Aquaculture 264 (2007) 140 – 147

www.elsevier.com/locate/aqua-online

Feeding with microbial flocs by tilapia in minimal discharge

bio-flocs technology ponds
Yoram Avnimelech
Department of Civil and Environmental Eng., Technion, Israel Inst. of Technology, 32000 Haifa, Israel
Received 7 August 2006; received in revised form 24 November 2006; accepted 25 November 2006

Abstract

The uptake of microbial flocs by tilapia was evaluated. Fish (tilapia Mozambique, 107 g) were stocked in 1 m3 tanks filled with
water from a limited exchange intensive tilapia producing pond (bio-floc technology, BFT system). Tagged ammonium nitrogen
(15NH4(SO4)2) and starch to ensure incorporation of the 15N into the bio-flocs, were added. Fish were held in the tanks for ca
2 weeks, not fed during a week period, when the only source of feed was microbial flocs. Floc volume, total suspended solids, as
well as total carbon and nitrogen in suspension were monitored.
The floc plug in settling cones contained 1.4% as dry matter.
Feed uptake, evaluated through the decrease with time in respect to 4 independently determined parameters, namely floc
volume, TSS, C and N in suspension, was found to be 10.3 ± 1.0 g kg fish− 1 day− 1, averaged for the computed values for these
parameters However, this preliminary evaluation was based on the assumption that fish harvesting is the only mechanism affecting
bio-flocs mass, neglecting biodegradation and production of flocs. Gross daily uptake of nitrogen as determined using 15N uptake
data was 0.25 gN kg− 1 (1.6 g protein), equivalent to the daily uptake of 6.2 g kg− 1 of dry bio-flocs, 60% of that computed by the
simplified mass balance approach. This difference may be attributed to microbial degradation of the bio-flocs.
Even the lower flux as evaluated through 15N uptake, constituted, in the specific case studied, almost 50% of conventional feed
ration.
© 2006 Elsevier B.V. All rights reserved.

15
Keywords: Microbial control; Intensive ponds; N uptake; Flocs; Bio-flocs technology

1. Introduction the water. As a result, intensive aquaculture industry

Intensive aquaculture systems are used to efficiently
produce dense biomasses of fish or shrimp. An intrinsic
feature of these systems is the rapid accumulation of
feed residues, organic matter and toxic inorganic ni-
trogen species. This cannot be avoided, since fish retain
only 20–30% of feed nutrients (e.g. Avnimelech and
Lacher, 1979; Avnimelech and Ritvo, 2003; Boyd,
1985). The rest is excreted and typically accumulates in
E-mail address: agyoram@tx.technion.ac.il.
0044-8486/$ - see front matter © 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.aquaculture.2006.11.025
faces two major problems. The first is the water quality
deterioration caused by the high concentrations of me-
tabolites and the second is the low feed utilization in
cases when high water exchange, within or outside the
pond system, is practiced.
The principles of growing fish or shrimp in limited
water exchange intensive ponds were developed simul-
taneously for shrimp in the Waddel Mariculture Center in
the USA and for fish, mostly tilapia, in Israel (Avnimelech
et al., 1989; Avnimelech et al., 1994; Hopkins et al., 1993;
Chamberlain and Hopkins, 1994) and practiced in the
 Y. Avnimelech / Aquaculture 264 (2007) 140–147
USA (Serfling, 2000), in the beginning of the 1990's. Bio-
floc technology, BFT, (called also active suspension
ponds, heterotrophic ponds, green soup and other terms)
was first developed to solve water quality problems.
Water quality management is based upon developing and
controlling dense heterotrophic bacteria within the culture
component. This goes in tandem with zero or minimal
water exchange rate. A microbial community develops,
reaching a density in the order of 107 colonies forming
units, CFU, ml− 1 (Burford et al., 2003), forming bio-flocs
that contain bacteria, other micro-organisms, protozoa
and zooplankton.
Accumulation of toxic inorganic nitrogen species
(NH4, NO2) is prevented in bio-flocs system by main-
taining a high C/N ratio and inducing the uptake of
ammonium by the microbial community (Avnimelech
et al., 1994; McIntosh, 2000). This process was quan-
titatively formulated (Avnimelech, 1999), verified and
practiced by farmers world-wide (Browdy et al., 2001;
Panjaitan, 2004).
