6143
DRAFT FINAL REPORT
TO
GREAT LAKES NATIONAL PROGRAM OFFICE
UNITED STATE ENVIRONMENTAL PROTECTION AGENCY
MODELING THE BEHAVIOR AND FATE OF NUTRIENTS
AND TRACE CONTAMINANTS IN THE UPPER GREAT
LAKES CONNECTING CHANNELS
NOVEMBER 1987
INTERAGENCY AGREEMENT DW 13931213-01-0
BETWEEN
GREAT LAKES ENVIRONMENTAL RESEARCH LABORATORY
NATIONAL OCEANIC AND ATMOSPHERIC ADMINISTRATION
ANN ARBOR, MICHIGAN
AND
GREAT LAKES NATIONAL PROGRAM OFFICE
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
CHICAGO, ILLINOIS
-------
U.S. DEPARTMENT OF COMMERCE
National Oceanic and Atmo&^t.ariu Administration
ENVIRONMENTAL RESEARCH LABORATORIES
Great Lakes Environmental Research Laboratory
2205 Commonwealth Blvd.
Ann Arbor, Michigan 48105-1593
December 8, 1987
R/E/GL
Mr. Anthony Kizlauskas, Project Officer
Great Lakes National Program Office
U.S. EPA
230 S. Dearborn
Chicago, IL 60604
Dear Tony,
As part of our interagency agreement (IAG No. DW13931213-01-0), I am pleased
to furnish you with a draft final report on research conducted during the
period December 1, 1984 - November 30, 1987.
We welcome your comments and suggestions.
Sincerely,
Thomas D. Fontaine
Head, Environmental Systems Studies
Enclosure
cc: Beeton
-------
DRAFT FINAL REPORT
TO
GREAT LAKES NATIONAL PROGRAM OFFICE
UNITED STATE ENVIRONMENTAL PROTECTION AGENCY
MODELING THE BEHAVIOR AND FATE OF NUTRIENTS
AND TRACE CONTAMINANTS IN THE UPPER GREAT
LAKES CONNECTING CHANNELS
NOVEMBER 1987
INTERAGENCY AGREEMENT DW 13931213-01-0
BETWEEN
GREAT LAKES ENVIRONMENTAL RESEARCH LABORATORY
NATIONAL OCEANIC AND ATMOSPHERIC ADMINISTRATION
ANN ARBOR, MICHIGAN
AND
GREAT LAKES NATIONAL PROGRAM OFFICE
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
CHICAGO, ILLINOIS
-------
TABLE OF CONTENTS
PAGE
EXECUTIVE SUMMARY 1
INTRODUCTION 6
REPORTS
Unsteady Flow Model of Entire St. Clair River 8
J. A. Derecki, L. L. Makuch, and J. R. Brook
St. Clair and Detroit River Current Measurements 92
J. A. Derecki, K. A. Darr, and R. N. Kelley
Development of a Shallow Water Numerical Wave Model for Lake St. Clair . 135
D. J. Schwab and P. C. Liu
Modeling Particle Transport in Lake St. Clair 145
D. J. Schwab and A. H. elites
Total Phosphorus Budget for Lake St. Clair: 1975 - 1980 161
G. A. Lang, J. A. Morton, and T. D. Fontaine, III
Phosphorus Release from Sediments and Mussels in Lake St. Clair,
with Notes on Mussel Abundance and Biomass 190
T. F. Nalepa, W. S. Gardner, and J. M. Malczyk
Sediment Transport in Lake St. Clair 213
N. Hawley and B. Lesht
Accumulation of Fallout Cesium-137 and Chlorinated Organic Contaminants
in Recent Sediments of Lake St. Clair 289
J. A. Robbins and B. G. Oliver
Toxicokinetics of Selected Xenobiotics in Hexagenia limbata: Laboratory
Studies and Simulation Model 357
P. F. Landrum and R. Poore
Modeling the Fate and Transport of Contaminants in Lake St. Clair . . . 386
G. A. Lang and T. D. Fontaine, III
-------
EXECUTIVE SUMMARY
The following points summarize the findings of our research:
Unsteady Flow Model of Entire St. Clair River
- This is the only flow model for the entire St. Clair River, including its
extensive delta.
- Model provides for flow separation around islands in the upper river and
through the main delta channels in the lower river.
- Three model versions provide information on river stages (profile),
discharge, or velocities along preselected portions or the entire river.
Daily or monthly tabulations of data with corresponding means can be
furnished to users.
- The model furnishes hydrological data needed to compute contaminant mass
balances.
St. Clair and Detroit River Current Measurements
- The measurements represent a unique time series of long-term, continuous
velocity measurements in the Great Lakes connecting channels in the lake.
- The measurements permit evaluation of winter ice effects and weed effects
during most of the year on the flows of the St. Clair and Detroit Rivers.
- The measurements permit evaluation of different types of current meters.
Single-point, contact sensors (electromagnetic meters) present problems
because of weed clogging effects; remote sensors (acoustic Doppler-shift
meters) are ideally suited for use in the river but are expensive. The
acoustic Doppler current profiler measures velocity distribution in the
overhead vertical water column, which could not be duplicated with a
string of point-measuring meters because of ice and navigation problems.
- The measurements provide data for adjusting winter flows in the St. Clair
and Detroit Rivers.
Shallow Water Wave Model for Lake St. Clair
- A shallow water version of our deepwater wave model tends to
underestimate the highest waves at all stations in the lake.
- The deep water version of the model provides quite acceptable estimates
of waveheight, even for the largest waves at the shallowest stations and
is therefore quite acceptable in Lake St. Clair.
-------
The wave model can be used to drive sediment resuspension in contaminant
fate models.
Particle Transport Model for Lake St. Glair
- Although the average hydraulic residence time for Lake St. Clair is about
nine days, residence times of water from individual tributaries range
from 4 to over 30 days in the absence of wind.
- East winds greatly decrease the expected residence time of water from the
Thames River, while northwest winds increase the residence time of water
entering from the St. Clair Cutoff and St. Clair Flats.
- Based on circulation patterns generated by 6 months of real wind
conditions, water entering the lake from the Thames River or one of the 3
lower St. Clair River channels has a large probability of mixing with
waters of the eastern two-thirds of the lake.
- Flows from the upper St. Clair River channels and the Clinton River
follow well-defined routes that are restricted to the western third of
the lake.
Circulation model predictions compare favorably with current measurements
made by drifters and current meters in Lake St. Clair in 1985, although
the model appears to be underestimating the current speeds.
Lake St. Clair Phosphorus Budget 1975 - 1980
- Lake Huron was the major source of phosphorus to Lake St. Clair,
accounting for approximately 52% of the total annual load. Hydrologic
area loads (diffuse and indirect point sources) contributed 43% of the
total load. The remaining 5% came from the atmosphere, shoreline
erosion, and direct point sources.
- The Thames and Sydenham areas of Ontario contributed 75% of the total
hydrologic area load (92% of the total Canadian hydrologic area load).
The Clinton and Black areas of the U.S. contributed 15% of the total
hydrologic area load (83% of the total U.S. hydrologic area load).
Eighty-five percent of the total hydrologic area load was from diffuse
sources.
- Averaged over the six-year study period, estimated external loads were
not significantly different from estimated outflow losses. Therefore,
there does not appear to be a significant net internal source or sink of
phosphorus in Lake St. Clair during 1975-80.
Sediment Phosphorus Dynamics and Mussel Populations
Soluble reactive phosphorus release from Lake St. Clair sediments
averaged 19 ugP/m^/day. This represents about 1% of the total
bioavailable phosphorus load that annually enters the lake from other
-------
sources. Mean maximum release was 47 ugP/m^/day, or about 2% of the
annual bioavailable load from other sources.
Mean mussel density was 2/m^ and mean biomass was 4.3 gDW/m^. Of the 20
species collected, Lampsilis radiata siliquodea and Leptodea fragilaris
were the dominant forms, accounting for 45% and 13% of the total
population. The former species was dominated by older individuals, which
may indicate the population is declining.
The mean excretion rate of phosphorus by mussels was 1.4 ugP/gDW/h. On a
lake-wide basis, this amounts to 5% of the bioavailable load from other
sources.
Sediment Transport in Lake St. Clair
Sediment resuspension in Lake St. Clair is due primarily to wave action.
- The initiation of sediment resuspension can be predicted using a simple
model with wave orbital velocity as the forcing function.
- Critical values of wave orbital velocity for resuspension range from 0.1
cm/s to 0.9 cm/s when calculated one meter above the bed.
- Variations in the critical orbital velocity may be related to substrate
characteristics, but this cannot be proven with the data available.
- The orbital velocities are calculated from the wave field produced by the
GLERL wave model. This model requires over-lake wind velocity
measurements as input.
Cesium-137 and Chlorinated Organics in Lake St. Clair Sediments
Eight percent of atomic fallout Cs-137 that entered Lake St. Clair mainly
in the mid 1960s is still present in sediments 20 years later (1985).
Comparison of total Cs-137 loading with 1985 Cs-137 storage indicates a
sediment residence time of about 5 years. This is consistent with
previously studied loss rates of mercury with chlorinated organics from
the bottom. Changes in total storage of the radionuclide between 1976
and 1985, however, implies a longer residence time of about 15 years.
Probably because of mixing and burial mechanisms, the residence time of
this tracer and other contaminants appears to increase over time.
Resuspension of sediments is continuing to supply radiocesium to the
water where it is exported from the lake.
- Chlorinated organic compounds, like Cs-137, are associated with fine
sediments and preferentially deposit in the deepest water.
-------
The sediments presently contain nearly 1 metric ton of HCB and PCBs and
0.2 metric tons of OCS, far higher contaminant masses than those found in
the St. Clair River.
Based on the behavior of Cs-137, crude estimates of total particle-
associated contaminant loads to Lake St. Clair are: HCB, 15 MT; QCB, 2.8
MT; HCBD, 3.3 MT; OCS, 4.3 MT; PCBs, 20 MT; total DDT, 5.4 MT.
Hexagenia and Contaminants
- Hexaeenia are an important food source for fish in the connecting
waterways of the Great Lakes. Because of Hexagenia's position in the
foodweb, their losses due to toxic effects of contaminants, or their
bioaccumulation of contaminants and subsequent transfer to fish, are
important topics to study.
- Contaminant uptake and depuration rates of Lake St. Clair Hexagenia
limbata vary seasonally. A simulation model suggests that temperature
mediated changes in the depuration rate constant is the term responsible
for most of the seasonal variation.
- The contaminant accumulation rate constant of Hexagenia limbata is
similar to that of other Great Lakes invertebrates, but its depuration
rate constant is much larger.
- Hexagenia limbata obtain a greater percentage of their contaminant body
burden from the sediment than do other Great Lakes invertebrates.
Generic Contaminant Fate and Transport Modeling
- A multi-segment model to simulate the distribution and transport of
contaminants in Lake St. Clair was developed based on the EPA's Chemical
Transport and Fate Model TOXIWASP. The model segmented Lake St. Clair
into 126 well-mixed segments: 42 water segments, 42, active sediment
layer segments, and 42 deep sediment layer segments.
- Simulations of the fate and behavior of chloride, cesium-137, and the
organic contaminants octachlorostyrene (OCS) and polychlorinated
biphenyls (PCB) were carried out.
- The ability of the model to accurately predict the movement and
concentrations of a conservative tracer was substantiated using chloride
data gathered during four, 1974 cruises.
- The ability of the model to accurately predict the distribution and
concentrations of highly partitioned contaminants was tested with Cesium-
137 data. Through calibration of the spatially-varying organic carbon
content of the sediments, the model matched the magnitude and
distribution of observed 1976 cesium-137 in surficial (0-2 cm) sediments.
The calibrated sediment organic carbon content values ranged from 0.14%
to 5% (areal mean = 1.27%) and are well within the range and spatial
distribution of observed values. Simulations were continued out to 1985
-------
for model verification purposes. The predicted distribution of sediment-
bound Cs-137 in the active layer compared well with the observed 1985
values .
Using an estimated load (1.9 Ibs/d) of DCS to Lake St. Clair, the model
was run until simulated DCS levels in the active layer (0-10 cm) agreed
with observed 1983 values. This occurred at 4500 days and implies that
DCS loadings began in the latter part of 1970. This result is consistent
with speculation that OCS was introduced to the lower Great Lakes
beginning in the 1970s. The model predicted 1983 mean and maximum OCS
sediment concentrations of 3.8 and 23.1 pg/kg, respectively. These
compare with observed values of 2.7 and 26.2 jjg/kg. The model predicted
1983 active layer bio-bound OCS levels of 0-96 A»g/kg dry weight, with a
mean concentration of 20 pg/kg. The observed range was 2-154 /ig/kg dry
weight (mean = 43 MgAg) measured in whole clam tissue collected in 1983.
Using observed 1970 PCB levels as initial conditions and an estimated PCB
load (5.1 Ib/d) , the model accurately predicted the 1974 PCB sediment
(0-2 cm) concentrations in Anchor Bay and open- lake sediments. However,
the model tended to overpredict the PCB values along the eastern and
western segments of the main lake, which may indicate additional or
increased PCB sources in these areas. The data showed a decline in mean
lake -wide sediment concentration of total PCB from 19 /ig/kg (max. = 40
/ig/kg) in 1970 to 10 pg/kg (max. - 28 pg/kg) in 1974. The model
simulated a similar 4 year decline in mean sediment concentration from
19.8 to 8.7 pgAg (max. from 39 to 26 pg/kg) . The model predicted a 1974
mean active layer bio-bound PCB concentration of 97 ^g/kg dry wt. (range
- 31-212 pg/kg) . This compares with observed mean values of 90.6 and
44.2 /ig/kg dry wt. (Aroclors 1254 and 1260, respectively) measured in
whole clam tissue in 1983.
-------
INTRODUCTION
[he Upper Great Lakes Connecting Channels Study (UGLCCS) is a multi-agency,
nulti-national study of the St. Marys River, the St. Clair River, Lake St.
Dlair, and the Detroit River. The goals of the study include:
1. Determining the present environmental status of the study area;
2. Identifying and quantifying sources of ecosystem degradation in the
study area;
3. Assessing the adequacy of existing or planned control programs;
4. Developing long-term monitoring programs for assessing the effectiveness
of control programs;
5. Facilitating the development of remedial action plans by the the
Province of Ontario and the State of Michigan.
Towards accomplishing these goals, the Great Lakes Environmental Research
Laboratory (GLERL) of the National Oceanic and Atmospheric Administration (NOAA)
designed modeling, field, and laboratory studies of the connecting channels
study area and processes therein. Through the Activities Integration Committee
or other less formal avenues, GLERL's studies were carefully coordinated with
proposed or ongoing studies of other agencies in order to maximize scientific
insights. A cross reference showing the correspondence between GLERL's studies
and UGLCCS activity numbers is provided in Table 1. In some cases the GLERL
studies were associated with more than one of the UGLCCS activities.
As part of the NOAA - EPA interagency agreement, GLERL provided the Great Lakes
National Program Office (GLNPO) with written quarterly reports describing
progress towards meeting the goals of the study, and oral presentations
summarizing each year's work. Scientists at GLERL are presently submitting the
results of their work to professional scientific journals. This draft final
report documents research conducted during the entire interagency agreement
period, December 1, 1984 through November 30, 1987
-------
?able 1. Correspondence of GLERL Activities with UGLCCS Activities.
5LERL Activity UGLCCS Activity No.
Jnsteady Flow Model of Entire St. Clair River C.5
;t. Clair and Detroit River Current Measurements C.5
)evelopment of a Shallow Water Numerical Wave Model
:or Lake St. Clair C.2
Modeling Particle Transport in Lake St. Clair C.2
[otal Phosphorus Budget for Lake St. Clair: 1975 - 1980 C.I
Phosphorus Release from Sediments and Mussels in Lake St. Clair
vith Notes on Mussel Abundance and Biomass H. 20
Sediment Transport in Lake St. Clair G. 3
Accumulation of Fallout Cesium-137 and Chlorinated Organic
Contaminants in Recent Sediments of Lake St. Clair G.1.G.2
Toxicokinetics of Organic Xenobiotics in the Hexagenia limbata:
laboratory Studies and Simulation Model H. 16
Modeling the Fate and Transport of Contaminants in Lake St. Clair . C.1,C.2,C.3
-------
UNSTEADY FLOW MODEL OF ENTIRE ST. CLAIR RIVER
Jan A. Derecki, Laura L. Makuch, and Jeffrey R. Brook
ABSTRACT
This report describes the development and calibration of an unsteady
flow model for the entire St. Clair River, from Lake Huron to Lake St.
Clair, to simulate hourly and daily flow rates. Unlike previous St. Clair
River hydraulic models that are limited to the upper single-stem river
channel, the present model versions (hourly or daily) provide flow
separation around Stag and Fawn Islands in the upper and middle river, and
through the main delta channels (North, Middle, South, and Cutoff) in the
lower river. The model provides three options for simulation of the river
stages (profile), discharge, or velocities, respectively. This information
is needed in order to predict the fate and transport of pollutants in the
entire river channel, including the island and delta obstructions to its
flow. The model can be run for the entire river or any preselected river
reach bounded by water level gages.
INTRODUCTION
A series of hydrodynamic models have been developed and used
extensively at the Great Lakes Environmental Research Laboratory (GLERL) to
-------
Consequently, interest in and applications of model-simulated flows are
expanding from total flow rates to flow distribution and localized effects.
The principal goal of this Upper Great Lakes Connecting Channels Study
(UGLCCS) activity was to develop the hydrodynamic models for selected upper
connecting channels of the Great Lakes. The St. Clair River was selected
because it forms the upper portion of the outlet through the St. Clair River
- Lake St. Clair - Detroit River system from the upper Great Lakes
(Superior, Michigan, and Huron). This unsteady flow model of the entire
St. Clair River, from Lake Huron to Lake St. Clair, provides flow separation
around Stag and Fawn Islands in the upper and middle river, and through the
main delta channels of the North, Middle, and South Channels and the St.
Clair Cutoff in the lower river (Figure 1). The forcing functions in the
model are the river stages recorded at the water level gages enclosing
preselected river reaches (entire river or longitudinal segments). These
water levels from the extreme gage locations form the model's boundary
conditions.
MODEL DEVELOPMENT
The present unsteady flow model for the entire St. Clair River is an
extension and modification of the existing GLERL upper river model versions.
All these models are driven by water level data taken from appropriate water
level gages along the river. The present model also uses the St. Clair
Shores gage in Lake St. Clair to indicate water levels at the mouth of the
10
-------
river. To permit near real-time model applications only the official
National Oceanic and Atmospheric Administration (NOAA) water level gage data
available at GLERL are used in the model.
Unsteady Flow Equations
The unsteady flow model is based on complete one-dimensional partial
differential equations of continuity and momentum. The momentum equation
includes the effects of motion but neglects the effects of wind stress and
ice. Except for short periods associated with storms, the wind stress
effects were found to be generally insignificant on the St. Clair River
flows in a previous study (Derecki and Kelley, 1981) and the wind data are
normally not available for real-time applications. The effects of
transient ice flows and resulting ice jams in the lower river are
significant and may be substantial during winter and early spring in the St.
Clair River. However, no tested method for including these effects is
presently available.
Expressed in terms of flow Q and stage Z above a fixed datum, the
equations of continuity and motion are as follows:
3t T 3X
1 3Q . 2QT 3Z + (g . Q2T) 3Z + gn^Q/Q/ _ _ 0 (2)
A 3t A2 3t A3 aX 2.208 A2RV3
11
-------
where X - discharge in the positive flow direction
t - time
A - channel cross-sectional area
T - top width of the channel at the water surface
g - acceleration due to gravity
R - hydraulic radius
n - Manning's roughness coefficient
3 - partial derivative function
// - absolute value.
Channel definition is shown in Figure 2. Equations (1) and (2) were placed
in finite difference form at point M in an X-t grid (Figure 3) to yield
respectively,
Zu' + Zd' - Zu - Zd . 6 (Qd' - Qu') + (1-9) (Od - Out _ Q (3)
2 At T AX
Qu' + Qd' - Qu - Qd . QT (Zu' + Zd' - Zu - Zd) +
2 A At A2 At
(g - 02T) . 6 f(Zd' - Zu') + (1-6)(Zd - Zu)1 +
A3 AX
gn2 0/Q/ _ 0 (4)
_2 4/3
2.208 A R
where a prime indicates location and overbars indicate mean, such that
9 = A£l (5)
At
12
-------
Q- 0.5 19 (Qu' + Qd') + (1-9) (Qu + Qd)] (6)
A - 0.5 [8 (Auf + Ad') + (1-6) (Au + Ad)] (7)
Solution of equations (3) and (4) by the implicit method forms the
basis of the numerical unsteady flow model. A stable solution for these
equations is provided by the weighting coefficient 9, which was selected
empirically (Quinn and Wylie, 1972) to be 0.75. Application of the
equations at the river's cross-sections selected to define the actual river
channel produces a set of nonlinear equations that are solved simultaneously
with linear approximations by the Newton-Raphson numerical iteration
procedure. In the initial St. Clair River model version an idealized river
channel, based on averaged river cross-sections for selected reaches, was
used. The use of idealized river channel simplifies simulation of
discharge but prevents valid velocity determination. Description of the
initial St. Clair River model, including calibration, sensitivity analysis,
program listings, and output samples, are given by Quinn and Hagman (1977).
The initial model has been revised by Derecki and Kelley (1981) to replace
the idealized river channel with the actual configurations of the river and
to include wind stress effects.
Mathematical Solution
A schematic diagram for the entire St. Clair River model, including its
delta, is shown in Figure 4. The model can be run for any river reach
containing at least three NOAA water level gages by specifying gage
locations for the beginning and ending boundaries of the reach; the mid-gage
13
-------
is used to check the accuracy of the computed river profile by comparing
deviations between computed and measured water levels. In the previous St.
Clair River models the mathematical solution of the model equations was
provided by using banded matrix, which is most efficient for solving single -
channel configurations.
However, this matrix is impractical for solving flow separations and
was replaced with sparce matrix in the present model. The Yale Sparce
Matrix package available in the GLERL computer library is used in the model.
All the St. Clair River model versions are the hydraulic transient models,
which differ from standard profile or backwater computations; the hydraulic
transient models include time dependent terms of mass continuity and
momentum, which allows the simulation of wave propagation as well as
profiles along the river.
Initial work on the new delta-model development included extending the
existing single-stem upper river model through the middle river (St. Clair
to Algonac), with modification to provide flows around the Stag and Fawn
Islands.
Computations for continuity and momentum around an island start at the
downstream channel junction or node, proceed along one side of the island to
a breakpoint section just below the upstream node, then return to the other
side of the island and proceed to its breakpoint; the breakpoints are then
combined at the upstream node and the computations continue upstream in the
single-stem channel (Figure 5) . Because of mass continuity at the nodal
14
-------
, representing channel separation or confluence, the water level is
;he same for the joint and separate channels, and the flow in the joint
:hannel is the sum of flows in separate channels. This provides additional
continuity equations for the water surface and flows at the nodes, which are
listed below.
WSn - WSsl - WSs2 (8)
Qn - Qsl + Qs2 (9)
The island-model version was extended to Lake St. Clair by treating the
upper delta as an island and the lower delta as a composite of two islands
with separate channels. The composite delta islands are terminated with
short imaginary channels in the lake. This treatment of delta flow
distribution provides separation of flows through the North Channel and
South Channel in the upper delta, and consequent respective separation
through the lower North and Middle Channels, and the lower South and St.
Clair Cutoff Channels in the lower delta, covering all the main delta
channels. To help the program converge on a solution more quickly, the
initial flow values around the islands and the delta are provided in the
model. These initial flows are the fixed percentages of the normal total
flow, based on existing measurements. Nearly all gages used in the model
were moved at various times and some with more recent moves may have
different numbers specified in the model. Occasionally, a gage in the same
location may have experienced a vertical movement, due to some corrective
measure performed on the gage; such corrections for the vertical movement
15
-------
are also specified in the model. These corrections are based on results
determined in previous studies and are included in previous model versions,
where appropriate. Because all physical and hydraulic input data are in
English system, the basic model computations are performed in the English
units and the final result-output converted to the SI system of units, if
desired.
To initialize the computations the model is operated until a steady
state is achieved, prior to simulation of actual data. This is
accomplished by successive iterations of the continuity and momentum
equations in their discrete form (finite difference) until an acceptable
tolerance is attained.
Based on previous models, the number of this "steady state" iterations
is preset at 12. Occasionally, the water level gages break down providing
erroneous or no water level records. However, these records are necessary
for the gages forming the model's boundary conditions. To permit
initiation of computations the missing data are estimated from long-term
means (or previous values within the run), which most likely would not be
sufficiently accurate. Initiation of computations in such cases may
require unreasonably large number of iterations to achieve "fictitious"
steady state (with erroneous results), and the preset number of 12
iterations eliminates such possibility. To make the user aware of possible
inaccuracies (along with causes) , all partial or missing/estimated water
level inputs are flagged in the model outputs.
16
-------
Model computations are performed for each river reach between
successive sections used to define the river channel. For the entire
river, with the delta, this involves 180 cross-sections of the river
channel. Input data for these computational reaches are obtained by
averaging records of the successive bounding sections. Except for the
starting and ending reaches, each reach contains four unknowns, which are
the upstream and downstream water surfaces and flows. The starting and
ending model reaches contain three unknowns because the upstream and
downstream sections in these reaches, respectively, correspond to the water
level gages with known water surfaces (model's boundary conditions). The
model equations are set-up in the sparce matrix as alternating rows of
continuity (odd rows) and momentum (even rows) equations, as shown in Figure
6. The matrix non-zero values are indicated by X's and are the partial
derivatives of the equations for indicated rows with respect to the
variables in indicated columns. The partial derivatives of the continuity
and momentum equations with respect to water surface and flow are listed
below.
-------
_2
3M - 0 T + (g - Q_I) - §_ (14)
5Zu' _2 _3 AX
A At A
3Zd' _2 _3 AX
A At A
3M - 1 - 6 T (Zu' + Zd' - Zu - Zd)
3Qu' _ _2
2 A At 2 A At
e Q T re (Zu' - zd') + a - euzu -zd)i +
_3
A AX
2 _
9 g n 0 (16)
_2 4/3
2.208 A R
3M - 3M (17)
3Qd' 3Qu'
MODEL CALIBRATION
Model Scope
As mentioned in the preceding discussion, the model has two versions
(with separate programs) for the simulation of hourly or daily river
profiles and resulting flows. The model can be operated for the entire
river, with separation of flows around the upper river islands and through
the main delta channels in the lower river, or for any river reach bounded
18
-------
by NOAA water level gages, which form the model's boundary conditions (a
minimum of three gages are employed, with the mid-gage used to provide a
check on simulation accuracy). Because all river profile and flow
information for the entire river will not fit on a computer page, three
separate options are provided for the river stages, discharge, and
velocities, respectively.
Hydraulic Parameters
The hydraulic parameters needed to operate the model are the river
stations, the top channel widths, the datum reference elevations, and the
base areas below the datum for each section used to define the river
channel. Because of the large number of sections (180), printout of this
information is normally suppressed in the model output but is contained in
the program and can be easily reinstated. Other hydraulic parameters
needed to run the model are the water surface elevations for the water level
gages and the roughness coefficients for the river reaches bounded by
successive water level gages (8). All other data needed in the
computations (total channel area, hydraulic radius, length of river reaches,
etc.) are determined from the above data.
Model Calibration
Calibration of the model consisted of adjusting the roughness
coefficients of the river channel, which is the unknown in the flow equation
during periods of flow measurement. The channel roughness coefficients
19
-------
rere determined for each river reach bounded by successive water level
jages, with separate coefficients for the North and South delta channels,
for a total of 8. The roughness coefficients were derived from 14 sets of
flow measurements on the St. Clair River conducted by the Corps of Engineers
during 1959-77. The equation used to compute the roughness coefficients is
the Manning equation, which is
2/3
n - 1.486 A R . (Zu - Zd + 0 AA ) (18)
Q L 3
g L A
The relationships between computed roughness coefficients for the 8
successive river reaches and either upstream or downstream river stages are
shown in Figures 7-14. These relationships normally represent the best-fit
lines derived by regression (least squares) for graphs indicating slope or
the arithmetic means for graphs in which plotted data did not indicate any
slope.
Thus, the roughness coefficient graphs for the upstream reaches (FG-DP
and DP-MBR) indicate positive slope; those for most of the single-channel
river (MBR-DD, DD-MV, MV-SC, and SC-AL) indicate no slope or change in
channel roughness with water level elevation; both delta reaches are between
AL-SCS but are designated AL-ND and AL-SD to indicate north and south delta
channels, respectively, and show negative slopes. In the lower delta four
separate channels are actually used but there were insufficient data to
derive separate roughness coefficients. The downstream river channel, from
the St. Clair City through the delta, was affected by regimen changes
20
-------
between 1959-63, when extensive dredging was conducted for navigation
improvements. For these downstream reaches separate roughness coefficients
were derived for each regime, representing pre-project conditions (through
1963) and current conditions (starting in 1964). These and other St. Clair
River regimen changes are extensively analyzed and documented in various
studies (Derecki, 1985). The computed roughness coefficients actually
represent channel roughness and combined effect of possible errors, such as
those in flow measurements and determination of channel parameters, and the
computed coefficients for some reaches were modified somewhat when such
change was strongly indicated during the model calibration process. Thus,
calibrated roughness does not always represent the best-fit line for plotted
data (Figures 9 and 11). The calibrated roughness coefficients for the 8
river reaches are summarized in Table 1.
The Ft. Gratiot and St. Clair water level gages were moved in 1970,
with apparent uncompensated hydraulic effects. These effects were
determined from a comparison study as a 0.055 m (0.18 ft) reduction for the
Ft. Gratiot gage levels and a 0.027 m (0.09 ft) increase for the St. Clair
gage levels (Quinn, 1976). The Ft. Gratiot gage was modified again in
1981, following blockage of its intake by silt, with apparent uncompensated
hydraulic effect which was determined as an increase in its elevation of
0.037 m (0.12 ft) from the preceding period (Derecki, 1982). Thus,
effective Ft. Gratiot uncompensated hydraulic effect was decreased in 1981
to a reduction of 0.018 m (0.06 ft) in the gage levels. These vertical
gage-record corrections are included in the model to provide unbiased
continuation of the water levels at the gages.
21
-------
Computer Programs
The St. Clair River model for the entire river, with flow separation
around islands and through main delta channels, uses water level data from
GLERL computer disk pack files (VAX) , as did preceding models. Two
generalized versions of the model for simulating hourly and daily flow
rates, respectively, were prepared and stored in the computer files. These
model versions operate on hourly or daily computational time scales and
provide summary tables for daily or monthly data, respectively. Each model
version has three options for the output of simulated river stages at the
gage locations and the total discharge or average velocities at selected
points, as well as separate values around islands and through the delta.
Separate output options were provided since all this information would not
fit on a single computer page. A check of model accuracy is provided in a
form of water level deviations between computed and measured water levels at
the water level gage sites. Basic model computations are listed in the
program in English units; the final results are printed in either English or
SI units, as specified in the output option. The hourly and daily model
versions are listed in the Appendix as Figures A-15 and A-16, respectively.
Examples of model outputs for the water levels, discharge, and velocity
options are shown in the Appendix Tables A-l through A-3 for the hourly
model, and Tables A-4 through A-6 for the daily model, respectively.
22
-------
CONCLUSIONS
This model was developed to correct or eliminate the shortcomings of
•revious model versions. It simulates the St. Clair River profile either
:or the entire length of the river or for selected segments, and provides an
accuracy check for the computed profile at the water-level-gage locations.
rhe model simulates flows (discharge or velocity) in all the more important
river channels, providing flow separation around the islands in the upper
and middle river, and through the main delta channels in the lower river.
As such, it should become a valuable tool for both hydraulic and other water
resource studies in the upper Great Lakes basin.
ACKNOWLEDGEMENT
The authors gratefully acknowledge work performed by D. L. Schultz in
the initial phases of model development.
23
-------
LITERATURE CITED
•erecki, J.A., 1982. Effect of the 1981 Fort Gratiot gage modifications on
the hydraulic regime of the St. Clair River. GLERL Open File Report,
NOAA, Great Lakes Environmental Research Laboratory, Ann Arbor, MI,
3pp.
Derecki, J.A., 1985. Effect of channel changes in the St. Clair River
during the present century. Journal of Great Lakes Research.
11(3):201-207.
Derecki, J.A., and R.N. Kelley, 1981. Improved St. Clair River dynamic
flow models and comparison analysis. NOAA Tech. Memo. ERL GLERL-34,
NOAA Great Lakes Environmental Research Laboratory, Ann Arbor, MI, 36
PP-
Quinn, F.H., 1976. Effect of Fort Gratiot and St. Clair gage relocations
on the apparent hydraulic regime of the St. Clair River. GLERL Open
File Report, NOAA Great Lakes Environmental Research Laboratory, Ann
Arbor, MI, 7pp.
Quinn, F.H., and J.C. Hagman, J.C., 1977. Detroit and St. Clair River
transient models. NOAA Tech. Memo. ERL GLERL-14, NOAA Great Lakes
Environmental Research Laboratory, Ann Arbor, MI, 45 pp.
24
-------
Juinn, F.H., and E.B. Wylie, 1972. Transient analysis of the Detroit River
by the implicit method. Water Resources Research. 8(6):1461-1469.
-------
TABLE 1. Roughness coefficients for the St. Clair River reaches.
Reach Roughness coefficients (n)
FR-DP n - 0.0033947 (FG) - 1.92253
DP-MBR n - 0.0002708 (DP) - 0.12683
MBR-DD n - 0.0221
DD-MV n - 0.0250
MV-SC a. Current regime (starting 1964): n - 0.0240
b. Pre-project regime (through 1963): n - 0.0260
SC-AL a. Current regime: n - 0.0230
b. Pre-project regime: n - 0.0235
AL-ND a. Current regime: n - -0.0017647 (SCS) + 1.04729
b. Pre-project regime: n - -0.0053707 (SCS) + 3.11427
AL-SD a. Current regime: n - -0.0011146 (SCS) + 0.66250
b. Pre-project regime: n - -0.0032907 (SCS) + 1.90968
26
-------
LIST OF FIGURES
1. St. Clair River with location of water level gages.
2. Channel definition sketch.
3. X-t grid for the implicit method.
4. Schematic model diagram.
5. Representation of an island.
6. Sparce matrix.
7. Roughness coefficient for FG-DP reach.
8. Roughness coefficient for DP-MBR reach.
9. Roughness coefficient for MBR-DD reach.
10. Roughness coefficient for DD-MV reach.
11. Roughness coefficient for MV-SC reach.
12. Roughness coefficient for SC-AL reach.
13. Roughness coefficient for AL-ND reach.
14. Roughness coefficient for AL-SD reach.
27
-------
Michigan
SCALE IN MILES
*>
KILOMETERS
20
10
20
30
Ft. Gratioti
Dunn Paper<
Mouth of Black River*
Dry
Ontario
St. Clair Shores
Grosse Point
LAKE HURON
23
-------
^*- Water Mass at t+At
datum
AX
-------
t+At
•
Q'u
Z'u
At
I
1 Qu
Zu
1
1
1
If
t 1
1 —
1
1 Ax *•
Q'd
Z'd
Qd
Zd
u d
X
30
-------
ST. OAIR SHORES
LAKE
ST. OAIR
ttARYStlUE (155)
MT DOCS (IS7
rlOUTH OF BLACK RIVER ( l»0)
DUNN PAPER (176)
ieo ) FORT 6RATIOT
~*-S (UU HURON)
ST CLAIR RIVER SCHEMATIC REPRESENTATION
31
-------
32
j^i;i;;;^;i^1SLAND'1i;;;iii;;^ 10 *—
FLOW »^;!iii;;j:ii!;i;;;;i;;;:i;i;i:!;Hi:iy^ FLOW
-------
c
•
c
•
c
1-2|
1-21
2-31
2-21
2-4 |
2-4 |
Cl
WS2
WS3 C3
x
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
33
-------
682n
581-
580-
579-
678H
577-
576
575-
574
0.02
N= 0.0033947 (FG)-192253
0.03 0.04 0.05
ROUGHNESS (MANNING'S N)
0.06
-------
682-,
581-
N= O.OXD2708 (DP)-0.12683
580-
579 -I
678-
577-
576-
575 J
574
1 1 1 1 1 1 1 1 1 1 r
-i 1 1 1 1 1 r
0.027 0.028 0.029 0.030 0.031 0.032
ROUGHNESS (MAILING'S N)
35
-------
581
580
579-
578-
577-
LU 676
575
574-
573
N = 0.0221
o
o
o
0.020
o
o
0.021 0.022 0.023 0.024
ROUGHNESS (MANNNG'S N)
0.025
36
f
-------
681
580
579-
od
578
3 677-
<
O
ULJ 576
575-
574
:
573
o
N = 0.0250
0.023 0.024 0.025 0.026 0.027 0.028
ROUGHNESS (MANMNG'S N)
37
10
-------
681
580H
679-
5 578-
a :
H
Uj 576-
D
575-
574-
R-7O -
O
N = 0.0240
(Current)
0 0
0 0
X
o
X
0
o
1 1 1 1 1 1 1 1 1
N = 0.0260
(pre-Project)
X
r i 1 1 i i i 1 i 1 i
0.023 0.024 0.025 0.026 0.027 0.028
ROUGHNESS (MANNWG'S N)
38
-------
578n
N - -0.001H46 (SCS) + 0.66250
(Current o)
574-
673-
572-
571-
N = -0.0032907(SCS) + 190968
(Pre-Prqject x)
570
0.018 0.020 0.022 0.024 0.026 0.028
ROUGHNESS (MANNING'S N)
0.030
-------
APPENDIX A.
HOURLY AND DAILY UNSTEADY FLOW MODELS:
1. Model Programs (Appendix Figures A-15 and A-16)
2. Model Outputs (Appendix Tables A-l through A-6)
42
-------
Program [HYDRO.JDSTCLR]HDELTA.FOR
This is the St Clair River Transient Model - Hourly Version.
It is set to run in BATCH MODE...
To run the program...
1. Set desired parameters in file [HYDRO.JDSTCLR]HDELTA.PAR
Line 1 - Starting and ending day, month and yr.
DA MO YR DA MO YR (12,IX,12.IX,12,IX,12,IX.12,IX,12)
Line 2 - Staring and ending points of model. (II,IX,II)
1 - Fort Gratiot
2 - Dunn Paper
3 - Mouth of Black River
4 - Dry Dock
5 - Marysville
6 - St Clair
7 - Algonac
8 - Lake St Clair
Line 3 - Output Option (II)
1 - Water levels and deviations.
2 - Total discharge and discharge around islands and, if
included, discharge in the delta channels.
3 - Velocity near the starting, ending and midpoint of
the simulated river and velocities around islands and,
if included, velocities in the delta channels.
Line A - Units Option (II)
0 - Metric units
1 - English units
2. Make sure that file [HYDRO.JDSTCLR1HDELTA.DAT is available.
3. Type: SUBMIT HDELTA/NOTIFY
4. When your request is completed the output will appear in file:
[HYDRO.J DSTCLR]ZHDELTA.OUT
Note: this file is 132 characters wide.
43
-------
PROGRAM ST_DELTA_HOURLY
c
C This is an hourly version of PROGRAM ST_DELTA_DAILY.
C Programmers QUINN LLM JRB
C
C THIS VERSION USES ROUGHNESS COEFFICIENTS CALCULATED FROM
C . . . '59-'77 VIA GUESSN.FOR, AND COMPARES ALL INTERMEDIATE GAGES.
C ... IT WAS ADAPTED FROM XX2.FOR, AND THEREFORE ZYX AND SCDQMOD TO
C ALLOW FOR ISLANDS; ie SPARGE MATRIX AND NON-CONSEC. STATIONING.
C FOR005) HRISLE.DAT: PHYSICAL DATA (STA,ABAS,DATU,&AT)
C FOR006) THE OUTPUT FILE HRDEL.OUT
C . . . FOR007) SYS$OUTPUT
C . . . FOR010) TT
IMPLICIT DOUBLE PRECISION (A-H.O-Z)
REAL NPERC.MPERC.PERC
LOGICAL NEWROW.MONFLAG
INTEGER RR(358)ICC(358),ESP,PATH,FLAG
CHARACTER*9 NAMMON(12) ,NAME(8)*20,NMM(8)*3,NOTE(8 . 24)*1 ,MARK(181)
>*2
COMMON /WAT001/IHOUR(24,31).MEAN(31),MEM,MAXV(31),MAXD(31) , I FLAG,
> MINH(31).MIND(31),MAXM(4),MINM(4).IC,IGEAGE.MONAA,IYRR,IDUM(142)
DIMENSION AA(180),ABAS(180),DATU(180),AT(180),X(180).STA(180)>
> WS(50>180),Q(55,180),YVECT(358),XMTRX(358,358),T(180),AN(180),
> A(180) ,U(180) ,R(180) ,QA(180) , SUM(44) ,AVE(44) ,ADJ(8) , IGAGE(8) ,
> OLD(8) ,LOC(8) ,DEV(8) ,WSSAV(50,8) ,Al(8) ,B1(8) .NODE(IO) ,NBR(5) ,
>IA(359),AVECT(1378),JA(1378),ICC(358),YV(358),RSP(5722),ISP(5722)
>,VEL(180)
EQUIVALENCE (ISP.RSP)
DATA NAMMON/' JANUARY',' FEBRUARY',' MARCH',' APRIL' ,
> • MAY',' JUNE',' JULY',' AUGUST',
> ' SEPTEMBER',' OCTOBER',' NOVEMBER' ,' DECEMBER' /
DATA NAME/' FT. GRATIOT ',' DUNN PAPER
> 'MOUTH OF BLACK RIVER' . ' DRY DOCK
> ' MARYSVILLE ' , ' ST CLAIR
> ' ALGONAC ','LAKE ST. CLAIR (SCS)'/
DATA NMM/' FG',' DP'.'MBR',' DD',' MV',' SC',' AL'.'SCS'/
DATA LOC/180,178,160,157,155,129,77,I/
DATA IGAGE/14098 ,14096 .14090.14087 ,14084,14080,14070 ,14052/
DATA OLD/580.92 ,580.43,580.09,579.53,579.01,578.16,576 . 60, 576. 23/
C Al IS SLOPE AND Bl IS INTERCEPT OF MANNINGS N'S FOR ALL REACHES
c this is for years 1959-1977
DATA Al/0.0033947,0.0002708,0. ,0. ,0. ,0. ,-0.0011146, -0.0017647/
DATA B1/-1.92253,-0.12683,0.0221,0.0250,0.0240,0.0230,
>0.66250,1.04729/
C
C . . ISLAND/DELTA SPECIFIC VARIABLE ASSIGNMENTS
DATA NODE/3,6.25,39,51,76,97,108,135,154/
DATA NBR/15,36,45,102,144/
EPERC1-.253
WPERC1-1.-EPERC1
EPERC2-.376
44
-------
WPERC2-1.-EPERC2
NPERC-.35
MPERC-.21
CNPERC-.56
FNPERO.56
SPERC-.21
COPERC-.23
CSPERC-.44
FSPERC-.44
NSP-5722
NISLANDS-5
ICOUNT-0
LRATIO-2
IFLAG-0
C
C . . . PHYSICAL DATA ACCESSED
DO 10 I-1,LOC(1)
10 READ(5,1020) STA(I),ABAS(I),DATU(I),AT(I)
C
C . . . PROMPT FOR AND READ BEGINNING AND ENDING DATES
C WRITE(7,2000)
READ (10.1000) MONA,IDAYA,IYRA,MONB,IDAYB,IYRB
IF(MONA.LE.O) THEN
MONA-12
IYRA-IYRA-1
END IF
IYRA-IYRA+1900
IYRB-IYRB+1900
C
C . . . PROMPT FOR LOCATION OF UPPER AND LOWER LIMITS TO BE RUN
C 9 WRITE(7,2020) (I,NAME(I), 1-1,8)
9 READ(IO.IOIO) IUP.IDN
IIDN-IDN-1
IIUP-IUP+1
IF(IUP.GE.IIDN) GO TO 9
C
C . . . PROMPT FOR OPTION NUMBER
C 616 WRITE(7,8001)
616 READ(10,8002)NOPT
IF(NOPT.GT.3.0R.NOPT.LT.l) GO TO 616
C
READ(10,8002) MUNITS
C
C . . . DEFINE AND ADJUST PARAMETERS, BASED UPON LIMITS FROM ABOVE
NRM-LOC(IUP)-LOC(IDN)
NMR-NRM+1
IF(LOC(IDN).EQ.l) GO TO 27
DO 24 I-l.NMR
STA(I)-STA(I-1+LOC(IDN))
ABAS (I)-ABAS(I-1+LOC(IDN))
DATU(I)-DATU(I -l-i-LOC(IDN) )
24 AT(I)-AT(I-1+LOC(IDN))
45 A-ts.z
-------
NODE(1)-NODE(1)-LOC(IDN)+1
NODE(2)-NODE(2) -LOC(IDN)+1
NODE(3)-NODE(3)-LOC(IDN)+1
NODE(4)-NODE(4) -LOC(IDN)+1
NODE(5)-NODE(5)-LOC(IDN)+1
NODE(6)-NODE(6)-LOC(IDN)+1
NODE(7)-NODE(7)-LOC(IDN)+1
NODE(8)-NODE(8)-LOC(IDN)-H
NODE(9)-NODE(9)-LOC(IDN)+1
NODE(10)-NODE(10)-LOC(IDN)+1
NBR(1)-NBR(1)-LOC(IDN)+1
NBR(2)-NBR(2)-LOC(IDN)+1
NBR(3)-NBR(3)-LOC(IDN)+1
NBR(4)-NBR(4)-LOC(IDN)+1
NBR(5)-NBR(5)-LOC(IDN)+1
DO 25 I-IUP.IDN
25 LOC(I)-LOC(I)-LOC(IDN)+1
C
C CALCULATE DISTANCES BETWEEN SECTIONS
27 DO 30 I-l.NRM
X(I)-STA(I+1)-STA(I)
30 IF(I.EQ.NBR(1).OR.I.EQ.NBR(2).OR.I.EQ.NBR(3).OR.I.EQ.
>NBR(4).OR.I.EQ.NBR(5)) X(I)-0.0
C
C The following lines which, have been COMMENTED with "cO", will
C . . . WRITE the BASIC PHYSICAL DATA. Should the user desire to see
C this data, the "cO's" would have to be eliminated and the
C program re-compiled/linked etc. TO RE-COMPILE submit HCDELTA.
C
cO WRITE (6,3000)
cO WRITE (6,3010)
cO DO 40 I-l.NMR
cO MARK(I)-' '
cO IF(I.EQ.NODE(3).OR.I.EQ.NODE(5).OR.I.EQ.NODE(6)
cO >.OR.I.EQ.NODE(7).OR.I.EQ.NODE(8).OR.I.EQ.NODE(9).OR.
cO >I.EQ.NODE(10)) MARK(I)-'<'
cO IF(I.GT.NODE(2).AND.I.LE.NBR(1))MARK(I)-'N '
cO IF(I.GT.NBR(1).AND.I.LT.NODE(3))MARK(I)-'M '
C0 IF(I.GT.NODE(3).AND.I.LE.NBR(2))MARK(I)-'UN'
cO IF(I.GT.NODE(4).AND.I.LE.NBR(3))MARK(I)-'S '
cO IF(I.GT.NBR(3).AND.I.LT.NODE(5))MARK(I)-'CO'
cO IF(I.GT.NODE(5).AND.I.LT.NODE(6))MARK(I)-'US'
c0 IF(I.GT.NODE(7).AND.I.LE.NBR(4).OR.I.GT.NODE(9).AND.I
cO >.LE.NBR(5))MARK(I)-'W '
c0 IF(I.GT.NBR(4).AND.I.LT.NODE(8).OR.I.GT.NBR(5).AND.I
cO >.LT.NODE(10))MARK(I)-'E '
cO IF(I.LE.NODE(2).OR.I.GT.NBR(2).AND.I.LE.NODE(4)) GO TO 40
cO WRITE(6,3020) STA(I),ABAS(I),DATU(I),AT(I),MARK(I)
cO DO 40 IJ-IUP.IDN
cO 40 IF(I.EQ.LOC(IJ)) WRITE(6,3030) NAME(IJ)
C
C THESE NEXT 2 MANNINGS N PRINTOUTS ARE HARDWIRED. IF THE
46
-------
C NUMBER OF STATIONS CHANGES, CHANGE THESE!
C
cO IF(IDN.EQ.S) VRITE(6,3040) Al(8),NMM(8),B1(8),STA(7),STA(36)
cO IFUDN.EQ.8) WRITE(6.3040) Al(7) ,NMM(8) ,B1(7) ,STA(40) ,STA(77)
cO DO 45 I-IIDN,IUP,-1
cO 45 IF(I.LT.7)WRITE(6,3040) Al(I),NMM(I),B1(I),STA(LOC(I+1) ) ,
cO >STA(LOC(I))
C
C . . . INITIALIZE AND ASSIGN ADDITIONAL VARIABLES
DO 50 1-1,4
50 ADJ(I)-0.
NVAR-NRM*2
ANC-1.
DT-ANC*3600.
DO 56 1-1,42
56 SUM(I)-0.
TH-.75
TH1-.25
MM-0
M-13
istart-13
iend-36
KDK-IDAYA
kkz-11
MON-MONA
IYR-IYRA
MONFLAG-.FALSE.
JRB-(IUP+IDN)/2
C
C
C Routine to read all of the water level data froa the disk and
C store it in another tempory file. This way the disk is not tied
C up for long periods of time when running the program.
C
C . . . READ WATER LEVELS FROM DISC.
C
52 CALL NODAYS( IYR.MON.1.NDM.NDY,JD)
DO 55 JJ-IUP.IDN
IW-1
IC-IGAGE(JJ)/10000
IGAG-IGAGE( JJ )-IC*10000
CALL GAGEIO( IW.IC.IGAG ,MON, IYR.IB, IT, IDA, IDB.IDC, IER)
IF(IER.NE.O) THEN
WRITE(6,3110) NAMMON(MON),IYR,IER
CALL EXIT
END IF
DO 54 J-l.NDM
DO 53 1-1,24
WRITE(9,1050) IHOUR(I.J)
53 CONTINUE
WRITE(9,1050) HEAN(J)
54 CONTINUE
-------
55 CONTINUE
C
IF(IYR-IYRB) 58,57,65
57 IF(MON-MONB) 58,65,65
58 CONTINUE
IF(MON-12) 60,59,59
59 IYR-IYR+1
60 CONTINUE
C UPDATE MONTH AND YEAR AND RECHECK IF MORE DATA SHOULD BE USED
MON-MON+1
IF(MON-13) 62,61,61
61 MON-1
62 IF(IYR-IYRB) 64,63,65
63 IF(MON-MONB) 64,64,65
64 CONTINUE
C
GO TO 52
65 CALL DISMOUNTPACK('WATER_LEVELS')
MON-MONA
IYR-IYRA
REWIND 9
C COME HERE EACH DAY AND PRINT TITLES AND HEADINGS FOR
C EACH OPTION
70 continue
if(Nopt.cq.1.and.iup.eq.1) iiup-iiup+1
C*********************OPTION FOR WATER LEVELS********************
IF(NOPT.EQ.l) THEN
WRITE(6,3000)
WRITE(6,3050) NAMMON(MON).kdk.IYR,NAME(IUP).NAME(IDN)
WRITE(6,3060) ANC.NRM
WRITE(6,3070) NMM(IDN).NMM(IUP),
> NMM(IUP),(NMM(I),I-IIDN,IIUP,-1)
WRITE(6,3071) (NMM(I),I-IIDN.IIUP.-1)
IF(NOPT.EQ.LAND. IUP. EQ.DIIUP-IIUP-1
C
C ******************OPTION FOR FLOWS WITHOUT DELTA*****************
C
ELSE IF(NOPT.EQ.2 .AND. IDN.NE.8) THEN
WR1TE(6,3000)
WRITE(6,8051)NAMMON(MON) ,kdk,IYR,NAME(IUP) .NAME(IDN)
WRITE(6,3060)ANC,NRM
WRITE(6,7020)
WRITE(6,7030)NMM(IDN),NMM(JRB).NMM(IUP).NMM(IIDN),NMM(JRB),
> NMM(IUP) .NMM(JRB)
WRITE(6,7040)
C
C*******************OPTION FOR FLOWS WITH DELTA********************
C
ELSE IF(NOPT.EQ.2 .AND. IDN.EQ.8) THEN
WRITE(6,3000)
WRITE(6,8051)NAMMON(MON) ,kdk,IYR,NAME(IUP) ,NAME(IDN)
WRITE(6,3060)ANC,NRM
48
-------
WRITE(6,8020)
WRITE(6,8030)NMM(IDN) ,NMM(IIDN-1) ,NMM(IUP) ,NMM(IUP) NMM(IIDN)
> NMM(IIDN-l)
WRITE(6,8040)
C
C ****************QPTION TOR. VELOCITIES WITHOUT DELTA***************
C
ELSE IF(NOPT.EQ.3 .AND. IDN.NE.8) THEN
WRITE(6,3000)
WRITE(6,9051)NAMMON(MON) ,kdk, lYR.NAME(IUP) .NAME(IDN)
WRJTE(6,3060)ANC,NRM
WRITE(6.9025)
WRITE(6, 7030)NMM(IDN) ,NMM(JRB) ,NMM(IUP) .NMM(IIDN) .NMM(JRB)
> NMM(IUP),NMM(JRB)
WRITE(6,9041)
C
C*****************OPTION FOR VELOCITIES WITH DELTA******************
C
ELSE IF(NOPT.EQ.3 .AND. IDN.EQ.8) THEN
WRITE(6.3000)
WRITE(6,9051)NAMMON(MON) ,kdk,lYR.NAME(IUP) .NAME(IDN)
WRITE(6,3060)ANC,NRM
WRITE(6,9020)
WRITE(6.9030)NMM(IDN) .NMM(IIDN-l) .NMM(IUP) .NMM(IUP) .NMM(IIDN)
> NMM(IIDN-l)
WRITE(6,9040)
END IF
71 Continue
C
C . . . ADJUST GAGE-SPECIFIC VARIABLES BASED UPON YEARS STUDIED
IF(IYR.LT.1970) IGAGE(1)-14099
IF(IYR.LT.1971) IGAGE(6)-14080
IF(IYR.GT.1970) ADJ(l)—.18
IF(IYR.GT.1981) ADJ(l)—.06
IF(IYR.LT.1971) ADJ(6)--.09
C
C
C . . . READ WATER LEVELS FROM UNIT 9. . .
C
CALL NODAYS( IYR.MON,l.NDM.NDY,JD)
C
DO 110 JJ-IUP.IDN
C
DO 77 J-l.NDM
DO 75 1-1,24
75 READ(9,1050,ENI>-78) IHOUR(I.J)
77 READ(9,1050,END-78) MEAN(J)
C
C THE BLOCK OF CODE FOR READING THE WATER LEVELS DISC
C IS DIFFERENT FROM THE DAILY VERSION OF THE PROGRAM.
C
78 CONTINUE
-------
DO 110 J-lstart,l«nd
IF(KXZ.EQ.ll) THEN
I^J-12
ELSE
I-J-1
END IF
C
C . . . FLAG MISSING DATA AND ASSIGN GAGE LEVELS TO WSSAV
IF( IHOUR(I.KDK).GT.O) THEN
HOTE(JJ.KDK)-1 '
ELSE
NOTE(JJ,KDK)-'E'
END IF
if(lhour(i,kdk).le.O) then
wssav(j.jj)-old(jj)
else
wssav(j,Jj)-(ihour(i,kdk)+ib)/100.+adj(;jj)
end if
old(Jj)-wssav(j,JJ)
IF(JJ.EQ.IUP.OR.JJ.EQ.IDN) WS(J ,LOC(JJ))-WSSAV(J ,JJ)
110 CONTINUE
IF(MONFLAG) GO TO 200
C
C . . . SET WS FOR 12 PREVIOUS TIME STEPS, TO ACHIEVE 'STEADY STATE'
DO 120 1-1,12
DO 120 J-1,8
VSSAV(I.J)-VSSAV(13,J)
120 IF(J.EQ.IUP .OR. J.EQ.IDN) US(I,LOC(J))-WSSAV(13,J)
C
C . . . ZERO MATRIX, SET CHANNEL PARAMETERS, & SET INITIAL CONDITIONS.
DO 130 I - l.NVAR
YVECT(I)-0.
DO 130 J - l.NVAR
130 XMTRX(J,I)-0.
XSUM-STA(LOC(IUP) ) -STA(LOC(IDN) )
SLOPE- (USSAV(1 , IUP) -WSSAV(1 , IDN) ) /XSUM
AA(1)-ABAS(1)*AT(1)*(US(1,1)-DATU(1))
DO 150 I-l.NRM
IF(I.NE.NRM) THEN
WS(1,I+1)-WS(1,I)+SLOPE*X(I)
IF(I.EQ.NBRd)) WS(1.I+1)-WS(1,NODE(2))
IF(I.EQ.NBR(2)) WS(1,I-H)-WS(1.SODE(1))
IF(I.EQ.NBR(3)) WS(1,1+1)-WS(1 ,NODE(4))
IF(I.EQ.NBR(4)) WS(1, U1)-WS(1 ,NODE(7))
IF(I.EQ.NBR(5)) WS(1 , 1+1)-WS(1 ,NODE(9) )
END IF
AA(I-i-l)- ABAS(I-»-l)+AT(I-H)*(WS(l,I+l)-DATU(I+l))
150 R(I)-A(I)/T(I)
Al(IIDN)*VSSAV(M,IIDN)+Bl(IIDN)
50
-------
QSTART-1.486*A(l)*R(l)**(2./3.)*(VS(l,2)-WS(1,!>>**. 5/AN(l)
c
c
. SPLIT FLOW INITIALLY AROUND THE ISLANDS FOR FASTER CONVERGENCE
DO 190 I-l.NMR
PERC-1.
IF(I.GT.NODEU) .AND. I. LE.NODE(2)) PERC-FNPERC
I.LE.NBR(l)) PERC-NPERC
I.LT.NODE(3))
I.LE.NBR(2))
I.LE.NODE(4))
. I.LE.NBR(3))
I.LT.NODE(S))
I.LT.NODE(6))
190
200
210
230
C
C
c
IF(I.
IF(I
.AND.
AND.
.AND.
AND.
.AND.
AND.
.AND.
.AND.
AND.
.AND.
AND.
