based on this number. According to the Council Regulation 850/98, however, the MLS for Whiting has been changed to 27 cm. For shrimp, the data were presented as number of animals discarded, number landed and the total by quarter. The results are given as number / 10,000 m2 and number and weight extrapolated to the fleet level. The weight of the shrimps caught was calculated by a conversion function:

W = 3.212 * IO'6 * T L3178 (Redant, 1978)

• W = fresh weight of a Brown Shrimp (g) • TL = total length of a Brown Shrimp (cm) Discarding by the shrimp trawler fleet 5-42 5.3 Results

5.3.1 National data

5.3.1.1 General A total of 21 sea trips and 108 fishing hauls were carried out for the Belgian share of the discard sampling prograimne. Details are given in Table 5-2.

Table 5-2: Overview of the sea trips carried out for the discard sampling programme

Quarter Trip No. Average Total time of Towing Beam Total surface duration tows tow measured speed length sampled duration hauls (knots) (m) (10.000m2) (minutes) (hours) 1996 01 24 hours 6 122 12.2 3.0 16 108 36 hours 6 180 18.0 3.0 16 160 Q2 1 night 5 87 7.3 2.7 16 58 1 night 5 105 8.8 2.9 16 75 1 night 5 69 5.8 3.3 16 56 Q3 1 night 5 71 5.9 2.9 16 51 1 night 5 82 6.8 3.0 14 53 1 night 6 84 8.4 3.0 16 75 1 night 6 89 8.9 3.0 16 79 1 night 2 78 2.6 3.0 16 23 1 night 6 67 6.7 3.6 16 71 1 night 2 65 2.2 2.8 16 18 1 night 6 78 7.8 3.7 16 87 Q4 1 night 6 95 9.5 3.0 16 84 1 night 5 64 5.3 3.1 16 49 1 night 6 99 9.9 3.4 16 100 1 night 5 100 8.3 3.6 16 89 1 night 3 72 3.6 3.4 16 36 1997 01 48 hours 8 158 21.0 3.0 16 185 24 hours 5 138 11.5 3.0 16 102 1 night 5 92 7.7 3.2 16 73 Total 21 trips 108 178 hours 1632

The trials started in Q1 (first quarter) of 1996. In Denmark and Germany, however, the sampling prograimne in Q1 encountered difficulties due to the long winter and icy conditions. Therefore, it was decided not to use the incomplete dataset for Q1 of 1996 and extend the program to Q1 of 1997. The seasonal coverage was good, with most of the sampling effort being concentrated in Q3 and Q4. Due to continuing bad weather conditions in Q4, however, only 5 trips instead of the planned 6 have been completed. It was anticipated, though, that bad weather conditions in Results 5-43 Q4 might cause problems and therefore, the team took advantage of good weather conditions in the second half of September to carry out two extra trips to compensate for possible deficiencies in Q4. The objectives, with respect to the minimum numbers of fish and shrimps to be measured, were largely met. Length and age frequency distributions with a low degree of scatter were obtained for all species except for the rarest (such as Turbot, Brill and Solenette). The numbers caught of a certain species varied strongly from one haul to another and from one trip to another. The average percentage over all hauls (95% confidence limits between brackets), of each catch fraction in the total catch was 45.1% (42.3% - 47.8%), 26.3% (24.1% - 28.5%) and 28.7% (26.3% - 31.0%) for the main by-catch, discard shrimp and coimnercial shrimp fractions respectively. For the frequency distribution of the discard ratios, a visual inspection and the normality tests indicated normality of the data. The mean discard ratio was 71.3 % (68.9%-73.7%) (Fig. 5-1) with a minimum and maximum of 37.1% and 99.0% respectively.

74 ' ------

gT 72 (/>q % "O S 70 Q

68 * * ------

Fig. 5-1 - Box and whisker plot for the haul by haul discard ratios observed during the Belgian discard sampling programme. The central dot gives the mean, the box presents the standard deviation and the whiskers the 95% confidence intervals.

5.3.1.2 Fish The length frequency data for the fish species that were observed in sufficient numbers are given in Fig. 5.2. The fishing effort deployed by the Belgian shrimp trawler fleet, expressed as hours fished and the swept area is given in Table 5.3. The numbers of fish caught, by species and by quarter, per 10,000 m2 and raised to fleet level are given in Table 5-4. This table also gives the proportion of fish < MLS per species and quarter and the pooled data for all quarters. The frequency distributions of the numbers of fish caught / 10,000 m2 (for each species, over a series of hauls by quarter or over the whole year) were, however, not normally distributed. This was obvious from a visual inspection of the data (e.g. Fig. 5-3). The skewness of the distributions was usually very high and significant and the normality tests indicated non-normality for each case. The outliers were considered valid, a measurement of a normal natural phenomenon, and not due to any error. In order to overcome the problem of a strong deviation from normality and to obtain confidence intervals necessary for the modelling work, the data were bootstrapped. The Discarding by the shrimp trawler fleet 5-44 upper and lower 5% percentiles and median for the arithmetic mean from a number of 1,000 repetitions in each bootstrap procedure are given in Table 5-4.

Table 5-3: Fishing effort by the Belgian shrimp trawler fleet

Month Hours fished Hours fished Swept area monthly quarterly (10,000 m2) April '96 1814 May '96 1579 June '96 3477 02 6870 59099 July '96 3328 August '96 4131 September '96 4597 03 12056 103712 October '96 5717 November '96 2988 December '96 3227 0 1 11932 102645 January '97 1143 February '97 20 March '97 935 0 1 2098 18052 Results for 8 selected species - Q uarter 2, 3, 4 in 1996 and quarter 1 in 1997 - as observed in the Belgian Belgian the in observed as - 1997 in 1 fleet ler quarter traw and 1996 shrimp in 4 Belgian 3, 2, the m2 in 10,000 / uarter fish Q - discarded species for selected data 8 for frequency Length - 5-2 Fig. discard sampling programme. Vertical black line indicates MLS. indicates line black Vertical programme. sampling discard Q u arter4 1996 Q u arte r3 -1996 Q u arte r2 -1996 Quarterl 1997 Q u arter4 1996 Q u arte r3 -1996 Q u arte r2 -1996 Quarterl 1997 X-axis: Lengthclass (cm) Y-axis: Numbers / 10,000 m2 10,000 Y-axis:Numbers / (cm) Lengthclass X-axis: MLS (Minimum landing size) indicated as vertical black line (cod: MLS = 45, bib, poor cod, gurnards and flounder: no MLS)no andflounder: gurnards cod, poor bib,MLS45, (cod: = line black as vertical indicated landing size) MLS(Minimum 0.0 0.1 0.2 0.3 6.0 0.0 0.1 0.2 0.3 0.1 - 0.2- 0.3 0.1 - 0.2- 0.3 0.0 2.0 4.0 6.0 0.0 2.0 4.0 6.0 0.0 2.0 4.0 6.0 0.0 2.0 4.0 Bib and poor cod poor and Bib e t 1. 2. 30.0 20.0 10.0 0 I 30.0 20.0 10.0 0 30.0 20.0 10.0 0 30.0 20.0 10.0 0 30.0 20.0 10.0 0 30.0 20.0 10.0 0 30.0 20.0 10.0 0 30.0 20.0 10.0 0 30.0 20.0 10.0 0 3 n |J Flounder 1 00 00 30.0 20.0 10.0 30.0 20.0 10.0 00 00 30.0 20.0 10.0 ____ 0.0 0.0 0.0 0.0 2.0 2.0 3.0 2.0 3.0 2.0 3.0 2.0 3.0 0.0 0.2 0.4 0.6 0.0 0.2 0.4 0.6 0.0 0.2 0.4 0.6 0.0 0.2 0.4 0.6 1.0 1.0 1.0 1.0

i at 00 00 30.0 20.0 10.0 30.0 20.0 10.0 30.0 20.0 10.0 00 00 30.0 20.0 10.0 Plaice i , Cod A , 0.0- 0.0 2.0 4.0 6.0 0.0 2.0 4.0 6.0 0.0 3.0 6.0 9.0 0.0 3.0 6.0 9.0 0.0 3.0 6.0 9.0 0.0 3.0 6.0 9.0 0.0 2.0 4.0 6.0 m 1. 2. 30.0 20.0 10.0 0 30.0 20.0 10.0 0 30.0 20.0 10.0 0 30.0 20.0 10.0 0 3 I hiting W \ \ 00 20.C 10.0 30.0 20.0 10.0 30.0 20.0 10.0 30.0 20.0 10.0 Dab , 1 L- .1 0 : “ - i 30.0 0.5 0.0 0.0 0.1 0.1 0.0 0.0 0.1 0.1 0.0 0.0 0.1 0.1 0.0 0.5 0.0 0.5 1.0 1.5 1.0 1.5 1.0 1.5 É 1. 2. 30.0 20.0 10.0 0 30.0 20.0 10.0 0 30.0 20.0 10.0 0 30.0 20.0 10.0 0 3 3 00 00 30.0 20.0 10.0 urnards G 1 00 00 30.0 20.0 10.0 30.0 20.0 10.0 30.0 20.0 10.0 I Sole ------E E 5-45 Discarding by the shrimp trawler fleet 5-46 Table 5-4: Numbers of fish caught during the Belgian discard sampling programme, split up by quarter and by species; nos / 10,000 m2 and nos raised to shrimp trawler fleet level. The numbers caught are presented as the median and the 5% and 95% percentiles as obtained from the bootstrap procedure. Percentage fish caught < MLS are given for the species having a MLS according to the EU Council Regulation 345/92. Nos split up by age group are given for the species for which reliable stock assessment parameters were available for the modelling work. Q1 -1997 Whiting Cod Bib Age o Age 1 Age 2+ All ages Ageo Age 1 Age 2+ All ages All ages Numbers of fish caught / 10,000 m 1 5% percentile 012 0.12 0 026 0 2 53 1.55 4.27 007 Median 0.17 0.19 0 0.36 0 3.62 4.7 8.21 0.23 95% percentile 022 028 0 047 0 4 77 927 1405 0 46

Numbers caught by the whole Belgian shrimp trawler fleet 5% percentile 2,166 2.166 0 4,694 0 45,672 27.981 77,082 1,264 Median 3,069 3,430 0 6,499 0 65,348 84,844 148,207 4,152 95% percentile 3,971 5.055 0 8,484 0 86,108 167.342 253,630 8,304 % < MLS 67% 99%

Q 2-1996 Whiting Cod Bib Age o Age 1 Age 2+ All ages Age o Age 1 Age 2+ All ages All ages Numbers of fish caught / 10,000 m 1 5% percentile 052 10.27 0.13 11.13 0 0 0 0 743 Median 1.78 16.83 0.27 18.96 0 0 0.01 0.01 10.91 95% percentile 3.72 24.88 0 45 28.72 0 0 0.03 0.03 14 52

Numbers caught by the whole Belgian shrimp trawler fleet 5% percentile 30.732 606.949 7.683 657.774 0 0 0 0 439.107 Median 105,197 994,639 15,957 1,120,520 0 0 591 591 644,772 95% percentile 219,849 1.470.387 26.595 1.697,328 0 0 1.773 1.773 858,120 % < MLS 58% 76%

Q 3-1996 Whiting Cod Bib Age o Age 1 Age 2+ All ages Age o Age 1 Age 2+ All ages All ages Numbers of fish caught / 10,000 m* 5% percentile 41 31 0.47 0 41.35 0 0 05 0 005 2557 Median 59.62 0.83 0.05 59.18 0 0.15 0 0.15 34.4 95% percentile 87 57 1.29 0.13 88 03 0 027 0 0.27 44 88

Numbers caught by the whole Belgian shrimp trawler fleet 5% percentile 4,284.332 48.745 0 4.288,480 0 5,186 0 5,186 2,651.909 Median 6,183,294 86,081 5,186 6,137,661 0 15,557 0 15,557 3,567,684 95% percentile 9.082.037 133.788 13,483 9,129.744 0 28.002 0 28.002 4.654,583 % < MLS 99% 84%

Q 4-1996 Whiting Cod Bib Age o Age 1 Age 2+ All ages Age o Age 1 Age 2+ All ages All ages Numbers of fish caught /10,000 m* 5% percentile 14 55 0.87 0 15 65 0 1 82 001 19 4 96 Median 19.79 1.49 0.03 21.31 0 3.34 0.03 3.37 10.14 95% percentile 253 2.24 008 27 63 0 4 97 0.07 497 16 87

Numbers caught by the whole Belgian shrimp trawler fleet 5% percentile 1.493,485 89.301 0 1,606,395 0 186,814 1.026 195.026 509,119 Median 2,031,345 152.941 3,079 2,187.366 0 342,834 3,079 345,914 1,040,821 95% percentile 2.596.919 229.925 8,212 2.836.082 0 510,146 7,185 510.146 1,731.622 % < MLS 54% 99%

Total (Q1+2+3+4) Whiting Cod Bib Age o Age 1 Age 2+ All ages Age o Age 1 Age 2+ Ali ages All ages Numbers o f fish caught / 10,000 m 1 5% percentile 21.89 2.81 0.03 24.70 0.00 0.90 0.11 1.04 13.57 Median 31.35 4.66 0.09 35.61 0.00 1.60 0.33 1.92 19.81 95% percentile 44 84 6.93 0.18 51.50 0.00 2 35 0.66 2.99 27 32

Numbers caught by the whole Belgian shrimp trawler fleet 5% percentile 5.810.715 747.160 7.683 6.557,343 0 237,671 29.007 277.293 3.601.399 Median 8,322,904 1,237,091 24,222 9,452,045 0 423,739 88,515 510,268 5,257,428 95% percentile 11,902.777 1.839 155 48,289 13.671,639 0 624,256 176,300 793.551 7.252 628 % < MLS 84% 98% Results 5-47 Table 5-4 (continued): Numbers of fish caught during the Belgian discard sampling programme, split up by quarter and by species; nos / 10,000 m2 and nos raised to shrimp trawler fleet level. The numbers caught are presented as the median and the 5% and 95% percentiles as obtained from the bootstrap procedure. Percentage fish caught < MLS are given for the species having a MLS according to the EU Council Regulation 345/92. Nos split up by age group are given for the species for which reliable stock assessment parameters were available for the modelling work.

Q1 -1997 Gurnards Dab Plaice All ages Age o Age 1 Age 2+ All ages Age o Age 1 Age 2+ All ages Numbers of fish caught / 10,000 m ‘ 5% percentile 0 0 4.5 5 08 10.58 0 7.11 0.59 772 Median 0 0 6.5 6.96 13.72 0 10.14 0.86 11.01 95% percentile 0 0 9.2 8 77 16.85 0 13 89 1.26 15 14

Numbers caught by the whole Belgian shrimp trawler fleet 5% percentile 0 0 81,234 91,704 190.990 0 128,350 10,651 139,361 Median 0 0 117,338 125,642 247,673 0 183,047 15,525 198,752 95% percentile 0 0 166.078 158,316 304.176 0 250,742 22,746 273.307 % < MLS 95% 99%

Q 2 -1996 Gurnards Dab Plaice All ages Age o Age 1 Age 2+ All ages Age o Age 1 Age 2♦ All ages Numbers of fish caught / 10,000 m 1 5% percentile 0 24 0 10 45 14 15 25 67 0 0 9 001 0 88 Median 0.48 0 14.12 19.27 33.64 0 1.69 0.06 1.77 95% percentile 0.77 0 17.8 25 69 42 47 0 2 73 0 12 2 77

Numbers caught by the whole Belgian shrimp trawler fleet 5% percentile 14,184 0 617,586 836,253 1,517.076 0 53,189 591 52,007 Median 28,368 0 834,480 1,138,841 1,988,096 0 99,878 3,546 104,606 95% percentile 45,506 0 1,051,965 1,518,258 2,509,942 0 161,341 7,092 163,705 % < MLS 74% 82%

Q 3- 1996 Gurnards Dab Plaice All aqes Aqe o Aqe 1 Aqe 2+ All aqes Aqe o Aqe 1 Aqe 2+ All aqes Numbers of fish caught /10,000 m2 5% percentile 0 1.46 4.12 0.1 6.52 4.02 1 32 0 534 Median 0 2.85 7.47 0.31 10.56 7.85 3.56 0 11.6 95% percentile 0 4.87 11.77 0 56 15.49 14.57 7.37 0 21 61

Numbers caught by the whole Belgian shrimp trawler fleet 5% percentile 0 151,419 427,292 10.371 676,201 416.921 136,899 0 553,821 Median 0 295,578 774,727 32,151 1,095,196 814,137 369,214 0 1,203,056 95% percentile 0 505,076 1,220,687 58,079 1.606.495 1,511,080 764.356 0 2,241,211 % < MLS 99% 99%

Q 4 -1996 Gurnards Dab Plaice All ages Age o Age 1 Age 2+ All aqes Age o Age 1 Age 2+ All aqes Numbers of fish caught / 10,000 m 1 5% percentile 0 7 8 1 94 0 74 11.06 5.67 0 22 0 6 Median 0.02 17.19 3.06 1.49 20.83 9.84 0.47 0 10.57 95% percentile 007 274 439 2 32 3287 15.5 0 81 0 17 04

Numbers caught by the whole Belgian shrimp trawler fleet 5% percentile 0 800,631 199,131 75,9571,135,254 581.997 22.582 0 615,870 Median 2,053 1,764,468 314,094 152,941 2,138,096 1,010,027 48,243 0 1,084,958 95% percentile 7,185 2.812.474 450,612 238,136 3.373.942 1.590.998 83.142 0 1.749,071 % < MLS 94% 99%

Total (Q1+2+3+4) Gurnards Dab Plaice All aqes Age o Age 1 Aqe 2+ All aqes Age o Age 1 Aqe 2+ All ages Numbers of fish caught / 10,000 m 1 5% percentile 0.05 3.59 4.99 3 82 13.26 3.76 1.28 0.04 5 13 Median 0.11 7.76 7.69 5.46 20.60 6.87 2.64 0.07 9.76 95% percentile 0.20 12.50 10.88 743 29.36 11.69 4 74 0.11 1668

Numbers caught by the whole Belgian shrimp trawler fleet 5% percentile 14,184 952,050 1,325.244 1.014.286 3,519.520 998.919 341,020 11,242 1,361,059 Median 30,421 2,060,047 2,040,639 1,449,575 5,469,061 1,824,164 700,382 19,071 2,591,372 95% percentile 52,692 3,317,550 2.889.342 1.972.7897.794.555 3,102.078 1,259,581 29,837 4,427.294 % < MLS 87.7% 98 3% Discarding by the shrimp trawler fleet 5-48 Table 5-4 (continued): Numbers of fish caught during the Belgian discard sampling programme, split up by quarter and by species; nos / 10,000 m2 and nos raised to shrimp trawler fleet level. The numbers caught are presented as the median and the 5% and 95% percentiles as obtained from the bootstrap procedure. Percentage fish caught < MLS are given for the species having a MLS according to the EU Council Regulation 345/92. Nos split up by age group are given for the species for which reliable stock assessment parameters were available for the modelling work. Q1 -1997 Solle Flounder Brill T urbot S o le nette Age o Aqe 1 Aqe 2+ All ages All ages All aqes All aqes All aqes Numbers offish caughti 10,000 m1 5% percentile 0 1.35 0 1.3 1 83 0 0 001 Median 0 2.54 0 2.52 3.77 0 0,02 0.02 95% percentile 0 4 09 0 4.14 6,15 0 004 0,05

Numbern caught by the whole Belgian shrimp trawler fleet 5% percentile 0 24.370 0 23.468 33,035 0 0 181 Median 0 45,852 0 45,491 68,056 0 361 361 95% percentile 0 73.833 0 74.735 111,020 0 722 903 % < MLS 100%

Q 2 -1996 Sole Flounder Brill Turbot So le nette Aqe o Aqe 1 Aqe 2+ All aqes All aqes All aqes All aqes All aqes Numbers offish caught /10,000 m! 5% percentile 0 2,47 004 2,63 0.34 0 0 0 Median 0.01 3,27 0.09 3,41 0,54 0 0 0 95% percentile 0.03 4.11 0.16 4.21 0.76 0 0 0

Numbers caught by the whole Belgian shrimp trawler fleet 5% percentile 0 145,975 2,364 155.431 20,094 0 0 0 Median 591 193,254 5,319 201,528 31,914 0 0 0 95% percentile 1.773 242.898 9,456 248.808 44,915 0 0 0 %

Q 3-1996 Sole Flounder Brill Turbot Solenette Age o Age 1 Age 2+ All ages All ages All aqes All ages All ages Numbers of fish caught / 10,000 m 3 5% percentile 3 08 2 52 0 01 6 48 0 79 0 0 0 Median 6.12 3.79 0.03 9,83 1.32 0 0 0 95% percentile 964 5.2 0.07 14.07 2.04 0 001 0

Numbers caught by the whole Belgian shrimp trawler fleet 5% percentile 319,432 261,354 1,037 672.052 81,932 0 0 0 Median 634,716 393,067 3,111 1,019,486 136,899 0 0 0 95% percentile 999,781 539,301 7,260 1,459.224 211,572 0 1,037 0 % < MLS 94%

Q4-1996 Sole Flounder Brill Turbot Solenette Age o Age 1 Age 2+ All ages All ages Alli ages All ages All ages Numbersoffish caught /10,000 m1 5% percentile 248 0 0 2.48 0.2 0 0 0 Median 4.02 0.03 0 4.09 0.48 0 0 0 95% percentile 595 0 08 0 5,96 0 97 0 0,01 0

Numbers caught by the whole Belgian shrimp trawler fleet 5% percentile 254,560 0 0 254.560 20,529 0 0 0 Median 412,633 3,079 0 419,818 49,270 0 0 0 95% percentile 610.738 8.212 0 611,764 99,566 0 1.026 0 % < MLS 99%

Total (Q1+2+3+4) Sole Flounder Brill Turbot Solenette Age o Age 1 Aqe 2+ All aqes All ages All ages All ages All aqes Numbers offish caught / 10,000 m* 5% percentile 2.16 1.63 0.01 4.16 0.59 0.00 0.00 0.00 Median 3.95 2.39 0.03 6.35 1.08 0.00 0.00 0.00 95% percentile 6.07 3.26 0.06 9.02 1.76 0,00 0,01 0,00

Numbers caught by the whole Belgian shrimp trawler fleet 5% percentile 573.992 431,699 3,401 1,105,510 155,590 0 181 Median 1,047,940 635,253 8,430 1,686,324 286,139 361 361 95% percentile 1,612.292 864,243 16,716 2,394.531 467,073 2.786 903 % < MLS 93.8% Results 5-49

Dab Age 1 -Q3 1996 Cod Age 1 -Q41996

14 10 >12 “ 10 ai 8 c- 6 c 4 S! 4 ai 2 £ 2 0 I... (N (N CO CO IO IO CD

Numbers 110,000 m2 Numbers 110,000 m2

Fig. 5-3 - Examples of the frequency distributions of the numbers of fish caught / 10,000 m2 over all hauls by quarter as observed in the Belgian discard sampling programme - Dab and Cod, Age 1, Q3 and Q4 1996 respectively.

The quarterly data (Fig. 5-2) revealed clear seasonal trends in both relative abundance and size composition of the species studied. These trends are related to migration patterns, the reproductive cycle, the time of the year when fish reach their size at first capture, the subsequent growth and mortality rates. Whiting, Dab and bib were the top three discarded fish in the Belgian shrimp fishery (Table 5-4, Fig. 5-2). Age 0 Whiting appeared in the catches in Q3. In Q2, mainly Age 1 fish were caught and in Q3 and 4, Age 1 fish catches dropped to a low level. The discard rates for Whiting were highest in Q2, 3 and 4 and were almost zero in Ql. Bib had a pronounced seasonal distribution, with nearly nil abundance in Ql, maximum abundance in Q3, and intermediate abundance in Q2 and 4. Dab was discarded all year through. Age 0 Dab first appeared in the catches in Q3 and their numbers considerably increased into Q4. They then became Age 1 fish in Ql. As their growth proceeded, they became increasingly available to capture up to Q2. Thereafter, the quantities of Age 1 Dab steadily decreased through Q3 and 4 as they moved to deeper waters. Dab of ages 2 and over were caught in quite large numbers in Q2. Plaice, Sole, Cod and Flounder held an intermediate position with regard to numbers tin-own overboard. Plaice of Age 0 was first seen in the catches in Q3 and the numbers caught gradually increased through Q4. In Ql, when these fish had grown to Age 1, they still appeared in the catches, but in somewhat lower numbers. Although Plaice was discarded all year through, the catches showed a clear minimum in Q2. Age 0 Sole first turned up in the catches in Q3 and their numbers gradually decreased into Q4. As these fish grew to Age 1, they kept on being caught through Ql, 2 and 3. In Q4, Age 1 fish were almost absent from the catches. Sole was discarded all year through with a clear peak in Q3. The discards for Cod were highest in Ql, at a lower level in Q4 and dropped to almost zero in Q2 and 3. Age 0 fish did not occur in the catches at all. Also for Flounder the highest catches were observed in Ql. Adult Flounder were caught over all quarters. For gurnards. Brill, Turbot and Solenette, the discard levels were very low. Dab, Flounder and Whiting of marketable size are regularly caught in the Belgian shrimp fishery and represent a significant part of the landings. Marketable Plaice, Sole and Cod are caught as well, although the numbers are relatively small. Discarding by the shrimp trawler fleet 5-50

5.3.1.3 Brown shrimp The length frequency data for the Brown Shrimp catches per quarter are given in Fig. 5-4.

Quarter 1 1997 Quarter 2 1996 2000 2000 g 1500 g 1500 o Discard o Discard o' 1000 o' 1000 ■ Comm. I Comm. y> 500 — Total y> 500 - Total o o

10 20 30 40 50 60 70 80 90 10 20 30 40 50 60 70 80 90 Lengthclass (mm) Lengthclass (mm)

Quarter 3 1996 Quarter 4 1996 2000 2000 N E o 1500 o 1500 o Discard o Discard o 1000 O 1000 ■ Comm. ■ Comm. — Total . — Total ■ Ô 5 0 0

10 20 30 40 50 60 70 80 90 10 20 30 40 50 60 70 80 90 Lengthclass (mm) Lengthclass (mm)

Fig. 5-4 - The length frequency data per quarter for the Brown Shrimp catches in numbers caught / 10,000 m2 as observed during the Belgian discard sampling programme

Shrimp Q2 1996 (all lengths) Shrimp Q4 1996 (all lengths)

incNiincoin^l-minincDmiY T^inCNiincoin^rininincDin

Numbers (thousands) 110,000 m2 Numbers (thousands) 110,000 m2

Fig. 5-5 - Example of two frequency distributions of the numbers of shrimp caught / 10,000 m2 over all hauls in Q2 (total) and Q4 (total) 1996 as observed in the Belgian discard sampling programme respectively - Q2 with a small and Q4 with a high degree of deviation form the normal.

The average numbers of discard and commercial shrimps caught and the totals (discard + coimnercial), by quarter, per 10,000 m2 and raised to fleet level are given in Table 5.5. Nonnality was tested for each dataset (i.e. nos of shrimp caught over a series of hauls by quarter or over the whole year). The tests indicated nonnality for the data in Ql and Q2 with Results 5-51 a low and non-significant skewness and kurtosis (e.g. Fig. 5-5 “Shrimp Q2 1996”). Visual inspection of the data showed frequency distributions with a small deviation from normality, without outliers. The data for Q3, Q4 and the whole year, on the other hand, were not normally distributed, as was indicated by the tests, and had a significant skewness and kurtosis for each dataset (e.g. Fig. 5-5 “Shrimp Q4 1996”). Therefore, the data for Q3, Q4 and the whole year were bootstrapped.

Table 5-5: The Brown Shrimp catches as observed during the Belgian discard sampling programme; undersized, commercial and total catch; nos / 10,000 m2, nos caught extrapolated to the whole fleet and weight caught by the whole fleet; by quarter and the whole year

Q 1-1997 Undersized Commercial Totals Q2-1996 Undersized Commercial Totals

Numbers c a u g h10,000 t/ m2 Upper 90% conf. limit 900 526 1536 1562 1243 2859 Mean 1229 722 1952 2033 1469 3502 Lower 90% conf. limit 1559 919 2367 2504 1696 4146

Numbers caught by the whole Belgian shrimp trawler fleet Upper 90% conf. limit 16,242,586 9,489,661 27,733,009 92,303,134 73,479,735 168,945,278 Mean 22,192,707 13,041,921 35,234,629 120,146,276 86,844,938 206,991,214 Lower 90% conf. limit 28,142,829 16,594,181 42,736,248 147,989,419 100,210,140 245,037,149 % discarded 64% (54-75) 56% (51-61)

Weight caught by the whole Belgian shrimp trawler fleet (kg) Upper 90% conf. limit 12,177 17,973 32,538 89,231 143,910 233,621 Mean 16,638 24,701 41,339 116,147 170,085 286,232 Lower 90% conf. limit 21,099 31,429 50,141 143,063 196,261 338,843

Q3-1996 Undersized Commercial Totals Q4-1996 Undersized Commercial Totals

Numbers c a u g h t/10,000 m2 5% percentile 5466 2171 7864 2440 1306 3808 Median 6466 2619 9151 3125 1610 4708 95% percentile 7552 3124 10552 4001 1933 5789

Numbers caught by the whole Belgian shrimp trawler fleet 5% percentile 566,939,190 225,113,591 815,600,532 250,496,984 134,024,642 390,825,058 Median 670,642,633 271,589,934 949,074,430 320,731,846 165,301,609 483,286,674 95% percentile 783,227,949 324,006,884 1,094,410,877 410,648,892 198,369,732 594,207,973 % discarded 64% (59-70)

Weight caught by the whole Belgian shrimp trawler fleet (kg) 5% percentile 373,880 343,694 736,407 185,167 249,662 440,739 Median 442,269 414,652 856,921 237,085 307,925 545,009 95% percentile 516,516 494,680 988,146 303,551 369,524 670,097

Total (Q1+2+3+4) Undersized Commercial Totals

Numbers ca ug h t/10,000 m2 5% percentile 3298 1598 4972 Median 3898 1826 5723 95% percentile 4518 2082 6515

Numbers caught by the whole Belgian shrimp trawler fleet 5% percentile 934,941,109 453,054,176 1,409,567,403 Median 1,105,153,599 517,569,241 1,622,623,612 95% percentile 1,280,809,442 590,391,087 1,847,184,573 % discarded 65% (63-68)

Weight caught by the whole Belgian shrimp trawler fleet (kg) 5% percentile 687,056 803,013 1,502,412 Median 812,139 917,363 1,729,502 95% percentile ______941,223______1,046,435______1,968,854 Discarding by the shrimp trawler fleet 5-52 It was clear that in wintertime, the densities of shrimps were very low, as were the catches and the discards. In spring, the catches and discards gradually rose together with rising shrimp densities to reach a maximum in summer. After that, densities and catches dropped in autumn, to almost half of the maximum. Shrimp discards contained more than double the amount of shrimps (in numbers) compared to shrimps landed. In weight, shrimp discards and landings were in the same order of magnitude.

5.3.2 North Sea data For completeness, the results of the other project partners are summarised in this section. The sampling effort for the participating nations is given in Table 5.6.

Table 5-6: Summary of the discard sampling programme B D Dk F NI UK Total

No. of hauls analysed 108 151 87 - 18 163 527 No. of trips carried out 21 34 13 - 6 30 104

The numbers of fish and shrimps caught / 10,000 m2 and the totals, extrapolated to the whole fleet, are given in Tables 5.7 and 5.8.

Table 5-7: Annually averaged numbers of fish caught / 10,000 m2 during the discard sampling programme, split up by species and country

Whiting Cod Bib & poor cod Country Dicarded Landed Totals Dicarded Landed Totals Dicarded Landed Totals Dk 2.74 0.00 2.74 7.44 0.00 7.44 0.00 0.00 0.00 D 13.50 0.00 13.50 23.84 0.00 23.84 0.00 0.00 0.00 NI - north n/a n/a n/a n/a n/a n/a n/a n/a n/a NI - south n/a n/a n/a n/a n/a n/a n/a n/a n/a B 32.41 3.70 36.12 1.71 0.24 1.94 17.92 0.04 17.95 UK 40.90 0.00 40.90 14.46 0.00 14.46 0.00 0.00 0.00

Gurnards Dab Plaice Country Dicarded Landed Totals Dicarded Landed Totals Dicarded Landed Totals Dk 0.04 0.00 0.04 128.42 0.00 128.42 82.54 0.00 82.54 D 0.25 0.00 0.25 114.45 0.00 114.45 997.53 0.00 997.53 NI - north n/a n/a n/a n/a n/a n/a n/a n/a n/a NI - south n/a n/a n/a n/a n/a n/a n/a n/a n/a B 0.12 0.00 0.12 15.83 2.02 17.85 7.86 0.15 8.01 UK 0.00 0.00 0.00 17.35 0.00 17.35 31.51 0.00 31.51

Sole Flounder Brill Country Dicarded Landed Totals Dicarded Landed Totals Dicarded Landed Totals Dk 0.27 0.00 0.27 0.13 0.00 0.13 0.25 0.00 0.25 D 11.96 0.00 11.96 5.91 0.00 5.91 0.05 0.00 0.05 NI - north n/a n/a n/a n/a n/a n/a n/a n/a n/a NI - south n/a n/a n/a n/a n/a n/a n/a n/a n/a B 5.36 0.35 5.71 0.61 0.34 0.95 0.00 0.00 0.00 UK 4.40 0.00 4.40 3.30 0.00 3.30 0.00 0.00 0.00

Turbot Shrimp (in 1000's) Country Dicarded Landed Totals Dicarded Landed Totals Dk 0.25 0.00 0.25 5.87 3.45 9.32 D 0.14 0.00 0.14 17.68 3.75 21.43 NI - north n/a n/a n/a n/a n/a n/a NI - south n/a n/a n/a n/a n/a n/a B 0.00 0.00 0.00 3.66 1.68 5.34 UK 0.00 0.00 0.00 4.34 4.15 8.49 Results 5-53 Table 5-8: Total annual numbers of fish caught extrapolated to national fleet levels, split up by species and country Whiting Cod Country Dicarded Landed Totals Dicarded Landed Totals Dk 1,118,000 0 1,118,000 3,037,000 0 3,037,000 D 9,808,000 0 9,808,000 17,322,000 0 17,322,000 NI - north 5,011,000 0 5,011,000 14,151,000 0 14,151,000 NI - south 17,333,000 2,432,000 19,765,000 2,731,000 140,000 2,871,000 B 9,189,000 1,050,000 10,239,000 484,000 67,000 551,000 UK 12,874,000 0 12,874,000 4,552,000 0 4,552,000 Total 55,333,000 3,482,000 58,815,000 42,277,000 207,000 42,484,000

Bib & poor cod Gurnards Country Dicarded Landed Totals Dicarded Landed Totals Dk 0 0 0 17,000 0 17.000 D 0 0 0 184,000 0 184,000 NI - north 0 0 0 74,000 0 74,000 NI - south 9,475,000 36,000 9,511,000 119,000 1,000 120,000 B 5,080,000 10,000 5,090,000 33,000 0 33,000 UK 0 0 0 0 0 0 Total 14,555,000 46,000 14,601,000 427,000 1,000 428,000

Dab Plaice Country Dicarded Landed Totals Dicarded Landed Totals Dk 52,440,000 0 52,440,000 33,706,000 0 33,706,000 D 83,148,000 0 83,148,000 724,734,000 0 724,734,000 NI - north 288,180,000 0 288,180,000 151,478,000 0 151,478,000 NI - south 13,570,000 2,033,000 15,603,000 6,025,000 132,000 6,157,000 B 4,489,000 572,000 5,061,000 2,228,000 43,000 2,271,000 UK 5,462,000 0 5,462,000 9,920,000 0 9,920,000 T ota I 447,289,000 2,605,000 449,894,000 928,091,000 175,000 928,266,000

Sole Flounder Country ______Dicarded Landed Totals Dicarded Landed Totals Dk 111,000 0 111,000 55,000 0 55,000 D 8,691,000 0 8,691,000 4,297,000 0 4,297,000 NI - north 554,000 0 554,000 366,000 0 366,000 NI - south 3,455,000 257,000 3,712,000 1,092,000 273,000 1,365,000 B 1,519,000 100,000 1,619,000 173,000 95,000 268,000 UK 1,386,000 0 1,386,000 1,040,000 0 1,040,000 Total 15,716,000 357,000 16,073,000 7,023,000 368,000 7,391,000

Brill Turbot Country Dicarded Landed Totals Dicarded Landed Totals Dk 103,000 0 103,000 104,000 0 104,000 D 33,000 0 33,000 100.000 0 100,000 NI - north 430,000 0 430,000 457,000 0 457,000 NI - south 0 0 0 4,000 0 4,000 B 0 0 0 1,000 0 1,000 UK 0 0 0 0 0 0 Total 566,000 0 566,000 666,000 0 666,000

Shrimp {¡n 1000's) Country Dicarded Landed Totals Dk 2,396,000 1.408,000 3,804,000 D 128,415,000 27,278,000 155,693,000 NI - north n/a n/a n/a NI - south n/a n/a n/a B 1,038,000 477,000 1,515,000 UK 1,367,000 1.306,000 2,673,000 Total 133,216,000 30,469,000 163,685,000

The overall differences between the national fleets as to numbers of fish caught per unit of swept area. broadly reflect the differences as to both density distribution and habitat preference of the species investigated. Species with a preference for open waters, such as Bib and Poor Cod, were much more evident in the catches of the Belgian shrimpers than in those of the other fleets. Conversely, species with a clearly estuarine distribution, such as Flounder and Age 0 Plaice, were found in much higher numbers in the catches of the Gennan and UK Discarding by the shrimp trawler fleet 5-54 trawlers, which mostly operate on the inshore and estuarine shrimp fishing grounds in respectively the Waddensea and the Wash and Humber estuaries. Most striking were the huge numbers of juvenile Plaice caught in the German shrimp fishery and high numbers of juvenile Dab caught all over the North Sea. The number of discarded Brill, Turbot and gurnards was low. The roundfish species and Flounder and Sole took an intermediate position.

5.4 Discussion Alversson et al. (1994b) estimated the discard ratio for the North Sea shrimp fishery at 59%. The present study, though, showed that a consistently higher proportion of the catch was thrown overboard, consisting of large amounts of undersized shrimps and commercial fish and a large variety of non-commercial fish and invertebrate species. On average, the catch consisted of 29 % commercial shrimps and the estimated discard ratio was 71.3%. With the exception of some outliers, the variability of the discard ratio was low. The observed numbers of fish caught at species level, however, were very variable from one haul to another. This variability arose from the variability of the distribution of the fish on the fishing ground and the variability of the catchability of a fish species at a particular moment. This variability was also found by Morizur et al. (1992) in the French small mesh fishery in the Western English Channel. Fish behaviour influences gear efficiency in three ways: through availability, accessibility and vulnerability (Harden Jones, 1974). First of all, many commercially exploited fish stocks are migratory and theiravailability will change as fish move from one area to another, usually depending on the season (Arnold et al., 1997). Not all fish available to a fishing gear, necessarily come within its zone of influence. Environmental factors may cause fish to aggregate in places not accessible to the gear (Harden Jones, 1974, Godo and Wespestad, 1993). A shrimp beam trawl has a limited vertical opening (58 cm) and gives little stimulation to fish dug in the seafloor. Moreover, only fish present within a narrow band in the water column will be accessible to capture. This accessibility can change constantly as was observed for Plaice, that shows a clear vertical migration pattern depending on the currents (Arnold et al., 1997). Vulnerability is the third factor affecting the interaction between fish and gear and is determined by the fishing power of the gear, which is subject to variability (Dahm, 1999), and the ability of fish to escape from a gear. Presumably, the observed variability in the data was a reflection of a real situation and a more intense sampling scheme would also have shown a similar high variability. The observation that catch composition and discards can be quite variable is confirmed when comparing the Belgian results with by-catch data observed by Walter (1997). The shrimp fraction (weight-based) in the catch was only 11% compared to 29% in the Belgian shrimp fishery. The undersized shrimp and other by-catch fraction were 64% and 25% respectively compared to 26% and 45% in Belgium. Berghahn and Vorberg (1997) reported 25-50% of edible shrimps in the catch of German shrimp trawlers and 20-35% other crustaceans and fish. These figures based on weight, however, can be ambiguous and the use of numbers of individuals is more functional for further calculations. It is clear from the Belgian discard data that commercial fish species such as Flounder, Brill, Turbot and gurnards will probably not suffer from the discarding practices in the Belgian shrimp fishery. For Whiting, bib, Dab, Plaice and Sole, on the other hand, the numbers discarded are quite high and indicate that there may be a serious problem affecting recruitment. Discussion 5-55 The numbers of fish discarded in the Belgian shrimp fishery have been estimated in the past, in 1932/33 (Gilson, 1935) and in 1963 (Leloup and Gilis, 1965). The data are not comparable because of differences in methodology and coverage. The fishing gears used and the area fished were, however, similar to the present study. For the sake of completeness, the data are presented here but should be interpreted with caution (Table 5-9). The numbers of fish discarded per hour are in the same order of magnitude. Whether differences are caused by differences in fishing power of vessel and gears, abundance of the fish or differences in methodology is unclear.

Table 5-9: Numbers of fish discarded in the Belgian Brown Shrimp fishery

Size or age group Numbers of fish discarded / year Numbers of fish discarded / hour 1932-33 1963 1996-97 1932-33 (*) 1963 1996-97 1932-33 1963 1996-97

W hiting <18cm <20cm Age 0 & 1 21,693,792 4,059,328 9,559,995 133 56 290 Dab <18cm <20cm Age 0 & 1 21,703,248 7,321,288 4,100,686 133 101 124 Plaice <24cm <2 5 cm Age 0 & 1 5,963,388 10,873,200 2,524,546 36 150 77 Sole <24cm <24cm Age 0 & 1 3,706,596 3,986,840 1,683,193 23 55 51 (*): Only the Oostende shrimp trawler fleet

Although the recent landings statistics clearly show that shrimp landings can differ quite strongly from one year to another (Fig. 5.2 C), the seasonal pattern of fishing effort (low effort in Q l & 2 and high effort in Q 3 & 4) is quite stable and was also observed in this study. Table 5.10 shows the effort pattern for the period when the sampling took place. An extreme low value was observed in wintertime, rising quite sharply in spring and rising further to a maximum in autumn. It is clear from this table that the density of shrimps on the fishing ground has an influence on the behaviour of the fishermen. This is not the only factor, though, since the highest effort was observed in Q4 when shrimp density was only 61% of the maximum (Table 5.10). Factors such as the market price for shrimp, which can fluctuate quite strongly, alternative fisheries and weather conditions also play a role in the fishing effort pattern.

Table 5-10: Fishing effort (source: national fishery statistics) and commercial shrimp densities (as observed in the Belgian discard sampling programme) as a percentage of the maximum observed. Q l 1997 Q2 1996 Q3 1996 Q4 1996 Fishing effort as a % of maximum 14% 68% 94% 100% observed Density of shrimps as a % of maximum 28% 56% 100% 61% observed

Quite a large fraction of the shrimp catch was discarded. On average (in numbers) this was 65% with a maximum in summer of 71% and a minimum in spring of 56%. Large amounts of shrimps are, thus, discarded all year through. Converted to weight, an estimated 812 tons / year of shrimps was discarded compared to the yearly estimated landings of 917 tons. The consequences of this discarding for the shrimp stock are, however, reckoned to be rather low Discarding by the shrimp trawler fleet 5-56 since the survival of shrimp discards has been estimated to be higher than 78% (Revill et al.. 2000). The length frequency distributions of shrimp catches for the different quarters show a remarkably similar shape, especially if expressed as relative frequencies (Fig. 5-6). Unlike the fish catches, different age groups could not be distinguished for shrimp, nor was there a clear shift of average length during the year. It must be noted, though, that high amounts of small shrimps escape through the meshes of the shrimp trawl (see Section 7). Consequently, a high mode of juvenile shrimps is unlikely to occur in the length frequency distribution of the catches due to the selective properties of the trawl.

0.25

> 0.2 o § 0.15

a> > I 0.05 £

0 50 100 Lengthclass (mm)

Fig. 5-6 - The relative length frequency data per quarter for the Brown Shrimp catches.

The issue of by-catches of juvenile fishes in the North Sea shrimp fishery has been addressed in several studies and was identified as a major problem. The present project provided basic data on fleets and discards for this fishery in all countries surrounding the North Sea. The data collection was done over a large area, in the same period of time, following a standard methodology. This was the first initiative to study this issue in such a wide context. The results, however, reflect the discards in one year only and other years would most likely show different results. The numbers caught and discarded of any species will depend on 1) its year class strength, 2) its geographical distribution and 3) the amount and geographical distribution of fishing effort of the shrimp trawler fleet. 1) Year class strength varies considerably for many fish species (ICES, 2001a). At first sight, one might conclude that this seriously undermines the findings of this project. Since, however, the time series of year class strength is well known for most of the species studied, the present results can be tuned and used as a guideline to estimate the impact of the shrimp fishing fleet on the juvenile fish stocks in other years. 2) The different age groups of fish species may be spread differently over the North Sea from one year to another (Van Beek et al., 1989). This source of variability lias a minor effect on the numbers of juvenile fish discarded by the shrimp trawler fleet as a whole, but it does have an effect on the role played by each of the national fleets. From the Conclusions 5-57 present figures, it looks as if the German fleet is responsible for almost 80% of the discards of juvenile Plaice. A more southerly distribution of the 1996 year class of Plaice might have shown a different picture. These trans-national movements of fish plead for an international approach, where all countries collectively assume responsibility to reduce the by-catch levels in the shrimp fishery. 3) It is apparent that the magnitude of discarding in the shrimp fishery is related to the variable magnitude of the fishing effort. Time series of fishing effort exist for this fishery (ICES, 2001b) so the results of the present study can easily be tuned to be applicable to other years. The fate of discards is an important issue. If the animals returned to the sea would all survive, discarding would not be a problem. This is, however, not the case. Kelle (1976/77) suspected that the catching process in conjunction with subsequent sorting of the shrimp catch is responsible for high mortality among flatfishes. Berghahn and Vorberg (1997) found mortality rates for Whiting and other gadoids as well as smelt of 100% and suspected comparable rates for clupeids. For flatfish the mortality ranged from 17 to 100%. For other species like hooknose and sculpin, mortality was only around 10%. As already found by Van Beek et al. (1990), discard mortality is quite variable and also depends on catch volume and composition, water/air temperature, maintenance on board etc. After being returned to the sea, discard survival will be further reduced by feeding activity of seabirds, seals and other predators. On top of this, injuries may lead to further mortality or diseases e.g. the black spot disease in shrimps (Knust, 1990). Any further study of the biological impact of discarding will have to take discard mortality rates into account.

5.5 Conclusions For the Belgian part of the discard sampling program, 21 sea trips and 108 hauls were carried out with a good seasonal coverage. The discard ratio observed, indicated that almost three quarters of the total catch volume is returned to the sea. Length frequency distributions by species for the numbers of fish caught by the Belgian shrimp trawler fleet are presented together with the total numbers caught, split up by age group, per 10.000 m2 and by the whole fleet. The quarterly data revealed clear seasonal trends in both the relative abundance and the size composition of the species studied. These trends are related to migration patterns, reproductive cycle, the time of the year when fish reach their size at first capture, the subsequent growth and mortality rates. The yearly totals gave a good indication of the relative importance of the different species discarded. For shrimp, the densities on the fishing grounds in wintertime were very low, as were the catches and the discards. In spring, the catches and discards gradually rose together with rising shrimp densities to reach a maximum in summer. After that, densities and catches dropped in autumn, to almost half of the maximum. Shrimp discards contained more than double the amount of shrimps (in numbers) compared to shrimps landed. All partners together did 104 sea trips and 527 hauls. A summary of the numbers of fish and shrimps discarded, by country and by species, is presented. The overall differences between the national fleets in the numbers of fish caught per unit of swept area, broadly reflect the differences in both the aerial distribution and habitat preference of the species investigated. Species with a preference for open waters, such as Bib and Poor Cod, were much more evident in the catches of the Belgian shrimpers than in those of the other fleets. Conversely, species with a clearly estuarine distribution, such as Flounder and Age 0 Plaice, were found in much higher numbers in the catches of the German and UK trawlers, which mostly operate on the inshore and estuarine shrimp fishing grounds in respectively the Waddensea Discarding by the shrimp trawler fleet 5-58 and the Wash and Humber estuaries. Most striking were the huge numbers of juvenile Plaice caught in the German shrimp fishery and high numbers of juvenile Dab caught all over the North Sea. Species like Brill, Turbot and gurnards were only discarded in low numbers. The roundfish species and Flounder and Sole took in an intermediate position. The issue of by-catches of juvenile fishes in the North Sea shrimp fisheries has been addressed in several studies and was identified as a major problem. The present project provided basic data on fleets and discards for this fishery in all countries surrounding the North Sea. The data collection was done over a large area, in the same period of time, following a standard methodology. This project was the first initiative to study this issue in such a wide context. The present study produced absolute numbers of discards. Despite the alarming nature of these figures, it would be incautious to call for immediate management action. Before doing so, it seems advisable to carefully examine these figures in relation to other factors that determine the composition of a fish stock. Discard survival is an important factor determining the significance of discarding. Natural mortality caused by severe winters, predation or competition also plays a part in the degree of decimation of a year class, before it has a chance to reach the age of first capture. Fishing and discarding are, however, the man-made effects that can be managed and reduced, provided there is a strong case to do so. Whether the case would be strong enough was evaluated in a follow-up project (See Section 6 ). Introduction 6-59

6 Biological and economic consequences of discarding in the Brown Shrimp fishery

6.1 Introduction The biological and economic impacts of the discarding of juvenile round- and flatfish in the Crangon fisheries in the North Sea have been studied in the EU-funded project ECODISC with the co-operation of following institutes: • ARBEE Computer Consultants, Grimsby, UK, • Bundesforschungsanstalt für Fischerei, Hamburg, Germany, • CEF AS, Lowestoft, UK • CEMARE, Portsmouth, UK, • Departement voor Zeevisserij, Belgium, • DIFMAR, Charlottenlund, Denmark, • IFREMER, Boulogne sur mer, France, • Rijksinstituut voor Visserij Onderzoek, Umuiden, Netherlands, • University of Humberside, Grimsby, UK. • University of Newcastle upon Tyne, Newcastle, UK, This project was a follow-up of the RESCUE project, essential to put the findings of that project into their proper perspective. The numbers of juvenile commercial fish discarded in the Crangon fisheries are impressive, but their impact remained largely unknown. In this project, the biological and economic significance of discarding has been studied, in terms of its potential impact on recruitment and spawning stock biomass of the fish species concerned, and hence on the lost potential landings and their associated market value. This analysis drew upon the joint expertise of fish and shrimp biologists, gear technologists and fisheries economists to ensure that the multitude of biological, technical and economic factors are taken into account. The task of the author was to provide the basic data in a suitable format, to comment on the newly designed model and to give advice on the fishing gear aspects of the exercise. Biologists and economists, however, carried out the modelling work and consequently, the model is not elaborated in detail in this report. Because of its importance for further work a short description of the model design and an executive summary of the results are included. Further details are given in Revill et al (1999).

6.2 Materials and methods Four fish species were selected to be included in the model calculations, i.e. Plaice, Sole, Whiting and Cod. The main reason for this limitation was, apart from their economic importance, the lack of sufficiently reliable basic biological data, necessary in the model, for the other species. The biological models used in this work are improved versions of models developed previously by Revill (1997). In principle, the model calculates how many fish would survive and contribute to the future stock and eventually be caught and landed by other fleets if discarding in the shrimp fishery would stop or be reduced. The biological consequences evaluated in the model were defined as: • the estimated loss to the spawning stock biomasses (SSBs) that arise from the estimated levels of discarding in the Brown Shrimp fisheries and • the estimated annual lost landings that arise from the estimated levels of discarding in the Brown Shrimp fisheries. Biological and economic consequences o f discarding in the Brown Shrimp fishery 6-60 The structure of the model is detailed in Fig. 6-1. The input for the model was the data collected in the RESCUE project, reworked as follows: • grouped according to the typical discard patterns in terms of age and species in the Crangon fisheries, • grouped according to the spatial distribution of Crangon directed fishing effort, • corrected for the typical annualCrangon directed fishing effort levels, • corrected for the annual local indices of relative abundance of the discard species and • corrected for the different discard survival rates. In the next steps, the following variables were incorporated within the model: • the losses which would arise from natural mortality (M), • the losses which would arise from subsequent discarding by other fleets (i.e. the whitefish directed fleets) (D), • the landings which would arise from current finfish directed fishing effort levels (F) and • age - weight relationships for the different species. The model output the amounts of fish lost due to the discarding practices. The model could also be used to calculate the likely biological benefits that would arise from the introduction of discard reduction techniques (selective gears, closed season).

Source Data Numbers of fish discarded / 10,000m2 on an individual tow basis.

Tows grouped into country of origin, age, species discarded, quarter & year of discarding.

Grouped tows bootstrapped to produce median and confidence limits

Data corrected to eliminate the influence of: Data raised from 1. Year class recruitment strength fluctuations Nos per 10,000m' 2. Crangon directed fishing effort fluctuations to fleet level

Data modelled to allow for losses that Data corrected would occur from natural mortality according to discard IM}______survival rates

Data modelled to allow for losses that would occur from subsequent discarding by other fleets (D) Output Losses due to (F) converted to Data modelled to allow for losses that landings (with confidence limits) would occur from fishing mortality using age- weight relationships ______

Fig. 6-1 - The structure of the biological model. Results 6-61 In the economic analysis, a number of components were included. The first stage involved the estimation of the economic value of the Crangon fisheries in Europe and the economic cost associated with the discarded fish. The whitefish fisheries in the North Sea incur these latter costs. The economic costs of discarding were estimated based on the predictions of the biological model of the total potential landings missed. The estimated increase in landings of each species through eliminating discarding was allocated to each country on the basis of their relative share of the total allowable catches for the North Sea. The estimated increase in landings in each country was multiplied by the average fish market price for each species in each country. From this, some conclusions could be drawn about the net economic value of the shrimp fishery. This was derived by subtracting the value of the potential landings from the estimate of the economic value of the fishery. The second stage of the analysis involved estimating the potential effects of a regulation making selective gear mandatory, on the economic performance of the shrimp fisheries. In many fisheries where technical measures have been introduced to reduce by-catch, this has also resulted in a reduction of the targeted catch. This causes a decrease in the fishermen’s income and an overall loss of profitability to the fishery. This was calculated for the shrimp fleet and involved developing a bio-economic model of the fishery and running a number of simulations with alternative assumptions about the impact of such a regulation.

6.3 Results

6.3.1 The biological and economic consequences of discarding in the Crangon fisheries The magnitude of estimated annual losses to the SSBs for the four selected fish species, arising from current levels of discarding in the North Sea Brown Shrimp fisheries are detailed in Table 6-1. The estimated annual lost landings for the EU fishing fleet caused by this discarding are given in Table 6-2.

