Final Investigation Report EVALUATING EFFICIENCY OF NUTRIENT AND SOLID WASTE RETENTION OF TRADITIONAL AND IMPROVED CAGES FOR FISH CULTURE IN DEEP LAKE WATERS Watershed and Integrated Coastal Zone Management 09WIZ03UM

1


Final Investigation Report


EVALUATING EFFICIENCY OF NUTRIENT AND SOLID WASTE RETENTION OF
TRADITIONAL AND

IMPROVED CAGES FOR FISH

CULTURE IN DEEP LAKE WATERS


Watershed
and

Integrated Coastal Zone Management

09WIZ03UM



Keith Hayse
-
Gregson and James S.
Diana

University of Michigan

Ann Arbor, Michigan, USA


Liu Liping

College of Fisheries and Life Science

Shanghai Ocean University

Shanghai, China


Wang Weimin

College of Fisheries

Huazhong Agricultural University

Wuhan, Hubei, China


Song Biyu
, He Xiang, Zhang Qian

School of Resource and Environmental Science

Wuhan University

Wuhan, Hubei, China



ABSTRACT



This study assessed the impacts of improved freshwater aquaculture cages designed to reduce waste
loadings into Longtan Reservoir in southern Guizhou Province, China. These experimental cages featured
a sediment collector under the cages, which allowed for

removal of feces and waste feed from the water
column. The cages were stocked with catfish and also featured an outer cage stocked with bighead carp
,
common carp, and tilapia that could feed off plankton and waste feed in the water column and could
improv
e water quality around the cages. The experiment began in May 2010, and data collection
continued until December 2010. Fish weight and length were measured monthly to establish growth rates.
Fish carcasses, feces, and fish feed were analyzed to determine t
he percent phosphorus. The
sedimentation rates were also measured by sampling the sediment from the sediment collector. Water
chemistry data was also collected: NO
3
, NO
2
, TN, TP, TSS, pH, Chl
-
a, NH
4
, temperature, and Secchi
depth. Phytoplankton and

zooplan
kton were also monitored in Longtan Reservoir.
Water s
amples were
collected inside each cage and 1m outside the cages at depths of 0.5, 5, and 15m. Additional samples were
collected 1km upstream and downstream of the cages, as well as in the bay in which t
he cages were
located.


Contrary to expectations, growth rates of catfish were lower in traditional

cages than in modified cages.
Feed conversion ratios ranged from 1.37
-
2.61
,

but were not statistically different between traditional,
modified
-
a, and modified
-
b groups. Slower growth rates resulted in lower catfish production from
traditional cages. As expected fish in the control cage grew slower than fish in the modified outer

cages.

Results from the mass balance model indicated that the outer cages retained 0.05
-
0.07 kg P ton
-
1

fish
produced and there was no difference in retention between the modified
-
a and modified
-
b cages. This
retention was smaller

than expected, equalin
g 0.034
-
0.05% of the total pho
sphorus input into the system.

2


ANOVA analysis of model outputs only found a difference in P retained in catfish between traditional
and modified cages.


There were no significant differences

between the surface and bottom wat
er
quality in
cage
s
.
Water
temperature ranged from 19.5 to 30.7
℃
, pH from 7.99 to 8.80, and DO from 4.76 to 8.71 mg/L
from June
to December 2010
. Ammonia and nitrite

accounted for
a small proportion

of the
total inorganic nitrogen
,

while
nitrate

accounted
for not

only the main part of
TIN
, but also
a major c
omponent of total nitrogen
.
During the growth period,
2.7t dry weight of waste was collected, which inhibited water eutrophication.
The
feces collected

contained 2.93% crude p
rotein, 0.29%TP, and 0.47% T
N.
There were no significant
differences among water quality of tr
aditional or experimental cages.
Longtan Reservoir was phosphorus
limited
, although t
he content of chlorophyll a showed a significantly positive correlation with TN. There
were no significan
t differences in TN and TP between water
from

experimental cage
s

and the reservoir.


There were 122 species of phytoplankton, which belonged to 49 genera and 7 phyla. The most dominant
species of phytoplankton in the reservoir were
Cyclotella comensis
,

Cyclotella stelligera
,
Navicula
exigua
,
Scendesmus bjjuga
,
Trionema minus
,
Merismopedia tenuissima
,
Crptomonas ovate
,
Chlorella
minimum
,
Crucigenia rectangularis
. Also there were 92 species of zooplankton including 26 Protozoans,
43 Rotifers,
14 Cladoceran
s and 9 Copepods.
The predominant species were mainly
Keratella cochlearis
,
Brachionus falcatus
,
Dicranophorus caudatus
,
Bosmina coregoni
, and
Paracyclops fimbriatus
.


Plankton
collections revealed no evidence of the cage system causing a change in the pla
nkton community.


INTRODUCTION


A
quaculture presents reservoir managers and fish farmers with a dilemma. How can production be
maximized while maintaining water quality for human use and ecological integrity? An aquaculture
system only retains ~ 24%
carbon, 31% nitrogen, and 31% phosphorus inputs within the fish biomass, the
rest is released into the water column (Troell and Norberg, 1998). In aquaculture, water quality
degradation primarily occurs from poorly designed cages, overcapacity of cages, or

improper feeding
strategies,
which highlights

the importance of understan
ding the water body’s

assimilation capacity (Wu
et al., 2000; Guo et al., 2009). When cages produce an excess of uneaten food or feces, nutrient levels
around the cages increase and
benthic habitats are disturbed, altering ecological relationships in the
reservoir (Bureau and Hua, 2010).


Solid waste from cage aquaculture affects the benthic habitat below the cages (Kalantzi and Karakassis,
2006).
High organic content in aquaculture w
astes enriches benthic sediments and can elevate benthic
biomass
however;

aquaculture enrichment also consistently reduces benthic biodiversity
(Mente et al.,
2006; Rooney and Podemski, 2009). Diversity decreases due to deoxygenation as organic material
de
composes (Kalantzi and Karakassis, 2006; Giles 2008; Rooney and Podemski, 2009). Mitigating
benthic degradation usually involves ceasing production over an area to allow the benthic community to
rehabilitate or mechanical sediment removal or oxygenatio
n

(A
ngel et al., 2005; Buryniuk et al., 2006).
These methods results in decreased farm production and additional management costs.


