PB85-121036
EPA-600/3-84-100
October 1984
BIOLOGICAL AVAILABILITY OF SEDIMENT PHOSPHORUS
INPUTS TO THE LOWER GREAT LAKES
by
Scott C. Martin, Joseph V. DePinto,
and Thomas C. Young
Department of Civil and Environmental Engineering
Clarkson College of Technology
Potsdam, New York 13676
Grant No. CR807155
Project Officer
William L. Richardson
Large Lakes Research Station
Environmental Research Station - Duluth
Grosse lie, Michigan 48138
ENVIRONMENTAL RESEARCH LABORATORY - DULUTH
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
DULUTH, MINNESOTA 55804
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
. REPORT NO.
EPA-600/3-84-100
2.
3. RECIPI
. TITLE AND SUBTITLE
Biological Availability of Sediment Phosphorus Inputs
to the Lower Great Lakes
5. REPORT DATE
October 1984
6. PERFORMING ORGANIZATION CODE
. AUTHOR(S)
S. C. Martin, J. V. DePinto, and T. C. Young
8. PERFORMING ORGANIZATION REPORT NO.
. PERFORMING ORGANIZATION NAME AND ADDRESS
Department of Civil and Environmental Engineering
Clarkson College of Technology
Potsdam, New York 13676
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
CR 807155
12. SPONSORING AGENCY NAME AND ADDRESS
Environmental Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Duluth, Minnesota 55804
13. TYPE OF REPORT AND PERIOD COVERED
14. SPONSORING AGENCY CODE
.EPA-600/03
15. SUPPLEMENTARY NOTES
16. ABSTRACT
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-HCl , as well as algal "bioassay analyses of sediment:
P bioavailability using the Dual Culture Diffusion Apparatus (DCDA) technique of
DePinto.
Sediments from several of the bioassay experiments were reconcentrated after the
bioassays and resubjected to the chemical fractionation sequence.
Several other forms of P inputs to the Lower Great Lakes were also analyzed for
chemical composition and/or bioavailability.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS C. COS AT I Field/Group
18. DISTRIBUTION STATEMENT
Release to public
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
187
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (R»». 4-77) PREVIOUS EDITION is OBSOLETE 1
-------
DISCLAIMER
"Although the research described in this article has been
funded wholly or in part by the United States Environmental
Protection Agency through grant number CR807155010 to Clarkson
College of Technology, it has not been subjected to Agency review
and therefore does not necessarily reflect the views of the
Agency and no official endorsement should be inferred."
11
-------
ABSTRACT
Early efforts to control eutrophication in the Lower Great
Lakes focused primarily on the reduction of phosphorus (P)
concentrations in wastewater treatment plant effluents discharged
to the Lower Lakes basin. Presently, consideration is also being
given to the reduction of diffuse (runoff) sources of P through
land management practices such as conservation tillage. In
assessing the merit of different load reduction programs, it is
important to consider the relative bioavailabilities of the
various forms of P entering the lakes. In this study, river
water samples were collected from several major tributaries to
the Lower Great Lakes (Maumee R., Sandusky R., Honey Cr.,
Cuyahoga R., Genesee R., and Cattaraugus Cr.) 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-HCl, as well as algal
bioassay analyses of sediment P bioavailability using the Dual
Culture Diffusion Apparatus (DCDA) technique of DePinto (1982).
NaOH-extractable P was the largest of the chemical fractions
measured for the Ohio tributaries. HC1-P was the largest
fraction for New York river sediments, reflecting a geology
higher in apatite minerals. The release of sediment P in the
presence of P-starved SelenastZlM capiic.o.j:.n:u£.um followed
approximately first-order kinetics. The mean percentages of
total sediment P (T-Sed-P) determined by the bioassay method to
be ultimately available were: Maumee R. - 25.0%;
Sandusky R. - 21.4%; Cuyahoga R. - 33.9%; Honey Cr. - 24.9%;
Genesee R. - 19.3%; and Cattaraugus Cr. - 7.7%. The chemical
fractionation and bioavailability of sediment P showed greater
variation between tributaries than among samples from any given
tributary. The rate of sediment P release showed relatively
little variation between tributaries, and averaged 0.182 day"
for all river sediment samples analyzed.
Sediments from several of the bioassay experiments were
reconcentrated and subjected to the chemical fractionation
sequence. P released during the bioassays was found to be
associated predominately with the reactive NaOH-P fraction
(R-NaOH-P). A number of the chemical fractions were strongly
correlated with the amount of sediment P ultimately available to
55l£Da^ijjjm (Pujt) • Regression equations of ?ult versus R-NaOH-P
and non-apatite inorganic P (NAIP = R-NaOH-P +• CDB-P) were
developed and shown to be significant at the 0.0001 level.
i
The model expressions currently used to describe P release
from suspended sediments in models of Great Lakes trophic status
overpredicted the bioassay data for tributary sediments by a
111
-------
considerable amount after a period of about 7 days. Coefficients
were calculated for two different models that more accurately
described the bioassay results. In one model, Puit was
considered to be released at a single first-order rate, while in
the other, pi was separated into rapidly and slowly released
components. Based on the latter model, the rapidly released
component comprised the following mean percentages of Puit for
the individual tributaries: Maumee R. - 24.3%; Sandusky R. -
29.1%; Cuyahoga R. - 8.6%; Honey Cr. - 17.0%; and
Cattaraugus Cr. - 19.4%.
Several other forms of P inputs to the Lower Great Lakes
were also analyzed for chemical composition and/or
bioavailability. Soluble reactive P in the tributary samples was
reduced to very low levels (2-5 jug/1) within 24 hours by
P-starved algae in bioassay experiments. Phosphorus in a sewage
treatment plant effluent discharged to the Sandusky River was
found to accumulate in river bottom sediments in a variety of
chemical forms. Other forms of particulate material investigated
include Detroit River suspended sediments, Lake Erie bottom
sediments, and material from eroding bluffs on the shoreline of
Lake Erie.
IV
-------
References 131
Bibliography of Related Publications 141
Appendix
A. The Power Law Function and Related Equations 142
B. Determination of Diffusion Coefficient for DCDAs 147
C. Derivation of an Equation for Phosphorus Uptake by 154
Algae on the Assay Side of DCDA Reactors (Based
on DePinto, 1982)
D. Complete Listing of Results Summarized in 158
Tables 7 and 8
E. Equations and Computer Algorithm Used in Calculating 164
Two-Component Sediment P Release Coefficients
VI
-------
LIST OF FIGURES
I Binuclear complex (from Kingston, e£ 3l./ 1974). 15
2 Location of tributary sampling sites. 32
3 General flow chart for analysis of tributary 38
samples.
4 Procedure for chemical f ractionation of sediment P. 46
5 Procedures for bioassay measurements of phosphorus 49
availability.
6 Assembled DCDA reactor. 50
7 Photograph of assembled DCDA reactor. 51
8 Cumulative release of sediment P versus time 67
measured in DCDA bioassays on samples 119 (Maumee
River) and #20 (Sandusky River) .
9 Comparison of current Great Lakes model predic- 81
tions of Prel versus time with actual bioassay
data and the one-component model predictions for
sample #19.
10 Comparison of one-component (Eq. 14) and two- 84
component (Eq. 16) sediment P release models
with bioassay data for sample #19.
11 Regression line of P ,, versus R-NaOH-P super- 92
imposed on data for tributary suspended sediments.
12 Regression line of Pult versus NAIP (non- 93
apatite inorganic P) superimposed on data for
tributary suspended sediments.-
13 Regression line of k versus Residual P super- 94
imposed on data for tributary suspended sediments.
14 Regression line of Pult versus the decrease in 100
R-NaOH-P during bioassays (-AR-NaOH-P) superim-
posed on data for tributary suspended sediments.
vn
-------
15 Cumulative particle volume versus particle 102
diameter for sample 148 (Honey Creek).
16 Power law distribution plot - log(&N/Ad ) 104
versus log d for sample #48.
17 Comparison of cumulative volume distributions 107
for sample #48 (Sandusky R.) with and without
30 urn aperture data.
18 Regression line of D^Q versus flow superimposed 111
on data for Sandusky River suspended sediments.
19 Comparison of trends in D50v/ 630' and T-Sed-P 112
with flow over the course of four runoff events
on the Sandusky River.
20 Relationship between normalized absorbance and 116
normalized suspended solids for sediments from
sample #44 (Sandusky River), based on a settling
column analysis (from Kwolek, 1982).
21 Comparison of trends in SV50f TSSf and k with 120
flow over the course of runoff events on the
Sandusky River.
A-l Hypothetical particle size distribution (from 143
Lawler, 1979) .
A-2 Significance of 0 values (from Lawler, 1979) . 145
B-l Results of first DCDA diffusion experiment. 150
Vlll
-------
LIST OF TABLES
Sage
I Tributary sampling stations and sample collectors 33
2 Dates and locations of sample collection 35
3 Analytical methods and references 39
4 Coulter counter and channelyzer settings 42
5 Electrolyte solution - 0.6% Isotone (from 44
Autenrieth, 1981)
6 Composition of algal growth medium (modified from 52
Guillard and Lorenzen, 1972)
7 Summary of physical and chemical measurements on 57
unfiltered tributary samples and filtrates
8 Summary of P fractionation results for tributary 58
suspended sediments
9 Results of chemical analyses on tributary 60
sediments from previous studies
10 Mean composition of the non-apatite inorganic P 64
(NAIP) fraction in tributary suspended sediments
11 First-order release coefficients calculated from 69
bioassay data by the Thomas method
12 Coefficients of variation for Pult measured by 72
DCDA bioassays
13 Two-component release coefficients (for Equation 76
12) calculated from bioassay results
14 Simple correlation coefficient matrix for one- 79
component and two-component sediment P release
coefficients
15 Bioavailability of soluble phosphorus in 86
tributary samples
16 Simple correlation coefficients between one- 88
component sediment P release coefficients and
chemical fractions for tributary suspended sediments
IX
-------
17 Partial correlation coefficients between biolog- 89
ically available P and chemical fractions for
tributary suspended sediments
18 Slopes and y-intercepts from linear regression 91
analyses of one-component sediment P release
parameters versus chemical fractions
19 Simple correlation coefficients between two- 96
component sediment P release coefficients and
chemical fractions for tributary suspended sediments
20 Post-bioassay sediment P fractionation results 98
21 Changes in chemical fractionation of sediment P 99
during bioassays compared with P release measured
in DCDAs
22 Median particle diameters and power law exponents 105
for tributary suspended sediments
23 Simple correlation coefficients between sediment 109
particle size parameters and variables related to
the flux and bioavailability of sediment P for
Sandusky River samples
24 Median particle settling velocities for tributary 114
suspended sediments
25 Simple correlation coefficients involving median 118
particle settling velocities
26 Chemical fractionation and bioassay results for 122
shoreline erosion (bluff) and lake bottom sediment
samples
27 Effects of a point source discharge on phosphorus 125
fractionation in Sandusky River bottom sediments
B-l Results of first DCDA diffusion experiment 149
B-2 Results of second DCDA diffusion experiment 152
D-l Results of physical and chemical measurements on 158
unfiltered tributary samples and filtrates
D-2 Results of P fractionation studies on tributary 161
suspended sediments
-------
ABBREVIATIONS AND SYMBOLS
UNITS
ft-ca
9
9/1
mg/1
.MgP/g sed
hr
1
ml
m
cm
mm
nm
min
N
Angstroms
Degrees Centigrade
foot candles
Gravitational acceleration
grams per liter
milligrams per liter
micrograms per liter
micrograms of phosphorus per gram of sediment
dry weight
hours
liters
milliliters
meters
centimeters
millimeters
micrometers (or microns)
nanometers
micromoles per liter
minutes
Normal (equivalents per liter)
ANALYTICAL CONCENTRATIONS
CDB-P
HC1-P
NAIP
PP
NR-NaOH-P
R-NaOH-P
Residual P
SRP
T-NaOH-P
TP
TPP
T-Sed-P
TSP
TSS or SS
Sediment P extracted by citrate, dithionite, and
bicarbonate reagents in reactive form
Sediment P extracted by hydrochloric acid reagent in
reactive form
Non-apatite inorganic P (-• R-NaOH-P + CDB-P)
Particulate phosphorus
Sediment P extracted by sodium hydroxide reagent in
non-reactive form
Sediment P extracted by sodium hydroxide reagent in
reactive form
Sediment P removed by persulfate digestion
after chemical fractionation
Soluble reactive phosphorus
Total sediment P extracted by sodium hydroxide
reagent
Total phosphorus
Total particulate phosphorus
Total sediment phosphorus
Total soluble phosphorus
Total suspended solids (>0.45
xi
-------
DIFFUSION EQUATIONS
A Open pore area in membrane
2
D Molecular diffusion coefficient (cm /day)
m
P Initial soluble P level on spiked side of DCDA
P Soluble P level on spiked side of DCDA
s
P Soluble P level on unspiked side of DCDA
a Bulk diffusion coefficient (day )
a Membrane thickness
STATISTICAL ANALYSES
C.V. Coefficient of variation
HQ: p=0 Null hypothesis
n Number of observations
r Correlation coefficient
a Probability of rejecting a true null hypothesis
PARTICLE SIZE ANALYSIS
A Coefficient related to total concentration of
particulate matter in suspension
BCT Base channel threshold
CN Channel number
d Particle diameter
Volume-weighted median sediment particle diameter
DSO Number-weighted median sediment particle diameter
N Particle number
S Particle surface area
SV5Q Median particle settling velocity based on 50%
decay in absorbance
SV2g Median particle settling velocity based on 20%
decay in absorbance
TF Threshold factor
V Particle volume
WW Window width
0 Power law exponent (unitless)
$ 20 Power law exponent obtained from Coulter counts
with 30 yum aperture
Xll
-------
3 5Q Power law exponent obtained from Coulter counts
with 50 ^m aperture
DCDA BIOASSAY ANALYSIS
C
Cl
C2
DCDA
rr
sr
m
pr
Prel
ult
V
w
OTHER
AAP
APHA
ATP
Co
EDTA
EPC
ER
K
1/m
M.W.
Constant of integration
Soluble P concentration in sediment vessel
Concentration of algal particulate P in assay vessel
Dual culture diffusion apparatus
First-order release rate of available P(day )
First-order release rate for rapidly available
component of sediment P (day )
First-order release rate for slowly available
component of sediment P (day )
Mass of sediments in sediment vessel
Amount of rapidly-released sediment P
Amount of P released from sediments at any given time
Amount of slowly-released sediment P
Total amount of sediment P ultimately bioavailable
Amount of P released from sediments, but remaining
in the sediment vessel
Amount of P taken up by algae in assay vessel of DCDA
Volume of liquid in DCDA vessel
Rate of P release from sediments in DCDA (mass/time)
Algal assay procedure
American Public Health Association
Adenosine triphosphate
Initial soluble P concentration (adsorption kinetics)
(Ethylenedinitrilo)tetraacetic acid
Equilibrium phosphorus concentration
Enrichment ratio
Constant in adsorption kinetics expression
Constant in adsorption kinetics expression
Molecular weight
Xlll
-------
NTA Nitrilotriacetic acid
O.D. Outside diameter
p.d. Potential determining
USGS United States Geological Survey
X Amount of P adsorbed per unit weight of adsorbent
A Change in a given parameter
xiv
-------
ACKNOWLEGMENTS
The authors would like to thank several persons who have
made valuable contributions to this project. Dr. David Baker and
the staff of the Heidelberg College Water Quality Lab provided
samples from the Ohio rivers. Dr. Steve Yaksich and Steve
Predmore of the U.S. Army Corps of Engineers, Buffalo District,
provided Cattaraugus Creek samples. The following Clarkson
College personnel conducted much of the experimental work
reported herein: Michael Bollenbacher and William Beckwith -
chemical fractionations; James Bonner - particle size
distribution software; John Kwolek - particle size and settling
velocity measurements; and Anne Marie Victore - chemical
fractionations and particle size measurements. Dr. Victor
Bierman served as the project officer during the organizational
and experimental phases of the study. Also, Nelson Thomas, Chief
of the EPA Large Lakes Research Station, demonstrated continual
interest in the project and contributed several valuable
suggestions.
xv
-------
SECTION 1
INTRODUCTION
BACKGROUND
Since the signing of the Great Lakes Water Quality Agreement
by the governments of Canada and the United States in April,
1972, an extensive effort has been directed at curbing the
problem of eutrophication in the 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 effective means of controlling eutrophication.
Accordingly, target total phosphorus loads were established based
on the reduction of treatment plant effluent phosphorus
concentrations to 1.0 mg/1. 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 improvements in phosphorus removal at wastewater treatment
facilities discharging to the Lower Lakes basin. Although
municipal loads have decreased considerably, it has now become
clear ,that 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 present programs. In addition, improvements in
the water quality of the lakes resulting from these load
reductions have, to date, been disappointing (Sugarman, 1980) .
Future courses of action being considered include further
reductions in municipal treatment plant effluent phosphorus
concentrations to 0.5 mg/1 and improvement of land management
practices (e.g. implementation of conservation tillage programs)
in order to reduce diffuse (runoff) sources of phosphorus 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 phosphorus from
different sources for supporting algal growth must be carefully
evaluated in developing a cost-effective management scheme for
the contol of eutrophication (Sonzogni, si_3l., 1982). Nearly
50% of the total P load to Lake Erie and 40% of the Lake Ontario
load are from diffuse sources (Chapra and Sonzogni, 1979). A
large fraction of this phosphorus is in a sediment-bound form and
may, therefore, not be readily available for algal uptake. In
fact, studies of several Great Lakes tributaries have indicated
that generally 40% or less of suspended sediment total P is
potentially bioavailable (Armstrong, si_5l., 1979; Logan,
-------
1979). On the other hand, much (often over 70%) of the
phosphorus in municipal treatment plant effluents is present in
easily assimilable forms (DePinto, si_3l.f 1980; Young, si_sl.f
1982) .
To facilitate the evaluation of various programs for
eutrophication control, a number of mathematical models have been
developed that provide predictions of the water quality
improvements that would result from these programs. Several of
these models (e.g. Bierman, 5i_sl., 1980; DiToro and Connolly,
1980; Thomann, 5±_3l., 1975) include explicit distinctions
between available and unavailable phosphorus forms. However, the
behavior of all particulate P in the water column is assumed to
be the same, whether it originates from allochthonous or
autochthonous sources. Estimates of the potential
bioavailability of particulate P inputs to the Lower Lakes have
been made by a variety of methods, including bioassays and
chemical extractions. However, the kinetics of conversion of
unavailable allochthonous phosphorus to available forms have not
been adequately quantified to date. It is possible that
particulate material may settle out of the water column before
all of its potentially available P is released (Sonzogni, £±_.al.f
1982; Logan, £i_5l.f 1979). In addition, the rate of phosphorus
release from allochthonous particulate material in the water
column may differ considerably from the regeneration rate for
autochthonous particulate P. Therefore, experimentation designed
to characterize the kinetics of phosphorus release from suspended
sediments under in-lake conditions could significantly improve
the predictive capabilities of trophic status models.
As phosphorus associated with suspended particulate material
enters the receiving water environment, it may become available
for algal uptake via such mechanisms as 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, hydrology and mixing
dynamics, size, shape and density of the particles, and the
chemical forms of phosphorus in the particles (Logan, £±_.3l. ,
1979). A more thorough understanding of the relationships
between these factors and the bioavailability of particulate P is
necessary in order to predict phosphorus release for a wide
variety of sediment and receiving water conditions.
OBJECTIVES OF STUDY
The general goal of this research was to obtain information
that would permit refinement of mathematical models of lake
trophic status through the development of a sediment phosphorus
bioavailability submodel. The specific objectives of the
experimental program were:
-------
1. to determine the rate and extent of sediment P
bioavailability for a variety of tributaries to the Lower
Great Lakes under receiving water conditions using a
bioassay procedure;
2. to physically and chemically characterize the tributary
suspended sediments with respect to factors influencing
phosphorus bioavailability;
3. to develop a methodology that may be used to estimate the
rate and extent of available P release from suspended
sediments based on the characteristics of the sediments;
4. to compare the P bioavailability and chemical
characteristics of tributary suspended sediments to other
potential sources of sediment P to the Lower Great Lakes.
SCOPE OF WORK
To accomplish the objectives of this study, river water
samples were collected from six Lower Great Lakes tributaries
(Maumee R., Sandusky R., Cuyahoga R., Honey Cr., Genesee R. , and
Cattaraugus Cr.) during storm runoff events. Suspended sediments
from these samples were concentrated and algal bioassays of
sediment phosphorus availability were conducted. The sediments
and river water were also subjected to a variety of chemical and
physical analyses, including a chemical fractionation sequence
for sediment phosphorus and particle size distribution. Kinetic
coefficients (rate and amount) for the release of sediment P were
determined and correlations between these values and the physical
and chemical parameters were examined. This was done in an
effort to develop a means of predicting the kinetics of sediment
P bioavailability based on a more convenient analytical procedure
than the sediment P bioassays.
Several other forms of particulate material from the Lower
Lakes basin were characterized in this study as well. These
included Sandusky River bottom sediments near a point source of
P, Lake Erie bottom sediments, Detroit R. suspended sediments,
and shoreline erosion (bluff) sediments.
-------
SECTION 2
CONCLUSIONS
1) The chemical fractionation and bioavailability of P bound to
tributary suspended sediments showed little variation among
samples from a particular tributary or between the two years
of the study (1980 and 1981) . Greater disparity was found
between tributaries having differences in soil type or land
use conditions. Suspended sediments from the Ohio rivers
contained higher concentrations of non-apatite inorganic P
(NAIP = Reactive NaOH- plus CDB-extr actable P) than those from
the New York rivers. Sediments transported by the New York
rivers contained more apatite P (HCl-extracfcable) than those
from the Ohio rivers.
2) Suspended sediments transported during storm runoff events in
Ohio tributaries (Maumee R. , Sandusky R. , Honey Cr./ and
Cuyahoga R. ) contained higher levels of bioavailable
phosphorus than those from New York tributaries (Cattaraugus
Cr. and Genesee R.). The mean concentrations of ultimately
available sediment P (Pu]_t) measured by DCDA bioassays ranged
from 38.8 ugP/g sed. (7.7% of total sediment P) for
Cattaraugus Cr. to 449.2 pgP/g sed. (33.9% of total sediment
P) for the Cuyahoga R.
3) The release of available phosphorus from tributary suspended
sediments in the presence of P- starved 5sl5Jja5ixiffij
followed approximately first-order kinetics.
Release rates (kr) ranged from 0.043 to 0.266 day" , and
averaged 0.182 day for all tributary samples.
4) An expression used to describe P release from water column
•particulate material in current Great Lakes models
(a first-order function of total particulate P) grossly
overpredicted the observed release from tributary suspended
sediments at times greater than about 7 days. A first-order
expression based on the ultimately available fraction of
sediment P provided a much more accurate prediction of P
release from tributary sediments.
-------
5) A model in which Pult is separated into rapidly released and
slowly released components predicted the experimentally
observed release slightly more accurately than a single
first-order expression based on ultimately available sediment
P.
6) Reactive NaOH-extractable P (R-NaOH-P) and NAIP were strongly
correlated with Puit* The following regression equations were
obtained for P ,. versus these two chemical fractions, based
on all tributary suspended sediment samples analyzed:
Pult = °*8824 (R-NaOH-P) + 17.1 (r = 0.779)
Pult = 0.6363 (NAIP) - 58.0 (r = 0.832)
7) Phosphorus released from tributary suspended sediments in the
presence of P-starved SslssastXlim £apiico,.rii.ui.um was
predominately associated with the R-NaOH-P fraction. The
change measured in the R-NaOH-P fraction during the course of
DCDA bioassays averaged over 90% of the P release measured
during the bioassays. On the other hand, the change in the
CDB-P fraction accounted for under 8% of sediment P release.
Changes in the non-reactive NaOH-P and Residual P fractions
were closely correlated with the first-order P release rate,
kr.
8) Phosphorus in sediments from eroding bluffs on Lake Erie was
predominately associated with apatite, and was completely
unavailable to SslsnastXJM over the course of bioassay
experiments.
9) The relative contributions of allochthonous and autochthonous
material to Lake Erie bottom sediments may vary markedly as a
function of location in the lake. In areas receiving
substantial autochthonous inputs, the bottom sediments may
experience a considerable enrichment in biologically available
forms of P. The amount of P released by such sediments upon
resuspension may be comparable to, or exceed, that released
from tributary suspended sediments (in^ugP/g sed) .
10) P discharged by a point source (municipal wastewater
treatment plant) to the Sandusky R. was found to accumulate in
the downstream bottom sediments in a variety of chemical
forms. Although all chemically extractable forms of sediment
P showed increases at the downstream sampling sites, by far
the greatest accumulation occurred in the R-NaOH-P and CDB-P
fractions.
-------
SECTION 3
RECOMMENDATIONS
1) The current models of Great Lakes trophic status should be
restructured to reflect the distinctly different
bioavailability kinetics of particulate P from allochthonous
and autochthonous sources.
2) The bioavailability of sediment P transported to the Lower
Great Lakes by tributaries (particularly the Detroit River)
should be more accurately characterized through additional
bioassay and chemical fractionation analyses. Resources could
be utilized most efficiently by concentrating the monitoring
on areas of differing soil and land use conditions.
3) Regression equations used to estimate the amount and rate of
sediment P release based on chemically extractable fractions
should be improved by the addition of new data as it becomes
available. A better understanding of the mechanisms
responsible for these relationships might be gained through
analysis of chemical extracts for metals typically active in
the binding of P (e.g. Fe, Al, Ca) and/or a measurement of
biological activity (e.g. ATP) in the sediments during
bioassay experiments.
4) Further experimentation is necessary to check the validity of
a two-component model for P release from tributary sediments.
More frequent sampling of the reactors during the first week
of the bioassays would be required.
5) Studies of the fractionation of P associated with various
microorganisms at different stages of growth and decomposition
would provide useful insights into relationships between
chemically defined fractions of sediment P (e.g. NR-NaOH-P and
Residual P) and the release rate of available P.
6) Models of Great Lakes water quality should include expressions
describing the resuspension of bottom sediments by
wind-induced wave action, coupled with a kinetic expression
for the release of P associated with these sediments.
7) Considering the diversity of lake bottom sediments indicated
by this study, it is important that the kinetics of sediment P
bioavailability be thoroughly characterized at locations where
resuspension is likely.
-------
8) Additional study is needed to ascertain the fate of P
discharged to Lower Great Lakes tributaries from point
sources. In particular, the effect that attachment to bottom
and suspended sediments has on the flux of this P to the
receiving lake and on its bioavailability should be
investigated.
9) The impact of programs aimed at the reduction of diffuse
sources of P to the Lower Lakes, such as reduced tillage,
should be monitored as these practices are implemented. This
would require quantification of the total P flux and the
bioavailability of the soluble P and particulate P components
before and after implementation of the diffuse source
controls.
-------
SECTION 4
REVIEW OF RELATED RESEARCH
INTRODUCTION
Phosphorus has long been recognized as a nutrient essential
to the growth of aquatic and terrestrial plants. As a result,
extensive information is available in the scientific literature
on the chemical forms and dynamics of phosphorus in lakes and
soils. However, the biological availability of particulate
phosphorus upon entering lakes has only recently been
investigated in detail. The objective in this chapter is to
summarize the literature that either contributes to the current
knowledge of sediment phosphorus availability or indicates a
direction for future research on the topic.
PHOSPHORUS TRANSPORT IN TRIBUTARIES
. njesh anisms
The effect that diffuse sources of sediment phosphorus have
on a lake's productivity depends to a large extent on the recent
history of the particulate material. Important factors include
the point of origin of the particles in the watershed or stream
and the conditions in the tributary during the period of
transport to the lake. Extensive investigations in tributaries
to the Great Lakes have provided some useful insights into the
transport of phosphorus and suspended solids (U.S. Army Corps of
Engineers, 1982) .
Most of the total phosphorus transport in streams appears to
occur during storm runoff events and in association with
suspended sediments (Verhoff, si_5l., 1979). As stream discharge
increases during (or following) a storm event, both total P and
suspended solids concentrations typically increase as well
(Cahill, £i_5l.r 1974; Baker, 1982a) . These increases may be
largely attributed to two important processes - erosion of soils
by surface runoff in the watershed and resuspension of stream
bottom sediments.
Rainwater falling on the surface of a watershed moves toward
the adjacent stream by two major pathways. During a
low- intensity storm, much of the water is able to infiltrate into
the soil, and travels laterally to the stream as subsurface flow,
or "interflow". This water often contains high concentrations of
-------
certain soluble species as a result of prolonged contact with the
soil, but is usually low in suspended solids. During
high-intensity storms, the infiltration capacity of the soil is
exceeded and water flows rapidly over the soil surface toward the
stream. This surface runoff water typically contains high
concentrations of suspended solids and low concentrations of
dissolved substances (Baker, 1982a). A number of studies have
shown that the majority of phosphorus in surface runoff is in a
sediment-bound form (Harms, 5i_3l., 1974; Burwell, £i_j&l., 1975;
Sharpley and Syers, 1979) .
Several authors have also suggested that resuspension of
bottom sediments may play an important role in the transport of
total phosphorus during storms. Keup (1968) concluded that
substantial amounts of phosphorus may be temporarily stored in
stream bottom deposits that are subsequently scoured from the
channel and rapidly discharged during periods of rising water
levels. This transport theory, whereby phosphorus moves from the
watershed to the receiving water body in a series of resuspension
and deposition steps initiated by storm events, has been termed
the Discontinuous Theory (Verhoff and Melfi, 1978) . By the
opposing view, the Continuous Flow Theory, total phosphorus is
considered to be washed from the land, through the river system
and into the receiving water body during one storm event.
Verhoff and Melfi (1978) developed a mass balance equation
for water and total phosphorus in rivers to investigate the
plausibility of these two transport theories. The model results
were compared to the characteristics of hydrographs and total
phosphorus chemographs observed during storms in four western
Ohio tributaries to Lake Erie. These observations revealed that:
-the peak of the total P chemograph almost always precedes
the hydrograph peak at the same station;
-the total P concentration declines to its approximate
steady-state value before the river discharge; and
-the peak total P concentration is not necessarily higher at
the downstream stations than at the upstream stations.
These characteristics could be reproduced only when a
resuspension/deposition term was included in the total P mass
balance. The results of Verhoff and Melfi (1978) provide strong
support for the Discontinuous Theory of total phosphorus
transport.
In a subsequent paper, Verhoff, 5±_5l. (1979) described a
technique for estimating the average distance traveled by total
phosphorus passing a given station based on hydrograph and
chemograph data. Application of the technique to the Sandusky,
Maumee and Gattaraugus Rivers revealed that, where the duration
of the hydrograph is short, as in a steeply sloped watershed, the
-------
travel distance of total P (and suspended sediments) is also
relatively short. However, in the mainstreams of the Sandusky
and Maumee Rivers, even rather routine storm events may produce
average travel distances on the order of hundreds of miles
(Verhoff, £i_3l., 1979).
Based on the large variability in nutrient and sediment
loads for storms of equal size, Baker (1982a) has suggested that
material export is not limited by the transport capacity of the
river, but rather, by the movement of the material from the land
surface to the stream system. This is in agreement with the
finding presented by Verhoff, jei_sl. (1979) that transport
distances tend to be shorter in the upper reaches of a watershed.
The total phosphorus and sediment transport considerations
presented above begin to assume importance in terms of
particulate phosphorus bioavailability when one considers the
potential reactions involving particulate P during transport in a
river system.
Suspended sediments eroded from a surface soil generally
have higher particulate phosphorus levels than the soil source.
This enrichment in P is attributed to the selective erosion of
finer particles, which tend to contain more P (Frink, 1969;
Armstrong, .g±_3l., 1979). For example, the clay-sized (<4 >jm)
fraction of suspended sediments collected from the Menomonee
River basin was found to account for an average of 77% of
sediment dry weight and 88% of total particulate P (Dong, 5i_5l.,
1983) . The effect of this selective erosion is often quantified
using a phosphorus enrichment ratio (ER) , calculated as the ratio
of the P concentration in the suspended sediments to that in the
source soil. ER may vary considerably depending upon soil
physical and chemical properties, land usage, soil slope and
rainfall intensity (Sharpley, 1980) .
During fluvial transport, suspended sediments may undergo
additional enrichment due to the adsorption of P from solution
(Sharpley, £i_5l . , 1981; Green, 5i_3l . , 1978). The adsorption
process may also continue after the sediments have been deposited
on the stream bottom. Bottom sediments in the vicinity of a
point source of phosphorus to the Sandusky River were found to
rapidly remove P from solution (Baker, 1980) . This phosphorus
remained in a sediment-bound form during subsequent runoff events
(Baker, 1982b) . Also, Dong, £±_&1. (1983) reported sizeable
increases in the P content of clay-sized bottom sediments near
various sources of P to the Menomonee River.
Green, e£_&l. (1978) reported that bottom sediments from the
Maumee River had higher adsorption capacities and higher
adsorption energies than suspended solids in the river. It was
proposed (by McCallister and Logan, 1978) that chemical changes
10
-------
in eroded soil material following deposition in the stream
accounted for the increased adsorption capacity. Formation of
amorphous iron oxide precipitates capable of adsorbing additional
P was cited as one possibility.
McCallister and Logan (1978) reported equilibrium phosphorus
concentration (EPC) values ranging from 8 to 54 ^u.g/1 (average =
24 ^g/1) for soil clay fractions (the component most likely
eroded during a storm event) and bottom sediments from the Maumee
River basin. The sediments should tend to adsorb P if soluble
levels exceed the EPC and desorb P if the soluble concentration
falls below this value. Flux weighted mean soluble reactive
phosphorus (SRP) concentrations in the Maumee River were on the
order of lOO^g/1 for 1975 to 1978 (Baker, 1982a) . Therefore,
these sediments could be expected to adsorb (rather than desorb)
phosphorus during their residence in the river, both when
suspended and when deposited on the river bottom. Upon entering
the western basin of Lake Erie, where SRP concentrations may
remain under 10 jqg/1 for much of the year (DiToro, 1980) , the
sediments would tend to desorb P to solution.
Some other processes that could potentially affect the
relative distribution of soluble and particulate phosphorus
components in streams were mentioned by Keup (1968) . These
include chemical combination of soluble P with metallic cations
to form precipitates and uptake of soluble P by plankton and
attached plants. Also, phosphorus assimilated by plants in the
stream may be subsequently released through microbial decay of
the organic matter (Keup, 1968).
The amounts and chemical forms of the phosphorus present
initially in eroded soil and of that accumulated during river
transport will have a major influence on the biological
availability of particulate P entering the water column of the
receiving water. Many investigations have been conducted into
the chemical forms of particulate P and their relative mobilities
as well as the effects d>t solution conditions on the uptake or
loss of phosphorus by sediments. This work is described in the
following sections.
PHOSPHORUS ADSORPTION-DESORPTION STUDIES
Studies of phosphate adsorption characteristics have often
been conducted using soils, lake bottom sediments and synthetic
minerals as the adsorbent. Since such particles are often
chemically similar to suspended and bottom sediments in
tributaries, the results of these studies are of interest here as
well.
