56 (2006) 317–328www.elsevier.com/locate/seares

Journal of Sea Research

Evaluation of habitat use by adult plaice (Pleuronectes platessa L.)using underwater video survey techniques

Richard Shucksmith a,⁎, Hilmar Hinz b, Melanie Bergmann c, Michel J. Kaiser a

a School of Ocean Sciences, University of Wales-Bangor, Menai Bridge, Anglesey, LL59 5AB, United Kingdomb Marine Biological Association, The Laboratory, Citadel Hill, Plymouth, PL1 2PB, Devon, United Kingdom

c Alfred-Wegner Institute, Am Handelshafen 12, Bremerhaven D-27570, Germany

Received 28 February 2006; accepted 16 June 2006Available online 5 July 2006

Abstract

Large-scale spatial surveys of fish species in relation to habitat have tended to focus on depth, sediment type and temperature asdescriptors of fish habitats. At a smaller scale, habitat parameters such as the relief of the sea floor, the presence of structuring faunaand prey availability may have a considerable influence on fish distribution, but often are not considered. In the present study weused video survey techniques to study habitat components in areas of the English Channel that were known to support consistentlyhigh densities of adult plaice. Habitat features were quantified and related to the density of adult plaice caught within the samestudy areas. To focus the study on habitat components other than sediment type all sites chosen had sandy substrata. The scale andspatial distribution and heterogeneity of physical and biological structures were quantified for each site and correlated to plaicedensities. Plaice densities correlated with the abundance of benthic fauna recorded. In particular the emergent tube-dwellingpolychaetes Lanice conchilega and Cheatopterus spp., which are a valuable food source for plaice, dominated some sites. Abiotichabitat features and habitat heterogeneity showed no clear relationships with respect to plaice densities at the scale of our surveys.This indicated that prey availability might be the driving force for habitat selection of adult plaice within sandy habitats and thatother habitat descriptors assume less importance at smaller spatial scales.© 2006 Elsevier B.V. All rights reserved.

Keywords: Pleuronectes platessa; Habitat heterogeneity; Habitat selection; Food availability; Benthos; English Channel

1. Introduction

There is a current desire to move towards moreecosystem-based approaches to achieve the goal ofsustainable fisheries (e.g. Link, 2002; Meester et al.,

⁎ Corresponding author. Present address: Scottish Association forMarine Science, Dunstaffnage Marine Laboratory, Oban, Argyll,Scotland, PA37 1QA, United Kingdom Tel.: +44 1631 559345; fax:+44 1631 559001.

E-mail address: richard.shucksmith@sams.ac.uk (R. Shucksmith).

1385-1101/$ - see front matter © 2006 Elsevier B.V. All rights reserved.doi:10.1016/j.seares.2006.06.001

2004). While fisheries scientists have striven to under-stand the population dynamics of target species, knowl-edge of the ecological requirements of the latter arepatchy in their coverage (Bigelow and Schroeder, 2002)for all but a selection of iconic species, e.g. cod Gadusmorhua. Identification of the habitat requirements of keylife stages and an understanding of how these affectdistribution patterns on various spatial and temporalscales is a necessary component of an ecosystem-basedapproach to management (McConnaughey and Smith,2000). For demersal fish species, in particular flatfishes

318 R. Shucksmith et al. / Journal of Sea Research 56 (2006) 317–328

that spend most of their life in close affinity with theseabed, an appreciation of habitat use needs to be viewedin the wider context of the ecological effects of fishingactivities that have the potential to change seabed habitatstructure (Jennings and Kaiser, 1998; Kaiser et al.,2002). The use of towed bottom fishing gear is known tochange and alter the structure and function of habitatsthat fulfil an important role in the life-history of fish mostclosely associated with these habitats (Auster andLangton, 1999; Kaiser et al., 2002; Ryer et al., 2004).

Studies of the environmental determinants of thedistribution pattern of flatfish mostly have beenconducted over large geographical or regional scalesand have used only the principal environmentaldescriptors such as depth, sediment type and tempera-ture as predictors of these patterns (Smale et al., 1993;Albert et al., 1998; Ellis et al., 2000; Amezcua andNash, 2001). Few have tried to consider and quantifyother environmental components of flatfish habitatssuch as substratum relief, the presence of structuringepifauna, or prey availability (but see McConnaugheyand Smith, 2000; Stoner and Titgen, 2003). Theseparameters may be important predictors of flatfishdistribution at a smaller (local) scale and may influencehabitat selection. The importance of some of thesehabitat parameters for juvenile flatfish (substratum reliefand structuring fauna) has been demonstrated already(Abookire and Norcross, 1998; Norcross and Mueter,1999; Stoner and Abookire, 2002; Stoner and Titgen,2003; Ryer et al., 2004); however, for larger flatfishabove the minimum landing size, the role of suchparameters is not yet fully understood (Stoner andTitgen, 2003).

In the present study, habitat components at sitesknown to have consistently high densities of adult plaice(Pleuronectes platessa L.) were studied using under-water video surveys. Sampling benthic habitats withvideo transects, unlike by traditional methods such asgrabs and beam trawls, permits real-time observation ofthe seabed and assessment of the spatial distribution andcoverage/density of both abiotic and biotic features.Substratum structures and habitat heterogeneity as wellas the density of potential prey items can be estimatedover scales most likely to be relevant for individuallyforaging fish. Although the composition of parts of thebenthic assemblage can accurately be determined usinggrab sampling, this technique does not permit thedetermination and adequate quantification of otherhabitat descriptors, such as substratum relief and morescarcely distributed epibenthic fauna. Conversely, whilebeam trawls sample epifauna more efficiently coveringlarger areas, they pool benthic fauna over the areas

sampled leading to a loss of information on small spatialscales.

