cxEPA
United States
Environmental Protection
Agency
Municipal Environmental Research
Laboratory
Cincinnati OH 45268
EPA-600/2 79-155
December 1979
Research and Development
Verification of
the Water Quality
Impacts of Combined
Sewer Overflow
-------
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-600/2-79-155
December 1979
VERIFICATION OF THE WATER QUALITY
IMPACTS OF COMBINED SEWER OVERFLOW
by
Thomas L. Meinholz
William A. Kreutzberger
Martin E. Harper
Kev i n J. Fay
Rexnord Inc., Environmental Research Center
Milwaukee, Wisconsin 53201
Grant No. R-80^518
Project Officer
John N. Eng1ish
Wastewater Research Division
Municipal Environmental Research Laboratory
Cincinnati, Ohio ^5268
MUNICIPAL ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO ^5268
-------
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.
i i
-------
FOREWORD
The Environmental Protection Agency was created because of increasing public
and government concern about the dangers of pollution to the health and wel-
fare 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
management of wastewater and solid and hazardous waste pollutant discharges
from municipal and community sources, for the preservation and 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 communication link between the
researcher and the user community.
This report describes water quality impacts associated with wet weather
discharges into the Milwaukee River and details the contribution of combined
sewer overflows to this impact. Through this project data are being obtained
to determine in a rational way the degree of national wet weather pollution
control required.
Francis T. Mayo, Director
Municipal Environmental
Research Laboratory
-------
ABSTRACT
The purpose of this study was to identify the source and mechanism of the
water quality impacts in the Milwaukee River following wet weather discharges.
The results of intensive field surveys were used to identify the river's
response in terms of dissolved oxygen (DO) and fecal coliform concentrations
to different rainfall events under a range of river flow conditions. River
dye studies indicate that velocities are extremely slow even during high
flow conditions in the lower portions of the Milwaukee River which are under
the influence of Lake Michigan. As a result of these slow velocities,
sediments accumulate in the lower river. Laboratory and field investigations
indicate that these bottom sediments are a significant sink for DO. Studies
of the in situ sediment oxygen demand (SOD) rates using respirometers and
chemical analyses of sediment core samples were utilized to develop
sediments maps illustrating the magnitude and distribution of these parameters.
SOD rates ranged from 1.8 to 6.7 gm 02/m2-day in the lower river. Bench scale
tests conducted with bottom sediments show that the oxygen demand of disturbed
or agitated sediments exceeded 1000 gm 02/m2-day at some locations. This is more
than 100 times greater than the SOD rates measured in situ.
Continuous monitoring of DO and temperature at several locations in the lower
Milwaukee River and the results of the intensive monitoring surveys have demonstrated
that there is often a rapid decline in DO following combined sewer overflow (CSO)
events. Water quality modeling of the river with Harper's water quality model
and modeling results of other investigators indicate that the loadings from combined
and storm sewer discharges are not sufficient to cause the observed rapid declines
in DO. The mechanism of this rapid decline is the scouring of sediment oxygen
demanding materials by submerged CSO outfalls. This was determined through
measurements of instream velocities near the bottom sediments resulting from
discharges from submerged outfalls and settling tests of sediments which indicate
the velocity required to scour sediments. Empirical equations were developed
using multiple regression analysis to predict the impact of sediment scouring
by CSO discharges on DO levels. An expression was also added to Harper's
water quality model to provide for time varying SOD rates which are required
to simulate the high oxygen demand of scoured sediments. This model was calibrated
and verified for dry and wet weather conditions in the study area of the river and
was used to determine the DO and fecal colfform impact which is attributable to
CSO.
This report was submitted in partial fulfillment of research grant no. EPA
R804518 by the Metropolitan Sewerage District of the County of Milwaukee under
the sponsorship of the U.S. Environmental Protection Agency. This report covers
the period of August 15, 1976 to December 31, 1978 and was completed December 31,
1978.
iv
-------
CONTENTS
Foreword [[[ i i i
Abstract [[[ i v
Figures [[[ vi
Tables [[[ xi
Acknowl edgment [[[ xi ii
I . Concl us ions [[[ I
2 . Recommendat ions .................................................. 2
3. I nt reduction [[[ 3
Project study area ............................................. *»
Project background ............................................. 16
Report organ izat ion ............................................ 17
*». Field Investigations ............................................. 1 3
Hydraul ic studies .............................................. 20
In tens ive surveys .............................................. 28
Continuous DO and temperature monitoring ....................... ^8
Sediment studies ............................................... 55
5. Analysis and Discussion .......................................... 70
Hydraul ic studies .............................................. 70
Sediment i nvestigation resul ts ................................. 71
Prediction of the DO impact .................................... 7^
Contribution of CSO ............................................ 77
Instream monitoring resul ts .................................... 79
6 . Mode ling [[[ 8l
STORM model [[[ 8l
-------
FIGURES
Number
page
I Illustration of the CSO drainage areas contributing to the
various receiving waters in Milwaukee, Wisconsin ................. 5
2 Monitoring locations in the study area of the Milwaukee
R i ve r [[[ 6
3 Photograph of the North Avenue Dam on the Milwaukee River ........ 7
A Mean monthly flows for the Milwaukee River at the Estabrook
Park USGS flow gauging station ................................... 8
5 Photograph and cross section of the Milwaukee River at St.
Pau 1 Avenue [[[ 9
6 Photograph and cross section of the Milwaukee River at
Walnut Street [[[ 10
7 Photograph and cross section of the Milwaukee River at
North Avenue [[[ I I
8 Photograph and cross section of the Milwaukee River at
Port Washington Road ............................................ 12
9 Illustration of the distribution of CSO outfalls and contribu-
tion from the CSO drainage area along the Milwaukee River ....... lA
10 Illustration of CSO and storm sewer areas tributary to the
Mi Iwaukee River ................................................. 15
II Illustration of dry weather (July 29-31, 1977) and wet
weather (August 3~5, 1977) DO levels at St. Paul Avenue in
the Mi Iwaukee River ............................................. 19
12 Illustration of dry weather (September 22-2**, 1976) and wet
weather (August k-6 , 1977) fecal coliforms levels at St. Paul
-------
FIGURES (continued)
Number Page
\k Monitored dye concentrations at Walnut Street and St. Paul
Avenue during the June 13, 1978 Milwaukee River dye study ......... 23
15 Monitored DO levels in the Milwaukee River during dry weather,
September 21-23, 1976 (Survey l) .................................. 31
16 Monitored fecal coliform levels in the Milwaukee River during
dry weather, September 2I-2A, 1976 (Survey I) ..................... 32
17 Monitored DO levels in the Milwaukee River during wet weather,
May 3 1 -June 2, 1977 (Survey 2) .................................... 3^
18 Monitored fecal coliform levels in the Milwaukee River during
wet weather, May 3 1 -June 2, 1977 (Survey 2) ....................... 35
19 Monitored DO levels in the Milwaukee River during wet weather,
June 17-20, 1977 (Survey 3) ....................................... 37
20 Monitored fecal coliform levels in the Milwaukee River during
wet weather, June 18-20, 1977 (Survey 3) .......................... 38
21 Monitored DO levels in the Milwaukee River during wet weather,
August 3-8, 1977 (Survey k) ....................................... 4|
22 Monitored DO versus distance in the Milwaukee River during wet
weather, August 3~6, 1977 (Survey k)
23 Monitored fecal coliform levels in the Milwaukee River during
wet weather, August 3-7, 1978
2k Monitored DO levels in the Milwaukee River during wet weather,
June 15-18, 1978 (Survey 5)
25 Monitored fecal coliform levels in the Milwaukee River during
wet weather, June 16-18, 1978 (Survey 5)
26 Monitored DO levels in the Milwaukee River during wet weather,
July 25-29, 1978 (Survey 6)
27 Monitored fecal coliform levels in the Milwaukee River during
wet weather, July 26-29, 1978 (Survey 6) .......................... 50
28 St. Paul Avenue continuous DO monitoring results for August 2-7,
1977 [[[ 51
29 St. Paul Avenue continuous DO monitoring results for July 26 to
-------
FIGURES (continued)
Number
30 Ammonia-nitrogen (mg/kg) values in the sediments of the
Mi Iwaukee River 57
31 Chemical oxygen demand (mg/kg) in the sediments of the
Mi Iwaukee River 58
32 Lead (mg/kg) values in the sediments of the Milwaukee River 59
33 Photograph of sediment oxygen demand respirometer for measuring
in situ SOD rates 66
3k Sediment sampling locations for bench scale sediment oxygen
demand determinations 68
35 The observed decline in DO, instream velocity from a submerged
CSO outfall, and rainfall volume at St. Paul Avenue in the lower
Milwaukee River on August 3~k, 1977 73
36 Methodology utilized to obtain ADO and duration values from
continuous DO records at St. Paul Avenue 75
37 CSO, storm sewer and river monitoring locations on the
Milwaukee River, Lincoln Creek, and Kinnickinnic River 83
38 Illustration of assumptions utilized in Harper's water quality
model for simulation of the Lake Michigan inflow 90
39 Influence of temperature on the maximum growth rate of
phytoplankton (U ) and phytoplankton endogenous respiration
rate (31) T?? 93
AO Characteristics of the time varying sediment oxygen demand 97
Al Example of temperature and dissolved oxygen simulations with
Harper's water quality model l°2
k2 Time step sensitivity analysis for St. Paul Avenue, September
21-23, 1976 (Survey I) lo6
^3 Dry weather verification results for St. Paul Avenue, September
21 -23, !976(Survey I) l07
M Dry weather verification results for Walnut Street, September
21-23, 1976 (Survey I) '08
kS Temperature verification results for North Avenue, September
21-23, 1976 (Survey I) 110
viii
-------
FIGURES (continued)
Number
k6 Temperature verification results for Capitol Drive,
September 21 -23, 1976 (Survey I) ill
hi Sensitivity analysis for the dry weather SOD at St. Paul Ave 113
48 Sensitivity analysis for the base extinction coefficient at
St. Paul Avenue. I 14
49 Wet weather calibration results for Wells Street, May 31-
June 3, 1977 (Survey 2) 118
50 Wet weather calibration results for Walnut Street, May 31-June
June 3, 1977 (Survey 2) 119
51 Wet weather calibration results for St. Paul Avenue, June 18-
21 , 1977 (Survey 3) 120
52 Wet weather calibration results for Walnut Street, June 18-21,
1977 (Survey 3) 121
53 Wet weather verification results for St. Paul Avenue, August 4-8,
1977 (Survey 4) 124
$k Wet weather verification results for Wells Street, August 4-8,
1977 (Survey 4) 125
55 Wet weather verification results for St. Paul Avenue, June 16-18,
1978 (Survey 5) - 126
56 Wet weather verification results for Walnut Street, June 16-18,
1978 (Survey 5) 127
57 Wet weather verification results for North Avenue, June 16-18,
1978 (Survey 5) 128
58 Wet weather verification results for St. Paul Avenue, July 26-
29, 1978 (Survey 6) 129
59 Wet weather verification results for Walnut Street, July 26-
29, 1978 (Survey 6) 130
60 Wet weather fecal coliform calibration results, June 18-20,
1977 (Survey 3) '32
61 Wet weather fecal coliform calibration results, August 4-7,
1977 (Survey 4) 133
62 Wet weather fecal coliform calibration results, August 4-7,
1977 (Survey 4) 134
-------
FIGURES (continued)
Number
63 Typical model output for determining frequency and magnitude
of dissolved oxygen violations 137
6A Sensitivity of dissolved oxygen results in CSO loads 138
65 Comparison of the instream model results for with and without
the time varying SOD (scour) for St. Paul Avenue, August 3~8,
1977 (Survey k) I kQ
66 Comparison of instream model results for with and without the
time varying SOD (scour) for St. Paul Avenue, July 16-18, 1978
(Survey 5) l*»l
67 Comparison of the instream model results for with and without
the time varying SOD (scour) for Walnut Street, May 3l~June 3,
1977 (Survey 2) I A3
68 Comparison of instream model results for with and without the
time varying SOD (scour) for Walnut Street, June 16-18, 1978
(Survey 5) IM
-------
TABLES
Number Page
I Milwaukee River Watershed Characteristics ....................... 7
2 Milwaukee River Flow Characteristics Determined By Dye
Tracings [[[ 2 A
3 Summary of Rainfall and River Flow Conditions For Intensive
Mon i to r i ng Su rveys ............................................. 30
A Rainfall Distribution in CSO Area During the Wet Weather
Intensive Survey on August 3-8, 1977 (Survey A) ............... AO
5 Variations in River Flow at the North Avenue Dam and
Estabrook Park During the Intensive Monitoring Survey on
August 3-8, 1977 (Survey 4) .................. ".
6 Summary of Variability of Chemical Parameters During the
Summer of 1977 in Different Reaches of the Milwaukee River ..... 60
7 Moisture Characteristics of Milwaukee River Sediments .......... 6l
8 Summary of Mass Balance Calculations for Milwaukee River
Sediments Prior to and Following Centri fugat ion ................ 62
9 Settling Characteristics of Milwaukee River Sediments From
Wisconsin Avenue Suspended in River Water, January, 1977 ....... 63
10 Settling Characteristics of Sediments from Junction of
Milwaukee and Memononee Rivers Suspended in River Water,
February , 1 978 ................................................ 63
II Variation of In Situ Sediment Oxygen Demand Rates in the
Mi Iwaukee River ................................................ 66
12 Bench Scale Determinations of Sediment Oxygen Demand Under
Undisturbed and Disturbed Conditions .......................... 69
13 Results of the Forward (Stepwise) Regression Analysis with
ADO As the Dependent Variable .................................. 76
I k Results of the Forward (Stepwise) Regression Analysis with T
-------
TABLES (continued)
Number
1 r
1 P
16
17
18
19
20
Pet- imat^H Annual Pn 1 1 1 it ant Load inas for the 1977 Water Year . . . ,
Estimated Loadings to the Sediments in the Milwaukee River
ItuHv Area for the 1977 Water Year
Comparison of Total Flow Predicted by the SWMM and STORM
Model s for the September 24 1 968 Storm
STORM Calibration Results for Suspended Solids and BOD
Compos i te Concent rat ions for CSO Discharges
STORM Calibration Results for Fecal Coliforms for CSO
D i s charges
Literature Values for Combined Wastewater and Separated
STORM Flow Discrete Oualitv
Page
78
....78
84
. . 86
86
....87
21 Concentrations for Monitored STORM Sewer Samples,
22 Range of Input Parameter Values Utilized for Dry Weather
Calibration of Harper's Model .................................... 115
23 CSO and Separate STORM Sewer Loadings to the Milwaukee River
for the Intensive Monitoring Surveys Predicted With STORM ........ 117
24 Sensitivity Analysis of Lumping of CSO Outfalls Into Point
Sources in Receiving Water Model ................................. 117
25 Calibration Values for Harper's Model Input Parameters for the
Mi Iwaukee River .................................................. 123
26 The Dissolved Oxygen Impact of CSO Loads Variations Using An
Extreme Runoff Year .............................................. '37
27 Results for Instream Flow Conditions Using an Extreme Runoff
Year [[[ 139
28 Days of Fecal Coliform Violations for Variations in CSO Load
and River Flow Conditions Using an Extreme Runoff Year ........... 145
-------
ACKNOWLEDGMENT
The completion of this report required input from many individuals, and the
authors gratefully acknowledge their help. The following personnel are
dully recogni zed:
Dr. Nicholas P. Kobriger
Mr. Richard Race
Mr. David Gruber
Mr. Richard Wullschleger
Mr. Joseph Kuderski
Statistical Analysis
Model Development
Field Mon i tor ing
Manager-Analytical Laboratory
Drafting
Thanks are extended to Mr. John English, Project Officer, for direction
and guidance throughout the course of this project, and also to Mr. Richard
Field of EPA who was involved in the initial discussions on the project
objectives. Mr. James Ibach of the Milwaukee Metropolitan Sewerage District
is also thanked for helpful suggestions and support.
XIII
-------
SECTION I
CONCLUSIONS
I. Sediment oxygen demand (SOD) is the major source of the dissolved oxygen
(DO) impacts in the lower Milwaukee River and has its greatest effect
during low river flow conditions.
2. The mechanism of the DO depletion observed in the Milwaukee River during
wet weather is the scouring and resuspension of the bottom sediments
by submerged combined sewer overflow outfalls.
3- The oxygen demand of disturbed sediments can be greater than 1000 gm/m2-
day which is more than 100 times greater than the demand of undisturbed
sediments.
A. Combined sewer overflows (CSO) contribute approximately 40 to 50 percent
of the annual loadings of oxygen demanding materials (in terms of
carbonaceous BOD) and suspended solids to the sediments in the lower
MiIwaukee River.
5. The relationship of rainfall, runoff and prestorm history to the
variable sediment oxygen demand can be predicted through the use of
a statistical procedure.
6. The major source of fecal coliforms in the Milwaukee River is combined
sewer overflow and removal of these overflows will nearly eliminate
fecal conform standards violations.
-------
SECTION 2
RECOMMENDATIONS
I. The significance of sediment oxygen demand as a sink for dissolved
oxygen in receiving waters should be investigated in other areas of the
United States.
2. Investigate the feasibility of the periodic implementation of a dredging
program in the Milwaukee River as a means of alleviating dissolved
oxygen problems.
-------
SECTION 3
INTRODUCTION
The presence of combined sewer systems in the United States and Canada has
been extensively documented in numerous literature sources (1,2). The
effects of discharges from these systems on the receiving waters has become
increasingly important to municipalities and regulating agencies because of
the directives of PL 92-500, the Federal Water Pollution Control Act
Amendments of 1972, which states:
"Wherever attainable, an interim goal of water quality which
provides for the protection and propagation of fish, shell-
fish, and wildlife and provides for recreation in and on the
water shall be achieved by July I, 1983". (Section lOla)
Significant amounts of data have been generated in studies of the quantity
and quality of combined sewer overflow (CSO) from locations throughout North
America (3,4). In addition, demonstration studies at numerous locations have
been completed which provide data on treatment or other control measures for
reducing the discharge loads from CSO areas (5~8) . More recently, comprehen-
sive projects have been proposed, or have already initiated major expenditures
of funds for the abatement of CSO. Typical of major projects are the
following cities efforts:
Chicago, IL - $1.8 billion program for CSO and related flood control (9).
Rochester, NY - $0.4 billion to abate CSO (9).
Milwaukee, Wl - $1.5 billion for the control of CSO and upgrading of
treatment facilities (10).
Each of these projects and countless more in other metropolitan areas will
be required to identify in various levels of detail, the improvements in
water quality as a function of dollars for CSO control (EPA PRM 75-3M. In
order to develop examples of the magnitude of the water quality impacts
associated with CSO, EPA has conducted or is in the process of conducting
site specific water quality studies at a few locations. This study,
"Verification of Water Quality Impact from CSO" is one of these projects
whose objectives are two-fold. The first objective was to use real-time data
collected through monitoring surveys to characterize the water quality
impacts associated with wet weather discharges. The monitoring surveys
included continuous data collection, intensive grab sample surveys lasting
3 or k days and special studies to establish the mechanism of the DO impacts.
The second objective was to decipher the contribution of CSO to this impact
so that a general relationship can be generated for the study area.
3
-------
PROJECT STUDY AREA
The Milwaukee area was selected as one of the sites for the impact evaluation
project because of its large combined sewer area, the availability of
previous and ongoing projects relating to CSO, and the significant water
quality impacts which occur as a result of wet weather discharges. A map
of the three rivers that enter Lake Michigan at the heart of the metropolitan
Milwaukee area is presented in Figure I. Each of the rivers has an associated
combined and storm sewer area which discharges during wet weather. For
purposes of this study, the investigations into the water quality impacts
were limited to the Milwaukee River for the following reasons:
I. It is the largest of the three rivers and it contains the largest
contribution of CSO.
2. It has a definite physical boundary which separates the lower reach
of the river into sections influenced and not influenced by
Lake Michigan.
3. Previous projects related to the CSO problem have been undertaken
on this river.
4. The other two rivers have similar downstream impact characteristics
which would provide no unique data to the analysis.
Mi 1waukee River
The drainage area characteristics of the Milwaukee River watershed are listed
in Table I. Figure 2 presents the location of the river monitoring points
and other important features of the Milwaukee River study area. The most
dominant feature of the river is the inflow from Lake Michigan to the lower
reaches of the river which influences flow conditions as far upstream as
the North Avenue Dam (Figure 3). The inflow of lake water results in the
slowing or reversal of flow in these reaches throughout the year. The lake
influence has been monitored at a few downstream locations with a sensitive
current speed and direction meter. The findings of these surveys have shown
the river to be flowing upstream at one depth in the water column and down-
stream at another depth with a complete reversal of this trend a few moments
later. Attempts to relate the inflows to wind speed, river flow and lake
level have proved fruitless.
The mean annual discharge for the river as measured by the USGS gauge
at Estabrook Park is 400 cfs (11.3 m3/sec) with the monthly variation in
flow shown in Figure 4. During the coldest portion of the winter months, the
river is ice covered in both the lower and upper reaches. The CSO area
that discharges to the river extends along both sides of the river with the
first outfall located at Capitol Drive as shown in Figure 2. A majority of
the outfalls in the lower portions of the river are completely or partially
submerged. Typical cross-sections of the river at upstream and downstream
locations and photographs of these sites are shown in Figures 5 through 8.
-------
1
0 AOOO 8000
(0) (1200) (2400)
SCALE, feet (meters)
x>
KINNICKINNIC
LAKE MICHIGAN
KINNICKI
RIVER
LAKE MICHIGAN
Figure 1. Illustration of the CSO drainage areas contributing
to the various receiving waters in Milwaukee, Wisconsin.
-------
£ USGS FLOW GAUGING STATION
A RAIN GAUGE
• CONTINUOUS DO MONITOR
PORT WASHINGTON ROAD
ESTABP.OOK PARK
MILWAUKEE RIVER
HOLTOM STREET >• A
MENOMONEE RIVER
WALNUT STREET
CHERRY STREET
WELLS STREET
ST. PAUL AVF.
BARTLETT AVENUE
NORTH AVENUE DAM
LAKE MICHIGAN
BROADWAY STREET
Figure 2. Monitoring locations in the study area of the Milwaukee River.
-------
TABLE I. MILWAUKEE RIVER WATERSHED CHARACTERISTICS
2 2
Total drainage area - mi (km )
Populat ion
Land use - percent of area
Res i dential
Commercial
Industrial
Agricultural
Open
,2
Combined sewer drainage area - mi
Estabrook Park flow - cfs (nr/sec)
Mean annual
Seven day - ten year low flow
100 year flood flow
(km2)
695 (1800)
530,000
6.1
0.3
0.4
61 .0
32.2
9.5 (24.6)
400.0 (11.3)
19-2 (0.5)
14000.0 (396.5)
Figure 3- Photograph of the North Avenue Dam on the Milwaukee River.
-------
oo
in
14-
o
I ,000
800
600
AOO
200
35
30
25
20
15 ^
10
JAN. FEB. MAR. APR. MAY JUNE JULY AUG. SEP. OCT. NOV. DEC.
Figure 4. Mean monthly flows for the Milwaukee River at the Estabrook Park USGS flow gauging
stat ion.
-------
WEST
EAST
% o (o)
^E
*JO (3)
11
I
20 (6)
0
(o)
100
(30)
200
(60)
WIDTH, feet (meters)
Figure 5. Photograph and cross section of the Milwaukee River
at St. Paul Avenue.
-------
WEST
- 20
i
LU
(0)
(3)
(6)
I
0
(0)
I
EAST
100
(30)
WIDTH, feet (meters)
J
200
(60)
Figure 6. Photograph and cross section of the
Milwaukee River at Walnut Street.
•
-------
•
I
'!'
WEST
0 (0)
S 10 (3)
i
o
(o)
1
1
100 200 300
(30) (60) (90)
WIDTH, feet (meters)
EAST
(120)
Figure ?• Photograph and cross section of the Milwaukee River
at North Avenue.
'= '
-------
SOUTH
NORTH
"£ 0 (0
-------
Combined Sewer Area
The combined sewer area that contributes to the Milwaukee River is approximat-
ly 6000 ac (2428 ha) with a total of 52 outfalls ranging in size from 12 in.
(30.5 cm) to a double 10 by 7.5 ft (3 by 2.3 m) box outfall. The drainage
areas for individual outfalls range in size from 5 to 702 ac (2 to 284 ha).
The entire combined sewer area has been modeled in a previous project using
the EPA Storm Water Management Model (SWMM) and the Army Corps of Engineers
Storage Treatment Overflow Runoff Model (STORM)(ll). The land use percentages
of the CSO area from the model data are as follows:
Land use
Single family residential
Multi family residential
Commerci al
Industri al
Pa rkland
Percent of CSO area
57.6
12.6
17.8
11.3
0.7
Figure 9 presents a representation of the CSO drainage area that contributes
to the river as a function of distance from Lake Michigan.
Storm Sewer Area
Those areas within Milwaukee County that drain to the Milwaukee River and are
served by storm sewers comprise approximately 27,000 ac (10,900 ha) of
drainage area. These areas contain the following land use breakdown:
Land use
Single family residential
Multi family residential
Commerci al
I ndustri al
Parkland - open area
Percent of storm sewer area
50.1
2.7
12.9
4.4
29-9
This area has also been modeled with the SWMM and STORM models in a previous
project (II). A large portion of this drainage area contributes to Lincoln
Creek (Figure l) which eventually discharges to the Milwaukee River.
Figure 10 illustrates a comparison of the storm sewer and CSO areas. Some of
the storm sewers within this area contain cross connections with the sanitary
sewers so that during wet weather the surcharged sanitary system may be
relieved through nearby storm sewers. Further descriptions of this system
will be developed in Section 6, Modeling Studies.
Upstream Areas
2 2
The Milwaukee River watershed is comprised of 644 mi (1668 km ) of drainage
area that lies to the north of the Milwaukee metro area. The predominant land
use within this area is agricultural in nature which comprises approximately
66 percent of the upstream watershed. Urban land use accounts for 1.5 percent
while open areas and woodlands make up the remaining area. in 1970, an
estimated 530,000 persons resided in the entire watershed including the area
within Milwaukee County.
13
-------
100
80
2 60
an cj
< z
si AO
O OQ
u. o:
o H-
H o 20
z <_>
LU
52
50
OQ
O
O
U.
O
cc.
LU
OQ
z>
30
20
10
O LU
mi
(ki
i-
Q.
<
\f"i
>
0£
O
J
lometers) '
7
h-
Qi
O
Z
t 3
<
a
1 1 1
6 5 *»
DISTANCE
1- 1-
ID LU
Z LU
_l OL
< 1-
2 1
1 1
3 2
WELLS
STREET
1
<
a.
I-
y>
0
1
0
z>
z
LU
<
Figure 9. Illustration of the distribution of CSO outfalls and
contribution from the CSO drainage area along the Milwaukee River,
-------
I
MILES
km = mi x 0.62
OZAUKEE COUNTY
GREENFIELD AVENUE
KINNICKINNIC RIVER
Figure 10. Illustration of CSO and storm sewer areas tributary
to the Milwaukee River.
15
-------
PROJECT BACKGROUND
Two projects within the Milwaukee area that have dealt with some aspect
of the CSO problem have preceded the subject project. Some of the findings
and data from these projects were utilized in the subject project to verify
these investigations. In order to acquaint the reader with this background
history, the following brief discussions are presented.
Humboldt Avenue Detention Tank Project (12)
Within this demonstration project, the City of Milwaukee evaluated the merits
of detention tanks as a practical method for the abatement of CSO. A 3-9
million gal. (IA,800 m3) CSO detention tank was constructed to intercept
overflow from a 570 ac (230.8 ha) segment of the Milwaukee River CSO area.
As part of the evaluation of the facility, an extensive sewer and river
monitoring program was conducted in conjunction with a detention tank and
river modeling program. This project was carried out under partial sponsor-
ship of EPA and was completed in late 1972.
Milwaukee Combined Sewer Overflow Pollution Abatement Project (I I)
The Milwaukee Metropolitan Sewerage District in late I97A initiated a study
of the entire CSO area within Milwaukee to assess the impact of CSO on the
area rivers and to develop the most cost effective CSO abatement alternative
to improve water quality. This project was partially funded by EPA as a
Section 201 Construction grant with the firm of Stevens, Thompson and Runyan
(now STRAAM Engineers) as the prime consultant. The 3.25 million dollar
project included water quality monitoring and modeling to define the
frequency and magnitude of any instream impacts. The modeling of the entire
CSO area using the SWMM and STORM models was completed in this project as
well as the instream modeling of the Milwaukee River using Harper's water
quality model. The water quality evaluations of this project were limited
to the determination of impacts associated with various levels of protection
for selected abatement alternatives using design storm analyses. The project
is presently evaluating the abatement strategy of conveyance-storage-treatment
using deep tunnel conveyance and mined underground storage.
An additional portion of the Milwaukee CSO project was the modeling require-
ments necessary to meet the PRM 75-3^ requirements of EPA which states that
the selection of the most cost effective CSO abatement alternative will be
based upon a comparison of the relationship between the water quality
improvement associated with each abatement measure and its resulting cost.
For this project the Environmental Sciences Division of Envirex (now the
Environmental Research Center of Rexnord Inc.) with assistance from Harper-
Owes Consultants was contracted to carry out the above analyses (13). In
order to evaluate the frequency and magnitude of water quality standards
violations, Harper's model was modified to simulate long term periods (years)
of rainfall. The results of this portion of the CSO project have shown the
conveyance-storage-treatment concept to be the most effective in improving
the water quality of the Milwaukee River when it is designed for a 1/2 year
recurrence interval storm. Storms of lesser frequency provide little
additional improvement for a substantial increase in cost.
16
-------
REPORT ORGANIZATION
The remaining portions of this report will describe the data collection and
analyses used to define the CSO impact within the Milwaukee area. Section 4,
Field and Laboratory Investigations, will describe the monitoring tasks
within the project that were used to quantify the magnitude of the CSO
impacts during and after wet weather discharges. These investigations include
the instream and combined sewer monitoring that was used to verify the
model network used in the evaluation of the impacts. Section 5, Analysis
of Data, uses the data of Section 4 to determine the source and mechanism
of the wet weather impacts, Section 6, Modeling Studies, describes how the
model network was used to simulate the water quality conditions which led to
the evaluation of the CSO contribution to the individual impacts. Descrip-
tions of the instream model are also provided in this section. Section 7,
Evaluation of CSO Impact, contains the analysis of the CSO impact and relates
the improvements in water quality associated with control techniques. The
benefits of CSO control, as well as the relationship of this project to a
national evaluation are also included.
17
-------
SECTION k
FIELD INVESTIGATIONS
The water quality conditions within the Milwaukee River exhibit extreme
variations between wet and dry weather periods. Figure II presents data
from a continuous DO monitoring device located at St. Paul Avenue showing
the variation in DO. Figure 12 presents fecal coliform data from the same
site during two intensive monitoring surveys. The dramatic loss of dissolved
oxygen and rapid increases in fecal coliform concentrations were investigated
in the field monitoring portions of this project. These' consisted of
of comprehensive field surveys, the analysis of two years of data taken from
three continuous DO and temperature monitors (Figure 2), dye studies to
define the rivers hydraulic and mixing characteristics, and sediment
investigations.
This section of the report will describe these four components of the
field monitoring program which were used to decipher the contribution of
CSO to this impact. The term water quality impact for this study will be
limited to the violations of the DO and fecal coliform standards that have
been set by the Wisconsin Department of Natural Resources for the Milwaukee
River. The standards apply for both dry and wet weather flow conditions in
the river and are as follows:
Upstream of North Avenue Dam
I. Preservation and enhancement of fish and other aquatic life. The
DO content shall not be lowered to less than 5 mg/1 at any time.
2. Full body contact recreational use; the membrane filter fecal
coliform count (MFFCC) shall not exceed 200 per 100 ml as a
geometric mean based on not less than five samples per month.
Downstream of North Avenue Dam
I. Marginal conditions for fish and aquatic life. The DO shall not
be lowered to less than 2 mg/1 at any time.
2. Partial body contact recreational use; the MFFCC shall not exceed
1000 per 100 ml as a geometric mean based on not less than five
samples per month.
The difference in standards between the upper and lower portions of the
river reflect the physical and biological characteristics of these two
18
-------
RAINFALL
Dl
E 6
5
13
O
o 3
LU J
O
CO
12 I
O
I 0.36 in. (0.91 cm)
DRY WEATHER
1200 2400 1200
TIME, hours
2400
Figure II. Illustration of dry weather (July 29-31, 1977) and wet weather
(August 3-5, 1977) DO levels at St. Paul Avenue in the Milwaukee River.
E
O
O
\ I05
O
O
u.