The materials most frequently cited as the major P-sorbing
components of soils and aquatic sediments are iron and aluminum
11
-------
oxides. Amorphous complexes of hydrated Fe and Al oxides, in
particular, are characterized by numerous hydroxyl groups in
surface positions and readily participate in "exchange
adsorption" reactions with orthophosphate. These substances have
a much higher capacity to adsorb P than crystalline Fe and Al
oxides (Williams, si_3l., 1971a). Based on sorption experiments
performed on lake bottom sediments from 14 Wisconsin lakes,
Shukla, 5Jt_3l. (1971) postulated that most of the phosphorus
adsorbed by both calcareous and noncalcareous sediments was
associated with a gel complex of hydrated Fe oxide. The gel
complex was also thought to contain small amounts of Al-C^,
Si (OH), and organic matter.
Isotherm Equations —
Adsorption of phosphorus onto Fe and Al oxides frequently
follows a "high intensity", or "H-type", isotherm (Ryden, £i_al.,
1977a; Breeuwsma and Lyklema, 1973) . That is, plots of P
adsorbed per gram of adsorbent versus equilibrium soluble P
concentration show a steep rise (next to the ordinate) at low P
concentrations. Similar isotherms have been observed for
adsorption of phosphorus onto hydrous ferric oxide gel and a
variety of soils (Ryden, 5±_al., 1977a) , kaolinite, gibbsite
( Y-A1203' 3H20) and pseudoboehmite (A100H) (Muljadi, s±_al.,
1966a), and hematite (ot-Fe-O-j) (Breeuwsma and Lyklema, 1973).
The results of these and other studies have indicated that
the mechanism of phosphorus adsorption is not uniform over the
entire range of equilibrium P concentrations investigated.
Isotherms are often divided into two or more regions of
relatively constant adsorption behavior. These regions are then
described by separate isotherm equations. Some researchers have
combined the Langmuir isotherm with the Freundlich (Kuo and
Lotse, 1974) or linear (Muljadi, si_al., 1966a) equations, while
others (Ryden, jei_al., 1977a; Syers, s±_al., 1973) have used a
series of Langmuir expressions. An advantage of the latter
approach is that consistent estimates of the energy of adsorption
may be obtained for all equilibrium P concentrations (Syers,
l., 1973). Chen, s±_3l., (1973a) have indicated that, for P
-4
concentrations typical of natural waters (<10 moles/1) ,
adsorption results can be expected to conform to a Langmuir
expression.
Adsorption Mechanisms —
Adsorption at low equilibrium P concentrations is usually
dominated by a relatively strong binding of phosphate to the
adsorbent. Such high binding energies are characteristic of a
chemisorption, or "specific" adsorption, process (Breeuwsma and
Lyklema, 1973; Ryden, s±_al., 1977a) . Several authors have
12
-------
proposed mechanisms for the chemisorption of P on the surface of
Fe and Al oxides. Adsorption is commonly thought to involve an
exchange of phosphate ions for hydroxyl (OH) or protonated
hydroxyl (H20) groups in the first co-ordination sphere of the
metal ions on the oxide surface (Kingston, £±_.sl., 1967; Muljadi,
al.f 1966a and 1966b; Breeuwsma and Lyklema, 1973; Ryden,
l., 1977a) . Although representations of this exchange
reaction vary somewhat, they are typically of the form shown in
Equations 1 and 2 (from Ryden, e±_.al., 1977a).
Fe-OH
1+
Fe-OH
0
(1)
,Fe-OH
\
Fe-OH
/
\
Fe-OH
Fe-H2P04
+ OH
(2)
Hsu (1965) suggested that adsorption is a special case of
precipitation, and that both processes result from the same
chemical force.
Specific adsorption of phosphate on Fe and Al oxides is
generally enhanced by the protonation of surface hydroxyl groups
due to the resulting formation of positively charged sites. This
accounts for the increases in P sorption capacities often
observed with decreasing pH. However, there is some disagreement
as to the effect of phosphate adsorption on surface charge.
Kingston, ei_sl. (1972) proposed two P sorption schemes that
would always result in an increase in negative (or a
neutralization of positive) charge. On the other hand, the
mechanism suggested by Muljadi, ei_5l. (1966a) involves exchange
of H2P04
with OH only, and no net change in surface charge.
Using concentrations of sodium (Na) as an indicator of changes in
surface charge, Ryden, £i_.al. (1977a) found that, at low levels
of P addition (Region I), sorption decreased the surface charge,
while at slightly higher levels (Region II), phosphate adsorption
caused no change in surface charge. Thus, they concluded that P
was adsorbed by both of the mechanisms shown in Equations
1 and 2.
Plots based on a linear form of the Langmuir equation have
been used to evaluate the relative binding strength of P for the
different regions of adsorption isotherms (e.g. Syers, eJt_sl.,
1973; Ryden, e±_.al., 1977a) . The binding energy is generally
found to decrease as the density of P sorbed to the particle
surface increases. Ryden, e£_.al. (1977a) found that adsorption
13
-------
in Region I, thought to occur by exchange of H2P04 for H20
(Equation 1), was thermodynamically more favorable than Region II
adsorption, which was believed to involve exchange with OH
(Equation 2). They indicated that additional Region I adsorption
accounted for most of the observed increases in P sorption
capacity (of iron oxides and soils) resulting from decreases in
pH. In addition, this study revealed remarkable similarities in
the free energies of adsorption between a laboratory preparation
of ferric oxide gel and several soils of varying chemical
composition. Thus, it is likely that P is sorbed to all of these
particles by the same mechanism (Ryden, s£_.al., 1977a) .
A number of researchers have attributed the strong binding
of phosphate on Fe and Al oxide surfaces at low equilibrium P
concentrations to the formation of a binuclear complex such as
that shown in Figure 1 (Atkinson, £i_3l.f 1972; Kafkafi, s£_.al.f
1967; Hingston, £i_sl., 1974). Infrared spectroscopy results
obtained by Parfitt, £i_sl. (1975) indicated that this complex
was the principal mode of phosphate adsorption onto the surfaces
of several synthetic iron oxides. The formation of a binuclear
complex is consistent with the surface charge data collected by
Ryden, si_sl. (1977a). This could occur by a condensation
reaction involving singly co-ordinated phosphate at the oxide
surface as represented in Equation 3 (from Ryden, 5i_sl., 1977a).
/Fe-OH
\ Fe-H PO
Fe -„
\
Fe
0
0
(3)
Kingston, £i_5l. (1974) have characterized phosphate adsorption
as a ligand exchange reaction and designated the Ifeft-hand and
right-hand sides of Equation 3 as monodentate and bidentate
configurations, respectively.
Specific adsorption of P on Fe and Al oxide surfaces has
been found to be largely irreversible with respect to phosphate
concentration (Hingston, si_5l«, 1974; Kafkafi, si_sl., 1967),
although desorption may be induced by increases in pH (Muljadi,
5i_5l., 1966a). Limited desorption has also been reported for
soils (Ryden, s±_.al., 1977a) as well as river bottom (Oloya and
Logan, 1980) and suspended (Green, £±_.al., 1978) sediments at low
levels of surface coverage by phosphate. However, evidence of
the mobility of adsorbed P was presented by Li, si_.al. (1972) ,
who found that 19 to 34% of the total inorganic P in lake bottom
32
sediments was exchangeable (with P). The observed
14
-------
Oh
0
0
.0'
•OH.
Figure 1. Binuclear complex(from Kingston, et. al., 1974)
15
-------
irreversibility is often attributed to the presence of the highly
stable binuclear complex discussed above (Atkinson, s±_sl., 1972;
Kingston, £i_5l., 1974).
Ryden, si_al. (1977a) suggested that adsorption of phosphate
onto soils and Fe oxide gel was essentially irreversible at
equilibrium soluble P concentrations up to 25 >uM (775vugP/l),
while Muljadi, 5i_al. (1966a) saw evidence of irreversibility at
concentrations up to 100 /uM (3100 ^ngP/l) for adsorption onto Al
oxides and kaolinite. Phosphate adsorption at higher levels of
added P is generally dominated by a weaker, more reversible
binding mechanism. The increase in P sorption with increasing
ionic strength observed in this region of the isotherms implies
that phosphate acts as a potential-determining (p.d.) ion (Ryden,
£±_al., 1977a; Breeuwsma and Lyklema, 1973). This adsorption is
more physical, or "non-specific", in nature than adsorption at
lower equilibrium P concentrations. A p.d. sorption mechanism is
consistent with the linear portions of isotherms observed at high
P concentrations (Muljadi, si_sl., 1966a; Breeuwsma and Lyklema,
1973) since no absolute sorption maximum would be expected
(Ryden, 5i_sl./ 1977a) .
Parameters Affecting Phosphate Adsorption —
Some of the parameters that may affect P sorption on Fe and
Al oxides as well as soil and sediment particles have been
mentioned briefly above. The most important factors include the
pH and ionic composition of the solution and the surface area of
the adsorbent. At equilibrium P concentrations up to about 1
mg/lf decreases in pH have been shown to result in increases in
the adsorption of P onto iron oxide surfaces (e.g. hematite and
goethite) over a wide pH range (2-12) . Although adsorption also
exhibits a general increase with decreasing pH at higher P
concentrations, the opposite effect has been observed over narrow
pH ranges in the vicinity of pH=pK2 (=7.2 for phosphate)
(Breeuwsma and Lyklema, 1973) . A somewhat different behavior has
been reported for P adsorption onto several Al oxides and clays.
Adsorption of P from solution onto these materials increased as
pH was decreased, reaching a maximum at about pH 4-5. However,
further decreases in pH caused abrupt reductions in the amounts
of P adsorbed (Chen, 5i_al., 1973a; Edzwald, si_3l., 1976;
Muljadi, 5i_sl., 1966a) . Similar trends were seen for a wide
range of phosphorus concentrations. Edzwald, 5i_5l. (1976)
noticed a different pH dependency for the adsorption of phosphate
onto montmorillonite. They found a roughly linear increase in P
sorption with increasing pH (between 2 and 10) . This was
attributed to reaction of exchangeable calcium on the clay with P
to form an insoluble calcium phosphate phase.
The presence of certain ions in the surrounding solution (or
"support medium") has been shown to affect the adsorption of
phosphate on a variety of Fe and Al oxides and soils. Chen,
16
-------
2+ 2+
Si_sl. (1973a) found that some cations (Ca , La , and a
cationic polyelectrolyte) increased P adsorption onto alumina
(y-Al203) and kaolinite for pH>4. They indicated that cations
increase phosphate adsorption by reducing the repulsive force
between a negative surface and phosphate ions. However, since
other cations (Ni , Co , and Mg ) did not influence adsorption
perceptibly, it is not likely that the observed effects may be
explained entirely by electrostatic interactions. Helyar/
Si_3l. (£976) reported that divalent cations with ionic radii
near 1.0 A (e.g. Ca, Cd, and Sr) increased the adsorption of
phosphate by gibbsite, while larger or smaller ions (e.g. Mg, Zn,
Na, and K) showed little effect. These authors proposed the
formation of a surface complex composed of adsorbed P and a
divalent cation, and suggested that ion size is more important
than electronegativity in complex formation. Other studies have
shown that, while phosphate adsorption generally increases with
increasing ionic strength of the support medium, the introduction
2+
of some cations (e.g. Ca ) results in an additional enhancement
of P adsorption beyond that expected on the basis of ionic
strength alone (Robarge and Corey, 1979; Ryden and Syers, 1975) .
However, for phosphate adsorption onto soils, at least, it
appears that the ionic strength and the dominant cation often
affect only the kinetics of P removal from solution by the
adsorbent. Ryden and Syers (1975) found that, for low levels of
P addition, where chemisorption dominates overall sorption,
differences due to ionic strength and cation effects were
eliminated at equilibrium. At higher phosphate concentrations
(>30 yUM) , on the other hand, differences in adsorption similar to
those mentioned above were maintained at equilibrium (Ryden,
, 1977b).
The influence of various anions in the surrounding solution
on phosphate adsorption by clays and Al oxides has also been
investigated. Chen, _e£_.al. (1973a) reported that sorption of P
onto kaolinite and alumina decreased in the presence of certain
anions (e.g. F , humic acid) for pH<6, while others (e.g. HCO, ,
SiCKOHK ) showed little influence on adsorption. The same study
also demonstrated a decrease in phosphate sorption over a wider
pH range (pH<8) upon the addition of complexing agents such as
tartrate, EDTA, and citrate. The effects of Cl~ and S042~
additions on the adsorption of P by clays was used by Edzwald,
ji_ al . (1976) to assess the role of ion exchange reactions in
phosphate sorption. These anions caused no decrease in the
17
-------
CONTENTS
Disclaimer ii
Abstract iii
List of Figures vii
List of Tables ix
Abbreviations and Symbols xi
Acknowledgments xv
1. Introduction 1
Background 1
Objectives of study 2
Scope of work 3
2. Conclusions 4
3. Recommendations 6
4. Review of Related Research 8
Introduction 8
Phosphorus transport in tributaries 8
Phosphorus adsorption-desorption studies 11
Chemical fractionation of sediment phosphorus 20
Estimation of sediment P bioavailability 24
by chemical methods
Bioassay measurements of sediment P 26
availability
5. Experimental Methods 31
Sample collection 31
Sample handling and preparation 34
Analytical program 37
6. Results and Discussion 56
Routine analysis on unfiltered tributary 56
samples
Sediment P fractionation results 56
Bioassay measurements of phosphorus avail- 66
ability for tributary samples
Correlations between chemical fractiona- 87
tion and bioassay results
Changes in sediment P fractionation 97
during DCDA bioassays
Particle size distributions and settling 101
velocity analyses
Lake Erie bluff and bottom sediment 121
samples
Sandusky River bottom sediment samples 124
7. Implications of Results for Phosphorus Management 127
v
-------
amounts of P adsorbed, indicating that sorption is not driven by
a purely electrostatic mechanism.
The phosphate sorption capacity of any material depends to a
large extent on the surface area of the particles. Amorphous Fe
oxides and hydroxides have a considerably greater surface area
than substances with more crystalline structures, such as
goethite, hematite, gibbsite, clays and soils, and also exhibit
much higher P sorption capacities (Parfitt, .gi_3l., 1975; Ryden,
5i_5l-f 1977a; Muljadi, si_3l.f 1966a) . Bottom sediments in
lakes and rivers may be exposed to conditions (water saturation
and fluctuating redox potential) that favor the formation of
amorphous hydrated metal oxides. It has been proposed that
coatings of these materials can result in substantial increases
in the P sorption capacity of aquatic sediments (McCallister and
Logan, 1978; Shukla, .e±_al., 1971; Williams, .ei_.al., 1971a) .
Although the sorption of phosphate onto aquatic sediments
has been largely attributed to Fe and Al oxides on the particle
surfaces, there is evidence that calcite (CaCO^) may also
contribute to the P sorption capacity of some sediments (Green,
Si.fll.» 1978; Wentz and Lee, 1969). Green, £_t_sl. (1978) found
that the calcite content of suspended river sediments was
positively correlated with native total sediment P and negatively
correlated with adsorption energy. They concluded that, while P
can be adsorbed by calcite, the binding force is weak and P may
be easily desorbed. Other studies, conducted on lake bottom
sediments (Shukla, £±_3l., 1971) and on soils and river bottom
sediments (McCallister and Logan, 1978) have shown negative
correlations between CaCOg content and phosphate sorption for
calcareous as well as noncalcareous sediments.
Experiments have been conducted to characterize the
adsorption of phosphate onto commercially prepared calcium
carbonate. The isotherms consisted of a wide range of
equilibrium P concentrations (0 to 2500 ^qgP/1, or more) where
results conformed to the Langmuir equation, followed by a range
corresponding to calcium phosphate precipitation (Kuo and Lotse,
1972; Cole, 5i_3l., 1953). However, even at the upper end of the
Langmuir portion of these isotherms, the amounts of P adsorbed
were well below 100 ,ugP/g CaC03. This is small compared to the
total sediment P concentrations (mostly >1000 jugP/g sediments)
reported by Green, si_al. (1978). Thus, if calcite is to make a
significant contribution to the P sorption capacity of sediments,
it must certainly do so via a' different mechanism than was
operative in these studies.
18
-------
In assessing the impact of sediment-bound phosphorus loads
on receiving waters, the kinetics of interactions between P and
particle surfaces must be considered. An investigation of
phosphate adsorption onto alumina and kaolinite revealed a rapid
adsorption step for 12-24 hr, followed by a slower, first-order
rate process extending over the next 60 days (Chen, si_sl.,
1973b). Kuo and Lotse (1974) characterized the kinetics of both
adsorption and desorption of phosphate by lake sediments.
Adsorption was described by Equation 4, which was developed from
the Freundlich equation.
X » K CQ t1/m (4)
where: X = Amount adsorbed (jagP/g sediment)
C = Initial phosphorus concentration (ppm)
t = Reaction time
and, K and 1/m are constants.
The same authors reported that, although desorption of native
sediment P was negligible except in the presence of NaOH or EDTA,
significant percentages of freshly adsorbed P were rapidly
desorbed in accordance with a kinetic expression similar to
Equation 4. This is contrary to the results of Li, si_^l.
(1972), who indicated that native P in lake bottom sediments and
freshly sorbed P showed roughly the same degree of
32
exchangeability with P, and that exchange kinetics could be
resolved into three first-order reactions. In this study, the
majority (mean = 76%) of exchangeable P was found to participate
in a rapid reaction, at an average rate of 34.7 hr . Atkinson,
32
Si_al. (1972) concluded that the rate of P exchange at the
surface of goethite was first-order with respect to the
concentration of a Fe(III)-phosphate surface complex, and
independent of the solution P concentration. In another study,
conducted on river sediments and soils, two different pools of
desorbable native P were identified - one exhibiting a rapid
release independent of desorption time and another obeying
first-order kinetics (Oloya and Logan, 1980) . A technique for
estimating the amount of phosphate that would be desorbed as the
solution P concentration approached zero as a function of
desorption period was presented by Barrow (1979) and applied to P
desorption from a soil.
The kinetics of phosphate interactions with calcium
carbonate have also been investigated. Phosphate adsorption onto
the surface of calcite was described by second-order kinetic
expressions in two separate studies (Kuo and Lotse, 1972; Griffin
19
-------
and Jurinak, 1974). Kuo and Lotse (1972) reported that about 80%
of phosphate adsorption by CaCOj was completed in 10 seconds.
The formation of a calcium phosphate precipitate on the surface
of calcite at high soluble P concentrations was found to conform
to first-order kinetics. The desorption process (studied in the
presence of an anion exchange resin) could be described by two
simultaneous first-order expressions. It was proposed that one
reaction involved a rapid (rate = 0.049 hr~ ) dissolution of
surface-nucleated calcium phosphate minerals, while the other
-4 -1
involved a slower (rate = 2.4 x 10 hr ) desorption of
phosphate from calcite surface sites (Griffin and Jurinak,
1974).
The studies of phosphate adsorption and desorption discussed
in this section may provide a better theoretical understanding of
the processes-affecting the accumulation of phosphorus by
sediments during transport and the biological availability of
sediment-bound P upon entering the receiving water. However/
much of this work was conducted under conditions substantially
different from the aquatic environments to which sediments are
typically exposed. Thus, it is not possible to obtain accurate
quantitative estimates of sediment P bioavailability based on
information acquired from these investigations. For example,
whereas minimal desorption of both native and freshly sorbed P
has been observed when Fe or Al oxide, soil or sediment particles
are placed in P-free ionic media (Kingston, jei_sl.» 1974; Ryden,
5i_5l./ 1977a; Oloya and Logan, 1980), sediments incubated in the
presence of an active P-starved algal culture are often found to
release over 20% of native sediment P (Cowen and Lee, 1976a;
Williams, si_.al.f 1980a; Dorich, 5i_3l., 1980). Recently,
techniques have been proposed for quantitatively characterizing
the bioavailability of sediment-bound P. The evolution of these
techniques is reviewed in the following sections.
CHEMICAL FRACTIONATION OF SEDIMENT PHOSPHORUS
The development of techniques for the characterization of
sediment-bound phosphorus can be largely attributed to
researchers in the field of soil science. A variety of chemical
extraction procedures have been used to estimate the availability
of soil phosphorus to agricultural crops. These procedures
usually remove P by one or more of the following processes
(Logan, 1978) :
1. Displacement of soluble P held in soil pores;
2. Exchange of f^PO, adsorbed to soil surfaces with another
anion;
3. Dissolution or hydrolysis of soil P complexes.
20
-------
Several extraction methods originally designed for use on
soils have been applied to lake and river sediments as well.
These include the Bray PI (0.03 M NH4F + 0.025 N HC1) extraction
(Bray and Kurtz, 1945), the procedure of Olsen, ei_3l- (1954)
(0.5 M NaHC03), and extraction with acid ammonium oxalate
(Saunders, 1965) . Applications of these techniques to aquatic
sediments have been reported by McCallister and Logan (1978) ,
Romkens and Nelson (1974), and Williams, £i_al. (1971a),
respectively, as well as others. However, sediment P
bioavailability has more commonly been estimated using the
procedure recommended by Chang and Jackson (1957), or a
modification thereof.
Chang and Jackson (1957) first proposed a scheme for the
fractionation of sediment phosphorus into discrete chemical
forms. They suggested that treatment of soils with an extraction
sequence of NH.F, NaOH, H^SO., citrate-dithionite reagent and
NH.F again would selectively remove aluminum phosphate, iron
phosphate, calcium phosphate, occluded iron phosphate and
occluded aluminum phosphate, respectively. The occluded
phosphates were thought to be chemically similar to the iron and
aluminum phosphates extracted by NH.F and NaOH, but protected
from removal by iron oxide coatings that were insoluble in these
extractants. The authors also indicated that the NaOH step often
removed considerable organic phosphorus. This fraction was
eliminated from the extract solution (prior to analysis) by
acidification (to induce flocculation) followed by
centrifugation.
The fractionation sequence proposed by Chang and Jackson
(1957) was based on the view that aluminum- and iron-bound
phosphate in soils was predominately in the form of discrete
crystalline phases such as variscite (A1PO.) and strengite
(FePO.). This view, referred to as the "discrete phosphate"
theory (Williams and Walker, 1969), has been questioned in papers
by Bache (1963 and 1964), Bauwin and Tyner (1957), and Hsu
(1964). Williams, 5i_5l. (1971a) indicated that no direct
observations of discrete Al of Fe phosphates had been recorded in
unfertilized soils or lake sediments. A more common belief today
(called the "dispersed phosphate" theory) is that much of the
inorganic P in soils consists of phosphate ions chemisorbed onto
reactive surfaces such as iron and aluminum oxides or hydrous
oxides and amorphous alumino-silicates ("non-occluded" P), or
retained within the matrices of these materials or various soil
minerals ("occluded" P) (Williams and Walker, 1969; Logan,
1978). A certain amount of soil or sediment P may also be
21
-------
present in discrete phosphate minerals such as apatite
("discrete" P) or in organic forms such as esters of phosphoric
acid ("organic" P) (Williams, si_sl./ 1971a) . Although the
"discrete phosphate" theory is no longer widely accepted, the
work of Chang and Jackson (1957) has served as a basis for
additional research.
Some problems with the Chang and Jackson (1957)
f ractionation scheme were pointed out and some modifications
suggested by Williams, si_sl. (1971b and 1971c) . The results of
extractions on noncalcareous lake bottom sediments indicated that
the P removed by NH.F was largely present in association with Pe
rather than Al components of the sediment (Williams,
1971c) . In addition, calcium fluoride, formed during NH4F
extraction of calcareous materials, was found to "resorb"
inorganic phosphorus released during the same extraction
(Williams, 5i_5l., 1971b) . The authors suggested, therefore,
that the NH,F extraction could not be used to distinguish between
Al- and Fe-bound P, and proposed that this step be eliminated.
However, they indicated that this would not eliminate the problem
of resorption because CaC03 can sorb inorganic phosphorus
released during extraction with NaOH. In fact, the P extracted
from calcareous sediments by citrate-dithionite-bicarbonate (CDB)
reagent was thought to consist largely of P resorbed by CaC03
during a preceding NaOH extraction step. As a result of this
study, in which five different f ractionation schemes were
evaluated, a sequence of NaOH, CDB, and HC1 was recommended. The
sum of the NaOH- and CDB-extractable P fractions was said to
provide an estimate of Fe- plus Al-bound P, while treatment with
HC1 removed P present as apatite.
In subsequent f ractionation studies on Lake Erie bottom
sediments, Williams, _e±_.al. (1976a and 1976b) reversed the order
of the NaOH and CDB steps, presumably in order to avoid
resorption of P onto CaC03. The sum of CDB-P and reactive
NaOH-extractable P (R-NaOH-P) was termed non-apatite inorganic P
(NAIP) and that removed in the HC1 step was labeled apatite P.
These authors reported that the extra P extracted by the NaOH
step in this procedure was seldom greater than 4% of the CDB-P
value. They suggested, therefore, that a scheme employing only
CDB and HC1 could provide an adequate separation of NAIP from
apatite P.
Recent studies aimed at determining which forms of
phosphorus are released by the various extraction reagents
(Williams, £i_sl. , 1980b; Hieltjes and Lijklema, 1980) have been
conducted using discrete phosphate minerals of known
22
-------
compositions. This work has revealed still another potential
problem with the common f ractionation schemes. While most
calcium-bound phosphates and apatite minerals were not
solubilized appreciably by treatment with NaOH or CDB, a large
portion of the P associated with hydroxy-apatite (Ca,- (PO*) ^OH)
was extracted by the CDB step. Thus, in some cases/ it may not
be possible to separate NAIP from apatite P in sediments.
In both soils and aquatic sediments, organic P may
constitute a significant fraction (typically 10-60%) of total
sediment P (Frink, 1969; Syers and Walker 1969; Sommers, 5i_sl.f
1972; McCallister and Logan, 1978) . Organic P in soils or
sediments is often measured using the method of Mehta, £i_sl.
(1954), which involves successive extractions with concentrated
HC1, 0.5 N NaOH at room temperature, and 0.5 N NaOH at 90° C.
The extract solutions are combined, and organic P calculated as
the difference between the measured total P and inorganic
(reactive) P concentrations. Sommers, 5i_sl. (1972) proposed a
similar sequence of extractions, but analyzed each extract
separately to gain insights into the forms of organic P present.
They suggested that HC1 would extract low molecular weight
organic P compounds that are similar to inorganic P in their
nature of chemical association with sediments, while the NaOH
fractions consist largely of ill-defined humic-associated P
complexes in sediments. The cold NaOH step removed most (60-86%)
of the total organic P in lake bottom sediments, while the HC1
step extracted only 6-17% (Sommers, 5i_sl., 1972).
Based on a review of the phosphorus f ractionation
literature, it appears that, for certain types of soils or
sediments, an extraction sequence of NaOH-CDB-HCl may provide a
meaningful separation of sediment P fractions. Sediments
containing little CaCOj or hydroxy-apatite should release mostly
P sorbed to surfaces of iron and aluminum oxides and hydrous
oxides and organic P during the NaOH step. The CDB extraction
should then remove mostly P occluded within the matrices of iron
and aluminum oxides and minerals, with the HC1 step solubilizing
predominately apatite P. Unfortunately, the mineral composition
of river suspended sediment or lake bottom sediment samples is
often complex and difficult to determine. Thus, the exact origin
of P removed by different steps in the f ractionation schemes is
frequently unknown. Data obtained from sequential extractions
may be useful, nevertheless, as an indicator of the strength with
which P is bound to the sediments. Since the sequence proceeds
from the most selective extractant to the least selective, the P
removed in the earlier steps of the procedure is likely to be the
most weakly bound to the sediments, and thus, the most
environmentally reactive (Logan, 1978) .
Analyses of the total and/or extractable concentrations of
23
-------
certain metals (especially Fe, Al, and Ca) associated with P in
sediments may be used to suggest the likely origin of extractable
P fractions. For example, Williams, si^al. (1971a) found a
strong statistical correlation between Fe and P extracted from
lake bottom sediments by CDB. Metals analyses on the sediments
might also indicate the likelihood of encountering problems such
as resorption onto CaCO, or release of P from hydroxy-apatite by
CDB during the fractionation scheme.
ESTIMATION OF SEDIMENT P BIOAVAILABILITY BY CHEMICAL METHODS
The original phosphorus fractionation schemes were used by
soil chemists to study the reactions of phosphorus fertilizers in
soils. More recently, chemical extractions have also been used
to estimate the bioavailability of phosphorus associated with
sediments in aquatic systems. Estimates have also been obtained
using other chemical techniques, such as equilibration of
sediments with a P-free solution or with an ion exchange resin,
32
exchange with P, and incubation with chloroform.
Armstrong, s±_al. (1971) suggested that the potential
mobility of sediment inorganic P could best be interpreted in
terms of the amounts of non-occluded, occluded, and apatite P
present. They considered interstitial inorganic P (extractable
by neutral salts such as NaCl or NH,C1) to be chemically mobile
and that present in occluded forms (solubilized by reductants
such as dithionite) or as apatite (HCl-extractable) to be
essentially immobile. Non-occluded inorganic P
(NaOH-extractable) was considered to be in equilibrium with
inorganic P in solution and thus was labeled "potentially
mobile".
In subsequent work on sediments transported by five major
tributaries to the Great Lakes, Armstrong, s£_.al. (1979)
estimated the availability of sediment P by two different
chemical methods - extraction with 0.1 N NaOH and equilibration
with an anion exchange resin (Dowex 1X-8). The mean estimates of
available phosphorus for the individual rivers ranged from 14 to
37% of the total particulate P by the NaOH extraction method and
from 7 to 17% by the resin method.
Armstrong, 5±_sl. (1979) also separated river sediments into
three size fractions (0.2-2 yum, 2-20/am, and >20>Um) and analyzed
each for total and "available" P. They found that_total,
resin-exchangeable, and NaOH-extractable phosphorus
concentrations (per gram of sediment dry weight) generally
increased as the particle size decreased. Since efforts to
reduce sediment loads to the Great Lakes (by methods such as
no-till farming) are likely to favor removal of larger particles
from runoff, this finding suggests that the fractional reduction
24
-------
in the available P load will probably be somewhat less than the
reduction in sediment load.
The available P concentrations of suspended sediments from
over 30 tributaries in the Lake Erie basin have been estimated by
Logan (1978) using the NaOH-CDB-HCl extraction sequence of
Williams, .ei_al. (1971b). In this study, R-NaOH-P was considered
to be rapidly available for algal uptake, while the R-NaOH + CDB
fraction was assumed to represent the total potentially available
sediment P. R-NaOH-P was found to be on the order of 30-40% of
total sediment inorganic P for Michigan and Ohio streams, but
averaged only 14% for New York streams. (R-NaOH + CDB)-P ranged
from an average of 42.5% of total sediment inorganic P for New
York streams to an average of 88.6% for eastern Ohio streams.
The soluble reactive component of NaOH-extractable
phosphorus (R-NaOH-P) was used by Baker (1982b) to estimate the
bioavailability of P associated with suspended sediments in over
400 samples from ten sampling sites in the Sandusky River basin.
The R-NaOH-P fraction made up a relatively constant percentage of
the total sediment phosphorus (T-Sed-P) for medium to high flow
conditions and suspended solids concentrations. R-NaOH-P was
also approximately the same at both mainstream and tributary
stations, with averages ranging from 20.8 to 24.9% of T-Sed-P for
the individual sampling sites. Although point sources of P to
the Sandusky River have been found to move rapidly into the
stream sediments (Baker, 1.982a) , no significant increases in
R-NaOH-P downstream of point sources was detected under the
sampling and analytical conditions employed by Baker (1982b).
Reddy (1980), however, reported sizeable increases in
NaOH-extractable P in Genesee River bottom sediments due to point
source discharges.
Wildung and Schmidt (1973) measured the release of P from
lake bottom sediments by two different dialysis techniques. One
method involved the equilibration of sediments, enclosed in
cellulose acetate dialysis tubing, with water. The other
involved placement of sediment and water in two different glass
half-cells separated by a membrane filter with a pore size of
0.45 urn. Comparison of results using the dialysis methods and
measurements of P release in the presence of an anion exchange
resin led the authors to conclude that the presence of a
continuous P sink increased the release of P from sediments
considerably. Another resin method was used by Huettl, £i_5l.
(1979) to predict the availability of phosphorus in simulated
runoff suspensions. A cation exchange resin was saturated with
hydroxy-aluminum and equilibrated with the sediment suspensions.
The phosphorus adsorbed was considered to be the algal-available
fraction, and ranged from 20 to 40% of the total inorganic P.
The availability of particulate phosphorus in Lake
Memphremagog (Quebec-Vermont) and five of its tributaries was
measured by treating unfiltered samples with chloroform and
25
-------
monitoring increases in soluble P (Peters/ 1981) . Chloroform
stops biological uptake by killing cells, but does not inhibit
enzymatic breakdown of organic P substrates. The average P
release for the individual tributaries was reported to be between
17 and 66% of the total particulate P. The availability of P
contained in three other fractions (soluble P, small molecular
weight P and large molecular weight P) was derived from
particulate P availability estimates using the relative
32
exchangeability of these P fractions for PO. (Peters, 1981) .
Using a similar technique, Carignan and Kalff (1979) found the
amount of isotopically exchangeable P in lake sediment to be a
good measure of the P available to macrophytes.
Using chemical extraction procedures, it is possible to
obtain relatively quick and reproducible estimates of sediment P
availability. However, there are also serious disadvantages
associated with the use of these methods (Logan, si_sl., 1979).
In most cases, no attempt is made to simulate the receiving water
conditions under which release of P from suspended sediments will
occur. Also, the chemical methods supply little or no
information regarding the rate at which sediment P becomes
available. In order to predict the amount of phosphorus that is
released for algal uptake before sediments settle out of the
water column, it is necessary to characterize the kinetics of
sediment P release under receiving water conditions. Another
drawback of the chemical methods for estimating sediment P
availability is that often only inorganic P is considered (Logan,
5i_3l.r 1979). However, the results of Armstrong, £i_3l. (1979)
indicate that organic P may constitute a sizeable fraction of the
total sediment P in river suspended sediments. Rodel, £t_al.
(1977) reported that certain organic P compounds sorbed to lake
sediments were released via hydrolysis reactions, although rates
were slower than for soluble organic P. Organic P may,
therefore, make a significant contribution to bioavailable
sediment P. These difficulties can be overcome by the use of
bioassay methods to measure the rate and extent of sediment P
bioavailability.
BIOASSAY MEASUREMENT OF SEDIMENT P AVAILABILITY
Sediment P bioassay studies conducted to date have generally
involved the culturing of algae in a suspension of river or lake
sediments where the sediments provide the only source of P to the
algae and all other nutrients are in excess supply. Following an
incubation period, either the algal phosphorus content or algal
biomass (employing a P:biomass ratio for conversion) is used to
estimate the amount of sediment P that has become available
(Logan, 5±_5l., 1979). Some investigators have applied one or
more chemical methods in conjunction with bioassay measurements
of sediment P availability in an effort to find a correlation
that would allow prediction of biossay results by a less
26
-------
time-consuming method.
One of the first bioassay measurements of phosphorus
availability in aquatic sediments was performed by Golterman,
5i_al. (1969) on bottom muds from two Dutch lakes. They
monitored the increase in cell number of Ssg&s&SSKSS incubated in
the presence of lake muds as the only source of P. The apparent
phosphorus release was then calculated by comparing the log
growth rates with those observed for known orthophosphate
concentrations. By this method, algal uptake of P was estimated
at 7% of the total sediment P in one lake and about 29% in the
other. The authors also found that the decrease in the sum of
NaOH- and H2SO^-extractable P forms corresponded fairly well with
the algal P uptake calculated using a mean cell P content of
9.1xlO~8/igP/cell.
