&EPA
United States
Environmental Protection
Agency
Municipal Environmental Research
Laboratory
Cincinnati OH 45268
EPA-600/2-80-126
August 1980
Research and Development
Movement and
Effects of Combined
Sewer Overflow
Sediments in
Receiving Waters
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3, Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the ENVIRONMENTAL PROTECTION TECH-
NOLOGY series. This series describes research performed to develop and dem-
onstrate instrumentation, equipment, and methodology to repair or prevent en-
vironmental degradation from point and non-point sources of pollution. This work
provides the new or improved technology required for the control and treatment
of pollution sources to meet environmental quality standards.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/2-80-126
August 1980
MOVEMENT AND EFFECTS OF COMBINED SEWER OVERFLOW
SEDIMENTS IN RECEIVING WATERS
by
Stanley L. Klemetson
Colorado State University
Engineering Research Center
Fort Collins, Colorado 80523
and
Thomas N. Keefer and Robert K. Simons
The Sutron Corporation
Fairfax, Virginia 22030
Grant No. R806111
Project Officer
John N. English
Wastewater Research Division
Municipal Environmental Research Laboratory
Cincinnati, Ohio 45268
MUNICIPAL ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
-------
DISCLAIMER
This report has been reviewed by the Municipal Environmental Research
Laboratory, U.S. Environmental Protection Agency, and approved for publication.
Approval does not signify that the contents necessarily reflect the views and
policies of the U.S. Environmental Protection Agency, nor does mention of trade
names or commercial products constitute endorsement or recommendation for use.
ii
-------
FOREWORD
The Environmental Protection Agency was created because of the increasing
public and government concern about the dangers of pollution to the health
and welfare of the American people. Noxious air, foul water, and spoiled land
are tragic testimony to the deterioration of our natural environment. The
complexity of that environment and. the interplay between its components require
a concentrated and integrated attack on the problem.
Research and development is that necessary first step in problem solution;
and it involves defining the problem, measuring its impact, and searching
for solutions. The Municipal Environmental Research Laboratory develops new
and improved technology and systems for the prevention, treatment, and manage-
ment of wastewater and solid and hazardous waste pollutant discharges from
municipal and community sources, for the preservation attd treatment of public
drinking water supplies, and to minimize the adverse economic, social, health,
and aesthetic effects of pollution. This publication is one of the products
of that research, a most vital communications link between the researcher and
the user community.
This report investigates the current capability for determining the move-
ment, fate, and effects of the sediment material from combined sewer over-
flows (CSO's). First, the available literature describing the characteristics
of CSO sediments and their possible effects is reviewed. Next, the knowledge
of these characteristics is used in conjunction with a sediment transport model
to determine the movement of sediments in the Cuyahoga River between Akron
and Cleveland, Ohio. Experiments are described wherein the model is used to
predict the fate of sediment material from high flow bypass of the Akron
municipal treatment plant under various flow conditions,
Francis T. Mayo
Director
Municipal Environmental Research
Laboratory
iii
-------
ABSTRACT
The research work described here was a joint effort of Colorado State
University (CSU) and the Sutron Corporation. The study had two primary objec-
tives. The first objective was to determine from available literature the
characteristics of combined sewer overflow (CSO) sediments and the factors
affecting their transport properties. The second objective was to make use
of the information on characteristics to evaluate a current sediment model
capable of predicting the fate of CSO sediments.
CSU conducted the literature search and evaluation necessary to meet the
first objective. The Sutron Corporation selected a test study site, collected
limited field data, and used the characteristics of CSO sediments found by
CSU to evaluate a sediment model. Sutron also conducted a literature search
of sediment sampling and tracing techniques necessary for model application.
Combined sewer overflows are made up of urban surface runoff and sanitary
sewage. The contribution of sanitary sewage to the total flow is negligible
at times of peak flow in many cases, although its effects on time dependent
changes in the CSO's may be important.
Urban surface runoff makes up the majority of a CSO. The characteristics
of CSO sediment material for model development were thus taken to be those
of street surface solids. The size distribution of street surface solids
appear to be reasonably well defined log-normal distributions. In general,
sizes range from fine to medium sand. Makeup of street surface solids is
difficult to define from the available data, although sufficient data are
available to make an initial approximation. A wide variety of constituents
may be found, ranging from metallic elements to pesticides.
The effect of the chemical properties of the sediments on their transport
characteristics has not been addressed in the literature to any great degree.
Only broad general characteristics of the type of reactions, which might occur,
can be made. Even less appears to be known concerning the effect of deposited
or eroded materials on the biologic community. A good deal is known about what
types of animal and plant life exists in streams but few studies specifically
address the questions of being buried or deprived of light or any of the
dozens of other effects related to sediments.
Sutron investigated the feasibility of modeling the movement of CSO sedi-
ments on 64.37 km (40 mile) reach of the Cuyahoga River between Akron and
Cleveland, Ohio. The reach investigated was quite steep (slope = .00095). Two
small diversion structures formed sediment trap areas. The downstream end of
the reach included the Cuyahoga estuary into Lake Erie.
xv
-------
The sediment transport routines from a water shed model developed at CSU
were combined with a linear implicit finite difference flow model for use in
the feasibility study. Limited field data on sediment sizes and stream cross
sections were gathered by Sutron. Hypothetical sediment loads from the high
flow bypass of the Akron municipal treatment plant were developed from data
obtained by CSU's literature search.
Experiments indicate that the movement of hypothetical sediment loads may
be successfully modeled. Under normal low flow conditions, sediments accumulate
near the bypass outfall. A flood of five-year recurrence interval will move
course sediments from the outfall to the small trap areas. Fine material
moves through the reach and settles in the estuary under all flow conditions.
It was concluded from the model experiments that qualitative evaluation
can be made concerning the fate of CSO solids which are primarily noncohesive
sands. Semiquantitative evaluations could be made if proper data from a
particular CSO of interest could be obtained. Particularly important to the
model are the size distribution and settling velocity characteristics of the
CSO sediments. Several experiments were conducted with hypothetical sediments
with specific gravities similar to elemental heavy metals. These'experiments
indicated that qualitative prediction of the fate of such materials is also
possible. Flood frequency analysis combined with hypothetical flood hydro-
graphs provided a useful tool for analyzing the', fate and residence time of
CSO sediment material deposits. A verification study for the model would be
highly desirable. Most of the data used in this study was hypothetical. The
model's ability to predict the fate of sediments can only be conclusively
verified using data collected for that purpose.
Sutron's investigation of sampling and tracing techniques indicates that
considerable information exists on sampling the suspended portion of stream
sediment loads. No current network exists to accurately measure the portion
which moves along the channel bed by rolling and sliding. This implies that
reaches selected for model verification must be selected carefully to allow
measurement of all material in suspension.
Tracing the movement of CSO sediment materials is possible but may be
impractical. Radioactive tracers are highly developed and effective but
environmentally objectionable. Fluorescent dye methods could probably be
used but involve considerable labor and analysis costs. Results are quali-
tative in nature.
This report was submitted in partial fulfillment of Research Grant No.
R806111 by Colorado State University under the sponsorship of the U.S.
Environmental Protection Agency. This report covers the period of August 15,
1978 to August 14/1979 and was completed March 3, 1980.
v
-------
-------
CONTENTS
Disclaimer . , ii
Foreward ,,,...,.,,, iii
Abstract ...... iv
Figures • . ix
Tables . xi
Model Parameters in Order of Use ...... o xiii
Acknowledgment ................ .... XV£
1. Introduction . .......... 1
Study Background , 1
Objectives ..... 2
Scope of Work . .... 2
Organization and Conductance of Study 2
Organization of the Report ...... . 3
2. Conclusions and Recommendations 4
Introduction ........ 4
Findings and Conclusions . . . „ 4
Knowledge of the Characteristics of CSO
Sediments , 4
Interaction Between Sediment Materials and
Receiving Waters ..... ... 5
Interaction Between Sediments and the Biologic
Community .......... , . 5
Modeling the Movement of CSO Sediments 5
Sampling and Tracing Techniques . 6
Recommendations ........ 7
3. Sediment Characteristics and Potential Impacts ..... 9
Introduction 9
Characteristics of CSO Sediment Materials 9
Approach of This Subsection 9
General Characteristics , 10
Characteristics Related to Transport 21
Summary of Characteristics 36
The Receiving Water Environment 36
Introduction ... ..... 36
Potential Impacts . 38
Interaction of CSO Material With Receiving
Water 39
Biologic Community of Receiving Waters .... 46
Approach for Assessment of Environmental
Impacts 55
VII
-------
4. Modeling the Movement of CSO Sediments . . , 64
Introduction 64
Colorado State University Model and Modifications . 64
Original Watershed Model 64
Model Modifications 65
Cuyahoga River Study Reach 76
General Considerations ...,,....,., 76
Reasons for Selecting Reach 76
General Characteristics of Basin and River . . 76
Supplemental Data 78
Modeling Program Overview 79
Flow Model 79
Sediment Model , 86
5. Related Aspects of Sediment Transport Studies ..... 100
Introduction 100
Sampling Sediments 101
Bed and Bank Material Samples 101
Sampling Sediment Transported by the Flow . . 106
Determining Sediment Flow Rates 114
Characterizing Sediments. , . 116
Definitions 116
Sample Analysis 117
Tracing Sediment Movement 122
References 126
viii
-------
FIGURES
Number
1 Sanitary Sewage Flow Characteristics 11
2 Qualitative Description of an Urban Surface Runoff Event . 13
3 Settling Data for Stormwater Runoff and Sanitary Sewage. . 23
4 Particle Size Distribution of CSO and Street Surface
Solids 28
5 Size Distribution of Selected Constituents of Street
Surface 30
6 Time Dependence of Suspended Solids Concentrations in
CSO's 37
7 Schematic Diagram of CSO Solids Transport Process .... 42
8 Reaction Mechanisms and Material Interactions. 43
9 The Pyramidic Food Chain 48
10 Definition of Aquatic Ecosystem 50
11 Environmental Factors Which Produce Undesirable Effects in
Aquatic Life ' 54
12 Flow Chart for the Watershed Sediment and Routing Model . 66
13 Model Reach 77
14 Steady Flow Water Surface Profiles 80
15 Hourly Discharge as DO Data at Old Portage . . 82
16 Cuyahoga River Discharge Hydrographs from June 19, 1976
to July 18, 1976 83
17 Supplemental Discharge Data . 84
18 Dissolved Oxygen, Flow, and Sediment at Independence ... 85
19 Flood Frequency Data, Cuyahoga River at Old Portage. ... 87
20 Deposition and Erosion 91
21 Deposition and Erosion 92
22 Deposition and Erosion at STP Outfall 94
23 Deposition and Erosion Ten Miles Below Old Portage .... 95
24 Deposition and Erosion of Heavy Particles . 97
25 Deposition and Erosion of Heavy Particles at STP Outfall . 98
26 Deposition and Erosion of Heavy Particles Ten Miles Below
Old Portage 99
27 US BMH-53—Bed Material Sampler 104
28 Hand-Line, Spring-Driven, Rotary-Bucket, 30-Pound Bed
Material Sampler 105
29 US BM-54—Bed Material Sampler 105
30 Measured and Unmeasured Sampling Zones in a Stream
Sampling Vertical with Respect to Velocity of Flow and
Sediment Concentration 108
xx
-------
Number
31 Depth-Integrating, Suspended-Sediment, Wading-Type,
Hand Sampler, US DH-48 . .
32 Depth-Integrating, Suspended-Sediment, Hand-Type
Sampler, US DH-59 .'
33 Depth-Integrating, Suspended-Sediment, Cable and Reel
Sampler, US D-49
34 Components and Dimensions of the Basic Single-Stage,
Suspended-Sediment Sampler, US U-59 . .
35 Flow Chart for Bed and Bank Material Analysis . . , .
36 Flow Chart for Analysis of Suspended Sediment Samples
110
110
110
112
118
119
x
-------
TABLES
Number Page
1 Reported Parameter Concentrations for Typical Raw Domestic
Sanitary Wastewater Flows , . . . 12
2 Typical Parameter Concentration in Urban Surface Runoff , . 14
3 Observed Runoff Water Quality Concentrations in Urban
Street Runoff.in San Jose, California 15
4 Average Nation-Wide Pollutant Strengths Associated with
Street Surface Particulates ..... 16
5 Empirical Constants for the Equation C, /o\~^m( f •\t'aCh }
for Parameter Concentration in Urban Surface Runoff for
Third Fork Catchment, Durham, North Carolina 18
6 Comparison of Flow-Weighted BOD5 and Suspended Solids Means
and Standard Deviations by Land Use and Type of Sewage. . 19
7 Typical Parameter Concentrations for Sanitary Sewage, Urban
Surface Runoff, and Combined Sewer Overflows . 21
8 Particle Size Distributions for Urban Street Surface Partic-
ulates 25
9 Particle Size Distribution for Street Solids Samples from
Chicago, Illinois . . . .• 25
10 Particle Size Dis tribution for Street Solids Samples from
Washington, D.C. . . , 26
11 Particle Size Distribution of Suspended Solids in CSO's in
Lancaster, Pennsylvania 26
12 Particle Size Distribution of Suspended Solids in CSO's in
San Francisco, California 27
13 Particle Size Distribution in Percent of Suspended Solids
in CSO's in San Francisco, California by Catchment
Location. , . 27
14 Fraction of Constituent Associated with Street Solids
Particle Size Ranges 29
15 Fraction of Constituents Associated with Street Solids
Particle Size Ranges (Washington, D.C.) 31
16 Typical Specific Gravities of Material in Street Surface
Solids , . . 32
17 Average Daily Accumulation on Roadways in Materials in the
"Dust and Dirt" Fraction ...... 32
18 Summary of Data on Components of Street Litter in Chicago,
Illinois 33
xi
-------
Number
19
20
21
22
23
Specific Gravity of "Dust and Dirt" Street Surface Samples
in Chicago, Illinois , . . 33
Predicted Data for Composition and Specific Gravity of
CSO's .................. 34
Size Distribution of Suspended Sediment Load at Old Portage . 88
Distribution of Sizes in Bed Sediment 89
Distribution of Sizes in Akron STP Flow ; 89
xii
-------
MODEL PARAMETERS IN ORDER OF USE
EQUATIONS (1), (2), AND (3)
G = total sediment transport rate by volume, volume/time
s
3x = section length
C = Sediment concentration by volume, volume/volume
A = river cross-section area, area
9t = unit time
P = wetted parameter, cross-section length
z = net depth of loose soil, length
g = lateral sediment inflow, volume/unit length/time
S
Q = water flow rate, volume/time
EQUATION (4)
T = sediment shearing stress, force/area
c
<5 = empirical constant, range 0.01 to 0.06, used 0.047
S
Y = specific weight of sediment, weight/unit volume
S
Y = specific weight of water, weight/unit volume
d = particle diameter, length
S
EQUATIONS (5), (6), AND (7)
q = bed load transport rate, volume/unit width
T = boundary shear stress acting on sediment particle, force/area
T = critical sediment tractive force, force/area
c
a = empirical constant, used 8/(vp (Y -y)
b = empirical constant, used 1.5
p = density of water, mass/volume
f = Darcy-Weisbach friction factor, dimensionless
V = average flow velocity, length/time
EQUATIONS (8), (9), (10), AND (11)
C,. = sediment concentration at a distance above the bed, weight/volume
xiii
-------
C * - known, sediment concentration at distance "a*" above the bed,
weight/volume
R = hydraulic radius, area/wetted perimeter, length
g = height above bed, length
a* = reference height above bed, length
w = parameter based on settling velocity and shear velocity,
dimensionless
V = settling velocity of particle, length/time
s
UA » shear velocity, length/time
TA - specific shearing stress, force/area
EQUATIONS (12) THROUGH (17)
U = point mean velocity at a distance above the bed, length/time
B, - empirical constant dependent upon roughness, dimensionless
n = Manning's number range 0.01 to 0.1, used 0.035
S
a = ratio of sample distance above bed to hydraulic radius,
dimensionless
G « ratio 'of reference distance above bed to hydraulic radius,
dimensionless
a - thickness of bed layer equal to twice the size of the sediment,
length
EQUATIONS (18) THROUGH (23)
J = Einstein's integrals, dimensionless
q - suspended sediment transport rate, volume/unit width
S
q = total sediment load, volume/unit width
G - sediment transporting capacity, volume/time
EQUATIONS (24) THROUGH (28)
a
Z
F
adjusted fraction of sediment in the ith size
loose sediment depth in each size fraction, length
original bed material percentage in each size fraction,
dimensionless
)QA - size of'sediment for with 84 percent of the sample is finer,
length
xiv
-------
EQUATION (29)
P
AZ = total potential changes in loose soil storage, length
D = total amount of detached soil, length
T>f = detachment coefficient, range 0.0 to 1.0
p
C = potential sediment load, concentration, volume/volume
n = time index
j = location index
EQUATION (30)
0 = travel time per distance, time/length
a^ = space weight factor, usually set to 0,5, dimensionless
b = time weight factor, usually set to 0.5, dimensionless
xv
-------
ACKNOWLEDGMENTS
Colorado State University and The Sutron Corporation gratefully
acknowledges the cooperation of the United States Geological Survey in
obtaining data for this study and the helpful comments on the preliminary
report by Mr. John English and Mr. Doug Ammon of the Environmental Protection
Agency. Additional assistance on this report was obtained from Drs. G.
Fred Lee and R. Anne Jones, who prepared the section of the report on'the
approach for assessment of environmental impacts of the combined sewer over-
flow sediments.
xvi
-------
SECTION 1
INTRODUCTION
STUDY BACKGROUND
Considerable effort has gone Into the study of sewer systems and sewage
treatment. Less is known,, however, about the impact on receiving waters of
material which escapes the sewer via urban stormwater runoff and combined
sewer overflows (CSO's) during storm events. The study reported here investi-
gates the current technology for modeling the movement of sediment materials
from CSO's.
The original impetus for a study such as this was provided by a 1974
EPA report authored by the North Carolina Water Resources Research Institute
(1). An intensive study was made of the runoff from a 1,67 square mile urban
watershed in Durham, North Carolina. The urban runoff yield of chemical
oxygen demand (COD) was equal to 91 percent of the raw sewage yield. The
biochemical oxygen demand (BOD) was equal to 67 percent, and the urban runoff
suspended solids yield was 20 times that contained in raw municipal waters
for the same area. The study identified the,"first flush" phenomena, wherein
water quality may deteriorate drastically in the early period of storm run-
offs as builtup pollutants are flushed from the system. The importance of
sediments as a source of organic and inorganic pollutants was emphasized
by the facts that plain sedimentation of the runoff resulted in 60 percent
COD removal, 77 percent suspended solids removal, and 53 percent turbidity
reduction.
The Durham study was limited to direct urban land 'runoff. When this
runoff is collected in a combined sewer system and routed to a treatment
plant, additional problems are encountered. It is uneconomical to design
treatment facilities large enough to handle a once in 100 years storm flow
plus the normal municipal sewage load. Thus, at some high flow rate pro-
visions must be made to bypass the treatment facilities with a mixture of
sanitary sewage plus urban runoff. This combined sewer overflow (CSO)
material is characteristically dumped directly into a receiving water. The
Durham study illustrates that discharging the CSO mixture is not very differ-
ent from discharing raw sewage in the receiving water.
Field, Tafuri, and Masters (2) draw on the Durham study and cite an
ongoing R & D study in Milwaukee, Wisconsin, which defines some of the CSO
impact on receiving waters. Strong evidence is present that CSO discharges
intensify dissolved o:xygen (DO) sag and increase fecal coliform concentration.
The Milwuakee study again defined the need for study of the impact of CSO mat-
erial on receiving waters.
-------
The adsorptive and absorptive capacities of CSO sediments has a signifi-
cant effect on the pollution potentials of these sediments during periods of
re-entrainment. Pitt and Field (3) have reported that little is known about
either the short or long term toxic effects of urban stormwater runoff in a
variety of waters and ecosystems. Since in some instances, large amounts of
toxic materials such as heavy metals, pesticides, and PCB's are being dis-
charged along with nontoxic biological and chemical materials, it is desirable
to trace the route these materials take through a receiving water system.
OBJECTIVES
The studies described above indicate a need to study the paths by which
CSO sediments and sediment materials move through receiving waters. The
objectives of the study are to meet this need.
The two primary objectives of the study are to: (1) define the state
of knowledge concerning the sediment transport characteristics of CSO sedi-
ments and (2) evaluate the capability of a sediment transport model to predict
the fate of CSO sediments. Two secondary objectives are to: (1) summarize
current knowledge of the impact of sediment materials on receiving waters
and (2) describe sampling and tracing techniques which may be used to study
the fate of CSO sediment materials.
SCOPE OF WORK
The scope of this study consists of five general areas'with major emphasis
on two. The two areas with major emphasis are: (1) a literature search to
define the current state of knowledge of the quantities and characteristics
of CSO sediments, and (2) an investigation of a combined flow-sediment trans-
port model. The three remaining research areas are: (1) a literature search
for knowledge concerning the impact of CSO sediments on receiving waters,
(2) a literature investigation of the chemical interactions between various
CSO sediment constituents and the receiving water, and (3) a literature
investigation of sampling and tracing techniques.
ORGANIZATION AND CONDUCTION OF STUDY
Colorado State University conducted the major portion of the literature
survey for CSO sediment characteristics and effects. The Sutron Corporation
conducted the model investigation and the literature survey of sampling and
tracing techniques.
Colorado State University used standard library search methods as well
as computer keyword searches to identify pertinent references. All material
concerning the transport characteristics of CSO sediments were summarized
and forwarded to Sutron for use in the model study.
Sutron selected a 64.37 kilometer (40 mile) reach of the Cuyahoga River
between Akron and Cleveland, Ohio for model application based on data availa-
bility and results of previous study (4). Sediment and flow data were
obtained from USGS research. CSO flows from the Akron municipal treatment
plant bypass were estimated based on the information gathered by CSU.
-------
While Sutron was completing the model investigation, CSU summarized the
available knowledge on interacting of CSO sediments and receiving waters.
Sutron conducted a limited survey of sampling and tracing methods after com-
pletion of the model study.
Sutron combined all of the information provided by the literature survey
with the results of the model studies to form the initial draft report. CSU
reviewed the draft report and made final corrections.
ORGANIZATION OF THE REPORT
The report presentation is organized into the five general areas in the
scope of work. The first portion of the report concentrates on defining the
characteristics and quantities of solids and sediment materials which come
from treatment plant and combined sewer overflows. Consideration is given
to the possible interaction of these materials with the receiving water
environment. The second portion of the report deals with existing technology
for modeling sediment movement. Emphasis is placed on the application of this
technology to the prediction of the fate of sewer-related Pediments. Sedi-
ment transport routines from a watershed sediment model developed at Colorado
State University were combined with a finite difference flow model of the
Cuyahoga River below the Akron, Ohio, sewage treatment plant as described
in the study organization. The results of these tests are used to define
the usefulness of current modeling knowledge. Weaknesses in the model are
identified and areas for further research are defined. The final portion
of the report deals with related aspects of sediment studies. These include
sampling and analysis procedures plus a short discussion of sediment tracer
studies.
-------
SECTION 2
CONCLUSIONS AND RECOMMENDATIONS
INTRODUCTION
The conclusions from this study fall into five general areas. These
concern the characteristics of CSO sediments, the interaction of the sedi-
ments with the receiving waters, the interaction of the sediments and the
biologic community, modeling of sediment movement, and sampling and tracing
of sediments. These areas correspond in general to elements of the scope of
work. The general findings and conclusions from each of the five areas are
described below. The general recommendations made as a result of the study
are presented last*
FINDINGS AND CONCLUSIONS
Knowledge of the Characteristics of CSO Sediments
Characterizing CSO sediments was a major work element. It was found that
considerable information is available concerning the nature of the sediment
material from CSO's. In general, this information has been collected near
major urban areas such as San Francisco, Washington, D,C,, or Philadelphia
as part of site-specific studies, Ample evidence exists to show the undesir-
able nature of CSO sediments. Pollutants ranging from lead and mercury to
pesticides may be found. In terms of modeling sediment movement, the litera-
ture is sparce on useful data. The data suggest that during times of storm
flow the urban surface runoff is the major contribution to CSO flow. Thus
surface runoff sediments are a major portion of CSO sediments. Street sur-
face solids comprise the major portion of urban surface runoff. These solids
range from .063 to 2 or 3 mm in diameter with the distribution of sizes
roughly normal. The median diameter varies with geographic location. It
would be highly desirable to obtain further information on both the size
distribution and settling characteristics of the solids from CSO's. Such
information should be routinely gathered as part of any site-specific CSO
sediment model study.
Only general conclusions can be drawn regarding the characteristics of
CSO sediments. The approximate concentrations and the general size and com-
position are known. Specifically lacking for modeling purposes are:
—detailed information on the size distribution and settling character-
istics of the sediments over the course of a storm event;
-------
—general information on the geographic differences on the size dis-
tribution and settling characteristics of CSO sediments; and
—additional information concerning the variety and nature of pollutants
associated with sediments.
Studies to obtain some or all of the above at a variety of geographic locations
would be desirable.
Interaction Between Sediment Materials and Receiving Waters
The literature search resulted in almost no information concerning the
chemical interaction of CSO sediments either with themselves or the receiving
water. Chemical reactions may change the sediment transport characteristics
of CSO sediments, particularly the particle size and weight. The magnitude
of such changes and their effect on predictions of sediment models is
unknown.
Because of the complexity and number of possible reactions, it is doubt-
ful that a meaningful sediment model, including chemical reactions, could be
developed on purely theoretical grounds. A highly useful area' of research
would be to conduct size distribution and settling tests on actual CSO
sediment materials. These tests should include time variation in settling
(say two weeks to one month) and settling in different native waters (differ-
ent pE, alkalinity, etc.). Such tests would establish the variation in
settling properties and provide reasonable grounds for adjusting models to
account for these factors.
Interaction Between Sediments and the Biologic Community
Volumes of information are available concerning the variety of plant
and animal life which live in receiving waters. With the exception of fish
spawning beds, little has been done to determine the effects of sediment on
the biologic community. A number of general statements can be found regard-
ing turbidity and reduction of light to plants. It is certain that sediment
deposits cause biologic communities to move from one area of a stream to
another. Certain plant species may be killed by burial or conversely fer-
tilized by shallow deposits. A major interdisciplinary study would be
required to determine the effects for any specific stream reach. Some general
information could be obtained by investigating the sensitivity or depth' of
burial. Such investigations might prove useful at a later time when the
ability to predict sediment deposit areas and rates has been firmly estab-
lished. At this time, further study of the biologic community is not
warranted.
Modeling the Movement of CSO Sediments
The feasibility of modeling the movement of CSO sediments was investi-
gated on a 64.37 (40 mile) reach of the Cuyahoga River between Akron and
Cleveland, Ohio. The reach investigated was quite steep (slope = .00095).
Two small diversion structures formed sediment trap areas. The downstream
end of the reach included the Cuyahoga estuary into Lake Erie.
-------
The sediment transport routines from a watershed model developed at CSU
were combined with a linear implicit finite difference flow model for use in
the model study. Limited field data were gathered on sediment sizes and
stream cross sections. Flow and suspended sediment data were obtained from the
U.S. Geological Survey. Inflows to the study reach from the Akron treatment
plant high flow bypass (a CSO) were estimated using data from CSU's litera-
ture search.
Experiments indicate that the movement of the hypothetical CSO sediment
loads could be successfully modeled. Under normal summer low flow conditions,
sediment accumulates near the outfall. A flood of five-year recurrence inter-
val will move coarse sediments from the outfall to the trap areas. Fine
material moves through the reach and settles in the estuary under all flow
conditions.
It was concluded from the model experiments that qualitative and per-
haps quantitative predictions can be made concerning the fate of CSO sedi-
ments. Qualitative evaluation will require that the model be verified on
data collected for that purpose. Particularly important for a verification
study will be site specific values of size distributions and settling char-
acteristics of CSO materials.
Several experiments were conducted with hypothetical sediments with
specific gravities similar to elemental heavy metals. Qualitative predictions
concerning the movement of these materials is also possible. In the study
reach used here, heavy materials settled out at the outfall and were moved
downstream by a five-year flood.
The lack of site specific information at the Akron bypass prevented
accurate verification in this study. The model currently available will
work for materials with variable soecific gravities and sizes ranging from
silt to gravel. Currently, the model is in two separable pieces, flow and
sediment transport. Modifications may be required for directly coupling
the flow portion of the model to the sediment transport portion if large
changes in bed elevation (10 cm or so) occur in the channel bed. Such
large changes would change the water surface profile and transport velocities
thus requiring the direct coupling.
The current model considers all sediments to be noncohesive. This
generally limits the applicability to particles silt size and larger (generally,
.063 mm or larger). Enough data are available in the literature to extend
the model to noncohesive clays. The model is primarily limited by the amount
and quantity of input data. Particularly important for sediment studies are
settling velocity, data from the particular outfall or outfalls in question,
accurate data concerning the longitudinal slope of the channel bed, and the
characteristics of the bed and suspended sediments.
Sampling and Tracing Techniques
A sizable body of literature exists concerning sampling and tracing of
sediments. Suspended sediment can be sampled with a fair degree of accuracy
-------
using standard samplers developed by government agencies such as the U.S.
Geological Survey (USGS) and U.S. Agricultural Research Service (ARS). Stan-
dard samplers are also available for sampling bed and bank material, It is
not possible at the time of this writing to determine the quantity of sediment
which rolls, slides or bounces along the stream (the bed load). It is neces-
sary to find stream reaches where all the sediment is carried in suspension for
accurate measurement. The most useful analysis techniques for investigators
contemplating model studies are sieving and settling tests using the visual
accumulation tube. The former is widely used to determine size distributions.
