. '-"'- \V/ VVj

VlAAMS JNSTITUUT VOOR ~

FLANDERS MARINE INSTI\fISHERIES

Oostende - Belgium cpESEARCHELSEVIER Fisheries Research 40 (1999) 135-152

Fishing effects in northeast Atlantic shelf seas: patterns in fishingeffort, diversity and community structure. IV. Can comparisons of

species diversity be used to assess human impacts ondemersal fish faunas?

S.I. Rogersa.*, D. MaxweUa, A.D. Rijnsdorpb, U. Dammc, W. Vanheed

'CEF:4.S.Lowestoft Loboratory, Lowestoft, UKbNetherlands Institute for Fisheries Investigations. PO Box 68. 1970 AB Ijmuiden. Netherlands

cBundesforschun~sallstaltfiir Fischerei. 22727 Hamburg. GermanydFisheries Research Station. Ankerstraat I. 8400. Oostende, Belgium

Abstract

Patterns in the abundance of commercially important and non-target demersal fish species collected by beam trawl surveyfrom the coastal waters of the northeast Atlantic are described. Catches were dominated by a small number of species. whichoccurred in large numbers and at high biomass. The most abundant species (plaice and dab) were typical of shallow, uniformsandy and muddy seabed which occurred extensi\'eIy throughout the southern North Sea, and to a limited extent in UK westernwaters. Renyi's diversity index family was used to rank the diversity of coastal sectors throughout the region. The less species-rich North Sea fauna. partly a result of the uniform nature of the seabed. was largely responsible for lower diversity of NorthSea coastal faunas compared to those in the Channel and west of the UK. West of the Dover Strait. the more heterogeneoussubstrate supported a more diverse fauna of smaller sized fish, with the occurrence of southern species such as red gurnard andthickback sole and an increasing abundance of elasmobranchs. In the Irish Sea, fish biomass was dominated by plaice and dab,but to a lesser extent than on the continental coast of the North Sea. Sole, lesser spotted dogfish and cod were also important inthis assemblage. Patterns in community structUre over such a wide spatial scale, and without historical perspective, can beexplained by biogeographic factors, seabed structUreand the influence of regional hydrography. Inferring from these patternsan impact of anthropogenic factors (such as towed fishing gears) is unlikely to be achieved. Identifying vulnerable species, anduse of fishing effort distribution data of high resolution, may be a more fruitful approach. Crown copyright (Q 1999Publishedby Elsevier Science B.V. All rights reserved.

Keywords: Diversity; Fishing impact; Assemblage: Demersal; Beam trawl

1. Introduction

.Corresponding author. Tel.; +44-01-502-5622-44: fax: +44-01-502-5138-65; e-mail: s.i.rogers@cefas.co.uk

Recent interest in the impact of human activities onmarine demersal communities, particularly by toweddemersal fishing gears, has focused on the results of

0165-7836/99/$- see front matter Crown copyright!: 1999 Published by E1sevier Science B.v. All rights reserved.PH: SOI65-7836(98)00209-4

136 51 Rogers et al./Fisheries Research 40 (1999) 135-/52

experimental studies quantifying impacts on a rangeof seabed faunas (Bergman and Hup, 1992; Kaiser andSpencer, 1996; Collie et aI., 1997), and on interpretingspatial and temporal changes to seabed assemblages(Pope and Knights, 1982; ICES, 1992; Greenstreet andHall, 1996; Rijnsdorp et aI., 1996b). Considerableeffort has also been devoted to developing moresophisticated methods for describing communitystructure and identifying subtle changes to it, inaddition to the classical diversity indices which arealready widely used by ecologists. For example,improvements in the measurement and interpretationof diversity have recently been made using methods ofdiversity ordering (Tothmeresz, 1995), where a rangeof diversity indices within a family show varyingsensitivities to rare and abundant species. The useof the taxonomic relatedness of species within thecommunity has also proved to be an effective means ofidentifying changes to some infaunal communities(Somerfield and Clarke. 1995). Despite this progress,basic descriptions of demersal fish faunas comprisingboth commercially exploited species as well as thenon-target component, are not yet widely available formany coastal and offshore regions (but see for exam-ple Fargo and Tyler. 1991; Farina et a1., 1997),although their value with respect to understandingpossible fishing effects may be high. The distributionof selected species has been presented in most detailfor the North Sea (Daan et a1., 1990: Knijn et aI.,1993), but difficulties with combining catches bygears with different selectivity have prevented morewidespread analyses beyond this region.

One of the most likely direct effects of towed gearson seabed communities will involve changes in therelative abundance of predators and/or their prey, andthe reduction in abundance of vulnerable species. Thesize structure of the fish assemblage will also bestrongly affected by fishing pressure (Rice and Gisla-son, 1996: Rijnsdorp and Leeuwen. 1996). Analysis ofNorth Sea populations (Greenstreet and Hall, 1996)has shown that while changes in the demersal com-munity since the 1930s were evident, only subtlechanges had occurred to the non-target assemblage.Analysis of North Sea survey data since the 1970s alsoshowed changes in diversity and evidence of somecommunity-level effects of fishing, especially duringthe 1980s (Rice and Gislason, 1996). A decline in the

abundance of non-target species such as grey gurnard

Eutrigla gurnardus has been observed in the NorthSea (Rijnsdorp et aI., 1996b), and also in some elas-mobranch species including, for example, spurdogSqualus acanthias (Brander, 1981; Heessen and Daan,1996; Walker and Heessen, 1996). Some of the elas-mobranchs are considered to be particularly vulner-able to increased fishing mortality because of their lowfecundity and high age and length at maturity (Walkerand Heessen, 1996). Finding causal relationshipsbetween human impacts such as trawling or eutrophi-cation, and temporal trends in abundance of species,has been difficult. Interpreting the likely effects ofthese impacts on community structure over a largespatial scale, covering many degrees of latitude andlongitude, will be even more complex.

There are many factors which will influence thedistribution and abundance of fish, and one of the most

fundamental is regional biogeography. Although it isassumed that the tropics contain by far the greatestdiversity of genera and species of fish (Norman. 1963),there are few references to the precise rate of decreasein species richness between ul.e equatorial and polarfaunas (Ekman. 1967: Henderson, 1989). Indeed.other marine groups, such as seaweeds, reach theirgreatest diversity in temperate regions, and there areequal numbers of infaunal invertebrate species in softsubstrates in Arctic and tropical regions (Gee andWarwick, 1996). So, an understanding of the biogeo-graphic influences which control fish population diver-sity will underlie subsequent interpretations whichrelate to possible changes caused by human activity.

