2nd International Sustainable Marine Fish Culture Conference and Workshop
October 19-21, 2005
Design and Operation of a Zero-Exchange Mixed-Cell Raceway Production System
James M. Ebeling1, Carla F. Welsh1, Kata L. Rishel1, Michael B. Timmons2
1
The Conservation Fund’s Freshwater Institute, 1098 Turner Rd., Shepherdstown, WV 25443, USA
e-mail: j.ebeling@freshwaterinstitute.org
2
Department of Biological and Environmental Engineering, Cornell University,
Ithaca, NY 14853, USA e-mail: mbt3@cornell.edu
1.0 Introduction
The use of the zero-exchange systems has become a viable alternative to traditional pond
methods of intensive aquaculture production. Recently, the concept of zero-exchange systems
has been applied to indoor tank based production systems for marine shrimp, Penaeus vannamei.
In addition, raceways have been used for years for the production of a wide range of aquaculture
species, but their use has been significantly curtailed by the need for high volumes of high
quality water and difficulties treating their effluent discharge. An improvement to the
conventional raceway design is to convert the raceways into a series of counter-rotating mixedcells that act as hydraulically separated round tanks that have the inherent advantage of being
self-cleaning. Combining these two production concepts, yields a system that is economical to
construct, efficient to harvest, has excellent solids management abilities, and high potential
product (or biomass) yield.
Recently, a new production strategy has emerged called intensive zero exchange systems. In
these systems, ammonia-nitrogen is controlled by the manipulation of the carbon/nitrogen ratio
in such a way as to promote the growth of heterotrophic bacteria (Avnimelech, 1999, McIntosh,
2001). As a result, the ammonia-nitrogen is removed from the system through assimilation into
microbial biomass. As a bonus, for some aquaculture species (marine shrimp and tilapia), this
bacterial biomass produced in the intensive zero-exchange systems can be an important source of
feed protein, reducing the cost of production and thus improving the overall economics (Moss,
2002).
Raceways have been used for years for the production of salmonids and other species by federal
and state agencies for stocking purposes and commercial growers. Where there are large ground
water resources, raceways are the most common rearing tank design being used to grow the
majority of rainbow trout produced in the USA. One significant advantage of raceways is their
better utilization of floor space and easier handling and sorting of fish as compared to traditional,
circular tanks. Their primary disadvantages are the large volume of water required, high
turnover rates, and limited self-cleaning ability (Timmons et al. 1998). Raceways depend upon
high turnover rate to maintain acceptable water quality. Water enters the raceway at one end and
flows through the raceway in a plug-flow manner. As a result, the best water quality is found at
the head of the raceway where the water first enters and then it is deteriorated steadily towards
the raceway outlet. Because of the low velocities through the raceway (2-4 cm/s), removal
efficiency of settled solids is very poor, requiring frequent cleaning and maintenance tasks
(Timmons et al. 1998). This is due to the hydraulic design being based on oxygen design
2nd International Sustainable Marine Fish Culture Conference and Workshop
October 19-21, 2005
requirements, rather than cleaning requirements, that result in the much lower velocities. In
practical terms, raceways are incapable of producing the optimum water velocities recommended
for fish health, muscle tone, and respiration (Timmons et al. 1998). Even using lower exchange
rates and lower velocities, use of raceways is being severely limited due to the unavailability of
large quantities of high quality water, increased concern about their environmental impacts on
receiving waters, and the difficulty in treating large flows from effluent discharge. One solution
to this problem is to convert raceways into a series of counter-rotating mixed-cells (Watten et al.
2000). The concept of a mixed-cell raceway was first proposed by Watten et al. (2000) to
eliminate metabolite concentration gradients, increase current velocities, and improve solids
scour at low water exchange rates. Watten modified a standard raceway section 14.5 m long and
2.4 m wide by creating a series of six counter-rotating cells, each 2.4 m by 2.4 m. A series of
vertical pipe sections with jet ports were installed in the corners of each cell and water directed
tangentially to create rotary circulation. Water was withdrawn from centrally located floor
drains. Mixed-cell raceways can be managed as either partial reuse systems (Summerfelt et al.
2004) or intensive recirculation systems, allowing for substantial increase in production
(Timmons, et al., 2002).
