Aquaculture 314 (2011) 87–99

Contents lists available at ScienceDirect

Aquaculture

j ourna l homepage: www.e lsev ie r.com/ locate /aqua-on l ine

Modeling ecological carrying capacity of shellfish aquaculture in highly flushedtemperate lagoons

Carrie Byron a,b,⁎, Jason Link c, Barry Costa-Pierce a,d, David Bengtson a

a University of Rhode Island, Department of Fisheries, Animal and Veterinary Sciences, Kingston, RI 02881, USAb Gulf of Maine Research Institute, 350 Commercial Avenue, Portland, ME 04101, USAc National Marine Fisheries Service, Northeast Fisheries Science Center, 166 Water Street, Woods Hole, MA 02543, USAd Rhode Island Sea Grant College Program, University of Rhode Island, Narragansett, RI 02882, USA

⁎ Corresponding author at: University of Rhode IslAnimal and Veterinary Sciences, Kingston, Rhode Island

E-mail addresses: carriebyron@my.uri.edu, cbyron@jason.link@noaa.gov (J. Link), bcp@gso.uri.edu (B. Costa(D. Bengtson).

0044-8486/$ – see front matter © 2011 Elsevier B.V. Aldoi:10.1016/j.aquaculture.2011.02.019

a b s t r a c t

a r t i c l e i n f o

Article history:Received 31 August 2010Received in revised form 3 February 2011Accepted 12 February 2011Available online 18 February 2011

Keywords:Carrying capacityAquacultureShellfishModelingEcopathLagoons

Lagoons are some of the most productive systems on the planet — not only for aquaculture, but for fisheries,recreation, and as nurseries for many important species. These systems are also highly susceptible todegradation. Aquaculture is a rapidly increasing industry capable of impacting these sensitive systems. Densehuman populations and intensive multiple uses of ecologically sensitive coastal waters have forced resourcemanagers to evaluate the impact of expanding shellfish aquaculture to ensure sustainable development. Amass-balance ecosystem model of highly flushed temperate lagoons in Rhode Island, USA was constructed tocalculate the ecological carrying capacity for shellfish aquaculture. Cultured oyster biomass is currently12 t km−2 live weight and could be increased to 62 times this value before exceeding the ecological carryingcapacity of 722 t km−2. The lagoons were found to be a detritus-dominated system with high energythroughput which may permit the high capacity of the system for additional shellfish biomass allowingmanagers to consider expansion of shellfish aquaculture to densities much higher than comparativelyoligotrophic systems could support. Managing by ecological carrying capacity follows an Ecological Approachto Aquaculture bymaintaining ecological integrity and the sustainability of the aquaculture industry as well asother human uses of the lagoon system.

and. Department of Fisheries,02881, USA.

gmri.org (C. Byron),-Pierce), bengtson@uri.edu

l rights reserved.

© 2011 Elsevier B.V. All rights reserved.

1. Introduction

Coastal lagoons are inland shallow water bodies connected to theocean by an inlet (Kjerfve, 1994). High productivity of estuarinelagoons supports myriad biological process and human uses (Kjerfve,1994; Mann, 2000; Nixon, 1982). Proximity to shore, shallow depthand protected shorelines make most lagoon systems accessible tohumans and susceptible to anthropogenic degradation.

Of the many human uses and industries dependent on healthylagoon systems, aquaculture deserves notable attention. Success of ashellfish aquaculture industry is dependent on high productivity oflagoon systems. Cultured shellfish feed on ambient plankton anddetritus, requiring no inputs of feeds, drugs or fertilizers. Shellfishfarms are directly linked to the greater lagoon-wide system and donot operate independently of natural ecosystem processes andconditions. While it is widely recognized that shellfish aquacultureprovides many positive ecosystem services, there is a point where an

excessive biomass of cultured shellfish can be predicted to threatenthe ecological integrity of the ecosystem.

Bivalve aquaculture is increasing at a rapid rate in concentratedareas around the world (Costa-Pierce, 2008; Costa-Pierce et al., 2010;FAO, 2009) and has a long history in estuarine lagoons (Kurlansky,2006; Valiela, 2006). Unearthed oystermiddens in theNorthesternU.S.dating back to the 1600s are evidence of substantialwild harvests fromcoastal waters (Kurlansky, 2006). Substantial harvests in Rhode Island(RI) lagoons continued into the 1950s fromwild and cultivated oystersuntil breachways were stabilized (Lee, 1980). Breachways wereconstructed with the purpose of improving the shellfish industry byincreasing flushing and salinity (Lee, 1980). Stabilized breachways didalleviate symptoms of stagnation and eutrophication but also hadmany adverse effects including: loss of brackish water fisheries,increased sedimentation, and changes in species composition (Lee,1980). Human use of the RI coastal environments has increaseddramatically in recent decades. An increased number of septic systemsalong with a series of severe storm events most likely led to thepermanent decline of the shellfishery (Lee, 1980). Today naturalpopulations of bay scallops and oysters exist in small isolated patchesthrough heavily monitored and maintained restoration efforts(DeAngelis et al., 2008; Hancock, et al., 2007). Rhode Island lagoonsare subject to cumulative anthropogenic impacts which have greatly

88 C. Byron et al. / Aquaculture 314 (2011) 87–99

altered the system, nutrient regimes, sediment geochemistry, and thedelicate balance of the many species that rely on this habitat (Lee,1980).

Today, the cultivated oyster industry has effectively eclipsed thewild harvest production industry on account of its decline. While athriving wild harvest clam industry still exists, oysters (Crassostrea-virginica) make up 99% of the aquaculture in the state. In fact, theaquaculture industry has been doing extremely well in recent yearsand has increased from a $130,000 to a $1,600,000 industry in only sixyears, effectively doubling the space utilized for the industry to 28farms or 50 ha statewide. Half of the state's aquaculture is in lagoons(Alves, 2007). On a global scale, this industry is still quite small.However, the lagoons are also quite small ranging from 0.72 to7.85km2 and located in the smallest and second most denselypopulated state in the USA. As such, the rapid rate of increase isquite notable to commercial clam harvesters and other users of thelagoons. Aquaculture is only one of several industries and recreationaluses of this area. It is in the best interest of all users to mitigate userconflicts and maintain ecological integrity.

Proper management is prudent to avoid perturbations of thesedelicate, heavily used and ecologically important systems. AnEcological Approach to Aquaculture (EAA) is one way to preservethe anthropogenic and ecological resources of this system (Soto,2010). An EAA is guided by three principles: “aquaculture isdeveloped in the context of ecosystem functions and services withno degradation of these beyond their resilience capacity, aquacultureshould improve human-wellbeing and equity for all relevantstakeholders, and aquaculture should be developed in the context ofother relevant sectors” (Soto, 2010). EAA acknowledges equity amonghuman uses while maintaining ecological sustainability.

An EAA can be applied using carrying capacity. Carrying capacityhas been adapted into four types of carrying capacity appropriate forbivalve aquaculture (Inglis et al., 2002):

1. Physical — “total area of marine farms that can be accommodatedin the available physical space”,

2. Production — “the stocking density of bivalves at which harvestsare maximized”,

3. Ecological — “the stocking or farm density which causes unaccept-able ecological impacts”,

4. Social — “the level of farm development that causes unacceptablesocial impacts.”

While physical and production carrying capacity are useful on afarm-scale, acknowledging that the farm is only a part of a largerecosystem requires consideration of ecological and social carryingcapacities. In order to take an ecological approach to aquaculture(Soto, 2010), it is helpful to consider ecological carrying capacity.

Both the ecological and social carrying capacities are defined bythe acceptability of change and, therefore, depend on social values(McKindsey et al., 2006). McKindsey et al. (2006) explained thatsociety defines the variables of interest and howmuch those variablescan change. Therefore, society has a part in defining acceptability.Society can determine the acceptability of alterations to sustainedecological function, species biomasses and energy flows betweentrophic levels. Stakeholders in RI wanted to calculate carrying capacityofbivalve aquaculture for current ecosystem conditions and weretherefore, unwilling to accept any substantial change in ecosystemfunction, biomasses, or energy flows (Byron et al., 2011). After thegoals and definition of carrying capacity are clarified, ecologicalcarrying capacity can be calculated using mass-balance modeling(Byron et al., 2011; Jiang and Gibbs, 2005; McKindsey et al., 2006).

1.1. Modeling

Ecopath is a static, mass-balance ecosystem-based modelingsoftware that focuses on energy transfer between trophic levels and

is widely used in fisheries management (www.ecopath.org). Ecopathhas been used for modeling a wide range of systems andmanagementscenarios (Christensen, 1995; Christensen and Pauly, 1993; Monacoand Ulanowicz, 1997; Vasconcellos et al., 1997) including the carryingcapacity of bivalve aquaculture (Jiang and Gibbs, 2005). It differs frommost other modeling approaches because it incorporates the fulltrophic spectrum, which is what makes it truly an ecosystem modelappropriate for ecosystem-based management and determiningecological carrying capacity. Conversely, models of productioncarrying capacity are most applicable at the farm-scale and do notnecessarily include all trophic levels (Bacher et al., 1998; Carver andMallet, 1990; Nunes et al., 2003; Raillard and Ménesguen, 1994). Thisapproach is useful for determining production potential on a farm butis shortsighted for ecosystemmanagement where several user groupsdepend on the stability and sustainability of the entire system.Furthermore, Ecopath provides a methodology to standardize modeloutputs, thereby making it easy to compare across systems. Overall,Ecopath is a good balance between simplicity and the complexity ofother ecosystem models.

Since Ecopath is a static model, its complexity is limited comparedto dynamic models that are capable of tracking variability in fluxesacross different temporal and spatial scales. Some critics may view thestatic nature of Ecopath as a limitation. However, in circumstanceswhere long-term data sets are not available or data are disparate,Ecopath provides a standardized methodology for developing amodel. As data improve over time, Ecopath can be expanded toEcosim and Ecospace which are capable of tracking fluxes over timeand space, respectfully.

