Aquacult Int

DOI 10.1007/s10499-016-9990-3

EUROPEAN AQUACULTURE DEVELOPMENT SINCE 1993

Recirculation nursery systems for bivalves

Pauline Kamermans1 • Ainhoa Blanco1 • Sandra Joaquim2 •

Domitı́lia Matias2,3 • Thorolf Magnesen4 • Jean Louis Nicolas5 •

Bruno Petton5 • Rene Robert5,6

Received: 25 March 2015 / Accepted: 26 February 2016

Ó Springer International Publishing Switzerland 2016

Abstract In order to increase production of bivalves in hatcheries and nurseries, the

development of new technology and its integration into commercial bivalve hatcheries is
important. Recirculation aquaculture systems (RASs) have several advantages: high den-
sities of the species can be cultured resulting in a cost-effective production system; optimal
temperature maximizes production and allows rapid turnover of the product; stable water
quality improves growth rate and minimizes stress and potential loss by diseases. Pilot
RAS systems were developed for seed rearing of oysters (Crassostrea gigas), scallops
(Pecten maximus), mussels (Mytilus edulis) and clams (Ruditapes decussatus). Optimal
feed addition and waste matrix were determined. Based on this, system flow rates were
designed. Seed growth in the pilot RAS systems was compared at different renewal rates
and with growth in flow-through systems (FTS). All four species can be reared in RAS and
showed similar growth in RAS and in FTS or in RAS with a higher renewal rate. RAS can
keep O2, nitrogen and pH within the desired range. Temperature was generally higher in
RAS than in FTS, probably due to heat induced by the pump circulating the water. The

Guest editors: Elena Mente and Aad Smaal/European Aquaculture Development since 1993: The benefits of
aquaculture to Europe and the perspectives of European aquaculture production.

& Pauline Kamermans

pauline.kamermans@wur.nl
1
IMARES, Wageningen UR, PO Box 77, 4400 AB Yerseke, The Netherlands
2
IPMA, Portuguese Institute for the Ocean and Atmosphere, I.P. Av. 5 de Outubro s/n,
8700-305 Olhão, Portugal
3
CIIMAR – Interdisciplinary Centre of Marine and Environmental Research, University of Porto,
Rua dos Bragas 177, 4050-123 Porto, Portugal
4
Department of Biology, University of Bergen, PO Box 7800, 5020 Bergen, Norway
5
Ifremer, Laboratoire de Physiologie des Invertébrés Marins (UMR 6539, LEMAR),
29280 Plouzané, France
6
Ifremer, Unité Littoral, Centre Bretagne - ZI de la Pointe du Diable - CS 10070, 29280 Plouzané,
France

123
 Aquacult Int

supply of sufficient amount of food in combination with a desire to reduce the renewal rate
calls for use of concentrated feed in RAS.

Keywords Bivalve seed  Recirculation aquaculture system  Crassostrea gigas  Mytilus

edulis  Pecten maximus  Ruditapes decussatus

Introduction

In order to stabilize and increase survival of bivalves in hatcheries and nurseries, the
development of new technology and its integration into commercial bivalve hatcheries is
needed. The technology used in bivalve hatcheries has not progressed very much since the
birth of this industry and still essentially relies on methods developed in the 1960s by
Loosanoff and Davis (1963) for Crassostrea virginica and Mercenaria mercenaria and
Walne (1974) for Ostrea edulis. These methods have gradually been adapted to most
cultured bivalves though without any great changes (Helm et al. 2004).
Bivalve nurseries generally make use of a flow-through system where food is added
continuously. The seawater and energy needed can potentially be reduced considerably if
the flow-through system is linked to a recirculation system. The concept of recirculation
aquaculture systems (RAS) is to reuse a volume of water through continuous treatment and
delivery to organisms being cultured. A properly designed and operated recirculation
system requires 5–10 % of the total system volume as a daily input of water. In addition,
the use of a RAS provides better water quality control, since water will be taken in only
once per rearing cycle and is thus not subject to fluctuations in the quality of the natural
water supply. Furthermore, well-managed RAS have a stable water quality and beneficial
microbiota (Blancheton et al. 2013) that are hardly influenced by the low water intake.
Introduction of RAS technology and the use of different water treatments, such as biofilters
and protein skimmers, will provide the opportunity to increase the biomass cultured and
improve growth and survival without any considerable increase in the costs (Dunning et al.
1998; Martins et al. 2010; Besson et al. 2014); better water quality management and animal
welfare will also therefore be attained. It is nowadays commonly accepted in finfish culture
that this type of system contributes to increased survival of the larvae in hatcheries and
during later grow-out stages (Blancheton 2000; Attramadala et al. 2012a; Faulk and Holt
2005). Overall, RAS reduces stress and improves health (Martins et al. 2010; Attramadala
et al. 2012b). Other advantages include: low land requirements (high densities of the
species cultured resulting in a cost-effective production system); water temperature control
(optimal temperature maximizes production and allows rapid turnover of the product;
species exotic to the area can be cultured); and independence from weather conditions
(Timmons and Ebeling 2007). However, there are disadvantages as well. Bivalves are fed
microalgae that are suspended at a predefined concentration in the water. The biofilter and
protein skimmer of RAS remove the algae. Thus, part of the algae is wasted before they are
consumed. However, in a flow-through system not all algae are consumed either, before the
water leaves the system. Another disadvantage of recirculating the water is the risk of
carbonate depletion. Bivalves need carbonate for shell production. Thus, in a long-term set-
up carbonate needs to be added.
At present, there are no recirculation systems for bivalves in commercial use, and only a
limited number of studies have been performed on marine species (e.g. Widman 1998;