Phosphorus generally limits phytoplankton biomass in temperate freshwater lakes (Schindler et al., 1978;
Bureau and Hua, 2010).
However, tropical systems exhibit a more variable relationship between total
phosphorus and chlorophyll
-
a levels (Huszar et al., 2006). If nutrient levels, especially phosphorus,
surpass certain thresholds then phytoplankton blooms can have detrimental
e
ff
ects on the water column
leading to hypoxic conditions in the hypolimnion, fish kills, reduced water clarity,
and
cyanobacteria
blooms that degrade the taste of fish (Buyukates

et al., 2000) and reduce nutritional quality (Mares et al.,
2009). Therefore, nutrients released from cage aquaculture into the environment must be kept below
thresholds prone to induce negative biological
events
. Dilution has long been the human solutio
n to
nutrification of waters, but in many areas increased anthropogenic nutrient loading has exceeded the
assimilative capacities of freshwater systems (
Halwart et al.
, 2007; Troell et al., 2009). How can
aquaculture systems be engineered to profitably cul
ture fish while reducing their ecological footprint?


3


Polyculture, long practiced in aquaculture, offers a potential solution. Earliest aquaculture systems in
China involved culturing organisms at different trophic levels to manage waste products, for examp
le
raising fish in conjunction with rice paddies (Beveridge and Little, 2002). Most polyculture systems are
pond based and relatively little research has been done on freshwater cage polyculture.


This experiment assessed the implementation of a polycultu
re freshwater cage system. A new cage design
with two modifications
was tested f
or reductions in waste releases
compared to traditional cage designs.
The first modification involved a sediment collection cone underneath the cage

to capture

particulate
wast
e feed and feces. The second modification was an outer cage
to hold

filter feeding fish
which
were
stocked and not fed
. The objectives of this study were to determine the effectiveness of the new cage
design by quantifying nutrient releases from modified
and traditional cages. This was done by: 1)
evaluating growth rates of fish in the inner and outer cages
,

2) using a mass balance model to quantify
phosphorus dynamics of both types of cages
,

3) collecting water samples to assess any changes in water
chemi
stry between
t
he modified and traditional cages and the reservoir
, and 4) to quantify plankton
populations in the reservoir to evaluate effects of cages on the reservoir
.


MATERIALS AND METHODS


Cages

for this exper
iment were situated in Longtan R
eservoir

on the Hongshui Rive
r in southwestern
China
. The reservoir straddles southern Guizhou province and northern Guangxi province. With a surface
area of 360
-
540 km
2

and a depth at the facility

that
surpassed 50m
,
Longtan is characterized as a narrow
and deep
reservoir surrounded by a karst landscape.
During the testing period, water temperature was 19
-
31

C,
pH was generally above 8
, and DO was higher than 5

mg/L.


The facility used for
our

experiment
had

approximately 50 cages of which 10 were dedi
cated to
the
experiment
: 6 modified cages divided into 2 sets of 3. Modified
-
a cages contain an outer cage and no
sediment trap while modified
-
b cages contain both an outer cage and sediment trap. In addition to
modified cages were 3 traditional cages, and 1 contro
l cage.
M
odified cages
were 12x12 m in surface area
and 6m deep, not including the sediment collector below. The traditional cages were 5x5 m in surface
area and 5 m deep, and the control cage was 3x3 m in surface area and 3 m deep.


On

24 May

2010 channe
l catfish
(
Ictalurus punctatus
) were stocked in inner
modified cages and
traditional cages at a density of
160 fish m
-
2
, equaling 16,000 fish in modified cages and 4,000 catfish in
traditional

cages. M
odified outer cages were stocked with 350

kg
of
bighead

carp (
Hypophthalmichthys
nobilis
)
, 100

kg

of

Nile
tilapia (
Oreochromis nilocticus
)
, and 50

kg
of
common carp (
Cyprinus carpio
)
;
this equates to a total density of 11.36 kg m
-
2
. Twice a day (07:30
and

19:30) fish

in the inner cages were
fed
by hand with
Tongwei Company Feed. Outer cage

and control
fish were not fed at all, but left to
consume waste drifting out of inner cage
s

as well as natural food in the water column.


To test the efficacy of outer cages in removing effluent waste, an additional control

cage was used to
determine growth rates of fish supplied with only natural food. This

control cage was stocked with 42

kg
of
bighead carp, 24

kg
of
tilapia, and 6

kg
of
common carp for a total density of 8 kg m
-
2
, slightly less
than the d
ensity of modifie
d outer cages.
The cage was situated adjacent to a facility building
approximately 10

m away from all cages. To estimate growth rates each fish was weighed at stocking and
a sample of 44 fish from this cage sampled on 27 December 2010 and wet weight measur
ed. Average
individual
wet weight at stocking for bighead carp equaled 312

96g, tilapia 339

98g, and common carp
466

80g.
Fish were removed from the cages five times when they were anesthetized, measured for wet
weight and total length, and returned to the

cage.


Proximate composition, including moisture, crude protein, crude fat, phosphorus, crude fiber, amino acids
and crude ash
were measured
in the feed and fish carcasses.

Crude protein and phosphorus in the fe
ces
were also determined.
Dry matter was mea
sured by

weight after drying at 105°C.
Crude protein was
measured by Kjeldahl method, fat content by Soxhlet method with ether extraction,
ash by combustion at
550°C in a muffle furnace, crude fiber by filtration, and amino acid by automatic analyzer
.
All
of the

4


methods were
according to Chinese standard methods (GB/T 9695.4 Standardization Administration of
the People’s Republic of China, 2009).



