EPA-600/2-76-145
August 1976
Environmental Protection Technology Series
                  METHODOLOGY  FOR THE STUDY  OF
                           URBAN STORM  GENERATED
                             POLLUTION AND CONTROL
                                 Municipal Environmental Research Laboratory
                                      Office of Research and Development
                                      U.S. Environmental Protection Agency
                                             Cincinnati, Ohio 45268

-------
                                             EPA-600/2-76-1^5
                                             August 1976
              METHODOLOGY FOR THE STUDY OF

                  URBAN STORM GENERATED

                  POLLUTION AND CONTROL
                           by

                 Richard E. Wullschleger
                   Alphonse E. Zanoni
                    Charles A. Hansen
            Envirex Inc. (A Rexnord Company)
             Environmental  Sciences Division
               Milwaukee, Wisconsin  5321A
                 Contract No. 68-03-0335
                    Project Officers

            Richard I. Field and Chi Yuan Fan
            Storm and Combined Sewer Section
Municipal Environmental Research Laboratory (Cincinnati)
                Edison, New Jersey  08817
       MUNICIPAL ENVIRONMENTAL RESEARCH LABORATORY
           OFFICE OF RESEARCH XAND DEVELOPMENT
          U.S. ENVIRONMENTAL PROTECTION AGENCY
                 CINCINNATI, OHIO  *f5268

-------
                              DISCLAIMER
This report has been reviewed by the Municipal Environmental Research
Laboratory, U.S. Environmental Protection Agency, and approved for
publication.  Approval  does not signify that the contents necessarily
reflect the views and policies of the U.S. Environmental Protection
Agency, nor does mention of trade names or commercial products
constitute endorsement  or recommendation for use.
                                 11

-------
                                 FOREWORD

     The Environmental Protection Agency was created because of
increasing public and government concern about the dangers of pollu-
tion to the health and welfare of the American people.   Noxious air,
foul water, and spoiled land are tragic testimony to the deterioration
of our natural environment.  The complexity of that environment and
the interplay between its components require a concentrated and
integrated attack on the problem.

     Research and development is that necessary first step in problem
solution and it involves defining the problem, measuring its impact,
and searching for solutions.  The Municipal Environmental Research
Laboratory develops new and improved technology and systems for the
prevention, treatment, and management of wastewater and solid and
hazardous waste pollutant discharges from municipal and community
sources, for the preservation and treatment of public drinking water
supplies, and to minimize the adverse economic, social, health, and
aesthetic effects of pollution.  This publication is one of the
products of that research; a most vital communications  link between
the researcher and the user community.  It contains recommended stan-
dard procedures to be used in studies of urban storm generated pollu-
tion and control.  Although the study was sponsored by  the Storm and
Combined Sewer Section and recommendations are made with such an appli-
cation in mind, this report is much more general.  It is hoped that it
will be of interest and helpful to anyone active in the water pollution
control field.
                                  Francis T. Mayo
                                      Director
                   Municipal Environmental Research Laboratory
                                   111

-------
                               ABSTRACT


This report contains recommendations for standard procedures to be
followed in the conduct of projects dealing with pollution assessment
and abatement of storm generated discharges.  The purpose of this
project was to develop standard procedures needed to insure that all
discharges and treatment processes coul.d be evaluated by the same means.
The procedures chosen were those found to be the most applicable and
optimum for the field of storm and combined sewer overflow pollution
control.

The project efforts were devoted to the major areas listed below.

     1.  Recommended methods for sampling and sample preservation.

     2.  Appropriate monitoring instrumentation available.

     3.  The choice of quality parameters to be utilized.

     4.  The analytical procedures to be followed.

     5.  The methods for evaluating storm generated discharge pollution.

     6.  The standard procedures for evaluating treatment processes
         treating storm generated flows.

Choice of the recommended procedures was based upon the U.S. EPA research
and demonstration project reports in this and associated fields, other
published literature, ongoing U.S. EPA funded projects, and the
contractor's experience in the field of stormwater pollution control.

This report was submitted in fulfillment of Contract No. 68-03-0335 under
the sponsorship of the Environmental Protection Agency.  The project was
performed by the Environmental Sciences Division of Envirex Inc.  Work
was completed as of December 1974.
                                   IV

-------
                                CONTENTS

                                                               Page

Abstract                                                        jv

List of Tables                                                 vi i

List of Figures                                                  x

Acknowledgments                                                 xv


Sections
t       SUMMARY                                                  1

           Sampling Procedures and Considerations                1
           Monitoring  Instrumentation                            3
           Quality Characteristics and Laboratory Studies        k
           Describing  Storm Generated Discharges                 8
           Methods for Evaluation of Treatment Processes         9
              for Storm Generated Discharges

II      RECOMMENDATIONS                                         11

III     INTRODUCTION                                            13

IV      SAMPLING PROCEDURES AND CONSIDERATIONS                  16

           Sampling Storm Generated Discharges                  16
           Types of Sampling Programs                           25
           Sample Handling and Preservation                     32
           Automatic Sampling Equipment                         35
           Recommended Sampling Programs                        ^5
           Sampling Accumulated Roadway Material                4?

V       MONITORING INSTRUMENTATION                              55

           Flow Measuring Equipment                             55
           In Situ Monitoring Equipment                         81
           Raingage Networks                                    90

-------
Sections                                                         Page

  VI       QUALITY CHARACTERISTICS  AND  LABORATORY  PROCEDURES        118

              Quality Characteristics  for  Storm Generated
                Discharges                                         118
              Past Practices  Relative  to Storm Generated
                Discharges                                         '22
              Recommended Oxygen Demand Potential  Indicator        128
              Recommended Measurement  of Particulate
                Concentration                                     1^7
              Recommended Choice of  Pathogenic Indicator           158
              Recommended Measurement  of Eutrophication
                Potential                                          172
              Recommended Choice of  Metal Analyses                 182
              Pesticides and  Polychlorinated Biphenyls             189
              Recommendation  of Other  Characteristics              200
              Analysis of Sludges                                  222
              Analysis of Accumulated  Roadway Material             225
              Summary                                             230


  VII      DESCRIBING  STORM GENERATED FLOWS                         233

              Explanation and Discussion of Terms and Units        233
              Developing Background  Data                           236
              Recommended Method of  Evaluating Storm Generated
                Discharges                                         2k$
              Evaluation of Storm Generated Discharges             257


  VIII    METHODS  FOR EVALUATION OF  STORMWATER TREATMENT
           PROCESSES                                              260

              Brief Inventory of Control and Treatment Methods     261
              Evaluation of Treatment  Processes                    271
             Economic Decision Making Considerations              299

  IX      REFERENCES                                               302


 X       GLOSSARY
                                 VI

-------
                                TABLES


 No>                                                               Page

IV-1    DIFFERENT TYPES OF SAMPLES                                  26

IV-2    RECOMMENDED PRESERVATION METHODS                            33

IV-3    PURPOSE OF DIFFERENT PRESERVATIVES                          36

IV-4    AUTOMATIC SAMPLER COSTS AND SAMPLE TYPES OBTAINABLE         37

IV-5    AUTOMATIC SAMPLER CAPABILITIES AND SAMPLING METHODS         kO

IV-6    SAMPLING PROCEDURE FOR THE COLLECTION OF STREET SURFACE
          CONTAMINANTS                                              50

IV-7    DISTRIBUTION OF POLLUTANTS BETWEEN DUST AND DIRT AND
           FLUSH SAMPLE FRACTIONS                                   52


V-1     RUNOFF COEFFICIENTS                                         96

V-?     NUMBER OF RAINGAGES NECESSARY TO PREDICT RUNOFF FOR
          CYCLONIC STORMS                                          102

V-3     SIMPLE CORRELATION COEFFICIENT RANGE BETWEEN RAINGAGES     103

V-*»     DISTRIBUTION OF SEWERAGE DRAINAGE CATCHMENT SIZES IN
          SOME MAJOR CITIES                                        106

V-5     RAINGAGE DENSITIES FOR DIFFERENT STUDIES                   10?

V-6     RECOMMENDED RAINGAGE DENSITY FROM LITERATURE               115

V-7     RAINGAGE NETWORK NEEDS FOR DIFFERENT STUDY AREAS           116
VI-1    CLASSIFICATION, EXAMPLES, AND PRIMARY SOURCES OF
          POLLUTION FOUND IN STORM RUNOFF AND COMBINED SEWER
          OVERFLOW                                                 123

VI-2    USAGE FREQUENCY OF VARIOUS QUALITY CHARACTERISTICS IN
          PAST STUDIES INVOLVING STORM GENERATED DISCHARGES        12*»

VI-3    USAGE FREQUENCY OF VARIOUS QUALITY CHARACTERISTICS IN
          PAST STUDIES INVOLVING TREATMENT OF STORM GENERATED
          DISCHARGES                                               125

                                   vii

-------
                          TABLES  (continued)
   No.
 VI-4     USAGE  FREQUENCY OF VARIOUS QUALITY CHARACTERISTICS IN
            PAST STUDIES  INVOLVING ONLY IMPACT ON RECEIVING
            WATER                                                   ' 26

 VI -5     SUMMARY COMPARISON OF VARIOUS POTENTIAL OXYGEN DEMAND
            TESTS

 VI -6     EFFECT OF  IMPORTANT VARIABLES ON POTENTIAL OXYGEN
            DEMAND TESTS

 VI -7     COMPARISON OF NUTRIENT LOADINGS FROM SEPARATE STORM
            AND  COMBINED SEWERS, DETROIT - ANN ARBOR AREA            179

 VI -8     COMPARISON OF NUTRIENT LOADINGS FROM SEPARATE STORM
            AND  COMBINED SEWERS, WASHINGTON, D.C. AREA               179

 VI -9     COMPARISON OF AVERAGE ANNUAL NUTRIENT CONCENTRATIONS
            IN URBAN RUNOFF                                          180
 VI-10    HEAVY METALS LOADING INTENSITIES

 Vl-ll    INFORMATION SURVEY FOR VARIOUS HEAVY METALS                185

 VI-12    EFFECT OF EXPOSURE OF PESTICIDES TO MERCURY AND COPPER     195

 VI -13    POLLUTANT CHARACTERISTICS OF URBAN RUNOFF                  201

 VI-H    DEBRIS AND BULK SOLIDS FOUND IN STORMWATERS                207

 VI -1 5    OILS AND GREASE IN STORM GENERATED WASTEWATERS             208

 VI-16    SCREENING REMOVAL EFFICIENCIES FROM LITERATURE             210

 VI -17    WET SIEVE ANALYSIS RESULTS FROM TWO COMBINED SEWER
           OVERFLOWS                                                211

VI-18    PARTICLE SIZE DISTRIBUTION OF SOLIDS, SELECTED CITY
           COMPOSITES                                               213

VI -19    RECOVERY OF ORGAN I CS FROM WATER USING POROUS POLYMER
           PACKINGS                                                 218

VI -20    SUMMARY OF CRITERIA FOR ANALYZING THE RECOMMENDED
           CHARACTERISTICS                                          232
                                  Vlll

-------
                            TABLES (continued)
  No.                                                              Page

VI1-1    LOCAL CLIMATOLOGICAL DATA, ANNUAL SUMMARY WITH
           COMPARATIVE DATA, 1968                                   239

VI1-2    RAINFALL HYETOGRAPH                                        251
VI I 1-1   RATIOS DEVELOPED FROM  INFORMATION IN FIGURE VI I I-6         275

VI I 1-2   HYPOTHETICAL DATA ANALYSIS  (DISSOLVED-AIR FLOTATION
           SYSTEM)                                                  281
                                     IX

-------
                           LIST OF FIGURES
 No.

IV-1       Example of variance  in runoff quality and quantity
             with time                                              '8

IV-2       Example of variance  in runoff quality and quantity
             with time                                              '9

IV-3       Suspended solids variation with time, Racine,
             Wisconsin                                              20

IV-4       Influence of sampling velocity on suspended solids        2k

IV-5       Random grab sample collection                             27

IV-6       Method of compositing samples on a fixed volume-fixed     27
             time interval basis

IV-7       Method of taking discrete samples at constant flow
             volume Increments                                      28

IV-8       Method of compositing samples proportional to flow
             volume at constant time interval                       28

IV-9       Method of compositing samples proportional to flow
             rate                                                   29

IV-lO      Method of compositing samples of equal volume at
             equal  increments of flow                               29

 V-l       Level gaging dipstick                                     56

 V-2       Float driven recorder with clock drive                    56

 V-3       Recorder with scow float used in a sewer manhole          56

 V-4       Air bubbler level gage                                    57

 V-5       "Dipper" level gage based on electrical contact with      58
            the sewage surface

 V-6       Portable ultrasonic flow gage                             59

V-7       Dual range Parshall flume                                 62

V-8       Lagco flume                                               63

V-9       Float actuated flow recorder                             6^

-------
                           FIGURES (continued)

No.

 V-10     Scow actuated flow indicator                             65

 V-ll     Portable bubbler actuated flow recorder and              66
             total 1izer

 V-12     Dipper level transmitter                                 67

 V-13     Total flow computer-dipper system                        68

 V-1A     Portable flowmeter-dipper system                         68

 V-15     Flow recorder, indicator and total 1izer actuated         69
             by ultrasonic level measurement

 V-16     Flume-tracer dilution combination for sewer flow         70
             gaging

 V-17     Ultrasonic flow velocity and level  gaging on a raw       71
             sewage  channel,  (coarse screened)

 V-18     Ultrasonic velocimeter probes installed on inside        72
             surface of a sewer wall

 V-19     Pitot-type electromagnetic flowmeter transmitter         73

 V-20     "Velmeter" electromagnetic flowmeter transmitter         73
             and indicator

 V-21     Installation sketch,  propeller meter for sewage flow     75
             gaging

 V-22     Manhole insert housing and propeller meter for sewage    76
             flow gaging

 V-23     Doppler ultrasonic  flowmeter                             76

 V-2A     Portable doppler ultrasonic flowmeter                    77

 V-25     Electromagnetic flowmeter for full  pipe application       78

 V-26     Ultrasonic flowmeter  for full pipe  applications          79

 V-27     Signal  processing unit,  ultrasonic  flowmeter             80
                                   XI

-------
                           FIGURES  (continued)

No.

 V-28     TOD analyzer                                              83

 V-29     TOD analyzer                                              83

 V-30     TOC analyzer                                              85

 V-31     TOC analyzer                                              85

 V-32     Dissolved  oxygen  sensor                                   86

 V-33     Self-cleaning,  submersible  suspended  solids  sensor       88

 \l-3k     Cross section of  suspended  solids  sensing  head            88

 V-35     Control  unit for  suspended  solids  monitor                 88

 V-36     Surface  scatter turbidimeters                             89

 V-37     pH analyzer                                              90

 V-38     Conductivity analyzer                                     90

 V-39     Intensity  - duration   -  frequency  curves  for             95
             Bucyrus, Ohio

 V-40     Distribution of rainfall  intensity with  respect         110
             to time - from reference 20

 V-^l     Distribution of rainfall  intensity with  respect         110
             to accumulated rainfall  - from  reference  20

 V-^2     Graphs from Racine, Wisconsin  illustrating the           112
             effects of multiple raingages on the U.S. EPA
             Stormwater Management  Model

 \l-k3     Comparison of raingage densities recommended in         11A
             the literature with those in storm generated
             discharge studies
                                   XII

-------
                          FIGURES (continued)
No.
VI-1    Illustration of the danger In estimating the ultimate
          demand without knowledge of the rate of deoxygena-
          tion                                                      132

VI-2    Illustration of how the BOD20 lessens the effect of
          different rates of deoxygenatlon (k)                      132

VI-3    Illustration of the COD value becoming asymptotic with
          time when held at 20°C                                    134

VI-4    Drawing of a recommended procedure for obtaining proper
          elution rates                                             197

VI-5    Subjective check-off sheet for characterizing various
          contaminants                                              202

VI-6    Suspended solids removal for different size screens from
          storm generated discharge projects                        210

VI-7    Schematic diagram of a cell used to collect headspace
          vapors from solids and liquids                            217

VI-8    Extractor design for solvent llgher than water              220

VI-9    Extractor design for solvent heavier than water             220

VI-10   Standard curve - rubber In dust and dirt                    227
VI1-1   Average flow rates for twelve discrete samples              235

VI1-2   Pollutant mass loading curve                                235

VI1-3   Division of Bloody Run sewer watershed into different
          land uses                                                 238

VI1-4   Division of Bloody Run sewer basin into subareas and
          numbering of the sewer system's elements                  246

VI1-5   Recording ralngage chart                                    250

VI1-6   Rainfall hyetograph - Racine, Wisconsin, April 26, 1974     253

VI1-7   Pollutograph                                                255
                                  Kill

-------
                           Figures  (continued)


 to.	

 VI 1-8       Suspended  solids and flow variation with time             255


 VI I 1-1      Storage  process diagram                                   263

 VIII-2      Physical-chemical process diagram                         266

 VI 11-3      Dissolved  air flotation process diagram                   268

 VI I 1-4      Contact  stabilization process diagram                     270

 VI 11-5      Generalized combined sewer overflow treatment system      273

 VI11-6      Graphical  illustration of a hypothetical flow rate
             to a treatment system                                   274

 VIM-7      Graphical  illustration of a hypothetical flow rate
             to a storage tank                                       274

 VI11-8      Hypothetical mass loading on two different types of
             treatment processes and the resultant effluent
             quali ty pattern                                         278

 VI 11-9      Flow diagram (total  flow to a treatment unit)             282

 VIII-lO     Hydraulic loading rate to a flotation unit (split flow)
             indicating change in recycle percentage                 283

 VI I 1-11     Solids loading  as a  function of time                      284

VI I 1-12    Pollutograph for suspended solids (loading vs time)       284

VI I 1-13    Overall  efficiency at each time increment                 286

VI I 1-14    Hypothetical example of the economic solution
             methodology approach                                    301
                                  xiv

-------
                            ACKNOWLEDGEMENTS
In addition to the primary authors named on the cover page, the
following employees of the Environmental Sciences Division of Envirex
made significant contributions to this report:  Dr. R. W. Agnew,
R. J. Fulk, M. J. Clark and K. R. Huibregtse.  Dr. William Genthe of
Rexnord Inc., and Dr. John L. Carter of Michigan Technological
University also authored portions of the report.  The information for
developing the sections of this report dealing with the sampling and
analysis of roadway debris was supplied by Robert Pitt, James D. Sartor
and Gail B. Boyd of Woodward-Clyde Consultants, Donald A. Shaheen of
Biospherics,  Inc. and Francis J. Condon of the U.S. EPA Office of
Research and  Development.

Review and assistance in the preparation of this manuscript in final
form was supplied by the people listed below:


Mr. Dwight G. Ballinger, USEPA, Acting Director-Environmental  Monitoring
                         and Support Laboratory, Cincinnati, Ohio
                         (formerly USEPA, Director-Methods Development
                         and Quality Assurance Research Laboratory,
                         Cincinnati, Ohio)

Mr. Robert T. Williams,  USEPA, Chief-Waste Identification and Analysis
                         Section, AWTRL, Cincinnati, Ohio


                   Pathogenic Indicator Portion Only


Dr. Edwin E.  Geldreich,  USEPA, Research Director-Microbiological
                         Treatment Branch, Water Supply Research Division,
                         Cincinnati, Ohio (formerly USEPA, Research
                         Director-Microbiological Quality Control in
                         Water, Water Supply Research Laboratory,
                         Cincinnati, Ohio)

Mr. Cecil  W.  Chambers,   USEPA, Research Microbiologist-Biological
                         Treatment Section, TPOP, AWTRL, Cincinnati,
                         Ohio
                                    xv

-------
Mr. Francis T. Brezenski, USEPA, Chief-Surveillance and Monitoring Branch,
                          Region II, Edison, New Jersey

Mr. Irwin Katz,  USEPA, Microbfologist-Survei1 lance and Monitoring Branch,
                 Region II,  Edison,  New Jersey

Dr. Edwin C. Tifft,  Jr., O'Brien and Gere Engineers

Dr. Vincent P. Olivier!, The Johns Hopkins University
Sincere appreciation is  expressed  to  the  entire staff of the Storm and
Combined Sewer Overflow  Technology Branch  in  Edison,  New Jersey for
their assistnace.   Special  thanks  are given  to Mr.  Chi  Yuan Fan, the
Project Officer,  and Mr.  Richard Field, Project Officer for their
continued help and interest during the entire  project  duration.
                                  xvi

-------
                        SECTION I  - SUMMARY
SAMPLING PROCEDURES AND CONSIDERATIONS

I.  The importance of proper sampling is paramount since the decisions
    which will be based upon the samples can be no more accurate than
    the sample itself.

2.  Great variance in flow patterns and quality characteristic patterns,
    together with the spatial differences in storm generated discharges
    makes obtaining of truly representative samples almost impossible.

3.  When locating samplers for storm generated discharge sampling,
    certain trade offs have to be made between ideal samples and intake
    location and accessibility.

4.  A single grab or random sample is not sufficient for characterizing
    the average quality of a storm generated discharge.

5-  There are four common methods of compositing storm generated
    discharges.  They are:

         a.  Constant time - constant volume:  samples of equal volume
             are taken at equal  increments of time and composited to make
             an average sample.

         b.  Constant time - volume proportional to flow increment:
             samples are taken at equal  increments of time and are
             composited proportional to  the volume of flow since the
             last sample was taken.

         c.  Constant time - volume proportional to flow rate:  samples
             are taken at equal  increments of time and are composited
             proportional to the flow rate at the time each sample was
             taken.

-------
        d.   Constant volume  -  time  proportional  to flow increment:
            samples  of  equal volume are  taken  at  equal  increment of
            flow and composited.

Although this study  has shown  the method  described by  c.   is the  best,
both b. and d.  are considered  acceptable.

6.  At least 2000 ml of sample are  required  to perform the recommended
    analyses contained  in this report.

7.  Although there are  many  automatic  samplers on the  market,  the ideal
    sampler for sampling storm generated  discharges still  does not
    ex i s t.

8.  Theoretically, at least  four  separate sample  bottles  would be
    needed  per sample so that  the various preservatives required for
    certain constituents could be added  to the sample  bottle at the time
    of sampling.

9.  The following steps must be taken  when locating a  sampler for storm
    generated discharge sampling:

         1.  Maximum accessibility  and safety  - Manholes  on busy
             streets should  be avoided if possible; shallow depths
             with manhole steps in  good  condition are  desirable.  Sites
             with a  history  of surcharging and/or submergence by
             surface water should  be avoided if possible.  Avoid loca-
             tions which may tend  to invite vandalism.

         2.  Be sure that the  site  provides the information desired -
             Familiarity with  the  sewer  system is necessary.  Knowledge
             of the  existence  of  inflow  or outflow between the sampling
             point and  point of data use is essential.


         3-  Make certain the  site  is  far enough downstream from trib-
             utary  inflow to ensure mixing of  the tributary with the
             main sewer.

         4.  Locate  in  a straight  length of sewer, at  least six sewer
             widths  below bends.

         5.  Locate  at  a point of maximum turbulence,  as found in sewer
             sections of greater  roughness and of probably higher
             velocities.  Locate  just  downstream from  a drop or
             hydraulic  jump, if possible.

         6.  In all  cases, consider the cost of installation, balancing
             cost against effectiveness  in providing the data needed.

-------
10.   Plastic sample containers should be used in all  sample collection
     except where grease and oil or pesticides are to be analyzed.

11.   All samples should be kept in a refrigerator or ice cooled container
     during and after sampling.

12.   Bacterial analyses run on composite samples can only be used for
     system control and not for "reportable" values because of the
     cross contamination which will occur.  If a quantitative test is
     required by regulatory agencies, then the sampling and preservation
     techniques mandated by the agency should be followed.
MONITORING INSTRUMENTATION

1.  The construction and performance features of ultrasonic level
    gaging equipment makes it suitable to the measurement and main-
    tenance requirements of level gaging.  With a cost in the $800-
    $1,200 range, and at least twelve domestic manufacturers now offer-
    ing them, the ultrasonic level gage should find increasing use for
    tracing sewer levels during and after storm events, for infiltration
    studies, for determining discharge flows in overflow sewers, and
    for determining storage capacity, routing programs, and gate
    control in "in-line" storage systems.

2.  There  is presently no generally accepted sewer flow gaging technique
    which  provides accurate measurement over the entire range of sewer
    flows  from minimum dry weather flows to surcharge conditions.   There
    are, however, several promising methods which must yet be fully
    demonstrated in actual installation.

3.  In sewers not subject to surcharging, where hydraulics allow
    flumes to operate within acceptable submergence, these structures
    are suitable and flow head can be measured by ultrasonic level
    gages, floats, scows, capacitance gages, bubblers, lead cell gages
    and dippers.

k.  In situ monitoring equipment with instream sensing is available for
    measurement of turbidity, suspended solids and nitrates, and side
    stream analyzers are available for TOD, orthophosphate, nitrates
    and nitrites.  Other m situ monitors for parameters characteristic
    of particular storms are also available.  In situ moni toring,  at
    this time, is not recommended to replace analytical laboratory
    measurements, however, in si tu moni tors are recommended for flow
    control purposes, quick indicators of treatment efficiency, or for
    monitoring changes in sewage characteristics.

5.  The most common errors arising in raingage measurement are (1)
    mistakes in reading the scale of the stick or chart, (2) evapora-
    tion of water in the gage, (3) improper placement of the gage and

-------
     (4)  instrument error.

 6.   Optimum raingage location is simply on level ground with no trees or
     structures  in a proximity such as to affect the capture of rainfall.

 7.   The  use of  raingages in storm generated discharge studies is
     dependent upon the purpose of the project as shown below.

            RAINGAGE NETWORK NEEDS FOR DIFFERENT STUDY AREAS

                                                         Need for
                                                         Network

 To define areal rainfall and its variations               yes

 Mathematical Model verification                           yes

 Sewer, Treatment Plant, or System Design                  varies

 Operation of Storm Generated Discharge Systems            no
  (Treatment or  Holding Basin)

 Operation of In-line Storage or Selective                 yes
  Discharge Systems

 8.   The  density of raingages is not clearly formulated and is dependent
     upon project objectives.  However, at least three raingages should
     always be used when precipitation monitoring is being performed, so
     that the Theissen method may be used.


 QUALITY  CHARACTERISTICS AND LABORATORY STUDIES

 1.  The peculiar characteristics of storm generated discharges preclude
     the direct usage of commonly used characteristics for municipal dry-
    weather flows without some changes or modifications.  This fact
    applies not only to the characteristics themselves, but also to the
    analytical   laboratory techniques used.

2.  The many variables which can affect the quality of storm generated
    discharges such as time between occurrences, rainfall amount,
     intensity and duration and drainage area surface conditions make
    the use of the term "typical" to describe the quality of storm
    generated discharges a nuisance.

3.  The following water quality parameters have been recommended for
    routine use in storm generated discharge projects: Total Oxygen
    Demand, Five Day Biochemical Oxygen Demand, Fecal Coliform, Total
    Oxidized Nitrogen, Total Reduced Nitrogen (Kjeldahl), Total

-------
     Phosphorus and  pH.   In addition to the recommended procedures,
     periodic analyses for certain metals,  pesticides and oils and
     greases  should  be performed.

 A.   In a review of  past studies the most commonly used parameters for
     describing storm generated discharges  were BOD and SS,  followed by
     total  coliform, COD,  Kjeldahl nitrogen,  total phosphorus and fecal
     coliform.  In studies evaluating treatment processes, the BOD
     analysis was used considerably more than the COD analysis, but for
     studies  on the  receiving water quality,  the BOD and COD tests were
     used about the  same.

 5-   Because  of the  high percentage of potential oxygen demanding
     materials that  are in the particulate form and the possible
     presence of oils and toxic materials together with the possible
     presence of materials such as silt, dirt, and wood that may not
     exert an immediate oxygen demand but will have an ultimate demand,
     the conventional oxygen demand tests can be seriously affected.

 6.   When determining the effectiveness of an oxygen demand indicator, it
     must be  considered that the material may be in the aquatic environ-
     ment for a long period, thus an oxygen demand test of limited time
     is not appropriate.

 7-   Although the conventional BOD(- test is considered to have many
     disadvantages,  it has been recommended for use because of the
     existing historical data available, regulatory requirements, and the
     wide knowledge and utilization of the test.

 8.   The range of particulate concentration can vary greatly in storm
     generated discharges, depending upon the characteristics of the
     drainage area,  the intensity and duration of the rainfall,  and the
     time between rainfall events.

 9.   In most  applications, the removal of the entire suspended residue
     fraction from combined sewer overflows and storm runoff would be
     considered an ideal achievement or goal.  Anything less than this
     total would be a partial success which can be effectively evaluated
     by making the appropriate suspended residue analyses.

10.   It is recommended that the settleable residue be determined for
     those applications where gravity separation is involved, or where
     an assessment must be made of bottom deposit buildup directly below
     or surrounding  an outfall.

11.   The 4.7  cm glass fiber filter is recommended for the suspended
     residue  analysis because of  its lower cost (when compared to
     cellulose acetate filters), ease of use, and resistance to quick
     blinding.  The  major disadvantage is that the filter must be pre-
     washed to remove fines prior to use.  The likelihood of an

-------
      appreciable amount  of  sand and grit  in storm generated discharge
      samples precludes  the  use of pipettes  in aliquot transfer.

 12.   Although volatile suspended residue  is not recommended for every
      analysis,  it can be used  in certain  situations for additional  in-
      formation  and if so,  it should be  insured that a large enough
      sample is  used  so that enough weight loss can be measured and  that
      a high degree of accuracy is achieved  in the gravimetric analysis.

 13.   Because of a blinding  problem, the "weight" method as found  in
      Standard Methods should be used  instead of the Imhoff Cone method
      for settleable  residue determination.

 1A.   It has been demonstrated historically that the only practical way
      of making  an overall health assessment in water supply and
      pollution  control  is through the use of an indicator organism  rather
      than analyzing  for  the actual pathogens themselves.

 15.   The most commonly used indicators of pathogenic pollution are  the
      total  coliform, fecal  coliform,and  fecal streptococci.   All  of the
      above analyses  have disadvantages, the most common is the lack of
      positive correlation with the presence of specific pathogens
      themselves.

 16.   Until  a  better  indicator system becomes available, the assumption
      can  be made  that if fecal coliforms are absent so are the pathogens
      of  all waterborne diseases,  including viruses.  Likewise, if fecal
      coliforms  are present, the likelihood of some microbial pathogens
      being  present is high, particularly bacteria of the salmonella
      group.

 !?•   The  use of the fecal coliform analysis is sufficient for character-
      izing  the  effectiveness of pathogen destruction of any treatment
      process applied to either combined sewer overflows or stormwater
      runoff.

 18.   In the case of storm generated discharges,  gross bacterial counts
      indicating relative differences are usually more important than the
     absolute numbers.   Thus,  it is more  important that the procedure
     employed be convenient to run, economical,  and one in which
     results are obtained as quickly as possible, rather than one where
     accuracy is the main criterion.  Of course, this is not true in
     studies  in the areas of research or public health where precision
     is required.

19.  Optimum sampling for bacteriological analysis is to collect  a
     discrete or grab sample in a sterile bottle and  return  to the
     laboratory  for immediate  analysis.   Whenever automatic  samplers are
     used, the sampler  head and sample lines should be disinfected.

-------
20.  Samples for fecal coliform analysis should be chilled at 10° C or
     below during sample collection, storage and transit to the laborato-
     ry, and refrigerated at the laboratory for less than two hours
     before processing.  Freezing of the samples shall be avoided,
     since this will cause a destruction of the microorganisms.

21.  When utilizing 2k hour composite samples, it is obvious that the
     first sample will have been stored 2k hours before transport to
     the laboratory begins.  Die-off of microorganisms can be kept to
     a minimum by proper cooling of the sample during the entire time
     the composite is  stored in the sampler.   Relative gross  counts  will
     be unaffected as  long  as  all  samples  are  handled in  a  similar
     fashion.

22.  The membrane filter technique for fecal coliform and fecal strep-
     tococcus analyses has many advantages over the MPN technique, even
     for the case of chlorinated samples.

23.  From available evidence at the present time, it appears that either
     nitrogen and/or phosphorus are the limiting nutrients, and as such
     both should be used as a measure of eutrophicat ion potential.
2k.  Analysis of both the oxidized form of nitrogen (NO^ and NO?) , by
     the superior nitrite analytical method, and the reduced form of
     nitrogen (ammonia and organic) by the Kjeldahl procedure are
     recommended .

25-  Because of the many forms  in which the phosphorus fraction can
     occur  (up to fourteen), the total phosphorus analysis is rec-
     ommended utilizing the persulfate digestion method.

26.  When sampling for heavy metals, the bottles must be absolutely
     clean and inert to the metal being determined, all glassware must
     be scrupulously clean, extreme care must be taken in analytical
     preparation, and the digestion procedures must be followed careful
     ly to  insure no loss of metal.  These steps are critical due to
     the fact that the metals may be found in the yg/l (ppb)  range, and
     slight analytical errors can greatly affect the concentration
     determinations.

27-  It is  recommended that composite samples of storm generated dis-
     charges be analyzed for lead, zinc, copper, chromium, mercury,
     cadmium, arsenic, nickel, and tin four times per year (seasonally)
     Based upon these tests, decisions can be made on the need for
     routine analysis.

28.  Analyses of samples for heavy metals should always include the
     total amount of the metal present and the samples digested
     prior to analysis.   Samples for metal analysis can be preserved
     with nitric acid to pH2 and refrigerated for up to 6 months.

-------
29.  For various reasons, no pesticides are recommended for routine
     analysis.  However, it is recommended that when evaluating the
     quality of a storm generated discharge,  a study of the drainage
     areas should be made to determine the possibility and type of
     pesticide use.

30.  Other parameters which may be of interest in storm generated
     discharge studies for various reasons include asbestos, asphalt
     and road materials, color,debris and bulk solids, density, specific
     gravity, solids settling  velocity distribution, oils and grease,
     odor,  particle size distribution,  pH, rubber,  specific conductance,
     sulfates, temperatures and trace organics.   Of the above, only pH is
     recommended for routine analysis.

31.  The sludges produced from treatment of storm generated discharges
     should be measured for total  residue if  concentrated and suspended
     residue if dilute.  Various  other analyses  may be required depend-
     ing upon the method of sludge treatment  and disposal.
DESCRIBING STORM GENERATED DISCHARGES

 1.   The following parameters  are  recommended  for  presentation and
     quantification when  reporting background  data on  storm generated
     discharge projects:  (l) drainage  area,  (2)  land usage, (3)  popula-
     tion density, (^)  median  residential  incomes, (5)  climate,  (6)  top-
     ography,  (7)  pervious and impervious  areas,  (8) street and  curb
     miles,  (9)  average daily  traffic,  (10)  methods and frequency of
     street  cleaning, and  (11)  lengths,  sizes  and  slopes of sewers.

 2.   Generally,  storm generated  discharges monitoring  programs fall  into
     one of  the  following  categories:

       - Evaluation  of  the impact  on receiving water quality
       - Facilities  design studies
       - Process evaluation
       - Compliance  with regulatory requirements

3-  The evaluation  of  the impact  on receiving water quality almost
    always  requires the generation of new data since most  of the
    existing data  is gathered regardless  of storm occurrence.

4.   In  facilities design  programs, special  characteristics  must  be
    monitored because of  the difference in  characteristics  between
    storm generated discharges and dry weather domestic sewage.

5.  It  is anticipated that in the future, storm generated discharges
    will require some type of monitoring  in which case automatic
    sampling equipment and composite samples  should prove to be
    adequate.

-------
METHODS FOR EVALUATION OF TREATMENT PROCESSES FOR STORM
GENERATED DISCHARGES

1.  Existing methods of storm generated flow treatment can be
    generally classified as follows:

       1.  Storage

           a.   ln-1ine
           b.  Off-line (with and without sedimentation)

       2.  Physical and Physical/Chemical

           a.  Coarse Screening
           b.  Fine Screening
           c.  Gravity Sedimentation
           d.  Swirl Concentration
           e.  Tube Settling
           f.  Sedimentation-Fi1tration-Adsorption
           g.  High Rate  Filtration
           h.  Adsorption-Coagulation-Tube Settling-
               Mult i Media  Filtration
           i.  Coagulation-Tube Settling
           j.  Series High  Rate Filtration
           k.  Screening/Dissolved-Air Flotation

       3.  Biological

           a.  Contact Stabilization
           b.  Trickling  Filter
           c.   Biological Contactors
           d.   Lagoons

       k.  Disinfection

           a.  Chlorination
           b.   Sodium and Calcium Hypochlorite
           c.  Chlorine Dioxide
           d.  Ozonation


2.  When evaluating a prototype treatment system, the entire pollutions!
    load,  including bypass  must be considered.   However, when evaluating
    a "pilot" type  system,  only the efficiency of the process units
    themselves should be  considered.

3.  When evaluating a treatment system, various  types of efficiencies
    may be determined.  These  include volumetric efficiency  (how much of
    the contaminant(s) is prevented from overflowing to the  receiving
    stream?) and  treatment  efficiency  (what portion of the contaminant(s)
    is removed  in each unit process of the treatment scheme?).  Both
    volumetric efficiency and overall efficiency must take into account

-------
    all flows bypassing the storage/treatment unit.

*».  In order to determine the efficiency of a system, it is recommended
    that samples be taken at constant time intervals and when composited,
    they may be proportioned either according to the flow rate at the
    time the sample was taken or according to the volume of flow since
    the last time the sample was taken.  Either of these methods will
    provide a sufficiently representative sample.  A third method of
    sampling and compositing that is sufficient is to take samples of
    equal  volume at equal intervals of flow volume.  However, this type
    of sampling requires a more flexible sampler since the time interval
    for sampling will be variable.  The ideal method of sampling would
    be to draw a sample continuously, with the flow of sample being
    proportional to the flow rate.  However, an automatic sampler of
    this type with proven field reliability for storm generated
    discharge studies  may not be practical to require at this time.

5.  For almost all types of treatment systems, only two sampling and
    flow measurement locations are needed.  These are at the discharge
    point itself and at the effluent.  The bypass can be calculated as
    the difference of the two.  To determine the total mass  of a
    contaminant removed, a single composite of the influent and effluent
    is necessary.  With this type of sample, the determination of over-
    all efficiency is relatively simple.

6.  Efficiencies of treatment plant performance should be calculated as
    an arithmetic average of percent removals achieved and as the
    weighted average based upon total mass treated and total mass
    removed during the period of study.

7.  When evaluating the efficiency of a storm generated discharge treat-
    ment system, any deleterious effect on the dry-weather treatment
    plant or any pollutants removed at the discharge site that escape
    in the effluent of the dry-weather plant must be subtracted from the
    treatment efficiency.

8.  So that all treatment processes evaluated may be compared on an equal
    basis, the following data should be reported following an eval-
    uation study:

         1.  Rainfall data
         2.  Drainage area description
         3.  Sewerage system data
         4.  Physical description of the treatment system
         5.  System operation
         6.  System costs
         7.  Sludge handling facilities
         8.  Dry-weather treatment plant data
         9.  Combined sewage data (for combined systems only)

     The general information required is the same for all  types of treat-
     ment  systems.


                                    10

-------
                     SECTION II - RECOMMENDATIONS
It is recommended that:

1.  The standardized procedures contained in this report be utilized
    in all future projects dealing with storm generated discharge.

2.  An automatic sampler, suited specifically for sampling storm
    generated discharge, be developed.

3.  The recommended sampling programs as found in this report be made
    more specific as to number of samples, sample frequency, duration of
    sampling program, etc., by means of a study based primarily upon a
    statistical approach.

k.  Continued research and studies be undertaken on the development/
    improvement of wide range flow measuring devices for sewer applica-
    tions.

5.  A standardized procedure for determining the number of raingages
    per area to be used in storm generated discharge projects be
    developed.

6.  Further work on sample preservation methods be conducted, with
    emphasis on the maximum holding time for the BOD analysis and on
    the feasibility of freezing as a preservation means.

7-  Further study be performed on the development of an oxygen demand
    potential analytical technique that fulfills the criteria for the
    ideal test.

8.  The long term oxygen demand of a number of storm and combined sewer
    overflow discharges be determined in an effort to develop firm  data
    on the ratio of the long term to BODr values in storm generated
    discharges.

9.  Analytical  techniques be developed for easier and more definite
    detection of specific pathogenic organisms themselves or for
    pathogenic indicators having a more positive correlation with
    pathogens themselves.
                                   11

-------
 10.  Research, development, and demonstration of new more reliable,
     interference free instruments, especially for organic loading, sus-
     pended solids, total phosphorus, and pathogenic indication be
     conducted.

 11.  A study be conducted to determine the true expected life of facil-
     ities built for storm generated discharge applications.   Because of
     periodic usage, the 20 or 25 year amortization periods may be too
     short and can result in extremely high unit operating costs.

 12.  Since the findings  presented in this report are based on knowledge
     and experiences to  date,  future changes can be expected  as addi-
     tional  research and  experience occurs.

13-  Immediate action  be  undertaken to determine how storm generated
     discharges will  be monitored when they eventually  fall  under  some
     type of regulatory  restrictions.
                                 12

-------
                      SECTION II I  - INTRODUCTION
This study, EPA Contract No. 68-03-0335, having the exact title,
"Standardize and Universalize Procedures for the Analysis/Evaluation of
Stormwater and Combined Sewage Characteristics/Treatment" was begun in
June 1973 and completed in December 1974.  The purpose of the project
was to develop standard procedures  recommended for use when conducting
projects associated with the evaluation and abatement of storm generated
d i scha rge pol1ut ion.

The results of this project, as presented in this report, will have an
extremely significant  impact on all ongoing and future projects deal ing
with storm generated discharges.  Development of these recommended
standard procedures was necessary because of the wide variation in
methods being used to  sample, analyze, and evaluate storm generated
discharges and associated treatment processes.  Not only were the
standard procedures needed to insure that all discharges and treatment
processes be evaluated by the same means, but also because these
procedures were found  to be the most applicable and optimum for use in
this field.

The project efforts were devoted to the major areas 1isted below.

     1.  Recommended methods for sampling and sample preservation.
     2.  Appropriate monitoring instrumentation available.
     3.  The choice of quality parameters to be utilized.
     4.  The analytical procedures to be followed.
     5.  The methods for evaluating storm generated discharge pollution.
     6.  The standard  procedures for evaluating treatment processes
         treating storm generated discharges.

The recommended method of sampling was necessary because of the many
different methods presently being used, that is, grab, volume
proportional to flow rate at equal time intervals, equal volume at
constant flow increment, etc.  The need for a uniform sampling program
is pointed out and the rationale used in selecting sampling procedures
is discussed.
                                   13

-------
 An entire section was devoted to montirong  instrumentation  finding  use
 in storm generated discharge projects.   This  included  level  and flow
 measurement instrumentation, _m situ parameter  monitoring equipment and
 raingages.

 The choice of quality parameters was divided  into seven  general  areas.
 These were:

      1.  Oxygen demand potential  indicator.
      2.  Particulate concentration.
      3.  Pathogenic indicator.
      4.  Eutrophication potential.

      5-  Heavy metals.
      6.  Pesticide and Polychlorinated  Biphenyls.

      7.  Other characteristics.

 The report also discusses  the relative  merits and disadvantages  of  the
 BODij,  BOD20,  BODX,  COD,  TOC,  and  TOD tests, and explains why the TOD was
 chosen  as best for use in  dealing with  storm and combined sewer
 discharges.   The same is done for the other pollution  parameters.   For
 example,  an extensive literature search was necessary  to document the
 choice  of fecal  coliform as  the best indicator  of the  presence  of
 pathogens.  This test was  found to be more advantageous  than the total
 coliform, fecal  streptococcus or  analysis for any specific  pathogen.  If
 uniform quality  parameters are to be used, they only have meaning for
 comparison  if  the  laboratory  procedures used  in analysis are the same.
 Therefore,  for each  quality  parameter selected, the exact laboratory
 procedures are not only  spelled  out  for common  analyses  such as
 suspended solids but also  for the more  exotic heavy metals  and  pesticide
 analyses.

 The method of  determining  the impact or pollutional load of  storm
 generated discharges  concerns itself primarily  with the  units of
 description for  a  discharge.   Various reports of storm generated
 discharge quality  use units  of  mg/1,  kg/min (Ibs/min), kg/year  (Ibs/yr),
 and so  forth.  Also,  some  studies report pollutional loadings based  on
 source  loadings  such  as kg/curb  m  (Ib/curb mile),  kg/ha/day  (Ibs/acre/
 day) and  g/axle  m  (Ib/axle mile).  Not  only does  this make relative
 comparison difficult,  but  the periodic  frequency  of these discharges
 and their acute  loading  makes description such  as  mg/1  and kg/year
 (Ibs/year) useless for any purpose other than the  specific design of a
 treatment process  component once  the  basic process has been selected.
Therefore, the description which  provides a common base for  comparison
of all discharges and  indicates  the  possible effect of  a discharge on
a receiving body of water  is  presented  and explained.
                                   14

-------
Finally, since many municipalities and industries are beginning to plan
for implementation of some type of treatment process to abate their
storm generated discharges, it is imperative that process efficiency be
determined by the same means.   Although it is obvious that true effi-
ciency  is best measured by the difference of total mass into and out of
a process, many other considerations must also be made.  The report
discusses these.  For example, true efficiency must include the contam-
inants  bypassed when the process is running at full capacity.  If
captured flow is bled back to a conventional treatment plant, then the
effect  of this extra flow on plant performance must be calculated.  In
essence, this discussion specifies the precautions that must be taken
to insure that the true net process efficiency is actually determined
and reported.
                                   15

-------
           SECTION  IV  -  SAMPLING PROCEDURES AND CONSIDERATIONS


 SAMPLING STORM GENERATED  DISCHARGES

 Purpose of Water and  Wastewater Sampling

 Sampling can  be defined as  the act of obtaining a small volume of water
 that  represents a much  larger body or flow or water, e.g., a  lake or
 wastewater treatment  plant  influent or effluent.  The water quality of
 this  small  volume  is  then characterized  in terms of BOD,  TOD, suspended
 solids,  etc.,  and  is  assumed to be the same as the larger body of water
 from  which the sample was obtained.  This water quality information is
 then  used  to  accomplish one or more of the purposes listed below.

      1.   Problem identification.

      2.  Wastewater treatment plant evaluation and process control.
      3.  Treatment plant  design - modification.

      4.  Receiving water quality determination.

      5.  Criteria to  be used for enforcement of water quality and
         effluent standards.

      6.  Determination of waterborne health hazards.

      7.  Application  to basic research.

      8.  Regulatory agency  requirements.

      9.  Others.


The key  to accomplishing  these sampling objectives is to  obtain a
"representative" sample; one that reflects the true condition of the
larger water volume from which the sample is taken.  This is a difficult
task  and will  be discussed  in some detail in this section.  Improper
sampling can  lead to  very serious consequences, in that no reliable
decisions  regarding problem solving can be made if the samples do not
reflect  the quality of the  larger volume of water.  There are a variety
of methods or  guidelines available for accomplishing reliable sampling,
dependent on the situation encountered.  In order to obtain samples
that will accurately  describe the main body of water it is necessary to
                                  16

-------
know something about the water characteristics being sampled.  Is the
water highly polluted with BOD, TOD and suspended solids or toxic
substances?  How do the wastewater or stream characteristics vary with
time or location in the stream?  How does the flow or volume vary?  Are
floating materials present?  Are heavier materials slowly being carried
along the bottom of the stream flow?  There are many other qualitative
characteristics that will be helpful in accurately sampling a stream of
water or wastewater.  Some of these important characteristics will be
known from an understanding of the origin of the water flow.  Municipal
and industrial wastewaters have certain known characteristics and flow
fluctuation trends that can be used to help obtain representative
samples.  In many cases it may be necessary to determine some of these
water quality trends by means of a well scheduled series of grab samples
before  the sampling program is initiated.  Storm generated discharge
sampling  is complicated because the source of pollutants that character-
ize the wastewater may contain both municipal and industrial wastes.
The water quality may be affected by other factors such as street
cleaning  frequency.  It will also depend on rainfall intensity and dura-
tion, antecedent rainfall, degree of urbanization, geologic and
hydraulic characteristics of the area, the sewerage system and many
other factors.

Characteristics of Storm Generated Discharges

Variation of Wastewater Flow and Contaminant Concentration -  It  is well
documented that the quality and flow rate of a storm generated discharge
at a specific location vary significantly during a storm event (1)(2)(3).
Figures  IV-1 and IV-2 show typical variations in storm generated
discharges.  The greatest concentration of suspended solids and oxygen
demanding material often occurs during the early period of the storm
runoff  as seen  in these data.  As the storm generated discharge causes
increased flow  in the sewers, accumulated dry weather solids are flushed
from the  sewers, and washed or eroded from the tributary land areas.
These  initial flows containing high concentrations of contaminants are
often  referred  to as the "first flush".  Contaminants may increase  in
concentration as continued flows tend to flush out the material  that has
settled out  in  the sewers since the last rainfall.

When rainfall events occur close together the runoff may keep the sewer
lines and tributary areas clean enough so that first flush characteris-
tics may  not be observed  (k).  Figure  IV-3 shows additional data from
the U.S.  EPA sponsored combined sewer overflow project  in Racine,
Wisconsin (5) that demonstrates the first flush phenomenon.   It can be
appreciated  that taking one discrete grab sample during this storm event
would  tell nothing about  the average sewage quality, and there is some
question  as  to  the meaning of  the average of these data.  Sampling  tech-
niques  will  be  discussed  later.
                                    17

-------
              800

              700

              6oo

              5oo

              1,00

              300

              200

              100
                     2300
                            2330    2WO
                                         0030
COD
BOD
                     2300     2330    2^00    0030
                     2300
                           2330
                                  2
-------
« "°
£ 100
o
o
v>
   90
   70
  J°°
§ 200
  JOOO
  2000
 .1000
          0300
          0800
0830
                               0900
                                         0930
                     0830
                              0300
                                         0330
                                                   1000
                                                   1000
1030
                                                              1030
          0800       0830       0900        C930       1000
                         Time (clock hours)
                                                              >TS
                                                             _L*JS
                                                              1030
       Figure  IV-2.   Example of  variance  in  runoff quality
          and quantity with time,  taken from  reference  2
                                  19

-------
   600 r
   500
01  400
to
Q
"»  300

Q
U)
Q
2
LU
O.

ID
   200
    100
                30
60        90       120       150       180

    TIME, minutes since start of overflow
210
2^0
 Figure IV-3.   Suspended solids variation with time, Racine, Wisconsin  (EPA Grant No. S80074M

-------
The concentrations of pollutants may vary across the section of the
sewer or stream.  Factors influencing this concentration variability
include the degree of turbulence in the wastewater stream and the fact
that velocities naturally vary within the cross section due to friction
losses and boundary effects, density currents, temperature differences,
the settling velocity and structure of suspended solids particles, etc.
All these factors tend to influence the concentration of pollutants
across the sewer cross section.  Understanding these factors helps to
appreciate the difficulties in sampling storm generated discharges
representatively.

Where two streams or sewers come together density currents may occur
such that pollutant concentrations may vary significantly throughout the
cross section of the sewer.  Sampling should be done downstream from
these points such that sufficient turbulence and time is allowed for
thorough mixing of the two stream flows.

The more turbulence there is the better the mixing will be and hence
the sooner sampling will give a representative sample characteristic of
the sewage.  The more thoroughly mixed the wastewater, the easier it is
to obtain a representative sample of the cross section.  It may be
necessary to obtain a series of grab samples along the length of the
sewer and at different points  in each cross section in order to
establish where permanent sampling stations should be located.