123
Aquacult Int

Pfeiffer and Rusch 2000; Suantika et al. 2000; Xiongfei et al. 2005; Zohar et al. 2005;
Merino et al. 2009; Magnesen and Jacobsen 2012; Robert et al. 2013; Blanco Garcia and
Kamermans 2015; Joaquim et al. 2014).
Seed rearing generally concerns a higher rearing biomass than larval rearing and a lower
sensitivity to stress during this stage (Helm et al. 2004). Improvements obtained through
implementation of RAS will be particularly important for M. edulis, as this species has a
lower market value than the other bivalves cultured in hatcheries. RAS is also an inter-
esting technique for P. maximus in the northern European countries where hatchery run-
ning costs are closely related to seawater heating costs, as well as for R. decussatus in
Portugal where production is still dependent on catch of wild juveniles, which present a
large recruitment fluctuation.
In the framework of the EU FP7 project REPROSEED (reproseed.eu), RAS systems
were developed for seed rearing of oysters (Crassostrea gigas), scallops (Pecten maximus),
mussels (Mytilus edulis) and clams (Ruditapes decussatus). Results of these trials are
presented here. The main purpose of the experiments was to test RAS and compare results
at different renewal rates and with FTS for different species. Thus, for each species a
comparison between RAS and FTS, or RAS at a different renewal rate can be made.

Methods

RAS needs to accommodate the input of high amounts of feed and high stocking densities.
The first step in a successful implementation of RAS for the rearing of bivalve spat is the
determination of the optimal feed addition. The second step is the determination of the
waste matrix. This can be done by quantifying the feed addition, the solid waste production
and the dissolved waste production. Based on these figures, the necessary treatment steps
can be chosen (type of tank, type of mechanical filtration, need for nitrification, denitri-
fication or for fine particle removal, e.g. by skimming and ozone, need for routine cleaning
of settled algae, etc.). Based on these decisions, the system flow rates can be designed (seed
rearing tank, settlement basin, protein skimmers and biofilter). The last step is the building
of a pilot system followed by testing of system management. Tests with the four species
were carried out at four institutes involved in REPROSEED. This led to slightly different
approaches for the different species. For oysters and scallops, the pseudo-faeces threshold
and clearance rate were not established. For clams, the waste matrix was not determined.

Optimal feed addition

Diet tests were carried out for oysters. Five diets based on Tisochrysis lutea formerly
Isochrysis affinis galbana or T-Iso CCAP 927/14 (T) and Chaetoceros neogracile UTEX
LB2658 (Cg) assemblages were tested in triplicate for oyster spat (C. gigas): 100 T, 75 T/
25 Cg, 50 T/50 Cg, 25 T/75 Cg and 100 Cg. The five diets were tested in triplicate per
condition in 500-ml cylindro-conical Perspex tubes. Seawater flow rate depended on spat
biomass and due to its increase with time it ranged from 5 to 30 l/h. Each cylinder received
7 g of spat (total weight) corresponding to 26,500 young postlarvae (&265 lg/individual).
For the other three bivalve species, the diet was based on earlier studies. For mussels (M.
edulis), a mixture of the algal species Pavlova lutheri CCAP 931/6 and Chaetoceros
muelleri (Stichting Zeeschelp) proved to be the best diet for M. edulis spat in the BLUE
SEED project (blueseedproject.com). Postscallop spat (P. maximus) were fed with a

123
 Aquacult Int

mixture of the single celled algae T. lutea. (T-iso strain CCAP 927/14), P. lutheri CCAP
931/1, C. muelleri CCAP 1010/3 and Skeletonema marinoi CCAP 1077/5 at a ratio of
3:2:3:1 (Magnesen and Jacobsen 2012). Clam seed (R. decussatus) were fed with
microalgae T. lutea (T-iso) and Chaetoceros calcitrans (C. cal) (1:1 in cells/ml) (Fer-
nandez-Reiriz et al. 1999).
The pseudo-faeces threshold (maximum concentration of microalgae cells that seed can
consume without producing pseudo-faeces) was determined experimentally. The juveniles
were separated in three size classes (0–5; 5–9.5; 9.5–15 mm). Before this experiment, the
seed was maintained in a ‘‘batch’’ system with a seawater temperature of 18 °C, unfed
during one to 2 days to empty the stomach contents. After this fasting period, 10–20
bivalves per size class were placed in triplicate beakers with filtered seawater at 18 °C. For
R. decussatus, a mixed diet composed of two microalgae species T. lutea (T-iso) and C.
calcitrans (C. cal) (1:1 in cells/ml) was added at concentrations from 40,000 up to
100,000 cell/ml. For M. edulis, a diet composed of P. lutheri was added at concentrations
from 10,000 up to 100,000 cell/ml. The presence of faeces and pseudo-faeces was checked
regularly under a binocular microscope during a period of 3–4 h. When pseudo-faeces
(fluffy material, in contrast to solid faeces) was observed in one of the triplicate beakers a
positive score was noted. The concentration below this concentration was considered the
pseudo-faeces threshold. Every hour, a water sample was taken to count the algae con-
centration, when necessary, algae were added to restore the desired concentration.
Clearance rate on algae was measured only for the largest size classes of clam seed
(9.5–15 mm) and mussel seed (18.6 mm). A reservoir was filled with the algal diet that
was used in the growth experiments in a concentration just below the pseudo-faeces
threshold. The microalgae (same diets as in pseudo-faeces threshold determination) were
pumped into flow-through chambers with bivalve seed and controls without seed). The
seed was allowed to acclimatize for 1 h. The microalgae concentration in the outflow of
each chamber was measured in triplicate, every 15 min during 3 h. Clearance rate was then
calculated for each 15 min with the average of triplicates using the formula described by
Petersen et al. (2004):
 