The latter is used to determine fall velocity of particles. Both are impor-
tant to model work. Radioactive tracers are the most effective but least
environmentally acceptable. Fluorescent dyed particles of sand size could
probably be used to trace CSO sediments. Results are highly site specific
and qualitative in nature. Considerable sampling and analysis costs are
involved with the fluorescent dye tracers.
RECOMMENDATIONS
The most important general conclusion from the above section is that the
movement of CSO sediments may be successfully modeled. However, the Cuyahoga
River study was not an accurate verification of the model used here. The
following recommendations are designed to provide sufficient information to
make models quantitatively useful and broaden their range of application:
-.-First, it is recommended that a study site be selected and
sufficient information gathered to accurately verify the model
developed in this study. This information should include:
-water discharge into and out of the reach;
-instream sediment discharge into and out of reach;
-water and solids discharge from CSO outfall or outfalls;
-accurate time of travel study:
-identification of key deposition and erosion zones;
-monitoring of bed elevation at the above key zones; and
-samples of stream bed* bank, instream, and CSO sediments.
If possible, the verification study should be conducted in conjunction with
a study of other water quality parameters to help establish the correlation
of their behavior with Sediment movement.
—Second, it is recommended that as part of the above verification
study that steps be taken to provide empirical information con-
cerning the interaction between the receiving water and the CSO
solids. Information should be gathered on both the changes in
transport characteristics as a function of time, and on the effect
of the solids on the biologic community. The following activities
would be particularly useful:
-Collect large samples of CSO solids over the course of a
storm event. Divide the sample into several portions and
-------
analyze for size distribution and settling characteristics
as a function of time (that is, analyze one of the split
samples every 2 or 3 days). The results would help identify
changes in transport characteristics in deposits.
-Collect samples of CSO solids and determine their settling
characteristics in a variety of receiving water environments.
This could be done by using a standard visual accumulation
tube and varying the pH, alkalinity, temperature, salinity,
and other qualities of the fluid in the column. Such infor-
mation would help establish the range of error due to such
effects.
-As part of the verification study, monitor the biologic
community at the sediment ranges. Changes in biota could
be correlated with the change in depth and location of
deposits. Although such information would be highly site
specific, it would aid in establishing the harmful or bene-
ficial effects of CSO sediment deposits.
-Third, it would be worthwhile to investigate qualitatively the move-
ment of CSO sediments under a variety of stream conditions. This
could easily be done using the sediment model. The reach of the
Cuyahoga investigated in this study was rather steep, and contained
several sediment traps. Ranges of slopes from very flat (say equal
to the lower Mississippi) to that of the Cuyahoga and with and with-
out traps could be modeled. It would also be desirable to model a
variety of bed and bank conditions. The Cuyahoga has a coarse,
armored bed; nearly a rigid boundary. Reaches with fine sand beds
similar to the CSO material will make it much harder to track sedi-
ment movement.
-If the model verification study proves to be successful, final
modifications should be made to the model and documentation provided.
This would make the technology available to other investigators.
-------
SECTION 3
SEDIMENT CHARACTERISTICS AND POTENTIAL IMPACTS
INTRODUCTION
This section of the report is divided into two subsections. The purpose
of the first subsection is to determine from existing studies the character-
istics of the sediment material from combined sewer overflows. Emphasis is
placed on determining those characteristics most important to sediment model-
ing. The purpose of the second subsectipn is to discuss in general the
receiving water environment. First, -the potential interaction between the
receivlngvwater and the sediment material is considered. Next,,the biologic
community is discussed along with the possible impacts caused by the sediment,
Much of the second subsection is speculative in nature because of an
almost complete lack of data. Areas of potential research could be identi-
fied but very few hard facts are available. The most factual information is
contained in the first subsection. Sufficient information was found to
conduct the modeling experiments which are described in Section 4.
CHARACTERISTICS OF CSO SEDIMENT MATERIALS
Approach of This Subsection
The term characteristics, as used in the above heading, means both the
physical characteristics such as size and composition as well as the time
varying characteristics such as quantity and chemical properties. In order
to successfully model the fate and effects of the sediment material from
sewer and combined sewer outfalls, it is necessary to have information on
both.
No field or laboratory data concerning CSO material were collected as
part of this study. A thorough search was made of current literature con-
cerning urban runoff. All relevant information on the physical character-
istics, chemical properties, and time variation of sewer related sediment
material was gathered. Those portions of this information relevant to model-
ing are presented below. The data vary widely in quality, but wherever
possible, limited generalizations were drawn.
CSO's are composed of sanitary sewage and surface runoff typically from
urbanized areas. The relative amounts of these two types of flows will
depend on a variety of factors such as the intensity of rainfall and the
time of day and year. Because the characteristics of sanitary sewage and
-------
urban runoff are quite different, it should be expected that characteristics
of CSO's for any particular area would fluctuate markedly. There is a sig-
nificant body of data characterizing the two types of flows individually
and from these data it is possible to infer the composition of CSO's.
The approach taken here will be to characterize sanitary sewage and urban
runoff individually and then to extrapolate these characteristics to obtain
CSO characteristics as related to physical transport of CSO solids. Because
the impact of CSO solids on water quality may be related more to the materials
attached to the transported solids, an indication of the types and character-
istics of these materials will also be presented. It should be noted that
many reactions can occur that will not significantly affect the transport
potential. For a complete water quality modeling effort, such reactions
would need to be considered. Some of the potential interactions are con-
sidered in the next subsection.
General Characteristics
The characteristics of CSO's will vary significantly- from location to
location and will depend on a variety of factors. Because of this, only
approximate descriptions of the characteristics can be presented in a summary
of this type. In order to develop a model of a specific area; the data
given here should be supplemented.
Sanitary Sewage—
Flows—Sanitary sewage flows vary from location to location and also
vary with time. Typical flow characteristics are presented in Figure 1 (5).
Two distinct daytime peaks typically occur and the ratio of maximum and mini-
mum flows over the course of a day typically ranges from three to five.
Parameterconcentrations—As with 'flow, sewage strength varies over the
course of a day. The fluctuations for many of the constituents are not as
pronounced as the fluctuations in flow and average values are useful (although
site specific data are essential for characterizing individual locations).
Average concentration of selected parameters are listed in Table 1.
Urban Surface Runoff—
Characteristics of urban surface runoff are more difficult to quantify
than are those of sanitary sewage. This is due, in large part, to the diffi-
culties encountered in obtaining representative samples. It is also due to
the fact that concentrations of the various parameters of interest in urban
surface runoff vary markedly with time^ and the runoff events are often of
short duration. There are two stages in gathering data on urban stormwater
characteristics. The first of these is to measure the accumulation and compo-
sition of dust, dirt, and other materials on street surfaces. The second
method is the end of pipe measurement of flow and pollutant concentration.
The street solids make up one of the portion of the solids in urban run-
off and CSO's however, there are other contributions that are not negligible
such as: eroded material from pervious areas, solids washed from nonstreet
impervious areas, re-entrainment of previously deposited materials (e.g., dry-
weather deposition) in the conveyance system, atmospheric washout, and the
10
-------
Average
Flow
I
ctf
V)
0)
Maximum
o
•U
•3
4-J
Average
Minimum
10 20
Average Daily Flow, MGD
30
Figure 1.
Sanitary Sewage Flow Characteristics [After Metcalf and
Eddy, 1972 (5)]
11
-------
TABLE 1. REPORTED PARAMETER CONCENTRATIONS FOR TYPICAL RAW
DOMESTIC SANITARY WASTEWATER FLOWS
Parameter
Average Concentrations, mg/£
Total Solids
Total Suspended Solids
BOD5
COD
Total N as N
Total P as P
Cl~
Pb
Zn
Coliforms (MPN/100 ml)
860
160
150
320
30
8
50
34
7
106
Source: Manning, et al., (6)
sanitary portion of the overflow. Furthermore, the street sampled material
(removed by sweeping, vacuuming, or flushing techniques) may not be related
in the same proportions by the storm washoff process (i.e., the particles
distribution of street sampled material is not necessarily equal to the dis-
tribution of street solids in the runoff). In other words, the street solids
assumption is not that great for urban runoff let alone CSO's, Ammon (7).
While the end of pipe method is the most representative of the actual char-
acteristics of CSO's during a given storm event, the data is difficult to
obtain and is also site specific; therefore, the street surface data is often
more useful for predictive modeling, Berwick et al., (8). A qualitative
description of a runoff event in terms of flow rate, parameter concentration
in the surface runoff, and total parameter load transmitted by the runoff
is given in Figure 2, Amy et al., (9).
From Figure 2, it can be seen that each of the quantities plotted
varies with time. The hydrograph, or "flow" line, reaches a peak value
at some time after the flow begins. This is related to the "time of concen-
tration," the time required for runoff generated in the drainage areas to
reach the flow measurement point. The concentrations of many parameters
of interest decrease with time, a phenomenon that has been indicated by
many researchers and is referred to as the "first-flush" effect. This
phenomenon may become less pronounced as the size of the drainage basin
and intricacy of interconnecting sewers increases, Wanielista, (10). The
product of the flow and the concentration is the "mass rate," which is
seen to reach a peak at some time after runoff occurs.
Another phenomenon that has been reported and that is not indicated on
the figure is variability in characteristics as a function of time of year.
Kleusener and Lee (11) observed this to be particularly true in the case
of certain nutrients. It may be true for other parameters. For example,
suspended solids concentrations may be higher in the spring before grass is
established than in the summer.
12
-------
(SSBR)
(3UITI/SSBH)
4-i
8
o
c
0)
a
cd
MH -H
-I
o v-'
•H
4-1
a.
•H
n
u
w
cu
cd
•u
•H
^H
cd
C'
CN
0)
3
60
AO •[,,).
13
-------
As can be seen from Figure 2, average parameter concentrations for sur-
face runoff, as are often reported, are probably not very meaningful. However,
average values are derived from analyses of samples collected during runoff
events and as such may give some indication of the types and concentrations
of parameters that may be found. Typical values for concentrations in urban
surface runoff are given in Tables 2, 3, and 4.
Parameter concentrations are a function of, among other things, volu-
metric flow rate, time from beginning of the storm, time from the last storm,
and time from the last peak. Colston (1) presented data characterizing a
watershed in Durham, North Carolina. His findings, while very site specific,
Indicate that observed concentrations of various parameters are primarily
functions of the time from the beginning of the storm and volumetric flow
rate, with past history of the drainage area being of lesser importance.
However, the volumetric flow rate is a function of the size of the catch-
ment. He presented equations of the form:
C = b Q
m n
Where:
C = concentration, mg/£
Q = volumetric flow rate, cfs
t - time from the beginning of the storm, hrs
b, m, n = empirical constants
TABLE 2, TYPICAL PARAMETER CONCENTRATIONS IN URBAN SURFACE RUNOFF
Parameter
Concentration, mg/£*
Number of Studies, n
Total Solids
Suspended Solids
BOD5
COD
TOC
NO -N
TKN
NH3-N
Ortho PO.-P
ci-
496
210
14
87
31
0.50
0.72
0.39
0.25
9
20
20
28
25
15
4
16
4
19
15
*Geometric mean of n studies
Data Source: Manning et al., (6)
Empirical values for the constants, determined on the basis of 36 sampled
storms, are presented in Table 5.
14
-------
TABLE 3. OBSERVED RUNOFF WATER QUALITY CONCENTRATIONS IN
URBAN STREET RUNOFF IN SAN JOSE, CALIFORNIA
Parameter, Units*
Number of
Analyses
Minimum
Maximum
*mg/£ unless otherwise noted
**Nephelometric turbidity units
Source: Pitt (12)
Average
Common Parameters and Major Ions
pH, pH units
Oxidation Reduction Potential,
mV
Temperature, °C
Calcium
Magnesium
Sodium
Potassium
Bicarbonate
Carbonate
Sulfate
Chloride
Solids
Total Solids
Total Dissolved Solids
Suspended Solids
Volatile Suspended Solids
Turbidity, NTU**
Specific Conductance, ymhos/cm
88
. 39
11
5
5
5
5
5
5
5
5
20
20
20
10
88
88
6.0
40
14
2.8
, 1.4
, <0.002
1.5
<1
<0.001
6.3
3.9
110
22
15
5
4.8
20
7.6
150
17
19 ..;.
6.2
0.04
3.5
150
0.005
27
18
450
376
845
200
130
660
6.7
120
16
13
4.0
0.01
2.7
54
0.019
18
12
310
150
240
38
49
160
Oxygen and Oxygen Demanding Parameters
Dissolved Oxygen
Biochemical Oxygen (5-day)
Chemical Oxygen Demand
Nutrients
Kjeldahl Nitrogen
Nitrate
Orthophosphate
Total Organic Carbon
Heavy; Metals
Lead
Zinc
Copper
Chromium
, Cadmium
Mercury
11
13
13
13
5
13
5
11
11
11
11
11
11
5.4
17
53
2
0.3
0.2
19
0.10 •
0.06
0.01
0.005
<0.002
<0.0001
13
30
520
25
1.5
18
290
1.5
0.55
0.09
0.04
0,006
0.0006
8.0
24
200
7
0.7
2.4
110
0.4
0.18
0.03
0.02
<0.002
<0.0001
15
-------
H
W
in
CO
IS
M
P
W
^H
M
O
O
CO
CO
CO co
O <|
j|j ri
jyj tT~
co H
53
^4
•*sj"
W
1
H
TJ d
1-1 rt
rt a>
•2 S
pi
rt o
OT d
m o
0 -H
4-J
o rt
•H -H
4-1 >
rt 0>
M P
Tj O
J-4 -H
rt 4-1
rrj CO
d 'r^
rt !>
CO P
f-*
e'S
1 M
« n
S co
1-
O OO
g d
•H Q)
d 5j
S CO
td
4-J
d bO
rt d
Ql fl)
S !-l
CO
Q)
4J
o
d
J-l CO
cu rt
CO 4J
g Cl4
rt co
^ O
rt X
PM co
CO
H
•£>
CM O CM r- r~- oo
P«| T— j f__| fT) ^^ ^— ) LO C^ i-O T~H IO rH ^^ LO i~H t-O O"\ vD ^^ ^O O^ C1^ ^^ ^^
,_(,_),_( CO rHrHOrHOrHOrHrHO iHOOrHrHCMCMCMCOrH
r~~ co co
o co o VD r-~
CO CM CM O O O VO
VD OrHOOOrHOr-
OOO O OCOOOOOOOOrHO OOOOOOOtHOO
OOO O O rHOOOCMCOCsIrH OO
OO«r-~rH
c^j in oo rH r~~
c^ -*
O O CO
•* OOO r-
o o ooo o
CO rH • • • •
OO ^-1 4-1
V^ / S_^ I4_i <4_J (4_( r^ r^ r£) x"*\ Q ^
J3 *H *H L~ ' x™\ ^^-^ ^^^ *"^ ^**s *T^ i — I M
X""N "si* *s}" T~H T^ O *"O ^"^ ,<~| Cti '^^
ro Q o t-H O O CO v — ' O E~^ fd ^^ cj P^ nj
^^-Np-jpM/^^^^ OCJO plO«H CUP^.^^
LO ,d td rH s*-/ ^^^^ '""^ ^"^ s"x X— ** ^"^ Sw*/ ^"^ OJ Cti ClJ (-O £L< PJ i — I $-1 *"CJ rC »JM •"
1 § S S IWSs fi3fi£^S£^§l1lp:p:sl3ll§
16
-------
W
0
•^
cS
£3
C/2
H
W
Pi
H
O3
IT]
EH
H
^
p
E~ '
<| ^
rH rrj
o m
0 3
si
•<3 4J
CO §
EH N^X
CS
M OT
P^ H
CO ^
H o
J7*j . I
*~M r i
& P^
3 2i
l_3
O
W
£H*
HH
^j
1
|55
o
M
§5!
£5
W
O
•^
PH
pi]
^»
•^
.
-. M
n d
TO cd
rH
cd cd
0 &
4-1 4-1
•H
o
p ?"•»
cd cd
^ 13
^^M >JL
W rH Cd
0) -rl 13
4-J rQ rH
d v< -H
4-1 O 1
rH rQ H
rH rH 3
o o-
p, O ~-~
tO ,J2
4-J rH rH
0
CO O
6 cd r--
cd Is ro
!>-]
6C d MH
o o o
O 4-1 d
•H cd O
3 4-1
11 6 cd
S o '>
& o CD
A, cd T3
'cd'
CO
co
•H
J3
•P
CO
£"»
d
g
IH
o
CO
4J
3
CO
cu
cu
4J
m
o
d
o
i \ ^^
Cd rH
•H x~\
p, •> O
B d rH
0 0
0 4-1 d
co cd
Cd x-x rH ,d
i3 co ai -H
v-' >, CO rH H
O ' — ' — cd
^ FQ >
• CO O CO
rH rH 13 O CO CO
Cd •— ' d O rH Ml
cd ft M
4-1 d rt\N S cd
0) CO (-1 Cd rH
a) o II co
>> cd w P & !-i
S ^3 cd O CO CO
tP O 13 CO MH 60
x™\
CN
rH
4-1
4-1
•H
P-<
CO
o
O
CO
17
-------
TABLE 5. EMPIRICAL CONSTANTS FOR THE EQUATION
*, i /-.in . n
'(mg/Ji)
(hrs)
FOR PARAMETER CONCENTRATION IN URBAN SURFACE RUNOFF FOR
FOR THIRD FORK CATCHMENT, DURHAM, NORTH CAROLINA
Parameter
m
n
COD
TOG
TS
TVS
TSS
VSS
TKN
Total P
Al
Ca
Co
Cr
Cu
Fe
Pb
Mg
Mn
Ni
Zn
113
32
420
130
222
44
0.85
0.80
10
12.5
0.07
0.18
0.08
4.6
0.27
10.0
0.45
0.12
0.22
0.11
0.0
0.14
0,09
0.23
0.18
0.87
0.03
0.05
-0.4
0.18
-0.04
0.10
0,24
0.125
-0.02
0,11
0.03
0.10
-0.28
-0.28
-0,18
-0.11
-0.16
-0.17
-0.29
-0.29
-0.15
-0.09
+0.13
+0.06
+0.09
-0.18
-0.29
-0.16
-0.27
-0.01
-0.22
Source: Colston (1)
Colston's results (1) indicate that for nearly all of the parameters
studied a peak concentration occurs during the initial "rising leg" of the
hydrograph. This peak typically occurs at the beginning of the storm,
although for several of the parameters studied, there is an initial increase
in concentration with time followed by a decrease. Such a peak corresponds
to the first-flush effect referred to earlier. The magnitude of the effect
appears to vary among parameters.
Work was completed recently by Huber et al., (15) for the actual storm-
water quality data for CSO's for 35 storms and eight locations. Some of the
data was evaluated by statistical analysis for BOD,, and suspended solids.
While this work is still in its preliminary stages, future regression analysis
work on the presented data will attempt to find the causative relationships
among the water quality parameters and hydrologic and demographic factors.
A summary of the water quality characteristics for the flow-weighted 600$ and
suspended solids data by land use and type of sewerage is presented in Table 6.
18
-------
}-l CO
CU 4-1
fl M-l Cd
6 o cu
3 >
& &
CO
•H
rH
O
CO
.
ft
CO
3
co
cd
CO
CO
m
0
0
rl CO
CU 4-1
O t[ | £*
a o cu
13 W
tO
O
ffi
cd
o
rH
CO
O
O
M TJ ;
CU 60
oor-~inoorHrHoocNOOcoTJ4-14-i>-i!-i T3rHCN IO
rH!>> > CU -H -H -H C 1 rH
IM>JI iTjTSTa-HpiPicficdlsft
rrj"^ CUm-etJ-^clCHpIpInc) i — 1 I-H .cd.
•H UCU4-1 uucdcdcdcdis^cUrE! g
CO 4-ICO- ^3CUCU4l
CU "•rll»>»"«O\COOO«d'rl'H(dlHoocovor^(j>!>i-3fd4-)S
OIFMFJOOI 1 1 1 1 1 LL4-1
^ COM 3cocop3pWp]3p£j' rH
U CO
4-1
CO
•H
Q
•
CO M
3' CU
pQ 4J
id
i-H CU
cd o
M r^d
4-1 4-1
C 3
cu o
o co
1 1
<^ ^
is is
Q) CU*
rH rH
4J 4-1
4-1 4-1
cd cd
cu cu
CO CO
19
-------
CO
o
H
H
^t
H4
5
p
Q
2
^«
p
2j
o
CO
M
3
O
Q^
^
VO
W
PQ
^2
g_f
S-i co
CU 4-1
,0. IW 0
g o cu
3 >
is (S
CO
•rl
rH
O
CO
•
ft
CO
3
CO
cd
O
CO
m
0
o
•
^.
Q) x-v
p <=*
. 00
13 g
4-1 ^"^
CO
0 °^
cd -^
Sjf
S-l CO
CU 4-1
rO MH 0
g o cu
£ &
*
in
p
o
m
cd
o
rH
CO
o
0
^>
Q) ^-^
Q o^
"*^^
. 00
4J **~s
CO
/^*N
0 °^
cd *****
0) W)
s &
V^/
HH T3 CU
O CU 60
0 cd
g -H S-l
P f CU
§J5
CU
co o co
4-J
0
1
O
4-1
cd
0
K^
4-1
•H
0
O"v
OJ
CO
VO
rH
CM
• -H
O *£)
O EH
CO 1 0
•H CJ O
0 !S 4J
0 ftrW
cd « g o
s-i g cd o
[Xi Cd ,0 S-l
,0 JJ PQ
0 S-i S-i
CO P 13
•
CO
0 CO
o
•H S-l
4-J O
cd ft
CU g
cd
>-. CO
0
13 r-1
cu
•H 0
g cd
cu m
4-1 P
Q) O
13 PQ
CO S-l
Tj O
•H MH
O CO
CO CU
rH
13 ft
CU g
13 cd
0 CO
CU
ft OO
CO
3 "
CO 0
o
•H
4-1
cu cd
•H
•r4 S-l
cd
cT
CO g
m fi
O 0
0 4-1
CO
MH l
0 0
•H
13 r0
cd 4->
CU -H
4J &
CO
00-^
•rl O m
CU 13 13 ^
4-1 CU CU
CU 4-1 • CO «
g S-i 0 cd
Cd O O rQ rH
S-i ft -rl cd
Cd CU 4-1 0
ft S-l cd O 4-1
S-l -H CU
S-l 0\ 4-1 4-J
O O1 rH Cd S-l
<+-! CN -H -H CU
O M-l > rQ
cu r~~ cu 3
13 MH TJ ffi
O S-l O
O CU 13
4-> 13 S-l
H cu cd cd cu
w g cu 13 a
P4 Cd 4-1 0 S-l
o s-i co cd 3
H cd 0 4J 0
CO PH -H CO CO
X~S X*~S X~N
Cd rQ O
20
-------
Summary—
Typical data for sanitary sewage, urban surface runoff, and combined
sewer overflows are presented in Table 7. This table gives an indication
of the qualitative effects of combining sanitary sewage with urban surface
runoff. The differences noted between these values and those reported in the
previous tables support the need for future work to determine which factors
affect the water quality.
TABLE 7. TYPICAL PARAMETER CONCENTRATIONS FOR SANITARY SEWAGE, URBAN
SURFACE RUNOFF, AND COMBINED SEWER OVERFLOWS
Parameter
TS
TSS
BOD5
COD
Total N
Orthor PO, as P
Concentration, mg/£
Sanitary
Sewage
700
200
200
500
40
7
Urban
Surface Runoff
496
415
20
115
i
3 to 10
0.6
Combined
Overflows
589
370
115
375
9 to 10
1.9
Source: Manning et al., (6)
Metcalf and Eddy (5)
It is readily apparent that the characteristics of CSO's will be depen-
dent on the relative volumes of urban surface runoff and sanitary sewage.
These, in turn, will depend on the variety of factors including time of day,
intensity of rainfall, time since the beginning of the rainstorm, and time
of year among others.
What can be noted from Table 7 is that sanitary sewage can be expected
to significantly influence the concentrations of most of the parameters. The
important exception, which is notable in light of the purpose of this report,
is suspended solids. The suspended solids concentrations for sanitary sewage
and urban surface runoff are roughly equivalent so if urban surface runoff
contributes the majority of the volume of a CSO, as would typically be the
case, the suspended solids characteristics of the CSO would be expected to
be similar to those of the urban surface runoff. .
Characteristics Related to Transport
Because CSO's can affect water quality in receiving waters, movements of
CSO's .are important. The water quality effects observed will be dependent on
both the location of the CSO material at any time and the characteristics of
21
-------
the material that affect water quality.
ment of CSO solids.
The present study focuses on move-
Sanitary Sewage—
The contributions of solids by sanitary sewage may be small and these
solids will probably have a minor effect on the outfall characteristics, of
CSO solids. Little research effort,, however, has been directed towards
quantifying the relative significance of sanitary sewage and urban surface
runoff on CSO's.
Time dependent changes in the physical characteristics of CSO solids
may be significantly affected by sanitary sewage. For instance, sanitary
sewage will play an important role in any reactions involving nutrients 'that
lead to changes in the physical characteristics of the CSO solids. The
effects of these reactions might be completely different if only urban surface
runoff were involved.
Combined sewer systems are generally designed to handle the peak sani-
tary sewage flow plus a part of the urban runoff flow, A common basis for
design is to provide for a maximum flow of two to four times the average
dry weather flow. This provides a margin of safety for the anticipated peak
sanitary sewage flow. On the average, therefore, overflow from a combined
sewer will occur only when the urban runoff is from one to three times
the sanitary sewage flow. Significant overflow, on the average, will occur
only for urban runoff volumes greater than these values. It is assumed
for this study that CSO's consist of primarily urban surface runoff,
Urban Surface Runoff-*-
Urban surface runoff accounts for the bulk of the solids in CSO's,
(Although concentrations are roughly equal, the surface runoff is much
greater in volume,) Colston (1) found that for one urban watershed, urban
surface runoff accounted for ninety-five percent of the total annual sus-
pended solids loading on the receiving water. Randall et al., (16) obtained
a similar result for a larger watershed comprised mainly of urbanized areas.
Because such a large percentage of the suspended solids load is contributed
by urban surface runoff, the characteristics of the solids in CSO's might
be expected to be similar to those of urban surface runoff alone.
In modeling the transport of CSO solids, two basic classes of processes
must be considered. First, transport depends on the quiescent settling char-
acteristics of the solids. Second, any time dependent changes in the settling
characteristics, mass rate, or other transport variables" must be included
in the analysis. These two classes of processes are discussed in the follow-
ing sections.
Reported Settling Data,—In terms of quantifying the physical character-
istics that relate to transport, the most direct method may be to measure
settling velocities under quiescent conditions. Unfortunately, data of this
type are limited. Dalrymple et al., (17) presented settling data for both
stormwater runoff and sanitary sewage, and these data are reproduced in
Figure 3. For the samples tested, the stormwater solids settled more slowly
22
-------
cd
Q
c
Ol
u
S-
CO
(U
3
bO
•H
CO
O
O
s/uio 'A'^
23
-------
than the sanitary sewage solids. NO information was presented characterizing
the drainage basin or indicating the condition under which the samples were
collected. Therefore, the data are probably not representative of CSO's in
general.
One important conclusion can be drawn from the data of Figure 3. The
stormwater sample was split and analyzed at two different times and an
increase in the rate of settling was observed following "aging" of the sample.
Apparently time dependent processes occurred to change the physical character-
istics of the sample. These types of proceses will very likely be important
in characterizing the transport of CSO's.
Generated settling data—Few definitive data are available describing
the settling characteristics of CSO solids. This information is essential
for developing a transport model for the materials in CSO's. Obviously one
approach could be to collect samples from each overflow of interest and char-
acterize the material in the laboratory. This would be a difficult and
expensive procedure. A second alternative would be to attempt to generate
settling data from the data that are available. An approach to generating
the required data will be presented in this section, along with an outline
of research needs to develop a procedure for estimating the characteristics
of CSO's without extensive field work for locations of interest.
The variables affecting quiescent settling of discrete nonreactive
particles are:
—particle size,
—particle specific gravity,
—particle shape,
—fluid density, and
—fluid viscosity.
These types of data are available from a variety of sources and from them it
is possible to generate data that may be at least representative of the
types of characteristics to be expected for CSO solids.
Particle size data have been obtained in several studies. Solids in
urban surface runoff originate mainly from impervious surfaces (9). Because
of the inherent difficulties in obtaining representative samples of CSO's,
a number of researchers have taken street surface sample data to be represen-
tative of CSO's.
Sartor and Boyd (14) presented data for street surface samples from five
cities. Averages of these values are presented in Table 8. Similar data
are presented in Table 9 for Chicago (15) and in. Table 10 for Washington,
B.C. (13).
Dalrymple et al., (17) presented results from previous studies for
particle size distributions in CSO's. These data are summarized in Tables
11 and 12. It should be noted that the data of Table 11 were for solids
24
-------
retained in a catch basin from which some particles especially those in the
smaller size fractions and less dense, could be lost,
TABLE 8, PARTICLE SIZE DISTRIBUTIONS FOR URBAN STREET SURFACE
PARTICULATES
Size Range
Percent Distribution by Weight
>4,800 microns
2,000 to 4,800
840 to 2,000
246 to 840
104 to 246
43 to 104
30 to- 43
14 to 30
4 to 14
<4
5.9
is; 7
14.0
22.2
17.6
11.0
7.3 .
3.8
2.1
0.4
Source: Sartor and Boyd (14)
TABLE 9. PARTICLE SIZE DISTRIBUTION FOR STREET SOLIDS SAMPLES
FROM CHICAGO, ILLINOIS
Size Range
2,000 microns
1,190 to 2,000
840 to 1,190
590 to 840
840
Percent
Commercial
Site
5.8
7,8
5.2
6.6
74.6
Distribution by
Weight
Industrial
Site
3.4
7.0
6.4
12.8
70.4
Average
4.6
7.0
5.8
9.7
72.5
Source: APWA, 1969 (18); from Manning et al., (6)
The data of Tables 8 through 13 are presented graphically in Figure 4.
