Effects of Fertilization and Planktivorous Fish on Epilimnetic ...

opossumoozeMechanics

Feb 21, 2014 (3 years and 3 months ago)

58 views

Ef i eds
of
Ferti
ization
and
P
anktivorous
Fish
on
Epi
Phosphorus
and
Phosphorus
Sedimentation in Large
Enc
Asit
Mazumder,
W.
D.
Taylor,
D.
1.
McQueenT,
D.
R.
S.
LeanZ
Department
sf
Biology,
University
of Waterloo,
Waterloo,
One.
N2L
3C1
Mazurnder,
A.,
W. D.
Taylor,
D.
1.
McQueen,
and
D.
R.
%.
Lean.
1989.
Effects of fertilization and
planktivorous
fish
on epilimnetic phosphorus and phosphorus sedimentation in large enclosures. Can.
1.
Fish.
Aquat.
Sci,
46:
1735-1 742.
Enclosure experiments were conducted during 1986 and
1987
to test the hypothesis that predation-induced shifts
in plankton size-structure affect epilimnetic
total
phosphorus
(TP)
through changes in sedimentation rates, both
with and without external nutrient loading. Fertilization and/or
planktivorousfish
were applied to eight enclosures
in a
2
x
2
factorial design.
Epilirnnetic
TP
was higher with fish in the unfertilized enclosures and Bower with fish
in the fertilized enclosures; that is, fertilization increased epilimnetic
TP
only in enclosures without fish, and
much less than expected based on the amounts added. The fraction of
TP
sdirnenting
per unit time, and the
spring to summer decline in TP, was consistently lower with fish. Sedimentation rates were
poorly
correlated
with epilirnnetic
TP,
but were more strongly
correlated
with
TP
decline rates and particulate phosphorus larger
than
20
pm.
Presence
or
absence of planktivorous fish can alter the trophic status
of
planktonic systems, and the
impact of external nutrient loading
by
changing the size-distribution and biomass of plankton.
En 1986 et 1987, des experiences ont
4t6
reali&es
en
enclos
poker
verifier
i'hypoth2se
que
les
modifications de
la structure par taille
du
plancton
dues
2
la predation influent sur
le
phospkore total
(PP)
de
l'epilimwion
en
provoquant
des
changements dans
les
taux de
s6dimentation,
qu'il
y
ait
ou
non
apport externe
de
substances
nutritives.
Dans
ie
cadre d'un plan
d1exp6riences
factoriel
2
x
2, on a
mis
des
fertilisants
et
su
des poissons
planctivores
dans
huit
enclose
Le
PP de
I'epilimnisn
4tait
plus
eleve
quand
il
y avait
du
poisson dans les enclos
non
fertilis4s
et plus
faible
quand
il
y
avait eu
du
poisssn
dans les enclos
fertilis6s;
autrement dit, la
fertilisation
a
fait
accroitre
le
PT
de
If6pilimnion
seulernent
dans
les
enclos ne contenant pas
de
psisson
et
beaucoup
rnoins
que
prevu
d'apres
les
quantit6s
ajout6es.
La
partie
de
PT
qui
se
depose
par unit4 de temps et la baisse de
BT
sbserv6e
du
psintemps
A
Pet6
etaient
de
faqon
constante plus
faibles
lorsqu'il
y avait du poisson. Les taux de
sedimentation
corr6Ies
avec
le
PP
de
1'4pilimnion,
mais
il
I'4taient
plus fortement avec
les
taux
de
diminution
du
PT
et
le
phosphote
particulaire
d'un
diarn2tre
superieur
2
20
pm.
La
pr6sence
ou
['absence
de
poissons
planctonivores peut modifier
le
statut trophique des
csmmunaut6s
planctoniques et
les
r4percussions
de
I'apport
externe de substances nutritives en changeant la
repartition
par taille et la
bismasse
du
plancton.
Received
March
38,
T
988
Accepted May
29,
1989
(J9658)
R
educing
phosphoms
input
has been used
successfully
to
reverse lake eutrophication
(VoIIenweider
1968:
Edmcsndson
1969;
Schindler
1975,
1877;
Hurley
et
al.
1986).
However,
the
emphasis on external phosphorus loading
as a
determinant
of the
phosphoms
content
of
lakewaters
and
eutrophication has
obscured
the
hport mce
of
food web inter-
actions. Wright
md
S
hapiro
(1
984)
reported
that
epilimetic
total
phosphoms (TP) declined following a shift
from
small-
bodied
to large-bodied herbivorous
zoopla&ton,
and con-
cluded
that vertical migration of
zoopl dt on
may be
respon-
sible for
the
transport
of
phosphoms
to the
hypl i mi on.
How-
ever, they
did
not
measure
the sedimentation of
material
in their
manipulation
experiments.
Earlier studies have suggested that sedimentation
of
particles
from the
epilimnion
is a major process
reguiding
epilimetic
'IT
(White
and
Wetzel
1975;
Chalton
1975;
Fallon
md
Brock
1988),
but the mechanisms
contro1linag
sedimentation
rates have
rarely
ke n
studied. Sinking
of
algae
has
been
reported
as
an
important
factor
controlling
their
abundance in the
epilimnion,
Regu
le
30
mars
7988
Accept6
le
29
mai
7989
especially in the spring when diatoms predominate
(Ulen
1978;
S o me r
1984;
Scavia and
Fahnenstiel
1984). Grazing by her-
bivorous
zooplankton
is also
known
to effect the sedimentation
of algae (via
faecal
pellet
production)
from the
epilimion
(Biirgi
et
al.
1978;
Scavia
and
Fahnenstiel
1987;
UehIinger
and
Bloesch
1984).
Although some
work
has been done
on
sedimentation in
lakes
as
affected
by
increased
external
loading
(Chaltow
1975) and
by changes in
cornunity
structure
(Biirgi
et
d.
1979;
Uehlin-
ger
and
Bloesch
1987),
the importance of sedimentation in
cow-
trolling epilimnetic
TP,
and
the
mwhmisms
eontroling
sedi-
mentation,
are
pooriy
understood.
Here
we
examhe
temporal
changes
in epilimnetic TP, and
their
relationship to phosphoms
sedimentation, in
large
enclosures.