Cin  Cout
CR ¼ Q
Cs
where CR is clearance rate, Cin average algae concentration (cells/ml) in the outflow of the
control chambers, Cout average algae concentration (cells/ml) in the outflow of the bivalve
chambers, Cs concentration in the chamber (here considered to be equal to the outflow
concentration from the bivalve chambers and Q flow speed to the chambers.
Dry weight and ash-free dry weight of the shellfish used in this experiment were
determined. The clearance rate values were used in the mass balance spreadsheet as input
for the ‘‘feed rate as body weight percentage’’.

Waste matrix

Waste matrix calculations are used in the design of RAS to identify and quantify inputs,
outputs and internal changes in the systems (e.g. Losordo and Hobbs 2000; Timmons and
Ebeling 2007). A design based on the waste matrix (total ammonia nitrogen
(TAN = NH3 ? NH4-), dissolved oxygen (DO) and solids mass calculations) tries to
estimate the water flow requirements and the proper sizing of the biological filter. The
design of RAS should ensure that the important parameters affecting water quality and

123
Aquacult Int

system productivity (e.g. oxygen, ammonia, solids) are properly linked. This requires
calculating the value of each of these parameters independently to determine the thresholds
of each. For this, oxygen consumption and nitrogen production of the mussel seed were
determined in short-term experiments in batch mode (see Blanco Garcia and Kamermans
2015). Temperature and salinity were in the same range as in the culture systems (Table 2).
Groups of animals were utilized. Stocking density for mussels was three groups of twenty
mussels each of 15–20 mm. If some of the animals were not pumping, this was considered
to represent the situation in the culture system that was used for the growth experiments.
For oysters and scallops, oxygen consumption and nitrogen production were derived from
standard conditions at the hatchery. Once the measurements and calculations are done, the
system is operated at the maximum flow rate possible while maintaining a specific
parameter at or below its maximum tolerable level. The water quality criteria for nitrogen
in fish aquaculture are \1 mg TAN/l, \1 mg NO2/l and \400 mg NO3/l (Timmons and
Ebeling 2007). For oxygen a value of[7.64 mg/l was set (Timmons and Ebeling 2007) and
a seawater pH of 8. The optimal flow for one parameter may be too high or too low for the
other so the same waste matrix approach can be done for all the variables affecting water
quality. Losordo and Hobbs (2000) designed a computer spreadsheet for water flow and
biofilter sizing of recirculation aquaculture systems. This spreadsheet was adjusted for the
bivalve seed recirculation system necessities (Table 1). In this study, we calculated the
waste matrix for mussel spat, scallop spat and oyster spat.

Pilot systems

The RAS included culture and sump units (Fig. 1). The sump consisted of the water inlet
section, decanting cabinet, biofilter section and resting cabinet. Water from the culture unit
was collected in the inlet section of the sump and flowed upwards into the decanting
cabinet. Thereafter, water flowed from the top to the biofilter section. This was composed
of different components; a biological trickling filter and filter fleece bags that work as a
sedimentation unit; or bio balls layer, filter wool (saltwater proof synthetic fibre) and a
sponge layer. The filtered water passed to the resting cabinet and was then pumped through
a UV and a protein skimmer and then to the culture unit. For mussels, an algae aqua feed
regulator was connected to a peristaltic pump that could pump microalgae from a
microalgae reservoir into the rearing tank. To mature the biological filter, 1 lm filtered and
UV-sterilized seawater was added to the RAS system 8 weeks before the beginning of the
experiment. Furthermore, NH4Cl and NaNO3 were added as feed for the bacteria (except
for the pilot system for clams). The colonization process and the maintenance of the
biological balance of the filter were guaranteed by the addition of nitrifying and hetero-
trophic bacteria and microalgae to the system. Except for clams, the RAS were compared
with flow-through systems (FTS). They were the same as the RAS, but without the sump
unit. The flow for FTS was adjusted from Rico-Villa et al. (2006).
Different methods were used to deliver the algae to the spat in both RAS and FTS.
Oysters spat were continuously fed T. lutea (mean cell volume of 40 lm3) and C. neo-
gracile (80 lm3) at different single and bispecific microalgal diets, on the basis of a
constant concentration of 1500 lm3/ll at the exit of each rearing tanks. Feeding adjust-
ment was made twice a day after measuring phytoplankton concentration, by means of an
electronic coulter counter equipped with a 100-lm aperture tube, at the inlet and outlet of
each spat container. For scallops, algal cell density was monitored 1–2 times a week and
added manually directly to the tank to keep the desired algal concentration level in the
tank. For clams, food was provided using a peristaltic pump multichannel to the RAS at 2 h

