Effects of Hydraulic Dredging for Mercenaria mercenaria , Northern Quahog, on Sediment Biogeochemistry

JOURNAL OF THEWORLD AQUACULTURE SOCIETY

Vol. 45, No. 3June, 2014

doi: 10.1111/jwas.12114

Effects of Hydraulic Dredging for Mercenaria mercenaria, NorthernQuahog, on Sediment Biogeochemistry

Shannon L. Meseck1, Renée Mercaldo-Allen, Julie M. Rose, Paul Clark,Catherine Kuropat, Jose J. Pereira, and Ronald Goldberg

NOAA Fisheries, Northeast Fisheries Science Center, Milford Laboratory, 212 Rogers Avenue,Milford, Connecticut, 06460, USA

AbstractA before-after-control-impact (BACI) experiment was conducted to examine the effects of hydraulic

clam dredging on sediment biogeochemistry of a leased shellfish bed of Mercenaria mercenaria, northernquahog, over the course of an entire growing season. Six study plots (0.67 ha each), three dredgedand three not dredged, off of Milford, Connecticut, in Long Island Sound, were sampled from May toOctober 2009 for porewater fluxes of total ammonia, oxygen, and hydrogen. Particulate samples werealso analyzed for grain size, total nitrogen, total carbon, total sulfur, and organic carbon. Statisticalanalysis indicated no significant difference between dredged and not dredged sites. Grain size andoxygen flux explained 22% of the variation in the total benthic species assemblages; grain size and eithertotal carbon or organic nitrogen explained 18% of the variation in molluscan abundance. Our studydemonstrates that one-time hydraulic shellfish harvesting had minor effects on the sediment chemistryof a leased clam bed.

With increased shortages from capturefisheries and a growing human population,aquaculture has become one of the fastestgrowing, food-producing sectors in the world(FAO 2010). Annual aquaculture production,which represented less than 1 million m.t. in1950, has risen at an annual growth rate of 8.3%and reached 52.5 million m.t. by 2008 (FAO2010). Although bivalve aquaculture currentlyrepresents only 25% of the world’s aquacul-ture production (FAO 2010), over 80% of theshellfish that are consumed are obtained throughaquaculture practices, which have led to highscrutiny concerning potential effects of bivalveaquaculture on the environment (Shumway2011). Shellfish farming practices can varywidely depending on a variety of factors includ-ing species, tradition, environmental conditions,social acceptance, and local regulations (Ferreiraet al. 2011).

Molluscs have long been harvested fromestuaries by native people and colonial settlersfor hundreds of years (MacKenzie et al. 2002a,

1 Corresponding author.

2002b). Historically, the states of Florida, Vir-ginia, and Connecticut are the highest producersof molluscan shellfish (MacKenzie et al. 2002b),with Crassostrea virginica (the eastern oyster)and Mercenaria mercenaria (northern quahog)the two most frequently harvested species.Increased disease occurrence (i.e., Dermo), lossof habitat, and other environmental stressorshave resulted in a decline in harvests of C. vir-ginica since the 19th century (Mackenzie 1996).This reduction in oyster production contributedto increased commercial harvest of northernquahog during the 1920s, which became apromising commercial enterprise, following theintroduction of the hydraulic dredge (1958 inConnecticut) and seed hatcheries in the 1960s(Mackenzie et al. 2002b). As the dominantproducer of northern quahog, the USA har-vested over 4.1 million m.t. during 2011, valuedat US$87 million (http://www.fishwatch.gov/seafood_profiles/species/clams/species_pages/northern_quahog_clam_farmed.htm).

As the aquaculture industry continues to grow,there is an increased need to ensure that thecultivation and harvest of northern quahog have

Published 2014. This article is a U.S. Government work and is in the public domain in the USA.

301

302 MESECK ET AL.

minimal negative effects on the environment.Hard clam aquaculture may interact with theecosystem through (1) food consumption andwaste production and (2) harvesting and bedmaintenance (Dumbauld et al. 2009). In LongIsland Sound in Connecticut, the seafloor issurveyed and leased to harvesters, who relyon natural recruitment of hard clams. All theavailable shellfish beds in the state (28,328 ha)are currently under lease with income fromproduction rising from US$3.5 million in 1990(146,250 bags) to US$17.4 million (425,294bags) by 2010 (http://seagrant.uconn.edu/whatwedo/aquaculture/production.php, ConnecticutDepartment of Agriculture). In the adjacent NewYork waters of Long Island Sound, clams areharvested either from open grounds or, in somecases, young hatchery-reared clams are seededto populate leased beds for eventual harvest.

In Connecticut, towed hydraulic dredges areused to harvest hard clams. These dredges usehigh-pressure water jets to loosen the sedimentsand dislodged clams are collected in mesh bagsas the dredge bar passes over the fluidized bot-tom (MacKenzie et al. 2002a). Water pressure issufficient to remove clams without shell dam-age (Jolley 1972). Because hard clams growslowly, cultivated shellfish beds are dredgedevery 3–5 years to allow clams to reach har-vestable sizes (MacKenzie et al. 2002a, 2002b).The use of towed fishing gear elicits some con-cerns because of potential damage to non-targetbenthic organisms, chronic effects to diversity inthe dredge track, and potential biogeochemicalchanges in the sediments (Levy 1998; Watlingand Norse 1998; Watling 2005). The effectsof dredging on leased beds are considered lessextensive than those on wild beds because clam-mers know when to harvest in order to max-imize the catch of market size northern qua-hog, thereby reducing tow length and mortalityof nontarget organisms (Stokesbury et al. 2011).Sediment type, dredging gear, depth of watercolumn, currents, tides, and time of year areamong the factors that influence dredging out-comes (Falcão et al. 2003). Potential effects tothe benthic environment can include changes inthe biological community (Mercaldo-Allen andGoldberg 2011; Goldberg et al. 2012) and in