Table 6-1: Estimated annual losses to SSBs arising from current levels of discarding in the North Sea Crangon fisheries

Plaice Sole Cod Whiting Totals % of SSB Totals % ofSSB Totals % ofSSB Totals % of SSB (tonnes) (tonnes) (tonnes) (tonnes) 5% percentile 14,329 6.2 226 0.4 460 0.5 1,603 0.6 Median 23,851 10.3 711 1.1 881 1.0 2,837 1.1 95% percentile 37,378 16.2 1,430 2.2 1,507 1.8 4,409 1.7

The net economic value of the Crangon fisheries was estimated by subtracting the cost imposed on the whitefish fisheries (i.e. the estimated fish market value of annual lost landings due to the discarding; Table 6-2) from the economic performance indicators for the Crangon fisheries (Table 6-3). The key measures are full equity returns to owners (which represents the total level of profits to the owners) and the gross value added (i.e. the full equity returns and crew costs). From Table 6-3, it can be seen that discarding in the Crangon fishery reduces the value of the recorded landings by almost 30 %. Discarding may also result in a net loss in full equity returns to the combined whitefish/Crangon fisheries. Gross value added, however is likely to be positive for mid or lower estimates of discards. Biological and economic consequences o f discarding in the Brown Shrimp fishery 6-62 Table 6-2: Estimated annual lost landings for the EU fishing fleet caused by discarding in the Brown Shrimp fishery for four selected species:

Species Type of Estimated annual lost Estimated lost landings as a % Estimated fish market estimate landings (tonnes) of the 2000 North Sea TAC value of annual lost landings (million Euro) Cod 5% percentile 3,198 3.9 % 4.6 Median 1,890 2.3 % 2.7 95% percentile 997 1.2% 1.4

Whiting 5% percentile 2,372 7.9 % 1.9 Median 1,525 5.1 % 1.2 95% percentile 871 2.9 % 0.7

Plaice 5% percentile 18,749 19.2% 27.8 Median 12,066 12.4 % 17.9 95% percentile 7,349 7.5 % 10.9

Sole 5% percentile 1,355 6.9 % 9 Median 588 3.0 % 3.9 95% percentile 153 0.8 % 1

Table 6-3: Estimated net economic value of the European Crangon fisheries (with confidence limits) taking into account the cost of discarding (units in Million Euro).

Value of Cost to whitefish fisheries Net Value Crangon fisheries upper median low upper median low

Value of recorded landings 92.2 43.3 25.7 14.0 48.9 66.5 78.2 full equity return to owner 18.1 33.9 21.9 10.9 -15.8 -3.8 7.2 gross value added 43.1 43.3 25.7 14.0 -0.2 17.4 29.1

6.3.2 The benefits arising from the introduction of selective fishing gears In order to calculate the theoretical benefits if more selective gears were to be used in the shrimp fishery, the characteristics of a 12 mm selective sorting grid, studied by Graham (1997) in the UK, were used. The results apply to any selective device with similar selective characteristics. It should be borne in mind, however, that this aspect of the modelling is for illustrative purposes only. In reality, practical developmental work would be required to tailor and modify any such selective device to local conditions and to overcome problems such as debris / weed build up, commercial acceptability etc. Because of the limited selectivity data available at the time of the study, the modelling was performed only for two ‘representative’ fish species, namely Plaice and Whiting. The selectivity ogives of the grid for Plaice and Whiting are given in Fig. 6-2. These indicate the estimated percentages of fish at length that would escape if the grid were installed in the net, compared to the catch of a standard shrimp beam trawl. Results 6-63 The shrimp fisheries of Belgium and Germany were selected for this aspect of modelling due to the very striking differences in their discard patterns. The Belgian shrimp fishery is similar to the UK and Netherlands (South) fisheries, in that the most significant and important discarded fish species examined are the Age 1 Plaice. The German fishery and other Waddensea fisheries are strikingly different, in that the most significant discarded species examined are the Age 0 Plaice. This has major implications for the efficacy of selective gears and devices.

100%

ui 75% ui

Plaice 50% W hiting

25%

0%

Length (cm)

Fig. 6-2 - Selectivity ogives for the UK 12mm selective sorting grid .

Table 6-4 gives the following results: • The estimated annual lost landings arising from the current levels of discarding of Plaice and Whiting in the Belgian and German shrimp fishery. These numbers were already calculated in the biological model and refer to the fishery with standard nets in the two countries. • The estimated annual increase in landings arising from a theoretical introduction of a selective sorting grid (and thus reduction of discards) in the shrimp fisheries of Belgium and Germany. • The estimated efficacy of that type of sorting grid in reducing the discards and thus lost landings. The selective gears developed by Graham (i.e. a 12mm grid) could recover more than 90 % of the lost Plaice landings attributable to shrimp related discards in the Belgian fisheries. In the Waddensea shrimp fisheries, only about 20 % of lost Plaice landings could be recovered if such gear were used. The variation in recovery rates results from the size differences of the by-catches in the different fisheries. It is clear that the European whitefish fisheries would benefit from the introduction of a selective device in the shrimp fishery. The level of the benefits would, however, depend on the selective properties, i.e. the amounts of fish saved by the selective fishing gear. It was expected that the introduction of a selective grid would reduce the fishing power of the shrimp trawl for Crangon resulting in an overall decrease in Crangon landings. As the price flexibility for Brown Shrimp was, however, estimated to be approximately -1, prices were estimated to increase by the same proportion as catches decreased. Since in Germany Biological and economic consequences o f discarding in the Brown Shrimp fishery 6-64 and Denmark, contrary to Belgium, the Netherlands and the UK, many vessels already used a selective device to reduce discards, the effect of the assumed regulation on the shrimp catches was estimated to be rather low and the higher price for shrimps could increase fishing effort levels. In the case of Denmark, the potential increase in effort was limited as the fishery is constrained by limited entry. In Belgium, the Netherlands and the UK, the reduced catchability would lead to an increased cost of harvest and thus lower marginal revenue resulting in a reduction of fishing effort.

Table 6-4: The efficacy of selective fishing gear (12mm UK grid) in the Belgian and German Crangon fisheries

Country Species Age group Estimated annual lost Estimated annual increase Efficacy of landings arising from in landings arising from selective gear current levels of the introduction of discarding (tonnes) selective gears (tonnes) (% reduction)

Belgium Plaice 0 11 4 36% 1 272 257 95 % 2 11 11 100 % All ages 294 272 93 % Belgium Whiting 0 49 39 80% 1 143 143 100 % 2 8 8 100 % All ages 201 191 95% Germany Plaice 0 7,556 1,181 16% 1 1,157 643 56% 2 13 13 100 % All ages 8,727 1,838 21 %> Germany Whiting 0 29 3 10% 1 365 293 80% 2 6 6 100 % All ages 400 303 76%o All efficacy estimates are based on the median range estimates

The results of the modelling exercise demonstrated that vessels already using a selective device (and hence not facing reduced catchability) would obtain higher revenues if a selective fishing gear was enforced upon the whole fleet and the overall price level for shrimps increased. Vessels not yet using selective gear would have reduced revenues since the reduction in catches would be relatively greater than the price increase. Since the use of selective gear is relatively least in Belgium and obligatory in Denmark, Belgium is likely to be the most adversely affected in percentage terms, whilst Denmark would be the greatest beneficiary. From this, it might be concluded that introducing uniform gear restrictions may result in a transfer of benefits from some countries to others. It can be argued, however, that as the main beneficiaries had largely adopted selective gear prior to the simulated uniform introduction, these countries had been actually subsidising production in the other countries. This is because the reduced catches they faced resulted in a higher price for all producers. Introducing a uniform restriction addresses this imbalance. Conclusions 6-65 6.3.3 The effects of a closed season in the German fishery A single month’s closure of the German shrimp fishery was modelled to investigate its likely effect. A 4 weeks’ closure in quarter 3 was used, primarily as during this period the discard levels are at a peak due to the settling of juvenile 0 group flatfish in the German shrimp fishing grounds. The model examined the landings that would arise from the temporary cessation of discarding in this period of closure during quarter 3. Table 6-5 gives following results: • the estimated annual lost landings arising from the current levels of discarding of Plaice, Whiting, Cod and Sole in the German shrimp fishery: these numbers were already calculated in the biological model and refer to the fishery with standard nets during one year, • the estimated annual increase in landings arising from a theoretical closure of 1 month of the fishery (and hence cessation of discarding) in the German shrimp fisheries, • the estimated efficacy of a fishery closure as to reducing discards and consequent lost landings.

Table 6-5: The efficacy of a single month closure in quarter three in the German Crangon fisheries Species Age Estimated annual lost Estimated annual increase Efficacy of closure group landings arising from in landings arising from current levels of discarding the single month closure (% reduction) (tonnes) (tonnes)

Plaice 0 7,556 968 13 % 1 1,157 67 6% 2 13 1 8% All ages 8,727 1,037 11 % Whiting 0 29 1 24% 1 365 56 15% 2 6 1 17% All ages 400 64 16% Cod 0 253 66 26% 1 136 25 18% 2 3 1 33% All ages 392 92 23 % Sole 0 401 116 29% 1 36 2 6% 2 5 0 0% All ages 442 118 27% All efficacy estimates are based on the median range estimates

6.4 Conclusions Under the present assumptions with respect to year class strengths, natural and discard mortality rates, the annual lost landings arising from the current levels of discarding in the European shrimp fisheries are estimated to be around 2,000t for Cod, l,500t for Whiting, 12,000t for Plaice and 600t for Sole. The estimated market value of these landings is over 25 Million €. To realise these potential landings in full a management policy aimed at zero Biological and economic consequences o f discarding in the Brown Shrimp fishery 6-66 discarding in these fisheries would have to be implemented. Partial gains could be realised by the implementation of discard reduction measures such as the use of more selective fishing gears in these fisheries. Except for Plaice, where the potential gains amount to 12 % of the year 2000 North Sea TAC, the expected gains in the round- and flatfish directed fisheries are rather low (2.3% for Cod, 5.1% for Whiting and 3.0% for Sole). The full biological and economic benefits would not be realised for a period of 4-5 years after implementation of a discard reduction measure. The biological simulations show that 72% of the potential gains in Plaice landings, following a reduction of their discard mortalities to zero, would come from the Age 0, and 28% from the Age 1 group. The major part of the Age 0 population that is affected by shrimp fishing, is geographically located in the shallow areas (<10m) of the German and Dutch Waddensea, where they abound between June and December (which includes the major shrimp fishing season). Hence, the major part of the expected gains in Plaice landings coming from the Age 0 group (about 92% or approx. 8,000t out of a total of 8,700 t), could be achieved by a cessation - or at least a drastic reduction - of discarding in these areas. The potential gains in Plaice landings coming from the Age 1 groups (in total approx. 3,300t), on the other hand, are much more evenly distributed between the shrimp fisheries of the different nations. This is related to the wider dispersion of Age 1 group Plaice and their preference for slightly deeper waters. The biological modelling incorporating selective gear in the shrimp fishery demonstrated the efficacy of a selective grid (or similar selective device) in reducing the discarding of 1+ group fish. In the modelling, more than 90% of the lost landings could be ‘recovered’ if the selectivity device was used in waters where Age 1 discards predominate. Conversely, it was shown that where Age 0 fish predominate, the selective device was considerably less effective (particularly for flatfish) and only 21% of the lost German Plaice landings could be ‘recovered’ using this device. Because of the demonstrable spatial difference in the age composition of the discards, a single management regulation that does not take these differences into account is unlikely to realise fully the predicted potential benefits to the Plaice stocks and to their fisheries. One possible approach to resolve the Plaice discard problem, could be the introduction of closed seasons and/or areas aiming at maximum protection of the Age 0 Plaice. However, the modelling of a single month closure in the German fishery during quarter three only ‘recovered’ 11% of the estimated lost German Plaice landings. Increased ‘recovery’ of lost landings can only be achieved through longer periods of closure. Another approach could involve the introduction of species-selective gears, where the emphasis would be on the protection of Age 0 Plaice in certain areas (the flatfish nurseries in the Waddensea area), and on that of Age 1 Plaice in other areas (the coastal waters outside the nurseries). Although this study has produced substantial evidence for the introduction of selective devices with different performance criteria for the different areas, this may lead to difficulties in the legislation. The bio-economic modelling indicates that overall, the introduction of selective gears into these fisheries will represent a net benefit to the EU fisheries, although there will be some individual winners and losers and transfer of profitabilities. The ECODISC project also identified that the common current practice of quantifying discards in fisheries in terms of ‘numbers / weights / percentage of catch’ etc, may be misleading and may not identify the most important elements within a discarded population. For instance, at first glance the numbers of Whiting discarded in the European Crangon fisheries may seem significant at 55 million discarded fish during 1996. The modelling showed that the biological and economic significance of these Whiting discards is relatively small. What is apparent from the modelling is that discard reduction measures in this fishery Conclusions 6-67 should focus on reducing Plaice discards, particularly Age 0 fish in the Waddensea, and Age 1 groups elsewhere (e.g. Belgium). It was identified that these age groups of Plaice are the most important elements of the European Brown Shrimp related discard populations. Protective measures (such as closed areas, closed seasons or selective devices) that primarily aim at a reduction of the unwanted Plaice by-catch, will also reduce the by-catches of other species, both commercial and non-commercial. In doing so, they may contribute to the protection of the marine environment as a whole. The possible establishment of closed areas however, is likely to lead to increased fishing pressure in adjacent areas that remain open to the fisheries, thus countering the benefits that were envisaged by these closures. The bio-economic analysis showed that the European Crangon fishery represents a unique, rather than a common situation. The imposition of inefficiencies through the introduction of more selective gears into the Crangon fisheries is unlikely to reduce profitability. In fact, there is likely to be a net benefit to the EU fishery as a whole from such an introduction. This phenomenon is not likely to be observed in other EU fisheries, and is an artefact of the unique price-quantity relationship that exists forCrangon. Most of the benefits of any increase in white fish landings arising from the introduction of discard reducing measures will go to the flatfish directed fleets, particularly in the Netherlands, the UK and Denmark. This is because of the nature of the current North Sea TAC allocation. Selectivity o f the shrimp beam trawl 7-68

7 Selectivity of the shrimp beam trawl

7.1 Introduction Fishery management measures can be classified into three types. Output controls aim at constraining the amounts of fish caught or landed. Input controls aim at constraining the economic and physical inputs used to catch fish. Technical measures aim at influencing relationships between inputs and outputs. In the EU Coimnon Fisheries Policy, technical measures are important management tools to conserve fish stocks by controlling technical aspects of their capture (e.g. EU regulation 850/98). A substantial part of these measures aim at steering the selection of fish by fishing gears. The selection of fish by fishing gear in the widest possible sense, given in Wileman et al. (1996), is: the process which causes the catch of a gear to have a different composition from that of the fish population in the geographical area in which the gear is being used. There are many causes of these differences with chance playing a big part in the capture process. Gears will select by species and for each species there will also be size selection. Good fisheries management requires that fishing gears should catch the large adult fish while small juveniles are allowed to escape (Annstrong et al., 1990), although other options exist. In Anon. (1994), the main methods to improve the selectivity of towed fishing gears are grouped as 1) cod-end geometry, 2) selector panels, 3) rigid grids and 4) separator panels (Fig. 7.1). These methods all aim at filtering animals already caught in the net. In addition to these, alterations to the front part of the trawl, e.g. the groundrope, can modify the composition of animals entering the trawl, making it less likely that fish incur injury by contact with the net meshes or other parts of the trawl.

Fig. 7.1 - Examples of 1: cod-end 2: selector panel, 3: sorting grid and 4: separator panel.

Traditionally, the selectivity of towed gears has been controlled by prescribing the minimum mesh size of the netting (MacLennan, 1992). Increasing the mesh size in the cod-end is the established method to improve size selectivity, but many other factors affect size selection, only some of which can be controlled. It is the actual opening of the mesh in a cod-end during fishing, as opposed to the nominal mesh size, which determines fish escape. The design and shape of the cod-end (Isaksen and Valdemarsen, 1990; Hickey et al., 1995, Reeves et al., 1992), catch size (Erickson et al., 1996), twine size and stiffness (Ferro and O’Neil, 1994; Lowry and Robertson, 1996) all affect cod-end mesh opening. Fish girth and Introduction 7-69 swimming ability change seasonally and also affect escape and selectivity (Özbilgin, 1996). Also the vessel type can have an influence (Tschemij and Holst, 1999). Regulations exist to control mesh size, cod-end geometry, twine size and potentially obstructive attachments used to aid handling and prevent wear but this cannot prevent selectivity being variable. As the main aim of this project was to improve the selectivity of the shrimp beam trawl, estimating the selective characteristics of the standard trawl was an essential step. Based on existing selectivity parameters for commercial fish species (Wileman, 1992), it can be concluded that the shrimp beam trawl is nearly non-selective for fish. Due to the small body size of the shrimps, small meshes (minimum legal mesh size 16mm - EU regulation 850/98) are, however, necessary to obtain commercially viable catches and increasing the mesh size is not an option to improve selectivity for the by-catch. The application of selector panels (Fig. 7.1) is based on the differences in behaviour of the species entering a trawl. Some species will show the appropriate behavioural response to a panel and escape through its meshes while other species will not react and be caught. Several authors have shown that a square mesh panel rigged in the rear top panel of the net can be quite effective for releasing Whiting and Haddock {Melanogrammus aeglefinus) (e.g. Armstrong et al., 1998; Graham and Kynoch, 2001). This selective device has proved effective and is part of the EU technical measures for towed gears. The selector panel concept as a tool to improve the selectivity of trawls is not a recent invention but was already tested by Ridderstad (1915) in the Kattegat and the Baltic Sea. It is, however, quite unlikely that selector panels would be successful in the Brown Shrimp fishery. Whiting may use the opportunity to escape since it tends to swim upwards in a trawl (Main and Sangster, 1982) but flatfish, on the other hand, will show the opposite response (Fonteyne et al., 1997). In addition, some part of the commercial shrimp catch may be released through a selector panel. Rigid grids (Fig. 7.1) in trawls aim at separating the catch based on differences in physical and behavioural characteristics of species. In principle, these devices fdter the catch; animals with a certain shape can pass through the grid and the others are guided to an outlet or other compartment of the trawl. Different rigid grids have been developed for several by-catch problems in fisheries (Anon., 1996). The devices have been quite successfully implemented in the northern shrimp fisheries (see Section 8). In the Norwegian and CanadianPandalus fishery, grids are compulsory to reduce by-catches of undersized fish. Although the situation in the North Sea Brown Shrimp fishery is quite different, especially with regard to catch composition, it was thought, based on the success in the Pandalus fishery, worthwhile investigating this device (see Chapter 8). The general aim of separator panels as a selectivity tool is to take advantage of differences between the species in a trawl catch. These differences may be morphological, i.e. differences in physical size and/or shape, e.g. the separation of crustacean speciesNephrops ( norvegicus, Parapenaeus longirostris and Aristeus antennatus) from fish species like hake {Merluccius merluccius) in the Portugese fishery (Campos et al., 1996). These differences may also be behavioural, e.g. in the separation of Whiting, Haddock and Cod in the North Sea otter trawl fishery (Main and Sangster, 1985; Arkley and MacMullen, 1996). By exploiting these differences, the various species of a mixed catch can be segregated in order that different selection methods can be applied to optimise the selectivity of individual or similar species. In Anon. (1994), it is stated that one of the fisheries that were demonstrated to benefit from the separator principle is the shrimp fishery. This, together with the observation that differences in physical size and shape exist in the species mix caught in the Crangon fishery, led to a further exploration of this concept (see Chapter 9). Selectivity o f the shrimp beam trawl 7-70 A study of the selectivity of the commercially used Brown Shrimp beam trawl was an essential step in the search to find technical solutions for altering the trawl to improve the selectivity and reduce the discards. In general, selectivity studies concentrate on the cod-end. Observations made by divers and towed underwater vehicles show that large amounts of fish escape from the cod-end and for most species this is where the main mesh selection is thought to occur (Wileman et al., 1996). Hillis and Earley (1982), however, demonstrated that the selectivity of the net can be far more important than cod-end selectivity when they found that over 40% of Nephrops norvegicus entering a typical Irish prawn otter trawl escaped through the net body meshes, compared to only 10% through the cod-end meshes. A study by Thorsteinsson (1981) showed that many small shrimps(.Pandalus borealis) escaped through the side panels of the net. Experiments carried out by Bohl (1963) comparing two nets with identical cod-ends but with different mesh sizes in the net also demonstrated that some part of the shrimps (C. crangon ) escaped through the net meshes. Sakhno and Sadokhin (1980) on the other hand, found that only 5% of the shrimps(,Pandalus borealis) escaped through the net parts in front of the cod-end. To clarity the situation for the Brown Shrimp beam trawl, it was decided to study whole trawl selectivity and not only cod-end selectivity. To complete the picture, the selectivity of the groundrope was included in the study. The selectivity experiments were carried out during six sea trips in May and November ’95, March and November ’96 and December ’97 and were entirely carried out by the author and the staff of DvZ.

7.2 Materials and methods

7.2.1 The fishing area, the vessel and the trawl The sea trials were carried out on board of the RV (research vessel) “Belgica” which has an overall length of 50.9 m, a GRT of 765 t and an engine power of 1154 kW. Since a large team was necessary to handle the gear, empty the cod-ends and pockets and measure the many catch fractions from each haul, it was not feasible to use a commercial vessel for the selectivity trials. A commercial skipper, however, was hired to select the fish tracks and to guide the fishing operations in order to match as closely as possible commercial conditions. The shrimp fishing grounds on the “Vlakte of the Raan”, located in Belgian coastal waters in ICES sub-division IVc, were fished during the six sea trips. Due to safety measures and limited depth of the commercial fishing grounds, the experiments were carried out in this limited area. The towing speed was between 2.5 and 3 knots and the warp length was three times the water depth. The gear studied was a commercial shrimp beam trawl (Fig. 7-2) with a beam length of 8m and a vertical net opening of 0.5 m. The lengths of the headline and the groundrope were 7.8m and 9.8m respectively. The bobbin rope consisted of rubber bobbins with a diameter of 21cm rigged on steel axles with a diameter of 20mm. The net was made of knotted polyamide netting with nominal mesh sizes of 28mm in the front part decreasing to 22mm in the aft part. The cod-end was made of knotted polyamide netting with a nominal mesh size of 22mm and protected by a polyamide lifting bag with a nominal mesh size of 80mm. A net plan is given in Fig. 7-3. Materials and methods 7-71

cod-end cover half hoop

beam

bobbin rope

trawlhead

Fig. 7-2 - A commercial shrimp beam trawl.

76 1N2B 2T2B

Ssr.jird tonton (m | Agricultural Research Centre Ghent

Ankerstroat 1. 6400 Oostende. Belgium i shwnp (Crangon crangon) CopyngN d j logo«! CENTRE NATIONAL 0E LA MER / IFREMER

Fig. 7-3 - Plan of the shrimpnet as used during the experiments. Selectivity o f the shrimp beam trawl 7-72 During each haul the following variables were recorded: haul duration (minutes), speed through the water (knots), state of the sea, total catch volume (liters), clogging of the cod-end meshes, water temperature (°C) and light conditions (W/m2). The state of the sea was expressed as the standard deviation of the wave heights recorded each second by the heave compensator on the vessel’s echo sounder. The clogging of the cod-end meshes was a subjective measure of the amounts of seaweed and hydroids that were stuck on the mesh bars, thus decreasing the opening of the meshes. This was a number between 0 (fully open meshes) and 10 (fully clogged meshes), always recorded by the same person.

7.2.2 The cod-end and the net covers The cod-end selectivity was determined with the covered cod-end technique (Wileman et al., 1996). The cover was constructed of knotted polyamide netting with a nominal mesh size of 11 îmn, was 800 meshes on the circumference and had a sufficient length to leave an open space of 1.5m behind the aft end of the cod-end. The cover was held open by a half hoop over the top panel of the cod-end and had a diameter of 1.5 m. The application of full hoops was not feasible due to frequent damage caused by the close bottom contact of beam trawls. To study whole trawl selectivity, the net was subdivided into different sections (Fig. 7-4) to be sampled for escaping shrimps and fish. Longitudinally as well as cross-sectional (vertically), the net was subdivided into four sections. Each section was named with a letter, A, B, C or D, indicating the cross-sectional (vertical) position and a number, 1, 2, 3 or 4, indicating the longitudinal position. The cross-sectional subdivision was based upon the shape of the netting observed during earlier flume tank trials. Thus, the top panel as well as the belly of the net was divided in flat (A and D) and rounded (B and C) pieces of netting. For the longitudinal subdivision, the first section was located in the area of the bobbin rope. The net behind the bobbin rope was divided in three parts, based on the different mesh sizes.

8m

A,B,C,D,E: vertical subdivision

1, 2, 3, 4: longitudinal subdivision

Fig. 7-4 - The sub-division of the net into the sections that were sampled with small mesh pockets.

Each of the sixteen sections was provided with a small mesh pocket to collect the escapees. The material and mesh size of the pockets were the same as the cod-end cover. The size of the pockets was kept small in order to keep the change in the waterflow through the net meshes as small as possible. Each pocket covered a rectangular area of the net. The opening of the pockets measured 120 by 69 pocket meshes or 75 by 69 pocket meshes depending on whether they covered the flat net sections (A and D) or the rounded net sections (B and C), Materials and methods 7-73 the latter being too narrow to hold a wide pocket. The number of net meshes covered by the opening of the pockets ranged from 1100 to 2800 depending on the size of the pocket and on the mesh size in the net. The total number of meshes in a section divided by the number of net meshes covered by a pocket, rigged to that section, was the raising factor used to estimate from the pocket catch, the escapes from the whole section. The pockets were rigged with small floats or weights, if necessary, to keep the masking of the net meshes as small as possible. Just behind the groundrope, underneath the belly of the net, two small mesh nets were rigged to catch the animals escaping underneath the groundrope. Each net covered one metre of the path of the trawl. The net, the cover and the pockets were observed with an underwater camera. The netting of the cover and the pockets was seen to be well away from the net meshes during the fishing operation. Due to the small size of the meshes, the ICES mesh gauge could not be used to measure the opening of the meshes. A calliper was used, stretching the meshes with a weight of 2 kg. The weight of 2 kg was chosen according to the rules set up in EU-regulation 2108/84 for measuring meshes with a size below 35 mm. The mesh openings were measured on several occasions during the sea trips in series of 25 meshes per panel. The averages and standard deviations for the different net panels and covers are given on the net plan (Fig. 7-3).

7.2.3 Data collection and analysis The species investigated were Brown Shrimp, Whiting, Cod, Plaice, Sole and Dab, provided sufficient numbers were available in the catch. For a description of the general method of data collection, see section 5.2.3 in this document. The fish and the shrimps from the cod-end cover were manually sorted. All fish were measured and a sample of 1.5 litres of shrimps was taken to be measured in the laboratory. The catches in the small mesh pockets were collected in numbered buckets. All fish as well as all shrimps were measured immediately after the haul on board of the vessel. The cod-end selectivity was investigated for the four species measured. The SELECT model was chosen to describe the selectivity. The standard methodology for selectivity of fishing gears is described in Wileman et al. (1996). Based on the deviance residuals obtained when calculating the selection curves, the logistic function was chosen as a link function to fit the retention points for each species and fitted the data very well. This function is the cumulative distribution function of a logistic random variable and is specified by the following equation:

RR(TL) = exp (a + b • TL) / (1 + exp (a + b • TL)) where RR(TL) is the probability that an animal of length TL (Total Length) is retained in the cod-end. a and b, which are the two parameters to be estimated, represent the intercept and the slope, respectively, after a logit transformation. These parameters were estimated with the maximum likelihood method by the CC software (Constat, Denmark). L25, L50 and L75 are the body lengths at which 25%, 50% and 75% of the shrimps are retained in the cod-end. SF is the selection factor and is the L50 divided by the mesh size. SR is the selection range and is equal to the difference between L75 and L25 and gives an idea of the slope of the curve. Selectivity o f the shrimp beam trawl 7-74 Single hauls were combined by the variance component analysis method of Fryer (1991) by the CC software. 95% confidence limits of the selection parameters are given in brackets, in the text as well as in the tables. A data exploration of the selectivity parameters was conducted to get an idea of the variability and correlations with the covariates haul duration, speed through the water, state of the sea, total catch volume, clogging of the meshes, water temperature and light conditions. To determine whole trawl selectivity the numbers of shrimp and fish escaping from the different net sections and underneath the groundrope were calculated for each observed length class.

7.3 Results

7.3.1 Fishing conditions A log of the sea trials is given in Table 7-1. For shrimp, 63 successful hauls were conducted, of which 59 had a cod-end cover to determine cod-end selectivity. For 39 hauls the catches of the pockets were measured and these were valid for the whole trawl selectivity calculations. Shrimps with lengths within the SR were caught in sufficient numbers in each haul to calculate cod-end selectivity. Shrimps below 35mm were caught in low numbers relative to the total catch. In absolute numbers, however, in most hauls there were sufficient shrimps to calculate reliable retention points in the lower length range. In the upper length range numbers were always high. The length frequency distributions of all shrimps used to calculate cod-end selectivity and whole trawl selectivity are given in Figs 7-5 and Fig. 7-6 respectively. A total of 23 hauls was carried out to determine the cod-end selectivity for fish (Table 7-1). Dab and Plaice occurred in all 23 hauls, but Sole was only found in 8 hauls. Whiting and Cod were not caught in sufficient numbers to carry out any selectivity calculations. The length distributions of Dab, Plaice and Sole in the cod-end and cod-end cover are given in Fig. 7-7, Fig. 7-8 and Fig. 7-9 respectively. For Dab, 8 hauls contained sufficient numbers of fish in the cod-end and cover to compute single haul selectivity. The other 15 hauls were pooled to determine one combined ogive. For none of the hauls, however, were the numbers of fish in the cod-end cover sufficiently high to establish reliable confidence limits. No Plaice were found in the cod-end cover. For Sole, 7 hauls were used for selectivity calculations of which one contained not enough fish in the cover to calculate confidence limits. For 39 hauls, the small mesh pockets on the net were analysed for the presence of fish to determine the whole trawl selectivity (Table 7-1). Fish escaping through the meshes of the net body, however, were very rarely observed and negligible compared to the total catch. Results 7-75 Table 7-1: Log of the experimental hauls

haul date Cod-end Whole trawl L50 (mm) SR (mm) Cod-end number selectivity selectivity selectivity shrimp shrimp shrimp shrimp Fish 95/13-01 29-05-95 ves no 43.4 ('42.7-44.1 ~) 7.9 66.9-8.90 no 95/13-03 30-05-95 yes no 45.9 (26.4-65.4) 8.0 (-0.2-16.1) no 95/13-04 30-05-95 yes no 18.0 (12.7-23.3) 26.9 (20.3-33.6) no 95/13-05 30-05-95 yes no 34.4 (32.4-36.4) 13.1 (11.1-15.2) no 95/13-06 30-05-95 yes no 21.6 (15.0-28.2) 31.7 (21.8-41.7) no 95/13-08 30-05-95 yes no 21.0 (17.6-24.5) 27.2 (21.6-32.8) no 95/13-09 30-05-95 yes no 0. 8 (-76-9.2) 41.8 (17.1-66.5) no 95/13-12 31-05-95 yes no 30.9 (29.5-32.4) 17.1 (15.1-19.2) no 95/13-13 31-05-95 yes no 39.3 (38.1-40.4) 17.2 (15.0-19.4) no 95/13-14 31-05-95 yes no 39.9 (37.8-42.1) 15.6 (12.7-18.5) no 95/13-15 31-05-95 yes no 33.4 (31.0-35.7) 19.8 (15.9-23.6) no 95/13-16 31-05-95 yes no 23.4 (19.6-27.2) 17.0 (13.7-20.4) no 95/27-03 13-11-95 yes yes 43.1 (42.3-44.0) 11.4 (10.1-12.8) no 95/27-05 13-11-95 yes yes 39.4 (38.7-40.1) 8.8 (7.9-9.7) no 95/27-06 13-11-95 yes yes 36.1 (35.1-37.2) 11.2 (10.0-12.5) no 95/27-07 13-11-95 yes yes 38.5 (37.8-39.3) 8.8 (7.9-9.7) no 95/27-08 14-11-95 yes yes 34.8 (33.5-36.1) 12.1 (10.6-13.5) no 95/27-09 14-11-95 yes yes 35.0 (33.5-36.5) 11.3 (9.6-13.0) no 95/27-10 14-11-95 yes yes 41.8 (41.2-42.5) 8.1 (7.3-9.0) no 95/27-11 14-11-95 yes yes 37.2 (36.3-38.1) 12.5 (11.3-13.8) no 95/27-12 14-11-95 yes yes 28.5 (25.1-31.8) 17.0 (13.3-20.6) no 95/27-13 14-11-95 no yes (-) (-) no 95/27-14 14-11-95 no yes (-) (-) no 95/27-15 15-11-95 yes yes 36.5 (35.4-37.6) 11.1 (9.7-12.5) no 95/27-16 16-11-95 no yes (-) (-) no 95/27-17 16-11-95 yes yes 35.1 (33.3-36.9) 13.1 (10.8-15.5) no 95/27-18 16-11-95 yes yes 38.0 (37.1-38.9) 7.7 (6.6-8.7) no 95/27-19 16-11-95 yes yes 41.1 (40.3-42.0) 8.2 (6.9-9.5) no 95/27-20 16-11-95 yes yes 25.7 (22.6-28.9) 10.3 (8.4-12.1) no 95/27-21 16-11-95 yes yes 38.9 (37.7-40.0) 11.8 (9.7-14.0) no 95/27-22 16-11-95 yes yes 35.8 (34.8-36.9) 14.9 (12.8-16.9) no 95/27-23 16-11-95 yes yes 39.9 (39.1-40.7) 9.7 (8.3-11.2) no 96/05-02 04-03-96 yes yes 42.1 (41.3-43.0) 10.5 (9.0-12.0) no 96/05-03 04-03-96 yes yes 42.6 (42.1-43.1) 8.5 (7.6-9.3) no 96/05-04 04-03-96 yes yes 36.8 (-3.2-76.8) 14.6 (-10.-39.9) no 96/05-05 05-03-96 yes yes 45.6 (45.1-46.1) 8.0 (7.1-9.0) no 96/05-06 05-03-96 yes yes 40.8 (39.9-41.6) 11.6 (9.9-13.2) no 96/05-07 05-03-96 yes yes 42.5 (41.8-43.2) 9.0 (7.9-10.1) no 96/05-08 05-03-96 yes yes 42.3 (41.5-43.0) 10.4 (9.1-11.6) no 96/05-10 06-03-96 yes yes 39.3 (38.5-40.1) 9.6 (8.3-10.9) no 96/05-11 06-03-96 yes yes 47.7 (47.1-48.2) 8.9 (8.0-9.9) no 96/05-12 06-03-96 yes yes 41.7 (41.0-42.3) 10.0 (8.9-11.1) no 96/05-13 06-03-96 no yes (-) (-) no 96/05-14 06-03-96 yes yes 45.9 (45.1-46.7) 7.9 (6.9-8.9) no 96/25-01 04-11-96 yes yes 40.2 (18.3-62.1) 9.4 (-1.2-20.1) yes 96/25-03 04-11-96 yes yes 42.4 (12.0-72.8) 8.2 (-5.2-21.5) yes 96/25-04 04-11-96 yes yes 40.4 (39.4-41.4) 9.8 (8.4-11.2) yes 96/25-05 04-11-96 yes yes 43.4 (42.3-44.5) 12.8 (11.0-14.6) yes 96/25-06 04-11-96 yes yes 40.4 (38.1-40.9) 10.7 (10.1-14.6) no 96/25-07 04-11-96 yes yes 41.6 (40.6-42.5) 10.2 (8.8-11.6) yes 96/25-08 04-11-96 yes yes 39.8 (14.8-64.7) 8.4 (-3.1-19.9) yes 97/28-01 01-12-97 no no (-) (-) yes 97/28-02 01-12-97 no no (-) (-) yes 97/28-03 01-12-97 no no (-) (-) yes 97/28-04 01-12-97 no no (-) (-) yes 97/28-05 01-12-97 no no (-) (-) yes 97/28-06 01-12-97 yes no 44.8 (43.6-46.0) 12.5 (9.8-15.1) yes 97/28-07 02-12-97 yes no 42.6 (41.3-43.9) 15.3 (12.7-17.9) yes 97/28-08 02-12-97 yes no 43.7 (42.2-45.1) 22.1 (17.7-26.6) yes 97/28-09 02-12-97 yes no 37.4 (36.4-38.3) 10.1 (8.8-11.5) yes 97/28-10 02-12-97 yes no 43.8 (41.9-45.7) 18.4 (13.8-23.0) yes 97/28-11 02-12-97 yes no 45.9 (45.0-46.7) 9.8 (8.4-11.2) yes 97/28-13 03-12-97 yes no 47.8 (47.1-48.5) 9.2 (8.1-10.3) yes 97/28-15 03-12-97 yes no 44.0 (43.0-45.0) 11.7 (9.7-13.7) yes 97/28-16 03-12-97 yes no 35.3 (32.5-38.1) 23.0 (17.6-28.4) yes 97/28-17 03-12-97 yes no 44.0 (42.8-45.2) 13.6 (11.1-16.1) yes 97/28-18 03-12-97 yes no 41.6 (40.3-42.8) 14.5 (11.7-17.3) yes 97/28-19 03-12-97 yes no 46.8 (46.1-47.4) 8.6 (7.7-9.5) yes Selectivity o f the shrimp beam trawl 7-76 Table 7-1 (continued): Log of the experimental hauls

haul date Whole trawl L50 (mm) SR (mm) L50 (mm) SR (mm) number selectivity

Fish Dab Dab Sole Sole 95/13-01 29-05-95 no t-) M M n 95/13-03 30-05-95 no (-) (-) (-) (-) 95/13-04 30-05-95 no (-) (-) (-) (-) 95/13-05 30-05-95 no (-) (-) (-) (-) 95/13-06 30-05-95 no (-) (-) (-) (-) 95/13-08 30-05-95 no (-) (-) (-) (-) 95/13-09 30-05-95 no (-) (-) (-) (-) 95/13-12 31-05-95 no (-) (-) (-) (-) 95/13-13 31-05-95 no (-) (-) (-) (-) 95/13-14 31-05-95 no (-) (-) (-) (-) 95/13-15 31-05-95 no (-) (-) (-) (-) 95/13-16 31-05-95 no (-) (-) (-) (-) 95/27-03 13-11-95 yes (-) (-) (-) (-) 95/27-05 13-11-95 yes (-) (-) (-) (-) 95/27-06 13-11-95 yes (-) (-) (-) (-) 95/27-07 13-11-95 yes (-) (-) (-) (-) 95/27-08 14-11-95 yes (-) (-) (-) (-) 95/27-09 14-11-95 yes (-) (-) (-) (-) 95/27-10 14-11-95 yes (-) (-) (-) (-) 95/27-11 14-11-95 yes (-) (-) (-) (-) 95/27-12 14-11-95 yes (-) (-) (-) (-) 95/27-13 14-11-95 yes (-) (-) (-) (-) 95/27-14 14-11-95 yes (-) (-) (-) (-) 95/27-15 15-11-95 yes (-) (-) (-) (-) 95/27-16 16-11-95 yes (-) (-) (-) (-) 95/27-17 16-11-95 yes (-) (-) (-) (-) 95/27-18 16-11-95 yes (-) (-) (-) (-) 95/27-19 16-11-95 yes (-) (-) (-) (-) 95/27-20 16-11-95 yes (-) (-) (-) (-) 95/27-21 16-11-95 yes (-) (-) (-) (-) 95/27-22 16-11-95 yes (-) (-) (-) (-) 95/27-23 16-11-95 yes (-) (-) (-) (-) 96/05-02 04-03-96 yes (-) (-) (-) (-) 96/05-03 04-03-96 yes (-) (-) (-) (-) 96/05-04 04-03-96 yes (-) (-) (-) (-) 96/05-05 05-03-96 yes (-) (-) (-) (-) 96/05-06 05-03-96 yes (-) (-) (-) (-) 96/05-07 05-03-96 yes (-) (-) (-) (-) 96/05-08 05-03-96 yes (-) (-) (-) (-) 96/05-10 06-03-96 yes (-) (-) (-) (-) 96/05-11 06-03-96 yes (-) (-) (-) (-) 96/05-12 06-03-96 yes (-) (-) (-) (-) 96/05-13 06-03-96 yes (-) (-) (-) (-) 96/05-14 06-03-96 yes (-) (-) (-) (-) 96/25-01 04-11-96 yes 4.4 (-) 0.2 (-) 7.9 (-) 0.3 (-) 96/25-03 04-11-96 yes 4.8 (-) 0.4 (-) 7.9 (7.2-8.3) 1.8 (1.0-2.7) 96/25-04 04-11-96 yes 4.3 (-) 0.6 (-) 8.3 (7.7-8.8) 1.3 (0.6-2.1) 96/25-05 04-11-96 yes 4.9 (-) 0.4 (-) 8.6 (7.7-9.3) 1.7(0.8-2.7) 96/25-06 04-11-96 yes (-) (-) 7.8 (7.1-8.3) 1.7(0.8-2.7) 96/25-07 04-11-96 yes 5.2 (-) 0.6 (-) 7.8 (7.1-8.3) 1.7(0.9-2.5) 96/25-08 04-11-96 yes combined 7.5 (5.6-8.1) 2.3 (0.8-3.9) 97/28-01 01-12-97 no com bined (-) (-) 97/28-02 01-12-97 no com bined (-) (-) 97/28-03 01-12-97 no com bined (-) (-) 97/28-04 01-12-97 no com bined (-) (-) 97/28-05 01-12-97 no com bined (-) (-) 97/28-06 01-12-97 no com bined (-) (-) 97/28-07 02-12-97 no 5.1 (-) 0.3 (-) (-) (-) 97/28-08 02-12-97 no com bined (-) (-) 97/28-09 02-12-97 no com bined (-) (-) 97/28-10 02-12-97 no com bined (-) (-) 97/28-11 02-12-97 no com bined (-) (-) 97/28-13 03-12-97 no 4.8 (-) 0.4 (-) (-) (-) 97/28-15 03-12-97 no com bined (-) (-) 97/28-16 03-12-97 no com bined (-) (-) 97/28-17 03-12-97 no com bined (-) (-) 97/28-18 03-12-97 no 4.6 (-) 1.5 M (-) (-) 97/28-19 03-12-97 no com bined (-) (-) Combined hauls (*) 4.8 (-) 0.3 (-) (*) : the selectivity data for the combined hauls marked as combined in the table Results 7-77 7.3.2 Cod-end selectivity The L50 for shrimp for all hauls combined was 39.4 mm (37.0 - 41.8) and the selection factor 1.82 (1.71 - 1.93). The selection range was 11.6 ïmn (10.2 - 13.0). The selection ogive is shown in bold light blue in Fig. 7-5.

50000 - 100%

37500 75%

25000 50%

12500 25%

0% 20 30 40 50 Length class (mm TL)

Fig. 7-5 - The estimated overall selectivity ogive (in bold light blue) together with the 59 selection ogives of the single hauls and the length distribution of the cod-end + cod-end cover catch for shrimp.

60000

40000

20000

20 30 40 50 70 80 Length class (mm TL)

Fig. 7-6 - The length distribution of all shrimps in the trawl path of the 39 hauls used in the whole trawl selectivity calculations.

The selection ogives for the individual hauls are shown in Fig. 7-5 and the L50s and SRs are given in Table 7-1. Most of the L50s lay within the 35-45 ïmn range. Nine hauls had a low L50, below 30 ïmn. Most of the SRs lay within the 5-15 ïmn range. Eight hauls had a high Selectivity o f the shrimp beam trawl 7-78 SR, above 20 nun. Six of these hauls also had a very low L50 which pointed at a possible high association between L50 and SR. The correlation coefficient (R = -0.79) was highly significant (p < 0.001) and denoted a negative association between L50 and SR. A linear regression between L50 and SR indicated a slope significantly different from 0 (p < 0.001), also without the outlier (0.8 , 41.8) and an Rousted = 0.63. A visual inspection of the data, however, showed a strong scatter around the regression line and a limited number of observations in the lower L50 range, which weakened the conclusion of a clear relationship. The correlation coefficients of L50 and SR with the covariates are given in Table 7-2. Light was not significantly correlated with L50 nor with SR. Catch volume and clogging were not significantly correlated with SR. All other covariates were associated with L50 or SR at a p- level < 0.05 (Table 7-2) and had an influence on the cod-end selectivity for shrimps.

Table 7-2: The correlation coefficients of L50 and SR with the possible explanatoryvariables

L50 p-level SR p-level

Haul duration 0.41 0.001 -0.31 0.016 Speed through the water 0.32 0.013 -0.38 0.003 State of the sea 0.47 0.000 -0.38 0.003 Catch volume -0.49 0.000 0.16 0.220 Clogging of tile meshes -0.78 0.000 0.79 0.101 Water temperature -0.46 0.000 0.34 0.008 Light conditions 0.10 0.436 -0.19 0.151

Correlations in italics were significant

For Dab the L50 for all hauls pooled was 4.5 cm and the selection factor 2.1. The selection range was 0.7 cm. The selection ogive is shown in bold light blue in Fig. 7-7. The selection parameters for the 8 single hauls and the 15 combined hauls are given in Table 7-1. The selection ogives are given in Fig. 7-7. The range of the selectivity parameters was too low to find any relation with the covariates.

400 100%

300 75%

200 50%

100 25%

0%

Length class (cm TL)

Fig. 7-7 - The estimated selectivity ogives for Dab together with the length distribution of the cod- end + cod-end cover catch (the overall selectivity ogive is drawn bold). Results 7-79 Plaice with lengths down to 5 cm were caught but no cod-end selection was observed. The length frequency distribution of the cod-end (and thus total) catch is given in Fig. 7-8. For Sole the L50 for all hauls pooled was 7.9 cm and the selection factor 3.6. The selection range was 1.7 cm. The selection ogive is shown in bold light blue in Fig. 7-9. The selection parameters for the 7 single hauls are given in Table 7-1. The selection ogives are given in Fig. 7-9. As for Dab, the range of selectivity parameters was too low to find any relation with the covariates.

160

120

80

40

20 Length class (cm TL)

Fig. 7-8 - Length distribution of the cod-end catch for Plaice (no Plaice observed in cod-end cover).

200 100%

150 75%

100 50%

25%

0%

Length class (cm TL)

Fig. 7-9 - The estimated selectivity ogives for Sole together with the length distribution of the cod- end + cod-end cover catch (the overall selectivity ogive is drawn bold light blue).

7.3.3 Whole trawl selectivity In Fig. 7-10A and B the percentages of shrimps escaping from the different net sections are shown based on the cross-sectional and longitudinal subdivision of the net respectively. The areas indicate the relative amounts of shrimps escaping from each section. A moving average Selectivity o f the shrimp beam trawl 7-80 of three values has been applied to smooth the charts. The bar charts in Fig. 7-11 show each net section separately based on 3 mm length classes. Table 7-3 shows the percentages of shrimps < 45 ïmn TL, > 45 ïmn TL and all length classes combined escaping from the different net sections.

A - All sections - based on vertical subdivision 100%

80%

60%

40%

20%

0% o o lf} o o co "3- ITJ to r- Length class (rnn 71_)

B Codend a *e r □ Coderd □ Secbcn A □ Section B □ Section C □ Secbon D ■ Gand ope

Length class (rrmTL)

Fig. 7-10 - Percentage of the total number of shrimps in the trawl path escaping from the different net sections and retained in the cod-end cover.

It is clear from Fig. 7-10 that most of the larger shrimps end up in the cod-end. It is striking, however, that for the smaller shrimps the net selectivity is far more important than the cod- end selectivity. Of all shrimps < 45 mm TL, 44 % escape through the net meshes and only 23 Results 7-81 % escape through the cod-end meshes (Table 7-3). The selection through the meshes of the net is quite length dependent (Fig. 7-1 IB). For the smallest shrimps measured, over 60 % escape through the net meshes. This steadily decreases to reach a very low selection for the highest length classes. About 8 % of all shrimps escape underneath the ground rope (Table 7- 3), without a clear length dependence (Fig. 7-11 A). The data for the cross-sectional sub-division of the net (Fig. 7-10A) indicate that net section C, which is the rounded side part of the belly, shows the highest escapes (24 %). Section A, B and D, which are the upper parts of the net and the flat section of the belly of the net, are of very low importance for the selection of shrimps, especially when compared to Section C (Fig. 7-11C, D, E and F). Less than 6 % of the escapes occur through these three sections together. Except for Section A, it is evident that small shrimps can escape more easily than larger ones. Selectivity o f the shrimp beam trawl beam shrimp f the o Selectivity Fig. 7-11 - Percentages of the total number of shrimps in the trawl path escaping from each net each from escaping path trawl the in shrimps of number total the of Percentages - 7-11 Fig. section separately. section Percentage Percentage Percentage 100 25 50 75 68 1 4 7 0 3 6 9 2 5 8 1 4 7 0 3 76 73 70 67 64 61 58 55 52 49 46 43 40 37 34 31 2628 2628 31 34 37 40 43 46 49 52 55 58 61 64 67 70 73 76 2628 31 34 37 40 43 46 49 52 55 58 61 64 67 70 73 76 Underneath groundrope Underneath 3m m length class (mm TL) (mm class length m 3m Section C Section Section A Section I l R fl 1 T l H □ H B fl c E f 100 100 25 75 75 25 50 2628 31 34 37 40 43 46 49 52 55 58 61 64 67 70 73 76 26 28 31 34 37 40 43 46 49 52 55 58 61 64 67 70 73 76 l i S S i f S S fi i fl d I 3m m length class (mm TL) (mm class length m 3m Net (without cod-end) (without Net eto B Section eto D Section R i I...I 1 n n □ ...... „ D . _ D n „ □ n n » » n n D B F r

7-82

Results Fig. 7.11 (continued - Percentages of the total num ber of shrimps in the traw l path escaping from from escaping path l traw the in shrimps of ber num separately. total the section of net each Percentages - (continued 7.11 Fig. Percentage Percentage Percentage 100 I—100 25 75 68 1 4 7 0 3 6 9 2 5 8 1 4 7 0 3 76 73 70 67 64 61 58 55 52 49 46 43 40 37 34 31 2628 76 73 70 67 64 61 58 55 52 49 46 43 40 37 34 31 2628 76 73 70 67 64 61 58 55 52 49 46 43 40 37 34 31 2628 68 1 4 7 0 3 6 9 2 5 8 1 4 7 0 3 76 73 70 67 64 61 58 55 52 49 46 43 40 37 34 31 2628 in I ______i l i l 3m m length class (mm TL) (mm class length m 3m . i Section 3 Section 4 Section 3 Section 2 Section 1 Section Cod-end ...... i i L i L 1 fl ...... fl flfl fl n / G K Í 100 1-100 100 25 ■ 50 75 75 25 50 68 1 4 7 0 3 6 9 2 5 8 1 4 7 0 3 76 73 70 67 64 61 58 55 52 49 46 43 40 37 34 31 2628 68 1 4 7 0 3 6 9 2 5 8 1 4 7 0 3 76 73 70 67 64 61 58 55 52 49 46 43 40 37 34 31 2628 1 I ¡ . 3m m length class (mm TL) (mm class length m 3m Cod-end cover Cod-end l fifl a ¡ t

...... „ □ Mean ean-SD M ean+SD M ...... H J 7-83 Selectivity o f the shrimp beam trawl 7-84 Table 7-3: Percentages of shrimps, < 45 mm TL, > 45 mm TL and all length classes, escaping from the different net sections (with 95 % confidence limits between brackets)

<45 mm TL (%) >=45 mm TL (%) All (%) Cod-end cover 22.6 (19.5 -■25.6) 6.9 ( 5.6- ■ 8.2) 13.4 (11.3 - 15.6) Cod-end 24.2 (21.5 -■26.9) 65.0 (61.6 - ■68.3) 48.4 (44.7 - - 52.0) Net (without cod-end) 44.3 (42.4 -■46.2) 19.4 (17.4- ■21.4) 29.6 (27.5 - 31.6) Section 1 17.8 (16.3 -■ 19.4) 11.3 ( 9.6- ■ 13.0) 13.9 (12.5 - 15.3) Section 2 10.7 ( 9.8-■ 11.6) 3.7 ( 3.0-■ 4.3) 6.4 ( 5.8- 7.1) Section 3 13.8 (12.3 -■ 15.4) 4.2 ( 3.3-■ 5.0) 8.2 ( 7.0- 9.4) Section 4 2 ( 1.4-■ 2.5) 0.3 ( 0.2-■ 0.3) 1.0 ( 0.7- 1.4) Section A 0.4 ( 0.2-■ 0.6) 0.7 ( 0.4-■ 1.0) 0.6 ( 0.3- 0.8) Section B 2.5 ( 1-5-■ 3.5) 1.2 ( 0.8-■ 1.6) 1.8 ( 1-1- 2.5) Section C 37.4 (35.6- ■39.3) 14.8 (13.3 -■ 16.4) 24.0 (22.5 - 25.5) Section D 4 ( 3.3-■ 4.7) 2.7 ( 2.1-■ 3.3) 3.2 ( 2.6- 3.8) Underneath groundrope 8.8 ( 7.5- ■ 10.2) 9.0 ( 7.2- ■ 10.8) 8.8 ( 7.4- 10.3)

Based on the longitudinal subdivision (Fig. 7-10B) the differences in selectivity between sections are less than between the cross-sectional ones. Fig. 7-11G, H, I and J, however, indicate that escape opportunities for shrimps decrease gradually from the anterior part to the posterior part of the net, especially for the larger animals. This is because the mesh size decreases and because meshes tend to close more towards the aft end of the net. Section 1 accounts for 14 % of the escapes (Table 7-3). Contrary to the other sections, the smallest shrimps do not escape as easily as the mid length range from this anterior section. Sections 2, 3 and 4 account for 6 %, 8 % and 1 % respectively (Table 7-5). This very low selectivity in Section 4 is probably because this is the section with virtually closed meshes due to the tension in the net. The bulk of the larger shrimps (65 %) finish up in the cod-end (Fig. 7-1 IK) (Table 7-3), and a substantial amount (24 %) of non-marketable shrimps (< 45 mm) are caught in the cod-end. About 35 % of the marketable shrimps do not end up in the commercial catch and escape underneath the groundrope (9 %), through the meshes of the net (19 %) or through the cod- end meshes (7 %) (Table 7-3).