In constructing this figure, it was assumed that the geometric mean of each
size range represented the midpoint of the weight of material within the range.
For example, in Table 8, the size fraction 2,000 to 4,800 microns contained
15.7 percent of the sample by weight with 5,9 percent of the sample larger
than 4,800 microns. It was then computed that:
100 - (15.7/2) - 5.9 = 86.2 percent
25
-------
of the sample had sizes smaller than
/(2.000) (4,800) = 3,100 microns.
It is useful for reference to later sections of the report that 1 mm = 1,000
microns; thus, 3,100 microns = 3.1 mm.
TABLE 10. PARTICLE SIZE DISTRIBUTION FOR STREET SOLIDS SAMPLES
FROM WASHINGTON, D.C.
Size Range
1,700 to 3,350
microns
850 to 1,700
420 to 850
250 to 420
150 to 250
75 to 150
45 to 75
45
Arterial
Roadway
3.2
7.1
19.4
25.2
19.1
17.6
7.6
0,6
Percent
Urban
Highway
8.7
9.6
14.4
14.3
12.3
17.2
13.4
10.0
Distribution
Shopping
Center
1.8
6.3
19.7
25.4
15.4
16,4
10.8
4,3
by Weight
Commercial
Street
5.5
8.0
18.6
23.0
16.3
17.0
10,6
1.0
Average
4.8
7.8
18.0
22.0
15.8
17.0
10.7
4.0
Source: Shaheen, (13); from Manning et al., (6)
TABLE 11. PARTICLE SIZE DISTRIBUTION OF SUSPENDED SOLIDS IN CSO's
IN LANCASTER, PENNSYLVANIA
Size Range
Percent Distribution by Weight
9,525 microns
4,760 to 9,525
2,000 to 4,760
1,190 to 2,000
590 to 1,190
420 to 590
210 to 420
149 to 210
74 to 149
44 to 74
44
1.77
1.06
1.40
1.88
3.10
2.78
7.01
5.19
20.10
23.80
31.90
Note: These data represent material retained in a catch basin rather than
actual CSO's.
Source: Krants and Russell, (19); from-Dalrymple et al., (17)
26
-------
TABLE 12. PARTICLE SIZE DISTRIBUTION OF SUSPENDED SOLIDS IN CSO's
IN SAN FRANCISCO, CALIFORNIA
Size Range
Percent Distribution by Weight
3,327 microns
991 to 3,327
295 to 991
74 to 295
74
5.1
8.8
15.9
21.8
48.3
Source: Envirogencies Co
TABLE 13. PARTICLE SIZE DISTRIBUTION IN PERCENT OF SUSPENDED
SOLIDS IN CSO's in SAN FRANCISCO, CALIFORNIA BY
CATCHMENT LOCATION
Catchment Location
>75y
14-75y
5-14y
61.9
26.9
7,6
0,45-5y
Baker Street
Mariposa Street
Brotherhood Way
Vicente St. North
Vicente St. South
44.6
20.0
79.7
78.7
86.2
32.6
74,1
8.6
11.2
7.6
17.0
3.4
6.1
6.4
4.8
11.8
4.0
7.8
3.5
4.4
6.4
Source: Huber et al., (15)
Inspection of Figure 4 indicates reasonable correspondence between the
size distributions of CSO solids and street surface solids. The data of
Tables 8 and 11 more or less bound the distribution curves. It was noted in
Table 11 that those data represent solids retained in a catch basin. It is
reasonable to expect that some of the smaller sized particles were washed out
of the basin. Removing more fines from the sample would shift the distribu-
tion curve up to the left. The size distribution of the original CSO solids
would probably fall to the right of the curve shown. As an initial estimate,
it is reasonable to assume that CSO solids and street surface solids have the
same characteristics, at least when the solids are first picked up by the
rainfall generated runoff.
Several researchers have presented size distributions for certain materials
found in street surface solids. Data from two of these types are presented in
27
-------
id4
1C
I03
c
s_
u
01
'o
102
1O1
KEY
CSO SOLIDS
STREET SURFACE SOLIDS
[3 Table ff Data
A Table 9 Data
Q Table 1«Pata
<£> Table 11 Data
N7 Table 1* Data
10 20 30 40 50 60 7O 80 90 95 98
Percent by Weight. Less Than Corresponding Size
Figure 4. Particle Size Distributions of CSO and Street Surface
Solids
28
-------
Tables 14 and 15. The data for Table 14 were from a large number of cities,
while the data for Table 15 were for several types of roadways in Washington,
D.C. Selected data from these tables were plotted on probability paper in
an attempt to find any trends. Several of the plots are shown in Figure 5.
The close agreement with linearity in all cases indicates that the constitu-
ents shown, over the size ranges of the available data, have approximately
log-normal size distributions. Other constituents from the two data sets
were tested in a similar way and in nearly all cases the agreement with
linearity was good.
TABLE 14, FRACTION OF CONSTITUENT ASSOCIATED WITH STREET SOLIDS
PARTICLE SIZE RANGES
Total Solids
VS
BOD5
COD
TKN
N03
P°4
THM*
T Pesticides
Cr
Cu
Zn
Ni
Hg
Pb
Particle
>2,000 840-2
24.4% 7.
11.0 17.
7.4 20.
2.4 2.
9.9 11.
8.6 6.
0 0.
16.3 17,
0 16.
26.1 13.
22.5 20.
4.9 25,
26.2 14.
16.4 28.
1.7 2.
,000
6
4
1
5
6
5
9
5
0
6
0
9
2
8
6
Size, Microns
246-840
24.
12.
15.
13.
20.
7.
6.
14.
26.
16.
16.
16.
15.
16.
8.
6
0
7
0
0
9
9
9
5
3
5
0
3
4
7
104-246
27.
16.
15.
12.
20.
16.
6.
23.
. 25.
16.
19,
26.
17,
19.
42.
8
1
2
4
2
7
4
5
8
3
0
6
2
2
5
43-104
9.
17.
17.
45.
19.
28.
29.
7
9
3
0
6
4
6
27.
31.
27.
22.
26.
27.
19.
44.
<43
5.
25.
24.
22.
18.
31.
56.
8
7
7
0
6
1
2
5
9
6
3
7
7
9
2
*Total Heavy Metals
Source: Sarton and Boyd (14) •
These observations tend to support the hypothesis that the size distri-
bution of street solids are log-normal distributions. The mean, of course,
would be a function of the geographical location. It might also be hypothe-
sized that the size distributions are relatively independent of geographical
locations and dependent on such variables as traffic volume and street sweep-
ing schedules. If these hypotheses could be verified, it should then be
possible to generate valuable CSO transport data in terms of particle size
distributions as a function of the above mentioned variables. These con-
cepts are discussed in more detail in Appendix A,
29
-------
1O3
0)
.a
£
in
10
1O1
KEY
- Table 14 Data
- Table 15 Data
<$> Volatile Solids
O
A
Metals
Total Kjeldahl Nitrogen
10 20 3O 4O 5O 6O 7O 80
Percent Less Than Corresponding Size
9O 95 93
Figure 5. Size Distribution of Selected Constituents of Street
Surface Sol ids
30
-------
TABLE 15. FRACTION OF CONSTITUENT ASSOCIATED WITH STREET SOLIDS
PARTICLE SIZE RANGES (WASHINGTON, D.C.)
Dust and Dirt
Volatile Solids
BOD
COD
PO, as P
NO as N
NO as N
Total Kjeldahl
Grease
Petroleum
Asbestos
Rubber
Pb
Cr
Ni
Zn
Cn
Particle Size, Microns
3,350-850
13,6
17.0
14.7
12.8
9.4
13.6
24.5
N 20.1
11.6
10.8
13.0
3.0
6.5
16.8
25.9
7.2
8.1
850-420
19.8
14.8
16.7
13.4
13.7
13.3
13.6
26.0
10.3
9.1
15.5
5.4
18.3
13.2
11.8
13.9
11.7
420-250
23.9
13.1
20.2
14.7
19.3
15.7
9.8
17.5
12.5
12.5
20.5
11.3
15.5
16.6
16.2
24.9
13.8
250-75
31.3
33.33
29.4
36.8
37.9
36.1
21.8
23.9
40.1
39.9
39.6
37.8
42.8
36.8
29.4
40.4
44.2
<75
11.4
21.8
19.0
22.3
19.7
21.4
30.3
12.6
25.5
27.7
11.4
42.5
16.9
16.6
16.7
13.6
22.2
Source: Shaheen, (13); from Manning et al., (6)
Particle specific gravity depends on the composition of the particle.
CSO particles probably tend to change in composition as was indicated by
the settling data of Dalrymple et al., (15) presented earlier. Many processes
such as agglomeration, adsorption, and biological degradation may occur.
These processes are time dependent as. well as being functions of variables
other than time. As a first approximation, it is assumed that CSO particles
are discrete and inert. This assumption is obviously not valid. The amount
of error introduced cannot be determined without obtaining additional data.
As a first approximation, the assumption is considered reasonable.
Specific gravity data for materials that might be expected to be found
in street surface (and thus CSO) samples, are given in ,Table 16. In order
to use these data to derive settling data, knowledge of the relative amounts
of the various materials is necessary.
Tables 14 and 15 list some of the materials that might be expected to
be found in CSO's, but they do not indicate the relative amounts. Manning
et al., (6) compiled available data on the loading on street surfaces of many
of these types of materials and some of this information is summarized in
31
-------
TABLE 16. TYPICAL SPECIFIC GRAVITIES OF MATERIAL IN STREET
SURFACE SOLIDS
Material
Specifc Gravity
Material
Specific Gravity
Asphalt
Brick
Cardboard
Cement
Clay
Creosote
Glass
Paper
Rubber, hard
Silica
Tar
Wood
1.1 to 1.5
1.4 to 2.2
0.69
2.7 to 3.0
1.8 to 2 6
1.04 to 1.10
2.4 to 2.8
0.7 to 1.15
1.19
2.07 to 2.21
1.02
0.11 to 1.33
Cr
Cu
Fe
Hg
Mn
Ni
Pb
Zn
7.18 to 7
8.96
7.87
13.55
7.21 to 7
8.90
11.35
7.13
.20
.44
Source: CRC, (21)
Table 17. Data were collected for a variety of other materials, but the
loadings were extremely small in most cases and much less than those listed
in all cases. What Table 17 indicates is that the bulk of street surface
solids fall into categories other than those listed. A more complete anal-
ysis of the makeup of street surface solids is necessary in order to apply
the available density data in a reasonable way to predict settling velocities.
TABLE 17. AVERAGE DAILY ACCUMULATION ON ROADWAYS IN MATERIALS IN
THE "DUST AND DIRT" FRACTION
Material
Accumulation, kg/curb-km-day
BOD
COD
Total N as N
Total Kjeldahl N
Cr
Fe
Mn
Pb
Zn
Other*
Total Dust and Dirt
0.23
2.08
0.02
0.03
0.01
0.95
0.02
0.09
0.02
42.00
45,0
*Estimated by author
Computed from data summary of Manning et al., (6)
Sources: APWA, (198); Shahee, (13); Sharton and Boyd (14); and
Amy et al., (9)
32
-------
In an early study, street surface samples were collected from twenty
test sites in Chicago and analyzed for seven classes of materials. A summary
of the results is presented in Table 18. In the same study, the specific
gravities of street surface samples were determined, and the results of which
are given in Table 19. The portion of the street surface material classified
in this study consists of particles larger than about 3 mm, and the "dirt"
fraction was not based on composition but was made up of the material less
than about 3 mm in size.
Data on the composition of the small sized street surface particles that
will make up CSO solids are incomplete. It will be assumed, for the purpose
of this study, that the makeup of this material is the same as that given
in Table 18, neglecting the "dirt" fraction. From the limited data available,
this appears reasonable. For example, from Table 17, the accumulation of
metals is 2,4 percent of the total which corresponds to the value in Table
18. The actual material composition may be a function of such variables as
geographical location and time of year. Development of an empirical relation-
ship to predict the makeup of CSO's would be a valuable contribution. Dis-
cussion of this concept is included in Appendix B.
TABLE 18.- SUMMARY OF DATA ON COMPONENTS OF STREET LITTER IN
CHICAGO, ILLINOIS
Material Class
Percent Distribution by Weight
Rock
Metal
Paper
Dirt
Vegetation
Wood
Glass
Other
10.3
2.4
7.6
59,6
13.6
0.5
1,9
4,1
Computed from data of APWA, (15); from Manning et al., (6)
TABLE 19. SPECIFIC GRAVITY OF "DUST AND DIRT" STREET SURFACE
SAMPLES IN CHICAGO, ILLINOIS
Land Use
Specific Gravity Range
Commercial
Industrial
2.2 to 3.0
2.5 to 2.6
Source: APWA (18); from Manning et al,, (6)
Taking the data of Table 18 as being representative of CSO's in general,
specific gravities of the fractions can be estimated or taken directly from
Table 16. Data considered representative of CSO's for the purpose of this
33
-------
study, in terms of composition and density, are given in Table 20, The
specific gravity values are estimates and are probably variable. The weighted
average specific gravity is less than the values given in Table 19 and a
higher value may be more appropriate in some cases.
TABLE 20. PREDICTED DATA FOR COMPOSITION AND SPECIFIC GRAVITY OF
CSO'S
Material Class
Rock
Metal
Paper
Vegetation
Wood
Glass
Other*
Percent Distribution
by Weight
25.5
5.9
18.8
33.7
1.2
4.7
10.2
Specific
Gravity
2.6
8.0
1,1
1.1
0.9
2.6
1.2
100.0 Weighted Average =2.0
*Assumed to be rubber
Particle shape is the third particle variable that must be considered
in computing settling velocities. Particle shape is used in computing the
drag coefficient which is incorporated into a force balance on the particle.
A relationship for the drag coefficient on spherical particles has been
presented (22) as:
Where:
N.
+ 0.34
R
(1)
C_ = drag coefficient
K, = Reynolds number = ——
R J p
V = fall velocity, ft/sec 2 ^
p = particle density, Ib-sec /ft
d = particle diameter,ft „
p — fluid dynamic viscosity, Ib-sec/ft
4
This relationship is reported to be valid to a Reynolds number of about 10 .
A relationship that appears to fit the experimental data more closely (23)
is:
r - . 4.
~
+ 0.26
(2)
34
-------
Experimental data are available for two other idealized shapes, discs, and
cylinders. Relationships that fit these data reasonably well are:
Discs: C =
24
N.
R
+ 1,14
(3)
Cylinders:
10
CNR)
0,289
1.14
(4)
Portions of the CSO particles will fall into each of these classes, and
portions will not be described adequately by any of the three idealized
shapes. Initially, it is probably reasonable to assume that all the particles
are spherical. As data are collected, this assumption may be modified as
necessary.
Fluid density and fluid viscosity depend primarily on the temperature
of the fluid and the suspended solids load. In most cases it is probably
sufficiently accurate to consider these variables to be functions of tern-,
perature only. An error in the density value of five percent corresponds
to a suspended solids load on the order of 10 mg/£, a solution of about ten
percent solids. The error in assuming the fluid transporting CSO solids to
be pure water will, in most cases, be negligible.
Time can be an important factor in modeling the transport of CSO solids.
The effects of time on settling characterisitcs as indicated earlier (17),
may be significant. Time may be important is establishing the characteristics
of the street surface solids that constitute CSO solids as indicated previously
and discussed in the appendices. However, initial estimates for settling
characteristics of CSO solids were given in the previous section under the
assumption that time dependent changes are negligible.
Storm events that result in the urban surface runoff are time dependent
as indicated in the first part of this subsection. The time of year may be
important in areas where the frequency and magnitude of rainfall vary
seasonally. Also, the type of ground cover on previous surfaces will influ-
ence the amount of runoff generated as well as its size distribution and
the ground cover varies throughout the year, Time of year and frequency
of rainfall, as well as other time related variables will have an effect
on the makeup and physical characteristics of CSO solids, as discussed in
the appendices. For :the purpose of the model study presented in Section
4, all of these time related variables will be neglected.
The concentrations of most, if not all, of the constituents of CSO's
vary over the course of a runoff event, as demonstrated by Colston (1). The
empirical equations relating some of the constituents to the elapsed time
during a runoff event were presented earlier in Table 5. Because the concen-
tration may vary markedly, the time dependent nature of storm events has to
be incorporated into the transport model.
35
-------
Several workers have presented data for the variations of constituent
concentrations in CSO's during runoff period. Some of the data for suspended
solids are presented in Figure 6 along with three hypothetical curves obtained
from Table 5. It is quite evident that the suspended solids concentrations,
and therefore loads, vary with time. The equation obtained by Colston does
not appear to fit these data as there is evidently a more rapid tailing off
of the concentration of suspended solids in all cases than predicted by the
hypothetical equation. Developing a more raliable equation would require
collecting a significant amount of additional field data,
One equation that qualitatively fits the limited data available is:
°'85
Where:
TSS - 35Z
TSS = total suspended solids, mg/£
Q = flow rate, cfs
t = time from storms start, hrs
(5)
Also,
1'85 -1'
Where:
TSS = 1.6 Q
TSS = total suspended solids, Ib/hr
(6)
As a first estimate, lacking additional data, either these equations or the
equations obtained from Table 5 may be used to estimate the time dependence
of suspended solids in CSO's. However, when the regression analysis work
on the data by Huber et al., (15) is completed, better data should be
available for the time dependence of a variety of water quality parameters
in CSO's.
Summary of Characteristics
Combined sewer overflows are made up of urban surface runoff and sani-
tary sewage, The contribution of sanitary sewage to the total flow may be
negligible in -many cases, although its effect on time dependent changes in
the CSOrs -may be important.
Urban surface runoff makes up the majority of a CSO in most cases.
The characteristics of urban surface solids may, as a first approximation,
be considered similar to those of street surface solids.
THE RECEIVING WATER ENVIRONMENT
Introduction
In the previous subsection the characteristics of CSO sediments relative
to transport were considered. In this subsection, the impact of these
sediments on the receiving water and vice-versa will be considered. The
36
-------
Ill
u
o:
s-
3 CO
O +•>
i— tn
•-E
-0 +-> T-
Q
in
ro
CM E
o
8
CM
O
0
m
o
o
o
§
CO
o
C_>
0
CO
a
o
•rl
J-l
IS'
JJ
§
o
a
o
CO
t)
•H
O
CTJ
CU
-d
M-l
o
CD
o
0
0)
T3
•8
Pn
Q)
a)
VD
0)
t>0
•H
L/t5ui
papuadsnj
37
-------
chemical aspects of the interaction between the receiving water and the sedi-
ment material will be presented first. Then the potential impact of the
sediment material on the biologic community will be presented. The last part
in this section will present an approach for assessment of the environmental
impact of the combined sewer overflow sediments.
The content of this subsection is more speculative in nature. Very
little, if any, data have been collected to specifically identify the inter-
action between sediment materials and the environment. The goal is to
identify areas where knowledge exists and areas where more knowledge would
be beneficial.
Potential Impacts
Suspended Solids—
One of the readily identified impacts of CSO's in the receiving waters
is the increase in suspended solids load. Randall et al., (16) quantified
sediment loads from runoff events and found that for the particular drainage
basin studies, only about one percent of the total amount of suspended solids
load transported by the receiving stream was associated with base flows.
The remaining 99 percent was associated with runoff events.
Solids associated with CSO's can have direct impact on receiving water
qualtiy by decreasing the depth of light penetration and thereby decreasing
photosynthesis by modifying the thermal and hydraulic characteristics of
the stream and by physically interfering with biological processes (for
instance, by burying the bottom dwelling or benthic organisms). CSO solids
generally include materials such as heavy metals, pesticides, and organics
which can affect the biology and chemistry of the receiving water. These
types of impacts, if they occur, may be much more significant than the
direct physical impact of the solids themselves. Impact analysis is some-
what complex, however, since the mere presence of a material does not
necessarily imply that is has an impact.
Physical and Chemical Impacts—
Adverse impacts of CSO's could include physical and chemical impacts.
In evaluating potential impacts, three types of relationships are quite
important. These are:
(a) the specific form of the material or characteristic of the parameter
responsible for the impact,
(b) the magnitude-duration of impact relationships, and
(c) the environmental fate-time relationship for the material on
parameter of interest.
An example of these types of relationships is the following:
A particular sediment size fraction (a) may have an adverse impact
on algal production and fish life when above a certain concentration
for a certain amount of time (b), and may be removed from suspension
by physical transport processes at a certain rate (c). Relationships
38
-------
of these types are important in assessing the potential impacts of
any parameter of concern. Unfortunately, few such relationships
have been investigated.
Physical impacts of CSO's could include hydraulic effects and impacts
due to sudden increase in suspended solids loads. Increased flows could
sweep organisms, both free-swimming or floating and attached, downstream.
Increased suspended solids loads could reduce the depth of light penetration,
potentially affecting the population dynamics, and could physically bury
benthic organisms. Both transport from urban surfaces and resuspension of
bottom sediments could account for the increased suspended solids loads.
The importance of CSO's in producing these types of impacts would
depend largely on the hydrology of the drainage basin. The proportion of
the total streams flow contributed by CSO's would be an important variable.
In assessing the physical impacts of CSO's on receiving waters, the actual
effects on the ecosystem should be determined in terms of the three types
of relationships discussed earlier.
Adverse chemical impacts could include detrimental effects to the eco-
system of the receiving waters, a fish kill, for example, and effects
impairing a downstream beneficial use, such as an increased salt content
making the water unacceptable for irrigation. It is well known that the
types of relationships involved in creating the adverse impact, listed
earlier, are important here. For instance, only certain chemical forms of
particular materials affect organisms. The organisms must be exposed to
certain concentrations for certain time periods, and the chemical forms
can change over time to forms which may have increased or decreased effect.
These types of phenomena should be considered in predicting or assessing
impacts.
Interaction of GSO Material With Receiving Water
Chemical interaction of CSO material and the receiving water depends
on the types of material present. Numerous kinds of materials have been
identified in CSO's and data characterizing CSO's in terms of composition
are presented in the previous subsection of this report. In terms of
overall transport, CSO's are characterized by the mass of material in each
of seven categories: rock, metal, paper, vegetation* wood, glass, and
rubber.
Chemical Composition of CSO Solids—
Data have not been collected to completely describe the chemical compo-
sition of CSO solids. The composition would vary over time and from location
to location, so data of this type would probably not be useful if they were
available. A reasonable approach is probably to describe CSO's in terms of
the seven classes of material listed above and then to characterize a par-
ticular CSO in terms of any individual materials of interest by onsite
sampling and analysis.
39
-------
The material characteristics that may be important include:
—particle size,
—particle shape,
—particle charge,
—material density,
—material coordination number,
—oxidation state, and
—material concentration.
Receiving Water Characteristics—
The general receiving water characteristics that may be important
include:
~pH,
—temperature,
—hardness,
—alkalinity,
—nutrient level,
—dissolved oxygen level,
—oxidation-reduction potential,
—turbidity,
—density, and
-T-viscosity.
Potential Interactions—
Reactions that will be important in governing the transport of CSO
materials in the receiving waters will include those that change the physical
characteristics of individual particles. These reaction mechanisms will
include:
—sorption,
—solubility,
—oxidation-reduct ion,
—complexation, and
—biochemical reactions.
In modeling the transport of discrete, inert particles, only the physical
characteristics of the particles are important. The transport variable of
most importance is the particle fall velocity. In the case of CSO's addi-
tional transport variables must be considered. These might be considred
as "process" variables and include particle cohesiveness and particle
degradation of aggradation potential. These process variables involve the
reaction mechanisms previously listed but are considered separated partly
because they will be more easily observed.
Models of sediment transport incorporate certain physical characteristics
of the river or stream. These characteristics include:
—flow velocity,
—flow depth, and
—channel dimensions.
40
-------
Interaction Process Description—
Having identified most of the important variables involved in transport
of CSO solids, the next step is to organize these variables into a systematic
process description. Simplifying assumptions, in addition to those made in
selecting the important variables, might then be made in a logical way, with
some understanding of their effects on the final results.
The transport process is depicted schematically in Figure 7, The solids
and the receiving water interact according to the reaction mechanisms identi-
fied previously to yield a variety of types of particle physical character-
istics. These particle characteristics are further changed by the physical
characteristics of the receiving water body to result in the transport
variables. The transport variables govern the resulting CSO solids transport.
It is not known at this time if CSO solids transport differs to any great
degree from conventional sediment transport.
The interaction between the material and water characteristics and
the reaction mechanisms may perhaps be seen more easily in the form of a
matrix. Such a matrix is shown in Figure 8, which estimates of the relative
importance of the various interactions are given. These estimates are quite
important because they could serve as a basis for simplifying an obviously
complex system, and their validity should be investigated further. There
will be other interactions to consider for each material (such as whether
a particular reaction involving one material or water characteristic will
affect another characteristic) as well as interactions among the various
materials.
The reaction matrix is sufficiently complex and data so totally lacking
that at this time the best approach seems to be to ignore the complexity
altogether. A good data set will have to be collected which can be analyzed,
to determine if it is adequate to assume that all CSO material is noncohesive
and inert. If such an assumption is not adequate, then the interaction
matrix approach may be used to investigate better assumptions.
Impact Analysis—
In general, the criteria for identifying streams which are potentially
hazardous with, respect to the deposition of CSO material are not simple and
clearcut. More often than not, a hazardous situation will involve the
combined action of several factors. Taken individually, none of the factors
might cause objectionable conditions; taken collectively, however, the
addition or subtraction of a single factor might radically alter the entire
picture. '
The development of a hazardous situation obviously implies the require-
ments that there be (1) uptake and concentration of significant -quantities
of pollutants by CSO and stream sediments, (2) stream geometry and sediment
conditions conducive to extensive local deposits of contaminated sediments,
and (3) a rate of buildup of pollutants in deposits which exceeds the rate
of natural reentrainment and dispersion..
41
-------
Material
Charact eris tic s
(M±)
Water
Characteristics
(M±)
Reaction
Mechanisms
(b.)
Transport
Variables
CA±)
CSO Solids
Transport
Figure 7. Schematic Diagram of CSO Solids Transport Process
42
-------
rH
Cd
O
•ri
6
o
O
•H
pq
•H
4-1
cti
H
&
U
O
cu
•H
•H
fi
O
•H
4-1
PL,
O
Cft
•H
^ti C^ C.3 ^^ ^^ ^ti •^ri
rH rH CO rH rH rH rH
CJ C_> O O O O pq
rH tH CM CM CM CO CO
O
rH CO CO CO CO rH CO
pq O pq pq pq U pq
CM rH CM CM CM CM CO
pq o pq u o u
rH H rH CM CO CO
rJ 4-1 co
4-1 CO -H 4-1
Cti CO fi fi
!H Q) -H (U
CU fi rH -H
P. 13 cd !-<
Q W ^ 4-1
CU Cti rH 3
, H K <1 a
•^ri -^ri ^j fV) -^ri rQ ^^
H rH rH rH CO CO rH
O U U
rH rH rH CO CO rH rH •
~. 4-1
riS Cfl f-J
fi O
13 CO
CU !>^ C
4-1 Cd O S
0 |5
fll I ji rft
VJJ r^ -I— I UJ
"4-1 Cd fi 4J
IH cu 0
cd 4-1 13 >~, cu
co c rH m
co o a) cu m
•H 6 p, M cd
rH CU Cd
C cd -d w 0
o o
•H III -H
4-1 4-1
O rH CM CO O
Cd Cd
cu cu
P3 fv|
• •
K^>
CU
w
rH
cd
•H
H rl
CU
4-1
cd
e
CO
H ^ fi
cd o
15
rH 4-J
c^ pj
QJ
4_J rj3 ^
CO pj rH
§flj QJ
fX S-i
rH Q) Cd
cd nd S-i
i I 1
<3 PQ O
CO
fi
o
•H
4-1
o
cd
%
4-1
fi
M
rH
cti
•H
S-t
0)
4-1
3
CO
6
CO
•rH
fi
o
cu
fi
o
•H
4-1
o
cd
cu
43
-------
With respect to the sediment itself, the total available surface area
of the suspended-sediment particles is the most obvious index of hazard
potential. Concentration and size distribution are fairly reliable indicators
of surface area.
Fortunately, the greater sorption capacity of the fine sediments is to
some extent offset by the tendency of colloidal materials to remain in sus-
pension except when flocculation occurs. Heavy concentrations of contaminated
fine sediments might occur in density currents at the bottom of reservoirs
and pools.
The concentration and load of suspended sediment in streams can be
considered one of the important criteria for indentifying potentially
hazardous streams because of the effect on amount of available surface area
for sorption. Further, it is important as a factor in the nature and extent
of deposition and erosion within the stream system.
The nature of the concentration and (or) load variation with respect
to time may be important because it is expected to indicate the capacity
or lack of capacity of the stream to transport a given input of waste. For
example, a stream with a relatively uniform water discharge and high concen-
tration of fine sediment would probably maintain the waste in suspension on,
or with, the fine sediment. On the other hand, a stream with a large range
of water discharge and sediment load receiving the same concentration of
waste would be expected to carry the waste in "slugs" because the waste not
having an abundance of fine suspended sediment for sorption during low flow
periods would probably be sorbed on the particles in the stream bed. Then,
during the relatively short periods of storm runoff with high concentrations
of sediment, the waste held on the bed may be exchanged to "free" suspended
particles and (or) moved downstream considerable distances as bedload.
Conditions conducive to the formation of sediment deposits containing
large amounts of CSO discharge wastes are more likely to be found on the
inside of a bend on a stream near the source of waste disposal rather than on
the banks of a straight stream channel, on a stream exhibiting a considerable
range of sediment concentration and water discharge rather than a stream of
nearly uniform water discharge and sediment concentration, and on a stream
with man-made channel controls rather than on an uncontrolled stream. (The
Cuyahoga River, evaluated in the next section of the report, is a good
example of a stream with man-made controls.)