We test the hypothesis that
predat i on-i ndu
shifts in
plankton
size-structure
affect
sedi-
mentation rates
and,
therefore,
epilimetig:
TP,
both with
md
without external nutrient loading.
Materids
and
Methods
'Department
of
Biology,
York
University,
North
York,
Ont.
~ ~ i ~ ~ ~ ~ l
~~~i~~
M3J
183.
'National
Water
Research
Institute,
Box
5050,
Burlington,
Ont.
This study was
conducted
during the summers
of
1986
md
L7W
4A6.
1987
a
Lake
St.
George,
Ontario.
Eight
large
experimental
Can.
J.
Fish.
Aqua.
Scial
Vd.
46,
6989
1735
Can. J. Fish. Aquat. Sci. Downloaded from www.nrcresearchpress.com by UNIV VICTORIA on 07/13/11

For personal use only.

enclosures (8 m diameter,
15
m deep, and open
a
the sediment)
were used to evaluate two treatments, nutrient
and/or
plmktivorous
fish addition, in a
2
X
2
factorial design (see
Mmumder
et al. 1988 for details).
kt ai l s
of
the enclosures we
described
in
Post and
McQueen
(1987). Prior to the
expe~ments,
fish were removed from
all
enclosures with a
purse-seine
md,
to create similar
preexperimental
conditions,
the enclosures were flushed with
600
r nhf
water from the
lake.
Depths
of the thermoclines in the enclosures, defined as depth
of maximum temperature
change,
were between
4
md
7
rn
in
late
spring and summer.
The
enclosures were divided into two groups,
nonfertilized
md
fertilized. Nutrients
(N
and P,
13:
1) were
dissolved
in water
md
were added to the epilimnion
(0-4
m)
using a
garden
hose.
The hose was kept in
constant
motion,
both vertically and
ho~zontally,
to ensure even distribution of nutrients in the entire
epilimiom.
The
fertilized enclosures received weekly additions
of
3e69
rng
Pm-'-d-'
(added as
H,PO,)
and
48.3
mg
Nmm-2*d-1
(added as
NaNO,)
from April
28
to August 12 in
1986, and May 6
b
August
10
in 1987. Each weekly
addition
was enough to increase
epilimnetic
TP
and total nitrogen (TN)
by
6.46
gdlg
P-L-l
and 84.54
pg
NoL-'.
Two
enclosures
from each group (unfertilized and
fertlized)
were stocked
with
62.848.6
kg
-ha
-
'
sf
I
+
yellow perch
(86-
94
fish per enclosure
with
a mean wet weight of 3.65
g)
on
May 31 (1986) and with 35-38
kg-ha-"(50-55
fish per
enclosure with mean wet weight of 2.92
g)
on
April
27 (1987).
All added fish were
marked,
and were recaptured
t hee
times
during each summer. Fish killed by handling were replaced with
similar-sized fish from the
lake.
No fish were lost, and only
three small
unmaked
fish were found and removed.
Unfertilized enclosures with and
without
fish were designated
as
'
+
F9
and
6controH9
',
while fertilized enclosures with and
without fish were designated as
'
-+
9NF9
',
and
'
'
9
Nu'.
Details
of
experimental
design
are
described
in Mazumder et al.
(1988).
Determination
of Epilimnetic
Phosphoms
(TP)
Dissolved
phosphoms
and
pagtiiculate
phosphoms
in differ-
ent
size-fractions
were measured in replicate
0-4
m
tube sam-
ples
(6.5
cm
diameter) on six sampling dates
in
1986 and
four
dates
in
1987 (May
though
August).
Dissolved
and particulate
phosphoms in
10
size-classes
( c8.2
pm
to
>400
pm)
were
analyzed
following filter fractionation (see
Mazumder
et
al.
1988
for details). Concentration of
phosphorus
in each
size-
class was
detemined
after oxidation under pressure
(Menzel
and
Cowin
1965) with the ascorbic acid
modification
of the
molybdemm
blue method
(Stsickland
and
Pwsons
1972.
Epi-
limnetie
TB
concentration was
determined
by adding the con-
centrations of phosphorus in all size-classes.
TP
was also deter-
mined for
triplicate
whole water samples collected with Van
Dom
bottles from
8
m.
Accumulation of phosphorus by fish
was measured by analyzing
phosphoms
contents
of
15 fish
on
three dates in 1986 and
198'7
(initial, 17 July, and
25
August).
Loss of
phosphoms
into
pehiphyton
growth was measured by
analyzing the
phosphoms
content
of
vertically-oriented
sub-
strata (made of the same material as the enclosure
wall)
on
t h e
dates in
198'7
(25 June, 15 July, and 12 August). See Mazumder
et
al. (1989) for details on
periphyton
phosphoms.
Phosphoms
Sedimentation
Sedimentation of material from the epilimnion is usually
measured using sediment traps. These may accurately measure
sedimentation in certain
lakes,
especially
in deep
stratified
lakes.
Chalton
(1975) demonstrated that sedimentation rates of
phosphoms measured in
luge
enclosures
are
consistent with
those measured in
open
lake
waer.
Enclosures used in this study
are deep and show
strong
thermal
stratific2tim
during
late
spring and summer
periods.
Therefore, cylinder-type sediment
traps
(Bloesch
and Bums 1980) should provide a reliable esti-
mate of the
actual
loss of material from the epilirnnion
to
the
hypolimnion.
Four (two pairs) cylindrical sediment traps
(6.5
cm diameter,
40 cm
long,
aspect ratio of
116)
were set below the
themmline
(8
m)
in each
encl ssm
in
1986
(June 3-August
19).
Six traps
(one
central
pair floating at 8
m
fmm
m
anchor, one central
pair hanging from the top and one pair hanging
B
m away from
the
enclosure
wdl)
per enclosure were used in I987 (June
25-
August
10)-
Each
pair of traps was attached to the ends
of
a
1
-
m steel bar suspended with
a
nylon
rope
tied to a
float
at the
surface. The pairs of
tmps
suspended from
m
anchor, rather
than from the surface, and the pair of traps hanging from the
wall, were used to ensure that sedimentation rates were not
higher
below
the
ropes or
meax=
the walls.