Ecopath, like any model, does not come without shortfalls andlimitations (Plagányi and Butterworth, 2004). Most shortcomings areattributed to user error such as uncritical use of Ecopath defaultsettings — not all species groups should be treated equally eventhough the default settings are the same (Plagányi, 2007). Forexample, the default value for unassimilated consumption is 0.2which underestimates egestion for herbivores and detritivores. Amore reasonable rate for these groupswould be 0.4 (Christensen et al.,2005). Ecopath settings used in this model are further described in theModel considerations section of the Materials and methods. Perhapsthemost unavoidable shortfall of any ecosystemmodel is the quantityand quality of data available to feed the model. This study attemptedto minimize this shortfall by using data collected at the modellocations to calculate input parameters and by employing a series ofdiagnostic tests (Link, 2010) that were used to evaluate dataparameterization and will flag areas of data weakness that mayneed further investigation prior to model balancing.

2. Materials and methods

2.1. Study area

A string of nine barrier beach lagoons line the southern shore ofRhode Island (W71°30′ N41°15′). The ecology and natural history ofthese lagoons have been described extensively by others since the1940s (Beutel, 2009; Brown, 1962; Burnett, 2007; Conover, 1961;Crawford, 1985; CRMC, 1999; DeAngelis et al., 2008; Ernst et al., 1999;Lee, 1980; Lee and Olsen, 1985; Lee et al., 1997;Masterson et al., 2006;Nixon and Buckley, 2007; Satchwill and Gray, 1990,1991; Satchwilland Sisson, 1990; Thorne-Miller and Harlin, 1984; Thorne-Miller et al.,1983; Wright et al., 1949). These ecosystems are brackish withvarying salinity (4–31 ppt) and flushing rates (Hougham and Moran,2007; Meng et al., 2004; Olsen and Lee, 1982; Thorne-Miller et al.,1983). Nutrient cycles in these lagoons are driven by flushing of non-performing septic systems. Five of these lagoons have stabilizedbreachways and thus high flushing rates (residence timeb5 days) andsalinity (27–31 ppt) (Point Judith, Winnipaug, Quonochontaug,Potters, Ninigret). They range in size from 1.5 to 7.8 km2. The other

89C. Byron et al. / Aquaculture 314 (2011) 87–99

four have intermittent or no breachways resulting in low flushingrates (residence timesN10 days) and salinity (2–26 ppt) (Green Hill,Cards, Trustom, Maschaug) (Hougham and Moran, 2007; Houghamet al., 2008; Lee, 1980). These physical differences create ecologicaldifferences in the structure and functioning of the systems. Biomassesfor species and interactions between species are expectedly differentand therefore demand separate attention in ecosystem modeling.Only highly flushed lagoons were considered for this study becausethey are the only lagoons with aquaculture.

2.2. Functional groups

A model was constructed for a typical highly flushed RI lagoon. Noone particular lagoon was used for data inputs, rather all highlyflushed lagoons were considered where data were available. Since noprior modeling efforts have been made for the lagoons, a conceptualdiagram was constructed from expert knowledge [academics, coastal

Table 1Key species of RI highly flushed lagoons considered in this model.

Functional group Species common name

DetritusBenthic microalgae Diatoms, dinoflagellates, cyanobacteriaBenthic bacteriaDeposit feeders Polychaete worms

Benthic copepodsMacroalgae Brown algae

Red algae

Green algaeEelgrass EelgrassEpibenthic invertebrates Crabs

Grass shrimpAmphipodsJuvenile lobsterMud snail

Benthic feeding fish TautogCunnerBlack sea bassScupWinter flounder

Filter feeders OystersClamSlipper limpetAscidians/tunicates/sea squirtsClam

Cultured oysters OystersPhytoplanktonHolozooplankton Pelagic copepodsMerozooplankton Crustacean larvae (Nauplii)

Bivalve larvaeFish larvae

Planktivorous fish SilversidesMenhadenAlewife and hickory shadMackerelMummichog and striped killifishSheepshead minnowsCtenophoresLions mane jelly

Piscivorous fish BluefishStriped bass

Birds Geese and swansDabblingsDiversSeaducks

MergansersCormorantsGullsWaterbird

managers, and members of the Working Group on AquacultureRegulations (WGAR) led by the Coastal Resources ManagementCouncil (CRMC)] to represent the key functional groups in a typicalhighly flushed lagoon. All trophic levels were given equal consider-ation. Sixteen functional groups and all major species within eachfunctional group were defined and compiled in a list (Table 1). Thisconceptual diagram and species list was the basis for data acquisitionand input parameters of the model.

2.3. Input parameters

Ecopath requires input of threeparameters [biomass (B), production/biomass (P/B), consumption/biomass (C/B)] for every defined functionalgroup in the system (Christensen et al., 2005). From these threeparameters, one can calculate the fourth main parameter required forbalancing, ecotrophic efficiency (EE), which is a measure of the amountof production eaten in the system. The final two input components

Scientific name

Families Capitellidae, Maldanidae, NereidaeOrder HarpacticoidaFucusdistichus, F. vesiculosus, AscophyllumnodosumGracilariaspp, Chondruscrispus, Ceramiumvirgatum,Polysiphonia sppEnteromorphaspp,Ulvaspp, Codium fragile, CladophorasppZostera marinaLibiniaemarginata, Carcinusmaenas, HemigrapsussanguineusCancerirroratus, C. borealisPalaemonetessppAmphipodaHomarusamericanusLittorinaspp, IllyanassaobsoletaTautogaonitisTautogolabrusadspersusCentropristisstriataStenotomuschrysopsPseudopleuronectesamericanusCrassostreavirginicaMercenariamercenariaCrepidulafornicataClass AscidiaceaMyaarenariaCrassostreavirginica

Class CopepodaSubphylum CrustaceaClass Bivalvia

MenidiamenidiaBrevoortiatyrannusAlosapseudoharengus, A. mediocrisScomberscombrusFundulusheteroclitus, F. majalisCyprinodonvariegatusMnemiopsisleidyi, BeroëovataCyaneacapillataPomatomussaltatrixMoronesaxatilisBrantaalbeola, B. canadensis, Cygnus spp.Anasrubripes, A. americana, A. strepera, A. platyrhynchosAythyamarilaHistrionicushistrionicus, Clangulahyemalis, Somateriamollissima,Bucephlaclangula, Melanittaperspicillata, M. nigra, M. fuscaMergusserrator, LophodytescucullatusPhalacrocoraxspp.Larus spp.Gaviaimmer, Podicepsauritus

90 C. Byron et al. / Aquaculture 314 (2011) 87–99

whichmust be entered into themodel for every functional group aredietcomposition and fisheries removals (i.e. aquaculture harvest or wildcatch).

2.4. Data

Data for input parameters were taken from existing surveys andmonitoring effortsmade by a variety of government, non-government,and academic groups. Recent (2005–2008) data from studiesconducted in highly flushed RI lagoons were given priority. Despitethe attention these lagoons have received through the years, very fewmonitoring and survey efforts had been conducted over extendedtemporal and spatial scales. Combining the data for several of thehighly flushed lagoons renders a more complete database anddescription that can be used to parameterize the model. Considerablyless data, surveys, and information were available for the poorlyflushed lagoons and these datawere not considered in parameterizingthis model. Where specific data were lacking, we were forced toassume that species composition and food web structure were similaramong the five nearby lagoons with similar physical characteristics.Several local biologists suggested that this assumptionwas reasonable,allowing us to use a single genericmodel to characterize thefive highlyflushed lagoons.

The few efforts that were taken by others to survey the highlyflushed lagoons were typically not continued through the winterleaving data gaps in the year-round examination of the ecology of thisecosystem. Since Ecopath is a static snapshot that captures theaverage of year-round conditions, a yearly time series was needed.When a full year of data was not available, data were inferred fromseasonal conditions in nearby locations where year-round data wereavailable.

All units were converted to grams dry meat weight per metersquared (g DWm−2) according to McKinney et al. (2004). A 50:1carbon to chlorophyll-a ratio (Smayda and Borkman, 2008) and a 50%carbon to dry weight conversion (Link et al., 2006; Taylor et al., 1999)were assumed. Respiration for each functional group was calculatedas 65% of assimilated consumption (Link et al., 2006). Respiration,although not a required Ecopath input, is a useful parameter fordiagnostics. Most vital rates, if not otherwise specified, were inferredfrom other Ecopath models in similar systems (Christensen et al.,2009; Frisk et al., 2011; Monaco and Ulanowicz, 1997; Rybarcyk andElkaim, 2003; Rybarcyk et al., 2003).

Table 2(a) Input diet matrix. All diet matrix parameters represent percent (%) composition of diet ancultured oysters. (b) Mass-balanced input parameters. Values solved by Ecopath are in bold. Trconsumption/biomass (C/B), ecotrophic efficiency (EE), production/consumption (P/C). See Taestimated by Ecopath during the auto-balancing routine.

(a) Diet matrix

Prey/pred. 1 2 3 4 5 6 7 8 9 10

1 Brd – – – – – – – – – –

2 PiF – 0.045 – – – – – – – –

3 BFF – 0.04 0.018 – – – – – – –

4 PkF 0.115 0.889 0.001 0.005 – – – – – –

5 COy – – – – – – – – – –

6 FF 0.375 0.022 0.537 0.008 – – – – – –

7 EI 0.085 0.004 0.413 – – – 0.1 – – –

8 DF 0.025 – 0.001 – – – 0.1 – – –

9 Mer – – – 0.001 – – – – 0.1 0.110 Hol – – 0.03 0.986 0.1 0.1 – – – –

11 BB – – – – 0.2 0.2 0.2 0.25 0.4 0.212 Eel 0.05 – – – – – 0.1 – – –

13 Mac 0.075 – – – – – 0.1 – – –

14 BM 0.13 – – – 0.2 0.2 0.2 0.1 – –

15Phy – – – – 0.2 0.2 – – 0.15 0.516Det 0.145 – – – 0.3 0.3 0.2 0.65 0.35 0.2

aRI DEM 2006–2008; bBeutel, 2009; cSchult, 2010; dDurbin and Durbin, 1981; eMann, 2000; fTdata;hWarwicket al.,1979; hSchwinghamer et al., 1986; hLeloup et al., 2008; hRybaczyket al

Local surveys of birds were conducted on a single day in Januaryfrom 2006 to 2008 in highly flushed lagoons (Ninigret, Winnapaug,Quonochontaug, Point Judith, Potter) by the Rhode Island Departmentof Environmental Management (RIDEM). Since bird biomass andproduction vary greatly with season, landscape and human activity(McKinney et al., 2005), these single day counts were adjusted toreflect a year-round average based on monthly patterns of birddensity observed from year-round surveys (1992–1994) conductedby RI Fish and Wildlife in Trustom Lagoon. Counts of individuals fromsingle-day January surveys were converted to live weights using anaverage weight for each species (http://www.birds.cornell.edu/AllAboutBirds/BirdGuide/). The population of breeding adults wasestimated based on gull and cormorant measurements made in theSouthern Gulf of St. Lawrence (Savenkoff et al., 2004). It was assumedthat diet composition roughly corresponded to the gizzard contents ofRI waterfowl surveyed in Nov–Jan 1954–1957 (Cronan and Halla,1968). Only species that were found during the January survey ofhighly flushed lagoons in 2006–2008 (Ninigret, Winnapaug, Quono-chontaug, Point Judith, Potter) (RIDEM) were considered in compos-ing the diet matrix (Table 2a).