Irregular Event Occurrence - Storm generated discharges by definition
occur when a sufficient amount of rainfall has fallen to cause runoff.
As will be discussed later, rainfall can vary significantly from both a
time and spatial standpoint.  Therefore, the storm generated discharges
to be sampled will also occur periodically.  Because of the variable
nature of these discharges, it is difficult to have personnel immediately
available at preselected sampling sites to manually begin taking
samples or to turn on automatic samplers.  At times the duration of an
overflow event may be so short that the designated sampling personnel
may not reach the site until the discharge is completed, or at least
well past the first flush occurrence.  Conversely, long duration, low
intensity storms may result in discharges lasting 2k hours or more.
Consequently, it is desirable that automatic samplers be employed for
storm generated discharge sampling.  These samplers should be capable
of starting their sampling cycle automatically when a rainfall produc-
ing runoff occurs.  In this manner, all discharge events will be
sampled from the beginning of discharge through the entire duration of
the event.  This assumes the samplers are reliable and the proper
preventative maintenance has been performed.  In some cases the storm
may last sufficiently long so that all the sample containers are filled.
The operating personnel should be aware of the approximate starting
time of each storm so that sampler containers can be replaced when
necessary, or relief or standby samplers put into operation for a
long duration event.
                                    21

-------
 a
Floating Substances and  Coarse  Bottom  Sediments - Contaminants that have
a specific gravity significantly  different  from water cause special
problems in sampling.  This  has recently  been discussed.

     "Suspended solids heavier  than water have their  lowest
     concentrations near  the surface and  the concentration  increases
     with depth.   Near the bottom of the  sewer may occur a  'bed  load1
     composed  almost entirely of  heavier  solids. 'This may  'slide1
     along the bottom, or, with insufficient flow velocity, may  rest
     on the bottom.  As  the  velocity and  turbulence  increase,  the  'bed
     load1  may be  picked  up  and suspended in the sewage.

     "At the beginning of storm runoff, as  water picks up solids which
     have accumulated  in  the sewer upstream during periods of  no rain-
     fall,  the flow may  be composed largely of sewage solids,  or 'bed
     load1,  which  appears to be pushed ahead by the water.

     "Suspended materials lighter than water, such as oils and greases,
     float  on  the  surface, as do  leaves,  limbs, boards, bottles, and
     cloth  and paper materials.   Other small, light particles  are
     moved  randomly within the  flow by turbulence.  These may  be well
     distributed  in the  cross-section without significant effect of
     variable  velocity within the section."
Unless a sampler  intake  is  located near  the surface of the wastewater
and/or the bottom of a storm or combined sewer  flow having floating
materials or coarse sediments, these sewage constituents will not be
present  in the samples taken.  Consequently,  important pollutants that
can cause significant stream quality deterioration, or have major
effects on the treatment plant may not be  taken  into account.  Oils and
grease, coarse floatables,  and heavy solids may  cause unpredicted
overloaded conditions in the treatment plant  (grit can present specific
problems) and deposition or unsightly conditions  in stream beds.
Special methods can be adopted to specifically  sample for these consti-
tuents so that their quantity and quality  can be  determined and it can
be decided whether or not a problem with these materials exists.  Grab
samples may be used to determine the presence of  these types of
materials.  If floating or  especially heavy materials are frequently
found in the storm generated discharges  special  samplers may have to be
designed based on the type of pollutants that is  most significant.
Alternately, improved sampler inlet devices should be developed.

The relative amounts of these materials can be determined by placing a
representative sample of storm generated discharge in a one liter
graduated cylinder and allowing it to settle for  a predetermined time
interval.  The volume of settled and floated solids can be measured.
The settling time interval  should be relatively short so that the
condition in the graduated cylinder resembles the sewer rather than the
sedimentation basin.  However,  as discussed in a  later part of this
report, this test is not quantitative,  and  is  only intended to provide
an estimation of floatable or settleable water quality.
                                   22

-------
Excess grit from storm flow can be determined by measuring grit
removed at the sewage treatment plant grit chambers during wet-
weather flow periods.

Suspended Solid Particle Size and Sampling Velocity - The size and
specific gravity of the suspended solid particulates in storm generated
discharges influence the type of sampling equipment used and location
of intake orifices; the sampling velocity (velocity through sampler
orifice, tubing, and other critical points)  and intake orifice
dimensions directly influence the size of suspended solids that will be
collected.  When particle size and specific gravity is sufficiently
large for suspended materials to settle out during an overflow discharge,
bottom sludge may build up in the sewer or move slowly along the
bottom of the sewer.  If an increased rainfall intensity causes higher
velocities in the sewer, a resuspension or scouring may occur.  The
sampling velocity must be sufficiently high to ensure that the most
rapidly settling particle will not settle out in the sample tubing or in
other sections of the sampler train.  Otherwise, some of the suspended
solids will not reach the sample container and a representative sample
will not be obtained.  If any particles are found settling out in the
sample lines, the flow velocities must be increased or a more suitable
sampler should be used.

It has been determined that the velocity of sewage flowing through the
sampling orifice influences the concentration of suspended solids
obtained  in the sample (6).  This has a greater effect in air sampling
because of the large difference in densities between air and the
particles.  However, a similar phenomenon occurs in water samples.
This is shown in Figure \\l-k and is a result of the larger, heavier
particles continuing to flow  in a straight line path while the smaller
particles tend to follow changes in flow caused by the sampling orifice.
To eliminate these  interferences the sampling velocity should be the
same as the velocity in the stream being sampled.  This intake velocity
is apparently more  critical than the angle of the intake orifice with
respect to the flow direction.


Sample Site Consideration

When establishing a sampling station for water quality monitoring or
storm generated discharge sampling, it is important that the location
can easily accommodate an automatic sampler and provide enough
accessibility to maintain and operate the samplers.  If the sampler is
located in the sewer or low in the manhole chimney, then the sampler
should be waterproof, as the sewer can be expected to flow completely or
surcharge full during some storm events.  Location in a sewer or man-
hole is advantageous from the standpoint that it is out of sight and
will not attract vandals, though this sacrifices some advantages  in
ease of maintenance and accessibility.  After each storm event, an
automatic sampler must have clean bottles installed, be reprogrammed
                                   23

-------
                               Velocity   2000  ft/min
•y_	veioc I ty=
          ft/mini
                               Nozzle
               Gas  Stream
                            Velocity=1000 ft/min
                                   I I
                             =2000 ft/min
                                                          V  < V
                                                           s    w
                                  V  > V
                                   s    w
                Gas'Stream     Nozzle
       1.   When sampling  velocity  (Vs)  is  less  than  water  velocity  (Vw),
           then a  disproportionate  number  of  large participates  enter  the
           sampling  orifice,  and a  number  of  small particulates  that should
           enter the orifice  bypass it.

       2.   When V  > Vw,  a  disproportionate number of  small  particulates enter
           the sampling orifice, and  a  number of  large particulates  that
           should  enter the orifice bypass it.


          Figure \l\-k.   Influence of  sampling velocity on  suspended  solids
                                (from Reference 6)


and readied for the  next storm event.  Also, if  bacteriological tests are
to be performed, cleaning  and disinfection of  the sampling  lines is
recommended.   Consequently,  if it  is located in  a hard to reach area, it
will be difficult to perform these necessary functions, unless  it is
portable and can be  removed  from its sampling  position.

The suitability of  a storm generated discharge sampling site  is largely
determined by  how representative the samples are; the  samples should
have the same characteristics as larger' flow streams.   Generally,  it is
recommended that in:

     ". .  .sewers and in deep, narrow channels,  samples should  be
     taken from a point one-third  the water  depth from the  bottom.
     The velocity of flow  at the sample point  should,  at all  times,
     be sufficient  to prevent deposition  of  solids.  When collecting
     samples,  care  should  be taken to avoid  creating excessive
     turbulence which may  liberate dissolved gases  and yield  an un-
     representative  sample.  Additional considerations are  as
     follows:   (1)  the site  should provide maximum  accessibility
     and safety; (2) it should be  a  sufficient distance downstream
     from  the  nearest tributary inlet to  ensure  complete mixing of
     the two flows;  and  (3)  there  should  be a  straight length of
     pipe at least  7 sewer diameters upstream of the site." (7)
                                     24

-------
If all the above conditions are met, then one can be reasonably certain
that the flow is of relatively uniform quality across the sewer cross
section, thus making it easier to obtain a "representative" sample.
Although it is not always possible to meet the above conditions for
selecting a sampling site, these requirements should be followed as
closely as possible.

Samplers that are located in sewers will be exposed to very humid and
corrosive conditions.  Thus, the sampler should be constructed of non-
corrodible material, such as fiberglass.  If the sampler is capable of
corroding, there is a good possibility that any samples obtained will be
contaminated by the corrosion products, or samples may be lost completely.
If the sampler is not located in a sewer or manhole, but instead is
located at ground level,  it is necessary that some type of housing be
available to protect against both climate conditions and vandalism.
Common types of housing  which are sufficient include such things as
large diameter concrete sewer tiles and commercially available metal
"garden" sheds.

Just as  important as the consideration of sampler housing is the method
of protection of the sampler tubing leading from the sampler to the
sampler  intake.  Sample tubing located  in the sewer itself must be
protected from objects such as wood, branches, gutters, etc. contained
in the flow, together with the high velocities experienced.  Also, this
tubing must be buried or otherwise protected (such as enclosure in
conduit), from the point of exit from the sampling house to the point of
entrance  in the sewer or manhole.

The sampler should also be able to withstand freezing conditions.   In
northern climates discharges caused by freezing rains or snow melt can
occur while the sampler is under a freezing condition.  Similarly,  if
the sampler is permanently installed, it must resist freezing conditions
during  inactive months.  The location of the sampler should minimize the
freezing of samples and be accessible to check for freezing conditions
which require frequent maintenance.


TYPES OF SAMPLING PROGRAMS

There are a variety of types of water samples that can be obtained;
however, each type belongs to one of two main categories.  These
categories can be classified as discrete or composite.  Discrete samples
characterize water quality for a particularly short period  in time.
Composite sampling  is an attempt to synthesize a sample which will
represent the average discharge characteristics over a period of time.
The various types of samples that can be obtained are shown in
Table IV-1  (4).

In the discrete sampling  techniques described above, the random grab
samples are the easiest and most economical, but the least  reliable  in
                                    25

-------
               Table  IV-1.   DIFFERENT TYPES OF SAMPLES
                           Discrete samples
  Random grab  - one sample obtained at any  time at any
               point  by any available method (Figure  IV-5).

  Unit  time  -  a series of samples  taken during a discharge  event
              at equal time  intervals  (Figure  IV-6).

  Unit   volume - a series of samples taken  during a discharge event
                at equal volume  increments.  A flow recorder and
                totalizer is  required for  this type of sampling.
                (See Figure  IV-7.)
	ComposIte samp 1es	

 Constant time - constant volume:  samples of equal volume are taken
 at equal increments of time and composited to make an average sample
 (similar to  Figure  IV-6).

 Constant time - volume proportional to flow  increment:  samples are
 taken at equal increments of time and are composited proportional
 to the volume of flow since the last sample was taken. (See
 Figure  IV-&)

 Constant time - volume proportional to flow  rate:   samples are
 taken at equal increments of time and are composited proportional
 to the flow  rate at the time each sample was taken. (See Figure  IV-9.)

 Constant volume - time proportional to flow volume increment:  samples
 of equal volume taken at equal  increment of flow volume and composited,
 (See  Figure  IV-10)
                                 26

-------
          DENOTES COLLECTION OF SAMPLES  OF
          DIFFERENT VOLUMES AT RANDOM  TIMES
         Figure IV-5.   Random grab sample  collection
5 -
2  •
1   .
                  DENOTES SAMPLES OF EQUAL
                 VOLUME (SAME LENGTH ARROWS)
                    AT EQUAL TIME INTERVALS
         Figure IV-6.  Method of compositing samples
         on a fixed volume-fixed time interval  basis
                           27

-------
o/,
                                TIME

       Figure  IV-7.   Method of  taking discrete  samples
           at  constant  flow volume  increments
                           TIME
           Figure  IV-8.  Method of compositing samples
     proportional  to flow volume at constant time interval
                               28

-------
                   DENOTES COLLECTION OF A SAMPLE
                 WHERE VOLUME IS PROPORTIONAL  TO THE
                RATE OF FLOW.  THE INDIVIDUAL  SAMPLES
                  ARE COMPOSITED INTO ONE CONTAINER
Figure IV-9.
                                TIME
                     Method of compositing  samples  proportional
                              to flow rate
LJ
                                       VARIABLE
                               DENOTES  SAMPLES  OF  EQUAL VOLUME
                               (SAME  LENGTH  ARROWS) AT CONSTANT
                               FLOW INCREMENTS  (VARIABLE TIME)
                            TIME

    Figure  IV-10.   Method  of compositing  samples of equal
              volume at  equal  increments  of  flow
                                29

-------
terms of representing the wastewater flow characteristics.   As
previously discussed, storm discharges can be expected to vary in both
flow and strength; obviously a random sampling procedure as shown in
Figure  IV-5 cannot possibly describe the wastewater characteristics.
However, when series of discrete samples are taken in chronological
order or for a fixed interval of waste flow as shown in Figure IV-6 and
IV-7, it is possible to determine the waste characteristics.  These two
types of discrete samples can be taken either manually or with automatic
samplers.  In order to accurately describe a waste discharge using
discrete samples, the time or flow interval must be sufficiently short
to obtain the important wastewater characteristics.  These two types of
sample  techniques are indispensable in describing the waste characteris-
tics throughout any particular storm event.  Peak or maximum wastewater
characteristics associated with the first flush can be determined and
frequently these characteristics are of major importance in storm
generated discharge systems.

Composite sampling will provide information about the average storm
generated discharge flow but not maximum or minimum values.  The constant
time-constant volume composite sample approach will provide information
about average characteristics only if the waste flow is relatively
constant.  Where the waste flow is not approximately constant, one of the
other types of compositing approaches must be used to determine average
wastewater characteristics.  A composite samp1 ing procedure may consist
of drawing discrete samples into individual containers which can be
added together manually, or a series of discrete samples may be automat-
ically  mixed in a single container to make up the composited sample,
The two main devices used to gather the wastewater samples are a dipping
mechanism and a pumping system.  The pumping type sample collection
systems use either a vacuum arrangement or some type of centrifuge or
positive displacement pump.  Details of different automatic samplers have
recently been discussed in reference 4.

In evaluation of a recent review of storm generated discharge wastes (k)
it was  found that discrete random grab samples were used in four dif-
ferent  studies, automatic-discrete samplers were used in five studies,
and automatic-composite samplers were used in six studies.  In four of
these projects, both automatic and manual sampling were used.  This
evaluation points out the importance of using discrete sequential samples
to obtain time-series data.  In one study, automatic-composite samples
were used in the first phase (8), and in a later phase of the project
automatic discrete samples were used (9) in order to obtain the detailed
characteristics of the storm generated discharges.

In other studies grab samples were used to verify discrete samples taken
with automatic samplers (3)(19)(11).  Another obvious advantage of
automatic samplers that obtain individual discrete samples (in some
cases sample volumes may be automatically proportioned to flow rate or
volume)  is that each individual sample may be examined (12).  In this
manner  it is often possible to visually determine if time series
                                    30

-------
analyses should be made or if composite sample analyses to determine the
average characteristics will be sufficient.

The various types of composite sampling techniques available have both
advantages and disadvantages when dealing with storm generated discharges.
The four types of sample compositing procedures have been evaluated using
a mathematical approach for five different concentration variations.  The
constant time-volume proportional to flow rate at the time of sample
drawing composite sampling approach was found to give a sample concentra-
tion most nearly the same as the true weighted average waste concentra-
tion.  However, the difference found between the variable volume or time
sampling techniques was quite small and in most cases any of these proce-
dures approached the true value of the stream characterized.

The time constant-volume proportional to flow rate is a commonly used
compositing method.  This type most closely approximates continuous
sampling proportional  to flow rate which is considered  ideal.  This
knowledge  indicates that the smaller the time increment between samples,
the more closely the sample will approximate the actual stream.  Another
advantage  to  this sample is that a sample need not be directly connected
to the flow monitor if discrete samples are taken.  The compositing
technique of  time constant-volume proportional to flow since last sample
can also be obtained without direct connection to a flow meter.  The
accuracy of samples taken would be dependent upon the time  interval and
the variability of the discharge.

The volume constant-time proportional to volume since the last sample is
more difficult  to obtain due to the necessity of a flow monitor
 incorporated  into the  sampling mechanism.  However, this may be a
valuable technique for some storm generated discharge sampling.  The
phenomenon of first flush, as discussed previously, may be  better
characterized using this technique because there may be more samples
taken when the  flow rate is high.  Therefore, the loading parameters,
which may  vary, will be approximated more closely.  Then as the flow
rate  lessens  the sample  interval will widen and only a  few  samples will
be taken of the tail end flow.  This method has inherent difficulties
due  to the lack of manual monitoring.  A major problem would be to
determine  the volume  increment of flow needed to obtain a representative
sample without  overruning the bottle supply or not collecting a suffi-
cient number  of samples for analyses.  This decision must be left to the
discretion of the operator.

There  is a need for the flow monitoring equipment to be reliable and
accurate so that proper composites can be obtained.  For some composite
types  (Figure IV-8 and IV-10) a totalizer or  integrator will be needed.
 It will also  be advantageous to take samples  into discrete  containers
for manual compositing and for those cases when discrete analyses are
performed.  As discussed later, this discrete sampling may  be neces-
sitated with  sampling  storm generated discharges having first flushes
of flows and  for some  physical or chemical treatment processes.
                                     31

-------
 Very little  information  is available concerning the use or success of
 automatic  switching devices  to activate samplers at the beginning of an
 event.   Two  studies have  reported successful use of rising water  levels
 to actuate switching devices and start sampling operations (13)0*0-   A
 third study  reported that their automatic starting devices did not work,
 and it  was necessary to manually start each sampler at the beginning of
 each rainfall  (10).  Satisfactory results have been obtained using a
 float switch in a wet well at Racine to initiate sampling  (15).  The
 same switching system also starts the treatment process through a series
 of time delay devices.  The advantages of automatic switching to start
 sampling equipment are obvious and should be used  in every case where
 possi ble.

 SAMPLE  HANDLING AND PRESERVATION

 Volume  of  Sample Required

 ASTM Dl*t96 specifies that at least 2 liters of sample should be collected
 for  evaluation of industrial wastewater.  It is also noted that A liters
 is  preferable and 20 liters may be necessary (16).  Tables have been
 published  to give an estimate of the sample volume that is necessary for
 any  given  analytical testing (k)(16).  Table IV-2  is reproduced for the
 convenience  of the reader in determining sample volumes.  Note that for
 total metals there is no preservative technique for hexavalent chromium.
 This  analysis must be performed within 2k hours.  Also, ferrous iron
 must  be  adjusted to pH <6 and filtered, followed by the addition of HNOo
 to the  filtrate to pH <3.  The actual volume needed depends on the
 characteristics of the water and the number and type of analyses that
 must  be  determined.  Using Table IV-2 for estimated sample size for
 testing  wastewaters for BOD, TOD (assuming same sample volume as
 required for organic carbon), fecal coliform, total phosphorus, nitrate,
 total Kjeldahl nitrogen, and suspended solids (see Section VI) would be
 2700 ml.   This number can be considered conservative (high) since
 1000 ml  each for the BOD and suspended solids tests are extremely high.
 It  is the  authors' opinion that all of these tests could easily be
 performed  with 2000 ml  of sample.  However, in the case of most samplers
when  the discrete  samples are composited, the composited volume will  be
many  liters.   It is good practice with storm generated discharges to
 bring approximately 4 liters (M gal.) of sample to the laboratory under
normal circumstances.

 Sample Container Characteristics

 Disposable containers may prove to be more easily used and economical
 than  reusable containers that must be cleaned after each use.  However,
 samplers that require special shaped containers will  likely necessitate
 the use of reusable containers.   Also, for such constituents  as oils
and grease,  pesticides, TOC, etc.,  plastic containers cannot  be used.  A
 study of plastic and glass containers has shown that  storage  in Pyrex
                                   32

-------
                                      Table  IV-2.   RECOMMENDED  PRESERVATION  METHODS
CO
CO


Parameter
Acidity-alkalinity
Biochemical Oxygen Demand (5-day)
Calcium
Chemical Oxygen Demand
Chloride
Bacteria - fecal coll form, total
coliform, or fecal streptococcus
Color
Cyanide
Dissolved Oxygen
Fluoride
Hardness
Metals, total*
Metals, dissolved
Nitrogen, ammonia
Nitrogen, Kjeldahl
Nitrogen, nitrate - nitrite
Oil and grease
Organic carbon (total and dissolved)
pH
Phenol I cs
Phosphorus
Solids (total, dissolved, suspended.
volatile)
Specific conductance
Sulfate
Sulflde
Threshold odor
Turbidity


Preservative
Refrigeration at 4°C
Refrigeration at 4°C
None required
2 ml HjSO^ per 1 Iter
None requl red
Maintain temperature as at source -
usually requires refrigeration
Refrigeration at 4°C
NaOH to pH 10
Determine on site
None required
None required
5 ml HNO, per liter
Filtrate: 3 ml 1:1 HNO, per liter
40 mg HgCl2 per liter at4°C
40 mg HgCI2 per liter at 4°C
40 mg HgCl2 per llterat4°C
2 ml H2SOJ( per liter at4°C
2 ml HjSO^ per liter (pH 2)
Determine on site
1 . 0 g CuSO^ .j + H PO^ to pH 4.0 - 4°C
40 mg HgCl2 per liter at4°C
None aval table

None required
Refrigeration at 4°C
2 ml Zn acetate per liter
Refrigeration at 4°C
None aval lable

Maximum
holding period
24 hours
6 hours
Indefinite
7 days
Indefinite
8 hours

24 hours
24 hours
No holding
7 days
7 days
6 months
6 months
7 days
Unstable
7 days
24 hours
7 days
No holding
24 hours
7 days
7 days

7 days
7 days
7 days
24 hours
7 days
Approximate
required
sample size, ml
50-100
1000
50
50
50
200

50
500
250-300
200-300
25
100
100
100
100
100
1000
100
~
500
200
1000

—
100
1000
200
1000
                 a.  Sum of  the concentrations of metals In both the dissolved and suspended fractions.

-------
 and polyethylene did  not  significantly change  the silica, sodium,  total
 alkalinity,  chloride,  boron,  specific conductance, pH or hardness  of
 water samples  for a storage period of about  5  months  (17)-  Soft glass
 containers  did show an increase  in silica after 2 to  3 weeks, and  after
 longer periods increased  sodium  and  hardness were found.  All new
 containers  should be  thoroughly  washed and allowed to soak  for  several
 days in order  to remove as much  water soluble  material as possible. Old
 beverage or  other similar type containers should not  be utilized for
 wastewater  sampling.

 The proper method for  cleaning sample containers will depend on the
 contaminants to  be analyzed.  A  good soap washing will normally be used
 first.   When samples are  to be tested for phosphorus, heavy metals,
 pesticides,  and  grease and oil additional procedures  should be  used (18)


                  Sample Container Cleaning Procedures

      1.   After soap washing,  use cleaning solution made by adding
          1  liter  of concentrated sulfuric acid to 32 ml of a
          saturated water  solution of sodium  dichromate.

      2.   Rinse out cleaning solution thoroughly using generous  amounts
          of  tap water  followed by distilled  or demineralized water.

      3.   For samples to be analyzed  for heavy metals, an additional
          nitric acid rinse is needed.  Rinse containers with a nitric
          acid  solution made up of 1  part concentrated nitric acid  and
          k parts  water.

      4.   Complete container preparation by thoroughly rinsing with
          distilled or  demineralized water followed by drying.

Containers to be  used when testing for pesticides and bacteriological
analyses  may require additional  special preparation as will the sample
lines themselves  (discussed later).

Sample Preservation

If possible, samples should be analyzed immediately after collection and
no preservative  is needed.  However, from a practical standpoint both
composite and a series of sequential discrete samples cannot be
analyzed  immediately.   In some cases, for example, samples taken for
bacteriological testing,  it is necessary to  take short term composite
or grab samples  in order  to ensure reliable  results [(14), page 81].  A
number of preservation methods have been suggested (1)(3)04)(15)(17)
(18)  for  different constituents.   Review of the methods of preservation
(from the above references) shows that conflicts exist as to the type of
preservation to use and the length of storage allowed before significant
changes occur  in  the sample.   For example,  maximum recommended  sample
                                   34

-------
holding times for the BOD range from 6 hours  (16) to 2k hours  (19) when
the sample has been stored at k°C.  If the maximum storage time  is
2k hours, it is obvious that 2k hour composite samples cannot  be used.
Recommended preservatives and holding times shown in Table IV-2 can be
used as guidelines for storm generated discharge analysis.  Where several
constituents that require different preservatives must be determined,
multiple samples will have to be collected.  The purpose of different
preservatives are shown in Table  IV~3 (20).  As can be seen in Table  IV-2,
the most universal preservation method for a  large number of tests is
refrigeration.   In all cases where storm generated discharge samples are
not analyzed immediately, refrigeration or icing should be used to
maintain samples at k°C.  Preservation techniques for the specific rec-
ommended quality characteristics, after the sample has arrived in the
laboratory, are  presented in Section VI.

Sample  Identification

Samples should be carefully  identified using  tags or by printing directly
on the  container wall with a grease pencil to ensure that the  date of
collection and the location of sample is known.  If plastic containers
are used, masking tape should be applied with the marking put  on the
tape so it can be reused.  Discrete samples should also be identified by
the time of their collection.  When preservatives are added prior to
sample  collection, containers should also be marked with the type of
preservative used and analysis to be run.  Code or abbreviations may be
used in sample identification, but these should be standardized with a
known format.
AUTOMATIC SAMPLING EQUIPMENT

Survey of Samplers

A  published U.S.  EPA  report  (k)  has  summarized  the characteristics of k]
automatic samplers commercially  available  from  18 different companies,
as well as  12  custom  designed  samplers.  Tables  \\l-k and  IV-5 present a
summary of  the sampler  characteristics outlined  in that report.
Table  \\l-k  lists  the  costs of  the  commercially  available  samplers and
distinguishes  what type of samples each  sampler  is capable of obtaining.
An extra sample type  has  been  added  to those  previously discussed — a
continuous  sample --  whereby the waste discharge has a sample stream
removed continuously  and  composited  in a suitable sample  container.
Those  samplers that could be modified to obtain a continuous flow type
sample have been  noted  by the  footnote letter "a".  Table IV-5  presents
the  same sampler's capabilities  with reference  to sample  preservation
by   refrigeration, applicability to  hostile environments, and sample
contamination  plausibility.
                                     35

-------
                Table  IV-3.   PURPOSE OF DIFFERENT PRESERVATIVES
 Preservative

HgCl2


Acid (HNO.)


Acid (H2SOA)
         Action
Bacterial  inhibitor
Metals solvent, prevents
   precipitation

Bacterial inhibitor
        Applicable to;
                  Salt formation with
                     organic bases

Alkali (NaOH)     Salt formation with
                     volatile compounds

Refrigeration or  Bacterial inhibitor
   freezing
Nitrogen forms,
Phosphorus forms

Metals
Organic samples (COD, oil
   & grease, organic carbon,
   etc.)

Ammonia, amines
                           Cyanides, organic acids
                           Acidity - alkalinity,
                              organic materials, BOD.,
                              color, odor, organic P,
                              organic N, carbon, etc.,
                              biological organisms
                              (coliform, etc.)
                                   36

-------
                           Table  IV-4.   AUTOMATIC  SAMPLER  COSTS AND  SAMPLE TYPES OBTAINABLE
CO
Sampler
BIF
Brailsford DC-F
Brailsford EV-P
BVS PP-100
BVS SE-1»00
BVS SE-600
Chicago "Tru-Test"
Hydra-Numatic
Inf i Ico
isco 1391
Lakeside T-2
Markland 1301°
Markland 101°
Markland 102C
Markland 10ATC
N-Con Surveyor
N-Con Scout
N-Con Sentry
N-Con Trebler
N-Con Sentinel
August 1972,
Cost - $
595
281
583
1,000
2,600
2,800
2,578
1,800
A.AOO
1,095
1,577
1,210
980
1,265
M15
275
*»50
895
1,560
2,350
Discrete/ Discrete/ Comp,/
time volume Flowrate



X
X
X
X
X
X
XX X
b
X
X


X
X

b
X
X
Comp./
vo 1 ume



X
X
X
X
X
X
X
X


X
X


X
Comp./
time Continuous
X
X
X
X
X
x a
X
a
X
X
X
X
X
X
X
X
X
X
X
X
       a.   Can be easily modified to sample continuously.
       b.   Flow must be proportional to depth as in a Parshall  Flume,
       c.   Representativeness of sample questioned in Reference 1.

-------
                    Table IV-4 (continued).   AUTOMATIC SAMPLER COSTS AND SAMPLE  TYPES  OBTAINABLE
to
00
Sampler
Phipps & Bird
Protech CG-125C
Protech CG-125FPC
Protech CG-150C
Protech CEL-300
Protech DEL-2405
QCEC CVU
QCEC E
SERCO NW-3°
SERCO TC-2
Sigmamotor WA-I
Sigmamotor WA-3
Sigmamotor WDPP-2C
Sigmamotor WM-1-24C
SIRCO B/ST-VS
SIRCO B/IE-VS0
SIRCO B/DR-VS
SIRCO PM-A
SONFORD HG-4°
TMI
TMI Mark 38
August 1972 Discrete/ Discrete/
Cost - $ time volume
1,H5
723
1,038
895
M50
5,606 x x
940
875
920 x
2,495
700
700
680
1,525 x x
2,950 x x
2,850 x x
2,640 x x
1,387
495
660
685 x
Comp . /
Flowrate
x

x
x
x
x
x
x

x


x

x
x
x

X


Comp./
volume
x

x
x
x
X
X
X

X


X

X
X
X

X


Comp . /
time Continuous
x
x
x
x
a
x
x a
x
x

x
x
x


x
x
x
x
x
x

       a.  Can be easily modified  to sample continuously.
       b.  Flow must be proportional to depth as  in a Parshall  Flume.
       c.  Representativeness of sample questioned  In Reference 1.

-------
                    Table IV-4 (continued).   AUTOMATIC SAMPLER COSTS AND SAMPLE TYPES OBTAINABLE
oo
Sampler
AVCOC
Springfield
Milk River
Envi rosenics
Ron re r 1
Weston
Pavia-Byrne
Rexnord
Colston
Rohrer 1 1
Near
Freeman
August 1972, Discrete Discrete Comp. Comp. Comp.
Costs - $ time vol. flowrate volume time Continuous
d
X
d
X
d
X XX
x - 1 sample
onTyK
d
X XX
d
X X
d
X
d
X
d
X
d
X X
d b
X XX
d
X
        a.   Can be easily modified to sample continuously.
        b.   Flow must be proportional  to depth  as  in a Parshall  Flume.
        c.   Representativeness of sample questioned in Reference 1.
        d.   Custom designed for sampling combined  sewage.

-------
                                        Table  IV-5.   AUTOMATIC  SAMPLER  CAPABILITIES AND SAMPLING METHODS
*-
Abi 1 i ty to sample


Sampler
BIF
Braitsford DC-F

Bra! Isford EV-F
BVS PP-100
BVS SE-lfOO
BVS SE-600



Chicago "Tru-Test"

Hydro-Numatlc

Infllco

ISCO 1391
Lakeside T-2
Markland 130la
Markland 101a
Markland I02a
Markland lO^T3
N-Con Surveyor
N-Con Scout
N-Con Sentry
N-Con Trebler


Method of sample transport
Dipper
Positive displacement pump

Vacuum pump
Pneumatic ejection
Submersible pump
None provided - solenoid
valve diversion from
sample linec

Dipper from sample chamber
provided by customer
Centrifugal pump

Dipper from sample chamber
provided by customer

Peristaltic pump
Dipper
Pneumatic ejection
Pneumatic ejection
Pneumatic ejection
Pneumatic ejection
Centrifugal pump
Peristaltic pump
Peristaltic pump
Dipper

Refrigeration
provided
yes
no

no
yes
yes


yes


yes
no

yes

Ice cavity
yes
no
yes
yes
yes
no
no
no
yes

Abl 1 Ity to
Freezing
no
no

no
yes
yes


yes


no
yes

no

no
no
yes
yes
yes
yes
no
no
no
no

withstand
Immersion
no
no

no
no
no


no


no
no

no

yes
no
no
no
no
no
no
no
no
no


Floatables
no
no

no
yes
no

b


b
b


b

yes
some
no
no
no
no

no
no
some
Coarse
bottom
sediments
no
no

no
no
no

b


b



b

no
some
no
no
no
no

no
no
some
Self-
cleaning
features
none
continuous
flow
backflush
air purge
continuous


continuous
flow

continuous
continuous
flow
cont i nuous
flow
backflush
none
none
none
ai r purge
none
gravity drain
backflush
backflush
none
     a.  Representativeness of sample questioned In Reference I
     b.   Depends on how user arranges sampler Intake.
     c.  Continuous flow sample line provided by user.
     c.  Refrigeration provided only for stationary

-------
Table IV-5    (continued).   AUTOMATIC  SAMPLER CAPABILITIES AND SAMPLING METHODS
Sampler
N-Con Sentinel
Phtpps 6 Bird
Protech CG-1253
Protech CG-125FPa
Protech CG-1503
Protech CEL-300

Protect DEL-2405

0_CEC CVE
QCEC E
SERCO NW-3
SERCO TC-2


Sigmamotor WA-la
Sigmamotor WA-3a
Sigmamotor WOPP-28
Sigmamotor WM-l-24
SIRCO BST-VS
SIRCO B/IE-VS3
SIRCO B/DP-VS



SIRCO PII-A
Method of sample transport
Dipper from sample chamber
provided by customer
Dipper
Pneumatic ejection
Pneumatic ejection
Pneumatic ejection
Submersible pump

Submersible pump

Vacuum pump
Dipper
Evacuated bottles
Dipper from sample cahmber
provided by customer**

Peristaltic pump
Peristaltic pump
Peristaltic pump
Peristaltic pump
Vacuum pump
Dipper
None provided - solenoid
valve diversion from
sample line

Vacuum pump
Refrigeration
provided
yes
no
yes
d
d
d

yes

yes
no
ice cavity

yes

yes
no
no
yes
yes
yes


yes

no
Ability to Withstand
freezing
no
no
yes
yes
yes
yes

yes

yes
no
no

no

no
no
no
no
yes
yes


yes

no
Immersion
no
no
no
no
no
no

no

no
no
no

no

no
no
no
no
no
no


no

no
Ability
Floatables
b
no
no
no
no
no

no

no
no
no

b

no
no
no
no
b
no


no

b
to sample
Coarse
bottom
sediments
b
no
no
no
no
no

no

no
no
no

b

no
no
no
no
b
no


no

b
Self-
cleaning
features
continuous
flow
none
backflush
backflush
backflush
continuous
flow
continuous
flow
air purge
none
none

continuous
flow
backflush
none
none
backflush
backflush
none


continuous
flow
backflush

-------
                                    Table IV-5  (continued).   AUTOMATIC  SAMPLER CAPABILITIES AND SAMPLING METHODS
to
Sampler
Sonford HG-*8

TMI
TM! Mark 38
AVCOa
Springfield



Milk River
Envl rogenics
Rohrer 1

Weston



Pavla-Byrne

Rexnord
Colston


Rohrer II

Near
Freeman

Method of sample Transport
Sample tube which fills
by gravity
Pneumatic ejection
Evacuated bottles
Peristaltic pump
Dipper from sample chamber
provided with continuous
flow by sc rewrote r pump

Submerged pump
Mechanical - gravity
Diaphragm pump

Evacuated bottle from smaple
cha ber provided with
continuous flow

Screw pump

Positive displacement pump
Serco NW-3 samples from a
flume provided with
continuous flow
Diaphragm pump

Piston in tube
None provided - solenoid
valve diversion from
sample llnec
Refrigeration
provided

ice gravity
no
no
no


yes

yes
no
no



yes

yes

no


no
no

no

no
Ability to
Freezing

no
yes
no
no


no

—
yes
yes



no

no

no


no
yes

no

yes
wl thstand
immersion

no
no
no
no


no

—
yes
no



no

no

no


no
no

no

no
Floatables

no
no
no
some


some

—
yes
fa





no

b


B°
u

yes

b
Coarse
bottom
sediments

no
no
no
no


some

—
yes
b





no

b


no
b

no

b
Self-
cleaning
features

none
non*
none
none


continuous
flow
—
na
continuous
flow


continuous
flow
continuous
flow
backflush
none for
SERCO par

continuous
flow
none

continuous
flow
     a.  Representativeness of sample questioned In Reference I.
     b.  Depends on how user arranges sampler Intake.
     c.  Continuous flow sample line provided by user.
     c.  Refrigeration provided only for stationary

-------
Of the k\ commercially available samplers, only 9 are capable of
obtaining a number of discrete samples and only 6 of these accept
signals from flow monitoring equipment.  Two of those six have the rep-
resentativeness of the samples obtained questioned.  Those models that
are capable of doing this are generally the more expensive samplers.
It may be noted that of those samplers developed specifically for
sampling combined sewage, 7 of 12 were designed to obtain a quantity of
discrete samples.  It is evident from this that those persons with
expertise in combined sewage sampling prefer this method of discharge
cha rac ter i za t i on.

Table IV-5 shows that 21 of k] commercially available samplers are
refrigerated to provide sample preservation.  This is strongly
desirable for storm generated discharges  in that they are routinely
analyzed for biological and pathogenic bacterial indicators such as
fecal coliforms.  Four of 12 custom designed samplers were refrigerated.
Of all the samplers, only two are capable of withstanding total
immersion (which could occur in storm or combined sewers when samplers
are placed in manholes) and 22 of 53 are capable of operating in freezing
weather.  Ability to operate in freezing weather is not a necessary pre-
requisite for sampling combined sewage due to the infrequent overflows
during winter conditions.  However, it can be a desirable feature.

The dipper type samplers do not have provisions for keeping the dipper
clean, i.e., free of entanglement with rags or other debris.  This is a
serious drawback to these type of samplers, since storm generated
discharges can contain large quantities of debris which could cause
failure during overflow events.  Other sampler types have backflush
(sample flow reversal) or air purge systems to help unclog sample lines
and prevent sample line blockage.

The samplers that use vacuum or peristaltic pumps to withdraw water
samples usually obtain sample volumes  inversely proportional to the lift
height.  To increase the sample volume for a greater lift it is nec-
essary to sample for a longer period of time.

Very few of the samplers have provisions  for sampling floatable or
coarse bottom sediments.  Quantification of these materials cannot be
reliably done by automatic methods, and need to be accomplished using
manual grab sampling techniques.

Desirable Characteristics of Ideal Sampler

There are a number of different types of  samplers with a large number of
different characteristics (Table  IV-*t and  IV-5).  The ideal sampler for
storm generated discharge studies still does not exist.  Some sampler
characteristics are mutually exclusive and compromises must be made.
The following is a partial list of the most desirable sampler
characteristics for storm generated discharge studies.
                                    43

-------
 1.  Ability to take a sequential  timed (variable timer) series of
     discrete samples.  It should  be possible to use an external signal
     to allow sample volumes to be taken proportional to flow rate or
     increments of flow.   Five minutes should be the minimum sampling
     interval and should  be adjustable.

 2.  Four different sample containers should be filled at each sampling:
     (a) for solids and BOD testing to hold no preservatives, (b) for
     metals and TOD analysis acid  added to preserve sample, (c) for
     nitrogen and phosphorus HgCl2 added,  and (d) sterilized containers
     for bacterial analysis.  The  fourth set of containers could also be
     used for grease and  oil, pesticides or some other tests.

 3.  Capability of using  1 to 3 liter sample containers so that
     individual discrete  sample analysis can be made.

 4.  Capability of programming the time interval at which samples are
     taken, so the sampling interval  can be short during the early
     stages of the storm  and longer intervals automatically used as the
     storm continues.

 5-  To hold 96 sample containers  - this would allow sampling every five
     minutes for two hours, or ten minutes for four hours where longer
     duration discharges  are expected.

 6.  Refrigeration capabilities to hold samples at A° C.

 7-  Capability of lifting sample  7.6 m (25 ft) or more without
     affecting sample size.

 8.  Have a self-contained power source.

 9-  Be automatically activated to sample at beginning of storm.

10.  Inlet and inlet line to be sufficiently large to eliminate
     problems of plugging.

11.  Inlet orifice velocity to be  sufficiently high to keep heavy
     particles in suspension throughout their flow to the sample
     container.

12.  Inlet device of such a configuration to allow obtaining a rep-
     resentative sample throughout the depth of the stream flow.  Light
     floating material and heavy bottom sludge should be included in
     each sample.

13.  Inlet device does not plug easily and is self-cleaning.  Sample
     lines are self-purging with cleanser and/or disinfectant so that
     next sample is not included in any of the previously taken sample.
                                   44

-------
The aforementioned described ideal sampler does not currently exist;
however, improved samplers are being developed  (21).  Operating
personnel may have to modify or construct adequate samplers for their
application.  Continued development work is needed to provide new and
improved sampling devices for storm generated discharge projects.


RECOMMENDED SAMPLING PROGRAMS

Types of Sampling Programs

The various types of sampling programs have been discussed earlier  in
this section.  The application of these different types of sampling
programs are presented  in Section VII for evaluating storm generated
discharges and in Section VIII for evaluating storm generated treatment
processes.  Regardless of what type of sampling is being done, automatic
samplers are recommended and these samplers should have automatic start-
up capabilities.  Samplers taking discrete samples are recommended.
Each sample bottle should be at least 500 ml in volume, preferably closer
to 4 1 ,  if possible.

Samp] ing Location

In order to ensure characteristic samples and optimum sampler accessibil-
ity, several considerations are recommended
      1.  Maximum accessibility and safety — Manholes on busy streets
         should be avoided  if possible; shallow depths with manhole
         steps  in good condition are desirable.  Sites with a history
         of surcharging and/or submergence by surface water should be
         avoided if possible.  Avoid locations which may tend to invite
         vandal ism.

      2.  Be sure that the site provides the  information desired —
         Familiarity with the sewer system is necessary.  Knowledge of
         the existence of inflow or outflow  between the sampling point
         and point of data  use is essential.

      3-  Make certain the site is far enough downstream from tributary
         i nf 1 ow to ensure mixing of the tributary with the main sewer.

      k.  Locate in a straight length of sewer, at least six sewer
         widths below bends.

      5-  Locate at a point  of maximum turbulence, as found in sewer
         sections of greater roughness and of probable higher
         velocities.  Locate just downstream from a drop or hydraulic
         jump,  if possible.  However, if sampler is tied into flow
         measuring equipment, this must be taken into consideration.
                                   45

-------
      6.   In all cases, consider the cost of Installation, balancing
           cost against effectiveness in providing the data needed.

 Knowledge of the sewer system or stream flow is valuable in selecting
 the sampler site.  This will also help in establishing where to place
 the sampler Inlet device.  When I tern 5 above is met and the wastewater is
 sufficiently mixed that a uniform composition exists throughout the
 depth of  the waste flow, the sampler inlet device should be located at
 mid-depth.  Manual grab samples should be collected periodically at
 different depths and analyzed for suspended solids and TOD.  These
 analyses should be compared with results from the automatic sampler to
 ensure the sample site  is located in  a thoroughly mixed portion of the
 sewer.   Where stream or lake analyses  are made for impact studies,
 samples should also be taken periodically at different depths to
 establish that the samples are representative of the main water body.

 Sample  Handling and Preservation

 Plastic sample containers should be used for all  sample collection except
 where grease  and oil  or pesticides and other trace organics are to be
 analyzed.   Hard glass containers should be used where these analyses will
 be made due to the problems with adsorption.   Sample containers should
 be thoroughly  washed  as previously Indicated.   The possibilities of the
 container  affecting the sample analyses should  be checked  periodically.
 Distilled  or demineralized  water should be placed in a typical  container
 for a period of time  similar to that of a  normal  sample.   Then  the
 particular constituent  of interest should  be measured in  the water from
 this blank.  Checks for sample adsorption  on the container  should  also
 be made by placing  a  known  amount  of a  particular constituent in a
 typical  container.  After a specified  holding  time,  analyses should be
 made to  determine  if  any  of the material was adsorbed into  the  container
 or changed in any other manner.  These  checks should  be done after
 sample  bottles  have been  used  for  a series  of storms.   In  this  way the
 cleaning techniques used  can be tested  for  thoroughness.

 All samples should  be kept  in  a  refrigerated or  ice  cooled  container
 during and after sampling.  This cooling will slow  the  biological
 reactions  and prevent significant  changes  in the  nature of  the  sample.
 Although the ideal  procedure for preservation would be  to add chemicals
 prior to sample collection,  this is not possible  for  storm generated
 discharges.  Therefore cooling  is  recommended until  the sample  is  brought
 to  the  laboratory for analysis.  At this time, the volume needed for a
 specific analysis will be split off of  the  total  sample and  either
 analyzed Immediately or preserved  (as detailed  in Section VI) for  later
work.

 It  Is important to realize  that the biological changes will affect  the
 results of certain tests  if they are not run immediately.  This  is
especially true of BOD which should be  run within 6 hours  after
                                   46

-------
collection of the sample.  A maximum holding time of 2k hours may be
allowed if the sample is kept refrigerated.

Bacterial  analyses are also affected by biological processes and other
factors.  If bacterial analyses are to be run on the composite samples,
this can only be used as a control for the system and not as a
reportable value.  The cross contamination due to a single sampling line
and funnel in the sample and the lack of a dechlorinating agent in the
sample bottle can invalidate the test.  Bacterial tests on composite
samples can be used if "in house" control is desired to get relative
numbers, but should not be considered an absolute value.  If a quanti-
tative test is required by regulatory agencies, then the sampling and
preservation techniques mandated by the agency should be followed for
bacterial  analyses.

SAMPLING ACCUMULATED ROADWAY MATERIAL

Samples of materials deposited on roadways are collected using a
combination of sweeping, vacuuming, and water flushing techniques.
Each sample will consist of three fractions:  litter, dust and dirt,
and water flush.  The particulate materials collected by sweeping and
vacuuming are separated on the basis of particle  size into a litter
fraction and dust and dirt fraction.  The litter fraction consists of
that portion of  the particulates retained by a U.S.A. No. 6 sieve,
greater than 3-35 mm  in diameter.  This fraction is usually composed of
stones, gravels, wood fragments, and other larger-sized materials in
addition to bottles, cans, paper production, etc. which are normally
thought of as litter.  The dust and dirt fraction will contain partic-
ulates smaller than 3-35 mm in diameter.  The water flush fraction
contains those components of the dust and dirt fraction which were not
picked up at high efficiencies by the sweeping and vacuuming techniques.
The flush plus the dust and dirt constitute a total dust and dirt
fraction which is the major source of water pollutants found in runoff
from urban roadways (22).

Collection of Samples

If a physical and chemical description of the street surface contam-
inants  is needed, the sample should be collected by hand sweeping,
followed by flushing.  All of the dry solid material collected from the
test area should be placed in clean containers and shipped back to the
laboratory.  There it should be air dried thoroughly and sealed for
storage until analyzed.  All of the flushed material should be
measured for volume, but only a portion of it need be retained for
analysis.   The liquid sample should be stored in clean containers
(glass, if pesticide analyses are to be made) and cooled to <4° C., if
possible.   The analyses of the liquid fraction should be made as soon
possible after collection.  To reduce the number of chemical analyses
required,  the dry and liquid samples can be combined on an equal sample
                                   47

-------
area basis before the analyses  are performed (23).

If only physical  loading information (such as kg [Ibs] of solids per curb
km [mile]) is needed, hand sweeping is probably sufficient.  In most
cases, the additional quantity  of material that can be obtained by
subsequent vacuuming and/or flushing is insignificant.  If information
regarding particle size distribution is required,  then the sample should
be collected using a combination of hand sweeping  and dry vacuuming.
The vacuum is more efficient in removing the fine  particles which are
needed for size distribution analyses.  If size distribution of the
solids in the wet phase is needed, then flushing will also be required
(23).

The basic procedures for the collection of samples  are:

     Hand sweeping - Hand sweeping for dry solids  collection should
     utilize a standard sti/ff-bristled push broom.   The sweeping
     pattern should be from the center of street or from one edge
     of the test area towards the gutter or opposite side of the
     test area.  After concentrating the material  along this edge,
     the sample should be collected, using a whisk broom and
     dustpan.

     Vacuuming - Vacuuming the  test area usually removes more
     smaller-sized particles than is possible by only using
     sweeping techniques.  The  vacuuming pattern should approxi-
     mate the pattern described for hand sweeping.   An industrial
     wet/dry "shop" vacuum cleaner with a 5-7.6 cm (2 in. to 3 in.)
     diameter hose is recommended.  Other types of units, ranging
     from small household vacuums to large motorized vacuum
     sweepers, may also be satisfactory, depending on the size of
     the test area.

     Flushing -  The test area  can be flushed with water after
     hand sweeping to remove soluble films and other nonsweepable
     material.  The materials removed with this method more closely
     resemble those which are removed by a runoff  event.  The test
     area is first slightly wetted to soften and facilitate
     removal of soluble materials.  It is then flushed with a
     stream of water from a garden hose and spray  nozzle connected
     to a fire hydrant or other water supply.  Begin at the road
     crown and flush toward the edge.   The downslope gutter is
     dammed with sandbags to create a collection area.  A small
     vacuum collector is used before an industrial  wet/dry vacuum
     cleaner to remove the sample water from the collection area.
     All  water and contaminants are collected using this vacuum-
     operated collector trap.  This is an air-tight box or drum
     with a capacity of several gallons to several  hundreds of
     gallons (depending upon specific test procedures), outfitted
     to function as a "trap" in a vacuum  line.  The inlet hose of
                                   48

-------
     the collector trap has a pickup nozzle on the open end.  The
     outlet hose of the collector trap is connected to an  industrial
     shop vacuum.

Table IV-6 (22) gives a specific stepwise sampling procedure for the
collection of street surface contaminants.

The vacuum cleaner used for collection of roadway particulates consists
of a pick-up head attached to a 38 1 (10 gallon) canister on the top of
which is mounted an exhaust motor.  Exhaust ports from the canister
leading to the motor are covered by a filter bag to retain solids picked
up during the vacuuming operations.  Since the finer particles found on
roadways are relatively more heavily laden with water pollutants,
experiments have been performed to determine the retention of smaller-
sized particles by the filter bag  (22).  Recoveries of 99, 93, and 9*»
percent were obtained using a new filter bag with each sampling run.
These tests indicate satisfactory retention of fine particulates by the
filter bags as well as quantitative removal and recovery of vacuumed
particles from the canister walls and bags.

The water flush procedure has also been tested in the field (22).  It
was found that a roadway area of 92 sq m (1000 square feet) could be
thoroughly flushed with about 95 I  (25 gallons) of water.   In most
cases, over 50 percent of the applied flush was recovered  by vacuuming
of the impounded water along the curb.

The data  in Table  IV-7 lists average percentages found in  the flush
fractions for specific components of dust and dirt.  The standard
deviations are also listed to indicate the constancy of this fraction.
Arbitrarily selecting 80 percent or better as satisfactory  recovery, it
is apparent that most parameters are adequately recovered with the dust
and dirt  fraction.  The dry weight, heavy metals, asbestos, oil and
grease fractions,  COD and others are all found largely in  the dust and
dirt fraction.  However, a considerable percentage of BOD, Kjeldahl-N,
water soluble amines  and microorganisms are recovered with the water
flush.

It is concluded from these data that flush fractions must  be collected
in order  to obtain accurate values for some pollutants.

Samp1 ing  Site Selection

Sampling  sites should be chosen that represent the range of conditions
that occur  in the  area.   Important variables may include land use,
average daily traffic, type of adjacent landscaping and street surface
material.   It  is recommended that at least a single complete analysis be
made for  each land use area, with  total solids analyses being made on
samples representing other identified variables.   If several sampling
sites are established in each land use area, a portion of  each sample
                                    49

-------
      Table IV-6.   SAMPLING  PROCEDURE  FOR THE  COLLECTION
                      OF  STREET  SURFACE  CONTAMINANTS
	     Equipment   	    .	