Sediment trap samples were collected every 2-3
wk
in both
years. At the time of
collection,
the traps were pulled up very
slowly, and the water in the traps was decanted to leave 1
L,
stirred
thoroughly, and two
35-mL
subsamples were collected
in
clean
screw-capped test tubes.
Smples
(1W
d)
were
also
collected (1987 only) and
preserved
with
Lugol's
iodine solu-
tion for microscopic
observation.
Samples collected during the
luly-August
per i d
of 1987 were used for
microph~tography;
5-raaE
subsamples
were settled in
Utermdhl
chambers and pho-
tographed at
200
x
using
an
inverted phase microscope.
Con-
centrations
of
phosphoms in
the
subsamples
(two
per trap) was
measured following the
method
described above. Measured TP
concentrations in the sediment trap samples were
corrected
for
the ambient
TP
at 8
ma.
Determination
of
Total Epilimnetic
Calcium
Concentrations of total calcium (milligrams per
litre)
were
determined
from
0-4
integrated water samples. Collected sam-
ples were analyzed
d
Water Quality National Laboratory, Envi-
ronment Canada, Burlington,
Ontario.
Statistical Analysis
Analyses of
variance
(ANOVA)
were
p d o me d
for
TP
(mil-
ligrams phosphoms per
square
metre) and sedimentation rates
(milligrams
phosphoms
per
square
metre per day
and
percent
TP
per
day)
for all sampling dates. Repeated measures
ANOW
(Model
III
ANOVA)
were
dso
performed
for all dates
taken
together for each year.
Prior
to calculating
confidence
intervals
or
standard
errors,
variance ratio tests
(Zar
1984) were done
to
examine whether the
variances
were
different
between replicate
enclosures or
overall-
Simple regressions were
p d o me d
to
test the relationships between sedimentation rates and
TP,
and
particulate
phosphoms
(PI?)
>20
pm,
md
multiple regressions
were
p e ~ o me d
to
predict sedimentation rates from the
size-
distribution of
PP
and
hypolimetic
TP.
Epilimnetic
TP
In 1986, the seasonal mean
TPs
were 74.4, 88.6, 114.2,
a d
88.5
mg
JP*m-"n
the control,
+F9
+N,
and
+NF
Can.
J.
Fish.
Aqwab.
Sci.,
hi.
46,
8989
Can. J. Fish. Aquat. Sci. Downloaded from www.nrcresearchpress.com by UNIV VICTORIA on 07/13/11

For personal use only.

FIG.
1.
Epilimnetic
TB
in four treatments.
Bars
represent
95%
c~
of
h e
pol ed
mean of four (two
per
replicate
enclosure, degree of free-
dom- 2)
measurements.
TP
(mg
P~r n- ~)
=TP
(pg
P.L-9
x
4
rn.
Numbers
associated
with
sapl i ng
ddes
represent the
number
of days
lapsed between fertilization of enclosures and
sampling
for
TP.
enclosures, respectively.
TP
was higher (Fig.
1)
with
fertilization than without only on July
4,
JuBy
21, and August
18
(ANOVA,
p<O.05).
On
other dates
(May
21,
June
3, 21,
August
61,
dthoaagh
TP was highest in the
+N
enclosures,
significant effects of fertilization were not obtained because
h e
TP
in
+
NF
enclosures
was lower
than
in the control
and/or
+
F
enclosures. The effect
of
fish
ow
TP was reversed by
fertilization. As a result, the effect of fertilization
md
fish on
TP
over the whole season had significant interaction
(repeated
measures
ANOVA,
p<OnQ4;
Fig.
I).
There
was
a decline in TP
[TP
decline
=
((TP,
-
TP,
+
(TP,
-
TP,)
+
. .
.)
+
t;
where
TP,
to
TPn
-
epilimetic
TP
on first to nth sampling dates, and
t
=
total
number
of days
from
first
to
nth sampling date, during the spring
to
summer
experimental
period in
dl
treatments. Although the
timing of
this
decline
varied
among treatments,
md
between
yeam
for
h e
control enclosures, the
TP
declines
during
the
experiments were higher in the
fishless
enclosures;
0.65,0.48,
1.15,
mdO.3'7
mg
~.m- ~.d -'
for
control,
+F,
+N,
and
+NF
enclosures, respectively.
In
f
987,
egilimnetic
TP
trends
among treatments were
remakably
similar to those in 1986. However,
seasond
mean
TB
was significantly higher with fertilization (repeated
measures
ANOW;
p
=
0.004).
Seasonal
TBs
were 69.4,
9
1
-4,
111.4,
md98.2mgP.mx2inthecontrol,
+F,
+N,
md
+NF
enclosures, respectively.
On
one date (August
18)
TP
was higher
with
fertilization
(p=O.Wl)
in enclosures both with and
without fish.
On
the other dates,
+
F
enclosures
had
TB
greater
than at least
one
of
the
two fertilization treatments. As a result,
fertilization and fish again had significant interaction on
thee
dates
(8.019<P<0.03,
Fig.
1).
The rates of
TP
decline in the
different treatments were
similar
to those in 1986, but slightly
higher in the unfertilized
enclosures;
0.77,8.74,
1.13,
md
0.36
rng
P.n~- ~.d-
'
for
control,
+
F,
+
N,
and
+
NF
enclosures,
respectively.
The
fishless
enclosures
(control
md
+
N) again
had
higher
TP
declines.
FIG.
2. Sedimentation rate of
phosphoms
in
four
treatments
from
June
to August,
1986
and
1987.
Emor
bas
represent
95%
cca
of the pooled
means
(df=
6
for
1986,
and
18
for 1987).
*
=
95%
cr
was calculated
from the means
of
replicate
enclssures
with one degree of freedom,
because the
varimces
were different between
replicate
enclosures
within
treatment.
FIG.
3.
Sedimentation
rate in four treatments from June
to
August,
1986
and
1987,
expressed as
a
percent
of
h e
TP
at the beginning of
the measurement
period.
Enor
bas
represent
95%
er
of the
pol ed
mews.
*
is
same
as in
Fig.
2.
Sedimentation
of
Phospkoms
The
rates of TP decline
(PD)
were
strongly
related
with phos-
phorus sedimentation rates
(SR)
(SR
=
%
.66
+
3.67 PD,
?