123
 Aquacult Int

Table 1 Spreadsheet for flow rate estimation and biofilter sizing for the bivalve seed recirculation system
(RAS)

Tank size and biomass Tank water depth (m)

Tank radius (m)

Tank volume (m3 )

Maximum culture density (kg/m3)

Spat biomass (kg)

Spat weight (g)

Feed rate as body weight percentage (%)

Feed rate (kg/day)

TAN mass balances calculations Feed protein content (%)

TAN production rate (kg/day)

Percentage TAN from feed

Desired TAN in RAS (mg/l)

Passive nitrification (%)

TAN available after passive nitrification (kg/day)

Passive denitrification (%)

Maximum nitrate concentration desired (mg/l)

New water required to maintain nitrate concentration (l/day)

TAN available to biofilter after effluent removal (kg/day)

Biofilter efficiency from TAN removal (%)

Flow rate through biofilter to remove TAN to desired concentration (l/min)

Biofilter sizing calculation Estimated nitrification rate (gTAN/m2/ day)

Active nitrification surface required (m2)

Surface area of media (m2/m3)

Total volume media (m3 )

Media depth (m)

Volume/depth yields face area (m2)

Diameter of biofilter (m)

Oxygen mass balances calculations Submerged filter? (1 = yes and 0 = no)

Oxygen used per kg feed (%)

Oxygen used by feed addition (kg/day)

Desired oxygen concentration in tank (mg/l)

Dissolved oxygen concentration supplied to tank (mg/l)

Oxygen used by passive nitrification (kg/day)

Oxygen used for nitrification in biofilter (kg/day)

Total oxygen used (kg/day)

Estimated flow rate through system to keep oxygen at desired concentration (l/min)

Cells for input values are shown in white and cells for calculated parameters in grey

123
Aquacult Int

Culture unit

Water flow
Water flow

U.V.
Biofilter tube

Water flow

Filter bag
Air flow

Pump
Protein Skimmer

Decanting Resting
cabinet cabinet

Fig. 1 Schematic representation of the recirculating systems used for spat rearing

intervals. For mussels, a newly developed algae aqua feed regulator was used (Blanco
Garcia and Kamermans 2015). The actual algal cell density in the rearing container of both
RAS and FTS determines the addition of algae to the seed. The method makes use of a
feedback mechanism through a fluorescence sensor and thus controls the incoming food
levels.
During the experimental period, the daily maintenance of RAS consisted in changing
the seawater to offset the increase in salinity caused by evaporation and cleaning the
production tanks walls and the filter bag, to avoid clogging and overflowing. Daily
renewal rates are presented in Table 2. In the RAS, for scallops calcium carbonate
(oolitic coral sand) was added in the recirculation process and alkalinity measured.
During the trials, temperature, pH, TAN, NO2–N, NO3–N, NH4 and dissolved oxygen
levels were recorded daily in each of the rearing tanks. Total weight was measured each
week until the end of the experimental period, whereas survival as a percentage (number
of observed live spat/initial number of spat 9 100), individual shell length and weight
was measured on samples of 50 individuals from each replicate. Growth performance
was expressed as specific growth rate (SGR) according to the following formula
(Frandsen and Dolmer 2002):

123
 Table 2 Results of RAS and FTS comparison for four bivalve species
Bivalve species Oyster Scallop Mussel Clam
Crassostrea gigas Pecten maximus Mytilus edulis Ruditapes decussatus

123
Total RAS volume (l) 43 4898 127 115
Rearing tank volume RAS and FTS (l) 0.5 3500 32 0.5
Initial seed density RAS and FTS (g) 6 75,000 spat 450 100
Algal diet RAS and FTS Tisochrysis lutea and T. lutea, Pavlova lutheri, P. lutheri and C. muelleri T. lutea and Chaetoceros
Chaetoceros neogracile Chaetoceros muelleri and (1:1 cells/ml) calcitrans (1:1 cells/ml)
(50/50 equivalent cell volume) Skeletonema marinoi
at a ratio of 3:2:3:1
Flow rate of RAS (l/min) 0.08–0.50 3.72 0.96 0.13 (1600 %) and 0.06 (800 %)
Water renewal RAS (%/day) 120–480 20 10 10
Water renewal FTS (%/day) 2400 600 520 No data
Temperature in RAS (°C) 23.7 ± 0.2, n = 12 17.0 ± 0.7, n = 3 20.42 ± 0.14, n = 3 18.9 and 23.8 °C
Temperature in FTS (°C) 23.7 ± 0.3, n = 4 16.6 ± 0.6, n = 3 18.00 ± 0.08, n = 3 No data
Salinity in RAS (PSU) 33–34 34.5 ± 0.0 34.73 ± 0.12, n = 3 33 ± 1
Salinity in FTS (PSU) 33–34 33.4 ± 0.4 33.18 ± 0.04, n = 3 No data
pH in RAS 7.5 ± 0.2, n = 12 8.0 ± 0.1, n = 3 8.03 ± 0.10, n = 3 7.95–7.32
pH in FTS 8.0 ± 0.1, n = 4 8.1 ± 0.1, n = 3 8.16 ± 0.01, n = 3 No data
O2 in RAS (mg/l) 6.5 ± 0.4, n = 12 9.5 ± 0.3, n = 3 7.96 ± 0.06, n = 3 7.69–9.26
O2 in FTS (mg/l) 6.9 ± 0.0, n = 4 9.7 ± 0.3, n = 3 7.77 ± 0.03, n = 3 No data
NH4 in RAS (mg/l) –a No data 0.05 ± 0.01, n = 3 Below 0.1
NH4 in FTS (mg/l) –a No data 0.06 ± 0.02, n = 3 No data
NO3 in RAS (mg/l) –a No data 9.71 ± 3.98, n = 3 2.31 ± 0.17, n = 3
NO3 in FTS (mg/l) –a No data 0.96 ± 0.10, n = 3 No data
NO2 in RAS (mg/l) –a No data 4.30 ± 1.40, n = 3 Below 0.1
Aquacult Int
 Table 2 continued

