Marine Pollution Bulletin 57 (2008) 652–661

Contents lists available at ScienceDirect

Marine Pollution Bulletin

journal homepage: www.elsevier .com/locate /marpolbul

Amelioration of marine farming impact on the benthic environment usingartificial reefs as biofilters

Qin-Feng Gao a, Paul K.S. Shin b,c, W.Z. Xu b, S.G. Cheung b,*

a College of Fisheries, Ocean University of China, 5 Yushan Road, Qingdao City, Shandong Province 266003, Chinab Department of Biology and Chemistry, City University of Hong Kong, Tat Chee Avenue, Kowloon, Hong Kong SAR, Chinac Centre for Coastal Pollution and Conservation, City University of Hong Kong, Tat Chee Avenue, Kowloon, Hong Kong SAR, China

a r t i c l e i n f o

Keywords:

Fish farmingArtificial reefNutrient enrichmentBenthic communityBiofiltration

0025-326X/$ - see front matter � 2008 Elsevier Ltd.doi:10.1016/j.marpolbul.2008.02.033

* Corresponding author. Tel.: +852 2788 7749; fax:E-mail address: bhsgche@cityu.edu.hk (S.G. Cheun

a b s t r a c t

An in situ monitoring of the sediment characteristics and macrobenthic communities was undertaken at amarine fish culture site in subtropical waters of Hong Kong before and after the deployment of biofilterswhich were made of cement concrete artificial reef (AR) structures. According to the distance to theboundary of the fish cages, 6 points were selected as sampling stations: 2 at the fish cages, 2 near theboundary of the fish culture area, and 2 reference sites further away from the culture area. Bimonthlysediment samples were collected for analysis of silt-clay fraction (SCF), moisture content (MC), totalorganic carbon (TOC), total Kjeldahl nitrogen (TKN) and total phosphorus (TP). The macrobenthos(>0.5 mm) present in the sediment were sorted, identified and enumerated. TOC, TKN and TP levels atthe fish cage stations were consistently higher than those at the reference stations over the 1-yearpre-AR and 2-year post-AR deployment monitoring period. The diversity of macrofauna was significantlyreduced at the fish cage stations relative to the reference sites. The intermediary stations near the fishculture area showed a transitional state of disturbance. Over the 2-year post-AR deployment period,TOC, TKN and TP showed a decreasing trend at the fish cage and intermediary stations. More diverse mac-rofaunal communities were recorded at the fish cage stations, with species diversity H0increasing from 0–1 at the beginning of the AR deployment to H0 > 2 at the end of the study. The present results demon-strated that artificial reefs can improve the benthic abiotic environment and biotic conditions beneathfish rafts which are deteriorated due to farming activities.

� 2008 Elsevier Ltd. All rights reserved.

1. Introduction

Fish farming activities release a large amount of wastes in theform of feed residue, faeces and excreta. The wastes either suspendor dissolve in the surrounding waters at the beginning stage. Aconsiderable part of such farming wastes, however, will eventuallysettle and be accumulated on the seabed beneath or near the fishculture zone. Subsequent decay of the deposited wastes consumessubstantial dissolve oxygen and releases various substances harm-ful to the culture species, resulting in the deterioration of the farm-ing environment (see Wu, 1995; Pearson and Black, 2001 forreview).

Filter feeders, such as bivalves, sponges, hydrozoans, sea anem-ones, hydroids, anthozoans and polychaetes, dominating sublittor-al benthic communities on hard substrata can capture largequantities of particles from the water column owing to their denseabundance and high filtration efficiency. During the feeding activ-ities of these sessile organisms, suspended particulate matter is re-

All rights reserved.

+852 2788 7406.g).

assembled and re-processed, and the biodeposits produced bythese organisms lead to changes in physico-chemical characteris-tics of the seabed sediment once settling from the water column(Jørgensen, 1990; Graf and Rosenberg, 1997). Jordan and Valiela(1982) investigated the nitrogen budget of the ribbed mussels Geu-kensia demissa in a marsh ecosystem and found that nearly half ofthe nitrogen filtered from the water column by the mussel popula-tion was absorbed. Thus G. demissa increased retention of nitrogenwithin the marsh by filtering particulate nitrogen from surround-ing waters and stored nitrogen as biomass. Bivalves and other fil-ter-feeding organisms can selectively absorb more nutritionalconstituents from the filtered matter, and transfer such nutrientsas biodeposition on the seabed in the form of pseudofaeces andfaeces (Gao et al., 2002).

In aquaculture zones, Angel et al. (2002) reported that resi-dent organisms on artificial reefs which were deployed under fin-fish cages could effectively remove particulate matter fromsurrounding waters, resulting in marked reduction in chlorophylla when water traversed the artificial reefs. In an integrated aqua-culture system combining fish and filter-feeding mussels, themussels can ecologically function as a self-regulator for cycling

Q.-F. Gao et al. / Marine Pollution Bulletin 57 (2008) 652–661 653

of particulates, reducing organic pollution caused by wastes fromfish farming activities (Folke and Kautsky, 1989). Our previousstudies at the present experimental site also showed that filer-feeding mussels could efficiently remove fish farming wastes andother particulate matter from the surrounding water column(Gao et al., 2006, 2008).

Benthic invertebrates are ideal indicators for environmentalconditions because they are relatively non-mobile, and theirwell-being and diversity thus tend to reflect the status of the areabeing sampled (Reynoldson and Day, 1993). Since benthic organ-isms are directly associated with the seabed habitats, any changesin sediment characteristics will affect the structure and composi-tion of the benthic communities subsequently.

In the present study, artificial reefs (ARs) were deployed at KauSai fish culture zone of Hong Kong in April 2002. Baseline investi-gations on the biotic and abiotic aspects at the experimental sitebefore the AR deployment were conducted for 1 year. Furtherpost-AR deployment monitoring was continued. The results ofbaseline investigations before the AR deployment, especially thoseregarding the effects of farming activities on benthic communitystructure and nutrient enrichment, have been reported (Gaoet al., 2005). The present paper focuses on the changes in bioticand abiotic conditions in terms of sediment characteristics andbenthic community structure after the AR deployment comparedwith those of the pre-AR deployment period, so as to investigatethe effects of AR structures on the seabed conditions at the studysite.

2. Materials and methods

2.1. Sampling site and AR design

The study site Kau Sai marine fish culture zone which is locatedin the eastern waters of Hong Kong (22�210N and 114�190E) is asemi-enclosed bay (Fig. 1). The fish culture zone is confined tothe inner part of the bay and the total area is approximately

Fig. 1. A Hong Kong map showing the s

4.6 ha with water depth of 11–16 m. The total fish stock is�500 t with average density of 4.5 kg m�3. Cultured species in-clude grouper (Epinephelus areolatus), snapper (Lutjanus russellii)and seabream (Acanthopagrus latus). Small trash fish (mainlyanchovies, Thryssa spp.) is used as fish feed. Daily feed supply is3–5% total stock, i.e., 15–25 t day�1.