PERC-KPERC
PERC-CNPERC
PERC-FSPERC
PERC-SPERC
PERC-COPERC
PERC-CSPERC
I.LE.NBR(4)) PERC-WPERC1
I.LT.NODE(8)) PERC-EPERC1
I.LE.NBR(S)) PERC-VPERC2
I.LT,NODE(10)) PERC-EPERC2
IF
-------
IU-1D+1
DO 255 J-1,5
JJ-NBR(J)
KK-NODE(2*J)
LL-NODE(2*J-1)
IFdDN.NE.S.AND.ID.EQ.LL.OR.IDN.NE.S.AND.IU.EQ.KX.OR.
> IDN.EQ.8.AND.I.GT.106.AND.ID.EQ.LL.OR.IDN.EQ.8.AND.I.GT.106
> .AND.IU.EQ.KX.OR.IDN.EQ.8.AND.ID.EQ.6.0R.IDN.EQ.8.AND.ID.EQ.
> 24.0R.IDN.EQ.8.AND.ID.EQ.39.0R.IDN.EQ.8.AND.ID.EQ.50.0R.
> IDN.EQ.8.AND.ID.EQ.3) THEN
YVECTd)—(WS(N,IU)+WS(M,IU)-WS(N,ID)-WS(M,ID))/2.
XMTRXd, I-D —0.5
XMTRX(I,I-H)-+0.5
GO TO 260
ELSE IFUD.EQ.JJ.AND.IDN.NE.8.0R.ID.EQ.JJ.AND.IDN.EQ.8
> .AND.J.CE.4) THEN
YVECTd)--(VS(N.KX)+WS(M,KK)-WS(N,JJ)-WS(M,JJ))/2.
XMTRX(I,I-1)--0.5
XMTRX(I,2*KK-2)-+0.5
GO TO 260
ELSE IF(IDN.EQ.8.AND.1D.EQ.NBR(1)) THEN
YVECT(I)—(VS(N,ID)+WS(M,ID)-WS(N,NODE(3))-WS(M,NODE(3)))/2.
XMTRX(I,28) - +.5
XMTRX(I,48) - -.5
GO TO 260
ELSE IF(IDN.EQ.8.AND.ID.EQ.NBR(2)) THEN
YVECT(I)--(WS(N,ID)+WS(M,ID)-WS(N,NODE(6))-WS(K,NODE(6)))/2.
XMTRX(I,70)- +.5
XMTRX(I,150)- -.5
GO TO 260
ELSE IF(IDN.EQ.8.AND.ID.EQ.NBR(3)) THEN
YVECT(I)—(WS(N,ID)+WS(M,ID)-WS(N,NODE(5))-WS(M,NODE(5)))/2
XMTRX(I,88)- +.5
XMTRX(I.IOO)- -.5
GO TO 260
END IF
255 CONTINUE
YVECT(I)— ((WS(N,ID)+WS(N.IU)-WS(M,ID)-WS(M,IU))/(2.*DT) +
> (TH*(Q(N,ID) -Q(N,IU))+TH1*(Q(M,ID)-Q(M,IU)))/(T(ID)*X(ID)))
XMTRXd, I)-TH/(T(ID)*X(ID))
XMTRXd , 1+2)—XMTRX(I, I)
IF(I.EQ.l) THEN
XMTRXd, 2 )-l./(2-*DT)
ELSE
XMTRXd, I-1)-1./(2.*DT)
XMTRXd, 1+1)-!./(2. *DT)
IF(I.EQ.NRD) XMTRX(I,I+l)-XMTRX(I,I-(-2)
IF(I.EQ.NRD) XMTRXd,H-2)-0.
END IF
260 CONTINUE
C
C . . . MOMENTUM EQUATIONS
52
-------
DO 280 I-2.NVAR.2
ID-I/2
IU-ID+1
DO 265 J-1,5
JJ-NBR(J)
KK-NODE(2*J)
LL-NODE(2*J-1)
IF(IDN.NE.8.AND.ID.EQ.LL.OR.IDN.EQ.8.AND J GE 4 AND
> ID.EQ.LL) THEN ' ' '
YVECT(I)-(Q(N,JJ+1)^5(M,JJ+1)4<3(N>IU)4<5(M,IU)-Q(N,ID).Q(M.ID))
XMTRXd.I-l) — 0.5
XMTRX(I,I+1)-+0.5
XMTRX(I,2*JJ+1)-+0.5
GO TO 280
ELSE IF(IDN.NE.8.AND.ID.EQ.JJ.OR.IDN.EQ.8.AND J GE 4 AND
> ID.EQ.JJ) THEN
YVECT(I)-(VS(N,IU)+WS(M,IU)-WS(N,LL)-VS(M LL))/2
XMTRX(I,2*LL-2)--0.5
XMTRX(I.I)-+0.5
GO TO 280
ELSE IF(IDN.NE.8.AND.IU.EQ.KK.OR.IDN.EQ.8.AND J GE 4 AND
> IU.EQ.KK) THEN
YVECT(I)-(Q(N,KK)4.5
GO TO 280
ELSE IF(IDN.EQ.8.AND.ID.EQ.NODE(1)) THEN
YVECT(I)- -(Q(N,NODE(1))-K3(M,NODE(1))-Q(N,NODE(1)-H)
> -Q(M,NODE(l)+l)-Q(N,NBR(2)-H)-Q(M,NBR(2)+l))/2.
XMTRX(I,I-1)- +.5
XMTRX(I,I+1)- -.5
XMTRX(I,73)- -.5
GO TO 280
ELSE IF(IDN.EQ.8.AND.ID.EQ.NODE(2)) THEN
YVECT(I)-(Q(N,NODE(2))-K5(M,NODE(2))-Q(N,NODE(2)+1)
> -Q(M,NODE(2)+l)-Q(N,NBR(l)-H).Q(M,NBR(l)-H))/2.
XMTRX(I,I-1)- +.5
XMTRX(I.I-H)- -.5
XMTRX(I,31)- -.5
GO TO 280
ELSE IF(IDN.EQ.8.AND.ID.EQ.NODE(4)) THEN
YVECT(I)-(Q(N,NODE(4))+Q(M.NODE(4))-Q(N,NODE(4)-H)
> -Q(M,NODE(4)+l)-Q(N.NBR(3)+l)-Q(M,NBR(3)+l))/2.
XMTRX(I.I-l)- +.5
XMTRX(I,I+1)- -.5
XMTRX(I,91)- -.5
GO TO 280
ELSE IF(IDN.EQ.8.AND.ID.EQ.NBR(1)) THEN
YVECT(I)—(WS(N,NBR(1)+1)+WS(M.NBR(1)+1)-WS(N,NODE(2))
> -WS(M,NODE(2)))/2.
53
-------
XMTRX(I.I)- +.5
XMTRX(I.IO)- -.5
GO TO 280
ELSE IF(IDN.EQ.8.AND.IU.EQ.NODE(3)) THEN
YVECT(I)— (Q(N,NODE(3»+Q(M,NODE(3))-Q(N,NODE(3)-1)
> -Q(M.NODE(3)-l)-Q(N,NBR(l))-Q(M,NBR(l)))/2.
XMTRX(I,I+D- +.5
XMTRX(I.I-l)- -.5
XMTRX(I,29)- -.5
GO TO 280
ELSE IF(IDN.EQ.8.AND.ID.EQ.NBR(2)) THEN
YVECT(I)— (WS(N.NODE(1))+WS(M,NODE(1))-WS(N,NBR(2)+1)
> -WS(M,NBR(2)+l))/2.
XMTRX(I,72)- -.5
XMTRX(I,4)- +.5
GO TO 280
ELSE IF(IDN.EQ.8.AND.ID.EQ.NBR(3)) THEN
YVECT(I)— (WS(N,NODE(4))+WS(M,NODE(4))-WS(N,NBR(3)+1)
> -WS(M,NBR(3)+l))/2.
XMTRX(I,76)- +.5
XMTRXU.90)- -.5
GO TO 280
ELSE IF(IDN.EQ.8.AND.IU.EQ.NODE(5)) THEN
YVECT(I)— -Q(M,NODE(5)-l)-Q(N,NBR(3))-Q(M.NBR(3)))/2.
XMTRX(I.lOl)- +.5
XHTRX(I,99)- -.5
XMTRX(I,89)- -.5
GO TO 280
ELSE IF(IDN.EQ.8.AND.1U.EQ.NODE(6)) THEN
YVECT(I)—(Q(N,NODE(6))-K)(M,NODE(6))-Q(N,NODE(6)-1)
> -Q(M,NODE(6)-l)-Q(N,NBR(2))-Q(M,NBR(2)))/2.
XMTRX(I,15D- *-5
XMTRX(I,71)- -.5
XMTRX(I,U9)- -.5
GO TO 280
END IF
265 CONTINUE
Z41—QA(ID)*T(ID)*(WS(N,IU)+WS(N,ID)-WS(M,IU)-WS(M,ID))/
> (2.*DT*A(ID)**2.)*2.
Z61-(Q(N,ID)-KKN,IU)-Q(M.ID)-Q(M,IU))/(2.*DT*A(ID))
Zll-32 17*AN(ID)**2.*QA(ID)*U(ID)/(2.21*A(ID)**2.*R(ID)**(A./3.))
Z21-32.17*(TH*(WS(N,ID)-WS(N,IU))+TH1*(WS(M,ID)-WS(M,IU)))/X(ID)
Z31—(QA(ID)**2.*(AA(ID)-AA(IU))/(A(ID)**3.*X(ID)»
YVECT(I)— (Z1UZ21+Z31+Z4UZ61)
XMTRX(I.I)—(32.2-QA(ID)**2.*T(ID)/(A(ID)**3.))/X(ID)*TH-QA(ID)
> *T(ID)/(DT*A(ID)**2.)
IF(I.LT.A) GO TO 270
XMTRX(I,I-2)-(32.2-QA(ID)**2.*T(ID)/(A(ID)**3.))A(ID)*TH-QA(ID)
> *T(ID)/(DT*A(ID)**2.)
270 ZZZ1-ABS(QA(ID))
Pl-32.2*AN(ID)**2.*ZZZl*TH/(2.2082*A(ID)**2.*R(ID)**(4./3-))
- AT.//
54
-------
IF (YV(I).LT.-20.) YV(I)—20.
II-I/2-H
WS(N,II)-WS(N,II)+YV(I)
310 CONTINUE
C
C . . . ASSIGN MAXIMUM DELTA VALUE, AND TEST FOR EXCESSIVE ITERATIONS
DO 319 I-2.NNR.2
319 IF(ABS(YV(I)).GT.YVMAX) YVMAX-YV(I)
DO 320 I-2.NNR.2
IF(ABS(YV(I)).GT..002) THEN
IFOTER.GT.20) THEN
WRITE(6,1133) ITER.YVMAX.I/2
1133 FORMATC ',12,' ITERATIONS. YVMAX -',E10.3,'AT I-',12,/)
GO TO 500
END IF
ITER-ITER+1
GO TO 210
END IF
320 CONTINUE
C
C . . . DETERMINE THE DEVIATION OF CALCULATED FROM MEASURED LEVELS
JB-N+1
DO 330 IJ-IIUP.IIDN
IF(WSSAV(N.IJ).LT.50.) WSSAV(N,IJ)-WS(N,LOC(IJ))-.0001
330 DEV(IJ)-WS(N,LOC(IJ))-WSSAV(N,IJ)
MM-MM+1
NM-MM-kkz
IF(NM.LE.O) GO TO 340
IF(NOPT.EQ.3) THEN
DO 6262 I-l.NRM+1
6262 VEL(I)-Q(N,I)/AA(I)
END IF
C , . . METRIC OPTION FOR OUTPUT . .
IF(MUNITS.EQ.O)THEN
DO 6700 MC - IUP.IDN
6700 WSSAV(N.MC) - WSSAV(N,MC)/3.28083
DO 6720 MC - IIUP.IIDN
WS(N,LOC(MC)) - WS(N,LOC(MC))/3.28083
6720 DEV(MC) - DEV(MC)/3.28083
C
Q(N,LOC(IUP)) - Q(N,LOC(IUP))*0.02832
Q(N,LOC(IIDN)) - Q(N,LOC(IIDN))*0.02832
Q(N,NBR(5)+5) - Q(N,NBR(5)-i-5)*0.02832
Q(N.NBR(5)-4) - Q(N,NBR(5)-4)*0.02832
Q(N,NODE(10)+1) - Q(N,NODE(10)+1)*0.02832
Q(N,NBR(4)+3) - Q(N,NBR(4)+3)*0.02832
Q(N,NBR(4)-2) - Q(N,NBR(4)-2)*0.02832
Q(N,NODE(8)+1) - Q(N,NODE(8)+1)*0.02832
Q(N.NBR(l)-3) - Q(N,NBR(1)-3)*0.02832
Q(N,NBR(l)+6) - Q(N,NBR(1)46)*0.02832
Q(N,NBR(3)-3) - Q(N,NBR(3)-3)*0.02832
-------
Q(N,NBR(3)+3) - Q(N,NBR(3)+3)*0.02832
Q(N,LOC(JRB)) - Q(N,LOC(JRB))*0.02832
C
VEL(LOCUUP)) - VEL(LOC(IUP))/3.28083
VEL(LOC(IIDN)) - VEL(LOC(IIDN))/3.28083
VEL(NBR(5)+5) - VEL(NBR(5)+5)/3.28083
VEL(NBR(5)-4) - VEL(NBR(5)-4)/3.28083
VEL(NBR(4)+3) - VEL(NBR(4)+3)/3.28083
VEL(NBR(4)-2) - VEL(NBR(4)-2)/3.28083
VEL(NBR(l)-3) - VEL(NBR(l)-3)/3.28083
VEL(NBR(l)+6) - VEL(NBR(l)+6)/3.28083
VEL(NBR(3)-3) - VEL(NBR(3)-3)/3.28083
VEL(NBR(3)+3) - VEL(NBR(3)+3)/3.28083
VEL(LOC(JRB)) - VEL(LOC(JRB))/3.28083
END IF
C
C
C . . . PRINT OUTPUT.
if(NOPT. EQ.1.and.lup.eq.1) iiup-iiup+1
C*********************QPTION FOR WATER LEVELS********************
C
IF(NOPT.EQ.l) THEN
WRITE(6,3120) nm,WSSAV(N,IDN) ,NOTE(IDN,KDK) ,
> (WS(N,LOC(I)).I-IIDN,IIUP,-1)
WRITE(6,3121) WSSAV(N,IUP),NOTE(IUP,KDK),
> Q(N.LOC(IUP)),
> (WSSAV(N,I).NOTE(I,KDK),DEV(I), I-IIDN, HUP,-1)
C
C ****************QPTION FOR FLOW WITHOUT DELTA******************
C
ELSE IF(NOPT.EQ.2.AND.IDN.NE.8) THEN
WRITE(6,7050)NM,WSSAV(N,IDN),NOTE(IDN,KDK),WS(N,LOC(JRB)),
>WSSAV(N,IUP),NOTE(IUP,KDK),Q(N,LOC(IIDN)),Q(N,LOC(JRB)),
>Q(N,LOC(IUP)),Q(N,NBR(5)-t-5)IQ(N1NBR(5)-4),Q(N,NBR(4)+3),
X3(N,NBR(4)-2),WSSAV(N,JRB),NOTE(JRB>KDK),DEV(JRB)
C
C *******************OPTION FOR FLOW WITH DELTA*****************
C
ELSE IF(NOPT.EQ.2 .AND. IDN.EQ.8) THEN
WRITE(6,8050)NM,WSSAV(NfIDN),NOTE(IDN,KDK)>WS(N,LOC(IIDN-l)),
>WSSAV(N,IUP).NOTE(IUP.KDK),Q(N.LOC(IUP)),Q(N.LOC(HDN)).
>Q(N,NBR(5)-«-5),Q(N,NBR(5)-4),Q(N>NBR(4)+3),Q(N,NBR(4)-2),
>Q(N,NBR(l)-3) ,Q(N,NBR(l)+6) ,Q(N,NBR(3)-3) ,Q(N,NBR(3)+3) .
>WSSAV(N,IIDN-1),NOTE(IIDN-1.KDK),DEV(IIDN-1)
C
C******************OPTION FOR VELOCITIES WITH DELTA*************
C
ELSE IF(NOPT.EQ.3 .AND. IDN.EQ.8) THEN
C WRITE(*,*)'NM- '.NM
WRITE(6,9050) NM,WSSAV(N,IDN) ,NOTE(IDN,KDK),WS(N,LOC(IIDN-1))
>WSSAV(N,IUP) ,NOTE(IUP,KDK) ,VEL(LOC(IUP)) ,VEL(LOC(IIDN)) ,
>VEL(NBR(5)+5) ,VEL(NBR(5) -4) ,VEL(NBR(4)+3) ,VEL(NBR(4) -2) ,
57
-------
>VEL(NBR(1)0),VEL(NBR(1)^),VEL(NBR(3).3),VEL(NBR(3)-K3)
>WSSAVWSSAV(N,IUP),NOTE(IUP.KDK),VEL(LOC(IIDN)),VEL(LOC(JRB))
>VEL(LOC(IUP)),VEL(NBR(5)+5) ,VEL(NBR(5)-4) ,VEL(NBR(4)+3) '
>VEL(NBR(4)-2),WSSAV(N,JRB),NOTE(JRB,KDK),DEV(JRB)
W
END IF
C
if(Nopt.eq.l.and.iup.eq.l) iiup-iiup-l
331 continue
C
C . . . SHOW HOUR ON TERMINAL SCREEN AND COMPUTE MEAN VALUES
WRITE(7,2030)MON,KDK,m-1900,NM
SUM(1)-SUM(1)+WSSAV(N,IUP)
DO 334 1-2,7
334 SUM(I)-SUM(I)+WS(N,LOC(I))
SUM(8)-SUM(8)+WSSAV(N,IDN)
DO 335 1-1,6
335 SUM(I+8)-SUM(I+8)+WSSAV(N,Ul)
DO 336 1-1,6
336 SUM(H-14)-SUM(I+14)+DEV(I+1)
C
C
C . . . THESE ARE THE CALCULATIONS FOR FLOW AVERAGES
SUM(21)-SUM(21)-KKN,LOC(IUP))
SUM(22)-SUM(22)+Q(N,NBR(5)+5)
SUM(23)-SUM(23)-KKN,NBR(5) -4)
SUM(24)-SUM(24)-K)(N,NODE(10)+1)
SUM(25)-SUM(25)-KXN,NBR(4)+3)
SUM(26)-SUM(26)-K}(N,NBR(4).2)
SUM(27)-SUM(27)-H5(N,NODE(8)+1)
SUM(28)-SUM(28)-K5(N,LOC(IIDN))
SUM(29)-SUM(29)-KJ(N,NBR(l)-3)
SUM(30)-SUM(30)-KKN,NBR(l)+6)
SUM(31)-SUM(31)-Kj(N,NBR(3)-3)
SUM(32)-SUM(32)-K3(N,NBR(3)+3)
SUM(43)-SUM(43)-K3(N,LOC(JRB))
SUM(33)-SUM(33)-«-VEL(LOC(IUP))
SUM(34)-SUM(34)+VEL(LOC(IIDN))
SUM(35)-SUM(35)+VEL(NBR(5)-»-5)
SUM(36)-SUM(36)+VEL(NBR(5)-4)
SUM(37)-SUM(37)+VEL(NBR(4)+3)
SUM(38)-SUM(38)-KVEL(NBR(4)-2)
SUM(39)-SUM(39)-»-VEL(NBR(l)-3)
SUM(40)-SUM(40)^-VEL(NBR(l)+6)
SUM(41)-SUM(41)-fVEL(NBR(3)-3)
SUM(42)-SUM(42)+VEL(NBR(3)+3)
58
A 'IS.
-------
c
c
6730
6740
SUM(44)-SUM(44)+VEL(LOC(JRB) )
. CONVERT FROM METRIC BACK TO ENGLISH UNITS
IF(MUNITS.EQ.O)THEN
DO 6730 MC - IU?,IDN
VSSAV(N.MC) - WSSAV(N,MC)*3.28083
DO 6740 MC - IIUP.IIDN
WS(N,LOC(MC)) - WS(N.LOG(MC))*3.28083
DEV(MC) - DEV(MC)*3.28083
Q(N,LOC(IUP)) - Q(N,LOC(IUP))/0.02832
Q(N,LOC(IIDN)) - Q(N,LOC(IIDN))/0.02832
Q(N.NBR(5)+5) - Q(N,NBR(5)+5)/0.02832
Q(N,NBR(5)-4) - Q(N,NBR(5)-4)/0.02832
Q(N.NODE(10)+1) - Q(N,NODE(10)-H)/0.02832
Q(N,NBR(4)+3) - Q(N.NBR(4)+3)/0.02832
Q(N,NBR(4)-2) - Q(N,NBR(4)-2)/0.02832
Q(N,NODE(8)+1) - Q(N,NODE(8)+1)/0.02832
C
C
Q(N,NBR(l)-3)
Q(N,NBR(l)+6)
Q(N,NBR(3)-3)
Q(N,NBR(3)+3)
Q(N,LOC(JRB))
Q(N,NBR(l)-3)/0.02832
Q(N,NBR(l)+6)/0.02832
Q(N,NBR(3)-3)/0.02832
Q(N,NBR(3)+3)/0.02832
Q(N.LOC(JRB))/0.02832
VEL(LOC(IUP)) - VEL(LOC(IUP))*.28083
VEL(LOC(IIDN)) - VEL(LOC(IIDN))*3.28083
VEL(NBR(5)+5) - VEL(NBR(5)+5)*3.28083
VEL(NBR(5)-4)
VEL(NBR(4)^-3)
VEL(NBR(4)-2)
VEL(NBR(l)-3)
VEL(NBR(l)+6)
VEL(NBR(3)-3)
VEL(NBR(3)+3)
VEL(LOC(JRB))
END IF
VEL(NBR(5)-4)*3.28083
VEL(NBR(4)+3)*3.28083
VEL(NBR(4)-2)*3.28083
VEL(NBR(l)-3)*3.28083
VEL(NBR(l)+6)*3.28083
VEL(NBR(3)-3)*3.28083
VEL(NBR(3)+3)*3.28083
VEL(LOC(JRB))*3.28083
340 DO 350 I-l.NMR
350 Q(JB,I)-2.*Q(N,I)-Q(M,I)
DO 360 1-2,NRM
360 WS(JB,I)-2.*WS(N,I)-WS(M,I)
write(7,2030)mon,KDK,iyr-1900,NM
M-M+1
IF(M-kb)2333,2333,370
2333 continue
. . . HOURLY RETURN LOOP
IF(nM-24.Lt.O) GO TO 200
370 DO 380 1-1,44
380 AVE(I)-SUM(I)/NM
C
C
59
-------
c
C . . . PRINT DAILY MEAN VALUES
if (Nopt.eq.1.and.iup.eq.1) iiup-llup+1
C
C*********************OPTION FOR WATER LEVELS*********************
C
IF(NOPT.EQ.l) THEN
WRITE(6,3130) AVE(8) , (AVE(I) ,1-1IDH, IIUP, -1)
WRITE(6,3131) AVE(1),AVE(21),
> (AVE(1+7),AVE<1+13), I-IIDN, IIUP,-1)
if(Nopt.eq.1.and.iup.eq.1) iiup-iiup-1
C
C ******************OPTION FOR FLOW WITHOUT DELTA*****************
C
ELSE IF(NOPT.EQ.2.AND.IDN.NE.8) THEN
WRITE(6,7060)AVE(8),AVE(3),AVE(1),AVE(28).AVE(43),AVE(21),
>AVE(22) ,AVE(23) ,AVE(25) ,AVE(26) ,AVE(10) ,AVE(JRB+13)
C
C********************OPTION FOR FLOW WITH DELTA*******************
C
ELSE IF(NOPT.EQ.2 .AND. IDN.EQ.8) THEN
WRITE(6 ,8060)AVE(8) .AVE(IIDN-l) ,AVE(1) ,AVE(21) ,AVE(28) ,AVE(22) ,
>AVE(23) ,AVE(25) ,AVE(26) ,AVE(29) ,AVE(30) ,AVE(31) ,AVE(32) ,
>AVE(IIDN+6),AVE(IIDN+12)
C
C******************OPTION FOR VELOCITIES WITH DELTA***************
C
ELSE IF(NOPT.EQ.3 .AND. IDN.EQ.8) THEN
WRITE(6,9060) AVE(8).AVE(IIDN-l),AVE(1) ,AVE(33) ,AVE(34),AVE(35) ,
>AVE(36) ,AVE(37) ,AVE(38) ,AVE(39) ,AVE(40) ,AVE(41) ,AVE(42) ,
>AVE(IIDN+6) ,AVE(IIDN+12)
C
C *****************OPTION FOR VELOCITIES WITHOUT DELTA************
C
ELSE IF(NOPT.EQ.3 .AND. IDN.NE.8) THEN
WRITE(6,9065)AVE(8) ,AVE(10) ,AVE(1) ,AVE(34) ,AVE(44) ,AVE(33) ,
>AVE(35) ,AVE(36) ,AVE(37) ,AVE(38) ,AVE(3) ,AVE(JRB+13)
C
END IF
C
381 continue
MONFLAG-.TRUE.
IF(IYR-IYRB) 388,386,500
386 IF(MON-MONB) 388,387.500
387 IF(KDK-IDAYB) 388,500,500
388 CONTINUE
if(kdk-ndm)390,333,333
333 IF(MON-12) 390,389,389
389 IYR-IYR+1
mon-0
390 CONTINUE
DO 392 1-1,42
60
-------
392 SUM(I)-0.
DO 400 I-l.NMR
400 Q(2,I)- 'BASIC DATA'//, 23X, 'STATION', 5X, 'ABASE', 5X, 'DATUM', 5X, 'WIDTH',/)
3020 FORMAT(20X,F10.0,F10.0,F9.2,F8.0,1X,A2)
3030 FORMATC '+',62X,A20)
3040 FORMAT (/,12X,' MANNING n -'.F11.8,' * WS@',A3,' +'
> F10 7,' FOR STATIONS' ,F7.0,' THRU '.F7.0)
3050 FORMAT(///45X, ' ST. CLAIR RIVER HOURLY TRANSIENT MODEL' ,//,52X,A9,
> Ix i2,',',i5,/,10X,A20,' to '.A20.40X, 'WATER LEVELS VERSION',//)
3060 FORMAT (36X.F5.1, IX, 'HOUR TIME INCREMENTS' ,11X, 13, LX, 'REACHES' ,//)
3070 FORMAT(6X,A3,4X,'| ........ COMPUTED LEVELS ......... | ' ,2X,A3,5X,A3,
> 4X,'| ........ MEASURED LEVELS AND COMPUTED DEVIATIONS (C-M) ......
'HR' ,3X,'MEAS.' ,5(4X,A3))
61
-------
3071 FORMAT C + M50,' MEAS. COMP. Q' .IX, 4(IX, ' | ----', A3 ,'--- 1' ) ,1X,
>M. ___ • A3, '---I',/)
3110 FORMAT(ix,A9,lX,I4,lX/ ERROR OF TYPE ',12)
3120 FORMAT(1X,I2,F8.2,A1,5(F7.2))
3121 FORMATC + ',T48,F7.2,A1,F8.0,5(F7.2,A1,1X,F4.2))
3130 FORMAT (/, IX, 'AVE ' ,F6 .2,1X,5(F7.2))
3131 FORMATC + ',T48,F7.2,F8.0,1X,5(F7.2,2X,F4.2))
6190 FORMAT (' FLAG-', 15,' ESP-', 15,' PATH-' ,15)
C7000 FORMAT (/' ENTER OPTION NUMBER'/' 1. OUTPUT SHOWS WATER LEVELS
C > AND DEVIATIONS'/' 2. OUTPUT SHOWS FLOWS AROUND ISLANDS')
7010 FORMAT(Il)
C8001 FORMATC Enter option for delta output:'//' 1. Output shows Water
C >Levels and Deviations'/1 2. Output shows flows around delta and i
C >slands'/' 3. Output shows velocities around delta and islands'/)
8002 FORMAT(Il)
8051 FORMAT (///,45X/ ST. CLAIR RIVER HOURLY TRANSIENT MODEL' ,//,52X,A9 ,
>lx,i2,',',i5,//,10X,A20/ to '.A20.40X, 'RIVER DISCHARGE VERSION1,/
'
8020 FORMAT (//,5X/ | ---- RIVER PROFILE ..... | ' ,2X, ' j --TOTAL FLOW--1',3X,
>'| .......... ISLAND FLOWS .......... |'.3X,'| ........ DELTA FLOWS .....
> ---- |'2X,'|---DEV ---- |',//)
8030 FORMATC HR' ,4X,2(A3 , 5X) ,A3,6X,A3,6X,A3,6X, 'STAG E',3X,'STAG W'
> .4X/FAWN E',3X,'FAWN W' ,3X, 'N.CH. ' , 3X, 'M.CH. ' ,3X, 'S .CH. ' , 3X,
>'CUTOFF',3X,A3,4X,'DEV')
8040 FORMAT(6X, 'MEAS. ' ,4X, 'COMP. ' ,3X, 'MEAS. ' ,4X, 'FLOW' , 5X, 'FLOW ,8X, 'Q
>' ,8X, 'Q' ,8X, 'Q' ,8X, 'Q' ,8X, 'Q' ,7X, 'Q' ,7X. 'Q' ,8X,'Q' ,5X, 'MEAS. • ,2X,
>'C-M',/)
8050 FORMAT(1X,I2,F8.2,A1,1X,F7.2,F8.2,A1,1X,F8.0,U,F8.0,2X,F8.0,2X,
>F7.0,1X,F8.0.2X,F8.0,1X,F7.0,1X,F7.0,1X,F7.0,1X,F7.0,
>1X,F7.2,A1,F5.2)
8060 FORMAT(/,1X,'AVE',F7.2.2X,F7.2,F8.2,2X,F8.011X,F8.0,2X,F8.0,2X>
>F7. 0,1X^8.0,2X^8.0,1X^7.0,1X^7.0,1X^7.0,11^7.0,1X^7. 2, IX,
>F5.2)
9020 FORMAT (//,5X,' | ---- RIVER PROFILE ---- | ' , 3X, ' | --TOT. VEL.--|',4X,
>'| ...... MID ISLAND VELOCITIES ..... | ' ,3X, ' | ---MID DELTA VELOCITIES -
, ,
9030 FORMATC HR' ,4X,2(A3 ,5X) ,A3,6X,A3 ,6X,A3,6X, 'STAG E',3X,'STAG W
> ,4X,'FAWN E',3X,'FAWN W' ,3X, 'N.CH. ' ,3X, 'M.CH. ' ,3X, 'S.CH. ' ,2X,
>'CUTOFF',3X,A3,5X,'DEV')
9040 FORMAT(6X, 'MEAS. ' ,4X, 'COMP. ' ,3X, 'MEAS. ' ,4X, 'VEL. ' ,5X, 'VEL. ' ,8X,
>'V ,8X, 'V ,8X, 'V , 8X, 'V ,8X, 'V ,7X, 'V ,7X, 'V ,7X, 'V ,5X, 'MEAS. ' .
,,
9050 FORMAT(1X>I2,F8.21A1,1X,F7.2,F8.2,A1,1X,F6.2,3X,F6.2,4X,F6.2>4X,
>F5.2.4X.F6.2,3X,F6.2f3X,F5.2,3X.F5.2.3X.F5.2,3X.F5.2.2X,F7.2.Al.
>1X,F5.2)
9051 FORMAT(///, 4 5X,' ST. CLAIR RIVER HOURLY TRANSIENT MODEL' ,//, 52X.A9 ,
>lx,i2,',',i5,//,10X,A20,' TO ' .A20.40X, 'RIVER VELOCITIES VERSION'
9060>FORHAT(/,f AVE' ,F7. 2.2X,F7.2,F8.2,2X,F6.2,3X,F6.2 .4X.F6.2.4X,
>F5.2,4X.F6.2,3X,F6.2,3X,F5.2,3X,F5.2,3X,F5.2,3X,F5.2,2X,F7.2,2X,
>F5.2)
7020 FORMAT (//,17X,' | ..... RIVER PROFILE ..... |',4X,'|--- TOTAL DISCHA
62
-------
>RGE ---I | ISLAND FLOWS | | DEV !',/)
7030 FORMAT(8X,'DAY',7X,A3,8X,A3,8X,A3,8X,A3,6X,A3,6X,A3,3X,
>'STAG E',2X,'STAC V',2X,'FAWN E',2X,'FAWN V,4X,A3,7X,'DEV')
7040 FORMAT(ISX.'MEAS.1,6X,'COMP.',6X,'MEAS. • ,6X,2('FLOW',5X),'FLOW',3X
>,4('FLOW'14X),lX,'KEAS.'f4X,'(C-M)',/)
7050 FORMAT(9X,I2,6X,F6.2,A1,4X,F6.2,5X,F6.2,A1,1X,3(2X,F7.0),
>4(1X,F7.0)>2X,F6.2,A1,2X,F6.2)
7060 FORMAT(/,8X,'AVE',6X,F6.2,5X,F6.2,5X,F6.2.2X,3(2X,F7.0),4(1X,F7.0)
>.2X,F6.2,4X,F5.2)
9025 FORMAT (//,17X, '| RIVER PROFILE |',5X,'|-- TOTAL VELOC IT
>IES --I I-- MID ISLAND VELOCITIES ---| | DEV 1'/)
9041 FORMAT (ISX.'MEAS.1 .6X/COMP.' ,6X,'MEAS.' , 2X, 3(5X,'VEL. ' ) ,2X,
>'VEL.',3(4X,'VEL.'),5X,'MEAS.',4X,'(C-M)f,/)
9055 FORMAT(9X,I2,6X,F6.2,A1,3X,F6.2,6X,F6.2,A1I3(5X,F4.2),3X,F4.2,
>3(4X,F4.2),4X,F6.2)A1,2X,F6.2)
9065 FORMAT(/,8X,'AVE'.6X(F6.2,4X,F6.2,6X,F6.2,IX,3(5X,F4.2)I3XIF4.2,
>3(4X,F4.2),4X.F6.2,4X,F5.2)
C
C
END
63
-------
Program [HYDRO.JDSTCLR]DDELTA.FOR
This is the St Clair River Transient Model - Daily Version.
It is set to run in BATCH MODE...
To run the program...
1. Set desired parameters in file [HYDRO.JDSTCLR]DDELTA.PAR
Line 1 - Starting and ending month and yr.
MO YR MO YR (12 , IX, 12 ,IX, 12 , IX, 12)
Line 2 - Staring and ending points of model. (II,IX,II)
1 - Fort Gratiot
2 - Dunn Paper
3 - Mouth of Black River
4 - Dry Dock
5 - Marysville
6 - St Clair
7 - Algonac
8 - Lake St Clair
Line 3 - Output Option (II)
1 - Water levels and deviations.
2 - Total discharge and discharge around islands and, if
included, discharge in the delta channels.
3 - Velocity near the starting, ending and midpoint of
the simulated river and velocities around islands and,
if included, velocities in the delta channels.
Line 4 - Units Option (II)
0 - Metric units
1 - English units
2. Make sure that file [HYDRO.JDSTCLRJDDELTA.DAT is available.
3. Type: SUBMIT DDELTA/NOTIFY
4. When your request is completed the output will appear in file:
[ HYDRO. JDSTCLR ] ZDDELTA. OUT
Note: this file is 132 characters wide.
64
-------
>IA(359),AVECT(1378),JA(1378),ICC(358),YV(358),RSP<5722),ISP(5722)
>,VEL(180)
EQUIVALENCE (ISP.RSP)
DATA NAMMON/' JANUARY' ,' FEBRUARY' , ' MARCH',' APRIL',
> ' MAY' , ' JUNE' , ' JULY' ,' AUGUST' ',
> 'SEPTEMBER',' OCTOBER',' NOVEMBER',' DECEMBER'/
DATA NAME/' FT. GRATIOT ',' DUNN PAPER
> 'MOUTH OF BLACK RIVER' . ' DRY DOCK ' ,
> ' MARYSVILLE ',' ST CLAIR '\
> ' ALGONAC ','LAKE ST. CLAIR (SCS)'/
DATA NMM/' FG',' DP','MBR',' DD',' MV',' SC', ' AL'.'SCS'/
DATA LOC/180,178,160,157,155,129,77,1/
DATA IGAGE/14098,14096,14090,14087,14084,14080,14070,14052/
DATA OLD/580.92,580.43.580.09,579.53,579.01,578.16,576.60,576.23/
C Al IS SLOPE AND Bl IS INTERCEPT OF MANNINGS N'S FOR ALL REACHES
DATA Al/0.0033947,0.0002708,0.,0.,0.,0.,-0.0011146,-0.0017647/
DATA B1/-1.92253,-0.12683,.0221,.0250,.0240,.0230,
>0.66250,1.04729/
C
C . . . ISLAND/DELTA SPECIFIC VARIABLE ASSIGNMENTS
C
DATA NODE/3,6,25,39,51,76.97.108,135,154/
DATA NBR/15,36,45,102,144/
C
EPERC1-.253
WPERC1-1.-EPERC1
EPERC2-.376
WPERC2-1.-EPERC2
NPERC-.35
MPERC-.21
CNPERC-.56
FNPERC-.56
SPERC-.21
COPERC-.23
CSPERC-.44
FSPERC-.44
NSP-5722
ICOUNT-0
LRATIO-2
IFLAG-0
C
C . . . PHYSICAL DATA ACCESSED
C
DO 10 I-l.LOC(l)
10 READ(5.1020) STA(I),ABAS(I),DATU(I),AT(I)
C
C . . . PROMPT FOR AND READ BEGINNING AND ENDING DATES
C
C WRITE(7,2000)
READ (10,1000) MONA.IYRA.MONB.IYRB
IF(MONA.LE.O) THEN
MONA-12
66
-------
IYRA-IYRA-1
END IF
IYRA-IYRA+1900
IYRB-IYRB+1900
C
C . . . PROMPT FOR LOCATION OF UPPER AND LOWER LIMITS TO BE RUN
C
C 9 WRITE(7,2020) (I,NAME(I), 1-1,8)
9 READ(IO.IOIO) IUP.IDN
IIDN-IDN-1
IIUP-IUP+1
IF(IUP.GE.IIDN) GO TO 9
C
C . . . PROMPT FOR OPTION NUMBER
C
C 616 WRITE(7,8001)
616 READ(10,8002)NOPT
IF(NOPT.GT.3.0R.NOPT.LT.l) GO TO 616
C
READ(10,8005) MUNITS
C
C . . . DEFINE AND ADJUST PARAMETERS, BASED UPON LIMITS FROM ABOVE
NRM-LOC(IUP)-LOC(IDN)
NMR-NRM+1
IF(LOC(IDN).EQ.l) GO TO 27
DO 24 I-l.NMR
STA(I)-STA(I-1+LOC(IDN))
ABAS (I) -ABAS (I - 1+LOC (IDN ) )
DATU (I) -DATU (I - 1+LOC (IDN) )
24 AT(I)-AT(I-1+LOC(IDN))
NODE(1)-NODE(1) -LOC(IDN)+1
NODE(2)-NODE(2) -LOC(IDN)+1
NODE(3)-NODE(3) -LOC(IDN)+1
NODE(4)-NODE(4) -LOC(IDN)+1
NODE(5)-NODE(5) -LOC(IDN)+1
NODE(6)-NODE(6) -LOC(IDN)+1
NODE(7)-NODE(7) -LOC(IDN)+1
NODE(8)-NODE(8) -LOC(IDN)+1
NODE(9)-NODE(9) -LOC(IDN)+1
NODE (10)-NODE (10) -LOC(IDN)+1
NBR(1)-NBR(1)-LOC(IDN)+1
NBR(2)-NBR(2)-LOC(IDN)+1
NBR ( 3 ) -NBR( 3 ) - LOG (IDN)+1
NBR(4)-NBR(4)-LOC(IDN)+1
NBR(5)-NBR(5)-LOC(IDN)+1
DO 25 I-IUP.IDN
25 LOC(I)-LOC(I)-LOC(IDN)+1
C
C . . . CALCULATE DISTANCES BETWEEN SECTIONS
27 DO 30 I-l.NRM
X(I)-STA(I+1)-STA(I)
30 IF(I.EQ.NBR(1).OR.I.EQ.NBR(2) .OR. I .EQ.NBR(3) .OR. I .EQ.
67
-------
>NBR(4).OR.I.EQ.NBR(5» X(I)-0.0
C
C . . . The following lines which, have been COMMENTED with "cO", will
C . . . WRITE the BASIC PHYSICAL DATA. Should the user desire to see
C . . . this data, the "cO's" would have to be eliminated and the
C . . . program re-compiled/1inked etc. TO RE-COMPILE submit DCDELTA
C
cO WRITE (6,3000)
cO WRITE (6,3010)
cO DO 40 I-l.NMR
cO MARK(I)-' '
cO IF(I.EQ.NODE(3).OR.I.EQ.NODE(S).OR.I.EQ.NODE(6)
cO >.OR.I.EQ.NODE(7).OR.I.EQ.NODE(8).OR.I.EQ.NODE(9).OR.
cO >I.EQ.NODE(10)) MARK(I)-'<'
cO IF(I.GT.NODE(2).AND.I.LE.NBR(1))MARK(I)-'N '
cO IF(I.GT.NBR(1).AND.I.LT.NODE(3))MARK(I)-'M '
cO IF(I.GT.NODE(3).AND. I .LE.NBR(2) )MARK(I)-'UN'
cO IF(I.GT.NODE(4).AND.I.LE.NBR(3))ttARK(I)-'S '
cO IF(I.GT.NBR(3).AND.I.LT.NODE(5))MARK(I)-'CO'
cO IF(I.GT.NODE(5).AND.I.LT.NODE(6))MARK(I)-'US'
cO IF(I.GT.NODE(7).AND.I.LE.NBR(4).OR.I.GT.NODE(9).AND.I
cO >.LE.NBR(5))MARK(I)-'W •
cO IF(I.GT.NBR(4).AND.I.LT.NODE(8).OR.I.GT.NBR(5).AND.I
cO >.LT.NODE(10))MARK(I)-'E '
cO IF(I.LE.NODE(2).OR.I.GT.NBR(2).AND.I.LE.NODE(4)) GO TO 40
cO WRITE(6,3020) STA(I),ABAS(I),DATU(I),AT(I),MARK(I)
cO DO 40 IJ-IUP.IDN
cO 40 IF(I.EQ.LOC(IJ)) WRITE(6,3030) NAME(IJ)
C
C THESE NEXT 2 MANNINGS N PRINTOUTS ARE HARDWIRED. IF THE
C NUMBER OF STATIONS CHANGES, CHANGE THESE!
C
cO IF(IDN.EQ.S) WRITE(6,3040) Al(8),NMM(8),B1(8) ,STA(7),STA(36)
cO IF(IDN.EQ.S) WRITE(6,3040) Al(7),NMM(8),B1(7) ,STA(40),STA(77)
cO DO 45 I-IIDN.IUP.-l
cO 45 IF(I.LT.7)WRITE(6,3040) Al(I),NMM(I),B1(I),STA(LOC(I+1)),
cO >STA(LOC(I))
C
C . . . INITIALIZE AND ASSIGN ADDITIONAL VARIABLES
DO 50 1-1,4
50 ADJ(I)-0.
NVAR-NRM*2
ANC-24.
DT-ANC*3600.
DO 51 1-1,42
51 SUM(I)-0.
TH-.75
TH1-.25
MM-0
M=13
istart-13
iend-43
68
-------
kkz-11
MON-MONA
IYR-IYRA
MONFLAG-.FALSE.
JRB-(IUP+IDN)/2
C
C Routine to read all of the water level data from the disk and
C store it in another tempory file. This way the disk is not tied
C up for long periods of time when running the program.
C
C . . . READ WATER LEVELS FROM DISC.
C
52 CALL NODAYS( IYR.MON,1.NDM.NDY, JD)
DO 55 JJ-IUP.IDN
IU-1
IC-IGAGE(JJ)/10000
IGAG-IGAGECJJ)-IC*10000
CALL GAGEIO( IW.IC.IGAG ,MON,IYR,IB.IT.IDA.IDB.IDC,IER)
IF(IER.NE.O) THEN
WR1TE(6.3110) NAMMON(MON).IYR,IER
CALL EXIT
END IF
DO 54 J-l.NDM
DO 53 1-1,24
WRITE(9,56) IHOUR(I.J)
53 CONTINUE
WRITE(9,56) MEAN(J)
54 CONTINUE
55 CONTINUE
56 FORMAT(I6)
C
IF(IYR-IYRB) 58,57,65
57 IF(MON-MONB) 58,65,65
58 CONTINUE
IF(MON-12) 60,59,59
59 IYR-IYR+1
60 CONTINUE
C . . . UPDATE MONTH AND YEAR AND RECHECK IF MORE DATA SHOULD BE USED
MON-MON+1
IF(MON-13) 62,61,61
61 MON-1
62 IF(IYR-IYRB) 64,63,65
63 IF(MON-MONB) 64,64,65
64 CONTINUE
C
GO TO 52
6 5 CALL DISMOUNTPACK( ' WATER_LEVELS' )
MON-MONA
IYR-IYRA
REWIND 9
C
C COME HERE EACH MONTH AND PRINT TITLES AND HEADINGS FOR
69
-------
C EACH OPTION
70 continue
if(Nopt.eq.1.and.iup.eq.1) iiup-iiup+1
C
C ********************QPTION FOR WATER LEVELS********************
C
IF(NOPT.EQ.l) THEN
WRITE(6,3000)
WRITE(6,3050) NAMMON(MON) , IYR,NAME(IUP) .NAME(IDN)
WRITE(6,3060) ANC.NRM
WRITE(6,3070) NMM(IDN),NMM(IUP),
> NMM(IUP),(NMM(I),I-IIDN,IIUP,-1)
WRITE(6,3071) (NMM(I),I-IIDN.IIUP,-1)
IF(NOPT.EQ. LAND. IUP. EQ. 1)IIUP-IIUP-1
C
C ******************QPTION FOR FLOWS WITHOUT DELTA*****************
C
ELSE IF(NOPT.EQ.2 .AND. IDN.NE.8) THEN
WRITE(6,3000)
WRITE(6, 8051 )NAMMON(MON) ,lYR.NAME(IUP) .NAME(IDN)
WRITE(6,3060)ANC,NRM
WRITE(6,7020)
WRITE(6,7030)NMM(IDN) .NMM(JRB) .NMM(IUP) .NMM(IIDN) ,NMM(JRB) ,
> NMM(IUP),NMM(JRB)
WRITE(6,7040)
C
C ******************QPTION FOR FLOWS WITH DELTA********************
C
ELSE IF(NOPT.EQ.2 .AND. IDN.EQ.8) THEN
WRITE(6,3000)
WRITE(6,8051 )NAMMON(MON) ,IYR,NAME(IUP) ,NAME(IDN)
WRITE(6,3060) ANC.NRM
WRITE(6.8020)
WRITE(6,8030)NMM(IDN) .NMM(IIDN-l) ,NMM(IUP) .NMM(IUP) ,NMM(IIDN) ,
> NMM(IIDN-l)
WRITE(6,8040)
C
C ****************OPTION FOR VELOCITIES WITHOUT DELTA***************
C
ELSE IF(NOPT.EQ.3 .AND. IDN.NE.8) THEN
WRITE(6,3000)
WRITE(6,9051)NAMMON(MON) , lYR.NAME(IUP) .NAME(IDN)
WRITE(6,3060)ANC,NRM
WRITE(6,9025)
WRITE(6,7030)NMM(IDN) ,NMM(JRB) .NMM(IUP) ,NMM(IIDN) ,NMM(JRB) ,
> NMM(IUP),NMM(JRB)
WRITE(6,9041)
C
C ****************OPTION FOR VELOCITIES WITH DELTA******************
C
ELSE IF(NOPT.EQ.3 .AND. IDN.EQ.8) THEN
WRITE(6,3000)
70
-------
WRITE(6,9051)NAMMON(MON) ,IYR,HAME(IUP) .NAME(IDN)
WRITE(6,3060)ANC,NRM
WRITE(6,9020)
WRITE(6,9030)NMM(IDN) .NMM(IIDN-l),NMM(IUP) ,NMM(IUP) .NMM(IIDN) ,
> NMM(IIDN-l)
WRITE(6.9040)
END IF
C
71 continue
C ADJUST GAGE-SPECIFIC VARIABLES BASED UPON YEARS STUDIED
IF(IYR.LT.1970) IGAGE(1)-14099
IF(IYR.LT.1971) IGAGE(6)-14080
IF(IYR.GT.1970) ADJ(l)—.18
IF(IYR.GT.1981) ADJ(l)—.06
IF(IYR.LT.1971) ADJ(6)—.09
C
C READ WATER LEVELS FROM UNIT 9 ...
C
CALL NODAYS( IYR.MON,1.NDM.NDY,JD)
C
DO 110 JJ-IUP.IDN
C
DO 77 J-l.NDM
DO 75 1-1,24
75 READ(9,1050,END-78) IHOUR(I.J)
77 READ(9,1055,END-78) MEAN(J)
C
78 CONTINUE
KK - 1
DO 100 J-ISTART.IEND
DO 80 1-1,24
C
C FLAG MISSING DATA AND ASSIGN GAGE LEVELS TO WSSAV
IF(IHOUR(I,KK).GT.O) THEN
NOTE(JJ.KK)-' '
ELSE
NOTE(JJ.KK)-'*'
GO TO 90
END IF
80 CONTINUE
90 WSSAV(J.JJ)-0.0
IF(MEAN(KK).LE.O) THEN
WSSAV(J,JJ)-OLD(JJ)
NOTE(JJ,KK)-'E'
ELSE
WSSAV(J , JJ)-(MEAN(KK) +IB)/100.0+ADJ (JJ)
END IF
OLD(JJ) - WSSAV(J.JJ)
IF(JJ.EQ.IUP.OR.JJ.EQ.IDN) WS(J ,LOC(JJ))-WSSAV(J,JJ)
100 KK - KK+1
110 CONTINUE
IF(MONFLAG) GO TO 200
71
-------
c
C . . . SET WS FOR 12 PREVIOUS TIME STEPS, TO ACHIEVE 'STEADY STATE'.
DO 120 1-1,12
DO 120 J-1,8
WSSAV(I.J)-WSSAV(13,J)
120 IF(J.EQ.IUP -OR. J.EQ.IDN) WS(I,LOC(J) )-WSSAV(13, J)
C
C . . . ZERO MATRIX, SET CHANNEL PARAMETERS, & SET INITIAL CONDITIONS.
DO 130 I - 1,NVAR
YVECT(I)-0.
DO 130 J - l.NVAR
130 XMTRX(J,I)-0.
XSUM-STA(LOC( IUP) ) - STA(LOC{ IDN) )
SLOPE-
-------
200 CONTINUE
N-M+1
ITER-1
210 YVMAX-0.
DO 230 I-l.NRM
QA(I)-TH/2.^Q(N,I)4 IDN.EQ.8.AND.I.GT.106.AND.ID.EQ.LL.OR.IDN.EQ.8.AND.I.GT.106
> .AND.IU.EQ.KK.OR.IDN.EQ.8.AND.ID.EQ.6.0R.IDN.EQ.8.AND.ID.EQ.
> 24.0R.IDN.EQ.8.AND.ID.EQ.39.0R.IDN.EQ.8.AND.ID.EQ.50.0R.
> IDN.EQ.8.AND.ID.EQ.3) THEN
YVECT(I)— (WS(N,IU)+WS(M,IU)-WS(N,ID)-WS(M,ID))/2.
XMTRX(I,I-1)— 0.5
XMTRX(I,I+1)-+0.5
GO TO 260
ELSE IF(ID.EQ.JJ.AND.IDN.NE.8.0R.ID.EQ.JJ.AND.IDN.EQ.8
> .AND.J.GE.4) THEN
YVECT(I)— (WS(N,KK)+WS(M,KK) -WS(N.JJ) -WS(M,JJ))/2 .
XMTRX(I,I-1)— 0.5
XMTRX( I , 2*KK- 2 ) -+0 . 5
GO TO 260
ELSE IF(IDN.EQ.8.AND.ID.EQ.NBR(1)) THEN
YVECT(I)--(VS(N,ID)+WS(M,ID)-WS(N,NODE(3))-WS(M,NODE(3)))/2.
XMTRX(I,28) - +.5
XMTRX(I,48) - -.5
GO TO 260
ELSE IF(IDN.EQ.8.AND.ID.EQ.NBR(2)) THEN
YVECT(I)--(WS(N,ID)+WS(M,ID)-WS(N,NODE(6))-WS(M,NODE(6)))/2.
XMTRX(I,70)- +.5
73
-------
XMTRX(I,150>- -.5
GO TO 260
ELSE IF(IDN.EQ.8.AND.ID.EQ.NBR(3)) THEN
YVECT(I)— (WS(N, ID)+WS(M. ID) -WS(N.NODE(5)) -WS(M,NODE(5)) )/2 .
XMTRX(I,88)- +.5
XMTRXd.lOO)- -.5
GO TO 260
END IF
255 CONTINUE
YVECT(I)— ((WS(N,ID)+WS(N,IU)-WS(M,ID)-WS(M,IU))/(2.*DT) +
> (TH*(Q(N,ID) -Q(N,IU))+TH1*(Q(M,ID)-Q(M,IU)))/(T(ID)*X(ID)))
XMTRX(I,I)-TH/(T(ID)*X(ID))
XMTRX(I,I+2)--XMTRX(I,I)
IF(I.EQ.l) THEN
XMTRX(1,2)-1./(2.*DT)
ELSE
XMTRX(I,I-1)-1./(2.*DT)
XMTRX(I, I+D-1./(2 . *DT)
IF(I.EQ.NRD) XMTRX(I,I+l)-XMTRX(I,I+2)
IF(I.EQ.NRD) XMTRX(I,I+2)-0.
END IF
260 CONTINUE
C
C . . . MOMENTUM EQUATIONS
DO 280 I-2.NVAR.2
ID-I/2
IU-ID+1
DO 265 J-1,5
JJ-NBR(J)
KK-NODE(2*J)
LL-NODE(2*J-1)
IF(IDN.NE.8.AND.ID.EQ.LL.OR.IDN.EQ.8.AND.J.GE.4.AND.
> ID.EQ.LL) THEN
YVECT(I)—(Q(N,JJ-H)4<5(M,JJ+1)4<3(N>IU)-K}(M,IU)-Q(N,ID)-Q(M,ID))
> /2.
XMTRX(I,I-1) —0.5
XMTRX(I,I+D-+0.5
XMTRX( 1, 2*J J+l)-+0. 5
GO TO 280
ELSE IF(IDN.NE.8.AND.ID.EQ.JJ.OR.IDN.EQ.8.AND.J.GE.4.AND.
> ID.EQ.JJ) THEN
YVECT(I)—(WS(N,IU)+WS(M,IU)-WS(N>LL)-WS(M,LL))/2.
XMTRX(I,2*LL-2)--0.5
XMTRX(I,I)-+0.5
GO TO 280
ELSE IF(IDN.NE.8.AND.IU.EQ.KK.OR.IDN.EQ.8.AND.J.GE.4.AND.
> IU.EQ.KK) THEN
YVECT(I)— (Q(N,KK)-KHM,KK) -Q(N,ID)-Q(M,ID) -Q(N, JJ) -Q(M, JJ)
XMTRX(I,2*JJ-1)--0.5
XMTRX(I,I-D —0.5
XMTRX(I,H-1)-+0.5
GO TO 280
-------
ELSE IF(IDN.EQ.8.AND.ID.EQ.NODE(1)) THEN
YVECT(I)- -(Q(N,NODE(1))-KKM,NODE(1))-Q(N,NODE(1)+1)
-Q(M,NODE(l)+l)-Q(N,NBR(2)+l)-Q(M,NBR(2)+l))/2.
XMTRX(I,I-D- +.5
XMTRX(I,I+1)- -.5
XMTRX(I,73)- -.5
GO TO 280
ELSE IF(IDN.EQ.8.AND.ID.EQ.NODE(2)) THEN
YVECT(I)--(Q(N,NODE(2))-KKM,NODE(2))-Q(N,NODE(2)+1)
-Q(M,NODE(2)-H)-Q(N,NBR(l)+l)-Q(M,NBR(l)+l))/2.
XMTRX(I.I-l)- +.5
XMTRX(I,I+1)- -.5
XMTRX(I,31)- -.5
GO TO 280
ELSE IF(IDN.EQ.8.AND.ID.EQ.NODE(4)) THEN
YVECT(I)—(Q(N,NODE(4))4Q(M,NODE(4))-Q(N.NODE(4)+1)
-Q(M,NODE(4)+l)-Q(N,NBR(3)+l)-Q(M,NBR(3)+l))/2.
XMTRX(I.I-l)- +.5
XMTRX(I,I+1)- -.5
XMTRX(I,91)- -.5
GO TO 280
ELSE IF(IDN.EQ.8.AND.ID.EQ.NBR(1)) THEN
YVECT(I)—(WS(N,NBR(1)+1)+WS(M,NBR(1)+1)-US(N,NODE(2))
-WS(M,NODE(2)))/2.
XMTRX(I.I)- +.5
XMTRX(I,10)- -.5
GO TO 280
ELSE IF(IDN.EQ.8.AND.IU.EQ.NODE(3)) THEN
YVECT(I)—(Q(N,NODE(3))-KJ(M,NODE(3))-Q(N,NODE(3)-1)
-Q(M,NODE(3)-l)-Q(N,NBR(l))-Q(M,NBR(l)))/2.
XMTRX(I.H-l)- +.5
XMTRX(I.I-l)- -.5
XMTRX(I,29)- -.5
GO TO 280
ELSE IF(IDN.EQ.8.AND.ID.EQ.NBR(2» THEN
YVECT(I)--(WS(N,NODE(1))+WS(M,NODE(1))-WS(N,NBR(2)+1)
-WS(M,NBR(2)+l))/2.
XMTRX(I,72)- -.5
XMTRX(I,4)- +.5
GO TO 280
ELSE IF(IDN.EQ.8.AND.ID.EQ.NBR(3)) THEN
YVECT(I)—(WS(N,NODE(4))+WS(M,NODE(4))-WS(N,NBR(3)+1)
-WS(M,NBR(3)+l))/2.
XMTRX(I.76)- +.5
XMTRX(I,90)- -.5
GO TO 280
ELSE IF(IDN.EQ.8.AND.IU.EQ.NODE(5)) THEN
YVECT(I)—(Q(N,NODE(5))-H5(M,NODE(5))-Q(N,NODE(5)-1)
-Q(M,NODE(5)-l)-Q(N,NBR(3))-Q(M,NBR(3)))/2.