In a later study, Golterman (1976) found that the phosphorus
fraction removed from sediments by three successive extractions
with 0.01 M NaNTA was similar to that removed by S££Q£&£SIR£S • He
also found that the sequential CDB and NaOH extractions proposed
by Williams, £i_.al. (1976a) removed slightly more phosphorus than
the three NTA extractions. A later study of lake bottom
sediments conducted by Allan and Williams (1978) revealed a good
correlation between the CDB and NTA extractions. As a result,
the authors suggested that the P fraction extracted by CDB alone
could be considered available.
Fitzgerald (1970) investigated the growth of 5sl£Da5iijjm and
£1.3dQ.ph.Q.r..a sp. exposed to lake bottom muds (in dialysis tubes)
as the only source of phosphorus using absorbance at 750 run as a
biomass indicator. Although significant algal growth was caused
by soluble P or sediments spiked with soluble P in dialysis
tubing, sediments alone resulted in no growth response. However,
Hegemann, 5i_5l. (1982) indicated that bioassays in which soil
samples were enclosed in dialysis tubing showed much lower P
release than those in which the soil was in direct contact with
the algal culture. It is possible that, since the contents of
the dialysis tubes were not continuously mixed in these studies,
diffusion of soluble P to the surrounding culture was slow,
providing an opportunity for released P to resorb to the
sediments.
The availability of particulate P in urban runoff from
Madison, Wisconsin was measured by Cowen and Lee (1976a) using a
bioassay procedure. The test alga, 5£lSU55iJTJjm £spj:i£oxuijiijm was
incubated for a period of 19-22 days in assay flasks containing
AAP medium minus P and runoff sediments. Available P was then
determined by comparing cell counts to a standard curve of 18-day
cell counts versus initial orthophosphate concentration.
Available P measured by this procedure ranged from 8 to 55% of
the total particulate P (TPP) , with a mean of 30%. The
27
-------
NaOH-extractable fraction, which ranged from 22 to 27%,
corresponded most closely with the bioassay results.
Exchangeable P (anion exchange resin-extractable), which ranged
from 13 to 17% of particulate P, slightly underestimated the
algal-available fraction, while an acid extraction, which removed
from 33 to 46%, overestimated bioavailability.
Cowen and Lee (1976b) also used the procedure described
above to measure the availability of P in sediments transported
to Lake Ontario by the Genesee River. They found that 6% or less
of the sediment P was available based on the growth of
SslSnasS-Xm* In this case, both the NaOH-extractable (11-28% of
TPP) and resin-extractable (6-31% of TPP) fractions overestimated
the bioassay results.
Williams, s±_sl. (1980a) cultured several forms of
sedimentary material from the Lower Great Lakes basin with
P-starved Sssus^ssmus gjjad.ri.ca.uda, and used a standard curve of
cell counts versus initial orthophosphate added to calculate the
availability of sediment P. The fractionation scheme outlined by
Williams, 5i_sl. (1976a) was also applied to the sediments. By
comparing the chemical and biological measurements, the authors
concluded that apatite P and organic P in the sediments did not
support algal growth, while from 38 to 83% of the nonapatite
inorganic P became available during 12-day incubations.
There are several problems associated with the use of algal
biomass changes, rather than actual P uptake, to assess sediment
P availability (Logan, £±_jSl., 1979; Hegemann, si_sl., 1982).
Most of the problems stem from the fact that algae are capable of
taking up P in excess of their immediate metabolic needs, which
can result in a highly variable Prbiomass stochiometry. In
addition, it is quite possible that algae would respond
differently in terms of growth rate and cell P quota to a high
initial spike of orthophosphate (as used in developing standard
curves) than to a gradual release (typical of a sediment P
source) of the same amount of P (Logan, £i_5l., 1979). Further
complications could arise if the P uptake and growth response in
the synthetic medium used for the standard curve and/or sediment
bioassays differs from the response under receiving water
conditions.
Sagher (1976) found evidence of inaccuracies associated with
the use of constant P:biomass ratios to interpret sediment P
bioassay data. He compared algal uptake calculated from
55lSD55iJrJJffl cell numbers with a measurement in which NaOH and HC1
extractions were used to monitor the decrease in total inorganic
P (assumed to be incorporated into organic forms within algal
cells) in the sediment-algae suspensions. The estimate of
sediment P availability obtained by the chemical method (37% of
sediment inorganic P) was much more realistic than that obtained
using cell numbers (91% of sediment inorganic P). The bioassay
results obtained by Sagher (1976) corresponded very closely with
28
-------
estimates of sediment P availability based on a resin exchange
procedure (Huettl, .e£_.al., 1979).
In bioassay experiments on sediments suspended in runoff
water, Dorich, si_5l. (1980) used a method similar to that of
Sagher (1976). They suspended sediments in a nutrient medium,
inoculated them with SelSES-S^ISR and incubated them for two
weeks. The amount of phosphorus assimilated by the algal cells
was then calculated by comparing the results of sequential
extractions (NH4F-NaOH-HCl) applied before and after incubation
to determine the decrease in sediment inorganic P. They reported
that an average of about 20% of the total sediment P was
immobilized by the algae. Dorich, £i_sl« (1980) also conducted
bioassay experiments using both sterilized and nonsterile
sediment samples in order to determine the role of native
microorganisms in the removal of P from sediments. They found
that the sterilization process did not greatly affect the
measured release of sediment P and suggested, therefore, that
bacteria, protozoa and microflora are not essential for releasing
P to algae. It should be noted, however, that the chemical
measurements of availability by Dorich, £i_.al. (1980) and Sagher
(1976) would not detect the mineralization and release of
sediment organic P, which could be mediated by these organisms.
/
The relative bioavailabilities of P from various sources
(including river water, Grand River suspended sediments,
shoreline bluff material and tertiary treated sewage effluent) to
the Great Lakes were estimated by Millard, £i_sl. (1979) using
lake column simulators. Bioavailability was measured by
monitoring the response of five different algal biomass
indicators to continuous inputs of the P sources to the columns,
and comparing these responses to that for an equivalent input of
orthophosphate. The authors indicated that P in the sewage
effluent was essentially all available for algal uptake, while
shoreline erosion (bluff) material contained negligible amounts
of available P. Difficulties encountered in interpreting the
results for Grand River suspended sediments again demonstrated
the problems associated with the use of algal biomass to estimate
the availability of P sources.
Most of the bioassay studies discussed above were conducted
using a single incubation period between 12 and 28 days in
duration and an indirect measurement of algal uptake of P.
Recently, bioassay methods have been developed which allow
periodic harvesting of the algae and direct measurement of algal
P uptake. The release of P from one acid and one neutral soil
was found to continue for up to 112 days when the algal culture
was replaced with a fresh inoculum every 14 days (Hegemann,
5i_5l.f 1982). In another study, Verhoff, s£_.al. (1978)
investigated the removal of phosphorus from Lake Erie tributary
sediments by the indigenous phytoplankton population. Their
procedure involved periodically decanting the water from the
29
-------
incubation tanks, leaving most of the river sediments at the
bottom of the tank. Following retention of an aliquot for total
P analysis and removal of the algae by centrifugation, the water
was returned to the tank and the incubation continued. A mass
balance performed on total phosphorus and total inorganic solids
allowed calculation of the rate of conversion to available P.
The authors concluded that a linear availability rate of 0.2-0.4%
of total sediment P per day could be expected. This slow rate of
release over long time periods suggests that the rate of P
release is more important than the ultimate bioavailability
(Logan/ jei_al., 1979). However, few reliable measurements of
sediment P release rates are presently available.
DePinto (1982) has recently proposed an experimental
apparatus, called a Dual Culture Diffusion Apparatus (DCDA), for
investigating both the rate and extent of sediment P
bioavailability. The DCDA consists of two glass vessels that are
clamped together and separated by a polycarbonate membrane with a
0.4 ,0m pore size. By placing a sediment suspension in one vessel
and a P-starved algal culture in the other, mixing of the algae
and sediment particles is prevented while exchange of soluble
substances between the vessels is possible. Since the P-starved
algal culture maintains the soluble P level in the algal (or
"assay") vessel near zero, phosphorus released from the sediments
diffuses across the membrane where it is taken up by the algae.
The contents of the assay side are removed periodically for
analysis of the P uptake and are replaced with a fresh P-starved
culture. The kinetics of sediment P release are then quantified
by performing a mass balance on phosphorus on both sides of the
DCDA. This technique allows a reasonable reproduction of the
receiving water conditions under which P release from suspended
sediments actually takes place as well as a direct measurement of
algal uptake of released P. The procedure of DePinto (1982) was
used for all bioassays of algal-available P in this study.
30
-------
SECTION 5
EXPERIMENTAL METHODS
SAMPLE COLLECTION
In order to characterize the bioavailability of sediment P
inputs to the Lower Great Lakes, river water samples were
collected from five major tributaries to the lakes during storm
runoff events. The rivers selected (Maumee, Sandusky, Cuyahoga,
Cattaraugus (South Branch), and Genesee) represent a variety of
soil types (i.e. geochemistries and textures) and land use
conditions (Logan, 1978; U.S. Army Corps of Engineers, 1982).
The location of these tributaries within the Lower Lakes basin is
shown in Figure 2.
Sampling for the first year of this project was conducted on
all five rivers during the late winter and spring (March - June)
of 1980. Grab samples were taken at mid-depth in the rivers at
sampling stations located as close as possible to the mouth
without influence from the lake. The locations of the stations
are listed in Table 1 along with the USGS identification numbers
and the field personnel who conducted the sampling. All sampling
sites were located at or near USGS guaging stations, so a record
of flow at the time of sampling could be obtained.
During the second year of the study, a more diversified
sampling program was undertaken. The individual components of
the program were as follows:
1. The event-oriented sampling of spring 1980 (i.e. one
sample per storm event) was continued on the Maumee and
Cuyahoga Rivers during the spring of 1981;
2. River water samples were collected from the Sandusky
River at various times during each of four different
storm runoff events in the spring of 1981. Four to six
samples were taken per event in order to study trends in
available P discharge.
3. Samples were also collected from Honey Creek, a tributary
to the Sandusky, during high flow periods. The Honey
Creek watershed was the site of a three year (1979-1981)
pilot program to demonstrate and promote the use of
conservation tillage for the reduction of soil and
phosphorus losses. The present study should provide
31
-------
ONTARIO
Rochester
^Buffalo
Geoesee R.
[Cattaraugus
LAKE
ERIE
Clevelqn
Cuyahogq R.J
X <
00.
Haumee R.
Honey Cr.
dusky R.
Figure 2. Location of tributary sampling sites.
-------
TABLE 1. TRIBUTARY SAMPLING STATIONS AND
SAMPLE COLLECTORS
Tributary
Sampling
location
uses
sampling
station
I . D . no .
Sample
collector
Ohio Rivers:
Cuyahoga River
Maumee River
Sandusky River
Honey Creek
New York Rivers:
Cattaraugus Creek
(South Branch)
Genesee River
Independence, Ohio 04208000
Waterville, Ohio 04193490
Fremont, Ohio 04198000
Melmore, Ohio 04197100
Water Quality Lab
Heidelberg College
Tiffin, Ohio
Otto, New York
Rochester, New York
04213490
U.S. Army Corps of
Engineers
Buffalo, New York
William Prytherzh
Rochester Gas and
Electric Co.
Rochester, N.Y.
33
-------
background information on sediment P availability that
may be useful in assessing the effectiveness of such land
management practices at a future date.
4. Six longitudinally-spaced bottom sediment samples were
taken from the Sandusky River in August, 1981. Three of
the sites were upstream from the Bucyrus, Ohio, sewage
treatment plant, and three were downstream.
5. Suspended sediment samples were obtained from four
locations on a transect across the Detroit River north of
Grosse lie, Michigan.
6. Four bottom sediment samples were collected from the
western basin of Lake Erie - two near the Toledo, Ohio,
water intake and two near the Monroe, Michigan, water
intake.
7. Three shoreline erosion (bluff) samples were obtained -
two from Lake Erie (at Rondeau Park and Port Stanley,
Ontario) and one from Lake Superior (near Bardon Creek,
Ontario).
A summary of the locations and dates of collection for all
samples is presented in Table 2.
SAMPLE HANDLING AND PREPARATION
For each of the runoff event samples listed in Table 2, an 8
to 10 liter volume of river water was transferred on site to a
polyethylene cubitainer. The cubitainers were then cooled to
4 C as soon as possible, packed with frozen ice packs in large
insulated transportation containers, and shipped to Potsdam, NY,
via the quickest carrier service available. Upon receipt at
Clarkson College, the samples were stored at 4 C until
preparation and analysis.
The contents of the cubitainers were mixed thoroughly, and a
one liter aliquot of each sample was retained for physical and
chemical characterization of the unfiltered river water. The
remainder of the sample was filtered under positive pressure
through a 0.45 ,um pore diameter Nuclepore nitrocellulose
membrane. A one liter aliquot of the filtrate was also saved for
analysis. The non-filterable residue was removed from the
membrane using a stainless steel spatula, and resuspended in a
phosphorus-free, synthetic medium (see foonote in Table 6) at a
suspended solids concentration of about 10 g/1. This sediment
concentrate was stored at 4 C until its subsequent use in
particulate P fractionation and algal bioassay experiments.
Studies showed that the release of inorganic P from sediments
34
-------
TABLE 2. DATES AND LOCATIONS OP SAMPLE COLLECTION
U)
tn
Sample
number
River
(A) Tributary Samples - All
1
2
3
4
. 5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
Cuyahoga
Sandusky
Genesee
Cuyahoga
Maumee
Cattaraugus
Sandusky
Maumee
Cattaraugus
Cattaraugus
Sandusky
Cuyahoga
Maumee
Cuyahoga
Maumee
Sandusky
Maumee
Sandusky
Maumee
Sandusky
Maumee
Cattaraugus
Cattaraugus
Sandusky
Sandusky
Sandusky
Maumee
Date
collected
samples collected
3/8/80
3/9/80
3/20/80
3/22/80
3/20/80
3/23/80
3/22/80
3/25/80
4/6/80
4/14/80
4/15/80
4/15/80
4/15/80
5/28/80
. 5/28/80
5/27/80
5/20/80
5/19/80
6/3/80
6/5/80
6/10/80
6/20/80
6/20/80
6/2/80
6/9/80
4/14/81
4/14/81
Sample
number
at stations
28
29
30
31
32
33
34
35
36
37
38A
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
River
listed in TABLE 1.
Sandusky
Cuyahoga
Sandusky
Sandusky
Honey Creek
Honey Creek
Sandusky
Sandusky
Sandusky
Sandusky
Sandusky
Sandusky
Sandusky
Sandusky
Maumee
Maumee
Honey Creek
Sandusky
Sandusky
Sandusky
Sandusky
Honey Creek
Sandusky
Maumee
Sandusky
Sandusky
Date
collected
4/15/81
4/15/81
4/18/81
4/13/81
4/14/81
4/30/81
4/30/81
5/1/81
5/2/81
5/3/81
5/14/81
. 5/15/81
. 5/16/81
5/17/81
5/5/81
5/11/81
6/10/81
6/9/81
6/10/81
6/11/81
6/12/81
6/14/81
6/14/81
6/16/81
6/15/81
6/16/81
(continued)
-------
TABLE 2 (continued)
U)
Sample Name of
number sampling site
Direction Distance
from treat- from treat- Latitude
ment plant ment plant N
(B) Sandusky River Bottom Sediments - All samples
1
2
3
4
5
6
(C) Detroit
Locations:
Locustgrove
McCurdy
Weiss
Kestetter
Shupp
lOd
River Suspended
upstream 9
upstream
upstream
downstream
downstream
downstream
Sediments - Collected
Longitude
W
collected on 8/13/81.
miles
7
3
o;5
. 5
10
during
40°50'10"
40°49'50"
40°49'00"
40°48'30"
40°46I20"
40°44'30"
Summer, 1981.
82052'30"
82°54I00"
82°56'20"
83°00I20"
83°03'00"
83°06'30"
Four points along a transect of the river north of Grosse He, Michigan. Samples
were labeled as follows:
Fox Island
Trenton Channel
Livingston Channel
Amherst Channel
(D) Lake Erie Bottom Sediments
Sampling Locations
Toledo, Ohio Water Intake (2 samples)
Monroe, Michigan Water Intake (2 samples)
Date Collected
10/1/80
11/10/80
(E) Shoreline Erosion (Bluff) Sediments
Sampling Location
Rondeau Park, Ontario - Lake Erie
Port Stanley, Ontario - Lake Erie
2 miles west of Bardon Creek, Ontario -
Lake Superior
Date Collected
9/16/80
9/16/80
9/11/80
-------
stored at 4 C as a concentrate for prolonged periods (up to 6
months) averaged less than 2% of the total sediment P.
The Detroit River suspended sediments were collected by
personnel from the EPA Large Lakes Research Lab at Grosse lie,
Michigan, via pressure filtration of river water at the EPA lab.
Filter membranes (0.45 JJKI pore size) containing the sediments
were placed in plastic screw-cap bottles and refrigerated prior
to transport to Potsdam. Due to the low quantities of sediments
on the membranes, it was necessary to composite sediments from
all Detroit River sites into one sediment concentrate for
analysis.
The bottom sediments from the Sandusky River and Lake Erie,
and the shoreline erosion (bluff) samples were all grab samples
that were shipped to Clarkson College in insulated coolers.
Sediment concentrates were prepared by resuspending several
representative fine-grained portions of each sample in the P-free
media. Two sediment concentrates were made from each of the Lake
Erie bottom sediment and bluff samples to check the likelihood of
sample inhomogeneities.
ANALYTICAL PROGRAM
A summary of the analytical program conducted on tributary
samples is presented in Figure 3. Analyses performed on
unfiltered river water samples included total suspended solids
>0.45 ,um (SS), total phosphorus (TP) , and particle size
distribution, while filtrates were analyzed for total soluble
(i.e. filterable) P (TSP) and soluble reactive P (SRP). During
the first year of the project, suspended solids removed by a
20>um nylon mesh and pH were also determined on several of the
river water samples. In addition, a number of measurements of
alkalinity and soluble P bioavailability (by bioassay) were made
on the filtrates. During the second year, analyses of sedimept
particle sinking velocities were conducted on many of the ;
unfiltered tributary samples as part of a related research •
project. Total suspended solids (SS) analyses were performed on
the sediment concentrates to allow calculation of the volumes
necessary for chemical fractionation and algal bioassay studies.
Sediments were reconcentrated following bioassay experiments on
several of the 1981 samples, and subjected to the chemical
fractionation sequence. The total phosphorus (TP) concentration
of sediment concentrates was also measured. The ratio TP/SS for
sediment concentrates is equal to the total sediment phosphorus
(T-Sed-P) concentration in^ugP/g sediment. The procedures used
in performing the various physical, chemical, and biological
analyses mentioned above are listed, with the corresponding
references, in Table 3. Most of these are also described in more
detail in the next section.
For bottom sediment (L. Erie and Sandusky R.), bluff, and
37
-------
Tributary
Sample
V
Pressure filtra-
tion through
0.4 Sum membcane
Filtrate
Total Soluble P
Soluble Reac-
tive P
Alkalinity
Soluble P
Bioassay
V N/
Sediment
Concentrate
Suspended
Solids-
0 .4 5
Total P
Phosphorus
.Fractionation
Sequence
/\
Sediment P
Bioassay
Reconcentration
of Sediments
Unfiltered
Aliquot
Suspended
Solids -
0.45ym S 20um
fractions
Total P
Particle
Size Distri-
bution
Particle
Settling
Velocity
Figure 3. General flow chart for analysis
of tributary samples.
38
-------
TABLE 3. ANALYTICAL METHODS AND•REFERENCES
Analysis
A) Physical:
1) Suspended Solids
2) Particle Size Distribution
3) Particle Settling Velocity
B) Chemical:
1) Reactive Phosphorus
Preliminary Digestion
2) Alkalinity
3) pH
4) Sediment Phosphorus Frac-
tionation
C) Biological:
1) Sediment P Bioassay
2) Soluble P Bioassay
Method and Reference
Gravimetric; Drying at ^
103°C(APHA/ et al., 1976)
Coulter Counter (Bonner,
et al., 1983)
Spectrophotometric (Kwolek,
1982)
Ascorbic Acid Method
(Murphy and Riley, 1962)
Persulfate (APHA, et al.,
1976)
Acidimetric Titration(APHA,
et al., 1976)
Glass Electrode
NaOH-CDB-HCl (Logan, 1978)
Dual Culture Diffusion
Apparatus (DePinto, 1982)
Batch Incubation with Sele-
na strum (Young, et al.,
1982)
APHA, et al. (1976) is referred to as Standard Methods in the
text.
39
-------
Detroit River samples, only analyses appropriate to sediment
concentrates were applied (i.e. center branch of Figure 3).
Total P, suspended solids and P f ractionation were measured on
all of these samples, while bioassays were conducted on Lake Erie
bluff and bottom sediment samples only.
Suspended Solids —
Total suspended solids (SS) was measured by a procedure
similar to that described in 5Jt3J3.dajd_]ieJ;.h.od.s , Section 208D. A
known volume of sample was filtered through a preweighed 0.45
pore size Nuclepore nitrocellulose membrane (47 mm diameter) .
After drying at 103 °C, the weight of sediment retained on the
membrane was determined by weighing on a Mettler Model A30
top-loading electronic balance. The concentration of suspended
solids >20 jam was measured in the same manner, except that the
0.45/im membrane was replaced by a 47 mm diameter disc cut from a
sheet of Nitex nylon screening fabric with a pore size of 20 jm
(from Tetko, Inc., Elmsford, NY).
pH and Alkalinity —
Alkalinity and pH measurements were made using an Orion
Research Model 901 "microprocessor ionalyzer" with a glass
combination pH electrode. Sections 402 and 403 of
5±3B^a.r.d_JU£±iJ.ods. were used as guidelines for the alkalinity
analysis. A 50 ml aliquot of filtered river water was titrated
with a standardized acid (0.0099 N H2S04) solution to below pH
4. The endpoint of the titration was taken to be the inflection
point in the curve of pH versus volume of acid added, which
generally occurred at a pH of 4.8-5.0.
Reactive Phosphorus —
All reactive phosphorus concentrations were determined
colorimetrically using the ascorbic acid method proposed by
Murphy and Riley (1962) and modified slightly in 5iJ.J3daid
U.eiJ3i£d.s , Section 425F. Following addition of the mixed reagent
to samples, at least 10 min. was allowed for color development.
The sample was placed in a glass cell with a 10 cm path length
and absorbance was measured at 880 nm using a Bausch and Lomb
Spectronic 710 spectrophotometer . A set of standard solutions
ranging from 0 to 200yUgP/l were prepared and analyzed along with
each group of samples to provide a standard curve of reactive P
concentration versus absorbance. Deionized water was used as the
medium for all standards except those serving to calibrate the
analysis of CDB extraction solutions. The procedures used to
prepare CDB standards are discussed later in this section.
Preliminary digestion for total P analysis —
Samples analyzed for total phosphorus were subjected to a
persulfate digestion (.Stands j£L.Mfiih.o.ds , Section 425C.) prior to
40
-------
analysis for reactive P. Separate standard curves were developed
for the calculation of total P concentrations by digesting a set
of standards along with the samples.
Procedures —
The distributions of sediment particle sizes in the
unfiltered aliquots of river water were determined using a Model
ZBI Coulter Counter in series with a 100 channel pulse height
amplifier, or "Channelyzer" (both from Coulter Electronics/ Inc. ,
Hialeah, Florida) . The Channelyzer was interfaced to a Horizon
North Star microprocessor, which was used to transform
calibration and channel count information into real particle
number and volume distributions. Many of the procedures used for
particle sizing in this study were first reported by Rodgers
(1979) and have been automated by Bonner, £i_sl. (1983).
The sizing of particles with the Coulter Counter is based on
changes in resistance caused by the presence of particles in the
path of an electrical current passing through an electrolyte. A
flow of current is maintained between two electrodes, one
immersed in the sample and the other located in the aperture tube
into which the sample is drawn. As the sample containing the
particles passes through the aperture (a small opening at the
base of the tube) , voltage pulses that are proportional to the
volumes of individual particles are registered by the
Channelyzer.
In order to develop correlations between Coulter Counter
control settings and actual particle volumes, the instrument was
calibrated using counts performed on monodisperse suspensions of
polystyrene latex particles of known sizes. Calibration involves
the determination of a "threshold factor" that is required to
calculate the particle volume corresponding to each channel. By
performing counts with both 30,0m and 50 jm apertures, it was
possible to obtain accurate results for particles in an effective
diameter range of 0 .82-15.8 xum. At least three particle ranges
were used with each aperture to optimize the resolution of the
counts.
The instrument settings used for the various size ranges
examined in this study are summarized in Table 4. The volume of
individual particles associated with a particular channel is
given by the equation (Coulter Electronics, 1979) :
V = ((CN x WW/100) + BCT) x TF (5)
where V = volume of individual particles in a certain channel;
CN = the channel number (from 0 to 99) ;
WW = the window width setting;
41
-------
TABLE 4. COULTER COUNTER AND CHANNELYZER SETTINGS
Rangetf
30 urn
1
2
3
50 urn
£ 1
2
3
4
1
Amplification
aperture
1/4
1
2
aperture
2
2
8
32
1
Aperture
Current
1/2
1/2
1
1/2
2
2
2
Lower
Threshold
31.5
20.8
25.3
7
26
26
26
Upper
Threshold
79
97.9
101.9
104
105
105
105
Base
Threshold Channel
Factor Threshold
.009425
.0377
.1508
.3025
1.21
4.84
19.36
30
19
23.5
5
25
25
25
Window
Width
50
80
80
100
82
82
82
Diameter
(Mm)
0.
1.
1.
1.
3.
6.
9.
8188
1255
913
513
908
203
93 -
-
- 3
- 3
- 6
- 9
15
Range
1.1245
1.914
.085
.904
.244
.912
.768
-------
BCT = the base channel threshold setting;
and TF = the threshold factor determined by calibration with
standardized particles.
In this study, the effective particle diameter corresponding to
each channel was calculated from this volume by assuming a
spherical particle geometry.
The electrolyte in which the particles were suspended for
analysis was a 0.6% sodium chloride solution. A thickening
agent, polyvinylpyrollidone, was added to slow the passage of
particles through the aperture, thus providing greater accuracy
in the measurement of particle volumes. The composition of the
electrolyte solution and the preparation procedure are given in
Table 5. Samples were prepared for counting by diluting aliquots
of unfiltered river water with the electrolyte solution.
Inaccuracies may result in counting samples with high
particle concentrations due to the simultaneous passage of two or
more particles through the aperture. In addition, excessive
dilution of samples may result in a particle size distribution
that is not representative of the original sample. In this
study, dilutions were chosen to provide as high a particle
concentration as possible, without exceeding a maximum cumulative
count of 200,000 particles/ml in each diameter range. Counts
were performed on 0.1 ml aliquots of the appropriate sample
dilution. Depending on the reproducibility of cumulative
particle counts, between two and five replicate measurements were
made in each diameter range. These were averaged and used to
calculate the frequency distribution of sediment particle sizes.
Data Analysis—
The particle size distribution data for each tributary
sample was analyzed using both a power law distribution function
and a plot of cumulative particle volume versus particle
diameter. The size distribution of aquatic suspensions can often
be described by a mathematical relationship of the form (Lerman,
Sl., 1977; Lawler, 1979):
dN/d(dp) = A dp"6 (6)
where d(d ) = the particle diameter range being examined;
dN = the number of particles falling within that
diameter range;
A = a coefficient related to the total concentration of
particulate matter;
and 3 = the power law exponent.
43
-------
TABLE 5. ELECTROLYTE SOLUTION - 0.6% ISOTONE
(FROM AUTENRIETH, 1981)
. Compound
NaCl
Sodium azide
*
Polyvinylpyrrolidone-PVP- 40T
Na~HPO, • 12H-0
24 2
NaH2P04 • H20
NaOH
2/1
6.0
1.0
35.0
0.32
0.04
0.04
*
Technical grade from Sigma Chemical Co.,
St. Louis, Missouri.
Procedures -
1) Add the first three ingredients to 1 liter
of deionized water. Stir magnetically on
very low heat until all three ingredients
have been added.
2) Very slowly add the polyvinylpyrrolidone.
It should take 45 to 60 minutes to add this
compound (for a 1 liter batch).
3) When the polyvinylpyrrolidone has dissolved,
add the remaining ingredients. Stir until
completely dissolved.
4) Filter solution through a 0.2 um filter and
store in an acid-washed container.
44
-------
The power law exponent, g , is equal to the negative slope of a
log-log plot of dN/d(d ) versus d . In this study/ such plots
were used to obtain two values of 3 for each tributary sample,
one based on particle counts made with the 3Qjam aperture and the
other on counts with the 50 ^m aperture. The theoretical
development of the power law function and the significance of the
value of B are discussed in greater detail in Appendix A.
The Coulter Counter data was also used to generate plots of
cumulative particle volume versus particle diameter. From these
plots, the contribution of each diameter range to the total
volume of particles could be examined. Also, the plots were used
to evaluate the volume-weighted median particle diameter, i.e.
the diameter below (and above) which 50% of the total particle
volume is contained.
The sediment phosphorus f ractionation scheme employed in
this study was based on a procedure originally outlined by
Williams, si_5l. (1971b) involving sequential extractions with
NaOH, CDB, and HC1 reagents. This sequence has been used more
extensively and described in detail by Logan (1978) . A summary
of the procedure is shown in Figure 4. First, the volume of
sediment concentrate equivalent to 100 mg (dry weight) of
sediments was pipetted into a centrifuge tube and centrifuged at
900 x g for 20 min. in an International Equipment Co. Model UV
centrifuge with a swinging bucket rotor. The extractions were
then carried out as indicated in Figure 4. After both the NaOH
and CDB extraction steps, the sediments were resuspended and
rinsed with the same P-free media used to prepare the sediment
concentrates (see footnote in Table 6) .
During the second year of the project, the quantity of
sediments and volumes of extractants used were modified to
accomodate the use of a Beckman Model J2-21 centrifuge with a
fixed angle rotor. The following changes were made in the
procedure:
1) Dry weight of sediments - 60 mg
2) Volumes of extraction reagents:
a) 0.1 N NaOH - 37 ml
b) 1.1 M NaHC03 - 4 ml
c) 0.25 M Na-citrate - 33 ml
d) 1.0 N HC1 - 37 ml
3) Centrifugation - 31,000 x g for 30 min.
The extract solution from each step of the f ractionation was
decanted into an erlenmeyer flask (following centrifugation) for
storage prior to analysis. Aliquots of the NaOH and HC1 extracts
were neutralized to just below the phenolphthalein endpoint
45
-------
(0
4J
e
A
a,
0)
•H
4J
U
50ml 0.1N NaOH
25°C,. 17 hr.
Shaker table
Supernatant
Suoernanant
50ml l.ON HG1
1 hr., 25°C.
shaker table
Supernatant
Resuspended
Residue
Sediment
Concentrate -
0.1 grains sed.
Discard
Supernatant
5 ml 1.1M NaHCO- .x- ^
45ml 0 . 2.5M Na-citrate<^entrifugation
Ig Na2S204; 85°C, 15 mint*
water bath,
frequent mixing
P-free
media
Figure 4. Procedure for chemical fractionation of
sediment P.
46
-------
(pH 8.3) with 6 N R2S°4 and/or 2 N NaOH, and diluted by an
appropriate factor, before analysis of reactive phosphorus to
determine the R-NaOH-P and HC1-P fractions. Total phosphorus
analyses were performed on the NaOH extracts and the resuspended
sediment residue remaining at the end of the f ractionation
sequence (T-NaOH-P and Residual-P fractions) . Suitable volumes
of each were pipetted into erlenmeyer flasks, brought up to about
50 ml with deionized water, and digested via the persulfate
method. No adjustment in pH was made prior to the addition of
the digestion acid to these samples.
Before measuring the reactive P concentrations of CDB
extracts, it was first necessary to eliminate interferences
caused by both sodium citrate and sodium dithionite. This was
accomplished by following the procedures suggested by Weaver
(1974) . Excess dithionite was oxidized by bubbling moist air
through the CDB extract solutions for at least four hours. The
extracts were then brought up to a volume of 50 ml with deionized
water, and a 5 ml aliquot was removed for the reactive P
analysis. The citrate interference was eliminated by the
addition of 4.3 ml of 4% ammonium molybdate to the 5 ml of CDB
extract. This mixture was diluted to a final volume of 50 ml in
a volumetric flask and analyzed for reactive P by the ascorbic
acid method. Blank and standard solutions used to develop
standard curves for the analysis of CDB-P were prepared in a
medium containing 1 g of sodium dithionite and 25 ml of 0.25 M
sodium citrate per 50 ml of final solution volume. Interferences
were eliminated from the standards by bubbling and by the
addition of 3.75 ml of 4% ammonium molybdate to 5 ml aliquots of
solution before dilution to a 50 ml final volume.
Measurements of median particle settling velocities were
made for 28 unfiltered tributary samples using a
spectrophotometric technique. Absorbance readings were used as
an indicator of suspended solids concentration at a fixed height
in the sample cell. Measurements were made at a wavelength of
750 nm on a Perkin-Elmer 559-A OV/VIS spectrophotomet-er. Median
particle settling velocities were calculated based on the time
required for decays in absorbance to 20% and 50% of the initial
reading and the distance from the top of the sample cell to the
center of the light path. A detailed description of this
procedure and its applicability was presented by Kwolek (1982) .
The accuracy of the results appear to be adversely affected by
changes in the absorbance characteristics of sediments as a
function of particle size and by vibrations in the instrument.
Kwolek (1982) found that the measured settling velocities for
spherical glass particles with volume-weighted median particle
diameters of 3.78yum and 7.60 yum were 43% and 23% less,
respectively, than the velocities predicted by Stoke 's Law.
However, the procedure may be useful in identifying changes in
47
-------
the settling behavior of suspended river sediments during the
course of storm runoff events.
Procedures —
The bioavailability of sediment-bound phosphorus was
measured using the bioassay technique proposed by DePinto .
(1982) . A summary of the procedure is shown on the right-hand
side of Figure 5. The Dual Culture Diffusion Apparatus (DCDA)
reactor vessels were manufactured from 70 mm O.D., standard wall
Pyrex tubing. Half of these were painted black in order to
suppress photosynthetic activity by algae indigenous to the
tributaries that may have been present in the sediment vessels.
The sediment and assay vessels were separated by a 90 mm diameter
Nuclepore polycarbonate membrane with a 0. 4^*1111 pore size and were
held together by a hinged clamp made from two modified 57 mm
McCarter clamps. Rubber gaskets with a light coating of silicone
stopcock grease were used to provide a water-tight seal between
the glass flanges of the reactor vessels and the membrane. A
drawing of an assembled DCDA reactor is presented in Figure 6,
and a photograph is shown in Figure 7.
The assay organism used in this study was P-starved
Sslsn&S&XUR £3pli£5jrim£.umf a green alga, obtained from the
University of Texas at Austin collection (UTEX 1648) . Stock
cultures of Sslsnastxim were grown in deionized water spiked with
the necessary macro- and micro-nutrients. The concentrations of
nutrients added are listed in Table 6. This mixture is a
modification of the synthetic growth medium proposed by Guillard
and Lorenzen (1972), and contains 200 jigP/1. These cultures were
grown in 250 ml erlenmeyer flasks with cotton stoppers on a
shaker table under 300 ft-ca of continuous cool-white fluorescent
lighting. New stock cultures were started in autoclaved media
every 7 to 14 days. Algal cultures for the bioassays were grown
in Lake Erie water that was collected at the City of Buffalo
water treatment plant and filtered through a 0.2^um pore size
Nuclepore nitrocellulose membrane. This water was spiked with 20
jagP/1 (0.1 ml of stock solution 8 per liter) and all other
nutrients at the levels indicated in Table 6. The total
concentration of phosphorus in the spiked Lake Erie water was
typically 25-30 ^ug/1.