The 2006 RIDEM seine survey of the lagoons (Burnett, 2007) wasused in conjunction with the Narragansett Bay trawl survey (Longval,2009) to create an annual average biomass estimate of fish. Samplingtechniques employed when conducting seine surveys alone greatlyunderestimate actual biomass. The Narragansett Bay trawl survey isthe nearest and most representative survey of the lagoons since notrawls or surveys of adult fish are conducted in the lagoons. Countswere converted to mass based on the text series ‘Development ofFishes of the Mid-Atlantic Bight’ (Jones et al., 1978), Fishbase(Fishbase.org), and expert advice (David Beutel personal communica-tion). Specieswere grouped into benthic feedingfish, planktivorousfish,and piscivorous fish (Table 2b) for a total live biomass. Ctenophoreshave a similar ecological function in this system to planktivorous fish asa major consumer of zooplankton. Therefore, ctenophores wereincluded in the planktivorous fish group, even though they are notfish. Ctenophore biomasswas estimated from literature (Kremer, 1976)since they were not recorded in any seines or trawls and are a keyspecies in the planktivorous fish group (Table 1).

Vital rates vary across fish groups (Table 2b). A reasonable P/C ratiofor top predatory fish such as the piscivorous fish group is 0.1(Christensen et al., 2005; Link et al., 2006, 2008). Reasonable P/C ratiosfor benthic feeding and planktivorous fish is 0.15 (Christensen et al.,

d are taken from the original Monaco and Ulanowicz (1997) model with the exception ofophic level (TL), biomass (g DWm−²) (B), production/biomass (P/B) (g DWm−²year−1),ble 3 for prey/predator group abbreviations; detritus (Det). Numbers in parenthesis were

(b) Parameter inputs

11 16 TL B P/B C/B EE P/C

– 0 2.92 0.028a 0.17 3.4 0 0.05– 0 4.34 0.057a 0.471 4.71 0.45 0.1– 0 3.41 0.23a 0.595 3.97 0.198 0.15– 0 3.31 0.27a 2.643 17.62 0.391 0.15– 0 2.33 0.23b 3.1 15.5 0.123 0.2– 0.003 2.33 2.66c 4.14 20.7 0.217 0.2– 0.01 2.47 10.64c 3.87 19.35 0.51 0.2– 0.003 2.25 4.95c 4.35 21.75 0.95 0.2– – 2.56 0.52d 20 (22.794) 60 0.881 0.333– 0.022 2.31 1.733d 30 90 (65.61) 0.202 0.333– 0.317 2 27.5e 49.5 123.75 0.085 0.4– 0.147 1 27.58f 33 – 0.023 –

– 0.298 1 50.76g 35.88 – 0.011 –

– 0.101 1 16h 42 – 0.095 –

– 0.097 1 0.49i 244.9 – 0.629 –

1 – 1 60j – – 0.651 –

horne-Miller et al., 1983; f Thorne-Miller and Harlin, 1984; gThornber et al. unpublished., 2003; iURI Watershed Watch Program; jPauly et al., 1993.

91C. Byron et al. / Aquaculture 314 (2011) 87–99

2005; Link et al., 2006, 2008) meaning fish consume 6.7 to 10 timeswhat they produce. C/B values were averaged for each group (Benthic,Planktivorous, Piscivorous) based on individual species C/B valuescalculated by Fishbase.org using the default settings (Oct 2009). Dietcomposition of piscivorous, benthic and planktivorous fish wasdetermined from the NEFSC bottom trawl survey of Southern NewEngland (Azarovitz, 1981) (Table 2a).

Biomasses were estimated based on infauna surveys conducted onJuly 2006 at four sites in Narragansett Bay (Schult, 2010) anddiagnostic testing (Table 2b, Fig. 1). Infauna data were not availabledirectly from the lagoons and Narragansett Bay had the most similarinfauna and closest in proximity to the lagoons. Clam harvests wereestimated from the 2008 Standard Atlantic Fisheries InformationSystem (SAFIS). The biomass of cultured oysters was estimated usingplanting and harvest estimations from the farmers and usingallometric conversions derived from a 2008 survey conducted oncultured oysters in the lagoons (Markey, 2009). The harvest biomassof farmed cultured oysters was calculated from the number of oysterssold from the lagoons in 2008 (Beutel, 2009).

Biomass of zooplankton in estuaries varies widely and can rangefrom 1 to over 100 mg dry weight m−3 (Mann, 2000). Sincezooplankton were not measured in the lagoons, biomass wasestimated from neighboring Narragansett Bay (Durbin and Durbin,1981) and then decreased according to diagnostic tests (Fig. 1) toreflect lower zooplankton biomasses typical in lagoons with narrowocean inlets. Meroplankton was approximately 20% of the totalzooplankton (Durbin and Durbin, 1981).

Primary producers comprise three groups; eelgrass,macroalgae, andphytoplankton. Eelgrass estimates were adjusted from 1979 summer(Jul–Aug) surveys conducted in the lagoons (Ninigret, Greenhill, PointJudith, Potter) (Thorne-Miller et al., 1983) to reflect current year-roundconditions using sonar mapping in 2007 (MapCoast). Macroalgaebiomass was surveyed in the summer (May–Aug) of 2008 in PottersPond (Thornber et al., unpublished data). Macroalgae productionwas examined in the summer (Jul–Aug) of 1979 in Ninigret Pond(Thorne-Miller et al., 1983) and increased to reflect current conditionsas indicated by diagnostic testing and anecdotal observations.Phytoplankton biomass was calculated from chlorophyll fluorescence(chl-a) which was measured in highly flushed lagoons in 2006(Winnipaug, Greenhill, Ninigret, Quonochontaug Ponds) and in 2007(Point Judith Pond) from May to October. This warm weatherestimate was then weighted against a year-round survey (2008–2009) conducted in Narragansett Bay (http://www.narrbay.org) toobtain a chl-a value characteristic of a yearly average in the lagoons.Generally, the average phytoplankton productivity in an estuarinesystem is about 255 g C m−2 year−1 and above 350 g C m−2 year−1 inaneutrophied estuary (Mann, 2000).

Typical benthic bacteria abundances in estuarine systems range from5 to 50 g dry weight m−2 (Mann, 2000). Benthic bacteria diets wereassumed to consist solely of detritus (Table 2a). Macrodetritus wasmeasured in Potters Pond in 1980 and had a biomass of 20–100 g dryweight m−2 (Nowicki and Nixon, 1985). Further details on data sourcesand model parameterization were reported by Byron (2010).

2.5. Diagnostics

Prior to balancing the model, additional non-Ecopath deriveddiagnostics were performed to evaluate the validity of the data.Diagnostics check and aid in balancing the model independent ofEcopath assumptions and prior to the mandatory Ecopath automatedbalancing routine (Fig. 1). Diagnostic tests allow evaluation of thecohesiveness of the data despite the natural discrepancies that occurwhen using myriad data sources measured across varying scales.Prebalancing diagnostics allow the examination of the system holisti-cally instead of piecemeal in the way the individual data sources werecollected.

Diagnostic testswere performed on: biomass, the ratio of biomass toprimary production, the vital rates [production (P), consumption (C),respiration(R)], the ratios of vital rates (e.g. P/C) and the totalconsumptive removals. For detailed description on the diagnostic testssee Link (2010). In brief, each functional group was plotted alongthe x-axis in order of decreasing trophic level to allow easyvisualization of trophic relationships (Link, 2010).

• Biomass and vital rates should decrease with trophic level.• Ratios of biomass to primary production should remain less thanone.

• Vital rate ratios should be within ecologically acceptable limits(i.e. P/Cb0.5).

• Within each functional group, consumption should be higher thanproduction.

• Total consumptive removals should be lower than production.

When simple ecological and physiological ‘rules’ were not met asevident from these diagnostic tests, parameters were corrected forimproved ecological integrity and validity. The parameter in questionwas increased or decreased by a multiple of 0.2, 0.5, 1.333, 2, 3, 10, or15 until the parameter conformed to the expectations of thediagnostic test (Fig. 1). Performing diagnostic tests allowed greatercontrol over the mass-balancing of the model by manually bringingthe parameters closer to mass-balance instead of relying solely on theautomated Ecopath mass-balance routine.

In addition to diagnostic tests, Ecopath is equipped with tools thatcan be used to address uncertainty in the data, thereby furtherimproving the quality of the parameter inputs through the mass-balancing routine. The Pedigree routine allows for the entry of a rangefor each parameter input which then evaluates statistical uncertainty.Pedigree allows the user to mark the sources of data for each parameterand has a ranking scheme for the ‘goodness’ of that data giving eachparameter confidence intervals. These confidence intervals were thenused by the Ecorangermodule to give a probability distribution for eachparameter using a Monte-Carlo parametric (Christensen et al., 2005).Additionally, a Sensitivity Analysis was performed to evaluate the effectof each of the entered input parameters on all of the ‘missing’ EEparameters for each group in the system by varying each inputparameter from −50% to +50% (Christensen et al., 2005). Diagnostictests, Pedigree Analysis, and Sensitivity Analysis were all performed toevaluate the input parameters.