XMTRX(I.lOl)- +.5
XMTRX(I,99)- -.5
XMTRX(I,89)- -.5
75
-------
GO TO 280
ELSE IF(IDN.EQ.8.AND.IU.EQ.NODE(6)) THEN
YVECT(I)— (Q(N,NODE(6))4Q(M,NODE(6))-Q(N,NODE(6)-1)
> -Q(M,NODE(6)-l)-Q(N,NBR(2))-Q(M,NBR(2)))/2.
XMTRX(I,151)- +.5
XMTRX(I,71)- -.5
XMTRX(I,149)- -.5
GO TO 280
END IF
265 CONTINUE
241— QA(ID)*T(ID)*(WS(N,IU)+WS(N,ID)-WS(M,IU)-WS(M,ID))/
> (2.*DT*A(ID)**2.)*2.
Z61-(Q(N1ID)-K}(N,IU)-Q(MIID)-Q(M,IU))/(2.*DT*A(ID))
Zn-32.17*AN(ID)**2.*QA(ID)*U(ID)/(2.21*A(ID)**2.*R(ID)**(4 /3 )
Z21-32 . 17*(TH*(WS(N, ID) -WS(N, IU) )+THl*(WS(M, ID) -WS(M, IU) ) ) A(ID)
Z31— (QA(ID)**2.*(AA(ID)-AA(IU))/(A(ID)**3.*X(ID)))
YVECT(I)--(Z11+Z21+Z31+Z41+Z61)
XMTRX(I,I)— (32.2-QA(ID)**2.*T(ID)/(A(ID)**3.))/X(ID)*TH-QA(ID)
> *T(ID)/(DT*A(ID)**2.)
IF(I.LT.4) GO TO 270
XMTRX(I,I-2)-(32.2-QA(ID)**2.*T(ID)/(A(ID)**3.))A(ID)*TH-QA(ID)
> *T(ID)/(DT*A(ID)**2.)
270 ZZZ1-ABS(QA(ID))
Pl-32.2*AN(ID)**2.*ZZZl*TH/(2.2082*A(ID)**2.*R(ID)**(4./3.))
> -QA(ID)/A(ID)**3.*(AA(ID)-AA(IU))*TH/X(ID)-TH*T(ID)*(WS(N ID)
> +WS(N,IU) -WS(M, ID) -WS(M, IU))/(2 .*DT*A(ID)**2. )
XMTRX(I,I-1)-1./(2.*A(ID)*DT)+P1
XMTRX(I , I+D-XMTRXU , I- 1)
IF(I.EQ.NVAR) XMTRX(I,I)-XMTRX(I,I-H)
IF(I.EQ.NVAR) XMTRX(I,I+1)-0.
280 CONTINUE
C
C . . .SET UP AVECT, JA, AND IA, AND ACCESS SPARCE MATRIX SOLVER
DO 290 I-l.NVAR
290 IA(I)-0.
DO 291 1-1,900
AVECT(I)-0.
291 JA(I)-0.
K-0
DO 292 I-l.NVAR
NEWROW-.TRUE.
DO 292 J-l.NVAR
IF(XMTRX(I,J).NE.O.) THEN
K-K+1
AVECT(K)-XMTRX(I , J)
JA(K)-J
IF(NEWROW) THEN
NEWROW-. FALSE.
76
-------
END IF
END IF
292 CONTINUE
IA(NVAR+1)-K+1
NSP-100+8*NVAR+2+2*K
PATH-2
IF(ICOUNT.EQ.O) PATH-1
ICOUNT-ICOUNT+1
CALL CDRVCNVAR.RR.CC.ICC.IA.JA.AVECT.YVECT.YV.NSP ISP RSP ESP
> PATH,FLAG)
IF(FLAG.NE.O) THEN
WRITE(6,6190) FLAG,ESP,PATH
GO TO 500
END IF
C
C . . . DETERMINE NEW FLOWS AND LEVELS
NNR-NVAR+1
DO 300 I-1.NNR.2
II-I/2+1
IF(II-NMR.EQ.O) Q(N,NMR)-Q(N,NMR)+YV(NVAR)
300 CONTINUE
NNR-NVAR-2
DO 310 I-2.NNR.2
IF (YV(I).LT.-20.) YV(I)—20.
II-I/2+1
WS(N,II)-WS(N,II)+YV(I)
310 CONTINUE
C
C . . . ASSIGN MAXIMUM DELTA VALUE, AND TEST FOR EXCESSIVE ITERATIONS
DO 319 I-2.NNR.2
319 IF(ABS(YV(I)).GT.YVMAX) YVMAX-YV(I)
DO 320 I-2.NNR.2
IF(ABS(YV(I)).GT..002) THEN
IF(ITER.GT.20) THEN
WRITE(6,1133) ITER,YVMAX,I/2
1133 FORMATC ',12,' ITERATIONS. YVMAX -' ,E10.3 'AT I-' 12 /)
GO TO 500 '
END IF
ITER-ITER+1
GO TO 210
END IF
320 CONTINUE
C
C . . . DETERMINE THE DEVIATION OF CALCULATED FROM MEASURED LEVELS
JB-N+1
DO 330 IJ-IIUP.IIDN
IF(WSSAV(N,IJ).LT.50.) WSSAV(N, IJ)-WS(N,LOC(IJ)) - .0001
330 DEV(IJ)-WS(N,LOC(IJ))-WSSAV(N,IJ)
MM-MM+1
NM-MM-kkz
IF(NM.LE.O) GO TO 340
77
-------
IF(NOPT.EQ.3) THEN
DO 6262 I-l.NRM+1
6262 VEL(I)-Q(N,I)/AA(I)
END IF
C
C . . . METRIC OPTION FOR OUTPUT . .
IF(MUNITS.EQ.O)THEN
DO 6700 MC - IUP.IDN
6700 WSSAV(N.MC) - WSSAV(N,MC)/3.28083
DO 6720 MC- IIUP.IIDN
WS(N,LOC(MC)) - WS(N.LOC(MC))/3.28083
6720 DEV(MC) - DEV(MC)/3.28083
C
Q(N,LOC(IUP)) - Q(N,LOC(IUP))*0.02832
Q(N,LOC(IIDN)) - Q(N,LOC(IIDN))*0.02832
Q(N,NBR(5)+5) - Q(N,NBR(5)+5)*0.02832
Q(N,NBR(5)-4) - Q(N.NBR(5)-4)*0.02832
Q(N,NODE(10)+1) - Q(N,NODE(10)+1)*0.02832
Q(N,NBR(4)+3) - Q(N,NBR(4)+3)*0.02832
Q(N,NBR(4)-2) - Q(N,NBR(4)-2)*0.02832
Q(N,NODE(8)+1) - Q(N,NODE(8)+1)*0.02832
Q(N,NBR(l)-3) - Q(N,NBR(1)-3)*0.02832
Q(N,NBR(l)+6) - Q(N,NBR(1)+6)*0.02832
Q(N,NBR(3)-3) - Q(N,NBR(3)-3)*0.02832
Q(N,NBR(3)+3) - Q(N,NBR(3)+3)*0.02832
Q(N,LOC(JRB)) - Q(N,LOC(JRB))*0.02832
C
VEL(LOC(IUP)) - VEL(LOC(IUP))/3.28083
VEL(LOC(IIDN)) - VEL(LOC(IIDN))/3.28083
VEL(NBR(5)+5) - VEL(NBR(5)+5)/3.28083
VEL(NBR(5)-4) - VEL(NBR(5)-4)/3.28083
VEL(NBR(4)+3) - VEL(NBR(4)-t-3)/3 .28083
VEL(NBR(4)-2) - VEL(NBR(4)-2)/3.28083
VEL(NBR(l)-3) - VEL(NBR(l)-3)/3.28083
VEL(NBR(l)+6) - VEL(NBR(l)+6)/3.28083
VEL(NBR(3)-3) - VEL(NBR(3)-3)/3.28083
VEL(NBR(3)+3) - VEL(NBR(3)+3)/3.28083
VEL(LOC(JRB)) - VEL(LOC(JRB))/3.28083
END IF
C
C . . . PRINT OUTPUT.
if(NOPT.EQ.l.and.iup.eq.l) iiup-iiup+1
C
C ********************QPTION FOR WATER LEVELS********************
C
IF(NOPT.EQ.l) THEN
WRITE(6,3120) NM.WSSAV(N,IDN),NOTE(IDN,NM),
> (WS(N.LOC(I)),I-IIDN,IIUP,-1)
WRITE(6,3121) WSSAV(N,IUP),NOTE(IUP,NM),
> Q(N,LOC(IUP)),
> (WSSAV(N,I),NOTE(I,NM),DEV(I), I-IIDN, IIUP,-1)
A-tUV
-------
C ****************OPTION FOR FLOW WITHOUT DELTA******************
C
ELSE IF(NOPT.EQ.2.AND.IDN.NE.8) THEN
WRITE(6,7050)NM,WSSAV(N,IDN),NOTE(IDN,NM),WS(N,LOC(JRB))
>WSSAV(N,IUP)1NOTE(IUPINM),Q(N,LOC(IIDN)),Q(N,LOC(JRB))
XKN.LOC(IUP)) ,Q(N,NBR(5)+5) ,Q(N,NBR(5) -4) ,Q(N.NBR(4)+3) ,
>Q(N,NBR<4)-2),WSSAV(N,JRB),NOTE(JRB,NM).DEV(JRB)
C
C *******************QPTION FOR FLOW WITH DELTA*****************
C
ELSE IF(NOPT.EQ.2 .AND. IDN.EQ.8) THEN
WRITE(6,8050)NM,WSSAV(N.IDN),NOTE(IDN,NM),WS(N,LOC(IIDN-1))
>WSSAV(N,IUP),NOTE(IUP,NM),Q(N,LOC(IUP)),Q(N,LOC(IIDN)),
>Q(N,NBR(5)+5).Q(N,NBR(5)-4),Q(N,NBR(4)+3),Q(N,NBR(4)-2),
>Q(N>NBR(l)-3)IQ(N,NBR(l)+6),Q(N1NBR(3)-3)1Q(N,NBR(3)+3),
>WSSAV(N,IIDN-1),NOTE(IIDN-1,NM),DEV(IIDN-1)
C
C *****************QPTION FOR VELOCITIES WITHOUT DELTA**********
C
ELSE IF(NOPT.EQ.3 .AND. IDN.NE.8) THEN
WRITE(6,9055)NM,WSSAV(N,IDN),NOTE(IDN,NM),WS(N,LOC(JRB)),
>WSSAV(N,IUP) ,NOTE(IUP,NM) ,VEL(LOC(IIDN)) ,VEL(LOC(JRB) ) ,
>VEL(LOC(IUP)),VEL(NBR(5)+5),VEL(NBR(5)-4).VEL(NBR(4)+3),
>VEL(NBR(4)-2),WSSAV(N,JRB),NOTE(JRB,NM),DEV(JRB)
C
C *****************QPTION FOR VELOCITIES WITH DELTA*************
C
ELSE IF(NOPT.EQ.3 .AND. IDN.EQ.8) THEN
WRITE(6,9050) NM,WSSAV(N,IDN),NOTE(IDN,NM),WS(N,LOC(IIDN-1))
>WSSAV(N,IUP),NOTE(IUP,NM),VEL(LOC(IUP))1VEL(LOC(IIDN)),
>VEL(NBR(5)+5) ,VEL(NBR(5) -4) ,VEL(NBR(4)+3) ,VEL(NBR(4) -2) ,
>VEL(NBR(l)-3),VEL(NBR(l)+6),VEL(NBR(3)-3),VEL(NBR(3)+3),
>WSSAV(N,IIDN-1),NOTE(IIDN-1,NM),DEV(IIDN-1)
END IF
if (Nopt.eq. Land, iup.eq.l) iiup-iiup-1
333 continue
C
C . . . COMPUTE MEAN VALUES.
SUM(1)-SUM(1)+WSSAV(N,IUP)
DO 334 1-2,7
334 SUM(I)-SUM(I)+WS(N,LOC(I))
SUM(8)-SUM(8)+WSSAV(N,IDN)
DO 335 1-1,6
SUM(I+8)-SUM(I+8)+WSSAV(N,I+l)
335 SUM(H-14)-SUM(1+14)+DEV(H-l)
C
C . . . THESE ARE THE CALCULATIONS FOR FLOW AVERAGES
SUM(21)-SUM(21)+Q(N,LOC(IUP))
SUM(22)-SUM(22)+Q(N,NBR(5)+5)
SUM(23)-SUM(23)+Q(N,NBR(5)-4)
SUM(24)-SUM(24)+Q(N,NODE(10)+1)
SUM(25)-SUM(25)4Q(N,NBR(4)+3)
-------
SUM(26)-SUM(26)-KJ(N,NBR(4)-2)
SUM(27)-SUM(27)-KKN,NODE(8)+1)
SUM(28)-SUM(28)-K}(N,LOC(IIDN))
SUM(29)-SUM(29)4<3(N.NBR(l)-3)
SUM(30)-SUM(30)+Q(N.NBR(l)+6)
SUM(31)-SUM(31)+Q(N,NBR(3)-3)
SUM(32)-SUM(32)-KKN,NBR(3)+3)
SUM(43)-SUM(43)-KKN,LOC(JRB))
SUM(33)-SUM(33)+VEL(LOC(IUP))
SUM(34)-SUM(34)+VEL(LOC(IIDN))
SUM(35)-SUM(35)+VEL(NBR(5)+5)
SUM(36)-SUM(36)+VEL(NBR(5)-4)
SUM(37)-SUM(37)+VEL(NBR(4)+3)
SUM(38)-SUM(38)+VEL(NBR(4)-2)
SUM(39)-SUM(39)+VEL(NBR(l)-3)
SUM(40)-SUM(40)+VEL(NBR(l)+6)
SUM(41)-SUM(41)+VEL(NBR(3)-3)
SUM(42)-SUM(42)+VEL(NBR(3)+3)
SUM(44)-SUM(44)+VEL(LOC(JRB))
C
C . . CONVERT BACK TO ENGLISH UNITS . .
IF(MUNITS.EQ.O)THEN
DO 6730 MC - IUP.IDN
6730 WSSAV(N.MC) - WSSAV(N,MC)*3.28083
DO 6740 MC - IIUP.IIDN
WS(N,LOC(MC)) - WS(N,LOC(MC))*3.28083
6740 DEV(MC) - DEV(MC)*3.28083
C
Q(N,LOC(IUP)) - Q(N,LOC(IUP))/0.02832
Q(N,LOC(IIDN)) - Q(N,LOC(IIDN))/0.02832
Q(N,NBR(5)+5) - Q(N,NBR(5)+5)/0.02832
Q(N,NBR(5)-4) - Q(N,NBR(5)-4)/0.02832
Q(N,NODE(10)+1) - Q(N,NODE(10)+1)/0.02832
Q(N,NBR(4)+3) - Q(N,NBR(4)+3)/0.02832
Q(N,NBR(4)-2) - Q(N,NBR(4)-2)/0.02832
Q(N,NODE(8)+1) - Q(N,NODE(8)+1)/0.02832
Q(N,NBR(l)-3) - Q(N,NBR(l)-3)/0.02832
Q(N,NBR(l)+6) - Q(N,NBR(l)+6)/0.02832
Q(N,NBR(3)-3) - Q(N,NBR(3)-3)/0.02832
Q(N,NBR(3)+3) - Q(N,NBR(3)+3)/0.02832
Q(N,LOC(JRB)) - Q(N,LOC(JRB))/0.02832
C
VEL(LOC(IUP)) - VEL(LOC(IUP))*3.28083
VEL(LOC(IIDN)) - VEL(LOC(IIDN))*3.28083
VEL(NBR(5)+5) - VEL(NBR(5)+5)*3.28083
VEL(NBR(5)-4) - VEL(NBR(5)-4)*3.28083
VEL(NBR(4)+3) - VEL(NBR(4)+3)*3.28083
VEL(NBR(4)-2) - VEL(NBR(4)-2)*3.28083
VEL(NBR(l)-3) - VEL(NBR(l)-3)*3.28083
VEL(NBR(l)+6) - VEL(NBR(l)+6)*3.28083
VEL(NBR(3)-3) - VEL(NBR(3)-3)*3.28083
VEL(NBR(3)+3) - VEL(NBR(3)+3)*3.28083
80
-------
VEL(LOC(JRB» - VEL(LOC(JRB))*3.28083
END IF
C
340 DO 350 I-1,NMR
350 Q(JB,I)-2.*Q(N,I)-Q(M.I)
DO 360 1-2,NRM
360 WS(JB,I)-2.*WS(N,I)-WS(M,I)
write(7,2030)mon,nm,iyr-1900
M-M+1
IF(M-kb)2333,2333,370
2333 continue
C
C . . . DAILY RETURN LOOP.
IF(NM-NDM.LT.O) GO TO 200
370 DO 380 1-1,44
380 AVE(I)-SUM(I)/NM
C
C . . . PRINT MONTHLY MEAN VALUES
C
if(Nopt.eq.1.and.iup.eq.1) iiup-iiup+1
C
C ********************OPTION FOR WATER LEVELS*********************
C
IF(NOPT.EQ.l) THEN
WRITE(6.3130) AVE(8),(AVE(I),I-IIDN,HUP,-1)
WRITE(6,3131) AVE(1),AVE(21),
> (AVE(I+7),AVE(I+13), I-IIDN,HUP,-1)
if (Nopt. eq. Land, iup.eq. 1) iiup-iiup-1
C
C ******************OPTION FOR FLOW WITHOUT DELTA*****************
C
ELSE IF(NOPT.EQ.2.AND.IDN.NE.8) THEN
WRITE(6, 7060)AVE(8) ,AVE(3) ,AVE(1) ,AVE(28) ,AVE(43) ,AVE(21) ,
>AVE(22) ,AVE(23) ,AVE(25) ,AVE(26) ,AVE(10) ,AVE(JRB+13)
C
C *******************OPTION FOR FLOW WITH DELTA*******************
C
ELSE IF(NOPT.EQ.2 .AND. IDN.EQ.8) THEN
WRITE(6, 8060)AVE(8) .AVE(IIDN-l) ,AVE(1) ,AVE(21) ,AVE(28) , AVE(22) ,
>AVE(23) ,AVE(25) ,AVE(26) ,AVE(29) ,AVE(30) ,AVE(31) ,AVE(32) ,
>AVE(IIDN-i-6) ,AVE(IIDN+12)
C
C *****************OPTION FOR VELOCITIES WITHOUT DELTA************
C
ELSE IF(NOPT.EQ.3 .AND. IDN.NE.8) THEN
WRITE(6,9065)AVE(8) ,AVE(10) ,AVE(1) ,AVE(34) ,AVE(44) ,AVE(33) .
>AVE(35) , AVE(36) ,AVE(37) ,AVE(38) ,AVE(3) ,AVE(JRB+13)
C
C *****************OPTION FOR VELOCITIES WITH DELTA***************
C
ELSE IF(NOPT.EQ.3 .AND. IDN.EQ.8) THEN
WRITE(6,9060) AVE(8) ,AVE(IIDN-1) ,AVE(1) ,AVE(33) ,AVE(34) ,AVE(35)
-------
>AVE(36) ,AVE(37) ,AVE(38) ,AVE(39) ,AVE(40) ,AVE(41) ,AVE(42) ,
>AVE(HDN+6) ,AVE(IIDNf 12)
END IF
C
381 continue
MONFLAG-.TRUE.
IF(IYR-IYRB) 388,386,500
386 IF(MON-MONB) 388,500,500
388 CONTINUE
IF(MON-12) 390,389,389
389 IYR-IYR+1
390 CONTINUE
DO 392 1-1 ,42
392 SUM(I)-0.
DO 400 I-l.NMR
400 Q(2,I)-Q(JB,I)
DO 410 1-2, NRM
410 WS(2,I)-WS(JB,I)
C
C . . . UPDATE MONTH AND YEAR AND RECHECK IF MORE DATA SHOULD BE USED
MON-MON+1
IF(MON-13) 414,412,412
412 MON-1
414 IF(IYR-IYRB) 418,416,500
416 IF(MON-MONB) 418,418,500
418 CONTINUE
kkz>0
kb-36
lstart*2
iend-32
MM»0
M-l
C
C . . . MONTHLY RETURN LOOP AND END PROGRAM LOCATION
GO TO 70
500 CALL EXIT
C
C . . . FORMAT STATEMENTS.
1000 FORMAT(I2,3(1X,I2))
1010 FORMAT(I1,1X,I1>
1020 FORMAT(F8.0,F8.0,F8.2,F8.0)
1050 FORMAT(I6)
1055 FORMAT(I6)
C2000 FORMAT(/' ENTER BEGINNING AND ENDING DATES'/' MM/YY-MM/YY' )
C2020 FORMAT(/' ENTER STARTING AND ENDING STATIONS: ' /8( 12 ,') ' ,A20/))
2030 FORMATC ' ,12 ,2( ' /' ,13))
3000 FORMAT(lHl)
3010 FORMAT(///,26X,'ST. CLAIR RIVER TRANSIENT MODEL' ,/ ,36X,
> 'BASIC DATA' //,23X,' STATION' ,5X, 'ABASE' ,5X, 'DATUM', 5X, 'WIDTH1 ,/)
A-ICJ*
82
-------
3020 FORMATC20X,F10.0,F10.0,F9.2,F8.0,IX,A2)
3030 FORMATC'+',62X,A20)
3040 FORMATC/,12X,' MANNING n -',F11.8,' * WS€',A3,' + '
> ,F10.7,' FOR STATIONS',F7.0,' THRU ',F7.0)
3050 FORMATC///45X,'ST.CLAIR RIVER TRANSIENT MODEL',//,52X,A9 ,15,//,
> 10X.A20,' to ',A20,40X,'WATER LEVELS VERSION',//)
3060 FORMAT(36X,F5.1,IX,'HOUR TIME INCREMENTS',11X,13,IX,'REACHES',//)
3070 FORMAT(6X,A3,4X,'1 COMPUTED LEVELS V ,2X,A3,5X,A3,4
>X,'1 MEASURED LEVELS AND COMPUTED DEVIATIONS (C-M)
>1'/6X,'MEAS.',5(4X,A3))
3071 FORMATC'+',T50,'MEAS. COMP.Q1 ,1X,4(1X,'1 ',A3,' 1'),1X,
> '1 ',A3,' 1',/)
3110 FORMAT(1X,A9,1X,I4,1X,' ERROR OF TYPE ',12)
3120 FORMAT(1X,I2,F8.2,A1,5(F7.2))
3121 FORMATC'+* ,T48,F7.2,A1,F8.0,5(F7.2,A1,1X,F4.2))
3130 FORMAT(/,1X,'AVE ' ,F6.2,1X,5(F7.2))
3131 FORMATC-I-' ,T48,F7.2 ,F8.0,1X,5(F7 .2.2X.F4.2))
6190 FORMATC FLAG-',15,' ESP-',15,' PATH-',15)
C7000 FORMATC/' ENTER OPTION NUMBER'/' 1. OUTPUT SHOWS WATER LEVELS
C > AND DEVIATIONS'/' 2. OUTPUT SHOWS FLOWS AROUND ISLANDS')
7010 FORMAT(II)
7020 FORMATC//,17X/1 RIVER PROFILE 1',4X,'1 TOTAL DISCHA
>RGE 1 1 ISLAND FLOWS 1 1 DEV 11,/)
7030 FORMATCSX.'DAY' ,7X,A3 ,8X,A3,8X,A3 ,8X,A3 ,6X,A3,6X,A3,3X,
>'STAG E',2X,'STAG W',2X,'FAWN E',2X,!FAWN W',4X,A3,7X,'DEV )
7040 FORMATC18X,'MEAS.' ,6X,'COMP.' ,6X,'MEAS.',6X,2C'FLOW' ,5X),'FLOW' ,3X
>,4C'FLOW',4X),1X,'MEAS.' ,4X,'CC-M)' ,/)
7050 FORMATC9X,I2,6X,F6.2,A1,4X,F6.2,5X,F6.2,A1,1X,3(2X,F7.0),
>4C1X,F7.0),2X,F6.2,A1,2X,F6.2)
7060 FORMATC/,8X,'AVE' ,6X,F6.2,5X,F6.2,5X,F6.2,2X,3C2X,F7.0) ,4ClX,F7.0)
>,2X,F6.2,4X,F5.2)
C8001 FORMATC' Enter option for delta output:'//' 1. Output shows Water
C >Levels and Deviations'/' 2. Output shows flows around delta and i
C >slands'/' 3. Output shows velocities around delta and islands'/)
8002 FORMATC ID
8005 FORMATCII)
8020 FORMATC//,5X,'1 RIVER PROFILE 1' ,2X,' 1—TOTAL FLOW 1',3X,
>i^ ISLAND FLOWS 1',3X,'1 DELTA FLOWS
> 1' ,3X,'1 DEV 1' ,//)
8030 FORMATC7X,2CA3,5X),A3,6X,A3,6X,A3,6X,'STAG E',3X,'STAG W',4X,'FAWN
> E',3X,'FAWN W',SX.'N.CH.'.SX.'M.CH.'.SX.'S.CH.'^X,'CUTOFF',4X,
>A3,5X,'DEV')
8040 FORMATC6X,'MEAS.',4X,'COMP.',3X,'MEAS.',4X,'FLOW1,5X,'FLOW',8X,'Q
>' ,8X,'Q' ,8X,'Q' ,8X,'Q' ,8X,'Q' ,7X,'Q' ,7X,'Q' ,8X,'Q' ,6X,'MEAS.' ,3X,
>*C-M' /)
8050 FORMATC1X,I2,F8.2,A1,1X,F7.2,F8.2,A1,F8.0,1X,F8.0,2X,F8.0,2X,
>F7.0,1X,F8.0,2X,F8.0,1X,F7.0,1X,F7.0,1X,F7.0,2X,F7.0,
>2X,F7.2,A1,1X,F5.2)
8051 FORMATC///,45X,'ST.CLAIR RIVER TRANSIENT MODEL',//,52X,A9 ,15,//,
MOX.A20,' to ',A20,40X,'RIVER DISCHARGE VERSION',//)
8060 FORMATC/,1X,'AVE',F7.2,2X,F7.2,F8.2,1X,F8.0,1X,F8.0,2X,F8.0,2X,
>F7.0,1X,F8.0,2X,F8.0,1X,F7.0,1X,F7.0,1X,F7.0,2X,F7.0,2X,F7.2,2X,
-"-ft
-------
>F5.2)
9020 FORMAT(//,5X,'1 -- RIVER PROFILE - 1' ,3X,' 1— TOT. VEL.— 1',4X,
>'1 -- MID ISLAND VELOCITIES - 11 ,3X,'1 - MID DELTA VELOCITIES-
> - 1 ' ,2X, ' 1 -- DEV --- 1 ' ,//)
9025 FORMAT(//,17X,'1 --- RIVER PROFILE - 1',5X,'1— TOTAL VELOCIT
>IES —1 1— MID ISLAND VELOCITIES --- 1 1 -- DEV - I'/)
9030 FORMAT(7X,2(A3,5X),A3,6X,A3,6X,A3,6X,(STAG E',3X,'STAG W',4X,'FAWN
> E',3X,'FAWN W'^X.'N.CH.'.SX.'M.CH.'.SX.'S.CH.'^X, 'CUTOFF', 3X,
>A3,5X,'DEV)
9040 FORMATCex.'MEAS.' ,4X,'COMP.' ,3X,'MEAS.' ,4X, 'VEL. ' ,5X, 'VEL. ' ,8X,
'V ,8X,'V ,8X,'V ,8X,'V ,8X,'V' ,7X,fV ,7X,'V ,7X,'V' .SX.'MEAS.1 ,
9041 FORMAT (18X,1 ME AS.1 ,6X,'COMP.' ,6X,'MEAS.' ,2X,3(5X,'VEL.' ) ,2X,
>'VEL.' ,3(4X,'VEL.'),5X,'MEAS.' ,4X,'(C-M)' ,/)
9050 FORMAT(1X,I2,F8.2,A1,1X,F7.2,F8.2,A1,1X,F6.2,3X,F6.2,4X,F6.2,4X,
>F5.2,4X,F6.2,3X,F6.2,3X,F5.2,3X,F5.2,3X,F5.2,3X,F5.2,2X,F7.2,A1,
MX.F5.2)
9051 FORMAT(///,45X,! ST. GLAIR RIVER TRANSIENT MODEL1 ,//,52X,A9 ,15 ,//,
>10X,A20,' to ' ,A20,40X,' RIVER VELOCITY VERSION',//)
9055 FORMAT(9X,I2,6X,F6.2,A1,3X,F6.2,6X,F6.2,A1,3(5X,F4.2),3X,F4.2,
>3(4X,F4.2),4X,F6.2,A1,2X,F6.2)
9060 FORMAT(/,' AVE' ,F7.2,2X,F7.2 ,F8.2,2X,F6.2 ,3X,F6.2,4X,F6.2,4X,
>F5.2,4X,F6.2,3X,F6.2,3X,F5.2,3X,F5.2,3X,F5.2,3X,F5.2,2X,F7.2,2X,
>F5.2)
9065 FORMAT(/,8X,IAVE',6X,F6.2,4X,F6.2,6X,F6.2,1X,3(5X,F4.2),3X,F4.2,
>3(4X,F4.2),4X,F6.2,4X,F5.2)
C
END
C
A-K.U,
-------
TABLES
1. Roughness coefficients for the St. Clair River reaches.
A-2. Hourly model output - water levels option.
A-3. Hourly model output - river discharge option.
A-4. Hourly model output - river velocities option.
A-5. Daily model output - water levels option.
A-6. Daily model output - river discharge option.
A-7. Daily model output - river velocities option.
85
-------
Table A-l. Hourly model output - water levels option.
FT. CRATIOT to LAKE ST. CLAIR (SCS)
1.0 HOUR TIME INCREMENTS
ST.CLAIR RIVER HOURLY TRANSIENT MODEL
MAY 8, 1987
WATER LEVELS VERSION
179 REACHES
cr c
HR
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
JV..J
MEAS.
176
176
175
175
175
176
176
176
175
176
176
175
176
176
178
176
176
176
176
176
176
176
176
176
.46
.47
.47
.47
.47
.44
.47
.46
.47
.47
.48
.48
.47
.48
.47
.48
.46
.46
.46
.46
.46
.46
.46
.46
i
1 •
176
176
176
175
175
175
176
176
175
176
175
176
175
176
175
176
176
176
176
175
175
176
176
176
_ _, rnuDiiTcri i cuci c i
AL
.62
.62
.63
.63
.63
.61
.62
.62
.62
.63
.64
.64
.63
.63
.63
.63
.63
.62
.62
.61
.61
.61
.61
.61
SC MV DO
176.01
176.02
176.02
176.03
176.03
176.02
176.02
176.02
176.02
176.04
176.06
176.04
176.04
176.04
176.03
176.03
176.03
176.03
176.03
176 .02
176.02
176.02
176.01
176.00
176.29
176.30
176.30
176.31
176.31
176.31
176.31
176.31
176.31
176.33
176.33
176.32
176.33
176.32
176.31
176.32
176.31
176.32
178.32
176.31
176.31
176.30
176.29
176.28
176.42
176.43
176.43
176.44
176.45
176.44
176.44
176.44
176.44
176.47
176.46
176.44
176.46
176.46
176.44
176.46
176.43
176.46
176.46
176.44
176.44
176.44
176.42
176.41
MBR
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
.53
.65
.64
.56
.66
.65
.66
.66
.55
.59
.68
.66
.59
.66
.66
.67
.66
.67
.67
.66
.56
.66
.63
.62
fC
r Ij
MEAS.
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
.76
.77
.76
.79
.78
.77
.78
.78
.77
.83
.79
.77
.82
.77
.78
.79
.76
.81
.79
.78
.79
.77
.74
.74
FG
COMP.Q
6168.
6214.
6162.
6276.
6217.
6186.
6270.
6246.
6219.
6450.
6245.
6162.
6369.
6139.
6240.
6266.
6143.
6344 .
6272.
6222.
6289.
6230.
6124.
6162.
1
MCACiiDcri i c\/ci c Akin rnuc
| AL — -|
176.60
176.69
176.60
176.61
176.61
176.61
17S.61
175.61
175.61
176.60
176.62
176.61
176.62
176.62
176.62
175.61
176.62
176.60
176.61
176.62
176.61
175.63
175.62
176.61
0.01
0.03
0.02
0.02
0.02
0.00
0.01
0.01
0.01
0.03
0.02
0.03
0.01
0.02
0.01
0.02
0.01
0.01
0.01
0.00
0.00
-.02
-.01
-.01
jwnc.1/ UE-VtUs) I^TW *.
| SC---I I— -
176.02
176.03
176.03
176.03
176.06
176.04
176.04
176.03
176.04
176.07
1 76 . 06
176.06
176.06
176.07
176.01
176.06
1 76 . 06
176.04
176.06
176.06
170.04
176.07
176.07
176.04
- .01
- .01
- .01
- .01
-.02
-.02
-.02
- .01
-.02
-.03
- .01
-.02
- .01
-.03
0.02
- .02
-.03
- .01
- .02
- .03
-.02
-.06
-.06
- .04
176.
176.
176.
176.
176.
176.
176.
176.
176.
176.
176.
176.
176.
176.
176.
176.
176.
176.
178.
176.
176.
176.
176.
176.
uiTcrt r\cuT*TTnwc /r_u\_
, VffT' W < fc.1/ 1/Ur.A'llAl'i*** \ V
- MV | | DO---)
27
27
26
28
29
28
28
29
28
31
30
29
30
29
29
29
29
29
30
29
29
29
30
28
0.02
0.03
0.04
0.04
0.03
0.03
0.03
0.02
0.02
0.02
0.04
0.02
0.03
0.02
0.02
0.02
0.02
0.03
0.02
0.01
0.02
0.01
- .01
0.00
176
176
176
176
176
176
176
176
176
176
176
176
176
178
176
178
176
176
176
176
176
176
176
176
.41
.42
.39
.41
.42
.41
.41
.42
.41
.46
.43
.42
.43
.42
.41
.43
.38
.42
.43
.42
.42
.42
.43
.40
0.01
0.01
0.04
0.03
0.02
0.03
0.03
0.02
0.02
0.01
0.04
0.02
0.03
0.03
0.03
0.02
0.06
0.03
0.02
0.02
0.03
0.02
-.02
0.01
l
"> i
| MBR---)
176.52
176.53
176.53
176.66
176.64
176.54
176.64
176.54
176.55
176.67
176.56
176.65
176.56
176.54
176.64
176.56
176.64
176.57
176.66
176.66
176.66
176.66
176.63
176.63
0.01
0.01
0.01
0.01
0.02
0.01
0.02
0.02
0.00
0.02
0.02
0.01
0.03
0.02
0.02
0.01
0.00
0.00
0.01
0.01
0.01
0.00
0.00
-.01
AVE 176.47 17S.62 176.03 176.31 176.44 176.56 176.78
6233. 175.61 0.01 176.05 -.02 176.29 0.02 176.42 0.02 176.56 0.01
-------
Table A-2. Hourly model output - river discharge option.
FT. CRATIOT
ST.CLMR RIVER HOURLY TRANSIENT MODEL
MAY 8, 1987
to LAKE ST. CLAIR (SCS)
RIVER DISCHARGE VERSION
1.0 HOUR TIME INCREMENTS
179 REACHES
| RIVER PROFILE | I--TOTAL FLOW--|
co
HR
SCS
MEAS.
1
2
3
4
6
6
7
8
9
10
11
12
13
14
16
16
17
18
19
20
21
22
23
24
AVE
176
176
176
176
176
175
176
176
176
175
176
175
176
176
176
176
176
176
176
176
176
175
176
175
176
.46
.47
.47
.47
.47
. 44
.47
.46
.47
.47
.48
.48
.47
.48
.47
.46
.46
.45
.46
.46
.46
.46
.45
.46
.47
SC
COMP.
176
176
176
176
176
176
178
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
.01
.02
.02
.03
.03
.02
.02
.02
.02
.04
.06
.04
.04
.04
.03
.03
.03
.03
.03
.02
.02
.02
.01
.00
.03
FG
MEAS.
176
176
176
176
176
176
176
176
176
176
178
176
176
176
176
176
170
176
176
176
176
176
176
176
176
.76
.77
.76
.79
.78
.77
.78
.78
.77
.83
.79
.77
.82
.77
.78
. 79
. 76
.81
.79
.78
.79
.77
.74
.74
.78
FG
FLOW
6168.
6214 .
6162.
6276.
6217.
6186.
6270.
6245.
6219.
6460.
6246.
6162.
6369.
6139.
6240.
0206.
6143.
6344.
6272.
6222.
6289.
6230.
6124.
6162.
6233.
AL
FLOW
6156.
6146.
6168.
6184 .
6217.
6299.
6234.
6227.
6234.
6240.
6269.
6266.
6266.
6266.
0247.
6227 .
6740.
6281 .
6284.
6303.
6272.
6269.
6261.
6173.
6237.
,
STAG E
Q
2467.
2467.
246S.
2480.
2487.
2480.
2493.
2496.
2486.
2624.
2E19.
2480.
2602.
2490.
2481 .
2494.
2477.
2499.
2614.
2499.
2601.
2498.
2471.
2467.
STAC W
Q
3700.
3716.
3711.
3730.
3746.
3733.
3764.
3768.
3742.
3806.
3792.
3728.
3769.
3747.
3732.
3760.
3726.
3764 .
3787.
3760.
3766.
3760.
3716.
3698.
FAWN E
Q
777 .
777 .
778.
782.
786.
793.
789.
786.
786.
790.
793.
788.
789.
790.
786.
706.
766 .
792.
796.
796.
791.
790.
787.
777 .
,
FAWN W
Q
6380.
6380.
6386.
6400.
6433.
6471.
6469.
6447.
6447.
6468.
6488.
6468.
6466.
6469.
6462.
E44S.
6446.
6471 .
6496.
6498.
6481 .
6474.
6462.
6399.
1
N.CH.
Q
1973.
1962 .
1977.
1982.
1996.
2043.
1973.
2001 .
1991 .
1992.
1998.
2011 .
2014.
2001.
2004.
1989.
?00«.
7021 .
2007.
2020.
1999.
2011.
1997.
1963.
M.CH.
Q
828.
822.
830.
832.
838.
869.
826.
840.
834.
836.
838.
845.
846.
839.
841.
834 .
842.
848.
841 .
860.
837 .
843.
836.
821.
r LUfij — — —
S.CH.
Q
1008.
997 .
1009.
1010.
1020,
1061.
996.
1031 .
1010.
1014.
1011 .
1024.
1026.
1014.
1019.
1004 .
1022.
1033.
1020.
1041 .
1016.
1032.
1018.
994.
CUTOFF
Q
2349.
2340.
2347 .
2362.
2369.
2413.
2362.
2391 .
2377.
2382.
2379.
2400.
2400.
2396.
2399.
2387.
239B.
2410.
2396.
2416.
2396.
2409.
2399.
2374.
1 i/v.
sc
MEAS.
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
.02
.03
.03
.03
.06
.04
.04
.03
.04
.07
.06
.06
.06
.07
.01
.06
.06
.04
.05
.06
.04
.07
.07
.04
i i
OEV
C-M
-0.01
-0.01
-0.01
-0.01
-0.02
-0.02
-0.02
-0.01
-0.02
-0.03
-0.01
-0.02
-0.01
-0.03
.02
- .02
- .0*
- .01
- .02
- .03
- .02
- .06
-0.06
-0.04
2488.
3746.
787.
5449.
1997.
838.
1018.
2386. 176.05 -0.02
-------
Table A-3. Hourly model putout - river velocities option,
FT. GRAT10T
ST.CLAIR RIVER HOURLY TRANSIENT MODEL
MAY 8, 1987
TO LAKE ST. CLAIR (SCS)
RIVER VELOCITIES VERSION
1.0 HOUR TIME INCREMENTS
179 REACHES
| RIVER PROFILE I
|--TOT. VEL.--I
Hft
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
SCS
MEAS.
175.46
176.47
178.47
175.47
17S.47
176.44
175.47
176.46
176.47
176.47
176.48
176.48
176.47
175.48
176.47
178 .48
176.46
176.46
176.46
176.46
176.46
176.46
176.46
176.46
SC
COMP.
176.01
176.02
176.02
176.03
176.03
176.02
176.02
178.02
176.02
176.04
176.06
176.04
176.04
176.04
176.03
178.83
176.03
176.03
176.03
176.02
176.02
176.02
176.01
176.00
FC
MEAS.
176.76
176.77
176.76
176.79
176.78
176.77
176.78
176.78
176.77
176.83
176.79
176.77
176.82
176.77
176.78
178 .79
176.76
176.81
176.79
176.78
176.79
176.77
176.74
176.74
FC
VEL.
1 .06
1 .06
1 .06
1 .07
1 .06
1 .05
1.06
1 .06
1.06
1 .09
1.06
1 .06
1.08
1 .04
1 .06
1 .06
1 .04
1.08
1 .06
1 .06
1.07
1 .06
1 .04
1 .06
AL
VEL.
0.67
0.66
0.67
0.67
0.67
0.68
0.68
0.67
0.67
0.67
0.68
0.66
0.68
0.68
0.68
0.67
0.67
0.68
0.68
0.68
0.68
0.68
0.68
0.67
1 1
STAC E
V
0.98
0.98
0.98
0.99
0.99
0.99
0.99
0.99
0.99
1.00
1.00
0.98
0.99
0.99
0.99
0.99
0.98
0.99
1.00
0.99
0.99
0.99
0.98
0.98
fflk> launnu
STAG W
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
V
.86
.86
.86
.86
.86
.86
.87
.87
.86
.88
.87
.86
.87
.86
.86
.87
.86
.87
.87
.87
.87
.87
.86
.86
1 VCl.UV.4 1 11
FAWN E
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
V
.49
.49
.49
.49
.49
.60
.49
.49
.49
.49
.49
.49
.49
.49
.49
.49
.49
.49
.50
.60
.49
.49
.49
.49
-3 !
FAWN W
V
0.93
0.93
0.93
0.93
0.93
0.94
0.94
0.94
0.94
0.94
0.94
0.94
0.94
0.94
0.94
0.94
0.94
0.94
0.94
0.96
0.94
0.94
0.94
0.93
1 ""J.
N.CH.
V
0.78
0.78
0.78
0.79
0.79
0.81
0.78
0.80
0.79
0.79
0.79
0.80
0.80
0.79
0.79
0. 79
0.80
0.80
0.80
0.81
0.80
0.80
0. 79
0. 78
\J l/CV. 1 «
M.CH.
V
0.62
0.62
0.62
0.53
0.63
0.64
0.62
0.63
0.63
0.53
0.53
0.63
0.63
0.63
0.63
0.63
0.63
0.64
0.63
0.64
0.63
0.63
0.63
0.52
VC.UU*. A 1 J
S.CH.
V
0.36
0.36
0.35
0.35
0.35
0.37
0.36
0.36
0.36
0.36
0.36
0.36
0.36
0.36
0.36
0.35
0.36
0.36
0.36
0.36
0.35
0.36
0.36
0.36
.C J f
CUTOFF
V
0.76
0.76
0.76
0. 77
0.77
0.79
0.77
0. 76
0.77
0.78
0.77
0.78
0.78
0.78
0.78
0.78
0. 78
0.79
0.78
0.79
0.78
0.79
0.78
0.77
r •"•
SC
MEAS.
178.02
176.03
176.03
176.03
176.06
176.04
176.04
176.03
176.04
176.07
176.06
176.06
176.06
176.07
176.01
176.06
176.06
176.04
176.06
176.06
176.04
176.07
176.07
176.04
DEV
C-M
-0.01
-0.01
-0.01
-0.01
-0.02
-0.02
-0.02
-0.01
-0.02
-0.03
-0.01
-0.02
-0.01
-0.03
0.02
-0.02
-0.03
-0.01
-0.02
-0.03
-0.02
-0.05
-0.06
-0.04
AVE 175.47 176.03 176.78
1 .06
0.67
0.99
0.87
0.49
0.94
0.79
0.63
0.36
0.78 176.06 -0.02
-------
Table A-4. Daily model output - water levels option.
oo
VO
FT. CRATIOT
ST.CLAIR RIVER TRANSIENT MODEL
JUNE 1986
to LAKE ST. CLAIR (SCS)
WATER LEVELS VERSION
24.0 HOUR TIME INCREMENTS
179 REACHES
1
2
3
4
6
6
7
8
9
10
11
12
13
14
16
16
17
18
19
20
21
22
23
24
26
26
27
28
29
30
AVE
bLi
MEAS.
176.66
176.66
175.66
176.64
176.66
176.70
175.69
176.69
176.69
176.69
176.71
175.74
175.76
175.77
176.78
176.74
176.76
176.75
175.76
176.78
175.77
176.74
175.74
176.73
176.74
175.72
176.71
176.73
176.73
176.73
176.72
1
AL
176
176
176
176
176
176
175
176
176
176
176
175
175
176
176
175
176
175
176
176
176
175
176
176
176
176
175
176
176
176
176
.82
.82
.81
.80
.83
.86
.84
.84
.84
.83
.86
.87
.89
.91
.91
.88
.90
.89
.89
.92
.90
.88
.88
.88
.88
.86
.86
.88
.88
.88
.87
LUMri
SC
176.26
176.26
176.23
176.22
176.26
176.26
176.24
176.26
176.26
176.24
176.26
176.26
176.30
176.31
176.31
176.30
176.32
176.29
176.30
176.33
176.29
176.28
176.30
176.32
176.31
176.27
176.28
176.30
176.32
176.31
176.28
n c.u LtvtLa
MV DD
176.66
176.66
176.62
176.52
176.66
176.66
176.63
176.55
176.55
176.52
176.64
176.52
176.59
176.69
176.58
176.59
176.61
176.57
176.58
176.62
176.57
176.66
176.59
176.62
176.60
176.66
176.68
176.60
176.62
176.61
176.67
176.69
176.69
176.66
176.66
176.69
176.68
176.66
176.69
176.68
176.65
176.67
176.66
176.72
176.72
178.71
176.73
176.76
176.70
176.71
176.76
176.69
176.69
176.72
176.76
176.73
176.68
176.71
176.73
176.76
176.76
176.70
|
MBR
176
176
176
176
176
176
176
176
176
178
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
178
176
176
176
.81
.81
. 77
.78
.81
.80
.77
.81
.60
. 77
.79
.76
.84
.83
.82
.86
.87
.82
.83
.87
.81
.81
.84
.89
.85
.80
.83
.86
.69
.87
.82
rvj
MEAS.
177
177
177
177
177
177
177
177
177
177
177
176
177
177
177
177
177
177
177
177
177
177
177
177
177
177
177
177
177
177
177
.06
.06
.01
.02
.06
.03
.01
.05
.04
.00
.02
.98
.07
.06
.05
.09
.11
.06
.06
.11
.03
.04
.08
. 14
.10
.03
.08
.10
.14
.12
.06
rv»
COMP.Q
6690.
6690.
6639.
6604.
6647.
6662.
6486.
6638.
6582.
6480.
6488.
6319.
6690.
6507.
6454.
6664.
6666.
6527.
6662.
6658.
6414.
6500.
6633.
6816.
6671.
6497.
6663.
6704.
6812.
6740.
6693.
1 MCAaune
| AL— | I--
176.78
176.77
176.79
176.79*
176.81
175.83
175.82
175.83
175.82
175.82
176.85
176.91
175.90
176.90
176.91
175.92
175.89
176.89
175.90
176.90
176.90
176.90
175.89
176.88
176.88
176.88
175.88
175.87
175.86
176.87
176.86
0.04
0.06
0.02
0.01
0.02
0.02
0.01
0.01
0.02
0.01
0.01
-.03
- .01
0.01
0.01
-.04
0.02
0.00
-.01
0.02
0.01
-.02
-.01
0.00
0.01
-.02
-.02
0.00
0.02
0.01
0.01
176
176
176
176
176
176
176
176
176
176
176
176
176
178
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
U LtVtLi «IW
-- SC---I )--
.27
.28
.26
.26
.28*
.28
.28
.29
.29
.27
.30
.32
.36
.35
.34
.35
.36
.33
.33
.36
.32
.32
.34
.36
.34
.32
.34
.35
.35
.34
.32
- .02
0.00
- .02
- .03
- .02
- .02
- .03
-.03
- .03
- .03
- .04
-.07
- .06
-.04
-.03
-.05
- .03
- .04
- .04
- .03
- .02
- .04
- .04
-.03
-.04
-.06
-.06
-.04
- .03
-.03
- .03
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
v.uiwru icf L>CVI« i iumo i>_-
-- MV — -| | DD---I
.62
.61
.49
.49
.62
.62
.50
.63
.52
.50
.63
.53
.68
.57
.56
.59
.69
.56
.67
.60
.56
.66
.68
.60
.68
.56
.67
.69
.60
.69
.66
0.03
0.04
0.03
0.02
0.03
0.03
0.02
0.03
0.03
0.02
0.02
-.01
0.01
0.02
0.02
0.01
0.02
0.01
0.01
0.02
0.02
0.01
0.01
0.02
0.01
-.01
0.00
0.01
0.02
0.02
0.02
178
176
178
176
176
176
176
176
176
176
176
176
176
176
176
178
176
176
176
176
176
176
176
176
176
176
176
178
176
176
176
.66
.66
.63
.63
.67
.66
.64
.67
.67
.64
.66
.66
.71
.70
.69
.72
.72
.69
.71
.73
.68
.68
.71
.75
. 72
.68
.70
.72
.74
.73
.69
0.03
0.04
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.00
0.01
0.01
0.01
0.01
0.02
0.01
0.01
0.02
0.01
0.01
0.02
0.01
0.01
0.00
0.01
0.01
0.02
0.02
0.02
-iw; |
| MBR---I
176.79
176.78
176.76
176.76
176.79
176.78
176.76
176.80
176.79
176.76
176.78
176.78
176.84
176.83
176.82
176.84
176.86*
176.81
176.82
176.86
176.80
176.81
176.83
176.87
176.85
176.80
176.84
176.86
176.87
176.86
176.81
0.02
0.03
0.02
0.02
0.02
0.02
0.01
0.01
0.01
0.01
0.00
-.01
0.00
0.01
0.01
0.01
0.02
0.01
0.01
0.02
0.00
0.00
0.01
0.02
0.01
0.00
0.00
0.01
0.02
0.02
0.01
-------
Table A-5. Daily model output - river discharge option,
FT. GRATIOT
ST.CLAIR RIVER TRANSIENT MODEL
JUNE 1986
to LAKE ST. CLAIR (SCS)
24.0 HOUR TIME INCREMENTS
179 REACHES
RIVER DISCHARGE VERSION
I RIVER PROFILE | (--TOTAL FLOW---J
AL
FLO*
6090.
6690.
6561 .
6602.
6631.
6557.
6490.
6627 .
6587 .
6487.
6476.
6325.
6565.
6513.
6454 .
6666.
6657.
6544 .
6553.
6643.
6438.
6497.
6626.
6808.
6681 .
6511 .
6651 .
6699.
6806.
6746.
6592.
SCS
MEAS.
1
2
3
4
6
6
7
8
9
10
11
12
13
14
16
16
17
18
19
20
21
22
23
24
26
26
27
20
29
30
AVE
176
17S
175
176
175
175
175
176
175
175
176
175
175
175
175
176
176
175
176
175
175
176
176
176
176
176
176
176
175
175
176
.66
.00
.00
.64
.68
. 70
.69
.69
.69
.69
.71
.74
.76
.77
.78
.74
.76
.76
.76
.78
.77
.74
.74
.73
.74
.72
.71
.73
.73
.73
.72
SC
COMP.
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
176
.26
.26
.23
.22
.26
.26
.24
.26
.26
.24
.26
.26
.30
.31
.31
.30
.32
.29
.30
.33
.29
.28
.30
.32
.31
.27
.28
.30
.32
.31
.28
FG
MEAS.
177
177
177
177
177
177
177
177
177
177
177
176
177
177
177
177
177
177
177
177
177
177
177
177
177
177
177
177
177
177
177
.06
.00
.01
.02
.05
.03
.01
.06
.04
.00
.02
.98
.07
.06
.06
.09
. 11
.06
.06
.11
.03
.04
.08
.14
. 10
.03
.08
. 10
. 14
.12
.06
FG
FLOW
6690.
0690.
6639.
6604 .
6647.
6552.
6485.
6638.
6582.
6480.
6488.
6319.
6690.
6607.
6464.
6664 .
6666.
6527.
6562.
6658.
6414.
6600.
6633.
6816.
6671.
6497.
6663.
6704.
6812.
6740.
6693.
,
STAG E
q
2676.
2676.
2616.
2639.
2656.
2621.
2693.
2662.
2633.
2592.
2593.
2529.
2631.
2605.
2682.
2664 .
2666.
2613.
2623.
2661.
2B70.
2598.
2652.
2724 .
2671.
2600.
2662.
2681.
2724.
2698.
"--- AJ1_«I
STAG W
q
4016.
4016.
3928.
3963.
3987.
3934.
3893.
3980.
3952.
3891.
3892.
3796.
3947.
3907.
3873.
3997 .
3998.
3920.
3936.
3991.
38S4.
3899.
3978.
4086.
4006.
3902.
3995.
4022.
4086.
4046.
l_un;> •
FAWN E
q
838.
838.
820.
828 .
830.
820.
812.
830.
824.
812.
810.
790.
821.
812.
805.
833.
831 .
817.
819.
829.
803.
812.
828.
861 .
834.
813.
832.
837.
860.
842.
FAWN W
q
6852.
5862.
6730.
6773.
6805.
6737.
6677.
6798.
6763.
6674.
5669.
6635.
5747.
6702.
6649.
6830.
5829.
5724.
6736.
6817.
6632.
6684.
5799.
6967.
6847.
6696.
5820.
5864.
6956.
6903.
1
N.CH.
q
2119.
2119.
2077.
2087 .
2090.
2069.
2047.
2091 .
2074 .
2048.
2040.
1988.
2049.
2038.
2024.
2094 .
2070.
2061.
2053.
2077.
2022.
2040.
2074.
2136.
2094.
2046.
2086.
2097.
2138.
2119.
l/CL. 1 «
M.CH.
q
901.
901 .
683.
886.
889.
881.
872.
890.
883.
872.
869.
849.
878.
872.
867.
896.
886.
881.
877.
889.
865.
872.
886.
912.
894.
873.
890.
896.
913.
906.
S.CH.
q
912.
912.
897.
896.
890.
849.
828.
857.
843.
836.
829.
771.
728.
714.
708.
740.
716.
748.
766.
734.
709.
736.
764.
788.
779.
766.
791.
789.
809.
808.
i
CUTOFF
q
2768.
2768.
2096.
2735.
2756.
2768.
2744.
2789.
2786.
2732.
2734.
2716.
2910.
2887.
2866.
2936.
2982.
2869.
2866.
2940.
2844.
2853.
2912.
2974.
2913.
2829.
2886.
2917.
2946.
2912.
i "«-
MEAS.
176.27
176.26
176.26
176.20
176.28*
176.28
176.28
176.29
176.29
176.27
176.30
176.32
176.35
176.36
176.34
170.36
176.36
176.33
176.33
176.36
176.32
176.32
176.34
176.35
176.34
176.32
176.34
176.35
176.36
176.34
¥ — |
DEV
C-M
-0.02
0.00
-0.02
-0.03
-0.02
-0.02
-0.03
-0.03
-0.03
-0.03
-0.04
-0.07
-0.06
-0.04
-0.03
-0.0S
-0.03
-0.04
-0.04
-0.03
-0.02
-0.04
-0.04
-0.03
-0.04
-0.06
-0.06
-0.04
-0.03
-0.03
2637.
3956.
824.
5769.
2072.
884.
797.
2639.
176.32 -0.03
-------
Table A-6. Daily model output - river velocities option.
FT. GRATIOT
ST.CLAIR RIVER TRANSIENT MODEL
JUNE 1986
to LAKE ST. CLAIR (SCS)
RIVER VELOCITY VERSION
24.0 HOUR TIME INCREMENTS
179 REACHES
| RIVER PROFILE I
I--TOT. VEL.--I
DEV 1
1
2
3
4
6
e
7
B
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
26
29
27
28
29
30
SCS
MEAS.
175.68
176.66
175.66
175.64
176.68
176.70
175.69
175.69
175.69
175.69
176.71
175.74
175.7S
176.77
175.78
175.74
175.76
175.76
176.76
176.78
176.77
176.74
175.74
175.73
176.74
175.72
176.71
176.73
175.73
175.73
SC
COMP.
176.26
176.25
176.23
176.22
176.26
176.26
176.24
176.26
176.26
176.24
176.26
176.26
176.30
176.31
176.31
176.30
176.32
176.29
176.30
176.33
176.29
176.28
176.30
176.32
176.31
176.27
176.28
176.30
176.32
176.31
FC.
MEAS.
177.06
17 7.06
177.01
177.02
177.05
177.03
177.01
177.06
177.04
177.00
177.02
176.98
177.07
177.06
177.05
177.09
177.11
17 7.06
177.06
177.11
177.03
177.04
177.08
177.14
177.10
177.03
177.08
177.10
177.14
177.12
FC
VEL.
1 . 11
1.11
1 .00
1 . 10
1 .10
1 .09
1.08
.10
.09
.08
1.08
.05
.09
.08
1.07
1.10
1.10
1
1
1
08
09
10
1.06
.08
.10
1.12
AVE 176.72 176.28 177.06
1 . 10
1.08
1.10
1.11
1.12
1.11
1 .09
AL
VEL.
0.71
0.71
0.69
,70
.70
0
0.
0.
0.69
0.69
70
0.70
0.69
0.68
0.67
0.69
0.68
0.68
0.70
0.70
0.69
0.69
0.70
0.67
0.68
0.70
0.72
0.70
0.69
0.70
0.70
0.72
0.71
0.69
1 »
STAG E
V
1.03
1 .03
1 .01
1 .02
1 .03
1 .01
1 .00
1 .02
1 .02
1 .00
1 .00
0.98
1 .01
1 .00
0.99
1 .02
1 .02
1.01
1 .01
1 .02
0.99
1 .00
1 .02
1 .04
1 .03
1 .00
1 .03
1 .03
1 .04
1 .03
1.02
«1U 13U«WU
STAC W
0.
0.
