United States
Environmental Protection
Agency
Environmental Research
Laboratory
Duluth MN 55804
Research and Development
EPA/600/S3-84/100 May 1985
&EPA Project Summary
Biological Availability of
Sediment Phosphorus Inputs to
the Lower Great Lakes
Scott C. Martin, Joseph V. DePinto, and Thomas C. Young
In this study, river water samples
were collected from several major
tributaries to the Lower Great Lakes
during storm runoff events in the
spring and early summer of 1980 and
1981. Suspended sediments from
these samples were subjected to a
chemical fractionation sequence of
NaOH-CDB-HCI, as well as algal bio-
assay analyses of sediment P bioavail-
ability using the Dual Culture Diffu-
sion Apparatus (DCDA) technique of
DePinto. Sediments from several of
the bioassay experiments were
reconcentrated after the bioassays
and resubjected to the chemical frac-
tionation sequence. Several other
forms of P inputs to the Lower Great
Lakes were also analyzed for chemical
composition and/or bioavailability.
This Project Summary was devel-
oped by EPA's Environmental Re-
search Laboratory, Duluth, MN, to an-
nounce key findings of the research
project that is fully documented in a
separate report of the same title (see
Project Report ordering information at
back).
Introduction
Since April, 1972, when the govern-
ments of Canada and the United States
signed the Great Lakes Water Quality
Agreement, extensive efforts have been
made to curb eutrophication in the Lower
Great Lakes. This attention has been
focused primarily on the Lower Lakes
(Erie and Ontario), where the most severe
problems exist. Reductions in the
phosphorus (P) loadings to these lakes are
generally considered to be the most effec-
tive means of controlling eutrophication.
Accordingly, target total P loads were
established based on the reduction of
treatment plant effluent P concentrations
to 1.0 mg/l. These were later revised
through the use of water quality models
to predict the load reductions necessary
to achieve the desired water quality for
the Great Lakes (Task Group III, 1978).
Billions of dollars were committed to im-
proving P removal at wastewater treat-
ment facilities discharging to the Lower
Lakes basin. Although municipal loads
have decreased considerably, the revised
target loads of 11,000 metric tons/year for
Lake Erie and 7,000 metric tons/year for
Lake Ontario cannot be met via the pres-
ent programs. In addition, improvements
in the water quality of the lakes resulting
from these load reductions have, to date,
been less than anticipated. Future courses
of action being considered include further
reductions in municipal treatment plant ef-
fluent P concentrations to 0.5 mg/l and
improvement of land management prac-
tices (e.g. implementation of conservation
tillage programs) in order to reduce dif-
fuse (runoff) sources of P to the lakes
(U.S. Army Corps of Engineers, 1982).
Consideration of these options has
caused the utility of target loads based on
total P to be questioned. It is now
recognized that the relative availability of
P from different sources for supporting
algal growth must be carefully evaluated
in developing a cost-effective manage-
ment scheme for the control of eutrophi-
cation. Nearly 50% of the total P load to
Lake Erie and 40% of the Lake Ontario
load are from diffuse sources. A large
fraction of this P is in a sediment-bound
-------
form and may, therefore, not be readily
available for an algal uptake. Studies of
several Great Lakes tributaries have in-
dicated that generally 40% or less of
suspended sediment total P is potentially
bioavailable. However, the kinetics of
conversion of unavailable allochthonous P
to available forms have not been ade-
quately quantified. It is possible that par-
ticulate material may settle out of the
water column before all of its potentially
available P is released. In addition, the
rate of P release from allochthonous par-
ticulate material in the water column may
differ considerably from the regeneration
rate for autochthonous particulate P.
As P associated with suspended par-
ticulate material enters the receiving water
environment, it may become available for
algal uptake via desorption, dissolution,
and microbial decomposition of organic
matter. The degree of particulate P
availability depends on the physical,
chemical and biological characteristics of
the particles and the receiving water.
Among the most important factors are the
lake's productivity level, soluble inorganic
P concentration in the water column,
temperature, lake morphometry, hydrol-
ogy and mixing dynamics, size, shape,
and density of the particles, and the
chemical forms of P in the particles. A
more thorough understanding of the rela-
tionships between these factors and the
bioavailability of particulate P, in order to
predict P release for a wide variety of
sediment and receiving water conditions,
was the overall goal of this study.