A by-product of this is the growth of the microbial
community and the production of microbial protein.
Harvesting of the bio-flocs by fish serves as an addition
of high value feed, recycling of the non-utilized fraction
of the feed and was shown to double the utilization of
protein and feed by fish or shrimp (Avnimelech et al.,
1989; McIntosh, 2000; Velasco et al., 1998).
The fast spread and the large number of BFT shrimp
farms induced significant research effort of processes in
BFT shrimp production systems (e.g Browdy et al.,
2001; Burford et al., 2003; Taw, 2005). Yet, practically
no work was done in BFT fish production units.
The biomass in fish producing BFT is in the range of
10–40 kg m− 3, as compared to a range of 1–2 kg m− 3
for shrimp. It is expected that process rates and
concentrations in the fish ponds will be higher than in
shrimp ponds and easier to evaluate. Thus, besides its
importance for fish culture, research in such ponds can
reveal processes that are more difficult to evaluate in
shrimp ponds.
It is clear that bacteria, metabolizing carbohydrates,
take up dissolved inorganic nitrogen (mostly NH4) and
produce protein. Microbial flocs of different sizes can be
taken up by fish or shrimp and serve as a feed source
(Avnimelech et al., 1989; Burford et al., 2004; Tacon
et al., 2002) However, a quantitative and predictive study
of this is very complicated. A number of processes are
taking place simultaneously, formulated schematically as:
d½BF=dt ¼ BFproduction −ðBFharvesting þ BFdegradation Þ
ð1Þ
141
Where d[BF] / dt is the bio-flocs concentration
change with time, as affected by production, harvesting
by fish and biodegradation. Each of the processes given
in Eq. (1) is complex and depends on a variety of
factors:
a. Production of bio-flocs depends on the supply of
organic substrates to the microbial community, both
external sources (feed supply, algal activity) or by the
excretion of un-utilized feed components by fish. In
addition bio-flocs production most probably depends
on the quality of the added substrates, its C/N ratio,
bio-availability and other factors.
b. Uptake of the bio-flocs by fish depends most
probably on the fish species and feeding traits, fish
size, floc size and floc density. It is possible that bio-
flocs harvesting depends also on the presence and
rate of formulated feed added to the pond. In
addition, feed eaten by the fish may be utilized and
accumulate in the fish, or excreted and serve as a
substrate for the production of more bio-flocs.
c. Biodegradation of the floc depends on the microbial
community associated with the bio-flocs, be it
bacteria, protozoa or others.
d. Finally, all of these processes may be affected by
environmental and operational conditions such as
temperature, water salinity, water exchange rate
(affecting floc mean residence time, FMRT), mixing
intensity and many other.
The study of this complex system is difficult, since
it is practically impossible to study all factors
simultaneously and since the monitoring of any of
the processes given in Eq. (1) is technically difficult.
We consider this work as a preliminary work aimed at
the development of methodology and conceptual
approach toward the quantitative evaluation of bio-
flocs technology.
The goal of this work was to reveal some of the
processes involved in the bio-floc system, to get some
quantitative data on harvest of flocs by tilapia and to
evaluate some methods that can be used in future
research and monitoring. This work is based, on one
hand, upon the evaluation of the material balance of
suspended components along a two weeks experimen-
tal period. The second approach was to evaluate
nutrients uptake from flocs through the evaluation of
15
N tagging of the flocs and its uptake by fish. The 15 N
uptake was used as a means to evaluate the
contribution of microbial protein to the nutrition of
shrimp by Parker and Anderson (1989), Cam et al.
(1991) and Burford et al. (2004). Avnimelech et al.
142 Y. Avnimelech / Aquaculture 264 (2007) 140–147
(1989) used the distribution of the naturally occurring
13
C in feeds and fish to evaluate the contribution of
microbial proteins to fish nutrition.
2. Materials and methods
2.1. Facility
The experiment was conducted in Pacific Aquafarms
in summer 2004, using water from a pond in that farm.