The main aim of this study was to describe habitatfeatures using video survey techniques in relation to thedensity of plaice of a legally landable size, sampled inthe near vicinity. In this way it was possible toinvestigate which habitat features were correlated withcatches of plaice and hence may influence habitatselection and overall distribution patterns. In theanalysis of video transects particular emphasis wasplaced on the quantification of abiotic features that addtopographic structure to the seabed (e.g. sand waves,bedrock, and cobbles) and on organisms that areresponsible for structuring the seabed (e.g. emergentepifauna, burrowing fauna) or may represent animportant food source. Besides studying the overallrelationship of habitat parameters to plaice densitiesamong sites, the importance of spatial variability offeatures within sites was investigated to examine therole of habitat heterogeneity.

2. Methods

2.1. Site selection

The study was undertaken in the English Channelwhere major fisheries for plaice (Peuronectes platessa)and common sole (Solea solea) exists. Site selectionwas based on a grid of sites regularly surveyedannually as part of a groundfish survey conducted bythe Centre for Environment, Fisheries and AquacultureScience (CEFAS, Lowestoft) for fish stock assessmentpurposes (Rogers et al., 1998; Kaiser et al., 1999). Asubset of nine of the sites that consistently had highdensities of adult plaice was identified from thegroundfish survey data using the method of Hinz et al.(2003). This methodology was applied as it ensures anobjective selection process of such sites from largedatasets rather than a choice based on subjectivejudgment.

The groundfish survey data spanned a nine-yearperiod from 1990 to 1998. Only plaice over theminimum landing size (>26 cm) collected during theautumn groundfish survey cruises were considered foranalysis. Plaice in the English Channel spawn betweenFebruary and May, hence the groundfish survey datarelate to the distribution of plaice outside the spawningseason. We restricted the analysis to plaice over theminimum landing size (>26 cm) because they are ofgreatest commercial relevance and close to maturity atthis size. Male plaice tend to reach maturity at a size ofapproximately 24 cm, while females are mature at a size

319R. Shucksmith et al. / Journal of Sea Research 56 (2006) 317–328

somewhat larger than the minimum landing size. Foreach of the 133 groundfish survey sites used for thisinitial analysis, the percentage of the total number of fishsampled in that year was calculated. This value wascalculated for each site for every year in the dataset. Thisconversion was performed to prevent a bias in theanalysis caused by the occurrence of an exceedinglyhigh abundance of plaice at a particular site in any oneyear. The percentage data (+1) was then Log10transformed to achieve a normal distribution. Fromthis, the long-term mean percentage of the sampledpopulation was calculated for each site and plottedagainst its corresponding standard deviation. Theresultant scatter plot was overlaid with a fittedregression and corresponding 95% Predictor Intervals(PI) generated from a regression analysis of a rando-mised version of the dataset. Within each year, every sitewas assigned a percentage of the sampled populationallocated at random from within the range of the data forthat year. For a more detailed description and discussionof this method see Hinz et al. (2003). Thus, sites thatwere plotted below the lower 95% PI represented sites atwhich plaice occurred more consistently (high mean:S.D. ratio) than predicted from the randomised data.Thirty one of the sites from the groundfish survey fellbelow the 95% PI of which nine had similar sedimentcomposition (sandy substrata) and depth characteris-tics. The long-term mean percentage abundance of

Fig. 1. Location of study sites

plaice varied among these sites, which enabled us toexamine the relationship between habitat characteris-tics and plaice densities.

Previous studies of the large-scale distributionpatterns of plaice have shown that they were primarilyfound over sandy substrata (Amezcua and Nash, 2001),so in the present study we limited our investigation tothis type of habitat to increase our chances of elucidatingthe relationship between habitat characteristics andplaice density.

The sites sampled were distributed across threeregions of the English Channel (see Fig. 1). Two sites(W1 andW2) were located off Start Point, Devon. Threesites were positioned off the English coast (NE1–NE3),between Brighton and Hastings. A further four siteswere located off the east coast of France near Boulogne-sur-Mer (SE1–SE4).

2.2. Fish survey

Four daytime tows of 20 min duration each weremade with a 4 m beam trawl, towed at a speed ofapproximately 4 knots, to quantify the densities of fishat each of the nine sites in August 2002. The beam trawlwas fitted with a chain-matrix and an 82 mm diamondmesh cod end with a 40 mm square meshed liner. Allfish were sorted into species, counted and weighed. Thetotal length of each plaice was recorded to the nearest

in the English Channel.

320 R. Shucksmith et al. / Journal of Sea Research 56 (2006) 317–328

cm. Plaice catches were standardised to a tow length of250 m covering an area of 1000 m2.

Differences in plaice abundances between sites wereanalysed by the non-parametric Kruskal-Wallis test, asthe data did not fulfil the assumptions of ANOVA. Totalabundances of plaice caught, as well as the abundance ofonly plaice over the minimum landing size (>26 cm),were analysed for differences among sites.