According to the distance to the boundary of fish farms, 6 pointswere selected as sampling stations: two reference stations KS1 andKS6, two intermediary stations KS2 and KS5, and two fish cage sta-tions KS3 and KS4. The distances of reference and intermediary sta-tions to the boundary of fish farms were 600 and 100 m,respectively (Fig. 1). In April 2002, 16 AR sets made of cement con-crete (Fig. 2) with dimensions of 3 m (length) � 3 m (width) � 4 m(height) and total surface area of 250 m2 per AR were deployedaround the fish culture zone boundary.

Bimonthly field sampling was conducted in the pre-AR deploy-ment period from August 2001 to April 2002 and the post-ARdeployment period from August 2002 to April 2004. For each sam-pling, sediment was collected from the 6 stations using a 0.05 m2

van Veen grab with 7 L collection volume. A preliminary samplingprior to the present study showed that more than 90% species pres-ent at the study sites could be collected with a pool of five grabreplicates. Therefore, five grabs of sediment samples were col-lected at each of the six stations for benthic faunal analysis andthree additional grabs were collected for physical and chemicalanalyses (see below for details). Sediment samples were containedin individual plastic bags and transferred to the laboratory within2 h after field collection.

2.2. Sorting and identification of macrobenthos

In the laboratory, sediment samples for faunal analysis wereimmediately sieved through a screen with 500 lm mesh size. Res-idues remaining on the screen were fixed in 5% borax-buffered for-malin and stained with 1% Rose Bengal. The residues were hand-sorted, and all macrobenthic specimens were picked and preserved

tudy site and six sampling stations.

Fig. 2. An artificial reef set deployed at the study site.

654 Q.-F. Gao et al. / Marine Pollution Bulletin 57 (2008) 652–661

in 70% alcohol. All animals were taxonomically identified to thelowest level, if possible, and enumerated using a dissecting and acompound microscope, as necessary.

2.3. Measurements of sediment physico-chemical characteristics

Sediment samples for physico-chemical analyses were kept in�20 �C freezer prior to laboratory tests. Sediment physico-chemi-cal characteristics, including silt-clay fraction (SCF), moisture con-tent (MC), concentrations of total organic carbon (TOC), totalKjeldahl nitrogen (TKN) and total phosphorus (TP), were deter-mined following the methods of Mudroch et al. (1997). In sum-mary, triplicate wet sediment samples with known weight weredried at 40 �C for 72 h and the final dry weight was measured tothe nearest 0.1 mg, and the MC was calculated. SCF (<63 lm) ofthe sediment was determined by dry sieving method. TOC wasmeasured with the wet digestion method. Organic carbon in sedi-ment samples with specific weight was oxidized with acidified(concentrated sulphuric acid) dichromate (K2Cr2O7) followed bytitration with ferrous sulphate (FeSO4). To determine the concen-tration of TKN and TP, the sediment was digested with concen-trated sulphuric acid using a digestion block, at 200 �C for 30 minfor digestion of detritus and 370 �C for 2 h for digestion of other or-ganic nitrogen constituents. Copper sulphate (CuSO4) was used ascatalyst and potassium sulphate (K2SO4) was added to raise theboiling point of the digesting acid. The concentrations of TKNand TP were determined with a Flow Injection Analyzer (FIA, La-chat, QuikChem 8000).

2.4. Data analysis

Species number and abundance of each species were countedfor each sample. Species diversity of the 5-pooled samples at eachsampling station was calculated using Shannon-Wiener’s index(H0) (Shannon and Weaver, 1963). Similarity of benthic communitybetween sampling stations was calculated using the Bray-Curtiscoefficient (Bray and Curtis, 1957). The data of benthos abundancefor each monitoring month were pooled and analyzed with multi-dimensional scaling (MDS) ordination and Analysis of Similarity

(ANOSIM) test to explore the changes in benthic communities ofpost-AR deployment vs. those of pre-AR deployment (Clarke andWarwick, 2001). To reduce the disturbance of redundant specieson the multivariate analyses, those species of which total individ-ual number for all samples were less than 0.1% of the total abun-dance for all species were regarded as rare species and were notincluded in the analysis (Burd et al., 1990).

To explore the effects of AR deployment, the sediment datawere compared between the same months of pre-AR deployment(year 0), and 1st and 2nd year (referred to years 1 and 2, respec-tively) of post-AR deployment using repeated-measures multivar-iate analysis of variance (MANOVA). Once an overall difference wasfound with repeated-measures MANOVA, each single parameterfor each sampling station was compared between years 0, 1 and2. The percentage values were arcsine-transformed prior to com-parison analysis (Zar, 1999).

3. Results

3.1. Sediment characteristics

Sediment characteristics, including silt-clay fraction (SCF),moisture content (MC), total organic carbon (TOC), total Kjeldahlnitrogen (TKN) and total phosphorus (TP), of the 6 sampling sta-tions throughout the 1-year pre-AR and 2-year post-AR deploy-ment period are illustrated in Fig. 3. SCF at the fish cage stations(KS3 and KS4) was consistently lower relative to the intermediary(KS2 and KS5) (t5 = �10.5, p < 0.05) and the reference (KS1 andKS6) (t5 = �16.4, p < 0.05) stations. In contrast, the MC values ofKS3 and KS4 was higher than the intermediary (t5 = 5.1, p < 0.05)and reference stations (t5 = 7.4, p < 0.05). Annual average concen-trations of TOC, TKN and TP at fish cage stations were higher thanintermediary (TOC, t5 = 5.1, p < 0.05; TKN, t5 = 4.3, p < 0.05; TP,t5 = 3.1, p < 0.05) and reference (TOC, t5 = 7.4, p < 0.05; TKN,t5 = 4.5, p < 0.05; TP, t5 = 3.2, p < 0.05) stations.