The recent development within the InternationalCouncil for the Exploration of the Sea (ICES) of anextensive database describing the beam trawl catchesof demersal fish throughout the coastal waters of thenortheast Atlantic (Rogers et aI., 1998), has provided aunique opportunity to further develop our knowledgeof fish populations. Results from an initial analysisgrouped assemblages into broad categories, two ofwhich represented areas of relatively high diversity inthe western waters of the UK and off Denmark, andother less diverse assemblages in the eastern Channeland the southern North Sea (Rogers et aI., 1998).

This paper develops this initial classification anddescribes the results of more detailed comparisons ofthe diversity of demersal assemblages betweenregions. Its purpose is to identify aspects of commu-nity structure which. on a large spatial scale, are most

q'",-o-",",

5.l. Rogers et al./ Fisheries Research 40 (1999) 135-152

likely to reveal evidence of anthropogenic impact.Together with more detailed information of the dis-tribution of trawl effort, these data could in future be

used for more precise examination of casual relation-ships.

2. Materials and methods

2.1. The surveys

The beam trawl surveys of the UK, Netherlands,Germany and Belgium which contribute to this data

50'

SS'

56'1

54'

52"

50"

CelticSea

49'

(a)

137

set cover an area comprising the southern and easternNorth Sea (ICES Division !Vb and IVc), the EnglishChannel (ICES Division VIId and VIle), the BristolChannel (ICES Divisions VIIf and VIlg) and the IrishSea (ICES Division VIIa). Although the earliest beamtrawl survey began in the North Sea in 1985, aconsistent data set for all participating countries isonly available from 1990. For the analyses presentedin this paper, data for the entire area from 1990 to 1996inclusive have been used, and the ICES Divisions in

which sampling occurred have been subdivided intosmaller coastal sectors which generally group fishingstations with the nearest coastline (Fig. l(a».

Fig. 1. Map of the study area showing ICES Divisions subdivided into coastal sectors, and place names referred to in the text (a), and theposition of sampling stations (b).

138 S.l. Rogers et aUFisheries Researr:h40 (1999) /35-/52

58

57

56

55

48'1

-8 -4 -2-6

(hI

x xx xx x xx x x

x xx x

x x ,x x x

x

x xx x x

x x XX X XX X X X XX X XX X X X

xxxxxx x x x x x x

x x x x x xx x x x x x x

x x x x x xx x X X X XX X X X X X

X X x.r~-'~

~

!~ ~

x xx xxx,

0 6

I

i!II

I

II

~

102 4 8

LongitudeFig. L (continued)

Although the precise details of the sampling metho-dology have been described elsewhere (ICES. 1994;Rogers et aI., 1997), the following summary describesthe basic sampling procedure.

Beam trawls were used to survey the demer sal fishpopulations of the coastal waters (10-120 m depth) ofnorthwestern Europe. and different sized trawls wereused during the surveys. depending on the ability ofthe different vessels to deploy them. and the varyingnature of the seabed in different parts of the region(Rogers et al.. 1998). All samples were collectedduring the third quarter of the year. In general. fishing

stations sampled were at fixed positions, except inDutch and German surveys where the station positionswere stratified by ICES rectangle and selectedannually on a pseudo-random basis. The positionsof the fixed stations in UK waters were stratified bydepth band and by ICES rectangle. with the numberofstation positions in each depth band allocated inproportion to the catch of pre-recruit flatfish duringpreliminary surveys. Station positions of the Belgiansurveys in the North Sea were stratified by rectangle.In the North Sea, where station positions fished bysome countries were not fixed from year to year, all

C~_C'ccC --'_c ~~ ~..,c.~.

51. Rogers et aLIFisheries Research 40 (1999) 135-152

hauls were grouped by the quarter ICES rectangle inwhich the shooting position occurred, and the meancatch by species of these hauls was allocated to anominal fixed position at the centre of the quarterrectangle (Fig. l(b)).

At each station the fin-fish catch was identified to

species where possible, measured, and the numbersrecorded. The weight of each species caught at eachhaul during UK surveys was recorded. The meanweight of an individual of each of these species, basedon the total weight and total number of the speciescaptured, was used to estimate the weight of catches inNorth Sea surveys. Larger catches were subsampled,and more complex strategies adopted when fishingwith two beam trawls towed in parallel (ICES, 1994).The purpose of all sampling methods was to providecatch rates of each species raised to a standard 8 mbeam trawl tow, and for a duration of I h. Mean catch

rates of each species (number 8 m beamtrawl-l h tow-l and weight 8 m beamtrawl-l h tow-I) at each fixed station position forthe period 1990-1996 were calculated and used inall analyses. Only those fish species which wereclassified as commercially important, or as demersalnon-target species, were used in analyses of commu-nity structure (Rogers et aI., 1998).

2.2. Statistical methods

To describe the distribution of fish biomass with

depth, species which have been identified as importantcontributors to the total demersal fish fauna were

combined into six groups, and the total weight of fishin these groups was calculated for I m depth intervalswithin groups of ICES Divisions, over the entire depthrange of the survey (Rogers et aI., 1998). Theseweights were then expressed as a cumulative fre-quency. The six fish groups used were rays (Rajidae),gadoids (gadidae), sharks (all otherelasmobranchsexcluding the rays), all flatfish, all weevers/blenniesand gobies (Families Trachinidae, Blenniidae, Sti-chaeidae, Pholidae, Callionymidae and Gobiidae)and all gurnards. (Families Triglidae, Cottidae andAgonidae.)

Diversity profiles, a graphical display of a family ofdiversity indices, were calculated for the fish catches(number 8 m beam trawl-I h-I) within each of thesectors identified in the coastal waters of the northeast

139

Atlantic. This family of diversity indices was obtainedby changing the scale parameter (a). There are severalforms available, but one that is recommended for largedata sets is Renyi's diversity index family Ha (Patiland Taillie, 1982; Tothmeresz, 1995):

Ha = (logpn/(l - a),

where Pi is the proportion of individuals belonging tothe ith species. When substituting 0, I and 2 for thescale parameter a, Ha will be directly related to thespecies richness (i.e. a is the log of the speciesnumber), Shannon's entropy and Simpson's domi-nance index, respectively (Hill, 1973). Thus for anear zero, richness will have more effect on Ha, butfor larger values of the scale parameter, species even-ness has more effect. For scale parameters whichincrease from I to 4 the influence of rare species willbe gradually replaced by the influence of dominantspecies. One community is more diverse than anotherif its diversity profile is equal to or above that ofanother. over the whole range of the scale parameter. Ifthe two profiles intersect at any point then they can beconsidered non-comparable (i.e. different diversityindices would rank the communities differently).

Relative abundances and biomass of demersal spe-cies (k-dominance curves) were superimposed usingthe ABC method of Warwick (1986) to provide infor-mation on the size of the most dominant species.