Sediment of the coarser fractions, usually deposited on the point bar
on the inside of a stream bend, is derived mostly from that being cut from
the outside bank of the opposite bend immediatley upstream. The finer
fractions, usually deposited only at the upper elevations of the bar, or
during the recession of the storm runoff event, may be derived from many
sources including bank cutting in the immediate channel, or gully and sheet
erosion in the upper part of the drainage basin. The point bar at the first
bend immediately downstream from a source of waste may then be built of a
large mass of contaminated sand. In this discussion, it is assumed that the
waste is sufficiently mixed with stream flow so that contact with the sediment
in transport is relatively uniform.
44
-------
Streams of widely varying discharge are conducive to bank and flood plain
deposits. The most rapid deposition on the inside bank of the meandering
channel occurs during the high rates of flow; especially so, for the fines at
the top of this deposition bank. Flood plain deposits obviously cannot occur
without the generally unusual high rates of flow. Deposits of fines along
stream banks other than point bars usually occur during the recession of a
storm event for which relatively large concentrations of fines are carried
These may be of only minor significance because they are transported and
deposited when stream flow and concentrations are considerably greater than
average and, therefore, the concentration of waste in the deposit is likely
to be low. This again assumes a uniform rate of waste disposal.
Deposits of contaminated sediment may accumulate within the controlled
channel when flow and sediment concentration are low for long periods of
time; and thus, with a constant rate of waste disposal, the concentration of
pollutants in the water and on the sediment would be high. Similar to
channel control, the most critical deposits in the pool-and-riffle type of
stream fines are deposited in the slack water resulting in an intensive
concentration of CSO material in the pool. With this type of stream the
deposit is usually removed during the rising stage of the next period of
storm runoff.
The hazard potential of CSO material in streams may be considered in view
of differences in stream flow, in channel geometry, and in the concentration
and character of sediment as they may affect deposition of sediment and
consequent waste. However, streams are very complex with respect to the mag-
nitude and variation of these features; and, therefore, classification
according to hazard potential is difficult and indeed almost remote in view
of limitations concerning basic knowledge of the relation of these features.
For example, a stream with relatively large water and sediment discharges
may be conducive to extensive deposition and erosion within the channel;
but, because of the high rates of water and sediment movement, the concen-
tration of waste may be very low. Further, a meandering channel results in
a nearly continuous, building of a deposit as a point bar, but most of the
building occurs during high rates of water discharge when the concentration
of waste in the water is low.
The sand-bed stream of good alignment and carrying a relatively uniform
water discharge and concentration of fine sediment would be the least hazard-
ous kind of stream to transport a given uniform amount of CSO material with-
out accumulating extensive and (or) concentrated deposits. Such a stream
would have relatively minor and uniformly concentrated accumulation of waste
in relation to both time and space in the bed and along the banks.
One of the most hazardous conditions may occur either on the pool-and-
riffle type of stream or a large river controlled with low dams for navi-
gation purposes. For these situations, the pool or pools immediately down-
stream from the disposal outlet may accumulate extensive deposits in the
bottom of the pools. This results from the relatively low water discharge
and the very low concentration of sediment, and consequently, a very high
45
-------
concentration of material accumulating for long periods of low water flows.
When a relatively large increase in flow occurs after weeks of months of
accumulation, most of the contaminated deposit will move downstream and thus
become available for bed, bank, or flood-plain deposition (27).
Only after the movement and resting places of the CSO sediments have been
identified will it be possible to determine biologic input, For instance,
deposition rates in shallow areas can be compared to allowable rates for fish
spawning. Deposition depths and residence thus can be used to determine if
certain plant species would proliferate or die out, A knowledge of the
oxygen demanding -material present in the sediments, along with transport
information, could be used to assess DO impact. None of these things can
be done with certainty at this time.
In the following section of the report, the current state of sediment
transport modeling is discussed. Such models, along with suitable verification
data sets, are vital to address the data question. Once transport paths and
data are identified, then impacts can be intelligently addressed.
Biologic Community of Receiving Waters
Introduction—
The purpose of this section is to identify both the plant and animal
communities near combined sewer overflows and to briefly describe some of
their relevant characteristics (28,29). The discussion will be -centered
around rivers and streams since this is where most sewer overflows are
received. Characteristics will be limited to those that may in some way
affect or be affected by sediment transport.
General Description—
The aquatic environment is comprised of water, its chemical impurities,
and its various life forms including bacteria, phytoplankton, zooplankton,
fish and benthic animals. There are a number of approaches to the systematic
description of plants and animals living in the aquatic environment. In
this section, plants will be described by categorizing them into five basic
divisions: phytoplankton, flowering plants, ferns, moss and liverworts, and
fungi. An alternate method of description that will only be given a cursory
treatment, is to categorize plants by way of zonation, i.e., what depth of
water they have adapted to. Animals will be described by their level of
development, i.e., trophic levels.
The flora and fauna of rivers change in character as one moves from the
headstream to the mature lowland river. Animals in headstreams, for example,
must either have hooked appendages or suckers or be sufficiently strong
swimmers to maintain their position against the high velocity of the water.
Animals in mature rivers do not need appendages or strong swimming muscles
but must be tolerant of turbid and silting conditions, and must be resistant
to conditions of low oxygen tension and of high and variable temperature.
Transition in river characteristics are not abrupt; however, there are
four zones that may be identified. The first is the "very rapid" zone
found generally in areas of steep gradient; the second is the "moderately
46
-------
swift" zone usually with a bottom of stones and boulders; the third is the
"moderate current" zone, usually with a gravelly bed; and finally, the
"medium to slow" zone generally found in lowlands,
From headstream to mature river, a decreasing gradient is observed;
a gradual passing occurs through the four zones. During this transition,
one notes: a decrease in current velocity; a change in the bottom deposits
as a result of the decrease in erosion and transport capabilities of the water
from large rocks inheadstreams through lesser sizes of gravel to sands, and
finally silts in the mature river; increasing average temperature; and
increasing range in daily and yearly temperature; and a decreasing dissolved
oxygen content of the water. The changes are most pronounced in the range
just below the source.
In general, the human development of urban areas is restricted to
relatively flat locations. As a result, the "moderate current" and "medium
to slow" zones will be of greatest interest.
The Plant Community—
In a simplistic model of life, plants convert inorganic nutrients into
organic matter by using energy from the sun. Plants are consumed by herbi-
vores (plant eaters) which, in turn, provide food for carnivores (meat eaters)
which die and are decomposed to, inorganic matter by bacteria and fungi.
This constant recycling of matter is known as the food web.
Plants occupy the primary positions of the food web. The main divisions
of the plant kingdom are the algae (or phytoplankton), flowering plants,
ferns, mosses and liverworts, and fungi, all of which are represented in
the aquatic environment.
Phy toplankton—
In general, these are the most important organisms in the aquatic world.
Phytoplankton are actually comprised of photosynthetic microorganisms of
which diatoms (single celled algae with hard outer coverings of silica) are
the most abundant. Phytoplankton are important because they are a primary
producer of organic matter in the pyramidic food chain (Figure 9) and are a
major source of dissolved oxygen in natural waters. Physical factors
influencing river algae are the right size of the stream, current rate,
water level, depth, temperature, light, and turbidity,
Turbidity has a great effect on algae. Studies in Georgia and Alabama
compared algal populations in a very turbid stream and in a relatively clear,
associated stream. The algal population in the turbid stream was between
126 and 422 cells/-m£, and the genera did not exceed two. In the clear stream
sampled the same day, the algal population numbered 6,075 cells/mil and
genera, excluding diatoms, numbered 10. Increased complexity in a food web
can create an increased stability of a stream community by offering alterna-
tive food sources to consumers; thus excessive turbidity can destroy this
complexity by decreasing the diversity of species. Turbidity is associated
with the extremely fine clay range of sediment sizes. This is an area where
modeling capabilities are lacking at this time.
47
-------
*0uaternary Consumers
Secondary
Consumers
vo
Primary Consumers: \v>
Protozoa
Primary Producers:
Algae and Green.Plants
DECOMPOSERS
Figure 9. The Pyramidic Food Chain (28)
Flowering plants—
Flowering plants are another major primary producer. It is convenient
to categorize flowering plants into three types: (1) emergent plants, (2)
floating plants, and (3) submersed plants. Plants that grow in the soil
with only their lower portions submersed are called emergent vegetation.
Less of the plant body is supported by buoyancy and more synthetic activity
is invested in the buildup of supporting tissues containing cellulose. These
plants decay at a lower rate since most aquatic herbivores cannot metabolize
cellulose directly. Examples of emergent vegetation include cattails, arrow
arum, wild calla, and arrowhead.
Floating plants are generally rooted with just their leaves 'on the
water's surface, but a few are actually free-floating. Examples include
duckweeds, wolfiella, and hyacinths.
Submersed plants are those that are completely covered with water. They
form large, dense masses, especially in late summer, that are extemely
important as a source of shelter, food, and materials for nest building
for life beneath the surface. Examples include the water cress, water worts,
alligator weeds, mare's tail, and beggar-ticks.
The abundance of aquatic plants is very important to the productivity
of natural water. The existence of invertebrates is closely related to
aquatic plants; they tend to select those plants with compact, finely
branched leaves. So important are aquatic plants that their abundance can
be used as an "index of productivity" for fish.
48
-------
TKe interaction between the plant community and sediment transport is
intuitively obvious but poorly investigated. Plant species are one important
component of resistance to flow. Conversely, sediment deposits can smother
low growing vegetation. Turbidity can shut off light supplies for photo-
synthetic activity. No quantitative studies of these interactions exist.
Ferns—
The third major plant division is the ferns. Ferns are a widespread
class of nonflowering plants having roots, stems, and fronds (leaves) that
reproduce by spores instead of seeds. Aquatic examples include the quill-
wort, pillwort, and horsetail. The rest of the 8,000 species of ferns are
shade-loving terrestrial plants.
Mosses and liverworts—
These are occasionally called the amphibians of the plant world. They
are more advanced than the algae but do not have conducting and stiffening
tissue or true roots as do the higher plants. They often form extensive
"green carpets" where the substream is rocky or stony and provide a foot-
hold for animals which would otherwise be swept away by the current. The
water must be relatively clear for them to carry on photosynthesis at the
bottom of the stream. They are thus quite sensitive to turbidity or buried
by large scale deposition.
Fungi—
The fifth division is represented by many small and inconspicuous
species including molds, mildews, mushrooms, rusts, and smuts. These are
all parasites on living organisms or. feed upon dead organic material. Thus,
they play an important part in the decomposition of organic matter and as*
disease-producers in higher plant forms.
The Animal Community—
Recent literature has begun to abandon the rather simplistic "food
chain" concept and is adopting what is called the "food web." This latter
term suggests a more complex interaction and is shown in a simplified
matter in Figure 10 (29). The animal kingdom and its members appear near the
bottom of the food web. The members will be discussed individually.
Zooplankton—
At the lowest level of development in the animal community are the zoo-
plankton or the animal portion of the plankton. This term includes protozoa
rotifers, and other microscopic creatures. As with their plant counterparts,
zooplankton are of greatest importance in large, slow-moving rivers because
of the tendency of rapid current to wash them downstream.
Rotifers—
Rotifers make up the greatest share of zooplankton. They feed mostly
on minute organic particles; however, a few are predatory on other organisms
and a few feed on the fluid contents of filamentous algal cells. Rotifers
are largely associated with the substrata.
49
-------
MAN-INDUCED
WASTE LOADS
PHYTOPLANKTON AND
GREEN PLANTS
.BENTHIC
ANIMAL
FOOD
FISH
Figure 10. Definition of Aquatic Ecosystem (29)
50
-------
Protozoa—
The Protozoa are perhaps the least understood group of. aquatic animals.
A majority of Protozoa are attached to the substrata and are particularly
abundant in habitats of active decomposition. They are extremely important
as a major metabolizing organism of dissolved and particulate organic matter
in sewage treatment facilities and organically polluted streams. The very
large and varied feeding capabilities (algae, bacteria, particulate deritus,
and other protozoa) together with their large population densities on aerobic,
organic-rich sediments are all indicative of their significant metabolic role
in freshwater systems.
Planktonic crustaceans—
The groups Cladocerans and Copepods nearly dominate the truly planktonic
(freeliving) crustaceans. The seed shrimp is an example of a crustacean
that is primarily a benthic animal* They resemble tiny clams and are usually
around 1 to 3 min long. For the most part, seed shrimp are scavengers, filter-
ing the water to remove bacteria, molds, algae, and general detritus. They
generally creep about in the algae or over vegetation, or burrow in the
mud and ooze of the bottom. Many can survive long periods of stagnation with
little oxygen. Crustaceans are likely to be associated with backwater
areas and slow moving portions of rivers where fine material deposits form.
Mollusca
The phylum Mollusca, which includes the chitons, oysters, clams, mussels,
snails, whelks, slugs, and other invertebrates, is one of the least advanced
phylums but is very important in the aquatic environment. They make up
a large percentage of the benthic animals and together with insect larvae and
nymphs, crustaceans, worms, leeches, and sponges make up the bulk of the
intermediate trophic levels. The largest number of benthic animals are
found above the compensation level (where photosynthetic activity is equal
to respiration).
Snails and limpets—
Snails and limpets move by sliding across the substratum and maintain
their position against the force of current by viscosity. Snails and lim-
pets are able to remain submerged indefinitely. They feed on the micro-
scopic algae that coat submerged surfaces, filamentous algae, and other
green vegetation as well as on dead plant and animal matter (detritus).
They require a high oxygen content (limpets even require oxygen-saturated
water) so are seldom found in polluted water. They are also rare in swift
streams where the bottom is sandy and gravelly.
Bivalves—
The bivalves include clams and mussels and are most common- in the muddy
bottoms of large rivers. They range in size from 2 mm to 250 mm. Bivalves
move in much the same way as snails, and feed on microscopic plankton and
microscopic organic debris washed into the water. Fish include them in
their diet.
51
-------
Worms—
Small, often microscopic worms crawling over vegetation or debris, make
up another large portion of the- benthic animals and are of tremendous impor-
tance in the economic life of any body of water. They are scavengers and
feed on bits of organic material or upon small plants, transforming these
substances into animal tissue that can nourish larger invertebrates and even
small vertebrates. Occasionally, worms will feed directly upon other small
animals. They are preyed upon by other worms, crustaceans, insects, and
fish. Most worms are negatively phototactic and thus occur as much as 4 cm
under sediments and debris and in shaded parts near rocky substrata and macro-
phytes. A number of worms, in particular flatworms and nematodes, greatly
increase as a stream or lake becomes polluted with organic matter.
In the Milwaukee River at Milwaukee, Wisconsin, sludgeworms attained
populations of 84,000 per square foot of river bottom due to an inexhaustible
organic food supply. Immature stages of mayflies, caddises, and hellgrammites,
were eliminated. In the Brule River bordering Michigan and Wisconsin, where
man-associated organic wastes are not a problem, clean water larval caddis-
fly populations number around 1,100 per square foot.
few
Sponges—
The sponge is a plantlike animal usually found in the ocean. A
freshwater species do exist—however, when conditions are favorable.
Sponges feed by maintaining a current of water through their bodies by means
of flagella. Microscopic organisms are filtered out and used.
Insects
Insects are animals of the phylum arthropoda, having six legs and
usually one or two pairs of wings. Insects are a mjor importance in the
balance of nature. They aid bacteria and fungi in the decomposition of
organic matter and in soil formation. For example, the decay of carrion
brought about mainly by bacteria, is accelerated by the maggots (larva) of
flesh flies and blowflies.
Insects that have adapted to the aquatic world have had to make sub-
stantial modifications in their respriration in order to survive. Many rise
to the water surfaace and take air into their tracheal systems9 some prolong
submergence by trapping air among their surface hairs, and other insects
have adapted to the point of being able to obtain all of their oxygen from the
water (e.g., midge larvae, amyfly larvae, and dragonfly larvae).
Spiders and mites—
Arachnids (i.e., spiders and mites) differ from arthropods in having no
antennae, four pair of long six-segmented legs, and only two pair of mouth-
parts. Spiders are not truly aquatic although a few are consistently
associated with freswhaters. They are normally insectivores.
Mites, on the other hand, creep about on the bottom and on vegetation.
They are carnivorous or parasitic, feeding on insects and worms, piercing
their victim to draw out the juices. Insects and fish include mites in their
diet.
52
-------
Fish—
The most advanced phylum in the aquatic world is Ghardata, which
includes all those.animals with a backbone: fishes, various amphibians,
birds, and mammals.
(30):
Freshwater fish may be conveniently grouped as to their feeding habits
(1) Parasites. Feed on blood and bloody fuilds or other fish, e.g.,
Lampreys.
(2) Plankton feeders^ Having close-set gill-rakers. They often
show a migratory pattern, following the fluctuating plankton
population, e.g., Golden Shiner, Minnows, Paddlefish, Young Bass
and Trout. '•''.-'•..
(3) Bottom feeders. Obtain food by sucking. Many locate their food
by sensitive barbels hanging below the head and are often seen
swimming a short distance above the bottom with barbels dragging.
The stoneroller, however, picks up food with its protruding
lower jaw, e.g., Catfishes, Paddleifhs, Sculpins, Stoneroller,
Sturgeon, Suckers, Sunfishes, Whitefish.
(4) Vegetation feeders. Spend much of their time browsing or nibbling
on vegetation. They also feed on invertebrates, e.g., Cap, Creek
Chub, Eels, Golden Shiner, Goldfish, Killifishes, Minnows, Sculpins.
(5) Invertebrate feeders. Prey on insects and crustaceans mainly.
They also feed on Mulluscas. Includes the majority of freshwater
fish. ' :
(6) Invertebrate feeders. Also prey on fish and other vertebrates,
e.g., Bass, BoWfins, Turbot, Darters, Eels, Sculpins, Sunfishes,
Trout, Walleye, Yellow Perch.
(7) Vertebrate feeders. Use their long sharp teeth for grasping frags,
snakes,, and even turtles as well as many fish, e.g. , Gars,
Pickerel, Pike, Muskellunge.
(8) Notorious omnivores. e.g., Catfish.
It should also be noted that fish such as the trout are extemeley sensitive
to reduced dissolved oxygen, whereas buffalo fish, sticklebacks, the carp,
and the gar are quite tolerant of this condition.
Bacteria-^
Bacteria, as well as fungi, ase the recyclers, breaking down animal
and plant tissue into forms available for the lower trophic .levels. With-
out them, life would be impossible.
53
-------
r
In unpolluted waters, away from the more nutrient shores, the number of
bacteria is small on the order of a dozen per milliliter. The kinds of
bacteria present is greatly influenced by the amount of dissolved oxygen
present. Another important factor is the type of food present at the time.
For example, if a tree falls into the water, the whole flora is changed;
cellulose-digesters and fermentative types thrives as do saprohpytes. The
concentration may reach 100,000 per milliliter unitl an equilibrium is again
reached.
In streams polluted with organic material, EsoheTickia ooT/i and other
EnteTbaotevutceae3 as well as fecal streptococci and various species of
intestinal Clostridium are present in large numbers. Many soil saprophytes,
yeasts, and molds find organic wastes excellent food. The number of micro-
organisms in a heavily polluted stream may reach into the millions per milli-
liter. In the process of decomposing organic matter, aerobic bacteria con-
sume oxygen and may cause,a DO deficit if too much feed is available to them.
Summary of Impacts on Biologic Community—
The food web in the .aquatic environment is an extremely complex mecha-
nism. Its balance may be upset in a number of ways, many of which are
natural. Figure 11 shows some of the ways in which man made pollution may
affect life in rivers. Sediment transport may play a role in virtually all
the illustrated effects.
At this time it seems
sediments on aquatic life.
only partially understood,
ment approaches.
CAUSING EXTREME
pH CHANGE
impractical to define precise impacts of CSO
This is because the fate of the sediments is
However, the next section will consider assess-
LOWERING OF DO
INCREASING—
TURBIDITY
DEPOSITING
SETTLEABLE
SOLIDS
CAUSING
TOXICITY
NURTURING
UNDESIRABLE
GROWTH
INCREASING
TEMPERATURE
TAINTING FISH FLESH
Figure 11. Environmental Factors Which Produce Undesirable Effects in
Aquatic Life
54
-------
Approach for Assessment of Environmental Impacts
The U.S.. is entering an era when the costs of control of contaminants
from various sources will be weighed against potential environmental impacts
of failing to control these contaminants, or only partially controlling them.
Within this framework, it is important to ask: What are the real environ-
mental impacts of combined sewer overflow? How significant a problem is
this today? Can treatment, where needed, be directed toward those particular
contaminants of greatest concern? The concept of separating storm and
sanitary sewers in major municipalities has largely been abandoned by those
responsible for developing these programs because of the very high costs.
Therefore, the cost-effective ecologically sound approach which should be
used for control of the overflow is one of evaluating where there are real
environmental problems of major significance caused by combined sewer over-
flow from a particular municipality, and then taking steps to eliminate these
problems. Ultimately, if sufficient funds are made available, some of the
instances of combined sewer overflow of lesser importance in terms of environ-
mental degradation, can be corrected.
Basically, what is needed in the solution of combined sewer overflow
problems where they exist today is the development 'of an approach that will
allow a case-by-case evaluation of the environmental significance and cost
associated with controlling a particular combined sewer overflow that is
occurring at a specific site or within a limited region of a municipality.
This section addresses the information available today on the significance
of combined sewer overflow as a cause of water quality deterioration, with
particular emphasis given to the significance of suspended solids in the CSO
in affecting receiving water qualtiy in the region of the combined sewer
overflow point of entry to the environment.
The actual environmental impact of any particular combined sewer over-
flow is difficult to discern since usually these overflows occur in or to
waterways which would likely be heavily impacted by municipal and industrial
wastewater discharges to the waterway from sources upstream of the combined
sewer overflow point of entry. Therefore, even without the combined sewer
overflow, water quality in the waters receiving the overflow would usually
be degraded in many municipalities. There are few studies which specifically
single out combined sewer overflow as a cause of water qualtiy deterioration.
The lack of studies does not mean that there has not been significant water
quality deterioration because of CSO's. It is just that it has been difficult
to document this deterioration because of lack of funds for field studies
and the lack of attention being given to this area by those responsible for
control programs within urban centers. The situation with respect to the
environmental impact of combined sewer overflow solids is even worse. To
the knowledge of the authors there has not been a single study devoted
specifically to this topic area, where a quantifiable environmental degra-
dation has been traced back to a contaminant load from combined sewer over-
flow. While there have been no studies specifically directed to this topic
from the nature of combined sewer overflow contaminants, it is possible to
gain some understanding of potential environmental degradation that can
occur as the result of CSO's. Most importantly, guidance can be provided
to those responsible for assessing the significance of CSO's in a particular
55
-------
locality on how they should proceed to make this assessment and thereby judge
the potential benefits that will be derived from controlling the combined
sewer overflows to various degrees in a particular locality,
While there is little or no information on the direct environmental
impact of solids associated with combined sewer overflows, some inference
on the potential impact can be obtained through the studies that have been
conducted over the past half-a-dozen years as part of the Corps of Engineers'
Dredged Material Research Program. In the early 1970's, considerable concern
was voiced about the significance of chemical contaminants associated with
U.S. water way sediments during dredging and dredged materials disposal opera-
tions. There was a strong move made by pollution control regulatory agencies
to impose restrictions on the amount of open water disposal of dredged sedi-
ments because of these contaminants. The alternate methods of disposal were
often significantly more expensive than the then currently used open water
disposal techniques. This led to the federal Congress appropriating $30
million for a five-year study devoted to evaluating the environmental impact
of various methods of dredged material disposal. This study was completed
in 1978 and the results have been published in a series of reports by the
Corps of Engineers Dredged Material Research Program located at the Waterways
Experiment Station, Vicksburg, Mississippi. As part of this study, the
authors of this section of the report conducted a five-year, over $1 million
investigation in which a combination of laboratory and field studies were
conducted at a variety of sites across the U.S. in order to evaluate the
potential environmental impact of open water disposal of dredged sediments.
The findings of these studies were published in a two-volume report authored
by Jones and Lee (32) and Lee et al., (33).
The senior author of this section of the report was also involved as
an advisor to the Corps of Engineers in helping to establish the overall
Dredged Material Research Program and in reviewing the study results and,
therefore, is familiar not only with the studies being conducted under his
supervision but also with the studies conducted by other investigators as
part of the DMRP.
There is considerable analogy between the environmental impact of open
water disposal of dredged sediments and that associated with combined sewer
overflow. Both systems have considerable amounts of solid material con-
taining contaminants. Typically, however, the dredged sediment would tend
to be inorganic in nature while the combined sewer overflow solids should
tend to be more organic in nature. Further, both of these processes tend
to be intermittent in which there is a relatively large input of particulates
for a short period of time and then there is a period of time in which there
is little or no input.
For both open water disposal of dredged sediments and combined sewer
overflow, there are three basic environmental problems that must be con-
sidered. The first of these is the chemical contaminants associated with the
liquid phase (i.e., the solution). The environmental impact of these con-
taminants can be estimated to some extent based on the concentration in the
combined sewer overflow, the duration of overflow, the mixing-dispersion-
dilution that occurs within the waterbody, and the ambient concentrations of
56
-------
the contaminants of concern in the waterbody. It is important to note that the
US EPA July 1976 (34) water quality criteria (Red Book), or state standards
based on these criteria, were developed for worst case situations which involve
chronic exposure of aquatic organisms to available forms of the contaminants.
This fact should be kept in mind when using these criteria as a basis for
judging the significance of contaminants in the solution phase. The typically
intermittent occurrence of both dredged material disposal and combined
sewer overflow, coupled with the physical and chemical characteristics of
the materials, create situations where large portions of the contaminants
are in forms that are unavailable to affect aquatic life. Except in
rare instances where the receiving waters have limited ability to dilute
the contaminants, there will not be a chronic exposure of organisms.
The second area of concern with respect to the environmental impact of
combined sewer overflow is the physical effects of the solids on benthic
and epibenthic organisms in the region of the discharge, as well as any
watercolumn organisms present in the discharge area. Considerable work
has been done over the past several years as associated with the discharge
of mining waste to the environment, which has shown that the impacts of the
suspended solids associated with dredged material or tailings disposal are
rarely of significance to watercolumn organisms. The American Fisheries
Society and others have taken a less conservative attitude on the criteria
for suspended solids than that contained in the US EPA Red Book which limits
the suspended solids increase in water to a five percent change in compen-
sation depth, i.e., the depth of the watercolumn at which photosynthesis equals
respiration. Plumb and Lee (35) have discussed this point in detail in their
review of the potential effects of suspended solids on phytoplankton.
Considerable work has also been done [see Pedicord et al,, (36)] on the
significance of supended solids to both freshwater and marine animals. It
was found that many of these organisms can tolerate very high, i.e.,_grams
per liter concentrations of suspended solids for extended periods of time
without any adverse effects. Few, if any, problems would be expected to
result from the presence of suspended solids in the watercolumn associated
with open water dredged material disposal other than an aesthetic problem
of turbid-cloudy water for a period of time, usually a few hours to a day
or so, which might be readily noticeable in an aquatic environment which has
a low ambient suspended solids concentration. The likelihood of problems
due to suspended solids in combined sewer overflow would be even less than
that for dredged sediment because of the fact that the solids content of
dredged sediments is typically on the order of 25 percent sediment at the
point of release. Typical combined sewer overflow would be expected to have
considerably lower concentrations of solids and, therefore, except for the
potential for short term aesthetic problems, would not likely be a signifi-
cant factor in impairing the beneficial uses of the waterbody, as is the
case with dredged sediments, does not represent a long term, continuous
increase in the cloudiness of the water. Therefore, even if potentially
adverse suspended and solids levels were found at the time of the discharge,
because of the intermittent nature of these discharges, organisms would not
57
-------
likely be exposed to them for sufficient periods of time to be adversely
affected. Some organisms also exhibit an avoidance behavior; they would
leave an undesirable area or not enter it, thereby not being exposed to
adverse conditions.
There is one important physical effect that the solids in both dredged
sediments and combined sewer overflow have that can cause a significant
detrimental environmental impact. This is the burying or smothering benthic
and epibenthic organisms. There is no question about the fact that disposal
of dredged sediment in a region where there are coral reefs, oyster beds,
fisher production areas where the eggs are laid on the bottom, or in some
aquatic organism nursery areas, can have a very significant deleterious
effect on aquatic organisms in those regions. It is important that open
water dredged sediment disposal take place in such a way so as to avoid these
kinds of problems. This can usually be done through a site-specific investi-
gation designed to detect the presence of areas ecologically sensitive to
suspended solids deposition.
Since combined sewer overflow systems are no longer being constructed
and ecologically sensitive discharge areas have long ago been wiped out,
about all that a site-specific investigation of an area might conclude is
that if it were not for the suspended solids deposition in a region, the
region might have certain types of benthic and epibenthic habitat. At this
time, however, aquatic biologists'-ecologists' ability to predict a type of
habitat that would exist given different environmental conditions, such
as with or without a combined sewer overflow, is quite limited. Normally,
even without the combined sewer overflow, the habitat has been so drastically
altered by other physical structures and altered flow regimes, etc., that
there is little likelihood that one could predict with a high degree of
reliability the potential for reestablishing certain types of aquatic habi-
tats by elimination of a combined sewer overflow. This is an area, however,
that does need research and which could be cost-effective in helping water
pollution control officials to decide whether or not the control of the
solids associated with combined sewer overflow situation could represent a
cost-effective, ecologially sound approach toward improving aquatic habitat
of a region.