To prepare the algal growth media for inoculation, all of
the nutrients except FeCl3 and the tri-vitamin solution
(solutions #9 and #10 from Table 6) were first added to 8 1 of
filtered Lake Erie water. This was then filtered under positive
pressure through an autoclaved 0.2 jam pore size Nuclepore
nitrocellulose membrane and into an autoclaved 9 1 glass vessel.
Following addition of the iron and vitamins, the media was
inoculated with 8 ml of a stock culture of Sglsns-Stxum . The
48
-------
Tributary Sample
Note: Dashed lines
indicate repetitive
steps.
1 Filtration
through 0.4 Sum
membrane j
Sediment
Concentrate
Filtered
(0.2um) Lake
Erie water
Sediment
side con-
tents (lOOmg
Filtrate
per DCDA)
N03-N and
trace
nutrients
P-starved
Algal
Inoculum
Light incuba-
tion - ISL
reactors
Sediment
Vessel
(dark)
Assay
Vessel
(light)/
Termination
Harvest-drain
vessel
Q Harvest
i.45vm fil-
tration
Termination
Algae
Discarded
Reconcentration
of sediments
filtrate
Phosphorus
Fractionation
Sequence
Figure 5. Procedures for bioassay measurements of
phosphorus availability.
49
-------
O
COTTON-PLUGGED
GLASS TUBE
RUBBER
STOPPER
0 4«Ti NUCLEPORE MEMBRANE
BOROSILICATE
CORNING no. 770
REGENERATION VESSEL
/-MAGNETIC
/ STIRRER
Figure 6. Assembled DCDA reactor.
-------
Ln
i- •-
Figure 7. Photograph of Assembled DCDA Reactor,
-------
TABLE 6. COMPOSITION OF ALGAL GROWTH MEDIA
(MODIFIED FROM GUILLARD AND LORENZEN, 1972)
Stock
Concentration
Chemical Compound (M.W. ) (g compound/1)
1)
2)
3)
4)
5)
6)
7)
3)
9)
10)
CaCl2 • 2H20 (147.02)
MgS04 • 7H20 (246.365)
NaHC03 (83.9998)
Na2EDTA
Trace Metal Solution
CuS04 (159.54)
ZnS04 • 7H20 (287.44)
CoCl2 • 6H20 (237.84)
MnS04 • H20 (168.998)
Na2Mo04 • 2H20 (218.930)
H3B03 (64.81)
KN03 (101.096)
Na2HP04 -12H20 (357.953)
FeCl3 • 6H20 (270.206)
Tri-Vitamin Solution
Biotin
B12
Thiamine
36.755
36.970
12.600
4.36
0.0064
0.023
0.0119
0.152
0.0066
1.00
26.288
2.328
3.1605
0.0005
0.0005
0.10
ml of stock
added per
1 of media
1
1
1
2
1
1
1
1
1
1
- All stock solutions were made up in deionized water.
-' The P-free media used to make up sediment concentrates and
to rinse sediments during fractionations contained solutions
#1, 2, and 3 only.
52
-------
cultures were placed in an incubator at 20 °C with continuous
cool-white fluorescent lighting (500 ft-ca) and magnetic
stirring, and were utilized for bioassay experiments after a
growth period of between 5 and 12 days.
The suspensions placed in the sediment vessels of the DCDAs
were prepared by spiking filtered Lake Erie water with nutrient
solutions #4, 5,6,7 and 9 from Table 6 (to ensure P-limited
growth throughout the bioassays) and adding an aliquot of
sediment concentrate. In most cases, these suspensions contained
100 mg of sediments per 400 ml of final volume, although larger
amounts were used in analyzing Cattaraugus Creek and Lake Erie
bluff sediments. Sediment suspensions were usually made up in
1.5 1 batches to allow triplicate DCDA bioassays and storage of
an aliquot for suspended solids and total P analyses.
Following assembly of the DCDAs, 400 ml aliquots of
well-mixed sediment suspension and 5slej335ij:ijm_£3pii£SJiiyi.um
assay culture were measured out in graduated cylinders and added
to the dark and light sides, respectively, of each reactor. A
rubber stopper with a cotton-filled glass tube through the center
was placed in the opening of both sediment and assay vessels.
The reactors were then placed on magnetic stir tables under
cool-white fluorescent lighting (300 ft-ca) at 20 °C and stirred
continuously with 31.8 mm (length) egg-shaped magnetic stir bars
placed in both vessels.
The DCDAs were sampled periodically by first working the
stir bar around the inside of the assay side to remove any
attached algae from the walls of the vessel, and then decanting
the entire assay side contents into an erlenmeyer flask. This
was replaced by 400 ail of a fresh P-starved assay culture and the
incubation was continued. The reactors were generally sampled at
3, 7, 14, and 21 days after startup of the bioassay run. Several
of the reactors were also sampled at 28 days. Upon termination
of the experiments, the contents of the sediment side were
decanted and retained for total P analysis as well. The release
of phosphorus from the sediments was monitored by measuring the
differences in total P concentrations of the assay cultures
before and after incubations in the DCDAs. This approach is
contingent upon the rapid diffusion of soluble phosphorus across
the membrane to the assay side for algal uptake. In a
preliminary experiment, diffusion was found to occur at an
average first-order rate of 2.065 day (base e). A detailed
description of this experiment and the results obtained are
presented in Appendix B.
Data Analysis—
Experimental observations have indicated that the uptake of
phosphorus by the assay algae in DCDA bioassays can be described
53
-------
mathematically by a first-order saturation function having the
following form:
Q2(t) = Pulfc [1 - exp(-krt)] (7)
where Q2(t) = tne total amount of P taken up by the algae in the
assay vessel (vessel £2) at time t (jugP/g
sediment);
Pult = the total amount of sediment P ultimately available
for algal uptake (yagP/g sediment);
and k = the first-order P release rate (day ), base e.
DePinto (1982) has derived a similar expression by writing mass
balance equations for phosphorus in the two DCDA vessels and
making appropriate assumptions. The derivation of Equation 7 is
presented in Appendix C. This mathematical development also
shows that, under certain conditions, the expression for algal
uptake of P on the assay side is identical to that for the
release of P from the sediments.
The bioassay experiments provided phosphorus release data in
the form of Q2(t) for various values of t. From this data,
values of Pult and k were calculated using the Thomas Method,
which is based on the similarity of two series functions. This
method involves the determination of the slope (b) and intercept
1/3
(a) of a linear regression of (t/Q2(t)) ' versus Q2(t). Pult
and k are then calculated using the equations (Thomas, 1950) :
kr = 6b/a (8)
Pult -
During the first year of this study, an additional 2 1 of
filtered river water was collected during pressure filtration of
14 of the tributary samples to allow bioassay analysis of the
availability of soluble (filterable) P. The bioassay technique
54
-------
used was similar to that described by Young/ .gJt.al. (1982) . A
summary of the procedure is shown on the left-hand side of Figure
5. Two 1 liter aliquots of filtered river water were placed in
1 liter glass bottles and spiked with nutrients #4, 5,6,7 and 9
to ensure P-limited algal growth. These bottles, or "batch
reactors", were inoculated with 3 or 4 ml of a concentrated
suspension of j?£leBas£jum_£apiic3j.n.u£.uffi, and placed on a shaker
table at 20 °C under fluorescent lighting. Concentrated inocula
were obtained by centrifuging P-starved assay cultures (described
earlier), discarding the supernatant and resuspending the algae
in about 20 ml of P-free media per liter of initial culture
volume. An aliquot of each inoculum was saved and analyzed to
determine the amount of particulate P added to the batch
reactors. The initial TSP and SRP concentrations of the filtered
river water were measured as a part of the routine analyses
performed on all tributary samples.
After 1 day of incubation, 100 ml of the batch reactor
contents was-removed and analyzed for particulate/ total soluble,
and soluble reactive P fractions. Subsequent samplings were
conducted at 3 to 5 day intervals thereafter, and involved the
withdrawal of 50 ml for particulate P and total soluble P
analyses. In some cases/ one of the samplings was accompanied by
removal of the algae via filtration through a 0.45>om Nuclepore
nitrocellulose membrane and addition of a fresh P-starved
inoculum to the reactors.
55
-------
SECTION 6
RESULTS AND DISCUSSION
ROUTINE ANALYSES ON UNFILTERED TRIBUTARY SAMPLES
The results of TP, TSP, and SRP analyses and suspended
solids (>0.45yum and >20^um) measurements performed on unfiltered
aliquots of the tributary samples are summarized in Table 7. Mean
values of tributary discharge/ pH, and filtrate alkalinity are
also presented. A listing of all individual measurements of
these parameters is given in Appendix D. For all of the
tributaries except the Genesee, the samples collected represent a
wide range of flow conditions. However, all samples were
collected during storm runoff events, so low (or baseline) flows
were not included. Total suspended solids (SS) concentrations
increased markedly with increases in flow for all of the
tributaries. As a result, total P concentrations generally
showed greater variability than soluble P levels. The mean
concentrations of soluble P in the Ohio tributaries were
considerably higher than in the New York tributaries. Therefore,
suspended sediments would be more likely to accumulate P by
adsorption during transport in the Ohio rivers.
SEDIMENT P FRACTIONATION RESULTS
The data obtained from the chemical extraction of tributary
suspended sediments are summarized in Table 8. A listing of
results obtained for the individual samples is presented in
Appendix D. The mean extractable fractions for each tributary
are expressed both in^xig of P extracted per gram of sediment and
as a percentage of total sediment P (T-Sed-P) . Two separate
measurements of the total P content of sediments are given in
Table 8. The first of these is based on TP and suspended solids
analyses performed on sediment concentrates (T-Sed-P = TP/SS) ,
while the second is calculated from the TP, TSP, and SS
concentrations of raw river water samples (i.e. (TP-TSP)/SS) .
Unless otherwise indicated, the former values are used as the
basis for the following discussion, as well as for all
calculations.
T-Sed-P concentrations showed little variation among samples
from a given tributary. The coefficients of variation ranged
from 6.2% for the Cuyahoga River to 16.5% for Honey Creek.
56
-------
TABLE 7. SUMMARY OF PHYSICAL AND CHEMICAL MEASUREMENTS ON UNFILTERED
TRIBUTARY SAMPLES AND FILTRATES
Flow*
Tributary (m /s)
PH
Filtrate
Alkalinity
(mg/£ as
CaCO3)
Suspended solids
Total % < 20 urn
Phosphorus Cones
TP TSP
SRP
Ul
Maumee River;
Mean 647.0(11)
C.V. (%)** 75.3
Sandusky River;
Mean 222.7(27)
C.V. (%) 83.1
Cuyahoga River;
Mean 108.6(5)
C.V. (%) 50.6
Cattaraugus Creek;
Mean 29.5(5)
C.V. (%) 75.3
Honey Creek:
Mean 51.5(4)
C.V. (%) 84.5
Genesee River;
n = 1
ft
7.88(7) 119.5(5) 310.5(11) 86.1(6) 501.2(11) 111.8(11)
3.6 26.6 78.6 23.7 58.9 15.5
7.84(7) 90.0(7) 701.8(27) 85.0(11) 851.6(27) 81.0(27)
3.8 48.9 124.7 13.6 103.0 28.3
7.56 (4)
1.9
7.80(5)
4.5
65.3(3)
4.0
256.2.(5)
82.3
67.8(5) 1707.3(3)
35.1 133.6
62.3(4) 412.2(5) 77.9(5)
13.0 44.2 61.3
62.8(3) 1090.7(3) 13.9(5)
21.8 138.2 19.2
892.2(4) 91.8(2) 1071.6(4) 96.3(4)
96.6 4.2 69.3 52.6
7.35
80.2
156.0
88.5
192.0
37.0
97.4 (11)
15.1
66.9(27)
31.0
63.3(5)
84.4
6.3(5)
49.2
79.8 (4)
58.2
18.3
* Values given are instantaneous flows at time of sample collection, except those for
the Cuyahoga River, which are average daily flows.
t Amount retained on a membrane filter with a pore size of 0.45 \im.
** Coefficient of variation (C.V.) = (Standard Deviation/Mean) x 100%.
tt Values in parentheses are the number of measurements made.
-------
TABLE 8. SUMMARY OF P FRACTIONATION RESULTS FOR TRIBUTARY SUSPENDED SEDIMENTS
Ul
CO
(TP-TSP)
Triburary T-SED-P SS
Maumee River (n=9) :
Mean 1,259 1,244
C.V. (%) 14.3 18.9
Sandusky River (n=25) :
Mean 1,188 1,193
C.V. (%) 11.7 13.6
Cuyahoga River (n=4) :
Mean 1,315 1,246
C.V. (%) 6.2 9.9
Cattaraugus Creek (n=5) :
Mean ' 637 588
C.V. (%) 11.0 12.0
Honey Creek (n=4) :
Mean 1,189 1,218
C.V. (%) 16.5 14.5
Genesee River (n=l) :
T-NaOH-P
403(32.0)*
32.4
449(37.7)
20.8
575(43.8)
13.4
76(11.9)
33.2
505(42.5)
19.1
Concentrations in ygP/g sediment
R-NaOH-P NR-NaOH-P CDB-P IICH-P
273(21.7)
20.9
298(25.1)
20.8
442(33.6)
21.4
46(7.2)
27.2
356 (30.0)
23.5
131 (10.4)
61.9
151(12.7)
31.1
133(10.1)
29.6
30(4.7)
46.5
149 (12.5)
11.3
238(18.9)
17.5
237(20.0)
21.0
280(21.3)
16.6
82(12.9)
22.7
203(17.1)
22.2
117(9.3)
32.5
77(6.5)
35.6
188(14.3).
6.3
305(47.9)
12.9
56(4.7)
50.7
Residual P
174 (13.8)
42.2
194 (16.3)
42.5
73(5.6)
63.9
48(7.5)
8.7
265 (22.3)
21.7
957
Detroit River (n»l)
1,424
t
994 240(25.1) 170(17.8) 69(7.2) 187(19.5) 273(28.5) 77(8.0)
596(41.9) 305(21.4) 291(20.4) 177(12.4) 323(22.7) 170(11.9)
* Values in parentheses are means expressed as a percentage of T-Sed-P.
t Sediments composited from four sampling locations.
-------
Samples were collected from three of the tributaries (Maumee,
Sandusky, and Cuyahoga) during both 1980 and 1981. The mean
T-Sed-P concentrations for 1981 samples slightly exceeded the
1980 values for all rivers where a comparison was possible.
However, greater differences in sediment P concentrations were
observed between tributaries. The coefficient of variation
(C.V.) calculated from the mean T-Sed-P levels for all
tributaries was 23.1%.
Suspended sediments from Cattaraugus Creek contained
markedly lower levels of P than those from any of the other
rivers. The T-Sed-P concentration measured for the only Genesee
River sample collected was well above the mean for Cattaraugus
Creek, but still considerably less than the Ohio tributary
averages. Differences in the total sediment P levels among the
Ohio rivers were relatively small. Cuyahoga River sediments did
contain slightly more P than the others, possibly due to a
greater degree of influence from urban sources. Point source
inputs of P were probably also responsible for the fairly high
levels of T-Sed-P found in the Detroit River composite sample.
In comparing the findings of this study and previous ones,
consideration should be given to the techniques used to determine
total sediment P concentrations and discrepancies that may exist
between the various methods. A summary of data obtained by other
researchers on the rivers sampled in this study is presented in
Table 9 for comparison to the results in Table 8. Logan (1978)
found considerable disagreement among different measurements of
total sediment P. Perchloric acid digestions on freeze-dried
sediments gave T-Sed-P results somewhat higher than persulfate
digestions on freeze-dried sediments, but still well below the
values calculated from the TP, TSP, and SS concentrations of raw
river water. In the present study, total sediment P
concentrations obtained by the latter method agreed reasonably
well with those derived from persulfate digestion of sediment
concentrates. T-Sed-P concentrations reported by Logan (1978),
Baker (1982a), and Armstrong si_.al. (1979) were mostly greater
than those obtained in this study, while those reported by Baker
(1982b) and Reddy (1980) were somewhat lower.
Flux-weighted mean T-Sed-P concentrations computed by Baker
(1982a) indicate that the phosphorus content of sediments
transported by a particular river may vary considerably from year
to year. As was mentioned previously, only small fluctuations
were seen between 1980 and 1981 samples in the present study. In
addition, T-Sed-P levels appeared to remain relatively constant
over a wide range of flow conditions and suspended solids
concentrations for any given tributary. Based on a large number
of samples from Honey Creek, however, Baker (1982a) demonstrated
an inverse relationship between T-Sed-P (i.e., PP/SS) and SS
concentrations.
59
-------
en
o
TABLE 9. RESULTS OF CHEMICAL ANALYSES ON TRIBUTARY
SEDIMENTS FROM PREVIOUS STUDIES
A) Logan (1978) - All values are expressed in ngP/g sed. Values in parentheses are
extractable fractions as a percent of perchloric T-Sed-P.
I "DQV/~»V»I<~*V--I*-I i
T-Sed-P* Persulfate ' Perchloric
River/Site n (PP/SS)
T-Sed-P
T-Sed-P R-NaOH-P
CDB-P
HC1-P Residual P
Maumee River
Waterville 3
All Sites 5
Sandusky River
Fremont 3
All Sites 7
Cattaraugus Creek
South Branch 1
All Sites 3
1792
1648
1448
1865
1129
737
1142
1079
1092
1011
614
640
1234
1235
1265
1195
724
693
334(27
315(25
389(30
362(30
76(10
61(8.
.1)
.5)
.8)
.3)
.5)
8)
423(34.
380(30.
302(23.
299(25.
131(18.
133(19.
3)
8)
9)
0)
1)
2)
157(12
147(11
109(8.
140(11
271(37
326(47
.7)
.9)
6)
.7)
.4)
.0)
91(7.
107(8.
113(8.
95(7.
69(9.
73(10
4)
7)
9)
9)
5)
.5)
Honey Creek
Melmore 4 1555
Cuyahoga River
Independence 1 7404
915
1195
400(33.5) 286(23.9) 52(4.4) 80(6.7)
1332
1446
179
108
* TP and TSP measured on raw river water using perchloric acid digestion.
I Based on analysis of freeze-dried sediment.
(continued)
-------
TABLE 9 (continued)
(Tl
B) Baker (1982b) -
River Site
Sandusky /Fremont
Honey Creek/Melmore
R-NaOH-P is given
previous column.
n
64
61
as a percent of the
T-Sed-P*
(PP/SS)
1,077 ng/g
1,040
* TP and TSP measured on raw river water using persulfate
C) Baker (1982a) -
River Site
Maumee/Waterville
Honey Creek/Melmore
Sandusky /Fremont
Flux-weighted mean
1975
1660
-
1520
total sediment P value
R-NaOH-P
(%)
23.0%
22.7
digestion.
in the
sediment P concentrations.
1976
1420
1990
1660
1977 1978
1590 2060
2210 2520
1940 1910
1979
-
1630
1570
* Based on TP and SRP analyses (i.e., PP=TP-SRP) . TP digestion was performed by
persulfate method.
D) Armstronget al. (1979) - Mean values (in ngP/g sediment) for >0.2 urn suspended
sediment fraction.
T-Sed-P , A*
River Site n (PP/SS)* R-NaOH-P HCA-P
Genesee*1 15 833 162(19.4) 284(34.1)
Maumee/Waterville 4 1398 469(33.5) 273(19.5)
Values in parentheses are R-NaOH-P as a percent of T-Sed-P.
* Based on TP, SRP and SS analyses on river water, using 0.2 ym as the cutoff between
dissolved and particulate material.
I Performed at 2000:1 solution: sediment ratio rather than the 500:1 used in this study
and by Logan (1978) .
** Performed immediately after the NaOH extraction.
It All samples were taken upstream of Rochester, NY.
-------
Comparing the means of the extractable fractions in Table 8,
T-NaOH-P was found to be the largest fraction for all of the Ohio
tributaries sampled, ranging from 32.0% of T-Sed-P for the Maumee
River to 43.8% for the Cuyahoga River samples. In the Ohio
rivers, the CDB-P fraction consistently accounted for about 20%
of T-Sed-P, while HC1-P was the smallest fraction for all of
those tributaries except the Cuyahoga. On the other hand, HC1-P
was the largest fraction for all of the samples obtained from New
York rivers. The T-NaOH-P and CDB-P fractions for the Genesee
River sample were only slightly less than HC1-P, while
Cattaraugus Creek sediments were found to contain roughly four
times as much HC1-P as either T-NaOH-P or CDB-P. The
concentrations (in^/ugP/g sed) of both T-NaOH-P and HC1-P in the
Detroit River composite sediments exceeded the mean levels for
all other tributaries. The magnitude of the Residual P fraction
showed considerable variability among the tributaries sampled,
ranging from the smallest of all fractions measured for the
Cuyahoga River to the second largest for Honey Creek.
Concentrations of most of the extractable fractions were
consistent among the samples for a given tributary. Residual-P
exhibited substantial fluctuations for samples from the Maumee,
Sandusky and Cuyahoga Rivers, while for the Cattaraugus Creek and
Honey Creek samples, this fraction was relatively constant.
Variations in the other fractions were generally limited to one
or two unusually high or low values. Increases in all fractions
were observed from 1980 to 1981 for samples from the Maumee and
Sandusky Rivers. This is most likely due to a reduction in
sediment loss during the extraction sequence, achieved through
the use of a high-speed centrifuge (Beckwith, 1982). Good
agreement was obtained between the two years for Cuyahoga River
samples. Again, as with total sediment P concentrations, greater
variability in the extractable P forms was seen between
tributaries than among samples from any one tributary. The
coefficients of variation for the individual fractions,
calculated from the means for all tributaries, were:
T-NaOH-P - 46.4%
R-NaOH-P - 47.6%
NR-NaOH-P - 60.0%
CDB-P - 31.4%
HC1-P - 58.0%
Residual P - 54.1%
Thus, it appears that temporal variations in the forms of
phosphorus 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 8 can provide
insights into the forms and mobility of phosphorus associated
62
-------
with suspended sediments from the different tributaries. The sum
of the R-NaOH-P and CDB-P fractions may be considered to
represent the concentration of non-apatite inorganic P (NAIP) in
the suspended sediments, while HC1-P provides an estimate of
apatite P. Thus, the Ohio tributaries contained high levels of
NAIP and only small amounts of apatite P, while Cattaraugus Creek
sediments were rich in apatite P and low in NAIP. The Genesee
River and Detroit River samples contained intermediate levels of
both NAIP and apatite P.
The R-NaOH-P component of NAIP is thought to be mostly
adsorbed to Fe and Al oxides near the surface of sediment
particles, while the CDB-P fraction is occluded within the
crystalline matrices of similar compounds. Logan 5i_3l. (1979)
used R-NaOH-P as a measure of short-term bioavailable sediment P
and NAIP (R-NaOH-P + CDB-P) as a measure of the total amount
potentially releaseable over a long period of time. Thus, based
on the percentages of T-Sed-P present as NAIP, the ultimate
bioavailability of sediment P for the seven tributaries studied
would be expected to fall in the order: Cuyahoga > Honey Creek >
Sandusky > Maumee > Genesee > Detroit > Cattaraugus. It is
interesting to note that the mean ratio of (R-NaOH-P)/(CDB-P) for
these rivers generally tended to increase as the mean NAIP
concentration increased (as shown in Table 10). R-NaOH-P exceeded
CDB-P for all of the Ohio rivers, while the reverse was true for
the New York rivers. A likely explanation for this observation
is that the sediments with the greatest capacity for accumulating
P from anthropogenic sources are those that can sorb P rapidly at
the particle surface, possibly due to the presence of high
surface area adsorbents such as ferric oxide gel complexes.
Sediments that fix P predominantly by the slower process of
incorporation into the matrix of the soil particle (i.e. as
CDB-P) exhibit a lower overall tendency to accumulate P. The
relative importance of these mechanisms is strongly influenced by
the geologic origin of the suspended sediments. In addition, the
trends seen in Table 10 are no doubt accentuated by the fact that
the Ohio tributary sediments are generally exposed to higher
levels of soluble P and greater applications of P as a
fertilizer.
The majority of the organic P associated with suspended
sediments should be released as non-reactive P during NaOH
extractions (NR-NaOH-P). This is calculated as the difference
between T-NaOH-P and R-NaOH-P and was highest in the Detroit
River sediments. Since the Detroit River water originates almost
exclusively from the water column of Lake Huron (via Lake St.
Clair), it is likely to contain significantly higher
concentrations of algal biomass than the other tributaries.
Significant quantities of organic solids may also be attributable
to municipal wastewater treatment plant discharges from the
extensive urban areas along the Detroit River. While the mean
NR-NaOH-P concentration (in^ugP/g sediments) varied widely among
the other rivers sampled, this fraction comprised a relatively
63
-------
TABLE 10.. MEAN COMPOSITION OF. THE NON-APATITE INORGANIC
P.(NAIP) FRACTION IN TRIBUTARY.SUSPENDED.
SEDIMENTS
Tributary
Maumee River
Sandusky River
Cuyahoga River
Honey Creek
Genesee River
Cattaraugus Creek
*
Detroit River
Mean. NAIP Mean ratio of
(=R-NaOH-P + CDB-P) (R-NaOH-P) / (CDB-P)
511 ugP/g sed.
535
721
559
357
128
482
(40.6%) *
(45.1)
(54.9)
(47.1)
(37.3)
(20.1)
(33.8)
1.147
1.257
1.575
1.754
0.909
0.561
1.723
* Values in parentheses are. NAIP expressed as a percentage of
T-Sed-P.
64
-------
constant percentage of T-NaOH-P (from 23.1% for Cuyahoga River
sediments to 39.4% for Cattaraugus Creek). So, there appears to
be a positive correlation between the amount of organic material
present in association with tributary suspended sediments and the
magnitude of J?o.£.h the NR-NaOH-P and R-NaOH-P fractions. This is
consistent with the findings of Williams si_al. (1980a) , who
reported a close correlation between non-apatite and organic
forms of P associated with a variety of sediments from the lower
Great Lakes basin. It is possible that orthophosphate is
adsorbed onto particulate organic matter by relatively weak,
physical forces and released in a reactive form during the NaOH
extraction (i.e. as R-NaOH-P) . Other P may be bound to the
organic substances by stronger chemical forces, and released in a
non-reactive form (NR-NaOH-P) . The R-NaOH-P fraction should be
more readily available for algal uptake than the NR-NaOH-P
fraction.
The chemical f ractionation results obtained in this study
may be compared to those reported by Logan (1978) for all
tributaries except the Genesee and Detroit Rivers. The mean
concentrations of R-NaOH-P and CDB-P (Table 8) were lower than
those of Logan (1978) (Table 9) for all rivers. If Logan's one
Cuyahoga River sample (with unusually high T-Sed-P) is eliminated
from the calculations, his R-NaOH-P and CDB-P values (in^ugP/g
sed) average 32.6% and 51.4% greater, respectively, than those
measured at the same sites in this study. This may be partly due
to the fact that the samples analyzed by Logan (1978) were
generally collected under conditions of lower flow and suspended
solids concentration than those in this study. As mentioned
previously, data from Baker (1982a) have shown an increasing
trend in T-Sed-P (PP/SS) with decreasing SS. When f ractionation
results are expressed as a precentage of T-Sed-P based on
analysis of raw river water, the R-NaOH-P and CDB-P
concentrations from the two studies are much more consistent. It
is also possible that the observed disparity in the chemical
fractions between the two studies arises from differences in
techniques for concentrating and storing the suspended
sediments. The f reeze-drying process used by Logan (1978) may
have caused sediment particles to fracture, making sediment P
more susceptible to removal by the extraction reagents. On the
other hand, sediments stored as a concentrate in P-free media may
release small amounts of P to solution. This would not be
accounted for in any of the chemical fractions as measured in
this study.
Closer agreement was obtained with the results of Logan
(1978) for the HC1-P and Residual P concentrations. The two
studies yielded roughly equivalent quantities of HC1-P for
sediments from the Cuyahoga River, Cattaraugus Creek and Honey
Creek, while the present study showed lower amounts for the
Maumee and Sandusky Rivers. The mean Residual P levels listed in
65
-------
Table 8 exceeded those of Logan (1978) for the Maumee River,
Sandusky River and Honey Creek, but were slightly less for
Cattaraugus Creek and the Cuyahoga River.
Analyses on samples from the Sandusky River and Honey Creek
by Baker (1982b) yielded mean R-NaOH-P fractions of 23.0% and
22.7% of T-Sed-P, respectively, while the present study showed
.25.1% and 30.0% in this fraction for the same sampling sites. It
should be noted, however, that the results of Baker (1982b) were
for samples collected during periods of higher flow and SS
concentration. Armstrong, £i_sl. (1979) performed extractions on
suspended sediments from the Genesee and Maumee Rivers using a
different procedure than applied in this study. The NaOH step
was conducted at a higher solutionrsediment ratio (2000:1 as
opposed to 500:1) and was directly followed by extraction with
HC1. This procedure yielded a much higher mean concentration of
R-NaOH-P for Maumee River sediments, both in^ugP/g sediment and
as a percentage of T-Sed-P, than was found in the present study.
In fact, the R-NaOH-P levels (in^ugP/g sediment) reported by
Armstrong si_.al. (1979) for this river averaged only slightly
less than the sum of R-NaOH-P and CDB-P listed in Table 8. While
R-NaOH-P for the Genesee River, in /igP/g sediment, is comparable
for the two studies, it constitutes a greater fraction of T-Sed-P
for the samples of Armstrong si_5l. (1979). As expected, the
HC1-P fraction proved to be considerably smaller in the present
study since this step was preceded by a CDB extraction.
BIOASSAY MEASUREMENTS OP PHOSPHORUS AVAILABILITY FOR TRIBUTARY
SAMPLES
The release of available P from river sediments in the DCDA
reactors was measured by monitoring P uptake by SslsnastXUK in
the assay vessel. Plots of cumulative P release versus time
obtained from bioassays on two of the samples (#19 and #20) are
shown in Figure 8. The form of these release curves is typical of
the results observed for the majority of samples analyzed. In
Figure 8, the error bars about each data point represent one
standard deviation for the triplicate bioassays, while the solid
lines illustrate the amount of sediment P release predicted by
the first-order equation:
Prel(t) - *uit[1 ~ exp(-krt)] (10)
where P , (t) = the amount of sediment P released at time t
sed) , and P -,t(t) and k are as defined in Equation 7.
The equivalency of Equations 7 and 10 is discussed in detail in
-------
Ok
(fl
0)
\
CL
La
a:
428
CO
<
ta
A Samp I« No. 19 CMaumQQ R.)
kr= 3.171 day"1 ; PU;J t = 294 ,ug P/g
228 -
128. —•
S Sample No. 28 CSondusky R.)
= 8.161 day'1 ; Pu|t = 174 jug P/g
I 1 ! I L 1
i a
2S
35
TIME
Figure 8. Cumulative release of sediment P versus
time measured in DCDA bioassays'on Samples
#19 (Maumee R.) and #20 (Sandusky R.).
67
-------
Appendix C. For samples 119 and #20, this equation tends to
underpredict the observed P release for approximately the first
five days of the DCDA experiments, and to overpredict the results
obtained for the remainder of the bioassays. Similar tendencies
were noticed for most of the other samples as well. The data
suggest that a single first-order mechanism may not be
responsible for all of the sediment P release. It is possible
that two separate components of available sediment P may be
released at different first-order rates. This possibility is
discussed in greater detail later in this section.
Values calculated for the first-order release coefficients,
Pult and kr/ by the Thomas Method are listed in Table 11 along
with means for all tributaries and for 1980 and 1981 samples.
Pult *s a^so 9iven as a percentage of the T-Sed-P concentration
measured on the sediments placed in the DCDAs. The order of mean
bioavailable sediment P concentrations for the tributaries was
Cuyahoga > Maumee > Honey Creek > Sandusky > Genesee >
Cattaraugus. The mean values of Puit ranged from a low of 38.8
,ugP/g sed (7.7% of T-Sed-P) for Cattaraugus Creek to a high of
449.2yugP/g sed (33.9% of T-Sed-P) for the Cuyahoga River. Mean
first-order release rates (k ) ranged from 0.131 day" for
Cattaraugus Creek to 0.264 day for the Genesee River.
The variability of the bioassay results was comparable to
that observed for the sediment P f ractionation data. That is,
while significant differences in Pult were seen among samples
from a certain tributary, greater variability was generally found
between tributaries. The coefficients of variation (C.V.)
calculated for Puit in the individual tributaries and for the
mean P ,. values from all six tributaries are listed in Table 12.
The C.V. for the tributary means exceeded that for all of the
individual tributaries except Cattaraugus Creek, which contained
unusually low levels of bioavailable sediment P. In addition,
differences between the mean ? concentrations for 1980 and
1981 samples (Table 11) for a particular river were small
compared to the differences between rivers. The k values were
much less variable than P , ., both among samples for a particular
tributary and between tributaries.
68
-------
TABLE 11. FIRST-ORDER RELEASE COEFFICIENTS CALCULATED FROM BIOASSAY DATA
BY THE THOMAS METHOD
CTl
Tributary and
sample no.
Maumee River:
5
8 -
13
15
17
19
21
27
41
42
50
Mean (1980)
Mean (1981)
Mean (All)
Sandusky River:
2
7
11
16
20
25
26
28
30
31
44
45
46
T-Sed-P*
in DCDAs
(ngP/g sed)
1192
1254
1126
1646
1379
1270
1074
1683
1301
1287
1174
1277
1361
1308
1253
1144
1136
1018
984
1028
1205
1314
1405
1373
972
1022
1081
Sediment P
released
in DCDAs
(^igP/g sed)
300.2
203.6
165.2
529.6
406.2
262.0
172.1
530.6
248.7
183.9
267.7
291.3
307.7
297.3
321.2
151.3
144.4
195.9
154.9
219.6
224.1
257.0
300.6
285.0
158.7
216.1
218.9
Ultimately
Bioavailable
Sediment P
pnlt (ygP/g»sed)
350.2
230.6
187.3
596.2
432.3
295.3
201.3
629.3
281.6
213.3
293.2
327.6
354.3
337.3
352.3
164.6
163.4
218.6
173.6
226.1
231.2
307.3
343.0
343.5
174.1
239.6
240.2
Pult as
a % of
T-Sed-P
29.4
18.4
16.6
36.2
31.3
23.3
18.7
37.4
21.6
16.6
25.0
24.8
25.1
25.0
28.1
14.4
14.4
21.5
17.6
22.0
19.2
23.4
24.4
25.0
17.9
23.4
22.2
Release
Rate, kr
(day'1)
0.229
0.144
0.204
0.082
0.101
0.177
0.195
0.244
0.201
0.221
0.203
0.162
0.217
0.182
0.209
0.103
0.207
0.173
0.165
0.105
0.140
0.191
0.160
0.067
0.221
0.225
0.226
(continued)
-------
TABLE 11 (continued)
Tributary and
sample no.
T-Sed-P
in DCDAs
(MgP/g sed)
Sediment P Ultimately
released Bioavailable
in DCDAs Sediment P
(ygP/g sed) Pnii- (ugP/g sed)
P , . as
ult
a % of
T-Sed-P
Release
Rate, kr
(day'1)
Sandusky River: (cont'd.)