Pacific Aquafarms is a commercial farm located near
Salton Sea, in the Imperial Valley, CA. The farm consists
of about 20 grow-out ponds (ca 1000 m2 each). The daily
water exchange rates used in most ponds was 20%.
However, limited exchange BFT ponds (daily exchange
ca 5%) were initiated and the tendency is to change all
production ponds toward BFT systems. Water for the
ponds is extracted from a saline thermal water aquifer.
Water temperature at the source is 60 °C and its salinity is
3 ppt. All ponds are fully lined with soil cement mixture,
and have an average water depth of 100 cm.
The research was conducted using water from pond
T2, a pond that also served as a reference pond. The
pond was stocked on May 11, 2004 with 41,000 tilapia
Mozambique with an average weight of 47 g each.
Limited water exchange of about 6% per day was
practiced. The pond was aerated with paddlewheels
(13 HP) creating a circular current. Sludge that
accumulated in the pond centre, as a result of deposition
of waste products of feeding and other detritus, was
removed daily via a central drain.
A 20% protein feed was used in the pond, to
minimize ammonia accumulation and maximize protein
utilization (Avnimelech, 1999). Total ammonium nitro-
gen (TAN) concentration in the water was determined
twice a week. Starch, (pre-gelatinized wheat starch), at a
rate of 45 kg was added daily whenever TAN
concentration rose to values of above 7.0 mg l− 1.
2.2. Experimental system and details
Three 1 m3 fiber-glass tanks were placed near pond
T2 and filled on September 14, 2004, by pumping pond
water (800 l) into the tanks. Pond water con-
tained ammonium and nitrate-N at levels of about 5
and 40 mgN l− 1, respectively. Air stones were placed to
ensure proper aeration and mixing of the water. In ad-
dition, tanks were manually mixed twice daily, bottom
sediments resuspended and walls cleaned, using a mop,
to prevent the accumulation of organic-rich and oxygen-
poor environments and to ensure mixed and uniform
water body before sampling.
Five g 15NH4Cl (60% 15N enrichment), dissolved in
water was distributed to the 3 tanks. Nitrogen addition
amounted to about 0.5 mg l− 1. One hundred g starch in
suspension was added to each tank to enhance the
assimilation of the background and the added ammoni-
um into the microbial biomass. Two days later an
additional portion of 50 g starch was added to each tank.
The amount of starch added was in excess of theoretical
amount needed to immobilize the added ammonium, to
ensure a complete immobilization into the microbial
biomass. TAN concentration following the addition of
the starch was maintained at low levels (0.1–0.2 mgN
l− 1) all along the experimental period. Twenty fish at an
average weight of 107 g were introduced to each tank on
September 17. The experiment lasted for 2 weeks. No
fish mortality was detected.
Fish were not fed during a 6 days period along which,
the only feed source was the microbial suspension. A
portion of 100 g feed was added on September 23,
followed by daily additions of 30 g (ca 2% of fish body
weight), dropping to 25 g according to the decrease of
fish biomass due to sampling. Feed, as 20% protein
floating pellets, was distributed 3 times a day.
Oxygen and temperature were determined daily.
Oxygen level was on the average 8.2 mg l− 1 and did not
drop to values below 6 mg l − 1 . Average water
temperature was 26.2 °C. Tanks were covered with a
shade net to prevent excessive heating. Floc volume
(FV) was determined, in most cases, daily. Fish were
sampled at stocking and then every 3 days, 2 per tank, 6
replicates in total. Fish were weighed and chilled. Head
and fins were removed prior to grinding (guts were not
removed). Ground fish samples were then dried to
constant weight and kept frozen. Samples of fish were
placed at 60 °C for 2 days to determine dry weight.
Water was also sampled at the same days to
determine soluble nitrogen species, total suspended
matter, suspended nitrogen and carbon and 15 N enrich-
ment in the suspended matter.
2.3. Analyses
FV was determined by sampling 1000 ml pond water
into a series of Imhoff cones (Eaton et al., 1995). The
volume of the floc plug accumulating on the bottom of
the cone was determined 15 min following sampling. A
period of 10–15 min was long enough to get a stable FV.