0.
e.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
V
90
90
89
89
90
89
88
90
89
88
88
86
89
87
87
89
89
88
88
89
86
88
89
91
90
88
90
90
91
90
89
VC.LUV-1 1 1
FAWN E
V
0.51
0.61
0.60
0.50
0.50
0.60
0.49
0.60
0.60
0.49
0.49
0.48
0.60
0.49
0.48
0.60
0.60
0.49
0.49
0.60
0.4B
0.49
0.60
0.61
0.60
0.49
0.60
0.51
0.61
0.61
0.B0
CO |
FAWN W
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
' 0
0
0
0
1
0
0
V
.99
.99
.97
.97
.98
.96
.96
.98
.97
.95
.96
.93
.96
.95
.94
.98
.97
.96
.96
.97
.94
.96
.97
.00
.98
.96
.98
.98
.00
.99
.97
| nnx
N.CH.
V
0.82
0.82
0.80
0.81
0.81
0.80
0. 79
0.80
0.80
0.79
0.78
0.76
0.78
0.78
0.77
0.80
0.79
0.79
0.78
0. 79
0.77
0.78
0.79
0.82
0.80
0.78
0.80
0.80
0.82
0.81
0. 79
V 1/U.U 1 r-*
M.CH.
V
0.56
0.56
0.64
0.66
0.66
0.54
0.53
0.54
0.54
0.63
0.53
0.52
0.53
0.63
0.62
0.64
0.54
0.53
0.63
0.64
0.62
0.63
0.54
0.66
0.64
0.63
0.64
0.64
0.66
0.66
0.64
T »_t_ V^ A » *
S.CH.
V
0.31
0.31
0.31
0.31
0.30
0.29
0.28
0.29
0.29
0.28
0.28
0.26
0.26
0.24
0.24
0.26
0.24
0.25
0.26
0.25
0.24
0.26
0.26
0.27
0.26
0.26
0.27
0.27
0.27
0.27
0.27
WW |
CUTOFF
V
0.88
0.88
0.86
0.87
0.88
0.88
0.87
0.89
0.88
0.87
0.87
0.86
0.92
0.91
0.90
0.93
0.94
0.90
0.90
0.93
0.89
0.90
0.92
0.94
0.92
0.89
0.91
0.92
0.93
0.92
0.90
1
SC
MEAS .
176.27
176.26
176.26
176.26
176.28*
176.28
176.28
176.29
176.29
176.27
176.30
176.32
176.36
176.36
176.34
176.36
176.36
176.33
176.33
176.36
176.32
176.32
176.34
176.36
176.34
176.32
176.34
176.36
176.36
176.34
176.32
DEV
C-M
-0.02
0.00
-0.02
-0.03
-0.02
-0.02
-0.03
-0.03
-0.03
- .03
- .04
- .07
- .06
- .04
- .03
- .06
- .03
- .04
- .04
- .03
- .02
- .04
- .04
- .03
-0.04
-0.06
-0.09
-0.04
-0.03
-0.03
-0.03
-------
ST. CLAIR AND DETROIT RIVER CURRENT MEASUREMENTS
Jan A. Derecki, Kathleen A. Darr, and Raymond N. Kelley
ABSTRACT
Velocities in the unregulated Great Lakes connecting channels, the St.
Clair and Detroit Rivers, were continuously measured with current meters
during an experimental field program. The program was initiated to improve
determination of winter flows, when the accuracy of normal flow
determinations is affected by ice. This study describes the experimental
results of continuous flow measurements using electromagnetic current meters
and an acoustic Doppler current profiler meter during the 1983-87 period of
data collection. Verification of current meter results was provided by
model-simulated flows during open-water periods and flow transfer between
the rivers during winter, when at least one of the rivers was ice-free.
Results indicate that accurate estimates of mean river velocities (and
consequently discharge) can be obtained with a single well-placed current
meter. However, the electromagnetic current meters are a direct-contact
single-point sensors that are affected by frazil ice during winter and weed
effects during most of the year, producing frequently questionable or
erroneous data. The acoustic profiler is a remote sensor of velocities in
the overhead water column and is not affected by the frazil ice and weed
problems, producing superior data.
92
-------
INTRODUCTION
Flows in the unregulated Great Lakes connecting channels, the St. Clair
and Detroit Rivers (Figure 1), are normally determined by either stage-fall-
discharge equations or unsteady flow numerical models. The calibration of
both the equations and models is based on periodic discharge measurements
taken over the years by the Corps of Engineers (COE) during the open-water
seasons (spring, summer, and fall). Consequently, the calculated flows
normally exhibit good accuracy during ice-free periods, but may contain
large errors during winter months with extensive ice cover. The St. Clair
River is particularly prone to large ice jams because of practically
unlimited ice flow supply provided by Lake Huron and an extensive river
delta that retards the passage of these ice flows. Large ice retardation of
flows in the St. Clair River is relatively frequent during winter months.
The magnitude of larger ice retardations generally approaches about 20% of
normal flow (1,100 c3/s), but in extreme cases has been observed to approach
50%. The ice conditions in the St. Clair and Detroit Rivers are different,
because of large difference in the upstream lakes and ice supplies, and the
ice problem also contributes to large discrepancies in the simulated flows
for these rivers. In some cases, these flow discrepancies exceed 20% of
total flow, which exceeds acceptable errors by an order of magnitude.
The St. Clair and Detroit River flows are determined at the Great Lakes
Environmental Research Laboratory (GLERL) with unsteady flow models, with
the current model versions described by Derecki and Kelley (1981) for the
93
-------
St. Clair River, and Quinn and Hagman (1977) for the Detroit River. These
rivers generally do not freeze over and are frequently free of ice during
winter. During such ice-free periods the models are adequate for the
simulation of winter flow rates. However, the models are not calibrated for
additional flow resistance due to ice, because of lack of proper data, and
tend to greatly overestimate the river flows during heavy ice accumulations.
The ice covers in these rivers are transient in nature, formed by the
consolidation of ice flows supplied by the upstream lakes due to ice break-
up by winter storms or spring thawing. In both instances proper
meteorological conditions are required to produce heavy ice concentrations
in the rivers. Generally southern storms are needed to destroy ice bridges
(Figure 2) which normally form at the heads of the rivers and help keep the
rivers free of ice, with subsequent shift to northerly winds that can force
large amounts of ice flows into the river channels.
Knowledge of accurate flows during both open-water and ice-covered
periods in the St. Clair and Detroit Rivers is needed for a variety of
scientific and water resource studies, ranging from water balance and
chemical/biological loadings to lake regulation, lake level forecasts, and
winter navigation. Large discrepancies in winter flows are associated with
abnormal river profile on the St. Clair and/or Detroit Rivers that exist
during ice conditions. A lack of measured data on river velocities and
water levels at critical interim points makes it impossible to determine
whether one or both river models are in error during winter. To address
this problem, a field measurement program was implemented in the St. Clair
and Detroit Rivers. The program tests the applicability of using
94
-------
continuously recording current meters to provide accurate velocity
measurements on an ongoing basis, independent of river ice conditions.
FIELD MEASUREMENT PROGRAM
The velocity and thus flow (discharge) of the upper St. Clair and
Detroit Rivers were continuously measured using current meters. Practical
requirements dictated the use of current meters without moving parts (to
avoid clogging), that are capable of prolonged operation (six months) at
frequent sampling rates. The initial phase of the field program on the St.
Clair River contained a pilot study, started in 1981, which provided for
familiarization and field testing of equipment. The actual data collection
program started in 1983, following additional resolution of encountered
problems. During the Upper Great Lakes Connecting Channels Study (UGLCCS)
activities, begun in 1985, the field program was in its second phase,
started in 1984, with simultaneous measurements of point-velocities in both
rivers and selective measurements of vertical velocity profiles in one of
the rivers. The current-meter stations were located in the upper portions
of both rivers, close to the COE flow measurement sections. These sections
of the rivers have fairly steep hydraulic gradients and are normally free
from consolidated ice cover. The meters were permanently deployed at the
river bottom and connected by cable to shore-located data recording
stations. This arrangement permitted remote access by telephone via
teletype-recorder to both the meters and their individual data records. The
operation of the current meters was monitored daily to detect and correct
95
-------
any instrument problems in order to eliminate or reduce data gaps. Ice
conditions in the rivers were also monitored during winter and checked as
needed by periodic ice surveys conducted by car or plane. Collected
velocity data were stored in computer files and subjected to routine
preliminary analysis, including comparison with model-simulated flows.
Point Measurements
After examination of the types of meters available, an electromagnetic
(EM) current meter was selected (Marsh McBirney, Model 585) for the first
phase of the program limited to the St. Clair River. The standard meter was
modified to include an externally located recording system (Figure 3), which
provided unlimited continuous operational capacity at a cable-connected
recording system (cassette tapes) located on the shore. After field testing
and several meter modifications, the in situ field operations were started
in September 1981, with the deployment of two EM current meters in the upper
St. Clair River, near the head of the river at Port Huron, MI (Figure 2).
The meters were installed on the United States side of the river, outside of
navigation channel about 50 and 70 m from shore, in 13 and 15 m of water.
Meter sensors were positioned 2 m above the bottom. Deployment and
subsequent removal of meters took place with the assistance of the USCGC
Bramble and a commercial diver, who guided the underwater operation. The
Detroit District of COE also participated in the project by making discharge
measurements during the open-water seasons. These measurements were
intended to provide data for calibration of the point-velocities measured by
the meters with the mean river velocity at the meter location. They were
96
-------
not used in this study, because several conducted measurements either
encountered operational problems or indicated considerable discrepancy in
the data.
The EM current meters in this study sampled ambient river velocity at
one-second intervals for Y- and X-axis velocity components, and an azimuth
angle. These raw data were converted to the north and east velocity
components, which were recorded with an accompanying azimuth angle at
15-minute intervals. The 15-minute input data were stored in a computer
file and converted to hourly and/or daily resultant velocity magnitude and
direction. The field seasons during the first phase of the program normally
covered late fall, winter, and spring months (November-June). The meters
were redeployed for the 1982-83 and 1983-84 winter seasons. However,
velocity measurements during the first two seasons contained some unresolved
problems and questionable data, and were excluded from this study. High
quality river velocity measurements during the 1983-84 season (obtained with
one of the meters) coincided with the record St. Clair River ice jam of
April 1984. This jam, which lasted nearly the entire month (April 5-29),
established records for both magnitude and lateness of occurrence, and
provided an excellent opportunity for testing the current-meter program.
This record ice jam and other aspects of the field measurement experiment
are discussed in previously published papers (Derecki and Quinn, 1986a,
1986b, and 1987).
The second phase of the study included simultaneous velocity
measurements in both rivers, starting in November 1984, with redeployment of
97
-------
meters in the St. Clair River for the 1984-85 winter season. The Detroit
River installation consists of two EM current meters, which were deployed in
August and tested during the summer of 1984 in the upper portion of the
river at Fort Wayne COE Boatyard (Figure 4). The two meters were installed
outside of navigation channel about 60 and 90 m from the United States
shore. Meters were placed in 12 and 14 m of water with upward positioned
sensors 2 m above the bottom. Similar operation and deployment procedures
were used on both rivers, with the USCGC Mariposa or the USCGC Bristol Bay
providing assistance in the Detroit River. During this phase of the study
the meters were not removed for the summer but were left operating
throughout the year to test the effects of weed transport and accumulation
on the velocity measurements.
Vertical Profile Measurements
Initial point-velocity measurements indicated a need for vertical
distribution of velocities, and recent advances in acoustical
instrumentation (Doppler-shift sensors) made such measurements practical.
Consequently, the St. Clair River installation was augmented during the
November 1984 redeployment with one acoustic Doppler current profiler (ADCP)
meter (RD Instruments, Model 1200 RDDR), which permits measurements of
velocities at approximately 1 m intervals in almost the entire vertical
water column (Figure 5). The ADCP meter was installed between the two EM
current meters, about 60 m from shore in 14 m of water (Figure 3). The
meter housing was oriented horizontally and the upward-looking sensor was
connected by a 90-degree elbow about 0.5 m above the bottom. The ADCP meter
98
-------
samples remotely vertical velocities in the overhead water column with
continuous sound waves (pings) from four beams at a rate of five times per
second, starting about 1 m above the sensor. The raw data from the four
beams are averaged to produce Y- and X-axis velocity components, along with
an azimuth angle, for approximately 1 m increments of depth to the surface.
These data are converted to the north and east velocity components for the
1-m progressive data segments, and indicate velocities at the mid-points of
each vertical segment. In a total water depth of about 14 m, this procedure
provided vertical velocity and direction values for 11 levels between
approximately 2.5m above the bottom and 0.5 m below the surface. The
surface readings are eliminated because of large data scatter at the air-
water interface (sound speed is about 5 times faster in water than in air).
The data were recorded at a cable- connected shore station at 15-minute
intervals (similar to the EM current meters).
These remote-sensing instruments are expensive but provide continuous
measurement capability that can not be duplicated with a string of point-
measuring meters because of navigation and ice problems near the surface.
The ADCP meter was removed from the St. Clair River in April 1986. It was
used during summer on the Detroit River in a demonstration of moving-boat
measurements (in June), and later (November) deployed in a normal-bottom
position on that river in place of the outer EM meter, 90 m from shore
(Figure 4). The ADCP meter change was made to provide vertical velocity
profile measurements in both rivers. Additional requirement for such data
in the Detroit River are the reversals of its flow, which occur occasionally
99
-------
because of the combined effects of storms on Lake Erie and ice jams in the
St. Clair River.
ST. CLAIR RIVER DATA
Electromagnetic Current Meter Records
During the period of study, data were collected from the EM current
meters for nearly three and a half years on the St. Clair River. These data
underwent preliminary analysis and comparison with model simulated flows.
The meters' operation was monitored daily to detect and correct any
instrument problems in order to eliminate or reduce data gaps. The water
level gages on the river were also monitored daily to detect ice effects on
the river's profile and several ice surveys were conducted, when ice
problems were indicated by this process.
Operation of the current meter program and monitoring of the meter
records indicated that frazil ice affects the operation of the EM current
meters. Although frazil ice episodes in the St. Clair River (later
confirmed on the Detroit River) are relatively infrequent (about 5 to 10
occurrences on each river per winter), they drastically affect the meter
data, which have to be corrected by elimination of bad data records. The
formation of frazil ice is a supercooling phenomenon, with distinct
characteristics, and can be easily identified. During cold spells in the
winter months (December-February), an additional sudden drop in temperatures
100
-------
causes the formation of frazil ice. This jelly-like ice formation is sticky
and adheres to objects; it coats the meter sensors, reducing their
sensitivity and producing low readings, at times approaching zero. The
sudden drop in the EM meter velocities associated with frazil ice normally
starts after sunset (before midnight) and disappears rapidly after sunrise
(before noon). However, severe episodes of frazil ice may last continuously
for a few days at a time.
Serious weed effects on the EM current meter operations were not at
first apparent during the initial phase of the program, limited to the St.
Clair River, because the meters were deployed in late fall (November), the
Lake Huron water is relatively clean, and the water velocity is high in the
upper river. These factors contributed to reduced weed accumulation around
the sensors. However, definite weed problems were encountered during the
subsequent prolonged operations, particularly in the summer and fall seasons
during continuous annual operations of the meter program. Weed accumulation
reduces meter readings and requires divers to inspect and clean the sensors
at frequent intervals for reliable data records. The EM current meter
velocity records taken immediately before and after cleaning of sensors by
divers (on both rivers) indicate that weed accumulation may reduce meter
velocities by as much as 25-50%. The records also show that this weed
accumulation may occur in only a few days, following deployment or cleaning
of meters. However, weed accumulation is generally gradual and difficult to
identify during initial stages. Since diver operations are expensive and at
times not feasible, this type of meter is not generally suitable for
101
-------
prolonged/continuous operations in rivers with high weed content,
particularly during the high weed transport season.
Discussion and presentation of the EM current meter results in this
report are limited to selected episodes which illustrate the nature and
quality of data. The first drastic episode of this type is the record April
1984 ice jam on the St. Clair River. This jam vividly demonstrates the
effectiveness of the in situ current meter velocity measurements in
estimating the river flows. The collection of high quality-current meter
data during the ice jam represents a major accomplishment and invaluable
information on the winter flow regime of the St. Clair River. Results from
one of the current meters in operation at that time are indicated in Figure
6, which shows the effect of the ice jam on the upper river flows (velocity
and direction). The meter velocity was reduced by about 50% during most of
April, changing near the river bottom at the meter location from about 1.0
to 0.5 m s~l. Higher velocities at the beginning of May, following the jam
breakup, were produced by the increased head (water level difference)
between Lakes Michigan-Huron and St. Clair. Records from the second meter
during this period showed other/additional effects, which were later
determined as weed effects.
Verification of the current meter results on the St. Clair River during
the ice jam episode is provided by flow transfer from the Detroit River,
which was free of ice during April and provided accurate flow simulation
with a numerical model. Conversely, good agreement in derived flows by two
independent methods demonstrates that the St. Clair-Detroit River flow
102
-------
transfer method is a very useful technique, provided that one of the rivers
is free of ice problems. Comparison of flows transferred from the Detroit
River with the St. Clair River flows derived from the current meter
measurements is shown in Figure 7. Extrapolation of the average river
velocity and discharge from the current-meter point-measurements is
discussed in the following paragraphs. The transfer factor, shown in the
figure, represents a summation of the hydrologic factors that determine the
difference between the flows in the St. Clair and Detroit Rivers, namely,
the precipitation on Lake St. Clair plus tributary runoff minus evaporation
from the lake and the storage of water on the lake. The agreement between
the meter and transferred flows is good during most of the March-May period,
particularly during the ice jam in April. In the few instances when the two
sets of flows deviate substantially, it is probably the transferred flows
that are in error. Thus, the high peak in transferred flow at the beginning
of May is caused by an extremely high storage of water on Lake St. Clair,
which appears to be overestimated. Larger deviations at the beginning and
during the second week of March appear to be caused by model oversimulation
of the Detroit River flows, probably due to the presence of some ice in the
river (March ice cover was not observed).
Comparison of the current meter velocities with the St. Clair River
numerical model results during the first deployment period (November 1983
-July 1984 field season), expressed as a ratio of model to meter velocity,
is shown in Figure 8. As expected, the figure shows a complete breakdown of
the St. Clair River model following the development of the ice jam in April.
During other times, the normal model-meter relationship is reasonably
103
-------
consistent and first-cut estimates of the average river velocity at the
meter location could be obtained by applying the velocity ratio to the
point-measurements of the meter (CM#1) . The other meter (CM//2) shows weed
effects during March-June period, with reduced meter readings and
exaggerated model ice effects in April. The relationship between the normal
model and the weed-free meter velocities for the 1983-84 deployment period,
after elimination of the bad model results in April, is indicated in Figure
9. The two equations shown in the figure are for a linear regression of the
data points (least squares) and for a velocity forced through a zero-
intercept. The equations agree closely and either one could be used to
produce acceptable average river velocities. The equation constant from the
zero-intercept equation also agree very closely with the reciprocal of the
average model-to-meter ratio (Figure 8). The high correlation coefficient
(0.94) indicates that over 88% (R squared) of the variation between the
average river velocity (simulated by model) and the current-meter velocity
measured at a single point near the river bottom is explained by a simple
regression.
Determination of river discharge, based on measurements (Figure 7), was
made by multiplying derived average river velocities from current meters by
the corresponding cross-section areas, obtained from model computations.
These areas were readily available, since in either velocity extrapolation
method (ratio or regression) the flows (discharge or velocity) were also
simulated by the models. At the meter location, most changes in the river
discharge are produced by corresponding changes in velocity, and errors
introduced in the derived discharge due to omission of the corresponding
104
-------
cross-section area changes are relatively small. In the most extreme cases,
connected with prolonged-massive ice jams, the velocity and corresponding
discharge changes (reduction) could exceed SOX; similar cross-section area
and corresponding discharge changes would be under 5X. During large ice
jams, to which the St. Clair River is particularly prone, the above meter-
derived flows represent a tremendous improvement over uncorrected model
results, which may oversimulate actual flows by a factor of two (CM#1 in
Figure 8). Availability of similar measurements during such ice jam
episodes (provided meter readings are not affected by weeds), especially in
conjunction with flow transfers (if feasible), may provide acceptable flow
estimates.
Simultaneous operation of the current meter program on both rivers
throughout the year and monitoring of the meter records indicated the
seriousness of weed effects on the EM current meters. Severe weed effects,
especially after storms or other sudden surges, can be as dramatic as those
of frazil ice, but generally weed accumulation is gradual and may fluctuate
in severity. An attempt was made to keep the EM current meters free of weed
problems with periodic cleaning of meter sensors by divers, but was
generally unsuccessful. Primarily because of weeds, the EM current meter
field program was generally unsuccessful on the St. Clair River for
prolonged periods during the second continuous deployment spanning several
field seasons and a few meter changes because of instrument problems
(November 1984-June 1987). This is indicated by the model-meter velocity
ratios shown in Figure 10. Drastic weed-effect problems during a summer
season (May-October, 1986) are indicated in Figure 11, which shows the
105
-------
effects of small but gradually increasing weed accumulation during May and
June, some recovery in July, and a massive-sudden weed clogging of the
meters' sensors in mid-August that remained in effect until the cleaning of
meters in November.
Acoustic Current Profiler Records
The ADCP meter was deployed in the St. Clair River in November 1984 and
operated until April 1986, for nearly a year and a half long data period.
Data collected with this instrument are unaffected by the frazil ice and
weed problems, most likely because of the meters' physical characteristics.
Both the outgoing and reflected sound waves travel through any frazil ice
coating the sensor. The same applies to weed accumulation. Meter
characteristics also permit its deployment in a low-profile horizontal
position on a support structure designed to reduce weed accumulation. Very
little weed accumulation was actually observed by divers during inspections.
This eliminates data gaps during winter and questionable or outright
erroneous data periods during heavy weed transport/accumulation (summer-fall
and after storms). The upper St. Clair River vertical velocity profile
measurements obtained with this meter represent high quality, unique data
not previously available on the Great Lakes connecting channels. The
profiler is expensive but produces a data set which could not be duplicated
with a dozen of the EM current meters, since they could not be deployed at 1
m intervals and operated continuously near the surface throughout the year
(navigation and ice problems). The quality of ADCP meter data is also
better. The following discussion and presentation of the profiler results
106
-------
is limited to a few data samples that illustrate the nature and quality of
collected data.
The vertical distribution of velocity in the water column measured with
the profiler during June 1985 is indicated in Figure 12. It gives the
progression of daily velocities at 11 levels with 1-m depth increments
between approximately 2.5m above the bottom and 0.5 m below the surface,
the practical limits of vertical measurements for the water depth of about
14 m. The figure shows a high degree of consistency between velocities at
different depths throughout the month. This consistency indicates that good
estimates of velocities in the entire water column or at different depth
levels could be obtained with single point-measurements, such as those made
with the EM current meters (provided problems are eliminated) . Highest
velocities normally occur near the surface, with a smooth progression of
increasing velocities from the bottom towards the surface, unless surface
flow is opposed by substantial wind shear. With strong counter-current
winds (southerly), which are generally limited to relatively short periods,
the velocity near the surface is occasionally retarded sufficiently so that
the highest velocity occurs 2-3 m below the surface. A more frequent
occurrence is the nearly uniform velocity in the top water layer spanning a
few (occasionally several) meters.
The smooth transition of velocities between progressive water layers is
indicated even more vividly in Figure 13, which shows two vertical velocity
profiles. A typical high-velocity profile is shown by June 10, 1985, which
was selected because of sharp increase in velocities on that day (Figure
107
-------
12); the March 5, 1985, profile was added to show a typical low-velocity
profile. Despite rapid change in velocities on June 10, the graph shows an
extremely smooth transition in the vertical distribution of velocities. The
use of daily velocities provided some smoothing of the graphs, but generally
similar profiles are obtained for shorter periods (hourly and 15-minute
data). To extend the profiles to the bottom and the surface, where
velocities could not be measured, these points were estimated and
incorporated in the graphs. The surface point was estimated by extending
the curve indicated by the preceding three measured points to the surface.
The bottom point was estimated by forcing a similar curve near the bottom
through a maximum-depth and zero-velocity intercept. The profiles show that
the vertical velocity distribution is definitely exponential (logarithmic),
which agrees with theoretical derivations for turbulent flow (Prandtl, 1925;
vov Karman, 1934). This includes most of the depth but excludes the
boundary layer, where the distribution can not be logarithmic because of
theoretical considerations.
Verification of the high consistency of velocities at different levels,
indicated in the preceding figure, is shown in Figure 14 by a statistical
relationship between profiler velocities near the bottom (bin 1) and the
integrated average velocities (11 bins) for the eighteen-month period
(November 1984 - April 1986). The two equations shown in the figure are for
a linear regression of the measured data points (least squares) and for a
velocity forced through zero-intercept, which are nearly identical. Either
equation could be used to provide good estimates of the average vertical
velocity. The extremely high correlation coefficient for the least squares
108
-------
linear regression (R-0.99) indicates that almost all (nearly 992) of the
variation between the average vertical velocity and a single point
measurement near the river bottom is explained by a simple regression.
Comparison of the profiler velocities (near-bottom and integrated
average) with the St. Clair River average values (at this location), derived
with the unsteady flow numerical model for the November 1984 - April 1986
period and expressed as ratios of these velocities, is shown in Figure 15.
Larger variations or disagreements are seen during January and February,
when the model results contain substantial errors because of ice effects.
During other times the agreement is reasonably good and first-cut estimates
of the average river velocity (or eventually discharge) could be obtained by
applying the velocity ratios to the profiler measurements. The relationship
between normal model and profiler velocities (excluding the bad ice-affected
model results) for the same period is shown in Figure 16. Comparison of
results presented in Figures 14 and 16 indicates a considerable loss of
accuracy (23Z) for the estimates of average river velocity. However, the
correlation coefficient for these estimates is still reasonably high
(R=0.87). Even these estimates, obtained with a single meter, represent a
large improvement over the model-simulated results during winter months with
significant ice problems. The 1984-85 and 1985-86 winter seasons were
relatively uneventful (without large ice jams) and the ice effect indicated
in Figure 15 for the model-simulated flows represents approximately average
ice conditions.
109
-------
DETROIT RIVER DATA
Electromagnetic Current Meter Records
The period of record for the EM current meter data on the Detroit River
covered about 3 years (August 1984-June 1987). Problems with weed
accumulation for the EM current meters became readily apparent on the
Detroit River during the second phase of the program, when continuous meter
operation throughout the year were begun in the summer of 1984. The weed
content in the Detroit River is higher and the river velocities are lower
than in the St. Clair River, contributing to more weed accumulation and
higher weed effects. With the higher weed content, more problems were
encountered in the operation of the EM current meters on the Detroit River.
Primarily because of weeds, the EM current meter field program was generally
unsuccessful on the St. Clair River during most of the year (summer and
fall), but at least partially successful during winter, while on the Detroit
River it was completely unsuccessful throughout the three annual periods.
This is indicated in Figure 17, showing the model-meter velocity ratios for
the period of record. Generally, these ratios are not stable for any
extended period of time.
Acoustic Current Profiler Records
The quality of the ADCP meter data collected on the Detroit River
remained high, similarly to that from the St. Clair River. Profiler
operations were similar on both rivers, with about the same water depths and
110
-------
the vertical velocity measured at 11 levels or bins approximately 1 m in
depth. The quality of the Detroit River profiler data is indicated in
Figure 18, showing the model-profiler velocity ratios for the period of
study (November 1986 - June 1987). The figure shows small ice-effect
problems affecting model-simulated velocities during winter, which is
typical for the Detroit River. Large ice jams occurred during this winter
on the St. Clair River, but its EM current meters were generally affected by
weeds and indicate biased ice effect.
The current meter field experiment was terminated in June 1987 with the
removal of the EM current meters in both rivers. The ADCP meter was left in
place in the Detroit River to continue the study of its flow reversals.
More detailed data analysis from the experimental field program will be
conducted next year. Its primary purpose will be to develop a method for
correcting the unsteady flow model simulation during winter periods with
substantial ice problems.
SUMMARY AND RECOMMENDATIONS
Flows in the St. Clair-Detroit River system, the outlet from the upper
Great Lakes, are needed for a variety of hydraulic and water resource
studies. Applications include hydrologic water balance, lake regulation,
lake level forecasts, navigation, transport of pollutants, recreation, and
consumptive water use. During the open-water season, acceptably accurate
estimates for these flows are provided with available mathematical unsteady-
Ill
-------
flow models. However, these models may produce large errors during winter
months when rapid transport of ice flows causes formation of ice jams in the
lower river reaches. The St. Clair River is particularly prone to large ice
jams because of the potentially large ice flow supply from Lake Huron and an
extensive river delta which retards the passage of these ice flows. Flow
estimates during ice conditions can best be obtained from in situ current
meter measurements.
Analysis of data collected during the 1983-87 period indicates that
acceptable estimates of river flows can be obtained with a single, well-
placed current meter. However, the EM current meters are susceptible to
frazil ice problems during winter, and to weed effects during most of the
year, making them of dubious value on rivers with high weed content, such as
the Detroit River. These problems can be avoided with the ADCP meter, which
is not affected by the frazil ice and weed effects and produces better
quality data for nearly the entire water column. The vertical velocity
profiles measured with the ADCP meter show a high consistency in an
exponential vertical distribution of velocities.
Because of the high quality of data for the overhead water column,
deployment of the ADCP meters should be considered by agencies responsible
for flow measurements in large rivers, such as COE for the Great Lakes
connecting channels. Data from such meters could be collected either
continuously or on demand. With proper calibration, the ADCP meters may
provide a suitable substitute for the labor-intensive periodic measurements
112
-------
now conducted by these agencies. The quality of such measurements would
also be higher.
ACKNOWLEDGEMENTS
From the beginning of the current-meter field experiment a number of
people contributed to the successful operation of the project. The authors
express their thanks to all the contributors and acknowledge specifically
their contributions. Several people from GLERL Instrument Lab, namely, H.K.
Soo, R.D. Kistler, R.W. Muzzi, T.C. Miller, and J.E. Dungan are acknowledged
for their extensive involvement in instrumentation and field operations
throughout the project or its portions. The study was conducted by GLERL's
Lake Hydrology Group with Dr. F.H. Quinn as Head, initially by A.J. Potok
with assistance from J.J. Kolodziejczak. The initial program was
extensively modified and field tested prior to data collection phase. At
different times throughout the study assistance in data analysis was
provided by M. Moliassa, S.R. Bonema, W.P. Moore, B.M. Slizewski, B.E.
Short, and D.A. Buckwald, all part-time temporary employees. Finally, the
authors thank the U.S. Ninth Coast Guard District for the deployment support
provided on both rivers, without which this project would not have been
possible.
113
-------
LITERATURE CITED
Derecki, J.A., and R.N. Kelley. 1981. Improved St. Clair River dynamic
flow models and comparison analysis, NOAA Tech. Memo. ERL GLERL-34.
Derecki, J.A., and F.H. Quinn. 1986a. Natural regulation of the Great
Lakes by ice jams: a case study. Proceedings, 4th Workshop on
Hydraulics of River Ice, Service Hydraulique, Ecole Polytechnique de
Montreal, Quebec, June 19-20, Vol. 1, pp. F4.1-F4.24.
Derecki, J.A., and F.H. Quinn. 1986b. The record St. Clair River ice jam
of 1984, J. Hyd. Eng.. 112(12):1182-1194.
Derecki, J.A., and F.H. Quinn. 1987. Use of current meters for continuous
measurement of flows in large rivers. Water Resour. Res.,
23(9):1751-1756.
Prandtl, L. 1925. Bericht uber utersuchungen zur ausgenbildeten
turbulenz, Z. angew. Math, u. Mech.. 5(2):136.
Quinn, F.H., and J.C. Hagman. 1977. Detroit and St. Clair River transient
models, NOAA Tech. Memo. ERL GLERL-14.
von Karman, Th. 1934. Turbulence and skin friction, J. Aeronaut. Sci..
114
-------
LIST OF FIGURES
1. St. Glair-Detroit River system with location of water level gages.
2. The point-measuring EM current meter and support.
3. Locations of St. Glair River current meters and ice bridge.
4. Location of Detroit River current meters.
5. The ADCP meter with diagram showing its remote-sensing operation.
6. St. Glair River EM meter data centered around ice jam, March-May, 1984.
7. St. Clair and Detroit River flows, March-May, 1984.
8. St. Clair River ratios of model to EM meter velocity during first
deployment period, November 1983 - July 1984.
9. St. Clair River ratios of model to EM meter velocity during second
deployment period, November 1984 - June 1987.
10. Relationships between St. Clair River normal model and EM meter
velocity, November 1983 - July 1984.
11. St. Clair River EM meter and model velocities showing weed-effect
problems, May - October 1986.
12. Vertical distribution of daily velocity, June 1985.
13. Vertical velocity profiles, March 5 and June 10, 1985.
14. St. Clair River profiler relationship between near-bottom and average
vertical velocity, November 1984 - April 1986.
15. St. Clair River ratios of model to profiler velocity during deployment
period, November 1984 - April 1986.
16. Relationship between St. Clair River normal model and profiler
velocity, November 1984 - April 1986.
115
-------
17. Detroit River ratios of model to EM meter velocity during deployment
period, August 1984 - June 1987.
18. Detroit River ratios of model to profiler velocity during the period of
study, November 1986 - June 1987.
116
-------
.LAKE HURON
Ft. Gratioti
Dunn Papers
Mouth of Black River<
Dry
fPt Edward
Marysvillei
SCALE IN MILES
•:
KILOMETERS
Michigan
St. Claire
\o
JK
CO
^ '.-• Roberts;//".'.
^Landing//. .
'Port Lambton
3\lgonacJ
K
St. Clair Shores^
Grosse Point <
Windmill Point.
LAKE
ST. CLAIR
Ft.
>La Salle
WyandotteW. Mpr
Tecumsetv-
ERIE
\
FBelle River .
Ontario
-------
PORT
HURON
Blue Wtter
Current
113
-------
• 4 Probe
EM Sensor
EM Meter
- Support Clamp
Support Legs
Power/Data
Cable
3 RR Wheel Support
119
-------
DETROIT
Fort Wayne
120
-------
Horizontal
VWocJty
Components
Resolution
and
Bin Range
Cells (-
121
-------
200
08
U
O
i i
i
> 0.4
0.0
VELOCITY
DIRECTION
12O
11 II 11 10 20
MARCH APRIL
1984
-------
8000
4000-
— 0
^ 3000
2
o
rr1 -3000
"• 8000
4000-
DETROIT (MOOEL)
11 21
MARCH
A ^ A
V v
TRANSFER FACTOR
y— . .-w_..
10 20
MAY
123
-------
or
LJ
U_
O
5.0
4.5
4.0
3.5
3.0
2.5
2.0
1.5
1.0-
0.5-
0.0-
CM*{
1 6 11 16 2126 1 6 11 16212631 5 10 1520?5304 9 M 1924295 10 15202530 4 9 14 192*29 4 9 U 1924293 8 13182328 3 8 13182328
NOV DEC JAN FEB MAR APR MAY JUN JUL
1983 1984
-------
1.5
o>
> i
U
o
_l
ui
ac
OJ
= '0.941
0.867*X
0434- 0.832*X
o.o
0.6 1.0 1.5
MODEL VELOCITY (m/s)
2.0
-------
LJ
cr
LJ
LJ
d
§
b
Q
5.0
4.5-
4.0-
3.5-
3.0-
>
r^ 2.5 H
2.0 H
1.5
1.0 H
0.5
0.0
MODEL/C.MJ1
MODEL/C.M.#2
NDJFMAMJJASONDJFMAMJJASONDJFMAMJ
1984 1985 1986 1987
-------
1.75
1.50-
1.25-
1.00-
0.75-
0.50-
0.25
0.00
A \ *V * I'S \
/ \ A- .V/-.Y • • y .
CM#1 VELOCITY
CM#2 VELOCITY
MODEL VELOCITY
i i i i i i i T r r T i T i i i i I i i i i r i i i r i i I i i v i T r
1 6 11 16 21 26 31 5 10 15 20 25 30 5 10 15 20 25 30 4 9 14 19 24 29 3 8 13 18 23 28 3 8 13 18 23 28
MAY JUNE JULY AUGUST SEPTEMBER OCTOBER
1986
-------
u
3
IU
NEAR SURFACE
THRU
NEAR BOTTOM
11 II 1?
JUNE
21 » It 27 2ft
128
-------
. MARCH 5, 1985
JUNE |10, 1985
0.6 1 1.5
VELOCITY (m/s)
129
-------
<0 2.0-
O
o
LU
O
h-
o
CO
oc
LU
z
1.5-
R = 0.993
Y = 0.879*X
Y =-0.011 4- 0.888*X
0.0 0.5 1.0 1.5 2.0
AVERAGE VERTICAL VELOCITY (m/s)
-------
-------
2.0n
t 1.6H
O
O
—I
uu
> 1.0H
cc
uu
0.5-
LL
O
oc
Q.
0.0
0.0
x
X
R = 0.869
Y = 0.973*X
Y = 0.117 + 0.882*X
0.5 1.0 1.5
MODEL VELOCITY (m/s)
2.0
-------
5.0
CO 4.5
UJ
t:
Q 4.0
3.5-
3.0-
LJ
0
3
LU
_J
LU
Q
O
U_
O
-------
5.0
4.5-
4.0-
3.5-
3.0-
2.5-
2.0-
1.5-
1.0-
0.5-
0.0
SOLID LINE - MODEL/BIN #1 RATIO
DASHED LINE - MODEL/INTEGRATED AVG. RATIO
1 6 11 16 21 26 1 6
NOV
1986
11 16 21 26 31 5 10 15 20 25 30 4 9 M 19 2* 1
DEC JAN FEB
11 16 21 26 31 5 10 15 20 25 30 5 K) 15 20 25 30 * 9 M 19 24 29
MAR APR MAY JUN
1987
-------
DEVELOPMENT OF A SHALLOW WATER NUMERICAL WAVE MODEL FOR LAKE ST. CLAIR
David J. Schwab and Paul C. Liu
In the summer and fall of 1985, the National Water Research Institute
(NWRI) and GLERL participated in an extensive field measurement program in
Lake St. Clair. The availability of high-quality, over-lake meteorological
data and the other physical measurement systems in a relatively flat and
shallow basin with sufficient fetch to generate substantial waves provided a
perfect opportunity to measure wave dissipation and the effect of waves on
resuspension in shallow water. On June 19, 1985, scientists from GLERL,
NWRI, and AES (Atmospheric Environmental Service) met and planned a joint
program entitled Wave Attenuation, Variability, and Energy Dissipation in
Shallow Seas (WAVEDISS '85) for the fall of 1985 in Lake St. Clair to make
these measurements.
An inventory of readily available equipment limited the number of
recording stations to six, three each from GLERL and NWRI. The GLERL
stations consisted of a single Zwarts transmission wave staff and a Datawell
Waverider radio transmitter. The NWRI stations consisted of a triangular
array of capacitance wave gauges, a cup anemometer, and a SeaData recording
package. The instrumentation systems were attached to towers which were
guyed to a base anchored with railroad wheels. The towers were deployed
along a transect parallel to the prevailing storm wind direction (Fig. 1).
One tower was deployed on a perpendicular leg to give an indication of
135
-------
cross-transect variability. The separation distances between the towers
were based on expected differences in energy for a 10 m/s wind compared to
the sampling variability of the estimates of wave energy. A sampling rate
of 4 Hz was selected based upon expected wave periods. Wave record length
was set at 4096 samples, or about 17 minutes at 4 Hz sampling rate. This
record length was a compromise between recording capacity of the SeaData
recorders and the statistical reliability of wave energy estimates.
The GLERL instruments transmitted continuously to a shore station at
Stoney Point, Ontario where a microcomputer stored one wave record from each
station per hour. The microcomputer was periodically interrogated by
telephone to retrieve the data on the GLERL VAX computer. The NWRI
instruments recorded burst samples on even-numbered hours when the
anemometer reading exceeded a present threshold value. The triangular array
of capacitance gauges provided wave direction estimates in addition to the
wave energy spectrum. The experiment ran from September 20 - December 2,
1985. Nearly continuous measurements of significant wave height and wave
period (at hourly intervals) were obtained from the GLERL towers. The NWRI
towers recorded data only when the wind speed exceeded a preset threshold
(usually 7 ms'l) so the data covers only the higher wave height periods.
Meteorological data consisting of hourly values of wind speed, wind
direction, air temperature, and water temperature were obtained from the
NWRI meteorological buoy at 42.5 degrees north, 82.8 degrees east. The
meteorological data was used to drive the GLERL-Donelan numerical wave
prediction model on a 1.2 km numerical grid. This model was developed by
Donelan (1977) and used successfully to predict wave height and wave
136
-------
direction in Lake Erie (Schwab et al., 1984) and Lake Michigan (Liu et. al.,
1984) . The model is a parametric model based on a momentum balance equation
for the wave field. The model predicts the two components of the wave
momentum vector and the phase speed of the peak energy waves. From these
variables, significant waveheight, wave period, and wave direction are
derived. In the mathematical formulation of the numerical model, the waves
are assumed to obey the deep water dispersion relation. Refraction and
bottom dissipation are ignored.
As part of the UGLCCS, the GLERL-Donelan model has been modified to
account for the effect of finite water depth on wave propagation by
incorporating the Kitaigorodskii et al. (1985) shallow water wave spectrum
along with a depth-dependent group velocity and a simple form of bottom
friction. This shallow water version of the model was also run with the
same grid and same wind input as the deep water version. The results for
hourly values of significant waveheight are compared to observations at the
six towers in Figures 2 and 3. The statistical comparison in terms of root
mean square error and correlation coefficient is presented in Table 1.
As can be seen in Table 1 and Figures 2 and 3, the deep water version
of the model provides quite acceptable estimates of waveheight, even for the
largest waves at the shallowest stations. The shallow water version of the
model tends to underestimate the highest waves at all stations. The shallow
water model could be adjusted to better match the observed waveheights by
decreasing the bottom friction parameter, but the best it could do would be
no better than the deep water model.
137
-------
Work is underway now to determine why the deep water model works in
Lake St. Clair. One possibility is that the wind momentum input function in
the model is oversimplified and if it were formulated more realistically,
the deep water model would tend to overestimate the highest waves. This
possibility is being investigated, but for now, the deep water model appears
to be quite acceptable for providing wave height estimates in Lake St.
Clair.
138
-------
LITERATURE CITED
Donelan, M. A., 1977. A simple numerical model for wave and wind stress
prediction. Unpublished manuscript. National Water Research
Institute, Burlington, Ontario, Canada. 28pp.
Kitaigordskii, S.A., V.P. Krasitskii and M.M. Zaslavskii, 1975: On
Phillips' theory of equilibrium range in the spectra of wind-generated
gravity waves. J. Phys. Oceanogr.. 5, 410-420.
Liu, P. C., Schwab, D. J. and Bennett, J. R., 1984. Comparison of a two-
dimensional wave prediction model with synoptic measurements in Lake
Michigan. J. Phys.Oceanogr. 14:1514-1518.
Schwab, D.J. Bennett, J.R., Liu, P.C. and Donelan, M.A., 1984. Application
of a simple numerical wave prediction model to Lake Erie. J. Geophys.
Res., 89(C3):3586-3592.
139
-------
Table 1. Comparison of Deep and Shallow Water Model Predictins with
Measured Significant Waveheight
STATION
C3
Depth (m)
Data Points
Deep Water Model
rmse (m)
corr. coeff.l
Shallow Water Model
rmse (m)
corr. coeff.l
6,
.7
244
0,
0,
0,
0,
.11
.89
.13
.88
U2
7
.0
1366
0
0
0
0
.11
.93
.13
.93
C2
6
.4
237
0
0
0
0
.16
.88
.20
.85
Ul
5
.5
1478
0
0
0
0
.09
.93
.11
.91
Cl
3.
7
153
0.
0.
0.
0.
10
94
18
92
U3
4.4
1539
0.08
0.94
0.09
0.93
•"•correlation between predicted and measured values.
140
-------
LIST OF FIGURES
Fig. 1. Station locations for GLERL and NWRI wave towers during WAVEDISS
'85.
Fig. 2. Comparison of observed and predicted wave height for the deep
water model.
Fig. 3. Comparison of observed and predicted wave height for the shallow
water model.
141
-------
LAKE ST CLAIR
WAVE STATIONS
(
11
GLERL
U2".NWRI
C2-
-------
Lake St. Clair 1985 Station VI
CM
10
•g
m
13 O
CL
1=
0 0.5 1 1.5 2
Observed Waveheight (m)
Lake St. Clair 1985 Station C /
-------
Lake St. Clair 1985 Station
CM
*'
Q)
.n
O
0 0.6 1 1.6 2
Observed Waveheight (m)
Lake St. Clair 1985 Station Cl
\A^
I-
s.,
3 O
a
E
T*^
•».
0 0.6 1 1.6 2
Observed Waveheight (m)
Lake St Clair 1985 Station o
.g>52
s
0 0.6 1 1.6 2
Observed Waveheight (m)
Lake St. Clair 1985 Station CZ_
''
0 0.5 1 15 2
Observed Waveheight (m)
Lake St. Clair 1985 Station
S1" U5
O
Q.
0 0.5 1 1.6 2
Observed Waveheight (m)
Lake St. Clair 1985 Station
.
0)
rs o
Q.
0 05 1 1.5 2
Observed Waveheight (m)
-------
MODELING PARTICLE TRANSPORT IN LAKE ST. CLAIR
D. J. Schwab and A. H. Clites
The results of a numerical circulation model of Lake St. Clair are
used to describe particle transport pathways in the lake in terms of
residence time and variability due to wind-induced circulation.
Specifically, we address the following questions. 1) What path does water
entering Lake St. Clair from one of the tributaries follow through the lake
before leaving at the Detroit River? 2) How long does it take? 3) How is
the path changed by wind-induced circulation in the lake? 4) For the
meteorological conditions during the summer and fall of 1985, what are
typical statistical distributions of these pathways?
In order to answer these questions, a numerical circulation model of
the lake was used. The numerical model is the same time-dependent rigid-
lid model developed by Schwab et al. (1981) and used by Schwab (1983) in
Lake Michigan and Schwab and Bennett (1987) in Lake Erie. The basic
assumptions of the model are that the circulation is barotropic and non-
divergent, and that nonlinear acceleration and horizontal diffusion of
momentum are negligible compared to first-order acceleration and Coriolis
forces. A simple quasi-linear formulation is used to describe bottom
friction. The model is forced by the hydraulic flow through the lake and
by time-dependent wind stress at the surface. The hydraulic flow is
assumed to be constant in time at 5700 m3 s'l and the inflow is divided
145
-------
among the tributaries as follows: North Channel, 35Z; Middle Channel, 20Z;
St. Clair Flats, 20X; St. Clair Cutoff, 20X; Bassett Channel, 5X.
Currents are calculated on a 1.2 km grid approximating the shape of
Lake St. Clair (Fig. 1). Simons and Schertzer (1986) and Ibrahim and
McCorquodale (1985) have also developed numerical circulation models for
Lake St. Clair and have obtained essentially similar results for the wind-
induced circulation. However, the circulation patterns generated by the
numerical models often differ considerably from the results of Ayers'
(1964) physical model of the lake for some wind directions.
After currents are calculated for each grid square in the numerical
model, another model is used to move tracer particles through the lake.
The particles are assumed to follow exactly the vertically uniform currents
without sinking or diffusing. The numerical model for particle trajectory
calculations was developed by Bennett et al. (1983) and used by Schwab and
Bennett (1987) in Lake Erie. The model uses a second-order method to
compute particle trajectories and takes special care to realistically
represent the currents near the shoreline. As shown by Bennett and Clites
(1987) , this method is far more accurate than simple first order methods
yet is only slightly more complex computationally. This particle
trajectory model is also used in the "Pathfinder" trajectory prediction
system (Schwab et al., 1984) that is used by the National Weather Service
and the U.S. Coast Guard for tracking hazardous spills and search and
rescue missions on the Great Lakes.
146
-------
In order to answer questions about residence time for water entering
the lake from the various tributaries, several model runs were carried out
with idealized wind conditions. First, the hydraulic flow in the absence
of wind was calculated (Fig. 2). This pattern was then used as the initial
condition for a series of simulated storms with wind directions from the
eight compass points and peak wind speeds of either 10 or 20 ms'l. Figure
3 shows a graph of the time dependence of the wind for the storms. The
wind speed increases linearly from zero to its peak value over a period of
two days and then ceases. Particles from the five tributaries listed above
plus the Clinton River, the Clinton River Cutoff, and the Thames River were
released into the resulting circulation pattern starting at the beginning
of the second day, when the wind speed was at exactly half its peak value.
One particle was released every three hours for one day. These particles
were then tracked through the lake and their residence time was compared to
the residence time for particles in the absence of wind (purely hydraulic
flow). The results for the 10 ms"1 storms are summarized in Figure 4. The
dashed line in the panel for each tributary represents the residence time
for water entering there in the no-wind case. It can be seen that even
though the average hydraulic residence time for Lake St. Clair is about
nine days (based on hydraulic flow of 5700 m3s-l-, surface area of 1100 km2,
and a mean depth of 4 m), residence time for water from the individual
tributaries ranges from 4.1 days for the Middle Channel to over 30 days for
water from the Thames River. The calculated residence times for all the
tributaries in the no-wind case are:
147
-------
North Channel: 4.5 days
Middle Channel: 4.1 days
St. Clair Flats: 5.0 days
St. Clair Cutoff: 8.3 days
Bassett Channel: 24.1 days
Clinton River: 6.9 days
Clinton Cutoff: 9.5 days
Thames River: over 30 days
The effect of wind on the residence time is greatest for the Thames
River and the Bassett Channel where a 10 ms'l wind from the SE, E, or NE
can decrease the expected residence time to less than 10 days for Bassett
Channel or 15 to 20 days for the Thames. A 20 ms'l wind reduces the
residence time even more. Winds from W, NW, or N tend to increase the
residence time of water from St. Clair Flats and more considerably from the
St. Clair Cutoff and Bassett Channel by moving the water eastward into the
relatively stagnant eastern part of the lake. The residence time of water
entering the lake from the Clinton River is increased by NE, E, SE, and S
winds and decreased by SW, W, and NW winds. Wind from almost any direction
except NW tends to decrease the residence time of water from the Clinton
Cutoff.
The idealized wind condition calculation give some indication of the
variability in residence time that can be expected in Lake St. Clair due to
wind-induced circulation, but what kind of variability actually occurs
during the summer and fall? To answer this question, hourly values of wind
148
-------
speed and wind direction recorded at the CCIW met buoy at 42.5 degrees
north, 82.8 degrees east for the period 23 May, 1985 - 1 December, 1985
were used to drive the numerical circulation model. The resulting
circulation patterns were stored and then used with the particle trajectory
model to calculate the paths of tracer particles released at the mouth of
each tributary once every six hours during the entire six month period.
The calculated tracks of the particles were then used to develop
probability plots of the likelihood of a parcel of water emanating from one
of the eight tributaries passing through a given area of the lake during
this period. These plots quantify the wind-induced variability in the
pathway that water from one of the tributaries takes through the lake. The
results of these calculations are presented in Figure 5 in terms of
probability contours. The outermost line is the 99.9X contour, i.e., 99.9X
of the conservative particles released from that tributary remained within
this contour for the 6 months simulation. The next contour delineates the
area in which 90X of the particles remained. Remaining contours are at 10%
intervals.
Most of the water from the St. Glair River enters the lake through the
North Channel (35X). According to the calculations, this water tends to
flow down the western shore of the lake and never gets into the central or
eastern parts of the basin. Water from the Middle Channel tends to remain
in the western third of the lake, almost never entering the eastern half.
Water from St. Clair Flats and the St. Clair Cutoff can be dispersed almost
anywhere in the lake to the south of the shipping channel which connects
the St. Clair Cutoff with the Detroit River. A small amount of the St.
149
-------
Clair inflow (5X) enters through Bassett Channel. This water can pass
through any part of the eastern half of the lake depending on the wind
conditions. The Thames inflow tends to be confined to the eastern and
southern shores before reaching the Detroit River and it can take a very
long time to get there (see Fig. 4). Water from the Clinton River and
Clinton Cutoff is most likely to follow the western shore of the lake
southward with the most probable paths within 3 km of the western shore.
Water quality measurements made in Lake St. Clair by Leach (1972 and
1980) showed two distinctly different areas in the lake. In the
southeastern part of the lake, the water quality is dominated by the Thames
inflow, which is a major source of phosphate and other dissolved and
suspended material. The central and western parts of the lake were more
similar to Lake Huron in terms of water quality than to the southeastern
part of the lake. The pattern of water mass distribution mapped in Leach's
(1980) Figures 1-4 is very close to the combined patterns of the four main
St. Clair River inflows and the Thames inflow in our Figure 5. Bricker et
al. (1976) examined the distribution of zooplankton in the western half of
the lake. They distinguished an area of biological and physiochemical
similarity along the western shore of the lake that appeared to be
influenced more by the Clinton River than the St. Clair River. The shape
of this area matches quite well with the distribution pattern for water
from the Clinton River in Figure 5 here.
To verify the circulation model and lend credence to the calculated
currents used in this study, the model was tested by comparing model output
150
-------
to actual current data measured in Lake St. Clair in 1985. Two separate
current data bases were gathered. One involved the use of 5 drifting buoys
which were repeatedly launched and tracked in the lake. The other was the
result of several synoptic current surveys utilizing electromagnetic
current meters.
Currents predicted by the circulation model were used to simulate 16
drifter tracks. Most of the tracks are about 2 days in length from various
portions of the lake. In most cases, the model simulated the tracks
extremely well. For the entire data set, the mean root mean square (rms)
of the drifter was 25Z greater than that of the calculated current track.
The directions compared favorably except for a few tracks near the mouth of
the Bassett Channel, where the model prediction was over 90 degrees
different in direction when compared with the observed track.
The comparisons between current meter measurements and model-predicted
currents were even better. In nearly 100 comparisons, 60% of the variance
is explained by the model prediction. The model again seems to under-
predict the current speeds, here by about 30%.
The results presented in this report are not water quality
calculations. They only track conservative, non-dispersive tracers from
the mouths of the tributaries through the lake under various wind
conditions. Work in another GLERL UGLCCS project couples the circulation
patterns calculated here with the TOXIWASP water quality model for Lake St.
Clair. However, based on the comparisons with actual current measurements
151
-------
presented above, the calculated currents provide a realistic depiction of
the wind-induced circulation in Lake St. Clair.
152
-------
LITERATURE CITED
Ayers, J. C. 1964. Currents and related problems at Metropolitan beach,
Lake St. Clair. Great Lakes Res. Div. Spec. Rpt. No. 20, Univ. of
Mich., Ann Arbor, Michigan.
Bennett, J. R. and elites, A. H. 1987. Accuracy of trajectory calculation
in a finite difference circulation model. J. Comp. Physics,
68:272-282.
Bennett, J. R., Clites, A. H., and Schwab, D. J. 1983. A two-dimensional
lake circulation modeling system: programs to compute particle
trajectories and the motion of dissolved substances. NOAA Tech. Memo.
ERL-GLERL-46, 51pp.
Bricker, K. S., Bricker, F. J., and Gannon, J. E. 1976. Distribution and
abundance of zooplankton in the U.S. waters of Lake St. Clair, 1973.
J. Great Lakes Res. 2:256-271.
Ibrahim, K. A. and McCorquodale, J. A. 1985. Finite element circulation
model for Lake St. Clair. J. Great Lakes Res. 11:208-222.
Leach, J. H. 1972. Distribution of chlorophyll a and related variables in
Ontario waters of Lake St. Clair, pp. 80-86. In Proc. 15th Conf. Great
Lakes Res. Internat. Assoc. Great Lakes Res.
153
-------
Leach, J. H. 1980. Limnological sampling intensity in Lake St. Clair in
relation to distribution of water masses. J. Great Lakes Res.
6:141-145.
Simons, T. J. and Schertzer, W. M. 1986. Hydrodynamic models of Lake St.
Clair. NWRI Contribution #86-xxx, National Water Res. Inst.,
Burlington, Ontario.
Schwab, D. J. 1983. Numerical simulation of low-frequency current
fluctuations in Lake Michigan. J. Phys. Oceanogr. 13:2213-2224.
Schwab, D. J. and Bennett, J. R. 1987. Lagrangian comparison of objectively
analyzed and dynamically modeled circulation patterns in Lake Erie. J.
Great Lakes Res. (in press)
Schwab, D. J., Bennett, J. R., and Jessup, A. T. 1981. A two-dimensional
lake circulation modeling system. NOAA Tech. Memo. ERL-GLERL-38, 79pp.
Schwab, D. J., Bennett, J. R., and Lynn, E. W. 1984. "Pathfinder"--a
trajectory prediction system for the Great Lakes. NOAA Tech. Memo.
ERL-GLERL-53, 37pp.
154
-------
LIST OF FIGURES
Figure 1. The 1.2 km numerical grid for Lake St. Clair. Arrows show the
inflows and outflow used in the numerical model.
Figure 2. Modeled hydraulic flow in Lake St. Clair. Streamlines indicate
10Z increments of stream function from -2850 to 2850 ms'l.
Figure 3. Wind history for residence time calculations.
Figure 4. Calculated residence time for water from eight tributaries for
simulated storm winds of 10 and 20 ms'l from eight compass
points. Dashed lines indicate no-wind residence time.
Figure 5. Probability distributions of water masses from eight tributaries
in Lake St. Clair based on model simulations of particle
trajectories for the period 23 May - 1 Dec, 1985.
Figure 6. Verification of circulation model based on drifter and current
meter measurements gathered in 1985. Directions compare
favorably except near the mouth of the Bassett Channel. Model
current speeds seem to be slightly underestimated based on the
data.
155
-------
Numerical Grid for
Lake St. Clair
No Wind
(2)
156
-------
Wind History for Residence Time Simulations
Ui
Particles Released
1 2 3
Time (days)
-------
Residence Time (days) in Lake St. Clair
for a 10 ms"1 storm
(dashed line indicates no-wind residence time)
North J
Channels
o
m M+ » it S
Middle g
Channels
o
St. Clair "
_. «
Flats o
o
/H» XNI * S
St. Clair g
Cutoff o
o
•
-
-•
...
—
_,
-
...
...
...
"
"1
-1
...
™
...
Bassett "
Channels
Clinton j
River s
x^i. O
Clinton ;
Cutoff o
-»-i O
Thames "
River s
o
.
-
-
-
'
••4
.«
i«_
Residence Time (days) in Lake St. Clair
for a 20 ms"1 storm
(dashed line Indicates no-wind residence time)
E S W N
North ;
g
Channels
o
fc 4* _1 Jl «
Middle p
Channels
Q
fH. /^| • S
St. Clair ;
Flats o
o
-
-
•
;
..j
,
;
>
.~>
i
i»"l
!