Overall repeated-measures MANOVA showed that the sedimentcharacteristics between the 3 sampling years were significantlydifferent (Wilk’s k = 0.26, p < 0.001). SCF, MC, TOC, TKN and TPwere subsequently compared for each separate station betweenyears 0, 1 and 2. SCF and MC did not significantly change at anyof the 6 sampling stations at p = 0.05 level. TOC levels did not dem-onstrate marked differences at the reference stations (KS1 andKS6) and intermediary stations (KS2 and KS5) between pre- andpost-deployment (p > 0.05). At the fish cage stations KS3 andKS4, TOC levels showed some trend of decrease after AR deploy-ment; the results, however, were not statistically significant atp = 0.05 level. For TKN, the concentrations at KS1, KS2 and KS6did not show significant changes (p > 0.05) while at KS3, KS4 andKS5, i.e., the 3 stations inside or near the FCZ where ARs were de-ployed, a decrease in TKN concentrations after AR deployment rel-ative to that before AR deployment was statistically significant (forKS3, year 0 vs. year 1, t4 = 5.9, p < 0.05, year 0 vs. year 2, t4 = 3.8,p < 0.05, year 1 vs. year 2, t4 = 0.57, p > 0.05; for KS4, year 0 vs. year1, t4 = 4.5, p < 0.05, year 0 vs. year 2, t4 = 13.1, p < 0.05, year 1 vs.year 2, t4 = 4.3, p < 0.05; for KS5, year 0 vs. year 1, t4 = 5.4,p < 0.05, year 0 vs. year 2, t4 = 3.2, p < 0.05, year 1 vs. year 2, t4

= �1.2, p > 0.05). A similar trend was found for TP. At the referenceand intermediary stations, TP levels did not show significantchanges after AR deployment (p > 0.05). At KS4, the reduction inTP concentration during post-AR deployment period comparedwith that before AR deployment was significant (for year 0 vs. year1, t4 = 4.3, p < 0.05; for year 0 vs. year 2, t4 = 5.6, p < 0.05; year 1 vs.year 2, t4 = 3.0, p < 0.05). TP levels at KS3 also showed a decreasingtrend after AR deployment, but the reduction was not statisticallysignificant at p = 0.05 level.

Station

KS1 KS2 KS3 KS4 KS5 KS6

KS1 KS2 KS3 KS4 KS5 KS6

KS1 KS2 KS3 KS4 KS5 KS6

KS1 KS2 KS3 KS4 KS5 KS6

SC

F (

%)

0

20

40

60

80

100

120Year 0Year 1Year 2

Station

MC

(%

)

0

20

40

60

80

100

Year 0Year 1Year 2

Station

TO

C (

mg

g-1

)

0

10

20

30

40

50

60

70

Year 0Year 1Year 2

Station

KS1 KS2 KS3 KS4 KS5 KS6

TK

N (

mg

g-1

)

0

1

2

3

4

5

6

Year 0Year 1Year 2

a

b b

a

b

c a

bb

Station

TP

(m

g g

-1)

0

2

4

6

8

10

12

14

Year 0Year 1Year 2

b

c

a

a

c d

e

b

Fig. 3. Comparisons on sediment characteristics of six sampling stations between pre-AR deployment (year 0), and 1st year (year 1) and 2nd year (year 2) of post-ARdeployment. (a) SCF = silt-clay fraction (%); (b) MC = moisture content (%); (c) TOC = total organic carbon (mg g�1); (d) TKN = total Kjeldahl nitrogen (mg g�1); (e) TP = totalphosphorus (mg g�1). The columns labelled with different letters mean that the values are significantly different between years 0, 1 and 2 at the significance level of p < 0.05.Non-labelled columns mean no between-year difference at the significance level of p < 0.05. Values are presented as annual means of the 5 monitoring months, i.e., August,October, December, February and April, and the error bars represent 1SD.

Q.-F. Gao et al. / Marine Pollution Bulletin 57 (2008) 652–661 655

3.2. Benthic communities

Dominant species before AR deployment were presented at Gaoet al. (2005). After AR deployment, a total of 3223 individuals

belonging to 101 species were collected and identified. The 5 mostdominant species at each station for each sampling month arelisted in Table 1. Dominant species at the reference, intermediaryand fish cage stations were found to vary over the study period

Table 1The top five dominant species collected at each station for each sampling month after AR deployment (number = % of total individual recorded)

Month Reference stations Intermediary stations Fish cage stations

KS1 KS6 KS2 KS5 KS3 KS4

August 2002 Aglaophamus lyrochaeta 50.00 Apionsoma trichocephalus 26.67 Theora lata 33.33 Aglaophamus lyrochaeta 40.00 Sigambra tentaculata 100 Sigambra tentaculata 88.89Amphioplus laevis 18.75 Lumbrineris shiinoi 20.00 Aglaophamus lyrochaeta 22.22 Orbinia sp. 13.33 Apionsoma trichocephalus 11.11Gobiidae sp. 6.25 Amphioplus laevis 13.33 Sigambra tentaculata 22.22 Tellina cygnus 13.33Notomastus latericeus 6.25 Byblis sp. 6.67 Ophelina acuminata 11.11 Alpheus brevicristatus 6.67Orbinia sp. 6.25 Anomalifrons lightana 6.67 Thalamita sima 11.11 Glycera decipiens 6.67

October 2002 Aglaophamus lyrochaeta 21.86 Tellina cygnus 12.12 Aglaophamus lyrochaeta 20.00 Aglaophamus lyrochaeta 19.35 Sigambra tentaculata 100 Sigambra tentaculata 100Ophelina acuminata 21.86 Leptochela aculeocaudata 9.09 Sigambra tentaculata 20.00 Sigambra tentaculata 19.35Amphioplus laevis 12.50 Amphioplus laevis 9.09 Leptochela aculeocaudata 11.43 Theora lata 9.68Leptochela aculeocaudata 9.38 Sigambra tentaculata 9.09 Lumbrineris shiinoi 8.57 Ceratoplax sp. 6.45Sigambra tentaculata 9.38 Alpheus brevicristatus 6.06 Orbinia sp. 8.57 Glycera decipiens 6.45

December 2002 Aglaophamus lyrochaeta 13.33 Nemertinea sp. A 17.65 Theora lata 37.50 Sigambra tentaculata 33.33 Nil species Sigambra tentaculata 75.00Schizaster lacunosus 13.33 Aglaophamus lyrochaeta 11.76 Sigambra tentaculata 25.00 Aglaophamus lyrochaeta 20.00 Aglaophamus lyrochaeta 25.00Sigambra tentaculata 13.33 Ceratoplax sp. 11.76 Aglaophamus lyrochaeta 8.33 Glycera decipiens 20.00Ceratoplax sp. 6.67 Amphioplus laevis 11.76 Nemertinea sp. A 8.33 Glycinde sp. 6.67Eunice indica 6.67 Sigambra tentaculata 11.76 Apionsoma trichocephalus 4.17 Leptochela aculeocaudata 6.67

