Content uploaded by Hüseyin Sevgili
Author content
All content in this area was uploaded by Hüseyin Sevgili on May 08, 2018
Content may be subject to copyright.
This article appeared in a journal published by Elsevier. The attached
copy is furnished to the author for internal non-commercial research
and education use, including for instruction at the authors institution
and sharing with colleagues.
Other uses, including reproduction and distribution, or selling or
licensing copies, or posting to personal, institutional or third party
websites are prohibited.
In most cases authors are permitted to post their version of the
article (e.g. in Word or Tex form) to their personal website or
institutional repository. Authors requiring further information
regarding Elsevier’s archiving and manuscript policies are
encouraged to visit:
http://www.elsevier.com/authorsrights
Author's personal copy
Ecological
Indicators
36 (2014) 672–
681
Contents
lists
available
at
ScienceDirect
Ecological
Indicators
jou
rn
al
hom
epage:
www.elsevier.com/locate/ecolind
Parameters
selection
for
water
quality
index
in
the
assessment
of
the
environmental
impacts
of
land-based
trout
farms
Mehmet
Ali
Turan
Koc¸
er∗,
Hüseyin
Sevgili
Mediterranean
Fisheries
Research
Production
and
Training
Institute,
Antalya,
Turkey
a
r
t
i
c
l
e
i
n
f
o
Article
history:
Received
17
February
2013
Received
in
revised
form
25
September
2013
Accepted
28
September
2013
Keywords:
Water
quality
index
Aquaculture
Environmental
impact
Indicator
parameters
Ammonia
Organic
nitrogen
a
b
s
t
r
a
c
t
This
study
aimed
to
check
the
effectiveness
of
water
quality
indices
(WQIs)
in
the
specific
assessment
of
trout
culture
impacts
on
a
stream
water
quality
by
selecting
the
parameters
in
various
approaches.
Water
quality
was
monitored
monthly
for
a
period
of
1
year
in
one
reference
point
and
four
affected
stream
reaches
in
which
discharges
from
intensive
trout
farms,
and
rural
and
agricultural
activities
were
present.
The
objective
WQI
calculation
using
24
parameters
and
the
minimum
WQI
(WQImin)
using
dissolved
oxygen,
biochemical
oxygen
demand,
total
suspended
solids,
total
phosphorus,
ammonia
nitro-
gen
(NH4+-N),
and
total
nitrogen
as
major
indicators
in
trout
farm
effluents
could
not
distinguish
the
aquaculture-impacted
stream
reaches.
However,
WQImin calculation
with
NH4+-N,
total
organic
nitrogen
(TON),
soluble
reactive
phosphorus,
and
total
organic
phosphorus
which
were
selected
using
the
principal
component
analysis
findings
meaningfully
classified
the
sampling
points.
Further
reduction
of
parame-
ters
to
NH4+-N
and
TON
in
WQImin calculation
achieved
a
similar
successful
classification
of
the
sampling
points.
This
study
showed
that
WQImin calculated
using
NH4+-N
and
TON
is
a
useful
and
easily
applicable
methodology
in
the
assessment
of
the
impacts
of
trout
farm
effluents
on
the
stream
water
quality.
© 2013 Elsevier Ltd. All rights reserved.
1.
Introduction
The
use
of
surface
waters
for
various
purposes
threatens
the
integrity
of
aquatic
ecosystems
as
a
result
of
changing
its
quality
and
quantity.
Therefore,
representative
and
reliable
monitoring
and
assessment
of
water
quality
are
critical
(Massoud,
2010).
Surface
water
quality
is
traditionally
assessed
by
water
quality
standards
and
objectives
(Rosemond
et
al.,
2009).
However,
this
traditional
approach
cannot
provide
sufficient
information
on
the
overall
qual-
ity
of
water
or
the
spatial
and
temporal
trends
(Kannel
et
al.,
2007).
Although
dynamic
mathematical
modeling
or
multivari-
ate
statistics
are
the
best
approaches
to
determine
these
trends
(Boyacıo˘
glu
and
Boyacıo˘
glu,
2007),
they
require
overmuch
effort,
financial
resource,
and
expertness
and
they
are
not
easily
appli-
cable
or
cognizable.
Therefore,
researchers
and/or
environmental
authorities
have
strived
to
derive
a
simple
expression
of
the
gen-
eral
quality
of
surface
waters
by
a
single
number,
that
is,
the
water
quality
index
(WQI)
(Debels
et
al.,
2005).
There
are
several
WQIs
using
different
parameters
depending
on
the
water
quality
objectives
all
over
the
world
(CCME,
2001;
Debels
∗Corresponding
author
at:
Mediterranean
Fisheries
Research
Production
and
Training
Institute,
Yes¸
ilbayır
Mah.
Akdeniz
Bul.
No.
2,
07192
Dös¸
emealtı,
Antalya,
Turkey.
Tel.:
+90
242
2510585;
fax:
+90
242
2510584.
E-mail
addresses:
matkocer@akdenizsuurunleri.gov.tr,
matkocer@hotmail.com
(M.A.T.
Koc¸
er).
et
al.,
2005).
The
WQIs
are
commonly
used
in
either
the
classifica-
tion
of
surface
waters
(Boyacıo˘
glu,
2010;
Lermontov
et
al.,
2011)
or
the
assessment
of
beneficial
use
(Said
et
al.,
2004)
and
water
pollu-
tion
(Akkoyunlu
and
Akıner,
2012;
Bakan
et
al.,
2010;
Kannel
et
al.,
2007;
Zhang
and
Zhang,
2007).
For
instance,
Pesce
and
Wunderlin
(2000)
calculated
the
objective
WQI
(WQIobj)
on
20
parameters
and
the
minimum
WQI
(WQImin)
on
three
key
parameters
(dissolved
oxygen,
electrical
conductivity
or
total
dissolved
solids,
and
tur-
bidity)
to
assess
the
effect
of
urban
discharge
on
a
receiving
river
water
quality.
They
suggested
that
the
latter
was
sufficient
as
much
as
the
former
in
the
assessment,
resulting
in
a
decrease
in
analyt-
ical
cost,
which
is
a
limiting
factor
in
water
quality
assessments.
Kannel
et
al.
(2007)
also
showed
that
the
WQImin on
five
param-
eters
(temperature,
pH,
dissolved
oxygen,
electrical
conductivity
and
total
dissolved
solids)
could
be
useful
for
the
periodic
routine
monitoring
program
of
urban
impacts
on
a
river.
The
environmental
sustainability
of
aquaculture,
the
fastest-
growing
food
industry
due
to
the
expansion
of
an
annual
average
growth
of
8.3%
for
the
past
three
decades
(FAO,
2010),
is
a
concern
because
it
can
have
serious
negative
impacts
on
aquatic
ecosys-
tems
as
a
result
of
nutrient
and
organic
matter
enrichment
(Folke
and
Kautsky,
1992).