"
:— •
Bassett "
~>uw W >_r kV Q
Channels
-*... o
Clinton J
River o
Clinton ;
Cutoff o
o
*
-
-
•I
i
!
I
...
""""
St. Clair
Cutoff
Thames
River
8 -
158
-------
North Channel
BasseU. Channel
SL Clair Cutoff
159
-------
LAKE ST. CLAW
DRFTER STUDY
-Observed
D- Predicted
3-6 September 1985
LAKE ST. CLAIR
CURRENT SURVEY
Observed
Predicted
5 June 1985
Depth Integrated Currents
-------
TOTAL PHOSPHORUS BUDGET FOR LAKE ST. CLAIR: 1975 - 1980
Gregory A. Lang, Julie A. Morton, and Thomas D. Fontaine, III
ABSTRACT
As part of the U.S. - Canadian Upper Great Lakes Connecting Channels
Study a total phosphorus budget was developed for Lake St. Clair. An
unbiased ratio estimator technique was used to estimate annual loads and
variances from monitored hydrologic areas. During the 1975-80 period, Lake
Huron was the major source of phosphorus to Lake St. Clair, accounting for
approximately 52% of the total annual load. Hydrologic area loads, which
include diffuse and indirect point sources, contributed approximately 43% of
the total annual load. The remaining 5% came from the atmosphere, shoreline
erosion, and direct point sources. Of the hydrologic area loads, 85% could
be attributed to diffuse sources. The Thames area contributed 58% of the
total hydrologic area load, followed by the Sydenham (17%), the Clinton
(9%), the Ruscom (7%), the Black (6%), the St. Clair (3%), and the Rouge
(0.4%). Over the entire six year period examined, the lake's total input
and output of phosphorus were nearly equal. It was concluded that there was
no significant net source or sink of phosphorus in Lake St. Clair during the
1975-80 period.
161
-------
INTRODUCTION
Phosphorus budget calculations for the Great Lakes have generally
overlooked the dynamics of phosphorus transport into and through Lake St.
Clair. The sum of Lake St. Glair's point and non-point inputs have simply
been included as part of the total load to the Western Basin of Lake Erie; the
assumption has been that the entire input of phosphorus to Lake St. Clair is
transported, unaltered, through the lake and into the Detroit River. Recent
attention to the connecting channels of the upper Great Lakes, however, has
served as the impetus for calculating a phosphorus mass balance for Lake St.
Clair. The intent of this study was two fold: 1) to estimate the total
phosphorus budget for 1975-80, and 2) to determine whether or not the lake is
a net source or sink for phosphorus.
Lake St. Clair is unique against the pelagic backdrop of the Great Lakes
because it is shallow (average depth = 3.3 m) and has a very short average
hydraulic retention time (9 days). Lake St. Clair has a surface area of about
1100 km^ and a drainage area of about 15500 km^. We calculated the annual
phosphorus loads entering this shallow, quick flushing lake from a variety of
sources and compared them to the amount of phosphorus leaving the lake through
the Detroit River. From these comparisons along with estimates of data and
load uncertainty, we were able to draw certain conclusions concerning the
likelihood of net sources or sinks of phosphorus in Lake St. Clair.
162
-------
METHODS
The total phosphorus budget for Lake St. Clair includes external loads,
internal sources/sinks, outflow losses, and change in mass storage. At steady
state, the sum of the external loads and internal sources/sinks should balance
the outflow loss. External phosphorus loads to Lake St. Clair come from Lake
Huron, shoreline and streambank erosion, atmospheric sources, direct point
sources, and hydrologic area loads. Internal sources and sinks include
particle settling and resuspension, groundwater input, bioturbation, and
aquatic macrophyte uptake and release. The outflow loss is through the
Detroit River.
External Loading Estimates
Lake Huron loads were taken directly from Yaksich et. al. (1982). They
estimated the average daily loading for the head of the St. Clair River by
multiplying the flow weighted mean yearly phosphorus concentration by the
reported average daily flow per month. This daily load was adjusted by the
number of days in the month to yield a monthly load. Loading from shoreline
erosion was estimated by multiplying the length of Lake St. Clair shoreline by
the annual loading rate of phosphorus per kilometer of shoreline for the Lake
St. Clair basin (Monteith and Sonzogni 1976). Loading from streambank erosion
(along the St. Clair River) was assumed to be negligible because the total
loading to the Great Lakes from streambank erosion is < 4% of that contributed
by shoreline erosion on the U.S. side of the Great Lakes (Knap and Mildner
1978).
163
-------
Direct atmospheric loading was estimated to be the average (14.0 MT/yr)
of values obtained during the field seasons of 1975 (Delumyea and Petel 1977,
measurements made at stations around Southern Lake Huron) and 1981
(Klappenbach 1984, measurements made at Mt. Clemens, Michigan). Direct and
indirect point source load estimates were compiled from the International
Joint Commission (1982), the Great Lakes Basin Commission (unpub. data), and
the Ontario Ministry of the Environment (1985).
Hydrologic Area Loads
Seven hydrologic areas were defined using the convention of Hall et al
(1976): the Black, St. Clair complex, Clinton, and Rouge complex (does not
include the Rouge River) areas in Michigan, and the Ruscom, Thames, and
Sydenham areas in Ontario (Fig. 1). Hydrologic area loads include diffuse
loading from land areas that drain into a tributary or directly into Lake St.
Clair, and indirect point source inputs. Indirect point source loads are
those which discharge upstream of a river monitoring station. Their entire
load was assumed to be transported to Lake St. Clair. Direct point sources
were defined as those which discharge downstream of a river monitoring station
or directly into the St. Clair River or Lake St. Clair. Input from these
sources were not included as part of the hydrologic area loads. Several small
(< 1 MGD) point source discharges may have been omitted from our hydrologic
area analysis since scant information existed for them during 1975-80.
However, analysis of recent municipal point source data from STORET (U.S. EPA)
and industrial point source data from the Ontario Ministry of the Environment
(1985) shows that these sources presently represent about IX of all of the
164
-------
point source flows. Therefore, their omission should have little or no effect
on calculation of loads from hydrologic areas.
Each hydrologic area load equaled the sum of its monitored and unmonitored area loa
Monitored areas. Hydrologic area loads from monitored areas were calculated using
estimate the average daily load at the mouth, adjusted to minimize the
variance associated with the flow component, as follows:
Sxv/(nMvMx))
Ux • My/Mx • (1)
(1 + Sx2/(nMx2)
where Uy is the unbiased estimate of the daily load at the mouth, Ux is the
mean daily flow for the year, My is the mean daily loading for the days for
which concentration data exists, Mx is the mean daily flow for the days for
which concentration data exists, n is the number of days for which
concentration data exists, and
n
- nMyMx
_
Sxy = (2)
n-1
165
-------
n
S Xj.2 - nttx2
i-1 _
SX2 - -
n-1
where Xj. is the individual measured flow for each day for which concentration
data exists and Yj. is the loading for each day for which concentration data
exists (calculated as the product of the individual measured flow and the
total phosphorus concentration) . Since this load includes any indirect point
sources, the diffuse component of the hydrologic area load from a monitored
area was estimated by subtracting the indirect point source inputs from the
total calculated load.
The estimated mean square error of the estimated load (the square root of
which is the estimated standard error of the mean) was also calculated using
the ratio estimator method,
MSE - My2 - [ 1/n • (SX2/MX2 + Sy2/My2 - Sxy/(MxMy))
+ l/n2 • (2-(Sx2/Mx2)2 - 4.Sx2/Mx2.Sxy/(MxMy)
+ (Sxy/MxMy)2 + Sx2/Mx2.Sy2/My2)] (4)
where E is the estimated mean square error of the load estimator and Sy2 is
calculated analogously to SX2 . In an attempt to quantify some of the
uncertainty of these loads and to statistically compare the total annual
external load with the outflow loss, the root mean square error (RMSE) is used
in a later section to estimate 90% confidence intervals around the individual
tributary loads, the total external loads, and the outflow losses.
166
-------
Three hydrologic areas (Clinton, Thames, Sydenham) were 100X monitored
during the entire study period. Percent monitored represents the percent of
the hydrologic area that is monitored. The Black area was not monitored for
years 1975-77 and 100X monitored for years 1978-80. The St. Clair and Ruscom
areas were 36% (Belle R.) and 18Z (Ruscom R.) monitored, respectively, during
the study period. The Rouge area was not monitored. On an areal basis, 832
of the Lake St. Clair basin was monitored during the study period.
Flow data were obtained from the U.S. Geological Survey (1976-82) and the
Water Survey of Canada (1976-81). Flow measurements recorded at gaging
stations located upstream of the river mouth were corrected to the river mouth
by multiplying the gaged flow by the ratio of the entire drainage area to the
gaged drainage area as described in Sonzogni et al. (1978). Phosphorus
concentrations were obtained from the U.S. Environmental Protection Agency's
STORET system, the U.S. Geological Survey (1976-82), and the Ontario Ministry
of the Environment (1975-80). In most cases, water quality monitoring
stations were located at or near the river mouth. It was assumed that these
phosphorus concentrations equaled the concentrations at the mouth.
Unmonitored areas. Some hydrologic areas and individual tributaries were
not sufficiently monitored (fewer than 6 samples per year) over some or all of
the 1975-80 period. In addition, some land areas drain directly into Lake St.
Clair or the St. Clair River, and therefore cannot be monitored. Diffuse
loads for these watersheds were estimated by calculating a diffuse unit area
loading (UAL) (the diffuse load per unit area per unit time) for a monitored
area with basin characteristics similar to the unmonitored area. The
167
-------
selection of a representative monitored area was based on soil texture,
surface geology, runoff characteristics, land use, and proximity to the
unmonitored area. The diffuse load for the unmonitored area was estimated as
the product of the diffuse unit area load from the monitored area and the
unmonitored area as described in Sonzogni et al. (1978). Indirect point
source discharges in unmonitored areas were added to the estimated diffuse
load to yield total loads for the unmonitored area.
Total External Load
The total external phosphorus load to Lake St. Clair equaled the sum of
loads from the total hydrologic area, atmospheric sources, shoreline erosion,
direct point sources, and the Lake Huron. The variance associated with the
total external load was not known because some hydrologic areas were partially
or totally unmonitored and variance estimates were not available for the
atmospheric, erosion, direct point, or Lake Huron loads. However, an estimate
of the variance associated with the total external load was made using the
following procedure. Annual load probability distributions for each monitored
tributary and the outflow from Lake Huron were constructed with means set
equal to their estimated annual loads. The variances of the tributary load
distributions were calculated from their estimated RMSE (eq. 5). The variance
of the Lake Huron load distributions was assumed to be constant and equal to
the variance of the Huron load averaged over the six-year period. For each of
the six years, values from each distribution were randomly selected and summed
to give a single estimate the total annual external load (Figure 2). This
168
-------
procedure was repeated 12 times to yield a combined probability distribution
for each year's total external load. A sample size of 12 corresponds to the
average number of times per year that each tributary was sampled. The
standard error of the total annual external load was calculated from each
year's combined probability distribution for the purpose of statistical
comparison with the outflow data.
Outflow Loss
Annual phosphorus loss through the Detroit River and the mean square
error associated with this loss were estimated using the unbiased ratio
estimator technique. Phosphorus concentrations were measured along a transect
at Windmill Point near the head of the Detroit River by the Michigan
Department of Natural Resources (DNR) and recorded in EPA's STORET data base.
There are ten stations, each representing 10X of the flow of the river. The
data from these stations were limited to non-winter months (usually April
through October). Phosphorus data for the winter months were measured at the
Detroit municipal water treatment facility at Water Works Park and recorded in
the U.S. Geological Survey (1976-82). Water quantity data were taken from the
results of a hydrodynaroic model of the Detroit River (F. H. Quinn, Great Lakes
Environmental Research Laboratory, personal communication).
Internal Sources and Sinks
Scant quantitative information exists concerning the annual internal
input or loss of phosphorus in Lake St. Glair. It is estimated that some
169
-------
sources and sinks vary seasonally (e.g., particle settling and resuspension
and aquatic macrophyte uptake and release), while others remain more or less
constant throughout the year (e.g., groundwater input and bioturbation). For
the purpose of annual budget calculations, the amount of phosphorus added to
or lost from the lake via internal sources and sinks was estimated as the
difference between annual external loads and annual outflow losses.
RESULTS
The annual estimated river mouth phosphorus loads from monitored
hydrologic areas to Lake St. Clair as well as the annual estimated Detroit
River outflow losses are presented in Table 1. Table 1 also includes the
estimated RMSE of the estimated load and the 90% confidence intervals around
the estimated load. The RMSE and the Student's t-statistic were used to
calculate the 90% confidence intervals (Remington and Schork 1985).
The total annual phosphorus budget for Lake St. Clair is presented in
Table 2. On the average, Lake Huron contributed a major portion (about 52%)
of the external total phosphorus load. The contribution of hydrologic areas
was also significant and equaled approximately 43% of the external phosphorus
load to Lake St. Clair. The remaining 5% of the total annual load came from
the atmosphere (0.5%), shoreline erosion (2.6%), and direct point sources
(1.91). Average phosphorus dynamics of Lake St. Clair during the 1975-80
period are summarized in Figure 3 In the diagram, the relative proportion of
inflows and outflows are approximated by the thickness of the arrow shafts.
170
-------
Hydrologic area loads originating from Canadian sources averaged 82% of
the total hydrologic area load to Lake St. Clair. The Thames and the Sydenham
areas contributed 92% of the total Canadian hydrologic area load. The Black
and Clinton areas were responsible for 83% of the total U.S. hydrologic area
load. The largest individual hydrologic area loads originated from the Thames
area. In all years except 1975, loading from the Thames exceeded 50% of the
total hydrologic area load (the six-year average contribution was 58%). The
Thames area load was followed by the Sydenham (17%), the Clinton (9%), the
Ruscom (7%), the Black (6%), the St. Clair (3%), and the Rouge (0.4%).
Over the six year period investigated, about 85% of the total hydrologic
area load was calculated to originate from diffuse sources. The remaining 15%
came from indirect point sources. The diffuse portion of the Thames and
Sydenham area loads accounted for most of the total diffuse load (62% and 20%,
respectively). The Clinton and Thames areas contributed a majority of the
total indirect point source load (43% and 35%, respectively).
Internal Sources and Sinks
Assuming estimated loads presented in Table 2 are an accurate
representation of the actual loads and assuming steady state conditions, then
the difference between external loads and outflow losses is a measure of the
internal sources/sinks of phosphorus in Lake St. Clair. The total annual
incoming load falls within the 90% confidence interval around the Detroit
River outflow loss for all years except 1976 (Figure 4). The six-year mean
external load (3133 MT/yr) and loss (3148 MT/yr) were not found to be
171
-------
statistically different at the 10X level of significance (t-test, Remington
and Schork 1985). The root mean square difference between annual external
loads and outflow losses was only 7X of the mean external load (5* excluding
1976 values). Therefore, given the above assumptions and results, there
appeared to be no net internal source or sink of total phosphorus in Lake St.
Clair during 1975-80.
DISCUSSION
The second objective of this study was to determine whether or not Lake
St. Clair is a net source or sink for phosphorus. The total incoming load
fell within the 90% confidence interval around the Detroit River loss for all
years except 1976; and, averaged over 6 years, external loads were not
significantly different from outflow losses; implying no net internal source
or sink of phosphorus in the lake during 1975-80. Because the net internal
sources/sinks were calculated as the difference between the incoming loads and
outflow losses, their validity is limited by the accuracy and precision of the
load/loss estimates. If loads and losses are sufficiently accurate and
precise, then the conclusion that net internal sources/sinks were negligible
during the study period seems reasonable.
It is difficult to determine how accurate the estimated loads and losses
are since actual loads and losses can never be known exactly. Sonzogni et.
al. (1978) note that the RMSE terms in Table 1 are useful for statistical
comparisons, but they do not necessarily reflect how close the estimated load
172
-------
is to the true load. The RMSE is an estimate of the error determined from a
limited number of daily samples, based on the premise that the true annual
load can be determined by sampling flow and concentration at the river mouth
each day of the year. In addition, the method assumes that instrument and
measurement errors can be neglected and that instantaneous flow/concentration
measurements are true representations of the tributary conditions on that day.
As briefly stated previously, Dolan et. al. (1981) evaluated 10 tributary
load estimator methods and concluded that the ratio estimator method used in
the present study was the most suitable for application in the Great Lakes
basins. They recommended that this method be used to estimate tributary loads
for total phosphorus when concentration data are limited and the daily flow
record is available. The accuracy of the streamflow data used in this study
was rated "fair" to "good" by the U.S. Geological Survey. "Good" means that
about 95% of the reported daily discharges are within 10% of the actual values
and "fair" within 15%. Chemical data represent (as much as possible) water
quality conditions at the time of sampling, consistent with available sampling
techniques and methods of analysis. The phosphorus loads from Lake Huron,
estimated by Yaksich et. al. (1982), were consistent (trends and magnitude)
with other external loads to Lake Erie for 7 water quality constituents during
1970-80 and are thus, for lack of better criteria, considered representative
of the true Lake Huron loads. Therefore, given that 90% of the hydrologic
area load was from monitored tributaries and that 95% of the total external
load was from hydrologic areas and Lake Huron, we conclude that the calculated
total external phosphorus loads and losses are at least a fair representation
173
-------
and at best an accurate estimate of the actual loads and losses to and from
Lake St. Clair.
Finally, does this conclusion (i.e., negligible net internal
sources/sinks during the study period) seem reasonable given the nature of the
internal sources and sinks? Groundwater input of phosphorus is assumed to be
insignificant compared to that entering the lake from other sources. The
biological release of phosphorus from sediments and mussels to the overlying
water has also been shown to be negligible (Nalepa et. al. 1987). Assuming an
average apparent settling rate of 16 m/yr (Chapra 1977) and a range of lake-
averaged phosphorus concentrations measured during eight cruises in 1975 (US
EPA's STORET data base) yields a range of 253-321 MT of phosphorus potentially
lost to the sediments annually. These values represent only about 10X of the
total external load for any one year and would be lost in the variability
between input and output. Given the shallowness and high wave energy of Lake
St. Clair, sediment resuspension would reduce the impact of particle settling.
Robbins (1987) indicates that the net deposition of particulate matter to the
lake's sediments is small; the sediment thickness above post-glacial clay
ranges from 0 to 30 cm, corresponding to a net sedimentation rate of only
0.1-0.2 cm/yr. At this rate, a range of 0.01-0.04 MT of phosphorus/year would
be lost to the sediments, representing less than 1% of the total incoming load
for any one year. Therefore, it does seem reasonable that over a six year
period there would be no significant net source or sink of phosphorus in Lake
St. Clair.
174
-------
SUMMARY
The objectives of this study were two fold: 1) to estimate and present
the total phosphorus budget for 1975-80, and 2) to determine whether or not
the lake is a net source or sink for phosphorus. Lake Huron contributed over
half (522) of the lake's load while the seven hydrologic areas contributed 43%
of the remainder. About 92% of the total Canadian hydrologic area load is
attributable to the Thames and Sydenham areas of Ontario. The Clinton and
Black areas of Michigan were responsible for 83% of the total U.S. hydrologic
area load. Were reduction of phosphorus loads to Lake St. Clair deemed
desirable, control efforts might be best focused in these four hydrologic
areas. Because 85% of the total hydrologic area load is from diffuse sources,
a non-point source reduction plan might be most appropriate. Reduction of
municipal point sources along the Thames River and the Clinton River may also
be important as these sources contributed a majority of the remaining 15% of
the total hydrologic area phosphorus load.
Over the six year study period, 1975-80, the mean external load and
outflow loss of phosphorus were not found to be statistically different at the
10% level of significance. Assuming accurate estimates of the loads/losses
and steady state conditions, we conclude that there was no apparent net source
or sink of phosphorus in Lake St. Clair during the study period.
175
-------
ACKNOWLEDGEMENTS
This is GLERL contribution No. 511. This work was partially funded by
interagency agreement DW 13931213-01-0 with the Great Lakes National Program
Office, U.S. Environmental Protection Agency, Chicago. We thank Barry M.
Lesht, Douglas Salisbury, and two anonymous reviewers for reviewing an earlier
version of the manuscript.
176
-------
LITERATURE CITED
Beale, E.M.L. 1962. Some uses of computers in operational research.
Industrielle Organisation. 31:51-52.
Chapra, S.C. 1977. Total phosphorus model for the Great Lakes. J.
Environmental Engineering Div. ASCE. 103(2):147-161.
Delumyea, R.G, and Petel, R.L. 1977. Atmospheric Inputs of Phosphorus to
Southern Lake Huron. April-October 1975. U.S. Environmental Protection
Agency, Report number 600/3-77-038, Environmental Research Laboratory,
Duluth, Minnesota, 53 pp.
Dolan, D.M., Yui, A.K., and Geist, R.D. 1981. Evaluation of river load
estimation methods for total phosphorus. J. Great Lakes Res. 7:207-214.
Hall, J.R., Jarecki, E.A., Monteith, T.J., Skimin, W.E. and Sonzogni, W.C.,
1976. Existing River Mouth Loading Data in the U.S. Great Lakes Basin.
Pollution From Land Use Activities Reference Group Report, International
Joint Commission, Great Lakes Regional Office, Windsor, Ontario, 713 pp.
International Joint Commission. 1982. 1981 Municipal and Industrial Phosphorus
Loadings to the Great Lakes. Report of the Great Lakes Water Quality
Board to the International Joint Commission, Windsor, Ontario.
177
-------
Klappenbach, E. 1984. 1981 Atmospheric Loading for Lake Huron. Report to the
U.S. Environmental Protection Agency Great Lakes National Program Office,
Chicago, Illinois.
Knap, K.M., and Mildner, W.F. 1978. Streambank Erosion in the Great Lakes
Basin. Report of the International Reference Group on Pollution from Land
Use Activities to the International Joint Commission, Windsor, Ontario.
Monteith, T.J., and Sonzogni, W.C. 1976. United States Great Lakes Shoreline
Erosion Loadings. Report of the International Reference Group on Great
Lakes Pollution from Land Use Activities to the International Joint
Commission, Windsor, Ontario.
Nalepa, T.F., Gardner, W.S., and Malczyk, J.M. 1987. Phosphorus release from
sediments and mussels in Lake St. Clair, with notes on mussel abundance
and biomass. Upper Great Lakes Connecting Channel Final Report. Great
Lakes Environmental Research Laboratory, Ann Arbor, Michigan.
Ontario Ministry of the Environment. 1975. Water Quality Data: Ontario Lakes
and Streams. Volume 10. Report of the Water Resources Branch, Toronto,
Ontario.
. 1976. Water Quality Data: Ontario Lakes and Streams. Volume 11. Report
of the Water Resources Branch, Toronto, Ontario.
178
-------
1977. Water Quality Data: Ontario Lakes and Streams. Volume 12. Report
of the Water Resources Branch, Toronto, Ontario.
1978. Water Quality Data: Ontario Lakes and Streams. Volume 14. Report
of the Water Resources Branch, Toronto, Ontario.
1979. Water Quality Data: Ontario Lakes and Streams. Volume 15. Report
of the Water Resources Branch, Toronto, Ontario.
1980. Water Quality Data: Ontario Lakes and Streams. Volume 16. Report
of the Water Resources Branch, Toronto, Ontario.
1985. Upper Great Lakes Connecting Channels Study: Canadian Point Source
Discharge and Combined Sewer Overflow Activities. Report of the
Environmental Protection Service, Ontario Region, Environment Canada,
Ontario Ministry of the Environment, Windsor, Ontario.
Remington, R.D. and Schork, M.A. 1985. Statistics with Applications to the
Biological and Health Sciences. Prentice-Hall, Inc. Publ., Englewood
Cliffs, New Jersey. 415 pp.
Robbins, J.A. 1987. Accumulation of fallout cesium-137 and chlorinated organic
contaminants inn recent sediments of Lake St. Clair. Can. J. Fish, and
Aouatic Sci. In Press.
179
-------
Sonzogni, W.C., Monteith, T.J., Bach, W.N., and Hughes, V.G. 1978. United
States Great Lakes Tributary Loadings. Great Lakes Pollution from Land
Use Activities Reference Group, International Joint Commission, Technical
Report, Great Lakes Regional Office, Windsor, Ontario.
U.S. Geological Survey. 1976. Water Resources Data for Michigan. Water Year
1975. Report of the U.S. Geological Survey, Water Resources Division,
Lansing, Michigan.
. 1977. Water Resources Data for Michigan. Water Year 1976. Report of the
U.S. Geological Survey, Water Resources Division, Lansing, Michigan.
. 1978. Water Resources Data for Michigan. Water Year 1977. Report of the
U.S. Geological Survey, Water Resources Division, Lansing, Michigan.
. 1979. Water Resources Data for Michigan. Water Year 1978. Report of the
U.S. Geological Survey, Water Resources Division, Lansing, Michigan.
. 1980. Water Resources Data for Michigan. Water Year 1979. Report of the
U.S. Geological Survey, Water Resources Division, Lansing, Michigan.
. 1981. Water Resources Data for Michigan. Water Year 1980. Report of the
U.S. Geological Survey, Water Resources Division, Lansing, Michigan.
. 1982. Water Resources Data for Michigan. Water Year 1981. Report of the
U.S. Geological Survey, Water Resources Division, Lansing, Michigan.
180
-------
Water Survey of Canada. 1976. Surface Water Data. Ontario. 1975. Report of the
Inland Waters Directorate, Water Resources Branch, Water Survey of
Canada, Ottawa, Quebec.
. 1977. Surface Water Data. Ontario. 1976. Report of the Inland Waters
Directorate, Water Resources Branch, Water survey of Canada, Ottawa,
Quebec.
. 1978. Surface Water Data. Ontario. 1977. Report of the Inland Waters
Directorate, Water Resources Branch, Water survey of Canada, Ottawa,
Quebec.
. 1979. Surface Water Data. Ontario. 1978. Report of the Inland Waters
Directorate, Water Resources Branch, Water survey of Canada, Ottawa,
Quebec.
_. 1980. Surface Water Data. Ontario. 1979. Report of the Inland Waters
Directorate, Water Resources Branch, Water survey of Canada, Ottawa,
Quebec.
. 1981. Surface Water Data. Ontario. 1980. Report of the Inland Waters
Directorate, Water Resources Branch, Water survey of Canada, Ottawa,
Quebec.
Yaksich, S.M., Melfi, D.A., Baker, D.B. and Kramer, J.W. 1982. Lake Erie
Nutrient Loads. 1970-1980. Lake Erie wastewater management study. U.S
Army Corps of Engineers District, Buffalo, New York.
181
-------
Table 1. 1975-80 River Mouth Loadings. The load is presented in metric tons
per year, followed by the root mean square error in metric tons per year,
followed by the 90X confidence interval in Metric tons per year, followed by
the number of phosphorus samples.
Monitored
Tributary
Name 1975
Black NA1
Belle2 31.1
4.9
[22.2,
40.0]
11
Clinton 198.4
22.3
[158.3,
238.4]
12
Ruscom 3.8
0.1
[3.7,
3.9]
7
Thames 418.3
60.4
[309.9,
526.8]
12
Sydenham 196.0
28.6
[142.9,
249.1]
9
Detroit 2769.3
289.7
[2244.2,
3294.4]
11
1976
NA1
28.7
8.7
[13.1,
44.4]
12
143.8
24.0
[100.8,
186.9]
12
6.0
0.5
[5.1,
6.9]
11
690.8
160.6
[396.5,
985.1]
10
195.5
46.9
[110.5,
280.5]
12
3935.9
301.5
[3383.3,
4488.6]
10
Year
1977 1978
NA1
7.1
0.9
[5.5,
8.6]
12
118.3
16.9
[88.0,
148.7]
12
16.3
1.3
[14.0,
18.6]
11
1391.8
427.4
[625.1,
2158.4]
12
494.3
137.1
[248.4,
740.2]
12
3304.7
393.4
[2603.6,
4005.9]
13
47.4
2.9
[41.9,
52.9]
8
10.4
0.9
[8.8,
12.0]
12
112.7
23.3
[70.9,
154.5]
12
22.3
6.0
[11.5,
33.1]
12
643.3
58.1
[537.9,
748.7]
11
94.4
18.6
[61.1,
127.8]
12
3090.2
250.3
[2644.1,
3536.2]
13
1979
90.8
38.9
[21.0,
160.6]
12
11.9
2.1
[8.0,
15.7]
12
77.8
8.9
[62.0,
93.7]
12
9.0
1.8
[5.7,
12.3]
12
917.7
193.3
[587.0,
1248.4]
25
241.1
27.1
[195.0
287.2]
27
2879.7
247.9
[2448.5,
3310.9]
18
1980
88.4
29.0
[34.5,
142.4]
9
17.4
4.7
[8.7,
26.2]
9
114.0
28.9
[62.2,
165.8]
12
36.3
20.7
[0.0,
75.4]
8
663.8
102.1
[493.8,
833.8]
75
170.3
19.2
[137.7,
202.9]
31
2908.7
273.3
[2427.3,
3390.1]
15
^Phosphorus data not available for years 1975-77,
2within St. Clair Hydrologic Area.
182
-------
Table 2. Annual phosphorus budget (MT/yr) for Lake St. Glair (1975-80).
Values in parentheses represent percent of total load from diffuse sources,
Percent monitored refers to percent of hydrologic area that is monitored.
Source
Hydrological Areas
Black
0,100% monitoredb
St. Clairc
36% monitored
Clinton
100% monitored
Rouged
0% monitored
Ruscom6
18% monitored
Thames
100% monitored
Sydenham
100% monitored
Total HA Load
Atmospheric
Erosion
Direct Point
Lake Huron
1975
IHAI
115. 5a
(74)
75.4
(90)
198.4
(48)
10.8
(100)
21.8
(100)
418.3
(85)
196.0
(97)
1036.2
(80)f
14.0
82.5
55.8
2022
EXTERNAL LOAD 3211
90% C.I. [3014,
3408]
1976
84. 5a
(64)
68.7
(89)
143.8
(26)
4.2
(100)
34.1
(100)
690.8
(90)
195.5
(97)
1221.7
14.0
82.5
64.5
1373
2756
[2485,
3027]
1977
41. 9a
(28)
7.8
(3)
118.3
(21)
2.9
(100)
92.2
(100)
1391.8
(94)
494.3
(99)
2149.2
(90)f
14.0
82.5
58.9
1187
3492
[3070,
3914]
1978
47.4
(36)
17.0
(56)
112.7
(37)
4.7
(100)
126.4
(100)
643.3
(91)
94.4
(92)
1045.9
(83)f
14.0
82.5
57.7
1613
2813
[2648,
2978]
1979
90.8
(67)
21.2
(64)
77.8
(18)
1.6
(100)
51.1
(100)
917.7
(92)
241.1
(99)
1401.3
(87)f
14.0
82.5
59.2
1703
3260
[2987,
3533]
1980
88.4
(63)
38.2
(80)
114.0
(39)
5.0
(100)
205.7
(100)
663.8
(90)
170.3
(99)
1285.4
(86)f
14.0
82.5
54.0
1827
3263
[3136,
3390]
OUTFLOW LOSS
90% C.I.
IN - OUT
aUnit Area Load
UAL of Clinton
2769 3936 3305 3090 2880 2909
[2244, [3383, [2604, [2644, [2449, [2427,
3294] 4489] 4006] 3536] 3311] 3390]
441
-1180
186
-277
380
354
(UAL) assumed to equal average of UAL of St. Clair HA and
HA.
183
-------
bOX for years 1975-77, 100Z for years 1978-80.
CUAL assumed to equal UAL of Belle R. within St. Glair HA.
dUAL assumed to equal UAL of Clinton HA.
eUAL assumed to equal UAL of Ruscom R. within Ruscom HA.
^Percentages weighted with respect to percent of total annual input that is
attributable to a given area.
184
-------
LIST OF FIGURES
Figure 1. Lake St. Clair Basin showing hydrologic areas used in calculating
load estimates. Note: the Thames hydrologic area extends to
approximately 43" 30' latitude, 80° 30' longitude.
Figure 2. Schematic displaying procedure used to estimate the combined
probability distribution for each year's total external load.
Figure 3. Summary of Lake St. Clair average phosphorus dynamics during the
1975-80 period. Values are inMT/year. The relative proportion of
inflows and outflows are approximated by the thickness of the arrow
shafts.
Figure A. External phosphorus loads to Lake St. Clair (closed circles and
vertical bars represent 90X confidence interval around annual load,
MT/yr) and Detroit River outflow loss (shaded region represents 90%
confidence interval around annual loss, MT/yr).
185
-------
83°00'
82°30'
43°30'
43°00'
42°30'
82°00'
42°00'
Complex
136
-------
Clinton
Ruscom
Thames
Load
-------
Lake St. Clair average phosphorus loads and losses
during the 1975-'80 period (metric tonnes per year)
U.S.
Hydrologic
Areas
Atmospheric,
Erosion, Lake Huron
Direct Point 1,621
155™
Black
78
St. Clair
Complex 381
LAKE ST. CLAIR
Net Loss =
Canadian
Hydrologic
Areas
i232 Sydenham
i 788 Thames
Detroit River Outflow 3,148
188
-------
5000
ca
tn *•
2«
O o
4000-
« o
O -5=
p 3000
"5
_j i
Outflow Loss i
Load
1975 76 77 78 79
Time (years)
'80
189
-------
PHOSPHORUS RELEASE FROM SEDIMENTS AND MUSSELS IN LAKE ST. GLAIR,
WITH NOTES ON MUSSEL ABUNDANCE AND BIOMASS
T. F. Nalepa, W. S. Gardner, and J. M. Malczyk
INTRODUCTION
Since phosphorus is known to be the critical element in controlling
eutrophication, a thorough understanding of phosphorus dynamics is essential
for effective control strategies. The sediments play an important role in
phosphorus cycling, serving as either a sink or a source of phosphorus to
the overlying waters. The processes affecting the net flux of phosphorus
from the sediments are complex, depending on such factors as resuspension,
sedimentation, sorption, oxygen concentrations, and invertebrate activities.
Under oxic conditions, as found in most near-bottom waters of the Great
lakes, benthic invertebrates play a major role in phosphorus release from
the sediments (Gallepp 1979; Graneli 1979; Holden and Armstrong 1980;
Quigley and Robbins 1986). By their constant burrowing and feeding
activities, benthic invertebrates increase the rate of exchange between
nutrient-rich pore waters and overlying waters. Also, these organisms
ingest organic material and subsequently excrete remineralized nutrients in
forms readily available for further use by phytoplankton. In nearshore Lake
Michigan, excretion by benthic invertebrates was sufficient to account for
all the phosphorus released from the sediments (Gardner et al. 1981). Of
190
-------
the various invertebrate groups, unionid bivalves (mussels) in particular
can have a significant impact on nutrient cycling in a given body of water
(Lewandowski and Stanczykowska 1975; Walz 1978; Stanczykowska and Planter
1985; Kasprzak 1986; James 1987). These large filter-feeders have the
capacity to remove great amounts of organic material from the water. Thus,
they enhance nutrient mineralization either directly through excretion, or
indirectly by depositing the material on the sediment surface as faeces or
pseudofaeces and making it available to deposit-feeding forms.
The purpose of this study was (1) to quantify the rate of phosphorus
flux between the sediments and overlying waters in Lake St. Clair and (2) to
determine the rate of phosphorus excretion by the mussel population. The
significance of both sediment release and mussel excretion was subsequently
assessed by comparing these phosphorus sources to other sources of
phosphorus into in the lake. In addition to measuring excretion, the
abundance, biomass, species composition, and production of mussel
populations in the lake were also determined. Accurate estimates of mussel
biomass were, of course, essential to assessing the importance of phosphorus
excretion on a lake-wide basis.
METHODS AND MATERIALS
Intact sediment cores were collected by divers at five sites in Lake
St. Clair in May and September, 1985. The sites were chosen to be broadly
representative of different areas and sediment types (Fig. 1). The core
191
-------
tubes (4.2 cm diameter and 10 cm long) were inserted into the sediment about
5 cm, stoppered at both ends, and carefully brought to the surface. The
cores were kept upright in a cooler during transport back to the laboratory
and then placed in an incubator set at the in situ temperature. Aeration
lines were placed through the top stopper and air was slowly bubbled into
the overlying waters. This kept the water well-mixed and also kept
dissolved oxygen concentrations at near-saturation levels. All core tubes
and aeration lines were made of high-density linear polyethylene to minimize
phosphorus adsorption. Samples for phosphorus determinations were taken
every 3-4 days by drawing out 1 ml of water through a sampling port in the
top stopper. Phosphorus concentrations (SRP) were determined with an
AutoAnalyzer as described by Gardner and Malczyk (1983). Phosphorus levels
in lake-water controls were also measured on each sampling day. The volume
of overlying water was kept constant by adding 1 ml of lake water after each
sample was drawn. The incubation period lasted between 65 and 70 days.
A total of 6-8 replicates cores were collected at each station on each
sampling date. Since Lake St. Clair is shallow and bottom sediments are
easily resuspended, the impact of resuspension on sediment phosphorus
release was estimated by mixing one-half of the replicates at the beginning
of the incubation period to create a sediment slurry with the overlying
waters. The sediments were mixed again every 10 days until the end of the
incubation period. Phosphorus release rates in these mixed cores were
compared to release rates in cores that were left undisturbed.
192
-------
Mussels for phosphorus excretion determinations were collected on a
monthly basis from May to October in both 1985 and 1986. An epibenthic sled
was towed behind the vessel until enough individuals were collected. In
1985, mussels were collected from only one site (Station 72) and excretion
measured on several different species. In 1986, there were two collection
sites (Station 72 and Station 24) and rate measurements were made on only
one species, Lampsilis radiata siliquodea. The two sites had contrasting
substrate types, with sandy silt the dominant substrate at Station 72 and
silt dominant at Station 24. Excretion rates were determined on at least
four individuals from each of the two stations except in May when rates were
determined on only two individuals from each station. Individual mussels
were gently scrubbed and immediately placed in polyethlene containers having
2 liters of low-nutrient culture water (Lehman 1980). The containers were
placed in large coolers and the culture water was maintained at the in situ
temperature. The incubation period lasted 4 hours with 1-ml samples drawn
at 0, 2, and 4 h. Phosphorus concentrations were determined as in the
sediment cores. Dry weights of the mussels (soft tissue) were determined
after drying at 60 C for at least 48 h.
To determine the density, biomass, and species composition of mussels
in the lake, a population survey was conducted in September, 1986. Divers
placed a 0.5 m^ frame on the bottom and all shells within the frame area
were placed in a mesh bag. A total of 10 separate replicate samples were
collected at random at each of 28 different stations (Fig. 1). All live
mussels were immediately shucked and the soft tissue placed in preweighed
aluminum planchets. Dry weights of both the shell and soft tissue were
193
-------
obtained after drying at 60 C for at least 48 h. Individuals of the two
most abundant species, Lamosilis radiata siliquodea and Leptodea fragilaris.
were aged by counting the number of annual growth rings on the external
shell. The annual production rate of L. r. siliquodea was estimated from
the sum of the increase in weight of each of the different age groups
(Magnin and Stanczykowski 1971).
RESULTS AND DISCUSSION
Phosphorus Release From the Sediments
Sediment phosphorus release rates at each of the five stations on the
two sampling dates is given in Table 1. Rates were calculated from the net
increase in phosphorus concentrations in the overlying waters from day 10 to
day 65-70 for the May cores, and from day 3 to day 65-70 for the September
cores. In some instances, the increase in phosphorus was most rapid at the
beginning of the incubation period and then remained relatively constant
thereafter. For these cores, rates were recalculated based on the time
interval of greatest release and are included in Table 1 to provide an
estimate of maximum potential release. The mean release rates in this table
include values from all replicates at a given station since there were no
significant differences (t-test; P < .05) between release rates of mixed and
unmixed cores at any of the five stations.
194
-------
In both May and September, release rates at Stations 71 and 84 were
generally lower than release rates at the other three stations (Stations 4,
14, and 24). The former two stations were located in the northwestern
portion of the lake near the mouth of the St. Clair River where both
nutrient levels and algal productivity tend to be lower then areas more to
the southeast (Leach 1972). This portion of the lake is dominated by low
nutrient water from Lake Huron, while areas farther south are more
influenced by enriched waters from Ontario tributaries (Leach 1980). Also,
given the dominant current patterns and wind direction, very little
deposition of suspended material occurs in the northwestern portion of the
lake (Anne Clites, GLERL, per. commun.). Release rates at Stations 71 and
84 were similar in both May and September, but release rates at the other
stations were higher in May than in September. The settling and subsequent
mineralization of the spring phytoplankton bloom likely contributed to the
higher release rates at these stations in May. Overall, the highest release
rates occurred at Station 24; this station was located in the area of
greatest deposition.
The release of SRP from Lake St. Clair sediments was generally lower
than sediment release rates in other areas of the Great Lakes. The mean
release rate in this study was 19 ugP/m^/day with a mean maximum release
rate of 47 ug P/m^/day. This compares to release rates of 170-570
ugP/m^/day in nearshore Lake Michigan (Quigley and Robbins 1986) and 30-800
ugP/m^/day in Lake Ontario (Bannerman et al. 1974).
195
-------
To determine the significance of sediment phosphorus release in Lake
St. Clair, the annual net release from the sediments was compared to other
input sources, i.e. Lake Huron, tributaries, the atmosphere, and direct
point sources. The total mean load from these latter sources during the
1975-80 period was about 3,100 MT/year (Tom Fontaine, GLERL, per. commun.).
Of this amount, about 40% or 1200 MT can be considered bioavailable; that
is, available for algal uptake and not bound to particulates (Sonzogni et
al. 1982). Assuming that the mean release rate of phosphorus at the five
stations is representative of the entire lake, release from the sediments
amounts to about 8 MT/year or less than 12 of the total bioavailable
phosphorus load. Maximum sediment release amounts to only about 20 MT/year
or 22 of the total bioavailable load. Based on these calculations, the
sediments appear insignificant as a source of phosphorus in Lake St. Clair.
Phosphorus Excretion by Mussels
Mean rates of phosphorus excretion on the eleven sampling dates in 1985
and 1986 are given in Table 2. Rates from the two stations sampled in 1986
were combined since significant station differences were not apparent (t-
test; P < .05) for any of the sampling dates. Seasonal trends in phosphorus
excretion were similar for the two years. Rates were high in the spring,
declined in the summer, and then increased in the fall to reach peak values.
Reasons for this seasonal trend are not clear, but may be related to changes
in the gametogenic cycle of the organisms. An increase in ammonia excretion
in the summer (Table 2) indicates an increase in gamete production at this
time. Active protein catabolism (and hence increased ammonia excretion)
196
-------
occurs when the glycogen normally used for metabolism is used instead for
gamete production (Gabbott and Bayne 1973). While ammonia excretion would
increase, phosphorus excretion would likely decrease, since a greater
portion of assimilated material would be used for reproductive activities
and not metabolism. Seasonal changes in phosphorus excretion were
apparently unrelated to the nature of available food; the amount of
particulate phosphorus in the near-bottom water was constant throughout the
sampling period (Nalepa, unpublished). Mussel excretion rates were
generally lower than those of other Great Lakes benthic organisms (Table 3).
This may be expected, however, since the rate of phosphorus excretion per
unit weight increases as body weight decreases (Johannes 1964). The dry
weight of mussels in this study ranged from 1 to 4 g, while dry weights of
the other organisms shown in Table 3 have dry weights of less than 2 mg.
Based on estimates of biomass from the September 1986 population survey
(see below), the mussels in Lake St. Clair excrete about 59 MT of phosphorus
per year or 5% of the annual bioavailable load from other sources. In
addition to excreting phosphorus, mussels are active filter feeders and may
remove large amounts of particulate phosphorus from the lake water during
feeding. Based on preliminary estimates of mussel filtration rates
(Vanderploeg and Nalepa, unpublished) and amounts of particulate phosphorus
in Lake St Clair water, mussels are capable of filtering 220 MT of
phosphorus from the water on an annual basis; this amounts to 7% of the
total annual load.
197
-------
Mussel Population Survey
The overall mean abundance of mussels was 2/m2 (range 0-8/m2) and the
mean biomass was 4.3 g/m2 (range 0-19.4 g/m2). In general, both abundance
and biomass increased from the mouth of the St. Clair River to the head of
the Detroit River. This corresponded to previously noted trends in water
column productivity. A total of 287 individuals representing 20 different
species were collected. The three most abundant species, Lampsilis radiata
siliquodea. Leptodea fragilaris. and Proptera alata accounted for 45Z, 13%,
and 10Z of the total population. The former species was the most widely
distributed, being collected at 22 of the 28 stations. Abundances found in
this survey were lower than abundances reported from other areas in the
Great Lakes. For instance, densities of 7/m2 (Wood 1963) and 10/m2 (McCall
1979) have been reported from western Lake Erie, while Pugsley (unpublished
data) reported a mean density of 7/m2 in the southwestern portion of Lake
St. Clair. A direct comparison can be made between this survey and the
Pugsley survey since three of the sampling stations were the same and
sampling techniques were similar. Abundances in this survey were
significantly lower at two of the three stations (Table 4). It is not clear
whether these lower abundances are are a result of an actual decline in the
population or an artifact of horizontal patchiness. Considering the
stability for mussel populations over the short-term, such a decline in
abundances over just a 3-year period seems unlikely unless, of course,
environmental conditions have recently become unfavorable. Although an
unusually high number of dead mussels have been observed on Lake St. Clair
beaches over the past few years (Tom Freitag, US Army Corps of Engineers,
198
-------
per. commun.) only through long-term monitoring efforts can definite trends
in abundances be discerned.
In western Lake Erie, mussel density and diversity have apparently
declined over the past few decades (Mackie et al. 1980). Unfortunately,
historical records of mussel densities in Lake St. Glair are lacking.
However, mussel diversity and composition appear little changed since 1893.
Reighard (1894) reported finding 20 species in Lake St. Clair with Lampsilis
radiata siliquodea being very " widespread and abundant" and Proptera alata.
Liguinea nasuta. Anodonta grandis being found "frequently". The most
apparent difference between this survey and the 1893 survey of Reighard was
the relative abundance of Leptodea fragilaris: this species was reported
being "scarce" by Reighard but in this survey, it was the second most
abundant species.
The age structure of L. r. siliquodea and L. fragilaris is given in
Figure 3. For L. fragilaris. the age structure of the population was quite
similar to that found in other freshwater systems (Strayer et al. 1981;
Paterson 1985) and reflects low adult mortality and yearly variation in
recruitment. However, for L. jr. siliquodea. the average individual was
almost 10 years of age and few younger individuals were found. The reason
for this lack of recruitment is not clear, but may indicate that either the
adult population is under some sort of stress (low reproductive capacity) or
that mortality of the young is increasing. Populations of fish species
which serve as host for the glochidia of L. r. siliquodea (yellow perch,
smallmouth bass, largemouth bass, bluegill, and crappie among others) have
199
-------
remained stable over the years (Bob Haas, Michigan DNR, per.commun.).
Because the population is dominated by older individuals, the annual
turnover rate (production/biomass) of L. r. siliquodea was only 0.13; this
value is lower than found for mussels in most other freshwater lakes (Table
5).
200
-------
LITERATURE CITED
Bannerman, R. T., D. E. Armstrong, G. C. Holdren, and R. F. Harris. 1974.
Phosphorus mobility in Lake Ontario sediments (IFYGL), pp. 158-178.
Proc. 17th Conf. Great Lakes Res., Int. Assoc. Great Lake Res.
Gabbot, P. A. and B. L. Bayne. 1973. Biochemical effects of temperature
and nutritive stress on Mytilus edulis L. J. Mar.Biol. Ass. U. K.
53:269-286.
Gallepp, G. W. 1979. Chironomid influence and phosphorus release in
sediment-water microcosms. Ecology 60: 547-556.
Gardner, W. S., T. F. Nalepa, M. A. Quigley, and J. M. Malczyk. 1981.
Release of phosphorus by certain benthic invertebrates. Can. J. Fish.
Aquat. Sci. 38:978-981.
Gardner, W. S. and J. M. Malczyk. 1983. Discrete injection flow analysis
of nutrients in small-volume water samples. Anal. Chem. 55:1645-1647.
Graneli, W. 1979. The influence of Chironomus plumosus on the exchange of
dissolved substances between sediment and water. Hydrobiologia
66:149-159.
Holden, G. C. and D. E. Armstrong. 1980. Factors affecting phosphorus
release from intact sediment cores. Environ. Sci. Technol. 14:79-87.
201
-------
James, M. R. 1987. Ecology of the freshwater mussel Hyridella menziesi
(Gray) in a small oligotrophic lake. Arch. Hydrobiol. 3: 337-348.
Johannes, R. E. 1964. Phosphorus excretion and body size in marine
animals: microzooplankton and nutrient regeneration. Science
146:923-924.
Kasprzak, K. 1986. Role of Unionidae and Sphaeriidae (Mollusca, Bivalvia)
in the eutrphic lake Zbechy and its outflow. lint. Revue ges.
Hvdrobiol. 71: 315-334.
Leach J. H. 1972. Distribution of chlorophyll a and related variables in
Ontario waters of Lake St. Clair, pp.80-86. In Proc. 15th Conf. Great
Lakes Res., Int. Assoc. Great Lakes Res.
Leach, J. H. 1980. Limnological sampling intensity in Lake St. Clair in
relation to distribution of water masses. J. Great Lakes Res. 6:
141-145.
Lehman, J. H. 1980. Release and cycling of nutrients between planktonic
algae and herbivores. Limnol. Oceanogr. 25: 620-632.
Lewandowski, K. and A. Stanczykowska. 1975. The occurrence and role of
bivalves of the family Unionidae inMiklajskie Lake. Ekol. Pol. 23:
317-334.
202
-------
Mackie, G. L., D. S. White, and T. W. Zdeba. 1980. A guide to freshwater
mollusks of the Laurentian Great Lakes with special emphasis on the
genus Pisidium. EPA-600/3-80-068, Environmental Protection Agency,
Duluth, Mn. 144p.
Magnin, E. and A. Stanczykowska. 1971. Quelques donnees sur la croissance,
la biomass.et la production annuelle de trois mollusquesUnionidae de la
region de Montreal. Can. J. Zool. 49: 491-497.
McCall, P. L., M. J. S. Tevesz, and S. F. Schwelgien. 1979. Sediment
mixing by Lampsilis radiata siliquodea (Mollusca) from western Lake
Erie. J. Great Lakes Res. 5:105-111.
Nalepa, T. F., W. S. Gardner, and J. M. Malczyk. 1983. Phosphorus release
by three kinds of benthic invertebrates: effects of substrate and water
medium. Can. J. Fish. Aquat. Sci. 40:810-813.
Paterson, C. G. 1985. Biomass and production of the unionid, Elliptic
complananta (Lightfoot) in an old reservoir in New Brunswick, Canada.
Freshwat. Invertbr. Biol. 4: 201-207.
Quigley, M. A. and J. A. Robbins. 1986. Phosphorus release processes in
nearshore southern Lake Michigan. Can. J. Fish. Aquat. Sci. 43:
1201-1207.
203
-------
Reighard, J. E. 1894. A biological examination of Lake St. Clair. Bull.
Mich. Fish Comm. No.4. 61p.
Sonzogni, W. C., S. C. Chapra, D. E. Armstrong, and T. J. Logan. 1982.
Bioavailability of phosphorus inputs to lakes. J. Environ. Qual. 11:
555-563.
Stanczykowska, A. and M. Planter. 1985. Factors affecting nutrient budget
in lakes of the R. Jorka watershed (Masurian Lakeland, Poland) X. Role
of the mussel Dreissena polmorpha (Pall.) in N and P cycles in a lake
ecosystem. Ekol. Pol. 33:345-356.
Strayer, D. L., J. J. Cole, G. E. Likens, and D. C. Busco. 1981. Biomass
ans annual production of the freshwater mussel Elliptic complanata in
an oligotrophic lake. Freshwat. Biol. 11: 435-440.
Walz, V. N. 1978. Die produktion der Dreissena-population und deren
bedeutung im stroffkreislauf des Bodensees. Arch. Hydrobiol.
82:482-499.
Wood, K. G. 1963. The bottom fauna of western Lake Erie, 1951-52. Great
Lakes Res. Div., Univer. Michigan, Publ. No. 10, pp. 258-265.
204
-------
Table 1. Mean (+ SE) rates of phosphorus release from
Lake St. Clair sediments at each of the stations on
the two sampling dates in 1985. Maximum mean release
rates are given in parentheses. Rates are given as
ugP/m2/day.
Station
Sampling Date
May1
^-Water temperature = 13 C.
2water temperature = 22 C.
September^
4
14
24
71
84
31
15
31
4
8
.4 ±
• 5 ±
-2 ±
.2 ±
• 3 +
4.9
6.0
9.4
0.8
3.0
(40
(49
(38
(18
(15
.0)
.0)
.2)
.6)
.0)
11
11
22
5
8
• 2 ±
• 2 ±
.9 +
.2 ±
• 0 ±
6
1
5
1
3
.3
.6
.5
.2
.9
(32
(78
(58
(16
(32
.3)
.4)
.5)
.3)
-3)
205
-------
Table 2. Mean (± SE) phosphorus and ammonium excretion
rates of mussels in Lake St. Clair in 1985 and 1986.
Rates given in ug/gDW/h.
Sampling
Date
1985
May 9
May 14
Jul 16
Sep 3
Sep 19
1986
Apr 30
May 19
Jul 10
Aug 4
Sep 16
Oct 15
n
5
7
6
7
7
9
4
9
10
10
10
Excretion
Phosphorus
0.9 ± 0.1
0.9 + 0.3
0.7 + 0.5
3.9 ± 0.9
2.0 ± 0.4
1.1 + 0.3
0.6 ± 0.1
0.5 ± 0.1
1.5 + 0.3
1.9 ± 0.4
1.9 ± 0.4
Rate
Ammonium
24.6 ± 3.7
12.8 ± 2.1
49.1 + 3.2
25.3 ± 2.4
27.6 ± 4.5
23.3 ± 2.3
20.9 + 4.0
50.1 ± 3.7
52.3 ± 8.5
40.8 ± 6.2
21.3 ± 2.2
N:P
Ratio
27
14
69
7
14
20
35
109
35
21
12
206
-------
Table 3. Mean (± SE) phosphorus and ammonium excretion rates
(nmol/gDW/h) for some common benthic invertebrates occurring in
the Great Lakes. Data compiled from Nalepa et al. (1983), Gardner
et al. (1983), and Gauvin (unpublished).
Benthic Excretion Rate N:P
Organism Phosphorus Ammonium Ratio
Chironomidae 690 11,300 16
Oligochaeta 150 8,100 54
Pontoporeia 90 1,090 12
Unionidae 50 1,020 23
207
-------
Table 4. Comparison of mean mussel abundances (number per
square meter) at three stations in Lake St. Clair in 1983
(Pugsley, unpublished) and in 1986 (this study). Standard error
in parenthesis. * - Densities significantly different at the
0.05 level (t-test).
1983 1986
Station (Pugsley 1986) (This study)
3 13.8 (1.7) 7.8 (1.5)*
21 9.8 (2.1) 2.2 (0.9)*
66 2.4 (0.6) 2.0 (0.7)
208
-------
Table 5. Turnover ratio (production/biomass) of unionids from various
lentic environments.
Water Body
P/B
Reference
Lake Zbechy
Lake Mikolajskie
Lac des Deux
Montagnes
Morice Lake
Lake St. Clair
Mirror Lake
Lac Saint Louis
0.45
0.35
0.20
0.19
0.13
0.12
0.10
Kasprzak (1986)
Lewandowski and
Stanczykowska (1975)
Hagnin and
Stanczykowska (1971)
Paterson (1985)
This Study
Strayer et al. (1981)
Magnin and
Stanczykowska (1971)
209
-------
LIST OF FIGURES
Fig. 1. Sampling stations in Lake St. Clair. Sediment phosphorus
release rates were determined at Stations 4, 14, 71, 84, and
24. Mussel populations were sampled at all the stations
except the first four stations given above.
Fig. 2. Age structure of the two most abundant species, Lampsilis radiata
siliquodea and Leptodea fragilaris.
210
-------
10
m
Figure 1,
211
-------
25r-
£ 20
o
O
O
10
L r. siliquodea
0 2 46 8 10 12 14 16
Age (years)
25
20
O
JD
O 15
O
I 10
L fragilaris
o
•-•••••••—
4 6 8 10 12 14 16
Age (years)
Figure 2.
212
-------
SEDIMENT TRANSPORT IN LAKE ST. CLAIR
Nathan Hawley and Barry Lesht^
ABSTRACT
In order to study resuspension in Lake St. Glair, bottom-resting
instrumented tripods were deployed at various sites in the lake during 1985
and 1986. The tripods recorded time series measurements of current
velocity, temperature, and water transparency. The measurements were then
used to calculate the parameters of a simple flux model in order to
determine the criterion for sediment resuspension. Since most resuspension
in the lake is due to wave action, wave orbital velocity was used as the
forcing function. These velocities were calculated by running the GLERL
wave model using on-lake wind measurements to determine the wave climate,
and then calculating the orbital velocities using intermediate-water wave
theory. The pattern of the calculated velocities shows good agreement with
both the measured standard deviation of the current velocity (although the
magnitudes are somewhat different) and the total suspended material.
Critical values of wave orbital velocity (the threshold for resuspension)
range from 0.1 to 0.9 cm/s. The model shows good predictive capability and
is relatively insensitive to changes in the settling velocity and
resuspension coefficient. Some of the variability in the critical velocity
may be due to changes in substrate characteristics.
^Argonne National Laboratory
213
-------
INTRODUCTION
As part of the Upper Great Lakes Connecting Channels study, we
undertook to develop an empirical relation between flow activity and
sediment resuspension in Lake St. Clair. To do this we deployed bottom-
resting instrument packages in the lake at various times and sites which
recorded time series measurements of water transparency, flow velocity, and
water temperature. The measurements were then modeled using a relation
suggested by Simons and Schertzer (1986) to develop a criterion for sediment
resuspension. Measurements of total suspended material (TSM) and vertical
profiles of water transparency and water temperature were also made. A
meteorological tower measured on-lake weather conditions from July to
September, 1986. Plans to use a bottom-resting flume to measure erosion
thresholds had to be abandoned because a suitable vessel was not available
for its deployment. Several other programs, in particular the wave study by
the Great Lakes Environmental Research Laboratory (GLERL) and the Canadian
Centre for Inland Waters (CCIW) and the sediment transport study by CCIW
have provided valuable supporting data.
