WATER POLLUTION CONTROL RESEARCH SERIES
11024 EXF 08/70
Combined Sewer Overflow
Abatement Alternatives
Washington, D.C.
ENVIRONMENTAL PROTECTION AGENCY • WATER QUALITY OFFICE
-------
WATER POLLUTION CONTROL RESEARCH SERIES
The Water Pollution Control Research Reports describe the results and progress
in the control and abatement of pollution of our Nation's waters. They provide
a central source of information on the research, development and demonstration
activities of the Water Quality Office of the Environmental Protection Agency,
through in-house research and grants and contracts with the Federal, State
and local agencies, research institutions, and industrial organizations.
Previously issued reports on the Storm and Combined Sewer Pollution Control
Program:
11023 FDB 09/70
11024 FKJ 10/70
11024 EJC 10/70
11023 12/70
11023 DZF 06/70
11024 EJC 01/71
11020 FAQ 03/71
11022 EFF 12/70
11022 EFF 01/71
11022 DPP 10/70
11024 EQG 03/71
11020 FAL 03/71
11024 FJE 04/71
11024 DOC 07/71
11024 DOC 08/71
11024 DOC 09/71
11024 DOC 10/71
11040 QCG 06/70
11024 DQU 10/70
11024 EQE 06/71
Chemical Treatment of Combined Sewer Overflows
In-Sewer Fixed Screening of Combined Sewer Overflows
Selected Urban Storm Water Abstracts, First Quarterly
Issue
Urban Storm Runoff and Combined Sewer Overflow Pollution
Ultrasonic Filtration of Combined Sewer Overflows
Selected Urban Runoff Abstracts, Second Quarterly Issue
Dispatching System for Control of Combined Sewer Losses
Prevention and Correction of Excessive Infiltration and
Inflow into Sewer Systems - A Manual of Practice
Control of Infiltration and Inflow into Sewer Systems
Combined Sewer Temporary Underwater Storage Facility
Storm Water Problems and Control in Sanitary Sewers -
Oakland and Berkeley, California
Evaluation of Storm Standby Tanks - Columbus, Ohio
Selected Urban Storm Water Runoff Abstracts, Third
Storm Water Management Model, Volume 1 - Final Report
Storm Water Management Model, Volume II - Verification
and Testing
Storm Water Management Model, Volume III -
User's Manual
Storm Water Management Model, Volume IV - Program Listing
Environmental Impact of Highway Deicing
Urban Runoff Characteristics
Impregnation of Concrete Pipe
To be continued on inside back cover...
-------
Combined Sewer Overflow
Abatement Alternatives
Washington, D.C.
by
ROY F.WESTON, INC.
Environmental Scientists and Engineers
West Chester, Pennsylvania
for the
WATER QUALITY OFFICE
ENVIRONMENTAL PROTECTION AGENCY
Program No. 11024EXF
Contract No. 14-12-403
August 1970
For sale by the Superintendent of Documents, U.S. Government Printing Office, Washington, D.C. 20402 - Price $2
Stock Number 5501-0102
-------
EPA/WQO Review Notice
This report has been reviewed by the Water Quality Office
of the Environmental Protection Agency and approved for
publication. Approval does not signify that the contents
necessarily reflect the views and policies of the
Environmental Protection Agency, nor does mention of trade
names or commercial products constitute endorsement or
recommendation for use.
-------
ABSTRACT
Objectives of the project were: 1) define the characteristics of urban runoff; 2) investigate
the feasibility of high-rate filtration for treatment of combined sewer overflow; and 3)
develop and evaluate alternative methods of solution.
Investigative activities included: review of pertinent reports and technical literature; field
monitoring of combined sewer overflows and separated storm water discharges at three
sites; laboratory studies of ultra-high-rate filtration of combined sewer overflow;
hydrological analysis; and evaluation of feasible alternatives (based on conceptual designs,
preliminary cost estimates, and other factors).
Reservoir Storage, Treatment at Overflow Points, Conveyance Tunnels and Mined Storage,
and Sewer Separation were the approaches considered sufficiently promising for detailed
evaluation. Tunnels and Mined Storage with treatment at the Blue Plains plant and at
Kingman Lake after subsidence of the storm is recommended. Estimated capital costs (based
on the 15-year, 24-hour storm) are $318,000,000 (ENR=1800) with annual operation and
maintenance costs of $3,500,000. This approach also was preferable to the others on the
basis of systematic evaluation of reliability, flexibility, public convenience and other
non-quantifiable factors.
This report was submitted in fulfillment of Contract 14-12-403 (11024 EXF) between
the Environmental Protection Agency-Water Quality Office and Roy F. Weston, Inc.
-------
CONTENTS
Section Page
ABSTRACT iii
CONTENTS v
FIGURES vii
TABLES ix
I CONCLUSIONS 1
II RECOMMENDATIONS 3
Ml INTRODUCTION 5
General 5
Project Objectives and Scope 6
IV PROBLEM DEFINITION 7
Description of Study Area 7
Description of Present D.C. Sewer System 12
D.C. Sewage Treatment Facilities 19
Peak Rate and Volume of Overflow 21
Field Monitoring Program 29
Impact of Storm Water Discharges 42
V INVESTIGATION OF POTENTIAL ABATEMENT MEASURES 47
Review of Approaches Tried at Other Cities 47
Ultra-High-Rate Filtration 55
VI DEVELOPMENT OF FEASIBLE ALTERNATIVES 57
Alternative Approaches 57
Application of Alternative Approaches to D.C. 64
VII EVALUATION OF FEASIBLE ALTERNATIVES 79
General 79
Cost Ana lysis 80
Impact on Water Quality 99
Comparison of Non-Quantifiable Factors 106
Indicated Appropriate District-wide Solution 112
VIII ACKNOWLEDGMENTS 115
-------
CONTENTS
(Continued)
Section Page
IX REFERENCES 117
X APPENDICES 121
Appendix A - Geological and Other Natural Conditions 125
Appendix B - Hydrologic Analysis 137
Appendix C - Monitoring Equipment 139
Appendix D List of Previous Reports on Sewer System 145
Appendix E Results of Monitoring Program 157
Appendix F - Ultra-High-Rate Filtration 203
Appendix G Kingman Lake Project 241
VI
-------
FIGURES
Page
1 District of Columbia 8
2 Generalized Geologic Map 11
3 Combined Sewer Study Area 13
4 Combined Sewer Study Area and Combined System Separation Schedule 18
5 Hyetograph for Various Rainfall Frequencies 22
6 Standard Infiltration Capacity Curves for Pervious Surfaces 24
7 Determination of Infiltration Offset - 5-Year Rainfall Frequency - Pervious 25
Residential Area
8 Determination of Point of Intersection of Infiltration Capacity and 26
Precipitation Rate - 5-Year Rainfall Frequency - Pervious Residential
Area
9 Combined Sewer District B-4 32
10 Combined Sewer District G-4 33
11 Good Hope Run Sewer District 34
12 Monitoring Equipment 37
13 Annual Pollution Loads Discharged to Streams from Storm Water 44
14 Storage Reservoir Schematic (Typical) 58
15 Treatment at Overflow Points - Treatment Facility Schematic (Typical) 60
16 Tunnels and Mined Storage in Anacostia River Basin 65
17 General Tunnel Location - Northeast Boundary Trunk Sewer 66
18 Anacostia River Tunnel Profile 68
19 Maximum Use of Storage Reservoirs in Anacostia River Basin 69
20 Treatment at Overflow Points in Anacostia River Basin 71
21 Tunnels and Mined Storage in Upper Potomac-Rock Creek Basin 73
22 Rock Creek Tunnel Profile 74
VII
-------
FIGURES
(Continued)
Page
23 Upper Potomac Tunnel Profile 75
24 Maximum Use of Storage Reservoirs in Upper Potomac Rock Creek Basin 76
25 Treatment at Overflow Points in Upper Potomac - Rock Creek Basin 78
26 Total Construction Costs of Storage Reservoirs and Appurtenant Equipment 81
27 Unit Construction Costs of Tunnels 83
28 Total Construction Costs of Mined Storage 85
29 Total Construction Costs of Treatment Plants versus Capacity 86
30 Investment versus Design Storm Return Frequency for Alternative 93
District-wide Approaches
31 Percent Reduction in Pollution Load from Various Storms vs Alternative 101
Design Events
32 Effect of Investment on Discharge of BOD Loading for Various Storm 103
Return Frequencies
33 Seasonal Probability of Intense 24-Hour Rainfall 105
VIM
-------
TABLES
Page
1 Population of the District of Columbia 7
2 Combined Sewer Districts Scheduled for Separation after 1975 14
3 Deficiency in Sewer Capacities 15
4 Daily Pollution Loads from Blue Plains Plant 19
5 Combined Sewer Overflow Characteristics 27
6 Physical Descriptions of the Monitoring Sites 31
7 Storms Monitored 35
8 Major Storms Missed 36
9 Comparison of Characteristics of Combined Sewer Flow and Separated 40
Storm Sewer Flow
10 Expected Pollution Loads from Combined and Separated Storm Sewers 45
in District of Columbia
11 WQO Combined Sewer Overflow Treatment Projects 53
12 Maximum Use of Storage Reservoirs - Estimated Capital Costs 38
13 Treatment at Overflow Points - Estimated Capital Costs 83
14 Conveyance Tunnels and Mined Storage - Estimated Capital Costs 90
15 Sewer Separation Costs - Estimated Capital Costs 91
16 Comparison of Total Project Costs 92
17 Summary of Annual Operating Costs for Maximum Use of Storage 94
Reservoirs for 15-Year, 24-Hour Storm
18 Summary of Annual Operating Costs for Treatment at Overflow Points 95
for 15-Year, 24-Hour Storm
19 Summary of Annual Operating Costs for Conveyance Tunnels and Mined 96
Storage for 15-Year, 24-Hour Storm
20 Comparison of Annual Operating Costs Based on 15-Year, 24-Hour Storm 97
IX
-------
TABLES
(Continued)
Page
21 Comparison of 24-Hour Pollution Discharges from Various Return 102
Frequency Storms for Various Design Frequencies
22 Evaluation of General Factors 110
23 Estimated Capital Costs of Indicated Appropriate Solution 114
-------
SECTION I
CONCLUSIONS
1. The indicated appropriate solution to the problem of combined sewer overflows in
Washington is one that provides a network of large tunnels and mined areas to convey
and store overflow, with treatment of stored overflow at both the Blue Plains plant
and at a facility near Kingman Lake.
2. The information developed in this study further points to a design capacity equal
to the overflow from the 15-year, 24-hour storm. This design would provide effective
treatment of more than 99 percent of the long-term-averaged annual volume of
overflow that now contributes significantly to the pollution of the Potomac and its
tributaries. Even with this design, the 25-year, 24-hour storm would result in a total
24-hour BOD loading of 70,000 pounds, well over four times the recommended
maximum allowable daily loading from all BOD sources in the entire metropolitan
Washington area.
3. A network of tunnels and mined storage, plus an additional facility near Kingman
Lake, represents the least-cost alternative designed for the 15-year, 24-hour storm.
It would cost $318,000,000 (ENR=1800) to construct and $3,500,000 per year to
operate and maintain this system; in contrast, a program of complete sewer separation
would cost about $610,000,000 (ENR=1800).
4. An evaluation of non-quantifiable factors such as reliability, flexibility, land
requirements, public convenience, implementation, and solids removal suggested that
an approach incorporating tunnels and mined storage offers more advantages than
an approach based on storage reservoirs, treatment at overflow points, or sewer
separation.
5. Based on the results of a field monitoring program and a detailed rainfall-runoff
analysis, the overflow from combined sewers within Washington, D.C. discharges each
year the following pollutant loads to the Potomac and its tributaries:
Biochemical Oxygen Demand 3,200,000 pounds
Total Phosphorus 500,000 pounds
Total Nitrogen 500,000 pounds
Suspended Solids 59,000,000 pounds
6. Overflows from combined sewers do not occur as continuous steady discharges, but
rather as slug loadings. This characteristic, in combination with the long effective
residence times of estuarine waters, explains the particularly serious impact combined
sewer overflows have on water quality. For example, the 24-hour BOD load expected
in the 616 million gallon overflow from the 2-year, 24-hour storm (for the existing
sewer system) is 160,000 pounds, nearly ten times the recommended maximum
allowable daily loading from all BOD sources in the entire metropolitan Washington
area.
7. Any abatement alternative utilizing storage or in-line treatment will remove practically
all of the 59,000,000 pounds of suspended solids now discharged yearly as overflow
to the Potomac. The resulting quantity of sludge (30,000 tons per year on a dry
-------
basis) is greater than the quantity of sludge generated at the Blue Plains plant. Besides
this, the generation of this sludge is concentrated in the 50 to 60 overflows per
year, rather than coming on a uniform, daily basis as at Blue Plains. In light of
the general problems of solid waste disposal in the Washington area, a high level
of sophistication and a high degree of planning will be required to handle the combined
sewer sludge as is employed to handle the sewage treatment plant sludge. The handling
and disposal of this sludge will probably cost $750,000 per year.
8. A limited six-month field monitoring program of combined sewer overflow from two
combined sewer drainage basins and separated storm sewer discharge from one
separated sewer drainage basin during storm conditions suggested the following mean
values for certain significant waste constituents:
Combined Separated Storm
Sewer Flow Sewer Flow
Biochemical Oxygen Demand 71 mg/L 19 mg/L
Settleable Solids 229 mg/L 687 mg/L
Fecal Coliform 2,400,000/100 ml 310,000/100 ml
Total Phosphate 3.0 mg/L 1.3 mg/L
Values of BOD as high as 470 mg/L in combined sewer flow and as high as 90 mg/L in
separated storm sewer flow were detected during the monitoring program.
9. With reference to the laboratory investigations of ultra-high-rate filtration, the
following conclusions are drawn:
a. Tri-media filtration at rates less than 10 gpm/sq.ft. provides a satisfactory
effluent.
b. Ultra-high-rate filtration (15 gallons or more per minute per square foot) of
combined sewer overflows is not technically feasible for upflow filtration through
a garnet bed, because of poor effluent quality.
c. The fiberglass filter operated successfully at filtration rates of 15-30 gpm/sq.ft.,
with removals of 90 percent suspended solids and 70 percent non-soluble BOD.
d. The addition of flocculant aids did not improve the removal characteristics of
the fiberglass filter.
e. Soluble BOD^ was not significantly reduced by the addition of low dosages
of activated sludge to the filter influent.
f. The economic feasibility of the fiberglass medium will depend upon the extension
of the useful life of the medium and the improvement of backwash techniques.
10. The selection of the design storm event is properly made only in the context of
basin-wide quality water management. Many factors beyond the scope of this study
such as the impact of wastewater discharges outside the District of Columbia low-flow
augmentation, etc., have too great an influence to be ignored.
-------
SECTION II
RECOMMENDATIONS
Based on the findings and conclusions developed in the course of this combined sewer
overflow study, the following actions are recommended.
1. Discontinue the current sewer separation program and develop pollution abatement
programs for both the combined and separated sewer areas if the pollutional
characteristics of storm water determined in the current study are confirmed in other
areas of the District.
2. Proceed with engineering and construction of conveyance tunnels, mined storage, and
a treatment facility near Kingman Lake, somewhat like the facilities described in
reference (6), in conjunction with the following:
a. Conduct the sub-surface investigations necessary to confirm the assumed bedrock
characteristics.
b. Initiate a long-term monitoring program using monitoring equipment developed
for the current project to obtain long term comparative and detailed information
to define long-term impact of pollution loads on required treatment facilities
and receiving waters.
c. Confirm the selection of the design storm frequency in the context of basin-wide
water quality management.
3. Implement additional EPA-WQO sponsored laboratory and full-scale development
studies on the use of fiberglass as a filter medium with the following objectives:
a. To develop an improved filter bed design with respect to density gradation, depth,
and combination with granular media.
b. To refine the backwashing techniques, including underdrain design, stagewise
removal of backwash effluent, and regeneration of the fiberglass filter medium.
c. To optimize design parameters.
4. Re-evaluate, under EPA-WQO sponsorship, the validity and accuracy of the present
conventional approach to the determination of sewer flow rates (combined-sewer
areas) during the surge period associated with each storm.
-------
SECTION III
INTRODUCTION
General
Recent efforts stimulated by the Federal Water Pollution Control Act of 1956 (amended
in 1961, 1965, 1966, and 1970) have brought to light the significance of combined sewer
overflows as a source of pollution.
Most United States cities today are served both by combined sewers and by separate
sanitary and storm sewer systems. As determined from a 1967 survey sponsored by the
Water Quality Office of the Environmental Protection Agency, approximately 29 percent
of the total sewered population of the United States is served by combined sewer systems.
Approximately three percent of the total annual flow of sewage and as much as 95 percent
of the sewage produced during periods of rainfall is carried with combined sewer overflows
to the surface waters.
The District of Columbia follows this pattern, with an area of approximately 20 square
miles (one third of the total area of the District) being served by combined sewers. The
hydraulic capacity of the system is often exceeded during periods of precipitation, and
raw sewage mixed with surface runoff is discharged to the watercourses of the District.
The Potomac Estuary is polluted and continues to experience problems with low dissolved
oxygen, excessive algal growths, sediments, high concentrations of fecal bacteria and
repulsive floating matter. All of these problems, except sediments, are complicated by
combined sewers. Although combined sewer overflow adds to the sediment load, the
primary source of sediment is the heavy silt load included in the runoff from areas with
significant agricultural and construction activity. In addition to the effects of combined
overflow on the Potomac River, overflows into Rock Creek detract from the natural and
recreational features of this small, scenic stream flowing through the District.
In 1957, the District prepared a separation schedule for conversion of all combined sewers
into separate storm and sanitary systems. This conversion, however, would not be
completed until after the year 2000. The cost of this program is extremely high, estimated
at $610 million (ENR=1800), and budgeting and other problems have delayed progress.
In the meantime, the overflows from unseparated sewers will continue to contribute to
pollution problems in the Potomac Estuary.
-------
Project Objectives and Scope
The primary objectives stipulated for this study were to: define the problem of combined
sewer overflows within the District of Columbia; investigate the feasibility of high-rate
filtration (greater than 15 gallons/square foot/minute) for treatment of combined sewer
overflows; and study alternative methods of solution to the problem
The general scope of the project is outlined below as taken from the Statement of Work
of Contract No. 14-12-403 (between the Federal Government and Roy F. Weston, Inc.)
and as stated in the latter's proposal "Study of Pollution Abatement from Combined Sewer
Systems" dated 21 November 1967:
1. Develop a quantitative definition of the combined sewer pollution problem by
review of reports, studies, and data concerning water quality and rainfall
intensities, surface runoff, frequency of occurrence, and hydrologic data pertinent
to assessing methods for controlling combined sewer system pollution.
2. Collect data and subsequent water quality analyses of continuous rainfall and
sewer flow measurements made at three selected sites.
3. Evaluate by laboratory research the possibility of removing flocculated solids
and associated BOD by high-rate filtration of combined sewer overflows.
4. Define the technical and economic feasibility of flocculation and high-rate
filtration as a means of treating combined sewer overflows.
5. Investigate other selected practical solutions to provide a meaningful evaluation
of the relative merits of high-rate filtration and to provide a comprehensive study
of the combined sewer problem of Washington, D.C.
6. Present generalized capital cost and annual cost estimates for alternatives which
appear to be technically feasible for the Washington, D.C. system.
7. Develop technical, economic, and operational comparisons of practical
alternatives.
8. Further develop the general formulation and application of a methodology of
analysis for determining feasible solutions for eliminating combined sewer
pollution from municipal systems.
-------
SECTION IV
PROBLEM DEFINITION
Description of Study Area
General Conditions
The District of Columbia lies largely along the "Y" formed by the junction of the Anacostia
and Potomac Rivers, extending eleven miles along the northeasterly side of the Potomac
and straddling the Anacostia for six miles. Figure 1 illustrates the general area. The average
flow of the Potomac River is 13,300 cfs, whereas the Anacostia averages 150 cfs. The
total land area of the District is approximately 61 square miles, and the water area within
the District is some seven square miles. The District is the center of a metropolitan area
which includes parts of Maryland and extends across the Potomac River into Virginia.
The central and older portion, comprising approximately one-third the total land area
of the District, is the principal area of interest for the present investigation. This area
rises gradually from the confluence of the Potomac and Anacostia Rivers to the encircling
hills, and was originally drained by several sizable streams discharging to these two rivers.
The metropolitan Washington area has experienced a phenomenal growth in population
since 1930. Although the population of the area surrounding the District continues to
increase significantly each year, the District itself may have reached its peak population,
for some time, in 1950. Out-migration since then has outweighed population increases
from in-migration and the birth-less-death increases. It is uncertain whether or not
population will continue to decrease or increase; nevertheless, existing and planned land
use will probably impede any drastic increases in population. The population of the District
from 1900 to 1970 is given in Table 1. The possible future populations are indicated
in a February, 1957 Board of Engineers report to the District on improvements to the
sewerage system (1); the report projected a population of 1,125,000 in the year 2000.
However, the same report projected 1,010,000 population in 1970.
Table 1
Population of the District of Columbia
Year Population Year Population
1900 278,718 1940 663,091
1910 331,069 1950 802,178
1920 437,571 1960 763,956
1930 486,869 1970 746,1691
1 Preliminary value from 1970 Census.
-------
00
DISTRICT OF COLUMBIA
1500 0 1500 3000 4500 4000
SCALE IN FEET
FIGURE 1
-------
District activities largely concern government offices, public buildings, and related activites;
there are no large industries within its borders. The many large government buildings in
the District cover extensive areas and generate heavy concentrations of employees.
Considerable areas of the District are occupied by parks, recreational areas, institutional
grounds, and cemeteries; for example, in 1966, approximately 7,830 acres (20 percent
of the total land area of the District) were occupied by parks and playgrounds.
The District had a very low ratio of runoff to rainfall when the original combined sewer
system was installed, primarily because of the numerous parks and other unpaved areas.
However, as the population density increased, as more and more buildings were erected,
as street pavings grew wider and wider, and as parking areas were paved, the rainfall-runoff
coefficient has increased tremendously. This has contributed to the present inadequacy
of all storm water facilities in the central portion of the District.
The land surface in the District varies in elevation from the low areas (elevations of about
4 feet, USGS Datum) adjacent to the Potomac River and in East and West Potomac Park,
along the Washington Channel, and along the Anacostia River, to elevations as high as
410 feet near the intersection of Wisconsin and Nebraska Avenues NW. The area west
of 16th Street NW, in Rock Creek Park, is quite rugged, with a number of places where
the land rises to elevations of 300 feet and more.
Southeast of the Anacostia River and east of the Potomac River in the southern park
of the District, the land surface is low and flat along the rivers, as in the area of Boiling
Air Force Base and the Anacostia Naval Air Station. The remainder of this southeast
part of the District is hilly, reaching elevations of 200-300 feet in several areas.
Except for the low-lying lands along the rivers, the terrain of the District favors the design
of gravity sewers. The District generally drains well because of the good slopes that can
be obtained. There are a few bowl-like depressions surrounded by higher territory; however,
except for the area in the vicinity of 5th and Ingraham Streets NW, none is subject to
severe flooding in times of heavy rains. Some basement flooding has been experienced
from backup of storm water in surcharged sewers. This problem is a function of the sewer
system itself and will be discussed in greater detail in a subsequent section. Surface water
draining from the land areas of the District is carried off by the Potomac River, the
Anacostia River, Rock Creek, Foundry Branch, Oxon Run, and their minor tributaries.
Geological Features
Information on the geologic and other natural conditions is essential to a complete
evaluation of the various possible approaches to the abatement of pollution caused by
the inadequacies of combined sewers, including the feasibility of tunneling beneath the
City of Washington to provide for storage of combined sewer overflows.
Data were obtained from published geologic reports, from engineering reports prepared
under the auspices of the Metropolitan Area Rapid Transit Authority, and from verbal
communication with personnel of the Authority. The published geologic data are fairly
broad in coverage, yet sufficiently detailed to provide information helpful in forming
conclusions in regard to the feasibility of tunneling. The engineering reports are extremely
detailed and provide an abundance of data concerning the soils, bedrock, and rock
mechanics in certain restricted areas developed for the purpose of underground tunneling
for rapid transit. Although somewhat restricted as to area-wide application, the data
included in these reports provide enough coverage to be representative of the general
Washington area.
-------
The District of Columbia lies within portions of two physiographic provinces; the
southeastern portion is located within the Coastal Plain province, which consists of
relatively flat-lying sediments overlying deep bedrock, and the northwestern portion is
in the Piedmont province, which in general is characterized by a thin layer of overburden
covering crystalline bedrock. The Fall Line separating the two provinces extends roughly
southwest from Blair Park in the northeast through Farragut Square and on toward the
Pentagon. Figure 2 presents the generalized geology of the District of Columbia.
Previous sub-surface investigations in the District have resulted in grouping the materials
into five major categories: bedrock, Cretaceous sediments, Pleistocene terrace deposits,
recent river alluvium, and drainage channels and man-made fills. These major categories
of materials in various parts of the District are found in the following five vertical profiles:
1. Recent alluvium over bedrock or Pleistocene terrace deposits.
2. Overburden of Pleistocene terrace and Cretaceous coastal plain soils above deep
bedrock.
3. Comparatively thick cover of Pleistocene terrace and Cretaceous coastal plain
soils above deep bedrock.
4. Thin to moderately thick cover of Cretaceous coastal plain materials above
decomposed rock and bedrock.
5. Relatively thin cover of man-made fill and decomposed rock over bedrock at
shallow to moderate depths.
Geological and related natural conditions in the Washington, D.C. area and the implications
for tunneling are described in more detail in Appendix A.
10
-------
DISTRICT OF COLUMBIA
GENERALIZED GEOLOGIC MAP
LEGEND
PRIMARILY ALLUVIUM AND ARTIFICIAL FILL
UNDIFFERENTIAIED CRETACEOUS, TERTIARY AND
QUATERNARY FORMATIONS CONSISTING OF
CLAY, SILT, SAND AND GRAVEL
RIVER TERRACE DEPOSITS CONSISTING OF
GRAVEL, SAND AND LOAM
iS-I-S-a PRIMARILY CRETACEOUS CLAY, SAND AND GRAVEL
MASSIVE LIGHT GRAY, COARSE -TEXTURED GRANITE
GRANITE GNEISS WITH LAYERS OF SCHISTOSE
GRANITES, GNEISS AND SILICEOUS MICA SCHISTS
1500 0 1500 3000 4500 6000
SCALE IN FEET
FIGURE 2
-------
Description of Present D.C. Sewer System
Historical Development
Several studies have been made of the District of Columbia sewerage system and sewage
treatment facilities. A list of pertinent reports and related technical articles, which were
reviewed to obtain detailed information relevant to the present pollution problem, is
presented as Appendix B.
The installation of culverts and drains in the District began as early as 1810. In 1840,
the first piped water supply to a few homes began;, it was followed by a few sewers
discharging to the nearby culverts. This was the beginning of the combined sewer system
in Washington, D.C. With the introduction of the Potomac River water supply in 1859,
the extent of the combined sewer system discharging into nearby watercourses increased
very rapidly, and by 1874 there were approximately 80 miles of sewers. The sewers drained
into the principal watercourses and thence to the Potomac River. The discharges to the
watercourse from the combined sewers resulted in pollution of the streams and created
unsanitary conditions along the waterfront.
Upon the recommendations of the Hering-Grey-Sterns Report, a comprehensive system
of sewer construction began in 1890. Interceptor sewers were then installed to collect
all of the dry-weather flows and some surface-water runoff from the combined sewers
for conveyance to an outfall sewer for ultimate discharge into the deep water of the
•Potomac River.
Through 1890, the sewers constructed in the District were combined sewers. Subsequently,
however, the approach generally adopted was that combined sewers were tolerable in the
areas already sewered but that any additions (new areas of the District) must be separated
storm and sanitary sewers. By 1929, the construction of separated sewer systems in the
new, outlying areas was common practice.
Existing Sewer System
The present District sewer system is designed to serve an area of approximately 725 square
miles, comprising the entire District of Columbia and adjoining areas in Maryland and
Virginia. The U. S. Congress authorized the District to serve the upstream adjacent areas
of the two states so that the streams flowing through the District could be protected
from pollution.
Approximately 12,000 acres of the sewage-producing areas of the District are served by
the existing combined sewer system Substantially all of this acreage is in the central part
of the District. The remainder of the District and substantially all of the adjoining areas
of Maryland and Virginia are served by separated sewers. The existing sewer system is
shown in Figure 3. Table 2 lists the acreage of the combined sewer districts scheduled
for separation after 1975.
The District of Columbia has been divided into some 93 sewer drainage districts, arranged
in 11 groups under the names of the principal sewers to which the districts are tributary.
A tabulation of the statistics pertinent to these districts is presented in Chapter 6 of
the investigative report of June 1955 (2). The existing and required capacities of the
intercepting sewers are described in Chapter 16 of the same report.
12
-------
^M»U*ND\ PtiNCi_r,'\ Gtoasts COUNT'
\«._ . . •• ', DIStRlCT OF COLUMUA
\ - yjs
» .;V*
';i*l
jg^i5-. JjJ| CUPPER ANACOSTIA HIVES,.''
-------
Table 2
No.
A-10
A-11
A-12
B-3
B-4
B-5
B-6
B-7
C-23
C-24
C-25
C-26
C-28
C-29
D-4
D-5
D-6
D-7
D-8
D-9
D-10
D-11
D-12
D-13
D-14
D-15
D-16
D-17
E-2
E-3
F-1
G-2
G-3
G-4
G-5
G-7
Name
Combined Sewer Districts
Scheduled for Separation after 1975
Net Area
37th. St.-Georgetown
Georgetown
KSt. -Wisconsin Ave.
Q Street
Q Street -31st. St.
Olive Street 29th St.
MSt. 27th. St.
28th. St. - Wisconsin Ave.
Kalorama Circle, East
Kalorama Circle, West
Slash Run
N St. 25th. St.
KSt. -Penna. Ave.
I St. - 22nd. St.
Piney Branch
Oak St. - Mt. Pleasant
Ingleside Ter.
Park Road
Lament Street
Ken yon Street
Irving Street
Quarry Road
Ontario Road
BiltmoreSt.
Belmont Road
Mass. Ave. 24th Street
Northwest Boundary
26th. St. - M St.
Easby Point
Trunk Sewer
Tiber Creek
Northeast Boundary
Barney Circle
14th. St. -Penna. Ave.
12th. St.-9th. St.
6th. St. 7th. St.
(acres)
19
183
32
8
105
14
35
13
8
14
417
12
20
95
2,175
26
18
17
17
17
76
36
22
21
44
70
534
6
518
375
1,000
3,728
41
252
151
121
Drainage
Basin
Upper Potomac
Upper Potomac
Upper Potomac
Rock Creek
Rock Creek
Rock Creek
Rock Creek
Rock Creek
Rock Creek
Rock Creek
Rock Creek
Rock Creek
Rock Creek
Rock Creek
Rock Creek
Rock Creek
Rock Creek
Rock Creek
Rock Creek
Rock Creek
Rock Creek
Rock Creek
Rock Creek
Rock Creek
Rock Creek
Rock Creek
Rock Creek
Rock Creek
Upper Potomac
Anacostia
Anacostia
Anacostia
Anacostia
Anacostia
Anacostia
Anacostia
Source: Board of Engineers (Greeley, S.A. et al), "Report to District of
Columbia Department of Sanitary Engineering on Improvements
to Sewerage Systems, " February 1957.
14
-------
There are three principal interceptors which serve the major portion of the combined
sewer area: Upper Potomac Interceptor, Rock Creek Main Interceptor, and East Side
Interceptor. All three serve large areas of separated systems along the upstream reaches
and essentially are separate trunk sewers in these areas. They become interceptors in the
true sense in the downstream reaches in the combined sewer areas, and convey this flow
through an outfall sewer to the sewage treatment plant at Blue Plains.
The District sewer system is generally in good physical condition; however, many of the
combined and separated sewers and certain intercepting sewers do not have adequate
capacity. As development in the combined-sewer area of the District (39 percent of the
land area in 1957) increased, it was noted that the hydraulic capacity of the system was
exceeded during periods of precipitation. In order to prevent local flooding and the spilling
of sanitary sewage onto the surface of the ground, overflow structures and interceptor
chambers were built to relieve the excessive sewer flow by discharging it directly into
the natural watercourses.
In the past, the District has conducted detailed studies to evaluate the contribution to
the problem of pollution at these overflow structures. Nine automatic,
continuous-recording, depth-of-flow gauges were installed at various critical storm water
points to obtain data on the frequency and duration of overflows. It was found that
some of these overflow structures discharge as many as 40 to 50 times a year, while
others discharge only 2 to 3 times a year, depending on the drainage area characteristics
and the available interceptor capacities.
One of the most significant and comprehensive studies was the 1957 investigation of the
Board of Engineers. The Board of Engineers study indicated that many of the combined
and separate sewers were inadequate by generally accepted design standards. Considerable
surcharging occurred in the combined sewers, and there were excessive overflows of mixed
sewage and storm water to the streams, even including some overflows during dry weather.
It was evident that certain interceptors needed relief.
The situation relative to the inadequate capacities of various trunk sewers is summarized
in Table 3. This gives a picture of the magnitude of the problem. The required capacities
were based on conditions of development estimated for the year 2000.
Table 3
Deficiency in Sewer Capacities
Percent of
Total Length Present Capacity
Deficient in as Percent of
Description of Sewer Capacity Required Capacity
Upper Potomac Interceptor 98 40
Rock Creek Main Interceptor 77 54
East Side Interceptor 69 71
Piney Branch Trunk Sewer 100 56
Northwest Boundary Trunk Sewer 92 57
Slash Run Trunk Sewer 85 53
B St.-New Jersey Avenue
Trunk Sewer 98 49
Easby Point Trunk Sewer 93 47
Tiber Creek Trunk Sewer 70 50
Northeast Boundary Sewer 86 54
15
-------
To evaluate the problem of pollution of streams attributable to the overflows, the Board
of Engineers collected data on the frequency and duration of overflows at critical points
for a period of one year. It was found that the average number of overflows per month
at the various locations ranged from 5 to 16.8 during the summer months, and from
3.8 to 4.7 during the winter months. The average duration of overflows per month ranged
from 24 to 110 hours during the summer months and from 26 to 38 hours during the
winter months.
Sewer Separation Programs
Studies by Eddy, Gregory and Greeley in 1933 and by Sherman and Horner in 1935
recommended separation of the combined sewer system in several smaller areas of the
city. However, it was not until the 1957 Board of Engineers study that a unified program
for the elimination of combined sewers was presented.
The Board of Engineers outlined several alternative plans for abating or eliminating the
pollution that results from combined sewer overflows and from inadequate capacities of
combined sewers and interceptors. Two of the alternatives deemed to have the greatest
merit were those designated as Project A and Project C. Both of these alternatives require
the conversion of some combined sewers into separate sewers.
Project A was defined to convert all combined sewer areas to the separated system and
to provide interceptors to carry all sewage to the water pollution control plant. At the
completion of Project A, there would no longer be any overflows of untreated sewage
into the river. However, no appreciable reduction of pollution in the river would result
until substantially all of the work outlined in Project A was completed. Minimum
construction time required to complete Project A was optimistically estimated at twenty
years, and the estimated cost for this project was 238 million dollars at 1957 prices.
Under Project C, only 10 percent (approximately) of the combined sewer area was to
be separated initially, but new interceptors would be provided to carry additional quantities
of mixed sewage and storm water flows to the pollution control plant. Wet-weather
discharges to the river would occur even with these enlarged interceptors, but only during
heavier rainfall and on a scale substantially below present volumes. It was estimated that
Project C could be completed in ten years or less at an estimated cost of 72 million
dollars at 1957 prices. Between July 1957 and July 1970 the Engineering News Record
Construction Cost Index has increased from 725 to 1414, an increase of 94 percent; thus,
in terms of 1970 dollars, Project C costs would be approximately 140 million dollars.
The Board had recommended Project C over other alternatives because it was the least
expensive and could be completed in about ten years, and because the benefits (in terms
of reduction of pollution) could be realized as the work progressed.
The District of Columbia approved the Board's recommendations and later modified Project
C, the principal change being to design a two-level conduit for the Upper Potomac
Interceptor Relief Sewer and two force mains from the Potomac River Sewage Pumping
Station to carry the flow to the sewage treatment plant. One reason for conveying the
flow in two conduits was to prevent a complete breakdown of the system in case of
any problems. Another factor was the advantage, from the long-range standpoint, of being
able to separate the flow of sewage from the combined flow of sewage and storm water.
By this modification, all sewage originating upstream of Georgetown (in Maryland, Virginia,
16
-------
and the District) would be handled by an interceptor serving only separated system areas.
No overflow structures would be provided in this interceptor, and all the facilities could
be so designed that most of the objectionable overflows would be eliminated at an early
date.
As the sewers in the older section of the city are gradually converted to a separated
system under the Storm Sewer Separation Program the two-line interceptor would have
the built-in capacity to carry the entire flow of sanitary sewage to the treatment plant.
Eventually, Project C would achieve the same result as Project A but at reduced cost.
When the work outlined under modified Project C is completed, there will be no overflows
to the streams until the average sewage flow is diluted at least five times by storm water.
Work outlined under the separation program has been initiated, and conversion, until
recently, has kept pace with the schedule. Budgeting problems have delayed the progress
of this program. Figure 4 shows the District's present combined sewer separation schedule.
A large segment of the central section is not scheduled for separation until after the year
2000, but it is possible that areas listed to be separated after 2000 could be acted on
sooner, depending upon other improvement programs and the availability of funds.
Nevertheless, the problem of combined sewer overflows still exists. In 1961, there were
approximately 86 permanent built-in storm water overflow structures (some combined
sewer districts have more than one overflow structure) that discharge to the watercourses.
A few of these overflow structures have been plugged since then, but about 60 are still
operative, and a mixture of combined sewer overflows and storm water discharges through
them with each rainfall.
17
-------
00
\ I
i i t.-'/~ """"• •""
i »• >t- / '">•" -«'" ""
DISTRICT OF COLUMBIA
COMBINED SEWER STUDY AREA AND
COMBINED SYSTEM SEPARATION SCHEDULE
LEGEND
POtlTICAl BOUNDARY
COMBINED SEWERS
1 SANITARY SEWERS
, OTHER EXISTING SEWERS
(COMBINED AND SEPARATE |
SEWER DISTRICT NUMBER
• LIMITS OP COMBINED-SEWER SYSTEM IN 1970
. SEWER DISTRICT BOUNDARIES
PUMPING STATIONS
COMPLETION
AREA DATE
I I BEFORE 1973
IS 1985
E3 2000
E2 AFTER 2000
COMBINED SEWER OVERFLOW DISTRICTS |1975)
UPPER POTOMAC ROCK CREEH
SAUL
37TH ST.-O1OIOITOWN
OEOROETOWN
K ST.—WISCONSIN AVI
M ST -17TH ST
JITH ST-WISCONSIN AV
SLASH IUN(PANn
PINET ItANCH
D INOLESIDE TEI.
D PAU ID.
D LAMONT ST.
D KENYON ST
D 0 IRVING ST
D 1 QUAIIT ID
D 1 ONTARIO tO
E EAUY POINT
ANACOSTIA RIVER
NO NAMt
1-3 I ST.-NEW JEIIEY AVI. TIUNK MWI« 0/1 FOITKM
F.I TIIER CREEK
O-3 NORTHtAST IOUNDAIT
O-l IARNEY CIRCLE
Q.i UTH ST.-PENNA. AVE
O-J 13TH IT.-VTH }T.
Q-7 6TH ST.—7TM IT
0 1500 30C9 4500 6000
SCALE IN FEET
FIGURE 4
-------
D.C. Sewage Treatment Facilities
The first primary treatment plant to serve the District of Columbia was put into operation
in 1938, and secondary treatment facilities were added by 1959. The present treatment
facilities consist of: a pumping station; screening and grit removal units; primary
sedimentation, aeration, and secondary sedimentation tanks; chlorination facilities; effluent
outfall into the river; and sludge heating, thickening, digesting, elutriating, and vacuum
filtration facilities. The secondary (biological) treatment process is modified aeration.
The most recent improvements and additions to the plant include: secondary treatment
by modified aeration and sedimentation; new thickeners and heat exchangers; a doubling
of the elutriation tanks; mixing of the contents of the sludge digestion tanks; and a series
of improvements to the sludge-gas piping.
The treatment plant has the capacity to handle peak flows up to 300 mgd and in the
present setup, flows in excess of 300 mgd would bypass the treatment plant. In reviewing
the performance of the plant (3), operating data of the past three years (fiscal 1966-1968)
were utilized. It was observed that the plant was providing 70 to 80 percent reduction
of BOD and suspended solids, with removals lowest for the year 1968. For fiscal year
1969, the average flow to the treatment plant was 248 mgd. Table 4 lists the 1969 average
daily pollutant loadings discharged in the effluent from the Blue Plains Plant. Also listed,
for purposes of comparison, are the maximum allowable loadings recommended by the
conferees of the May 8, 1969 conference on pollution of the Potomac River (4).
Table 4
Daily Pollution Loads from Blue Plains Plant
BOD SS TP TN
Ibs/day Ibs/day Ibs/day Ibs/day
Average 1969 Effluent 124,000 139,000 87,000 75,000
Recommended Maximum
Allowable Discharge 12,700 -- 560 6,130
The District has made concerted efforts to maintain plant efficiency under increased flow
conditions and has had to develop specialized operating procedures and techniques.
However, major modifications and additions to the plant are required to meet the
recommended waste loadings. The District has prepared a plan to complete these
improvements by the year 1974. The requirements for advanced waste treatment and the
limited land area available restrict the ultimate capacity of the plant to 309 million gallons
per day. The plant will be constructed to handle this 309 mgd capacity by 1974. However,
in addition to the 309 mgd flow, the plant will be constructed to provide complete
treatment for an incremental increase of 289 mgd for a period not to exceed 400 hours
per year. For short durations, the plant will also be able to provide partial treatment
of flows 2 to 5 times the 309 mgd flow.
19
-------
Although the sewage flow from the District is not expected to increase significantly in
the future, the sewage flow from the suburbs served by the plant is increasing at such
a rate that the 309 mgd capacity will be exceeded in 1977. Therefore, the responsible
agencies outside of the District have scheduled the construction of additional regional
plants.
20
-------
Peak Rate and Volume of Overflow
General Basis of Data
The significance of the combined sewer problem is measured by the pollutant loadings
discharged in the overflow to the District's watercourses. The selection of the capacities
of facilities required in alternative solutions to the problem must rest, in part, on parameters
such as volume of overflow and peak flow rate. Particular values of each of these parameters
vary with each storm, and it is necessary to perform an analysis of rainfall-runoff
relationships to define the frequency and range of values. This type of information would
provide a basis for evaluating alternatives.
To perform this analysis, extensive data are required concerning rainfall intensities and
frequencies, characteristics of each drainage basin and the existing sewer system, average
dry-weather flow, etc. Most of the required data have been developed in previous, unrelated
investigations. To avoid the duplication of previous efforts and to assure that the best
available data were used, the analysis was coordinated with District officials. In this analysis,
only the areas scheduled for separation after the year 1975 were considered in the
determination of the magnitude and intensity of the combined sewer overflow problem.
For the purpose of this investigation, an attempt had been made to use the existing design
storm basis for the Washington, D.C. sewer system to evaluate the peak flow rates and
volume of runoff. This design storm, one which occurs once every fifteen years with a
duration of one hour, provides a basis for calculating peak flow rates, but does not provide
a sound basis for calculating the volume of runoff. For example, the amount of rainfall
associated with the 15-year, one-hour storm is 2.66 inches, whereas the maximum daily
rainfall occurring once in fifteen years for a 24-hour duration is 5.5 inches. It is obvious
then that all further considerations of volume of overflow should be based on a 24-hour
duration. The peak flow rates, however, should be based on storms of shorter duration,
since the mean intensity is higher.
Rainfall-Runoff Analysis
The hydrologic analysis of the current study involved two procedures. The first procedure,
the Rational Method, is limited to estimation of peak flow rates; it cannot provide the
second necessary component of the analyses, i.e., volume of runoff. In the second
procedure, for determination of the volume of overflow, several different approaches were
examined. Some methods seemed to provide reliable results, but were too elaborate,
requiring lengthy and involved data collection and analysis, to be completed within the
time and budget limitations of this study. Some simpler approaches such as the unit
hydrography method seemed to provide unreliable results. After a thorough examination,
an approach was selected similar to the approach used in another combined sewer study
(5) in Washington, D.C. This approach incorporates two methods. In the first method,
the volume of surface runoff of rainfall into combined sewers was calculated by subtracting
the losses due to infiltration into pervious soils and retention in surface depressions from
the total amount of precipitation. An explanation of the methodology used for the
calculations of runoff is given in detail in Appendix C as "The Hyetograph
Method Volume of Overflow". In general, the methodology includes:
1. Construction of hyetographs for storms of various frequencies (Figure 5).
21
-------
ro
NJ
10.0
8.0
8.0
oc
<= 7.0
UJ
"" 6.0
«/>
UJ
«->
= 5.0
»-
£ 4.0
UJ
>—
I 3.0
i.o
0.0
FIGURE 5
HYETOGRAPH FOR VARIOUS RAINFALL FREQUENCIES
0 10 20 30 40 50 60 70 80 90 100 110 120
TIME FROM BEGINNING OF SIGNIFICANT RAINFALL, MINUTES
130 140 150
160
-------
2. Determination of the relative percentages of pervious and impervious areas of
each sewer district.
3. Determination of the relationship of infiltration capacity to time (Figure 6).
4. Construction of "accumulated mass rainfall curves" and "actual accumulated
mass infiltration curves" and subsequent determination of rainfall in excess of
infiltration (Figures 7 and 8).
5. Determination of the volume of rainfall retained in surface depressions.
6. Use of an area-depth correction factor to account for application of point rainfall
data to the drainage basin.
The calculated volume of runoff accounted for all runoff from pervious areas during the
period of a storm when rainfall was in excess of infiltration capacity, and from impervious
areas during the 24-hour period of extreme rainfall. The rainfall intensity values were
read from the updated Washington, D.C. intensity-duration-frequency curves (recorded
values 1896-1897, 1899-1950, 1951-1969).
This method provides reliable values for the volume of runoff from any sewer district;
however, this value may differ from the resulting volume of overflow. To determine the
volume of overflow, a method of hydrograph routing was employed. Appendix C likewise
presents a discussion of this method. In general, this method used the Rational Method
to predict the peak flow rate, which was plotted at the time of concentration for each
drainage basin. Based on the peak runoff and the total volume of runoff previously
calculated, a simple triangular hydrograph was assumed and plotted.
The results of the hydrograph routing were smoothed and extended to reflect the pattern
of actual hydrographs, yet the volume of runoff determined by the hyetograph method
was maintained. Following this, a cursory examination was made of each point of overflow.
Proposed interceptor capacity, dry-weather flow, the particular operation of the diversion
structure (when it closes, when it opens), etc. were accounted for to determine the peak
rate of overflow and the volume of overflow. Table 5 lists the results for four storm
frequencies.
Not all of the existing combined sewer districts are included in Table 5. Two districts,
A-9 and part of B-2, were scheduled for separation prior to 1975. It should also be noted
that overflow will not occur at 13 combined sewer districts along the lower reaches of
Rock Creek, even during storms as intense as the 25-year storm. There will be no overflow
at these districts due to the relief provided by the capture of combined sewer flow upstream
(any District-wide solution must deal with the overflow along the upstream reaches of
Rock Creek) and the large capacity of the relief interceptors in the lower reaches of Rock
Creek. While the relief interceptors have the capacity to contain the peak rates of runoff
along Rock Creek, the pumping station near the mouth of Rock Creek does not have
the capacity to force the peak rates to the Blue Plains plant. For example, the Rock
Creek interceptors have a capacity of over 1,200 mgd, while the Potomac Sewage Pumping
Station has a capacity of less than 500 mgd available to pump the Rock Creek flow
plus sanitary sewage from Maryland and Virginia and combined sewer flow from the Upper
Potomac. It is obvious that although some districts will not experience overflows within
the districts themselves, their runoffs will overflow at a subsequent point in the sewer
system.
23
-------
FIGURE 6
STANDARD INFILTRATION-CAPACITY CURVES FOR PERVIOUS SURFACE
NJ
-Pi.
INDUSTRIAL AND COMMERCIAL AREAS
O
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180
TIME, MINUTES
SOURCE: DESIGN AND CONSTRUCTION OF SANITARY AND STORM SEWERS. ASCE, MOP.NO.37, NEW YORK, I960.