Bivalve species Oyster Scallop Mussel Clam

Crassostrea gigas Pecten maximus Mytilus edulis Ruditapes decussatus
Aquacult Int

NO2 in FTS (mg/l) –a No data 0.02 ± 0.00, n = 3 No data

Specific growth rate of seed in RAS 2.23 mg/day 27 lm/day 5.30 ± 0.48 %/day, n = 3 0.144 ± 0.03 mg/day for 1600 % and
(10 % treatment) 0.062 ± 0.06 mg/day for 800 %
Specific growth rate of seed in FTS 2.74 mg/day 26 lm/day 5.08 ± 0.02 %/day, n = 3 No data
a
For C. gigas, the RAS system has been previously developed for larvae before its adaptation to spat. Chemical parameters were studied during the larval experiment (Robert
et al. 2013) and were not recorded for spat trials

123
 Aquacult Int

   
% LnðWf Þ  LnðWi Þ
SGR ¼  100
day Time ðdays)
where Wf and Wi were the average final and initial weight of the bivalve seed, respectively.
Time (days) represents the duration of the experiment.
Implementation of RAS was tested at commercial scale for scallop seed (3500-l rearing
tanks) and oyster seed (840-l rearing tanks). For the other two species, there was no
commercial hatchery available.

Results

Optimal feed addition

The preferred diet for C. gigas spat was a balanced mixed diet 50 % T. lutea/50 % C. neogracile
or a diet with more diatoms 25 % T. lutea/75 % C. neogracile. With an induced growth
increment of 0.75 mg/day and 0.18 mm/day, the 50/50 bispecific diet was accordingly used.
Mussel seed until 5 mm filtered 40,000 cells/ml without producing pseudo-faeces, the
threshold was 50,000 cells/ml for 5–10 mm, 70,000 cells/ml for 10–15 mm and
80,000 cells/ml for 15–20 mm (Blanco Garcia and Kamermans 2015). The clearance rate
for mussel spat was highest at microalgae concentrations of 31,000 and 51,000 cells/ml
(13.23 ± 0.18 and 12.31 ± 0.12 l/h/g AFDW, respectively) and lowest at 87,000 cells/ml
(2.6 ± 0.11 l/h/g AFDW).
Scallops optimal feed addition was based on earlier work (Magnesen and Jacobsen
2012).
The concentration of 40,000 cell/ml was found to be excessive for clam spat until
5 mm. The concentrations 80,000 and 90,000 cells/ml should be used to feed 5–9.5 and
9.5–15 mm size classes spat, respectively. The clearance rate of R. decussatus spat was
higher in the first 15 min (23.14 ± 1.9 l/h/g AFDW), and after this time this rate decreased
until after 2 and 3 h, when were registered two more peaks of the clearance rate
(23.85 ± 1.9 l/h/g AFDW and 28.43 ± 2.32 l/h/g AFDW).

Determination of waste matrix

Waste matrix calculations for bivalve spat were performed including tank size and bio-
mass, TAN calculations, biofilter sizing calculations, solids and oxygen calculations. Based
on this an estimated required flow rate of 0.3 l/min was found for oyster spat in a 0.5-l
system, 0.96 l/min for mussel spat in a 127-l system with 32-l rearing tanks (Blanco Garcia
and Kamermans 2015), 3.72 l/min for scallop spat in a 4898-l system (Table 2).

Implementation of RAS at small and commercial scale

In the transition from FTS to RAS, different renewal rates were tested varying from 2400
to 10 %/day (Table 2).

Oyster

To evaluate the biomass that can be reared in a recirculating system without any dys-
function of the system (such as clogging, waste excess, spat loss), two spat densities at the

123
Aquacult Int

10

FT

Mean individual spat length (mm)