To compliment these descriptions, mean speciesabundances (numbers and biomass 8 m beamtrawl-I h-I ) were ranked and presented graphicallyfor the 18 species which comprised 95% of the totalcatch number throughout the survey area. The pre-sentation of abundance data for individual named

species allows the community structure to be inter-preted in terms of the contribution of individualspecies, rather than anonymous ranked species abun-dances. This can be used to distinguish identicalprofiles of diversity, or of ranked abundance such asABC plots, in two communities which have identicalrelative abundances, yet in which the species them-selves are different.

A similarity of percentages analysis (SIMPER)identified those species which typified the catcheswithin sectors, and also those species most responsiblefor the dissimilarity in catches between sectors, usingraised data. A full description of SIMPER analysis isprovided by Clarke (1993). Non-metric multi-dimen-

140 S.l. Rogers et al.! Fisheries Research 40 (1999) 135-/52

sional scaling (MDS) was used for ordination andclustering of the data to describe the inter-relation-ships between sectors, taking into account individualspecies abundances (Clarke and Ainsworth, 1993). ABray-Curtis similarity matrix of root transformedmean catch numbers and weight in each sector wascalculated. This allowed the species abundances to bedescribed in terms of their ranked similarities, and

these were plotted as Euclidean distances in a two-dimensional plot, where the distance apart of eachsample represented their relative similarity. A lowvalue of stress for such a plot «0.1) indicated thatthe technique gave a good representation of relation-ships between sectors and that only two dimensionswere required for displaying the ordination.

3. Results

The coastal waters of the northeast Atlantic cover a

wide range of marine environments, from shallow,sheltered areas in the North Sea with several largeestuaries, to exposed and more steeply shelving areason the southwest coast of the UK (Fig. lea»~. Thedepth distribution of groups of demersal speciesshowed that up to 60% of the total biomass of somegroups sampled occurred within the depth range 10-30 m, but this proportion was highly variable anddepended on the depth profile of each region andthe prevailing environmental conditions. The mostnotable differences in cumulative biomass between

ICES Divisions were evident for 'rays' in the NorthSea, which were restricted to the deeper parts ofDivision IVb (Fig. 2(a», in contrast to the Channeland western waters (Fig. 2(b) and (c». In the Channel,the biomass of demersal species in the 'shark' and'gurnard' groups was more associated with deep waterthan in either the North Sea or the Irish Sea (Fig. 2(b».

3.1. The fish assemblage: variation within sectors

Throughout the entire survey area, 39 demersalcommercial and non-target species provided 99% ofthe total numerical abundance of fish, and 38% of thistotal consisted of dab Limanda limanda and plaicePleuronectes platessa, while poor cod, dragonet andsolenette were also important contributors (Table I).When expressed in terms of biomass, the total con-

100 r a)80

60

40 ICES divisions

IVb & IVc

ICES divisionsVlid & Vlie

100 r C)80 f-

I

60 ~

''-

ICES divisions

Vlia, Vllt & VlIg

30 40 50 60

Depth (m)

70 80 90 100

Fig. 2. The cumularive frequency of tota! fish weights by I mdepth band. for blennies and gobies (fine line), flatfish (short dash),gadoids (long dash), gurnards (dot dash), rays (bold short dash),and sharks (bold line), in ICES Divisions IVb and IVe (a), VHd and

e (b), and VHa, f and g (c), see text for further details of species

within each group.

tribution of plaice and dab increased to over 50%, andthese species dominated the fauna. The relative rank ofspecies changed slightly when their percentage con-tribution was expressed in terms of biomass, due todifferences in the mean weight of each species. Forexample, larger species, such as sole, lesser spotteddogfish, and thornback ray all showed increased pro-portions of the total biomass and increased their rankorder. Smaller species such as poor cod and solenette,decreased in rank, as illustrated by the negative values

0

- 100 r b)"'"'reE 800;e.>-

600c:":Jer"40.;:

".2:1ij:; 20E:JU

0

S.l. Rogers et al.l Fisheries Research 40 (1999) 135-152 141

Table I

Comparison of total catch number and total catch weight for selected species from all regions, which comprise 99% of the total abundance ofall fish

for the difference in percentage contribution(%W-%N) (Table 1).

The results of a SIMPER analysis of the fish catcheswithin each sector identified those species whichcontributed a large proportion to the total within sectorsimilarity, and also those species which were found atconsistent abundance amongst all catches and whichthus typified the catches within each sector (Table 2).

The mean species abundances (number and biomass)are also illustrated in Fig. 3. In the North Sea (sectorsIVb(e), IVc(e) and IVc(w)), plaice, dab, sole, dragonetand scaldfish Amoglossus latema were important intypifying the demersal community, while very fewother species were important contributors to withinsector similarity of catches. In terms of species bio-mass, plaice and dab formed over 70% of the total

No. Percentage of total Percentage of total %W-%N

catch number (%N) catch weight (% W)

I Dab Limanda limanda 27.1 19.7 -7.32 Plaice Pleuronectes platessa 11.8 32.6 20.83 Poor cod Trisopterus minutus 8.7 1.6 -7.14 Dragonet family Callionymiidae 8.6 3.4 -5.2

5 Solenette Buglossidium luteum 8.3 I.l -7.26 Whiting Merlangius merlangus 4.6 3.3 -1.3

7 Sole Solea solea 4.1 7.4 3.38 Hook-nose Agonus cataphracrus 3.9 0.4 -3.6

9 Grey gumard Eutrigla gurnardus 3.1 1.2 -1.910 Scaldfish Arnoglossus laterna 3.0 0.5 -2.511 Bib Trisopterus luscus 2.8 1.8 -1.0

12 Goby family Gobiidae 2.0 0.1 -2.013 Lesser we ever Echiichthys vipera 1.9 0.4 -1.5

14 Thick back sole Microchirus variegatus 1.7 0.5 -1.2

15 Lesser spotted dogfish Scyliorhinus canicula 1.2 7.3 6.1

16 Lemon sole Microstomus kitt 0.9 1.4 0.6

17 Cod Gadus morhua 0.8 2.4 1.618 Red gumard Aspitrigla cuculus 0.7 0.8 0.2

19 Long rough dab Hippoglossoides platessoides 0.7 0.2 -0.520 Tub gumard Trigla lucerna 0.4 0.9 0.521 Thomback ray Raja clavata 0.4 2.8 2.4

22 Bull-rout Myoxocephalus scorpius 0.3 0.2 -0.123 Flounder Platichthys flesus 0.3 0.9 0.624 Haddock Melanogrammus aeglefinus 0.3 0.5 0.225 Angler Lophius piscatorius 0.2 1.6 1.326 Spotted ray Raja montagui 0.2 0.8 0.627 Brill Scophthalmus rhombus 0.2 1.0 0.828 Hake Merluccius merluccius 0.1 0.4 0.2

29 Norwegian topknot Phrynorhombus norvegicus 0.1 0.0 -0.130 Witch Glyptocephalus cynoglossus 0.1 0.2 0.0

31 Four-bearded rockling Enchelyopus cimbrius 0.1 0.0 -0.1

32 Cuckoo ray Raja naevus 0.1 0.6 0.533 Red mullet Mullus surmuletUs 0.1 0.1 0.0

34 Turbot Scophthalmus maximus O.! 0.8 0.735 Greater pipefish Svngnathus aCllS 0.1 0.0 -0.136 Blonde ray Raja brachYllra 0.1 0.4 0.337 John dory ZellS faber 0.1 0.1 0.138 Streaked gumard Trigloporus lastoviza 0.1 0.1 0.039 Buttertish Pho/is gunellus 0.1 0.0 -0.1

The difference in percentage contribution when using weight and abundance. is shown.