The
impacts
may
show
clear
temporal
and
spatial
variations
depending
on
species,
culture
method,
stock-
ing
density,
feed
type,
hydrologic
properties,
farm
capacity
and
husbandry
practices
(O’Bryen
and
Lee,
2003;
Tacon
and
Forster,
2003;
Wu,
1995).
There
are
some
studies
using
the
WQIs
in
the
assessment
of
environmental
impacts
of
aquacultural
activities.
1470-160X/$
–
see
front
matter ©
2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.ecolind.2013.09.034
Author's personal copy
M.A.T.
Koc¸
er,
H.
Sevgili
/
Ecological
Indicators
36 (2014) 672–
681 673
Ferreira
et
al.
(2011)
applied
the
Canadian
WQI
to
assess
the
effects
of
shrimp
farm
activities
on
two
coastal
environments
using
18
parameters
representing
physicochemical
and
biological
variations
and
bacteriological
contamination.
The
investigators
did
not
detect
a
significant
influence
in
the
sites.
Simões
et
al.
(2008)
used
U.S.
National
Sanitation
Foundation
WQI,
the
minimum
operator
con-
cept,
and
the
WQImin in
the
assessment
of
the
effects
of
pond
aquaculture
effluents
on
stream
ecosystem
and
suggested
that
the
deterioration
due
to
aquacultural
activity
can
be
easily
inferred
with
WQImin on
three
parameters:
dissolved
oxygen,
turbidity
and
total
phosphorus.
The
question
of
which
parameters
should
be
con-
sidered
and
how
they
are
selected
in
the
calculation
of
WQI
for
specific
assessments
of
the
impacts
of
aquaculture
effluents
on
stream
water
quality
remains
to
be
answered.
Therefore,
the
objec-
tives
of
this
study
were
to
determine
the
water
quality
of
stream
reaches
exposed
to
point
discharges
of
flow-through
trout
farms
and
diffuse
discharges
of
rural
community
and
agricultural
activi-
ties
and
to
show
a
way
of
selecting
the
indicator
parameters
for
the
WQIs
to
assess
the
trout
farming
impacts.
2.
Materials
and
methods
2.1.
Study
site
and
sampling
points
Es¸
en
Stream,
one
of
the
main
running
waters
of
the
western
Mediterranean
basin
in
Turkey,
discharges
into
the
Mediterranean
Sea.
The
streamflow
has
been
altered
markedly
for
the
purposes
of
hydroelectric
power
generation,
flood
control
and
irrigation.
The
flow
is
diverted
into
two
hydroelectric
power
plants
(HEPPs)
from
the
upstream
with
a
water
pipeline.
The
upper
reach
of
the
midstream
is
mainly
fed
by
a
ground
water
resource
(C¸
aygözü).
A
common
outlet
of
the
HEPPs
is
added
below
1
km
from
the
resource.
At
approximately
1
km
below
this
point,
a
control
gate
(Regulator)
dams
up
the
stream
and
diverts
the
water
into
an
open
channel
for
agricultural
irrigation
and
a
third
HEPP.
An
overflow
of
the
open
channel
capacity
discharges
into
the
main
streambed.
The
stream
taking
tributaries
of
the
midstream
and
the
downstream
finally
runs
into
the
sea
(Fig.
1).
The
basin
of
the
stream
is
a
significant
land-based
trout
culture
site.
There
are
more
than
50
flow-through
rainbow
trout
farms,
with
a
total
licensed
capacity
of
7700
tons
per
year
of
market
size
and
213
million
fry
per
year.
However,
the
farms
with
high
produc-
tion
capacities
are
generally
located
near
the
main
stream
channel
around
the
midstream
region.
Their
effluents
were
directly
dis-
charged
to
the
stream
without
any
solid
removal
and
treatment
during
the
study
period.
Nine
single-pass
flow-through
trout
farms
with
a
total
capacity
of
4400
tons
per
year
are
located
along
a
reach
of
2
km
between
C¸
aygözü
and
Regulator.
C¸
aygözü
represents
the
unaffected
reference
sampling
point
(S1),
whereas
the
regulator
site
represents
stream
reach
(S2)
affected
by
nine
intensive
trout
farms.
The
third
sampling
point
(S3)
is
fed
by
a
few
brooks
and
the
overflow
of
the
open
channel.
S3
is
largely
under
the
effects
of
rural
activities.
There
is
a
tributary
with
steep
slope
(Sögütlüdere
Brook)
discharges
between
S3
and
the
fourth
sampling
point
(S4)
where
a
trout
farm
with
a
capacity
of
950
tons
per
year
is
located.
Therefore,
S4
represents
a
combined
effect
of
steep
environmental
gradients
and
the
trout
farm
discharges.
The
fifth
(S5)
was
selected
from
downstream
reaches
as
a
self-purification
zone,
which
is
within
the
vicinity
of
intensive
agricultural
areas
plus
rural
settlements
(Fig.
1).
2.2.
Water
quality
analysis
The
samples
were
taken
monthly
between
March
2008
and
February
2009.
Temperature
(T),
dissolved
oxygen
(DO),
oxygen
Fig.
1.
Location
of
Es¸
en
Stream
and
sampling
points
(map
on
left),
and
a
schematic
view
of
the
nine
flow-through
trout
farms
between
first
and
second
sampling
points
(plot
on
right).
saturation
(SAT),
pH
and
electrical
conductivity
(EC)
were
mea-
sured
in
situ
by
YSI
55
and
YSI
63
field
instruments
(Yellow
Springs
Instrument
Co.,
Yellow
Springs,
OH).
Turbidity
(TUR)
was
measured
using
Hach
2100AN
model
benchtop
turbidimeter
(Hach
Company,
Loveland,
CO).
Total
suspended
solids
(TSS)
and
total
dissolved
solids
(TDS)
were
determined
by
filtration
and
then
dried
at
103–105 ◦C.
Calcium
(Ca2+),
magnesium
(Mg2+),
and
chloride
(Cl−)
were
determined
by
volumetric
titrimetry,
and
sulphate
(SO42−)
was
determined
by
spectrophotometry
using
the
turbidimetric
method.
Ammonia
nitrogen
(NH4+-N),
nitrite
nitrogen
(NO2−-N)
and
nitrate
nitrogen
(NO3−-N)
were
determined
by
phenate,
colorimetric,
and
cadmium
reduction
methods,
respectively.
Total
nitrogen
(TN)
was
analyzed
by
cadmium
reduction
after
persulfate
digestion.
Soluble
reactive
phosphorus
(SRP)
in
filtered
samples
and
total
inorganic
phosphorus
and
total
phosphorus
(TP)
after
hydrolysis
and
digestion
in
unfiltered
samples,
respectively,
were
determined
by
ascorbic
acid
method.
Total
organic
nitrogen
(TON)
and
total
organic
phosphorus
(TOP)
were
calculated
by
subtracting
the
inor-
ganic
fractions
from
total
concentrations.