7.4 Discussion Reported selectivity parameters forC. crangon are rare. Selection factors obtained by Bohl and Koma (1962) for polyamide cod-ends lie in the range of 2.0 to 2.8. Recent experiments carried out by Graham (1997) with cod-ends similar to those used in this study gave a selection factor of 1.6. The selection factor obtained in the present study (1.82) lay in between the results of these two authors. There is no legal minimum landing size for Brown Shrimps in the EU. There is however a minimum market size. According to EU-regulation 2406 shrimps should be graded into marketable and non-marketable shrimps on a sieve with 6.5 mm bar spacing. The minimum carapace width of 6.5 mm compares to a total length of 45 mm. This length lies very close to the mean L75 (45.2 mm TL) of the 22 mm shrimp cod-end and consequently quite a lot of non-marketable shrimps can be caught by this type of cod-end. This result was also found by Revill et al. (2000) indicating that high numbers of small shrimps are caught by a 20mm cod- end. According to EU regulation 850/98, the legal minimum mesh size for Brown Shrimps has been decreased to 16mm. The same study showed that with this mesh size even more Discussion 7-85 small shrimps are caught but that somewhat more commercially sized shrimps occur in the catches. This points at a rather wide selection range, which was also found in the present study. Revill concluded in his study that a 26mm cod-end could be an effective management tool to reduce the levels of discarding of C. crangon with no apparent detriment to the landings. The results on whole trawl selectivity in the present study confirm the conclusion reported by Bohl in 1963 that some shrimps escape through the net meshes. Revill et al. (2000) carried out experiments with larger meshes in the net body in front of the cod-end. The catches of undersized shrimps were indeed reduced by the better selection in the body of the net. The results of the present study not only demonstrate the existence of these escapees but also stress the relatively higher importance of net selectivity compared to cod-end selectivity. Graham (1997) reported an L50 of 4.6 cm and a SR of 0.6 cm for Dab in a comparable shrimp trawl, which is very close to the results of the present study. The selection factor for Dab of 2.1 found in this study is identical to the summary selection factor for Dab in beam trawls reported in a review by Wileman for the European Commission (Wileman, 1992). The selection factor for Sole, however, was 3.6, which is higher than the 3.2 mentioned in this review. The high variability observed in the L50 values of the cod-end could for a large part be explained by the clogging of the meshes by hydroids, the catch volume and the state of the sea. Logically, L50 was negatively correlated with clogging. The more hydroid threads that stick to the meshes, the smaller the opening will become for the animals to escape, thus reducing selectivity. Also total catch volume was negatively correlated with the L50. Higher catch volumes are often due to higher amounts of benthic animals in the catch, like crabs and starfish. Shrimps that end up between high amounts of benthic animals in the cod-end do not always have the opportunity or the strength to find their way towards the open meshes. In a clean, low volume catch the encounters with the cod-end meshes will be much higher, thus increasing escape opportunities and selectivity. A consequence of this could be that in trawls with selective devices, like sorting grids or sieve nets that have cleaner catches of less volume, cod-end selectivity for shrimps will be higher. A similar effect of catch volume on the selectivity of Brown Shrimps was found by Bohl and Koura (1962). The trials were not carried out on the traditional commercial fishing grounds and catch volume was relatively low compared to normal commercial catches. As a consequence, the selectivity of the shrimp trawl used in commercial conditions will probably be somewhat lower compared to the result obtained in this study. Sea state was positively correlated with the L50. Vessel motion, which depends on sea state, creates a pumping movement of the net, during trawling or during the hauling operation, when the trawl is heaving up and down alongside the vessel. These movements of the net may provoke the opening and closing of the meshes, which can induce a higher escapement of shrimps from the cod-end. A similar correlation between the sea state and cod-end selectivity was described for Nephrops norvegicus by Polet and Redant (1994) on a commercial vessel. For the whole trawl selectivity study it was assumed that the total number of shrimps in the trawl path was known. The reliability of this number, however, depends upon the catching efficiency of the collecting bags underneath the ground rope. These were constructed to have a very good bottom contact but it was not possible to estimate their catching efficiency. Therefore, the estimated number of shrimps escaping underneath the ground rope and consequently the total number of shrimps encountering the gear could be slightly underestimated. Selectivity o f the shrimp beam trawl 7-86 The selectivity experiments taught us that a significant increase of the mesh size to improve the selectivity for non-target animals is not an option because of high expected losses to the target species. A 26mm cod-end would not have a detrimental effect on the commercial shrimp catch, but would not be sufficient to let fish escape. Another option would be to insert large mesh escape windows in the net. Escape windows are based on behavioural differences and should be positioned in a part of the net where the target species does not attempt to escape. For the shrimp beam trawl, this is the front part of the top panel. Since flatfish are the main issue in the discard puzzle, a window should be positioned in a way to let especially flatfish escape, i.e. the lower parts of the net. The whole trawl selectivity study, though, showed that shrimps mainly escape through the meshes of the lower parts of the net. Inserting a large mesh window, would be an escape route for commercially sized shrimps especially since such a window would have a stronger water flow through its meshes, taking relatively more shrimps with it. Any selective device that would increase the water flow through it, will lead to commercial shrimp losses. The insertion of a filter in the net has better potential. The filter should be positioned in a way that the whole catch is forced to encounter the filter. Animals small enough, like shrimps, should pass through the filter and be caught in the cod-end. Larger animals should be guided to an outlet. A sorting grid and a sieve net should be appropriate for this purpose (see Chapters 8 and 9). Another option would be an alteration to the groundrope, especially if an alternative stimulation could be found with a high fishing power for shrimps and high selectivity for other species. Electric fishing may be good tool to achieve this goal (see Chapter 10).

7.5 Conclusions For the sake of having a good description of the selective properties of the shrimp beam trawl and as a preparation of the experiments with selectivity improving devices, a detailed study was made of the selectivity of the shrimp trawl cod-end, the net and the groundrope. An attempt was also made to explain the variability of the selectivity by covariates like haul duration, speed through the water, state of the sea, total catch volume, clogging of the meshes, water temperature and light conditions. The cod-end selectivity parameters found for Brown Shrimp were: L50 = 39.4 mm (37.0 - 41.4), the selection factor = 1.82 (1.71 - 1.91) and the selection range = 11.6 mm (10.2 - 13.0). The cod-end selectivity of the beam trawl for shrimps was found to be very variable. Clogging of the meshes, catch volume and the state of the sea contributed significantly to this variability. On average, the selectivity was rather poor for shrimp allowing relatively high quantities of non-marketable shrimps to be retained. The selectivity of the net body, however, was quite important and allowed more shrimps to escape than the cod-end. It was mainly the rounded lateral part of the net belly that contributed to this selectivity. Due to the small mesh size, cod-end selectivity for fish was very poor. Only a small part of even the 0-age group fish could escape through the meshes. In addition, the other parts of the net did almost not allow any fish to escape. Consequently, the shrimp beam trawl will inevitably catch high amounts of small fish in coastal areas and estuaries, where juveniles prevail. An increase in mesh size to improve this situation would shift the selection curve for shrimp to the right, what means that more marketable shrimps would be lost. Escape windows are not an option since these will most likely lead to commercial shrimp loss. Selective devices, like sorting grids or sieve nets, to separate the shrimps from the fish and/or benthos by-catch are therefore probably a better way to improve species selectivity of shrimp beam trawls. Also an alteration to the groundrope may show potential, especially if an alternative Conclusions 7-87 stimulation could be found with a high fishing power for shrimps and high selectivity for other species. The design of the fishing gear studied was comparable to that of most shrimp beam trawls used in the North Sea (van Marlen et al., 1997b) and hence the results in this report have a wider applicability. Evaluation o f the Nordmore sorting grid as a selectivity improving device 8-88

8 Evaluation of the Nordmore sorting grid as a selectivity improving device

8.1 Introduction The first large scale research into sorting grids concerned sea turtle by-catch. In the late 1970s, the incidental capture and mortality of sea turtles in shrimp trawling gear in the south­ eastern U.S.A. was determined to be a major threat to the survival of some endangered sea turtle stocks. Research was initiated to develop solutions to the problem and the TED (turtle excluder device) seemed to be the answer (Watson and Seidel, 1980). The TED is a simple soft or rigid grid device, which deflects the turtle to an opening in the trawl, while allowing shrimp to pass through it and end up in the cod-end. The TED has been further developed to reduce discarding of fish and invertebrates in tropical shrimp fisheries (NOAA, 1995; Foster and Watson, 2003). Today, many countries’ legislation contains the compulsory use of TEDs, especially driven by the U.S. embargo on import of shrimp caught in a way that may adversely affect sea turtles (U.S. Public Law 101-162, Sec. 609). In the late 1980s, a great deal of research was initiated on the application of sorting grids in the northern Pandalus fishery for the reduction of the by-catches in this small mesh fishery (Anon., 1998). Especially in Canada and Norway, different types of sorting grids were tested to improve species and size selectivity in that fishery. The Nordmore grid proved to be quite effective in separating shrimps from finfish although some part of the commercial shrimp catch was often lost (Canada e.g. Albert, 1992; Boudreau, 1998 - Norway: Larsen, 1991; Isaksen and Valdemarsen, 1994). Experiments were also carried out with a secondary Nordmore grid, rigged behind the primary grid, to improve size selection of the shrimps (Brothers, 1995). The next step was to apply the sorting grid to improve selectivity for finfish, e.g. to improve the size selectivity for Cod (Larsen and Isaksen, 1993) or to improve species selectivity, e.g. the reduction of Cod and Haddock by-catches in the silver hake (Merluccius bilinearis) fishery (Cooper, 1991). The success of the Nordmore grid is reflected in its mandatory use in several shrimp fisheries, e.g. in the Gulf of St. Lawrence (Canada) in 1993 and in all Canadian northern shrimp fisheries since 1997 (Brothers, 1998). The Nordmore grid, being the first developed and simplest grid, was soon followed by other more complex grids like the Sort-X and Sort-V grids giving a solution to specific problems and exploiting typical behaviour of species (Larsen 1992). Madsen and Hansen (2001) carried out experiments with grids in the North Sea Pandalus fishery with good results. In the North Sea Brown Shrimp fishery, however, the use of sorting grids was still in the experimental phase at the time of this study, with tests carried out in the United Kingdom (Graham, 1997) and Germany (Wienbeck, 1997). Because of the simple design of the Nordmore grid and its success in the northern shrimp fishery, this type of grid was chosen for further evaluation in the study presented in this report. The experiments consisted of three phases: a first phase in which a grid design similar to those tested by Graham and Wienbeck was studied, a second phase with alternative designs and a third phase, in the frame of the DISCRAN project, in which one selected new design was explored. The EU-funded project DISCRAN was performed with the co­ operation of the following institutes: • Bundesforschungsanstalt für Fischerei, Hamburg, Germany, • Departement voor Zeevisserij, Belgium, • Rijksinstituut voor Visserij Onderzoek, IJmuiden, Netherlands, • University of Newcastle, Newcastle, UK. Materials and methods 8-89 All work presented in this Section was carried out by the author and the staff of DvZ.

8.2 Materials and methods

8.2.1 Fishing area, vessels and trawls • The trials for Phase 1 were carried out during two cruises aboard the A.962, RV “Belgica” (length = 50.9 m, GRT = 765 t, engine power = 1154 kW) in November 1996 and April 1997 and three omises on the commercial shrimp trawler Z.582, “Asanai” (length = 19.8 m, GT = 49 t, engine power = 107 kW) in July and September 1997. The first trip on the RV “Belgica” was used for preliminary experiments to optimise the rigging of the grid whereas the data collected in the other trips were used to estimate the selectivity of the sorting grid. • During Phase 2, three omises were carried out aboard the RV “Belgica” in November 1997 and in April and November 1998, mainly to explore new grid designs. • Between July 1999 and January 2000 (Phase 3), thirteen sampling trips were undertaken on two commercial vessels: the commercial shrimp trawlers 0.700 “Bisiti” and 0.101 “Benny” (length = 16.8 m, GT = 33 t, engine power = 177 kW). One sampling trip was undertaken aboard the research vessel Belgica. The purpose of these trips was to test and optimise one selected new grid design. The A.962 “Belgica” is a governmental oceanographic research vessel, fishing from the stem. Each trip, a fishery skipper was hired to select fish tracks and guide the experiments to have practices as close as possible to commercial conditions. It has to be noted, however, that due to safety regulations and restricted access of the research vessel to shallow fishing grounds, the traditional shrimp fishing grounds could not be visited and the experiments were mainly carried out in the area “Vlakte van de Raan” (flat, hard sandy grounds) giving a rather low number of species in the catch. The commercial vessels were specifically fitted for catching shrimps (rotating shrimp riddle, cooker etc.) and had a three-man crew. In contrast to the research vessel, the commercial boats towed two beam trawls simultaneously, one at each side of the vessel. For the sea trials, the choice of fishing grounds and fishing practices was left to the skipper to ensure commercial conditions. The traditional fishing grounds on the Flemish Banks off the Belgian coast were fished (ICES sub-division IVc). In the area fished, a wide variety of sediments is found (Bastin, 1974). Van Lancker (1999) clearly demonstrated that this is an area with typical long stretched sand banks having irregular relief with varying slopes and continuously dredged navigation channels. A wide range of sand dunes, from small to large, occurs in this area. This large variation in sediments and topography affects the behaviour of the fishing gear. The towing speed was between 2 and 3 knots on the research as well as on the commercial vessels. The variability in towing speed was somewhat higher on the commercial vessels due to the influence of tide but the differences were quite small. The warp length was three times the water depth. The same trawl was used as in the selectivity experiments (see Section 7). Details about the trawl are given in Section 7.2.1. On the commercial vessels as well as on the research vessel, the skipper kept a record of abiotic and biotic data: • name of the fishing ground • starting time of the haul • ending time of the haul • geographical position at the beginning of a haul Evaluation o f the Nordmore sorting grid as a selectivity improving device 8-90 • geographical position at the end of a haul • wind direction • wind speed in Beaufort scale • direction of the current • state of the sea • precipitation or no precipitation • light intensity - bright/dusk/dark • water depth • length of warp paid out • towing direction • towing speed in knots Aboard the RV Belgica, only one gear could be towed. Therefore, the cod-end of the experimental net was covered or blinded with 11 mm mesh netting and the grid outlet was covered with an 11 mm outlet cover to retain the escapees from the grid outlet. For catch comparison aboard the commercial vessels, the experimental net was towed from the starboard side of the vessel while the standard net was towed from the port side. At regular time intervals the skippers were asked to check during the commercial fishery whether the two sides of the vessel had the same catching efficiency, to avoid bias. The catching efficiency of both sides was quite similar. Commercial mesh sizes were used. Unlike the shrimp fishery elsewhere in the North Sea, the Belgian shrimpers catch marketable sized fish and are allowed to land these fish. Financially the Belgian shrimp fleet depends strongly on the income from fish landings. The use of a sorting grid or a sieve net that directs the larger elements of the catch towards an outlet would lead to a substantial loss of commercial catch. Therefore, the trials aboard commercial vessels were always carried out with a grid outlet cover. This cover had a nominal mesh size of 80 mm, which is the legal minimum mesh size for the Sole fishery. The cover was attached around the outlet at a sufficient distance from the outlet to avoid distortion, although problems were not expected since catch volumes in the cover would usually be quite low because of the large mesh size. The sorting grid used for Phase 1 of the study was 60 cm wide and 80 cm long with bars parallel to the 80 cm side and parallel to the longitudinal axis of the net (Fig. 8-1A). The bars had a diameter of 6 mm and were made of polyethylene. The grid-frame consisted of plastic tubes with a diameter of 30 mm during the tests on the research vessel but were replaced by stainless steel tubes with a diameter of 18 mm for the trials on the commercial vessel. The latter grid was rigged with two 1 1 floats to prevent scraping the bottom. The bar spacing was 12 mm during the experiments in November 1996 and was increased to 14 mm for the remainder. The grid was braided to the net and rigged at an angle of 45° sloping backwards, with an outlet along the top of the grid with a width of 15 cm (Fig. 8-1B and C). To overcome the problems encountered during Phase 1 with the grid, three new grid designs (Fig. 8-ID, E and F) were tested during Phase 2: • A grid, identical to the Phase 1 grid, rigged in a metal frame for stability during fishing, • a combination of a primary wide-spaced grid (bar spacing 25 mm, angle 35°, 60 cm wide and 100 cm long) and a secondary grid, identical to the Phase 1 grid, and • a grid, similar to the Phase 1 grid, sloping downwards and with the outlet below. The design of the sorting grid used during Phase 3 (Fig. 8-1G and H) was based on the third option tested in Phase 2. It was 60 cm wide and 90 cm long with bars parallel to the 90 cm side and parallel to the longitudinal axis of the net. A stainless steel grid was used because of its rigidity. The grid consisted of a hollow tube frame (18 mm diameter) with solid bars of 6 Materials and methods 8-91 mm diameter and 20 mm bar spacing. Five floats (in a later trial three) of 1 1 were attached to the top bar of the grid to prevent scraping on the bottom. The grid was fitted in the net at an angle of approximately 47°. The angle was in that way oriented so that the top comes in front of the underside. The part of the catch that did not pass through the grid was directed to the bottom of the net and the escape outlet. As the sorting grid was not used commercially at the time of the trials and no standard design existed, the design of the grid was changed throughout the sea trips to optimise it. Several solutions were sought to reduce clogging and the loss of coimnercial shrimps.

The selective sorting arid Phase 1 - Rigging of grid and covers on RV Belgica Phase 1 - Rigging of grid end cover on Z.582

small mesh outlet cover

small mesh cod-end cover

Phase 2 - Alternative design with metal frame Phase 2 - Alternative design with two grids Phase 2 - Alternative design with outlet below

guiding panel matai frame with grid secondary grid primary grid

Phase 3 - Rigging of and and cover on RV Belgica Phase 3 - Rigging of grid and covers on commercial vessel 7 3 11 m m co d -en d blinder _

guiding panel 11 mm outlet cover.

Fig. 8-1 - Overview of the selective sorting grids studied.

8.2.2 The cod-end and the outlet covers During Phase 1, the cod-end selectivity of a shrimp beam trawl rigged with a grid was studied. This was done by applying the covered cod-end technique (Wileman et al., 1996). The cover was constructed of knotted polyamide netting with a nominal mesh size of 11 mm, was 800 meshes on the circumference and had a sufficient length to leave an open space of 1.5 m behind the aft end of the cod-end. The cover was held open by a half hoop over the top panel of the cod-end and had a diameter of 1.5 m (Fig. 8-1B). The application of full hoops was not feasible due to frequent damage caused by the close bottom contact of beam trawls. To study the selectivity of the grid on the research vessel, the outlet was covered by a small mesh cod-end (called the outlet cover) made of the same netting as the cod-end cover. To prevent obstmction of the outlet, the outlet cover was rigged with four 11 floats (Fig. 8-1B). On the coimnercial vessel, the outlet was covered by a large mesh (80 nun) cod-end to catch the marketable fish (Fig. 8-1C). A small mesh cover was not used because catch comparison was possible between the standard gear on one side and the experimental gear on the other side of the vessel. The absence of covers also better reflects real coimnercial conditions. To study grid selectivity on the research vessel during Phases 2 and 3, a small mesh blinder was inserted in the cod-end (Fig. 8-1G). The covered cod-end technique and the use of a blinder may have disadvantages, like a reduction of water flow, masking of the cod-end meshes and influence on the behaviour of the fish (Wileman et al., 1996). On the research vessel this Evaluation o f the Nordmore sorting grid as a selectivity improving device 8-92 method was, however, the best option although escapement of animals through the cod-end meshes and the grid outlet may have been underestimated. Due to the small size of the meshes, the ICES mesh gauge could not be used to measure the mesh opening. A calliper was used instead, stretching the meshes with a weight of 2 kg. The weight of 2 kg was chosen according to the rules set up in EU-regulation 2108/84 for measuring meshes with a size below 35mm. The mesh openings of the large mesh outlet cover were measured with the ICES mesh gauge set at 4 kg pretension. Throughout the experiments, the mesh openings of the cod-ends and covers used were regularly measured and showed very little variation. The average cod-end mesh opening varied between 21.3 and 21.7 mm with standard deviation varying between 0.58 and 0.81 mm. The average mesh openings of the covers and blinders varied between 10.5 and 10.7 mm with standard deviations between 0.54 and 0.79 mm. The average mesh opening of the large mesh outlet cover varied between 77.5 and 77.8 mm with a standard deviation between 2.16 and 2.24 mm.

8.2.3 Data collection and analysis The species investigated were Brown Shrimp, Plaice, Sole, Dab, Flounder, Whiting, Bib (and Poor Cod) and Cod. Only those species present in sufficient numbers in the catch were included in the data analysis. Besides these commercial species, non-commercial fish and invertebrates were also studied. Detailed descriptions of the catch handling and data analysis are given in Section 5.2.3 and 7.2.3. To estimate the amount of non-commercial fish and invertebrates, each haul, a 6 to 10-1 sample was taken from the main by-catch fraction of the catch, after the fish were sorted out. In the laboratory the animals were weighed and counted. This was also done for the non­ commercial fish and invertebrates in the Brown Shrimp samples. All these data were then recombined and raised to estimate the composition of the total catch. The selectivity of the grid was expressed in terms of animals escaping through the grid outlet. In the case of the research vessel trials, this was calculated as the ratio of the numbers in the outlet cover to the total numbers in the cod-end, the cod-end cover and the outlet cover. In the case of the commercial vessel, this was calculated as the ratio of the difference in numbers in the cod-ends of the standard and the experimental gear, to the numbers in the cod-end of the standard gear. Where relevant, the percentages of lost animals due to the grid were compared between grid designs. For this purpose, a t-test for independent samples was used and backed up with the non-parametric Mann-Whitney U-test if samples were small and normality of the data was not convincing. If possible, a selection ogive was fitted to the length selectivity data.

8.3 Results

8.3.1 Narrative of the sea trials A log of the sea trials is given in Table 8-1. Results 8-93 Table 8-1: Log of the experimental hauls with the selective sorting grid

Phase Vessel Date & code Number Rigging Action of hauls

1 RV Belgica November 1996; 96/25 5 grid - outlet on top optimise rigging 1 RV Belgica April 1997; 97/08 24 grid - outlet on top cod-end & grid selectivity - cover method 1 Z.582 August and October 10 grid - outlet on top grid selectivity - catch 1997; 97/A, B & C comparison 2 RV Belgica December 1997; 97/28 8 grid in metal frame evaluate new design 2 RV Belgica February 1998; 98/04 13 two grids evaluate new design 2 RV Belgica December 1998; 98/27 8 grid - outlet below evaluate new design 3 0.700 & July 1999 - January 45 grid - outlet below grid selectivity - catch 0.101 2000; 99/A-J, 00/A & B comparison 3 RV Belgica December 1999; 99/26 4 grid - outlet below grid selectivity - cover method

P hase 1 During the first cruise with RV Belgica (96/25), five hauls were carried out to optimise the rigging and the handling of the grid and the covers. A preliminary estimate was made of the loss of marketable shrimps. During the second trip with RV Belgica (97/08), 24 hauls were carried out with a trawl rigged with the sorting grid, a cod-end cover and an outlet cover. Shrimps with lengths within the SR were caught in sufficient numbers in each haul to calculate cod-end selectivity. The fish in the catch were, however, not small enough to collect sufficient numbers of escapees in the cod-end cover to calculate any cod-end selectivity ogive. For the calculation of the grid selectivity ogives for the five fish species caught, all hauls were pooled by species since numbers were too small to determine the ogives at the haul level. Consequently no confidence limits are given. Compared to coimnercial conditions, the catches were rather clean, i.e. with low amounts of benthos and debris. Ten hauls were carried out with the standard gear at one side of the vessel and the experimental net with a sorting grid on the other side during the three trips on the coimnercial vessel (97/A, B & C). Only five hauls were used for selectivity calculations, since clogging of the grid occurred in the five other hauls. For the calculation of the grid selectivity ogives for the four fish species analysed, again all hauls were pooled. The volumes of benthos and debris in the catches were large compared to the research vessel trips. The towing direction was kept constant during each haul. Due to obstacles on the seafloor and the nature of the fishing ground, the maximum length of the tracks differed and haul duration could not be kept constant. P hase 2 Three trips were carried out ou RV Belgica (97/28, 98/04 & 98/27) to evaluate three new grid designs (Fig. 8-ID, E and F, Table 8-1). An estimate was made of the loss of marketable shrimps. P hase 3 Thirteen sampling trips were undertaken on coimnercial vessels (99/A-J, 00/A & B) and 1 sampling trip was undertaken on board of RV Belgica (99/26) (Table 8-1) between July 1999 Evaluation o f the Nordmore sorting grid as a selectivity improving device 8-94 and January 2000. A total of 49 hauls was carried out with the sorting grid of which 35 were analysed. For the other 14 hauls, only the volumes of the different catch fractions were recorded. The purpose of these trips was to test and optimise the third alternative grid design tested in Phase 2, which at the moment gave promising results and seemed to be less susceptible to clogging with starfish. An overview of the different grid configurations tried out is given in Table 8-2. Rather high commercial shrimp losses were again recorded during the first four sea trips. Therefore, a guiding panel was inserted in front of the grid to avoid part of the catch going through the outlet before coming into contact with the grid. This, however, caused catch to accumulate in front of the grid in some hauls during trip 99/E. At the request of the skipper, the guiding panel was removed, but the results did not improve. On the seventh trip, the guiding panel was inserted again and the outlet was enlarged. This caused a higher shrimp loss. During the next trips, many alterations were made to the outlet to find a compromise between shrimp loss and release of debris and large items in the catch like jellyfish. From trip 99/J onwards, the design was kept constant for the rest of the trials. The methodology to determine grid selectivity was identical to the methodology used in Phase 1. In each haul, sufficient numbers of shrimps were caught for data analysis. The numbers of fish in the catch, though, varied quite strongly, depending on the fish species, time of day, season and fishing ground. To maximise the information retrieved from the experiments, datasets with low numbers of fish were also included in the analysis which sometimes led to a poor fit of selectivity ogives to the retention points (see Section 8.3.3 - Phase 3). As in Phase 1, for the calculation of the grid selectivity ogives for fish, all hauls were pooled by species since numbers were too small or retention points too irregular to determine the ogives at the haul level.

Table 8-2: Overview of the different net configurations with the sorting grid during Phase 3 of the project.

Date Vessel Guiding panel Shielding device in outlet Outlet

30/07/1999-99/A commercial no no bottom 11/08/1999-99/B commercial no no bottom 3 & 4/09/1999-99/C& D commercial no no bottom 8/09/1999-99/E commercial yes no bottom 9/09/1999 - 99/F commercial no no bottom 15/09/1999 - 99/G commercial yes no bottom 13/10/1999 - 99/H commercial yes yes bottom 28/10/1999-99/1 commercial yes yes bottom 3/11/1999 -9 9 /J commercial yes yes bottom 25/11/1999-99/26 RV Belgica yes yes bottom 10/01/2000-00/A commercial yes yes bottom 24/01/2000 - 00/B commercial yes yes bottom

8.3.2 Cod-end selectivity The cod-end selectivity was only studied during Phase 1 of this study. The main objective was to find out whether the lower catch volumes due to grid selection would influence selectivity. The L50 for Brown Shrimp for all hauls, on the second research vessel trip, combined was 44.1 mm (42.2 - 46.0 mm) and the selection factor 2.03 (1.94 - 2.12). The selection range was 12.7 mm (10.9 - 14.5 mm). The selection ogive is shown in bold light Results 8-95 blue in Fig. 8-2 together with the length distribution for all Brown Shrimps that entered the cod-end.

100 35000

30000

25000

20000 50 15000

10000 25 5000

20 40 50 80 Length class (mm)

Fig. 8-2 - The estimated overall selectivity ogive (bold light blue line) together with the 24 selection ogives (dotted lines) of the single hauls and the length frequency distribution of Crangon crangon of the cod-end + cod-end cover catch (bold deep blue line); (RV Belgica trip 97/08).

It has to be noted that the catch volumes during these trials were very low. Catch volumes comparable to coimnercial conditions, would probably have produced selectivity parameters closer to the selectivity of the standard trawl. The selection ogives for the individual hauls are also shown in Fig. 8-2 and the L50 and SR values in Table 8-3. Most of the L50 values lay within the 40-50 ïmn range. The SR values were rather variable ranging from 8.5 to 29 ïmn. For fish, no cod-end selectivity could be calculated. Apparently, the cod-end meshes were too small to allow any selection to take place for fish. Evaluation o f the Nordmore sorting grid as a selectivity improving device 8-96 Table 8-3: Cod-end selection parameters and 95 % confidence limits (between brackets) for Brown Shrimp (97/08 = RV Belgica trip; 97A, B & C = commercial vessel trip) Haul number Date Haul start Haul duration Cod-end L50 Cod-end SR time (h) (mm) (mm) 97/08 - 01 14/04/1997 10:45 1:15 45.9(45.1 -46.8) 8.5 (7.1-9.9) 97/08 - 02 14/04/1997 12:20 1:10 43.4 (42.7-44.1) 9.2 (7.7-10.8) 97/08 - 03 14/04/1997 13:50 1:05 52.4 (50.9- 53.8) 9.3 (7.5-11.1) 97/08 - 04 14/04/1997 16:00 1:30 35.8(32.2- 39.4) 20.7(15.1 -26.2) 97/08 - 05 14/04/1997 17:45 1:35 39.9(37.9-42.0) 12.7 (9.8-15.6) 97/08 - 06 14/04/1997 19:45 1:10 39.4 (35.9-42.9) 25.2(16.8- 33.7) 97/08 - 07 14/04/1997 21:25 1:00 34.2 (32.2 - 36.2) 11.7 (9.6-13.9) 97/08 - 08 15/04/1997 6:30 1:30 41.0 (39.1 -43.0) 19.7(14.8-24.6) 97/08 - 09 15/04/1997 8:15 1:40 43.6 (42.7-44.5) 9.5 (7.6-11.4) 97/08 - 10 15/04/1997 10:10 1:10 56.6 (52.4-60.8) 25.1 (15.2- 35.1) 97/08- 11 15/04/1997 15:00 1:30 44.4 (43.6-45.2) 9.4 (7.9-10.9) 97/08 - 12 15/04/1997 19:40 1:30 41.6 (40.3 -42.9) 12.7(10.1 - 15.2) 97/08- 13 15/04/1997 21:30 1:00 35.2 (30.6- 39.7) 20.5 (13.5 -27.6) 97/08 - 14 15/04/1997 23:00 1:30 42.5 (41.1 -44.0) 15.1 (11.6-18.5) 97/08 - 15 16/04/1997 7:30 1:30 40.3 (38.8-41.9) 15.0 (11.9-18) 97/08 - 16 16/04/1997 11:40 1:35 42.4 (41.5 -43.2) 9.0 (7.7-10.2) 97/08 - 17 16/04/1997 13:35 1:25 47.9(46.9-48.8) 12.4(10.6-14.2) 97/08 - 18 16/04/1997 15:20 1:25 42.4 (3.70- 81.1) 8.9 (-8.8-26.6) 97/08- 19 16/04/1997 17:10 1:25 43.7(42.5 -44.8) 15.6(12.4-18.7) 97/08 - 20 16/04/1997 18:55 1:30 47.1 (46.1 -48.1) 10.0 (8.5-11.6) 97/08-21 16/04/1997 20:45 1:25 57.0 (53.6-60.3) 29.1 (20.6- 37.7) 97/08 - 22 17/04/1997 7:30 1:30 49.4 (48.6- 50.1) 8.8 (7.6-10.1) 97/08 - 24 17/04/1997 12:30 1:30 47.1 (46.1 -48.2) 12.3(10.2-14.5) 97/08 - 25 17/04/1997 14:20 1:20 41.8(40.0-43.5) 18.2 (14.4-22) 97A - 01 8/07/1997 19:35 1:25 (-) (-) 91A - 02 8/07/1997 23:05 1:20 (-) (-) 97B - 03 10/07/1997 19:25 1:10 (-) (-) 97B - 04 10/07/1997 22:50 1:20 (-) (-) 97B - 05 10/07/1997 2:00 1:15 (-) (-) 97C - 06 10/09/1997 16:45 2:00 (-) (-) 97C - 07 10/09/1997 19:10 1:32 (-) (-) 97C - 08 10/09/1997 23:40 1:05 (-) (-) 97C - 09 10/09/1997 1:10 1:15 (-) (-) 97C - 10 10/09/1997 3:05 1:20 (-) (-)

8.3.3 Sorting grid selectivity Phase 1 During the first research vessel trip (96/25), the bar spacing of 12 mm proved to be too small. Almost 50 % of the Brown Shrimps escaped through the outlet. It was decided therefore to increase the bar spacing to 14 mm for the next trip. The bar spacing increase improved the grid selectivity for Brown Shrimp during the second research vessel trip (97/08). The percentage lost decreased to 14% (10-18). The reduction in catch of under-sized Brown Shrimps (<45mm) was somewhat higher (17% (10-24)) compared to the reduction of commercially sized shrimps (13% (10-16)), though a paired t- test did not show a significant difference (p > 0.05). In Fig. 8-3, the average loss of Brown Results 8-97 Shrimps, over all hauls, is given in relation to the body length with the 95% confidence limits as error bars.

100 O3 <11 .c .c U3 i o .c cU i 'oO) <11 U i Bc dl o ai o.

20 30 40 50 60 70 80 Length class (mm TL)

Fig. 8-3 - Percentage of Crangon crangon lost, with the 95% confidence limits indicated as error bars (RV Belgica trip 97/08).

The data points give the percentage of animals that did not pass through the grid spacings and escaped through the outlet in front of the grid. There was clear length dependence in the relative amounts of Brown Shrimps escaping. In the mid-range of the length classes (35 - 65 ïmn) about 15% of the animals escaped. Above 65 ïmn, the losses increased to almost 40 %. Below 35 ïmn the losses were even higher and increased to over 90 %. The overall catch reduction in numbers of coimnercial fish species due to the grid was 75% (70-80). The grid selection ogives and the length frequency distributions (total number in the catch) for Dab, Plaice, Sole, Whiting and Cod caught during the second research vessel trip are shown in Fig. 8-4A to E. The selection ogive gives the percentage of fish that do not go through the grid and escape through the outlet. The minimum landing size (MLS) is given as a dashed vertical line, except for Dab for which there is no MLS. The L50 and SR values are given in Table 8-4.

Table 8-4: Grid selection parameters for fish (RV Belgica 97/08; commercial vessel, 97/A, B & C) Species Grid L50 (nun) Grid SR (nun) Grid L50 (nun) Grid SR (nun) R V Belgica Z. 582 - commercial vessel Dab 8.6 11.5 -9.4 24.0 Plaice 9.4 9.7 9.5 7.3 Sole 9.9 5.2 N.A. NA. Whiting -0.9 17.3 6.7 10.5 Cod 14.2 15.2 N.A. NA. Evaluation o f the Nordmore sorting grid as a selectivity improving device 8-98

A: Dab B: Plaice 100 250 100 120 200 100 '5 S 75 o £ 75 => = 80 ¡2 d) O 150 «D d) O

Length class (cm TL) Length class (cm TL)

C: Sole D: Whiting 100 400 100 -, 300

Ö S 75 300 ui'S 200 ¡2 d) o S’ £ 50 200 100 z 100

Length class (cm TL) Length class (cm TL)

E: Cod 100

Öo> £= 75 dJ o 30 a> S’s. 50 20 §

Length class (cm TL)

Fig. 8-4 - Grid selection ogive and length frequency distribution for Dab, Plaice, Sole, Whiting and Cod (RV Belgica trip 97/08).

During the trips on the commercial vessel (97/A, B & C), five hauls had to be discarded due to clogging of the grid with starfish and/or debris, preventing nonnal passage of the Brown Shrimps. This resulted in a loss of the coimnercial catch ranging between 39 and 90 %. Over the ten hauls, the average loss of shrimps <45mm, >=45mm and all length classes was 67% (49-86), 40% (16-64) and 55% (35-75) respectively. The variability in the performance of the grid was high. For the five successful hauls the reduction in the coimnercial Brown Shrimp catch was between -9% and 26 % with an average of 12% (-5-29). The average loss of undersized shrimps was 46% (24-69). In Fig. 8-5 the average loss of Brown Shrimps, over the five hauls, is given in relation to the body length with the 95% confidence limits as error bars. Again, there is clear length dependence in the relative amounts of Brown Shrimps lost, although the error bars indicate a high variability. For lengths > 48 ïmn, between 10 - 20 % of the animals escaped. For lengths < 48 ïmn, losses increased with decreasing length with a reduction around 90% for the smallest animals. Results 8-99

100

40 Length class (mm)

Fig. 8-5 - Percentage of Crangon crangon lost, with the 95% confidence limits indicated as error bars (commercial vessel, 97/A, B & C).

For the successful hauls, the overall catch reduction in numbers of coimnercial fish species due to the grid was 74% (61-87). In Fig. 8-6A to D, the grid selection ogives and the length distributions (total number in the catch) in the standard (solid line) and experimental (dashed line) gears for Dab, Plaice, Sole and Whiting are given. The L50 and SR values are given in Table 8-4. It was not possible to calculate a selection ogive for Sole because of the high reduction in catch of the smallest fish.

A: Dab B: Plaice 100 250 100 500 200 400 tn 150 5 3 0 0 al .= 100 zi 200 = 50 100

10 20 Length class (cm TL) Length class (cm)

C: Sole D: W hiting 100 150 100 -, 1000

750 100 0)2 E 500 50 i 250

30 Length class (cm) Length class (cm)

Fig. 8-6 - Grid selection ogive (fine solid line) and length frequency distributions in the standard (bold solid line) and experimental (bold dashed line) gear for Dab, Plaice, Sole and Whiting during the trips with Z.582 (commercial vessel, 97/A, B & C). Evaluation o f the Nordmore sorting grid as a selectivity improving device 8-100 Overall, data for twelve non-commercial fish and invertebrate species were analysed. The average catch reduction was 61 % (43-79). For most of the species, the sorting grid was very effective in releasing the animals and the reduction ranged from about 5 % for Liparis liparis to almost 100 % for Ophiura spp. (Fig. 8-7).

Sepiola atlantica Pomatoschistus spp. Pagurus bernhardus Ophiura spp. Myoxocephalus scorpius Liparis liparis Liocarcinus holsatus Ensis directus Ciliata mustela Asterias rubens Anemones Agonus cataphractus 25% 50% 75% 100% Percentage going through outlet

Fig. 8-7 - The percentage reduction of the different benthic species in the catch of the experimental net (commercial vessel, 97/A,B & C).

Phase 2 Since the first grid design led to high losses of coimnercial shrimps, three alternative designs were tried out. An obvious cause of the catch losses could be clogging of the grid, as observed during the coimnercial vessel trips 97/A, B & C. Another possibility could be instability of the grid angle during the fishing operation. In the first new design (Fig. 8-ID), the existing grid was rigged into a metal frame, aiming at a stable grid angle. In the second design, a primary grid with a wider bar spacing and a lower angle was inserted to divert the larger elements of the catch (that often caused the clogging) to the outlet. It was anticipated that the wider bar spacing would still allow the shrimps to pass through the grid in spite of the low angle. The secondary grid was inserted to select out the smaller elements of the unwanted by-catch, e.g. juvenile fish. In the third design, the grid slope was downwards, leading to an outlet in the belly of the net. It was anticipated that clogging was less likely to occur with a downward slope. The losses of coimnercial shrimps with the grid in a metal frame were comparable to the losses observed during Phase 1 (Table 8-5). A t-test did not indicate a significant difference. Clogging of the grid still occurred and the metal frame caused practical problems while shooting or hauling the net and was damaged when coming into contact with the vessel. The catch losses using the second design with a secondary grid were quite high and statistically significant. The best results were obtained with the third grid design. Overall, the losses were significantly lower compared with the losses observed during Phase 1 and clogging of the grid was not observed. The Mann-Whitney U-tests confirmed the results of the t-tests, except for the third grid design, >=45 ïmn, where the difference was significant. Results 8-101 Table 8-5: The percentage of shrimps lost due to grid selection for < 45 mm, > 45 mm and all length classes (95% CL between brackets) and the p-values of the t-test for independent samples comparing the losses during Phase 1 and the new designs in Phase 2 (RV Belgica trips 97/28, 98/04 and 98/27).

Design % lost p-value % lost p-value % lost p-value <45 mm TL <45 nun TL >= 45 mm TL >=: 45 mm TL all shrimps all shrimps grid - outlet on top 17 (10-24) I 13 (10-16) 14(10-18) I (Phase 1) grid in metal frame 15(12-19) 0.39 12 (9-14) 0.80 13(11-16) 0.56 two grids 43 (36-51) <0.01 30 (23-37) <0.01 37(30-43) <0.01 grid - outlet below 6 (1-12) <0.01 5(0-9) 0.06 5 (1-10) 0.03

Phase 3 A haul log is given in Table 8-6. For brevity, only those hauls are given of which the catch was fully analysed. The good results obtained with this design on the research vessel during Phase 2 (trip 98/27, Table 8-5), could not be replicated during Phase 3. High shrimp loss, which occurred in most of the hauls on the coimnercial vessels, was mostly related to clogging of the grid, hindering the selection of the catch by the grid. This was not caused by starfish but by hydroids, seaweed, jellyfish, plastic bags, woodblocks and other debris. The irregular relief of the coimnercial fishing grounds, with small and large sand dunes, may also have contributed to the losses. The rather high variability in the data also points at an unstable functioning of the grid caused by varying conditions on the fishing grounds. Because of the continuing clogging problem, the design was often changed. Changes were made to the guiding panel in front of the grid and to the outlet.

Table 8-6: Haul log (commercial vessel 99/A-D & F, i.e. Set A; commercial vessel 99/J and 00/A & B, i.e. Set B; research vessel 99/26, i.e. Set C) Haul Date Haul start Haul Haul Date Haul start Haul number time duration (h) number time duration (h)

99/A - 01 30/08/1999 19:15 1:15 99/H -21 13/10/1999 21:11 1:04 99/B - 02 11/08/1999 18:30 0:30 99/1 - 22 28/10/1999 19:40 1:00 99/B - 03 11/08/1999 20:30 1:05 99/1 - 23 28/10/1999 20:50 1:00 99/C - 04 3/09/1999 19:35 1:25 99/1 - 24 28/10/1999 23:40 1:00 99/C - 05 3/09/1999 22:25 1:00 99/J - 25 3/11/1999 15:20 1:05 99/C - 06 3/09/1999 23:40 1:20 99/J - 26 3/11/1999 16:45 1:15 99/D - 07 4/09/1999 19:10 1:05 99/J - 27 3/11/1999 18:10 1:20 99/D - 08 4/09/1999 21:40 1:05 00/A - 33 10/01/2000 15:45 1:30 99/D - 09 4/09/1999 22:55 1:20 00/A - 35 10/01/2000 19:20 1:40 99/E - 10 8/09/1999 20:50 1:00 00/A - 37 10/01/2000 23:00 2:00 99/E - 11 8/09/1999 22:00 1:10 00/B - 38 24/01/2000 11:20 1:30 99/E - 12 8/09/1999 1:50 1:00 00/B - 39 24/01/2000 13:00 1:30 99/F - 13 9/09/1999 19:25 1:00 00/B - 42 24/01/2000 18:40 1:20 99/F - 16 9/09/1999 2:00 1:00 00/B - 44 24/01/2000 22:20 2:00 99/G- 17 15/09/1999 18:40 1:15 99/26 - 01 25/11/1999 7:40 0:30 99/G- 18 15/09/1999 20:15 1:00 99/26 - 02 25/11/1999 8:22 1:00 99/H- 19 13/10/1999 18:45 1:00 99/26 - 03 25/11/1999 9:35 1:35 99/H - 20 13/10/1999 19:55 1:05 99/26 - 04 25/11/1999 12:20 0:30 Evaluation o f the Nordmore sorting grid as a selectivity improving device 8-102 For presentation of the results, the hauls in Phase 3 were grouped according to the alterations made. The first group (referred to as Set A) contains the hauls with the same grid design as used in Phase 2 during the RV trip 98/27, i.e. seatrips 99/A-D & F. The second group contains the hauls with a grid with an enlarged outlet (Fig 8-8) shielded with a sheet of netting that opens when a slight pressure is exerted to allow large debris to leave the net, i.e. seatrips 99/J and 00/A & B (referred to as Set B). This arrangement was also rigged with a guiding panel in front of the grid. The same design was used on the research vessel (sea trip 99/26) but the data were treated separately because of the absence of clogging and are referred to as Set C.

Guiding panel Selvedge

Cod-end Net- 90cm, opening

Shielding device

10 meshes meshes

Top v ie w meshes meshes

Cod-end Netopening

Shielding device ' 5 floats) Outlet Grid 5 floats False rigging of 10 m eshes- '— False rigging of 40 m eshes

Fig. 8-8 - Schematic view of the grid section in the shrimp beam trawl - design with outlet below and shielding device (commercial vessel trips 99/J, 00/A & B and RV Belgica trip 99/26).

During trips 99/E and G to I, changes were frequently made to the guiding panel and outlet to try to optimise it. All these alterations were later abandoned. For brevity, the detailed results for these trips are not presented. On the coimnercial vessels the experimental net caught less of each of the three unsorted catch fractions, i.e. on average over all hauls 39% (28%-50%) less trash, 36 % (27%-45%) less discard shrimp fraction and 32% (26%-38%) less coimnercial shrimp fraction. On the RV the results were much better, with less than 5 % loss of coimnercial shrimp and a reduction of the trash fraction of 44 % (30%-58%). Table 8-7 gives the percentages of lost shrimps and the p-value of the difference between Sets A, B and C and the results with the grid with outlet below in Phase 2. For Set A and B, about one quarter of the coimnercial shrimp catch was lost, which was significantly higher compared to Phase 2. The difference between Set A and B was not significant (p > 0.05). For Set C, the grid selectivity for shrimp was comparable to the results in Phase 2 (p > 0.05) but significantly different from Set B and C (p<0.01). In Fig. 8-9, the average loss of Brown Shrimps, over all hauls in each Set, is given in relation to the body length with the 95% confidence limits as error bars. The length dependence observed in Phase 1 (Fig. 8-3) was far less apparent in Phase 3 (Fig. 8-9). Results 8-103 Table 8-7: The average percentages of shrimps lost due to the grid selection for < 45 mm, > 45 mm and for all length classes (95% confidence limits between brackets) and the p-values of the t- test for independent samples comparing the losses during Phase 2 and the three Sets in Phase 3 (commercial vessel 99/A-D & F, i.e. Set A; commercial vessel 99/J and 00/A & B, i.e. Set B; research vessel 99/26, i.e. Set C). Design % lost p-value % lost p-value % lost p-value <45 mm TL < 45 mm TL >= 45 nun TL >= 45 nun TL all shrimps all shrimps grid - outlet 6 (1-12)1 5(0-9) 5 (1-10) below (Phase 2) Set A 24(16-3 1) <0.01 24(15 - 32) <0.01 24(17-3 1) <0.01 Set B 30(18-41) <0.01 26(1 9- 33) <0.01 28 (20 - 36) <0.01 Set C 7 (-2 - 16) 0.48 4 (-3-11) 0.49 6 (-2 - 13) 0.95

SETA SET B SET C

75%

50%

25%

0% 80 20 40 60 80 20 Length class (mm TL) Length class (mm TL) Length class (mm TL)

Fig. 8-9 - Percentage of Crangon crangon lost with the 95% confidence limits indicated as error bars (commercial vessel 99/A-D & F, i.e. Set A; commercial vessel 99/J and 00/A & B, i.e. Set B; research vessel 99/26, i.e. Set C).

The variability in the data points was much higher on the coimnercial vessels compared with the research vessel. This was probably caused by the irregularity of clogging, as observed on the coimnercial vessels and by the irregular relief of the sea floor during the coimnercial trials. For the same reasons, there were high losses on the coimnercial fishing grounds, which were not observed on the research vessel (Table 8-7). The overall catch reduction in numbers of coimnercial fish due to the grid was 34% (20%- 48%), 32% (25%-40%) and 16% (13%-19%) for Set A, B and C respectively. The grid selection ogives and the length frequency distributions for Dab, Plaice, bib and Whiting are shown in Fig. 8-10A to K. The L50 and SR values are given in Table 8-8. The grid is highly selective for marketable fish with about 75 % of the sized fish guided to the outlet by the grid. The loss of marketable fish was, however, low at only 7 %, thanks to the presence of the 80 nun outlet cover that retained most of the marketable fish. The grid was most selective for marketable roundfish like Cod and Whiting. For flatfish the grid was somewhat less selective. In general the grid was much less selective for undersized fish. Animals below 10 cm length especially, could easily pass through the grid bars in large proportions. Selection started to improve between 10 and 20 cm. Evaluation o f the Nordmore sorting grid as a selectivity improving device 8-104 Table 8-8: Grid selection parameters for fish (commercial vessel 99/A-D & F, i.e. Set A; commercial vessel 99/J and 00/A & B, i.e. Set B; research vessel 99/26, i.e. Set C). Species Grid L50 Grid SR Grid L50 Grid SR Grid L50 Grid SR (mm) (mm) (mm) (mm) (mm) (mm) SetA Set B Set C Z. 582 - commercial vessel Z. 582 - commercial vessel R V Belgica Dab 9.5 15.1 15.9 13.2 26.3 13.8 Plaice 16.4 23.3 14.8 12.9 NA. NA Bib 21.7 20.1 19.4 21.6 N.A N.A Whiting 18.9 10.9 26.8 27.2 24.2 8.9

Dab below 10 cm was caught in high numbers in all experiments. The RV trials clearly showed that the reduction of small Dab (< 10cm) lay below 10 % (Fig. 8-10). This was confirmed in most of the commercial trials. Above 10 cm, selection started to improve and most fish above 20 cm could escape through the outlet. Due to the rather low numbers of Plaice caught, the results for this species were less clear, although they seem comparable to the results for Dab. On average 20 % of the soles escaped. Most of the fish caught had lengths in a narrow range around 10 cm length. For these length classes, the selection looked slightly better than for Dab and Plaice with reductions around 50 %, although the data points were rather scattered. The grid was, however, less selective for the larger fish (< 20 cm) compared to Dab and Plaice since many animals passed through the grid bars into the cod- end. The lengths of bib caught ranged from about 10 cm to 20 cm. The data points on catch reduction were rather scattered and difficult to interpret. The trials showed that bib of 10 cm could easily pass through the grid bars. Selection improved as fish got bigger, and reached about 100 % at 20 cm length. For Whiting, the catch reduction below 20 cm was rather low. Even above 20 cm many fish could pass through the grid bars into the cod-end. Overall, twelve non-commercial fish and invertebrate species were analysed. The average catch reduction was 60% (46-73) and 46% (29-63) for Set A and B respectively. For Set C, insufficient numbers of animals were caught to produce reliable results. For most of the species, the sorting grid was effective in releasing the animals. The reduction ranged from about 25% for Liocacinus holsatus to 90% for Ensis directus. (Fig. 8-11). Results 8-105

A: S et A - Dab B: S et A - Plaice 1 0 0 "i------3000 100 ~ r * ------250 200 2000 150 « ™ 50 100 3 1000

Length class (cm TL) Length class (cm TL)

C: Set A -B ib D: Set A-W hiting 100 T----- 120 1 0 0 -i------400

Ö £ 75 300 m — -E 50 200 100

Length class (cm TL) Length class (cm TL)

E: Set B -Dab F: Set B - Plaice 100 -,------600 100 ! -----

Length class (cm TL) Length class (cm TL)

G: Set B -Bib H: Set B - Whiting 100 -,------120 1 0 0 -i------500 100 '5 £ 75 400 m — 300 « ™ so 200 s 100

Length class (cm TL) Length class (cm TL)

I: Set C -Dab J: Set C -Whiting 100 - |------100 100 -----

Ö £ 75

■E 50

Length class (cm TL) Length class (cm TL)

Fig. 8-10 - Grid selection ogive (fine solid line) and length frequency distributions in the standard (bold solid line) and experimental (bold dashed line) gear for Dab, Plaice, bib and Whiting (commercial vessel 99/A-D & F, i.e. Set A; commercial vessel 99/J and 00/A& B, i.e. Set B; research vessel 99/26, i.e. Set C). For Set C, only the length frequency distribution for whole catch is given as a solid line. Evaluation o f the Nordmore sorting grid as a selectivity improving device 8-106

Pomatochistus spec. Pagurus bernhardus Ophiura spec. Macropodia rostrata Liparis liparis Liocarcinus holsatus Liocarcinus depurator Ensis directus Ciliata mustela Asterias rubens Anthozoa Agonus cataphractus

0 25 50 75 100 Percentage going through the outlet

Fig. 8-11 - The percentage reduction of the different benthos species in the catch of the experimental net (commercial vessel trials, Set A and B).