the biogeochemistry of the sediments (Mayeret al. 1991; Falcão et al. 2003). Some possiblebiogeochemical changes include resuspensionof ammonia from sediment porewater (Fanninget al. 1982), changes in redox potential (Aller1980; Ingall and Jahnke 1994), changes in sedi-ment grain size and porosity (Lenzi et al. 2005),changes in sediment particulate carbon/nitrogen(Lenzi et al. 2005), and release of anoxic sedi-ments to the surface (Badino et al. 2004). Theeffects on the benthic community can vary fromone location to another depending on the dredg-ing gear being used and on the physical, chemi-cal, and biological components of the local envi-ronment (Stokesbury et al. 2011).

Although the practice of hydraulic dredgingfor hard clam harvest has been conducted inConnecticut since the 1960s, information con-cerning biogeochemical effects and effects tothe benthic community from dredging on leasedbeds is limited. Before-after-control-impact(BACI) studies are frequently used to distinguishnaturally occurring environmental changes frommanmade activities (Underwood 1994; Dameet al. 2000, 2002; Hewitt et al. 2001; Stokesburyet al. 2011). In a BACI design, the control andimpact areas are assumed to behave similarlyexcept for any harvest-caused disturbances(Green 1979). Marine environments can behighly variable on small spatial scales; therefore,statistics that can address complex inequalitiesare utilized (Black and Miller 1991, 1994;Underwood 1991, 1992, 1994; Rangeley 1994).Assuming the null hypothesis, that there are nodifferences between the control (non-dredged)and impact (dredged) site, the experimental pro-tocol involves sampling preimpact, immediatelypostimpact, and continued sampling during therecovery phase (Stokesbury et al. 2011).

We designed a BACI study to assess thebiochemical effects of hydraulic dredging ona historically cultivated clam bed in LongIsland Sound. We measured sediment grain size,porewater fluxes (total ammonia, hydrogen,and oxygen flux), sedimentary total carbon,nitrogen, sulfur, and organic carbon to comparethe biogeochemistry of the sediments betweendredged and not dredged sites and before andafter dredging.

HYDRAULIC DREDGING AFFECTS SEDIMENTS 303

Figure 1. Study site in Long Island Sound (inset) off thecoast of Milford, Connecticut, USA. Projection showsa schematic of the six experimental plots (82× 122 meach), indicating dredged (D) and not dredged (ND)plots. Depths at mean low water are indicated in feet(1 ft= 0.31 m) “hrd S”= hard sand, “sft S”= soft sand.The numbers in the brackets are the mean surface(0–2 cm) grain size with standard error. Each letter rep-resents a different homogenous group.

Materials and Methods

The commercial clam bed was located in LongIsland Sound at 41∘11′N, 73∘5′W offshore ofMilford, Connecticut, USA (Fig. 1). A com-mercial clammer provided us with 4 ha on hisleased bed, where northern quahogs were lastharvested in 2007, to conduct the BACI experi-ment. According to nautical charts, this area wasexpected to have fairly uniform grain size witha tidal cycle water depth varying from 4.9 to6.1 m. The 4 ha plot was divided into six, 0.67 haboxes, which consisted of three control areas (1,3, 5) and three impacted plots (2, 4, 6). To facili-tate spatial randomization, each plot was furthersubdivided into nine boxes, and on each collec-tion date, one box was randomly selected forsampling.

The BACI design requires preimpact, postim-pact, and recovery sampling. Beginning onMay 28, 2009, samples were taken on bothcontrol and impacted sites on a biweekly

schedule aboard the NOAA Fisheries R/V“Victor Loosanoff.” One-time dredging of the“impacted” sites was conducted by the JesseD. Shellfish Company on July 6, 2009, usinga hydraulic clam dredge weighing 204 kg at atowing speed of 1.2–1.6 knots. This dredgingmethodology is typical of harvesting practicesconducted on leased commercial beds. A total of10 sampling trips were completed over a 24-wktime period.

Sediment Sampling

Sediment cores were obtained randomly fromdifferent locations in the study area (Fig. 1)using the sediment corer described in Alix et al.(2013). The coring device used was a gravitysurface corer that allowed for the recovery ofthe sediment–water interface and the sedimentimmediately below, with no disturbance tothe bottom (Hongve 1972). At each station,two sediment cores were collected for analysisof porewater pH, total ammonia, and oxygenconcentrations.

Fine resolution sediment core profiling (1 mm)yields better flux calculations and resolution thansectioning; however, the technology to accom-plish fine resolution sampling was not alwaysavailable. An oxygen micro-optical, 140-μmprobe in a needle (Loligo Systems, Tjele,Denmark2) was used for oxygen profiling. Theprobe was attached to a micro-manipulator formillimeter-scale resolution of oxygen profiles.An MI-414 pH electrode in a 16-gauge needlewas attached to the micro-manipulator to obtainmillimeter-scale resolution for pH. Oxygen andpH measurements were determined in 1 mmincrements in the top 10 mm of each core.As microelectrode probes for total ammoniawere not available, we obtained porewater fromcores that were sectioned at 2-cm intervals.Each section was placed in a 50-mL centrifugetube and tubes were centrifuged at 1000 g for20 min. The porewater was decanted, filteredthrough a 0.45-μm filter, and the effluent wascollected in a 15-mL centrifuge tube placed

2 The use of any trademark material or equipment does notindicate an endorsement by the federal government for thisequipment.