 Hard bristle broom,  rake,  shovel,  and foxtail  or paint brush

 Alternator power plant,  3500 watt, Dayton  Electric  Manufactuing
  Company,  Model  1W832A

 Two wet and dry  vacuum cleaners,  10  gallon,  Dayton  Electric
  Manufacturing Company,  Model  22612, with  sufficient filter
  bags.   A new filter bag for each  sampling (three vacuum passes)

 Steel drum, 208  1 (55 gallon)  with lid and rim lock, containing
  151-189 1 (40 to 50 gallons)  of water.

 Rotary screw pump, 3-5 amperes, Dayton Electric Manufacturing
  Company,  Model  No.  3P569

 Garden hose, 46  m (150 feet)

 Galvanized garbage can with clamp  fitting  lid, one  or more

 Dual Motor shop  wet and  dry vacuum,  Dayton Electric
  Manufacturing Company,  Model  No.  3Z107 mounted on  a 208  I
 (55 gal Ion)steel drum

 Sand bags - 5 to 7,  .014 cu m (1/2 cu ft)  bags
                           Procedure

1.  Select a roadway sampling site 30.48 continuous curb meters
    (100 feet)  or more.   The street surface and curbing should
    be in relatively good condition.  Mark the limits of the
    sampling length selected.

2.  Rake and/or brush along the curb for 3.0 or 4.6 m (10 or
    15 feet) from the limit markings away from the section to
    be sampled.

3-  Knock the brush clean.   Rake and/or brush from the higher
    elevation limit.  Shovel bulk litter plus swept dust and
    dirt into clean galvanized garbage can.
                               50

-------
  Table  IV-6 (continued).   SAMPLING PROCEDURE FOR THE COLLECTION
                              OF STREET SURFACE CONTAMINANTS

	Procedure	

 k.   Vacuum along the entire curb length of the roadway sampling
     site out to a distance of four to five feet from the curb.
     Three vacuumings of the site should be carried out to
     collect the dust and  dirt sample fractions.  Two vacuum
     cleaners are used simultaneously to speed up the operation
     with particular attention at the litter pick up point.

 5.   Position several sand bags at the curb of the lower limit of
     the sampling area to  impound the flush water.

 6.   Place the nozzle of the dual motor shop vacuum at a low point
     in  front of the sand  bags so as to suck water into the  208  1
     (55~gallon) drum.

 7.   Place the intake hose from the rotary screw pump into the
     208 1 (55~gallon) drum filled with water and begin flushing
     the roadway using the garden hose.

 8.   Flush the entire roadway surface area toward the curb and
     finish by flushing the gutter toward the sand bags.

 9.   Approximately 57 to 95 1 (15 to 25 gallons) of water are
     required to flush 56-93 sq m (600-1000 square feet) of
     roadway.  Generally greater than 50 percent of the flush
     water applied is recovered by the vacuum.
                  At the laboratory or on-site

 1.  Take out the filter bags and shake well  into garbage can
     with bulk material.  Save the bags.

 2.  Empty vacuum canisters into garbage can.  Brush canisters
     we) 1.

 3.  Take combined litter and dust and dirt in garbage can and
     the flush fraction to the laboratory.  Other equipment may
     proceed to next sampling site.
                                51

-------
Table IV-?.  DISTRIBUTION OF POLLUTANTS BETWEEN DUST
              AND DIRT AND FLUSH SAMPLE FRACTIONS
Parameters
Dry Weight
Volatile Sol ids
BOD
COD
Grease
Petroleum
n-Paraff ins
Total PO^-P
PO^-P
N03-N
N02-N
Total Kjeldahl-N
Chloride
Asbestos
Fecal Coliforms
Fecal Strep
Lead

Chromium
Copper
Nickel
Zinc
Avg. % in Flush
7
20
36
16
19
19
19
15
43
69
97
33
43
13
76
4**
h
"

17
5
5
2
% Standard Deviation
8
13
22
12
15
13
14
15
42
24
7
23
33
31
40
39


15
4
2
1
                        52

-------
could be combined for complete composite chemical analysis rep-
resenting that land use.

For a 12-month field study in Washington, D. C.  (22), seven area
roadways were chosen for the field study based primarily upon the
range of average daily traffic levels and road use categories
encompassed.  Other factors considered in the roadway selections
were speed limit and roadway surface material -  Satisfactory condition
of the street surface and a sufficient length of curb against which
the sample could be deposited and collected were important factors in
selection of the specific sampling sites on the area roadways chosen.

In general, the following information should be collected for a sampling
site:  sampling location; date; local land use; parking restrictions;
traffic characteristics; composition, type and condition of the
street, gutter and curb; the size of the test area; and a description
of the adjoining area.  Photographs of the area are often valuable.
Data concerning the cleaning frequency, the date of the last recorded
cleaning, and the recent rainfall history should also be obtained for
each test area (2k).

In addition, if the selected study area is subject to vehicular
traffic,  it will be necessary to establish some type of traffic
control for the protection of the field workers.  Flagmen and traffic
cones are probably a minimum precaution which should be used in all
areas.

The type of study area  (street surface, parking  lots, or other large
surfaces) and sampling objectives will determine the size of sampling
area.  A  typical secondary street can usually be sampled using a single
test area of about 93 sq m, 7.6 m x  12.2 m  (1000 sq ft) (25 x kO ft).
Large paved surfaces may be better sampled using several smaller test
areas  (0.9 sq m [10 sq ft])  and averaging the results.  Experimental
design procedures should be incorporated to determine the necessary
types of  study areas to sample to satisfy specific study objectives.
The published results of previous sampling programs (2k, 25, 26) may
be useful in this design process.

Frequency of Sampling

As with the selection of the study area, the frequency of sampling will
depend on the objectives of the sampling program.  For the Washington,
D. C. study (22) a schedule was set up early in the program such that
the roadways were sampled during several seasons of the year in order
that seasonal effects on pollutant  'deposition rates might be studied.
However, during the winter season, freezing conditions prevented the
collection of some of the flush fractions.
                                   53

-------
Sampling periods were scheduled to begin on a Monday and end one week
later on the following Monday.  Sample collections were planned to  be
carried out in the following manner:

     1.  An initial sample was obtained by cleaning the roadway surface
         and quantitative collection of materials initially found on the
         site.  No measurements of traffic were taken to correspond with
         the initial sample; however, records of precipitation and  dates
         of the most recent antecedent cleaning of the roadway surfaces
         were maintained throughout the 12-month field study.

     2.  The site was sampled a second time after an accumulation period
         of approximately 2k hours during which time a measured volume
         of traffic passed the roadway site.  As many as four samples
         having a one-day accumulation period were taken during the
         remainder of the week.  Traffic counts were taken with each
         one-day sample.

     3.  The final sample of the period was gathered following the
         weekend.  Ideally then, a sampling period consisted of an
         initial  sample, four one-day samples and a weekend sample
         with traffic data for all samples except the initial one.

     k.  Precipitation frequently interrupted the planned pattern of the
         sampling periods.  Samples were gathered after rainstorms  in
         a few cases;  however, it was felt that such samples would be
         atypical; and,  therefore, collections after runoff events were
         abandoned early in the program.   The roadway site was cleaned
         as soon  as convenient after  precipitation had ceased and a
         new sample accumulation period begun.  Sampling periods were
         extended in some instances in order to make up for loss of
         samples  due to  precipitation.

Experimental  design procedures should be incorporated ta determine the
required sampling frequency and sample numbers to satisfy specific
study objectives.   Again, the published results of previous sampling
programs may  be useful  in this design process.
                                   54

-------
                 SECTION V - MONITORING INSTRUMENTATION
FLOW MEASURING EQUIPMENT

Level Gaging

Devices for gaging wet weather sewage level in storm and combined
sewerage systems vary in complexity from a dipstick or chalked length
of rope to fairly sophisticated electromechanical  and electronic
instruments.  Float gages and bubble tubes have found most common use to
date, but ultrasonic level gages and conductivity gages are also
beginning to be used in storm generated discharge measurement applications.
Level determinations are important for tracing sewer levels during and
after storms as a function of storm intensity, duration and location,
for  infiltration studies, for determining discharge flows in overflow
sewers, and for determining storage capacity,  routing programs, and
gate control in "in-line" storage systems.

Dipstick and Chalked Ropes - The dipstick and  flashlight (Figure V-l)
while an inexpensive and reliable method, is best suited to slowly
changing dry-weather flows and Is obviously not adequate to measure the
multitude of rapid level changes in a combined sewerage system or storm
sewer under storm conditions.  Some studies have used a chalked length
of rope fastened vertically in a sewer to measure the peak level reached
during a storm event.  The chalk dust is washed off the submerged rope
by the sewage, but remains on the exposed length.   However to obtain a
continuous record of level, more sophisticated devices must be used.

Floats and Scows - A buoyant float or scow connected via a cable to a
clock driven drum recorder (Figure V-2)  is a commonly used method.  A
simple float can be used in sewage applications if a stilling well
provides an undisturbed liquid surface.  Measurements directly in the
sewage stream require a buoyant scow attached  to a swivel as sketched in
Figure V-3.

Recorder clock drives are either spring driven or synchronous motor
driven.  Spring drives provide continuous unattended operation for periods
from four hours to eight days as required.  Although direct and reliable,
simple floats require a fairly costly stilling well, and they as well as
scows are subject to fouling and submerging by passing debris, and
therefore require frequent inspection and cleaning.
                                    55

-------
   705
   60-
   50-
  .  30*
   1   •
     2*.
     Figure V-2.  Float driven recorder with
clock drive (courtesy of Leopold & Stevens,  Inc.)
                        Figure V-3.  Recorder with scow float  used  in  a
                        sewer manhole, (courtesy Leopold  £  Stevens,  Inc.)
Figure V-l.
Level gaging
dipstick
(courtesy of
Leopold & Stevens, Inc.)
                                   56

-------
Bubbler Level Gages - The bubbler level gage sketched in Figure V-A
requires less maintenance because of self-cleaning features.  A bubbler
dip tube, often a 1.8 cm (0.5 in.) pipe cut off at a right angle, is
immersed in the flow stream or inserted through the sewer wall and a
constant flow of air (or other gas, e.g., nitrogen) is forced through
the tube and bubbles to the surface.  The back pressure in the tube  is
proportional to level, provided the sewage density is constant.  A
constant air flow can be provided by a rotameter/differential pressure
regulator combination which maintains a constant differential pressure
across the rotameter regardless of back pressure.  The back pressure can
be read out  in a bourdon spiral pressure gage, or any of a variety of
mechanical, pneumatic or electrical pressure transducers, and recorded.
While widely used in sewage treatment  plant applications where a supply
of instrument air  is readily available, bubblers are not so commonly
used  in  sewerage systems because  of the need for a motor-compressor or
compressed nitrogen  tanks for a gas supply.  Occasionally pressure
transducers  are used to sense  level head directly, but an inlet port or
slack diaphragm at the sewer invert is easily fouled by sediment so  thi's
method  is  troublesome.  The air flow  through the bubbler dip  tube keeps
the exit port clear  and eliminates  the fouling problem.  However, the
method  is  subject to error due to a pressure reduction from the fluid
velocity past the bubbler exit (Bernoulli effect), and occasional
occlusion  of the dip tube by solids which crystallize at the  exit port.
                                                         0IAI 0*

                                                          DICATOM
                                                            Air Bubbler
                   Figure V-A.  Air  bubbler  level gage
                     (courtesy  Fischer 6  Porter  Co.)
                                    57

-------
 Conductivity  Level Gages - Another  level gage designed  to minimize
 fouling  problems  is  the "dipper", a device which  lowers a thin metal
 probe  to the  sewage  surface, by unwinding a motorized wire  spool.
 When the probe makes electrical contact with the  sewage surface,  the
 spool  motor  is reversed and the probe  is retraced.  Then it  is  lowered
 again  until electrical contact is made and then again retraced.   This
 dipping  action is continuous at five second intervals with  the probe
 position maintained  just above the sewage surface.  As  in a  float gage,
 spool  position is a measure of liquid  level.  The probe is  lowered  from
 the bottom of a unit which houses the dipper mechanism and a  level
 recorder as pictured in Figure V-5.  The dipper spool drive  is battery
 powered  and the recorder is spring powered with either a 2k  hour or 7
 day clock drive.

 A  second electrical contact with the sewage stream, a ground  return,  is
 also required.  A problem encountered  in some installations  is the  coat*
 ing of the ground return by grease which interrupts electrical contin-
 uity.  Also, passage of large pieces of debris floating on the sewage
 surface  can snag or hook onto the dipper probe, causing an erroneous
 reading,  or in extreme cases, damaging the mechanism.
      Figure V-5.  "Dipper" level  gage based on electrical contact
     with the sewage surface.   (Courtesy Manning Environmental Corp.)

Ultrasonic Level Gages - A growing number of domestic companies are
offering ultrasonic level gaging equipment for wastewater measurement
applications.  The major advantage is that no contact, electrical or
mechanical,  is required with the flow stream.  A typical gage (Figure
V-6) consists of an ultrasonic transceiver, a signal processor and a
recorder.  The gage operates by generating pulses of sound energy at a
frequency above audibility and directing the pulses at the liquid
surface from above.  Each pulse is reflected by the liquid surface and
a portion of the energy is received by the transceiver.  The time
required for the transit of each pulse from the transceiver to the
surface and  back again is measured by the signal processor.  This time
                                   58

-------
01
                                      Figure  V-6.   Portable ultrasonic flow gage
                                             (courtesy  of Rexnord Inc.)

-------
duration is a measure of liquid depth, and is converted to an
electrical output signal and recorded.

While more complex electronically than the other level gages described
earlier, the method has attained a cost and reliability over the last
several years which suit it to unattended sewage gaging applications.
The portable unit pictured in Figure V-6 can be mounted in manholes for
storm generated discharge and infiltration studies by affixing the trans-
ceiver probe to the manhole wall perpendicular to the sewage surface,
and hanging the signal processor/recorder unit and battery on ladder
rungs above.  The gage is also being used in permanent level gaging in-
stallations in sewerage systems (e.g., a large installation in Rochester,
N.Y.) and  in treatment plant applications.

Summary, Comparison and Recommendations - In the preceding discussion of
level measurement devices useful in storm generated discharge applica-
tions, little mention was made of other techniques which are in common
and successful use for head meausrements in the process industries, such
as force-balance and motion-balance differential  pressure transducers,
slack diaphragms, bourdon tubes and manometers.  These devices are in
general not well suited to sewage and storm generated discharge applica-
tions because of difficulties in keeping pressure sensing ports or
surfaces free from fouling.

Continuous level recordings can be provided by floats, scows, bubblers,
dippers and ultrasonic level gages.  Float gages require stilling wells
and in such installations serve reliably over the full range of sewage
depths provided the well does not become fouled with grease, scum or
other debris.  Instrument cost is lowest in the group of continuous level
recording gages, but capital cost of the stilling well must be added.
Scows are most often used in connection with head measurements over a
flume or weir, and like these flow structures, they operate over a
limited range of depths and cease to function accurately when submerged.
Discharge measurements over flumes and weirs become inaccurate when flow
in the downstream channel becomes sufficient to reduce velocity and
increase depth.  In the absence of regular maintenance, scum and debris
accumulation on scows can occasionally cause them to submerge completely
in the flow stream during normal flows.  A typical scow level sensor
with a 25.^ cm (10 in.) circular chart recorder sells for about $950.
Bubble tubes require a continuous gas supply and are not therefore very
well  suited to remote or standby service, although they are being
operated with a small air compressor or nitrogen tanks in some applica-
tions, and they have the virtue of relative simplicity.  Selling for
about $1100,  the "dipper" level  gage appears best adapted to portable
uses  where levels are to be gaged for comparatively short periods of time
at multiple sites,  as in infiltration studies.  The ultrasonic level  gage
also sells for about $1100 and is suited to both portable use and
permanent installations.  Because of this versatility and its lack of
need  for physical  or electrical  contact with the sewage, it is the
                                   60

-------
recommended method for level gaging in storm generated discharge
projects.  When evaluating storm generated discharge treatment processes
many of the gages are applicable depending on the exact situation.  It
is expected that the bubble tube, float, scow (for flume measurements)
and the ultrasonic level gage will all find various applications
depending upon the sewer or treatment process configuration.

Flow Gaging
Gaging of sewage flow In storm and combined sewer systems is the single
most Important measured parameter for overflow characterization and
control.  As such, sewer flow gaging has been the subject for considerable
RSD effort in  Industry, government and universities here and abroad over
the last several years, and a number of new measurement methods have
been developed.  In earlier years, most flow gaging research was
directed toward metering In full pipes, particularly for process Industry
applications.  The added complexities associated with open channel
hydraulics, entrained solids and Irregular sewer wall features make
sewage gaging  a more difficult problem, one that has been less
susceptible to successful solution by conventional instrumental techniques,

Sewage flow is most commonly gaged using weirs  (rectangular, V-notch,
Cipoletti) and venturi  flumes (Parshall, Palmer-Bowlus).  Discharge
through  these  structures is gaged  in  terms of the head over the structure,
which  is in turn measured by means of the  level sensing devices described
in the preceding section.   Relationships between discharge and head for
these  structures are generally a power  law function and are described
in the general engineering  literature.  Many level gaging devices are
available which automatically generate  linearized flow  information based
on weir or flume depth measurements using electrical or mechanical
function generators.   In uniform sewers with a  straight section at least
100 m  (300 ft)  long, relatively steady  flow can be estimated from the
slope  of the sewer and  sewage depth using  the Manning formula.  However,
the sewer  length chosen must be free of sudden expansions, contractions
or drops and must have a uniform wall roughness.  Because of the diffi-
culty  in satisfying the criteria for application of the Manning formula,
it is  seldom possible  to compute flow much better than about 10% by this
method.  A properly applied weir or flume under free flow conditions on
the other hand, can provide flow measurement accuracies of three to five
percent.

Full pipe raw  sewage flows are being measured successfully by electro-
magentic meters.  A drop into a U-shaped configuration  in the sewer line,
typically at a treatment plant  inlet, is used to obtain full pipe flows.
Also tracer and chemical dilution  techniques are  in use for intermittent
measurement.

Several new flow gaging methods for sewage applications have been
advanced in recent years.  These  include open channel flow velocity
gaging using ultrasonic techniques and electromagnetic  techniques.  The
new activity in seeking improved sewer  flow gaging technology has been
stimulated by  a growing need for  gaging in connection with infiltration,
inflow and overflow studies, and automatic  remote  control  flow routing.

                                    61

-------
planning, justification, design and operation of added treatment plant
and sewage collection system capacities, and are of direct importance
in storm generated discharge considerations.

Weirs and Venturi  Flumes - Weirs and flumes are flow structures
designed to provide a known, repeatable relationship between flow and
depth.  Upstream flow is backed up due to the structure's contraction of
the channel.  The  structure has a regular shape chosen to provide a
convenient function relating flow and depth, e.g.,  rectangular weirs,
trapezoidal weirs, and Parshall  flumes:   approximately 3/2 power law;
V-notch weir:   5/2 power law.   The chief disadvantage of the weir
relative  to a flume is  its higher headloss and greater tendency to
collect settled solids and debris upstream of the weir bulkhead.

In the case of flumes, there is  only a slight depression in the floor of
the flume, so most debris  is carried right through,  thereby minimizing
cleanout requirements.  A  dual-range (or nested)  Parshall  flume is
pictured in Figure V-7.
                Figure V-7.  Dual range Parshall flume
                   (courtesy Fischer 6 Porter Co.)
                                  62

-------
This flume is designed for application where there  is a sizeable
difference between initial and ultimate flow rates as occurs  in sewers
subject to storm flows.  Flow is  initially gaged by measuring head over
the small flume.  When this flume's capacity is exceeded,  it  is removed
and the outer flume  is used.  A disadvantage of the Parshall  flume in
sewer manhole installations is its relatively  long converging approach
section and diverging head recovery section which require  a sizeable
laying length.  A number of shorter form flumes have been  introduced
which can be more conveniently installed in manhole accesses  (the
Palmer-Bowlus flume).  An example  is pictured  in Figure V-8.
                        Figure V-8.   Lagco flume
                   (Courtesy F. B.  Leopold Co.  Inc.)
 Flow  from a  line  discharging  from  above  into  a  lower  receiving  stream
 can be metered  using  a weir,  an  open  flow  nozzle or an  H-flume.  The
 H-flume  (Plasti-Fab  Inc.)  has a  wide  flow  range, e.g.,  a  0.8 m  (2.5 ft)
 flume handles  flows  from  0.05 to 30 cm/min (0.03 to 17-6  cfs) with
 accompanying head range of 0.03  to 0.75  m  (0.1  to  2.^4 feet).

 Linear Flow  Indication from Weirs  and Flumes  -  Head,  H  over a weir  is
 usual ly  measured  a distance VHmax  upstream from the weir.  Head over a
 Parshall flume  is measured at a  point 1/3  of  the distance into  the
 converging section.   Head measurement pressure  ports  are  visible  in
 Figure V-7.  Head measurements in  the Palmer-Bowlus and Lagco flumes are
 measured just  upstream of the circular to  trapezoidal or  circular  to
 rectangular  transition section.  These heads  can be measured by any of
 the variety  of  level  gages described  in  the last section.  Some manufac-
 turers provide  flumes with built in bubble tubes,  capacitance level gages
 or ultrasonic  level  gages. The  measured depth  generally  has a  power  law
                                    63

-------
relationship to flow rate,  so some sort of linearizing device must be
used to extract linear flow rate information from the head measurement.
Pictured in Figure V-9 is a float actuated flow computer which uses a
characterized cam for converting head to flow rate.   The float measures
level in a stilling well  which Is conversant with the bottom of the
stream at one of the measurement points mentioned above.  The float cable
drives a drum and the cam is rotated by the drum.  Cam rise is linear
with flow rate and may be used to drive a recorder or indicator via a cam
rider and linkage, or it  can be integrated to register totallized flow.
Note the counter in Figure  V-9.
              Figure V-9.  Float actuated flow  recorder
                    (courtesy Fischer & Porter Co.)
                                   64

-------
A similar flow computer actuated by an  in-stream flow scow  is pictured
in Figure V-10.  This device can be mounted directly on  top of a  flume
structure or adjacent to a flow channel upstream from a  weir.
                Figure  V-10.   Scow actuated  flow  indicator
                    (courtesy  Fischer  6  Porter  Co.)
Parshall flumes are available with bubble  tubes built  in  to a vertical
depression  in  the side wall of  the converging section  2/3 of the way
upstream from  the entrance  to the flume  throat.  A dished bottom section
in the flume floor  is provided  for bubble  egress.  A portable bubbler
unit for flow  measurements  on weirs or flumes is pictured in Figure V-ll.
In this device the bubbler  gas  is supplied  from freon  cylinders and flow
linearization  is performed  electronically.  Various flume and weir level
to flow relationships can be accommodated  using dial adjustments, and
accuracy is given as within ±2% of the theoretical curve  for the struc-
ture in use.
                                    65

-------
Figure V-ll.   Portable bubbler actuated flow recorder
     and totallizer (courtesy Sigmamotor,  Inc.)
                          66

-------
The conductivity or "dipper" type of level gage has also been adapted to
flow gaging.  Figure V-12 and V-13 are respectively the dipper level
transmitter and total flow computer.  A battery operated portable unit
incorporating the functions of the transmitter and computer is pictured
in Figure V-1A.  The dipper transmitter is mounted above the flume or
weir at the appropriate head measurement position, and a signal propor-
tional to level is transmitted to the flow computer.   In the total flow
computer, the flow equation is characterized digitally in a plug-in
read-only memory unit which generates an output signal which is linear
with flow rate.  In the portable unit, Figure V-14. linear flow informa-
tion is generated by means of an electronic servo.  Accuracy for the
dipper transmitter is stated as:  linearity ±1% of reading; repeatabil-
ity ± 0.003 m  (0.01 feet) and resolution ±0.5% of full scale.
                                            I
                Figure V-12.  Dipper level transmitter
                (courtesy Manning Environmental Corp.)
                                    67

-------
                manmngaare
                •MMMBMMBMMMMO TUTAL rww • •
             Figure  V-13.  Total flow computer-dipper system
                 (courtesy Manning Environmental Corp.)
              Figure V-lA.   Portable  flowmeter-dipper  system
                 (courtesy  Manning  Environmental Corp.)

Level over a flume or weir  measured ultrasonically  is  converted  to  flow
rate in the instrument pictured  in  Figure V-15.  Flow  linearization is
performed with solids state electronics  in  this  instrument, and  resolu-
tion, repeatability and linearity are each  given as within  \%.

Another technique for generating linear  flow information  from a  flume  is
the use of a characterized  capacitance element mounted within a
depression in the flume wall  or  molded within the  flume wall  itself.  The
shape of the capacitance element is designed to  provide a capacitance
change with changing head which  is  directly and  linearly  proportional  to
                                   68

-------
flow through the flume.  Associated  indicating and recording circuitry
can therefore be linear.

Effects on accuracy of buildups of grease or other materials on the probe
if in contact with the sewage stream can apparently be minimized by
measuring the admittance of the sensing probe  (the "Comad" circuit,
Drexelbrook Engr. Co.), rather than  its capacitive impedance.
         Figure V-15-  Flow recorder,  indicator and totallizer
               actuated by ultrasonic  level measurement
    (courtesy of Environmental Measurement Systems Division of Wesmar)


Effects of Submergence and Surcharge on Flume Performance - When flow
through a weir or flume exceeds a critical value, flow resistance  in the
downstream channel reduces the velocity, and backwater level begins to
approach crest level.  The ratio of backwater level to crest level  is
defined as submergence.  Because of largely unknown effects of sub-
mergence on weir accuracy, weirs are generally operated  in a free-flow,
non-submerged condition.  Flumes however retain acceptable accuracy in
the presence of considerable submergence, i.e., with submergence ratios
up  to 70% in Parshall flumes 0.3 to 2.4 m (1 to 8 ft) wide.  For sub-
mergences up to 95%> the Parshall flume can still be used for flow gaging
if  crest and backwater levels are both gaged, and the discharge calcula-
tion is adjusted as a function of the  submergence ratio.  The Palmer-
Bowl us flume tolerates relatively higher submergence and has a shorter
lying length than the Parshall flume,  so it is recommended over the
Parshall flume (and over weirs) when flows are within a  flume's operating
range.
                                    69

-------
-d
o
                                               WATER LEVEL
                                               RECORDER
                             WATER LEVEL
                             GAUGE (CAPACITANCE)
                                                                             TRACER FEED q,Cn

                                                                                      STORM WATER
                                                                                      ~Q.Ci
COLLECTION OF
SAMPLES  FOR
WATER QUALITY
ANALYSIS  AND
TRACER ANALYSIS
 VENTURl  FLUME
(PALMER- BOWLUS)
                                                                                      (REMOTELY ACTIVATED
                                                                                       WHEN WATER IN THE
                                                                                       SEWER REACHES
                                                                                       DEPTH d)
                              Y////////A
                                 Figure  V-16.  Flume-tracer dilution combination for
                                  sewer  flow gaging  (courtesy J. Marsalek,  Canada
                                               Centre for  Inland Waters)

-------
However, if a sewer is approaching a surcharged flow condition, the
flume structure itself is submerged, and obviously can no longer be used
for gaging.  Surcharging occurs commonly in storm generated discharge
measurements, as do rapidly changing flows, so a number of techniques
have been advanced to maintain accurate measurement under these condi-
tions.  For example, the installation sketched in Figure V-16 combines
a tracer dilution technique with a flume/capacitance level gage.

When the flume head measurement indicates that submergence is about to
affect flume accuracy, an upstream tracer feed pump is turned on, and
stormwater samples are collected at the flume for later analysis after
the storm event.  Tracer concentration  in the samples carries the flow
rate information.

Open Channel Flow Velocity Gaging - Another approach to circumventing
flume submergence problems is  to eliminate the flume structure and
directly measure average stream velocity and depth instead.  Flow can
then be calculated automatically by multiplying average velocity by the
cross-sectional area of the flow stream.  One such system based on ultra-
sonic measurements of flow velocity and level was demonstrated in several
sewers  in Milwaukee under U.S. EPA sponsorship.  Pictured in Figure V-17
is an open channel  installation, the 2.kk x 3-05 m (81 x  10') main
influent channel at the Milwaukee Jones Island Wastewater Treatment Plant
         Figure V-17.  Ultrasonic flow velocity and  level gaging
                on a  raw sewage channel,  (coarse screened)
                       (courtesy Badger Meter  Inc.)
                                    71

-------
Ultrasonic transceivers are located obliquely opposite each other on
opposite sides of the channel  wall  at a depth equal to about bQ% of the
mean stream depth.  Bursts of  ultrasound are first sent in the downstream
direction from one transceiver to the other, with the reception of each
pulse triggering the transmission of the next.  Moving generally down-
stream, the propagation of the pulses is aided by the motion of the
stream.  The "singaround" or pulse repetition frequency established
thereby is counted for a fixed time period.   Then pulses are transmitted
upstream generating a lower singaround frequency, and are counted for
the same fixed period.  The difference between the upstream and down-
stream singaround frequencies  is  directly proportional to the velocity
of the liquid averaged over the imaginary line connecting the two
transceivers, and is independent  of the velocity of sound in the
measured liquid.  Depth is measured using a  conventional ultrasonic
level  gage,  and the velocity times  area multiplication to calculate flow
is performed electronically, and  indicated,  recorded and totallized.
Flow measurement accuracy is in the three to five percent range.  In-
stallation of ultrasonic velocimeter transceivers on the walls  of a
circular sewer is pictured in  Figure V-18.  A  single pair of  transceivers
         Figure V-18.  Ultrasonic velocimeter probes  installed
                   on  inside surface of a sewer wall
                     (courtesy Badger Meter  Inc.)
                                   72

-------
is customarily used, placed at a level corresponding to 25% to bQ% of
maximum flow.  At low flows when the  transceivers are close to the
surface, the Badger unit automatically reverts to a flow calculation
based on level only (Manning's equation).  This  is a reasonably good
approximation because flow more nearly reaches free flow conditions for
low depths.

The electromagnetic flow measurement  technique is based on the fact that
the motion of a conductive liquid  in  a magnetic  field will generate a
voltage in the liquid which is proportional to flow velocity and  in a
direction orthogonal to the field  and flow directions.  Embodiments of
this principle in flow gages suited to open channel gaging are pictured
in Figures V-19 and V-20.

The pitot-style transducer in Figure  V-19  is 25.^ cm (10  in)  in diameter
and has AC field coils built into  its cylindrical wall which generate a
vertically oriented, sinusoidally  varying magnetic field.  The output
voltage sensed by electrodes in the horizontal plane is a flow modulated
carrier in which the amplitude is  proportional to flow velocity.  Stated
accuracy is  ±0.5% for maximum flow velocities between 0.92 and 9-5
cm/sec  (3 and 31 ft/sec) flowing through the sensor, and ±\% for maximum
velocities between 0.305 and 0.92  cm/sec (1 and  3 ft/sec).

A smaller electromagnetic velocity sensor  is pictured in Figure V-20.
The probe wand shown in the figure foreground  incorporates a solenoid
which generates an  internal magnetic  field parallel to the probe axis
and an external field which is circularly  symmetric about the probe axis.
One or  two sets of electrodes are  mounted  in the probe, each of which
   Figure  V-19.   Pitot-type
   electromagnetic  flowmeter
     transmitter  (courtesy
      Fischer-Porter  Inc.)
Figure V-20.   "Velmeter" electro-
 magnetic flowmeter transmitter
 and indicator (courtesy Gushing
        Engineering Inc.)
                                   73

-------
picks up an output signal  proportional  to the flow component which  is
perpendicular to the plane formed by the axis of the electrode pair and
the probe axis of symmetry.  With two electrode pairs, two components
of flow velocity are measured simultaneously.  Full  scale velocities
are selectable between 0.092 and 9.2 cm/sec  (0.3 and 30 ft/sec).
Linearity Is stated as ±U of full  scale and zero flow offset as 0.003
m/sec (0.01  ft/sec).  The Instrument measures liquid velocity In the
immediate vicinity of the probe, and traversals can  be made to measure
average area velocity.

In each case, the flow sensor represents an obstruction to flow, and
liquid level must be measured by other means to determine total flow.

Sewage Flow Gaging  in Full Pipes -  A great many different types of flow
meters are available for gaging in full pipes, but few are well suited
to continuous or even intermittent measurements on sewage because of
entrained debris and grease.  An adaptation of a propeller which has
been applied in New York for intermittent sewage gaging in manholes is
sketched in Figure V-21.

An outer housing is wedged in place at the bottom of the manhole sealing
the sewer inlet to the housing.  Then a hinged flap gate is lowered into
the housing causing the housing to fill and overflow onto the manhole
floor and into the manhole outlet.  Finally, the propeller flow element
mounted within a cylindrical flow guide is lowered into the housing.
The flow guide is completely submerged in the flowing sewage, and all
flow passes through it,  so the propeller 'is operating in a full pipe of
known cross-section, and so that it can be calibrated directly  in units
of flow volume.  The unit, pictured in Figure V-22,  can be moved between
similar manholes, and total installation time is said to be under three
minutes.  Stated accuracy is three percent on totallized flow.

Ultrasonic and electromagnetic flow gaging techniques are well
established in full pipe flow metering applications.  A Doppler ultra-
sonic flowmeter is pictured in Figure V-23.

The acoustic probe containing two transducers is positioned in  the flow.
It is normally oriented  to point upstream, although the probe is bi-
directional.  One transducer projects an acoustic signal, which is
reflected by waterborne  particles and disturbances.   The reflected
signal,  frequency shifted proportionally to  the velocity of the fluid,
is received by the other transducer.  An electrical  signal proportional
to fluid velocity is converted to flow rate  by means of processing
circuits contained in the indicating receiver.  The actual point of
intersection of the two  beams is 0.305 m (one foot)  upstream from the
probe, so that measurements are made in an undisturbed region of the
flow field.   Because the processing circuits are not sensitive  to ampli-
tude variations,  the flowmeter is immune to  changes in impurity concen-
tration  and peripheral noise.
                                   74

-------
Figure V-21.   Installation sketch,  propeller meter for
  sewage flow gaging (courtesy Min-Ell  Company Inc.)
                          75

-------
Figure V-22.   Manhole insert housing and propeller meter
 for sewage flow gaging (courtesy Min-Ell  Company Inc.)


INDICATING
RECEIVER
r-*-1-
                         DIRECTION
                            OF
                           FLOW
      Figure V-23-   Doppler ultrasonic flowmeter
              (courtesy Edo Corporation)
                          76

-------
An automatic cleaning mechanism periodically withdraws the probe and
reinserts  it through wiping glands to  remove any accumulated deposits.
A portable version of this flowmeter is pictured in Figure \l-2k.  Both
units are  designed for service  in pipes from 0.15  to 2.A m (0.5 to 8 ft)
in diameter, measuring flow velocities between 0.03 to ^.6 m/second (0.)
to 15 ft/second) with a stated accuracy of  1.5%.

In Japan,  a flowmeter of  this general  design has been successfully
applied to measurement of biological sludges from  clarifiers.
            Figure V-2A.  Portable doppler ultrasonic flowmeter
                       (courtesy of Edo Corporation)
                                    77

-------
The electromagnetic flowmeter (Figure V-25) is in fairly widespread use
for sewage flow gaging in permanent, full pipe installations in diameters
from 0.15 to 2.^ m (0.5 to 8 ft).  Because the measurement is dependent
upon sensing a voltage induced in the stream, it is essential that the
pick-up electrodes be kept free of insulating deposits such as grease.
Various methods of electrode cleaning have been employed by the manufac-
turers of this type of meter including mechanical scrubbers, heaters and
ultrasonic cleaners.   Stated accuracies are ±k% at 0.15 m/sec (0.5
ft/sec), ±2% at 0.3 m/sec (1 ft/sec), ±\% at 0.6 m/sec (2 ft/sec) and
±0.5% at flow velocities from 1.22 to 7-62 m/sec (k to 25 ft/sec).
              Figure V-25.  Electromagnetic flowmeter for full
            pipe application (courtesy Fischer 6 Porter Company)


The singaround type ultrasonic flowmeter described earlier  is also used
in full pipe gaging applications.  The meter pictured in Figure V-26
makes use of ultrasonic transceivers strapped to the exterior surface of
the pipe wall.  Deposits on interior pipe walls have not caused diffi-
culties.  Although flowmeters of this type were not extensively used in
this country until recently because of circuit complexity and cost,
several hundred are in service in Japan and Western Europe.

The use of modern electronic integrated circuits has markedly reduced
the cost and improved the reliability of the ultrasonic flowmeter.  The
signal processing circuitry of one such meter is pictured in Figure V-27.
Stated accuracy and resolution are ±U of full scale, and bi-directional
flows can be metered.
                                   78

-------
Figure V-26.  Ultrasonic flowmeter for full  pipe
   applications (courtesy Badger Meter Inc.)
                      79

-------
            Figure V-27.   Signal  processing unit, ultrasonic
               flowmeter  (courtesy Badger Meter, Inc.)


Summary, Comparison and Recommendations - Achievement of accurate flow
gaging in a combined sewer system during and after a storm event
requires the flow gages to perform accurately over flow ranges as high
as 100:1, and abrupt changes in flow rate and level.  While many of the
metering devices described in the preceding sections perform well in
measuring slowly changing dry-weather sewage flows, their performance
suffers during storm flows.  Ideally the flow gage should continue to
function accurately from  minimum dry-weather flow to full pipe, sur-
charged conditions.  This requirement would clearly not be met by flumes
or weirs which are totally submerged during surcharged conditions, nor by
full  pipe flowmeters which cannot handle flow in partially filled
conduits.  This ideal requirement can, however,  be met by some of the
techniques under certain  conditions, and less stringent but still useful
performance requirements  can be met by others.

For example, if the hydraulics of a particular  installation allowed flow
to be gaged sufficiently  accurately by means of  level alone, using the
Manning equation, and if  the level probe (ultrasonic, dipper, bubbler,
scow, float-drywel1) had  sufficient range to encompass the full pipe
measurement, flow could be gaged over its entire range.  Also, if flow
average velocity were to  be measured with sufficient accuracy, say by a
multiplicity (three pairs) of ultrasonic transceivers, and flow area by
an ultrasonic level gage, then flow could be measured over the entire
range by multiplying velocity by area.
                                    80

-------
Another technique researched at the University of  Illinois  is a flume
structure designed to generate a head differential useable  for gaging
under both partially filled and full pipe, pressure flow conditions.  The
structure consists of cylindrical segments whose diameter is smaller than
that of the sewer, and which are attached to the sides of the sewer,
leaving the invert and crown clear.  It functions as a critical depth
flume under open channel flow, producing critical flow and  functions as
a conventional Venturi meter under full flow conditions.  Although
demonstrated  in laboratory scale, this method has not yet been evaluated
in a full scale application on sewage.  The method combining a flume for
open channel  flows, with automated tracer dilution and sampling for
storm flows also offers flow information over the full range of flows,
although attainment of adequate information rate for the storm event may
well require  the taking of a great many samples.  The measurement based
on Manning's  formula could likely provide accuracies no better than \Q%,
the combined  ultrasonic velocity/area measurement requires  equipment
costing in the  $12,000-$l8,000 range, and the universal flume has not
yet been fully demonstrated, so it must be concluded that no generally
applicable technique is presently available which fulfills  the ideal
requirement.  However, as costs for ultrasonic signal processors continue
their trend downward, this method appears to have the potential for
meeting the full range of metering requirements at acceptable cost.  It
has the further advantages of no moving parts, and no requirement for
electrical contact with the flow stream.  The method seems  well suited to
measurements  on grease and debris laden sewage, and for extended life in
the sewer environment.  A limitation of singaround type ultrasonic
velocimeters  is their susceptibility to entrained air bubbles.  An excess-
ive concentration of bubbles renders the stream opaque to ultrasound and
the velocimeter inoperative.  Accordingly the transceivers  must be
installed at  sites where flow is fairly quiescent and sufficiently down-
stream from sidestreams falling into the main stream from above.  Whether
they would continue to function at most sites under the effects of
agitation and air entrainment due to multiple stormwater inflows remains
to be demonstrated.

In sewers not subject to surcharging where hydraulics allow the Palmer-
Bowlus flume  to operate within acceptable submergences, these structures
are fully suitable, and flow head can be measured by any of the level
gages described in the preceding section.  The recommendations made there
apply also when they are used for level gaging in flow measurement
applications.

IN SITU MONITORING EQUIPMENT

There are two general categories of  Instruments suited for  continuous,
In-situ analysis of water quality parameters:  those that make use of an
in-stream sensor Incorporating an optical, electrochemical, ultrasonic
or electronic sensing device, and those that require the pumping of a
side stream to a close-by analyzer which employs electrochemical,
                                    81

-------
chemical/photometric, oxidation,  respirometric, optical, biochemical or
other automated analytical  methods.   Parameters for which there are
available commercial in-stream sensors  are the following:

      Chloride                        PH
      Conductivity                    Salinity
      Dissolved Oxygen                Suspended Solids
      Fluoride                        Turbidity
      Nitrate                         Temperature
      Oxidation Reduction Potential

Water quality parameters for which there are available commercial side-
stream analyzers are:

      Acicity                         Silica
      Alkalinity                      Silver
      Ammon i a                         Sod i urn
      Ammonia Mitrogen                Sulphide
      Biological Respiration          Sulphite
      Bromide                         TC
      Cadmium                         TOC
      Chromate                        TOD
      COD                             Total Chromium
      Copper                          Turbidity
      Cyanate                         Total Hardness
      Cyanide                         Nitrite
      Iodide                          Oil
      Iron                            Orthophosphate
      Lead                            Residual Chlorine
      Suspended So!ids

The key chemical and biological measurements to be made  regularly on
storm generated discharges as recommended  in Section VI  include  potential
oxygen demand (TOD and BOD^). particulate concentration  (suspended
solids), a pathogenic indicator (fecal  coliform), and eutrophication
potential (N02, NO^, TKN and TP).   Several of  the  in situ monitors  listed
above are in these general measurement categories, and a number  of  others
can be used to monitor specific parameters known to be  important at
particular sites.

Potential Oxygen Demand

The TOD analyzers pictured  in Figures V-28 and V-29 function  by  completely
oxidizing the entrained oxidizable matter  in a sewage sidestream and
measuring the amount of oxygen depleted  in the reaction.  The analyzer
pictured in Figure V-28 makes use of a A ml/min sidestream  pumped to the
unit where it is combined with a metered air stream.  The mixture is
delivered to a reaction chamber maintained at  8500 C, and the combustion
products then enter a liquid/gas separator where condensible  vapors are
                                     82

-------
oo
CO
  Figure  V-28.   TOD  analyzer (courtesy Astro
                Ecology  Corp.)
Figure V-29.   TOD analyzer (courtesy
          Phillip,  Holland)

-------
condensed and removed.  The remaining gases consisting mainly of original
metered air less oxygen depleted in the combustion process are routed to
a solid electrolyte fuel cell  where the oxygen depletion is measured and
translated into mg/1 TOD.  This unit has TOD ranges from 0-100 to
0-20,000 mg/1 full scale, has  a ten minute response time, and stated
repeatability is ±2% of full  scale.

The TOD analyzer pictured in Figure V-29 is functionally similar but makes
use of two zirconium oxide cells, the first to impart a known quantity of
oxygen to a nitrogen carrier gas, and the second to restore the amount of
oxygen to its original level  and thereby measure the amount depleted in
the combustion reaction.  Sample size in this instrument is 10 microliters.
Operation can be made stepwise continuous by means of a pump and automatic
injection device.  TOD ranges  are 0-100 up to 0-10,000 mg/1 ful1 scale,
measuring cycle  in five minutes, and stated accuracy and reproduceabi1ity
are ±10% of full scale.

Total organic carbon can be measured continuously by the instrument
pictured in Figure V-30  and in discrete samples by the unit  in Figure
V-31.  Total carbon is measured by oxidation of the sample at 850° C and
subsequent measurement of the  resulting carbon dioxide by an  infrared
analyzer.  To measure TOC, the sample is first acidified to remove the
inorganic carbon prior to combustion (Figure V-30). A second method  is to
measure the carbon content of  the C02 released due to acidification and
subtract this measurement from the measured total carbon to yield a
measure of organic carbon (Figure V-31).  For the Astro Ecology unit,
sample flow is k ml/min, stated repeatability is ±2% of full scale,
response time is five minutes, and full scale ranges vary from 0-25 to
0-5000 mg/1 of carbon.  In the Beckman unit, sample size is 250 micro-
liters, response time is two to four minutes per sample, full scale
ranges vary from 0-5 to 0-^000 mg/1 of carbon, and stated repeatability
is ±2% of full scale from 50 to 4000 mg/1  of  carbon  and ±5% at 5 mg/1
carbon.

Dissolved oxygen can be measured continuously by means of polarographic
or galvanometric cells.  One galvanometric type pictured in Figure V-32
makes use of a lead anode and  platinum cathode immersed in a potassium
iodide electrolyte which is retained by a teflon membrane.  Since teflon
is permeable to gases, oxygen  diffuses to the cathode at a rate that is
proportional to the partial pressure exerted on the membrane by the
oxygen dissolved in the liquid.  The cell generates a galvanic potential
of 0.5 volts and if the circuit is closed through a resistor, a current
flows which is proportional to the rate at which oxygen is reduced at
the cathode.  This rate is proportional to partial pressure and the
dissolved oxygen level at a particular temperature.  As temperature
decreases,  oxygen partial pressure decreases for a particular dissolved
oxygen concentration,  and resistance to oxygen diffusion through the
membrane increases.  Both these effects reduce output current as temper-
ature decreases, so temperature effects are compensated electronically
                                    84

-------
00
Ol
          Figure  V-30.   TOC  analyzer (courtesy
                 Astro  Ecology Corp.)
                                                                                                 j
Figure V~31.   TOC analyzer (courtesy
    Bechman Instruments, Inc.)

-------
            MODEL A-40
            D. O. PROBE
             INLET-
             OUTLET
             BAFFLE
             MODEL A-25
             CLEANER-
             AGITATOR
             SAMPLER
Cross-section view of
Model A-40 Probe and
Model A-25.
Figure V-32.  Dissolved oxygen  sensor
       (Courtesy  Rexnord  Inc.)
                    86

-------
based on a temperature measurement made with a thermistor.  For a 1 mil
membrane instrument response time is 30 seconds upscale and 90 seconds
downscale.  Ranges are available from 0-15 parts per billion to 0-15 parts
per million, and stated accuracy is one percent of reading.

Pa r t i c_u 1 a te Concen t ra t J on

A number of electro-optical instruments are available for measurement of
turbidity, or suspended solids, some making use of an in-stream sensor,
and others a side stream to the instrument.  Pictured in Figures V-33.
V-3^, and V-35 is an instrument with a self-cleaning in-stream sensor
that makes use of both measured light transmi ttance and light scattering
to obtain a linear response and to eliminate color interference over a
large range of suspended solids concentration.  The probe  is pictured  in
Figure V-33, the optical and self-cleaning features in Figure V-3^, and
the readout unit in Figure V-35-  Various models of this instrument from
0-30 to 0-3000 mg/1 suspended solids with a stated resolution of ±\% of
full scale are available.
Another  instrument developed under USEPA Contract No.  14-12-^9^ and now
being demonstrated under USEPA Grant No. 5802^00 makes use of an  incident
beam of  polarized  light and measures the degree of polarization in light
multiback scattered  (150°) from  the  incident beam.  Earlier tests of this
technique by American Standard Inc. on  raw  sewage and  biological  sludges
inidcated that the method could  be independent of particle size and color
over useful ranges of solids loadings  (53)-

Turbidimeters which  operate on a side stream are pictured  in Figure V-36.
The units pictured in Figure V~36 use a "surface scatter" principle where
light scattered from an incident beam  impinging on a  liquid surface
exposed  at  the top of an overflowing tube  is measured with a photocell.
Ranges for  this instrument vary  between 0-0.2 and 0-5000 nephelometr ic
turbidity units.  Side stream flow is  1 to  2 liters/minute (1A to 1/2
gpm) , response time  is 30 seconds, and  standardization is performed with
a standard  reflectance plate having a fixed MTU value.

Eutrophication Potential

Nitrate  concentration can be measured directly  in the  10   M to 10  M
range with  a liquid  ion exchange membrane electrode usually mounted in a
side stream analyzer.  Total organic nitrogen can be measured directly as
ammonia  with a gas sensing electrode after  the sample  is digested by the
Kjeldahl method.  Orthophosphate, nitrates  and nitrites can all be
measured instrumental ly using automated wet chemical methods.

Monitoring  of Other  Specific Parameters

As mentioned earlier in this section, these are  in-stream  sensors
available for pH, electrical conductivity,  ORP, salinity,  fluorides and
                                     87

-------
Figure V-33-  Self-
cleaning, submersible
suspended sol ids
sensor (courtesy
Biospherics,  Inc.)
                                                      RECIPROCATING
                                            PHOTOCELL    PISTON
                              SAMPLING
                              CHAMBER,
                                             V   \ \ \ \ \\\ \ \
                                                LIGHT
                                                SOURCE
                             Figure V-3^.   Cross section of  suspended
                                   solids  sensing head (courtesy
                                         Biospherics, Inc.)
                               Figure  V-35-   Control unit  for  suspended
                                       solids monitor (courtesy
                                          Biosphertcs,  Inc.)
                                   88

-------
Figure V-36.  Surface scatter turbidimeters (courtesy Hach Chemical Co.)


chlorides.  Typical analyzers for pH and conductivity are pictured in
Figures V-37 and V-38.  A wide variety of in-stream, side stream
electrode, and automated wet chemical analyzers are available for other
specific parameters, e.g., oil concentration by ultraviolet absorption,
oxygen demand by bacterial respiration, live cell concentration by
bioluminescent reaction of ATP with  luciferin and luciferinase, chemical
and elemental ions by solid state, liquid, gas and flow-through
electrodes.

In summary, of the water quality measurements on storm generated
discharges recommended in Section VI, in-stream monitors are available
for measurement of turbidity, suspended solids and nitrates, and side
stream analyzers are available for TOC, TOD, orthophosate, nitrates and
nitrites.  Other in-situ monitors for parameters characteristic of
particular storm generated discharges are also available.  There
continues to be a considerable need however for research, development
and demonstration of new, more reliable, interference-free instruments,
especially for organic loading, suspended solids, total phosphorus and
pathogenic microorganism Indication.   Therefore, no In-situ measurements
are recommended to replace any of the laboratory procedures recommended
                                   89

-------
                                                      CONDUCTIVITY
     Figure V-37-   pH analyzer
       (courtesy Leeds 6
             Northrup)
Figure V-38.  Conductivity analyzer
(courtesy Ecologic Instrument Corp.)
in Section VI.  However, the in-situ monitors are recommended  if they are
used for flow control purposes, quick indicators of treatment  efficiency,
or for monitors of changes in sewage characteristics.


RAINGAGE NETWORKS

Rainfa 11-Runoff Re 1 at ionsh i p

Precipitation is a general term used by hydrologists to describe all types
of moisture that can fall from the clouds to the ground.  Rainfall  is the
primary form of precipitation which is of concern in storm generated
discharge projects.   Although a correlation has been found between  the
snow that fell when the ground was frozen and the early spring runoff
(1), the term precipitation will be used interchangeably with  the term
rainfal1.

For precipitation to occur,  water vapor must be present in the atmosphere.
Something must bring about a cooling of the air so that the moisture will
condense to form water droplets.  Cooling of large regions of  air is nec-
essary for any significant amount of precipitation to occur and this is
                                    90

-------
usually achieved by a lifting of the air.  Classification of precipita-
tion can be made based on the factor causing the air lifting phenomenon.
These classifications are, 1) cyclonic precipitation,  2) convective
precipitation, and  3) orographic precipitation (2).