The specific objectives of the ex-
perimental program were:
1. to determine the rate and extent of
sediment P bioavailability for a varie-
ty of tributaries to the Lower Great
Lakes under receiving water condi-
tions using a bioassay procedure;
2. to physically and chemically
characterize the tributary suspended
sediments with respect to factors in-
fluencing phosphorus bioavailability;
3. to develop a methodology that may
be used to estimate the rate and ex-
tent of available P release from
suspended sediments based on the
characteristics of the sediments; and
4. to compare the P bioavailability and
chemical characteristics of tributary
suspended sediments to other poten-
tial sources of sediment P to the
Lower Great Lakes.
Experimental Approach
River water samples were collected
from six Lower Great Lakes tributaries
(Maumee, Sandusky, Cuyahoga, and
Genesee Rivers, and Honey and Cat-
taraugus Creeks) during storm runoff
events. Over 50 tributary suspended sedi-
ment samples were collected during
March-June, 1980 and April-June, 1981.
Suspended sediments from these samples
were concentrated and algal bioassays of
sediment P availability were conducted
using the Dual Culture Diffusion Ap-
paratus (DCDA) method of DePinto
(1982). In addition, a variety of chemical
and physical analyses were performed, in-
cluding sediment particle size distribution
(results presented in the full report) and a
chemical fractionation sequence for sedi-
ment P. Kinetic coefficients (rate and
amount) for the release of sediment P
were determined and correlations between
these values and the physical and chem-
ical parameters were examined. This was
done to develop a means of predicting the
kinetics of sediment P bioavailability
based on a more convenient analytical
procedure than the sediment P bioassays.
Results
Tributary Suspended
Sediments
The P fractionation procedure used in
this study consisted of the sequential
measurement of reactive (R-NaOH-P) and
nonreactive (NR-NaOH-P) sodium hydrox-
ide extractable P, citrate-dithionite-
bicarbonate extractable phosphorus (CDB-
P), HCI extractable P (HCI-P) and the
residual fraction after the above se-
quence. A summary of the P fractionation
results for the tributary suspended sedi-
ment samples is shown in Table 1. Con-
centrations of most of the extractable
fractions were consistent among the
samples for a given tributary. Greater
variability was seen between tributaries
than among samples from any one tribu-
tary. Thus, it appears that temporal var-
iations in the forms of P bound to
suspended sediments in streams are of
minor importance compared to differences
in soil type and land use, at least for
periods of high flow associated with
stormwater runoff.
The chemical fractionation results in
Table 1 can provide insights into the
forms and mobility of P associated with
suspended sediments from the different
tributaries. The sum of R-NaOH-P and
CDB-P fractions may be considered to
represent the concentration of non-apatite
inorganic P (NAIP) in the suspended sedi-
ments, while HCI-P provides an estimate
of apatite P. Thus, the Ohio tributaries
contained high levels of NAIP and small
amounts of apatite P, while Cattaraugus
Creek sediments were rich in apatite P
and low in NAIP. Samples from the
Genesee and Detroit Rivers contained in-
termediate levels of both NAIP and
apatite P.
The release of available P from river
sediments in the DCDA reactors was
measured by monitoring P uptake by
Se/enastrum capricornutum in the assay
vessel. An example of a plot of cumula-
tive P release versus time obtained from
bioassays on two samples (#19 and #20) is
shown in Figure 1. The form of these
release curves is typical of the results
observed for the majority of samples
analyzed. In Figure 1, the error bars about
each data point represent one standard
deviation for triplicate bioassays, while the
solid lines illustrate the amount of sedi-
ment P release predicted by the first-order
equation:
Prel(t) = PultH - exp(-Krt)] (1)
where Prei(t) = the amount of sediment P
released at time t (jigP/g sed), Puit = the
total amount of sediment P ultimately
available for algal uptake (/tgP/g sed), and
kr = the first-order P release rate coeffi-
cient (day1, base e).
A summary of the tributary mean
values calculated for the first-order release
coefficients, Puit and k, are presented in
Table 2. The order of mean bioavailable
sediment P concentrations for the trib-
utaries was Cuyahoga > Maumee >
Honey > Sandusky > Genesee > Cat-
taraugus. These results confirm that the
ultimately available fraction of tributary
suspended sediments is much less than
that of autochthonously produced par-
ticulate P (e.g. phytoplankton), while its
release rate is much faster.