February 2003 Echinoidea sp. 33.96 Amphioplus laevis 22.92 Aglaophamus lyrochaeta 17.14 Echinoidea sp. 24.64 Theora lata 46.15 Theora lata 39.39Philine sp. 9.43 Echinoidea sp. 12.50 Theora lata 11.43 Aglaophamus lyrochaeta 15.94 Sigambra tentaculata 41.48 Sigambra tentaculata 30.30Amphioplus laevis 9.43 Nemertinea sp. A 8.33 Prionospio malmgreni 8.57 Theora lata 13.04 Zeuxis sp. 3.02 Ophelina acuminata 9.09Tellina cygnus 7.55 Aglaophamus lyrochaeta 6.25 Echinoidea sp. 8.57 Tellina cygnus 10.14 Eurythoe sp. 1.65 Aglaophamus lyrochaeta 4.55Aglaophamus lyrochaeta 5.66 Prionospio malmgreni 6.25 Harmothoe sp. 5.71 Glycinde sp. 2.90 Orbinia sp. 1.10 Echinoidea sp. 4.55

April 2003 Echinoidea sp. 29.79 Aglaophamus lyrochaeta 16.13 Theora lata 27.03 Sigambra tentaculata 26.67 Theora lata 58.39 Sigambra tentaculata 40.54Amphioplus laevis 10.64 Anisocorbula cuneata 16.13 Sigambra tentaculata 21.62 Theora lata 21.67 Sigambra tentaculata 24.16 Theora lata 40.54Sternaspis scutata 8.51 Cylichna sp. 12.90 Eurythoe sp. 10.81 Aglaophamus lyrochaeta 10.00 Echinoidea sp. 10.07 Eurythoe sp. 8.11Cylichna sp. 6.38 Nemertinea sp. A 12.90 Tellinides chinensis 8.11 Glycera decipiens 8.33 Lovenia subcarinata 2.01 Nemertinea sp. A 2.70Theora lata 6.38 Amphioplus laevis 9.68 Nemertinea sp. A 5.41 Tellinides chinensis 5.00 Tellina cygnus 2.01 Ophiocnemis marmorata 1.35

June 2003 Amphioplus laevis 35.00 Aglaophamus lyrochaeta 50.00 Sigambra tentaculata 27.78 Sigambra tentaculata 28.57 Sigambra tentaculata 43.78 Sigambra tentaculata 33.33Aglaophamus lyrochaeta 20.00 Ceratoplax sp. 10.00 Aglaophamus lyrochaeta 22.22 Eurythoe sp. 21.43 Eurythoe sp. 39.13 Aglaophamus lyrochaeta 20.00Sigambra tentaculata 20.00 Nemertinea sp. A 10.00 Nemertinea sp. A 11.11 Aglaophamus lyrochaeta 14.29 Theora lata 8.70 Tellina cygnus 13.33Amphipoda sp. A 15.00 Alpheus brevicristatus 5.00 Theora lata 11.11 Ceratoplax sp. 7.14 Hexapus sp. 4.35 Theora lata 13.33Alpheus brevicristatus 5.00 Sternaspis scutata 5.00 Glycera decipiens 5.56 Glycera decipiens 7.14 Scacella sp. 4.35 Glycera decipiens 6.67

August 2003 Aglaophamus lyrochaeta 25.00 Aglaophamus lyrochaeta 20.00 Aglaophamus lyrochaeta 26.92 Aglaophamus lyrochaeta 37.50 Sigambra tentaculata 92.86 Sigambra tentaculata 80.00Tellinides chinensis 13.89 Tellinides chinensis 20.00 Tellinides chinensis 26.92 Eurythoe sp. 18.75 Tellinides chinensis 7.14 Eurythoe sp. 10.00Alpheus brevicristatus 8.33 Lumbrineris shiinoi 20.00 Eurythoe sp. 19.23 Sigambra tentaculata 12.50 Aglaophamus lyrochaeta 5.00Cylichna sp. 8.33 Alpheus brevicristatus 10.00 Amphioplus laevis 7.69 Tellina cygnus 12.50 Tellinides chinensis 5.00Sternaspis scutata 5.56 Ceratoplax sp. 10.00 Sigambra tentaculata 7.69 Tellinides chinensis 6.25

October 2003 Aglaophamus lyrochaeta 13.33 Processidae sp. 22.73 Nemertinea sp. A 22.22 Sigambra tentaculata 31.82 Sigambra tentaculata 53.45 Sigambra tentaculata 71.43Nemertinea sp. A 13.33 Aglaophamus lyrochaeta 18.18 Sigambra tentaculata 22.22 Prionospio malmgreni 24.24 Minuspio cirrifera 27.59 Nemertinea sp. A 14.29Prionospio malmgreni 13.33 Nemertinea sp. A 18.18 Eurythoe sp. 22.22 Tharyx sp. 10.61 Aglaophamus lyrochaeta 6.90 Notomastus latericeus 14.29Alpheus brevicristatus 6.67 Amphioplus laevis 9.09 Alpheus brevicristatus 11.11 Eteone sp. 7.58 Crustacea sp. 3.45Cossurella dimorpha 6.67 Prionospio malmgreni 9.09 Notomastus latericeus 11.11 Notomastus latericeus 4.55 Alpheus brevicristatus 1.72

656Q

.-F.Gao

etal./M

arinePollution

Bulletin57

(2008)652–

661

Dec

embe

r20

03Ph

oron

issp

.32.

29A

mph

iopl

usla

evis

13.5

1Ph

oron

issp

.31.

34Si

gam

bra

tent

acul

ata

17.8

6G

lyci

nde

sp.2

0.00

Prio

nosp

iom

alm

gren

i32

.56

Thal

asse

ma

sabi

num

31.2

5A

glao

pham

usly

roch

aeta

10.8

1Eu

ryth

oesp

.8.9

6A

glao

pham

usly

roch

aeta

14.2

9M

inus

pio

cirr

ifer

a20

.00

Theo

rala

ta9.

30A

mph

iopl

usla

evis

4.17

Love

nia

subc

arin

ata

10.8

1G

lyce

rade

cipi

ens

5.97

Ogy

ride

sor

ient

alis

7.14

Siga

mbr

ate

ntac

ulat

a13

.33

Nem

erti

nea

sp.A

6.98

Agl

aoph

amus

lyro

chae

ta3.

13N

emer

tine

asp

.A8.

11Lo

veni

asu

bcar

inat

a5.

97Pr

iono

spio

mal

mgr

eni

7.14

Sthe

nole

pis

yhle

ni13

.33

Siga

mbr

ate

ntac

ulat

a6.

98Lo

veni

asu

bcar

inat

a3.

13Si

gam

bra

tent

acul

ata

8.11

Not

omas

tus

late

rice

us5.

97Eu

ryth

oesp

.7.1

4Ph

iline

sp.6

.67

Capr

ella

sp.6

.98

Febr

uar

y20

04Th

alas

sem

asa

binu

m37

.66

Thal

asse

ma

sabi

num

65.8

4Eu

ryth

oesp

.37.