8 RAS 5%
RAS 10%
RAS 20%
6

0
A
28 38 48
Time from ferlizaon (day)

50
Mean individual spat wet weight (mg)

40 FT
RAS 5%
30 RAS 10%
RAS 20%

20

10

0
B
28 38 48
Time since ferlizaon (days)

Fig. 2 Growth in length (a) and wet weight (b) of Crassostrea gigas spat reared in experimental RAS with
renewal of fresh seawater varying from 100 (FT = flow through = control) to 5 % (only 5 % of fresh
seawater were added in the recirculated circuit per hour: RAS)

onset of the experiment (standard = ds = 7 g and 21 g = 3 ds) combined with three

seawater flow rates (1/2 standard = ‘ rs = 2.5–15 l/h; 1 standard = rs = 5–30 l/h; 2
standard = 2 rs = 10–60 l/h) were tested without any connection to the sump unit.
Negative effects of high seawater flow rate (10–60 l/h) on spat growth (linear and whole
weight) were found (Fig. 2). The low seawater flow rate (0.5 rs: 2.5–15 l/h) led to the best
spat growth performances with 1.48 mg/day and 0.35 mm/day. For seawater flow rate
\2 rs, higher spat density (21 vs. 7 g originally) had negative effects on whole spat weight
but not on linear growth. Spat reared at high density was accordingly lighter with a similar
length than spat reared at standard density. With spat growth performances of 1.12 mg/day
and 0.20 mm/day, the combination of seawater renewal (5–30 l/h) and spat density (7 g)
appeared to be a good compromise between biological performances and technical pos-
sibilities. Indeed it was problematic to maintain 2.5–15 l/h seawater flow in our equipment.
Thus, 7 g initial spat biomass was subsequently reared in 500 ml at 5–30 l/h flow rate.

123
 Aquacult Int

20 weight of the spat (Kg)

18
16
14
12
10
8
6
4
2
0
22/3 1/4 11/4 21/4 1/5 11/5 21/5

Fig. 3 Growth of Crassostrea gigas spat in the Satmar hatchery

The RAS was thereafter implemented with three biofilters connected to 500-ml seed
rearing systems using the previous results (7 g initial biomass in each 500-ml unit, flow
rate 5–30 l/h). Two-week-old postset larvae were distributed in each unit at 6 g total
biomass (&6500–8500 individuals) and reared in RAS at three seawater renewal rates, 20,
10 and 5 %/h, and compared to flow through (FT) at 100 % seawater renewal. Spat was
reared in 1 lm filtered and UV-treated seawater at 24 °C temperature, ambient salinity
(&34.5 ppm) and fed T. lutea and C. neogracile at 50/50 (equivalent cell volume) at
1500 lm3/ll (Table 2). After 2 weeks of development, C. gigas young spat reared in RAS
10 and 20 % exhibited lower spat size with 0.22–0.24 mm/day than FT or RAS 5 % with
0.28–0.30 mm/day for FT and RAS 5 %, respectively (Fig. 2a). However, one-way
analysis of variance showed no significant effect of flow rates on spat length on day 44.
Spat reared in FT exhibited the highest weight growth with 2.74 mg/day, whereas spat
cultivated in RAS 20 % and RAS 10 % showed similar weight increments with a daily
growth of 2.16–2.23 mg/day (Fig. 2b). In contrast, weight growth was lower when sea-
water renewal decreased in RAS. Thus, daily spat growth was only 1.45 mg/day at 5 %
fresh seawater renewal (Fig. 2b). One-way analysis of variance showed a significant effect
of flow rate on spat weight (F = 85.15; df = 3, p \ 0.001) with, however, no significant
differences occurring at 10 and 20 % using a Tukey comparison test. Satisfying weight
growth development was recorded at 10–20 % fresh seawater renewal/h close to FT
performances (20 % loss), but spat size was lower (Fig. 2). 5 % seawater renewal/h in
RAS is not recommended because it induced a loss of &50 % of spat weight growth
despite a similar length development. 10 % renewal per hour appeared to be a good
compromise in RAS for oyster spat, but further economic studies should be performed to
evaluate the benefit on heat economy and food.
In Satmar, the cost of the water treatment is much lower at the larval stage than that
at the micronursery stage. This is the reason why RAS for rearing small spat (500 lm
to 2 mm) is more interesting for a commercial hatchery. An experiment was carried out
at Satmar with oyster spat (500 ml of oyster spat at 800 lm size) in a RAS including a
filtration system, 840 l of media for biofiltration and a UV treatment. The flow rate
varied from 1.5 to 2 m3/h in function of biomass. The renewal rate of seawater was
fixed at 10 %, i.e. 150–200 l/h. The experiment lasted 2 months. The spat was healthy
and the growth and algal consumption equivalent to those in classical system (the grow
rate of the first week was 31 and 9.3 %/day after). The seed reached 3–12 mm for a
total weight of 22.5 kg (Fig. 3). The biofilter was prepared in maintaining oyster adults
in the system 2 weeks before the experiment. The nitrite did not exceed 2 mg/l while

123
Aquacult Int

Fig. 4 Shell length (mm) and

ash-free dry weight (mg) of
mussel seed after 4 weeks of
rearing in a recirculation system
(RAS) and flow-through system
(FTS)

nitrate attained 25 mg/l at the end of experiment. The temperature was maintained
around 22 °C.