\0L
O
O
10-
I I
1200 2400 1200 2400 1200
TIME, hours
Figure 12. Illustration of dry weather (September 22-24, 1976) and wet weather
. (August 4-6, 1977) fecal coliforms levels at St. Paul Avenue in the Milwaukee
River
19
-------
reaches. The upper portions above the dam are free flowing and not under
the influence of Lake Michigan while the lower portions are those reaches
which exhibit some degree of inflow from the lake. Figure 9 presents the
distribution of the CSO drainage area as a function of distance along the
river. The North Avenue Dam is located where approximately 55 percent of
the CSO area has contributed to the upstream portions of the river, but
only 12 of the 52 outfalls along the river discharge upstream of this location.
The field investigations conducted as part of this project include four
components which were carried out concurrently. The first component was
the hydraulic studies to characterize the flow and dispersion characteris-
tics of the river and of selected CSO's. The second was the intensive
sampling of the river for three or more days following rainfall events. The
third was continuous monitoring of DO and temperature. The final component
was the sediment investigations which were initiated after the source of the
DO impact was identified in previous studies (ll).
HYDRAULIC STUDIES
As was discussed earlier in Section 3, the portion of the Milwaukee River
in the study area is characterized by two hydraulically unique reaches. The
upstream reach, stretching from Estabrook Park to the North Avenue Dam, is a
relatively shallow, free flowing stream containing several areas with rapids.
Conversely, the downstream reach from the North Avenue Dam to the river mouth
is a deep, sluggish stream greatly influenced by the backwater effects of
Lake Michigan. These distinct differences result in widely varying stream
parameters, such as velocity, travel time, and dispersion. Variations in
these parameters greatly influence the impact of both storm and CSO discharges
in the Milwaukee River.
In order to investigate the stream characteristics, two types of dye studies
were undertaken. The first type involved the monitoring of dye in both the
upstream and downstream reaches of the river. This study enables the comput-
ation of actual instream dry weather velocities, travel times and dispersion
coefficients and permitted quantitative comparison of the two reaches. The
second set of dye studies was aimed at measuring CSO mixing characteristics
in the river. Both studies have justified and verified several of the
assumptions and coefficients used during the computer modeling. Likewise,
they have illustrated several river flow and CSO mixing patterns and the
sampling techniques required to adequately measure their interaction. In
addition to the dye studies, velocity measurements near the sediments in
the river were conducted to determine the scouring potential of river
velocities and velocities due to CSO discharges.
Dye Sampling Program
Both the river and CSO dye studies were undertaken with Rhodamine WT dye.
This dye was used since it is not affected appreciably by organics, and it
is approved by the U.S. Geological Survey (I1*)- For the river surveys,
dye was injected as a single slug at Estabrook Park for the upstream reach
and at the North Avenue Dam for the downstream reach. The CSO dye studies
20
-------
began by injecting dye in the outfall sewer at the intercepting devices
(dividing structure for flow to treatment plant and overflow to river) to
insure proper mixing before any backwater effects from the river were
expressed. In both dye studies samples were taken at downstream bridges.
Initial samples taken near the dye injection sites were gathered both
cross-sectionally and with depth until complete lateral mixing was observed.
The sampling at downstream sites was limited to surface and bottom samples
at the center of the stream.
The actual sample gathering was done manually and with automatic sampling
devices. Fluorometer readings taken at the laboratory were used to determine
the dye concentrations. Sampling began early enough to capture the dye flow
at each monitoring site and was continued until the entire dye plume had
passed.
Instream Results
The first of three river dye studies began on November 8, 1976. Dye was
injected at Estabrook Park and sampled at Capitol Drive and the North Avenue
Dam. Figure 13 illustrates the monitored concentrations at the two sites.
The last two studies were completed on the downstream reach of the Milwaukee
River. On March 30, 1978, dye was injected at the North Avenue Dam with
samples subsequently collected at Humboldt Avenue, Walnut Street, Kilbourn
Boulevard, and St. Paul Avenue. On June 13, 1978, a second dye injection
was made at the dam and samples were collected at Walnut Street, Wells
Street and St. Paul Avenue. Figure \k shows representative dye concentration
curves from these injections.
Based on the sampling results, both the upstream and downstream reaches
were divided into two smaller reaches for more detailed analysis. Using
the methods outlined in Godfrey and Frederick (15) and Fisher (16) the
mean velocities, travel times and dispersion coefficients of each reach
were computed. The characteristics of the four reaches; Estabrook Park to
Capitol Drive, Capitol Drive to North Avenue Dam, North Avenue Dam to
Walnut Street and Walnut Street to St. Paul Avenue, for the three surveys
are summarized in Table 2.
An examination of the overall trends shown in Table 2 supports the conditions
that would be expected from a stream with the physical characteristics of
the Milwaukee River. The reaches with the least downstream hindrance,
Estabrook Park to Capitol Drive and North Avenue Dam to Walnut Street, have
the highest velocities during both high and low flow conditions. The two
reaches below these are influenced from backwater at North Avenue Dam and
Lake Michigan and accordingly have lower velocities.
As would be expected, the backwater influence from Lake Michigan is less
during high flows. During the June 13, 1978 survey, a flow of 215 cfs
(6.1 m3/sec) was recorded. The resulting velocity from Walnut Street to
St. Paul Avenue was only 38 percent of the velocity from the North Avenue
Dam to Walnut Street. This ratio increased to 51 percent during the
1700 cfs C»8.l m3/sec) high flow condition of the March 30 survey.
21
-------
350
- 300
CT!
3.
AT CAPITOL DRIVE
o
z
o
250
200
150
100
50
d = 91 cfs
(2.6 m3/sec)
2 4 6 8 10
TIME AFTER DYE INJECTION, hrs
01
3.
2 30
20
10
0
Q = 91 cfs
(2.6 m3/sec)
AT NORTH AVENUE DAM
10
20
25
30
35
TIME AFTER DYE INJECTION, hrs
Figure 13. Monitored dye concentrations at Capitol Drive and North Avenue
Dam during the November 8, 1976 Milwaukee River dye study.
22
-------
200 r~
LU
O
§100
Q-215 cfs
(6.1 m3/sec)
dbdh
AT WALNUT STREET
' I I I I I I I 1 I
6 8 10 12
TIME AFTER DYE INJECTION, hrs
60 I—
en
<_>
o
>•
o
20 __
10 —
AT ST. PAUL AVENUE
TIME AFTER DYE INJECTION, hrs
Figure lA. Monitored dye concentrations at Walnut Street and St, Paul Avenue
during the June 13, 1978 Milwaukee River dye study.
-------
TABLE 2. MILWAUKEE RIVER FLOW CHARACTERISTICS DETERMINED BY DYE TRACINGS
N>
Tracing date
Hovembe r 8 ,
1976
March 30,
1978
June 13,
1978
Flow1
rate,
cfs Reach
91 Estabrook Park-
Capitol Drive
Capitol Drive-
North Ave. Dam
1700 North Ave. Dam-
Walnut Street
Walnut Street-
St. Paul Ave.
215 North Ave. Dam-
Walnut Street
Walnut Street-
St. Paul Ave.
Reach
length,
ft
5,000
1 2 , 1 70
4,650
6,750
4,650
6,750
Cross-sectional2
area,
ft2
_
_
2,380
3,230
2,380
3,230
Estimated
mean
velocity,
ft/sec
.
-
0.71
0.53
0.09
0.07
Mom toredS
mean
velocity,
ft/sec
0.31
0.09
0.78
0.40
0.16
0.06
Estimated
travel time,
hrs
.
_
1.82
3-54
14.4
26.8
Mon i tored*
travel time,
hrs
4.58
39.5
1.67
4.67
8. OS
29- 1
Longi tud tnal
dispersion coefficients
Godfrey-Frederick,
ft2/sec
30
24
30
50
46
4
Fischer,
ft2/sec
49
15
78
124
63
i
As measured at the USGS stream gauge at Estabrook Park.
Average cross-sectional areas measured during typical flow conditions.
Mean velocity and travel tine based on the time between the centroids in the dye concentration curves measured at the sampling sites.
Conversions: cfs x 0.02332 « raVsec
ft2 x 0.3048 - m
ft2 x 0.0929 - m2
ft/sec x 0.3048 - m/s«c
ft2/sec x 0.0929 - ro2/sec
-------
The same trends exhibited in velocities are followed inversely by the
computed tracer travel times; the greater the velocity a reach has the
shorter its tracer travel time. As a rough check of the dye study
velocities and travel times below the North Avenue Dam, estimated values
were computed from the reach lengths and the continuity equation:
Q = VA
Where: Q = Flow rate.
V = Velocity.
A = Typical reach cross-sectional area.
The flow rates used in the continuity equation were measured upstream of
the study area and were not adjusted for downstream flow contributions for
two reasons. First, all of the river dye studies were conducted during
dry weather and flow monitoring at the North Avenue Dam indicates that there
is little discernible change in flow between Estabrook Park and the dam
during these conditions. Secondly, there are no major tributaries to the
river within the study area. The cross-sectional areas used for the estimated
velocities and travel times were rough estimates and were not adjusted for
the different flow rates. This is not a significant factor because the
cross-sectional areas downstream of the North Avenue Dam are determined by
the water level of Lake Michigan. Due to these factors, the estimated
velocities and travel times can only serve as a yardstick with which to
compare the dye study values. Nevertheless, the values compared well and
indicate a reliable set of dye study data.
The most significant of the dye study results are the dispersion coefficients.
These dispersion coefficients are a relative measure of the significance
of longitudinal (or downstream) dispersion in the Milwaukee River. Longitu-
dinal dispersion is defined as the action by which a flowing stream spreads
out and dilutes a mass of pollutant (17). A pollutant such as CSO enters
a stream and instead of traveling downstream as a slug, spreads itself
out along the length of the stream as some parts travel faster and some
slower than the mean flow velocity.
The rate of longitudinal dispersion is an important parameter in the
modeling of pollutant loads to receiving streams. If longitudinal dispersion
is significant, complex dispersion terms must be included in the modeling
transport equations. The longitudinal dispersion coefficients for the
Milwaukee River were computed over a wide range of flow conditions by the
two well regarded methods presented earlier. The values are listed by
individual reaches and flow conditions in Table 2.
The computed longitudinal dispersion coefficients are relatively low values.
This indicates that longitudinal dispersion is a minor component of the
total transport mechanism of the Milwaukee River and need not be modeled.
A low longitudinal dispersion is typical of streams under the backwater,
estuarial influence of a large body of water. The Clinch River in Tennessee
is similar to the Milwaukee River in both width and depth, yet, as opposed
to the Milwaukee River, it is a free flowing stream without backwater
25
-------
influences (15)- In a dispersion study under similar discharges, the
Clinch River had a longitudinal dispersion coefficient of 500 ft2/sec
(^6.5 m2/sec) at l800 cfs (51.0 m3/sec) and a coefficient of 150 ft2/sec
(13.9 m2/sec) at 323 cfs (9.1 m3/sec). Comparison with the values in
Table 2 illustrates the minor influence of longitudinal dispersion in the
MIIwaukee River.
The relative importance of even a minor longitudinal dispersion transport
does depend somewhat on the type of pollutant load to the river. An
instantaneous loading comparable to an accidental fuel or toxic waste
spill would illustrate significant longitudinal dispersion characteristics
even in the Milwaukee River. However, in the study and modeling of CSO
and storm sewer discharges which can last several hours and are distributed
over a large stretch of the river, longitudinal dispersion ,would still be
minor in comparison to convective transport.
CSO Results
The aim of the CSO dye studies was to measure the mixing capabilities of
the CSO under different discharge magnitudes. Dye was added as a slug
and the instream mixing was measured by sampling at a downstream bridge at
three equidistant sites across the bridge and with depth. Dye was injected
into the 96 in. (2kk cm) outfall upstream of Walnut Street and into the
Ax5 ft (1.2x1.5 m) outfall upstream of Cherry Street. Both outfalls are
submerged and are located on the west bank of the river.
The two dye injections upstream of Walnut Street were made under low
discharge velocities. Visual tracking of the dye showed no lateral mixing
and the vast majority of the dye moved slowly down the west bank of the
river. Subsequent analysis of the collected samples verified the visual
conclusions. In fact, during the hour of sampling after each injection
no dye at all was measured on the east side of the river. Only a minor
concentration was found at mid-depth of the central sampling site during
one injection. It is apparent that at low discharge flow rates mixing
does not occur rapidly.
Rapid mixing was observed during the dye injection at the *tx5 ft (1.2x1.5 m)
outfall. This outfall was discharging at a high rate and the dyed discharge
plume could be seen traveling across the entire 250 ft (76.2 m) of channel
width. This almost instantaneous lateral mixing was verified with the
samples collected at Cherry Street.
The effect of the submerged outfalls is dependent upon discharge rates and
the relative depth of the outfalls. Both outfalls studied are located
slightly below the mid-depth of the channel. At the low discharge rates,
most of the dye remains near the outfall and spreads fairly uniformly
from mid-depth to the surface. Little discharge appears to affect the
bottom. At the high discharge rates, the entire section from the bottom
to the surface is influenced.
26
-------
It follows that there is a general tendency for CSO to rise to the surface.
Consequently, a surface outfall will tend to show less mixing and less
bottom sediment resuspension than a submerged outfall in spite of similar
high discharges. With most of the CSO outfalls below North Avenue Dam
being submerged, good mixing of discharges, as well as river bottom sediment
resuspension can be expected.
Velocity Measurements
Physical measurements of velocity in the river near the bottom sediments
were taken in order to determine whether velocities due to increased river
flow or CSO discharges are sufficient to scour these sediments. This was
investigated due to the possible impact of resuspended sediments on DO
levels in the river. The instrument utilized for these measurements
recorded both velocity and direction so that the source of the changes in
velocity could be determined.
Velocity measurements taken at a bridge located 300 feet (91 rn) downstream
from St. Paul Avenue indicate that the river velocities in the proximity of
the sediments are generally quite low. During low flow conditions of about
100 cfs (2.8 m3/sec) and following 0.30 in. (0.76 cm) of rainfall on
May 31, '977, the measured velocities were negligible. The velocities
generally ranged from 0.0 to 0.25 ft/sec (0.0 to 0.08 m/sec), however, the
direction of this velocity oscillated between the upstream and downstream
directions. Following the 0.71 in. (1.80 cm) of rainfall on August 3
through August 5, 1977 and river flows reaching 498 cfs (14.1 m3/sec) at
the North Avenue Dam, there was still very little measurable velocity near
the bottom sediments in the lower river. Values during this survey
occasionally were as high as 0.40 ft/sec (1.2 m/sec) but were generally
in the range of 0.0 to 0.25 ft/sec (0.0 to 0.8 m/sec). The direction
of these velocities were quite variable as during the previous survey. This
is understandable since the velocity near the bottom sediments is generally
estimated as 25 percent of the mean stream velocity (18).
Velocity measurements in the vicinity of submerged.CSO outfalls showed
different results. Two outfall locations were monitored approximately
15 ft (4.6 m) from the outfall in the river and about 1.0 ft (0.30 m) above
the sediment surface. An outfall located in the east bank of the Milwaukee
River at St. Paul Avenue was monitored for ten separate events. This
outfall is a completely submerged A.0x8.5 ft (1.2x2.6 m) box. A second
xitfall located just upstream of Walnut Street on the west bank of the
river was monitored for six storm events. This outfall is also completely
submerged having a diameter of 8.0 ft (2.4 m). The invert of the pipe
entering the river is approximately 12 ft (2.7 m) below the water surface.
Following an extremely intense rainfall event on August 3, 1977, a velocity
of about 12.0 ft/sec (3.7 m/sec) was measured near the outfall at St. Paul
Avenue. There was 0.22 in. (0.56 cm) of rainfall during a 10 minute period
in the vicinity of this site. On July 31, 1978 a velocity of over 5-0 ft/sec
(1.5 m/sec) was measured at the Walnut Street outfall. Approximately 0.40 in.
(I.01 cm) of rainfall was recorded during a 10 minute time period during
this storm. Both of these measurements were made approximately 1.0 ft (0.3 m)
27
-------
above the sediment surface. There was severe clogging of the velocity and
direction meter during both of these events which may have resulted in
lower measured values than the actual velocities. The force of the discharges
was so large as to actually push the velocity and direction meter out away
from the outfall and lift it off the bottom. Other measurements of velocity
generally ranged from 1.0 to 5-0 ft/sec (0.3 to 1.5 m/sec). This data
indicates that there is considerable potential from submerged CSO outfalls
to scour river sediments.
INTENSIVE SURVEYS
Each of the intensive sampling surveys is related to the volume, duration
and distribution of rainfall in the study area. Other factors that were
evaluated include the locations in the river where the impacts are present,
the river flow as related to the magnitude and duration of the impacts,
and the prestorm conditions which affect the magnitude of the impacts.
Because of the ice covered conditions of the Milwaukee River during the
winter, only the warm weather periods were evaluated in the monitoring
program.
The six intensive surveys conducted during this project were used to define
the changes in the quality of the Milwaukee River at five locations shown
in Figure 2. These locations are described as follows:
I. Capitol Drive - Port Washington Road: This site is situated
immediately upstream of the CSO area and is used to quantify the
incoming loads to the study area. Survey I uses the Capitol Drive
location while subsequent surveys use Port Washington Road because
of a more accessible location.
2. North Avenue: The second monitoring location is situated
immediately above the North Avenue Dam, providing data on those
portions of the river that are influenced by CSO but are above
the lake inflows. A continuous DO and temperature monitor is
located here.
3. Walnut Street: This monitoring site is situated within the lake
influenced areas of the river approximately ^600 ft (\kOO m)
downstream of the dam.
4. Wells Street: Another site within the lower portions of the river.
5. St. Paul Avenue: This site is the final location within the lake
influenced portions of the river which lies immediately above the
confluence of the Menomonee River. A continuous DO and temperature
monitor is located at this site.
The intensive surveys were conducted for a period of three to five days
with samples taken at intervals from three to six hours to try and quantify
the changes in water quality after various rainfall events. One survey
was conducted during dry weather to establish the base line conditions
which exist in the study area after the effects of previous wet weather
28
-------
discharges have abated. All samples for chemical analysis were collected
from bridges in cross section and composited before being transported to
the laboratory for analysis. Fecal coliform samples were collected at
mid-channel approximately 2 ft (0.6 m) below the water surface. Flow
measurements used in the surveys were available from the USGS gauge
located at Estabrook Park and the stage recorder installed at the North
Avenue Dam during the final year of the project. Each of the surveys will
be individually described in the remaining portions of this section.
Analysis of the data from these surveys is discussed in Section VI. Data
summaries for these surveys are listed in Tables A-l through A-13 in the
Appendix.
Survey I - Dry Weather Flow
The first intensive monitoring survey was conducted on September 21-23, 1976
to define the dry weather conditions within the CSO influenced portions of
the river. Table 3 lists the rainfall and river flow conditions during
each of the six intensive surveys that are part of the project. Survey I
was an ideal dry weather survey because the flows in the Milwaukee River
were extremely low after a long, dry summer season. Figure 15 presents the
DO conditions monitored at the five locations previously described for the
intensive surveys. The DO levels observed at the upstream boundary (Capitol
Drive) and at North Avenue range from 5*7 to 9-1 mg/1. These values
represent 60 and 95 percent satuation, respectively. The physical
characteristics of the upper reach of the river tend to make the trends in
DO at these two river sites appear quite similer. As shown previously in
the cross section figures of Section 3, the upper portions of the river
are extremely shallow. This results in substantial growth of macrophytic
plants which cause the observed diurnal variations in DO at these sites.
In the lower portions of the river, the DO levels are significantly depressed
when compared to the upper reaches. At Walnut Street, which is approximately
*»600 ft (lAOO m) downstream of North Avenue, the DO values ranged from
1.0 to A.O mg/1 less than the values measured in the upper river. At
St. Paul Avenue, the DO levels were less than 2.0 mg/1 during most of the
survey. Also, there is only a slight diurnal variation in the DO values in
the lower river during this survey. This portion of the river is much
deeper than upstream and thus, any changes in DO due to photosynehetic
activity must be due to phytoplankton. Measured secchi depths range from
approximately 18 to 30 in. (/*5-7 to 76.2 cm) and measured chlorophyll a_
values indicate that a substantial population of phytoplankton are present
at the time of the year this survey was conducted (Appendix Table 14). The
short photoperiod in September and the depth of the river precludes any
observable impact on DO levels.
The fecal coliform data of Figure 16 indicates that residual effects of
previous overflows from upstream were present during this survey. The
trend of decreasing coliform counts at four monitoring locations as the
survey progressed indicates that these sites could have been influenced by
upstream discharges that occurred during a rainfall that did not cover
the study area. The Walnut Street data does not show this relationship,
29
-------
TABLE 3. SUMMARY OF RAINFALL AND RIVER FLOW CONDITIONS FOR INTENSIVE MONITORING SURVEYS
Average rainfall
Survey
number
1
2
3
*»
5
6
Date survey
ini t iation
9-21-76
5-31-77
6-18-77
8-3-77
6-16-78
7-26-78
Survey duration ,
days
3
3
3
5
A
3
Total
vol
in.
0.0
0.30
0.5^
0.71
1.86
0.63
survey
ume ,
(cm)
(0.76)
(1.37)
(1.80)
(4.72)
(1.60)
Duration ,
hrs
0.0
1.70
0.97
3-19
8.80
1.84
Antecedent
dry per iod ,
days
II
40
6
10
4
5
Peak flow at
Estabrook Park,
cfs (m3/sec)
85 (2.41)
154 (4.36)
955 (27.05)
498 (14.10)
31 12 (88.13)
599 (16.96)
-------
10
D)
LJ
X
o
o
UJ
>
_l
O
CO
CO
ESTABROOK PARK FLOW:
70 - 80 cfs.
(19-8 - 22.7 mVsec)
.CAPITOL DRIVE
NORTH AVENUE
WALNUT STREET
ST. PAUL AVENUE
>400 1200 2400
21 SEPTEMBER
1200
22 SEPTEMBER
24jOO
r
1200 2*400
23 SEPTEMBER
TIME, hours and days
Figure 15- Monitored DO levels in the Milwaukee River during dry weather, September 21-23, 1976 (Survey 1)
-------
10,000
O
o
O
C_J
cc
o
o
o
o
UJ
1 ,000
100
' CAPITOL DRIVE
I ^_^_- NORTH AVENUE
— WALNUT STREET
ST. PAUL AVENUE,
\
2400
1200 2400 1200 2400
21 SEPTEMBER I 22 SEPTEMBER
1200 2400 1200
23 SEPTEMBER ' 24 SEPTEMBER
Figure 16.
TIME, hours and days
Monitored fecal conform levels in the Milwaukee River during dry weather
September 21-24, 1976 (Survey I).
-------
but rather shows a continuous high level of coliforms except for one value on
September 23rd. No sanitary overflows are known to exist in this area but
an obvious source of fecal coliforms must be present to provide these high
values. Dry weather surveys conducted previously as part of the Milwaukee
CSO project (II) did not show these same elevated values. The final day
of the survey shows the coliform concentrations leveling off to approximately
200 counts per 100 ml. The high counts at the upstream boundary of the CSO
area throughout the survey seems to reinforce the concept that a storm
passed through the upper portions of the basin without affecting the study
area raingauges.
Survey 2 - Wet Weather Flow
The first wet weather data collected as part of this project was during
Survey 2 which took place on May 31 to June 2, 1977. The survey was initiated
as 0.30 in. (0.76 cm) of rainfall was recorded for the area on May 31. The
river flows before and during the survey were again extremely low tor this
event because of the long dry period before this survey. The previous rainfall
event which was greater than O.I in. (0.25 cm) was recorded on April 20,
1977 which means approximately *»0 days of dry weather preceeded this event.
The DO and fecal coliform conditions are presented in Figures 17 and 18. The
upstream boundary station for the wet weather surveys was changed to Port
Washington Road rather than Capitol Drive because of accessibility and the
location of a combined sewer at the Capitol Drive location which could
affect the sampling data. The DO data illustrated for the North Avenue Dam
and St. Paul Avenue sites prior to the storm event on May 3' was obtained
from the continuous DO and temperature monitors.
The DO data of Figure 17 shows the upstream DO to range from 5-2 to I 1.0 mg/1.
This represents approximately 60 and 120 percent of saturation with a major
portion of the variability being associated with photosynthesis. On June I,
there is little variation in DO with the values being around 6.0 mg/1 which
is a result of the 100 percent cloud cover and two percent sunshine on this
day. At the North Avenue Dam site, there was a rapid decline in DO during
the first 8 hours after the rainfall event followed by a steady recovery in
DO for the remainder of the survey. The DO at this location was below the
5-0 mg/1 standard for portions of all three days of this survey. The pre-
storm DO data was obtained from the continuous DO monitor.
At Walnut Street, the DO values were in the range of 2.3 to 3-9 mg/1 during
the entire survey. Wet weather impacts are difficult to discern at this
site since the DO values are only slightly lower than observed in the dry
weather survey. The Wells Street location exhibited a decline of DO from
2.6 to nearly 0.0 mg/1 within 18 hours after the rainfall. The DO values
remained below 1.0 mg/1 until late on June 2. Recovery of DO after this
event was extremely slow due to the low river flow conditions. The St. Paul
Avenue site exhibited DO levels that were between 1.0 and 3-7 mg/1. For
much of the survey the DO at this site was in the same range as at Walnut
Street. This observation was probably .due to the effect of Lake Michigan
in the lower reaches of the river. It must be mentioned that the DO
values of Figure 17 were averaged over depth for each location. At Walnut
33
-------
VjO
-e-
PORT WASHINGTON ROAD
NORTH AVENUE
WALNUT STREET
WELLS STREET
ST. PAUL AVENUE
12
10
X
o
o 6
LU
>
_J
O
to
O 4
0
RAINFALL
I
fO.30 in.
(0.76 cm)
I
0.0*1 in.
(0.10 cm)
i
i
i
* i I . — '—I ••"" I
•T •!• I — I
f
400 1200
30 MAY
2^00
1200
2 JUNE
2400
1200 2400 1200
31 MAY J 1 JUNE
TIME, hours and days
Figure 17- Monitored DO levels in the Milwaukee River during wet weather, May 3l-June 2, 1977 (Survey 2)
-------
PORT WASHINGTON ROAD
NORTH AVENUE
•WALNUT STREET
,WELLS STREET
ST. PAUL AVENUE
RAINFALL
'°°.°°°PI^V"»)
10,000
o
o
o
U.
,000
o
_i
<
100
1
0.04 in.
(0.10 cm)
/ /«
2^00
1200
31 MAY
2400
2^00
1200
1 JUNE I
TIME, hours and days
1200
2 JUNE
2^00
Figure 18. Monitored fecal coliform levels in the Milwaukee River during
wet weather, May 3l~June 2, 1977 (Survey 2).
35
-------
Street the DO only varied by less than 1.0 mg/1 with depth while at St. Paul
Avenue the variation was as high as 5.0 mg/1. Significant temperature
variations were also found at this location which indicates the cooler, high
DO lake water is probably masking the normal DO sag for this location.
The fecal coliform results for Survey 2 are presented in Figure 18. The
most noticeable feature of this figure is the very low counts at the upstream
boundary of the CSO area throughout the duration of the survey. The North
Avenue Dam site exhibits the highest concentrations during the first two days
of the survey but then rapidly decline on the third day to levels very close
to the upstream boundary (Port Washington Road). St. Paul Avenue remains
fairly low in fecal coliform levels throughout the first two days of the
survey and then rapidly increase to much higher levels. The low flow
conditions during this event could have allowed the lake influence to be
much more pronounced at this location, thus diluting the coliform levels.
The remaining downstream sites are fairly consistent throughout the survey
with values in the I02 to 103 range which are an order of magnitude less
than at the North Avenue site. Again, the lake influence on the Walnut and
Wells sites would tend to dilute the coliform concentrations even though a
majority of the CSO discharges occur below the North Avenue Dam.
Survey 3 - Wet Weather Flow
The second wet weather event was conducted on June 18 through June 21, 1977.
This survey is an example of a wet weather event which included a substantial
rainfall but little observable impact on DO levels in the river. The rain-
fall data from this survey is unique in that the average volume of rain which
fell in the Milwaukee area was 0.5*» in. (1.37 cm) with one gauge recording
0.50 in. (1.27 cm) in a 15 minute period. The average daily river flows
during this survey ranged from I51* to 256 cfs (A.k to 7.2 nP/sec) with the
peak river flow occurring approximately 3 hours after the rainfall event
ended with a flow of 955 cfs (27 m3/sec) at Estabrook Park.
The DO conditions in the river during this survey are presented in Figure 19.
These indicate the same trends of decreasing DO concentration.s with •
distance downstream as observed during Surveys I and 2. At Port Washington
Road the DO levels were between 8.0 and 13-0 mg/1 throughout the survey.
The North Avenue site exhibited a large diurnal variation in DO with the
values ranging from 5^ to (60 percent of saturation. The sites in the lower
river also exhibited the same variation in DO, however, this variation
becomes more attenuated with distance downstream.
The data from the continuous DO monitors at North Avenue and St. Paul provide
the change in conditions before the rainfall event began. The North Avenue
site shows only a slight decrease in the diurnal peaks due to the storm
event. St. Paul Avenue presents a general decrease in DO from prestorm
to poststorm conditions.
The fecal coliform results for this survey are presented in Figure 20. At
all monitoring sites, the coliform concentrations rapidly decrease during the
second and third day of the survey. The upstream locations at Port Washington
36
-------
RAINFALL
10.53 in.
(1.35 cm)
PORT WASHINGTON ROAD
NORTH AVENUE
WALNUT STREET
WELLS STREET
ST. PAUL AVENUE
2400
1200
17 JUNE
Figure 19-
2400
1200
18 JUNE
2400
1200 2400
19 JUNE
1200
20 JUNE
2400
TIME, hours and days
Monitored DO levels in the Milwaukee River during wet weather,
June 17-20, 1977 (Survey 3).
-------
PORT WASHINGTON ROAD
RAINFALL
100,000
10,000
o
o
-1,000
a:
o
LL.
O
O
O
LU
100
NORTH AVENUE
WALNUT STREET
WELLS STREET
ST. PAUL AVENUE
2400
1200
18 JUNE
2400
1200
19 JUNE
2400
1200
20 JUNE
TIME, hours and days
Figure 20. Monitored fecal coliform levels in the Milwaukee River
during wet weather, June 18-20, 1977 (Survey 3).
38
-------
Road and the North Avenue Dam generally have the lowest concentrations
during the survey. The Port Washington Road location shows a rapid rise in
count during the first day of the survey which might be a function of cross-
connected storm sewers in the upper portions of the Milwaukee River drainage
area. Throughout the duration of this survey, the St. Paul Avenue site
contains the highest concentrations of fecal coliforms. This is the exact
opposite of Survey 2 when it exhibited the lowest levels. The difference
may be due to the higher river flows present during Survey 3- With
these conditions, the lake influence was not as prominent as in the
past events. One other point of interest in this survey is the very low
values for the Walnut Street site during the second day of the survey.
Survey k - Wet Weather Flow
This survey is unique to the remaining wet weather monitoring events because
of the continuous rainfall that was recorded on different days throughout
the five days of the survey. The initial rainfall on the study area on
August 3, 1977 averaged about 0.33 in. (0.84 cm). Additional rainfall
volumes are listed in Table k for the remaining days of the survey. The
DO data from this survey is presented in Figure 21 and in a slightly
different manner in Figure 22 in order to show the longitudinal trends. The
North Avenue Dam site had fluctuating DO levels between 2.0 to 5.0 mg/1
during the first day (August 4) after the start of the survey. The DO
recovered on August 5 and did not violate the 5-0 mg/1 standard for the
remainder of the survey. At the lower river sites, the DO decreased
progressively during the first 12 hours of the survey and then dropped to
nearly 1.0 mg/1 at these lower river sites by 1800 hours on August 4th. This
drop to 1.0 mg/1 is likely due to a second rainfall and overflow event which
occurred at 1500 hours on August 4th.
Between the Walnut Street and Cherry Street sites which is a distance of
approximately 1125 ft (343 m), the observed differences in DO levels
was as much as 5-5 mg/1 at 1800 hours on August 4th. Downstream from the
Cherry Street location, the DO remained near 0.0 mg/1 for nearly a full day.
Two additional storms on August 5th further suppressed the DO in the lower
river until August 6th when the DO recovered quite dramatically. Much of
the recovery can be attributed to the change in flow rate in the river.
Table 5 lists the average daily flow and range in flow at the
North Avenue Dam and Estabroofc Park during this survey. The high flows
towards the end of the survey may account for the rapid DO recovery observed
in the lower river sites.
The fecal coliform data for this survey is presented in Figure 23- Similar
trends as the previous wet weather survey are exhibited in this figure with
the St. Paul Avenue site showing the highest coliform levels and the upstream
boundary of the CSO area having the lowest levels. The rainfall throughout
the survey maintains the levels of fecal coliforms and prevents the flushing
of the coliforms by the higher than normal river flows which are present
during the final portions of the survey.