-------
01
3.5
3.0
2-0
CJ
C-3
i. o
FIGURE 7
DETERMINATION OF INFILTRATION OFFSET
5-YEAR RAINFALL FREQUENCY
PERVIOUS RESIDENTIAL AREA
ACCUMULATED MASS OF RAINFALL
RAINFALL LESS INFILTRATION
ACCUMULATED MASS OF INFILTRATION
(BEFORE OFFSET)
ACCUMULATED MASS OF INFILTRATION
(OFFSET 36 MINUTES)
40 50 60 70 80 90 100 110 120
TIME FROM BEGINNING OF SIGNIFICANT RAINFALL, MINUTES
OFFSET 36 MINUTES
POINT OF TANGENCY
48.5 MINUTES
-------
FIGURE 8
DETERMINATION OF POINT OF INTERSECTION
OF INFILTRATION CAPACITY AND PRECIPITATION RATE
5-YEAR RAINFALL FREQUENCY
PERVIOUS RESIDENTIAL AREA
NJ
O
HYETOGRAPH
INFILTRATION CAPACITY
FOR
RESIDENTIAL AREA
(OFFSET 36 MINUTES)
JF LTRATION
CAPACITY -- PRECIPITATION RATE
10 20 30 40
00
50 60 70 80 90 100 110 120 130
TIME FROM BEGINNING OF SIGNIFICANT RAINFALL
40 150 160
70
-------
Table 5
Combined Sewer Overflow Characteristics
(After 1975)
NJ
Location of Overflow
District
No. District Name
A-10 37th St. - Georgetown
A-11 Georgetown
A-12 K Street - Wisconsin Ave.
B-6 M Street - 27th Street
B-7 28th Street - Wisconsin Ave.
C-25 Slash Run (Part) at 22nd and M Sts.
D-4 Piney Branch
D-5 Oak Street - Mt. Pleasant
D-6 I ngleside Terrace
D-7 Park Road
D-8 Lamont Street
D-9 Kenyon Street
D-10 Irving Street
D-11 Quarry Road
D-12 Ontario Road
E-2^ Pre-Potomac Sewage Pumping Station
Sub Total
E-3 0 Street Pumping Station
F-1* Main Pumping Station
G-2 Northeast Boundary
G-3 Barney Circle
G-4 14th Street - Penna. Ave.
G-5 12th Street - 9th Street
G-7 6th Street - 7th Street
Sub Total
Districtwide Total
Inches of Rain
Storm
2-Year,
Volume
mil. gallons
0.2
11
2
0.0
0.0
1
64
1
0.9
0.8
0.5
0.4
2
1
0.5
97
182
58
70
280
0.2
15
6
5
434
616
24-Hour
Peak Rate
mgd
25
330
81
0
0
120
1,500
45
29
24
15
13
51
14
8
450
1,500
1,600
3,600
31
420
200
190
5-Year,
Volume
mil. gallons
0.4
15
3
0.3
0.3
4
100
2
1
1
0.7
0.6
4
2
1
158
293
79
96
370
0.4
20
9
7
581
874
24-Hour
Peak Rate
mgd
32
400
100
50
30
240
2,100
65
40
34
26
19
96
34
20
620
1,800
2,000
4,400
38
550
250
230
15-Year,
Volume
mil. gallons
0.5
20
4
2
1
8
160
3
2
2
1
1
6
2
1
214
428
100
130
490
0.8
28
13
10
772
1,200
24-Hour
Peak Rate
mgd
42
480
110
90
40
380
2,700
71
47
39
29
19
120
45
25
810
2,200
2,400
5,500
54
660
320
290
25-Year,
Volume
mil. gallons
0.8
24
4
2
2
11
190
4
2
2
1
2
6
3
2
260
516
120
150
560
1
32
15
12
890
1,406
24-Hour
Peak Rate
mgd
45
500
120
110
50
440
2,900
81
50
41
32
26
126
50
27
890
2,300
2,500
6,000
62
630
340
310
3.3
4.3
5.5
6.0
^Overflow results from runoff in all sewer districts in Rock Creek Basin. Overflow occurs through overflow structures located between
Rock Creek Pumping Station and Potomac Sewage Pumping Station. Sewer District E-2 is the nearest district and the major contributor
to overflow.
^Overflow results from runoff in sewer districts along the Anacostia River. Overflow occurs at Main Pumping Station, and Sewer
District F-1 is the major contributor to overflow.
-------
In all, there would still remain many points of overflow beyond 1975. The largest volume
of overflow was observed from the Northeast Boundary Trunk Sewer, which would have
490 million gallons of overflow with a peak rate of 5,500 mgd for a storm of 15-year
frequency. For this storm the volume of overflows, in general, ranged from 0.1 to 35
million gallons for small drainage districts, up to 490 million gallons for the Northeast
Boundary area, and the peak rates ranged from 19 mgd to 5,500 mgd. As can be noticed,
the peak rates of overflow were generally high, even for the drainage districts where the
volume of overflow was small.
28
-------
Field Monitoring Program
General
A significant secondary objective of this study was to obtain data concerning actual
overflows from some of the combined sewers in the District. Data concerning actual rates
of overflow, pollutant concentrations versus time, etc. provide a basis to predict the
characteristics of combined sewer overflows at other locations in the District.
The comprehensive field survey conducted as part of the definition of the pollution problem
involved three automated monitoring systems installed in different sewer districts; two
for combined sewer overflows and one for separated storm water discharges. The monitoring
systems were operated during the period April 1, 1969 to September 23, 1969.
Selection of Sites
Three sampling sites were selected at different geographical areas in Washington, D.C.,
based on the following considerations:
1. Size of Drainage Area-The drainage area of the sewer district must be large
enough to have a good combination of different surface development but also
small enough to be monitored economically. The optimum size was determined
to be in the range of 100 to 300 acres.
2. Population Density--The population density within the monitored district should
be reasonably close to the overall population density of the entire combined
sewered area, which was estimated to be about 33-35 people per acre.
3. Integrity and Simplicity of the Sewer System-Multiple diversion or intercepting
points in the system to be monitored interfere with valid correlation between
runoff and rainfall. All the sewage flow and the surface runoff of the sewer
district must converge into one trunk line equipped with one overflow point.
4. Feasibility of Construction and Equipment Installation-The geographical
configuration at the monitoring site must be flat, a power source must be
available at or near the site, and there must be reasonable expectation of
obtaining permission to use the site.
5. Traffic and Public Impact-The monitoring site must not involve any public
annoyance or traffic congestion attributable to the installation of the monitoring
equipment. Exposure to malicious vandalism must also be avoided.
6. Parity Between the Sites for the Combined and Separated Sewer Systems-The
site for the separated sewer system and one of the sites for the combined sewer
system should be as similar as possible. Their drainage areas and population
densities should also be similar. This makes comparison of the hydrological,
geological, geographical, and meteorological data for both kinds of monitored
areas more meaningful.
7. Size of the Trunk Sewer-In view of the necessity of installing various items
of equipment in the sewer, the trunk sewer must be spacious enough to work
inside. Minimum usable size was set at three feet in diameter.
29
-------
8. Underground Installation-The existence of any sizable underground structures
or installations in the vicinity of the monitoring stations is undesirable because
of possible interference with the underground conduits required for the
monitoring operation. The information concerning different utility lines such
as gas, water, and electricity, must be reviewed and their exact locations noted.
Based on the detailed study of various pertinent information and field observations, three
sewer districts were selected as meeting these criteria:
1. Sewer District B-4
2. Sewer District G-4
3. Sewer District of Good Hope Run
Sewer Districts B-4 and G-4 (as designated by the D.C. Department of Sanitary Engineering)
are in the combined sewer system whereas the Good Hope Run Sewer District is in the
separated sewer system. Sewer District B-4 is in the Georgetown area, and G-4 and Good
Hope Run are along the Anacostia River. The basic data for the physical descriptions
of the sewer districts selected for monitoring are summarized in Table 6. Figures 9, 10,
and 11 show the major sewers in each of these districts and the locations of monitoring
equipment.
Operation of Monitoring Systems
The monitoring systems were operated by ROY F. WESTON personnel, with the assistance
of personnel of the Sewer Operation Division of the Washington, D.C. Sanitary Engineering
Department. Storms of varying intensities were monitored at each of the three sites; seven
in Sewer District B-4, seventeen in Sewer District G-4, and ten in Good Hope Run Sewer
District. Starting time, duration, total rainfall, maximum intensity, numbers of samples
collected, and monitoring sites involved for each storm are listed in Table 7. Storms were
missed at each of the monitoring sites from time to time, mostly because of storm caused
damage to equipment located in the sewers; these storms are listed, along with the problems
involved, in Table 8.
The stations were designed to operate automatically, because of the difficulties of predicting
the beginning of significant precipitation, assembling personnel on short notice, and
assembling personnel during the night hours. A schematic diagram of the monitoring
equipment is shown in Figure 12.
Selection of a satisfactory technique for flow measurement presented a problem. A weir
setup could not be used, because backwater elevations would have caused surcharging and
flooding at the anticipated high flow rates. Depth-of-flow measurements with the use of
one of the steady state empirical equations (Manning, Kutter, etc.) for calculating flow
would not be applicable, since flow conditions were not steady-state during periods of
precipitation. The approach finally selected for measuring flow rates was to use a tracer
solution and a form of the continuity equation; this procedure is described mathematically
in subsequent paragraphs of this section. Physically, it involves the release of a tracer
solution (lithium chloride in this case) of known concentration and feed rate at an upstream
manhole, sampling of combined sewer overflows at a downstream manhole, and analysis
for tracer concentration. From the upstream feed rate and concentration and from the
downstream concentrations, the flow can be calculated accurately. Previous experience
with this method on steady-state flows has indicated that accuracy within ±4 percent
30
-------
CO
Sewer
District
B-4
G-4
Table 6
Physical Descriptions of the Monitoring Sites
Location
Rose Park Playground
Rock Creek Parkway
26th - 0 St.
14th-LSt. S. E.
14th-MSt. S. E.
14th-LSt.S. E.
Good Hope Run 17th - Minn. St. S. E.
1630 16th S. E.
16th - Minn. St. S. E.
Monitoring
Operation
Lithium Chloride
Release
Sample Collection
Rainfall Measurement
Lithium Chloride
Release
Sample Collection
Rainfall Measurement
Lithium Chloride
Release
Sewer
System
Combined
Combined
Separated
Storm
Population
Drainage Density
Area 1970
acres persons/acre
105 43.6
252 52.6
265 37.6
Trunk
Sewer
Size
4' Diam.
5'-6"Di£
6'x6'
Sample Collection
Rainfall Measurement
-------
25TH ST.
FIGURE 9
COMBINED SEWER DISTRICT B 4
RAINFALL MEASUREMENT
LITHIUM CHLORIDE RELEASE
SAMPLE COLLECTION
500 250
500
SCALE IN FEET
RF
32
-------
FIGURE 10
COMBINED SEWER DISTRICT
XS
CO
CO
500 250 0 500 1000
SCALE IN FEET
RF RAINFALL MEASUREMENT
LiCI LITHIUM CHLORIDE RELEASE
SC SAMPLE COLLECTION
-------
FIGURE 11
GOOD HOPE RUN SEWER DISTRICT
O
400 200 0 400
SCALE IN FEET
RF RAINFALL MEASUREMENT
LiCI LITHIUM CHLORIDE RELEASE
SC SAMPLE COLLECTION
-------
Date
Starting
Time
Site
Table 7
Storms Monitored
Total
Rainfall
(inches)
Duration
(min.)
May 9
May 19
May 20
June 1
June 2
June 3
June 8
June 8
June 15
July 6
July 27
July 27
July 28
July 28
July 28
July 28
July 28
Aug. 1
Aug. 1
Aug. 2
Aug. 2
Aug. 2
Aug. 3
Aug. 9
Aug. 9
Aug. 9
Aug. 9
Aug. 9
Aug. 10
Aug. 19
Sept. 4
Sept. 17
Sept. 17
Sept. 20
9:25 a.m.
1:42 a.m.
11:42 p.m.
7:25 p.m.
7:45 p.m.
12:25 a.m.
5:44 p.m.
5:50 p.m.
2:00 p.m.
7:40 p.m.
11:35 p.m.
11:35 p.m.
2:30 a.m.
11:30 a.m.
1:20 p.m.
1:28 p.m.
5:00 p.m.
7:30 p.m.
8:45 p.m.
8:05 p.m.
8:10 p.m.
8:17 p.m.
10:30 p.m.
9:20 p.m.
9:20 p.m.
11:00 p.m.
11:20 p.m.
11:22 p.m.
12: 25 a.m.
6:40 p.m.
3:45 p.m.
8:15 p.m.
8:20 p.m.
3:00 p.m.
G-4+
G-4
G-4
B-4°
B-4
B-4
G.H.R.*
G-4
G-4
G-4
G-4
B-4
G-4
G-4
G.H.R.
G-4
G.H.R.
G-4
G.H.R.
G-4
B-4
G.H.R.
B-4
G-4
G.H.R.
G.H.R.
G-4
B-4
G.H.R.
G-4
G-4
G-4
G.H.R.
G.H.R.
0.8
0.4
0.6
1.4
0.9
0.95
0.7
0.7
0.7
0.4
2.1
1.3
0.6
0.6
1.6
1.3
0.2
0.6
0.5
2.8
3.9
2.9
0.4
1.1
1.5
0.4
1.6
1.6
0.65
1.35
3.4
0.7
0.6
0.1
20
3
7
22
20
15
6
13
40
20
39
20
22
15
40
35
30
210
135
65
50
73
30
17
25
30
10
15
20
13
75
5
100
90
Maximum
Intensity
(in./hr.)
2.4
3.0
6.0
7.2
6.0
6.0
6.0
2.8
6.0
0.8
8.0
7.0
4.6
4.0
6.0
5.6
0.4
1.0
1.6
6.0
7.5
5.6
4.0
8.0
7.2
2.0
7.2
8.4
0.6
• 8.0
6.0
4.2
0.6
0.1
Number of
Samples
Collected
1
4
3
6
3
3
1
3
4
2
12
2
6
6
22
5
8
1
1
8
6
12
3
5
7
5
2
3
2
4
1
1
4
1
G-4+ Combined Sewage in Sewer District G-4
B-4° Combined Sewage in Sewer District B-4
G.H.R.* Storm Runoff in Sewer District Good Hope Run
35
-------
Date
June 1
June 2
June 1
June 1
June 2
June 15
June 15
June 18
July 20
July 22
Aug. 19
Sept. 4
Sept. 4
Sept. 8
Sept. 8
Starting
Time
10:40 p.m.
12:40 a.m.
10:18 p.m.
7:20 p.m.
7:45 p.m.
1:53 p.m.
12:45 p.m.
9:30 p.m.
5:38 p.m.
7:00 p.m.
6:50 p.m.
3:52 p.m.
3:45 p.m.
1:10p.m.
12:55 a.m.
Site
G-42
G-4
G.H.R.3
B-44
B-4
G.H.R.
B-4
All Sites
All Sites
All Sites
G.H.R.
G.H.R.
G-4
G-4
B-4
Table 8
Major Storms Missed^
Loss Involved
LiCI and Sample
LiCI and Sample
Sample
Instruments Flooded
Instruments Flooded
Sample
Sample and LiCI
Sample and LiCI
Sample and LiCI
Sample and LiCI
Sample and LiCI
Sample and LiCI
Sample and LiCI
Submersible Pump
None
Sample and LiCI
Problems
Submersible Pump Clogged
Submersible Pump Clogged
Bubbler Line Broken
Poor Drainage at Site
Poor Drainage at Site
Bubbler Line Broken
Electrical System Flooded
In Process of Repairing
Equipment in Sewers
Submersible Pump Clogged
In Process of Repairing
Equipment
Sampling Head not in Barrel
Sampler Switch System
Broken Down
Submersible pump Washed Away
Submersible Pump and Cage
Being Repaired
Submersible Pump Clogged
11ntensity-duration information for many of the missed storms can be found in Tables E-1, E-2 and E-3 of
Appendix E.
^Combined Sewer District G-4.
^Separated Sewer District - Good Hope Run
4Combined Sewer District B-4.
36
-------
FIGURE 12
MONITORING EQUIPMENT
Lid RELEASE STATION
ELECTRIC CLOCK
LiCI METERING PUUP-
L
/DIFFERENTIAL
PRESSURE
REGULATOR
, JESSURE
REGULATOR
ELECTRIC POHER
SUPPLY AND
T8IGGERINC
SrSTEN
-KERCURt PRESSURE SWITCH
-PURGEMETER
PROCESS PRESSURE TO
CURRENT TRANSMITTER
ELECTRONIC
STRIP-CHART
RECORDER
3-COH1UCTOR
JACKETED CABLE
SAMPLE COLLECTION STATION
LiCI RELEASE
LIKE
ELECTRIC CONDUIT
UPSTREAM
•SEKEfl MANHOLE
,RAMSET STUDS
= ELECTRIC tONDUIT
— -— , TSIB'IS, |
i*
1>
n DOINSTDEAM
-SEIER MANHOLE
LEGEND
'/ HASTEHATE8 OR LiCI RELEISE L
' AIR LINE
, ELECTRICIir LINE
fl
BUBBLER LINE
RI»L PLUGS
12
-------
can be obtained. It should be pointed out that the method has never been verified for
runoff flow measurements; however, because the tracer dilution method is based on
continuity of mass rather than energy, unsteady-state conditions should not introduce
errors.
Flow rate estimates based on depth-of-flow measurements and the Manning formula were
compared to the results of the tracer method. The depth-of-flow estimates showed only
a general correlation, having a significant spread, in comparison with the tracer results.
Three possible sources of error explain these differences:
1. Incomplete mixing of tracer solution
2. Inaccurate measurement of depth-of-flow
3. Assumption of steady-state conditions in using Manning's formula.
Sufficient steps were taken to assure practically complete mixing of the tracer solution,
and therefore, the other two sources of error explain the differences in flow estimates
and the relative inaccuracy of the depth-of-flow procedure.
During dry-weather periods, the operating effort was minimal, because the stations operated
only during periods of precipitation. Thus, dry-weather periods provided adequate time
for maintenance work, routine operating checks, and preparation of equipment for
forecasted storms. The triggering depth at each site was set according to the minimum
depth requirement for operation of the submersible pump, the maximum dry-weather flow
depth, and the expected magnitude of the impending storm; the metering rate for lithium
chloride was also pre-set on the basis of expected storm magnitude.
For the short intense storms, the concentration of each waste constituent was observed
to increase with the discharge rate in the sewer; the peak concentrations of many waste
constituents were concurrent with the peak flow. The concentrations were significantly
high and remained so throughout most of the monitoring period.
The time variation of the quality and quantity of wastewater generated by a long-duration,
high-intensity storm shows that the flow rate of the wastewater generated by this long,
intense storm is much higher than that of short storms. However, the concentrations of
various constituents were observed to be lower, which is an anticipated result of high
dilution by storm water.
Chemical oxygen demand and suspended solids concentrations of wastewater from a long,
intense storm are reduced to approximately one-third of the comparable values for a short,
intense storm. The biochemical oxygen demand for the long, intense storm wastewater
were about one-seventh of that in short-storm wastewater. This also implies that a higher
fraction of surface material was eroded by the long, intense storm. However, the ratio
of COD to BODs was essentially the same as that for short-storm wastewater.
The characteristics and quantities of wastewater generated by four consecutive storms
during July 27-28, demonstrated the initial flushing effect in all cases, even when the
storms were only a few hours apart. However, the average concentrations of specific
contaminants were observed to follow a decreasing trend with consecutive storms. For
example, the average concentration of chemical oxygen demand (COD) decreased from
307 mg/L for the first storm to 154 mg/L for the fourth storm. This decrease may be
attributed to the reduction of waste material accumulation on the surface and in the
38
-------
collection system. The ratio of COD to 6005 was essentially the same as the comparable
ratios for short storms and for long-duration, high-intensity storms. The characteristics
of long-duration, low-intensity storm wastewater probably will be similar to that observed
for consecutive storms.
The average organic and nutrient concentrations in separated storm water discharges were
observed to be approximately one-third of those in combined sewers, but the solids
concentrations (especially non-volatile solids and settleable solids) were much higher than
those in combined sewer overflow. However, this phenomenon may be attributed to the
differences in surface development in the monitored areas. Silt was found to be the
dominant factor in separated storm water discharges.
In general, the time variation in characteristics of wastewater from separated storm sewers
is very similar to that from combined sewers. One significant difference observed for
separated storm sewer discharge is the broader range of COD to BOD$ ratio (1.8 to 32).
The bacteriological examinations were made on selected individual samples of combined
sewer flow and on storm-composite samples. The range of variation and the mean values
of different bacteria species are summarized in Table 9. The bacteriological counts varied
with the flow rate of combined sewer overflow during each storm and the initial flushing
had a significant effect on various bacteria counts. The variation of bacteriological data
for the separated storm sewer discharge was similar to the variation in the combined sewer
discharge. Bacteria counts were high at the beginning of the storm and then decreased
as the storm progressed. Also, bacteria were present in storm runoff in fairly significant
amounts although less than in combined sewage, as cited in Table 9.
To assess the relative contribution to pollution throughout the duration of a storm, the
total waste materials discharged were expressed in units of pounds per unit time. As
expected, these waste loading expressions indicate that most waste materials were carried
by the initial flushing and scouring of the sewer. Waste loadings carried by the secondary
flushing in a prolonged storm were limited.
The rain gauges were operated by a local Washington, D.C. subcontractor. The waste-level
recorder charts required changing only once every 30 days, but the readings were observed
and the equipment was checked on a routine weekly basis. Equipment installed in the
sewer was inspected at least once a week.
Detailed discussions of monitoring practices and techniques are presented in Appendix D.
Monitored Wastewater Flows and Characteristics
As shown in Table 7, storms of varying intensities and durations were monitored during
the six-month program. Wastewater flows and characteristics under storm conditions were
studied both in the combined sewer districts and in the separated storm sewer district.
The effects of short-duration storms, long-duration, high-intensity storms, and consecutive
storms were observed and evaluated. The range and mean value of each of the waste
constituents in storm wastewater (for both combined sewer and separated storm sewer
flow) are presented in Table 9.
Appendix E presents a detailed comparison of the characteristics of the combined sewer
and separated storm sewer flows generated by storms representative of the wide range
39
-------
Table 9
Comparison of Characteristics of Combined Sewer Overflow
and Separated Storm Water Discharge
Waste Constituents ^
Chemical Oxygen Demand
Biochemical Oxygen Demand
Total Solids
Total Volatile Solids
Suspended Solids
Volatile Suspended Solids
Settleable Solids
Total Phosphate
Total Nitrogen
Orthophosphate
Ammonia Nitrogen
Total Coliform 6
Fecal Coliform^
Fecal Streptococcus^
o
Combined Sewer Overflow
Range
80 - 1 ,760
10- 470
120-2,900
40- 1,500
35 - 2,000
10- 1,280
0 - 1 ,308
0.8 - 9.4
1.0- 16.5
0.1 -5.0
0-4.7
5.6-6.7
Mean
382
71
883
344
622
245
229
3.0
3.5
2.0
1.5
6.3
420,000 - 5,800,000 2,800,000
240,000 - 5,040,000 2,400,000
1,000- 49,000 17,200
Separated Storm Water Discharge"
Range Mean
29- 1,514
3- 904
338- 14,600
12- 1,004
130- 11,280
0- 880
0- 7,640
0.2 - 4.5
0.5-6.5
Not Sampled
Not Sampled
7.2-6.0
335
194
2,166
302
1,697
145
687
1.3
2.1
Not Sampled
Not Sampled
6.5
120,000 - 3,200,000 600,000
40,000 - 1,300,000 310,000
3,000- 60,000 21,000
1mg/L unless otherwise noted.
on analysis of 94 samples
%ased on analysis of 64 samples
^Excluding sample from June 8 storm, which had a BOD concentration of 600 mg/L.
5pH units.
6Countsper 100ml.
-------
of durations and intensities. Appendix E includes figures that relate measured
concentrations and loadings of various contaminants with time, cumulative rainfall, and
sewer flow rate. Although Appendix E presents a detailed and complete discussion of
the results of the monitoring program, it is convenient to highlight the more significant
results.
41
-------
Impact of Storm Water Discharges
Annual Discharges
Results of the 1956-1958 studies of the Washington Sewer System indicated that an average
of 3.31 million gallons per day (mgd) of sanitary sewage had overflowed from the
combined-system areas during rain storms and 0.34 mgd during dry weather. In terms
of percentages, about 3 percent of the annual sanitary sewage produced was discharged
directly to the streams during rain storms and about 0.31 percent during dry weather.
Estimates were also made for the future, when the program of improvements (Project C
as previously discussed) would be completed. At that time, the overflow from rain storms
would amount to an average of 0.38 mgd, equivalent to 0.42 percent of the sewage flow
from the areas which would still have combined sewers.
The approach to assessment of the effects of overflows used in the present study is slightly
different. The effect on receiving streams was determined in terms of pollution load
discharged into the streams. The parameters used were BOD, suspended solids, total
phosphates, and total nitrogen. The amounts of these elements discharged into the streams
from storm water and combined sewer overflow were calculated separately.
From a probability plot of the average rainfall for the period 1900-1968 based on Weather
Bureau records, the annual rainfall with a 50 percent probability of occurrence was derived.
This value, 40.2 inches, was used to calculate the volume of total storm water reaching
the streams from the entire area of the District. After obtaining the median annual rainfall,
certain basic assumptions were made to determine the quantity of storm water and sanitary
sewage that would overflow into the streams. The parameters used in the calculations
were obtained from prior reports, from literature review, from current study data, or from
conservative estimates based on general technical knowledge.
The first assumption was that 90 percent of the runoff will reach the streams. Other
assumptions, based on the field monitoring program were made for concentrations of
pollution parameters. The following are the values of these parameters used for runoff
in this assessment of pollution effects, as taken from Table 9.
BOD 19mg/L
Suspended Solids 622 mg/L
Total PO4 1.3 mg/L
Total Nitrogen 2.1 mg/L
The suspended solids concentration of 1,697 mg/L suggested for storm water runoff by
Table 9 was not used. In comparison with the suspended solids concentrations measured
in the two combined sewer districts and with the concentrations reported in other studies,
the value of 1,697 mg/L seems extremely high and unreasonable. A value of 622 mg/L,
an average measured for the two combined sewer districts, should be more valid.
The overall runoff coefficient for the entire District had to be determined in order to
estimate the total volume of storm water runoff reaching the streams. The overall runoff
coefficient value was derived by calculating a weighted average of the runoff coefficients
obtained for individual drainage districts. The annual volume of storm water runoff was
then calculated by multiplying 0.9 (only 90 percent was assumed to reach the streams)
by the overall runoff coefficient, the total area in acres of the entire district, and the
42
-------
median annual rainfall. The average values of pounds of BOD, Suspended Solids, Total
Phosphate, and Total Nitrogen were then calculated on the basis of this annual volume
and the listed concentrations of pollution parameters.
Pollution load data for sanitary sewage discharged through the overflow structures were
calculated separately. One of the assumptions made was that approximately 95 percent
of the sanitary sewage in the combined sewers would overflow into the streams during
periods of overflow. The data for average flow per capita and for sewage characteristics
were obtained from the February 1969 (3) report on the District of Columbia's sewage
treatment plant. Duration of overflow for an average year for various interceptor sewers
was derived from the Board of Engineers' 1957 report (1).
The volume of sanitary sewage discharged directly into the stream through the overflow
structures was then calculated to be an average of 545 million gallons in 1970 and 346
million gallons in 1975 (when less overflow would result because of improvements in the
sewer system). In veiw of the delays encountered in the sewer improvement program,
it appears to be more realistic to use the 1970 estimate to assess the pollution load to
the river. In comparison with the value obtained in the prior studies, the current values
are approximately 1.0 percent of the total annual sewage flow to the treatment plant
or about 3.4 percent of the total annual sewage flow from the combined-system areas.
The average annual quantities of BOD and other pollution elements discharged to the
streams through dry-weather overflow were calculated from the volume of overflow and
concentrations of these elements present in sewage. Upon obtaining the data on pollution
loads from storm water and sanitary sewage overflows, the total quantities of BOD, SS,
P04 and Total Nitrogen discharged into the streams annually were calculated. The values
obtained initially were for the 50 percent probability of occurrence. However, once these
values were obtained, similar data for other probabilities were calculated by applying
appropriate factors from the probability plot of average annual rainfall. Probability plots
for BOD, SS, Total P04 and Total Nitrogen discharged annually to the streams in million
pounds were plotted. These plots are shown on Figure 13, and the data for 50 percent
probability are summarized in the first part of Table 10 (Annual Loads).
Individual Storm Loadings
A definition of the combined sewer problem based solely on annual pollution loadings
is incomplete. Overflows from combined sewers do not occur as continuous steady
discharges, but rather as slug loadings. This characteristc, in combination with the long,
effective residence times of estuarine waters, explains the particularly serious impact
combined sewer overflows have on water quality.
The sources of pollutants in combined sewer overflow are the storm runoff from the
urban area, the sanitary sewage mixed with the runoff, and the initial flushing action
in the sewers. The average concentrations of pollutants in storm runoff from an urban
area have been established in this study. It is reasonable to assume that pollutant
concentrations of sanitary sewage included in combined sewer overflows are typical of
normal domestic sewage. An independent study (6) of the flow in combined sewers B-4
and G-4 during storm conditions estimated that flushing action results in a total added
BOD loading of 3 pounds per acre of drainage area. An extension of this calculation
to phosphate and nitrogen pollution indicates that flushing action results in total
phosphorus and total nitrogen loadings of 0.2 Ibs. and 0.3 Ibs. per acre, respectively.
43
-------
FIGURE 13
ANNUAL POLLUTION LOAD DISCHARGED
TO STREAMS FROM STORM WATER
12.0
1.0
10 0
I
9.0
6.0
7.0
6.0
5.0
4.0
250
225
200
175
150
125
100
75
50
25
2.0
1.0 L
0.2
O.I
59.2 53.0
ANNUAL RAINFALL , INCHES
46.4 40.2
35.0 30.7
0.2
12 5 I" 20 30 40 50 60 70 BO 90 95 98 99
PROBABILITY OF OCCURRENCE
44
FIG.
13
-------
Table 10
Expected Ppllution Loads
from Combined and Separated Storm Sewers
in District of Columbia
Based on Annual Rainfall of 40.2 Inches
Overflows from
Combined Sewer Areas1
Storm Water
Sanitary Sewage
Sub-Total
Direct Discharge from
Separated Storrn Sewers
TOTAL
Discharge from D.C. Combined Sewer District
for 2-Year, 24-Hour Storm
Discharge from D.C. Combined Sewer Districts
for 5-Year, 24-Hour Storm
Discharge from D.C. Combined Sewer Districts
for 15-Year, 24-Hour Storm
Discharge from D.C. Combined Sewer Districts
for 25-Year, 24-Hour Storm
Districts4
Recommended Maximum Allowable
Volume
mil.gal./yr.
1 1 ,000
1,000
12,000
16,000
28,000
BOD
mjl.lb./yr.
1.8
1.4
3.2
2.5
5.7
SS
mil.lb./yr.
57
2
59
83
142
24-Hour Pollution
TP2
mil,lb./yr.
0.1
0.4
0.5
0.2
0.7
Loads from
Various Return Frequency Storms
Volume BOD TP2
ter District
»er Districts
ler Districts
/er Districts
bined Sewer
e
Washington Area
million gallons
616
874
1,190
1,396
33
Ib.
160,000
200,000
250,000
280,000
8,000
16,500
T67~
12,000
13,000
14,000
15,000
1,400
740
Tl\|3
mil.lb./yr.
0.2
0.3
0.5
0.2
0.7
Tl\|3
Ib,
19,000
24,000
29,000
33,000
1,400
8,000
1Combined Sewer Overflows occur approximately 50-60 times per year.
^Total Phosphorus.
^Total Nitrogen,
4Determined by dividing annual loads by 365 days/year.
45
-------
The previous analysis of rainfall-runoff relationships to determine volume of overflow
provides a sufficient basis to determine pollutant loadings. The total pollution load
overflowing from the combined sewer district during a storm is estimated by adding the
following values:
1. Volume of storm water runoff times the average pollutant concentrations
previously established.
2. Acreage of the drainage district times the flushing unit loadings.
3. Typical sanitary sewage pollutant concentrations times the average dry-weather
flow generated in the district.
The 24-hour pollution loads calculated for the combined sewer system for storms of various
return frequencies are presented in the second half of Table 10. Also included for purposes
of comparison are the maximum allowable contaminant loadings if water quality objectives
in the Potomac are to be met. Examination of these figures shows that the BOD and
phosphorus discharges, even from a 2-year, 24-hour storm are from ten-fold to twenty-fold
greater than the recommended loadings in the entire metropolitan Washington area. In
fact, even the average daily loadings (determined by dividing annual loadings by 365
days/year) in combined sewer overflows account for a good portion of or exceed the
recommended loadings. It is pointed out that there is no uniform daily discharge of
combined sewer overflow; the annual loadings are concentrated unevenly in the fifty to
sixty overflows that occur each year. It is this characteristic combined with the long
residence times of estuarine waters that explains the particularly serious impact of combined
sewer overflows on the water quality of the Potomac.
46
-------
SECTION V
INVESTIGATION OF POTENTIAL ABATEMENT MEASURES
Review of Approaches Tried at Other Cities
Many municipalities and some private companies have been and are continuing to investigate
methods for dealing with the problem of combined sewers. The methods being considered,
separately or in combination, can be classified under four headings: 1) sewer separation;
2) off-system storage; 3) treatment processes; and 4) miscellaneous.
Sewer Separation
Complete separation of sanitary and storm sewers is an enormous project. It requires the
following:
1. Separate storm and sanitary sewers
2. Separate roof drains and downspouts
3. Separate yard and areaway drains
4. Separate air conditioning and cooling system drains
5. Separate foundation drains
6. Separate catch basin inlets
The costs of complete separation of any extensive existing combined sewer system are
prohibitively high. In a recent study (7), the American Public Works Association estimated
that it would cost 30 billion dollars to separate all the combined public sewers in the
country, and another 18 billion to make the related plumbing changes on private properties.
Costs for partial separation, involving separation of only the most troublesome sewers
and the related storm water drains would obviously be less, but still would be substantial.
According to District of Columbia Department of Sanitary Engineering figures, the typical
cost encountered within the District of Columbia in 1957 was $18,000 per acre.
There are other complications. Complete separation projects would take many years to
complete, even if concentrated construction effort were applied. Communities and
businesses would suffer inconveniences and losses during construction, when streets would
be closed to traffic. Extensive policing would be required to ensure the separation of
all storm water connections (e.g. roofs, yards, etc.). Furthermore, recent studies (8, 9,
10) reveal that storm water runoff itself is polluting the streams, and, in many situations,
should receive treatment. This has also been confirmed in this study.
Even with these complications, sewer separation is by far the most commonly used remedial
method in the 900 communities surveyed by the APWA; however, separation has generally
been confined to portions of sewer systems and has not very often been applied to an
entire system. Only rarely have all roof drains, air-conditioning drains, etc., been separated.
The method of separation which appears to be the most practical is to construct a sanitary
sewer within the larger existing combined sewer. This was accomplished in Ottawa, Canada
using 15-inch cast iron pipe at a cost of $20-25 per foot in place (11). Minneapolis installed
a flattened 42-inch corrugated steel sanitary sewer along the invert of a 102-inch tunnel,
at a low but undisclosed cost (12). A different approach is used in the ASCE Combined
Sewer Separation Project, which involves pumping comminuted sewage from individual
buildings through pressure tubing to pressure conduit installed within existing combined
sewers (13).
47
-------
Off-System Storage
The use of storage (surface or sub-surface) in connection with the problem of combined
sewers is being applied in two general ways:
1. Temporary storage, with return of the retained flow to the sewer system for
conventional treatment when the storm subsides; and
2. Temporary storage, followed by discharge of the retained flow directly to the
watercourses.
The latter approach may be acceptable in certain situations, because some degree of solids
removal will occur while the wastewater is stored; however, application is limited to
situations where water quality requirements are not severe and/or when other pollutant
concentrations (dissolved materials) are low.
Chicago Underflow and Deep Tunnel Plans (14) (15)
Several cities have or plan to have facilities for temporarily storing overflow for later
release to a conventional waste treatment plant. The Chicago Underflow and Deep Tunnel
Projects are the biggest and best known of these plans. For years, Chicago has experienced
basement and underpass flooding, as well as storm and beach pollution resulting from
combined sewer overflows. The Underflow-Storage Plan and the Deep Tunnel Plan represent
viable alternatives for solving the problems of flooding and water pollution by handling
the runoff from a 100-year storm.
"The Underflow-Storage Plan proposes the construction of a pattern of large tunnels in
the dense Niagaran limestone rock formation, 200 to 300 feet below the surface waterway
system. These tunnels would be sized to provide a linear distribution of storage volume
and conveyance capacity in a pattern which would intercept all of the approximately
400 outfalls of the existing combined sewers. The tunnels would be sloped down to low
points and pumping facilities opposite the existing sewage treatment plants. Overflow from
the combined sewers, during storm periods, would drop through shafts to the large storage
tunnels. In the post storm period, the tunnels would be dewatered by pumping directly
to the existing treatment works."
"The Underflow-Storage Plan takes advantage of the lower water level to be established
in the Illinois Waterway at Lockport, Illinois, for improvement of navigation and flood
control of the waterway system. The new water level, 70 feet or more below the level
of Lake Michigan, will allow the construction of tunnels with large underflow conveyance
capacity to Lockport and provide flood protection for the largest storm of record."
"Storage of 18,000 acre-feet or 1.12 inches of runoff in the tunnel system will provide
98.5 percent reduction of pollutants entering the waterway from combined sewer spillages."
(This represents sufficient capacity to handle the 100-year frequency storm.)
"During the study of the Underflow-Storage Plan, it was decided to modify a large relief
sewer proposed by the City of Chicago, as an Underflow Sewer similar to the Metropolitan
area-wide plan but on a much smaller scale. The Underflow Sewer would be constructed
in solid rock, 250 feet below the ground surface. This sewer is now under construction
with a portion being funded by a demonstration grant from EPA-WQO. Two additional
Underflow Sewers are also under construction by the Metropolitan Sanitary District at
48
-------
widely separated locations in the Chicago area in the same dolomitic limestone rock
formation. Each of the three Underflow Sewers are being mined by a machine of different
manufacture. The construction of these Underflow Sewers has confirmed the structural
integrity and the dense impermeability of this underlying rock blanket throughout the
entire Chicago area."
"Prices per cubic yard of rock excavation vary from $60.00 for a 10-foot diameter single
tunnel with two headings to $5.65 per cubic yard for cavern (room and pillar) excavation
with multiple headings. The principal governing factor appears to be the size or face area
of the headings. For the combined Underflow-Storage tunnels, with 26-foot wide by 50-foot
high tunnel faces, the estimates are $8.81 per cubic yard for single tunnels and $8.03
per cubic yard for twin tunnels."
"The Deep Tunnel Plan is a multi-purpose plan, including hydroelectric power development,
with a "pumped-storage" scheme, now widely used throughout the world as adjuncts to
hydro-power developments on surface streams or to thermal power plants. In the Deep
Tunnel Plan, storage for hydro-power would be provided in rock caverns, 600 feet or
more below the surface and in surface reservoirs above ground in the vicinity of the
underground caverns. Reversible pump-generator units would be used intermittently to
move water upward and to develop power during downflow. Power would be generated
and sold daily during the hours of peak demand for electricity. Power would be purchased
for pumping daily, during the periods of low demand for other uses in the Metropolitan
area. Based on an estimated net revenue, in excess of cost of operation, revenue bonds
would be sold by the Metropolitan Sanitary District to provide capital for a portion of
the multi-purpose project."
"The underground caverns and the surface reservoir would be over-sized beyond the needs
for power development to provide for entrapment and storage of excess spillage from
the combined sewer outlets. Primary sedimentation would be provided underground at
the entrance to the caverns, and the sediment pumped to the existing treatment works.
Controlled outflow from the surface storage would also be directed to the existing major
treatment works."
"The total volumes of the proposed multi-purpose storage is 35,000 acre-feet below ground
and 45,000 acre-feet above ground, or a total in the system of 80,000 acre-feet, of which
20,000 acre-feet was considered to be normally needed for power development, leaving
60,000 acre-feet normally available for pollution and flood control."
"The tunnel system to deliver the combined sewer spillage to the storage and power
development site or sites would be generally of the same pattern as for the Underflow
Plant."
The underflow tunnels under construction are compatible with future extensions along
the conceptual lines of the Deep Tunnel Plan.
Other Storage Applications
Most of the applications of storage have been for treatment followed by release to receiving
waters rather than by return to the sewer systems. The treatment may be removal of
solids, or possibly some degree of stabilization, with or without disinfection by chlorination.
49
-------
This type of storage was originally used at points along a sewer system where overflows
would normally enter streams that have relatively low flows. Columbus, Ohio has been
employing uncovered standby surface storm water tanks for 35 years. The tanks hold
only part of the combined overflow, with the excess released to nearby streams (16).
Because no means for frequent removal of settled solids was originally provided, odor
problems were experienced when anaerobic decomposition occurred. Also, periodic cleaning
of the settled sludge resulted in a serious overload on the activated sludge treatment plant,
unless the cleaning operation could be stretched out over a week or so. However, facilities
for regular removal of solids have been installed, and the odor and treatment plant overload
problems have been minimized.
Halifax, Nova Scotia began the construction of two 1,000,000-gallon surface storage tanks
in 1965 at a cost of $400,000 each (17). In contrast, complete separation was estimated
at 4.3 million dollars, and partial separation (which would not prevent overflows in large
storms) was estimated to cost 1 million dollars. These storage tanks, designed to provide
15 minutes retention at peak flow, will reduce the average frequency of overflow during
the 4-month swimming season from fifteen overflows to two, and the volume of total
overflow will be reduced by about 85 percent. Chlorination facilities will provide a dosage
of 30 ppm for flows up to 40 cfs and a constant chlorine flow for the more diluted,
higher flows. The city is now considering partial separation as a supplement to surface
storage. This could be accomplished for one million dollars, and the retention tanks would
then have capacity for the 5-year storm.
Other applications of storage for treatment are Grosse Point Woods, Michigan (18) (7.5
minutes detention of 1-year storm in surface tank, sludge removed by flushing with water);
and Johnson County, Kansas (7) (30 minutes detention at maximum design rate). Planned
applications are Milwaukee (19, 20) (underground, concrete tank with 3,900,000-gallon
capacity providing 15 minutes of detention, screening of influent, and chlorination of
effluent before discharge to river; Boston (19) (10-minute minimum
sedimentation-chlorination, with chlorinated effluent discharged to river); and New York
City (19) (system of basins designed to contain 25 of 40 summer storms, with overflow
chlorinated to protect beaches). The reports of these planned applications did not identify
the design storm.
For many years, British cities have been employing surface storm water tanks to store
and treat combined sewer overflows. Their general practice is as follows (21), using British
terminology:
1. Flows up to three times the normal dry-weather flow are treated at the municipal
treatment plant.
2. Flows between three and six times the normal dry-weather flow are diverted
to storm water tanks for sedimentation prior to discharge to the receiving stream.
Any flow in excess of six times the normal dry-weather flow is discharged to
the stream without any treatment.
3. Influent to the storm water tanks is screened (6-inch and 1-inch openings).
Effluent is not chlorinated.
4. Minimum number of storm water tanks is two, and capacity must be sufficient
to provide at least a two-hour detention period at the maximum flow (not
disclosed) or must equal one half of a day's average dry-weather flow.
50
-------
5. Settled solids and remaining wastewater are returned to the sewer systems for
treatment when the storm subsides. Sludge can be removed by mechanical
equipment or by flushing with municipal water.
Current studies supported by EPA/WQO research grants are investigating the use of
polyelectrolytes and tube-type clarifiers to improve sedimentation in storage tanks (19).
A study in Cleveland involves the feasibility of using storage basins for stabilization of
combined sewer overflows (19); secondary effluent and flow from polluted streams would
also be diverted to these basins. Little information is available now but there is enough
to indicate that a higher degree of pollution abatement can be obtained with
stabilization-retention basins than with sewer separation, and at approximately one-third
of the cost of separation (22). The amount of land required, however, prohibits this
application in many cities.
Effectiveness of Storage in Abating Pollution
Studies (10) at the EPA/WQO Robert A. Taft Sanitary Engineering Center (Cincinnati
Water Research Laboratory) on sedimentation and concurrent sedimentation-disinfection
of urban storm water runoff from a separately sewered residential and light-commercial
area indicate that removal of 55 percent of suspended solids (normally expected of sanitary
sewage in primary treatment) is not obtained within 20 minutes, but takes one hour of
settling. The variations in removal appeared to be independent of the seasonal occurrence
of the storms.
Studies in England and Canada (21, 23, 24) indicate similar results with combined sewer
overflow. One consulting firm in Toronto, Canada maintains that a one-hour detention
of the one-year storm overflow can effect a BOD removal of 30 percent and a suspended
solids removal of 60 percent (23). For storms of less intensity, BOD reduction could
be as high as 45 percent, and suspended solids removal could be 65 percent. None of
these studies used polymeric flocculants or any other chemical additives.
Eliassen (25) investigated the aftergrowth of coliforms resulting from average discharge
of combined sewer overflows into a brackish water tidal basin. He reported that,
"Chlorination of the overflows to the 15-minute chlorine demand...will limit the peak
aftergrowth to approximately 500,000 per 100 ml in 40 hours...Chlorination to the chlorine
demand will result in average MPN aftergrowth values in the basin of from 10 percent
to 30 percent of those which would develop if unchlorinated overflow were discharged
to the river in the normal ranges of summer dilutions."
All studies stressed that their results, in general, cannot be applied to other sewerage
systems. The composition of combined sewer overflow is extremely variable, varying from
drainage area to drainage area as well as with intensity, duration, and frequency of storms
and probably with the time of day. No data could be obtained which would correlate
variations in composition with intensity and duration of rainfall. Suspended solids and
BOD removal by sedimentation depend on concentration as well as on the period of
detention. This further complicates establishing a pattern for Chlorination efficiency,
because it depends in part on the nature and concentration of the suspended solids.
Combined Sewer Overflow Treatment Processes
There are approximately 15 current EPA/WQO-funded R and D projects for investigation
and/or demonstration of varied, non-conventional processes for treating combined sewer
51
-------
overflows; however, these projects have not been completed, and the available information
is not conclusive (19). Very few of these treatment methods involve biological processes;
most are concerned with screening or filtration. Some of the more interesting projects
are listed in Table 11.
Microstraining, a form of simple filtration using specially-woven wire fabric mounted on
the periphery of a revolving drum, is one of the more promising approaches. In this process,
wastewater passes through a partially submerged drum from inside to outside. Downstream
water is used to flush any solids that build up on the inside.
Preliminary results of microstraining at a demonstration site in Philadelphia indicated that
a 30 to 60 percent removal of both BOD and suspended solids could be obtained with
a 35-micron fabric; filtering rates were approximately 10gpm per square foot (submerged).
The contractor feels that better results are obtainable, because the overflows tested were
pumped to the Microstrainer, and pumping should be avoided since it tends to break
up any fragile solids, thereby reducing the efficiency of their removal.
The advantages of microstraining are: compactness; continuous and automatic operation
(backwash is continuous and employs a non-clogging jet); the low head-loss of 12 to 18
inches; and economical operation. For best performance, a holding tank providing 10 to
20 minutes detention should precede the Microstrainer; this tank would tend to equalize
the flow rates and would catch any heavy material, but it should not be designed as
a sedimentation tank.
Miscellaneous Solutions
Besides sewer separation, off-system storage and in-line treatment, there are other possible
solutions, which can be categorized as in-system storage, land-use improvements, and
monitoring and regulations.
In-system storage is concerned with using the storage capacity of the sewer conduits
themselves to hold the storm water and deliver it to the treatment plant over an extended
period. This can be accomplished by reducing infiltration or friction to increase sewer
capacity. It can also be accomplished by adding inter-connecting branch sewers or relief
sewers. In-system storage will serve to prevent overflows from smaller storms, but is not
practical for larger storms.
Some problems associated with combined sewers result from changes in zoning and land
use, which in turn change surface characteristics to conditions for which existing sewer
systems were not designed. Advanced urban planning on an area-wide basis and stricter
control of land-use practices would assure continued efficient use of a sewer system and
should lower the incidence of pollution in urban storm water runoff. A study to develop
such a relationship has been conducted in Tulsa, Oklahoma (19).