Cyclonic precipitation is caused by the movement of air masses from high
pressure to low pressure areas.  These pressure differences are caused
by unequal heating of the earth's surface.  Cyclonic precipitation can
also be categorized as frontal or nonfrental.  There are two types of
frontal cyclonic storms.  In the warm front type, warm air replaces cold
air; the second type of front is where cold air replaces warm air and is
called a cold front.  If there is no movement of the front, it is called
a stationary front (3)-

Convective precipitation is caused by the heating of the air near the
earth's surface.  The heated air expands and water vapor is taken up.
As the warm moist air rises and  is surrounded by cold dense air,
precipitation occurs.  These types of storms are usually quite variable
or spotty and may produce light  showers or high intensity rains often
called thunder storms.  Convective precipitation, because of its
variability,  is often the most difficult to accurately record and usually
is the limiting factor in raingage network design.

Orographic precipitation is produced by  rising air caused by the top-
ography of the land.  This type  of rainfall can vary significantly in
intensity and quantity.  These variations are particularly pronounced
in mountainous and hilly regions of the country.  A rise of air on the
windward side of a slope causes  warm air to move upward and precipita-
tion forms as the warm moist air comes  into contact with the cooler air
at higher altitudes.  Thus, rainfall tends to occur on the windward side
of major slopes or mountains.

Precipitation can also be classified as  to form and intensity.  A "trace"
is recorded when precipitation is  less  than 0.127 mm (0.005 in.).  A
"drizzle" is usually  less than 1.016 mm/hr (0.04  in./hr) and is made up
of water drops under  0.508 mm  (0.02 in.) diameter.  "Rain" is classified
as greater than 1.016 mm/hr  (0.04  in./hr) and consists of drops larger
than 0.508 mm (0.02  in.)  (2).  "Thunderstorms" are high  intensity short
duration  (15~30 min.) forms of precipitation that are usually localized.

Both spatial and temporal variations in  precipitation events cause major
problems  in measuring rainfall from which storm generated discharges
result.  A common term "areal rainfall"  is defined as the average precip-
itation over an area  during a given time period (4).  Raingage data
is usually weighted using the Theissen method  (discussed later) (2)(5) or
averaged to arrive at the areal  rainfall value for a given storm.
Variations in rainfall require an adequate raingage network to arrive at
an areal rainfall value that will accurately predict the resultant storm
generated discharges.
                                    91

-------
A greater raingage density is necessary to accurately describe precipi-
tation  in mountainous terrain than is needed in flat or gently undulat-
ing terrain.  Mountains affect the quantity and distribution of  rainfall
as well as the areal variability.  Hutchinson (6) studied the effect of
topography on the areal variability in order to initiate a more  quanti-
tative  procedure for determining necessary raingage densities.   He was
primarily concerned with improving the understanding of the areal
variability of precipitation.  According to Huff  (7) factors that can
contribute to significant variations in storm measurements include   1)
stage of development of the storm,  2) size and complexity of the storm
system,  3) rainfall type,  A) storm type,  5) location of sampling area
with respect to the storm center, and  6) movement of the storm  system
across  the sampling region.  Many of these variations are due to the
frontal variations and cellular nature of convective rainstorms.  Huff
(7) in  studying heavy storm characteristics described four major types
of storm patterns as,  1) closed elliptical,  2) open elliptical,  3)
multi-cellular and  4) bonded.  He was interested in modeling storms for
hydrologic applications.   The best correlation between actual storm data
and predicted values from the model were obtained by relating the per-
cent of storm rainfall to the percent of total storm time; the storms
were then classified according to the quart!le of heaviest rainfall.  In
developing this model, Huff defined each storm as the rain period which
is separated from previous or successive rain by six hours or more.  He
also required that all raingages measure at least 12.7 mm (0.5 in.) and
one gage had to measure more than 25-4 mm (1.0 in.).

A number of attempts have been made to predict various types of  rainfall
so that mathematical rainfall models can be used to study different
characteristics of runoff hydrology (8)(9)(10).   It has been necessary
to develop these types of synthetic data because of the lack of  adequate
historical data for the spatial  and temporal  variations that often occur
in convective storms.  In order to accurately monitor these rainfalls,
very dense raingage networks are necessary, and yet few of these types
of networks are available.  In runoff predictions an average or  areal
rainfall is determined over the watershed area in question.  The average
rainfall over an area can be determined by an arithmetic mean, the
Thiessen method or the isohyetal  method (2).  The arithmetic mean tech-
nique is useful  only where the gages are uniformly dispersed and the
rainfall does not vary excessively throughout the area.  The Thiessen
method usually predicts a more accurate reading of areal rainfall.   The
polygons formed in the Thiessen method require that at least three
raingages be spread over  the watershed for proper weighting of the rain-
fall  data.   Every time a  change is made in the raingage network, a new
Thiessen diagram must be  developed.  The isohyetal method is the most
accurate method for determining areal  precipitation.  However, an
extensive raingage network is necessary and the analyst must carefully
account for the watershed topography and the rain storm characteristics.
                                   92

-------
When insufficient data are available, rainfall formulas may be used  (k).
These formulas relate rainfall intensity  (mm/hr  [in./hr])with storm
duration by use of constants that are tabulated  for different regions.
Areas of very high rainfall  intensity variation  use special charts and
figures to determine rainfall intensity as a  function of return frequency
and duration for storm sewer design.  Other methods have been established
for predicting rainfall data where  insufficient  raingage information  is
available (7)(9)-  In all cases, great care must  be taken when designing
combined sewer overflow systems and predicting rainfall data  in the
absence of historical data.  Storm variations can produce major errors
in the prediction techniques and the use  of historical data cannot be
over emphasized for good designs.

In storm generated discharge studies, the primary interest in rainfall
lies in the runoff that is produced.  Combined sewer overflows caused
from runoff in urban areas are significantly  different than runoff from
natural or rural areas, primarily because of  the  impervious nature of the
watershed.  Significantly  less infiltration of rainfall to groundwater
will occur in urban areas.

Several methods have been used to predict runoff  for storm sewer design
from precipitation data (4).  A number of empirical formulas have been
developed to relate one or more parameters with  expected runoff rates
and volumes (11).  The Burkli-Ziegler and the McMath formulas relate
runoff with maximum rate of  rainfall, slope of the ground, area of
watershed, and a factor to express  the nature of  the ground surface.

     Burkl i-Ziegler:     0. * CRA    /s7A~

     McMath:             Q - CRA  5/s7A


     where:  Q = runoff, cfs
             R = maximum rate of runoff,  in,/hr  over entire area
             S = slope of  the ground surface, ft/1000 ft
             A » area, acres
             C = nature of ground surface or  relative imperviousness

Runoff from areas greater  than ^05  ha  (100 acres) has been related to the
size of the drainage area  (II).  Several  are  listed below.
     Metcalf and Eddy:    0_ =  -jc +  '5


     Fuller:              Q = CM0'8

     Fanning:             Q = ZOOM5
     Talbet:              Q = 500MI/2*
                                    93

-------
     where:   A = area in acres
             M = area in square miles

These rational type formulas vary greatly from each other
because they are developed for  specific areas or conditions.  In most
cases the frequency of the computed flow Is not known   The use of these
formulas has greatly declined in favor of the Rational Method.

These designs do not  relate storm flows  to actual  rainfall  frequency.
Factors such  as variations  in population have appreciable effects on
this type of  design.
The Rational Method of storm and combined sewer system design relates
runoff to rainfall intensity, area of watershed, and a runoff coefficient
or  imperviousness factor.  The Rational Method has been expressed as
Q = ciA and as Q = AIR (11) (13)-  These two equations are equivalent
since c = I, i = R, and  in both cases A = A.  The form Q = ciA will be
used in this text, where the units are runoff, Q(m3/hr[cfs]), intensityj
i(cm/hr[in./hr]),  and area, A(ha[acres]).  The rainfall intensity factor
i depends on the frequency of the storm selected and the duration of the
storm which is chosen to be equal to the time of concentration of each
drainage area.  The time of concentration is defined as the maximum time
required for a drop of runoff to flow from the furthest point in the
drainage area to the point in question (13)-  This is usually divided
into the inlet time, a time of flow across the drainage area, and the
time of flow in the sewer.  The time or concentration is used to
determine rainfall intensity values from intensity-duration-frequency
curves similar to Figure V-39.  Frequencies of 1 to 10 years are commonly
used where residential areas are to be protected.  Higher return period
storms may be used where high value districts are involved.  The runoff
coefficient c  must be selected by the engineer, and requires some judg-
ment to correlate*,local  conditions.  Typical values of G are shovm in
Table V-l.
The Rational Method should usually be used where drainage areas are
smaller than 0.810 ha (2 sq mi).  It should be emphasized that the
Rational Method only predicts peak runoff flow rates.  The unit hydro-
graph method has been used because it predicts the runoff characteris-
tics throughout a storm.  In the combined sewer overflow study at
Bucyrus, Ohio (14) the Rational  Method was not adequate to predict peak
runoff values and a modified unit hydrograph was used.

Although the unit hydrograph is  the most common hydrograph approach to
rainfall runoff analyses, there  are other approaches that have been used
for specific types of problems.   The unit hydrograph is linear and linear
analyses are often used in other hydrograph techniques.  The primary
                                    94

-------
                            RETJRM
ERI
)D-Y
EARS
     20       kO      60      80      100
               RAINFALL DURATION,  m\n
           20
Figure V-39-   Intensity - duration - frequency
           curves for Bucyrus,  Ohio

              From Reference 1^
                    95

-------
                   Table V-l.   RUNOFF COEFFICIENTS Ca
     Surface Type                                      C  Value

 Bituminous streets                                  °-70  to  °-95
 Concrete streets                                    0.80  to  0.95

 Driveways, walks                                    0-75  to  0.85

 Roofs                                               0-75  to  0.95

 Lawns;  sandy so?1
      Flat, 2 percent                                0.05  to  0.10
     Average, 2  to 7 percent                        0.10  to  0.15
     Steep, 7 percent                               0.15  to  0.20

 Lawns;  heavy soil
      Flat, 2 percent                                0.13  to  0.17
     Average, 2  to 7 percent                        0.18  to  0.22
     Steep, 7 percent                               0.25  to  0.35


 a.  From reference 13


concept involving the linear analysis of hydrographs is that output is
directly proportional to the input.  In the case of runoff analyses, the
peak discharges  of two hydrographs are proportional to the amounts of
rainfall producing runoff that constitutes the hydrograph.  Mitchell (15)
evaluated a large number of small  drainage basins for culvert design
where a more rapid method of analyses than the unit hydrograph approach
was needed.  He  described linear hydrographs using five parameters:
 1) the size of  the drainage area,  2) the amount of rainfall producing
runoff  (rainfall  excess),  3)  the duration of the rainfall producing
runoff,  4) the  lapsed time from the end of the runoff produced by rain-
fall to the point of inflection on the recession limb of the hydrograph
(translation time) and  5) a storage Index.  Using these parameters and
a set of predetermined tables, any linear flood hydrograph for a given
drainage basin can be determined.

A somewhat different approach was developed using a regression analysis
of many urban basins throughout the United States (16).  Hydrograph
parameters were  described by five equations using the area of the water-
shed, length of  the main drainage channel, slope of the subarea and the
impervious nature of the surface.   These equations, however, do not take
into account the rainfall Intensity, duration, or antecedent rainfall.
Eagleson and Schaake (17) have also used linear hydrograph analyses to
analyze drainage basins.  However, they converted from the time domain
to the frequency domain using Fourier transformations and examined
                                    96

-------
drainage basins as low pass frequency filters (analogous to electronic
circuits).  This approach permits a unique viewpoint and expands the
knowledge of rainfal1-runoff relationships.  However, the difficulties
in describing necessary hydrologic relationships mathematically makes
this approach difficult to apply in storm generated discharge studies.
Application of the unit hydrograph technique (subtracting Infiltration
and retention capacities) to predict the frequency of storm runoff of
various magnitudes from rainfall or snowmelts on small drainage basins in
various stages of urbanization was successfully performed in Michigan
under USEPA Grant No.  R-8009^1(110^0 DRS).  (54)

The unit hydrograph approach can be used in runoff studies or design and
has been utilized in computer models that will be described briefly later.
The primary disadvantage of the unit hydrograph method lies in that both
rainfall and runoff data are needed for  its derivation.  Also, it is
limited to a particular drainage basin  (13) •  The unit hydrograph
represents the runoff volume of 2.5^ cm  (1.0 in.) from a particular
drainage area for a specified storm duration.  Theoretically a different
unit hydrograph  is required for different  length storms.  However, unit
hydrographs for  short duration  rainfalls can be combined into hydro-
graphs for longer storms.  Other hydrographs for the same drainage area
can be derived from the unit hydrograph  using the following relationship.


                   Unknown discharge at  tx  (m /hr[cfs])
              Volume of Rainfall producing  runoff  (cm[in7j)

                discharge from unit hydrograph at  tx
                          2.54 cm  (1.0  in.)

Thus the runoff at any time tx can be determined for any rainfall
volume that has the same duration as that used to  produce the unit hydro-
graph.  Unit hydrographs can be established for several different storm
durations and maintained on file.  The  basic steps used to produce a
unit hydrograph are (13):

     1.  Analyze the sewer flow hydrograph  to permit separation of
         normal domestic wastewater flow from the  storm generated
         runoff.

     2.  Measure the total volume of direct runoff from the storm
         producing the original hydrograph.  The volume is equal to
         the area under the hydrograph  after the domestic wastewater
         flow has been subtracted.

     3.  Divide the ordinates of the direct runoff produced by the
         storm by the total volume (cm[in.]) of stormwater runoff.  A
         plot of these values versus time is the unit hydrograph.
                                   97

-------
     k   The effective duration of the stormwater runoff producing
         rain for this unit hydrograph is found by evaluating the hye-
         tograph of the rainfall.

Modifications of the unit hydrograph and methods of compositing indivi-
dual hydrographs for different sewers and catch basins are described in
the Los Angeles Method (18) and the Chicago Method (19).  Unit hydro-
graphs were used to evaluate storm discharges that flow to a river (14).
The Rational Method was evaluated  in this study but assumed runoff co-
efficient values (C) appeared to be in error.  Peak discharges were
predicted that were in excess of the sewer capacity.  A straight line
relationship was found between maximum rainfall intensity and peak over-
flow to the river for given storm  durations.

The  Rational Method was successfully used in the Des Moines, Iowa study
where the C values were determined by comparing plots of rainfall in-
tensity (in./hr) versus time and runoff flow (cfs/acre) versus time (20).
The runoff coefficient C was determined as the ratio of runoff per acre
to the rainfall intensity.  It is  necessary to determine C at the point
of maximum rainfall intensity and  to check that the storm duration is
greater than the time of peak flow.  Thus, the entire watershed will  be
contributing to the runoff flow and the peak flow can be assumed to
occur at the time of concentration.

In most combined sewer systems the design and evaluation involving the
use of rainfall-runoff data is complicated.  Many other parameters are
involved and simple evaluation using the  Rational  Method or  the
Unit Hydrograph Method is  inadequate.  Hence, several computer models
have been developed which can calculate rainfall-runoff relationships,
stormwater quality characteristics and other design and operational in-
formation.  These models have been summarized  (21) and are shown below.

     1.  HydrpgraphrVo1ume Method  - The Hydrograph-Volume Method (Ritter,
         1971) was deve1oped in Germany (22).  This model calculates the
         dry-weather flow and storm runoff, and routes the combined
         flows through a complex sewerage network.   Its benefits appear
         to be in conduit sizing and design.  The model does not simulate
         flow quality or control regulators.

     2.  Road Research Laboratory Hydrograph Method - The British Road
         Research Laboratory (RRL) Method uses storm rainfall to
         provide a stormwater runoff hydrograph for the purpose of storm
         drainage design (23).  Rainfall  is applied to the paved area of
         the drainage basin which is directly connected to the storm
         drainage system.  Travel  time to the nearest storm drainage
         inlet is computed for various  increments of the total paved
         area that  is directly connected.  From this time-area informa-
         tion, the surface hydrograph arriving at  the  inlet  is computed.
         The surface hydrograph is then routed through the storage
         available  in a particular section of pipe.  The surface
                                    98

-------
hydrograph at the next downstream inlet is added, and the
combined hydrograph is routed on downstream.  Thus, the
successive addition and routing of surface hydrograph produces
an outflow hydrograph at the downstream discharge point.  In
the RRL method, the quality associated with the runoff  is not
computed.  The application and use of the method is described
in a recent USEPA report (2*0 .

Stanford Watershed Model (Hydrocomp Simulation Program) - The
Stanford Watershed Model (Crawford and Lins ley, 1966)(25) along
with its commercial successor, the Hydrocomp Simulation Program,
is a comprehensive mathematical model that simulates watershed
hydrology and flow routing.  This model has been used
extensively to simulate existing and .planned surface water
systems.  Recently, it has been expanded to include water
quality computations.   It does not perform cost calculations,
however, and has not been .adapted to sewerage systems»

Urban Runoff Model - The Urban Runoff Model PDP-9 (UROM-9) was
developed at the University of Minnesota for the Metropolitan
Sewer Board, St. Paul, Minnesota (26).  The purpose of  this
model is to predict discharges.to the Minneapolis-St. Paul
interceptor sewers, given rainfall readings at various  points
around  the Twin Cities.  The model computes storm runoff from
major catchments, combines the.runoff with estimates of dry-
weather flow, routes the combined flows through the interceptor
system  to the treatment plant, and computes overflows to the
receiving water at control regulators.  It uses monitored
rainfall and flow level data from various points for real-
time control of the overflows.  The model has not been  adapted
to consider water quality aspects of the overflows.   It is not
intended for use for design purposes.

Urban Wastewater Management Model - The Urban Wastewater
Management Model (Ba t teI ]e-Northwest and Watermation, Inc..,
1972) is a comprehensive mathematical model developed to
continuously simulate time-varying wastewater flows and
qualities in complex metropolitan combined .sewerage systems
(27) (28).  The model simulates major sewerage, system components,
such as trunk and interceptor sewers, regulators, storage
facilities, and treatment plants.  It provides a means  of
evaluating the time-varying performance of a planned or exist-
ing sewerage system under a variety of rainfall conditions
(considering both time and spatial rainfall variations) without
simulating every pipe or manhole.  The model simulates  seven
wastewater quality parameters:  SS, BODr, COD, phosphate,
nitrate, ammonia, and Kjeldahl nitrogen.  The required  opera-
tion of control regulators during real-time rainstorm events
to minimize overflows is modeled.  The model can also be used
                         99

-------
         for design and planning studies.   It computes sizes and costs of
         structural sewer system modifications, such as sewers,
         regulators, and storage and treatment facilities, that will
         result in the least-cost combination of alternatives for improv-
         ing system performance.

     6.  Storm Water Management Model  - The Storm Water Management Model
         (SWMM) (Metcalf 6 Eddy, Inc., University of Florida, and Water
         Resources Engineers, 1971)  was developed under the sponsorship
         of the U.S. EPA (29) (30),(31 ).(32) -   It is a comprehensive math-
         ematical  model capable of representing urban stormwater runoff,
         storm sewer discharge, and  combined sewer overflow phenomena.
         The SWMM has been demonstrated at  more than 20 sites throughout
         the country ranging from 76 to 8,100 ha (187 to 20,000 acres).
         During demonstration,  the SWMM has been verified to be capable
         of representing the gamut of  urban stormwater runoff phenomena
         for various catchment  systems (30).  This includes both quantity
         and quality from the onset  of precipitation on the basin,
         through collection, conveyance,  storage and treatment systems,
         to points downstream from outfalls that are significantly
         affected by storm discharges.  The SWMM is intended for use by
         municipalities, governmental  agencies, and consultants as a
         tool for evaluating the pollution  potential of existing systems,
         present and future, and for comparing alternative courses of
         remedial  action.  The  use of  correctional devices in the catch-
         ment, along with evaluation of their cost effectiveness, has
         also been demonstrated.

Ra i n fa 11  Mea s uremen t

There are basically two types of raingages.  The non-recording type
measures  total precipitation and is  manually read using a measuring
stick which is calibrated to 0.25^ mm  (0.01 in.) of rainfall.  The
standard  raingage used by the U.S. Weather  Bureau has a 20.32 cm (8 in.)
diameter  collector which directs the rain  into a measuring tube.  The
measuring tube is 1/10 the area of the collector, thus 2.5*t mm (0.1 in.)
rainfall  will fill the measuring tube  to a  depth of 25.4 mm (1 in.).

The recording type raingages measure the intensity of rainfall by
continuously recording the amount of rainfall throughout the storm.  This
is achieved by use of a tipping bucket gage (U.S. Weather Bureau) or a
weighing-type gage.  The tipping bucket gage measures rainfall intensity
by recording the number of times a small  bucket is filled and tipped or
dumped into a reservoir.  Two buckets  are used so that one bucket is
emptying  into the reservoir while the  other is filling from the rainfall.
Each bucket is calibrated so that 0.254 mm  (0.01 in.) of rainfall will
tip the bucket into the reservoir.  The weighing type raingage utilizes
a weighing mechanism to continuously measure the rainfall as it passes
into a large bucket.  The increase in  the weight of the bucket is related
to the amount of rainfall.  Heating  devices or other types of raingages
                                   100

-------
are used where excessive snowfalls contribute to a major amount of
moisture to the precipitation.

Errors that arise from raingage measurements include   1) mistakes  in
reading the scale of the stick or chart,  2) evaporation of water  in the
gage,  3) improper placement of the gage, and  4) instrument error.  It
has been suggested that 6 hour recording gages be used  instead of 2*t hour
to facilitate the reading of short time  interval rainstorms when weigh-
ing type raingages are used  (33).  It was found that errors caused by
skewness of the charts was significantly reduced for three different
types of recording raingages when six hour charts were  used.  Errors in
translating recorded data to runoff should be random in nature and tend
to cancel each other out over a period of time.  Most  evaporation errors'
would be insignificant in the high intensity, short duration, storms of
importance in storm generated discharge  studies.

The best site for raingage location is on level ground with no trees or
buildings in a proximity such as to affect the capture of rainfall.
Distant trees and buildings  that tend to break up wind  currents are
advantageous.  Errors due to wind affecting measured rainfall have been
rated as the most serious problem in precipitation measurement (2).  Wind
shields are available and should  be used where winds  are expected to
cause measurement errors.  Errors due to gage location  have been
minimized by choosing raingages with the lowest variance as permanent
stations (3*0.

Raingage Density

Variation in precipitation with topography, season, time, type of  rain-
fall, and other factors has  been briefly discussed.  Because of these
spatial variations, the concentration of raingages for  any given water-
shed must be carefully evaluated if accurate rain patterns are to be
measured.  A precise method  of determining the proper  number of raingages
for a given area has not been developed.  It has been  suggested that a
large number of temporary gages may be required before  the unnecessary
gages can be determined and  a selection made for the permanent stations
(35).   It is important in establishing an optimum raingage network to
carefully consider the limits and requirements that the rainfall data
will be used for.  What are  the objectives of the project?  Climatic
conditions, topography, and  normal type of rainfall should also be
considered in establishing raingage networks.

A number of investigations have been conducted to specify an adequate
raingage density to characterize precipitation patterns and/or runoff
from watersheds.  These studies follow two basic approaches:   1) math-
ematical approach to describe the watershed basin (36)  and  2) empirical
approach to correlate historical data with equations or curves (3*0 (35)
(37).  Examples of the different design methods will be discussed next.
                                   101

-------
Correlation fields (distance plotted for equal values of the inter-
station correlations of daily precipitation for each of the raingage
stations) were developed by Hendrick and Comer from extensive historical
data (35).  Raingage distance, azimuth, and minimum precipitation were
used to derive a combined correlation coefficient.  Using this combined
correlation coefficient, 0.9 correlation ellipses were used to graph-
ically determine the gage network that would be required for the 112
sq km (l»3 sq mi) watershed to predict summer precipitation.  When
distance was considered alone, the raingages needed to be about 1.21 km
(0.75 miles) apart for a correlation of 0.80.  The correlation coeffi-
cient decreased with distance at a rate of 0.062 per km (0.1 per mile)
for summer storms.  This resulted in a O.OAA5/sq km (0.21/mi) raingage
density requirement for the 112 sq km (^3 sq mi) basin.  These authors,
however, point out that major errors in precipitation pattern measure-
ment may occur due to periodic severe storms.  Correlation criteria used
in developing raingage networks is useful in minimizing errors over
extended periods of time, but the networks may not accurately measure
any one specific storm.

Taking a mathematical  approach, Eagleson (36) devised several  equations
for which graphs were presented to relate raingage density requirements
with watershed area and storm size.   Table V-2 below shows the number of
ratngages that were necessary to predict the error in maximum runoff
discharge for different size cyclonic storms.
                     Table V-2.  NUMBER OF RAINGAGES
            NECESSARY TO PREDICT RUNOFF FOR CYCLONIC STORMS3
Runoff
Error
5%
10%
15*
Ratio of
1
No. Gages
k
2
]
Watershed Length
2
No. Gages
7
k
2
to Effective
k
No. Gages
2k
10
5
Storm Radi i
10
No. Gages
k6
]k
7
a.  From reference 36.


Eagleson pointed out that fewer raingages are required to predict runoff
from watershed dynamics than when rainfall patterns are to be determined.
No actual verification was made to determine if the mathematical
predictions of Eagleson were actually valid.  However, it is important
that the number of gages required increases as the watershed area becomes
greater than the effective storm area.  Thus more raingages are required
for thunderstorms than for uniform rainfall.  Eagleson stated that two
properly located raingages are adequate to determine long term average
rainfall at many watersheds.
                                   102

-------
Osborn, et al, (37) used a 150 sq km  (58 sq mi) watershed  in  southeastern
Arizona with 95 recording raingages to study  the optimum  raingage  density
for thunderstorms.  They were  interested in the raingage  network nec-
essary to measure the rain volume and  intensity for  individual  storms  and
to predict the runoff from specific watersheds.  Osborn used  the maxi-
mum 15 minute rainfall and the total storm rainfall  to  investigate  the
optimum raingage density required to describe  individual  storms and  to
correlate rainfall with runoff.  Table V~3 shows the  results  expressed
as the range of simple correlation values between each  raingage for
several different watersheds.

  Table V-3-  SIMPLE CORRELATION COEFFICIENT  RANGE BETWEEN RAINGAGES3
Correlation coefficient based
Raingage spacing,
feet
506
2100
4200
3800
926
- 3,000
- 8,550
-19,300
Total
ra infal 1
0.91
0.82
0.00
0.00
- 0.96
- 0.91
- 0.71
- 0.80
on:
1 5 minute
maximum rainfal 1
0.86 -
0.73 -
0.00 -
0.17 -
0.92
0.90
0-57
0.84
a.  From reference 37-


Each watershed was chosen based on the distance between raingages.  A
plot of all their data showed that the simple correlation decreased
roughly by 10% for every 304.8 m  (1000 ft) of distance between rain-
gages.  Their results showed that gages should be placed every 548.4 m
(1800 feet) to describe the total rainfall for a correlation of 0.90.
To describe the maximum 15 minute rainfall, gages must be every
304.8 m (1000 feet) for a correlation of 0.90 and about 1400 gages would
be required for the 150 sq km (58 sq mi) watershed or about 9-3
gages/sq km (25/sq mi).  The costs for this extensive raingage network
would be prohibitive and the real concern  in most cases is the accurate
prediction of storm runoff, so another approach was taken.

Eagleson (36) and others have shown that the runoff from watersheds tend
to smooth out rainfall variations.  Osborn therefore decided to in-
vestigate the relationships between rainfall, raingage density, and
storm runoff using a multiple linear regression program.  From the basis
of this study, they concluded that the following gage densities are
necessary to relate watershed runoff to rainfall:

     1.  One centrally located gage for watersheds up to 48.5 ha
         (0.48 sq km)[120 acres(0.l87 sq ml)].

     2.  Three evenly spaced gages for a 260 ha (2.6 sq km)[640 acres
         (1 sq mi)] watershed with a length to width ratio of 4.

                                    103

-------
     3.  Five evenly spaced gages for a 2600 ha (26 sq km) [6*»00 acres
         (10 sq mi)] watershed with gages spaced about every 2.^2 km
         (] .5 miles).

These conclusions are not a great deal  different from the results of
Hendrick and Comer (35)  when only interstation distance was considered
for a correlation of 0.8.  However, Osborn,  et al  (37) developed these
raingage density guidelines to predict runoff while Hendrick and Comer
were concerned with accurate precipitation analysis.  Osborn, et al,
also found that areal rainfall estimated by the Theissen weighting
procedure produced better correlations than average raingage data.

Radar patterns were compared with data from twenty-one recording gages
and eight non-recording  gages in a California watershed.  Twenty of the
recording gages were located in approximately a 26 sq km (10 sq mi) area
that included stations in the urban area of the City of Davis (3^0 •  The
maximum distance between gages was 4.5 km (2.8 miles) and the minimum
distance was 0.8 km (0.5 miles).  Data from individual gages were
compared with the mean using a simple regression equation of the form:

                           X" = b  + b.  x.
                                o    1   i

              where:  X" = the mean precipitation of the entire array

                     x.  = cumulative precipitation value of an
                          individual gage at each 5 minute time interval
                 b , bj  = regression coefficients


The gages which produced the most consistent data from storm to storm
were selected as permanent raingage stations.  These authors found that
corrections should be made for systematic errors due to instrument
characteristics, gage location and other factors.   After corrections
were made for the systematic errors, runoff hydrographs were made based
on the mean network hyetograph and on individual gage hyetographs.
Comparison of these two runoff hydrographs showed that similar runoff
hydrographs were obtained for each of the corrected gage data because
of attentuation characteristics of the watershed.   They concluded that
any single gage record was as representative of the area mean as any
gage after each gage was corrected for systematic errors.  Also, it was
suggested that since any gage or group of gages were equivalent, lumped-
parameter methods for storm runoff could be used in design applications.
While these results were derived for topography variations that included
both rural and urban characteristics, it is likely that mountainous and
other varied terrain will require more sophisticated raingage networks.
Also, more extensive raingage networks will  be necessary for watersheds
that receive convective precipitation or thunderstorms.
                                   104

-------
Raingage Density in Storm Generated Discharge Projects

The number of raingages necessary for any network  is related to the size
and number of the catch basins or runoff areas  in  a municipality.   In
reviewing ASCE's Program Technical Memorandum No.  10 (38),  it was noted
that for four cities one-fourth of the catchment basins are smaller
than 2 to 26 ha (5 to 65 acres); one-half are smaller than 2 to 81  ha
(5 to 200 acres); and three-fourths are smaller than 20 to 200 ha (50
to 500 acres) (see Table V-4)  (39).   It was suggested that pilot rain-
gage networks over 20 ha (50 acres) would be representative of a large
number of catchments particularly in  suburbs beyond the corporate city
1imits.

Table V-5 shows a list of raingage densities for different storm
generated discharge studies.   Six studies had raingage network densities
of 3-85 gages/1000 ha (1 gage/sq mi)  or more.  With a uniform grid  of
raingages this would put the maximum  distance between stations at 2.24
km (1.4 miles).  The lowest raingage  density network was for the
southerly district of Cleveland, Ohio (40) where one gage was used  with
difficulty   to correlate wet weather  flow.   It was noted that heavy
rainfalls near the treatment plant could produce significant combined
sewer overflows when no rainfall was  being recorded at the raingage
station.

The rainfall data at Minneapolis-St.  Paul (41) was used to determine
when to  inflate or deflate fabridams  in order to direct stormwater  run-
off to maximize storage of the runoff in the sewers.  Rainfall data was
also used in the mathematical  model to determine overflow volumes and
other rainfall-runoff information.  This information was coupled with
monitoring  information on the  depth of flow  in  the interceptor sewers
so as to maximize flow routing and storage.  During this study the
thawing of  snow and ice caused runoff to the sewers that was relatively
easy to handle because the thaw process is slow and predictable.

Seattle sewers (42) were pumped out before storm runoff arrived to maxi-
mize in-sewer storage.  However, at least a  two hour notice was needed
in order to empty the sewers;  this was achieved by using rainfall data
from the outlying regions of the metropolitan area.  Rainfall data
were also used to predict sewer flows and estimate runoff volumes.  At
the time of this report (1974), correlation studies were to be made to
determine if more stations are needed.  Eventually rainfall data will
provide  input to a model that  will be used t!o develop system hydrographs.
These hydrographs will be used to route stormwater discharges through
the sewers  and maximize in-line storage.

In a Cleveland study (43) rainfall data were used  to characterize varia-
tions in rainfall over a large area and design  storm rainfall data  were
developed from these characteristics.  Runoff discharges were compared
for a 1 year storm rainfall using the Rational Method, the Routed
                                    105

-------
         Table V-*».  DISTRIBUTION OF  SEWERAGE DRAINAGE CATCHMENT SIZES  IN SOME MAJOR CITIES
Total area
of city
City
San FrancI sco
Washington
Milwaukee
Houston
sq
km
114
158
246
1160
sq
mi
1)1*
61
95
OH
No. of
caBch-
ments
1)2
93
465
1283
Largest
drainage
area,
ha acres
830 2050
2500 6180
71)0 1820
1040 2550
Average
drainage
area,
ha
227
152
38
26
acres
560
375
95
65
Median
drainage
area,
ha
77
26
10
2
acres
190
65
25
.4 6
Upper
quart! le
drainage
area,
ha
181*
108
28
19
acres
455
265
70
46
Lower
quart! le
drainage
area,
ha
26
10
4
0.8
acres
65
25
10
2
Upper 1 i mi-t of
size of catchments
representing
hal f of total
drained area of city,
ha
850
912
1 74
.34
acres
2100
2250
430
330
a.  From Basic Information Needs in Urban Hydrology  (39)-

-------
                     Table V-5.  RAINGAGE DENSITIES FOR DIFFERENT STUDIES


Drainage area
Minneapolis, St. Paul
Seattle
Cleveland
Cleveland Southerly Dist.
Bucyrus, Ohio Basin No. 8
Basi n No. 17
Basin No. 23
Roanoke, Virginia
Murray Run
24th Street
Trout Run
Philadelphia, Cobbs Creek
Boneyard Drainage
Des Moines
Detroit
Milk River Drainage
Racine, Wisconsin
San Francisco
Kenosha, Wisconsin
Humboldt Avenue, Milwaukee
a. One was an official U.

Area
ha
58,200
21,000
15,600
25,000
72
182
151

736
419
404
4.5
1,040
19,800
1,610

325
11,300
540
230
S. Weather
b. Three additional raingages outsi


sq mi
225.00
81 .20
60.60
96.90
0.28
0.71
0.59

2.84
1.62
1.56
0.017
4.00
76.56
6.23

1.25
43.70
2.08
0.89
B ureau
de study
No. of
recording
raingages
ioa
6b
6a
lc
1
1
1

1
od
od
1
4
8e
3

3
19
3
2
Gage.
area.




Gage density, Refer-
no/ 1000 ha
0.017
0.029
0.038
0.004
1.390
0.550
0.660

0.140
0
0
22.200
0.380
0.040
0.190

0.920
0.160
0.560
0.870
d. One
e . Two
no/sq km
0.172
0.286
0.385
0.040
13.890
5.490
6,620

1.360
0
0
222.000
3.850
0.404
1.860

9.230
1.550
5.600
8.700
non- record!
were offici
no/sq mi ence
0.04
0.07
0.10
0.01
3-57
1.41
1.69

0.35
0
0
58.82
1.00
0.10
0.48

2.30
0.43
1.45
2.25
ng raingage.
al U.S. Weather
41
42
43
40
14
14
14

45
45
45
46
47
20
48

49
50
51
52

Bureai
c.  Only one gage was used In study
Gages

-------
Hydrograph Method (Modified Chicago Hydrograph) and the Unit Hydrograph
Method.  The Routed Hydrograph Method was found to be the best method  to
obtain peak discharge rates and total volumes of runoff.  Areal rainfall
was determined using the Isohyetal Method.  Incremental rainfall was
determined between each pair of isohytes and summed.  Weighted average
rainfall was then calculated by dividing by the total area.  Then using
the design storm rainfall,  hydrographs were developed for 1, 3, and 5
year storms and used in the design of the collection system for combined
sewer overflow.  The retention basin volume was designed on the basis of
expected storm runoff from the combined sewer overflows and streams
caused by a three consecutive day rainfall that had a three year
recurrence interval.  Design volume was adjusted for expected wave
action from Lake Erie.

In another Cleveland study (*»0) rainfall data was used from a single gage
16.1 km (10 mi) from the treatment plant.  Due to the large size of the
drainage area, rainfall data could not be correlated with the combined
sewer overflow.  This was not an important aspect because this was a
pilot plant study and the primary emphasis was in the evaluation of the
treatment process.  Combined sewer overflow was pumped into two 16.^4
cu m (5,000 gal.) tanks for use in extended filtration runs.

Rainfall intensity data were plotted in the form of hyetographs and
compared with overflow (from the sewer system) hydrographs at Bucyrus,
Ohio (1A).  The time between the start of significant rainfall and the
time overflow began was determined.  Although a raingage was used in each
of the three watersheds, data was taken from the "Rainfall Frequency
Atlas of the United States" (M») to plot rainfall depth-duration-
frequency curves.  The unit hydrograph was used to determine the shape of
the overflow hydrographs.  Then the peak runoff and total runoff volume
were determined by the Rational Method and the Hydrograph Method.  From
this comparison a modified hydrograph was adopted for designing the
stormwater facilities.   Antecedent rainfall was considered and straight
line relationships were found for curves of peak overflow and overflow
volume plotted against peak rainfall intensity.

In the Roanoke, Virginia (^5) study two non-recording raingages and one
recording raingage was used.  The non-record ing raingages were found to
be inadequate for measuring area!-variations in rainfall intensity.
Rainfall tended to follow the mountain ridges in the summer months but
large rainfall variations were found.  Considering the total area
involved, the recording raingage density was 0.6k gages/1000 ha (0.17
gages/sq mi).  Hyetographs and sewer hydrographs were plotted and the
relationship between surface runoff and total rainfall were determined
for each study area.  The intensity and duration of rainfall was also
used as an input in a computer program that determined the location and
amount of overflow in the interceptor for each drainage area.  Rainfall
data from the "Local Climatological Data" published by the U.S. Weather
Bureau of the U.S. Department of Commerce were used to determine total
overflows for different total rainfall amounts.  Based on the rainfall
                                   108

-------
studies and other water quality studies alternate methods were
considered to reduce or treat sewer overflows.

The small area of Cobbs Creek drainage  in Philadelphia  (46) was studied
using one raingage.  Because of the small area a large  raingage density
existed.  The treatment system studies  required 0.51-0.76 cm/hr (0.2  to
0.3 in./hr) for one hour  in order to produce enough overflow for a test
run.  Rainfall intensity-duration was measured with the single raingage
located 91-4 m (100 yards) from the treatment plant.  The runoff coeffi-
cient was determined for  the study area.  The primary interest in this
study was the treatment system and the  rainfall data were used to
correlate overflow data after the treatment plant had been  installed.

The Boneyard Drainage Basin  in Champaign,  Illinois  (47) has been used for
a number of studies.  The data reported  in Table V~5 were developed for
the evaluation of methods for determining cumulative storm generated
runoff volume and rates (EPA Contract 68-03-0302).

In  the Des Moines,  Iowa study (20) eight  raingages were spread around an
area of  19,800 ha  (76 sq  mi).  However,  four study areas  (total study
area was 1,420 ha  [5-5 sq mi]) actually monitored did not have raingages
within their boundaries.  Hence a modified Thiessen method was used to
determine  rainfall  in the study areas.   Intermediate points called
"dummy rainfall stations" were established on a line connecting existing
raingages.  Rainfall at the dummy rainfall stations were determined by
interpolation; these data were checked  to see that  relatively uniform
rainfall occurred at each raingage station.  The total  depth of rainfall
was found  to be the best  parameter to correlate with the volume of run-
off.   Curves of rainfall  intensity versus percent clock hours  in which a
given  intensity is equaled or exceeded  (Figure V-40) were used to
determine  the amount of combined sewer  overflow that could be expected
annually from a given watershed as described below.

     "Given the area of the watershed,  the capacity of  the combined
       sewer, and coefficients of runoff  as previously described, the
       intensity of  rainfall  that can be  contained by the system can
       be calculated by the Rational Formula.  From  this, the percent
       of time annually that overflow will occur can be  predicted by
       use  of a curve such as shown  in Figure V-40.  For example, if
       it were determined  that a given combined sewer would contain
       the  runoff from a rainfall of up  to 0.8  inches per hour  intensity
       before overflowing, overflow would  be expected to occur O.I per-
       cent of the clock hours or approximately 9 hours  annually.  From
       this, an estimate of sanitary sewage overflow is  possible."
       (20)

Rainfall  intensity was also  related to  the total rainfall depth which
occurs at  or greater than a given  intensity  (Figure V-41).  As shown  in
this figure, two sources  were used  to develop this  curve.
                                    109

-------
  2.Or
  1.6
o
o
             1        23^56
     PERCENT  OF  CLOCK  HOURS WITH RAINFALL  INTENSITY EQUAL TO OR
                    GREATER THAN GIVEN  INTENSITY
           Figure  V-^0.  Distribution of  rainfall  intensity
  ,0 u          with  respect  to time --from reference 20
     0      20       kO       60      80       100
      PERCENT OF  ACCUMULATED  RAINFALL EQUAL  TO
          OR GREATER THAN  GIVEN  INTENSITY

     Fiugre  V-Al.   Distribution  of  rainfall  intensity with  respect
            to accumulated  rainfall  -  from  reference 20
                                110

-------
     "The curve shown in Figure V-41 may be used in much the same
      manner as that developed for the clock hours exceeding a given
      intensity.  For example, if a given facility has the capacity to
      handle the runoff resulting from a rainfall of 1.0 inch per hour
      intensity, approximately 22 percent of the average March to
      November rainfall would be expected to occur at intensities greater
      than this.  From this the quantity of overflow can be evaluated."
      (20)

The rainfall data taken during this study were used along with historical
data from the U.S. Weather Bureau to determine the basis of design for
storm flow treatment facilities and for evaluating the effectiveness of
alternate systems.

During eight years of studying the Milk River Drainage Basin in Detroit
(48), long term average rainfall data were found to vary by 50?.  Because
of the extreme variations in storm patterns, runoff data were analyzed
for selected storms rather than for average hydroldgical data.  Data
from these raingages were averaged using the Thlessen weighting method.
It appeared that overflows at the Milk River basin would be more a
function of total rainfall than intensity of rainfall.  Combined and
storm sewers in the Milk River Drainage Basin were sized using the
Rational Method to determine runoff.

Racine, Wisconsin (49) used three raingages for 325 ha (1.25 sq mi) so
that the  Thiessen Method could be used to determine average rainfall.
It was necessary to use the Thiessen Method to determine the a real rain-
fall for use in the U.S. EPA Stormwater Management Model (see Figure
V-42).  The model was used to correlate predicted runoff hydrographs
with measured hydrographs.  Interestingly, it was found in these
comparisons that the resulting hydrographs were very sensitive to
changes in rainfall input being changed from a change in the number of
raingages.  In San Francisco 19 gages were used to study rainfall-runoff
relationship over the 11,300 ha (43.7 sq mi) area (50)-  Twenty storms
were monitored for six basins for which the land use varied from single
family residential to industrial.  Profiles of quantity and quality of
combined sewer overflows occurred for rainfall intensities greater than
0.51 mm per hour (0.02 inches per hour).  A one year analysis of rainfall
intensity and duration was made for 1969*1970 and assumed to be about
normal for that area.  These and other data were used to evaluate
several possible methods of handling their combined sewer overflow.

A contact stabilization biological treatment system was evaluated at
Kenosha, Wisconsin (51)-  Three raingages were used for the 540 ha (2.1
sq mi) area so the Thiessen Method of determining areal rainfall could
be used.  During this five month study twenty-five rainfall events were
recorded for volumes over 0.25 cm (0.1 in.).  No direct correlation
between rainfall volume or intensity with overflow volume or flow rate
could be determined.
                                   Ill

-------
        STORM  \kt  SITE  II
            STORM  H»,  SITE  I
  35
  30

u-
"20

315
u.
  10.

   5
                 	 ONE  GAGE
                 ---  TWO  GAGES
 100    200    300   ^»00     500
  TIME AFTER START  OF  STORM,  min
   12
M  10
u-
°.  8
I  6
                A
                M
                                                   	 ONE GAGE
                                                   	 TWO GAGES
    100    200    300    400    500
    TIME AFTER START OF STORM, min
12
10

 8
 6
        STORM  16,  SITE  II
	 ONE GAGE ty\
-- TWO GAGES|  \
            I   i
            I
            I
     100   200    300    1»00
  TIME AFTER START OF STORM, min
                                      35

                                      30
                                    i/>
                                   M-
                                    °. 25
                                   o  20
                                   u.
                                      15
                                      10

                                       5
                                              STORM  16,  SITE  II
                        	 ONE GAGE
                        --  TWO GAGES
       100     200     300
   TIME AFTER START OF STORM,  min
 Figure V-^2.  Graphs from Racine,  Wisconsin illustrating the effects
   of multiple raingages on the U.S.  EPA Stormwater Management Model
                                112

-------
A study  in Milwaukee  to evaluate storage of  combined  sewer overflows
utilized three raingages for a  230  ha  (0.89  sq mi)  area  (52).   Studies
were planned to evaluate the time of day storms  occur, antecedent  rain-
fall and influence of  intensity and duration of  storms on the  average
overflow quantity and  quality.  The effects  of these  parameters on  the
variation  in overflow  quantity  and  quality were  determined.

High initial concentration or first flush conditions  were noted during
this study.  No other  correlation was  found  between combined sewer  over-
flow qualities and specific storm characteristics.  During this study,
runoff coefficients were found  ranging from  0.3  to  0.8.  BOD and sus-
pended solids were subjected to linear regression analysis for 97  storm
events.  Correlation  coefficients between 0.81 and  0.92  were obtained
for BOD  and suspended  solids data taken at the beginning of the storm,
the maximum point and  for zero  to 30 minute  average values.  No correla-
tion was found  in combined sewage quality variations  with storm charac-
teristics  and a very  low correlation coefficient was  found relating the
maximum  BOD to  the  interval between storms.

A  summary  of the  raEngage densities recommended  in  the  literature  (both
mathematical and  empirical studies) is shown in  Table V-6.  It  is signi-
ficant  that the  recommended  raingage densities necessary to describe
areal  rainfall  or runoff  increase as the  runoff  area  decreases. This
would occur  if  any  fixed number,  for example three, raingages  were used
for different  runoff  basin areas.   However,  the  number of gages required
for any  fixed area  also  tends  to  increase as shown  in Table V-6.  The
fourth  study shown  in Table  V-6 is  out of  line with the  other  recommended
raingage densities  because  it was conducted  in a part of California where
relatively uniform  rainfalls are  experienced.  In comparing these  data
with  the raingage densities  for storm  generated  discharge studies  with
drainage areas  less  than 25-9  sq  km (10 sq mi),  it  can be seen that six
out of  eight raingage networks  were essentially  the same as recommended
 in the  literature (see Figure  V-43)•   In many cases,  the raingage  net-
works  in Table  V-5  were  used  to establish  rainfall  relationships of
'existing pilot  and  full  scale  treatment plants.  Because evaluation of
the treatment system  was the primary focal point of the  investigation,
the raingage data were not as  complete or thoroughly  studied as they
would have been  if  the studies  had  been made to  design  the treatment
system.   In the  Detroit  and  Racine  projects, raingage densities were
exactly  the same  as  that recommended  in the  literature.  On the other
hand,  the  Cleveland  study at  the  southerly district was  a pilot project
to evalute a treatment system,  and  a very  low raingage density was  used.

Recommended Raingage  Density

In recommending a raingage density  or  approach for  arriving at a density
to be used in storm  generated discharge projects,  it  is  necessary  to
establish  clearly the use of  the  precipitation data.  When modeling
studies  are being conducted, extensive raingage  networks will  be nec-
essary  to  accurately  measure areal  precipitation patterns  in order to
                                    113

-------
E
cr
cr
1/1
O
Q

UJ
     2.31
    (6)
     1.15
    (3)
     0.77
    (2)
£    0.38
    (1)
                2.6
                (I)
                        RECOMMENDED  RAINGAGE  DENSITY  FROM THE LITERATURE
                                            RAINGAGE DENSITIES  RECOMMENDED
                                        ©  FROM  MATHEMATICAL AND  EMPIRICAL
                                            STUDIES
                                            RAINGAGE DENSITIES IN STORM
                                            GENERATED DISCHARGE STUDIES
                                                                                 -0
5.2
(2)
                              7,8
                              (3)
10. k
(M
12.9
(5)
15.5
(6)
18.1
(7)
20.7
(8)
23.3
(9)
25-9
do)
          Figure  V-^*3.   Comparison  of  raingage  densities  recommended  in  the  literature
                       with  those  in storm  generated  discharge  studies

-------
	Table V-6.  RECOMMENDED RAINGAGE  DENSITY  FROM  LITERATURE

Runoff Area, sq km  (sq mi)   gages/sq  km     gages/sq mi      Reference
1.
2.
3.
4.
5.
6.
7-
8.
9-
10.
25-9
(10)
111.4
(43)
64.8
(25)
25.9
(10)
25-9
(10)
10.4
(4.0)
2.5
(1.0)
2.6
(i.o)
0.65
(0.25)
0.49
(0.19)
0.054
0.081
0.096
0.039
0.193
0.386
0.772
1.157
1.54
2.045
0.14
0.21
0.25
0.1
0.5
1.0
2.0
3.0
4.0
5-3
39
35
39
34
37
39
39
37
39
39
 verify  the  model  which  ties  subcatchments  together.   Extensive  histor-
 ical  precipitation  data is essential  in  either storm  sewer  design  or
 combined  sewer  overflow treatment  plant  design.   In most  cases  histor-
 ical  data can be  obtained  from the periodical  "Local  Climatological
 Data" which is  published by  the U.S.  Weather Bureau of  the  U.S.  Depart-
 ment  of Commerce.   More detailed rainfall  intensity-duration  data  can
 be obtained from  the  "Rainfall  Frequency Atlas of  the United  States"
 Technical Paper No. 40, published  for the  Department  of Agriculture by
 the U.S.  Department of  Commerce (44). These historical data  can be very
 useful  in storm flow  design  projects  where peak flow  is determined using
 the Rational Method.   If these sources of  data are not  sufficient, it  is
 possible  in many  cases  to  obtain copies  of the original raingage charts
                                   115

-------
from the U.S.  Department of Commerce,  NOAA, Environmental Data Service,
Asheville, North Carolina 28801  (National  Climatic Center, Federal
Building).  However, the historical  data must include sufficient rainfall
intensity-duration data to develop the "i" value in the Rational^Method
equation, otherwise additional  data  may have to be taken from raingage
networks.

Rainfall data alone is not sufficient  to use the hydrograph methods
previously described.   These procedures require a known rainfal1-runoff
relationship; hence, raingage networks will be necessary if the entire
runoff hydrograph is needed.  The length of a study relating rainfall to
runoff may be too short to ensure that the correct storm frequency has
been encountered for design.  For example, a 2 year study may not measure
a 5 or 10 year storm.   Thus, it is likely that a combination of actual
rainfall data and historical data will be necessary for storm generated
discharge design projects.  When sufficient historical data is not
available, longer term rainfall  data from extensive raingage networks
will be necessary.

The need for rainfall  intensity-duration data is also related to the type
of project or to its extended use.  Table V~7 shows four main study
areas and their need for raingage networks.