8.4 Discussion The selectivity of the cod-end for Brown Shrimp in a net with a sorting grid is higher compared to that of a cod-end in a net without a grid. Previous cod-end selectivity experiments (see Section 7.3) with a coimnercial Brown Shrimp beam trawl without grid, gave an L50 of 39.4 ïmn (37.0 - 41.4 ïmn) and a selection factor of 1.82 (1.71 - 1.91). This is significantly lower (p < 0.01) than the present results (see section 8.2.2.2). This is probably due to the cod-end catch size in a net with a grid being smaller and the catch composition more homogenous with less large benthic animals and debris that can hinder shrimp escapes. It should be noted, though, that the trials were carried out on a research vessel and low catch volumes were recorded. In coimnercial conditions, i.e. with higher catch volumes, the difference in cod-end selectivity would probably be less pronounced. In the literature the relation between catch size and selectivity differs. Erickson et al. (1996), Hodder and May (1964) and Madsen and Moth-Poulsen (1994) found a negative correlation between the selection factor and the size of the catch for roundfish. Dalun (1991) and Suuronen et al. (1991) did not find a significant relationship between catch size and L50 for Herring in pelagic trawls. O’Neil and Kynoch (1996), on the other hand, found that L50 increased with increasing catch size. In a previous Belgian experiment (see Section 7.3) with a Brown Shrimp beam trawl, however, it was found that catch size and clogging of the meshes by seaweed and hydroids had a significant, negatively correlated effect on the cod-end selectivity of Brown Shrimp. Catches in a net with a grid are lower in volume and show less clogging of the cod-end meshes. Consequently, the higher selectivity parameters found with the grid were to be expected from and confirm the previous results. During the cod-end selectivity trials in trip 97/08 almost no fish escaped through the cod-end meshes. In a previous study (see Section 7.3) the cod-end L75 for Dab and Sole was 4.8 and 8.7 cm respectively and no Plaice with length above 4 cm escaped from the cod-end. The Dab, Sole and Plaice caught during this cruise were bigger than 5 cm, 7 cm and 5 cm Discussion 8-107 respectively and, therefore, had very little opportunity to escape through the cod-end meshes. Consequently, it was not possible to calculate the selectivity ogives. The percentages of lost commercial Brown Shrimp catch, as given e.g. in Table 8-5, should be interpreted with caution. These numbers are an average over a pooled range of length classes. Since the losses of shrimps were often observed to be length-dependent (e.g. Fig. 8- 3), the “percentage” loss with the same net will differ according to the catch composition. A higher relative amount of large shrimps in a catch would give a higher percentage lost, since large shrimps tend to escape relatively easily. A standardisation of the numbers caught per length class would have been an option. It was decided, however, that the numbers had to reflect a real situation. With the conversion formula to calculate the carapace width of a Brown Shrimp from its total length (Redant, 1978), it was found that the largest shrimps in the catches in this experiment had carapace widths below 11.5 mm. Theoretically these shrimps should be able to pass through the bar gaps of a grid with a 12 mm bar spacing. The tests have shown, however, that this was not the case. Even with a 20 mm bar spacing, a significant amount of the larger shrimps escaped through the outlet. Probably for some of the animals encountering the grid with their bodies not parallel to the gaps, time was too short to obtain the right orientation to pass through the gaps, or other material lying against the grid hindered their passage. The observation that relatively more small shrimps escaped (Fig. 8-3 and Fig. 8-5), however, seems illogical since they should pass through the gaps more easily than the bigger shrimps. A possible explanation could be that the smallest animals are easily taken by the waterflow through the outlet due to their low body weights. This water flow can be quite significant as was observed in a Canadian study with trawl models (Anon., 1996), showing that the waterflow behind a grid can be reduced to about half the normal flow. The reduction in waterflow in that area goes together with an increased waterflow in areas in front of the grid, through net meshes but especially in net parts without obstruction like the outlet. The total numbers of shrimps caught in the lowest and highest length classes, though, were low and the confidence limits wide, which results in less firm conclusions on this matter. The grid selectivity ogives for fish (e.g. Fig. 8-4A to E) demonstrate that the length classes within the selection range lie far below their minimum landing sizes. The grid gives the opportunity to the larger fish (age 1 and higher) to escape through the outlet and almost all marketable fish will be lost if no second large mesh cod-end is attached over the outlet. Age 0 fish, however, still pass quite easily through the grid bars (often 50 % and more) and are caught in the cod-end. This has consequences for the applicability of the sorting grid. If the unwanted by-catch of fish consists mainly of the larger fish, the grid can be useful. If mainly small fish are caught, as in nursery areas, the effectiveness of the grid is low. When comparing the results in Phase 1 from the research vessel to these of the commercial vessel, the variability in the data is higher in the latter, due to the rather low number of successful hauls. The grid selectivity results for Brown Shrimp and Plaice were confirmed in commercial conditions (Fig. 8-3, Fig. 8-4B, Fig. 8-5 and Fig. 8-6B). For Dab, however, more small fish escaped in the case of the commercial vessel (Fig. 8-4A and Fig. 8-6A). Although the opposite seems to be true for Whiting, such a comparison is difficult to make since small fish were not present in the catches on the research vessel (Fig. 8-4D and Fig. 8-6D). The results for Sole were unexpected, since the smallest fish (< 8 cm) showed a higher escape rate than the mid range of the length classes. It is possible that the smallest fish were more easily taken by the waterflow through the outlet. It is, however, also possible that, due to the lower catch volume in the cod-end of the experimental gear, the selectivity for Sole improved when compared to the standard gear. The structure of the comparative fishing experiment, however, does not allow attributing the higher escapes of the small soles to one Evaluation o f the Nordmore sorting grid as a selectivity improving device 8-108 or the other possible cause. For Phase 3, the selectivity retention points were quite scattered which made comparison of designs difficult. During the RV-trials it was clear that more shrimps and fish went through the grid, into the cod-end, compared to the commercial trials. This was probably caused by the absence of clogging on the RV. Comparing the grid designs used during Phases 1 and 3, it was clear that the grid with the outlet below saved remarkably less fish compared to the design with the outlet on top. The loss of marketable Brown Shrimp recorded during this study, when no clogging occurred, was comparable to the 12 % loss found by Graham (1997) with a 14 mm bar spacing grid. Wienbeck (1997) recorded a 15 % loss with a 20 mm bar spacing grid. Both authors also found length dependence in the Brown Shrimp selectivity of the grid with higher losses for the larger animals. The high loss of the smallest Brown Shrimps as recorded in the Belgian experiments was not observed. Many reports on the application of sorting grids in the Pandalus fishery describe some loss of commercial shrimps (e.g. 5 % without length dependence; Isaksen et al., 1992) or even an increase in shrimp catch (Madsen and Hansen, 2001). In each of these experiments, though, clogging was never mentioned as a problem. Graham and Radcliffe, on the other hand, reported severe clogging problems in the English Brown Shrimp fishery with the seaweed Ulva spp. in 20 to 25% of the hauls. The authors also reported handling problems with the grid and concluded that it is unlikely that the device would be taken up on a purely voluntary basis. Graham (1997) found L50 values for a 14 mm grid (outlet on top) for Whiting and Plaice of 11.8 cm and 10.4 cm respectively, which is, in comparison to the present study (Phase 1), higher for Whiting and comparable for Plaice. Clogging by starfish was the key problem for the functioning of the grid with the outlet on top (Phase 1). These animals have a tendency to stick their arms between the bars and hold on. This reduces the selective area of the grid and decreases the opportunities for Brown Shrimps to pass through the bar spacing. Any object, plant or animal that accumulates on the grid will have such an effect. It would also be plausible, however, that an unstable grid angle would cause the catch losses of commercial shrimps. A rolling grid could cause an irregular water flow through the grid outlet. A stronger water flow could take larger amounts of shrimps with it through the outlet and reduce the catch. The first alternative grid in a frame (Phase 2) aimed at reducing this instability. The losses of shrimps were, however, as high as with the frameless grid. Together with the impracticalities of the frame, this led to the discontinuation of these trials. The second design (2 grids), aiming at a reduction of clogging, led to much too high losses of commercially sized shrimps. Clogging was, indeed, reduced but apparently, the water flow was altered in a way that many shrimps were taken through the outlets. The third design, with the outlet below, finally reduced the clogging of the grid by starfish to an acceptably low level. Starfish were found in the outlet cover and did not stick to the grid. It has to be mentioned, though, that the fishery was carried out on a flat fishing ground with rather clean catches. Seaweed, hydroids, big jellyfish, plastic bags, pieces of wood, etc., are very common in certain seasons or throughout the year in the southern North Sea catches. These elements led to a flawed functioning of the grid during the commercial trials in Phase 3 because of clogging. Alterations to the outlet (large enough to allow debris to pass) and the introduction of a guiding panel made clogging less problematic but could only reduce the losses to a small extent. The larger size of the outlet was probably responsible for a higher outflow of water taking part of the catch with it before encountering the grid for selection. The trials with the same design, on the RV, were quite successful, with very low losses of large shrimps. This shows that no errors were made in the rigging of the device. The main difference with the commercial trials was that another fishing ground was used with 1) cleaner catches with less clogging and 2) a flat and hard sandy bottom. It is Conclusions 8-109 obvious that less clogging reduces the loss of shrimps. The effect of the topography of the fishing ground can, however, not be neglected. The commercial fishing grounds off the Belgian coast often have small and large sand dunes. Underwater observations were not possible due to the very low visibility in the water but it seems logical to assume that the up and down movement of the net caused instability of the outlet and an irregular flow of water through the outlet. Probably, part of the catch went through the outlet without encountering the grid and no selection could take place. The rather high variability in the data also points at an unstable functioning of the grid caused by varying conditions on the fishing grounds. Because of the large shrimp losses, during Phase 3, changes were made to the grid on almost every sea trip. Consequently, many experiments were carried out with few hauls per experiment and the data were quite scattered. It was therefore difficult to draw firm conclusions on the selective properties of the grid. The losses of shrimp also pointed to the fact that the grid was not functioning properly and probably part of the catch did not even encounter the grid before it went through the outlet. This probably also contributed to the scattering of the selectivity retention points. The RV data give more reliable selectivity data for shrimp and fish, though these were based on only 4 hauls. This information is, however, hypothetical for the Belgian situation since the sorting grid does not seem to function properly in a commercial situation. On all commercial sea trials, problems were observed with the grid, leading to a loss of commercially sized shrimps. Therefore, a sorting grid as used in this experiment, with a small selection area, a tendency to accumulate part of the catch and susceptible to instability will probably not be accepted by the fishermen in the Southern North Sea. This is especially the case because the vast majority of the fishing grounds within the range of activity of the Belgian coastal shrimp trawler fleet have an irregular relief and contain material that can cause clogging almost all year through.

8.5 Conclusions Several experiments with different sorting grid designs were carried out in the Belgian shrimp fishery to investigate the potential to reduce by-catch. The general layout of the grid was based on the Nord more grid. In a first Phase, a design with an escape outlet on top was studied. If the catch composition did not cause clogging problems, the sorting grid met its purpose. The reduction of age 1 and older fish in the catch was satisfactory. Age 0 fish were sorted out as well, but to a lesser extent. Most benthic animals were selected out by the grid. The commercial Brown Shrimp catch was reduced but usually by not more than 15%. The cod-end catch consisted mainly of Brown Shrimps and required less sorting and the cod-end selectivity for Brown Shrimp improved. In this case, the sorting grid seemed a very favourable choice to improve the selectivity of the Brown Shrimp beam trawl. If, however, there was material in the catch that caused clogging of the grid, the commercial Brown Shrimp catch soon fell below the level acceptable to commercial fishermen. Starfish and debris were the main causes of clogging during these experiments. Seaweeds, hydroids, jellyfish and other debris, however, often occur in the Brown Shrimp catches in the southern North Sea and can be expected to cause the same problems. In this case it can be anticipated that acceptance by the fishing industry will be difficult to achieve. Clogging of the grid with starfish (which occurs in almost every catch) was a constant problem. Several designs were tried out, in Phase 2 of the project, to overcome this problem but none was successful. All these tests were done with a design with the outlet on top of the grid. Subsequent exploratory tests with the outlet below were quite successful and clogging Evaluation o f the Nordmore sorting grid as a selectivity improving device 8-110 with starfish was reduced significantly. Therefore it was decided to adopt this design in a third project Phase. Research vessel trials were successful each time, with low losses of commercial shrimp. It can be concluded, however, that the functioning of the grid, with the outlet below, on board of commercial vessels was not favourable. After various alterations to the grid configuration, the following conclusions could be drawn: • The grid was still very sensitive to clogging material that occurs very frequently in the catches off the Belgian coast; • Large debris and jellyfish often caused problems with a small outlet; • The outlet had to be of a substantial size to allow jellyfish and debris to pass through the opening. The large size, however, caused large shrimp losses. Shielding of the outlet with a rectangular piece of netting that opened when a slight pressure was exerted, thereby allowing bigger pieces of the catch to escape, did not reduce shrimp catch loss sufficiently; • The experiments did not find a compromise between an opening large enough to allow jellyfish and debris to escape and small enough to prevent losing large parts of the commercial catch; • The large loss of commercial shrimp observed on board of the commercial vessels was not observed during the RV trials. Explanations for this difference may be: o The catch composition strongly influences the functioning of the grid. The RV catches were always much cleaner compared to the commercial vessel catches. o The experiments on board the RV were carried out on a flat and hard sandy bottom. Commercial vessels fish on a variety of grounds, not accessible to the RV, with small and large sand dunes and sometimes softer sediment. This can influence the functioning of the grid. It is plausible that when fishing on an irregular bottom, a rigid device like the grid bumps up and down. The outlet can become unstable thereby allowing part of the catch to escape before encountering the grid. • Even when functioning properly, the grid did not allow significant amounts of juvenile fish to escape. Above 10 cm length, selection started to improve. Although the selective grids have some clear advantages, like catch reduction of Age 1+ fish, non-commercial fish and invertebrates and a better cod-end selectivity, the device is too susceptible to malfunction. In the particular Belgian situation, as regards catch composition and fishing ground characteristics, the sorting grid is a difficult device to operate and unlikely to be accepted by fishermen. Introduction 9-111

9 Evaluation of the sieve net as a selectivity improving device

9.1 Introduction The discard problem in the Brown Shrimp fishery has been recognised for a long time (e.g. Gilson, 1935). Several early attempts were made to reduce the by-catch, by increasing the mesh size (Roelofs, 1950, Gilis, 1951, Mistakidis, 1958). This resulted in a reduction of undersized shrimps in the catch, often accompanied with a reduction of the commercial catch. In 1965, the first results were reported on trials with a new selective device for Brown Shrimp trawls to reduce the by-catches (Kuro et al., 1965). The device was based on the separator panel principle. Contrary to the above mentioned experiments, the aim was not only to reduce by-catches of undersized shrimps, but mainly to reduce discarding of undersized fish. These trials were soon followed by Dutch experiments (Boddeke, 1965) with the same device and later on with an early version of the sieve net used nowadays. The report concluded that the catch of juvenile fish and small shrimps was reduced, that the quality of the catch improved and that the workload on board was reduced. Comparable experiments were also carried out in Belgium (Van Middelem and Cleeren, 1967) but in some cases a significant commercial catch loss occurred. Also in the United States similar experiments were carried out with promising results (Anon., 1968). Based on the fairly successful developments of separator panels in shrimp trawls and the perception of the seriousness of the discard problem in shrimp fisheries, FAO organised an expert consultation workshop on the issue (Anon., 1973). In the Netherlands, part of the shrimp fleet adopted the sieve net on a voluntary basis to improve selectivity (Besançon, 1973). The report mentioned, as one of the disadvantages of the system, the loss of commercial shrimps when hydroid polyps were present in the water and stated that the selective trawl should only be used in the absence of material that can clog the sieve net meshes. In France there were plans to enforce the sieve net (Brabant, 1973). The report, however, described the poor selectivity of the device for the smallest flatfish. In the same period, experiments were carried out with separator panels in the northern shrimp fishery, in Norway (Rasmussen, 1973) and Iceland (Thorsteinsson, 1973). Although the fishery had some different characteristics, the issue was identical, i.e. separating fish from shrimps during the capture process. A similar device has also been tested successfully as a turtle excluder device in a shrimp trawl (Kendall, 1990).In several North Sea areas, fishermen started using the sieve net on a voluntary basis. An example is Germany where the device is used dependent on area and season (Mohr and Rauck, 1979). These authors clearly stated that application in the fishery was not based on protecting young fish but on the practicalities for the fishermen, i.e. saving time and labour while sorting and continuing to fish when e.g. large quantities of jellyfish occured in the catches. By the start of the present study in 1995, Denmark was the only country where the sieve net was legally enforced in the Brown Shrimp fishery (van Marlen et al., 1997b). Based on the studies available, the sieve net seems to be effective in releasing part of the discards, fish as well as invertebrates. It is not made of rigid material and therefore it is more acceptable to fishermen than a rigid sorting grid. This, together with the observation that fishermen tend to use the device voluntarily in certain circumstances, makes the sieve net an obvious choice for further study. The sieve net was studied in the frame of the DISCRAN project (see Section 8.1). The trials were carried out by the author and the staff of DvZ on a research vessel and commercial vessels over a one-year period. Evaluation o f the sieve net as a selectivity improving device 9-112 9.2 Materials and methods

9.2.1 Fishing area, vessels and trawls The trials were carried out on board of the RV “Belgica” during two cruises in February- March and November 2000 and on the coimnercial shrimp trawler 0.700, “Bisiti” during fourteen omises between April 2000 and January 2001. A total of 93 hauls was carried out with the sieve net. The purpose of these trips was to evaluate an existing sieve net design in the Belgian shrimp fishery. Details on the fishing area, the vessels and the trawls can be found in Section 8.2.1.

76 1N2B : 2T2B \ 28 135 1N2B :AN

74 1N2B \

70 60 \

1N10B ' '

\ 200 / \ /

\ 120 /

Net: 7.80 m / 9.30 m Vessel: Ropes Sea Fisheries Department Length Material Diameter Ankerstraat 1, 8400 Oostende, Belgium Beam trawler - shrimp a 7.80 m mixed PE Tel. +32 59 342253 - Fax +32 59 330629 Target species : brown shrimp{Crangon crangon ) b 0 80m Source : Lucien Desmil 18.00 Copyright du logiciel: CENTRE NATIONAL DE LA MER / IFREMER Date: 30/03/95 d 4.50 m mixed 18.00

outlet cover

Fig. 9-1 - The experimental net rigged with a sieve net and its netplan. Results 9-113 Sieve nets are rarely used in the Belgian shrimp fishery (van Marlen et al., 1997b). For the design of the sieve net to be used in the Belgian trials, the net plans used by the Dutch (for a 9 m beam) and the UK (for a 6 m beam) project partners were used and scaled to the size of an 8 m beam. Both sieve net designs were being used in the commercial fishery in the Netherlands and the UK at the time of the project. After scaling, both sieve nets were almost identical. Also the position in the net was copied from the other partners. The net plan and the positioning are given in Fig. 9-1. The sieve net had a nominal mesh size of 70 mm, was 116 meshes wide at the front part, 16 meshes wide at the rear end and 60 meshes deep. The outlet of the sieve net was positioned in the belly of the net, close to the cod-end. As the sieve net was already in use in the commercial fishery, no further design and optimisation work has been done. This design was used throughout the trials to test its performance in different conditions. The trials were grouped according to the season and whether they were carried out on a commercial or a research vessel.

9.2.2 The cod-end and the outlet covers The cod-end used was a standard commercial cod-end. Aboard the commercial vessel, the outlet was covered with an 80 mm cover to retain commercially sized fish. On the RV Belgica the outlet was covered with an 11 mm outlet cover to retain all the fractions of the catch that would normally escape. The cod-end was blinded with an 11 mm bag. More details on the methodology are given in Section 8.2.2. The average cod-end mesh openings varied between 21.5 and 21.8mm with standard deviation varying between 0.62 and 0.90mm. The average mesh openings of the cover and blinder were 10.9 and 10.6mm with standard deviations of 0.63 and 0.78mm respectively. The average mesh opening of the large mesh outlet cover was 77.9mm with a standard deviation of 2.30mm.

9.2.3 Data collection and analysis Data collection and analysis was carried out under the same principles as laid out in Section 8.2.3.

9.3 Results

9.3.1 Narrative of the sea trials The first trials with a sieve net were conducted on board of RV “Belgica” for a preliminary assessment of its selective properties at the end of February and the start of March 2000. The first series of trials with a sieve net on board of a commercial vessel was conducted in April, referred to as “CV-Spring” trials. The next series of sea trips was carried out in July and September, referred to as “CV-Summer. The “CV-Autumn-Winter” trials were executed between November and January. A second trial was performed on RV “Belgica” in November to check whether the good results obtained on the research vessel in the off­ season could be replicated in a season when large catches prevail. As these results were comparable, the experiments were grouped and referred to as RV-Trials. An overview is given in Table 9-1. Evaluation o f the sieve net as a selectivity improving device 9-114 Table 9-1: Log of the experimental hauls with the sieve net

Code Vessel Period Number Rigging Action of hauls

RV-Trials RV Belgica February-March and 21 standard sieve sieve selectivity - November 2000 net cover method CV-Spring 0.700 April 2000 15 standard sieve sieve selectivity - net catch comparison CV-Summer 0.700 July and September 2000 20 standard sieve sieve selectivity - net catch comparison CV-Autmn- 0.700 November and December 37 standard sieve sieve selectivity - Winter 2000, January 2001 net catch comparison

In each haul, sufficient numbers of shrimps were caught for data analysis. The numbers of fish in the catch varied strongly, depending on the fish species, time of day, season and fishing ground. For many hauls, though, the numbers caught were sufficient to calculate selectivity ogives at haul level. This was the case for Dab, Plaice, bib and Whiting (except for bib in CV Spring). For each series of experiments and each fish species, the selectivity parameters and corresponding confidence limits were then calculated for the combined hauls. Cod was only caught during the RV-trials. Due to the low numbers in the catches, the selectivity ogive was calculated on the pooled hauls. Sole did not show the normal selection pattern as observed for the other species and the typical ogive could not be fitted.

9.3.2 Sieve net selectivity A log of the fully analysed hauls is given in Table 9-2. On the commercial vessel, the experimental net caught less compared to the standard net for each of the three “unsorted catch fractions”, i.e. on average over all hauls 40% (29%-51%) less main by-catch, 44% (32%-55%) less discard shrimp fraction and 23% (15%-30%) less commercial shrimp fraction. On the RV the results were much better, with less than 8% (7%- 8%) loss of commercial shrimps. The catch reduction of unwanted by-catch, on the other hand, was less satisfactory with 17% (15%-20%) loss of discard shrimps and a reduction of the main by-catch fraction with 28% (24%-33%). Results 9-115 Table 9-2: Haul log (RV-Trials, CV-Spring, CV-Summer, CV-Autmn-Winter) Haul number Date Haul start Haul Haul number Date Haul start Haul time duration time duration (h) (h)

RV-Spring - 1 28/02/2000 18:48 1:57 CV- Aut-Win - 42 15/11/2000 19:40 1:20 RV-Spring - 2 28/02/2000 20:55 2:05 CV- Aut-Win - 43 15/11/2000 21:10 1:25 RV-Spring - 3 02/03/2000 9:43 1:47 CV- Aut-Win - 44 15/11/2000 23:00 1:35 RV-Spring - 4 02/03/2000 11:53 2:07 RV-Autumn - 9 20/11/2000 22:20 1:10 RV-Spring - 5 02/03/2000 14:20 1:20 RV-Autumn - 10 21/11/2000 6:32 1:28 RV-Spring - 6 02/03/2000 15:55 1:35 RV-Autumn - 11 21/11/2000 8:18 1:57 RV-Spring - 7 02/03/2000 19:55 0:50 RV-Autumn - 12 21/11/2000 10:27 2:03 RV-Spring - 8 02/03/2000 21:00 0:45 RV-Autumn - 13 21/11/2000 12:45 2:00 CV-Spring - 1 03/04/2000 21:25 1:35 RV-Autumn - 15 21/11/2000 17:12 2:03 CV-Spring - 2 03/04/2000 23:05 1:25 RV-Autumn - 16 21/11/2000 19:30 2:03 CV-Spring - 3 03/04/2000 0:50 2:00 RV-Autumn - 23 23/11/2000 8:54 1:21 CV-Spring - 4 06/04/2000 9:50 1:10 RV-Autumn - 24 23/11/2000 10:27 1:33 CV-Spring - 5 06/04/2000 13:40 1:45 RV-Autumn - 25 23/11/2000 14:32 1:39 CV-Spring - 6 06/04/2000 1:10 1:50 RV-Autumn - 26 23/11/2000 16:25 1:35 CV-Spring - 7 07/04/2000 17:30 1:30 RV-Autumn - 27 23/11/2000 18:18 1:05 CV-Spring - 8 07/04/2000 19:10 1:20 CV- Aut-Win - 45 18/12/2000 15:30 1:30 CV-Spring - 9 07/04/2000 0:10 1:50 CV- Aut-Win - 46 18/12/2000 17:10 1:20 CV-Summer - 16 18/07/2000 18:45 1:00 CV- Aut-Win - 47 18/12/2000 18:40 2:00 CV-Summer - 18 18/07/2000 21:15 1:15 CV- Aut-Win - 49 18/12/2000 22:40 1:35 CV-Summer - 19 18/07/2000 22:35 1:25 CV- Aut-Win - 51 19/12/2000 18:30 1:25 CV-Summer - 20 18/07/2000 0:05 1:25 CV- Aut-Win - 52 19/12/2000 20:05 1:40 CV-Summer - 22 19/07/2000 18:45 1:00 CV- Aut-Win - 53 19/12/2000 21:50 1:40 CV-Summer - 23 19/07/2000 19:55 1:05 CV- Aut-Win - 56 20/12/2000 18:15 1:30 CV-Summer - 25 19/07/2000 22:30 1:25 CV- Aut-Win - 57 20/12/2000 19:50 1:40 CV-Summer - 28 19/07/2000 4:45 0:15 CV- Aut-Win - 58 20/12/2000 21:40 1:25 CV-Summer - 29 09/09/2000 18:20 1:00 CV- Aut-Win - 62 21/12/2000 17:50 1:30 CV-Summer - 30 09/09/2000 19:30 1:00 CV- Aut-Win - 63 21/12/2000 19:10 2:00 CV-Summer - 32 09/09/2000 22:00 1:00 CV- Aut-Win - 64 21/12/2000 21:19 1:51 CV-Aut-Win - 36 14/11/2000 14:55 1:35 CV- Aut-Win - 67 06/01/2001 18:40 2:20 CV-Aut-Win - 37 14/11/2000 16:40 1:35 CV- Aut-Win - 68 06/01/2001 2:10 2:50 CV-Aut-Win - 39 14/11/2000 20:10 1:50 CV- Aut-Win - 69 08/01/2001 16:55 1:35 CV-Aut-Win - 40 14/11/2000 22:10 1:50 CV- Aut-Win - 70 08/01/2001 21:10 2:20 CV-Aut-Win - 41 15/11/2000 18:25 1:05 CV- Aut-Win - 71 08/01/2001 23:40 2:15

9.3.2.1 Crangon crangon Based on the “numbers of shrimps” in the catches, the RV-trials gave very promising results, with only 7 % loss of the commercial shrimps and a catch reduction for the discard shrimps of 14 % (Table 9-3). These catch reductions were considered to be low, but still highly significant (p < 0.01). The sea trips in CV-Spring more or less confirmed these results with again a rather low loss of the commercial shrimp fraction (13%) and a 34% reduction of the discard shrimps. The difference with the RV-trials was, however, highly significant. The losses of shrimps in CV-Summer were somewhat higher compared to CV-Spring, i.e. 15% and 35% for the commercial and discard shrimps respectively. The difference with the RV- trials was highly significant but not significant compared to CV-Spring. The results for the Evaluation o f the sieve net as a selectivity improving device 9-116 autumn and winter trials were consistently different from the other coimnercial trials and the RV experiments, with very high losses of shrimps. An average loss of coimnercial shrimp catch of over 31 % was observed. Also the catch reduction of the discard shrimps went up to 58%. The differences with all other trials were highly significant. For each of the differences, the t-test results were confirmed with the Mann-Whitney U-test.

Table 9-3: The average percentages shrimps lost due to the sieve net selection for the fraction smaller than 45 mm, the fraction larger than 45 mm and for all length classes (95% confidence limits between brackets) Dataset % lost % lost % lost <45 mm TL >= 45 nun TL all shrimps RV-Trials 14% (10%-18%) 7% ( 6%- 8%) 10% ( 8%-13%) CV-Spring 34% (28%-41%) 13% ( 8%-19%) 20% (15%-25%) CV-Summer 35% (24%-47%) 15% (12%-18%) 26% (19%-33%) C V -Autmn-W:inter 58% (52%-65%) 31% (28%-35%) 44% (39%-48%)

20 40 60 80 20 40 60 80 Length class (mm TL) Length class (mm TL)

100% 100% CV Summer CV Autumn-Winter

£ 75%

50% 50%

25% 25%

0% 0% 20 40 60 80 30 40 50 60 Length class (mm TL) Length class (mm TL)

Fig. 9-2 - Percentage of C. crangon lost with the 95% confidence limits indicated as error bars. Results 9-117 The percentages of lost shrimps by length class are given in Fig. 9-2. For the RV-trials, escapes were low over all length classes. A length effect was quite clear but not pronounced, with somewhat more small shrimps escaping. For the CV-Spring and CV-Summer trials, losses of shrimps were more evident. For lengths above 50 mm, a length effect was not clear and catch reductions were roughly 10 %. Below this length, though, losses increased with decreasing size of the animals and even exceeded 50% for CV-Spring. For the Autumn- Winter trials, a length effect was observed over the whole length range. For the largest animals, losses were close to zero. Contrary to the CV-Spring and CV-Summer trials, the catch reduction decreased steadily from 60% for the smallest shrimps to almost zero for the largest shrimps.

9.3.2.1 Commercial fish species The selective properties of the sieve net for fish in the different trials are given in Table 9-4 and presented graphically in Fig. 9-3 to Fig. 9-6. The selection ogives and parameters by species for all hauls combined over all experiments are given in Fig. 9-7. The selective properties of the sieve net for each of the different fish species were comparable between the research and the commercial vessels and over the seasons (Table 9- 4). Although the catch reductions for shrimps were variable over the different trials, the selectivity of the sieve net for fish was more constant. The sieve net was very selective for marketable fish. The marketable Dab, Plaice, Sole, Bib and Cod catch was almost entirely sorted out (Fig. 9-3 to Fig. 9-6). Only marketable Whiting succeeded in penetrating the sieve net meshes to some extent with retention in the cod-end of less than one quarter of the fish. The 80mm outlet cover retained most of the marketable fish after they escaped through the outlet. The loss of marketable fish due to selection of the 80mm outlet cover was low during the summer and autumn-winter trials with 9% and 19% loss respectively. The losses during the spring trials were higher, i.e. 35%.

Table 9-4: Sieve selection parameters for fish (95% confidence limits between brackets). Species Sieve L50 Sieve SR Sieve L50 Sieve SR Sieve L50 Sieve SR Sieve L50 Sieve SR (mm) (mm) (mm) (mm) (mm) (mm) (mm) (mm) R V Trials CV Spring CV Summer CV Autumn- Winter

Dab 11.3 8.6 9.0 6.4 9.7 6.8 11.6 7.0 (10.7-12.0) (6.2-10.9) (8.1-9.7) (4.5-8.3) (7.7-12.9) (4.6-9.0) (9.9-13.2) (6.0-8.1) Plaice 10.0 5.9 9.0 8.3 10.2 4.9 11.1 5.3 (9.4-10.5) (4.6-7.3) (7.3-10.8) (5.6-10.9) (7.5-15.9) (0.2-9.7) (9.9-12.2) (3.9-6.8) Bib 14.9 8.6 11.7 11.5 14.6 8.4 (13.0-16.6) (6.2-10.9) (-) (-) (7.8-13.6) (5.9-17.1) (12.8-15.9) (6.3-10.4) Whiting 22.7 13.0 18.8 12.2 18.7 8.3 20.3 13.0 (21.5-24.2) (11.0-15.0) (17.7-20.1) (9.3-15.2) (15.7-22.0) (6.0-10.6) (18.6-22.0) (10.9-15.0) Cod 19.2 3.9 (-) (-) (-) (-) (-) (-) (-) (-)

As for the grid, the sieve net showed a very poor selection for all commercial fish species with lengths below 10cm (i.e. Dab, Plaice, Sole, Bib, Whiting and Cod)( Fig. 9-3 to Fig. 9- 6). Usually less than one quarter of these animals managed to escape through the outlet. Above 10cm, selection improved and a higher proportion of the fish could escape. For the Evaluation o f the sieve net as a selectivity improving device 9-118 flatfish species Dab and Plaice, the selectivity was better compared to the roundfish species Bib, Cod and especially Whiting that showed a very poor selectivity. The sieve net had very similar selective properties for Dab and Plaice (Fig. 9-7). Below 10cm length the catch reduction was very low. The selection rose very steeply between 10 and 15cm to reach 100% catch reduction around 20cm. For Sole, the selection pattern deviated from the other flatfish species. For the smallest animals, the sieve net sorted out a rather high proportion. Around 10 cm, the number of escapees dropped to a minimum and increased again for the larger length classes. 100% catch reduction was only reached at about 25cm length. The smallest bib (about 10cm) could easily pass through the sieve net meshes into the cod- end. Above that length selection improved and reached 100% around 25cm length. For Whiting a much shallower selection pattern was observed. As for bib, almost all fish of 10cm length passed through the sieve net meshes but also many fish above 20 cm ended up in the cod-end. The 100% retention was only reached above 30cm. For Cod, the data were scarce and the results should be interpreted with caution. The selection ogive was quite steep compared to the other roundfish species but L50 was comparable to the L50 of Whiting.

30 Length class (cm TL) Length class (cm TL)

E: Whiting F: Cod 1 0 0 -i 3000 100

ö S 75 u i'3 d) o £ £ .5 0 1500

o 4- 30 Length class (cm TL) Length class (cm TL)

Fig. 9-3 - Sieve net selection ogives (black lines for single hauls, light blue line for combined hauls) and length frequency distributions (deep blue line) for Dab, Plaice, Sole, Bib, Whiting and Cod during the RV trials. Results 9-119

A: Dab B: Plaice 100 300 100 120

200 ¡2

40 2

Length class (cm TL) Length class (cm TL)

C: Sole D: Whiting 100 500 100 -, 750

500 ¡2 250'

25

10 20 Length class (cm TL) Length class (cm TL)

Fig. 9-4 - Sieve net selection ogives (black lines for single hauls, light blue line for combined hauls) and length frequency distributions in the standard (deep blue line) and experimental (deep blue dashed line) gear for Dab, Plaice, Sole and Whiting during the CV-Spring trials.

A: Dab B: Plaice 100 1500 100 150

1000 100

500 z

Length class (cm TL) Length class (cm TL)

C: Sole D: Bib 100 600 100 800

400 400

Length class (cm TL) Length class (cm TL)

Fig. 9-5 - Sieve net selection ogives (black lines for single hauls, light blue line for combined hauls) and length frequency distributions in the standard (deep blue line) and experimental (deep blue dashed line) gear for Dab, Plaice, Sole and Whiting during the CV-Summer trials. Evaluation o f the sieve net as a selectivity improving device 9-120

E: Whiting 100 -, 300

200

Length class (cm TL)

Fig. 9-5 (continued) - Sieve net selection ogives (black lines for single hauls, light blue line for combined hauls) and length frequency distributions in the standard (deep blue line) and experimental (deep blue dashed line) gear for Dab, Plaice, Sole and Whiting during the CV- Summer trials.

A: Dab B: Plaice 100 4500 100 1500

3000 1000

500 z 25

10 20 10 20 Length class (cm TL) Length class (cm TL)

C: Sole D: Bib 100 600 100 3000

400 2000

25

Length class (cm TL) Length class (cm TL)

E: Whiting 100 -, 4500

3000

25

30 Length class (cm TL)

Fig. 9-6 - Sieve net selection ogives (black lines for single hauls, light blue line for combined hauls) and length frequency distributions in the standard (deep blue line) and experimental (deep blue dashed line) gear for Dab, Plaice, Sole and Whiting during the CV-Autumn-Winter trials. Results 9-121

All hauls Dab Plaice - - - Bib - - - Whiting # Sole Cod 100 -, D ab:

Plaice:

SR: 5.7 (4,9-6,6)

Bib:

Whiting: L50: 20,8 (19.9-21.

Cod:

Length class (cm TL)

Fig. 9-7 - Sieve net selection ogives for Dab, Plaice, Whiting and Cod and retention points for Sole - all hauls combined over all sieve net experiments on commercial vessels.

9.3.2.2 Non-commercial fish species and invertebrates The average reduction of the unsorted main by-catch fraction (mainly containing invertebrates and debris), due to sieve net selection, on the RV trials was 28% (24%-33%). The reduction for the CV-Spring, CV-Summer and CV-Autumn-Winter trials was 49% (44%-55%), 29% ( 16%-41%) and 41% (34%-48) respectively. The catch reductions of non­ commercial fish and invertebrates at species level are given in Fig. 9-8. For clarity, the results for all coimnercial hauls have been pooled and are presented together with the results for the RV-Trials. Evaluation o f the sieve net as a selectivity improving device 9-122

Pomatochistus spec. Callionymus spec Liparis liparis Agonus cataphractus Myoxocephalus scorpius Syngnathus spec. Ciliata mustela Trisopterus luscus Engraulis encrasicholus Ophiura spec. Asterias rubens Sepiola atlantica Sepia officinalis □ RV-Trials Ensis directus Mactra corallina ■ CV-Trials Abra alba Spisula subtruncata Pecten maximus Nassarius reticulatus Liocarcinus puber Liocarcinus depurator Liocarcinus holsatus Liocarcinus arcuatus Macropodia rostrata Pagurus bernhardus Pectinaria koreni Anthozoa 100%

Fig. 9-8 - The percentage reduction of the different non-commercial fish and invertebrate species in the catch of the experimental net (All CV-Trials combined and RV-Trials).

9.4 Discussion Based on the amounts of lost shrimps, the experiments with the sieve net can be sub-divided into three groups: a) all RV-trials - very low shrimp loss, b) spring and smmner trials - low shrimp loss, c) autumn-winter trials - high shrimp loss. The high shrimp losses for the latter came as a surprise since the previous coimnercial vessel experiments and the RV-trials gave good results, which were quite similar to the results obtained by the other project partners (Van Marlen et al., 2001). For a number of hauls, the poor operation of the sieve could be attributed to clogging with seaweed and hydroids but this was not always the case. Throughout the trials, the same net and sieve net were used, without alterations. In addition, the coimnercial vessel and its skipper and crew were the same throughout the trials. The reason for this sudden change in coimnercial shrimp loss was not obvious. For a number of hauls with high coimnercial shrimp losses, the skipper aboard the fishing vessel noticed gilled fish in the sieve net meshes around the outlet after hauling the net. Quantification and measurement of these fish was not possible since after bringing the net on board, these fish had usually already dropped out of the meshes. Based on the experience of the skipper, however, it was estimated that these fish were large (> 30 cm) Dab and Whiting. The presence of these fish around the outlet would quite likely distort the outlet of the sieve net. Possibly a higher water flow was then directed through the outlet, taking part of the catch with it. In order to check this hypothesis, the correlation between the catch reduction of shrimps and the numbers of large Dab and Whiting was calculated. The relation between both variables is presented graphically in Fig. 9-9. From this graph, it is obvious that no clear linear relation exists between the variables. What this graph does demonstrate, however, is that if there are no large Dab and Whiting in the catch, there is a low probability for loss of shrimps (< 20%). Higher numbers of these fish lead to a higher probability for loss of shrimps (> 30%). Discussion 9-123

Correlation: r = 0.49

• •

9 • • ♦ • • Ä g* *> • • ♦ + y * * « * ------»«C 9*9 9 -»*----- ,--- . * » .» 0% 10% 20% 30% 40% 50% 60% Percentage shrimps lost

Fig. 9-9 - The relation between the loss of shrimps (commercial and non-commercial) and the numbers of large (> 30 cm) Dab and Whiting in the catch.

The early experiments done with sieve nets (Boddeke, 1965; Kuro et al., 1965) in the Netherlands and France, did not indicate reductions of commercial catch. The Belgian trials, however, (Van Middelem and Cleeren, 1967) did pose problems for the scientists being confronted with catch losses of commercially sized shrimps for three of the six sieve net designs tested. The authors could not explain why but it was suggested that the problems might have been caused by the design of the net, although the differences in design were quite small. The catch composition was not studied in detail. Possibly, the causes of the catch losses were the same as in the present study but due to a lack of historic data, this cannot be checked. It is striking that both in the sixties and in the nineties, the application of the sieve net was problematic in the southern North Sea, while it worked well on the other shrimp grounds. Revill et al. (2000) tested four designs of sieve nets and found a length effect for Brown Shrimps. For each design, the losses of undersized shrimps were higher compared to commercial shrimps. For all experiments with a sieve net during the present study, a comparable length effect was observed for the catch reductions of shrimps, with higher losses for the smaller animals (Fig. 9-2). This effect may partly be explained by slightly better cod-end selectivity in the cod-end of the experimental trawl. The effect was small for the RV-trials, medium for the CV-Spring and CV-Summer trials and high for the CV- Autumn-Winter trials. A Canadian study with shrimp trawls (Anon., 1996) demonstrated that the waterflow in the net can be considerably altered by inserting a selective device in the trawl. By inserting a sieve net in the shrimp trawl and cutting an outlet in the belly of the net, part of the water flow is deviated towards the outlet while fishing (minor though due to the high mesh size of the sieve net). This water current takes part of the catch with it through the outlet. It is plausible that larger, heavier shrimps are not easily deviated from their track by this flow. Smaller shrimps, on the other hand, possibly are taken more easily with the water. This could explain the length effect observed and the difference in length effect in the different trials. During the RV-trials, the sieve net worked properly and no clogging was observed. In this case the water flow through the outlet can be expected to be minimal as Evaluation o f the sieve net as a selectivity improving device 9-124 were the shrimp losses. The presence of a small mesh outlet cover also may have further reduced this water flow. During the CV-Spring and CV-Summer trials, the catch losses (and the length effect) were higher, probably because of some degree of clogging causing a somewhat higher water flow through the outlet. During the CV-Autumn-Winter trials, catch losses (and the length effect) were highest due to clogging and gilled fish causing a rather high water flow through the outlet. Consequently, the length effect as observed during the trials strengthens the hypothesis that the malfunctioning of the sieve net was caused by clogging. For the largest shrimps, no increased catch reductions were observed, as often was the case with the sorting grid. The small mesh outlet cover used during the RV-trials may have masked the outlet giving a possible under estimation of the selection process in the sieve net. The results for the RV- Trials should therefore be interpreted with caution and catch reductions may have been under rated. Most of the marketable fish were selected out by the sieve net and were led through the outlet. Without outlet cover, all these fish would escape and be lost to the fisherman. Therefore, it is considered essential that legislation will allow a large mesh outlet cover to be fitted over the outlet to retain the marketable fish and reduce a possible loss of income for the fisherman. The sieve net has proved to be unable to save a significant amount of fish below 10cm, i.e. mainly Age 0 fish. This was also found by the project partners (Van Marlen, 2001) and in early French (Brabant, 1973) and German experiments (Mohr and Rauck, 1979) and stated as a major disadvantage of the sieve net. Above 10cm, selection improves and a large proportion of the Age 1 and older fish escape through the outlet. The biological and economic modelling exercise (see Section 6) has shown that for the typical Belgian catch composition in the shrimp fishery, a reduction in discarding of Age 1 and older fish can benefit the fish stocks. The possible benefits of saving Age 0 fish were considered negligible in this area. Therefore, despite the low selectivity of the sieve net for Age 0 fish, this device would be valuable if adopted by the shrimp trawler fleet in Belgium, rather than in the Wadden Sea area where Age 0 fish are predominantly present. The reduction of the unsorted main by-catch fraction as a whole (mainly containing invertebrates and debris) on the commercial vessel, by sieve net selection, was significant. It was highest in spring, autumn and winter and close to 50%. In summer, this catch reduction was lower (29%), caused by the incidental capture in a number of hauls of very high numbers of unusual small crabs (Liocarcinus holsatus)-, small enough to pass through the sieve net meshes into the cod-end. This catch decreased the average catch reduction. Excluding these hauls gave an average catch reduction of the main by-catch fraction of 53% (45%-61%), which was comparable with the other commercial vessel trials. At species level, the catch reduction of invertebrates and non-commercial fish varied quite strongly, i.e. between 23% and 100%. For the majority of the species, though, this reduction was higher in the commercial vessel trials compared to the catch reductions observed for the same species during the RV-Trials. This also was the case for the reduction of the unsorted main by-catch fraction. Possibly, the absence of clogging during the RV-Trials and masking of the outlet by the small mesh cover may have caused these lower discard reductions. The losses of commercial shrimps with the sieve net, as observed in commercial conditions in spring and summer, were significantly lower (Table 9-5) compared to the losses obtained with the selective sorting grid (Phase 1 and Phase 3, Set A & B). For CV-summer, the conclusions are less firm because the t-tests were not confirmed by the non-parametric tests. The losses observed in autumn with the sieve net were lower compared to the Phase 1 results Conclusions 9-125 with the grid. Comparing the autumn results with the Phase 3 results for the grid, the tests do not convincingly indicate a difference (Table 9-5). It can thus be concluded that, overall, the sieve performed better than the sorting grid.

Table 9-5: The p-values of the t-tests (and Mann-Whitney U-tests between brackets) indicating the significance of the differences between commercial shrimps lost comparing the sieve net and the sorting grid (only commercial vessel trials).

Sorting grid Sieve net Phase 1 Phase 3, Set A Phase 3, Set B CV-Spring 0.03 (0.04) 0.04 (0.04) <0.01 (<0.01) CV-Summer 0.02 (0.08) 0.05 (0.14) <0.01 (<0.01) CV-Autumn-Winter 0.20 (0.78) 0.04 (0.08) 0.10 (0.05)

The selective properties for fish of grids and sieve nets are different. For the sieve net, selectivity was constant over the seasons and trials. With the grid, the results were very variable and probably influenced by the clogging of the grid (see Section 8). This makes a comparison uncertain and of little use, since a properly working sieve net would be compared with a grid of which the selective properties were altered by external factors. The sieve net has proved to be an easy to fit and robust device. Since it had already been used in certain areas in the commercial shrimp fishery, it was well known in the fisheries community. The fishermen had a rather neutral attitude towards this device. As any alteration to the traditional shrimp beam trawl to increase selectivity would probably complicate the fishery and increase costs and labour, fishermen generally dislike selective devices. Several contacts with the fishermen, however, demonstrated that the sieve net could be acceptable. It was feared, though, that clogging in certain seasons could be a problem leading to loss of commercial catches and fishing time. The experiments showed that loss of commercial shrimps is indeed inevitable with the sieve net. These catch reductions usually exceeded 10% and in cases of clogging or presence of gilled fish could exceed 30%. It can, however, be expected that technical alterations to the sieve net outlet could reduce the clogging problem significantly.

9.5 Conclusions During a one-year period, a sieve net has been tested for its selective properties and an evaluation was made of its operational characteristics. The sieve was based on commercially used designs in the Netherlands and the UK. This selective device is being used on a voluntary basis by several shrimp fishermen, although the application in Belgium is rare. The main reason for using it is not the protection of juvenile fish but the practicalities for the fishermen, i.e. saving time and labour while sorting and allowing to keep on fishing when e.g. large quantities of jellyfish occur in the catches. The loss of commercial shrimp catch when using a sieve net was 15% or less in favorable conditions. Certain components of the catch can, however, lead to a distortion of the outlet with a reduction of the commercial catch of over 30%. It is likely that technical alterations to the outlet can prevent this. The sieve net showed very poor selective properties for commercial fish species with a length below 10cm. Above 10cm the selection improves with increasing length. Especially for Agel and older fish, this selective device serves its purpose. Fish of marketable size were led to the outlet and escaped. As the length frequency distributions of fish throughout this report (Sections 5, 7, 8, 9 and 10) indicate, many fish of Evaluation o f the sieve net as a selectivity improving device 9-126 marketable size are caught in the Belgian Crangon fishery. Belgian shrimp fishermen depend for a significant part of their income on fish (Revill et al., 1999). It is justified therefore for legislation to allow the selective shrimp trawl to be fitted with a large mesh outlet cover, allowing marketable fish to be caught whilst allowing undersized fish to escape. The attitude of fishermen towards selective devices is not positive because any alteration to their traditional nets is prone to cause practical problems. The sieve net seemed, however, to be a more or less acceptable device to improve selectivity of their nets. In general it can be concluded that the sieve net was less susceptible to clogging compared to the grid and performed better in different conditions. It was also rarely subject to damage. The loss of commercial shrimps was lower compared to the grid although in certain conditions losses could rise to an unacceptably high level. A length effect in the catch reduction of shrimps was observed. Undersized shrimps escaped to a higher degree compared to the large ones, which reduced shrimp discards. The selectivity of the sieve net for Age 0 fish is very low. Therefore, this device is of rather low value in areas where large amounts of these small fish are caught. The biological and economic modelling exercise has shown that a reduction of discarding of Age 0 fish in Belgian waters has limited benefits for the commercial fish stocks. Saving Age 1 and older fish in the Belgian shrimp fishery, though, could effectively benefit the stocks. Since the sieve is indeed selective for these fish, it can be considered as a valuable tool in reducing fish discards and contribute to reduce the pressure on the fish stocks. On top of this, the application of the sieve net also leads to a significant reduction in unwanted by-catch of invertebrates and non-commercial fish, which would reduce the impact of the shrimp fishery on the marine environment in general. It is, however, important to acknowledge that losses of commercial shrimp in certain seasons and areas, lead to financial losses for the shrimp fishermen. Introduction 10-127

10 Electric fishing

10.1 Introduction Most selectivity enhancing measures concentrate on the net part of the trawl (Walsh et al., 2000). These aim at catch separation or improved filtering of the catch, the disadvantage being that the animals are exposed to net meshes or other parts of the net before they can escape. Damage incurred by contact, or stress caused during the capture and escape process may lead to mortality amongst escapees (ICES, 2000d). A better approach to improve selectivity, if at all possible, is to try to avoid unwanted sizes and species to enter the net. To attain this goal, it is necessary to find alternative means of stimulation in the net mouth, i.e. a stimulus inducing the desired reaction from the target species without stimulating unwanted animals. The traditional way to stimulate Brown Shrimp in a trawl is mechanical, i.e. by the bobbin rope (see Section 4). According to Berghahn et al. (1995), it is mainly the turbulence in the water that causes Brown Shrimps to tail flip. This stimulus, together with vibrations in the sediment and direct contact with the bobbins, indiscriminately stimulate all animals that come into contact with it and offer little opportunity to alter the selectivity of the net. Several other means of stimulation exist, (Van Marlen, 1997a), like light or sound. Many fish species are attracted to light, a response that has been exploited to enhance capture in fisheries all over the world (e.g., Mubamba, 1992). Visual stimuli are also quite important in the fish capture process in trawls (Wardle, 1993), i.e. the effect that different gear parts have on guiding the fish towards the net mouth. Vision can also be important in inducing an escape response, e.g. with square mesh windows made of contrasting coloured netting (Glass et al., 1995). No studies have been done, however, on the stimulation of Brown Shrimps by using light for the purpose of capture. In fact, it is unlikely that light would stimulate Brown Shrimps since in strong light, the animals go into an inactive phase and hide in the sand. Van Marlen (1997a) states that sounds may have either attractant or repellent effects on fish and that a combination of the two might conceivably be exploited to improve the selectivity of fishing gear. There is, indeed, much information and literature available on the reaction of fish to sound. Several studies pointed at the avoidance reactions of fish to vessel noise (e.g. ICES, 1994), the effect on cod-catches of seismic shooting (e.g. Lokkeborg and Soldai, 1993) and the use of sound as a repellent in acoustic barriers (e.g. Knudsen et al., 1994). Sound can also be used to attract fish, e.g. the attraction of fish to feeding sounds (Takemura et al., 1988) and the luring effect of natural sounds like those of feeding and splashing (Maniwa, 1976). No study was found on the application of sound to influence behaviour in the mouth of a net. In Van Marlen (1997a), the olfactory and gustatory senses are also indicated as alternative means of stimulation, though not for application in trawl fisheries. The idea of using electricity in fishing is very old. De Groot and Boonstra (1974) mentioned a reference to Job Baster stating as early as 1765 that electricity might affect shrimps and that this should be investigated. Better known is the work done by Bary (1956) on the behaviour of roundfish in electric fields. Research on the application of electric fields in shrimp trawling started in the late 1960s. De Groot and Boonstra published a first report on an electrified shrimp trawl in 1970. Some promising results were obtained, but after a disappointing trial in 1976, the work on shrimps was terminated. In the same period, electro­ trawls for Brown Shrimps were tested in Belgium (Vanden Broucke, 1972), the United Kingdom (Baker, 1973) and Germany (Horn, 1976) and most of the work pointed at good Electric fishing 10-128 prospects for this type of fishery. The main objective of the work usually was to reduce fuel consumption and to increase the commercial catches with no or very little attention to by- catches. Some experiments, however, already pointed at possibilities for selective fishing with electricity (Stewart, 1975). Research on electric fishing continued into the 1980s but then stopped almost simultaneously in all North Sea countries. This was mainly caused by national bans on electric fishing driven by the fear of overfishing. In other parts of the world, however, interest in the fishing method, especially with an application for shrimps, was maintained. In 1987 experiments with Brown Shrimps in electric fields were reported in Lithuania (Burba and Petrauskiene, 1987). In the USA, a selective electrified shrimp trawl was developed (Holt, 1992), although commercial application was not reported. Also in India experiments were carried out with electric fishing (Van Marlen, 1997a). In 1997, Willy Versluys, a Belgian fishing vessel owner, visited China and reported that over 2000 fishing vessels were using electric pulses to catch penaeid shrimps. He brought a Chinese pulse generator back to Belgium. This renewed the interest in the method. The potential of electric pulses as a means to develop a species selective electro-shrimp trawl was studied in the project “Development of an environment friendly fishing method for the shrimp fishery based on stimulation by electric pulses”. The basic idea was to invoke selectively a startle response for shrimp with electric ticklers and to allow non-reacting species to escape underneath a raised groundrope. The project was a national cooperative feasibility study with the following partners: • The fishermen’s organisation “Rederscentrale”, • ship owner “Brevisco” and • the Sea Fisheries Department. The scientific work was carried out by the author and the staff of DvZ. The project was financed by the Flemish Community, the Belgian Ministry of Agriculture and the European Union and was set up as a feasibility study aiming at the development of a shrimp trawl with the following aims: • improving species and length selection, • reducing discards, • reducing the impact of shrimp trawling on the environment and • improving the quality of the commercial catches. It consisted of two phases, (1) laboratory experiments for the study of the characteristics of the equipment and the behaviour and survival of shrimps, fish and invertebrates in the electric field; (2) sea trials to test a preliminary design of electro-shrimp trawl.

10.2 General arrangements Pulse generators have been studied quite intensively in the past for application in sea fisheries. The main interest was usually focused on increasing catching efficiency of the fishing gear. In view of the state of the stocks in the North Sea and the growing concern about the marine environment, catching efficiency only plays a minor role in research today. The idea that electric pulses might be used as an alternative stimulation to invoke selectively a reaction from shrimps without stimulating fish and other invertebrates, led to this project. The basic idea was to raise the groundrope and stimulate shrimps to jump over this raised groundrope. Animals not stimulated could then escape underneath the net. If effective, this could be used to increase the species selectivity of the groundrope of a towed fishing gear. The project “Electric fishing” consisted of following phases: Phase 1 - Laboratory experiments Materials and methods 10-129 • Preliminary observation tests to assess the applicability of the available pulse generators. • Study of the pulse generators and the electric field: some basic measurements have been executed to increase the understanding of functioning of the electronic equipment and the characteristics of an electric pulse field in seawater. These measurements were not meant to be a thorough study but were used as guidelines to steer the experiments. A more detailed study was not possible due to lack of time and finances to carry out tests, buy equipment and attract the necessary expertise. • Study of the reaction of Brown Shrimps, other invertebrates and fish in the electric field. These experiments were meant to assess the feasibility of electric pulses as a tool to increase the species selectivity of the shrimp trawl. • Survival experiments to determine whether the pulses cause mortality. Phase 2 - Sea trials • Alteration of the existing shrimp trawl to fix electrodes to the groundrope to fish selectively. • Comparative sea trials. • Defining the optimal characteristics of a pulse generator and a selective shrimp trawl as a basis for a follow-up project.

10.3 Materials and methods

10.3.1 Phase 1 - laboratory experiments

10.3.1.1 Preliminary observation tests The pulse generators available commercially were a product of China and developed for penaeid shrimps, which are large, compared to the Brown Shrimps in the North Sea. Nothing was known about the response of North Sea fish and invertebrates to the electric pulses of this apparatus. For the purpose of the project, it was important, however, that the response of shrimps would be maximal and the response of other animals minimal. If this would not be the case, a new pulse generator would have to be developed. It was therefore necessary to perform some preliminary tests to assess the response of North Sea animals to the Chinese pulse generators. The preliminary observation tests were carried out in an aquarium on Brown Shrimps, Dab, Plaice and Sole.

10.3.1.2 Aquaria and instrumentation For the trials in the laboratory, an aquarium infrastructure had to be built to keep a supply of laboratory animals and to perform the behaviour observation and survival tests. For the storage, seven aquaria with a total volume of 6000 litre were available. The behaviour observation tests were performed in a Plexiglas tank of 210cm L x 110cm W x 60cm H. The animals from the survival experiments were kept in a series of 18 aquaria of 80cm L x 60cm W x 32cm H. The bottom of the aquaria was covered with sand and the daylight was obscured to simulate natural conditions. Close attention was given to the quality of the seawater. It was regularly tested and replaced with fresh seawater obtained through a pipeline connection with the sea. All aquaria were provided with a biological and a UV- filter. Depending on the species, the animals were fed with mussels and pieces of fish or pellets. Electric fi s hi ng 10-130 Laboratory animals were collected in tidal waters or during fishing operations at sea. In tidal waters, a light man-operated beam trawl was towed for about 5 minutes. The animals collected in buckets were transferred to the institute as soon as possible. At sea, short (10 min.) fishing hauls were carried out with a shrimp beam trawl to obtain fish and invertebrates. The animals were stored in closed containers with a continuous water flow. Once ashore, the animals in good condition were transferred to the aquaria in the institute. The animals usually adapted well to the new enviromnent and remained in good condition and feeding for a long time with very low mortality. The pulse generators used in this project are made in China. Two types of generators exist, i.e. the transformer and the capacitor pulse generator. With the former, a high-tension voltage is transformed to the desired pulse voltage. Interruptions in the circuit determine the frequency and the duration of the pulse. This is the standard pulse applied in the Chinese fishery. The main disadvantage of this system is the safety risk in working with high voltage aboard a fishing vessel. The capacitor pulse generator, which was in the development phase at the time of the project, lias some clear advantages. Energy consumption is lower and it is claimed to be more efficient in stimulating shrimps (pers. connu, prof. Zhong Wei Guo). The pulse is bipolar, thus avoiding corrosion of the electrodes.