39
-------
TABLE 4. RAINFALL DISTRIBUTION IN CSO AREA DURING THE WET WEATHER
INTENSIVE SURVEY ON AUGUST 3-8, 1977 (SURVEY 4) '
Date,
1977
8/3
8/4
8/5
8/5
Time
2130
1515
1 145
1*420
Broadway
Rainfall
vol ume,
in. (cm)
0.36 (0.91)
0.23 (0.58)
0.3^ (0.86)
0.20 (0.50
Street
Duration,
hours
0.75
0.75
0.25
1.18
Hoi ton
Rainfal 1
vo 1 ume ,
in. (cm)
0.50 (1.27)
0.20 (0.51)
0.03 (0.08)
NA2
Street
Duration,
hours
0.83
0.83
0.50
Bartlett
Rainfall
vo 1 ume ,
in. (cm)
o.u (0.36)
0.08 (0.20)
0.00
0.00
Avenue
Duration,
hours
0.83
1.17
Location of raingauges is illustrated in Figure 2.
Not available.
-------
PORT WASHINGTON ROAD
NORTH AVENUE
WALNUT STREET
WELLS STREET
ST. PAUL AVENUE
16
12
r- 10
X
o
o
l/>
5
RAINFALL
10.33 in. I
(0.8*» cm) I
0.17 in.
(0.^3 cm)
I
0.22 in.
(0.56 cm)
I
0.07 in.
(0.18 cm)
2*»00 1200 2*400 1200 2^00
3 AUGUST
AUGUST
1200
5 AUGUST
2ifOO
1200
6 AUGUST
2^00 1200
21*00
7 AUGUST
1200
8 AUGUST
TIME, hours and days
Figure 21. Monitored DO levels in the Milwaukee River during wet weather, August 3-8, 1977 (Survey
-------
LU
13
o
Q
LU
>
_1
O
in
o
2300 HOUR 8-3-77
0600 HOUR 8-4-77
1200 HOUR 8-4-77
10 r—
1800 HOUR 8-4-77
2400 HOUR 8-4-77
1200 HOUR 8-5-77
DISTANCE, feet
m = ftxO.30
DISTANCE, feet
Figure 22. Monitored DO versus distance in the Milwaukee River during wet weather,
August 3-6, 1977 (Survey 4) (continued).
-------
12 r—
1500 HOUR 8-5-77
2100 HOUR 8-5-77
0600 HOUR 8-6-77
12
10
C3
I 6
O
o
to
to
••••••**»**•
1200 HOUR 8-6-77
1500 HOUR 8-6-77
21*00 HOUR 8-6-77
I I I I I I
o
o
o
o
er\
o
o
o
o
QC
o
CJ
a:
o
a.
o
o
o
UJ
UJ
O
O
I- >-
3: =3 o: to o. a:
»- z a: _j H
a: _i LU —J • =3
o < 3: uj I— o
z 3 o -3. to z
DISTANCE, feet
= ftxO.30
Figure 22 (continued).
DISTANCE, feet
-------
Xr
-C-
I .000.000
100,000
o
o
_ \
10,000
01
o
o
o
1,000
I
RAINFALL
0.33 in. • 0.17 in.
0.84 cm) 1(0.43 cm)
I
0.22 in.
(0.56 cm)
PORT WASHINGTON ROAD
NORTH AVENUE
WALNUT STREET
WELLS STREET
ST. PAUL AVENUE
2^00
1200
AUGUST
1200
7 AUGUST
2400
I
Figure 23-
5 AUGUST I 6 AUGUST
TIME, hours and days
Monitored feca] coliform levels in the Milwaukee River during wet weather, August 3-7, 1978
(Survey 4).
-------
TABLE 5. VARIATIONS IN RIVER FLOW AT THE NORTH AVENUE DAM AND
ESTABROOK PARK DURING THE INTENSIVE MONITORING SURVEY ON
AUGUST 3-8, 1977 (SURVEY 4)
North Avenue Dam
Date
1977
8/3
8/4
8/5
8/6
8/7
8/8
Average
flow,
cfs (m3/sec)
145
(4.1)
220
(6.2)
371
(10.5)
479
(13.6)
487
(13.8
419
(11.9)
Range in flow,
cfs (m3/sec)
97-523
(2.7-14.8)
1 12-534
(3.2-15.1)
178-440
(5.0-12.5)
342-522
(9.7-14.8)
441-514
(12.5-14.6)
371-507
(10.5-14.4)
Estabrook Park
Average
f low ,
cfs (m3/sec)
130
(3-7)
170
(A. 8)
344
(9.7)
461
(I3.D
448
(12.7)
383
(10.8)
Range in flow,
cfs (m3/sec)
105-61 1
(3.0-17-3)
98-418
(2.8-11.8)
176-405
(5-0-1 1.5)
312-498
(8.8-14.1)
399-491
(M.3-13.9)
33^-498
(9-5-14.1)
Survey 5 - Wet Weather Flow
In order to quantify the effects of wet weather discharges on the Milwaukee
River, the fifth survey was undertaken on June 16, 1978 while the flow in
the river reached over 3000 cfs (84 m3/sec). The average river flows
during the survey were 453 cfs (12.7 tn3/sec) on June 16th, 1623 cfs (45-4
m3/sec) on June 17th and 1645 cfs (46 m3/sec) on June l8th. The DO values
in the river during this survey are presented in Figure 24. An average
of 1.86 in. (4.72 cm) of rain fell on the study area during the survey
with one gauge registering 2.3 in. (5-8 cm). The resulting DO values in
the river show very little sag from the start to the end of the survey. By
the third day of the survey each monitoring location had approximately the
same DO level. These results are likely due to the rapid rate at which the
water was transported through the study area during the high flow conditions.
The fecal coliform concentrations in the river during Survey 5 are presented
in Figure 25. The high flows that occur during this survey seem to have
less of an impact on the coliform levels than in previous surveys. Again
the upstream sites show the lowest concentrations but they remain relatively
constant after the first day of the survey. The highest concentrations
were obtained at the Wells Street site especially during day 3 of the
-------
PT, WASHINGTON RD.
NORTH AVENUE
WALNUT STREET
— WELLS STREET
— ST. PAUL AVENUE
RAINFALL
10 i—
0.18 in.
(0.46 cm)
C3
X
o
0
2400
IB 0.44 in.
I (1.12 cm)
10.85 in.* 0.25 in.
(2.16 cm) I (0.64 cm)
./ \ r
1200
15 JUNE
Figure 24.
2400
1200 2400 1200
16 JUNE | 17 JUNE
TIME, hours and days
2400
1200
18 JUNE
Monitored DO levels in the Milwaukee River during wet weather
June 15-18, 1978 (Survey 5).
2400
-------
PORT WASHINGTON ROAD
NORTH AVENUE
WALNUT STREET
WELLS STREET
•••——ST. PAUL AVENUE
1,000,000
100,000
o
o
o
10,000
o
o
1 ,000
IT 0.18 in.
-f |
-f |
0.^6 cm)
O.U in.
( 1.1 2 cm)
RAINFALL
§0.85 in. |0.25 in.
2.16 cm) §0.64 cm)
1
I
I
I
I
2400 1200 2400 1200 2
I 16 JUNE I 17 JUNE
TIME, hours and days
1200
18 JUNE
Figure 25. Monitored fecal coliform levels in the Milwaukee River during
wet weather, June 16-18, 1978 (Survey 5).
-------
survey. Instead of flushing out the high coliform levels, the high river
flows continue to bring in significant levels of coliforms from upstream
throughout the duration of the storm. Also, the rainfall through day 2
of the survey contributed to the high concentrations throughout the survey.
Survey 6 - Wet Weather Flow
The final intensive survey was conducted on the 26th through 29th of July,
1978. An average of 0.63 in. (1.60 cm) of rain fell during the period with
the flow in the river peaking at 600 cfs (16.8 nrvsec) . The DO concentra-
tions monitored during this survey are presented in Figure 26 which show?
that there is very little change at a particular site during the three days
of the survey. The trend of decreasing DO as distance downstream increases
is again present during this survey. The DO approaches zero at the
St. Paul Avenue site at the start of the second day of the survey, but
rapidly increases to the 2.0 to 4.0 mg/1 range shortly after.
The fecal coliform concentrations during this survey are graphically
presented in Figure 27- The Port Washington site which is above the CSO
area contains the lowest coliform levels while the St. Paul and Wells
Street sites contain the highest. Very little change in concentrations is
found between the start and end of the survey at a particular site.
CONTINUOUS DO AND TEMPERATURE MONITORING
Data from the continuous DO and temperature monitors located at the North
Avenue Dam, Cherry Street and St. Paul Avenue was utilized to identify the
impact of CSO on DO levels in the Milwaukee River. These monitors measured
changes in DO at mid-channel and mid-depth in the river. From this data it
was possible to examine three major trends in DO:
I. Dry weather variation prior to rainfall events.
2. Decline in DO following rainfall events.
3. Recovery of DO to dry weather levels.
It was also possible to contrast the changes in DO levels to the changes in
river flow following rainfall events.
Figure 28 illustrates the changes in DO levels at St. Paul Avenue, and
changes in flow at Estabrook Park and at the North Avenue Dam during
August 2-7, 1977- This is one of the few periods during the study when
the level recorder at the North Avenue Dam was operating properly. This
figure shows data for a period which also includes a portion of Survey b
(August 3-8, 1977) • During the dry weather period on August 2 and 3, 1977
the DO levels ranged between 2.7 and 6.6 mg/1. Following rainfall, the
DO declined 2.5 mg/1 during a two hour period and then fluctuated around
1.0 mg/1 until the next rainfall event on August *t, 1978. The DO then
remained near 0.0 mg/1 for nearly two days.
As expected the changes in river flow at the North Avenue Dam are slightly
out of phase with the changes at Estabrook Park. There was an initial
rapid increase in the flow rates, and then following rainfall on August 4
-------
0.22 in.
0.22 in.
RAINFALL
u.zz in.» • u./z in.
(0.56 cm)| I (0.56 cm)
PORT WASHINGTON ROAD
NORTH AVENUE
WALNUT STREET
WELLS STREET
ST. PAUL AVENUE
-e-
vo
X
o
Q
Ul
o
00
in
^ 6
10
0
2400 1200
25 JULY
Figure 26.
\\"
2400
1200
26 JULY
2400
1200
27 JULY
2400
1200
28 JULY
2400 1200
29 JULY
TIME, hours and days
Monitored DO levels in the Milwaukee River during wet
weather, July 25-29, 1978 (Survey 6).
-------
100,000
RAINFALL
0.22 in.B • 0.22 in.
(0.56 cm) ...
(0.56 cm)
B
|
PT. WASHINGTON RD.
NORTH AVENUE
10,000
o
o
o
o
DC
o
o
o
,000
2400
I
WALNUT STREET
WELLS STREET
ST. PAUL AVENUE
\
I
I
I
I
1200
26 JULY
2^00
|
1200
27 JULY
2^00
I
1200
28 JULY
2^00
TIME, hours and days
Figure 27- Monitored fecal coliform levels in the Milwaukee River during wet weather,
July 26-29, 1978 (Survey 6).
-------
.DISSOLVED OXYGEN
(NORTH AVENUE DAM FLOW
'ESTABROOK PARK FLOW
• E
z u
C£. -C
^
C
0.4 (1.0)
0.8 (2.0)
01
E
C3
x 9
o i
a
LU
>
s °
to
- 240T
vnn*r.nt IT.TI ts ry.rr.sr. rT^s.f.rs.T»srrs«.
3 —
O i/>
_l 14-
LJ- U
600 (17.0)
400 (11.3)
200 (5.7)
1200
5 AUGUST
2400
TIME, hours and days
Figure 28. St. Paul Avenue'continuous DO monitoring results for August 2-7,
1977- a. Dry weather durations.
b. .Wet weather DO variation.
c. DO recovery, (continued).
51
-------
DISSOLVED OXYGEN
••- NORTH AVENUE DAM FLOW
— ESTABROOK PARK FLOW
u
0)
ui
1/1
M-
U
o
600(17-0)
400(11.3)
200( 5.7)
0
2400
1200
6 AUGUST
2400
1200
7 AUGUST
2400
TIME, hours and days
Figure 28 (continued).
and August 5, 1977 the flow increased to between 400 and 500 cfs (11.3 and
14.2 m3/sec) at both of the flow monitoring sites. At a flow rate of 450 cfs
(12.7 mVsec) the travel time from the North Avenue Dam to St. Paul Avenue
is approximately 20 to 24 hours. The DO levels at St. Paul Avenue begin to
recover approximately 24 hours after the flow at North Avenue reaches the
400 cfs (11.3 cu m/sec) level. The short duration increases in flow on
August 4, 1977 appear to have little effect on DO in the lower river because
the effect of these increases is dampered by the influence of Lake Michigan.
When the flow increases to a constant level, the effect of the CSO is
eventually flushed from the river and the DO at St. Paul Avenue recovers
dramatically to nearly 8.0 mg/1.
Figure 29 illustrates changes in DO at St. Paul Avenue from July 26 to
August 2, 1976. The first portion of this figure shows the dry weather changes
of DO on July 26 and 27, 1976 followed by a gradual decline in DO after a
small rainfall event. The effect of this event is very dramatic because the
previous rainfall event on the CSO area was recorded on June 18, 1976. The
river flow was also at an extremely low level of 85 cfs (2.4 m3/sec) prior
to this storm. On July 30, 1976 there was an extremely large storm of 1.60 in.
52
-------
DISSOLVED OXYGEN
ESTABROOK PARK FLOW
•> E
— 0.4(1 .0)
_0.8(2.0)
en
E
X
O
co 0
CO
I
I
0
0)
I/I
JE
1/1
u-
o
400 (11.3)
200(5.7)
2400
1200
26 JULY
1200
27 JULY •
- E
z u
JC
•^.
•
c.
0.4(1.0)
0.8(2.0)
B
r- 6
en
x
o
o
LU
>
_l
o
CO
CO
O
(U
I/)
i/l
u-
o
O
400(1 I.3)
200(5.7)
0
1200
- 28 JULY-
2400
1200
29 JULY
2400
TIME, hours and days
Figure 29. St. Paul Avenue continuous DO monitoring results for July 26 to
August 2, 1976. a. Dry weather DO variations.
b.&c. Wet weather DO variations.
d. DO recovery. (continued) .
53
-------
•DISSOLVED OXYGEN
•ESTABROOK PARK FLOW
1200 2400
— 30 JULY—\ 4*
1200
31 JULY-
2400
1 AUGUST-
0.4 (1.0)
0.8 (2.0)
o
0)
u
o
1200
00
1200
2 AUGUST
2400
— 400 (11.3)
200 (5.7)
0
1200
TIME, hours and days
Figure 29 (continued).
54
-------
(4.06 cm) of rain which caused a dramatic increase in flow at Estabrook Park
to more than 2000 cfs (56.6 mVsec). This increase does effect the DO at
St. Paul Avenue. The DO recovers for a short period of time due to the
increase in flow and then drops rapidly back to 0.0 mg/1. The recovery of
the DO back to dry weather levels occurs about 2 days later on August 2, 1976.
These were just two examples of the DO impact in the lower Milwaukee River
following rainfall events as recorded with the continuous monitoring devices.
Many other events equally as dramatic were recorded. The magnitude of the
DO impact, duration of the depressed DO, rainfall and flow data was tabulated
for several rainfall events. The analysis of this data is presented in
Section 5.
SEDIMENT STUDIES
The DO conditions occurring within the lower portions of the river during
the monitoring activities of this project have shown a wide range of values
and types of response. In some overflow events the river has exhibited a
dramatic loss of DO to near 0.0 mg/1 and in other events the DO gradually
decreases only a few mg/1 and then recovers. The mechanism of this impact
was thought to be related to the bottom sediments in the lake influenced
portions of the river.
The purpose of the sediment investigations was to characterize these bottom
materials as to their chemical and oxygen demanding potential. Once these
parameters were known, further studies could be conducted to determine the
mechanism of the rapid DO loss during selected overflow events. Following
the identification of the mechanism of this impact, the source of these
materials could be determined and related to the contribution from CSO. The
monitoring of the sediment oxygen demand (SOD) was conducted to provide a
means for the instream water quality model to duplicate the observed changes
in DO during the intensive surveys. Further details of the modeling
of the sediments will be discussed In Section 6, Modeling Studies.
Chemical Analyses
The sediments in the Milwaukee River were characterized during the summer
of 1977 at four locations upstream of the CSO area contribution (Capitol
Drive), three locations within the CSO area but upstream of the North
Avenue Dam, and five locations in the lake influenced portion of the river.
Sediment sampling consisted of collecting a sediment core of 20 in. (51 cm)
in length at mid-channel and a grab sample at each of the quarter points of
the river at each location. The sediment core was split in half in order
to determine differences between the top portion of the sediment and the
lower portions of the sediment. Cores could not be collected above the
North Avenue Dam and at two lower river sites because the sediment deposits
were not sufficiently deep.
Table A-15 in the Appendix lists the results of the chemical analyses
performed on the sediment samples. There was no general trend in the
differences in concentration between the top and bottom portion of the
55
-------
sediment cores. At one location, the top portion of the core had considerably
higher concentrations of all parameters than the bottom portion of the core.
However, at the other two locations where cores were collected, there was
little difference between the top and bottom.
The variability in the concentrations of the parameters observed at the
quarter points of a particular sampling location are likely due to differences
in the water velocity, dredging, and the scouring and deposition effects of
combined sewer outfalls. An excellent example of variability in the deposits
in a cross section of the river is at the junction of the Milwaukee and
Menomonee Rivers (Site I). The middle portion of the channel was dredged
during the summer of 1977- No sample could be obtained from the eastern
quarter point of the river because the substrate was extremely hard clay,
and only a small grab sample could be obtained at mid-channel.
It is difficult to determine whether this lack of soft bottom material was
due to dredging or the scouring effect of the flow from the Menomonee River
entering the Milwaukee River. The western quarter point was located on the
inside corner of the junction of the two rivers, therefore, the sediment
contained more of the flocculant material that is characteristic of most of
the samples from the lake influenced portions of the rivers. The mid-channel
and western quarter point samples differed mainly in the moisture contents
of the sediments and the concentrations of ammonia-nitrogen, total
phosphorus and heavy metals.
Figures 30 to 32 illustrate the general trends in the sediment concentra-
tions of chemical oxygen demand (COD), ammonia-nitrogen (NH3~N) , and lead
(Pb) in the various reaches of the river. Figures A-l to A-6 in the Appendix
illustrate the observed trends in the other parameters. The values shown
represent averages of the quarter point and core samples. In each of these
figures there are obvious differences in the sediment concentration of the
parameters in each reach of the river. Upstream of Capitol Drive, the
ammonia-nitrogen values are approximately 10 percent of those observed
between the North Avenue Dam and Capitol Drive. The physical characteristics
of these reaches of the river are quite similar in terms of river velocity
and sediment deposition rates. The only difference is the contribution of
CSO beginning at Capitol Drive. Downstream of the North Avenue Dam, the
sediment concentrations of ammonia-nitrogen were about twice those observed
between Capitol Drive and the North Avenue Dam. This is likely the result
of the slower river velocities with the subsequent build-up of more sediments.
Also, the percentage of the CSO area contributing to the river increases
almost linearly with distance downstream (see Figure 9). Similar differences
in the sediment concentrations of COD and lead in the different reaches of
the river can be identified in Figures 31 and 32.
Table 6 is a summary of the observed range of all of the measured parameters
in the three reaches of the Milwaukee River. These ranges were developed
from the raw data and not the averages presented in previous figures. For
every parameter, without exception, the observed range below the North Avenue
-------
SILVER SPRING DRIVE
CAPITOL DRIVE
MILWAUKEE RIVER
NORTH AVENUE
Ul
LU
CtL
^-
to
i
ur
WISCONSIN AVENUE
•^
I MENOMONEE RIVER
(LJ
km = ml x 0.62
0 1 2
MILES
KEY
1. RIVER JUNCTION
2. MARINE BANK
3. HIGHLAND BLVD.
4. WALNUT STREET
5. HUMBOLDT AVENUE
6. NORTH AVENUE
7. BURLEIGH STREET
8. HUBBARD PARK
9. LINCOLN PARK - SOUTH
10. LINCOLN PARK - CENTRAL
11. LINCOLN PARK - NORTH
12. DEAN ROAD
Figure 30. Ammonia-nitrogen (mg/kg) values in the sediments of the Milwaukee River.
-------
\n
co
SILVER SPRING DRIVF
WISCONSIN AVENUE
MILES
KEY
1. RIVER JUNCTION
2. MARINE BANK
3. HIGHLAND BLVD.
k. WALNUT STREET
5. HUMBOLDT AVENUE
6. NORTH AVENUE
7- BURLEIGH STREET
8. HUBBARD PARK
9. LINCOLN PARK - SOUTH
10. LINCOLN PARK - CENTRAL
11. LINCOLN PARK - NORTH
12. DEAN ROAD
Figure 31. Chemical oxygen demand (mg/kg) in the sediments of the Milwaukee R
i ver,
-------
SILVER SPRING DRIVE
CAPITOL DRIVE
MILWAUKEE RIVER
NORTH AVENUE
WISCONSIN AVENUE
MENOMONEE RIVER
km B mi x 0.62
0 I 2
KEY
I. RIVER JUNCTION
2. MARINE BANK
3. HIGHLAND BLVD.
k. WALNUT STREET
5- HUMBOLDT AVENUE
6. NORTH AVENUE
7. BURLEIGH STREET
8. HUBBARD PARK
9- LINCOLN PARK - NORTH
10. LINCOLN PARK - CENTRAL
I I. LINCOLN PARK - SOUTH
12. DEAN ROAD
MILES
Figure 32. Lead (mg/kg) values in the sediments of the Milwaukee River.
-------
TABLE 6. SUMMARY OF VARIABILITY OF CHEMICAL PARAMETERS DURING THE
SUMMER OF 1977 IN DIFFERENT REACHES OF THE MILWAUKEE RIVER
North Avenue "~ "
Below North Dam to Capitol Upstream
Avenue Dam Drive Capi to 1 Drive
Volatile total solids (%)' 1.88-5.12 1.16-3-50 0.35-K75
Chemical oxygen demand (mg/kg)2 85,983-380,000 8,820-195,000 577-35,186
Ammonia - N (mg/kg)2 60-743 20-334 9-85
Nitrite + nitrate-N (mg/kg) 0.03-0.85 0.01-0.09 0.01-0.29
2
Total phosphorus (mg/kg) 475-2,840 69-2,006 18-290
Iron (mg/kg)2 11,000-26,300 6,800-25,000 4,300-34,000
2
Cadmium (rng/kg)
2
Zinc (mg/kg)
Lead (mg/kg)2
Copper (mg/kg)
9-37
53-826
34-6,350
21-528
3-16
110-492
49-668
3-141
2-8
21-171
24-239
2-36
2 Values are percent of total sample weight as volatile total solids.
Values are r,ig of measured parameter per kg of total solids.
Dam was considerably higher than in the other two sections of the river. The
range of values for the reach between Capitol Drive and the North Avenue Dam
generally overlapped the range observed downstream, however, the maximum
values observed below the Dam were considerably higher than those upstream.
Upstream of Capitol Drive the sediment concentrations of all the parameters
except iron were considerably less than in the sections of the river receiving
loadings from CSO.
The sediments in the Milwaukee River were also characterized during December
1976. Samples were collected at mid-channel only in the lower portion of
the river. The results of the chemical analyses of these samples is listed
in Table A-16 of the Appendix. These results are quite comparable to
those observed during the summer of 1977. All of the measured parameters
were in the same range during both surveys.
Three of the sediment samples collected during the winter were analyzed
using a centrifuge in order to determine whether the pollutant parameters
were associated with the solids or the interstitial water of the sediments.
Samples collected at three sites on the Milwaukee River exhibit similar
characteristics for centrifuge tests. These characteristics are the
60
-------
transfer properties of various pollutants from the sediments during the
centrifuging process. The transfer properties are similar even though the
moisture contents of the sediments are substantially different as shown in
Table 7.
TABLE 7- MOISTURE CHARACTERISTICS OF MILWAUKEE RIVER SEDIMENTS
Total solids, % Volume of centrate,
Sediment location Uncentrifuged Centrifuged 1 i ters*
Harbor TS77 *»8.5 Ol
St. Paul Avenue 29.1 50.1 I.M»
Holton Street 3^.0 5^.9 0.73
Volume of centrate per kg (dry weight) of centrifuged sediment.
The results in Table 8 indicate that nearly all parameters measured during
the centrifuge tests remained with the solids of the sediments. Only
ammonia-nitrogen appears to be readily soluble and thus transferred to the
centrate. The heavy metals were expected to remain closely associated with
the solids. However, the retention of BOD2Q. COD and TOC with the solids
was not expected as greater amounts of these parameters were thought to be
soluble.
Inspection of the Table 8 values suggest that the sensitivity and accuracy
of some tests may be inadequate for mass balance calculations. For example,
the mass of TOC calculated to exist per kilogram of dry solids following
the centrifugation of sediments exceeds the mass calculated to exist prior
to centrifugation. In other words, the quantity of TOC contained in
centrifuged sediments and centrate exceeds the quantity of TOC contained
in the uncentrifuged sediments. Such generation of mass is not possible
and is attributed to test procedure limitations. Regardless of such
inconsistencies, the points made in the previous paragraphs remain valid
due to the extremely small amounts of mass transferred to the centrate
relative to the mass in the uncentrifuged sediments.
Sett!ing Tests
Settling tests were conducted on Milwaukee River sediments suspended in
river water in January, 1977 and February, 1978. The sediment samples in
1977 were from Wisconsin Avenue and the 1978 samples were collected at the
junction of the Menomonee and Milwaukee Rivers. Settling tests were run on
four different sediment/water concentrations as listed in Tables 9 and 10.
The settling curves for the sediments suspended in river water in 1977
(Appendix Figure A-7) indicates at least two phases of settling occur.
Suspended solids concentrations rapidly decrease by at least an order of
magnitude in the first hour of settling for the range of conditions conducted
during the laboratory tests. Suspended sediment concentrations continued
to decrease, but at steadily decreasing rates for the next 10 to 15
61
-------
TABLE 8. SUMMARY OF MASS BALANCE CALCULATIONS FOR MILWAUKEE RIVER SEDIMENTS
PRIOR TO AND FOLLOWING CENTRIFUGATION
Weiqht of constituent (ma) oer ka dry wt.
Harbor Sediments
Parameter
Total solids
Volatile solids
BOI)20
M COD
TOC
Aimonia-N
Total -P
Iron
Cadmium
Lead
Copper
Zinc
Un-
eentrifuged
1,000,000
I9*,900
47,900
260,800
80,800
4, 200
6,470
24,000
39
285
165
705
Centrifuged
996,900
143,600
19,400
687,160
223,000
2,690
7,780
-
-
-
-
-
Cent rate
3.070
857
656
487
173
1,180
1.7
21.6
0.4
0.4
0.3
<0. 1
St.
Un-
eentrlfuged
1.000,000
118,000
16,100
160,000
61,900
269
1.560
18.000
10
W5
III
410
. of sediments
Paul Sediments
Centrifuged
998,800
137.800
61,800
r, 009, 800
260.500
40
14,680
-
-
-
—
-
Centrate
1,160
307
144
180
82
137
0.8
12.4
<0.l
<0.l
<0.l
<0.l
Hoi ton Sediments
Un-
centrifuged
1,000,000
97,000
20,000
152,000
64,700
590
1,530
18,000
II
685
112
455
Centrifuged
999,300
110,900
16,800
226,800
252,100
470
8.090
-
-
-
-
-
Centrate
630
138
22
45
23
45
<0.l
2.3
<0.l
<0.l
<0.l
<0.l
-------
hours. Settling of solids appeared to have ceased beyond 15 hours as solids
concentrations are generally no longer decreasing.
TABLE 9. SETTLING CHARACTERISTICS OF MILWAUKEE RIVER SEDIMENTS FROM
WISCONSIN AVENUE SUSPENDED IN RIVER WATER, JANUARY, I9771
Sediment
mixture^,
ml
25
50
75
100
Percent
of initial suspended sed
iments having
velocity (V) (in./hr) or greater
20
92
94
97
89
10
95
95
98
98
5
97
96
99
99
settl ing
2
98
97
99
99
cm = 0.39 in.
Estimated from settling plots.
2
Volume of sediment added to river water for a total volume of one liter.
TABLE 10. SETTLING CHARACTERISTICS OF SEDIMENTS FROM JUNCTION OF
MILWAUKEE AND MENOMONEE RIVERS SUSPENDED IN RIVER WATER, FEBRUARY, I9781
Sediment
mixture^,
ml
25
50
75
100
Percent
20
87
86
90
82
of initial suspended sediments having settling
velocity (V) (in./hr) or greater
10
88
92
93
94
5
93
94
95
96
2.5
94
.97
98
38
1
98
99
99
99
cm = 0.39 in.
Estimated from settling plots.
2
Volume of sediment added to river water for a total volume of one liter.
These observations were illustrated by settling velocities calculated for
each of the sediment and river water mixtures. In general, in excess of 90
percent of the sediments had settling velocities greater than 20 in./hr
(51 cm/hr) (Table 9). The amount of suspended sediment that remained in
suspension ranged from I to 3 percent of the original amount of sediments.
The settling tests were repeated in early 1978 using identical procedures.
The settling curves for the settling tests conducted in 1978 also indicated
at least two phases of settling (Figure A-8 in the Appendix). Similar to
the 1977 test, rapid settling occurred for at least the first hour.
Continually decreasing rates of settling occurred after approximately four
hours of settling. However, contrary to the 1977 tests, settling of solids
was continuing after 40 hours.
63
-------
The settling rates of solids were slightly less for the 1978 tests compared
with the 1977 settling tests. In general, greater than 80 percent of the
sediments suspended in river water had settling rates exceeding 20 in./hr
(51 cm/hr) for the 1978 tests (Table 10) compared with 90 percent for the
1977 test (Table 9).
Despite the slightly slower settling rates for the 1978 tests, the total
percentage of solids settling within the test period was greater for the
1978 tests than for those conducted in 1977- Only I to 2 percent of the
solids remained in suspension after 40 hours of settling during the 1978
tests. The differences in these settling test results are minor and are
likely due to the different sediment sampling locations for the two tests.
Scouring velocities can be estimated from the settling data through two
approaches presented by Fair, Geyer and Okun (19)- One approach is to
use the equation for scouring velocity based upon settling velocity:
Vd = (8/f)l/2Vs
Where;
Vd = Scouring velocity.
Vs = Settling velocity.
f = Friction factor ranging from 0.02 to 0.04.
Using a settling velocity of 20 in./hr (51 cm/hr) the range of scouring
velocity is 0.006 to 0.009 ft/sec (0.18 to 0.27 cm/sec).
A second estimate of scouring velocity is based upon the following equation
(19):
Vd = 1(8 k/f) g (s-1) d]l/2
Whe re:
k = 0.04 for sand and > 0.06 for sticky material.
g = 32.2 ft/sec2 (9.8 m/sec2).
s = Specific gravity of material.
d = Diameter of particle being settled.
To estimate an equivalent particle diameter, Stoke's Law is assumed to
apply (i.e., laminar flow conditions and the settling particle is a sphere).
Using a settling velocity of 20 in./hr (51 cm/hr) and specific gravity of
particles of 1.05, the equivalent sphere diameter is approximately 0.01 cm.
The particle with this diameter would have a scouring velocity of 0.08
ft/sec (2.A cm/sec).
In summary, the scouring velocity range is approximately 0.01 to O.I ft/sec
(0.30 to 3.0 cm/sec) for the particle having settling velocity of 20 in./hr
(51 cm/hr). This range is calculated assuming the particle is nearly
spherical in shape, that unhindered settling occurs and that laminar flow
64
-------
conditions prevail. Irregular shaped particles would have lower settling
velocities, which would result in lower scouring velocities than previously
estimated. Thus, at least 10 to 20 percent of the Milwaukee River sediments
have scouring velocities lower than O.I ft/sec (3.0 cm/sec). During summer
conditions the scouring velocity is probably somewhat faster than indicated
by these tests. The sediment samples for these tests were obtained during
the winter months when the river flow was less than 100 cfs (2.8 mVsec) and
the river velocities were quite slow. Because of these slow velocities,
the sediments probably contained a higher percentage of the particles with
low scouring velocities than found during higher flow conditions in the
spring and summer. These tests still provide a relative indication of
the velocity required to scour sediments in the lower Milwaukee River.
Sediment Oxygen Demand Tests
In order to determine the significance of the sediments on the dissolved
oxygen balance in the Milwaukee River, both in situ measurements and
laboratory testing was conducted. The in situ technique for measuring
sediment oxygen demand (SOD) rates was developed by Lucas and Thomas (20).