A number of performance indicators were calculated for the fish, including:


Daily Growth Rate (DGR) =

ΔW

×

100 /

[(t
2
-
t
1
)

×

W
1
]


Specific Growth Rate (SGR) =

[
(LnW
2

-

LnW
1
)

×

100
]

/

(t
2
-
t
1
)


Fatness (Kcp) =

100

×

W/L
3


Feed
Conversion Rate

(FCR) =

W
F

/

(ΔW

×

%

Survival
)


Nutrient Utilization Rate (PNV) = G
nutrition

/ C
nutrition

×

100


Weight increment of
catfish

=

[
W
f

×

(
N×
%Survival)
]

–

(
W
i
×N
)


Crude fiber = 100

-

[
100

×

F
1

×

N
2
/(N
1

×

F
2
)
]


w
here

W
1
and W
2

represent weight of fish at t
1

and t
2
; W
F

the feeding amount during a
time

period; ΔW
the increased weight of fish during the
time

period; W
i

and W
f
body weight of fish at stocking and
harvest; N number of catfish in the cage; V volume of inner cage, G
nutrition

and C
nutrition
growth nutrients,
including protein, fat, phosphorus, nitrogen free extract and amino acid.


We also estimated crude
digestibility of N and P using the formula:


Apparent Crude Digestibility
(%)

=

100

–

[(
100×F
1
×N
2
)
/(N
1
×F
2
)
]

for either N or P


w
here

F
1

represents

percen
tage of crude fiber in feed,

F
2
crude fiber in
feces
;
N
1
percent

nitrogen or
phosphorus in feed; and
N
2
percent

nitrogen or phosphorus in
feces
.


A mass balance model was used to quantify phosphorus dynamics of traditional and modified cages and
to estimate difference
s

in phosphorus loadings to the reservoir for each design.
Phosphorus discharged
from cages
into the water column was modeled by the following equation:


(P
waste

+

P
feces
+ P
excretion
) = P
feed

-

P
carcass



P
waste

= M
ass of phosphorus in uneaten food (g)


P
feces

=
M
ass of in
digestable

phosphorus egested as feces (g)


P
excretion
= M
ass of excreted soluble phosphorus (g)


P
feed

=
M
ass of phosphorus in fish feed (g)


P
carcass

=
M
ass of phosphorus retained by fish (g)


The model used
a number of equations to determine these parameters, derived
from Reid and Moccia
(2007)
.



To corroborate the mass balance model
,

water samples were also taken around the facility to assess if
there was a reduction in phosphorus around the modified cages.

Water samples were taken 1

m outside
each modified and traditional cage

at depths of 0.5, 5, and 15

m. Samples were also taken inside
each
traditional cage at 0.5 m deep, and inside

each

modified cage

at 0.5 and 5

m deep
. An additional 3

5


samples
were also

taken approximately 1
-
2

km away from the cages to establish background
reservoir
water characteristics independent of
any

experimental

influence
s
.
S
amples were taken
at

three

depths

at

0.5, 5, and 15

m. Water samples were processed
at

Guizhou Normal Unive
rsity for total phosphorus (TP)
,
chlorophyll
a

(chla), and total nitrogen (TN).
TN

and TP
were

determined with persulfate digestion in an
autoclave
,

and concentrations measured colorimetrically (Chinese standard methods of water quality
analysis, GB3838
-
20
02). Chla samples were filtered through a 0.45µm glass fiber paper,

then

the filter
was steeped in 90% acetone for 24 hours and chla concentration

determined by spectrophotometer (Lin

et
al.
, 2005).


Additionally, phytoplankton and zooplankton samples were

collected in cages and also at the outside
locations where wat
er quality samples were taken.
Samples were collected

every

two weeks between 15

July

and 15 November 2010. All collections began at
0
8:00 and finished before 12
:00

the same day.
Samples were v
ertical tows from 5 m to the surface, with a mesh size of
157 µ
for phytoplankton and
78
µ
for zooplankton.
Samples were then preserved for examination in the lab, where they were enumerated
to species and estimates of numbers, biomass, and species diversi
ty were made. Species

diversity was
evaluated u
sing the Shannon
-
Wiener Index.


Due to seasonal changes in water chemistry
,

some
variables

violated
the
assumption that all data came
from an identical distribution
,

so
statistical tests for water quality data
were done using a

sign test to
assess difference in means. Nutrient
concentrations

inside and outside cages were subtracted from one
another, and then sign test used to
determine

if
the difference

was
significantly

greater or lesser than
zero
.
One
-
way analysis of variance (ANOVA) tests were conducted on final estimates from
the
mass balance
model

as well as fish growth data
. When significant differences were detected
,

Tukey HSD
test was used
to establish which
group
s

were significantly different from each other
. All statistics were computed on
R:
A Language and Environment for Statistical Computing

(R Development Core Team, 2011).

For
statistical tests alpha was set at 0.05
.


RESULTS


Contrary to expectations, growth rates of catfish were different between traditional and modified cages
(Table 1). Even though there was no difference in the feed available between the traditional, modified
-
a,
and modified
-
b cages, fish in the traditional
cages were smaller at harvest (Table 2) than fish in the
modified cages. Feed conversion ratios (FCR) ranged from 1.37
-
2.61 but were not statistically different
between traditional, modified
-
a, and modified
-
b groups. Slower growth rates resulted in lower c
atfish
production from traditional cages.


As expected fish in the control cage grew slower than fish in the modified outer cages (Figure 1). The
average weight of tilapia and bighead carp on 27 December in the outer cages was 702 and 424 g,
respectively,

while in the control cage the weight was 339 and 304 g, respectively. Common carp were
not sampled in the outer or control cages on 27 December and without FBW estimates were not included
in growth analysis.


Taking fibrin as an indigestible indicator,
apparent digestibilit
y of nitrogen and phosphorous

was not
significantly different in the two types of modified cages (
P
>0.05). The apparent digestibility of nitrogen
and phosphorous in Modified
-
a cages was 92.5%±0.2% and 79.5±1.0% respectively, while the
apparent
digestibility of nitrogen and phosphorous in Modified
-
b cages was 91.2%±0.6% and 82.1%±2.5%,
respectively (Table 3). On 1 September, the amount of dry matter and nitrogen collected from the
sedimentation cone accounted for 18.8% and 6.5% of the fe
ed applied, respectively (Table 4). On 7
November 7, it represented 21.8% and 16.2% of the feed applied, respectively.