The implications of not distinguishing
between external ultimately available P
and that produced within the lake in phy-
toplankton models have been reported
elsewhere. When sediments of tributary
origin remain suspended in the water col-
umn of a lake for longer than about 7
days, the current modeling approach
would tend to overestimate the input of
soluble P to the lake. For models
calibrated primarily on the basis of algal
biomass, this would likely cause an over-
prediction of soluble P concentrations in
the water column. In turn, this might lead
to either an overestimation of the P reten-
tion time, or an unrealistic adjustment of
other model coefficients in order to effect
a calibration. It might also lead to inac-
curate conclusions about the relative con-,
-------
tributions of the various P sources to
biological productivity in the lake.
In order to investigate the possibility of
using a chemical method as a surrogate
measure of bioavailable P, simple product-
moment correlation coefficients (r) were
determined between all chemically extract-
able sediment P fractions (Table 1) and
the one-component P release coefficients,
Puit and kr (Table 2).
All of the chemical fractions except
Residual P were significantly correlated (a
< 0.05) with PU|t. The highest correlation
coefficient obtained was between Puit and
non-apatite inorganic P (NAIP =
R-NaOH-P + CDB-P); however, the in-
organic component of base-extractable
sediment P (R-NaOH-P) is more closely
linked to bioavailability than any other
single chemical fraction measured in this
study. Correlation coefficients and regres-
sion equations are presented in the full
report.
Observation of the changes in sediment
P fractionation during DCDA bioassays
confirmed that the major contribution of
the various chemically extractable forms
of sediment P to the release during
bioassays came from R-NaOH-P. This was
accomplished by reconcentrating the
sediments after a DCDA run and resub-
jecting them to the same chemical frac-
tionation sequence used for initial
characterization. Seventeen of the 52
tributary samples were analyzed in this
manner. The most noticeable feature of
the data from this analysis is that
decreases in R-NaOH-P during the
bioassays consistently exceeded the
changes in all other chemical fractions. A
large and relatively constant percentage of
the R-NaOH-P initially present in the
DCDAs was released for all of the
samples analyzed. This percentage ranged
from 66.0 to 80.1%, with a mean of
70.8% and a coefficient of variation of
only 6.5%. In addition, AR-NaOH-P was
very closely correlated (r = 0.896) with
the amount of sediment P taken up by
the assay culture in the bioassays.
Shoreline Bluff and
Bottom Sediment Samples
The erosion of shoreline (bluff)
sediments and the resuspension of lake
bottom sediments during high winds both
may result in substantial inputs of par-
ticulate P into the water column of Lake
Erie. The biological availability of sediment
P in samples from these sources was in-
vestigated using chemical fractionation
and bioassay analyses. The results are
presented in Table 3. These data are
especially interesting since they encom-
pass much wider extremes of sediment P
fractionation and bioavailability than the
tributary data.
The eroding bluff material contained ex-
tremely low levels of NaOH- and CDB-
extractable P. In addition, the T-Sed-P
concentrations of the Port Stanlfy and
Rondeau Park samples were well below
the lowest value measured for tributary
sediments. A large percentage (mean =
78.3%) of T-Sed-P in the bluff samples
was removed in the HCI extraction step
and, therefore, was most likely present in
the form of apatite P. By comparison,
although concentrations (In ftgP/g sedi-
ment) were similar for Cattaraugus Creek
sediments, HCI P made up only 47.9% of
T-Sed-P. In DCDA bioassays on t!iese
samples, none of the sediment P became
available to algae during the course of the
experiments. This is consistent with the
assertion made by other researchers that
apatite P is virtually unavailable tc aquatic
organisms and that recessional shoreline
sediments contain very low levels of
bioavailable sediment P.
The Monroe bottom sediment samples
represent the opposite extreme from the
bluff sediments. All concentrations of
T-Sed-P, T-NaOH-P, and CDB-P in the
Monroe #2 sample were more than twice
as large as the highest values observed
for tributary suspended sediments. Bio-
assays showed that PU|t for both Monroe
samples was also over twice the largest
value measured for tributary sediments.