2.0 Design Concept
The mixed-cell raceway acts as a series of hydraulically separated round tanks. The basic design
concept of the mixed-cell raceway (Watten et al. 2000) is to operate it as a series of adjacent
counter rotating square/octagonal tanks, each having a center drain for continuous removal of
solids and sludge, Figure 1. Early research on mixed-cell raceways by Watten et al. (2000)
examined their use in retrofitting existing raceways at federal and state hatcheries and was
reflected in the overall small systems size, 22.7 m3. In contrast, this works started with the
design of a small commercial production system, 108 m3. In addition, Timmons, et al., (1998)
recommended tank diameter to depth ratios for good self-cleaning capability from 5:1 to 10:1,
compared to 3.7 for Watten et al. 2000 vs. 5.5 in this current study.
Sludge
Pu
mp
Pu
mp
Pu
mp
SR
TANK
Sump &
Settling
Harvest
Cage
Pu
mp
Pu
mp
Pu
mp
captur
e
Figure 1. Mixed-cell raceway layout and flow pattern.
A prototype raceway was constructed in one bay of an existing greenhouse at The Conservation
Funds Freshwater Institute with approximated dimensions of 16.3 m x 5.44 m x 1.22 m (54 ft
x18 ft x 4 ft), which created three mixed-cells. Each cell received water from four vertical
manifolds (downlegs) extending to the raceway floor and located in the corners of each cell and
2
2nd International Sustainable Marine Fish Culture Conference and Workshop
October 19-21, 2005
at the intersection between adjacent cells (Figure 2); four of the manifolds supply water to two
cells concurrently. Water is pumped through several orifice discharges (or jet ports) from each
of the downleg pipes to establish rotary circulation in the cell, with adjacent cells rotating in
opposite directions. Each cell had a bottom drain located at the center of the cell connected to a
drain line, which discharged solids and sludge to a settling sump. A small fraction of the total
recirculated flow, e.g., 10 to 20%, is withdrawn from this sump and returned to the raceway,
creating a “Cornell” dual drain system (Timmons et al. 1998).
Usually an engineering design for an aquaculture system starts with the water exchange rate
required for the production tank, usually determined by the dissolved oxygen requirements based
on loading densities or some other limiting parameter. For this design analysis, a water
exchange rate of 0.5, 1.0, and 2.0 exchanges per hour was chosen, which would correspond to a
zero-exchange low biomass density (5 to 10 kg/m2), a moderate biomass density (30 to 50 kg/m3)
and a high biomass density (50 to 100 kg/m3). The mixed-cell raceway had a volume of 88.8 m3
(23,500 gallons), using a water depth of 1.0 m (3.3 ft). The required total flow rate was 0.74
m3/min (195 gpm), 1.48 m3/min (390 gpm), and 2.95 m3/min (780 gpm) for 0.5, 1.0, and 2.0
exchanges per hour. This included a withdrawal rate from the center drains 10 to 15% of the
total flow.
The rotational velocity in the cells can be controlled by the design of the orifice discharges,
either by increasing the flow rate, the discharge velocity, or the total number of orifices (Ebeling,
et al., 2004, Labatut, et al., In press). Research has shown that for round tanks, the water
rotational velocity is fractionally proportional to the discharge velocity from the orifices in the
water inlet structure. Timmons et al. (1998) reported that this proportionality constant between
tank rotational velocities is generally 15 to 20% of the inlet velocity, although this applied to
smaller diameter round tanks. Watten et al. (2000) reported that the grand mean of current
measurements corresponded to 3.7% of the inlet jet velocities, calculated based on orifice
diameter and head. This was substantially smaller than the value reported for round tanks and
was attributed to the increased drag associated with forced distortion of circulating cells within
the raceway rectangular boundaries (Watten et al. 2000).
2.0 Materials and methods
2.1 Construction details
The mixed-cell raceway was constructed as an above ground tank with a width of one
greenhouse bay. This resulted in a raceway with dimensions of 16.3 m x 5.44 m x 1.22 m (18 ft
x 54 ft x 4 ft). Sidewall modules (1.22 m x 2.44 m) were prefabricated of 2x6 construction studs
on 0.61 m (24 in) spacing and covered with 12 mm (½ in) plywood sheeting. These sidewall
modules were supported on a 6x6 treated wood beam ‘foundation’ and connected together with
12 mm (½ in) lag bolts. In addition, a top plate was added to the sidewall modules to provide
additional rigidity, see Figure 2. Normally such a raceway would be constructed below grade,
allowing for structural support of the walls with backfill material. To provide this structural
3
2nd International Sustainable Marine Fish Culture Conference and Workshop
October 19-21, 2005
manometer
5 cm unions
0.75 kW pump
top plate
5 cm vertical manifold
5 cm gate
valve
Water Depth
check valve
threaded bushing 3.18 cm by 2.54 cm
with a 2.54 cm threaded plug - drilled
10 cm manifold
1.22 m
15 cm drain with 5 cm orifice
20 ml HDPE liner
Sidewall modules
15 cm drain line
5.44 m
Figure 2. Cross-section of mixed-cell raceway showing construction details, pipe manifolds,
vertical manifolds and drain details.