The third area, which is probably of greatest concern associated with
combined sewer overflow, is that of the significance of chemical contaminants
associated with the solids that are discharged to the environment, Combined
sewer overflow solids represent a mixture of domestic wastewaters, industrial
x*aste, and urban stormwater drainage, All three of these sources likely have
very high concentrations of contaminants associated with their solid phase,
Even if the contaminants are discharged to the sewerage system in. a soluble
form, the relatively high concentrations of the solids present in the waste-
waters and urban drainage would readily sorb many of these rendering them
less available to affect aquatic life.
The Dredged Material Research Program provides considerable information
that is'directly pertinent to this topic area. The studies by Lee et al.,
(33) and Jones and Lee (32) as well as those of others, have shown that a
58
-------
relatively simple leaching test - the elutriate test - provides a reliable
indication of' the potential for contaminants associated with waterway sedi-
ments to be released to the watercolumn upon open water disposal of the sedi-
ments. A test like the elutriate test needs to be developed for combined
sewer overflow in which well-defined leaching tests are run on samples of
the overflow material; the amount of leached contaminants is then interpreted
in terms of the concentraion of contaminants in the ambient water and the
dilution-dispersion of these contaminants that takes place in the receiving
waters. Since it is conceivable that contaminants associated with solids
in the combined sewer overflow could be and would be released to the water-
column under the more dilute situations that occur in the aquatic environ-
ment, a leaching test should be used. However, it is important that an
approach similar to that advocated by Jones and Lee (31) be used in interpre-
tation of the leaching test results. As noted, by Jones and Lee, the
significance of contaminants released from dredged sediments, and this would,
certainly be true for combined sewer overflow as well, must be judged in
terms of a critical concentration-duration of exposure relationship for
available forms of the contaminant. Available forms are those forms that
can have an adverse effect on aquatic organisms. Since some so-called
soluble forms are unavailable to aquatic organisms, it is important that the
leaching test properly consider available forms of the contaminants of
interest. As advocated by Lee and Jones (37) in their November 1978 state-
ment to the Colorado Commission on Water Quality, for interpretation of data
such as may be developed from combined sewer overflow leaching tests, the
US EPA water quality criteria should be used to flag potential problems.
Upon detection of an apparently excessive concentration of contaminants
based on US EPA criteria (which were typically developed by chronic
expsoure of organisms to 100 percent available forms of contaminants) the
source of the contaminant, or in this case the municipalities, would under-
take a field study to determine if the contaminants released from the solids
as well as those that are present in the aqueous phase of the combined sewer
overflow do, in fact, have an adverse effect on water quality, i.e., bene-
ficial uses in the waters receiving combined sewer overflow. Specific
guidance can be provided on how municipalities can conduct studies of this
type at a minimal cost, which would still provide the necessary information.
Another aspect of the contaminants associated with combined sewer over-
flow solids that needs to be addressed is that of the direct uptake of these
contaminants by benthic and epibenthic organisms which could result in there
being a detrimental effect on the organisms or on higher food web organisms,
including man who might consume the organisms. The latter area is of partic-
ular concern because of the fact that there are a number of chemical con-
taminants present in combined sewer overflow which would tend to be associ-
ated with solids which would cause edible fish and shellfish in the region to
become contaminated to the point where they would not be suitable as a
source of food for man, based on the Food and Drug Administration limits.
These chemicals include mercury, PCB's, and others.
The results of the Dredged Material Research Program and related
studies show that one cannot predict, based on the total contaminant content
or leachable fraction of contaminants from a particular sediment, the amount
59
-------
of a contaminant that will be taken up directly by aquatic organisms living in
contact with these sediments. While there are situations where there is- direct
transfer of contaminants from these solid phases to the organisms, situations
occur [see Neff et al., (38)] where large concentrations of contaminants
present in sediments, are not transferred to organisms living in contact with
them. In fact, organisms may actually be "cleaned up" by the sorption capacity
of the sediments, in which the net gradient of the transfer of contaminants
is from the organisms to the sediment rather than the reverse.
Jones and Lee (32) have suggested that the appropriate approach to take
with respect to dredged material disposal operations where there is concern
about the transfer of contaminants from the solids to benthic and epibenthic
organisms is to not try to predict such transfer, but to measure whether such
transfer is occurring at this time. This can be done by analyzing the tissu'e
of organisms in the vicinity of a disposal site being used to dispose of simi-
lar sediments. Therefore, in a similar way studies of organisms from the
areas where combined sewer overflow solids are accumulating, would likely show
whether or not there are any potential problems from future accumulations of
CSO solids in that region. In general, it is important to note that the ener-
getics in terms of hydrodynamic regimes of combined sewer overflow areas are
such that periodically the solids of the region would be flushed out or dis-
persed to deeper or other waters. This is exactly the same situation that
occurs in dredged sediment disposal and, therefore, like with the dredged sedi-
ment "liquid phase" there is a dilution of the contaminated sediments with
lesser contaminated sediments of the region. While dilution does not always
result in less contamination of organisms, ultimately since the sediments of
many waterways have appreciable sorption capacity for contaminants, it is
likely that upon dispersal of the sediments from either dredged sediment or
combined sewer overflow, few problems related to the chemical contaminats
associated with the CSO solids would likely be encountered.
The Dredged Material Research Program has shown that very little is
known today about the effects of contaminants associated with sediments on
benthic and epibenthic organisms. It has been well established that many of
the contaminants associated with sediments are not available to organisms.
However, under certain conditions, there is transfer of contaminants from the
sediment. At this time no one is in a position to interpret the significance
of the uptake of contaminants to benthic and epibenthic organisms. We do not
know the critical body tissue concentrations and therefore we cannot, by
analyzing organisms, discern with any degree of reliability the potential impact
of a CSO sediment-derived contaminant. As noted above, the only guidelines
that are available today in this area are the FDA limits for those organisms
which serve as human food. It is unlikely that this situation will change in
the foreseeable future and,' therefore, the focal point of any contaminant
control program for the CSO sediment contaminants must be directed toward
edible organisms and those contaminants in which there is a definable limit
on critical concentrations, i.e., the FDA limits. It is important that the
bulk sediment criteria advocated by some researchers be evaluated before
used since these have been clearly shown to be invalid as a basis for pre-
dicting adverse effects to benthic organisms. The approach that was developed
as part of the Dredged Material Research Program of using bioassays (see
60
-------
Jones and Lee (32) and Lee et al., (39) for discussion of this area) to detect
the problem areas, is a technically valid approach that should be considered.
If interpreted properly, bioassays of the type developed by Lee et al., (33),
Jones and Lee (32), and Lee et al., (39) can be adapted to the CSO situation
to allow municipalities to detect potentially significant problem areas
associated with CSO contaminants. Chemical tests will not properly predict
organism uptake.
An important aspect of the bioassay program is the proper interpretation
of the data. As discussed by Lee et al., (39) the approach used by the US
EPA and the Corps of Engineers in their July 1977 bioassay manual may not
•be technically valid. However, it is being used by a number of environmental
quality regulatory agencies. Guidance on the approach that should be used in
the interpretation of benthic organism bioassays is provided by Mariani, et
al., (40),
It is important to note that the current pollution control efforts of
the federal, state, and local agencies will tend to reduce the amounts of
highly toxic contaminants entering the combined sewers. The pretreatment
requirements for industrial waste will go a long way toward reducing the
concentrations of various types of contaminants in combined sewer overflow
and, therefore, in time, the environmental impact of combined sewer overflow
will become more and more restricted to the physical effects of these solids
and the classical pollutants such as oxygen demand and turbidity, associated
with domestic wastewaters. Since in general, these effects are of much lesser
significance than those of the hazardous chemical contaminants, it is, there-
fore, important for those who are encountering combined sewer overflow
problems to eliminate as soon as possible the "exotic" chemicals entering the
combined sewers. Many municipalities have found that by investigating the
heavy metal, chlorinated hydrocarbon, and other contaminant content of their
domestic wastewaters, they can detect abnormally high concentrations of
contaminants which can then be traced back to particular sources and control
the source of input. Such a program could readily greatly lessen the environ-
mental impact of combined sewer overflow in many municipal and urban areas.
From an overall point of view, it may be concluded that while no work
has been done on the significance of CSO solids as they may cause environ-
mental quality degradation, based on the nature of the materials and the
information available from other areas, principally the Dredged Material
Research Program, it is apparent that both adverse physical and chemical
effects can occur, and that these must be evaluated on a site-by-site basis.
Further, the results of the Dredged Material Research Program provide
valuable guidance to the approach that should be used in developing a CSO
contaminant control program. It is evident that because of the high cost
associated with the complete control of combined sewer overflow, careful
site-by-site or at least limited-region evaluations should be made by
municipalities to be certain that the initiation of such control programs
is cost-effective in terms of improving the environmental quality of a
particular region that is receiving CSO at this time. There is little
point in conducting very expensive programs to control CSO-associated
61
-------
contaminants which are unavailable to affect aquatic life and/or water
quality or that have such a limited extent of effect that their elimination
would be judged to be essentially inconsequential compared to the cost of
their removal. An example of this type of problem occurs with the phosphorus
control in the Great Lakes. Sufficient information has been accumulated
over the years on the sources and significance of phosphorus in the Great
Lakes to enable critical evaluations to be made'of the benefits that may
be derived by controlling phosphorus from any particular source. One of the
potential sources that is of concern is combined sewer overflow. It has
been estimated that about five to ten percent of the phosphorus entering
Lake Michigan is from combined sewer overflow, Lee (41). While this might
be judged to- be a significant aspect of the Great Lakes phosphorus control
program based on the total phosphorus input, it is apparent that only a
small part of the phosphorus associated with combined sewer overflow would
be available to adversely affect Great Lakes water quality. Lee et al.,
(42) have recently completed a comprehensive review of the information
available today on the available phosphorus in domestic wastewaters as well
as other sources such as urban stormwater drainage. Coincident with
developing an understanding of how much of the phosphorus that is present
in various sources entering the Great Lakes is likely to become available
to stimulate algal growth within these waterbodies, information has been
developed on load-response relationships in which it is now possible to
predict the impact of altering phosphorus load to one of the Great Lakes,
or for that matter, to many other waterbodies based on the results of the
OECD eutrophication study that has been conducted over the past several
years. Rast and Lee (43), Lee et al,, (44), Lee and Jones (45), Jones
et al., (46), and Lee et al., (47), have presented the relationships
between the reduction in phosphorus loads to each of the Great Lakes and their
resulting eutrophication-related water quality. The Lee et al., (44) review
has shown that significant reductions in the available phosphorus loads must
occur for each of the Great Lakes, especially the lower lakes, if there is ^
to be any improvement in water quality that would be discerned by the public.
It is quite apparent that elimination of the combined sewer overflows to
Lake Michigan would, even if all of the phosphorus present in the CSO's
were avalaible, have little or no impact on Lake Michigan's overall eutro-
phication-related water quality. When this is considered in light of the
fact that upwards of possibly 50 or so percent of the combined sewer over-
flow would be in a form that would not be available to support algal
growth, it is highly questionable whether the control of combined sewer
overflow as a means of phosphorus control is cost-effective and technologically
valid. As discussed by Lee and Jones (45), there may be situations where
control of phosphorus in combined sewer overflow in a particular municipality
might, on a localized basis, have a significant beneficial effect on water
quality. These situations require a case-by-case evaluation.
It is impossible to generalize on the benefits that would be derived
by large-scale combined sewer overflow control. These systems must be
examined individually using the techniques available today to assess load-
response relationships and availability to contaminants to affect water
quality in terms of the typical or normal situations that occur in a
particular combined sewer overflow discharge area. It is conceivable that
62
-------
some of the techniques that have been utilized for lake restoration, such as
direct iron or alum addition to the lake to control phosphorous might be highly
beneficial in eliminating adverse effects of contaminants associated with
combined sewer overflow. Lee et al,, (42) suggested that the use of water
treatment plant alum sludge might be a highly cost-effective means of control-
ling the impact of urban stormwater drainage on urban waterbodies. It is
conceivable that in those areas where it is judged to be too expensive to
try to eliminate the combined sewer overflow or to install treatment works,
the simple practice of feeding iron or alum to the combined sewer overflow
at the time that it is discharged to the environment, including the potential
for using wastewater treatment plant sludge, should be investigated. There
are few situations in the environment where the addition of a small amount
of solids which would occur as the result of allowing the sludge to settle
in the receiving waters and thereby carry with it many of the contaminants,
would result in a significant environmental degradation. In fact, the
high sorption capacity of hydrous metal oxides, such as iron and aluminum
(Lee, 48) is such that in many cases the addition of these materials directly
to the environment may prove to be highly beneficial in eliminating the
problems not only of contaminants present in combined sewer overflow, but
also of contaminants derived from other sources as well. It is evident
that there is need to investigate the actual environmental impact of combined
sewer overflows and to develop guidance that can be used by municipalities
to determine for a particular situation, the benefits that can be derived
from spending the necessary funds to completely eliminate CSO through combined
sewer separation, to treat combined sewer overflows through conventional
treatment works, or to partially treat utilizing the environment as part of
the treatment system, or from completely failing to do anything about the com-
bined sewer overflow, The ability to provide guidance of this type and to
develop systems which are cost-effective, technically valid, and ecologically
sound,'have improved significantly over the past several years. This improved
technology should be brought to bear on the combined sewer overflow problem.
63
-------
SECTION 4
MODELING THE MOVEMENT OF CSO SEDIMENTS
INTRODUCTION
As discussed in the previous section, the key to determining the impact
of CSO sediment material is to determine its fate. The current state-of-
the-art of sediment tracing is such that direct measurement is not practical
(discussed in Section 5). The most sensible means of determining where CSO
sediments come to rest and how they travel is by means of mathematical
models.
The purpose of this section of the report is to present an evaluation
of a current sediment model for studying the fate of CSO sediments. The
model is developed from transport routines in Colorado State University's
watershed sediment model. These routines are combined with a flow model
developed by the U.S. Geological Survey. The evaluation is designed to
determine the facility with which such models might be used to analyze
the fate and effects of CSO sediments. A reach of the Cuyahoga River
between Akron and Cleveland, Ohio is used in the evaluation. The reach
includes the bypass and outfall of the Akron municipal treatment plant.
This reach was identified as one in which dissolved oxygen deficits
frequently accompany storm events (4).
A brief discussion of the Colorado State University model and the
modifications made by Sutron is presented first. This is followed by a
complete description of the study reach and the available data. Model
calibration and experiments are presented last.
COLORADO STATE UNIVERSITY MODEL AND MODIFICATIONS
Original Watershed Model
As originally configured, the Colorado State University sediment model
was designed to predict the sediment yield as a function of time for small
forested watersheds. In fact, the original development was funded through
EPA with funds from the U.S. Forest Service, Department of Agriculture,
Flagstaff, Arizona. The original model is described in reference (49).
The original model simulated the land surface hydrologic cycle, sedi-
ment production and water sediment movement. Conceptually, the watershed
is divided into an overland flow part and a channel system part. Different
physical processes are important for the two different environments. In the
64
-------
overland flow part, processes of interception, evaporation, .infiltration,
raindrop impact detachment of soil, erosion by overland flow, and overland
flow water and sediment routing to the nearest channel are simulated. The
overland flow portion was discarded for this study. In the channel system
part, water and sediment contributed by overland flow are routed and the
amount of^channel erosion or sediment deposition through the channel system
is determined. The main functions in the original model are shown in Figure
First, the excess rainfall rate is determined by subtracting the
interception storage of water by vegetation, the evaporation rate and the
infiltration rate from the total rainfall rate. Excess rainfall or the
amount of rainfall that runs off the land surface is then routed through
the overland flow units downhill to the nearest channel unit. Again, this
portion of the model was discarded, Once the water reaches a channel unit,
it is routed on to the outlet of the watershed through additional channels
based on the principles of continuity and momentum. The amount of sediment
transported is governed by the available supply of sediment, transport
capacity of the flow of water and the principle of continuity. The continuity
(conservation of mass) principle for water and sediment simply states that
whatever water or sediment that comes into a stream segment must be either
stored there or conveyed out the downstream end in such a way that mass is
neither created nor destroyed, The subroutines comprising the channel
routing portion of the model were retained and modified for this study,
Model Modifications
Water Routing—
Water routing in the original model predicted the response of the water-
shed to an input of water which was in the form of excess rainfall. The
watershed model printed out the discharge hydrograph at the outlet of the
watershed. The water routing formulation was a second-order nonlinear-
solution of the continuity equation under a.kinematic wave assumption. The
water routing procedure was also discarded. This was done because the
kinematic wave formulation would not work under adverse slope (channel bed
elevation increasing downstream) conditions. Such conditions are occasionally
formed behird small dams or natural rock outcrops. A linear implicit finite
difference flow model developed by Reefer and Jobson (50) was used instead.
The flow model uses a finite-difference solution of the one-dimensional
continuity and momentum equations for gradually varied flow. The forms used
here are identical to those presented by Amein and Fang (34). The tech-
nique used to solve the equations is referred to as fully-forward, linear,
implicit. In this scheme, the finite-difference approximations of spatial
derivatives are written at the forward time level so that the solution for
the unknowns, at the end of the time step, can be obtained without iteration.
The linear implicit technique is one of the most stable of the finite differ-
ence flow models.
The channel reach to be modeled is represented by a fixed number of
grid points up to 100. The channel cross-section at each grid point is
65
-------
r
1 GEOMETRY
DATA
1
[ SOIL
DATA
1
fc.RO
CANC
DATA
'BASIN
\RACTER)
ISTICS
DATA
[ FLOW
RESISTANCE
I DATA
f SEDIMENT
ROUTING
I DATA
1
r INITIAL
INTERCEPTION
STORAGECONTENT
Figure 12. Flow Chart for'the Watershed Sediment and Routing Model
66
-------
represented by up to 20 pairs.of elevations and distances. Resistance to flow
is set by Manning's
"n" value at each cross-section.
T.he Keefer-Jobson model was used without modification. The program
was designed to accept a variety of upstream and downstream boundary condi-
tions as well as tributary flow. Velocities and depths at all model cross-
sections can be written on a direct access disc file. The disc file acts
as the coupling mechanism between the flow and sediment models. The sedi-
ment model reads the velocities and depths rather than computing them as in
the original watershed model. An initial run is made with the flow model to
predict velocities and depths. The sediment model can then be run with a
wide variety of boundary and initial conditions without rerunning the flow
model. It should be noted that if substantial changes in cross-section
shapes occurred due to aggradation or degradation, direct coupling of the
two models would be necessary. Large changes in the bed slope would even-
tually cause substantial changes in the flow field. For this preliminary
evaluation, indirect coupling was considered adequate.
Sediment Transport Routine-—
The sediment model will be discussed in greater detail, since it is
of primary concern to this study. The sediment routing portion of the model
is highly process-oriented. Such fundamental mechanisms of transport such
as settling or detachment are handled by a separate equation. The model is
believed to be one of the best currently available.
The movement of sediment in a channel is governed by the equation of
continuity for sediment and sediment transport equations (such as fall
velocity and critical shear stress). The amount of sediment that could be
transported is described by equations of sediment detachment by the flow.
The equations used in the model are described below.
The equation of continuity for sediment can be expressed as (see
reference 51):
3G
s
9x
3CA + 3Pz
3t
3t
= g (volume/unit length/time) (1)
Where:
C =
(volume/volume)
(2)
and G is the total sediment transport rate by volume, C is the sediment
concentration by .volume, z is the net depth of loose soil, P is the wetted
perimeter, and g' is the lateral sediment inflow.
S
The sediment load can be broken into two main categories; bed material
load and suspended load. Bed load consists of sediment particles that move
by saltation (jumping) or rolling along the stream bed. Suspended load
consists of particles that are transported above the bed by the turbulent
nature of the flow.
67
-------
In order to simulate the actual grain size distribution found in soil
samples, the sediment load may be broken into any specified numbers of size
fractions. The sediment continuity equation is then written using arrayed
variables according to sediment size. The percentage of sediment in each
size fraction is accounted for in the transport equations.
36s(I) 3C(I)A , 3Pz(D _ m
ax at at ~ 8U;
(volume/unit length/time) (3)
in which I indicated which size fraction is being calculated (I
of size fractions, currently limited to 10 in the model).
number
The sediment transport equations are used to determine the sediment
transporting capacity of a specific flow condition. Different transporting
capacities are expected for different sediment sizes. The transporting rate
of each sediment size can be divided into the bedload transport rate and the
suspended-load transport rate. Before a particle can be transported, how-
ever, it must be detached from the channel bed. (Note that in all cases,
"particle" will refer to spheres with specific gravities of 2.65. The
model was subsequently modified to accept other specific gravities, but
this will be discussed later.)
When a river flows over its bed, it exerts a tractive force on the bed
in the general direction of the flow. This force is called the boundary
shear stress and may or may not be large enough to cause sediment particles
of various sizes to move. The shear stress at which a given particle
begins to move is the critical shear stress. Critical shear stress depends
mainly on the specific gravity and diameter of the particle and is given by
the following equation:
(force/area)
(4)
Where:
Y = the specific weight of sediment
s
Y = specific weight of water
d = particle diameter
a constant
6 s =
The general form of this equation is attributed to Shields, who com-
pared the ratio of gravitational forces holding a particle down to the
inertial forces of the flow wanting to carry it away. Analyses comparing
the ratio of the energy to cause particle motion and to resist motion given
similar results. Laboratory experiments have shown that this beginning of
motion criteria is valid for particles with specific gravities from 0.25
up to 8. There is little reason to suspect heavier particles will not also
follow this relationship. The constant 8 has been reported to be 0.06 by
Shields (52) and 0.047 by Meyer-Peter andSMuller (53).
68
-------
Shields' critical shear criterion is generally accepted for cohesion-
less particles from 0.0675 mm sand sizes on up. Sediment which consists
of silt and clay particles show greater resistance to erosion than coarser
sediment and are not considered in the modes used here.
The critical shear criteria for cohesive sediments are a more difficult
problem. A task committee for the American Society of Civil Engineers in
the text of their sedimentation manual (54) state the following:
"The behavior of fine sediments under the attack of flows is
complex and depends on many factors including the electro-
chemical environment of the sediment. Few studies have been
made of this problem and knowledge of this phase of sedimenta-
tion is therefore in a primitive state."
Despite this, the manual contains some basic information on critical
shear criteria of cohesive praticles. Relationships are given based on
experimental data that demonstrate that the critical shear stress increases
dramatically as the percentage of clay or the plasticity index of the soil
in question increase.
Unfortunately, the critical shear stress of cohesive materials depends
on such things as the amount of time that the material has been sitting in
its present position. The amount of salt content in a clay specimen can
change the strength by a factor of 1,000. Such difficulties make the pre-
diction of the critical shear of cohesive materials.highly speculative. Use
of a given equation that was valid for certain specimens may give tremendously
inaccurate results when used in general. Additional research is needed to
understand the critical shear and transport of cohesive sediments.
Equations describing the bed load transport generally follows the form
given by DuBoys (55) and is closely related to the critical shear stress
criteria. These equations often are written as:
Where:
lv = a(f. - T ) (volume/unit width/time) (5)
T = the boundary shear stress acting on a sediment particle
a and b = constants
The boundary shear stress can be expressed by:
-» 1*1
(force/area)
8
(6)
Where:
f '•= a Darcy-Weisbach friction factor due to grain resistance
p = the density of water
V = the average flow velocity
69
-------
Again, numerous engineers have conducted laboratory experiments to
determine a and b. A simple and widely used bed load transport equation is
the Meyer-Peter, Muller equation (56):
.1.5
- Y)
(volume/unit width)
(7)
where q is the bed load transport rate in volume per unit width. A dis-
cussion of various bed load equations is found in (55). The Meyer-Peter,
Muller bed load equation is incorporated in the model at present but any
other formulation could be used if proven more acceptable for the particular
type of modeling to be done. Reference (57) gives a complete description
of numerous other forumalations and their limitations.
The suspended load plus the bed load gives the total sediment load
carried by the stream sediment that is carried in suspension consists
usually of smaller sized particles continuously supported by turbulence.
Settling velocities for suspended loads are usually quite small.
One of the most widely recognized methods of estimating suspended load
was developed by Einstein (58) and was modified by Colby and Hembree (59).
The modified Einstein procedure is incorporated in the model and is described
below.
The sediment concentration profile which relates the sediment concen-
tration with distance above the bed (49) can be written as:
R - a*
w
(dimensionless)
(8)
Where:
C = the sediment concentration at a distance £ from the
C bed
3 A = the known concentration at a distance "a*" above the
a bed
R = the hydraulic radius
w ~ paramater defined as:
V
w = s
(dimensionless)
0.4IL
(9)
Here V is the settling velocity of the sediment particles, and U^ is the
shear velocity defined as:
11/2
(length/time)
(10)
70
-------
in which specific shearing stress, TA, is defined as:
,2
_ Tr
- fpV
(force/area)
(11)
A logarithmic velocity profile is commonly adopted to describe the
velocity distribution of turbulent flow and can be written:
r^2- = Bn + 2.5 Hn
U^. 1 n
(dimensionless)
(12)
Where:
U_ = point mean velocity at a distance £ above the bed
B.. = a constant dependent on roughness and n is the rough
ness height s
The integral of suspended load above "a*" level in the flow is obtained
by combining Equations (8) and (12):
(volume/unit length/time)
Ca*U*
a*
R - a*'
(13)
Let
(dimensionless)
and
R
(dimensionless)
Substitution into Equation (13) yields:
qs =
(1 - G)
w
[-1 + 2'5 "(tj/
2'5
Ana
1 --a
w
da
(volume/unit length/time) (14)
According to Einstein (58) , the sediment concentration near the bed layer
C £ is related to the bed load transport rate q, as:
a ^b
71
-------
q, = 11.6 C *U,.a (volume/unit width)
(15) .
in which a is redefined as the thickness of the bed layer which is twice the
size of the sediment.
The average flow velocity V is defined by the equation:
(length/time)
'
V
R
(16)
J
Using Equation (12):
•2- = B, + 2.5
- 2.5 (dimensionless) (17)
Einstein (41) defined the two integrals in Equation (14) as :
1 / Y7
-j
J = /
da
(dimensionless) (18)
and
'2 ' G
i / \w
'l (L^JL) .
\ a /
Unada
(dimensionless)
(19)
The intervals J- and J« cannot be integrated in closed form for most values
of w so a numerical method of determining J, and J~ developed by Li (49) is
adopted in this study.
Substitution of Equations (15), (17), <18), and (19) into Equation (14)
s the exression iven b Simons et al., (37):
ouDBUitutJuoii or nquatJ-uiia \J-JJ , \J-ij, \J-OJ , a-ii
yields the expression given by Simons et al., (37):
q w-1 r, , "1
n - ^ G f V . o A T -LOST
Q — ^rr; ^ 11 7^ ~"~ ^-.3? J^ ^ ^.J «Jol
Hs 11.6 . .w \U. /I 21
V.J. ~ \j) L, " J
(volume/unit width)
The total sediment load per unit width is:
qt = qb + qs (volume/unit width)
and the sediment transporting capacity of the section G is:
Q = Pg (volume/time)
(20)
(21)
(22)
72
-------
When considering transport by different sizes, Equation (22) should be modi-
fied as follows:
G (I) = PF (I)
c a
q. (I)
t
(volume/time)
(23)
in which F (I) is the adjusted fraction of sediment in the ith size.
3.
The percentage in each size fraction on the surface changes over time
due to armoring. The water transports the smaller sizes more easily and
leaves the larger size fractions behind. Thus, the percentages of sur'face
material need adjustment each time step. If the total loose soil depth is
greater than D
g,
the adjusted percentages, F (I)
3.
(the size of sediment for which 84% of the sample is finer) ,
can be written as:
Fa(I)
M
I Z(I)
1=1
(dimensionless)
(24)
M
If the total loose soil depth J Z(I) is less than DR,, then the adjusted
1=1 bif
percentages must account for the layer of undisturbed soil which is distributed
according to the original percentages plus the loose soil which covers it.
'84
84
M
- I
1=1
(25)
(dimensionless)
Often a size class or type of sediment particle is not found initially
in the bed material but is tranported into the reach of the water flowing in
the channel. For example, the transport of heavy metals in a CSO may affect
size distribution in the bed. In this situation, the percentages of incoming
material into a channel reach are used to further modify the adjusted per-
centages of size classes found in the bed. This modification was added by
Sutron as part of this study.
The amount of sediment detachment from surface bed runoff is determined
by comparing the sediment transporting capacity of the total available amount of
loose soil. By substituting the sum of the transporting eapacities,
M
I
1=1
.CD.
(given by summing the transport rates for M size fractions) into the trans-
porting rate given by Equation (1), the total potential changes in loose soil
storage is determined:
P _ 32
AZ =
9t
(At)
(length)
(26)
73
-------
If AZ >. -Z, the loose soil storage is enough for transport and no detach-
ment of soil by surface runoff is expected. Soil is detached if AZ™ < -Z and
the amount of detachment is:
D = -D [AZ
+ Z]
(length)
(27)
in which D is the total amount of detached soil and D. is defined as a detach-
ment coefficient with values ranging from 0.0 to 1.0 depending on soil credi-
bility. As an example, if the flow were over a nonerodible surface, the value
for D,. would be zero. If one were considering flow in a river where the
riverbed is always loose, the value for
D would be unity.
The new amount of loose soil should be further modified as follows:
Z(I) = Z(I) + D[F(I)] (length)
(28)
in which Z(I) is calculated for each size fraction of sediment. The numerical
procedure used for sediment routing is now presented. The transporting
capacity is determined by using Equation (23) and the computed flow conditions
from the water routing model. The potential sediment load concentration for
a given size fraction is then:
(D
n+1
0
(volume/volume)
(29)
These qualities are at time n + 1 and space j 4- 1 in the space-time plan.