Keeping the sample for longer periods of time led to the
formation of gas bubbles in the floc plug and floatation
phenomenon.
The apparent dry density of the plug was evaluated
following a careful siphoning of the overlying water
 Y. Avnimelech / Aquaculture 264 (2007) 140–147
with minimal disturbance of the plug. The remaining
suspension was transferred to a bottle wherein the
volume was adjusted to 100 ml. Total Suspended Solids
(TSS) in the suspension were determined and the
original volume of the floc plug and its dilution were
considered to calculate dry matter concentration in the
original plug. TSS were determined by filtration of a
sample of pond water or floc plug suspension through a
pre-weighed glass filter (GFA). The filter was re-
weighed following oven drying at 100 °C up to a
constant weight.
Filtered water samples were frozen and later used for
a colorimetric determination of TAN, NO2 and NO3,
using an auto-analyzer, following standard methods
(Eaton et al., 1995). Homogenized water samples were
filtered on 2.3 cm diameter GFA, dried and sent to a
laboratory (UC Santa Barbara Marine Science Inst.
Analytical Laboratory) to determine total suspended
carbon and nitrogen, using a CHN analyzer. Similar
samples were sent to UC Davis Stable Isotope Facility
for 15N enrichment determination.
3. Results and discussion
3.1. Fish growth and feeding in pond T2
Fish growth was followed at fortnight intervals in
pond T2, the commercial reference pond. Average
specific daily fish growth along the growing season was
1.39%, better than that in control ponds. Fish in the
control pond responded to feed application by jumping
toward the floating pellets. The floating pellets dis-
appeared within a period of less than 10 min. Feed
application to the bio-flocs technology pond (BFT)
(with fish of about the same size and density), did not
lead to a dramatic response. Fish slowly nibbled at the
feed that remained floating on the water for an average
period of about 30 min. Fish in BFT pond T2 grew
better than those in the control ponds, thus the slow
feeding does not seem to be due to stressed fish but
probably stemmed from the fact that fish ate perma-
nently, harvesting the microbial flocs.
3.2. Floc volume
The volume of flocs settled in the Imhoff cones (FV) as
a function of time, averaged for the 3 replicated tanks, is
shown in Fig. 1. The FV increased during the 2 days prior
to fish stocking, a period when rather high levels of starch
were added to the tanks. (The value in day 0, when fish
were stocked was probably too high, since the FV was
evaluated shortly following starch addition). Subsequent-
143
ly, there was a clear diminution in the FV in the first
6 days, when feed was not given, from about 30 ml l− 1 to
20 ml l− 1. The daily change of FV (omitting the probably
erroneous point of day zero) was:
FV ðml l−1 Þ ¼ 31−1:714 t; R2 ¼ 0:686 ð2Þ
Where t is the time in days.
From day 6 onward, FV rose up to an average level of
about 27 ml l− 1.
Ten replicated samples of settled flocs were collected
on days 6 and 12 of the experimental period. The
average concentration of dry solids in the floc plug was
14.2 ± 4.3 mg l− 1, i.e. the floc plug contained 1.4% dry
solids. Using these results, the daily change of FV can
be given as a change in the dry weight of the flocs, a
daily drop of 24.3 mg l− 1.
3.3. TSS, suspended carbon and suspended nitrogen
Total suspended matter, suspended carbon and sus-
pended nitrogen concentrations along the experimental
period are given in Figs. 2, 3 and 4, respectively and in
Table 1. Both TSS, carbon and nitrogen concentrations
decreased during the non-fed period. An increase of these
components started slowly from day 9 onward.
TSS dropped during the no-feed part of the experiment
from an initial value of 582 mg l− 1 to 460 mg l− 1 in day 6,
i.e. a daily decrease of about 20 mg l− 1. Later on, when
feed was applied, TSS increased up to 643 mg l− 1.
The decrease of suspended carbon and nitrogen
during the non-fed period followed a linear pattern:
Suspended C ðmgl−1 Þ ¼ 168−6:61 t R2 ¼ 0:986
ð3Þ
Suspended N ðmgl−1 Þ ¼ 24:8−0:868 t R2 ¼ 0:987
ð4Þ
Where t is the time in days.