To evaluate whether plaice abundances at the ninesites followed the same trends as those predicted by thelong-term groundfish survey data, the mean number offlatfish caught above the minimum landing size (MLS)was correlated with the mean Log10-percentage plaiceabundance sampled by CEFAS over nine years (Pearsonproduct-moment correlation). All further analyses intothe relationship of habitat characteristics identified fromvideo tows and plaice densities were based on fishabundance as recorded concomitantly with the habitatcharacteristics.

2.3. Video survey

To assess the composition of habitat structure andbenthic faunal assemblages an underwater camerasystem was deployed prior to sampling with the 4 mbeam. At each site, one 30 min camera tow along theseabed was conducted with a video sledge. The sledgewas mounted with a UWTV video system Photosea1000 arranged with the camera pointing downwards atan angle of 45° and with the lens approximately 0.7 mabove the seabed. The field of view spanned an area of0.2 m2. Lights were mounted at 60° to the camera. Thevideo images were recorded using a Hi-8 video cassetteplayer.

The video sledge was towed at the speed of thecurrent acting upon the vessel. The average videotransect was 561±214 m and towing speeds rangedbetween 3.8 cm and 62.9 cm s−1.

Digital images were freeze-framed and extractedfrom the video recordings made during individual towsat intervals of 20 s. Ninety images were extracted foreach video tow. To standardise for the differences in towlength and allow analysis of images from equaldistances traversed across the seabed, images wereallocated into consecutive 50 m sections that weredetermined from the vessel speed and time of recording.Each 50 m section thus had a different number of imagesand was therefore sub-sampled. For each 50 m sectionof each tow, five images were randomly selected fordetailed analysis. Five images proved to conservesufficient habitat information for the purposes ofquantifying habitat characteristics. This was determined

by using cumulative sample curves calculated forhabitat characteristics. The use of fewer than fiveimages did not identify all quantifiable features withina 50 m bin interval. To make the data of each towcomparable, the overall tow length for each tow wasstandardised by randomly selecting an equal number of50 m intervals for each site. As a result an equal towlength of 200 m and 20 images per site was used foranalysis of physical substratum features, so that scaleand sampling effort were held constant.

Extracted video images were analysed for differencesin substratum type and structures by projecting theimage on a screen with a grid divided into 100 equalcells. The percentage cover of sediment types per imagewas estimated by calculating the number of cellscovered by a particular sediment type (Table 1).Different sediment types were identified by their colourand grain size as seen in the image and verified by grainsize analysis of sediment samples collected with a Daygrab at each site. Sand ripples were recorded as presentor absent and all other physical structures were counted(Table 1).

The habitat characteristics data were subjected to acluster analysis to generate a dissimilarity matrix(PRIMER v.5) after calculating the standardised Eucli-dian distance between each pair-wise comparison ofsites. Those habitat features that contributed most to thesimilarity among groups of sites (identified a posteriorifrom the cluster analysis) were determined usingSIMPER analysis (PRIMER v.5).

Habitat heterogeneity within sites was determinedto assess spatial differences in the distribution ofphysical features between sites (Table 1). At each sitea dissimilarity matrix was calculated from the 20video images based on Euclidean distance for eachpair-wise comparison of video images. From thismatrix the mean dissimilarity and its correspondingstandard deviation was calculated. The mean dissim-ilarity of all images of one site was then used as anapproximation of habitat heterogeneity. Sites with alow mean image dissimilarity indicated homogeneoussubstratum characteristics, while sites with high meanimage dissimilarity indicated heterogeneous substra-tum characteristics. To test whether mean substratumdissimilarity (heterogeneity) was correlated with plaiceabundance the Pearson's product-moment correlationtest was used. Both heterogeneity and plaice densitywere Log10-transformed prior to analysis. The rela-tionship between depth and flatfish abundance wastested in a separate correlation analysis.

From the video images, benthic organisms wereidentified to the lowest taxonomic level possible and

Table 1Summary of plaice caught, substratum characteristics and benthic fauna at respective sites caught

English West English East French East

W1 W2 NE1 NE2 NE3 SE1 SE2 SE3 SE4

PlaiceTotal number of plaice caught 21 14 9 22 13 330 76 189 26Mean number of plaice per tow 5.2 3.5 2.2 5.5 3.2 82.5 19 47.2 6.5Mean number of plaice >26 cm per 1000 m2 0.33 0.27 0.22 0.17 0.28 5.32 1.56 0.18 0.44Mean length in cm 30 34 32 24 29 32 32 20 33Min length in cm 26 28 26 18 24 22 31 12 26Max length in cm 38 44 38 32 35 42 43 35 48

Substratum characteristics⁎ Mean % cover of shelly gravelly sand 52.3 1⁎ Mean % cover of sand 100 87 96.8 97.7 100 93 36.3 100 99⁎ Mean % cover of small and broken shell 13 3.2 2.2 11.4⁎ Small sand ripples 0–5 cm (% P/A) 100 70 100 55 100 5⁎ Medium sand waves >5 cm (% P/A) 1 0.4⁎ Large stones >5 cm (C) <0.1 0.4⁎ Stones 0–5 cm (C) 16.8 7 0.3⁎ Small burrow 0–2 cm (C) 0.5⁎ Large burrow 2–5 cm (C) 0.7

Benthic fauna per 50 m sectionPorifera (C) 0.06Anthozoa (C) 0.06Urticina felina (C) 0.2Polychaeta tube structures (E) ≈2.05 ≈0.04 ≈478 ≈57.5 ≈73.5Pagurus spp. (C) 0.05Asterias rubens (C) 0.05 0.23 0.2Ophiura spp. (C) 0.06Ophiothrix fragilis (E) ≈21.6Echinocardium spp. (C) 0.05