In-system monitoring, with regulation and diversion, is one method being explored by
Seattle, Detroit, and the Minneapolis-St. Paul Sanitary District (19). It assures maximum
utilization of available capacity and minimizes the pollutional effects of any overflows.
With proper instrumentation, the large flows could be diverted to those sewers with low
flows; this is essentially a refinement of in-system storage. Instrumentation required includes
flow measurement, rain-gauge telemetering, conduit liquid-level sensing, and remote
operation of diversion gates. Computer-controlled operation would be required for rapid
transmittal, recording, and feedback of data.
52
-------
Table 11
WQO Combined Sewer Overflow Treatment Projects
Contractor
City of San Francisco
Rand Corp.
American Process
Equipment Corporation
Cornell, Howland,
Hayes & Merryfield
Autotrol
Rex-Chainbelt
Crane Co.
Fram Corp.
Hercules, Inc.
Hydrotechnic
Battelle-Memorial
Project Site Description of Treatment
Short-term sedimentation followed
by dissolved air flotation and chlorina-
tion
Cleveland, Percolation through a shallow bed
Ohio of coal to filter coarser materials.
Coal-solids mixture is later incinerated.
Los Angeles, Ultrasonic filtration.
California
Portland, High-rate, fine-mesh, vibrating
Oregon screens for removal of solids.
Milwaukee, Biological treatment by a rotating
Wisconsin biological contactor.
Milwaukee, Fine screening and dissolved air f lota-
Wisconsin tion for solids removal.
Philadelphia, Microstraining, ozonation, and
Pennsylvania chlorination
Providence, Strainer followed by self-cleaning
Rhode Island diatomaceous earth filter.
Cumberland, Self-cleaning filter of a flexible, filament-
Maryland wound structure that will flex during
storm flow to become permeable.
New York, Multi-media filtration
New York
Richland, Activated carbon adsorption, chemical
Washington coagulation, sedimentation
53
-------
Water quality monitoring of such characteristics as turbidity could assure that less-polluted
flows are selected for discharge. This concept can be further refined by including techniques
to provide selective discharge of overflows at different points in a manner which minimizes
their pollutional effects on a waterway. Such a system is being studied in Seattle.
Another alternative method is to flush the sewerage system intermittently during
dry-weather conditions with municipal water. This would flush out any solid matter which
may have settled along the bottom of the sewer during the slow, dry-weather flows. The
flushing flow must be controlled so as to not overload the hydraulic capacity of the sewage
treatment plant. This method is being studied by FMC (19).
Summary of Previous Approaches
It is highly improbable that a community would have the economic capability to completely
separate its combined sewers. Furthermore, complete separation could be accomplished
only over a long period of time at great inconvenience to urban existence. Partial separation
is less costly, but it probably would not provide the abatement required in many situations.
Extensive land areas are required to provide surface storage sufficiently large either to
hold the larger overflows for later return to the sewerage system or to detain the flow
long enough for proper sedimentation. Sub-surface storage appears to be far more
promising, and it can also provide additional benefits beyond those of pollution control.
It, too, would be an enormous project and feasible only if the bedrock is suitable.
Some treatment processes appear promising, but there is a dearth of basic information.
EPA/WQO is supporting continuing studies in an effort to develop basic data which can
be used to help work out economical solutions; however, few of these projects have yet
produced definitive results. Even so, the extremely variable character of combined sewer
overflows will inhibit general application of these data.
No particular method of solution has been proven to be the least costly for general
application. Each city must evaluate its local situation and determine the method or
combination of methods (and their priority) which will reduce the polluting impact of
this problem to the desired level at the lowest cost and with the least inconvenience to
the citizenry.
54
-------
Ultra-High-Rate Filtration
General Discussion
High-rate filtration has been extensively applied for the removal of suspended impurities
from raw water or wastewater, especially when the impurities are primarily non-volatile
discrete particulates, such as the wastewater from steel mills. In combined sewer overflow
and storm water discharge, large fractions of the suspended waste constituents are
recognized to be non-volatile discrete solids; thus, high-rate filtration may be an effective
treatment method. Despite the many studies that have been undertaken, the status of
filtration development is still in transition from an art to a science. The practical design
parameters for application of the filtration process to treatment of a specific wastewater
must still be determined from results of specific laboratory or pilot-scale investigations.
Moreover, ultra-high-rate filtration (greater than 15 gpm per square foot) must be applied
in order to cope economically with the unique hydraulic characteristics of the combined
sewer overflow or storm water discharge-high discharges within a short time period. This
adds another dimension of uncertainty to the development of a feasible filtration process
to treat the excess urban wastewater derived from intense storms. Therefore, a filtration
study was conducted with the following objectives:
1. To evaluate the applicability of ultra-high-rate filtration to the treatment of
combined sewer overflows.
2. To determine the flocculation effects of chemical additives on the removal of
solids and organic material.
3. To provide a conceptual design basis for pilot-scale or full-scale treatment units.
The principal process variables evaluated in the laboratory program were:
1. Filter media, including type, depth, size, and arrangement.
2. Filtration rates.
3. Effects of addition of flocculants and flocculant.
4. Variation of solids concentration in the wastewater.
5. Backwash rate and quantity.
6. Air-souring rate, duration, and sequence in the backwash procedure.
7. Effluent quality characteristics, including suspended solids, COD, total five-day
BOD, and soluble five-day BOD.
8. Length of filter run.
9. Head loss requirements.
55
-------
Detailed discussion and results of ultra-high-rate filtration studies are presented in
Appendix F and are summarized as follows.
Summary of Findings
The laboratory test program indicated that ultra-high-rate filtration, at rates of 15 or more
gallons per minute per square foot, is a technically feasible process for the removal of
solids and non-soluble BOD from combined sewer overflow. While actual combined sewer
overflow was not used in these tests, a synthetic waste was made using an appropriate
mixture of sanitary sewage, silt from the District area, and lake water. Of the three filter
systems tested, the fiberglass filters performed best, achieving at least 90 percent removal
of suspended solids and 70 percent removal of non-soluble BOD^ at filtration rates of
15-30 gpm/sq.ft. and with reasonably long filter runs (1-3 hours). Comparable effluent
quality was not achieved in tri-media filter runs at filter rates above 10 gpm/sq.ft. Upflow
filtration through a garnet bed was unsatisfactory, largely on the basis of poor effluent
quality.
Soluble BOD removal was negligible in all three filter systems. Even the addition of
activated sludge to the influent wastewater did not significantly improve soluble BOD
removal. The organic content (6005, COD) of the influent wastewater appeared to have
a greater impact on head-loss building than did the suspended solids content. The addition
of flocculants and flocculant aids was not effective in improving the performance of the
fiberglass filters.
The economic feasibility of the fiberglass filter process for ultra-high-rate filtration will
probably require extending the useful life of the fiberglass medium beyond the limits
indicated by the laboratory tests. Improvement of the backwash operations through
modification of underdrain design, stagewise removal of backwash effluent, the use of
air-scouring driven backwash, etc., and development of improved fiberglass bed designs
and fiberglass filter regeneration techniques appear to be promising approaches to extension
of filter life.
56
-------
SECTION VI
DEVELOPMENT OF FEASIBLE ALTERNATIVES
Alternative Approaches
After thorough review and interpretation of the pertinent information, four methods of
abating pollution from combined sewer overflows appeared to offer sufficient promise
to justify consideration as alternative approaches for the District:
1. Sewer Separation.
2. Storage Reservoirs (with treatment at the Blue Plains Plant or other centralized
facilities after the storm subsides).
3. Treatment at Overflow Points.
4. Tunnels and Mined Storage (with treatment at the Blue Plains Plant or other
centralized facilities after the storm subsides).
Sewer Separation
o
As the name implies, this approach consists of complete separation of storm and sanitary
sewers. The initial steps of this program involve continuing the sewer separation program
of Project C as modified by the City of Washington. The details of Project C are contained
in the 1957 report (1) of the Board of Engineers. Storm water would be discharged to
the surface streams through what is now the combined sewer system, and sanitary
wastewater would be conveyed through a new sewer system to the Blue Plains sewage
treatment plant for treatment.
Storage Reservoirs
The concept in this approach is to provide sufficient underground storage volume to hold
the combined sewer overflows caused by each storm until the storm subsides and then
to pump the stored wastewater back into the sewerage system for conveyance to a
centralized treatment plant. This would be accomplished by the construction of shallow,
multi-cell, concrete tanks located five to ten feet underground at reservoir top. Depth
of each reservoir would depend on land availability and volume to be stored; however,
there are various technical, economic, and aesthetic factors that prevent the use of storage
reservoirs in certain locations.
Each reservoir would be compartmented. The initial compartment would function as a
settling chamber to remove grit and heavy solids, with the overflow going to other
compartments for storage. The number of compartments used during any storm would
depend on the rainfall intensity, the duration of the storm and the amount of stored
water remaining from any previous storm A schematic of a typical installation is shown
in Figure 14.
Each reservoir would be equipped with trash racks, a system for flushing sediment from
the reservoir bottom, a ventilation system, pumps for returning stored wastewater into
the regular sewer system and special pumps for transferring accumulated sludge back into
the system or to tank trucks for disposal.
57
-------
FIGURE U
STORAGE RESERVOIRS
STORAGE RESERVOIR SCHEMATIC (TYPICAL)
en
00
SCREENINGS
TO LANDFILL
EXISTING OVERFLOW
TO STREAM
GO
X
CONTROL
EQUIPMENT
VENTILATION
EQUIPMENT
FRESH AIR
INTAKE
S SETTLED OVERFLOW
S PUMPS
CO
s
CO
-------
Due to the extensive urban development within the District, little vacant land is available
near overflow points for storage reservoirs. In certain areas of the District scheduled for
redevelopment (e.g., Georgetown Waterfront), storage reservoirs could be constructed under
planned open spaces and coordinated with the razing phase. In most other overflow areas,
the only available vacant land is park land under the jurisdiction of the National Park
Service; full cooperation of the Park Service would be necessary to use these sites for
reservoir purposes. Since the reservoirs themselves would be underground, only relatively
small pump houses would extend above grade, and the surface area could be used for
parks and playgrounds. Sufficient care to preserve the natural park landscape by proper
architectural design of both the pump houses and maintenance-access structures would
tend to offset objections of the National Park Service.
Treatment at Overflow Points
In this method, there would be a treatment facility at each existing overflow point, except
where conditions either prevent this or dictate that certain overflow points be combined.
The treatment sequence would be sedimentation in the storage reservoirs, followed by
ultra-high-rate filtration and chlorination. A schematic flow diagram for a typical treatment
facility is presented in Figure 15. The individual filters would be ultra-high-rate pressure
filters similar to those used in the treatment of industrial water supplies. Each installation
would include a battery of filters, the number of which would vary depending on the
design flow rate at each location. In addition to the filters, various pumping,
chemical-mixing, and flocculant-aid equipment would be needed.
The filtered wastewater would be disinfected (by chlorination) and discharged to the surface
streams. However, a portion of the filtrate would be stored for use as backwash, initiated
when a predetermined head is reached at any individual filter. The backwash from the
filters would be pumped back into the regular system for conveyance to the Blue Plains
treatment plant. Backwash requirements were estimated to be approximately eight percent
of the forward flow.
In addition to actual treatment facilities, each installation would require equalization
storage because of the extremely high flow rates that would be encountered. (The smallest
sewer district would require a treatment facility with a capacity of several million gallons
per day if no equalization storage were provided.) Even with the high rates (15 gpm/sq.ft.)
involved in ultra-high-rate filtration, the capacity required for equalization would be
essentially the same as that required to capture the entire storm overflow. Therefore, the
actual treatment facilities in this alternative would have to be supplemented by the same
storage capacity required in the storage reservoir method.
The capacity of any treatment facility, since it determines the drawdown rate, is related
to the probability of overflow from a subsequent storm exceeding the capacity of a storage
facility. Regardless of any selected capacity of a storage facility or treatment facility,
it is possible that the combined effect of two or more storms, occurring within a relatively
short period of time, will result in runoff in excess of the reserve storage capacity, even
though the separate runoff from each storm is less than the design capacity. The capacity
of the treatment facility should be sufficiently large so that the probability of two
consecutive storms overflowing a storage facility is reasonably low. A preliminary statistical
analysis suggests that the maximum capacity that can be considered reasonable is one
that draws down a filled storage facility within five days. A drawdown rate of five days,
as compared to 10 days, certainly reduces the possibility of malodors resulting from long
residence times.
59
-------
FIGURE 15
TREATMENT AT OVERFLOW POINTS
TREATMENT FACILITY SCHEMATIC (TYPICAL)
CHLORINE
CONTACT
O)
o
UJ CO
;LUDGE
PUMP
UNDERGROUND STORAGE
-BACKWASH
SUPPLY
^VENTILATION
BACKWASH DISCHARGE TO INTERCEPTOR
SCREENINGS
TO LANDFILL
FIBERGLAS FILTERS
BACKWASH
PUMP
TO RECEIVING STREAM
EXISTING OVERFLOW
VENTILATION
EQUIPMENT
FLUSHING
WATER SUPPLY
STORAGE ZONE
SLUDGE DISCHARGE
-------
Tunnels and Mined Storage
In this concept, combined sewer overflow would drop through a vertical shaft down to
an underground system of tunnels and mined storage. The tunnels would convey the
overflow at high velocity to mined storage. After the storm subsides, the retained overflow
would be pumped back into the regular sewers for conveyance either to the Blue Plains
sewage treatment plant or to a separate, centralized facility constructed specifically to
treat combined sewer overflow.
Vertical Shaft Considerations
In a tunnel-mined storage system, each of the overflow points would have a vertical shaft
for dropping the overflow to the underground system. Certain vertical shafts may be used
for purposes other than the primary function to convey overflow to the tunnels. The
shafts used for access of men and equipment and for removal of drilling muck would
be larger than those to be used for conveyance of overflow.
The larger-diametered (20 and 30-feet) access shafts would be constructed by conventional
techniques (i.e., by drilling and blasting, and by lining with either jacking or slip-form
construction). The smaller (5-foot diameter) shafts would be constructed by simply augering
down from grade to the storage tunnels or by raise-boring of a pilot hole from the
tunnel, followed by pulling a larger-diameter auger up from the tunnel.
Each shaft would be concrete-lined (for structural purposes and to prevent ground water
from entering the system) and would have a baffle to permit the escape of entrained
air to the atmosphere. The tunnel bottom at the junction with each vertical shaft would
be designed to dissipate the energy of the falling water and to provide transition from
the vertical shaft to the conveyance tunnels.
Land requirements for shaft construction (and, therefore, tunnel construction) would be
limited to the land needed for storing construction equipment and a small amount of
excavated material. Preliminary site investigation has shown that sufficient land is available
at nearly all reasonable vertical shaft sites and that only slight modifications to the existing
sewer system would be required to connect the overflow points to the appropriate vertical
shafts.
Tunnel Considerations
The tunnels would be concrete-lined and constructed in bedrock at sufficient depths to
assure the structural integrity of the foundations of all existing bridges, buildings, or
monuments. They would also be located so as to avoid interference with any planned
underground facilities of the Washington Metropolitan Area Transit Authority.
In general, the bedrock geology of the area involved in the proposed storage tunnel
construction meets the significant criteria for successful rock tunneling. As depth increases,
the rock types tend to become less weathered, and the secondary openings become tighter
and less frequent. The Washington area is relatively stable tectonically, and no problems
traceable to rock deformation or faulting are anticipated. The geology of the area and
its impact on tunneling are discussed in more detail in Appendix A.
61
-------
Material excavated during tunneling would be transported to the surface through the vertical
shafts. Depending on its quality, this material could be used for landfill or as concrete
aggregate. If so used, it could substantially reduce the construction cost of this alternative,
but in the cost comparisons to be presented in a later section of this report, the excavated
material will be considered as being hauled away as waste.
Recent developments in the tunnel construction and the tunneling equipment fields
contribute to the feasibility of this approach and to the realistic consideration of available
maximum tunnel diameters. Tunneling equipment, or "moles", capable of boring 28-foot
diameter tunnels are already in use, and those providing 32-foot diameter bores are in
the design stage and should be ready within a few years. Therefore, the size of the
conveyance tunnels would present no added problem in the drilling operations.
The Conveyance Tunnel/Mined Storage approach has many of the conceptual features of
the Chicago Deep Tunnel Plan. The Chicago Plan also incorporates a pumped storage
hydroelectric generating system to provide additional electric power during the peak
consumption periods. However, preliminary investigation has indicated that such
supplementary power generation is not economically feasible for the Washington D.C. area.
Therefore, power generation has not been included in the Conveyance Tunnel/Mined
Storage alternative.
Mined Storage Considerations
The mined storage would consist of a network of criss-crossed chambers at atmospheric
pressure. Conventional deep-mining methods (i.e., drilling and blasting to remove the rock)
and mine-railroad or rubber-tired earth-moving equipment to convey the excavated material
to the access shaft would be used in construction of this mined-storage area. Storage area
would be compartmented, with some compartments used as settling basins to facilitate
solids removal. The chambers would be concrete-lined only in areas where faults or fractures
would otherwise permit excessive inflow of ground water. Final design, as with the final
design of tunnels, would depend on data collected during sub-surface geological
investigation.
Pumping Station Considerations
A pumping station would be included with each storage system either at the bottom
of one of the vertical shafts, or in a specially-excavated chamber nearby. The shaft involved
would have elevators, discharge piping, ventilation, and electrical conduits. The design of
the pumping system would be based on the capability of the existing sewer system to
handle the stored overflow after the storm subsides. This capability in turn would be
influenced by variations in the dry-weather flow, and the pumping of the stored overflow
would have to be programmed to coordinate with the off-peak hours of normal dry-weather
flow.
62
-------
Solids Removal Considerations
If the tunnels are constructed on a reasonable slope, sufficient flow velocity will result
to convey most of the solids to the mined-storage area. The initial chambers in the
mined-storage area will provide sedimentation and will concentrate the deposition of solids
there rather than uniformly over the entire bottom of the mined-storage area. This should
permit effective removal of at least the lighter solids by pumping. If the removal of heavier
solids in this manner presents a problem, their removal may be effected at the point
of overflow by using vortex separators (compact, cyclone-type equipment that remove
grit with centrifugal force). The separated grit would be discharged as a slurry to a classifier
which washes and dewaters the grit for pickup by truck.
63
-------
Application of Alternative Approaches to D.C.
General
It is especially fitting to evaluate each of the four approaches by individual sewer drainage
basins because the method selected to deal with the combined sewers in one part of the
District does not have to be the same method used in another part. Actually, the
appropriate solution to the total combined sewer problem in the District may include
features from each of these four methods.
Although it is possible to apply each of these four approaches at any sewer drainage
district, physical and economic factors dictate that only one or two methods be employed
for each district. The very large volume of overflow from certain districts in combination
with existing land use requires the use of mined storage rather than reservoir storage if
a storage method is used. The proximity of some overflow points or existing land use
suggests that certain districts be interconnected. Because the available or proposed
interceptor capacity is quite large in places, it is appropriate to use this capacity to convey
combined sewer flow from certain districts during storm conditions to a more centralized
storage facility at a subsequent point in the sewer system.
For the purpose of this evaluation, the District was divided into two large drainage basins,
Anacostia and Rock Creek-Upper Potomac. The basic layout of the alternative systems
was based on retaining and treating the overflow from the 15-year, 24-hour storm (see
Table 5); however, the same design should apply to other major return frequencies with
the exception that certain dimensions (e.g., tunnel diameter) will change.
Anacostia River Area
Conveyance Tunnels and Mined Storage
The basin layout for this approach is shown in Figure 16. Essentially the plan proposes
the dropping of all overflows through vertical shafts into conveyance tunnels that empty
into a single mined-storage area near the Robert F. Kennedy Stadium. As mentioned in
the description of the present D.C. sewer system (in Section IV of this report) and as
shown in Table 2, the Northeast Boundary Trunk Sewer does not have sufficient capacity
to convey the runoff from a 15-year storm. The District of Columbia currently has
preliminary plans to construct a relief sewer in the Northeast Boundary area at a cost
of $33,600,000 (1968 dollars, ENR = 1117); however, the relief sewer would behave as
a combined sewer for a considerable period of time until sewer separation was completed.
Even after complete separation, the relief sewer would discharge with each storm a
considerable pollution load represented in the runoff from a drainage area one half as
large as the G-2 district.
The construction of storage facilities in the upper and central reaches of the G-2 sewer
system would not only provide additional storage capacity, but more importantly, could
also provide sufficient relief to ease surcharging during the 15-year storm. In contrast
with the relief sewer, the relief/storage system has the added feature of preventing discharge
of untreated runoff and sanitary sewage to surface streams.
A recent study (6) proposed a combination relief/storage tunnel in the central part of
the sewer district as presented in Figure 17. This tunnel would reduce most of the
64
-------
o
CJ1
EAST SIDE INTERCEPTOR
RELIEF SEWAGE PUMPING
STATION
LEGEND
" — 'POLITICAL BOUNDARY
—— COMBINED SEWERS
llhllUIIHIMIH SANITARY SEWERS
,_._,_.» OTHER EXISTING SEWIRS
(COMBINED AND SEPARATED)
A-10 SEWER DISTRICT NUJMCt
—_-^--^ LIMITS OF COMBINED SEWER SYSTEM IN 1970
-— SEWER DISTRICT BOUNDARIES
D PUMPING STATIONS
»^~^~^« TUNNEL
£ VERTICAL SHAFT
• PUMPING STATION
MINED STORAGE
NO.
NAME
E-3 B ST.-NEW JERSEY AVE. TRUNK SEWER D/S PORTION
F-l TIBER CREEK
G-2 NORTHEAST BOUNDARY
G-3 BARNEY CIRCLE
G-4 UTH ST.-PENNA. AVE.
G-5 12TH ST.-9TH ST.
G-7 6TH ST.-7TH ST.
1500
1500 3000 4500
SCALE IN FEET
6000
DISTRICT OF COLUMBIA
TUNNELS AND MINED STORAGE IN
ANACOSTIA RIVER BASIN
FIGURE 16
-------
FIGURE 17
LOCATION OF NORTHEAST BOUNDARY RELIEF/STORAGE TUNNEL
MICHIGAN
AVE.
MC MILLIAN
RESERVOIR
UNION
STATION
LEGEND
^??' - TUNNEL AREA
- VERTICAL SHAFT
- PUMPING STATION
NO SCALE
PUMPING
STATION
66
17
-------
surcharging in this sewer district; however, it would not provide total relief. It is conceivable
that construction of only part of the proposed relief sewer or some strategically located
storage reservoirs, in combination with the relief/storage tunnel, could prevent surcharging
at a lower cost than the entire relief sewer. It is also possible that a different tunnel
layout, with tunnels running perpendicular to the trunk sewer but beneath critical branch
sewers, may also prevent surcharging at a lower cost.
The other basic components of the tunnel and mined-storage approach include:
1. A 20- to 22-foot I.D. tunnel extending from the Main Sewerage Pumping Station
on the west bank of the Anacostia River northeastward (parallel to the river)
to the vicinity of the Robert F. Kennedy Stadium. The total length of the
Anacostia conveyance tunnel would be approximately 10,600 feet. The first
3,900 feet (starting from the Main Sewerage Pumping Station) would be 20
feet I.D., and the remainder would be 22 feet I.D. Figure 18 provides a profile
of this tunnel.
2. A 3,700,000-cubic yard mined-storage area in bedrock near the Robert F.
Kennedy Stadium. This underground storage facility, to be located beneath the
parking lot of the Robert F. Kennedy Stadium, would eliminate the need for
a separate tunnel or other facility to convey the extremely large flow from the
Northeast Boundary Trunk Sewer. This location for the mined storage would
also relieve portions of the Northeast Boundary Trunk Sewer during periods
of high flow.
3. An underground pumping station and a series of vertical overflow and access
shafts.
Storage Reservoirs
Figure 19 shows the basic layout of a storage system incorporating shallow underground
reservoirs wherever they are reasonable. Storage reservoirs are not applicable throughout
the Anacostia River area; they would be attractive only for the Sewer Districts with smaller
flows, namely G-3, G-4, G-5, and G-7. The required capacities for these concrete tanks
range from 800,000 gallons to 28,000,000 gallons.
The very large volume (490 million gallons) and overflow rate (5.5 billion gallons per
day) from sewer district G-2 dictate the use of an underground mined-storage chamber
beneath the overflow point near Robert F. Kennedy Stadium as the most effective means
of capturing the overflow from the Northeast Boundary Trunk Sewer. This mined-storage
area is similar to the mined-storage area proposed before, except that the capacity is not
as large.
For Sewer Districts F-1 and E-3, existing land use and the high volume (230 million gallons)
of overflow for the outfall at the Main Sewerage Pumping Station dictate that the overflow
at these two structures be conveyed to the east side of the Anacostia River for retention
in a mined-storage facility. These mined-storage areas are suggested only if storage reservoirs
are utilized in the Anacostia River; otherwise, a single larger mined-storage area near the
stadium is suggested.
67
-------
FIGURE 18
CONVEYANCE TUNNELS AND MINED STORAGE
ANACOSTIA RIVER TUNNEL PROFILE (SEE FIGURE 16 FOR PLAN)
VERTICAL
SHAFTS
EAST SIDE
INTERCEPTER
ROBERT F. KENNEDY E»ST SIDE
STADIUM /"RELIEF
INTERCEPTER
-100
-300
MAIN SEKERAGE
PUHPING STATION
RAVEL,SAND.
SILT.AND CLAY
0 SEA LEVEL
-350
-400
G-5 G-4 C-3
20 1.0. TUNNEL
PUMP STATION
WITH ACCESS
SHAFT
7
G-2
MINED
STORAGE
== -450
-500
-550
-600
li' i 0 TUNNEL
FIG
68
18
-------
•£ -
EAST SIDE INTERCEPTOR
RELIEF SEWAGE PUMPING
STATION
LEGEND
POLITICAL BOUNDARY
COMBINED SEWERS
IIIIIMllll SANITARY SEWERS
,_._._,. OTHER EXISTING SEWERS
(COMBINED AND SEPARATED)
A-10 SEWER DISTRICT NUMBER
,~-»_-^ LIMITS OF COMBINED SEWER SYSTEM IN 1970
_ SEWER DISTRICT BOUNDARIES
D PUMPING STATIONS
STORAGE RESERVOIRS
STORAGE TUNNELS
MINED STORAGE
NO.
E-3
F-l
G-2
G-3
G-4
G-5
G-7
B ST.-NEW JERSEY AVE. TRUNK SEWER D/S PORTION
TIBER CREEK
NORTHEAST BOUNDARY
BARNEY CIRCLE
14TH ST.-PENNA. AVE.
12TH ST.-9TH ST.
6TH ST.-7TH ST.
1500
1500 3000 4500
SCALE IN FEET
6000
DISTRICT OF COLUMBIA
MAXIMUM USE OF STORAGE RESERVOIRS IN
ANACOSTIA RIVER BASIN
FIGURE 19
-------
The relief/storage tunnel in the central reach of the Northeast Boundary Trunk Sewer
is required in view of the surcharging frequency of this sewer; however, in this approach,
total use of storage reservoirs rather than a tunnel may be desirable.
An independent study (6) of the G-2 Sewer District has identified a feasible location
for a large storage reservoir near the Robert F. Kennedy Stadium. This location could
be developed into a storage capacity of 235 million gallons. While this is not sufficient
for the overflow from the 15-year, 24-hour storm, this storage capacity plus the capacity
of the relief tunnel would be more attractive than mined storage for design storms of
less overflow volume.
Treatment at Overflow Points
The conceptual design of this approach essentially follows the layout for the storage
reservoir approach, except that treatment plants are required near points of storage. The
locations of the treatment facilities are shown in Figure 20. As stated before, the treatment
plant capacities are sufficiently large to draw down each filled storage facility within five
days.
The most attractive feature of this method is that is provides facilities to treat stored
overflow. As mentioned in the discussion of the D.C. sewage treatment facilities, the average
dry-weather flow to the only sewage treatment facility within the District will exceed
the ultimate capacity at this plant in 1977. While the Blue Plains plant can handle for
short durations flow in excess of this ultimate capacity, the plant will not be able to
fully treat all increased flows when stored overflow is returned to the sewer system unless
some of the dry-weather flow to this plant is diverted and treated elsewhere. Hence,
treatment facilities at overflow points will probably be needed.
A recent study (6) has suggested an approach for the Northeast Boundary Trunk Sewer
District (G-2) similar to the method of treating at overflow points, except that the project
would serve many purposes other than pollution abatement. The project is referred to
as the Kingman Lake Project. It envisions the retention of overflow from the G-2 sewer
district in a surface reservoir, followed by treatment of the stored overflow and reuse
for recreational purposes, e.g., swimming and fishing. The degree of treatment required
at the Kingman Lake treatment facility is obviously higher; but other than this, the features
of the Kingman Lake project follow exactly the basic concepts suggested in this study,
i.e., relief tunnel and surface storage, possibly augmented with mined storage. The
EPA/WQO, the National Park Service, and the District of Columbia have shown considerable
interest in this project since it would bring a much needed recreational park to the urban
dweller and would rid the urban area of a serious source of pollution.
Another attractive feature of local treatment, is that treatment plants can be located
underground near some of the artificial pools in Washington that sometimes are polluted
by storm water runoff. Treatment of the pool water will restore their beauty rather than
let debris and biotic growths accumulate.
70
-------
•f) -^
EAST SIDE INTERCEPTOR
RELIEF SEWAGE PUMPING
STATION
"**-\
*v
LEGEND
' •— POLITICAL iOUNDARY
——— COMIINED SEWERS
NIIIIIIIIIIHIIII SANITARY SEWERS
,_,„,_,. OTHER EXISTING SEWERS
ICOMIINED AND SEPARATED)
A- 10 SEWER DISTRICT NUMBER
*-.—«-—•. LIMITS OF COMBINED SEWER SYSTEM IN 1970
- • -- SEWER DISTRICT BOUNDARIES
D PUMPING STATIONS
TREATMENT FACILITIES
STORAGE TUNNELS
MINED STORAGE
NO.
NAME
E-3 B ST.-NEW JERSEY AVE. TRUNK SEWER D/S PORTION
F-l TIBER CREEK
G-2 NORTHEAST BOUNDARY
G-3 BARNEY CIRCLE
G-4 UTH ST.-PENNA. AVE.
G-5 12TH ST.-9TH ST.
G-7 6TH ST.-7TH ST.
1500
1500 3000 4500
SCALE IN FEET
6000
DISTRICT OF COLUMBIA
TREATMENT AT OVERFLOW POINTS IN
ANACOSTIA RIVER BASIN
FIGURE 20
-------
Rock Creek Upper Potomac Area
Conveyance Tunnels and Mined Storage
The basic layout for this approach is shown in Figure 21. Essentially the plan proposes
the dropping of all overflows through vertical shafts into conveyance tunnels that empty
into a single mined-storage area near Water Gate. The basic components of this system
include:
1. A 15- to 18-foot I.D. tunnel extending from the outfall of the Piney Branch
Trunk Sewer in a southerly direction (approximately parallel to Rock Creek)
to the vicinity of the Water Gate. The total length of the Rock Creek conveyance
tunnel would be approximately 19,800 feet, of which the first 4,600 feet (starting
from D-4) would be 15 feet I.D. The profile of this tunnel is shown in Figure
22.
2. A 2- to 7.5-foot I.D. sewer/tunnel extending from the intersection of 36th Street
NW and the Potomac River eastward to intersect with the Rock Creek Tunnel
near New Hampshire Avenue NW (see Figure 23). The Upper Potomac tunnel
would be 5,300 feet long, of which the first 1,600 feet would be 24-inch sewer
pipe and the remainder 7.5-foot I.D. tunnel.
3. A 1,100,000-cubic yard mined-storage area in bedrock beneath Water Gate plus
a pumping station.
4. A vertical shaft at each point of overflow. Three 30-foot diameter access shafts
are contemplated, one each in sewer districts A-10, D-4, and E-2. The E-2 shaft
would have the additional function of providing space for the pumping
equipment. There will be no overflow from a number of sewer districts along
the eastern lower part of Rock Creek (e.g., Districts D-13 through D-16 and
part of C-25) due to the large interceptor capacity available and the relief
resulting from capture of flow upstream. Any plan in the Rock Creek Area
should propose routing this flow directly to the Blue Plains plant rather than
retain it in any storage facility.
Storage Reservoirs
Figure 24 shows the basic layout of a storage system incorporating as many underground
storage reservoirs as reasonable. Storage reservoirs offer promise in the vicinity of
Georgetown; however, in other parts of the Rock Creek-Upper Potomac area, tunnel tunnel
storage and mined storage are indicated.
Storage reservoirs would be attractive only for the sewer districts with smaller flows: A-10,
A-11, A-12, B-6, B-7, and C-25. Because of the present land use in the vicinity of the
A-11 overflow point, it would be necessary to convey the overflow to the A-10 overflow
point for reservoir storage or treatment. Likewise, districts B-6 and B-7 should be combined.
The required capacities for these concrete tanks range from 3,000,000 to 21,000,000
gallons.
The large volume (160 million gallons) and high flow rate (2.7 billion gallons per day)
from the overflow point at Piney Branch (D-4) indicate that mined storage or a network
of underground tunnels would be the most effective method of handling the storm
72
-------
-J
CO
Illlllllllllllllfl
A-10
DISTRICT OF COLUMBIA
TUNNELS AND MINED STORAGE IN
UPPER POTOMAC-ROCK CREEK BASIN
LEGEND
POLITICAL BOUNDARY
COMBINED SEWERS
SANITARY SEWERS
OTHER EXISTING SEWERS
(COMBINED AND SEPARATED)
SEWER DISTRICT NUMBER
LIMITS OF COMBINED SEWER SYSTEM IN 1970
SEWER DISTRICT BOUNDARIES
PUMPING STATIONS
TUNNEL
VERTICAL SHAFT
PUMPING STATION
MINED STORAGE
NO.
A-10
A-ll
A-12
B-6
B-7
C-25
D-4
D-5
D-6
D-7
D-8
D-9
D-10
D-ll
D-12
E-2
NAME
37TH ST.-GEORGETOWN
GEORGETOWN
K ST.-WISCONSIN AVE.
M ST. -27TH ST.
28TH ST.-WISCONSIN AVE.
SLASH RUN (PART]
PINEY BRANCH
OAK ST.-MT.PLEASANT ST.
INGLESIDE TER.
PARK RD.
LAMONT ST.
KENYON ST.
IRVING ST.
QUARRY RD.
ONTARIO RD.
EASBY POINT
1500
1500 3000
SCALE IN FEET
4500 6000
FIGURE 21
-------
200
FIGURE 22
CONVEYANCE TUNNELS AND MINED STORAGE
ROCK CREEK TUNNEL PROFILE (SEE FIGURE 21 FOR PLAN!
XISTING POTOMAC RIVER
FORCE MAINS
UNCONSOLIDATEO
(30' DIAMETER
FOR ACCESS)
ZONE OF
DISINTEGRATED ROCK
r~^-^-^rI.~=--^=r=-- SHIFTS
7.5' I.D.
UPPER POTOMAC
__-: / TUNNEL
:SEE FIGURE 4B:
(30' DIAMETER
FOR ACCESS)
-250-
E-
C- 5
22
D 11 D 9
-------
FIGURE 23
CONVEYANCE TUNNELS AND MINED STORAGE
UPPER POTOMAC TUNNEL PROFILE (SEE FIGURE 21 FOR PLAN)
UNCONSOLIDATED
MATERIAL
24"I.D.SEWER
BEDROCK
(CAM EXPECT SOME OUTCROP
OF BEDROCK,AND ZONES
OF UNCONSOLIDATED
MATERIAL APPROX.30-40'
THICK)
ZONE OF DISINTEGRATED
ROCK
S= 0 SEA LEVEL
8' I.D. ROCK CREEK
TUNNEL
(SEE FIGURE 22)
150
^f^_:: .IT .~^C_! ~^ n.--^i:=L ^ 7.5 • I.D. TUNNEL
A-10
A-ll A-12 B6.B7
r~—"^ -200
75
23
-------
-J
O)
LEGEND
-— POLITICAL IOUNDARY
COMIINED SEWERS
IIINNHMNMI SANITARY SEWERS
1 —,_1 —.. OTHER EXISTING SEWERS
ICOMUNED AND SEPARATED)
A-10 SEWER DISTRICT NUMtER
*—~*~,-~* LIMITS OF COMIINED SEWER SYSTEM IN 1970
— SEWER DISTRICT lOUNDARIES
Q PUMPING STATIONS
H STORAGE RESERVOIRS
STORAGE TUNNELS
MINED STORAGE
NO.
A.10
A-ll
A-12
B-6
B-7
C-25
D-4
D-5
D.6
D-7
D-8
D-9
D-10
D-ll
D-12
E-2
NAME
37TH ST.-GEORGETOWN
GEORGETOWN
K ST.-WISCONSIN AVE.
M ST. -27TH ST.
28TH ST.-WISCONSIN AVE.
SLASH RUN (PART)
PINEY (RANCH
OAK ST.-MT.PLEASANT ST.
INGLESIDE TER.
PARK RD.
LAMONT ST.
KENYON ST.
IRVING ST.
QUARRY RD.
ONTARIO RD.
EASBY POINT
1500
1500 3000
SCALE IN FEET
4500 6000
DISTRICT OF COLUMBIA
MAXIMUM USE OF STORAGE RESERVOIRS IN
UPPER POTOMAC-ROCK CREEK BASIN
FIGURE 24
-------
overflow. Since the D-5 outfall is only about 300 feet downstream of D-4 and since the
D-5 overflow volume would be small, it is logical to combine the overflows from these
two districts and collect them in the underground storage required for D-4. Likewise,
mined storage is indicated for sewer district E-2 by itself if storage reservoirs are used
in the Georgetown area. The close proximity of the points of overflow from sewer districts
D-8 through D-12 and the relatively low hydraulic characteristics make it advantageous
to connect these overflow points by means of tunnels (or conventional cut-and-fill sewer
construction). Likewise, districts D-6 and D-7 should be connected in a similar manner.
Having connected these districts, simply expanding the connecting system into shallow
tunnel storage is probably less expensive than construction of storage reservoirs.
Nevertheless, the small volume of overflow (less than three million gallons) from these
seven districts indicates that the costs to control their overflow is minor in comparison
to other sewer districts.
Treatment at Overflow Points
Again, the conceptual design of this approach follows the layout for the storage reservoir
method, except that treatment facilities are required near points of storage. The locations
of the required treatment facilities are shown in Figure 25. Again, the most attractive
feature of this approach is that it provides facilities for treating stored overflow.
77
-------
00
LEGEND
' --- — POLITICAL lOUNOAiY
COMIINED SfWfIS
IIIIMMIMIimil SANITAIY SfWEIS
._._,_.. OTMfl EXISTING SEWERS
(COMBINED AND SEPARATED)
A. 10 SEWER DISTRICT NUMIER
«— «^_—». LIMITS OF COMUNED SEWER SYSTEM IN 1970
- ---- SEWEi DISTRICT IOUNDARIES
O PUMPING STATIONS
TREATMENT FACILITIES
STORAGE TUNNELS
MINED STORAGE
NO.
A-10
A-11
A-12
B.6
B-7
C-25
D-4
D-5
D-6
D-7
D-8
D-9
D-10
D-ll
D-12
E-2
NAME
37TH ST.-GEORGETOWN
GEORGETOWN
K ST.-WISCONSIN AVE.
M ST. -27TH ST.
28TH ST.-WISCONSIN AVE.
SLASH RUN (PART]
PINEY BRANCH
OAK ST.-MT.PLEASANT ST.
INGLESIDE TER.
PARK RD.
LAMONT ST.
KENYON ST.
IRVING ST.
QUARRY RD.
ONTARIO RD.
EASBY POINT
1500
1500 3000
•SCALE IN FEET
4500 6000
DISTRICT OF COLUMBIA
TREATMENT AT OVERFLOW POINTS IN
UPPER POTOMAC-ROCK CREEK BASIN
FIGURE 25
-------
SECTION VII
EVALUATION OF FEASIBLE ALTERNATIVES
General
The preceding sections of this report have defined the variability and impact of combined
sewer overflows in the District and have presented four approaches for dealing with the
problem. From these approaches, it is possible to develop numerous alternative strategies
within the District for collecting, storing, and treating overflow. For example, one strategy
may involve conveyance tunnels and mined storage for the overflow from the 2-year,
24-hour storm in the Anacostia River area while calling for storage reservoirs and treatment
for the overflow from the 25-year 24-hour storm in the Rock Creek-Upper Potomac area.
To provide a sound basis for decision-making, it is necessary to define the consequences
of promising alternatives in detailed and comparative terms. Costs and impact on water
quality demand the most attention, but non-quantifiable aspects, such as public
convenience, also influence the selection of the appropriate solution. The four District-wide
approaches previously presented in Figures 16 thru 25 are considered sufficiently varied
to offer a wide choice of strategies throughout the District. The approaches presented
were:
1. Maximum use of storage reservoirs
2. Treatment at points of overflow
3. Conveyance tunnels and mined storage
4. Sewer separation
Again, it is pointed out that the approach selected for one sewer basin does not have
to be used in the other basins; the appropriate District-wide solution may incorporate
features from each of these four approaches.
The following sections present the costs and facilities associated with individual sewer
districts for each approach. The indicated reduction in pollutant loadings for alternative
design-storm frequencies is also indicated. Finally, to permit a balanced assessment,
non-quantifiable factors (those other than cost and pollutant reductions) are discussed
and compared.
79
-------
Cost Analysis
Construction Costs
Cost is an essential factor in selecting the appropriate alternative for abating combined
sewer overflows. Estimates have been made for the costs involved in the alternatives which
have been specifically discussed. A construction cost and escalation contingency of 15
percent has been included in all estimates, as well as 6 percent for engineering and 15
percent for resident inspection, bonding, etc.; but it should be noted that the cost estimates
do not include Surveying, Soils Investigations, Land Acquisition, or Rights-of-Way.
If it will take the responsible agencies two years from the time of release of this report
to act on appropriation of funds for any alternative (which is a reasonable assumption
in view of the magnitude and complexity of the project), construction would probably
start about the middle of 1973. To assure that the cost estimates are consistent with
that date, the total cost estimates reflect an Engineering News-Record Construction Cost
Index of 1800, the projected index for June, 1973. The construction costs specific to
any date can be determined by multiplying the costs presented herein by the ratio of
the then current ENR index to 1800.
General Procedure
The same general procedure was used to estimate the construction costs of the alternative
methods of storage reservoirs, treatment at overflow points, and tunnel and mined storage.
First, a detailed cost estimate was made of the facilities required to handle the 15-year,
24-hour storm at each sewer district. This was done for each of the three approaches
illustrated in Figures 16 through 25. This cost analysis included preliminary on-site
inspection to consider land availability, terrain, etc. in selecting sites for facilities. All
the unit cost estimates are based on December 1969 costs for similar construction, with
the total costs increased by the ratio of the 1800 ENR index to 1300, the national average
during December 1969.
Following the analysis of the 15-year, 24-hour storm graphs were made of total costs
versus capacity relationships for the following components:
1. Storage reservoirs and appurtenant equipment
2. Tunnels
3. Mined storage
4. Vertical shafts
5. Treatment plants
6. Pumping stations
The many different capacities required for the various sewer districts provided a suitable
span of plotting points. These costs curves were then used to estimate the costs for handling
the overflow from each sewer district for the 2-year, 5-year, and 25-year frequency, 24-hour
duration storms.
Storage Reservoir Construction Costs
The storage reservoirs would be shallow-underground, multi-cell, concrete tanks. Figure
26 shows the relationship for construction costs of reservoirs and appurtenant equipment
versus capacity. Typical costs range from $900,000 for a 600,000-gallon reservoir to
80
-------
FIGURE 26
TOTAL CONSTRUCTION COSTS OF
STORAGE RESERVOIRS AND APPURTENANT EQUIPMENT
ENR= 1800
20
10
1
.2 .3 .4 .5 .6 .8 1 2 3 4 5 6 8 10 20 30
STORAGE CAPACITY, MILLION GALLONS
81
26
-------
for a 32,000,000-gallon reservoir. The construction costs include the costs of site
preparation, excavation, concrete, backfill, buildings, landscaping, roadways, and
mechanical and electrical construction.
Costs are based on simultaneous construction of all reservoirs. Some advantage may be
gained by building only a portion of the reservoirs at the beginning of the program. The
knowledge and experience gained during the design and construction of the initial reservoirs
could be used to reduce the construction costs of subsequent reservoirs.
Tunnel Construction Costs
These costs were based on up-dated bid prices of current tunnel construction projects
in Chicago and California in which the boring method is utilized. Information obtained
from one of the largest manufacturers of drilling equipment indicated that the rapid
development of boring equipment will tend to moderate increases in these prices even
in the face of rising costs. Figure 27 defines the unit cost versus diameter relationships
for tunnel construction. The following important assumptions have been made with regard
to construction costs for tunnels:
1. One machine would be used to construct the Upper Potomac-Rock Creek tunnels,
and a separate machine for the Anacostia River tunnel.
2. All tunnels would be concrete lined. (12" lining at an unit cost of $100 per
cubic yard of concrete).
3. Material excavated from the the tunnel would be hauled away as waste at a
cost of $5.00 per cubic yard.
4. Boring would be at an average rate of 36 feet per day.
5. There would be no salvage value for the boring machine or mucking equipment.
If the quality of material excavated from tunnel and mined-storage construction is such
that it could be used as a building material, the construction costs would be reduced
substantially. The structural characteristics of the material would be determined during
the sub-surface investigation, and the physical properties (size and shape) of the aggregate
would be determined at the time of construction.
Although boring rates of approximately 200 to 300 feet per day have been considered
possible in rock similar to that underlying Washington, an average of 36 feet per day
over the construction period seems more reasonable for purposes of cost estimating.
Boring machines are generally custom built for a particular project and may not be usable
on a different project. For this reason, no salvage value has been allowed. Other
equipment-muck cars, locomotives, dust, cable, etc.,-will have some small salvage value,
but this will be negligible compared to total construction costs.
The construction costs also include power and labor. Labor costs and concrete costs are
the most significant, accounting for over 60 percent of all costs.
82
-------
FIGURE 27
UNIT CONSTRUCTION COSTS OF TUNNELS
ENR = 1800
1500
c/o
I—
CO
1000
900
800
700
600
500
7.5
10
15 20
DIAMETER, FEET
25
30 32
83
27
-------
Mined Storage Construction Costs
Mined storage construction costs are based on similar costs for underground mining and
include labor, equipment, and handling of excess material. The unit cost for mined-storage
construction is about $20 per cubic yard but varies depending on the volume to be
excavated. Figure 28 illustrates the relationship between construction costs and
mined-storage capacity. This curve accounts only for the cost of mining and does not
include pumping stations and vertical shafts.
Vertical Shaft Construction Costs
Three basic types of vertical shafts are expected:
1. Relatively small shafts (for the smaller sewer districts) drilled from the ground
surface to the tunnel.
2. Slightly larger shafts, five to six feet in diameter, (for the larger sewer districts)
to be constructed by the raise-boring method. The unit cost of these shafts
is about $200-$250 per vertical foot.
3. Large-diameter (30 feet) shafts, to be used for access by men and equipment
as well as for diverting flow to the tunnel. Construction techniques would be
conventional slurry trench or jacking methods. Large-diameter shafts would also
be used for pumping and appurtenant equipment. The unit cost of these shafts
is about $5,000-$6,000 per vertical foot.
Pump Station Construction Costs
Pump station construction cost estimates include the cost of pumping equipment, controls,
sludge facilities, superstructures, heating, ventilating and electrical work, instrumentation,
metering, inside piping and valves, and sitework. The cost of the access shaft for pumping
equipment is included in the vertical shaft construction costs.