Mussel

No mortality was recorded in RAS and FTS. Mussel seed reached similar shell lengths in
FTS (26.56 ± 0.69 mm) compared to RAS (27.16 ± 1.25 mm) (Fig. 4). Final length and
weight did not differ significantly between RAS and FTS (ANOVA, p [ 0.05). Specific
growth rate did not differ significantly between RAS and FTS (Table 2, ANOVA,
p [ 0.05). Water quality parameters were within the desired range, except for nitrite
(Table 2). pH, oxygen and ammonium concentration were not significantly different
between RAS and FTS (ANOVA, p [ 0.05). Temperature was significantly higher in RAS
(ANOVA, p \ 0.05). This is probably caused by the heat given off by the pump for
recirculation the water. Nitrate and nitrite were also significantly higher in RAS (ANOVA,
p \ 0.05). Nitrate levels were above the set concentration. However, the seed showed no
mortality and good growth, which indicates that mussel seed can tolerate nitrate levels
above 1 mg/l. Epifanio and Srna (1975) found a high nitrite tolerance in the hard clam (M.
mercenaria) and the oyster (C. gigas) juveniles when exposed to concentrations as high as
2415 mg NO2/l. After a 96 h exposure period, the 50 % lethal concentration (LC50) for
nitrite ranged between 329 and 735 mg NO2/l.

Scallop

Experiments on scallop postlarvae in downwelling sieves containing postlarvae showed

that no significant differences were found for growth rate in the two systems (Fig. 5; t test,
p = 0.168). In the large-scale experiments, the alkalinity (CaCO3) was stable between 165
and 170 ppm. No pH reduction was found (Table 2), but at times increased nitrite con-
centrations were observed. RAS functions well for scallop postlarvae with regard to

123
 Aquacult Int

Fig. 5 Shell height (lm) of scallop spat in flow-through system (FTS) and a recirculation system (RAS)

settlement, survival and shell growth. The high potential for RAS to reduce water and
energy use in scallop spat nursery therefore clearly exists. The economic evaluation
showed a reduction of approximately 10 % of the total costs compared to the flow-through
system. An additional point is the stabilization of water quality meaning reduced labour for
the same production or increased production for same labour. There is a need to go further
to improve biofilter system set-up and find limits regarding maximum biomass capacities
for bivalves.

Clam

There was no statistically significant difference in seawater temperature, pH and dissolved

oxygen between the two tested treatments (K–W, p [ 0.05). The seawater temperature
increased between 18.9 and 23.8 °C, along the experimental period (Table 2). Dissolved
oxygen remained stable between 7.69 and 9.26 mg/l, except in the last 7 days of culture
when a rise of temperature causes a decrease in this parameter. The pH decreased along the
experimental period from 7.95 to 7.32 in all treatment. At the end of the experiment, the
ammonia increased in the 1600 % flow treatment. However, concentration remained below
0.1 mg/l in all treatments throughout the experimental period (Table 2). The nitrites
concentration increased throughout the experiment, especially in the flow 800 %
(0.071 mg/l). The content of nitrates increased at the end of the experiment and was similar
in the two treatments (2.44 mg/l). These values do not exhibit toxicity for bivalves.
During the 60 days that the experiment lasted, juveniles did not grow significantly,
either in length or in weight (RAS 1600 %—growth 179.3 lm in length and 8.7 mg in
weight; RAS 800 %—growth 19.3 lm in length and 3.7 mg in weight) (Fig. 6), and no
significant differences were found among treatments and replicates of the same treatment
in the RAS system (K–W; p [ 0.05). Nevertheless, no mortality was observed in the two
flows tested.
A mixed diet composed of T. lutea and C. calcitrans (1:1 in cells/ml) has shown to be a
good ration for R. decussatus juvenile development. The clearance rate of R. decussatus
spat also provides substantial information on the microalgae consumption. Results showed
that the dimensions RAS system were probably undersized for the rearing of R. decussatus
juveniles with the size used in the experiment. Although no imbalance (physical, chemical
or microbiological) in the system has been found, clams did not grow significantly. We

123
Aquacult Int

Fig. 6 Growth in weight (mg) 50

and shell length (mm) of 40
Ruditapes decussatus rearing in
RAS system with two different 30

(mg)
flows of seawater per day (1600 20 1600%
and 800 %)
10 800%
0
0 60
Days of culture

7
6
5
4

(mm)
3
2
1
0
0 60
Days of culture

supposed that the food consumption by the system can be the cause of this failure.
However, the capacity of the system did not allow the introduction of more food.

Conclusions

All four species can be reared in RAS and showed similar growth in RAS and in FTS or in
RAS with a higher renewal rate. RAS can keep O2, nitrogen and pH within the desired
range. Temperature was higher in RAS than in FTS, probably due to the heat induced by
pump circulating the water. The supply of a sufficient amount of food in combination with
a desire to reduce the renewal rate calls for use of concentrated feed in RAS.

Acknowledgments The study was supported by the EU Grant No. 245119 REPROSEED. We would like
to thank Blandine Diss for providing information on the RAS pilot performed at Satmar.

References
Attramadala KJK, Salvesenb I, Xuec R, Øieb G, Størsethb TR, Vadsteind O, Olsena Y (2012a) Recircu-
lation as a possible microbial control strategy in the production of marine larvae. Aquacult Eng
46:27–39
Attramadala KJK, Øieb G, Størsethb TR, Alverc MO, Vadsteind O, Olsena Y (2012b) The effects of
moderate ozonation or high intensity UV-irradiation on the microbial environment in RAS for marine
larvae. Aquaculture 330–333:121–129
Besson M, Komen H, Aubin J, de Boer IJM, Poelman M, Quillet E, Vancoillie C, Vandeputte M, van
Arendonk JAM (2014) Economic values of growth and feed efficiency for fish farming in recirculating
aquaculture system with density and nitrogen output limitations: a case study with African catfish
(Clarias gariepinus). J Anim Sci 92:5394–5405
Blancheton JP (2000) Developments in recirculation systems for Mediterranean fish species. Aquacult Eng
22(1–2):17–31