Lake St. Clair is a large (approximately 40 km wide), shallow (maximum
depth 7m) lake located between Lakes Huron and Erie. As such it receives
the entire outflow from the upper Great Lakes - approximately 5300 m^/s.
Since the residence time of the water in the lake is only about 7 days, much
of the sediment carried into the lake is almost immediately swept out again.
However much of the lake is covered by sand and silt which can be
resuspended. Because of its shallow depth and large fetch, we felt that
214
-------
wave action might be a prime cause of resuspension in Lake St. Clair. Thus
most of our current velocity measurements were in burst mode. This allowed
us to measure not only the mean current, but also its standard deviation,
which we felt could serve as a measure of wave action.
DATA COLLECTION
One or more tripods were deployed five times during 1985 and 1986 (Fig.
1). One of these tripods, the one deployed by Barry Lesht of Argonne
National Laboratory (ANL), was equipped with a Marsh-McBirney current meter
and 25 cm pathlength Seatech transparency meter, as well as with a
temperature probe. The other two tripods were deployed by Nathan Hawley
(GLERL) and were equipped with only a temperature probe and 25 cm Seatech
transparency meter. Although we had planned to have current meters on these
two tripods as well, the meters we bought never worked properly. Details of
the tripod locations and deployment periods are given in Table 1. In all
cases the GLERL transparency meters were 0.9m above the bottom. Both they
and the temperature sensor, which was located 1.2m above the bottom, took a
60 second sample at 1 Hz every 15 minutes. The mean and standard deviation
of the transparency (TSD) were recorded along with the average temperature.
The transparency meter on the ANL tripod was 0.9m above the bottom, the
current meter 0.7m and the temperature sensor 1.0m. During all but one of
the deployments the current meter recorded a 75 second burst at 3.4 Hz.
During July-August of 1985 and May-June, 1986 these measurements were taken
every 45 minutes, during July-August, 1986 every 60 minutes, and during
215
-------
October, 1986 every 30 minutes. For all but the last of these deployments
only an average transparency and temperature were recorded. During the
October, 1986 deployment transparency was recorded in the same manner as the
current velocity. During the September-October, 1985 deployment continuous
5 minute averages of current velocity, water transparency, and temperature
were made.
In order to calibrate the transparency meters, measurements of
transparency and TSM were made in triplicate at 25 different stations (Fig.
2). For 110 measurements, TSM (measured in mg/1) is related to water
transparency (measured as the fraction of the transmittance in air) by
TSM - -8.33* Ln(Tr)-1.96 r2=0.92 (1)
The measurements used to establish this relationship show no geographic or
temporal trends. Use of (1) allowed TSM to be calculated from the
transparency measurements. Figure 3 shows the predicted values and the
measurements used in the regression. The two curved lines represent the 95%
confidence interval for the predicted values. It should be noted however,
that all of the calibration measurements were made during fairly calm
conditions. During resuspension events far more sand-sized material is
likely to be in the water column. Because of its high density, a given
weight of sand will attenuate the light beam far less than the same weight
of more porous, far less dense, floes which form from cohesive material.
Thus, the TSM calculations must be underestimates of the true loading. In
216
-------
the model we used, however, concentration is essentially a surrogate for
transparency. Since the model parameters are all internally estimated, they
will still be consistent. However, the actual rates of settling and
resuspension measured on a mass basis will probably be somewhat different
from those calculated.
MODEL DESCRIPTION
Simons and Schertzer (1986) have proposed a simple model which relates
changes in the vertically integrated (assuming vertical uniformity)
suspended sediment concentration (C) to changes in the balance between the
rates of sediment resuspension and sediment settling. This is just a flux
model in which the net flux (the left-hand side of the equation) is equal to
the difference between the upward and downward fluxes. The resuspension rate
is assumed to be a linear function of the forcing function (F) above some
critical value (Fc) and zero otherwise, with R being the resuspension
concentration. The settling rate is the product of the settling velocity
(S) and the ambient concentration (Ca) which is assumed to always be
present. Thus the model is
D*dC = - S(C-Ca) for F < Fc (2)
dt
and
D*dC = R(F-FC) - S(C-Ca) for F > Fc (3)
dt
217
-------
Using the results from our deployments we could first solve equation 2 for S
and Ca for low values of dC, and then use these values in equation 3 to
solve for R and Fc. However, we have used the time integrated form of
equation (3) to estimate the sets of model parameters that best reproduce
the observed time series of sediment concentration. The model results were
then evaluated using Wilmot's (1984) criteria. Obviously, since the model
does not take advection into account, sediment concentration changes due to
advection must at least be identified, and if possible removed from the
record, prior to determining the parameters. This requires a set of
criteria to differentiate between advection and local resuspension events.
One obvious criteria for local resuspension is that the increase in
sediment concentration should occur at the same time as an increase in the
forcing function. Thus, if resuspension is due to wave activity, increases
in sediment concentration should occur when wave action is greatest (during
storms). We have used high values of Speed Standard Deviation (SSD) as an
indication of wave activity. Since waves are wind-generated, in the absence
of direct or indirect wave measurements wind records might be useful in
predicting resuspension activity. There should also be (in the absence of
advection) a characteristic decay time of suspended sediment concentrations
which depends on the settling speed of the sediment and the height to which
the sediment was resuspended. All of the observations in Lake St. Clair
indicate that it is vertically well mixed, particularly during resuspension
events, so differences in the decay time should depend only on the settling
rates. These variations are expected to be small. Chriss and Pak (1976)
have also suggested that the standard deviation of water transparency should
218
-------
be higher during resuspension events than during advection because the water
has not had as much time to become well mixed
Ironically, the data sets that most unambiguously show local
resuspension were collected last. Figure 4 shows the sediment concentration
record measured at station 42 during October, 1986. The record extends only
to day 301 because the other two stations were retrieved at that time. Thus
the data collected on days 303-309 are not discussed here. The gaps in the
record are due to instrument problems; the recording tape was filled on day
296 and not replaced until day 301. The very high concentrations (off the
scale) occurred when the transparency was zero. For zero transparency,
equation 1 gives a TSM of of infinity. We have used a value of 68 mg/1 -
equivalent to the lowest transparency we could measure. The TSM record is
very well correlated with both the mean speed and SSD (Figs. 5 and 6). In
fact the plots of the mean speed and its standard deviation are almost
identical. All three plots show major changes beginning on day 287 and
extending until the measurements were interrupted on day 289. The high TSM
levels, however, continue for several more days - until day 293 - even
though both the mean speed and the SSD are at background levels. Most
probably the high TSM values during this time are due to advection of
material resuspended elsewhere.
Figure 7 shows wave orbital velocities calculated for the site. These
velocities were calculated by running the GLERL wave model using wind data
collected by CCIW near the station. The wave model results were then used
as input to calculate the maximum orbital velocity one meter above the
219
-------
bottom. The pattern of the results is very similar to that of the mean
measured speed and the SSD, although there is a difference in magnitude of
about two. The calculated velocities also decay more quickly than do the
measured values. This is probably because the measured parameters also
include the effects of currents generated by the storm. In spite of these
differences, the generally good agreement suggests that at least most of the
TSM record can be accounted for by local wave-induced resuspension. The
temperature record (Fig. 8) shows a marked decrease during and after the
storm, another indication that advection was present. The standard deviation
of the transparency (TSD) in Figure 9 shows a peak during the storm, but
also several other peaks when no wave action was present. Apparently, this
measurement is not an unequivocal indicator of local resuspension, at least
in Lake St. Clair. The very large peak on day 293 is particularly
interesting because it does not appear to be related to any measure of flow,
either observed or calculated.
Figure 10 shows the results from the model when the calculated orbital
velocity data is used as the forcing function. The fit is remarkably good,
as evidenced by the high values of Wilmot's (1984) index of agreement (d)
and the unsystematic mean-squared error (MSEU, Table 2). Here d is a
relative index of agreement between the model and the observations based on
the summed square error. MSEU is that portion of the root mean-squared
error that is unsystematic. Note that the model predicts a higher
concentration than the maximum measured, but recall that the transparency
meter was saturated during this period. Also note that the elevated TSM
measurements on days 289-301 are not predicted by the model, as would be
220
-------
true if they are due to advection. Although the main peak (on days 287-299)
is predicted quite well, the peak on day 308 is substantially less than the
predicted concentration. This suggests that either R is too high or that
the resuspension flux is not a linear function of Fc but possibly a power
function of some sort. The results agree quite well with those obtained
using SSD as the forcing function (Fig. 11), so we have used the wave model
results at the other two stations as the forcing function since we have no
current measurements at those sites.
Figure 12 shows the TSM record at station 71 for the same period. The
large increase on day 287 is also present here, as is the long decay. In
addition, however, there is a pronounced peak on day 283. The calculated
orbital velocities (Fig. 13) also show a marked peak on day 283, in contrast
to the record at station 42, where the peaks in both sediment concentration
and wave orbital velocity are much smaller. Again, as at station 42, cooler
water is present during the storm and its aftermath (Fig. 14). The TSD
(Fig. 15) shows two peaks - one during the storm and the second just prior
to the dramatic lowering of the concentration on day 291. Since this second
peak is not associated with wave action, it may indicate mixing of clearer
water from upstream with the more turbid water in the lake. Comparison of
the model results with the observations are shown in Figure 16. Again the
model slightly overpredicts the peak concentration, although in this case
the transparency meter was not saturated. The model also underpredicts the
peak concentration during the storm on day 283, and predicts several small
peaks that were not observed. Overall, however, the model does a good job
of predicting the actual measurements.
221
-------
Results of the TSM measurements made at station 1 are shown in Figure
17. The biggest difference between this station and the other two is the
two peaks on day 285. It is very reassuring to find that a peak in
calculated wave orbital velocities also occurs on that day (Fig. 18), in
addition to the peaks on days 283 and 287. The effects of advection after
the storm on day 287 are not as evident at this station, and there is no
peak in the TSD on day 293 (Fig. 19). There are however several peaks later
in the record which are not correlated with wave activity. These may be due
to inhomogeneities in the water entering the lake. The temperature record
(Fig. 20) is similar to those at the other stations. The observed and
modeled results are shown in Figure 21. The model accurately predicts all
three peaks, although it either overestimates or underestimates their
magnitude, and the decay curves are reasonably close to the observed ones.
Model results for the stations are tabulated in Table 2, along with the
percent mud (less than 60 microns) of the bottom sediments as measured by
the University of Windsor (1985). The stations for this deployment were, in
fact, chosen so that the sand percentages were approximately equal since one
of the goals was to investigate the effect of differing wave climates on
similar substrates. Although the model results from stations 1 and 71 are
in good agreement, the results from station 42 are somewhat different - all
the parameters but S are much higher. Some of this discrepancy may be an
artifact of the data however. The higher ambient concentrations are
probably due mostly to the outflow of the Thames River, which has a high
suspended load. If Ca is artificially high, then bottom resuspension may be
occurring at lower values of Fc than is evident from the data. An
222
-------
artificially high value of Fc could in turn lead to a higher value of R, if
in fact the resuspension rate is not linear as assumed here but is a power
function of F.
The spring, 1986 deployment was designed to examine the effects of
changes in substrate on resuspension. During this period the tripods were
all located within 8 km of each other in the northwest portion of the lake.
We hoped that putting the tripods close together would minimize the
differences in wave climate between stations. Since we did not yet have our
meterological tower available, we put the tripods near the St. Glair Shores
Coast Guard station, which makes weather observations. However, a
comparison of the weather records from St. Glair shores with on-lake
observations later in the year showed substantial differences between the
two. Until a valid transfer function can be developed, wave climates
estimated from the St. Glair Shores weather data are not accurate. This
means that we could not use calculated wave orbital velocities as the
forcing function in the model for this deployment. Since only the current
meter at station 3 worked, only that data set could be analyzed, since we
felt that, given the variability in the TSM measurements between the
stations, it would not be justifiable to apply the current data from one
station to another site. Once a suitable transfer function is developed, we
will be able to assess the effects of substrate variability, but this cannot
be done yet.
The TSM record for station 3 (Fig. 22) shows several pronounced peaks,
one of which lasted for several days. Unfortunately the mean speed record
223
-------
(Fig. 23) is so noisy that it is hard to see any obvious correlation. The
record of the SSD (Fig.24) shows a correlation with some, but not all, of
the TSM peaks and also shows peaks where no TSM increase occurred. We
believe the three TSM peaks after day 152 and the one on day 136 are due to
wave resuspension since they correlate with peaks in the SSD and decay
quickly. The long episode starting on day 142 is probably due to advection
since both the speed and SSD are low when it begins. The peak on day 137 is
hard to explain because it is not correlated with a peak in SSD but decays
quickly. The temperature record (Fig. 25) does not help much in
interpreting the results. The model results, using SSD as the forcing
function, show a reasonable agreement with the data (Fig. 26) though not as
good as for the fall. The most obvious failure is on day 149 (May 29) when
the model predicts a non-existent peak in TSM. This is because of a peak in
SSD that day. The model parameters (Table 2) are fairly consistent with the
results from the fall deployment. Ca is lower, reflecting the relative
absence of sediment in the water, and although Fc is higher, it is a
different parameter than in the fall. The values of R and S are close to
those for the other western stations.
The results from station 5 (Fig. 27-29) are very much like those from
station 3. The same TSM peaks are seen, although their form is slightly
different, particularly for the (assumed) advection event. The TSD record
correlates well with some TSM peaks, particularly those late in the record,
but overall TSD does not seem to be a reliable indication of local
resuspension. The temperature record show the same general pattern as at
station 3.
224
-------
The tripod at station 1 worked for only part of the deployment. The
TSD peaks (Fig. 31) correlate extremely well with the early peaks in TSM
(Fig. 30) but not with the TSM peak on day 142. This latter peak, although
it looks like the one in the middle of the records at stations 3 and 5,
actually begins about a day earlier. The temperature record (Fig. 32) shows
a minimum just before the peak in TSM further supporting the hypothesis that
it is due to advection.
Stations 1, 5, and 71 were occupied during the summer of 1986.
Stations 1 and 5 had been occupied earlier in the year, and station 71 was
near both our meteorological tower and a wave rider deployed by NESDIS of
Canada. However, several problems complicate the interpretation of the data
from this deployment. First, algal growth during the deployment period
fouled the transparency meters. We endeavored to calibrate the fouling by
taking vertical profiles with a clean meter once a week at each station, but
unfortunately two of the tripods stopped during the deployment so we have
only one complete calibration curve, and the results vary from station to
station. The TSM data shown has been corrected to the best of our ability,
but is not perfect. In addition, we also found that although there were
very few significant wave events, movement of turbid bottom water occurred
fairly persistently, particularly at station 71, which is where our current
measurements were made. It is thus extremely difficult to say with any
great confidence when resuspension occurred during this deployment.
The TSM record from station 71 is much noisier than that during the
fall deployment (Fig. 33) and does not correlate very well with either the
225
-------
mean speed (Fig. 34) or the SSD (Fig. 35). Nor does it seem to correlate
very well with the calculated orbital velocities (Fig. 36). In fact the
best correlation seems to be with temperature (Fig. 37) which shows
extraordinary variability. In this case, sharp drops in TSM occurred
simultaneously with abrupt increases in temperature. These temperature
increases are in turn correlated with periods of high wave orbital
velocities. Thus, rather than causing resuspension, wave action appears to
be associated with minimums in TSM. The explanation appears to be that
during the summer there is a thin bottom layer of more cooler, more turbid
water underlying the warmer, clearer water. Wave action mixes the two and
brings the warmer water down nearer the bottom. Vertical profiles taken
during this period frequently show this cooler, turbid layer (Fig. 38). In
addition, the water temperature sensor on our meteorological station, which
was 3m below the surface, recorded temperatures between 22 and 24 degrees
during the deployment. It seems likely then that most of the TSM signal is
not due to resuspension but to vertical movement of the upper surface of
this bottom turbid layer past the sensor. Not surprisingly, the model does
not do very well with this data set.
The results from stations 5 and 1 are somewhat less noisy. At station
5 the TSM (Fig. 39) shows 3 pronounced peaks, and a noticeable minimum
beginning on day 201. This minimum correlates with a rise in temperature
(Fig. 40) so it may also be due to a thinning of a bottom turbid layer. The
first two peaks are associated with peaks in orbital velocity (Fig. 41), but
the last is not: it may be due to advection. Again, the TSD record (Fig.
42) is not much help in distinguishing resuspension events.
226
-------
The TSM record at station 1 (Fig. 43) shows several pronounced peaks
which decay very quickly. These peaks are well correlated the the
calculated wave orbital velocities (Fig. 44) and also have high TSD values
(Fig. 45). There does not appear to be a consistent correlation with
temperature (Fig. 46), so it seems likely that these events are in fact due
to wave resuspension. We have run the model for this station using the
parameters from the fall deployment with the wave orbital velocities as the
forcing function. The results, shown in Figure 47) are surprisingly good.
The model accurately predicts the occurrence, if not the actual
concentrations, for several of the TSM peaks. The most noticeable defect is
the overly long decay times, which indicates that S should be increased.
The relatively poor values for the index of agreement and MSEU are not too
surprising when one considers that most of the summer data is in the range
of the ambient concentration for the fall deployment.
Both of the 1985 deployments were exploratory, and neither has been
analyzed in any detail. During the summer deployment, one of the axes of
the current meter failed, so we have no record of either mean speed or SSD.
The transmittance and temperature records are shown in Figures 48 and 49.
During the fall deployment, the tripod was placed near one of the wave
stations deployed by CCIW. The current meter and transparency meter were
both set up to log continuous 5 minute averages. These records and the
temperature are shown in Figures 50-52. Although preliminary examination of
both data sets shows a correlation between TSM and wave orbital velocities,
the model results for these deployments are not yet available.
227
-------
DISCUSSION
It appears that most, if not all, resuspension in Lake St. Glair is due
to wave action. Although it is difficult in some of the records to
distinguish between resuspension and advection, those events which can be
unambiguously identified are almost always associated with wave activity.
The good fits obtained from a very simple model which totally ignores both
advection and resuspension due to currents also indicates that wave action
is the primary cause of resuspension. The wave orbital velocities
calculated from the results of the GLERL wave model serve very well as the
forcing function in the model. These velocities are the maximum values
calculated using intermediate-water wave theory for a height one meter above
the bottom in a total depth of 6.5m. Since the wave model gives significant
wave height and period as the output, the orbital velocities are not the
absolute maximum velocities, but the maximums for the significant waves,
which are somewhat smaller than the peak waves. It is thus not surprising
that the calculated orbital velocities are somewhat smaller than the actual
measured velocities. The good agreement between the patterns of calculated
orbital velocities and the measured standard deviations of the speed
indicate that in general the latter is a good analog for the former.
However, there are times when high values of SSD are not correlated with
high orbital velocities, so the analogy is not exact.
We had hoped that the standard deviation of the transparency (TSD)
would be a good indicator of local resuspension, but we found frequent
instances of high TSD values which did not correlate with resuspension
228
-------
events. It appears that in Lake St. Clair the time and length scales for
resuspension and advection are too similar for TSD values to serve as a
useful distinguishing criterion. This means that resuspension was identified
primarily by the simultaneous occurrence of a rise in TSM and in orbital
velocity.
The various model results show the most consistency for the values of S
and R. However the predicted concentrations are relatively insensitive to
the values of these parameters. Examination of the predictions indicate
that a higher value of S may improve the fit in several cases by shortening
the time required to return to ambient conditions. This in turn would
require an increase in R in order to keep the peak concentrations the same.
Another solution would to be to use a more complicated model, possibly one
in which the resuspension rate is a power function of F, as proposed by
Lavelle et al (1984). However, given the limitations of our measurements
and the good agreement between the model and measured concentrations, a more
complicated model may not provide much more insight.
When wave orbital velocities are used as the forcing function in the
model, the critical value above which resuspension occurs is less than 1
cm/s (except at station 42 where we believe that the very high ambient
concentrations mask the actual initiation of resuspension). Although an
extrapolation of F to the bottom is fraught with peril (recall that it is a
calculated number - not measured), it appears that sand-sized material is
unlikely to be resuspended at these low values. Since sand is resuspended
during at least some of the resuspension events (as evidenced by the sand
229
-------
found in the CCIW traps), there is probably a second, higher value of F,
which applies to the coarser material. The value determined from the model
is for the finer material which causes the decrease in transparency. The
variability in Fc between stations may be due to differences in substrate
characteristics, but the only such measures available (X sand, gravel, and
mud) are not sufficient to explain the differences. Although there is
considerable variation in Fc between stations, the model parameters appear
to be fairly constant through time. The good agreement between the model
results for the summer deployment at station one, which were obtained using
the parameter values calculated from the fall deployment, and the
observations, shows that the model has good predictive capability and lends
credence to the other results. Further tests of the model using the other
summer and spring data will be attempted in the future.
CONCLUSIONS
Our results show that resuspension in Lake St. Glair is due mainly to
wave action. When wave orbital velocities are used as the forcing function
in a simple model, the agreement between the predicted and observed
instances of sediment resuspension is quite good. Critical values of the
orbital velocity are less than one cm/s when calculated at one meter above
the bottom. These velocities can be calculated from the results of the
GLERL wave model if on-lake wind records are available. Variations in the
critical velocity between sites may be due to differences in substrate
characteristics, but there is no adequate data to test this.
230
-------
LITERATURE CITED
Chriss, T.M. and H.J. Pak, 1978, Optical evidence for sediment
resuspension-Oregon continental shelf, EOS, 59, p 410.
Great Lakes Institute, University of Windsor, 1985, A case study of
selected contaminants in the Essex Region, Vol 1: Physical Sciences,
contract report for DSS contract UP-175.
Lavelle, J.W., Mofield, H.O., and E.T. Baker, 1984, An in situ erosion
rate for a fine-grained marine sediment, Jl. Geophys.
Res.,89,6543-6553.
Simons, T.J. and W.M. Schertzer, 1986, Modeling wave-induced sediment
resuspension in Lake St. Glair, unpublished MS, NWRI.
Wilmot, C. J.,1984, some comments on the evaluation of model performance,
Bull. Am. Meteor. Soc., 63, 1309-1313.
231
-------
Dates
7/11/85 - 8/8/85
9/10/85 - 10/9/85
5/15/86 - 5/24/86
5/15/86 - 6/6/86
5/15/86 - 6/6/86
7/08/86 - 7/27/86
7/08/86 - 7/31/86
7/08/86 - 8/8/86
10/10/86 - 10/28/86
10/10/86 - 11/11/86
10/10/86 - 10/28/86
TABLE
TRIPOD DEPLOYMENTS
Station #
42
71
1
3
5
1
5
71
1
42
71
Location
42°23'45nN
82°42'03"W
42°24'55"N
82°41'45"U
42*31' 18-N
82°44'48"W
42"29'42"N
82°47'42"W
42°28'06"N
82°47'24"W
42°31'18"N
82°44'48"W
42°28'00"N
82°47'18"W
42'25'00-N
82"40'48"W
42'31'11-N
82°44'49"W
42°23'08BN
82-32'28-W
42°24'58'tN
82°40'38"W
Tripod
ANL
ANL
GLERL
ANL
GLERL
GLERL
GLERL
ANL
GLERL
ANL
GLERL
All moorings were in 6-7m of water.
232
-------
Table 2
Model Results
Station
% Mud
(mg/1) (cm/s) (cm/s) (mg/1)
MSEU
42-fall
71-fall
1-fall
3-spring
5-spring
1-spring
1-summer
71-summer
5-summer
25
35
30
39
53
30
30
35
53
6.4
3.8
4.3
1.3
4.3
2.8
0.9
0.1
2.9
0.1
0.0055
0.0063
0.0100
0.0033
0.100
0.17
0.05
0.05
0.02
0.05
.947
,969
.922
.870
.616
.999
1.00
.995
.980
.757
233
-------
LIST OF FIGURES
Figure 1. Deployment locations.
Figure 2. Transparency calibration locations.
Figure 3. Transparency-TSM calibration curve.
Figure 4. TSM plot, station 42, October, 1986
Figure 5. Speed plot, station 42, October, 1986
Figure 6. Standard deviation of the speed, station 42, October, 1986
Figure 7. Calculated wave orbital velocity, station 42, October 1986
Figure 8. Temperature, station 42, October, 1986
Figure 9. Standard deviation of the transparency, station 42, October, 1986
Figure 10. Comparison of the observed and calculated TSM using wave orbital
velocity as the forcing function, station 42, October, 1986
Figure 11. Comparison of the observed and calculated TSM using the standard
deviation of the speed as the forcing function, station 42,
October, 1986
Figure 12. TSM, station 71, October, 1986
Figure 13. Calculated wave orbital velocity, station 71, October, 1986
Figure 14. Temperature, station 71, October, 1986
Figure 15. Standard deviation of the transparency, station 71, October,
1986
Figure 16. Observed and calculated TSM using the wave orbital velocity as
the forcing function, station 71, October, 1986
Figure 17. TSM, station 1, October, 1986
Figure 18. Calculated wave orbital velocity, station 1, October, 1986
Figure 19. Standard deviation of the transparency, station 1, October, 1986
Figure 20. Temperature, station 1, October, 1986
Figure 21. Observed and calculated TSM using wave orbital velocity as the
forcing function, station 1, October, 1986
234
-------
Figure 22. TSM, station 3, May-June, 1986
Figure 23. Speed, station 3, May-June, 1986
Figure 24. Standard deviation of speed, station 3, May-June, 1986
Figure 25. Temperature, station 3, May-June, 1986
Figure 26. Observed and calculated TSM using the standard deviation of the
speed as the forcing function, station 3, May-June, 1986
Figure 27. TSM, station 5, May-June, 1986
Figure 28. Standard deviation of the transparency, station 5 May-June
1986
Figure 29. Temperature, station 5, May-June, 1986
Figure 30. TSM, station 1, May-June, 1986
Figure 31. Standard deviation of the transparency, station 1, May-June
1986
Figure 32. Temperature, station 1, May-June, 1986
Figure 33. TSM, station 71, July-August, 1986
Figure 34. Speed, station 71, July-August, 1986
Figure 35. Standard deviation of the speed, station 71, July-August, 1986
Figure 36. Calculated wave orbital velocity, station 71, July-August, 1986
Figure 37. Temperature, station 71, July-August, 1986
Figure 38. TSM profile, station 71, July 22, 1986
Figure 39. TSM station 5, July-August, 1986
Figure 40. Temperature, station 5, July-August, 1986
Figure 41. Calculated wave orbital velocity, station 5, July-August, 1986
Figure 42. Standard deviation of the transparency, station 5 July-August
1986
Figure 43. TSM, station 1, July-August, 1986
Figure 44. Calculated wave orbital velocity, station 1, July-August, 1986
Figure 45. Standard deviation of the transparency, station 1 July-August
1986
235
-------
Figure 46. Temperature, station 1, July-August, 1986
Figure 47. Observed and calculated TSM using the wave orbital velocity as
the forcing function, station 1, July-August, 1985
Figure 48. Transparency.station 42, July-August, 1985
Figure 49. Temperature, station 42, July-August, 1985
Figure 50. Transparency, station 71, September, 1985
Figure 51. Speed, station 71, September, 1985
Figure 52. Temperature, station 71, September 1985
236
-------
237
-------
138
-------
7)
£
I/O
TSM CRLIBRRTION CURVE
I)
•i'
til
li-
ft
r
0.0
5.0
10. 0
15.0
20.0
TSM (MG/L)
25.0
30.0
35.0
•10.0
-------
STA 42, OCT 10-NOV 8, 1986
o
Si
3H
q
81
q
n-
ho
*-
o
q
iri-
q
£H
q
a
q
id
283.0 285.0 287.0 289.0 291.0 293.0
JULIAN DAY 1986
295.0
297.0
299.0
301.0
-------
STA 42, OCT 10-NOV 8, 1986
7)
cs'
283.0
205.0
287.0
269.0
291.0 293.0
JULIAN DAY 1986
295.0
297.0
299.0
301.0
-------
STA 42, OCT 10-NOV 0, 1906
283.0 285.0 207.0 280.0 201.0 203.0
JUUAN DAY 1986
305.0
207.0
209.0
301.0
-------
STATION 42 10/1-11/15/86 STARTS AT 0000 ESTT
o
M-
o
-J-
o
b-
NJ
4^
LO
1SJ
3-
q
6-
263.0 285.0 287.0 289.0 291.0 293.0
JULIAN DAY 1986
295.0
297.0
299.0
301.0
-------
STA 42, OCT 10-NOV 0,1906
q
iri-
q
ri-
q
si-
o
6-
q
o>-
a,
3-
SH
o
evfH
§.
—I
285.0
—I
287.0
—r
289.0
1
297.0
283.0
201.0 293.0
JULIAN DAY 1986
295.0
299.0
—l
301.0
-------
STA 42, OCT 10-NOV 8, 1986
8-
S-i
o
Q o-
CO 6
283.0 885.0 207.0 269.0 291.0 293.0
JULIAN DAY 1986
295.0
297.0
299.0
301.0
-0
-------
80-i
CK
LJ
Ld
0,
CO
13
CO
O
60
O 40
UJ
O
20-
0
STATION H2 - OBSERVED AND MODELED TSM
FLOW PARAMETER - MODELED WAVE ORBITAL SPEED
MODEL PARAMETERS
CA=6.4 FPC=2.8 S=0.0055 R=0.17
OBSFRVFD
PREDICTED
283 287 291 295 299 303
JULIAN DAY - 1986
307
311
o
-------
NJ
80-i
U)
£
60-
Ld
!<
O
UJ
O
40-
LJ
Q_
(f)
ID
20-
o
0
283 287
STATION Ml - OBSERVED AND MODELED TSM
FLOW PARAMETER - SPEED STANDARD DEVIATION
MODEL PARAMETERS
CA=8.0 FPC=6.5 S=0.006 R=0.225
I—i—i—i—|—i—i—i—|—i—i—i—|—r
291 295 299 303
JULIAN DAY - 1986
307 311
-------
STATION S710CT
q
s
K3
4>
CO
q
a-
q
8
q
to
O
a
o
o
e
cd
O
ol
o
6
383.0 285.0 287.0 289.0 201.0 293.0 295.0
JULIAN DAY 1986
297.0
299.0
301.0
-------
STATION 7110/1-11/15/86 STARTS AT 0000
283.0
286.0
287.0
289.0
V?
201.0 293.0
JULIAN DAY 1988
296.0
297.0
299.0
301.0
-------
STATION S710CT
1-
fO
^n
O
283.0 285.0 287.0 269.0 291.0 293.0 295.0
JUUAN DAY 1986
297.0
299.0
301.0
-------
STATION S710CT
ho
Ui
8
6
283.0 285.0 287.0 289.0 291.0 293.0
JULIAN DAY 1986
295.0
297.0
299.0
301.0
-------
NO
l_n
KJ
O)
O
LJ
Q
CL
V)
Q
25-1
20-
15-
10-
283
STATION 71 - OBSERVED AND MODELED TSM
FLOW PARAMETER - MODELED WAVE ORBITAL SPEED
MODEL PARAMETERS
CA=3.8 FPC=0.9 S=0.0063 R=0.050
1 - 1
287
,
295
291
JULIAN DAY - 1986
299
OBSERVED
PREDICTED
-•—I
303
-------
STATION Sl-OCT
ISJ
o
M-
p
O
283.0 285.0 287.0 289.0 291.0 293.0 2950
JULIAN DAY 1986
297.0
299.0
301.0
-------
STATION 110/1-11/15/86 STARTS AT 0000
NJ
(J\
ff
S-,
o
ffs.
283.0 2B5.0 287.0
280.0 291.0 203.0
JULIAN DAY 1986
205.0
297.0
200.0
301.0
-------
STATION Sl-OCT
Ol
3
d'
2
6
283.0
—I
285.0
287.0 289.0 291.0 293.0
JULIAN DAY 1986
295.0
297.0
299.0
301.0
-------
STATION Sl-OCT
q
ri-
q
N-
o
j.
q
d-
q
o>
Ul
P
-------
O>
a:
o
Q
Q_
o
25-1
20-
15
10
13
00 5-
0
STATION 1 - OBSERVED AND MODELED TSM
FLOW PARAMETER - MODELED WAVE ORBITAL SPEED
MODEL PARAMETERS
CA=4.3 FPC=0.1 S=0.01 R=0.05
283
OBSERVED.
PREDldTED
287 29T 295
JULIAN DAY - 1986
—i—
299
303
u
-------
STATION S3 MAY-JUNE, 1906
CD
135.0 137.0
130.0
141.0 143.0 145.0 147.0 149.0
JULIAN DAY 1986
151.0
153.0
155.0
157.0
N
-------
STATION S3 MAY-JUNE, 1986
o
0-1
K3
Ui
_, , . p.
13S.O 137.0 139.0 141.0 143.0 145.0 147.0 149.0
JULIAN DAY 1986
101.0
153.0
155.0
1
157.0
-------
STATION S3 MAY-JUNE, 1986
135.0 137.0
139.0
141.0
143.0 145.0 147.0 149.0
JULIAN DAY 1986
151.0
153.0
155.0
157.0
N
-c
-------
STATION S3 MAY-JUNE, 1986
q
oo
:
2-
—I 1
139.0 141.0
—i 1 r
143.0 145.0 147.0 149.0
JULIAN DAY 1986
135.0 137.0
151.0 153.0 155.0 157.0
KJ
-------
12-j CONCENTRATION
<
4-1
PRB5KTED CONCENTRATION (fflg/U
a-
4-
16
MAY
1966
T
19
23
i
27
31
4
JUNE
8
262
-------
STATION S5-MJ
S.
o
ri
o
Si-
O
CV1-
135.0 137.0 130.0 141.0 143.0 145.0 147.0 149.0
JULIAN DAY 1986
151.0
153.0
155.0
157.0
-------
STATION S5-MJ
NJ
7)
o'
135.0 137.0 130.0 141.0 143.0 145.0 147.0 140.0
JULIAN DAY 1986
151.0
153.0
155.0
—l
157.0
-------
STATION S5-MJ
p
in-i
N3
o
ri-
sl-
o
3-
135.0 137.0 139.0 Ul.O
143.0 145.0 147.0 149.0
JULIAN DAY 1986
151.0
—I 1 1
153.0 155.0 157.0
—n
-------
STATION Sl-MJ
S.
135.0 137.0 139.0 141.0 143.0 145.0 147.0 149.0
JULIAN DAY 1986
151.0
153.0
155.0
157.0
-------
STATION Sl-MJ
NJ
&u
o
CM
*•
O
o
o
§
d
s
d
s
6
3
o
135.0 137.0 130.0 141.0 143.0 145.0 147.0 149.0
JULIAN DAY 1986
151.0
1530
155.0
157.0
-------
STTATION Sl-MJ
i-o
ON
00
£3-
Si 9
71
cs>
135.0 137.0 139.0 141.0 143.0 145.0 147.0 149.0
JULIAN DAY 1986
151.0
153.0
155.0
157.0
-------
STA 71, JULY 6- AUGUST 7,
NJ
ox
189.0 191.0 193.0 195.0 197.0 199.0 201.0 203.0 205.0 207.0 209.0 211.0 213.0 215.0 217.0 219.0
JULIAN DAY 1986
-------
STA 71, JULY 6- AUGUST 7,
to
•vj
o
SH
3j
o
6-
o
oi-
W o
U iC
(X,
189.0 191.0 193.0 105.0 107.0 190.0 201.0 203.0 205.0 207.0 209.0 211.0 213.0 215.0 217.0 210.0
JULIAN DAY 1986
-------
STA 71, JULY 6- AUGUST 7,
9
8-
3.
N
o
81
65*
189.0 191.0 193.0 195.0 197.0 199.0 201.0 203.0 205.0 207.0 209.0 211.0 213.0 215.0 217.0 219.0
JULIAN DAY 1986
-------
STATION 71 7/10-9/27/86 STARTS AT 0600
Ni
•-J
N5
o
ri
fc
jjj
.. it
18901010 1930 195.0 197.0 199.0 201.0 203.0 205.0 207.0 209.0 211.0 213.0 215.0 217.0 219.0
JULIAN DAY 1986
1
-------
STA 71, JULY 6- AUGUST 7,
189.0 191.0 193.0 105.0 107.0 199.0 201.0 203.0 205.0 207.0 209.0 211.0 213.0 215.0 217.0 219.0
JULIAN DAY 1986
-------
KJ
Station: STA
• i
1
I-
- t
a
I
o a 10 i& a> a so an 40 4» BO u to
TRIMS M
-------
STATION S5JA
NJ
S-,
o
ci
o
N
§8-
S 9
-- r-
189.0 191.0 193.0 105.0 197.0 199.0 201.0 203.0 205.0 207.0 209.0 211.0 213.0 215.0 217.0 219.0
JULIAN DAY 1986
-------
STATION S5JA
q
n -
N
14
q
o»-
a,
o
189.0 191.0 193.0 195.0 197.0 199.0 201.0 203.0 205.0 207.0 209.0 211.0 213.0 215.0 217.0 219.0
JULIAN DAY 1986
-------
STATION 5 7/10-9/27/36 STARTS AT O600
o
rj
q
a
JLL
lA A.
*t* " ' '"f^1* '" iif -Tilt «T. ^ IT I .t-j- j ......
189.0 191.0 193.0 195.0 197.0 199.0 201.0 203.0 205.0 207.0 209.0 211.0 213.0 215.0 217.0 219.0
JULIAN DAY 1986
es
-t
-------
STATION S5JA
K)
vj
00
2
6
o
6
8.
8
b
3
6
o
8
189.0 191.0 193.0 195.0 107.0 199.0 201.0 203.0 205.0 207.0 209.0 211.0 213.0 215.0 217.0 210.0
JULIAN DAY 1986
-------
STATION STA1-J
q
ri
q
Ji
to
890 191.0 193.0 195.0 197.0 199.0 201.0 203.0 205.0 207.0 209.0 211.0 213.0 218.0 217.0 219.0
JULIAN DAY 1986
-------
STATION 1 7/IQ-9/27/Q6 STARTS AT 0600
o
n
NJ
00
O
109.0 191.0 193.0 195.0 197.0 199.0 201.0 203.0 205.0 207.0 209.0 211.0 213.0 215.0 217.0 219.0
JULIAN DAY 1986
-------
STATION STA1-J
ISi
oo
i
890 191.0 193.0 195.0 197.0 199.0 801.0 203.0 205.0 207.0 209.0 211.0 213.0 215.0 217.0 219.0
JULIAN DAY 1986
-------
STATION STA1-J
o
id
q
sl-
00
INJ
Cu
189.0 191.0 193.0 195.0 197.0 199.0 201.0 203.0 205.0 207.0 209.0 211.0 213.0 215.0 217.0 219.0
JULIAN DAY 1986
-------
CO
LO
12
D)
E 10
8-
6-
UJ
I
o
o
z
UJ
OL
CO
ID
CO
O
0
STATION 1 - OBSERVED AND MODELED TSM
FLOW PARAMETER - MODELED WAVE ORBITAL SPEED
MODEL PARAMETERS
CA=4.3 FPC=0.1 S=0.01 R=0.05
DERIVED FROM FALL DATA
OBSERVED
PREDICTED
190
195 200
JULIAN DAY - 1986
205
210
-------
NJ
oo
60-1
8 60
Z
<
t
t: 40-
30-
co
z
QC
h-
Z 20
LLJ
O
QC
LLJ 10
CL
0
LAKE ST. CLAIR
850203
11
JULY
1985
16
~r~r
19
23
-J
27
31
4
AUG
8
-------
24-
00
Oi
O 23
UJ
cc
D
!<
n:
IU
CL
LU
h-
22-
21-
20
LAKE ST. CLAIR
860203
I " -" r T-
11
JULY
1085
1 1 '
15
1 1 '
10
• • , -r
23
1 ' I ' '
27
1 1 '
31
4
AUG
8
-------
00
ox
UJ
O
LU
O
CC
UJ
O.
75-i
E 60 H
2!
25-
0
D
11
FPT
i
16
21
i
26
1
OCT
i
6
11
-------
15-1
oo
Q
in
in
Q.
CO
LAKE ST. CLAIR
860206
n
SEPT
1086
— 1 — 1 — p^ — I — 1
16
i.ji.i
21
. I . .
26
1 ' I ' '
1
OCT
• • I • •
6
1 • i
11
-------
22-1
00
CO
o
O
LU
CC.
DC
UJ
UJ
18-
14-
10
LAKE ST. CLAIR
860206
11
SEPT
1985
T
16
T"
21
26
T
T—r
1
OCT
"T
6
11
-------
ACCUMULATION OF FALLOUT CESIUM-137 AND CHLORINATED ORGANIC
CONTAMINANTS IN RECENT SEDIMENTS OF LAKE ST. GLAIR
John A. Robbins and Barry G. Oliver*
ABSTRACT
Cesium-137 originating from atmospheric nuclear testing, lead-210,
potassium-40 and several chlorinated organic compounds (HCB,
Hexachlorobenzene; OCS, Octachlorostyrene; PCBs, Polychlorinated Biphenyls;
HCBD, Hexachlorobutadiene; QCB, Pentachlorobenzene; TCB, Total
Trichlorobenzene; TeCB, Total Tetrachlorobenzene, and Total DDT) have been
measured in a set of sediment cores collected by diver from Lake St. Clair
in 1985. Distributions of Cs-137 often show subsurface peaks which
apparently correspond to peak testing in the mid 1960s. Excess lead-210
distributions are similar to those previously encountered elsewhere in the
Great Lakes, possessing a zone of constant activity extending about 3 cm
down from the surface with exponential fall off below. Mixed depths and
sedimentation rates inferred from lead-210 profiles are consistent with Cs-
137 profiles and indicate sedimentation rates of the order of 0.1-0.2 cm/yr.
Observed profiles can also be correctly predicted by mixing processes
(rather than sedimentation) in which mixing below the surface region is
National Water Research Institute, Burlington, Ontario, Canada
289
-------
characterized by constant eddy diffusion coefficients of .2 to 4. cm2/yr.
Because of the evidence of extensive biological activity down to at least 15
cm, mixing is favored as the mechanism producing observed radionuclide and
contaminant distributions. Comparison of total loading of Cs-137 to the
lake (471 Ci in 1985) to actual storage (37 Ci) indicates a sediment
residence time of about 5 years. This value is consistent with that
inferred from previously observed changes in surficial sediment levels of
mercury and pesticides between 1970 and 1974. Changes in surficial sediment
Cs-137 concentrations between 1976 and 1985 are less than expected on the
basis of a five year residence time and suggest that the residence time of
particle-associated contaminants increases with the amount of time the
contaminant has remained in the system. Preliminary measurements of Cs-137
in trap samples collected by others at two sites in the lake indicate that
the isotope may be used to distinguish between particle-associated
contaminants resuspended from the bottom and new contributions from the St.
Clair River. Concentrations of chlorinated organic compounds in surface
sediments are well-correlated with each other and the pattern over the lake
bottom is closely related to the thickness of recent deposits. In some
cores profiles of total DDT and PCBs (as well as compounds with significant
local sources such as HCB, QCB and DCS) apparently reflect the history of
loading to the lake. Also, profiles of HCB/OCS and HCB/QCB ratios, used to
distinguish between alternative industrial sources of the chemicals,
indicate the changing history of waste management practices. Total storage
of contaminants shows that Lake St. Clair sediments are a significant
repository of chemicals passing through the Lake. As of 1985 it is
290
-------
estimated that 960 Kg of HCB, 870 Kg of PCBs and 210 Kg of DCS are contained
in the sediments.
INTRODUCTION
The enormous flow of water water from the upper Great Lakes proceeds
out of Lake Huron through the St. Clair River and into a shallow heart-
shaped body of water, Lake St. Clair. Urban and industrial activity on the
shores have made the river and this lake one of the most heavily
contaminated regions of the Great Lakes. Although the lake has an extremely
short mean hydraulic residence time, about 9 days, sediments manage to
acquire significant burdens of contaminants originating from tributary
sources, especially the St. Clair River. Previous studies of mercury
(Thomas et al., 1977) and chlorinated organic compounds (Frank et. al, 1977)
indicated that contaminants passing through the lake are temporarily
retained by sediments. The marked reduction which took place during the
four year period between 1970 and 1974 in the concentration of these
constituents in surficial sediments indicated that the sediment reservoir
was not a permanent sink however probably because resuspension in this
shallow lake (3m mean depth) ultimately exports materials out of the lake
and down the Detroit river.
These qualitatively characterized properties of the sediment reservoir
can be more quantitatively treated by determining the storage and
distribution of particle-associated radiotracers whose loading histories, in
291
-------
contrast to the above contaminants, are well-known. Fallout cesium-137 is
especially useful for this purpose since it is a tracer of fine-grained
(clay sized) constituents and has an accurately known input history. A
second isotope, lead-210 is also of considerable use, since it is delivered
to the lakes at a virtually constant, well-determined rate and has no known
anthropogenic sources. Thus Cs-137 can serve to illustrate the long-term
response of sediments to a pulse of radioactivity passing through the system
in the mid 1960s while lead-210 characterizes the steady-state response.
The recent severe contamination of the St. Glair River with chlorinated
organics from industrial activity in the Sarnia area has been documented
(Environment Canada, 1986, J. Water Poll. Res., 1986). It is known that
Lake St. Clair retains some portion of these contaminants since they have
been found at significant concentrations in the Lake's sediments (Frank et
al., 1977; Pugsley et al.,1985; Oliver and Bourbonniere, 1985). This report
also examines the distributions of the contaminants in surficial sediments
of the Lake, estimates the mass of contaminants stored in sediments, and
discusses contaminant trends in sediment cores.
METHODS
Sediment cores were collected by diver at the sites shown in Fig la.
during May and September 1985. The locations were chosen to coincide with
those occupied previously by Pugsley et al. 1986. Locations possessing sand
or coarser materials are unsuited for hand insertion of core tubes and, at
292
-------
least for cesium-37, are known to possess insignificant amounts of the
isotope (cf Robbins, 1986). Two 10.1 cm i.d. cores were collected at each
location with no observed disturbance of the sediment-water interface and
were suitably spaced from one another (ca 50 cm) to avoid interferences.
Cores were extruded hydraulically in the field and sectioned in 1 cm
intervals to 10 cm and in 2 cm from 10 cm to the bottom of the core or in
many cases to the interface with gray glacial clay. Sections from one core
were subdivided for radionuclide and metals analysis while sections from the
replicate core were subdivided for organics and additional metals analysis.
Because the accuracy of sediment tracer and contaminant inventories depends
directly on the extent of recovery of material collected in a tube of known
area, care was taken to collect all sediment from each section. Subsamples
for organics analysis were stored frozen in precleaned glass jars prior to
analysis.
Samples for radionuclide analysis were weighed, freeze dried and
reweighed to obtain the fractional dry weight of sediment and the mass of
dry sediment per unit area of the core. Dried samples were lightly
disaggregated and placed in vials of standard geometry for gamma counting.
The activity of Cesium-137 (661.6 KeV) and K-40 (1460.7 KeV) were determined
using a lithium-drifted germanium detector coupled to a multichannel
analyzer. Absolute detector efficiencies were determined by counting
sediment samples of equivalent geometry spiked with a mixture of gamma
emitting radionuclides of known activity. Uncertainties in sample
activities were generally well under ten percent. Samples for organics
analysis were soxhlet extracted prior to analysis by capillary, electron
293
-------
capture detector, gas chromatographic analysis. Procedure details have been
published (Oliver and Bourbonniere, 1985).
In addition, sediment trap samples were analyzed for Cs-137 from two of
nine sites occupied during 1985 by Charlton and Oliver (1986). Trap
materials were collected at nine (about equally spaced) time intervals
during the period from early June to mid-November. The sites 10 and 13 are
situated approximately 1 Km north of coring locations 24 and 28 respectively
(See Fig la.). Samples were freeze dried and lightly disaggregated prior to
analysis. Details of the collection and sample preparation as well as
supplementary analytical results are given in the above reference.
RESULTS AND DISCUSSION
Characteristics of Sediment Cores
In general sediment cores collected from this Lake evidence significant
stratigraphic inhomogeneity. Gelatinous, flocculent material resides on the
uppermost centimeter or so of sediment. The surface, particularly in the
deepest parts of the Lake is perforated by many circular holes about 0.5 cm
in diameter which extend as burrows into sediments as deep as 16 cm.
Densities of perforations reached as high as 800 M"2. The burrows,
surrounded by about a 2 cm diameter halo of oxidized sediment were often
inhabited by Hexagenia larvae. In regions of the lake with high clay
content, the upper 10-20 cm of sediments consisted of brownish gray silty
294
-------
material with some fine sand and occasional layers of shell fragments. In a
few of the cores a transition layer was encountered between 10 and 20 cm
consisting of clay with a high water content. An admixture of clay and sand
extended below this layer becoming consolidated and possessing increasing
sand content. In nearly all cores, at sufficient depths, the sediment
composition shifted abruptly to fine post-glacial clay. Sediments above
this layer may be considered to be of recent origin. The thickness of the
recent sediments, shown in Fig. lb., corresponds roughly with lake depth and
reaches a maximum of over 30 cm in a narrowly defined area. Cores collected
in shallower parts of the lake with low clay content occasionally possessed
macrophyte fragments and remnants of plants rooted in the sediments
collected. A large live clam (Unionid) was found in the region of 2-6 cm
depth in a sediment core (discarded) collected at site 39.
Sources of Radiocesium in Lake St. Clair
Cesium-137 in the Great Lakes originates almost exclusively from
atmospheric nuclear testing. There is no natural source of the radionuclide
and contributions from reactors situated along the shores are many orders of
magnitude less than direct fallout loadings. More than two decades of
routine monitoring of monthly fallout by U.S. and Canadian authorities have
provided the basis for developing estimates of time varying loadings which
greatly surpass, both in quantity and quality, our knowledge of any other
contaminant entering the lake system. The most detailed records are for Sr-
90 which is co-produced with radiocesium in nuclear detonations as a fission
product. Cesium-137 and Sr-90 have comparable half-lives (0.2 and 28.1
295
-------
years respectively) and the ratio of the two isotopes in air and in fallout
is well-determined and largely time-invariant. For this reason loadings of
Cs-137 may be reliably inferred from Sr-90 records as well as from direct
measurement. Data for both Sr-90 and Cs-137 from all monitoring sites in the
Great Lakes region have been combined with empirical models relating
atmospheric concentrations to over-lake and drainage basin deposition
(Robbins, 1985a). This regional source function is used to predict loadings
to Lake St. Clair using a lake system response model.
The response of the Great Lakes to fallout radionuclide loadings was
first examined by Lerman and Taniguchi (1972) for Sr-90 and subsequently by
others for the two long-lived fallout isotopes which are much more strongly
associated with particulate matter, Cs-137 and plutonium. Wahlgren et al.
(1980) related plutonium isotope loadings to measured aqueous concentrations
using a long term fate model which included removal to sediments through an
apparent settling velocity. Thomann and DiToro (1983) enlarged the
treatment to include the effects of equilibrium partitioning of plutonium
between water and suspended matter and included an explicit term for the
resuspension of the isotope. Tracy and Prandtl (1983) reported on new Sr-90
and Cs-137 concentration data for several of the Great lakes and illustrated
the need to consider resuspension in accounting for the long-term behavior
of cesium-137. Recently, Robbins (1985b) has developed an equivalent
formalism to calculate the response of the lakes to Cs-137 loadings which
includes an improved source function and adds the contributions of the
isotope from land drainage. This latter component in the model is based on
the semi-empirical relations developed by Menzel (1974) and was calibrated
996
-------
for the Great Lakes through periodic measurement of the concentration of the
isotope in tributary rivers (cf. Nelson et al.1984). Details of the
calculation are given in Robbins (1985b).
The results of the calibrated model calculation are shown in Figure 2.
Contributions from the watershed comprise a small (5X) virtually constant
fraction of the loading. Direct atmospheric contributions are initially
about 40% of the total load but decrease in recent years to less than 20X.
Inflow from Lake Huron contributes the largest portion of the loading: about
60X during the period of maximum fallout and increasing to more than 80X at
the present time. This increase reflects the persistence of small amounts
of the isotope in Lake Huron water. It is fortunate that lake Huron
provides the major input to Lake St. Clair as the most comprehensive set of
measured concentrations exist for this lake (Barry, 1973; Dasgupta, 1975;
Durham and Tamro, 1980; Alberts and Wahlgren, 1981). As a result, the
estimate of Cs-137 loadings are only weakly dependent on the assumptions of
the long term response model.
The validity of the loading estimate rests on several assumptions.
First, it is assumed that measured concentrations (of total Cs-37)
adequately represent the concentration in water exiting from the lake Huron.
Reported values (especially the pre-1970 data) are provided by Barry (1973)
and represent hypolimnetic waters. The values reported by Alberts and
Wahlgren (1981) are for the summer when the lake is stratified but no
significant difference exists between epi- and hypoliomnetic water. This is
important because concentrations of some elements (e.g. plutonium) decrease
297
-------
significantly in the epilimnion as stratification develops. If this were
the case for Cs-137, surface waters which presumably exit preferentially
from the lake would contribute less to Lake St. Clair during the summer
months than predicted by the model. Additionally, it assumed that the river
system conveying the radionuclide to the lake has no sinks. This is
probably true for the St. Clair river which has a coarse grained, scoured
bottom. However the delta and marshy areas which receive inflowing waters
could more plausibly store cesium-137. While we were unable to collect
suitable cores in these areas there are reasons why the effect should be
small. Deposits in the delta are coarse, sand-sized materials which
undoubtedly contain negligible amounts of the isotope. Finer grained
materials which carry the radionuclide accumulate in the open lake beyond
the delta area (Thomas et al., 1975). The marshy areas may trap fines
efficiently but the proportion of water which flows through them is very
small in comparison with open channel flow.
Vertical Distributions of Cesium-137
Representative profiles of Cesium-137 are shown in Figure 3 for sites
along a transect between stations 65 and 28 (see Fig. !.)• Concentrations
are generally low in comparison with surface sediments in the main lakes
where in Lake Huron values are typically around 20 dpm/g and 5-10 dmp/g in
Lake Erie. In most cases there is a peak in the activity which occurs 2 to
10 cm below the sediment surface. The extent of penetration of cesium-137
is generally shallow ranging from 2-4 cm toward the ends of the transect and
reaching a maximum of about 14 cm around the transect midpoint. The
298
-------
occurrence of subsurface maxima suggests that some information about the
history of deposition may be preserved in these deposits although the
correspondence between loading history and profile shape is generally weak
at best.
Profiles of cesium-137 at sites 69 and 38 exhibit the best defined
maxima of any from this Lake. As shown in Figure 4, both cores possess high
values of cesium-137 at about 10 cm with secondary maxima at about 25 cm and
15 at sites 69 and 38 respectively. If cesium-137 reached these sites by
direct transfer without sediment mixing or other integrative processes the
expected profile would appear as the dashed line in each case.
Sedimentation rates have been chosen to place the theoretical profile at the
position of the maximum activity. Inferred rates are 0.54 cm/yr and 0.43
cm/yr respectively for site 69 and 38. In both cases the calculated
distribution agrees very poorly with observation. The deeper maximum is not
reproduced nor are concentrations predicted successfully elsewhere away from
the primary peak region.
These profiles are apparently artifacts of stratigraphic
inhomogeneities as can be seen in the companion plots in Figure 4. In each
case the valley in the Cs-137 distribution is associated with (1) an
decrease in the sediment solids content (FDW-fraction dry weight) as
verified both by field observation and measurement, (2) an increase in the
amount of material which dissolves on treatment with 10% HCl(FSOL=fraction
soluble), (3) a marked increase in the K-40 content over an interval of a
few centimeters. Potassium-40 is a direct measure of the total potassium
299
-------
content of sediments and indicates the amount of clay sized minerals
present. These results in combination with field data indicate that such
discontinuities as seen in the Cs-137 profiles probably result from
interlayering of old fine-grained materials of relatively high water
content. As the two sites in question are relatively near the shipping
channel (Fig. 1.) it is plausible the such layers as well as the material
comprising the secondary peaks could originate from dredging operations.
The only other core showing double Cs-137 peaks is 71 (cf Fig. 5f) which is
also adjacent to the channel. Adjacent cores 65,68 and 70 do not have
double peaks but the activity of cesium-137 in these cores penetrates only a
few centimeters and may be subject to mixing of near surface sediments which
might obliterate such structure.
Vertical Distributions of Lead-210
Distributions of excess lead-210 are shown for eleven cores in Figure 5
a-f. Excess lead-210 is computed by subtracting the activity of supported
lead-210 from each value. Supported lead-210 is taken as the average of the
lowest values of lead-210 in each core, 1.5 +/- °-5 dpm/g- In estimating
the supported level, activities of the isotope within glacial clay were not
included inasmuch as the supported lead-210 content of the clay may not be
representative of background levels in recent sediments. In these cores the
K-40 content is essentially uniform down to the gray clay interface
indicating that distributions of lead-210 as well as other contaminants are
probably not artifacts of grain-size dependent sorting processes. In every
case distributions of excess lead-210 are characterized by a zone, extending
300
-------
down from the sediment water interface to varying depths, possessing a
constant activity. Below this depth the activity decreases virtually
exponentially. Exceptions to this pattern occur in cores 38, and 69 which
exhibit disturbances in the vicinity of the stratigraphic anomalies
discussed above.
The remaining distributions have the form seen elsewhere in the Great
Lakes, both in the open lake environment (Robbins and Edgington, 1975;
Robbins et al., 1978; Robbins, 1980) and in the silty, shallower recent
deposits of lower Saginaw Bay. In previous discussions such profiles have
been generally treated in terms of the rapid steady state mixing model
(RSSM) in which instantaneous mixing of sediment and lead-210 is postulated
to occur within a discreet zone at the sediment water interface. Sediments
below this zone are free of mixing or other disturbance and accumulate at a
constant rate. Details of the mathematical model are provided elsewhere
(Robbins et al., 1977). The solid lines shown in Figures 5 a-f are based on
the RSSM model, in which values for the depth of mixing (expressed in g/cm^)
and sedimentation rate (expressed in g/cm^/yr) are chosen to yield the best
weighted least-squares fit. Inspection of the figures shows that the
agreement between observation and the model is excellent except for the
previously discussed cores 38 and 69.
Apart from cores 38 and 69 (and 65) mixed depths have a relatively
narrow range, averaging about 3.5 (+/-1.3) cm. Sedimentation rates are also
narrowly confined, 0.19 (+/-0.06) cm/yr. If the process is correctly
characterized by the RSSM model then the ratio of the mixed depth( g/cm^) to
301
-------
the sedimentation rate (g/cm2/yr) is a measure of the time resolution for
reconstructing time histories from the sediment record. The mean ratio
(excluding 38,65 and 69) is 19 (+/- 12) yr.