47
49
51
52
Mean (1980)
Mean (1981)
Mean (All)
Cuyahoga River:
1
4
12
29
Mean (1980)
Mean (1981)
Mean (All)
Genesee River:
1109
1174
1092
1152
1094
1173
1145
1294
1334
1210
1419
1279
1419
1314
258.2
209.3
247.3
227.5
197.9
236.6
222.9
340.6
569.5
261.3
421.9
390.5
421.9
398.3
273.4
225.6
271.2
253.6
216.4
263.9
247.1
383.5
645.5
279.8
487.8
436.3
487.8
449.2
24.7
19.2
24.8
22.0
19.7
22.4
21.4
29.6
48.4
23.1
34.4
33.7
34.4
33.9
0.201
0.192
0.200
0.221
0.160
0.186
0.177
0.231
0.189
0.145
0.189
0.188
0.189
0.188
900
141.9
173.8
19.3
0.264
Cattaraugus Creek;
Mean
6
9
10
22
23
(All)
362
562
612
648
610
559
53.
0.
0.
24.
79.
31.
7
0
0
4
3
5
61.
0.
0.
27.
104.
38.
9
0
0
6
5
8
17
0
0
4
17
7
.1
.0
.0
.3
.1
.7
0
0
0
0
.172
-
—
.177
.043
.131
(continued)
-------
TABLE 11 (continued)
Tributary and
sample no.
Honey Creek:
32
33
43
48
Mean (All)
T-Sed-P*
in DCDAs
(ygP/g sed)
1337
1383
1094
978
1198
Sediment P
released
in DCDAs
(ugP/g sed)
297.9
279.0
262.6
211.3
262.7
Ultimately
Bioavailable
Sediment P
pult (iigP/g sed)
346.5
320.7
284.2
240.8
298.0
t> a o
Pult as
a % of
T-Sed-P
25.9
23.2
26.0
24.6
24.9
Release
Rate, kr
(day" )
0.224
0.212
0.199
0.223
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 12. COEFFICIENTS OF VARIATION FOR P , MEASURED
BY DCDA BIOASSAYS
Coefficient of variation for:
Pult in Pult as a
ygP/q sed • % of T-Sed-P
Values included
Mauraee River samples 45.5%
Sandusky River samples 25.0%
Cuyahoga River samples 34.7%
Cattaraugus Creek samples 115.2%
Honey Creek samples 15.4%
Means for all rivers 54.9%
30.4%
17.8%
31.7%
113.7%
5.3%
39.1%
72
-------
The bioassay and chemical f ractionation data collected in
this study indicate that many of the characteristics (e.g.
chemical form and biological availability) of sediment P
transported by storm runoff events do not vary substantially over
the course of a year or from year to year. Instead, these
characteristics appear to remain relatively constant for a
particular set of soil and land use conditions. This observation
has important implications regarding efforts to characterize the
bioavailability of sediment P inputs to a lake basin. In cases
where it is not possible to perform bioassay or chemical
f ractionation measurements on sediments from all streams in the
basin, it may be reasonable to assume similiar behavior for
sediment P transported by several streams within a large,
geographically distinct region.
The shape of the sediment P release curves observed in this
study is similar to that reported by other researchers using
bioassay techniques with repetitive sampling (e.g. Hegemann
5i_5l.r 1982; Verhoff, 5±_al., 1978). The release of P from
sediments occurred more rapidly than in previous studies
utilizing direct contact between algae and sediments. This is
most likely due to the fact that the bioassay reactors employed
by Hegemann si_al. (1982) and Verhoff .ei_al. (1978) did not
receive continuous mixing, while the sediments and algae in the
DCDAs did.
Although the DCDA bioassay procedure differs markedly from
most of the methods used in past investigations, some of the
results are quite comparable. For example, Verhoff £i_al. (1978)
found that about 20-25% of total P was removed by algae from
Honey Creek suspended sediments during 80-day incubations, while
Cattaraugus Creek sediments supported negligible algal growth.
Both of these findings are consistent with the bioassay results
of the present study. However, the mean P value of 21.4% of
T-Sed-P calculated from DCDA bioassay data on Sandusky River
sediments is considerably greater than the 6-7% release measured
by Verhoff 5i_3l. (1978) for this river. Dorich si_al. (1980)
reported an average release of 20.7% of T-Sed-P for storm runoff
sediments collected from Black Creek, a tributary to the Maumee
River. This agrees fairly well with the mean Pult value of 25.0%
obtained for the Maumee River in this study. Bioassays conducted
by Williams _ei_al. (1980a) on suspended sediments from Oakville
and Hillman Creeks in Ontario, Canada (tributaries to the Lower
Great Lakes) yielded estimates of 24% and 37% of T-Sed-P,
respectively, as available for algal uptake. Considering that
these values were obtained using incubation periods of only 12
days (Oakville Creek) and 18 days (Hillman Creek), they would
appear to fall at the high end of the range of bioavailabilities
observed in the present study. On the other hand, Cowen and Lee
(1976b) reported that less than 6% of particulate P in Genesee
73
-------
River water was available to SslsnastzlWt while 15.8% was
released in the DCDA bioassays and Pult was estimated at 19.3% of
T-Sed-P in this study. In general, the results presented in
Table 11 tend to agree more closely with those of previous
studies that utilized a direct mesurement of algal uptake of P
than those that relied upon cell counts to estimate P uptake.
As was mentioned previously, attempts to describe the
release of sediment P in the DCDAs by a single first-order
expression resulted in an underprediction of the observed results
early in the bioassays and overprediction later in the runs. In
order to achieve a better fit with the DCDA data, an
investigation was conducted into the possibility that two
separate components of sediment P were released at different
first-order rates. Logan s±_5l. (1979) have suggested that
R-NaOH-P may be released at a more rapid rate than
CDB-extractable sediment P. Such a release may be described
mathematically by the equation:
Prel(t) = Pr[l-exp(-krrt)]
Pg[(l-exp(-ksrt)J
(11)
where
and
P , P
the amount of sediment P released (in^ugP/g sed)
via the rapid and slow mechanisms, respectively
the first-order rates for the rapid and slow
release mechanisms, respectively (day~ ).
By substituting this expression into mass-balance equations for P
in both vessels of the DCDAs, an expression was also obtained for
the concurrent uptake of P by algae on the assay side of the
DCDAs as a funciton of time.
A computer algorithm was developed to calculate the four
release coefficients in Equation 11 based on optimization of the
correlation between observed and predicted values of P uptake.
Application of this algorithm to the bioassay results yielded
values of the rapid component release rate, krr/ in excess of 1.0
day" for roughly half of the samples. Since this rate is on the
same order as the bulk diffusion rate measured for the DCDA
reactors (a= 2.065 day" ), the accuracy with which krr can be
quantified by this method is questionable. However, precise
74
-------
estimation of krj. may not be essential, inasmuch as these
preliminary calculations indicate that the time required for the
release of most of the "rapid component" of sediment P is
considerably shorter than the expected residence time of the
suspended sediments in the water column of the receiving lake.
It was decided that, for all practical purposes, the rapid
component could be considered to be released from the sediments
essentially instantaneously upon entering the water column (or
the DCDAs). In this case, the release of sediment P can be
described by the equation:
Prel(t) = Pr + Ps[(l-exp(-ksrt)] (12)
Using a mass-balance approach, the following expression was
derived for the corresponding uptake of P on the assay side of
the DCDAs:
aPs kP
sr (13)
Pr[l-exp(-at)]
where Q^ft) = cumulative P uptake by agae on assay side (in
/igP/g sed) at time t.
The derivation of this equation is presented in Appendix E.
Again, a computer algorithm was developed to calculate values of
P ,P , and k based on the bioassay data. A listing of the
L S 5 U
BASIC computer program used is also presented in Appendix E. The
release coefficients obtained for each of the samples bioassayed
are presented in Table 13. Values are also tabulated for the
total amount of sediment P that would ultimately be released
(Pult = Pr + Pg) , along with mean values of all parameters for
the individual tributaries.
The mean value calculated for the rapidly released component
of sediment P (Pr) was considerably smaller than the mean of the
slowly released component (P ) for all of the rivers except the
S
Genesee. Although P comprised 95.2% of Pult for the one Genesee
75
-------
TABLE 13. TWO-COMPONENT RELEASE COEFFICIENTS (FOR EQUATION 12)
CALCULATED FROM DCDA BIOASSAY RESULTS
cr>
Tributary and
sample number
Maumee River
5
8
13
15
17
19
21
27
41
42
50
Mean
Sandusky River
7
11
16
20
25
26
28
30
31
44
45
46
Total ultimate
sediment P
release, Pult
(\igP/g sed)
301.2
206.6
176.2
599.6
457.2
273.0
177.1
562.6
252.7
186.9
301.7
317.7
162.3
156.4
198.9
162.9
270.6
259.1
263.0
308.6
332.0
175.7
250.1
267.9
Rapidly
released
component, Pr
(ngP/g sed)
117.6
66.4
44.2
1.2
2.1
67.5
55.3
267.8
51.3
66.5
110.5
77.3
29.4
20.0
13.3
56.7
9.0
33.7
140.2
154.9
0.9
67.7
81.8
104.6
Slowly
released
component , PS
(ygP/g sed)
183.6
140.2
132.0
598.4
455.1
205.5
121.8
294.8
201.4
120.4
191.2
240.4
132.9
136.4
185.6
106.2
261.6
225.4
122.8
153.7
331.1
108.0
168.3
163.3
Release rate
for slow
component, ksr
(day"1)
0.261
0.122
0.101
0.076
0.077
0.105
0.118
0.096
0.164
0.156
0.084
0.124
0.081
0.099
0.163
0.096
0.057
0.060
0.112
0.106
0.070
0.086
0.078
0.057
(continued)
-------
TABLE 13 (continued)
Tributary and
sample number
Total ultimate
sediment P
release, Pult
(ygP/g sed)
Cuyahoga River
4
12
29
Mean
573.5
278.3
433.9
428.6
Rapidly
released
component, P
(yiqP/g sed)
75.2
28.3
7.5
37.0
Slowly
released
component, Ps
sed)
498.3
250.0
426.4
391.6
Release rate
for slow
component, k
(day"1)
sr
Sandusky River
47
49
51
52
Mean
(cont'd. )
337.2
256.3
276.3
259.5
246.0
95.2
92.1
115.6
131.7
71.7
242.0
164.2
160.7
127.8
174.4
0.053
0.061
0.083
0.067
0.083
0.227
0.106
0.117
0.150
Genesee River
142.9
136.0
6.9
0.0917
Cattaraugus Creek
6 54.7
22 25.4
23 198.4
Mean 92.8
Honey Creek
43.9
10.0
0.1
18.0
10.8
15.4
198.3
74.8
0.119
0.0968
0.0177
0.078
32
33
43
48
Mean
299.9
281.0
284.6
213.3
269.7
49.1
37.0
75.9
21.6
45.9
250.8
244.0
208.7
191.7
223.8
0.196
0.196
0.106
0.214
0.178
-------
River sample, this component averaged only 8.6 to 29.1% of
for the other tributaries. The mean values of Pult obtained
using the two-component algorithm were less than the mean
calculated by the Thomas Method for the Maumee, Cuyahoga, and
Genesee Rivers and Honey Creek, but were greater for the Sandusky
River and Cattaraugus Creek. Also, as expected, the tributary
means for the slow component release rate (ksr) were all
substantially lower than the mean Thomas Method release rates (k
in Table 11) .
A correlation coefficient matrix containing all of the
one-component and two-component release parameters is presented
in Table 14. The values of Puit calculated by the two different
methods were closely correlated. Fluctuations in Pult appear to
be associated with changes in the slowly released component (P )
5
rather than the rapidly released component (P ) . However, the
overall rate of sediment P release is influenced by the size of
the rapidly released component, as evidenced by the fairly strong
correlation between Pr and kr. Although the slow component
release rate (k ) also showed a significant correlation with k ,
SL IT
the relationship was not as close as that between P and k
JL L 9
It should be noted that, since the sediment P release curves
obtained from bioassay experiments consisted of only four or five
data points, there is a good deal of uncertainty involved in the
calculation of the two-component release coefficients.
Nevertheless, there seems to be some potential for improving the
description of sediment P release using the two-component
approach. In an effort to gain further insights into the
mechanisms of sediment P release, the two-component release
coefficients were included in the analyses presented in the
following subsections.
The amount of particulate phosphorus becoming available for
algal uptake while river sediments are suspended in the water
column of the receiving lake depends primarily upon the relative
rates of P release and settling of particles from the water
column. The western basin of Lake Erie, for example, has a mean
depth of about 7.3 m and a mean hydraulic retention time of about
78
-------
TABLE 14.
-j
10
SIMPLE CORRELATION COEFFICIENT MATRIX FOR ONE-COMPONENT AND
TWO-COMPONENT SEDIMENT P RELEASE COEFFICIENTS
Pult
Pr Ps <=pr+ps)
P 1.0000* -0.2306 0.2147
r 0.0000 0.1637 0.1956
P 1.0000 0.9009
S 0.0000 0.0001
P , 1.0000
(=P +P ) 0.0000
k
sr
P ,.
lilt"
(one-comp. )
k
r
ksr
0.0129
0.9387
0.0830
0.6203
0.0891
0.5949
1.0000
0.0000
Pult
(one-comp. )
0.2754
0.0942
0.8433
0.0001
0.9693
0.0001
• 0.2674
0.1046
1.0000
0.0000
kr
0.5836
0.0001
-0.3886
0.0159
-0.1297
0.4378
0.4134
0.0099
0.0266
0.8704
1.0000
0.0000
* Top value is the simple correlation coefficient (r). Bottom value is a, the
probability of Type I error (i.e., rejection of a true null hypothesis, H i p=0) .
-------
1.5 months. Measurements performed in this study (presented
later in this section) indicate that sediment particles entering
the western basin from the Maumee River during high flow events
have a volume-weighted median particle diameter of approximately
3 yUm. Assuming a spherical particle geometry and a uniform
density of 2.5 g/cm , Stoke 's Law predicts that 50 percent of the
particle mass would settle into the bottom sediments within about
10 days. Although this is a rather crude estimate of sediment
removal rate, comparison with available phosphorus release rates
measured in this study indicates that tributary suspended
sediments will probably settle to the lake bottom or be
transported a considerable distance before all of the potentially
available phosphorus has been released. It seems essential,
therefore, that dynamic water quality models properly account for
the rate at which allochthonous particulate P is made available
in the water column of a lake.
Most deterministic nutrient-phytoplankton models currently
being applied to the Great Lakes do include a mechanism for the
conversion of allochthonous "unavailable" P to available P
(Thomann, 5i_sl., 1975; Bierman, £i_sl., 1980; DiToro and
Matystik, 1980; DiToro and Connolly, 1980; Rodgers and Salisbury,
1981) . However, particulate P inputs from tributary sources are
usually lumped together with P present in decomposing aquatic
organisms (e.g. phytoplankton, zooplankton) in the water column
to make up the total "unavailable" P pool. As illustrated in
Figure 8, the data collected in this study suggest that the rate
of release of available P from tributary suspended sediments may
be represented as a first-order function of the amount of
potentially available phosphorus remaining on the sediments. On
the other hand, in the models currently employed, the rate of
release is considered to be a first-order function of the total
particulate P remaining in the "unavailable" pool. No
distinction is made between the rate of conversion of
allochthonous "unavailable" P (primarily inorganic sediment P)
and the rate of conversion of autochthonous "unavailable" P
(primarily algal-bound P) in these models. The rate coefficients
used to characterize release from the "unavailable" P pool are
generally lower than those measured for tributary sediments in
this study (Table 11) .
A comparison of the two approaches to predicting the
cumulative P release from tributary sediments, applied to sample
119 (Maumee River) , is presented in Figure 9 along with the DCDA
data. Using the one-component release coefficients calculated by
the Thomas Method (Table 11) the predicted release of available P
versus time would be given by the equation:
Prel(t) = 295.3 [l-exp(-0.177t)] (14)
80
-------
609
CO
LJ
-------
where P^ = fc^e cumulative available P released from the
sediments (jugP/g sed) .
295.3 = ultimately available sediment P calculated for
Maumee River sample f 19 (jjgP/g sed) .
0.177 = available P release rate calculated for sample
#19 (day'1) .
t = incubation time of sediments in the lake water
column (days) .
Assuming a water temperature of 20 C in a highly productive
systems such as Lake Erie, the approach used by DiToro and
Connolly (1980) would result in the following equation for
predicting available P release from sample #19 sediments:
Prel(t) = 1286 [l-exp(-0.03t)] (15)
where 1286 = the total particulate P concentration of Maumee
River sample 119 sediments (jigP/g sed) (Table 8)
and 0.03 = the rate of conversion of "unavailable" P to
available P at 20 °C (day*1) .
These two equations provide similar predictions for the first 5-7
days. Thereafter, however, Equation 15 predictions exceed both
the observed data and the predictions of Equation 14 by a
considerable amount. Virtually all of the samples analyzed in
this study behaved in a manner similar to sample #19 when the two
methods for predicting sediment P release were compared. The
deviation between the two methods is primarily a function of the
difference between the amounts of total particulate P and
ultimately available sediment P.
The implications of the above analysis are quite obvious.
For systems in which the tributary sediments do not reside long
in the water column (less than about 7 days) , either Equation 14
or 15 would yield reasonable results. On the other hand, when
tributary sediments remain suspended in the water column for
longer than about 7 days, the current modeling approach (Equation
15) would tend to overestimate the input of soluble P to the
lake. For models calibrated primarily on the basis of algal
biomass, this would be likely to cause an overprediction of
, soluble P concentrations in the water column. In turn, this
might lead to either an overestimation of the phosphorus
retention time, or an unrealistic adjustment of other model
• coefficients in order to effect a calibration. It might also
lead to inaccurate conclusions about the relative contributions
of the various phosphorus sources to biological productivity in
the lake.
82
-------
As was mentioned previously, a third possible approach to
the prediciton of P release from tributary sediments is the use
of a two-component release expression such as Equation 12.
Substituting the release coefficients calculated for sample #19
(Table 13) into this equation yields:
Prel(t) = 67.5 + 205.5 [l-exp(-0.105t)] (16)
where 67.5 = the amount of rapidly released sediment P
calculated for Maumee River sample #19
(jugP/g sed) .
205.5 = the amount of slowly release sediment P
calculated for sample #19 (,ugP/g sed) .
0.105 = the first-order rate coefficient for the slowly
released component (day" ) .
A comparison between the predictions of this equation and
Equation 14 is presented in Figure 10 , superimposed on the
observed sediment P release data from bioassays on sample #19.
The predictions of the two-component equation show a slightly
better agreement with the observed data than those based on the
first-order function of ultimately available sediment P.
Equation 16 corresponds almost exactly to the observed P release
at times greater than about 10 days. However, as with Equation
14 f the predictions fall below the DCDA data during the early
portion of the incubations. It is possible that a more accurate
quantification of the two-component release coefficients could be
achieved if the DCDAs were sampled more frequently, especially
during the first week of the bioassays.
Although close agreement with the bioassay data is certainly
important, that alone is not necessarily justification for
adopting a particular model of sediment P release over others.
It is also desirable to establish a connection between the model
coefficients and the physical, chemical and biological
characteristics of the sediments. That is, evidence should exist
to indicate that the model provides a description of the actual
mechanisms of sediment P release. In this way, it may be
possible to estimate the model coefficients based on certain
characteristics of the sediments. An investigation into
relationships between the coefficients calculated for one- and
two-component sediment P release (Tables 11 and 13) and the
various chemical measurements performed on the same samples
(Table 8) is presented later in this section.
A sizable portion of the total phosphorus input to the Lower
Great Lakes via diffuse sources is in soluble rather than
83
-------
483
388 -
-------
particulate form. Based on loading data provided by the U.S.
Army Corps of Engineers, Buffalo District, Salisbury jei_al.
(1983) indicated that SRP comprised about 25% of the tributary
total P load to Lake Erie (excluding the Detroit River load) for
the years 1970-80. This component is generally considered to be
readily available for uptake by algae or other aquatic
organisms. Bioassays were conducted on several of the tributary
samples collected in this study to measure the availability of
soluble P forms to Ssl§£a.s£££R• The results are presented in
Table 15. Over 90% of SRP was found to be available within 24
hours for all samples except those from Cattaraugus Creek. The
algal inoculum rapidly reduced SRP to a nearly constant level
(usually from 2 to 5^ug/l) regardless of the initial
concentration present. This factor, in combination with an
unusually low initial soluble P content, accounts for the low
percentage utilized by the algae for Cattaraugus Creek samples.
TSP showed a slightly lower degree of availability than SRP
for most of the samples analyzed. This was expected, since the
TSP fraction includes non-reactive forms of phosphorus that may
not be utilized by algae as readily as soluble orthophosphate.
Although most of the TSP uptake occurred during the first 24
hours of incubation, several samples showed additional reductions
in this fraction at the second sampling (4 to 6 days of
incubation). Thus, it appears that a small portion of the
soluble P, although not immediately available, may be broken down
into an assimilable form within a few days following exposure to
an active P-starved algal population. The TSP concentration in
some of the bioassay reactors increased between the first and
second samplings. This occurred only when the algae were not
harvested (removed by filtering) after 24 hours, and may have
resulted from the excretion of organic compounds containing P by
the algal cells following a period of rapid growth and P uptake.
A third sampling at 12-14 days was included in several of the
soluble P bioassay experiments. However, this did not reveal any
further decreases in TSP. As with SRP, the algal inoculum
reduced TSP to a relatively constant level (4-S^ug/l) regardless
of the initial concentration present in the tributary sample.
Since the samples from the Ohio rivers contained high levels of
TSP, in most cases, 90% or more was found to be bioavailable.
Soluble P concentrations were lower for samples from the New York
rivers, and as a result, the fraction taken up by Sslenastxy.® was
also smaller.
85
-------
TABLE 15. BIOAVAILABILITY OF SOLUBLE PHOSPHORUS IN TRIBUTARY SAMPLES
00
a\
Tributary
Maumee River
Sandusky River
Cuyahoga River
Cattaraugus Creek
Genesee River
Sample
number
5
8
13
Mean
2
7
11
16
Mean
1
4
12
Mean
6
9
10
Mean
3
Initial cone.
-------
CORRELATIONS BETWEEN CHEMICAL FRACTIONATION AND BIOASSAY RESULTS
Simple product-moment correlation coefficients (r) were
determined between all chemically-extractable sediment P
fractions (Table 8) and the one-component P release coefficients,
P , t and k (Table 11) . The results are listed in Table 16 along
with values of a (the probability of rejecting a true null
hypothesis, HQ: p=0) . All of the chemical fractions except
Residual P were significantly correlated (a < 0.05) with
The highest correlation coefficient obtained was between Pult and
non-apatite inorganic P (NAIP = R-NaOH-P + CDB-P) . The only
parameter significantly correlated with k was Residual P,
although the relationship was not a strong one (r = 0.3480,
a = 0.0323) .
The high degree of correlation observed between T-Sed-P and
Pult kas limited applicability beyond the scope of this study.
Total phosphorus in suspended sediments from natural systems is,
typically, not a single, uniform substance. Rather, it consists
of several components that differ in their proportions and in
their relative availabilities. In particular, the results of
Table 16 indicate that the inorganic component of
base-extractable sediment phosphorus (i.e. R-NaOH-P) is more
closely linked to bioavailability than any other single chemical
fraction measured in this study. Other authors (Sagher, 1976;
Williams, £i_5l.r 1980a) have reported the same finding for
particulate matter from a variety of sources, although different
chemical extraction procedures were used.
In addition to R-NaOH-P, the CDB-P fraction, and to a lesser
extent, NR-NaOH-P also appear to contribute to bioavailable
sediment P. It is possible, however, that intercorrelations
between these fractions magnified the apparent relationships with
ultimately available P (Puj_t) • R-NaOH-P displayed a fairly
strong correlation with both CDB-P (r = 0.6285) and NR-NaOH-P
(r = 0.6111) .
The association between a chemical fraction and available P
may be examined independently by mathematically partialling out
the effects of these intercorrelations. This was done for the
three fractions under discussion and the results are given in
Table 17. The partial correlation coefficients between Pult and
the non-apatite inorganic fractions (R-NaOH-P and CDB-P) were
87
-------
TABLE 16. SIMPLE CORRELATION COEFFICIENTS
BETWEEN ONE-COMPONENT SEDIMENT P
RELEASE COEFFICIENTS AND CHEMICAL
FRACTIONS FOR TRIBUTARY SUSPENDED
SEDIMENTS
Chemical
fraction
One-component
release coefficients
ult
k
T-Sed-P
T-NaOH-P
R-NaOH-P
NR-NaOH-P
CDB-P
r
a
0.8006
0.0001
0.8084
0.0001
0.7790
0.0001
0.6487
0.0001
0.7229
0.0001
0.1446
0.3862
0.1610
0.3341
0.1639
0.3255
0.1040
0.5344
-0.0509
0.7679
HC1-P
Residual P
NAIP(R-NaOH-P
CDB-P)
•0.3160
0.0470
0.2182
0.1761
0.8317
0.0001
•0.1690
0.3104
0.3480
0.0323
0.0885
0.6079
88
-------
TABLE 17. PARTIAL CORRELATION COEFFICIENTS
BETWEEN BIOLOGICALLY'AVAILABLE P
AND CHEMICAL FRACTIONS FOR
TRIBUTARY SUSPENDED SEDIMENTS
Chemical
Fraction
One-Component
Release Coefficient
Pult
R-NaOH-P
NR-NaOH-P
CDB-P
0.7407
0.0001
0.5454
0.0005
0.7071
0.0001
Top number is partial correlation coefficient.
Bottom number is a, the probability of Type I
error (i.e., the rejection of a true null
hypothesis, HQ:p=0).
89
-------
only slightly less than the corresponding simple correlation
coefficients. The coefficient between NR-NaOH-P and PUI t/ on the
other hand, showed a more sizable decrease. Thus, R-NaOH-P and
CDB-P appear to make significant contributions to bioavailable
sediment P, while the contribution of the organic component of
base-extractable sediment P (i.e. NR-NaOH-P) may be of minor
importance.
The highly significant correlation coefficients seen in
Tables 16 and 17 indicate a strong potential for the development
of equations to predict the bioavailability parameters,
k , based on chemical fractions of sediment P. Linear regression
analyses were performed for several pairs of parameters, with
Pult or ^r as t^ie Dependent variable and the chemical fraction as
the independent variable. The regression coefficients obtained
are listed in Table 18. As an example, if R-NaOH-P were used to
predict ultimately bioavailable sediment P, the appropriate
equation would be:
Pult = °'8824 (R-NaOH-P) + 17.1 (17)
where P . , R-NaOH-P, and 17.1 all have units of ,ugP/g sediment.
The regression line corresponding to this equation is
superimposed on the tributary sediment data in Figure 11.
Comparisons between the regression equations and the observed
data are also presented for Puj_t versus NAIP in Figure 12 and for
k versus Residual P in Figure 13. From these plots, it is
apparent that the amount of sediment P that ultimately becomes
available may be predicted from f ractionation results with a much
greater degree of accuracy than the rate at which this phosphorus
is released.
Another possible means of predicting the bioavailability
parameters, Pu^t and kr, is using an equation based on a linear
combination of two or more of the chemical fractions measured.
Considering the close correlations between P •, t and both R-NaOH-P
and CDB-P, a regression analysis was performed to obtain
coefficients for a predictive model based on these two
fractions. This yielded the following equation:
90
-------
TABLE 18. SLOPES AND Y-INTERCEPTS FROM LINEAR
REGRESSION ANALYSES OF ONE-COMPONENT
SEDIMENT P RELEASE PARAMETERS VERSUS
CHEMICAL FRACTIONS
Independent
Variable
T-Sed-P
R-NaOH-P
CDB-P
Residual P
NAIP(=R-NaOH-P
+ CDB-P)
Dependent Variable
P * • ]f T
pult fr
Slope y-Int. Slope y-Int.
0.4383
0.8824
1.3622
0.6363
-238.3
17.1
-42.9
0.0001812 0.1566
-58.0
*
y-Intercept values have units of ugP/g sediment.
Values of slope are unitless.
t y-Intercept values have units of day . Values of
slope are in [day'1/(ugP/g sediment)].
91
-------
688
-------
683 -
-------
a.3
I
a a.2
TJ
v-/
LU
-------
Pult = °'6083 (R-NaOH-P) + 0.6880 (CDB-P) - 61.7 (18)
where, again, all terms have units of_,ugP/g sediment. The
correlation coefficient for this regression was r = 0.8316, which
is comparable with the highest value found for the regressions
involving a single chemical parameter (i.e. r = 0.8317 for Pult
versus NAIP) . From this analysis, it appears that correlation
equations based on both the R-NaOH-P and the CDB-P fractions,
either alone or summed, provide the most accurate predictions of
the ultimate availability of P associated with tributary
suspended sediments.
The rationale behind the investigation of a two-component
sediment P release model was based on the assertion by Logan
5i_5l. (1979) that R-NaOH-P represents the short-term available
P, while the sum (R-NaOH-P + CDB-P) represents the total sediment
P that ultimately becomes available. If this were the case, one
would expect a significant correlation between the rapidly
released component of bioavailable sediment P (Pr in Table 13)
and R-NaOH-P (Table 8) as well as between the slowly released
component (P_) and CDB-P. Simple correlation coefficients were
s
calculated between all of the two-component release coefficients
and the chemical fractions for the tributary suspended sediment
samples. The results are presented in Table 19.
The correlations observed between ?ult (= PC + PS) and the
chemical fractions followed trends similar to those between the
one-component parameter, P , ., and these same fractions.
Although the correlation coefficients were somewhat lower than in
Table 16, PUTI. nevertheless showed highly significant
correlations with all of the fractions listed except HCl-P and
Residual P. These correlations appear to be largely attributable
to close relationships between the slowly released component (P_)
s
and several of the chemical fractions. P was strongly
S
correlated with CDB-P (r = 0.6234), but it was even more closely
related to R-NaOH-P (r = 0.6798). The rapidly released component
of sediment P (Pr) was not significantly correlated with either
R-NaOH-P or CDB-P. Based on comparisons with the two-component
release coefficients calculated in this study, no evidence was
obtained to indicate that the R-NaOH-P and CDB-P fractions
95
-------
TABLE -19. SIMPLE CORRELATION COEFFICIENTS BETWEEN
TWO-COMPONENT SEDIMENT P RELEASE
COEFFICIENTS.AND.CHEMICAL.FRACTIONS FOR
TRIBUTARY SUSPENDED SEDIMENTS
Chemical
fraction
T-Sed-P
T-NaOH-P
R-NaOH-P
NR-NaOH-P
CDB-P
HC1-P
Residual P
NAIP
( =R-NaOH-P+CDB-P )
Two- component release coefficients
. Pr Ps pult-VEs ksr
r=0,3555
a=0.0334
0.2796
0.0986
0.1566
0,3616
0.4263
0.0095
0.2122
0.2140
-0.1137
0.5089
0.2551
0.1332
0.1978
0.2474
0.5581
0.0004
0..6404
0.0001
0.6798
0.0001
0.3690
0.0268
0.6234
0.0001
-0.0979
0.5701
0.0138
0.9364
0.7365
0.0001
0.6844
0.0001
0.7211
0.0001
0.6952
0.0001
0.5483
0.0005
0.6719
0.0001
-0.1458
0.3962
0.1403
0.4144
0.7673
0.0001
0.2790
0.0993
0.2037
0.2335
0.2615
0..1234
0.0308
0.8584
0.1801
0.2933
-0.0617
0.7209
0.0452
0.7936
0.2577
0.1291
96
-------
corresponded to rapidly released and slowly released components
of bioavailable sediment P, respectively. The correlation
coefficients in Table 19 reveal no significant relationship
between the release rate for the slow component (k__.) and any of
O £
the chemical fractions.
It should be pointed out that even a two-component release
model is still a considerable simplifiction of the actual
processes controlling sediment P availability. In reality,
several different chemical forms of phosphorus may be released at
a variety of rates. It is possible that R-NaOH-P and CDB-P were,
in fact, released at different rates in the bioassay experiments,
but that the sampling schedule and the model chosen did not
permit adequate resolution to detect this difference. Further
differentiation of bioavailable sediment P into various
components would probably require more frequent sampling and a
longer incubation period in the DCDAs.
CHANGES IN SEDIMENT P FRACTIONATION DURING DCDA BIOASSAYS
In order to further assess the contributions of the various
chemically extractable forms of sediment P to the release
observed during bioassays, sediments were reconcentrated after
several of the DCDA runs and subjected to the chemical
fractionation sequence. The extraction results are presented in
Table 20 for the 17 samples analyzed in this manner. Changes in
the chemical fractions during the bioassays are tabulated and
compared to the measured release of sediment P to Sslsn3.stxu.in. in
Table 21.
The most noticeable feature of the data in Table 21 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. This
relationship is shown graphically in Figure 14.
The mean values calculated for the changes in the other
fractions all showed average decreases of less than 20.AigP/g sed
(and less than 10% of the initial concentration), compared with a
mean decrease of 233 ^ugP/g sed for R-NaOH-P. In general, small
decreases in both NR-NaOH-P and HC1-P were observed during the
bioassays, although a few samples exhibited small increases in
one or the other of these fractions. On the other hand, ACDB-P
and Residual-P were much more variable. Both of these fractions
showed sizable increases for some samples and sizable decreases
for others, although on the average, the changes were small. The
occasional increases measured in the CDB-P and Residual P
97
-------
TABLE 20. POST-BIOASSAY SEDIMENT P FRACTIONATION RESULTS
CD
Tributary and
sample number
Maumee River
41*
42*"
50
Sandusky River
26
28
44
45
46
47
49
51
52
Honey Creek
32
33
43
48
Cuyahoga River
29
T-Sed-P
(pgP/g sed)
925
925
912
1,069
1,123
781
814
815
880
832
894
901
901
997
792
787
1,007
R-NaOH-P
102
102
85
128
116
63
57
64
81
76
85
87
118
122
100
86
113
Sediment P
NR-NaOH-P
114
114
108
121
135
103
99
110
127
124
133
117
135
140
104
118
109
fractions
CDB-P
131
131
219
192
223
216
216
220
216
259
240
245
127
131
223
180
231
(ljgP/g sed)
HCA-P
131
131
84
120
108
54
46
48
68
43
44
45
68
74
34
27
153
Residual P
278
278
198
406
446
202
196
204
214
176
177
185
288
302
205
159
228
* In order to provide enough sediment for duplicate chemical fractionations, sediments
from samples #41 and #42 were combined into the same sediment concentrate after
bioassays. Since the sediment chemistries of these two samples were similar, it is
felt that the post-DCDA fractionation results obtained are representative of both
samples.
-------
TAUI.E 2 1.