The respirometer consisted of a plexiglass chamber with a DO probe attached
inside and a small 12 v submersible pump attached outside for circulation
of water (Figure 33). The SOD rates were determined from the decline in
DO over time after the chamber was sealed on the sediment surface. The
in situ SOD rate was calculated using the following formula:
SOD = (Ci-Cf)V
tA
Where:
2
SOD = Sediment uptake rate in gm 02/m -day.
V - Volume of confined water in m3.
A = Bottom area within chamber in m .
t = Test period in days.
Cj = Initial measured DO in chamber in mg/1.
Cf = Final measured DO in chamber in mg/1.
This formula provides the SOD rate on an areal basis and can be converted
to a volumetric rate in the river by division by the mean depth at the
sampling location.
The SOD rate was determined twice at five locations in the lower Milwaukee
River and once at three locations upstream of the North Avenue Dam. The
measured SOD rates are listed in Table II. These rates have been corrected
for temperature differences using a Van't Hoff-Arrhenius expression as
determined by McDonnell and Hall (21). The two rate measurements conducted
at each location were generally quite close despite being taken at intervals
up to one and one-half months apart. There was a substantial difference
in the observations only at Humboldt Avenue. This may have been due to the
variability in the sediment depth at this site which is in close proximity
to the North Avenue Dam. Except for the Walnut Street site, there appears
to be a general trend of increasing SOD with distance downstream. The SOD
65
-------
-igure 33- Photograph of sediment oxygen demand respirometer for measuring
in situ SOD rates.
\RIATION OF
Location
River Junction
St. Paul Avenue
Highland Blvd.
Walnut Street
Humboldt Avenue
Burleigh Street
Lincoln Park
Dean Road
IN SITU SEDIMENT
MILWAUKEE RIVER
Distance upstream',
km
0.0
0.3
•
2.3
3.<<
5.6
11.8
20.9
OXYGEN
Date
7/8/77
8/1/77
6/23/77
8/1/77
6/27/77
8/1/77
6/27/77
8/1/77
6/27/77
an/n
9/U/77
7/IV77
7/IV77
DEMAND RA1
SOD
gm/m2-day
6.6
6.7
5.3
5.5
3.7
M
6.1
6.7
3-7
1.8
*.5
3.0
0.0
The distance upstream Is listed from the junction of the Milwaukee
and Menomonee Rivers.
66
-------
rates measured at Walnut Street were in the same range as those observed at
the junction of the Menomonee and Milwaukee Rivers.
Laboratory measurements of the undisturbed and disturbed SOD rates in the
river were conducted in a plexiglass cylinder lined with aluminum foil to
avoid light penetration. The cylinder was 8.5 in. (21 cm) in diameter and
4.2 ft (1.3 m) in height. A two liter sediment sample was placed in the
bottom of the cylinder and overlaid with 21 liters of river water. This
system was allowed to settle and then was aerated with compressed air in
order to saturate the water with oxygen. An air stone was utilized to
diffuse the air to avoid disturbance of the sediments. DO readings
were taken intermittently at mid-depth in the cylinder with a portable
DO probe and meter. This continued for a five day period while the system
remained quiescent. The decline in DO during this period was used as a
verification of the SOD rates determined with the in situ technique.
Following the measurement of the undisturbed SOD, the system was aerated
again. The decline in DO was then measured as the sediments were
disturbed by circulating the river water with a centrifugal pump. The SOD
rates measured under these conditions were utilized as a good approxima-
tion of the demand exerted in the Milwaukee River when the sediments are
scoured.
Sediment samples for laboratory SOD rate determination were collected in
each of the three rivers entering Lake Michigan and the inner harbor as shown
in Figure 34 and Table 12. The undisturbed rates measured in the laboratory
are considerably smaller than those measured in situ. This is likely due to
oxidation of the sediments during sampling, storage, and set up of the
laboratory system. The aeration of the system is also a likely cause of
the reduced undisturbed values.
The two sites in or near the inner harbor (Sites 2 and 3) had significantly
larger disturbed SOD values than the other sites on the three rivers. The
lower values obtained further upstream in each of the three rivers were
quite similar. All three rivers contribute to the sediments in the inner
harbor, therefore, it is feasible that the largest disturbed SOD values
were obtained from the sample collected in this area. There did not
seem to be any relationship between the undisturbed SOD and the disturbed
SOD. The site with the second largest undisturbed SOD (Site 4) had one
of the smallest disturbed SOD rates.
In order to determine whether the density of the sediments affected the
SOD measured, the rates were also calculated on a weight basis (gm 02/qm
sediment-day). Since all of the sediment samples were approximately
the same density and of similar moisture content, this change in units
had little effect on the previously mentioned trends in the SOD values
in the rivers. The areal expression of SOD values is the conventional
form for these rates.
-------
HAMPTON AVENUE
GREENFIELD AVENUE
I
L
A
K
E
M
I
C
H
I
G
A
N
KINNICKINNIC RIVER
1
MILES
km = mi x 0.62
Figure 3^- Sediment sampling locations for bench scale sediment oxygen
demand characteristics.
68
-------
TABLE 12. BENCH SCALE DETERMINATIONS OF SEDIMENT OXYGEN DEMAND
UNDER UNDISTURBED AND DISTURBED CONDITIONS1.
Map location
no.
1
2
3
k
5
6
Description
Kinnickinnic River
at Fi rst Street
Mooring Basin
in Inner Harbor
Milwaukee River
at RR Bridge
Menomonee River
at Great Lakes Coal
Milwaukee River
at Highland Avenue
Mi Iwaukee River
at Hubbard Park
Undisturbed SOD,
gm/m2-day
0.65
1 .40
2.10
1.70
1 .kO
0.33
Disturbed SOD,
gm/m2-day
430
1,370
800
270
360
66
The laboratory sediment oxygen demand determinations were carried
out at a temperature of 20 ^0.5 C.
The results of
(as determined
the sediment oxygen demand
by in situ measurements)
tests indicate that the SOD rates
in the lower Milwaukee River range
between 1.8 and 6 . 7 gm/mZ-day and the average SOD rate for this portion of the
river was 5-0 gm/mz-day. The laboratory measurements of the SOD rates indicate
that the oxygen demand of the sediments in the rivers and inner harbor in
Milwaukee may exceed 1000 gm/m2-day when the sediments are disturbed or agitated.
This represents more than a 100 fold increase in the SOD rate. The significance
of these m. situ and disturbed SOD rates on the DO balance of the Milwaukee River
is discussed in Section 5 of this report.
69
-------
SECTION 5
ANALYSIS AND DISCUSSION
The field data collected as part of this project was used in this section to
define the source and the mechanism of the change in water quality following
overflow events. A methodology was developed to predict the DO impact and
input to a receiving water model. Therefore, the relationships between the
various elements and sources of the water quality impact could be evaluated
in the following sections of this report.
HYDRAULIC STUDIES
In addition to providing data on travel times within the river during
different flow conditions, the results of the hydraulic studies have further
implications. The most significant finding was that the velocities in the
lower portions of the river were not fast enough to cause the scouring of
the bottom sediments during low flow conditions. During river flows of
approximately 200 cfs (5-7 m3/sec), the measured velocity between Walnut
Street and St. Paul Avenue was just 0.06 ft/sec (0.018 m/sec). This velocity
was in the range where slight sediment scouring could occur as determined
with the sediment settling tests. However, the scouring velocities determined
in the laboratory must be considered rough estimates and the velocities near
the bottom sediments were much lower than the average river velocity. The
results of the velocity monitoring near St. Paul Avenue indicated that
velocities were not only quite low near the sediments, but the direction was
quite variable. Some of the velocities measured near the river bottom during
flows of 100 cfs (2.8 m3/sec) and flows reaching ^50 cfs (12.7 m3/sec) were
sufficient to cause some sediment scouring. However, the oscillating nature
of the direction of the measured velocities indicated that there was little
net downstream velocity near the sediments.
The trends in river velocity are important to the analyses which follow since
they indicate that the scour of bottom sediments in these portions of the
river does not occur primarily as a function of instream flows. The
continuous DO recorders in these reaches have shown that rapid instream
changes in flow have little effect on the downstream DO conditions. This fact
was tested in early Spring of 1977, when the flood control dam above
Estabrook Park was opened to allow the spring snowmelt flows to be released
to the downstream reaches. Although the flows in the river increased by
more than 300 cfs (8.5 cu m/sec), the DO record downstream indicated no
change as a function of sediment scour. This is reasonable since the
estimated change in the mean velocity in the lower river due to this change
in flow is only 0.10 ft/sec (0.03 m/sec). The increase in the velocity
near the sediments would be much smaller.
70
-------
A second finding of the hydraulic studies that has provided insight into
the mechanism of the water quality impacts during wet weather is the
dispersion characteristics of the CSO outfalls. The dye injection at a CSO
that occurred during low CSO discharge flows tended to remain at the top
portion of the instream water column and the dye also remained on the same
side of the river that the outfall was situated. This characteristic is
important to remember since the submerged outfalls do not cause rapid mixing
actions within the river during low CSO discharge flow conditions. When
the outfall was flowing at higher velocities, then the dye plume was well
dispersed across the river almost instantaneously after the injection.
This rapid dispersion was thought to be the mechanism through which the
bottom materials were scoured when the instream velocities were insufficient
to affect this change. The data on velocities in the river generated from
CSO discharges support this theory. This will be discussed in more detail
in relation to the sediment investigations.
SEDIMENT INVESTIGATION RESULTS
The data generated within the sediment investigations portion of this project
are the most important to deciphering the contribution of CSO to the DO
impacts of the Milwaukee River. The first finding of note in these investi-
gations is the vast differences in the upstream versus downstream portions
of the river. For example, the ammonia-nitrogen concentrations of the
sediments above the CSO area are significantly less than the concentrations
in the CSO influenced portions of the river. As the distance downstream
increases, the area of CSO contributing to the river also increases.
Because of this, the ammonia-nitrogen concentrations below the North Avenue
Dam were approximately double the concentration in the CSO influenced area
above the dam. This is an indication of the influence of CSO on the bottom
materials from a chemical characteristic basis. The same trend can be
found in the SOD rate measurements taken in these areas. The general trend
of these measurements was an increase in SOD with distance downstream.
The data must be examined to determine the significance of SOD on DO levels
during dry weather and during wet weather conditions. Even during the dry
weather survey (Survey I) the DO in the lower portions of the river, partic-
ularly at St. Paul Avenue, approaches the 2.0 mg/1 DO standard. This was
partially due to the low flow conditions during this survey when the average
flow was y» cfs (2.1 m3/sec). The significance of the dry weather or un-
disturbed SOD can be illustrated with the following hypothetical case. For
example, if the river flow is 100 cfs (2.8 m3/sec) and the SODjs taken as an
average of the measured values, the SOD can account for a decline in DO
of approximately k.k mg/1 between the North Avenue Dam and the junction of
the Milwaukee and Menomonee Rivers. Calculated on a volumetric basis, this
means a rate of 0.04 mg 02/l-hr in the lower portions of the river. Like-
wise, if the ultimate BOD in the river is assumed to be 5.0 mg/1 as
estimated with data from the dry weather survey, the soluble load can
account for a decline in DO between the same locations of approximately
2.1 mg/1. This is assuming a BOD reaction rate coefficient of O.I per day
as utilized in the receiving water model for this portion of the river.
The volumetric rate of this soluble demand is approximately 0.02 mg 02/1-hr.
This indicates that the SOD is twice as significant during dry weather as
71
-------
the soluble demand.
The relationship of SOD to the DO conditions within the river was not
fully understood until the laboratory studies of the undisturbed and
disturbed demand were completed. These studies indicated that the potential
exists in the sediments to cause the rapid decline in DO that had been
observed with the continuous DO monitors. In addition, although the
downstream reaches were low in DO during dry weather, they still remained
above the 2.0 mg/1 standard for these areas. If SOD was the source of
the rapid DO sags, then the undisturbed demand could not account for this
mechanism. Implementation of the undisturbed SOD rates into the receiving
water quality model only served to lower the DO values in the river in both
dry and wet weather conditions. The mechanism by which the sediments were
disturbed and how this mechanism could be implemented in the water quality
model led to the investigation of velocities at the bottom of the river
during wet weather events.
The scour of bottom materials within the river by discharges from submerged
CSO outfalls was verified with the velocity studies and with visual observa-
tions. The velocities measured in the vicinity of the submerged outfalls
were extremely large and capable of scouring considerable quantities of
sediment. The effect of the scouring can be illustrated with the rainfall,
DO and CSO velocity data for a storm occurring on August 3, 1977- Figure 35
depicts the response of the river at St. Paul Avenue in terms of DO to this
rainfall event and the subsequent scouring of sediments due to CSO discharges.
An instream CSO velocity of nearly 12.0 ft/sec (3'.7 m/sec) was measured
1.0 ft (0.3 m) above the sediment surface at a distance of approximately 30 ft
(9-1 m) from a submerged CSO outfall at St. Paul Avenue. This velocity was
measured just after discharges from this CSO outfall began and the direction
of the velocity was perpendicular to the river. After discharge from the
outfall ceased, the river velocity was variable in direction and very small
in magnitude. The rapid decline of DO shown in Figure 35 following the
rainfall event is the result of the scouring or agitation of bottom sediments
throughout the lower Milwaukee River from discharges from submerged CSO
outfalls. This CSO velocity measurement, as well as other measurements
indicate that sediment scouring from submerged CSO discharges does occur
in the lower portion of the river.
Visual observations in the lower river have also indicated the significance
of sediment scouring. During severe storm events, the plumes from the
numerous outfalls and the resulting scour of sediments were observed. The
overall significance of scouring from CSO discharges is evident when one
considers that there are approximately *»0 outfalls discharging into the
Milwaukee River below the North Avenue Dam. Although some of these outfalls
have irregular shapes, the majority are circular and range from 36 in.
(0.9 m) to 96 in. (2.7 m) in diameter. Also most of them are submerged.
Averaging the number of outfalls evenly over the length of this portion of
the river, approximately one CSO outfall exists every 300 ft (95 m) of
river. Considering that many of the outfalls enter the river in pairs
from opposite banks, and considering frequency of the outfall occurrence,
the scouring potential from these outfalls can be quite significant.
72
-------
m/sec = ft/sec x 0.30
cm = in. x 2.5^
0.6 L—
8
6
3 AUGUST
2400
TIME, hours and days
Figure 35- The observed decline in DO, instream velocity from a submerged
CSO outfall, and rainfall volume at St. Paul Avenue in the lower Milwaukee
River on August 3-k, 1977.
73
-------
PREDICTION OF THE DO IMPACT
Due to the severe impact which sediment scouring has on DO levels following
CSO events, the data collected during the study had to be analyzed in order
to develop a consistant methodology to adjust SOD parameters for the
calibration of Harper's water quality model. For this methodology an
empirical expression was developed utilizing a forward (stepwise) multiple
regression program. Data on DO impact from the continuous monitoring
devices, rainfall and river flow were used for this analysis. Two different
parameters were utilized as the dependent variables. The first was the
change in DO (ADO) expressed in gm-02/m2 at a continuous monitoring site
following a rainfall event. The areal expression for ADO was utilized so
that it could be easily related to the SOD rate which also was expressed
on an areal basis. The other parameter was the duration of the DO decline
(T) in hours.
Only the data on DO from the St. Paul Avenue continuous monitoring site
was utilized for this analysis for two major reasons:
I. The most severe impact on DO levels was generally observed at
this site.
2. If the impact could be predicted at this site, the decline in
DO at other sites could be predicted by assuming that'the magnitude
of the wet weather demand increased with distance downstream from
the North Avenue Dam.
The major assumption in the development of this procedure for predicting
the DO impact is that the majority of the observed DO decline is due to
sediment scouring. Some of the decline in DO is due to the organic loading
from CSO, storm sewers and upstream loads. However, the rates of DO decline
far exceed the rates which could be due to the measured 6005 and TOC
concentrations in the river. Previous modeling studies have also been
unable to account for the large wet weather demand (II, 12). In order
to minimize correlating the decline in DO due to the soluble oxygen demanding
materials, only the rapid decline portion of the DO curves were utilized in
determining the ADO values. Figure 36 illustrates the methodology used to
obtain these values.
Several independent variables were selected for the regression analysis.
These variables were as follows:
I. Rainfal1 volume.
2. Average rainfall intensity.
3- Peak rainfall intensity.
k. CSO volume.
5. Peak river flow.
6. Change in river flow.
7- Antecedent dry period.
8. Natural log of the antecedent dry period.
-------
2400
1200
1200
TIME, hours
Figure 36. 3, _.
duration values from continuous
Methodology utilized to obtain A DO and
f-™, r.— *: D0 records at St. Paul Avenue.
The rainfall data used was from the Broadway Street precipitation gauge
(Figure 2) which is located in the drainage basin tributary to the lower por-
tions of the river. The CSO volume was estimated using the calibrated STORM
and the antecedent dry period was defined as the time since the previous
rainfall event with a volume greater than 0.10 inches (0.25 cm).
Only four of the eight parameters were entered into the multiple regression
equations developed with the forward regression program for the predictions
of ADO or T. Table 13 lists the sequence of variables added in building
the regression equation. The first variable entered was the CSO volume
which accounts for the largest amount of variance in the dependent variable
for a one variable model. This variable accounted for approximately 67
percent of the variance in the ADO data as indicated by the coefficient
of determination, R^. This result is reasonable since the change in DO due
to sediment scouring should be highly related to the volume of CSO discharged,
The second variable entered into the equation was the natural log of the
antecedent dry period. This two variable model accounted for about 76
percent of the variance in the dependent data. The rainfall volume was the
75
-------
TABLE 13. RESULTS OF THE FORWARD (STEPWISE) REGRESSION
ANALYSIS WITH ADO AS THE DEPENDENT VARIABLE
Independent
No. variables variables
in equation entered
1 CSO volume
2 CSO volume
Loge dry period
3* CSO volume
Loge dry period
Ratnfal 1 volume
4 CSO volume
Log dry period
Rainfall volume
Dry period
Variable
Intercept coefficient
Bo B,
7.62
-3-49
-0.90
-4.65
26.22
27.56
2.30
76.95
2.77
82.99
4.07
-39.15
0.01
Coefficient of
determination
R2 Prob > F
0.673 0.0.01
0.761 0.001
0.810 0.001
0.820 0.001
"Selected model for prediction of ADO.
TABLE 14. RESULTS
ANALYSIS
Independent
No. variables variables
in equation entered
1 Rainfall volume
2 Rainfall volume
Dry period
3* Rainfall volume
Dry period
CSO volume
4 Rainfall volume
Dry period
CSO volume
Loge dry period
OF THE FORWARD (STEPWISE) REGRESSION
WITH T AS THE DEPENDENT VARIABLE
intercept
Bo
15-58
7-89
5-90
13.70
Variable
coefficient
Bi
23-33
27.68
0.02
43-57
-23-37
0.02
53-54
-36.70
0.03
-2.44
Coefficient of
determination
R2 Prob > F
0.287 0.022
0.425 0.016
0.428 0.045
0.^34 0.095
*Selected model for prediction of T.
-------
third parameter added to the equation and the three variable model accounts
for 81 percent of the variance in ADO. The four variable model accounted
for 82 percent of the variance with the fourth variable being the antecedent
dry period without the log conversion. No other variables met the 0.5
significance level for entry into the model. The values of the regression
F-test (Probe > F column) indicate that the null hypothesis (that there is
no correlation) can be rejected at the 99 percent confidence level for each
of the models listed.
The three variable model was selected in order to predict ADO for two
reasons:
I. The four variable model accounted for little additional variance
in the dependent variable.
2. It did not seem practical to include the antecedent dry period
and the natural log of the same data in the same predictive equation.
Prediction of the duration of the DO decline (T) was not as successful as
for ADO. The duration is more difficult to predict due to changes in
river flow and the influx of water from Lake Michigan. Table 14 lists the
results of the regression analysis. In this case, the rainfall volume was
the first variable entered into the model accounting for only 29 percent
of the variance in the T data. The antecedent dry period was the second
parameter added to the model. This two variable model accounted for 42
percent of the variance in T. The addition of the CSO volume and the
natural log of the antecedent dry period explained little additional
variance in the dependent variable. The four variable model was eliminated
from consideration for predictive purposes because the null hypothesis
could not be rejected at the 95 percent confidence level. The three variable
model was the selected model because it did meet this criteria.
From the regression analysis, it was possible to predict the magnitude and
the duration of the DO impact due to sediment scouring. The incorporation
of this methodology into the receiving water quality model is discussed
in Section 6.
CONTRIBUTION OF CSO
In order to fully evaluate the water quality impact of CSO, the pollutant
contribution to the Milwaukee River from CSO was identified. In addition,
the evaluation of pollutant loadings to the river also required considera-
tion of the impact of CSO on the sediments, since both monitoring and
modeling studies indicated that sediments are a major water quality influence
in the Milwaukee River.
The pollutant loadings from CSO were first compared to the pollutant load
entering the study area at the USGS gauging station. The CSO loadings were
estimated with STORM while the upstream loadings were estimated from monthly
sampling data and average monthly river flows from USGS (22). The 1977
water year (October, 1976 through September 1977) was evaluated since
77
-------
complete water quality data for the Estabrook Park site was available from
USGS. The estimated annual loadings of selected pollutants is presented
in Table 15-
TABLE 15. ESTIMATED ANNUAL POLLUTANT LOADINGS FOR THE 1977 WATER YEAR
Source
Upstream^
CSO area
Suspended
sol ids ,
IbxIO^
1420
810
Carbonaceous
BOD1.
IbxIO^
314
263
Total
ni trogen,
IbxIO2*
89
22
Total
phosphorus,
1 bx 1 04
7
9
Carbonaceous BOD is estimated as 1.5 times BOD5 values for sanitary sources
including CSO.
Loadings represent values for the Milwaukee River upstream of the CSO
area contributions. kg = 2.2xlb
Calculation of the annual pollutant loadings which contributed to the sediments
was more difficult and various assumptions were required throughout the
evaluation. An attempt was made to be consistent in all calculations and to
provide a conservative or worst case estimate of the loadings. Therefore,
it must be emphasized that the estimates are an approximation of the
sediment loadings which are consistent with the other results of this study.
Suspended solids contributions to the sediments were calculated from results
of 48-hour settleabi 1 i ty tests on river water samples and composited CSO
samples. The settleable percentage was assumed to represent the fraction
of the suspended solids load that would contribute to the sediments in the
study area of the Milwaukee River. The settling tests indicated that an
average of 50 percent of the solids settled from ten river samples while
an average of 98 percent of the solids settled from five CSO samples.
Therefore, estimated suspended solids loadings to the sediment (Table 16)
indicate that the upstream portion of the river and CSO contribute
approximately equal proportions of suspended solids to the sediments.
TABLE 16. ESTIMATED LOADINGS TO THE SEDIMENTS IN THE MILWAUKEE RIVER
STUDY AREA FOR THE 1977 WATER YEAR
Pollutant
Upstream
CSO area
In situ
productivity
Suspended solids,
710
790
Carbonaceous BOD,
lbx!04
72
210
118-249
kg = 2.2xlb.
78
-------
The loadings of oxygen demanding material were estimated in terms of
carbonaceous BOD (CBOD) . The relationships of soluble and insoluble BODc;
and settling solids were used to estimate the BOD contributions to the
sediments (Table 16). Total and soluble BODij values were determined in river
water and flow composited CSO samples. An average of 80 percent of the
BOD in the CSO samples was associated with suspended material while about
25 percent of BOD in the river water was solids associated. The data and
settling results established final CBOD loadings. This approach may over-
estimate the upstream contribution of BOD to the sediments since no data
is available to establish whether the suspended BOD is associated with the
settleable or non-settleable suspended solids.
Phytoplankton and macrophytic plants also contribute significant amounts
of oxygen demanding materials to the sediments. On the basis of studies
conducted by Bothwell (23) and field productivity measurements, the BOD
contributed to the sediments from phytoplankton within the study area of
the Milwaukee River has been estimated (2^). It was assumed that all
production reached the sediments, however, contributions of upstream
sloughed plant material and phytoplankton were not added. The range of
calculated BOD contribution from in situ productivity is also presented
in Table 16.
These calculations have shown that roughly kQ to 50 percent of the annual
loadings of oxygen demanding materials (in terms of CBOD) to the Milwaukee
River are due to CSO. These data are consistent with the results of the
chemical analyses of sediment samples, which indicate that the chemical
composition of the sediments changed abruptly where the river enters the
CSO area. However, since there are significant assumptions involved in
the calculations, the values should be considered rough estimates with a
wide confidence range. The numbers can be used to compare the loadings
contributions from the different pollutant sources.
INSTREAM MONITORING RESULTS
The analysis of the data from the intensive surveys conducted during this
project can be related to the findings of the sediment investigations for
the DO results and to the dye studies for the fecal coliform results. Thus,
low intensity rainfalls that cause overflows which do not result in wide-
spread scour of the bottom materials will cause only a gradual decline in
DO conditions in the lower reaches of the Milwaukee River. High intensity
storms with long periods of dry weather prior to the event can cause the
rapid loss of DO that is common to the lower reaches of the river. Further
details of the individual monitoring events will be discussed in Section-7,
Evaluation of CSO Impact.
The fecal coliform data obtained during the intensive monitoring surveys
is not affected by the sediment investigations or the mechanisms explained
in the preceding sections. However, the dye studies allow a better insight
into the relation of coliform concentration and river flow. Of course, the
number of coliforms discharged for each event is a function of the rainfall
and prestorm history. But the length of time that the coliform standards
79
-------
are violated is a function of how fast the river flow is capable of flushing
these organisms out of the lower portions of the river. The dye studies
have indicated that the lake influence is minimal during the high flow
conditions within the river. This accounts for the rapid decrease in
coliform levels after the overflows and upstream contributions to the lower
reaches of the river have ceased. Implementation of the mixing characteris-
tics of the CSO outfalls into the coliform routines of the instream model
could not be accomplished in this project. The instream sampling techniques
near CSO outfalls was modified to insure that the entire cross section of
the river was sampled and not just one location which could be influenced by
a nearby outfall. The contribution of CSO to the coliform levels observed
in the monitoring program is discussed in Section 6.
80
-------
SECTION 6
MODELING
STORM MODEL
The Corp of Engineers Storage Treatment Overflow Runoff Model (STORM) was
used in this study to simulate the existing sewer system (25). STORM
is a "pipeless" model which does not route flow and quality through an
actual sewer network but rather estimates flow and quality from a specified
watershed by calculating total surface runoff and adding this flow to the
watershed's total dry weather flow. Surface runoff and the associated pollu-
tant washoff from the watershed are calculated by the model on an hourly
basis, using a continuous hourly precipitation record, available surface
pollutants and watershed characteristics (types of land uses within the
watershed). A treatment rate, representing the flow diverted to the sewage
treatment facilities, is applied to the total estimated flow (dry weather
flow plus surface runoff) and any flow exceeding the treatment rate is
considered overflow. An overflow event begins with the first hour in
which overflow occurs and ends with the hour in which no overflow occurs.
A rain storm lasting several hours could cause one or several overflow
events, depending on the treatment rate, total rainfall for each hourly
segment and watershed characteristics.
Each hour of overflow has an associated pollutant load, which depends on
the surface pollutants available at the beginning of each hour and the
hourly surface runoff rate. The pollutant loads simulated with the STORM
model for this study were suspended solids, carbonaceous BOD (CBOD) and
fecal coliform. CBOD was assumed to be approximately 1.5 times the BODc for
sanitary sources including CSO. The CBOD for separate storm sewer discharges
was assumed to be slightly more than the 8005 for modeling purposes.
Rainfall
The results of the STORM simulations presented in this report were obtained
by using continuous hourly precipitation data obtained from a tape supplied
by the Southeastern Wisconsin Regional Planning Commission (SEWRPC). This
tape contained 38 years of rainfall data for the period 19^0 to
1977. The rain data was collected and recorded by the National Weather
Service Office at Mitchell Field in Milwaukee.
Precipitation records from city raingages were used in lieu of the SEWRPC
tape for the calibration process for simulations involving individual
outfalls. The response of each outfall's drainage area to a particular
81
-------
storm event could best be defined using the city raingage located within
each drainage area. A valid comparison between monitored data and model
simulation results could then be made. Therefore, calibration was performed
using the raingauge in the closest proximity to the outfall being evaluated.
Calibrat ion
Before using STORM to simulate various sewer systems analyzed
in this study, certain variables within the model had to be adjusted or
calibrated so that the model output reflected characteristics which are
specific to the Milwaukee area. The calibration procedure consisted of
the following steps:
I. Collecting and tabulating monitoring data of CSO's in the
MiIwaukee area.
2. Applying STORM to the representative drainage area and
rainfalls for which monitoring data exists.
3. Comparing output of the model with actual monitoring data.
A. Adjusting certain parameters in STORM until the output from the
model comes close to actual measured values.
The monitoring data available for the calibration was from the following
CSO sites:
27th Street on the Kinnickinnic River
Burleigh Street on the Milwaukee River
Edgewood Street on the Milwaukee River
Humboldt Avenue Detention Tank on the Milwaukee River.
These sites are shown in Figure 37.
Because STORM is a "pipeless" model and uses a single treatment rate to
characterize a drainage area, a treatment rate for combined sewer overflow
based on existing sewer system diversion structure capacities was impossible.
The problem involves trying to characterize a sewer system which has dynamic
flow dividers and diversion structures with a single treatment rate. A
treatment rate equal to the peak dry weather flow was, therefore, used in
simulating existing CSO conditions. This treatment rate was calculated using
the formula for dry weather flow provided in the STORM manual (25). The
calculation of dry weather flow is based on area size, population, land use
types within a watershed, and fnfi1tration. Population within each water-
shed modeled was obtained from Milwaukee census daJia and acreages used
were derived by design engineers for the CSO system and from City of
Milwaukee sewer maps. The land use types within a watershed were obtained
from zoning maps. Five categories of land use types are commonly used
and these include:
82
-------
oo
00
SILVER SPRING DRIV
MENOMONEE I
X
KEY
1. BROWN DEER ROAD SITE
2. SILVER SPRING SITE
3. 2VTH STREET STORM SEWER
k. 2VTH PLACE STORM SEWER
5. 54TH STREET STORM SEWER
6. PORT WASHINGTON ROAD SITE
7. KERN PARK CSO
8. EDGEWOOD CSO
9. AUER AVE. CSO
10. BURLEIGH STREET CSO
11. NORTH AVENUE DAM SITE
12. HUMBOLDT CSO
13. WALNUT SITE
l*t. WELLS SITE
15. ST. PAUL SITE
16. 27TH STREET CSO SITE
1 . t
MILES
km = mi x 0.62
Figure 37- CSO,storm sewer and river monitoring locations on the Milwaukee River, Lincoln Creek,
and Kinnickinnic River.
-------
I. Single family residential.
2. Multiple family residential.
3. Commercial.
k. Industrial .
5. Parks and open areas.
The infiltration rate provided in the STORM User's Manual (25) was used.
The calibration of STORM for the quantity of CSO for given rainfall events
was not accomplished using monitored flows for several reasons. First,
only "rough" flow measurements with stage boards were taken during monitoring.
Secondly, modeling a single outfall drainage system with STORM is nearly
impossible because STORM is a "pipeless" model. Nearly all of the CSO sewer
networks are interconnected, making accurate predictions for a single out-
fall nearly impossible with STORM. Finally, it was not possible to monitor
all of the outfalls with interconnected sewer networks. Therefore, STORM
was calibrated for CSO quantity by comparing the predicted values tor total
overflow from STORM with values predicted by SWMM. This model was calibrated
for the CSO drainage area as part of the Milwaukee Combined Sewer Overflow
Pollution Abatement Project (II). Predictions from the two models for
individual outfalls and groups of outfalls for several storms were compared.
The runoff coefficients in STORM were adjusted so that the predicted over-
flow volumes were reasonable for several storm events.
Table 17 lists the predicted CSO volumes for individual CSO outfalls for a
storm event on September \k, 1968. This storm was selected for comparison
because it is a long duration event with variable rainfall intensities
which produced 1.2 inches of rain in seven hours. STORM predictions of
overflow volume were high for the Kern Park outfall and low for the Auer
Avenue and Burleigh Street outfalls. However, the predictions seem reasonable,
TABLE 17. COMPARISON OF TOTAL FLOW PREDICTED BY THE SWMM AND STORM
MODELS FOR THE SEPTEMBER 24, 1968 STORM.
Location
Kern Park outfal 1
Auer Avenue outfall
Burleigh Street outfall
Total flow,
ft3xlo6
SWMM STORM
0.35 0.41
0.87 0.60
1.49 1.28
Storm sewer areas tributary to the Milwaukee River upstream of the CSO
area contribution were also simulated with STORM. Figure 10 in Section 3
depicts the CSO and storm sewer areas simulated for input to Harper's
water quality model.