14Sc

hiza

ster

lacu

nosu

s33

.93

Min

uspi

oci

rrif

era

43.6

1Si

gam

bra

tent

acul

ata

29.7

5Lo

veni

asu

bcar

inat

a11

.69

Prio

nosp

iom

alm

gren

i6.

93N

emer

tine

asp

.A8.

57G

lyce

rade

cipi

ens

10.7

1Si

gam

bra

tent

acul

ata

27.8

2M

inus

pio

cirr

ifer

a20

.25

Prio

nosp

iom

alm

gren

i6.

49Si

gam

bra

tent

acul

ata

4.95

Siga

mbr

ate

ntac

ulat

a8.

57Si

gam

bra

tent

acul

ata

10.7

1Pr

iono

spio

mal

mgr

eni

18.8

0N

otom

astu

sla

teri

ceus

11.3

9A

glao

pham

usly

roch

aeta

5.19

Agl

aoph

amus

lyro

chae

ta3.

96Lo

veni

asu

bcar

inat

a5.

71A

glao

pham

usly

roch

aeta

8.93

Agl

aoph

amus

lyro

chae

ta2.

26Eu

ryth

oesp

.10.

13N

emer

tine

asp

.A5.

19Eu

ryth

oesp

.2.9

7A

mph

iopl

usla

evis

5.71

Thar

yxsp

.8.9

3St

heno

lepi

syh

leni

2.26

Agl

aoph

amus

lyro

chae

ta7.

59

Apr

il20

04Th

alas

sem

asa

binu

m52

.55

Thal

asse

ma

sabi

num

57.8

6Sc

hiza

ster

lacu

nosu

s18

.84

Schi

zast

erla

cuno

sus

68.1

8Si

gam

bra

tent

acul

ata

50.0

0N

otom

astu

sla

teri

ceus

33.3

3Pr

iono

spio

mal

mgr

eni

12.4

1Sc

hiza

ster

lacu

nosu

s7.

86Eu

ryth

oesp

.17.

39Lu

mbr

iner

issh

iinoi

7.58

Echi

noid

easp

.23.

53Si

gam

bra

tent

acul

ata

25.6

4A

glao

pham

usly

roch

aeta

5.11

Gly

cera

deci

pien

s4.

29Lo

veni

asu

bcar

inat

a13

.04

Prio

nosp

iom

alm

gren

i7.

58Th

eora

lata

11.7

6M

inus

pio

cirr

ifer

a12

.82

Har

mot

hoe

sp.2

.92

Nem

erti

nea

sp.A

4.29

Thal

asse

ma

sabi

num

8.70

Not

omas

tus

late

rice

us6.

06G

lyce

rade

cipi

ens

2.94

Prio

nosp

iom

alm

gren

i10

.26

Not

omas

tus

late

rice

us2.

92Lo

veni

asu

bcar

inat

a3.

57Pr

iono

spio

mal

mgr

eni

5.80

Siga

mbr

ate

ntac

ulat

a3.

03Lo

veni

asu

bcar

inat

a2.

94G

lyce

rade

cipi

ens

5.13

Q.-F. Gao et al. / Marine Pollution Bulletin 57 (2008) 652–661 657

after the AR deployment. At the reference stations, the polychaeteAglaophamus lyrochaeta was numerically abundant for most of thesampling months. However, in the later several months (Februaryand April, 2004), its top abundance was replaced by the EchiuranThalassema sabinum. At the intermediary stations, dominant spe-cies represented both those occurring at the reference and fish cagestations. For example, at these stations, the polychaete A. lyrocha-eta was still ranked within the top 5 dominant species. Other abun-dant species included the polychaete Sigambra tentaculata, whichwas also registered in high numbers at the fish cage stations. Again,in the later several months (February and April, 2004), the seaurchin Schizaster lacunosus became more numerically abundantthan the above polychaete species. At the fish cage stations, theretended to have less species dominant in the first few samplingmonths (August–December, 2002) after the deployment of theARs. The dominant species was the polychaete S. tentaculata. Start-ing from the sampling month of February 2003 onwards, a morediverse benthic infauna appeared to emerge at these stations, withmore species sharing the top 5 dominant positions (Table 1). Thesespecies included the bivalves Theora lata, Tellina cygnus, Scacella sp.and Tellinides chinensis, the gastropod Zeuxis sp., the polychaetes A.lyrochaeta, Eurythoe sp., Ophelina acuminata, Glycera decipiens,Notomastus latericeus, Minuspio cirrifera, Glycinde sp., Prionospiomalmgreni and Sthenolepis yhleni, the crab Hexapus sp., and thesea urchins Lovenia subcarinata and Echinoidea sp.

As shown in Fig. 4, Shannon-Wiener’s diversity indices (H0) ofboth the reference (KS1 and KS6) and intermediary (KS2 andKS5) stations were relatively higher than those at the fish cage sta-tions (KS3 and KS4). H0 showed a range between 1.80 and 4.01 overthe sampling months. Except for one sampling month (April 2004at KS5) with relatively lower H0 of 1.80, other sampling data re-vealed H0 in the range of >2 to >3. At the fish cage stations, diversityindex showed an increase towards the end of the sampling period.In particular, H0 increased from 0–1 in the first few months (Au-gust–December 2002) after AR deployment to H0 > 2 at the end ofthe study (December 2003–April 2004).

Changes in benthic communities after AR deployment for thefish cage, intermediary and reference stations are shown in theMDS ordination plots according to the sampling year (Fig. 5). Re-sults of ANOSIM indicated that benthic communities of 1st year(year 1) and 2nd year (year 2) after AR deployment were differentfrom those before AR deployment (year 0) (R = 0.46, p < 0.05). TheR and p values of ANOSIM for subsequent pairwise comparisonsbetween the communities of years 0, 1 and 2 for reference, inter-mediary and fish cage stations are listed in Table 2. At the refer-ence and intermediary stations, results of ANOSIM showed thatthe community structure was significantly different (p < 0.05)among the 3 years. At the fish cage stations, results of ANOSIMshowed that significant changes (p < 0.05) in the benthic structurewere noted between pre-AR deployment (year 0) and post-ARdeployment (years 1 and 2), but not between years 1 and 2.