=
0.67,
n
=
16). However, when unfertilized
and
fertilized
enclosures were analyzed
separately,
the relationship for fertil-
ized enclosures was
SR
=
2.36
+
3
- 3
1
PD
(9
=
0.87),
while a
much
weaker
relationship
(SR
=
0.44
+
4.83
PD,
1-2
=
0.47)
with a smaller y-intercept and a
lager
regression
cmfficient
was found for unfertilized enclosures.
Sedimentation
rate (milligrams
phosphorus
per
square metre
per
day), as measured by the traps, was
usuaIly
higher with fish
Can.
J
Fish.
Aquar.
Sci..
Val.
46,
1989
Can. J. Fish. Aquat. Sci. Downloaded from www.nrcresearchpress.com by UNIV VICTORIA on 07/13/11

For personal use only.

"
0
3
20
40
60 80
100
120
140
160
W
m
TOTAL
PHOSPHORUS
(mg
PenfZ
)
Fna.
4.
Relationship
between
sedimentation rate
(SR)
and
epilimnetic
TP.
Data include seven collection
periods
(1986
and 1987) for
four
treatments.
!==4
?
SR
=
1-23
c
O.lt4PP
>
20
pm,
r
=
0.66,
n
=
112
f-a
FIG.
5. Relationship between
SW
and
PP>20
pm.
Data
include
seven
dates (1986
and
1987)
for
four
treatments.
in the unfertilized
(
+
F)
enclosures, and lower with fish in
the
fertilized
(
+
NH)
enclosures (Fig. 2). As a result, significant
interactions
for the effects of fertilization and fish were obtained
on most dates
(ANOVA,
0.001
<p<0.034),
On average, sedi-
mentation rates were 4.56, 3.86, 7.83, and 3.35 in 1886
and
2.17,2.70,5.72,
and
3.39
rng
~ m - ~ a d -'
in
1989
for
control,
+
H,
+
N,
and
+
NF
enclosures, respectively.
The interaction of fish
md
fertilization was removed
by
expressing sedimentation as a fraction of the
epilimetic
TP
(percent
TP
per
day) (Pig.
3).
The
effect of fertilization on
fractional sedimentation rates
(1986
and 1887) was not signif-
icant
(repeated measures
ANOVA,
p
=
8.16
11,
but the effect
of
fish was significant
(P
=
0.026). The fraction of TP sedimented
was significantly lower
(0
-00
1
<P<0.05)
in the enclosures
with fish
(+
F
and
+
NF) on all dates except July 21,
1986
in
the
+
F enclosures. On this date, the two
replicate
+
F
enclo-
sures had
significantly
different means
and
variances,
md
for
that reason
confidence
intervals
were
cdculated
using the
means
of
the
two replicate enclosures with one
degree
of
freedom
(Figs. 2 and
3).
This
may
have produced significant interaction
for
the effects of fertilization and addition of fish on that date
@I
=
0.087).
Sedimentation rates were positively
reIated
with
epilimnetic
TP
(Fig.
4).
However, this relationship explained only 23%
@<0.61)
of
the
variation
in
sedimentation
rate.
Sedimentatam
rate was more strongly related with the
)20
pm
fraction of
PP
(Fig. 5). There
was
no
significmt
relationship between sedi-
mentation rate and
PP
<20
pm.
Multiple
regression was used
in an attempt to relate
SR
to
the amount of phosphoms in var-
ious size-fractions. A regression coefficient of near zero was
found for
PP
8.2-20
pm,
and the coefficient for
PP
2&2W
pm
was
higher
than that for
PP>280
pm.
The model
(I)
SR
=
1.167
+
0.001(BP
0.2
-
28)
did not Improve the prediction of sedimentation rate (multiple
?
=
0.45,
n
=
I 12). The residuals between the measured and
predicted
SR
were only 5 to
7%
for the unfertilized
encl6_asures,
but were I0 to 25% for fertilized
enclosures.
Examination of
the residuals indicates that the
relationship
differs within treat-
ments. The
strength
of the
correlation
is primarily among, rather
than within, treatments.
The high regression coefficient for the relationship between
SR
md
TP decline, indicating that
SW
is more
than
adequate
to account for the decline in
epilirnnetic
TP, and the low coef-
ficient of
detemination
of the relationship between
SR
and
TP
fractions, suggests there is a
source
or component of sedimen-
tation not
included
in this analysis. To investigate whether
TP
between the
epilimircen,
as defined by sampling regimen, and
the sediment trap at
8
m
was the missing source,
we
repeated
the multiple regression equation (1) between
SR
and
TP
frac-
tions adding TP at 8
rn
as
rn
additional
independent variable.
The multiple regression equation
+
0.482(PP20
-
200)
+
0,176
(BP>
200)
+
0.024
(TP
8m),
had a significantly higher
?
(0.75). This multiple regression
model
was used to
correct
sedimentation rate for the
contri-
bution sf
metalimnetic
TP
by setting TP at
8
m
to
0
and pre-
dicting
SR.
To verify our
rate
estimates, we attempted a mass
bdmce
of
phosphoms in
our
enclosures for
1987,
when we estimated the
mount
of phosphoms
lost
as
periphyton
growth and fish
growth. Wet deposition of
phosphoms
in the vicinity of
Lake
St. George is also known
(0.05
mg
P ~ m- ~ d -';
Tang et al.
1986). The net imbalance in the
phosphoms
budget is expressed
as internal loading (Table 1). Calculated
internal
loading was
higher
in
the fertilized enclosures, enclosures
with
fish had
higher
internal
loading than in the enclosures without fish.
Sedimenting
Materials
Photographs of sediment trap material from different treat-
ments illustrate the qualitative effects of the treatments on the
particles
sedimentiwg
from
the
epilimnion
(Fig. 6).
Sediment-
ing
particles in
fishless
enc~osures
were
larger
than those in the
enclosures with fish. Aggregated materials, possibly
faxes,
were
c omon.
Identifiable algae from the control trap were
mostly
Aaekdstrodesmess
fa&catusas,
the diatom
Achnanthes,
and
Di nob~on
loricae.