When computing the potential sediment transport, the excess shear may be
less than or equal to zero indicating that at that section of channel that
particular sediment particle will settle out. Even though the excess shear
is negative, some particles may be transported downstream because their
settling time may be too slow as compared with the time it takes the par-
ticle to move downstream at the average stream velocity. Thus, a certain
minimum transport rate is maintained for that particular class of particles.
This minimum rate may be near zero when settling velocities are large enough.
This capability was also added to the model by Sutron as part of this study.
The potential change in loose soil storage for sediment in a given
size fraction is:
AZ(I) =
Gs(I)At -
- a)
- b)
- C(I)n An)
3 J'
(30)
(length)
74
-------
If AZ (I) is positive, that size of sediment is aggrading on the bed; and if
negative, that size of sediment is being transported off the bed.
The actual transport rate depends on both the availability of material
and the transporting capacity of the flow. If AZP(I) >_ -Z(I), the availability
is greater than the transporting capacity. Thus, the transport rate for
material in size fraction I is equal to its transporting capacity or:
and the actual change in Z(I) is:
AZ(I) = Zr(I)
(volume/volume)
(length)
(31)
(32)
If AZ (I) < -Z(I), the availability of material is less than the
transporting capacity. The transport rate is limited by the availability
of loose soil and the bed material concentration is therefore:
n+1
j+l
- b) -
(b)] + g(i)At - e[-Gs(i)n+1 (i - a) + (GS(I)J+I
j) (a)]
PZ(I)
,
- a)
(1 - a)
(length)
(33)
and
AZ(I) = -
(length)
(34)
The sediment transport rate
, , _ , x
^ j+l ~ ^ j+l
1^
is determined by Equation (2)
(volume/time)
(35)
75
-------
and the amount of loose soil available at the next time increment is:
Z(I) = Z(I) + AZ(I) (length) (36)
A four point noniterative finite difference scheme is used to solve the
above equations at each cross-section at each time step. This concludes the
presentation of the model theory. The following section describes the Cuyahoga
River study reach where the model was tested.
CUYAHOGA RIVER STUDY REACH
General Considerations
The river reach selected for this study is illustrated in Figure 13. The
figure is taken from reference (60). The reach extends from the U.S. Geological
Survey guying station, "Cuyahoga River at Old Portage" downriver to the Cuyahoga
Estuary at Lake Erie. A straight line distance of slightly over 40.23 km
(25 miles) is involved. The distance along the river is 64.37 km (40 miles)
counting meanders.
Reasons for Selecting Reach
Two basic reasons for selecting this study reach were data availability
and a history of water pollution problems. The USGS stream gauges at Old
Portage and Independence have been in existence for many years, Four para-
meter water quality monitors (HO, pH, temperature, and conductivity) are
also maintained at both gauges. Most important for this study, sediment
samples had been collected at both gauges by continuous pumping samplers.
The Akron sewage treatment plant is located 4.83 km (3 river miles) below
the Old Portage gauge. An additional stream gauge, "Cuyahoga River at Ira,"
is located close below the treatment plant outfall. Records from the USGS
gauges provided most of the data necessary to develop the sediment model.
The Cuyahoga estuary is widely known for having caught fire in the
1960's. In a previous study [EPA Contract No. 68-03-1630 (4)], Sutron identi-
fied the study reach as one in which a high probability existed for DO
deficits to occur at times of urban storm runoff.
General Characteristics of Basin and River
A good general description of the runoff and sedimentation characteris-
tics of the Cuyahoga basin is contained in (60). The, Cuyahoga River Basin
lies in northeastern Ohio and drains an area of about 810 square miles. There
is a breakwater protected outer harbor at Cleveland of about 526 hectares
(1,300 acres) and the 9.33 km (5.8 miles) of channel near the mouth have been
improved for navigation. The basin contains portions of the cities of Akron
and Cleveland and is one of the most heavily industrialized areas in the
United States. The basin has a humid climate with precipitation distributed
fairly uniform throughout the year. Mean monthly precipitation values at
Cleveland vary from a minimum of 5.92 cm (2.33 inches) in February to a
maximum of 8.86 cm (3.49 inches) in May. The average annual precipitation
76
-------
1 V) t
tu
=
UJ
f- Q
< UJ
O N
Q >- e ^-
Hits
5 g g =
u 5 § §
-o- -J^- .•! S
z
o
£E
**
f-
J-
LU
_J
D
0
UJ
_J
1NGFIELD
IT
O.
Z
o
_J
(/>
p.
t—
or
>
te
<
o
0
X
>- o
D tc
o s:
UJ **
ig
t 0
_i o:
UJ
(E
o
c
0
s
^
(T
>
5
<
0
r
o
UJ
H
10
o
Q.
tr
s
cc
r
K-
tc
UJ
AHOGA R
o
^
5
« j
tf
— **
< I
o *"
>. «
D ^
0 <
N^ I! ^>^Lj3i\r
o
OJ
0)
pi
18
60
77
-------
at Cleveland, based on a 96-year period, is 89.31 cm (35.16 inches). A large
portion of the basin has been urbanized and the average population density
is about 115 persons per square kilometer (300 per mile2). Flood peaks near
the mouth resulting from rainfall over the entire basin occur principally
as a result of runoff from the downstream portion of the basin. In recent
years, runoff above Old Portage has contributed only 10 to 20 percent
of the maximum discharges recorded in the downstream basin.
Streamflow in the Cuyahoga River Basin follows a characteristics seasonal
pattern. Fall and winter flows are generally low. There is a marked rise
in discharge during March and April by runoff from the winter's melting snow-
pack and ice cover. Runoff is normally well sustained during April, May, and
June. During late summer the streamflow is quite low. Heavy rains may cause
sharp rises during any of the spring, summer, and fall months. Flows at
Independence have varied from a minimum of 0.4 m^/sec (14 cfs) on November 30,
1930 to a maximum of 701.8 m3/sec (24,800 cfs) on January 22, 1959. The
average annual runoff from the basin upstream from the gauge is about 34.56
cm (14 inches).
In 1952 the Department of Agriculture made a study of the erosion and
sediment damage which occurs in the Cuyahoga River watershed. The results
of the watershed study shows that 28.1 percent of the total sediment reaching
Cleveland Harbor comes from a stream bank erosion; 15.5 percent is contributed
by sheet erosion; 8.1 percent comes from flood plain scour, and 0.3 percent
is from valley trenching and gully erosion. The remaining 48 percent was
estimated to be supplied by municipal and industrialCSO wastes. The Department
of Agriculture study also showed that 6.5 percent of the total sediment load
upstream from the harbor originated as industrial and domestic waste,
Presumably, some of all or this would have to originate at Akron and the
small towns in between.
Supplemental Data
Some field data were required to develop the sediment model. These
data include information on the nature of the sediments in the study reach
and descriptions of the river cross sections.
A field trip was conducted during the last week of November 1978 to
obtain the required data. Sutron personnel traveled to Cleveland and sur-
veyed 11 cross sections of the river at all highway bridges from Old Portage
to Independence. At the same time, representative samples of the bed and
bank material were collected. On route to Cleveland, the USGS office^in
Columbus, Ohio, was visited and published flow and sediment data obtained.
While in Cleveland, the U.S. Corps of Engineers was contacted and a number
of additional river cross sections obtained for the reach of the river below
Independence into Cleveland Harbor. Operating personnel at the Akron
sewage treatment plant were contacted. They indicated that little was known
about the sediment load from their outfall, particularly at times of high
flow. Turbidity records were available but these were considered to be of
little value. The plant operators indicated that bypassing of the plant was
78
-------
not uncommon during wet weather. Bypassing activities resulted in sizable
quantities of raw waste being dumped in the river. Holding ponds were under
construction to help alleviate this problem.
Modeling Program Overview
It was realized from the outset that an absolutely accurate model of
the Cuyahoga could not be developed with the available budget. The objective
was to create a reasonable model which could be used to assess the capability
to predict the fate and effects of CSO sediments. In this case, the CSO
resulted from the bypassing of the Akron treatment plant at as many as 37
overflow plants.
The modeling program proceeded through a series of distinct stages. The
first step was to analyze the field data and create a flow model of the study
reach. Next, several velocity-depth files were created for use in the sedi-
ment model. Finally, a number of experiments were conducted with the
sediment model.
Initial sediment model experiments concentrated on determining if it was
possible to figure out where sediment material from the Akron plant would
deposit. Next, experiments were conducted to determine how the deposits
moved under unsteady flow conditions. Finally, experiments were conducted
to determine how particles with specific gravities other than that of sand
(2.65) would move down the river. The flow and sediment modeling efforts
will be described in detail in the following sections.
Flow Model
The writer's experience indicated that the first step in developing a
successful unsteady flow model is to obtain successful backwater (steady
flow) profiles of the study reach. The velocities and depths from these
study flow calculations are from the initial conditions for the unsteady flow
model. Areas of possible instability in the unsteady flow model can usually
be identified by careful examination of the steady flow profile for anomalous
depths or unusually high velocities.
The cross section information gathered by Sutron was combined with the
data from the Corps of Engineers to create a complete bed profile of the
Cuyahoga from Old Portage to the estuary. When cross sections were widely
spaced (greater•than 3 miles) additional sections were interpolated. A
stable, reasonable appearing backwater profile was obtained using 29 cross
sections at from 0.4 to 3.22 km (1/4 to 2 miles) spacing. Two profiles
for 3.54 and 25.5 m3/sec (125 to 900 cfs) are illustrated in Figure 14.
These flow values were typical of those observed during the summer.
The three most significant features of the backwater profile are the
two small impoundments at 16.1 km and 30.6 km (10 miles and 19 miles) below
Old Portage plus the estuary which begins at 46.7 km (29 miles). These
areas of increased depth and lower velocity act as sediment traps and are
very important in considering the fate of sediment materials from the STP,
as will be seen shortly.
79
-------
S-
OJ
. o
o
-------
With steady flow established, the next step in developing the flow model
was to select a period of historic record for us-e in calibration and testing.
A period was selected from the summer of 1976, Sutron had processed a month
of hourly flow and dissolved oxygen data for both Old Portage and Independence
for this time period as part of an earlier study. This period of record for
the gauge at Old Portage is illustrated in Figure 15. Discharge only for the
same time period at both Old Portage and Independence is illustrated in Figure
16. The period selected for use in the model is shown on both figures. There
were two reasons for selecting this time period. First, sizable DO deficits
occurred at Independence at the time of the small hydrograph peaks. The
writers wished to examine the sediment discharge at that time. Second, the
much larger hydrographs beginning on the 22nd day exceeded bank fill stage.
This was beyond the range of the surveyed cross sections. An additional
advantage of this time period was the period of nearly steady flow from day
14 onward. This meant that the reach was nearly in equilibrium which is
what was assumed for model initial conditions.
Additional discharge information was required to complete the flow model.
The flow from the Akron STP was required as well as the tributary inflow
from Tinkers Creek, a large stream which enters above the Independence gauge.
Estimates of these flows were obtained from USGS daily flow records. Figure
17 illustrates the daily average flow at Old Portage, Ira, and Independence,
as well as the discharge from Tinkers Creek and the sediment flow at Old
Portage. The difference in flows between Old Portage and Ira was used as
the treatment plant outflow. This is quite reasonable as no major tributaries
other then the STP outfall enter the Cuyahoga in that reach and no bypasses
occurred during the model period. The use of the sediment data will be
explained later.
The very steep reach between kilometer 16.1 and 32.2 (miles 10 and 12)
as well as the sharp breaks in slope at the small control structures required
small time stages in the flow model. One half hour was used with good
results. All flow values were interpolated, to this time increment for use
as input data.
No great effort wasmade to calibrate the flow model. The daily average
values from Tinkers Creek filtered out all detail. A number of other siz-
able tributaries as well as withdrawal points occurred in the reach. No
data from these were available. A comparison was made, however, to see that
the predicted stage changes at Independence were reasonable. The results
of this comparison are illustrated in Figure 18. Also illustrated on the
figure is the dissolved oxygen level and the saturation dissolved oxygen
level plus the predicted sediment load. The figure indicates that the pre-
dicted stage variations are at least reasonable, if not perfect. The
inability of the model to predict the second small rise in stage is felt to
be due to the absence of hourly data from Tinkers Creek (that is, real
hourly data, not the one-half hour interpolated values). The sediment and
dissolved oxygen values will be discussed later.
After the flow model was successfully developed, it was desired to have
a larger size hydrograph for use in the sediment model.- The idea was to
determine what magnitude of flows might cause sediment deposits at the outfall
81
-------
oo
I
13
3
H
o
s
en
I
m
I
OJ
60
0)
ctf
4-J
cfl
O
w
nJ
•s
CO
•H
P
M
I
M
•rl
oa
ivs
oa
CXI '
82
-------
M-
O
a)
I
0
0
w>!
oo
>
O
60
•H
(puooss uaj
9uaBijosi.Q
83
-------
I
200
150
100
50
Daily Oischaraes (June 19, 1976 to July 18, 1976)
Upper Line - Cuyahoqa at IRA
Lower Line - Cuyahoqa at Old Portaqe
, Model ^H
Period H~^
eooo
5000
-4000
^
t/i ro
•r- T3
3000
^ £ 2000
T3
OJ
O'*
1000
Cuyahoqa at
Old Portaoe
-Model.
Period
200
150
co 100
VI
o
50
Upper Line - Cuyahoqa at Indeoendence
Lower Line - Tinkers Creek
^_ Model .
Period
O 2 4 68 10 12 14 16 18 20 22 24 26 28 30
Time From Start of Period, Days
Figure 17. Supplemental Discharge Data
84
-------
OJ
>
d)
r>
8
13
0)
O
CO
CO
•H
p
oo
iH
(U
O CO
-------
and other locations to be reentrained and move on downstream. In order to
select a reasonable hydrograph, a simple flood frequency analysis was
run on the daily average maximum flows at Old Portage. This analysis is
illustrated in Figure 19.
3
Initially, two peak flows were selected. These were 84.95 m /sec „
(3,000 cfs) with a recurrence interval of roughly five years and 127.4 m /sec
(4,500 cfs) with a recurrence interval of roughly 50 years. Unfortunately,
the 50-year flow exceeds bank full stage by a sizable margin. This exceeded
the capabilities of the cross section in the flow model. An artificial
hydrograph based on Pearson Type III statistical distribution with a peak
flow of 84.95 m^/sec (3,000 cfs) and an overall time from rise to recession
of 48 hours was used in the high flow experiments.
The net results of the flow model development was disc file containing
velocities and depths at all 29 model cross sections for a period of 8-1/3
days. The first six days were "real," 'representing the calibration data.
The remaining 2-1/3 days were the artificial hydrograph with a recurrence
Interval of five years. The philosophy was that a qualitative determination
could be made of when and to where sediment deposits formed during the "real"
period moved during the five-year flood.
Some general observations from the flow modeling portion of the study
are:
—The bed profile of the river is of great importance and should
be established as accurately as possible.
—Major manpower efforts are required to obtain flow data at
small time intervals.
—If steep reaches such as below mile ten are of interest, a revised
backwater program would be desirable. Considerable difficulty was
encountered getting the flow model to start correctly there
because of superficial flow. The backwater program used in this
study assumed all flows to be subcritical and computefi all profiles
working upstream,
—Other than the above, the technology for modeling flow is highly
developed and reasonably easy to apply. Costs are reasonable,
being roughly $25 to $40 for the 8-1/3 day period of 1/2 hour
values.
The following section describes the sediment model development and experi-
mental results.
Sediment Model
Model Setup—
The biggest single difficulty encountered with the sediment model was
lack of information. A great deal had to be assumed or calculated based on
reasonable guesswork.
86
-------
\
s
o
in
c\j
ri
CD
CM
LO
CD
-.
« «
be 6
J§ 1
. 0}
4-1 cd
n) !-(
CD g
O
O
a)
I
•H
87
-------
The sediment model required much the same information as the flow model
plus additional information on sediment inflows. The stream bed profile
and cross section information were available from the flow model, as well
as the disc file of velocity and depth data. The sediment inflow was developed
from USGS data and from information presented in Section 6.
The sediment inflow at the Old Portage gauge was fairly easy to obtain.
This was plotted in Figure 16. There were two disadvantages. The data
were daily averages and only total load was given. Only isolated samples
were analyzed for size distribution. The model required 1/2-hour inputs
of sediment in each desired size class (six size classes were used). These
two disadvantages were overcome by developing six sediment rating curves.
A year of daily sediment discharges was plotted versus daily discharges.
This gave a total suspended load rating curve (note that since no bed load
data were available, the suspended load was assumed to be the total load).
Table 21 from USGS analysis lists the distribution of sizes found in
the suspended load.
TABLE 21. SIZE DISTRIBUTION OF SUSPENDED SEDIMENT
LOAD AT OLD PORTAGE
Sieve Size, mm
Percent
Classification
Less than .002
.002
.004
.008
.016
.031
.062
.125
35 CLAY-SILT
7
9
13
13
5 ,
s
7 SAND
1
These data were used to develop rating curves for each size class by multi-
plying the total load rating, curve by the percentage in each class. In
the model, no sieve size larger than 0.062 was used. Thus, 89 percent of the
total load was assigned to the 0.62 mm and smaller sizes and 11 percent was
carried in the geometric mean of .062 and .125 size classes (0.18 mm = /.062*.125)
The sediment model also required an initial condition. This was the
percentage of sediment present in each size class at each model cross
section. These data were obtained from the sediment samples collected by
Sutron. Since the samples were few and widely scattered, they were all
88
-------
averaged and the same size distributions used at all cross sections. Table
22 lists the values used. In the absence of actual data, the size distri-
bution of the sediment material from the treatment plant had to be estimated
Based on the data from Section 3, District of Columbia street solids were
selected as typical. Table 23 lists the percentages which were assigned to
the sediment from the Akron STP.
TABLE 22. DISTRIBUTION OF SIZES IN BED SEDIMENT
Size Class, mm
Percent Assigned
to Class
Classification
,063
.180
.350
1.770
7.070
100.000
.52
1.07
3.97
12,78
40.00
41.66
a
|
SAND
1
I
T
GRAVEL-COBBLE
^
TABLE 23. DISTRIBUTION OF SIZES IN AKRON STP FLOW
Size Class, mm
Percent Assigned
to Class
Classification
.063 28 1
.180 20 SAND
. 135 37 •
1.770 15
I/
The important thing to note about the three size distributions is their
sizable difference. The upstream inflow is virtually all fine material.
The inflow from the treatment plant is primarily sand (remember, this is
hypothetical). The channel bed is primarily gravel and cobble sizes.
The method of determining the quantity of sediment discharges from the
treatment plant should be mentioned. The data in Section 3 indicated that a
89
-------
concentration of 589 mg/& is typical of CSO solids. The STP flows (Cuyahoga
at Ira minus Cuyahoga at Old Portage, Figure 16) were mulitplied by this
concentration times the appropriate percentage and converted to pounds per
second inflow. At the end of the calibration flow (six days) the sediment
flow from the STP was held constant. A worthwhile program modification would
be a sediment rating curve for the STP flow,
One of the most unfortunate aspects of the entire study was the lack of
sediment outflow data from the study reach. Initial conversations with
the USGS left the impression that sediment data were collected at both Old
Portage and Independence. They are, but unfortunately not in the same year.
This made it impossible to calibrate the sediment transport rate,
CSO Sediment Transport Experiments—
Normal sediment experiments—^
The initial experiments performed with the sediment model were to deter-
mine if the model could distinguish between sediment deposits resulting from
the STP flow and deposits resulting from the natural sediment flow at Old
Portage. The results of these experiments are illustrated in Figures 20 and
21. Figure 20 illustrates the deposition and erosion patterns in the channel
at three different times with zero sediment flow from the STP. These are
at 1,800 minutes (1-1/2 days), 9,000 minutes (6-1/4 days), and 12,000
minutes (8-1/3 days) after the beginning of the model. These times correspond
to the end of the period of study flow just prior to the first "real" hydro-
graph, the time immediately before passage of the 3,000 cfs, five-year flood
hydrograph, and the end of the five-year flood hydrograph. Deposition/erosion
depths are shown at each cross section for each of the six size classes
modeled.
The results indicate that during the steady flow period preceding the
hydrographs, no sediment was transported. That is, the solid lisie coincides
with the graph axis. The passage of the "real" hydrograph moved fine mater-
ial C-063 mm) from the controls for the two small impoundments and the steep
slopes downstream. Small amounts of movement occurred at the control points
in the 7,07 mm size class. Note that .001 ft. of the degradation is less
than one particle diameter and thus is not measurably significant. The
passage of the five-year flood results in further erosion in the same location.
The higher velocities move the fine material deposits further out into the
estuary. In viewing these graphs, recall that virtually all the natural
inflow was fine material and the bed material is mostly course.
Figure 21 illustrates the same information as Figure 20 but with sedi-
ment inflow from the STP included. The deposit areas for the STP material
are easily identified because of the sizes involved. Most of the material
was in the .18, .35, and 1,77 mm size classes. Thus, any deposition is due
to the STP. The results indicate that prior to the passage of the "real"
hydrograph, the STP sediments deposited at the outfall. The passage of the
real hydrograph moved some of all the .18 - 1.77 mm material down to the
first small retention structure at mile ten. The passage of the five-year flood
hydrograph resulted in the .18 mm size class of the STP sediments reaching
90
-------
UJ
c
0)
u
220
2OO
18O
160
' Trap
Area
1.5,
J
03
0
-03,
Q6
O
-O.6
0-3
O
-O.3
0.03
i-
\
}
-0.03
0,03
O
-O.03
CUYAHORA RIVER
•Approximate Water Surface
Trap
Area
Estuarv
' ?
KRON STP
o 10
I
I
2O 30 4O 50 60
Distance Below Old Portage (km)
Sediment Size
.-*i^:^_ 0.063 mm
Fines Hove Out Into Estuary
Afer Flood -^—^—^—^
mm
.1770mm
.707O mm
JOQQGOmm
—•—— Tipie 5= 1,800 minutes
,.,.,.,. Time = 9,000 minutes
_.—._ Time s 1,200 minutes
Without Lateral Sediment Inflow
Figure 20.Deposition and Erosion
91
-------
s~
e = 9,000 minutes
Time = 12,00n minutes
V'ith Lateral Sediment Inflow
Figure 21. Deposition and.Erosion
92
-------
the second retention structure at mile 20. These experiments indicate that
it is qualitatively possible to predict the fate of the STP sediments.
The behavior of the channel bed at the STP outfall for the entire time
period of the model is illustrated in Figure 22. Shown in the figure are
the overall channel, the discharge hydrographs (bottom), and the cumulative
deposition or erosion at the treatment plant due to each size fraction. The
behavior of the various size classes is clearly evident. No fine material
(0.63 mm) is ever deposited. At the beginning of the "real" hydrograph,
fine material begins to erode. The passage of the five-year flood accelerates
the erosion. The STP sediments CL8 to 1.77 mm) accumulates rapidly prior
to the "real" hydrograph, cease accumulating or erode slightly (1.77 mm) as
the hydrograph passes, and erode away as the five-year flood passes. The inflow
rate of .35 mm sand was sufficient to prevent complete scouring of the
deposits by the five-year flood. The behavior of the channel bed at cross
section 11, the first small impoundment, is illustrated in Figure 23.
It was concluded from these experiments that the model, coupled with
flood frequency analysis, gave a reasonably powerful tool for analyzing the
fate and resting time of deposits formed from the STP sediments. One
condition which should be further investigated is the difficulty of separat-
ing the STP deposits when the STP sediments are identical to those in the
flow or the channel bed or both. The Cuyahoga may be unique in the way the
three size distributions vary. No general conclusions should be drawn about
the fate of STP solids from these experiments. The Cuyahoga is fairly steep
by stream standards. The movement pattern in a river of very flat slope
comparable to that of the lower Mississippi with different sediment sizes
should be investigated. Each STP and CSO will certainly be unique.
Particles with other specific gravities—
After determining that the model was capable of qualitative fate pre-
dictions for sand-like particles (recall that up to now all particles are
quarts spheres of 2.65 specific gravity) experiments were conducted with
sediments of varying specific gravities. The model was modified so that a
specific gravity could be read in for each size class. The subroutines which
calculated critical shear stress and fall velocities were modified accordingly.
The reasoning behind the modification was to determine if the model could also
be used to make qualitative judgements concerning the fate of heavy metal
pollutants on other materials different from sands.
After the model was modified, small quantities of 0.63 mm particles
with the specific gravities of lead and iron were introduced at the STP.
The small size was used because pollutants are often adsorbed to fine particles.
Flow conditions were identical to those used in earlier experiments. Note
that under no circumstances do these experiments imply the presence of lead
or iron in the outflow from the Akron STP. No actual data were available.
Also, it should not be assumed that lead or iron would normally be tound in
their pure form. These elements were picked strictly because they represent
a wide range of specific gravity.
93
-------
s-
-------
ai
220
200
180
16O
0.06.
0-
-0.06.
0.15-
CUYAHORA RIVER AT X-SEC 11
(With Lateral Sediment Inflow)
Trap
Area
Estuary
6 1O 20 30 '4O 50 6O
Distance Below Old Portane, km
Particle Size
i.0.083 mm
.0.180 mm
i_Q350mm
.1.770 mm
ZO70 mm
2000 4000 6OOO 800O 1000O 1200O
Time, min
Figure 23, Deposition and Erosion Ten Miles Below Old Portage
95
-------
Figures 24 through 26 illustrate the results of the experiments. Figure
24 illustrates the deposition areas for the two types of particles prior to
the "real" hydrograph, after its passage, and after the five-year flood. The
particles with the specific gravity of iron accumulate at the outfall and at
the 16,1 km (10 mile) point prior to the real hydrograph. The "real"
hydrograph transports most of them to the 16.1 km (10 mile) point and the
five-year flood carries them to the estuary. The particles with the _specific
gravity of lead accumulate at the outfall and move to the 20 mile point with
the passage of the five-year flood.
Figures 25 and 26 further illustrate the movement of the hypothetical
particles. Figure 25 illustrates the aggradation at the STP as a function
of time. The hydrograph is also shown. Figure 24 illustrates the aggra-
dation and erosion at the 16.1 km (10 miles) point. The more ready move-
ment of the particles with iron's specific gravity can be clearly seen.
From these experiments it was concluded that the model was capable of
qualitative predictions of the fate of heavy particles. Quantitative
predictions will require data which identify the size characteristics and
specific gravities of pollutant material characteristics of the outfall
(STP or CSO) being modeled. The following section describes methods of
obtaining required data.
96
-------
o
+J
m
22O
2OO
18O
160
Trap
Area
CUYAHORA RIVER
Aonroximate Water Surface
Trap
Area
. AKRONSTP
INDEPENDENCE
0 1O 20 30 40 5O 6O
Distance Below Old Portane, km
Delta Z, centimeters
O.06-I
oi-
1
-0.061
O.06,
nl-
1
-O.06J
Specific Hracjty of Iron
A ^
Specific r*avitv of Lead
•*.
Heavy Particle Inflow at STP
Particle Size
Time = 1,800 minutes
Time = 9,000 minutes
Time = 12,000 minutes
Figure 24, Deposition and Erosion of Heavy Particles
97
-------
22O
2OO
18O
160
Trap
Area
CUYAHORA RIVER AT AKRON STP
(With Lateral Sediment Inflow)
•Annroximate Water Surface
Tran
Area'
O 1O 2O 3O 40 50 6O
Distance Below Old Portape, km
01
u
Q06,
°r
•OQQl
OD6i
O-
HEAVY PARTICLES
Snecific Gravity
Iron
Particle Size
_O.O63mm
Soecific Rravity of Lead
D063mm
E
C
20-
2000 4000 6000 8000 100QO 12000
Time, minutes
Figure 25. Deposition-and Erosion of Heavy Particles at STP Outfall
98
-------
01
e
220
200
180
160
Trap
Area
CUYAHOGA RIVER AT X-SEC 11
(Hith Lateral Sediment Inflow)
Approximate Mater Surface
Trap
Area'
0 10 20 30 40 50 60
Distance Below Old Portane, km
HEAVY PARTICLES
8
-------
SECTION 5
RELATED ASPECTS OF SEDIMENT TRANSPORT STUDIES
INTRODUCTION
The bulk of this report has dealt with three subjects. These are
characterizing the sediment materials which come from combined sewer over-
flows, assessing their impact, and evaluating the technology for modeling
the movement of these materials. This final section of the report deals
with data problems.
Sediment modeling requires a significant amount of data on channel
geometry, streamflow, sediment characteristics, and sediment movement. The
techniques for obtaining geometry and streamflow information are widely
understood. Stream geometry is obtained by standard surveying techniques
or from maps. Streamflow data are routinely collected by the USGS and
others by means of depth recorders and rating tables based on current meter
measurements. Methods for measuring the rate of flow and movement of sedi-
ment are less well known.
The purpose of this section is to briefly summarize the techniques
used to:
—sample sediments
—determine sediment flow rates
—characterize sediments
—trace sediment movement
The discussions are fairly brief and are designed to aid someone unfamiliar
with sediment studies with an understanding of how to collect and use data
for model studies.
Numerous references are available concerning sediment transport. The
following discussions are based on several books in the writers' personal
library and on publications of the United States Geological Survey.
Probably the best general reference book on sediment transport is (54) from
the American Society of Civil Engineers. Chapter III is devoted entirely
to sampling techniques and their limitations. A second good book is (61)
by Graf. Graf's Chapter 13 covers essentially the same materials as (5)
but in greater detail. The book also has a somewhat more international
flavor. Other good references are (61) and (63) by Simons and Senturk and
100
-------
Bogardi. Bogardi gives one of the few descriptions of tracer techniques.
The best detailed references on sediment analysis techniques are published
by the USGS. References (.64 and 65) on sampling techniques and analysis
procedures are particularly useful.