304 MESECK ET AL.

immediately on ice in the dark, for deter-mination of subsequent total ammonia. Thesediment remaining in the centrifuge tube wasanalyzed for particulate nitrogen, carbon, andsulfur.

Total ammonia was determined within 24 hof sample collection using a QuAAtro autoana-lyzer (Seal Analytical, Mequon, WI, USA) usingthe Berthelot reaction, as described by Hansenand Koroleff (1999). The detection limit of theinstrument was 0.05 μM.

Particulate carbon, nitrogen, and sulfur weredetermined with a Costech ECS 4010 CHNS ele-mental analyzer (Valencia, CA, USA). All sam-ples were dried in an oven (60 C) overnight.Sediments were ground using a Retsch PM 200grinder (Newton, PA, USA) to a size of 63 μm.Approximately 3-μg subsamples of sedimentwere weighed into tin boats with 0.5 μg of vana-dium oxide added for the determination of totalcarbon, nitrogen, and sulfur. A subsample ofeach sediment section was also acidified for thedetermination of organic carbon using the sameelemental analyzer. A standard reference mate-rial (SRM 8704 Buffalo River Sediment) wasanalyzed with the samples, with a reported totalcarbon value of 3.351%. Recovery of total car-bon measured 3.150± 0.327% (n= 50), withinthe reported value range.

Flux Calculations

Based on the oxygen micro-profiles, it appearsthat bioturbation effects in the sediments wereminor. As no oxygen could be detected beyond1 cm, sediment ammonia, hydrogen, and oxy-gen fluxes were calculated using Fick’s FirstLaw of diffusion (Berner 1980). This methodis commonly used for shallow-water, estuarinesediments (Emerson et al. 1984, Hammond et al.1985) where molecular diffusion represents themajor component during exchange of dissolvedsubstances between bottom sediments and over-lying water and is expressed by the formula:

J = −φmDs∂C∂z

where J is the flux, φ is the porosity, m has avalue of 3 for these surface sediments (Ullman

and Aller 1982), Ds is the effective diffusioncoefficient, and ∂C

∂zis the observed concentra-

tion gradient of porewater. Molecular diffusioncoefficients in seawater were corrected for thein situ, bottom-water temperature. Positive num-bers indicate a net flux into the sediment whilenegative numbers indicate a net flux out of thesediments.

Data Analysis and Statistics

A standard BACI-style statistical analysis ofmain effects and interactions was performedusing the permutational multivariate analysis ofvariance (PERMANOVA) add-on to the statis-tical software Primer v6 (Anderson et al. 2008;Clarke and Gorley 2006). PERMANOVA hasthe advantage of no assumptions of normal-ity and tests all main effects and interactionsfrom the BACI style of experimental design.PERMANOVA was used to determine the maineffects of treatment (impact versus control), timeperiod (predredging versus postdredging), sam-ple date (nested in time period), and plot (nestedin treatment). This analysis was completed forall measured chemical parameters. A draftsmanplot was examined prior to analysis to ensure thatvalues for each variable were evenly distributed,that is, not heavily skewed or containing extremeoutliers.

Benthic assemblage data from the samesamples are described thoroughly in Gold-berg et al. (2012). The species compositiondata from Goldberg et al. (2012) were usedin a distance-based redundancy analysis toinvestigate the relationship between sedimentchemistry and benthic assemblages. The anal-ysis was performed using the DISTLM routinewithin the PERMANOVA program (McArdleand Anderson 2001; Anderson et al. 2008).Benthic-assemblage data were square-roottransformed and the Bray–Curtis resemblancemeasure was used to generate similarity matri-ces for the total benthic assemblage and forthe subset of the assemblage belonging to thephylum Mollusca. As most of the Connecti-cut shellfish industry relies on natural set, wewere interested in identifying the relation-ships between the abundance of molluscs and

HYDRAULIC DREDGING AFFECTS SEDIMENTS 305

our measured physical and chemical param-eters. Ratios of total and organic carbon tonitrogen were not included among the predictorvariables. A draftsman plot was used to check formulti-colinearity among the remaining chemicalvariables, and because all correlations were wellbelow the recommended cutoff of 0.95, thesevariables were included in the DISTLM routineas predictor variables. The “Best” selectionprocedure was used for model-building, whichexamined all possible combinations of predictorvariables. Both AICc and BIC were used asselection criteria, and models that had AICc orBIC values within 2 units of the best model areincluded here (Schwartz 1978; Sugiura, 1978;Anderson et al. 2008).

Results

Table 1 reports the results of the BACI-stylestatistical analysis of the data. Significant maineffects of sampling date and plot were observedin several of the chemical parameters. No maineffects of dredging were detected. Short-termeffects of dredging were observed for organicnitrogen and hydrogen flux (detailed below), butthese differences disappeared within weeks ofthe dredging event.