In
the
+N
trap, dominant genera were
Gieocystds
md
Ooeystie's.
Mmy
of the cells had
prominent
sheaths, and
appeared
to be resistant stages. In the
+
F
enclo-
Can.
J.
Fish.
Aquaf.
Sci.,
V01.
46,
1989
Can. J. Fish. Aquat. Sci. Downloaded from www.nrcresearchpress.com by UNIV VICTORIA on 07/13/11

For personal use only.

FIG.
6.
Photographs of sediment
trap
material
from
July-August
1987
preserved
with
Lugal's
iodine.
Each
row
is
from
one
treatment;
from
top
to
bottom, control,
+F9
+
NNp
md
+NF.
Smallest
unit in
the
scde
is
equal
to
10
pm.
SURS,
most
sf
the
material
in the sediment trap examined was
nmoplmkton.
These were
predsminandy
Scenedesmus
a d
other green algae
in
+
F
enclosure, while
Lagerheiasn'a
md
C~~QOCOCCUS
dominated
the
+
NF
sample. Pa the
+
NF
ewclo-
swes,
sedimewtiing
dgal
cells were generally
lager
than
those
in
the
+F
enc%oswes.
Most
of the cells
from
both
+F
and
+
NF traps
appeared
to
have
intact chloroplasts.
Aggregations
of cells, possibly
faecal
materid,
were more
cor non
in the
+
NF
treatment.
'There
were
treatment differences in the amount of
epilim-
wetic
ealciurn
in
both
yeas
(Fig-
7).
Calcium was less
in
the
presence
s f
fish
and
with
fertilization.
White
sediment
that
we
believe to be calcite was observed in
+
F,
+
N,
and
+
NF
traps
in summer.
Discussion
Fish
reduced
the
fraction
sf
epilimnetic
TP
sedirnenting
per
day,
md
the
seasonal
decline in
TP,
both
with and
without
fer-
tilization.
However,
there
was
strong interaction between
our
two treatments; fertilization was
more
effective in increasing
epilimetic
TP
in
the
absence
of
fish,
but
addition
of
fish was
associated
with
an
increase in
'IF
in
unfertilized
enclosures.
Fertilization
and
addition of
fish
also
had
intractive
effects
om
Can.
J.
Fish.
Aquaaaf.
Sci.,
Vd.
46,
1989
1738
Can. J. Fish. Aquat. Sci. Downloaded from www.nrcresearchpress.com by UNIV VICTORIA on 07/13/11

For personal use only.

FIG.
7.
Temporal changes in the
total
epilirnnetic
(0-4
m)
calcium
concentrations
in
four
treatments
on
four dates
in
1986.
Each point is
a mean
of
two
deteminations
(one per enclosure).
sedimentation rate. However, these
interactions
were removed
when sedimentation rates were expressed as fraction of
TP
sed-
imenting
per unit
time.
Our results
from
unfertilized
enclosures
support Wright
md
Shapiro's
(1984) observation that
TP
can be reduced
through
a
shift from small to large grazers,
alhough
the opposite result
was obtained in fertilized enclosures,
In
enclosures without
plmktivorous
fish,
body
lengths of
eladscerans,
especially
Dqhni a,
were greater (A. Mazumder, unpubl. data), and about
54-57%
of the
total
PP
was present as
microplanktonic
a d
mesopla&tonic
particles
(20-208
prn
and
>2Kl
pm,
Mazum-
der
et
al.
1988).
In
contrast, in
the
fish enclosures
(
+
F
and
+NF),
only
3637%
of the total epilimnetie particulate phos-
phorus was
comprised
of these
lager
particles, and most
of
the
particulate phosphorus was
pieo-
and
nanoplankton
(8.2
to
20
pm).
We
observed higher seasonal epilimnetic
TP
declines in
the
fishless
enclosures with larger
zoopHankton.
Our
results suggest that
fractional
sedimentation of
TB
can
be affected by food web changes. Fractional phosphoms sedi-
mentation was reduced by
planktivoroaas
fish, but
s w
measured
rates
are
within the range reported from
lakes
(White and
Wetzef
1975;
Fallon
and
Brock
1986%;
Blsesch
and Uehlinger 1986;
Scavia and
Fahnenstiel
1987)
and
enclosures
(Chxlton
19'75;
Uehlinger and Bloesch
1
987).
For example, our
corrected
mean
sedimentation rates were
1
.O
B
and 3.23
%
TH"P.8-
"in
the Control
and
+ N
enclosures, while
Charlton
(1975)
reported
2.1%
TP-d-"5..9
mg
Pm-2.d-')
and
2.8%
TPd-'
(13.6
rng
P-m-'sd-
')
in control and fertilized enclosures, respectively.
Higher phosphoms sedimentation rates
(4--5%=d-
')
were esti-
mated from a budget
model
by
Staraffer
(1985)
for two calcar-
eous
lakes
in Wisconsin.
In
Toussaint
Lake and Lindsey Pond,
the daily
losses
of phosphoms
were
1
md
2%-d-'
(Rigler
1973). We suggested these rates
may
be similar among
lakes.
Vollenweider
(1969a,b)
assumed a loss of phosphoms propor-
tional
to
epilimnetie
TP
in his phosphorus
budget
model for
lakes.
Although
Wright
and
Shapiro
(1984) speculated that
enhanced vertical
transport
by herbivores was responsible for
TP decline,
Nilssen
(14898)
and Taylor (1984) suggested that
large
herbivorous zooplankton
enhance
sedimentation, while
Biirgi
et
al.
(1
979) and Uehlinger and Bloesch
(1
98'7)
demon-
strated that higher
sedimentation
rates were associated with
abundant large herbivores. In
Like
Michigan, changes in the
abundance of
luge
gazers
(Dwphni~),
associated
with changes
in
planktivore
predation, may
have
caused
changes
in
epilim-
TABLE
1. Sources and losses
for
phosphoms
(seasonal
means)
for
four
treatments
in
1987.
All values except
TP
at
8
rn
are
in
rng
B*m-'*d-'.
TP loading
=
wet
depositon
+
weekly additions.
Expected
net loss
from
the
epilimnion
=
loading
-
(loss
of
pekiphytsn
+
Ash)
+
TB
decline.