The
Helios-␣
model
UV–vis
spectrophotometer
(Thermo
Scientific,
Cambridge,
United
Kingdom)
was
used
for
analyzing
of
nutrient
forms.
Biochemical
oxygen
demand
(BOD5)
was
determined
by
five
days
incubation
and
chemical
oxygen
demand
(COD)
with
open
reflux
method.
Total
coliforms
(TC)
and
fecal
coliforms
(FC)
were
determined
using
membrane
filtration
methods.
All
water
quality
analyses
were
per-
formed
according
to
Standard
Methods
for
the
Examination
of
Water
and
Wastewater
(APHA,
AWWA,
WEF,
1998).
Author's personal copy
674 M.A.T.
Koc¸
er,
H.
Sevgili
/
Ecological
Indicators
36 (2014) 672–
681
Table
1
The
relative
weights
and
the
normalization
factors
of
the
parameters.
VariablesaWeight
(Pi)
Normalization
factor
(Ci)
100
90
80
70
60
50
40
30
20
10
0
T
1
21/16
22/15
24/14
26/12
28/10
30/5
32/0
36/−2
40/−4
45/−6
45/<-6
pH
1
7
7–8
7–8.5
7–9
6.5–7
6–9.5
5–10
4–11
3–12
2–13
1–14
EC
1
<750
<1000
<1250
<1500
<2000
<2500
<3000
<5000
<8000
≤12,000
>12,000
DO
4
≥7.5 >7
>6.5
>6
>5
>4
>3.5
>3
>2
≤1
<1
TUR
2
<5
<10
<15
<20
<25
<30
<40
<60
<80
≤100
>100
TSS
4
<20
<40
<60
<80
<100
<120
<160
<240
<320
≤400
>400
TDS
2
<100
<500
<750
<1000
<1500
<2000
<3000
<5000
<10,000
≤20,000
>20,000
Ca2+ 1
<10
<50
<100
<150
<200
<300
<400
<500
<600
≤1000
>1000
Mg2+ 1
<10
<25
<50
<75
<100
<150
<200
<250
<300
≤500
>500
Cl2+ 1
<25
<50
<100
<150
<200
<300
<500
<700
<1000
≤1500
>1500
SO42−2
<25
<50
<75
<100
<150
<250
<400
<600
<1000
≤1500
>1500
SRP
1
<0.025
<0.05
<0.1
<0.2
<0.3
<0.5
<0.75
<1
<1.5
≤2
>2
TOPb1
<0.025
<0.05
<0.1
<0.2
<0.3
<0.5
<0.75
<1
<1.5
≤2
>2
TP
1
<0.2
<1.6
<3.2
<6.4
<9.6
<16
<32
<64
<96
≤160
>160
NH4+-N
3
<0.01
<0.05
<0.1
<0.2
<0.3
<0.4
<0.5
<0.75
<1
≤1.25
>1.25
NO2−-N
2
<0.005
<0.01
<0.03
<0.05
<0.1
<0.15
<0.2
<0.25
<0.5
≤1
>1
NO3−-N
2
<0.5
<2
<4
<6
<8
<10
<15
<20
<50
≤100
>100
TONc2
<0.05
<0.1
<0.2
<0.3
<0.4
<0.5
<0.7
<1
<2
≤3
>3
TNd2
<0.8
<3.8
<7.5
<13
<18
<27
<48
<85
<149
≤265
>265
BOD53
<0.5
<2
<3
<4
<5
<6
<8
<10
<12
≤15
>15
COD
3
<5
<10
<20
<30
<40
<50
<60
<80
<100
≤150
>150
TC
3
<50
<500
<1000
<2000
<3000
<4000
<5000
<7000
<10,000
≤14,000
>14,000
FC
3
<5
<50
<100
<200
<300
<400
<500
<700
<1000
≤1400
>1400
Adopted
from
Pesce
and
Wunderlin
(2000)
and
Kannel
et
al.
(2007).
aValues
in
mg/L,
T
as ◦C,
EC
as
S/cm,
TUR
as
NTU,
TC
and
FC
as
colonies/100
mL.
bNormalization
factor
and
relative
weight
selected
same
with
SRP
values.
cNormalization
factor
adopted
from
Kjeldahl
nitrogen
values
of
Regulation
on
Water
Pollution
Control
of
Turkey
(RWPC,
2004)
and
relative
weight
selected
same
with
TN
values
dCalculated
sum
of
inorganic
nitrogen
forms
and
total
organic.
2.3.
Water
quality
indices
WQIobj,
created
by
Rodriguez
de
Bassoon
(Pesce
and
Wunderlin,
2000),
was
used
for
assessment
of
water
quality
in
the
stream
reaches.
WQIobj =n
i=1CiPi
n
i=1Pi
where
n
is
the
total
number
of
parameters,
Ciis
the
value
assigned
to
parameter
i
after
normalization
and
Piis
the
relative
weight
assigned
from
1
to
4
to
each
parameter
that
has
the
most
impor-
tance
for
aquatic
life
preservation
(Table
1).
T,
pH,
EC,
DO,
TUR,
TSS,
TDS,
Ca2+,
Mg2+,
Cl−,
SO42−,
SRP,
TP,
NH4+-N,
NO2−-N,
NO3−-N,
TN,
BOD5,
COD,
TC
and
FC
were
used
for
calculation
of
the
index.
WQImin,
originally
derived
from
WQIobj (Kannel
et
al.,
2007;
Pesce
and
Wunderlin,
2000),
was
calculated
to
determine
the
Table
2
The
changes
of
the
water
quality
variables
in
the
sampling
points
in
Es¸
en
Stream
(data
are
means
±
SD
of
12
monthly
replicates;
T
in ◦C,
EC
in
S/cm,
TUR
in
NTU,
TC
and
FC
in
colonies/100
mL,
other
values
in
mg/L).