support, a series of polypropylene-impregnated wire ropes were run across the top of and below
the raceway at five equally spaced intervals along the sidewalls. In addition, a single cable was
strung the length of the raceway on the top at the center and two cables were strung lengthwise
below the insulated floor. These cables were secured into the sidewall top plates and the 6x6
treated wood beam foundation with 18 mm (3/4 in) diameter eyebolts and forged galvanized
steel hook and eye turnbuckles to allow adjusting and tightening. Lighter gauge materials and
less reinforcing were found by experience to be inadequate to prevent wall movement.
The floor of the raceway was covered with 5 cm of fine sand and graded to provide a 2% slope to
the three center drains. Walls were insulated with 2.54 cm x 1.22 m x 2.44 m (1 in x 4 ft x 8 ft)
of foam insulation board to minimize heat loss through the sidewalls. The outside perimeter of
the floor was covered with 5.0 cm x 1.22 m x 2.44 m (2 in x 4 ft x 8 ft) of insulation board and
the center strip of the floor with 2.54 cm x 1.22 m x 2.44 m (1 in x 4 ft x 8 ft) of insulation board.
All insulation board was expanded polystyrene rigid smooth finish (Rvalue of 0.9 m2K/W per 2.5
cm of thickness). The raceway was lined with a 20 ml high-density cross-laminated
polyethylene (HDPE) raceway liner from Permalon, Reef Industries, Inc.
A 15.24 cm (6 in) drainline with three outlet drains (tee fittings) centered on each of the three
cells was buried along the longitudinal axis of the raceway. A standard flange socket fitting was
modified by boring out the center to allow either a standpipe or a screened inlet and installed on
each of the three tee fitting outlets. A concentric ring of PVC sheet materials was used to secure
the liner to the flange flat surface to provide a water tight seal at each drain. To provide for
uniform bottom withdrawal from each of the three cells, a 5.0 cm (2 in) orifice plate was
installed in each flange to create a restricted outlet flow. This was designed based on a desired 5
to 15% bottom withdrawal rate and a 30 cm (1 ft) head across the orifice to equalize flow rates.
A 0.6 m diameter circular plate was installed approximately 2.5 cm (1 in) above the orifice to
force water to move parallel with the bottom into the outlet drain lines. The three drains
discharged into a 1.83 m x 1.83 m x 1.83 m (6 ft x 6 ft x 6 ft) fiberglass sump tank. The sump
tank had both a standpipe for water level control and a drainline to flush the system. The sump
tank was intended to fulfill several roles, including solids management by acting as a solids
4
2nd International Sustainable Marine Fish Culture Conference and Workshop
October 19-21, 2005
settling basin, water level set point with the standpipe, and as a harvesting basin by flushing the
production raceway through a screened harvesting cage.
Eight 0.75 kW pumps (1 hp) were installed along the outside walls on platforms and discharged
into a 10 cm (4 in) schedule 40 PVC manifold that encircled the entire raceway. These pumps
effectively simulated a dual-drain system, where most of the recycled water is removed from the
top sides of a tank and only a small fraction is removed from the center drains. Two of the
pumps were located on the sump collection tank, while the remaining six were placed equal
distant along the length of the raceway outside walls. The inlet of the suction lines with check
valves were located approximately 75 cm (30 in) from the floor of the raceway. The pump
discharges were connected to the manifold with a 5.08 cm (2 in) flexible PVC hose with a bronze
gate valve to control flowrate. Later, the 10 cm manifold was replaced with a 7.6 cm (3 in)
schedule 40 PVC manifold placed on top of the wall to reflect a commercial scale operation.