Benthic fauna per 1 km towPorifera (C) 1.3 2.2 4.2Anthozoa (C) 15.5 1.8 5.1Urticina felina (C) 7.1 5.1Polychaeta tube structures (E) ≈6 ≈1 ≈44886 ≈16539 ≈3724Pagurus spp. (C) 2.5 5 10.8 1.5 5.9Liocarcinus spp. (C) 1.4 1.9Goneplax rhomboides (C)Mytilus edulis (C) 268.8Asterias rubens (C) 5.7 46.5 22.8 3.9 4.1Ophiura spp. (C) 1.4 159.4Ophiothrix fragilis (E) ≈2795Echinocardium spp. (C) 1 51 3.5Attached epifauna (unidentified) (C) 1.3 1.4 5.1Fish (unidentified) (C) 1.9 1.1

Substratum type cover was given as mean percentage per analysed image for each site. Sand ripples and waves were recorded as percentage presenceor absence from images taken at each site (% P/A) while stones were counted (C). Epifauna was counted (C) or estimated (E). Mean abundance ofepifauna per 50 m section of a 200 m tow and for the whole tow and standardised to a transect distance of 1 km. Asterisks indicate the category used tocalculate the heterogeneity indices.

321R. Shucksmith et al. / Journal of Sea Research 56 (2006) 317–328

counted (Table 1). The mean number of benthic animalswas calculated for each 50 m section (mean of 5 images)of each tow. From this the overall mean abundance per200 m camera tow was determined. While the largerepifauna were easily counted, smaller and hyper-

abundant species were difficult to quantify. Due to thelow quality of some images it was often difficult tocount protruding worm tubes or brittle stars in all partsof an image. Therefore worm tube density was estimatedfor most images using the screen grid as an aid to

322 R. Shucksmith et al. / Journal of Sea Research 56 (2006) 317–328

extrapolate numbers from well-defined areas to lessdefined ones.

Owing to the low number of taxa observed in six outof the nine sites, an additional analysis was performedthat involved analysis of the entire video transect (90images) for each site. The abundances of benthic specieswere subsequently standardised to a common transectdistance of 1 km. These semi-quantitative data wereincluded as they contained valuable information on thebenthic assemblages found at each of the sitesinvestigated that otherwise would have been lost.

The mean number of benthic animals per 50 m sectionfor each site (200 m tow) was correlated with plaice den-sities by Pearson product-moment correlation. The totalabundance of benthic animals and the number of taxaobserved for the complete tow (semi-quantitative data)were also correlatedwith plaice densities (Pearson product-moment). The abundances of benthic organisms and thenumber of taxa were Log10-transformed prior to analysis.

Fig. 2. Scatter plots of mean Log10-abundance of plaice >26 cm per 1000 m2

surveys (CEFAS), (b) mean Log10-dissimilarity (heterogeneity) of substratumnumber of benthic taxa fauna recorded at each site.

3. Results

3.1. Plaice densities

The median number of plaice caught at each of thenine sites differed significantly (K–W, H=27.06, d.f.=8, p=<0.001). Sites SE1, SE2 and SE3 off the Frencheast coast had significantly higher median plaicedensities than all other sites. The median abundance ofplaice per tow above the 26 cm minimum landing size(MLS) differed significantly among the nine sitessampled (K–W, H=27.79, d.f.=8, p=0.001). SitesSE2 and SE1 had significantly higher median abun-dance of plaice above the 26 cm MLS per tow comparedwith all other sites. This indicated that at sites SE1 andSE2 a greater median number of larger plaice werecaught than at any of the other stations (Table 1).

There was a significant correlation between CEFASlong-term mean Log10-percentage abundance of plaice

versus: (a) mean Log10-abundance of plaice from long-term groundfishcharacteristics, (c) mean Log10-abundance of benthic fauna, (d) Log10-

323R. Shucksmith et al. / Journal of Sea Research 56 (2006) 317–328

over nine years and the mean Log10-abundance of plaicecaught per tow (r=0.68, d.f.=8, p=0.04, Fig. 2). Thisindicated that the abundance of fish sampled during thepresent study in general could be predicted from thelong-term CEFAS groundfish survey data. However, thecorrelation was strongly influenced by the three siteslocated off the French coast: SE1, SE2 and SE4. Thesesites had the highest mean abundances of plaice (aboveMLS) in the long-term data of CEFAS, which matchedwith the abundances of plaice caught at these sites by thepresent study.

3.2. Substratum characteristics, habitat heterogeneityand depth

The analysis of the substratum characteristics showedthat although the seabed at all sites was primarilycomposed of sand, subtle habitat differences betweensites were apparent (Table 1 and Fig. 3). Sites clusteredinto two main groups at a dissimilarity level of 65%.Group one consisted of sites SE1, SE3, SE4 and NE3.The SIMPER analysis showed that these sites werecharacterised by a high mean percentage cover of sand(mean±SD 98±3.6) and the presence of medium sandwaves (mean percentage occurrence 36.2±47.5). Themean number (±SD) per video frame of small (1.83±3.4) and large stones (0.13±0.21) also characterisedthese sites. The second group consisted of sites NE1,

Fig. 3. Cluster analysis of substratum characteristics showing dissimilaritieidentified by SIMPER analysis were summarised below station clusters, withAbundance of plaice caught at each site per 1000 m2: SE1 (5.32); SE3 (0.18);SE2 (1.56).