Variables
S1
S2
S3
S4
S5
T
13.0
±
0.3
13.8
±
1.9
16.6
±
3.5
17.4
±
3.6
18.5
±
4.5
pH
8.1
±
0.1
8.1
±
0.1
8.1
±
0.1
8.1
±
0.1
8.0
±
0.1
EC
267
±
7
271
±
13
304
±
28
330
±
67
375
±
65
DO
9.4
±
0.4
9.7
±
0.4
9.8
±
0.9
9.7
±
1.3
9.6
±
0.7
SAT
90
±
4
94
±
5
98
±
12
101
±
18
102
±
11
TUR
1
±
0
8
±
10
6
±
5
149
±
201
109
±
149
TSS
1
±
1
20
±
27
7
±
6
144
±
179
125
±
185
TDS
193
±
27
212
±
31
237
±
49
249
±
59
248
±
58
Ca2+ 52.5
±
5.5
51.0
±
7.0
54.0
±
5.9
57.9
±
5.5
48.0
±
6.4
Mg2+ 14.1
±
2.7
12.3
±
2.2
12.9
±
3.7
15.9
±
6.2
15.8
±
5.7
Cl−2.5
±
0.3
3.6
±
0.7
4.4
±
1.7
4.6
±
0.9
5.6
±
1.3
NH4+-N 0.004
±
0.007 0.284
±
0.298
0.028
±
0.029
0.024
±
0.032
0.006
±
0.007
NO2−-N
0.001
±
0.001
0.016
±
0.017
0.019
±
0.011
0.022
±
0.012
0.007
±
0.011
NO3−-N
0.346
±
0.114
0.528
±
0.225
0.640
±
0.304
0.776
±
0.318
0.795
±
0.296
TON
0.103
±
0.054
0.553
±
0.351
0.232
±
0.136
0.402
±
0.217
0.277
±
0.167
TN
0.461
±
0.051
1.408
±
0.459
1.035
±
0.251
1.285
±
0.489
1.181
±
0.393
SRP
0.011
±
0.007
0.060
±
0.063
0.028
±
0.026
0.029
±
0.024
0.018
±
0.008
TOP
0.016
±
0.021
0.240
±
0.138
0.152
±
0.114
0.213
±
0.085
0.160
±
0.098
TP
0.038
±
0.020
0.341
±
0.213
0.245
±
0.179
0.366
±
0.133
0.240
±
0.094
SO42−8.3
±
1.1
7.4
±
0.7
8.7
±
2.0
13.9
±
3.3
14.5
±
3.8
BOD54.6
±
0.8
6.2
±
0.3
6.1
±
0.9
6.0
±
0.6
5.6
±
0.8
COD
4.5
±
0.7
12.0
±
3.7
10.6
±
3.0
11.9
±
3.8
10.4
±
2.5
TC
340
±
285
11,765
±
8621
16,363
±
12,808
22,051
±
22,508
7581
±
3530
FC
0
±
0
32
±
37
33
±
32
33
±
15
42
±
31
Author's personal copy
M.A.T.
Koc¸
er,
H.
Sevgili
/
Ecological
Indicators
36 (2014) 672–
681 675
Table
3
Variation
explained
with
PCA
of
log-transformed
data
(,
eigenvalue;
Cum.
var.,
cumulative
variation).
Axis
High-flow
Low-flow
Moderate-flow
Annual
period
Cum.
var.
(%)
Cum.
var.
(%)
Cum.
var.
(%)
Cum.
var.
(%)
1
0.242
24.2
0.259
25.9
0.228
22.8
0.181
18.1
2
0.164
40.5
0.194
45.3
0.158
38.6
0.136
31.7
3
0.139
54.4
0.163
61.6
0.120
50.6
0.109
42.7
4
0.098 64.3 0.088
70.4
0.114
62.0
0.075
50.1
effects
of
trout
culture
effluents
as
an
alternative
to
WQIs
requiring
numerous
parameters.
WQImin =n
i=1CiPi
n
The
process
of
WQImin was
carried
out
with
three
different
ways.
In
the
first
way
(WQImin-a),
a
set
of
six
parameters
(DO,
BOD5,
TSS,
TP,
NH4+-N
and
TN)
that
are
mostly
associated
with
trout
farm
effluents
by
previous
studies
(Sindilariu,
2007;
Sindilariu
et
al.,
2009a,b;
Stewart
et
al.,
2006;
Tello
et
al.,
2010)
was
used.
The
sec-
ond
and
third
ways
(WQImin-band
WQImin-c,
respectively),
on
the
other
hand,
were
based
on
a
sequential
use
of
various
statistical
approaches
in
the
parameter
selection
as
detailed
further.
Both
WQIobj and
WQImin grade
the
water
quality
to
five
classes
in
point
of
index
scores:
excellent
(91–100),
good
(71–90),
medium
(51–70),
and
bad
(26–50)
and
very
bad
(0–25)
(Kannel
et
al.,
2007;
Pesce
and
Wunderlin,
2000).
2.4.
Statistical
analysis
The
data
were
first
stratified
according
to
flow
conditions,
i.e.
high,
low,
moderate-flows
and
annual
periods.
Average
flow
rate
during
the
high-flow
period
was
8.9
m3per
second
at
the
last
samp-
ling
point.
This
period
included
December,
January,
February
and
March.
Low-flow
period
was
representative
of
summer
and
early
autumn
months
(June,
July,
August
and
September),
and
average
flow
was
1.2
m3per
second.
Moderate-flow
period
was
a
combi-
nation
of
the
spring
and
autumn
months
(April,
May,
October
and
November)
with
an
average
flow
of
3.6
m3per
second
at
the
last
sampling
point.
We
used
the
principal
component
analysis
(PCA)
to
identify
the
strength
and
the
direction
of
the
variations
in
the
water
quality
parameters.
For
this
purpose,
the
normality
of
data
was
tested
for
each
parameter
using
the
Shapiro–Wilk
test.
Water
quality
data
which
were
not
normally
distributed
were
logarithmically
trans-
formed.
Parameters
that
were
loaded
as
the
main
gradients
of
PCA
were
selected
for
WQImin-band
WQImin-ccalculations.
The
relation
Fig.
2.
Coordinates
of
principal
component
axis
based
on
the
monitored
variables
and
the
sampling
points
(dots:
blue,
red,
green,
gray
and
black
denote
the
samples
of
sampling
points
1,
2,
3,
4
and
5,
respectively).
(For
interpretation
of
the
references
to
color
in
this
figure
legend,
the
reader
is
referred
to
the
web
version
of
this
article.)
Author's personal copy
676 M.A.T.
Koc¸
er,
H.
Sevgili
/
Ecological
Indicators
36 (2014) 672–
681
Fig.
3.
WQIobj scores
in
sampling
points
(values
not
sharing
a
common
letter
were
significantly
different
(P
<
0.05)).
Fig.
4.
WQImin-ascores
in
sampling
points
(values
not
sharing
a
common
letter
were
significantly
different
(P
<
0.05)).
Author's personal copy
M.A.T.
Koc¸
er,
H.
Sevgili
/
Ecological
Indicators
36 (2014) 672–
681 677
Fig.
5.
WQImin-bscores
in
sampling
points
(Values
not
sharing
a
common
letter
were
significantly
different
(P
<
0.05)).
of
water
quality
parameters
and
the
resulting
index
scores
was
checked
using
the
linear
regression.
WQI
scores
of
the
sampling
sites
were
compared
using
a
repeated
measures
ANOVA
assuming
an
additive
model
of
factors
Site
(fixed,
5
levels)
and
Time
(fixed,
4
levels
for
flow
periods
and
12
levels
for
the
whole
year).
Multiple
comparisons
between
factors
(Sites)
were
performed
by
Tukey’s
HSD
test.
GraphPad
Prism
5
(GraphPad
Software,
Inc.,
La
Jolla,
CA)
and
CANOCO
4.5
(Microcomputer
Power,
Ithaca,
NY)
were
used
for
data
processing
and
analysis.