Results from the mass ba
lance model
(
Table 5
) indicate that t
he outer cages retained 0.05
-
0.07 kg P ton
-
1

fish produced and there was n
o difference in retention between the modified
-
a and modified
-
b cages
. This
retention

was

smaller than expected
,

equaling

0.034
-

0.05% of the total phosphorus input into the system


6


(Table
6
).
ANOVA analysis of model outputs only found a difference

in
P retained in catfish between
traditional and modified cages.


Contrary to expectations, water quality measurements showed no depression of TP, TN, and chla
concentrations
inside or
outside the modified cages. Water quality measurements showed no statistical
difference when measured at 0.5m and 5m and so these data were pooled to create a larger sample size.
Once pooled, TP, TN, and chla concentrations were the same inside and outside

the cages for both the
traditional, modified
-
a, and modified
-
b cages (Figure 2). Additionally, samples were also the same
between inside, outside, and reference sites for all cages. This indicates that TP, TN, and chla were not
elevated near the aquacultu
re facility.


There were 122 species of phytoplankton collected, belonging to 49 genera and 7 phyla.

Altogether, 23
genera and 56 species of Chlorophyta, 10 genera and 37 species of Bacillariophyta, 8 genera and 18
species of Cyanophyta, 2 genera and 4 spe
cies of Cryptophyta, 2 genera and 3 species of Euglenophyta, 2
gener
a and 2 species of Pyrrophyta,
and
2 genera and 2 species of Xanthophyta were identified.
Phytoplankton in the three types of cages and the control water were mainly composed of Chlorophyt
a,
Bacillariophyta and Cyanophyta, which were typical phytoplankton species found in f
reshwater lakes and
reservoirs.

Traditional cages had greater number of species than modified a or b cages, and all cages had
more s
pecies than reservoir samples.
However
, biomass of phytoplankton was fairly similar between the 3
cage types, with traditional cages ranging from 0.35
-
2.98 mg/L and an average of 1.47 mg/L; modified b
cages ranging from 0.47
-
2.47 mg/L and averaging 1.36 mg/L; and modified a ranging from 0.38
-
2
.19
mg/L, averaging 1.44 mg/L.


The dominant species of phytoplankton changed over time (Table

7
).
Cyclotella comensis

was dominant
mostly in mid
-
July,
Cyclotella stelligera

from mid
-
July to mid
-
November,
Navicula exigua

in mid
-
July
and mid
-
September,
Sce
nedesmus bijuga

and
Tribonema minus

from mid
-
July to mid
-
August,
Merismopedia tenuissima

appeared in late August,
Cryptomonas rostrata

from mid
-
October to mid
-
November, both
Tetraedron minimum

and
Crucigenia quadrata

were most abundant in mid
-
November.
Phytoplankton biomass also increased as water temperature increased. For example, the phytoplankton
biomass in traditional cages was 2.98

mg/L in mid
-
July and only 0.35

mg/L in mid
-
November. The
Shannon
-
W
i
ener index H
′ for traditional cages averaged 2.15, for modified b cages it averaged 2.10, and
for modified a cages it averaged 2.06. Overall the cages showed slightly elevated phytoplankton
populations compared to open water, and no differences among cage type.


There

were 92 species of zooplankton identified i
ncluding 26 species of Protozoa
, 43 of Rotifera, 14 of
Cladocera and 9 of Copepoda. Rotifera were most abundant, and Copepoda least. Dominant species
included
Keratella cochlearis
,
Brachionus falcatus
,

Brachionus

forficula,

Keratella valga
,
Dicranophorus caudatus
,
Brachionus donneri
,
Bosmina coregoni

and
Paracyclops fimbriatus
.
Once
again, biomass trends in the 3 cage types were similar

(Table 8)
, and densities of zooplankton were also
fairly similar among cage ty
pes. Zooplankton density ranged from
48
-
678

ind./L in traditional cages and
averaged
287

ind./L, modified b cages ranged from
41
-
570

ind./L and averaged
176

ind./L, while
modified a cages ranged from 40
-
490

ind./L and averaged
239

ind./L.
For all cage
types, density was
maximum in mid
-
September and minimum in mid November.

The Shannon
-
Weaver index H′ for
traditional cages averaged
1.
78, while modified b
and a
cages
both

averaged 1.74.

Once again for
zooplankton, populations inside cages were slightly el
evated above background reservoir levels, and did
not differ largely among cage types.


The net income ratio of modified cages a and b were 74.3% and 73.8%, respectively. Cost of the waste
collection de
vice accounted for only
0.6% of the total cost. The re
venue of channel catfish in modified
cages a and b accounted for 96.8% and 96.4% of the total revenue, respectively. The revenue of fish in
the outer cages of modified cages a and b accounted for 3.2% and 3.6% of the total revenue, respectively.


7




DISCUSSION


This study tested the efficacy of planktivorous fish to retain nutrients from cage aquaculture. Fish in outer
cages retained <1% of total phosphorus input from cages. This retention was not enough to influence
nutrient concentrations between
modified, traditional, and reference sites. This raises questions about the
availability of waste nutrients from intensively fed cages to consumption of materials by filter feeding
fish. Since fish in outer cages grew faster than fish in the control cage,
they must have had access to feed
energy drifting out of the inner cage, which elevated their growth rates. However, they may not have had
access to elevated plankton populations

because plankton populations were only slightly elevated in

cages
compared to

open water.
Some studies have shown elevated production near aquaculture facilities (Angel,
2002; Spanier et al., 2003; MacDonald et al., 2011), and other authors have found no increase in
productivity around cages (e.g.
, Navarette
-
Mier et al., 2010).
In
our case, the waste feed appeared to
increase fish growth near the facility, but there was no
major

increase in natural foods there.


The growth of channel catfish demonstrated obvious seasonal changes, which were related to the changes
in water temperatu
re and possibly DO. Compared with other cage culture experiments on channel catfish,
this experiment had higher yield per unit time and unit volume of the cage (Ding et al., 1999; Qu and Du,
2002; Zhu et al., 2004; Kelimo et al., 2005; He and Pei, 2007).