I "III 142 S.l. Rogers et al.! Fisheries Research 40 (1999) 135-152

Table 2

Percentage contributions (>6%) of demersal species to the average within-sector similarity, identified using SIMPER analysis of Bray-Cudissimilarities (Clarke, 1993)

Within-sector similarity (%) 68.75

"Species consistent in typifying the community (meanlSD>4).

biomass in North Sea catches (Fig. 3(b». Plots of k-dominance for these three North Sea sectors,expressed as cumulative ranked abundance and bio-mass, confirmed this dominance, and also showed that

these dominant species generally occupied a largerproportion of the total when expressed as biomassrather than as numerical abundance. due to differences

in the mean weight of each species (Fig. 4(a)-(c».In both northern and southern parts of the Channel

(Division VIld(n) and VIld(s», small species, such asdragonet, solenette and bib Trisopterus luscus hadhigh mean abundance, but poor cod were also espe-cially abundant in the north (Fig. 3(a) and (b». Bothdragonet and hook-nose Agonus cataphractus, typi-fied these demersal populations (Table 2). In contrastto the North Sea, the k-dominance plots for theChannel showed an inverted pattern with the abun-dance curve more elevated than the biomass curve,

confirming that these numerically most abundant spe-cies are small in size (Fig. 4(d) and (e». At highspecies rank, both curves were steeper in gradientthan those in the North Sea, indicating that morespecies occurred in moderate to high abundance.

In the western Channel and Bristol Channel (Divi-sions VIle, VIIf and VIlg, respectively), althoughflatfish still fonned part of the catch, dragonets andmembers of the Gadidae, particularly poor cod and

73.66 60.97 51.10 53.88 53.87 54.61 58.44 45.32

whiting Merlangius merlangus, contributed most twithin-sector similarity (Table 2). In both sectors thlesser spotted dogfish had the greatest mean weight (all species sampled (Fig. 3(b». This mixture of largand small species, each with a different rank in termof abundance and biomass. produced k-dominanccurves which crossed for those catches taken frorthe Bristol Channel, and in the western Channe

resembled those further to the east (Fig. 3(a) an(b); Fig. 4(0 and (g».

In the Irish Sea (sectors VIla(e) and VIl(w», neonly were gadoids important contributors to th,within-sector similarity (Table 2), but dab, plaicepoor cod, and dragonet each provided more tha]10% of the total mean abundance (Fig. 3(a». Biomasin the Irish Sea was dominated by dab and plaice, buto a lesser extent than on the Continental coast of th

North Sea, while sole, lesser spotted dogfish and C04were also important (Fig. 3(b». Plots of k-dominanclreflected the similarity with North Sea catches, abiomass curves were generally more elevated thaJthose for abundance (Fig. 4(h) and (i». The absence 0typifying species throughout western waters of the DJ'compared to the North Sea (Table 2), and the relatively low within-sector similarity of these sectorsindicated the heterogeneous nature of the populationsampled.

IVb(e) IVc(e) IVc(w) Vlld(n) VIId(s) VIle Vllf&g Vlla(e) Vlla( 1

Dragonet family Callionymiidae 8.03 8.26" 8.27 12.1I" 15.01" 12.20 7.72 8.91 9.05Dab Limanda limanda 20.05 13.13" 9.66 - - - 6.76 8.63Plaice Pleuroneetes platessa 12.74" 9.26" 8.17 - 6.44 7.36 - 6.24 8.20Sole Solea solea - 7.12" 11.00 6.28 8.59 7.82 - 7.42 -Hook-nose Agonus eataphraetus 6.20 - 7.84 6.75 8.97" - - - 6.40Poor cod Trisopterus minulUs - - - 10.05 - 6.56 10.75 10.41 14.43Whiting Merlangius merlangus 8.70 - 7.19 - - 6.36 - 10.51Grey gumard Eutrigla gumardus 8.37 - - - - - 7.49 6.40 7.38Lesser spotted dogfish Seyliorhinus eanieula - - - - - 7.30 6.90 6.06ScaIdfish Amoglossus latema 6.52 8.85" - -Bib Trisopterus luseus - - 8.26 9.89Red gurnard Aspitrigla eueulus - - - - - 8.80Lesser weever Eehiiehthys vipera 9.82Solenette Buglossidium luteum - 9.02Lemon sole Mierostomus kitt - - 6.68

(a)

51 Rogers et al.lFisheries Research 40 (1999) 135-152

IVb(e)

0c:-5<ii"'0;ft

Observed population composition

IVe(e)

143

IVe(w)

. .. .

Species reference number

Vlla(e)

.§ 40

I! 20 ......-.~ 0 2 4 6 8 '0 12 14 16 18

species reference number

Vlld(s)

species reference number

Vlla(w)

I : I.." 0---

2 4 6 8 1012 '4 1618

species reference number

Vile

species relerence number

(bl

IVb(e)E0>

.~ 40 ..! 2: I.;ft

-2 4 6 8 10 12 14 16 18

species reference number

Vlla(e)E

.~ 40

Ii 20 ..'0 0;ft

I2 4 6 8 10 12 14 16 18

spedes reference number

Vlld(s)E

~ 40

Ii 20 ..~ 0...

L. . ~

2 4 6 8 10 12 14 16 18

species reference number

spectes reference number

Catch weight composition

IVe(e)E0>

I ~ I~...

.2 4 6 8 10 12 14 16 18

species reference number

L.-

2 4 6 8 10 12 14 16 18

species reference number

Vlld(n)

species reference number

~a;'" 40.c* 20"'0...

i 40 1i 20 J.a'0 0...

ES!'..