3.
Results
and
discussion
The
parameters
pH,
DO
and
SAT
were
comparable
in
all
samp-
ling
points.
T,
TDS,
EC,
Cl−and
SO42−showed
an
increasing
trend
downstream.
Mg2+ increased
in
S4
and
S5,
but
Ca2+ decreased
in
S5.
TUR
and
TSS
increased
drastically
in
S4
during
the
rainy
season
as
a
result
of
high
suspended
matter
load
from
the
tributarial
stream.
TN,
TP,
BOD5,
and
COD
were
lower
in
the
reference
station
than
the
other
sampling
points.
NO3−-N
gradually
increased
downstream,
suggesting
a
growing
agricultural
rather
than
aquacultural
impact.
However,
NH4+-N,
TOP
and
TON
were
the
highest
in
S2.
FC
was
the
highest
in
S5
due
to
an
increased
rural
impact
(Table
2).
Four
main
axis
of
PCA
were
able
to
explain
only
less
than
three-
fourths
of
the
total
variation
due
to
a
complex
data
set
(Table
3).
The
fisrt
two
axis
provided
a
general
view
on
the
temporal
and
spa-
tial
trends
of
water
quality
(Fig.
2).
A
close
look
at
the
figure
reveals
that
the
samples
of
S1
characterized
with
the
lowest
concentrations
during
all
three
flows
and
annual
periods
were
plotted
at
the
oppo-
site
side
of
the
other
samling
points
(Fig.
2).
The
first
axis
of
PCAs
for
different
flow
conditions
and
annual
period
was
mostly
loaded
by
the
parameters
of
suspended
solids
(TSS
and
TUR),
organic
mat-
ter
(BOD5and
COD),
nitrogen
and
phosphorus
(NO2−-N,
NO3−-N,
TON,
TN,
SRP,
TOP
and
TP)
and
fecal
indicators
(TC
and
FC)
which
were
more
associated
with
downstream
reaches
(S4
and
S5)
but
slightly
with
S2.
The
second
axis
were
largely
loaded
by
T,
DO,
SAT
and
dissolved
constituents
(EC,
TDS,
Ca2+,
Mg2+,
Cl−,
SO42−),
which
were
closely
related
to
the
samples
of
S3,
S4,
and
S5.
The
scores
of
the
samples
of
S3
were
close
to
the
axis
origines,
implying
that
its
contribution
to
the
total
variation
was
less
than
the
other
sampling
points.
There
was
an
indication
that
the
samples
of
S2
were
separated
from
the
others,
and
mainly
associated
with
NH4+-N
but
to
a
lesser
extent
with
other
nitrogen
and
phosphorus
fractions
(TON,
TN,
SRP,
TOP,
TP).
The
coincidence
of
these
parameters
and
S2
may
be
specif-
ically
associated
with
the
impacts
of
intensive
trout
farms
in
the
study
area.
In
line
with
this
view,
PCA
suggests
an
overlap
of
S4
and
S5
with
NO3−-N,
FC,
COD,
Cl−,
and
TC,
implying
that
the
occur-
rence
of
agricultural
and
rural
impacts
in
S4
and
S5
may
be
the
case.
At
this
juncture,
it
can
also
be
suggested
that
the
aquacultural
impacts
in
S2
was
masked
by
the
agricultural
and
rural
influence
in
farther
downstream
reaches.
Because
of
this
masking
effect,
WQIobj did
not
clearly
discrim-
inate
S2
from
the
downstream
sampling
points
(Fig.
3).
WQIobj
classified
the
reference
point
as
“excellent”
with
average
index
scores
of
in
a
range
of
92
and
93.
S1
was
significantly
differ-
ent
from
the
other
sampling
points
(with
the
exception
of
S3
in
high-flow
period)
during
high,
low
and
moderate-flow
and
annual
periods
(F4,12 =
9.8,
F4,12 =
9.7,
F4,12 =
13.8,
F4,44 =
24.1,
respectively,
P
<
0.001).
However,
the
overall
water
quality
of
the
affected
samp-
ling
points
from
water
and
land
uses
was
classified
as
“good”
in
the
range
of
72
and
85
average
index
scores.
WQIobj could
not
detect
a
significant
difference
between
the
affected
sampling
points
in
general
when
the
differences
between
S3
and
S4
in
high-flow
and
annual
periods
were
neglected.
WQIobj did
not
support
the
results
of
PCA
simply
because
it
is
calculated
using
many
variables
to
assess
the
overall
water
quality.
This
speculation
is
further
supported
by
Author's personal copy
678 M.A.T.
Koc¸
er,
H.
Sevgili
/
Ecological
Indicators
36 (2014) 672–
681
Fig.
6.
Regression
diagrams
between
WQImin-band
the
parameters
used
in
calculation.
the
findings
of
Ferreira
et
al.
(2011),
who
estimated
water
quality
using
the
WQI
of
the
Canadian
Council
of
Ministers
of
the
Environ-
ment
by
including
a
large
set
of
parameters
and
did
not
detect
a
significant
difference
between
qualities
of
affected
and
unaffected
sites
from
shrimp
farms.
Therefore,
there
is
a
clear
need
for
the
parameter
selection
and
data
reduction
to
clarify
the
aquacultural
impacts,
which
is
the
main
impetus
for
the
present
study.
Data
selection
and
reduction
by
omitting
the
unnecessary
vari-
ables
can
be
an
option
to
make
WQIobj simpler.
In
this
study,
we
arbitrarily
removed
the
unnecessary
parameters
and
left
only
six
parameters
(DO,
BOD5,
TSS,
TP,
NH4+-N
and
TN)
to
calculate
WQImin-aconsidering
that
these
variables
are
major
impacts
of
intensive
trout
farming
on
the
effluent
water
quality
(Bergheim
and
Brinker,
2003;
Sindilariu
et
al.,
2009a;
Koc¸
er
et
al.,
2013).
The
WQImin-ain
this
way
yielded
more
and
less
similar
scores
to
those
by
WQIobj.
However,
WQImin-awas
not
able
to
separate
the
reference
point
and
the
others
to
a
large
extent,
though
some
sig-
nificant
differences
for
the
four
sets
of
data
was
the
case
(Fig.
4).
This
finding
was
further
supported
with
a
highly
significant
cor-
relation
(r2=
0.72;
P
<
0.001;
N
=
59)
between
the
WQIobj and
the
WQImin-aindex
scores
(not
shown).
One
explanation
of
this
can
be
that
since
both
urban/rural
and
agricultural
activities
are
also
the
major
sources
of
the
parameters
included
in
the
calculation
of
WOImin-ain
rivers
and
streams
(Carpenter
et
al.,
1998;
Jarvie
et
al.,
2008;
Mander
et
al.,
2000;
Moss,
2008;
Neal
and
Jarvie,
2005),
it
may
not
have
clearly
separated
the
affected
sampling
points
in
the
present
study.