The average carbon content of the suspended solids
was 24.6(± 4)% and that of suspended nitrogen was 3.7
(± 0.55)%. The average C/N ratio in the suspended
matter was 6.6, a rather uniform value all along, with a
coefficient of variation (CV) of only 2.3%.
The drop of suspended carbon and nitrogen can be
expressed as equivalent TSS units by dividing the carbon
or nitrogen daily change by their appropriate concentra-
tions in the TSS. The daily TSS decrease, calculated using
the daily rates of carbon and nitrogen decrease is equal to
−26.9 or −23.5 mg l− 1, respectively.
144 Y. Avnimelech / Aquaculture 264 (2007) 140–147
Fig. 1. Floc volume (FV, ml/l) in the experimental tanks as a function of
time from fish stocking. [Error bars represent standard deviation (n=3)].
By using a common denominator, TSS, to relate the
concentration changes of the 4 components it was
possible to compare the 4 fluxes. The daily changes in
concentration of suspended matter (dry weight) calculated
from the change of FV, TSS, C and N, attributed to the
uptake by the fish are given in Table 2. The daily loss of
material was calculated by multiplying the daily concen-
tration change (mg l− 1 day− 1) by the volume of water in
the tank (800 l). The uptake attributed to fish was
calculated by the dividing these values by fish weights in
the tanks for the non-fed period:
−1
Feed retained by fish ðg  kgfish Þ
¼ DS  V  FW−1 ð5Þ
Where ΔS is the daily change in concentration of FV,
TSS, C or N (normalized to TSS equivalents), V is the
volume of the tank (800 l) and FW is the weight of fish in
each tank.
The four different estimates were found to be very
similar. The daily uptake by 1 kg fish was, on the average,
10.28 + 1.03 g, as dry suspended matter.
Fig. 2. Total suspended solids (TSS, mg/g)) in the experimental tanks
as a function of time from fish stocking. [Error bars represent standard
deviation (n = 3)].
Fig. 3. Suspended carbon, mg/g, in the experimental tanks as a
function of time from fish stocking. [Error bars represent standard
deviation (n = 3)].
3.4. Changes with time of 15N enrichment in suspended
solids and fish
The enrichment of 15N (%) in the suspended solids,
and its dependence on time, are given in Fig. 5. A
logarithmic decrease with time (t, days) of the 15N
enrichment in the suspended matter, was detected:
%15 N ðflocsÞ ¼ 1:614−0:2033 ln t R2 ¼ 0:9341
ð6Þ
15
The rate of increase with time of N in the fish
tissues (Fig. 5), followed:
%15 NðfishÞ ¼ 0:3728 þ 0:011 lnðtÞ R2 ¼ 0:9588 ð7Þ
The 15N enrichment in the first day of the experiment,
0.37%, represents the natural 15N enrichment of fish
before the exposure to the tagged suspension.
The measured, net uptake, of 15N by the fish, (Fig. 6),
was calculated by multiplying the change in %15N by the
nitrogen concentration in the fish (40 g N kg− 1). The net
Fig. 4. Suspended nitrogen, mg/g, in the experimental tanks as a
function of time from fish stocking. [Error bars represent standard
deviation (n = 3)].