£ 40

I;;; 20

~o'";ft

species reference number

IVe(w)

ol~2 4 6 6 10 12 14 16 18

spectes reference number

Vlld(n)

. . L-2 4 6 8 10 12 14 16 18

species reference number

VIIf&g

... L2 4 6 8 10 12 14 16 16

spectes reference number

Fig. 3. Mean species abundances, expressed as numbers (a) and biomass (b) per 8 m beam trawl h-1, by coastal region, for the 18 specieswhich comprised 95% of the total catch number throughout the survey area. Species reference numbers on the x-axis can be interpreted usingTable 1.

speciesreference number

VileE0>

! 40

I

B 20

~0..LIL- _.J.;ft 2 4 6 8 10 12 14 16 18

species reference number

IL--0 0

40c:

40c:

40 j'"

IL-'"B B

20 '"20 20"

'0 - .8 '00 .0 ;ft 0 .'2 4 6 8 10 12 14 16 18 2 4 6 8 10 12 14 16 18

0 0c:40 c 40

I.J.--5

I...;;;" 20 " 20'0 '0 . -... 0 " 0

2 4 6 8 '0 12 14 '6 18 2 4 8 8 10 12 '4 16 '8

Vlla(w)E0>;;; 40

I.L.3'-5 20

L.'0 0... 2 4 6 8 10 12 14 16 18

0

£ 40 I

;;; 20

1416180 0 4 6 8 10 1;ft 2

VJIf&g

0c 40.c

I.J..- - .B

20'0 -;ft 0 2 4 6 8 10 12 '4 16 18

144 S.l. Rogers et at.! Fisheries Research 40 (1999) 135-152

~~ t'L

20 ~

0 [ '-'-

/

..

1ooc1! I Cc l-

f 8°f

'

F8O~ r~ 40~

IV':'

J2l/0'----100r

80~ e ~~ I/~80. ', /t /

40v:::/20V

O"1 10

:~g ~t- -""I ,rf- . /

" 80 ~ / ~ /g .,'

i 40- //iJ Y//" /.:?20~u~ 0I 1OOr... . I~ r<3 80.

m" /:vot

1 '0Speciesrank

[ d

/1

Species rank

10

V/

'"10

Spec,es rank

Fig. 4. Cumulative biomass (triangle) and cumulative abundance (filled circle) for ranked species (expressed on a logarithmic scale) in csectors IVb(e) (a); IVc(e) (b); IVc(w) (c); Vlld(s) (d); Vlld(n) (e); Vllf and g (f); VIle (g); VIIa(w) (h) and Vlla(e) (i).

S.l. Rogers et al./ Fisheries Research 40 (1999) 135-152

Table 3

A comparison of Renyi diversity profiles shown in Fig. 2. which identifies those sectors which are non-comparable x (i.e. the profileswithin the range of the scale parameter used) and those which have ditt'erent diversity

3.2. Variation between sectors

Diversity profiles for each sector were used toidentify those demersal communities which weremore or less diverse than each other. and those which

were non-comparable (Fig. 5(a)-(c»). A summary ofthese diversity relationships between each sector(Table 3) showed that the samples from the continen-tal coast of the North Sea (sectors IVb( e) and IVc( e J)had Iow diversity, both because of the limited totalnumber of species present (Iow value at 0'=0), and theeffect of high dab and plaice abundance creating apopulation in which species abundances wereunevenly distributed (Iow evenness at 0'=3). Thesetwo sectors had lower diversity than all others exceptthe Continental coast of the eastern Channel. VIId(s)(Table 3).

Those species responsible for discriminatingbetween sectors were identified by their contributionto the average dissimilarity, using SIMPER analysis(Table 4). The between-sector dissimilarity value forNorth Sea sectors IVb(e) and IVc(e) of 35.27% wasthe lowest of all comparisons made, confirming thesimilarity of the demersal faunas in each sector. Thelesser weever, with greater abundance in sector IVc(e),was the only species which consistently discriminatedbetween these two communities (Table 4). The highermean catch rate of bib and lemon sole Microstomus

kitt in the western part of the southern North Sea,IVc(w), contributed to the discrimination betweenthese catches and those to the east IVc(e), and also

to its greater diversity (Tables 3 and 4).

In contrast to the North Sea, communitieseastern Irish Sea (VIIa(e)) were more diverse th

scctors except those in the northern part of the e:Channel, VIId(n), the Bristol Channel, VIIf andand the English coast of the southern NorthI\'c(w), which were all non-comparable. The e:pan of the Irish sea was more diverse than theVlIa(w), in part due to the relatively higher calsole and solenette, and lower mean catch of ]

(Fig. 3(a); Table 4)..-\It other sectors identified in Fig. 1 had div

profiles which occurred between those of the Irisand Continental North Sea coasts. This reflectec

the range in total species richness values forsector, and the way in which the abundances ofspecies were distributed relative to each(Table 3).

On the English coast in the southwest, commuin the western Channel (VIle) were less clthan those to the east (VIld(n)) and to the(VIII' and VIIg). This was mainly caused by a sIlower species richness for this sector when cornto VlId(n)), even though communities were of SIevenness (Fig. 5). In sector VUf and VIIg,were more species with a moderate (5-10%) pttage of the catch, and this was partly due to thepercentage of dragonet and solenette pi(Fig. 3(a)).

The inter-relationships between sectorsfurther described by MDS ordination and c1usplots using both mean numerical abundance andspecies weight. The distinction between tht

Group I Group 2

VIla(e) VIld(n) VIIf&g VIIa(w) VIld(s) VIle IVc(w) IVb(e)

VIla(e) x x > > > x >

VIld(n) x - x x > > x >

VIIf&g x x > > > x >

Vlla(w) < x < - > x >Vlld(s) < < < x - x x >VIle < < < < " - x >

IVc(w) x x x x " x - >

IVb(e) < < < < ..;: < < -

IVc(e) < < < < < < x

Communities in sectors in group I were less diverse than those in group:: f".. combinations marked with '<', and more diverse when'>'.

~. --"'-"""-"--'--

:j;:~

Renyi Diversity Profiles Renyi Diversity Profiles Renyi Diversity Profiles

VII.(.)

",:-.~c""'"~~

Vlld(n)Vlld(,)VII.

VIII&,VII.(w)

IVe(w) 'Vb(.)'Ve(.)

Hlu) Hlu)

Vlld(n)VII.

V"'&,VII.(.)VII.(w)

;:."-~;:-'".,,,''"

Hlu)IVe(w)

Vlldl')

~'"'"'"::,ti;:-.J>..C

IVel')

'~IVbl') ~.....\Q\Q

~.....u"V.I

.....

~

(a) (b) (c)

Fig. 5. Families of Renyi's diversity indices for fish catches (number per 8 m beam trawl h-') within each of the coastal sectors. These profiles display graphically a family ofdiversity indices obtained by changing the scale parameter lX,which for values of 0, I and 2 will be directly related to the species richness (i.e. Q is the log of the species number),Shannon's entropy and Simpsons dominance index, respectively. One community is more diverse than another if its diversity profile is equal to or above that of another, over thewhole range of the scale parameter. Diversity profiles are shown for sectors in the North Sea (a), the Channel (b) and western waters of the UK (c).