NE2, W1, W2 and SE2. Site SE2 clustered separatelyfrom the highly similar cluster of the remaining sites andwas analysed independently (Fig. 3). Sites NE1, NE2,W1 and W2 had a slightly lower percentage sand coverthan group one sites (mean±SD 95.3±5.7), while videoimages were also characterised by a low meanpercentage cover of small and broken shells (4.6±5.7).Small sand ripples on average occurred in 81% (±22.5)of the images from these sites. Site SE2 showed sandripples (100%) and a lower mean percentage cover ofsmall and broken shells (11.5±27.9). Site SE2 alsodiffered from the other sites in its mean percentage coverof sand (36± 44) and it was the only site where apercentage cover of shelly gravelly sand (mean±SD 52±49.2) was recorded. The percentage contributions ofeach substratum characteristic to the percentage similar-ity among sites in each cluster are summarised in Fig. 3.There was no clear relationship between the posterioiridefined site (habitat) groupings (Fig. 3) and plaicedensities recorded during this study. Thus sites with ashigh and low densities of plaice occurred within bothgroupings of sites defined on the basis of substratumcharacteristics.

Calculation of the mean dissimilarity of images ofeach of the sites as an approximation of physical habitatheterogeneity showed that there were distinct differ-ences in habitat heterogeneity among sites (Table 2).There was, however, no significant correlation between

s of station clusters in percentages. Characteristics of station clusterstheir corresponding percentage contribution to that cluster (contrib. %).NE3 (0.28); SE4 (0.44); NE1 (0.22); W1 (0.33); W2 (0.27) NE2 (0.17);

Table 2Site, location, depth and a general description of the habitats recordedby video tows

Station Area Deptm

Station description Mean dissimilarity/heterogeneityIndex±S.D.

W1 Englishwest

70 Sand substratum,small sand rippleswith detritus inthe troughs.

0±0

W2 Englishwest

68 Sand substratumwith small shellcontent. Smallsand ripples.

27.93±45.05

NE1 Englisheast

16 Sand substratumwith small shelland stones withsand ripples andoccasional burrows.Occasional clumpof Mytilus edulis.

12.54±8.16

NE2 Englisheast

20 Sand substratum,some broken shell,and occasionalsand wave. Someburrows presentswith occasionalworm tubes.

3.37±2.96

NE3 Englisheast

21 Sand substratumwith small sandripples. Occasionalhermit crab.

1.16±1.26

SE1 Frencheast

22 Sand substratumwith some patchesof stones witha tube mat cover.

18.86±39.27

SE2 Frencheast

29 Sand substratumwith patchy tubemat cover. Patchesof shelly gravellysand. Some smallsand ripples.

81.79±60.13

SE3 Frencheast

20 Sand substratumwith worm tubesat low density.

0.14±0.42

SE4 Frencheast

52 Sand substratumwith Ophiothrixfragilis beds at thestart of the tow.Medium sand waves.

4.59±7.93

Mean dissimilarity (heterogeneity indices) and S.D. of physicalfactors. The higher the mean dissimilarity the greater the variation inphysical features encountered within a transect.

324 R. Shucksmith et al. / Journal of Sea Research 56 (2006) 317–328

mean number of plaice caught and habitat heterogeneityat each site (r=0.504, d.f.=8, p=0.166, Fig. 2).

No significant correlation could be found betweenthe water depth (Table 2) and the abundance of plaice ateach site (r=−0.195, df=8, p=0.615, Fig. 4). This is

not surprising as we deliberately chose sites from withina narrow depth range to eliminate the influence of thisenvironmental variable in the context of the presentstudy.

3.3. Benthic fauna

There was a significant positive correlation betweenthe mean abundance of benthic fauna recorded per 50 mat each site and plaice densities (r=0.75, d.f. =8,p=0.02, Fig. 2). Likewise, the total abundance ofbenthic organisms per video tow correlated significantlywith plaice abundance (r=0.676, d.f. =8, p=0.045). Thenumber of taxa sampled for the whole camera tow didnot significantly correlate with plaice densities(r=0.614, d.f.=8, p=0.079).

The correlations between the abundance of benthicfauna and plaice abundance were strongly influenced bythe sites from the east coast of France (SE1–SE4, seeFig. 2). These sites had the highest level of communitydiversity and the largest quantity of benthic fauna,which corresponded well with the abundance of plaiceat these sites. In particular, plaice abundance seemed tobe related to the presence of high-density polychaetetube mat structures at sites SE1–SE4 (Tables 1 and 2,Fig. 5). The tube mats were composed of thepolychaetes Lanice conchilega and a previously unde-scribed species of Cheatopterus sp. (Rees et al., 2005).