Treatment Plant Construction Costs
The treatment process recommended is filtration preceded by sedimentation and followed
by disinfection. Three cost estimates were developed from the costs of the major
components of 1-mgd, 10-mgd, and 100-mgd plants. Figure 29 illustrates the relationship
of total construction costs to capacity as suggested by these three estimates. For a 10-mgd
plant, the costs (ENR = 1300) of various plant components were as follows:
Construction
Component Costs
$ 500,000
Cnlormator and Contact Tank gg
'
Pumps 40'OQO
Chemical Feed and Mixing gg QQQ
Instruments, Electrical, Piping 2?o'oOO
Building 250:000
„ . $1,240,000
Contingency 190.000
TOTAL $1,430,000
84
-------
FIGURE 28
TOTAL CONSTRUCTION COSTS OF MINED STORAGE
ENR= 1800
200
CO
<=>
100
90
80
70
60
50
40
30
20
0
50 60 70 80 90 100 200 300 400 500
CAPACITY, MILLION GALLONS
800 1000
85
28
-------
00
o>
ro
>o
FIGURE 29
TOTAL CONSTRUCTION COSTS OF TREATMENT PLANTS VERSUS CAPACITY
ENR= 1800
10
9
8
7
6
5
3 4 5 6 7 8 910 20 30
TREATMENT PLANT CAPACITY, MILLION GALLONS PER DAY
40 50 60 70 80 100
90
-------
Two important assumptions have been made in determing the flow rate of the treatment
facility to be located at each overflow point:
1. The maximum residence period in a storage structure will be five days.
2. The minimum capacity of a treatment facility will be 0.2 mgd.
The relationship illustrated in Figure 29 was used to determine the costs of the centralized
facilities required to treat stored overflow, in addition to being used to define the costs
of treatment facilities at points of overflow.
Total Construction Costs
The .four alternative District-wide approaches are as follows:
1. Maximum use of storage reservoirs
2. Treatment at points of overflow
3. Conveyance tunnels and mined storage
4. Sewer separation
Tables 12, 13, 14, and 15 present the construction costs by individual sewer drainage
basins for the four approaches. It must be emphasized again that the appropriate
District-wide solution may include features from each of these approaches. To facilitate
the selection of this solution, the costs and facilities associated with each approach are
identified by sewer district.
As stated before, these costs correspond to an ENR index of 1800. Also, these costs
exclude the costs of dealing with those sewer districts scheduled for separation prior to
1975. However, budgeting problems have stopped the program of separating those sewer
districts: therefore, in the final analysis, the total costs in Tables 12 to 15 will have to
be increased. The incremental costs attributable to those sewer districts is probably slight
because their total area is about 400 acreas, representing less than 4 percent of the collective
area of all combined sewer districts. The added costs would be from $6,000,000 to
$12,000,000, depending on the design storm, for projects not involving separation, and
$27,000,000 for the separation project.
In the approach proposing treatment at overflow points, equalization storage in the amount
equal to the total volume of overflow would be required because the rate of overflow
(even from the smallest drainage area) is extremely high. The cost of constructing treatment
facilities is, therefore, in addition to the reservoir storage construction costs. The difference
in costs between this approach and the reservoir approach is explained by the economies
of scale resulting from constructing a single centralized treatment facility. Inclusion of
the Kingman Lake project in the approach proposing treatment at points of overflow
would increase the costs of the G-2 sewer district from $45,000,000 to $56,700,000 for
a 2-year, 24-hour design event. The relatively small increase in cost for incorporating an
approach that proposes complete treatment is explained by the fact that, for the 2-year
design storm, it is feasible to construct a sufficiently large storage reservoir. In contrast,
the cost estimate listed in Table 13 for District G-2 was based on more costly mined
storage.
87
-------
Table 12
Maximum Use of Storage Reservoirs'
Sewer
District
A 10
All
A 12
B-6
B7
C25
D-4
0-5
D-6
D7
0-8
D-9
D-10
D-11
D 12
E-2
E-3
F 1
G-2
G-3
G-4
G-5
G-7
Storage Total
Northeast Boundary
Relief/Storage Tunnel
Centraliz-d Treatment
Plant
Total Construction Cost
Engineering Design at * 6%
Other Costs (resident
inspection, legal, bonding.
and administrative)
at J 15%
Total Project Costs^
Type of
Storage
Reservoir
Reservoir
Reservoir
Reservoir
Mined
Tunnel/Reservoir
Tunnel/Reservoir
Mined
Mined
Mined
Reservoir
Reservoir
Reservoir
Reservoir
2 Years • 24-Hour
Volume Cost
million
gallons
11. S 7,600,000
2 2,200.000
-
1 1,300,000
65 15,200.000
1.7 2,000.000
4.4 3.900.000
97 o 18,300,000
128 23,300,000
280 45,000,000
0.2 400,000
15 9,600,000
6 4,800,000
5 4,300,000
$138,000.000
27,000,000
10.000.000
175,000,000
1 1 ,000,000
26,000,000
$212.000,000
Estimated Capital Costs
IENH Index = 1800)
Storm Frequency and Duration
5 Years • 24-Hour 15Yei
Volume Cost Volume
million million
gallons gallons
Rock Creek - Upper Potomac E
15. $ 9,600,000 21
3 2,900,000 4
0.6 900.000 3
4 3,600.000 8
104 21.500,000 168
2 2,200,000 4
8.3 5,600,000 1 1
158 27,700,000 214
Anacostia River Basin
175 30,500.000 230
370 57,000,000 490
0.4 700.000 0.8
20 1 2,000,000 28
9 6,600,000 13
7 5.400,000 10
$186,000,000
27.000,000
10,000.000
$226,000,000
14,000,000
34,000,000
$274,000,000
ITS • 24-Hour
__Cast
lasln
$ 12.200.000
3,600,000
2,900,000
6.000,000
30,900.000
3.600.000
7,600,000
35,500,000
38,300.000
73.000.000
1,100.000
15,000,000
8,600,000
7,200,000
$246,000,000
27,000,000
16.000.000
$289,000.000
17,000,000
43,000.000
$349,000.000
25 Years - 24-Hour
Volume Cost
million
gallons
24 $ 13,800.000
4 3,600,000
4 3.600.000
11 7,600,000
190 33.300.000
4 3,600.000
14 9.100,000
260 41.700,000
270 43,900.000
560 82,000.000
1 1,300,000
32 17,000,000
IS 9,500,000
12 8,000,000
$278.000,000
27.000.000
18.000,000
$323,000,000
19,000,000
48,000,000
$380,000.000
'Basic layout illustrated in Figures 19 and 24.
'' ValU" ""' '"" '° "
re"e" "V «.«W»W««O.WO. to reflec. the costs o, dealing with those sewer district, scheduled for
88
-------
Table 13
Treatment at Overflow Points1
Sewer District
A-10, A-11
a. Storage
b. Treatment
A-12
a. Storage
b. Treatment
B-6, B-7
a. Storage
b. Treatment
C-25
a. Storage
b. Treatment
0-4, 0-5
a. Storage
b. Treatment
0-6, 0-7
a. Storage
b. Treatment
0-8, 0-9, 0-10,0-11,0-12
a. Storage
b. Treatment
E-2
a. Storage
b. Treatment
2 Years.
$ 7,600,000
600,000
3,200,000
170,000
...
...
1 ,300,000
100,000
15,200,000
2.100,000
2,000,000
150,000
3,900,000
300,000
18.300,000
2,900,000
24-Hour
$ 8,200,000
2,370,000
1 ,400,000
17,300,000
2,150,000
4,200,000
21,200,000
Estimated Capital Costs
(ENR Index = 18001
Storm Frequency and Duration
SYeai
Rock Creek •
$ 9,600,000
780,000
2,900,000
220,000
900,000
100,000
3,600,000
280,000
21,500,000
3.000,000
2,200,000
170.000
5,600,000
460,000
27,700,000
4,000,000
-s, 24-Hour
Upper Potomac Basin
$ 10,380,000
3,120,000
1,000.000
3,880,000
24,500,000
2,370,000
6,060,000
31,700,000
15Yei
$12,200.000
960,000
3,600,000
280,000
2,900,000
220,000
6,000,000
400,000
30,900,000
4,100,000
3,600,000
280,000
7,600,000
570,000
35,500,000
4,900,000
srs, 24-Hour
$ 13,160.000
3,880,000
3,120,000
6,400,000
35,000,000
3,880,000
8,170,000
40,400.000
25 Years,
$13,800,000
1,100,000
3,600,000
280,000
3,600,000
280,000
7,600,000
560,000
33,300,000
4,400,000
3,600,000
280,000
9,100,000
690,000
41,700,000
5,600,000
24-Hour
$ 14,900,000
3,880,000
3,880,000
8,160,000
37.700,000
3,880,000
9,790,000
47,300,000
Anacostia River Basin
E-3, F-1
a. Storage
b. Treatment
G-2
a. Storage
b. Treatment
G3
a. Storage
b. Treatment
G-4
a. Storage
b. Treatment
G-5
a. Storage
b. Treatment
G-7
a. Storage
b. Treatment
Storage and Treatment
Construction Cost
Northeast Boundary
Relief/Storage Tunnel
Total Construction Cost
Engineering Design at ± 6%
Other costs (resident
inspection, legal, bonding
and administrative) at ± 15%
Total Project Costs2
23,300,000
3,400,000
a
45,000,000
5,800,000
400,000
100,000
9,600,000
730,000
4.800,000
370,000
4,300,000
360,000
26,700,000
50,800,000
500,000
10,330,000
5,170,000
4,660,000
$155,000,000
27.000.000
$182,000,000
11,000,000
27.000,000
$220,000,000
30,500,000
4,200,000
57,000,000
7,300,000
700,000
100,000
12,000,000
900,000
6,600,000
540,000
5,400,000
440,000
34,700,000
64,300,000
800,000
12,900,000
7,140,000
5840QOO
$209,000,000
27.000.000
$236,000,000
14,000,000
35,000,000
$285,000.000
38,300,000
5,200,000
73,000,000
8,400,000
1,100,000
100,000
15,000.000
1,300,000
8,600,000
680,000
7,200,000
580,000
43,500,000
81,400,000
1,200.000
16,300,000
9.280,000
7,780,000
$273,000,000
27.000.000
$300,000,000
18,000,000
*t5,ooc.ooo
4363,000,000
43,900.000
5,800,000
82,000,000
9,500,000
1,300,000
100,000
17,000,000
1,500,000
9,500,000
750,000
8,000,000
650,000
49,700,000
91,500,000
1,400,000
18,500,000
10,250,000
8,650,000
$309,000,000
27,000,000
$336,000,000
19,000,000
50.000,000
$405,000,000
^ Basic layout illustrated in Figures 20 and 25.
2ln the final analysis, those values will have to be increased by $6,000,000-$! 2,000,000, to reflect the costs of
dealing with those districts scheduled for separation before 1975.
89
-------
Table 14
Conveyance Tunnels and Mined Storage'
Overflow Volume, million
gallons
Rock Creek
Upper Potomac
Anacostia River
Costs
1. Tunnel Construction
a. Rock Creek
b. Upper Potomac
2. Vertical Shafts
a. Rock Creek
b. Upper Potomac
3. Mined Storage
4. Pumping Station
Costs
1. Tunnel Construction
2. Vertical Shafts
3. Mined Storage
4. Pumping Station
5. Northeast Boundary
Relief/Storage Tunnel
Sub-Total
Centralized Treatment Plant
Total Construction Costs
Engineering Design at 6%
Other Costs (resident
inspection, legal, bonding
and administrative) at ± 15%
Total Project Costs
Estimated Capital Costs
(ENR Index = 1800)
Storm Frequency and Duration
2-Year, 24-Hour
169
13
434
Rock Creek
$ 14,900,000
1,800,000
4,400,000
2,100,000
26,500,000
2,600,000
5-Year, 24-Hour
275
18
581
• Upper Potomac Area
$ 16,800,000
1 ,900,000
4,400,000
2,100,000
40,000,000
4,300,000
15-Year, 24-Hour
404
24
772
$ 18,200,000
2,000,000
4,400,000
2,100,000
57,000,000
6,200,000
25-Year, 24-Hour
487
29
890
19,000,000
2,100,000
4,400,000
2,100,000
67,000,000
7,600,000
Anacostia River Area
9,400,000
5,800,000
58,000,000
6,100,000
27,000,000
$159,000,000
10,000,000
$169,000,000
10,000,000
25,000,000
$204,000,000
10,400,000
5,800,000
74,000,000
8,300,000
27,000,000
$195,000,000
13,000,000
$208,000,000
12,000,000
31,000,000
$251,000,000
11,100,000
5,800,000
94,000,000
10,900,000
27,000,000
$239,000,000
16,000,000
$255,000,000
1 5,000,000
38,000,000
$308,000,000
11.500,000
5,800,000
108,000,000
12,600,000
27 000 000
$267,000,000
18000000
$285,000,000
17,000,000
43,000,000
$345,000,000
Basic layout illustrated in Figures 16 and 21,
*ln the final analysis, these values will have to be increased by $6,000,000-$ 12,000,000, to reflect the cost of dealing with those sewer districts
scheduled for separation before 1975.
90
-------
Table 15
Sewer Separation Costs
(Based on the 15-Year Storm)
Estimated Capital Costs
(ENR Index = 1800)
Sewer District Separation Cost
A-10 $ 900,000
A'11 9,400,000
A-12 700,000
B-3 500,000
B-4 4,300,000
B-5 600,000
B-6 1,500,000
B-7 300,000
C-23 200,000
C-24 500,000
C-25 (Small Part) 4,100,000
C-25 (Large Part) 20,500,000
C-26 600,000
C-28 1,500,000
C-29 6,900,000
D-4 97,000,000
D-5 1,400,000
D-6 1,200,000
D-7 1,000,000
D-8 800,000
D-9 1,000,000
D-10 5,600,000
D-11 2,500,000
D-12 1,900,000
D-13 1,500,000
D-14 2,300,000
D-15 2,200,000
D-16 35,000,000
D-17 300,000
E-2 28,000,000
E-3 18,000,000
F-1 49,000,000
G-2 177,000,000
G-3 1,300,000
G-4 13,000,000
G-5 7,700,000
G-7 6,600,000
Total Construction Costs $507,000,000^
Other Costs (resident inspection, legal, bonding,
and administrative) at + 15% 76,000,000
Total Projects Cost $583,000,0002
1 Includes Engineering Costs
2ln the final analysis, this value will have to be increased by $27,000,000, to
reflect the costs of separating those sewer districts scheduled for separation
before 1975.
91
-------
The construction costs for sewer separation have been estimated from cost information
in the 1957 Board of Engineers Report (1). The estimated costs, updated to an ENR
index of 1800, for separating those sewer districts scheduled after 1975 is $583 million.
Table 16 and Figure 30 summarize the total construction costs of each District-wide
approach. The costs of sewer separation is considerably higher than the costs of the three
other approaches, which are about the same. The alternative of total use of tunnels and
mined storage represents about a 10 percent cost savings over the next lowest approach.
Table 16
Preliminary Estimate
Comparison of Total Project Costs
Design Storm Frequency
Districtwide Alternative 2-Year 5-Year 15-Year 25-Year
Maximum Utilization of1 $212,000,000 $274,000,000 $349,000,000 $380,000,000
Storage Reservoirs
Treatment at Points of1 220,000,000 285,000,000 363,000,000 405,000,000
Overflow
Total Use of Conveyance1 204,000,000 251,000,000 308,000,000 345,000,000
Tunnels and Mined Storage
Sewer Separation2 583,000,000
11n the final analysis, these values will have to be increased by $6,000,000-$ 12,000,000, to reflect the
cost of dealing with those sewer districts scheduled for separation before 1975.
^In the final analysis, this value will have to be increased by $27,000,000, to reflect the cost of separat-
ing those sewer districts scheduled for separation before 1975.
Operating Costs
Preliminary estimates of annual operating costs have been prepared for each of the various
alternatives and are presented in Tables 17 through 20. These estimates are prepared for
the 15-year, 24-hour storm only. The annual operating costs for various alternatives are
similar except for a significant difference in maintenance costs. The maintenance costs
for the reservoir approach and the local treatment approach are higher because there are
fourteen storage or storage-treatment facilities to be maintained, as compared to only two
central storage facilities for the tunnel approach. The treatment plant cost for the local
treatment approach is higher because this approach requires many local plants rather than
a single centralized facility.
The handling and disposal of sludge represents a significant cost for all the approaches
except sewer separation. As listed in Table 10, combined sewer overflows will contain
approximately 30,000 tons of dry suspended solids per year. In any system utilizing storage
or in-line treatment, most of these solids will be removed, and consequently will require
handling and disposal.
92
-------
FIGURE 30
INVESTMENT VERSUS DESIGN STORM RETURN FREQUENCY
FOR ALTERNATIVE DISTRICTWIDE APPROACHES
CD
CO
CJ
o
450,000
400,QOO
e/o
I—
00
a
CJ
350,000
_, 300,000
250,000
200,000
2-YR.
5-YR. 10-YR. 15-YR.
DESIGN STORM FREQUENCY (24-HR DURATION)
TREATMENT AT POINTS
OF OVERFLOW
MAXIMUM USE OF
STORAGE RESERVOIRS
TUNNELS AND MINED
STORAGE
25-YR.
-------
Table 17
Preliminary Estimate
Summary of Annual Operating Costs
for
Maximum Use of Storage Reservoirs
for 15-Year, 24-Hour Storm
Administration $ 290,000
Labor
Permanent Staff $ 200,0001
Auxiliary Staff 160,0001 360,000
Maintenance
Structures $ 370,000
Mechanical 2,000,000
Electrical &
Instrumentation 730,000 3,100,000
Utilities
Electrical 20,520 x 0.746 x 952 x $0.010/KWH 150,000
Treatment Plant (Additional Facilities to treat
storm water flows at Blue Plains Plant or
other location)
Sludge Handling and Disposal 750,000
Operation and Maintenance - 4% of Capital Cost 640,000
Sub-Total $5,290,000
Operating Contingency at 5 Percent 270,000
TOTAL $5,560,000
1
Includes overhead and benefits
94
-------
Table 18
Preliminary Estimate
Summary of Annual Operating Costs
for
Treatment at Overflow Points
for 15-Year, 24-Hour Storm
Administration $ 300,000
Labor
Permanent Staff $ 170,0001
Auxiliary Staff 200,0001 370,000
Maintenance
Structures $ 370,000
Mechanical 2,100,000
Electrical &
Instrumentation 750,000 3,220,000
Utilities
Electrical - 20,520 x 0.746 x 952 x $0.010/KWH 150,000
Filter System (per 1,000 gals.)
Filter Replacement and
Installation Cost $0.08
Operation arid Maintenance 0.10
Power and Backwash 0.02
975,000 x $0.20 = 195,000 200,000
Treatment Plants
Sludge Handling and Disposal 750,000
Operation and Maintenance - 4% of Capital Cost 1,500,000
Sub-Total $6,490,000
Operating Contingency at 5 Percent 330,000
TOTAL $6,820,000
11ncludes overhead and benefits.
95
-------
Table 19
Preliminary Estimate
Summary of Annual Operating Costs
for
Conveyance Tunnels and Mined Storage
for 15-Year, 24-Hour Storm
Administration $ 250,000
Labor
Permanent Staff $ 140,000]
Auxiliary BO^OO1 200,000
Maintenance
Structures $400,000
Access Shafts 60,000
Mechanical 450,000
Electrical &
Instrumentation 370,000 1,280,000
Utilities
Electrical - 23,850 x 0.746 x 952 x $0.010/KWH 170,000
Treatment Plant (Additional Facilities to treat
storm water flows at Blue Plains Plant or
other location)
Sludge Handling and Disposal 750,000
Operation and Maintenance - 4% of Capital Cost 680,000
Sub-Total $3,330,000
Operating Contingency at 5 percent 170,000
Total $3,500,000
'Includes overhead and benefits.
96
-------
Table 20
CO
Preliminary Estimate
Comparison of Annual Operating Costs
Based on 15-Years, 24-Hour Storm
Administration
Labor
Maintenance
Utilities
Sludge Handling and Disposal
Treatment 0 & M Costs
Sub-Total
Operating Contingency at 5 Percent
TOTAL
Maximum Use of
Storage Reservoirs
$ 290,000
360,000
3,100,000
150,000
750,000
640,000
$5,290,000
270,000
Treatment at
Overflow
Points
$ 300,000
370,000
3,220,000
150,000
750,000
1,500,000
$6,490,000
330,000
Conveyance Tunnels
and
Mined Storage
$ 250,000
200,000
1,280,000
1 70,000
750,000
680,000
$3,330,000
170,000
$5,560,000
$6,820,000
$3,500,000
-------
This quantity is greater than the quantity of solids generated at the entire Blue Plains
plant (26); therefore, the quantity of sludge from combined sewer overflows probably
could not be handled entirely at the Blue Plains plant. However, in view of the general
problems of solid waste disposal in the Washington area, perhaps the same order of
sophistication will be required to handle the combined sewer sludge as is employed to
handle the sewage treatment plant sludge. To handle and dispose of one ton of dry
suspended solids at the Blue Plains plant costs about $25.00 (26). This suggests that annual
costs of sludge disposal for a combined sewer pollution abatement approach (other than
separation) will be $750,000.
98
-------
Impact on Water Quality
Indicated Requirements
As previously presented in this report, the Water Quality Office of the Environmental
Protection Agency has determined the maximum allowable contaminant loadings to the
Potomac if water quality objectives are to be met. These recommended loadings are quite
stringent, requiring, at all treatment plants, the removal of at least 96 percent of 5-day
BOD, 96 percent of phosphorus, and 85 percent of nitrogen. These removals are based
on existing flows to these plants; as population increases, even greater degrees of treatment
will be required.
These reductions assume no other source of pollutants to the Potomac and no allowance
for combined sewer overflow loadings. Even if these reductions are achieved at the
treatment plants, the combined sewer overflows (occuring fifty to sixty times a year)
will in themselves result in loadings greater than recommended. At times, these overflow
loadings will be from ten to thirty times greater than the recommended loadings. It is
apparent that some measures must be taken to reduce the discharge of untreated combined
sewer overflow, or the water quality objectives will almost certainly fail to be achieved.
These pollutant reduction percentages should not apply directly to combined sewer
overflows. A statistical analysis (6) indicates that a system of storage and treatment facilities
sized for the overflow from the one-year, 24-hour storm will retain over 99 percent of
the long-term-averaged annual volume of overflow. Nevertheless, infrequent storms will
at times have an overflow volume far in excess of this storage capacity. This will result
in the discharge of a significantly high shock loading to the Potomac, enough to interfere
with many uses of the river, possibly enough to result in the mass destruction of a desirable
fish population which may have reappeared with the cleansing of the Potomac. The primary
decision focuses not on selecting a reduction percentage but on selecting the storm event
and degree of control for which storage and treatment facilities will be designed.
Reduction in Pollutant Loadings
The selection of the design frequency is properly made only in the context of basin-wide
water quality management. The Potomac pollution problems to which combined sewers
contribute are interstate in scope, and require evaluation on a regional basis. Many factors
beyond the scope of this study, e.g. flow augmentation, financing, population growth,
have too great an influence to be ignored. To provide part of the information basis upon
which this decision is made, this section of the report indicates the reduction in pollutant
loadings expected from storage and treatment facilities based on various design storm
frequencies.
Although sewer separation will stop the discharge of untreated sanitary sewage into the
Potomac, a significant loading will still occur with each rainfall in the form of storm
water runoff. Table 10 shows that, on an annual basis, storm water runoff from the
combined sewer districts accounts for more organic loading than the sanitary sewage in
the overflow. This effect, in combination with the extremely higher costs, the long
implementation period, and the public inconveniences associated with sewer separation,
tends to disqualify the sewer-separation approach as a feasible District-wide solution.
Therefore, the remaining discussion of impact on water quality focuses on the other three
approaches.
99
-------
A two-year frequency is suggested as the minimum design frequency. While a two-year
design frequency may seem small, it will retain and treat over 99 percent of the expected
annual volume of overflow on a long-term basis. Two reasons explain this seemingly high
value:
1. Practically all of the fifty to sixty overflows occurring yearly have volumes less
than the volume of the overflow from the 2-year, 24-hour storm.
2. The volume of overflow does not differ greatly among significant storms, e.g.
a facility designed to retain the overflow from the 2-year, 24-hour storm could
retain 70 percent of the overflow from the 5-year, 24-hour storm.
Regardless of the selected design-storm frequency, there exists the possibility, however
remote, of occurrence of a storm which would result in an overflow in excess of the
design storage and treatment capacities. Nevertheless, implementation of a program based
on a design storm whose recurrence interval is less than 25 years may represent an
injudicious use of public funds. On the other hand, the very high storm recurrence intervals
(50 years and more) are generally associated only with projects involving immediate risk
of human life and/or catastrophic property damage.
The abatement of pollution from combined sewers requires storage facilities. Since any
stored flow is to be treated, the quantity of retained overflow represents a practical
measurement of the reduction in the pollution load. Figure 31 illustrates the reduction
in BOD, SS, PO/j, or N corresponding to various retention capacities. It was assumed
the percentage reduction in any pollution load is equal to the percentage of the total
overflow collected, multiplied by the percentage removal efficiency of the subsequent
treatment method (assumed to be 85 percent). Examination of Figure 31 shows that
facilities designed to handle a 2-year, 24-hour storm (and provide 85 percent treatment)
will provide 60 percent treatment for a 5-year, 24-hour storm but only 38 percent for
a 25-year, 24-hour storm. Table 21 summarizes some of the information illustrated in
Figure 31.
Figure 31 and Table 21 illustrate the necessity of evaluating combined sewer overflows
in terms of their true impact as shock loadings. For example, effective storage capacity
equal to the overflow from the two-year, 24-hour storm probably will reduce the expected
annual loadings by just under 85 percent (the maximum possible). Nevertheless, with a
storage capacity equivalent to the two-year, 24-hour storm it is expected that once in
five years a loading equal to or exceeding 40 percent of the total loading in the five-year,
24-hour storm will be bypassed to the river. In terms of BOD, this represents a loading
of 80,000 pounds, almost five times the maximum allowable daily loading in the entire
metropolitan Washington area. Again, it is pointed out that this shock form of loading
in combination with the long effective residence times of estuarine waters explains the
serious impact of combined sewer overflows in the Washington areal
In making the final analysis of alternative strategies. Figure 31 can be used with the
information in Tables 5 and 10 (in the Problem Definition section) to determine the
pollutant loadings from any sewer district. While this procedure requires some simplifying
assumptions, it will provide a sufficiently accurate estimate of pollutant loadings for
comparison with other factors such as costs. Figure 32 represents the relationship between
investment and the discharge of BOD for three intense-storm return frequencies. The
100
-------
FIGURE 31
PERCENT REDUCTION IN POLLUTION LOAD FROM VARIOUS STORMS
VS
ALTERNATIVE DESIGN EVENTS
100
80
60
40
20
MAXIMUM REDUCTION
= 85%
V
\
X^
\ /
» » s
V*S? ^
» cT>
^i^p
*^'
»
V
v^
\'*
""•f '^!
m* ^
X
»
^ »
\\
X
25 15 5
2
YEARS YEARS YEARS YEARS
DESIGN STORM EVENT
101
31
-------
Table 21
Comparison of 24-Hour Pollution Discharges from Various Return Frequency Storms
for Various Design Frequencies
Design Storm
o
N)
Storm
2-Year, 24-Hour Storm
BOD, Ibs.
TP, Ibs.
5-Year, 24-Hour Storm
BOD, Ibs.
TP, Ibs.
15-Year, 24-Hour Storm
BOD, Ibs.
TP, Ibs.
25-Year, 24-Hour Storm
BOD, Ibs.
TP, Ibs.
No
Treatment
1 60,000
1 2,000
200,000
13,000
250,000
14,000
280,000
15,000
2-Year,
24-Hour
24,000
1,800
80,000
5,000
140,000
8,000
1 70,000
9,000
5-Year,
24-Hour
24,000
1,800
30,000
2,700
93,000
5,000
130,000
7,200
15-Year,
24-Hour
24,000
1,800
30,000
2,700
38,000
2,100
76,000
4,000
25-Year,
24-Hour
24,000
1,800
30,000
2,700
38,000
2,100
42,000
2,300
-------
FIGURE 32
EFFECT OF INVESTMENT ON DISCHARGE OF
BOD LOADING FOR VARIOUS STORM RETURN FREQUENCIES
350
300
o
CO
250
CO
t—
GO
200
150
20,000
40,000
60,000
80,000
100,000
120,000
140,000
160,000
180,01
24-HR. BOD LOADING, POUNDS
CJ
-------
construction costs correspond to the approach which employs conveyance tunnels and
mined storage District-wide. Note that each even-increment increase in investment results
in a far greater proportionate decrease in BOD loading from intense storms. For example,
an increase of investment from $250,000,000 to $300,000,000 results in a decrease of
BOD loading, due to the 25-year, 24-hour storm from about 130,000 pounds to about
84,000 pounds.
Besides examining the costs and possible pollutant loadings when determining the design
frequency, consideration must be given to the probability of an intense storm occurring
during critical stream conditions. The recommended maximum pollutant loadings for the
metropolitan Washington area were determined for summer conditions, when fresh water
flow is low (less dispersion of waste), temperature is high (higher temperatures increase
biological activity and reduce the dissolved oxygen saturation concentration), and plant
respiration overnight has somewhat depleted the available dissolved oxygen. Water quality
conditions usually are not as critical during other seasons of the year. A graph in a U. S.
Weather Bureau Technical Paper (27) is reproduced as Figure 33 to provide insight as
to the seasonal probability of intense storms. Note that in the months of July, August,
and September there is a 15 percent probability of obtaining a rainfall equal to or exceeding
that corresponding to a 5-year return period, while the probability of a similar occurrence
during the remaining nine months is only 5 percent.
The preceding discussion has focused on the removal of those contaminants that deplete
dissolved oxygen levels. However, the overflow from combined sewers also contributes
to high bacteria levels and repulsive floating matter in the Potomac and its tributaries.
While only 85 percent BOD removal can probably be effected from any retained overflow,
total removal of floating matter and almost total disinfection is possible. Therefore, facilities
based on a 15-year design frequency would rid the Potomac of high bacteria levels and
repulsive floating matter from combined sewers for 14 years out of a 15-year period,
and for practically all of the recreational season in the fifteenth year.
104
-------
FIGURE 33
SEASONAL PROBABILITY OF INTENSE 24-HOUR RAINFALL
MIDDLE ATLANTIC REGION
PROBABILITY IN PERCENT OF OBTAINING
A RAINFALL IN ANY MONTH OF A PAR-
TICULAR YEAR EQUAL TO OR EXCEEDING
THE RETURN PERIOD VALUES.
J F M A M ) J A S 0 N D
MONTH OF YEAR
SOURCE: U.S. WEATHER BUREAU TECHNICAL PAPER NO.40
105
33
-------
Comparison of Non-Quantifiable Factors
To determine which of the feasible alternative approaches would offer the best opportunity
for abatement of the combined sewer overflow problem requires careful consideration of
many factors. Cost comparison obviously is essential, but there are other factors which
also have a significant bearing on the selection of the most favorable plan. The most
important of these other factors are Reliability, Flexibility, Land Requirements, Public
Convenience, Implementation, and Solids Deposition and Gas Production. In general, the
effects of these factors on the various alternatives are either non-quantifiable or are difficult
to quantify. The relative impact on each alternative is discussed in the following paragraphs.
Reliability
An important consideration is the reliability of any system to perform its designed function
whenever a storm occurs; there is no time to get things ready after the onset of a storm
In regard to collection of the storm waters, all four approaches are equally (and strongly)
reliable, because all rely primarily on gravity for capture and collection.
However, efficient capture and collection presupposes sufficient storage capacity, which
means that the wastewaters collected from previous storms must have been moved out
of the storage spaces to the treatment facilities or to the streams. This obviously is
dependent on the readiness and capacity of the pumping facilities (except for sewer
separation). In this regard, there is a difference, however small, between the various
alternatives. Sewer separation would be the most reliable overall, because it would involve
no dependence on pumping equipment. Conveyance tunnels and mined storage would be
the next most reliable. Its requirement for two large pumping stations in comparison with
many smaller pumping stations for the reservoir and local treatment approaches constitutes
a reliability advantage, because of the inherent higher reliability of a larger installation
and the relative ease of routine preventive maintenance.
In addition to the same number of pumps required for the reservoir approach, treatment
at overflow points would involve equipment at treatment facilities, and this equipment
would also have to be constantly maintained and operated to assure reliability. This
dependence on additional equipment makes this approach not quite as realizable as the
others. Nevertheless, no severe problems are anticipated with regard to reliability.
Flexibility
In addition to random occurrence of storms and the variations in storm duration and
intensity for the Washington, D. C. area as a whole, there will be variations within the
area. These local variations could be significant.
Each storage reservoir (in the storage reservoir approach) would serve a discrete sub-drainage
area, and the probability of occurrence of a storm of such duration and intensity as to
cause overflow is greater for any of the smaller discrete areas than for the overall area.
Treatment at overflow points has reservoirs based on the same discrete drainage areas
and is subject to the same degree of inflexibility as the reservoir approach, and the same
is true for sewer separation.
A network of tunnels and mined storage would provide the greatest flexibility. Variations
in storm occurrence, intensity, and duration within the Washington D. C. area would not
106
-------
per se consititute a problem, because flow from all sub-drainage districts would be dropped
to the mined-storage chambers and equalized over the whole study area.
Land Requirements
Site investigations have been made to determine whether land is available at outfall points
for construction of facilities required under the various approaches. In general, there is
vacant land available at each of the overflow points. However, it should be noted that
most of the potential sites are under the jurisdiction of the National Park Service, and
their full cooperation would be necessary to obtain the sites.
Since construction at storage reservoirs would be underground, the only land required
would be for access and maintenance. The ground above the reservoirs could be used
for parks and playgrounds, with only relatively small pump houses extending above grade.
Architecturally, the pump houses could be built to preserve the national park landscape.
The land required would be greater for the approach proposing treatment at overflow
points. The additional housing for the pressure filters and appurtenances would require
construction of buildings at each overflow point. Again, the structures could be built to
preserve the aesthetics of the land under the jurisdiction of the National Park Service.
Land required for tunnel and mined-storage construction would be minimal. Only the
land necessary for temporary occupancy during shaft construction would be required. Upon
completion of construction, the work area would be landscaped to restore the original
condition.
Sewer construction under a separation program would be in the City streets and
rights-of-way. Therefore, for purposes of this report, no land is required for this approach.
Public Inconvenience
The disruption of public travel in Washington during the construction period is a major
factor in the selection of an alternative approach. In fact, this may be the most important
factor, possibly carrying more weight than costs or water quality. The present construction
of the Washington Metropolitan Transit Authority subway illustrates the severe impact
of a construction program that requires the closure of streets to traffic and the erection
of barricades to pedesterians. Many businesses have had to close for indefinite periods,
since customers refuse to frequent establishments on the disrupted streets. Traffic problems
worsen as fewer avenues of travel and fewer parking spaces remain available.
Reservoir construction would be off the public streets. Traffic disruptions would be caused
only by the construction vehicles required to move equipment and materials from the
reservoir site.
For the local treatment approach, traffic disruption again would be only for movement
of construction equipment and materials. However, due to additional equipment and more
complex construction, the construction period and attendant traffic disruption would be
for a longer period of time.
107
-------
The tunnel and mined-storage approach would require virtually no disruption or
inconvenience to residents and visitors. All construction would be underground and off
the City streets. The only movement of vehicles would be to remove drilling muck. This
operation could be done during the off-peak traffic hours; however, if moling progresses
at a fast rate, continuous muck removal may be required.
Sewer separation would be the most disruptive to public travel. Virtually every street
in the study area would be excavated at some point in time. The ramifications of such
disruption would be difficult to assess completely, but it is certain that the disruption
would be much more serious than with any of the other approaches. Besides this, sewer
separation would require considerable work on private property to separate roof drains
and other plumbing.
Implementation
Each of the four approaches can be implemented, but some will be easier and/or quicker
than others. Considerations affecting the relative ease of implementation are:
1. Public attitudes and reaction.
2. Cooperation of other District and Federal Agencies.
3. Political implications and restraints.
4. Staging of construction and possible benefits with phased construction.
5. Financial schedules and monetary assistance programs.
Construction schedule, i.e. time required to complete each of the alternatives, is one of
the critical elements involved in evaluating the adaptability of each approach to the
establishment of stricter requirements, particularly in view of the strong pressures to
improve the environment at the earliest possible date.
Before final design of storage reservoirs could be started, additional studies would have
to be made to determine their exact location and the design parameters of each. The
necessity to design reservoirs unique to each drainage district complicates the design; and
the number of facilities required extends the construction period. It would take
approximately eight years to complete the facilities required if this approach is applied
District-wide where feasible.
Additional studies and individual design would also be necessary for treatment at overflow
points. Due to the necessity to design special equipment, and taking into account the
need to fabricate and install the special equipment, the time for completion would be
longer than for the storage reservoir approach. It is estimated that a District-wide approach
of treatment at overflow points could be completed in ten years.
The District-wide tunnel and mined-storage approach would take about eight years.
However, the boring techniques are sufficiently advanced to allow relatively rapid
construction.
Sewer separation would take the longest time, by far, to complete. On the basis of data
presented in the 1957 Board of Engineers Report, it is estimated that the sewer separation
program would be completed sometime after 2000.
108
-------
Solids Deposition and Gas Production
Another major consideration in the selection of an approach is the problem of solids
deposition and concomitant gas production. Solids deposition would have a three-fold
effect:
1. Deposition of solid material could reduce storage capacity.
i
2. Organic solids may decompose to form malodorous compounds thus causing a
nuisance condition.
3. Organic solids may decompose to form explosive organic gases.
Settleable solids deposited in the settling zone of each reservoir could be pumped back
to the existing system for ultimate disposal at Blue Plains. Solids carried over to the storage
zone would be pumped to Blue Plains through the raw sewage pumps. If the reservoirs
are emptied relatively soon after the storm subsides and regular maintenance by flushing
is provided, neither septicity nor gas production is expected to be a problem. To insure
that gas production is not a problem each of the reservoirs would be equipped with
adequate ventilating and air-pollution control devices.
The solids deposition and gas production would have about the same impact on the
approach using treatment at overflow points as for the reservoir approach. In addition
to solids removed in the settling zone of the storage structures, backwash from the fiberglass
filters would be discharged to the existing system and conveyed to Blue Plains.
The field studies presented in the first part of this report have determined that a high
concentration of solids occurs during the first few minutes of the storm Therefore, the
major portion of the solids will be carried into the tunnels (under this approach) during
the first few minutes. This flushing action will carry the solids to the low point and
facilitate solids removal. The mined storage would be compartmentized and the first few
compartments used as settling basins to facilitate removal of solids. However, the great
depths and vast extent of these tunnels may present problems in regard to effective solids
removal.
With a separation program, solids carried ..into the storm sewer would be discharged to
the receiving stream. As determined during the field studies, the amount of solids delivered
to the stream would be significant.
Evaluation of Factors Other Than Cost
The results of the evaluation of comparative reliability, flexibility, land requirements, and
other non-cost factors are summarized in Table 22. In view of the difficulty in properly
assessing the relative worth of each factor in quantitative terms, no quantifiable
measurement is employed in this presentation. Instead, a measurement scale of X's indicates
the degree of the problems possible with each approach. An overall evaluation indicates
a clear advantage for conveyance tunnels and mined storage over the other three approaches,
provided the solids question can be answered.
109
-------
Table 22
Evaluation of General Factors
Maximum Use of
Factor
Reliability
Flexibility
Land Requirement
Public Inconvenience
Implementation
Solids Deposition and
Gas Production
Storage
Reservoirs
X
X
XX
X
XX
Treatment at
Overflow Points
X
X
XX
X
XX
Conveyance Tunnels
and Mined Storage
X
—
—
—
X
Sewer
Separation
—
X
—
XXX
XXX
X
X
XXX
Key: — No Problem
X Possible Minor Problem
XX Possible Moderate Problem
XXX Possible Major Problem
-------
As previously pointed out in the comparison of the cost estimates presented in Tables
15 though 20, the conveyance tunnel-mined storage approach would also involve the lowest
construction cost of any of the four approaches and lower operating costs than either
storage reservoir or treatment at overflow points (Operating costs, which would be primarily
maintenance, were not calculated for sewer separation). Thus, the evaluation of general
factors reinforces the selection (on a cost basis) of conveyance tunnels and mined storage
as the approach to be adopted.
111
-------
Indicated Appropriate District-wide Solution
The preceding sections of this report illustrate how it is possible to develop within the
District numerous strategies for collecting, storing, and treating combined sewer overflow.
Each strategy offers a different level of costs, pollutant loadings, and benefits. While the
selection of the appropriate strategy must rest, in part, on factors beyond the scope of
this study, the information presented in this report indicates that the appropriate
District-wide solution may be one which centers on a single approach, the use of mined
storage and conveyance tunnels.
The recreational and other values of the Potomac River and its tributaries demand that
these waters be protected from pollution. Further study is needed to predict the effect
of shock contaminant loadings to the Potomac and its tributaries. Therefore, the selection
of the appropriate design frequency requires a somewhat arbitrary decision at this time.
It is common practice to use a 10-year minimum flow recurrence frequency as the design
basis for treatment facilities on fresh water streams. While the situation for estuarine waters
is different, this basis can be interpreted to mean that all combined sewer overflows from
storms-of-lower magnitude than the design storm must be entirely intercepted, stored, and
treated. Quite possibly, extending this to the 15-year design frequency represents a
reasonable and conservative decision in view of the small increase in costs (see Figure
30) and the significant decrease in pollution loadings from the intense storms (see Figure
32). Selection of the 15-year, 24-hour design event is further supported by consideration
of the water-contact activities proposed for the Potomac River and its tributaries. The
indications are that facilities based on this design event would permit all the desired
recreational activities with practically no interruption.
Table 10 in this report shows that the expected annual volume of combined sewer overflow
from the District is 12 billion gallons. The Blue Plains plant can completely treat 289
mgd above the average dry-weather flow for 400 hours per year. This is equivalent to
an annual reserve capacity of 5 billion gallons. While the Blue Plains plant can partially
treat some of the remaining 7 billion gallons, it is obvious that treatment at some other
location is required. The conceptual engineering study (6) of the Kingman Lake project
has identified the feasibility of constructing a multiple-purpose treatment plant in that
area. Actually, as brought out in that study (see Appendix G), this location for a treatment
facility offers a wide number of benefits, probably more than any other location in the
District. The original concept of a 50-mgd plant in this area could probably be expanded
to a 150-mgd plant (the 100-mgd increase could be directed simply to filtration and
chlorination rather than to complete treatment). This would handle in about five days
the 772 million gallons of overflow from the Anacostia River area for the 15-year, 24-hour
storm. This single facility, in combination with the reserve hydraulic capacity of the Blue
Plains plant, would provide sufficient capacity to treat the annual volume of combined
sewer overflow from the entire District.
Concentrating treatment capacity at these two locations, rather than at many more
locations, affords another advantage. Even with complete retention of overflow from the
15-year, 24-hour storm, the assumed 85 percent treatment efficiency would still allow
a 24-hour discharge of about 38,000 pounds of BOD to the Potomac. This loading equals
twice the recommended maximum daily loading in the Metropolitan Washington area, and
perhaps this would deplete dissolved oxygen below the minimum acceptable levels.
However, the treatment provided at the Blue Plains advanced plant and at the Kingman
Lake reclamation plant would result in an overall treatment efficiency higher than 85
percent, and less than 38,000-pound loading would result.
112
-------
A network of conveyance tunnels and mined storage is suggested because no other approach
offers capacity for the overflow from the 15-year, 24-hour storm at a lower cost or at
a higher level of benefits. The total project costs of the indicated appropriate solution
is about $318,000,000, as shown in Table 23. Note that this cost includes the costs of
dealing with those sewer districts scheduled to be separated before 1975 whose separation
has been delayed indefinitely. A solution incorporating mined storage in the Anacostia
River area and with both storage reservoirs and mined storage in the Rock Creek-Upper
Potomac area (see Figure 24) comes somewhat close to this figure, with a cost of
$325,000,000. By inspection of Tables 12 and 15, it is possible to reduce this figure
by $4,800,000 if the method of separation rather than storage reservoirs is applied to
Sewer Districts A-10, A-11, and A-12.
Actually, a cost less than $318,000,000 is possible with a solution designed for the 2-year,
24-hour storm in one area and the 20-year, 24-hour storm in another. This represents
both overdesigning and underdesigning, and produces the same reduction in pollutant
loadings from the 25-year storm as a uniformly applied 15-year design frequency.
Economies of scale explain this cost difference. However, this scheme would allow a higher
loading from the 5-year storm than would a uniformly applied 15-year design frequency.
Although the selection of the appropriate solution must consider some factors beyond
the scope of this study, the information developed herein does point to a particular
solution. In summary, this solution is one that provides a network of tunnels and mined
storage, with treatment at the Blue Plains plant and at a facility near Kingman Lake.
A 15-year, 24-hour storm event (or one based on a longer recurrence interval) is suggested
at the design basis, in the public's best interest.
113
-------
Table 23
Estimated Capital Costs of Indicated Appropriate Solution
(ENR lndex = 1800)
Feature Costs
Rock Creek Upper Potomac
1. Tunnel Construction
a. Rock Creek $ 18,200,000
b. Upper Potomac 2,000,000
2. Vertical Shafts
a. Rock Creek 4,400,000
b. Upper Potomac 2,100,000
3. Mined Storage 57,000,000
4. Pumping Station 6,200,000
Anacostia River Area
1. Tunnel Construction 11,100,000
2. Vertical Shafts 5,800,000
3. Mined Storage 94,000,000
4. Pumping Station 10,900,000
5. Northeast Boundary Re lief/Storage Tunnel 27,000,000
6. 150-MGD Treatment Plant at Kingman Lake"1 12,000,000
Sewer Districts Scheduled for Separation prior to 1975 10,000,000
Total Construction Costs 261,000,000
Engineering Design at 6% 16,000,000
Other Costs (resident inspection, legal, bonding, and
administrative at 15%) 41,000,000
Total Project Costs $318,000,000
'Costs based on 150-MGD plant incorporating processes depicted in Figure 15
(filtration and chlorination) rather than complete water reclamation processes.
114
-------
SECTION VIM
ACKNOWLEDGMENTS
Roy F. Weston, Inc. gratefully acknowledges Mr. Paul Freese and Mr. George A. Moorehead,
of the Department of Sanitary Engineering of the District of Columbia, for general
assistance and particularly for their efforts in connection with the construction of field
monitoring stations and installation of equipment.
Acknowledgment is made to Mr. John F. Miller, Special Studies Branch, United States
Weather Bureau, for furnishing local climatological data and for providing assistance in
interpretation of those data.
Acknowledgment is made to Mr. Al Cywin, Director of the Division of Applied Science
and Technology, Mr. William Rosenkranz, Chief of the Storm and Combined Sewer Branch,
Dr. Henry R. Thacker, Project Officer, and Mr. Darwin R. Wright, Project Manager, all
of the WATER QUALITY OFFICE, for their interest and guidance during the course
of this project.
Acknowledgment is also made to the members of the staff of Roy F. Weston, Inc. who
have participated in this project.
Roy F. Weston, P.E., President
J. E. Germain, P.E., Vice President Concept Technology Division
V. T. Stack, P.E., Principal Consultant Consulting Division
J. A. DeFilippi, P.E., Project Manager
M. S. Neijna, P.E. Concept Technology Division
P. L. Buckingham, P.E.
E. F. Gilardi, Ph.D., P.E.
C. S. Shih, Ph.D., P.E.
U. C. Mankad, P.E.
J. A. Lee
C. J. Cahill
R. E. Coleman, P.E. Engineering Design Division
T. E. Taylor
J. G. Ryan
J. K. Kane Geology Section
115
-------
Special acknowledgment is made to the Technical Editing, Report Preparation, Duplicating,
and Report Graphics groups.
J. L. Simons
J. P. Jarosh
A. M. Tocci
N. L. McVey
J. L. Smith
D. L. Rimel
S. M. Jones
W. McDermott
J. W. Hitzelberger
J. C. Allison
D. 0. Thompson
D. W. Syphard
W. J. Costello
E. J. Burke
D. Angevine
R. E. Towson
116
-------
SECTION IX
REFERENCES
1. Board of Engineers (Greeley, S. A. et al), "Report to District of Columbia Department
of Sanitary Engineering on Improvements to Sewerage Systems," February 1957.
2. Metcalf & Eddy, Inc., Engineers, "Report Upon Investigation of Sewerage System,"
June 1955.
3. Metcalf & Eddy, Inc., Engineers, "A General Plan for Development of Water Pollution
Control 1969-1972," 1969.
4. U. S. Department of the Interior, Federal Water Pollution Control Administration,
"Potomac River Water Quality, Washington, D.C. 1969," 1969.
5. Tholin, A. B. and Keifer, C. J., "The Hydrology of Urban Runoff," Journal of the
Sanitary Engineering Division, ASCE, March 1959.
6. U. S. Department of the Interior, Federal Water Quality Administration, "Conceptual
Engineering Report, Kingman Lake Project," by ROY F. WESTON, Water Pollution
Control Research Series, 11023 FIX 08/70, U. S. Government Printing Office,
Washington, D.C., 1970.