Grain size varied consistently among the plots,with fine sand located near shore, becoming veryfine sand moving offshore (Fig. 1). There wasno significant effect of grain size on any of thechemical parameters measured (Table 1). Therewas concern that grain size differences could bemasking minor, but significant, effects of dredg-ing because there were differences in grain sizeamong the plots. Thus, the PERMANOVA anal-ysis was performed first using grain size as acovariate and then also using organic carbon asa covariate. In general, running PERMANOVAwith grain size as a covariate eliminated signifi-cant plot effects that were observed for several ofthe chemical parameters, but in no instance didit alter the results of the test of treatment (i.e.,dredged versus not dredged). Running organiccarbon as a covariate had no effect on the resultsof the PERMANOVA analysis.

The oxygen flux varied from a mean of210 mmol/m2/d to 1104 mmol/m2/d at the

control site, with a similar range observed atthe impact site (Fig. 2). For both the controland impact sites, and throughout the samplingseason, oxygen was fluxing into the sediments.Significant effects of sampling date weredetected (P< 0.01). From May through August,the oxygen flux increased with the highestconcentration occurring in August, and valuesdecreasing after. No significant differenceswere observed between the control and impactsites (P= 0.66), predredged versus postdredgedsamples (P= 0.78), nor any of the interactionterms (all P> 0.11, Table 1).

Total ammonia flux varied from−21.51 mmol/m2/d to −65.90 mmol/m2/d atall sites (Fig. 2). Significant effects of samplingdate were detected (P= 0.04). Total ammoniafluxed out of the sediments during the samplingperiod with the largest fluxes observed duringthe month of August. There was no significantdifference between the control and impact sites(P= 0.21), predredged versus post dredgedsamples (P= 0.56), or any of the interactionterms (all P> 0.26, Table 1).

Unlike the flux from total ammonia andoxygen, which was consistently unidirec-tional throughout the season, the hydrogenflux was more variable, ranging from negative−0.52 mmol/m2/d to 0.58 mmol/m2/d (Fig. 2).A significant interaction was observed betweensampling date and treatment (P= 0.04). Signif-icant differences between control and impactsites were observed on sampling dates May 18,2009 (P= 0.01), and July 7, 2009 (P= 0.03).For both these sampling dates, the control hada lower flux of hydrogen than the impact sites.On May 18, the control site had a hydrogen fluxof −0.39± 0.27 mmol/m2/d, while the impactsite had a flux of −1.06± 0.65 mmol/m2/d. OnJuly 7, a day after dredging, a hydrogen fluxof+ 0.02± 0.13 mmol/m2/d was measured atthe control site while the impact site had aflux of +0.29± 0.08 mmol/m2/d. There was nosignificant difference between predredged andpostdredged samples (P= 0.55), or for the inter-action between predredged and postdredgedsamples with plot (P= 0.93).

Particulate organic carbon ranged from 2.10to 6.61 mg/g for the control sites, while the

306 MESECK ET AL.

Table 1. Results of BACI-style statistical analysis. Main effects include treatment (control versus impact), samplingperformed pre dredging versus postdredging, sample date, and plot.a

Factor (df)Mean

grain sizeOxygen

fluxAmmonia

flux Hydrogen fluxOrganicnitrogen

Organiccarbon

Totalnitrogen

Totalcarbon

Totalsulfur

Treatment (1)P 0.52 0.66 0.21 0.46 0.30 0.67 0.20 0.63 0.59MS 0.56 0.11 152.63 0.45 0.03 0.69 0.70 1.35 0.01F 0.44 0.20 1.85 0.64 1.22 0.19 1.98 0.24 0.32Predredging versus postdredging (1)P 0.12 0.78 0.56 0.55 0.12 0.10 0.60 0.22 0.55MS 0.43 <0.01 35.17 0.43 0.65 23.44 0.05 17.16 0.15F 3.66 0.08 0.39 0.43 4.14 5.24 0.33 2.04 0.41Sample date (8)P 0.10 <0.01 0.04 <0.01 0.01 0.06 0.06 0.15 0.18MS 1.03 1.23 632.94 1.40 0.08 4.26 1.00 8.93 0.61F 0.44 11.60 2.41 7.39 3.31 2.18 2.10 1.65 1.56Plot (4)P <0.01 <0.01 0.92 0.06 0.73 0.01 0.63 <0.01 0.08MS 1.44 0.95 60.65 0.50 0.01 7.60 0.31 23.48 0.93F 14.59 8.96 0.23 2.64 0.50 3.88 0.66 4.35 2.38Predredging versus postdredging× treatment (1)P 0.18 0.11 0.26 0.34 0.50 0.81 0.38 0.15 0.70MS 0.05 0.44 163.14 0.46 0.075 1.39 0.49 9.06 0.27F 1.75 2.32 1.43 1.21 0.86 0.50 1.11 1.93 0.62Predredging versus postdredging× plot (4)P 0.78 0.72 0.73 0.93 0.03 0.89 0.31 0.81 0.15MS 0.04 0.06 133.35 0.04 0.08 0.52 0.59 2.05 0.72F 0.43 0.54 0.51 0.21 3.22 0.26 1.24 0.38 1.84Sample date× treatment (8)P 0.89 0.14 0.75 0.04 0.21 0.01 0.76 0.46 0.52MS 0.04 0.18 165.35 0.49 0.04 6.16 0.28 5.39 0.36F 0.43 1.70 0.63 2.61 1.45 3.14 0.60 1.00 0.91

AICc=Modified Akaike Information Criterion; BIC=Bayesian Information Criterion; BACI= before-after-control-impact; MS=mean squares; RSS= residual sum of squares.

aInteractions of main effects are also included. Degrees of freedom for each test are inside the parentheses. Significantdifferences (P< 0.05) are indicated in bold.

impact sites ranged from 2.56 to 6.12 mg/g(Fig. 2). A significant interaction was observedbetween sampling date and treatment (P= 0.01).Significant differences between control andimpact sites were observed on July 7, 2009(P= 0.02), and August 4, 2009 (P< 0.01).These observed differences between control andimpact sites were not consistent in direction;control sites had a lower mean particulate car-bon on July 7 (4.07± 0.78 and 6.12± 0.78 mg/g,respectively), but impact sites had a lower meanparticulate carbon on August 4 (2.56± 0.97 and5.52± 0.24 mg/g, respectively). No significantdifferences were observed related to dredg-ing treatment (P= 0.67) or predredged versuspostdredged sampling (P= 0.10).