CHANGES IN CHEMICAL FRACTDONATION Of SEDIMENT P DURING DCDA
B10ASSAYS COMPARED WITH P RELEASE MEASURED IN DCDAs
Tributary and
sample number
Maumee River
41
42
50
Sandusky River
26
2B
44
45
46
47
49
51
52
Honey Creek
32
33
43
48
Cuyahoga River
29
Mean
(All samples)
AT-Sed-P
-280(23.2)
-333(26.4)
-237 (20.6)
-230(17.7)
-243(17.8)
-155(16.5)
-204 (20.0)
-240(22.7)
-292 (24.9)
-142 (14.6)
-184(17.1)
-199(18.1)
-445(33.0)
-370(27.0)
-262(24.8)
-201(20.3)
-419(29.4)
-261 (22.0)
Changes in extractable sediment
AR-NaOll-P ANR-NaOlI-P ACDB-P
-213(67
-210(67
-204 (70
-251 (66
-273(70
-145 (69
-220(79
-220(77
-249(75
-153(66
-193(69
-169(66
-308(72
-301(71
-222 (68
-168(66
-455 (80
-233(70
.6)
.3)
.6)
.2)
.1)
.7)
.4)
.5)
.4)
.8)
.4)
.0)
.3)
.0)
-9)
.1)
.1)
.8)
- 9(7.3)
- 2(1.7)
- 9(7.7)
+ 21 (21.0)
+ 7(5.5)
-25(19.5)
-17(14.6)
-12(9.8)
- 8(5.9)
+ 3(2.5)
+ 4(3.1)
-21(15.2)
-27(16.6)
-22(13.6)
-23(18.1)
-26(18.1)
0
-10(6.9)
-80(37.9)
-74(36.1)
-33(13.1)
-70(26.7)
-47 (17.4)
+96(80.0)
+36(20.0)
- 5(2.2)
-62(22.3)
+42(19.4)
+ 8(3.4)
+10(4.3)
-74(36.8)
-131(50.0)
+ 71 (46.7)
-18 (9.1)
+ 1 (0.4)
-19(4.6)
P fractions (pgP/g sed) *
AIIC1-P AResidual-P £ (AFrac
+ 8(6.5)
-64(32.8)
- 3(3.4)
-18(13.0)
-10(8.5)
- 4(6.9)
-11(19.3)
- 7(12.7)
- 3(4.2)
- K2.3)
- 4(8.3)
- 1(2.2)
- 8(10.5)
-11 (12.9)
- 3(8.1)
0
-27(15.0)
-10(9.0)
0
+ 3(1.1)
- 42(17.5)
+109(36.7)
+134 (42.9)
- 83(29.1)
- 99(33.5)
- 64(23.9)
+ 29(15.7)
- 20(10.2)
- 17(8.8)
- 17(8.4)
- 19(6.2)
- 3(0.9)
- 58(22.0)
- 25(13.6)
+126(123.5)
- 3(2.9)
-294(28
-347 (31
-290(29
-209(17
-189(15
-161(20
-311(33
-308(32
-293(29
-130(16
-202(22
-198(22
-436(37
-468(37
-235(26
-239(29
-355(29
-274(27
Sediment P 1st Order
release release
. measured rate
.) in DCDAs (days"1)
.0)
.5)
.5)
.8)
.5)
.2)
.6)
.3)
.3)
.1)
.9)
.6)
.2)
.8)
.1)
.5)
.9)
.0)
249
184
268
224
257
159
216
219
258
209
247
227
298
279
263
211
422
246
0.201
0.221
0.203
0.140
0.191
0.221
0.225
0.226
0.201
0. 192
0.200
0.221
0.224
0.212
0.199
0.223
0.189
0.205
* + indicates an increase during bioassays; - indicates a decrease. Values in parentheses represent the percent
change in that chemical fraction during bioassays.
I E(AFractions) = (AR-NaOH-P) + (ANR-NaOil-P) + (ACDU-P) + (AHC1-P) + (AResidual-P).
-------
583
483
TJ
0)
0)
CL
183
3
I f ~~
N. » 17 ; r = 3.98S8
——— 95 X Confidence Interval
79.
I I I II I I I I I I I I I I I I I ! I I I I
3
183 238 338 480 S88
-AR-NaOH-P Cjug P/g sedO
Figure 14.
Regression line of Puj_t versus the decrease
in R-NaOH-P during bioassays (-AR-NaOH-P)
superimposed on data for tributary suspended
sediments. •
100
-------
fractions indicate that a rearrangement of sediment P forms may
have occurred during bioassays on some samples. This is
consistent with the results of Verhoff .e±_al. (1978), who
reported large increases in CDB-P during longer bioassays.
Other authors (e.g. Logan, .e±_3l > 197 9 ; Allan and Williams,
1978). have suggested that all or part of the sediment P
extr actable by CDB reagents may be available for algal uptake.
Based on the decreases in CDB-P observed during bioassays for
several samples in this study, it appears that, in some cases,
this fraction may make a significant contribution to
algal-available P. However, this contribution tends to vary
considerably, as indicated by the lack of correlation between P
release in the DCDAs and ACDB-P (r = 0.238). In any event, the
vast majority of bioavailable sediment P released during 3-4 week
bioassays appears to originate from the R-NaOH-P fraction rather
than the CDB-P fraction. Thus, the strong correlations noted
earlier between R-NaOH-P and both Pult and Pg may be attributed
to a direct cause-effect relationship. The data provide no
evidence of such a mechanistic relationship between CDB-P and
sediment P bioavailability.
The overall first-order rate of available P release in the
DCDAs (kr in Table 11) was found to be closely correlated
(r = -0.6882, a = 0.0023) with the change in the Residual P
fraction during the bioassays. This result is consistent with
the significant correlation (Table 16) observed between kr and
Residual P measured before the bioassays. Release rate was even
more strongly correlated (r = -0.7975, cc= 0.0001) with the
change in the NR-NaOH-P fraction, although the relative magnitude
of this change was small for all samples analyzed. It seems
likely that the concentrations of Residual P and NR-NaOH-P were
influenced by the level of microbial activity in the DCDAs.
However, the proposal of a mechanism responsible for the observed
correlations would require additional information on the
f ractionation of P associated with various microorganisms at
different stages of growth and decomposition.
PARTICLE SIZE DISTRIBUTIONS AND SETTLING VELOCITY ANALYSES
Three types of plots were used to analyze the particle size
distribution data obtained for tributary suspended sediments in
this study. Plots of cumulative particle volume versus particle
diameter were generated for all tributary samples, and the volume
median particle diameter (D50v) was determined from these as
shown in Figure 15. For samples collected during the second year
of the study, cumulative particle number was also plotted against
101
-------
199
LU
ID
O
LU
CJ
H
< 40
-------
particle diameter (d ) and the number median particle diameter
(D5Q ) obtained in a similar manner. In addition, plots of
log (AN/ Ad ) versus log d (see Equation 6) were used to calculate
values of the power law exponent, 3 • An example of such a plot
is presented in Figure 16. Separate estimates of 8 were obtained
using counts made with 30 _/um and 50/um apertures (83Q and 85Q,
respectively) on the Coulter Counter. As seen in Figure 16 , no
particles were registered in some of the larger size ranges.
These data points were eliminated from the linear regression
analysis used to compute 6.
The sediment particle size parameters for all tributary
samples are listed in Table 22. Since the computer software used
to produce the power law distribution plots had not yet been
developed, 8 was not evaluated for samples #1-10. Also, the
values for samples #1-10 are probably overestimated in Table 22
since counts were made with a 50 ^um aperture only. An example of
the discrepancy to be expected is shown in Figure 17. For sample
#25, a volume median diameter of 2.70_yum was obtained using a 50
jum aperture only, while counts made with both 30^um and 50 /am
apertures yielded a D5Qv of 2.40^um.
The only particle size parameter available for comparison
among all tributaries is the volume median particle diameter,
D50v* Based on these values, samples taken from the two New York
rivers (Genesee and Cattaraugus) were found to contain larger
suspended sediment particles than those from all of the Ohio
rivers except the Cuyahoga. These results are consistent with
information on surface soil texture in the Lake Erie basin, which
indicates that, while the Cuyahoga River and Cattaraugus Creek
drain almost exclusively silty loam soils, the Maumee River and
Sandusky River watersheds contain extensive areas of finer,
clay-sized soils (U.S. Army Corps of Engineers, 1982). The
particle size data is also in agreement with the measurements of
suspended solids <20 Jim presented in Table 7, which also indicate
that Cuyahoga River sediments were substantially larger than
those from the other Ohio rivers. Coefficients of variation for
DCQV values from individual tributaries ranged from 16.9% for the
Cattaraugus Creek samples to 87.7% for the Cuyahoga River, while
the C.V. among the mean D5Qv values for all tributaries was
42.3%. All of the values obtained for the number median particle
volume, D/ fell within the narrow range of O.Se^um to 1.07
yum. The power law exponents also showed little variability, both
103
-------
8
Q.
"D
CD
O
0
B
SAMPLE * 48 (Honey Creek)
/330=4.320
I1 1 1 1
II I
aperture
50^im aperture
1 1 1 1
1 II ill 1 1 1 1 1
L l k
II
-0.2S
Q. 00
Q.2S Q . 5Q
Q . 75
1.00
| . 25
1 .50
LOG PARTICLE DIAMETER, dp C/jm)
Figure 16. Power law distribution plot - log (AN/Ad ) versus log d
for Sample j}48 P I
-------
TABLE 22.
MEDIAN PARTICLE DIAMETERS AND POWER LAW EXPONENTS
FOR TRIBUTARY SUSPENDED SEDIMENTS
O
U1
Tributary Sample #
Maumee River 5
8
13
15
17
19
21
27
41
42
50
Mean
Sandusky River 2
7
11
16
20
24
25
26
28
30
31
34
35
36
37
38
38A
39
Volume
Median
Particle
Diameter
3.15 jam
2.73
2.60
2.30
3.50
2.55
2.05
6.16
3.60
3.48
1.30
3.04
2.95
3.00
1.70
2.60
1.70
3.50
2.40
2.95
2.85
2.90
1.75
3.97
2.69
2.08
3.21
4.25
3.22
4.50
Number
Median
Particle
Diameter
—
-
-
-
-
-
-
0.87 vim
1.06
1.00
0.90
0.96
—
-
-
-
-
-
-
0.98
1.03
0.86
0.87
0.95
0.91
0.87
1.01
1.03
0.87
1.06
Power Law Exponent, 3
30 ym 50 vim
—
-
4.109
4.250
4.064
4.302
4.777
5.274
4.105
4.224
5.520
4.514
—
-
5.018
4.067
4.311
3.838
4.342
4.309
4.071
4J326
3.657
3.630
5.470
4.074
4.365
4.261
3.864
3.964
-
-
4.743
4.433
4.726
5.782
4.757
3.135
5.102
4.329
4.297
4.589
-
-
4.795
4.397
5.830
4.239
5.048
4.878
4.728
3.937
3.096
3.886
5.244
4.985
4.193
4.640
4.174
4.338
(continued)
-------
TABLE 22 (continued)
Tributary
Sandusky River
(cont'd)
Cuyahoga River
Genesee River
Cattaraugus Creek
Honey Creek
Sample #
40
44
45
46
47
49
51
52
Mean
1
4
12
29
Mean
3
6
9
10
Mean
32
33
43
48
Mean
Volume
Median
Particle
Diameter
4.50 yin
1.41
2.42
3.39
1.85
4.70
1.45
1.84
2.84
5.65
15.30
2.90
2.95
6.70
5.60
4.65
6.10
6.50
5.75
2.78
2.76
1.56
2.21
2.33
Number
Median
Particle
Diameter
1.05 vim
1.02
1.01
1.00
0.97
1.05
1.02
1.00
0.98
—
-
—
0.87
0.87
-
_
-
-
—
0.96
0.96
0.99
1.07
0.99
Power Law Exponent, 3
30 vim 50 ]im
3.980
4.503
4.638
4.840
5.189
4.208
5.231
5.307
4.394
_
-
4.013
5.280
4.646
-
_
-
-
-
4.908
4.703
5.231
4.320
4.790
4.525
5.051
4.896
4.046
5.213
5.150
5.288
5.128
4.654
_
-
4.618
4.935
4.776
-
_
-
-
-
4.819
5.095
5.650
5.179
5.186
-------
"D
V!
68
O
H
H-
C£
<
Q-
48
h-
O
28
U.
Q
8
a
a
a
SAMPLE. * 48 CHoney
30 jum and 50^am ap«rtur«s
50jum aperture only
I I ! I I I I I I I
1- 1 1
! 1 1
8
!8
28
PARTICLE DIAMETER, d
Figure 17. Comparison of cumulative volume distributions
for Sample #48 (Sandusky R.) with and without
30um aperature data.
107
-------
among samples for a given tributary and between tributaries,
although no figures were available for samples from the New York
rivers. 83Q and BJ-Q both averaged over 4.0 for all of the Ohio
tributaries. Thus, total particle number and surface area were
predominately associated with smaller particle diameters, while
particle volume was more evenly distributed among all size
ranges. However, even the distribution of particle volume was
skewed somewhat toward the smaller particle diameters. Despite
substantial disagreement between $30 and 85Q for some individual
samples (e.g. sample $27) , the mean values calculated for each
tributary differed by less than 10%. In all cases, the mean B^g
was slightly greater than the mean g3Q
In previous studies, Baker (1982a) found that sediment
phosphorus concentrations (PP/SS) decreased as tributary
suspended solids concentrations increased, while Armstrong £i_5l.
(1979) reported an inverse relationship between total sediment P
(in ugP/g sediment) and sediment particle size. Based on these
results, one might expect to find a shift toward higher
percentages of large particles during periods of high flow.
However, analyses conducted by Armstrong, 5i_sl. (1979) indicated
that particle size distributions remained relatively uniform over
wide ranges of total suspended solids concentration and flow for
a particular tributary. Since Coulter counts were performed on
26 samples from the Sandusky River in the present study, it was
possible to investigate correlations between the particle size
parameters in Table 22 and other variables such as flow, total
suspended solids concentration (TSS) , T-Sed-P, ?ult and release
rate (k ) . The simple correlation coefficients obtained from
analysis of Sandusky River samples are listed in Table 23.
Based on the samples collected in this study, TSS in the
Sandusky River was directly related to flow. Also, as reported
by Baker (1982a) , T-Sed-P showed a strong negative correlation
with TSS. However, trends in sediment particle size are more
difficult to interpret. Volume medium particle diameter,
was not significantly correlated with any of the parameters in
Table 23. On the other hand, the number median diameter,
displayed significant inverse correlations with T-Sed-P and
ultimately bioavailable sediment P (Pult) i as well as a direct
relationship with flow. It is likely that these correlations
were inflated somewhat by the narrowness of the range of
values, in conjunction with wider variations in the other
108
-------
TABLE 23. SIMPLE CORRELATION COEFFICIENTS BETWEEN
SEDIMENT PARTICLE SIZE PARAMETERS AND
VARIABLES RELATED TO THE FLUX AND
BIOAVAILABILITY OF SEDIMENT P FOR
SANDUSKY RIVER SAMPLES
D50v
D50n
S30
*50
Flow
T-Sed-P
Total SS
o
'ult
Flow
r = -0.1469
a= 0.4739
0.5330
0.0188
0.4822
0.0170
0.5084
0.0112
1.0000
0.0000
T-Sed-P
0.2451
0.2376
-0.5439
0.0161
-0.3895 '
0.0662
-0.6141
0.0019
-0.5511
0.0043
1.0000
0.0000
i
Total SS
-0.1337
0.5149
0.4046
0.0857
0.1936
0.3646
0.3936
0.0570
0.7042
0.0001
-0.7196
0.0001
1.0000
0.0000
Pult
0.1102
0.6737
-0.7344
0.0101
0.2379
0.3933
-0.6148
0.0147
-0.1086
0.6782
0.6124
0.0090
-0.3885
0.1233
1.0000
0.0000
kr
-0.0159
0.9517
0.7149
0.0134
0.6766
0.0056
0.4411
0.0998
0.5189
0.0328
-0.5008
0.0406
0.4386
0.0782
-0.0863
0.7418
109
-------
parameters. An example of this effect is given in Figure 18,
which is a plot of Dcnn versus flow, and is indicated by a nearly
horizontal regression line. Nevertheless, since particle number
and surface area are predominately associated with smaller
diameter particles, it is reasonable to expect that D-Q would be
more closely related than D-Q to the bioavailability parameters,
P ,. and k . Both of the power law exponents, B3Q and B5Q, were
also well correlated with flow. However, the correlation
coefficients were positive, characteristic of a shift toward
smaller particle sizes with increasing flows, rather than
negative as would be expected. Similar results were found for
the correlations between B5Q and both T-Sed-P and P , ,, where the
relationships were apparently close, but the signs of the
correlation coefficients were again opposite to what was
expected. Fluctuations in $30 and 650 values for the Sandusky
River samples were small. Coefficients of variation for these
two parameters were only 12.0% and 12.8%, respectively.
Therefore, correlation coefficients involving the power law
exponents may also have been enhanced by the relative uniformity
of these values.
Correlation analyses based on all Sandusky River samples
revealed no consistent relationships between sediment particle
size and flow, T-Sed-P, TSS, or sediment P bioavailability. The
possibility of such relationships was investigated further by
comparing trends in these parameters over the course of
individual storm runoff events. Volume median particle diameter
(D5Qv) and the power law exponent, &2Q' have been plotted against
sampling date along with flow and T-Sed-P for four 1981 runoff
events on the Sandusky River (Figure 19). Evidence of the
expected negative correlation between 63Q and flow was seen for
the first five samples taken during the June 9-16, 1981 period.
The correlation coefficient for these parameters was r = -0.9677,
although addition of data from the last two samples (June 15 and
16) lowered this considerably (to r = -0.2363). D5QV did not
show a consistent trend as a function of flow for this event, but
did for the period of May 14-17, 1981.
Based on Figure 19, it appears that over the course of some
individual storm runoff hydrographs, as flow increases, sediment
particle size may also increase while sediment P levels
decrease. However, these relationships are generally not the
same from one event to the next. For example, flows in the
Sandusky River during the period of June 9-16 were considerably
110
-------
1.6
5-'-'
l.Z
LJ
h-
LU
-------
^1388
^1288
-------
higher than for the three other storms monitored/ and, as
expected/ T-Sed-P concentrations were lower. * Ultimately
bioavailable sediment P/ Puitf displayed trends similar to
T-Sed-P, and was also lower for the June samples. But, 3 3Q
values measured for this event were noticeably higher/ indicating
a higher proportion of fine-grained particles compared to the
earlier storms. Thus/ other factors appear to exert a greater
effect than particle size on the total and bioavailable P content
of suspended tributary sediments. These might include the
residence time of sediments in the tributary system and the SRP
concentrations to which they are exposed during this period. In
addition, the distribution of particle sizes for suspended
sediments in a tributary is influenced by other factors besides
flow. The relative amounts of rainfall transported to the river
by surface flow and "interflow" can also be important/ as well as
the quantities and sizes of bottom sediments available for
resuspension during high flows. Both of these factors depend
upon the recent history of precipitation patterns in the river
basin. With respect to the 1981 Sandusky River samples/ it is
possible that/ during the low flows characteristic of the winter
months/ smaller sediment particles were resuspended from the
river bottom and transported downstream/ while coarse-sized
particles remained intact. Then, the first high flow events in
the spring capable of resuspending these larger sediments would
tend to exhibit a greater proportion of coarse-grained particles
than events later in the year. This could account for the lower
830 values obtained for high runoff periods in April and May of
1981. Also, the prolonged exposure of bottom sediments to the
overlying river water during winter low flow conditions may have
allowed an accumulation of phosphorus (by adsorption) , resulting
in higher levels of T-Sed-P and Pult for the early spring
events.
The median particle settling velocities measured for 28 of
the tributary samples are listed in Table 24 (from Kwolek, 1982).
These values were first calculated from the time required for a
reduction in absorbance (at 750 nm) to 50% of the initial
reading/ and are denoted by the symbol SVcg- However, a given
mass of coarse-grained particles in suspension will block less of
the incident light than the same mass of finer sediments. Thus,
a 50% decrease in absorbance corresponds to a larger decrease in
suspended solids concentration (SS) . The relationship between
absorbance and SS was investigated for sample 144 by monitoring
these two parameters at a fixed height in a large (8.6 cm
diameter x 145 cm height) settling column. The results/ shown in
Figure 20, indicate that a decrease of 50% in the mass of
113
-------
TABLE 24.
MEDIAN PARTICLE SETTLING VELOCITIES FOR TRIBUTARY
SUSPENDED SEDIMENTS
Tributary
Maumee River
Sandusky River
Sample
number
27
41
42
50
Mean
26
28
30
31
34
35
36
37
38
38A
39
40
44
45
46
47
49
51
52
Mean
Sampling
date
4/14/81
5/5/81
5/11/81
6/16/81
4/14/81
4/15/81
4/18/81
4/13/81
4/30/81
5/1/81
5/2/81
5/3/81
5/15/81
5/14/81
5/16/81
5/17/81
6/9/81
6/10/81
6/11/81
6/12/81
6/14/81
6/15/81
6/16/81
Median
Based on 50%
abs. decrease
(m/day)
1.36
0.24
0.30
0.16
0.51
0.17
0.22
0.43
0.40
0.25
0.25
0.18
0.26
0.63
0.34
0.22
0.31
0.15
0.10
0.10
0.13
0.23
0.18
0.14
0.25
particle settling
Based on 20%
abs. decrease DC
(m/day)
4.54
0.90
0.93
0.60
1.74
0.64
0.92
1.15
1.47
1.00
1.25
0.83
0.81
2.22
1.11
0.83
1.35
0.51
0.44
0.39
0.50
0.88
0.53
0.47
0.91
velocity
Calculated from
;QV an<^ Stoke's Law
(m/day)
2.66
0.91
0.85
0.12
1.13
0.61
0.57
0.59
0.21
1.11
0.51
0.30
0.72
1.27
0.73
1.42
1.42
0.14
0.41
0.81
0.24
1.55
0.15
0.24
0.68
(continued)
-------
TABLE 24 (continued)
Median particle settling velocity
Tributary
Cuyahoga River
Honey Creek
Sample
number
29
32
33
43
48
Mean
Sampling
date
4/15/81
4/14/81
4/30/81
5/10/81
6/14/81
Based on 50%
abs. decrease
(m/day)
0.60
0.17
0.21
0.08
0.14
0.15
Based on 20%
abs. decrease
(m/day)
3.15
0.68
1.03
0.31
0.54
0.64
Calculated from
D50v anc^ Stoke 's Law
(m/day)
0.61
0.54
0.53
0.17
0.34
0.39
Ul
-------
CTl
1.0
SUSPENDED SOLIDS CONC. (mgSS/l)
0 120 239 359 478 598
0
0
0.2
0.4 0.6
SS/SSo
0.8
£
c.
o
tr>
LlJ
O
z:
<
OQ
C£
O
10
CD
1.0
Figure 20
Relationship between normali2ed absorbance and normalized
suspended solids concentration for sediments from Sample #44
(Sandusky River), based on a settling column analysis.
-------
suspended solids at the sampling port resulted in only about a
20% decrease in absorbance. Although this percentage may vary
somewhat from sample to sample, for the purpose of this study, it
was assumed to be typical of all samples and a second series of
median settling velocities was calculated based on the time
required for a 20% decrease in absorbance (denoted by SV20). In
addition, another estimate of median particle settling velocity
was obtained from the measured volume median particle diameter
(DSQV) by applying Stoke's Law. For this calculation, it was
assumed that the sediment particles had a density of 2.5 g/cm
and settled in water at 20 °C. Both SV2Q and the Stoke's
velocity are also listed in Table 24.
The median velocities based on a 20% decrease in absorbance
were generally about 3.5 to 4 times as large as those based on a
50% decrease. For the majority of samples, the settling velocity
calculated from DJJQV using Stoke's Law fell between SVcQ and
^20.
An important application of settling velocity measurements
is in predicting the residence time of tributary suspended
sediments in the water column of the receiving water to which
they are discharged. In the numerical example presented
previously, a median residence period of 10 days was used in
comparing model predictions of sediment P release in the western
basin of Lake Erie (mean depth = 7.3 m). Although the median
settling velocities obtained in this study were quite variable,
0.73 m/day would be a typical value, indicating that a 10 day
residence time is a reasonable basis for model comparison.
Simple correlation coefficients between these median
settling velocities and the sediment particle size parameters
were calculated using the data from all samples, and are listed
in Table 25A. As anticipated, particle settling velocity showed
a fairly strong positive correlation with D5Qv and a significant
negative correlation to B5Q. S^CQ exhibited slightly stronger
correlations with these two parameters than did SV20* However,
the implications of this result are uncertain, since the two
settling velocities were found to be very closely related to one
another (r = 0.9634). The correlations between the Stoke's Law
settling velocity calculated from D5Qy and the velocities
measured by Kwolek (1982) were highly significant. Nevertheless,
117
-------
TABLE 25. SIMPLE CORRELATION COEFFICIENTS INVOLVING MEDIAN
PARTICLE SETTLING VELOCITIES
A)
B)
Correlations with Particle Size Parameters -
All Tributary Samples
Median Settling Velocity
Parameter
SV
50
SV
20
D50v
Stoke 's Law
Settling Velocity
D50n
830
B50
r = 0.6332
a = 0.0003
0.7205
0.00.01
-0.4137
0.0286
0.0727
0.7131
-0.5732
0.0014
0.6010
0.0007
0.6696
0.0001
-0.4366
0.0202
0.1160
0.5565
-0.4748
0.0107
Correlations with Sediment P Flux and Bioavailability
Parameters - Sandusky River Samples Only
Median Settling Velocity
Parameter
Flow
TSS
T-Sed-P
Pult
kr
SV50
r = -0.4691
a = 0.0427
-0.3836
0.1050
0.5031
0.0281
0.7957
0.0067
-0.7642
0.0062
SV20
-0.5010
0.0289
-0.4094
0.0817
0.4911
0.0327
0.7380
0.0095
-0.8492
0.0009
118
-------
the agreement between these values was somewhat disappointing.
The observed discrepancies were most likely due to the effects of
dispersion in the spectrophotometer cell in which settling
velocities were measured and to variations in the relationship
between suspended solids concentration and absorbance from sample
to sample. Neither DcQn nor B^Q displayed the expected
relationship with settling velocity.
Correlation coefficients between the median settling
velocities and flow, T-Sed-P, TSS, Pult and kr for the Sandusky
River samples were also determined. These values are presented
in Table 25B. The results followed essentially the same trend as
was found for correlations between these parameters and 3cn (see
Table 23). That is, although relationships appeared to exist
with some of .the parameters (flow, T-Sed-P, and P ,.), the
correlation coefficients were opposite in sign to what was
expected based on previous studies. One exception is the strong
inverse correlation between settling velocity and the first-order
release rate of bioavailable sediment P. Rapid settling
velocities would typically be displayed by larger particles.
Provided other factors, such as chemical composition and level of
biological activity, were relatively constant, these larger
particles should release sediment P more slowly than fine-grained
particles since they -have a smaller surface area (per unit weight
of sediment) at which the release reaction(s) may occur. The
significant positive correlation between Q^Q anc^ ^ (Table 23)
constitutes additional evidence that larger sediment particles
released P more slowly in this study, at least for samples from
the Sandusky River.
Trends in SV5Q over the course of individual runoff events
on the Sandusky River are shown in Figure 21, and compared with
changes in flow, TSS, and kr. These results are in agreement
with those in Figure 19. Although samples from the June 9-16
runoff event were taken at higher flows and contained higher TSS
concentrations than those from the earlier events, the suspended
sediment particles in the June samples were generally smaller and
settled more slowly. This is surely responsible for many of the
unexpected correlation coefficients seen in Tables 23 and 25. It
is possible that these correlations would change if data from a
larger number of Sandusky River samples were included. However,
the information presented here provides a strong indication that
the relationship between sediment particle size and tributary
flow may vary widely from one storm runoff event to the next, as
may the relationship between sediment particle size and sediment
phosphorus content.
119
-------
0 9-28
i
Q- O. I O
3 a,la
jt" a.as
2 3288
^2488
-1688
3 B-69
5 9-4S
g 8-38
£ a.is
c^T 4S8
w 388
1 193
u.
i t i • * i
1 t I 1
I II t
in
CS
u?
U7 U?
-------
LAKE ERIE BLUFF AND BOTTOM SEDIMENT SAMPLES
The erosion of shoreline (bluff) sediments and the
resuspension of lake bottom sediments during high wind periods
both may result in substantial inputs of particulate P into the
water column of Lake Erie (Sonzogni, .ei.sl., 1982; Lam and
Jaquet, 1980) . The biological availability of sediment P in
samples from these sources was investigated using chemical
fractionation and bioassay analyses. The results are presented
in Table 26. These data are especially interesting since they
encompass much wider extremes of sediment P fractionation and
bioavailability than do the tributary data.
The eroding bluff material contained extremely low levels of
NaOH- and CDB-extractable P. In addition, the T-Sed-P
concentrations of the Port Stanley 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 HC1 extraction step, and therefore, was most
likely present in the form of apatite P. By comparison, although
concentrations (in,ugP/g sediment) were similar for Cattaraugus
Creek sediments, HC1-P made up only 47.9% of T-Sed-P. In DCDA
bioassays on these 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 (e.g.
Williams, 5i_3l., 1976; Logan, 1978; Young, 5i_3l.f 1982) that
apatite P is virtually unavailable to aquatic organisms.
Previous studies conducted on recessional shoreline sediments
have also indicated that they contain very low levels of
bioavailable sediment P (Sonzogni, £i_sl., 1982).
The Monroe bottom sediment samples represent the opposite
extreme from the bluff sediments. Concentrations of T-Sed-P,
T-NaOH-P, and CDB-P in the Monroe 12 sample were all more than
twice as large as the highest values observed for tributary
suspended sediments. Bioassays showed that P , , for both Monroe
samples was also over twice the largest value measured for
tributary sediments. The percentage of T-Sed-P available to
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.
Lake bottom sediments consist of a combination of
allochthonous and autochthonous particulates which have, during
and after deposition, undergone a variety of physical, chemical,
and biological transformations. The results obtained for Toledo
bottom sediments suggest that they are composed largely of
allochthonous material. Much of this material probably
originates from the Maumee River, which delivers large quantities
of suspended sediments to Lake Erie at Toledo. In fact, the
characteristics of the Toledo bottom sediments closely resemble
121
-------
TABLE 26. CHEMICAL FRACTIONATION AND BIOASSAY RESULTS FOR. SHORELINE EROSION (BLUFF) AND
LAKE BOTTOM SEDIMENT SAMPLES
A) Chemical Fractionations (Values in ngP/g sed.):
Sample (Type) T-Sed-P T-NaOH-P R-NaOH-P NR-NaOH-P CDB-P HC1-P Residual P
Monroe jfl (bottom)
Monroe #2 (bottom)
Toledo #1 (bottom)
Toledo #2 (bottom)
Port Stanley (bluff )
Rondeau Park (bluff)
L. Superior (bluff )
' 2266
3551
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 #1
Monroe #2
Toledo #1
Port Stanley
Rondeau Park
T-Sed-P
in DCDA ' s
(ngP/g sed. )
2656
3044
947
662
557
Sediment P
Released in
in DCDA's
(pgP/g sed.)
964.2
1341.3
99.3
0
0
Ultimately
Bioavailable
Sediment P,
pult(|jgP/g sed.)
1435.0
1482.4
99.3
0
0
Pta a
i i do
ult
a % of
T-Sed-P
54.0
48.7
10.5
0
0
Release
Rate, kr
0.042
0.080
0.140
—
-------
those of Maumee River sediments from which 50-75% of the'
ultimately available sediment P has been released. puit ^or tne
Toledo #1 sample was measured at 99.3/igP/g sediment, compared to
a mean of 337.3/igP/g sediment for Maumee River suspended
sediments, while k was found to be very close to the mean slow
component release rate (k,,,.) for Maumee River samples (Table
S L
13). In addition, the commonly employed chemical measures of
available P (i.e. R-NaOH-P and NAIP) were substantially lower for
the Toledo bottom sediments than for Maumee River suspended
sediments.
As mentioned previously, the Monroe bottom sediments
contained much higher levels of T-Sed-P than any of the tributary
sediments analyzed. Based on the sampling location from which
the Monroe bottom sediments were obtained, it is likely that the
deposition of autochthonous particulate material, such as
phytoplankton, is largely responsible for the observed enrichment
in sediment P. This enrichment occurred almost exclusively in
the R-NaOH-P pool, and to a lesser extent in the CDB-P fraction
(see Table 26), or the non-apatite inorganic components of
sediment P. Therefore, it must have taken place via a
decomposition-mediated mineralization of autochthonous P at the
lake bottom, followed by reentry into the bottom sediments
through adsorption or precipitation.
The results of DCDA bioassays performed on Monroe bottom
sediments also suggest the influence' of autochthonous particulate
material at this site. First-order release rates measured for
these samples were less than the mean slow component release
rate, k__, for suspended sediments from any of the tributaries
S L
(Table 13). Bottom sediments receiving a steady influx of
phytoplankton biomass would no doubt develop a large decomposer
population. It is possible that P released from the Monroe
sediments in the DCDAs was immobilized temporarily by such
bacteria. This P would be re-released upon subsequent
decomposition of (or endogenous respiration by) these organisms,
resulting in a slower measured release rate.
The data of Table 26 indicate that the chemical nature of P
in Lake Erie bottom sediments may vary considerably as a function
of location in the lake. The amount of available P released from
bottom sediments during periods of resuspension'will depend
primarily on the point of origin within the lake and the duration
of residence in the water column. Sediments receiving large
influxes of autochthonous material may make substantial
contributions of available P to the water column upon
resuspension. For example, based on bioassay results, the Monroe
bottom sediments would release as much sediment P during a 3 day
resuspension period as would be released by Maumee River
123
-------
suspended sediments during a 7 day period.
SANDUSKY RIVER BOTTOM SEDIMENT SAMPLES
The effect of a point source of phosphorus on the sediment P
chemistry was also briefly examined in this study. The chemical
fractionation sequence was applied to bottom sediment samples
from three sites upstream and three sites downstream from the
Bucyrus, Ohio wastewater treatment plant discharge (on the
Sandusky River). The results are given in Table 27. Substantial
increases in the R-NaOH-P and CDB-P fractions occurred downstream
of the point source. The NR-NaOH-Pf HC1-P, and Residual P
fractions were also found to increase, but by smaller amounts.
This movement of phosphorus into the sediments is consistent with
the finding by Baker (1982a) that soluble reactive P
concentrations decreased rapidly with distance downstream from
the Bucyrus treatment plant discharge. However, Baker (1982b)
has suggested that P from such point sources is not exported as
R-NaOH-P, and thus, does not contribute significantly to the
input of bioavailable particulate phosphorus to the receiving
water. The data presented in Table 27 indicate that, although
much of the point source P may be tied up in relatively
unavailable forms such as CDB-P, HC1-P, and Residual P, a sizable
amount is also present as R-NaOH-P, and is therefore potentially
bioavailable.
It appears that phosphorus from the point source is capable
of assuming a number of different chemical forms in association
with sediments. Baker (1982b) suggested that biological
processes were more important than physical adsorption in the
processing of point source inputs. The increases in R-NaOH-P and
CDB-P observed in the present study, however, indicate the
activity of physicochemical mechanisms. Biological activity also
appears to play a role, although a less important one, in the
fixation of point source P, as suggested by the large increase in
NR-NaOH-P for site #6. It should be noted that the work of Baker
(1982b) was conducted on suspended sediments from the Sandusky
River, rather than on bottom sediments. The processing of P
inputs by the indigenous algal population would be a more
important factor in the water column of the river than in the
bottom sediments. However, the environment should also be
favorable for an increase in the R-NaOH-P concentration of
suspended sediments by adsorption of SRP from solution. CDB-P,
on the other hand, would be more likely to accumulate in the
bottom sediments than in suspended sediments, since higher levels
of soluble species such as Fe and Al are typically found at
the sediment-water'interface. These may result in the formation
of iron or aluminum oxide coatings on the surface of sediment
particles, and thus, the occlusion of surface-bound phosphorus.