123
 Aquacult Int

Blancheton JP, Attramadal KJK, Michaud L, Roque d E, Vadstein O (2013) Insight into bacterial population
in aquaculture systems and its implication. Aquacult Eng 53:30–39
Blanco Garcia A, Kamermans P (2015) Optimization of blue mussel (Mytilus edulis) seed culture using
recirculation aquaculture systems. Aquacult Res 46(4):977–986
Dunning RD, Losordo TM, Hobbs AO (1998) The economics of recirculating tank systems: a spreadsheet
for individual analysis. SRAC publication no. 456
Epifanio CE, Srna RF (1975) Toxicity of ammonia, nitrite ion, nitrate ion, and orthophosphate to Merce-
naria mercenaria and Crassostrea virginica. Mar Biol 33:241–246
Faulk CK, Holt GJ (2005) Advances in rearing cobia Rachycentron canadum larvae in recirculating
aquaculture systems: live prey enrichment and greenwater culture. Aquaculture 249:231–243
Fernandez-Reiriz MJ, Labarta U, Albentosa M, Pérez-Camacho A (1999) Lipid profile and growth of the
clam spat, Ruditapes decussatus (L), fed with microalgal diets and cornstarch. Comp Biochem Physiol
B Biochem Mol Biol 124(3):309–318
Frandsen RP, Dolmer P (2002) Effects of substrate type on growth and mortality of blue mussels (Mytilus
edulis) exposed to the predator Carcinus maenas. Mar Biol 141:253–262
Helm MM, Bourne N, Lovatelli A (2004) Hatchery culture of bivalves—a practical manual. In: FAO
fisheries technical paper 471, p 177. Food and Agriculture Organization of the United Nations
Joaquim S, Matias D, Matias AM, Leitão A, Soares F, Cabral M, Chı́charo L, Gaspar MB (2014) The effect
of density in larval rearing of the pullet carpet shell Venerupis corrugata (Gmelin, 1791) in a recir-
culating aquaculture system. Aquacult Res 1–12. doi:10.1111/are.12561
Loosanoff VL, Davis HC (1963) Rearing of bivalve molluscs. Adv Mar Biol 1:1–136
Losordo TM, Hobbs AO (2000) Using computer spreadsheets for water flow and biofilter sizing in recir-
culating aquaculture production systems. Aquacult Eng 23:95–102
Magnesen T, Jacobsen A (2012) Effect of water recirculation on seawater quality and production of scallop
(Pecten maximus) larvae. Aquacult Eng 47:1–6
Martins CIM, Eding EH, Verdegem MCJ, Heinsbroek LTN, Schneider O, Blancheton JP, d’Orbcastel ER,
Verreth JAJ (2010) New developments in recirculating aquaculture systems in Europe: a perspective
on environmental sustainability. Aquacult Eng 43:83–93
Merino G, Uribe E, Soria G, von Brand E (2009) A comparison of larval production of the northern scallop,
Argopecten purpuratus, in closed and recirculating culture systems. Aquacult Eng 40:95–103
Petersen JK, Bougrier S, Smaal AC, Garen P, Robert S, Larsen JEN, Brummelhuis E (2004) Intercalibration
of mussel Mytilus edulis clearance rate measurements. Mar Ecol Prog Ser 267:187–194
Pfeiffer TJ, Rusch KA (2000) An integrated system for microalgal and nursery seed clam culture. Aquacult
Eng 24:15–31
Rico-Villa B, Le Coz JR, Mingant C, Robert R (2006) Influence of phytoplankton diet mixtures on
microalgae consumption, larval development and settlement of the Pacific oyster Crassostrea gigas
(Thunberg). Aquaculture 256:377–388
Robert R, Petton B, Mingant C, Le Souchu P (2013) Effects of flow renewals in a RAS on Crassostrea gigas
larval development and seawater quality. Aqua 2013 ‘‘Making Sense of Science’’. Trondheim (Nor-
way), 9–12 August 2013, (communication)
Suantika G, Dhert P, Nurhudah M, Sorgeloos P (2000) High density production of the rotifer Brachionus
plicatilis in a recirculation system: consideration of water quality, zootechnical and nutritional aspects.
Aquacult Eng 21:201–214
Timmons MB, Ebeling JM (2007) Recirculating aquaculture. Cayuga Aqua Ventures, LLC, New York
Walne PR (1974) Culture of bivalve molluscs: 50 years experience at Conwy. Whitefriars Press Ltd, Fishing
New Books, London
Widman JC (1998) Design of recirculating nursery culture system for the bay scallop Argopecten irradians.
J Shellfish Res 17:364
Xiongfei W, Zhidong Z, Deshang L, Kangmei C, Zhuanshang T, Liegang S, Kaichong X, Bailin G (2005)
Closed recirculating system for shrimp-mollusk polyculture. Chin J Oceanol Limnol 23:461–468
Zohar Y, Tal Y, Schreier HJ, Steven C, Stubblefield J, Place A (2005) Commercially feasible urban
recirculated aquaculture: addressing the marine sector. In: Costa-Pierce B (ed) Urban aquaculture.
CABI Publishing, Cambridge, pp 159–171

123