Fig. 10-1 - The 4 different pulse generator types available to the project, i.e. MJX-50 (left top), Tongfa-98 (right top), LWY with cable winches (left bottom) and LPG (right bottom).

Four types of pulse generators were used in the project, all of the transformer type (Fig. 10- 1), i.e.: • Pulse generator type MJX-50 : Commercially used instrument developed by the Zhoushan Department of the Oriental Science Industry and Trade Company Group (OSITCG) (Zhoushan, People’s Republic of China). This apparatus consisted of a control unit, to be installed aboard the vessel, the pulse generator, the connecting cables and the electrodes. Materials and methods 10-131 • Pulse generator type Tongfa-98: Commercially used instrument developed by the Zhoushan Import and Export Corporation of Zhejiang (Zhoushan, People’s Republic of China). This apparatus consisted of a pulse generator, with batteries, not connected to the fishing vessel. • Pulse generator type LWY : two prototype pulse generators developed by the Ningbo Haitian Group Corporation (Ningbo, People’s Republic of China) to meet the higher pulse voltage needs for the Crangon fishery. These apparatuses consisted of a control unit, to be installed aboard the vessel, the pulse generator, the connecting cables and the electrodes. • Pulse generator type LPG: A laboratory pulse generator specially developed for laboratory tests in this project. The apparatus had an adjustable pulse frequency and voltage. This generator was developed as an initiative of the Sea Fisheries Department with co-operation of the following partners: o Department of Electrical energy, systems and automation - Prof. Dr. ir. A. Van den Bossche: electronic design and follow-up. o KHBO-Katholieke Hogeschool Brugge Oostende - Bart Huyghebaert: construction. All details can be found in Huyghebaert (1999). As none of the characteristics of the sea going pulse generators were known, some rudimentary measurements of the pulses and the electric field were made. This was done in the observation tank with short electrodes (lm long at 50cm distance) using an oscilloscope (Fig. 10-2). For measuring the so-called head-tail voltage (potential difference between the head and the tail of a shrimp in the electric field) a simple bipolar probe connected to the oscilloscope was used. As it was suspected that the pulse generators would behave differently with a full load at full scale, the operational characteristics were also measured on a full-scale electrode array (Fig. 10-2) in seawater. For this purpose, the pulse amplitude was measured with an oscilloscope at several positions in the electric field. The current through the different cables was also determined. Since the spacing between the electrodes throughout the project usually was 50 cm, the pulse has usually been described with the tension generated (V) instead of the electric field strength (V/cm).

observation aquarium pulse generator Non-insulated anode 'Jon-insulated cathode nsulated anode nsulated cathode

electrode

electrode pulse generator

Fig. 10-2 - Arrangement for the measurement of the pulse generator characteristics in the laboratory (left) and at full scale (right).

10.3.1.3 Minimum pulse amplitude In order to determine the minimum pulse amplitude at which shrimps are startled, the following test was made: Electric fishing 10-132 • A homogeneous field was created in the observation tank by fixing two copper plates parallel to each other at a distance of 50cm. These plates were connected to the output of the pulse generator LPG. • Small shrimps (± 3cm) as well as large shrimps (± 6cm) were positioned in the observation tank and put through an electric pulse test. The tests started with a pulse amplitude < IV and were gradually increased in steps of IV. This was done until all shrimps responded. • For each test, the reactions were recorded for an orientation parallel and perpendicular to the electrodes. For each amplitude step, animal size class and orientation, 30 animals were tested.

10.3.1.4 Observation tests - Brown Shrimp The purpose of the observation tests was to determine what percentage of shrimps in the trawl path could be caught with electric pulses as an alternative stimulation if the groundrope was raised at a certain distance from the seafloor. For this purpose, the following parameters were measured: • the percentage of shrimps being startled, • the height of the startle response, • the time elapsed between the start of the pulse and the maximum startle response and • the duration of the swimming phase. All these characteristics were measured in relation to the following variables: pulse amplitude and frequency, shrimp size, light conditions and water temperature. The observation tank is shown in Fig. 10-2. Two threadlike electrodes were used, laying 50cm apart. For each combination of variables, six tests with 10 shrimps were carried out. The animals were transferred from the holding tank to the observation tank with minimal handling. There they were given sufficient time to settle before the electric field was switched on. The tests were continued for about 30 seconds. Each test was video taped with a Hi-8 camera. This allowed analysis of the pictures at 25 images a second. For each shrimp, the distance from the bottom was determined as a function of time after the start of the pulses. The tank was sub divided into five vertical zones, 0-10cm, 10-20cm, 20-30cm, 30-40cmand > 40cm. For each unit of time, the numbers of shrimps in each vertical zone were counted. For interpretation of the data, these numbers were presented graphically for each test. The time necessary to obtain a maximum reaction and the numbers of shrimp above 10 cm at that time were determined for analysis. The variables water temperature and light conditions were a limiting factor for the experiments. The temperature prevailing in the aquaria was 12°C. Reducing that temperature was very demanding for the cooling equipment and could not be maintained for a long time. Observation tests in low light conditions were very time consuming when analysing the images. Therefore, the variables pulse amplitude and frequency and shrimp size were tested to their maximum range at a 12° water temperature and during daylight conditions. For other water temperatures and dark conditions, the other variables were tested only over a limited range. Materials and methods 10-133 10.3.1.5 Observation tests - flatfish and other demersal species The purpose of these tests was to determine the reactions in an electric field of commercial flatfish species, non-commercial fish species and invertebrates. Their reactions are of the utmost importance for the selective groundrope to work properly. It is based on a strong startle response for shrimps that jump higher than the groundrope and can be caught. The other species need to escape underneath the groundrope so their reaction should be minimal. The aquarium arrangement used was the same as for shrimps and is shown in Fig. 10-2. Two threadlike electrodes were used, laying 50cm apart. The animals were transferred to the observation tank and were given sufficient time to settle. The video camera was switched on and 15s later the pulses were started. After a further 15s, the pulses were switched off and the reactions of the animals were recorded for another 15s. The experiments were carried out in dusk conditions, water temperature of 12°C and pulses of 65 V and 6Hz.

10.3.1.6 Survival experiments The survival tests were intended to determine whether the electric pulses would induce mortality amongst the most important fish and invertebrate species in the catches of Belgian shrimp trawlers. For this purpose, the animals were exposed to the electric field in the observation tank for at least 15s (Fig. 10-2). Two threadlike electrodes were used, laying 50cm apart. The results of these tests can be considered as quite conservative since at sea, the animals would only be exposed to the pulses for less than 4s. After the tests, the animals were transferred to holding tanks and observed for 10 days to 1 month. In order to reduce stress, the densities in the holding tanks were kept low. As a rule of thumb, this density was at most half the density in the aquaria before the test. For each tested group of animals, a control group was put through identically the same procedure, except for the pulses. This was necessary since mortality is likely to occur with captive animals, even without being exposed to experimental treatments. The survival of the test group was expressed as the percentage of animals surviving compared to the control group. After each test, the numbers of dead animals were recorded. The behaviour and feeding activity of both experimental and control groups was also observed.

10.3.2 Phase 2 - sea trials

Plan of the field experiments The aim of the field experiments was to try and find ways to alter the shrimp beam trawl to let invertebrates, non-commercial fishes and undersized commercial fishes escape underneath the net while keeping the losses of commercial shrimps minimal. This had to be done within narrow time and financial margins, characteristic to a feasibility study. Hence, it was impossible to design and test a new shrimp trawl from scratch, built for the purpose of electro-fishing. Therefore, the existing shrimp trawl had to be adapted to allow electrodes to be fitted and the groundrope to fish selectively. In the traditional shrimp trawl, the groundrope fishes rather close to the seafloor and lies about 40cm behind the centre of the bobbins (Fig. 10-3). According to fishermen, it is the tickler effect of the bobbins that stirs up animals from the seafloor. Once stimulated, the distance between the bobbins and the groundrope gives the animals sufficient time to reach a height above the groundrope. For the electro-fishing trials, this groundrope was raised, Electric fi s hi ng 10-134 leaving a vertical gap of 10-15 cm, and was rigged shorter so it would lie against the bobbins (Fig. 10-3).

Electrode array A Electrode array B

anode cathod' anode cathode

Schematic presentation of a standard shrimp trawl Schematic presentation of the adapted shrimp trawl with raised groundrope area of electric field

$10-15 cm bobbin trawlhead normal position belly of the net raised groundrope of the groundrope

Fig. 10-3 - A shrimp beam traw l altered for electro fishing.

The gap provides an escape route underneath the net and the contact between bobbins and groundrope eliminates the stimulation effect of the bobbins. To prevent shrimps from escaping underneath the net, an electric pulse field was generated in the net opening, sufficient to initiate a startle response for the shrimps to be caught. Because the pulse amplitude of the generator used (type LWY) drops from 120 V to 44 V at full load, usually two generators were used simultaneously in order to attain an amplitude of 64V (see Section 10.4.1.2). Two electrode arrays were tested: (1) parallel anodes and cathodes, gliding over the seafloor and (2) cathodes gliding over the bottom and anodes, rigged perpendicularly in the top panel of the net (Fig. 10-3). The latter gave a slightly better startle response for shrimps during the observation tests but had the disadvantage of being more difficult to fit in the net.

10.3.2.1 Vessels, fishing gear and fishing grounds A first exploratory sea trip was carried out on RV Belgica between 22 and 26 November 1999. The purpose of this trip was to try out the pulse generators and the rig of the electrodes and groundrope to find any practical problems. To compare the experimental catch with a nonnal coimnercial catch, it was arranged that a coimnercial shrimp trawler would fish alongside RV Belgica. Two members of the DvZ team embarked on board of this vessel (Z.582 - Asanat) to analyse the catches. Results 10-135 Experimental electro-fishing in commercial conditions was carried out in the year 2000. The vessel chartered was the 0.700 (see Section 5.2.2). A total of 12 sea trips and 90 hauls were carried out. The experimental trawl was towed at one side of the vessel and the standard trawl at the other side, allowing comparison of catches obtained in the same conditions. Ideally, a fishing gear developed especially for the purpose of electro-fishing should have been used. Due to the circumstances in the project, the existing standard shrimp beam trawl was used and modified. The pulse generators (type LWY) were fitted to the beam of the net and connected to the electrodes rigged into the net mouth. The two electrode arrays given in Fig. 10-3 were tried out. Between the two beam trawl shoes, a chain was rigged for the connection of the electrodes, lying 50cm apart. The electrodes were partly loaded with light chains for a moderate bottom contact. At sea, the bottom contact was regularly checked by looking at the abrasion of the chains. The electrodes in the top panel were woven through the net meshes and lay 50cm apart. The principle outlined in Fig. 10-3 for a selective groundrope was used. To raise the groundrope and bring it closer to the bobbins, the groundrope itself was shortened. At first, the connectors between the groundrope and the bobbins were shortened but were still long enough to allow some movement of the groundrope. This resulted, however, in a rather unstable groundrope often rising too high above the bobbins. Therefore, later in the project, these connectors were further shortened, pulling the groundrope tight to the bobbins. During most of the trials, the net itself was not altered, except for three trials when an 11mm mesh piece of netting was rigged to the front part of the top panel (28mm meshes). Aboard the vessel, the control units verified the functioning of the generators. Two man- operated portable winches were installed aboard for the shooting and hauling of the cable between the power supply and the generators. The traditional shrimp fishing grounds off the Belgian coast were fished and the choice of area was left up to the skipper to meet commercial conditions. Further details on the vessels, the fishing gear and the fishing grounds can be found in Section 8.2.2.1. Several series of trials were carried out with different net configurations.

10.3.2.2 Data collection and analysis A detailed description of the catch handling and data analysis is given in Sections 5.2.3, 7.2.3 and 8.2.2.3. To evaluate the efficiency of the electro-net, the catches of this net were expressed as a percentage relative to the standard net, by species. The significance of differences was assessed with t-tests. For Brown Shrimps, length frequency distributions were calculated together with the percentage difference in catch by length class. For commercial fish species, fish below and above MLS were distinguished. For non-commercial fish and invertebrates, total numbers were used to evaluate the differences in catch.

10.4 Results

10.4.1 Phase 1 - laboratory experiments

10.4.1.1 Preliminary observation tests The results of these preliminary tests were positive with respect to the purpose of the project. All shrimps reacted to the pulses and made an upwards or sideways movement. Plaice and Sole did not react to the pulses. Dab did react by swimming upwards at a steep angle. Electric fi s hi ng 10-136 This meant that the Chinese pulse generators could be used in the project. It had to be borne in mind, though, that the pulses of these generators were not adjustable. Further development of electronic equipment might still be necessary.

10.4.1.2 Instrumentation The characteristics of the pulse generators, as observed in the laboratory (Fig. 10-2), are given in Table 10-1.

Table 10-1: Characteristics of the pulse generators used in the project. Pulse Input voltage Link between Input voltage Pulse Pulse Pulse generator control unit vessel and generator amplitude frequency duration (V) generator (V) (V) (Hz) ' (ms) MJX-50 24 Cable 400 65 5 0.4 Tongfa-98 (-) None (fed by 48 45 5 0.5 batteries) LWY 24 Cable 350 123 5 0.6 LPG (-) None (only for 230 0 -2 0 0 1 -9 0.6 laboratory use)

140 120 100 — 120 «¡ 100

— Tongfa-Ç — LPG ■■■LWY I — MJX-50

0.5 500 1000 1500 2000 Time (ms) Current (A)

Fig. 10-4 - Graphical representation of the pulse. Fig. 10-5 - The measured relation between the output current and the pulse amplitude.

The graphical representation of the pulses is given in Fig. 10-4. The full-scale measurements (Fig. 10-2) demonstrated that the pulse amplitude drops quite sharply at full load. The measured relation between the output current (depending on the configuration and length of the electrodes) and the pulse amplitude for three generators is given in Fig. 10-5. For generator LWY the amplitude dropped from 123 V to only 44V when the full-scale electrode array was fed by the generator instead of the short laboratory electrodes. The pulse frequency and duration remained constant. The current at the output of the generator increased sharply from less than 250A with short electrodes in the laboratory, to 1200A at full scale. Due to the high current in the cables, the pulse amplitude measured at the electrodes was somewhat lower compared to the amplitude at the output of the generator (i.e. IV less at the start and 2V less at the end of the electrodes). When half of the electrodes were disconnected from the generator, the drop in pulse amplitude was not as substantial. The amplitude then measured was 64 V with a current of 900A. Results 10-137 The head-tail potential difference over 1cm in the electric field for generator LWY, measured perpendicular to the electrodes (which lay 50cm apart - Fig. 10-2) in 5 vertical positions is given in Fig. 10-6. The linear potential difference is 1/50 of the pulse amplitude, i.e. 2.5V/cm. Because of the shape of the electrodes, the electric field is not linear and much higher values were recorded close to the electrodes compared to other positions in between the electrodes. In relation to the vertical position, the highest values were recorded at the bottom position. The higher up, the more the voltage gradient dropped. Outside of the electrodes, the voltage quickly dropped to a value close to zero. In addition to the electrode array considered above, seven others were tried out. For brevity and since most of these were not used further in the project, the results are not included in this report but can be found in Polet (2001).

100%

bottom 10cm 20cm 30cm Large, + 40cm Small, + Large, // Small, //

■20.0 Cath0de 30.0 anode 0 50 100 150 200 250 300 Horizontal position (cm from cathode) Pulse amplitude (mV/cm)

Fig. 10-6 - Potential difference over 1 cm, Fig. 10-7 - The pulse amplitude in relation to the percent- measured perpendicular to the electrodes, age response of shrimps for 1) large shrimps perpendicular to the electrodes, 2) small shrimps perpendicular, 3) large shrimps parallel and 4) small shrimps parallel.

10.4.1.3 Minimum pulse amplitude The response results are given in Fig. 10-7. The lowest head-tail voltage that invoked a response for a fraction of the animals in the test was 40mV/cm, in the case of large shrimps perpendicular to the electrodes. At 80mV/cm, all animals were stapled. For small shrimps, a somewhat higher voltage was needed for the first reaction, i.e. 60mV/cm. At 120mV/cm all animals responded. For large and small animals parallel to the electrodes, 180 and 240mV/cm respectively was needed to invoke a 100% reaction.

10.4.1.4 Observation tests - Brown Shrimps After the shrimps were transferred from the holding tank to the observation tank for testing, they settled on the sand and usually dug into the sand until often only the eyes and antennae were visible. As soon as a sufficiently strong pulse field was switched on, a startle reaction was observed for all shrimps at the first pulse. The body movement was a contraction of the abdomen, referred to as tail-flip. With the following pulses, the shrimps continued the startle reaction and maintained a swimming phase until so-called pulse fatigue set in. The tail-flips weakened and the animals slowly sank to the bottom. It took at least 15s for pulse fatigue to show. After a few minutes, the shrimps recovered and dug themselves into the sand. A detailed analysis of the tail flip movement showed that the frequency of the contraction coincided with the frequency of the pulse. A low frequency (< 3Hz) resulted in a discontinuous tail flip with a short rest between the contractions. A frequency between 5 and 6Hz resulted in continuous and complete contractions of the body. Higher frequencies also Electric fi s hi ng 10-138 gave a continuous tail flip but the contractions were incomplete, resulting in a slower movement.

~ 60% ou Large - day -1 2 ° - 6Hz * 40% a Large - dark - 9° - 4Hz V) Large - day -1 2 ° - 4Hz c Large - day -1 2 ° - 6Hz ■ Large - dark - 9° - 6Hz o * Large - dark -1 2 ° - 4Hz Large - day -1 2 ° - 8Hz □ Large - dark -1 2 ° - 6Hz « 20% Large - day -1 7 ° - 6Hz Small - day -1 2 ° - 6Hz

Shrimp size Large Large Large Large Large Small Large Light Day Day Day Day Day Day Day

Temperature 1 2 ° 1 2 ° 1 2 ° 17° 17° 12° 12°

Frequency 4Hz 6 H z 8H z 6 H z 8H z 6 H z 6 H z

Pulse amplitude Percentage shrimps above 10cm at the time of maximum response (95% CL between brackets)

15 5% (-3-13) 12% (6-17) 20% (13-27) 1 2 % (6-19) 18% (3-32) 1 2 % (9-15) 1 0 % (7-13) 25 22% (17-27) 30% (27-33) 35% (27-43) 32% (21-43) 30% (24-36) 29% (23-35) 2 0 % (16-24) 35 38% (35-41) 48% (44-52) 51% (41-60) 48% (35-62) 47% (39-54) 34% (29-39) 47% (46-48) 45 40% (27-53) 48% (42-54) 57% (46-69) 53% (41-65) 51% (41-61) 39% (35-43) 70% (68-72)

55 43% (24-61) 50% (35-65) 53% (43-63) 55% (41-69) 60% (56-65) 43% (34-52) 6 8 % (62-73)

65 44% (38-50) 50% (45-54) 48% (41-55) 56% (45-67) 64% (54-74) 47% (39-55) 65% (60-70) 75 47% (39-55) 52% (43-62) 47% (35-59) 59% (47-71) 70% (59-81) 49% (38-60) 54% (47-60) 85 49% (40-58) 52% (47-56) 47% (24-69) 64% (56-72) 75% (65-85) 51% (40-62) 43% (36-49) 95 56% (48-64) 59% (46-73) 45% (29-61) 69% (61-77) 72% (65-79) 53% (43-63) (-) 105 64% (57-71) 58% (54-62) 44% (29-59) 75% (69-81) 6 8 % (60-76) 51% (42-60) (-)

115 57% (51-63) 5 4 % (48-60) 43% (32-54) 71% (65-77) 61% (55-67) 48% (44-52) (-) 125 49% (40-58) 50% (38-62) 42% (36-48) 6 6 % (62-70) 59% (51-67) 49% (45-52) (-)

Shrimp size Large Large Large Large Small Small Small Light Dark Dark Dark Dark Dark Dark D ark

Temperature 9° 9° 1 2 ° 1 2 ° 9° 1 2 ° 1 2 °

Frequency 4Hz 6 H z 4Hz 6 H z 6 H z 4H z 6 H z Percentage shrimps above 10cm at the time of maximum response (95% CL between brackets)

65% (62-67) 61% (43-79) 71% (64-77) 70% (63-76) 52% (47-57) 65% (59-71) 60% (54-65)

Fig. 10-8 (+ data table) - The percentage of shrimps being startled by electric pulses above 10cm height at the time of maximum response after the start of the pulse field. The co-variables taken into account were shrimp size, light conditions, water temperature, pulse frequency and pulse amplitude. The 95% confidence limits are given between brackets. The data indicated with (*) gives the results for a homogeneous field. The startle response for Brown Shrimps in relation to the co-variables (shrimp size, light conditions, water temperature, pulse frequency and pulse amplitude) is given in Fig. 10-8. For readability of the graphs, the confidence limits are given in the data table below the figure. Over the whole range of the co-variables, the pulse amplitude was the most determining factor (Fig. 10-8). At a low amplitude (< 20V), the startle response was minimal (< 20% of the shrimps above 10cm height). Increasing the amplitude from 20 to 30V roughly doubled the nmnbers of shrimps jumping over 10cm high. The response further increased quite Results 10-139 strongly up to 35V. With a further increase of the pulse amplitude, the curves flattened but the response increased until a maximum was reached. The maxima depended quite strongly on the co-variables. Beyond the maximum, the response decreased with increasing amplitude. For the curve “Large - day - 12° - 6Hz”, a maximum was reached at 95V but the response did not vary much between 35 and 120V. At a higher frequency (8Hz) the maximum was reached at 45V. A pulse amplitude over 45V had a negative effect on the response. At a lower frequency (4Hz), the maximum response was observed at 105V. The maximum response for the three frequencies was quite comparable but the frequency mainly determined the amplitude at which this maximum was reached. Small shrimps responded to a lesser extent compared to large animals, although the difference was not high and for most data points not significant. The effect of the water temperature was also clear from Fig. 10-8. A higher temperature resulted in a better response over the whole pulse amplitude range and the maximum was reached at a higher amplitude. At 9°, the response was weaker compared to both 12° and 17°. The temperature effect was the same for each value of the other co-variables. The tests at low light intensity were only carried out at a pulse amplitude of 65V (Fig. 10-8). It was clear, however, that all tests in dark conditions resulted in a better response compared to light conditions. 100%

75%

■ 40cm 25% ■ 30cm □ 20cm □ 10cm

20 Time (s)

Fig. 10-9 - The time between the start of the electric Fig. 10-10 - Graphical representation of pulses and the maximum response. The 95% confl- the response of shrimp in an electric dence limits are given as error bars. M ark that the field, data points have been shifted left or right to allow better readability of the error bars.

The time span needed to have maximum response, i.e. a maximum number of shrimps above 10cm, is given in Fig. 10-9. Mark that the time for maximum response should be linked with the data in Fig. 10-8; a rapid response does not necessarily mean a strong response. At a frequency of 4Hz and a low pulse amplitude, it took about 9s for the maximum response to be reached. Increasing the amplitude reduced that time to about 4s at 125 V. At 6Hz, this time was usually below 4s, also at low amplitude. High amplitude, however, seemed to have a negative effect and resulted in a slower response. The time for maximum response was minimal at a high frequency, with the better results for a medium amplitude. The other co­ variables did not have a clear effect on the time for maximum response (results not shown). One example of the graphical representation of the startle response of shrimps in an electric field is given in Fig. 10-10. This graph gives the percentage of shrimps present in a certain zone above the bottom over the whole time span of the test (30s in this case). For each Electric fi s hi ng 10-140 combination of variables given in this section, such a graph was made. Fig. 10-10 gives a typical evolution of the startle response. After a short time, shrimps will leave the bottom and obtain their maximum response and will then gradually fall back to the bottom when pulse fatigue overtakes their reactions. The time of maximum response in this graph is about 3s. This number should, however, be interpreted with caution since already after Is, almost the same number of animals was above 10cm. This was observed many times. The response of shrimps in a homogeneous field (Fig. 10-8 (*)) was quite strong compared to the standard electrode array. The maximum was already reached at only 45V but then quickly dropped at higher amplitudes. Since it was often observed that shrimps had a slight tendency to be attracted by the anodes, a number of other electrode arrays were tested. The two most relevant alternatives are given in Fig. 10-11. For large shrimps during clear light, at 12°C, 65 V and 6Hz, with the standard electrode array (Fig. 10-2), the startle response was 50% (45-54) (Fig. 10-8). The alternative array B gave almost the same response, i.e. 49% (44-53). The alternative array C, on the other hand, resulted in a slightly higher response, i.e. 56% (45-67), although not significantly different from the results with the standard array (p=0.27). Any attraction of shrimps to the anode could thus not be detected from the data.

Fig. 10-11 - A selection of alternative electrode arrays tested (anode = black, cathode = gray). A: standard array, B: alternative array with perpendicular anodes and cathodes, C: alternative array with parallel electrodes.

10.4.1.5 Observation tests - flatfish and other demersal species Plaice (Pleuronectes platessa) After the settling period. Plaice dug into the sand until only the head was partly visible. After the start of the pulses, the fish body started vibrating gently to the rhythm of the pulses without leaving the buried positions. This lasted for the full 15s and after the pulses were switched off, the animal remained where it was.

Sole (Solea solea) Sole showed comparable reactions to Plaice in most of the tests. In about one quarter of the cases, however, the fish swam up from its buried position and started swimming around intensely. Dab (Limanda limanda) Like the other flatfish. Dab was buried in the sand before the start of the tests. When the pulses started, however, this fish showed a vigorous reaction. Sometimes it swam along the bottom and sometimes it swam straight upwards to the surface (50cm in this arrangement). Results 10-141 After a few seconds in the less strong electric field at the surface, Dab tried to return to the bottom where it re-entered the stronger electric field and swam straight up. After the pulses stopped, the fish returned to the bottom. Turbot (Psetta maxima) Turbot showed the same reaction as Plaice. Ray (Raya spp.) Ray showed the same reaction as Plaice. Armed Bullhead (Myoxocephalus scorpius) At rest, these animals lay on the bottom without digging in the sand. Under the influence of the pulses, the fish showed light vibrations, but did not move their position. Dragonet (Callionymus spp.) While resting, Dragonets were partly dug into the sand. After the pulses started, their bodies showed strong uncontrolled jolts and short displacements over the bottom. They stayed, however, close to the bottom. Pogge (Agonus cataphractus) Like the Armed Bullhead, most Pogges lay on the bottom before the pulses started. Some of the fish, however, kept on swimming around in the tank. When the pulses started, the fish lying on the bottom started moving around slowly while their bodies vibrated at the frequency of the pulses. The fish up in the water column immediately returned to the bottom. Five Bearded Rockling (Ciliata mustela) Before the pulses started, these fish rested on the bottom or swam slowly around over the sand. During the electric stimulation, the Five Bearded Rockling agitatedly swam around close to the bottom. Shortly after the pulses were switched off, these fish soon returned to the behaviour they showed before stimulation. Swimming Crab (Liocarcinus holsatus) While resting, the crabs were dug into the sand. When stimulated by the pulses, they started walking around agitatedly on the bottom. After the test, the animals quickly sought protection in the sand. Shore Crab (Carcinus maenas) The behaviour of the Shore Crab was comparable to that of the Swimming Crab. Other A number of other, less mobile species like Hermit Crab (Bernhardus pagurus), Common Starfish (Asterias rubens), Spisula subtruncata and Brittle Star (Ophiura spp.), were also tested. Usually, no visible change in behaviour was observed.

10.4.1.6 Survival experiments For Brown Shrimps, survival tests comprising a total of about 700 animals were carried out. Different values of the co-variables (pulse amplitude, frequency, electrode array) were used. For the combination of co-variables to be used at sea (65 V, 6Hz, threadlike electrodes), the survival was 100.2% (92.6-107.8). For the other tests, the survival was also close to 100%, often slightly higher, due to higher mortality in the control group. Survival tests were also carried out for the following species: Pandalus montagui, Plaice, Sole, Dab, ray, Turbot, Cod, Armed Bullhead, Dragonet, Pogge, Five Bearded Rockling, Electric fishing 10-142 gobies (Pomatoschistus spp.), Swimming Crab, Shore Crab, Hermit Crab and Spisula subtruncata. For these species, lower numbers were tested (between 15 and 30 per species). The survival each time was 100%, indicating a very low or absent impact of the electric pulses. No differences in behaviour and feeding activity were observed between the experimental and control groups.

10.4.2 Phase 2 - sea trials The sea trips carried out and the configurations tested are given in Table 10-2. The catch results for the different series of sea trials are given in Tables 10-3, 10-4 and 10-5. For each series and each species, the table presents four numbers, i.e. 1) the total numbers of animals caught in the standard net, 2) the percentage animals caught in the experimental net relative to the standard net indicating the catch loss (or increase) due to gear modifications, 3) the 95% confidence limits for this difference and 4) the p-value of the t-test indicating whether the catch difference was significant or not. A graphic representation of the length frequency distributions of the shrimp catches in both nets and the percentages shrimps caught in the experimental net relative to the standard net by length class is given in Fig. 10-12.

Table 10-2 - Overview of the sea trips and the configurations tested with the electro-trawl. Series Vessel Pulses No. of Electrode Groundrope Connectors Small meshes generators array in top panel

1 Belgica no 2 A/B shortened short no 2 Belgica yes 2 A shortened short no 3 Belgica yes 2 B shortened short no

4 0.700 no - A/B standard standard no

5 0.700 no - A/B shortened long no

6 0.700 no - A/B shortened short no 7 0.700 yes 1 A standard standard no 8 0.700 yes 2 A standard standard no 9 0.700 yes 2 A standard standard yes 10 0.700 yes 2 A shortened long no 11 0.700 yes 2 A shortened short no 12 0.700 yes 2 A shortened short yes 13 0.700 yes 1 B standard standard no 14 0.700 yes 2 B standard standard yes 15 0.700 yes 2 B shortened short no

During the first three series of experiments on RV Belgica, the shrimp trawl with shortened groundrope was used, i.e.: • Series 1: without pulses The loss of shrimps was 90% for small and 76% for large shrimps. This demonstrates that the raised groundrope creates an escape route through which most of the shrimps escape when no pulses are used. From Fig. 10-12 (Series 1) a clear length effect was observed, i.e. an increasing escape rate with decreasing length. Results 10-143

• Series 2: with pulses and electrode array A By switching the pulse generator on in the same configuration as in Series 1, the loss is reduced. This is especially the case for the larger shrimps with 31% shrimp loss relative to the standard net. This was significantly more (p-value = 0.019) than in Series 1. For the smaller shrimps, the catch loss was 76%, which was not significantly higher than in Series 1. Fig. 10-12 (Series 2) again shows a length effect, indicating that for shrimps larger than 60mm there was no catch loss and even a slight catch increase. • Series 3: with pulses and array B By using electrode array B, the catch losses were further decreased. The catches for small and large shrimps were 50% and 76% respectively relative to the standard net, both significantly higher compared to Series 1 (respective p-values 0.008 and 0.004). Again the losses of larger animals were smaller compared to the smaller shrimps. The first three series of trials on the commercial vessel were used to identify the effect of fishing gear alterations without electric pulses although the electrodes were rigged into the net opening for each series. • Series 4: standard shrimp trawl During these trials, the experimental net was identical to the standard net, with the exception of the presence of electrodes in the net opening (not generating pulses). With this Series, the tickler effect of the electrodes could be evaluated. The shrimp catches in the standard and the experimental net were comparable. For the smaller shrimps a small reduction was observed, although not significant. For most of the different fish and invertebrate species, a catch increase was observed which was, however, never significant. • Series 5: shortened groundrope with long connectors The fishing gear in this set of trials was identical to the previous Series with the exception of a raised groundrope with long connectors. The losses of shrimps were substantial and significant for small as well as large shrimp. No clear length effect was observed. The losses of fish and invertebrates were also substantial. It is clear that the raised groundrope creates a good escape route for all animal species caught by shrimp trawlers. • Series 6: shortened groundrope with short connectors. The only difference between this Series and the previous one was the short connectors, resulting in a more stable groundrope sticking close to the bobbins. The catch reduction for shrimps was significant for both small and large animals, although the losses were significantly less (p-value = 0.004) compared to Series 5. For all fish and invertebrate species, large catch reductions were observed but for many species the reductions were not significant.

The following six series of trials were carried out to evaluate electrode array A, i.e.: • Series 7: standard shrimp trawl, with pulses generated by 1 pulse generator, It was originally planned to carry out trials with two generators, but due to a breakdown, the sea trip was continued with only one functioning generator. A notable loss of shrimp catch was observed. The small shrimps only attained about 50% and the large only 80% compared to the standard net. The length effect was quite clear, with high losses of small shrimps and for the largest animals (> 70mm) a slight tendency of a catch increase. Since the trials were carried out with a standard net, the catch loss must have been caused by the electric pulses. Electric fishing 10-144 For the commercial fish species, the only significant difference recorded was the catch reduction for large Whiting. For the other species, a significant catch increase was found for non-mobile species like the Common Starfish (Asterias rubens) and Ensis directus. For gobies (Pomatoschistus spp.) a significant catch reduction was obtained. For the other species, the results were variable. • Series 8: the same as the previous one but with 2 pulse generators operating in parallel Also in this Series, important catch reductions were obtained for Brown Shrimps, although less pronounced compared to the previous Series. A concurrent length effect was observed. For commercial fish, the results were again rather inconclusive. The catch reduction for undersized Whiting was significant. For large Sole, the catch was higher in the electro-net although the significance could not be tested since this species was only caught in one haul. The results for the non-commercial fish and invertebrates were quite species dependent. The significant catch reduction for Pogge (Agonus cataphractus) and gobies may point at a repellent effect of the pulses for these animals. The significant increase in catch of the Common Swimming Crab (Liocarcinus holsatus) may, however, point at a stimulating effect of the pulses for crabs. • Series 9: the same as the previous one but with small meshes in the top panel Because it was not expected that the shrimp catches would decrease with electric pulses (Series 7 and 8), it was suspected that shrimp made use of another escape route. It might be that due to the extra stimulation, shrimp would tail flip up to the top panel where the larger meshes (28mm) would allow escape. Therefore, the front part of the top panel was shielded with a small mesh panel (10mm mesh opening). The catch of undersized shrimps was almost identical in both nets. The catch of large shrimps was 12% higher in the electro-net, although the difference was not significant. A clear length effect was not observed. The catch differences for the other species were again species dependent and to a great extent the same as in Series 8. • Series 10: shortened groundrope with long connectors, no small meshes in the top and pulses generated by 2 generators The configuration used in this Series was identical to Series 5, but with pulses. High shrimp losses were observed, although not as high as in Series 5. The catch losses were significantly less compared to Series 5 (p-value 0.007 and 0.001 for small and large shrimps respectively). Strong reductions in fish and invertebrate catches were also observed. • Series 11: the same as the previous one but with short connectors The catches of small and large shrimps were, respectively, 81% and 92% relative to the standard net. The catch losses were somewhat higher for the small shrimps but a length effect was not pronounced. For almost all fish and invertebrate species, important catch reductions were observed. The results for this Series are less reliable since only one valid haul was carried out. • Series 12: the same as the previous one but with small meshes in the top panel A catch loss of 12% for large shrimps was observed, although not significant. For small shrimps, a significantly higher catch was obtained in the electro-net (116% relative to the standard). The length effect observed in the trials without small meshes in the top was absent and had reversed to some extent. The catch differences for commercial fish were quite Results 10-145 variable and only for undersized Plaice was a significant reduction noted. For most of the non-commercial fish and invertebrate species catch reductions were seen. Only for the Common Swimming Crab (Liocarcinus holsatus) and Callionymus spp. was the catch higher in the electro-net. The last three series of trials were carried out to evaluate electrode array B. The number of different configurations tried out was not as complete as for electrode array A due to time restrictions. • Series 13: standard shrimp trawl, with pulses generated by 1 pulse generator A significant catch reduction was observed for small shrimps and as for the other trials without small meshes in the top, the length frequency distribution showed a clear length effect. For large shrimps the catch reduction was small and non-significant. The catch differences for the other species were quite variable and species dependent. • Series 14: the same as the previous one but with small meshes in the top panel and 2 generators The shrimp catches in the electro-net were not significantly different from those in the standard net. The length effect seen in the trials without small meshes in the top panel was absent. For most of the commercial fish species, a catch reduction was seen, although only significant for undersized Whiting. For Sole and Dab the catches were higher in the electro- net but the differences were not significant. The results for non-commercial fish and invertebrates were quite variable and species dependent. • Series 15: shortened groundrope with short connectors, no small meshes in the top and pulses generated by 2 generators, A significant catch reduction was seen for small shrimps, but for large shrimps the catches remained the same. For almost all commercial fish the catches were lower in the electro-net, but only significant for undersized Plaice. With the exception ofCallionymus spp. all non-commercial fish and invertebrate species were caught in lower numbers in the electro-net. Electric fishing 10-146 Table 10-3: Results of the sea trials for Brown Shrimp - total numbers in standard net (*); percentage animals in experimental relative to the standard net (**); 95% confidence limits for the percentage difference (***); p-value for the difference in catch between the standard net and the experimental net (****). Series: 1 2 3 4 5 6 1 8

Brown shrimp * 18556 18556 18556 189513 40596 172535 23442 499735 < MLS ** 10% 24% 50% 94% 31% 54% 48% 53% ***(.4 24) (-5_ 53) (31 69) (85 104) (15 48) (47 60) (2 93) (42 63) **** 0.000 0.002 0.000 0.467 0.003 0.002 0.039 0.001 Brown shrimp 15503 15503 15503 47526 47711 38386 29934 143947 > MLS 24% 69% 76% 100% 31% 57% 79% 86% (-8_55) (39 99) (56 96) (89 110) (14 48) (46 67) (70 87) (72 100) 0.004 0.047 0.025 0.410 0.003 0.003 0.008 0.056

Series: 9 10 11 12 13 14 15

Brown shrimp 25445 60978 48825 129883 30534 211612 58875 < MLS 99% 52% 81% 116% 74% 110% 72% (70 128) (46 58) (-) (102 130) (72 75) (80 139) (54 91) 0.673 0.001 0.032 0.000 0.355 0.022 Brown shrimp 25380 61208 19888 56038 40956 55798 29314 ox 0 0 0 0 > MLS 112% 73% 92% 91% 98% 96% (69 155) (59 88) (-) (74 103) (74 108) (91 105) (82 110) 0.330 0.015 0.096 0.303 0.762 0.323 Results 10-147

Table 10-4: Results of the sea trials for commercial fish species - total numbers in standard net; percentage animals in experimental relative to the standard net; 95% confidence limits for the percentage difference; p-value for the difference in catch between the standard net and the experimental net. Series: 4 5 6 7 8 9 10 11 12 13 14 15 Whiting 154 584 120 863 799 588 862 198 962 2834 210 329 < MLS 104% 69% 66% 93% 67% 86% 65% 85% 85% 90% 58% 90% (0 208) (49 90) (25 107) (42 144) (47 87) (77 94) (32 97) (-) (46 124) (1 180) (46 70) (-25 205) 0.985 0.023 0.165 0.614 0.007 0.051 0.043 0.888 0.42 0.021 0.903 11 22 20 v v ë “I S S S 100% 32% 144% (-136 336) % (28 36) (57 231) 0.678 0.009 0.206 Poor cod 2368 8 1081 56 615 2972 70 3567 36 99 1030 < MLS 130% 80% 102% 111% 97% 98% 103% 104% 207% 93% 92% (62 198) (-193 353) (62 143) (2 221) (83 112) (63 132) (-) (89 119) (-63 478) (83 103) (66 118) 0.291 0.717 0.45 0.292 0.61 0.425 0.421 0.301 0.881 0.988 Sole 259 138 110 211 394 271 92 14 977 392 106 84 < MLS 104% 41% 47% 103% 76% 76% 60% 81% 87% 74% 88% 78% (15 193) (37 46) (-41 135) (92 114) (55 98) (-10 162) (41 80) (-) (64 110) (31 116) (26 150) (-3 160) 0.537 0.001 ___ 0.253 . 0.51 0.11 0.213 0.012 ______0_.842 0.096 0.903 0.363 Sole 6 17 8 12 2 . 24 2 9 > MLS 286% F 112% 375% 208% 0% i 142% 250% 100% (-) (41 182) (-) (-) (-) (48 235) (-) (-) 0.238 ; 0.547 0.315 \ 0.421 0.5 Plaice 214 62 124 35 236 140 97 22 309 1004 66 23 < MLS 155% 54% 66% 92% 94% 89% 40% 65% 81% 70% 74% 41% (-2 312) (8 100) (52 79) (37 147) (57 131) (-33 210) (38 41) (-) (63 99) (57 83) (52 95) (4_78) 0.333 0.218 0.042 0.438 0.236 0.675 0.003 0.009 0.013 0.739 0.045 Plaice 12! 29 26 2: 32 > MLS 25%: 153% 50%: 173% CO ro O

(-200 250)' (13 294) IO (-): (131 216) 0.1 r 0.244 0.023 Dab 321 131 459 15 530 9 79 13 84 1584 255 < MLS 103% 49% 100% 67% 61% 282% 61% 38% 157% 65% 135% (51 155) (-12 109) (-223 424) (-16 149) (32 89) (-54 618) (59 63) (-) (78 235) (57 73) (106 163) 0.778 0.079 0.544 0.225 0.704 0.175 0.003 0.867 0.005 0.176 Dab 27 6: 10 18 32 547 > MLS 17% 0%' 130% 29% 79% 83% (1_33) (-): (-) (21 38) (-197 356) (31 135) 0.005 0.691 0.184 0.721 0.186 Flounder 7 26: . 32 27: 36 179: > MLS 57% 29%: i 120% 30%: 69% 37%: (-228 341) (-26 84): : (-59 299) (io so): (-37 176) (10 65): 0.288 0.042: 0.678 0.543: 0.824 0.01 : Electric fishing 10-148

Table 10-5: Results of the sea trials for invertebrates and non-comm. fish species - total numbers in standard net; percentage animals in experimental relative to the standard net; 95% CL for the percentage difference; p-value for the difference in catch between the standard net and the experimental net. Series: 4 5 6 7 8 9 10 11 12 13 14 15 Anemones 0 16 10; 0% 58% 420%: o (-) (-): Q.687.. . Pagurus 43 37 449 77 26 123 50 105 79 65 43 spp. 42% 65% 181% 105% 115% 1% 30% 38% 237% 113% 50% (-63 148) (-256 385) (77 286) (-) (-) (1 2) (-) (-) (136 338) (99 128) (-4 104) ...... Q.432...... 0 .333...... 0.0.70...... Q.24S ...... J3.Q25...... P...... _Q,4Q5...... 0.023...... Q.249. . . ___ _0,QZ4 Liocarcinus 18013 3660 10937 1202 23391 10456 8477 6493 20210 3365 8756 10724 holsatus 116% 33% 87% 102% 127% 180% 33% 56% 160% 47% 110% 59% (90 142) (37 38) (28 147) (82 123) (120 134) (99 261) (5 60) (-) (30 290) (-47 141) (73 148) (56 63) 0.081 0 0.464 0.269 0.023 0.066 0.009 0.183 0.135 0.531 0 Spisula 47 56 14 3735 335 22 56! ■ 15 subtruncata 167% 35% 81% 142% 101% 69% 17%; 0% (-565 900) (-11 81) (-) (94 190) (66 135) (-) (-); (-) 0.458 0.025 0.083 0.412 0.124; 1 Abra spp. + | 39 131 995 14 0 214: 1091 Angulus spp. 21% 39% 91% 105% 12% 77% i 161%; (-) (-103 181) : (40 141) (-) (-) (-); (-): 0.165 : 0.996 0.344' 0 .8 6 3 ^ „ Ensis 465 10 87 4082 202 219 1151 710 79 162 102% 53% 69% 131% 80% 113% 44% 112% 38% 51% (67 136) (-) (-361 499) (126 135) (12 149) (61 165) (3 86) (82 143) (-) (18 83) 0.456 0.728 0.031 0.389 0.186 0.052 0.22 0.029 Sepiola 84 0 70 106 245I 125 atlantica 98% 0% 25% 117% 63%; 34% (-911 1107) (-) (4 46) (77 156) (5 122); (-25 94) ...... QJ5P2 ...... P.Q14...... P.ZQS...... _Q,Q4i: ....0 ,0 4 2 Allotheutis o; 40 w 69 21 subulata o%; 0% E 128% 0% (-): (-) (-) (-) Asterias 14 83 2 59 26 91 13; 451 121% 36% 50% 128% 123% ; 25% : 61%; 1% (-) (-18 89) (-) (126 130) (-) (-2_52) (-) 0.046 0.014 0.007 Ophiura spp. 4788 2137 675 32035 9937 1063 4141 90 4805 1259 2489 1472 99% 32% 86% 96% 96% 106% 12% 71% 79% 108% 119% 79% (14 183) (25 38) (-22 194) (74 117) (72 120) (43 169) (2 23) (-) (6 152) (73 142) (116 121) (37 121) 0.728 0.001 0.715 0.429 0.396 0.725 0.001 0.724 0.435 0.038 0.197 Results 10-149

Table 10-5 (continued): Results of the sea trials for invertebrates and non-commercial fish species - total numbers in standard net; percentage animals in experimental relative to the standard net; 95% CL for the percentage difference; p-value for the difference in catch between the standard net and the experimental net. Series: 4 5 6 7 8 9 10 11 12 13 14 15 Ciliata mustela 17 s; 63% 66 %: (10 116) (-)! 0.093 0.3292 Eutrigla 0 92 0% 0% (-) (-) Trigla lucerna 0 50! 141 0% o%: 7% (-) (o_o); (-4 18) 0.003 Myoxocephalus 8 2 scorpius 0% 133% (-) (-) Agonus 2623 58 534 22 3138 4603 10 136 1938 5 1591 493 cataphractus 150% 41% 62% 9% 68% 82% 54% 51% 72% 0% 128% 87% (0 299) (7_75) (-39 164) (-30 48) (44 91) (68 97) (33 75) (-) (59 84) (-) (-) (-18 192) ...... 0,214 ...... Q0.1S ...... Q.3J 2 ...... 0.98.2 ...... OD.4.1...... Q.QÇ2.. ____0.105 ...... Q.00S ______Q.762 Liparis liparis 27 2 10 0% 0% 0% (-) (-) (-) Trachinus vipera 261 0 81 68 395I 569 666 60% 0% 13% 44% 45%; 51% 57% (-238 359) (-) (-) (-) (-122 212); (18 84) (47 67) ...... ÍUQ.I ...... 0.405...... Q.633: ...... Q.D28 ...... Q.QQ3 Callionymus spp. 105 181 179 287 7 97 186 204 19 106% 46% 99% 133% 505% 46% 205% 242% 117% o (-1 94) (48 149) (90 177) (176 834) (0 93) (137 274) (153 330) (-) 0.521 0.05 0.914 0.764 0.034 0.667 0.48 0.207 smelt 66 203 55 21 31 6 2% 0% 0% 0% 69% 0% (-5_8) (-) (-) (-) (-) (-) 0.015 0.242 Pomatoschistus 13366 885 6259 242 35689 4381 456 4504 3330 950 11223 1711 spp. 105% 38% 74% 38% 60% 52% 38% 54% 67% 61% 65% 28% (77 133) (17 60) (53 96) (28 47) (40 79) (8 95) (-36 112) (-) (41 94) (59 63) (58 73) (22 33) 0.293 0.008 0.074 0.001 0.001 0.059 0.068 0.046 0.001 0.005 0 Electric fishing 10-150

Series 1 Series 2

2500 100% 2500 150% ▲ 2000 80% 2000 120% I £2„ £2 " 1500 60% " 1500 Aa a a 90% Q. \ A 1000 40% I 1000 60% " 500 .aa^ X a 20% Z 500 'A 30% s? ^ \ 0 0% 0 0% 0 50 100 50 100 Lengthclass (mm) Lengthclass (mm)

Series 3 Series 4

25% 36000 150%

100% " 24000 100% 2 1500

12000

50 100 0 50 100 Lengthclass (mm) Lengthclass (cm)

Series 5 Series 6

100% 32000 100%

in 6000 24000

16000

0 50 100 0 50 100 Lengthclass (mm) Lengthclass (cm)

Series 7 Series 8

200% 100000 125% •M m 80000 100% 2 150% ë £2 ó . ^ 60000 I 2500 100% g 1 40000 50% ^ 0 s Z 20000 0% 0 50 10 0 50 100 Lengthclass (cm) Lengthclass (cm)

Fig. 10-12 - The length frequency distributions for Brown Shrimp caught in the standard and the experimental net and the percentage in the experimental net catch relative to the standard net catch. Results 10-151

Series 9 Series 10

12000 150% ■4-» 6000 300% (1) 10000 125% o g A C 8000 100% 5 4000 200% Q. E 0) 6000 ^ / * \ A z 2000 100% C 50% .E £ 0 0% 0 50 100 0 50 100 Lengthclass (cm) Lengthclass (cm)

Series 11 Series 12

175% 25000 200% 150% 'S 125% c S2 18750 150% c 100% g 12500 100% * 50% -

50 100 50 100 Lengthclass (cm) Lengthclass (cm)

Series 13 Series 14 7500 150% 30000 200%

5000 100% ^ S? 22500 150% 15000 100% = 2500 50% .E 50% ” 0 0% 0 50 100 0 50 100 Lengthclass (cm) Lengthclass (cm)

Series 15

8000 200% ▲ : standard net r/s. \ a ^ A 0) S2 6000 150% c 0) Q. '*V\4éA a a X : electro-net I 4000 a V y a 100% 0) *A A / f VàâV \ c z 2000 A V 50% ” A : % in electro-net 0 0% 20 40 60 80 100 Lengthclass (cm)

Fig. 10-12 (continued): The length frequency distributions for Brown Shrimp caught in the standard and the experimental net and the % in the experimental net catch relative to the standard net catch. Electric fishing 10-152 10.5 Discussion

10.5.1 Phase 1 - laboratory experiments The equipment had a fixed pulse and none of the characteristics were adjustable at the control unit. It was therefore essential to use the newly developed generator LPG with adjustable amplitude and frequency for the observation tests in the laboratory. Within the financial means of the project, it was, however, impossible to have the pulse shape and pulse duration also adjustable. These two characteristics were considered to be of less importance. The pulse shapes of the four generators available were different, but no difference in response was observed between the generators at equal amplitude and frequency. Observation tests on the effects of electric pulses on the behaviour of shrimps have been done several years ago (Kessler, 1965), but not for Brown Shrimps. Despite the difference in species - Kessler observed Penaeus duorarum -the results were comparable. The minimum pulse for reaction as well as the effect of shrimp size, orientation and water temperature were remarkably similar. Several electrode configurations were tested in the laboratory. Some of these performed badly in stimulating shrimps. Others performed as well as the electrode array given in Fig. 10-2. The latter, however, had the advantage that it was easy to rig into a towed fishing gear. It had also proven to be efficient in many experiments (Van Marlen, 1997a). The drop in amplitude when the generators were applied at full power, had consequences for effectiveness. De Groot and Boonstra (1974) found that a pulse amplitude below 10V was too low for shrimps and that the amplitude should not be much below 65 V (Boonstra, 1976). The generator MJX-50 emitted a very low pulse amplitude of less than 30V when used at full power. Generator Tongfa-98 even stopped operating. Even for the most powerful generator LWY the emitted pulse had an amplitude below 50V. Whether these amplitudes would be sufficient to invoke the desired response for shrimps became clear from the observation experiments. With threadlike electrodes, the electric field is not homogeneous (Fig. 10-5). The closer to the bottom and to the electrodes, the higher is the head-tail voltage. Therefore, the strength of the stimulus will depend upon the horizontal and vertical position of the shrimp in the field. On top of this, the orientation of the animals with respect to the electrodes will also influence the head-tail voltage it will experience; minimum if the body lies parallel and maximum if perpendicular to the electrodes. This means that with a pulse amplitude of 120V, a shrimp of 3cm lying perpendicular to and with its head against an electrode, experiences a head-tail voltage of 20V (Fig. 10-6). The same shrimp, lying in the middle between the electrodes, experiences a head-tail voltage of about 6V or IV for a perpendicular or parallel position in relation to the electrodes (the carapace width of a shrimp is about 1/6 of its length, without taking antennae and legs into account). A consequence for the sea trials is that animals in the electric field of the fishing gear can experience a wide range of strength of the pulse stimulus. It is important that a pulse amplitude is chosen for all animals in the gear path to undergo a stimulus strong enough to induce the desired reaction. In that case, many animals will undergo a pulse amplitude that is much higher than the minimum needed. Based on the results of the minimum pulse measurement (10.4.1.3), 240mV/cm should be sufficient to invoke a startle response from shrimp in each size class and orientation on the bottom level. Based on the measurement of the electric field (Fig. 10-6), a generator producing a pulse of 120V gives a head-tail field of 2V/cm in the centre between the electrodes. If the generator would, however, only produce a 40V pulse, this strength would drop to about 660mV/cm. This still seems to be high enough to startle all shrimps. It should Discussion 10-153 be borne in mind, though, that the reaction at minimum pulse level was only a noticeable body movement and not a contraction of the body producing a clear movement of the animal. For a clear movement, a higher pulse amplitude is necessary. The exact amplitude could not be determined precisely because, as soon as the pulse starts, a shrimp will start “swimming” in a random direction and the animal will depart from its original fixed position and orientation. It is roughly estimated that at least 400 mV/cm is needed for shrimp clearly to change its position and orientation. This was confirmed by the observation tests (10.4.1.4) which show that below a pulse amplitude of 30V (500mV/cm, middle electrodes, bottom level) the response of the shrimps was low with incomplete tail flips. The observation tests were carried out in a small tank, tiny compared to the conditions at sea, where there is no space confinement, often strong currents, a fishing gear passing over etc. Therefore, the reactions observed in the laboratory may well deviate from fishing conditions. The results should thus mainly be interpreted as a comparison and not as a prediction in absolute numbers of the situation at sea. The orientation and mechanism of escape swimming has been well described in Neil and Ansell (1995) who indicated the tail-flip as the main propulsion when trying to escape. The observation tests in the present project showed that, during the swimming phase in the electric field, the shrimps jumped in random directions, sometimes rather high, but often also close to the bottom. Research carried out by Amott et al. (1998) with Brown Shrimps exposed to predators demonstrated that the shrimps tried to escape from the predators with lateral movements in random directions in 84% of the cases; the rest escaped upwards. So escape swimming may incorporate unpredictability as a predator-avoidance mechanism. The randomness in choice of direction seems also to take place in an electric field. Not all shrimps moved upwards after being startled by the pulses. A consequence for the sea trials is that it will not be possible to bring 100% of the shrimps present in the trawl path above the groundrope since many will move along the seabed. Consequent to the behaviour of the shrimps, an electro-shrimp trawl will never have a 100% efficiency for catching shrimps. The observation tests also showed that shrimps were slightly attracted to the anodes. Tests with the cathodes on the bottom and the anodes higher up gave a slightly better response, although not significantly different. This array could not prevent a certain percentage of the shrimps from tail-flipping along the bottom. In theory, larger animals show stronger reactions in an electric field because they experience a larger potential difference. Maksimov et al. (1987) found that the product of fish length * voltage was constant, indeed indicating that larger fish need a lower voltage to show an equal response. The results of the laboratory shrimp tests indicate that larger shrimps have a somewhat stronger response to the electric field, although the results were only significant with part of the tests. Based on the results of Jeffery and Revill (2002), the higher vertical movement of large Brown Shrimps when being startled, is inherent to the behaviour of the species and probably not only caused by the electric field. The maximum startle response for shrimps was almost the same for the three pulse frequencies tested. The frequency had, however, an influence on the pulse amplitude necessary to obtain this response. With a higher frequency, a lower amplitude was needed. This has consequences for the energy consumption and can be important for a generator fed by batteries and should be taken into account for further development of the equipment. Important for the fishing power of an electro-trawl is the time that elapses between the moment of first impact with the electric field and the moment the animal makes contact with the groundrope. Based on an average towing speed of a shrimp trawler of 2.5 knots and with a standard shrimp trawl, this time is less than 2.5s. This means that with the pulse generators Electric fishing 10-154 available this time is too short, as at least 3 s is needed to obtain the maximum response. The expected loss of shrimps due to the short response time available is expected to be low because of the wide response peak observed (Fig. 10-9). A longer exposure time is however necessary to avoid this loss, so a shrimp trawl with a longer distance between the beam of the trawl and the groundrope should be developed. Typical behaviour of shrimps is that these animals dig into the sand during daytime and in clear water (Hagerman, 1970). If visibility decreases (by night or in turbid water), they leave their refuge and become active. This is why the shrimp fishery only obtains good shrimp catches by night or in turbid water. This typical behaviour protects the animals against predators. According to Hagerman, water temperature (season) has no effect on this behaviour. From the observation tests it was also clear that the response was better when the light intensity was lower. This was confirmed by Jeffery and Revill (2002) who observed that Brown Shrimps showed a more pronounced vertical response to a sampling trawl in dark or dusk conditions. In daylight, the shrimps tail-flipped along the bottom and only moved upwards when an obstacle forced them to do so. It can therefore be expected that electro fishing will be more efficient in dark conditions. It is, however, also likely that the relative catch increase compared to a standard trawl, when applying electric pulses, would be higher in light conditions. In the present project, catch comparison was not carried out in light conditions because in the experimental design it was decided to follow commercial practice by fishing in turbid water or dark conditions. Further trials in clear water were not possible within the financial and time constraints of the project. New experiments in clear water may be advisable, although consequences relating to increased fishing effort if day fishing would become profitable should be taken into account. The observation tests also showed that besides shrimps, Sole and Dab none of the other tested animals left the seafloor after electric stimulation. For the purpose of the project, this was quite important since based on that behaviour, catch separation by the groundrope while fishing seems to be possible. If a fishing gear can be made operational with a groundrope fishing at 10cm above the sea floor, animal species like Plaice, armed bullhead, dragonet, pogge, rockling, starfish, shellfish, crabs and to a certain extent also Sole could escape underneath the net. The critical point, though, is that electric pulses should be the only stimulation. Very little is known about the stimulating effect of the sound of the towed fishing gear, sand clouds stirred up by the net and vibrations that may also aid the catching of bottom dwelling species. Together with a possible tickler effect of the electrodes, these may reduce the selective potential of electric fishing.