The total carbon concentration varied from5.00 to 12.47 mg/g for both the control andimpact sites with no significant treatmenteffect (P= 0.63, Fig. 2). No significantdifferences were observed for total carbonbetween predredged and postdredged samples(P= 0.22) sampling date (P= 0.15), or any ofthe interaction terms (all P> 0.15, Table 1).

Total nitrogen varied from 0.54 to 2.10 mg/gacross all samples with no significant treat-ment effect observed (P= 0.20, Fig. 2). As withtotal carbon, there was also no observed sig-nificant differences between predredged versuspostdredged samples (P= 0.60) sampling date(P= 0.06), or any of the interaction terms (allP> 0.31, Table 1).

HYDRAULIC DREDGING AFFECTS SEDIMENTS 307

Figure 2. Diffusive calculated porewater fluxes for each of the sampling dates for oxygen, total ammonia, hydrogen ion,and sedimentary organic carbon, total carbon, nitrogen, and sulfur. The arrow indicates when dredging occurred on thetime line.

Total sulfur content varied from 0.45 to5.69 mg/g across all sites and dates (Fig. 2), withno significant differences observed betweencontrol and impact sites (P= 0.59), betweenpredredged and postdredged samples (P= 0.55),sampling date (P= 0.18), or any of the interac-tion terms (all P> 0.15, Table 1).

Benthic invertebrate species compositiondata were reported previously (Goldberget al. 2012). Some of the dominant speciesincluded amphipods Ampelisca spp., Calliopius

laeviusculus and Leptocheirus pinguis; poly-chaetes Glycera spp., Clymenella torquata andNephtys spp.; crustaceans Pagurus longicarpus,Pinnixa spp., and Crangon septemspinosa;and bivalves M. mercenaria and Yoldia limat-ula (Goldberg et al. 2012). Distance-basedredundancy analysis, a form of multivariatemultiple regression, was used to partition thevariance observed in the biological assemblagesusing the multivariate matrix of chemical data(McArdle and Anderson 2001; Anderson et al.

308 MESECK ET AL.

Table 2. Model results for chemical parameters aspredictors of benthic assemblage species composition.

AICc BIC R2 RSS Variables

30.54 335.45 0.22 52,648 Grain size, oxygenflux

330.56 336.9 0.26 49,989 Grain size, oxygenflux, organicnitrogen

331.16 337.5 0.25 50,651 Grain size, oxygenflux, hydrogen flux

331.20 337.54 0.24 50,696 Grain size, oxygenflux, organic carbon

331.37 337.71 0.24 50,883 Grain size, oxygenflux, total nitrogen

Table 3. Model results for chemical parameters aspredictors of molluscan assemblages.

AICc BIC R2 RSS Variables

323.50 326.88 0.14 47,491 Grain size323.91 328.82 0.18 45,580 Grain size, total carbon323.97 328.88 0.18 45,641 Grain size, organic nitrogen

2008). Marginal tests for grain size and oxygenyielded significant P-values (<0.001) for thetotal benthic assemblage. Marginal tests foroxygen flux and ammonia flux yielded signifi-cant P-values (<0.001 and 0.03, respectively)for the assemblage of molluscs alone. Both theAkaike and Bayesian model indicated that grainsize and oxygen flux were the best predictors ofcommunity composition, but these two factorsexplained just 22% of the observed variance(Table 2). The addition of organic nitrogen,hydrogen flux, organic carbon, or total nitrogento the model improved the explanatory powerby an additional 2–4%. The multiple regressionanalysis was also performed with the molluscassemblages. Grain size was again the best pre-dictor of mollusc assemblage, explaining 14%of the variation in abundance, and the addition ofeither total carbon or organic nitrogen increasedthe predictive ability to 18% (Table 3).

Discussion

The chemical effects of harvest dredging onsediment chemistry are highly specific to localhabitat, environment, and the level of physical

impact. Disruption of sediments during har-vest can potentially release porewater and freesequestered nutrients (Coen 1995), resultingin short-term oscillations in chemistry at thesediment surface. We found that the one-timedredging event on July 7, 2009, on a shellfish bedin Long Island Sound did not have a significant,lasting effect on any of the chemical parame-ters that we measured (Table 1). There was aninteraction between sampling date and treatmentfor the hydrogen ion flux and particulate organiccarbon (Table 1), but this was short lived anddissipated within the next one or two samplingintervals.