7. "Problems of Combined Sewer Facilities and Overflows, 1957," American Public
Works Association for the U. S. Department of the Interior, FWPCA Publication
WP-20-11, December 1967.
8. Burm, R. J. et al, "Chemical and Physical Comparison of Combined and Separate
Sewer Discharges," JWPCF, January 1968, p. 112.
9. Wiebel, S. R. et al, "Characterization, Treatment, and Disposal of Urban Stormwater,"
Proceedings 3rd International Conference on Water Pollution Research, 1967, p. 327.
10. Evans, F. L. et al, "Treatment of Urban Storm-Water Runoff," Cincinnati Water
Research Laboratory Pre-Publication Copy, January 1967.
11. "Sewer Within a Sewer," Water Works/Wastes Engineering, February 1964, p. 36 (from
Selected Urban Stormwater Runoff Abstracts).
12. "Combination Sewer Separated into Sanitary and Storm Sewers at Low Cost," Civil
Engineering, December 1967, p. 61.
13. McPherson, M. B., "Progress Report-ASCE Combined Sewer Separation Project," Civil
Engineering, December 1967, p. 61.
14. Koelzer, V. A. et al, "The Chicagoland Deep Tunnel Project," The Metropolitan
Sanitary District of Greater Chicago, September 1968.
15. Cellini, W. F. et al, "Underflow Plan for Pollution and Flood Control in the Chicago
Metropolitan Area," Paper presented at the EPA-WQO Symposium on Storm and
Combined Sewer Overflows, Chicago, Illinois, June 22-23, 1970.
117
-------
16. Gregory, J. H. et al, "Intercepting Sewers and Storm Stand-By Tanks at Columbus,
Ohio," ASCE Proceedings, October 1933, p. 8.
17. Waller, D. H., "One City's Approach to the Problem of Combined Sewer Overflows,"
Water and Sewage Works, March 1967, p. 113.
18. Him, W. C., "Providing Primary Treatment for Storm Sewage Overflows," Wastes
Engineering, September 1962.
19. "Research and Development Programs," FWPCA, August 1968.
20. "Milwaukee Seeks to Solve Pollution Problems," Civil Engineering, September 1967,
p. 79.
21. Romer, H. and Klashman, L. M., "The Influence of Combined Sewers on Pollution
Control," Public Works, October 1961, p. 129.
22. Simpson, G. D., "Treatment of Combined Sewer Overflows and Surface Water at
Cleveland, Ohio," Technical Paper, 41st Conference, WPCF, September 1968.
23. Dunbar, D. D. and Henry, J. G., "Pollution Control Measures for Stormwaters and
Combined Sewer Overflows," JWPCF, January 1966, p. 9.
24. Gameson, A. L. H. and Davidson, R. N., "Storm Water Investigations at
Northampton," Institute Sewage Purification Journal, 1963 (from Selected Urban
Stormwater Runoff Abstracts).
25. Eliassen, R., "Coliform Aftergrowth in Chlorinated Storm Overflow," Journal of the
Sanitary Engineering Division, ASCE, April 1968, p. 371.
26. Board of Engineers (S. A. Greeley et al) "Planning Study of Sewage Treatment Plant
to the Year 2000," August 1964.
27. "Rainfall Frequency Atlas of the United States," U.S. Weather Bureau Technical Paper
No. 40, January 1963.
28. Kiefer, C. J. and Chu, H. H., "Synthetic Storm Pattern," Journal of Hydraulics
Division, ASCE, Vol. 83, August 1957.
29. Horner, W. W. and Jens, S. W., "Surface Runoff Determination from Rainfall Without
Using Coefficients," Transactions, ASCE, Vol. 107, 1942.
30. "Design and Construction of Sanitary Sewers," ASCE MOP No. 37, New York, 1960.
31. Allen and Taylor, "The Salt-Velocity Method of Water Measurement," Mechanical
Engineering, ASME, Vol. 46, 1938, p. 13-16.
32. Gnoat, B. F., "Chemi-hydrometry and Its Application of Testing of Hydroelectric
Generators," Transactions of ASCE, Vol. 80, 1915, p. 951-1305.
33. Anon., "Fluorometry in Studies of Pollution and Movement of Fluids," Fluorometry
Reviews Ace. No. 9941, February 1968.
118
-------
34. Whitney, W. H. and Wilson, G. G., "Effluent Distribution Studies on the Miramichi
and Restigouche Estuaries."
35. Fewerstein, D. L. and Selleck, R. E., "Fluorescent Tracers for Dispersion
Measurements," Journal of the Sanitary Engineering Division, ASCE, Vol. 89, No.
SA-4, Proc. Papers 3586, August 1963, p. 1-21.
36. Replagle et al, "Flow Measurements with Fluorescent Tracers," ASCE, Proc. 92 (Hy-5,
No. 4895), September 1966, p. 1-15.
37. Barnwell, E. L., "Multiple Tracers Establish Waterflood Flow Behavior," Oil and Gas
Journal, Vol. 64, N-28, 1966, p. 76-79.
38. Cederwall, K. and Hansen, J., "Tracer Studies on Dilution and Residence Time in
Receiving Waters," Water Research Pergamon Press, Vol. 2, 1968, p. 297-310.
39. Cederwall, K. and Hansen, J., "Dispersion Studies by Means of Tracers," (In Swedish),
Vag-Och Vattenbygganen, No. 1-2.
40. Cederwall, K. and Hansen, J., "Surface Water Studies by Means of Parallel Injected
Tracers," (In Swedish) Vag-Och Vattenbygganen, No. 10.
41. Kilpatrick, F. A., "Flow Calibration by Dye Dilution Measurements," Civil
Engineering, Vol. 28, 74-76F, 1968.
42. Agg, A. R., Mitchell, N. T., and Mitchell, G. E., "Use of Lithium as a Tracer for
Measuring Rates of Flow of Water and Sewage," Water Pollution Research Labs,
Stevenage, England, Ind. Sewage Purif. J. Proc., 1961, p. 240-5.
43. Spencer, E. A. et al, "Flow Measurement by Salt Dilution Method," /. Ind. Water
Eng., Vol. 14, 215-35, 1960.
44. "Selection and Handling a Radio Tracer for Study Sewage Distribution," Ind. Eng.
Chem., 50, 210-211, 1958.
45. Burgece, S. G. et al, "Determination of Flow Characteristics in Sewage Works Plants,"
Sewage Purif. J. Proc., P+3, 206-15, 1957.
46. Simpson, E. S. et al, "Radio Tracer Experiments in the Mohawk River, New York
to Study Sewage Path and Dilution," Geophysical Union, 39, 427-33, 1958.
47. Montens, A., "The Use of Radioactive Isotopes for Water Flow and Velocity
Measurements, Precautions and Possibilities of Use," Wasserfach, 93, 427-33, 1958.
48. Truesdale, G. A., "Measurement of Sewage Flow with Radio Active Tracers," Water
and Sanitary Eng., 4, 93-8, 1953.
49. Montens, A., "The Use of Radioactive Isotopes for Water Flow and Velocity
Measurements," Proc. 2nd Radio Isotope Conf., Oxford 2, Phys. Sciences and Ind.
App., 169-80, 1954.
50. Meek, R. L., Indianapolis Makes a Sewage Analysis, "Flow Velocities by Electrometric
Determination of CaCl2," Am. City, 43, (1), 96-8, 1948.
119
-------
SECTION X
APPENDICES
Appendix Page
A GEOLOGICAL AND OTHER NATURAL CONDITIONS 125
Introduction 125
Geologic Setting 125
Principal Unconsolidated Formations 126
Artificial Fills 126
Bedrock 127
Hydrology 129
Tunneling 130
Figure A-1: Generalized Geologic Map 134
Figure A-2: Contours of Bedrock Surface 135
B LIST OF PERTINENT STUDIES AND REPORTS REVIEWED 137
C INVESTIGATION OF RAINFALL RUNOFF RELATIONSHIPS. .. 139
Rainfall Data 139
Peak Rate of Overflow 139
Volume of Overflow 140
D MONITORING EQUIPMENT, INSTRUMENTS, AND PROCEDURES 145
Discussion of Facilities and Operating Procedures 145
Descriptions of Principal Equipment and Instruments 145
Lithium Chloride Flow Measurement 148
Monitoring Practices 150
Figure D-1: Monitoring Equipment Diagram 152
Figure D-2: Electric Power Supply and Triggering System
Diagram 153
Figure D-3: Sampling Components 154
Figure D-4: Rain Gauge 155
E MONITORED WASTEWATER FLOWS AND CHARACTERISTICS . 157
Comparison of Representative Storms 157
o Bacteriological Quality of Storm Wastewater 159
Waste Loadings from the Monitored Sewer Districts 160
Total Waste Load in Storm Water Runoff 161
Comparison of Flow Rate Measurements 163
Tables E-1
through E-3: Characteristics of Dry-Weather Flow for
Combined and Separated Sewer Districts 165
Table E-4: Characteristics of Storm Runoff in Sewer
District Good Hope Run 168
Table E-5: Waste Loadings in Combined Sewer Overflow
and Separated Storm Water Discharge 169
Table E-6: Total Waste Loadings Generated by
Different Storms 170
121
-------
Appendix
E
Table E-7:
Figures E-1
and E-2:
Figures E-3
through E-7:
Figures E-8
and E-9:
Figures E-10
through E-13:
Figure E-14:
Figures E-15
through E-22:
Figures E-23
and E-24:
Figures E-25
through E-27:
Figures E-28
and E-29:
Figure E-30:
Comparison of Actual Total Flow
Measurement and Volume Obtained by
Hydrograph Analysis
Characteristics of Dry-Weather Flow
in Combined and Separated Sewer
Districts
Characteristics of Combined and
Separated Sewer Flows for Storms of
Various Intensities and Durations . .
Characteristics of Sewer Flow during
Consecutive Storms in Combined Sewer
District G-4
Page
171
172
174
179
Quality and Quantity of Runoff for
Storms of Various Intensities,
Durations and Intervals of
Occurrence
Representative Bacteriological Data
for Combined Sewer Overflow
181
185
Waste Loadings Associated with
Storm of Various Intensities,
Durations and Intervals of
Occurrence
Correlations between COD and
Suspended Solids Loadings and
Total Rainfall
186
194
Representative Rainfall and Runoff
Measurements for Storms of Various
Intensities, Durations and
Intervals of Occurrence
Correlations between Water
Quality Parameters for Various
Sewer Districts
Comparison of Flow Measurement
Techniques for Combined Sewer
Overflows
196
199
201
122
-------
Appendix Page
F ULTRA-HIGH-RATE FILTRATION 203
Introduction 203
Theoretical Background 203
Considerations for Practical Application 209
Description of Laboratory System 210
Discussion of Laboratory Procedures and Data 213
Table F-1: Characteristics of Fiberglass Plug
Filter Beds 224
Table F-2: Summary of Filter Cartridge
Configurations 225
Table F-3: Characteristics of Media Used in the
Tri-Media Filters 226
Table F-4
and F-5: Filtration Study Data Summary for
Fiberglass and Tri-Media Filters 227
Figure F-1: Schematic Diagram of Filtration
System 229
Figure F-2
through F-7: Results of Fiberglass Filter Evaluation 230
Figure F-8
through F-10: Results of Tri-Media Filter Evaluation 236
Figure F-11: Upflow Filter Performance 239
Figure F-12: Filtration System Components 240
G KINGMAN LAKE PROJECT 241
123
-------
APPENDIX A
GEOLOGICAL AND OTHER NATURAL CONDITIONS
Introduction
Information on the geologic and other natural conditions is essential to a complete
evaluation of the various possible approaches to the abatement of pollution caused by
the inadequacies of combined sewers, including the feasibility of tunneling beneath the
City of Washington to provide for storage of combined overflows.
Data were obtained from published geologic reports, from engineering reports prepared
under the auspices of the Metropolitan Area Rapid Transit Authority, and from verbal
communication with personnel from the Authority. The published geologic data are fairly
broad in coverage, although sufficiently detailed to provide information helpful in forming
conclusions in regard to the feasibility of tunneling. The previously-mentioned engineering
reports are extremely detailed and provide an abundance of data concerning the soils,
bedrock, and rock mechanics along certain restricted areas for the purpose of underground
tunneling for rapid transit. Although somewhat restricted as to area-wide application, the
data included in these reports provide enough coverage to be representative of the overall
Washington area.
Geologic Setting
The District of Columbia lies within portions of two physiographic provinces; the
southeastern portion is located within the Coastal Plain Province, which consists of
relatively flat-lying sediments overlying deep bedrock, and the northwestern portion is
in the Piedmont Province, which in general is characterized by a thin layer of overburden
covering crystalline bedrock. The Fall Line separating the two provinces extends roughly
southwest from Blair Park in the northeast through Farragut Square and on toward the
Pentagon. Figure A-1 presents the generalized geology of the District of Columbia.
Previous subsurface investigations throughout the District have resulted in grouping the
materials into five major categories: bedrock, Cretaceous sediments, Pleistocene terrace
deposits, recent river alluvium, and drainage channels and man-made fills. These major
categories of materials in various parts of the District are found in the following five
vertical profiles:
1. Recent alluvium over bedrock or Pleistocene terrace deposits,
2. Overburden of Pleistocene terrace and Cretaceous coastal plain soils above deep
bedrock,
3. Comparatively thick cover of Pleistocene terrace and Cretaceous coastal plain
soils above deep bedrock,
4. Thin to moderately thick cover of Cretaceous coastal plain materials above
decomposed rock and bedrock,
5 Relatively thin cover of man-made fill and decomposed rock and bedrock at
shallow to moderate depths.
125
-------
Principal Unconsolidated Formations
Cretaceous Sediments
Lower Cretaceous sediments belonging to the Potomac group overlie weathered bedrock
throughout most of the downtown area of Washington. These sediments dip gently to
the southeast, thickening in that direction. In general, they vary in lithology from weathered
soft clayey materials to dense sand, silt, and gravel. In other areas along the Atlantic
Coastal Plain, the Potomac group is overlain by upper Cretaceous and Tertiary sediments.
However, in downtown Washington, these younger formations plus a considerable thickness
of the Potomac group have been removed by past erosion. The sediments occur over
sufficient areas and at elevations appropriate for tunnel storage.
Pleistocene Terrace Deposits
Pleistocene terrace deposits are a succession of river deposits, which generally overlie the
Cretaceous sediments or decomposed rock. They consist of a heterogeneous mixture of
interbedded sandy clays, sand, and gravelly sands.
Previous investigations have determined that the terraces occur at several characteristic
elevations in the Washington area, e.g. the "25-foot terrace", "50-foot terrace", and the
"90-foot terrace". The terraces characteristically show a change in gradation (in vertical
profile) from coarse-grained soils at their base to fine-grained sands, silts, and clays at
shallower depths. In general, the Pleistocene deposits occur at elevations which are
considered too high for storage tunnels.
Recent Alluvium
In comparison to the other unconsolidated formations in the Washington area, the recent
alluvium is relatively restricted in areal extent. It occurs primarily along the Potomac River
as far west as the mouth of Rock Creek, and along the flats bordering the Anacostia
River. The alluvial material consists of fine-grained organic sand and silty to sandy clay
with lenses of peat, and may have a thickness of as much as 25 feet.
Artificial Fills
Cutting and filling of irregular natural topography has been quite extensive throughout
the Washington area. Reference to old maps indicates that extensive filling has taken place
along the low areas bordering the Potomac and Anacostia Rivers, and in East Potomac
Park, National Airport, the Navy Yard, and the Southwest Mall area. In addition, the
drainage systems of the Tiber and St. James Creeks and of Slash Run, which were originally
located in downtown Washington, have been covered with fill.
Several hills composed of Pleistocene terrace materials have been removed and the material
utilized to fill nearby low areas. Extensive cuts and fills have been made along Connecticut
Avenue north of Rock Creek. Due to the lack of continuity of the fills and to the elevation
at which they occur, they are not considered suitable of tunneling.
126
-------
Bedrock
Bedrock in the Washington area consists of crystalline metamorphic schists and gneisses
of Paleozoic age. It originated in upgrading by metamorphism of wide areas of sandy
and clayey sediments to form the lithologic complex known as the Wissahickon formation,
which is a part of the extensive band of crystalline rocks that extend from New England
to Alabama. The general category is schistose gneiss, which is essentially a metamorphosed
sedimentary rock with zones of various mineral compositions. In many areas of Washington,
rock is overlain by compact residual soil which is derived from weathering and
decomposition of the underlying parent rock. Northwest of the Fall Line, remnants of
residual soil over 60 feet thick have been encountered, while southeast of the Fall Line
erosion has removed much of this material leaving an average thickness of only five feet.
General Depths of Bedrock
The bedrock surface southeast of the Fall Line dips southeastward at a rate of
approximately 60-125 feet per mile, thus occurring at increasing depths toward the
southeast. In the area of the Anacostia River, it occurs between 300 and 400 feet below
mean sea level. Northwest of the Fall Line, bedrock begins to outcrop at the surface,
although in many areas it is overlain by a cover of unconsolidated material. In this area,
bedrock generally occurs between the surface and a depth of about 75 feet. Figure A-2
depicts the general configuration of the bedrock surface underlying the City of Washington.
Depth of Rock Weathering
A study of the logs of bore holes drilled for the Metropolitan Area Transit Authority
indicates that the upper surface of the bedrock generally is weathered to some extent.
The extent of weathering is largely dependent on rock type and on the frequency of
jointing. The logs indicate the general depth of weathering to be on the order of 5 feet,
although weathered thicknesses up to 17 feet were noted in several of the borings.
Rock Types
Through subsurface investigations, three interfingered general bedrock types have been
identified along Connecticut Avenue. These include schistose gneiss, chlorite schist, and
quartz-diorite gneiss; these are described below, since each has different characteristics
in regard to tunnel construction. These rock types are considered to be generally
representative of the rock and conditions underlying Washington.
Schistose Gneiss
The schistose gneiss includes three rocks of distinctive mineral composition: hornblende
gneiss, quartz-hornblende gneiss, and quartz-biotite gneiss. The hornblende gneiss is
characterized by its dark color and high content of hornblende crystals, combined with
lesser amounts of quartz and feldspar. Compared with the other two rock types, schistose
gneiss is more subject to weathering and lower in compressive strength and modulus of
elasticity. The quartz-hornblende gneiss is distinguished by its "salt and pepper" appearance,
and has the highest compressive strength and elastic modulus of the schistose gneisses.
The quartz-biotite gneiss is similar in character to the quartz-hornblende gneiss, but has
slightly less compressive strength and modulus of elasticity.
127
-------
Chlorite Schist
The chlorite schist is a gray-green to dark green rock, which corresponds to the soapstone
that has been previously mapped from outcrops northwest of the Fall Line. The occurrence
of talc and the presence of joint planes in which the chlorite schist has weathered to
clay make the formation structurally the weakest of the bedrock types. It has been
encountered in the valley of Rock Creek and in other areas lying to the north. Because
of the weakness of the chlorite schist, tunnels would probably require structural support
whenever this rock type is encountered. The existing data indicate that this rock type
is of relatively limited extent.
Quartz-Diorite Gneiss
The quartz-diorite gneiss (also referred to as granite gneiss or biotite gneiss) is a
light-colored, coarse-grained gneiss with prominent flakes of biotite mica. Eighty to ninety
percent of the rock is composed of feldspar and quartz. It has been encountered north
of Klingle Creek, along Connecticut Avenue. In general, the quartz-diorite gneiss is the
most structurally favorable bedrock from the standpoint of tunnel construction. It is the
strongest and least weathered of the bedrock types and has the highest compressive strength
and modulus of elasticity. Tunnels constucted in this rock type would probably require
the least amount of structural support.
Structure of Bedrock
Faults and Folds
The crystalline bedrock floor underlying the Coastal Plain sediments has been folded and
faulted to some extent during the geologic past. Since early Cretaceous time, much of
the region has been uplifted and depressed many times, accompanied by folding and faulting
of the bedrock surface. However, due to the thickness (and in some cases, partial removal
of the overlying sediments), data on the directional trend and amount of movement of
these folds and faults are extremely difficult to obtain. The predominant trend of these
movements and resultant structures has been along northeast-south-west axes, although
localized variations, both in trend and amount, do exist. Several faults in bedrock and
the overlying sediments have been observed in the northwestern part of Washington, but
the displacements appear small and the faults short. One fault, which was exposed in
a trench on 18th Street near California Avenue, showed a displacement of 40 feet where
Potomac group sediments and crystalline bedrock had been thrust in contact with each
other. Another fault, on Adams Mill Road, has a vertical displacement of only 8 feet.
Joints and Foliation
Previous investigations have been made in regard to determining the general foliation pattern
and the primary and secondary jointing pattern of the bedrock underlying Washington.
Foliation is a layering within a rock caused by segregation of various constituent minerals
due to metamorphism. It is an important factor to consider in regard to tunnel construction
in that rock breakage may be controlled by the attitude of foliations. Joints are fractures
or partings which interrupt the physical continuity of a rock mass. The attitude and
frequency of joints are extremely important criteria in the construction of tunnels, and
zones of intense jointing are areas which may require special structural treatment. A detailed
examination of outcrop in Rock Creek Park was made by R. E. Fellows in 1950. Another
investigation made by the Metropolitan Area Transit Authority along Connecticut Avenue
128
-------
utilized bore hole photography to determine the attitude of foliation and jointing. The
results of both investigations show a good correlation of attitudes and indicate a definite
consistency throughout the study areas, with little effect from lithologic variations. Based
on the results of the above-mentioned investigations, the prinicipal conclusions regarding
the attitude of foliation and jointing are as follows:
1. The general foliation pattern trends directionally between the true north and
N20°E, and dips at angles from the horizontal between 45° and 75°, with an
average angle of 60° to the northwest.
2. The primary or major jointing pattern trends directionally between I\I70°W and
IM80°W, with the joint surface dipping to the northeast between 45° and 75°
from the horizontal.
3. The secondary jointing pattern trends directionally between N20° and N40°E,
with dips ranging between 30° and 60° to the southeast.
Frequency of Secondary Openings
For the purpose of this study, secondary openings are considered to be joints which are
defined as fractures or partings which interrupt the physical continuity of a rock mass.
Joint frequency is a criterion by which the quality of a rock mass can be judged; the
greater the frequency of joints, the lower the quality of the rock mass. In brief, frequency
is based on the number of joints occurring within a given interval of rock core or bore
hole, computed on the basis of joints per foot. Frequency of joints was studied during
the Metropolitan Area Transit Authority investigations, which included the use of bore
hole photographic logs and of core samples recovered from the bore holes. The logs
indicated a frequency of joints of one for every two feet of depth. Examination of core
samples showed a frequency range of 1.3 to 1.9 joints per foot of depth. Concentrations
of closely spaced joints in broken zones were omitted in making the frequency counts.
It should be assumed that localized shear zones with more intense jointing may be
encountered, particularly in contact zones between different rock types. In regard to tunnel
concentration, zones of intense jointing might necessitate structural support or sealing off
of the zone to prevent inflow of ground water.
Hydrology
Present and Potential Ground-Water Use
Ground water in the Washington area is obtained from both the unconsolidated Coastal
Plain sediments and from the crystalline rock of the Piedmont sediments and from the
crystalline rock of the Piedmont province. In general, the upper portion of the Coastal
Plain sediments are clayey and contain few highly productive aquifers; the older and deeper
strata are sandier and furnish moderate yields to wells. As of 1960, approximately 2 million
gallons per day were being pumped from wells located within Washington; this is lower
than the pumpage rates of earlier years. Most of the well water pumped in 1960 was
from wells located at the railroad terminal and at other industrial and commercial sites
in the area east of North and South Capitol Streets. Since the Piedmont section of
Washington is largely residential and is supplied with surface water, very little ground
water is utilized there. However, ground-water studies of nearby Montgomery and Howard
Counties in Maryland indicate that the crystalline rock of the Piedmont is capable of
yielding as much as 180 gallons per minute per well.
129
-------
It is estimated that several additional million gallons per day of ground water can be
withdrawn from the aquifers underlying Washington without appreciably affecting the
present hydrologic balance. The general water supply for Washington is from surface
sources, in particular the Potomac River, and is considered adequate for present and future
demands. However, serious consideration must be given to protecting the valuable
ground-water resource from pollution and depletion, in case further demands and changes
in the water budget necessitate its use.
Ground-Water Levels
Ground-water levels in the Washington area are dependent on several factors, such as
seasonal changes of infiltration, presence of storm sewers to carry water away, temporary
or permanent pumping associated with construction, and variation in river levels. Three
observation wells used by the U. S. Geological Survey indicate that the typical yearly
variation, from a high in February or March to a low in August or September, is
approximately 5 feet. A study of well data and water levels in the borings made by the
Metropolitan Area Transit Authority shows that the ground water level in Washington
generally occurs at depths between 16 and 38 feet.
Permeability of Rock and Unconsolidated Sediments
Permeability tests performed on the various lithologies encountered indicated that, in
general, the overall permeability decreases with depth and increasing age of the deposit,
although local exceptions to this should be expected.
The overall media permeability coefficients for the various lithologies are as follows:
1. Pleistocene sands and gravelly sands 5 x 10'4 feet/minute (fpm)
2. Cretaceous sands and gravelly sands 3 x 10~4 fpm
3. Decomposed rock 1 x 10"^ fpm
4. Bedrock of all types 4 x 10'° fpm
Tunneling
Types of Tunnels
Two general tunnel types, earth and rock, have been considered for the purpose of storage.
The earth tunnel, which would be located in the area of coastal plain sediments, would
require a complete lining. It is anticipated that the rock tunnel type would be constructed
in areas where depth to bedrock is not too great. This type would probably require a
minimum of lining and structural support.
Certain desirable geological criteria reguarding the feasibility of rock tunneling have been
established as follows:
1. Rock type capable of being tunneled with a minimum of structural support.
2. Bedrock with a minimum number of shear and fault zones, which would require
additional structural support.
3. Bedrock of high stability and relative insensitivity to tectonic disturbances.
130
-------
4. Bedrock of low porosity and permeability, to minimize the amount of water
seepage into tunnels.
5. Location below the water table, to prevent contamination of ground water.
6. Sufficient depth, to avoid any effects of overlying structures or nearby existing
tunnels.
7. Construction in fresh unweathered rock.
8. Location as shallow as possible, yet conforming to the above criteria, to keep
the economics of construction and operation at a minimum
Undesirable conditions for rock tunneling are as follows:
1. Bedrock types with thick or irregular zones of weathering.
2. Bedrock susceptible to rapid weathering.
3. Rock types of high permeability.
4. Areas of variable rock types.
5. Easily deformable bedrock.
6. Bedrock that has been subject to much faulting and shearing.
7. Unstable areas subject to tectonic disturbances.
Experience has shown that the rock types underlying the Washington area tend to become
less weathered with increasing depth. In addition, as depth increases, secondary openings
become tighter and occur with less frequency. Tunnels constructed for the storage of
the combined flow will probably be located at such a depth that most of the weathered
zones will be avoided.
Existing Tunnels
Earth Tunnel
An earth tunnel was recently constructed in southeast Washington for the District Sanitary
Engineering Department. It is a 12-foot diameter sewer tunnel extending from L and Half
Streets, S.E. to the District pumping station at the Navy Yard. The tunnel was excavated
through recent alluvium and terrace deposits utilizing a shield-driving method. During
construction of the tunnel, dewatering was provided by deep wells spaced at minimum
intervals of 300 ft. Although the quantity pumped was small, it was felt that sloughing
of cohesionless sand was significantly reduced.
Rock Tunnel
A rock tunnel was completed in 1966 for the National Park Service on Rock Creek and
Potomac Parkway. The tunnel is located near the Zoo and about 1/2 mile from Connecticut
131
-------
Avenue. It is 780 feet long with a 30-foot diameter, and extends through a rock ridge
at a distance of 80 to 100 feet from ground surface to tunnel base. The weathered condition
of the rock encountered during construction necessitated that a concrete lining be used
for the entire length. Problems from heavy overbreak in the rock occurred during
construction. Seepage of ground water caused little difficulty, except where bore holes
had not been grouted prior to construction to provide storage of the combined sewer
flows, because most of the construction will be in massive unweathered bedrock.
A more typical example of rock tunneling was encountered in the construction of the
Lydecker Water Tunnel (1888-1902). This tunnel extends from Georgetown Reservoir to
MacMillan Park Reservoir, and runs at depths ranging from-60 to 170 feet. The tunnel
has a horseshoe-shaped section and is brick-lined for most of its length; however, the
section underlying Rock Creek is circular and steel supported. Except for some minor
cracks and bulges that developed near the Georgetown end of the tunnel (which
subsequently needed additional support), no evidence has been reported of swelling or
breaking rock.
Rock Drillability
Laboratory tests, performed on the different rock types for the purpose of estimating
the practicability of using tunnel excavation machines in the Washington area, measured
various factors relating to rock hardness and strength. A combination of rebound hardness
and abrasion hardness, known as "total hardness", is thought to be the most closely related
to machine drillability. Past performance indicates that rock with total hardness values
less than 120 has been successfully drilled by machines presently in use. The results of
the tests are summarized in the following tabulation:
Compressive Modulus of Total
Strength Elasticity Hardness
p.s.i. p.s.i.
Chlorite Schist 5,000 2 x 106 20-30
Schistose Gneiss 10,000 5 x 106 90
Quartz-diorite Gneiss 15,000 8 x 106 110 130
It appears that the chlorite schist and schistose gneiss can be quite easily tunneled by
the machines now in use. On the other hand, the quartz-diorite gneiss would probably
present difficulty with respect to maintenance, drilling progress, and cost.
Estimates of the rate of drilling, based on laboratory tests and on a measurement index
known as "Reed" drillability, indicate the rate of mechanical boring in feet per hour.
The following rates have been estimated for the three general rock types:
quartz-diorite-gneiss 3 ft/hr.; schistose gneiss 3-1/2 4-1/2 ft/hr.; chlorite schist 8
10 ft/hr.
Feasibility of Tunneling
Based on existing geologic literature and the Transit Authority investigations, tunnels for
storage of combined sewer flows appear to be feasible. The bedrock geology of the
Washington area in general meets most of the criteria necessary for successful rock
tunneling. Although there are areas where the condition of the bedrock will require much
132
-------
additional structural support, most of the bedrock should be capable of being tunneled
with a minimum of lining and structural support. It is probable that the principal mode
of support would be a combination of rock bolting and several layers of
pneumatically-applied "shotcrete" covering the roof area of the tunnel. Because of the
low permeability of the bedrock, seepage of ground water will probably be insignificant
except in areas where the rock has been intensely sheared. In these cases, the shear zones
will have to be sealed to prevent the inflow of large quantities of ground water into
the tunnel. In general, the Washington area is relatively stable tectonically; therefore, no
particular problems are anticipated in regard to rock deformation and faulting due to
tectonic activity.
Earth tunnels are considered feasible from a geologic standpoint. At this stage of the
investigation, the best sediments in the Washington area in which to construct lined-earth
tunnels appear to be the more dense Cretaceous clays, sands, and gravels. They are
considered best because of their greater horizontal continuity and their occurrence at more
suitable elevations. Even though earth tunnels are feasible geologically, the fact that a
complete lining is required may inhibit their use for sub-surface storage, for economic
reasons.
133
-------
DISTRICT OF COLUMBIA
GENERALIZED GEOLOGIC MAP
LEGEND
PRIMARILY ALLUVIUM AND ARTIFICIAL FILL
UNDIFFERENTIATED CRETACEOUS, TERTIARY AND
QUATERNARY FORMATIONS CONSISTING OF
CLAY, SILT, SAND A?4D GRAVEL
RIVER TERRACE DEPOSITS CONSISTING OF
GRAVEL, SAND AND LOAM
PRIMARILY CRETACEOUS CLAY, SAND AND GRAVEL
MASSIVE LIGHT GRAY, COARSE-TEXTURED GRANITE
GRANITE GNEISS WITH LAYERS OF SCHISTOSE
GRANITES, GNEISS AND SILICEOUS MICA SCHISTS
1500 0 1500 3000 4500 6000
SCALE IN FEET
FIGURE A-l
-------
CO
Ul
DISTRICT OF COLUMBIA
CONTOURS OF BEDROCK SURFACE
1500 0 1500 3000 4500 6000
SCALE IN FEET
FIGURE A-2
-------
APPENDIX B
A LIST OF PERTINENT STUDIES AND REPORTS REVIEWED
1. Metcalf and Eddy, "Report Upon Investigation of Sewerage System", 1955.
2. Board of Engineers (Greeley, S. A., et al), "Report to District of Columbia Department
of Sanitary Engineering on Improvements to Sewerage Systems", 1957.
3. Greeley, S. A., et al, "Report to District of Columbia Department of Sanitary
Engineering on Improvement to Sewage Treatment Plant", 1964.
4. Metcalf and Eddy, Engineers, "A General Plan for Development of the Water Pollution
Control 1969-1972), (1969).
5. Moorehead, G. J., "Overflows from Combined Sewers in Washington, D.C.," JWPCF,
July 1961, p. 711.
6. Johnson, C. F., "Equipment, Methods, and Results from Washington, D.C., Combined
Sewer Overflow Studies", JWPCF, July 1961, p. 721.
7. The Potomac Interceptor Symbol of Metropolitan Cooperation.
8. Project C. Improvements to the Sewerage System of the Nation's Capital.
9. Storm Sewer Preparation Program.
10. Water Pollution Control Plant of the District of Columbia, 1967.
11. Eddy-Gregory-Greeley Report, 1933.
12. Sherman-Horner Report, 1935.
137
-------
APPENDIX C
INVESTIGATION OF RAINFALL RUNOFF RELATIONSHIPS
Rainfall Data
Rainfall data specific to the Washington, D.C. area were used in this investigation.
Intensity-duration-frequency relationships have been formulated and plotted in a previous
study using rainfall data collected during a 60-year period from 1894 to 1954. These
relationships are defined in the following equation for return frequencies from 2 years
to 100 years.
'avg (td + b) c
where:
t^ = duration of storm corresponding to a period of maximum rain-
fall, minutes
'avg = avera9e intensity during particular duration, in./hr.
a, b, c = constants for particular return frequencies
In addition, a plot of these relationships has been updated by the Special Studies Branch
of the Office of Hydrology to reflect additional data collected at the National Airport
for 1951-1969. These updated relationships have only been plotted and not formulated
in the form of equation (C-1).
Peak Rate of Overflow
The rational method was used to determine the peak rate of runoff in each sewer district
and employed the following equation:
Q = CIA
where:
I
Q = rate of runoff in cu.ft./sec.
C = runoff coefficient
I = rainfall intensity in in./hr.
A = drainage area in acres
The areas used were the net sewage-producing areas of each basin. An average runoff
coefficient was determined for each sewer district from zoning maps and runoff coefficients
applicable to each type of zoning. This required planimetering the maps of the Zoning
Regulations of the District of Columbia (effective May 12, 1958} to determine the area
occupied by each type of zoning. Coefficients for individual types of zoning were provided
by Washington, D.C. The overall coefficient for each drainage basin was then determined
by a summation of the products of the appropriate average coefficient for each type of
zoning multiplied by the fraction of the products of the appropriate average coefficient
for each type of zoning multiplied by the fraction of the total area occupied by that
type.
139
-------
Different intensities were used in equation (C-2) together with a constant area and runoff
coefficient to determine the rate for runoff produced in each district for different storm
return frequencies. The intensity used for each storm was the average intensity in equation
(C-1) which corresponded to the period of maximum rainfall equal to the time of
concentration of each district.
The rational method determines the peak rate of runoff, not the peak rate of overflow.
To determine this, a cursory examination was made of the hydraulic characteristics, control
structures, and average dry-weather flow at each point of overflow. The rate of flow not
overflowing, but continuing on in the sewer system was subtracted from the peak rate
of runoff to determine the peak rate of overflow. At overflow points with complete
diversion of all flow, the rate of sewage flow had to be added to determine peak overflow
rates. The specific information relative to capacity of interceptors, operation of diversion
structures, etc. was obtained from the District officials, as well as from data included
in previous reports. Table 5 of the main report lists the peak rate of overflow at various
diversion structures.
Volume of Overflow
Hyetograph Method Volume of Runoff
The volume of the surface runoff of rainfall from a particular area is determined by
subtracting from the total amount of precipitation the losses due to interception by
vegetation, infiltration into permeable soils, retention in surface depressions, and
evaporation. Of these, infiltration and retention in depressions are the only losses of
significance in urban drainage. Numerous runoff coefficients accounting for the combined
effect of these losses have been widely reported; however, the use of these coefficients
calls for extensive judgment because the coefficient will vary throughout the duration
of the storm. It is more precise to account for these losses separately and to examine
their variability through time.
The methododogy employed to account for these losses involves the development of
hyetographs. Using differential and integral calculus, the equation for the hyetograph is
deriued from the equation (C-1) to be as follows:
(C-3a)
' — (C-3b)
where:
tj, = time before the peak intensity, minutes
tg = time after the peak intensity, minutes
i[j = instantaneous intensity before the peak intensity, in./hr.
ia = instantaneous intensity after the peak intensity, in./hr.
r = portion of any duration of maximum rainfall occurring before
the peak intensity
Note: tb = rtd, ta (l-r)td
a,b,c constants from equation (C-2)
140
-------
A study (28) of the storms occurring in the Chicago area showed that r in equation B-2
has a weighted average value of 3/8. A detailed examination of twelve consecutive months
of rainfall as recorded at the Washington National Airport Station indicates that a value
of 3/8 is reasonable for the Washington, D.C. area. Hyetographs for the relevant storm
frequencies were developed on the basis of this value and of the values for a, b, and
c in equation (C-1). As an example. Figure 5 in the main report is the hyetograph for
various rainfall frequencies.
Research (29) indicates that the capacity which a soil exhibits for infiltration varies little
with surface slope and depends primarily on soil porosity, ground cover, and antecedent
rainfall. Infiltration capacity curves have been developed for various soil and surface
conditions, and two of the mostly commonly used curves are shown in Figure 6 of the
main report.
The capacity for infiltration is relatively high at the beginning of precipitation and decreases
rapidly to a rather definite minimum value. During periods when the rate of precipitation
is less than the infiltration capacity, all precipitation is infiltrated; when precipitation is
greater than the infiltration capacity, excess runoff is produced in an amount equal to
the difference between the precipitation rate and infiltration capacity. Consequently, an
analysis of runoff requires the simultaneous examination of precipitation rates and
infiltration capacities as they vary through the duration of the storm.
Figure 7 of the main report is a graph of the accumulated mass of rainfall and the
accumulated mass of infiltration for residential type pervious surfaces versus time from
beginning of significant rainfall. For storms in which the precipitation rate is initially less
than the infiltration capacity, it is reasonable to assume the same amount of infiltration
will occur as if the infiltration capacity was exceeded, but at some later time in the course
of the storm Therefore, the time of the start of excess is found by shifting the accumulated
mass of infiltration curve along the time axis until it is tangent to the accumulated mass
of rainfall curve. The actual mass infiltrated follows the trend of the accumulated mass
of rainfall from zero time to the point of tangency, and from there it follows the offset
curve of accumulated mass of infiltration up to the point when infiltration again equals
the rate of precipitation. To define the latter point, the infiltration capacity curve is plotted
with the hyetograph but at a time-offset equal to the offset of the accumulated mass
of infiltration curve. The point at which infiltration again equals precipitation is defined
as the second intersection of their curves. This procedure is illustrated in Figure 8 of
the main report.
After the precipitation rate again drops below the infiltration capacity, the deficit is
satisfied in part by infiltration from runoff traveling overland. It is conservative, but
reasonable, to assume that for an urban area of moderate grade and intense development
such as the Northeast Boundary Trunk Sewer basin, the time of overland travel on a
pervious surface is so short that this type of infiltration is insignificant. On this basis,
the rainfall in excess of infiltration is defined as the difference between the accumulated
mass of the rainfall curve and the actual accumulated mass infiltration curve at the point
when the rate of precipitation equals, for the second time, the infiltration capacity. The
rainfall in excess of infiltration is shown in Figure 5 of the main report.
In this study, the rainfall in excess of infiltration was determined for pervious surfaces
characteristic of residential and commercial-industrial areas. It is assumed in this study
that infiltration is negligible for impervious surfaces and that runoff from an impervious
141
-------
onto a pervious surface is insignificant, or at least the effect of such is negligible. A past
study by the D.C. Department of Sanitary Engineering provides a sound basis for
determining the area of pervious and impervious surfaces. In the study, the pervious and
impervious areas of typical blocks of the city were reported as follows:
Block Type Pervious Impervious
Residential, single house 54.1 percent 45.9 percent
Residential, row house 35.8 percent 64.2 percent
Commercial, neighborhood 25.8 percent 74.2 percent
The overall surface characteristics of the drainage basins were determined by assuming
that zoning categories R-1 and R-2 correspond to the single house residential block type;
R-3, R-4, and R-5 correspond to row house residential block type; C-1, C-2, C-M, and
M correspond to neighborhood and commerical block type; park area is totally pervious;
and C-3 and C-4 are totally impervious. On this basis, the percent impervious, percent
residential-type pervious, and percent commercial-type pervious was determined for each
sewer district.
It should be noted that any hyetograph is asymptotic to the time axis. However, in this
study the hyetograph was not extended beyond the point at which rainfall intensity is
less than 0.2 inches/hour. For pervious surfaces, the minimum infiltration is greater than
0.2 inches/hour and there is no need to examine rainfall past this point. Nevertheless,
prior to the peak of a storm the mass of rainfall preceding an intensity of 0.2 inches/hour
may be significant. For impervious surfaces, the total amount of precipitation for each
rainfall frequency examined was the 24-hour duration value read from the updated
Washington, D.C. intensity-duration-frequency curves published by the Special Studies
Branch of the Office of Hydrology of the U. S. Weather Bureau.
o
A significant volume of the rainfall in excess of infiltration is retained in surface depressions,
and it either evaporates or infiltrates after the storm subsides. Observations made during
periods of heavy rainfall for the flat topography of Chicago suggest that the overall average
depth of depression storage is 0.25 inches on pervious areas and 0.06 inches on paved
areas (5). In a widely accepted design manual (30), various investigators have observed
that in urban areas of moderate grade the overall average depth of surface depressions
is about 0.05 inches for impervious surfaces and about 0.10 inches for pervious surfaces.
The values reported in the design manual were assumed to apply to the Washington, D.C.
area and were subtracted from the mass of rainfall in excess of infiltration.
The application to a large drainage area of rainfall data collected at one point outside
the area requires careful interpretation of two phenomena:
a. The difference in prevailing physical conditions at two separate, but proximate,
locations may result in different extreme rainfall data.
b. Average depth of rainfall over a large area is less than maximum point rainfall.
As for the former, the atmospheric forces of extreme storms are greater than the local
effects of terrain and thermal patterns in the Washington area, and there should be no
difference in extreme rainfall data (i.e. intensity versus frequency) at two separate sites
in the area. To account for the latter, a correction factor of 98 percent was extrapolated
142
-------
from the area-depth relationship used by the D.C. Department of Sanitary Engineering.
The volume of runoff from each drainage basin is determined by multiplying the mass
of rainfall (less abstractions) by the appropriate surface area, summing, and reducing by
the appropriate percent value to account for the depth-area relationship.
Hydrograph Routing Volume of Overflow
The hyetograph method provides reliable values for the volume of runoff from any sewer
district; however, this value may differ from the volume of overflow. To determine the
volume of overflow, a method of hydrograph routing was used.
For hydrograph routing studies, the hydrographs obtained by the rational method were
then further simplified by treating them as triangular hydrographs with their peaks located
at the time of concentration of each drainage basin. Some of the critical
time-of-concentration values were obtained from the District officials, while the remaining
were calculated by considering the drainage area characteristics of each drainage district.
A time offset method of routing was utilized to route the hydrographs down the interceptor
sewer to ascertain the cumulative effect at any point along the interceptor. This permitted
development of a volume and peak rate relationship. A velocity of three feet per second
was assumed along the interceptor, and the appropriate hydrographs from the tributary
districts were plotted. Each drainage district hydrograph was offset an amount equal to
the period of time required to reach the point in question. The resulting hydrograph was
then found by summing up the ordinates of the individual hydrographs at the required
points.
The hydrograph developed from hydrograph routing was compared with the volume of
runoff determined by the hyetograph method. Using a heuristic procedure, the two types
of information were combined to develop a synthetic hydrograph equal in volume to the
runoff determined by the hyetograph method but smoothed and extended to follow the
general shape of actual hydrographs. The dry-weather flow was added to the synthetic
hydrograph to provide a plot of flow rate versus time at each point of overflow
To determine the volume of overflow, a cursory examination was made at each point
of overflow. This examination practically duplicated the examination made to determine
the peak rate of overflow, i.e. interceptor capacity and type of diversion structure operation
were accounted for. Table 5 lists the volume of overflow (by point of overflow)
corresponding to four different storm frequencies.
143
-------
APPENDIX D
MONITORING EQUIPMENT, INSTRUMENTS, AND PROCEDURES
Discussion of Facilities and Operating Procedure
At each monitoring site, two fenced metal sheds, 400 to 800 feet apart, were erected
along the sewer trunk line in the vicinity of the overflow point of the monitored sewer
district. The upstream shed was set up as the lithium chloride release station and the
downstream shed as the sample collection station. Each pair of stations was integrated
electrically into an operating system by an underground conduit installed in the sewer.
During each storm, the system was actuated by the back pressure of a bubbler system
installed at the manhole near the lithium chloride release station. Lithium chloride was
injected as a tracer for accurate measurement of the unsteady storm discharges in the
sewers. With the lithium chloride continuously released at a specified metering rate into
the sewer, wastewater samples were collected intermittently at predetermined sampling
intervals varying according to the duration of storm. All the collected samples were
refrigerated at 3°C until they were ready for analysis. Discharge flows were computed
on the basis of the lithium concentration detected in the collected samples and on
application of the sample mass-conservation relationship. Major equipment and instruments
installed at each site are shown in Figure D-1 as a schematic flow diagram.
Descriptions of Principal Equipment and Instruments
Lithium Chloride Release Station
The lithium chloride release station included a continuously-operated air bubbling system,
a lithium metering system, and an electricity triggering system A metal shed 6'4" x 3'4"
x 6'11" high housed the equipment and instruments. To provide additional protection,
a transformer-cage type of structure with a 9-gauge cyclone fence on four sides and top
was constructed outside the metal shed.
Air-Bubbling System
The air-bubbling system consisted of an air pump, a rotameter, a pressure regulator, a
mercury pressure switch, and a stainless steel bubbling-tube assembly. Air from the pump
continuously bubbled at the bottom of the sewer through the pressure regulator, flow
rotameter, and the 1/4"-stainless steel tubing system. The back pressure of the bubbling
system actuated a mercury pressure switch, which sends proper signals to the relay of
the electrical triggering system. During a storm an increase of the water depth in the
sewer activated the mercury pressure switch which, through the relay, triggered the circuits
for the lithium chloride metering system and the sample collection system.
Lithium Chloride Metering System
The lithium chloride metering system consisted of a metering pump, a vacuum breaker,
and a polyethylene piping assembly. The metering pump was a piston-diaphragm pump
with positive displacement, with output adjustable by turning the handwheel which in
turn sets the length of the stroke. Maximum theoretical metering rates were 11.9 gallons
per hour for the pump used at the B-4 Sewer District monitoring station and 26.0 gallons
145
-------
per hour for the pumps used at the other sites. The initial setting of each metering pump
was determined by the expected intensity of rainfall and by the maximum dry-weather
flow. The major moving parts of the metering pump were kept in an oil bath constantly.
The level of the oil in the gear box and compression reservoir was always maintained
at 3/4" below the top.
The vacuum breaker was an anti-siphon fixture (Watts Model No. 288A) installed for the
prevention of back-siphonage of lithium chloride through the pump head after pumping
action stopped. It contains a light-weight disc-float, which opened and closed the
atmospheric vent according to the pressure downstream of the outlet of the metering
pump. The minimum negative pressure head required in the discharge line was 6 inches;
at each monitoring site, this requirement was met all the time. The vacuum breaker was
located above the possible highest level of the lithium chloride in the storage tanks. In
order to prevent caking of lithium chloride residue in the pump, a water-rinse line was
attached to the piping ahead of the inlet to the metering pump. The lithium chloride
injection line downstream of the vacuum breaker was 1/2"-diameter polyethylene piping.
At each lithium release station, one electric clock wired into the storm time operating
circuit provided an accurate estimate of total storm operation time.