10.5.2 Phase 2 - sea trials A disadvantage of the generators MJX-50 and LWY for application at sea, was the cable for the connection between the power supply and control unit on board and the generator rigged on the fishing gear. This made shooting and hauling the gear difficult. The high electric voltage in the cable was also a safety risk for the crew. An advantage of the cable connection was the ability to check the functioning of the generator while fishing. The generator Tongfa- 98, fed by batteries, solved the problem of a cable connection, but had the disadvantage of not being sufficiently powerful to feed the full electrode array.

10.5.2.1 Brown shrimp From the preliminary experiments on RV Belgica (Series 1-3), it was clear that the pulses had an effect on the shrimp catches. A catch loss for large shrimps of 76% was reduced to a loss of 31% and 24% for electrode arrays A and B respectively. Other trials demonstrated the Discussion 10-155 same effect of the pulses, e.g. comparing Series 6 (no pulses) with a catch loss of 43% and Series 12 (with pulses) showing a non-significant catch loss of 12%. Apparently, the stimulation of the electric pulses was strong enough to bring the majority of the shrimps within reach of the net. In a number of experiments, however, a clear length effect was observed when pulses were used, i.e. the catch reduction of small shrimps was higher compared to large shrimps. At first, it was thought that the larger shrimps showed a stronger startle response compared to the small ones. The observation tests in the laboratory, though, only indicated a rather small difference (10% or less) in startle response. When a small mesh panel was rigged into the top panel, the length effect was not observed. This indicated that some of the shrimps were tail-flipping as high as the top panel and, in the absence of small meshes, could escape through the top panel meshes. Clear length selectivity occurred in that area, with more small shrimps escaping than larger ones. De Groot and Boonstra (1974) obtained 39% and 20% higher commercial and undersized shrimp catches with an electro-net (i.e. standard shrimp trawl with pulses as extra stimulation). There seems to be a length effect in the shrimp catches, although not as pronounced as in this study. Shrimp nets then were usually made of smaller netmeshes. Due to the lower selectivity, it is logical that the length effect is less clear. The relative catch increase for commercial shrimps of 39% is markedly higher compared to the 12% observed in Series 9 of this study. It may be so that electric pulses enhance the escapement of shrimps through the meshes of the net body. Smaller meshes will release less shrimps, leading to a stronger relative catch increase, as observed by De Groot and Boonstra. It may also be so that the difference in environmental conditions during the sea trials caused this difference in relative catch increase but due to a lack of data, no decisive answer can be given. Similar catch increases were observed by Delanghe et al. (1988) fishing with an electrified shrimp beam trawl. A possible disadvantage of the electro-net as used in this study was the tickler effect of the electrodes gliding over the seafloor. The impact of this tickler effect on the catches when fishing with raised groundrope is difficult to estimate with the available data set. An alternative electrode array with electrodes not touching the seafloor might further decrease the unwanted by-catches. Future experiments should pay attention to this issue.

10.5.2.2 Commercial fish species For Whiting catch, reductions were obtained for most of the Series, although only significant in four trials. Escapes were recorded with the raised groundrope, but also when the standard net was fished with pulses. Apparently, the electric field had a scaring effect on this species inducing their escape from the net. Data on by-catches in electro-trawls in the literature are quite scarce, but for Whiting and also for Cod the same effect was reported by Vanden Broucke and Van Hee (1977). For Poor Cod, on the other hand, no significant reductions were recorded, indicating that this species was not scared out of the net and did not even escape underneath the raised groundrope. For Sole, the results were quite variable and inconclusive. Undersized Plaice, on the other hand, clearly used the escape route underneath the raised groundrope. There was even a slight indication of catch reduction due to the electric field when fishing with the standard groundrope. For undersized Dab, catch reduction was obtained when fishing with electric pulses, both with the raised and the standard groundrope. This might indicate that these fishes did not use the escape route underneath the net. Only when small meshes were rigged in the top panel of the net, were the catches not lower in the electro net. This indicates that this species had a similar behaviour to shrimps, i.e. an upwards movement in the electric Electric fishing 10-156 field. The fishes that were small enough could escape through the larger net meshes in the top panel. This is backed up by the observation tests that showed that Dab swam straight up after the electric field was switched on.

10.5.2.3 Non-commercial fish and invertebrate species Heavier animals like Hermit Crabs clearly benefited from the escape route underneath the net reaching catch reductions from 35% up to almost 100%. Similar good catch reductions were obtained for shellfish species like Spisula subtruncata, Abra spp.,Angulus spp. and Ensis directus. Also for the other invertebrate species, strong catch reductions were observed with the electro-net. Often, these were not significant due to the rather low numbers of observations per Series. For Liocarcinus holsatus the same behaviour was recorded with one exception. In Series 12, despite the raised groundrope, higher catches were found in the electro-net. In these hauls, very high numbers of small crabs (about 1cm carapace width) were caught. Apparently, these small animals did not show the same behaviour as the larger animals of the same species that are more usual in the catches. The small crabs did seem to react to the pulses. The results for the non-commercial fish species followed the same line. Species likeAgonus cataphractus and especiallyPomatoschistus spp. also showed catch reductions in the electro- net rigged with a standard groundrope. For these two species, the pulses seemed to have a scaring effect resulting in lower catches.

Comparison with the other selective devices In Section 9 it was already concluded that the sieve net would be the better choice for the Belgian shrimp fishery compared to the sorting grid. Comparing electric fishing with the sieve net is difficult. Electric fishing is a method still further to be developed. Some problems with the rigging have already been identified and it is probable that future research and commercial testing will lead to improved selectivity. The sieve net, on the other hand, has long past the development phase and has been in commercial use for several years. The design as it stands now is close to optimal for commercial application. For the sake of completeness, a comparison is given below and some general conclusions can be drawn. The results should be interpreted with the remarks of the previous paragraph in mind. The principle of electric fishing, i.e. selection of animals before entering the net, has clear advantages compared to the filtering principle of the sieve net. The contact of escapees with the electro-net is minimal since they escape underneath the net, so escapee mortality is likely to be lower. In the previous chapter, it was already demonstrated that the sieve net performed better compared to the sorting grid for the loss of commercial shrimps. For this catch fraction the electro-net (Series 12) outperforms the sieve net (Table 10-6, All CV-Trials). The catch loss was not different when comparing the electro-net with the CV-Spring and CV-Summer trials. So, if the technical problems with the sieve net observed in the CV-Autumn-Winter trials could be resolved or if a seasonal enforcement of the sieve net would be implemented, the catch loss of the sieve and the electro-net would be comparable. It is, however, expected that further optimisation of the electro-net could improve the performance for commercial shrimps. So, it is likely that in the future, electric fishing would be the better choice for the fisherman. The good catch reductions for commercial fish

Electro-net - Series 12 Sieve net CV-Spring 0.75 (0.72) CV-Summer 0.42 (0.42) CV-Autumn-Winter <0.01 (<0.01) All CV-trials 0.02 (0.03)

The future of electric fishing The results of this study indicate that electric fishing has good potential to improve the selectivity of the shrimp beam trawl. The sea trials have been carried out, though, in a limited variation of circumstances like seasons, vessels and fishing grounds. Therefore, the potential for a wider applicability in Brown Shrimp trawling should be further studied. Since the electric pulses have a clear selective effect on different fish and invertebrate species, other applications should be possible as well. Any trawl fisheries with a mixed catch, facing problems with discards, could apply the method to prevent certain species to enter the trawl, on the condition that the target species have a different response compared to the by-catch species. Catch separation within the net is also feasible. An electric field could enhance the effectiveness of selective devices like a separator panel in the net. Another application could be the reduction of catch losses of target species in an outlet of e.g. a sorting grid. A recent example is the problem of losses of penaeid shrimps through the outlet of a sorting grid aiming at the reduction of fish by-catch (Foster and Watson, 2003). The grid appears to be very effective in releasing finfish but also allows shrimps to escape. A small electric field could prevent this. The effectiveness of square mesh windows could be improved with electric pulses. A problem with these devices is that they offer an escape route to fish but the animals sometimes do not use the opportunity because of lack of stimulus. Pulses could offer this stimulus in some cases. This list is not exhaustive. An acute problem with electric fishing is legislation. Council Regulation 850/98 prohibits the use of electricity in any fishery in EU waters. The reason is that electricity can be (mis)used in different ways to catch fish, often quite detrimental. Before the method can be applied, it is inevitable that steps will have to be taken to alter the regulation. A consensus on the merits of the method should therefore be achieved amongst fishermen, managers and scientists before a realistic attempt can be made to do so. DG XlV(fisheries) of the European Commission has already expressed a positive attitude towards electric fishing and stated that legislation can be altered on condition that a substantial report with sufficient arguments can be presented.

10.6 Conclusions Experiments with electric pulses have been carried out in the past in many areas in the world. The main purpose of these experiments was to obtain higher catches of the target species. The idea in the present project was to use electric pulses to improve the selectivity of the shrimp beam trawl, i.e. reduce the discards while maintaining the target species catches. The project was initiated by the observation that in the People’s Republic of China, over 2000 vessels were fishing with electrified nets for penaeid shrimps. In a first phase of the project, detailed observation and survival tests were carried out as a preparation of the sea trials. The Electric fishing 10-158 second phase consisted of sea trials to compare the catches of a standard shrimp trawl with the electrified net in commercial conditions. The following conclusions can be drawn from the observation tests: • Brown shrimps react quite intensely to electric pulses. • The optimal pulse amplitude lies between 40 and 110V for a 50cm electrode spacing. A higher and lower amplitude has a negative effect on the response of shrimps. • The pulse frequency has a limited effect on the maximum response of shrimps but the choice of a certain frequency is important for the energy consumption of the generator. • The startle response for small shrimps is slightly lower compared to that of large animals. • A higher water temperature results in a stronger response. • A low light intensity results in a stronger response. • The maximum response is usually obtained within 4s after the start of the pulses. Fish and other invertebrates, with the exception of Dab and Sole, almost do not react to the pulses or if a reaction is observed, the animals keep close to the bottom. This means that, in principle, species selective fishing for Brown Shrimps should be possible with electric pulses as an alternative stimulation. The best startle response for shrimps was obtained in a homogeneous electric field at low pulse amplitude (Fig. 10-8). If a homogeneous field could be generated in a fishing gear, good shrimp catches would be possible with minimum energy consumption. This involves practical problems, viz. the capacity of the generators and the difficult application of plate electrodes in a fishing gear. Therefore, rope-like electrodes were chosen for the sea trials. The generator type LWY applied at full power with rope-like electrodes generated a pulse with amplitude of 44V at 6Hz. This was considered rather low for a good shrimp response. Therefore, it was decided to use two generators, giving a 64V pulse at 6Hz, for the sea trials. This amplitude and frequency combination do not give the best response, but it was the best possible with the available equipment and only little below the optimum. The average water temperature on the shrimp grounds at the Belgian coast is about 12° C (data provided by the “Coastal Waterways Division” - Ministry of the Flemish Community). With the co-variables set at these values, the expected relative amounts of shrimp that can be caught based on the observation tests would be: • large shrimps, clear conditions: 50% (45-54) • large shrimps, dusk conditions: 70% (63-76) • small shrimps, clear conditions: 47% (39-55) • small shrimps, dusk conditions: 60% (54-65) If a pulse generator was available generating a more powerful pulse, these percentages could increase by roughly 10%. The observation tests indicated that the response for shrimps depended on a number of coincidences. During the same test, some shrimps showed a minimum and others a strong response. The position of the animal in the field and its condition probably played an important role and cannot be controlled at sea. The fact that the tests were carried out in an aquarium with captive animals also adds bias to the results. With this in mind, a projection of these results for the expected catch rates at sea is speculative but should be sufficient to Conclusions 10-159 judge whether electro-fishing might be feasible. To make this projection, it should be taken into account that 9.0% (7.2 - 10.8) of the shrimps in the trawl path escape underneath the groundrope with traditional shrimp fishing gear (see Section 7.3.3). For a shrimp fishery by night, the percentage of large shrimps in the trawl path that could be caught with an electro­ trawl would thus be 79% (70-87) compared to the traditional fishery. The price for a possible environmental benefit in terms of reduction of discards and reduced sea floor disturbance would therefore be a reduction in commercial shrimp catch of roughly 20%. Due to the fact that the price coefficient for Brown Shrimps is about -1, a decrease in shrimp landings should go together with a relatively equal rise in price of the product. So, if applied by the whole fleet, the overall income of the shrimper fleets would stay roughly unchanged. The results of the sea trials matched these projections quite closely. With the selective electro-net with electrode array A, catch losses of large shrimps were 12% and not significant. For small shrimps, a catch increase was even observed. The sea trials also indicated that the raised groundrope created an escape route for most of the species regularly caught in shrimp trawls, and also for Brown Shrimps. The electric field, however, made the shrimps tail-flip high enough to be caught. The response was even strong enough to make the shrimps jump as high as the top panel, leading to catch losses. To prevent this, the insertion of a piece of small mesh netting into the top panel was necessary. It can be concluded that the electro-net with raised groundrope and small meshes in the top panel gave satisfactory results. The losses of commercial shrimp catches were small or even non-existent. Part of the undersized commercial fish catch could escape and especially non­ commercial fish and invertebrates were caught in lower numbers compared to the standard net. Future work should pay attention to the design of an electro-trawl with a larger net opening to fit long electrodes and a new type of bobbin rope with less bottom contact. Bearing this in mind, electric fishing should be a feasible alternative to the standard shrimp trawl and could be an acceptable alternative between the economic interests of the fishermen and the ecological demands of the marine ecosystem. It should, however, be borne in mind that the sea trials in this project only covered a short time range and a narrow range of conditions such as water temperature, currents, degree of activity of the shrimps etc. An extensive range of sea trials on commercial vessels in different conditions should precede commercial application. A future pulse generator should have the following characteristics: • compact and able to fit into the beam of the trawl, • robust and resistant to conditions at sea (shocks, vibrations, water, etc.), • fed by batteries making the cable connection between vessel and gear redundant, • a pulse frequency of 6 Hz, • a pulse amplitude of minimum 60 V, rather 95 V with electrode spacing of 50cm, • a pulse duration of 0.5ms and • a control unit aboard the vessel to control the functioning of the generator via an acoustic link. Since the electric pulses have a clear selective effect on different fish and invertebrate species, other applications should be possible as well. Any trawl fisheries with a mixed catch, facing problems with discards, could apply the method to prevent certain species to enter the trawl, provided that the target and by-catch species have a different response to the pulses. The functioning of selective devices like separator panels, square mesh windows, grids etc. could also be enhanced in certain circumstances. Electric fishing 10-160 The final obstacle on the way to a possible commercial application is Council Regulation 850/98 that prohibits the use of electricity in any fishery in EU waters. Before the method can be applied, it is inevitable that steps will have to be taken to alter this regulation. A consensus on the merits of the method should therefore be achieved amongst fishermen, managers and scientists before a realistic attempt can be made to do so. 11-161

11 Comparison of the different trawls and devices tested

All details on the selective properties of the standard shrimp beam trawl, the selective sorting grid, the selective sieve net and the electro-net can be found in Sections 7 to 10. An overview and comparison of the characteristics of the different selective trawls is given in summary in this Section (Table 11-1). For clarity and brevity, the characteristics are described only qualitatively. Only results obtained from sea trials in commercial conditions are presented. Several characteristics are relevant in assessing the effectiveness of a selective trawl: • The potential loss of commercial catch is an important factor for the fishermen and should be kept as low as possible. Which level of catch reduction would be endurable for the Belgian shrimp fleet is difficult to assess and would be a subject for a separate study. The study by Revill et al. (1999), however, indicated that the revenues of the Belgian shrimp fleet are low compared to the other North Sea shrimp fleets and any reduction in catchability of the shrimp trawl would lead to a decline of the fleet. • A second factor is the catch reduction of non-commercial catch, which has been the focus of this study. This catch consists of undersized commercial fish and shrimp, non­ commercial fish species and non-commercial invertebrate species. • Thirdly, costs, labour and practice should be considered: the introduction of a new selective device may increase the cost of the trawl, increase the maintenance and cause practical problems during operation. • Once a new device enters the legislation it will need to be enforced. The ease and straightforwardness of inspection is important for the device to be used correctly and to reach its full potential in reducing discards. This is especially the case if commercial catch losses are likely to occur. For the comparison, the following devices were selected: the sorting grid with outlet on top, the grid with outlet below (Set B), the sieve net and the electro-net. Differences in catches have been expressed relative to the standard commercial catches obtained in the simultaneous fishing operation during the experiment. The loss of commercial shrimp catch is very high for both sorting grids. Many alterations were tried to improve this but none was successful. It is therefore unlikely that the sorting grid would be acceptable to the fishermen. For the sieve net, the commercial catch losses were moderate for shrimp and low for fish. Further technical alterations to the sieve net, avoiding gilled fish around the outlet, will probably reduce these losses. The problem of clogging with the sieve net, which also causes loss of commercial catch, can be avoided by a temporal enforcement over the year. This is already implemented in the present legislation, which was welcomed by the fishermen. A further fine-tuning of the period of enforced use may be necessary to maximize the use of the sieve net. With the electro-net, the catch reduction for commercial shrimps was low. For commercial fish, the numbers caught were too low to draw firm conclusions, but the data hint at a higher loss compared to the sieve net. The electro-net is in its first stage of development and future work should optimize this fishing gear. A larger, rectangular electric field and an improved groundrope will probably reduce the shrimp losses to a minimum and may even lead to higher catches compared to the traditional shrimp trawl. The amounts of undersized commercial fish caught with the sorting grids were low compared to the traditional net. Especially for fish below 10cm length, the grids performed better compared to the sieve net and the electro-net. The catch reductions for these small fish were Comparison o f the different trawls and devices tested 11-162 only moderate for the sieve net and with the electro-net often a catch increase was observed. Since the importance of these fish in the discards of Belgian shrimp trawlers is low, this is not considered problematic. The selective properties of the sieve net and the electro-net for fish with lengths between 10cm and minimum landing size (MLS), on the other hand, were good. Thus, for the catch composition on Belgian fishing grounds, the grid and the sieve net, and to a lesser extent the electro-net, can be considered very good tools for reducing the discard problem for coimnercial fish species. In areas where small fish (Age 0) prevail, on the other hand, like in the Waddensea, these devices cannot solve the discarding problem.

Table 11-1: Overview of the characteristics of the different selective trawls and devices Characteristic Sorting grid, Sorting grid, Sieve net Electro-net outlet on top outlet below Brown shrimp < MLS Brown shrimp ~> MLS ' Dab < 10cm 10cm -M LS "> MLS ' Plaice < 10cm iÖ cm -M LS ">MLS” Sole < 10cm IÖcm-M LS ">MLS” Bib < 10cm IÖcm-M LS ">MLS” Whiting < 10cm ÏÖcm-M LS > M L S ' Cod < 10cm 10cm -M LS ">"m l s " Non-commercial fish______Non-commercial invertebrates Material cost < 250 € < 250 € < 250 € < 5000 € Labour - installation < 1 day < 1 day < 1 day < 2 days Labour - maintenance low low low moderate Vulnerability high high moderate moderate Complexity low moderate low moderate Major problem clogging, clogging, clogging failure of damage damage equipment Likely acceptance by fishermen very low very low fair fair Enforcement easy easy moderate difficult

Colour scale: indicates the catch reduction or catch increase of a selective trawl compared to the traditional shrimp beam trawl used in coimnercial conditions______-50 -25 -15 15 25 50

insufficient data

Each of the devices tested had very good selective properties for non-commercial fish and invertebrates. The catch reductions were species dependent but the majority of the animals were sorted out. The sieve net and the grid with outlet on top performed best, followed by the 11-163 electro-net and the grid with outlet below. Again, further optimization of the electro-net will probably improve these catch reductions. The cost and labour for the installation and maintenance of sorting grids and sieve nets is low. Compared to this, the electro-net is expensive and installation and maintenance takes more time. The sorting grid was very vulnerable at sea and often damage was observed. Compared to the grid, damage incurred with the sieve net and the electro-net was rare. As for complexity, the sorting grid with outlet on top and the sieve were considered as simple devices. The pulse generators, the electrodes, the raised groundrope, the cable connection with the vessel etc. necessary for electro-fishing, made the electro-net more complex. Especially the installation of the device needed good guidance. At sea, the system was, however, quite manageable for non-experts. The main problems observed with the grids and the sieve net was clogging. Since the selectivity of these devices was based on filtering the catch, any object sticking to the fdter reduces its filtering capacity. Especially for the grid, this was problematic since the filtering surface is quite small. For the sieve net, this surface is large and partial clogging usually did not cause commercial catch loss. The main problem observed with the electro-net, was failure of the equipment. This only happened a few times, but any breakdown leads to significant loss of fishing time. Robust equipment is thus necessary. Combining all these considerations, it can be concluded that the sorting grid can be rejected especially because of its vulnerability, liability to clogging and loss of commercial catch. The sieve net has good selective properties, is cheap and easy to install and maintain and seems to be the better choice with the present state of knowledge. The electro-net shows very good potential, but further research and optimization is needed. It is also more complex and expensive compared to the sieve net. The advantage of the electro-net compared to the sieve is that it is likely that the commercial catch loss can be reduced to zero or even increased. The fate of the escapees is also better with electric fishing because the physical contact of the animals with the fishing gear is minimal. With the sieve net, escapees come into contact with the net and the outlet cover and can be damaged before they can escape. The reaction of fishermen follows the same lines. The sorting grid is rejected mainly because of the handling problems, clogging and loss of commercial catch. The sieve net seems to be acceptable. Mainly clogging of the sieve in certain periods of the year is considered as a serious drawback but the temporal enforcement by the legislation gives in to the objections. Fishermen are very interested in electric fishing - the idea to start research with electro trawls for shrimps was initiated by a Belgian fishing vessel owner. But there is reluctance to adopt a new trawl that is only in its first stage of development. Interest has, however, been expressed for cooperation on a project basis to further develop the system. The sieve net is a simple device so enforcement should be fairly straightforward but inspection at sea can be difficult. The sieve net is rigged inside the shrimp trawl and is rather large so it can be difficult to have visual contact with all parts of the sieve. Doing measurements or counting meshes is even more difficult while the trawl is hanging alongside the vessel, especially in rough weather. Even when the whole trawl is taken on board, inspection is not easy. Enforcement of an electro-net is considered even more difficult. Catch composition strongly depends on the rigging of the electrodes and the groundrope. Small changes can affect the catches and it can be hard for an inspector to assess whether a rigging is optimal for good selectivity. Measuring the characteristics of the pulses is not straightforward at sea and needs sensitive electronic equipment. It is therefore suggested that trawls like the electro-net should not be enforced but could be introduced on a voluntary basis. Inspection of the catches rather Comparison o f the different trawls and devices tested 11-164 than inspection of the fishing gear could be more effective. Thresholds of amounts of by- catches could be set up for allowing the use of this trawl. If the thresholds are not met, the standard selective trawl (the sieve net) should be used. As a conclusion, all arguments point at the sieve net as the most obvious selective trawl to be introduced into the shrimp fishery. The sorting grid does not meet the requirements and has more drawbacks than benefits. The electro-net is in a too early stage of development for commercial use but has good potential for the future. 12-165

12 General discussion and conclusions

Global production from capture fisheries and aquaculture is currently the highest on record and remains very significant for global food security. However, as fishing pressure continues to increase, many fisheries are facing a crisis. Several stocks are depleted or recovering from depletion and most of the healthy stocks are at their maximum exploitation level. Overall, the international community is well aware of the crisis facing fishery resources and fisheries, and has taken a number of important steps towards improved global management of fisheries. They all focus on reductions in fishing mortality, fishing effort and discarding and the growing care about the impact that fishing may have on the structure and function of ecosystems as a whole. The challenge for the 21st century is to implement management measures successfully and in particular, to translate policy to the level of application by the stakeholders. The study presented in this report concentrates on one small segment that contributes to the fishery crisis, i.e. the demersal fishery for Brown Shrimps (Crangon crangon ) in the North Sea and the associated by-catch and discard issues. The North Sea Brown Shrimp trawler fleet is, on average, quite old and consists of about 650 vessels with an engine power below 221 kW. About 75% of the fleet is based in Germany and the Netherlands. The fishing gear used is a simple beam trawl with a similar design over the whole fleet. Due to the location of the fishing grounds and the size of the target species and the associated small mesh size, the Brown Shrimp fishery faces a serious discarding problem. The fishery itself is in a healthy condition but the discarding practices may have consequences for other fish stocks and the health of the ecosystem where the fishery is carried out. The information collected in the first phase of this project on fishing effort and landings, vessels and operational characteristics of the fleet, fishing gear and fishing grounds was a sound basis for further activities in the project such as the discard sampling programme and the selectivity experiments. The results of the Brown Shrimp fishery discard sampling program revealed a discard ratio of 71%, indicating that almost three quarters of the total catch volume is returned to the sea. The quarterly data revealed clear seasonal trends in both the relative abundance and the size composition of the species caught. These trends are related to migration patterns, reproductive cycle, the time of the year when fish reach their size at first capture, the subsequent growth and mortality rates. The yearly totals gave a good indication of the relative importance of the different species discarded. For shrimp, the catches and the discards in wintertime were very low. In spring, they gradually rose and reached a maximum in summer to drop in autumn, to almost half of the maximum. Shrimp discards contained more than double the amount of shrimps (in numbers) compared to shrimps landed. The overall differences between the national fleets (B, D, Dk, Nl, UK) in the numbers of fish caught and discarded per unit of swept area, broadly reflect the differences in both the geographical distribution and habitat preference of the species investigated. In Belgium, mainly species with a preference for open waters were abundant in the catches, contrary to e.g. the Waddensea containing species with a clearly estuarine distribution. Most striking were the huge numbers of juvenile Plaice caught in the German shrimp fishery and high numbers of juvenile Dab caught all over the North Sea. Species like Brill, Turbot and gurnards were only discarded in low numbers. The roundfish species and Flounder and Sole took an intermediate position. The data collection in the discard sampling program was done over a large area, in the same period of time, following a standard methodology. This project was the first initiative to General discussion and conclusions 12-166 study this issue in such a wide context. The data presented were absolute numbers of discards. Before interpreting the results, it was necessary to examine these figures carefully in relation to other factors that determine the composition of a fish stock. This was done by processing the data through a newly developed biological and economic model. This model produced estimates of the annual lost landings for four selected species arising from the current levels of discarding in the European shrimp fisheries. These were around 2,000 t for Cod, 1,500 t for Whiting, 12,000 t for Plaice and 600 t for Sole. The estimated market value of these landings is over 25 Million €. To realise these potential landings in full, a management policy aimed at zero discarding in these fisheries would have to be implemented. Partial gains could be realised by the implementation of discard reduction measures such as the use of more selective fishing gears in these fisheries. Except for Plaice, where the potential gains amount to 12 % of the year 2000 North Sea TAC, the expected gains in the round- and flatfish directed fisheries are rather low (2.3% for Cod, 5.1% for Whiting and 3.0% for Sole). The model results indicated that for open waters like in the Belgian shrimp fishery, most gains could be achieved by reducing the discards for Age 1 and older fish. Any selective device should thus concentrate on these fish. In other areas like the Waddensea, mainly Age 0 fish should be reduced in the by-catches. Management regulations should consider these differences. The bio-economic modelling indicated that overall, the introduction of selective gears into the North Sea shrimp fisheries will represent a net benefit to the EU fisheries, although there will be some individual winners and losers and transfer of profitabilities. On average, the selectivity of the shrimp trawl cod-end was rather poor for shrimp allowing relatively high quantities of non-marketable shrimps to be retained. The selectivity of the net body, however, was quite important and allowed more shrimps to escape than the cod-end. Due to the small mesh size, cod-end selectivity for fish was very poor. Only a small part of even the 0-age group fish could escape through the meshes. In addition, the other parts of the net did almost not allow any fish to escape as well. Consequently, the shrimp beam trawl inevitably catches high amounts of small fish in coastal areas and estuaries, where juveniles prevail. An increase in mesh size to improve this situation would lead to losses of marketable shrimps. The selectivity experiments also indicated that escape windows or separator panels are not an option since these will most likely lead to high commercial shrimp losses. Selective devices, like sorting grids or sieve nets, to separate the shrimps from the fish and/or benthos by-catch are therefore probably a better way to improve species selectivity of shrimp beam trawls. In addition, an alteration to the groundrope may show potential, especially if an alternative stimulation could be found with a high fishing power for shrimps and high selectivity for other species. Several experiments with different Nordmore type sorting grid designs were carried out. If the catch composition did not cause clogging problems, the sorting grid performed well. The reduction of age 1 and older fish in the catch was satisfactory. Age 0 fish were sorted out as well, but to a lesser extent. Most benthic animals were selected out by the grid. If, however, there was material in the catch that caused clogging of the grid, the commercial Brown Shrimp catch soon fell below the level acceptable to commercial fishermen. Starfish, seaweeds, hydroids, jellyfish and debris were the main causes of clogging. Several alterations were made to the grid and the outlet, but a good by-catch reduction together with a low shrimp loss was never obtained. Although the selective grids have some clear advantages, like catch reduction of Age 1+ fish, non-commercial fish and invertebrates and better cod-end selectivity for shrimp, the device is 12-167 too susceptible to malfunction. In the particular Belgian situation, as regards catch composition and fishing ground characteristics, the sorting grid is a difficult device to operate and unlikely to be accepted by fishermen. As an alternative, a sieve net was tested for its selective properties and an evaluation was made of its operational characteristics. The sieve was based on commercially used designs. The loss of commercial shrimp catch when using a sieve net was 15% or less in favorable conditions. Most of the Agel+ fish, non-commercial fish and invertebrates were sorted out by this device. Also marketable fish, significantly contributing to the fishermen’s income, escaped. It is therefore important that legislation would allow the selective shrimp trawl to be fitted with a large mesh outlet cover, allowing marketable fish to be caught whilst allowing undersized fish to escape. The attitude of fishermen towards selective devices in general is not positive because any alteration to their traditional nets is prone to cause practical problems. The sieve net seemed, however, to be a more or less acceptable device to improve selectivity of their nets. In general, it can be concluded that the sieve net was less susceptible to clogging compared to the grid and performed better in different conditions. It was also rarely subject to damage. Because the sieve net mainly saves Age 1+ fish, this device is of rather low value in areas where large amounts of smaller fish are caught. This is the case in e.g. the Waddensea. In Belgian waters, where Age 1+ fish are far more important for the fish stocks, the sieve can indeed be considered as a valuable tool in reducing fish discards and contribute in reducing the pressure on the stocks. On top of this, the application of the sieve net also leads to a significant reduction in unwanted by-catch of invertebrates and non-commercial fish, which would reduce the impact of the shrimp fishery on the marine environment in general. It is, however, important to acknowledge that losses of commercial shrimp in certain seasons and areas, lead to financial losses for the shrimp fishermen. Experiments with electric pulses have been carried out in the past in many areas in the world. The main purpose of these experiments was to obtain higher catches of the target species. The idea in the present project was to use electric pulses to improve the selectivity of the shrimp beam trawl, i.e. reduce the discards while maintaining the target species catches. The observation tests in the laboratory demonstrated that Brown Shrimps react quite intensely to electric pulses. Fish and other invertebrates, with the exception of Dab and Sole, almost do not react to the pulses or if a reaction is observed, the animals keep close to the bottom. This means that, in principle, species selective fishing for Brown Shrimps should be possible with electric pulses as an alternative stimulation. By raising the groundrope, non­ target animals could escape underneath the groundrope. The shrimps, being startled up in the water column by the pulses, would lift over the groundrope and be caught by the trawl. From the sea trials, it can be concluded that the electro-net with raised groundrope and small meshes in the top panel gave satisfactory results. The losses of commercial shrimp catches were small or even non-existent. Part of the undersized commercial fish catch could escape and especially non-commercial fish and invertebrates were caught in lower numbers compared to the standard net. Future work should pay attention to the design of an electro­ trawl with a larger net opening to fit long electrodes and a new type of bobbin rope with less bottom contact. With this in mind, electric fishing should be a feasible alternative to the standard shrimp trawl and could be an acceptable alternative between the economic interests of the fishermen and the ecological demands of the marine ecosystem. Further work is, however, necessary. The study presented in this report has attempted to make a vertical integration of disciplines to reach a well-defined goal, i.e. reduce the discards in shrimp fishing. The inventory of General discussion and conclusions 12-168 operational characteristics of the fleet and features of the fishing gears (Section 4) was a sound basis for the consecutive projects that led to this goal. The discard sampling program (Section 5) and the biological and economic modelling work (Section 6) put the discarding in its correct perspective. With these data, a tool was available to convince fishermen that a reduction of discards was necessary. The request of environmentalists to close fisheries or seriously decrease fishing effort could be countered. A compromise could then be sought by seeking technical solutions to reduce discarding while allowing fishermen to maintain their income and way of living. The detailed study of the selectivity of the traditional shrimp trawl (Section 7) was an intermediate step that prepared further species selectivity work. Next, technical alterations to the trawl were tried out to reduce the discards while maintaining the commercial catches (Section 8-10). The final step in the project was giving the necessary support to have the successful devices implemented. Studies often concentrate on one step in this sequence and lack of integration prevents implementation. The necessity to get the consecutive projects financed each time to prevent stopping the sequence was a heavy burden on the responsible researchers. Perseverance and cooperation but also a portion of luck were necessary to obtain this. The international cooperation and integration of expertise from different fields (technology, biology, and economy) added to credibility of the work and gave it a wide support. The manager’s confidence that was obtained by this approach became clear when the European Commission postponed its legislation for the North Sea shrimp fishery for two years, i.e. until after finalisation of the DISCRAN project. Based on the conclusions of the project, the sorting grid and/or sieve net were enforced. The EC allowed countries to define a period when a selective device should not be used if clogging would prevent fishing. As for the Belgian implementation of the EU-legislation, good cooperation between managers and scientists was established. In Belgium, management considered the conclusions of the project and chose the sieve net while rejecting implementation of the sorting grid. It should also be noted that throughout the project, cooperation with fishermen was very good. Despite the fear that their future could be jeopardised, we experienced a positive and open attitude. The new technical measures were established in 2002 (Belgisch Staatsblad, 13.07.2002). Without doubt, these measures will reduce the impact that shrimp fishing has on the fish stocks and on the ecosystem as a whole. Nevertheless, where fishing occurs, impact on the ecosystem is inevitable. The Sea Fisheries Department has the intention to continue to strive to achieve more environmentally friendly fishing. 13-169

13 Abstract

The Brown Shrimp (Crangon crangon) fishery in the North Sea, the focus of this study, is carried out with small meshed nets in vulnerable areas like coastal zones and estuaries. The discarding practices associated with it have been regarded as a problem for many years. The discussion, however, was difficult since no sufficiently reliable discard data were available. The need for data on this issue and a solution for the discard problem was the starting point of the study presented in this report. The main objectives of the study were to quantify the biological and economic consequences of discarding in the Brown Shrimp fishery and to evaluate possible technical alterations to the shrimp beam trawl to reduce discarding in this fishery. This study has focused on the Belgian shrimp fishery, but, where relevant, data collected by other project partners was used to broaden the picture and present information for the whole North Sea Brown Shrimp fishery. In order to obtain detailed information on the structure, characteristics and exploitation patterns of the shrimp fishery in the North Sea, a thorough inventory was carried out. To fill the gap in knowledge on discarding practices in the North Sea Brown Shrimp fishery, a number of institutes in the major Brown Shrimp fishing nations agreed to identify the magnitude of the problem in a cooperative discard sampling programme. In this study, absolute numbers of discards were produced. These were carefully examined in relation to other factors that determine the composition of a fish stock by using a newly designed biological and economic model. The annual lost landings arising from the current levels of discarding in the European shrimp fisheries, as calculated by the model, are estimated to be around 2,000 t for Cod, 1,500 t for Whiting, 12,000 t for Plaice and 600 t for Sole. The estimated market value of these landings is over 25 Million €. The effectiveness of selective devices, closed areas and seasons was also evaluated. A next step in the project was a detailed study of the selectivity of the shrimp trawl cod-end, the net and the groundrope for Brown Shrimp and commercial fish species. This was done for the sake of having a good description of the selective properties of the shrimp beam trawl and as a preparation for the experiments with selectivity improving devices. Three techniques were selected for study: 1) a selective sorting grid, 2) a selective sieve net and 3) electric pulses as an alternative stimulation. Although the selective grids have some clear advantages, like catch reduction of Age 1+ fish, non-commercial fish and invertebrates and better cod-end selectivity, they were found to be too susceptible to malfunction. In the particular Belgian situation, as regards catch composition and fishing ground characteristics, the sorting grid is a difficult device to operate and unlikely to be accepted by fishermen. The sieve net on the other hand, seemed to be a more acceptable device to improve selectivity of their nets. In general, it can be concluded that the sieve net was less susceptible to clogging compared to the grid and performed better in different conditions. It was also rarely subject to damage. The loss of commercial shrimps was lower compared to the grid. The sieve can be considered as a valuable tool reducing fish discards and contributes to reduce the pressure on the fish stocks if used in Belgian waters. On top of this, the application of the sieve net also leads to a significant reduction in unwanted by-catch of invertebrates and non-commercial fish, which would reduce the impact of the shrimp fishery on the marine environment in general. It is, however, important to acknowledge that losses of commercial shrimp in certain seasons and areas, lead to financial losses for the shrimp fishermen. Abstract 13-170 The potential of electrical pulses as an alternative stimulation was chosen as the third option to develop a species selective shrimp trawl. The basic idea was to invoke selectively a startle response for shrimp with electric ticklers and to allow non-reacting species to escape underneath a raised groundrope. The laboratory observation tests indicated that Brown Shrimps reacted quite intensely to electric pulses. Fish and other invertebrates, with the exception of Dab and Sole, almost did not react to the pulses or if a reaction was observed, the animals kept close to the bottom. This means that, in principle, species selective fishing for Brown Shrimps should be possible with electric pulses as an alternative stimulation. From the sea trials, it can be concluded that the electro-net with raised groundrope and small meshes in the top panel gave satisfactory results. The losses of commercial shrimp were small or even non-existent. Part of the catch of undersized commercial fish species could escape and especially non-commercial fish and invertebrates were caught in lower numbers compared to the standard net. Future work should pay attention to a new design of an electro­ trawl New technical measures were established in 2002 and included, for Belgium, the enforcement of the use of sieve nets in the Brown Shrimp fishery. Without doubt, these measures will reduce the impact that shrimp fishing has on the fish stocks and on the ecosystem as a whole. Nevertheless, where fishing occurs, impact on the ecosystem is inevitable. The Sea Fisheries Department has the intention to continue to strive towards more environmental friendly fishing. 14-171

14 Samenvatting

De visserij op grijze garnaal (Crangon crangon ) in de Noordzee, het onderwerp van deze studie, gebeurt met fijnmazige netten in kwetsbare gebieden zoals kustzones en estuaria. De teruggooipraktijken die ermee samengaan zijn al jarenlang een probleem. De discussie is altijd moeilijk geweest doordat niet voldoende betrouwbare gegevens over de teruggooi beschikbaar waren. De behoefte aan gegevens hierover en aan een oplossing voor het teruggooiprobleem was het uitgangspunt van deze studie. De studie beoogde vooral de biologische en economische gevolgen van teruggooi in de grijze-gamalenvisserij te kwantificeren en mogelijke selectieve aanpassingen aan de gamalenboomkor te evalueren. Deze studie heeft zich vooral toegespitst op de Belgische garnalenvisserij, maar waar relevant, werden gegevens van andere projectpartners aangewend om het beeld te verruimen en informatie te verschaffen over de garnalenvisserij in de gehele Noordzee. Om gedetailleerde informatie te verzamelen over de structuur, eigenschappen en exploitatiepatronen van de garnalenvisserij in de Noordzee werd een grondige inventaris opgemaakt. Teneinde de omvang van het teruggooiprobleem vast te stellen, werd door enkele instituten een gezamenlijk bijvangstbemonsteringsprogramma uitgevoerd. In deze studie worden absolute teruggooicijfers gegeven. Hun belang werd onderzocht met behulp van een nieuw ontworpen biologisch en economisch model, rekening houdend met andere factoren die de samenstelling van een visbestand bepalen. Het model berekende dat het jaarlijks vangstverlies, veroorzaakt door de huidige omvang van de teruggooi binnen de Europese garnalenvisserij, rond de 2.000 ton ligt voor kabeljauw, 1.500 ton voor wijting, 12.000 ton voor pladijs en 600 ton voor tong. De geschatte economische waarde van dit verlies is meer dan 25 miljoen euro. De efficiëntie van selectieve tuigen, gesloten gebieden en seizoenen werd ook geëvalueerd. Een volgende stap in het project was een gedetailleerde selectiviteitsstudie van de kuil van het gamalennet, het net en de onderpees voor grijze garnaal en commerciële vissoorten. Dit werd uitgevoerd om een goede beschrijving te hebben van de selectiviteitseigenschappen van het gamalennet en ook ais voorbereiding voor de experimenten met de selectieve tuigen. Er werden drie technieken geselecteerd om de teruggooi te verminderen: 1. een selectief rooster 2. een selectief zeefnet 3. elektrische pulsen ais alternatieve stimulus Hoewel het selectief rooster enkele duidelijke voordelen heeft, zoals verminderde vangst van vis van leeftijdsklasse 1+, niet-commerciële vis en ongewervelden en een betere selectiviteit in de kuil, functioneert het dikwijls onvoldoende. Vooral in de Belgische situatie qua vangstsamenstelling en eigenschappen van visgronden, is het rooster een moeilijk instrument om mee te werken en zou het bovendien moeilijk aanvaard worden door de vissers. Het zeefnet daarentegen lijkt beter aanvaardbaar ais middel om de netten selectiever te maken, hoewel problemen onderkend worden. Over het algemeen was het zeefnet minder onderhevig aan verstopping in vergelijking met het rooster en het bleef beter functioneren onder verschillende omstandigheden. Het liep bovendien zelden averij op. Het verlies aan commerciële garnaal was ook lager in vergelijking met de verliezen met het rooster. Het zeefnet is een waardevol middel om teruggooi te verminderen en de druk op de visbestanden te helpen lenigen in Belgische wateren. Daarenboven leidt het gebruik van het zeefnet tot een aanzienlijke afname van ongewenste bijvangst van ongewervelden en niet-commerciële vis. Dit vermindert in het algemeen de impact die de garnalenvisserij heeft op het mariene milieu. Samenvatting 14-172 Toch moeten we toegeven dat er ook verlies is aan commerciële garnaal in bepaalde seizoenen en gebieden, wat een financieel verlies voor de gamalenvissers betekent. Elektrische pulsen ais alternatieve stimuli werden bestudeerd ais derde optie voor het ontwikkelen van een soortenselectief gamalennet. Het basisidee was om selectief een schrikreactie uit te lokken bij garnaal met elektrische pulsen en om soorten die niet reageren de kans te geven te ontsnappen onder de verhoogde onderpees. De laboratoriumtesten wezen uit dat grijze garnaal hevig reageert op elektrische pulsen. Vis en andere ongewervelden, met uitzondering van schar en tong, reageren bijna niet op de pulsen of ais ze al reageren, blijven ze dicht bij de bodem. Dit betekent dat soortenselectief vissen op grijze garnaal in principe mogelijk moet zijn met elektrische pulsen ais alternatieve stimulus. De zeeproeven toonden aan dat het elektrisch net met de verhoogde onderpees goede resultaten gaf. Het verlies aan commerciële garnaal was klein of zelfs onbestaand. Een gedeelte van de ondermaatse commerciële vissoorten kon ontsnappen en vooral niet-commerciële vis en ongewervelden werden in kleinere aantallen gevangen, vergeleken met de vangst van het standaardnet. Verder onderzoek moet de nodige aandacht geven aan een nieuw ontwerp voor het elektrisch net. Voor de gamaalvisserij werden in 2002 nieuwe technische maatregelen opgelegd door de Europese Commissie. Voor België hielden die o.a. het verplicht gebruik in van het zeefnet. Deze maatregelen zullen ongetwijfeld de invloed van de garnalenvisserij op de visbestanden en op het ecosysteem verminderen. Niettemin, waar gevist wordt, is invloed op het ecosysteem onvermijdelijk. Er kan gezocht worden naar een compromis die de belangen verzoent van de visser en van het mariene milieu waarin de doelsoorten van de visserij leven. Het Departement Zeevisserij zal blijven streven naar milieuvriendelijker vistechnieken. 15-173