Indeed,
downstream
reaches
showed
lower
index
scores
than
S2
probably
due
to
high
contribution
of
NO3−-N
to
TN
concentrations.
Moreover,
high
TSS,
originated
from
the
tributary,
seems
to
be
one
of
the
main
contributors
for
the
index
scores
of
the
downstream
reaches.
Therefore,
TSS
could
be
neglected
because
its
concentrations
may
be
dictated
by
natural
dynamics,
as
previously
mentioned
for
this
study
and
elsewhere
(Meybeck,
2003).
Instead
of
an
arbitrary
selection
of
the
parameters
in
the
cal-
culation
of
WQImin,
we
assumed
that
PCA
results
would
be
more
appropriate
way
to
select
the
representatives
of
aquacultural
impacts
in
the
study
site.
The
results
suggested
that
SRP,
TON,
and
TOP
along
with
NH4+-N
were
more
associated
with
the
sampling
point
(S2)
as
a
result
of
the
trout
farm
discharges
as
previously
stated.
A
calculation
of
WQImin based
on
these
four
parameters
(WQImin-b)
classified
water
quality
in
S2
as
“medium”
(average
scores
between
58
and
63
in
three
different
flow
conditions
and
annual
period)
whereas
other
impacted
points
as
“good”
(the
index
scores
between
76
and
88).
These
results
seemed
to
be
a
more
accurate
in
terms
of
characterizations
of
the
farm
effluents
(Fig.
5).
However,
there
was
not
a
clear
statistically
significant
discrimina-
tion
of
S2
from
the
other
sampling
points
for
all
flow
conditions
probably
due
to
variations
in
amounts
of
fish
production
and
flow
regime.
In
fact,
annual
trout
production
in
the
study
site
is
per-
formed
in
two
periods.
Juvenile
trout
with
an
average
individual
weight
of
4
g
are
stocked
into
the
raceways
twice
a
year
at
early
win-
ter
and
summer,
and
the
biomass
reaches
the
market
size
within
6–7
months.
It
means
that
higher
amounts
of
feed
are
used
dur-
ing
the
spring
and
autumn
(i.e.
moderate-flow
period),
when
more
waste
is
generated,
while
a
lower
feed
use
and
waste
generation
take
place
high
flow
and
low-flow
periods.
Therefore,
the
seasonal
Author's personal copy
M.A.T.
Koc¸
er,
H.
Sevgili
/
Ecological
Indicators
36 (2014) 672–
681 679
Fig.
7.
WQImin-cscores
in
sampling
points
(values
not
sharing
a
common
letter
were
significantly
different
(P
<
0.05)).
variability
in
WQI
scores
was
inevitable,
and
thus
lower
scores
for
S2
were
generally
recorded
during
moderate-flow
period.
But,
it
should
be
underlined
that
we
could
not
directly
correct
WQI
scores
depending
on
the
seasonal
changes
in
the
farm
activities
as
we
did
not
reach
the
pertinent
records.
There
were
strong
and
significant
correlations
between
NH4+-
N
and
TON
concentrations
and
the
index
scores,
indicating
that
they
make
a
remarkable
contribution
to
WQImin-b(Fig.
6).
Even
though
an
interference
of
point
or
nonpoint
sewage
and
agricul-
tural
discharges
may
not
be
ruled
out,
WQImin-bprovided
a
more
precise
approach
than
WQIobj and
WQImin-ain
assessing
the
effects
of
trout
farm
effluents
on
the
water
quality
of
the
receiving
stream,
particularly
in
agriculture-dominated
rural
catchments.
The
bet-
ter
performance
of
WQImin-bcould
be
because
NO3−-N
and
SRP
are
the
dominant
fractions
of
TN
and
TP,
respectively,
in
catch-
ments
under
the
effects
of
the
diffuse
discharges
from
agriculture
as
well
as
treated
and
untreated
sewage
wastes
of
rural
community
(Jarvie
et
al.,
2008;
Neal
and
Jarvie,
2005).
In
effluents
of
trout
farms,
these
two
nutrient
forms,
although
variable
to
a
certain
degree
(Sindilariu,
2007),
are
dominated
by
NH4+-N
as
well
as
TON
and
TOP,
respectively
(Foy
and
Rosell,
1991;
Roque
d’Orbcastel
et
al.,
2008).
TP
in
trout
farm
effluents
can
be
removed
up
to
80%
via
sedimentation
and
mechanical
filtration,
but
the
removal
of
TN
is
highly
difficult
(Sindilariu,
2007;
Sindilariu
et
al.,
2009b);
thereby
the
ratio
of
NH4+-N
to
TN
may
increase
up
to
79%
(Dalsgaard
and
Pedersen,
2011).
Nitrogen
excretions
in
the
form
of
urea
and
amino
acids
may
comprise
a
considerable
amount
of
organic
nitrogen
(Kajimura
et
al.,
2004).
In
fact,
the
ratios
of
NH4+-N
and
TON
to
TN
Fig.
8.
Regression
diagrams
between
WQImin-band
WQImin-cscores
(left:
based
on
data
of
all
the
sampling
points;
right:
based
on
data
of
S2).
Author's personal copy
680 M.A.T.
Koc¸
er,
H.
Sevgili
/
Ecological
Indicators
36 (2014) 672–
681
Fig.
9.
Regression
diagrams
between
WQImin-band
WQINH4and
WQITON scores
(left:
based
on
data
of
all
the
sampling
points;
right:
based
on
data
of
S2).
were
reported
to
range
between
50%
and
65%
in
trout
farm
efflu-
ents
in
our
study
site
(Koc¸
er
et
al.,
2013).
That
is
why
the
linear
regression
detected
higher
correlations
of
NH4+-N
and
TON
con-
centrations
with
WQImin-bfor
all
sampling
periods
compared
with
SRP
and
TOP
in
the
presently
reported
study.
The
results
suggest
that
WQImin-bwould
be
a
useful
approach
for
assessment
of
trout
farm
impacts
on
the
receiving
stream
reaches,
and
NH4+-N
and
TON
seem
to
be
better
indicators
for
aquacultural
impacts
as
opposed
to
the
suggestion
of
Simões
et
al.
(2008),
who
noted
that
TP
could
be
the
typical
indicator.
Therefore,
as
a
next
step
in
the
parameter
selection
we
cal-
culated
WQImin based
on
NH4+-N
and
TON
(WQImin-c).
WQImin-c
produced
highly
similar
scores
for
all
sampling
points
to
WQImin-b
(Fig.
7).
A
significant
differentiation
between
S2
and
the
other
affected
sampling
sites
was
the
case
only
for
annual
data
(F4,44 =
14.4;
P
<
0.001).