51. Rogers et al.lFisheries Research 40 (1999) 135-152 147

Table 4

Percentage contributions (>3.5%) of discriminating demersal species to the average dissimilarity between selected sectors, identified usingSIMPER analysis (Clarke, 1993)

Between-sector

dissimilarity (%)

"Species consistent in discriminating between sectors (mean/SD> 1.5).

SolenettePoor codScaldfishDabBibLesser weever

Grey gumardLemon soleSoleThick back soleHook-nosePlaice

Red gumardGoby family

Buglossidium luteum

Trisopterus minutus

Amoglossus latemaLimanda limanda

Trisopterus luscusEchiichthys vipera

Eutrigla gumardusMicrosromus kitt

Solea solea

Microchims variegatus

Argol1us cataphractusPleuronectes platessa

Aspitrigla cuculusGobiidae

IVb(e)&IVc(e)

5.73

4.7911.32"4.00

5.38

35.27

diverse North Sea faunas, and more diverse western

UK faunas, was clearly identified in clustering plotsfor both numerical and weight data, where sectorswere grouped approximately according to their geo-graphical distance apart (Fig. 6). When using fish

Stress = 0.04

(a)IVb(e)

IVc(e)

IVc(e)&IVc(w)

5.114.294.82

5.465.57

5.354.23

4.19

4.274.26

4.36

4.41

41.65 53.13 49.93 52.28 47.34

biomass. the distinction between the southern North

Sea fauna on the English coast (IVc(w)) was lessclearly differentiated from the westerly faunas(Fig. 6(b)). The predominance of small speciesthroughout the Channel distinguished this assemblage

Stress = 0.04

(b)

IVb(e)

Vlla(w) IVc(e)

VIIf&g Vlla(e)

IVc(w)

VileVlld(n)

Vlld(s)

Fig. 6. Multi-dimensional scaling ordination plots for mean catch number (a) and mean catch weight (b), of all fish samples caught within the

nine sectors illustrated in Fig. I.

Vlla(w)

Vllf&gVlla(e) IVc(w)

Vile Vlld(n)Vlld(s)

VIIa( e)& VIId(n)& VIIe& VIIe&

VIIa(w) VIId( s) VIIf&g VIId(n)

4.77 4.47 4.16 4.68

4.5 4.71 4.36 4.76-

4.564.00

(a)

S.l. Rogers et al./Fisheries Research 40 (1999) /35-/52

IVb(e)

148

60 70 80 90

Bray-Curussimilarity

IVc(e)

IVc(w)

Vlld(n)

Vlld(s)

Vile

Vlla(w)

Vlla(e)

70 80 90

Bray-Curtis similarity

, Vllf&g100

Fig. 7. Bray-Curtis similarity dendrograms for mean catch number (a) and mean catch weight (b). of all fish samples caught within the ninesectors illustrated in Fig. I.

from others to the west where flatfish. gadoids andother larger species. such as elasmobranchs, becameincreasingly abundant. In the west, ordination of IrishSea and Bristol Channel faunas distinguished the lessdiverse community in the western Channel (VIle),although the degree of similarity within this groupvaried depending on whether numerical or biomassdata were used to determine sample similarity(Fig. 7(a)-(b».

3.3. Distribution of the elasmobranchs

The distribution of the five elasmobranch specieswhich occurred in the top 36 ranked species (Table 1),the lesser spotted dogfish, thornback ray Raja clavata,spotted ray Raja montagui, cuckoo ray Raja naevus,and blonde ray, Raja brachyura, is shown in Fig. 8(a)-(e)). Without exception the centre of abundance ofthese species is to the west of the UK. Dogfish arewidespread throughout the area sampled and evenoccur in abundance in the southern North Sea off

the English coast, but appear to be rare or absentelsewhere in the North Sea (Fig. 8(a». Thornback,blonde and spotted rays show a similar distribution butare generally less abundant, especially the blonde ray,and more closely associated with coastal areas(Fig. 8(b)-(d)). The cuckoo ray is confined to thewestern waters of the UK, and is found almost entirelyin the Irish Sea and Bristol Channel (Fig. 8(e».

4. Discussion

These analyses have allowed us for the first time tocompare directly the diversity of the demersal fishassemblage of the northeast Atlantic shelf seas, usingdata collected by beam trawl surveys. The results showthat catches were dominated by a few species, whichoccurred in large numbers and at a high biomass. Themost abundant species were those which were typicalof shallow, uniform sandy and muddy seabed whichoccurred predominantly in the southern North Sea, butalso to a lesser extent in the Irish Sea and BristolChannel (Knijn et aI., 1993). The high level of dom-inance of catches by plaice and dab, and the limitedtotal number of species present, were partly respon-sible for the generally low diversity of demersalassemblages on the Continental coast of the NorthSea. compared to those in the Channel and west of theUK. One of the few species which helped to differ-entiate bt:tween these populations in the North Seawas the lesser weever, which occurred in lower abun-dance in the waters off Germany and Denmark, whileoff the coast of southeast England, bib and lemon soleoccurred in larger numbers.

West of the Dover Strait, towards regions whichwere more influenced by southern fish faunas, andwhich supported a generally more heterogeneoussubstrate, the total number of demersal speciesincreased (Fig. 5) (Henderson, 1989; Rogers et al.,

IVb(e) (b)

IVc(e)

IVc(w)

Vlld(n)

Vlld(s)

Vile

Vlla(e)

Vlla(w)

,Vllf&g ,100 60

S./. Rogers et al./ Fisheries Research 40 (/999) 135-/52

581.~

~""

J .:::j;r'# ...

\)I ~. : .:,::

" ... ::: :.:::'c ~ ..::::::.: ,.1 w '\ "::::::::~:.'.t:

~~ ? .. :.::::::" .,;;;r-..,;:..'~::""'~.. ..W') ", , . . V.'-~ ".'..,..;;: ~"'...H~~.. ~fij? numbern~~. --<'p per hour

... ~

~. ... . :. x 0

.. '..::..~~:F . 25...p... \1 ..".: \ '---y. . 150~ . 100048

(a) 0

!

581 ~

~~ ...:'(

I1~'/ ":::::1'... ~ :.::...,

\ i'~\" ::.:.::. I

~. 1" .:tH\j?;s I). a . ". : :: :,r"-i,..j!

j r: : . :: ,7 ""'"j:::;:::~~" "I ~..~ 'f ,..:I..., . .