      Table V-?.  RAINGAGE NETWORK NEEDS FOR DIFFERENT STUDY AREAS	

                                                               Need for
                                                               Network

     To define areal rainfall and its  variations                  yes
     Mathematical Model verification                              yes

     Sewer, Treatment Plant, or System Design                     varies

     Operation of Storm Generated Discharge Systems               no
      (Treatment or Holding Basin)

     Operation of In-line Storage or Selective                    yes
      Discharge Systems
In cases where the interest is primarily in rainfall characteristics, it
is necessary to determine spatial rainfall  variations or areal rainfall
and extensive networks will be necessary.  A minimum raingage density
for this type of study would be given by the curve  in Figure V-^3.  A
minimum of three raingage stations should be used in any network system
so the Thiessen weighting method can be used.  A digital computer can be
applied to the Thiessen system thereby reducing the inherent problem of
establishing a new weighting system every time the  rainfall stations are
changed (5).
                                   116

-------
Where mathematical models are used to predict  runoff volumes and other
characteristics, raingage networks are also needed, because of the
major contribution rainfall has to these characteristics.  When modeling
studies are being conducted in large municipal areas,  the cost of raingage
installation and maintenance can become excessive.  Models can be
verified by selecting smaller catchment areas  for detailed study.  Rain-
gages can be located  in one or several of  these smaller areas and the
storm generated discharge can be characterized to establish the model's
parameters.  After the model has been verified in these small catchment
regions, it should be checked at several points throughout the municipal
watershed before it  is used.

The need for raingage networks in the design of sewers or conveyance
systems, treatment plants or holding basins  is not as  clear cut as the
previous two types of studies.  These systems  usually  are designed on
the basis of total volume and peak runoff  rate.   If sufficiently
accurate C values can be predicted for the runoff area and historical
rainfall intensity data  is available for the design storm, then the
Rational Method can  be used.  However, if  sufficient historical data is
not available to determine the rainfal1-intensity-duration curves, rain-
gage networks will be necessary.  The curve  in Figure  V-43 provides a
suggested raingage density, but  in system  design a somewhat lower density
may be  satisfactory.  This  is especially true  in  regions of the country
where uniform rainfall patterns are encountered.

Hydrograph methods of predicting runoff may be used for system design in
areas where runoff surfaces are not expected to change.  However, in
newly developed areas or where expansion of peripheral suburban areas
are  involved, the runoff will be difficult to  measure  and rainfall-run-
off  relationships will be  impossible to predict from hydrograph tech-
niques.  In this case, values can be established  for the expected land
use of  the drainage  area and the Rational  Method used  for design.

Raingage networks are not needed for the operation of  storm generated
discharge facilities, treatment systems or holding basins (except of
course  where the  rainfall pattern  is used  to operate the sewerage system
flow control system).  A non-recording raingage at the treatment plant
will be sufficient to determine daily and  yearly average rainfall.  Where
storm generated discharge treatment plants are expected to be built  in
the  future, more extensive  raingage networks may  provide essential data
for design.
                                   117

-------
               SECTION VI  - QUALITY CHARACTERISTICS AND
                        LABORATORY PROCEDURES
Virtually any water resources problem or study which involves storm
generated discharges depends greatly upon an accurate knowledge of the
physical, chemical  and biological  characteristics of these flows.
Experience in past  years with the  analysis of wastewater, plant effluents
and river, estuary  and lake waters is very helpful in suggesting the
characteristics which most Tikely  are of primary concern.  The peculiar
properties of combined sewer overflows and strom runoff preclude the
direct usage of these commonly used characteristics without some changes
or modifications.   As an example,  while coliform counts have been and are
still  being used to assess the sanitary quality of a water body for the
purpose of a municipal water supply source, the use of this same charac-
teristic to assess  the sanitary quality of runoff from an urban area
could result in very misleading conclusions.

What is stated above applies not only to the characteristic Itself, but
also to the procedure used for Its analysis.  The experience with the
laboratory analysis of combined sewer overflows and storm runoff is much
more limited than  the case of other liquid and sludge samples in the
general  water resources and water  pollution control field.  Thus the use
of "standard" laboratory procedures as is done with most wastewaters may
not produce satisfactory results in some stormwater cases.  The procedures
or modifications of established procedures recommended in this section
take into account actual laboratory experiences in the analysis of
combined sewer overflow and storm  runoff samples.

QUALITY CHARACTERISTICS FOR STORM  GENERATED DISCHARGES

Purpose

Knowledge of the quality characteristics of storm generated discharges
is useful for a number of reasons.  First of all this information is
necessary to assess the impact of  these discharges on the aquatic
environment.   As in the case of wastewaters and treatment plant effluents,
storm generated discharges can cause an alteration in the physical,
chemical and biological quality of receiving waters.   A significant
alteration in water quality can result in a situation which is harmful
to the indigenous biota or impair  the use of the water for drinking,
                                    118

-------
recreational, industrial and other purposes.  Combined sewer overflows
and storm runoff contain constituents that do affect the criteria fre-
quently used in water quality standards.

Quality characteristics are also required for the specific design of
facilities used for the treatment of combined sewer overflows and storm
runoff.  Some design procedures require that the characteristics be ex-
pressed in terms of a concentration unit such as mg/1 or a gross loading
unit such as kilograms  (pounds) per unit time.  It  is also often helpful
to know the variation of the quality characteristics with time of rain-
fall.  Data obtained in this form are commonly referred to as "time
series".
fn addition to the design of treatment facilities, quality characteristics
are needed to evaluate the treatment process.  Naturally, influent and
effluent quality characteristics must be obtained to make this appraisal.
It is important to establish treatment unit efficiencies under various
loading or flow conditions.  These data are also essential for developing
appropriate cost information and as a guide for the selection, design, and
operation of additional or furture combined sewer overflow or storm runoff
treatment facilities.

Pertinent Quality Characteristics

If one considers all of the organic and inorganic compounds which can get
into wastewaters of various types and in storm generated flows from a
multitude of different sources,  it is apparent that the list of possible
quality characteristics which can be run on these samples ts almost end-
less.  However, experience with wastewaters through the years plus the
more recent experience with combined sewer overflows and storm runoff
have suggested a limited number of these characteristics to be particularly
useful.  It will be noted that these quality characteristics are not
necessarily specific compounds or elements but rather refer to gross
groupings or general categories.  For example, the particulate concentration
in a given combined sewer overflow sample consists of a myriad of different
organic and inorganic undissolved solid particles.  In most cases the major
concern is the total concentration of the particulates which are present in
the samples, and possibly, the overall percent which falls in the organic
and inorganic categories.  Identification of the specific organic and in-
organic compounds is normally not of interest, though the possiblility
should be kept open that in certain restricted applications it may be.

Presented below is a brief description of the quality characteristics con-
sidered to be important for storm generated discharge applications.  More
detailed descriptions plus recommendations regarding laboratory procedures
are presented in the subsequent portions of this section.  Also, at this
point, the characteristics are discussed under broad categories rather
than specific analyses.  The latter will be presented in subsequent por-
tions of the section also.
                                    119

-------
Oxygen Demand - One of the most important quality characteristics  in a
receiving body of water is the dissolved oxygen concentration.  The
dissolved oxygen concentration has a direct bearing on the quality and
natural balance of much of the aquatic biota.  Dissolved oxygen concen-
tration can also have an effect on the recreational and aesthetic uses
of a body of water.  Storm generated discharges that contain organic
and inorganic compounds that exert a demand for the oxygen dissolved in
water can be considered pollutional discharges in the same sense as dry-
weather municipal wastewaters.  The oxygen demand is exerted by 1) organic
compounds that undergo biochemical oxidation as a result of microbial
activity and 2) by the immediate demand exerted by the chemical oxidation
of inorganic reduced compounds.

Particulate Concentration - The solid matter present in storm generated
discharges can be divided into two major categories; namely, particulate
solids and dissolved solids.  Particulate solids are Important in combined
sewer overflows and storm runoff applications because they usually
represent a large fraction of the total solids.  Also, these solids are
generally removed from the flow by physical treatment processes such as
sedimentation, screening, flotation, and filtration—the type of processes
most commonly used for storm generated discharges.  It Is, in fact, the
relatively high concentration of particulate solids in these flows which
makes such processes attractive.

Pathogenic Microorganism  Potential - Any discharge which includes some
amount of domestic wastewater or waters which have come into contact with
excrement from warm blooded animals of any type should be considered as
having the potential for conveying pathogenic bacteria, viruses, protozoa
and other contagions.  It is extremely difficult  if not logistically
impossible to monitor these discharges for the many pathogens themselves.
This problem was recognized in the water supply field many years ago and
has led to the almost universal usage of the coliform group of bacteria
as the indicator or measure of the sanitary quality of water.  The coli-
forms themselves are not necessarily pathogenic but the!r,presence should
infer the possible presence of pathogens.  For a number of reasons which
will  be considered in detail further on In this section, the coliform
group is not necessarily the  most sensitive indicator as far as storm
generated discharges are concerned.

Eutrophlc Potential - In addition to sunlight and carbon dioxide, aquatic
plants like terrestlal plants require nutrients and trace salts.  The
principal nutrients are compounds which contain the elements phosphorus,
nitrogen and potassium.   The proliferation of aquatic plants in most water
bodies is undesirable.  The term "eutrophic" refers to a condition in a
water body where copious plant growth has resulted in an undesirable or
unsightly situation of accelerated lake deterioration.

Toxicants - The presence of high concentrations of toxic materials in
combined sewer overflows and storm runoff is obviously undesirable.  Toxic
materials In these discharges can have a deleterious effect on the aquatic
                                   120

-------
biota.  Some toxic materials are more harmful to higher life forms such
as crustaceans and fish while others may have a greater effect on
plankton or possibly aquatic insects.  Both categories are of concern
If one considers the entire ecological balance in a water body.  The
disruption of one portion of the life chain eventually has an effect on
the entire biological balance.

Higher life forms including man can also be adversely affected by the
presence of toxic materials in storm generated discharges.  Receiving
water used for water supplies and for recreational purposes such as
swimming, fishing and boating must be carefully monitored for the
presence of toxic materials.

A larger number of compounds of varying toxicity and concentration are
likely to be found in combined sewer overflows and storm runoff.  How-
ever, the toxicants of major concern can be divided into the general
categories of heavy metals, pesticides and herbicides.

Other Characteristics -  In addition to the quality characteristics al-
ready noted, there are a number of relatively simple analyses that are
used  frequently to characterize the physical state or condition of water
and wastewater samples.  These same may be of interest in the case of
storm generated discharges.  A number of these analyses are run routinely
on water samples.  The other characteristics most likely of concern are
pH, temperature, conductivity, color, odor, and oils and grease among
others.

Origin in Storm Generated Discharges

The fact that discharges from combined sewer overflows can be grossly
polluted is obvious because of the mixing of the storm runoff with
municipal sewage.  Discharges from separated storm sewers and direct
runoff can also be polluted.  From the time rainwater falls on an
urbanized area until it  Is ultimately discharged to a receiving body
of water it encounters and conveys pollutants from many different
sources.  Rainwater  itself has been found to become polluted as It
passes through the hydrologlc cycle.  The suspended solids and COD
loading, g/day/sq m  (Ib/acre/day) from rainwater can be higher than
sanitary sewage during actual periods of precipitation (1).  Overland
travel of the rainwater, or runoff, results In contact with rooftops,
gutters, lawns and parklands, streets, alleys, sidewalks, parking lots,
etc.  Following the overland travel the runoff may then enter ditches,
culverts, catch basins,  sewers, flood channels, or holding tanks before
discharging to a watercourse.

It  is during these travels that the various contaminants and pollutants
are gathered.  The pollutlonal effect from these matelals Is measured
by the presence of pathogenic bacteria, oxygen demanding materials,
toxic substances, nutrients, inorganic salts, particulates and coarse
solids as described previously.  Of course, many materials can be
                                   121

-------
classified under more than one of the above classes.  For example,
pieces of plastic, wood, or rubber will eventually exert an oxygen
demand.  However, when in the particulate or coarse solid form  they
are of more concern because of their undesirable effect on the  visual
quality of a water.  Table Vl-l contains a partial classification of
pollutants and contaminants found in storm runoff and combined  sewer
overflows, along with examples and their primary sources.

The fact that many pollutants in storm generated discharges originate
from the land surface of urban areas introduces much variation  in the
resulting quality characteristics.  Raw wastewaters, while also
exhibiting variation in characteristics, do have fairly predictable
diurnal and weekly patterns in their characteristics.  Storm generated
discharges, on the other hand, originate from many different types of
land surface ranging from relatively clean residential areas to grossly
littered industrial work and storage yards.   Thus a myriad of different
materials which randomly end up on the earth's surface as a result of the
usual activities of man,deliberate or accidental spillage, natural forces,
etc., will be found in storm generated discharges over extremely wide
ranges of concentration.  Superimposed on these variations is  the varia-
tion introduced by the hydrological  cycle itself.  Runoff from an
urban area following a long dry period has the potential for conveying
large quantities of pollutants "stored" in the area during that period.
And if the precipitation producing the runoff is of high intensity and
short duration, the concentration of the various constituents can be
much higher than is ever experienced in municipal wastewaters during
dry-weather.  This is even further compounded if there is a build-up of
materials in the combined  sewer during dry-weather.  It is for these
reasons that the term "typical" when applied to the quality character-
istics of storm generated  discharges can be a serious misnomer.

PAST PRACTICES RELATIVE TO STORM GENERATED DISCHARGES

Quality Characteristics

A review of past investigations in the area of storm generated discharges
can provide a valuable insight as to the quality characteristics con-
sidered to be the most useful  to researchers from different parts of the
country.   With this objective  in mind, a survey was made of thirty-eight
reported studies to establish  the frequency with which various charac-
teristics were employed (1-38).   Of the 38 studies, 31 were reported in
U.S. EPA final  reports and 7 were reported in technical  journals.
Thirty-five of the 38 studies  were published in the 1969 to 1973 period,
and of the remaining three, one was published in 1968, the other in 1966
and the earliest in 19*42.

The results of the survey  are  summarized in Tables VI-2, VI-3, \l\-k.
Table VI-2 presents the results of all 38 studies, that is, studies
which involve some treatment facilities, as well as studies which were
simply concerned with the  quality of storm generated discharges.  As
                                   122

-------
                   Table VI-].  CLASSIFICATION, EXAMPLES, AND PRIMARY SOURCES OF POLLUTION
                               FOUND  IN STORM RUNOFF AND COMBINED SEWER OVERFLOW
          Cl ass i ficat i on
                                          Examples
                                              Primary Sources
to
CO
        Nutr ients

        Organic matter
Inorganic salts

Toxic substances

   Heavy metals



   Pesticides


Pathogenic bacteri\a


Part iculates
                                Nitrogen,  phosphorus

                                Waste food,  excreta, decaying
                                   vegetation,  humus

                                Oil,  gasoline,  other hydro-
                                carbons3,  and  various organo-
                                    industrial  wastes
                                        CaCl2, NaCl
Zinc, copper, lead,  nickel,
   mercury, chromium


Chlorinated hydrocarbons,
  organo-phosphorus

Fecal, total  coliform,  and
  fecal  streptococci0

Glass, stones,  plastics, metals,
aggregates from construction,
inorganic portion of dust  and
    di rt
                                     Lawn fertilizers, excreta

                                     Litter,  animals,  lawns


                                     Motorized  vehicles,  industrial
                                     discharges and  spillage


                                     Roadway  deicing
Motorized vehicles, atmosphere
washout, metallic corrosion and
      industrial discharges

Spraying of lawns and gardens
                                                                             Excreta, garbage, soil
                                                                             Streets,  sidewalks,  buildings,
                                                                             litter and construction sites
         a.   Primary concern of these  parameters  is  the  aesthetic nuisance and danger to macroorganisms
             rather than oxygen demand.

         b.   In separate storm sewers  these  are the  sources of contamination;  in combined systems these
             are added to those already  in  the municipal wastewater.
         c.   Microorganism groups  which  serve  as  indicators of  possible pathogens.

-------
                          Table VI-2.  USAGE FREQUENCY OF VARIOUS QUALITY CHARACTERISTICS
                                IN PAST STUDIES  INVOLVING STORM GENERATED DISCHARGESa
             Quality  characteristic
                                   No. of times
                                     employed
                 Qua 1ity characteristic
                                  No.  of times
                                   employed
to
Biochemical oxygen demand  (BOD)
Suspended solids  (SS)
Volatile suspended solids  (VSS)
Total coliforms
Chemical oxygen demand  (COD)
Total sol ids
Total volatile solids
Settleable sol ids
Total kjeldahl nitrogen
Total phosphorus
Fecal col I forms
pH
Fecal streptococci
Oi 1  and grease
Ammonia nitrogen
Nitrate nitrogen
Temperature
Chlorides
Total organic carbon
27
25
22
20
19
19
18
18
17
17
11
10
 8
 6
 6
 6
 6
Turbidi ty
Organic nitrogen
Conductivi ty
Soluble phosphorus
Chlorine demand
Floating material
Total alkalini ty
Particle size distribution
Total hardness
Chlorine residual
Ortho phosphorus
Heavy metals
Phenols
Pesticides
Calcium hardness
Total  inorganic carbon
Dissolved BOD
Dissolved organic carbon
Di ssolved sol ids
5
5
k
3
3
3
3
2
2
2
2
2
2
2
1
1
1
       a.  Survey of  38  studies.

-------
                       Table VI-3.  USAGE FREQUENCY OF VARIOUS QUALITY CHARACTERISTICS IN PAST
                              STUDIES INVOLVING TREATMENT OF STORM GENERATED DISCHARGES3
             Quality characteristic
                                  No. of times
                                    employed
              Quality characteristic
                                   No. of times
                                     employed
to
C71
Biochemical oxygen demand (BOD)
Suspended solids (SS)
Volatile suspended solids(VSS)
Total coliforms
Settleable solids
Total solids
Total volatile  solids
PH
Chemical oxygen demand (COD)
Fecal coliforms
Total kjeldahl  nitrogen
Total phosphorus
Oil & grease
Fecal streptococci
Turbidi ty
Temperature
Total organic carbon
22
22
19
16
13
13
13
12
II
10
 9
 8
 7
 6
 A
 k
Ammonia nitrogen
Chlorine demand
Conductivity
Nitrate nitrogen
Particle size distribution
Floating material
Chlorides
Total hardness
Organic nitrogen
Soluble phosphorus
Calcium hardness
Total inorganic carbon
Chlorine residual
Ortho phosphorus
Total alkalinity
Dissolved BOD
Dissolved organic carbon
3
3
3
2
2
2
2
2
2
2
1
1
1
1
1
I
1
         a.   Survey of  22  studies.

-------
                  Table Vl-'i.  USAGE FREQUENCY OF VARIOUS QUALITY CHARACTERISTICS
                       IN PAST STUDIES  INVOLVING ONLY  IMPACT ON RECEIVING WATER3
      Quality characteristic
 Biochemical oxygen demand  (BOD)
 Suspended solids  (SS)
 Chemical oxygen demand  (COD)
 Total phosphorus
 Total coliforms
 Total kjeldahl nitrogen
 Volatile suspended solids  (VSS)
 Fecal co15forms
 Total sol ids
 Settleable sol ids
 Total volatile solids
 Fecal streptococci
 Ammonia nitrogen
 PH
 Nitrate nitrogen
 Chlorides
                                             No.  or times
                                               employed
                  Quality characteristic
                                  No. of times
                                    employed
to
12
12
II
10
 9
 9
 8
 7
 7
 6
 6
 5
 5
Oil & grease
Organic nitrogen
Temperature
Total organic carbon
Total alkalinity
Heavy metals
Phenols
Pesti cides
Soluble phosphorus
Floating material
Turbidi ty
Conduct ivi ty
Chlorine residual
Ortho phosphorus
Dissolved sol ids
3
3
2
2
2
2
2
2
I
1
1
a.  Survey of 16 studies.

-------
might be expected, the BOD and SS were the most commonly used analysis,
reflecting, no doubt, the common practice employed  in the case of waste-
waters and effluents.  The second group  in preference includes the total
coliform, COD, kjeldahl nitrogen, total  phosphorus, fecal coliform and a
number of the solids analyses.  Many of  the remaining analyses were pro-
bably employed in individual studies for reasons specific to the given
study.  For example, heavy metal analyses were employed on only two of
the occasions presumably because there was an  interest  in toxicant levels.
The results in Tables VI-3 and \l\-k are  much the same.  It is interesting
to note that in the case of the former,  the BOD analysis was employed
considerably more than the COD analysis  for studies involving treatment
of storm generated discharges.  In the case of studies  involving only im-
pact on receiving waters, the BOD and COD analyses were employed with
about the same frequency.

It is not being suggested that because certain quality characteristics
were most commonly employed in the past  that they necessarily become the
characteristics of choice for future work.  Most storm generated dis-
charge studies are of fairly recent origin.  In comparison, quality
characteristics have been used in the area of water resources and water
pollution control for many years.  It would stand to reason that when
interest in storm generated discharges began to increase sharply, analysts
and other workers in the field looked to these related areas for guidance
in the selection of quality characteristics.  Experience in the last half
dozen years or so has demonstrated that  characteristics which have been
suitable in the areas of water resources and water pollution control are
not necessarily so for storm generated discharges.  It  is primarily this
past experience which serves as the basis for the recommendations which
will follow in subsequent portions of this section.  It should also be
kept in mind that the recommendations presented in this section reflect
the results of experiences up to this point.   It is reasonable to expect
that future recommendations will change  as additional  experience with
quality characteristics is gained and as environmental concerns and
objectives are changed or altered.

Quality Characteristics Analysis Procedures

The usual approach for an analyst who is to conduct a given analysis on
a water, sludge, or related sample for the first time is to first examine
a procedure which has been recommended for the particular analysis.  After
examining the procedure critically he must determine if it is applicable
to the sample involved.  The physical characteristics of the sample and
interfering materials often preclude the use of the recommended procedure
as written.  The analyst must modify the procedure to fit the situation
at hand.  Modifications to recommended procedures are developed by con-
sulting recommended procedures used for other types of but related samples,
by examining the current technical literature for published modifications,
through the use of ancillary studies set up by the analyst himself in
which a "recovery" approach is employed, by direct communications with
other researchers or workers in the field, and by consulting with manu-
facturers of equipment used for laboratory analyses.
                                    127

-------
The above approach would be employed  In the case of storm  generated
discharges and residues associated with these discharges.   The  re-
commended procedure which most logically would be examined  initially
 is the one which appears in the latest edition of Standard  Methods  (39).
An excellent companion reference which should be consulted  at the early
stages is the U.S. EPA publication  entitled, Methods  for  Chemical
Analysis of Water and Wastes (40).   Many of the procedures  in  this
manual are based on the recommendations found in Standard  Methods  (39)
and ASTM Standards (41),  but modifications are made in those  instances
where experience gained in federal water quality laboratories indicates
an improvement in the accuracy and precision of a given analysis.   In
addition to the above, a number of other sources are available  for the
examination of recommended procedures.  The Manual of Sea Water Analysis
(42) may be helpful in those instances where high salinity or brackish
waters are involved.    For the analysis of residues resulting from
storm runoff and combined sewer overflow applications, it might be
helpful to consult the Methods of SoiIs Analys is (43) and  the Methods
for the Collection and Analysis of Water Samples (44).  The latter,
published by the U.S.  Geological  Survey, includes well proven analytical
procedures for relatively unpolluted waters.  In addition  to the above,
two recent publications of the U.S.  EPA (45) (46) include  useful infor-
mation on the analysis and monitoring of wastewaters, though the
information is not as detailed in the area of specific analysis as  in
the other references  cited.

Recommended laboratory procedures are included in the discussions on the
various quality characteristics which follow.  For the most part the
procedures recommended follow those  presented in Standard Methods (39)
and Methods for Chemical  Analysis of Waters and Wastes (40).In each
case modifications and cautionary comments are made based on the exper-
ience derived in recent years working with combined sewer overflows and
storm runoff samples.

RECOMMENDED OXYGEN DEMAND POTENTIAL  INDICATOR

The level  of oxygen is critical for  many types of aquatic  life.   A
small  decrease in the amount of dissolved oxygen (D.O.) in natural
waters can result in  stream deterioration and the disappearance of
higher species of fJsh.  A serious lack of dissolved oxygen can result
in fish kills, odors  and  unsightly appearance of the body of water.
An oxygen demand test is  used as  an  indirect measure of the degradation
of both chemical  and  biological matter by microorganisms and other
chemical  reactions.  Since the effect of the decomposition  is to use
oxygen and this oxygen consumption has a critical effect on the envir-
onment, the oxygen demand test will  measure the amount of oxygen consumed
during microorganism  activity and chemical  oxidation.   Although streams
do have a limited capacity to absorb the oxygen demand of discharged
materials through natural  processes, there may be an excess of demand
which results in a drop in the D.O.  level in the stream itself.   There-
                                   128

-------
fore it is critical to have a method of measuring this demand so that
the effect on the receiving stream may be evaluated.  This  is extremely
critical in warm weather or climates where the rates of the oxidation
reactions are increased and the solubility of the oxygen  in the water
is decreased.

Storm generated discharges have certain characteristics different from
municipal sewage that affect not only the D.O. level in the receiving
waters, but also the conventional tests used to measure oxygen demand.
Since combined sewer overflows have a variety of sources other than just
municipal sewage, the discharges may contain materials that cause special
problems.  During dry-weather when flow through a combined sewer system
Is low, solids settle out.  At the start of a storm, the first flush of
water through the system may have a high concentration of solids that
affects the demand characteristics of the waste.  It has been found that
the fraction of BOD in the particulate form can range from 69-87 percent
(A7) which is considerably higher than the 30-50 percent present in most
municipal wastewaters.  Also, combined sewer overflows from Industrial
areas and urban runoff may contain oils, toxic materials and chemicals
which are foreign to the natural environment and interfere with tradi-
tional oxygen demand tests.  Finally storm generated discharges contain
a large amount of natural materials such as silt, vegetation, wood and
other materials such as plastic that may not exert an immediate demand
but will eventually use the oxygen required for decomposition.  These
characteristics cause these discharges to be different from that waste
normally encountered in sanitary analyses.

When determining the effectiveness of the oxygen demand indicator it
must be considered that the material may be in the aquatic environment
for a considerable amount of time.  Solids are deposited and resuspended
over and over again and each time they may exert a reduced demand.
Therefore a test of limited time may not indicate the true demand of the
material.

Consideration of the importance of the oxygen demand test has led to the
establishment of two criteria that must be satisfied to get a reasonable
estimate of the oxygen demand.  First the method should indicate the
effect of the addition of the discharge on the oxygen content of the
receiving water.  That is, it should give an indication of the total
oxygen demand on the water.  Second, the method should have a standard
procedure and precision so that valid comparisons may be made for over-
flow samples entering widely different types of receiving waters.

Due to  Its present and historical importance, oxygen demand should be
comparable to results previously obtained.  The oxygen demand tests in
this report will be evaluated according to these criteria and the final
selection of method will  then be made.
                                   129

-------
Tests Available for Use as Potential Oxygen Demand  Indicators

BODr - This test is the most widely used in the analysis of oxgyen
demand .  BOD is defined as the amount of oxygen required by bacteria to
stabilize decomposable organic matter under aerobic conditions  (48) .
The standard procedure consists of placing a known amount of sample  In
a bottle and diluting it with specially prepared water containing appro-
priate nutrients and saturated with D.O.  The initial dissolved oxygen
content is noted and then compared to that measured after five days of
incubation in the dark at 20°C.  The amount of oxygen consumed by the
microorganisms is assumed to be directly related to the organic material
decomposed in the waste and therefore represents the oxygen demand on
the simulated environment.

The advantages of the test are the following:

     I.   The biological  conditions are somewhat similar to
         environmental  conditions.  The BOD test is the only
         biochemically oriented of the common oxygen demand
         procedures.  This is the only test that approximates
         what may happen in the natural environment.

     2.   The test has been used for many years so a large
         amount of data is available for comparative work.

     3.   Little sophisticated instrumentation is needed.   An
         incubator,  standard size bottles and an oxygen measuring
         device are  all  of the special  equipment that is required.

     A.   The test has achieved wide acceptance in the field of
         environmental  control.  Therefore almost all wastewater
         laboratories are equipped to analyze for BOD and they
         have a system of control  based on this procedure.

However,  the disadvantages must also be considered:

     1.   Toxic materials may suppress the BOD result.  Since
         storm generated discharges may contain a large amount
         of heavy metals (11)  and other materials which are
         toxic to the biochemical  processes,  the BOD may  be
         lower than  the actual oxygen demand.
2.  The test has a five day lead time; thereby maki
    immediate process control  impossible.
                                                        ng
     3.   The dilution theory of analysis is based on the
         assumption that  the metabolism is  proportional  to the
         amount of material  available to the microorganisms
         rather than the  concentration of this material
                                  130

-------
A.  Seed bacteria are required perferably acclimated to
    the waste and this is not generally available with
    storm generated discharge samples.

5.  Participate matter may exert a large percentage of the
    BOD, and the oxygen demand of the particulates that
    settle during the incubation period may not be completely
    measured (^7).   Also, variation of particle-size and
    density causes different rates of deoxidation due to
    metabolic surface to weight or volume relationships  of  the
    particles.

6.  The nitrification demand may or may not be included in
    the standard five day test period.

7.  The test may not be a good indication of what will
    actually occur in the environment.  The turbidity and
    nutrient content of the receiving stream are different
    from the artificial aquatic environment created in the
    dilution water and the affect of sunlight and natural
    biological  population are not simulated.

8.  Since the test is run only five days, the actual per-
    centage of the demand measured is unknown so the total
    oxygen demand is not determined.  However, various
    techniques have been developed that will allow the
    calculation of the ultimate BOD and the reaction rate
    constant.  A first order reaction equation is generally
    used to describe the BOD progression with time; namely,


                   y = L(l-lo"kt)


    in which, y = BOD, L = carbonaceous ultimate demand,
    k = reaction rate constant and t = time of sample
    incubation in days.  Using this equation and a series
    of BOD determinations at varying time intervals, the
    values of "L" and "k" can be estimated by a number of
    different methods which vary in complexity and degree
    of accuracy  (50)(51)(52)(53)(5*0 .

    Although it has been historically assumed that the k
    value for domestic waste is approximately 0.1 and there-
    fore that the 5 day BOD is nearly 67 percent of the
    ultimate, this has been found inaccurate for a number
    of wastes.   Therefore estimations of this type are in-
    correct and the only way to find the actual value of
    the ultimate BOD is to have a series of values deter-
    mined at varying times.  Figure Vl-l illustrates how
    different wastes could have the same BOD-, but because
    of different rates of deoxygenation, the ultimate demands
    are significantly different.
                             131

-------
Q
o
CO
                               10
                            Time, days
                                                       20
          Figure Vl-l.   Illustration of the danger  in
            estimating  the  ultimate demand without
          knowledge of  the  rate of deoxygenation  (k)
     300 -
      250 -
                                            15
                                                        20
                              10
                           Time, doys
     Figure VI-2.  illustration  of  how the BOD20
lessens the effect of different  rates  of deoxygenation  (k)

      (Figures based on those  in reference A8.)
                             132

-------
When evaluating the advantages  listed,  it seems the most  important
advantage is that the test  is well known.   It has been standardized
and large amounts of historical data are available.   It must also be
recognized that the performance of this test  is required  for many
regulatory agencies.  Another  important advantage is  that the BODr
test directly measures the  biochemical demand, and an equation can be
used to compensate for temperature changes  in the receiving water.

However, ther are certain disadvantages that are more critical with
storm generated discharges.  The presence of toxic material can cause
a reduction in the total demand without the knowledge of  the analyst,
since the seed does not have enough time in the five days to acclimate
sufficiently to exert the full  BOD.  Also the particulates in the sample
may not be completely included  in the result and the nitrogenous demand
may not be included.  The knowledge of the exact percentage of total
demand exerted is unknown and the artificial environment  created may
cause the biochemical demand measured to be different from the actual
demand under natural conditions (^8).  Therefore the criteria of a
measure of the total demand for oxygen  is not sufficiently met.  -How-
ever, the need for a standard procedure has been accomplished.

BOD?Q - The BOD2Q test is used  to better estimate the ultimate or final
oxygen demand of a waste.   The main advantage is that the importance of
estimating the correct k value  is reduced.  It can be seen in Figure VI-2
that as the BOD approaches  the  ultimate value, the variation caused by
different k values is lessened.  Therefore, the advantage of a longer
incubation time is apparent.  Other advantages include the fact that a
significant percent of the  nitrogenous  demand is included and the
effects of toxicity are less since the organisms have had sufficient
time to adapt to the environment  (^9) -  The major disadvantage is the
time needed to achieve results.

BOD  - The BOD  test can be run for times less than five days or greater
than twenty days.  For the  longer period of time, the advantages and dis-
advantages of a BOD  are the same as for BOD2Q with one exception.  The
longer the time allowed, the closer the test approaches the ultimate
value of BOD and therefore  eliminates the guesswork in determining the
final effect on the environment.

Short term tests are used to allow the BOD to be estimated over a
shorter time and possibly used  for control purposes.  This has been
found successful  in certain areas but considerable testing is usually
necessary to allow correlations to standard tests.  Another problem is
that this is not a standard procedure and good comparative data are not
available.

ACOD Test - The delta COD test  is used to give a better measure of the
true biochemically oxidizable matter without having some of the inherent
difficulties of the standard BOD test (11)  (55).  In the ACOD test a
                                    133

-------
 wastewater  is  put  in  an  Erlenmeyer  flask with  a  stirring  magnet,  and
 the  flask  is  incubated at  20°C.  By measuring  the  COD  initially and
 at various  time  increments  it  is possible  to get the decrease  in  CCD
 with  time,  and determine a  rate constant  (k).  The  COD will  decrease
 to a  certain  level and then decrease at an extremely slow rate  there-
 after.   The difference between the  initial COD and  the COD when the
 decrease becomes asymptotic is considered  to be  the biodegradeable COD
 and  analagous  to  the BOD.  Figure V|-3 below  illustrates this  concept.
                '400
                300
       COD
      Values
       mg/1
                200
                100
Biodegradable COD
(Estimate of BOD)
                           Non-biodegradable  COD
                                                                j	I
                                     6810

                                    Time  (Days)
                        12
                   Figure  VI-3.   Illustration  of the COD value
                      becoming  asymptotic  with time  when
                               held  at  20°C
This test was developed as a substitute to the- BOO to eliminate problems
with variable dilutions causing inconsistency in data.  However, it must
be remembered that since no dilution water is added in the ACOD test, the
ultimate concentration of toxic materials would be higher, further slowing
                                   134

-------
the degradation process.  Therefore the potential toxicity problem
would also be greater.  The dilution factor as discussed by  Colston
(11) may become an asset when  it  is considered that once a material
is discharged into an aquatic  environment  it  is naturally diluted.
Therefore the dilution which is apparent  in the BOD test may more
closely simulate the environmental conditions.  Finally  it must be
remembered that this test is subject to the limitations of the COD
test which can be considerable in storm generated discharges as dis-
cussed  in this section.

Other Biochemical  Demand Tests  - Other tests are also used to determine
total oxygen demand and employ  the use of various instruments.  The
Warburg test and other short term uptake tests allow an unlimited supply
of oxygen with intermittent readings taken in order to determine the
rate curve.   Also BOD tests with continuous mixing (e.g. Hach Method)
have been developed and will consistently yield higher values of demand
since fresh interface is constantly available to the microorganisms
(56).  However, all of these tests are subject to the same type
limitations as the BOD and have the additional disadvantage of not being
routinely used or standardized.

COD  (Chemical  Oxygen Demand) -  The COD test is not an attempt to repro-
duce the biological conditions  of oxidation, but instead the wastes are
chemically oxidized.  There is  no inherent relationship between the COD
and BOD tests because the mechanisms of reaction are different.

The actual test is based on the principle that most types of organic
matter are decomposed when boiled in a mixture of potassium dichromate
and sulfuric acid.  Known amounts of potassium dichromate and sulfuric
acid are refluxed with the samples for two hours.  Mercuric sulfate
is added to complex the chlorides and silver sulfate to catalyze the
hydrocarbon oxidation.  The excess dichromate is titrated with ferrous
ammonium sulfate and the amount of dichromate consumed is related to the
oxygen demand of the material.

The following list of advantages will indicate why the COD test has
received relatively wide acceptance.

     1.  The short amount of time for analysis (two to four
         hours).

     2.  Particulates have little effect.

     3.  Relatively little attention or equipment maintenance
         requi red.

     4.  Not affected by toxic wastes.

However, as  with the BOD test,   there are serious disadvantages which must
be considered.
                                   135

-------
     1.   The most critical  problem has been the chloride inter-
         ference.  The chloride ion will  react with the dichro-
         mate in the following  manner:
                       12C1~ + 14 H+ -»  2Cr3+ + 2C1
         Although the effect is stoichiometric and can be cor-
         rected, there are other mechanisms which complicate the
         system.  A silver sulfate catalyst is added to allow the
         complete oxidation of the short chain hydrocarbons and
         the chloride ion will  form a silver chloride complex that
         is not completely oxidized.   Mercuric sulfate is used to
         tie up the chlorides as mercuric chloride (57)(58) , a
         soluble complex, however the actual percentage of the
         chlorides affected is dependent on the waste and the
         apparatus.  At low COD values that may be present in
         storm generated treatment process effluents, the pre-
         sence of chloride ions (especially from salt laden ice
         and snow melt) will significantly affect the results
         (59)(60).  Also, some inorganics, such as bromides and
         iodides, which would normally not cause an oxygen demand
         are oxidized and included in the COD value.

     2.  Not all of the organics are  oxidized, notably the
         aldelydes and fatty acids.

     3.  The nitrogenous  demand of the material is not included
         in the determination.

     k.  Chemical costs are high, and toxic chemicals such as
         mercury and silver are wasted to the environment.  Also,
         operating and labor costs outweigh  the gain in less
         expensive equipment costs.

In evaluation of the COD test,  the advantage of relatively wide field
acceptance is important.  In storm generated flow work the fact that the
COD test is relatively unaffected by  toxic materials, particulates and
materials with variable rates of deoxygenation is a major advantage.
However, the problem with the chloride interference may be paramount.
Although many studies have been made, the use of mercuric sulfate is still
the recommended way to remove the chloride interference and this is not
acceptable when dealing with storm generated flows.  The nitrogenous de-
mand is not satisfied and some inorganics are included that would not
exert a demand.  Therefore the actual significance of the value determined
is not known.
                                   136

-------
TOC (Total Organic Carbon)- The TOC analysis  is a measure of the amount
of organic carbon in the sample and this theoretically can be related
to an oxygen demand.  In the TOC analysis, the sample  is  injected  into
a stream of oxygen passing through an oven at 950°C.  The carbonaceous
material  is immediately oxidized to CC>2 and 1^0.  The C0£ is measured
by an infrared analyzer or converted to methane where  it  is measured by
a flame lonization detector.  The  inorganic carbon  is either removed by
acid pretreatment or analyzed in a separate channel to allow calculation
of organic carbon by difference.  There  is an automatic recording  of all
readouts  so a continuous record is available  (61)(62).  The advent of
sophisticated instrumentation allows this analysis  to be performed in a
matter of minutes.

The advantages of the TOC are listed here.

      1.  Quick method of analysis.

      2.  Simple procedure.

      3.  Accurate  in measuring the carbon since  there are
          few interferences.

      *t.  Not affected by toxic materials.

The disadvantages are:

      1.  No direct  relationship to oxygen demand since
          only carbon  is measured.

      2.  The nitrogenous demand and  inorganic oxygen
          demand are not included.  This  is especially true
          of sulfides and nitrites that can be present in
          storm flows.

      3.  It  is assumed all organics are oxidized.

      4.  PartJculates may  interfere with the operation of the
          instrument.

      5.  Expensive  equipment is needed which may require con-
          siderably  maintenance.

Although  TOC  is an accurate method of measuring carbon with reproducible
results,  the exact  relation to oxygen demand  is unknown.  Even  though
the theoretical relationship of carbon to oxygen  is straight forward,


              C + 02 •*• C02      and      02/C - 2.66
                                    137

-------
 the  actual  calculation  is complicated  because carbon  is  available in
 many forms,  some of which include oxygen.  Therefore  the theoretical
 value  cannot be found unless the exact chemical  formula  of  the  con-
 stituents of the waste  is known.  This lack of a direct  relationship
 to oxygen is the most critical  factor  in evaluating this analysis and
 is the biggest disadvantage.  The major disadvantages  are the high
 equipment cost, the possibility of extensive maintenance, and the fact
 that any demand exerted by a material other than carbon  cannot  be
 measured.

 TOD  (Total  Oxygen Demand) - The TOD analyzer is an  instrument that has
 been developed to measure total oxygen demand directly.   The sample  is
 injected into a stream of nitrogen and a known amount  of oxygen.   As
 the  material passes over a platinum catalyst at 900°C, the oxidizable
 portions of the sample are oxidized and there is a momentary depletion
 in the  oxygen on the platinum surface.  This is replenished by  the gas
 stream and  simultaneously the reduction of oxygen in the gas is measured
 by a silver-lead fuel cell detector and recorded.  The amount of  oxygen
 needed  to oxidize the materials in the sample is used  to determine the
 Total  Oxygen Demand (63).

 The  advantages of this analysis are:

       1.  The actual total oxygen demand is measured.

       2.  Common ions, including chloride, have little or
          no effect.

       3.  All organics are oxidized over the catalyst.

       4.  Results can be obtained in minutes.

      5.  Accurate and precise measurement is  possible.

      6.  Toxic materials have no effect.

The  disadvantages are:

       1.  More materials may be oxidized than  will  actually
          exert a demand on a receiving stream.

      2.  Expensive instrumentation is needed.

      3.  Particulates may clog instrument thus  necessitating
          blend i ng.

When  the criteria for optimum oxygen demand tests are examined, the TOD
satisfies the need for complete demand measurement.   The  direct measure
of oxygen demand is a great advantage and  the  speed of analysis enhances
the  procedure.   The high cost of equipment and  of possible maintenance
                                  138

-------
may be a problem.  The fact that this  is a relatively new type of
instrument can be a disadvantage since large amounts of comparative
data are not available.  Another problem is that the TOD may over-
estimate the oxygen demand since it  is a chemical rather than a
biochemical test.

Other Chemical Oxidation Demand Tests -  In order to reduce the time
needed for traditional COD tests, a  number of techniques have been
developed.  Twenty and thirty minute tests have been tried and they
result in standard values for certain materials.  An instrument which
can establish a COD value in two minutes is termed the C02D analysis
(64).  In this system the sample is  injected into a carrier gas of
carbon dioxide in a combustion furnace where the organics are oxidized
to carbon dioxide, carbon monoxide and water.  The water is removed and
the carbon dioxide converted to carbon monoxide which is measured by an
infrared detector.  The value is directly related to the COD of the
sample and is interpreted as COD through a calibration chart.  Another
device pyrolyzes the sample and uses a gas chromatographic analysis of
the constituents (65). The  major problem with these tests is that they
must be standardized against the normal COD test which has well
documented disadvantages as discussed earlier.

Recommendat ions

When the two established criteria, 1) the measurement of a total oxygen
demand on the environment and 2) a standardized test procedure, are
applied to the tests outlined as oxygen demand indicators,  it becomes
apparent that none of the avialable  analytical tests alone can satisfy
both requirements.  Therefore,  it  is recommended that two tests, the TOD
and BODr, be used to give a satisfactory measure of the potential oxygen
demand and to allow the desired historical data comparisons.  The TOD
test was chosen as the indicator of  total potential oxygen demand and the
BODr test, run with minor modifications for storm flows, satisfied the
need for a comparative>and biochemical test.  Although it is not con-
venient to run both tests as an estimate of a single parameter, the impor-
tance of the oxygen demand indicator needs special treatment.

The rationale behind the choice of TOD  is based on the theoretical concept
of oxygen demand.  Most of the tests  out 1ined equate the carbonaceous
oxygen demand to the total oxygen demand and this is not necessarily true.
Organic nitrogen, sulfur, ammonia, sulfides, nitrites and other reduced
compounds exert an oxygen demand and they can be present in storm flows.
Therefore the COD and TOC tests which do not include these materials can-
not be correlated with a TOD in many cases.  Another consideration is
that a large amount of particulate organic matter will be suspended and
resuspended over long periods of time  in the aquatic environment, and this
may cause a considerable change in the D.O. content of a receiving body
of water.  Therefore, the test; must  include the long term demand to allow
correct determination of the effects on the aquatic area to which the
material has been discharged.
                                   139

-------
The TOD test has other advantages,  especially the lack of Interference
which is a serious problem with other tests,  notably COD.  Although the
chloride ion affects TOD results at chloride  values greater than 20,000
mg/1, compensation may be made by spiking the standards with the chloride
ion at the same concentration.  Therefore the troublesome ion has little
permanent effect.

The disadvantages  of TOD are mainly the high  cost of equipment and the
possibility of extensive maintenance.  Another disadvantage is the lack
of a distinct stoichiometric relationship between nitrites and sulfides
and the oxygen demand (66).   However, since most other tests neglect
these demands completely, at least  the TOD includes some measure of them
so that a certain  effect can be determined.

The BODr test is also recommended because the TOD test cannot satisfy the
need for a standardized, wel1-documented test for oxygen demand.  The use
of BODr is so widespread and so often required by regulatory agencies
that elimination of this test  would be difficult.  Therefore, the 8005
test, with minor variations for storm flows which would allow the measure
of BODr to be more representative,  is recommended.  These modified pro-
cedures include the use of wide mouth or broken tip pipettes, specific
dechlorination and desaturation procedures and well mixed samples.
Although the test  does have numerous disadvantages, results can be com-
pared with the immense amount  of histroical data available and this
yields valuable information.  The problem of  toxicity effects on the
BODr can be used to advantage.  Comparison of the BODr and TOD results
could yield some qualitative information about the degree of toxicity
present in a sample and the possible effect on the natural environment.

A comparative representation in summary form  of the qualitative merits
of the various available tests discussed is found in Table VI-5.  This
table includes the major parameters that are  important in determining
a potential oxygen demand indicator.  Table VI-6 presents an arbitrary
representation of certain key variables on the ideal test for oxygen
demand.   It can be seen that  the TOD and BOD^ tests had scores closest
to the ideal.  However, it  is interesting to  note that the lowest and
highest rating are different by the same amounts as the ideal rating
and the best rating.

Procedures for Oxygen Demand  Indicator Tests

The following procedures are recommended for  the determination of oxygen
demand.  For the  instrumental  method  tests,  operation of the instrument
according to the manufacturers specifications  is  recommended.

General Sample Storage - Sample storage for oxygen demand tests  is
generally refrigeration at 4°C.  Acid pretreatment may be used but  it
should be avoided in samples that are to be biochemically analyzed
(specifically the BOD tests), in samples where an emulsion  is broken
with acid addition, or  in samples where the nature of the partlculate
                                   140

-------
Table VI-5.  SUMMARY COMPARISON OF VARIOUS POTENTIAL OXYGEN DEMAND TESTS
Parameters
Organics
Inorganics
Biochemical demand
Chemical demand
Toxic materials
Interferences
Initial cost
Operating cost
Maintenance cost
Routine analysis
Nitrogenous demand
Particulate effect
Sand/grit effect
Kinetics
Time for analysis
P re treatment
Storage time
BOD5
i ndi rectly
no
yes
some
yes
toxic
medium
average
low
yes
no
partly included
no
yes
days
none
6-24 hours
BOD2Q
indi rectly
no
yes
some
yes (less than BODr)
toxic
med i urn
average
low
yes
yes
partly included
no
yes
many days
none
6-24 hours
COD
indi rectly
yes
no
yes
no
Cl
med i urn
high
some
yes
no
incl uded
no
no
hours
blending
preferred
1 week w/acid
preservation
TOC
di rectly
no
no
no
no
none
hi gh
low
high
no
no
included
yes
no
minutes
blending/
microblend-
ing
2 wks w/ac
or freezi
TOD
indirectly
some
no
yes
no
none
hi gh
low
high
no
yes
i ncl uded
yes
no
minutes
blending/
micro-
b lend ing
id 2 wks w/
ng acid or
freezing

-------
Table VI-6.  EFFECT OF IMPORTANT VARIABLES ON POTENTIAL  OXYGEN  DEMAND  TESTS
Variable
Ideal oxygen
demand test S
Biodegradeable organics
measured
Non-biodegradable organics
(cellulose, plastics, etc.)
measured
Reduced inorganics
(N0~, S0°, S etc.)
Kjeldahl nitrogen
measured


Inert inorganics measured
(Cl~, $04, NOj, Na , CS++,
etc.)
Toxic materials effect
(Hg, Pb, CN, Phenol , etc.)
Time of analysis
Historical data aval
Total points
Rating points: 1 .
2.
3.

lable

1.
1.
1.
1.
1.
1.
1.
1.
8
Measure al 1
0_ needed
Should have
no effect
Measure al 1
0. needed
Measure al 1
0- needed
Should have
no effect
Should have
no effect
Results In
minutes
Much data

3. Some-not all 2.
(approx 70?)
1.
1.
3.
1.
4.
3-
1.
17
No effect 1.
Measures all 1 .
0_ needed
May measure 2.
some 0» needed
Little or no 1 .
effect
Serious effect 3.
Results In 4.
days
Much data 4.
18
BOD20
Most-not 2.
all 02 needed
No effect 3.
Measures all 1 .
0. needed
Measures almost 4.
all 0. needed
Little or 3.
no effect
Serious effect 1 .
Results In 2.
weeks
Little data 2.
18
COD TOC
Most-not I. Measures all
all 0. needed 02 needed
Most-not all 4. Measures all
0. needed 0, needed
Measures all 4. No effect
0. needed
Measures none 4. No effect
of 0. needed
Measures some 1 . No effect
chlorides &
other ha 1 ides
No effect 1 . No effect
Results in 4. Results In
hours minutes
Some data 4. Little data
20
TOD
1 . Measures al 1
0. needed
4. Measures all
0, needed
1 . Measures al 1
0. needed
1. Measures all
0_ needed
1. No effect
1 . No effect
1. Results In
In minutes
4. Little data
14
Matches ideal test.
Slightly different than ideal.
Greatly different than ideal.
Completely opposite from ideal.

-------
matter has an effect on the analysis.  Generally, for any procedure that
is affected by the addition of acid, the sample should not be preserved
with acid.  Therefore, only refrigeration  is recommended.

Specific Procedures - TOD Analysis -

     1.  Sample pretreatment - Samples should be homogenized
         at nameplate high speed  in a Waring blender  immedi-
         ately after being received  in the laboratory in order
         to reduce the size of the particulate matter to less
         than the size of the syringe (150 microns).  A tissue
         grinder can be used if the blender  is not satisfactory.

     2.  Special considerations for storm  flow samples.

             a.  Prior to actual analysis  of the sample
                 in the instrument, the material should
                 be further homogenized using a tissue
                 blender.

             b.  It is critical that when  removing a
                 sample from the  large container, that
                 the material be well mixed  to insure
                 inclusion of a homogenous mixture of
                 all solids that exert an  oxygen demand.

     3.  Sample storage and preservation.

             a.  Samples should be analyzed  as soon as
                 possible and can be held  ^8 hours with
                 refrigeration at ^°C in most cases.

             b.  Sulfuric acid or freezing can be used to
                 preserve the sample up to two weeks  if this
                 produces no apparent physical change in the
                 sample.  The formation of free oil or the
                 precipitation or coagulation of solids is
                 to be avoided.

     b.  Analysis for oxygen demand.

             a.  Follow manufacturer's instruction for
                 analys i s.

             b.  Calibration curves are recommended for all
                 ranges of TOD used.
                                  143

-------
Specific Procedures -  BOD Analysis  -

     1.   Sample pretreatment  -  none.

     2.   Special  considerations for storm flows.

            a.   Samples  should  be well  mixed  before they are
                removed  from  the original  containers.   Some
                type-of  stirrer or  shaker is  recommended when
                the aliquot  is  being withdrawn so that a re-
                presentative  amount of  a  homogenous mixture
                of all solids is included in  the  analysis.

            b.   Large  mouth or  broken tip pipettes should be
                used to  insure  inclusion  of large part iculates.

            c.   Homogen izat ion  of the sample  is not recommended
                for routine analysis.  This process could change
                the characteristics of  the waste  and give an
                erroneous representation  of demand.

            d.   The recommended seed is untreated storm flow
                that has been settled for one hour.  The
                normal concentration is 1  ml/BOD  bottle or
                approximately 3 ml/1.