2.6. Ecopath outputs

Ecopathalsoproduces several outputs thatwereused to characterizethe system. Similar to the Sensitivity Analysis, theMixed Trophic ImpactAnalysis was used to evaluatewhich functional groupsweremost likelyto be impacted by slight perturbations by measuring the impact of aninfinitesimally small increase in group biomass on other groups. MixedTrophic Impact is themeasure of direct or indirect influence a group (onthe left of the matrix) has on another group (at the top of the matrix)based on food web characteristics. Ecopath produces a matrix thatreports the measured direct and indirect increase or decrease in everygroup's biomass parameter caused by an infinitesimally small increasein every other group's biomass.

Ecopath Summary Statistics provide informative systems measuressuch as throughput, cycling index, and number of pathways that can beused to characterize the ecosystem. Total system throughput is the sumof all flows in a system: consumption, export, respiration, and flows todetritus (Christensen et al., 2005). It represents the size of the systemexpressed as a flow (Ulanowicz, 1986). Cycling index is the percentageof energy throughput in the system that is recycled and correlates tosystemmaturity, resilience and stability (Christensen et al., 2005; Finn,1976; Odum, 1969). The number of pathways is indicative of theredundancy and stability of the system.

Fig. 1. Pre-balancing diagnostics depicting B, P/B, and C/B values used as Ecopath input parameters prior to automated mass-balancing. See Tables 2 and 3 for group name abbreviations. In some cases, groups were summed; all primaryproducer groups (PP), all zooplankton groups (Zoo). (a) Trophic decline in biomass of species groups. (b) Species group biomass and production ratios to total primary production showing all levels remained at or below one. (c) Vital ratesshowing relative decline with trophic level and higher consumption than production and respiration. (d) Vital rate ratios. (e) P/C was less than P/R. (f) Total consumptive removals which were less than production and consumption.

92C.Byron

etal./

Aquaculture

314(2011)

87–99

Table 3Ecopath Pedigree Index for the lagoon model. Higher numbers signify greaterconfidence.

Group B P/B Q/B Diet Catch

Birds (Bir) 4 3 2 1 —

Piscivorous fish (PiF) 4 3 3 4 —

Benthic feeding fish (BFF) 4 3 3 4 —

Plantivorous fish (PkF) 4 3 3 4 —

Cultured oysters (Coy) 4 2 2 2 5Filter feeders (FF) 4 2 2 2 5Meroplankton (Mer) 4 2 2 2 —

Holoplankton (Hol) 4 2 2 2 —

Epibenthic invertebrates (EI) 4 2 2 2 —

Deposit feeders (DF) 4 2 2 2 —

Benthic bacteria (BB) 4 3 2 2 —

Eelgrass (Eel) 4 5 — — —

Macroalgae (Mac) 4 5 — — —

Benthic microalgae (BM) 4 5 — — —

Phytoplankton (Phy) 5 7 — — —

Ecopath Pedigree Index: 0.485.Number of living groups: 15.Measure of fit: 2.00.

93C. Byron et al. / Aquaculture 314 (2011) 87–99

2.7. Carrying capacity

Ecological carrying capacity was calculated following the methodsof Jiang and Gibbs (2005). Cultured oyster biomass and proportionalcultured oyster harvest were increased in consecutive static modelsuntil the system no longer represented its present state. Culturedoyster harvests are currently, and have historically been, 38% ofstanding stock biomass and were expected to increase in the sameproportion total biomass (Pietros and Rice, 2003). Specifically, whenthe model was no longer balanced and ecotrophic efficiency value(s)were pushed to 1.00, the system was said to have exceeded itsecological carrying capacity for cultured shellfish. The point just priorto this change, at which the model was still balanced and EEs wereless than 1.00, was reported as the ecological carrying capacity.

Zooplankton are major competitors with oysters for food (Jiangand Gibbs, 2005). Removing zooplankton from the system wouldallow greater food source for oysters thereby maximizing oysterproduction. Production carrying capacity was calculated by removingall zooplankton groups from the model, rebalancing the model usingcurrent levels of shellfish biomass and harvest, and then iterativelyincreasing cultured shellfish biomass and proportional culturedshellfish harvest until the model became unbalanced. The biomassvalue just prior to causing the zooplankton-less model to becomeunbalanced while restricting EEs to less than one was the productioncarrying capacity. Because zooplankton groups were removed fromthe diet matrix, effectively thatmeant that the proportion of each preyitem consumed by zooplankton predators needed to be redistributedso that all prey items consumed by a given predator would still add to100% of their diet. The percentage of zooplankton in each predator'sdiet was evenly distributed among all other prey groups prior tobalancing the model.

Robustness of biomass parameters while the system was atecological carrying capacity was examined by changing individualbiomass values of each group by factors of 0.01, 0.5, 2, 10 and 100 oftheir original value when the model was balanced at carryingcapacity. Only one value was changed at a time while all otherbiomass values remained constant. The degree to which the biomassvalue could vary while the model remained balanced was an indicatorof the robustness of that functional group at carrying capacity againstperturbations. Additionally, balanced biomass parameters at ecolog-ical carrying capacity were then rebalanced using EE as primaryconstraint (b0.95) under varying cultured oyster biomass scenarios:0.001, 0.01, 0.1, 10, 100, and 1,000 times the ecological carryingcapacity biomass (i.e. Link et al., 2009). This scenario testing allowed avisual representation of the influence of different levels of oysterculture on other types of species in the system. Finally, the ecologicalcarrying capacity of cultured oysters in the lagoons was compared tofilter feeder biomasses in Ecopath models of other similar systems.

2.8. Model considerations

Ecopath mass-balances the model by slightly adjusting the inputparameters within their confidence limits according to two masterequations so that the energy input and output are equal for eachgroup (Christensen et al., 2005). Energy flows between groups arelinked through the diet matrix. To avoid the primary error in Ecopathmodeling – uncritical use of default settings – some default settingswere changed to more accurately represent each species group.Unassimilated consumption value of 0.2 was used for every groupexcept the two zooplankton groups, in which case 0.4 was used toavoid underestimation of egestion.

Ecopath provides a range of setup options for the automatedmass-balancing routine. All modeling was done using EwE5 softwarepackage (available at www.ecopath.com). Five-thousand (5000)iterations per run were chosen and the EE goal was forced to 0.95.The confidence intervals were allowed to be set by the Pedigree with a

lower limit set at 10% of the original. For “perturbation method”,neighborhood perturbations of 10%B, 20%DC (diet composition), 10%P/B and 10%C/B were chosen. The initial conditions for each run wereset to continue with B and DC values at end of the last run. Finally,both “reduce sum of excess EE” and “reducemax EE”were selected forthe iteration design logic. The model was considered balanced whenall EEs were below 1.0, all P/C values were below 0.5, and there was nonegative respiration.

3. Results

3.1. Parameter quality

The quality of data sources is quite high based on the calculatedPedigree Index of 0.48 (Morissette, 2006) (Table 3). Biomassparameters contributed the most to the high Pedigree Index valueas they had higher confidence than vital rate parameters (P/B and C/B)and diet estimates (Table 3). Most data for biomass estimates camefrom surveys conducted in the lagoons or neighboring NarragansettBay. Vital rates and diet composition values came from a combinationof surveys and values reported in other models. Pedigree confidencewas lowered in cases where no recent information was available fromRhode Island waters or the prebalancing diagnostics varied greatlyfrom the data-derived input value.

According to the Sensitivity Analysis, the B and C/B inputparameters of the Epibenthic Invertebrates group have the greatestimpact on C/B parameters of Eelgrass and Macroalgae groups and theEE of the Deposit Feeders group. Piscivorous Fish C/B has an impact onthe EE of the same group. Cultured oyster inputs (B and P/B) onlyinfluence parameters (EE) of their own group. These results were thesame at both the current biomass and at the ecological carryingcapacity biomass of cultured oysters.

3.2. Summary statistics

The most striking observation from the Mixed Trophic ImpactAnalysis was the lack of impact the cultured oysters group had on anyother group (Fig. 2). The relatively low biomass of cultured oysters(Table 2b, Fig. 1a) did not have any quantitative impact on the othergroups. An increase in biomass of a group resulted in a negativeimpact on that same group for all groups. There were particularlylarge impacts observed for piscivorous fish, filter feeders, benthicbacteria, deposit feeders, the harvest of clams (“Harvesters”) and theextraction of farmed oysters (“Farmers”). An infinitesimal increase in

Fig. 2.Mixed trophic impact index of balancedmodel. An infinitesimally small increase in biomass of the groups on the left column impacts the biomasses of the groups across the toprow. A bar extending downwards symbolizes a negative impact and an upwards rising bar symbolizes a positive impact. Catch is noted for wild clams (Harvesters) and oysters(Farmers).

94 C. Byron et al. / Aquaculture 314 (2011) 87–99

filter feeder biomass had the largest positive indirect impact on theextracted biomass of cultured oysters (“Farmers”).

Reverse trophic cascade effects on deposit feeders and planktivorousfish are evident by a large positive indirect response from aninfinitesimal increase in phytoplankton and deposit feeder biomasses,respectively (Fig. 2). An infinitesimal increase in detritus had a largepositive direct impact on benthic bacteria, holoplankton, and epibenthicinvertebrates and a negative impact on benthic microalgae.

RI lagoons were greatly heterotrophic with a total primaryproduction / total respiration (P/R) ratio of 2.2. Transfer efficiencyfor the lagoons was 5%. Piscivorous fish were at the top of the foodweb (TL 4.34) followed by benthic feeding fish (TL 3.41) andplanktivorous fish (TL 3.31) (Fig. 3). As expected, higher trophiclevels had a higher number of pathways with which energy couldtravel up the food web. Piscivorous, benthic, and planktivorous fishgroups had some of the highest number of pathways (83, 46 and 19,

Fig. 3. Ecopath output of the trophic structure of Rhode Island Lagoons, USA.

respectively). Birds (TL 2.92) did not have the highest trophic level asexpected, but did have a high number of pathways (44). The cyclingindex calculated by this model was 25.66%. Over 40% of the energythroughput was diverted to detritus (Table 4). Low trophic levelscontributed most to detritus and also had the highest throughput andexport (Table 4). Seventy one percent (71%) of the total energy floworiginated from detritus. The total system biomass, excluding detrituswas 144 g dry weight m−2 and the net system production was1912 g dry weight m−2year−1. The complete suite of Ecopath outputswere reported by Byron (2010).