Each hydraulically separated cell created within the raceway measured approximately 5.44 m x
5.44m (18 ft x 18 ft) as determined by the downleg placements. Four 5.08 cm (2 in) vertical
manifold downlegs were placed in the four corners of the raceway and four 7.62 cm (3 in)
vertical manifold downlegs were located along the sidewalls constructed of PVC schedule 40
pipe, dividing the raceway into three equal cells. Two of the 7.62 cm (3 in) vertical manifold
downlegs had 10 orifices; five discharging parallel to the walls in opposite directions and the
other two vertical manifold discharge downlegs had five orifices each and discharged
perpendicular to the raceway walls. The orifices were spaced 15.24 cm (6 in) apart on the
vertical manifolds, starting 5.0 cm off the bottom. Orifice openings were constructed by welding
a threaded bushing 3.18 cm by 2.54 cm (1 ¼ x 1 in), which allowed a 2.54 cm (1 in) threaded
plug to be inserted. This allowed for easy modification of the orifice sizes and plugging of
unused orifice openings, Figure 3. A clear section of rigid tubing was attached at the top of the
downleg to act as a pieziometer tube manometer to measure vertical manifold backpressure.
Figure 3. Vertical manifolds arranged along the sidewalls of the mix-cell raceway. On the far
right, a single-sided manifold composed of five nozzles directed across the width of the raceway.
Middle and far right pictures show the double-sided manifolds composed of ten nozzles directed
tangentially to the raceway wall.
5
2nd International Sustainable Marine Fish Culture Conference and Workshop
October 19-21, 2005
2 .2 Analysis
A SonTek Argonaut Acoustic Doppler Velocimeter from Yellow Springs Instruments (1725
Brannum Lane, Yellow Springs, OH 45387 USA) was utilized in this study to measure speed
and direction within the hydraulically separated cells. The SonTek velocity meter is a singlepoint, 3D Doppler current meter that measures water velocity via the Doppler shift in frequency
of sound from a moving object, in this case small particulate matter in water current. The
SonTek probe assembly was mounted on a rigid aluminum beam supported above the width of
the raceway, which allowed the probe to be moved across the raceway width in a repeatable
fashion. The probe was lowered into the raceway to specified depths of 10 cm off the bottom,
mid-depth, and 10 cm below the water surface, along a 0.5 m horizontal-squared-grid measuring
system. Measurements were taken for a 20 seconds averaging interval at each of grid points and
the values averaged for a numerical value used to plot the results. The data collected form the
SonTek system was downloaded into Microsoft Excel for processing and contour graphing of the
velocities was created with Sigma Plot.
3.0 Results
3.1 Zero-exchange low biomass density (5 to 10 kg/m2) - 0.5 tank exchanges/hr
In a zero-exchange system, ammonia-nitrogen is removed via the growth of heterotrophic
bacteria, stocking densities are low (5 to 10 kg /m3) and oxygen requirements are modest, so
high tank exchange rates through biofilters, and oxygenators are not necessary. At an exchange
rate of 0.5 tank volume/hr, a flow rate of only 0.74 m3/min (200 gpm) is required. This was
accomplished using two 0.75 kW (1 hp) pumps which removed water at the two ends of the
raceway and injected it into a 3 inch manifold that circled the top of the raceway. Water was
withdrawn either from a 5 cm (2 in) PVC pipe inlet located approximately 25 cm below the
surface at one end or an end sidewall discharge drain at the other. Approximately 15% of the
flow 0.11 m3/min (30 gpm) was from the three bottom drains, using a smaller 0.375 kW (1/2 hp)
submersible pump in the sump drain. A small diameter orifice (10 mm) was used to provide a
high orifice discharge velocity to insure adequate rotational velocities to move waste particles
and uneaten food to the center drains.
Figure 4 shows the contour plot for cell #3 (end cell) at velocity intervals of 5 cm/sec,
approximately 5 cm off the bottom. The pressure head was approximately 1.00 m gauge (1.5
psi). As can be seen in this figure, relatively high scouring velocities are found at the outside
perimeter of the cell (20 to 24 cm/sec), lower velocities near the center (10 to 6 cm/sec) and very
low velocities at the center of the cell. The velocity profiles shown in Figure 5 were created by
averaging the velocities at each grid point in an annular ring 0.5 m wide starting at the center at
each of the three depths. This graph shows the almost linear velocity profile as a function of
distance from the center drain and significant cleaning velocities in the corners of the cell. The
velocities in the z-direction were very small, just above the drain values of -2.2 to 2.8 cm/s was
measured.