4. Discussion

Filter-feeding organisms on ARs consume phytoplankton andother suspended particulate matter from the water column. Sus-pended particulates removed from the ambient waters are partlyrejected as pseudofaeces, or if ingested but not absorbed, are even-tually eliminated as faeces (collectively termed biodeposits). Inturn, matter absorbed from the ingested food can be lost as excre-tion, releasing ammonium and orthophosphate as products ofmetabolism. Thus, filter feeders function as material processorsin aquatic ecosystems and transfer nutrients from the water col-umn to sediment and biomass of filter feeders themselves. In thisway, filter-feeding organisms are a direct trophic link betweenthe water column and benthos. When filter feeders occur in

Sampling Month

Aug01

Oct01

Dec01

Feb02

Apr02

Aug02

Oct02

Dec02

Feb03

Apr03

Jun03

Aug03

Oct03

Dec03

Feb04

Apr04

Sh

ann

on

-Wie

ner

's In

dex

(H

')

0

1

2

3

4

5

KS1KS2KS3KS4KS5KS6

AR deployment

Fig. 4. Shannon-Wiener’s diversity index of the reference (KS1 and KS6), intermediary (KS2 and KS5) and fish cage (KS3 and KS4) stations over the sampling period of pre-ARdeployment from August 2001 to April 2002 and post-AR deployment from August 2002 to April 2004.

Fig. 5. MDS ordination plots showing changes in benthic communities between pre- and post-AR deployment for (a) reference stations (KS1 and KS6); (b) intermediarystations (KS2 and KS5); (c) fish cage stations (KS3 and KS4). Each square point labelled with 0, 1 and 2 in the plots represents the community structure of each monitoringmonth, i.e., August, October, December, February and April, before AR deployment (year 0), the first (year 1) and second (year 2) year after AR deployment, respectively.

658 Q.-F. Gao et al. / Marine Pollution Bulletin 57 (2008) 652–661

abundance, they may structurally and functionally influence theirecosystems.

Early estimates of the effects of filter feeders on sedimentationin marine ecosystems have been reported by Verwey (1952). Ver-

Table 2R and p values of analysis of similarity (ANOSIM) for pairwise comparisons betweenbenthic communities of years 0, 1 and 2 for reference, intermediary and fish cagestations

Station Comparison group R-value p-Value

Reference stations year 0–year 1 0.82 <0.05year 0–year 2 0.87 <0.05year 1–year 2 0.18 <0.05

Intermediary stations year 0–year 1 0.42 <0.05year 0–year 2 0.66 <0.05year 1–year 2 0.32 <0.05

Fish cage stations year 0–year 1 0.34 <0.05year 0–year 2 0.51 <0.05year 1–year 2 0.10 0.09

Q.-F. Gao et al. / Marine Pollution Bulletin 57 (2008) 652–661 659

wey speculated that biodeposition due to bivalve filtration was amajor factor determining the flux of materials into seabed sedi-ments and biodeposition by filter feeders changed the characteris-tics of the sediments. Rodhouse and Roden (1987) computed thecarbon budget of blue mussels Mytilus edulis and concluded thatmussels would extensively crop both the resident and tidal inputof offshore suspended particulate matter, leading to more sedi-mentation under the mussels and subsequent changes in thebenthos.

Generally, during the feeding and absorption processes, en-larged size and density of compact biodeposits enhance the set-tling of particles in the area adjacent to the assemblages of filterfeeders. The chemical constituents of biodeposits will be changeddue to such biological reworking. As a result, the sedimentatedsubstances are quantitatively and qualitatively modified (Grafand Rosenberg, 1997). Navarro and Thompson (1997) determinedthe concentration of organic carbon and organic nitrogen in thefaeces and pseudofaeces of the horse mussels Modiolus modiolusand found that nutrient content in biodeposits was reduced com-pared with that in natural seston which was taken up by the mus-sels. Other studies at the present experimental site also showedthat mussels (Gao et al., 2006, 2008) and other filter-feeding organ-isms (Gao, 2005) were able to efficiently remove particulate organ-ic matter from the water column where the ARs were present. As aresult of nutrient removal due to biological reworking by filterfeeders in the AR systems, it appears that the seabed environmenthas shown some degree of reduction in nutrient levels at the sam-pling stations in the vicinity of AR deployment although the extentof reduction for each nutrient was different. Among the 3 nutrientsmonitored in the present study, nitrogen is the most motile ele-ment because particulate organic nitrogen which is filtered by fil-ter feeders from the water column can be mineralized toammonium, which may undergo nitrification, with oxidation ofammonium to nitrate. Denitrification of nitrate may also occur,especially under the hypoxic, reducing conditions in the summerseason at the study site. Reduction of nitrate in the process of deni-trification releases nitrogen in various gaseous forms such as NO,NO2 and eventual N2. All such gaseous substances are more readilyemitted away (Kaspar et al., 1985). Accordingly, a marked decreasein nitrogen levels at stations KS3, KS4 and KS5 occurred after ARdeployment relative to that before AR deployment (Fig. 3d). In con-trast, phosphate, as the metabolic product of aquatic organisms,tends to be combined with calcium (Ca2+), magnesium (Mg2+)and iron (Fe2+) ions, which are abundant in marine environments,rendering it insoluble and to be accumulated in seabed sediments(Holby and Hall, 1991). As a result, pronounced reduction in phos-phorus loadings during post-AR deployment period only occurredat KS4 (Fig. 3e). TOC levels at the 2 fish cage stations KS3 andKS4 also showed some trend of decrease after AR deployment.However, since carbon is the most abundant element in sediments,the proportion of organic carbon removed by filter feeders is rela-

tively small and the reduction is not statistically significant(Fig. 3c).

In the fish culture zone, it is of concern that the increase indeposition of farming wastes to the bottom leads to accumula-tion of organic matter under fish cages. This may create adverseenvironmental effects such as anoxic conditions, production oftoxic gases, etc. (Chua, 1993). Lara and Moreno (1995) reportedthat filter feeders help prevent the accumulation of particulatematter that occurs under the fish cages since they take up thesuspended particles in the water column before their settlementon the seabed. In freshwater salmon farms, Soto and Mena(1999) also demonstrated that high densities of mussels Diplodonchilensis were able to efficiently mitigate salmon farming impactsthrough their filtration function, resulting in changes in salmonfarms from a hypereutrophic situation to an oligotrophic one.ARs also provide additional habitats and shelter for pelagicfishes. The aggregates of the carnivorous fishes might functionas consumers taking up sessile organisms on the ARs and utiliz-ing the farming wastes directly (Relini et al., 2002). As a result,the farming wastes were converted into the biomass of pelagicichthyofauna, and hence reduce the impacts of farming wasteson the environment.