Electricity Triggering System
The Electricity Triggering equipment included one magnetic contactor, two manual motor
starters, underground conduit, and liquid-tight junction boxes, as shown in Figure D-2.
The magnetic contactor provided a safe and automatic connection for the time-operated
circuits. The magnet coil was wired to be energized by the current through the pressure
switch.
The entire assembly was housed in a water-tight enclosure on the panelboard in the lithium
chloride release station. Also on this panelboard were manual starters used for
supplementary control of the lithium metering pump and water-level recording systems.
Underground conduit was installed to integrate the operations of the lithium release and
sample collection systems; this consisted of a 3-conductor, neoprene-jacketed cable,
pre-assembled in a 1-1/2"-diameter polyethylene pipe. The conduit was fastened along
the center top of the sewer with power-charged Ramset studs and EMT strips.
Flow Depth Recording System
At one of the sampling sites in a combined sewer district, an electronic flow-depth recording
system was installed. Integrated with the bubbler tubing assembly, the flow recording
instruments included a process pressure-to-current transmitter and an electronic strip chart
recorder.
The process pressure-to-current transmitter was an electronic instrument operating on the
principle of force balance. With the bellows as the pressure element, the transmitter
produced an electrical output signal in milliamps. The monitoring range of the pressure
was fully adjustable between zero and the maximum pressure, which in turn could be
adjusted between 15 and 150 inches of water. In the present instance, the transmitter
was calibrated for a pressure range 0-100 inches of water.
146
-------
The water-level recorder was a 3" x 6" strip-chart electronic recorder. It accepted electrical
analog signals representing pressure as input, recorded the input on a 4" strip chart, and
displayed the value on a vertical indicating scale. The input signal ranged between 4 and
20 milliamps DC. An internal power supply (117 volts AC) provided operating power
for the transmitter. The accuracy of the recorder was ±0.5 percent of the range, and
repeatability was ±0.2 percent of the range. Recording chart speed was 7/8" per hour,
and the useful length of each roll of chart was sufficient to permit continuous operation
for more than a month.
Sample Collection System
The sample collection system consisted of a submersible pump, a refrigerated
vacuum-charged sampler, a wastewater retention tank, a portable vacuum pump, and the
piping system associated with wastewater flow. During a storm the sample collection
system and the lithium chloride release station were triggered by the increased back-pressure
of the bubbler line resulting from the increased depth of combined sewer overflow or
separated storm water discharge. The wastewater was pumped to the surface continuously
by a submersible pump anchored onto the floor of the sewer and flowed through the
wastewater retention tank (retention time normally less than one minute). The sample
was collected by the vacuum-charged sampler according to a predetermined sampling
interval. The wastewater retention tank, the refrigerated sampler, and the piping all were
housed in a 7' x 5'3" x 6'6" metal shed, protected by a 9-gauge cyclone fence around
its four sides and above its top.
Submersible Pump
The submersible pump was a heavy-duty, manually-controlled sewage pump. It could be
operated for long periods with the motor housing out of water, without damage. In view
of the impacts generated by the hydraulic surge and the heavy solids flowing down the
sewer during a storm, a solid, rigid anchorage for the submersible pump was designed
and constructed. To provide additional protection against large pieces of solid waste, an
expanded metal cage was installed for the containment of the pump. The legs of the
pump were fastened onto the bottom plate of the prefabricated cage with three 1/2-inch
machine bolts, and the cage was anchored onto the bottom of the sewer with four 1/2-inch
diameter Rawl plugs. To furnish additional stability for the anchorage, angle irons were
installed to brace the rigid pipe extending from the outlet of the submersible pump. The
capacity of the pump was 60 to 80 gallons per minute, depending on the pumping head
required at the different sites.
Automated Sampler
The sampler installed at each site was an automated sampling device equipped with a
portable vacuum pump and an electric timer to adjust the sampling interval to the desired
period (5 minutes minimum). Twenty-four bottles, each with an individual sampling line
and the control switch, were furnished for each sampler and installed in a refrigerated
enclosure. The intake ends of sampling lines (sampling head) were in the wastewater
retention tank, and the sample was drawn into the bottle by vacuum when the control
switch was released by a tripper arm operated in conjunction with the timer. Photographs
of the sampling components are presented in Figure D-3.
147
-------
Rain Gauges
The rain gauges were spring-driven, weighing-recording instruments, with 12 inches of range
and a 24-hour time scale. The precipitation was measured in a prebalanced collection system
by a weighing mechanism, which guided the pen to trace on a paper chart. The chart
drive was spring-wound, and it rotated a cylinder on which the chart paper was held.
The accuracy of the rain gauges in the first traverse (0" to 6") was ±0.03", and ±0.06"
in the second traverse (6" to 12").
A dashpot was provided to minimize irregularities in the chart trace. In view of the
requirements for truly representative records and to protect against vandalism the gauges
were installed on the roof of the public schools in the vicinity of the sampling points.
Their locations were: 1) Francis Junior High School, at IN and 25th Streets, NW; 2)
Chamberlain Vocational High School, at 14th and Potomac Avenue, SE; and 3) Anacostia
Senior High School, at R and 16th Streets, SE.
Each of the rain gauges was fastened to a solid foundation, and the base was carefully
leveled. To minimize possible damage from snow drifts or floods, each gauge was set up
on a wooden platform 15 inches above the roof. For each rain gauge, windbreakers of
2" x 12" x 1" white pine were installed to provide additional stability for the measurement
of rainfall, as illustrated in Figure D-4.
Lithium Chloride Flow Measurement
General Procedures and Principles
The lithium dilution procedure provides an accurate and simple technique for the
measurement of wastewater flows which may be applied when:
1. Flow measurement devices do not exist,
2. Temporary installation of flow measurement devices is difficult, and
3. Suitable conditions for application exist in the wastewater system
The lithium dilution procedure was a modification of the long-applied salt dilution
procedure. The use of a lithium salt improves the technique because the background of
lithium in wastewater is usually low, and because lithium concentrations at fractional parts
per million levels can be accurately and conveniently determined by atomic absorption
or flame emission spectroscopy.
148
-------
The technique involves preparing a lithium salt at a known concentration and then
delivering to the wastewater stream at a continuous precise rate. At a point downstream
where the lithium salt has become homogeneously mixed with the flow of wastewater,
the wastewater is sampled and the concentration of lithium is determined. The flow of
wastewater is then calculated as:
Ci Co
where:
Qw = the flow rate of wastewater in the sewer, in gallons per minute (gpm)
QI = the metering rate of the lithium chloride solution, in gpm
C-| = the lithium concentration in the lithium chloride solution, in milligrams/
liter (mg/L)
C2 = the lithium concentration in the collected sample in mg/L, and
Co = the lithium concentration in the raw wastewater, in mg/L
Since repeated analyses showed that the lithium concentration in the raw wastewater was
consistently at undetectable level. Equation D-1 was simplified by eliminating Co, as
follows:
Qw = Ql x
C2
which rearranges to:
Qw + Q-| = Q-] x
However, during a storm the metering rate of lithium chloride (Q-|) is negligible compared
to the wastewater flow rate (Qw), and thus it can be eliminated from the left-hand term
of Equation 2 without affecting the accuracy of subsequent calculations. This permits
further simplification of the equation to:
Qw = Q! x
C1
Precautions
The preparation of the lithium salt solution requires a high degree of mixing to insure
that the lithium is homgeneously dispersed. Because of the density of the salt solution,
stratification can occur, if mixing is inadequate. Once the solution is prepared, it will
remain homogeneous.
Mixing in the wastewater stream after the lithium solution is added must adequately achieve
a homogeneous dispersion of lithium throughout the wastewater stream before the
149
-------
downstream sample is taken. If the wastewater stream is essentially in laminar flow the
salt dilution procedure cannot be employed.
Application to D.C. Combined Sewer Study
Lithium chloride solutions varying in strength from 60,000 to 78,000 mg/L were purchased
from the Foote Mineral Company. The actual lithium concentration in each batch was
determined by ROY F. WESTON personnel using an atomic absorption spectrometer. The
lithium chloride solution was introduced into the sewers (by means of previously calibrated
pumps) at flow rates as high as 260 milliliters per minute. Lithium chloride solution volumes
were determined before and after each dosing period, to serve as a check on the dosing
rate. Since lithium is detectable in concentrations as low as 0.005 mg/L, flows as high
as 1,000,000 gallons per minute could be measured with reasonable accuracy.
Monitoring Practices
Storm-monitoring activity was triggered automatically by the increase of depth of water
in the sewer. Lithium chloride release, submersible pump operation, and the
sample-collection system were started by this initial impulse. However, a time lag of 5-10
minutes was built into the program for the start of actual sampling, to give enough time
for the injected lithium chloride to mix with the wastewater and to flow to the downstream
manhole. The operational metering rate for lithium chloride solution was calculated from
the storage tank levels before and after release, and from electric clock measurements
of initiation and completion of release.
After each monitored storm, the collected samples were immediately brought to the ROY
F. WESTON laboratory in West Chester, Pennsylvania for analysis. The characteristics
determined for each sample included: BOD, COD, settleable solids, total solids, volatile
solids, suspended solids, volatile suspended solids, lithium chloride, total coliform fecal
coliform, and fecal streptococci. Total phosphate, total nitrogen, and other nutrient
analyses were performed by ROY F. WESTON personnel at Baltimore's Back River Sewage
Treatment Plant.
The lithium concentration was analyzed in an atomic absorption spectrometer. All analyses
of the other chemical and physical characteristics were conducted according to the
appropriate procedures in "Standard Methods for the Examination of Water and
Wastewater", Twelfth Edition, 1967. The bacteriological analyses were basically in
accordance with the membrane filter technique suggested in "Standard Methods", with
various equipment and procedural improvements as recommended by the filter
manufacturer.
After each storm operation, the lithium chloride release lines, the metering-pump head,
the wastewater retention tank, the sample collection piping, and the sampling lines were
rinsed clean with tap water. The area around the submersible pump was cleaned of debris,
rags, and large solids. Before the systems were set up for the next storm all equipment
and instruments were turned on by hand to check for proper functioning.
150
-------
During the monitoring program, several operating problems were encountered, the most
significant of which were:
1. Difficulties in Pumping Wastewater Up from the Sewer-The submersible pump
anchored to the bottom of the sewer was often clogged by solid wastes (such
as cans, rags, wire, wood chips, tree stems, gravel, sand, etc.) and stopped
working. There were also some pump stoppages during low-intensity storms,
probably because of insufficient water depth in the sewer.
2. Physical Damage to Equipment Installed in the Sewer-During intense storms,
heavy solid wastes (such as tires, concrete slabs, 55-gallon drums, mattresses,
automobile radiators, chains, etc.) slammed into the protective cages of the
submersible pumps and caused extensive damage to various equipment items.
Bubbler lines were broken and torn loose; the protective cage was severely
deformed and even disintegrated; pump braces were sheared off; pumps were
washed away in the sewer; and the electric conduit was pulled out of its fastening
studs.
3. Flooding of Lithium Chloride Release Station at Rose Park Playground-This
station was sunk half way into the ground, at the request of the local residents.
Consequently, the lower part of the structure was inundated by excess storm
water runoff. This flooding caused minor damage to the bubbler instruments,
the lithium chloride release system and the pressure-to-current transmitter. This
problem was overcome by reinstalling the equipment above grade.
The equipment malfunctions and physical damage described above prevented complete
coverage of all the storms that occurred during the monitoring period.
151
-------
FIGURE 0-1
MONITORING EQUIPMENT
I CONDUCTOR
imtTfU CIBIE
StIPLE COLLECTION SIITIOK
SdPLIMC LINIS-
LIU! - -
~V • CLIII S«ltCHtS-v ^ [ | 1
fucmc COIDUII
•SEIER MIHHOU
-lii mn COKOUM H
I
J&-
_
1 DOINMKI III
' -\l Ilk IINHOU
UUEHll
f «»SIE«IIIR M Lii:i RILH5I I
' MR L1NI
, ELECIRICITT UNI
RIGID STEEL
EIPIHOEO-NEIIL
CUE
SUBNEBCIBLE PUNP
Rill PLUGS
D-l
-------
FIGURE D-2
ELECTRIC POWER SUPPLY AND TRIGGERING SYSTEM
14>-IOO AMPS SERVICE
R
s
o
SERVICE
SWITCH
/-
<8
)30
AMPS
MAGNETIC CONTROL
CONTACTOR
(CUTLER HAMMER)
#9560-X25)
MANUAL MOTOR [ZI
STARTER
(CHOUSE-HINDS
DS-I2B)
WATER TIGHT
JUNCTION BOX
(CROUSE-HINDS
JGRX239)
METERING
[PUMP
METER
i>30T
AMPS|_
AIR PUMPS
LIGHTS
LIQUID-TIGHT JUNCTION BOX
^
SC
<8
n n
y?
MANUAL
STARTER
(G.E. CRIOI
N700J)
ELECTRONIC
RECORDER
ft
PRESSURE TO
CURRENT
TRANSMITTER
• 31/C #6 IWG NEOPREME-JACKETED CABLE
.(GENERAL CABLE TYPE », USE)
FROM LiCI
RELEASE
STATION
LIQUID-TIGHT
JUNCTION BOX(CROUSE-HINDS »GRX239)
•»• TO SAMPLE COLLECTION STATION
N
ELECTRIC CLOCK
r^n
MERCURRY PRESSURE
PUMP
\FOUR-GANG BOX ((/LIQUID-TIGHT COVER
(HUB8ELL S286-RL-2)
TKIST-LOCR PLUGS
(HUBBEL #7736)
D-2
-------
FIGURE D-3
SAMPLING COMPONENTS
REFRIGERATED AUTOMATIC SAMPLER
INSIDE OF SAMPLER
PLATE SWITCHES OF SAMPLER
WASTEWATER RETENTION
TANK AND PIPING
154
D-3
-------
FIGURE D-4
RAIN GAUGE
RAIN GAUGE AND WIND BAFFLE
155
D-4
-------
APPENDIX E
MONITORED WASTEWATER FLOWS AND CHARACTERISTICS
Comparison of Representative Storms
The characteristics of the combined sewer overflows and of the separated storm water
discharges generated by representative storms are described herein, according to the
duration and intensity of the storms monitored.
A detailed examination of some graphs reveals an apparent discrepancy, at times, between
rainfall intensity and flow rate. For example, flow may be decreasing while it is still raining
at a significant rate. Inconsistencies such as this reflect the non-uniform spatial and temporal
distribution of rainfall over the drainage basin. Although rainfall was measured at a location
closest to the sampling point in each of the districts, the rainfall intensity may vary
considerably between that measured at the point of sample collection and that which
is actually falling in upper areas of the drainage basin.
Dry-Weather Conditions
In order to provide a more accurate definition of the pollution problem of combined
sewer overflows, a 24-hour dry-weather flow study was conducted for Sewer District G-4
on September 22, 1969. The time variations of the characteristics and of the flow rate
of the dry-weather flow from this combined sewer area are presented in Table E-1 and
in Figure E-1. The range and mean value of each of the waste constituents in dry-weather
sewage are presented in Table E-2. The characteristics of the dry-weather flow were
consistent with those expected for municipal sewage.
The dry-weather flow for Good Hope Run Separated Storm Sewer District was also
monitored; a very small flow of significantly polluted water was observed. Quality and
flow data for a 24-hour dry-weather flow study are shown in Table E-3 and Figure E-2.
Hydraulically, the variation of flow was minor. The biochemical oxygen demand was
generally between 10 and 20 mg/L; comparative chemical oxygen demand concentrations
were 40 to 100 mg/L. The bacteriological analyses show that the mean values of total
coliform and fecal coliform were, respectively, 24,000 counts/100 ml and 21,000
counts/100 ml.
Whereas the dry-weather flow from the combined sewer district exhibited characteristics
which were expected, the flow from the separated storm sewer district indicated relatively
high BOD, COD, and fecal coliform levels. The BOD's and COD s could be explained
by decaying organics in catch basins, inlets, manholes, and deposition in the pipe. The
fecal coliform level is less than one percent of the corresponding value in the combined
sewer district, and is an indicator that either some sanitary sewage or animal fecal matter
is entering the storm sewer.
Short-Duration Storms
The data derived from the samples collected for the storms on June 8 and June 15, 1969
indicate the combined sewer overflow quality and quantity in Sewer District G-4 for storms
of short duration. Figures E-3 and E-4 show the starting time, the duration, and the total
rainfall of these storms. Because samples were collected only during the period of high
157
-------
flow in the sewer (the monitoring equipment automatically shuts-off when the flow
subsides), the fact that samples were not obtained indicates that the flow rate is quite
low at the end points of the graphs, with corresponding concentrations of pollutants.
The total rainfall of each storm was the same, 0.7 inches, but the peak flow rate during
the storm of June 8, 1969 was much higher than that of June 15, 1969. The characteristics
(other than the settleable solids concentration) of the combined sewer overflow generated
by these two storms were almost identical because both storms occurred during the same
season and both had comparable waste accumulation times, as measured by the dry-weather
period before precipitation.
For the short intense storms, as shown in Figures E-4 and E-5, the concentration of each
waste constituent was observed to increase with the discharge rate in the sewer; the peak
concentrations of many waste constituents were concurrent with the peak flow.
The concentrations were significantly high and remained so throughout most of the
monitoring period. For example, for the intense storm of June 8, 1969, the flow rate
was relatively constant during the 30-minute storm discharge period, but the concentrations
of pollutants peaked in the early period of discharge and decreased to approximately one
half of the peak value after two thirds of the discharge period had elapsed. The observed
peak values for biochemical oxygen demand, chemical oxygen demand, and suspended
solids were 405 mg/L, 1,358 mg/L, and 1,268 mg/L, respectively. The fraction of the
fixed or inorganic solids was about one half of the total suspended solids, and this
essentially constant relationship reflected the composition of surface flushings.
Characteristically, the ratio of COD to BODs in the wastewater during this short storm
period varied between 3.0 and 20.
Long-Duration, High-Intensity Storms
The time variation of the quality and quantity of wastewater generated by a long-duration,
high-intensity storm at Sewer District G-4 is indicated by the data for the storm on August
2, 1969 (Figure E-6). This shows that the flow rate of the wastewater generated by this
long, intense storm is much higher than that of short storms. However, the concentrations
of various constituents were observed to be lower, which is an anticipated result of high
dilution by storm water.
Similar results were observed at Sewer District B-4, as shown in Figure E-7. In general,
the concentrations of specific waste constituents measured at District B-4 were lower than
those at District G-4. This may be explained by the fact that higher sewer flows were
required before sampling could be triggered.
Concentration data in Figures E-3 and E-6 show that chemical oxygen demand and
suspended solids concentrations of wastewater from a long, intense storm are reduced
to approximately one third of the comparable values for a short, intense storm The
biochemical oxygen demands for the long, intense storm wastewater were about one seventh
of that in short-storm wastewater. This also implies that a higher fraction of surface material
was eroded by the long intense storm. However, the ratio of COD to BOD5 was essentially
the same as that for short-storm wastewater.
Consecutive Storms
The characteristics and quantities of wastewater generated by consecutive storms can be
represented by the data of the four storms during July 27-28, 1969, as presented in Figures
158
-------
E-8 and E-9. The initial flushing effect was observed in all cases, even when the storms
were only a few hours apart. However, the average concentrations of specific contaminants
were observed to follow a decreasing trend with consecutive storms. For example, the
average concentration of chemical oxygen demand (COD) decreased from 307 mg/L for
the first storm to 154 mg/L for the fourth storm This decrease may be attributed to
the reduction of waste material accumulation on the surface and in the collection system.
The ratio of COD to BODs was essentially {he same as the comparable ratios for short
storms and for long-duration, high-intensity storms. The characteristics of long-duration,
low-intensity storm wastewater probably will be similar to that observed for consecutive
storms.
I
Separated Storm Water Discharge
Nine storms, of varying size, were monitored at the Good Hope Run Sewer District. Table
E-4 and Figures E-10, E-11, E-12 and E-13 illustrate the quality and quantity of the
runoffs from a short intense storm a long intense storm and two consecutive storms.
The average organic and nutrient concentrations in separated storm water discharges were
observed to be approximately one third of those in combined sewers, but the solids
concentrations (especially non-volatile solids and settleable solids) were much higher than
those in combined sewer overflow. However, this phenomenon is attributed to the
differences in the urban development on land surfaces in the monitored areas.
In general, the time variation in characteristics of wastewater from separated storm sewers
is very similar to that from combined sewers. These figures also show the effects of the
intensity, duration, and frequency of rainfall on wastewater characteristics. When the
interval between storms is short, the accumulation of contaminants on the surface is less,
and storm water is less contaminated.
One significant difference observed for separated storm sewer discharge is the broader
range of COD to 6005 ratio (1.8 to 32). The higher ratio probably reflects a significant
level of chemically oxidizable inorganic material introduced by erosion.
Bacteriological Quality of Storm Wastewater
Combined Sewer
The bacteriological examinations were made on selected individual samples. The range of
variation and the mean values of different bacteria species are summarized in Table 9
in the body of the report. As presented in Figure E-14, the bacteriological counts varied
with the flow rate of combined sewer overflow during each storm The initial flushing
also had a significant effect on various bacteria counts.
The mean values for total and fecal coliforms were, respectively, 2,800,000 and 2,400,000
microorganisms/100 ml.
Separated Storm Sewer
The variation of bacteriological data for the separated storm sewer discharge was similar
to the variation in the combined sewer discharge. Bacteria counts were high at the beginning
of the storm and then decreased as the storm progressed. For instance, the average bacteria
counts in the early part of the storm runoff on August 9, 1969 were 1,340,000/100
159
-------
ml, 1,300,000/100 ml, and 31,000/100 ml, respectively for total coliform, fecal coliform,
and fecal streptococci, whereas in the later part of the storm they averaged only
540,000/100 ml, 160,000/100 ml, and 17,400/100 ml, respectively.
Bacteriological examinations were made for selected samples, and their results indicated
that microorganisms were present in storm runoff in fairly significant amounts although
less than in combined sewage as cited in Table 9. The mean values of total and fecal
coliforms in separated storm sewer discharge were 600,000 and 310,000
microorganisms/100 ml.
Waste Loadings from the Monitored Sewer Districts
Waste loading during a storm is the total waste material discharged, as pounds per unit
time, and is computed from the measured flow rate and the concentration of each waste
constituent. Expression of the waste loading in terms of pounds of specific waste
constituent per minute provides an integrated assessment of the pollution potential
associated with the wastewater generated by different storms. The range and mean value
of waste loadings from the monitored storms and from the dry-weather flow in Sewer
District G-4 are summarized in Table E-5.
Waste Loading Associated with Combined Sewer Overflow
The variation in waste loadings of combined sewer overflow was found to be highly
dependent on the fluctuation of the flow rate in the sewer, which changed with the rainfall
intensity during each storm. For an impulse-type short storm, the waste loading reached
the peak value immediately and decreased very rapidly with time (see Figure E-15). For
a long intense storm (with more than one discharge peak) the loadings of various
constituents varied with the flow rate during the initial peak as anticipated (see Figure
E-16). When a second peak occurred, the loadings increased with discharge only through
the first few minutes and then dropped rapidly. Thus, most waste materials were carried
by the initial flushing and scouring of the sewer. Waste loadings carried by the secondary
flushing in a prolonged storm were limited.
The waste-loading time-variations in combined sewer overflow associated with the
consecutive storms are presented together with the flow rate and the rainfall data in Figures
E-17 and E-18. As anticipated from flow and concentration data, an initial impulse loading
was always observed for the short intense storms even though they were only a few hours
apart. Also as anticipated, the magnitude of the initial loading was found to be proportional
to the length of the dry-weather period between the storms. The organic loading after
the initial flushing of each storm was almost always low; for instance, the COD loading
at the end of each consecutive storm was between 20 and 60 pounds per minute on
the monitored areas, while the peak loadings varied between 105 and 346 pounds per
minute. This implies that a certain base value may exist for the average organic loading
in combined sewer overflow in each sewer district.
The mean waste loading of the dry-weather flow (sanitary sewage from District G-4) was
calculated to demonstrate the relationship between the pollution loads of dry- and
wet-weather flows. Because of the significant increase in flow rates during wet weather,
the pollution load during such a time may be as much as three orders of magnitude higher.
160
-------
Waste Loading Associated with Separated Storm Water Discharge
The waste loadings contained in separated storm water discharges varied with time in the
same manner as combined sewer overflows. Generally, the peak organic loading in separated
storm water discharges was lower than in combined sewer overflow, but the peak solids
loading was higher (see Figures E-19, E-20, E-21 and E-22). However, these deviations
may be attributed to the difference in surface runoff characteristics of the monitored
sewer districts. Silt was found to be the dominant factor in storm water discharges from
the Good Hope Run Sewer District.
Total Waste Load in Storm Water Runoff
The pollution constituents found in storm water runoff constitute inputs to the
environment or disruptions of the environment. These include:
1. Depositions on the surface from air pollution,
2. Depositions on highways from vehicular traffic or dust control activities,
3. Practiced depositions from such activities as plant fertilization or insect control,
4. Natural degradation for flora and fauna,
5. Soil erosion, and
6. Air pollutants washed from air during rainfall.
The waste load in a particular storms depends upon: 1) the amount of material which
has accumulated or developed on surfaces since the last storm and 2) the efficiency of
the washing action accomplished by the storm
If the storm waters are transported through combined sewers, materials which have
accumulated between storms may be flushed from the sewer. Any attempt to predict
waste loads which will be generated by a specific storm must be based upon knowledge
of the inputs to the drainage area and the washing efficiency that the storm will develop.
Waste loadings generated by storms in combined and in separated storm sewer districts
are presented in Table E-6. The total waste loading generated by each storm was computed
by integrating the waste-loading time relationship of each storm Since the total waste
loading carried by each storm could be a function of intensity of rainfall, quantity of
rainfall, variation of the surface runoff characteristics, duration of the dry-weather period
before the storm, and factors related to surface deposition, it would be very difficult
to derive any definitive correlation between any type of total waste loading and the
variables mentioned above. Examination of the data indicates that depositions from air
pollution are very important in waste generation, because the waste loading is appreciably
larger for a storm interval of even a few days (that is, larger than for a one-day or shorter
interval). The major source of the rather rapid accumulation of materials is logically from
air pollution; other sources are catch basins, inlet boxes, and street sweepings.
Two storms which provide a good comparison occurred on June 8 and 15, 1969 in
combined Sewer District G-4 (refer to Figures E-3 and E-4). Rainfall in each of the two
storms was the same (0.7 inches), the dry-weather period before each storm was about
161
-------
the same (6-7 days), and the storms were one week apart, which should eliminate any
seasonal differences. Also, the dry-weather flow should have been the same since both
storms occurred in the afternoon of the same day of the week. In addition to this, the
characteristics of the rainfall event preceding the June 8th storm were approximately the
same as those during the June 8th storm. However, the maximum intensities of the two
storms were different, being 4.0 inches per hour on June 8 and 2.5 inches per hour on
June 15. The more intense storm produced about twice the waste load, as measured by
chemical oxygen demand. Maximum storm intensity and waste loading appear to correlate,
which suggests that depositions to be carried away by the storm flow were equal and
that the different results were due to washing efficiency. A potential fallacy in this
reasoning is related to depositions on surfaces. Depositions which existed before the more
intense storm may have been twice as large. The number of days of dry weather is not
in itself an adequate criterion for estimating the deposition on a given area from air
pollution. Wind direction and velocity as related to pollution sources are potentially more
important.
Examination of the data in Table E-6 shows that there are general trends which are
influenced by the length of dry-weather periods, total rainfall, and rainfall intensity. The
variations of the total loadings of COD and suspended solids with total rainfall of each
storm are presented in Figures E-23 and E-24. The relationships were fitted to the data
using the method of least squares. Although the results tend to be inconclusive and there
is considerable data scattering, Figures E-23 and E-24 show a recognizable difference
between the separated and the combined sewer districts; however, they do not show the
consistent pattern that might have been expected for the two combined sewer districts.
Despite the inconsistencies in the available data, an attempt has been made to establish
a correlation between various parameters. The average correlation between the total COD
loading and total rainfall for Sewer District G-4 can be derived from Figure E-23, as follows:
L = 1,000 x e1-47F , with 0.5> F > 3.0
where L = total COD loading, in pounds per storm, and
F = total rainfall of the storms, in inches
Similarly an average suspended solids correlation can be derived:
S = 600 x e°-44F , with 0.5> F > 3.0
where S = total suspended solids loading, in pounds per storm
The variation of runoff with the progress of the storm showed a reasonably consistent
relationship to the change in rainfall intensity. Figures E-25, E-26, and E-27 depict typical
variations of wastewater depth in the sewer, rainfall intensity, and wastewater flow as
the storm progresses; these plots show the respective effects of a short intense storm,
a long intense storm, and consecutive storms in a combined sewer district. Rainfall and
runoff data were also compared to develop the coefficient used in the calculation of volume
of storm runoff.
In order to generalize the quality of both combined sewer overflow and separated storm
water discharge, the concentrations of different wastewater quality parameters were
compared and correlated. A comparison between BOD^ and COD is shown in Figure E-28,
162
-------
in which the relationships were fitted to the data by inspection. The ratios between average
BOD5 and average COD are 0.20, 0.14, and 0.11 for Sewer Districts G-4, B-4, and Good
Hope Run, respectively. This implies that the organic materials in separated storm water
discharge do not degrade as rapidly or as completely as the organic materials in combined
sewer overflow. The average ratio of 6005 to COD for the combined sewer overflow
from Sewer District B-4 falls between the ratio for combined sewer overflow from Sewer
District G-4 and the ratio for the separated storm water discharge from the Good Hope
Run District, possibly because the combined sewer overflow monitored at Sewer District
B-4 was more highly diluted by storm runoff than was the combined sewer overflow in
District G-4. As shown in Figure E-29, the correlation (developed by fitting the relationships
to the data by inspection) between volatile suspended solids and total suspended solids
averaged 0.50, 0.30, and 0.052, respectively, for G-4, B-4, and Good Hope Run. This
indicates that the organic proportion of the solids varied in much the same manner as
the BOD5-to-COD ratio for the wastewater collected at the different sewer districts. The
ratios of the volatile suspended solids to total suspended solids also indicate that most
of the solids contained in storm runoff are non-volatile silt and sand.
The results of the various comparisons and correlations indicate that there are definite
overall trends within these relationships; however, they are very loose. Although it is
believed that a relationship could be developed for surface characteristics and depositions
to relate to waste loading characteristics, sufficient data were not collected to define this
relationship.
Comparison of Flow Rate Measurements
Use of the steady-state formula and flow depth measurement has been the basic approach
of many engineers to estimation of the flow rate of storm runoff. This method may be
valid for the flow in streams or rivers, where flow variations generally are gradual. The
flows of interest in the current study, however, vary drastically with time, especially during
the surge period of a storm. Thus, a formula for calculation based on steady-state flow
was considered not applicable. The lithium dilution tracer method as a direct measurement
technique was selected, therefore, as the primary flow measurement method in this study.
Tracer dilution methods are recognized as reliable and practical flow measurement
techniques in both laboratory and industrial operations. Application of the lithium dilution
method to a variety of industrial wastewater stream flow measurements by ROY F.
WESTON has shown an accuracy of ±4 percent (6). A detailed outline of the lithium
chloride flow measurement technique is presented in Appendix D.
For purposes of comparison, flow estimations were made at Sewer District G-4 by assuming
steady-state conditions (Manning equation) and measuring the depth of flow. A plot of
the "depth" method versus the reference lithium chloride method is shown in Figure E-30.
Although a detailed statistical evaluation was not made, the data were observed to be
a general cluster around the 45° line of exact correlation. The maximum predicted gravity
flow using the Manning equation (n = 0.015, s = 0.0459 ft/ft, Diam = 5.5 feet) is
approximately 280,000 gpm.
It is felt that the variability in the flow rate estimated by depth measurements is due
to the lack of applicability of the steady state flow formula to the estimation of storm
runoff. However, the depth-of-flow measuring instrument may have malfunctioned (leak
in the line, insufficient air pressure etc.) even though this was not detected. Nevertheless,
163
-------
the tracer method may introduce some errors into the results unless feed rates, remote
operations, degree of mixing, and concentration of tracer during storage are carefully
checked or controlled. Besides this, the tracer method gives no indication of flow between
samples.
In order to verify the unit hydrograph approach to the estimation of total runoff discharge,
the flow data generated by the lithium chloride dilution method in District G-4 for various
monitored storms were compared with unit hydrographs constructed for the same storms.
The actual total discharge of each storm was determined by integrating the measured flow
rates with time, with peak rates taken directly from the field data. The unit hydrograph
methodology was discussed in a preceding section and in Appendix C. Table E-7 summarizes
the results of this comparison. For each storm compared, the total measured runoff was
within ten percent of the amount established from the corresponding unit hydrograph.
164
-------
Sampling
Date
Sept. 22, 1969
O)
01
Sept. 23, 1969
Schedule
Time
Sampling
Interval
minutes
10:50 a.m.
11:48 a.m.
12:46 p.m.
1:44 p.m.
2:42 p.m.
3:40 p.m.
4:38 p.m.
5:36 p.m.
6:34 p.m.
7:32 p.m.
8:30 p.m.
9:28 p.m.
10:26 p.m.
11:24 p.m.
12:22 a.m.
1:20 a.m.
2: 18 a.m.
3:16a.m.
4:14 a.m.
5:12 a.m.
6:10a.m.
7:08 a.m.
8:06 a.m.
9:04 a.m.
10:02 a.m.
11:00 a.m.
11:58 a.m.
58
58
58
58
58
58
58
58
58
58
58
58
58
58
58
58
58
58
58
58
58
58
58
58
58
58
58
Table E-1
Dry Weather Flow For Combined Sewer District G-4
Flow
gpm
1,000
1,000
940
955
900
882
940
985
1,000
1,015
985
955
940
711
872
604
542
516
477
438
457
542
780
1,156
940
956
928
£H
6.5
6.7
6.7
6.6
6.8
6.8
6.8
6.6
6.4
6.7
6.6
6.8
6.4
6.4
6.6
6.6
6.6
6.7
6.8
6.6
6.4
6.4
6.8
6.7
6.8
6.6
6.7
COD
mg/L
360
340
330
360
360
370
420
440
400
510
560
490
430
340
320
320
320
260
170
184
275
255
357
418
439
337
306
BOD
mg/L
204
119
....
108
111
128
126
135
117
171
192
183
153
117
188
—
84
71
49
54
58
86
103
91
168
123
105
Total
Solids
mg/L
680
540
610
570
830
540
600
650
550
680
660
720
680
580
650
610
610
740
470
550
780
540
570
610
620
590
540
Total
Volatile
Solids
mg/L
330
160
180
180
210
210
200
230
300
250
280
300
290
200
250
260
220
230
210
220
760
180
300
240
240
200
190
Suspended
Solids
mg/L
170
130
130
170
170
190
190
210
170
240
220
260
200
160
190
190
190
200
90
120
240
200
160
180
160
110
100
Volatile
Suspended
Solids
mg/L
130
130
120
156
170
182
182
190
168
230
178
230
154
160
174
168
190
200
88
120
20C
180
140
156
158
110
90
Settleable
Solids
mg/L
90
30
22
70
88
110
90
102
50
100
100
140
60
60
110
130
118
160
50
100
200
168
92
70
12
-------
Waste Constituents. ma/L
Chemical Oxygen Demand
Biochemical Oxygen Demand
Total Solids
Total Volatile Solids
Suspended Solids
Volatile Suspended Solids
Settleable Solids
Total Phosphate
Total Nitrogen
Orthophosphate
Ammonia Nitrogen
pH
Table E-2
Characteristics of Dry-Weather Flow
Combined Sewer District G-4
"Range A/lean
170-560
49 - 204
470 - 830
30 - 760
90 - 260
88 - 230
12-300
358
120
621
243
176
161
95
Separate Storm Sewer in
Good Hope Run
Sewer District
Range
10- 184
8- 58
456-1,950
158-1,780
20 - 260
0- 64
2- 180
Mean
69
18
639
345
68
28
47
6.4-6.8
6.6
6.6-6.1
6.35
Total Coliform, counts/100 ml
Fecal Coliform, counts/100 ml
Fecal Streptococcus, counts/100 ml
3,900,000
2,900,000
64,000
24,000
21,000
800
-------
Table E-3
Storm Sewer Dry-Weather Flow for Separated Sewer District - Good Hope Run
Location of Sampling Site - 17th St. - Minnesota Ave. S.E.
Drainage Area - 264 acres
Sampling
Date
Sept. 22, 1969
O)
Sept. 23, 1969
Schedule
Time
3:15 p.m.
4:15 p.m.
5:15 p.m.
6:15 p.m.
7:15 p.m.
8:15 p.m.
9:15 p.m.
10:15 p.m.
11:15 p.m.
12:15 a.m.
1:15 a.m.
2:15 a.m.
3:15 a.m.
4:15 a.m.
5:15 a.m.
6:15 a.m.
7:15 a.m.
8:15a.m.
9:15 a.m.
10:15 a.m.
12:15 p.m.
1:15 p.m.
2:15 p.m.
3:15 p.m.
Sampling
Interval
minutes
60
60
60
60
60
60
60
60
60
60
60
60
60
60
60
60
60
60
60
60
60
60
60
60
60
Flow
gpm
92.0
90.0
86.0
83.7
116.0
90.0
90.0
90.0
90.0
88.1
90.0
91.0
91.0
91.0
89.2
91.0
92.0
92.0
138.0
92.0
109.0
88.0
88.0
91.0
89.0
PH
6.5
6.4
6.4
6.4
6.4
6.4
6.4
6.3
6.4
6.4
6.4
6.4
6.4
6.5
6.2
6.2
6.2
6.1
6.2
6.2
6.2
6.4
6.4
6.4
6.6
COD
mg/L
41
82
51
10
61
42
71
61
61
51
82
82
82
51
82
51
51
82
184
102
71
71
102
102
BOD
mg/L
37
—
13
11
10
8
8
13
9
12
20
14
17
12
11
18
15
31
19
58
22
18
18
16
20
Total
Solids
mg/L
730
630
630
590
640
610
540
620
550
560
560
1,950
534
550
564
560
514
630
456
920
470
464
546
554
610
Total
Volatile
Solids
mg/L
240
300
310
230
270
180
190
280
290
280
320
1,780
264
334
314
308
192
270
232
764
250
158
296
308
258
Suspended
Solids
mg/L
170
60
60
70
260
130
50
80
200
90
52
60
68
52
60
28
30
98
30
46
40
20
32
36
30
Volatile
Suspended
Solids
mg/L
50
40
34
30
64
50
10
50
64
60
44
18
40
6
0
14
10
18
20
16
20
0
22
16
10
Settleable
Solids
mg/L
50
16
12
34
180
62
14
32
152
42
50
40
50
6
20
14
2
90
22
46
20
20
12
36
12
-------
Table E-4
Characteristics of Storm Runoff in Sewer District Good Hope Run
Location of Sampling Site • 17-Minn. and 16 S.E.
Storm
Total
Rainfall
inches
1.6
Date
Time
July 28
1:20-2:00 p.m.
July 28
5:00-5:30 p.m.
0.20
August 2
8:17-9:30 p.m.
2.9
August 9
9:20-9:45 p.m.
1.5
Sampling
Interval
minutes
5
5
5
5
5
5
5
5
5
5
5
5
10
10
10
10
35
10
10
10
30
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
40
10
10
10
20
Flow
gpm
21,000
65,600
75,000
57,900
47,700
43,300
57,900
15,400
12,500
10,200
7,900
6,090
4,570
5,000
3,740
3,620
4,770
2,020
2,625
2,190
3,180
1,140
2,020
2,500
4,040
2,640
3,600
4,200
3,090
1,640
2,020
34,400
16,800
10,100
5,660
4,400
3,520
2,470
1,960
1,995
1,601
1,410
1.340
2,205
2,030
50,000
16,150
7,500
5,000
2,860
_EH
6.2
6.2
6.1
6.0
6.1
6.0
6.0
6.2
6.2
6.3
6.2
6.3
6.4
6.3
6.6
6.8
6.9
7.0
6.3
7.0
7.0
7.1
7.0
6.9
7.0
7.1
7.1
7.1
7.2
7.1
6.3
6.3
6.2
6.2
6.5
6.6
6.8
6.9
7.0
7.0
7.0
7.1
6.3
6.3
6.1
6.2
6.1
6.2
6.2
COD
mg/L
430
400
280
170
310
300
370
240
230
210
210
230
150
120
140
120
58
48
67
77
67
29
58
96
77
77
48
106
86
77
38
400
259
216
140
184
119
108
140
129
65
86
54
180
200
160
390
270
180
150
BOD
mg/L
13
15
11
16
15
15
5
8
13
15
16
4
15
17
14
14
12
4
4
5
8
3
4
6
7
5
5
5
5
4
3
16
12
36
16
36
17
12
31
13
40
14
17
20
29
19
28
20
16
17
Total
Solids
mg/L
14,600
12,560
6,638
5,830
10,002
10,632
10,242
8,676
7,198
6,092
4,898
4,598
3,908
2,898
2,310
1,670
1,454
1,140
770
944
776
778
578
488
446
530
1,070
1,842
1,580
1,984
1,240
10,346
6,626
4,290
3,318
2,478
1,836
1,090
1,290
1,342
680
1,130
910
458
1,094
2,374
13,590
10,674
5,462
3,988
Total
Volatile
Solids
mg/L
912
996
278
268
600
484
512
488
460
390
288
378
284
228
200
110
136
136
138
136
76
142
90
90
154
96
120
136
84
12
106
538
368
250
226
188
130
40
184
178
164
200
72
190
272
126
826
604
248
260
Suspended
Solids
mg/L
9,600
11,200
6,050
5,520
9,020
10,010
9,170
8,150
5,560
5,900
4,620
3,920
3,140
2,160
1,920
1,020
1,160
640
480
720
480
520
380
320
300
340
920
1,500
1,300
1,740
980
9,568
6,560
4,210
2,610
1,200
1,550
1,278
910
840
200
548
416
100
1,000
2,020
11.28C
8,50C
5,100
3,400
Volatile
Suspended
Solids
mg/L
880
860
60
40
430
370
380
410
460
210
180
280
300
180
200
50
100
120
100
120
100
100
100
120
100
120
180
160
180
140
524
210
250
50
70
60
232
60
20
0
40
12
10
80
92
720
450
150
150
Settleable
Solids
mg/L
6,756
7,640
3,330
2,660
6,528
6,906
5,702
6,662
2,912
2,332
2,530
3,616
2,792
1,016
1,036
360
524
—
__
396
280
200
248
144
192
157
472
812
708
964
496
5,353
4,760
2,370
1,290
710
1,050
490
662
700
40
400
268
80
280
396
460
1,676
1,640
1,212
Total P
mg/L
4.5
2.8
1.5
1.8
2.4
2.0
2.6
1.6
1.8
2.2
2.0
1.6
1.6
1.5
1.4
2.1
1.0
1.0
0.5
1.0
1.0
0.4
0.3
0.4
0.4
1.8
1.0
0.8
0.4
0.2
0.?
2.0
1.8
1.5
1.0
1.0
1.0
1.0
1.0
1.4
0.4
0.2
0.4
1.4
1.0
1.0
1.4
1.0
1.0
0.6
Total N
mg/L
4.0
2.8
2.5
2.5
4.0
2.5
3.0
2.5
2.0
3.2
3.0
2.0
2.2
1.8
2.0
2.0
1.6
1.4
1.5
2.4
1.2
1.6
2.0
1.6
1.4
3.4
1.2
1.2
1.2
1.2
1.0
4.0
4.0
2.5
2.0
2.0
1.5
2.0
1.6
1.0
1.0
1.0
0.6
1.4
1.2
0.3
1.6
1.4
1.0
0.8
-------
Table E-5
Waste Loadings in Combined Sewer Overflow and
Separated Storm Water Discharge
05
CO
Waste Constituents
Chemical Oxygen Demand
Biochemical Oxygen Demand
Total Solids
Total Volatile Solids
Suspended Solids
Volatile Suspended Solids
Settleable Solids
Total Phosphate
Total Nitrogen
Combined Sewer Overflow
from Sewer District G-4
(Ibs./min.)
Range Mean
0.2-1,359
0.6
10.4
3.0
7.3
1.5
3.4
0.0
0.1
-298
- 2,552
-876
- 2,268
-652
- 1 ,996
- 38.75
-15.37
157.4
28.3
419.9
125 .4
322.7
85.8
165.7
1.82
1.56
Separated Storm Water
Discharge from Good Hope
Run Sewer District
(Ibs./min.)
Range Mean
0.2-219 19.0
<0.1-8.7 1.7
3.7 - 6,872 333.0
0.2 - 545 23.8
0.9-6,128 293.0
<0.1 -471 16.5
<0.1 -4,180 10.2
<0.1 -1.53 .09
<0.1 -1.5 .11
-------
Table E-6
Total Waste Loadings Generated by Different Storms
Storm
Date
1969
Combined Sewer District G-4
May 19
May 20
June 8
June 15
July 6
July 27
July 28
July 28
July 28
August 2
August 9
August 9
August 19
Combined Sewer District B-4
June 1
June 2
June 3
July 27
August 2
August 3
August 9
Separate Sewer District -
Good Hope Run
July 28
July 28
August 2
August 9
August 9
August 10
September 17
Time
Total
Rainfall
1:42- 1:45 a.m.
11:42-1 1:49 p.m.
5:50- 6:03 p.m.
2:20- 2:40 p.m.
7:40- 8:20 p.m.
11:35-12:14a.m.
2:30- 2:52 a.m.
11:30-1 1:45 a.m.
1:20- 2:00 p.m.
8:05- 9: 15 p.m.
9:20- 9:37 p.m.
11:20-1 1:30 p.m.
6:40- 6:53 p.m.
7:25- 7:45 p.m.
7:45- 8:05 p.m.
12:25-12:40 a.m.
11:35-1 1:55 p.m.
8:10- 9:00 p.m.
10:30-1 1:40 p.m.
11:22-1 1:37 p.m.
1:20- 2:00 p.m.
5:00- 5:30 p.m.
8:17- 9: 30 p.m.
9:20- 9:45 p.m.
11:00-1 1:30 p.m.
12:25-12:45 a.m.
8:20-10: 00 p.m.
0.4"
0.6"
0.7"
0.7"
0.4"
2.1"
0.6"
0.6"
1.3"
2.8"
1.1"
1.6"
1.35"
1.4"
0.9"
0.95"
1.3"
3.9"
0.4"
1.6"
1.6"
0.2"
2.9"
1.1"
0.8"
0.65"
0.6"
COD
5,423
155
2,474
3,749
193
178
1,906
BOD
Total
Total Volatile
Solids Solids
Total Loadings, pounds
Suspended
Solids
Volatile
Suspended
Solids
Settleable
Solids
Total
P
1 'oofl 0
1 ,9£U.U
2,890.0
1,415.0
838 0
Otju.vs
9,538.0
2,009.0
2,696.0
3,109.0
44,815.0
7,624.0
673.0
3,213.0
T/ \J
152
1 Oi
894
302
160
1,554
326
526
390
10,109
1,272
115
354
R ORQ
\jf\j*)*y
3,947
2,410
1,304
26^886
4,777
5,539
7,895
101,817
29,176
2,476
14,218
2 309
£.rtj\jy
2,250
1,130
637
10,951
2,075
2,071
1,379
24,010
10,136
736
3,569
4 294
^f AOt
2,521
1,328
1,060
20J01
3,187
3,844
6,431
75,657
18,374
2,061
11,667
l'fi34
1 ,UOt
1,390
473
561
4,877
974
1,200
1,506
14,399
4,635
289
1,089
*J/ I
fifi?
UU/
880
1,007
14,374
2,107
2,969
4,434
61,705
3,331
153
9,052
11.13
5.75
37.89
9.11
13.40
25.61
421.44
59.99
5.11
13.30
20.52
5.19
104.18
13.66
18.21
23.90
382.48
88.02
4.43
61.29
3,120.0
155.0
137.5
866.0
4,608.0
280.0
5,848.0
316
28
29
85
1,309
105
1,063
6,335
345
405
2,694
18,831
454
14,692
1,392
166
170
3,671
252
4,041
4,306
78
73
2,004
18,145
444
13,469
1,249
40
13
1,011
2,709
156
3,722
4,205
48
38
1,822
16,162
39
886
7.86
21.87
222.72
14.18
15.37
15.41
74.84
10.63
252
10
160
388
16
40
208
160,253
2,167
57,556
81,193
425
1,884
15980
9,415
215
3,180
4,709
76
359
1.841
142,958
1,732
55,000
68,150
272
1,737
143R3
6,621
268
2,604
3,679
3
56
1 9R7
90,845
924
30,503
9,021
148
433
sru-*
41.60
1.37
13.47
20.03
0.93
1.83
54.69
3.06
25.91
18.01
0.90
1.16
-------
Table E-7
Comparison of Actual Total Flow Measurement and
Volume Obtained by Hydrograph Analysis
Actual Measurement Volume From
Storm __ Volume Hydrograph Analysis
million gallons million gallons
Combined Sewer District G-4:
July 27-28, 1969 4.36 3.93
11:35p.m.-0:14a.m.