Site SE1 had the highest densities of emergentpolychaetes (mean±SD, 2390±470 m−2, Figs. 4 and 5)and had the highest densities of plaice caught during thisstudy (Table 1). On average 5.7 (±0.74) plaice werecaught 1000 m−2 with the majority of individuals (5.3±0.72) above the MLS (Fig. 5). Densities of polychaeteswere lower at site SE2 (288±575 m−2) and more patchyin their distribution. Adult plaice densities also werelower at this site (1.56±0.86 m−2). The density ofemergent polychaetes was similar at site SE3 and siteSE2 (367±185 m−2). Overall densities of plaice wererelatively high (3.55±2.08). Here, catches mainlyconsisted of small plaice below MLS (3.36±2.06, Fig.4). No emergent tubeworms were observed at site SE4(Fig. 5) which coincided with the lowest density ofplaice encountered at any of the French coast sites (0.46,S.D.±0.20). The majority of plaice caught were,however, above MLS (0.44±0.23). SE4 was dominatedby the brittlestar Ophiothrix fragilis which occurred inlarge aggregations at the beginning of the video cameratow (Table 2). At the remaining sites (W1, W2, NE1,NE2 and NE3) benthic organisms were only rarelyobserved (Table 1). The only other site with aconsiderable abundance of benthic epifauna was site

Fig. 4. Mean abundance of two different size classes of plaice eitherbelow or above legal commercial landing size (juvenile <26 cm, adult>26 cm) per 1000 m2 at the French coast sites SE1–SE4. Meanabundance of protruding Polychaeta tubes per 50 m section for a towof 200 m. Error bars represent S.D.

325R. Shucksmith et al. / Journal of Sea Research 56 (2006) 317–328

NE1, which was dominated by the common musselMytilus edulis (Table 1). Plaice densities at these siteswere much lower compared to the French coast sites and

Fig. 5. Selected representative images from the French coast stations SE1–camera on the video sledge. Images of SE1 and SE2 showing dense tube matmat cover was less dense and the sandy sediments at this site appear to be cowaves with shell fragments and small stones deposited in the troughs.

ranged on average between 0.17–0.33 animals per1000 m2 (2–5 fish per 20 min tow, see Table 1).

4. Discussion

4.1. Site selection and plaice abundance

The results of this study demonstrated that usinglong-term datasets from groundfish surveys to identifysites of consistent plaice abundance in the EnglishChannel was a relatively accurate predictor of therelative magnitude of replicated fish catches amongdifferent sites. Thus it would appear that someenvironmental component, or habitat feature at thesesites, or behaviour of flatfish in relation to these habitats,is relatively consistent across time.

4.2. Habitat descriptors and plaice densities

The present study indicated that high densities ofplaice were associated with areas that had a rich benthicfauna typified by dense mats of the tube-buildingpolychaetes Lanice conchilega and Cheatopterus spp.

SE4 taken with a stills camera mounted together with the underwatercover of Lanice conchilega and Chaetopterus spp. At site SE3 the tubevered with a thin layer of finer deposits. The image of SE4 shows sand

326 R. Shucksmith et al. / Journal of Sea Research 56 (2006) 317–328

However, this relationship was only observed at the sitesoff the French coast (SE1–SE4) as all other sites hadgenerally low abundances of tubicolous polychaetes andother benthic organisms. Therefore it should be notedthat the interpretation of these results relied profoundlyon these sites. Nevertheless, the corresponding gradientof benthic fauna and plaice densities observed withinthese four sites suggests the existence of a relationshipbetween fish abundance and prey abundance beyond acritical threshold. Tube-building polychaetes are knownto be an important component of plaice diet (Jones,1952; Wyche and Shackley, 1986; Piet et al., 1998;Rijnsdorp and Vingerhoed, 2001). Moreover, as tubi-colous polychaetes occurred in high abundances, searchtimes when feeding will be minimal, conserving energyand ultimately yielding a higher rate of food intake.Likewise at the site with the lowest polychaete tubecover (SE3), abundances of adult plaice (>MLS) werelow, while smaller plaice were found in high abun-dances. This aggregation of smaller size fish, on habitatsof lower prey resource value could be related to intra-specific competitive exclusion by the large plaiceoccupying prime feeding habitats (such as SE1 andSE2). The presence of mainly small plaice could also bedue to different habitat requirements between juvenileand adult flatfish. Predator avoidance is probably a morepressing issue for juvenile than adult flatfish (Gibson,1994), since a variety of crustaceans and fish prey uponthem leading to high mortality rates (Ansell and Gibson,1993). Protection from predators is achieved by thecryptic body colouration and the ability to bury into thesediment (Gibson and Robb, 1992, 2000; Ansell andGibson, 1993). The open spaces left by the less densepolychaete tube mat cover at site SE3 and the sandysubstratum may have fulfilled both habitat requirements(cover and food) of smaller plaice. Conversely, protec-tion for larger flatfish is thought to be less crucial as thenumber of potential predators decreases with increasingbody size (Gibson and Robb, 1992, 2000; Stoner andAbookire, 2002).

Using density of worm tubes, recorded at one point intime as a prognostic tool to predict distribution patternsof plaice for longer time periods, may prove to bedifficult. Populations of Lanice conchilega for examplehave an ephemeral and patchy distribution (Zühlke,2001) and densities are likely to change through time.However, there is evidence that the area off the Frenchcoast had a similar faunal composition in the early 1970s,indicating some degree of long-term faunal stability.Sanvicente-Anorve et al. (2002) analysed dredgesamples collected in 1971–1975 in the eastern EnglishChannel and reported a distinct Abra alba assemblage

which was characterised by a high species richness andhigh abundances of Abra alba and Lanice conchilega.The long-term benthic productivity within the area islikely to be linked to local hydrodynamic regimes. Thenutrient-rich discharges from the rivers Seine andSomme, together with the eddy caused by the ContentinPeninsula that entrains these nutrient-rich and productivewaters (Tappin and Reid, 2000), generate a rich foodsupply for the resident benthic fauna (Hoch and Garreau,1998).