3.3. Carrying capacity

Cultured oyster biomass is currently at 0.233 g DWm−2 and couldincrease 62 times this value without exceeding the ecological carryingcapacity to 14.5 g DWm−2 (Table 5). Assuming only 2% dry meatweight from of a whole live animal (Rheault, unpublished data), thattranslated to 11.65 t km−2 live weight of cultured oysters currently inthe system and a potential of 722 t km−2at ecological carryingcapacity. Current cultured oyster harvest was 38% of total biomasswhich equated to 274.5 t km−2 at ecological carrying capacity. Atecological carrying capacity, the smallest lagoon, Potters, was capableof having a total of 962 t cultured shellfish biomass which equated to269 t cultured shellfish harvest annually (Table 6). The largest lagoon,Ninigret, could have 5001 t biomass or 1400 t cultured shellfishharvest annually at ecological carrying capacity (Table 6).

Upon surpassing ecological carrying capacity in these lagoons,phytoplankton were overgrazed in the model which resulted in an EEequal to or greater than one (Table 5). Ecotrophic efficiency was theconstraining parameter indicatingmodel imbalance. Meroplankton andholoplankton constrained the model, meaning that at ecologicalcarrying capacity even a small change inmeroplankton or holoplanktonbiomasses caused the model to become unbalanced (Tables 5 and 7).Primary producers were more robust as the model remained balancedeven when any one of the primary producer groups' biomassesincreased or decreased 10-fold (Table 7). An exception was thatphytoplankton biomass could not be reduced without unbalancing the

Table 4Energy flows (g DW m−2 year−1) by trophic level or all groups combined for the lagoon model.

Trophic level (TL)\flow Imported Consumed Exported Flows to Detritus Respired Throughput

VII 0 0 0 0 0VI 0 0 0.002 0.007 0.009V 0.009 0.005 0.123 0.25 0.387IV 0.387 0.046 5.149 5.789 11.372III 11.372 0.519 58.629 73.085 143.604II 143.604 1.332 2069.559 1532.393 3746.888I 0 3746.888 1909.918 3342.885 0 8999.691Sum 0 3902.261 1911.821 5476.348 1611.523 12901.95Percent of total throughput 30% 15% 42% 13%

95C. Byron et al. / Aquaculture 314 (2011) 87–99

model. The model also remained balanced when one biomass of thehigher trophic level consumer groups was halved or doubled (Table 7).

Production carrying capacity of the lagoons was calculated to be31.2 g DWm−2 or 1561 t live weight km−2. Exceeding productioncarrying capacity resulted in phytoplankton ecotrophic efficiencyexceeding one (Table 5). The lagoons can support this high level ofbiomass production across 46% of surface area before exceeding theecological carrying capacity.

Upon surpassing the production carrying capacity in the model,phytoplankton EE increased to unrealistic levels. The model could notwithstand an increase of piscivorous fish, cultured oysters, filterfeeders, or epibenthic invertebrates without it becoming unbalanced(Table 7). Decreases in planktivorous fish, deposit feeders andphytoplankton also unbalanced the model at production carryingcapacity (Table 7). However, the production carrying capacity modelwas extremely robust in other groups withstanding a 10-fold increaseand decrease in birds, eelgrass, or detritus and a 2-fold increase ordecrease of most other functional groups (Table 7).

Changing the oyster biomass resulted in minor changes in modeloutputs of other groups. Scenarios of oyster biomass two to threeorders of magnitude above the ecological carrying capacity resulted inperturbations of less than one order of magnitude in other groups(Fig. 4). All groups were affected at the 1000× ecological carrying

Table 5Changes in Ecopath lagoon model during estimationsof carrying capacity using theautomated mass-balancing routine. The bold-type numbers are the calculated ecologicalcarrying capacity and production carrying capacity biomass values for cultured oysters.Any biomass value below the carrying capacity did not affect the balanced model. Anybiomass value above the carrying capacity unbalanced the model by unrealisticallyincreasing the ecotrophic efficiency (EE) of another group to one (1) or higher signifyingthat the entire group was consumed.

Multiplier B Catch Mass-balance changes in model

1 (Current condition) 0.233 0.0886

Ecological carrying capacity2 0.466 0.1772 Balances50 11.65 4.43 Balances60 13.98 5.316 Balances61 14.213 5.4046 Balances62 14.446 5.4932 Balances63 14.679 5.5818 Phytoplankton EE=1.00365 15.145 5.759 Phytoplankton EE=1.015100 23.3 8.86 Phytoplankton EE=1.225

Production carrying capacity125 29.125 11.075 Balances130 30.29 11.518 Balances133 30.989 11.7838 Balances134 31.222 11.8724 Balances135 31.455 11.961 Phytoplankton EE=1.005150 34.95 13.29 Phytoplankton EE=1.105200 46.6 17.72 Phytoplankton EE=1.440500 116.5 44.3 Phytoplankton EE=3.4461000 233 88.6 Holoplankton EE=7.140, Benthic

Microalgae EE=1.169,Phytoplankton EE=6.642, DetritusEE=1.080

capacity scenario where benthic and zooplankton groups decreasedslightly and detritus, primary producers, and fish groups increasedslightly. The biggest impact was to benthic invertebrates suggestingcompetitive interaction with the oysters for food resources.

4. Discussion

Despite the rapid increase in oyster aquaculture, total biomass andimpact to the ecosystem is still quite negligible. Oyster aquaculturehas the potential to continue on its current trend of increase (Alves,2007; Beutel, 2009) without ecologically altering the lagoon system.The mixed trophic impact analysis shows little impact of oysters onother groups at current levels and the scenario tests show onlymodest biomass changes in other key groups when the model isrebalanced at 100–1000× the ecological carrying capacity limit.Although bottom-up indirect effects are possible (Steele, 2009; Steeleet al., 2007), they are practically undetectable due to the relatively lowcultured oyster biomass (Table 5, Fig. 2). Those groups that were mostlikely to be impacted by high cultured oyster biomass were lowtrophic levels, specifically plankton groups (Tables 5 and 7). Highconcentration of bivalves in small systems with rapid water massturnover rates, such as these lagoons, is not uncommon (Dame andPrins, 1998).

Cultured oysters are in direct competition with zooplankton forfood (Gerritsen et al., 2002). Under current conditions, zooplanktonare the dominant grazers of primary production. However, if culturedoysters were to become the dominant grazers on phytoplankton,higher trophic levels could perish or move to other locations wherefood was more abundant. This scenario has been modeled in othersystems (Jiang and Gibbs, 2005) and although possible in RI lagoons, isunlikely. The high throughput of energy to detritus and low efficiencyof the system suggests that much energy is going unutilized and isbeing stored in the detrital pool. Bivalves feed on both phytoplanktonand detritus and the relative contribution of each to bivalve energybudgets varies widely depending on a number of factors. In somesystems, bivalves remove four times as much detritus as they doprimary production (Ferreira et al., 2007). Increasing the biomass ofcultured bivalves may improve efficiency of energy cycling byincreasing the benthic pelagic link by filtering material out of thewater column and repackaging it so it is biologically available forplants and benthic consumers (Newell, 2004; Peterson and Heck,1999; 2001).

Table 6Lagoon surface area, cultured oyster total biomass and harvest (28% of total biomass)biomass at ecological carrying capacity.

Lagoon Surface areaa (km2) Total biomass (t) Harvest biomass (t)

Point Judith 6.37 4601 1288Potter 1.33 962 269Ninigret 6.92 5001 1400Quonochontaug 2.93 2133 592Winnipaug 1.89 1304 365

a Lee, 1980.

Table 7Robustness of biomass (B) values at ecological carrying capacity and production carrying capacity of the lagoons. Biomass values of each groupwere multiplied by factors of 0.01, 0.1,0.5, 2, 10 and 100. Only one valuewas changed at a timewhile all other biomass values remained constant. Values in italics indicate the factor at which themodel became unbalancedand represents low robustness. Values in bold in shaded boxes signify that the model remained balanced representing high robustness in those groups at carrying capacity. SeeTables 2 and 3 for group name abbreviations.

Ecological Carrying Capacity Group Name Production Carrying Capacity

0.01 0.1 0.5 Biomass (gDWm−²) (gDWm−²)

2 10 100 0.01 0.1 0.5 Biomass 2 10 100

0.00028 0.0028 0.014 0.028 0.056 0.28 2.8 Bir 0.00028 0.0028 0.014 0.028 0.056 0.28 02.80.00057 0.0057 0.0285 0.057 0.114 0.57 5.7 PiF 0.00057 0.0057 0.0285 0.057 0.114 0.57 05.70.00232 0.0232 0.116 0.232 0.464 2.32 23.2 BFF 0.00232 0.0232 0.116 0.232 0.464 2.32 23.20.00265 0.0265 0.1325 0.265 0.53 2.65 26.5 PkF 0.00344 0.0344 0.172 0.344 0.688 3.44 34.40.14446 1.4446 7.223 14.446 28.892 144.46 1444.6 Coy 0.31222 3.1222 15.611 31.222 62.444 312.22 3122.20.02659 0.2659 1.3295 2.659 5.318 26.59 265.9 FF 0.02659 0.2659 1.3295 2.659 5.318 26.59 265.90.10641 1.0641 5.3205 10.641 21.282 106.41 1064.1 EI 0.10641 1.0641 5.3205 10.641 21.282 106.41 1064.10.0495 0.495 2.475 4.95 9.9 49.5 495 DF 0.0495 0.495 2.475 4.95 9.9 49.5 4950.00517 0.0517 0.2585 0.517 1.034 5.17 51.7 Mer0.01733 0.1733 0.8665 1.733 3.466 17.33 173.3 Hol0.275 2.75 13.75 27.5 55 275 2750 BB 0.275 2.75 13.75 27.5 55 275 27500.27578 2.7578 13.789 27.578 55.156 275.78 27578 Eel 0.27578 2.7578 13.789 27.578 55.156 275.78 2757.80.5076 5.076 25.38 50.76 101.52 507.6 5076 Mac 0.5076 5.076 25.38 50.76 101.52 507.6 50760.16 1.6 8 16 32 160 1600 BM 0.16 1.6 8 16 32 160 16000.0049 0.049 0.245 0.49 0.98 4.9 49 Phy 0.0049 0.049 0.245 0.49 0.98 4.9 490.6 6 30 60 120 600 6000 0.6 6 30 60 120 600 6000Det

96 C. Byron et al. / Aquaculture 314 (2011) 87–99

The high throughput in this model indicates that the lagoons aresystems with high overall activity. However, these systems are notvery efficient in transferring energy between trophic levels. The lowtransfer efficiency of the lagoons (5%) was still within reasonablerange. Typical estuarine environments have transfer efficiencies of 5–20% with an average of 10% (sensu Odum; Zajak et al., 2009). Below-average transfer efficiency suggests that the lagoons have a generalist-based food web structure. Low efficiency may be attributed to highnutrient inputs compared to oligotrophic systems with higherefficiencies (Shannon et al., 2003). High nutrient inputs support thehigh ecological and production carrying capacity estimates.