With respect to benthic communities, both the reference andintermediary stations had relatively higher species diversity thanthe fish cage stations before and after the AR deployment (Fig. 4).This might be attributed to the adverse impact of nutrient enrich-ment beneath and in the vicinity of fish farming waters on thebenthos (Gao et al., 2005). At reference stations KS1 and KS6where the sediment quality was not affected by fish farming, spe-cies diversity, H0, was recorded within a range between 2.18 and4.01 over the sampling months, indicating that the benthic infau-na was diverse similar to other soft bottom benthic communitiesin Hong Kong (Shin et al., 2004). The fish cage stations tended tohave less species dominance in the first few sampling monthsafter the deployment of the ARs. The hypoxia-tolerant polychaeteS. tentaculata was the most dominant species at fish cage stations,probably due to the low dissolved oxygen caused by organicenrichment (Nichols-Driscoll, 1976). Results of ANOSIM showedthat community composition at the reference and intermediarystations was different among the 3 years throughout the monitor-ing period (Table 2), indicating that inter-annual variations oc-curred as reflected in the benthic assemblages. As for the fishcage stations KS3 and KS4, significant changes (p < 0.05) in ben-thic structure were noted between year 0 (pre-AR deployment)and years 1 and 2 (post-AR deployment), but not between years1 and 2, suggesting that benthic structure before and after ARdeployment was different. The insignificant difference in commu-nity structure between years 1 and 2 (post-AR deployment) at thefish cage stations also suggested that the benthic communitymight become more stable with the presence of ARs in responseto fluctuations in natural conditions, which were revealed by thechanges in benthic structure between years 1 and 2 at referenceand intermediary stations. In addition, species diversity H0 atthe fish cage stations showed an increasing trend during thepost-AR deployment period, from 0 to 1 in the first few months(August–December 2002) after AR deployment to H0 > 2 at theend of the study (December 2003–April 2004) (Fig. 4), indicatingan enhancement in the diversity of benthic communities after ARdeployment at the fish cage stations where AR were present.However, H0 values of reference and intermediary stations keptrelatively constant and did not show marked changes over thepost-AR deployment period (Fig. 4).

Natural and artificial rocky reefs provide habitats for diverseassemblages of aquatic organisms. Numerous studies have focusedon the colonization and succession of fish and epibenthic animalsin such systems (Bohnsack and Sutherland, 1985; Parrish, 1989).

660 Q.-F. Gao et al. / Marine Pollution Bulletin 57 (2008) 652–661

Previous studies have shown that the presence of hard reef con-structions potentially alter the composition and abundance of ben-thic assemblages in neighbouring sandy bottoms (Davis et al.,1982; Ambrose and Anderson, 1990; Posey and Ambrose, 1994;Barros et al., 2001). These studies suggested that the modificationof adjacent benthic communities associated with reef systemsmight be attributed to the changes in the hydrographic conditions,sedimentation of suspended particulate matter, characteristics ofsediments and predation stresses. The present study also revealedthat the deployment of artificial reefs led to improved speciesdiversity in surrounding infaunal communities.

The temporal pattern of increasing species diversity at the fishcage stations KS3 and KS4 after AR deployment is related to acorresponding reduction in nutrient levels at these stations. Inaddition, biodeposition from mussels and other filter feeders onthe ARs has the potential to modify the bottom micro-habitatsby changing surface roughness and by construction of mounds,pits, and settlement of tubes, resulting in elevated heterogeneityof seabed habitats and subsequent enhanced capability of accom-modating more morphologically, physiologically and behaviour-ally diverse organisms (Rowden and Jones, 1994). The settledbiodeposits are more edible and digestible to the benthic depositfeeders in the soft bottom as such biodeposits have been digestedby the filter feeders, stimulating the production of resident spe-cies and attracting the colonization of immigrants (Graf andRosenberg, 1997). It has been acknowledged that deployment ofARs might also have an important function to slow down currentspeed and change current direction, improving ecological stabilityand providing sanctuary for the residents, including sessileorganisms on the ARs and infaunal benthos in the bottom habi-tats within the AR area (Turner et al., 1969; Davis et al., 1982).Such protection functions of ARs, together with the increasedheterogeneity of seabed habitats, offer ideal refuge for manyorganisms against natural disastrous events, especially the mon-soon-driven waves which frequently occur in tropical and sub-tropical regions.

Fabi et al. (2002) reported that the Shannon-Wiener diversityindex of benthic community was relatively higher at the artificialreefs than at a control site owing to changes in physical factorssuch as water current and sediment characteristics. Similarly, Bar-ros et al. (2001) investigated the influence of reef habitats in Aus-tralia on structure of benthic macrofauna in nearby soft-sedimentsalong transects from the stations far away from rocky reefs tothose near to the reefs. The results demonstrated that infaunal spe-cies richness at the stations close to rocky reefs was higher com-pared with those far from the reefs. All these studies and thepresent investigation suggest that the presence of man-made ornatural reef structures can exert direct or indirect effects on softbottom conditions, leading to modifications in benthic assem-blages in adjacent areas.

Acknowledgements

The work described in this paper was fully supported by a Stra-tegic Research Grant (Project No. 7002069) from City University ofHong Kong. Assistance in artificial reef deployment by the Agricul-ture, Fisheries and Conservation Department, Hong Kong SAR Gov-ernment is greatly appreciated.

References

Ambrose, R.F., Anderson, T.W., 1990. Influence of artificial reef on the surroundinginfaunal community. Marine Biology 107, 41–42.

Angel, D.L., Eden, N., Breitstein, S., Yurman, A., Katz, T., Spanier, E., 2002. In situbiofiltration: a means to limit the dispersal of effluents from marine finfish cageaquaculture. Hydrobiologia 469, 1–10.

Barros, F., Underwood, A.J., Lindegarth, M., 2001. The influence of rocky reefs onstructure of benthic macrofauna in nearby soft-sediments. Estuarine, Coastaland Shelf Science 52, 191–199.

Bohnsack, J.A., Sutherland, D.L., 1985. Artificial reef research: a review withrecommendations for future priorities. Bulletin of Marine Science 37, 11–39.

Bray, J.R., Curtis, J.T., 1957. An ordination of upland forest community of SouthernWisconsin. Ecological Monographs 27, 325–349.

Burd, B.J., Nemec, A., Brinkhurst, R.O., 1990. The development and application ofanalytical methods in benthic marine infaunal studies. Advances in MarineBiology 26, 169–247.

Chua, T.E., 1993. Environmental management of coastal aquaculture development.In: Pullin, P.S.V., Rosenthal, H. (Eds.) Environment and Aquaculture inDeveloping Countries. Manila, Philippines, International Center for LivingAquatic Resources Management, Frankfurt/Main, Federal Republic ofGermany, Deutsche Gesellschaft fur Technische Zusammenarbeit (GTZ)GmbH, pp. 199–212.

Clarke, K.R., Warwick, R.M., 2001. Changes in Marine Communities: An Approach toStatistical Analysis and Interpretation (2nd ed.). PRIMER-E Ltd., Plymouth, UK.