Although an indirect point source of phosphorus may cause a
substantial increase in the amount of available P bound to
124
-------
TABLE 27. EFFECTS OF A POINT SOURCE DISCHARGE ON PHOSPHORUS
FRACTIONATION IN SANDUSKY RIVER BOTTOM SEDIMENTS
K)
Ul
Site
no.
1
2
4
6
Miles
from
point
source
9
7
0.5
10
Sandusky River
T-Sed-P
777
644
A *7 A
4 /U
1510
1039
1404
Concentrations
in ygP/g
T-NaOH-P R-NaOH-P NR-NaOH-P
284
226
C K
3D
528
97
596
215
177
4-3
J
434
320
305
69
49
94
77
291
sediment
CDB-P
214
160
452
308
177
HC1-P
125
123
O C. 1
ZD J
249
165
323
Flow
Residual P
107 —
m —
yi
132
119
170 —
I
— 0
A
""* MA ^^""
r
-•
Point
Source
Discharge
-------
sediments, on the whole, attachment of this P to particulate
material causes a decrease in its potential bioavailability.
DePinto £i_^l. (1980) performed bioassay measurements of
phosphorus availability on effluents from four municipal sewage
treatment plants in the Lower Great Lakes basin. They reported
that an average of 72% of the total P - 55% of the particulate P
fraction and 82% of the total soluble fraction - was available
for uptake by Sssns^ssi^^s sp. Incorporation of this highly
available phosphorus into the CDB-P, HC1-P, and Residual P
fractions of the bottom sediments essentially immobilizes a
portion of the point source P. Based on the data collected in
this study, however, it is not possible to quantify the reduction
in the bioavailability of P from the Bucyrus treatment plant due
to processing in the river. This would require characterization
of P associated with suspended particulates as well as soluble P
concentrations as a function of distance downstream from the
point source discharge.
126
-------
SECTION 5
IMPLICATIONS OF RESULTS FOR PHOSPHORUS MANAGEMENT
As part of the effort to reduce the productivity of the
Lower Great Lakes, target phosphorus loads were established by
Task Group III (1978). These were based on total P and were set
at 11,000 metric tons/year for Lake Erie and 7,000
metric tons/year for Lake Ontario. However, the results of this
study indicate that the various forms of P present in inputs to
these lakes often have vastly different degrees of availability
to aquatic organisms. For example, sediment P associated with
Lake Erie bluff material was totally unavailable to J3.el.eiia.six.um,
while SRP inputs from the Ohio tributaries were nearly all
rapidly available. In the existing phosphorus budgets for the
Lower Great Lakes, inputs are broken down into two forms -
soluble and particulate. Such budgets may not provide an
entirely accurate picture of the relative contribution of each
source to productivity in the receiving water. An improvement
could be made by separating the particulate P inputs into
ultimately available and unavailable fractions. Based on the
results of the present study, Salisbury 5i_5l. (1983) have made
preliminary revisions of loading estimates by the U.S. Army Corps
of Engineers for Lake Erie to reflect the bioavailability of PP.
Expressing target loads for the Lower Great Lakes in terms of the
sum of bioavailable fractions should result in a clearer idea of
the steps necessary to achieve the desired improvements in water
quality. In this way, a more efficient appropriation of
resources for upgrading the trophic condition of these lakes
might be realized.
i
The existing target total P loads were developed through the
application of mathematical models of lake trophic status. If
external loads are separated into three different forms - SRP,
ultimately available PP, and refractory PP - as suggested above,
certain changes in the structure of these models would be
necessary. The concept of differing bioavailability of P sources
is already built into the current Great Lakes models to a certain
extent. For example, phytoplankton in the water column are
assumed to utilize only phosphorus present as SRP. Also, forms
of sediment P having an extremely low degree of bioavailability,
such as inputs from Lake Erie bluff erosion, are often eliminated
from the loading data. Particulate P in the water column is
assumed to be released to solution (as SRP) at a fixed
first-order rate.
127
-------
In the existing models, all P associated with suspended
particulate material is considered to be capable of conversion to
SRP. The rate at which this conversion takes place is assumed to
be the same for all forms of PP regardless of its origin.
However, the results of this study indicate that a sizable
fraction of PP in tributary suspended sediments is unavailable
for algal uptake. The percentage available varies among
tributaries with differences in soil and land use conditions. In
addition, it appears that allochthonous particulate material
exhibits much different P release kinetics than autochthonous
material in the water column. Sediment inputs from tributary
sources generally release PP at a faster first-order rate, but
less completely than decomposing algal biomass. Lake bottom
sediments, which contain both types of material, exhibit
bioavailability characteristics that are influenced by both
components. Modification of the existing models to account for
the different bioavailability kinetics of allochthonous and
autochthonous particulate P would result in a more realistic
mechanistic description of in-lake processes. This, in turn,
would tend to minimize any artificial adjustments in the model
coefficients that may be necessary to calibrate the existing
models. With this change, the models should also provide more
accurate predictions of the response of lake water quality to
reductions in the phosphorus loads from various sources.
Preliminary work conducted by Salisbury s£_al. (1983) using the
model developed by DiToro and Matystik (1980) for Lake Erie has
indicated that implementation of the changes mentioned above is
feasible.
Lake bottom sediments are composed of a combination of
allochthonous and autochthonous particulate material. The
results obtained in this study suggest that, where influxes of
the latter are relatively large, the bottom sediments may undergo
a significant enrichment in sediment P. Also, much of this may -
be present in.biologically available forms. Therefore, the
resuspension of bottom sediments during high wind periods in the
shallower basins of Lake Erie represents a potentially important
source of available P. The Great Lakes models currently
employed, in general, do not include a term to .account directly
for the input of P via resuspension of bottom sediments. This
phenomenon may be partially accounted for by adjustment of
particulate settling velocities to effect a calibration for
particulate P concentrations in the water column. However,
inclusion of an expression describing sediment resuspension, in
combination with the availability kinetics of the associated P,
should result in an improvement in the model predictions.
Differentiation between allochthonous and autochthonous
sources of particulate P in Great Lakes water quality models
requires that the bioavailability of sediment P inputs be
characterized. This may be accomplished most accurately using
bioassay measurements such as the DCDA technique employed in this
study. Since these methods tend to be quite time-consuming,
128
-------
however, chemical extractions are often used to estimate the
bioavailability of sediment P. In this study, P released during
the bioassays on tributary suspended sediments was found to be
associated predominately with the R-NaOH-P fraction. Other
chemical fractions (such as CDB-P and T-Sed-P) were also strongly
correlated with Pu-it. However, it is felt that predictions of
Pult snould be based on either R-NaOH-P or NAIP (= R-NaOH-P +
CDB-P), since the data indicate a mechanistic as well as a
statistical connection between R-NaOH-P and bioavailability.
None of the chemical fractions measured in this study provided
accurate predictions of the first-order sediment P release rate,
kr. Release rate did not exhibit much fluctuation either among
samples for a given tributary or between tributaries.
Based on bioassays and extractions conducted during the
first year of this study, Salisbury si_sl. (1983) considered the
following percentages of particulate P inputs to be ultimately
available for algal uptake in the water column of Lake Erie:
1) Detroit River - 20% (from regression of P , . versus
R-NaOH-P);
2) Western Basin (excluding Detroit R.) - 25% (from
bioassays on Maumee R. samples);
3) Central Basin - 28% (from bioassays on Sandusky R. and
Cuyahoga R. samples);
4) Eastern Basin - 8% (from bioassays on Cattaraugus Cr.
samples).
These authors assumed the rate of sediment P release to be the
same for all sources - 0.154 day , based on the average of all
bioassay measurements. When the bioassay results from the second
year of this study were added to the data set, the same averages
were obtained for the percentages of T-Sed-P available from
tributary sources in all basins. However, the average release
rate for all Lake Erie tributary samples from both years was
found to be 0.180 day . The bioavailability kinetics for P
associated with Lake Erie bottom sediments were found to vary
considerably as a function of location in the lake.
The information on bioavailability of sediment P inputs
gathered in this study can serve as a preliminary basis for
implementing the changes in phosphorus loading calculations and
model structure suggested above. However, this should be
supplemented with additional data from the tributaries studied
and others in the Lower Great Lakes basin, as well as for bottom
sediments at locations where resuspension is likely.
Characterization of Detroit River suspended sediments is
129
-------
especially important, since this source constitutes over 40% of
the total PP input to Lake Erie.
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 bioavailability. The results presented in this report
indicate that bioavailability remains relatively constant for a
given set of soil type (i.e. geochemistry and texture) and land
use conditions. Evidence of this 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 anthropogenic influence, displayed a higher level of
sediment P bioavailability. Relatively little fluctuation in
sediment characteristics was observed over time for any given
tributary. Based on these findings, it is felt that future
monitoring programs should concentrate on identifying differences
between regions of distinct soil type and land use. This can be
accomplished most readily using routine chemical extraction
analyses on sediments transported during storm runoff events
supplemented by less frequent bioassay measurements of sediment P
bioavailability.
130
-------
REFERENCES
Allan, R.J., and J.D.H. Williams, "Trophic Status Related to
Sediment Chemistry of Canadian Prairie Lakes." J. Environ.
Qual., 7, 99-106 (1978).
APHA, AWWA, and WPCP, Standard Methods for the Examination of
Water and Waste vat ei*"!14th Edition, Washington, DC (.1976) .
Armstrong, D.E., R.F.Harris, and J.K. Syers, "Plant Available
Phosphorus Status of Lakes." Final Technical Completion
Report, OWRR B-022-WIS, 29 pp. (1971).
Armstrong, D.E., J.J. Perry, and D.E. Flatness, "Availability of
Pollutants Associated with Suspended or Settled River
Sediments Which Gain Access to the Great Lakes." USEPA
Technical Report No. 90'5/4-79-028, Region V, Chicago,
Illinois (1979).
Atkinson, R.J., A.M. Posner, and J.P. Quirk, "Kinetics of
Isotopic Exchange of Phosphate at the QC-FeOOH-Aqueous Solution
Interface." J. Inorg. Hucl. Ghem., 34, 2201-2211 (1972).
Autenrieth, R.L., "The Effects of Acid Lake Recovery by Chemical
Manipulation on Algal Growth Potential." M.S. Thesis, Glarkson
College of Technology, Potsdam, NY (1981j.
Bache, B.W., "Aluminum and Iron Phosphate Studies Relating to
Soils: I. Solution and Hydrolysis of Variscite and
Strengite." J. Soil Sci. , \A, 113-123 (1963).
Bache, B.W., "Aluminum and Iron Phosphate Studies Relating to
Soils: II. Reactions Between Phosphate and Hydrous Oxides."
J. Soil Sci. , 15, 110-116 (1964)."
Baker, D.3., "Upstream Point Source Phosphorus Inputs and
Effects." Seminar on Water Quality Management Tradeoffs:
Point Source vs. Diffuse Source Pollution, Chicago,. Illinois,
EPA-905/9-30-009, 227-239 (1930).
Baker, D.B., "Fluvial Transport and Processing of Sediments and
Nutrients in Large Agricultural River Basins." LEWMS Technical
Report, U.S. Army Corps of Engineers, Buffalo, N.Y. (I982a).
Baker, D.B., "Characteristics of the Transport of Bioavailable
Particulate Phosphorus in the Sandusky River Basin." Final
Report to Proctor and Gamble Co., Cincinnati, Ohio (!982b).
131
-------
Barrow, N.J., "The Description of Desorption of Phosphate from
Soil." • J. Soil Sci., 30, 259-270 (1979).
Bauwin, G.R., and S.H. Tyner, "The Nature of Reductant-Soluble
Phosphorus in Soils and Soil Concretions." Soil Sci. Soc. Am.
Proc., 21_, 250-257 (1957).
Beckwith, V.R., "Chemical Fractionation as an Indicator -of
Phosphorus Bioavailability in Suspended Solids from
Tributaries to Lake Erie." Master's Thesis, Clarkson College
of Technology, Potsdam, NY (1982).
Bierman, V.J., Jr., D.M. Dolan, E.P. Stoermer, J.E. Gannon, and
V.E. Smith, "The Development and Calibration of a Spatially
Simplified Multi-Class Phytoplankton Model for Saginaw Bay,
Lake Huron." Great'Lakes Environmental Planning Study (GLEPS),
Contribution No. 33, Great Lakes Basin Commission (1980).
3onner,J.S., J.V. DePinto, and P.V. Rodgers, "Analysis of Aquatic
Particle Size Distributions Using a Computer-Interfaced
Coulter Counter." Clarkson College of Technology, In
preparation (1983).
Bray, R.H., and L.T. Kurtz, "Determination of Total, Organic, and
Available Forms of Phosphorus in Soils." Soil Sci., 59, 39-45
(1945). ;~~
Breeuwsma, A., and J. Lyklema, "Physical and Chemical Asorption
of Ions in the Electrical Double Layer on Hematite (tf-Fe-O*)."
J. Colloid Interface Sci., 43, 437-448 (1973).
Burwell, R.S., D.R. Timmons, and R.P. Holt, "Nutrient Transport
in Surface Runoff as Influenced by Soil Cover and Seasonal
Periods." Soil Sci. Soc. Am. Proc., 39, 523-528 (1975).
Cahill, T.H., P. Imperato, and P.H. Verhoff, "Evaluation of
Phosphorus Dynamics in a Watershed." J. Environ. Eng. Div.,
Proc. Am. Soc. Civ. Bng., 100, No. SS2, 459-458(1974;•
Carignan, R., and J. Xalff, "Quantification of the Sediment
Phosphorus Available to Aquatic Macrophytes." J. Fish. Res.
3d. Can., 36., 1002-1005 U979)-
Chang, S.C., and M.L. Jackson, "Practionation of Soil
Phosphorus." Soil Sci. , 84, 133-U4 (1957).
Chapra, S.C., and W.C. Sonzogni, "Great Lakes Total Phosphorus
Budget for the Mid-1970'a." J. Water Pollut. Control Fed.,
51, 2524-2533 (1979).
Chen, Y.R., J.N. Butler, and W. Stunm, "Adsorption of Phosphate
on Alumina and Kaolinite from Dilute Aqueous Solutions." J_._
Colloid Interface Sci., 43, 421-436 (I973a).
132
-------
Chen, Y.R., J.N. Butler, and W. Stumm, "Kinetic Study of
Phosphate Reaction with Aluminum Oxide and Kaolinite."
Environ. 3ci. Technol., 7, 327-332 (1973b).
Cole, C.V., S.R. Olsen, and C.O. Scott, "The Nature of Phosphate
Sorption by Calcium Carbonate." Soil Sci. Soc. Am. Proc., 17,
352-356 (1953).
Coulter Electronics, Inc., "Operator's Reference. Manual - Coulter
Counter Model ZBI." Hialeah, Florida (1979).
Cowen, W.F., and G-.F. Lee, "Phosphorus Availability in
Particulate Materials Transported by Urban Runoff." J. Water
Pollut. Control Fed., 48, 580-591 (I976a).
Cowen, W.F., and G-.F. Lee, "Algal Nutrient Availability and
Limitation in Lake Ontario During IFYGL: Part I. Available
Phosphorus in Urban Runoff and Lake Ontario Tributary Waters."
Ecological Research Series, EPA-600/3-76-094a, 218 pp.
(1976b).
DePinto, J.V., J.K. Edzwald, M.S. Switzenbaum, and T.C. Young,
"Phosphorus Removal in Lower Great Lakes Municipal Treatment
Plants." USSPA Resort No. 600/2-80-177, Cincinnati, Ohio
(1980).
DePinto, J.V., "An Experimental Apparatus for Evaluating Kinetics
of Available Phosphorus Release from Aquatic Particulates."
Water Res., 1_6, 1065-1070 (1982).
DiToro, D.M., "The Effect of Phosphorus Loadings on Dissolved
Oxygen in Lake Erie." In: Phosphorus Management Strategies
for Lakes, R.C. Loehr, C.S. Martin, and W. Rast (Eds.), Ann
Arbor Science Publishers, Inc., Ann Arbor, Michigan, 191- 205
(1980).
DiToro, D.M., and W.F. Matystik, Jr., "Mathematical Modlels of
Water Quality in Large Lakes, Part 1: Lake Huron and* Saginaw
Bay.". US3PA Report No. 600/3-80-056 (1980).
DiToro, D.M., and J.P. Connolly, "Mathematical Models of Water
Quality in Large Lakes, Part 2: Lake Erie." USEPA Report No.
600/3-80-065 (1980).
Dong, A., G-.V. Simsiman, and G-. Chesters, "Particle-size
Distribution and Phosphorus Levels in Soil, Sediment, and
Urban Dust and Dirt Samples from the Menomonee River
Watershed, Wisconsin, USA." Water Res., 17, 569-577 (1983).
Dorich, R.A., D.W. Nelson, and L.E. Sommers, "Algal Availability
of Sediment Phosphorus in Drainage Water of the Black Creek
Watershed." J. Environ. Qual., 9, 557-563 (1980).
133
-------
Sdawald, J.K., D.C. Toensingi and M.C. Leung, "Phosphate
Adsorptio.n Reactions with Clay Minerals." Environ. Sci.
Technol., HD, 485-490 (1976).
Fitzgerald, Or.P., "Aerobic Lake Muds for the Removal of
Phosphorus from Laice Waters." Limnol. Oceanog., 15» 550-555
(1970).
Prink, C.R., "Chemical and Mineralogical Characteristics of
Eutrophic Lake Sediments." Soil Sci. Soc. Am. Proc., 33. >
369-372 (1969).
Golterman, H.L., "Sediments as a Source of Phosphorus for Algae
Growth." In: Interactions Between Sediments and Freshwaters
(Proc. SIL-UN33CO Conference), K.L. Golterman (.Bd.j, Junk and
Pudoc, The Hague, Netherlands (1976).
.Golterman, H.L., C.C. Bakels, and J. Jakolbs-Mogelin,
"Availability of Mud Phosphates for the Growth of Algae."
Yerh. Int. Ver. Limnol., JJ7, 467-479 (1969).
Green, D.3., T.J. Logan, and N.E. Smeck, "Phosphate Adsorption-
Desorption Characteristics of Suspended Sediments in the
Maumee River Basin of Ohio." J. Environ. Qual., 7, 208-212
(1973). ~
Griffin, R.A., and J.J. Jurinak, "Kinetics of Phosphate
Interaction with Calcite." Soil Sci. Soc. Am. Proc., 38,
75-79 (1974). ~~
Guillard, R.R.L., and C.J. Lorensen, "Yellow-Green Algae with
Chlorophyllide C." J. Phycol., 8, 10-14 (1972).
Harms, L.L., J.N. Dornbush, and J.R. Anderson, "Physical and
Chemical Quality of Agricultural Land Runoff." J. Water
Pollut. Control Fed., 46, 2460-2470 (1974).
Hegemann, D.A., J.D. Keenan, and A.H. Johnson, "Determination of
Particulate-Nutrient Availability." J. Environ. Eng. Div. ,
Proc. Am. Soc. Civ. Bng. , 108, 1312-1325 (J982).
Helyar, K.R., D.N. Munns, and R.G. Burau, "Adsorption of
Phosphate by Gibbsite: II. Formation of a Surface Complex
• Involving Divalent Cations." J. Soil Sci., 27, 315-323
(1976).
Hieltjes1, A.H.M. , and L. Lijklema, "Fractionation of Inorganic
Phosphates in Calcareous Sediments." J. Snviron. Qual., 9,
405-407 (1980). ;~
Kingston, F.J., R.J. Atkinson, A.M. Posner, and J.P. Quirk,
"Specific Adsorption of Anions." Nature, 215, 1459-1461
(1967).
134
-------
Hingston, P.J., A.M. Posner, and J.P. Quirk, "Anion Adsorption by
Goethite and Gibbsite: I. The Role of the Protons in
Determining Adsorption Envelopes." J. Soil Sci., 23, 177-192
(1972).
Hingston, P.J., A.M. Posner, and J.P. Quirk, "Anion Adsorption by
Goethite and Gibbsite: II. Desorption of Anions from Hydrous
Oxide Surfaces." J. Soil Sci., 25, 16-26 (1974).
Hsu, P.H., "Adsorption of Phosphate by Aluminum and Iron in
Soils." Soil Sci. Soc. Am. Proc., 28, 474-478 (1964).
Hsu, P.H., "Fixation of Phosphate by Aluminum and Iron in Acidic
Soils." Soil Sci., 99, 398-402 (1965).
Huettl, P.J., R.C. Wendt, and R.B. Corey, "Prediction of Algal-
Available Phosphorus in Runoff Suspensions." J. Environ.
Qual., 8, 130-132 (1979).
Kafkafi, U., A.M. Posner, and J.P. Quirk, "Desorption of
Phosphate from Kaolinite." Soil Sci. Soc. Am. Proc., 3.1.>
348-353 (1967).
Keup, L.E., "Phosphorus in Plowing Waters." Water Res., 2_,
373-386 (1968).
Kuo, S., and E.G. Lotse, "Kinetics of Phosphate Adsorption by
Calcium Carbonate and Ca-Kaolinite." Soil Sci. Soc. Am.
Proc., 3i, 725-729 (1972).
Kuo, S-. , and S.G-. Lotse, "Kinetics of Phosphate Adsorption and
Desorption by Lake Sediments." Soil Sci. Soc. Am. Proc., 38,
50-54 (1974).
Kwolek, J., "Particle Size and Settling Velocity Analysis of
Suspended Solids from ?our Lake Erie Tributaries." M.S.
Thesis, Clarkson College of Technology, Potsdam, NY (1982).
Lam, D.C.L., and J.-M. Jaq.uet, "Computations of Physical
Transport and Regeneraion of Phosphorus in Lake Erie, Fall
1970." J. Fish. Res. Bd. Can., 33., 550-563 (1976).
Lavler, D.P., "A Particle Approach to the Thickening Process."
Ph.D. Dissertation, University of North Carolina, Chapel Hill
(1979).
Lerman, A., K.L. Carder, and P.R. Betzer, "Elimination of Pine
Suspensions in the Oceanic Water Column." Earth and -Planetary
Science Letters, 37, 61-70 (1977).
Logan, T.J., "Chemical Extraction as an Index of Bioavailability
of Phosphate in Lake Erie Basin Suspended Sediments." LEWS
Report, U.S. Army Corps of Engineers, Buffalo, N.Y., 49 pp.
(1978).
135
-------
Logan, T.J., P.H. Verhoff, and J.V. DePinto, "Biological
Availability of Total Phosphorus." LEWM3 Report, U.S. Army
Corps of Engineers, Buffalo, N.Y., 62 pp. (1979).
Li, W.C., D.E. Armstrong, J.D.H. Williams, R.P. Harris, and J.K.
Syers, "Rate and Extent of Inorganic Phosphate Exchange in
Lake Sediments." Soil Sci. Soc. Am. Proc., 36, 279-285
(1972).
McCallister, D.L., and T.J. Logan, "Phosphate Adsorption-
Desorption Characteristics of Soils and Bottom Sediments in
the Maumee River Basin of Ohio." J. Environ. Qual., 7, 87-92
(1978).
Mehta, N.C., J.O. Legg, C.A.I. Goring, and C.A. Black,
"Determination of Organic Phosphorus in Soils: I. Extraction
Method." Soil Sci. Soc. Am. Proc. , j_8, 443-449 (1954).
Millard, E.S., C.C. Charlton, and G.A. Burnison, "Availability of
Phosphorus in Different Sources Entering the Great Lakes for
Algal Growth." Report to Task Group D (Canadian Section),
PLUARG, 38 pp. (1979).
Muljadi, D., A.M. Posner, and J.P. Quirk, "The Mechanism of
Phosphate Adsorption by Kaolinite, G-ibbsite and
Pseudoboehmite: Part I. The Isotherms and the Effect of pH on
Adsorption." J. Soil Sci. , 17, 212-229 (1966a).
Muljadi, D., A.M. Posner, and J.P. Quirk, "The Mechanism of
Phosphate Adsorption by Kaolinite, Gibbsite, and
Pseudoboehmite: Part II. The Location of the Adsorption
Sites." J. Soil Sci., ]_7, 230-237 (I966b).
Murphy, J., and J.P. Riley, "A Modified Single Solution Method
for the Determination of Phosphate in Natural Waters." Anal.
Chim. Acta, 27, 31-36 (1962).
01oya,.T.O., and T.J. Logan, "Phosphate Desorption from Soils and
Sediments with Varying Levels of Extractable Phosphate." J_._
Environ. Qual., £, 526-531 (1980).
Olsen, S.R., C.V. Cole, P. Watanabe, and L.A. Dean, "Estimation
of Available Phosphorus in Soils by Extraction with Sodium
Bicarbonate." USDA Circular 939 (1954)-
Parfitt, R.L., R.J. Atkinson, and R.St.C. Smart, "The Mechanism
of Phosphate Fixation by Iron Oxides." Soil Sci. Soc. Am.
Proc., 39, 837-841 (1975).
Peters, R.H., "Phosphorus Availability in Lake Memphremagog and
its Tributaries." Limnol. Oceanogr., 26, 1150-1161 (1981).
136
-------
Reddy, M.M., "Sediment Phosphorus Characterization in the G-enesee
River Watershed, New York." In: Contaminants and Sediments;
Volume 2. Analysis, Chemistry, Biology, R.A. Baker (Ed.),Ann
Arbor Science Publishers, Inc., Ann Arbor, Michigan, 119-136
(1930).
Robarge, W.P., and R.B. Corey, "Adsorption of Phosphate by
Hydroxy-Aluminum Species on a Cation Exchange Resin." Soil
Sci. Soc. Am. J., 43., 431-487 (1979).
Rodel, M.G., D.E. Armstrong, and R.F. Harris, "Sorption and
Hydrolysis of Added Organic Phosphorus Compounds in Lake
Sediments." Limnol. Oceanogr., 22, 415-422 (1977).
Rodgers, P.W., "Kinetic Studies of Phytoplankton Decomposition."
Ph.D. Dissertation, Clarkson College of Technology, Potsdam,
NY (1979).
Rodgers, P.W., and D.K. Salisbury, "Modeling Water Quality.in
Lake Michigan and the Effect of the Anomalous Ice Cover of
1976-1977." A Great Lakes Environmental Planning Study Report,
Great Lakes Basin Commission (1981).
Romkens, M.J.M., and D.W. Nelson, "Phosphorus Relationships in
Runoff from Fertilized Fields." J. Environ. Qual., 3, 10-14
(1974). "
•Ross, S.L., "Introduction to Ordinary Differential Equations."
Isrox College Publishing, Lexington, Massachusetts (1974).
Ryden, J.C., and J.K. Syers, "Rationalization of Ionic Strength
and Cation Effects on Phosphate Sorption by Soils." J. Soil
Sci., 26, 395-406 (1975).
Ryden, J.C., J.R. McLaughlin, and J.K. Syers, "Mechanisms of
Phosphate Sorption by Soils and Hydrous Ferric Oxide Gel." J.
Soil Sci., 28, 72-92 (I977a).
Ryden, J.C., J.K. Syers, and J.R. McLaughlin, "Effects of Ionic
Strength on Chemisorption. and Potential-Determining Sorption
of Phosphate by Soils." J. Soil Sci., 28, 62-71 U977b).
Sagher, A., "Availability of Soil Runoff Phosphorus to Algae."
Ph.D. Dissertation, University of Wisconsin - Madison, 176
pp. (1976).
Salisbury, D.K., J.V. DePinto, and'T.C. Young, "Impact of Algal-
Available Phosphorus on Lake Erie Water Quality: Mathematical
Modeling." USEPA Environmental Research Laboratory - Duluth,
Minnesota (1983)•
137
-------
Saunders, W.M.H., "Phosphate Retention by New Zealand Soils and
its Relationship to Free Sesquioxides, Organic Matter, and
Other Soil Properties." N. Z. J. Agric. Res., 8, 30-57
(1965).
Sharpley, A.N., "The Enrichment of Soil Phosphorus in Runoff
Sediments." J. Environ. Qual., 9, 521-526 (1980).
Sharpley, A..N., and J.K. Syers, "Phosphorus Inputs into a Stream
Draining an Agricultural Watershed: Amounts and Relative
Significance of Runoff Types." Water, Air, Soil Pollut . , 9.,
417-428 (1979).
Sharpley, A.N. , R.G. Menzel, S.J. Smith, -E.D. Rhoades , and A.E.
Olness, "The Sorption of Soluble Phosphorus by Soil Material
During Transport in Runoff from Cropped and Grassed
Watersheds." J. Environ. Qual., 10, 211-215 (1981).
Shukla, S.S., J.K. Syers, J.D.H. Williams, D.E. Armstrong, and
R.F. Harris, "Sorption of Inorganic Phosphate by Lake
Sediments." Soil Sci. Soc. Am. Proc. , 35, 244-249 (1971).
Sommers, L.E., R.P. Harris, J.D.H. Williams, D.E. Armstrong, and
J.K. Syers, "Fractionation of Organic Phosphorus in Lake
Sediments." Soil Sci. Soc. Am. Proc. , 36_, 51-54 (1972).
Sonzogni, W.C., S.C. Chapra, D.E. Armstrong, and T.J. Logan,
"Bioavailability of Phosphorus Inputs to Lakes." J. Environ.
Qual., n_, 555-563 (1982). *
Sugarman, R.J., "Conference Objectives." In: Phosphorus
Management Strategies for Lakes, R.C. Loehr, C.S. Martin, and
W. Rait (Eds. ) , Ann Arbor Science Publishers, Ann Arbor,
Michigan, 7-10 (1930).
Syers, J.K., and T.W. Walker, "Phosphorus Transformations in a
Chronosequence of Soils Developed on Wind-31own Sand in New
Zealand: I. Total and Organic Phosphorus." J. Soil Sci . , 20,
57-64 (1969). ~~
Syers, J.K., M.G. Browman, G-.W. Smillie, and R.B. Corey,
"Phosphate Sorption by Soils Evaluated by the Langmuir
Adsorption Equation." Soil Sci. Soc. Am. Proc., 37, 358-563
—
Task Group III, "Fifth Year Review of Canada-United States Water
Quality Agreement." Report of Task Group III, A Technical
Group to Review Phosphorus Loadings, U.S. Department of State,
Washington, D.C. (1978).
138
-------
Thomann, R.V., D.M. DiToro, R.P. Winfield, and D.J. O'Connor,
"Mathematical Modeling of Phytoplankton in Lake Ontario: I.
Model Development and Verification." USEPA Report No.
600/3-75-005, Ecological Research Series, Large Lakes Research
Lab, Grosse He, Michigan (1975).
Thomas, H.A., Jr., "Graphical Determination of BOD Curve
Constants." Vater and Sevage Works, 97> 123 (1950).
U.S. Army Corps of Engineers, "Lake Erie Wastewater Management
Study - Final Report." Buffalo, N.Y. (1932).
Verhoff, F.H., and D.A. Melfi, "Total Phosphorus Transport During
Storm Events." J. Environ. Eng. Div., Proc. Am. Soc. Civ.
Eng., 104, 1021-1026 U978J.
Verhoff, F.H., D.A. Melfi, and S.M. Yaksich, "Storm Travel
Distance Calculations for Total Phosphorus and Suspended
Materials in Rivers." Water Resour. Res., 15, 1354-1360
(1979).
Verhoff, F.H., M. Heffner, and W.A. Sack, "Measurement of
Availability Rates for Total Phosphorus from River Waters."
LEWMS Technical Report, U.S. Army Corps of Engineers, Buffalo,
N.Y. (1978).
Weaver, R.M., "A Simplified Determination of Reductant-Soluble
Phosphate in Soil Phosphate Fractionation Schemes." Soil Sci.
Soc. Am. Proc., 38, 153-154 (1974).
Wentz, D.A., and G.F. Lee, "Sedimentary Phosphorus in Lake Cores
- Observations on Depositional Pattern in Lake Mendota."
Environ. Sci. Technol., 2» 754-759 (1969).
Wildung, R.E., and R.L. Schmidt, "Phosphorus Release from Lake
• Sediments." USEPA Report No. EPA-R3-73-024, 185 pp. (1973).
Williams, J.D.H., and T.W. Walker, "Fractionation of Phosphate in
a Maturity Sequence of Hew Zealand Basaltic Soil Profiles: I."
Soil Sci., 107, 22-30 (1969).
Williams, J.D.H., J.K. Syers, S.S. Shukla, R.F. Harris, and D.E.
Armstrong, "Levels of Inorganic and Total Phosphorus in Lake
Sediments as Related to Other Sediment Parameters." Environ.
Sci. Technol., 5., 1113-1120 (1971 a).
Williams, J.D.H., J.K. Syers, R.F. Harris, and D.E. Armstrong,
"Fractionation of Inorganic Phosphate in Calcareous Lake
Sediments." Soil Sci. Soc. Am. Proc., 35, 250-255 d971b).
Williams, J.D.H., J.K. Syers, D.E. Armstrong, and R.F. Harris,
"Characterization of Inorganic Phosphate in Noncalcareous Lake
Sediments." Soil Sci. Soc. Am. Proc., 35, 556-561 d971c).
139
-------
Williams, J.D.H., J.M. Jaquet, and R.L. Thomas, "Forms of
Phosphorus in the Surficial Sediments of Lake Erie." J. Fish.
Res. Bd. Can., 33., 4-13-429 (1976a).
Williams, J.D.H., T.P. Murphy, and T. Mayer, "Rates of
Accumulation of Phosphorus Forms in Lake Erie Sediments." J_._
Fish. Res. 3d. Can., 33., 430-439 0976b).
Williams, J.D.H., H. Shear, and R.L. Thomas, "Availability to
Scenedesmus quadricauda of Different Forms of Phosphorus in
Sedimentary Materials from the Great Lakes." Limnol.
Qceanogr., 25, 1-11 (1980a).
Williams, J.D.H., T. Mayer, and J.O. Nriagu, "Extractability of
Phosphorus from Phosphate Minerals Common in Soils and
Sediments." Soil Sol. Soc. Am. J. , 44, 462-465 (19801)).
Young, T.C., J.V. DePinto, S.B. Flint, M.S. Switzenbaum, and J.K.
Edzwald, "Algal Availability of Phosphorus in Municipal
Waatewater." J. Water Pollut. Control Fed., 54, 1505-1516
(1932). ~
140
-------
BIBLIOGRAPHY OF RELATED PUBLICATIONS
Bollenbacher, M. , "Ion Exchange Resins as Estimators of
Algal-Available Sediment Phosphorus from Five Lower. Great
Lakes Tributaries." M.S. Thesis, Clarkson College of
Technology, Potsdam, NY (1981) .
DePinto, J.V., T.C. Young, and S.C. Martin, "Algal-Available
Phosphorus in Suspended Sediments from Lower Great Lakes
Tributaries." J^_fi££3i_LaJS£.s_I&es.a/ 1, 311-325 (1981)
DePinto, J.V., T.C. Young, S.C. Martin, and J.S. Bonner, "Release
of Available Phosphorus from Particulate Matter in the Great
Lakes." Submitted to J^_£jT55i_L3JjS5_R5^ (1983).
Victore, A.M., "Physical and Chemical Characterization of
Algal-Available Phosphorus in Suspended Sediments from Lower
Great Lakes Tributaries." M.S. Thesis, Clarkson College of
Technology, Potsdam, NY (1981) .
Young, T.C. , and J.V. DePinto, "Algal-Availability of Particulate
Phosphorus from Diffuse and Point Sources in the Lower Great
Lakes Basin." in 5sdimsj)i/I'JC55i)W5tsx_JBiSJ5£ii5U, P.G. Sly
(ed.), Dr W. Junk Publishers, 111-119 (1982).