Internal
loading
=
corrected
sedimentation rate
-
expected net
loss
from
the
epilimnion.
Control
+ F
+N
+NF
TP loading
Lass
to
fish
Loss
to
peri-aphytsn
Observed
TP
decline
Expected
net loss
from
the
epilirnnicpn
Measured
SR
TP
at
8
m
(pg
PsL
Corrected
SW
for
TB
at
8
m
Internal
loading
or net imbalance
neeie
TP
and sedimentation rate (Scavia et
al.
1986;
Seavia
and
Fahnenstiel
1987).
Our sedimentation rates overestimated phosphorus loss
fmrn
the
epilirnnion
because
biological
ineovoration
of dissolved
phssphoms
into
particulate
phosphorus in the
4-43
m zone con-
tributed to sedimentation at
8
m. Both sedimentation rates and
metaliwanetie
TP
were highest in the unfertilized enclosures with
fish
(
+
F)
and
in fertilized
enclosures
without fish
(
9
N).
In
the
+
N
enclosures, high
TP
at
8
rn
may have been
due
to
development
sf
phytoplankton as the water was
oxic
and light
intensity was high.
In
the
+
F
enclosures, where this depth was
anoxic
and light intensity was
%ow,
there was a dense population
of photosynthetic bacteria (A. Mazumder, unpubl.
dzta).
Con-
centration of
TB
at this depth was higher than the
epiliimnetic
TB
among all treatments, the highest
csncentration
was
obsewed
in the
+
F
enclosures (Table 1).
Changes in epilimnetic
TB
seem
largely
detemined
by the
rate of phosphoms loss
from
the epilimnion via sedimentation;
64%
of the variation
in
rate of
TP
decline is explained by the
sedimentation rates in the unfertilized
and
fertilized enclosures.
Although
fertilized enclosures received weekly nutrient addi-
tions, they showed maximum spring to summer declines of
epi-
Iimnetic
TP,
and
a strong correlation between
TP
decline
md
sedimentation rate.
External
loading was only moderately
effective in increasing epilirnnetic
TB
despite
our
adding
enough
PO4
(332
mg
PmP2
in
90
d)
to triple the total concentrations.
Earlier (Mazumder
et
al.
1988)
we demonstrated
opportunistic
uptake
sf the added phosphate
by
Ewge
gkytopla&ton,
which
probably have high
sinking
rates (see equation
(I);
Uehlinger
and
Blmsch
198'4)
and
are
abundant
in
trap material (Fig.
6).
This
could be the mechanism by which the added phosphate
is
removed from the epilimnion.
The mechanism by which fish affect sedimentation may
involve
the size-distribution
of
particles,
which
was strongly
affected
by the treatments (Mazumder
et
al.
1988). This is sup-
ported
by
a strong relationship between
PP>20
pm
and
SR
(Fig.
51,
md
the lack
of
a
relationship
with
PP<20
prn
(slope
=
6.081,
equation (1)).
Similarly,
Uehlinger and Bloesch
(1
987)
reported specific sedimentation rate of
only
0.
007sd
-
'
for
PP
<
%
2
pm,
and
mentioned
thzt
although
suspended
seston
consisted
mainly
of
small
particles,
matefial
in sediment
traps
was dominated by large
particles.
There are
various
reports
regarding the types of
material
that
eontribute
more
to
sdi -
Cm.
J~
Fi.fh.
Aqua.
Scd.,
Vo&.
46,
1989
Can. J. Fish. Aquat. Sci. Downloaded from www.nrcresearchpress.com by UNIV VICTORIA on 07/13/11

For personal use only.

mentation.
Some
have reported that most of the
sedimented
material
is faecal pellets produced by
calmoid
coppods
(Fa-
range
and
R&
19%8),
while others have reported
that
most sf
the sedimenting
materid
is of algal origin
(Ulen
1978).
Sedi-
mentation rates
reported
here were not
correlated
with biomass
of
copepods
(cyclopoid,
calmoid,
or
bt h)
or other
zooplank-
tow.
Scavia
and
Fhnenstiel
(1
987)
found that in
Lake
Michi-
gan,
direct
sinking
(nonpredatory
loss) of algae,
especially
dia-
toms, was the important component sf sedimenting
material
in
spring, while
algae
egestd
by
zoopl dt on
were
important
in
summer.
The
sedimentation rates estimated
during
our
expek-
ments
were higher in
the
spring when grazing activity of
zoo-
plankton was low.
The observation that
particle
size-distribution only
partly
explains sedimentation rates probably reflects qualitative as well
as quantitative differences
in
the
PP
size-fractions. Our frac-
tional
sedimentation rates can be translated into settling
velsc-
ities
by multiplying them by 8
m,
the depth of the trap. Mean
settling velocities in the treatments
rmged
between
8.
I5
to
0.18
mad-',
and between 0.22 to 0.26
mad-
I,
in the fish and
fishless
enclssures,
respectively. These
are
similar
to
velocities
deter-
mined for
phytopldton
in
Lake
St. George by a different
method
(Bums
md
Rosa
1980).
The nature of the
particles,
such as
the
shape, ornamentation, density, motility,
etc.
are
also
important.
For
example,
So er
(1984)
found diatoms have
very
high
(3.8
to 16.6
mad-
')
sinking
rates. Green algae and
Aphanizomensn
have lower settling velocities (0.03 to
2.1
rnd-
I),
and
flagellates may have
z m
velocities
(Stabel
1987).
Bums
and
Rosa
(1980) found a
mg e
of
Isw
velocities for
Lake
St. George
phytoplakton,
which
are
only loosely related to
size, and lower
than
for
detritus.
Qualitative
differences in the
sinking
particles
are
obvious
from
Fig. 6, and were
also
sug-
gested for the important
microplmkton
fraction by the inter-
active nature of the treatments on this fraction
(Mazumder
et
d.
1988).
The
net
imbalance
in
h e
phosphoms
budget is expressed as
internal loading. Few studies have attributed the discrepancy
between
epdimnetic
TP
changes
and sedimentation
expricitly
to
internal
loading, but internal loading
cm
be
luge
for lakes
with
anoxic
hyporimia
(Skuffer
1985;
Lehrnan
and
Nau-
mowski
1986).