 Y. Avnimelech / Aquaculture 264 (2007) 140–147 145

Table 1
Suspended solids, suspended carbon and nitrogen concentrations, averages of data along the experiment
Day TSS Carbon Nitrogen Carbon in TSS Nitrogen in TSS C/N
mg/l, (SD) mg/l, (SD) mg/l, (SD) % %
0 582 (13) 163 (7) 24 (0.8) 28.0 4.1 6.75
3 555 (22) 146 (6) 22 (0.8) 26.3 3.9 6.67
6 460 (53) 129 (2) 20 (0.3) 28.1 4.3 6.58
9 612 (40) 119 (25) 19 (3) 19.5 3.1 6.34
12 643 (20) 137 (20) 21 (2.8) 21.3 3.2 6.68
Average 24.6 3.7 6.6
Standard deviation 4 0.55 0.15
CV % 16.2 14.8 2.3

uptake of 15N is assumed to be equal to the gross uptake
minus the subsequent excretion of 15N. The measured
uptake as above represents the net change. To evaluate the
gross 15N uptake, it was assumed as an approximation,
that there was no excretion of newly consumed 15N
during the first day. Since there was probably some
excretion during day 1, our calculated gross uptake is a
low estimate. Further, it was assumed that the 15N uptake
is proportional to the TSS, the concentration of flocs in the
water, multiplied by the 15N enrichment of the suspended
solids, both lowered with time. The product of the two
parameters, TSS and % 15N, defining the 15N stock in the
water, was calculated for each day. Multiplying the 15N
uptake in day 1 (GU1) by the 15N stock in day n divided
by the stock at day 1, gave the assumed gross 15N uptake,
GU(n):
Assumed GU ðnÞ ¼ GU1  ð15 N stockday nÞ
=ð15 N stockday 1Þ
The difference between the net and the assumed
gross uptake represents the 15N excretion. The excretion
of 15N varied only slightly with time (Fig. 6).
3.5. Combining the different uptake indicators
The changes in concentrations of several components
suspended in the water-FV, TSS, suspended carbon (C)
and suspended nitrogen (N), were determined. These
four components were determined independently (C and
N were determined in the same samples), and thus each
represents an experimentally independent estimate. The
number of points to determine the rate of decrease with
time of these parameters was limited, yet, the consis-
tency of all independent determinations adds strength to
the results.
The decrease of these parameters, during a week
without feeding, when the sole source of feed was the bio-
flocs suspended material, was considered as a preliminary
means to estimate bio-flocs uptake by fish.. At this stage it
was assumed that the bio-flocs are not bio-degraded or
produced. This certainly is just a first approximation.
A different approach to evaluate uptake of microbial
proteins by the fish was through the accumulation of 15N
in the fish. Since the measured accumulation is a function
of both uptake and excretion, we used here the 15N uptake
during the first day, 0.455 mg kg− 1, as the best estimate of
daily 15N uptake. This is probably a low estimate of the
ð8Þ true uptake flux, since some excretion could take place
even in the first day. The total computed N uptake by fish
for the first day, derived from the 15N uptake data, was
254 mgN kg− 1. By dividing this value by the nitrogen
concentration in the suspended matter at that period,
4.1%, we find that daily bio-floc uptake by fish was 6.2 g
kg− 1. This value is about 60% of the attributed TSS
uptake by fish, 10.3 ± 1.03, computed from the average for
the diminution of FV, TSS, N and C.
The fact that we got a smaller value using the 15N
uptake is reasonable, both due to the fact that the 15N
uptake was a low estimate and that diminution of bio-

Table 2
Daily decrease of TSS, FV, suspended carbon and nitrogen, during the no-feed period
TSS Floc volume Carbon Nitrogen
Daily measured change 20 mg/l 1.74 ml/l 6.61 mg/l 0.87 mg/l
Equivalent dry SS change (mg/l) 20 24.3 26.9 23.5
Daily uptake by fish as equivalent SS (g/kg fish) 8.91 10.79 11.03 9.66
Changes in equivalent suspended solids (SS) concentration (as dry weight) and daily uptake attributed to 1 kg fish (107 g average fish weight).
146 Y. Avnimelech / Aquaculture 264 (2007) 140–147
Fig. 5. 15N enrichment (%) of suspended matter and fish as a function
of time from fish stocking. [Error bars represent standard deviation
(n = 3)].
flocs was attributed only to fish harvesting, disregarding
microbial degradation. Considering these un-certainties,
the differences between the values obtained using the
different methods are not too far apart.
However, even the low estimate of fish bio-flocs
harvesting may represent a very significant feed source,
constituting about 50% of the regular feed ration of fish
this size (assuming daily feeding of 2% body weight,
i.e. 20 g kg− 1). The daily uptake of nitrogen was about
0.3 g N per kg fish, an amount equivalent to 2 g protein,
a significant portion (close to 50%) of protein feed
requirements of the fish. The estimated contribution of
the daily nitrogen retention from natural biota in shrimp
(1 to 9 g) was 18–29% of the total nitrogen uptake
(Burford et al., 2004). However, the suspended nitrogen
in the shrimp ponds studied was in the range of
7–11 mgN l− 1 , less than a half of suspended N
concentrations in the present system. Considering this
difference, the results reported by Burford et al. (2004)
are in line with findings of this experiment.