July 28,1969 2.43 3.22
1:28 p.m. 2:03 p.m.
August 2, 1969 21.10 20.20
8:05 p.m.-9:10 p.m.
July 6, 1969 0.12 0.19
7:40 p.m. - 8:20 p.m.
July 28, 1969 1.09 0.96
2:30 a.m.-2:52 a.m.
Combined Sewer District B-4:
August 2, 1969 4.40 5.12
8:10 p.m. -9:00 p.m.
171
-------
FIGURE E-1
DRY WEATHER FLOW IN COMBINED SEWER DISTRICT G-4
SEPTEMBER 22. 1969 110 50 >«) DRY REtTHER FLO*
.200 _
1.000 _
3 600 -
200 -
10 12 14 16 18 20 22 24
SEPTEMBER 22. 1969 (10:50 AN) OR* WEATHER FLOW
250
0246
10 12 i4 16 IB
TIDE .HOURS'
20 22 24 26 28
172
E-1
-------
FIGURE E-2
DRY WEATHER FLOW IN GOOD HOPE RUN SEPARATED STORM SEWER
SEPTEMBER 22-23,1969 (3:15 PM) DRI HEATHER FLOK
2,000-
i.aoo
1,600-
1.400-
~l,200-
\
~.000 -
400-
200-
0-
0246
10 12 U 16
TIME (HOURS)
20 22 24 26
SEPTEMBER 22-23, 1969 (3:15 PM) DRV HEATHER FLOW
160 -
140 -
120 '
IlOO
§
^80
360 '
40 '
20'
0
0246
10 12 14 16
TIME (HOURS)
20 22 24 26
173
E-2
-------
FIGURE E-3
SHORT INTENSE STORM IN COMBINED SEWER DISTRICT G-4
IIME I. IBM (9:50 M) 13 Ml. - 0.7"
20,000-
2,000-
10,000" j
i.ooo-
0-1
2.000
£ 1.000
10 20
TIDE (MINUTES)
JUNE 8, 1969 (550 W) 13 III. - 0.7"
1.25
174
E-3
-------
FIGURE E-4
LOW INTENSITY STORM IN COMBINED SEWER DISTRICT G-4
i
i
a
g
JUNE 15, 1969 (2:00 PI) 40 UN. 0.7'
2,000 -
.1,000 -
TIKE (MINUTES)
JUNE 15, 1969 (2:00 PI) 40 UN. - 0.7"
10.000 -
= 5,000 -
TIME (MINUTES)
175
E-4
-------
FIGURE E-5
SHORT INTENSE STORM IN COMBINED SEWER DISTRICT B-4
100 -i § i.ooo
50-
0.
180,000-1 600-1 1,500
150,000-
120,000-
;90,000-
60, 000-
30. 000-
0-1
500-
400-
300-
200-
100-
1,250
1,000
JUNE I, 1969 7:25 PM 22 MIN. 1.4"
10 15 20
TIME (MINUTES)
30
IUNE I, 1969 (7:25 PM) 22 MIN 1.4"
10 15 20
TIME (MINUTES)
25 30
176
E-5
-------
FIGURE E-6
LONG INTENSE STORM IN COMBINED SEWER DISTRICT G-4
600,
500-
400-
300-
i 200-
100-
660,000 GPM
HUGUST 2, 1969 (8: OS PN) 65MIN. -2.8
0-"- 0
40 50
TIME (MINUTES)
100 -
ll.OOO
5D-
p 500
0 J
AUCOST 2, 1969(8:05 PM)65MIK. 2.8"
10 20 30
40 50
TIME (MINUTES)
177
E-6
-------
FIGURE E-7
LONG INTENSE STORM IN COMBINED SEWER DISTRICT B-4
AUGUST 2. 1969 (B 10 PM) 50 HIM - 3.9"
200-
150-
100-
50-
0 -1
100 -
J5 -
50 -
25 -
10 15 20
TIKE (MINUTES)
1320,000
AUGUST 2, I9G9 (8:10 PH) 50 MIN - 3.9"
200 000-
I
100 000-
oj o
500
10 15
20 25
TIKE (HINUTES)
30 35 40 45 50
178
E-7
-------
FIGURE E-8
CONSECUTIVE STORMS IN COMBINED SEWER DISTRICT G-4
JULY 27-28, 1969
1.820
JULY 27.1969 (I!:35 PH) 39HIN.-2.1''
150,000 - I5QO
125,000 -
100.000 -
4
£ 75.000 -
50,000 -
25,000 -
20 30 40
TIME (MINUTES)
JULY 28, 1969 (2:30 AM) 22 DIN. 0.6"
500 -
f
250 -
ion,ooo-
50.000-
179
E-8
-------
FIGURE E-9
CONSECUTIVE STORMS IN COMBINED SEWER DISTRICT G-4
JULY 28, 1969
IULT 28, 1969 (11:30 AH) 15 DIN - 0.6"
100.000 -
i 50.000
JULY 28. 1969 (I 28 PN) 35 MIN 1.3"
180
E-9
-------
FIGURE E-10
SHORT INTENSE STORM IN SEPARATED SEWER DISTRICT
AUGUST 9. 1969 (9:20 PM) 25 MIN.-I.5"
12,000-1
10.000-
8 000-
6.000-
4 000"
2,000-
0-1
1,000-
600-
3 400-
200-
0 20
40 60
TIKE CHINUTES)
120
AUCUST 9. 1969 (9:20 PM) 25 MIN.-l
50,000-
40,000-
!=30,DOO-
20.000-
1.000-
aoo-
^ 600-
400-
10.000- 200-
0-1 0-1
15,000
12,500
TIME (MINUTES)
181
E-10
-------
FIGURE E-11
LONG INTENSE STORM IN SEPARATED SEWER DISTRICT
100 —i I 000 —i 10.000
- 500 -
30,000-
20.000.
Ci
g
10.000-
1.500.
1.000-
500-
15,000
10,000
5.000
AUGUST 2 1969 (8 17PM) I HO. 13 KIN -2.9"
34.400
TIME (HINUUi)
AUGUST 2. 1969 (6 17 PM) I Hft. 13 MIN 2.9"
100 120
TIME (MINUTES)
182
E-11
-------
FIGURE E-12
CONSECUTIVE STORMS IN SEPARATED SEWER DISTRICT
JULY 28, 1969 1:20 PM
10.000 -1 14.
50.000 10.000
I DO 120 HO
TIKE (MINUTES)
160 I BO 200
Kill 28. 1969 (1'20 PM) 40 MIN - I.I
.400-r 700 T 14
1.000
4
100 120
TIME (MINUTES)
140 160 180 200
183
E-12
-------
FIGURE E-13
CONSECUTIVE STORMS IN SEPARATED SEWER DISTRICT
JULY 28, 1969 5:00 PM
6.000
' 4.000
51.000
I"
i
- 20
IUIT 2!. 1969 (5 00 PI) 30 KINS - 0 20
40 50
TIKE I III NOTES)
IU11 26. 1969 (5 00 Pll) 30 MINS - 0 ZO' '
ft
^ 1 000
£600
40 SO
TIKE llimiTES)
184
E-13
-------
FIGURE E-14
REPRESENTATIVE BACTERIOLOGICAL DATA FOR COMBINED SEWER OVERFLOW
IN WASHINGTON, D.C.
62.500-
" 37.500
3 25.000
12.500
g
I 2.1
1UU 21. 1969 (11:30 I.II ) 15 IIH. 0 I
40,000
-30.000
10.000
20 25
TIME (MINUTES)
(11:35 P.I.) 39 DIN. - 2 I"
30 40
TIME (MIKUTES)
185
E-14
-------
FIGURE E-15
WASTE LOADING ASSOCIATED WITH SHORT INTENSE STORM IN
COMBINED SEWER DISTRICT G-4
24-, 240
20 -
15 -
10 15
TIME (MINUTES)
1UNE 8 1969 (5:50 PM) 13 KINS. - 0.7"
15 20
TIKE (MINUTES)
186
E-15
-------
FIGURE E-16
WASTE LOADING ASSOCIATED WITH LONG INTENSE STORM IN COMBINED SEWER DISTRICT G-4
»UGUST 2, 1969 (8:95 PH) 65 MINS - 2.9"
• 2,4
± 1.290
50
TIME (MINUTES)
AUGUST 2, 1999 (9:05 PM) 79 BINS. -2.8"
19 29 39
50 60
TI«E (MINUTES)
187
E-16
-------
FIGURE E-17
WASTE LOADING ASSOCIATED WITH CONSECUTIVE STORMS IN COMBINED SEWED DISTRICT G-4
JULY 27, 1969 . CUMULATIVE RAINFALL, FLOW RATE, COO
- 0 6 '
111 900
E-17
-------
FIGURE E-IS
WASTE LOADING ASSOCIATED WITH CONSECUTIVE STORMS IN COMBINED SEWER DISTRICT G 4
JULY 27, 1969 . SUSPENDED SOLIDS, TOTAL VOLATILE SOLIDS, VOLATILE SUSPENDED SOLIDS.
15 KINS - 0
E-18
-------
FIGURE E-19
WASTE LOADING ASSOCIATED WITH SHORT INTENSE STORM IN SEPARATED SEWER DISTRICT
45 BO 75
TIKE (IIIIIIITES)
105 120
2 0-, 35.000-, ,n
AUGUST 9. 1969 (9:10 P «.) <5 KIN -15
-20.000-
15 30 15 60 15 90 IDS liO
190
E-19
-------
FIGURE E-20
WASTE LOADING ASSOCIATED WITH LONG INTENSE STORM IN SEPARATED SEWER DISTRICT
28 000-. l«o_, I
34.400
• UGOST 2. 1963 (B I) P M I 73 SitN - 2 90
ID 20 30 40 50
1600 -i 160
•OCOST 2 1969 (8 17 P « II. HIM -i 31
0 50
TIME (HIHUTES)
I 0 !2D
191
E-20
-------
CO
ro
3 HI
3
FIGURE E-21
WASTE LOADING ASSOCIATED WITH CONSECUTIVE STORMS IN SEPARATED SEWER DISTRICT
JULY 28, 1969 - CUMULATIVE RAINFALL, FLOW RATE, COD
E-21
-------
40 HIHS - 1.6 INCHES
CO
CO
FIGURE E-22
WASTE LOADING ASSOCIATED WITH CONSECUTIVE STORMS IN SEPARATED SEWER DISTRICT
JULY 28, 1969 TOTAL SOLIDS; SUSPENDED SOLIDS, TOTAL VOLATILE SOLIDS, VOLATILE SUSPENDED SOLIDS
E-22
-------
FIGURE E-23
CORRELATION BETWEEN COD WASTE LOADING AND TOTAL RAINFALL
100. 000
70. 000
GO 000
50 000
.10.000
jO. 000
20. 000
10.000
7.000
6,000
5,000
4,000
3.000
2,000
1 .000
700
GOO
500
400
300
200
100
COMBINED SEWER DISTRICT B-4 (— •)
COMBINED SEWER DISTRICT G-4 ( •)
SEPARATE SEWER DISTRICT GOOD HOPE RUN (-
- -A)
A/
0.5
1.0
1.5 2.0 2.5
TOTAL RAINFALL ( INCHES)
3.0
3.5
4.0
194
-------
FIGURE E-24
CORRELATION BETWEEN SUSPENDED SOLIDS WASTE LOADING
AND TOTAL RAINFALL
200 000
100, 000
70 000
60,000
GO, 01)0
40, onn
30, 000
20, 000
1 10,000
o
1 —
co
\
£ 7.000
" 6,000
CO
- 5,000
— 1
4,000
a
LU
E 3,000
O_
CO
to
2, 000
1,000
700
600
500
400
300
20
100
•
'/
J*
jt
(
A
/
/
/
/
• / /
•/
k^
/ A
/
•
>
x^
/
r
A.
m
A
/
/
/
/
/
/
/
/
/
/
/ •
/
/
x
/
/
•
x^
A /
/
/
/
/
• /
f
t ^^
SUSPENDS
r n MR i H F r
COMBINE!
. SEPAR.AT
/
/
/
s
%/
/
/
X
X
s®
s
/
/
D SOLIDS
SEWER DISTRICT B-4 (— ^— • )
SEWER DISTRICT G-4 ( •)
SEWER DISTRICT GOOD HOPE RUN (
^r u
• -A)
I
s
0 0.5 .0 .5 2.0 2.5 3.0 3.5 4.0
TOTAL RAINFALL (INCHES)
E-24
-------
FIGURE E-25
REPRESENTATIVE RAINFALL AND RUNOFF MEASUREMENTS FOR
SHORT INTENSE STORM IN COMBINED SEWER DISTRICT G-4
(SQNVSnOHi Nl HdO) MOId 33VM3S Q]NI9W03
i r
T
•••• CBH/S3H3NI)
JO A1ISN31NI
(S3HONI) H3M3S Nl Hld30 MOU
196
E-25
-------
FIGURE E-26
REPRESENTATIVE RAINFALL AND RUNOFF MEASUREMENTS FOR
LONG INTENSE STORM IN COMBINED SEWER DISTRICT G-4
'(SQNVSfiOHi N! WdD) Mdld 33VM3S Q]NI8WOO
••••• Cm l S3H3NI) 11VJNIVU JO AHSN31NI
(S3HONI) H3M3S Nl Hid3Q M01J
197
E-26
-------
FIGURE E-27
REPRESENTATIVE RAINFALL AND RUNOFF MEASUREMENTS
FOR CONSECUTIVE STORMS IN COMBINED SEWER DISTRICT G-4
"***:
•
•••• —
•••• ('8H/S3H3ND 11VJNIVM dO AilSNBiNI
(S3HONI) H3M3S Nl Hid3fl MOU
1 1
.(Nd9) MOld 33VM3S Q3NI8W03
198
E-27
-------
6D
50
40
s
1 30
20
10
0
NOTE: LINES OF
FIGURE E-28
CORRELATION BETWEEN BOD AND COD
COMBINED SEWER DISTRICT B 4
•
X
J**^ •
0
600
500
400
-1
•
= 300
t
200
IDD
0
IDO
90
80
70
60
3 50
0
"~ 30
20
10
BEJ
m
'
'. «.. s
^
/
•
m
^
.
IOD 200 300 400
CID (HE'D
COMBINED SEVER DISTRICT G-4
'3r*P
•
#£•''
'>"
.
jX**""^ •
• .
'
/^
'•
^
^
•
^
•
200 4DO 600 BDO I.ODO 1,200 1,400 .GOD 1.
COD (MS/L)
SEPARATE STORM SEVER DISTRICT
»^^"
'
•
i***-
^^
:
"
^
/
*
!
•
^
X^
.
^X
^
'
^X
^
'/"
500
BOO
lm 2DO 300 400 500 600 100 300 900
COD (MC LI
,T FIT 199 E-28
-------
FIGURE E-29
CORRELATION RETWEEN SUSFENDfD SOLIDS AND VOLATILE SUSPENDED SOLIDS
COKBtltD UVCB DiSTIICT 8 t
n SEDER QlSIEiCT H •
| NOTE LINES OF BEST FIT
200
E-29
-------
FIGURE E-30
COMPARISON OF FLOW MEASUREMENT TECHNIQUES FOR COMBINED SEWER OVERFLOWS
LOW
, uuu
7on
RQfl
400
300
200
100
70
50
40
30
20
10
7
5
3
1
/
/
/
•
•
/
/
•
»A
V
•
/
i
^
>
*
L/
• —
—
i
S
•
9~
t
?
•m
/
/
/
/
/
-
(
y
*
j
2 3457 10 20 30 40 50 70 100 200 300 400 500 700 1,000
FLOW RATE MEASURED WITH LITHIUM TRACING (GPM IN THOUSANDS)
E-30
201
-------
APPENDIX F
ULTRA-HIGH-RATE FILTRATION
Introduction
High-rate filtration has been extensively applied for the removal of suspended impuritites
from raw water or wastewater, especially when the impurities are primarily non-volatile
discrete particulates, such as the wastewater from steel mills. In combined sewer overflow
and separated sewer discharge, large fractions of the suspended waste constituents are
recognized to be non-volatile discrete solids; thus, high-rate filtration may be an effective
treatment method. Despite the many studies that have been undertaken, the status of
filtration development is still in transition from an art to a science. The practical design
parameters for the filtration process to treat a specific wastewater must still be determined
from results of specific laboratory or pilot-scale investigations. Moreover, ultra-high-rates
of filtration (greater than 15 gpm per square foot) must be applied in order to cope
economically with the unique hydraulic characteristics of the combined sewer overflow
or separated sewer discharge-high discharges within a short-time period. This adds another
dimension of uncertainty to the development of a feasible filtration process to treat the
excess urban wastewater derived from intense storms. Therefore, a filtration study was
conducted with the following objectives:
1. To evaluate the applicability of ultra-high-rate filtration to the treatment of
combined sewer overflows,
2. To determine the flocculation effects of chemical additives on the solids and
organic material removal and
3. To provide a conceptual design basis for pilot-scale or full-scale treatment units.
The principal process variables evaluated in the laboratory program were:
1. Filter media, including type, depth, size and arrangement;
2. Filtration rates;
3. Effects of addition of flocculants and flocculant aids;
4. Variation of solids concentration in the wastewater;
5. Backwash rate and quantity;
6. Air-scouring rate, duration, and sequence in the backwash procedure;
7. Effluent quality characteristics, including suspended solids, COD, total five-day
BOD, and soluble five-day BOD. (Soluble BOD$ was measured by performing
BOD5 analyses upon the filtrate produced by vacuum filtration of the wastewater
sample passed through a diatomaceous earth filter);
8. Length of filter run; and
9. Head loss requirements.
Theoretical Background
The basis for this discussion of technical background was a comprehensive review of the
technical literature on removal mechanisms in the filtration process, analysis of filter
performance, and considerations for practical filter applications.
203
-------
Particle Removal Mechanisms
The possible mechanisms for the removal of particulate material from water by filtration
through porous media may be categorized into two types: physical and chemical removal
mechanisms.
Physical Removal
Physical removal mechanisms include: 1) straining, 2) sedimentation, 3) inertial
impingement and centrifugal collection, 4) Brownian movement, 5) physical contact caused
by the convergence of fluid streamlines, and 6) diffusion of particulate materials. They
are all dependent in varying degrees on different physical and operation variables such
as size of the media, size of the particulates, filtration rate, temperature, etc.
Straining: Straining (or direct sieving) has been considered as the primary mechanism to
remove suspended solids in the traditional sand-filtration process. The probability of
removal for suspended solids by straining, Ps, was defined as:
Ps~e>
in which D and d are, respectively, the diameters of the suspended particle and the sand.
Sedimentation: The ability of the void spaces in a rapid sand filter to act as settling
basins was first postulated by Hazen. Stanley found that the sand filter can be expected
to remove particulates which are one-twentieth the diameter of the particles removed by
a settling process at the same hydraulic loading. It has also been found that the removal
increases with the square of the particulate size and the difference in density between
the particle and the fluid. Increases in filtration rate and the fluid viscosity may impair
the particle removal rate. Particle growth by flocculation within the pores of a filter
increases the effectiveness of the straining and sedimentation removal mechanisms. The
mean velocity gradient affects the rate of flocculation in the filter and the total amount
of flocculation produced is proportional to the product of mean velocity gradient and
average detention time.
Inertial Impingement: As the suspension flows around the filter media with continual
changes of direction, the momentum associated with the inertial force may cause the
particles to impinge on the surface of the filter media. Chen describes the probability
of impingement for the air flow through a fibrous medium as:
pD2v
where:
PJ is the probability of impingement,
D is the diameter of the particle,
p is the density of the particle,
v is the particle velocity,
f is the fluid viscosity, and
d is the diameter of the filter medium fibers.
204
-------
Brownian Movement: This removal mechanism is not too significant in rapid sand filtration.
It would provide the contact between the particulates and the surface of the filter media,
but would have little effect for the particles larger than two microns, based on theoretical
calculations.
Physical Contact: Physical contact between the particles and the surface of the filter media
will be generated as a result of convergence of the fluid streamlines. Stein proposed the
following relationship:
P „ 21
PC d3
where:
Pc is the probability of removal of a suspended particle,
D is the diameter of the particle, and
d is the diameter of the filter medium grains.
Diffusion: The diffusion of suspended particles into the "dead spaces" of the filter medium
was proposed as a removal mechanism by Hunter and Alexander. The dead spaces are
those regions where the fluid flow is essentially zero. The colloidal particles diffuse across
the stream-line driven by the particle concentration gradient. Because of the tendency
to migrate to the regions of low shear, the colloidal particle concentration in the dead
space may become considerably higher than that in the ambient fluid.
Chemical Removal
Review of filter performance in the treatment of uncoagulated and coagulated suspensions
indicated that the physical removal mechanisms are inadequate to expain the entire
filtration process. Thus, chemical removal mechanisms were postulated. Basically, these
all stem from the electric charges on the surface of particles.
All the electrostatic charges on the surface of the particles in water are derived from
one or more of the following sources: 1} ionization of molecules at the particle surface,
2) imperfections of the crystal lattice, 3) direct chemical reaction with specific ions in
the water, and 4) physical adsorption of ions from the water solution. The principal
chemical removal mechanisms are discussed in the following paragraphs.
Electrokinetic Effects: From observations of suspension penetration in filter beds, different
investigators have concluded that the electrokinetic forces were the primary removal
mechanism for charged particles. Additional support for this conclusion was found from
filter bed examinations which showed the floe had no preference for horizontal surfaces
or pore interstices. Using positively pre-charged media. Hunter and Alexander were able
to improve the filter efficiency for negatively-charged clay particles. O'Melia and Crapps
found that the floe suspensions with zeta potential ranging from low negative to high
positive would allow little bed penetration in sand filters; they also suggested that rapid
sand filtration could be chemically controlled through the influence of specific chemicals
on the adsorption of particulates to the filter media.
205
-------
Van der Waal Forces: These are molecular cohesive forces between particles. The intensity
increases drastically as the particles approach each other. Between two atoms, these forces
are proportional to r^, where r is the distance between the two atoms. For large multi-atom
particles, the Van der Waal forces are proportional to r'3. Mackrles contended that these
attractive forces were not affected by the electrokinetic nature of the particles but
depended almost completely on their density.
Mathematical Relationships
Many different mathematical models have been developed to describe the behavior of
filtration through granular media for the removal of suspended solids from water. Iwasaki
first formulated the filtration phenomenon in terms of a first-order equation as:
ff=-Xc (F-D
where:
c is concentration of suspended solids,
I is depth of filter layer from top surface, and
X is impediment modulus, or filter coefficient.
Iwasaki suggested another relationship to express the removal of suspended solids as the
increase of deposition onto each layer:
3c _ I da
dl ~ ~ v 3t
where:
a is the specific deposit on the filter,
t is the filtration time, and
v is filtration rate.
These two equations have been widely used in filtration studies for many years. Although
other researchers have applied these equations to evaluate filtration behavior, no basic
modifications have been made.
The filter coefficient, X, varies with the filtration time, t, due to progressive clogging of
the filter pores, and may be defined by the specific deposit,a, as suggested by Iwasaki:
X = X0 + ka (F_3)
where:
X is initial filter coefficient, and
k° is constant.
206
-------
Ives later suggested another equation to describe the variation of the filter coefficient,
X, with the specific deposit, a, as
Po - a (F-4)
where:
P0 is the porosity of clean bed, and
0 is a filtration parameter.
In Inves1 formula, the increase of the interstitial velocity and the reduction of available
grain surface area near the end of filter run were considered.
Sholji generalized the Ives equation based on his theoretical and experimental analyses
and expressed the filter coefficient in terms of different operation variables as:
x~vaVv" (F-5)
where:
v is approach velocity of filtration,
do is geometric mean size of media,
M is dynamic viscosity of influent fluid, and
a1/t>1,ci, are constants.
However, Fox and Cleasby found that Ives' equation could not adequately describe the
behavior of filtration of hydrous ferric floe suspensions, but they did verify the initial
linear relationship ofX versus t. The failure to fit the experimental data with Ives' equation
could be attributed to the electric potential associated with the hydrous floe, which was
not considered in the derivation of the equation. Recently, Deb suggested a new set of
filtration equations which have incorporated the unsteady component of local variation
of suspension concentrations with time into the non-dimensional forms. Specifically,
1~ps 9a = xc (F-7)
c0(Po - a> 91
207
-------
where:
c
c = — , dimensionless concentration ratio (F-8)
~ co
\f
_t_ = — t, dimensionless time value (F-9)
I = — , dimensionless depth (F-10)
L
co is initial concentration of suspended solids,
L is total depth of filter bed, and
ps is porosity of deposited material.
The filtration behavior of unsized Fuller's earth could be defined satisfactorily. However,
the application of this model to the removal of flocculated material was not considered.
Considering the probabilistic nature of particle transport and attachment to a granular
medium, Hsiung proposed a new filter performance prediction theory based on a "Random
Walk" analogy. If P represents the probability of penetration of a unit of filter depth
by a particle, the probability of a particle moving down to depth I at time t will be:
The total number of particles deposited between I and l+ 1 can be represented by:
Am ~ P (I: vt, p) Qc0 At (F-12)
Then, from equation (F-2), the following can be derived:
|f- vAt ~P(I: vt,p) vc0At (F-13)
t ff~P(l:vt'P>
Equation (F-I3) indicates that the particle removal per unit depth of filter could be
characterized by a probability function. Based on the Chi-square distribution, Hsiung
proposed a deposit index, U, to translate filtration performance data for practical design.
This probability theory application permits description of the random nature of the
suspension, of the filter media, and of the transport and removal mechanisms (including
both physical and chemical mechanisms) in a relatively practical and simple formula.
208
-------
Considerations For Practical Application
Filtration efficiency is a function of many design variables: filtration rate, media size,
filter depth, and the properties of the suspension. With the use of the deposit index,
U, in the "random walk" analogy, Hsiumg also studied the relationship between different
design varables for high-rate filtration, and found that any increase in filtration rate must
be accompanied by a decrease of grain size in order to produce the same effluent quality.
On the other hand, he also pointed out that the grain size of the meia should be increased
in order to provide comparable filtration results for the same influent at a higher filtration
rate at the same allowable head loss. Two types of performance curves were proposed
for evaluation of various ranges of grain size, flow rate, and influent suspended solids
concentration; one based on required effluent quality and other on terminal head loss.
For the multi-media filter, Conley and Hsiumg showed the application of the random
walk analogy by using an equivalent grain size, de, as:
de = xi di + x2 d2 + x3 d3 (F-14)
where:
xt. x2, and x3 are the percentages by volume of individual
media, and
dj, d2, and d3 are the mean grain size of individual media,
respectively.
Hudson and Hsuing suggested the following relationship for head loss and filtration time:
L0.5t
L O O xj 1 Q \r~~IOj
where:
Ht is head loss at filtration time t, and
Ho is head loss at beginning of filtration
through a clean bed.
Thus, the head requirement will increase with filtration rate and initial suspension
concentration very rapidly. Accordinq to the relationship between the depth of filter, L,
and the lumped data of G (G = v0-2*3 d°-62t) it is suggested that, for the same filtration
efficiency and same length of filter run, the filter bed depth requirement increases with
the filtration rate.
Therefore, in order to facilitate ultra-high-rates of filtration, (v>15 gpm/sq.ft.), the filter
bed must be deep, the total head available must be high, and the size of the filter media
must not be too small. However, for an influent containing a high solids concentration,
such as combined sewage, the grain size of the media cannot afford to be too large either.
In various types of filter applications, the arrangement of media is still an art. Oeben
et al found that the reverse-graded filter will permit decreasing the head requirement or
increasing the filter run. Upflow with deep depth of sand has been applied to facilitate
209
-------
utilization of the entire filter bed. Mixed-media filters were reported as being able to
provide longer filter runs or better effluent quality than regular sand filters. Filter media
made of fiberglass were also reported to have potential application for ultra-high-rate
filtration when the suspended solids concentration is high in the influent.
The backwash after each filter run has gradually attracted the attention of various
researchers. In general, it was found that the air scouring would improve the backwash
efficiency and reduce the backwash water requirements. The air may be introduced at
the same time as the low-rate backwash water, but generally the air scouring is applied
in alternation with water at 3-5 cfm for 5 minutes every cycle.
For the treatment of steel mill industrial wastewater, ultra-high-rate filtration has been
applied. The filtration rate could be as high as 20 gpm/sq.ft. with influent solids
concentration of 50 mg/L. Some deep-bed pressure filter applications for steel mill
wastewater treatment reported similar results.
Description of Laboratory System
Three independent laboratory filtration systems, identical except for the filter media, were
constructed for this program. As shown on the schematic diagram (Figure F-1), the principal
components were the filter column, storage tank, transmission facilities, flocculant supply
system, and various controls and safeguards. Photographs of these principal components
are shown in Figure F-2.
Filter Column
A filter column was required with associated instrumentation to monitor and/or control
filtration rate, operating pressure, total head loss, and fluid temperature. The column had
to have adequate strength to withstand elevated pressures, adequate depth for deep bed
filtration, and easy disassembly for the purpose of changing or modifying the filter bed.
Initially, each system included a glass column as described below (a second column was
subsequently provided for the fiberglass system when a new form of fiberglass medium
was introduced). The column is a 9-foot, jointed, glass pipe with a four-inch inside diameter
(Figure F-1). Enclosed at the top by a one-inch thick PVC plate, the column will withstand
internal pressures up to 35 psig. The underdrain also consists of a one-inch thick PVC
plate, perforated with 30 to 40 evenly-distributed 1/2- and 1/4-inch holes. Mechanical
joints consisting of cast iron flanges and Teflon gaskets are located 60, 74, and 86 inches
above the underdrain, connecting the four sections of conical pipe. The underdrain is
attached to the bottom section of the column so that one or more sections may be removed
from the frame, retaining the media while making it accessible for purposes of replacement
or modification. The maximum available depth for a fluidized granular bed during backwash
is 92 inches.
The top section of the column is a cross tee, with 2-inch diameter branches. When the
column is operated as a downflow filter, the wastewater enters through one of the branches
and exits through the underdrain. The other branch of the cross accommodates the
backwash effluent, as well as filter effluent when the apparatus is used for upflow operation.
Midway through the study, when a new form of fiberglass medium was introduced, a
special filter column, constructed by Owens-Corning Fiberglas Corporation, was installed
210
-------
in the system to accommodate the new fiberglas-reinforced-plastic (FRP) cartridges. This
new column is constructed of plexiglass in five sections. Each end section contains a splash
plate, a perforated baffle and 1/2-inch pipe connections. The three intermediate sections
each can accommodate cartridges of 12 to 24 inches in depth. The total available depth
for the fiberglass medium is 60 inches, and no space is required for backwash expansion.
Taps located at each end of each section of the column permit multi-depth sampling and
measurement of head loss distribution with depth. Fabrication of the column is such that
it may be inserted in place of the glass column without disruption or major alteration
of the appurtenant piping. Consequently, the appurtenant piping and instrumentation are
the same for both types of columns.
Filtration rate is controlled by a constant-pressure differential flow controller located on
the effluent line and capable of maintaining a constant flow as long as there is a constant
pressure on the downstream side of the controller. Flow can be regulated and automatically
controlled between 0.4 and 4.4 gallons per minute, with a minimum pressure differential
of 6 psi across the controller. Filtration rate and backwash rate are measured by a rotameter
located in-line between the filter and the flow controller.
Pressure taps are located at the top of the filter and at the underdrain, and a precision
pressure gage is connected to either end (depending upon whether the filter is operated
downflow or upflow) to monitor influent pressure.
Head loss across the filter bed is measured directly with a mercury manometer having
an upper limit of 66 inches of mercury. The temperature of the wastewater in the glass
filter column is measured with a dial-type thermometer located 67 inches above the
underdrain in a tee-section of the filter column.
The filter bed is backwashed with tap water supplied by a connection below the underdrain.
Air scouring of the bed may be used in the backwash process. The air supply is also
connected to a tap located below the underdrain. Distribution of the air bubbles is
accomplished by the underdrain and by the supporting gravel layer of granular beds. Air
pressure may be regulated at any level to 35 psig. An air relief valve is connected to
the top of the filter to permit evacuation of accumulated air in the filter column prior
to bringing the filter up to operating pressure. Filtered water samples are collected from
the sampling valve downstream of the flow controller.
Storage Tank
Storage of sufficient capacity was required to provide for the maximum anticipated volume
of wastewater which the system may process in a single filter run.
The storage tank has a gross capacity of 1,120 gallons and is capable of sustaining a
filtration rate of 15 gpm per square foot for 12 hours, or a filtration rate of 50 gpm
per square foot for 3-1/2 hours. The fiberglass-lined wooden tank is 5 feet square and
7 feet deep, and is enclosed on top by wooden hatches and insulated on the sides and
bottom by 5 inches of fiberglass insulation. A thermostat-controlled 2,000-watt immersion
heater compensates for heat loss to maintain the desired water temperature. The wastewater
is continuously mixed by an electric, propeller-type mixer suspended over the center of
the tank.
211
-------
Transmission Facilities
Transmission facilities between the storage tank and the filter were required to deliver
the wastewater over the desired ranges of flow and pressure without materially affecting
its quality.
Wastewater is transmitted by a centrifugal pump from the storage tank through a 1/2-inch
PVC pipe to the filter. At an operating pressure of 35 psig at the filter, the pump delivers
more than 10 gpm. System pressure is controlled by a back-pressure relief valve located
on the pump throttle line. All pump discharge in excess of the filtration rate is returned
to the tank. The small (1/2-inch diameter) transmission line was chosen to minimize the
settling of solids without excessive head loss due to friction.
Flocculant Supply
Separate systems were required for the simultaneous injection of the flocculant and
flocculant aid in the transmission line so as to avoid contact between the chemicals before
mixing with the wastewater.
Each injection system consists of a metering pump which feeds a PVC injection nozzle
strategically located in the transmission line near the filter (Figure F-1). Mixing is achieved
by three mechanisms: turbulence induced by the projection of the nozzle into the flow
of the wastewater, turbulence occurring at three to six elbows between the injection point
and the filter, and longitudinal mixing in non-laminar pipe flow.
System Controls and Safeguards
Coordination and protection of the filtration system is provided by the devices listed below.
Safeguards
Pressure regulator ahead of filter.
Low water-level auto-shutoff for tank heater and mixer, with alarm and warning light.
Monitors
Tank water-level indicator.
Pressure gauges on filter and backwash air and water lines.
• Flow meters for air and water.
Thermometer on filter column.
Controls
Back-pressure relief valve on transmission line.
Flow controller.
Pressure Regulators for air and backwash water.
Theremostat on tank heater.
Purgemeter with needle valve for air scour.
212
-------
Discussion of Laboratory Procedures and Data
Outline of Laboratory Investigation
The purpose of the laboratory investigation was to evaluate the technical feasibility of
ultra-high-rate filtration for the treatment of combined sewer overflow. Minimum
performance levels could not be prescribed, because the state of the art would not support
such assumptions with any degree of confidence. Consequently, a more generalized
approach was applied to the investigation.
Technical feasibility depends primarily on three broad conditions: 1) filtration capacity,
2) effluent quality, 3) operating efficiency. These conditions were evaluated in terms of
filtration rate and duration of filter run; removal of 6005, COD, and suspended solids;
and quantity of filtrate required for backwashing. !
Three types of filters were operated at rates of 15 or more gallons per minute per square
foot, utilizing a simulated wastewater and both inorganic and polyelectrolyte flocculants.
Three filters, with different types of filter media and filter-bed arrangements, were evaluated
in this study:
1. A newly-developed fiberglass filter medium previously untested in application
to high-rate filtration of combined sewer overflow, provided by Owens-Corning
Fiberglas Corporation.
2. A deep-bed tri-media filter specially designed for this application, provided by
Neptune-Microfloc Corporation.
3. A garnet bed operated as an upflow filter.
Alum was used as the inorganic flocculant and Rohm & Haas Primafloc C5 as the cationic
polyelectrolyte flocculant.
Filter runs were also conducted with excess activated sludge added both to use as a
flocculant and to exploit its biosorptive capacity in an attempt to effect significant removal
of soluble BOD.
The study was conducted within a two-phase framework. The first phase consisted of
a series of runs to determine a suitable flocculant for each filter. Separate runs were made
with:
1. Plain wastewater (no flocculant dosage).
2. Wastewater plus alum.
3. Wastewater plus alum and Rohm & Haas Primafloc C5.
4. Wastewater plus activated sludge.
These runs were performed at the lower end of the range of filtration rates to be covered,
i.e., at 10 and 15 gallons per minute per square foot (gpm/sq.ft.).
The second phase of the study consisted of a series of runs designed to evaluate the
performance of each filter over various ranges of filtration rate and suspended solids
concentration. In each case, the flocculant used was the one found to be most suitable
213
-------
to each filter as determined in the first phase of the study. The use of air scouring and
agitation was evaluated as an aid to backwashing of clogged filters in an attempt to
economize on backwash requirements.
Laboratory System Design and Operation
Process and System Parameters
Wastewater characteristics which were controlled and/or observed were: size and
concentration of suspended solids, 6005 and COD concentrations, and temperature. The
operating variables were: filtration rate, pressure, backwash rate and quantity, air-scouring
rate and duration, and flocculant dosage. The design variables included the type, depth,
size, and arrangement of filter media.
Three identical independent filtration systems were constructed at laboratory scale, as
shown in the schematic flow diagram, Figure F-1. The description in the preceding section
of this appendix is typical of each system exclusive of the filter media.
Characteristics of Wastewater Feed
The combined sewer overflow used in this study was a simulated combined sewer
wastewater, composed of diluted raw domestic sewage and silt. A wastewater with
suspended solids and BOD5 concentrations of 400 mg/L and 40 mg/L, respectively, was
desired for the purpose of uniformity in the flocculant evaluation and for some other
comparisons. The intended procedure for synthesis of the waste was to dilute the domestic
sewage with fresh water to attain the desired 6005 concentration and then to add a
slurry of silt to the diluted sewage to bring the suspended solids concentration to the
desired level. However, the character of the domestic sewage was too variable to permit
determination of an appropriate dilution ratio; therefore, the wastewater formulation
became largely a matter of adjusting the suspended solids concentration.
Soil was oven-dried at 250°C and then pulverized. A slurry of the resulting silt was prepared
at a concentration of 100 grams per liter. After thorough mixing (at least 10 minutes)
the slurry was allowed to settle for one minute. The supernatnat was then used (at a
suspended solids concentration of 30,000 mg/L) to adjust the suspended solids
concentration of the simulated wastewater.
Sewage was delivered to the laboratory by truck from a local sewage treatment plant.
Several loads were anaerobic upon delivery, and on at least one occasion, the pH was
a high as 10 (pH affects flocculant action and the attachment of solids to the filter medium,
the fiberglass being especially sensitive to pH variation).
Large particles in the sewage were removed during the formulation of the wastewater,
to prevent subsequent clogging of pumps, transmission lines, regulators, and valves. Most
of this solid matter consisted of gravel and grease which was retained in a residual sludge
in the delivery truck. Attempts to flush this material from the tank truck before loading
with sewage at the treatment plant were unsuccessful. The problem was resolved by
pumping the first 50-100 gallons from the truck into 55-gallon drums. The sewage was
passed through a wire cloth-lined basket to remove the large but less dense solids. The
effect of this screening procedure was basically the same as that of a grit chamber except
that large buoyant solids were also removed.
214
-------
Substantially different suspended solids and BOD5 concentrations were encountered when
the three storage tanks were filled one at a time from the truck. This difference was
attributed to the density stratification of solids in the tank truck during transport. It
was therefore necessary to fill all three tanks simultaneously so as to obtain uniform
wastewater characteristics. This was achieved by pumping from the truck into a small
dosing tank which drained into all three storage tanks.
Filter Media
Fiberglass Filter-Two basic forms of fiberglass were utilized in this study: fiberglass plugs
and fiberglass-reinforced-plastic (FRP) cartridges. Several configurations of each form,
differing in depth, density, fiber diameter, fiber orientation, density stratification, and
combination with granular media, were tested.
The fiberglass plugs were cut from laminated fiberglass boards, with the fibers sharing
a common alignment (either parallel or perpendicular to the direction of fluid flow). Plugs
with parallel fiber orientation were 3" long and slightly more than 4" in diameter so
that they fit tightly inside the 4" glass column. Insertion of the plugs inside the column
required approximately 5 percent compression of the medium Plugs with perpendicular
fiber orientation were actually disc-shaped, having a 4" diameter and a 5/8" thickness.
The disc were used for one run only and were found to be unsuitable because of high
resistance to flow.
The first seven runs were performed on beds consisting of plugs in layers of two densitites.
The upper layer in all cases had a density of 5 Ibs. per cu.ft., although the layer thickness
and fiber diameter varied among the runs. The lower layer, also varying in thickness and
fiber orientation, was in all cases 10 Ibs. per cu.ft. The compositon of the bed for each
test run is presented in Table F-1.
Each of the first five runs was attended by a media-collapse phenomenon, wherein the
individual plugs were observed to collapse inward and to separate from the glass wall
of the filter column. At the termination of each run, the collapsed plugs were no longer
suitable for use, because backwashing did not restore the original structure of the plug.
Two subsequent runs were performed on beds 3" shallower than those of the earlier runs.
The collapse phenomenon was not evident in these runs, but recycling on the same medium
was not attempted. On several occasions (including Runs 6 and 7), the plugs were displaced
by the backwash water.
Runs 8 through 16 were performed on FRP cartridges, fiberglass modules consisting of
an inner core of fibers bonded to a rigid resinous casing. This form was introduced in
an attempt to circumvent the collapsing phenomenon observed with the fiberglass plugs
and to achieve more effective backwashing. Two densities of fiberglass medium were tested
in the FRP cartridge form.
Each cartridge had a 4"-square cross section. The 5 Ibs/cu.ft. cartridges were 24" deep,
and the denser medium was 10" deep.
Five cartridge configurations were tested. Three cycles (Runs 8 through 10) were made
on Configuration A, which consisted of 48" of 5 Ibs/cu.ft. material with .00150 fiber
diameter over 10" of 10 Ibs/cu.ft., .00050 material. Two attempts were made at running
on Configuration B, which consisted solely of a 10" depth of 10 Ibs/cu.ft. cartridge; both
215
-------
attempts were unsuccessful because of excessive head loss, and no effluent samples were
collected nor measurements taken. Cartridge Configuration C was the first dual-media filter
utilizing fiberglass. It consisted of 10" of 10 Ibs/cu.ft. material, 24" of 5 Ibs/cu.ft. (.00150)
material, and a cover layer of 6" of coarse and 6" of fine garnet. The garnet was chosen
for its high specific gravity to permit high backwash rates without necessitating special
measures for preventing loss of media. The filter performance was the same as Configuration
B and is not reported in the results of the study. In Configuration D, the garnet was
replaced by coarse anthracite, and the performance was more satisfactory (Run 11).
Configuration E was used in five cycles (Runs 12 through 16); the only distinction between
Configurations D and E was in the fiber diameter of the 5 Ibs/cu.ft. material. The latter
contained .00110 fibers, and the former contained .00150 diameter fibers. The anthracite
was contained in a 4"-diameter PVC pipe covered with a perforated plate to prevent loss
of media during backwash.
The collapse problem appears to have been solved by the use of the FRP cartridges,
although internal splitting of the medium may be another manifestation of the same basic
circumstance.
A summary of cartridge configurations for which runs are analyzed in this report appears
in Table F-2.
Tri-Media Filter-The tri-media filter consisted of an anthracite-sand-garnet bed on a 12"
gravel and coarse garnet base. The depth, grain sizes, and specific gravities of the various
constituents are listed in Table F-3.
The gravel and garnet were placed in 3" layers. After placing 25" of the intermediate-sand,
fine-garnet mixture, the bed was backwashed and 1" of media was syphoned from the
surface. Similarly, after placement of about 40" of anthracite, the bed was again
backwashed and 4" of media syphoned off the surface. After an initial trial run, which
lasted only 15 minutes, the anthracite depth was decreased to 30" by syphoning off the
top of the surface. After Run 5, the anthracite depth was increased to 36" in an attempt
to improve effluent quality. The depth was again decreased to 30" after Run 10, when
the bed was repacked after a failure of the glass column.
Dp-Flow Filter~The up-flow filter contained a 48" garnet bed on a 12" gravel, coarse-garnet
base identical to that of tri-media filter. The garnet bed consisted basically of two grain
sizes (0.707 mm and 1.19 mm) in equal volumes, except for the first run, when the
bed was composed of the finer grain size only.
Results of Filter Runs
Methods and Criteria of Evaluation-The effluent quality aspect of filter performance was
evaluated in terms of the percentage removal of 6005, COD, and suspended solids from
the raw wastewater; the effects of differences in concentrations of these materials in the
raw wastewater were also considered. Soluble 6005 was monitored for the purposes of
characterizing the wastewater and evaluating the usefulness of activated sludge as a soluble
6005 removal agent. Raw wastewater samples were collected from the pump throttle
lines at the storage tanks. Filtrate samples were taken from the sampling valve located
downstream from the flow controller. Streaming current and pH were measured on a
majority of the samples.
216
-------
The capacity evaluation was based upon filtration rate and length of filter run. The length
of run as reported herein does not necessarily refer to the actual duration of filter operation.
Time designated as "T^" is the time to the point at which head loss equalled 15 psi.
The Jength of run characterized by a breakthrough of suspended solids is designated by
"Tb" and is defined as the time to the midpoint of a breakthrough. The breakthrough
was taken as a significant upturn in the effluent concentration from a lower stable state.
This definition was preferred to the more traditional definition based on a percentage
increase in concentration, because it is more characteristic of the filter and is less of a
reflection of an imposed effluent criterion. The length of run is assumed to be the lesser
of T-|5 and Tj-,. Consequently, the average effluent concentration (and percent removal)
is computed from the analytical results of samples jtaken prior to breakthrough or head
loss in excess of 15 psi. Therefore, TI-, tends to be a conservative estimate of run time;
backwash requirements (based on actual backwash and filtrate volumes) also tend to be
conservative.
The backwash requirements were evaluated independently of the other parameters as the
search for an economic backwashing procedure was carried out. Backwash requirements
are expressed as percent of filtrate volume and as gallons of backwash water required
per pound of solids removed from the wastewater during the run.
Fiberglass Filter Flocculant Evaluation (Runs 1, 3, 5, 6)-The performance of the fiberglass
filter was not enhanced by the use of flocculants. Four filter runs at 15 gpm per sq.ft.
were compared in this evaluation (Figure F-3 and F-4); Run 1 was made with plain
wastewater (no flocculant was added); Run 3 with a dosage of 150 mg/L of alum injected
continuously in-line; Run 5 with Primafloc C5 (4 mg/L), a cationic polyelectrolyte, used
in addition to the alum; and Run 6 with 50 mg/L of activated sludge added to the
wastewater.
The effluent quality of the run with plain wastewater was superior to that of each of
the comparable runs. With the apparent exception of BOD5 removal in the alum-C5 run,
the removal of COD, 6605, and suspended solids was unexcelled by any of the runs
utilizing a flocculant. However, the influent BOD5 concentration for the plain run was
suspiciously low with respect to the influent COD concentration (Table F-4). While the
ratio of COD to BQD$ concentration for the plain 1.8 to 4.0, the ratio in Run 1 was
about 38. (In all of the subsequent runs the COD/BOD5 ratio was still below 5.5 except
in Run 11, where the ratio was 42.) Apparently there was an erroneous influent BOD^
measurement in the case of Run 1 and 11. If in fact the influent BOD5 in the plain
run were of the order of 100 mg/L (giving a COD to BOD5 ratio of 7.5), then the degree
of removal would amount to 95 percent. On the basis of effluent quality, the plain and
the alum-C5 runs were significantly more effective. Of the two, the plain run was considered
to be better because of greater COD removal and longer run time.
Breakthrough of suspended solids was observed in each of the four runs. The time to
breakthrough (TD) was nearly 85 percent longer in the case of the alum and the activated
sludge runs as compared to the 3-hour plain run. However, the 35 percent higher solids
and 140 percent higher COD of the influent of the plain run undoubtedly contributed
significantly to the shorter run time.