141
-------
APPENDIX A
THE POWER LAW FUNCTION AND RELATED EQUATIONS
In analyzing polydisperse suspensions, a mathematical
description of the distribution of particle sizes is often
useful. The three most common measures of particle size are
particle diameter (d ), surface area (S), and volume (V). An
example of a particle size distribution based on particle volume
is shown in Figure A-l. In this plot, the cumulative number
concentration (N), in units of number/cm , is equal to the total
concentration of particles with a volume equal to or less than V
(jum ) . Alternatively, the distribution of particle sizes could
be represented by a plot of N versus d or S. The slopes of
these curves can be described by mathematical expressions known
as particle size distribution functions (Lawler, 1979).. These
functions may be defined as follows:
AN/ AV = dN/dV = n(V) (number/cm3um3) (A.I)
AN/ AS = dN/dS = n(S) (number/cm3um2) (A.2)
AN/Ad = dN/d(d ) = n(d ) (number/cm3urn) (A.3)
where n(V), n(S), and n(d ) represent the particle size
distribution functions.
As was mentioned in the text of Chapter 3, for aquatic
suspensions, the particle size distribution function often takes
the form:
n(dp) = A dp" (A. 4)
142
-------
f
Nc
cr.
LJ
QQ
O
CJ
PARTICLE VOLUME
Figure A-l.
Hypothetical particle size distribution
(from Lawler, 1979).
143
-------
where A = a coefficient related to the total particle
concentration;
and 6 = the power law exponent.
The value of 8 for a particular sample may be determined
experimentally from the slope of a log-log plot of n(d ) versus
d . For naturally occurring particles, values of & typically
range from 2 to 5 (Lawler, 1979) . Smaller values of 8 are
characteristic of samples that contain a relatively large
proportion of coarse-grained particles. Samples containing
higher concentrations of fine-grained particles exhibit steeper
slopes for log-log plots of dN/d(d ) vs. d , and thus, higher
values of 3 .
Lawler (1979) used Equation A. 4, known as the power law
function, and the assumption that particles were spherical in
shape to derive the following relationships:
(4-6)
ddogd ) 6 p
= 2.3TT A - d - (A. 6)
ddogd ) p
dN
d(logdp)
= 2.3 A • dr,(1~6) (A.7)
These equations can provide insights into the significance of 6
values in terms of the distribution of particle volume, surface
area, and number among the range of particle diameters
investigated. The results are particularly interesting when the
exponent of d in one of these equations is equal to zero (i.e.
when 8 = 1, 3, or 4). Plots of the left-hand sides of Equations
A.5-A.7 versus log d for 6 values of 1, 3, and 4 are shown in
Figure A-2.
A suspension that can be descibed by the power law function
with an exponent, 8 , of 1.0 would have an equal number of
particles in each logarithmic size interval. This is
demonstrated by the horizontal line for particle number
144
-------
CO
CO
E
u
CO
o
01 I
^ .o E o E
«- 7 3,7 a.
a.
•a
01
o
O.
•a
o»
a
a.
-a
or
o
D = =>
m CD 03
to co in
5 Q 5
o: LU
LU ^
5« iS
UJ
-3 O O O
^^ ^"* ^^^ ^^^ ^^
X (/) X -> X
SURFACE
NUMBER
.5 0 0.5 1.0 1.5
LOG OF PARTICLE DIAMETER
n
Figure A-2. Significance of 6 values (from Lawler,
1979).
145
-------
distribution in Figure A-2a. In such a suspension, the coarser
sized particles would make a greater contribution to the total
particle volume and surface area than the smaller diameter
particles.
From Figure A-2b, it can be seen that, for a sample with a
particle size distribution characterized by a 6 value of 3.0,
the total particle surface area is uniformly distributed among
the various size intervals. In this case, the total number of
particles in the suspension is concentrated in the fine-grained
region of the particle size distribution, while the coarse-sized
fractions account for the majority of the total particle volume.
Finally, as shown in Figure A-2c, a particle suspension
exhibiting a 8 value of 4.0 would have the total particle volume
equally distributed throughout all of the logarithmic size
intervals. Also, both the total particle number and surface area
would be predominately associated with the finer-sized fractions
of the size distribution. Thus, experimental evaluation of the
power law exponent, 8 , for a sample can provide useful
information on the contribution of various particle size
(diameter) fractions to the total volume, surface area, or number
of particles in the suspension.
146
-------
APPENDIX B
DETERMINATION OF DIFFUSION COEFFICIENT FOR THE DCDA
Two experiments were conducted to provide data for the
calculation of a diffusion coefficient for the DCDA reactors used
for bioassays in this study. In both cases, a series of DCDAs
was assembled and the dark side of each was filled with deionized
water and spiked with a known concentration of soluble
orthophosphate. For the first experiment, the opposite side of
each reactor was filled with deionized water containing no
soluble phosphate. In the second experiment, 400 ml of a one
week old P-starved (25 /ugP/1) culture of SslgBSStziHQ.
£3pri£fijr.n.y£.um was added to the light side. The DCDAs were
stirred continuously on a magnetic stir table for various lengths
of time. The reactors were then sampled and the soluble P levels
in both vessels were determined. Diffusion rates were calculated
according to the mathematical analyses presented below.
FIRST EXPERIMENT
A mass balance on P in the spiked vessel yields:
dPs/dt = - a (Ps - Pu) (B.I)
where PS = soluble phosphorus concentration on the spiked side
P = soluble phosphorus concentration on the unspiked side
and a = the bulk diffusion coefficient (day" ) .
Since no soluble P sink was present in the first diffusion
experiment, the total mass of soluble P must be conserved. So,
Vs Ps + Vu Pu = Vs Po (B'2)
where V_, V,, = volumes of spiked and unspiked vessels,
s u
respectively;
147
-------
and P = initial soluble P concentration in spiked side
(/Kj/1)
Since V^ = V = 400 ml, P + P = P , or
S u S u O
Pu = PQ - Ps (B.3)
Thus, Equation B.I can be rewritten as
dP /dt = - a (2P= - P J (B.4)
S SO
or, dP /dt + 2 a P = a P (B.5)
S 5 w
Using the initial condition that P = P at t = 0 to evaluate the
' so
constant of integration, this linear first-order differential
equation is found to have the solution:
Ps = PQ/2 (1 + e~2ct fc) (B.6)
Based on Equation B.6, a plot of In (2P /P - 1) versus t should
s o ,
have a slope of -2 a .
The results of the first diffusion experiment are presented
in Table B-l and are also shown graphically in Figure B-l. These
data indicate that diffusion is consistent from one DCDA to the
next and that the vessels are approaching an equilibrium level of
125jjgP/l at a rapid rate. Regression analysis of
In (2PS/PQ - 1) vs. t yields a straight line with a slope
= - 4.131 day" and a correlation coefficient, r = - 0.971.
Therefore, the bulk diffusion coefficient, a , for these DCDA
systems was 2.065 day . Predicted values of P based on
S
equation B.6 with a = 2.065 day" and PQ = 250 ^ug/1 are plotted
along with the observed data in Figure B.I.
148
-------
TABLE B-l. RESULTS OF FIRST DCDA DIFFUSION EXPERIMENT
Measured SRP
cone, at sampling
DCDA Time of sampling Spiked, side Unspiked side In ( (2Pe./P^)-1)
f hours days (PQ) (Pn)
-
1
2
3
4
. 5
0
4.75
4.75
7.0
10.0
10.0
0
0.198
0.198
0.292
0.417
0.417
250 ug/1
175.6
170.. 5
159.9
152.1
142.0
0 ug/1
76.7
74.4
88.2
98.6
105.6
0
-0.904
-1.011
-1.276
-1,529
-1.995
- Initial SRP concentration on spiked side, P = 250 ug/1.
- Contents of unspiked side - Deionized water.
149
-------
250
CL
o>
200
< 150
d
o
Q
— 100
07
ID
tr
o
CL 50
05
O
Prediction with a-2.065 days"
Spiked Side O
Equilibrium Level
Unspiked Side D
2.5
5.0
7.5 10.0
12.5
TIME (hours)
Figure B-l. Results of first DCDA diffusion experiment.
150
-------
SECOND EXPERIMENT
For the second experiment, in which the unspiked sides of
the DCDAs contained P-starved Sslsn&stiUK , Equation B.I again
results from a mass balance on soluble P in the spiked vessel.
That is,
dP/dt = - a (P - P ) (B.I)
S 5 U
However, since the algae represent a sink of soluble P, it is not
possible to simplify this using the conservation of mass equation
as in the first experiment. Equation B.I can be rearranged to
the form:
dP /dt + a p = a p (B.7)
S S U
Assuming that the algal culture maintains PU at a relatively low
and constant level, this may be solved as a linear, first-order
differential equation. Again using the condition that PS = PQ at
t = 0, the solution to this equation is found to be:
(Ps - V = {Po - V e"at (B'8)
The bulk diffusion coefficient, a , may be determined from a plot
of In [(P_ - PM)/(P_ - P..)] versus t. This plot should yield a
5 U O tl
straight line with a slope of - a .
The results of the second diffusion experiment are presented
in Table B-2. The difference between the initial spike of 125
,ugp/l and the sum of the last two columns in Table B-2 represents
the amount of phosphorus taken up by the algae on the unspiked
side of the DCDAs. Based on the nonzero SRP concentrations
observed in the unspiked vessels, it appears that as a result of
the large P spike and high diffusion rate, P was supplied to the
algae faster than it could be taken up. The value of PU does
decrease steadily with time, however, indicating that this
condition is a transient one.
From the data in the last column of Table B-2, the
assumption that P = constant = 10 xug/l appears to be reasonable
151
-------
TABLE B-2. RESULTS OF SECOND DCDA DIFFUSION EXPERIMENT
DCDA
#
1
2
3
4
. 5
6
7
8
Incubation
time
(hours)
4.42
4.42
8.26
8.26
12.0
12.0
20.0
20.0
SRP on spiked
side after
incubation, P
(yg P/D
77.5
78.5
62.3
66.2
47.5
50.4
29.1
26.6
SRP on unspiked
side after
incubation, P
(ug P/D u
13.5
22.4
11.8
14.6
13.6
13.6
10.1 .
10. ..1
- Initial SRP concentration on spiked side, P = 125 yg/1.
S
Note: Initial SRP concentration on unspiked side was 0.3 ug/1.
152
-------
for this experiment. Thus, a plot of In [(P - 10)/(P - 10)]
s o
versus t should give a straight line with a slope of - a .
Plotting the data in Table B.2 in this manner results in a bulk
diffusion coefficient, a = 2.105 day" and a linear correlation
coefficient, r = -0.9917. This value is very close to that
obtained in the first diffusion experiment.
MOLECULAR DIFFUSION COEFFICIENT
The bulk diffusion coefficient, a , for the DCDAs may be
ed to
expression:
related to the molecular diffusion coefficient, D , by the
a = Dm A/ a V (B.9)
2
where D = molecular diffusion coefficient (cm /day)
2
A = open pore area in the membrane (cm )
a = thickness of the membrane (cm)
and V = volume of liquid in the spiked vessel (cm ).
By rearranging this to the form of Equation B.10, D may be
calculated from the value of a obtained in the diffusion
studies.
D = a a V/A (B.10)
For the DCDA reactors used in this study, V = 400 cm . The
Nuclepore polycarbonate membranes used to separate the two
vessels of the DCDAs have a thickness, a = 10~ cm and a pore
8 2
density of 10 pores/cm . The value of D calculated using this
—1 2 —6
information and a = 2.065 day is 0.218 cm /day, or 2.52 x 10
2
cm /sec, which is definitely within the typical range for
molecular diffusion of salts in water (DePinto, 1982) .
153
-------
APPENDIX C
DERIVATION OF AN EQUATION FOR PHOSPHORUS UPTAKE BY ALGAE
ON THE ASSAY SIDE OF DCDA REACTORS (BASED ON DEPINTO, -1982)
A mass balance for so liable phosphorus on the sediment vessel
of the DCDA results in the following equation:
.Accumulation rate. _ .Rate of release, _ .Rate of diffusion^
of soluble P from sediments across- membrane
or
-dC
V = W(t) - VaC- (C.I)
where C, = Soluble P concentration in sediment vessel
W(t) = Rate of P release from sediments (yg/day) .
V = Sediment vessel volume (I) . •
a = Bulk diffusion coefficient for soluble P across
the DCDA membrane (days"~l) .
If the bioavailable sediment P is released at a first-order rate
in the. sediment vessel, then
Prel(t> =
where P ,(t) = the quantity of P released form the sediments at
time t (ugP/g sediment)
P , = the total amount of P released from the
sediments at t=°° (ygP/g. sediment)
and k = the first-order sediment P release rate (days" ),
The rate of release is given by:
dP x(t) -k t
W(t) = m —i|| = m kr Pult e r (C.3)
154
-------
where m = total mass of sediments in the sediment, vessel (g).
Substituting Equation C.3 into C.I and rearranging terms:
dCx -k t
V — -f VctC = m k ' P e (C 4)
dt 1 r Ult ^'*>
C1V
Since Q, = • = the quantity of P that has been released from
the sediments, but remians in the sediment
vessel (ugP/g sediment)
if all terms in Equaiton C.4 are divided by m, the result is
dQ, -k t
dlT + aQl - kr Pult e (C'5)
Applying an integrating factor of ea gives the solution:
Pult -krt -at (c '
e + c e (C-5)
l ~ (a-kr)
Using the initial condition that Q,=0 at t=0 to solve for C,
this becomes
'- «-ot>
i
Performing a mass balance for particulate (algal) P on the
assay side of the DCDA yields:
.Rate of accumulation. _ .Rate of diffusion.
of algal P across membrane
or
dC2
V dt^ = VctCl (C*8)
155
-------
where C~ = the concentration of algal particulate P in the
assay vessel that originated as sediment P in the
sediment vessel (ug/&).
Using the relationships:
C2V
= Q- = the quantity of P released from the sediments and
.taken up by the algae on .the assay, side (ugP/g
sediment)
ClV
and Ql = — '
if both sides of Equation C.8 are divided by m, then
= aQ, (C.9)
Substituting Equation C.7 into C.9 gives:
Pult -V _ akr Pult -at
e • e
dt (a-kr) c (a-kr
Integration and application of the initial condition Q2=0 at t=0
yields the solution:
n - Pult . -at.
Q
'2vl" (a-kr) ^ ' ' (a-kr]
For cases where a » k ,
a 51 and -; r—r a 0
(ct-kr) - * — (a-kr)
and Equation C.ll may be simplified to
Q2(t) =
156
-------
The right-hand side of this equation is identical to Equation
C.2. This indicates that, provided the diffusion rate is much
greater than the sediment P release rate, algal uptake of P on
the assay side can be used directly as a measurement of sediment
P release in the regeneration vessel.
157
-------
TABLE D-l.
RESULTS OF PHYSICAL AND CHEMICAL MEASUREMENTS ON
UNFILTERED TRIBUTARY SAMPLES AND FILTRATES
tn
oo
Tributary
and
sample
no . - •
Maumee River
5
8
13
15
17
19
21
27
41
42
50
Flow*
(m3/s)
•
803.
872.
691.
118.
348.
898.
339.
578.2
246.7
345.2
1878.0
PH
7.35
7.80
8.13
8.13
7.84
7.82
8.11
Filtrate
alkalinity
(mg/fc as
CaCO3)
105.9
91.1
137.6
166.3
96.5
Suspended
(rag/
Total1
390.
421.
278.
55.5
111.
453.
348.
146.
146.2
149.0
918.
solids
A)
%<20ym
97.8
96.6
93.9
83.7
98.6
45.9
Phosphorus cones.
TP TSP
841.
602.
437.
188.
266.
668.
496.
312.5
279.0
272.0
1152.0
133.3
97.5
102.5
103.
99.9
145.3
124.4
114.4
100.9
120.0
88.4
SRP
109.4
88.2
93.1
95.7
107.8
120.5
99.5
88.1
91.1
111.2
66.5
Sandusky River:
2
7
11
16
18
20
24
25
26
28
30
31
34
35
36
566.
362.
245.
59.1
73.5
257.
22.5
60.1
240.2
161.8
43.3
85.1
101.4
102.0
50.8
7.60
7.69
7.60
7.83
7.62
8.22
8.30
48.5
69.3
94.0
103.9
30.2
153.4
130.7
1028.
690.
557.
535.
72.
1020.
183.
432.
520.
247.
105.
131.
220.
232.
205.
96.1
96.1
96.2
94.8
95.0
77.8
83.0
74.3
61.8
83.2
77.2
1350.
875.
624.
686.
165.
1030.
302.
530.
749.
440.5
196.5
291.5
365.5
398.5
362.5
129.0
91.5
77.4
97.3
60.8
74.8
107.8
86.1
125.6
120.6
56.1
91.0
90.4
82.6
86.2
99.9
83.3
59.2
83.9
57.7
56.8
95.6
77.1
96.9
104.8
40.7
77.5
71.0
69.5
77.3
(continued)
-------
TABLE D-l (continued)
Ul
V£>
Tributary
and
sample
no .
Flow*
(m3/s)
PH
Filtrate suspended solids
aT /a1 Y (mg/i) Phosphorus cones
(rag/ X- as -t-
CaCOs) Total %<20ym TP TSP
. (yg/£)
SRP
Sandusky River:
37
38A
38
39
40
A A **
44
45
46**
47
49
51
52
31.0
85.5
122.2
136.0
189.4
419.9
383.1
282.2
206.3
630.3
576.2
521.9
123.
129.
212.5
268.7
223.7
3753.
1432.
1095.
574.
. 3120.
1002.
838.
248.5
204.5
359.0
425.0
366.0
3890.0
1581.0
1271.0
831.0
3222.0
1205.0
1023.0
56.6
73.3
83.9
80.8
86.7
59.4
72.9
83.1
81.2
50.4
36.2
46.4
Cuyahoga River:
1
4
12
14
29
Cattaraugus
6
9
10
22
23
142
133
130
10
127
(S.
64
32
8
12
30
•
•
•
.8
.0
Br. ) :
.3
.6
.10
.4
.0
7
7
7
7
7
7
8
8
7
.55
.37
.60
.71
.63
.38
.05
.26
.70
66
67
62
47
43
65
100
82
.3
.3
.4
.5
.6
.3
.0
.7
590.
278.
195.
12.0
206.
4325.
647.
150.
70.8
. 66.9
58.5
52.9
48.2
64.8
75.3
716.
446.
303.
275.
321.0
2825.
346.
101.
39.5
51.4
75.3
160.
63.4
11.4
14.6
10.8
16.6
16.1
45.8
62.6
, 63.0
66.5
83.4
47.5
61.5
62.5
68.4
41.7
22.7
30.4
i
32.1
30.7
47.8
157.7
48.1
11.4
4.2
3.9
5.0
6.9
(continued)
-------
CTl
O
TABLE D-l (continued)
Tributary
and
sample
no.
Flow*
(m3/s)
Filtrate
alkalinity
(mg/fl. as
pH CaCOi)
Suspended solids
(mg/Jl)
Total1 %<20um
Phosphorus
TP
cones.
TSP
SRP
Genesee River:
3
7.35 80.2
156. 88.5
192.
37.0
18.3
Honey Creek;
32
33
43
48
25.
14.
53.
112.
8
6
7
0
404.
220.
815.
2130.
89.
94.
1
5
679.
398.
1115.
2094.
0
5
0
0
165
98
75
46
.2
.6
.4
.1
145.
79.
56.
38.
0
2
4
8
* Values given are instantaneous flows at time of sample collection, except those for
the Cuyahoga River, which are average daily flows.
t Amount retained on a membrane filter with a pore size of 0.45 pm.
** Due to leaks in the cubitainers in which samples #44 and 46 were shipped, the total
suspended solids and TP concentrations listed are probably somewhat high.
-------
TABLE D-2.
RESULTS OF P FRACTIONATION STUDIES ON TRIBUTARY
SUSPENDED SEDIMENTS
CTl
Sample
no.
Maumee River
5
8
13
19
21
27
41
42
50
Mean (1980)
Mean (1981)
Mean (All)
T-Sed-P
1,181
1,362
1,118
1,286
1,088
1,682
1,205
1,258
1,149
1,207
1,323
1,259
(TP-TSP)
SS
1,815
1,198
1,203
1,154
1,068
1,357
1,218
1,020
1,159
1,244
Mean (% of T-Sed-P)
Sandusky River
2
7
11
16
20
25
26
28
30
31
34
35
36
37
38A
1,253
1,319
1,071
1,209
940
1,057
1,299
1,366
1,405
1,373
1,232
1,260
1,276
1,309
1,316
1,188
1,136
981
1,100
936
1,028
1,199
1,295
1,251
1,531
1,250
1,362
1,348
1,560
1,017
T-NaOH-P
289
412
363
304
278
710
438
428
406
329
495
403
32.0
451
422
367
395
318
232
479
517
578
637
563
525
406
509
498
Concentrations in ygP/g sediment
R-NaOH-P NR-NaOH-P CDD-P HCl-P
203
284
225
227
223
376
315
312
289
232
323
273
21.7
259
287
249
249
227
155
379
389
369
355
359
290
265
319
332
86
128
138
77
56
334
123
116
117
97
172
131
10.4
192
135
118
146
91
77
100
128
209
282
204
235
141
190
166
292
244
218
250
171
301
211
205
252
235
242
238
18.9
—
215 .
217
272
143
337
262
270
245
335
238
249
298
251
238
107
73
126
125
75
141
123
195
87
101
136
117
9.3
75
55
58
53
43
56
138
118
105
105
82
87
75
80
101
Residual P
121
155
69
131
134
162
278
275
240
122
239
174
13.8
165
143
123
30
102
89
297
312
125
146
129
124
147
143
300
(continued)
-------
TABLE D-2 (continued)
Saitiple
no.
T-Sed-P
(TP-TSP)
SS
T-NaOH-P
Concentrations in
R-NaOH-P NR-NaOH-P
pgP/g sediment
CDB-P HC1-P Residual P
Sandusky River (cont'd.)
Mean
Mean
Mean
Mean
38
39
40
44
45
46
47
49
51
52
(1980)
(1981)
(All)
(% of
1,214
1,241
1,239
936
1,018
1,055
1,172
974
1,078
1,100-
1,141
1,203
1,188
T-Sed-P)
1,295
1,281
1,249
1,021
1,053
1,085
1,306
1,017
1,166
1,165
1,193
508
516
542
336
393
406
465
350
407
394
364
475
449
37.7
373
357
369
208
277
284
330
229
278
256
238
317
298
25.1
135
159
173
128
116
122
135
121
129
138
126
158
151
12.7
211
212
217
120
180
225
278
217
232
235
237
238
237
20.0
113
118
100
58
57
55
71
44
48
46
57
83
77
6.5
291
299
287
285
295
268
185
196
194
202
109
221
194
16.3
Cuyahoga River
Mean
Mean
Mean
1
' 4
12
29 .
(1980)
(All)
(% of
1,294
1,310
1,229
1,426
1,278
1,315
T-Sed-P)
1,147
1,419
1,168
1,250
1,246
518
513
593
677
541
575
43.8
339
420
442
568
400
441
33.6
178
92
152
1*09
141
133
10.1
—
322
287
230
304
280
21.3
196
176
200
180
191
188
14.3
108
77
6
102
64
73
5.6
Genesee River
% of
3
T-Sed-
957
P
994
240
25.1
170
17.8
69
7.2
187
19.5
273
28.5
8.0
(continued)
-------
TABLE D-2 (continued)
CTl
U)
Sample
no.
T-Sed-P
(TP-TSP)
SS
T-NaOH-P
Concentrations in ygP/g sediment
R-NaOH-P NR-NaOH-P CDB-P HC1-P Residual P
Cattaraugus Creek
6
9
10
22
23
Mean (All)
Mean (% of
604
673
733
624
549
637
T-Sed-P)
651
512
601
588
111
79
76
40
73
76
11.9
59
51
51
26
43
46
7.2
52
28
25
14
30
30
4.7
57
88
70
105
88
82
12.9
242
308
348
303
323
305
47.9
49
55
46
44
47
48
7.5
Honey Creek
32
33
43
48
Mean (All)
Mean (% of
1,346
1,367
1,054
988
1,189
T-Sed-P)
1,272
1,363
1,276
961
1,218
588
585
449
398
505
42.5
426
423
322
254
356
30.0
162
162
127
144
149
12.5
201
262
152
198
203
17.1
76
85
37
27
56
4.7
307
305
263
184
265
22.3
Detroit River
*
% of T-Sed-
1,424
P
596
41.9
305
21.4
291
20.4
177
12.4
323
22.7
170
11.9
* Sediments were composited from 4 sampling locations.
-------
APPENDIX E
EQUATIONS AND COMPUTER ALGORITHM USED IN CALCULATING
TWO-COMPONENT SEDIMENT' P RELEASE COEFFICIENTS
In this, study, two-component release of sediment P was first
assumed to follow the equation:
Prel(t) = Pr(l-e ) + Psd-e ) (11)
where prel ^ ~ t^e aroount °f sediment P released (in
sed) at time t.
P /P = the amount of sediment P released (in ugP/g
sed) via the rapid and slow mechanisms,
respectively.
and krr/Jcsr = tlie first~orc^er release rates for the rapid
and slow mechanisms (in day'1).
t = time (days).
A computer algorithm was developed to calculate the four release
coefficients (P , P , k and. k ) from the bioassay data
(values of Prel versus t). Since many of the k values obtained
were comparable in magnitude to the bulk diffusion rate
(a=2.065 day~ ) measured for the DCDA reactors, the rapid
component (P ) was assumed to be released instantly and Equation
11 was rewritten in the form:
Prel(t) = pr + Ps(1"e
The rate of sediment P release, W(t), is given by:
W(t) = m - = m ksr Ps e (E.I)
164
-------
where m = the mass of sediments in the DCDA (g) .
and W(t) has units of yg/day.
If the. soluble P concentration on the assay side of the DCDA is
assumed to be maintained at zero by the P-starved algal culture,
a mass balance for soluble P performed about the sediment, side of
the DCDA yields:
where C, = the soluble P concentration on the sediment, side
V = the volume of the. sediment vessel (i).
and a = the bulk diffusion coefficient for the DCDAs (day )
or,
3F1 + acl " " V ^ ^^^
C1V
Setting Q, =
i m
where Q, = the quantity of P released from the sediments, but
remaining in the sediment vessel (ygP/g sed).
results in the equation:
Ps
This is a linear differential equation that may be solved by
applying an integrating factor of ea . The resulting solution
is (Ross, 1974):
165
-------
-k t
Ql = e"at[/ eCtt ksr Ps e Sr dt + C]
where C = constant of integration.
-at , (a-ksr}t
Or, Qx = e at[/ ksr Pg e sr dt + C] (E.6)
Performing the integration yields:
n sr s Q sr . r a-at « -.
Ql = (a-ksr) e + Ce (E'7)
Using the initial condition: Q, = P at t=0 to solve for the
integration constant gives
Jc P
Substituting this into Equation .7:
C_ may be defined as the concentraiton (ugP/£) of particulate
P in the assay vessel attributable to release from the seidments
(i.e., C_ = total PP-initial PP at t=0). Assuming that soluble
P is taken up immediately by the assay algae, a mass balance for
C- about the assay vessel of the DCDA yields:
dC2
166
-------
Multiplying both, sides by V/ra and letting Q_ = VC2/m, where
Q2 = cumulative uptake of sediment P by algae (ygP/g sed) :
dO
k P — k t
sr
Integrating both, sides
"a P
_
Q2 ~ (a-k
Si
ksr Ps -at _ -at
(a-k ) r e
S JT
where C = constant of integration.
Since C- = 0 at t=0 ,
a P
C =
C
k P
- - - - sr - 5.
(o-Jcgr) (a-ksr)
and, substituting into Equation E.12,
The computer algorithm used to calculate the two-component
release coefficients sought to optimize the correlation between
the measured values of Q2 (from the bioassays) and the predicted
values given by Equation E. 14. The procedure involved an
iterative search for the value of P
Since Pult = Q2 at t==°, from Equation E. 14,
(=P + P ).
IT s
a P
ult
167
-------
and
• TH=CT e -ifidre +pre (E-16>
S JT S.L
If a » k , at high values of t, e~a s o and a/(a-k ) s l, so
o^ S i
ln(Pult-Q2) = In Pg - ksrt (E.17)
Therefore, if the early data points are eliminated from the
bioassay data, a plot of In(P lt~Q2) versus t should have a
slope of -k and an intercept equal to In P . Then p can be
5 i S i
obtained from the equation P = P ,.. - P . A listing of the
computer program used to execute these steps is also presented
in this Appendix.
168
-------
20 PX = 0 : PY = 0 : r:=G : KY = 0 : KZ = 0 : :'M = Q
30 PT=0 : Pl=0 : ?>0 : Kl=0 : K2 = 0 : ?. = 0
40 DIM P(20) ,T(20) ,XCO) ,Y(20)
60 REM PHOSPHORUS AVAILABILITY KINETICS: ONE INSTANTANEOUS COMPONENT
65 REM AND ONE SLOWER FIRST-ORDER COMPONENT.
70 REM READING IN INFORMATION THAT STAYS THE SAKE FOR ALL SAMPLES
80 REM CC = CONVERGENCE CRITERION (ACCEPTABLE INCREMENTAL INCREASE
85 REM IN CORRELATION BETWEEN PREDICTED AND OBSERVED P UPTAKE)
90 REM IGMIN = MINIMUM NUMBER OF POINTS TO BE IGNORED IN
100 REM CALCULATING SLOW RELEASE COEFFICIENTS
110 REM IGMAX = MAXIMUM NUMBER OF POINTS TO BE IGNORED
120 REM NOSAM = NUMBER OF SAMPLES TO BE ANALYZED
125 REM ALPHA = THE DIFFUSION COEFFICIENT FOR THE DCDA'S
130 READ CC,IGMIN,IGMAX,NOSAM,ALPHA
135 PRINT "INPUT DATA: CC=";CC;" IGMIK=" ; IGMIN; M IGMAX=";IGMAX;
136 PRINT " NOSAM=";NOSAM
140 REM OPENING UP A FILE TO STORE RESULTS
150 OPEN "0" fn,nI^RELCOn
160 REM PROGRAM WILL NOW READ IN DATA ON ONE SAMPLE AT A TIME,
170 REM PERFORM CALCULATIONS, STORE THE RESULTS IN A FILE NAMED
180 REM "INRELCO", AND GO ON TO THE NEXT SAMPLE.
190 FOR K = 1 TO NOSAM
200 REM NS = SAMPLE I.D. NUMBER; N = NUMBER OF DATA POINTS
210 READ NS,N
215 PRINT "VALUES READ IN: NS=";NS;" AND N=";N
220 FOR J = 1 TO N
223 REM READING IN TIKE AND CUMULATIVE P RELEASE AT EACH SAMPLING:
230 READ T(J),P(J)
233 PRINT "TIME *";J;'=";T(J);" P RELEASE =";P(J)
240 NEXT J
340 FOR M = IGMIN TO IGMAX
350 IT = 0
360 PX=0 : PY=0 : PZ=0 : KY=0 : KZ=0 : RX=0
370 PT=0 : Pl=0 : P2=0 : Kl=0 : K2=0 : R=0
380 PT = P(N) + 1
390 GOSUB 1140
400 IT = IT + 1
405 REM GENERATING DATA FOR LINEAR REGRESSION TO FIND SLOW RELEASE
406 REM COEFFICIENTS (P2 AND K2) AND ULTIMATE RELEASE (PT).
410 FOR I = M+l TO N
420 Y(-I) = LOGCPT - P (I) )
430 X(I) = T(I)
440 GOSUB 1160
450 NEXT I
460 N = N - M
465 REM NOW CALCULATING LINEAR REGRESSION COEFFICIENTS:
470 GOSUB 1220
480 N = N + M
495 REM NOW CALCULATING SLOW RELEASE COEFFICIENTS (K2 AND P2):
500 REM SLOW RELEASE RATE, K2 (PER DAY):
510 K2 = -1 * Bl
520 REM SIZE OF SLOWLY RELEASED COMPONENT, P2 (UG P/G SED):
530 P2 = EXP(BO)
540 REM NOW CALCULATING THE SIZE OF THE RAPIDLY RELEASED COMPONENT,
545 REM PI, IN UNITS OF UG P/G SED :
550 PI = PT - P2
169
-------
550 IF PI <= o T;:L;N GOTO 770
605 REM NOW CHECKING CORRELATION BETWEEN GBljERVLD AND
610 REM PREDICTED RELEASE:
620 REM DEFINING FUNCTIONS TO 3E USED IN PREDICTING P UPTAKE ON THE
625 REM ASSAY SIDE VERSUS TIME:
630 DEF FNP1(T) = ((ALPHA*P2)/(ALPHA-K2) ) *(1-EXP(-(K2*T)))
631 DEF FNP2(T) = ((K2*P2)/(ALPHA-K2) )*(1-EXP(-(ALPHA*T)))
632 DEF FNP3(T) = PI*(1-EXP(-(ALPHA*T) ) )
640 GOSUB 1140
645 REM CALCULATING THE PREDICTED CUMULATIVE UPTAKE OF P ON THE
643 REM ASSAY SIDE FOR EACH SAMPLING DATE:
650 FOR I = 1 TO N
653 A = FNP1(T(I))
655 B = FNP2(T(IJ)
657 C = FNP3(T(I))
660 Y(I) = A - B - C
670 X(I) = P(I)
675 REM DETERMINING CORRELATION COEFFICIENT (R) BETWEEN
677 REM PREDICTED (Y) AND OBSERVED (X) UPTAKE:
680 GOSUB 1160
690 NEXT I
700 GOSUB 1220
740 REM LIMITING THE MAXIMUM NUMBER OF ITERATIONS TO 100:
745 IF IT = 100 THEN GOTO 790
747 REM CHECKING TO SEE IF CONVERGENCE CRITERION HAS BEEN MET.
748 REM IF NOT, PT IS INCREASED BY 1 AND ROUTINE CONTINUED:
750 IF (R - RX) < CC THEN GOTO 790
760 PX=PT : PY=P1 : PZ=P2 : KY=K1 : KZ=K2 : RX=R
770 PT = PT + 1
780 GOTO 390
790 PT=PX : P1=PY : P2=PZ : K1=KY : K2=KZ : R=RX
795 REM COPYING RESULTS TO DISK:
800 PRINTS!,NS;M;IT;PT;P1;P2;K2;R
810 NEXT M
820 NEXT K
830 CLOSE
840 END
1120 REM THIS IS THE LINEAR REGRESSION PORTION OF -THE PROGRAM:
1140 XX = 0 : YY = 0 : X2 = 0 : Y2 = 0 : XY = 0
1150 RETURN
1160 XX = XX -i- X(I)
1170 YY = YY + Y(I)
1180 X2 = X2 + X(I)~2
1190 Y2 = Y2 + Y(I)"2
1200 XY = XY + X(I) * Y(I)
1210 RETURN
1220 SX = (X2 - XX~2 / N) / (N - 1)
1230 SY = (Y2 - YY~2 / N) / (N - 1)
1240 SP = (XY - XX * YY / N) / (N - 1)
1250 Bl = SP/SX
1260 BO = (YY - Bl * XX) / N
1265 V = B1/ABS(B1)
1270 R = (Bl / ABS(Bl)) * SQR(SP * 2 / (SX * SY))
1280 RETURN
1290 END
170
------- |