The percentage of T-Sed-P available to
Table 1. Summary of P Fractionation Results for Tributary Suspended Sediments
Tributary T-SED-P
(TP-TSP)
SS
T-NaOH-P
R-NaOH-P
Concentrations in ngP/g sediment
NR-NaOH-r CDB-P
HCI-P
Residual P
Maumee River (n=9l:
Mean 1,259
C.V. (%) 14.3
Sandusky River (n =25):
Mean J, 188
C.V. (%) 11.7
Cuyahoga River (n = 4):
Mean 1,315
C.V. 1%) 6.2
Cattaraugus Creek (n = 5):
Mean 637
C.V. (%) 11.0
Honey Creek (n = 41:
Mean 1,189
C.V. (%) 16.5
Genesee River (n = 1):
957
1,244
18.9
1,193
13.6
1,246
9.9
588
12.0
1,218
14.5
994
Detroit River In = 1)t:
1,424
403 (32.01*
32.4
449 (37.7}
20.8
575 (43.8)
13.4
76(11.9)
33.2
505(42.5)
19.1
240 125.1)
596 (41.9)
273 (21.71
20.9
298 (25.1)
20.8
442 (33.6)
21.4
46 (7.21
27.2
356(30.0)
23.5
170(17.8)
305 (21.41
131 (10.41
61.9
151 (12.7)
31.1
133 (10.1)
29.6
30(4.7)
46.5
238 (18.9)
17.5
237(20.0)
21.0
280 (21.3)
16.6
82 (12.9)
22.7
149 (12.5) 203 (17.1)
11.3 22.2
69 (7.2)
291 (20.4)
187(19.5)
177(12.4)
117(9.3)
32.5
77 (6.5)
35.6
1881/4.3!
6.3
305(47.9)
12.9
56 (4.7)
50.7
273 (28.5)
323 (22.7)
174 (13.8)
42.2
194(16.31
42.5
73 (5.6)
63.9
48(7.5)
8.7
265 (22.3)
21.7
77(8.0)
170(11.9!
^Values in parentheses are means expressed as a percentage of T-Sed-P.
^Sediments composited from four sampling locations.
-------
400 f
300 —
200 —
CL
.5
100
A Sample No. 19 (Maumee R.)
, = 0.171 day'\- PuH = 294 ug p/g
Sample No. 29 (Sandusky R.)
Figure 1.
Time (days)
Cumulative re/ease of sediment P versus time measured in DCDA biassays on
Samples #75 (Maumee R.) and #20 (Sandusky R.).
algae (over 50%) was greater as well.
Bottom sediments from the Toledo sites,
on the other hand, exhibited chemical
fractionation and bioavailability
characteristics well within the ranges
measured for tributary suspended
sediments.
Sandusky River Bottom
Sediment Samples
The effect of a point source of P on the
sediment P chemistry was also briefly ex-
amined. Our chemical fractionation se-
quence was applied to bottom sediment
samples from three sites each upstream
and downstream from the Bucyrus, Ohio
wastewater treatment plant discharge (on
the Sandusky River). The results are given
in Table 4. Substantial increases in the
R-NaOH-P and CDB-P fractions occurred
downstream of the point source. The NR-
NaOH-P, HCI-P, and Residual P fractions
increased by smaller amounts. This move-
ment of P into the sediments is consistent
with the finding that soluble reactive P
concentrations decreased rapidly with
distance downstream from the Bucyrus
treatment plant discharge. The data in
Table 4 indicate that, although much of
the point source P may be in relatively
unavailable forms such as CDB-P, HCI-P,
and Residual P, a sizable amount is also
present as R-NaOH-P, and is therefore
potentially bioavailable should these
sediments become resuspended and car-
theless, these data suggest that the P
discharged from an indirect point source
undergoes a reduction in bioavailability in
transit to the receiving water body relative
to that discharged from a direct point
source.
Conclusions
An analytical program combining the
accuracy of bioassay measurements with
the efficiency of chemical fractionation
studies has been shown to be an effective
means of quantifying sediment P bioavail-
ability. The information gathered in this
study can, therefore, serve as a pre-
liminary basis for evaluating the impact of
P management strategies on bioavailable
P loads to the lower Great Lakes. It also
provided the necessary process ex-
perimental information for implementing
changes in phytoplankton model struc-
tures to permit differentiation of
allochthonous and autochthonous sources
of bioavailable P.
The results presented in the full report
indicate that bioavailability remains
relatively constant for a given set of soil
type (i.e. geochemistry and texture) and
land use conditions. Evidence is seen in
the similarity between the bioassay and
fractionation results for the Maumee and
Sandusky Rivers. Also, Cattaraugus Creek
sediments, which were much higher in
apatite content, showed a distinctly lower
level of P availability. The Cuyahoga
River, which experiences a greater an-
thropogenic influence, displayed a higher
level of sediment P bioavailability. Little
fluctuation in sediment characteristics was
observed over time for any given tribu-
tary. Based on these findings, it is felt
that future monitoring programs should
concentrate on identifying differences be-
tween regions of distinct soil type and
land use. This can be accomplished most
readily by using routine chemical extrac-
tion analyses on sediments transported
during storm runoff events supplemented
by less frequent bioassay measurements
of sediment P bioavailability.
ried into the receiving water body. Never-
Table 2. Mean First-Order Release Coefficients Calculated from Bioassay Data
Tributary
Maumee River
Sandusky River
Cuyahoga River
Genesee River
Cattaraugus Creek
Honey Creek
T-Sed-P*
in DCDAs
1308
1145
1314
900
559
1198
Ultimately
Bioavailable
Sediment P
337.3
247.1
449.2
173.8
38.8
298.0
Puit as
a % of
T-Sed-P
25.0
21.4
33.9
19.3
7.7
24.9
Release
Rate, k,
(day"1!