15 Executive summary

World capture fisheries production has shown a steady increase from the mid 20th century. It remained rather stable during the last decade and is at a level close to the historical maximum. This generally stable situation for global catches masks regional disparities. As fishing pressure continues to increase, about 47% of the world’s major fish stocks are fully exploited and have no room for further expansion. Another 18% are over-exploited and need remedial action and 10% have been depleted or are recovering from depletion. In the North Sea, the state of the stocks of most round- and flatfish species has deteriorated during the last decade. Some of these stocks have reached a historical low. On several occasions, concern about the situation of fish stocks and the urgent need to take management actions has been expressed. Recommendations all focused on reductions in fishing mortality, reductions of fishing effort and implementation of management measures to reduce the amounts of juveniles caught and discarded. In addition to the concerns expressed about individual fish stocks, there is a growing worldwide interest in and concern about ecosystems and the impact that fishing may have on their structure and function. The Brown Shrimp (Crangon crangon) fishery in the North Sea, the focus of this study, is carried out with small meshed nets in vulnerable areas like coastal zones and estuaries. The discarding practices associated with it have been regarded as a problem for many years. The discussion, however, was difficult since no sufficiently reliable discard data were available. There was no discussion on the issue that large amounts of by-catch were returned to the sea, but the seriousness of the problem and the possible consequences for the fish stocks and the ecosystem were minimised by the fishermen. Environmentalists claimed the opposite. The need for data on this issue and a solution for the discard problem was the starting point of the study presented in this report. The main objectives of the study were to quantify the biological and economic consequences of discarding in the Belgian Brown Shrimp fishery and to evaluate possible technical alterations to the shrimp beam trawl to reduce discarding in this fishery. It was carried out in five consecutive steps: • the collection of data on the characteristics of the Belgian shrimp trawler fleet, • the collection of data on the discarding practices in the Belgian Brown Shrimp fishery, • estimation of the biological and economic significance of the discarding, • study of the selectivity of the shrimp beam trawl and • evaluation of discard reducing technical alterations to the shrimp beam trawl as possible management tools. This study has focused on the Belgian shrimp fishery, but, where relevant, data collected by other project partners was used to broaden the picture and present information for the whole North Sea Brown Shrimp fishery. In order to obtain detailed information on the structure, characteristics and exploitation patterns of the shrimp fishery in the North Sea, a thorough inventory was carried out. Crangon crangon is caught by trawlers, generally rigged for twin beaming. The design of the net is quite similar over the whole shrimp trawler fleet. Due to the small mesh size necessary to catch the shrimps and the location of the fishing grounds, there is a significant by-catch of juvenile finfish and invertebrates. The unwanted by-catch is discarded overboard and may suffer from high mortality. The Belgian shrimp fishery is a typical seasonal fishery with peak landings and effort in the period August-October, although part of the fleet targets shrimps all year round. In 1995, there were respectively 51 and almost 650 vessels engaged in the Executive summary 15-174 Belgian and North Sea Crangon fisheries. About 75% of the fleet was based in Germany and the Netherlands. The North Sea Brown Shrimp fishery is in a good condition although the statistics indicate declining catch and effort levels for the Belgian fleet. The LPUE’s are quite stable in general but decreasing in Belgium. The information collected on the shrimp trawler fleet was a sound basis for the further activities in the project such as the discard sampling programme and the selectivity trials. In order to fill the gap in knowledge on discarding practices in the North Sea Brown Shrimp fishery, a number of institutes in the major Brown Shrimp fishing nations agreed to identify the magnitude of the problem in a cooperative discard sampling programme. For the Belgian part of the discard sampling program, 21 sea trips and 108 hauls were carried out with a good seasonal coverage. The discard ratio observed, indicated that almost three quarters of the total catch volume is returned to the sea. For shrimp, the densities on the fishing grounds in wintertime were very low, as were the catches and the discards. In spring, the catches and discards gradually rose together with rising shrimp densities to reach a maximum in summer. After that, densities and catches dropped in autumn, to almost half of the maximum. Shrimp discards contained more than double the amount of shrimps (in numbers) compared to shrimps landed. The Belgian data revealed clear seasonal trends in both the relative abundance and the size composition of the by-catch species studied. These trends are related to migration patterns, reproductive cycle, the time of the year when fish reach their size at first capture, the subsequent growth and mortality rates. All project partners (B, D, Dk, Nl, UK) together did 104 sea trips and 527 hauls. The overall differences between the national fleets in the numbers of fish caught per unit of swept area, broadly reflect the differences in both the aerial distribution and habitat preference of the species investigated. Most striking were the huge numbers of juvenile Plaice caught in the German shrimp fishery and high numbers of juvenile Dab caught all over the North Sea. Species like Brill, Turbot and gurnards were only discarded in low numbers. The roundfish species and Flounder and Sole took an intermediate position. The issue of by-catches of juvenile fishes in the North Sea shrimp fisheries has been addressed in several studies and was identified as a major problem. The present project provided basic data on fleets and discards for this fishery in all countries surrounding the North Sea. The data collection was done over a large area, in the same period of time, following a standard methodology. This project was the first initiative to examine this issue in such a wide context. In this study, absolute numbers of discards were produced. Despite the alarming nature of these figures, it would be incautious to call for immediate management action. Before doing so, it seems advisable to carefully examine these figures in relation to other factors that determine the composition of a fish stock. Discard survival is an important factor determining the significance of discarding. Natural mortality caused by severe winters, predation or competition also plays a part in the degree of decimation of a year class, before it has a chance to reach the age of first capture. Fishing and discarding are, however, the man-made effects that can be managed and reduced, provided there is a strong case to do so. Whether the case would be strong enough was evaluated by a newly designed biological and economic model. The annual lost landings arising from the current levels of discarding in the European shrimp fisheries, as calculated by the model, are estimated to be around 2,000 t for Cod, 1,500 t for Whiting, 12,000 t for Plaice and 600 t for Sole. The estimated market value of these landings is over 25 Million €. Other by-catch species were not studied due to the lack of sufficiently reliable biological data. 15-175 The modelling exercise identified that the common current practice of quantifying discards in fisheries in terms of ‘numbers / weights / percentage of catch’ etc, may be misleading and may not identify the most important elements within a discarded population. For instance, at first glance the numbers of Whiting discarded in the European Crangon fisheries may seem significant at 55 million discarded fish during 1996. The modelling showed that the biological and economic importance of these Whiting discards is relatively insignificant. What is apparent is that discard reduction measures in this fishery should focus on reducing Plaice discards, particularly Age 0 fish in the Waddensea, and Age 1 groups elsewhere (e.g. Belgium). It was identified that these age groups of Plaice are the most important elements of the European Brown Shrimp related discard populations. Protective measures (such as closed areas, closed seasons or selective devices) that primarily aim at a reduction of the unwanted Plaice discards, will also reduce the discards of other species, both commercial and non­ commercial. In doing so, they may contribute to the protection of the marine environment as a whole. The possible establishment of closed areas is likely to lead to increased fishing pressure in adjacent areas that remain open to the fisheries, thus countering the benefits that were envisaged by these closures. Another way to resolve the discard problem could be the introduction of closed seasons aiming at maximum protection of the Age 0 Plaice. However, the modelling of a single month closure in the German fishery during quarter three only ‘recovered’ 11% of the estimated lost German Plaice landings. Increased ‘recovery’ of lost landings can only be achieved through longer periods of closure. Another approach could involve the introduction of species-selective gears, where the emphasis would be on the protection of Age 0 Plaice in certain areas (the flatfish nurseries in the Waddensea area), and on that of Age 1 Plaice in other areas (the coastal waters outside the nurseries, e.g. Belgium). The biological modelling incorporating selective gear in the shrimp fishery demonstrated the efficacy of a selective grid (or similar selective device) in reducing the discarding of Age 1+ group fish. In the modelling, more than 90% of the lost landings could be ‘recovered’ if the selectivity device was used in waters where Age 1 discards predominate (e.g. Belgium). Conversely, it was shown that where Age 0 fish predominate (e.g. the Waddensea), the selective device was considerably less effective (particularly for flatfish) and only 21% of the lost German Plaice landings could be ‘recovered’ using this device. Because of the demonstrable spatial difference in the age composition of the discards, a single management regulation that does not consider these differences is unlikely to realise fully the predicted potential benefits to the Plaice stocks and to their fisheries. The bio-economic analysis showed that the European Crangon fishery represents a unique, rather than a common situation. The imposition of inefficiencies through the introduction of more selective gears into the Crangon fisheries is unlikely to reduce profitability. In fact, there is likely to be a net benefit to the EU fishery as a whole from such an introduction. This phenomenon is not likely to be observed in other EU fisheries, and is an artefact of the unique price-quantity relationship that exists forCrangon. Most of the benefits of any increase in white fish landings arising from the introduction of discard reducing measures will go to the flatfish directed fleets, particularly in the Netherlands, the UK and Denmark. This is because of the nature the current North Sea TAC allocation. A next step in the project was a detailed study of the selectivity of the shrimp trawl cod-end, the net and the groundrope for Brown Shrimp and commercial fish species. This was done for the sake of having a good description of the selective properties of the shrimp beam trawl and as a preparation of the experiments with selectivity improving devices. The cod-end selectivity parameters found for Brown Shrimp were: L50 = 39.4 mm (37.0 - 41.4), the selection factor = 1.82 (1.71 - 1.91) and the selection range = 11.6 mm (10.2 - Executive summary 15-176 13.0). The cod-end selectivity of the beam trawl for shrimps was found to be very variable. Several variables like clogging of the meshes, catch volume and the state of the sea contributed to this variability. On average, the selectivity was rather poor for shrimp allowing relatively high quantities of non-marketable shrimps to be retained. The selectivity of the net body, however, was quite important and allowed more shrimps to escape than the cod-end. It was mainly the rounded lateral part of the net belly that contributed to this selectivity. Due to the small mesh size, cod-end selectivity for fish was very poor. Only a small part of even the Age 0 group fish could escape through the meshes. In addition, the other parts of the net hardly allowed any fish to escape. Consequently, the shrimp beam trawl will inevitably catch high amounts of small fish in coastal areas and estuaries, where juveniles prevail. An increase in mesh size to improve this situation would shift the selection curve for shrimp to the right, what means that more marketable shrimps would be lost. Escape windows are not an option since these will most likely lead to commercial shrimp loss. Selective devices, like sorting grids or sieve nets, to separate the shrimps from the fish and/or invertebrate by- catch are therefore probably a better way to improve species selectivity of shrimp beam trawls. Also an alteration to the groundrope may show potential, especially if an alternative stimulation could be found with a high fishing power for shrimps and high selectivity for other species. Based on these results, three options were selected for further study: 1) a selective sorting grid, 2) a selective sieve net and 3) electric pulses as an alternative stimulation. Several experiments with different sorting grid designs were carried out in the Belgian shrimp fishery to investigate the potential to reduce by-catch. The general layout of the grid was based on the Nordmore grid. In a first Phase, a design with an escape outlet on top was studied. If the catch composition did not cause clogging problems, the sorting grid met its purpose. The reduction of invertebrates and age 1 and older fish in the catch was satisfactory. The commercial Brown Shrimp catch was reduced but usually by not more than 15%. If, however, there was material in the catch that caused clogging of the grid, the commercial Brown Shrimp catch soon fell below the level acceptable to commercial fishermen. Clogging of the grid with starfish (which occurs in almost every catch) was a constant problem. Other grid designs with the outlet on top of the grid were tried out, in Phase 2 of the project, to overcome this problem but none were successful. Subsequent exploratory tests with the outlet below were quite successful and clogging with starfish was reduced significantly. Therefore, it was decided to adopt this design for further study in a third project Phase. Although research vessel trials were successful each time, the functioning of the grid on board of commercial vessels was not favourable. Clogging with starfish was seldom observed but the grid was still very sensitive to other clogging material (seaweed, hydroids, plastic bags, large debris...) that occurs very frequently in the catches off the Belgian coast. Large debris and jellyfish often caused problems with a small outlet. No compromise was found between an opening large enough to allow jellyfish and debris to escape and small enough to prevent losing large parts of the commercial shrimp catch. Although the selective grids have some clear advantages, like catch reduction of Age 1+ fish, non-commercial fish and invertebrates and better cod-end selectivity, the device is too susceptible to malfunction. In the particular Belgian situation, as regards catch composition and fishing ground characteristics, the sorting grid is a difficult device to operate and unlikely to be accepted by fishermen. 15-177 Based on early studies, the sieve net seems to be effective in releasing part of the discards, fish as well as invertebrates. It is not made of rigid material and therefore it is more acceptable to fishermen than a rigid sorting grid. This, together with the observation that fishermen tend to use the device voluntarily in certain circumstances, makes the sieve net an obvious choice for further study. During a one-year period, a sieve net was tested for its selective properties and an evaluation was made of its operational characteristics. The sieve was based on commercially used designs in the Netherlands and the UK. The loss of commercial shrimp catch when using a sieve net was 15% or less in favorable conditions. Certain components of the catch can, however, lead to a distortion of the outlet with a reduction of the commercial catch of over 30%. It is likely that technical alterations to the outlet can prevent this. A length effect in the catch reduction of shrimps was observed. Undersized shrimps escaped to a higher degree compared to the large ones, which reduced shrimp discards. The sieve net showed very poor selective properties for commercial fish species with a length below 10cm. Above 10cm the selection improves with increasing length. Especially for Age 1 and older fish, this selective device serves its purpose. Many fish of marketable size are caught in the Belgian Crangon fishery. Belgian shrimp fishermen depend for a significant part of their income on fish. With a sieve net, fish of marketable size are led to the outlet and escape. It is justified therefore for legislation to allow the selective shrimp trawl to be fitted with a large mesh outlet cover, allowing marketable fish to be caught whilst allowing undersized fish to escape. The attitude of fishermen towards selective devices is not positive because any alteration to their traditional nets is prone to cause practical problems. The sieve net seemed, however, to be a more acceptable device to improve selectivity of their nets. In general it can be concluded that the sieve net was less susceptible to clogging compared to the grid and performed better in different conditions. It was also rarely subject to damage. The loss of commercial shrimps was lower compared to the grid. The selectivity of the sieve net for Age 0 fish is very low. Therefore, this device is of rather low value in areas where large amounts of these small fish are caught (e.g. the Waddensea). The biological and economic modelling exercise has shown that a reduction of discarding of Age 0 fish in Belgian waters has limited benefits for the commercial fish stocks because densities are relatively low. Saving Age 1 and older fish in the Belgian shrimp fishery, on the other hand, could effectively benefit the stocks. Since the sieve is indeed selective for these fish, it can be considered as a valuable tool in reducing fish discards and contribute in reducing the pressure on the fish stocks if used in Belgian waters. On top of this, the application of the sieve net also leads to a significant reduction in unwanted by-catch of invertebrates and non-commercial fish, which would reduce the impact of the shrimp fishery on the marine environment in general. It is, however, important to acknowledge that losses of commercial shrimp in certain seasons and areas, lead to financial losses for the shrimp fishermen. Most selectivity enhancing measures concentrate on the net part of the trawl. These aim at catch separation or improved filtering of the catch, the disadvantage being that the animals are exposed to net meshes or other parts of the net before they can escape. Damage incurred by contact, or stress caused during the capture and escape process may lead to mortality amongst escapees. A better approach to improve selectivity, if at all possible, is to try to avoid unwanted sizes and species entering the net. To attain this goal, it is necessary to find alternative means of stimulation in the net mouth, i.e. a stimulus inducing the desired reaction from the target species without stimulating unwanted animals. The potential of Executive summary 15-178 electrical pulses as an alternative stimulation was chosen as the third option to develop a species selective shrimp trawl. The basic idea was to invoke selectively a startle response for shrimp with electric ticklers and to allow non-reacting species to escape underneath a raised groundrope. Experiments with electric pulses have been carried out in the past in many areas in the world. The main purpose of these experiments was to obtain higher catches of the target species. The idea in the project presented in this report was to use electric pulses to improve the selectivity of the shrimp beam trawl, i.e. reduce the discards while maintaining the target species catches. The project was initiated by the observation by a Belgian vessel owner that in the People’s Republic of China, over 2000 vessels were fishing with electrified nets for penaeid shrimps. In a first phase of the project, detailed observation and survival tests were carried out as a preparation of the sea trials. The second phase consisted of sea trials to compare the catches of a standard shrimp trawl with the electrified net in commercial conditions. From the observation tests the following conclusions can be drawn: • Brown shrimps react quite intensely to electric pulses. • The optimal pulse amplitude lies between 40 and 110V for a 50cm electrode spacing. Higher and lower amplitude has a negative effect on the response of shrimps. • The optimal pulse frequency lies around 5-6 Hz. • The startle response for small shrimps is slightly lower compared to large animals. • A higher water temperature results in a stronger response. • A low light intensity results in a stronger response. • The maximum response is usually obtained within 4s after the start of the pulses. Fish and other invertebrates, with the exception of Dab and Sole, almost do not react to the pulses or if a reaction is observed, the animals keep close to the bottom. This means that, in principle, species selective fishing for Brown Shrimps should be possible with electric pulses as an alternative stimulation. The sea trials were carried out with two generators, giving a 64V pulse at 6Hz. This amplitude and frequency combination does not give the best response but it was the closest possible with the available equipment and only little below the optimum. For a shrimp fishery by night, the estimated percentage (based on the laboratory tests) of large shrimps in the trawl path that could be caught with an electro-trawl would be 79% compared to the traditional fishery. The price for a possible environmental benefit in reduction of discards and reduced sea floor disturbance would thus be a reduction in commercial shrimp catch of roughly 20%. If a pulse generator would be available generating a more powerful pulse, the predicted catches could increase by roughly 10%. Because the price coefficient for Brown Shrimps is about -1, a decrease in shrimp landings should go together with a relatively equal rise in price of the product. So, if applied by the whole fleet, the overall income of the shrimper fleets would stay roughly unchanged. The results of the sea trials matched these projections quite closely. With the selective electro-net with electrode array A, catch losses of large shrimps were 12% and not significant. For small shrimps, a catch increase was even observed. The sea trials also indicated that the raised groundrope created an escape route for most of the species regularly caught in shrimp trawls. The electric field, however, made the shrimps tail-flip high enough to be caught. 15-179 It can be concluded that the electro-net with raised groundrope and small meshes in the top panel gave satisfactory results. The losses of commercial shrimp catches were small or even non-existent. Part of the undersized commercial fish catch could escape and especially non­ commercial fish and invertebrates were caught in lower numbers compared to the standard net. Future work should pay attention to the design of an electro-trawl with a larger net opening to fit long electrodes and a new type of bobbin rope with less bottom contact. With this in mind, electric fishing should be a feasible alternative to the standard shrimp trawl and could be an acceptable alternative between the economic interests of the fishermen and the ecological demands of the marine ecosystem. It should, however, be borne in mind that the sea trials in this project only covered a short time range and a narrow range of conditions such as water temperature, currents, degree of activity of the shrimps etc. An extensive range of sea trials on commercial vessels in different conditions should precede commercial application. The study presented in this report has attempted to make a vertical integration of disciplines to reach a well-defined goal, i.e. reduce the discards in shrimp fishing. The inventory of operational characteristics of the fleet and features of the fishing gears was a sound basis for the consecutive projects that led to this goal. The discard sampling program and the biological and economic modelling work put the discarding in its correct perspective. With these data, a tool was available to convince fishermen that a reduction of discards was necessary. The request of environmentalists to close fisheries or seriously decrease fishing effort could be countered. A compromise could then be sought by seeking technical solutions to reduce discarding while allowing fishermen to maintain their income and way of living. The detailed study of the selectivity of the traditional shrimp trawl was an intermediate step that prepared further species selectivity work. Next, technical alterations to the trawl were tried out to reduce the discards while maintaining the commercial catches. The final step in the project was giving the necessary support to have the successful devices implemented. Studies often concentrate on one step in this sequence and lack of integration prevents implementation. The necessity to get the consecutive projects financed each time to prevent stopping the sequence was a heavy burden on the responsible researchers. Perseverance and cooperation but also a portion of luck were necessary to obtain this. The international cooperation and integration of expertise from different fields (technology, biology, and economy) added to credibility of the work and gave it a wide support. The manager’s confidence that was obtained by this approach became clear when the European Commission postponed its legislation for the North Sea shrimp fishery for two years, i.e. until after finalisation of the DISCRAN project. Based on the conclusions of the project, the sorting grid and/or sieve net were enforced. The EC allowed countries to define a period when a selective device should not be used if clogging would prevent fishing and also allowed the use of a large mesh outlet cover. As for the detailed national legislation, good cooperation between managers and scientists was established. In Belgium, management considered the conclusions of the project and chose the sieve net while rejecting implementation of the sorting grid. It should also be noted that throughout the project, cooperation with fishermen was very good. Despite the fear that their future could be jeopardised, we experienced a positive and open attitude. The new technical measures were established in 2002. Without doubt, these measures will reduce the impact that shrimp fishing has on the fish stocks and on the ecosystem as a whole. Nevertheless, where fishing occurs, impact on the ecosystem is inevitable. It is important that a compromise is found between the interests of the fishery and the quality of the ecosystem, the environment where the target species of the fishery live. The Sea Fisheries Department has the intention to continue to strive towards more environmental friendly fishing. Uitgebreide samenvatting 16-180

16 Uitgebreide samenvatting

Sedert de helft van de vorige eeuw heeft de visserij op wereldschaal een gestage groei gekend. In het vorige decennium stagneerde die groei om zich nu praktisch op een historisch maximum te bevinden. Dit schijnbaar evenwicht verbergt echter grote regionale verschillen in exploitatiedruk. De druk van de visserij neemt nog steeds toe terwijl ca. 47% van de belangrijkste visbestanden wereldwijd al maximaal geëxploiteerd wordt en dus niet extra bevist kan worden. 18% van de visbestanden wordt overbevist en heeft nood aan remediërende maatregelen en 10% van de bestanden is uitgeput of herstellende. Gedurende het laatste decennium zijn de meeste bestanden van rond- en platvissoorten in de Noordzee erop achteruit gegaan. Sommige hebben zelfs een historisch dieptepunt bereikt. De situatie is werkelijke verontrustend en er moeten dan ook dringend beheersmaatregelen worden genomen. Alle aanbevelingen richtten zich op de reductie van vissterfte, reductie van visserij-inspanning en het doorvoeren van beheersmaatregelen om het aantal juveniele vissen in de teruggooi te doen afnemen. Naast de noodkreet i.v.m. de individuele visbestanden, krijgt nu ook het ecosysteem wereldwijd meer aandacht en begint men zich zorgen te maken over de mogelijke invloed van de visserij op de structuur en functie van de mariene ecosystemen. De voorliggende studie concentreert zich op de visserij op grijze garnaal(Crangon crangon ) in de Noordzee. Garnalen worden gevist met fijnmazige netten in kwetsbare gebieden, zoals kustzones en estuaria. De teruggooi wordt al jaren ais een probleem beschouwd. Doordat er echter geen betrouwbare gegevens beschikbaar waren over de teruggooi in de garnalenvisserij, was dit een moeilijke discussie. Het stond buiten kijf dat grote hoeveelheden bijvangst in zee gegooid werden, maar de ernst van het probleem en de mogelijke gevolgen voor de visbestanden en voor het ecosysteem werden door de vissers geminimaliseerd. Vanuit de groene beweging werd net het tegenovergestelde beweerd. Deze studie werd uitgevoerd om aan de vraag naar de gegevens over teruggooi te voldoen en om een oplossing voor het probleem te bieden. De hoofdbedoeling van de studie was om de biologische en economische gevolgen van teruggooi in de grijze-gamalenvisserij te kwantificeren en om technische wijzigingen aan de boomkor door te voeren waardoor de teruggooi zou verminderen. De studie gebeurde in vijf stappen: • verzamelen van gegevens over de Belgische gamaalvloot • verzamelen van gegevens over de teruggooi in de Belgische grijze-gamalenvisserij • ramen van de biologische en economische betekenis van de teruggooi • onderzoeken van de selectiviteit van de gamalenboomkor • evalueren van de bijvangstverminderende technische wijzigingen aangebracht aan de gamalenboomkor ais mogelijke beheersmethode Deze studie heeft zich toegespitst op de Belgische garnalenvisserij, maar waar het interessant was het beeld te verruimen, werden gegevens van andere projectpartners aangewend en informatie gegeven over de grijze-gamalenvisserij in de gehele Noordzee. Om gedetailleerde informatie te verkrijgen over de structuur, kenmerken en exploitatiepatronen van de garnalenvisserij in de Noordzee werd een gedetailleerde inventaris opgemaakt. Crangon crangon wordt gevangen met treilers die doorgaans uitgemst zijn met twee boomkorren. Het ontwerp van het net is gelijkaardig voor de volledige gamalenvloot. De kleine afmetingen van de garnalen maken het noodzakelijk dat met 16-181 fijnmazige netten wordt gevist en de situering van de visgronden heeft tot gevolg dat de bijvangst van jonge dieren en ongewervelden aanzienlijk is. Deze bijvangst wordt overboord gezet waardoor de mortaliteit hoog kan oplopen. De Belgische garnalenvisserij is een typische seizoensgebonden visserij met piekvangsten van augustus tot oktober. Toch vist een gedeelte van de vloot het hele jaar door op garnalen. In 1995 waren 51 schepen actief in de Belgische garnalenvisserij en 650 in de gehele Noordzee. Ongeveer 75% van de vloot heeft zijn thuishaven in Duitsland of Nederland. De grijze-gamalenvisserij in de Noordzee doet het goed, maar toch wijzen de statistieken op een temgval voor de Belgische vloot. De vangst per eenheid van visserij-inspanning blijkt over het algemeen tamelijk stabiel te zijn, maar neemt af in België. De gegevens die verzameld werden i.v.m. de gamalenvloot vormen een goede basis voor het vervolg van het onderzoek dat o.a. bestond uit eenbijvangstbemonsteringscampagne en selectiviteitsproeven. Teneinde de omvang van het temggooiprobleem van de grijze-gamalenvisserij in de Noordzee in te schatten, werd door enkele instituten uit de “grijze-gamalenvisserijnaties” een gezamenlijk bijvangstbemonsteringsprogramma uitgevoerd. Voor het Belgische gedeelte van dit programma werden 21 zeereizen en 108 slepen uitgevoerd, gespreid over de vier seizoenen. Hiemii bleek dat bijna driekwart van het totaal gevangen vangstvolume temg in zee wordt gegooid. Wat garnaal betreft waren de visgronden dunbevolkt in de winter en vandaar ook dat de vangsten en de temggooi laag waren. In de lente namen de aantallen geleidelijk toe om in de zomer hun maximum te bereiken. Daarna zakten hun aantallen weer in de herfst tot ongeveer de helft van het maximum. De temggooi bevatte meer dan dubbel zo veel garnalen dan het aantal garnalen dat uiteindelijk aan land werd gebracht. De Belgische bijvangstgegevens vertoonden afhankelijk van de seizoenen duidelijke schommelingen qua aantallen en grootte van de bestudeerde soorten. Deze schommelingen zijn gerelateerd aan migratiepatronen, de voortplantingscyclus, het seizoen waarin de dieren commerciële afmetingen bereiken, de groei die daarop volgt en het sterftecijfer. De projectpartners (B, D, Dk, NI, UK) voerden samen 104 zeereizen uit en deden 527 slepen. De nationale verschillen in aantal gevangen dieren per eenheid van beviste oppervlakte zijn afhankelijk van de geografische spreiding en de habitat voorkeur van de onderzochte soort. Het verrassendst waren de enorme aantallen juveniele pladijs die de Duitse garnalenvisserij ving en de hoge aantallen juveniele schar die over de gehele Noordzee werden gevangen. Soorten ais griet, tarbot en poon werden slechts in kleine aantallen temggegooid. De aantallen rondvissoorten en schar en tong lagen daar tussenin. De bijvangst van jonge vissen in de garnalenvisserij in de Noordzee is een belangrijk probleem en er werden al verschillende studies aan gewijd. Het huidige project biedt basisgegevens over de vloot en de temggooi in de garnalenvisserij in alle Noordzeelanden. De gegevens werden voor een groot gebied in dezelfde periode en volgens eenzelfde methodologie verzameld. Dit project is het eerste dat de garnalenvisserij zo uitgebreid bestudeert. Dit rapport geeft absolute cijfers over temggooi. Deze cijfers zijn werkelijk verontmstend, maar toch zou het onvoorzichtig zijn om onmiddellijk beheersmaatregelen te nemen. Het is aan te raden eerst de gegevens grondig te onderzoeken en te vergelijken met andere factoren die de samenstelling van een visbestand bepalen. Het al dan niet overleven van de temggooi bepaalt mee de omvang van het temggooiprobleem. Natuurlijke sterfte, veroorzaakt door strenge winters, predatie en onderlinge concurrentie heeft ook een invloed op het aantal dieren dat kan doorgroeien tot commerciële afmetingen. De visserij-inspanning en de temggooi zijn echter de enige factoren die beheerd en verminderd kunnen worden, tenminste Uitgebreide samenvatting 16-182 ais daar een goeie reden toe is. Of de reden goed genoeg is, werd geëvalueerd met een nieuw ontworpen biologisch en economisch model. In het model werd berekend dat jaarlijks rond de 2.000 ton kabeljauw, 1.500 ton wijting, 12.000 ton pladijs en 600 ton tong verloren gaat door de temggooi binnen de Europese garnalenvisserij. De geschatte economische waarde van dit verlies is meer dan 25 miljoen euro. Andere soorten in de bijvangst werden niet bestudeerd wegens een tekort aan betrouwbare biologische gegevens. De modelberekeningen tonen aan dat het kwantificeren van de temggooi zoals dat doorgaans gebeurt, volgens aantal/gewicht/percentage van de vangst, misleidend kan zijn en niet altijd de belangrijkste soorten weergeeft die worden temggegooid. Op het eerste gezicht lijken bijvoorbeeld de aantallen wijting die in de Europese garnalenvisserij worden temggegooid aanzienlijk met hun 55 miljoen temggegooide dieren in 1996. Uit het model kan echter worden afgeleid dat het biologische en economische belang van temggegooide wijting relatief onbetekenend is. Wat ook blijkt, is dat de bijvangstverminderende maatregelen zich moeten concentreren op pladijs, vooral de leeftijdsgroep 0 in de Waddenzee en de leeftijdsgroep 1 in andere gebieden (zoals België). Er werd vastgesteld dat deze pladijs het grootste aandeel vormt van de temggooi binnen de Europese garnalenvisserij. Beschermende maatregelen (zoals gesloten visgronden en seizoenen of selectieve netten) die in de eerste plaats de temggooi van pladijs proberen te verminderen, zullen ook de temggooi van andere, zowel commerciële ais niet-commerciële, soorten verminderen en zo ook de bescherming van het marine milieu in het algemeen bevorderen. Het sluiten van visgronden zou waarschijnlijk de druk van de visserij opdrijven in de aanpalende gebieden die wel nog toegankelijk zijn en de voordelen die met het sluiten waren beoogd op die manier teniet doen. Een andere manier om het temggooiprobleem op te lossen, kan erin bestaan de visgronden gedurende bepaalde periodes te sluiten om zo bijvoorbeeld de leeftijdsgroep 0 van pladijs maximaal te beschermen. Er werd echter berekend dat een sluiting van 1 maand voor de Duitse visserij in het derde kwartaal slechts 11% van het geschatte verlies aan pladijs kon redden. Dit percentage opdrijven kan enkel door de gebieden voor langere periodes af te sluiten. Een ander middel is soortenselectieve netten te gebruiken die vooral pladijs van de leeftijdsgroep 0 moeten beschermen in de kraamgebieden voor platvis in de Waddenzee en leeftijdsgroep 1 in de kustwateren buiten de kraamgebieden (b.v. België). Het biologisch model toonde de doeltreffendheid aan van een selectief rooster (of gelijkaardig instrument) in de garnalenvisserij voor het verminderen van de temggooi van vis in leeftijdsgroep 1. In het model kon 90% van de verloren gegane vangst gered worden wanneer het selectief rooster werd gebmikt op plaatsen waar dieren van leeftijdsgroep 1 overheersten (b.v. in België), maar in gebieden waar vis van leeftijdsgroep 0 overheerste was dit instrument veel minder efficiënt (vooral voor platvis). Slechts 21% van de Duitse temggooi van pladijs kon worden gered met het rooster. Gezien het verschil per gebied in leeftijdssamenstelling van de temggooi, is het onwaarschijnlijk dat een enkele beheersmaatregel die geen rekening houdt met dit verschil, de voorspelde voordelen voor de pladijsbestanden en de pladijsvisserij kan waarmaken. De biologisch-economische analyse toont aan dat de Europese garnalenvisserij uniek is. De toepassing van selectieve vistuigen in de garnalenvisserij zal de winstmarge niet verkleinen. Er kan zelfs een winsttoename optreden binnen de Europese visserij in haar totaliteit. Dit fenomeen is echter onwaarschijnlijk voor andere Europese visserijen; dit is een gevolg van de unieke prijs-aanvoerverhouding die er voor garnalen bestaat. Een vermindering van de temggooi in de garnalenvisserij zal vooral ten goede komen aan de platvisvloot in de Noordzee, vooral in Nederland, Groot-Brittannië en Denemarken. De reden hiervoor is de verdeelsleutel voor de huidige Noordzee visquota. 16-183 Een volgende stap in het project was een gedetailleerde studie van de selectiviteit voor grijze garnalen en commerciële vissoorten van de kuil van het gamalennet, het net zelf en de onderpees. Deze studie werd uitgevoerd om een goede kennis te hebben van de selectieve eigenschappen van de gamalenboomkor en ais voorbereiding op de experimenten met selectieve tuigen. Voor grijze garnalen waren de selectiviteitsparameters van de kuil: L50 = 39,4 mm (37,0 - 41,4), de selectiefactor = 1,82 (1,71 - 1,91) en het selectiviteitsbereik = 11,6 mm (10,2 - 13,0). De selectiviteit van de kuil van de gamalenboomkor was erg variabel. Verschillende factoren, zoals verstoppen van de mazen, grootte van de vangst en de staat van de zee, hebben een invloed op de selectiviteit. Gemiddeld was de selectiviteit nogal laag voor garnaal; er bleven tamelijk grote hoeveelheden niet-commerciële garnaal in de vangst. De selectie van garnaal in het net (kuil niet inbegrepen) waren echter tamelijk aanzienlijk en meer garnalen konden ontsnappen via de mazen van het net dan uit de kuil. Door de kleine mazen was de selectiviteit van de kuil voor vis laag. Slechts een klein gedeelte, zelfs van de leeftijdsgroep 0, kon ontsnappen door de mazen; via de andere delen van het net kon de vis bovendien ook niet ontsnappen. Daardoor vangt de gamalenboomkor onvermijdelijk grote hoeveelheden kleine vis in kustzones en estuaria waar juveniele vis in grote dichtheden voorkomt. Het vergroten van de minimum maaswijdte om aan deze situatie te verhelpen, zou de selectiecurve voor garnaal naar rechts doen opschuiven, wat betekent dat meer commerciële garnalen verloren zouden gaan. Ook ontsnappingsvensters kunnen niet worden overwogen aangezien deze zouden leiden tot verlies van commerciële garnaal. Selectieve tuigen, zoals roosters of zeefnetten, die de garnalen scheiden van niet- commerciële dieren zijn daarom waarschijnlijk een beter instmment om de soortenselectiviteit van de gamaalboomkor te verbeteren. Een wijziging aan de onderpees is ook een mogelijkheid, vooral indien een alternatieve stimulus gebmikt wordt met een hoge visnamigheid voor garnaal en hoge selectiviteit voor andere soorten. Op basis van deze resultaten werden 3 opties geselecteerd voor verder onderzoek: 1) een selectief rooster; 2) een selectief zeefnet en 3) elektrische pulsen ais alternatieve stimulus. Om de potentiële vermindering van temggooi te onderzoeken werden meerdere experimenten in de Belgische garnalenvisserij uitgevoerd met verschillende ontwerpen van roosters. Het algemene ontwerp van het rooster was gebaseerd op het Nordmore-rooster. In een eerste fase werd een ontwerp bestudeerd met een ontsnappingsopening bovenaan in het net. Ais de samenstelling van de vangst geen verstoppingen veroorzaakte was het rooster vrij effectief. De afname van het aantal ongewervelden en vis leeftijdsklasse 1 en ouder in de vangst was beduidend. De commerciële vangst van grijze garnalen verminderde, maar meestal met niet meer dan 15%. Wanneer echter elementen in de vangst verstopping van het rooster veroorzaakten, behaalde het aantal commerciële grijze garnalen niet meer het niveau dat aanvaardbaar was voor vissers. Verstopping door zeesterren was een blijvend probleem. In de tweede fase van het project werden andere ontwerpen van roosters met ontsnappingsopening bovenaan uitgetest om dit probleem te verhelpen, maar geen enkel was succesvol. Experimenten met de opening onderaan waren succesvol en verstopping door zeesterren was vrijwel onbestaand. Daarom werd beslist dit ontwerp verder te bestuderen in een derde fase van het project. Proeven met onderzoekingsvaartuigen waren steeds succesvol. Toch functioneerde het rooster niet zoals verwacht aan boord van commerciële schepen. Verstopping door zeesterren gebeurde zelden, maar het rooster was zeer gevoelig voor verstopping door ander materiaal (zeewier, mosdiertjes, plastic zakken, groot afval, enz.) dat zeer veel voorkomt langs de Belgische kust. Groot afval en kwallen veroorzaakten dikwijls problemen wanneer een Uitgebreide samenvatting 16-184 kleine ontsnappingsopening voorzien was. De oplossing werd gezocht in een opening die groot genoeg is om kwallen en afval toe te laten te ontsnappen en klein genoeg om geen grote gedeelten van de commerciële gamalenvangst te verliezen. Toch werd die gulden middenweg niet gevonden. Hoewel de selectiviteitroosters enkele duidelijke voordelen hebben - er wordt minder vis van leeftijdsgroep 1+, niet-commerciële vis en ongewervelden gevangen en de selectiviteit van de kuil is hoger - functioneren ze niet naar behoren. Vooral in de Belgische situatie is het, door de vangstsamenstelling en de eigenschappen van visgronden, moeilijk een rooster te gebruiken en zeer onwaarschijnlijk dat vissers deze methode zouden aanvaarden. Volgens vroegere studies kan een zeefnet een gedeelte van de bijvangsten laten ontsnappen, of het nu vis of ongewervelden zijn. Het is niet vervaardigd uit hard materiaal en daarom makkelijker aanvaardbaar door vissers dan een rooster. Hierom en omdat vissers het instmment onder bepaalde omstandigheden vrijwillig gebruiken, werd het zeefnet voor verder onderzoek geselecteerd. Een zeefnet werd gedurende een jaar getest op zijn selectiviteit en er werd een evaluatie gemaakt van zijn eigenschappen. Het onderzochte zeefnet was een ontwerp dat in Nederland en Groot-Brittannië commercieel gebmikt wordt. Het verlies aan commerciële garnaal bij gebruik van een zeefnet was 15% ais de omstandigheden goed waren. Bepaalde elementen van de vangst kunnen echter de ontsnappingsopening vervormen waardoor het verlies aan commerciële garnaal kan oplopen tot meer dan 30%. Waarschijnlijk kan dit worden verhinderd door technische aanpassingen. Er werd ook een lengte-effect vastgesteld in de vangstvermindering. Ondermaatse garnaal ontsnapte gemakkelijker dan de grotere garnalen, waardoor temggooi verminderde. Het zeefnet was weinig selectief voor commerciële vissoorten met een lengte onder de 10 cm. Bij vissen vanaf 10 cm neemt de selectiviteit toe met de lengte. Vooral voor vis leeftijdsklasse 1 en ouder is dit instmment doeltreffend. Er wordt veel vis van commerciële grootte gevangen in de Belgische garnalenvisserij. Belgische gamalenvissers halen een groot deel van hun inkomen uit deze visvangst. Met een zeefnet wordt commerciële vis naar de opening geleid en ontsnapt ze. Vandaar dat het verantwoord is dat de wetgeving toelaat dat de gamalennetten worden uitgemst met een grootmazige overkuil, waardoor commerciële vis wordt gevangen, terwijl ondermaatse vis kan ontsnappen. De houding van de vissers tegenover selectiviteitsvoorzieningen is niet positief omdat elke wijziging aan hun traditionele netten praktische problemen met zich meebrengen. Het zeefnet is echter een aanvaardbaarder instmment. Vergeleken met roosters zijn zeefnetten over het algemeen minder onderhevig aan verstopping en ze blijven langer functioneel onder verschillende omstandigheden. Bovendien zijn ze minder onderhevig aan averij en het verlies aan commerciële garnaal is lager dan met het rooster. De selectiviteit van het zeefnet voor vis van leeftijdsklasse 0 is erg laag. Vandaar dat dit tuig van weinig betekenis is in gebieden waar grote hoeveelheden van deze kleine visjes worden gevangen (b.v. de Waddenzee). Het biologisch-economische model toonde aan dat de vermindering van bijvangsten in leeftijdsklasse 0 in Belgische wateren slechts een beperkt voordeel oplevert voor de commerciële visbestanden omdat de densiteit relatief laag is. Vis van leeftijdsklasse 1 en ouder beschermen daarentegen kan in de Belgische garnalenvisserij een positieve impact hebben op de visbestanden. Aangezien het zeefnet erg efficiënt is voor deze vis, is het een waardevol instmment voor het verlagen van bijvangst; de inzet ervan in Belgische wateren kan de druk op de visbestanden verlagen. Daarenboven leidt het gebmik van het zeefnet ook tot een aanzienlijk kleinere bijvangst aan ongewervelden en niet- commerciële vis, wat de algemene impact van de garnalenvisserij op het marine milieu 16-185 verlaagt. Toch is het belangrijk te erkennen dat verlies van commerciële garnaal in bepaalde seizoenen en gebieden tot financieel verlies voor de vissers leidt. De meeste maatregelen die de selectiviteit bevorderen, concentreren zich op het net. Ze zijn gericht op scheiding van de vangst of een betere filtering van de vangst. Het nadeel ervan is dat de dieren met de mazen of andere delen van het net in aanraking komen voor ze kunnen ontsnappen. Schade door contact met het net of door stress tijdens het ontsnappingsproces kan tot mortaliteit leiden. Een betere aanpak om de selectiviteit te verbeteren zou zijn om te vermijden dat ongewenste groottes en soorten het net binnenkomen. Om dit te bereiken, moeten alternatieve stimuli in de netopening gevonden worden. M.a.w. stimuli die de gewenste reactie uitlokken van de doelsoort, zonder ongewenste dieren te stimuleren. Elektrische pulsen ais een alternatieve stimulus werden gekozen ais derde optie om een selectief gamalennet te ontwikkelen. Het basisidee was om selectief een schrikreactie bij garnaal uit te lokken met elektrische pulsen en ervoor te zorgen dat de soorten die niet reageerden, kunnen ontsnappen onder de verhoogde onderpees. In het verleden werden wereldwijd experimenten met elektrische pulsen uitgevoerd. Het hoofddoel van deze experimenten was om grotere hoeveelheden van de doelsoort te vangen. Het idee in dit project was om elektrische pulsen te gebruiken om de selectiviteit van het gamalennet te verhogen, d.w.z. de bijvangst verminderen terwijl de vangst van de doelsoort gelijk blijft. Het project werd gestart doordat een Belgische reder had gezien dat meer dan 2.000 vaartuigen in de Volksrepubliek China op penaeïde garnaal visten met elektrische pulsen. In een eerste fase van het project werden gedetailleerde observatie- en overlevingstests uitgevoerd ais voorbereiding op de experimenten op zee. De tweede fase bestond uit proeven op zee waarbij de vangst van een standaardgamalennet werd vergeleken met de vangst van een elektrisch net en dit onder commerciële omstandigheden. Uit de observatietests kon het volgende worden afgeleid: • grijze garnaal reageert hevig op elektrische pulsen • de optimale pulsamplitude ligt tussen 40 en 110 V (tussenafstand van de elektrodes 50cm). Hogere en lagere amplitudes beïnvloeden de reactie van de garnalen negatief. • een hogere watertemperatuur levert een heviger reactie op • een lage lichtintensiteit levert een heviger reactie op • de maximale reactie wordt doorgaans bereikt binnen de 4 sec. na de start van de pulsen Vis en andere ongewervelden, met uitzondering van schar en tong, reageren bijna niet op de pulsen of ais ze reageren, blijven ze dicht bij de bodem. Dit betekent dat soortenselectief vissen op grijze garnaal in principe mogelijk moet zijn met elektrische pulsen ais alternatieve stimulus. De experimenten op zee gebeurden met twee pulsgeneratoren die een puls van 64V en 6Hz gaven. Deze combinatie van amplitude en frequentie geeft niet de beste reactie, maar benaderde de ideale situatie zo dicht mogelijk met de beschikbare toestellen. Met een elektrisch net kan ’s nachts (volgens laboratoriumtests) in vergelijking tot de visserij op de traditionele manier 79% van de grote garnalen worden gevangen. De prijs die betaald moet worden voor het voordeel dat deze methode oplevert voor het milieu dankzij de vermindering van temggooi en de beperking van zeebodemverstoring, zou dus raw geschat zijn dat 20% minder commerciële garnaal wordt gevangen. Mocht een pulsgenerator beschikbaar zijn die een sterkere puls kan genereren, dan zouden de voorspelde vangsten toenemen met ongeveer 10%. Aangezien de prijscoëfficiënt voor de grijze garnaal ca. -1 is, moet een afname in de gamalenvangst samengaan met een proportioneel even grote toename Uitgebreide samenvatting 16-186 in prijs van het product. Dit betekent dat het inkomen van de gamalenvisser gelijk blijft, wanneer de volledige vloot het elektrisch vissen toe zou passen. De resultaten van de zeeproeven benaderden de voorspellingen sterk. Met het selectieve elektronet met elektrodeconfiguratie A bedragen de vangstverliezen 12% en zijn niet statistisch significant. Er werd zelfs een vangsttoename vastgesteld van kleine garnalen. De zeeproeven wijzen er ook op dat de verhoogde onderpees een ontsnappingsroute creëert voor de meeste soorten die regelmatig in gamalennetten worden gevangen. Het elektrisch veld zorgt ervoor dat de garnalen hoog genoeg springen om wel gevangen te worden. Het besluit is dat het elektrisch net met verhoogde onderpees goede resultaten oplevert. Het verlies aan commerciële garnaal is klein en vooral niet-commerciële vis en ongewervelden worden in kleinere aantallen gevangen, vergeleken met de vangst van het standaardnet. Verder onderzoek moet zich toespitsen op het ontwerp van een elektrisch net met een netopening waarin lange elektroden geplaatst kunnen worden en een alternatieve klossenpees. Ais we hiermee rekening houden moet het elektrisch vissen een haalbaar alternatief zijn en bovendien aanvaardbaar zowel wat de economische belangen van de vissers betreft ais de ecologische vereisten van het mariene ecosysteem. Er moet echter rekening mee worden gehouden dat de zeeproeven in dit project slechts over een korte periode doorgingen. Factoren ais watertemperatuur, stroming, graad van activiteit van de garnalen, enz.hebben hun invloed en werden niet over hun ganse bereik uitgetest. Vooraleer deze methode gecommercialiseerd kan worden, moeten uitgebreide zeeproeven gebeuren op commerciële schepen onder de meest uiteenlopende omstandigheden. Deze studie heeft gepoogd om een verticale integratie van disciplines te maken om zo een welafgelijnd doei te bereiken, nl. de temggooi in de garnalenvisserij verminderen. De inventaris van operationele eigenschappen van de vloot en de vistuigen was een goede basis voor het project. Het bijvangstbemonsteringsprogramma en het biologisch-economisch model deden ons de temggooi in het juiste perspectief zien. Met deze gegevens hadden we argumenten om de vissers ervan te overtuigen dat een vermindering van de temggooi noodzakelijk is. Het verzoek van milieuactivisten om visgronden te sluiten of de visserij drastisch te beperken kon worden ontkracht. Er werd een compromis bereikt waarbij technische oplossingen de temggooi verminderen terwijl de inkomsten en levensstandaard van de vissers op een vergelijkbaar niveau blijven. De gedetailleerde studie naar de selectiviteit van de traditionele garnalenvisserij was een tussenstap die ais voorbereiding diende voor verder onderzoek naar soortenselectiviteit. Daarna werden technische wijzigingen aan de netten uitgetest om de bijvangst te verminderen. De laatste stap in het project was de nodige steun te verlenen om de succesvolle methodes te implementeren. Veel studies focussen zich slechts op een van deze stappen en door een gebrek aan integratie gebeurt de implementatie niet. De noodzaak om financiering te vinden voor elk opeenvolgend studieonderdeel om zo telkens te verhinderen dat de studie onderbroken zou worden, was een zware belasting voor de verantwoordelijke onderzoekers. Doorzetting en samenwerking, maar ook een portie geluk, waren noodzakelijk om dit te bereiken. De internationale samenwerking en integratie van expertise van verschillende disciplines (technologie, biologie en economie) verhoogden de waarde van het werk en onderbouwden het. Het vertrouwen werd beloond toen de Europese Commissie de nieuwe wetgeving voor de garnalenvisserij in de Noordzee met twee jaar uitstelde, nl. tot na het afronden van het DISCRAN-project. Op basis van de conclusies van het project werden het rooster en het zeefnet verplicht. De EC liet toe dat de lidstaten een periode definieerden waarin de selectieve tuigen niet gebruikt moesten worden wanneer verstopping het vissen onmogelijk maakt. Tevens werd het gebmik van grootmazige overkuilen toegelaten. 16-187 De nationale wetgeving werd in goede samenwerking met wetenschappers door de beleidsmensen in detail uitgewerkt. In België werd rekening gehouden met de besluiten van dit project en werd voor het zeefnet gekozen, terwijl een implementatie van het rooster werd afgewezen. Merk ook op dat de samenwerking met de vissers gedurende de uitvoering van het project zeer goed was. Ondanks de angst dat hun toekomst op het spei stond, namen ze een positieve en open houding aan. De nieuwe technische maatregelen werden vastgelegd in 2002. Deze maatregelen zullen ongetwijfeld de impact van de garnalenvisserij op de visbestanden en op ons ecosysteem verminderen. Waar echter gevist wordt, is invloed op het ecosysteem onvermijdelijk. Van belang is dat een compromis gezocht wordt tussen de belangen van de visserij en de kwaliteit van het ecosysteem waar de doelsoorten van de visserij leven. Het Departement Zeevisserij heeft de bedoeling om verder te blijven streven naar milieuvriendelijker visserijtechnieken. References 17-188