Nearshore coastal sediments, prone to frequentnatural disruption, are likely to recover chem-ical equilibrium rapidly following brief, pulsedisturbance events. In these environments, theeffects of shellfish dredging on bottom sedi-ments may be indistinguishable from naturallyoccurring changes to the seabed (Constantinoet al. 2009; Sciberras et al. 2013). Falcão et al.(2003) found an immediate drop in porewa-ter chemical parameters including ammonium,nitrates, organic nitrogen, phosphate, and sili-cate just after clam dredging in Portugal, butthese values returned to original values withinminutes to hours. Other studies of clam har-vesting found no measurable changes to nitro-gen, sulfide, phosphate (Tarr 1977; Goodwin1977; Goodwin and Shaul 1980), dissolved oxy-gen (Tarr 1977), or total organic carbon (Sparsiset al. 1993) measurements related to cultivationof sediments. Similar to these other findings, ourstudy indicated that changes to sediment chem-istry related to dredging were generally imme-diate and short-lived, either resolving before ourfirst postdredge sampling on the following day orin the case of hydrogen flux and organic carbon,within several weeks of the dredging event.

Benthic invertebrate species compositioncan vary on large (e.g., bioregionalisation) orsmall (habitat) spatial scales. Sediment grainsize, environmental water chemistry, and habi-tat characteristics can influence the presenceof marine organisms. Sediment grain size isconsidered one of the most important variablesdefining habitat suitability for macrobenthicfauna (Ysebaert et al. 2002; Compton et al.

HYDRAULIC DREDGING AFFECTS SEDIMENTS 309

2009; Kraan et al. 2010). In our study, sedimentgrain size was the best single explanatory factorfor the observed variability in both the totalbenthic and molluscan assemblages, respec-tively, but explained just approximately 14% ofthe total variability for each of the two groups.We found that adding oxygenation of sedimentsto this multiple regression model improvedexplanatory power slightly for the total benthicassemblage, but not for the molluscan subset.A shift from oxygenated sediments to hypoxiahas been reported to result in a migration ofmobile epibenthic species with echinodermsand most crustaceans. Sedentary annelids,molluscs, and cnidarians are more tolerant ofoxygen depletion (Rosenberg et al. 1991; Diazand Rosenberg 1995). Occasionally, there can bea reduction in benthic species at higher oxygenlevels (>100 μM) that can be correlated with areduction in observed species due to subtoxiceffects (Thrush et al. 1992), but that was notobserved in these sediments. Green et al. (2013)suggest that calcite and aragonite saturationstate might be better indicators of molluscansettlement than the chemical parameters andgrain size classifications we measured. Furtherstudies are required to identify specific factorsthat may be influencing molluscan settlement atdredged sites.

Our BACI-style experiment found thatone-time hydraulic shellfish dredging, asconducted by a commercial harvester in Con-necticut, had minor effects on the sedimentchemistry of a leased clam bed, which resolvedwithin days or weeks. Sediment grain sizeand oxygen concentration influenced benthiccommunity structure and molluscan abundancemore strongly than any of the other chemicalparameters we measured.

Acknowledgments

We thank Captains R. Alix and W. Schreinerfor vessel operations; S. Auscavitch, S. DeCarli,M. Dixon, J. Esposito, J. Goggins, K. Harper,T. L. Nguyen, D. Redman, J. Reidy, G. Sen-nefelder, and A. Wu for technical support; andL. Williams of the Jesse D. Shellfish Companyfor dredging our study site and allowing us to

sample their leased clam beds. The State of Con-necticut, Department of Agriculture, Bureau ofAquaculture’s D. Carey and T. Barrell helpedcoordinate experimental dredging and shellfishmarking beds.

Literature CitedAlix, R. J., W. Schreiner, and S. L. Meseck. 2013.

Construction of an inexpensive gravity sedimentcorer for micro-profiling O2 and pH in marine sedi-ments. NOAA Tech Memo NMFS NE 224. NationalMarine Fisheries Service, Woods Hole, Massachusetts,USA 02543–1026. p. 13. Accessed August 2013 athttp://www.nefsc.noaa.gov/nefsc/publications/

Aller, R. C. 1980. Quantifying solute distributions in thebioturbated zone of marine sediments by defining anaverage microenvironment. Geochimica Cosmochim-ica Acta 44:1955–1965.

Anderson, M. J., R. N. Gorley, and K. R. Clarke. 2008.PERMANOVA+ for PRIMER: Guide to software andstatistical methods. Primer-E Limited, Plymouth, UK.

Badino, G., F. Bona, A. Mafiotti, O. Giovanardi, andF. Pranovi. 2004. Impact of mechanical clam harvest-ing on a benthic habitat: evolution by means of sedimentprofile imagining. Aquatic Conservation: Marine andFreshwater Ecosystems 14:S59–S67.

Berner, R. A. 1980. Pages 237. Early diagenesis: a theo-retical approach. Princeton University Press, Princeton,New Jersey, USA.

Black, R. and R. J. Miller. 1991. Use of the intertidal zoneby fish in Nova Scotia. Environmental Biology of Fishes31:109–121.

Black, R. and R. J. Miller. 1994. The effects of seaweed har-vesting on fishes: a response. Environmental Biology ofFishes 39:325–328.

Clarke, K. R. and R. N. Gorley. 2006. Primer v6: usermanual/tutorial. Primer-E Ltd., Plymouth, UK.

Coen, L.D. 1995. A review of the potential impacts ofmechanical harvesting on subtidal and intertidal shell-fish resources. South Carolina Department of NaturalResources, Marine Resources Research Institute. p. 46pp. + 3 appendices.

Compton, T. J., A. Tineke, A. Troost, J. Drent, C. Kraan,P. Bocher, J. Leyrer, A. Deking, and T. Piersma. 2009.Repeatable sediment associations of burrowing bivalvesacross six European tidal flat systems. Marine EcologyProgress Series 382:87–98.