Total phosphorus input into the system was not unusually high. Phosphorus loading at Longtan was
actually less than the 25
-
35 kg P ton
-
1

fish that Guo and Li (2003) and Guo et al. (2009) reported in
previous studies on cage aquaculture. Guo and Li’s (2003
) larger waste output can be explained by an
imprecise feeding regime for cultured fish in that study, consisting of forage fish, grass, and formulated
feed. Using nutritionally imprecise feeding can increase phosphorus loading from aquaculture (Cho

et al.
,
1994,
Cho and Bureau,
2001; Bureau

and Hua
, 2010). De

Silva

et al.

(2010) supported this conclusion by
finding that waste from artisanal feeds loaded more phosphorus than waste from commercial feeds. The
high ratio of intensively fed fish to filter feedi
ng fish (11.4:1), in this study may have resulted in lower
phosphorus retention rates than other studies have shown. Yi et al. (2003) cited a range of phosphorus
retention values from 0.86
-
17% by filter feeding fish in integrated pond systems. The stocking

ratios of
intensively fed to filter feeding fish ranged from 2.5:1
-
9:1 in that study, and lower ratios achieved higher
retention. In partitioned aquaculture systems a ratio of catfish to tilapia stocking of 4:1 was found to be
optimal to control algae gro
wth (Brune et al., 2003). The high ratio in our study probably reduced
retention capacity. Future research on optimal stocking ratios in reservoir systems may improve nutrient
retention.



Herbivorous and filter feeding finfish may retain <10% of P in rese
rvoir systems. This presents a
challenge to freshwater integrated aquaculture since marketable aquatic autotrophs are limited; the lack of
marketable freshwater autotrophs may limit the maximum nutrient retention possibilities for freshwater
integrated aqu
aculture.


Constraints in sampling ability and methodological assumptions limited conclusions drawn from the
results. The remote location of the facility made sampling infrequent, and the experiment occurred at a
commercial facility so it is unknown how
the proximity of the experiment to other cages at the facility
confounded results. Since sampling only occurred on one location outside of the cages, water flow could
influence the results of the water chemistry. Flow appeared negligible, but if physical p
rocesses were
moving nutrient loads from the cages away from our sampling site then we would under estimate cage
loading of nutrients. Other studies have shown cage aquaculture impacts extending up to 50m from cages
(Guo and Li, 2003). Finally, low stockin
g density made sampling outer cage fish very difficult, resulting
in small samples sizes.


Most integrated aquaculture research occurs in ponds and marine systems, many of these experiments
have successfully increased trophic efficiency (Yi
et al., 2003;
Angel et al., 2002
; Troell et al., 2009;
MacDonald et al., 2011), though some integrated cage mariculture publications find no enhanced growth

8


from integrated organisms (Navarett
-
Mier et al., 2010). The elevated growth rate of fish in outer cages
suggests
that pro
ductivity around cages increased
, but the mechanisms and magnitude of this increase
were not clear
.
The assumption that ecological productivity is elevated in the vicinity of the cages must
also be tested.


According to the analysis on economic be
nefits, the cost of
the solid
waste collection device only
accounted for 0.6% of the total cost, quite
a
low proportion,
and

it could effectively reduce the
environmental pollution of cage
culture. The costs of fish fry

and feed constituted the
main part o
f total
cost, and this

proportion
was

up to 95%. The revenue from channel catfish
was

the main
contributor to

total revenue. The revenue from fish cultured in
the
outer cage accounted for only a small proportion of
the total revenue.


LITERATURE CITED


Angel, D
.
L
.
, N. Eden, S. Breitstein, A. Yurman, T. Katz, and E. Spanier
,

2002. In situ biofiltration: A
means to limit the dispersal of effluents from marine finfish cage aquaculture.
Hydrobiologia
,

469:
1
-
10.

Angel, D
.
L
.
, T. Katz, N. Eden, E. Spanier
, and K
.
D
.

Black
, 20
05. Damage control in the coastal zone:
Improving water quality by harvesting

aquaculture
-
derived nutrients.
Strategic Management of
Marine Ecosystems
,

50:
53
-
64.

Beveridge, M., and D
.
C. Little
, 20
02. The history of aquaculture in tradi
tional societies. In
B.A. Costa
-
Pierce

(Editor).
Ecological aquaculture: The evolution of the blue revolution.

Blackwell Science
Ltd., Oxford, pp. 3
-
29.

Brune, D
.
E, G. Schwartz, A
.
G
.

Eversole, J
.
A
.

Collier, and T
.
E
.

Schwedler
, 20
03.

Intensification of pond
aquaculture and high rate photosynthetic systems.
Aquacultural Engineering
,

28:65
-
86.

Bureau, D., and K. Hua
, 20
10. Towards effective nutritional management of waste outputs in aquaculture,
with particular reference to salmonid aq
uaculture operations.
Aquaculture Research
,

41:777
-
7
92.

Buryniuk, M., R
.
J
.

Petrell, S. Baldwin, and K
.
V
.

Lo
, 20
06. Accumulation and natural disintegration of
solid wastes caught on a screen suspended below a fish farm cage.
Aquacultural Engineering
,

35:78
-
90.

Buyukates, Y., S. Rawles, and D. Gatlin
, 20
00. Phosphorus fractions of various feedstuffs and apparent
phosphorus availability to channel catfish.
North American Journal of Aquaculture
,

62:184
-
18
8.

Cho, C
.
Y
.
, and D
.
P
.

Bureau
, 20
01. A review of diet
formulation strategies and feeding systems to reduce
excretory and feed wastes in aquaculture.
Aquaculture Research
,

32:349
-
3
60.

Cho, C.Y., J.D. Hynes, K.R. Wood, and H.K. Yoshida
.

1994. Development of high
-
nutrient
-
dense, low
-
pollution diets and prediction of aquaculture wastes using biological approaches.
Aquaculture

124:293
-
305.

De Silva, S
.
S., B
.
A. Ingram, P
.
T. Nguyen, T
.

M. Bui, G
.

J. Gooley, and G
.

M. Turchini
, 20
10. Estimati
on
of nitrogen and phosphorus in effluent from the striped catfish farming sector in the
Mekong
delta,
Vietnam
.
Ambio
,

39:504
-
5
14.