6
2nd International Sustainable Marine Fish Culture Conference and Workshop
October 19-21, 2005
2 cm/sec
4
6
8
10
12
14
16
18
20
22
24
2.5
2.0
1.5
1.0
0.5
0.0
-0.5
-1.0
-1.5
-2.0
-2.5
-2.5
-2.0
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
2.0
2.5
Figure 4. Velocity contour plot for cell #3, with 10 mm orifice diameter, 100 cm pressure head,
recirculating 250 gpm or approximately 0.5 tank exchanges/hr and 15% withdrawal from the
center drains.
20
16.2
14.4
Velocity (cm/s) .
16
13.1
11.1
12
8.2
8
5.6
3.2
4
0.4
0
er
rn
co
7
2.
52.
5
2.
02.
0
2.
51.
5
1.
01.
0
1.
50.
5
0.
00.
r
te
en
C
m
m
m
m
m
m
Figure 5. Cell #3 mean velocity profile with 10 mm orifice diameter, 100 cm pressure head,
recirculating 255 gpm or approximately 0.5 tank exchanges/hr and 15% withdrawal from the
center drains.
7
2nd International Sustainable Marine Fish Culture Conference and Workshop
October 19-21, 2005
During a demonstration zero-exchange production trial, the mixed-cell raceway received from 4
to 5 kg of feed per day, plus supplemental carbon. Figure 6 shows the daily sludge removed
from the sludge tank during this production trial. Daily removal rates of uneaten feed, fecal
matter, and excess heterotrophic bacterial floc ranged from 1 to 8 kg dry weight per day,
demonstrating the solid removal effectiveness of the mixed-cell raceway.
Solids Removed (kg) .
10
8
6
4
2
0
0
10
20
30
40
50
60
70
80
90
100
Days
Figure 6. Solids removed by settling tank for demonstration zero-exchange marine shrimp
production, 4 to5 kg feed per day.
3.2 Moderate biomass density (30 to 50 kg/m3) - 1 tank exchanges/hr
At higher stocking densities, both biofiltration and oxygen requirements demand high
recirculation flow rates. In the research system, this was accomplished by adding two addition
0.75 kW (1 hp) pumps and increasing the orifice discharge diameter to 15 mm. Figure 7 shows
the velocity profiles measured at approximately 5 cm off the bottom. As can be seen, high
cleaning velocities are seen around the outer edges, with low velocities near the center drains.
Figure 8 again shows the linear velocity profile from the center to the outer edges.
8
2nd International Sustainable Marine Fish Culture Conference and Workshop
October 19-21, 2005
2 cm/sec
4
6
8
10
12
14
16
18
20
22
24
26
28
30
32
2.5
2.0
1.5
1.0
0.5
0.0
-0.5
-1.0
-1.5
-2.0
-2.5
-2.5
-2.0
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
2.0
2.5
Figure 7. Velocity contour plot for cell #3, with 15 mm orifice diameter, 120 cm pressure head,
recirculating 450 gpm or approximately 1 tank exchanges/hr and 10% withdrawal from the
center drains.
24
19.6
20
Velocity (cm/s) .
17.0
16
16.1
13.6
10.8
12
7.9
8
3.9
4
2.2
0
1.
01.
2.
51.
2.
02.
2.
52.
0
5
0
5
7
er
rn
co
1.
50.
r
5
te
en
0.
00.
C
m
m
m
m
m
m
Figure 8. Cell #3 mean velocity profile with 15 mm orifice diameter, 120 cm pressure head,
recirculating 450 gpm or approximately 1 tank exchanges/hr and 10% withdrawal from the
center drains.
9
2nd International Sustainable Marine Fish Culture Conference and Workshop
October 19-21, 2005
3.3 High biomass density (50 to 100 kg/m3) – 2 to 4 tank exchanges/hr
Finally at the highest biomass density, high exchange rates are required to provide for significant
dissolved oxygen demand, removal of ammonia-nitrogen and carbon dioxide and solids. This
was accomplished by increasing the orifice diameter to 20 mm and increasing the number of
pumps to seven 0.74 kW (1 hp) pumps. In a commercial application, two or three, low head,
high efficiency axial flow impeller pumps, would replace these pumps. Figure 9 shows the
velocity profile for this configuration and again shows the high cleaning velocities at the
perimeter and the low velocities near the center drain. Figure 10 shows the mean velocity profile
very similar to the others two, although with a slightly higher center drain velocity.