Constantino, R., M. B. Gaspar, J. Tata-Regala, S. Car-valho, J. Cúrdia, J. T. Drago, R. Taborda, and C. C.Monteiro. 2009. Clam dredging effects and subsequentrecovery of benthic communities at different depthranges. Marine Environmental Research 67:89–99.

Dame, R., D. Bushek, D. Allen, D. Edwards, L. Gregory,A. Lewitus, S. Crawford, E. Koepfler, C. Corbett,B. Kjerve, and T. Prins. 2000. The experimentalanalysis of tidal creeks dominated by oyster reefs: the

310 MESECK ET AL.

pre-manipulation year. Journal of Shellfish Research19:361–369.

Dame R., D. Bushek, D. Allen, A. Lewitus, D. Edwards,E. Koepfler, and L. Gregory. 2002. Ecosystemresponse to bi valve density reduction: managementimplications. Aquatic Ecology 36:51–65.

Diaz, R. J. and R. Rosenberg. 1995. Marine benthichypoxia: a review of its ecological effects and thebehavioral responses of benthic macrofauna. AnnualReview of Oceanography and Marine Biology33:245–303.

Dumbauld, B. R., J. L. Ruesink, and S. S. Rumrill. 2009.The ecological role of bivalve shellfish aquaculture inthe estuarine environment: A review with application tooyster and clam culture in West Coast (USA) estuaries.Aquaculture 290:196–223.

Emerson, S., R. Jahnke, and D. Heggie. 1984.Sediment-water exchange in shallow water estuarinesediments. Journal of Marine Research 42:709–730.

Falcão, M., M. B. Gaspar, M. Caetano, M. N. Santos,and C. Vale. 2003. Short-term environmental impactof clam dredging in coastal waters (south of Portu-gal): chemical disturbance and subsequent recovery ofseabed. Marine Environmental Research 56:649–664.

Fanning, K. A., K. L. Carder, and P. R. Betzer. 1982. Sed-iment resuspension by coastal waters: a potential mech-anism for nutrient re-cycling on the ocean’s margins.Deep Sea Research 29:953–965.

FAO. 2010. The State of World Fisheries and Aquaculture2010, Food and Agriculture Organization of the UnitedNations, Rome, Italy.

Ferreira, J. G., A. J. S. Hawkins, and S. B. Bricker. 2011.The role of shellfish farms in provision of ecosystemgoods and services. Pages 3–31 in S. Shumway, editor.Shellfish aquaculture and the environment. John Wiley& Sons, Inc., Oxford, UK.

Goldberg, R., R. Mercaldo-Allen, J. M. Rose, P. Clark,C. Kuropat, S. L. Meseck, and J. J. Pereira. 2012.Effects of hydraulic shellfish dredging on the ecologyof a cultivated clam bed. Aquaculture EnvironmentInteractions 3:11–21.

Goodwin, L. 1977. Some environmental impacts of subtidalhardshell clam harvest with hydraulic clam harvestersin a small experimental plot in Puget Sound. State ofWashington, Department of Fisheries, Olympia, Wash-ington, USA. p. 16.

Goodwin, L. and W. Shaul. 1980. Studies of mechanicalclam harvest on an intertidal beach near Port Townsend,Washington Progress Report 119. State of Washington,Department of Fisheries, Olympia, Washington, USA.

Green, R. H. 1979. Sampling design and statistical methodsfor environmental biologists. John Wiley & Sons, NewYork, New York, USA.

Green, M. A., G. G. Waldbusser, L. Hubazc, E. Cathcart,and J. Hall. 2013. Carbonate mineral saturation state asthe recruitment cue for settling bivalves in marine muds.Estuaries and Coasts 36:18–27.

Hammond, D. E., C. R. Fuller, D. Harmon, B. Hartman,M. Korosec, L. G. Miller, R. Rea, S. Warren, W.

Berelson, and S. W. Hager. 1985. Benthic fluxes in SanFrancisco Bay. Hydrobiologia 129:69–90.

Hansen, H. and F. Koroleff. 1999. Determination of nutri-ents. Pages 159–228 in K. Grasshoff, K. Kremling,and M. Ehrhardt, editors. Methods of seawater analysis.Wiley-VCH, Germany, New York, New York.

Hewitt, J. E., S. E. Thrush, and V. J. Cummings. 2001.Assessing environmental impacts: effects of spatialtemporal variability at likely impact scales. EcologicalApplications 11:1502–1516.

Hongve, D. 1972. En bunnhenter som er lett å lage. Fauna25:281–283.

Ingall, E. and R. Jahnke. 1994. Influence of water-columnanoxia on the elemental fractionation of carbon andphosphorus during sediment diagenesis. Marine Geol-ogy 139:219–229.

Jolley, J. W. 1972. Exploratory fishing for the sunray Venusclam Macrocallista nimbosa in northwest Florida.Florida Department of Natural Resources MarineResearch Lab Technological Services 67. pp. 1–42.

Kraan, C., G. Aarts, J. W. D. Meer, and T. Piersma.2010. The role of environmental variables in structuringlandscape-scale species distributions in seafloor habi-tats. Ecology 91:1583–1590.

Lenzi, M., M. G. Finoia, E. Persia, S. Comandi, V.Gargiuol, D. Solari, P. Gennaro, and S. Porrello.2005. Biogeochemical effects of disturbance in shallowwater sediment by macroalgae harvesting boats. MarinePollution Bulletin 50:512–519.