Other habitat parameters besides food availabilityhave been suggested as important for flatfish inparticular structuring components such as certain bedforms, e.g. sand waves (Norcross and Mueter, 1999) oremergent epifauna e.g. Porifera, Anthozoa or Bryozoa(Stoner and Titgen, 2003) which are thought to giveadditional protection from predators. Norcross andMueter (1999) showed that juvenile flatfish were oftenassociated with structures such as biogenic depressionsand troughs between sand waves. They also showed thatflatfish were randomly distributed over a uniform seabedbut had a clustered distribution on heterogeneoussubstrata. Heterogeneous sediments may affect buryingcapabilities and/or prey distribution, thus influencingoverall flatfish distribution patterns. No significantrelationship was apparent between plaice density andthe substratum features and heterogeneity recorded bythis study. However, the plaice sampled in the presentstudy were relatively large (>26 cm) and thereforeparameters important for predator avoidance may not besuch an important factor influencing their distribution(see above). Similar conclusions for adult plaice mayhold true for the protective function of structuringepibenthic fauna. Nevertheless, this habitat componentmay still be important for adult plaice as it may providehabitat niche spaces for potential prey organisms. Thediet of larger adult plaice, besides polychaetes, alsocontains a large proportion of epibenthic crustaceans andsmall fish (Carter and Grove, 1991; Piet et al., 1998)which may benefit from the habitat provided byemergent epibenthic fauna (Bradshaw et al., 2003).

4.3. Methodological evaluation and sampling scale

This study demonstrates the value of seabed imageryin the characterisation of habitat preferences of fish.Unlike other sampling methods, video transects allowreal-time sampling of the seabed and observation of thespatial distribution and density of both abiotic and bioticfeatures. However, the low resolution of the digitalcamera makes the detection and identification of smallorganisms such as amphipods, shrimps and protruding

327R. Shucksmith et al. / Journal of Sea Research 56 (2006) 317–328

worms sometimes difficult or impossible. Burrowingand small fauna may remain undetected amongst otherbenthic fauna. Video imagery can therefore not substitutesamples taken by trawls and grabs, which sample crypticfauna more efficiently. However, larger areas can becovered by this method and this study demonstrated thatmeaningful relationships between flatfish and habitatparameters could be established. As the location of thesites sampled were partly separated by considerablegeographical distances differences in flatfish densityobserved may not necessarily reflect differences inhabitat quality but indicate differences in populationswithin regional geographical areas. Indeed CEFAS(unpubl. data) assumes the presence of two distinctsub-stocks of plaice in the western and eastern EnglishChannel. These two stocks may vary in their populationsize and give rise to natural variation in the density offlatfish. One way to reduce the confounding variableswould be to compare areas of high and lower flatfishdensity in a more restricted geographical area, so as toreduce variability due to large-scale geographicaldifferences. This smaller-scale orientated approachseems to be supported by the data from the Frenchcoast. Sites were positioned in relatively close vicinity toeach other and the trends observed here between plaiceand benthic fauna may indicate that a sampling regimeon a smaller geographical scale may be more successfulin studying distribution patterns and thus habitatrequirements of plaice than the sea basin scale. However,the results of such a smaller-scale study would havelimited power as the conclusions would only beapplicable to a very restricted area. To date, it is unclearat what scale the mechanisms of habitat selection operate(Tyler and Hargrove, 1997). This demonstrates one ofthe inherent difficulties in the identification of essentialfish habitats and habitat requirements of fish in the fieldand needs further scientific attention.

Acknowledgements

This study was financially supported by the Depart-ment for Environment Food and Rural Affairs (DEFRA)contract MF0805. We would also like to thank Ivor Reesand Jan Hiddink for their advice and support throughoutthe project.

References

Abookire, A.A., Norcross, B.L., 1998. Depth and substrate asdeterminants of distribution of juvenile flathead sole(Hippoglossoides elassodon) and rock sole (Pleuronectes bilinea-tus), in Kachemak Bay, Alaska. J. Sea. Res. 39, 113–123.

Albert, O.T., Eliassen, J.E., Hoines, A., 1998. Flatfishes of Norwegiancoasts and fjords. J. Sea. Res. 40, 153–171.

Amezcua, F., Nash, R.D.M., 2001. Distribution of the orderPleuronectiformes in relation to the sediment type in the NorthIrish Sea. J. Sea. Res. 45, 293–301.

Ansell, A.D., Gibson, R.N., 1993. The effect of sand and light onpredation of juvenile plaice (Pleuronectes platessa) by fishes andcrustaceans. J. Fish. Biol. 43, 837–845.

Auster, P.J., Langton, R.W., 1999. The effects of fishing on fishhabitat. Am. Fish. Soc. Symp. 22, 150–187.

Bigelow, H.B., Schroeder, W.L., 2002. Fishes of the Gulf of Maine.Smithsonian Institution Press, Washington, DC.

Bradshaw, C., Collins, P., Brand, A.R., 2003. To what extent doesupright sessile epifauna affect benthic biodiversity and communitycomposition? Mar. Biol. 143, 783–791.

Carter, C.G., Grove, D.J., 1991. Trophic resource partitioning betweentwo coexisting flatfish species off the north coast of Anglesey,North Wales. J. Sea. Res. 27, 325–335.

Ellis, J.R., Rogers, S.I., Freeman, S.M., 2000. Demersal assemblagesin the Irish Sea, St George's Channel and Bristol Channel. Estuar.Coast. Shelf Sci. 51, 299–315.

Gibson, R.N., 1994. Impact of habitat quality and quantity on therecruitment of juvenile flatfishes. J. Sea. Res. 32, 191–206.