     3.   Sample storage  and preservation  - Samples should be analyzed
         within six hours if  possible or  twenty-four hours  maximum.
         They should not be preserved with acid.   Refrigeration  at
             is recommended.
         Analysis  for oxygen  demand.

            a.   This  procedure  is  found  in  Standard  Methods,
                p.  1*95,  Method  220,  (39).

            b.   The direct  pipetting  technique is  recommended
                for analysis  of storm flows.   The desired
                aliquot  should  be  directly  pipetted  into the
                BOD bottle  and  the material  diluted  to the
                specified volume with the  special  dilution
                water.

            c.   The initial DO  reading is  taken  using a
                membrane electrode with  bottle stirrer and
                meter.

                1)  The  meter should  be  standardized using
                   the  manufacturers instructions and
                   checked at  each  standardization  by
                                  144

-------
                titration against the Winkler/Azide
                method for dissolved oxygen.  Cor-
                relation should be within 0.1 mg/1.

            2)   If a probe is not avialable, dupli-
                cate bottles should be used and a
                Winkler/Azide modification titration
                run In one bottle while the other is
                incubated (Standard Methods, p. 477,
                Method 218B (39))-

        d.   A glucose - glutamic acid test should be run
            periodically to check the method.  It is
            important that this be run each time new
            nutrients are made and periodic checks are
            highly recommended.

5.   Special  attention should be given to the following.

        a.   Bottle Washing - Special care must be taken
            to insure that the bottles used in the
            analysis are free of all organic contamination.
            Chromic acid wash  is recommended followed by
            thorough rinsing.

        b.   Dilution Water - This is freshly prepared by
            adding the proper amount of nutrients to a
            high quality distilled or demineralized water,
            aerating, and allowing it to stand until it
            stabilizes.  The quality of the water used to
            prepare the dilution water should be checked
            periodically for toxic metals (e.g. Cu)  when
            distilled water Is used or for organics when
            demineralized water is used.  Water should be
            stored In dark containers.

        c.   Dechlorination - All samples should be spot
            checked for chlorine content.  The dechlorination
            procedure outlineed on page 491  of Standard
            Methods (39) should be followed.

        d.   Supersaturation Removal - The temperatures of the
            samples should be ckecked and If they are under
            20°C the material may be supersaturated with
            oxygen.  The sample should be warmed to 20°C and
            aerated to remove excess oxygen.

        e.   A water seal should be present on all bottles and
            a cover provided during incubation.  The bottles
            recommended are described in Standard Methods (39)
                              145

-------
Specific Procedures - COD An_a_1_y_sj_s -

     1.  Sample pretreatment - Sample should be homogenized  in
         a blender immediately after received in the laboratory.

     2.  Special consideration for storm flows.

             a.  Extra clean glassware is necessary at all
                 times because the percent of organic matter
                 oxidized must be the same and small conta-
                 mination can cause significant  error.
                 Therefore cleaning with chromic acid and
                 special  rinsing is recommended.  It is also
                 recommended that flasks and condensers used
                 for the COD digestion be segregated and not
                 used for other purposes involving organics.

             b.  A well  mixed sample should be taken before
                 and after homogenization to include a re-
                 presentative amount of solids in the aliquot.

     3.  Sample storage and preservation - Samples can be stored
         up to one week when preserved with 2 ml of concentrated
         ^SO^ per liter and refrigerated.  Acidification should be
         avoided, however, if free oil is present or if the sample
         is altered in some manner it is difficult to obtain a
         representative sample.

     k.  Analysis for oxygen demand.

             a.  The method as outlined in Standard Methods,
                 pp. ^95-500, Method 220, (39).

             b.  The dilute method on p.  ^98 (Section *»c)  of
                 Standard Methods is recommended for samples
                 having low COD values.

             c.  Be sure sample containers are very clean to
                 prevent  contamination.

Specific Procedures - TOC Analysis -

     1.  Sample pretreatment - Samples should be homogenized in
         a blender (in the same manner as the TOD test) and
         placed in test  tubes or small jars and  covered.

     2.  Special considerations for storm generated discharges.

             a.  Prior to analyzing the actual  sample,
                 homogenization in a tissue blender is
                 requ i red.
                                 146

-------
              b.  The sample must be kept well mixed
                  whenever an aliquot is withdrawn.

     3.   Sample storage and preservation.

              a.  An acidified sample, refrigerated,
                  can be stored up to two weeks.

              b.  Analysis should be done within forty
                  eight hours on refrigerated samples
                  with no acid pretreatment.

     k.   Analysis for oxygen demand.

              a.  For actual analysis, follow the pro-
                  cedure outlined by the manufacturer
                  of the instrument.

              b.  Calibration curves ,are necessary for
                  all ranges.

Analysis of Dissolved Portion - When it is desired to determine what
fraction of the demand is in the soluble form, the sample should be
double filtered since storm flow samples are quite often high in par-
ticulate concentrations.  First the sample should be filtered through
Whatman  kO or 5^*1 filter paper followed by a  second wash through .k$\i
Milipore filters that are prewashed with distilled water.

RECOMMENDED MEASUREMENT OF PARTICIPATE CONCENTRATION

One of the quality characteristics of greatest interest in the case of
storm flows is the particulate concentration.  For the purposes of this
discussion only, particulate matter is defined as solids which are re-
tained on a filter medium following passage of a water sample.  The frac-
tion retained depends to some extent upon the medium selected.  This
matter,  as well as other possible particulate parameters will be discussed
in a subsequent portion of this section devoted to recommended laboratory
procedure.  The solids which pass the filter  can be characterized as
dissolved and colloidal.  These solids are too small to be influenced by
the forces of gravity or buoyancy.  Most particulate solids on the other
hand are influenced by these forces, but there is a wide variation as to
the extent of this influence depending upon the size distribution and
specific gravity of the particulate fraction  involved.  These properties
of the particulate solids conveyed by combined sewer overflows and
storm runoff are of interest in that they have a significant effect on the
design and operation of treatment devices used, as well as the impact of
these  flows on receiving waters.
                                   147

-------
Mature of Participate Solids

There is a tremendous array of participate solids that can be conveyed
by storm generated discharges.  The solids can be divided  into the com-
mon broad categories of inorganic and organic.  Much of the  inorganic
particulate solids originate from storm runoff and infiltration.
Included in this category are the clays, the fine to coarse  sands, pea
gravel, weathering fragments from concrete and bituminous  pavements, walk-
ways and parking lots, metal filings and pieces, etc.  The scouring po-
tential of high intensity rainfall requires that stones, rocks»and even
larger inorganic debris be added to this particulate fraction.  The
wastewaters mixed with the storm runoff can also contribute  inorganic
particulates but this contribution is usually minor  in comparison to
storm runoff.  Zcinon i and Rutkowski (6?) for example, found  that the
total suspended solids in a strictly domestic wastewater averaged
175 mg/1 and of this total, only 19 mg/1 or 11% was  in the "fixed" or
inorganic category.

The organic particulate solids in storm flows can originate  from the dry-
weather wastewaters and storm runoff.   These solids are conveyed easily
by storm runoff because their specific gravity is generally  very close
to that of water.   Included are wood chips and fibers, pieces of leaves,
stems, grass particles and related plant parts, seeds, grain parts,
pollen, remnants of fruits and vegetables, paper and cardboard bits,
organic fragments from peat, loam and similar soils, animal  droppings,
industrial  spillage, vehicle residue plus an almost endless  list of re-
lated materials.  The extent of the contribution of organic  particulates,
from wastewater depends upon the percentage of flows from  this source in
the total  overflows, but usually it is comparatively small.  The organic
fraction in wastewaters originates from body wastes, food  scraps and
from numerous industrial  operations which might be included  in the system.
Breweries,  paper and pulp, and packinghouses are examples of industries
that could  contribute a sizable fraction of organic particulates.

The fact that it is the organic particulates which exert an  oxygen demand
is another reason why there is interest in identifying this  fraction.
The removal of organic particulates from storm generated discharges re-
sults in the removal of oxygen demanding constituents as well.  The work
of Hansen,  Gupta and Agnew (17)  in combined sewer overflow studies at two
Wisconsin cities demonstrated that the contact stabilization process and
a physical-chemical process were effective in removing BOD because in
both cases  approximately 70% of the organic pollutants were  of a parti-
culate nature.

Variation in Parttculate Concentration

One of the  important features of particulate concentrations  associated
with storm flows is the tremendous variation in possible values, even at
one outfall site.   There are a number of reasons for this, among the most
important of which are:
                                   148

-------
     1.  Intensity and duration of the rainfall.  High  intensity
         rains result in higher rates of overland flow with an
         associated increase in scour.   If the velocity is suf-
         ficiently high in the conduit,  the scoured particulates
         will be conveyed to an -outlet point.  If not, the par-
         ticulates settle to the bottom  of the conduit.  A num-
         ber of studies with storm flows have shown rather con-
         clusively that the particulate  concentration conveyed
         is directly proportional to rate of discharge, that  is,
         peak concentrations occurred during the peak discharge
         period (7)(9)(67) (68).  This relationship may be partly
         the result of conduit scour.  The velocity of flow in a
         circular conduit increases with each incremental  increase
         in discharge; the former increasing at a faster rate.
         These high velocities cause the scour of heavy particu-
         lates.  Only a few of these particulates in a sample can
         influence the resulting concentration value greatly.

     2.  Duration between rainfall  events.  Each day a certain
         amount of dirt, dust, particulates, litter and debris
         of all types accumulates on the land surface of urban
         areas.  Rainfall and thaws  "wash" this material into
         separate storm and combined sewers.  Naturally, the
         longer the time between runoff events, the greater
         will be the amount of particulates removed from the
         surface.   The higher "first flush" particulate con-
         centrations reported by some investigators (7) (17)(69)
         is no doubt the result of  the initial scour of the
         lighter fraction of this accumulated debris as well
         as the same material deposited  in combined sewers
         during dry-weather periods.

     3-  Characteristics of the drainage area.  Land runoff from
         industrial yards, material  storage areas, produce
         storage and industrial  centers, railroad yards, truck
         depots, erosion from construction sites, etc.  would be
         expected  to convey more particulate debris than from
         well kept residential  areas and parks.   Grass areas
         and other vegetated regions tend to retard flow,  increase
         infiltration and consequently reduce scour.  Central  city
         areas which include mostly paved areas and roof tops  can
         also be the origin of much particulate matter.  General
         community cleanliness including municipal  street  sweeping
         practices has a significant effect on the amount  of par-
         ticulates which will eventually end up in a storm
         generated discharge.

For the above reasons, it is difficult to predict the  concentrations of
particulate matter which one can expect  in combined sewer  overflows and
                                  149

-------
storm runoff.  Peak concentrations in the 10,000 to 20,000 mg/1  range
are not unusual.  After several hours of rainfall, concentrations
of  100 to 200 mg/1 are commonly reported.  The work of Burm £t^  a_l_ (8)
is particularly interesting in this regard since they compared storm
runoff concentrations with combined sewer overflow concentrations  in
the Detroit area for a number of constituents.  In the case of particulate
concentrations they reported the following:
                                                      Comb ined
                                 Storm                 Sewer
                                 System               System


    Suspended Solids (mg/1)

       Maximum                   11,900                804
       Annual Mean                2,080                27^

    Settleable Sol ids (mg/1)

       Maximum                   11 ,100                656
       Annual Mean                1,590                238
Weibel et_ aj_ (69)  found that particulate concentrations from a separate
storm sewer area in Cincinnati  exhibited the greatest variation of all
parameters evaluated, as well  as the only parameter which exceeded the
loading from a separate sanitary sewer area.  They obtained the following
summary results:


                                 Range                 Mean

    Turbidity, (JTU)            30-1000                 170

    Suspended Sol ids  (mg/1)      5-1200                 210
The values reported by Weston, Germain and Fiore (70) for the Washington,
D.C.  area are interesting when the particulate concentration for domestic
raw wastewater,  combined sewer overflows and storm runoff are compared.
The fact that the BOD of the latter two flows is very low compared to the
suspended solids indicates that a large fraction of the particulates are
inorganic, or if organic, resistant to microbial degradation.
                                  Raw       Combined Sewer*   Storm
                                Domestic       Overflow       Runoff
     Biochemical  Oxygen
        Demand (mg/1)              250              71            19

     Suspended Solids  (mg/1)       220             622          1697

        -Average of 18 storms


                                   150

-------
Finally, the results obtained at the Hawley Road combined sewer overflow
in Milwaukee (17) are interesting in that the particulate concentrations
obtained during "firstflushes"  are  compared  to  those obtained  during
"extended overflows".  In the case of the former, a mean suspended solids
value of 522 ± 150 mg/1   (95% confidence level) was obtained for 12 over-
flow events, whereas in the case of the latter the value was 166 + 26
mg/1 (95% confidence level) for the case of kk overflow events.  It is
noted the first flush concentrations are more inherently variable than
the extended overflow concentrations which is to be extpected.

Possible Particulate Analytical Procedures

There are a number of different analytical procedures which can be used
to quantitate the concentration of particulate matter in combined sewer
overflows and storm runoff.  No one procedure satisfies all the require-
ments desired for the determination in question.  One procedure favors
the lighter fraction of the particulates whereas another favors the
heavier fraction.  The procedure that is most desirable for evaluating
a particular treatment process  is not necessarily the best for assessing
the impact of a particular discharge on a receiving water body.  The
advantages and disadvantages of five possible procedures follow.

Total Residue - One possible measure of particulate matter is to determine
the total residue in a sample following evaporation of a known aliquot.
The advantages of this procedure are  that it is a measure of the total
residue present  in the sample, both undissolved or particulate and
dissolved, and that the  laboratory procedure  is basically a very simple
one.  The disadvantages are that one large particulate particle such as
a pebble has a disproportionately large effect on the final results, and
that generally, the dissolved fraction is of minor interest in most
combined sewer overflow and storm runoff applications.  Treatment devices
and systems which have been applied to these discharges have for the most
part been directed toward the removal  of as large a fraction of the par-
ticulates as possible.   Including the dissolved fraction in the analysis
procedure tends to reduce the sensitivity of  this parameter for such
applications.  It is true that the dissolved  fraction of residue in storm
flows is generally low in comparison with the undissolved fraction, but at
times it can be quite high.  Examples of the  latter are during the early
thaw period in northern communities when deicing chemical concentrations
are high, or in the case of the effluent of treatment devices which are
effective in removing the particulate fraction.

It  is common practice when a total residue analysis  is made to determine
the fractions of the total which are volatile and fixed.  The total
volatile residue in the sample  is of particular interest since it con-
sists primarily of organic matter.  In fact some investigators have
used the total  volatile residue determination as an  indirect measure of
the amount of oxygen demanding matter in a sample, in lieu of or in
addition to the BOD or the COD.  The use of the total volatile residue
                                   151

-------
determination as an oxygen demand parameter is reasonable as long as
the amount of inorganic salt decomposition during the high temeprature
analysis step is minimal.   This is presumed to be the case for the type
of total residue found in  most storm flow samples.

Settleable Residue - This  analytical  procedure is particularly useful
for those applications where gravity separation is used to remove
particulates from storm generated discharges.   It also provides a more
sensitive parameter to estimate the extent of potential bottom deposits
below an outflow point.  It is a cumbersome test to run in comparison
to other particulate analytical procedures, since a large sample must
be settled in the laboratory for a period of one hour or more.   The
test is not a particularly sensitive one for those situations where
physical treatment methods such as straining,  filtration or flotation
are employed.  The settleable solids concentration to these treatment
units will naturally be quite high, particularly during the "first
flush" period.   The effluent concentration, on the other hand,  would
be close to zero, and yet  these processes are designed to remove solids
in the size distribution smaller than what is  included in the settleable
fraction.  Again, the use  of this analytical  procedure is meaningful
where the treatment device involves gravity separation.

Flotable Residues - Particulates with a specific gravity less than that
of water represent a special class of particulates in storm generated
discharges.  All such flows include some particulates within this cate-
gory.  These generally include the organic substances such as grease
globules, twigs,  leaf and grain parts, fibers and related materials.
The amount of these flotables depends greatly on the characteristics of
the drainage area and the  rainfall pattern.  In terms of oxygen demand
it it possible  that this fraction might be more significant in  some areas
than the settleable fraction.   Usually from a  weight standpoint, the
settleable fraction is still considerably greater, but from the stand-
point of impact  on the environment, it can be less important than the
lighter flotable fraction.

The main objections to the use of the flotable particulate fraction
are the difficulty in getting a representative sample and in conducting
the analysis itself.   A two step operation must be used; first  a
quiescent storage step to  allow the flotables  to accumulate on  the
liquid surface  followed by a separation and weighing step.

Turbidity - The  principal  advantage of using turbidity as a measure of
particulate concentration  is that this parameter is amenable to contin-
uous automatic  monitoring.  Stegmaier (9) showed that the turbidity value
of storm runoff  correlated well with fixed residue and total residue
values in a number of time series studies conducted in the City of
Baltimore.  In  spite of this, the turbidity measurement is not  a sensi-
tive parameter  for evaluating the efficiencies of treatment processes
which would normally be involved.  For this reason this parameter has
never gained favor in the  wastewater treatment field as well.  The
                                  152

-------
participate fraction associated with turbidity  Is a very small percen-
tage of the total participate fraction of most  storm flows, and thus
this parameter Is of little Interest In most situations where  Impact on
receiving water  Is a major concern.  It Is conceivable, however, that a
situation might arise where It Is necessary to  treat storm generated
discharges to a point where low turbid Ittes are  desirable, as for example,
discharges Into certain recreational areas or fish stocking ponds.

Suspended Residue - The suspended or nonf11terable residue has been one
of the most common parameters used  for determining the fraction of par-
tlculate matter  In wastewaters.  As discussed earlier, It has also been
used extensively for storm flow applications.   The partlculates which
are included in this fraction depend upon the filter selected for use in
the analysis.  It is correct to assume, however, that regardless of the
filtering medium, all the settleable and  flotable  solids are included.
In addition a large percentage of finely suspended matter that would
take many hours to settle by gravity is also retained on the filter.  On
the other hand, all the dissolved solids as well as the very finely
suspended and colloidal solids which cause turbidity in water do not
contribute to the suspended residue fraction.

The principal advantage of using the suspended  residue as the particulate
parameter Is that it does include the total fraction of the residue that
isw'lthln this classification.  Particulate residue and nonf 11 terable
residue characterize the same fraction of suspended residue.  This same
desirable feature can become a definite disadvantage In certain applica-
tions, particularly where specific  treatment processes are involved.

As  In the case of the total residue, it Is common practice to determine
the volatile fraction of the total  suspended residue whenever the analysis
is made.  The assumption Is made that this fraction consists primarily of
organic matter and  is the fraction which will exert a demand for dissolved
oxygen when discharged  Into the aquatic environment.  Also It is the
volatile suspended  residue which contributes to the benthic layer or
botlom muds of a receiving water body, or which makes up the organic
fraction of sludges produced in the treatment of storm flows.

Recommended Particulate Concentration Analyses

With full consideration to the preceding discussion plus the experiences
of the authors of this  report and others who have worked with storm
generated discharges, It Is recommended that the suspended residue be
employed as the measure of particulate concentration.  In most appli-
cations the removal of  the entire suspended residue fraction from com-
bined sewer overflows and storm runoff would be considered an  Ideal
achievement or goal,  Anything less than  this total would be a partial
success which can be effectively evaluated by making the appropriate
suspended residue analyses.  The analysis Itself Is considered routine
and not as time  consuming and cumbersome  as some of the other particulate
                                    153

-------
analyses.  Only a small sample aliquot is needed in most cases.  With
a few additional steps it is possible to determine both the fixed and
volatile portion of the suspended residue, an additional useful piece
of information  in many instances.

In addition to the suspended residue concentration, it  is recommended
that the settleable residue be determined for those applications where
gravity or inertial separation is involved, or where an assessment
must be made of bottom deposit buildup directly below or surrounding
an outfall.  For these applications  the use of the suspended residue
concentration alone could result in  a distorted evaluation of the sit-
uation, primarily because of the heavy inert fraction of residues
generally associated with storm generated discharges.  Also, this
determination may point out the need for grit removal facilities if
treatment is being contemplated.

Conducting total residue, flotable  residue and turbidity analyses on
combined sewer overflows and storm runoff generally do not add much
additional information, and for this reason, the routine use of these
analyses is not recommended.

Recommended Laboratory Procedures

For the most part the procedures recommended in Standard Methods (39)
and the 1971 EPA manual CtO) can be  successfully employed for the
determination of suspended residue and settleable solids of combined
sewer overflow and storm runoff samples.  This statement applies par-
ticularly to such recommendations as drying and volatization times and
temperatures, type of equipment, weighing procedure, and filter medium
handling and preparation.  However,  the peculiar characteristics of the
samples involved do suggest a number of modifications and additions
which are disucssed below.  Preservation method for solids determination
is limited to refrigeration until the samples can be analyzed.  Therefore
as little delay as possible is best  to avoid dissolution of certain
materi als.

Suspended Residue Determination - The value of the suspended residue of
any sample depends primarily on the  filter medium selected for the
analysis.  One of the first media used consisted of an asbestos mat
formed  at the bottom of a "Gooch" curcible.  Because of the variability
in the  mat characteristics and time  required in its preparation, this
medium  has limited use today.   Other media which have been employed
for this analysis are cellulose acetate membranes with a pore opening
of 0.45 microns, glass fiber filters with an approximate pore opening
of 0.8  to 1.0 microns, and conventional filter paper.  The first two
have been used rather extensively for wastewaters, combined sewer over-
flows,  storm runoff and related water samples.  Filter paper, for a
number  of reasons, has had limited use and will no longer be considered
in this discussion.
                                  154

-------
Engelbrecht and McKfnney (71) deomonstrated the superiority of the
cellulose acetate filter over the gooch cruci.ble and aluminum dish
methods for determining the suspended residue of activated sludge
mixed liquor.  They also presented similar evidence for the case of
raw wastewater suspended residue analyses.  They felt that cost of
the filter was the principal deterrent to its routine use.  The
authors have had extensive experience with the use of these filters
in the analysis of samples from storm flows.  There is no question
of the cost disadvantage; the cellulose acetate filters cost almost
five times as much as the glass fiber filters.  Other disadvantages
are that the cellulose acetate filters are more difficult to handle,
having tendency to adhere to surfaces or to wrinkle easily.  More
handling is required for volatile residue determinations since the
weighed filters must be transferred to tared porcelain crucibles
following the total residue determination.  Also, because of the
small pore size the filters tend to plug up quickly thus limiting
the amount of sample which can be filtered and decreasing the
accuracy of the determination because of the small buildup of resi-
due on the filter.  Principal advantages of the cellulose acetate
filter are that prewashing with distilled water is not required and
the filter is completely destroyed in the volatile solids determination.

The glass fiber filter is highly recommended for the analysis in ques-
tion.  V/yckoff (72) demonstrated the superiority of this medium over
the cellulose acetate filter for the total and volatile residue analysis
of raw wastewater, activated sludge effluent and mixed liquor.  The
author  as well as others (8)(33)(35) has used glass fiber filters on
storm flow samples with much success.  In addition to being economical,
the filters are easy to handle, can accumulate a sizable residue load
before plugging and can be fired to 550°C without being altered.  One
disadvantage is that the filter must be pre-washed to remove fines
prior to use.  The glass fiber filters are available in a number of
sizes depending on the filtering assembly employed.  The 4.7 cm size
is recommended because it provides a sufficient area to obtain a good
buildup of residue, and because  it is the size used on the filtering
apparatus for cellulose acetate filters, which is currently readily
available in most water and wastewater analysis laboratories.  The
2.2 cm size can be placed in the same gooch crucibles which formerly
were used for the asbestos mats.  Though the smaller filters have
been successfully employed in the past, the larger ones are recommended
for the reasons cited above.

The characteristics of storm flow samples require that a few special
handling and preparation steps be employed.  It is not uncommon to find
large floating debris such as leaf parts, twigs, stems, cigarette filters,
pieces of paper, etc. on the surface of the sample to be analyzed.
Generally the laboratory analyst must exercise judgment as to whether
or not he removes the particles from the sample.  If it appears that the
particle is of large enough size to significantly influence the total
weight of residue removed on the filtering medium, it should normally be
                                    155

-------
 removed.   Or  if the particle  is of such a  size  that  it  causes  a difference
 in determinations of dupl icate samples amounting  to approximately 20% or
 more,  it  should be removed.  With experience, most analysts  can become
 fairly consistent in exercising this judgment.   If a  sufficient amount of
 this debris makes it cumbersome to remove  the individual  pieces by  hand,
 it is  recommended that the sample be passed through a No.  6  sieve (3-36 mm
 opening)  prior to analysis.  The analyst should  indicate  that  this  operation
 was applied to the sample  in any final recording of the  laboratory  results.

 The likelihood of an appreciable amount of sand and grit  in  the samples  in
 question  precludes the use of pipettes in  transferring an  aliquot to the
 filter assembly.  It is  recommended that 200 to 300 ml of  sample first be
 transferred to a 500 ml  beaker while at the same time the  sample is
 shaken vigorously.  The  sample should then be transferred  to a  graduated
 cylinder  while it is being stirred with a glass rod.  The  total  aliquot
 measured  in the cylinder should be passed through the glass  fiber filter.
 As a general  rule, a maximum of 10 minutes filtering time  under vacuum
 should be employed.   Values considerably below this amount indicate   that
 more sample should be passed through the filter to obtain  an accurate
 result.   Greater values  indicate that the residue concentration  (or
 characteristics)  was such that a smaller aliquot could have been  used.

 Generally with some experience analyzing combined sewer overflow and
 storm runoff samples,  the analyst will  gain proficiency in selecting  the
 most suitable aliquot.

 Vigorous  shaking and stirring as described above is all  that can  be  done
 with samples containing oil and grease.   Any quantity of these  materials
 which remains with the residue on the filter following drying becomes
 part of the total  suspended residue value.

 It is considered  good  practice to run duplicate analyses,  particularly
 considering the properties of the samples in question.  Variation of
 duplicate analyses within 10% or less would be considered an acceptable
 precision for samples  of this type.

While it  is not recommended that the volatile fraction of  the total
 suspended residue  be determined in  each  case,  as stated  previously,  this
added analysis can provide some additional  useful information in  certain
situations.  When  running the volatile fraction  it is important  that a
 large enough sample  be  passed through the glass  filter to  insure  that a
measurable amount  of weight loss is  obtained following ignition at 550°C
for 20 minutes.   Temperature control  is  critical since higher temperatures
can result in the  partial decomposition  of inorganic salts as well.
Because the volatile fraction of some samples  can be quite small, special
precautions should  be  taken that the usual  gravimetric analysis  steps
such  as  desiccation times, filter  handling and  transport, weighing
technique, etc.  are  carefully done.
                                    156

-------
Settleable Residue Determination - The most common method of determining
settleable residue in wastewater samples  in past years was by the  Imhoff
cone method.  The experience of this author with this method applied to
combined sewer overflow and storm runoff  samples has not been too  favor-
able.  Because of the nature of the solids, bridging generally occurs  in
the bottom portion of the cone; particularly when a twig or similar par-
ticle becomes wedged horizontally In the  settling matter.  Also numerous
particles tend to adhere to the side walls of the cone.

The "weight" method described  in Standard Methods (39) is felt to  be
superior  in spite of the fact  that  it  is  more time consuming than  the
cone method.  Basically the procedure  involves making a total suspended
residue determination on a well mixed  sample as described previously,
allowing a sufficiently large  sample to stand quiescent for the period
of time desired, and siphoning an aliquot for a second suspended residue
determination from the center  of the quiescent sample.  The concentration
of settleable matter is the difference between the two determinations.
An excellent technique for obtaining the  quiescent sample is to use a
200 ml broken-tip volumetric pipette with the mouth end immersed in the
sample and the tip end connected by rubber tubing to a water trap  and
vacuum source.  The rate of sample withdrawal is conveniently controlled
by pinching the rubber tubing.

Other Sol ids Analyses - Though the  routine use of total residue, flotable
residue and turbidity analyses as measures of particulate concentration
in storm  flow samples has not  been  recommended, there may be instances
where it  is useful to conduct  one or all  of these tests.  In general the
same precautions noted in the  case of  the suspended residue determination,
particulatly with regard to sample handling and obtaining a truly  repre-
sentative sample aliquot, apply to  the above anlayses as well.  The
Standard  Methods  (39) procedure for conducting total residue and total
volatile  residue appears to be suitable for most storm generated flow
samples.  Samples that include the  runoff from industrial yards and truck
loading areas may include grease and other petroleum products that would
begin to  volatilize at the  103-105°C temperature range used for the total
residue determination.  Care should be exercised in the interpretation
of the  results of such analyses.

There is  no "standard" procedure for determination of flotable  residue.
In those  situations where this parameter  may be useful, the following
procedure  is  recommended.   First allow one  liter of the sample  to  reamin
quiescent for one hour.  Decant the upper 800 ml quickly  into a second
beaker, discard the lower portion,  and allow the decanted portion  to
remain quiescent a second hour.  With  the inlet at the bottom of the
container,  carefully siphon from this  container until only 70 to 80 mg/1
of sample remains.  Pour the remainder into a tared evaporating dish,
rinse the container with a  small amount of distilled water, and pour  into
the  same  evaporating dish.  Determine  the total residue by the  procedure
described above.  During the final  stages of the siphoning procedure,  set
aside an  aliquot of the siphoned  liquid and determine the total  residue.
                                    157

-------
Subtract the total residue found in the siphoned liquid from the value
determined in the last 70-80 ml to obtain an estimate of the total
flotable residue.

The main problem with the turbidity analysis is obtaining an aliquot
of sample which does not include the flotable and settleable solids.
These solids do not add significantly to the turbidity of the sample,
but can cause annoying problems in any type of turbidity measurements.
It is recommended that the turbidity measurements be conducted on the
internatant aliquot siphoned from a quiescent sample as described under
the settleable residue determination.  The use of one of the proprietary
devices for turbidity measurement would be much more practical  than the
use of the Jackson candle turbidimeter for analysis of storm flow samples.

Although no analysis other than the suspended solids analysis has been
recommended for standard use (settleable solids may be needed in some
cases), Section V - Monitoring Instrumentation contains a discussion of
some of the available in-situ monitoring equipment.  This  includes
instruments for both the measurement of suspended solids and turbidity.

RECOMMENDED CHOICE OF PATHOGENIC INDICATOR

A quality parameter of particular concern in storm flows is the quantity
and types of microorganisms present, since this parameter is related to
the health and safety of humans and other animals.   Complete water
resources management near urban areas requires knowledge of the quantity
and type of disease causing microorganisms which enter the aquatic
environment from combined sewer overflows and storm runoff sources.  It
has been demonstrated historically from experiences gained in water supply
and water pollution control applications, that the use of an indicator
organism or organisms is the only practical  way of making an overall
health assessment.  It does not necessarily follow that the same indicator
system used in the past is directly applicable to the combined  sewer over-
flow and storm runoff situation.   Thus, it is the purpose of this portion
of the report to examine the literature critically in this regard, and on
the basis of this information plus experience gained from field surveys
and studies,  suggest an indicator system for pathogens that would serve
best for the applications noted above.

It should be also noted that there are other applications for use of a
suitable indicator system in addition to assessing receiving water quality.
The possible use of treatment processes and storage for the control of
combined sewer overflows and storm runoff will  require the use of a
number of parameters for evaluation of treatment and storage effectiveness.
Degree of microbial  destruction will no doubt be employed as one of the
parameters.   However, it does not necessarily follow that the same  indi-
cator system should be used for both receiving water and treatment
applicat ions.
                                    158

-------
Pathogens Associated with Storm Generated Discharges

The most important pathogens associated with the aquatic environment
originate in the feces of humans and other animals.  Because a combined
sewer overflow generally includes both domestic wastewater and storm
runoff, the likelihood that some pathogens are present  in this discharge
is high.  The same is not as apparent  in the case of storm runoff.

However, the runoff from an urban area can contain fecal matter from
dogs, cats, rodents, and other animals and in some instances even humans,
and thus this flow can also contain some disease-causing microorganisms.
Also,  illegal sanitary sewer connections to storm sewers can be a source
of pathogens.  Based on current knowledge, the pathogens of greateast
concern  in these flows are the bacteria, and to a lesser extent, the
viruses and a few other types of microorganisms.

Bacteria - Bacteria are single celled  microorganisms in the 1 to 10
micron  size which can be present in waters in either the vegetative or
spore  state.  The most common bacterial pathogens associated with fecal
discharges are:  Salmonella sp, the causative agents of typhoid and
paratyphoid fever, and gastroenteritis; Shigella, the causative agent
of bacillary dysentery;  Vibrio the causative agent of cholera; Leptospi ra,
the causative agent of leptospirosis;  enteropathogenic  Escherichia, the
causative agent of  intestinal disorders; and Streptococci sp which can
cause  a number of diseases  in both man and animals.  To a lesser extent
Mycobacterium and Francisella, the causative agents of  tuberculosis and
tularemia respectively, and Staphylococcus  aureus  and Pseudomonas
aeruginosa the pathogens involved  in most ot the eye, ear nose and skin-
type  infections,  have also been isolated from the fecal discharges of
warm  blooded animals.

Viruses  - Viruses are ultramicroscopic obiigate intracellular parasites
processing either DMA or RNA.  They are considerably smaller than the
bacteria and thus not visible in a conventional light microscope.  The
viruses  generally associated with  fecal discharges are  the causative
agents  of  infectious hepatitus, poliomyelitis, and a number of other
respiratary and gastrointestinal diseases.  Enteric viruses, or the
so-called "Enteravi ruses1',  which  are  a subgroup of the major group,
"Picornaviruses",  include  the Polioviruses,coxsackle viruses, ECHO
viruses, the virus(es) of  infectious hepatitis, and adenoviruses, and
the  reoviruses  (73)-

Others - A number of other  types of organisms present  in feces can
cuase  diseases.  E. histolytlca, the causative agent of amoebic
dysentery, is a protozoan,  the group containing the smallest form
of animal  life.  The hookworm diseases are caused by a  very thin worm
about  1.3 cm  (0.5  in) long which resides  in the intestines of  infected
persons.  The eggs  produced by the femaleworm are discharged  in the
feces  of humans and other  animals  and  are  the means of  producing sub-
sequent  infections  in healthy hosts.   In addition to the two organisms
above,  a few other  minor parasites which are found  in  feces can also
cause  diseases.
                                    159

-------
Water-Borne Diseases

The primary reason for interest in the presence of pathogens  in combined
sewer overflows and storm runoff is the fact  that  these pathogens can
enter private water supplies or open finished water reservoirs used for
potable water, or surface water bodies used for recreation by people and
water sources by animals and thus result in the transmission of a disease.
Though most water-borne diseases have been brought under control  in this
country, the threat always remains if pathogens are present and safeguards
break down.  An excellent review of water-borne diseases which includes
an extensive literature search was prepared by Geldreich (7*0-  According
to a study conducted in England, cholera and typhoid are diseases most
frequently associated with water supplies (75)-  A summary of the water-
borne disease outbreaks in the United States for the 19^6-1960 period is
presented below (76).
                  Disease                       Outbreaks
              Gastroenteritis                      126
              Typhoid                               39
              Infectious hepatitis                  23
              Diarrhea                              16
              Shigellosis                           11
              Salmonellosis                          *»
              Amebiasis                              2
              Other                               	7_

                                      Total	  228


All the above diseases can be contracted by people engaged in swimming
and other water contact sports in the event that some water containing
pathogens is swallowed while partaking in these activities.  The likeli-
hood of such an event occurring is rather remote, at least in the case
of the serious water-borne diseases.   However, such waters can serve as
the media for the transmission of eye, ear, nose, throat and skin
infections as a result of pathogens which originate from combined sewer
discharges and storm runoff.

Possible Pathogenic Indicators

It is apparent that in applications dealing with combined sewer over-
flows and storm runoff, there is need for an analytical procedure to
establish whether pathogens are present or absent.  Because of the nature
and origin of these discharges, the analytical procedure employed Is not
necessarily the one that works best for water supplies, raw wastewaters
and receiving waters.
                                  160

-------
Two types of procedures are possible when it comes to establishing the
presence of absence of pathogens in water.  The first would  involve
techniques for the detection of the pathogens themselves or  some by-
products traceable to specific pathogens.  The second would  involve the
use of an indicator system in which the presence of pathogens is inferred
by virtue of the presence of related microorganisms or by their by-pro-
ducts.  The first type provides direct proof of pathogen presence, the
second type provides only indirect proof of the same situation.  It would
appear that the first type is superior for most applications, however, as
will be discussed below, there are many factors which preclude the use of
this approach for routine work.

Specific Pathogens Themselves

Ideally the most satisfactory  indicator of the presence of pathogenic
organisms in waters is to find the specific pathogens themselves.  In
practice, however, this could never be a suitable approach because of
the variety of pathogens involved and the fact that the laboratory pro-
cedures are complex, time consuming, and for the most part rather insen-
sitive (77)(78).  Also, because of the comparatively low numbers of patho-
gens, very large water samples are needed for the detection procedure.
Except possibly for some recent work (79) which demonstrated that a
quantitative procedure is possible for the Isolation of Salmonella sp.
the prospect of using specific pathogens in combined sewer overflows and
stormwater runoff as indicators of unsanitary conditions Is  remote at the
present time, and probably for some years into the future.  Thus reliance
must be placed on some other group of microorganisms.

Total Coliforms - Since the latter part of the nineteenth century when the
coliform group was first isolated from human feces by Escherich, this
group of microorganisms has been the most important indicator of unsani-
tary or possible disease producing conditions in waters (80).  According
to Standard Methods (39), the coliform group comprises all of the aerobic
and facultative anaerobic, gram-negative, non-spore-forming, rod-shaped
bacteria which ferment lactose with gas formation within k8 hours at 35°C.
This group has been used as an indicator group through the years because
of the belief that it compares in many respects to the common enteric
pathogens.  According to Kabler (77), "This comparison Is suggested be-
cause it has been repeatedly observed that the coliform group and the
pathogenic enteric bacteria have survival rates in the same order of mag-
nitude under similar environmental conditons of temperature, pH, disin-
fection, or extended exposures to soil  or to fresh, polluted or salt waters.
Another reason why the coliform organism  has served as an  indicator  group
all these years is the tremendous numbers discharged in human feces,
estimated at 1.95 billion per person per day as an average (81).  Thus,  it
has become common practice that the presence of any members of the coliform
group in treated "potable" water is not acceptable regardless of the source
(80).  Recent reviews of the use and limitations of coliforms in the water
quality standards and other applications have been written by Wolf (82)
and Bott (83).
                                     161

-------
One of the principal problems in using the coliform group  for applications
other than potable water is that in addition to being found  in  the  feces
of warm blooded animals, coliforms can also be isolated  from the  gut of
cold-blooded animals, soils and many paints.  It must be recognized, as
demonstrated by many investigators, that there are both  coliforms of fecal
origin and non-fecal origin in virtually all natural waters  (80) (8k)(85).
Many investigators have demonstrated the presence of large coliform
counts in combined sewer overflow, urban runoff and natural water bodies
(86)(8?)(88)(89).  The questions are, in spite of these  high counts, do
these waters in actuality represent a health hazard to those who  come
into contact with them?  What is the seriousness of the  health  hazard?
Is  it possible to have relatively high coliform counts and yet  negligible
health hazards?  When applied to potable water supplies, the coliform
standard provides a safety factor which is highly desirable for obvious
reasons.  When applied to other applications, like combined sewer over-
flows and urban runoff, the same standard can result in confusion and
erroneous interpretations of conditions.

Fecal  Coliforms - There has been interest in distinguishing the fecal
from the non-fecal coliforms for many years.  In 190*1 Eijkman first pro-
posed an elevated temperature test for this purpose (81).  The  so-called
IMViC procedure described in Standard Methods (39) for differentiation
of the coliform group into Escherichla coli , Aerobacter aerogenes and
Escherichla freundii (intermediates) has been in various stages of  develop-
ment for a large number of years.  The primary reason for the strong in-
terest in the fecal group is the conviction held by most observers  that
these organisms represent a more sensitive measure of health hazard because
of their definite origin in the feces of warm blooded animals (39)(84)(90)
(91).   Geldreich  (92) found that the fecal  coliform group as an indicator
system has an excellent positive correlation with warm blooded animal
fecal  contamination.  He reported the following correlation percentages  in
the feces of various warm blooded animals.


              Sources of feces           % Positive Correlation

              Humans                              96.^4
              Livestock                           98.7
              Poultry                             93.0
              Cats, dogs and rodents              95-3


Geldreich went on to state that, "The fecal coliform test  is the most
accurate bacteriological measurement now available for detecting warm
blooded animal  feces in polluted water".  Other work by Geldreich et al
(93) showed that very few coliforms traceable to the gut of warm blooded
animals are associated with plants and insects,  which lends further sup-
port to the usefulness of this indicator group investigations and surface
water quality evaluations.
                                    162

-------
One important characteristic of the fecal coliform as an  indicator
group is that the behavior of these organisms  is very similar to that
of common enteric bacterial pathogens.  According to Van  Donsel et al
(3k), for example, there is a close relationship between  the growth
and survival of fecal coliforms and the pathogens, Salmonella and
Shigella.  This is one of the most  important features of  a  good indi-
cator system, especially when it  is not possible to use the pathogens
themselves.  Also, the fact that  the fecal coliform analysis does not
distinguish between fecal coliforms of human origin and fecal coliforms
of other warm blooded animals is  more of an advantage than  disadvantage.
The literature provides evidence  that human pathogens can be present in
the feces of poultry, livestock,  cats, dogs and wild animals (92).

Fecal Streptococci - The fecal streptococcus group, sometimes referred
to as the "enterococcus group", has been under consideration as an
indicator of fecal pollution for  about 25 years.  The primary interest
in this group is related to the fact that the normal habitat of the
bacteria included is the intestine of man and warm blooded animals.
However, studies have shown that  there is considerable variation in types
and numbers of fecal streptococci between various animals, even between
animals of the same species (95)(96).  One often cited advantage of the
fecal streptococcus group as an indicator is that certain species pre-
dominate in human feces and other species predominate in animal feces
(97).  For example S. faecal is strains are part of the former group
and S. bovis part of the latter group.  Another advantage is that
streptococci normally persist longer in surface waters than bacteria
in the coliform group, and apparently they do not multiply  in surface
waters (98).  Though streptococcus survival rates in natural waters may
not correlate with those of known pathogens, this situation provides
a safety feature for this  indicator system.

The principal disadvantage of the fecal streptococci is that some mem-
bers of the group are indigenous  to insects, vegetation, agricultural
soils and the feces of fresh water fish (99).  S. faecal is var. 1igui-
faciens  is an example of such a strain not limited to animal feces.
Obviously there is little sanitary significance to the group of
streptococci which do not originate in animal  feces.  It  is possible
that the use of this group alone  as a pollutional indicator in certain
situations may yield misleading results.

Salmonella Group - The reason for the interest in the salmonella group
of bacteria as a possible  indicator of pollution is obvious when one
considers that the causative agents of typhoid and paratyphoid fever,
and salmonellosis belong to this  important group.  Though there has been
consideration given to this group for a number of years (77), it has
only been recently that a fairly  reliable quantitative method for  its
detection has been developed (79)-  According  to Geldreich  and Van
Donsel (100) there are several hundred strains of Salmonella known to
be pathogenic and the presence of any one of these strains  in water
should be considered a serious health hazard.
                                   163

-------
The major problem with the use of the salmonella group is that the
occurrence of this population in polluted waters is highly variable.
It has been found, for example, that salmonella infection rates are
higher in animals than in human populat ion (101).   Even 13 percent
of clinically healthy cattle were found to be infected with Salmonella
in one study (101).  Thus the population of Salmonella in polluted
waters can be quite varied and the counts are considerably below those
of some of the other Indicator groups.   The summary results of some
work conducted in the Saline and Huron  rivers tributary to Lake Erie
demonstrate this point (102).
                     	Counts per 100 ml  relative to Salmonella
                                                Fecal         Fecal
    River            Salmonella   Coliforms    Coliforms   Streptococci
   Saline River           1          1472         22?          36?
   Huron River            1           274          24          109
Dutka and Bell (103)  concur that the introduction of salmonellae into
natural waters is intermittent and therefore natural dilution results
in low concentrations.   For this reason the above investigators recommend
a 50 gallon sampling  technique.  The undesirabi1ity of using such a
sampling technique for  routine application is obvious.

Other Possible Indicator Systems - Several other  indicator systems have
been suggested through  the years, namely the total  plate count, the
enteric viruses,  the  upper respiratory tract bacteria, and Adenosine
tri-phosphate (ATP)  measurements.  The total plate count is simply a
gross count of the bacteria present in a water sample.  Since many of the
bacteria are indigenous to natural waters, the count is not by any means
a sensitive indicator for the presence of enteric pathogens.  Some
indication of the degree of domestic sewage pollution can be obtained
by comparing the  total  counts which result from 35°C and 20°C incubation
temperatures.

There is much interest  in the enteric viruses since there are some that
are known to be the  causative agents of water-borne diseases.  The major
problem with the  use  of this group as a pathogen  indicator  ts that detec-
tion and enumeration  techniques are considerably  more difficult than the
case of bacterial indicators.  Investigators are  still searching for a
suitable technique that laboratory technicians could use routinely for
the detection of  small  amounts of viruses in large volumes of water (104).
A recent American Water Works Association committee report  (73) stressed
that since the ratio  of coliforms to viruses is approximately 92,000 to
1 in the case of  raw sewage and 50,000 to 1 in the case of polluted waters,
the coliform indicator  is superior to a virus  indicator.  This same obser-
vation is made in Standard Methods (39), that  is, if viruses are to be
                                    164

-------
used as Indicators, the major problem  in analysis  is  to concentrate
sufficient viruses for detection with  the result that  large volumes of
water must be processed to achieve a reasonable degree of precision.
Smith and McVey (105)  investigated the use of chlorine dioxide for
inactivation of viruses in synthetic combined sewer overflow samples.
The results of limited studies  indicated that a small animal-type
bacteriophage (virus) would not be as  good an indicator as the fecal
streptococci or total coltforms, but that a polioviru,s potentially could
be a better indicator than these bacterial indicators.  The authors felt
that disinfection treatment to  achieve a 5 log reduction in virus titer
would probably result  in the destruction of all viable viruses for most
situations relative to combined sewer  overflows.

In those cases where combined sewer overflow and surface runoff may in-
fluence the quality of waters used for swimming and other water contact
sports, indicator organisms that originate from the gastro-intestinal
tract may not be a sensitive parameter of potential health hazard from
diseases of the skin and upper  respiratory tract.  In this situation,
bacteria traceable to the upper respiratory tract or  the skin of humans
may be better since these are the ones involved  in eye, nose, and throat
infections.  S. aureus and P. aeruginosa are bacteria which have been
suggested for this purpose (39); however, a limited history of past usage
plus difficulty of routine analysis preclude the use of these organisms
at this point in time.

Moffa et al (106) used a luminesence biometer to analyze for ATP extracted
from synthetic combined sewer overflow samples as a means of monitoring
the effectiveness of various chlorine  and chlorine dioxide schemes for
disinfection purposes.  They obtained  good correlations between ATP reduc-
tions and both total coliform and fecal coltform reductions following
disinfection operations, which  prompted the authors to concluded that,
"...ATP represents a promising  possibility for more accurate and precise
control of disinfection".

Use of Indicator Count Ratios - In recent years, a very interesting and
useful  analytical tool has been developed for the  interpretation of
bacteriological data obtained from wastewaters and polluted receiving
waters.  The tool is simply to  determine the ratio of counts of bacterial
groups used for pollution indicators.  The ratios which have been employed
are:  total coliform (TC) to fecal streptococci  (FS) and fecal coliform
(FC) to fecal  streptococci.  However,  because of the wide distribution of
total coliforms in nature, only the latter ratio is considered to be of
any real  practical value.  The  case study reported by Geldreich (?8) in
which he employed FC/FS ratios  to establish the effect of heavy runoff on
a small lake, is a good example of how these ratios can be helpful in
assessing pollutional  loadings.

The usefulness of the FC/FS ratio is based on the  fact that fecal coliform
counts are considerably higher  in human feces than fecal streptococcus
                                    165

-------
counts.  The converse is true for most animal feces.  The FC/FS  ratios
reported for a number of different animal feces are as follows  (107):
Rat - 0.0*4, chipmunk - 0.03, rabbit - 0.0004, cat - 0.29, dog -  0.02,
and man - k.k.  Thus, the FC/FS ratios for domestic sewage would be ex-
pected to be considerably higher than for storm runoff.  The ratios for
combined sewer overflows would be somewhat in between the two extremes
above, depending upon the amount of storm runoff which is diluting the
raw domestic wastewater or the re-entry of the pollutional discharge.
Geldreich and Kenner (99) have investigated seven wastewater sources and
found that the FC/FS ratios varied from 4.0 to 27-9, and three stormwater
sources obtained had ratios from 0.04 to 0.26.  Following a review of the
literature, a leading manufacturer of membrane filters (108) accepted the
basic findings of Geldreich and Kenner (99) and suggested that when the
FC/FS ratio is equal to or greater than 4.0,  it is highly probable that
the pollution is from human origin.  On the other hand, when the ratio
is equal to or less than 0.7» it is highly probable the pollution source
is animal feces, such as livestock, poultry and household pets.  Un-
doubtedly, the low ratios usually found in urban runoff are due  to high
fecal streptococcus counts which originate from the feces of household
pets, rodents and bird droppings.

Selection of a Pathogenic Indicator for Combined Sewer Overflow and
Stormwater Runoff

The selection of a certain group of microorganisms as an indicator system
for pathogens should take into account the following traits or
character!sties:

     1.  The microorganisms should be present in large numbers in
         the feces of  humans and animals.

     2.  The microorganisms should always be  present when various
         types of enteric pathogens are also  present.

     3.  The growth and die-off characteristics of the micro-
         organisms in waters and wastewaters  should be as close
         to those of key pathogens as possible.  Ideally the
         indicator organisms should persist in nature a little
         16nger than all enteric pathogens.

     4.  The absence of the indicator microorganisms should mean
         that all enteric pathogen are also not present.

     5.  The laboratory detection procedure should be as routine,
         economical and reliable as possible.

It is obvious that what is described above is an ideal indicator system.
The variable characteristics of wastewaters,  combined sewer overflow and
                                   166

-------
storm runoff make it impossible to achieve completely all of the traits
listed.   The selection process is one of meeting them as closely as
possible.

peculiar Characteristics of Combined Sewer Overflows and Storm Runoff -
In contrast to wastewater flows, storm runoff only occurs during periods
of precipitation and thaws.  Even in well-watered regions, this type of
discharge occurs only 3 to 5 percent of the total time.  The character-
istics of urban runoff, including its bacterial constituents depend on
numerous factors such as intensity and duration of rainfall, time from
preceding rainfall, watershed characteristics, etc.  Usually the "first
slug" of urban runoff contains the highest microorganism densities.
Numerous studies have been reported  in which total coliforms, fecal
coliforms and fecal streptococci have been found in urban runoff (69)
(8?)(109)(110)(111)(112).  The population of total coliforms is generally
highest, followed by fecal  streptococci and fecal coliforms.  Though the
numbers vary considerably, it can be expected that the total coliform
counts per 100 ml are usually in the range of hundreds of thousands, the
fecal streptococci  in the tens of thousands, and the fecal coliform
usually less than ten thousand.