SAMPLING SEDIMENTS
In general, two different but related sampling problems will be encoun-
tered in a model study, The first problem is to determine what type and
quantities of material make up the channel bed and banks. The second
problem is to determine what type and quantities of material are transported
by the flow, Sampling the channel bed and bank material is fairly straight-
forward, as the composition is usually fairly stable with time. The quantity
and type of material in the flow varies constantly both in time and space.
In addition, a distinction is made between material actually suspended in
the flow and that which moves mostly in contact with the bed. The second,
is, thus, the more difficult problem. Both will be discussed here.
Obtaining samples from the bed and flow will be discussed first. The process
of analyzing the samples is largely the same regardless of the source and
will be discussed last.
Bed and Bank Material Samplers
The methods used to sample bed and bank material of streams are
generally unsophisticated„ The determining factor on the method used is
somewhat a function of the size distribution of the particles.
The common shovel is an adequate tool to obtain most bank material
samples in material up to 7 to 10 cm (3 or 4 inches) in diameter. Represen-
tative areas are picked and shovelfuls transferred to watertight sample
bags. While these samples are not taken underwater, the watertightness
is desirable to prevent the loss of the fine fractions. Note that a
shovel is not a generally satisfactory tool for sampling underwater. Most
of the fine material is washed away when the sample is withdrawn.
When bed material exceeds 7 to 10 cm (3 or 4 inches) sampling becomes
more difficult. The USGS (66) describes an optional method for use in
coarse sizes. All that is required is a photograph of the bed with a size
scale such as a survey rod or ruler included. It is unlikely that such
methods would be necessary in the average model study. The technology
for modeling transport of coarse materials is limited, A bed consisting
largely of 10 cm (4 inches) and up material would probably be treated as a
rigid boundary.
The problem of sampling a streambed beneath flowing water has been
studied in considerable detail. Several standard samples are available
for this purpose. Guy and Norman (64) and ASCE (54) provide good descrip-
tions. Reference (64) is considerably more detailed. Before describing
the samplers, a short discussion from Guy and Norman concerning the standard
codes used to identify approved sediment samples will be presented.
-------
An array of standard samplers and methods have been developed by the
Federal Inter-Agency Sedimentation Project (FIASP) of the Inter-Agnecy
Committee on Water Resources, located first at Iowa City, Iowa and since
1948 at the St. Anthony Falls Hydraulic Laboratory in Minneapolis, Minnesota,
Their reports cover almost all aspects of measurement and analysis of sedi-
ment movement in streams. A complete catalog of these reports is available
in reference (11). The reader should refer to the FIASP reports for further
background material and details on the standard samplers. Samplers carry
the following coded designation:
US - United States standard sampler (after first use in designating
a sampler in this chapter, it will usually no longer be included
as part of the designation)
D • - depth integrating
P - point integrating
H - hand held by rod or rope (for cable-and-reel suspension, the
H is omitted)
BM - bed material
U
- single stage
YEAR - year (last two digits) in which the sampler was developed.
The samplers described in this section are physically limited to those
capable of collecting bed-material samples consisting of particles finer
than about 30 or 40 mm in diameter. There may also be limitations with
respect to some very fine sediments for some of the samplers. The collection
and analysis of material larger than coarse gravel logically becomes more
difficult and costly because other techniques are required to avoid handling
heavy samples with larger and more expensive equipment. Because of the
difficulty in measuring large sizes, little information regarding size distri-
bution is available on streams having gravel, cobble, and boulder beds.
Bed material samplers, as first developed, may be divided into three
types: the drag bucket, grab bucket, and vertical pipe. The drag-bucket
sampler consists of a weighted section of cylinder with an open mouth and
cutting edge. As the sampler is dragged along the bed, it collects a
sample from the top layer of bed material. The grab-bucket sampler is
identical in principle to the drag-bucket sampler. It consists of a section
of a cylinder attached to a rod and is used in the collection of samples from
shallow streams. Both samplers are operated by dragging upstream so that the
mouth is exposed to the flow, which results in the loss of some fine material
while in transit from the stream bed to the water surface. The vertical-
pipe or core sampler consists of a piece of metal or plastic pipe that can
be forced into the stream bed by hand.
102
-------
The drag and grab-bucket samplers are either too cumbersome to handle
or do not obtain representative samples of the bed material. The vertical-
pipe samples is satisfactory for use in shallow streams (54).
The Federal Inter-Agency Project has developed three types of instru-
ments for sampling the bed material of streams where most of the material
is finer than medium gravel,' The smallest of the three, designated as the
"US BMH-53" (see Figure 27), is designed to sample the bed of wadable
streams. The collecting end of the sampler is a stainless steel thin-
walled cylinder 2 inches in diameter and 8 inches long with a right-fitting
brass piston. The piston creates a partial vacuum above the material being
sampled and thereby compensates in a reverse direction for some of the
frictional resistance required to push the sampler into the bed. This
partial vacuum also retains the sample in the cylinder while the sampler
is being removed from the bed.
The bed material of deeper streams or lakes can be sampled with the
US BHM-60 (see Figure 28). This is a hand-line sampler about 22 inches
(56 cm) long, made of cast aluminum, has. tail vanes, and is available in
weights of 30, 35, or 40 pounds (13.6, 15.9, or 18.2 kg). Because of its
light weight, its use should be restricted to streams of moderate depths
and velocities and whose bed material is also moderately firm and yet does
not contain much gravel.
The sampler mechanism of the US BMH-60 consists of a scoop or bucket
driven by a cross-curved constant torque, motor-type spring that rotates
the bucket from front to back. The scoop, when activated by release of
tension on the hanger rod, can penetrate into the bed about 1.7 inches (4.3
cm) and can hold approximately 175 cc of material. The scoop is aided in
penetration of the bed by extra weight in the sampler nose (67).
The bucket closes when the sampler comes to rest on the streambed. A
gasket on the closure plate prevents trapped material from contamination
or being washed from the bucket.
Except for streams with extremely high velocities, the 100-pound cable-
and-reel suspended BM-54 sampler (Figure 29) can be used for sampling bed
material of streams and lakes of any reasonable depth. The body of the BM-54
is of cast steel. Its physical configuration is nearly identical with the
cast aluminum BMH-60, 22 inches (56 cm) long and with tail vanes. Its
operation is also similar to the BMH-60 in that it takes a sample when
tension on the cable is released as the sampler touches the bed, The
sampling mechanism externally looks similar to that of the BMH-60, but its
operation is somewhat different.
In the event that core samples are needed in deep flowing water, a *
sampler has been developed and extensively used in studies of the Columbia
River Estuary by Pyrch and Hubbell (68). This cable-suspended sampler
collects a 4.76 cm (1-7/8 inch) diameter by 1.83 m (6 foot) long core by a
combination of vibration and an axial force derived by cables from a 113 kg
(250 pound) streamlined stabilizing weight.
103
-------
-*•-
•it
ed
•H
O)
PQ
ro
m
CO
CN
0)
S-i
M)
•H
104
-------
Figure 28.
Hand-Line Spring-Driven Rotary-Bucket 30-Pound
Bed-Material Sampler, US BMH-60 (11)
PLAty
ELEVATION
SECTION A-A
Figure 29. US BM-54—Bed Material Sampler (54)
105
-------
In terms of collecting data for model studies, the three US standard
samplers should be adequate in most cases. Long core samples would most
likely be required only when studying long-term deposition, rates in lakes
or other impoundments. Such measurements would be connected with studies
extending over several years,
A short discussion is in order concerning where to sample. Model
studies such as discussed earlier in this report descretize the stream into
fixed cross sections. Customarily, the sediment characteristics assigned to
a single cross section are assumed to be representative of the stream half^
way to the two adjoining sections. The objective of sampling is to produce
values consistent with the model assumptions. Economic consideration usually
dictate that as few samples as possible be taken. As a minimum, the writers
prefer at least one sample from the channel and one from the overbank area
at each cross section. If dollars allow, more samples should be taken in
the channel to define the variations along the cross section. This is more
important when the stream in question is always muddy and a visual determina-
tion of the nature of the bed is not possible.
Sampling Sediment Transported by the Flow
The objectives of any sampling of the material which flows by a sedi-
ment laden stream are twofold. First, it is usually desired to quantify
the rate at which sediment is being discharged, usually in terms of weight
per unit time. Second, it is usually desired to know what sizes of particles
are being transported and what quantity. In order to successfully accomplish
these objectives, it is necessary to understand how sediment materials are
distributed in flowing water.
Depending on the size of each sediment particle, the stream transports
the sediment by maintaining the particle in suspension with turbulent
currents or by rolling or skipping the particles along the streambed
(Saltation). The finer sediments move downstream at about the same velocity
as the water, whereas, the coarsest sediments may move only occasionally
and remain at rest much of the time (69), While material is transported
in suspension, saltation, and rolling and sliding on the bed is also occurring.
The different modes of transportation are closely related and it is difficult,
if not impossible, to separate them completely. The borderline between con-
tact load and saltation load is certainly not well defined. It is indeed
hard to picture a particle rolling, on the bed without at some time losing
contact with the bed and executing short jumps. In a similar manner, the
distinction between saltation and suspension is also not definite.
The term "bed load" is defined as material moving on or near the bed.
The total load is made up of the bed load and suspended load. In addition,
the total load is divided into "bed sediment load" and "wash load," which
are defined as being, respectively, of particle sizes found in appreciable
quantities and in very small quantities in the shifting portions of the
bed. Obviously, both the bed sediment load and the wash load may move
partially as bed load and partially as suspended load, although by definition,
practically all the wash load is carried in suspension.
106
-------
In terms of a model study, the term of most concern is "sediment dis-
charge" which is defined as the quantity of sediment per unit time carried
past any cross section of a stream* The term should be qualified. For
example, one may refer to the bed load discharge, the bed sediment discharge*
or the total sediment discharge* It will be noted that, as used herein,
the term "load" denotes the material that is being transported, whereas
the term "sediment discharge" denotes the rate of transport of the material
(5). The model described in this report prints 6tit the sediment discharge
for each size class considered. The bed load discharge is calculated
internally and not reported, although it could be if required. Our sampling
techniques are, thus, directed toward measuring the suspended and bed load
discharges or both, at once, if possible*
The distribution Of the suspended-pediment sizes in the vertical
direction may vary from stream to stream and from cross section to cross
section within the same stream* Generally, the finer sediments are dis^
tributed uniformly throughout the vertical, and the coarser particles are
concentrated near the streambed but with some coarse patfticles reaching
the water surface at times (.64), This behavior is illustrated in Figure
30. Figure 30 illustrates the difficulty of Sampling the various sediment
loads. Typical suspended sediment samples cannot reach the lower .09 to
.12 m (.0.3 to 0.4 foot) of flow.
The higher concentration and coarser sizes of sediment passing beneath
the sampler nozzle, in suspension and ott the bed, are difficult to measure.
This unmeasured part may or may not be a sizable part of the total sediment
(measured plus unmeasured) and is sometimes computed empirically for lack
of suitable sampling devices.
At some stream cross sections -, all the sediment sizes being transported
may be thrown into a fairly uniform suspension throughout the entire
vertical by natural or artificial turbulence. The measured part at these
sites is representative of the entire vertical and represents total sediment
discharge (64) . When locating field measuring sections, such places are
highly desirable. Locations where all the sediment is uniformly suspended
offer the maximum measurement accuracy and require the fewest samples.
Two types of samplers will now be discussed. The first are suspended
sediment samplers designed to measure from the water surface down to the
"unsampled zone." The second are bed load samplers, designed to' work at
or in the channel bed.
Suspended Sediment Samplers—
Two types of suspended sediment samplers are generally used, The
depth-integrating sampler collects and accumulates the sample as it is
lowered to the bottom of the stream and raised back to the surface* The
sampler must be moved at a uniform rate in a given direction but not
necessarily at equal rates in both directions.
107
-------
M) |
d e
•H O
iH CJ
6.
0 4->
cd fj
Cfl 0)
Cd T3
0) (1)
cti cd
.13 IS
«H O
tH
CO IJ4
0)
a IH
O O
Cs!
(30 4-1
a -H
•H O •
rH O /-N
ft rH iH
S 0) ,H
cd > v-x
Crt
o a
T3 4J O
0) -rl
5-1 4J -U
3 a cd
co o) }-<
cd ex +j
cu to a
g£ S
T3 4-J
p! -H
cd &
cu cd
5-1 O
3 -H
CO 4-1
cd H
O
CO
CD
60
•H
108
-------
The point integrating sampler, on the other hand, is designed to collect
a time integrated sample at a single point in the flow. It can be operated
to obtain a depth-integrating sample from deep or swift streams by holdling
the valve open while integrating the stream depth in parts.
Where streams can be waded, or where a low bridge is accessible, a
choice of two lightweight hand samplers can be used to obtain suspended-
sediment samples. The smallest of the two is designated "DH-48" (Figure
31). It consists of a streamlined aluminum casting 13 inches (33 cm) long
which partly encloses.the sample container. The container is usally a round
pint, glass milk bottle. A standard stream-gauging wading rod, or other
suitable handle, is threaded into the top of the sampler body for suspending
the sampler. The instrument can sample to within 3-1/2 inches (9 cm)
of the streambed.
The other lightweight sampler, designated "DH-59," (Figure 32) was
designed to be suspended by a hand-held rope in streams too deep to be
waded. It too only partly encloses the sample container. Because of.its
light weight, it is limited in use to streams with velocities less than
about 5 fps.
These two lightweight hand samplers are the most commonly used for
sediment sampling during normal flow in small and perhaps intermediate
sized streams. Because they are small, light, durable, and adapatable,
they are preferred by hired observers arid fieldmen on rountine or on
reconnaissance measurement trips. At most locations, a heavier sampler
will be needed only for high flow periods.
When streams cannot be waded, but are less than 15 to 18 feet deep,
depth-integrating samplers designated "D-49" can be used to obtain sus-
pended-sediment samples. The D-49 resembles a small bronze submarine.
It weights approximately 62 pounds (28 kg). The design is compatible with
cable and reel suspensions used with current meters.
Point-integrating samplers are more versatile and, consequently, more
complex than the simpler depth-integrating types. They can be used to
collect a sample that represents the mean sediment concentration at any
selected point beneath the surface of a stream except within a few inches
of the bed, and also to sample continuously over a range in depth. They
are used for depth-integration in streams too deep (or too swift) to sample
in a round-trip integration. In depth integration, sampling can start at
any depth and continue in either an upward or downward direction for a
maximum vertical distance of about 30 feet.
The US P-46 consists of a 100-pound (46 kg) streamlined cast bronze
shell, an inner recess to hold a round pint milk bottle, a pressure
equalizing chamber and a tapered three position rotary valve operated by
solenoid which controls the sample intake and air exhaust passages.
The 105 pound (48 kg) US P-61 (.Figure 33) is similar to the P-46,
but is simpler and somewhat less expensive. It can be used for depth
109
-------
Figure 31. Depth-Integrating Suspended-Sediment Wadiiig-
Type Hand Sampler, US DH-48
N
Figure 32, Depth-Integrating Suspended-Sediment Hand
Type Sampler, US DH-59
Figure 33t Depth-Integrating Suspended-Sediment Cable
and Reel Sampler, US D-49
110
-------
integration as well as for point integration to stream depths of at least
180 feet (55 m) «
The US P-63, a 200 pound (91 kg) electrically operated suspended Sedi-
ment sampler, is tetter adapted to very great depths and high velocities.
The P-63 differs from the the P-61 mainly in size*, weights and in the capacity
of the sample container that can be used* It has the capacity for a quart-
sized round milk bottle4 An adapter is furnished so that a round pint-sized
milk bottle can be used* The maximum sampling depth is about 180 feet
(55 m) with a point sample container and 120 feet (37 m) with a quart con-
'tainer.
The 300 pound (136 kg) US P-50 is designed for use in extremely deep
streams and high velocities» Its operating characteristics are similar to
the P-63 (48). All the point samples are designed for suspension with a
steel cable having an insulated inner conductor core. By pressing a switch
located at the operators station^ an electric current opens a valve to
allow the sample in.
Because of the complex nature of point integrating samples, the reader
may find it necessary to seek additional information from USGS or F1ASP
reports.
Occasionally the mQdeler will need to collect data on sediment inflow
from small tirbutaries* Such tributaries may be difficult to reach or
flow only intermittently. The singles-state sampler, US U-59, Was tested
by the FIASP to meet the needs for an instrument that would obtain some
sediment data on small fast-rising Streams where it is impractical to use
a conventional depth-integrating sampler*
The US U-59 sampler consists of a pint milk bottle or other sampler
container, a 3/16 inch inside diameter copper tube intake. Each tube is
bent to an appropriate shape arid inserted through a stopper which fits
tightly into the top Of the container. There are two general types of this
sampler, one with a vertical intake and the other with a horizontal intake.
The U-59 is illustrated in Figure 34*
When the stream surface rises to the elevation Of the intake nozzle,
the water-sediment mixture enters; and as the water surface continues to
rise in the stream, it also rises in the intake* When the water-Surface*,
elevation W reaches C, flow starts over the weir of the siphon, primes the
siphon, and begins to fill the sample bottle* The sampling operation just
described is Somewhat idealistic because in reality the operation is
affected by the flow velocity and turbulence which alters the effective
pressure at the nozzle entrance.
The U-59 has many limitations with respect to good sampling Objectives.
It must be considered a type of point sampler because it samples a single
point in the stream at whatever stage the intake nozzle is positioned before
a flow event occurs. Its primary purpose is to collect a sample automatically,
and it is used at stations on flashy streams or other- locations where extreme
111
-------
Inner leg
Inner .leg
Air exhaust
, Outer leg
Exhaust port
Crown
' Wier' _?
Intake
Intake nozzle
-D
t >-t
Surge
•w
Sample
container
-Inner end
Figure 34. Components and Dimensions of the Basic Single-
stage Suspended-Sediment Sampler, US U-59, (64).
112
-------
difficulty is encountered in trying to reach a station to collect samples
at appropriate times by the normal procedure with standard equipment.
A number of manufacturers have developed automatic point samplers. In
general, these use timed pumps which extract the sediment water mixture from
a point in the flow and transfer it to one of several bottles in a rotating
rack. These can be quite useful when continuous records fora model are
required. Supplemental measurements must still be made with samplers dis-
cussed earlier to aid in extrapolating the point data to an entire cross
section.
Bedload Samplers—
Up to this point we have discussed means for determining the quantities
of sediment in the upper portion of the depth of a stream (Figure 30). We
will now look at means for determining the quantity of .sediment which flows
near the bed.
At this time the reader may wish to note the difference between bedload
and unmeasured load. Bedload is the sediment that moves in the stream at
velocities less than the surrounding flow by sliding, rolling, or bouncing
on or very near the streambed. The size particles moving as bedload is
identical with samples of bed material in the, movable part of the stream-
bed. Unmeasured load is that sediment which is not measured with the sus-
pended-sediment samplers and consists of bedload particles and particles in
suspension in the flow below the sampling zone of the suspended-sedinrent
samplers (Figure 30).
Bedload is difficult to measure for several reasons. Any mechanical
device placed in the vicinity of the bed will disturb the flow and hence
the rate of bedload movement. Another reason why bedload is difficult to
measure is that the sediment movement and the velocity of water close to the
bed vary considerably with respect to both space and time; and, therefore,
if a good sample could be obtained at a given point, it may not be represen-
tative of the entire cross section for a reasonable interval of time (48).
The bedload discharge, therefore, cannot be determined in the same
manner as suspended-sediment loads through computation by use of suspended
concentration and water discharge data. Thus, a bedload sampler must be
able to effectively trap all particles moving along the bed when and if
they pass over an area of the undisturbed bed in a specified period of
time.
Many agencies all over the world have developed their own bed load
measuring devices. However, these devices are, at least in principle,
quite similar. They can be distinguished as (1) devices called bed load
samplers or catchers, (2) direct observation of the movement of bed forms,
and (3) sonic samplers. Rather than direct measurement, bed load move-
ment can be computed by one or more of several different formulas. CSU
sediment model already incorporates one of these, the Meyer-Peter and Mueller
method. In addition, periodic quantity surveys of sediment deposits may
give a rasonable estimate of the bed load rate.
113
-------
Similar to other hydrometric devices, the bed load samplers must be
calibrated. Calibration of a sampler consists of determining its efficiency
coefficient, that is, the ratio of bed load indicated by the sampler to the
true bed load. Efficiencies of bed load samplers have been determined in
laboratory flumes with fixed and movable beds.
The bed load samplers may be generally classified as the following
types: box- and basket-type samplers, pan-type samplers, and pit-type
samplers. The box-basket samplers consist of a pervious container where bed
load accumulates, of a supporting frame and cables to make the samples
portable, and of a vane to give the sampler the appropriate direction. The
pan-type samplers consist of apan with a bottom and two side-walls. Within
this pan there may or may not be a baffle system to retard the water-sediment
mixture and, thus, trap the sediment. The pit-type samplers consist of
a depression (pit) installed in the channel bottom to catch and accumulate
the bed load which is removed by a mechanical device to obtain a continuous
record. Among the three types of samplers, the box- and basket-type samplers
are most common (54).
The current state-of-the-art of bedload samplers is so poor that the
writers recommend they be avoided. At the time of this writing, the
procedures for calculating the bedload discharge based on the suspended
load discharge are just as accurate as sampling, If sampling is necessary,
up-to-date information on methods can be obtained by contacting the US
Geological Survey's Research Laboratory, St. Anthony Falls, Minneapolis,
Minnesota, and the Sediment Research Center in Denver, Colorado,
DETERMINING SEDIMENT FLOW RATES
It is one problem to obtain sediment samples from the flow field and
another to determine the discharge of sediment from these samples, Using
the suspended-load measuring devices discussed previously, one can obtain
suspended-load samples at any location and time. However, these samples
may or may not yeild concentrations truly representative of the mean sus-
pended-sediment concentration for, the entire cross section. Ideally, the
best procedure for sampling any stream for sediment concentration, as
related to discharge, would be to collect the entire flow of the stream for
a given period of time. For practical reasons such a method cannot be
employed. Consequently, some system for selecting limited numbers of
samples must be used (54),
The most common methods that have been used to locate the transverse
positions of sampling verticals are:
—single vertical at midstream,
—single vertical at thalweg or point of greatest depth,
—verticals at 1/4, 1/2, and 3/4 width,
—four or more verticals at midpoints of equal width sections, or
equal transit rate (ETR) at all verticals, and
—verticals at centroids of equal discharge increments (EDI) across the
stream.
114
-------
Obviously, the simplest practice is the selection of a single sampling
vertical at midstream or at thalweg.' However, a single vertical should only
be used in a very small stream or in certain types of routine sampling when
there is adequate mixing and accuracy is not important.
Selection of sampling at the 1/4, 1/2, and 3/4 width of a stream cross
section^is convenient and practical. This method provides more information
concerning the distribution and discharge of sediment than the single-vertical
method.
The ETR and the EDI methods are generally adopted in important investi-
gations. These are sufficiently complex that a discussion cannot be included
here. Reference (54) should be consulted for procedures and formulas. In
general, these require a sound knowledge, of stream gauging practice. In
terms of model studies, the ETR and EDI methods would probably only be used
at key points such as the downstream end of the model reach. Such detailed
measurements would be important to obtain accurate data for model calibration.
It is worth pointing out at this time that a knowledge of the water dis-
charge is vital to determining the sediment discharge. The general procedure
is to analyze a sediment sample for concentration of sediment by weight and
compute concentration. This is true whether one sample or several are
collected. In the simplest case, the water discharge may be read from a
rating curve at a gauging station. In the worst case, current meter measure-
ments must be made at the same time the samples are collected.
As noted earlier, the sediment sampling equipment is limited mostly
to the collection of suspended-sediment and bed-material samples from
streams. The suspended-sediment sampler can sample to only within a few
tenths of a foot of the streambed. The sampled part is referred to as
carrying the measured load and the unsampled zone as carrying the unmeasured
load (Figure 30). The unmeasured load contains both unmeasured and suspended-
sediment load and the unmeasured bedload. (Bedload is that material trans-
ported in a stream by sliding, rolling, and bouncing along the bed and
very close to it, that is, within a few grain diameters.) Total load, then,
is the sum of the measured and unmeasured load.
There are some streams with sections so turbulent that nearly all
sediment particles moving through the reach are in suspension. For instance,
the Cuyahoga River reach used for the present study had several reaches where
this was true. Sampling and suspended sediment in such sections with a
standard suspended-sediment sampler then very nearly represents the total
load.
Turbulence flumes or special weirs can be used to bring the total
load into suspension, Total load can usually be rather accurately sampled
where the streambed consists of an erosion resisting material such as bed-
rock or a very cohesive clay. In such situations, the majority, if not all,
the sediment being discharged is in suspension or the bed would contain a
deposit of sand. Most total load sampling is accomplished at the crest of
small weirs, dams, culvert outlets, or other places- where all sediment must
be in suspension.
115
-------
Where such conditions or structures are not present, the -unmeasured
load must be computed by various formulas. The reader concerned in detail
should consult reference (11),
CHARACTERIZING SEDIMENTS
It is beyond the scope of this report to present complete laboratory
procedures for analyzing sediment samples. Instead, the intent is to point
out those characteristics important to modeling and provide a brief descrip-
tion of how they are determined, The single best reference on sediment
laboratory procedures is (49) by Guy and Norman. The serious reader should
obtain a copy from the USGS to obtain full details on laboratory procedures.
Chapter II of (54) also contains much useful information. The discussion
here is edited from these two references.
Definitions
Before proceeding directly to analysis, it is worthwhile to review the
assumption within the model. At the same time, some standard analysis
terminology will be introduced.
The sediment model evaluated in this report assumes that all sediments
are noncohesive, spherical particles, with fixed specific gravities. The
spherical particles are divided into fixed sizes, That is, the 0.63 mm size
class is all assumed to be exactly 0,063 mm in diameter. Obviously, real
sediment particles are neither spherical or fixed in specific gravity. The
analysis problem then is to take a real sediment sample and analyze it in
such a way that information fixed to the model is consistent with the
assumptions. In order to accomplish this, the following standard terminology
is introduced (65):
—The nominal diameter of a particle is the diameter of a sphere
that has the same volume as the particle.
—The sieve diameter of a particle is the diameter of a sphere equal
to the length of the side of a square sieve opening through which
the given particle will just pass.
—The standard fall velocity of a particle is the average rate of fall
that the particle would attain if falling alone in quiescent, dis-
tilled water of infinite extent and at a temperature of 24°C,
—The standard fall dj-ameter, or simply fall diameter, of a particle
is the diameter of a sphere that has a specific gravity of 2.65 and
has the same standard fall velocity as the particle.
—The sedimentation diameter of a particle is the diameter of a sphere
that has the same specific gravity and terminal uniform velocity
as the given particle in the same sedimentation fluid.
116
-------
—The standard,sedimentation diameter of a particle is the diameter
of a-sphere that has the same specific gravity and has the same
standard fall velocity as the given particle.
—Size distribution,, or simply distribution, when applied in relation
to any of the size concepts, refers to distribution of material by.
percentages or proportions by'weight.
—Fall velocity and settling velocity are genrally terms which may
apply to any rate of-fall or settling as distinguished from
standard fall velocity.
By examining the above terminology, it can be seen that the model used in the
study could be operated in two quite different ways. First, all the
particle sizes could be assigned a specific gravity of 2.65. Sediment samples
would then have to be analyzed in such a way that the percent of material
having a certain standard full diameter could be determined. Using this
method, it would not matter how big a particle was or what is actual specific
gravity was but only how rapidly it would settle out of suspension. Hydraulic
analysis techniques which sort the sample by immersion in fluid are required.
The second method is to assign each of the size fractions in the model a
particular specific gravity. Samples must then be sorted by physical size
(nominal diameter) and then by weight within a size fraction. In terms of
fate and effect studies, the latter has an advantage when it is known that
particular pollutants (say heavy metals) attach to specific size fractions.
Sample Analysis
In the earlier discussion of sampling techniques, two different types
of samples were described. The first was a single material sample. These
are taken to characterize the bed and bank sediments. The second was a
suspended sediment sample. These are taken both to characterize the sedi-
ments being transported and to determine the transport rate.
Reference (65) provides two flow charts which describe the sequence
of analysis for the two types of samples. These two flow charts are
reproduced in Figures 35 and 36. Analysis of a bed or bank material sample
is somewhat simpler and will be discussed first. Each box in the flow
charts represents an anlaysis procedure. Boxes divided by horizontal lines
indicate that more than one procedure may be used to accomplish the same
result.
Bed Material Sample Analysis—
Note that the analysis of a bed material sample is generally divided
into three parts. There is a part dealing with coarse, medium, and fine
particles. Coarse, in this case, means particles greater than approximately
2 mm and fine means particles smaller than 0.063 mm. The reason for the
three parts is that various methods of analysis are limited in the range
of particle sizes they can accommodate.
117
-------
BED-MATERIAL SAMPLE
Weigh
(Unused)
(Unused)
' Hand remove
Coarse sieve
1
Split
(Coarse particles)
1
Sieve
Immersion
(Fine particles)
Split
(Medium particles)
(Unused)
Weigh
Figure 35, Flow Chart for Bed and Bank Material Analysis
118
-------
SUSPENDED-SEDIMENT SAMPLE
(Concentration)
Decant excess
supernatant
liquid
| Arrange in chronological order
Weigh water and sediment
Dark cool storage
Settle sediment
(Sand)
Dissolved
?nl!H<-
Otgi:
Evapoidlkm
method r
Wets
Gooch filter
method Drv s
VAt
| Oven dry
Oven
Desiccate
Wei
Weigh
nic
>val
ieve (Dispersed settling)
ieve
ube Organic
removal |
dry ZH
' Split
1 (dist lied)
gh ' -T- '
±_J |
Mechanical
and chemical
dispersing
Dissolved
« 1 solids 1
1 Pipet
1 BW tube
Evaporate
Oven dry
Desiccate
Weigh
(Particle size)
Mechanical
disaggregation
Wet sieve
(native)
(Silt-clay)
Split
(native)
(Sediment weight)
| Settle]
Dissolved
sol ds 1
| Decant [
Oven dry
Desiccate
Weigh
(Native settling
Split
(native)
I Dissolved
1 solids
Mild
mechanical
mixing
Pipet J
BW tube 1 .