,

""', ' IoI..~ f.,.." ;i ,1.:':" f

V::.:: numbern IY:"::~1.; --<; : :. per hour i... ...;.4. i

'.. ".'--~ iI '.., '0 ,I . ../ . . . .. "', !! ~,:. ':.:,;y . 25 Ii '::..,...~:;. !i 'o\'~ .150,

::;-~ .100048i

(c) 0 10

.-

58fr

~"" ~

J ':::::!,

-!! "', : ::.: ;"

...:;;;[ ::;',-" ':.:...:.,... . t:

. i ..::::::.: ~~

~ ~r \, '::"':::::.,' .' :\. ., . : . : : : : : . ',' L,-j 0,,-v.1. ... .. "~>sII (::':::7 ~::::::::CI ~:'-~ (~:::"/".,~ \

y""Jd::

~: J. "Y

,

=:;::' numbers.. . .. .. . hourI ... --<.:.... perI::':

~::jld .'I . :j ... . . .. x 0I .. ."'" .:;;, ../ .". .. :,;.- . 25

j . '::.<~~ 8 150I ,: ~ .1000i '!:48j

0(d) 10

14

581'1 l~ .:::j;(! ", I : .. .\

~: <:i!i;;~"'1. _.::\ .. ::.:::::;:i"'1,fr:.. ..;;. ~ ~ : : : : ~, . ~~.".~. 'r.. )oo ~:I::.' """

pt'::::

~:. :~:::. numbern,~.~~ ~... hour

~:

~~ --:::.. per..' ::.. ....". .' ....:. ' 0.:' """.':;J.." ':.'r .25

'.::.<\~ 8 150~ .,000

10

48

(b) 0 10

58~~~~( ~ ...::r!!1~~~ ~/ ::jj\i\i' ~ :: :j 'W; \ ':::::.: 'i.I~

d ~ ?" ~ ::j::::;::::~/:',. '\. .:.: :~

I

. t/. , V1 " .. : ~i>" .. .1 "" . . . . . . "', , """""'V11 ).:~ ~::::::~

:~~f:':

0k::':'/ barnI ' "Y num

!~'~.Y ;.;,:::. per hourj

':'. """"'4 ,..;.

I:: .'::.r-~:r. x 0

I .:l ":':':.'::; . 25' ~.. ;";':y. . ::"'\~ 8 150

~ . 1000:;-48

(e) 100

Fig. 8. The mean catch rate (numbers per 8 m beam trawl h-') of lesser spotted dogfish (a), thornback ray (b), spotted ray (c), blonde ray (d),and cuckoo ray (e), by quaner ICES rectangle.

150 S.l. Rogers et al./Fisheries Research 40 (/999) 135-152

1998). The more steeply shelving substrate here sup-ported a wider range of sediment types, and popula-tions of smaller fish dominated the demersal

assemblage. Still further west, towards the CelticSea, the presence of southern species such as redgurnard and thickback sole, the increasing abundanceof elasmobranchs, and the high catches of gadoids.increased species diversity of these assemblages com-pared to those to the east. Between the Bristol Channeland the Irish Sea this basic population structureremained but was increasingly dominated by largeflatfish populations, which, like those in the North Sea.were able to exploit areas of shallow sandy seabed(Symonds and Rogers, 1995).

These data summarise the diversity of demersalassemblages as sampled by beam trawl on a widespatial scale, and without a historical perspective.Explanations for the observed patterns in diversityinclude well-known principles involving the distribu-tion of warm-water and boreal faunas. and the hydro-graphic influence of the major water current systemsin the region (Ekman. 1967; Henderson. 1989; Pauly.1994). These factors can be assumed not only toinfluence the relative abundance of resident species.but also the richness of those fish faunas which are

accessible to southern migrants. An extensive area ofrelatively low diversity exists in the southern NorthSea. and the comparison of this with other regions.may suggest the involvement of some form of humanactivity. There is certainly a considerable amount ofanthropogenic impact in the region. with fishingactivity, eutrophication, pollution and infrequent per-iods of oxygen depletion (Riesen and Reise. 1982;Rijnsdorp and Leeuwen, 1996). This low diversity fishfauna may also, however, be related to the extremeenvironmental conditions in this relatively shallowarea. which is exposed to larger seasonal fluctuationsin temperature and salinity than in the other sectorsstudied.

A more fruitful approach to finding evidence ofimpact may not be to use spatially diverse communitylevel indices such as these, but to examine individual

parts of the assemblage which may show quantifiableeffects of exploitation. For example, the low fecundityand high length and age at maturity of elasmobranchsmakes this group particularly vulnerable to towedgears, and thus an additional source of evidence ofgear effects. Dogfish and four ray species were absent

from the southeast North Sea, where consistently highgear impact has occurred. Not all species are equallysensitive, however, and those with a relatively lowlength at maturity such as the starry ray Raja radiata,have maintained population size (Walker and Heessen,1996). The size structure of parts of the demersalpopulation is affected by selective removal of largerfish, and a comparison of these patterns in relation toknown levels of fishing effort, and at a range of depths,could also be informative (Pope and Knights, 1982;Rice and Gislason, 1996; Rijnsdorp et aI., 1996b;Rogers et aI., 1998).

In contrast to much of the fish fauna, towed gearshave shown quantifiable effects on invertebratebenthic communities, for example, experimentaltrawling altered the abundance of small invertebrates(Bergman and Hup, 1992), and significantly reducedthe number of species in a stable sediment community(Kaiser and Spencer, 1996). These direct effectsincluded the provision of food for other scavengerspecies through the damage or killing of benthicinvertebrates in the passage of the gear, and alteredthe proportions of short-lived and long-lived species.For example, large numbers of the bivalve Arcticaislandica have been reported to be damaged by ottertrawling in Kiel Bay, and during intense fishing activ-ity this species is more frequently found in the sto-machs of cod, Gadus morhua (Arntz and Weber,1970). Thus the possible impact of trawling on demer-sal fish may also occur through indirect effects via thebenthos. and so it has been difficult to attribute

observed changes in demersal fish populations tofishing effort. One of the few associations to bedocumented has involved changes in the growth rateof sole and plaice in the North Sea, where recent

increases in mean length at age were thought to becaused by a combination of beam trawl and eutrophi-cation effects (Millner and Whiting, 1996; Rijnsdorpand Leeuwen, 1996).

In subsequent investigations of the impact of trawl-ing on widespread demersal communities, it will bevery difficult to disentangle the effect of naturalprocesses from any potential influence of towed gears.Examining specific aspects of the assemblage, withemphasis on selected vulnerable species rather thanthe community as a whole, may be the best way tointerpret such spatially diverse data. The provision ofaccurate fishing effort data to a high level of resolution

S.l. Rogers et al./ Fisheries Research 40 (1999) 135-/52

will be a valuable next step in interpreting theseobservations (Rijnsdorp et aI., 1996a).

Acknowledgements

We would like to thank Dr. Peter Henderson who

introduced us to diversity profiles, and stimulated theanalyses in this paper. We are also grateful to all thescientific staff, ships' officers and crew who havehelped to collect data during the Belgian, Dutch,German and UK Beam Trawl Surveys. This workwas funded by the Ministry of Agriculture, Fisheriesand Food under MOU 'A'.