When STORM was used to simulate flow and quality from the storm sewers a
treatment rate was applied which was very close to zero. By using this
-------
treatment rate, essentially all surface runoff and its associated pollutant
load reach the receiving waters. The same values for the variables used to
control quantity of surface runoff in combined sewer simulations were
used in storm sewer simulations. In addition to treatment rate, another
hydrological difference in simulating storm sewers is the lack of dry
weather flow. The dry weather flow block of STORM is used only in simulating
combined sewer overflows.
STORM was set up using the dust and dirt method of pollutant accumulation
which assumes that all pollutants are associated with dust and dirt accumula-
tion in streets (lbs/100 ft of gutter length/day). A different dust and
dirt accumulation rate is assigned to each land use category. Each dust and
dirt accumulation rate has associated pollutant fractions, i.e., the fraction
of suspended solids, BOD and fecal coliforms associated with each 100 Ibs
of dust and di rt.
In addition to adjusting the variables controlling the dust and dirt method
of pollutant accumulation, the variables controlling the dry weather flow
block of STORM were also used to calibrate the combined sewer overflow
pollutant load predictions. The concentrations of pollutants in dry weather
flow were obtained from the literature (22)(2?). The model was calibrated
to simulate dry weather flow which has a concentration of approximately
200 mg/1 BOD, 250 mg/1 suspended solids and >IO°/IOO ml fecal coliforms.
After these values for dry weather flow were obtained, the dust and dirt
variables for surface runoff were then adjusted (calibrated) to obtain the
desired overflow concentrations.
STORM was calibrated for quality using estimates of the composite quality
concentrations for BODtj and suspended solids using the discrete field sample
measurements and the relative discharge rates based on stage height. Fecal
coliform counts were measured for discrete field samples and these provided
a range to which the predicted composite fecal coliform counts could be
compared.
The results of the quality calibration for CSO are presented in Tables 18
and 19- A wide variation in pollutant concentration exists for the monitored
data. On a composite basis, monitored suspended solids ranged from 113 to 950
mg/1 and 8005 ranged from 25 to 265 mg/1. These variations are due to
differences in prestorm history of pollutant accumulation (number of dry
days), rainfall volume and intensity, and the watershed characteristics of
individual outfalls. It is not surprising that measured and modeled data
do not always show a close fit. The model cannot be expected to predict
accurately for each individual outfall on a per event basis using a single
set of calibration variables. The model was calibrated to simulate over-
flow for the entire basin and should give reasonable predictions on an
annual basis. On a yearly basis the model was calibrated for concentrations
of approximately *tOO mg/1 suspended solids, 130 mg/1 CBOD and >IO°/IOO ml
fecal coliform. These values fall within the range of the monitored data
(Tables 18 and 19) and reported literature values (Table 20).
85
-------
TABLE 18. STORM CALIBRATION RESULTS FOR SUSPENDED SOLIDS AND BOD COMPOSITE
CONCENTRATIONS FOR CSO DISCHARGES
CSO
Outfall
Edgewood
Burleigh
Humboldt
27th and
KK
Date
7/6/77
7/17/77
8/4/77
7/23/75
5/31/77
6/17/77
6/24/77
6/28/77
7/6/77
7/17/77
8/4/77
9/12/72
7/17/77
7/11/75
7/23/75
SSj
Measured
210
122
113
145
950
524
393
418
405
297
281
228
525
540
265
mg/1
Modeled
70
247
70
42^
267
109
304
101
370
238
304
152
191
308
Measured (BODc) ,
mg/1
55
-
25
50
265
170
103
26
79
-
60
63
-
105
4l
. Modeled (CBOD)3,
ma/1
89
150
100
158
126
56
79
81
70
83
124
50
222
235
Assumed to be I.5 x BOD5 for sanitary sources including CSO.
TABLE 19- STORM CALIBRATION RESULTS FOR FECAL COL I FORMS FOR CSO DISCHARGES
Outfall
Burleigh
Date
1977
5/31
6/17
6/24
6/28
Fecal col i form,
Modeled
(composite)
6600
1800
2900
2100
[counts/100 ml] x 103
Measured
(discrete range)
18-650
75-12,000
3-TNTC3
15-600
Too numerous to count.
86
-------
CO
•vl
TABLE 20. LITERATURE VALUES FOR COMBINED WASTEWATER AND SEPARATED
STORM FLOW DISCRETE QUALITY
Combined wastewater
Location
Washington, DC (28)
Detroit (29)
Detroit (29)
Detroit (30)
BOD5,
mg/1
Range Mean
10-470 71
50
153
Suspended sol
mg/1
Range
35-2,000
23-1398
ids,
Mean
622
250
27*
150
Fecal col
[counts/100
Range
2*0-5,0*0
200-1 7, 000a
iform,
ml] x I03
Mean
2,*00
Separated storm flow
Washington, DC (28)
Detroit (29)
Cincinnati (29)
Ann Arbor (29)
Detroit (30)
3-90 19
1*7
19
28
130-11 ,280
900-2,062
1,697
1*7
210
2,080
*0-l ,300
8-l,ll5a
310
aMedian values from several samples taken during a particular flow event.
-------
The quality of storm sewer flow predicted by STORM is solely a function of
surface runoff quality which is determined by the variables controlling the
dust and dirt method of calculating pollutant accumulation. These variables
were adjusted during calibration. The model was calibrated for storm sewer
concentrations of approximately 100 mg/1 suspended solids, 30 mg/1 CBOD
and >I03/IOO ml fecal coliform. The CBOD and fecal coliform values are
consistent with the monitored data (Table 21) and literature values (Table 20)
However, the calibrated suspended solids concentrations are higher than
the measured values (Table 21). This is because all the monitored storm
sewers are in single family residential areas, which have lower suspended
solids concentrations than commercial or industrial locations. This is
supported by the fact that storm sewer suspended sol ids concentrations
reported in the literature were considerably higher. Therefore, a suspended
solids concentration of 100 mg/1 was used to characterize the area being
modeled by STORM.
TABLE 21. CONCENTRATIONS FOR MONITORED STORM SEWER SAMPLES
Location
54th Street
Date
1977
6/17
6/28
111
7/17
Suspended sol ids ,
mg/I (composite)
3k
1 1
31
3k
BOD5,
mg/1 (composite)
1 1
14
14
-
Fecal col i forms ,
[counts/100 ml]x!03
(discrete range)
1-70
3-18
24th Street
7/7
7/17
19
60
\k
:4th Place
111
7/17
18
45
15
RECEIVING WATER QUALITY MODEL
General Assumptions
The water quality model modified for this project is based on a model used
in previous studies on the Milwaukee River (I I). Similar to its predecessor,
the modified model is a series of equations applied to the section of the
Milwaukee River extending from Brown Deer Road to the confluence of the
Milwaukee River with the Menomonee River as shown in Figure 37- The modeled
section of the river is broken into several segments or reaches. The
characteristics of these reaches are assumed to be constant within each
88
-------
reach for a certain time period. Characteristics may change from reach to
reach or within the same reach after a certain period of time has passed.
The characteristics include physical and biological variables such as
sediment oxygen demand, reaeration and deoxygenation rate coefficients, mean
velocity and average water depth.
The water quality model is a one-dimensional, time-varying model. Complete
vertical and lateral mixing of all water quality parameters is assumed.
Thus, variations of dissolved oxygen or temperature, with depth, which have
been observed in the Milwaukee River, are not simulated. Similarly,
variations in water quality parameters across the river, perpendicular to the
direction of flow, are also not modeled. The model calculates changes in
water quality parameters along the length of the river, or longitudinally,
as a function of time.
Hydraulic Equations
Hydraulic data needed by the water quality model must be developed separately
and input to the model. This data includes the hydraulic characteristics
of the river reaches;
Mean width.
Depth.
Discharge.
When modeling unsteady flow conditions, these characteristics may change
during the simulation. This requires a new set of hydraulic data cards for
each flow condition. The hydraulic characteristics are held .constant within
each reach, however, until changed by a new set of hydraulic data cards.
Routines are programmed in the water quality model for linear interpolation
of hydraulic data. This interpolation is necessary to minimize numerical
errors caused by immediate changes of hydraulic conditions at the
boundaries of river reaches and when hydraulic conditions change for a
particular reach. If not corrected, numerical errors could result in
simulations in which large amounts of pollutants are either gained or lost
when, in fact, no change is desired.
Each river reach is divided into several units called computational elements
for solving the equations of the water quality model. Three hydraulic factors
may vary within the river reaches and are determined for each computation
element. These factors are inflow from Lake Michigan, overflow volumes from
combined sewers and inflow volumes from surface runoff.
Inflow from Lake Michigan is considered a uniformly Increasing discharge
beginning at some point downstream from the North Avenue Dam to a specified
value at the confluence of the Milwaukee and Menomonee Rivers (Figure 38).
The basis for the development of this relationship was trends in salt
water intrusion in estuaries. Also, monitoring data on specific conductance
and total dissolved solids during dry weather collected for a previous
project (II) are generally in agreement with this relationship. Evaluation
89
-------
of the rate of inflow from Lake Michigan is discussed later in this section
of the report.
2.0 _
60
o
o
r 20
1.5
o
<1> 1
t/> I .
— 0.5
1000
2000
meters
3000
4000
8000
4000
feet
DISTANCE
120 0
NORTH
AVENUE
DAM
Figure 38. 1.1 lustration of assumptions utilized in Harper's water
quality model for simulation of the Lake Michigan inflow.
Overflow volumes from combined sewers and inflow volumes from surface runoff
are calculated from data supplied by either SWMM or STORM. The water quality
model is capable of receiving inflow data from several storms and is presently
programmed for twenty storms. This restriction may easily be increased by
increasing storage in the appropriate statements in the program. Another
restriction is that the number of inflows (the sum of combined sewer outfalls
and storm runoff discharge points) must remain constant for the simulation
period and is limited to 80 in number.
Water Quality Equations
The water quality model is a series of equations which are linked to
calculate changes in several water quality parameters. These water quality
parameters are any conservative pollutant (i.e., total dissolved solids),
temperature, phytoplankton, biochemical oxygen demand, dissolved oxygen
and fecal coliform bacteria. Equations for growth, respiration and
sloughing of benthic algae are also included in the model.
90
-------
Basic Transport Equation--
The basic equation used to simulate transport of all materials in the
Milwaukee River is referred to as the convecti ve-dispersion equation:
3C_ +..9£ = D 92C ± S(x,t)
Where:
C = Water quality parameter concentration.
t = Time.
V = Stream velocity.
DL = Coefficient of longitudinal dispersion.
x = Longitudinal distance.
S = Source-sink term for the water quality parameter.
This equation describes the rate of change of a water quality parameter
based on the assumptions of complete lateral and vertical mixing as
previously discussed. For conservative parameters, such as total dissolved
solids, the source-sink term equals zero at all times and locations. The
source-sink term varies for each parameter according to the following
discussion.
Phytoplankton and Benthic Algae Source-Sink Equations —
The growth and respiration rates of phytoplankton (floating and suspended algae)
are assumed to apply for the growth and respiration of benthic (attached to
the river bottom) algae. Adjustments are made for the light available for
growth of benthic algae and for sloughing of benthic algae from the stream
bottom. The following discussion, however, basically applies to both
phytoplankton and benthic algae, even though only phytoplankton are mentioned.
The equation used to estimate the changes in phytoplankton concentrations is
a first order linear differential relationship:
(x,t)=dP = (U-R-Set)P 2
Where:
*[* »^* > w / "._'
P dt
P = Phytoplankton concentration, mg dry weight algae/1.
U = Phytoplankton growth rate, sec~l.
Set = Phytoplankton settling rate, sec~'.
R = Phytoplankton respiration rate, sec"'.
S (x,t) = Net phytoplankton growth, mg dry weight algae/1-sec.
Mathematical models have usually treated phytoplankton growth rate as a
function of several nutrients, light and temperature. The nutrients have
included nitrate, ammonia, phosphorus and carbon. These models have equations
which reduce the growth rate of phytoplankton from some maximum rate to a
lower rate depending upon the modeled concentrations of nutrients and
relative optimum levels required by algae. Light available for growth is
91
-------
normally treated in a manner similar to nutrients.
Existing nutrient concentrations in the Milwaukee River are normally higher
than concentrations usually considered limiting to algal growth. Consequently,
phytoplankton growth is assumed to be limited by available light and the
growth rate equation may be expressed as:
U = U (lave) 3
max
(K. + I )
I ave
Where:
Umax = Maximum growth rate, a function of temperature, sec"'.
K| - Half saturation constant for light, Langley sec~'.
'ave = Avera9e light intensity through the water column,
Langley sec"'.
Average light intensity through the water column is used in the growth rate
equation because phytoplankton are assumed to be uniformly distributed
through the water column consistent with the mixing assumptions. The average
light intensity (lave) can be shown to be related to the incident light
intensity (lo) by the relationship:
I =!/, -kh\ .
ave o (1 -e ; H
kH
Where;
I0 = Incident solar radiation, Langleys sec"'.
H = Mean stream depth, ft.
k = kb + (Kpnyto) (P) = extinction coefficient, ft"'.
kfc = Base extinction coefficient, ft"'.
Kphyto = Coefficient for variable extinction coefficient (mg/l-ft)"'.
The variable extinction coefficient is intended to simulate the self-shading
of algae. Through this mechanism, high phytoplankton concentrations will
absorb most of the light in the water column thereby reducing light available
for growth to levels which limit phytoplankton growth.
Both maximum growth rate, Umax and respiration rate, R, are assumed to be
functions of temperature. Linear relationships are used for both variables
based on data from literature (31) which are illustrated irr Figure 39.
The phytoplankton settling rate is a variable which is estimated for each
river reach. Establishment of this rate is part of the model calibration
procedure. It is noted that the net growth (U-R) P is used in calculation
of dissolved oxygen concentrations. In the sequence of computations,
phytoplankton growth and respiration occur prior to settling. This sequence
of calculations was found to be necessary during model calibration to obtain
92
-------
-
0.05
32
TEMPERATURE, F
60 80
10 20
TEMPERATURE, °C
30
100
Figure 39. Influence of temperature on the maximum growth rate of
Phytoplankton (U^) and phytoplanktor, endogenous respiration rate (3|)
93
-------
observed diurnal ranges in dissolved oxygen while maintaining acceptable
levels of phytoplankton.
The kinetics of benthic algal growth are similar to phytoplankton growth
kinetics. Major differences are:
I. Benthic algae losses are modeled by sloughing off rather than by
settling.
2. Light reaching the channel bottom is used in the growth rate
expression rather than average light intensity through the water
column.
A change in units from the phytoplankton growth equation (equation 2) is
necessary as benthic algae are expressed on a dry weight per unit area
basis rather than a dry weight per unit volume basis. That is:
Sb (x,t) =d£= (U-R) B 5
dt
Where:
B = Benthic algae concentration, mg/m2.
S[j (x,t) = Net benthic algae growth, mg/m2-sec.
Sloughing of benthic algae is simulated by restricting the maximum allowable
density of benthic algae. Thus, net growth may occur when the density is
at the maximum level, but the density will not increase. This maximum
level is presently programmed as a single value specified by the user for
the entire river being modeled.
Dissolved Oxygen Source-Sink Equations—
The sources and sinks of oxygen considered to be most significant in the
mathematical model are:
I. Reaeration.
2. Oxygen consumption by bottom deposits and suspended and dissolved
ma te r i a 1.
3. Algal production and respiration of oxygen.
Reaeration is assumed to be a first order process where the rate of transfer
across the air-water interface is proportional to the dissolved oxygen
deficit. The dissolved oxygen deficit is the difference between the actual
concentration and the dissolved oxygen concentration for saturation conditions.
If super-saturation conditions occur, the transfer of oxygen is from the
water column to the atmosphere. If undersaturation conditions occur, or if
a dissolved oxygen deficit exists, oxygen is transferred from the atmosphere
to the water column. This process may be expressed by:
d£ • k2 (Cs - C)
dt
-------
Where:
C =
Dissolved oxygen concentration, mg/1 .
Reaeration coefficient (a function of temperature) sec"'.
Saturation dissolved oxygen (a function of temperature), mg/1
Cs = 14.652 - 0.4I022T + 0.00799IT2 - 0.000077774T3,
Where:
T = Stream temperature, °C.
The reaeration coefficient may be estimated from several published experi-
mental equations. The O'Connor-Dobbins equation is used in this model (32).
Isotropic turbulence conditions or conditions when the Chezy coefficient is
greater than 20 must be assumed. This equation has been found to be generally
adequate for most river applications. Problems in using the equation have
been experienced for rivers having average depths less than one foot. In
these cases, the applicability of the equation depends on the time step used
to solve the equations of the model.
The reaeration coefficient equation for 20°C is:
3/2
Whe re :
m
V
Molecular diffusivity of oxygen at 20°C , ft2/sec.
Mean stream velocity, f ps .
The reaeration rate coefficient is corrected using a Van't Hoff-Arrhenius
type expression:
Where;
k2(T) and k2(20) = Reaeration coefficients at T°C and 20°C ,
respectively.
0| = A constant ranging from I. 01 to 1.047.
T = Stream temperature, °C.
The sediment oxygen demand (or benthal oxygen demand) is estimated differently
for dry weather conditions and wet weather conditions. This difference is
due to the immediately higher oxygen demand observed to occur following
combined sewer overflows. This higher demand cannot be solely attributed to
biologically controlled reactions which was explained in Section k and 5.
95
-------
Sediment oxygen demand during dry weather conditions is calculated by an
equation which sets the benthic demand constant until the dissolved oxygen
in the overlying water limits oxygen uptake (21):
.. _ . r0.3 n (T-20)
dC_=l<3c 92 10
dt fi
Where:
= Consumption rate coefficient, mg/l-sec.
= Dissolved oxygen concentration of overlying water, mg/1.
0i = Constant for temperature correction ranging from 1.067 to
1.078.
The coefficient k, is a function of the type, depth and age of deposited
material and of the types of organisms decomposing the organic fraction. A
Van't Hoff-Arrhenius equation is used to adjust the oxygen consumption as
the temperature changes.
The wet weather sediment oxygen demand is modeled as a time-varying demand.
This demand is initially calculated as a multiple of the dry weather demand
to approximate the immediately high demand required at the time a combined
sewer overflow event begins. The high demand is reduced at an exponential
rate to the dry weather sediment oxygen demand. The magnitude of the
increase in sediment oxygen demand and rate of decrease are specified by the
model user for each CSO event. The equation describing the time-varying
sediment oxygen demand is:
SOD wet weather = (SOD dry weather) (I + ae" l) II
Where;
a and b are constants which may be varied for each CSO
event and t is the simulation time after the CSO event
has begun.
The constants a and b are intended to reflect the range in wet weather SOD
rates. For example, a long dry weather period preceeding a CSO event would
probably cause a larger increase in wet weather SOD than a shorter dry
period due to more time :for production of gases and reduced chemicals in
the sediments under anaerobic conditions. Similarly, a more severe CSO
event would probably cause a larger wet weather SOD due to more disturbance
of the sediments. Thus, for these instances, the value of the "a" constant
would be larger than for shorter dry weather periods or less severe CSO events,
The "b" constant, which regulates the rate at which the wet weather SOD is
reduced to the dry weather SOD, may be similarly varient according to the
nature of the CSO event. The curves shown in Figure AO illustrate the
characteristics of the time varying SOD equation.
96
-------
USE OF CURVES:
bT (dtmensionless)
SUPPOSE
AND
THEN
AND
10 when T = 0
1 .5* when T =
-2T
(i.e. initial
1 .5 hours
SOD
hr
(1
+ lOe *')
-1
b = 0.5 hr
-1
I
y/x
-bT
2 3 *» 5
TIME AFTER CSO BEGINS (hours)
1 + ae (dimensionless)
Figure 40. Characteristics of the time varying sediment oxygen demand
-------
The water quality model also has the flexibility to restrict the operation of
the wet weather SOD equation to certain reaches. A river location may be
specified above which the wet weather SOD equation should not apply. More-
over, the length of time that the wet weather SOD equation applies may be
specified .
These factors in addition to the constants a and b are input to the model
for each combined sewer overflow event. Analysis of continuously monitored
dissolved oxygen conditions yields estimates of these factors for initial
model calibration simulations.
Oxygen consumption by suspended and dissolved material is assumed to be the
typical first order process:
,2
dt
Where:
k| = Deoxygenat ion constant, day"'.
L = Ultimate biochemical oxygen demand, mg/1.
83 = Constant for temperature correction approximately equal to
1.047.
The uptake and release of dissolved oxygen by phytoplankton and benthic
algae are assumed to be proportional to the net growth of phytoplankton and
benthic algae:
d£ = UQ_ (U-R) (P+B) ................................................. 13
dt H"
Where ;
C = Oxygen concentration, mg/1.
Uoz = Oxygen uptake rate coefficient, mg/mg.
U, R, P, B and H are as previously defined.
Adding all of the terms together yields the composite dissolved oxygen
source-sink equation:
-k.L67 (T"20) - Un (U-R) (P+B) ....................................... 14
13 uz. H
Temperature Source-Sink Equations--
In the preceeding sections, several rate coefficients have been given or
developed as functions of temperature, including the maximum algal growth
rate, algal respiration rate, saturation dissolved oxygen concentration
and the Van't Hof f-Arrhenius type temperature corrections in the oxygen
98
-------
source-sink equation. Because of these interactions, water temperature
is simultaneously modeled with the other water quality parameters.
One of the simplest methods available for temperature modeling is the heat
balance approach published by Raphael (33)- This procedure involves
estimation of the net heat transfer across the water surface to calculate
the change in temperature of the waterbody. The equation is:
ST-£-^ is
Where;
-2 -1
O_T = Net heat transfer across the water surface, Btu ft sec .
Cp = Specific heat of water, Btu lb~|0F~'.
Y = Unit weight of water, 62.4 Ibs ft"^.
ST = Rate of change in temperature, °F sec"1.
The net heat transfer consists of several components, each of which is
estimated independently. The individual components are incident solar
radiation (Qj), conductive heat transfer (QH), effective back radiation (QB),
and evaporative heat transfer (0_E) • Incident solar radiation may be
calculated theoretically using the solar altitude, the latitude of the site
and the cloud cover. Direct measurement of incident solar radiation is
also possible.
Evaporative heat transfer equations have been developed from Lake Hefner
evaporation data. Wind speed and vapor pressure differences are the major
variables used in the equation:
QE = KE U (ew - ea) 16
Where:
Kg = Experimental constant.
0 = Wind speed, knots.
e^, - Vapor pressure of water in saturated air at the temperature
of the water surface, inches of mercury.
ea = Ambient vapor pressure, inches of mercury.
The conductive heat transfer is estimated using the Bowen ratio of conducted
heat to energy utilized by evaporation, which leads to:
Qn = K UP (T - tj 17
™ n a
Where;
KH = Experimental constant
u = experimental constant.
P = Atmospheric pressure, inches of mercury.
t = Air temperature, °F.
d
99
-------
Effective back radiation has been defined as the difference between the long
wave radiation leaving a body of water and the long wave radiation from the
atmosphere being absorbed by the body of water. Radiation from both of these
bodies can be calculated using the Stefan-Boltzmann radiation law. Radiation
from the water surface is a function of the temperature of the water body
and the emissivity of the water surface. Atmospheric radiation is dependent
upon the air temperature and the moisture of the air. Energy transfer by
both mechanisms has been combined to yield the effective back radiation as:
OB'0'97*
-------
Whe re:
Umax = Maximum fecal coliform growth rate, sec
L = Biochemical oxygen demand, mg/1.
K-s = Half-saturation constant, mg/1.
KC = Die-off rate coefficient, sec"'.
Example of Temperature and Dissolved Oxygen Equations
The temperature and dissolved oxygen equations are sufficient to model
extremely complex aquatic systems. Some understanding of the response of the
equations may be gained by application of the equations to a simplified river
system. The results of such an application are illustrated in Figure k\.
Two simplified river systems are illustrated in Figure 41. The first
consists of a stream which receives a constant discharge of oxygen demanding
materials. The second system is a stream which supports algae; phytoplankton
or benthic algae. In both instances, stream temperature varies diurnally
with maximum temperatures occurring during afternoons and minimum tempera-
tures occurring during earlier morning hours (Figures ^l-a and Al-d).
For the system receiving the constant discharge, the impact of the discharge
can be evaluated by examination of dissolved oxygen conditions upstream
and downstream from the discharge. At the upstream station (Figure b\ -b),
dissolved oxygen concentrations are at saturation values, calculated
according to equations 6 and 7. The dissolved oxygen curve is a mirror
image of the temperature curve at the station. This is because cooler
water has a greater capacity for dissolved oxygen than warmer water. There-
fore, minimum dissolved oxygen concentrations occur at maximum temperatures
and maximum dissolved oxygen concentrations occur at minimum temperatures.
The mean dissolved oxygen (illustrated by the dashed line in Figure 4l-b)ls
the saturation concentration of the mean temperature (illustrated by dashed
line in Figure Al-a). The diurnal range in dissolved oxygen (A.) is caused
solely by temperature variations.
Downstream from the location where the discharge enters the river (Figure
k\-c) the entire dissolved oxygen curve has been moved downward by an. average
amount (A£) . This decrease is caused by the exertion of oxygen demanding
materials. These materials may include dissolved or suspended biochemical
oxygen demand from material discharged or released to the river (equation 12),
or may include a constant demand from the sediments such as during dry
weather (equation 10).
Phytoplankton and benthic algae will cause significant diurnal variations
of dissolved oxygen directly in phase with water temperature variations
(compare Figures k\-d and Al-f). This similarity is due to the importance of
incident solar radiation in the heat budget equation (equation 15) and
in the algal growth rate equation (equation 3).
The impact of wet weather discharges may also be discussed in terms of
these examples. A large discharge of oxygen demanding materials to the
101
-------
WITH CONSTANT
OXYGEN DEMAND
01
LU
D-
1200
00 1200
UPSTREAM STATION
X
o
Q
LU
>
o
CO
1200
2*tOO
T200
cs
x
O
o
LU
>
_l
O
CO
CO
DOWNSTREAM STATION
1200 2400 1200
TIME, hours
WITH ALGAL PHOTOSYNTHESIS
AND RESPIRATION
\
1200 2400 1200
UPSTREAM STATION
1200 2400
1200
— DOWNSTREAM STATION
I
1200 2^00
TIME, hours
1200
Figure 41. Example of temperature and dissolved oxygen
simulations with Harper's water quality model.
102
-------
river would cause a temporary effect similar to constant oxygen demand. The
dissolved oxygen concentrations would be depressed below the concentrations
shown for the downstream stations in both examples. The depression would
occur for a short period of time and conditions would recover to the dry
weather conditions. The large discharge of oxygen demanding materials may
occur from combined sewer overflows or disturbance of sediments. The
large demand may be caused by biological and/or chemical factors.
Coefficients in equation II are adjusted as part of model calibration to
simulate the observed conditions.
Understanding these relationships provides the basis for model calibration
and verification. Rate coefficients in the equations may be initially
estimated based on previous model applications. The water quality model
can be more easily calibrated and verified, however, if the nature and
importance of the various components of the equations are understood.
Receiving Model Set Up
Harper's water quality model was set up to simulate a l*» mi (22.6 km) por-
tion of the Milwaukee River extending from Brown Deer Road to the confluence
with the Menomonee River. For this study, calibration of the model was
for sites downstream from Port Washington Road. The area upstream of
the site was included in the set up in order to accurately simulate
loadings from storm sewer areas tributary to the Milwaukee River and
Lincoln Creek.
The river was segmented into 100 units, each being approximately 7*»0 ft
(225 m) in length. Input requirements for describing the river include
the following:
I. Distance or length of river simulated.
2. Flow rate.
3. Mean width of segment.
A. Mean depth of segment.
The mean width and depth did not have to be specified for each of the 100
segments. Rather, these dimensions are specified at the upstream boundary
and several locations downstream. Dimensions of the segments in between
these locations are calculated by the computer program using linear inter-
polation. River dimensions in the Milwaukee River were specified at
eight locations downstream of Port Washington Road and three locations
upstream. Less detail was required upstream since this area was mainly
utilized to input storm sewer loads. Specified dimensions in the model
can be changed during the simulation as the flow rate changes. This
provides the flexibility for accurate simulations of velocity in the river.
Other model input requirements for a given simulation include the
following;
I. Upstream quality concentrations.
Conservative parameter - usually total dissolved solids.
Temperature.
103
-------
Phytoplankton.
Biochemical oxygen demand.
Dissolved oxygen.
Fecal coliforms.
Benthic algae.
2. Initial quality conditions - several locations.
Parameters same as above.
3. Lake Michigan influx flow and quality.
Parameters same as above.
4. Rate coefficients.
Base extinction coefficient for light.
Sediment oxygen coefficient.
BOD resuspension coefficient.
Coliform decay.
Phytoplankton settling coefficient.
5. Weather data.
Solar radiation.
Wind speed.
Atmospheric pressure.
Cloud cover.
Dew point temperature.
Air temperature.
6. Phytoplankton parameters.
Self shading factor.
Half saturation constants.
Uptake rate coefficients.
7. Combined sewer and storm sewer loadings.
Input time.
Flow rate.
Conservative - usually suspended solids.
Temperature.
Biochemical oxygen demand.
Fecal coliforms.
8. CSO scour coefficients.
Distance where coefficients apply.
Magnitude coefficient.
Decline coefficient.
Decline time.
9. Size of time step.
The upstream quality is specified at time intervals throughout the simula-
tion. This is also true of the weather input data. The CSO loadings are
specified for the appropriate input time. Restrictions on the input time
are:
-------
I. That there must be an upstream quality specified at this time.
2. The input time must be a multiple of the time step.
The CSO scour coefficients are specified for each CSO loading. AM other
inputs remain constant throughout the simulation. Sensitivity analyses
on the size of the time step were conducted to determine how large the time
step could be without causing large computational errors. Figure A2
presents the results of this analysis for a one and four hour time step.
These results indicate that computational errors do not get extremely large
even when a time step of four hours is utilized. A three hour time step
was selected for modeling purposes. This allows suitable flexibility for
inputting CSO loads and for the length of simulations.
Receiving Model Calibration/Verification
Calibration/verification of Harper's model involved several steps:
I. Calibration of the water quality model for dry weather/steady
flow conditions.
2. Development of a time varying SOD rate for simulation of
sediment scouring and the subsequent release of oxygen demanding
materials.
3. Development of a predictive procedure for the determination of time
varying SOD coefficients based upon rainfall/runoff characteristics.
4. Calibration of the water quality model for wet weather/unsteady
flow conditions.
5- Verification of the water quality model for wet weather conditions.
The development of the time varying SOD equations has been discussed
previously in the discussion of Harper's water quality model. The develop-
ment of a predictive methodology for determination of the coefficients
of the time varying SOD equation was discussed in Section 5. The remaining
topics will be discussed in the following paragraphs.
Dry Weather Calibration—
Harper's model was originally calibrated for dry weather as part of the
Milwaukee Combined Sewer Overflow Abatement Project (I I) utilizing data
from an intensive dry weather survey in June, 1975- Data collected during
dry weather in September, 1976 (Survey I) for this project has been utilized
to verify the initial calibration of the model, as well as determine an
acceptable range for certain parameters for wet weather calibration. The
September, 1976 survey is for that portion of the river in the CSO area
(downstream of Capitol Drive).
Figures k3 and M and Appendix Figures A-9 and A-IO illustrate the verifica-
tion results for dissolved oxygen at sites in the CSO area portion of the
Milwaukee River for Survey I. The range in DO shown for the observed
105
-------
x
o
3.5
3.0
9, 2.5
E
2.0
1.5
o
to
?. 1.0
Q
0.5
0.0
ONE HOUR TIME STEP
FOUR HOUR TIME STEP
1200
21 SEPTEMBER
2400 1200
22 SEPTEMBER
ME, hours and days
2^00 1200
23 SEPTEMBER
Figure 42. Time step sensitivity analysis for St. Paul Avenue, September 21-23, 1976 (Survey 1)
-------
MONITORED RANGE
PREDICTED
3.5
3.0
1 2.5
z
LlJ
CS
£ 2.0
o
a
UJ
i K5
5 1.0
0.5
0
I
2400 1200
21 SEPTEMBER
2400
1200
22 SEPTEMBER
2400
1200
23 SEPTEMBER
TIME, hours and days
Figure 43. Dry weather verification results for St. Paul Avenue, September 21-23, 1976
(Survey I).
-------
MONITORED RANGE
PREDICTED
o
GO
x
o
o
10
_ I
a
2400
I
1
1200
21 SEPTEMBER
2400
1200
22 SEPTEMBER
TIME, hours and days
2400
I
1200
23 SEPTEMBER
Hgure V*. Dry weather verification results for Walnut Street, September 21-23, 1976
(Survey |).