The apparent deficiency in run time is not considered to be of sufficient importance to
outweigh the substantially higher effluent quality obtained from the plain run. All
subsequent runs on the fiberglass filter were performed without benefit of flocculant or
flocculant aid.
217
-------
Efflect of Solids Concentration (Runs 8 and 14)-Two runs at 25 gpm/sq.ft., with
wastewaters having high and low suspended solids concentrations are depicted in Figure
F-5 in terms of head loss (Runs 8 and 14). The rate of increase in head loss was
approximately 7 times greater for the run with the higher solids. The non-soluble BOD$
was the same for each wastewater (approximately 40 mg/L), but the COD concentration
was higher for the low solids run. Therefore, the higher rate of head loss is attributable
to the higher solids concentration rather than to higher organic content.
Filtration Rate (Runs 15, 13, and 12)-Three filter runs at 15, 35, and 50 gpm per square
foot, show a sharply reduced run time at the two higher rates (see Table F-4 and Figure
F-6). The run time at the lower of the three rates was 2 hours (T-J5 = 122 minutes),
while at rates of 35 and 50 gpm per square foot, the run times dropped to less than
one-half hour (T-J5 = 28 and 25 minutes, respectively).
Effluent quality was lower at the higher filtration rates also, but to a lesser extent. COD
removal showed the greatest reduction of the three effluent parameters evaluated. The
removal of COD at 15 gpm per square foot averaged 78 percent, whereas removals at
the higher filtration rates were 52 percent and 64 percent. Suspended solids removal
dropped from 95 percent to 87-91 percent, and 6005 removal dropped from 72 percent
at 15 gpm/sq.ft. to 58-68 percent at 35 and 50 gpm/sq.ft.
A significant change in filter performance occurs between 15 and 35 gpm per square foot.
The performance levels at the two higher rates were very nearly the same, and the runs
were very short. It should be pointed out that all three runs were performed on the
same fiberglass cartridges, but in order of decreasing filtration rate. Thus, the shorter run
times for the higher rates cannot be attributed to solids retained in the bed from previous
runs.
Fiber Backwash-A unique feature of the fiberglass medium undergoing backwash is that
the medium remains unexpended. Consequently, the solids, wherever they may be lodged
in the bed, must be forced through the medium through which they have penetrated
during the filtration cycle. Unlike a fluidized granular bed, the fiberglass retains the ability
to trap solids during the backwash cycle and to impede their passage, thereby limiting
backwash efficiency. The solids, wherever they may be lodged in the bed, must be forced
back through the medium which they have penetrated during the filtration cycle. This
limitation may be partially offset by providing some mechanism for releasing solid particles
so entrapped. The mechanism chosen for study was air agitation. Air was discharged through
the bed, both alternating with and concurrent with backwash water, and the process was
repeated until the backwash effluent remained clear at high backwash rates. The alternating
discharge method was the more effective technique.
Due to the high porosity and compressibility of the fiberglass medium fiber reorientation
is possible. An extreme case of such distortion occurs when fiberglass plugs collapse during
filtration. The media collapse phenomenon, in all cases, was observed to occur in a
progression from the upstream plug to successive plugs in a downstream direction. The
plugs collapsed anisotropically inward, normal to the planes of laminations.
The probable mechanism by which collapsing occurs is basically that as the filter run
progresses, a layer of solids builds up near the upstream face of the plug or directly on
the upstream face, thereby creating a sharp pressure differential within the plug.
Simultaneously, as wastewater seeps between the plug and the glass wall of the column,
218
-------
the flow acquires a radial component, and another layer of solids builds up within the
plug, concentric to the column. The pressure differential then increases until failure occurs
in the direction of the weakest plane. This process recurs in each successive plug.
Backwash requirements of the plug-type medium ranged from 6.5 percent to 14.5 percent
of the filtrate volume for first-run beds, or 13 to 41 gallons per pound of solids retained
(Table F-4). Backwashing of a second-run bed required 41 percent of the filtrate, or nearly
90 gallons per pound of solids.
Backwashing of fiberglass plugs was not effective. Although the collapsed plugs partially
recovered their original shape after the run terminated, the basic weakness of the plugs
remained, probably due to broken fibers. Only one recycle was attempted on plugs. Runs
3 and 4 constituted the first and second runs on one set of plugs. However, the first
run was terminated on the basis of head loss before all the plugs had collapsed. When
the second cycle began, the weakened plugs again collapsed almost immediately, but the
remaining plugs were effective for 1-1/4 hours, at which time breakthrough occurred.
In addition to the problem of media collapse, the plug-form of fiberglass was shown to
be susceptible to dislodging during backwash. This usually occurred in the denser layer
of the bed, and all the plugs above it were displaced upward.
The use of FRP cartridges, to avoid the media collapse phenomenon and to enhance the
washability of the bed, was largely successful. The bonding of the outer fibers to the
casing apparently prevented the build-up of the forces which are presumed to have caused
the plugs to collapse, and the medium also was held firmly in place during backwashing.
In a few of the cartridges, however, internal splits, which were visible at the surface,
developed. These less extreme fiber reorientations were relatively small, the largest opening
being less than 1/4 inch wide and approximately one inch long. The depth was not
measured, but obviously did not extend through the entire depth of the cartridge. The
split appeared to be the result of a weakness in the fabrication of the medium
The backwash requirements for the cartridge were not very different from those for the
plugs, and they ranged from 3.6 percent to 22 percent of the filtrate volume, or 11 to
43 gallons per pound of solids retained (Table F-4). However, backwash requirements
decreased with successive cycles. For example, only 3.6 percent of the filtrate was required
to backwash the medium after its fifth filtration cycle. This indicates an increasing degree
of fiber reorientation and that increased quantities of backwash water flows through
channels or "corridors" of high permeability. The degree to which the agitating air bubbles
contribute to such channeling effects was not estimated, but the effectiveness of air
agitation in loosening bound solids was repeatedly demonstrated.
The effectiveness of air agitation is shown in Figures F-7 and F-8. Five cycles are depicted
in terms of head loss and the removal of suspended solids, 6005, and COD. The initial
head loss increased from 3.5 psi to 7.5 psi between Cycle 1 and Cycle 5, at a filtration
rate of 50 gpm per square foot. (The effect of subsequent cycles on the initial head
loss was lessened somewhat by the difference in influent suspended solids concentration,
since a higher concentration was associated with the earlier cycle.) The removals of
suspended solids, BOD^, and COD were not greatly affected by repeated cycles, although
a slight decline in percentage removal appears to be related to the difference in influent
concentrations.
219
-------
Tri-Media Filter Runs-Flocculant Evaluation (Runs 1, 3, 4, 5)-The combination of alum
and Primafloc C5 was found to be the most suitable of the flocculants tested in terms
of both effluent quality and length of filter run. The flocculant evaluation was based
upon four runs at 10 pgm/sq.ft.: Run 1 with plain synthetic wastewater. Run 3 with
an alum dosage of 150 mg/L, Run 4 with the same dosage of alum plus 4 mg/L of Primafloc
C5, and Run 5 with activated sludge at a concentration of 55 mg/L.
The alum-C5 run did not demonstrate the highest degree of removal of either suspended
solids, BODs, or COD, nor did any other flocculant exhibit consistent removal superiority.
As may be seen in Figure F-9 and Table F-5, the alum C5 combination was generally
the second best of the flocculants with respect to effluent quality. Suspended solids removal
in excess of 90 percent was achieved in the alum C5 run, and was second only to the
98 percent removal in the laum run. The BOD5 removal of nearly 60 percent for this
run was exceeded only in Run 1 (plain wastewater), for which removal averaged slightly
less than 70 percent; COD removal was 40 percent. The alum-C5 combination was
considered to be the best avialable alternative, largely on the basis of greater length of
filter run, which was 80 to 400 percent longer than the others. Head loss of 15 psi was
reached after 3-1/4 hours.
Effect of Suspended Solids Concentration (Runs 4, 6, 10)--The effect of influent suspended
solids concentration on filter performance was demonstrated by Runs 4 and 10 (Figure
F-10 and Table F-5), with influent solids concentrations of 410 mg/L and 2,420 mg/L
respectively, and COD and BOD5 concentrations approximately the same in both runs.
The run with the lower solids concentration (No. 4) was nearly 90 percent longer than
Run 10.
Run 6 also appears to show the effect of the organic content of the wastewater (Figure
F-10). The wastewater used in Run 6 had a lower suspended solids concentration than
either of the other runs, but had a much higher organic content-200 mg/L 6005 and
560 mg/L COD in Run 6, versus 25 mg/L 6005 and 160 mg/L COD in Runs 4 and
10. The effect of the higher organic content was a 50 percent shorter run time than
that in the high-solids run (Run 10).
Filtration Rate (Run 8)-Operation of the tri-media filter at rates in excess of 10 gpm/sq.ft.
proved to be unsatisfactory with respect to effluent quality and length of run. Although
performance was erratic at rates above 10 gpm/sq.ft., Figure F-11 illustrates a typical
performance for Run 8 at 20 gpm/sq.ft. This run lasted less than one hour, and the removal
of suspended solids, 6005, and COD declined typically as the run progressed. The average
suspended solids concentration in the filter effluent was 240 mg/L (50 percent removal),
while removal of COD and BOD5 each averaged approximately 30 percent.
Backwash Requirements-The normal backwash requirements of the tri-media filter were
12 percent to 25 percent of the filtrate, or 40 to 75 gallons per pound of solids retained
(Table F-5). However, frequently more backwash was required (up to 250 gallons per
pound of solids) because of two factors: 1) fibrous solids clinging to the surface of the
bed during backwash, and 2) difficulties in fluidizing the bed. The fibrous solids could
be removed only by expanding the bed to the overflow level. Failure of the bed to break
up readily resulted in the entire sand and anthracite layers being lifted without fluidizing.
The anthracite layer was particularly difficult to break up, and the problem was usually
caused by the accumulation of a layer of tiny air bubbles at the sand-anthracite interface.
220
-------
Upflow Filter Runs-The upflow filter was limited to a maximum filtration rate of 15
gpm per square foot, because the bed (a garnet medium) would fluidize at higher rates.
Nine runs were performed at 5, 10 and 15 gpm per square foot (Table F-6).
Flocculant Evaluation-The use of flocculants did not appear to have a beneficial effect
upon effluent quality at either 15 or 10 gpm per square foot. Run 1, with no flocculant
addition, was terminated because of excessive head loss (15 psi) after less than 1-1/4 hours;
only 60 percent removal of suspended solids was achieved.
Since a more rapid head-loss buildup was anticipated with a flocculant added, the garnet
bed was modified after the first run. Fifty percent of the fine garnet medium (by volume)
was replaced by a garnet sand of slightly larger grain size. This was effective in increasing
the length of run, and head loss did not again reach 15 psi in any run until much later
in the study when the bed depth was increased.
The second run (at 15 gpm per square foot) was made with an alum dosage of 150 mg/L.
Compared with the first run, the performance was only slightly better. The percent removal
of suspended solids was about the same, with the influent solids concentration about 40
percent lower and the effluent concentration correspondingly low. The 6005 removal
was about the same and the COD removal (75 percent) was double that of the first run.
Two subsequent runs at 15 gpm per square foot were made with different dosages of
alum and Primafloc C5 polyelectrolyte. In both cases, the removals of suspended solids,
6005, and COD were all less than one-half the corresponding removal in Run 1.
The consistently poor performance of the upflow filter at 15 gpm per square foot and
the ineffectiveness of flocculant addition seemed to indicate that a lower filtration rate
was appropriate. This was reinforced by the knowledge that at 15 gpm per square foot
the bed was on the verge of expansion. The flocculant evaluation was repeated at a filtration
rate of 10 gpm per square foot; in this series, a run with activated sludge as a flocculant
was also included.
In general, the filter performance did not improve at the lower filtration rate. Runs with
plain wastewater and with the combination of alum and Primafloc C5 were slightly better
at 10 gpm per square foot than at 15 gpm per square foot. However, the effluent suspended
solids concentration never dropped below 100 mg/L, nor did the removal exceed 80
percent. The maximum BOD^ and COD removals were 75 percent. A final run at 5 gpm
per square foot with the addition of alum was not substantially better in terms of effluent
quality.
Filter performance at 5, 10, and 15 gpm per square foot is shown in Figure F-12. The
runs at 15 and 5 gpm per square foot (Runs 2 and 9) were both made with an alum
dosage of 150 mg/L, and the run at 10 gpm per square foot was made with no flocculant
added. Runs 2 and 9 were characterized by steadily decreasing removal efficiencies; thus,
the average concentrations are strongly affected by the length of run. On the other hand,
Run 6, with plain wastewater, exhibited relatively constant removal efficiencies.
The reduced fluid velocity in the filter column above the garnet bed was not sufficient
to keep all of the solids in suspension in the filter effluent. Consequently, the solids
concentration continuously increased in this portion of the column. Filter effluent samples
taken from a valve downstream from the column were not truly representative of the
221
-------
filter effluent. Only when the solids concentration in the column reached such a level
that the overflow concentration equalled the filtrate concentration would the data be
representative. Thus, the effluent concentrations were actually higher than the data indicate.
Backwashing the upflow filter was a formidable task; almost always, the bed was so bound
up with solids within the gravel layer that the entire bed was lifted as a solid plug when
backwashing was attempted. The most successful technique was to backwash at a low
rate (less than 15 gpm per square foot) to remove fine solids from the upper portion
of the bed. Air was then introduced at a pressure of 25 psig, and a drain valve at the
bottom of the filter was opened with the filter under pressure, resulting in a very rapid
downflow. The high velocity downflow stripped accumulated solids from the lower portion
of the bed. This procedure generally had to be repeated six or more times before the
bed could be backwashed by the more normal procedure.
Backwash requirements generally amounted to 13 to 18 percent of the filtrate volume.
However, these figures are of limited value in that the filter runs were not terminated
due to head loss or turbidity breakthrough. Effluent quality was so consistently poor
that the filter runs were terminated at the convenience of the operator.
Filter Runs With Activated Sludge-One run was performed on each filter with activated
sludge as a flocculant and as a potential removal agent for soluble 6005. (Run 6 in
the fiberglass filter series. Run 5 in the tri-media series, and Run 5 in the upflow filter
series.) In no case was thaactivated sludge determined to be the most suitable flocculant,
nor did the sludge have a noticeably detrimental effect on effluent quality or length of
filter run (see Tables F-4, F-5, and F-6). No significant removal of soluble 8005 was
observed. In fact, the soluble BOD5 decreased by only 1 mg/L on the average for each
run with activated sludge. (Average reductions of zero to 30 mg/L were observed for
other runs, by comparison.
Summary of Findings
The laboratory test program indicated that ultra-high-rate filtration, at rates of 15 or more
gallons per minute per square foot, is a technically feasible process for the removal of
suspended solids and associated non-soluble BOD from combined sewer overflow. Of the
three filter systems tested, the fiberglass filters performed best, achieving at least 90 percent
removal of suspended solids and 70 percent removal of non-soluble BOD5 at filtration
rates of 15-30 gpm/sq.ft. and with filter runs of 1 to 3 hours duration. The addition
of flocculants and flocculant aids was not effective in improving the performance of the
fiberglass filters. Comparable effluent quality was not achieved in the tri-media filter runs
at filter rates above 10 gpm/sq.ft. Upflow filtration through a garnet bed was unsatisfactory,
largely on the basis of poor effluent quality.
Soluble BOD removal was negligible in all three filter systems, even with the addition
of activated sludge to the influent wastewater. The non-soluble organic content of the
influent wastewater appeared to have a greater impact on head-loss building than did the
suspended solids content.
222
-------
From the results of these observations, it can be concluded that an improved effluent
quality (i.e. lower concentrations of suspended solids, BOD^ and COD) could be obtained
from the fiberglass filter by:
1. increasing total bed depth
2. increasing media density in the bottom layer
3. optimizing density gradation
A multi- or graded-density fiberglass bed is needed to retain large solids and to permit
the passage of smaller solids so as to make the most efficient use of pore space and
avoid premature clogging at the shallower depths of the bed. Turbidity of effluent from
the fiberglass filter was due to very fine particulates, and it is believed that a bottom
density in excess of 15 Ibs/cu.ft. could reduce effluent concentrations of suspended solids
to less than 40 mg/L.
The economic feasibility of fiberglass filter process for ultra-high-rate filtration may depend
on extending the useful life of the fibergalss medium beyond the limits indicated by the
laboratory tests. Improvement of the backwash operations through modification of
underdrain design, staged removal of backwash effluent, the use of air scouring during
backwash, and development of improved fiberglass bed designs and fiberglass filter
regeneration techniques appeared to be promising approaches to extension of filter life.
223
-------
Table F-1
Characteristics of Fiberglass Plug Filter Beds
1
Upper Layer, 5 Ibs./cu.ft.
Run Layer Fiber
No. Thickness Diameter
inches
inches
Lower Layer, 10 Ibs./cu.ft.
Layer Fiber
Thickness Diameter
inches
inches
Total
Depth
inches
1
2
3-4
5
6
7
48
54
54
54
51
51
.0011
.0015
.0011
.0011
.0011
.0011
12
6
6
6
6
6
.0005
.00052
.0005
.0005
.0005
.0005
60
60
60
60
57
57
1
All fibers aligned parallel to flow unless otherwise noted.
-Fibers aligned perpendicular to flow.
224
-------
Table F-2
Summary of Filter Cartridge Configurations
Fiber Diameter
Configuration
Configuration A (Run Nos. 8-10)
Upper Layer
Lower Layer
Configuration D (Run No. 11)
Upper Layer
K> Middle Layer
ui Lower Layer
Configuration E (Run Nos. 12-16)
Upper Layer
Middle Layer
Lower Layer
Medium
— .
Anthracite
Fiberglass
Fiberglass
Anthracite
Fiberglass
Fiberglass
Depth
inches
48
48
18
24
10
18
24
10
or Grain Size
1 Clinches
150
50
coarse
150
50
coarse
110
50
Density
Ibs/cu.ft.
5
10
5
10
,5
10
-------
Table F-3
Characteristics of Media Used in the
Tri-Media Filters
Grain Specific
Material Layer Depth Size Gravity
inches
Anthracite 30-36 2.00-2.83 mm 1.6
Mixed: Sand 15 0.50-1.00 mm 2.6
Fine Garnet 9 0.35-1.00 mm 4.2
Coarse Garnet 3 1.41-4.00 mm 4.2
Fine Gravel 3 4.00-8.00 mm 2.6
Medium Gravel 3 5/16-5/8 inch 2.6
Coarse Gravel 3 1-2 inches 2.6
226
-------
Table F-4
Filtration Study Data Summary
Fiberglass Filter
ro
Run No.
Nominal Flow per Unit Area, gpm/sq.ft.
Flocculant Dosage
Alum, mg/L
Primafloc C5, mg/L
Activated Sludge, mg/L
Sutpended Solids
Influent, mg/L
Effluent, mg/L
Percent Removal
BOD5
Influent, mg/L
Effluent, mg/L
Percent Removal
Soluble BOD5
n u , "W
P B 1
COD
Influent, mg/L
Effluent. mg/L
Percent Removal
pH
Influent
Effluent
Length of Filter Run, minutes
Head Lots
Initial psi
Final, psi
Filtrate Volume (actual), gallons
Backwash (actual)
Volume, gallons
Volume as Percent of Filtrate.
Volume per Pound of Solids Retained, gal/lb.
Influent Pressure, psig
Actual Run Terminated By:
Cycle
Form
— 1
15
0
0
0
670
10
98
19
5
72
7SO
68
91
6.8
7.0
187b
1.4
7.9
299.4
21.1
7.0
13.0
24
Media-
Collapse
Plugs
_2
15
0
0
0
465
51
89
58
32
45
214
118
45
6.7
6.8
149b
1.4
8.2
248.8
27
Media-
Collapse
and
Head Loss
Plugs
_3
15
ISO
0
0
500
60
88
97
34
65
288
112
«1
7.1
6.8
**b.K
0.9
15-0
463.5
30.1
6.5
17.9
25
Media-
Collapse
and
Head Loo
1st of 2
Plugs
_4
15
150
0
0
632
32
95
62
14
78
250
58
77
7.0
6.9
75b
1.7
9.1
173.7
71.2
41.0
88.4
25
Media-
Collapse
and
Head Loss
2nd of 2
Plu«s
_5
15
150
4
0
580
6
99
124
6
95
18
6
65
220
44
80
6.8
6.9
"«b
0.4
8.4
217.2
14.2
6.6
15.2
25
Media-
Collapse,
Head Lots,
and
Turbidity
Plugs
_6
15
0
0
50
506
73
84
110
23
77
12
11
10
320
144
55
6.8
6.8
330,,
0.5
9.3
510.9
73.8
14 5
40.8
25
Turbidity
{no collapse)
Plugs
_7
25
0
0
0
232
46
80
87
58
53
32
30
3
296
148
50
7.2
7.2
210,
0.8
6.9
590.0
.........
24
Shortage
of
Synthetic
Wntewanr
(DO collapse)
Plugs
J
25
0
0
0
176
26
85
105
85
19
67
54
19
350
178
49
6.8
6.8
220
2.2
15.0
544.6
21
Head Loss
1st of 3
FRP
J)
35
0
0
0
560
56
90
98
61
37
145)
(45)
0
410
287
30
6.9
7.0
28
5.3
15.0
160.3
25
Head Loss
2nd of 3
FRP
JIO
25
0
0
0
242
24
90
95
58
39
43
(43)
0
400
208
48
6.8
6.7
62
5.4
15.0
234.2
24
Head Loss
3rd of 3
FRP
VI
30
0
0
0
680
48
93
4
1
71
0
0
170
60
65
7.4
7.6
'So
2.1
9.3
234.9
51.2
21 8
43.2
23
Head Loss
FRP
!2
50
0
0
0
1,000
90
91
86
36
58
20
15
25
343
123
64
7.0
7.1
25
3.4
15.0
126.1
15.1
11.9
IB
Turbidity
and
Head Loss
1st of 5
FRP
13
35
0
0
0
760
99
87
48
15
68
13
1 1
B
212
102
52
7.2
7.3
28
3.5
15.0
1D2.3
9.2
9.0
16.3
21
Turbidity
and
Head Loss
2nd of 5
FRP
H
25
0
0
0
840
42
95
45
10
77
8
7
11
243
61
75
6.9
7.1
52
1.9
15.0
183.8
1X3
7 j
iO.9
22
Head Loss
3rd of S
FRP
16
15
0
0
0
888
44
95
96
27
72
1 1
18
435
78
82
7.1
7.1
122
0.4
15.0
201.2
15.9
79
28
Head Lost
4th of 5
FRP
S
334
40
B8
22
9
58
5
5
91
46
49
7.0
7.0
52
7.4
15.0
350.6
12.7
3.6
14.7
19
Head Loss
SThofS
FftP
b w b 15 = Breakthrough of Suspended Solids.
-------
Table F 5
Filtration Study Data Summary
Tri-M«di» Filter
NJ
Run No,
Nominal Flow per Unit Am, gpm/eq.ft.
Flocculenl Dotage
Alum, mg/L
Primafloc C5. mg/L
Activated Sludge. mg/L
Suspended Solids
Influent. mg/L
Effluent, mg/L
Percent R«movil
B005
Influent. mg/L
Effluent, mg/L
Percent Remove!
Soluble BOD6
Influent, mg/L
EHIuent, mo/L
Percent Removel
COD
Influent, mg/L
Effluent, mg/L
Percent Removel
pH
Influent
Effluent
1
10
0
0
0
390
23
94
31
10
68
110
6.8
6.8
_2
15
75
0
0
440
42
90
47
23
52
137
70
49
6.7
6.9
_3
10
150
0
0
418
8
98
40
18
54
._._...
—
108
77
29
7.0
7.0
4
10
150
4
0
410
29
93
25
11
57
107
65
39
6.9
6.9
_B
10
0
0
55
475
101
79
66
37
44
19
16
6
339
207
39
6.8
6.9
_6
10
150
4
0
218
87
60
196
137
30
103
(103)
0
560
431
23
6.8
6.9
-1
15
150
4
0
158
96
39
159
102
36
110
(110)
0
670
456
20
7.0
7.0
_a
20
150
4
0
528
243
54
89
60
33
69
46
33
360
256
29
6.7
6.7
_9
6
150
4
0
188
113
40
111
58
48
50
20
60
340
156
54
7.0
6.7
JLQ
10
150
4
0
2,420
436
82
19
8
59
1
3
19
162
32
80
7.1
7.2
11
10
0
0
0
620
112
82
5
1
83
1
0
100
120
80
37
7.9
7.8
12
20
150
4
0
640
173
68
29
16
48
11
6
41
172
69
60
7.0
6.5
11
16
150
4
0
640
90
86
42
8
80
9
5
48
2S3
56
78
7.0
6.4
J4.
10
160
4
0
720
161
79
65
4
94
IS
4
76
232
51
78
7.0
6.0
Jt
20
0
4
0
334
37
89
22
9
57
6
4
20
91
36
60
7.0
7.0
Length of Filter Run, minutes
Heed Lou
Initial. p«i
Final, pv
Filtrate Volume (actual), gallon!
BackwB* (actual)
Volume, gallons
Volume as Percent of Filtrate
Volume per Pound of Solids Retained, gal./lb.
Actual Run Terminated By:
46
1.5
15.0
51.8
13.8
26.6
87
42
2.3
1S.O
67.6
12.0
17.8
54
75
1.4
15.0
87.2
21.9
25.2
73
198
1.4
15.0
?08.3
26.0
12.5
40
111
1.3
9.1
91.1
20.8
22.8
54
1.3
15.0
61.6
17.5
28.4
248
38
1.6
15.0
57
3.1
15.0
123.2
16.5
13.4
61
120
0.7
3.0
61.2
11.2
177
105b
1.4
13.2
102.0
1.5
3.3
102.0
15.3
15.0
37
2.7
15.0
33.1
8.2
24.9
81
13
2.2
15.0
17.2
6.6
32.0
71
21
1.6
16.0
22.0
7.4
33.5
71
24
4.4
16.0
61.7
9.1
17.6
71
Heed Loss Head Loss Head Loss Head Loss Turbidity Head Loss Head Loss Heed LOB Turbidity Turbidity Turbidity Head Loet Head Lose Hied Lou Heed Lot*
D • Breakthrough of Suspended Solids.
-------
FIGURE F-l
SCHEMATIC DIAGRAM OF FILTRATION SYSTEM
WASTEWATER
STORAGE TANK
FILTER
COLUMN
PRECISION
PRESSURE GUAGE
BACKWASH WASTE OR UPFLW FILTRATE
THERMOSTAT
Ni
CO
TEMP.
SENSOR
MSTEMTER
LEVEL
INDICATOR
(lANMETER)
On
PURGEMETEfl
PRESSURE
DIFFERENTIAL
REGULATOR
LIQUID LINE
AIR LINE
ELECTRICAL LINE
F-l
-------
FIGURE F-2
FIBERGLASS FILTER: FLOCCULANT EVALUATION ( RUNS 1,3,5,6
-HEAD LOSS AND SUSPENDED SOLIDS REMOVAL
t HEAD LOSS
FILTRATION TIME. MINUTES
(FILTRATION RATE = 15 GPH SF)
SUSPENDED SOLIDS REMOVAL
90
• PLAIN
A ALUM (150 MG Ll + PRIMAFLOC C5 (4 MG L)
• ALUM (150 MG L)
# ACTIVATED SLUDGE (50 MG Li
120
150
160
210
240
270
300
330
360
390
FILTRATION TIME MINUTES
(FILTRATION RATE - 15 CPM SFI
230
F-2
-------
FIGURE F-3
FIBERGLASS FILTER: FLOCCULANT EVALUATION ( RUNS 1,3,5,6 )
BOD AND COD REMOVALS
C. BOD5 REMOVAL
« PLAIN
A ALUM (150 MG/L) + PRIMAFLOC C5 (4 MG L)
ALUM (150 MG'L)
* ACTUATED SLUDGE (50 MG L)
FILTRATION TIME, MINUTES
(FILTRATION RATE = I5GPM SF)
0. COD REMOVAL
A ALUM (150 MG.-L) 4- PRIMAFLOC C5 M MG L)
• ALUU (150 MG/L)
* ACTIVATED SLUDGE (50 MG L)
150 180 210
FILTRATION TIME. MINUTES
(FILTRATION RATE 15 GPM SF)
231
F-3
-------
FIGURE F-4
FIBERGLASS FILTER: EFFECT OF INFLUENT SOLIDS CONCENTRATION
ON HEAD REQUIREMENT (RUNS 8 , 14)
So =INFLUENT SUSPENDED
SOLIDS CONCENTRATION
So =840 HG/L
So =180 MG/L
FILTRATION TIME, MINUTES
(FILRATION RATE 25 GPM/SQ. FT.)
F-4
232
-------
FIGURE F-5
FIBERGLASS FILTER: PERFORMANCE AT 15-50 GPM/SQ. FT.
PERFORMANCE AT 50 GPM/SQ. FT. (RUN NO. 12)
PERFORMANCE AT 35 GPM/SQ. FT. (RUN NO. 13|
INFLUENT QUALITY:
• SUSPENDEU SOLIUS - 1.000 MG;L
A 6005-a6 MG L
*COD-340 MG/L
100
0 30
FILTRATION TIME. MINUTES
INFLUENT QUALITY:
• SUSPENDED SOLIDS =160 MG L
A BOD5 = 48 MG/L
# COD =210 MG/L
16
15
H
0 30
FILTRATION TIME. MINUTES
PERFORMANCE AT 15 GPM/SQ. FT. (RUN NO. 15)
100
'50
-40
HEAD
LOSS
INFLUENT DUALITY:
• SUSPENDED SOLIDS = 890 MG/L
A BOD5 = 96 MG/L
* COD = 435
60 9D 120
FILTRATION TIME, MINUTES
150
233
F-5
-------
FIGURE F-6
FIBERGLASS FILTER: FIVE CYCLES ON FRP CARTRIDGES
HEAD LOSS AND SUSPENDED SOLIDS REMOVAL
A HEAD LOSS
CICLC
IMC.CPI.'SF
I15 milUIES
15
CO
0.
2
o
1
50
25
i
/
/
/ ,
i
2
35
21
/
/
/
/
/
3
25
52
i
/
\
/
f
1
1
/
'
/
4
15
122
/
/
/
/
/
/
f
/
/
/
/
/
'
5
50
50
i
^
/
/
|/
^
/
/
0 30 60 90 120 150 160 210 240 270 300 330 360
FILTRATION TIME, MINUTES
B SUSPENDED SOLIDS REMOVAL
CYCLE
(UTE.GPH/SF
INFL.CQNC.tHGA
100
BO
BO
70
^60
UJ
«50
a
u
S40
30
20
10
0
\
50
1000
A
J
2
35
760
n
\
3
25
B40
4
15
888
/
f
5
50
334
_
^^^
-•— -1
0 30 BO 90 120 150 160 210 240 270 300 330
FILTRATION TINE. MINUTES
234
360
F-6
-------
FIGURE F-7
FIBERGLASS FILTER: FIVE CYCLES ON FRP CARTRIDGES
BOD5 AND COD REMOVALS
C. B005 REMOVAL
CYCLE
RATE. GPM/SF
INFL. CONC. CMG/L)
100
90
80
70
i 60
j
" 50
UJ
CJ
LL4
=-40
30
20
10
0
1
50
86
\
N;
2
35
48
X,
3
25
45
^
^
4
15
96
k
A. — ,
0 30 60 90 120 ISO 180 210 240 2
5
50
22
k
v
X,
^^
,-
0 300 330 36
FILTRATION TIME, MINUTES
D COD REMOVAL
CYCLE
RftTE . GPM/SF
NFL CONC. (KG/LI
100
90
80
70
LLJ
^ 50
LJJ
CJ
§ 40
30
20
10
0
1
50
343
.X
2
35
212
1
3
25
243
1
^
^
*+*
4
15
435
•—
1
-"^-~- .
1 u
— * '
i
•^
5
50
31
1
\
/
V
/
/
30 60 90 120 150 180 210 240 270 300 330 3
FILTRATION TIME, MINUTES
235
F-7
-------
FIGURE F-8
TRI-MEDIA FILTER: FLOCCULANT EVALUATION { RUNS 1,3,4,5 )
> mi LOSS
SumiDED SOLIDS IEIOIH
• Pill I
• ILUI (ISO IC/L)
A ILDI (ISO IC/L) < PIIIIFLOC
CS (4 IE/L)
ICTIIITED ILUOGE (55 IC/L)
60 BO IZO ISO
FIITR«TIO« HIE. HIIIIITES
(FUTRITIOI RITE " 10 CPI/SF)
• PLUM
• ILUI (150 1C L>
A HUH (ISO 11! L) 4 PHIimOC
CS (4 11! L)
* ICIIVIIED SLUDCE (Si 11! L)
00 120 ISO
FILTRATION TIKE. HIKUTES
(FILTRITIOK RITE = 10 CPU SF)
100
C BODj REIDVU
• PLAII
• ILUI (ISO IC/L)
A ILUI (ISO II/L) 4 milFLOC
CS (4 1C L)
ICTIIITEO SLUDCE (SS 1C LI
FILTRITIO» THE. IHUTES
(FILTRITIOI I1TE = ID GPI SF)
236
F-8
-------
FIGURE F-9
TRI-MEDIA FILTER: EFFECT OF INFLUENT SOLIDS CONCENTRATION
ON HEAD REQUIREMENT (RUN 4, 6, 10)
20
So - INFLUENT SUSPENDED SOLIDS
CONCENTRATION
So = 220 MG/L
So= 410 IG/L
So - 2,420 MG/L
FILTRATION TIME,MINUTES
(FILTRATION RATE = 10 GPM/SF)
237
F-9
-------
FIGURE F-10
TRI-MEDIA FILTER: PERFORMANCE AT 20 GPM/5Q. FT. (RUN 8)
INFLUENT QUALITY:
• SUSPENDED SOLIDS = 530 MG/L
ABOD5 = 89 MG/L
#COD =360 MG/L
FILTRATION TIME, MINUTES
238
F-10
-------
FIGURE F-ll
UPFLOW FILTER: PERFORMANCE AT 5,10,AND 15 GPM/SQ.FT.
PERFORMANCE AT 15 GPM/ SQ. FT. (RUN 2)
PERFORMANCE AT 10 GPM/SQ. FT. (RUN 6)
INFLUENT QUALITY:
I SUSPENDED SOLIDS = 420 MG/L
;=4I MG/L
* COD =160 H/L
100
FILTRATION TIME, MINUTES
12°
= 50
i40
30
20
INFLUENT QUALITY:
• SUSPENDED SOLIDS =560 MG/L
A BOD5 = 94 MG/L
4i COD =260 MG/L
30
- I :
90 120 150 180
FILTRATION TIME, MINUTES
210
240
PERFORMANCE AT 5 GPM/SQ. FT. (RUN 9]
100
90
80
70
60
i
£ so
1 40
30
20
10
\
INFLUENT QUALITY:
• SUSPENDED SOLIDS = 870 MG/L
A BOD5 = B7 MG/L
* COD = 330 MG/L
30 60 90 120
FILTRATION TIME, MINUTES
239
f"
-------
FIGURE F.I2
FILTRATION SYSTEM COMPONENTS
WASTE STORAGE TANK
WASTE FEED PUMPS
SAFEGUARD DEVICES
FILTER COLUMN
240
F-12
-------
APPENDIX G
KINGMAN LAKE PROJECT
Summary
This conceptual engineering study concerns the reclamation of combined sewer overflows
and utilization of the reclaimed waters in a major water-oriented recreational facility for
the District of Columbia. The investigation encompasses a comprehensive solution of
environmental problems by proposing multi-use objectives and facilities.
Principal objectives of the project included: 1) evaluation of rainfall runoff relationships
for sizing of storage and treatment plant capacities; 2) confirmation of treatment feasibility
using filtration and an activated carbon process; and 3) development of sufficient data
for preliminary design purposes.
Laboratory studies not only demonstrated process feasibility, but showed the need for
including flocculation and sedimentation for removal of minute particles, together with
chlorine and iodine addition for maximum disinfection. The recommended
storage/treatment plan provides for a 175 million gallon storage basin, a 50
million-gallon-per-day reclamation facility and two 46-acre swimming and boating lakes.
Cost effectiveness (Cost/Benefit Ratio) of the project, as envisioned, has been indicated
to be 1.6 at an estimated total project cost of $45,200,000, and an estimated annual
operating cost of $1,777,000. Implementation of the proposed plan would not only provide
a least-cost alternative over single-purpose projects to attain identical objectives, but would
also reduce the annual pollution now discharged by the Northeast Boundary Trunk Sewer
by approximately 99 percent.
This report was submitted in fulfillment of Program No. 11023 FIX under Contract No.
14-12-829 between the Federal Water Quality Administration and Roy F. Weston, Inc.
Conclusions
It is the considered opinion of Roy F. Weston, Inc. that this project in its entirety is
both technically and economically feasible. It is possible for the Federal Government to
demonstrate that wastewater and waste land can be reclaimed for the use and advancement
of society, and that a total environmental approach to the problem of pollution can be
effective.
Combined Sewer Overflow
1. The project conceived as a result of the study would be the largest control and
treatment works for combined sewer overflows in the United States and would elevate
the effluent quality above all other existing or planned combined sewer overflow
projects.
2. The Northeast Boundary Trunk Sewer serves approximately one third of the combined
sewer area of the District of Columbia.
241
-------
3. The methodology developed for this study provides a reasonably accurate definition
of the quantity, quality, and variability of overflow from the Northeast Boundary
Trunk Sewer.
4. The annual discharge of BODs in the overflow from the Northeast Boundary Trunk
Sewer accounts for over 25 percent of the recommended allowable waste loading
to the Potomac River in the Washington metropolitan area. Even the storm that is
expected to occur with a frequency of four times or more per year results in a
BOD loading of over twice the recommended allowable daily loading for discharges
from all waste treatment facilities in the metropolitan Washington area. The recent
Potomac Enforcement Conference held 21 and 22 May 1970 required that these
recommended allowable loadings be further reduced.
5. The impact of BOD loading in the overflow on dissolved oxygen levels is more serious
than the 25 percent value reflects, due to the combined effect of long residence
times of estuarine waters and the true form of overflows occurring at discrete impact
loadings.
Storage Alternatives
6. Storage is an essential element of any plan for the abatement of pollution from the
Northeast Boundary Trunk Sewer, because of the extremely high overflow rates.
7. Locating a surface storage basin in the vicinity of Kingman Lake is feasible and
desirable since all overflow can be collected there without extensive modifications
and without pumping, and also because much of the area is undeveloped, indicating
the release of open space for a storage basin is possible. The estimated cost for surface
storage was $14.3 million as compared to $33.4 million for comparable mined storage.
8. Preliminary soils investigation indicates that the construction of a storage basin in
the lower section of Kingman Lake is technically feasible.
9. The tunnel storage concept is applicable to the Northeast Boundary Trunk Sewer
and has the decided advantage of providing surcharge relief of the existing sewer
system.
Treatment Alternatives
10. Different levels of benefits are associated with each capacity of a wastewater
reclamation facility; these include allowable bather load, probability of not
overflowing the storage basin, and the additional storage capacity provided by the
plant's operating during the period of the storm.
11. A literature evaluation disclosed that there are some promising alternatives to those
unit processes investigated in this study; however, confirmation work is necessary
before any one can be applied totally to the Kingman Lake Project. Included in
this listing are fiberglass filtration media and microstraining for suspended solids
removal, U-tube aerators for aeration, and ozonation for odor control and disinfection.
12. The use of iodine in conjunction with chlorine will provide effective disinfection
of viruses as well as bacteria in the swimming lake.
242
-------
13. The use of chlorine in the swimming lake will prevent excessive algae growth, but
there may be an algae problem in the fishing-boating lake, which does not otherwise
require chlorination.
14. The recommended standards for the influent to the swimming lake are: pH 7.5
to 8.0; BOD 5.0 mg/L; Suspended Solids -15.0 mg/L; Total Phosphorus 0.05 to
1.0 mg/L; Free Chlorine 1.0 mg/L; Free Iodine 1.0 mg/L; Fecal Coliform 200
per 100 ml.
Laboratory Investigative Program
15. Coagulation-Sedimentation followed by multi-media filtration, activated carbon
adsorption, and disinfection will produce an effluent which meets the water quality
criteria objectives for swimming and fishing.
16. Filtration through fiberglass was shown to be an effective method for removal of
suspended solids; however, significant development work is required prior to any major
facility application.
Soils Investigation
17. Construction of a storage basin with vertical walls to a depth of minus 40 feet is
feasible.
18. Construction of a storage basin with sloping side walls is undesirable because of the
restrictions on side slopes, the period of construction, and the relatively small storage
capacity available.
19. Construction cost estimates indicate that significant economics can be effected through
use of the slurry wall construction method rather than the conventional approach
using sheeting and bracing.
20. The foundations for all structures must extend to the sand and gravel layer or to
the underlying stiff clay. For shallow structures, it will generally be most economic
to provide pile foundations extending to these strata. The deeper structures can be
founded directly on the above-mentioned strata or may be constructed on pile
foundations.
21. Problems concerning general side grading, seepage, and other work associated with
raising the fishing and swimming lakes to elevation +3.5 feet are not anticipated.
Selection of Alternative
22. The alternative scheme evaluated to have the highest cost effectiveness encompasses
a storage basin capacity of 175 million gallons and a reclamation plant capacity of
50 million gallons per day.
23. The alternative scheme selected will provide:
a. Ninety-nine percent reduction of annual pollution load from Northeast Boundary
Trunk Sewer, eighty percent reduction of pollution from a storm with a two-year
recurrence frequency, and sixty-one percent reduction of pollution from a storm
with a five-year recurrence frequency.
243
-------
b. Effective storage capacity to contain a storm with a recurrence frequency of
1.2 years.
c. A ninety-six percent probability that the reclamation facility will draw down
the storage basin prior to the recurrence of a second storm overflow (the volume
of which would exceed the remaining volume in the storage basin).
d. Sufficient treatment plant capacity to support a maximum of 30,000 bathers
per day in the swimming lake.
24. The covering of the storage basin with a parking roof is justified on the basis of
comparable land acquisition and development costs and estimated annual revenues.
25. The total project cost estimate of $45,200,000 compares extremely well with benefits
estimated at $72,755,000.
26. Operating costs are estimated to be $1,777,000 per year. Included in the operating
costs are: administration, labor, maintenance, utilities, chemicals, make-up carbon,
fuel oil, and an operating contingency.
Recommendations
1. Demonstrate how the treatment of combined sewer overflows can serve additional
beneficial uses.
2. Demonstrate how the FEDERAL GOVERNMENT can approach the total solution
to environmental problems.
3. Design and construct a 175,000,000-gallon storage facility, a 50,000,000-gallon per
day water-reclamation facility, and the associated swimming, boating, fishing, and
parking facilities.
4. Continue to investigate and gather additional data to refine this report and estimate
prior to and during the engineering design and construction phases of this project.
5. Provide sufficient space within the water-reclamation plant and establish a field test
facility to demonstrate new and promising processes for improving the treatment of
combined sewer overflows.
244
-------
A ccessiori1 Number
Snbj-e.ct Field & Group
SELECTED WATER RESOURCES ABSTRACTS
INPUT TRANSACTION FORM
Organization
ROY F. WESTON, West Chester, Pennsylvania
Title
COMBINED SEWER OVERFLOW ABATEMENT ALTERNATIVES:
WASHINGTON, D.C.
10
Authors)
Buckingham, Phillip L.
Shin, ChiaS.
Ryan, James G.
Lee, James A.
Kane, John K.
16
Project Designation
EPA, WOO Contract No. 14-12-403
21 Note
22
Citation
23
Descriptors (Starred First)
*Storm Runoff, *0verflow, * Flow Measurement, Underground Storage, *Filtration, Design Storm,
Depth-Area-Duration Analysis, Rainfall-Runoff Relationships, Organic Loading, Treatment Facilities,
Tunnel Design, Reservoir Design, Sewers, Capital Costs, Annual Costs, Comparative Costs, Geology,
Tracers, Analysis, Sludge, Hydrology.
25
Identifiers (Starred First)
*Combined Sewers, *Storm Water, Potomac River, District of Columbia
27
Abstract
Objectives of the project were: 1) define the characteristics of combined sewer overflow; 2) investigate
the feasibility of high-rate filtration for treatment of combined sewer overflow; and 3) develop and evaluate
alternative methods of solution.
Investigative activities included: review of pertinent reports and technical literature; field monitoring
of combined sewer overflows and separated storm water discharges at three sites; laboratory studies of
ultra-high-rate filtration of combined sewer overflow; hydrological analysis; and evaluation of feasible
alternatives (based on conceptual designs, preliminary cost estimates, and other factors).
Reservoir Storage, Treatment at Overflow Points, Conveyance Tunnels and Mined Storage, and Sewer
Separation were the approaches considered sufficiently promising for detailed evaluation. Tunnels and
Mined Storage with treatment at the Blue Plains plant and at Kingman Lake after subsidence of the storm
is recommended. Estimated capital costs (based on the 15-year storm) are $318,000,000 with annual
operation and maintenance costs of $3,500,000. This approach also was preferable to the others on the
basis of systematic evaluation of reliability, flexibility, public convenience and other non-quantifiable
factors.
Abstractor
Institution
ROY F. WESTON
SEND TO: WATER RESOURCES SCIENTIFIC INFORMATION CENTER
U.S. DEPARTMENT OF THE INTERIOR
WASHINGTON. D. C. 20240
« SPO: 1969-359-339
-------
Continued from inside front cover....
11022 08/67
11023 09/67
11020 12/67
11023 05/68
11031 08/68
11030 DNS 01/69
11020 DIH 06/69
11020 DES 06/69
11020 06/69
11020 EXV 07/69
11020 DIG 08/69
11023 DPI 08/69
11020 DGZ 10/69
11020¥EKO 10/69
11020 10/69
11024 FKN 11/69
11020 DWF 12/69
11000 01/70
,11020 FKI 01/70
11024 DOK 02/70
11023 FDD 03/70
11024 DMS 05/70
11023 EVO 06/70
11024 06/70
11034 FKL 07/70
11022 DMU 07/70
11024 EJC 07/70
11020 08/70
11022 DMU 08/70
11023 08/70
11023 FIX 08/70
11024 EXF 08/70
Phase I - Feasibility of a Periodic Flushing System for
Combined Sewer Cleaning
Demonstrate Feasibility of the Use of Ultrasonic Filtration
in Treating the Overflows from Combined and/or Storm Sewers
Problems of Combined Sewer Facilities and Overflows, 1967
(WP-20-11)
Feasibility of a Stabilization-Retention Basin in Lake Erie
at Cleveland, Ohio
The Beneficial Use of Storm Water
Water Pollution Aspects of Urban Runoff, (WP-20-15)
Improved Sealants for Infiltration Control, (WP-20-18)
Selected Urban Storm Water Runoff Abstracts, (WP-20-21)
Sewer Infiltration Reduction by Zone Pumping, (DAST-9)
Strainer/Filter Treatment of Combined Sewer Overflows,
(WP-20-16)
Polymers for Sewer Flow Control, (WP-20-22)
Rapid-Flow Filter for Sewer Overflows
Design of a Combined Sewer Fluidic Regulator, (DAST-13)
Combined Sewer Separation Using Pressure Sewers, (ORD-4)
Crazed Resin Filtration of Combined Sewer Overflows, (DAST-4)
Stream Pollution and Abatement from Combined Sewer Overflows •
Bucyrus, Ohio, (DAST-32)
Control of Pollution by Underwater Storage
Storm and Combined Sewer Demonstration Projects -
January 1970
Dissolved Air Flotation Treatment of Combined Sewer
Overflows, (WP-20-17)
Proposed Combined Sewer Control by Electrode Potential
Rotary Vibratory Fine Screening of Combined Sewer Overflows,
(DAST-5)
Engineering Investigation of Sewer Overflow Problem -
Roanoke, Virginia
Microstraining and Disinfection of Combined Sewer Overflows
Combined Sewer Overflow Abatement Technology
Storm Water Pollution from Urban Land Activity
Combined Sewer Regulator Overflow Facilities
Selected Urban Storm Water Abstracts, July 1968 -
June 1970
Combined Sewer Overflow Seminar Papers
Combined Sewer Regulation and Management - A Manual of
Practice
Retention Basin Control of Combined Sewer Overflows
Conceptual Engineering Report - Kingman Lake Project
Combined Sewer Overflow Abatement Alternatives -
Washington, D.C.
------- |