Our mass
$dance
for
phosphoms
indicates that
phosphoms
sedimentation, even
corrected
for
metalimetic
TB,
is more
than
adequate to explain
TP
decline in the unfertilized
enclosures. If we overestimated
periphyton
growth
by using
substrata that were initially free of
periphyta,
then our cal-
culated
internal
loading
may be even
smaller.
The fertilized
enclosures had slightly too little sedimentation to
account
for
TI?
decline,
after
correcting
sedimentation rates
for
metalim-
metic
contribution. However, the discrepancy is very
srnajl
.
The
relative
importance
of calcium minerals as
a
vehicle for
phosphoms
transport
in our enclosures is
u&nowpn.
However,
we
observed
that
the highest pH
md
lowest concentration in
epilimetic
calcium was associated with presence of fish
(Fig.
'3,
which may be due to the increased
abmdmces
of
pieo-
and
nanoplankton
(Mwumder
et al.
1988).
Murphy
et
al.
(1983)
reported that when calcite precipitation
occumed
during
an
Aphnkzomenon
bloom,
PO,
was completely removed from the
euphotic
zone.
Stabel
(1
987)
reported
that high sedimentation
in
Lake
Constance
was
associated
with calcite precipitation.
Phosphate
concen@ations
in the enclosures were always below
detection limit
(c0.5
pg
P04-P-L-
7,
md
as a
result
it is
not
possible
to estimate the amount of phosphoms that may be pre-
cipitating with
cdcium.
However, if this was a vehicle for ver-
tical
transport
of phosphorus, then its
effect
would be
to
reduce
the treatment effect (reduced percent sedimentation with fish)
we
observed.
We conclude that food web effects resulting from the pres-
ence of
planktivorous
fish can affect the
troghic
status
of
plank-
tonic systems,
md
that the efficacy of
fertilization
is dependent
on the food web.
Plaktivorous
fish can reduce the
loss
of phos-
phorus from the
epilimnion
by favoring the dominance
of
smaller
plankton
which contribute more to total biomass and
les
s to sediment
ation.
Acknowledgements
We
thank
D.
Scavia,
R.
Smith,
G.
Y.
Conan,
6.
Niknberg,
and
two
anonymous
reviewers
for their
cements
on
an
earlier draft of
the
manuscript.
We also
thank
N.
MacNeill
for identifying sediment-trap
algae,
and
the staff
at
L&e
St. George
Field
Station
(P.
Blanchfield,
M.
Sohmnes,
B.
Sameraik,
R.
Bake,
K.
Nymm)
for their assistance.
This
research
was supported by Natural Sciences
md
Engineering
Research
Council
operating
grants to
D.
J.
McQumn
and
W.
64.
Tay-
Isr,
and
by
Environment
Canada.
References
BLOESCH,
B.
AND
BURNS,
No
M.
1980. A
critical
review of sedimentation trap
technique. Schweiz.
Z.
Hydrol.
42:
15-55.
BLOESCH,
5.,
AND
U.
UEHLHNGER
1986. Horizontal sedimentation differences
in
a eutrophic Swiss
lake.
Lirnnsl.
Bcemogr.
3
1
:
1094-1
189.
Biiwor,
V.
H.
R.,
M.
BUHRER,
J.
BLOESCH,
AND
E.
SZABB.
1979. The influence
of
experimentally
varied
zooplankton density on
production
and sedimen-
tation
in
a highly
eaatrophic
lake.
Schweiz.
Z.
Hydml.
41:
384%.
BURNS,
N.
M.,
AND
F.
ROSA. 1980. In
situ
measurements sf settling
velocity
of
organic
carbon
particles and
10
species of phytoplankton.
Limn01.
Ocemog.
25:
$55464.
C H A ~ ~ I V,
M. N. 1975. Sedimentation: measurements in
experimental
enclo-
sures.
Verh.
Int.
Ver.
Li msl.
19:
267-242.
EDMONDSON,
W*
T.
1969. Eutrophication in
North
America,
p.
124-149.
blm
Eutrophication:
causes, consequences,
comectives.
Nat~
Acad.
Sci.
Wash.
B.C.
FALLON,
R.
D.,
AND
T.
D.
BRWK.
1980.
Planktonic
blue
green algae: pro-
duction,
sedimentation,
and
decamposition
in
Lake
Mendota,
Wisconsin.
Li mol.
Bceansgr.
25: 72-88.
PA~ANTE,
5.
G.,
AND
D.
J.
PTAK.
1978.
Hetemtrophic
bacteria
associated
with
the degradation of zooplankton faecal pellets in Lake Michigan. J. Great
Lakes
Res. 4: 221-225.
H ~ E Y,
B.
A.,
W.
J.
CHRISTIE,
C.
K.
MINNS,
E.S.
MILLARD,
J.
M.
COOLEY,
M.
G. JOHNSON,
K.
H,
NIGHOLLS,
G.
W.
ROBINSON,
G.
E.
OWEN,
P.
G.
SLY,
W.
T.
GEILING,
AND
A.
A.
CROWDER.
1986.
Trophic
structure
and
inkraceions
in the Bay of Quinte,
Lake
Ontario,
before
md
after
pi nt -
source
ptaosphoms
control.
In
Minns
et
al. [ed.] 1986.
koj wt
Quinte:
point-source
phosphsms
control
and ecosystem response in the Bay sf
Quinte,
M e
Ontario.
Can.
Spec.
Publ.
Fish.
Aqua.
Sci.
86:
259-270.
LENMAN,
J,
T.,
AND
T.
NAUMOWSKI.
1986.
Net
community
production and
hyplimnetic
nutrient
regeneration in a Michigan
lake.
Lirnnol.
Oceanogr.
31: 788-797.
~ ~ MUMDER,
A.,
D.
J.
MCQUEEN,
W.
B.
TAYLOR,
AND
B.
W.
S.
LEAN. 1988.
Effects of fertilization and
plamktivorsus
fish (yellow
perch)
predztion
on
size-distribution of
particulate
phcssphoms
and assimilated
phosphate:
kxge
enclosure
experiments.
Limnsl.
Oceanogr.
33:
421438.
MAZUMDEW,
A.,
we
D.
TAYLOR,
D.