Cell #3 Velocity Profiles at 5 cm off of Bottom
2.5
2.0
1.5
1.0
0.5
5
10
15
20
25
30
35
40
0.0
-0.5
-1.0
-1.5
-2.0
-2.5
-2.5
-2.0
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
2.0
2.5
Figure 9. Velocity contour plot for cell #3, with 20 mm orifice diameter, 135 cm pressure head,
recirculating 760 gpm or approximately 2-tank exchanges/hr and 15% withdrawal from the
center drains.
10
2nd International Sustainable Marine Fish Culture Conference and Workshop
October 19-21, 2005
28
23.5
24
22.3
Velocity (cm/s) .
20.1
20
14.9
16
10.6
12
8
12.0
6.4
4
0
-2
-2
-2
-1
.7
.5
.0
m
m
m
m
m
m
.5
er
rn
co
5
2.
0
2.
5
1.
0
1.
1.0
50.
0.5
00.
Figure 10. Cell #3 mean velocity profile with 20 mm orifice diameter, 135 cm pressure head,
recirculating 450 gpm or approximately 2-tank exchanges/hr and 15% withdrawal from the
center drains.
5.0 Conclusions
In order to better understand the hydraulics of a large, mixed-cell raceway, a prototype raceway
was constructed in a research greenhouse at The Conservation Funds Freshwater Institute. The
raceway was constructed of simple, readily available materials at a low economic cost, allowing
for rapid start-up (or shut down) of production systems. Raceways have several inherit
advantages over circular tanks, including ease of sorting, grading, and handling fish and
optimization of floor space. Results of this study showed excellent bottom velocities for
scouring solids and moving them towards the center drains in each cell of the prototype mixedcell raceway at several different exchange rates. Mixed-cell raceways show excellent potential
for either retrofitting existing raceways or as a design for a new production system.
Acknowledgements
This work was supported by the United States Department of Agriculture, Agricultural Research
Service under Cooperative Agreement number 59-1930-1-130 and Magnolia Shrimp, LLC,
Atlanta, Georgia.
11
2nd International Sustainable Marine Fish Culture Conference and Workshop
October 19-21, 2005
References
Avinimelech, Y., 1999. Carbon/nitrogen ratio as a control element in aquaculture systems.
Aquaculture, 176:227-235.
Ebeling, J.M., Timmons, M.B., Joiner, J.A., Labatut, R.A., 2005. Mixed-Cell Raceway:
Engineering Design Criteria, Construction, Hydraulic Characterization. Journal of North
American Aquaculture, 67(3): 193-201.
Labatut, R.A., Timmons, M.B., Ebeling, J.M., Bhaskaran, R., 2005. Hydrodynamics of a LargeScale Mixed-Cell Raceway (MCR): Experimental Studies. In review: Aquacult. Eng.
Labatut, R.A., Timmons, M.B., Ebeling, J.M., Bhaskaran, R., 2005. Experimental Evaluation of
the Effects of Nozzle Diameter and Effluent Withdrawal Strategy on Tank Hydrodynamics in a
Large-Scale Mixed-Cell Raceway (MCR). In review: Aquacult. Eng.
McIntosh, R.P., 2001. High Rate Bacterial Systems for Culturing Shrimp. In S. T. Summerfelt
et al. (eds). Proceedings from the Aquacultural Engineering Society’s 2001 Issues Forum.
Shepherdstown, West Virginia, USA. Aquaculture Engineering Society, Pp. 117-129.
Moss, S.M., 2002. Dietary importance of microbes and detritus in penaeid shrimp aquaculture.
In Lee, C.-S. and O’Bryen, P. (eds), Microbial approaches to aquatic nutrition within
environmentally sound aquaculture production systems. The World Aquaculture Society, Baton
Rouge, Louisiana, USA
Timmons, M.B., Summerfelt, S.T., and B.J. Vinci. 1998. Review of circular tank technology
and management. Aquacultural Engineering 18:51–69.
Timmons, M.B., Ebeling, J.M., Wheaton, F.W., Summerfelt, S.T., Vinci. B.J., 2002.
Recirculating Aquaculture Systems, 2nd Edition. Cayuga Aqua Ventures, 769 pgs.
Watten, B.J., Honeyfield, D.C., and M.F. Schwartz. 2000. Hydraulic characteristics of a
rectangular mixed-cell rearing unit. Aquacultural Engineering 24:59–73.
12