The microbiological characteristics of combined sewer overflows are even
more variable than  storm runoff since the added variable of degree of
mixture with raw sewage is involved.  Depending upon the hydraulics and
design of the sewer system, direct overflow from a combined sewer system
may not occur with  the same frequency as urban storm runoff.  As expected,
numerous studies have also demonstrated the presence of total coliforms,
fecal coliforms and fecal streptococci in combined sewer overflows, but
the counts were greater than urban storm runoff in all categories (7)(17)
(1 13) (1 I1*) (1 15) (1 16) .  A study by Benzie and Courchaine (114) for example,
showed that the total coliform, fecal coliform and fecal streptococci
counts in combined  sewer overflows were 8, 33 and 4 times as great as
those normally found in storm runoff respectively.  Burns and Vaughan (113)
showed that the fecal streptococcus densities were about the same in com-
bined sewer overflows and stormwater  runoff.  The biggest bacteriological
difference between  combined sewer overflow and storm runoff was found to
be  in the fecal coliform category, with the higher densities in the
former discharge as expected.

Indicator System Selected - All the evidence available in the literature
and actual experience acquired in working with combined sewer overflows
and storm runoff, strongly suggest that the fecal coliform indicator
system is the superior one at the present time for assessing the sanitary
quality of these flows.  The fecal coliform test comes closest to meeting
all the traits of an ideal  indicator system listed previously.  As stated
by Geldreich (92),  "The fecal coliform test is the most accurate bacterio-
logical  measurement now available for detecting warm blooded animal feces
in polluted water".   Man and other animal feces can contain the caustive
agents of all the Important water-borne diseases.  Until a better
                                    167

-------
indicator system becomes available, the assumption can be made that  if
fecal conforms are absent so are the pathogens of all water-borne
diseases, including viruses.   Likewise, if fecal coliforms are present,
the  likelihood of some microbial  pathogens being present is high, par-
ticularly bacteria of the salmonella group.  The major exception to
the above statements is the cysts of E. hlstolytica which are resistant
to common water disinfection  procedures, and thus do not behave as the
common enteric bacteria.

For some applications it may  be very useful to also include the fecal
streptococcus analysis as well  and express the bacterial data in the
form of the ratio of fecal coliforms to fecal  streptococci (FC/FS).  Use
of this ratio would be helpful  in assessing the pollutional impact of
combined sewer and storm runoff flows directly at the outfalls.  As
indicated previously, ratios  in excess of approximately A.O indicate
pollution of a human origin,  whereas ratios less than 0.7 indicate
pollution of animal origin.  These ratios should be established only
at the outfalls or immediately downstream of an outfall.  Because of
variable die-off characteristics, it would be highly questionable to
apply the same reasoning to samples obtained after residence times of
one day or more in receiving  waters or from other locations.

The use of the fecal coliform analysis is sufficient for characterizing
the effectiveness of pathogen destruction of any treatment process
applied to either combined sewer  overflows or stormwater  runoff.  Little,
if any, additional information is gained by the use of the FC/FS ratio in
applications such as these, except possibly where packinghouse wastewaters
or feedlot runoff are involved.

Recommended Laboratory Procedure

A clear distinction should be made at the outset between conducting
bacteriological analyses on raw wastewaters, storm runoff and combined
sewer overflows on the one hand,  and potable waters and other high
quality waters on the other.   In  the case of the former the counts are
generally very high, frequently in the range of hundreds of thousands
to millions per 100 ml  of sample.  A serial dilution must be conducted
in the laboratory in order to conduct these counts.  While careful
laboratory techniques should  be used in any case, it is apparent that
with such high counts, high precision work of the type used in research
studies is simply not required, and for that matter, an unnecessary
expense.

In the case of storm generated discharges, gross counts of a relative
nature are generally of primary interest.  It may be of interest to
know, for example, the difference in fecal coliform counts between in-
fluent and effluent of a treatment process; the variation in fecal
coliform counts with time of  overflow at a specific discharge point,
the difference in fecal coliform  counts for one land use type versus
another type within the same  community; or a comparison of counts in
                                    168

-------
combined sewer overflows at a number of different communities.   In all
the'above situations the relative differences are of greater  Importance
than the absolute numbers.  Thus It is more important that the procedure
employed be convenient to run, be economical, and be one in which re-
sults are obtained as quickly as possible, rather than one where maximum
accuracy Is the main criterion.

In the case of raw or potable water supplies, such laboratory analysis
features as convenience and economy are subservient to accuracy and pre-
cision.  Values In the order of one to several fecal coliforms per 100 ml
are of significance.  It  is possible that under certain conditions this
level of accuracy is desirable even in the case of storm flows.  A good
example is an impounded area used for recreational purposes, even to the
extent of water contact activity and fishing, to which storm runoff is
directed.  It is also possible that well treated combined sewer overflows
are discharged to the same water bodies, in which case the most sensitive
pathogenic indicator analysts available is not only desirable but necessary.

Sampling Procedure - The most suitable way of obtaining a sample for
bacteriological analysis  Is to collect a discrete or grab sample in a
sterile bottle and return the sample under iced conditions to the labor-
atory for  Immediate analysis.  When chlorinated effluents are being
sampled, it will be necessary to dechlorinate as described on page 491 of
Standard Methods  (39).  It Is best practice to collect bacteriological
samples in separate bottles containing the appropriate concentration of
thlosulfate solution.  These samples should not be used for other analyses
except possibly certain select analyses where interferences are known to
be minimal.   If non-chlorinated discharges are involved, the samples
collected  for bacteriological analyses can be used for other analyses upon
completion of the initial work.

As a general  rule bacteriological samples should not be composited.  Fre-
quently, however, where storm generated discharges are involved, it is only
possible or economically  practical to obtain composited samples for such
analyses.  Though admittedly this is not an ideal situation,  it is often
the only sample source for analysis, and the results must be employed or
interpreted with full cognizance given to this practice.

It  is usually very difficult to achieve ideal sampling conditions in most
storm generated discharge applications, since automatic samplers are
generally  used to obtain  the necessary samples.  Samplers that have
individual suction  lines  leading to separate sample bottles are obviously
the most suitable.   In this case, sterile sample bottles and tubing
should be  used.

Autoclaving the bottles is the best way of Insuring that sterile con-
ditions are achieved.  This may not always be the most practical way,
considering the fact that, as stated previously, gross counts are
usually of interest.  Ancillary studies conducted by this author have
                                    169

-------
demonstrated that washing the bottles  in an automatic  laboratory dish-
washer, using a good quality detergent, effective rinsing, and water
temperatures of the 80°C to 90°C range, while not achieving  strict
sterile conditions, will destroy vegetative cells.  The automatic
sampler head and  individual sample lines have to be disinfected as
well.  This can be done using a chlorine solution  with approximately
50 mg/1 of free available chlorine to  flush out the lines, and to
clean the head and other parts.  It is imperative that these surfaces
be well rinsed with hot water to insure that no trace of chlorine re-
mains.  The sample bottles should be attached to the sampling lines
using aseptic techniques, perferably in the laboratory rather than the
field.

Many of the automatic samplers in use  have a single sampling line which
directs the samples into one bottle or into a series of bottles.  In
both cases, the sample bottles should  be washed as described previously.
Naturally, once the first sample is drawn, the line is contaminated and
will effect all subsequent samples.  Automatic samplers which provide ,
a thorough pre-flushing sequence with  the new sample water prior to
introduction of the sample into the bottle, reduce somewhat the problem
of carry-over contamination of the intake line.   However there is no
convenient way of completely eliminating this sampling problem, and thus
it should simply be accepted as one of the normal  features of conducting
bacterial analyses on storm generated discharge samples using single line
automatic samplers.

Because of the number of organisms involved and the fact that relative
counts are usually of primary interest, the error introduced by this
sampling technique is normally not considered to be significant.  This
situation should not preclude the use of acceptable sample handling and
setup procedures after the sample is collected.   Care should still  be
exercised in cleaning the sampling lines and bottles as described pre-
viously, as well as the use of other aseptic techniques common in
bacteriological-type analyses.

Analysis Procedure - Once fecal  coliforms and other enteric organisms
are discharged from the ideal  habitat of the intestines of warm blooded
animals to the more hostile aquatic environment, they begin to die-off.
Work by Hendricks  (117) and Klock (118) among others have shown that the
rate of die-off is retarded at lower temperatures.   Thus, for best   re-
sults, the samples should be chilled (iced If possible)to 10°C or below
during sample collection, storage and transit to the laboratory.
Freezing of the sample should  be avoided since this will  cause a
destruction of microorganisms.  As stated in Standard Methods (39)  samples
for bacteriological analysis should be refrigerated upon receipt in the
laboratory and processed within two hours.   When samples are being  com-
posited over a 2k  hour period, it is obvious that  the first sample
obtained of the series will  have been stored 2k hours before it is
possible to transport the composite to the laboratory.  This once again
                                   170

-------
is one of the unavoidable features of working with samples of this type.
Die-off of microorganisms can be kept to a minimum by proper cooling of
the sample during the entire time the composite is stored in the sampler.
Relative gross counts will be meaningful as long as all samples are
handled in a similar fashion.

Fecal  coliform counts can be made using either the multiple tube fermen-
tation procedure in which results are expressed in a form of a statistical
number called the most probable number  (MPN) , or the membrane filter
technique (MF) ,  in which a direct count is made of colonies which form on
the filter pad.   In both cases it is imperative that the incubation tem-
perature be rigidly controlled; ^.5 +  0.2°C being the specification
generally recommended.   In recent years there has been a definite trend
away from the MPN procedure  in preference to the membrane filter procedure.
The primary reasons for this trend are  the latter technique is easier to
set up, requires less time for results, causes fewer logistical problems
in the laboratory, results in a direct  count rather than a statistical
count, and on the whole  is more economical to run per analysis.  The two
enumeration procedures, however, do not yield the same results.  The work
of Little et al  (119) on total and fecal coliform removals in nine oxida-
tion ponds in the southern United States showed, for example, that the
MPN values were generally higher than the MF values.  They found that the
MF value fell within the MPNt95% confidence interval approximately 80
percent of the time.  For most of the applications associated with storm
generated discharges, this disparity is inconsequential, compared to the
many other favorable aspects of the MF  procedure.  In any case, because
of the high counts involved, sampling problems and dilutions required
during analysis, either enumeration procedure results in an estimate of
the "true" count.

Standard Methods (39) discourages the use of the MF procedure for
chlorinated samples because of the lower recovery as contrasted to the
multiple tube procedure.   Though the precise reasons are not given,
presumably it may be because the shorter MF incubation does not allow
sufficient time for the "aftergrowth" of fecal coliforms protected within
solids from the action of the chlorine, or that the liquid medium of the
MPN technique provides a greater "buffering" effect in stimulating the
growth of semi-viable or stressed organisms, or that the membrane medium
tends to plug up since larger aliquots  of debris laden sample must be
processed to obtain reliable counts.  Bordner (120) of U.S. EPA in Cincinnati
is working on a pre-incubation enrichment step which possibly could
be applied to chlorinated samples in order to minimize the erratic results
which apparently can occur with the standard single-step MF technique.
It is apparent that it is only a matter of time before the MF procedure
will be modified to compensate for the  problems noted above.

It is recommended that the MF technique be employed for conducting fecal
coliform and fecal streptococcus analyses on all samples associated with
combined sewer overflow and storm runoff applications.  The experiences
                                     171

-------
of this author clearly indicate that the numerous advantages of the MF
technique far outweigh the MPN technique, even for the case of chlorinated
samples.  The MF technique is sufficiently sensitive considering the high
counts normally encountered and the less than ideal sampling procedure
which by necessity must be employed.  In certain instances where high
quality waters are involved it may be advisable to use the MPN technique,
especially on chlorinated samples, or to establish correlations between
the MF and MPN results.  Decisions of this type can only be made by the
particular investigator in each case.

Standard Methods (39)  or the membrane manufactuer's brochures (108) should
be consulted for the most suitable growth media and techniques to be
employed.  Sodium thiosulfate solution must be added to the bottles used
to collect or store sample with chlorine residuals.  Because of the random
nature of storm flow samples, it is generally best practice to make up the
medium fresh for each sampling program or to purchase pre-made sterile
medium in convenient sized ampoules.  All pre-dilutions must be made with
sterile-buffered dilution water.  It Is not necessary to use a new sterile
membrane holder and funnel for each new sample.   A thorough wash with
sterile water will  suffice to remove bacteria on the funnel and include
these organisms in the sample aliquot being filtered.  It  is good practice
to run the samples in the order of lowest estimated density to highest
density to minimize carry-over from one sample to the next.  Floating
debris and heavy grit often associated with storm flows are normally not
a problem if the sample is mixed well  and the aliquot for analysis is
pipetted from the mid-height of the container.  Also the necessity for
sample dilution tends to minimize the effect of particulates in the
original sample.   Some investigators have suggested that use of short
term blending of turbid stormwater  samples prior to analysis by the MF
procedure to minimize the effects of particulates, however experience
of the a'uthor with storm generated discharge samples have not proven
this step to be necessary.

RECOMMENDED MEASUREMENT OF EUTROPHICATION POTENTIAL

There has probably been more attention given to the problem of eutrophi-
cation during the last fifteen years than to any other water resources
problem.  As stated in the National Academy of Sciences publication (121)
Eutrophicatton:  Causes, Consequences, Correctives, "The term 'eutrophic1
means wel1-nourished;  thus'eutrophication1 results from the proliferation
of plant life in water bodies as a result of the additon of plant
nutrients."  Algae are the category of plant life primarily involved in
most water bodies, though in some cases rooted aquatic plants are also
a manifestation of a eutrophic condition, particularly its advanced
stages.  Eutrophication conditions are generally associated with fresh
water lakes, but it is possible for rivers and estuaries to exhibit
some form of eutrophic behavior under the proper set of environmental
conditions.  The nutrients which are required to stimulate plant growth
can originate from numerous sources, namely, domestic and  industrial
                                   172

-------
wastewaters, urban and rural runoff, combined sewer overflows, under-
ground flow, and rainfall.  Thus these sources are the result of both
natural and man-originated activities.  The sources of primary interest
in this report are those associated with storm flows to water bodies.

It should be kept in mind thateutrophication is a natural process which
takes place in all water bodies.  The activities of man can greatly
increase the rate of this natural activity to the point where unde-
sirable results can occur.  Accelerated eutrophication or "cultural"
eutrophication can cause changes or shifts in the population of biota
in a water body.  It can also adversely influence the esthetics of a
water body, impair its use for recreational and water supply purposes,
and reduce the values of properties along the shoreline.  Excessive
rooted vegetation, also a manifestation of accelerated eutrophication,
fouls the air and uses up dissolved oxygen as a result of death and de-
compos11 i on.

Possible Nutrients Contributing to the Eutrophication Problem

During photosynthesis sunlight provides the energy for the production of
plant cellular materials from basic raw products including carbon
dioxide, water, nitrates, phosphates, plus a number of other trace salts
which include the elements potassium, iron, sulfur, calcium, magnesium
and cobalt among others.  Molecular oxygen is the major by-product of
this growth activity.  A chemical assay of plant material will show that
the dry wight will consist primarily of carbon, oxygen, hydrogen,
nitrogen, and phosphorus, plus a very small weight fraction of numerous
trace elements.

It is obvious that in order to have plant growth all  the above elements
must be present.  In agriculture when plant production is of primary
concern, it is common practice to carefully examine the soil environment
to determine if any key element or nutrient is present in a low concen-
tration or lacking, since it will be this particular element that Will
eventually limit plant growth.   This particular element would be con-
sidered the "limiting" one since at some point in the plant production
cycle it will  begin to control  growth.  Obviously, at this point,
fortifying the soil with the element will result in the resumption of
the plant growth.  In the case of eutrophication it is the opposite
approach which is employed; that is, to attempt to limit plant growth
by controlling one of the most critical  plant nutrients.  One of the
principal eutrophication research objectives through the years has been
directed toward determining which is the most effective nutrient to
control.  Because trace nutrients are available in most natural  waters
to support algal growth, very limited success has been achieved in
attempting to control eutrophication by limiting such elements as
potassium, iron, calcium, silicon and magnesium.  Most efforts have
centered around the control of various forms of carbon, nitrogen, and
phosphorus.
                                   173

-------
Carbon - The largest fraction of plant cellular material consists of the
element carbon.  The source of carbon in plant growth is carbon dioxide
which  is obtained directly from the atmosphere in the case of terrestrial
plants, and from that dissolved in water in the case of aquatic plants
like algae.  In the case of the latter, bound carbon dioxide  in the form
of bicarbonates and carbonates can also serve as carbon sources, but this
availability is tied into other chemical properties of water such as pH
and dissociation kinetics.  Gaseous carbon dioxide can also be  introduced
into water bodies as a result of the aerobic degradation of organic matter
directly in the receiving water body.  Carbon dioxide is one of the main
end products of this degradation.   Several  years ago Kuentzel (122) sug-
gested that in most situations biologically produced carbon dioxide de-
rived  from organic matter is the limiting nutrient in the eutrophication
process.  This work instigated the carbon dioxide versus phosphorus
limiting nutrient controversy which occurred during the early 1970's.
Presently it is generally accepted that while carbon dioxide is necessary
for plant growth, it is so readily available in all aquatic environments
and thus very unlikely to be limiting to plant growth.  As Fruh (123)
noted, carbon dioxide may cause growth stimulation without being the
limiting nutrient in the sense of  "Liebig's Law of the Minimum".

It would appear that the use of any type of carbon measurement on storm
flows would not provide a sensitive indicator or parameter of eutrophic
potential.   Maier and McConnell (12^) felt  that organic carbon type
measurements, as for example the TOC, conducted on lake and river waters
do provide an indication of the state of productivity of these waters.
Water bodies in a eutrophic state  as a rule exhibit high rates of organic
matter production usually expressed in grams of carbon per year per square
meter of water surface area.  Conducting TOC analyses on storm flows, on the
other hand, wi11  yield little information regarding the contribution of
these flows to a specific eutrophication problem.

Ni trogen - Approximately 6 to 8 percent of the total  dry weight of plant
cellular material consists of nitrogen, and for this reason is considered
one of the key nutrients of plant  growth.  Nitrogen quantities are always
considered whenever  nutrient loadings or nutrient budgets are conducted
on water bodies (125) (126).   Aquatic plants can utilize nitrogen in either
the nitrate or ammonium ion forms, both of which can be present in surface
runoff and combined sewer overflows.  There are some blue-green algae
which have the ability of utilizing gaseous nitrogen as their sole
nitrogen source in the event inorganic forms are not available.  For this
reason Fruh (123) felt that to try to control eutrophication by limiting
the supply of combined nitrogen alone is an exercise in futility.  He
added further that because nitrogen fixing algae are not indigenous to
high saline waters, it is possible and likely that nitrogen can be the
limiting nutrient in coastal water eutrophication problems.  Though
nitrogen may not be the limiting nutrient in all situations,  it is still
considered one of the key nutrients, and the availability of large amounts
of inorganic nitrogen will no doubt intensify the problem greatly.
                                    174

-------
Conversely its unavailability will limit algae growth in fresh waters to
the nitrogen fixing varieties which in itself is a form of eutrophication
1 imitations.

Phosphorus - Algal cells contain anywhere from 1 to 3 percent phosphorus
on a dry weight basis.  By far, the greatest amount of attention has been
given to this element as the key nutrient in the eutrophication process.
Phosphorus is available to plant life in the phosphate ion form only, and
unlike nitrogen,  it is never available to plants or present in nature in
the elemental form.  For this reason many researchers in the eutrophication
area agree with Fruh (123) that since the distribution of phosphorus in
nature is largely  influenced by the activities of man, and since this
element is an indispensable building material in all  living matter, its
control is the single most efficient means of limiting eutrophication in
fresh waters.

Others - As stated previously, in addition to the key elements of carbon,
nitrogen and phosphorus, aquatic plants require a large number of other
elements for optional growth.  Some of these elements, even though re-
quired in extremely low concentrations, are absolutely essential for
growth.  Iron, for example, is required for key enzymes involved in cel-
lular energy transformations.  If it were possible to totally eliminate
iron from the aquatic environment, presumably algae growth would cease.
Given the wide distribution of iron in nature, the prospect of such a
situation occurring in a body of water simply does not exist.   Thus even
though iron theoretically can be a limiting nutrient, its use as a
eutrophication parameter in any type of discharge would be completely
worthless.

In addition to iron, Stewart and Rohlich (127) have listed a number of
other elements that researchers have considered as possible eutrophteat ion
parameters.  These include magnesium, calcium, silicon, sulfur, manganese,
sodium, potassium and cobalt.  Some attention has also been given to cer-
tain vitamins and micronutrients.  At the present time, none of these
would be considered seriously as a sensitive measure of eutrophication
potent ial .

Selection of the Best Eutrophication Parameters

From the evidence available at the present time, analyses for nitrogen
and phosphorus in storm generated discharges will provide the most suit-
able parameters of the eutrophication potential  of such discharges.

Nitrogen - Nitrogen occurs in nature in five forms, namely, elemental
nitrogen, organic nitrogen, ammonia nitrogen, nitrite nitrogen and
nitrate nitrogen.  Analyzing water samples for elemental  nitrogen is
of little value.  Of the remaining forms, the reduced forms of nitrogen,
organic and ammonia, are found in all municipal  wastewaters, and to some
extent in surface runoff.  The oxidized forms of nitrogen, nitrite and
nitrate, are usually present in the effluents of biological treatment
plants and surface runoff.   These forms are normally not present in raw
                                    175

-------
municipal wastewaters.  The concentration of nitrate  is much higher than
nitrite  in most water samples, and for this reason the nitrite analysis
is often eliminated in the nitrogen series.  From the standpoint of
impact on a receiving water body, it is the total nitrogen being added
that is of primary interest, since all  four forms are eventually avail-
able to plant life following transformations in the receiving water body.
Unfortunately, a single analytical procedure is not available to Include
all the nitrogen forms.

It is recommended that basically two nitrogen analyses be conducted on
storm flows.  The first would be for the total  oxidized forms of nitrogen,
that is, the sum of the nitrite and nitrate forms.  This can be done by
oxidizing all the nitrite nitrogen to the nitrate form and conducting one
analysis on the latter ion.  Or it is possible to reduce all the nitrate
nitrogen £0 the nitrite form and just run the nitrite analysis.  Both of
the most popular nitrite and nitrate analyses are colorimetrlc procedures,
but of the two, the nitrite analysis is considered to be superior.  The
nitrate nitrogen analysis  is not as sensitive a procedure as the nitrite
analysis, and it is subject to more serious interferences, especially the
chloride ion (128).  Chloride is commonly found in storm generated dis-
charges, especially in cities along coastal areas and during the early
spring in northern cities where salt is used for deicing roads.  Thus it
is recommended that the nitrite analysis be used for quantitating the
total oxidized nitrogen in storm generated discharges.

The second nitrogen analysis would be for the reduced forms of nitrogen,
namely the total of ammonia and organic nitrogen.  It is recommended that
the Kjeldahl procedure be used for this purpose.  This procedure includes
all nitrogen in the amine  (-Nh^) form.   This form of nitrogen is a de-
sirable plant nutrient,.especially when present as ammonia.  Organic
nitrogen undergoes a biochemical process in all natural waters called
ammonification, in which the amine radical  is released from complex
organics as ammonia.

It is recommended that both the oxidized and reduced nitorgen analyses
be expressed in terms of the weight of nitrogen rather than the individual
ions.  In this way the two nitrogen values can  be added to arrive at the
total nitrogen content.  In many storm flow applications, only this one
value is actually needed.   In some applications it may be necessary to
know in what forms the nitrogen is present, as  for example, when certain
wastewater treatment processes are involved.  A good example is
chlorination.  Chlorine combines with organic and ammonia nitrogen to
form chloramines.  These compounds of chlorine and nitrogen which still
exert biocidal  action in receiving waters, are not derived from the
oxidized nitrogen forms.   Knowledge of the amount of Kjeldahl nitrogen
is also advantageous since a stoichiometric estimate of the nitrogenous
oxygen  demand (which usually does not appear in the BOD^ test) can be
made.
                                   176

-------
Phosphorus - All phosphorus in nature  Is present as  in the form of P0| = ,
its highest oxidized form.  The phosphate  ion can occur  in a variety of
soluble and insoluble inorganic compounds,  including orthophosphates and
polyphosphates or complex phosphates.   In addition,  the  phosphate ion can
occur in both living and dead organic matter.   According to Barth (129)
it 5s possible to obtain fourteen different phosphorus fractions depending
upon the sample processing technique and analytical  procedure  involved.
Because of these possibilities, Barth  recommends that the total phosphorus
analysis be conducted for situations involving the evaluation of treatment
processes.  The same can also be said where impact on a  receiving water
body is of concern.

It is recommended that the total phosphorus analysis be  conducted on storm
generated discharge samples, as the nutrient parameter for phosphorus.
Since the phosphate ion tends to combine readily with a  number of metalics
like iron, calcium and aluminum to form  insoluble compounds, in certain
treatment processes for storm generated discharges it may be very helpful
to run both the total soluble phosphate and the total phosphate.  This can
be accomplished by passing a sample through the same filtering medium re-
commended for suspended residue analysis, and conducting the total phos-
phorus analysis on both the original sample prior to filtration and on the
fl Itrate.

Nitrogen and Phosphorus Concentrations Associated with Storm Generated
Dl scharges

A  number of studies have been conducted  in  recent years  which already
demonstrate that storm generated discharges can contribute sizable amounts
of nutrients to receiving water bodies.  Weidner et  al (88) examined the
surface runoff quality from six different  rural sites around Coshocton,
Ohio and showed that the kg/ha  (Ibs/acre) of total solids, BOD, COD,
phosphate and total nitrogen varied considerably for various types of
agricultural crops and practices.  For example, on one particular plot
the phosphate loading varied from 2.3  to 310.2 kg/ha (1.1 to 27-7 Ibs/
acre) and the total nitrogen varied from 123-3 to 265^ kg/ha (11 to 237
Ibs/acre).  They concluded that runoff  is a factor in stream pollution
and that  it must be considered when one evaluates the quality of any
stream or receiving water.

Hetling and Sykes  (130) conducted an extensive study of  a nutrient
balance on Candarago Lake  in New York.  Under nutrients  they included
total and soluble phosphorus, the full nitrogen series,  magnesium,
potassium and chlorides.  They felt that none of the latter three
elements were algae limiting in the study  lake.  An  interesting finding
of their study was that two-thirds of  the nitrogen in streams tributary
to the lake was either in the nitrite or nitrate form.

Two investigations concentrated on the quality of urban  runoff.  An often
quoted study by Weibel et al (69) conducted on a 10.8 ha (27 acre) site
                                    177

-------
in Cincinnati over a 13 month period, resulted in the following nutrient
load ings:

             Nutrient         Range, mg/1        Mean, mg/1

             NO  - N          0.02 - 0.2            0.05

             NO  - N          0.1  - 1.5            0.40

             NH. - N          0.1  - 1.9            0.60
             Org - N          0.2  - 4.8            1.70

      Total   SOI.  POZ,         0.07 - 4.3            0.80
It is particularly noteworthy that the nitrite nitrogen concentration is
insignificant compared to the other nitrogen forms and that approximately
15 percent of the total nitrogen loading is in the oxidized form.  Sartor
et al  (131)  analyzed the street surface runoff from 12 cities in the U.S.
and concluded that storm runoff is a more serious source of pollutant in
many areas than municipal wastes.  They estimated that the kg/curb m
(Ibs/curb  mile) of phosphates, nitrates, and Kjeldahl  nitrogen were 0.30
(1.1), 0.027 (0.094) and 0.62 (2.2) respectively, and  that the largest
percentage of the nutrient loading was associated with the fine solids
(less than 246y) fraction of the street surface contaminants.  It should
be noted that the nitrite ion concentration was not included in this
study and  that the largest share of the nitrogen loading is in the amine
form rather than the oxidized form.

Several  studies have been conducted in  which pollutional loadings from
separate storm sewers were compared to those from combined sewer overflows
The study  conducted in the Detroit-Ann Arbor area was  very extensive and
included a comparison of nutrient loadings, the salient features of which
are summarized in Table VI-7 (8)(15)-  As noted, analyzing for the soluble
phosphate  alone can result in an erroneously low conclusion.  Also, the
nitrate fraction of the total nitrogen  loading can be  significant;
especially for a separate storm sewer or unsewered discharge.  Similar
comparison studies were also conducted  in Washington,  D.C. (7)(70), in
which it was concluded that the average nutrient concentrations in
separated  storm sewer runoff were approximately one-third of those in
combined sewer discharges.  Selected nutrient  loading  data are presented
in Table VI-8 for a six month period of testing.  These data demonstrate
the preference of running the total phosphate  analysis over that of one
of the other fractions.

Nutrient analyses were also conducted during combined  overflow studies
at two cities in Wisconsin (17).  The mean Kjeldahl nitrogen concentration
at the 95% confidence level was 17.6 + 3.1 mg/1 for 12 "first flush"
events and 5-5 + 0.8 mg/1 for 44 extended combined overflow events, at an
overflow site in Milwaukee.  At Kenosha, the mean Kjeldahl nitrogen and
PO^ as P concentration of 23 overflow events were 12.9 and 5-1  mg/1,
respect ively.
                                   178

-------
      Table VI-7.   COMPARISON OF NUTRIENT LOADINGS  FROM  SEPARATE
           STORM AND COMBINED SEWERS,  DETROIT - ANN ARBOR  AREA
Separate system
Cone. ,
Max
NH, -
Org -
NO- -
N
N
N
Soluble POj
Total
P04
2.
4.
3.
» 3<
16.
0
0
6
4
4
ms/10
Mean
1.0
1.0
1.5
0.8
5.0

Load! ng"
kg/ha
7.8
4.5
9.0
10.0
31.4
Combined system
Cone.
Ibs/acre max
0
0
0
0
2
.7
.4
.8
.9
.8
134
38
2.8
21.2
43.2
, mg/1D
mean
12.6
3.7
0.5
7.7
14.6
Loading"
kg/ha
69.4
17.9
1.7
62.7
123.2
Ibs/acre
6.2
1.6
0.15
5.6
11.0
a.  During three month summer period.
b.  Annual values
      Table VI-8.   COMPARISON OF NUTRIENT LOADINGS  FROM  SEPARATE
             STORM AND COMBINED SEWERS,  WASHINGTON,  D.C. AREA3

NH, - N
Total - N
Ortho - PO,
Total - PO^
Separate system,
Range
0.5-6.5
0.2-4.5
mg/1
Mean
2.1
1.3
Combi ned
Range
0-4.7
1.0-16.5
0.1- 5.0
0.8- 9.4
system, mg/1
Mean
1.5
3-5
2.0
3.0
a.  Six month spring and summer period.
                                    179

-------
 Two recent  studies  on  urban  runoff  quality  shed  new light on nutrient
 loadings  from this  source.   The  first  was conducted by Kluesener and
 Lee (132) on  a 50 ha  (123 acre)  residential  area site in Madison,
 Wisconsin.  They estimated that  27  percent  of  the area is impervious
 (streets, walks, roofs, etc.).   Thirty-four  storms were monitored.
 The results of their study compared  to those of  others are summarized
 in  Table  VI-9-
              Table VI-9.   COMPARISON OF AVERAGE ANNUAL
                 NUTRIENT CONCENTRATIONS IN URBAN RUNOFF
Source of
data3
Sawyer
Sylvester
Weibel
Bryan
AVCO
Kluesener
& Lee
Nitrogen as N,
mg/1
NH3 NO^
0.28
0.53
0.60 0.40
—
—
0.45 0.60

Org
--
2.0
1.7
—
0.8
3.5
Phosphorus as
mg/1
Dissolved
0.22
--
0.26
0.14
0.22
0.57
p,
Total
0.56
0.15
—
0.19
—
0.98
 a.   Refer  to  Reference  19,  p.  931.
As noted, the nitrate concentration exceeds the ammonia concentration.
They found that the concentration of all parameters was highest during
the first flush.  Extrapolating the results of the study to the entire
Lake Winona basin, they estimated that approximately 80 percent of the
total phosphorus and 35 to 40 percent of the total nitrogen arise from
urban runoff.  The second study conducted by Whipple et al (133) evalu-
ated the urban runoff from a 13-5 sq m (5.2 sq mi) mainly residential
area with a population of 11,700.  The nitrate and phosphate contributions
were 59 and 32 kg/day (130 and 70 Ibs/day), respectively over a 4 day low
rainfall period and 154 and 263 kg/day (340 and 580 Ibs/day), respec-
tively over a 7 day heavy rainfall period.   The authors concluded that
for rivers with high loadings of nutrients  emanating from urban runoff,
it would be entirely unrealistic to attempt to control  algae blooms by
treatment of dry-weather flow only.
                                   180

-------
Recommended Analyses for Nitrogen and Phosphorus

As noted previously, two analyses are recommended  for establishing the
nitrogen content of storm flow samples, namely, total Kjeldahl nitrogen
arid total oxidized nitrogen.  The total phosphorus analysis  is recommended
for quantitating the second nutrient of interest.  Salient features of
the analysis procedures recommended for these nutrients follow.

Nitrogen - The total Kjeldahl analysis as described  in Standard Methods
(39) or the 1971 EPA Manual  (^0)  is recommended for  the determination of
total nitrogen  in the reduced form.  This procedure  can be applied to
most combined sewer overflow and storm runoff samples without much
difficulty.  If the analysis is conducted within  1 or 2 days of the time
of collection, storage under refrigeration would  be  advisable.  During
this period little, if any at all, microbial nitrification would be
expected to take place.  The same would be true in the case of ammonifi-
cation, but even if some does occur, it would not  affect the final re-
sults since the analysis includes all reduced nitrogen forms.  If longer
storage  is anticipated it would be advisable to add  mercuric chloride
as a preservative to arrest bacterial activity.   This will insure that
the proportion of nitrogen in the reduced and oxidized forms will remain
as collected, and that nitrogen will not be  lost  through the microbial
process of denitrification.

When it  is necessary to run the ammonia nitrogen  analysis  it is recom-
mended that the distillation procedure be employed.  Experience with the
use of this procedure on combined sewer overflow  samples indicates that
the borate buffer recommended in the 1971 EPA Manual (40)  is preferred
to the phosphate buffer recommended in Standard Methods (39).  The latter
buffer solution may result in the formation of precipitates  if a high
concentration of hardness-causing cations is present in the sample.

The procedure which is recommended for total oxidized nitrogen is the
cadmium  reduction method presented by Strickland  and Parsons (42) and
in Standard Methods (39).  For most storm flow applications there is no
need to distinguish between the nitrite and nitrate  ion forms, therefore
an aliquot of sample can be passed directly through  the reduction column
and, as  recommended, analyzed for the nitrite concentration.  The colori-
metric analysis procedure  for nitrite  is a very sensitive one and not
subject  to many interferences.  It  is  important that a control be passed
through  the reduction column each time a new series  of samples is
analyzed, to make certain  that the reduction properties of the cadmium
surface  are intact.  The life of the column can be prolonged by passing
only clarified samples on  the cadmium surface.  Samples heavily laden
with organic and inorganic particulates should be clarified  by first
allowing the heavier particles to settle and passing the upper clarified
portion  through filter paper.  For highly turbid  samples,  the use of
zinc sulfate flocculent would be helpful for the  clarification process.
                                   181

-------
Phosphorus - Two basic analysis steps are  involved  in the total  phos-
phorus analysis.  The first involves a rigorous digestion step  in which
all phosphorus  in the sample is released from organic matter and hydro-
lyzed to the ortho-phosphate form, and the second  involves a colorimetric
analysis step where the ortho-phosphate ion is complexed with ammonium
molybdate to form a blue colored solution.  Standard Methods (33) suggests
three possible digestion procedures.  The most rigorous of the  three, and
the one used successfully by the Milwaukee Sewerage Commission  (13*0  in
a very extensive study on removing phosphorus from municipal wastewaters,
is the perchloric acid digestion method.  A major disadvantage  of this
technique is the explosion hazard of the perchloric acid mixture.  Of the
remaining two methods, namely,  the sulfuric acid-nitric acid digestion
and the persulfate digestion,  the latter is much more convenient to run
though it may not be as rigorous a digestion procedure as the former.
Zanoni (135) compared the persulfate digestion method to the ashing
method and found the results to be comparable for most liquid samples.
The ashing method is considered  to be an excellent control method, but
it is far too time consuming to use for routine analysis.  Therefore,
it is recommended that the persulfate digestion method be employed for
the routine analysis of combined sewer overflow and storm runoff samples.
In select cases wehre heavy sludges or samples containing much  organic
debris are involved, the ashing method is recommended.

Three different colorimetric methods are suggested  in Standard  Methods
(39)  for ortho-phosphate determination, any one of which could  be applied
to storm flow samples.  The authors have a preference for the ascorbic
acid method or the single reagent method of the 1971 U.S. EPA Manual  (*tO)
which also includes the use of  ascorbic acid as a reducing agent.

RECOMMENDED CHOICE OF METAL ANALYSES

The analysis of metals has been somewhat ignored in most storm  generated
discharge projects.  However,  the potential toxic effect of some metals
on the biological community merits more attention.  Since metals in
storm generated dishcarges could exist in relatively high concentrations,
depending upon the industrial  discharges (including those into  the air),
metallic corrosion and the vehicular traffic in the area, periodic analysis
is necessary.  If a certain constituent has a high concentration, more
frequent tests may be necessary.

With the development and improvements in atomic absorption, the analysis
of various metals has been simplified and the accuracy improved.  Atomic
absorption spectroscopy works  on the principle that a sample is aspirated
and then atomized in a flame.   The amount of 1ight absorbed by  a certain
element is proportional to the  concentration of the element  in  the sample.
Calibration curves are used and the method is relatively rapid  and free
of interferences.
                                    182

-------
In practice however, special care must be taken prior to actual analysis.
The first important aspect  is in the sampling procedure.  The bottles
must be absolutely clean and inert to the metallic compound being de-
termined so there is no leaching into or from the sides of the container.
Secondly, all glassware must be scrupulously clean for all samples and
standards and extreme care must be taken in standard preparation due to
the minute quantities being measured.  Finally, the digestion procedure
must be followed carefully to insure that there is no loss of metal in
this step.  All of these are critical due to the fact that the metals may
be found in concentrations as low as the ug/1 (ppb) range and a slight
error in analysis can result in a large error in sample concentration
determinations.

Heavy Metals

Heavy metals analyses were performed in only two of thirty-eight studies
of storm flow characterization and treatment of (1)(37).  It appears that
the concentrations of heavy metals have been considered relatively unim-
portant compared to potential oxygen demand and particulate concentration.
Several of the studies, however, indicate that problems with BOD results
may have been due to the presence of toxic heavy metals (11)(131)(136).
Heavy metals may enter the storm or combined sewer system from various
sources.  Runoff from streets and highways may contain lead, zinc and
copper and to a  lesser extent, chromium, mercury and nickel.  The average
amount of these metals along certain streets of six cities studied by
Sartor and Boyd  (2) are listed in Table VI-10.  Another possible source
of toxic metals  in storm generated discharges may be metallic corrosion
(i.e. Cu, Cr, Zn).  Drainage of industrial  areas may contain any of the
metals used  in that industry.

When studying the quality of storm flows or when operating a stormwater
treatment system, it is recommended that a composite sample of the flow
be analyzed for  lead, zinc, copper, chromium, mercury, cadmium, arsenic,
nickel, and tin four times a year (seasonally).  Based upon the results
of these tests, a decision can be made as to how often certain heavy metals
will have to be analyzed thereafter.  It is expected that lead, zinc,
copper and chromium may be measured routinely.  In certain combined sewer
areas serving known industries, or in certain storm sewer discharges from
areas of heavy vehicular traffic, it may be necessary to do more fre-
quent analysis.  Also, based on the concentrations found in the discharges
going to treatment, the sludge resulting from this treatment should be
analyzed accordingly.  This is especially important if the possibility
exists that the sludge may be discharged to a subsequent process sensi-
tive to heavy metal toxicity.

Table Vl-ll has been constructed for nine heavy metals indicating possi-
ble sources, effects, preservation and storage, and pretreatment methods,
and the analytical  method.  In cases where chromium concentrations are
high, it may be desirable to know whether the chromium is present in the
                                    183

-------
                              Table VI-10.   HEAVY METALS LOADING INTENSITIES
                                                    Ib/curb mile (kg/curb km)
     	Zinc	Copper	Lead	Nickel	Mercury	Chromium

     San Jose-I      1.40   (0.39)    0.49  (0.13)    1.85  (0.51)    0.19  (0.05)   0.200  (0.05)   0.100   (0.03)

     San Jose-ll     0.28   (0.08     0.020  (0.005)   0.90  (0.25)    0.085 (0.024)  0.085  (0.024)  0.140   (0.04)

     Phoenix-ll      0.36   (0.10)    0.058  (0.016)   0.12  (0.034)  0.038 (0.01)   0.022  (0.006)  0.029   (0.008)

     Milwaukee       2.10   (0.59)    0.59  (0.16)    1.51  (0.42)    0.032 (0.009)  ~     ~       0.047   (0.013)
i—>
™    Baltimore       1.30   (0.36)    0.33  (0.09)    0.4?  (0.13)    0.077 (0.02)   —     —       0.450   (0.13)

     Atlanta         0.11   (0.03)    0.066  (0.018)   0.077 (0.04)    0.021 (0.006)  0.023  (0.006)  0.011   (0.003)

     Seattle         0.37   (0.103)   0.075  (0.021)   0.50 .(0.14)    0.028 (0.008)  0.034  (0.009)  0.081   (0.023)


     Arithmetic
        means        0.75   (0.21)    0.21   (0.06)    0.68  (0.19)    0.060 (0.017)  0.080  (0.02)   0.12    (0.034)
            Taken from  reference 131

-------
                         Table Vl-ll.  INFORMATION SURVEY FOR VARIOUS HEAVY METALS
OT
Metal
Copper
Chromium
(see below)
Lead
Mercury
Nickel
Arsenic

Tin
Cadmi urn
Zinc
Chromium,
hexavalent
Source
industry
corrosion of pipe
water system
industry
corrosion
industry/mining
smel ter/pl umbl ng
industry
sea water
industry
industry
insecticides

industry
plating
industry
galvanized pipe
industry
deteriorated
galvanized pipe
Industry
Effects
not much
large doses cause
liver damage
carcinogenic
toxic
cumulative
bone poison
poison
(respiratory)
dermatitis
poison
carcinogenic

no effect
toxic/feed
poisoning
taste
carcinogenic
toxic
Preservation
acidify to less
than pH 3.0
with HNO,
acidify to less
than pH 3.0
with HNO,
acidify to less
than pH 3.0
acidify to less
than pH 3.0
with HNO-
acidify to less
than pH 3.0
with HNO,
acidify to less
than pH 3.0
with HN03
acidify to less
than pH 3-0
with HNO,
acidify to less
than pH 3.0
with HNO-
acidify to less
than pH 3.0
with HNO-
trcre
Storage
time
no limit
no limit
no 1 imit
as soon
as possible
no limit
no limit

as soon
as possible
no limit
no 1 Imlt
<2k hours
P re treatment
digestion
procedure outlined
for AA
digestion with
HNO -HCI
digestion with
HN03-HC1
digestion wi th
HNO, only
digestion with
HHOj-HCl
digestion with
3
digestion with
HNO-
digestion with
HNO -HCI
digestion -with
HNO--HC1
remove sol Ids
by centrifuge
Analysis
AA
AA
AA
flameless AA
AA
colorimetric

AA
AA
AA
colorimetric
(diphenyl
carbazlde)

-------
trivalent or the more toxic hexava lent form.  In this case a  portion of
the sample should be separated before addition of HNO^ preservative and
analyzed for hexavalent chromium by the diphenylca>bazide method within
2k hours.

Other Metals

Metals which  are not toxic to biological life or which are  not commonly
expected to be present need not be routinely analyzed.   In certain cir-
cumstances, however, analysis for some of these metals may be desirable.
Certain  industies contribute metals of a specific type and street salting
in northern climates can contribute sodium, calcium and  iron in various
concentrations.  The amount of material  present will indicate the fre-
quency of analysis needed.

Ferrous  iron is more soluble than ferric acid and may not be removed by
most storm generated discharge treatment processes.  However, oxidation
at a later time may cause formation of an unsightly ferric precipitate
in the treatment process effluent as in  the  receiving water.

Recommended Laboratory Procedures

Heavy Metals - Analysis of samples for heavy metals should always Include
the total amount of metal present.  Additional  analyses for  the soluble
or insoluble portions are optional and may be useful in certain instances
for evaluating the performance of some treatment units.  Evaluation of
the potential  effect of toxic metals on  the receiving water, however,
requires that  the total amount be measured.   Therefore, because of the
large amount of organics generally present, samples should be digested
prior to analyses.

It is recommended that the following procedures be used for digestion of
the samples and for analyses.  Reagent blanks should be run with each
set of samples from the digestion step on.  Samples should be analyzed
in duplicate starting with the digestion step.   Samples for metal  deter-
mination should be split off, preserved  with nitric acid to pH less than
2.0, and refrigerated until digestion.  They may be stored in this manner
up to 6 months.  The container used will depend upon the analysis to be
run and samples should be acidified soon after sampling to minimize
adsorption on  the contralner walls.  Polyethylene containers are recom-
mended for most metals other than mercury.

Listed below are the recommended digestion procedures for various heavy
metals.

     1.  Lead,  zinc, copper, chromium, cadmium, and nickel - The
         nitric acid-hydrochloric acid procedure (described  later)
         should be used.
                                   186

-------
Mercury and Tin - Mercury pretreatment will depend on
the type of sample involved.  Effluent samples of good
quality or any other sample that has little solid
material can be analyzed directly without digestion.
However, samples from storm generated discharges and
the resultant sludges from the treatment of these
discharges must be digested according to the nitric
acid procedure (described later).

Tin seems to cause problems due to volatility,
especially of SnCl2 and organotin compounds.  Normal
digestion procedures will cause the loss of tin.
Therefore, consideration should be given to using
the nitric acid digestion procedure that is used for
mercury for tin also.

Arsenic - The sulfuric acid-nitric acid procedure
described on page 420 of Standard Methods (39) should
be used.  Care should be taken to maintain oxidizing
conditions during digestion to prevent formation of
arsine and loss of arsenic (137).  After digestion,
all the nitric acid must be removed to prevent sub-
sequent interference with arsine evolution.  To insure
the absence of nitric acid, ammonium oxalate solution
may be added and heat applied until SO? fumes are
evolved.

The recommended anlaytical procedures are as follows:

     1.  After digestion, analysis for lead, zinc,
         copper, chromium, cadmium, nickel, and
         tin should be performed using atomic
         absorption spectrophotometry.  The instru-
         ment should be operated according to the
         manufacturers instructions.

     2.  After digestion, the analysis for mercury
         should be performed by flameless atomic
         absorption methods as described on page 121
         of WQO Methods  1971
         Analysis for arsenic  should be performed
         using the method by which arsenic  is reduced
         to arsine and  reacted with silver
         diethyldithiocarbamate  (SDDC) .  This method
         is described on page  62 of Standard Methods
         (39).
                         187

-------
When analyzing for hexavalent chromium the sample should not be digested.
The sample should beanalyzedas soon as possible after sampling to
minimize losses due to adsorption on container walls (use of new bottles
is recommended).  Analysis should be performed by the diphenylcarba-
zide method described on page 156 of Standard Methods (39).

Other Metals - For the analysis of aluminum, calcium, iron, magnesium,
manganese, potassium, silver, sodium, antimony, barium, beryllium,
molybdenum, selenium, thallium, titanium, and vanadium, digestion using
the nitric acid-hydrochloric acid procedure is recommended.  Special care
should be given to prevent possible loss due to volatilization when
digesting samples for Fe,  Mg, Mn, Mo and V analysis.  The digestates are
then analyzed by atomic absorption spectroscopy.   Boron is analyzed using
the curcumin method outlined in Standard Methods  p.  69, Method 107A (39).
Care must be taken to store the sample in polyethylene or boron free
glassware and to have thoroughly clean glassware  throughout the analysis.
Samples for analysis of ferrous ion must be specially prepared immediately
ater sampling.  The pH should be adjusted to > 6  to precipitate ferric
iron and filtered as soon  as possible to avoid oxidation.  The filtrate
pH should be reduced to <  3 and analyzed for iron by atomic absorption.

Nitric acid-hydrochloric acid digestJon procedure - Choose a volume of
sample that will yield the appropriate value of metals.  Place this
aliquot of well mixed sample into a beaker and add 3 nil of concentrated
nitric acid.  Place the beaker on a hot plate and evaporate to dryness
making certain the sample  does not boil.  Cool the beaker and add
another 3 ml portion of concnetrated HNOo.  Cover the beaker with a
watch glass and return to  the hot plate.  Increase the temperature of
the hot plate so a gentle  reflux action occurs.  Continue heating,
adding additional acid as  necessary until the degestion is complete,
generally indicated by a light colored residue.  Add sufficient dis-
tilled 1:1  HC1 and again warm the beaker to dissolve the residue.
Wash down the beaker walls and wash glass with distilled water and
filter sample to remove silicates and other insoluble material that
could clog the atomizer.  Adjust the volume to a  predetermined value
based on expected metal  concentrations.  This will  yield values of
the total metal concentration.

Nitric acid digestion procedure - A sample of suitable volume is placed
in a 250 ml  round bottom flask and 10 ml of concentrated nitric acid
added.  The flask is then  connected to a reflux condenser (about 60 cm
in length)  and heated with a heating mantle causing the acid to reflux
gently.  Continue heating  for 2 hours and cool the mixture.  Wash down
the column with about 60-70 ml  of distilled water.   Filter the sample
through Whatman No. 42 paper to remove insoluble  material and make
filtrate up to 100 ml with distilled water.  Take a suitable aliquot
and analyze.
                                   188

-------
PESTICIDES AND POLYCHLORIMATED BIPHENYLS

The controversy over the use of pesticides continues as the pros and cons
of pesticide effects are debated in courts and agencies.  However, their
use Is widespread and does merit special attention;  Pesticides can be
classified into several types of which the organochlorine compounds are
most dangerous due In part to their extremely long half life.  This fact
has caused many pesticide users to change to less objectionable types of
materials, such as organophosphates, in their present applications.
However, it is still necessary to analyze for the organochlorines since
they are very persistent and the toxic effects are important.

Although polychlorinated biphenyls (PCB's) are not pesticides they react
similarly to pesticides and have the same type of persistence.   These
chemicals accumulate in the fatty tissues of animals and have been found
in all parts of the earth.  PCB's have a wide range of applications from
plasttcizers to hydraulic brake fluid and the amount of these in the
environment can be quite large.  Therefore analysis for these materials
is also cri tleal.

Pesticide and PCB analyses can be broken into three parts:

      1.  Extraction
      2.  Cleanup
      3.  Measurement

The extraction  is done with a hexane-acetone mixture and all hexane
soluble materials are removed.  The cleanup steps remove materials that
may interfere with the final analysis or cause questionable results.
The measurement is done with electron capture gas chromatography and  this
step is very sensitive to minute quantities of interfering compounds.
Pesticide determination on storm flows is complicated due to the wide
range of materials present in the sample and the large percentage of orga-
nic solids.  The general analytical procedures are outlined below.  Due
to the nature of storm generated discharges, a special  procedure has  been
developed and a brief discussion of the modifications is presented here.
The specific procedure is outlined later.