-------
values illustrates the concentration variations laterally and vertically
at the sites. These results indicate that the model provides an accurate
estimate of variation in DO in the river under dry weather, steady
flow conditions. Predicted values were generally quite close to the
observed range at each of the illustrated locations.
The temperature prediction of the model are not quite as accurate as the
dissolved oxygen predictions. Figures k$ and 46 illustrate the observed
average in the temperature values versus the predicted temperature at the
North Avenue Dam and at Capitol Drive. Most of the variation between the
observed and predicted values is probably due to the wide range in
observed values during each time period. The largest difference between
observed averages and predicted values is 3°C which is acceptable
for modeling purposes.
Calibration for fecal coliforms with the data from the dry weather survey
(Survey I, September, 1976) was not possible. The concentrations observed
at all locations at the beginning of the survey were quite substantial.
The levels increased further during the first day and then declined (see
Figure !6 in Section 4). The only sources of fecal coliforms assumed in
this modeling effort were CSO, storm sewers and the upstream load. During
dry weather, the only source was the upstream load. Unless fecal
coliforms are allowed to have a net growth rate in the river, the observed
concentrations cannot be predicted. A net growth rate for fecal coliforms
in receiving waters is not feasible. Therefore, fecal coliforms are
assumed to be nearly conservative in the river with only a slight die-off
rate. This assumption was evaluated with data from the wet weather surveys.
Many of the parameters input to Harper's model were determined from
literature values or from the June 1975 calibration. There are a number
of parameters which were not measured during either of the intensive
surveys which can significantly influence simulation results for dissolved
oxygen. These are:
I. Lake Michigan inflow.
2. Phytoplankton concentration.
3. Phytoplankton self-shading factor and settling coefficients.
4. Benthic algae concentration.
5. Base extinction coefficients for light.
6. Sediment oxygen demand.
Periodic sampling for this project has provided additional data on all the
above parameters except for the benthic algae concentration, phytoplankton
settling coefficient and phytoplankton self-shading factor.
Inflow of Lake Michigan into the Milwaukee River was estimated based
upon changes in specific conductance values and total dissolved solids
between the North Avenue Dam and the inner harbor. This inflow is quite
variable since it is dependent on atmospheric pressure, wind speed and
direction and the river flow among other factors. Estimates for this
inflow at St. Paul Avenue are between 0 and 150 cfs (4.2 m3/sec). The
109
-------
-O MONITORED
-O PREDICTED
I8r—
17
15
en
LU
<
I
I
>400 1200
21 SEPTEMBER
2400 1200
22 SEPTEMBER
TIME, hours and days
I
62
60
DC
rs
<
LU
O.
Z
UJ
h-
LU
12
58
2^00 1200
23 SEPTEMBER
Figure 4$. Temperature verification results for North Avenue, September 21-23, 1976 (Survey 1)
-------
MONITORED
._.__<» PREDICTED
<
(£.
a:
20
19
18
17
16
15
13
12
1200
21 SEPTEMBER
2400
1200
22 SEPTEMBER
TIME, hours and days
—, 68
66
62 <
60
58
56
1200
23 SEPTEMBER
Figure 46. Temoerature verification results for Capitol Drive,
September 21-23, 1976 (Survey 1).
-------
value utilized in the receiving water model is 55 cfs (1.6 m3/sec) at the
junction of the Milwaukee and Menomonee Rivers. This value was in the
range estimated with field data and provides the best match with the
calibration data. Estimates of this inflow have also been obtained by
integrating velocity measurements taken from several depths (II). An
inflow value at St. Paul Avenue of 50 cfs (\.k m3/sec) was selected for
modeling purposes based upon this analysis.
Periodic sampling for chlorophyll a_ and measurements of secchi depth during
the summer of 1977 were utilized to estimate phytoplankton concentrations
and base extinction coefficients input to Harper's model. The results
of the sampling for chlorophyll a_ are listed in Table A-14 in the Appendix.
Secchi depth measurements were used to estimate the base extinction ,
coefficients by assuming that this depth is at 10-15 percent of the surface
1ight intensi ty.
The results of the sediment oxygen demand investigations have provided a
range of possible values for each area of the river. SOD rates utilized
in Harper's model were calibrated within the observed range. The SOD
data is presented in Section 4.
Benthic algae, which are assumed to represent attached algae and macrophytes
are only important in the shallow reaches of the Milwaukee River between
Port Washington Road and North Avenue. Values for this parameter between
1000 and 8000 mg/m2 have generally allowed for accurate simulation of DO
in the upper portions of the river with the particular concentrations
dependent upon the time of year under analysis. The phytoplankton
self-shading factor and settling coefficients have been adjusted so that
phytoplankton concentration remains in an acceptable range for the
time of year being simulated.
Sensitivity analyses have been conducted on a number of parameters in order
to determine an acceptable range for wet weather calibration. Figures kl
and A8 illustrate the results of these sensitivity analyses for the base
extinction coefficient for light and the sediment oxygen demand rate at
St. Paul Avenue in the Milwaukee River for the September 1976 survey. The
sensitivity analyses presented represent a ± 0.2/ft (0.06/m) change in the
base extinction coefficient and a ± O.I gm/m^-hr change in the SOD rate
throughout the modeled portion of the river. Sensitivity analyses
performed for.this project and calibration results for June 1975 have
been utilized to develop the range of input parameter values listed in
Table 22. Sensitivity analyses on the size of the time step have indicated
that a three hour time step can be utilized for simulations without
introducing significant errors.
Wet Weather Calibration/Verification--
The basic methodology employed in calibrating Harper's model for dissolved
oxygen during wet weather conditions involved three steps:
I. Determining the CSO and storm sewer loads using the STORM model.
112
-------
VjO
3.5
3.0
en
LU
CJJ
X
o
o
LU
O
CO
co
2.0
1.5
0.5
J_ MONITORED RANGE
3 = 0.3
„ = 0.2
I --
T... — j""" ...... ~"il
1200
21 SEPTEMBER
2400 1200
22 SEPTEMBER
TIME, hours and days
2400 1200
I 23 SEPTEMBER
Figure 47- Sensitivity analysis for the dry weather SOD at St. Paul Avenue
-------
o
o
LU
>
_l
o
t/1
O
3.5|—
3.0
2.5
2.0
1-5
1.0
0.5
I
MONITORED RANGE
kb = 1.6
kb = 1.8
~ kb =
2^00 1200 2^00 1200
I 21 SEPTEMBER I 22 SEPTEMBER
TIME, hours and days
2^00 1200
23 SEPTEMBER
Figure *»8. Sensitivity analysis for the base extinction coefficient at St. Paul Avenue.
-------
2. Determining the a, b and T coefficients from the values of ADO
and the decline duration predicted with the regression equations.
3. Calibrating the model within the acceptable range of values utilized
for dry weather calibration/verification.
Data from two intensive surveys of three days duration in May and June, 1977
were utilized to calibrate for wet weather conditions. These intensive
surveys were single storm events with the monitoring being initiated immedi-
ately after the storm. Three additional surveys in August, 1977, June, 1978
and July, 1978 were used to verify the model. Each of these surveys
encompassed multiple rainfall events which made the simulation of CSO impact
more difficult.
TABLE 22. RANGE OF INPUT PARAMETER VALUES UTILIZED FOR DRY WEATHER
CALIBRATION OF HARPER'S MODEL
Input
parameter
kb
phyto
U02
kl
k3
KCS
kc
^
Units
ft"1
(mg/l-ft)"'
mgO_/mg algae
per day
gm/m -hr
mg/1
per day
cfs
Description
Base extinction coefficient
Phytoplankton self-shading factor
Production coefficient for algal
growth
Deoxygenation coefficient
Sediment oxygen demand
BOD half saturation constant
for coliform growth
Col i form die-off coefficient
Lake Michigan inflow
Cal ibrat ion
range
0.5-1 .8
0.01-0. 10
1.5
0.05-0.10
0.0-0.4
10.0-20.0
0.0-10.0
25-100
in
m = O.SOxft
mVsec = 0.028xc.fs
The CSO and storm sewer loadings for each of the storm events utilized
wet weather calibration/verification are listed in Table 23- These
loadings were input to the receiving water model by dividing the CSO area
tributary to the Milwaukee River into six basins. Each of these basins
was modeled separately with STORM and the loads were then assumed to enter
this river as six point sources. Due to the relatively large size of the
completely mixed longitudinal segments (7^0 ft - 225 m) utilized in the
model for the Milwaukee River, this assumption should introduce no observable
errors.
15
-------
Sensitivity analyses were conducted in order to quantify any error due to
these simplifications. This was accomplished by lumping the CSO outfalls
into twelve point sources and then eighteeen point sources. Rather than
evaluate the differences with a single storm simulation, the differences
were evaluated using a one month simulation. The month utilized for this
analysis was June and the rainfall input to the STORM model was for June,
This was an extreme month of rainfall with 8.28 inches recorded by the
National Weather Service at Mitchell Field in Milwaukee. The differences
observed with the various number of point sources simulating CSO outfalls
was evaluated using the area below a DO value of 5-0 mg/1 and above the
predicted levels of DO. This area is actually the magnitude in mg-day/1
of the time the DO is below the level. Table 2k lists the results of this
analysis for several sites in the study area of the Milwaukee River. The
largest effect of the lumping of CSO outfalls was observed at Walnut Street.
By using eighteen point sources rather than six distributed along the river,
the magnitude of the area below 5-0 mg/1 increased by 5 percent. The
variations at the North Avenue, Wells Street and St. Paul Avenue sites were
much less. This analysis indicates that the errors introduced by the
lumping into only six point sources does not significantly effect model
results.
Calibration of Harper's model for the wet weather events mainly involved
determining a consistent method of obtaining the a and b coefficients from
the predicted dissolved oxygen decline and decline duration. The time vary-
ing SOD equation provides an exponential decay of the wet weather SOD rate,
therefore, it is possible to adjust the a and b coefficients from the extreme
condition of nearly all the oxygen demand being exerted in the first three
hour time step to a nearly linear exertion of the demand over many time
steps.
The observed rapid decline in dissolved oxygen following a runoff event can
generally be approximated as a linear decline. On the basis of this obser-
vation, the a and b coefficients have been determined so that the decline
is as linear as possible. This was accomplished by adjusting the coefficients
for a given storm event until the average SOD during the predicted decline
period matched the predicted decline in dissolved oxygen. The coefficients
were then further adjusted until the peak SOD rate during the first three
hour time step was between 1.0 and 1.5 times the average SOD rate. Storm
events with large predicted ADO values had the highest ratios of peak to
average SOD rate.
This methodology proved to be extremely successful for calibration of
Harper's model for wet weather conditions. Figure 49 through 52 illustrate
the calibration results for the surveys conducted in May and June, 1977
(Surveys 2 and 3), at several locations in the Milwaukee River. The May 31,
1977 storm event (Survey 2) was a small volume high intensity storm with
0.17 in. (0.43 cm) of rain recorded during a 25 minute period at the
Broadway Street precipitation gauge. This small event had a severe impact
on DO concentrations particularly at the Wells Street site. These low
DO concentrations appear to be due to three factors:
116
-------
TABLE 23. CSO AND SEPARATE STORM SEWER LOADINGS TO THE MILWAUKEE RIVER FOR THE INTENSIVE
MONITORING SURVEYS PREDICTED WITH STORM
CSO loadings
Survey
number
1
2
3
4
5
6
In? t iat ion
date
9-21-76
5-31-77
6-17-77
8- 3-77
6-16-78
7-26-78
Overflow
vol ume ,
mil. ga 1 .
5.4
69.2
97-5
209.3
54.7
Suspended
solids,
1bx!03
34
178
165
549
122
CBOD,
Ibxl03
17
52
101
210
64
Fecal
col i forms ,
counts
4.0xl0'5
7.7xl0'5
1 . 9x 1 0 1 6
3.lxl016
1 . fxlO16
Separate storm
Runoff
vol ume ,
mil. gal .
8.4
167.6
245.2
573.9
129-3
Suspended
sol ids ,
lbx!03
3
115
128
544
89
loadings
CBOD,
lbx!03
5
30
62
158
42
Fecal
col i forms ,
counts
3-lxlO12
1 .2x!Ql3
3-3x1013
7.!xlOf3
2.2xlO|3
kg = 0.454xlb.
TABLE 24. SENSITIVITY ANALYSIS OF
LUMPING OF CSO OUTFALLS
WATER MODEL
INTO POINT SOURCES IN RECEIVING
Magnitude below 5-0 mg/1,
mg-days/1
Site
St. Paul Avenue
Wells Street
Walnut Street
North Avenue
Distance upstream,
mi (km)
0.2
0.6
1.4
2.4
(0.3)
(1.0)
(2.3)
(3.9)
Number
6
N3.5
88.7
21 .1
0.0
of point sources
12 18
M5.5
91.8
21.7
0.0
116.9
93-9
22.2
0.0
-------
oo
en
E
o
>-
X
o
o
10
PREDICTED DO WITH TIME VARYING SOD
PREDICTED DO W/0 TIME VARYING SOD
I OBSERVED DO RANGE
RAINFALL
•0.30 in.
(0.76 cm)
I
I
0.0*» in.
(0.10 cm)
I JUNE
TIME, hours and days
Figure *»9- Wet weather calibration results for Wells Street, May 31 - June 3, 1977 (Survey 2).
-------
PREDICTED DO WITH
TIME VARYING SOD
PREDICTED DO W/0
TIME VARYING SOD
vo
_ 5
a
>•
x
o 2
CO
I
RAINFALL
0.30 in.
(0.76 cm)
OBSERVED DO RANGE
I
0.04 in.
(0.10 cm)
2400
I
1200
31 MAY
2400
1200
1 JUNE
2400
1200
2 JUNE
2400
TIME, hours and days
Figure 50. Wet weather calibration results for Walnut Street, May 31 - June 3, 1977 (Survey 2)
-------
PREDICTED DO WITH TIME VARYING SOD
NJ
O
C3
X
O
O
to
O
I0r—
I
RAINFALL
0.53 in.
(1-35 cm)
PREDICTED DO W/0 TIME VARYING SOD
OBSERVED DO RANGE
I
2400
1200
18 JUNE
2400
1200
I 19 JUNE
TIME, hours and days
2400
1200
20 JUNE
2400 1200
21 JUNE
Figure 5'• Wet weather calibration results for St. Paul Avenue, June 18-21, 1977 (Survey 3).
-------
RAINFALL
PREDICTED DO WITH
TIME VARYING SOD
PREDICTED DO W/0
10.53 in.
— . TIME VARYING SOD
(1.35 cm)
10
8
I1
»
5 6
o
X
o
a
o
10
0
2
0
—
w
.•«•
r*«J
\ x
\
V
•
2400
t
X
•
^^T
^^
•
•
II ^
T OBSERVED DO RANGE
T I _
T
i*^ ^J«^L
^ ^^^^ ^^^* ^^^ ^^ ^ ^^ ^^ \^^ "^^^
^^^^^. T.^^ ^^^^Ta ,-^ ^ ^^^^
^K^lx^ *— - * T ^^^^
J»
* •
T
•"• 1
~
-•
1
1200 2400 T200 2400 1200 2400 1200 24(
18 JUNE 19 JUNE 20 JUNE 21 JUNE
TIME, hours and days
Figure 52. Wfet weather calibration results for Walnut Street, June 18-21, 1977 (Survey 3).
-------
I. The low river flow - 100 cfs.
2. A large load of pollutants due to more than a month of dry weather
proceeding the storm.
3. Sediment scour from the submerged CSO outfalls in the lower Milwaukee
River.
Using the input parameters determined in the dry weather calibration/verifica-
tion and the a and b coefficients estimated as described previously, the
instream model provided a good prediction of observed DO concentrations
at most of the monitoring locations.
In order to demonstrate the sensitivity of the model to the time varying SOD
equation and the methodology utilized to adjust the a and b coefficients , al1
of the calibration/verification figures show the predicted curve if only
the dry weather SOD rate and soluble (and suspended) load is considered. It
is obvious from this comparison that the significance of this equation on
model output varies from survey to survey on the basis of the rainfall
volume, CSO volume and river flow. These results will be utilized in
Section 7 to evaluate the impact of CSO on DO in the Milwaukee River.
The model predictions for the June 17, 1977 survey (Survey 3) are not as
close to the observed values as those for the May wet weather survey
(Figures 51 and 52). The DO concentrations predicted by the model recover
at a faster rate than those observed during the survey. The results are
still quite acceptable for model calibration purposes.
Table 25 is a list of the calibrated input parameters used for Harper's model,
The only parameters that were assumed to vary throughout a runc/ff year
(April through October) were the phytoplankton concentrations and the benthic
algae concentrations.
In order to verify the model it was necessary to determine the accuracy of
the model prediction after a series of storm events. The August 3~8, 1977
survey (Survey 4) was utilized for the verification since there were four
separate storm events during the first three days of the five day survey.
Figures 53 and 5^ and Appendix Figures A-11 and A-12 illustrate the results
of this verification run. Harper's model accurately predicts the rapid
decline following the initial storm event at 2200 on August 3, 1977 and
the dissolved oxygen recovery beginning on August 6, 1977- The rapid
recovery is partially a function of DO production in the river by phyto-
plankton. The change in river flow from 105 cfs at the beginning of the
survey to about 450 cfs at the end of the survey as observed at the Esta-
brook Park USGS gauging station also accounts for a portion of the DO
recovery.
Figures 55 through 59 illustrate the verification results for the June 1978
(Survey 5) and July 1978 (Survey 6) wet weather surveys. The DO results at
122
-------
TABLE 25. CALIBRATION VALUES FOR HARPER'S MODEL INPUT PARAMETERS FOR THE MILWAUKEE RIVER
VA>
River reach
Brown Deer Road to
North Avenue
North Avenue to
Walnut Street
Walnut Street to
River Junction
All
All
Brown Deer Road to
Port Washington Road
Port Washington Road to
River Junction
Brown Deer Road to
Port Washington Road
Port Washington Road to
North Avenue Dam
North Avenue Dam to
River Junction
All
All
River Junction
Input
parameter
kb
u
kb
Kphyto
U02
Wi.
k
k,
k3
J
k3
j
k3
Kcs
I/O
kc
QL
Units
ft'1
ft'1
ft'1
(mg/l-ft)"'
mg02/mg algae
per day
per day
gm/m -hr
gm/m -hr
gm/m -hr
mg/1
per day
cfs
Description
Base extinction coefficient
Base extinction coefficient
Base extinction coefficient
Phytoplankton self-shading factor
Production coefficient for algal
growth
Deoxygenation coefficient
Deoxygenat ion coefficient
Sediment oxygen demand
Sediment oxygen demand
Sediment oxygen demand
BOD half saturation constant for
col i forms
Coliform die-off coefficient
Lake Michigan inflow
Cal i brat ion
value
0.9
1.4
1.6
0.05
1.5
0.05
0.10
0.10
0.20
0.30
20.0
5.0
55.0
m = O.SOxft.
m3/sec = 0.28xcfs
-------
PREDICTED DO WITH
TIME VARYING SOD
PREDICTED DO W/0
TIME VARYING SOD
I
OBSERVED DO RANGE
I
RAINFALL
0.33 in.
(0.84 cm)
I
0.17 in.
(0.43 cm)
I
0.22 in.
1(0.56 cm)
0.07 in.
(0.18 cm)
I
- 6
X
o
O
00 2
00 *•
2400
1200
AUGUST
2400
1200 2400
5 AUGUST
1200
6 AUGUST
2400 1200
7 AUGUST
2400
8 AUGUST
TIME, hours and days
Figure 53. Wet weather verification results for St. Paul Avenue, August 4-8, 1977 (Survey 4).
-------
.PREDICTED DO WITH
TIME VARYING SOD
.PREDICTED DO W/0
'TIME VARYING SOD
OBSERVED DO RANGE
I
RAINFALL
0.33 in.
(0.84 cm)
10.17 in.
(0.43 cm)
0.22 in.
(0.56 cm)
0.07 in.
(0.18 cm)
2400
1200 2400
4 AUGUST I
1200
2400
1200
2400
5 AUGUST | 6 AUGUST
TIME, hours and days
1200
7 AUGUST
2400
8 AUGUST
Figure 54. Wet weather verification results for Wells Street, August 4-8, 1977 (Survey
-------
PREDICTED DO WITH
TIME VARYING SOD
PREDICTED DO W/0
TIME VARYING SOD
OBSERVED DO RANGE
RAINFALL
1-0
0.18 In.
(0.46 cm)
12 I—
10
Ol
' 8
11
0.44 in.
(1 .12 cm)
Q
UJ
>
_l
O
6
4
2
0
I
10.85 in.
(2.16 cm)
I
0.25 in.
(0.64 cm)
2400
1200
16 JUNE
2400 1200
I 17 JUNE
TIME, hours and days
2400
1200
18 JUNE
2400
Figure 55. Wet weather verification results for St. Paul Avenue, June 16-18, 1978 (Survey 5).
-------
PREDICTED DO WITH
TIME VARYING SOD
PREDICTED DO W/0
TIME VARYING SOD
OBSERVED DO RANGE
to
RAINFALL
0.18 in.
(0.46 cm)
12 i—
10
o>
8
u>
i«
o
to
CO
I I
2400
0.44 in.
(1.12 cm)
I
0.85 in.
(2.16 cm)
10.25 in.
(0.64 cm)
1200
16 JUNE
2400
1
1200
~2400
1200
18 JUNE
2400
17 JUNE
TIME, hours and days
Figure 56. Wet weather verification results for Walnut Street, June 16-18, 1978 (Survey 5)
-------
•PREDICTED DO
NJ
oo
CD
z
LU
C3
X
o
o
CO
10
12
10
0.18 in.
(0.1*6 cm)
I I
RAINFALL
OBSERVED DO
2400
0.44 in.
(1.12 cm)
1
1200
16 JUNE
10.85 in. | 0.25 in.
(2.16 cm) 1(0.64 cm)
2400
i
1
2400
1
1200
18 JUNE
1200
17 JUNE
TIME, hours and days
Figure 57- Wet weather verification results for North Avenue, June 16-18, 1978 (Survey 5)
-------
PREDICTED DO WITH TIME VARYING SOD
RAINFALL
0.22 in.
(0.56 cm)
0.22 in.
(0.56 cm)
PREDICTED DO W/0 TIME VARYING SOD
I
OBSERVED DO RANGE
IVJ
VX3
2400 1200
26 JULY
2400 1200 2400 1200
' 27 JULY I 28 JULY
TIME, hours and days
2400 1200
' 29 JULY
Figure 58. Wet weather verification results for St. Paul Avenue, July 26-29, 1978 (Survey 6)
-------
PREDICTED DO WITH TIME VARYING SOD
10
o
o>
E
x
o
o
LU
>
_l
o
CO
co
PREDICTED DO W/0 TIME VARYING SOD
OBSERVED DO RANGE
I
RAINFALL
I6r—0.22 inf • 0.22 in.
I I
14
12
10
8
6
(0.56 cm) • |(0,56 cm)
2400
1200
26 JULY
2400
1200
2400
27 JULY I
TIME, hours and days
1200 2400 1200
28 JULY ' 29 JULY
Figure 59. Wet weather verification results for Walnut Street, July 26-29, 1978 (Survey 6)
-------
St. Paul Avenue for Survey 5 (Figure 55) indicate that the model predicts
fairly well for this event. The observed DO values are quite inconsistent
during the initial day of the survey which accounts for some of the
differences between observed and predicted values. The predicted DO
does not recover quite as rapidly as the observed DO, however, the general
fit of predicted to observed values is good. The predictions at
the Walnut Street and North Avenue sites (Figures 56 and 57) are also quite
good for this survey. During Survey 6, a marked diurnal variation in DO
was observed even in the lower portions of the river. The general trend in
DO levels was matched at the St. Paul Avenue site (Figure 58). However,
the diurnal variation was not predicted well. Variations of this magnitude
were not observed at this station during other surveys or from the continuous
monitoring results for similar times of the year. Unrealistically high
values of phytoplankton would need to be input to the model to match the
observed values. The verification results for DO at the Walnut Street
(Figure 59) and North Avenue (Appendix Figure A-15) sites are much better.
Despite some of the variability between observed and predicted values,
the model predictions are adequate to verify the model.
A source of error for all of the calibration/verification simulations was the
assumption of a constant inflow from Lake Michigan. This inflow was considered
a constant during all river flow conditions which may account for some of
the variation between monitored and predicted values. The inflow only has
a measureable impact on simulation results during low flow conditions.
During high flows, the inflow was still simulated, however, the impact on
predicted values was negligible. Although the model predictions are
generally good, accurate simulations of the Lake-river interchange during
all flow conditions will only be possible if a relationship between inflow
and river flow (or another parameter) can be developed.
As mentioned in the discussion of dry weather calibration, accurate
predictions of current levels of fecal coliforms was impossible due to the
assumptions utilized in determining the loadings from the storm sewer areas.
At the present time, there are numerous cross-connected storm sewers and
some combined sewers which discharge into Lincoln Creek. There are also
some cross-connected storm sewers discharging into the Milwaukee River
upstream of the CSO area. There is an on-going program to eliminate these
discharges. The STORM model was calibrated for these areas under the
assumptions that all the cross-connections will be eliminated and that the
combined sewers discharging to Lincoln Creek will also be eliminated.
Despite the assumptions utilized for the storm sewer loadings, the fecal
coliform concentrations predicted for monitoring sites in the CSO area
of the Milwaukee River are fairly good. Figures 60 through 62 illustrate the
model predictions of fecal coliform concentrations versus the monitoring
results. Figure 62 illustrates the predicted versus observed values at
Port Washington Road during the August 3~8, 1977 wet weather survey. This
site is directly downstream from the junction of Lincoln Creek and the
Milwaukee River. It is obvious from the illustration that the actual
load from Lincoln Creek is much higher than the predicted load.
131
-------
_0 MONITORED
--0 PREDICTED
100,000
10,000
O
O
I ,000
cc
O
O
o
<
o
100
10
17 18
JUNE JUNE
100,000
10,000
I ,000
100
19 20
JUNE JUNE
NORTH AVENUE DAM
10
100,000
t
6
17
JUNE
18
JUNE
19
JUNE
10,000
I ,000
100
20
JUNE
WALNUT STREET
171 IF
JUNE JUNE
19
JUNE
20
JUNE
ST. PAUL AVENUE
Figure 60. Wet weather fecal coliform calibration results, June 18-20, 1977 (Survey 3).
-------
1,000,000==
100,000 =
o
o
o
o
U.
U_
X
10,000 —
o
o
1,000 —
100
MONITORED
PREDICTED
1,000,000^
100,000 —
10,000 =
1,000 —
100
Figure 61. Wet weather fecal coliform calibration results, August 4-7, 1977 (Survey 4).
-------
MONITORED
. O PREDICTED
200,000
100,000
o
o
o 10,000
o
o
1,000
100
AUG.
PORT WASHINGTON ROAD
100,000 p=
10,000
,000 _
100
Figure 62. Wet weather fecal coliform calibration results, August A-7, 1977 (Survey k)
-------
The predicted peak concentrations of fecal coliforms are reasonably close to
observed values. The low upstream load of fecal coliforms has little
effect on the predicted concentrations in the lower river due to the large
load input from the CSO area. This lack of an upstream load of fecal
coliforms may be responsible for a faster predicted decline in fecal coli-
forms than observed. On the basis of these results, the assumption of a
slight die-off rate of fecal coliforms as discussed in the dry weather
calibration section, appears to provide a reasonable estimate of fecal
coliform trends in the Milwaukee River. Calibration for fecal coliforms is
considerably different than calibration for dissolved oxygen for two reasons:
I. The only adjustment factors are the loads and the growth/die-off
coefficient.
2. There is such a large amount of variability In fecal coliform
results that predicted values within an order of magnitude of the
observed values are acceptable.
135
-------
SECTION 7
EVALUATION OF CSO IMPACT
The contribution of CSO to the DO and fecal coliform levels in the Milwaukee
River can now be quantified using the relationships developed in the previous
sections of this report. The source of the water quality impacts for all
rainfall-runoff events can be deciphered and discussed through the long term
operation of the model network that is used to simulate instream water
quality conditions. Once the contribution is fully described, the methodolo-
gies will be related to other sites that do not contain downstream influences
such as Lake Michigan.
DISSOLVED OXYGEN IMPACTS
In order to evaluate the contribution of CSO to the DO impacts within the
Milwaukee River, a sensitivity analysis using the instream model network
with various CSO loads was conducted. The STORM and instream model were
run for one year of rainfall for the period April through October when
non-frozen ground and non-ice cover conditions are present. In order to
measure the differences in the DO curves that are computed by Harper's
model, the technique shown in Figure 63 was utilized. The shaded area below
5-0 mg/1 standard represents the magnitude of the dissolved oxygen impact
which has the units of mg-day/1. The magnitude measurement represents the
difference in water quality between a given standard and the predicted
values. Another measurement which is used to quantify the DO impact is the
number of days of violation that occur within the simulated time period.
Thus, any part of a day which contains a violation of the standard will be
listed as such. Figure 63 presents the data where there are eight DO
values listed per day in the model output. The 5.0 mg/1 standard is violated
during day 2 and 3 of the period with day 2 having four values below the
5.0 mg/1 standard and day 3 having one. Two days of violation will then be
listed along with the magnitude of the violations for this period.
In order to quantify the sensitivity of the river to the loads generated
by CSO, an extreme year of rainfall was selected from the 38 years of
record available for the Milwaukee area. The predicted flow and quality
from STORM for each storm event were input to Harper's model to generate
the frequency and magnitude of DO violations at each site for the
April through October period using mean monthly low flows. The DO standards
for determining the magnitude and frequency of violations were 5.0 mg/1
for sites upstream of the North Avenue Dam and 2.0 mg/1 for sites downstream
of the dam as discussed in Section 3-
136
-------
1Q—
Q tJL^
AT NORTH AVENUE DAM
X
0 4
0
LU
>
f \ / \ / STANDARDS
— ^&/y ^4 VIOLATION
2 —
I I I I 1 1 1 II I I I I I II II I I I I
DAY 1
DAY 2
TIME OF DAY
DAY 3
Figure 63. Typical model output for determining frequency
and magnitude of dissolved oxygen violations.
Next, the predicted pounds' bf BOD, suspended solids and numbers of fecal
coliforms for each overflow event during the year in the CSO area were
doubled and loaded into Harper's model with the same storm sewer and
upstream boundary loads as the previous model runs. This technique was used
to determine the chanqe in DO quality with this 200 percent increase in CSO
pollutant load. Similarly, the CSO load and flow was also completely removed
to simulate the complete elimination of CSO. ihese three simulations were
run using the CSO sediment scour potential for each runoff event and they
were run without the scour mechanism in order to quantify the differences
between these conditions. The results of these model runs are listed in
Table 26 for three sites within the study area. A graphical representation
of the differences in these simulations is shown in Figure 64 for both the
with and without scour mechanism.
TABLE 26. THE DISSOLVED OXYGEN IMPACT OF CSO LOADS VARIATIONS
USING AN EXTREME RUNOFF YEAR
DO impact magnitude,
mg-days/1
with scour
w/o scour
North
Exist-
ing
43.8
Avenue
200$
52.7
Zero
34.6
Walnut
Exist-
ing
17.2
5-3
Street
200%
30.7
18.1
Zero
0
0
St.
Paul Avenue
Exist-
ing
135.
12.
5
3
200%
165-
47.
1
3
Zero
5.0
0
Days DO violations
with scour
w/o scour
32
.0
33
.0
31.0
18.0
9.0
33-0
16.0
0
0
1 10.0
19.0
117.0
43.0
29.0
0
The data of Table 26 and Figure 64 provide a means of evaluating the contri-
bution of CSO to the impacts on DO at various locations within the Milwaukee
137
-------
KO;
1 J W
*V.
(0
In 100
»
o
o.
SOLVED OXYGEN 1
vn
O O
Q
19
3 w
54
ITH
SCOUR
u
n WITHOUT SCOUR
1
EX 1 ST
I
200%
I
ZERO
NORTH AVENUE
EXIST
p
1
200%
ZERO
WALNUT
%
/,
1
y.
V,
\
1
P3
^
^
X
|
V
y
\
m—
EXIST! 200% |ZERO
ST. PAUL
Figure 6A. Sensitivity of dissolved oxygen results in CSO loads.
River. For example, the North Avenue data shows that the doubling of the
CSO load results in an increase of approximately 20 percent in the magnitude
of the DO violations of the 5-0 mg/1 standard. Removal of all CSO at this
site results in a 21 percent decrease in the magnitude. Sediment scouring
from submerged CSO outfalls does not occur in the portions of the river
upstream of the North Avenue Dam. Therefore, the DO impacts at North Avenue
are strictly a function of the wet weather loads and the low flow conditions
that were used for this analysis. At Walnut Street, where sediment scouring
does occur, the differences are more pronounced. Doubling the CSO load
results in a 78 percent increase in the magnitude of the standard violations
of 2.0 mg/1. Complete removal of the CSO load results in zero magnitude below
the standard. The contribution of sediment scour to these violations, as
determined using the without scour results, is approximately 69 percent.