References

Armz. W.E., Weber. W. 1970. Cyprina is/andica (Mollusca):Bivalvia als Nahrung von Dorsch und Kliesche in derKieler Bucht. Ber. Dtsch. Wiss. Komm. Meeresforsch. 21.193-209.

Bergman, MJ.N., Hup, M., 1992. Direct effects of beam trawlingon macrofauna in a sandy sediment in the southern North Sea.ICES J. Mar. Sci. 49. 5-11.

Brander. K.. 1981. Disappearanceof the common skate Raja batis.from the Irish Sea. Nature 290. 48--49.

Clarke. K.R., 1993. Non-parametric multivariare analyses ofchanges in community structure. Aust. J. Eco!' 18. 117-143.

Clarke. K.R.. Ainsworth. M., 1993. A method of linking multi-variate community structure to environmental variables. Mar.Eco!' Progr. Ser. 92. 205-219.

Collie. 1.S., Escanero. G.A.. Valentine. P.C., 1997. Effects ofbottom fishing on the benthis megafauna of Georges Bank.Mar. Eco!' Progr. Ser. 155. 159-172.

Daan. N.. BromJey, P.1., Hislop. J.R.G.. Nielsen. XA., 1990.Ecology of North Sea fish. Nerh. J. Sea Res. 26. 343-386.

Ekman, S., 1967.Zoogeography of the Sea. Sidgwick and Jackson.London.

Fargo. 1.. Tyler. AY.. 1991. Sustainabiliry of flattish-dominatedfish assemblages in Hecate Strait. British Columbia. Canada.Nerh. 1. Sea Res. 27, 237-253.

Farina. A.C., Freire, J.. Gonzalez-Gurriaran. E.. 1997. Demersalfish assemblages in the Galician continental shelf and upperslope (NW Spain): Spatial strucutre and long-term changes.Est. Coastal Shelf Sci. 44, 435-454.

Gee. J.M.. Warwick, R.M., 1996. A study of global biodiversitypatterns in the marine motile fauna of hard substrata. J. Mar.BioI. Assoc. 76, 177-184.

Greenstreet, S.P.R., Hall, S.1., 1996. Fishing and the ground-fishassemblage structure in the north-western North Sea: ananalysis of long-term and spatial trends. J. Anim. Eco!' 65,577-598.

151

Heessen. H.1.L., Daan. N., 1996. Long-term trends in ten non-target North Sea fish species. ICES J. Mar.Sci. 53. 1060-1078.

Henderson, P., 1989. On the structure of the inshore fishcommunity of England and Wales. J. Mar. BioI. Assoc.. UK69. 145-163.

Hill. M.O., 1973. Diversity and evenness: a unifying notation andits consequences. Ecology 54. 427-432.

ICES. 1992. Report of the Study Group on Ecosystem Effects ofFishing Activities. ICES Coop. Res. Rep.. ICES, Copenhagen.

ICES, 1994. Report of the Study Group on beam trawl surveys inthe North Sea and Eastern Channel. ICES CM 1990/G:59.

Kaiser. M.1., Spencer, RE., 1996. The effects of beam-trawldisturbance on infaunal communities in different habitats. J.Anim. Eco!. 65. 348-358.

Knijn, R.J., Boon, T.W., Heessen. H.J.L., Hislop, J.R.G., 1993.Atlas of North Sea fishes. ICES Coop. Res. Rep. 194. 268.

Millner. R.S., Whiting, c.L.. 1996. Long-term changes in growthand population abundance of sole in the North Sea from 1940to the present. ICES J. Mar. Sci. 53, 1185-1195.

Norman. J.R., 1963. A History of Fishes. Ernest Benn. London.Patil, G.P., Taillie. C.. 1982. Diversity as a concepr and its

measurement. J. Am. Stat. Assoc. 77, 548-567.

Pauly, D., 1994. A framework for latitudinal comparisons offlatfish recruitment. Neth. J. Sea Res. 32. 107-118.

Pope, J.G., Knights. B.1., 1982. Comparison of the lengthdistributions of combined catches of all demersal fishes in

surveys in the North Sea and at Faroe Bank. In: Mercer, M.C.(Ed.), Canadian Special Publication of Fisheries and AquaticSciences. vo!. 59. pp. 116-118.

Rice. 1.. Gislason. H.. 1996. Patterns of change in the size spectraof numbers and diversity of the North Sea assemblage, asref1ected in surveys and models. ICES J. Mar. Sci. 53. 1214-1225.

Riesen. W.,Reise. K., 1982. Macrobenthos of the subtidal WaddenSea: revisited after 55 years. Helgolander Meeresunters 35.409--423.

Rijnsdorp. AD.. Buys. A.M., Storbeck. E. Visser. E.. 1996a.Micro-scale distribution of beam trawl effort in the southernNorth Sea between 1993 and 1996 in relation to the trawlingfrequency of the sea bed and the impact on benthic organisms.ICES CM 19961Mini: !l, 34 pp.

Rijnsdorp. AD.. Van Leeuwen. P.L 1996. Changes in growth ofNorth Sea plaice since 1950 in relation to density. eutrophica-tion. beam trawl effort, and temperature. ICES J. Mar. Sci. 53.1199-1213.

Rijnsdorp, A.D., Van Leeuwen, PI.. Daan, N., Heessen, HJ.L..1996b. Changes in abundance of demersal fish species iD theNorth Sea between 1906-1909 and 1990-1995. ICES J. Mar.Sci. 53, 1054-1062.

Rogers, S.l., Damm. U., Rijnsdorp, A.D.. Vanhee, W. 1997.Description and evaluation of the international beam trawlsurveys of the European continental shelf. ICES CM 1997/Y:23. 20 pp.

Rogers. S.l., Rijnsdorp, A.D., Damm, U., Vanhee. W, 1998.Demersal fish populations in the coastal waters of the UK andcontinental NW Europe from beam trawl survey data collectedfrom 1990 to 1995. J. Sea Res. 39, 79-102.

152 S./. Rogers et a/.I Fisheries Research 40 (1999) /35-/52

Somerfield, P.J., Clarke, K.R., 1995. Taxonomic levels, in marinecommunity studies, revisited. Mar. Eco!' Progr. Ser. 127, 113-119.

Symonds, D.J., Rogers, S.I.. 1995. The influence of spawning andnursery grounds on the distribution of sole So/ea so/ea (L.) inthe Irish Sea, Bristol Channel and adjacent areas. J. Exp. Mar.Bio!' Eco!' 190,243-261.

Tothmeresz, B., 1995. Comparison of different methods fordiversity ordering. J. Veg. Sci. 6, 283-290.

Walker, P.A., Heessen, H.J.L., 1996. Long-term changes in raypopulations in the North Sea. ICES J. Mar. Sci. 53, 1085-1093.

Warwick. R.M., 1986. A new method for detecting pollutioneffects on marine macrobenthic communities. Mar. Bio!' 92.557-562.