High ecological carrying capacity has strong implications formanagement. Ecosystem models, such as this one, are useful toolsfor management (Héral, 1993). It should be stressed that modeledcarrying capacity is a theoretical limit and should be viewed withcaution. Perturbations only slightly above the ecological carryingcapacity do have implications on the system (Table 5). Mostimmediately, phytoplankton is consumed as indicated by increase inEE at oyster levels above carrying capacity (Table 5). Natural andanthropogenic perturbations may also influence the carrying capacity

Fig. 4. Changes in biomass of key groups under different scenarios of cultured oysters.

but are not examined here due to limitations of static models. Also,ecological models do not account for myriad uses of these ponds orother societal restrictions. In other words, there simply may not beenough available space to accommodate aquaculture at its ecologicalcarrying capacity limit. Following the Ecological Approach toAquaculture, it is strongly advised that managers take precautionand limit aquaculture well below the modeled carrying capacity limitstated here.

Using a static model to calculate a dynamic process, such as carryingcapacity warrants caution when making management decisions.Temporal and spatial scales at which populations are measured aswell as climatic variability contribute to the variability of carryingcapacity. Climate change in coastal zones, predicted to be characterizedby increase temperature, possible decreases in salinity in the coastalzone, and decrease in pH (Anthony et al., 2009; IPCC, 2007), will greatlyinfluence the vital rates of bivalves (Davis, 1958; Miller et al., 2009;Motes et al., 1998) and subsequently their carrying capacity. Asuggested precautionary approach is to limit aquaculture to half thecalculated carrying capacity value. Half of carrying capacity is themaximum sustainable yield (MSY) where growth rate is high and is anaccepted management target for many fin fisheries (Mace, 2001).Managing at MSY will also allow for the inherently dynamic variabilityof carrying capacity that is not described by static models.

Aquaculture and coastal managers should consider carefully whatindicators tomonitor as to remainunder the ecological carrying capacitylimits. The RI lagoons are currently managed based on percent surfacearea (CRMC, 2008). Currently aquaculture is restricted to 5% of thesurfacewaters of eachwater body.Managing aquaculture-based surfacearea restrictions is dangerous due to varying levels of oyster biomassproduction. Productionbiomass calculatedbyEcopath suggests that 46%of the surface area of the lagoons can be farmedwhile remainingwithinthe ecological carrying capacity limit. However, the production biomasscalculated by Ecopath is high and unrealistic given farming techniques,equipment, and regulations in place today. A more conservativeestimate of production biomass reported by a bottom culture farmerin Point Judith Pond (1121 t km−2; Rheault, 2008) suggests up to 64% ofthe surface area can be farmed if farming techniques and equipmentremain the same. To ensure that farmers were remaining under theecological carrying capacity, managers could monitor one parameter —total biomass per unit area which is easily inferred from stocking andsales receipts.

97C. Byron et al. / Aquaculture 314 (2011) 87–99

Diagnostic testing was a particularly critical step in the assembly ofthis model on account of disparate data. Many of the parameters wereinferred from data measured in other systems with similar qualities.The spatial and temporal variation associated with using data from abreadth of resources can undoubtedly lead to inconsistencies inparameters that violate basic ecological principles. Diagnostic testingwas a way to flag and troubleshoot those ecological violations early inthe parameterizing process and avoid compounding the problem insubsequent steps of model preparation.

No other lagoon systems have been evaluated for the carryingcapacity of bivalve aquaculture using this methodology. However,Ecopath was used to calculate ecological carrying capacity in a largebay system in New Zealand (Jiang and Gibbs, 2005). RI lagoons had anecological carrying capacity that is an order of magnitude higher thancomparatively oligotrophic New Zealand bays (65 t km−2) (Jiang andGibbs, 2005). The ecological carrying capacity for bivalve culture in RIlagoons is more comparable to current suspension feeder biomass inLong Island Sound (Zajak et al., 2009). Estuarine systems are bettersuited for bivalve culture and tend to have higher bivalve carryingcapacities than more oligotrophic systems. The eutrophic Narragan-sett Bay located adjacent to the lagoons had less than half theecological carrying capacity of the RI lagoons (Byron et al., submittedfor publication) demonstrating that even among eutrophic systems, RIlagoons have a high ecological carrying capacity.

The expansion of aquaculture in many places worldwide coupledwith renewed political interest in some of these places makesaquaculture a coastal use of particular concern (FAO, 2009; NRC,2010; Soto, 2010). Of all coastal systems, barrier-beach lagoons maybe the most susceptible to anthropogenic and ecological impacts. Themethodology presented here is easily transferable to other denselypopulated systems affected by user conflict issues and can be appliedto any fisheries or aquaculture related industry in any marine oraquatic environment. Mass-balance modeling using Ecopath allowsthe modeler to dictate ‘acceptability’ limits for a particular systemwhen calculating ecological carrying capacity. Pre-balancing diagnos-tic tests aid in parameter preparation and promote ease balancing themodel. Myriad Ecopath outputs, together with the calculated carryingcapacity, provide managers with the necessary information toimplement an Ecological Approach to Aquaculture.

Acknowledgements

Robert Rheault, David Beutel and David Alves provided valuableassistance. This work was possible through the cooperation andsupport of all the agencies and labs that shared data used toparameterize the model. This work was funded by the NOAA NationalMarine Aquaculture Initiative #NA08OAR4170838, NSF IGERT #DGE-0504103, and a John Wald Science Grant.

References

Alves, D., 2007. Aquaculture in Rhode Island. 2007 Yearly Status Report, CoastalResource Management Council. Available at: http://www.crmc.state.ri.us/pubs/pdfs/aquareport07.pdf.

Anthony, A., Atwood, J., August, P., Byron, C., Cobb, S., Foster, C., Fry, C., Gold, A., Hagos,K., Heffner, L., Kellogg, D.Q., Lellis-Dibble, K., Opaluch, J., Oviatt, C., Pfeiffer- Herbert,A., Rohr, N., Smith, L., Smythe, T., Swift, J., Vinhateiro, N., 2009. Coastal lagoons andclimate change: ecological and social ramifications in U.S. Atlantic and Gulf coastecosystems. Ecol. Soc. 14, 8.

Azarovitz, T.R., 1981. A brief historical review of the Woods Hole Laboratory trawlsurvey time series. In: Doubleday, W.B., Rivard, D. (Eds.), Bottom trawl surveys:Can. Spec. Publ. Fish. Aquat. Sci., 58, pp. 62–67.

Bacher, C., Duarte, P., Ferreira, J.G., Héral, M., Raillard, O., 1998. Assessment andcomparison of the Marennes-Oléron Bay (France) and Carlingford Lough (Ireland)carrying capacity with ecosystem models. Aquat. Ecol. 31, 379–394.

Beutel, D., 2009. Aquaculture in Rhode Island. 2009 Annual Status Report. CoastalResources Management Council. Available at: http://www.crmc.ri.gov/aquaculture/aquareport09.pdf.

Brown, C.L.J., 1962. On the Ecology of Aufwuchs of Zostera marina in Charelstown PondRhode Island, Biol. Oceanogr. University of Rhode Island.

Burnett, M., 2007. Assessment of recreationally important finfish stocks in Rhode Islandcoastal ponds. Young-of-the-Year Survey in RI Coastal Ponds. Rhode IslandDepartment of Environmental Management Division of Fish and Wildlife MarineFisheries Laboratory.

Byron. 2010. Carrying capacity of bivalve aquaculture. Ph.D. Thesis. Department ofFisheries Animal and Veterinary Sciences, University of Rhode Island.

Byron, C., Bengtson, D., Costa-Pierce, B., Calanni, J., 2011. Integrating science intomanagement: ecological carrying capacity of bivalve shellfish aquaculture. Mar.Policy 35, 363–370.

Byron, C., Link, J., Costa-Pierce, B., Bengtson, D., submitted for publication. Calculatingecological carrying capacity of shellfish aquaculture using mass-balance modeling:Narragansett Bay, Rhode Island, USA. Ecol. Model.

Carver, C.E.A., Mallet, A.L., 1990. Estimating the carrying capacity of a coastal inlet formussel culture. Aquaculture 88, 39–53.

Christensen, V., 1995. Ecosystem maturity — towards quantification. Ecol. Model. 77,3–32.

Christensen, V., Pauly, D., 1993. Trophic models of aquatic ecosystems. InternationalCenter for Living Aquatic Resources Management (ICLARM); Copenhagen,Denmark: International Council for the Exploration of the Sea; Copenhagen,Denmark: Danish International Development Agency, 390 pp.

Christensen, V., Walters, C.J., Pauly, D., 2005. Ecopath with Ecosim: a user's guide,Fisheries CentreNovember 2005 edition. University of British Columbia, Vancouver.154 pp.

Christensen, V., Beattie, A., Buchanan, C., Ma, H., Martell, S.J.D., Latour, R.J., Preikshot, D.,Sigrist, M., Uphoff, J.H., Walters, C.J., Wood, R.J., Townsend, H., 2009. Fisheriesecosystem model of the Chesapeake Bay: methodology, parameterization andmodel exploration. NOAA Tech Memo. NMFS-F/SPO-106.