Evaporate
IU"
Oven dry
1 Desiccate |
["weigh]
Figure 36. Flow Chart for Analysis of Suspended Sediment
Samples (65)
119
-------
The first step involved in analysis is storage. Samples may be stored
either wet or dry. Wet storage is recommended when large quantities of fine
material are present, particularly clays. Dry storage of samples containing
clay may result in particles cementing together or hardening into large
aggregated lumps. These lumps are difficult to disperse and bias the
analysis.
The second step in a material sample analysis is weighing. This infor-
mation is used later in determining what percent by weight each fraction
represents. With the weight determined, it is then possible to divide the
sample into size fractions and proceed with detailed analysis.
Following weighing the very coarse particles (say 1/2 inch or greater)
are removed by sieving or by hand. These can then be weighed. In terms of
data for modeling, it will probably not be necessary to determine the
specific gravity of the very large particles. These are the least likely to
move and are seldom connected with pollutant transport. For practical pur-
poses, a large size class is established in the model (say 20 to 100 mm) and
all large particles are assigned to this class regardless of exact size or
specific gravity. The exact effect of this type assumption is not known
but doing anything else is practically infeasible.
With the coarse particles removed, the sample may be larger than
necessary. It may be split and the unused portion weighed. Splitting
the sample to obtain a representative smaller amount is particularly important
when using hydraulic methods to determine standard full diameter. These
methods will be discussed shortly.
Before proceeding further, the organic material should be removed from
the sample. Small root hairs, leaf matter, and other organics may bias
results of later tests. Large pieces of organic matter may be removed
manually. Smaller particles are normally oxidized with hydrogen peroxide.
After removal of organics, the medium and' fine fractions are separated.
This is normally done by sieving. The two fractions can then be analyzed
individually for either nominal size and specific gravity or for standard
full diameter.
The analysis of medium size particles may begin with an additional
splitting and organic material removal if required. Three options are
then available for particle classification. Two are sieve methods. If the
model is to be used for fate-type studies as was done in the present report,
the particles are sieved to separate them into suitable size fractions.
The sieving may be done by shaking dry material through or by washing the
material through with streams of water. The latter method is generally
faster and gives more consistent results. It is often assumed that sieve
diameter and nominal diameter are equal. The writers feel this is an
adequate assumption for modeling purposes. The third method of analysis
for medium particles is the visual accumulation tube (VA). A uniformly
mixed sample is released from the top of a long water column. The particles
sort themselves hydraulically and accumulate at the bottom of the column at
120
-------
times proportional to their standard full diameter (assuming the temperature
of the column is appropriate), An optical tracking device on the tube
coupled to a chart recorder allows direct determination of percent finer
versus full diameter curves. Use of the VA tube method requires that the
model be run using uniform specific gravities for all size classes. The
remaining step in the analysis of medium particles is weighing. In the case
of sieve analysis, the portions on each sieve are weighed and the percentage
of the total sample determined. In the case of the VA tube, the entire
sample is weighed as the percent-finer graph used to determine what percent
is what size.
The analysis of the fine particles is, in ways, quite similar to the
analysis of the medium particles. The sample is split, if required, the
the unused portion weighed, and the organic material removed..
Two methods are available for the analysis of the fine particles. These
are the pipet and bottom withdrawal (BW) tube methods. In the .pipet method,
a graduated cylinder filled with distilled water and a suitable volume of
sediment is agitated until the sediment is fully mixed, A special hydro-
meter is inserted and the specific gravity of the mixture as a function of
time recorded. From this information, a standard full diameter versus per-
cent finer curve can be calculated. The bottom withdrawal tube method is
similar in some ways, A full column filled with a uniform mixture of sedi-
ment and water are used. At time zero, the column is fully mixed. Samples
are withdrawn from the bottom at specified intervals, dried, and weighed.
A calculation procedure yields standard full velocity versus percent finer
data.
In terms of model studies, the techniques most likely to be used are
sieve analysis and the VA tube. In order to determine, for example, the
fate of 0.063 mm particles with specific gravities of 4.0,, it would first
be necessary to sieve the sample to determine what percent of the total
was 0.063 mm. Next, all the 0.063 mm particles or a representative sample
would be VA tube analyzed to see what percent had full velocities equivalent
to particles with specific gravities of 4.0.
Suspended Material Sample Analysis—
The various procedures for analyzing a suspended sediment sample are
illustrated in Figure 36. There are two major differences in the analysis
of a suspended material sample. First, one objective is to determine the
concentration of material by weight suspended in the flow. Second, there
are usually no particles larger than sand found in suspension. Thus, much
of the analysis concentrates on the fine material fraction (less than 0.063
mm). The determination of concentration will be discussed first, followed
by a discussion of the particle size diameter analysis.
If the concentration of a suspended material sample is to be determined,
it will be necessary to store both the particulate material and the water
from the samples. The first step in analysis is to determine the weight of
the sediment-water mixture. If long-term storage is required a cool, dry
place is recommended to discourage flocculation of-the clays and other adverse
chemical reactions.
121
-------
The determination of concentration is straighforward. Generally, the
sediment is allowed to settle and the excess liquid decanted. The liquid
may be analyzed for dissolved solids. The weight of sediment may be deter-
mined by either filtration or simple evaporation plus weighing. The concen-
tration is determined from the formula:
/£ weight of sediment x 1,000,000
»m§' weight of water sediment mixture
where f is a factor ranging from 1.0 to 1.34 depending on the weight ratio.
The factor accounts for the fact that part of the water sediment mixture is
occupied by the sediment and is used because weight is easier to determine
accurately than volume.
The particle size analysis for the suspended material begins with
disaggregation if required. This assumes that concentration was determined
and that the drying process has bonded many of ther.particles together due
to the presence of clays and silts. Disaggregation reseparates the particles
if required.
The remainder of the analysis of suspended sediment samples does not
differ a great deal from the analysis of bed material samples. The sand
fraction (greater than 0.063 mm) may be wet or dry sieved or a VA tube
analysis performed. The fine material may be simply dried and weighed
or it may be analyzed in detail using the BW tube or pipet technique.
The pipet technique has not been discussed yet. It is similar to the
B¥ tube method in that analysis starts at time zero with a uniformly mixed
suspension of fine particles. A pipet is used to withdraw samples from a
fixed level in the mixture at predetermined times. The concentration pro-
vides information on percent finer versus full diameter,
Given the current state of modeling, it is likely that simply weighing
the fine fraction to determine its percentage of the total would be adequate.
Sieve methods combined with the VA tube as described for the bed material
sample would meet the remainder of analysis requirements.
TRACING SEDIMENT MOVEMENT
The difficulty involved in tracing the movement of a sediment particle
is obvious. Movement is governed by particle shape, size, specific gravity,
the position relative to other particles, the temperature and properties
of the flow and many more. It is thus not surprising that tracer methods
are not widely used.
The purpose of this section is to provide a very brief look at the
types of studies which have been attempted. The coverage is not exhaustive
but should give the reader an idea of the problems involved and the type
of infomration attainable.
122
-------
Only two readily available references discuss tracer techniques in any
detail. These are references (61) and (63) by Graf and Bogardi, respectively.
Additional information is contained in the publications and work of the US
Geological Survey (References 69, 70, and 71). Much of this work has been
done by Hubbell.
Graf identifies the following desirable properties of tracer materials:
—A labeled and an unlabeled solid particle must react to the forces
responsible for sediment motion in the same way.
—The physical and/or chemical properties of the. traced particle(s)
must be distinguishable and/or detectable with appropriate equip-
ment. Depending on the desired accuracy of the sensitivity of the
measuring device, the amount of tracing material necessary must be
reasonable.
—The tracer on or in the solid particle should be durable, at least
for the time over which the experiments or study extends.
—The tracer should not be hazardous to the biological environment.
—The cost of producing traced particles should be reasonable.
At present, there are three different types of tracers in use: radioactive
tracers, paint and fluorescent tracers, and density tracers (heavy miner-
als) .
Three different methods are available for creating radioactive tracers.
One method is the irradiation method. If inactive isotope exists in a
natural or aritifical solid particle, neutron irradiation in a nuclear
reactor this isotope can be activated and then emits detectable radiation.
Artificial sediment has been made out of glass, having the same density as
quartz. Incorporated in the glass is an inactive isotope. The glass is
ground and sorted, and a size distribution is selected which matches the
sediment to be investigated. Immediately prior to the test, the ground
glass is irradiated. Some natural sediments contain already inactive
isotopes.
Another method is by sorption or coating of radioactive elements on
natural sand. One technique is to add a small amount of radioactive gold
chloride to the wet sediment, Hubbell and Sayve (44) used sand particles
labeled with iridium-192.
Investigators concerned with the movement of larger particles have used
another method. Tracers are inserted into holes which are drilled into the
natural or aritificial particles and then sealed in resin,
Graf (61) recommends that isotopes be used which emit hard gamma
radiation rather than soft beta radiation. Gamma radiation is only partly
absorbed by sand and water, and thus it facilitates an in situ radioactivity
measurement. The half-life, of the tracer should be long enough to conduct
123
-------
the experiment but short enough to minimize interference between successive
experiments. Elements known to be required in the metabolic processes of
organisms should be used with utmost care, if at all.
Paint and synthetic resin finishes may be used as traces, There are,
however, certain disadvantages involved. Although these finishes are
inexpensive, they are limited in their use to the larger fractions of the
sediment (greater than 0.2 mm). Silt and finer matter will begin to stick
together. These finishes are also subject to abrasion.
A fluorescent tracer is material attached to the solid grain made up
of organic dyes which fluoresce. A disadvantage of this tracing method lies
in the difficulty of obtaining and analyzing the samples, which often have
to be viewed in the dark with fluorescent lamps.
Among geologists, it is popular to trace heavy and light minerals, There
are about 30 heavy minerals which are reasonably diagnostic of geologic
source material. Some of the heavy minerals used for this purpose are horn-
blene, biotite, muscovite, augite, and zircon. The study of the distri-
bution of heavy minerals is helpful for the determination of sand movement
along coastlines. Heavy metals and minerals may also be introduced into
the flow. Their behavior, however, is not representative of sand in general
because of the greater density.
Ferromagnetic substances may also be used as tracers. Natural sand
is obtained from the point under study and is mixed with fine-grained ferro-
magnetic material. By spreading the mixture on the bottom, information on
the movement of bed load is obtained from differences in the magnetic
properties of the original and labelled bed load. The principal properties
of the magnetic tracer produced of the natural bed load are identical with
those of the original sand.
Detection and analysis techniques vary depending on the type of tracer.
Radioactive tracers may be followed in situ by suitably encased Geiger
counters or scintillation detectors. For all other tracers, it is generally
necessary to collect bed or suspended material samples. The tagged
particles are located by visual means only (paint or resins) or by viewing
under ultraviolet light (fIncrescents), Heavy mineral tracers may sometimes
be separated out using immersion in dense liquids such as Bromoform.
Special instrumentation similar to that used for radioactive tracing is used
to detect ferromagnetic material.
At this time, tracer studies seem to be relatively limited in their
application to CSO studies. Radioactive tracers are the most highly developed
and the .easiest to detect. At the time of this writing, the environmental
objectives to their use may be such that they are no longer practical. This
leaves fluorescent materials, paints, ferromagnet materials, and heavy
minerals. It appears from earlier portions of this report that most interest
in CSO sediments centers around the midsand sizes (say 1-2 mm) down to the
silt-clay range. Particles coated with fluorescent dyes could be used to
gain insight into the .2 mm to 2 mm range. Paint does not appear to be
124
-------
useful. No current technology appears applicable to smaller particles and
the silt-clay range. Dissolved fluorescent dyes such as Khodamine WT could
provide some insight into the movement of wash load (micron size) particles.
Ferromagnetics and heavy minerals would only be useful if equivalent full
diameter particles could be found for heavy metal pollutants such as lead
or mercury.
The most promising application of tracers to CSO studies would be model
verification. A study in which a large number of particles typical of a CSO
outfall were released into a flow and followed could provide a useful data
set. iuch a study would provide insight into the correctness of model
assumptions, particle transport rates, and patticle dispersion along the
receiving stream.
125
-------
REFERENCES
(1) Colston, N. V., Jr., Characterization and Treatment of Urban Land
Runoff, EPA-670/2-74-096, December 1974, 157 pp.
(2) Field, R., A. N. Tafuri, and H, E. Masters, Urban Runoff Pollution
Control Technology Overview, EPA 600/2-77-047, March 1977, 90 pp.
(3) Pitt, R. and R. Field, 'Water Quality Effect from Urban Runoff,
American Water Works Association J., 69 (8), 1977, pp, 432-436.
(4) Keefer, T. N., R. K. Simons, and R. S. McQuivey, Dissolved Oxygen
Impact from Urban Storm Runoff, Final Repqrt, EPA Contract No.
68-03-2630, March 1979.
(5) Metcalf and Eddy, Inc., Wastewater Engineering* Mc-Graw-Hill
Publishing Company, 1972.
(6) Manning, M. J., R. H. Sullivan, and T. M. Kipp, Nationwide Evaluation
of Combined Sewer Overflows and Urban Stormwater Discharges: Volume
III, Characterization of Discharges, EPA 600/2-77-064C, August 1977.
(7) Ammon, D., EPA, Edison, New Jersey, Personal Communication,
August 21, 1979.
(8) Berwick, R., M. Shapiro, J. Kuhner, D. Luecke, and J, J. Wineman,
Select Topics in Stormwater Management Planning for New Residential
Developments, EPA-Draft Report, Grant No. R-805238.
(9) Amy, G. F., R. Pitt, S. Rameshawar, W, L. Bradford, and M, B. LaBraff,
Water Quality Management Planning for Urban Runoff, EPA 440/9-75-004,
December 1974.
(10) Wanielista, M. P., Stormwater Management: Quantity and Quality.
Ann Arbor Science, 1978
(11) Kluesener, J.- W. and G. F, Lee, Nutrient Loading from a Separate
Storm Sewer in Madison, Wisconsin," Water Pollution Control
Federation J., _4j5 (5), May 1974, pp. 920-936.
(12) Pitt R. and R. Field, Water Quality Effect from Urban Runoff, American
Water Works Association J., 69 (8), 1977, p. 432-436.
126
-------
(13)
(14)
(15)
(16)
(17)
(18)
(19)
(20)
(21)
(22)
(23)
(24^
(25)
(26)
(27)
Shaheen, D. G., Contributions of Urban Roadway Usage to Water Pollution,
EPA-600/2-75-004, April 1975.
Sartor, J. B. and G, B, Boyd, Water Pollution Aspects of Street Surface
Contaminants, EPA-R2-72-081, November 1972.
Huber, W. C. , J. P. Heaney, K. J. Smolenyak, a^ ^. A. Aggidis, Urban
Rainfall-Runoff-Quality Data Base: Update with Statistical Analysis,
EPA Draft Report, Grant No. 68-03-0496.
Randall, C. W. T, J, Grizzard, and R. C, Hoehn, Runoff Sediment Loads
from Urban Areas, In: Proceedings of the International Symposium on
Urban Hydrology, Hydraulics, and Sediment Control, University of
Kentucky, July 18-21, 1977, pp. 317-323.
Dalrymple, R. J. , S. L, Hood, and D. C. Marin, Physical and Settling
Characteristics of Particulates in Storm and Sanitary Wastewaters,
EPA 670/2-75-011, April 1975,
American Public Works Association, Water Pollution Aspects of Urban
Runoff, EPA 11030 DNS01/69, January 1969.
Krantz, J. and D. L, Russell, Lancaster Silo Project: Particle Sizing
and Density Study, Meridian Engineers, Philadelphia, Pennsylvania,
January 1973.
Envirogenics Company, In-Sewer Fixed Screening of Combined Sewer
Overflows, EPA 11-24, FKJ 10/70, 1970.,
Chemical Rubber Company, Handbook of Chemistry and Physics, 55th
Edition, CRC Press, 1974.
Fair, G. M. , J. C. Geyer, and D. A, Okum, Water and Wastewater
Engineering, Volume II. John Wiley and Sons, New York, New York,
1966.
Rich, L- G. , Unit Operations of Sanitary Engineering, Rich Publishing,
Clemson, South Carolina, 1974.
Waller, D.. H. , Pollution Attributable to Surface Runoff and Overflows
from Combined Sewer Systems, Atlantic Industrial Research Institute,
Halifax, Nova Scotia, April 1971.
Burgess and Niple, Ltd., Stream Pollution and Abatement from Combined
Sewer Overflows, Bucyrus, Ohio, EPA 11024, FKN 11/69, November 1969.
Davis, P. L. and F. Borchardt, Combined Sewer Overflow Abatement Plant,
Des Moines, Iowa, EPA Report Number, EPA-R2-73-170, April 1974.
Sayre, W. W. , H. P. Guy, and A. R. Chamberlain, Uptake and Transport
of Radionuclides by Stream Sediments, US Geological Survey Professional
Paper 433-A, 1963.
127
-------
(28) Frobisher, M., R. D. Hindsdill, T. Koby, Crabtree, and C. R. Goodheart,
Fundamentals of Microbiology. W. B. Saunders Publishing Company,
1974.
(29)
(30)
(31)
(32)
(33)
(34)
(35)
(37)
Chen, C. W. and G. T. Orlob, Ecologic Simulation for Aquatic Environ-
ments. Prepared for the Office of Water Resources Research, US
Department of Interior, December 1972, PB 218 828.
Klots, E. B., The New Fieldbook of Freshwater Life,
Company, New York, New York, 1966.
Putman Publishing
Mackenthun,•J. M., Biological Problems in Freshwater Environments—
Their Identification, Investigation, and Control, US Department of
the Interior, Federal Water Pollution Control Admininstration.
Jones, R. A. and G. F. Lee, Evaluation of the Elutriate Test as a
Method of Predicting Contaminant Release During Open Water Disposal
of Dredged Sediment and Environmental Impact of Open Water Dredged
Material Disposal, Volume I: Discussion, Technical Report D-78-45,
US Army Corps of Engineers, WES, Vicksburg, MS, 1978,
Lee, G. F., R. A. Jones,. F. Y. Saleh, G. M. Mariani, D, H. Homer,
J. S. Bulter, and T; Bandyopadhyay, Evaluation of the Elutriate Test
as a Method of Predicting Contaminant Release During Open Water
Disposal of Dredged Sediment and Environmental Impact of Open
Water Dredged Materials Disposal, Volume II: Data Report, Technical
Report D-78-45, US Army Corps of Engineers WES, Vicksburg, MA,
1978, 1186 pp.
US Environmental Protection Agency, Quality Criteria for Water,
EPA-440/9-76-023, US Government Printing Office, Washington, D.C.
1976, 256 pp.
Plumb, R. H. and G. F. Lee, Response of Lake Superior Algae to
Nutrients in Taconite Tailings, Water Pollution Control Federation
J., 47, 1975, pp. 619-625.
Peddicord, R. K., V. A. McFarland, D. P. Belfioni, and T, E. Byrd,
Effects of Suspended Solids on San Francisco Bay Organisms, Physical
Impact Study, Appendix G, Dredged Sediment Disposal Study, San
Francisco Bay and Estuary, US Army Corps of Engineers, San Francisco,
July 1975.
(38) Neff, J. W. , R. S. Foster, and J. F. Slowey, Availability of Phosphorus
to Phytoplankton and its Implications for Phosphorus Management
Strategies. To be published In: Proceedings of IJC-Cornell University
Conference on Phosphorus Management Strategies for the Great Lakes,
April 1979.
128
-------
(39) Lee, G. F., R, A. Jones, and G. M. Mariani, Comments on US EPA
Corps of Engineers Dredged Sediment Bioassay Procedures, Colorado
State University, Environmental Engineering Program, Occasional Paper
No. 26, December 1977.
(40) Mariani, G. M., G. F. Lee, and R. A. Jones, Significance of Sediment-
Associated Contaminants. To be presented at: American Water Resources
Association National Meeting, Las Vegas, NV, September 1979.
(41) Lee, G. F., Phosphorus, Water Quality, and Eutrophication in Lake
Michigan. In: Proceedings of the Fourth Session, 1972 Conference
on Pollution of Lake Michigan and its Tributary Basin, Illinois,
Michigan, and Wisconsin, US EPA Region V, Chicago, Illinois 1974.
(42) Lee, G. F., R. A. JOnes, and W. Rast, Availability of Phosphorus
to Phytoplankton and its Implications for Phosphorus Management
Strategies, To be published In: Proceedings of IJC-Cornell University
Conference on Phosphorus Management Strategies for the Great Lakes,
April 1979.
(43) Rast, W. and G, F. Lee, Summary Analysis of the North American (US
Portion) OECD Eutrophication Project: Nutrient Loading-Lake Response
Relationships and Trophic State Indices. Report EPA-Corvallis,
EPA 600/3-78-008, 1978.
(44) Lee, G. F., W. Rast, and R. A. Jones, Use of the OECD Eutrophication
Modeling Approach for Assessing Great Lakes" Water Quality, Colorado
State University, Environmental Engineering Program, Fort Collins,
Colorado, Occasional Paper No. 42, 1979.
(45) Lee, G. F. and R. A, Jones, Water Quality Characteristics of the US
Waters of Lake Ontario during the IFYGL and Modeling Containinat Load-
Water Quality Response Relationships in the Nearshore Waters of the
Great Lakes, Report to National Oceanic and Atmospheric Administration
Great Lakes Research Laboratory, Ann Arbor, June 1979.
(46) Jones, R. A., W. Rast, and G. F. Lee, Relationship Between Mean and
Maximum Chlorophyll ja Concentrations in Lakes, Environmental Science
and Technology, 13, 1979, pp. 869-870,
(47) Lee, G. F., W. Raset, and R. A. Jones, Eutrophication of Waterbodies:
Insights for an Age-Old Problem, Environmental Science and Technology,
12, 1978 pp. 900-908.
(48) Lee, G. F., Role of Hydrous Metal Oxides in the Transport of Heavy
Metals in the Environment, In: Proceedings of Symposium of
Transport of Heavy Metlas in the Environment, In: Progress in Water
Technology, 17, 1975, pp. 137-147.
129
-------
(49) Simons, D. B., R. M. Li, and M. A. Stevens, Development of Models for
Predicting Water and Sediment Routing and Yield from Storms on Small
Watersheds, USDA Forest Service, Rocky Mountain Forest and Range
Experiment Station, Flagstaff, Arizona, August 1975.
(50) Keefer, T. N. and H. E. Jobson, River Transport Modeling for Unsteady
Flows, Hyd. Div, J,, ASCE, Volume 104, No, HY5, Proceedings Paper
No. 13735, May 1978.
(51) Amein, M. M. and C. S. Fang, Implicit Flood Routing in Natural
Channels, Hyd. Div. J., ASCE, Volume 96, No. HY12, Proceedings Paper
7773, December 1970.
V
(52) Shields, A., Anwendung der Aehnlichkeitsmechanik und Turbulenz forschung
auf die Geschiebewegung, Mitteilung Preussischen Versuchanstalt Wasser,
Erd, Schiffbau, Berlin, No. 26, 1936 (in German).
(53) Meyer-Peter, E. and R. Muller, Formulas for Bed-Load Transport,
Proceedings, III Cong. IAHR, Stockholm, 1948, pp. 39-64*
(54) Vanoni, V A., (editor), Sedimentation Engineering. Preparedly the
ASCE Tack Committee for the Preparation of the Manual on Sedimentation
of the Sedimentation Committee of the Hydraulics Division, 1975.
(55) Simons, D. B.;and F. Senturk, Sediment Transport Technology, Water
Resources Publications, Fort Collins, Colorado, 1977.
(56) US Bureau of Reclamation, Investigation of Meyer-Peter, Muller Bed-
load Formulas, Sedimentation Section, Hydrology Branch, Division of
Project Investigations, US Department of the Interior, Buruea of
Reclamation, June 1960.
(57) Shulits, S. and R, D. Hill, Jr., Bed Load Formulas, Part A, A
Selection of Bedload Formulas: Part B, Program Listing for Bedload
Formulas, NTIS, PB-194, 950 Pennsylvania State University, December
1968.
(58) Einstein, H. A., The Bedload Function for Sediment Transportation in
Open Channel Flows, US Department of Agriculture, Soil Conservation
Service, T. B. No. 1026, 1950.
(59) Colby, B. R. and C. H. Hembree, Computation of Total Sediment Discharge,
Niobrara River near Cody Nebraska, US Geological Survey, Water Supply
Paper No. 1357, 1955.
(60) US Army Corps of Enginers, Buffalo District; Flood Plain Infomration,
Cuyahoga River, Akron to Summit-Portage County Line, Ohio, May
1971.
130
-------
(61) Graf, Walter H., Hydraulics of Sediment Transport, McGraw-Hill
Publishing Company, New York, NewyYork, 1971, 513 pp.
(62) Simons, Daryl B, and F» Sentiirk, Sediment Transport Technology,
Water Resources Publications, Fort Collins, Colorado, 1977 807 pp.
(63) Bogardi, J,, Sediment Transport in Alluvial Streams,, Publishing House
of the Hungarian Academy of Sciences, Budapest, 1974> 826 pp.
(64) Guy, H. P. ,and V. Norman, Field Methods for Measurement of Fluvial
Sediments, USGS Techniques of Water Resources Investigations, Book
3, Chapter C2.
(65) Guy, H. P., Laboratory Theory and Methods for Sediment Analysis,
USGS Techniques of Water Resources Investigations, Book 5, Chapter
Cl.
(66) Ritter, J. R. and E, J. Helley, Optical Method for Determining
Particle Sizes of Coarse Sediments, USGS Techniques of Water Resources
Investigations, Book 5, Chapter C3,
(67) Federal Inter-Agency Sedimentation Project„ Instruments and Reports for
Fluvial Sediment Investigation, Inter-Agency Report Catolog, St.
Anthony Falls Hydraulics Laboratory, Minneapolis, MN, 1966, 67 pp*
(68) Prychj E. A. and D, W. Hubbell, A Sample for Coring Sediments in
Rivers and Estuaries, Geological Surwey of American Bulletin,
Volume 77, 1966 4 pp. 549-556.
(69) Rathbun, R. E,, V, C. Kennedy, and J, K, Culbertsorij Transport and
Dispersion of Fluorescent Tracer Particles for the Flat Bed
Condition, Rio Grande Conveyance Channel, Near Bernardo, New Mexico,
US Geological Survey Professional Paper No. 5621, USGPO, 1971.
(70) Hubbell, D. W. and W* W, Sayre, Application of Radioactive Tracers in
the Study of Sediment Motion. In: Proceedings of Fed. Inter-Agency
Sedimentation Conference, US Department of Agriculture, Misc. Pub.
970.
(71) Hubbell, D* W. and W. W. Sayre, Sand Transport Studies with Radioactive
Tracers. In: Proceedings of American Soceity of Civil Engineers,
' Volume 90, No* HY3.
131
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/2-80-126
3. RECIPIENT'S ACCESSION-NO.
4. TITLE AND SUBTITLE
MOVEMENT AND EFFECTS OF COMBINED SEWER OVERFLOW
SEDIMENTS IN RECEIVING WATERS
5. REPORT DATE
August 1980 (Issuing Date)
6. PERFORMING ORGANIZATION CODE
"Stanley L. Klemetson, Thomas N. Keefer and
Robert K. Simons
8. PERFORMING ORGANIZATION REPORT NO.
;9. PERFORMING ORGANIZATION NAME AND ADDRESS
Colorado State University
Engineering Research Ctr.
Ft. Collins, Colorado 80523
10. PROGRAM ELEMENT NO.
35B1C
11. CONTRACT/GRANT NO.
R-806111
12. SPONSORING AGENCY NAME AND ADDRESS
13. TYPE OF REPORT AND PERIOD COVERED
Municipal Environmental Research Laboratory—-Gin., OH
Office of Research & Development
U.S. Environmental Protection Agency
Cincinnati, OH 45268
FINAL 8/78 to 11/79
14. SPONSORING AGENCY CODE
EPA/600/14
15. SUPPLEMENTARY NOTES
Project Officer: John N. English Phone: (513) 684-7613
16. ABSTRACT
The research work described here was a joint effort of Colorado State University (CSU)
and the Sutron Corporation. The study had two primary objectives. The first objec-
tive was to determine from available literature the characteristics of combined sewer
overflow (CSO) sediments and the factors affecting their transport properties. The
second objective was to make use of the information on characteristics to evaluate a
current sediment model capable of predicting the fate of CSO sediments.
CSU conducted the literature search and evaluation necessary to meet the first objec-
tive. The Sutron Corporation selected a test study site on the Cuyahoga River between
Akron and Cleveland, Ohio; collected limited field data; and used the characteristics
of CSO sediments found by CSU to evaluate a sediment model. Sutron also conducted a
literature search of sediment sampling and tracing techniques necessary for model
application.
It was concluded from the model experiments that qualitative evaluation can be made
concerning the fate of CSO solids which are primarily noncohesive sands. Semi-quan-
titative evaluations could be made if proper data from a particular CSO of interest
could be obtained. Particularly important to the model are the size distribution and
settling velocity characteristics of the CSO sediments.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS C. COSATI Field/Group
Combined Sewers
**Sediments
**Mathematical Models
Water Quality
Cuyahoga River
Sediment Transport
Sediment Tracing
Sediment Sampling
13B
13. DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
148
20. SECURITY CLASS (This page)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
132
t, U.S. GOVEnNMENT PRINTING OFFICE: 1980-657-165/0140
------- |