Levy, S. 1998. Watery wastelands. New Scientist 2134:40–44.

MacKenzie, C. L. Jr.. 1996. History of oystering in theUnited States and Canada, featuring the eight greatestoyster estuaries. Marine Fisheries Review 58:1–78.

MacKenzie, C. L. Jr., A. Morrison, D. L. Taylor, V. G. Bur-rell Jr., W. S. Arnold, and A. T. Wakida-Kusunoki.2002a. Quahogs in eastern North America, Part I. Biol-ogy, ecology, and historical uses. Marine FisheriesReview 64:1–55.

MacKenzie, C. L. Jr., A. Morrison, D. L. Taylor, V. G. Bur-rell Jr., W. S. Arnold, and A. T. Wakida-Kusunoki.2002b. Quohogs in eastern North America, Part II His-tory by province and state. Marine Fisheries Review64:1–64.

Mayer, L. M., D. F. Schick, R. H. Findlay, and D. L. Rice.1991. Effects of commercial dragging on sedimen-tary organic matter. Marine Environmental Research31:249–261.

McArdle, B. H. and M. J. Anderson. 2001. Fitting mul-tivariate models to community data: a comment ondistance-based redundancy analysis. Ecology Letters82:290–297.

Mercaldo-Allen, R. and R. Goldberg. 2011. Review ofthe ecological effects of dredging in the cultivation andharvest of molluscan shellfish. US Department of Com-merce, NOAA Technical Memorandum NMFS-NE-220; 78p. Accessed at http://www.nefsc.noaa.gov/nefsc/publications/.

HYDRAULIC DREDGING AFFECTS SEDIMENTS 311

Rangeley, R. W. 1994. The effects of seaweed harvestingon fishes: a critique. Environmental Biology of Fishes39:319–323.

Rosenberg, R., B. Hellman, and B. Johansson. 1991.Hypoxic tolerance of marine benthic fauna. MarineEcology Progress Series 79:127–131.

Schwartz, G. 1978. Estimating the dimension of a model.Annals of Statistics 6:461–464.

Sciberras, M., H. Hinz, J. D. Bennell, S. R. Jenkins,S. J. Hawkins, and M. J. Kaiser. 2013. Benthiccommunity response to a scallop dredging closurewithin a dynamic seabed habitat. Marine EcologicalProgress Series 480:83–98.

Shumway, S. E. 2011. Preface. Pages XV-XVII in S.Shumway, editor. Shellfish aquaculture and the environ-ment. John Wiley & Sons, Inc., Oxford, UK.

Sparsis, M., J. T. DeAlteris, and M. A. Rice. 1993.Effects of bottom cultivation on quahogs and other bot-tom invertebrates in Narragansett Bay. Pages 63–78 in:M.A. Rice and D. Grossman-Garber, editors. Proceed-ings of the Second Rhode Island Shellfish Industry Con-ference, August 4, 1992, University of Rhode Island,Narragansett Bay Campus. Report no. RIU-W-92-002,Rhode Island Sea Grant, University of Rhode Island,Narragansett, Rhode Island, USA.

Stokesbury, K. D. E., E. P. Baker, B. P. Harris, andR. B. Rheault. 2011. Environmental impacts relatedto mechanical harvest of cultured shellfish. Pages319–338 in S. Shumway, editor. Shellfish aquacultureand the environment. John Wiley & Sons, Inc., Oxford,UK.

Sugiura, N. 1978. Further analysis of the data by Akaike’sinformation criterion and the finite corrections. Com-munications in Statistics – Theory and Methods7:13–26.

Tarr, M. 1977. Some Effects of Hydraulic Clam Harvestingon Water Quality in Killisut Harbor, Port Susan, andAgate Pass, Washington. Progress Report 22. State ofWashington, Department of Fisheries, Olympia, Wash-ington, USA. p. 82.

Thrush, S. F., R. Pridmore, J. E. Hewitt, and V. J. Cum-mings. 1992. Adult infauna as facilitators of coloniza-tion on intertidal sandflats. Journal of ExperimentalMarine Biology and Ecology 159:253–265.

Ullman, W. J. and R. C. Aller. 1982. Diffusion coeffi-cients in near-shore marine sediments. Limnology andOceanography 27:552–556.

Underwood, A. J. 1991. Beyond BACI: experimentaldesigns for detecting human environmental impactson temporal variations in natural populations. Aus-tralian Journal of Marine and Freshwater Research42:569–587.

Underwood, A. J. 1992. Beyond BACI: the detection ofenvironmental impacts on populations in the real, butvariable, world. Journal of Experimental Marine Biol-ogy and Ecology 181:145–178.

Underwood, A. J. 1994. On beyond BACI: sampling designsthat might reliably detect environmental disturbances.Ecological Applications 4:3–15.

Watling, L. 2005. The global destruction of bottom habitatsby mobile fishing gears. Pages 198–210 in E. A.Norse and L. B. Crowder, editors. Marine conservationbiology. Island Press, Washington, D.C.

Watling, L. and E. A. Norse. 1998. Disturbance of theseabed by mobile fishing gear: a comparison to forestclearcutting. Conservation Biology 12:1180–1197.

Ysebaert, T., P. Meire, P. M. J. Herman, and H. Verbeek.2002. Macrobenthic species response surfaces alongestuarine gradients: prediction by logistic regression.Marine Ecology Progress Series 225:79–95.