Since the model results refer to sediment processes and not lead-210
per se, a test of the self-consistency of the RSSM model may be developed by
applying it to the Cs-137 profiles using the mixed depth and sedimentation
rate derived from lead-210. The results of this exercise are shown in the
companion plots in Figures 5 a-f. Cs-137 distributions are plotted along
with the solid curve which is the unadulterated model fit. In general (with
the exception of 38 and 69 as usual) the agreement is good. Dashed curves
also shown result from allowing the mixed depth and sedimentation rate to
vary so as to produce an optimized fit. In general the predicted
distributions are not much better. The effect of unconstraining the
solution is to increase the model mixed depth so as to fit the region where
the activity of the isotope falls off rapidly. This occurs with a loss of
structure in the peak region. In the case of core 65, considerable mixing
is implied by the Cs-137 data and no mixing by the lead-210 data. Such an
inconsistency would occur if some recent event removed enough surface
material either by scouring or by disturbance during collection of the core.
The idea of surface sediment loss in core 65 is further substantiated
by the total Cs-137 to total excess lead-210 ratio provided for each core in
Table 1. The three cores (38,69 and 71) with stratigraphic discontinuities
have ratios of 1.0-1.6. The remaining cores have ratios ranging from 0.6 to
0.8 while core 65 has a ratio of 0.48. Loss of surface sediment would
302
-------
reduce total cesium-137 in preference to excess lead-210 and suppress this
ratio. The mean ratio of the seven remaining cores is 0.69+/-0.08.
While the RSSM model provides a self-consistent account of
sedimentation processes, it is not necessarily an accurate one. Several
factors argue against sedimentation as the main control on profile shapes
below the mixed zone. First, the lake is nondepositional in the sense that
at most 30 cm of material has accumulated in the post-glacial period. Thus
the long-term net sediment accumulation rate must be considerably less than
the mean measured value of about 0.2 cm/yr. It could be argued that recent
sedimentation rates are higher than historical values perhaps because of
increased erosion of watershed soils etc. However, horizontal patterns of
deposition might be considered to remain the same even with increased
sedimentation, in which case RSSM rates ought to at least correlate with the
thickness of recent sediments. There is no significant correlation (r-0.45,
N-9). Third, organisms and their burrows extend to at least 15 cm depth and
their long-term effect should be partial mixing of sediments to considerable
depths. In open lake sediments Robbins (1982) showed that the RSSM mixed
depth correlated well with the range of occurrence of Oligochaete worms, but
those deposits were devoid of organisms such as Hexagenia and the Unionid
clams whose scale of sediment mixing is considerably greater.
If the profiles are chiefly the result of sediment mixing, the RSSM
results can be reinterpreted to yield mixing rate constants. It is well
known (Robbins, 1978) that exponential excess lead-210 profiles can also
result from uniform eddy diffusive mixing of sediments. Robbins used this
303
-------
approach (Robbins et al. 1978) to reconcile distributions of cesium-137,
lead-210 and pollen species in sediment cores from the shallow perturbed
sediments of western Lake Erie. If the logarithmic decrease in lead-210
activity is due to the combination of sedimentation and radioactive decay
then the slope of the activity with depth is given by 1/w, where 1 is the
radioactive decay constant (0.03114 yr'l). If it is due solely to the
combination of eddy diffusive mixing and radioactive decay then it is given
by sqrt(l/Kb) where Kb is the eddy diffusive mixing coefficient in cm2/yr.
Since the " best" value of w is already determined, the equivalent value of
Kb is given by w2/!. These values are reported in Table 1 as the deep
sediment mixing model (DSMM) values. According to this view, sediments near
the surface experience intense mixing characterized by eddy diffusion
coefficients sufficiently large to produce complete homogenization. Below
the mixed zone, values decrease rapidly but continuously to values given by
the DSMM model. Such a depth-dependent eddy diffusion coefficient will also
produce self-consistent results when applied to vertical distributions of
Cs-137.
Horizontal Distributions of Cs-137 Based Quantities
The extent of penetration of CS-137 into sediments is a model
independent quantity which shows how far particle associated contaminants
introduced in the mid 1960s are found more than 20 years later. The contour
maps of the depth in cm (Fig. 6a) and in g/cm2 (Fig. 6b) show that the
penetration depth follows the thickness of the recent sediments quite
closely (r-0.80, N-31). This is also true for the relation between the
304
-------
thickness of recent sediments and the surface (0-1 cm) concentration of Cs-
137 (Fig. la., r-.66, N-33) and the total accumulation (vertically
integrated) Cs-137 (Fig 7b., r-0.82, n-35). The central area in the middle
of Fig. 7b corresponds to the region of the two cores with stratigraphic
anomalies (38 and 69) which possess unusually high amounts of Cs-137 (58 and
117 dpm/cm2 respectively). The total accumulation of Cs-137 in sediments of
the Lake as of 1985 is 37 Ci corresponding to an average of 7.4 dpm/cm2 of
sediment. This may be compared with a total loading (decay corrected to
1985) of 470 Ci. Thus, roughly 8X of all the Cs-137 which has passed
through the lake remains in sediments today.
The concentration of Cs-137 in surface (0-2 cm) sediments 9 years
earlier is shown in Fig. 8a. These samples were collected by Shipek grab
(Thomas et al., 1977) and represent the upper 2 cm of sediment if the sample
is undisturbed. In order to compare these results with 1985 data the effect
of grab sampling was simulated by vertically integrating measured
distributions of Cs-137 from 0 to 2 cm. The resulting contour is shown in
Fig. 8b. Also simulated was the possible effect of losses of surface
material in grab sampling. A 1 or 2 cm loss had only a small effect on the
estimate of accumulation (20X or less) partly because the higher water
content of surface materials reduces the amount which those levels
contribute to the total accumulation. Levels of Cs-137 are considerably
higher in 1976 than in 1985 partly because of radioactive decay which
accounts for 19Z of the difference.
305
-------
Since there has been virtually no new input to the system during this 9
year period, the ratio of 1985 surface Cs-137 to 1976 values (decay-
corrected to 1985) is a direct measure of the removal of the isotope from
surface layers .
The ratio, contoured in Fig 9, indicates that at selected sites 50Z or
more of the activity has gone in 9 years. Elsewhere within the main body of
the deposits the ratio is between 50 and 25Z removed while toward the
margins the ratio approaches 100Z retained. Sediments at the margins of the
deposit consist of coarse material (sand) with occluded fine grained
sediments containing Cs-137. Evidently, contaminants that manage to get
incorporated into coarse sedimentary materials possess a longer residence
time in the system.
Sediment Reservoir Residence Times
The changes in activity of surface materials over time, as well as the
retention of the isotope in sediments, are measures of the residence time of
contaminants in the sedimentary reservoir. Since the sediments are
essentially nonaccumulating, the reservoir can be treated as a system in
which the rate of mass input is exactly balanced by removal. The flux of
tracer out of the reservoir F' (dpm/cm^/yr) is given by
(F-F')/Tp -IF' (1)
306
-------
where F is the flux into the pool and Tp is the pool residence time (yr)
The amount of tracer in the pool at any time is given by
S(dpm/cm2)-F'*Tp,
and the concentration of tracer anywhere in the pool (considered to be
uniformly mixed) is
C(dpm/g)-F'/rs,
where rs is the mass flux (which equals the resuspension flux in g/cm2/yr)
Under steady state conditions which may be approximated for lead-210,
dF'/dt-O so that the amount in the pool is given by
S(lead-210)-F'*Tp=F/(l+lp)
where lp-1/Tp. Since in a closed system, the amount of lead-210 stored
would be F/l, the fraction of lead-210 retained in a system with a residence
time of Tp is
fr-lTP/(l+lTp)
For Cs-137 the fraction retained is similarly derived but must be evaluated
by numerical solution of Eq. 1. The resulting relations between fr and Tp
for both Cs-137 and lead-210 are shown in Fig. lOa. For Cs-137, fr, is of
course time -dependent and has been calculated for 1985. Also shown in this
figure is the value of fr based on total storage of Cs-137 in the Lake (8%).
307
-------
This value corresponds to a pool residence time of 6 years. Lead-210
retention is also about 8Z, but this value is based on a very limited number
of cores and may not be representative. Additionally the estimate of
loading of lead-210 to the lake is based on model estimates of the
concentration of the radionuclide in Lake Huron water. In the calculation
it is assumed that the Kd of lead-210 in water is 106. With these
assumptions the pool residence time based on lead-210 is about 3 years.
It can be seen from the above equations that if the pool contains a
given amount of contaminant and receives no further inputs, the
concentration of nonradioactive (and nondegrading) contaminant in the pool
will decrease according to the relation
(6)
so that the pool residence time may be calculated as
Tp-t/ln(c0/c) . (7)
Thus the changes in mean mercury concentrations (Thomas et al., 1977) and
chlorinated organic concentrations (Frank et al.,1977) in the four year
period between 1970 and 1974 lead to additional estimates of the pool
residence times. These values are given together with those for Cs-137 and
lead-210 in Table 2. In general values are very consistent and imply a mean
pool residence time of about 4 +/- 1 year.
308
-------
The pool residence time may also be calculated (from Eq. 1-5) by
comparing changes in Cs-137 in surface sediments between 1976 and 1985. The
ratio as a function of pool residence time, shown in Figure lOb, rises to a
value for very short residence times which reflects loading numbers
integrated over a very short time span. The measured mean ratio, 0.6,
predicts a pool residence time of 16 years as indicated by the dashed line
in the figure. This value differs appreciably from those calculated above.
Whether the difference is significant or not is uncertain. However this
ratio reflects changes taking place long after the initial pulse of Cs-137
passed through the lake. In contrast, the comparison of Cs-137 storage with
the total loading emphasizes the processes of removal occurring during the
pulse period of the mid 60s.
Since considerably less Cs-137 was lost from surface sediments between
1976 and 1985 than expected on the basis of a 4 year mean residence time,
perhaps the residence time of the tracer increases with pool contact time.
Certainly this is consistent with the observation that little or no change
took place in concentrations of the tracer toward the margins of the
deposit. It also can be shown that if depth-dependent mixing controls the
downward transport of cesium-137, the residence time will increase with age
because material reaching deep sediments will be kinetically impeded from
returning to the surface where it can be exported through resuspension.
Menzel (1974) has described a somewhat analogous situation in which the
residence time (reflecting losses through runoff) of Sr-90 stored on
watersheds increases over the years as the isotope migrates deeper into
layers not accessible to erosion.
309
-------
A further estimate of pool residence time can be be obtained from the
ratio of total Cs-137 to total excess lead-210 in individual cores. This
method does not rely on horizontal integration methods. Apart from the few
anomalies discussed above, the ratio shows little variability as seen in
Table 1. The theoretical ratio, shown in Figure 11 versus pool residence
time, is defined as an envelope to reflect the uncertainty in the estimate
of lead-210 loading from Lake Huron. Values of the distribution coefficient
used to calculate the expected concentration in Lake Huron water are assumed
to be comparable to that in Lake Michigan (cf Eadie and Robbins 1987). The
measured ratio (0.69+/--08) predicts a pool residence time of 11 +/- 5
years. With direct measurements of the lead-210 loading based on the 210pb
content of Lake Huron water much uncertainty in this estimate would be
removed.
Cs-137 in Sediment Traps
The activity of Cs-137 in trap samples (Fig. 12) contrasts markedly
between stations. At station 10 (near coring station 24 in Fig. la.) the
activity is about 2 dpm/g during the early period and rises to a maximum of
3.7 dpm/g by the end of July. By October activities are down again
approaching levels of 2 dpm/g. In contrast with site 10, little change
occurs in the activity of the radionuclide in trap material at site 13 (near
core station 28). There is no peak and activities decline steadily from
1.45 dpm/g in early June to 0.8 dpm/g by mid-September, thereafter rising
toward the previous early June value. In contrast to the behavior of Cs-
310
-------
137, K-40 is essentially invariant with time of year and with location.
Values shown in Table 3 indicate no significant differences in K-40 content.
Without additional study the cause for these variations remains
unclear. However it is likely that sediment traps collect two distinct
populations of particles: those resuspended from sediments and those
entering the lake for the first time primarily by inflow from Lake Huron.
The Cs-137 activity on particles in Lake Huron itself may be inferred from
measured total concentrations of the isotope in water as of 1985 (0.06
dpm/L, Robbins, 1985b), the well-known partition coefficient for Cs-137 in
open waters of the Great Lakes (3xl05 ml/g, Alberts et al. 1981) and the
total suspended matter in the lake (TSM-0.5 ppm). These values combine to
give a Cs-137 concentration on particles of about 20 dpm/g. This value
agrees well with concentrations of the isotope in surface sediments of the
lake (Robbins et al., 1977 and Robbins 1980). Because TSM probably
increases and water composition is altered in the St. Clair River,
partitioning of Cs-137 between particle and solution phases could change.
However it is likely that the activity of cesium-137 on particles entering
Lake St. Clair from the river greatly exceeds (by about a order of
magnitude) activities of the isotope on particles resuspended from the
bottom of the lake. Such a large difference can be exploited to infer the
admixture of particles originating from the two sources which end up in
sediment traps. The data are consistent with this interpretation.
During the summer months when the lake is less mixed physically,
resuspension plays a smaller role in loading traps, especially those located
311
-------
along the direct course of water flow between the St. Clair and Detroit
River. With less efficient horizontal averaging during the summer period a
site such as 10 which is closer to the streamlines will record a higher
proportion of particles from river inflow. A site such as 13 which is
further from the streamlines will be less sensitive to river contributions
and collect a greater proportion of resuspended materials. This idea is
also supported by a comparison of the Cs-137 concentration of the trap
samples with concentrations of the radionuclide in surface sediments
underlying the traps. Site 10 is situated in the middle of a broad region
where the surface activity (0-1) cm is 2.0 dpm/g. Thus, in early June and
again after mid-September, levels in traps at this site exactly match
surface sediment concentrations. In contrast Site 13 is situated over
sediments with an activity of 0.2 dpm/g. During the early part of the
observation period, values are close to those at Site 10 and in fact reflect
the mean concentration of Cs-137 in surface sediments over a wide area. As
the season progresses values decrease, presumably as a result of a decrease
in the extent of horizontal mixing, and approach a value of 0.8. However the
decrease in activity does not reach the value of 0.2 which would represent
concentrations directly under the trap. This occurs presumably because at
least some material is derived from inflow and because even under relatively
quiescent conditions, regions of the lake bottom with higher surface Cs-137
concentrations contribute to the trap as well.
While the results of Cs-137 measurements in trap materials from this
lake are only preliminary, it is clear that the radionuclide will be of
considerable use in characterizing the seasonal and perhaps shorter term
312
-------
processes of transport of contaminants through the system. Of particular
interest is the use of the isotope to sort out the relative contributions
that new (inflow) and old (resuspension) sources of particle-associated
contaminants make to concentrations measured in water at various places in
the Lake.
Chlorinated Organic Compounds in Surficial Sediments
Typical contaminant distribution maps for Lake St. Clair surficial
sediments (0-1 cm) are shown in Figure 13. The highest contaminant
concentrations are found near the center of the lake in the region of
greatest water depth. The sediment cores show the greatest accumulation of
fine-grained sediments and the thickest layer of recent sediments over
glacial clay in this vicinity as well. Some minor contaminated sediment
accumulation also occurs in Anchor Bay at the northern end of the lake. For
the most part the sediments in the rest of the lake are sandy and, like Cs-
137, contain very low concentrations of organic contaminants. Table 4 shows
the range and mean values for several halogenated organic compounds. These
concentrations are in good agreement with other previously published
surficial sediment data for the Lake (Frank et al., 1977; Pugsley et al.,
1985; Oliver and Bourbonniere, 1985). Although the mean contaminant
concentrations, with the possible exception of HCB, are not particularly
high compared to other areas in the Great Lakes Basin, the maximum
concentrations reach significant levels for many of the Sarnia-source
contaminants.
313
-------
Trends from Sediment Cores
Fig. 14 shows the concentration of HCB (hexachlorobenzene), DCS (octa-
chlorostyrene), PCBs (polychlorinated biphenyls), HCB (Hexachlorobutadiene),
QCB (pentachlorobenzene) and DDT (total DDT-sum of DDT and degradation
products) in sediment cores from the lake. The absolute concentrations of
the contaminants in the cores varies considerably with location. For
example the maximum HCB concentration is 180 ppb in core 38 and only 20 ppb
in core 84. Two of the shallower cores, 64 and 71 (ca 10 cm mud over glacial
clay) had similar contaminant profiles. Chemical concentrations were fairly
constant in the upper 6-7 cm for most compounds, presumably reflecting the
mixing processes evidenced in the radionuclide profiles. Below the constant
concentration zone, levels gradually decrease and approach near zero
(undetectable levels) at the silt/clay interface. Core 84 exhibited
peculiar behavior with several high concentration spikes for various
compounds appearing deep in the core. This unusual behavior might be due to
translocation of active surface sediments to discreet depths through the
reworking behavior and burrow infilling by Hexagenia larvae found in this
and other cores.
Cores at sites 38 and 69, possessing the greatest thickness of recent
sediments, provide some consistent information on the loading history of
certain compounds. The loading pattern of DDT to the Great Lakes is well-
known (Rappaport et. al., 1985) and is closely related to U.S. usage data.
DDT production began in 1944, peaked in 1959 and stopped in 1972. For core
38, a major subsurface maximum for total DDT is observed at 13 cm and the
314
-------
peak profile crudely follows production history. However an anomalous early
peak at 19 cm for total DDT and for some of the other chemicals may indicate
that some disturbance in the sedimentation process has occurred. The time
of this disturbance (likely due to dredging in the area) seems to be before
the peak DDT input to the lake or of the order of 25-30 years prior to 1985.
Despite this problem, the top 6 cm of the core still seems to provide a
useful indication of contamination trends in the lake. The PCS maximum in
this core occurs at a depth of about 8 cm. Peak PCB production in the
United States occurred in 1970 (Peakall, 1975). Contamination history can
be roughly inferred by using total DDT and PCB peaks as markers, 1959 (13
cm) and 1970 (8 cm) respectively, assuming production/usage scenarios can be
used to infer loadings to Lake St. Clair. OCS in core 38 peaks at 8 cm and
shows a significant decline in recent years. In contrast, HCB, HCBD and QCB
exhibit a fairly uniform concentration in the upper 6-8 cm of this core.
In core 69, very high concentrations of many contaminants are found
below 18 cm. This appears to be due to site disturbance at a time when DDT
input to the lake was high (probably around 1960). Normal contaminant
distributions are found above 16 cm. Again, PCB and OCS peaks occur at
about 8 cm depth with dramatic decreases in concentration towards the core
surface. HCBD and QCB concentrations are high and relatively constant in
the 0-8 cm range in the core. HCB concentrations seem to increase steadily
toward the core surface with the highest concentration at the surface.
Both DDT and PCBs entered the lake mainly from diffuse non-point
sources including the tributaries. The other four chemicals originate
315
-------
mainly from industrial activity in Sarnia. Like DDT and PCBs, OCS and QCB
exhibit higher concentrations deeper in the sediment. Reduced surface
concentrations apparently reflect the decrease in loading of the chemicals
which has likely occurred in recent years. Both HCB and HCBD concentrations
are increasing or are staying fairly constant near the top of the cores.
Loadings of these chemicals to the lake are evidently not dropping
significantly and may even be increasing.
Studies of sediments from the St. Clair River have shown that the
ratios of HCB to OCS and QCB are useful in tracking the source of
contaminants in the river (Oliver and Bourbonniere, 1985. The HCB/OCS and
HCB/QCB ratios are 1.3 and 4.0, respectively, for sediments near the Scott
Road Landfill, a site which contains waste byproducts from Dow's early
production of chlorine and chlorinated solvents. The HCB/OCS and HCB/QCB
ratios just below Dow's outfall and where non-aqueous wastes have leaked
into the river are 16 and 23 respectively. In 1980 Dow began treating the
Scott Road leachate by carbon filtration to reduce loadings from the site.
In cores 38 and 69 the HCB/OCS ratio changes from 2 lower in the core to 9
at the surface. Similarly the HCB/QCB ratio increased from 4 near the
bottom to 20 near the surface. Thus, the trends in the Lake St. Clair cores
are consistent with a diminishing contribution from the Scott Road Landfill
and contributions from DOW's outfall which have increased since the
beginning of the decade.
A quantitative treatment of these trends requires further measurement
of upstream loading and development of better models for characterizing the
316
-------
nature of transport processes in these sediments. While vertical transport
is probably dominated by a combination of physical and biological mixing
with little net sedimentation, changes in the loading to sediments can
nevertheless leave behind a record which to some extent corresponds to the
loading history as discussed above for the radionuclides and organics.
Interpretation of sedimentary profiles in this energetic, shallow water
environment must be approached with caution because of important processes
which can affect and perhaps override that of continuous mixing. These
include (1) intense resuspension which operates with varying intensity over
the lake bottom, (2) discontinuous (event-related) processes of
sedimentation and resuspension, (3) spatial heterogeneity in the biological
and physical mixing process, (4) horizontal reintegration of sediment loads,
(5) particle sorting effects, (6) and the effects of human activities such
as fishing, anchoring, boating, shipping operations and dredging. Sediment
core profiles are the cumulative end result of these many potential effects
and simple models, even self consistent ones, may well be inadequate.
Contaminant Storage and Loading
Depth-integrated samples (interval composites) were prepared and
analyzed for each core to estimate the mass of contaminants stored in the
sediments. Horizontal distributions in total storage have patterns which
are essentially congruent with the thickness of recent sediments and form
the basis for estimating total storage in the Lake (contour integration).
For the sandy non-accumulating areas, where cores were not collected, a
value of 5 ng/cm^ was used for PCBs and HCB and a value of 0.5 ng/cm^ was
317
-------
used for OCS. These approximations are not critical since the sandy areas
contributed less than 5X of the contaminant mass for these chemicals. Lake
St. Clair sediments presently contain about 960 Kg of HCB, 870 Kg of PCBs
and 210 Kg of OCS.
These values are much higher than the contaminant masses found by
Oliver and Pugsley (1986) for the St. Clair River sediments (69 Kg HCB, lOKg
OCS) indicating that Lake St. Clair is a more significant repository for
chemicals than the river itself. Recent loading estimates for HCB and OCS
in the combined dissolved and particulate fraction at Port Lambton in the
St. Clair River were 180 Kg/yr for HCB and 11 Kg/yr for OCS (Chan et al.
1986). At these rates Lake St. Clair sediments contain about a 5 year
supply of HCB and a 20 year supply of OCS. Thus, the sediments retain
significant fractions of these chemicals and, given the uncertainties in the
calculation, accumulation is consistent with sediment reservoir residence
times derived from historical studies of metal and organic chemicals in the
system and from the response of sediments to particle-associated
radionuclides.
A crude estimate of the total historical loading of particle-associated
organics can be calculated if it is assumed that their behavior is similar
to that of 13?cs. The following values were estimated by multiplying the
average ratio of organics (ng/cm2) to 137cs (dpm/cm^) for cores 64, 71, 84,
38 and 69, by the total mass of l^^Cs which was loaded into the system (470
ci converted to dpm): HCB 15 metric tons (MT); QCB 2.8 MT; HCBD 3.3 MT; OCS
4.3 MT; PCBS 20 MT; and total DDT 5.4 MT. These calculations show that
318
-------
considerable quantities of chemicals have been discharged into the system.
The total masses are probably on the low side for constituents which possess
a greater solubility or have a longer history of loading than radiocesium.
ACKNOWLEDGEMENTS
The authors wish to acknowledge the assistance of K. Hill and Technical
Operations staff of The Canada Centre for Inland Waters for help in the
collection of cores and to Mrs. Alena Mudroch for assistance in the field
sectioning of sediment cores. Thanks are due M. Charlton for the loan of
trap sample material for cesium-137 analysis. The help of R. Rossmann and
E. Meriweather in carrying out laboratory analysis of sediments for lead-210
and of K. Nichol for assistance in analysis of organic constituents is
gratefully acknowledged.
319
-------
LITERATURE CITED
Alberts, J. J. and M. W. Wahlgren. 1981. Concentrations of 239/240^f 137CS|
and 90sr in waters of the Laurentian Great Lakes: comparison of 1973
and 1976 values. Environ. Sci. Technol. 15: 94-98.
Anon. 1986. St. Clair River Pollution Investigation, Environment Canada and
Ontario Ministry of the Environment
Anon. 1986. St. Clair River Pollution, Water Pollution Res. J. Can. (J.
Lawrence, Ed.) 21: 283-459.
Barry, P. J. 1973. Estimating dose commitments to populations from
radioactive waste disposals to large lakes. In: Environmental Behavior
of Radionuclides Released in the Nuclear Industry. International Atomic
Energy Agency, Vienna, pp. 499-505.
Chan, C. H., Y L. Lau, and B. G. Oliver. 1986. Measured and modelled
chlorinated contaminant distributions in St. Clair River water. Water
Poll. Res. J. Can. 21: 332-343.
Charlton, M. N. and B. G. Oliver. 1986. Chlorinated organic contaminants on
suspended sediment in Lake St. Clair. Water Poll. Res. J. Can. 21:
380-389.
320
-------
Dasgupta, A. K. 1975. Radioactivity in Lake Huron water, summary of 1963 to
1973 data, and radioactivity in Lake Ontario water 1971-1973. In:
Radio-activity in the Great Lakes, a summary of radionuclide monitoring
and nuclear plant discharge data available Jan. 1975. Prepared by the
Radio-activity work group of the International Joint Commission, Great
Lakes Water Quality Board by M. S. Olijnyk and R. W.Durham.
Durham, R. W. and G.H. Jamro. 1981. Great Lakes radiological surveillance
1980. Environment Canada, National Water Research Institute,
Burlington, Ontario, Canada. July, 1981. 6 pp.
Eadie, B. J. and J. A. Robbins. 1987. The role of paryiculate matter in the
movement of contaminants in the Great Lakes. In Sources and Fates of
Aquatic Pollutants (R. A. Hites and S. J. Eisenreich, Eds.) Advances in
Chemistry Series 216, American Chemical Society, Washington, D. C., pp.
319-364.
Frank, R. , M. Holdrinet, H. E. Braun, R. L. Thomas, A. L. W. Kemp AND J.-M.
Jaquet. 1977. Organochlorine insecticides and PCBs in sediments of Lake
St. Clair (1970 and 1974) and Lake Erie (1971). Sci. Tot. Environ.
8:205-227.
Lerman, A. and H. Taniguchi. 1972, Strontium-90 in the Great Lakes:
concentration- time model. J. Geophys Res. 77: 3256-3264.
321
-------
Menzel, R. G. 1974. Land surface erosion and rainfall as sources of
strontium-90 in streams. J. Environ. Qual. 3: 219-223.
Nelson, D. M., D. N. Metta and J. 0. Kartunnnen. 1984. 239/240pu> 137cs, and
90 Sr in Tributaries of Lake Michigan. Environmental Research Division
Annual Report, Jan.-Dec. 1983, Argonne National Laboratory, Argonne,
II., Dec. 1984, pp.45-54.
Oliver, B. G. and R. A. Bourbonniere. 1985. Chlorinated contaminants in
surficial sediments of Lakes Huron, St. Clair and Erie: implications
regarding sources along the St. Clair and Detroit Rivers. J. Great
Lakes Res. 11: 366-372.
Oliver, B. G. and C. W. Puglsey. 1986. Chlorinated Contaminants in St. Clair
River sediments. Water Poll. Res. J. Can. 21: 368-379.
Peakall, D. B. 1975. PCBs and their environmental effects. CRC Crit. Rev.
Envrion. Cont. 5: 469-508.
Pugsley, C. W., P. D. N. Hebert, G. W. Wood, G. Brotea AND T. W. Obal. 1985.
Distribution of contaminants in clams and sediments from Huron-Erie
corridor. I-PCBs and octachlorostyrene. J. Great Lakes Res. 11:
275-289.
Rappaport, R. A., N. R. Urban, P. D. Capel, J. E. Baker, B. B. Looney, S. J.
Eisenreich AND E. Gorham. 1985. Chemosphere 14; 1167-
322
-------
Robbins, J. A. 1986. Sediments of Saginaw Bay, Lake Huron: Elemental
Composition and accumulation rates. Special Report 102, Great Lakes
Research Division, University of Michigan, Ann Arbor, Michigan. 102 pp.
(Appendix, 278 pp.)
Robbins, J. A. 1985a. Great Lakes regional fallout source functions, Great
Lakes Environmental Research Laboratory Technical Memo, ERL-GLERL-56,
Feb. 1985, 22 pp.
Robbins, J. A. 1985b. The coupled lakes model for estimating the long-term
response of the Great Lakes to time-dependent loadings of particle
associated contaminants. Great Lakes Environmental Research Laboratory
Technical Memo, ERL GLERL-57, Apr. 1985,41 pp.
Robbins, J. A. 1982. Stratigraphic and dynamic effects of sediment reworking
by Great Lakes zoobenthos. In: Developments in Hydrobiology 9,
Sediment/Freshwater Interaction, Ed. P. G. Sly. Proc. of the 2nd
International Symposium on Sediment-Water Interactions, Kingston, Ont.
Hydrobiologia 92 (1982) 611-622.
Robbins, J. A. 1980. Sediments of southern Lake Huron: elemental composition
and accumulation rates. U.S. Environmental Protection Agency,
Ecological Research Series, EPA-600/3-80-080. August, 1980. 309 pp
(Appendix, 198 pp.)
323
-------
Robbins, J. A. 1978. Geochemical and geophysical applications of radioactive
lead. In: Biogeochemistry of lead in the environment, Part A. (J. 0.
Nriagu, Ed.), Elsevier Scientific Publishers, Amsterdam, Netherlands,
Vol. 1A: 285-303.
Robbins, J. A. and D. N. Edgington. 1975. Determination of recent
sedimentation rates in LAke Michigan using Pb-210 and Cs-137. Geochim.
Cosmochim. Acta. 39: 285-304.
Robbins, J. A., D. N. Edgington AND A. L. W. Kemp. 1978. Comparative lead-
210, Cs-137, and pollen geochronologies of sediments from LAkes Ontario
and Erie. Quatern. Res. 10: 256-278.
Robbins, J. A., J. R. Krezoski and S. C. Mozley. 1977. Radioactivity in
sediments of the Great Lakes: postdepositional redistribution by
depositfeeding organisms. Earth Planet. Sci. Lett. 36: 325-333.
Thomann, R. V. and D. M. DiToro. 1983. Physico-chemical model of toxic
substances in the Great Lakes. J. Great Lakes Res. 9: 474-496.
Thomas, R. L. , J.-M. Jaquet, and A. Mudroch. 1977. Sedimentation processes
and associated changes in surface sediment trace metal concentrations
in Lake St. Clair, 1970-1974, Proc. Int. Conf. on Heavy Metals in the
Environment, Toronto, 1975, pp. 691-708.
324
-------
Tracy, B. L. and F. A. Prantl. 1983. Twenty-five years of fission product
input to Lakes Superior and Huron. Water, Air and Soil Poll. 19: 15-27
Wahlgren, M. A., J. A. Robbins and D N. Edgington. 1980. Plutonium in the
Great Lakes. In Transuranics in the Environment^ W. C. Hanson (Ed.),
Report No. DOE/TIC-22800, Technical Information Center, U. S. Dept. of
Energy, Washington D.C. pp 659-683.
325
-------
Table !• Summary of lead-210 and cesium-137 mixing and sedimentation model results.
RSSM Mixed depth
2
ite (g/cm ) (cm)
17 1.7 2.8
18 4.3 4.8
20 3.0 3.0
21 5.1 5.1
23 4.1 5.0
24 2.0 2.8
38. 10. 9.
64 2.8 3.1
65 0.2 0.2
69 ? ?
71 1.2 1.2
RSSM Sedimentation
2 Rate
(g/cm /yr) (cm/yr)
0.29
0.19
0.14
0.12
0.14
0.20
0.16
0.21
0.29
1.6
0.14
0.34
0.18
0.10
0.11
0.13
0.20
0.14
0.21
0.20
1.3
0.09
RSSM Time DMM
Resolution Rate
(yr) (cm /yr)
6.
22.
21.
44.
30.
10.
60.
13.
<1
?
9.
3.7
1.0
0.4
0.4
0.57
1.2
0.65
1.4
1.3
54.
0.23
Total
(dpm/cm )
29.6
27.3
22.7
30.7
34.3
36.5
49.5
36.7
18.7
75.6
12.1
Total Total
-------
Table 2. Sediment reservoir residence times inferred
from radionuclide storage as of 1985 and changes in mean
contaminant levels from 1970-1974
Constituent Residence time( yr) Reference
Cesium-137
Excess lead- 210
Mercury
DDE
IDE
DDT
Total PCBs
6.0
3.0
4.0
3.6
4.6
2.9
6.2
This work
ibid.
Thomas et al.(1974)
Frank et al . (1977)
ibid.
ibid.
ibid.
Mean
327
-------
Table -3 Cs-137 and K-40 in trap samples collected from Lake
St. Clair in 1985*.
Location-
Sample number
10-401
10-402
10-403
10-404
10-405
10-406B
10-408
10-410
Collection
Period
5/30-6/5
6/5-6/26
6/26-7/10
7/10-7/31
7/31-8/21
9/13-9/25
9/25-10/17
10/17-11/17
Cesium-137
(dpm/g)
1.93 ± 0.20
1.90 ± 0.15
2.73 ± 0.30
3.00 ± 0.30
3.70 ± 0.50
2.71 ± 0.34
2.10 ± 0.15
2.19 ± 0.13
Potassium-40
(dpm/g)
30 ± 4
34 ± 3
34 ± 10
22 ± 8
26 ± 10
57 ± 10
27 ± 3
37 ± 3
Mean 2.53 ± 0.62 32 ± 9
13-401
13-402
13-403
13-404
13-405
13-406
13-406B
13-408
13-410
5/30-6/5
6/5-6/26
6/26-7/10
7/10-7/31
7/31-8/21
8/21-9/13
9/13-9/25
9/25-10/17
10/17-11/17
1.45 ± 0.17
1.17 ± 0.06
1.33 ± 0.30
1.19 ± 0.16
1.01 ± 0.18
1.18 ± 0.19
0.80 ± 0.10
0.94 ± 0.10
1.33 ± 0.11
34 ± 4
33 ± 2
30 ± 9
33 ± 4
33 ± 6
33 ± 5
32 ± 3
30 ± 2
35 ± 3
Mean 1.16 ± 0.21 33 ± 2
* Trap collections by Murray Charleton, Canada Centre for Inland
Waters.
328
-------
Table 4. Chlorinated organic compounds in surficial (0-1 cm)
sediments of Lake St. Clair in 1985 (ng/g).
Compound Range Mean
Hexachlorobenzene (HCB) 0.4-170 32
Octachlorostyrene (OCS) n.d. - 21 4.8
PCBs n.d. - 60 19
Hexachlorobutadiene (HCBD) n.d. - 32 5.4
Pentachlororbenzene (QCB) n.d. - 8.7 3.2
Total Trichlorobenzene (TCB) n.d. - 28 4.3
Total Tetrachlorobenzene (TeCB) n.d. - 20 3.7
Total DDT and metabolites (SDDT) n.d. - 12 3.8
329
-------
LIST OF FIGURES
Fig. 1. (a) Locations of the 1985 coring sites in Lake St. Glair. Stations
correspond to those occupied previously in the U. Windsor study
(Pugsley et al., 1985).
(b) Thickness (cm) of recent sediments. Net post-glacial sediment
accumulation is exceedingly small.
Fig. 2. Loadings of Cs-137 to Lake St. Clair from the three principal
sources: inflow from Lake Huron, direct fallout and transfer from
the drainage basin. Loading is dominated by inflow from Lake Huron
which is a persistent low-level source.
Fig. 3. Vertical distributions of cesium-137 along an east-west transect
(Fig. 1). Activities are generally very low compared with open lake
depositional sites and distributions correspond only weakly to
loading history.
Fig. 4. (a) Cs-137 distributions in cores with distinctive stratigraphic
anomalies. The dashed line is the profile expected from direct
transfer of cesium to undisturbed sediments with peak positions
arbitrarily set.
(b) Distributions of quantities (see text) indicating bulk sediment
composition. A layer of old clay (dashed line) produces a valley in
the distribution of cesium-137.
330
-------
Fig. 5. (a-f). Distributions of excess lead-210 (top, log scale) and cesium-
137 (bottom) for 11 cores. Distributions of lead-210 are usually very
well characterized by the RSSM model (solid lines) showing intense
near surface mixing (ca 3 cm) and either sedimentation or uniform
mixing below. Distributions of Cs-137 are generally predicted well
by the RSSM model using mixing and sedimentation parameters set by
lead-210 (solid lines). When mixing and sedimentation parameters are
allowed to vary for Cs-137 the dashed line results. The anomalous
cores 38 and 69 give poor fits.
Fig. 6. Extent of penetration of Cs-137 (a, cm; b, g/cm2) into sediments.
Associated with fine-grained materials, Cs-137 has reached a maximum
depth of about 20 cm in roughly 25 years. The penetration depth
correlates with the thickness of recent sediments and bathymmetry.
Fig. 7. (a) Concentration of cesium-137 in surface sediments (0-1 cm) and (b)
total accumulation (dpm/cm**2). Anomalously high total Cs-137
(shaded region with >50 dpm/cm2 Cs-137) occurs in cores with marked
stratigraphic inhomogenieties.
Fig. 8. (a) Concentration of Cs-137 in surface sediments (0-2 cm) collected
by Shipex grab sampling in 1976. (b) Concentration in the 0-2 cm
interval calculated from profiles in cores collected in 1985. Only
19% of the decrease can be attributed to radioactive decay.
331
-------
Fig. 9. Ratio of Cs-137 activity in surface sediments (1985) to the activity
in 1976, decay-corrected to 1985. The ratio approaches unity toward
deposit margins where the net loss of the radionuclide is thus least.
Fig. 10. First order loss rate model calculations for a sediment contaminant
pool. (a) Fraction of Cs-137 and excess lead-210 retained in the
pool vs pool residence time. For both lead-210 and cesium-137 the
fraction retained is about 8% indicating pool residence times of 6
and 3 years respectively. Uncertainties in total excess lead
inventories are large. (b) Ratio of 1976 to 1985 concentrations of
Cs-137 in surface sediments vs pool residence time. The measured
value of 0.60 predicts a residence time of 16 years.
Fig. 11. Ratio of total Cs-137 (1985) to total excess lead-210 vs pool
residence time. The envelope reflects uncertainties in the value of
the partition coefficient for lead-210 used to calculate loadings of
the radionuclide from Lake Huron. The measured mean ratio of 0.69
predicts a pool residence time of 12 years.
Fig. 12. Seasonal variation in the concentration of Cs-137 in trap materials
collected at two sites in Lake St. Clair. Differences between sites
are ascribed to varying proportions of Cs-137 labeled particles from
two principal sources: resuspension of low activity sediments already
in the Lake and inflow of new, high activity material from Lake
Huron.
332
-------
Fig. 13. Distribution of chlorinated organic compounds in selected cores.
Significant concentrations of contaminants are confined to about the
upper 10 cm.
Fig. 14. Distribution of chlorinated organic compounds in cores for which
companion cores possess stratigraphic inhomogenieties.
Fig. 15. Concentrations of chlorinated organic compounds in surface (0-1 cm)
sediments. Concentrations are well correlated with the thickness of
recent sediments and associated with fine grained materials traced by
Cs-137.
333
-------
_AKE ST. CLAIR
Sediment Coring Sites (1985)
o \ -e
-------
LAKE ST. CLAIR
82°30'
Thickness of Recent Sediments (cm)
335
-------
Cs Loading to Lake St. Clair
Watershed
Watershed + Atmosphere
Watershed +Atmosphere
+Inflow from Lake Huron
1960 1970
Year
1980
-------
137.
Distribution of Cs on an East/West Transect
in Lake St. Clair (1985)
4 8 12 0
4 8 12 0 4 8 12 0
DEPTH (CM)
4 8 12
LSC-85 25
LSC-85 26
LSC-85 27
LSC-85 28
0 4 8 12 0 4 8 12 0 4 8 12 0 4 8 12
DEPTH (CM)
pi c\o^e- t) .
-------
u>
U>
00
'0 10 20 30
Depth (cm)
'0 10 20
Depth (cm)
Q.
\c\o^e,
-------
6
_ 5
t 4
Q.
3 3
J
2°- z
CVJ
^ 3
^
Q.
3 2
LSC-85 171
LSC-85 171
6
5
4
3
2
j LSC-85 181
10
15 0
Depth (cm)
jLSC-85 181
10
15
5".
339
-------
Depth (cm)
340
-------
0
10 15 W0 5
Depth (cm)
10
15
r
o C
-------
20
Depth (cm)
342
-------
5-
dLSC-8565
O.
-o
.a
Q.
0.5
fLSC-85 65
a.
(O
2
I
0.5
Q2
8
*• J LSC-86 69
10
15 0
20 30
Depth (cm)
t~ i 'v, o -f e.
343
-------
Depth (cm)
r t c\ o
344
-------
LAKE ST. CLAIR
LAKE ST. CLAIR
,
' 1 4 Ov/e 6 <*., b
-------
LAKE ST. CLAIR
Cs Surface (0-1 cm) Activity in I985(dpm/g)
\^o<«. 1
346
-------
_AKE ST. CLAIR
Total Cs Storage as of 1985 (dpm/cm )
< a -o r
.
-------
LAKE ST. CLAIR
l37Cs in Surface Grab Samples (0-2cm)
as of 1976 (dpm/g)
r i
348
-------
_AKE ST. CLAIR
l37Cs in Surface Grab Samples (0-2cm)
as of 1985 (dpm/g)
•: \ Q v
349
-------
_AKE ST. CLAIR
137
Cs Surface Activity Ratio: 1985/1976
-
.
350
-------
0.6
2 0.4
IT
c
o
I 0.2
III
0.0
o 0.6
0.2
o
O
0.0
rCs
0 10 20 30
Pool Residence Time (yr)
0
351
-------
O .0
Pool Residence Time
Uul
a;
F
352
-------
137
Cs in Lake St. Clair Trap Samples (1985)
_ 3
a.
SITE 10
SITE 13
May Jun. Jul. Aug. Sept. Oct. Nov.
Month
* -
i-
353
-------
LAKE ST. CLAIR _ HCBD
LAKE ST. CLAIR HCB
LAKE ST. CLAIR
LAKE ST. CLAIR ZDDT
LAKE ST. CLAIR
, /3
-------
Core 64
Core 71
Core 84
c
o
o
"c
o
c
o
o
10 '0 5
Depth in Core (cm)
n
<\
' »
355
-------
Core 38
Core 69
g
"o
o>
o
c.
o
o
TOTAL
DDT
Depth in Core (cm)
15"
356
-------
TOXICOKINETICS OF SELECTED XENOBIOTICS IN HEXAGENIA LIMBATA:
LABORATORY STUDIES AND SIMULATION MODEL
Peter F. Landrum and Ronald Poore
ABSTRACT
Understanding the role of benthos in the fate and transport of
toxicants requires understanding the toxicokinetics of those organisms for
both water borne and sediment associated compounds. This effort focused on
the toxicokinetics of Hexaeenia limbata as an important component of the
benthic community of the upper Great Lakes connecting channels. The
accumulation and loss of selected polycyclic aromatic hydrocarbon congeners
and a hexachlorobiphenyl congener were followed over the course of a season
inH. limbata collected from Lake St. Clair. Both the uptake and
elimination rate constants increased with increasing temperature through the
spring and summer. The elimination constant was relatively large compared
to other Great Lakes benthos. The uptake constant for sediment associated
compounds was essentially constant the two times it was measured. The
limits for the parameters to a deterministic simulation model were derived
from the laboratory kinetics constants. The model suggests that toxicant
concentrations in the organisms should decline as the temperature increases,
as a result of a greater increase in the elimination constant, and rise
again as the temperature declines. Based on the best estimates of
environmental concentration of the toxicants studied in both sediment and
357
-------
water, the model suggests that H. limbata should obtain greater than 90% of
its contaminant body burden from the sediment associated pollutants.
INTRODUCTION
Hexagenia limbata are an important food source for fish in the
interconnecting waterways of the Great Lakes (Hunt, 1958). Further,
g. limbata are extremely sensitive to pollution resulting in their
disappearance from some locations in bays and the connecting channels (Carr
and Hiltunen, 1965, Howmiller and Beeton, 1971, Schneider et al. , 1969,
Hiltunen and Schlosser, 1983). While the disappearance has been attributed
to low dissolved oxygen as a result of eutrophication, oil inputs into the
St. Mary's River below Sault Ste. Marie, MI, have eliminated the mayfly
larvae from stretches of the river that they previously inhabited (Hiltunen
and Schlosser, 1983). The mayfly larvae are beginning to repopulate areas
of the Great Lakes (Thornley, 1985). Since H. limbata lives and feeds in
the sediment, the mayfly larvae may mobilize contaminants from the sediments
up through the food chain. Because of their sensitivity to pollution and
their importance in the food web, I undertook to define the toxicokinetics
of these organisms for selected polycyclic aromatic hydrocarbons (PAH) and
polychlorinated biphenyl congeners (PCB) . The studies were run over the
course of a field season to determine the seasonal variability that occurs
in the toxicokinetics resulting from environmental variables such as
temperature and physiological variables such as changes in llpld content.
358
-------
MATERIALS AND METHODS
H_._ limbata were collected over the course of a field season beginning
in early May 1986 and continuing until November 1986 on approximately
monthly intervals. The collection site was near the middle of Lake St.
Clair, at Loran coordinates 49934 and 31128. The depth was approximately
6m. The animals were collected by ponar grab, gently screened from the
sediment and placed in a container of lake water and sediment for transport
back to the laboratory (generally about 2 h). The bottom temperature was
taken at the same time the animals were collected. The animals were
maintained in the laboratory in aerated (50 L) aquaria containing
approximately 3 cm lake sediment and 10 cm lake water at the temperature of
collection.
The compounds studied were 3H-benzo(a)pyrene (BaP)(specific activity
23.8 Ci/mmol, Amersham), l^C-phenanthrene (Phe)(specific activity 14
mCi/mmol, Pathfinder Laboratories) and 14C-2,4,5,2',4',5'-hexachlorobiphenyl
(HCB)(Specific activity 14.06 mCi/mmol, Pathfinder Laboratories). All
compounds were determined to be at least 98% radiopure prior to use by thin
layer chromatography (TLC) and liquid scintillation counting (LSC) . The
purity check was performed on silica gel plates (E Merck, 250 /rn) using
hexanerbenzene (8:2 V:V) as the solvent system. The plate was scraped at
the same Rf as a standard and several sections lower. These scrapings were
placed in scintillation vials and counted. The counts below the Rf of the
standard were assumed to be degradation products. All analytical procedures
359
-------
were performed under gold fluorescent lights (X > 500 ran) to minimize the
degradation of the PAH congeners.
Toxicokinetic studies were performed within one week of the time of
organism collection. The accumulation from water was performed as a
constant infusion, flow through, experiment. The temperature for each
experiment was the same as the environmental temperature measured at the
time of the collection (Table 1). Water, collected from Lake St. Clair at
the same time as the animals, was filtered prior to use through glass fiber
filters (Gelman AE) . The water was dosed in bulk (5 L) with the radiolabeled
compounds using a methanol carrier and allowed to equlibrate for 1 h prior
to starting the flow and introducing the organisms. Methanol as a carrier
is not expected to influence the kinetics (Landrum, 1983). The animals were
provided with artificial burrows to minimize the potential thigmotactic
stress (Henry et al. , 1986). The artificial burrows used in these studies
were made from stainless steel screen instead of glass tubes. Preliminary
studies with glass tubes indicated that depletion of compound concentration
within the tubes was occurring resulting in a very large variation in the
accunulation of compound by the organisms (Landrum, unpublished data) .
Organisms were removed after 1,2,4, and 6 h exposure, blotted dry, weighed
and placed in scintillation cocktail for radioanalysis. The remaining
organisms in the exposure chamber were removed and placed in uncontaminated
sediment for depuration. These animals were removed at approximately 1, 3,
5, 7 and 14 d, rinsed of sediment, blotted dry, weighted and placed in
scintillation cocktail for radioanalysis. Actual times were used for
kinetic determinations. The uptake and elimination rate constants were
360
-------
determined by fitting the data to a simple two compartment model using
initial rate assumptions for uptake and a simple first order decay model for
elimination (Landrum et al. , 1985) . The terms clearance constant and rate
constant to describe the uptake constants determined will be used
interchangeably throughout this paper. This uptake kinetics constant for
uptake actually describes the rate of clearance of the compound from water.
The accumulation from sediment was measured by sorbing the radiolabeled
compounds onto sediments in an aqueous slurry overnight. The sediment was
transferred to a large beaker and allowed to settle at 4°C for 24 h. The
overlying water was decanted and the sediment mixed and distributed to
individual beakers (approximately 75 g per beaker). The actual weights of
the sediment in each beaker were determined. Subsamples of the sediment
were taken for analysis of the concentration of radiolabeled compound and
for wet to dry weight of the sediment. Care was taken to insure that the
samples for the initial sediment concentration and for wet to dry weight
were taken at the beginning, middle and end of the distribution of material
for exposing the organisms to insure that there was no apparent bias in the
exposure concentrations. The beakers were placed in small aquaria, six per
aquaria, and overlying, Lake St. Glair, water added. The aquaria were
allowed to stand for an additional 24 h before adding the organisms. One iL.
limbata was added per beaker and screening was placed over the top of the
beaker until the organism had burrowed into the sediment. The burrowing was
almost immediate. Three organisms were removed for uptake, one from each of
three aquaria, at approximately 1, 3, 5, 7, and 14 d (actual times were used
for determination of kinetics). The animals were rinsed of sediment,
361
-------
blotted dry, weighed and placed in scintillation cocktail for radioanalysis.
Samples of sediment were taken for analysis of toxicant concentration and
for wet to dry weight of sediment. All sediment concentrations were based on
the dry weight of sediment. Accumulation kinetics were fit to a two
compartment model with a source function that could show exponential decay
to account. for changes in the concentration of the toxicant in the sediment
and for potential changes in the biological availability. This model had
previously been shown to be most appropriate for sediment accumulation of
PAH by Pontoporeia hoyi (Landrum, unpublished data).
- KsCse-At - KdCa (1)
Where:
Ca - concentration in the organism (ng g~^)
Cs - Concentration in the sediment (ng g~^)
Ks - Uptake rate constant from sediment (g sediment g"l animal h'l)
A - Rate constant for reduction in the bioavailability of the contaminant
in the sediment
K
-------
were obtained by allowing the samples to sit for 2 to 3 weeks after mixing
with the ethylacetate at the time of sampling and separating the solvent by
filtration. Subsequent soxhlet extraction after filtration did not yield
additional removal of radiolabeled material. Concentrations of the
toxicants were determined by radioanalysis of a known fraction of the
extracting solvent. The bulk of the extract was reduced in volume by a
combination of rotary evaporation and evaporation under a stream of nitrogen
to a final volume of approximately 500 pL. A portion of the extract was
chromatographed by TLC using hexane:benzene 8:2. The extent of degradation
was determined as described above for the pre-use purity check.
Organisms exposed through water were analyzed for biodegradation
products by extraction in ethylacetate:acetone 4:1 (2 X 20 mL) followed by
extraction with cyclohexane (1 X 20 mL) . The extraction was accomplished by
nacerating the animals in a tissue grinder with the extracting solvent. The
extracting solvents were combined and filtered from the residue of the
carcass. The solvent was reduced in volume and the biotransformation
determined as was done for the sediment extracts. The carcass residue was
analyzed for radioactivity as described for whole animals. The extent of
metabolism was determined as the amount of non-parent compound found by TLC
and the amount activity remaining in the residue.
Radiometric analyses were performed on an LKB 1217 scintillation
counter using the external standards ratio method. Additional measures made
over the course of the season included lipid content and wet to dry weight
aeasurements. The lipid content was determined with a microgravimetric
363
-------
method (Gardner e_t_al. , 1985a) . The wet to dry to ash free dry weight
measures were made by weighing animals wet within a few days after
collection. These animals were desiccated to a constant dry weight and
subsequently ashed at 500°C for one h and the ash weight determined. The
ashfree dry weight is the difference between the ash weight and the dry
weight.
Oxygen consumption and respiration rates were determined at the same
temperature as the kinetics experiments. One animal was placed in each of
five to ten 300 mL BOD bottles , the bottles capped and the oxygen
concentration determined after removing the animal. The duration of the
experiment was 24 h. The concentration in the water was compared to a
control containing the same water but no organisms. The oxygen consumption
normalized to biomass was determined by the Winkler titration using 0.005 M
thiosulfate (Grasshoff et al., 1976). Clearance constants for oxygen,
equivalent to the uptake clearance constants of the toxicants, were
calculated from oxygen consumption and the concentration of oxygen in the
control bottles.
A deterministic simulation model of the toxicokinetics for H. limbata
was developed to examine the predicted seasonality of body burdens. The
nodel used the general differential equation:
+ KSCS - KdCa (2)
364
-------
Where
-------
425 and 325 ng "1 respectively as determined in Lake Erie sediments (Eadie
et al., 1982).
RESULTS AND DISCUSSION
The dry to wet weight and ashfree to dry weight determinations were
made three times over the course of the season. The dry to wet weight ratio
was 0.18 ± 0.02 (mean ± sd, n-3) while the ash free dry to dry weight was
0.75 ± 0.10 ( mean ± sd, n-3). Each determination employed 5 to 10
organisms. The lipid content appeared to peak at the beginning of June
(Table 1) and ranged from about 4 to 6 % of dry weight from about August
through the fall (Table 1). The lipid content of H. limbata is low
compared to other Great Lakes invertebrates. The lipid content was most
comparable with the oligochaetes. The chironomid larvae, another insect
larvae in the Great Lakes, contain nearly twice the lipid content of JL.
limbata (Gardner et al., 1985b) . Oxygen consumption, respiration rates,
also appeared to peak in the summer and declined toward the fall (Table 1) .
Respiration rates in the spring and fall were significantly lower than rates
in the summer while respiration rates in the fall were significantly lower
than spring (p < 0.05, Students T test). The increase in oxygen
consumption in the summer, over those measurements in the spring and fall,
was probably due in part to increases in temperature. However, the low
oxygen consumption in October was at the highest environmental temperature
exaained. This low October value was probably due in part to the long hold
time in the laboratory before the measure was made (approximately 60 d)
366
-------
should not necessarily be considered representative for the time of year.
If the low value is incorrect then the oxygen consumption increases with
temperature as expected. These measures of oxygen consumption are similar
to those found in the literature for H. limbata (Zimmerman et al., 1975).
The toxicokinetics for water were measured on a monthly basis over the
course of the field season for 1986 (Table 2). For BaP at the same
temperature, the accumulation clearance constants from water for H. Limbata
were lower than those for M. relicta. a Great Lakes invertebrate of similar
size (Frez and Landrum, 1986). The elimination rate constants were greater
than those for either M. relicta or P. hovi although the study temperatures
for H. limbata were also greater. The relatively low lipid content of iL_
limbata probably accounts, in part, for the higher elimination rate
constants. The uptake from water generally seemed to rise in the summer and
decline again in the fall for all the compounds. The rise in uptake rate
constant for the HCB seemed to lag that of the PAH. The uptake rate
constants for HCB generally had larger variances than those for the PAH.
The reasons for this are not clear. The variances of all the accumulation
kinetics constants were generally greater than had been previously observed
for other Great Lakes invertebrates (Frez and Landrum, 1986, Frank et al.,
1986). The depuration rate constants also showed the same trend of an
increase in the summer and a decrease in the fall. These changes generally
tracked the changes observed in the temperature of collection, the
temperature at which the experiments were performed. Bioconcentration
factors (BCF) calculated from the kinetics in water showed a general trend
of higher BCF in the spring and summer for PAH while the BCF for HCB was
367
-------
essentially constant over the course of the season (Fig. 1). No
biotransformation was attributable to the organisms; although, some overall
degradation was observed, 4-5% of the total. The degradation did not
increase with time as is expected for PAH and has been observed with other
organisms (Leversee et al.. 1982). The degradation was not observed for
Phe.
The accumulation of neutral organic xenobiotics from water is assumed
to occur as passive diffusion across the respiratory membrane of the aquatic
organisms. A clearance rate for oxygen was determined from the oxygen
consumption experiments (Table 1) and can be compared with the uptake
clearance constants for the toxicants (Table 2). In most cases, the
toxicant was cleared from the water with a greater efficiency than the
oxygen. This increased efficiency suggests that the integument as well as
the respiratory membrane is a route for some of the uptake.
The sediments used in the studies had an organic carbon content of
6.98%. The size fractionation of the sediment indicated that 31.9% of the
sediment mass was <74 ^m diameter. This fine grained material had an
organic carbon content of 8.3%. The determination of the uptake rate
constant from sediment required the use of a model that could account for
changes in both the chemical and bioavailable concentration of the toxicant
in the sediment. The extractable concentration of the toxicants often
showed slight declines with time over the course of the studies. These
changes are accounted for with the A value in the kinetics model. The
accumulation from sediments was essentially constant the two times it was
368
-------
measured, (Table 2). This measurement was made only twice due to the time
required for performance of the experiments and the analysis of the samples
generated.
The assimilation efficiency from sediments can be estimated from
literature values of feeding rates (Zimmerman and Wissing, 1978) and the Ks
values. Our organisms were generally 20 mm in length or longer and Ks was
determined at 15 and 20°C. The feeding rates for the larger nymphs were
0.21 g sediment g'1 organisms h'1 at 15°C and 0.31 g g'1 h'1 at 20°C after
converting the feeding rates based on dry weight nymphs to a wet weight
basis and converting to hourly averages (Zimmerman and Wissing, 1978) .
These feeding rates have the same units as the Ks values. Thus, it is
possible to compare the feeding rates directly with the Ks values. Since
the Ks values are less than the feeding rate the implication is that all of
the contaminant is not being removed from the sediment as it passes through
the gut of the organisms. Therefore a ratio of Ks to feeding rates should
give the fraction of material assimilated or removed. With these feeding
rates the efficiency for assimilation for BaP ranged from 11.3 - 21%, Phe
from 13.6 - 31.3% and HCB from 14.4 - 29.1%. This estimated assimilation
efficiency is in the same range as that determined for obligochaetes for HCB
in an elegant dual labeled study (Klump et al., 1987).