Davis, N., Vanblaricom, G.R., Dayton, P.K., 1982. Man-made structures on marinesediments: effects on adjacent benthic communities. Marine Biology 70, 295–303.

Fabi, G., Luccarini, F., Panfili, M., Solustri, C., Spagnolo, A., 2002. Effects of artificialreef on the surrounding soft-bottom community (central Adriatic Sea). ICESJournal of Marine Science 59, S343–S349.

Folke, C., Kautsky, N., 1989. The role of ecosystems for a sustainable development ofaquaculture. Ambio 18, 234–243.

Gao, Q.F., 2005. Use of artificial reefs and green-lipped mussels (Perna viridis) forremoval of nutrients from marine fish farming. PhD Thesis, City University ofHong Kong, Hong Kong.

Gao, Q.F., Wang, Z.L., Wong, W.H., Cheung, S.G., 2002. Effects of food quality andquantity on feeding and absorption in black-ribbed mussels, Septifer virgatus(Wiegmann) (Bivalvia: Mytilidae) dominating wave-exposed habitats in HongKong. Journal of Shellfish Research 21, 51–57.

Gao, Q.F., Cheung, K.K., Cheung, S.G., Shin, P.K.S., 2005. Effects of nutrientenrichment derived from fish farming activities on macroinvertebrateassemblages in a subtropical region of Hong Kong. Marine Pollution Bulletin51, 994–1002.

Gao, Q.F., Shin, P.K.S., Lin, G.H., Chen, S.P., Cheung, S.G., 2006. Stable isotopic andfatty acid evidence for uptake of organic wastes from fish farming throughgreen-lipped mussels (Perna viridis) in a polyculture system. Marine EcologyProgress Series 317, 273–283.

Gao, Q.F., Xu, W.Z., Liu, X.S., Cheung, S.G., Shin, P.K.S., 2008. Seasonal changes in C, Nand P budgets of green-lipped mussels Perna viridis and oval of nutrients fromfish farming in Hong Kong. Marine Ecology Progress Series 353, 137–146.

Graf, G., Rosenberg, R., 1997. Bioresuspension and biodeposition: a review. Journalof Marine Systems 11, 269–278.

Holby, O., Hall, P.O.J., 1991. Chemical fluxes and mass balances in a marine fish cagefarm II, phosphorus. Marine Ecology Progress Series 70, 263–272.

Jordan, T.E., Valiela, E., 1982. A nitrogen budget of the ribbed mussel, Geukensiademissa, and its significance in nitrogen flow in a New England salt marsh.Limnology and Oceanography 27, 75–90.

Jørgensen, C.B., 1990. Measuring pumping rates. In: Bivalve Filter Feeding:Hydrodynamics, Bioenergetics, Physiology and Ecology, Olsen and Olsen,Helstedsvej, Denmark, pp. 35–40.

Kaspar, H.F., Gillespie, P.A., Boyer, I.C., Mackenzie, A.L., 1985. Effects of musselaquaculture on the nitrogen cycle and benthic communities of Kenepuru Sound,Marlborough Sounds, New Zealand. Marine Biology 85, 127–136.

Lara, G.P., Moreno, C.A., 1995. Efectos de la depredación de Aegla abtao (Crustacea,Aeglidae) sobre la distribución espacial y abundancia de Diplodon chilensis(Bivalvia Hyriidae) en el Lago Panguipulli, Chile. Revista Chilena de HistoriaNatural 68, 123–129.

Mudroch, A., Azcue, J.M., Mudroch, P., 1997. Manual of Physico-Chemical Analysis ofAquatic Sediments. CRC Press Inc., Boca Raton.

Navarro, J.M., Thompson, R.J., 1997. Biodeposition by the horse mussel Modiolusmodiolus (Dillwyn) during the spring diatom bloom. Journal of ExperimentalMarine Biology and Ecology 209, 1–13.

Nichols-Driscoll, J., 1976. Benthic invertebrate communities in Golfo Dulce, CoastaRica, an anoxic basin. Revista de Biologia Tropical 24, 281–297.

Parrish, J.D., 1989. Fish communities of interacting shallow-water habitats intropical oceanic regions. Marine Ecology Progress Series 58, 143–160.

Pearson, T.H., Black, K.D., 2001. The environmental impacts of marine fish cageculture. In: Black, K.D. (Ed.), Environmental Impacts of Aquaculture. SheffieldAcademic Press, Sheffeild, UK, pp. 1–31.

Posey, M.H., Ambrose Jr., W.G., 1994. Effects of proximity to an off-shore hard-bottom reef on infaunal abundances. Marine Biology 118, 745–753.

Relini, G., Relini, M., Torchia, G., De Angelis, G., 2002. Trophic relationships betweenfishes and an artificial reef. ICES Journal of Marine Science 59, S36–S42.

Reynoldson, T.B., Day, K.E., 1993. Freshwater sediments. In: Calow, P. (Ed.),Handbook of Ecotoxicology, vol. 1. Blackwell Science Publication, Oxford, pp.83–100.

Rodhouse, P.G., Roden, C.M., 1987. Carbon budget for a coastal inlet in relation tointensive cultivation of suspension-feeding bivalve mollusks. Marine EcologyProgress Series 36, 225–236.

Rowden, A.A., Jones, M.B., 1994. A contribution to the biology of the burrowing mudshrimp, Callianassa subterranea (Decapoda: Thalassinidea). Journal of theMarine Biological Association of the United Kingdom 74, 623–635.

Q.-F. Gao et al. / Marine Pollution Bulletin 57 (2008) 652–661 661

Shannon, C.E., Weaver, W., 1963. The Mathematical Theory of Communication.University of Illinois Press, Urbana, IL.

Shin, P.K.S., Huang, Z.G., Wu, R.S.S., 2004. An updated baseline of macrobenthicinfauna in subtropical waters of Hong Kong. Marine Pollution Bulletin 49, 128–135.

Soto, D., Mena, G., 1999. Filter feeding by the freshwater mussel, Diplodon chilensis,as a control of salmon farming eutrophication. Aquaculture 171, 65–81.

Turner, C.H., Ebert, E.E., Given, R.R., 1969. Man-made reef ecology. CaliforniaDepartment Fish Game Bulletin 146, 1–221.

Verwey, J., 1952. On the ecology and distribution of cockles and mussels on theDutch Wadden Sea, their role in sedimentation and the source of their foodsupply. Netherlands Journal of Zoology 10, 172–239.

Wu, R.S.S., 1995. The environmental impact of marine fish impacts: towards asustainable future. Marine Pollution Bulletin 31, 159–166.

Zar, J.H., 1999. Biostatistical Analysis, fourth ed. Prentice-Hall, New Jersey,USA.