The existing magnitude of DO violations is 17.2 mg-days/1 with scour. This
decreases to 5.3 mg-days/1 (69 percent decrease) when the scour mechanism
is removed. At higher CSO loads (200 percent), the sediment contribution
to the magnitude of DO violations is reduced to approximately 37 percent
at Walnut Street. This reduction in influence of sediments is due to the
higher dissolved and suspended loads utilized in the 200 percent CSO load
simulation.
The St. Paul Avenue site has the most drastic changes in water quality when
comparing the sensitivity of various loads. By doubling the CSO load, the
DO magnitude below 2.0 mg/1 increases by 22 percent while removing the CSO
138
-------
load reduces the impact by 96 percent. Removal of the scour mechanism
from the model network reduces the impact of the existing CSO load by 91
percent when compared to the with scour alternative. This result indicates
the significance of the scouring of sediments on DO conditions in the
lower portions of the Milwaukee River. In comparison, the CSO loads have
only a minor influence on DO levels during the low flow conditions used for
this simulation.
The sensitivity of the DO conditions in the Milwaukee River to the river
flow was evaluated using the mean monthly flows rather than the mean low
flows as in the analyses previously presented. The results of these
simulations are shown in Table 27.
TABLE 27- RESULTS FOR
INSTREAM FLOW CONDITIONS USING AN EXTREME
RUNOFF YEAR
DO
impact magnitude,
mg-days/1
with scour
w/o scour
North Avenue
Low Average
flows flows
A3. 8 0
Walnut Street
Low Average
flows flows
17.2 0
5-3 0
St. Paul Avenue
Low Average
flows flows
135.5 2A.3
12.3 0
Days DO violations
with scour
w/o scour
32.0
0
18.0
9.0
0
110. 0
19-0
35.0
0
Changes in the DO impact when instream flows are increased to average
values is as dramatic as the removal of the scour mechanism. For example,
the St. Paul Avenue site has a decrease in the magnitude of DO violations
of approximately 82 percent when the river flow changes from low flow to
mean flow conditions using the existing CSO loads. The ability of the river
to flush out the low DO conditions and replace them with oxygenated upstream
flow is evident from this analysis.
The previous simulations of water quality are a result of the predictions
using the instream water quality model. In order to check the validity
of the predictions of the model, the monitoring data generated in the
intensive surveys was used to find similar cases which would tend to back up
these findings. Figures 65 and 66 present the monitored and predicted DO
levels at St. Paul Avenue during Survey k and 5, respectively. As was
discussed in Section A, survey k occurred during low flow conditions with
over 1.0 in. (2.5 cm) of rainfall on portions of the CSO area. The dotted
line used in this figure represents the predicted DO level without the
sediment scour mechanism. The difference between these curves is what would
139
-------
PREDICTED DO WITH
TIME VARYING SOD
PREDICTED DO W/0
TIME VARYING SOD
I
OBSERVED DO RANGE
RAINFALL
10.33 in.
(0.84 cm)
10.17 in.
(0.43 cm)
0.22 in.
(0.56 cm)
0.07 in.
(0.18 cm)
I
8r-
-e-
o
o> 6
E
CJJ
24
a
ui
>
_i
O
2400
1200
AUGUST
2400
2400
JLiL
/i
i
1200
5 AUGUST
TIME, hours and days
1200
6 AUGUST
2400
1200
7 AUGUST
8 AUGUST
Figure 65. Comparison of t/ie.finstream model results for with and without the time varying SOD (scour)
for St. Paul Avenue, August 3-8, 1977 (Survey I*).
-------
PREDICTED DO WITH
TIME VARYING SOD
PREDICTED DO W/0
TIME VARYING SOD
OBSERVED DO RANGE
0.18 in.
(0.46 cm)
I I
RAINFALL
0.44 in.
(1.12 cm)
10.85 in. •
(2.16 cm) •
0.25 in.
(0.64 cm)
12 i—
10
. 8
LU
C3
X
o
a
O
to
CO
2400
1200
16 JUNE
2400 1200
I 17 JUNE
TIME, hours and days
2400
1200
18 JUNE
2400
Figure 66. Comparison of instream model results for with and without
the time varying SOD (scour) for St. Paul Avenue, July 16-18, 1978 (Survey 5),
-------
be expected when the scour mechanism is removed during these flow conditions.
Figure 66 represents the DO data from Survey 5 which occurred during high
flows following approximately 2.0 in. (5«l cm) of rainfall. Note that the
DO curves representing the with and without scour cases are closer together
and that the depressed DO levels shown in Figure 65 do not occur in this
event. The effects of sediments are masked by high flows that remove
the deoxygenated waters to the lake and replace them by upstream flows.
Figures 67 and 68 represent the monitored and computed DO levels for the
Walnut Street site during Surveys 2 and 5, respectively. Survey 2 was
conducted during low flow conditions while Survey 5 occurred during high
flows. The differences in the with and without scour computed curves is
most noticeable during the low flow conditions of Survey 2. The monitored
DO levels remain reasonably high for this site throughout both surveys
because of the reduced influence of sediments when compared to sites that
are further downstream.
The monitoring and modeling results were illustrated in the four previous figures
to verify the sensitivity of the river to the different flow rates that has
been developed in the preceeding sections of this report. Obviously, the
monitoring program could not collect data that represents the non-scour
simulation since this is physically impossible. But the sediment studies
and the verification of the instream model have provided an accurate means
of evaluating the CSO contribution to the DO impact in the river.
FECAL COL I FORM IMPACTS
The contributionsof CSO discharges to the fecal coliform levels In the
Milwaukee River were determined using the same techniques as with the DO
data. Thus, the sensitivity to CSO loads and the effect of river flows on
the number of days of standards violations were investigated to produce
an estimate of the source of high coliform levels observed in the monitoring
program. A major problem in deciphering the CSO contribution revolved around
the loads generated from the cross-connected storm sewers in the upper
reaches of the river. A decision was made in the Milwaukee CSO project (II)
to ignore these sanitary overflows because of an ongoing program of the
Milwaukee Metropolitan Sewerage District to remove these cross-connections
in the near future. It was decided to evaluate the fecal coliform impacts
under this same assumption in the subject project. This means that the
instream model was loaded with the CSO fecal coliform load as predicted by
the calibrated STORM model while the upstream contribution of fecal coliforms
was estimated as though the storm sewers in these areas were not cross-
connected. The differences between the instream model predictions and the
monitoring data from the intensive surveys can then be attributed to the
cross-connected storm sewers.
Table 28 lists the fecal coliform results of the instream model for
variations in CSO loads for the extreme runoff year during average and low
flow conditions. The number of days of standards violations takes into
account the differences in standards between the upstream and downstream
142
-------
PREDICTED DO WITH
TIME VARYING SOD
O)
ca
>-
X
o
O
UJ
>
_i
o
C/>
c/>
1
RAINFALL
0.30 in.
(0.76 cm)
•
|
0.04 in.
(0.10 cm)
PREDICTED DO W/0
— TIME VARYING SOD
OBSERVED DO RANGE
I
1
2400
1200
31 MAY
2400 1200
I 1 JUNE
TIME, hours and days
2400
1200
2 JUNE
2400
Figure 67
Comparison of the instream model results for with and without the time varying SOD
(scour) for Walnut Street, May 3 1 -June 3, 1977 (Survey 2).
-------
PREDICTED DO WITH
TIME VARYING SOD
PREDICTED DO W/0
TIME VARYING SOD
OBSERVED DO RANGE
0.18 in.
(0.46 cm)
RAINFALL
I I
0.44 in.
1(1.12 cm)
I
0.85 in.
(2.16 cm)
I
0.25 in.
(0.64 cm)
12 r-
10
\
* 8
43
x 6
o
o
1/1
VI
0
2400
1200
16 JUNE
2400
1200
17 JUNE
2400
1200
18 JUNE
2400
TIME, hours and days
Figure 68. Comparison of instream model results for with and without
the time varying SOD (scour) for Walnut Street, June 16-18, 1978 (Survey 5).
-------
reaches. Thus, *»0 days of violation at the North Avenue site is in reference
to the 200/100 ml standard while the lower reaches are in violation of the
1000/100 ml standard. The most significant result of these simulations is
the absence of standards violations when the entire CSO load is removed from the
the model network. Another point of interest is that doubling of the CSO
load results in an approximate 50 percent increase in the number of days
of violation during low flow conditions. The findings also indicate that
there are more days of violations at Walnut Street than at North Avenue
despite the difference in the standards. This supports the results of the
intensive monitoring surveys which showed increasing levels of fecal
coliforms as the distance downstream increased. This reflects a greater
proportion of the CSO area contributing to the river. The major source
of the fecal coliform levels in the river is thus, the CSO area. The
St. Paul Avenue site has less days of violation than the Walnut Street
site because of dilution from Lake Michigan.
TABLE 28. DAYS OF FECAL COLIFORM VIOLATIONS FOR VARIATIONS IN CSO LOAD
AND RIVER FLOW CONDITIONS USING AN EXTREME RUNOFF YEAR
Average
flow
Low flow
North
Existing
28
ko
Avenue
2002
NA1
57
Zero
0
0
Walnut
Existing
31
*»0
Street
200%
,,n
61
Zero
0
0
St. Paul Avenue
Existing
30
33
200%
NA
50
Zero
0
0
Not avallable.
The total numbers of fecal coliforms discharged to the river and arriving
at the upstream boundaries can be calculated for the year of analysis.
Non cross-connected storm sewers were assumed for these calculations. The
results are shown in Table 29.
TABLE 29. YEARLY FECAL COLIFORM LOADS
Source
Upstream (arriving)
Storm sewers
CSO
counts
3.1 x I0'3
5.3 x I0j!»
2.8 x I01/
These data show that the relative magnitude of the CSO contribution is much
greater than the sum of the other two sources.
APPLICATION TO OTHER STUDY AREAS
The results of this study are site specific because of the unique nature of
the receiving waters. In order to relate the findings of this effort to
other study areas, the specific results will be reduced to a more general
-------
consideration. The most significant aspect of the subject project is the
relationship of the bottom sediments in the river to the DO levels observed
during the monitoring program. Since the source of the sediments is partially
related to the CSO area (approximately 40 to 50 percent of the loading of
oxygen demanding materials to the sediments are attributable to CSO), the
possibility in other study areas for similar buildup of bottom materials
must be evaluated in the design of a monitoring program. Thus, larger
urban areas than the Milwaukee area may have a greater percentage of the
drainage area served by urban storm sewers. These discharges may produce
enough settleable material to cause the build-up of oxygen demanding
sediments in those portions of the receiving waters where scouring velocities
are not present during low flow conditions. These discharges coupled with
combined sewer overflows could produce a sediment oxygen demand which
would have a serious effect on the DO conditions of the river during wet
weather.
The contribution of CSO to the DO impacts in the Milwaukee River has been
quantified as being partially due to the dissolved and suspended loads and
significantly due to the scour of sediments by submerged CSO outfalls.
The most significant contribution is the scouring potential of the submerged
outfalls. This mechanism accounts for zero percent of the impacts at
the North Avenue site and 91 percent at the St. Paul Avenue site. The
lower reaches of the river represent a deep, slow moving portion of the
river while the upstream site is more shallow and represents a free flowing
stream. The Walnut Street site has conditions which range between
these two extremes since the river is not as deep nor is the influence of
the lake as pronounced as at sites further downstream. The contribution
of the scour mechanism to the DO impacts at this site ranges from 37 to
69 percent. The remainder of the impact is due to the CSO loads. The
fecal coliform impacts at all sites are solely attributable to the CSO loads.
The removal of submerged outfalls within the CSO drainage area of the
Milwaukee River was simulated by removing the scour mechanism from the
receiving water model. The change in DO impacts after the removal of this
mechanism was significant. In other receiving waters, the existance of
submerged outfalls may have effects on the DO levels of the stream which
should be investigated in the monitoring program before complex models
are applied to the stream to predict the DO concentrations. The scour and
dispersion of the CSO discharges may have dramatic effects on the quality
of the receiving stream which are normally not included in most analyses.
Investigations of the source and mechanism of water quality impacts in
receiving waters should consider the scouring potential of CSO discharges
and the dispersion of CSO during the sampling efforts.
146
-------
SECTION 8
REFERENCES
1. Sullivan, R. H., e_t^ aj_., "Nationwide Evaluation of Combined Sewer
Overflows and Urban Stormwater Discharges, Volume 1: Executive
Summary", EPA-60Q/2-77-064a, U.S. Environmental Protection Agency,
Cincinnati, Ohio, September 1977.
2. Lager, J. A., et^ aj_., "Urban Stormwater Management and Technology:
Update and Users Guide", EPA-600/8-77-011», U.S. Environmental
Protection Agency, Cincinnati, Ohio, September 1977.
3. Colston, N. V., "Characterization and Treatment of Urban Land Runoff",
EPA-670/2-7A-096, (NTIS-PB 240 987/AS), U.S. Environmental Protection
Agency, Cincinnati, Ohio, December 1974.
k. "Characterization and Control of Combined Sewer Overflows in San
Francisco", Water Research. Vol. 3, p. 531, 1969.
5. Agnew, R. W. , §_t al .. "Biological Treatment of Combined Sewer
Overflow at Kenosha, Wisconsin, EPA-670/2-75-019, U.S. Environmental
Protection Agency, Cincinnati, Ohio, April 1975.
6. Gupta, M. K., et^al., "Screening/Flotation Treatment of Combined Sewer
Overflows, VolumeT", EPA-600/2-77-069a, U.S. Environmental Protection
Agency, Cincinnati, Ohio, August 1977-
7. Crane Company, "Microstraining and Disinfection of Combined Sewer
Overflows", 11023EV006/70, (NTIS-PB 195 674), U.S. Environmental
Protection Agency, Cincinnati, Ohio, June 1970.
8. Simpson, G. D., and Curtis, L. W., "Treatment of CSO and Surface Waters
at Cleveland, Ohio, JWPCF, Vol. Al, No. 2, p. 151, 1969-
9. "Chicago Drives Large Bores to Control Combined Sewer Overflow",
Engineering News Record, McGraw-Hill, New York, February 3, 1977.
10. Milwaukee Metropolitan Sewerage District, "Cleaner Water for the Future
Begins Today", Public Involvement Bulletin, Vol. 1, No. 1, 1977.
11. Metropolitan Sewerage District of Milwaukee, and Stevens, Thompson
6 Runyan, Combined Sewer Overflow Pollution Abatement Project,
March 1974 to present.
-------
12. Consoer, Townsend and Associates, "Detention Tank for Combined Sewer
Overflow, Milwaukee, Wisconsin", Demonstration Project prepared for the
Milwaukee Department of Public Works, Wisconsin Bureau of Engineers,
EPA-600/2-75-071 (NTIS-PB 250 k2J) , U.S. Environmental Protection
Agency, Cincinnati, Ohio, December 1975.
13- Meinholz, T. L., e_t^ aj_. , "Water Quality Analysis of the Milwaukee River
to Meet PG-61 (PRM 75~34) Requirements", Prepared for the Milwaukee
Metropolitan Sewerage District, U.S. Environmental Protection Agency
Grant No. C550772-011, February 1978.
1^4. U.S. Department of the Interior, Geological Survey, "Fluorometer
Procedures for Dye Tracing", Techniques for Water Resources
Investigations of the United States, Geological Survey, Book 3,
Chapter A12, 1970.
15- Godfrey, H. G., and Frederick, B. J., "Dispersion in Natural Channels",
U.S. Department of the Interior, Geological Survey, Open-File Report,
Washington, D.C. 1963.
16. Fischer, H. B., "The Mechanics of Dispersion in Natural Streams",
JASCE-Hydraulics Division, Vol. 93, No. HY6, p. 87, November.
17- Holley, E. R., "Unifield View of Diffusion and Dispersion", JASCE-
Hydraul ics Division, Vol. 95, No. HY2, p. 621, 1969.
18. Chow, V. T., Open-Channel Hydraulics, McGraw-Hill, New York, 1959-
19- Fair, G. M., Geyer, J. C., and Okun, D. A., Water and Wastewater
Engineering; Volume 2, Water Purification and Wastewater Treatment
and Disposal, John Wiley and Sons, New York, 1968.
20. Lucas, A. M., and Thomas, N. A., "Sediment Oxygen Demand in Lake
Erie's Central Basin 1970", Proceedings of the lAth Conference on
Great Lakes Research, International Association for Great Lakes
Research, 1971.
21. McDonnell, A. J., and Hall, S. D., "Effect of Environmental Factors on
Benthal Oxygen Uptake", JWPCF, Vol. 41, No. 2, p. 353, 1969.
22. "Water Resources Data for Wisconsin, Water Year 1977", U.S.
Geological Survey Water-Data Report WI-77-1, 1977.
23. Bothwell, M. L., "Studies on the Distribution of Phytoplankton Pigments
and Nutrients in the Milwaukee Harbor Area and Factors Controlling
Assimilation Numbers", Ph.D. Thesis, University of Wisconsin - Madison,
December 1975.
2k. Zison, S. W., et a_]_., "Rates, Constants, and Kinetic Formulation in Surface
Surface Water Quality Modeling", Draft Report for U.S. Environmental
Protection Agency, Athens, Georgia, September, 1970.
148
-------
25. "Storage, Treatment, Overflow, Runoff Model-User Manual", Hydrologic
Engineering Center, U.S. Army Corps of Engineers, Davis, California,
1976.
26. Metcalf and Eddy, Inc., Wastewater Engineering, McGraw-Hill, New York,
1972.
27. "Wastewater: Is Muskegon County's Solution Your Solution?", EPA
EPA-905/2-76-00^, U.S. Environmental Protection Agency, Cincinnati,
Ohio, 1976.
28. DeFilippi, J. A., and Shih, C. S., "Characteristics of Separated
Storm and Combined Sewer Flows", JWPCF. Vol. *»3, No. 10, p. 2033, 1971.
29. Burm, R. J., Krawczyk, D. F., and Harlow, G. L., "Chemical and
Physical Comparison of Combined and Separate Sewer Discharges",
JWPCF. Vol. kO, No. 1, p. 112, 1968.
30. Benzie, W. J., and Courchaine, R. J., "Discharges from Separate Storm
Sewers and Combined Sewers", JWPCF. Vol. 38, No. 3, P- MO, 1966.
31. DiToro, D. M., O'Connor, D. J., and Thomann, R. V., "A Dynamic Model
of Phytoplankton Population in Natural Waters", Manhattan College,
6k pp., 1970.
32. O'Connor, D. J., and Dobbins, W. E., "Mechanism of Reaeration in
Natural Streams", JASCE-Transactions, Vol. 123, Paper No. 2931*,
PP. 6^1-666, 1958.
33. Raphael, J. M., "Prediction of Temperatures in Rivers and Reservoirs",
JASCE - Power Pi vis ton, Vol. 88, No. P02, pp. I57H81, 1962
pp. 157-tBt, 1962.
-------
SILVER SPRING DRIVE
CAPITOL DRIVE
0_.8
MILWAUKEE RIVER
vn
o
WISCONSIN AVENUE
KEY
1. RIVER JUNCTION
2. MARINE BANK
3. HIGHLAND BLVD.
4. WALNUT STREET
5. HUMBOLDT AVENUE
6. NORTH AVENUE
7. BURLEIGH STREET
8. HUBBARD PARK
9. LINCOLN PARK - SOUTH
10. LINCOLN PARK - CENTRAL
11. LINCOLN PARK - NORTH
12. DEAN ROAD
O X
c
CTJ
> <=
I 73
— m
vi in
MILES
Figure A-l. Total volatile solids (% of sample weight) values in the sediments of the Milwaukee River.
-------
SILVER SPRING DRIVE
i
1.
2.
3.
KEY
RIVER JUNCTION
MARINE BANK
HIGHLAND BLVD.
4. WALNUT STREET
5. HUMBOLDT AVENUE
6. NORTH AVENUE
7. BURLEIGH STREET
8. HUBBARD PARK
9. LINCOLN PARK ^ SOUTH
10. LINCOLN PARK - CENTRAL
11. LINCOLN PARK - NORTH
12. DEAN ROAD
MILES
Figure A-2. Phosphorus (mg/kg) values in the sediments of the Milwaukee River.
-------
SILVER SPRING DRIVE
WISCONSIN AVENUE
km = mi x 0.62
0 1 2
KEY
1 . RIVER JUNCTION
2. MARINE BANK
3. HIGHLAND BLVD.
A. WALNUT STREET
5. HUMBOLDT AVENUE
6. NORTH AVENUE
7. BURLEIGH STREET
8. HUBBARD PARK
9. LINCOLN PARK - SOUTH
10. LINCOLN PARK - CENTRAL
11. LINCOLN PARK - NORTH
12. DEAN ROAD
MILES
Figure A-3- Iron (mg/kg) values in the sediments of the Milwaukee River,
-------
vn
CO
SILVER SPRING DRIVE
WISCONSIN AVENUE
km = mi x 0.62
0 1 2
KEY
1. RIVER JUNCTION
2. MARINE BANK
3. HIGHLAND BLVD
It. WALNUT STREET
5. HUMBOLDT AVENUE
6. NORTH AVENUE
7. BURLEIGH STREET
8. HUBBARD PARK
9. LINCOLN PARK - SOUTH
10. LINCOLN PARK - CENTRAL
11. LINCOLN PARK - NORTH
12. DEAN ROAD
MILES
Figure A-4. Cadmium (mg/kg) values in the sediments of the Milwaukee River.
-------
SILVER SPRING DRIVE
WISCONSIN AVENUE
km = mi x 0.62
0 1 2
KEY
1. RIVER JUNCTION
2. MARINE BANK
3. HIGHLAND BLVD.
A. WALNUT STREET
5. HUMBOLDT AVENUE
6. NORTH AVENUE
7. BURLEIGH STREET
8. HUBBARD PARK
9. LINCOLN PARK - SOUTH
10. LINCOLN PARK - CENTRAL
II. LINCOLN PARK - NORTH
12. DEAN ROAD
MILES
Figure A-5- Copper (mg/kg) values in the sediments of the Milwaukee River.
-------
33
SILVER SPRING DRIVE
CAPITOL DRIVE
116
VAUKEE RIVER
VI
Vfl
KEY
1 . RIVER JUNCTION
2. MARINE BANK
3. HIGHLAND BLVD.
k. WALNUT STREET
5. HUMBOLDT AVENUE
6. NORTH AVENUE
7. BURLEIGH STREET
8. HUBBARD PARK
9. LINCOLN PARK - SOUTH
10. LINCOLN PARK - CENTRAL
11 . LINCOLN PARK - NORTH
12. DEAN ROAD
MILES
Figure A-6. Zinc (mg/kg) values in the sediments of the Milwaukee River,
-------
100,000
10,000
en
E
O 25 ml SEDIMENTS
Q 50 ml SEDIMENTS
£ 75 ml SEDIMENTS
y 100 ml SEDIMENTS
(SEDIMENTS + WATER
1.0 LITER)
1,000
o
in
o
to
o
LU
a
a.
to
10
&• —
V
a <
D l
A
i
Q
10 20 30
SETTLING TIME, hours
O
1*0
Figure A-?. Settling of Milwaukee River sediments in river water with
the samples collected at Wisconsin Avenue in January, 1977-
156
-------
100,000
10,000
O 25 ml SEDIMENTS
Q 50 ml SEDIMENTS
A 75 ml SEDIMENTS
y 100 ml SEDIMENTS
(SEDIMENTS + WATER
en
E
<
cc.
o
z
o
CO
o
CO
LU
a.
co
1,000
1 .0 LITER)
100 —
0 10 20 30 *»0
SETTLING TIME, hours
Figure A-8. Settling of Milwaukee River sediments in river water
with the samples collected at the junction of the
Menomonee and Milwaukee Rivers in February, 1978.
157
-------
I
MONITORED RANGE
PREDICTED
vn
oo
10 r—
8
X
o
UJ 4
_l
o
«/>
CO
^ 2
2400
J JL I .1 I
200
21 SEPTEMBER
T200
2400
22 SEPTEMBER
TIME, hours and days
I I I
2400
1200
23 SEPTEMBER
1500
Figure A-9. Dry weather verification results for Capitol Drive, September 21-23, 1976 (Survey 1)
-------
OBSERVED DO RANGE
PREDICTED DO
\n
10 _
en
X
o
6
4
o
to
CO 2
2400
I
I
I
1200
21 SEPTEMBER
2400 1200
I 22 SEPTEMBER
TIME, hours and days
I
24001200
23 SEPTEMBER
2400
Figure A-10. Dry weather verification res.ults for North Avenue, September 21-23, 1976 (Survey l)
-------
ON
O
PREDICTED DO WITH
TIME VARYING SOD
RAINFALL
12
10 —
e»
X
o
o
CO
to
o 4
I" 0.33 in.
(0.84 cm)
I"
(0
0.17 in.
.43 cm)
I
PREDICTED DO W/0
TIME VARYING SOD
0.22 in.
(0.56 cm)
OBSERVED DO RANGE
0.07 in.
(0.18 cm)
I
24(
)0
1
1 200 2k
4 AUGUST
DO
1 1
1200 2400
5 AUGUST 1
6
1
1200 24(
AUGUST
)0
1200
7 AUGUST
240
0
TIME, hours and days
Figure A-ll. Wet weather verification results for Walnut Street, August 4-8, 1977 (Survey 4)
-------
PREDICTED 00
OBSERVED DO
12
10
I1
X ..
o 6
RAINFALL
10.33 in.
(0.84 cm)
I
0.17 in.
(0.43 cm)
I
0.22 in.
(0.56 cm)
I
1200
I
4 AUGUST
2400 i
5 AUGUST
TIME, hours and days
24001200
6 AUGUST
I
0.07 in.
(0.18 cm)
2400 1200 2'
\ 7 AUGUST 18 AUGUST
Figure A-12. Wet weather verification results for North Avenue, August 4-8, 1977 (Survey 4).
-------
PREDICTED DO
OBSERVED DO
0.18 in.
(0.46 cm)
•
RAINFALL
0.44 in.
(1.12 cm)
I
0.85 in.
(2.16 cm)
I
0.25 in.
(0.64 cm)
O
>-
X
Q
UJ
12 r-
10
8
6
4
2400
Figure A-13-
I
I
I
I
I
1200
16 JUNE
2400
2400
1200
18 JUNE
2400
1200
17 JUNE
TIME, hours and days
Wet weather verification results for Port Washington Road, June 16-18, 1978 (Survey 5)
-------
.PREDICTED DO WITH
TIME VARYING SOD
RAINFALL
0.22 in.| |0.22 in.
(0.56 cm)
(0.56 cm)
I
PREDICTED DO W/0
TIME VARYING SOD
OBSERVED 00 RANGE
13
x
o
6
5
4
3
2
1
0
2400
1200
26 JULY
2400
1200 2400 1200
27 JULY I 28 JULY
TIME, hours and days
2400 1200
29 JULY
Figure A-14. Wet weather verification results for Wells Street, July 26-29
1978 (Survey 6).
-------
.PREDICTED DO
RAINFALL
i n.
0.22 in.B • 0.22
(0.56 cm) | I (0.56 cm)
OBSERVED DO
ca
x
o
o
to
a
I6i—
^ 12
i1
_- 10
8
6
2^00
1200
26 JULY
2^00
1200
27 JULY
2^00
1200
28 JULY
TIME, hours and days
1200
29 JULY
Figure A-15- Wet Weather verification results for North Avenue, July 26-29, 1978 (Survey 6)
-------
TABLE A-l.
APPENDIX TABLES
(A- I through A-16)
SUMMARY OF MILWAUKEE RIVER INTENSIVE MONITORING RESULTS FOR THE
DRY WEATHER SURVEY, SEPTEMBER 21-24, 1976 (SURVEY 1)
(DAILY AVERAGES)
vn
Date
Sept. 21
Sept. 22
Sept. 23
Location
Capitol Drive
North Ave
Humboldt Avenue
Walnut Street
St. Paul Avenue
Capitol Drive
North Ave.
Humboldt Avenue
Walnut Street
St. Paul Avenue
Capitol Drive
North Ave.
Humboldt. Avenue
Walnut Street
St. Paul Avenue
TS, SS, VSS, BOD5' COD, TOC,
mg/1 mg/1 mg/1 mg/1 mg/1 mg/1
2.5 -
4.0 -
5.0 -
3.0 -
3.0 -
3.0 -
_ _
- _
- _
- _
3.0 -
- _
- _
- _
- - _ _ _ _
NH--N,
mg/1
0.03
0.02
0.11
0.34
0.58
0.05
0.01
0.21
0.36
0.62
0.02
0.04
0.18
0.38
0.71
NO. &
NO|-N, Org.N,
mg/ I mg/ 1
0.02 -
0.02 -
0.06
0.09
0.17 -
0.02
OeOl
0.08
0.09 -
0.22
0.02
0.02 -
0.08
0.10
0.26 -
Total
P,
mg/1
0.12
0.15
0.20
0.19
0.12
0.13
0.14
0.23
0.20
0.13
0.13
0.16
0.22
0.20
0.14
Fecal
col [forms,
MPN/100 ml
2100
2470
5200
7450
4120
507
5630
11300
13100
708
207
252
2000
10300
210
-------
TABLE A-2. SUMMARY OF MILWAUKEE RIVER INTENSIVE MONITORING RESULTS FOR
DISSOLVED OXYGEN AND TEMPERATURE DURING DRY WEATHER, SEPTEMBER 21-24, 1976
(SURVEY I).
Dissolved oxygen
Date
September
21
September
22
September
23
Site
Capitol Drive
North Ave. Dam
Humboldt Avenue
Walnut Street
St. Paul Avenue
Capitol Drive
North Ave. Dam
Humboldt Avenue
Walnut Street
St. Paul Avenue
Capitol Drive
North Ave. Dam
Humboldt Avenue
Walnut Street
St. Paul Avenue
Average
7.0
7-1
6.9
4.9
1.6
6.3
6.7
5.9
4.4
2.0
7. A
7.9
6.3
5.4
2.1
Min imum
5.6
5.4
6.0
3.5
0.1
5.4
5.7
4.8
3.2
1.3
5.4
6.5
5.0
4.1
1.4
•* mg/l
Maximum
8.5
8.4
7.8
6.3
2.3
7.2
8.3
7.8
5.6
2.6
9.9
9.0
8.2
6.6
2.7
Temperature
Mi n imum
15.0
16.0
16.0
16.0
19.0
9-0
14.0
16.5
17.8
20.5
13-5
15.0
16.0
16.0
18.0
Maximum
19.5
18.0
19.0
20.0
24.0
18.5
17.0
IB. 5
19.5
23.0
18.5
lb.0
18.5
19-5
25.0
166
-------
TABLE A-3. SUMMARY OF MILWAUKEE RIVER INTENSIVE MONITORING RESULTS FOR THE
WET WEATHER SURVEY, MAY 31 - JUNE 3, 1977 (SURVEY 2)
(DAILY AVERAGES)
ON
Date
May 31
June 1
June 2
June 3
Location
Port Washington Rd
North Ave. Dam
Walnut Street
Wells Street
St. Paul Avenue
Port Washington Rd
North Ave. Dam
Walnut Street
Wells Street
St. Paul Avenue
Port Washington Rd
North Ave. Dam
Walnut Street
Wells Street
St. Paul Avenue
Port Washington Rd
North Ave. Dam
Walnut Street
Wei Is Street
St. Paul Avenue
TS,
mg/1
572
543
497
572
562
619
691
606
518
441
1990
590
591
497
440
1101
597
962
505
489
SS,
mg/1
46.8
35.8
16.0
22.5
21.3
20.8
24.8
17-0
10.8
26.0
14.8
18.0
12.8
13.0
12.3
1.0
13.0
8.0
17.0
~
vss,
mg/1
14.3
31.7
41.3
22.0
15.5
44.0
25.0
12.5
8.0
16.0
12.3
13.7
10.7
10.7
8.0
1.0
6.0
8.0
2.0
11.0
BOD5, COD,
mg/ 1 mg/ 1
3.3
4.7
-
-
3.3
3.7
5.3
-
3.7
3.5
4.0
-
-
4.0
4.0
3.0
-
3.0
38.3
38.8
40.0
34.5
26.5
36.0
42.8
41.5
33.5
24.3
38.0
33.8
33.3
27.0
21.3
36.0
35.0
80.0
25.0
16.0
TOC,
mg/1
22.0
21.0
22.0
18.0
16.0
16.0
17-0
16.0
14.0
10.0
18.0
18.0
16.0
12.0
10.0
15.0
30.0
19.0
13.0
NH -N,
mg/1
0.09
0.35
0.27
0.50
0.58
0.05
0.09
0.47
0.41
0.80
0.09
0.11
0.30
0.42
0.66
0.05
0.22
0.53
0.70
0.79
N02 £
N03-N, Org.N,
mg/1 mg/1
0.04
0.09
0.12
.87
1.19
.27
0.19 0.97
0.78 0.79
0.04
0.19
0.13 (
0.14
1.00
.29
).96
-15
0.19 0.87
0.04
0.09
1.14
0.10
0.12
0.02 1
0.03 o
0.19 3
0.13 0
0.12
.17
.63
.45
.52
.12
.30
.76
.27
.85
.11
Total
P,
mg/1
0.29
0.32
0.34
0.24
0.18
0.22
0.31
0.26
0.20
0.14
0.21
0.30
0.24
0.24
0.28
0.30
0.28
0.65
0.27
*r 9 «- /
0.16
Fecal
co 1 i f o rms >
MPN/100 ml
503
6200
11225
10450
3195
283
32300
4130
• • ^ w
8700
V / W
4630
563
3090
2110
5980
^ ./ **w
9000
250
740
/ ' w
320
270
16000
-------
TABLE A-**. SUMMARY OF MILWAUKEE RIVER INTENSIVE MONITORING RESULTS FOR
DISSOLVED OXYGEN AND TEMPERATURE DURING WET WEATHER, MAY 31-JUNE 2, 1977
(SURVEY 2)
Dissolved oxygen
Date
May 31
June 1
June 2
Site Average
Port
Washington Rd.