Using the same estimates of water and sediment concentrations as used
for the simulation model, calculation of the amount of compound accumulated
from water versus sediment at steady state was determined by the following
equations
369
-------
Css - (KwCw + KSCS)/ Kd
Cw -
Cas - KsCs/Kd
Where Css is the concentration of the toxicant in the organism from both a
water and sediment source at steady state, Caw is the concentration in the
animal at steady state from water (ng g'1), and Cas is the concentration in
the animal at steady state from sediment (ng g'1). From these equations the
fraction from water can be computed from Caw/Css. Similarly, the fraction
from sediment would be computed from Cas/Css.
For the two times that the HCB accumulation from sediment was measured,
the fraction of the body burden from the water was estimated to be 0.1 and
0.03. Thus the fraction from sediment was 0.9 and 0.97. The route of HCB
accumulation for this organism is apparently via the sediment. Similar
comparisons for the PAH yielded ratios for BaP of 0.9 and 0.96 and for Phe
0.95 for both determinations as the fraction of toxicant obtained from the
sediments. This determination is very dependent on the ratio of the
products of KWCW and KSCS; therefore, changes in environmental
concentrations without any change in rate constants would alter the fraction
obtained from a particular source.
Comparing the estimated role of sediment as a source with other
organisms, H. limbata obtains a greater fraction of its body burden from the
370
-------
sediment. Using the same water and sediment concentrations as for the
projections for H. limbata and the kinetics constants for the oligochaete,
S. heringianus. (Frank et al., 1986) and P. hoyi (Landrum et al., 1985), the
oligochaetes would obtain 34 to 67% of their BaP body burden from sediment
while P. hoyi would only obtain 39% from sediments. This suggests that the
role of sediments as a source will depend on the organism as as well as the
sediment characteristics.
Using the environmental data available, the deterministic simulation
model suggests that the highest concentration of the toxicants studied will
occur in the winter and early spring. There is a consistent peak in the
body burden that occurs in the spring for all the compounds (Figs. 2-4) and
is driven primarily by a reduction in K^ resulting in a reduced flux out of
the organism (Fig. 5). The organism concentrations then decline during the
summer (Figs. 2-4). In general the flux of compound into the organism is
primarily from the product of KSCS and under the conditions used in the
simulation for HCB would be 2.7 ng g~l h"l. This flux from the sediment is
augmented by the flux in from the water, KWCW, and is reduced by the flux
out of the organism, K
-------
impact the seasonal changes in body burden but there was insufficient data
to justify a change with season. The bioaccumulation factor (BAF) (organism
concentration/sediment concentration) was based on sediment concentration
because it was the predominant source for the organisms. The BAF ranged
from about 4.5 to 15.5 for HCB and reflects the generally low elimination
rate constant while PAH showed a much lower BAF, Phe 0.9 - 2.2 and BaP 1.5 -
3.8. The change with season of the BAF values tracks the change in organism
concentration. The BAF predicted for H. limbata are higher than those found
for oligochaetes for HCB in the field (Smith et al., 1985) and BaP (Eadie et
al., 1982) but were about the same as oligochaetes for Phe (Eadie et al.,
1982. Comparing H. limbata to P. hovi the range of BAF's are about the same
for the two PAH studied (Eadie et al., 1985).
In conclusion, H. limbata exhibits seasonal changes in the
toxicokinetics and these changes are expected to result in changes in the
BCF and BAF. The organism appears to obtain the preponderance of its body
burden for the toxicants examined from the sediment based on the
calculations from the toxicokinetics and sediment and water data from the
literature.
372
-------
ACKNOWLEDGEMENTS
This work was jointly supported by the Great Lakes Environmental
Research Laboratory, NOAA and by the U. S. Environmental Protection Agency
through interagency agreement No. DW13931213-01-01. I wish to thank Tom
Fontaine for his critical discussions and suggestions on the kinetics
modeling. I also wish to thank Brian Eadie, Tom Nalepa, Mike Quigley and
Guy Stehly for their critical review of this manuscript.
373
-------
Table 1. Oxygen consumption and lipid content for Hexagenia in 1986
Date Oxygen Consumption Oxygen Clearance1 %Lipid Temperature
-l'1 mL '1 h'1 °C
62 mg-lh'1 mL g' h'
May 0.327 ± 0.111 23.3 ± 7.9 7.8 ± 1.9 10
n - 5 n = 7
June • Lost 15.1 ± 2.6 15
n = 5
July 0.667 ± 0.296 41.4 ± 18.3 9.1 ± 3.4 15
n = 5 ti = 8
4.3 ± 1.82
August 0.435 ± 0.10 61.9 ± 13.5 3.6 ± 1.0 20
n _ 5 n = 7
September 0.25 ± 0.06 44.5 ± 12 6.0 ± 2.4 20
n = 10 n = 6
6.9 ± 1.92
October3 0.16 ± 0.10 18.6 ± 11.8 3.7 ± 1.2 20
n = 4 n = 7
3.3 ± 0.92
November ND 6.0 ± 1.4 10
n = 4
1. The n for the clearance determination is the same as the oxygen consumption
determination.
2. Samples collected in 1985.
3. Sample actually collected on 30 September 1986, Oxygen consumption was run
60 d after collection.
ND - not determined
374
-------
Table 2. Seasonal Uptake and Elimination Rate Constants
for Hexaeenia limbata
Month
May Ku*
Kd'
**
June Ku
Kd
Ks
***
Benzo(a)pyrene Phenanthrene Hexachlorobiphenyl Temp.
68.5 ± 11.2
0.011 ± 0.003
67.0 ± 28.0
0.006 ± 0.002
0.043 ± 0.005
0.025 ± 0.0042
131.1 ± 46.8
0.032 ± 0.004
43.3 ± 12.0
0.0076 ± 0.0016
0.065 ± 0.016
47.5 ± 23.9
0.007 ± 0.001
44.2 ± 8.0
0.005 ± 0.002
0.030 ± 0.01
15
July Ku
Kd
Aug Ku
Kd
Ks
Sept Ku
Kd
Sept3Ku
Kd
Nov Ku
Kd
101.9 ± 32.6
0.013 ± 0.002
65.1 ± 29.1
lost
0.035 ± 0.005
149.5 ± 29.0
0.016 ± 0.003
76.3 ± 41.0
0.028 ± 0.001
40.9 ± 30.6
0.010 ± 0.001
57.5 ± 5.0
0.029 ± 0.002
11.9 ± 4.0
lost
0.042 ± 0.008
56.3 ± 6.8
0.032 ± 0.004
33.0 ± 8.0
0.067 ± 0.008
34.2 ± 7.2
0.026 ± 0.002
40.8 ± 37.3
0.005 ± 0.001
40.8 ± 37.3
0.007 ± 0.001
0.09 ± 0.02
128.7 ± 20.3
0.015 ± 0.003
95.0 ± 17.3
0.017 ± 0.002
45.5 ± 16.1
0.004 ± 0.0006
15
20
20
20
10
* Ku has been corrected for sorption to dissolved organic carbon and has
units of mL g"*- h~l.
** Kd has units of h'l.
*** Ks has units of g dry sediment g'l animal h'l
1. Temperature is in degrees centigrade.
2. Uptake from sediment was measured twice for BaP
3. This collection was actually made on September 30, 1986.
375
-------
LITERATURE CITED
Carr, J. F. and J. K. Hiltunen. 1965. Changes in the bottom fauna of western
Lake Erie from 1930 to 1961. Limnol. Oceanogr. 10:551-569.
Eadie, B. J.. W. Faust, W. S. Gardner and T. Nalepa. 1982. Polycyclic
aromatic hydrocarbons in sediments and associated benthos in Lake Erie.
Chemosphere 11:185-191.
Eadie, B. J., W. R. Faust, P. F. Landrum, N. R. Morehead, W. S. Gardner and
T. Nalepa. 1983. Bioconcentrations of PAH by some benthic organisms of
the Great Lakes. In: Polynuclear Aromatic Hydrocarbons: Seventh
International Symposium on Formation. Metabolism and Measurement. M.
W. Cooke and A. J. Dennis, Eds. Battelle Press, Columbus, OH. pp. 437-
449.
Eadie, B. J., W. R. Faust, P. F. Landrum and N. R. Morehead. 1985. Factors
affecting bioconcentration of PAH by the dominant benthic organisms of
the Great Lakes. In: Polynuclear Aromatic Hydrocarbons: Eighth
International Symposium on Mechanisms. Methods and Metabolism. M. W.
Cooke and A. J. Dennis, Eds. Battelle Press, Columbus, OH. pp. 363-377.
Frank, A. P., P. F. Landrum and B. J. Eadie. 1986. Polycyclic aromatic
hydrocarbon rates of uptake, depuration, and biotransformation by Lake
Michigan Stylodrilius heringianus. Chemosphere 15:317-330.
376
-------
Frez, W. A. and P. F. Landrum. 1986. Species-dependent uptake of PAH in
Great Lakes invertebrates. In: Polvnuclear aromatic hydrocarbons:
Ninth International Symposium on Chemistry. Characterization and
Carcinoeenesis. M. W. Cooke and A. J. Dennis, Eds. Battelle Press,
Columbus, OH. pp. 291-304.
Gardner, W. S., W. A. Frez, E. A. Cichocki and C. C. Parish. 1985a.
Micromethod for lipid analysis in aquatic invertebrates. Limnol.
Oceanogr. 30:1099-1105.
Gardner, W. S., T. F. Nalepa, W. A. Frez, E. A. Cichocki and P. F. Landrum.
1985b. Seasonal patterns in lipid content of Lake Michigan
macroinvertebrates. Can. J. Fish. Aquat. Sci. 42:1827-1832.
Grasshoff, K., M. Ehrhardt and K. Kremling. 1976. Methods of Seavater
Analysis. Second Edition, Verlag Chemie, Federal Republic of Germany,
pp. 419.
Henry, M. G., D. N. Chester and W. L. Mauck. 1986. Role of artificial
burrows in Hexagenia toxicity tests: Recommendations for protocol
development. Environm. Toxicol. Chenu 5:553-559.
Hiltunen, J. K. and D. W. Schlosser. 1983. The occurrence of oil and the
distribution of Hexagenia nymphs in the St. Mary's River, Michigan and
Ontario. Freshwat. Invertebr. Biol. 2:199-203.
377
-------
Howmiller, R. P. and A. M. Beeton. 1971. Biological evaluation of
environmental quality, Green Bay, Lake Michigan. J. Wat. Pollut. Cont.
Fed. 123-133.
Hunt, B. P. 1958. The life history and economic importance of a burrowing
mayfly, Hexagenia limbata. in southern Michigan lakes. Bulletin of the
Institute of Fisheries Research. No. 4. Michigan Department of
Conservation, Lansing, MI pp 151.
Klump, J. V., J. R. Krezoski, M. E. Smith and J. L. Kaster. 1987. Dual
tracer studies of the assimilation of an organic contaminant from
sediments by deposit feeding oligochaetes. Can. J. Fish. Aquat. Sri
44:(In press).
Landrum, P. F. 1983. The effect of co-contaminants on the bioavailability of
polycyclic aromatic hydrocarbons to Pontoporeia hoyi. Polvnuclear
Aromatic Hydrocarbons: Seventh International Symposium on Formation.
Metabolism and Measurement. M. W. Cooke and A. J. Dennis eds. Battelle
Press, Columbus, OH. pp. 731-743.
Landrum, P. F. , B. J. Eadie, W. R. Faust, N. R. Morehead and M. J.
McCormick. 1985. Role of sediment in the bioaccumulation of
benzo(a)pyrene by the amphipod, Pontoporeia hovi. Polvnuclear Aromatic
Hydrocarbons: Eighth International Symposium on Mechanisms. Methods and
Metabolism. M. W. Cooke and A. J. Dennis Eds. Battelle Press, Columbus,
OH. pp. 799-812.
378
-------
Leversee, G. J.. J. P. Giesy, P. F. Landrum, S. Gerould, J. W. Bowling, T.
E. Fannin, J. D. Haddock and S. M. Bartell. 1982. Kinetics and
biotransformation of benzo(a)pyrene in Chironomus riparius. Arch.
Environ. Contain. Toxicol. 11:25-31.
Schneider, J. C., F. F. Hooper and A. M. Beeton. 1969. The distribution and
abundance of benthic fauna in Saginaw Bay, Lake Huron. Proc. 12th Conf.
Great Lakes Res. International Assoc. Great Lakes Res, pp. 80-90.
Smith, E. V., J. M. Spurr, J. C. Filkins and J. J. Jones. 1985.
Organochlorine contaminants of wintering ducks foraging on Detroit
River sediments. J. Great Lakes Res. 11:247-265.
Thornley, S. 1985. Macrozoobenthos of the Detroit and St. Clair rivers with
comparisons to neighboring waters. J. Great Lakes Res. 11:290-296.
Zimmerman, M. C., T. E. Wissing and R. P. Rutter. 1975. Bioenergetics of the
burrowing mayfly, Hexagenia limbata. in a pond ecosystem. Verh.
Internat. Verein. Limnol. 19:3039-3049.
Zimmerman, M. C. and T. E. Wissing. 1978. Effects of temperature on gut-
loading and gut clearing times of the burrowing may fly, Hexagenia
limbata. Freshwat. Biol. 8:269-277.
379
-------
LIST OF FIGURES
Fig. 1. Calculated bioconcentration factors in Hexagenia limbata over the
1986 field season calculated from the ratio of
Fig. 2. Simulation of the concentration of HCB in H. limbata through one
season based on the concentrations of HCB in sediments and water
found in the Detroit River..
Fig. 3. Simulation of the concentration of BaP in H. limbata through one
season based on the concentrations of BaP in Lake Erie sediments
and water concentrations found in the Great Lakes .
Fig. 4. Simulation of the concentration of Phe in H. limbata through one
season based on the concentrations of BaP in Lake Erie sediments
and water concentrations found in the Great Lakes .
Fig. 5. Simulated flux of HCB into the organism from water and out of the
organism through elimination.
380
-------
r'*
OJ
00
*_
+-»
o
03
LL
centration
c
o
O
O
in
•
H
^
^
\c. \J\J\J
10 000-
r
2T500-
• ^
IBenzo(a)nvrene**^
!*•/ vi i^uv/yd/|^Jr i wl l W|
Hexochlorobiphenylj ^
* .A A
*i* • Mft
^Jfr ^F
PhenanthrerieK •
^^^ ' •
^i i^ 5fe
ii n1' or1 -in'1 o ' -i ' nr ' -*-• > r« I '-±"1 ' 1 ^
Jl Fl Ml Al MI J1 IJ 'A 'S 'D TN fD
•»
••
••
Time (months)
-------
1000:
Lo
03
KJ
O)
O)
O
c
O
O
m
o
800-
6QQL
400^
200^
0!
—I I , , ,
601 120 160 i24Q] 300 £60
Time (Julian days)
-------
CD
2800;
0}
Q!
1?_Q 180 I24Q
Time (Julian days)
300
1360
-------
2400
to
oo
•C-
12Q1 1SQ: I240
Time Gulian days)
1300
!360
-------
U)
oo
Ol
O)
O)
CD
O
I
H—
O
X
E
1-
OH
=2r
~3r
0!
Water Uptake
Depuration
60i r120i__ J80 |240 300
Time (jtiilian days!
-------
MODELING THE FATE AND TRANSPORT OF CONTAMINANTS IN LAKE ST. CLAIR
Gregory A. Lang and Thomas D. Fontaine, III
INTRODUCTION
When looking at the Great Lakes System as a whole, two aspects of Lake
St. Clair stand out: 1) its average depth is only about 3 m (the next most
shallow lake is Lake Erie at 19 m) and 2) its theoretical hydraulic
retention time is only about 9 days (again, the next closest is Lake Erie at
about 3 years). Lake St. Clair has been described as simply a "wide part of
the river" that extends from mouth of Lake Huron to the head of Lake Erie,
and past Great Lakes budget calculations have generally overlooked the
dynamics of nutrient and contaminant transport into and through Lake St.
Clair. Sediment studies of Frank et. al. (1977), Pugsley et. al. (1985),
and Oliver and Bourbonniere (1985) document the levels of organic
contaminants, including polychlorinated biphenyls and octachlorostyrene, in
the surficial sediments of Lake St. Clair and the St. Clair and Detroit
Rivers. The concentrated chemical plumes located near the center of Lake
St. Clair, between the South Channel and the Detroit River, indicates the
ability the lake's sediments to trap particle-bound contaminants.
Recent attention to the connecting channels of the upper Great Lakes
and the recognized harmful effects to biota and the extreme environmental
386
-------
persistence of many of these hydrophobic chemicals have served as the
impetus for studying the dynamics of contaminant transport within Lake St.
Clair. The objectives of this study were two fold: 1) to develop a multi-
segment mass balance model to simulate contaminant fate and transport in
Lake St. Clair, based on the Environmental Protection Agency's model
TOXIWASP; and 2) to calibrate, verify, and apply the model using available
contaminant and tracer data. This report presents the model, briefly
describes the model processes affecting chemical and solids concentrations,
details the physical lake characteristics and model segmentation, describes
the tracer and contaminant input data, and presents the simulation results.
MODELING APPROACH
The mathematical model used in this analysis was based on the
Environmental Protection Agency's Chemical Transport and Fate Model TOXIWASP
(Ambrose et. al. 1983). TOXIWASP combines the kinetic formulations in the
Exposure Analysis Modeling Simulation, EXAMS (Burns et. al. 1982), and the
transport processes in the Water Analysis Simulation Program, WASP (DiToro
et. al. 1983). The mechanics of TOXIWASP were left relatively unchanged.
However, in an attempt to streamline the model, improve its execution time,
and to make it more specific to Lake St. Glair's physical and contaminant
data, some modifications were made. Numerous programming errors found in
the source code (in particular, subroutine SETTLE) were corrected. The EPA-
Athens modeling group was informed of all corrections needed.
387
-------
Two state variables are included in TOXIWASP, total organic chemical
and total sediment. Sediment concentrations are affected by advection,
dispersion, mass loading, settling, and resuspension. Chemical
concentrations are affected by these same processes, plus degradation,
sediment-water diffusion, and biological mixing to deep sediments. Chemical
degradation is due to hydrolysis, biolysis, photolysis, oxidation, and
volatilization. Sorption onto sediments and biota is calculated via
equilibrium kinetics using a chemical-specific partition coefficient and
spatially-varying environmental organic carbon fractions. These transport
and transformation processes are detailed fully in the EPA TOXIWASP Manual
(Ambrose et. al. 1983), EPA EXAMS Manual (Burns et. al 1982), and the EPA
WASP Manual (DiToro et. al. 1983).
Physical Lake Characteristics
The volume and surface area of Lake St. Clair were taken to be 3.75 km3
and 1065.7 km2, respectively. The study area was segmented into 126 well-
mixed segments: 42 water segments, 42 active sediment layer segments, and 42
deep sediment layer segments (Fig. 1). The segment numbers in Figure 1
correspond to water column segments; active layer and deep layer segment
numbers are determined by adding one and two, respectively, to the water
column numbers. The average segment size was about 5 km on a side. The
segmentation scheme was based on the results of two contributing projects.
One was an extensive cluster analysis performed on the available chemical,
physical, biological, and contaminant data in the water column and sediments
of Lake St. Clair (Rybczyk 1986). The other was the 1.2 km grid Lake St.
388
-------
Clair Rigid Lid Hydrodynamic Model (Schwab and Liu 1987) which generated the
wind-driven flow fields used in the present study to transport solids and
chemical throughout the water column. The 126-segment grid and the 1.2 km
grid were superimposed such that segment interfaces were shared. The
hydrodynamic model flows were summed along the larger TOXIWASP segment
interfaces yielding 75 net interface flows.
The volumetric inflow and outflow rate was constant at 5700 cms. The
depth of the water column segments ranged from 2 m to 5.2 m (areal average =
3.5 m) . The depth of the active sediment layer was uniform at 2 or 10 cm,
depending on the simulation. The depth of the deep sediment layer was equal
to the total depth of the "recent sediments" minus the active layer depth.
The "recent sediments" are defined by Robbins and Oliver (1987) as those
sediments deposited on top of the post-glacial till and range from 3-30 cm.
The rate of horizontal dispersion was assumed to be IxlO5 cm2/s (Great Lakes
Institute 1986). Pore water diffusion was set to IxlO'5 cm2/s.
Solids
Suspended solids from tributary sources were loaded at a constant rate
of 2700 MT/day into 8 water column segments. The St. Clair River load
entered segments 16, 25, 37, 49, and 55 (Fig. 1), the Sydenham River load
entered segment 58; the Clinton River load entered segment 19, and the
Thames River load entered segment 124. The loads were based on average
tributary concentrations and flows reported by the Great Lakes Institute
(1986) . The settling velocity of solids was set equal to 3 m/d, based on
389
-------
results of an empirical modeling procedure for Lake St. Clair (Simons and
Schertzer 1986). The resuspension velocity in each segment was calculated
such that the particulate settling and resuspension fluxes across each
segment's sediment/water interface were equal (i.e., the net sedimentation
rate was assumed zero) . The assumption of zero net sedimentation is based
on the relatively shallow depth of the recent, post-glacial sediments
(Robbins and Oliver 1987).
Under constant wind (and thus constant flow) conditions, the solids
concentration in each water column segment could be calculated from the
advective flux of solids to and from each segment, the dispersive solids
flux between segments, the rate of solids loading into each segment, and the
assumption of zero net sedimentation. The suspended solids concentrations
resulting from a steady 6 m/s wind from the southwest ranged from 4.7-10.5
mg/1 (volumetric average - 6.33 mg/1). These values are consistent with
those measured by Bukata et. al. (1987) along ship transects during three
separate cruises in Lake St. Clair in September, 1985. They reported a
range of 2.5-16 mg/1, with a mean of approximately 5 mg/1. The
concentration of solids in the active and deep layer model sediments was set
to 1.2xl06 mg/1, based on the average of 19 10-cm sediment cores from 1985
(Robbins 1986, Great Lakes Environmental Research Laboratory, personal
communication). The calculated resuspension rates ranged from 0.42-0.95
cm/yr.
The model does not distinguish between different types of sediment, a
particular concern when simultaneously modeling nearshore and offshore
390
-------
concentrations of organic, hydrophobic contaminants. However, it does
provide a parameter for defining the spatially-varying organic carbon
content of the sediment. It was assumed that the organic carbon content in
each segment would not vary over time; i.e., particulate matter resuspended
from the active layer would immediately be available for transport to
another segment, but the average organic carbon content of each segment
would be maintained over time.
TOXIWASP requires the user to quantify certain model parameters and
constants; parameters vary over space but not time, constants do not vary.
Ranges and values of Lake St. Clair parameters and constants used in the
present study are presented in Table 1. Constants associated with
particular contaminants are presented in separate tables.
Chemical
The chemical being modeled was loaded into Lake St. Clair from
atmospheric, land runoff, and/or tributary sources. A whole-lake areal
atmospheric load was calculated from available flux measurements and then
segmented according to each segment's surface area. The tributary chemical
loads entered the same segments as the solids loads (segments 16, 19, 25,
37, 49, 55, 58, and 124); although, not necessarily all eight. For example,
octachlorostyrene, thought to originate mainly from the Sarnia area, entered
the lake through segments 37 and 49 only. The total load from land runoff
sources was partitioned to all water segments bordering the shore according
to each segment's drainage area.
391
-------
The model assumed a local equilibrium between the dissolved, sorbed,
and bio-sorbed chemical as defined by the organic carbon content of
particles and octanol-water partition coefficients, KQC and KOW. KOW, held
constant throughout time and space, was multiplied by the varying organic
carbon content of the sediment, resulting in a spatial distribution of the
solid/water partition coefficient, Kp, that corresponded closely to the
distribution of fine-grained, organic-rich sediment. KOW was multiplied by
the organic carbon content of the biomass to yield an overall biota/water
partition coefficient, KD.
Vertical distributions of excess lead-210, measured from a set of
diver-collected sediment cores from Lake St. Clair in 1985, possessed a zone
of constant activity extending down to about 3 cm with exponential fall off
below (Robbins and Oliver 1987). Analyses of these distributions plus those
of cesium-137 indicate sedimentation rates on the order of 0.1-0.2 cm/yr.
Robbins and Oliver state that because of the evidence of extensive
biological activity down to at least 15 cm, biological mixing, and not net
particulate settling, is favored as the mechanism producing observed
vertical radionuclide and contaminant sediment distributions. Because there
is no mechanism for biological mixing in TOXIWASP, the model was modified to
account for constant, spatially-varying, burial of chemical from the active
sediment layer to the deep sediments. This rate was set to 0.1 cm/yr
throughout the lake.
392
-------
SIMULATIONS
A series of simulations were performed during this study for purposes
of model calibration, verification, and application. These include
simulations of the conservative ion chloride, the radionuclide cesium-137,
and the organic contaminants octachlorostyrene and polychlorinated
biphenyls. Descriptions of simulation conditions, available chemical
observations, loading functions, and chemical constants required for these
simulations are presented here; the results are presented in a later
section.
Chloride
Chloride, a conservative ion, was chosen to test the efficacy of the
transport mechanisms in TOXIWASP. Bell (1980) documents chloride and
meteorological data collected during a series of cruises in Lake St. Clair
during the summer of 1974. Cruises 3 (19-29 June), 5 (15-24 July), 6 (5-15
August), and 8 (16-25 September) were the most complete and were thus
selected for this study. At least 30 stations were monitored during each of
the 4 cruises. An average chloride load to the lake was calculated for each
cruise as the product of tributary flows and chloride concentrations. The
flow fields during each cruise were generated with the hydrodynamic model
using the cruise-averaged wind conditions (speed and direction). The model
flows were multiplied by a factor of 0.93 to account for the difference in
total inflow between the 1974 value (5300 cms) and that used in the
393
-------
hydrodynamic model (5700 cms). The horizontal dispersion coefficient was
set to 1x10^ cm^/s for all segment interfaces. Chloride is a dissolved
conservative substance; thus partitioning, settling, resuspension, and
degradation were not included.
Four steady-state simulations were performed, one corresponding to each
cruise; each with a different chloride load and wind-induced flow field.
The average wind conditions for cruises 3, 5, 6, and 8 were 4 m/s from the
northeast, 5 m/s from the north, 5 m/s from the east, and 6 m/s from the
west, respectively. The total chloride loads to Lake St. Glair during the
four cruise periods were 3.7xl06 kg/day, 3.4xl06 kg/day, 3.2xl06 kg/day, and
3.3x10^ kg/day, respectively. Eight additional simulations were generated
for each cruise; one for each wind direction other than the observed
direction and one for the no wind case.
Cesium-137
Cesium-137, a surrogate of many hydrophobic, organic contaminants, was
chosen to calibrate and verify the processes associated with sediment-bound
contaminant movement. A 35-year history (1950-1985) of Cs-137 loading to
Lake St. Clair (Figure 2) was extracted from Robbins and Oliver (1987).
During this period, the largest portion of loading has been inflow from Lake
Huron. A nearly equal contribution (about 40% of the total load) originated
from the atmosphere during the 1950s and early 1960s. Lake Huron presently
contributes about 75% of the total load, while the atmosphere contributes
less than 20%. The remainder of the load (about 5%) originates from land
394
-------
runoff sources. Observed cesium concentrations in the upper 2 cm of bottom
sediment were available for 1976 and 1985 (Robbins and Oliver 1987). In
both years, concentrations were closely related to the thickness of the
recent sediments, with the highest concentrations located in the corridor
between the mouth of the South Channel of the St. Clair River and the head
of the Detroit River. In addition, Robbins and Oliver estimated the total
accumulation of cesium-137 over the entire depth of sediments in 1985.
Cesium-137 has a half life of 30.2 years and a partition coefficient,
Kp, of about 29000 lw/kg. The partition coefficient was calculated from the
lake-averaged suspended solids concentration in Lake St. Clair and the
fraction of dissolved contaminant (0.847), assumed constant for all of the
Great Lakes (Robbins 1985). With an average organic carbon content of the
open-lake sediments equal to about 2.5%, Koc is calculated to be 1.16x!06
lw/kg- The active sediment depth was set to 2 cm to allow direct comparison
with the available 1976 and 1985 data.
The model was run for 9700 days, corresponding to the period 1950 to
mid 1976, with a constant flow field generated by 6 m/s wind from the
southwest. Wind frequency data from meteorological sampling stations in the
St. Clair Region during the period 1951-1980 show a predominantly moderate
wind speed (5-8 m/s) with a southwesterly flow in all seasons (Great Lakes
Institute 1986). Initial 1950 conditions of cesium-137 were set to zero in
all segments. The organic carbon content of each segment was adjusted
within the range of O.OX to 5% until the model results matched the 1976
cesium data. This calibration exercise was verified by running the model
395
-------
for an additional 3300 days (9 years) and comparing the results to the
observed 1985 data.
Octachlorostyrene (PCS)
DCS was first reported in the lower Great Lakes in 1980 (Kuel et. al.
1980); however, information regarding its sources and environmental effects
is limited. The only documented and recognized regional source of OCS, a
by-product of several industrial processes including chlorine and solvent
production, appears to be chemical companies in the Sarnia area (Great Lakes
Institute 1986, Oliver and Bourbonniere 1985). OCS concentrations in
sediments from the St. Clair River indicate that the source is located along
the eastern shore of the St. Clair River, possibly from drainage of the
Scott Road landfill and Dow Chemical's First Street Sewer discharge (Oliver
1987). Little historical data exists, however, and the magnitude of total
loading to Lake St. Clair is unknown. Limno-Tech, Inc. (1985) speculated
that OCS was introduced to the lower Great Lakes beginning in the 1970s as
industries manufacturing chlorine converted from a process using mercury to
one resulting in OCS as a by-product. Concern exists because OCS tends to
bioconcentrate and is a chlorinated hydrocarbon. The simulation of OCS in
Lake St. Clair is intended as an application of the model.
Pugsley et. al. (1985) presented the distribution of OCS levels in
surficial (0-10 cm), diver-collected sediments for 1983. The values ranged
from non-detectable to 26.2 pgAg- The distribution of OCS showed a
concentrated plume extending from the mouth of the South Channel towards the
396
-------
the head of the Detroit River, very similar to the pattern of cesium-137 in
the surficial sediments. Great Lakes Institute (1986) estimated the total
load of OCS to Lake St. Clair to be about 1.9 Ibs/day, based solely on model
results in which the load was adjusted until predicted sediment
concentrations matched the data. In the present study, the load was assumed
to enter the lake solely through the South Channel outflow because the
source of OCS is believed to be located along the eastern shore of the St.
Clair River.
The model was run for 5000 days with a constant flow field
corresponding to a 6 m/s wind from the southwest. Initial conditions of OCS
were set to zero in all segments based on the assumption that background
levels of OCS in Lake St. Clair prior to the 1970s were negligible. The
active layer depth was set to 10 cm for direct comparison with the data.
The chemical constants used in the OCS simulation, taken from Ibrahim
(1986), are presented in Table 2 (e.g., KOW, Koc, molecular weight, etc.).
Polychlorinated Biphenvls (PCBs)
As a final application, the model was used to simulate the fate and
distribution of total PCBs in Lake St. Clair. PCBs were first prepared in
1881 and had been manufactured and extensively used since 1930 for
industrial purposes where extreme thermodynamic conditions exist such as
dielectric fluids, heat transfer agents, and flame retardants (Limno-Tech,
Inc. 1985). They are formed by the chlorination of biphenyl in the presence
of an iron catalyst. PCBs are relatively nonvolatile, insoluble in water,
397
-------
soluble in organic compounds, have high dielectric constants, are relatively
inert towards acids, alkalies, and other corrosive chemicals, and are stable
towards oxidation (Roberts et. al. 1978). They are of particular concern
because of their known toxicity to biota, low solubility in water,
bioaccumulation, and extreme environmental persistence.
Frank et. al. (1977) measured PCB concentrations of surficial sediments
(0-2 cm) collected from Lake St. Clair in 1970 and 1974. They noted that
the distribution of PCB in 1970 showed a concentrated plume entering from
the St. Clair and Thames Rivers. Great Lakes Institute (1986) estimated the
annual load of total PCB to the lake to be 1860 Ibs (5.1 Ibs/day), based on
model simulations in which the load was adjusted to reach the best
comparison between model predictions and measured values in 1974. This
value is comparable to 1-4 Ibs/day, a Lake St. Clair PCB load estimate
calculated from Lake Huron outflow concentrations and average atmospheric
and tributary loading rates for Lakes Erie and Huron during the late 1970s
(Thomann and Mueller 1983). It was assumed that 18Z of the total load
entered through Thames River, 16% through the Clinton River, and 66% through
the St. Clair River. The atmospheric input of PCB was assumed negligible
based on an atmospheric loading rate of l.OxlO"10 Ibs/m2/day which is the
average value for Lakes Erie and Huron (Thomann and Mueller 1983).
The model was run for 4 years with a constant flow field corresponding
to a 6 m/s wind from the southwest. Initial conditions of PCB were set
equal to the 1970 values (Figure 3) . The active layer was set to 2 cm for
direct comparison with the data. The chemical constants used in the PCB
simulation, taken from Mabey et. al. (1982), are presented in Table 3.
398
-------
RESULTS AND DISCUSSION
Chloride
The magnitude and distribution of simulated chloride concentrations in
the water column of Lake St. Clair agree very well with the corresponding
data for cruises 3, 6, and 8. Figures 4 and 5 compare the observed
concentrations (left frame) and the predicted concentrations (right frame)
for all four cruises. The model did especially well in predicting the
chloride concentrations in Anchor Bay and in locating the 7, 7.5, and 8 mg/1
contour lines. Although it did quite well overall, the simulation
corresponding to cruise 3 (Figure 4a) tended to overpredict the
concentrations along the northeast and east shores of the main lake (from
Figure 1, segments 55, 58, 61, 79, 82, and 103). The simulation
corresponding to cruise 5 (Figure 4b) overpredicted the entire eastern half
of the main lake. This condition will be addressed in the next paragraph.
The simulation corresponding to cruise 6 (Figure 5a) slightly underpredicted
the values along the eastern shore between Anchor Bay and the main lake
(segments 25, 34, 37, 46, and 49). The simulation corresponding to cruise 8
(Figure 5b) was probably the best of the four, although it slightly
underpredicted the concentrations along the northwestern shore of the main
lake (segments 28, 40, and 64).
Examining the simulations generated using wind directions other than
the observed directions and the no wind case revealed that simulations using
the observed winds for cruises 3, 6, and 8 were unique solutions. They more
399
-------
closely matched the data than simulations using any other wind direction and
the no wind case. The simulated chloride concentrations for cruise 5 did
not agree well with the data using the observed average wind speed of 5 m/s
from the north (Figure 4b) . However, simulations generated using winds from
the NE, E, and SE agree much better with the data. Looking more closely at
the observed wind directions from cruise 5; 13 stations measured winds from
the north, 7 from the NE, 3 from the E, 4 from the SE, 1 from the S, zero
from the SW, 2 from the W, and 3 from the NW. Therefore, it is not
inconceivable that the NW, E, and SE simulations agree well with the data
since 14 of the 33 stations reported winds from these three directions.
Cruise 5 results may point out the importance of easterly winds in flushing
dissolved compounds from the relatively low flowing eastern portion of Lake
St. Clair.
In general, considering that the wind conditions were averaged over
10-11 days and over at least 30 stations, the distribution of simulated
chloride concentrations in Lake St. Clair were quite close the observed
data. The results of the chloride simulations tend to confirm the accuracy
of the wind-induced transport mechanisms used in the Lake St. Clair
Contaminant Fate and Transport Model.
Cesium-137
Model-simulated 1976 cesium-137 concentrations in surficial (0-2 cm)
sediments compare well with observed 1976 values for Lake St. Clair (Figure
400
-------
6a). Through calibration of the organic carbon content of the sediments,
the model was able to match the magnitude and distribution sediment-bound
contaminant in the active layer throughout the lake. The calibrated organic
carbon content values closely matched the observed distribution of fine-
grained surficial (0-10 cm) sediments in 1983 and 1984 and the observed
distribution of organic carbon content of surficial (0-10 cm), SCUBA diver
collected sediments in 1983 (Great Lakes Institute 1986). However, the
magnitude of the calibrated values ranged from 0.14X to 5X (areal mean -
1.27X), while the observed values ranged 0.06X to 1.9Z (mean - 0.58X).
Kaiser et. al. (1985) and Maguire et. al. (1985) reported values of percent
organic carbon in Detroit River sediments which were also consistently
higher than those reported in the Great Lakes Institute report. Ibrahim
(1986) and Great Lakes Institute (1986) used values of 5X for their model
simulations.
The calibrated organic carbon contents, combined with the organic
carbon partition coefficient, Koc, and the suspended solids concentrations,
yielded water column dissolved Cs-137 fractions of 0.77-0.99; the lower
values located in the zones of highest deposition (i.e. fined-grained
sediment, rich in organic carbon). The average value for the open-lake
segments was 0.84 (n-10). Robbins (1985) reported a constant value of the
fraction of dissolved Cs-137 equal to 0.847 for the open-waters of each of
the Great Lakes.
Without any further calibration, the model was run for an additional
3300 days, corresponding to 1985. The predicted magnitude and distribution
401
-------
of sediment-bound Cs-137 in the active layer compare well with observed 1985
values (Figure 6b). Cs-137 was concentrated in the sediments along the
South Channel-Detroit River corridor in 1976 and 1985, closely matching the
distribution of fine-grained, organic-rich sediments in the depositional
zones of Lake St. Clair. Modeled and observed maximum cesium-137
concentrations in the active sediment layer declined during this time period
from about 5 dpm/g to about 2 dpm/g due to decreased loading, radioactive
decay, particle resuspension, and burial to the deep sediments. In
addition, the model predicted the total lake-wide accumulation of Cs-137 in
Lake St. Clair sediments to be 41 Ci in 1985 (corresponding to an average of
8.6 dpm/cm2), which agrees reasonably well with the measured 1985 value of
37 Ci (Robbins and Oliver 1987).
During the 35 years of simulation, the model predicted a total Cs-137
loading of 800 Ci. Of the total load, 720 Ci or 90X, exited through the
Detroit River, 38.8 Ci or 5X, were lost due to radioactive decay, and 41.5
Ci or 5X, remained in the system. The calibration and verification
exercises performed during the cesium-137 simulations provided valuable
insight to the processes associated with sediment-bound contaminant
movement. Cesium-137 proved to be a unique data set in Lake St. Clair. The
loading function was well documented, the initial conditions were known, and
two sets of spatially-complete observations were available. The knowledge
gained from the Cs-137 simulations (e.g., organic carbon contents) was
applied to the following OCS and PCB simulations.
402
-------
Octachlorostyrene
The simulation of DCS in Lake St. Clair was intended as an application
of the model, not necessarily as a further test of the model's ability to
predict an observed distribution. A fundamental problem existed: the actual
time-variable loading function was unknown. However, on the basis of Great
Lakes Institute (1986) estimates, the load was assumed to be constant at 1.9
Ibs/day. The model was run until the simulated DCS levels in the active
layer (0-10 cm) agreed with the observed 1983 values. This occurred at 4500
days, or just over 12 years (Figure 7), implying that the load was first
introduced in the latter part of 1970. This result is consistent with
Limno-Tech, Inc.'s (1985) speculation that OCS was introduced to the lower
Great Lakes beginning in the 1970s.
Using the sediment organic carbon values from this study, the estimated
OCS load, and the constants in Table 2, the model was able to reproduce the
observed OCS pattern in the sediments; that of a concentrated plume of OCS
extending from the mouth of the South Channel to the head of the Detroit
River. However, the model-calculated plume was wider and slightly longer
than the measured plume. The model predicted that the mean and maximum
sediment concentrations in 1983 were 3.8 and 23.1 MgAg, respectively.
These compare with 2.7 and 26.2 ^gAg reported by Pugsley et. al. (1985).
The model predicted 1983 active layer bio-bound OCS levels of 0-96 ,gAg dry
wt., with a mean concentration of 20 ,g/kg dry wt. Pugsley et. al. (1985)
reported levels of 2-154 MgAg dry wt. (mean - 43 MgAg dry wt.) measured in
whole clam tissue (lamUit ^ia^ jlUuaidtt) collected in Lake St.
Clair during 1983.
403
-------
Even though 55X of the St. Clair River flow is directed through the
North and Middle Channels, the negligible levels of OCS in the sediments of
Anchor Bay tend to verify that the majority of the OCS load to Lake St.
Clair entered through the South Channel from its likely source in Sarnia.
Model-calculated loss rates ranged from 0.03-0.08 day"! in the water column
and 0.34x10-^ - 0.37xlO'3 day1 in the sediments.
During the 4500-day simulation period, the model predicted that 3.9 MT
of OCS entered Lake St. Clair. Of this total load, 2.6 MT or 68Z, were
flushed from the system through the Detroit River, 0.7 MT or 18X, were lost
due to biological degradation and volatilization, and 0.5 MT or 13X,
remained in the system. That which remained in the lake was concentrated in
the sediments between the South Channel and the Detroit River. It is worth
noting once again that the OCS simulation was performed without prior
knowledge of the OCS load to Lake St. Clair. Although we were able to
adequately reproduce the 1983 OCS observations using a constant load of 1.9
Ibs/day for 4500 days, this solution is not necessarily unique. There is no
reason to believe that the load remained constant during the entire
simulation period. Any number of time-varying load magnitude and duration
combinations could have produced similar results. However, with known
initial conditions (zero concentration in all segments), the model was able
to predict how much OCS (3.9 MT) had to be loaded into the system to produce
the 1983 observations. It also provided some insight on the origin of the
OCS load to Lake St. Clair.
404
-------
Polvchlorinated Biphenvls
The simulation of total PCB in Lake St. Clair was also intended as an
application of the model. However, as with the OCS simulation, the loading
function of PCB was unknown during the period 1970-74. Great Lakes
Institute (1986) estimated the load to be about 5 Ibs/day, based on model
simulations. This load was used in the present study. With initial
conditions set equal to conditions in 1970 (Figure 3), and using the
constants presented in Table 3, the model was able to reproduce the observed
sediment PCB distribution in 1974 fairly well (Figure 8). In general, the
model accurately predicted the 1974 PCB sediment concentrations in the
Anchor Bay and the open-lake sediments. However, the model tended to
overpredict the PCB values along the eastern and western segments of the
main lake, which may indicate additional or increased PCB sources in these
areas.
The data indicate a decline in mean lake-wide sediment (0-2 cm)
concentration of total PCB from 19 MgAg 1970 to 10 MgAg in 1974 and in
maximum sediment PCB concentration from 40 MgAg in 1970 to 28 ,g/kg in 1974
(Frank et. al. 1977). The model simulated a similar 4 year decline in mean
active layer sediment concentration from 19.8 MgAg to 8.7 Mg/kg -i* -
maximum sediment concentration from 39.0 ,gAg to 26.0 MgAg- The model
predicted 1974 active layer bio-bound total PCB levels of 31-212 ,gAg dry
wt with a mean concentration of 97 MgAg- *, comparable literature
estimates of bio-bound PCB were available for the early 1970s. However,
Pugsley et. al. (1985) reported mean values of 90.6 and 44.2 ,gAg whole
405
-------
clam tissue (L. radiata) for Aroclors 1254 and 1260, respectively, in Lake
St. Glair during 1983.
The model-calculated volatilization rates ranged from 0.11-0.18 m/d.
This range compares well with the theoretical rates for Aroclor 1242, 0.21
m/d, and Aroclor 1260, 0.17 m/d (Richardson et. al. 1983). The total loss
rate was calculated to be 0.04-0.11 day1 in the water column and 0.34x10'^
- 0.45x10'3 day"1 in the sediments.
From 1970 to 1974, the model predicted that the total system mass of
PCB decreased from 2.5 MT to 1.5 MT. During this time period, 3.4 MT of PCB
were loaded into the system, 2.3 MT were flushed from the lake through the
Detroit River, and 2.1 MT were lost due to biological degradation and
volatilization. Again, as with OCS, the PCB simulation was performed
without prior knowledge of the PCB load to Lake St. Clair. Thus, the 1974
solution, although an adequate reproduction of the data, is not necessarily
unique. Any number of time-varying load magnitude and duration combinations
could have produced similar results. However, with known initial conditions
(1970 values), the model was able to predict how much PCB (3.4 MT) had to be
loaded into the system to produce the 1974 observations. It also provided
some insight to the possibility of additional PCB sources along the eastern
and western main-lake segments of Lake St. Clair.
406
-------
LITERATURE CITED
Ambrose, R.B., Hill, S.I., Mulkey, L.A. 1983. User's Manual for the Chemical
Transport and Fate Model TOXIWASP. Version 1. EPA Report No. EPA-
600/3-83-005. Environmental Research Laboratory, Office of Research and
Development, U.S. Environmental Protection Agency, Athens, Georgia. 178
PP-
Bell, G.L. 1980. Lake St. Clair and St. Glair and Detroit Rivers Chemical
and Physical Characteristics Data for 1974. NOAA Data Report ERL GLERL-
12. Great Lakes Environmental Research Laboratory, Ann Arbor, Michigan.
10 pp.
Bukata, R.P., Jerome, J.H., and Bruton, J.E. 1987. Remote and In Situ
Optical Studies of Seston and Suspended Sediment Concentrations in Lake
St^ Clair. Preliminary Report. Rivers Research Branch, National Water
Research Institute, CCIW, Burlington, Ontario, Canada.
Burns, L.A., Cline, D.M., andLassiter, R.R. 1982. Exposure Analysis
Mode line System (EXAMS^) : Use^ Manual gnd System Documentation. EPA
Report No. EPA-600/3-82-023. Environmental Research Laboratory, Office
of Research and Development, U.S. Environmental Protection Agency,
Athens, Georgia. 145 pp.
407
-------
DiToro, D.M., Fitzpatrick, J.J., Thomann, R.V. 1983. Documentation for Water
Quality Analysis Simulation Program (WASP) and Model Verification
Program (MVP). EPA Report No. EPA-600/3-81-044. Environmental Research
Laboratory, Office of Research and Development, U.S. Environmental
Protection Agency, Duluth, Minnesota. 145 pp.
Frank, R. , Holdrinet, M., Braun, H.E., Thomas, R.L., Kemp, A.L.W., and
Jaquet, J.-M. 1977. Organochlorine insecticides and PCBs in sediments of
Lake St. Clair (1970 and 1974) and Lake Erie (1971). Sci. Tot. Environ.
8:205-227.
Great Lakes Institute. 1986. A Case Study of Selected Toxic Contaminants in
the Essex Region. Volume I: Physical Sciences. Parts One and Two. Final
Report. University of Windsor, Windsor, Ontario, Canada.
Ibrahim, K.A. 1986. Simulation of Pollutant Transport Responses to Loading
and Weather Variations in Lake St. Clair and the Connecting Channels.
PhD Dissertation, Department of Civil Engineering, University of
Windsor, Windsor, Ontario, Canada. 436 pp.
Kaiser, K.L.E., Comba, M.E., Hunter,H., Maguire, R.J., Tkaca, R.J., and
Platford, R.F. 1985. Trace organic contaminants in the Detroit River.
J. r.r**r T^kes Res. 11(3) : 386-399 .
408
-------
Kuel, D.W., Leonard, E.N. , Welch, K.J., and Veith, C.D. 1980. Identification
of hazardous organic chemicals in fish from the Ashtabula River, Ohio
and Wabash River, Indiana. Association of Official Analytical Chemists
Journal . 63:1238-1244.
Limno-Tech, Inc. 1985. Summary of the Existing Status of the Upper Great
Lakes Connecting Channels Data. Limno-Tech, Inc. , Ann Arbor, Michigan.
157 pp.
Mabey, W.R. , Smith, J.H., Podoll , R.T., Johnson, H.L. , Mill, T. , Chou,
T.-W., Gates. J., Waight Partridge, I., Jaber, H. , and Vanderberg, D.
1982. Aquatic Fate Process Data for Organic Priority Pollutants. EPA
Report No. EPA- 440/4 -81 -014. Monitoring and Data Support Division,
Office of Water Regulations and Standards, Washington, DC. 145 pp.
Maguire, R.J., Tkaca, R.J., and Sartor, D.L. 1985. Butyltin species and
inorganic tin in water and sediment of the Detroit and St. Clair Rivers.
T fir«at lakes Res. 11(3) :320-327 .
Oliver, B.C. 1987. ^ n«1r Riv-r sediments. A level II report for the
Upper Great Lakes Connecting Channels Study. ELI Eco Laboratories Inc.,
Rockwood, Ontario, Canada.
409
-------
Oliver, B.C. and Bourbonniere , R.A. 1985. Chlorinated contaminants in
surficial sediments of Lakes Huron, St. Clair, and Erie: implications
regarding sources along the St. Clair and Detroit Rivers. J. Great Lakes
Res. ll(3):366-372.
Pugsley, C.W., Hebert, P.D.N., Wood, G.W. , Brotea, G. , andObal, T.W. 1985.
Distribution of contaminants in clams and sediments from the Huron-Erie
corridor. I-PCBs and octachlorostyrene. J. Great Lakes Res.
ll(3):275-289.
Richardson, W.L., Smith, V.E., and Wethington, R. 1983. Dynamic mass balance
of PCS and suspended solids in Saginaw Bay- a case study. In Physical
Behavior of PCBs in the Great Lakes. Edited by Mackay, D. , Paterson, S.,
Eisenreich, S.J., and Simmons, M.S. Ann Arbor Science, Ann Arbor,
Michigan. 442 pp.
Robbins, J.A. and Oliver, E.G. 1987. Accumulation of fallout cesium-137 and
chlorinated organic contaminants in recent sediments of Lake St. Clair.
Can. J. FJ«h- and Ag^tic Sci. Submitted.
Robbins J A 1985. Th^ CoupleH T^kes Mndel for Estimating the Long-Term
i,,,^- nf t-.h» Great JLakes_1-o Tii^-nPT^dent Loadings of Particle-
AMOCi.ted Contaminants. NOAA Technical Memorandum ERL GLERL-57. Great
Lakes Environmental Research Laboratory, Ann Arbor, Michigan. 41 pp.
410
-------
Roberts, J.R., Rodgers, D.W., Bailey, J.R., andRorke, M.A. 1978.
Polychlorinated Biphenvls: Biological Criteria for an Assessment of
their Effects on Environmental Quality. Publication No. NRCC 16077.
National Research Council of Canada, NCR Associate Committee on
Scientific Criteria for Environmental Quality Ottawa, Canada. 172 pp.
Rybczyk, J.M. 1986. Cluster Analysis of Physical, Biological, and Chemical
Data of Lake St. Clair. Unpublished Data. Great Lakes Environmental
Research Laboratory, Ann Arbor, Michigan.
Schwab, D.J. and Liu, P.C. 1987. Development of a Shallow Water Numerical
Wave Model for Lake St. Clair. Upper Great Lakes Connecting Channels
Study Final Report. Great Lakes Environmental Research Laboratory, Ann
Arbor, Michigan.
Simons, T.J. and Schertzer, W.M. 1986. Modelling Wave-Induced Sediment
R^usoension in Lake St. Clair. Preliminary Report. Aquatic Physics and
Systems Division, National Water Research Institute, Burlington,
Ontario, Canada.
Thomann, R.V. and Mueller, J.A. 1983. Steady state modeling of toxic
chemicals-theory and application to PCBs in the Great Lakes and Saginaw
Bay. m m.y-^.1 R*h»vi~ ^ »™« *" the Great Lakes. Edited by Mackay,
D., Paterson, S., Eisenreich, S.J., and Simmons, M.S. Ann Arbor Science,
Ann Arbor, Michigan. 442 pp.
411
-------
Table 1. Lake St. Clair parameters and constants required by the TOXIWASP
model. Units are consistent with model specifications. Reference in
parentheses.
Variable
Parameter
TEMP
DEPTH
VELOC
WIND
BACTO
BIOMS
OCS
PCTWA
PH
US
CMPET
Constant
OCB
CLOUDG
LATG
Description
Average segment temperature
Depth of segment
Average water velocity
Average wind velocity
Bacterial population density
Total biomass in segment
Sediment organic carbon content
Percent water in sediments
Hydrogen ion activity
Settling rate in water column
Resuspension rate in bed
Contaminant burial rate
Light extinction coefficient
Biomass organic carbon content
Average cloud cover
Geographic latitude
• '
Units
degrees C
feet
feet s"1
meters s~l
cells ml'l (water)
cells 100 g'1 (bed)
mg 1~1 (water)
g m"2 (bed)
dimensionless
dimensionless
pH units
m dayl
cm yr'l
cm yr'l
m-1
dimens ionless
Value
13(1)
6.6-17.1(2)
0.48(2)
6(3)
106(4)
107-108(4)
10(4)
1-50(4)
0.14-5.0X(5)
1.67-1.71(5)
8.1(1)
3.0(6)
0.45-0.95(5)
0.1(7)
2.0(1)
«(5)
tenths of full cover 4(3)
degrees and tenths
43.2
ISTORET DATA
2schwab (1987)
3GLI (1986)
^Ibrahim (1986)
5This study
^Simons and Schertzer (1986)
Bobbins and Oliver (1987)
412
-------
Table 2. Octachlorostyrene constants required by the TOXIWASP model. Units
are consistent with model specifications. From Ibrahim (1986).
Constant Description Units Value
ROW Octanol water partition coefficient lw !Oct"^ 2.48x10**
KOC Organic carbon partition coefficient lw kg"1 1.20x10^
MWT Molecular weight g mole'1 300
HEN Henry's Law constant Atm m^ mole"1 10"^
VAP Vapor pressure torr 4x10"5
SOL Aqueous solubility mg I'1 0.02
413
-------
Table 3. Total polychlorinated biphenyl constants required by the TOXIWASP
model. Units are consistent with model specifications. From Mabey et. al.
(1982).
Constant
KOW
KOC
MWT
HEN
VAP
SOL
Description
Octanol water partition coefficient
Organic carbon partition coefficient
Molecular weight
Henry's Law constant
Vapor pressure
Aqueous solubility
Units
lw loct
lw kg-1
g mole'l
Atm m^ mole'l
torr
mg l"1
Value1
4.14xl05
2xl05
300
3.9x10-3
5x10 '4
5xlO-2
IWithin range of values for Aroclors 1232, 1242, 1248, 1254, and 1260. They
most closely resemble values for Aroclor 1248.
414
-------
LIST OF FIGURES
Fig. 1. Lake St. Clair numerical grid used in Lake St. Glair contaminant fate
and transport model, based on EPA's TOXIWASP. Segment numbers
correspond to water column segments. Active layer and deep layer
segment numbers are determined by adding one and two, respectively,
to the water column numbers.
Fig. 2. Loading of Cs-137 to Lake St. Clair during the period 1950-1985 from
three principal sources: inflow from Lake Huron, direct atmospheric
fallout, and land runoff from the watershed. Reprinted from Robbins
and Oliver (1987).
Fig. 3. Observed distribution and model initial conditions of PCBs (^g/kg) in
surface (0-2 cm) sediments in Lake St. Clair, 1970. Data from Frank
et. al. (1977).
Fig 4. (a) Comparison of observed and model-simulated concentrations (mg/1)
of chloride in the water column of Lake St. Clair during cruise 3,
19-29 June 1974. Wind speed: 4 m/s NE. (b) Comparison of observed
and model-simulated water column concentrations (mg/1) of chloride
during cruise 5, 15-24 July 1974. Wind speed: 5 m/s N. Data from
Bell (1980).
Fie 5 (a) Comparison of observed and model-simulated concentrations (mg/1)
of chloride in the water column of Lake St. Clair during cruise 6,
5-15 August 1974. Wind speed: 5 m/s E. (b) Comparison of observed
and model-simulated water column concentrations (mg/1) of chloride
during cruise 8, 16-25 September 1974. Wind speed: 6 m/s W. Data
from Bell (1980).
Fie 6 (a) Comparison of observed and model-simulated concentrations (dpm/g)
' of Cs-137 in surface (0-2 cm) sediments of Lake St. Clair in 1976.
Model was calibrated by adjusting the spatially-varying sediment
organic carbon content until the model results matched the data. (b)
Comparison of observed and model-simulated 0-2 cm concentrations
fdDm/R) of Cs-137 in 1985. The 1985 simulation verified the 1976
model results. Data from Robbins and Oliver (1987).
FiE 7 Comparison of observed and model-simulated concentrations of
g' octachlorostyrene (DCS) in surface (0-10 cm) sediments of Lake St.
Clair in 1983 The simulation was run until the model results agreed
wlS the 1983'data, which occurred at day 4500; implying that the
load was first introduced during the latter part of 1970. Data from
Pugsley et. al. (1985).
Fig 8 Comparison of observed and model-simulated concentrations (^g/kg) of
total PCBs in surface (0-2 cm) sediments of Lake St Clair in 1974
Simulation was run for four years -ing initial conditions presented
in Figure 3. Data from Frank et. al. (1977).
415
-------
Scale in Kilometers
I I I I I I I
048
i
' •
•
82°30'
-------
.
• •
Cs Loading to Lake St. Clair
Watershed
Watershed * Atmosphere
j Watershed + Atmosphere
1 -i-Inflow from Lake Huron
1960
197O
1980
Year
-------
Scale in Kilometers
I I I I I
0 4
i
i •
0
PCB Concentration 1970
Surficial (0-2 cm) Sediments
Q Not Sampled
D <
n?n 5-10
10-20 /L/g/kg
20-30 /ug/kg
>30 /43/kg
Initial Conditions
-------
Scale in Kilometers
I I I I I I I
0 4
i
i •
Cruise 3
19-29 June 1974
Wind - 4 m/s NE
-------
!
I '
Cruise 5
15-24 July 1974
Wind = 5 m/s N
-------
I
I 1
Cruise 6
5-15 August 1974
Wind = 5 m/s E
82°30'
_JL_
-------
Scale In Kilometers
II I I I I I
i
i i
Cruise 8
16-25 September 1974
Wind = 6 m/s W
82°30'
-------
6/iudpg<
6/Ludp Q-
6/tudp £-
6/u)dp 2-1.
6/tudp 1.-9 0
6/iudp g o>
sjuauijpas (wo 2-0)
9Z61
'• i
' i
i
-------
Scale in Kilometers
I I I I I I I
0 4 8 12
I
i
;
137Cs Concentration 1985
Surficial (0-2 cm) Sediments
<0.5 dpm/g
0.5-1 dpm/g
1-2 dpm/g
2-3 dpm/g
Simulation
-------
M ,
• I
r
-------
I I
i
PCS Concentration 1974
Surficial (0-2 cm) Sediments
Not Sampled
5-10 j/g/kg
10-20 ,ug/kg
20-30 /jg/kg
Simulation
------- |