Ding
,
Q
.
,
G.
Xu,
Y.
Wang,
D.
Yang,

and

C.
Shi
, 1999.
Tests on the high
-
yield techniques for the cage
culture of the channe
l catfish in reservoir
s
, Reservoir Fisheries, 19 (6): 22
-
22.


9


Giles, H
.
, 20
08. Using bayesian networks to examine consistent trends in fish farm benthic impact
studies.
Aquaculture
,

274:181
-
1
95.

Guo, L., and Z. Li
, 20
03.
Effects of nitrogen and phosphorus from fish cage
-
culture on the communities
of a shallow lake in middle
Yangtze River
basin of
China
.
Aquaculture
,

226:201
-
2
12.

Guo, L., Z. Li, P. Xie, and L. Ni
, 20
09. Assessment effects of cage culture on nitrogen and ph
osphorus
dynamics in relation to fallowing in a shallow lake in
China
.
Aquaculture International
,

17: 229
-
2
41.

Halwart, M.
, D.

Soto

and J.R.

Arthur,(eds.)
, 2007.

Cage aquaculture
–

Regional reviews and global
overview.
FAO Fisheries Technical Paper.

No. 4
98. Rome, FAO,

241pp.

He, G., and B. Pei, 2007. Cage aquaculture of the channel catfish in the reservoir. Reservoir Fisheries,
27(3): 51
-
52.

Huszar, V., N. Caraco, F. Roland, and J. Cole
, 20
06.

Nutrient
-
chlorophyll relationships in tropical
subtropical lakes: Do temperate models fit?
Biogeochemistry
,

79:239
-
250.


Kalantzi, I
.
, and I
.
Karakassis
,
2006. Benthic impacts of fish farming: Meta
-
analysis of community and
geochemical data.
Marine Polluti
on Bulletin
,

52:484
-
4
93.

Kelimo, M.C., J. Zhang, and E. Zhou, 2005. Small volume and high density cage culture of channel
catfish with soybean feed in Nanning, China. Fisheries, 3:84
-
85.

Lin
,

S.J., L.J. He, P.S. Huang
and

B.P. Han, 2005. Comparison and
improvement on the extraction
method for chlorophyll a in phytoplankton. Ecological
Science
,

24:9
-
11.

MacDonald,
B.

A., S
.

M.C. Robinson, and K
.
A. Barrington
, 20
11. Feeding activity of mussels (
Mytilus
edulis
) held in the field at an integrated
multi
-
trophic aquaculture (IMTA) site (
Salmo
salar
) and
exposed to fish food in the laboratory.
Aquaculture

314:244
-
2
51.

Mares, J
.
, M
.

Palikova, R
.

Kopp, S
.

Navratil, and J
.

Pikula
, 20
09. Changes in the nutritional parameters of
muscles of the common carp

(
Cyprinus
carpio
) and the silver carp (
Hypophthalmichthys
molitrix
)
following environmental exposure to cyanobacterial water bloom.
Aquaculture Research
,

40:148
-
1
56.

Mente, E
.
, G
.
J. Pierce, M
.
B
.

Santos, and C
.

Neofitou
, 20
06.

Effect of feed and feeding in the culture of
salmonids on the marine aquatic environment: A synthesis for
European
aquaculture.
Aquaculture
International
,

14:499
-
522.

Navarrete
-
Mier, F
.
, C
.

Sanz
-
Lazaro, and A
. Marin,

2010. Does bivalve mollusc polycultur
e reduce marine
fin fish farming environmental impact?
Aquaculture
,

306: 101
-
10
7.

Qu
,

W., and Z.
Du
, 2002.
Cage
aquaculture of the channel catfish in the reservoir
. Freshwater Fisheries,
32 (6): 22
-
23.

Reid, G., and R. Moccia
, 20
07.

Estimating aquatic phosphorus concentrations 30 metres down
-
current
from a rainbow trout cage array.
Journal of Environmental Monitoring
,

9:814.

Rooney, R
.

C., and C
.
L. Podemski
, 20
09. Effects of an experimental rainbow trout (
Oncorhynchus
mykiss
) farm o
n invertebrate community composition.
Canadian Journal of Fisheries and Aquatic
Sciences
,

66:1949
-
19
64.


10


Schindler,
D., E. Fee, and T. Ruszczynski, 1978.
Phosphorus input and its consequences for
phytoplankton standing crop and production in Experimental L
akes Area and in similar lakes.
Journal of the Fisheries Research Board of Canada
,

35:190
-
196.

Spanier, E.,
A.
Tsemel,
H.

Lubinevski,
A.
Roitemberg, A. Yurman,
S.
Breitstein,
D.
Angel,
N.

Eden, and
T. Katz
,
2003. Can open water bio
-
filters be used for the reduction of the environmental impact of
finfish net cage aquaculture in the coastal waters of Israel? Annales Ser
.

Hist
.

Nat
.

13: 25
-
28.

Troell, M., and J. Norberg,

1998. Modelling output and retention of s
uspended solids in an integrated
salmon
-
mussel culture.
Ecological Modelling
,

110:65
-
77.

Troell, M
.
, A
.

Joyce, T
.

Chopin, A
.
Neori, A
.
H. Buschmann, and J
.
G
.
Fang
, 20
09. Ecological engineering
in aquaculture
-

potential for integrated multi
-
trophic aquacul
ture (IMTA) in marine offshore
systems.
Aquaculture
,

297:1
-
9.

Wu, Z., J. Guo, and Y. Yang
, 20
00. Current status and sustainability of cage culture in reservoirs: A case
study in
China
. In

I.
C
.
Liao

and

C.K
. Lin (Editors)
.
International Symposium on Cage Aquaculture
in Asia, Manila, Philippines, pp.
259
-
265
.

Yi, Y., C
.
K
.

Lin, and J
.
S
.

Diana
, 20
03. Hybrid catfish (
Clarias
macrocephalus x C
-
gariepinus
) and
Nile
tilapia (
Oreochromis
niloticus
) culture in an integrated pen
-
cum
-
p
ond system: Growth performance
and nutrient budgets.
Aquaculture
,

217:395
-
408.