However,
WQImin-cclearly
separated
the
sampling
sites
with
respect
to
their
water
quality
status,
and
rated
S1,
S2
and
the
other
affected
sites
as
“excellent”,
“medium”
and
“good”
classes,
respectively
for
all
different
flow
conditions
and
annual
period.
The
good
consistencies
between
WQImin-band
WQImin-ccalculated
for
all
the
sampling
points
as
well
as
only
for
S2
were
further
supported
by
the
strong
linear
relation
(Fig.
8),
sug-
gesting
an
effective
minimization
of
the
parameters
using
NH4+-N
and
TON
is
possible
and
more
relevant.
An
attempt
at
the
index
calculation
(WQINH4,
WQITON)
using
only
a
single
parameter
(NH4+-N
or
TON),
which
is
actually
an
environmental
quality
standard
(EQS)
(see
Table
1),
resulted
in
higher
water
quality
classes
of
the
sampling
points
than
WQImin-b
(not
shown).
Moreover,
WQINH4 appeared
to
have
a
higher
cor-
relation
with
WQImin-bthan
WQITON (Fig.
9).
Although
WQINH4
had
a
good
association
with
WQImin-bat
S2,
there
was
no
a
reasonable
relation
for
all
the
sampling
points
when
compared
with
WQImin-c.
This
may
suggest
that
the
EQS
of
NH4+-N
alone
seemed
to
have
been
a
good
indicator
for
assessment
of
aqua-
cultural
impacts
at
the
discharge
site,
but
it
was
still
insufficient
for
assessment
of
downstream
impacted
sites
(Fig.
9)
probably
due
to
interference
of
several
other
factors
such
as
nitrifica-
tion,
dilution,
diffuse
loads
from
agriculture
and
rural
activities
etc.
4.
Conclusions
This
study
clearly
indicates
that
the
WQIobj calculation
based
on
too
many
variables
fails
to
differentiate
the
affected
stream
reaches
from
the
discharges
of
trout
farms.
An
arbitrary
selec-
tion
of
the
most
influenced
parameters
in
the
farm
effluent
for
WQImin calculation
also
appears
as
an
unsuccessful
attempt.
However,
a
selection
of
parameters
based
on
PCA
findings
for
the
index
calculation
is
much
more
decisive
in
discerning
the
affected
stream
reach
from
the
intensive
trout
farms.
WQImin
calculation
with
NH4+-N,
TON,
SRP
and
TOP
concentrations
is
a
suitable
method
for
assessment
of
aquacultural
impacts.
But,
NH4+-N
and
TON
are
more
representative
parameters
of
aqua-
cultural
impacts
compared
with
others.
WQImin calculation
using
these
two
parameters
will
be
a
cost,
time
and
effort-saving
way
that
is
a
fundemantal
aspect
of
effective
water
quality
monitoring
Author's personal copy
M.A.T.
Koc¸
er,
H.
Sevgili
/
Ecological
Indicators
36 (2014) 672–
681 681
programs.
This
index
is
also
a
better
assessment
method
for
aquaculture
affected
stream
reaches
than
an
EQS
of
a
single
param-
eter.
The
prensent
approach
can
find
a
widespread
use
in
the
assessment
of
the
specific
impacts
of
land
use
on
the
receiving
water
quality.
Future
studies
may
consider
an
improvement
of
the
index
weighting
on
seasonality
of
fish
biomass
and
stream-
flow.
Acknowledgements
This
study
was
supported
by
grants
from
The
Scientific
and
Technical
Research
Council
of
Turkey
(TUBITAK,
Project
No.
107Y084).
The
authors
thank
Ramazan
Uysal,
Faruk
Pak,
Adil
Yılayaz,
Mahir
Kanyılmaz,
Ahmet
Mefut,
Nesrin
Emre,
Gazi
Uysal,
Aybike
Topc¸
uo˘
glu
and
Gül
Tunc¸
Karaa˘
gac¸
for
their
invaluable
help
with
data
collection
and
analysis,
and
Prof.
Dr.
Ays¸
e
Muham-
meto˘
glu
and
Assoc.
Prof.
Yılmaz
Emre
for
their
suggestions
on
this
article.
References
Akkoyunlu,
A.,
Akıner,
M.E.,
2012.
Pollution
evaluation
in
streams
using
water
qual-
ity
indices:
a
case
study
from
Turkey’s
Sapanca
Lake
Basin.
Ecol.
Indic.
18,
501–511.
APHA,
AWWA,
WEF,
1998.
Standard
Methods
for
the
Examination
of
Water
and
Wastewater,
20th
ed.
American
Public
Health
Association,
Washington,
DC.
Bakan,
G.,
Özkoc¸
,
H.B.,
Tülek,
S.,
Cüce,
H.,
2010.
Integrated
environmental
quality
assessment
of
Kızılırmak
river
and
its
coastal
environment.
Turkish
J.
Fish
Aquat.
Sci.
10,
453–462.
Bergheim,
A.,
Brinker,
A.,
2003.
Effluent
treatment
for
flow
through
systems
and
European
Environmental
Regulations.
Aquac.
Eng.
27,
61–77.
Boyacıo˘
glu,
H.,
2010.
Utilization
of
the
water
quality
index
method
as
a
classification
tool.
Environ.
Monit.
Assess.
167,
115–124.
Boyacıo˘
glu,
H.,
Boyacıo˘
glu,
H.,
2007.
Surface
water
quality
assessment
by
environ-
metric
methods.
Environ.
Monit.
Assess.
131,
371–376.
Carpenter,
S.R.,
Caraco,
N.F.,
Correll,
D.L.,
Howarth,
R.W.,
Sharpley,
A.N.,
Smith,
V.H.,
1998.
Nonpoint
pollution
of
surface
waters
with
phosphorous
and
nitrogen.
Ecol.
Appl.
8,
559–568.
CCME,
2001.
Canadian
Water
Quality
Guidelines
for
the
Protection
of
Aquatic
Life:
Canadian
Water
Quality
Index
1.0
Technical
Report.
Canadian
Environmental
Quality
Guidelines,
Winnipeg,
Manitoba.
Dalsgaard,
A.J.T.,
Pedersen,
P.B.,
2011.
Solid
and
suspended/dissolved
waste
(N,
P,
O)
from
rainbow
trout
(Oncorynchus
mykiss).
Aquaculture
313,
92–99.
Debels,
P.,
Figueroa,
R.,
Urrutia,
R.,
Barra,
R.,
Niell,
X.,
2005.
Evaluation
of
water
quality
in
the
Chillan
River
(Central
Chile)
using
physicochemical
parameters
and
a
modified
water
quality
index.
Environ.
Monit.
Assess.
110,
301–322.
FAO,
2010.
The
State
of
World
Fisheries
and
Aquaculture
2010.
Food
and
Agricul-
ture
Organization
of
the
United
Nations,
Fisheries
and
Aquaculture
Department,
Rome.
Ferreira,
N.C.,
Bonetti,
C.,
Seiffert,
W.Q.,
2011.