5.
MCQUEEN,
AND
D.
R.
S.
LEAN.
1989.
Effects sf
nutrients
md
grazers
on
periphyton
phospkoms
in
large
enclo-
sures.
Freshwater
Biol.
(In press)
MENZL,
B.
W.,
AND
N.
CORWIN.
1965. The measurement of
total
P
in seawater
based on
ttae
liberation of organically
bund
fraction by
persulfate
oxi-
dation.
Ei mol.
Oceanogr.
10:
28gk282.
MURPHY,
T.
P.,
K.
Ja
HALL,
AND
I.
~ESAKI.
1983.
Csprecipitation
of
phosphate
with
cdcite
in
a naturally eutrophic
lake.
Limnol.
Oceanagr.
28:
58-49.
NILSSBN,
J.
P.
1948. Eutrophication,
mimute
algae, and
inefficient
grazers.
Mem.
%st.
Itd.
Idmbiol.
36:
121-138.
POST,
J.
R.,
AND
D.
J.
MCQUEEN.
1987. The impact of
plaktivorous
fish
sma
the
structure
of a
plankton
community. Freshwater Bisl. 17:
656467.
Can.
$.
Fish.
Aquat.
Sci.,
Vol.
46,
1989
Can. J. Fish. Aquat. Sci. Downloaded from www.nrcresearchpress.com by UNIV VICTORIA on 07/13/11

For personal use only.

~ G L E R,
E
H.
1973.
A
dynamic
view of the
phosphoms
cycle in
lakes,
p.
538-
572.
En
Edward
er
al.
[ed.]
Environmental
phosphoms handbook. John
Wiley
md
Sons, New
Yo&,
NY.
Seavr~,
B.,
AND
6.
L.
FAHNBNSTIEE.
1987.
Dynamics
sf
Lake
Michigm
phy-
toplankton:
mechmisms
controlling
epiliimnetic
communities.
J.
Great
Me s
Res.
13:
1643-120.
~ A V I A,
D.,
6.
L.
~ A ~ E N S ~ L,
M.
S.
EVANS,
D.
J.
JUDE,
AND
J.
T.
LEHMAN.
'1986.
Inhence
of
sdmonine
predation and weather
on
long-term
water
quality
trends
in Lake Michigan. Can.
1.
Fish.
Arguat.
Sci.
43:
4 3 5 4 3.
S ~ NDLER,
We
Do
1975.
Whole-lake
eutrophication
experiments with
phss-
phoms,
nitrogen,
md
cabon.
Verbe
In%.
Ver.
Limnol.
19: 3221-3231.
1977. Evolution of phosphoms limitation in lakes. Science (Wash.,
Be)
195:
260-262.
SBMMER,
U.
1984. Sedimentation of principal
phytoplankton
species in Lake
Constance.
J.
Plankton
Res.
6:
1-14.
STABEL,
H-H.
1987. Settling velocity and residence time of particles in Lake
Consme.
Schweiz.
Z.
My&ol.
49: 284-293.
SIFAUEER,
R.
E.
1985.
Nutrient
internal
loading
and
trophic
regulation
of
Green
Lake,
Wisconsin.
Limcrl.
Ommogr.
364:
347-363.
S ~ C K L A ~,
J.
D.,
AND
T.
R.
PARSONS. 1972.
A
practical
b n d b k
of sea-
water
mdysis.
Bull. Fish.
Res.
Board
Cm.
167:
310.
TANG,
A.
S.
L.,
Wg
H.
CHAN,
D.
H.
S.
CHUNG,
AND
M.
A.
LBJSLS.
1986.
Precipitation
and air concentration
md
wet and
dry
depsition
fields of
pollutants
in
Ontario,
1983.
Ontario
Ministry of Environment Rep.
ARB-
W8-86-AQM.
TAYLOR,
W.
D.
1984.
Phosphoms
flux
through
epilimetic
zmplmktora
from
Lake
Ontario: relationship with
body
size
md
sigmificmce
to
phytoplank-
ton.
Clan.
J.
Fish.
Aquat.
Sci. 41: 1702-1712.
UBHLMGW,
U.,
AND
I.
BLBEXH.
H
987.
The
influence of
crustacean
zosplmk-
ton on the size distribution of
dgal
biomass and
suspended
and
settling
seston
(biommipulationa
h
Iimmona1s
II).
Int.
Rev.
Gesmten.
Hydro-
biol.
72:
473-486.
ULEN,
B.
1978.
Seston
and
sediment
in
Lake
Nomiken
1.
Seston
composition
and sedimentation. Schweiz.
Z.
Hydrol.
40:
262-286.
VOLLEWEIDER,
R.
A.
1968. Scientific
fundamentals
of
the
eutrophication of
lakes
and
Wowing
waters,
with particular reference to
phssphoms
and
nitrogen as factors
in
eutrophication-
OECD
Tech. Rep.
DAStCSW68.27,
p.
159.
1969a.
M6glickeiBn
und
genzen
elemenmer
modelle
der
stoff-
bilmz
von
seen.
Arch.
HyBPsbisl.
$6:
1-36.
1969b.
Possibilities
and
limits of elementary models concerning the
budget of
substances
in lakes
(translated
from
Gemanan).
Cmda
Sec.
State-
tr-8343,
1969,
p.
1-53.
W m,
W.
S.,
AND
W.
G.
WETEL.
1975. Nitrogen, phosphoms, particulate
and
colloidal
c a h n
content of
sedimenting
materid
of a hard-water
lake.
Verh.
6nt.
V'r.
L i m~ l.
19:
330-339.
WRIGHT,
D.
I.,
AND
J.
SHAPWO.
1984.
Nutrient
reduction by
bbiommipaalattion:
am
unexpected
phenomenon
and
its
possible
cause.
Veh.
Int.
Ver.
Limnsl.
22:
518-524.
ZAR,
I.
H.
1984.
BisstatisBical
analysis.
Predce-Hdl.
New Jersey,
Nl
697 p.
Can.
J.
Fish.
Aquat.
Sei.,
Vol.
46,
I989
Can. J. Fish. Aquat. Sci. Downloaded from www.nrcresearchpress.com by UNIV VICTORIA on 07/13/11

For personal use only.