Zhu, S., F. Shi, K. Zhu, and Z. Kong, 2004. Cage aquaculture of the channel catfish. Scientific Fish
Farming, 7: 22
-
22.


11


Table 1
. Growth performance and production results

for each cage type
. Capital letter superscripts indicate a significant
difference between averaged cage values in each column

(p>0.05).


Cages

Feed Applied
(g/fish)

FBW (g)

Growth Rate (g/day)

Tons Produced
(ton=1000kg)

FCR

Trad1

1251

579.8

2.2

2.32

2.61

Trad2

1251

581.8

2.2

2.33

2.60

Trad3

1251

580.9

2.2

2.32

2.60

Avg.

1251.00

580.8
A

2.2
A

2.32
A

2.60







Mod
-
a1

1375

810.4

3.3

12.97

1.94

Mod
-
a2

2030

886.6

3.6

14.19

2.58

Mod
-
a3

1789

963.1

4.0

15.41

2.07

Avg.

1731.33

886.7
B

3.6
B

14.19
B

2.20







Mod
-
b1

1375

1106.3

4.6

17.70

1.37

Mod
-
b2

2030

1042.3

4.3

16.68

2.15

Mod
-
b3

1789

861.7

3.5

13.79

2.35

Avg.

1731.33

1003.5
B

4.2
B

16.06
B

1.96





Table 2.

Average weight (grams), one standard deviation, and sample size for all fish species at each sampling event. Different
capital superscripts denote a significant difference.


Date

a.catfish

±
SD

n

b.catfish

±
SD

n

Trad.

catfish

±
SD

n











16
-
Jul
-
10

154.67

61.60

94

151.57

83.93

75

154.52

103.89

70

30
-
Jul
-
10

238.44

78.03

74

154.23

61.68

63

189.47

80.81

102

30
-
Sep
-
10

542.13

142.78

60

588.20

164.6
7

60

487.02

149.91

60

7
-
Nov
-
10

718.57

171.09

60

801.45

216.4
5

60

583.53

147.21

60

27
-
Dec
-
10

886.70
A

258.90

104

1000.62
A

653.7
6

60

580.85
B

216.19

63











Date

a.bighead

±
SD

n

b.bighead

±
SD

n




27
-
Dec
-
10

424.91

151.21

32

522.50

211.7
3

22














Date

a.tilapia

±
SD

n

b.tilapia

±
SD

n




27
-
Dec
-
10

734.47

370.43

17

679.85

149.3
3

27















12



Table 3
. Conversion ratios of nutrients in different experimental cages.


Cage

Protein (%)

Fat (%)

Phosphorous (%)

Essential amino acid (%)

Modified a

31.06

65.79

51.02

62.89

Modified b

34.36

60.61

36.10

54.64






Table 4
. The ratio of collection of dry matter and nitrogen (mean ± SE) in material collected from the sedimentation cone.

Means with different superscripts within the same column indicate significantly different (
P
＜
0.05).

Date

Dry matter


Nitrogen

Feed (kg)

Collected amount (kg)


Feed (g)

Collected amount (g)

9.1

108.5

20.3±1.8


5936.3

384.1±48.2

11.7

94.4

20.6±1.2


4944.2

800.4±96.1


13



Table 5.

Mass balance model outputs. Capital letter superscripts indicate a significant

difference between averaged cage values in each column (p>0.05).



Cages

Total P Input

(kg / ton)

P Retained in
Catfish

(kg / ton)

Total Waste

(kg / ton)

Particulate Waste


(kg / ton)

Soluble Waste


(kg / ton)

P Retained in
Outer Cage

(kg / ton)

Unaccounted

(kg / ton)









Avg.
Trad

17.02

4.30
A

12.81

6.21

6.16

N/A

-
0.10









Trad1

17.04

4.30

12.84

6.22

6.18

N/A

-
0.10

Trad2

16.99

4.31

12.78

6.20

6.15

N/A

-
0.10

Trad3

17.01

4.30

12.81

6.21

6.16

N/A

-
0.10









Avg.
Mod
-
a

15.39

4.61
B

10.94

5.61

4.86

0.06

0.31









Mod
-
a1

13.40

4.56

9.04

4.89

3.69

0.07

0.27

Mod
-
a2

18.09

4.61

13.58

6.60

6.52

0.06

0.36

Mod
-
a3

14.67

4.66

10.19

5.35

4.37

0.06

0.29









Avg.
Mod
-
b

13.87

4.68
B

9.39

5.06

3.86

0.05

0.28









Mod
-
b1

9.82

4.73

5.37

3.58

1.31

0.05

0.20

Mod
-
b2

15.39

4.70

10.85

5.61

4.77

0.05

0.31

Mod
-
b3

16.40

4.60

11.94

5.98

5.49

0.06

0.33



14



Table 6.

Fate of phosphorus expressed as a percent of total P introduced into the system from
feed. See Table 4 for definition of categories.








P Retained in Catfish

Total Waste P

Particulate Waste

P

Soluble Waste P

P Retained in Outer
Cage







Avg. Trad

25.30%

72.70%

36.47%

36.23%

N/A

Trad1

25.25%

72.75%

36.47%

36.28%

N/A

Trad2

25.35%

72.65%

36.47%

36.18%

N/A

Trad3

25.30%

72.70%

36.47%

36.23%

N/A







Avg. Mod
-
a

30.42%

67.58%

36.47%

31.11%

0.41%

Mod
-
a1

34.01%

63.99%

36.47%

27.52%

0.50%

Mod
-
a2

25.51%

72.49%

36.47%

36.02%

0.34%

Mod
-
a3

31.76%

66.24%

36.47%

29.77%

0.38%







Avg.

Mod
-
b

35.59%

62.41%

36.47%

25.94%

0.41%

Mod
-
b1

48.18%

49.82%

36.47%

13.35%

0.50%

Mod
-
b2

30.56%

67.44%

36.47%

30.97%

0.34%

Mod
-
b3

28.03%

69.97%

36.47%

33.50%

0.38%


15


Table 7.