0.182
0.177
0.188
0.264
0.131
0.214
*The T-Sed-P values listed in this table are based on TP and SS analyses performed on the sediment
suspensions placed in the DCDAs.
-------
Table 3. Chemical Fractionation and Bioassay Results for Shoreline Erosion (Bluff) and Lake Bottom Sediment Samples
A) Chemical Fractionations (Values in =gP/g sed.):
Sample (Type)
T-Sed-P
T-NaOH-P
R-NaOH-P
NR-NaOH-P
CDB-P
HCI-P
Residual P
Monroe #1 (bottom)
Monroe #2 (bottom)
Toledo #1 (bottom)
Toledo #2 (bottom)
Port Stanley (bluff)
Rondeau Park (bluff)
L. Superior (bluff)
2266
355J
926
935
279
396
771
1321
1863
208
226
13
8
15
1226
1796
156
187
8
6
13
95
67
53
40
5
2
2
475
612
156
167
94
22
32
258
315
248
227
228
326
543
44
43
105
99
25
28
88
B) DCDA Bioassays (One-Component Release Coefficients Calculated by Thomas Method):
Sample
Monroe t>1
Monroe #2
Toledo tt1
Port Stanley
Rondeau Park
T-Sed-P
in DCDA's
(ngP/gsed.)
2656
3044
947
662
557
Sediment P
Released in
DCDA's
(ngP/g sed.)
964.2
1341.3
99.3
0
0
Ultimately
Bioavailable
Sediment P,
Pu/,(p.gP/gsedJ
1435.0
1482.4
99.3
0
0
Pu/tas a
% of
T-Sed-P
54.0
48.7
10.5
0
0
fie/ease
Rate, kr
(day-')
0.42
0.080
0.140
—
—
Table 4. Effects of a Point Source Discharge on Phosphorus Fractionation in Sandusky River Bottom Sediments
Miles
from
Site point
no. source
1 9
2 7
3 3
4 0.5
5 5
6 10
T-Sed-P
777
644
470
1510
1039
1404
T-NaOH-P
284
226
55
528
397
596
Concentrations in ngP/g sediment
R-NaOH-P NR-NaOH-P CDB-P
215
177
43
434
320
305
69
49
12
94
77
291
214
160
92
452
308
177
HCI-P
125
123
263
249
165
323
Sandusky River
Residual P Flow
107
91
54
132
119
170
Point
Source
Discharge
References
DePinto, J.V., "An Experimental Appara-
tus for Evaluating Kinetics of Available
Phosphorus Release from Aquatic Par-
ticulates." Water Res., 16, 1065-1070
(1982).
Task Group III, "Fifth Year Review of
Canada-United States Water Quality
Agreement." Report of Task Group III,
A Technical Group to Review Phospho-
rus Loadings, U.S. Department of
State, Washington, D.C. (1978).
U.S. Army Corps of Engineers, "Lake Erie
Wastewater Management Study - Final
Report." Buffalo, N.Y. (1982).
S. C. Martin, J. V. DePinto, and T. C. Young are with Clarkson College of
Technology, Potsdam, NY 13676.
W. L. Richardson is the EPA Project Officer (see below).
The complete report, entitled "Biological Availability of Sediment Phosphorus
Inputs to the Lower Great Lakes," (Order No. PB 85-121 036; Cost: $17.SO,
subject to change) will be available only from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Large Lakes Research Station
Environmental Research Laboratory—Duluth
U.S. Environmental Protection Agency
Grosselle.MI48138
U. S. GOVERNMENT PRINTING OfFICE: 1985/559 111/10846
-------
United Slates
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
Official Business
Penalty for Private Use $300
Oi
u ••>
p -,
•> .CNVl? PFJT-CTK'/v £r.fNCY
* o i: N ;• L : '• ? a v y
c T f !~ !; T
ft" 6 'J -'•"
------- |