The accumulation of 15 N in the fish, as a function of
time represents its uptake minus its excretion. The ex-
cretion of TAN by the fish was demonstrated through
the decrease of the 15 N enrichment in the water. The
uptake of tagged bio-flocs should result in the
diminution of the amount of 15 N in the water, not its
concentration. The lowering of 15 N enrichment was
most probably due to the fact that low enriched TAN
was excreted, diluting the 15 N concentration in the
water. Unfortunately, this process could not be followed
quantitatively, due to the high nitrate concentrations in
the water and the relatively high variability of its
concentrations, not enabling to clearly define and
compute nitrogen balances in the water. (Experiments
like the one reported here should be conducted keeping
as low as possible NO3 concentrations).
4. Conclusions
Microbial flocs developing in bio-flocs technology
(BFT) ponds are demonstrated to be an effective potential
food source for tilapia, and possibly other fish. The feed
contribution of microbial flocs in the tested ponds
contributed close to 50% of fish protein requirement.
The potential feed storage and dynamics of the pond water
can be easily determined in the field through the
monitoring of FV. Fish growing in the BFT pond did
not rush onto the added feed pellets, since the pond
contained flocs as a potential feed, feed that is available
24 h per day.
An important potential conclusion of this work is
the possibility to reduce feed rations in BFT systems.
Avnimelech et al. (1994) estimated that feed utilization
is higher in BFT systems, while tilapia in such ponds is
fed a ration 20% less than conventional one. Fish yields
in these ponds were high, yet the authors did not check
the possibility of feed reduction in a controlled,
quantitative way. The feed requirement of shrimp
growing in BFT models was studied recently by
Panjaitan (2004). It was found that lowering feed
application by up to 30% of conventional feeding
ration, did not lower shrimp growth, probably due to
the partial replacement of feed by the microbial flocs.
The recycling of protein and the lower demand for
external protein sources is an important environmental
advantage, lowering the fishing pressure on marine
resources.
The determination of FV can be practiced in the
farm. It is very easy to conduct and does not demand
sophisticated laboratory or expensive instruments. Yet,
it can be used to estimate the storage of feed equivalent
in the water. According to data obtained in the present
work, each cm3 of settled flocs contains about 14 mg
flocs as dry weight. A relatively low reading of 5 ml l− 1
FV is equivalent to an amount of 700 kg dry matter per
Fig. 6. Actual 15N uptake, computed gross uptake and excretion of 15N
tagged bio-flocs.
 Y. Avnimelech / Aquaculture 264 (2007) 140–147
ha. This is a very significant potential feed equivalent,
as compared to the amount of feed rations in ponds.
The results obtained in this work may be site
specific or may vary with the change of fish, pond and
other variables. The potential importance of the results
obtained here fully justify further research and
widening of the knowledge base to help utilize the
natural feed recycling potential of bio-floc systems. In
addition, methods used to quantitatively evaluate the
different fluxes taking place in bio-floc technology
systems should be improved and refined.
Acknowledgments
The execution of this work was possible only due to
the help of Pacific Aquafarm, headed by the president
Mr. Bill Engler and the manager Mr. Colin Bornia. The
help of Kent SeaTech farm, enabling the use of their
laboratory facilities is appreciated. The help of the UC
Santa Barbara Marine Science Inst. Analytical Labora-
tory in determining carbon and nitrogen is greatly
appreciated. Finally I want to thank Mr. Dean Farrell for
his encouragement both before and during this work,
and the help given by my wife, Mira in executing the
laboratory work.
References
Avnimelech, Y., 1999. Carbon/nitrogen ratio as a control element in
aquaculture systems. Aquaculture 176, 227–235.
Avnimelech, Y., Lacher, M., 1979. A tentative nutrient budget for
intensive fish ponds. Israeli Journal of Aquaculture-Bamidgeh 31,
3–8.