17 References

Albert, F., 1992. Experiments with a rigid separator grate in a shrimp trawl. DFO Gulf Region Project Summary No. 1. Alverson, D. L. and Dunlop, K., 1998. Status of world marine fish stocks. Fisheries Research Institute, University of Washington, School of Fisheries, Seattle. (FRI-UW- 9814). Alverson, D.L. and Larkins, P.A., 1994a. Fisheries: fisheries science and management. In C.W. Voigtlander (ed.), The state of the world’s fisheries resources. Proceedings of the World Fisheries Congress, Plenary Sessions, American Fisheries Society, New Delhi, India. Alverson, D.L., Murawski, S.A. and Pope, J.G., 1994b. A global assessment of fisheries bycatch and discards. FAO Technical Paper no. 339, Rome, 1994. Anon., 1968. New shrimp sorting trawl looks good in tests. Fish Business, Volume 3, No. 10, October 1968. Anon., 1973. Report of the expert consultation on selective shrimp trawls - Ijmuiden, The Netherlands, 12-14 June 1973. FAO Fisheries Report No. 139. Anon., 1987. Ministerial declaration of the second international conference on the protection of the North Sea, London, 24-25 November 1987. Anon., 1990. Ministerial declaration of the third international conference on the protection of the North Sea, The Hague, 8 March 1990. Anon., 1994. Consideration of recently tested methods for changing the selectivity of towed fishing gears. Final report to the European Commission, Contract No.AIR2 CT- 93-1656. Anon., 1995. Esjberg Declaration. Ministerial declaration of the fourth international conference on the protection of the North Sea, Esjberg, 8-9 June 1995. Anon., 1996. Construction and testing of a 1:2 model 1168 Terra Nova shrimp trawl cod-end with selectivity devices. Fishing Technology Unit report No 6/96, Memorial University of Newfoundland. Anon., 1997. Statement of conclusions from the intermediate ministerial meeting on the integration of fisheries and environmental issues. Intermediate ministerial meeting on the integration of fisheries and environmental issues, Bergen, Norway, 13-14 March 1997. Anon., 1998. Report of the study group on grid (grate) sorting systems in trawls, beam trawls and seine nets. ICES C.M. 1998/B:2. Anon., 2001a. 12th report of the scientific, technical and economic committee for fisheries. Report of the 12th meeting of STECF of the Commission of the European Communities in Brussels, 2-6 April 2001. Anon., 2001b. Green paper on the future of the common fishery policy. EC COM(2001) 135 final, ISSN 0254-1475. Anon., 2001c. Communication from the Commission to the Council and the European Parliament EC COM(2001) 143 final, ISSN 0254-1475. Anon., 2001d. Compilation of submitted inputs to the progress report to the fifth North Sea Conference. The North Sea Secretariat, 7 December 2001. Anon., 2002. Bergen Declaration. Ministerial declaration of the fifth international conference on the protection of the North Sea, Bergen, 20-21 March 2002. 17-189 • Arkley, K. and MacMullen, P., 1996. Commercial evaluation of a separator trawl in a North Sea fishery. Meeting of the ICES Working Group on Fishing Technology and Fish Behaviour - Woods Hole, USA, April 1996. • Armstrong, DW, RST Ferro, DN MacLennan, and SA Reeves. 1990. Gear selectivity and the conservation of fish. Journal of Fish Biology 37A, pp. 261-262. • Armstrong, M.J., Briggs, R.P. and Rihan, D., 1998. A study of optimum positioning of square-mesh escape panels in Irish SeaNephrops trawls. Fisheries Research, Vol. 34 (1998), pp. 179-190. • Arnold, G.P., J.D. Metcalfe, B.H. Holford and A.A. Buckley, 1997. Availability and accessibility of demersal fish to survey gears: New observations of "natural" behaviour obtained with electronic data storage tags. ICES C.M. 1997/W: 11. • Amott, S., Neil, D. and Ansell, A., 1998. Tail-flip mechanism and size dependent kinematics of escape swimming in the Brown Shrimp Crangon crangon. The Journal of Experimental Biology 201(1998), pp. 1771-1784. • Attrill & Power 2002. Climatic influence on a marine fish assemblage. Nature 417, pp. 275-278 • Bailly, D. and Collet, S. (1999). An ethical approach to re-thinking the present and the future of the Common Fisheries Policy in the European Union. Workshop in Political Theory and Policy Analysis. Bloomington: Indiana University. • Baker, C., 1973. Experimentally electric fishing voyage on DAFS research vessel Clupea. White Fish Authority FR 066. • Bary, Mck., B., 1956. The effect of electric fields on marine fishes. Marine Research, Scotland, 1. • Bastin, A., 1974. Regionale sedimentologie en morfologie van de zuidelijke Noordzee en van het Schelde Estuarium. Unpublished PhD Thesis, Katholieke Universiteit Leuven, Belgium • Berghahn, R., 1992. Gamelenfisherei, Seehunde und Möwen. Das Fischerblatt, 1992 no. 3, pp. 88-93. • Berghahn, R., Wiese, K. and Lüdemann, K., 1995. Physical and physiological aspects of gear efficiency in North Sea Brown Shrimp fisheries. Helgoländer Meeresuntersuchungen, 49 (1995), pp. 507-518. • Berghahn, R. and Vorberg, R., 1997. Shrimp fishery and nature conservation in the national park Wadden Sea of Schleswig-Holstein. Environmental Research Plan of the Federal Ministry of the Environment, Nature Conservation and Reactor Safety - Research Project 108 02 085/01 Texte 82/97. • Besançon, 1973. Review on the development of the selective shrimp trawl in the Netherlands. In: Anon., 1973. Report of the expert consultation on selective shrimp trawls - Ijmuiden, The Netherlands, 12-14 June 1973. FAO Fisheries Report No. 139. • Boddeke, R., 1965. Een beter gamaalnet. Visserijnieuws 18 (1), pp. 2-8 (1965). • Boddeke, R., 1989. Management of the Brown Shrimp (Crangon crangon) stock in Dutch coastal waters. In: J.F. Caddy. Marine Invertebrate Fisheries. Wiley & Sons, New York, pp. 35-62. • Bohl, H. and Koma, R., 1962. Mesh selection experiments with beam trawls on shrimps (Crangon vulgaris) off the German coast. ICES CM 1962, Comparative Fishing Committee, No. 43. • Bohl, H., 1963. Further investigations on the selectivity of beam trawls in the German shrimp fishery. ICES CM 1963, Comparative Fishing Committee, No. 55. • Boonstra, G.P., 1976. Ervaring onder bedrijfsomstandigheden met een elektrisch vissysteem voor de vangst van garnalen. RIVO Report T.O. 1976-06. References 17-190 • Botsford, L.W., J.C. Castilla and C.H. Peterson. 1997. The Management of Fisheries and Marine Ecosystems. Science vol. 277, pp. 509-15. • Boudreau, M., 1998. Performance of a modified Nordmore grate. DFO Laurentian Region. • Brabant, J.C., 1973. Devismes’s selective trawl for Brown Shrimp (Crangon crangon). In: Anon., 1973. Report of the expert consultation on selective shrimp trawls - Ijmuiden, The Netherlands, 12-14 June 1973. FAO Fisheries Report No. 139. • Brabant, J.C., 1973. Le chalut sélectif devismes pom la crevette grise (Crangon crangon). Rapports FAO sur le pêches,139, FIIG/R 139, pp. 30-33. • Brothers, G., 1995. Shrimp size selectivity. DFO Newfoundland Region CAFID Project summary No. 10. • Bundy, A., 2001. Fishing on ecosystems: the interplay of fishing and predation in Newfoundland-Labrador. Canadian Journal of Fisheries and Aquatic Sciences 58: pp. 1153-1167. • Burba, A. and Petrauskiene, L., 1987. Structure of the Baltic shrimp reactions in DC and pulsed current fields. Acta Hydrobiologica Litunica, Volume 6, ISSN 0208-2527, pp. 79-92. • Caddy, J.F. and Garibaldi, L., 2000. Apparent Changes in the Trophic Composition of World Marine Harvests: The Perspective from the FAO Capture Database. Ocean & Coastal Management 43 (2000), pp. 615-655. • Campos, A., Fonseca, P. and Wileman, D., 1996. Experiments with sorting panels and square mesh windows in the Portuguese crustacean fishery. Meeting of the ICES Working Group on Fishing Technology and Fish Behaviour - Woods Hole, USA, April 1996. • Carlton, J.T, Geller, J.B., Reaka-Kudla, M.L., and Norse, E. 1999. Historical extinctions in the sea. Annu. Rev. Ecol. Syst. 30, pp. 515-538. • Carstensen, P.H., 1992. Die Kutter zwischen Massen und Maschen. Das Fischerblatt, 1992 no. 7, pp. 213-218. • Clark, R.A. and Frid, C.L.J., 2001. Long-term changes in the North Sea ecosystem. Environmental Review 9, pp. 131-187. • Cook, Sinclair & Steffansson, 1997 Potential collapse of North Sea cod stocks. Nature 385, pp. 521-522. • Cooper, C., 1991. Experiments with a rigid separator grate in a shrimp trawl. DFO Scotia Fundy Region Project Summary No. 25. • Dahm, E., 1991. Doubtful improvement of the selectivity of herring midwater trawls by means of square mesh codends and constructional modifications of diamond mesh cod- ends. ICES C.M. 1991/B:2. • Dahm, E., 1999. Escape of fish species and fish of different size under the groundrope, a crosscheck to the bagnet method by alternate hauls with different roller gear. ICES Fishing Technology and Fish Behaviour Working Group, St. John’s, Canada, 19-23 April 1999. • De Groot, S.J. and Boonstra, G.P., 1974. Notes on the further development of an electrified shrimp trawl in The Netherlands. ICES C.M. 1974/B:5. • De Groot, S.J. and Boonstra, G.P., 1974. The development of an electrified shrimp trawl in The Netherlands. Journal of the Council for the Exploration of the Sea 35 (2), pp. 165-170. • Delanghe, F., Vanden Broucke, G. and Deschacht, W., 1988. Elektrische visserij op garnalen en platvis. Mededelingen van het Rijksstation voor Zeevisserij, no. 224, 1988. 17-191 • De Wilde, J.W., 1999. Effects of subsidies on the distant water and coastal fisheries of the Netherlands. In: Overcapacity, overcapitalisation and subsidies in European fisheries, Portsmouth. Proceedings of the first Concerted Action workshop on Economics and the Common Fisheries Policy: UK, 28-30 October 1998. CEMARE Miscellaneous Publication no. 44 (1999). • Dippner, J. W. 1997. Recruitment success of different fish stocks in the North Sea in relation to climate variability. German Journal of Hydrography, 49, pp. 277-293. • Erickson, D.L., Perez-Comas, J.A., Pikitch, E.K. and Wallace, J.R., 1996. Effects of catch size and cod-end type on the escapement of walleye Pollock Theragra ( chalcogramma) from pelagic trawls. Fisheries Research 28 (1996), pp. 179-196. • FAO, 2000. The state of world fisheries and aquaculture - 2000. ISBN 92-5-104492-9 . • FAO, 2002. The state of world fisheries and aquaculture - 2002. ISBN 92-5-104842-8. • Ferro, R.S.T. and O’Neil, F.B., 1994. An overview of characteristics of twines and netting that may change cod-end selectivity. ICES C.M., 1994/B:2. • Fonteyne, R., Polet, H., Van Marlen, B., Macmullen, Ph. And Swarbrick, J., 1997. Optimization of a species selective beam trawl. Abstract in: Report of the Working Group on Fishing Technology and Fish Behaviour, ICES CM 1997/B:2. • Foster, D.G. and Watson, J.W., 2003. The response of finfish to side opening grids. ICES Symposium on Fish Behaviour in Exploited Ecosystems, 23-26 June 2003, Bergen, Norway. • Francis, R. and Shotton, R., 1997. “Risk” in fisheries management: a review. Canadian Journal of Fisheries and Aquatic Sciences 54, pp. 1699-1715. • Fryer, R., 1991. A model of between-haul variation in selectivity. ICES Journal of Marine Science, 48, pp. 281-290. • Garcia, S.M. and Newton, C. 1997. Current situation, trends and prospects in world capture fisheries, pp. 3-27. In Global trends: fisheries management. Editors Pikitch, E.K., Huppert, D.D., and Sissenwine, M.P. American Fisheries Society Symposium 20, Bethesda, Maryland. • Gilis, C., 1951. L’effet de tamissage des mailles du châlut à crevettes. Annales Biologiques Vol. VIII, 1951, pp. 173-174, Belgium. • Gilson, G., 1935. Recherches sur la destruction du jeune poisson par la pêche crevettière sur les côtes de Belgique. Mémoire no. 3, Annales de l’Institut d’ Etudes Maritimes, Belgium. • Glass, C.W., Wardle, C.S., Gosden, S.J. and Racey, D.N., 1995. Studies on the use of visual stimuli to control fish escape from cod-ends. Fisheries Research 23 (1995), pp. 157-164. • Godo, O.R. and Wespestad, V.G., 1993. Monitoring changes in abundance of gadoids with varying availability to trawl and acoustic surveys. ICES Journal of Marine Science, 50(1), pp. 39-51. • Graham, N. 1997. Reduction of by-catch in the Brown Shrimp, Crangon crangon , fisheries of the Wash and Humber estuaries. PhD thesis at University of Lincolnshire and Humberside. • Graham, N. and Kynoch, R.J., 2000. Square mesh panels in demersal trawls: some data on Haddock selectivity in relation to mesh size and position. Fisheries Research, 49 (2000), pp. 207-218. • Hagena, W., 1991. Probleme und Ziele de Niedersächsigen Kleinen Hochsee- und Küstenfischerei. Das Fischerblatt, 1991 no. 12, pp. 349-359. • Hagerman, L., 1970. Locomotory activity patterns ofCrangon vulgaris (Fabr.). Ophelia, 8, pp. 255-266. References 17-192 • Haii, S., 1999. The effects of fishing on marine ecosystems. Blackwell Science Ltd. ISBN 0-632-04112-9. • Harden Jones, F.R., 1974. Objectives and problems related to research onto fish behaviour. In Harden Jones (Ed.), Sea Fisheries Research, Elek Science, London ISBN 0-470-35142-X, pp. 261-275. • Hatcher, A., 1999. The European Community’s structural policy for the fishing industry. In: Overcapacity, overcapitalisation and subsidies in European fisheries, Portsmouth. Proceedings of the first Concerted Action workshop on Economics and the Common Fisheries Policy: UK, 28-30 October 1998. CEMARE Miscellaneous Publication no. 44 (1999). • Hickey, W.M., Boulos, D.L. and Brothers, G., 1995. A study of the Influence of Lastridge Ropes on Redfish Selectivity in a Bottom Trawler. Canadian Technical Report of Fisheries and Aquatic Sciences No. 2076, November 1995. • Hillis, J.P. and Earley, J.J., 1982. Selectivity in the Nephrops trawl. ICES CM 1982/BT9. • Hodder, V.M. and May, A. W., 1964. The effect of catch size on the selectivity of otter trawls. ICNAF Research Bulletin,, pp. 28-35. • Holt, 1992. Innovative (non-traditional) marine technologies which utilize animal behaviour to enhance fishing selectivity and efficiency. Marine Technology Society MTS ’92 Washington DC 1992, Conference Proceedings Volume 1, pp. 347-353. • Horn, W., 1976. Rationalisiering der Seezungenfisherei durch Einsatz elektrifizierter Baumkurren. Informationen für die Fishwirtschaft 23(1), pp. 20-22. • Hudson, E., 1996. Red listing marine fishes. Newsletter of the Species Survival Commission, IUCN, The World Conservation Union, nos. 26-27, pp. 31-32. • Hutchings, 1996. Spatial and temporal variation in the density of northern cod and a review of hypotheses for the stock’s collapse. Canadian Journal of Fisheries and Aquatic Sciences 53, pp. 943-962. • Hutchings, 2001. Conservation biology of marine fishes: perceptions and caveats regarding assignment of extinction risk. Canadian Journal of Fisheries and Aquatic Sciences 58, pp. 108-121. • Huxley, T.H., 1884. Fisheries Exhibition Literature 4 (1884). • Huyghebaert, 1999. Het bouwen en uittesten van een impulsgenerator met een in frequentie en spanning instelbare impuls. Thesis for the degree of graduate in electricity, option electronics at the Katholieke Hogeschool Brugge - Oostende. • ICES, 1994. Report of the Study Group on research vessel noise measurement. ICES C.M.1994/B:5. • ICES, 1996. Report of the Study Group on Grid (Grate) Sorting Systems in Trawls, Beam Trawls and Seine Nets. ICES C.M. 1996/B:1. • ICES, 1999. Report of the ICES Advisory Committee on the Marine Environment, 1998. ICES cooperative research report no. 233. • ICES, 2000a. Report of the ICES Advisory Committee on the Marine Environment, 2000. ICES cooperative research report no. 241. • ICES, 2000b. Report of the Working Group on Seabird Ecology. ICES CM 2000/C04. • ICES, 2000c. Report of the Working Group on Crangon fisheries and life history. ICES CM 2000/G: 11, Ref. ACFM, B. • ICES, 2000d. Report of the FTFB Topic Group on unaccounted mortality in fisheries. ICES Fishing Technology and Fish Behaviour Working Group meeting, Haarlem, The Netherlands, 10-11 April 2000. 17-193 • ICES, 2001a. Report of the ICES Advisory Committee on Fishery Management, 2001. ICES Cooperative Research Report. • ICES, 2001b. Report of the working group on Crangon fisheries and life history. ICES C.M. 2001/G:10. • ICES, 2001c. Report of the Working Group on Marine Mammal Population Dynamics and Habitats. ICES CM 2001/ACE:01. • Isaksen, B. and Valdemarsen, J.W., 1990. Cod-end with short lastridge ropes to improve size selectivity in fish trawls. ICES, C.M. 1990/B:46 • Isaksen, B., Valdemarsen, J.W., Larsen, R.B. and Karlsen, L., 1992. Reduction of fish by-catch in shrimp trawl using a rigid separator grid in the aft belly. Fisheries Research 13 (1992), pp. 335-352. • Isaksen, B. and Valdemarsen, J.W., 1994. By-catch reduction in trawls by utilising behaviour differences. In: Fern, A. and Olsen, S., 1994. Marine fish behaviour in capture and abundance estimation. Fishing News Books, Oxford, 1994. ISBN 0-85238-211-1. • IUCN, 1996. 1996 IUCN red list of threatened animals. International Union for the Conservation of Nature (IUCN), London. • Jackson, J.B.C., Kirby, M.X., Berger, W.H., Bjomdal, K.A., Botsford, L.W., Bourgue, B.J., Bradbury, R.H., Cooke, R., Erlandson, J., Estes, J.A., Hughes, T.P., Kidwell, S., Lange, C.B., Lenihan, H.S., Pandolfi, J.M., Peterson, C.H., Steneck, R.S., Tegner, M.J., and R.R. Warner, 2001. Historical Overfishing and the Recent Collapse of Coastal Ecosystems. Science vol. 293, pp. 629-637. • Jackson, J.B.C. 2001. What was natural in the coastal oceans? Proceedings of the National Academy of Sciences 98(10), pp. 5411-5418. • Jeffery, S. and Revill, A., 2002. The vertical distribution of southern North Sea Crangon crangon (Brown Shrimp) in relation to towed fishing gears as influenced by water temperature. Fisheries Research 55 (2002), pp. 319-323. • Kaiser, M.J. and de Groot, S.J., 2000. Effects of fishing on non-target species and habitats. Blackwell Science Ltd. ISBN 0-632-05355-0. • Kelle, V.W., 1976/77. Sterblichkeit untermassiger Plattfische im Beifang der Gamelenfisherei. Meeresforschung 25 (1976/77), pp. 77-89. • Kendall, D., 1990. Shrimp retention characteristics of the Morrison soft TED: a selective webbing exclusion panel inserted in a shrimp trawl net. Fisheries Research 9 (1990), pp. 13-22. • Kessler, D.W., 1965. Electrical threshold responses of pink shrimpPenaeus duorarum, Burkenroad. Bulletin of Marine Science, 1965, 15(4), pp. 885-895. • Klyashtorin, L. B. (1998) Long-term climate change and main commercial fish production in the Atlantic and Pacific. Fisheries Research 37, pp. 115-125 • Knudsen, F.R., Enger, P.S. and Sand, O., 1994. Avoidance responses to low frequency sound in downstream migrating salmon smolt (Salmo salar). Journal of Fish Biology, 45 (2), pp. 227-233. • Kröncke, I., 1995. Long-term changes in North Sea benthos. Senckenberg. marit. 26 (1/2), pp. 73-80. • Kuro, G. and Fame, L., 1965. Un nouveau modèle de chalut sélectif pom la pêche des crevettes. ICES, Comp. Fish. Comm., report no. 56, pp. 1-9. • Kure, G., Fame, L. and Lament, T., 1965. Essaisd’un nouveau type de chalut équipé d’un dispositif de sélectivité pour la pêche des crevettes. Science et Pêche, Bulletin d’information et de documentation de l’Institut Scientifique et Technique des Pêches Maritimes, April 1965, No. 136. References 17-194 • Larsen, R.B., 1991. Selectivity experiments using a 55mm sorting grid in bottom fish trawls. Workshop on cod-end selectivity, SOAFD Marine Laboratory, Aberdeen, June 1991. • Larsen, 1992. A short history on the development of the Sort-X system for size selectivity of fish. Newfoundland and Nova Scotia Mobile Gear selectivity workshop, Canada, April 1992. Technical Report, Fisheries and Oceans, Ottawa, pp. 41-69. • Larsen, R.B and B. Isaksen 1993. Size selectivity of rigid sorting grids in bottom trawls for Atlantic cod (Gadus morhua) and haddock (Melanogramus aeglefinish). ICES marine Science Symposium 196, pp. 178-182 • Lindebo, E., 1999. A review of the fishing capacity concept. Presented at 11th EAFE conference, Dublin, Ireland, April 1999. • Lindeboom, H.J. and de Groot, S.J., 1998. Impact II - The effects of different types of fisheries on the North Sea and Irish Seabenthic ecosystems. NIOZ rapport 1998 - 1. • Lokkeborg, S. and Soldai, A.V., 1993. The influence of seismic ecploration with airguns on cod (Gadus morhua) behaviour and catch rates. ICES Marine Science Symposium, 11-13/06/1992, Bergen, Norway. • Lowry N. and Robertson J.H.B., 1996. The effect of twine thickness on cod-end selectivity of trawls for haddock in the North Sea. Fisheries Research 26 (1996), pp. 353-364. • Ludwig, D., R. Hilbom, and C. Walters. 1993. Uncertainty, resource exploitation, and conservation: lessons from history. Science vol. 260, pp. 17-36 • MacLennan, D.N., (1992) Fishing gear selectivity: an overview. Fisheries Research 13 (1992), pp. 201-204. • Madsen, N. and Moth-Poulsen, T., 1994. Measurement of selectivity of Nephrops and demersal roundfish species in conventional and square mesh panel cod-ends in the northern North Sea. ICES C.M. 1994/B:14. • Madsen, N. and Hansen, K.E., 2001. Danish experiments with a grid system tested in the North Sea shrimp fishery. Fisheries Research 52 (2001), pp. 203-216 • Main J. and Sangster G.I., 1982. A study of separating fish fromNephrops norvegicus L. in a bottomtrawl. Department of Agriculture and Fisheries for Scotland, Scottish fisheries research report no. 24, 1982. ISSN 0308 8022. • Main J. and Sangster G.I., 1985. Trawling experiments with a two-level net to minimize the undersized gadoid by-catch in a Nephrops fishery. Fisheries Research, 3 (1985), pp. 131-145. • Maksimov, Y., Malevicius, S. and Yudin, V., 1987. Selective effect of electrified field while fishing by electrotrawl. Acta Hydrobiologica Lituanica, Volume 6, ISSN 0208- 2527, pp. 54-74. • Maniwa, 1976. Attraction of bony fish, squid and crab by sound. In: Schuijf, A. and Hawkins, A.D., eds., 1976. Sound reception in fish. Elsevier, Amsterdam, pp. 271-283. • Meyer, Von, P.F. and Thiews, K., 1965. Der beifang in den Fängen der deutschen Gamelenfisherei in den Jahren 1954-1960. Berichte der Deutschen Wissenschaftlichen Kommission für Meeresforschung XVIII (1965), H., 1, pp. 13-78. • Milazzo, M., 1998. Subsidies in world fisheries: A re-examination. World Bank Technical Paper no. 406, Fisheries Series. World Bank, Washington DC. • Mistakidis, M.N., 1958. Comparative fishing trials with shrimp nets. Fishery Investigations, Series II, Vol. XXII, No. 1, 1958, Ministry of Agriculture, Fisheries and Food, UK. • Mohr, H. and Rauck, G., 1979. First results of German experiments with a selective shrimp trawl. ICES C.M. 1979/B:7. 17-195 • Myers, R.A. and Worm, B., 2003. Rapid worldwide depletion of predatory fish communities. Nature vol. 423: (15 May 2003), pp. 280-283 • Mubamba, R., 1992. The biology and exploitation of small pelagic fishes in Zambia. In: Papers presented at the symposium on biology, stock assessment and exploitation of small pelagic fish species in the African Great Lakes region. Bujumbura, Burundi, 25-28 November 1992. Marshall, B.E. and Mubamba, R., eds., FAO, Rome, 19. • NAFO., 2001a. Report of Scientific Council meeting, 31 May - 14 June 2001 of NAFO. NAFO Ses. Doc. 01/24, ser. no. N4462. • Neil, D.M. and Anseil, A.D., 1995. The orientation of tail-flip escape swimming in decapod and mysid crustaceans. Journal of Marine Biology 75 (1995), pp. 55-70. • Neilson, J. D., K. G. Waiwood and S. J. Smith. 1989. Survival of Atlantic halibut (Hippoglossus hippoglossus) caught by longline and otter trawl gear. Canadian Journal of Fisheries and Aquatic Sciences 46(5), pp. 887-897. • NOAA, 1995. Continued observer coverage of the Gulf of Mexico and South Atlantic shrimp fisheries to characterize the catch and evaluate the efficiency of by-catch reduction devices. A report of the Gulf and South Atalntic Fisheries Development Foundation, pursuant to NOAA Award No. NA47FM0131. • O'Brien, C., Fox, C., Planque, B. and Casey, J., 2000. Fisheries: Climate variability and North Sea cod. Nature 404, pp. 142 • O’Neill, F.G. and Kynoch, R.J., 1996. The effect of cover mesh size and cod-end catch size on cod-end selectivity. Fisheries Research 28 (1996), pp. 291-303 • Özbilgin H., Ferro R.S.T., Robertson J.H.B., Hutcheon J.R., Kynoch R.J. and Holtrop G., 1996. Seasonal variation in cod-end selectivity of haddock. ICES CM 1996/B: 18 • Paschen, M., Richter, U. and Köpnick, W., 2000. Trawl penetration in the seabed. University of Rostock, Final report to the European Commission, Contract No. 96-006. • Pauly, D. and Christensen, V., 1995. Primary production required to sustain global fisheries. Nature vol. 374: (6519), pp. 255-257. • Pauly, D., Christensen, V., Dalsgaard, J., Froese, R. and Torres, F., 1998. Fishing down marine food webs. Science vol. 279, pp. 860-863. • Pauly, D., Palomares, M.L., Froese, R., Sa-a, P., Vakily, M., Preikshot, D., and Wallace, S. 2001. Fishing down Canadian aquatic food webs. Canadian Journal of Fisheries and Aquatic Sciences. 58(1), pp. 51-62 • Polet, H. and Redant, F., 1994. Selectivity experiments in the Belgian Norway lobster {Nephrops norvegicus) fishery. ICES C.M. 1994/B:39 • Polet, H., 1998. Experiments with sorting grids in the Belgian Brown Shrimp{Crangon crangon) fishery. Meeting of the ICES Working Group on Fishing Technology and Fish Behaviour - La Coruna, Spain, April 1998. • Polet, H. and Redant, F., 1999. Effect of population structure, sampling strategy and sample size on the estimates of selection parameters for shrimp{Crangon crangon) trawls. Fisheries Research 40 (1999), pp. 213-225. • Polet, H., 2001. Ontwikkeling van een milieu-vriendelijke visserijmethode voor de gamaalvisserij, gebaseerd op stimulering door elektrische pulsen. Final report to the Flemish Community, Contract No. EOGFL 30-A41. • Rasmussen, B., 1973. Fishing experiments with selective shrimp trawl in Norway. In: Anon., 1973. Report of the expert consultation on selective shrimp trawls - Ijmuiden, The Netherlands, 12-14 June 1973. FAO Fisheries Report No. 139. • Redant, F., 1978. Konsumptie en produktie van post-larvale Crangon crangon (L.) (Crustacea, Decapoda) in de Belgische kustwateren. PhD thesis, Free University of Brussels. References 17-196 • Reeves, S. A., Armstrong D.W., Fryer R.J. and Coull K.A., 1992. The effects of mesh size, cod-end extension length and cod-end diameter on the selectivity of Scottish trawls and seines. ICES Journal of Marine Science 49 (1992), pp. 279-288. • Revill, A., 1997. Provisional study: The economic and biological impacts of discarding in the UK (East Coast) Crangon crangon fishery. MAFF funded study. University of Lincolnshire & Humberside. • Revill, A., Pascoe, S., Radcliffe, C., Riemann, S., Redant, F., Polet, H., Damm, Ü., Neudecker, T., Kristensen, P. and Jensen, D., 1999. The economic & biological consequences of discarding in the European Crangon fisheries. Final report to the European Commission, Contract No. 97/SE/025. • Revill A., Riemann S., Radcliffe C., Dutton K., Bower S., Jeffery S., Frid C. and Taylor K., 2000. The discarding of non marketable (undersized) Crangon crangon in the UK Crangon crangon fishery. Final report to the Chief Scientist Group of the Ministry Agriculture, Fisheries and Food (London, UK), MAFF CSG Project no. MF 0615. • Ridderstad, G., 1915. A new construction of trawl-net intended to spare under-sized fish. Svenske Hydrografisk Biologiska Kommisionens skrifter Vol 6. • Rijnsdorp, A.D., van Leeuwen, P.I., Daan, N., and Heessen, H.J.L., 1996. Changes in abundance of demersal fish species in the North Sea between 1906-1909 and 1990-1995. ICES Journal of Marine Science 53(6), pp. 1054-1062. • Robins, C.M., Wang, Y. and Die, D., 1998. The impact of global positioning systems and plotters on fishing power in the northern prawn fishery, Australia. Canadian Journal of Fisheries and Aquatic Sciences 55: pp. 1645-1651. • Roelofs, E.W., 1950. Mesh size and construction as a factor in releasing small fish and shrimp from trawl nets. U.S. Fish and Wildlife Service, Fisheries Leaflet 382. • Safina, C., 1995. The world’s imperilled fish. Scientific American, November 1995. • Sada, S., 1983. Importance and assessment of discards in commercial fisheries. UN/FAO, Rome, Italy. FAO Circ. 765, 62pp. • Sakhno, V.A. and Sadokhin, K., 1980. Investigation on bottom trawl selectivity in relation to the Barents Sea shrimp {Pandalus borealis). ICES C.M. 1980/B:7. • Stewart, P.A.M., 1975. Catch selectivity by electrical fishing system. Journal of the International Council for the Exploration of the Sea, 36(2), pp. 106-109. • Suuronen, P., Millar, R.B. and Jarvik, A., 1991. Selectivity of diamond and hexagonal mesh codends in pelagic herring trawls: evidence of catch size effect. Finnish Fisheries Research, 12, pp. 143-156. • Takemura, A., Nishida, N. and Kobayashi, Y., 1988. The attraction effect of natural feeding sound in fish. Bulletin of the Faculty of Fisheries, Nagasaki University. • Tschemij, V. and Holst, R., 1999. Evidence of factors at vessel-level affecting cod-end selectivity in Baltic cod demersal trawl fishery. ICES CM 1999/R:02. • Thatje, S. and Gerdes, D., 1997. The benthic macrofauna of the inner German Bight: present and past. Archive of Fishery and Marine Research 45/2, pp. 93-112. • Thorsteinsson, 1973. Selective shrimp trawl experiments in Icelandic waters. In: Anon., 1973. Report of the expert consultation on selective shrimp trawls - Ijmuiden, The Netherlands, 12-14 June 1973. FAO Fisheries Report No. 139. • Thorsteinsson, G., 1981. The effect of net slack in the sidepanels of shrimp trawls on the size distribution of the catch. ICES CM 1981/B:5. • Thorsteinsson, G., 1995. Survival of shrimp and small fish in the inshore shrimp fishery at Iceland. ICES Study Group on Unaccounted Mortality in Fisheries, Aberdeen, 17-18 April 1995. 17-197 • Valdemarsen, J.W., 2001. Technological trends in capture fisheries. Ocean & Coastal Management 44 (2001), pp. 635-651. • Van Beek, F. A., Van Leeuwen, P.I. and Rijnsdorp, A.D., 1990. On the survival of plaice and sole discards in the otter-trawl and beam-trawl fisheries in the North Sea. Netherlands Journal of Sea Research 26, pp. 151-160. • Van Beek, F.A., A.D. Rijnsdorp and R. de Clerck, 1989. Monitoring juvenile stocks of flatfish in the Wadden Sea and the coastal areas of the southeastern North Sea. Helgoländer Meeresuntersuchungen 43, pp. 461-477 • Vanden Broucke, G., 1972. Eerste resultaten in de electro-visserij. Mededelingen van het Rijksstation voor Zeevisserij, 68-TZ/50/1972. • Vanden Broucke, G. and Van Hee, J., 1977. Verder onderzoek over de elektrische visserij op garnalen. Mededelingen van het Rijksstation voor Zeevisserij, 133- TZ/82/1977. • Van Lancker, V., 1999. Sediment and morphodynamics of a siliciclastic near coastal area, in relation to hydrodynamical and meteorological conditions: Belgian continental shelf. Unpublished PhD Thesis, Ghent University, Belgium • Van Marlen, B., 1997a. Alternative Stimulation in Fisheries. Final Report to the European Commission, Contract No. AIR3-CT94-1850. • Van Marlen, B., Redant, F., Polet, H., Radcliffe, C., Revill, A., Kristensen, P.S., Hansen, K.E., Kuhlmann, H.J., Riemann, S., Neudecker, T. and Brabant, J.C., 1997b. Research into the Crangon fisheries unerring effect. RIVO-DLO rapport C054/97 • Van Marlen, B., de Haan, D.I., Revill, A.S., Dahm, E., Wienbeck, H., Purps, M., Coenjaerts, J. and Polet, H., 2001. By-catch reduction devices in the EuropeanCrangon fisheries. ICES C.M. 2001/RT0. • Van Middelem, A. and Cleeren, G., 1967. Een onderzoek van gamaalnetten in Beneluxverband. Werkgroep “Techniek in de Zeevisserij”, Proefstation voor Zeevisserij, rapport m. 24, 1967. • Walsh, S.J., Engas, A., Ferro, R.S.T., Fonteyne, R. and Van Marlen, B., 2000. Improving fishing technology to catch (or conserve) more fish: the evolution of the ICES Fishing Technology and Fish Behaviour Working Group during the past century. ICES Symposium on 100 years of science under ICES, 1-4 August 2000, Helsinki, Finland. • Walter, U., 1997. Quantitative analysis of discards from Brown Shrimp trawlers in the coastal area of the East Frisian islands. Archive of Fishery and Marine Research 45/1, pp. 61-76. • Wardle, C.S., 1993. Fish behaviour and fishing gear. In: Behaviour of Teleost Fishes. Pitcher, T.J., ed. Chapman and Haii, London: 609-643. • Watson, J.W. and Seidel, W.R., 1980. Evaluation of techniques to decrease sea turtle mortalities in the south-eastern United States shrimp fisheries. ICES C.M. 1980/B:31. • Wienbeck, H., 1997. First trials on the selection of a sorting grid in the commercial fishery for Brown Shrimps. ICES C.M. 1997/FF:8 • Wileman, D. A., 1992. Codend Selectivity: Updated review of available Data. Final report to the European Commission, Directorate General for Fisheries, Study Contract No. 1991/15. • Wileman, D.A., Ferro, R.S.T., Fonteyne, R. and Millar, R.B. (Eds.), 1996. Manual of methods of measuring the selectivity of towed fishing gears. ICES Cooperative Research Report. No. 215. • Will, K., 1993. Verbod der Krabbenfisherei im Nationalpark Hamburgisches Wattenmeer. Das Fischerblatt, 1993 no. 12, pp. 336-338. References 17-198 • Wulff, A., and Bückmann, A., 1935. Der Gammelfang der Gamelenfisher und die Bedeutung des Fortganges junger Plattfische für den marktfähigen Plattfishbestand in der Deutschen Bucht. Wissenschaftliche Meeresuntersuchungen, 19, pp. 1-16. 18-199

18 Appendix 1 - Glossary of terms and abbreviations

• Age class: A group of individuals of the same age in a population. The 0 group are the fish in their first year of life. A fish bom in April of a given year, remains in the 0 group until April of the following year. • Abundance: Density of a certain species on a fishing ground (e.g. no.s / 10.000 m2) • Bar spacing (for a sorting grid): The opening between two bars of a sorting grid, i.e. the gap between the bars through which animals that are small enough can penetrate. • Beam trawl: Trawl in which the horizontal opening of the net is provided by a beam, usually made of metal. Beam trawls are used mainly for flatfish and shrimp fishing. • Beam trawler: A trawler in which the fishing gear is towed from outrigger booms. Commonly used for shrimp and flatfish trawling. • Benthos; benthic organism: The biota living on or very near the bottom of the sea, river, or lake. • Bobbin rope: Groundrope rigged with bobbins (wooden or rubber cylindrical rollers). • By-catch: Part of a catch of a fishing unit taken incidentally in addition to the target species towards which fishing effort is directed. Some or all of it may be returned to the sea as discards, usually dead or dying. • Catchability: In general, the extent to which individual fish in a stock are susceptible to fishing. Catchability often increases with developments in fishing technology, and so needs to be monitored. It depends on the habits of the fish as well as on the type and deployment of fishing gear. It may also depend on the abundance of the fish (e.g. less abundant fish may be more catchable due to less saturation of gear or to concentration in schools). Specific climatic conditions may result in increased or decreased availability of the fish. This would lead to increased (decreased) catchability and, thus, increased (decreased) fishing mortality rate with the same fishing effort. • CFP: Common Fisheries Policy of the European Union • Closed area: a fishing ground where fishing is prohibited. • Closed season: a period in the year that fishing is prohibited. • Cod-end: Anterior part of the net where the catch accumulates. • Cover: Bag, similar to a cod-end, made of netting material to cover a cod-end or outlet and collect escaping animals. • CV: Commercial vessel. • D : The level of discarding by a certain fishing fleet for a certain species (%). • Demersal trawling: Bottom trawling. Operation of a trawl net designed for use on or near the bottom e.g. otter trawl, beam trawl. • DIFMAR: Danish Institute for Fisheries and Marine Research. • Discards: Part of the catch, which is not retained and is returned to the sea. Discard typically consists of "non-target" species or undersized specimens. While some species Appendix 1 — Glossary o f terms and abbreviations 18-200 (clams, sea stars, etc.) might survive the process, most fish will die. (FAO Fisheries Technical Paper 382 - X2465). • DISCRAN : Project title: Reduction of discards in Crangon trawls. • DvZ: Departement Zeevisserij; Sea Fisheries Department, Oostende, Belgium. • EC: European Commission. • ECODISC: Project title: Economic Consequences of Discarding in the Crangon Fisheries. • EU : European Union. • F: Fishing mortality, i.e. the level of fishing pressure on a certain stock. • Fishing capacity: The ability of a vessel or a fleet to catch fish. The exact fishing capacity indicator used will depend on the characteristics of the fishery or fleet and the availability of reliable data. • Fishing effort: A measure of the activity of fishing boats. Several definitions are used such as: fishing capacity x days fishing; total hours fishing per year; numbers of vessels etc. • Fishing gear: The equipment used for fishing. In the case of Brown Shrimp this usually is a beam trawl. • Fishing mortality (F). A mathematical expression of the part of the total rate of deaths of fish due to fishing. Fishing mortality is often expressed as a value that indicates the proportion of removals of the population in a year. Fishing mortality should reflect all deaths in the stock that are due to fishing, not just those fish that are actually landed. For management purposes, it is important to consider how F is distributed among age groups. • Fishing power: Of a boat or a fishing gear. Measured by its catch per unit of time, for a given density of aquatic animals. The fishing power depends on: (a) the area (or volume) affected by the gear, relative to the total area covered by the stock; (b) the number of animals present in that area (or volume), relative to the total stock ; and (c) the proportion of the animals present in that area (or volume) which can effectively be captured by the gear. • Fish stock: The living resources in the community or population from which catches are taken in a fishery. Use of the term fish stock usually implies that the particular population is more or less isolated from other stocks of the same species and hence self- sustaining. In a particular fishery, the fish stock may be one or several species of fish but here is also intended to include commercial invertebrates and plants. • Full equity return to owner (as used in Chapter 6): The profit of a vessel owner after subtracting all exploitation costs, inclusive salaries of the crew. • GPS: Global positioning system. • Gross value added (as used in Chapter 6): Is the value of the landings minus the exploitation costs (exclusive salaries of the crew). • Groundrope: Rope, usually of wire and protected by e.g. bobbins, attached to the front part of the lower panel of the net to shield the lower leading margine of a bottom trawl from ground damage whilst maintaining ground contact. • Haul: A single fishing operation, between veering and hauling the net. 18-201 • Headline: The upper frame rope to which the posterior part of the top panel of the net is attached. • ICES: International Council for the Exploration of the Sea. An international science forum founded in 1902. • IFREMER: Institut français de recherche pom l'exploitation de la mer • Industrial fishing : • IUCN: World Conservation Union (formerly the International Union for Conservation of Nature and Natural Resources). It aims to provide knowledge and guidance about conservation and the sustainable use of natural resources. • Juvenile: A young fish or animal that has not reached sexual maturity. • L25: Length at 25% retention. A parameter describing the selectivity of a device (cod- end, grid etc.), giving the length of the animal at which 75% of the animals escape while the other 25% are caught. • L50: Length at 50% retention. A parameter describing the selectivity of a device (cod- end, grid etc.), giving the length of the animal at which 50% of the animals escape while the other 50% are caught. • L75: Length at 75% retention. A parameter describing the selectivity of a device (cod- end, grid etc.), giving the length of the animal at which 25% of the animals escape while the other 75% are caught. • Landings: The part of the catch that is disembarked at a landing site. May be different from the catch (which includes the discards). • LOA: Length overall, a measure of the size of a fishing vessel • Lower panel: All net sections of the lower part of the net. • LPUE: The landings per unit of fishing effort. • M: Natural mortality (%) • MAGP: Multi-annual guidance programme. EU member states have over the years agreed on a series of multi-annual guidance programmes, aimed at reducing the EU fishing fleet's capacity to levels more in line with the opportunities to catch fish. • Mesh size: The size of holes in fishing net. Minimum mesh sizes are often prescribed by regulations in order to avoid the capture of the young of valuable species before they have reached their optimal size for capture. • Natural mortality (M): Deaths of fish from all causes except fishing (e.g. Ageing, predation , cannibalism, disease and perhaps increasingly pollution). It is often expressed as a value that indicates the proportion of fish dying in a year. • Numbers at age: The number of fish in each age class in the stock at a particular point intime. • PA: polyamide, nylon. • PE: polyethylene. • Pelagic trawling: Midwater trawling. Operation of a trawl net designed for use in the water column, without bottom contact. Appendix 1 — Glossary o f terms and abbreviations 18-202 • Price flexibility: A number indicating the relation between the amounts landed and the price. For Brown Shrimp this is “-1”, indicating that any increase in landings goes together with a similar drop in price and vice versa. • Quota: A share of the Total Allowable Catch (TAC) allocated to an operating unit such as a country, a vessel, a company or an individual fisherman (individual quota) depending on the system of allocation. • RESCUE: Project title: Research into Crangon fisheries unerring effect • RV: Research vessel • Selective gear: A gear allowing fishermen to capture few (if any) species other than the target species or few (if any) fish sizes below the minimum landing size. • Selectivity (a): Ability to target and capture fish by size and species during harvesting operations, allowing by-catch of juvenile fish and non-target species to escape unharmed. Often expressed as a relationship between retention and size (or age) with no reference to survival after escapement. • Selectivity (b): The process which causes the catch of a gear to have a different composition to that of the fish population in the geographical area in which the gear is being used. • Selectivity curve (ogive): The relationship between size (or age) and the probability of a fish escaping from the gear after having encountered it, e.g. swimming through the mesh of a net, the sorting grid of a trawl, or the escape gate of a trap. • SF: Selection factor. A parameter describing the selectivity of a device (cod-end, grid etc.), calculated as L50/mesh size. • SfD: See DvZ. Sea Fisheries Department. • Sieve net: A conical shaped net rigged in a trawl, in front of the cod-end, at the rear part connected to an outlet. The mesh size of the sieve net is larger than the mesh size in the net. (Fig. 9-1) • Sorting grid: A grate-like selective device inserted in a trawl to filter the catch and improve selectivity. (Fig. 8-1) • Spawning: Release of ova, fertilized or to be fertilized • SR: Selection range. A parameter describing the selectivity of a device (cod-end, grid etc.), calculated as L75 - L25. 18-203 • Species:

A bra alba (W. Wood) Agonus cataphractus (L.) Allotheutis subulata Asterias rubens (L.) Tellin Pogge (Lamarck) Common Starfish Witte dunschaal Hamasmannetj e European Common Squid Gewone zeester Kleine pijlinktvis

Callionymus spp. Carcinus meanas (L.) Ciliata mustela (L.) Crangon crangon (L.) Dragonet Shore Crab Five Bearded Rockling Brown Shrimp Pitvis Strandkrab Gewone meun Grijze garnaal

Ensis directus (Conrad) Eutrigla gurnardus (L.) Limanda limanda (L.) American Razor Clam Grey Gurnard Dab Amerikaans mesheft Grauwe poon Schar

Liocarcinus holsatus Melanogrammus aeglefinus Myoxocephalus scorpius (Fabricius) (L.) (L.) Swimming Crab Haddock Armed Bullhead Gewone zwemkrab Schelvis Zeedonderpad

Ophiura spp. Pagurus bernhardus (L.) Pandalus borealis (Kröyer) Platichthys flesus (L.) Brittle Star Hermit Crab Northern Shrimp Flounder Slangster Heremietkreeft Noorse garnaal Bot Appendix 1 — Glossary o f terms and abbreviations 18-204

Pleuronectes platessa (L.) Pomatoschistus spp. Psetta maxima (L.) Raya spp. Plaice Goby Turbot ray Schol, Pladijs Dikkopje Tarbot rog

Sepiola atlantica Solea solea (L.) Spisula subtruncata (da Echei ichthys vipera (d’Orbigny) Sole Costa) (Cuvier) Atlantic Bobtail Tong Cut Trough Shell Lesser Weever Dwerginktvis Afgeknotte strandschelp Kleine Pieterman

Trisopterus luscus (L.) Bib Steenbolk

Sources o f figures: o ADEMA, J.P.H.M. (1991). Krabben van Nederland en België. Publ. by Nationaal Historisch Museum, Leiden, The Netherlands. o HAYWARD, P.J. and RYLAND, J.S. (1990). The marine fauna of the British Isles and North-West Europe. Publ. by Oxford Science Publications, Oxford, England. o POLL, M. (1947). Poissons marins. Ed. Musée Royal d’Histoire Naturelle de Belgique, Brussels

• SSB: Spawning stock biomass. The total weight of all sexually mature fish in the population (both males and females). This quantity depends on the abundance of year classes, the exploitation pattern, the rate of growth, both fishing and natural mortality rates, the onset of sexual maturity, and enviromnental conditions. • Stock: see fish stock. • Stock assessment: A judgement made by a scientist or scientific body on the state of a resource, such as a fish stock (e.g., size of the stock, potential yield, whether it is over- or underexploited), usually for passing advice to a management authority. • Swept area: The area of the sea floor over which the gear (usually a trawl, or a dredge) is dragged during its operation . The area is equal to the effective horizontal opening of 18-205 the gear multiplied by the distance the gear has covered during the period of time considered. • TAC: Total allowable catch. The TAC is the total catch allowed to be taken from a resource in a specified period (usually a year), as defined in the management plan. The TAC may be allocated to the stakeholders in the form of quotas as specific quantities or proportions. • Target species: Those species that are primarily sought by the fishermen in a particular fishery. The subject of directed fishing effort in a fishery. There may be primary as well as secondary target species • TED: Turtle excluder device. Selective device often rigged in tropical shrimp trawls, guiding sea turtles to an escape outlet. • TL: total length • Top panel: All net sections of the upper part of the net. • Tow: see “Haul”. • Trawl: A eone or funnel-shaped net that is towed through the water by one or more vessels. • Undersized: Fish (caught) at a size smaller than the minimum size limit established by regulation. • W: Fresh weight. • Warp: Long flexible steel rope connecting the vessel to the trawl. Appendix 2 — Curriculum vitae 19-206

19 Appendix 2 - Curriculum vitae

Hans Polet (°1965) graduated at the Ghent University as MSc in Agriculture and Applied Biological Sciences (Forestry) in 1988 with a Masters thesis entitled “Biomass production of Prosopis juliflora (L.) in relation to ecological factors on the Cape Verde Islands”. After military service and 9 months interim work, he became a member of the research team of the Gear Technology Section at the Sea Fisheries Department (Agricultural Research Center, Ghent) in Oostende (Belgium) in 1990. In this function, he has been involved in research in the field of selectivity of fishing gear and discard reduction, environmental impact of sea fisheries and fishing techniques. Fishing gear research aims at the development of efficient trawls from a technical, biological, ecological and economical point of view. In this discipline, he has been an active researcher and coordinator of several EU and national research projects. Fieldwork aboard commercial and research vessels and contacts with fishermen and scientists has given him a wide expertise with fishing gear and fishing techniques. The impact research evaluates the ecological consequences of fishing activities. This is the basis for the development of fishing techniques that reduce the disturbance of the seafloor and the inhabiting fauna. He had an active role in the key-projects IMPACT I and II that studied the environmental impact of trawls in the North Sea and Irish Sea and that initiated a world wide interest in fisheries impact. Further work was done in consequent EU-projects dealing with seafloor and fauna disturbance of flatfish and shrimp beam trawls. The selectivity research aims at enhancement of the catches of trawls, i.e. maintaining the commercial catches while minimising the discards of non-commercial fish and invertebrates and undersized commercial fish. The main accomplishments were the development of a species selective flatfish beam trawl, the improvement of the flatfish beam trawl with reduced benthic by-catch, the study of sorting grids in the Nephrops fishery, the evaluation of sorting grids and sieve nets in the Brown Shrimp fishery and the development of an electro-shrimp trawl. Internationally, the research supports the recommendations of the International Council for the Exploration of the Sea (ICES) and the European Common Fisheries Policy, i.e. defining the technical measures for the protection of fish stocks. Besides the work focused on fishing gear, he has been active in fisheries research in a wider sense: the development of conceptual balancing model “Sustainable Management of the North Sea”; the design and implementation of a selectivity database; the development of a new net mesh measuring device. Throughout 13 years of research at the Sea Fisheries Department, he got actively involved in following projects: • SOBETRA I: FAR-TE 2.554: Improving the selectivity of flatfish beam trawls in the North Sea - Belgium, Netherlands, UK (1991-1992) • IMPACT I: FAR-MA 2.549: Environmental Impact of Bottom Gears on Benthic Fauna in Relation to Natural Resources Management and Protection of the North Sea - Belgium, Germany, Netherlands, UK (1992-1993) • EFFORT: AIR-CT92-0445: Investigation of the relative fishing effort exerted by towed Demersal gears on North Sea human consumption species" - Belgium, Denmark, UK (1992-1995) • NEPHROPS: BIO 1992.3: Selectivity and discards in the Belgian Nephrops fishery - Belgium (1993-1993) 19-207 • SOBETRA II: AIR2-CT93-1015: Optimisation of a species selective beam trawl" - Belgium, Netherlands, UK (1993-1996) • ALTSTIM: AIR3-94-1850: Alternative stimulation in fisheries - Belgium, Germany, Finland, Netherlands, Norway, UK (1994-1996) • IMPACT II: AIR2-94-1664: The effects of different types of fisheries on North Sea and Irish Sea benthic ecosystems" - Belgium, Germany, Ireland, Netherlands, UK (1994-1997) • RESCUE: BIO 94.C 144.04: Quantification of discards in the Brown Shrimp fishery - Belgium, Denmark, Germany, France, Netherlands, UK (1995-1997) • REDUCE: Fair CT97 3809: Reduction of adverse environmental impact of demersal trawls - Belgium, Germany, Ireland, Netherlands (1998-2000) • TRAPESE: Biological Study: Determining the trawl penetration in the seabed - Belgium, Germany, Netherlands (1997-1999) • SELDAT I: FAIR CT96 1531: Selectivity database - Belgium, Denmark, Germany, France, Greece, Netherlands, Norway, Portugal, Sweden, UK (1996-1998) • ELEKTRO: 5BW.EOGFL 30B-A.4.1: Ontwikkeling van een milieuvriendelijke visseriimethode voor de gamaalvisserii, gebaseerd op stimulering door elektrische pulsen - Belgium (1997-2000) • ECODISC: Economic Study - 97.SE.025: Economic consequences of discarding in the Crangon fisheries - Belgium, Denmark, Germany, UK (1998-1999) • NETRASEL: Fair PL-98-4164: Nephrops trawl discard reduction using activating selection grids - Belgium, Greece, Norway, UK (1999-2001) • DISCRAN: Biological Study - 98.012: Reduction of discards in crangon trawls - Belgium, Germany, Netherlands, UK (1999-2001) • SELDAT II: FAIR CT98-4044: Selectivity database 2 - Belgium, Denmark, Germany, France, Greece, Netherlands, Norway, Portugal, Sweden, UK (1999-2002) • OMEGA: QLRT-2001-01335: Development and testing of an objective mesh gauge - Belgium, Germany, France, Italy, Netherlands, Spain, UK (2002-2006) • RECOVERY: QLRT-2001-00935: Research on effective cod recovery measures by enhancing the selectivity of trawls - Belgium, Denmark, Netherlands, Norway, UK (2002- 2005) • BALANS: DWTC-project: Afweging van de menselijke activiteiten in het Belgisch deel van de Noordzee - Belgium (2002-2006) • NECESSITY (accepted proposal): Research on effective cod recovery measures by enhancing the selectivity of Nephrops trawls - Belgium, Denmark, Finland, France, Greece, Ireland, Italy, Netherlands, Norway, Portugal, Spain, Sweden, Turkey, UK

Publications

Chapters in books • Bergman, M.J.N., Craeymeersch, J.A., Polet, H. and Van Santbrink, J.W., 1998. Fishing mortality in invertebrate populations due to different types of trawl fisheries in the Dutch sector of the North Sea in 1994. In “The effects of different types of fisheries on the North Sea and Irish Sea benthic ecosystems”. NIOZ-Rapport 1998-1, RIVO-DLO Rapport C003/98 Appendix 2 — Curriculum vitae 19-208 • Polet, H., Ball, B., Blom, W., Ehrich, S., Ramsay, K. and Tuck, I., 1998. Fishing gears used by different fishing fleets. In “The effects of different types of fisheries on the North Sea and Irish Sea benthic ecosystems”. NIOZ-Rapport 1998-1, RIVO-DLO Rapport C003/98 • Polet, H. and Blom, W., 1998. Size of bottom trawling fleets. In “The effects of different types of fisheries on the North Sea and irish Sea benthic ecosystems”. NIOZ-Rapport 1998-1, RIVO-DLO Rapport C003/98

Articles published in international peer reviewed journals • Polet, H. and Redant, F., 1999. Effect of population structure, sampling strategy and sample size on the estimates of selection parameters for shrimp(Crangon crangon) trawls. Fisheries Research 40 (1999) 213-225 • Polet, H., 2000. Cod-end and whole trawl selectivity of a shrimp beam trawl used in the North Sea. Fisheries Research 48 (2000) 167-183 • Polet, H., 2002. Selectivity experiments with sorting grids in the North Sea Brown Shrimp (Crangon crangon) fishery. Fisheries Research, 54 (2) (2002) pp. 217-233 • Fonteyne, R. and Polet, H., 2002. Reducing the benthos by-catch in flatfish beam trawling by means of technical modifications. Fisheries Research, 55 (1-3) (2002) pp. 219-230

Articles published in journals without peer review • Fonteyne, R. en Polet, H., 1995. Ontwikkeling van een species selectieve boomkone. Mededelingen van het Rijksstation voor Zeevisserij, 1995, nr. 236. • Polet, H. en Fonteyne, R., 1995. Huidige vistuigen en visserijmethodes in de Belgische Zeevisserij. Mededelingen van het Rijksstation voor Zeevisserij, 1995, nr. 237. • Polet, H. and Fonteyne, R., 2001. Selectief vissen in het teken van de kabeljauwcrisis. Vis en Visie 2 (2), 2-4. • Polet, H., 2001. Roosters en zeefnetten in de gamaalvisserij. Vis en Visie 2 (3&4), 10-12 • Redant, F. en Polet, H., 2002. De gamaalvisserij: een kustgebruikersgroep met kopzorgen. De Grote Rede, Nr. 5, augustus 2002.

Articles published in proceedings of international congresses with oral presentation • Polet, H. and Fonteyne, R., 1993. Development of a species selective beam trawl. 2nd series of experiments - September-October 1992. ICES Fish. Technol. Fish. Behav. Work. Group Meeting, Gothenburg, Sweden, april 1993. • Van Marlen, B., Fonteyne, R., Polet, H. and Arkley, K., 1993. EC-Project TE-2-554 "Improved selectivity of fishing gears in the North Sea fishery - Beam trawling". ICES, C.M. 1993/B:13. • Polet, H. and Redant, F., 1994. Selectivity experiments in the Belgian Norway lobster (Nephrops norvegicus) fishery. ICES Fish. Technol. Fish. Behav. Work. Group Meeting, Montpellier, France, april 1994. 19-209 • Polet, H., 1994. Beam trawl selectivity experiments with cod-end covers equipped with hoops. ICES Fish. Technol. Fish. Behav. Work. Group Meeting, Montpellier, France, april 1994. • Polet, H. and Redant, F., 1994. Selectivity experiments in the Belgian Norway lobster (.Nephrops norvegicus) fishery. ICES C.M. 1994/B:39 • Redant, F. and Polet, H., 1994. Introduction on the finfish by-catches and discards in the Belgian Norway lobster (Nephrops norvegicus) fishery. ICES C.M. 1994/G:29 Ref. K • Polet, H. and Redant, F., 1996. Effect of population structure, sampling strategy and sample size on the estimates of selectivity parameters for shrimp Crangon ( crangon) trawls: preliminary results ICES Fish. Technol. Fish. Behav. Work. Group Meeting, Woods Hole, USA, april 1996. • Polet, H., 1996. Research programme on the reduction of by-catches in the Belgian shrimp fishery. Phase 2: Selectivity of the commercial shrimp beam trawl ICES Fish. Technol. Fish. Behav. Work. Group Meeting, Woods Hole, USA, april 1996. • Polet, H. and Redant, F., 1996. Effect of population structure, sampling strategy and sample size on the estimates of selection parameters for shrimp Crangon ( crangon) trawls ICES C.M. 1996/B:29 Ref. D, K • Fonteyne, R., Polet, H., Van Marlen, B., Macmullen, Ph. and Swarbrick, J., 1997. Optimisation of a species selective beam trawl. ICES Fish. Technol. Fish. Behav. Work. Group Meeting, Hamburg, Germany, april 1997. • Polet, H., 1998. Experiments with sorting grids in the Belgian Brown ShrimpCrangon ( crangon) fishery. ICES Fish. Technol. Fish. Behav. Work. Group Meeting, La Coruna, Spain, april 1998. • Polet, H., 1998. Cod-end and whole trawl selectivity of shrimp beam trawls. ICES Fish. Technol. Fish. Behav. Work. Group Meeting, La Coruna, Spain, april 1998. • Polet, H. and Ghesquiere, K., 1999. Possibilities to reduce discarding in the Belgian Brown Shrimp (Crangon crangon) fishery by means of electric pulses as an alternative stimulation. ICES Fish. Technol. Fish Behav. Work. Group Meeting, St. John’s, Canada, april 1999. • Fonteyne R., Polet H., 2000. Reducing the benthos by-catch in flatfish beam trawling by means of technical modifications. ICES Fish. Technol. Fish Behav. Work. Group Meeting, Haarlem, Netherlands, april 2000 • van Marlen, B., de Haan, DI, Revill, A.S., Dahm, E., Wienbeck, H., Purps, M., Coenjaerts, J. and Polet, H., 2001. By-catch reduction devices in the European Crangon fisheries. ICES C.M. 2001/R:10

Poster presentations • Polet, H., 1996. Research programme on the reduction of by-catches in the Belgian shrimp fishery. Phase 2: Selectivity of the commercial shrimp beam trawl ICES C.M. 1996/B:30 Ref. K - Poster • Polet, H., 2001. The discarding problem in the North Sea Brown Shrimp (Crangon crangon) fishery. Proceedings of the 7th PhD symposium, 10/10/01, Ghent University, Ghent. Fac. Landbouww. Univ. Gent, 66/4,2001. Appendix 2 — Curriculum vitae 19-210 • Polet, H., 2003. H. Development of a selective shrimp beam trawl with electric pulses as an alternative stimulation. ICES/FAO Symposium on "Fish Behaviour in Exploited Ecosystems", 23-26 June 2003, Bergen, Norway. (“Best Poster Award”)

Author or co-author of reports of contract studies • Anon., 1992. Improved Selectivity in the North Sea Fishery - Beam Trawling. Final report to the European Commission; FAR project Contract No FAR-TE 2.554. • Polet, H. and Redant, F., 1993. Selectivity and discards in the Nephrops fishery - Belgium. Final report to the European Commission; Biological Study Contract No 1992/003. • Anon., 1996. Optimisation of a species selective beam trawl. Final report to the European Commission; AIR project Contract No AIR2-CT93-1015. • Van Marlen, B., Redant, F., Polet, H., Radcliffe, C., Revill, A., Kristensen, P.S., Hansen, K.E., Kuhlmanii, H.J., Riemann, S., Neudecker, T. and Brabant, J.C., 1998. Research into the Crangon fisheries unerring effect. RIVO-DLO rapport C054/97 • Revill, A., Pascoe, S., Radcliffe, C., Riemann, S., Redant, F., Polet, H., Damm, U., Neudecker, T., Kristensen, P.S. and Jensen, D., 1999. Economic Consequences of Discarding in the Crangon Fisheries. Final report to the European Commission; Economic Study - 97/SE/025 • van Marlen, B., de Haan, DI, Revill, A.S., Dahm, E., Wienbeck, H., Purps, M., Coenjaerts, J. and Polet, H., 2001. Reduction of discards in crangon trawls (DISCRAN). Final report to the European Commission; FAIR project (EC) Contract No FAIR 98/012 • Radcliffe, C., Graham, N., Kalianotis, A., Polet, H., Galbraith, D., 2001. Nephrops trawl discard reduction using activating selection grids. Final report to the European Commission; FAIR project (EC) Contract No Fair CT-98-4164 • Polet, H. en Delanghe, F., 2001. Ontwikkeling van een milieu-vriendelijke visserij methode voor de gamaalvisserij, gebaseerd op stimulering door elektrische pulsen. Final report to the European Commission; 5b project Contract No 5BW/EOGFL 30B-A.4.1. • Anon., 2002. Reduction of the adverse environmental impact of demersal trawls. Final report to the European Commission; FAIR project (EC) Contract No FAIR CT-97-3809.

Other • Anon., 2002. Incidental catch of small cetaceans. Subgroup on Fishery and Environment (SGFEN) of the Scientific, Technical and Economical Committee for Fisheries (STECF), doc. SEC(2002) 376, Bmssels, 10-14/12/01.

Scientific activities

International congresses and symposia o Annual Science Conference of ICES: 19-211 La Rochelle, Frankrijk September - October 1991 Rostock, Duitsland September - October 1992 Dublin, Ireland September - October 1993 1 presentation St. John's, Canada September - October, 1994 2 presentations Aalborg, Denemarken September 1995 Rejkjavik, IJsland September - October, 1996 1 presentation, 1 poster Baltimore, VSA September - October 1997 rapporteur at the Fisheries Technology Committee Bmgge, België September 2000 Oslo, Noorwegen September - October 2001 co-author of 1 presentation

Working Group meetings (FTFB, FAST) of the ICES Fisheries Technology Committee: Ancona, Italië April 1991 Bergen, Norway June 1992 Gothenburg, Sweden April 1993 1 presentation + rapporteur Montpellier, France April 1994 1 presentation Aberdeen, UK, April 1995 active member of the sub-group “Evaluation of recent Nephrops selectivity experiments” Woods Hole, USA April 1996 1 presentation, 1 poster active member of the sub-group “Grid Sorting Systems” La Coruna, Spain April 1998 2 presentations active member of the sub-group “Grid Sorting Systems” St. John’s, Canada April 1999 1 presentation Haarlem, Netherlands April 2000 1 presentation Seattle, USA April 2001 Bergen, Norway June 2003

EC (FAR) Workshop on Codend Selectivity Aberdeen, UK, June 1991 ICES Symposium on Fish Behaviour in Relation to Fishing Operations. Bergen, Norway, June 1992 7de PhD symposium, 10/10/01, Faculty of Agricultural and Applied Biological Sciences Ghent University, Gent. 1 poster ICES/FAO Symposium on "Fish Behaviour in Exploited Ecosystems", 23-26 June 2003, Bergen, Norway. 1 poster - “Best Poster Award” International meetings for scientific advice organized by the European Commission Brussels, Belgium February 2001 Expert meeting on the hake recovery plan Brussels, Belgium May-June 2001 Expert meetings on cod-recovery plan Brussels, Belgium October 2001 Scientific expert meeting on future research in relation to additional improvement of the exploitation pattern of demersal species in the North Sea Appendix 2 — Curriculum vitae 19-212 Brussels, Belgium December 2001 SGFEN - Expert meeting on incidental catches of small cetaceans Bmssels, Belgium March 2002 Meeting on evaluation of recovery plans o Meetings of the Scientific, Technological and Economical Committee for Fisheries (STECF) of the European Commission November 1997, April and November 1998, April and November 1999, April and November 2000, April 2001, April and November 2002, April 2003.

Evaluation of project applications

o The EU FAIR-programmea (IVthe Framework programme), June 1996.

o The EU “Promotion of innovation and encouragement of SME participation” (Vthe Framework programme) March 2001.

o The EU “Marie Curie” programme (Vthe Framework programme) May 2001.

o The EU “Quality of Life” programme (Vthe Framework programme) November 2001.