Gibson, R.N., Robb, L., 1992. The relationship between body size,sediment grain-size and the burying ability of juvenile plaice,Pleuronectes platessa. J. Fish. Biol. 40, 771–778.

Gibson, R.N., Robb, L., 2000. Sediment selection in juvenile plaiceand its behavioural basis. J. Fish. Biol. 56, 1258–1275.

Hinz, H., Kaiser, M.J., Bergmann, M., Rogers, S.I., Armstrong, M.,2003. Ecological relevance of temporal stability in regional fishcatches. J. Fish. Biol. 63, 1219–1234.

Hoch, T., Garreau, P., 1998. Phytoplankton dynamics in the EnglishChannel: a simplified three-dimensional approach. J. Mar. Syst.16, 133–150.

Jennings, S., Kaiser, M.J., 1998. The effects of fishing on marineecosystems. Adv. Mar. Biol. 34, 201–352.

Jones, N.S., 1952. The bottom fauna and the food of flatfish off theCumberland Coast. J. Anim. Ecol. 21, 182–205.

Kaiser, M.J., Cheney, K., Spence, F.E., Edwards, D.B., Radford, K.,1999. Fishing effects in northeast Atlantic shelf seas: patterns infishing effort, diversity and community structure VII. The effectsof trawling disturbance on the fauna associated with the tubeheadsof serpulid worms. Fish. Res. 40, 195–205.

Kaiser, M.J., Collie, J.S., Hall, S.J., Poiner, I.R., 2002. Modification ofmarine habitats by trawling activities: prognosis and solutions.Fish Fish. 3, 114–136.

Link, J.S., 2002. Ecological considerations in fisheries management:when does it matter? Fisheries 27, 10–17.

McConnaughey, R.A., Smith, K.R., 2000. Associations betweenflatfish abundance and surficial sediments in the eastern BeringSea. Can. J. Fish. Aquat. Sci. 57, 2410–2419.

Meester, G.A., Mehrotra, A., Ault, J.S., Baker, E.K., 2004. Designingmarine reserves for fishery management. Manag. Sci. 50,1031–1043.

Norcross, B.L., Mueter, F.J., 1999. The use of an ROV in the study ofjuvenile flatfish. Fish. Res. 39, 241–251.

Piet, G.J., Pfisterer, A.B., Rijnsdorp, A.D., 1998. On factorsstructuring the flatfish assemblage in the southern North Sea. J.Sea. Res. 40, 143–152.

Rees, E.I.S., Bergmann, M., Hinz, H., Galanidi, M., Shucksmith, R.,Kaiser, M.J., 2005. An enriched Chaetopterus tube mat biotope inthe eastern English Channel. J. Mar. Biol. Ass. UK 84, 323–326.

328 R. Shucksmith et al. / Journal of Sea Research 56 (2006) 317–328

Rijnsdorp, A.D., Vingerhoed, B., 2001. Feeding of plaice Pleuro-nectes platessa L. and sole Solea solea (L.) in relation to the effectsof bottom trawling. J. Sea. Res. 45, 219–229.

Rogers, S.I., Rijnsdorp, A.D., Damm, U., Vanhee, W., 1998. Demersalfish populations in the coastal waters of the UK and continentalNW Europe from beam trawl survey data collected from 1990 to1995. J. Sea. Res. 39, 79–102.

Ryer, C.H., Stoner, A.W., Titgen, R.H., 2004. Behavioral mechanismsunderlying the refuge value of benthic habitat structure for twoflatfishes with differing anti-predator strategies. Mar. Ecol. Prog.Ser. 268, 231–243.

Sanvicente-Anorve, L., Lepretre, A., Davoult, D., 2002. Diversity ofbenthic macrofauna in the eastern English Channel: comparisonamong and within communities. Biodiv. Conserv. 11, 265–282.

Smale, M.J., Roel, B.A., Badenhorst, A., Field, J.G., 1993. Analysis ofthe demersal community of fish and cephalopods on the AgulhasBank, South-Africa. J. Fish. Biol. 43, 169–191.

Stoner, A.W., Abookire, A.A., 2002. Sediment preferences and size-specific distribution of young-of-the-year Pacific halibut in anAlaska nursery. J. Fish. Biol. 61, 540–559.

Stoner, A.W., Titgen, R.H., 2003. Biological structures and bottomtype influence habitat choices made by Alaskan flatfishes. J. Exp.Mar. Biol. Ecol. 292, 43–59.

Tappin, A.D., Reid, P.C., 2000. The English channel. In: Sheppard, C.(Ed.), The Seas at the Millenium: An Environmental Evaluation.Vol. I. Regional Chapters: Europe, the Americas and West Africa.Pergamon, Amsterdam, pp. 65–82.

Tyler, J.A., Hargrove, W.W., 1997. Predicting spatial distribution offoragers over large resource landscapes: a modeling analysis of theIdeal Free Distribution. Oikos 79, 376–386.

Wyche, C.J., Shackley, S.E., 1986. The feeding ecology of Pleuro-nectes platessa L., Limanda limanda (L.) and Scopthalmusrhombus (L.) in Carmarthen Bay, South Wales, U.K. J. Fish.Biol. 29, 303–311.

Zühlke, R., 2001. Polychaete tubes create ephemeral communitypatterns: Lanice conchilega (Pallas, 1766) associations studiedover six years. J. Sea. Res. 46, 261–272.