North Ave. Dam
Walnut Street
Wells Street
St. Paul Avenue
Port
Washington Rd.
North Ave. Dam
Walnut Street
Wei Is Street
St. Paul Avenue
Port
Washington Rd.
North Ave. Dam
Walnut Street
Wells Street
St. Paul Avenue
7
3
3
1
2
6
3
2
0
2
«
5
3
0
1
.8
.3
.6
.A
.5
.3
.8
.7
.1*
.1*
.6
.3
.2
.8
.9
Min
6
1
2
0
1
5
3
2
0
0
5
2
2
0
Q
imum
.8
.k
.3
.3
.6
.2
.1
.1
.0
.9
.2
.9
.1
.1
.6
- mg/1
Max
9
5
A
3
3
7
i»
3
2
6
1
6
A
2
3
imum
.A
.0
.3
.0
.5
.2
.0
.9
.*»
.2
1 .0
.8
.3
.it
.6
Temperature
Minimum
20
20
20
22
18
20
20
21
19
16
19
19
20
19
16
.0
.0
.0
.5
.0
.2
.0
.0
.2
.2
.8
.5
.0
.0
.0
Maximum
23
21
22
19
21
21
21
22
22
22
22
21
22
22
23
.0
.8
.0
.0
.5
.5
.0
.k
.5
.0
.0
.5
.2
.5
.0
168
-------
TABLE A-5. SUMMARY OF MILWAUKEE RIVER INTENSIVE MONITORING RESULTS FOR THE
WET WEATHER SURVEY, JUNE 18-20, 1977 (SURVEY 3)
(DAILY AVERAGES)
Date
June 18
June 19
June 20
Location
Port Washington Rd
North Ave. Dam
Walnut Street
Wells Street
St. Paul Avenue
Port Washington Rd
North Ave. Dam
Walnut Street
Wells Street
St. Paul Avenue
Port Washington Rd
North Ave. Dam
Walnut Street
Wells Street
St. Paul Avenue
TS,
mg/1
481
477
461
453
454
419
452
432
446
406
518
444
423
390
353
SS,
mg/1
43.3
40.3
23.0
23.2
23.2
29.2
33.6
26.6
20.6
18.4
19.0
25.8
26.2
14.4
12.8
VSS, B°°5, COD, TOC,
mg/1 mg/1 mg/1 mg/1
19.5 -
19.5 -
13.2 -
13.2 -
11.8 -
16.2 -
16.8 -
14.2 -
12.4 -
8.6 -
10.8 -
13.2 -
11.0 -
7.6 -
5.8 -
23.0
22.0
21.0
19.0
17.5
21.0
21.0
18.5
19.0
17.0
_
-
-
-
-
NH3-N,
mg/1
0.04
0.01
0.02
0.05
0.06
0.01
0.03
0.02
Ooll
0.21
0.02
0.02
0.10
0.23
0.28
N02 6
N03~N, Org.N
mg/1 mg/1
0.04 -
0.02
0.03 -
0.09 -
0.12
0.02
0.01
0.01
0.02
0.08 -
0.01
0.02
0.02 -
0.03
0.06
Total Fecal
, P, col I forms ,
mg/I MPN/100 ml
0.20
0.20
0.20
0.20
0.16
0.16
0.24
0.23
0.17
0.16
0.14
0.20
0.11
O.J2
OolO
8780
8880
27200
33200
51800
768
450
238
1260
11530
„
_
_
_
-
-------
TABLE A-6. SUMMARY OF THE MILWAUKEE RIVER INTENSIVE
MONITORING RESULTS FOR DISSOLVED OXYGEN AND TEMPERATURE
DURING WET WEATHER, JUNE 18-20, 1977 (SURVEY 3)
Dissolved oxygen - mg/1
Date Site
June 18 Port
Washington Rd.
North Ave. Dam
Walnut Street
Wei Is Street
St. Paul Avenue
June 19 Port
Washington Rd.
North Ave. Dam
Walnut Street
Wells Street
St. Paul Avenue
June 20 Port
Washington Rd.
North Ave. Dam
Walnut Street
Wells Street
St. Paul Avenue
Average
11.3
8.8
7.3
5.1
3.5
12.2
9.3
6.6
3.6
2.9
9.3
6.7
5.0
2.7
3.6
Minimum
3.**
5.5
3.1
3.6
0.2
10.9
A. 6
3. a
1 .6
0.9
it. 6
5-7
1.8
1.6
2.1
Maximum
12.5
13.4
9-8
7.2
6.6
13.3
13.3
8.8
6.0
4.6
13.3
7.8
6.7
4.8
3-8
Temperature
Minimum
22.5
22.0
21 .8
21 .0
19.0
24.0
23.8
23.5
23.5
20.0
24.0
22.0
23-5
20.0
16.0
Maximum
24.5
24.5
23.5
22.5
23.2
25.2
25.5
25.8
25.2
24.0
25.2
25.2
26.0
26.2
23.8
170
-------
TABLE A-7. SUMMARY OF MILWAUKEE RIVER INTENSIVE MONITORING RESULTS FOR THE
WET WEATHER SURVEY, AUGUST 4-8, 1977 (SURVEY 4)
(DAILY AVERAGES)
Date
Aug. 4
Aug. 5
Aug. 6
Location
Port Washington Rd
North Ave. Dam
Walnut Street
Wells Street
St. Paul Avenue
Port Washington Rd
North Ave. Dam
Walnut Street
Wells Street
St. Paul Avenue
Port Washington Rd
North Ave. Dam
Walnut Street
Wei Is Street
St. Paul Avenue
TS,
mg/1
490
513
470
438
434
466
493
478
469
429
435
444
440
431
405
SS,
mg/1
26.5
32.3
27.0
37.3
34.7
27.3
35.0
26.8
29.0
23.5
35.0
41.0
34.8
21.0
16.3
VSS, BOD5, COD,
mg/1 mg/1 mg/1
11.5 -
13.8 -
10.5 -
11.3 -
13.3 -
12.0
13.5 -
10.0
15.8 -
13. 3 -
12.3 -
11.5 -
12.5 -
9.0 -
8.8 -
TOC,
mg/1
22.3
26.8
21.3
26.3
21.7
24.8
21.3
21.0
25.8
23.8
18.3
23.0
21.5
20.3
20.5
MM Kl N02 &
NH3 N' N07-N, Org.N,
mg/1 mg/I -mg/1
0.10 -
0.24
0.29 -
0.10
0.22
0.11
0.16
0.24 -
0.10
0.10 -
0.12 -
0.19 -
0.20
0.22
0.19 -
Total
P,
mg/1
0.23
0.36
0.33
0.30
0.26
0.30
0.31
0.35
0.45
0.40
0.26
0.30
0.28
0.25
0.32
Fecal
col i forms,
NPN/100 ml
115000
61000
46100
25000
186000
21900
22400
41000
119000
500000
1310
1910
4700
192000
250000
(continued)
-------
TABLE A-7 (continued).
NJ
Date
Aug. 7
Aug. 8
Location
Port Washington Rd
North Ave. Dam
Walnut Street
Wells Street
St. Paul Avenue
Port Washington Rd
North Ave. Dam
Walnut Street
St. Paul Avenue
TS,
mg/1
523
541
560
514
515
515
500
460
487
SS,
mg/1
39.3
1*3.5
42.0
33.3
30.3
44.0
47.5
45.5
29.5
VSS,
mg/1
17.8
16.0
15.8
12.5
1J.3
17.0
13.5
19.5
13.0
BOD5> COD,
mg/1 mg/1
•• •
-
-
-
- -
- -
-
-
-
TOC,
mg/1
22.3
23.0
22.5
21.0
21.0
22.5
26.5
24.5
22.5
N03-N,
mg/1
0.10
0.10
0.10
0.18
0.26
0.01
0.01
0.01
0.23
N02 &
N03-N,
mg/1
-
-
-
-
_
-
-
-
Org.N,
mg/1
-
-
-
-
_
-
-
-
Total
P,
mg/1
0.24
0.29
0.29
0.26
0.24
0.24
0.22
0.27
0.20
Fecal
col i forms,
MPN/100 ml
295
405
645
175
4800
170
360
900
4800
-------
TABLE Ar8. SUMMARY OF THE MILWAUKEE RIVER INTENSIVE
MONITORING RESULTS FOR DISSOLVED OXYGEN AND TEMPERATURE
DURING WET WEATHER, AUGUST 4-8, 1977 (SURVEY 4)
Dissolved oxygen - mg/1
Date
August 4
August 5
August 6
August 7
Site
Port
Washington Rd .
North Ave. Dam
Walnut Street
Wells Street
St. Paul Avenue
Port
Washington Rd .
North Ave. Dam
Walnut Street
Wells Street
St. Paul Avenue
Port
Washington Rd
North Ave. Dam
Walnut Street
Wei Is Street
St. Paul Avenue
Port
Washington Rd .
North Ave. Dam
Walnut Street
Wei Is Street
St. Paul Avenue
Average
6.8
4.0
-
1 .2
1 .0
6.2
5.5
5.3
0.6
0.3
9-3
7.6
6.6
3-2
1.6
9-5
8.3
7.7
6.0
4.0
Mi nimum
5.6
2.1
-
0.1
0.0
5.4
3.4
3.3
0.0
0.0
5.5
4.9
4.0
0.4
0.0
8.0
5.8
6.2
4.7
1.8
Maximum
8.4
5.2
-
2.9
2.8
8.5
8.6
7.4
3.2
1.4
15.0
11.4
12.0
7.7
5.4
11.4
11 .0
9-3
7.5
6.2
Temperature
Minimum Maximum
21.5
22.2
19. »
19.5
20.0
21.5
21 .9
21.5
21.5
20.0
21 .0
21.5
21 .0
22.0
19-5
22.5
22.0
23.0
23.0
20.0
23.0
23.0
23.0
21.5
21.5
25.0
23-5
23.5
23-5
23.5
24.0
23-5
24.0
24.0
23.5
24.0
25.0
24.5
24.0
23.5
(contl
nued)
173
-------
TABLE A-8. (continued)
Dissolved oxygen - mg/1
Date
August 8
Site
Port
Washington Rd.
North Ave. Dam
Walnut Street
Wei Is Street
St. Paul Avenue
Average
8.6
7-1
8.0
7.1
5.2
Mini mum
7-5
5.**
7.3
6.0
3.8
Maximum
9.1,
9.0
8.8
8.0
6.2
Temperature
Min imum
22.0
22.0
22.0
23.0
18.0
Maximum
24.0
24.0
24.0
23-5
23.0
-------
TABLE A-9. SUMMARY OF MILWAUKEE RIVER INTENSIVE MONITORING RESULTS FOR THE
WET WEATHER SURVEY, JUNE 16 - JUNE 18, 1978 (SURVEY 5)
(DAILY AVERAGES)
Date
June 16
June 1 7
June 18
TS,
Location mg/1
Port Washington Rd
North Ave. Dam
Walnut Street
Wei Is Street
St. Paul Avenue
Port Washington Rd
North Ave. Dam
Walnut Street
Wells Street
St. Paul Avenue
Port Washington Rd
North Avee Dam
Walnut Street
Wei Is Street
St. Paul Avenue
ss, vss, BOD5,
mg/1 mg/1 mg/1
3
7
8
11
4
-
-
-
-
— — —
8
10
9
9
10
COD,
mg/1
32
37
38
45
41
39
39
37
53
44
39
39
37
41
41
TOC,
mg/1
30
38
40
42
37
13
11
11
19
14
12
12
21
12
13
NH3-N,
mg/1
<.10
.13
.17
.46
.67
.18
.22
.29
.35
.29
c23
.26
.29
.39
.34
N02 £
N03-N, Org.N,
mg/1 mg/1
.35 -
o22
.15 -
.14 -
.15 -
.58 -
.51 -
.54 -
o50 -
.50 -
K73 -
1.56 -
1.51 -
1.42 -
1.37 -
Total
P,
mg/1
o!4
.18
.18
.23
.24
.12
.22
.20
.37
.23
.20
.25
.20
.24
.27
Fecal
col i forms,
MPN/100 ml
1930
3000
10200
21500
13300
12100
18800
64000
94700
55700
9900
11000
9800
50800
13800
-------
Table A-10. SUMMARY OF THE MILWAUKEE RIVER INTENSIVE
MONITORING RESULTS, FOR DISSOLVED OXYGEN AND TEMPERATURE
DURING WET WEATHER, JUNE 16-19, 1978 (SURVEY 5)
Date Site
6/16/78 Port Washington Rd.
North Ave. Dam
Walnut St.
Wei Is St.
St. Paul Ave.
6/17/78 Port Washington Rd.
North Ave. Dam
Walnut St.
Wells St.
St. Paul Ave.
6/18/78 Port Washington Rd.
North Ave. Dam
Walnut Street
Wells St.
St. Paul Ave.
6/19/78 Port Washington Rs .
North Ave. Dam
Walnut St.
Wells St.
St. Paul Ave.
Ave.
7.8
7.5
8.0
6.0
3-9
6.3
7.2
7.5
6.7
5.7
6.7
6.8
7-7
7.3
6.3
7.6
7-2
8.2
7.6
7.1
Dissolved oxygen,
Min.
6.6
6.0
6.8
5.1
2.2
6.1
6.2
5.2
6.2
3.7
4.1
4.6
4.6
4.3
4.2
7.6
7.2
8.0
7.5
7.0
mq/1
Max.
8.1
9.8
9.4
7.2
5o6
8.8
8.2
8o5
7.7
7.0
8.3
8.5
9.1
8.1
7.7
7.6
7.2
8.5
7.9
7.5
Temperature
Min.
18.0
17.0
17.5
17.5
18.0
18,0
18.0
18.0
17.5
17.5
18.5
18.25
18.5
19.0
19.0
19.5
19.5
19.5
20.0
20.0
Maxo
20o5
20.5
21oO
21.0
20.75
20.5
20.0
20.5
20 00
20.0
22.0
2KO
22.0
21.0
21.0
19.5
21.0
21.0
21.0
21.0
-------
TABLE A-11.
SUMMARY OF MILWAUKEE RIVER INTENSIVE MONITORING RESULTS FOR THE
WET WEATHER SURVEY, JULY 26 - JULY 29, 1978 (SURVEY 6)
(DAILY AVERAGES)
Date
July 26
July 27
July 28
July 29
TS,
Location mg/1
Port Washington Rd
North Ave. Dam
Walnut Street
Wei Is Street
St. Paul Avenue
Port Washington Rd
North Ave. Dam
Walnut Street
Wei Is Street
St. Paul Avenue
Port Washington Rd
North Ave. Dam
Walnut Street
Wei Is Street
St. Paul Avenue
Port Washington Rd
North Avenue Dam
Walnut Street
Wells Street
St. Paul Avenue
ss, vss, BOD5,
mg/1 mg/1 mg/1
10
5
7
14
10
6
7
8
8
8
7
8
9
8
9
5
6
7
5
6
COD,
mg/1
45
49
47
56
51
44
46
46
42
44
47
47
44
39
34
41
47
42
41
34
TOC,
mg/1
16.5
15.5
16.5
20.5
19.0
14.5
15.0
15.5
14.0
14.8
16.5
15.3
14.8
14.3
14.0
17.0
18.0
16.0
17.0
13.5
NH3-N,
mg/1
.17
.22
.22
.44
.57
.25
.19
.38
.54
.35
.16
.20
.09
.23
.32
.25
.24
.04
.21
.51
N02 &
N03-N, Org.N,
mg/ 1 mg/ 1
.u6
.08
.12 -
.12 -
.06
.15 -
.12
.10 -
.12 -
.12 -
.12 -
.06
.06
.10
.08
<.01
.05 -
<.01
.05 -
.17 -
Total
P,
mg/1
.15
.24
.23
.24
.22
.15
.15
.20
.18
.23
.15
.19
.17
.18
.19
.14
.23
.18
.19
.16
Fecal
col i forms,
MPN/100 ml
680
28000
36700
26700
21700
6700
13800
15900
50200
38300
680
12500
3540
5870
33300
310
9000
1400
1300
30000
-------
TABLE A-12. SUMMARY OF THE MILWAUKEE RIVER INTENSIVE
MONITORING RESULTS FOR DISSOLVED OXYGEN AND TEMPERATURE
DURING WET WEATHER, JULY 26-29, 1978 (SURVEY 6)
CO
Date
7/26/78
7/27/78
7/28/78
7/29/78
Dissolved oxygen, mg/1
Site Ave. Min.
Port Washington Rd. 9.2 8.8
North Ave. Dam 7-7 7.2
Walnut St. 7-5 5.6
Wells St. 5.3
Wells St. ^ 2.8
St. Paul Ave. 2.7 0.0
Port Washington Rd. 12-] 8-2
North Ave. Dam 9.3 5.4
Walnut St. 7.4 3-7
Wells St. 5-0 3.2
St. Paul Ave. 3-4 0.7
Port Washington Rd . 13-3 12.1
North Ave. Dam 10.3 7-7
Walnut St. 8.1 7.4
Wells St. 6.1 5.4
St. Paul Ave. 3.6 2.7
Temperature
Max. Min. Max.
9.8 23.0 25.0
8.6 23.0 23.5
8.6 21.0 23-0
6.3 21.0 24.0
5.8 22.5 25.0
15 22.0 25.0
15 22.0 25.0
9.5 21.0 2*4.0
7.0 21.0 24.0
5.9 21.0 23.5
15 22.0 23.0
14.6 21.0 23.0
12.2 21.5 24oO
7.2 22.5 24.0
6.0 21.0 26.0
15 22.5 23.0
14.1 22.0 23.0
10.2 22.0 23.0
7.3 22.5 24.0
5.6 18.5 25.0
-------
TABLE A-13- ESTABROOK PARK RIVER FLOW DURING THE INTENSIVE MONITORING
SURVEYS OF THE MILWAUKEE RIVER
•vj
VJD
Survey
1
2
3
4
5
6
Date
9/21/76
9/22/76
9/23/76
5/31/77
6/1/77
6/2/77
6/3/77
6/18/77
9/19/77
6/20/77
8A/77
8/5/77
8/6/77
8/7/77
8/8/77
6/16/78
6/17/78
6/18/78
7/26/78
7/27/78
7/28/78
7/29/78
Average
flow,
cfs
79
73
70
89
7k
71
86
250
202
188
170
344
461
448
383
453
1623
1645
342
268
247
255
Maximum
flow,
cfs
85
77
72
15^
80
80
87
393
215
205
418
405
498
491
498
915
3112
2302
568
306
255
255
Minimum
flow,
cfs
77
72
68
56
70
64
77
181
181
163
98
176
312
399
334
206
372
933
275
240
240
255
NOTE: m3/sec = 0.028 x cfs
-------
TABLE A-14. SUMMARY OF MILWAUKEE RIVER CHLOROPHYLL a CONCENTRATIONS.
oo
o
Chlorophyll a., mg/m^
Location June 2k
Brown Deer 64.97
Silver Spring 37-85
Port Washington 33-08
North Avenue
Walnut
Wells
St. Paul 16-99
Inner Harbor
Outer Harbor
July 8
24.55
54.22
38.61
38.00
23.49
19.71
15-17
—
—
July 1J
69.5
68.7
38.7
75.3
81.3
59-4
22.0
--
--
J July 20
152.0
132.0
125.3
121.3
48.7
35.0
15-9
--
--
July 28 August 8
22.75
14.92
25-24
26.86 15-16
26.84 20.99
28.86
23-90 12.36
—
—
August 30
144.05
—
88.85
—
61 .00
—
16.50
17-13
3.48
September 27
16.31
20.19
13-36
19-71
19-46
16.40
17-93
—
--
-------
GO
TABLE A-15- RESULTS. OF CHEMICAL ANALYSIS OF MILWAUKEE RIVER SEDIMENTS
SAMPLED IN MAY AND JUNE, 1977
Location
River Junction
Parameter
Total Solids - %
Volatile Total Solids - %
COD-mg/kg
Nitrite+Nitrate Ni trogen-mg/kg
Ammon i a-N i t rogen-mg/kg
Total Phosphorus-mg/kg
1 ron-mg/kg
Cadmium-mg/kg
Zinc-mg/kg
Lead-mg/kg
Copper-mg/kg
PH
East Mid-Top
57.6
3.0
156,000
0.12
77
616
21 ,500
23
53
34
25
8.0
West
26.9
2.9
118,200
0.06
420
1,648
20,900
15
443
628
117
7.4
East
17.3
2.0
264,000
0.05
197
1,620
21,600
10
500
568
128
7.3
Marine
Mid-Top
14.9
2.1
167,400
0.37
271
1,493
16,400
9
296
6,350
70
6.6
Bank
Mid-Bottom
55-5
3.6
86,000
0.12
60
475
11,700
10
63
63
21
7.9
West
14.0
1.9
120,800
0.13
695
1,706
26,300
30
758
1,020
263
7.5
(continued)
-------
TABLE A-15 (continued).
GO
Location
Highland Blvd
Parameter
Total Solids - %
Volatile Total Sol ids-%
COD-mg/kg
Ni tri te+Ni trate Ni trogen-mg/kg
Ammonia-Ni trogen-mg/kg
Total Phosphorus-mg/kg
1 ron-mg/kg
Cadmium-mg/kg
Zinc-mg/kg
Lead-mg/kg
Copper-mg/kg
PH
East
18.0
2.2
236,000
0.05
302
1,510
19,900
13
459
57**
117
7.4
Mid-Top
37.8
A.I
159,000
0.0ft
390
1,630
20,900
13
ft95
434
107
6.7
Mid-Bottom
33.2
4.8
380,000
0.0ft
7ft3
2,840
19,800
23
783
623
222
6.7
West
30. ft
3.6
151,200
0.05
378
1,059
21,200
19
493
587
163
7. ft
East
36. ft
3. ft
127,000
0.85
ftl»0
1,510
16,500
11
ft33
824
122
7.3
Wai
Mid-Top
ftO.8
4.1
127,800
0.0ft
394
1,728
19,800
18
554
471
528
7.4
nut Street
Mid-Bottom
36.9
5-1
243,000
0.03
572
2,660
20,800
16
826
703
198
7.0
West
27.8
3-1
185,100
0.03
451
1,144
23,300
37
749
582
112
6.7
-------
TABLE A-15 (continued).
CO
Location
Humboldt Ave.
Parameter
Total Solids - %
Volatile Total Solids - %
COD-mg/kg
Ni tri te+Ni trate Nitrogen-mg/kg
Ammonia-Ni trogen-mg/kg
Total Phosphorus-mg/kg
1 ron-mg/kg
Cadmium-mg/kg
Zinc-mg/kg
Lead-mg/kg
Copper-mg/kg
pH
East
9.8
2.0
293,800
0.10
388
1,580
19,300
13
556
643
118
6.7
Mid
31.4
3.7
122,700
0.03
241
862
12,900
16
314
563
92
7.4
West
26.7
2.0
129,000
0.03
188
830
11,000
9
259
321
64
7.5
East
38.4
3-3
119,000
0.05
168
1,147
15,200
9
365
505
93
7.0
North Ave.
Mid
38.4
2.8
112,500
0.04
332
1,230
19,700
14
357
396
78
7.2
West
20.8
1.9
195,000
0.07
22
2,000
25,000
12
492
648
141
7.0
(cont inued)
-------
TABLE A-!5 (continued).
CO
-C-
Location
Burleigh St.
Parameter
Total Sol ids - %
Volati le Total Sol ids - %
COD-mg/kg
Nitrite+Ni trate Ni trogen-mg/kg
Ammonia-Ni trogen-mg/kg
Total Phosphorus-mg/kg
i ron-mg/kg
Cadmi um-mg/kg
Zinc-mg/kg
Lead-mg/kg
Copper-mg/kg
PH
East
33.9
3-5
150,000
0.02
163
1,243
19,000
9
438
530
3
6.7
Mid
51.6
1.2
41,770
0.03
130
407
7,00
16
no
186
44
7.5
West
21.2
2.2
166,000
0.09
33^
1,504
19,000
10
^32
668
107
7.3
Hubbard Park
East
40.4
3-3
148,600
0.06
163
1,476
16,300
6
425
716
123
6.9
Mid
28.9
2.9
148,400
0.03
210
1,708
22,800
14
393
457
89
7,4
West
78.6
2.0
8,820
0.01
20
69
6,800
3
137
49
125
7.6
-------
TABLE A-15(continued).
CO
vn
Location
Parameter
Total Sol ids - %
Volati le Total Solids - %
COD-mg/kg
Nitrite+Nitrate N i trogen-mg/kg
Ammoriia-Nitrogen-mg/kg
Total Phosphorus-mg/kg
I ron-mg/kg
Cadm?um-mg/kg
Zinc-mg/kg
Lead-mg/kg
Copper-mg/kg
pH
Lincol
East
30.7
0.4
3,390
0.05
14
32
34,000
7
90
67
32
5.7
n Park- South
Mid
35.4
0.4
2,160
2.12
9
28
9,700
6
86
26
10
8.0
West
62.7
1.8
33,600
0.02
33
248
15,500
4
171
239
36
7.0
Lincoln
East
81.7
0.8
2,510
<0.01
10
26
5,300
2
37
40
5
7.1
Park-Central
Mid
78.8
0.4
577
<0.01
10
18
4,770
5
29
38
5
7.9
West
80.7
1.5
14,800
0.02
17
147
6,600
3
50
58
9
6.9
-------
TABLE A-15 (continued).
oo
Locat ion
Lincoln Park-North
Parameter
Total Solids - %
Volatile Total Solids - %
COD-mg/kg
Ni tri te+Ni trate Ni trogen-mg/kg
Ammon i a-N i t rogen-mg/kg
Total Phosphorus-mg/kg
1 ron-mg/kg
Cadmi um-mg/kg
Zinc-mg/kg
Lead-mg/kg
Copper-mg/kg
PH
East
77.4
1.0
4,370
0.01
21
4o
5,100
2
28
25
5
7.7
Mid
41.7
1 .2
35,200
0.02
3^
290
17,600
5
48
27
16
7.7
West
85,5
1.3
8,940
0.29
9
72
4,300
3
24
29
5
7.3
East
76.3
1.0
11,800
0.01
85
133
5,570
7
42
33
2
7.5
Dean Rd.
Mid
61.7
0.4
2,000
0.01
41
23
5,360
8
21
24
2
7-2
West
82.4
1.2
2,800
0.01
34
42
6,590
4
31
33
6
8.1
-------
TABLE A-16. LABORATORY ANALYSIS RESULTS - SEDIMENT SAMPLES OF JANUARY 21, 1976
oo
Parameter
pH
Oxidation-reduc-
tion potential
Total solids
Volatile sol ids
BOD20
COD
TOC
Ammonia-N
Nitrate-N
Nitrite-N
Total phosphorus
Iron
Cadmium
Lead
Copper
Zinc
Units
millivolts
%
% (dry
mg/kg
nig/ kg
mg/kg
mg/kg
mg/kg
mg/kg
mg/kg
mg/kg
mg/kg
mg/kg
mg/kg
mg/kg
wt)
(dry
(dry
(dry
(dry
(dry
(dry
(dry
(dry
(dry
(dry
(dry
(dry
wt)
wt)
wt)
wt)
wt)
wt)
wt)
wt)
wt)
wt)
wt)
wt)
Harbor
6.6
-207
16.7
19-9
**7,900
260,800
80,800
4,200
<0.l
<0.1
6,470
24,000
39
285
165
705
St. Paul
6.7
-237
29.1
11.8
16,100
160,000
61 ,900
269
<0.1
<0.1
U560
18,000
10
405
111
410
Wei
7.1
Is
-238
22.
12.
28,
180
53,
321
<0.
<0.
2,1
5
7
400
,400
300
1
1
10
21,000
11
600
122
470
Cherry
6.7
-216
20.0
15.4
21,500
220,200
70,000
620
<0.1
<0.1
1,940
18,000
10
535
102
465
State
6.9
-220
24.2
12.6
18,600
185,300
66,100
432
<0.1
<0.1
1,870
21,000
11
465
121
470
Walnut
6.9
-256
58.4
5.5
7,000
117,600
22,230
68
<0.1
<0.1
616
8,200
6.1
310
73
255
Hoi ton
6.9
-230
34.0
9.7
20,000
152,000
64 , 700
590
<0.1
<0.1
U530
18,000
11
685
112
455
-------
TECHNICAL REPORT
(Please read Instructions on the reverse
DATA
before completing)
1. REPORT NO.
EPA-600/2-79-155
2.
3. RECIPIENT'S ACCESSION"NO.
4. TITLE AND SUBTITLE
5. REPORT DATE
December 1979 (Issuing Date)
VERIFICATION OF THE WATER QUALITY
COMBINED SEWER OVERFLOW
IMPACTS OF
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
Thomas L.
Martin E.
8. PERFORMING ORGANIZATION REPORT NO.
Meinholz, Wi11iam
Harper, and Kevin
A. Kreutzberger,
J. Fay (Rexnord Inc.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Metropolitan Sewerage District of the
County of Milwaukee
P. 0. Box 2079
Milwaukee, Wisconsin 53201
10. PROGRAM ELEMENT NO.
1BC822, SOS 1, Task
11. CONTRACT/GRANT NO.
R-80*»5l8
12. SPONSORING AGENCY NAME AND ADDRESS
Municipal Environmental Research Laboratory-Gin
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
.,OH
13. TYPE OF REPORT AND PERIOD COVERED
Final. 8/76-12/78
14. SPONSORING AGENCY CODE
EPA/600/Ik
15. SUPPLEMENTARY NOTES
Project Officer: John N. English
Phone: 513/68A-7613
16. ABSTRACT
The purpose of this study was to identify the source and mechanism of the water
quality impacts in the Milwaukee River following wet weather discharges. Monitoring
surveys have demonstrated that there is often a rapid decline in DO following CSO
discharge events in the lower portions of the river. Water quality modeling of the
river with Harper's water quality model indicate that the loadings from combined
and storm sewer discharges are not sufficient to cause the observed rapid declines
in DO. The results of laboratory and field investigations indicate that bottom
sediments within the lower river are a significant sink for DO and are linked to the
rapid loss of DO. The mechanism of this rapid DO decline is the scouring of sediment
oxygen demanding materials by submerged CSO outfalls. This was determined through
measurements of instream velocities near the bottom sediments resulting from
discharges from submerged outfalls. Empirical equations were developed using multiple
regression analysis to predict the impact of sediment scouring by CSO discharges on DO
levels. An expression was also added to Harper's water quality model to provide for
time varying SOD rates which are required to simulate the high oxygen demand of
scoured sediments. This model was calibrated and verified for dry and wet weather
conditions in the study area of the river and was used to determine the DO and fecal
coliform impact which is attributable to CSO.
7.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS C. COSATI Field/Group
Combined sewers
Water pollution
*Water quality
^Mathematical models
^Sediments
Metals
"Dissolved oxygen
Sediment oxygen demand
MiIwaukee River
Urban runoff
Fecal coliforms
Suspended sol ids
13B
18. DISTRIBUTION STATEMENT
Release to public
19. SECURITY CLASS (ThisReport)
Unclass ified
21. NO. OF PAGES
202
2O. SECURITY CLASS (This page)
Unclassified
22. PRICE
EPA Form 2220-1 (»-73)
• US GOVERNMENT PfUHTMO OFFICE: 1MO -657-146/5540
------- |