Dominant species of phytoplankton in each cage type, in water near each cage type, and in open

reservoir water in Longtan Reservoir during 2010.



mid
-
July

Trad. cages

Cyclotella comensis

Navicula exigua

Tribonema minus

Outside Trad.

Cyclotella comensis

Navicula exigua

Cyclotella stelligera

Mod. b

Cyclotella comensis

Navicula exigua

Cyclotella stelligera

Outside b

Cyclotella comensis

Navicula exigua

Tribonema minus

Mod. a

Cyclotella comensis

Cyclotella stelligera

Navicula exigua

Outside c

Cyclotella comensis

Navicula exigua

Cyclotella stelligera

Open water

Cyclotella comensis

Cyclotella stelligera

Navicula exigua

late July

Trad. cages

Cyclotella stelligera

Scenedesmus bijuga

Tribonema minus

Outside Trad.

Cyclotella stelligera

Tribonema

minus

Cyclotella comensis

Mod. b

Cyclotella stelligera

Tribonema minus

Cyclotella comensis

Outside b

Cyclotella stelligera

Tribonema minus

Cyclotella comensis

Mod. a

Cyclotella stelligera

Tribonema minus

Scenedesmus bijuga

Outside c

Cyclotella
stelligera

Tribonema minus

Navicula exigua

Open water

Cyclotella stelligera

Scenedesmus bijuga

Cyclotella comensis

mid
-
August

Trad. cages

Scenedesmus bijuga

Navicula exigua

Tribonema minus

Outside Trad.

Cyclotella stelligera

Tribonema minus

Scenedesmus bijuga

Mod. b

Tribonema minus

Scenedesmus bijuga

Cyclotella stelligera

Outside b

Tribonema minus

Scenedesmus bijuga

Navicula exigua

Mod. a

Tribonema minus

Navicula exigua

Scenedesmus bijuga

Outside c

Tribonema minus

Cyclotella
stelligera

Scenedesmus bijuga

Open water

Tribonema minus

Cyclotella stelligera

Scenedesmus bijuga

late August

Trad. cages

Cyclotella stelligera

Merismopedia tenuissima

Navicula exigua

Outside Trad.

Cyclotella stelligera

Navicula exigua

Merismopedia
tenuissima

Mod. b

Cyclotella stelligera

Merismopedia tenuissima

Navicula exigua

Outside b

Cyclotella stelligera

Merismopedia tenuissima

Navicula exigua

Mod. a

Cyclotella stelligera

Merismopedia tenuissima

Navicula exigua

Outside c

Cyclotella
stelligera

Merismopedia tenuissima

Cyclotella comensis

Open water

Cyclotella stelligera

Merismopedia tenuissima

Navicula exigua

mid
-
September

Trad. cages

Cyclotella stelligera

Navicula exigua

Tribonema minus

Outside Trad.

Cyclotella stelligera

Navicula exigua

Cryptomonas rostrata

Mod. b

Cyclotella stelligera

Navicula exigua

Tribonema minus

Outside b

Cyclotella stelligera

Navicula exigua

Scenedesmus bijuga

Mod. a

Cyclotella stelligera

Navicula exigua

Tribonema minus

Outside c

Cyclotella
stelligera

Navicula exigua

Scenedesmus bijuga

Open water

Cyclotella stelligera

Navicula exigua

Tribonema minus

mid
-
October

Trad. cages

Cyclotella stelligera

Rhodomonas lacustris

Cryptomonas rostrata

Outside Trad.

Rhodomonas lacustris

Cyclotella
comensis

Cyclotella stelligera

Mod. b

Cyclotella stelligera

Tribonema minus

Cryptomonas rostrata

Outside b

Cyclotella stelligera

Cryptomonas rostrata

Navicula exigua

Mod. a

Cyclotella stelligera

Cryptomonas rostrata

Tribonema minus

Outside c

Cyclotella stelligera

Cryptomonas rostrata

Coelastrum sphaericum

Open water

Cyclotella stelligera

Rhodomonas lacustris

Cryptomonas rostrata

mid
-
November

Trad. cages

Cryptomonas rostrata

Cyclotella stelligera

Rhodomonas lacustris

Outside Trad.

Cryptomonas rostrata

Cyclotella stelligera

Cyclotella meneghiniana

Mod. b

Cryptomonas rostrata

Cyclotella stelligera

Coelastrum sphaericum

Outside b

Tetraedron minimum

Cryptomonas rostrata

Cyclotella stelligera

Mod. a

Cryptomonas rostrata

Cyclotella
stelligera

Cyclotella meneghiniana

Outside c

Cryptomonas rostrata

Rhodomonas lacustris

Crucigenia quadrata

Open water

Cryptomonas rostrata

Rhodomonas lacustris

Cyclotella meneghiniana


16


Table 8.

The biomass (mg/L) of zooplankton
found in
three group
s

of cages and open reservoir water in Longtan Reservoir
during 2010.



Trad.
cages

Outside

Trad.

Mod. B

cages

Outside B

cages

Mod. A
cages

Outside A

cages

Open

Reservoir

mid
-
July

1.11±0.27

1.55

1.18±0.18

2.06

0.73±0.53

1.04

15.75

late July

4.60±1.64

4.95

4.01±2.80

7.24

2.89±0.28

4.76

7.59

mid
-
August

2.16±0.53

4.18

2.05±0.81

4.1

1.84±0.37

5.47

8.66

late August

2.88±0.61

1.53

2.64±0.77

0.95

1.65±0.30

2.78

9.86

mid
-
September

5.94±3.33

6.94

6.10±1.94

4.7

5.05±2.48

13.14

5.7

mid
-
October

1.40±0.31

1.16

2.11±0.82

0.92

2.28±0.56

2.61

2.25

mid
-
November

0.40±0.20

0.05

0.33±0.09

1.02

0.64±0.36

0.51

0.75

average

2.64±2.00

2.91±2.48

2.63±1.91

3.00±2.43

2.15±1.50

4.33±4.28

7.22±5.01



17



Figure 1.

Average weight of carp and tilapia in modified and control
cages at stocking and at harvest.


18




Figure 2.

Water quality measurements
collected
during fish culture inside

cages
, 1

m outside

cages
, and at reference
sites.