Coastal Resources Management Council (CRMC), 1999. Rhode Island's salt pond region:a Special Area Management Plan (SAMP)Available at: http://www.crmc.ri.gov/samp_sp.html.

Coastal Resources Management Council (CRMC), 2008. The State of Rhode IslandCoastal Resources Management Program as amended. Coastal Resources Manage-ment Council. Available at http://www.crmc.ri.gov/regulations/RICRMP.pdf.

Conover, R.J., 1961. A study of Charlestown and Green Hill ponds, Rhode Island. Ecology42, 119–140.

Costa-Pierce, B., 2008. Epilogue: aquaculture, innovation and social transformation. In:Culver, K., Castle, D. (Eds.), The International Library of Environmental, Agriculturaland Food Ethics. 238 pp.

Costa-Pierce, B.A., Jena, J.K., Kaushik, S.J., Hashim, R., Yakupitiyage, A., Rana, K., Hinshaw,J., Lemos, A.J., Hernandez, A.J., Bueno, P., Rutaisire, J., Greenhalgh, F., 2010.Responsible use of resources for sustainable aquaculture. Global Conference onAquaculture 2010. Food and Agriculture Organization of the United Nations (FAO,Rome, Italy, Phuket, Thailand.

Crawford, R.E., 1985. Retention of winter flounder larvae within a Rhode Island saltpond. Estuaries 8, 217–227.

Cronan, J.M., Halla, B.F., 1968. Fall and winter foods of Rhode Island waterfowl.Department of Natural Resources Division of Conservation. 25 pp.

Dame, R.F., Prins, T.C., 1998. Bivalve carrying capacity in coastal ecosystems. Aquat. Ecol.21, 409–421.

Davis, H.C., 1958. Survival and growth of clam and oyster larvae at different salinities.Biol. Bull. 114, 296–307.

DeAngelis, B., Hancock, B., Kocot, M., Turek, J., Lazar, N., 2008. North Cape ShellfishRestoration Program, 2007 annual report. Rhode Island Department of Environ-mental Management, National Oceanic and Atmospheric Administration, and U. S.Fish and Wildlife Service. 44 pp.

Durbin, A.G., Durbin, E.G., 1981. Standing stock and estimated production rates ofphytoplankton and zooplankton in Narragansett Bay, Rhode Island. Estuaries 4,24–41.

Ernst, L.M., Miguel, L.K., Willis, J., 1999. Rhode Island's Salt Pond Region: A SpecialArea Management Plan. Prepared for the Coastal Resources ManagementCouncil.

FAO, 2009. State of the World's Fisheries and Aquaculture 2008. Fisheries Department,Food and Agriculture Organization of the United Nations, Rome.

Ferreira, J.G., Hawkins, A.J.S., Monteiro, P., Service, M., Moore, H., Edwards, A., Gowen, R.,Lourenco, P., Mellor, A., Nunes, J.P., Pascoe, P.L., Ramos, L., Sequeira, A., Simas, T.,Strong, J., 2007. SMILE — Sustainable Mariculture in Northern Irish LoughEcosystems — Assessment of Carrying Capacity for Environmentally SustainableShellfish Culture in Carlingford Lough, Starngford Lough, Belfast Lough, LarneLough and Lough Foyle. IMAR - Institute of Marine Research. 100 pp.

Finn, J.T., 1976. Measures of ecosystem structure and function derived from analysis offlows. J. Theor. Biol. 56, 363–380.

Frisk, M., Miller, T.J., Latour, R.J., Martell, S.J.D., 2011. Assessing biomass gains frommarsh restoration in Delaware Bay using Ecopath with Ecosim. Ecol. Model. 220(1), 190–200.

Gerritsen, J., Holland, A.F., Irvine, D.E., 2002. Suspension-feeding bivalves and the fate ofprimary production: an estuarine model applied to Chesapeake Bay. Estuaries 17,403–416.

Hancock, B., DeAngelis, B., Turek, J., Lazar, N., 2007. North Cape Restoration Program,2006 annual report. Rhode Island Department of Environmental Management,National Oceanic and Atmospheric Administration, and U. S. Fish and WildlifeService. 64 pp.

Héral, M., 1993. Why carrying capacity models are useful tools for management ofbivalve culture. In: Dame, R. (Ed.), Bivalve Filter Feeders in Estuarine and CoastalEcosystem Processes. Springer-Verlag, Heidlelburg, pp. 455–477.

Hougham, A.L., Moran, S.B., 2007. Water mass ages of coastal ponds estimated using223Ra and 224Ra as tracers. Mar. Chem. 105, 194–207.

98 C. Byron et al. / Aquaculture 314 (2011) 87–99

Hougham, A.L., Moran, S.B., Masterson, J.P., Kelly, R.P., 2008. Seasonal changes insubmarine groundwater discharge to coastal salt ponds estimated using 226Ra and228Ra as tracers. Mar. Chem. 109, 268–278.

Inglis, G.J., Hayden, B.J., Ross, A.H., 2002. An overview of factors affecting the carryingcapacity of coastal embayments for mussel culture. Client Report CHC00/69. NIWA,Christchurch, p. 31.

IPCC, 2007. Climate Change 2007: The Physical Science Basis. Cambridge UniversityPress, Cambridge, UK. Available online at: http://www.ipcc.ch/.

Jiang, W., Gibbs, M.T., 2005. Predicting the carrying capacity of bivalve shellfish cultureusing a steady, linear food web model. Aquaculture 244, 171–185.

Jones, P.W., Martin, F.D., Hardy Jr., J.D., 1978. Development of Fishes of the Mid-AtlanticBight. US Department of the Interior. Fish and Wildlife Service, Solomons, MD(USA). 6 volumes.

Kjerfve, B., 1994. Coastal Lagoon Processes. Elsevier Science, Amsterdam. 577 pp.Kremer, P., 1976. Population dynamics and ecological energetics of a pulsed zooplankton

predator, the ctenophore Mnemiopsisleidyi. In: Wiley, M. (Ed.), Estuarine Processes:Uses, Stresses, and Adaptation to the Estuary, I. Academic Press, New York, pp.197–215.

Kurlansky, M., 2006. The big oyster. History on the Half Shell. Ballantine Books, NewYork. 307 pp.

Lee, V., 1980. An elusive compromise: Rhode Island coastal ponds and their people. In:Center, C.R. (Ed.), Marine Technical Report 73. University of Rhode Island.

Lee, V., Olsen, S., 1985. Eutrophication and management initiatives for the control ofnutrient inputs to Rhode Island coastal lagoons. Estuaries 8, 191–202.

Lee, V., Ernst, L., Marino, J., 1997. Rhode Island salt pond water quality. Salt PondWatchers monitoring data 1985–1994. Technical Report, October 1997. CoastalResources Center, University of Rhode Island and Rhode Island Sea Grant. http://seagrant.gso.uri.edu/coasts/RIsaltponddirectory.html.

Leloup, F.A., Desroy, N., Le Mao, P., Pauly, D., Le Pape, O., 2008. Interactions between anatural food web, shellfish farming and exotic species: the case of the Bay of MontSaint Michel (France). Estuar. Coast. Shelf Sci. 76, 111–120.

Link, J.A., 2010. Adding rigor to ecological network models by evaluating a set of pre-balanced diagnostics: a plea for PREBAL. Ecol. Model. 221, 1580–1591.

Link, J.A., Griswold, C.A., Methratta, E.T., Gunnard, J. (Eds.), 2006. Documentation for theEnergy Modeling and Analysis eXercise (EMAX). U.S. Department of Commerce,Northeast Fisheries Science Center. 166 pp. Available from: National MarineFisheries Service, 166 Water Street, Woods Hole, MA 02543–01026.

Link, J., O'Reilly, J., Fogarty, M., Dow, D., Vitaliano, J., Legault, C., Overholtz, W., Green, J.,Palka, D., Guida, V., Brodziak, J., Methratta, E., Stockhausen, W.T., 2008. Energy flowon Georges Bank revisited: the Energy Modeling and Analysis eXercise (EMAX) inhistorical context. J. Northwest Atlantic Fish. Sci. 39, 83–101.

Link, J., Col, L., Guida, V., Dow, E., O'Reilly, J., Green, J., Overholtz, W., Palka, D., Legault, C.,Vitaliano, J., Griswold, C., Fogarty, M., Friedland, K., 2009. Response of balancednetwork models to large-scale perturbation: implications for evaluating the role ofsmall pelagicsin the Gulf of Maine. Ecol. Model. 220, 351–369.

Longval, B.A., 2009. Biomass spectra in Narragansett Bay: from phytoplankton to fish.Ph.D. Thesis. Graduate School of Oceanography, University of Rhode Island.

Mace, P.M., 2001. A new role for MSY in single-species and ecosystem approaches tofisheries stock assessment and management. Fish Fish. 2, 2–32.

Mann, K.H., 2000. Ecology of Coastal Waters with Implications for Management, 2ndedition. Blackwell Science. 406 pp.

Markey, K.R., 2009. Performance of three lines of the Eastern oyster, Crassostreavirginica in Rhode Island Shellfish Farms. Masters Thesis. Fisheries, Animal andVeterinary Science Department, University of Rhode Island.

Masterson, J.P., Sorenson, J.R., Stone, J.R., Moran, S.B., Hougham, A., 2006. Hydro-geology and simulated ground-water flow in the salt pond region of southernRhode Island: U.S. Geological Survey, Scientific Investigations Report 2006–5271.56 pp.

McKindsey, C.W., Thetmeyer, H., Landry, T., Silvert, W., 2006. Review of recent carryingcapacity models for bivalve culture and recommendations for research andmanagement. Aquaculture 261, 451–462.

McKinney, R.A., Glatt, S.M., McWilliams, S.R., 2004. Allometric length–weight relation-ships for benthic prey of aquatic wildlife in coastal marine habitats. Wildl. Biol. 10,241–249.

McKinney, R.A., Glatt, S. M., McWilliams, S. R., 2005. Habitat use by waterfowl winteringin Narragansett Bay, Rhode Island. Ph.D. Thesis. Environmental Sciences, Universityof Rhode Island.