Hydrological
and
Water
Quality
Indices
as
management
tools
in
marine
shrimp
culture.
Aquaculture
318,
425–433.
Folke,
C.,
Kautsky,
N.,
1992.
Aquaculture
with
its
environment:
prospects
for
sus-
tainability.
Ocean
Coast
Manag.
17,
5–24.
Foy,
R.H.,
Rosell,
R.,
1991.
Fractionation
of
phosphorus
and
nitrogen
loadings
from
a
Northern
Ireland
fish
farm.
Aquaculture
96,
31–42.
Jarvie,
H.P.,
Withers,
P.J.A.,
Hodgkinson,
R.,
Bates,
A.,
Neal,
M.,
Wickham,
H.D.,
Har-
man,
S.A.,
Armstrong,
L.,
2008.
Influence
of
rural
land
use
on
streamwater
nutrients
and
their
ecological
significance.
J.
Hydrol.
350,
166–186.
Kajimura,
M.,
Croke,
S.J.,
Glover,
C.N.,
Wood,
C.M.,
2004.
Dogmas
and
controversies
in
the
handling
of
nitrogenous
wastes:
the
effect
of
feeding
and
fasting
on
the
excretion
of
ammonia,
urea
and
other
nitrogenous
waste
products
in
rainbow
trout.
J.
Exp.
Biol.
207,
1993–2002.
Kannel,
P.R.,
Lee,
S.,
Lee,
Y.-S.,
Kanel,
S.R.,
Khan,
S.P.,
2007.
Application
of
Water
Quality
Indices
and
dissolved
oxygen
as
indicators
for
river
water
classification
and
urban
impact
assessment.
Environ.
Monit.
Assess.
132,
93–110.
Koc¸
er,
M.A.T.,
Kanyılmaz,
M.,
Yılayaz,
A.,
Sevgili,
H.,
2013.
Waste
loading
into
a
regulated
stream
from
land-based
trout
farms.
Aquacult.
Environ.
Interact.
3,
187–195.
Lermontov,
A.,
Yokoyama,
L.,
Lermontov,
M.,
Machado,
M.A.S.,
2011.
A
fuzzy
water
quality
index
for
watershed
quality
analysis
and
management.
In:
Broniewicz,
E.
(Ed.),
Environmental
Management
in
Practice.
InTech,
Crotia,
pp.
387–410.
Mander,
Ü.,
Kull,
A.,
Kuusemets,
V.,
2000.
Nutrient
flows
and
land
use
change
in
a
rural
catchment:
a
modelling
approach.
Landsc.
Ecol.
15,
187–199.
Massoud,
M.A.,
2010.
Assessment
of
water
quality
along
a
recreational
section
of
the
Damour
River
in
Lebanon
using
the
water
quality
index.
Environ.
Monit.
Assess.
184,
4151–4160.
Meybeck,
M.,
2003.
Global
analysis
of
river
systems:
from
Earth
system
controls
to
Anthropocene
syndromes.
Philos.
T.
R.
Soc.
B
358,
1935–1955.
Moss,
B.,
2008.
Water
pollution
by
agriculture.
Philos.
T.
R.
Soc.
B
363,
659–666.
Neal,
C.,
Jarvie,
H.P.,
2005.
Agriculture,
community,
river
eutrophication
and
the
Water
Framework
Directive.
Hydrol.
Process.
19,
1895–1901.
O’Bryen,
P.J.,
Lee,
C.S.,
2003.
Management
of
aquaculture
effluents
workshop
discus-
sion
summary.
Aquaculture
226,
227–242.
Pesce,
S.F.,
Wunderlin,
D.A.,
2000.
Use
of
water
quality
indices
to
verify
the
impact
of
Cordoba
city
(Argentina)
on
Suqu ´
ya
River.
Water
Res.
34,
2915–2926.
Roque
d’Orbcastel,
E.,
Blancheton,
J.P.,
Boujard,
T.,
Aubin,
J.,
Moutounet,
Y.,
Przybyla,
C.,
Belaud,
A.,
2008.
Comparison
of
two
methods
for
evaluating
waste
of
a
flow
through
trout
farm.
Aquaculture
274,
72–79.
Rosemond,
S.,
Duro,
D.C.,
Dubé,
M.,
2009.
Comparative
analysis
of
regional
water
quality
in
Canada
using
the
Water
Quality
Index.
Environ.
Monit.
Assess.
156,
223–240.
RWPC,
2004.
Regulation
of
Water
Pollution
Control.
Official
Gazette
of
Republic
of
Turkey
No.
25687
In
Turkish.
Said,
A.,
Stevens,
D.K.,
Sehlke,
G.,
2004.
An
innovative
index
for
evaluating
water
quality
in
streams.
Environ.
Manag.
34,
406–414.
Simões,
F.D.,
Moreira,
A.B.,
Bisinoti,
M.C.,
Gimenez,
S.M.N.,
Yabe,
M.J.S.,
2008.
Water
quality
index
as
a
simple
indicator
of
aquaculture
effects
on
aquatic
bodies.
Ecol.
Indic.
8,
476–484.
Sindilariu,
P.D.,
2007.
Reduction
in
effluent
nutrient
loads
from
flow-through
facil-
ities
for
trout
production:
a
review.
Aquac.
Res.
38,
1005–1036.
Sindilariu,
P.D.,
Reiter,
R.,
Wedekind,
H.,
2009a.
Impact
of
trout
aquaculture
on
water
quality
and
farm
effluent
treatment
options.
Aquat.
Living
Resour.
22,
93–103.
Sindilariu,
P.D.,
Brinker,
A.,
Reiter,
R.,
2009b.
Waste
and
particle
management
in
a
commercial,
partially
recirculating
trout
farm.
Aquac.
Eng.
41,
127–135.
Stewart,
N.T.,
Boardman,
G.D.,
Helfrich,
L.A.,
2006.
Treatment
of
rainbow
trout
(Oncorhynchus
mykiss)
raceway
effluent
using
baffled
sedimentation
and
artifi-
cial
substrates.
Aquac.
Eng.
35,
166–178.
Tacon,
A.G.J.,
Forster,
I.P.,
2003.
Aquafeeds
and
the
environment:
policy
implications.
Aquaculture
226,
181–189.
Tello,
A.,
Corner,
R.A.,
Telfer,
T.C.,
2010.
How
do
land-based
salmonid
farms
affect
stream
ecology?
Environ.
Pollut.
158,
1147–1158.
Wu,
R.S.S.,
1995.
The
environmental
impact
of
marine
fish
culture:
towards
a
sus-
tainable
future.
Mar.
Pollut.
Bull.
31,
159–166.
Zhang,
M.,
Zhang,
M.,
2007.
Assessing
the
impact
of
leather
industries
on
the
qual-
ity
of
water
discharged
into
the
East
China
Sea
from
Wenzhou
Watersheds.
J.
Environ.
Manage.
85,
393–403.