Avnimelech, Y., Ritvo, G., 2003. Shrimp and fish pond soils: processes
and management. Aquaculture 220, 549–567.
Avnimelech, Y., Mokady, S., Schroeder, G.L., 1989. Circulated ponds
as efficient bioreactors for single-cell protein production. Israeli
Journal of Aquaculture-Bamidgeh 41 (2), 58–66.
Avnimelech, Y., Kochva, M., Diab, S., 1994. Development of
controlled intensive aquaculture systems with a limited water
exchange and adjusted carbon to nitrogen ratio. Israeli Journal of
Aquaculture-Bamidgeh 46 (3), 119–131.
Boyd, C.E., 1985. Chemical budget for channel catfish ponds.
Transactions of the American Fisheries Society 114, 291–298.
147
Browdy, C.L., Bratford, D., Stokes, A.d., McIntosh, R.P., 2001.
Perspectives on the application of closed shrimp culture systems.
In: Browdy, C.L., Jory, D.E. (Eds.), The New Wave, Proceedings
of the Special Session on Sustainable Shrimp Farming. World
aqua. Soc., Baton Rough. LA, pp. 20–34.
Burford, M.A., Thompson, P.J., McIntosh, P.R., Bauman, R.H.,
Pearson, D.C., 2003. Nutrient and microbial dynamics in high-
intensity, zero-exchange shrimp ponds in Belize. Aquaculture 219,
393–411.
Burford, M.A., Thompson, P.J., McIntosh, P.R., Bauman, R.H.,
Pearson, D.C., 2004. The contribution of flocculated material to
shrimp (Litopenaeus vannami) nutrition in a high-intensity, zero-
exchange system. Aquaculture 232, 525–537.
Cam, D., Rollet, P.E., Mariotti, A., Guillaume, J., 1991. The relative
contribution of natural productivity and formulated food in the
nutrition of Panaeus japonicum reared in semi-intensive condi-
tions. Aquatic Living Resources 4, 175–180.
Chamberlain, G.W., Hopkins, S.J., 1994. Reducing water use and feed
cost in intensive ponds. World Aquaculture 25, 29–32.
Eaton, A.D., Cleserci, L.S., Greenberg, A.E. (Eds.), 1995. Standard
Methods for the Examination of Water and Waste Water, 10th
edition. Amer. Pub. Health Assoc., Washington D.C.
Hopkins, J.S., Hamilton II, R.D., Sandifer, P.A., Browdy, C.L., Stokes,
A.D., 1993. Effect of water exchange rate on production, water
quality, effluent characteristics and nitrogen budgets of intensive
shrimp ponds. Journal of the World Aquaculture Society 24,
304–320.
McIntosh, P.R., 2000. Changing paradigms in shrimp farming: IV.
Low protein feeds and feeding strategies. Global Aquaculture
Advocate 3, 44–50.
Panjaitan, P., 2004. Field and laboratory study of Penaeus monodon
culture with zero water exchange and limited water exchange
model using molasses as a carbon source. Ph.D. Thesis, Charles
Darwin Univ., Darwin, NT, Australia.
Parker, P.L., Anderson, R.K., 1989. A 13C and 15N tracer study of
nutrition in aquaculture: Penaeus vannamei in pond growout
system. In: Rundel, P.W., Ehrlinger, J.R., Nagi, K.W. (Eds.), Stable
Isotopes in Ecological Research. Springer, New York, pp. 157–171.
Serfling, S.A., 2000. The Solar aquafarms story. Global Aquaculture
Advocate 3, 48–51.
Tacon, A.G.J., Cody, J., Conquest, L., Divakaran, S., Forster, I.P.,
Decamp, O., 2002. Effect of culture system on the nutrition and
growth performance of Pacific white shrimp Litopenaeus vannami
(Boone) fed different diets. Aquaculture Nutrition 8, 121–137.
Taw, N., 2005. Shrimp farming in Indonesia. Evolving industry
responds to varied issues. GAA Advocate 8, 65–67.
Velasco, M., Lawrence, A.L., Neill, W.H., 1998. Development of a
static-water ecoassay with microcosm tanks for postlarval Penaeus
vannamei. Aquaculture 161, 79–87.