Meng, L., Cicchetti, G., Chintala, M., 2004. Nekton habitat quality at shallow water sitesin two Rhode Island coastal systems. Estuaries 27, 740–751.

Miller, A.W., Reynolds, A.C., Sobrino, C., Riedel, G.F., 2009. Shellfish face uncertain futurein high CO2 world: influence of acidification on oyster larvae calcification andgrowth in estuaries. PLoS ONE 4, e5661.

Monaco, M.E., Ulanowicz, R.E., 1997. Comparative ecosystem trophic structure of threeU.S. mid-Atlantic estuaries. Mar. Ecol. Prog. Ser. 161, 239–254.

Morissette, L., 2006. The trophic role of marine mammals in the northern gulf of St.Lawrence. Mar. Mamm. Sci. 22, 74–103.

Motes, M.L., DePaola, A., Cook, D.W., Veazey, J.E., Hunsucker, J.C., Garthright, W.E.,Blodgett, R.J., Chirtel, S.J., 1998. Influence of water temperature and salinity onVibrio vulnificus in Northern Gulf and Atlantic Coast oysters (Crassostrea virginica).Appl. Environ. Microbiol. 64, 1459–1465.

National Research Council (NRC), 2010. Ecosystem Concepts for Sustainable BivalveMariculture. The National Academies Press, Washington, D.C.. 179 pp.

Newell, R.I.E., 2004. Ecosystem influences of natural and cultivated populations ofsuspension-feeding bivalve molluscs: a review. J. Shellfish Res. 23, 51–61.

Nixon, S.W., 1982. Nutrient dynamics, primary production and fisheries yields oflagoons. Oceanol. Acta 4, 357–372.

Nixon, S.W., Buckley, B.A., 2007. Nitrogen inputs to Rhode Island coastal salt ponds —

too much of a good thing. A white paper prepared for the Rhode Island CoastalResources Management Council.

Nowicki, B.L., Nixon, S.W., 1985. Benthic community metabolism in a coastal lagoonecosystem. Mar. Ecol. Prog. Ser. 22, 21–30.

Nunes, J.P., Ferreira, J.G., Gazeau, F., Lencart-Silva, J., Zhang, X.L., Zhu, M.Y., Fang, J.G.,2003. A model for sustainable management of shellfish polyculture in coastal bays.Aquaculture 219, 257–277.

Odum, E.P., 1969. The strategy of ecosystem development. Science 104, 262–270.Olsen, S., Lee, V., 1982. Inlet modification: an example of an holistic approach to the

management of lagoons, Oceanologica Acta. Proceedings International Symposiumon coastal lagoons, SCOR/IABO/UNESCO, Bordeaux, France, 8–14 September, 1981,pp. 373–382.

Pauly, D., Soriano-Bartz, M.L., Palomares, M.L.D., 1993. Improved construction,parameterization and interpretation of steady state models. In: Christensen, V.,Pauly, D. (Eds.), Trophic models of aquatic ecosystems: ICLARM ConferenceProceedings, 26. International Center for Living Aquatic Resources Management(ICLARM), Manila, Philippines, pp. 1–13.

Peterson, B.J., Heck, K.L., 1999. The potential for suspension feeding bivalves to increaseseagrass productivity. J. Exp. Mar. Biol. Ecol. 240, 37–52.

Peterson, B.J., Heck, K., 2001. Positive interactions between suspension feeding bivalvesand seagrass — a facultative mutualism. Mar. Ecol. Prog. Ser. 213, 143–155.

Pietros, J.M., Rice, M.A., 2003. The impacts of aquaculture oyster, Crassostreavirginica,(Gmelin, 1791) on water column nitrogen and sedimentation: results of amesocosm study. Aquaculture 220, 407–422.

Plagányi, É.E., 2007. Models for an ecosystem approach to fisheries. FAO FisheriesTechnical Paper Series 477. 108 pp.

Plagányi, É.E., Butterworth, D.S., 2004. A critical look at the potential of Ecopathwith Ecosim to assist in practical fisheries management. S. Afr. J. Mar. Sci. 26,261–287.

Raillard, O., Ménesguen, A., 1994. An ecosystem box model for estimating the carryingcapacity of a macrotidal shellfish system. Mar. Ecol. Prog. Ser. 115, 117–130.

Rheault, R.B., 2008. Carrying capacity, CRMC Working Group on AquacultureRegulations. Subcommittee on Biology. Report on Biological Impacts of Aquacul-ture. Coastal Resources Management Council, pp. 49–59. Available at: http://www.crmc.ri.gov/aquaculture/riaquaworkinggroup/workinggroup_bioimpacts_acqua-culture.pdf.

Rybarcyk, H., Elkaim, B., 2003. An analysis of trophic network of a macrotidal estuary:the Seine Estuary (Eastern Channel, Normandy, France). Estuar. Coast. Shelf Sci. 58,775–791.

Rybarcyk, H., Elkaim, B., Ochs, L., Loquet, N., 2003. Analysis of the trophic network of amacrotidal ecosystem: the Bay of Somme (Eastern Channel). Estuar. Coast. ShelfSci. 58, 405–421.

Satchwill, R., Gray, C.L., 1990. Estuarine sportfish investigation survey of eighteen tidalrivers, coastal ponds and embayments in Rhode Island. The fisheries resources ofCharlestown pond and Green Hill pond, Charlestown, and South Kingstown RhodeIsland 1990. Rhode Island. Department of Environmental Management.Division ofFish and Wildlife. Federal Aid in Sportfish Restoration.

Satchwill, R., Sisson, R., 1990. The fisheries resources of Quonochontaug pondWesterlyand Charlestown, Rhode Island, 1989. Rhode Island Department of EnvironmentalManagement. Division of Fish and Wildlife. Federal Aid to Fisheries Project.F-51-RRhode Island. Segment Report 1.

Satchwill, R., Sisson, R., 1991. The fisheries resources of Point Judith and Potter pondSouth Kingstown and Narragansett, Rhode Island, 1991. Rhode Island Departmentof Environmental Management. Division of Fish and Wildlife.Federal Aid toFisheries Project. F-51-R Rhode Island. Segment Report 3.

Savenkoff, C., Bourdages, H., Swain, D.P., Despatie, S.P., Hanson, J.M., Méthot, R.,Morissette, L., Hammill, M.O., 2004. Input data and parameter estimates forecosystemmodels of the southern Gulf of St. Lawrence (mid-1980 andmid-1990s).Can. Tech. Rep. Fish. Aquat. Sci. , pp. 25–29.

Schult, M., 2010. Exploration of factors influencing benthic macrofaunal assemblagepatterns (N2.0 mm) within Narragansett Bay, Rhode Island. Masters Thesis.Graduate School of Oceanography, University of Rhode Island.

Schwinghamer, P., Hargrave, B., Peer, D., Hawkins, C.M., 1986. Partitioning ofproduction and respiration among size groups of organisms in an intertidalbenthic community. Mar. Eco. Prog. Ser. 31, 131–142.

Shannon, L.J., Moloney, C.L., Jarre, A., Field, J.G., 2003. Trophic flows in the southernBenguela during the 1980s and 1990s. J. Mar. Syst. 39, 83–116.

Smayda, T.J., Borkman, D.G., 2008. Nutrient and plankton dynamics in Narragansett Bay.In: Desbonnet, A., Costa-Pierce, B. (Eds.), Science for Ecosystem-Based Manage-ment: Narragansett Bay in the 21st century. Springer, pp. 431–484.

Soto, D., 2010. Aquaculture Development 4. Ecosystem approach to aquaculture. FAOTechnical Guidelines for Responsible Fisheries. 5 (Suppl. 4), 53 pp.

Steele, J.H., 2009. Assessment of some linear food web models. J. Mar. Sci. 76, 186–194.Steele, J.H., Collie, J.S., Bisagni, J.J., Gifford, D.J., Fogarty, M.J., Link, J.S., et al., 2007.

Balancing end-to-end budgets of the Georges Bank ecosystem. Prog. Oceanogr. 74,423–448.

Taylor, D.I., Nixon, S.W., Granger, S.G., Buckley, B.A., 1999. Responses of coastal lagoonplant communities to levels of nutrient enrichment: a mesocosm study. Estuaries22, 1041–1056.

Thorne-Miller, B., Harlin, M.M., 1984. The Production of Zostera marina L. and othersubmerged macrophytes in a coastal lagoon in Rhode Island, U.S.A. Bot. Mar. XXVII,539–546.

Thorne-Miller, B., Harlin, M.M., Thursby, G.B., Brady-Campbell, M.M., Dworetzky, B.A.,1983. Variations in the distribution and biomass of submerged macrophytes in fivecoastal lagoons in Rhode Island, U.S.A. Bot. Mar. XXVI, 231–242.

99C. Byron et al. / Aquaculture 314 (2011) 87–99

Thornber, C., Yarish, C., Tobin, A. Unpublished data on algal biomass in Rhode Islandcoastal lagoons. University of Rhode Island.

Ulanowicz, R.E., 1986. Growth and Development: Ecosystem Phenomenology. SpringerVerlag, New York. reprinted by iUniverse, 2000 203 pp.

Valiela, I., 2006. Global Coastal Change. Blackwell Publishing Ltd. 368 pp.Vasconcellos, M., Mackinson, S., Sloman, K., Pauly, D., 1997. The stability of trophic

mass-balance models of marine ecosystems: a comparative analysis. Ecol. Model.100, 125–134.

Warwick, R.M., Joint, I.R., Radford, P.J., 1979. Secondary production of the benthos in anestuarine environment. In: Jeffries, R.L., Davy, A.J. (Eds.), Ecological Processes inCoastal Environments. Blackwell, London, pp. 429–450.

Wright, T.J., Cheadle, V.I., Palmatier, E.A., 1949. A survey of Rhode Island's salt andbrackish water ponds and marshes. In: Rhode Island Department of Agricultureand Conservation, Department of Fish and Game (Ed.), Pittman RobertsonPamphlet No. 2.

Zajak, R.N., Seals, K., Simpson, D., 2009. Food webs in Long Island Sound: review,synthesis and potential applications. Final Report — EPA Grant No. LI-97101401,Submitted to the USEPA Long Island Sound Study. 84 pp.