The extraction step is basically the same as a routine analysis, however,
the large amount of organic solids usually prevent the separate handling
of the solids and liquid portion.  The solid material is separated by
centrifugation.  The solids are dried with anhydrous sodium sulfate to
avoid codtstlllation with water that occurs with these organic  compounds
during air drying.  They are extracted from a glass thimble using a
Soxhlet extraction apparatus.  The liquid portion is extracted  in a sep-
arate funnel to Insure good contact between the phases.

The cleanup is divided into several steps due to the difficulty in removing
the interfering substances from the sample.  Because of the high concentra-
                                    189

-------
tion of these  interferences in storm generated discharges the  four
cleanup steps  are recommended.  Special care  is taken so the elements
are pure and that the concentrations are accurate.  The micro-scale
alkali cleanup/confirming procedure is a destructive step, so  only
portions of the sample are treated in this manner.  If sulfur  inter-
ference  is indicated, treatment with elemental copper is recommended,
however, alkali treatment also removes the sulfur.

Finally, the analysis is done on two gas chromatograph columns of dif-
ferent polarity to allow cross checks.  The peak post ions are  compared
to peaks produced by known materials for identification.   The  peak area
is used to determine the quantity of material measured.

Recommended Analyses

Because of the wide variability of pesticides in use, the periodic nature
of their application depending upon season and nature of the drainage
area, and the complexity of the laboratory analyses, no pesticides or
associated compounds are recommended for routine analysis.   However, it
is recommended that when evaluating the quality of a storm generated
discharge, a study of the drainage area should be made to determine the
likelihood of pesticide application (and the type) and if it is probable
that the storm flow may contain pesticides.  At least one discharge
should be analyzed to see if that pesticide is present.  Depending upon
this result a decision can be made as to whether more analyses are
needed.

For evaluating a treatment process, the same procedures should be used
as for evaluating a discharge.  In addition, if it is found that pes-
ticides are present in the treatment process, then the residual sludges
arising from these processes should also be analyzed.

Recommended Analytical Procedures

Scope and Application - This method describes the extraction and isolation
of organochlorine pesticides and certain PCB mixtures from storm sewer
discharges and combined sewer overflows, influents and discharges from
treatment processes.  The cleanup procedures permit the analyst to
eliminate those interferences which may be encountered and  allows for
separation of analogs ofArochlor #125^, #1260, and #1262  from organo-
chlorine pesticides.

PCB's and organochlorine pesticides are coextracted either  by  liquid-
liquid extraction or for samples of high solids by mixing with anhydrous
sodium sulfate and soxhlet extration.   A combination of the standard
Florisil  column cleanup and silicic acid column chromatography are employed
to separate PCB's from organochlorine pesticides (138).  Identification is
made with a gas chromatograph equipped with an electron capture detector
through the use of two or more unlike columns.  Further configuration by
chemical  modification using a micro scale alkali treatment  (139) is
recommended.
                                    190

-------
Sampling Preservation Volume and Container - As with all analyses, the
more quickly the test can be performed, the more accurate the results.
Pesticide samples must be taken  in glass containers and preserved by
refrigeration at 4°C.  Sludge samples may also be preserved by freezing.

Interferences - All glassware, solvents, reagents, and sampling hardware
must be demonstrated to be free of interferences under the conditions of
analysis.  It Is recommended that all glassware be fired at 230°C as
described by Lamberton et al (1^0).  Organochlorine pesticides and PCB's
are mutually interfering.  The silicic acid column will not separate
Arochlor #1221, #12^2, and #1248 completely from DDT and Its analogs.
(Early eluting peaks from the Aroclors may occur In the polar eluate.)
For this reason, the use of the chemical modification confirmation tech-
nique is recommended.

The sensitivity of this procedure  is predicated on the size of the aliquot
extracted, the organochlorine's  response on the EC detector and the back-
ground  interferences.  The cleanup procedure will eliminate most  of the
background interferences, however, great care should be exercised to
minimize organochlorine loss during cleanup.  Spiked samples are recom-
mended as a quality control check of organochlorine recovery.  As a
general rule, a 100 ml sample should yield quantifiable results for most
organochlorine pesticides with a sensitivity of 1 vg/1.

Apparatus -

      1.  Gas chromatograph equipped with recorder.
      2.  Detector, electron capture.
      3.  Gas chromatograph columns - two glass columns packed
          with nonpolar and semipolar adsorbents suitable for
          pesticide analysis.  Recommended column dimensions
          are 0.63 cm x 1.83 m (0.25 in. x 6 ft).  Suitable
          packings are 1.52 OV-17 and 1.952 OF-1 on 80-100 mesh
          Anakrom ABC (polar column) and 52 OV-210 on 80-100
          mesh Anakrom (semipolar columns).
      4.  500 ml Kuderna-Danish  glassware (Kontes K-570000).
      5.  Chromatographic column 400 x 22 mm  (Kontes K-420550,
          C-4) with adapter, hose connector type (Kontes
          K-185030).
      6.  Separating funnel 250 ml (Kontes K-633030).
      7.  Evaporative concentrator (Kontes K-5&9250).
      8.  Concentrator tube (Kontes K-570050) graduated in
          0.1 ml to  1 ml.
      9.  Separatory funnels  (125 ml, 1000 ml with teflon
          stopcocks).
                                   191

-------
     10.  Volumetric flask,  250 ml.
     11.  Florist1-PR Grade  (60-100  mesh)  prepared after the
          method of Hall  (141).
     12.  Silicic acid,Mallinckrodt 100 mesh "Especially Pre-
          pared for Chromatograph analysis by the Method of
          Ramsey and Patterson", see sample preparation later.

     13.  Glass wool - hexane  extracted.

     14.  Centrifuge tubes,  40 ml Pyrex.

     15.  Soxhlet extractor,  250 ml.
     16.  Magnetic stirrer with teflon control  bar, hexane
          extracted.
     17.  4 liter (1 gal)  sample bottles,  with  teflon caps.

     18.  Transfer pipette,  10 ml.
     19.  Celite 545, acid washed.
     20.  Air regulator.

Reagents, Solvents, and  Standards

      1.  Sodium chloride, ACS, saturated  solution.

      2.  Sodium sulfate,  ACS, granular anhydrous, conditioned
          for four hours  at  400°C.

      3.  Diethyl  ether  -  nanograde.

      4.  Hexane,  acetonitrile, methanol ,  methylene chloride,
          petroleum ether  (BP  30-60°C) -  pesticide grade.
      5.  Standards - appropriate organochlorine and arochlors
          for elements in  question.

CalJbration - Gas chromatograph conditions are  considered  acceptable when
response to heptachlor epoxide is 50% of full scale for <  1 ng injection
(full scale - 1 x 10~9 amp).   Detector response for quantitative work
must be in the demonstrated  linear range.   Standards are injected fre-
quently as a check on detector and column  stability.  For  quantitat ion,
noise should not exceed  2% of  full  scale.

Sample Preparation - Adjust  pH to near neutral.  If the solids content of
the sample is high (as with  sludges  and some influent samples) liquid-
liquid partition is not  possible due to emulsion formation.  Under these
conditions the sample aliquot  is centrifuged and the supernatant treated
as described below.  The  solids are  combined with anhydrous sodium sulfate
and extracted as described below also.
                                   192

-------
Extraction - Two methods of extraction may be employed depending on the
nature of the sample.  Unless the sample appears to be low  in solids and
organics (a well treated effluent sample for example)  it will be necessary
to separate the solids from the liquid and extract each separately.  The
extracts may then be combined and concentrated as a single  extract.

Liquid-liquid extraction is employed for samples of low solids and organic
content.  Place an aliquot of the sample in a one liter separatory funnel
and make the volume up to 500 ml using distilled water.  Add 30 ml of 15?
methylene chloride in hexane  (V:V) and shake vigorously for two minutes.
Allow the phases to separate and drain the water layer into a clean
Erlenmeyer flask.  Pass the organic layer through a 7.6-10.2 cm (3-4")
column of anhydrous sodium sulfate and collect in a 500 ml  flask.   Return
the water phase to the separatory funnel and rinse the Erlenmeyer with a
second 30 ml volume of solvent.  Add the solvent to the separatory funnel
and complete the extraction procedure.  The water phase should be extracted
with three 30 ml aliquots of solvent.  Concentrate the extract on a water
bath to 5 ml.  Note that if an emulsion is formed between the water and
solvent phases  it will be necessary to remove the solids as follows.

Samples of high solids content should be centrifuged in clean, hexane
washed glass centrifuge tubes.  Decant the supernatant Into a 1  liter
funnel and extract the pesticides as outlined above.  Remove as much of
the centrifuge cake as possible with a glass rod and combine  it with
hexane washed anhydrous sodium sulfate in a large mortar and pestTe.
Work the sample to free flowing dry state by continuously adding small
amounts of anhydrous sodium sulfate.  Add a small amount of sodium sulfate
to the centrifuge tube to dry any remaining sample and aid  In removing it.
Combine  all the dried sample and pour it into a glass Soxhlet extraction
thimble.  Place the filled thimble in a Soxhlet apparatus and wash the
mortar and pestle and centrifuge tube with three washings of 1:1 hexane,
acetone.  Carefully add the washings to the extraction apparatus by pouring
them through the filled extraction thimble.  Extract the sample for 6 to
8 hours.  Take  the extract just to dryness on a water bath  in a Kuderna-
Danish assembly, cool and wash the Kuderna-Danish assembly with hexane and
adjust sample volume to 5 ml.

Inject the concentrate in the gas chromatograph and determine:

      1.  If only organochlorIne pesticides are present.

      2.  If only PCB's are present.

      3.  Combination of 1 and 2.
      4.  If elemental sulfur Is present.
      5.  If response it too complex to determine 1, 2, or  3-

      6.  If 1, determine organochlorine pesticides according
          to Reference 142.
      7.  If 2, determine PCB's according to Reference 143,
          Section llff.
                                    193

-------
      8.   If 3, compare peaks obtained to standard Arochlors
           and determine which Arochlors are present.   If
           Arochlor peaks are analogs of #125^ and #1260,  the
           PCB's may be separated from DDT and its analogs
           by the combination of Florist 1 column and silicic
           acid column technique described later.  If other
           Arochlor analogs are present further confirmation
           with the microalkali technique in Appendix I I I of
           Reference 1^3 should be employed.

      9.   If 4, remove sulfur according to the procedure
           discussed later.

      10.   If 5» see the discussion below.  The selection of
           the various cleanup techniques can be simplified
           if a background knowledge of the storm generated
           discharge sample is available.

Cleanup and Separation Procedures -

Acetonltrile partition for removal  of fats and oils -  (Note:  Not all
pesticides are quantitatively recovered by this procedure.  Efficiency
of partitioning for pesticides of interest should be demonstrated.)
Transfer the concentrated extract to a 125 ml  separatory funnel using
hexane rinses to ensure complete transfer.   Final volume should be 15 ml.
Extract the sample with four 30 ml  portions of hexane saturated acetoni-
trile by shaking vigorously for one minute.   Combine and transfer the
acetonitrile phases to a one liter  separatory funnel  and add 650 ml of
distilled water.   Add  kO ml of saturated sodium chloride solution, mix
thoroughly and extract with two 100 ml  portions of hexane.  Combine the
hexane extracts in a one liter separatory funnel  and  wash with two 100
ml portions of water.   Discard the  water layer,  pass  the hexane layer
through a 7.6 to 10.2  cm (3-A In.)  anhydrous sodium sulfate column into
a Kuderna-Danish flask and rinse the funnel  and column with three 10 ml
portions of hexane.  Concentrate the hexane extracts  to 6-10 ml and
analyze by gas chromatography if further  cleanup is  not required.

Sulfur inte r fere nee -  Elemental  sulfur is encountered in most sediment
samples, marine algae  and some industrial  wastes.  The solubility of
sulfur in various  solvents is very  similar to the organochlorine and
organophosphate pesticides; therefore,  the sulfur interference follows
along with the pesticides through the normal  extraction and cleanup
techniques.  The  sulfur will  be quite evident in gas  chromatograms
obtained from electron capture detectors, flame photometric detectors
operated in the sulfur or phosphorus mode,  and Coulson electrolytic
conductivity detectors.   If the gas chromatograph is  operated at the
normal conditions  for  pesticide analysis, the sulfur  interference can
completely mas'k the region from the solvent  peak through pp'DDE.
                                  194

-------
This technique eliminates  sulfur  by  the  formation of  copper  sulfide
on the surface of  the copper.  There  are  two  critical  steps  that must
be followed to remove all  the  sulfur:   1)  the copper  must be highly
reactive; all oxides must  be removed  so  that  the copper has  a shiny,
bright appearance; and 2)  the  sample  extract  must be  vigorously agitated
with the  reactive  copper for at least one  minute.

It will probably be necessary  to  treat both the 6% and 15% Florisil
eluates with copper if sulfur  crystallizes out upon concentration of the
6% eluate.

Certain pesticides will also be degraded  by this technique, such as the
organophosphates,  chlorobenzilate and heptachlor, as  shown in Table
VI-12.
                   Table VI-12.  EFFECT OF EXPOSURE
                    OF PESTICIDES TO MERCURY AND COPPER

                                     Percentage recovery based on mean
                                             of duplicate tests	
     	Compound	              Mercury              Copper

     BHC                                 81.2                 98.1
     Lindane                             75.7                 9*».8
     Heptachlor                          39.8                  5.4
     Aldrin                              95.5                 83-3
     Hept. Epoxide                       69.1                 96.6
     p.p'-DDE                            92.1                102.9
     Dieldrin                            79.1                 94.9
     Endrin                              90.8                 89.3
     DDT                                 79.8                 85.1
     Chiorobenzilate                      7.1                  0
     Arochlor 1254                       97.1                104.3
     Malathion, diazlnon                  0                    0
     Parathion, Ethion
     Tr!thion
          Note:  If the microalkali dehydrochlorination procedure is used,
                 elemental sulfur  is removed  (from Reference 139 ).
However, these pesticides are not likely to be found in routine sediment
or sludge samples because they are readily degraded in the aquatic
envi ronment.

If the presence of sulfur is indicated by an exploratory injection from
the final extract concentrate (usually 5 ml) into the gas chromatograph,
                                   195

-------
proceed with removal as follows:

      1.  Under a nitrogen stream at ambient temperature,
          concentrate the extract in the concentrator tube
          to exactly 1.0 ml.

      2.  If the sulfur concentration is such that crystal-
          lization occurs, carefully transfer, by syringe,
          500 yl of the supernatant extract (or a lesser
          volume if sulfur deposit is too heavy) into a
          glass-stoppered, 12 ml graduated, conical  centri-
          fuge tube.  Add 500 yl of iso-octane.

      3.  Add about 2 mg of bright copper powder, stopper
          and mix vigorously  1  minute on a Vortex mixer.
          (NOTE:  The copper  powder as received from the
          supplier must be treated for removal of surface
          oxides with 6]£ HNO,.   After about 30 seconds of
          exposure, decant off acid, rinse several  times
          with distilled water and finally with acetone.
          Dry under a nitrogen stream.)

      *».  Carefully  transfer 500 yl of  the supernatant-
          treated extract into a 10 ml graduatedvevaporator
          concentrator tube.   An exploratory'Injection into
          the gas chromatograph at this  point will  provide
          information as to whether further quantitative
          dilution of the extract is required.  NOTE:   If
          the volume transfers given above are followed, a
          final extract volume of 1.0 ml will  be of equal
          sample concentration to a 4 ml concentrate of the
          Florisil cleanup fraction.

Florisil column cleanup - Place a charge of activated  Florisil (the
weight of the charge is determined by its Laurie Acid  Value - see
Reference 141) in the Chromaflex column  and settle by  gentle tapping.
Add a 1  cm layer of anhydrous sodium sulfate and pass  50-60 ml of
petroleum ether through the column.  When the petroleum ether is about
5 mm from the sodium sulfate  surface, transfer the sample extract by a
long stem funnel with petroleum ether washings to the  column and elute
with the following mixed ethers at 5 ml/minute.  (NOTE:  For both
column chromatography procedures the elution rate is important.  To
quickly adjust this rate the  lower part  of a broken 25 ml burette
equipped with teflon stopcock placed between the chromaflex column and
the receiving vessel is most  useful  in making repetitive(flow adjustments
without losing eluate.  See Figure V|-i>).  Collect each eTuate in a 500
ml Kuderna-Danish flask and concentrate  to 5 ml.
                                   196

-------
                                       Chromaflex column
                                    Lower section of a
                                    25 ml burette with
                                    teflon stopcock for
                                    adjusting elution
                                    rate
                                       Receiving vessel
              Figure Vl-^t.   Drawing of a recommended procedure
                   for obtaining proper elution rates
First elution (6% eluate) add 200 ml of 6%
ether (V/V).   Second elution  (15% eluate):
ether in petroleum ether.  Most pesticides
these eluates.  Refer to Reference  1^2 for
            ethyl ether in petroleum
             Add 200 ml of 15% ethyl
            of interest will  be in
            more details.   The pesticides
that may be expected in each of the eluates are listed below:
      Aldrin
      BHC
      Chlordane
   6% Eluate

Heptachlor
Heptachlor epoxide
Lindane
Strobane
Toxaphene
Tr i f1uralin
                                   197

-------
            ODD              Methoxychlor                   PCB's
            DDE              Mi rex
            DDT              Pentachloronitrobenzene


                                  15%  Eluate

            Endosulfan  I                                    Dichloran
            Endrin                                          Phtholate
            Dieldrin                                          esters


            Concentrate  the  eluates  and analyze  by  GLC.

Sijjcic Acid Column separation  procedure -

      1.   Silicic Acid Preparation:  Celite  5^5  must be  oven
          dried and free of  electron capturing  substances
          (acid washed).   Silicic acid:  oven dry  for a  mini-
          mum of seven hours at  130°C  to remove  water.   Cool
          the silicic acid,  weigh  into a glass  stoppered
          bottle and add 3%  water.   Stopper  bottle  and  shake
          well. Allow  15 hours  for  equilibrium  to  occur.
          Determine separation  achieved as described below
          by loading kO  yg of Arochlor #125^ and pp'DDE  in
          hexane on the  column.   Inadequate  separation  will
          mean readjustment  of  the water content of the  silicic
          acid in recommended increments of  0.5%.   More  water is
          required when  the  PCB  elutes in  the polar solvent with
          pp'DDE; less water when pp'DDE elutes  in  the  petroleum
          ether portion.   Standardization  is required for  each
          new lot of silicic acid purchased. Once  a batch of
          silicic acid  is hydrated activity  remains for  about
          5  days.

      2.   Column Preparation:   Weigh 5 g of  Celite  and  20  g of
          silicic acid and combine  in  a 250  ml  beaker.   Immedi-
          ately slurry with  80  ml of petroleum  ether.  Transfer
          the slurry to  the  chromatographic  column, keeping the
          stopcock open.   Stir  the slurry  in the column  to remove
          air bubbles,  then  apply air  pressure  to  force  the petro-
          leum ether through the column.   Do not allow  the column
          to crack or go dry and close the stopcock when air  pres-
          sure is not being  applied.   Stop the  flow when the
          petroleum ether level  is 3 mm above the  surface  of  the
          silicic acid.   The absorbant at  this  point should be
          firm and not  change shape  if tapped.
                                 198

-------
Elution Patterns:  Large amounts of PCB's or pes-
ticides placed on the column will result  in incom-
plete separation.  The extracted sample placed on
the_column should contain no polar solvents and
be < 5 ml in volume.  Place a 250 ml volumetric
flask beneath the column and carefully add a suit-
able aliquot of the 6% Florisil eluate, taking care
not to distrub the surface of the silicic acid.  A
long stem funnel is useful for this purpose.  Apply
slight air pressure until the solvent level is about
3 mm from the surface of the silicic acid.  Carefully
position the 250 ml separatory funnel containing 250
ml of petroleum ether on the column and allow the
petroleum ether to run down the sides of  the column
until the space above the silicic acid is one half
full.  Apply air pressure and adjust the  flow rate
to 5 ml/minute.  When exactly 250 ml are  collected re-
place the volumetric flush with 500 ml Kuderna-Danish
flask and elute at 5 ml/minute with 200 ml of
methylene chloride, hexane and acetonitrile (90:19:1,
V/V) to recover the pesticides.  Quantitatively trans-
fer the petroleum ether eluate containing the PCB's to
a 500 ml K-D and concentrate both eluates to 5 ml .
Analyze by gas chromatography .  NOTE:  The separation
between the PCB's and pp'DDE is very narrow; great
care should be exercised  in adjusting the elution
flow rate and volume of the petroleum ether portion.

                  Petroleum Ether Eluate

       Aldr in

       Arochlors       #1248a         #1252a
                       #1221a         #125*ta
                       #1252a         #1260a
                       #1258a         #!262a

       Hexach 1 orabenzene

 Polar Eluate (Acetonitrile, Methylene Chlroide,  Hexane)

       Arochlors   #1221a             Endrin
                   #12^2              Heptachlor
                                      Heptachlor epoxide
       BHC                            Lindane
       pp'DDE                         Toxaphene
       pp'DDT
       pp'DDD

 a.   These Arochlors divide between the two eluates.   The
     earliest evaluating GLC peaks may occur in the polar
     el uate.

                      199

-------
 Confirmation Techniques

 Qualitative confirmation  -  By comparing relative retention times of the
 constituents on two or more unlike columns  as  a minimum criteria for
 identification  after appropriate  cleanup  and column chromatography, it
 is felt that satisfactory qualitative  confirmation  can be performed.

 If an Arochlor  analog which does  not completely occur in the petroleum
 ether eluate is suspected,  the  alkali-dechlorination procedure is
 strongly  recommended (see Young et  al  (139))-   In any event  such con-
 firmation  techniques add greatly  to the reliability of the residue analy-
 sis in the  absence  of more  sophisticated  mass  spectroscopy instrumentation.

 Quan t i t a t i ve Pete rmi na t i on  -  For  organochlorine pesticides,  use  Reference
 142,  Section II.  For PCB's  use Reference 143,  Section II.   Report results
 in micrograms/1  without correction  for  recovery data.   For sludge samples
 it may  be necessary  to adjust for a density factor.

 RECOMMENDATION  OF OTHER CHARACTERISTICS

 Storm generated  discharges may contain  a number  of  contaminants  that are
 not often measured  in  most domestic wastewaters, but  may  be  important  in
 the characterization  and treatment of storm generated  discharges.   Many
 of these characteristics have been  indicated in  the  literature (7)(30)
 (144)(145).  A  recent  review of possible combined sewer overflow conta-
 minants has  been made  by Field and Tafuri (146).  Table VI-13 shows  some
 of  the  contaminants  that can be expected  in storm generated discharges.

 Surveys have been conducted  to determine how ordinances and street
 cleaning practices can best  minimize the contaminants  that build  up  in
 streets between  rainfalls (2)(l45).  Street surface contaminants  have
 been  characterized and quantified  in different areas of the country  (2).
 The largest portion of street surface contaminants consisted of  dust and
 dirt  (fine solids),  however, litter and other contaminants were  found
 from a number of other sources.  Many of these  contaminants were subjec-
 tive  in nature and  are not easily  quantified by available testing proce-
dures.  The proper  characterization of  the subjective waste contaminants
can be handled by use of a check off sheet with general categories as
shown in the tables  in Figure VI-5.  Also, entries should be made perio-
dically in a log book to describe  these characteristics in the operator's
own manner as he observes it.

The general  appearance of storm generated discharges will often  be very
 useful  in determining treatment plant problems.  Changes  in wastewater
color and extensive foaming  may indicate the presence of  large quantities
of soaps, detergents or other materials that reduce the surface  tension.
These types of materials can cause significant problems  in the physical
and chemical treatment of combined sewer overflows by  inhibiting  settling
and chemical coagulation.  An increase  in soluble and  total phosphorus
may also occur which will  increase thei nutrient  load to the receiving
 body of water.  Gasoline, solvents, and other industrial  chemicals that
                                   200

-------
Table VI-13.  POLLUTANT CHARACTERISTICS OF URBAN RUNOFF


  1.  Color  causing materials
  2.  Turbidity
  3.  Foam causing materials
  4.  Floating material
  5.  Street litter debris
  6.  Material from street or pavement surface
  7.  Debris from vacant  lands
  8.   Ice control chemicals
  9.  Pest control chemicals
 10.  Fertilizers
 11.  Droppings  from  animal or  bird sources
 12.  Lawn or garden  Utter
 13.  Household  or commercial refuse
 ]k.  Air deposited materials from precipitation
 15.  Twigs  and  leaves
 16.  Paper
 17.  Plastic materials
 18.  Tire and vehicular  exhaust  residue
 19.  Heavy  metals
 20.  Hazardous  material  spills
 21.  Other  large, heavy  items
                       201

-------
Time
Normal
   COLOR CHARACTERIZATION

Blackish'  Whlteish  Reddish
Foaming    Other
                        ODOR CHARACTERIZATION

                            Rotten  Egg      Hydrocarbon
Time Normal Septic H2$ Gasoline, Solvents

























Other





                       DEBRIS CHARACTERIZATION
Time
  Construction material,
   lumber,  bricks, etc.
                       Rags
         Other
             Figure  VI-5.   Subjective  check-off  sheet  for
                characterizing  various contaminants
                                202

-------
can cause ft res or explosions  in enclosed areas usually produce strong
distinctive odors that are readily detected by treatment plant operating
personnel.  Debris can plug  in.lets and damage treatment facilities.
Materials toxic to humans, fish, or aquatic flora and fauna may be pre-
sent in thes discharges.  A discussion of physical and other unusual
parameters that are  important  in characterizing storm generated dis-
charges or  in the treatment of these waters follows.

Potential Physical Parameters  in Storm Flow Characterization

Asbestos - The  importance of asbestos in storm generated discharges has not
been defined.   The primary potential source of asbestos is street and high-
way runoff where deposits may  be caused by asbestos brake shoes of vehicles.
In some areas of the country industries using asbestos as raw products may
contribute  to the presence of  asbestos in storm generated discharges.

The analysis and detection of asbestos in storm flow are in the formative
stages of development.  Many studies from manufacturing plants producing
asbestos products use the suspended solids test described in Standard
Methods (39) in characterizing their wastewaters.   The suspended solids
test is easily determined but  it does not differentiate between asbestos
or other fibers and suspended material  that can be present in a storm
generated flow; hence it is  inadequate for this purpose.  A dispersion
staining technique that relies on microscopic observation of fibers de-
posited on a filter has been developed (1^7).  It was possible to
differentiate between asbestos and other fibers by suspending the non-
filterable particles from the  filter in a liquid with a refraction index
that matches that of the asbestos fiber at a wavelength of 550 nanometers.
In the same reference a procedure for determining the total  fibers present
in a water sample by counting  fibers by a phase-contrast microscope was
described.  It was suggested that to determine the total number of asbestos-
like fibers present  in the water sample, a microscope with a device for
switching objectives could be employed.   Thus, the same field of view
could be observed using alternately the dispersion staining and then the
phase-constrast objectives.  In other studies identification and quanti-
fication has been accomplished with an electron microscope (148) (1^9) (150),
using  the electron microscope for measurement of asbestos fibers.  It was
reported that despite variations in geographical origin, asbestos
(chrysotlles) fibers have a narrow size distribution.  Ultrasonic treatment
was used to separate the asbestos material into individual fibers (150).

Since there are a large number of different types of suspended materials
in storm flows  there are many  interferences to asbestos identification
analytical techniques.  Ashing at ^50°C has been used to destroy organic
materials such as plant and bacterial debris that would interfere (1^8) .
It was shown that the addition of dibutylphthalate to the suspended
solids sample prior to ashing would reduce loss of asbestos due to ex-
plosive combustion of the membrane filters (150).  After ashing, the
samples were homogenized using an ultrasonic water  bath to resuspend the
remaining materials and to separate the asbestos into individual fibers.
                                    203

-------
X-ray diffraction has also been used for identification and quantitative
determination of different types of asbestos (151).   Fifty to 100 pg/1
of asbestos could be detected even when interferences were present.
The amount of asbestos in the filter must  remain small  in order to
reduce errors due to specimen thickening.   Studies with air samples have
shown the x-ray diffraction technique capable of detecting 0.1 pa/m3
of asbestos; the electron microscopic method can detect 0.1 ng/rrr.   Thus
the electron microscopic technique is about 1000 times  more sensitive.
Other methods of detecting asbestos include infrared spectrometry and
elemental analysis
Five hours are required for the analysis of each sample using the
electron microscopic technique.  However, actual  time required is longer
because of extensive filtration and ultrasonic resuspension techniques
required.  The routine testing  for asbestos is not recommended for storm
flows.   Periodic testing for asbestos  In storm flows should be conducted
in municipal  areas where there  are asbestos factories or industries
machining asbestos products.  In most  cases it will  be necessary to send
the samples to a regional  laboratory where x-ray diffraction or an
electron microscope is available.
Asphalt and Road Materials  - Street  surfaces have been found to be a
source of potential  contaminants in  storm generated discharges (152).
Included in these materials are asphalts and portland cement, their
various decomposition products and aggregate materials.   There may
also be some contribution of small amounts of road marking paints, crack
fillers and expansion joint compounds.   All  asphalt streets and those
streets with poor surface conditions were found to have a substantially
higher amount of loose particulate materials that could  end up in storm
generated discharges than all  concrete  streets and streets in good
condition (2).  Climate also was a major factor in the presence of
contaminants from streets due to freeze-thaw cycling and the use of
studded tires.  Leaks and spills of  fuels or oils hastened the
degradation of asphaltic pavement and resulted In a higher amount of
pollutants appearing in storm generated discharges.

Specific identification of  asphalt  is complicated by Its complex makeup
and the fact that many asphaltic functional  groups form intermolecular
complexes.  Infrared spectrometry has been used to identify key
asphaltic functional groups that have been Isolated using selective
chemical reagents and solvents  (150-   However,  no analytical procedure
has been developed for detection of  asphalt from different areas of the
country.  In stormwater studies the  presence of asphalt will be
included in other more general  analytical tests.   Small  particulate
materials from road  surfaces may enter  a large bore plpet and be
measured in the suspended solids determination.  In some cases the
presence of larger particles from road  and other sources will cause
large variations and erratic results in the suspended solids analysis.
These variations are caused by the  nonuniform presence of heavy
                                   204

-------
particles and the difficulty of obtaining a representative sampling of
these types of materials.

Asphaltic materials will also be measured to some extent where solvent
extractions are made to determine grease and oil.  While the specific
determination of asphaltic and other road materials will not be deter-
mined by the above tests, the general pollutional nature will  be
determined when suspended solids, grease and oil, and to some extent
when oxygen demand tests are conducted.  Since specific tests for all
types of asphaltic materials are not readily available and since the
asphaltic materials are probably quantified by the other analyses, the
routine analysis for asphalt is not recommended.
Color - Natural color in storm generated discharges will  cause these
wastewaters to exhibit different colors in different areas of the
country.  Natural color can be due to tannins, humic acid, humates from
the decomposition of 1ignin (48), iron, colloidal  clay and  other
inorganic materials.  Weeds, vegetable materials and industrial  wastes
may also contribute to the color of a wastewater.  Descriptions  of
materials collected on screens in sewers have included brownish  pulp
appearance, pinkish-gray appearance and light brown appearance (30).
Intentional and accidental industrial waste spills can have significant
effects on storm generated discharge characteristics.  Paint spills have
colored wastewaters and material  from screening operations have  brilliant
colors.  Oil dumps have occurred that produced a black color in  storm
generated discharges  (27).  The normal color of storm generated
discharges in the midwest has been found by the authors to vary  through-
out a storm and to depend, to some extent, on antecedent  rainfall.  The
first flush is usually black and gradually changes to a yellow-brown
color as the storm progresses.  After an extended period,  the water may
become relatively clear with very little color.  Combined  sewer
overflows at Racine, Wisconsin have contained a white, milk-like
color possibly due to soaps or detergents, milk products,  or some other
unknown contaminants  (153).  A gray cast to these discharges has also
occurred at the same plant during the first flush.  A milky-white
combined sewer overflow at Kenosha, Wisconsin was traced  to a local
industry (154).

While the characterization of color  in storm flow is often difficult and
somehwat subjective, the presence of color may indicate a  need to
determine specific tests that are not normally conducted.   Analyses
such as calcium, oil and grease, or TOD may provide  important informa-
tion in characterizing the storm generated discharges, in  evaluating
their treatabi1ity, and the future treatment needs.

True color is designated as that color exhibited by a water sample after
the turbidity has been removed.  Apparent color  is the color observed
from the whole sample including the suspended materials.   Two methods
are described in Standard Methods (39) for accurately specifying the
                                   205

-------
color of a wastewater.   Color is  not usually determined in the
characterization of sewages but may be important where Industrial  wastes
are a component of the wastewater.   The determination of color as
described in Standard Methods (39)  is not recommended for storm flows.
However, a qualitative description  of storm flows can be useful  and
should be conducted by the operating personnel.   Special  analytical
tests should be conducted  when the  color of storm flows continue to
deviate from the normally  expected  color.   Surveys of local  industries
may show possible problems with  industrial  waste controls that can be
corrected.
Debris and Bulk Sol ids  -  No analytical  laboratory test can be used  for
debris and bulk solids  found in storm generated  discharges.   Usually,
specific items in this  category can  be readily identified.  Table \l\-\k
shows a list of different types of  items  that  have been found in inlets
to storm generated discharge treatment plants.  Many of the  larger  debris
have caused damage to protective cages of submersible pumps  (7)  or
eventually stopped flow to the treatment  plant as rags and other smaller
items collected around  the larger bulk items  (153)-   Many of these  items
listed in Table VI-H are from highway or sewer  construction.   However,
in some cases intentional  disposal of wastes  from street and lawn
maintenance have occurred as individual  home  owners  use the  street  and
nearby catch basins for disposal. Many reports  on storm generated
discharge studies have  not found debris or other large items to  be  a
problem; however, most  of these studies are  pilot scale where combined
sewer overflows or storm  sewer discharges are pumped or drawn from  an
existing sewer in such  a  manner as to screen  out the larger  objects.
Where wetwells, sumps or  the entire  waste has  been treated,  debris  and
large items similar to  those in Table VI-14  have been ecountered.

It is expected that bar screens will  be an important design  consideration
in stormwater treatment systems because of these large objects.   A  log
book should be maintained to record  the presence of  all debris and
large objects found in  storm flows.   The  log  should  include  a description
of the effects of the objects and any damage  they may have caused.
Attempts should be made in the surrounding community to minimize the
loss of these items to  the sewer where possible.


Density/Specific Gravity  - Very little meaningful information can be
obtained from density measurements of storm  flow as  a whole  because of
the extremely high percentage of water contents.  Although the specific
gravity of the particles  in a wastewater  is  the  most significant
variable in the treatability of these materials, direct determination
is not necessary.  Instead, the settleability (or flotability) of the
solids in a storm generated discharge is  better  qualified by the
routine bench scale settling and flotation tests conducted in the
laboratory.  However, the density of the  sol ids  themselves has been
discussed and may be important in solids  handling and disposal (155).
                                  206

-------
Table VI-14.  DEBRIS  AND  BULK SOLIDS  FOUND  IN STORMWATER

1.
2.
3.
4.
5.
6.
7-
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19-
20.
Item
Droppings from animals, birds or humans
Lawn or garden litter
Remnants from household or commercial refuse
Twi gs
Cloth and rags
Plastic material
String
Tires
Concrete slabs
Steel drums
Mattresses
Automobile radiator
Chai ns
Wheel borrow
2V x 10' PVC pipe
4" x 50' long fine hose
Building bricks, cement
Logs, 18" x 5'
Street barricade with flashers
Rodents
Reference
30
30
30
144
144
144
144
145
145
145
145
145
145
153
153
153
153
153
153
153
                            207

-------
The density of clay is  about  0.96  kg/1  (60 Ib/cu  ft)  and for sand,  about
1 77 kg/1  (110 Ib/cu ft);  a  single measurement  of combined sewer
overflow solids measured  13.6 kg/1  (85  Ib/cu  ft).  The sludge volume
index (SVI) defined in  Standard  Methods (39)  gives the volume of settled
sludge (ml) occupied by one  gram of solids after  settling for 30 minutes.
This is the inverse specific  gravity which is equal  in magnitude to
density in the metric system  of  units.   Hence the density of activated
sludge would be l/(SVl) x  1  g/cc (62.4  Ib/cu  ft).  A  commonly accepted
SVI value for well  settling  sludge is 100  ml/gm which is equal  to a
sludge density of 0.1 g/cc (6.24 Ib/cu  ft).  The parameter, sludge
density index (SDI) has been  defined in Standard  Methods (39) as
SDI = 100 x I/(SVI), thus  SDI  measurements would  be equivalent  to
100 x the density of the  solids.  SVI or SDI  measurements are usually
made for biological treatment systems and  would be useful  in character-
izing biological  plants treating storm  generated  discharges such as the
contact stabilization plant  at Kenosha, Wisconsin (17).  Where biological
treatment systems are used to treat storm  wastewaters one of the above
tests should be used.  However,  no other density  parameters are
recommended for routine analyses.
Oils and Grease - Oils and grease are commonly found in storm flows as
seen in Table VI-15.
      Table VI-15-   OILS AND GREASE IN STORM GENERATED WASTEWATERS
Concentration in
Range
2 - 99
1.3 - 54.4
10.9 - 48.4
31 - I40a
--
33 - 95
inlet to treatment plant, mg/1
Mean
--
12.3
27.6
81
40
--

Reference
29
30
35
156
157
157
a.  Pressure sewer system


Oils and grease may be introduced into storm flows by overflow from
municipal treatment plants, by various industrial  sources, by individual
home owners and of course by roadway runoff.  Filling stations have
disposed of used crank case oil  in floor drains and In catch basins (19).
Laundries, car wash operations,  packing houses, and refineries can be
expected to contribute significant amounts of oils and grease to the
wastewaters when they are treated by combined municipal-industrlal plants,
                                   208

-------
It Is not necessary to routinely measure oils and grease in most areas
of the country.  However, oils and grease should be determined once at
the beginning of every storm generated discharge study.  Where oils and
grease are likely to be present In storm discharges, as discussed above,
analyses should be made for both dry-weather flows and for each storm
event analyzed.  Oils and grease are defined based on the method used
for analysis; hence the method used should always be specified with
these data.  The soxhlet extraction method described In Standard Methods
(39) Section 209A, page A09 should be used for investigating the presence
of oils and grease or where high solids are present in a wastewater
sample.  If light oil or soluble oils are possible constituents of a
sample the liquid-liquid extraction procedure described in Section 137,
page 25** of Standard Methods should be used.
Odor - Although a tentative test exists for determining odor intensity
in wastewater samples  (39), odor should be considered a qualitative
parameter only.  Storm generated discharges may smell of gasoline, oil,
chemical solvents, fertilizer, herbicides, insecticides, mustlness,
hydrogen sulfide, septicity or not at all.  Strong odors of herbicides,
Insecticides, and oil or gasoline have been detected In operating combined
sewer overflow treatment plants  (153).  The odor should be recorded in
the operating log for the plant and specific analytical tests should be
considered depending on the odor.  When strong solvent or gasoline odors
are detected, safety precautions should be taken and the source Identified
if possible.  No specific threshold odor tests used  in water quality
analyses are recommended for storm generated discharges (39)-


Particle Size Distribution - Particle size Is  important In several
treatment processes used in treating storm generated discharges.  Although
the density of particles also plays a large role In  solids removal, the
particle size will affect the settleabi1ity or floatabiltty of different
solids.  Thus, settling tanks, air flotation,  screening, and filtration
may be affected by particle size.  Suspended solids  removal by screening
is dependent on the particle size distribution as showr^ in Figure VI-6
and Table VI-16.  These data are from combined sewer overflow studies and
show a significant Improvement  In removal where small opening screens
were used, thus indicating the presence of a large amount of smaller
particle sizes.  A wet sieve analysis developed by Envirex Inc. has
been used to characterize two storm generated  discharges as shown  in
Table VI-17.   In the discharge at Hawley Road  (Milwaukee, Wisconsin) 3k%
of the solids were smaller than  84 y and at Racine,  Wisconsin about 12%
were smaller than Ik y.  Over 90%  (by weight)  of the solids at Hawley
Road were smaller than 37 y.  While 10% of the solids at Hawley Road
were larger than 37 y, 22%  (by weight) suspended solids removal was
achieved with 63 y screens  (159).   Increased solids  removal was achieved
because of the buildup of solids on the screen which made the effective
pore size smal1er.
                                    209

-------
                                                         ©
                                            200
                                      MICRONS
        Figure VI-6.  Suspended sol i'ds removal  for different size
              screens from storm generated discharge projects


      Table VI-16.  SCREENING REMOVAL EFFICIENCIES FROM LITERATURE

Phi ladelphia
Mi Iwaukee,
Haw ley Road
Haw ley Road
Philadelphia
Mi Iwaukee
Haw ley Road
Haw ley Road
Haw ley Road
Portland
Mi Iwaukee
Haw ley Road
Portland
Mi Iwaukee
Haw ley Road
San Francisco
Screen size,
p
23

23
23
35

63
63
63
73

149
167

841
3175
SS removal,
percent
72

70
54
44

28
32
25
35

22
31

8
2a
References
158

26
26
158

26
26
26
48

26
48

26
19
a.  Solids removal  based on total  solids.
                                 210

-------
                Table VI-17-   WET SIEVE ANALYSIS  RESULTS
                     FROM TWO  COMBINED SEWER OVERFLOWS3
                    Hawley Road combined  sewer overflow
Screen opening,  y _ ~ _         %  retained,  by weight

     841                                                      2.0

     149                                                      2.6

      84                                                      1.0

      37                                                      4.3

     <37                                                     90.1


                       Racine combined  sewer  overflow
Screen opening, y
2380
841
297
149
74
<74
% retained, by weight
1.2
2.8
6.7
10.6
1.4
72.3
 a.  Reference 159
                                    211

-------
Table VI-18 shows the particle size distribution of solids from street
surfaces for several  cities.   These data indicate that the potential
solids sizes from street surfaces are greater than the sizes found using
the dry sieve analysis tn Tables  VI-16 and VI-17-   A large fraction of
the solids In Table VI-18 is  due  to sand.   These particles are usually
heavy and may settle out either in the sewers or somewhere near the
head end of the treatment plant.   Also, sand  may be removed by street
washing operations.  The data In  Table VI-18  were taken using the dry
sieving particle size distribution technique  which may also be the cause
of the difference tn average  particle size as compared with Table VI-17.

Particle size distribution studies are not recommended except for general
evaluation of storm generated discharges unless screening is to be used
as a means of treatment or the efficiency  of  street cleaning is being
studied.  A minimum of three  particle size distribution analyses evenly
spaced throughout the project should be made  for these treatment plant
designs.  In some cases more  analyses will be required.  The wet sieve
method has been found by the  authors to provide the most realistic values
with a minimum of interference; hence it is recommended.
£H_ - The pH of a wastewater sample is a measure of the hydrogen Ion
activity and is used to determine if a water  is acidic*   pH <7, or
basic,  pH >7.  pH is defined mathematically  as the log  1/(H+), thus as
the hydrogen ion concentration (or activity)  increases,  the pH decreases.
Water with a pH of 7.0 is designated a neutral  solution.   pH is a
measure of the instantaneous hydrogen activity  in contrast to alkalinity
and acidity measurements which measure the capacity of a  liquid sample to
resist a change in pH.  The parameter pH is important in  all  environmental
projects because of its effects on treatment  plant operations, corrosion
to pipes and equipment, disinfection, water softening and receiving water
quali ty.

The pH of natural  waters Is usually between A to 9; most  waters will have
a pH slightly greater than 7 (slightly basic) because of  the presence of
bicarbonate and carbonate alkalinity.  In most  storm generated discharges,
the pH Is also slightly basic.   However, the  pH may differ significantly
from either very low or very high values when stormwaters come In  contact
with Industrial wastes.  Where stormwater runoff passes  over acid  mine
drainage areas and other industrial waste disposal sites, pH may be a
problem and should be monitored.   Where biological treatment is to be
used, the pH should be between 6 and 9.5.  Chemical coagulation of
natural waters is usually optimum at pH values  between 5.0 and 6.5  (152).
However, the optimum pH for coagulation will  usually vary depending on
the alkalinity, hardness, and other water quality characteristics.
Chlorinatlon is also affected by pH because the type of  chlorine residual
existing in a water is highly dependent on the pH of the solution   (48).
At pH values below 7-5 hypochlorous acid predominates and greater
bacterial killing power tends to exist than at  pH levels  above 7-5 where
the hypochlorite ion predominates.
                                   212

-------
         Table VI-18.   PARTICLE  SIZE  DISTRIBUTION OF SOLIDS^
                             SELECTED CITY  COMPOSITES
Size range,
M
<4,800
2,000 - A, 800
840 - 2,000
246 - 840
104 - 246
43 - 104
30 - 43
14 - 30
4 - 14
<4
Sand, %
43 - 4,800
Silt, *,
4 - 43
Clay, %t

-------
Currently pH is seldom adjusted in combined sewer and stormwater treatment
plants; however, because of the effects described above, pH should be
measured regularly throughout discharge.   In cases where industry or other
factors cause the pH to vary greatly outside the range of 6.0 - 9-5, pH
control will usually be necessary.  As the treatment of storm generated
discharges becomes more sophisticated, pH adjustment may frequently im-
prove treatment efficiencies and reduce chemical dosages.  Recommended
pH procedures are described in Section 1M, page 276 of Standard Methods
(39).

Rubber - Potential presence of rubber in  storm generated discharges
exists from the wear of vehicle tires which can be washed from streets.
Rubber is of concern as a pollutant because of its high BOD, taste, and
odor.  Rubber found in modern day tires is usually made up of a variety
of polymeric materials and presents a unique problem in analysis.  Dif-
ferent procedures such as thin layer chromatography, pyrolyses-gas chroma-
tography, nuclear magnetic resonance spectroscopy, infrared and other
physical and chemical  techniques for analysis of rubber have been recently
reviewed (160).  Reproducibi 1 ity of results between laboratories has been
a major problem and it is usually necessary to run standards for each
analysis.  Quantitative evaluation of natural rubber, styrene-butadiene
rubber, and ethylene-propylene-terpolymer rubber has been achieved using
pyrolyses-gas chromatography to eliminate interferences of carbon black
and other similar materials (161).  Analytical  techniques usually used are
based on the analysis of a single component of the specific rubber or one
of its pyrolysis products, there is no general  analytical test for identi-
fication or quantifying the different types of rubber into a single clas-
sification which can be called "rubber".   As has been pointed out, the
major monomer units of rubber are ethene  and propene which are produced
on the pyrolysis of almost all rubbers (161).  Hence it is possible that
analysis based on the evaluation of ethene or propene from the pyrolysis
of samples containing rubber could be used to group all types of rubber
into a single test.  Additional research  needs to be conducted in this
area before analyses of rubber can be considered for storm generated
d i scharges.

Specific Conductance - Specific conductance measurements are often used
in water and wastewater characterization  to obtain an estimate of the
amount of dissolved solids in solution.  This measurement determines the
ability of a water sample to carry an electrical current.  The dissolved
solids value can be determined by multiplying an empirical factor times
the specific conductance of that sample.   It Is important to note,
however, that only ionized material will  contribute to the specific
conductance measurement.  Partially ionized organic acids (the fraction
that is not  ionized) and nonionized materials, such as glucose and
benzene, will not be detected by specific conductance measurements.  On
the other hand, specific conductance measurements are easily made and
continuous monitoring is possible.  High  salt content in stormwater runoff
from ice control can be detected by specific conductance measurements.
                                   214

-------
However, because of the lack of ability to consistently measure the
dissolved solids content of a sample, interpretation of specific con-
ductance data is very difficult and  it  is not recommended for routine
analyses in storm generated discharge studies.   In specific cases where
it is possible to correlate specific conductance with chloride content
or other ice control materials, it may  be useful.  When specific con-
ductance is used the technique described in Standard Methods (39) Section
\5k, page 323, should be followed.

Sulfates - Sulfate concentrations  in storm generated discharges are of
interest primarily from the receiving stream water quality standpoint.
U.S. Public Health Service Standards recommend an upper concentration
of 250 mg/1 in any water to be used  for human consumption.  Also, sul-
fates are detrimental because of  the odors produced when anaerobic con-
ditions exist.   In the absence of  oxygen, sulfates are reduced by anaero-
bic bacteria to  hydrogen sulfide with the resulting objectionable odor.   In
some cases the hydrogen sulfide that  is produced under anaerobic con-
ditions can be oxidized by a special group of bacteria in another part
of the  sewer, usually the crown,  and produce sulfuric acid  (48).  When
sulfuric acid comes  in contact with  the cement  in concrete sewers,
corrosion of the sewer results.   High sulfate concentrations could also
result  in a digester going sour.

In most surface  waters sulfate  is  not a problem.  For example, Lake
Michigan has a sulfate concentration  less than 20 mg/1 and Lake Superior
is about 3 mg/1  (162).  Hence the  addition of storm generated discharges
to surface waters will not increase  the sulfate  concentrations to levels
of concern in most cases.

The methods for  the analysis of sulfates have been presented in Standard
Methods  (39) and discussed as to  their  relative  merits.  Routine analysis
of sulfate is not recommended.  However, in the  special case where sulfate
is known to be a problem or potential problem, the Gravimetric Method
with Ignition of Residue  is suggested as described in Standard Methods,
Section 156A, page 331 (39).

Temperature - Temperature is important  in both biological and chemical
reactions  in that the higher the  temperature, within a given range, the
the faster the rate of reaction will be, and the lower the temperature
the slower the rate of reaction.   The solubility of chemicals in chemical
treatment processes and of oxygen  in water are also affected by temperature
Solubility of oxygen at 10°C is given as 11-3 mg/1 when no chloride is
present  in the sample compared to  4.5 mg/1 at 50°C (39).  The temperature
in most storm generated discharges will range somewhere between that of
the air and the  ground surrounding the  sewer lines.  Temperature measure-
ments are not recommended in storm generated discharge studies because of
their relatively constant temperature and the impracticality of making
any major changes in these temperatures.  However, in some cases where
disinfection of  storm flows is being performed temperature monitoring may
be important.
                                    215

-------
Trace Organics- Trace organics in storm generated discharges may come
from natural substances, pesticides, insecticides, fertilizers, herbicides
and industrial or other sources.   Some of these materials are very resis-
tant to biological activity and may remain in the water for long periods
of time even after treatment.

Pesticides, polychlorinated biphenyls and phthalic acid esters have been
found to be widely distributed in the environment (163).  Using computer-
ized gas chromatography-mass spectrometry, plasticizers and dye carries
were identified in two New England rivers at levels from 0.1 to 30 ppb
(parts per billion).   In another study a column of polystyrene macroret-
icular resin was used to isolate hydrocarbons in well  water (164).  The
hydrocarbons apparently originated from a coal  tar pit used in the 1920's
for disposal.

Organic materials are of particular concern when they are carcinogenic in
nature, or when they  tend to build up to toxic levels in cellular material
of man and animals.  Very small quantities of some of these materials may
be very significant and it is  necessary to be able to detect a few micro-
grams per liter or parts per billion concentration.   Very few compounds
can be detected at these low levels hence concentration steps are necessary.
During the concentration of wastewater samples the concentration of many
other organics and inorganics  will also be increased and these will often
interfere with analyses.  Hence,  isolation techniques may also be necessary.

Standard Methods (39), Section 139, page 259, describes procedures that may
be used to obtain a measurement of organic contaminants.  Two procedures
described use activated carbon to adsorb the trace organics from large
quantities of wastewater followed by solvent extraction to remove these
materials.  The solvent is finally evaporated and the weight of the residue
determined.  Three basic steps have been used in more specific analysis of
organics in air and water samples (165)-  These steps are:

      1.   Organics are concentrated and isolated from the matrix.

      2.   Constituents of interest are identified.
      3.   Amounts of  the identified compounds are measured.

Gas chromatography is used in  the second step of identification.  Two pri-
mary objections to the use of  activated carbon are presented.  The recovery
of organic materials  from the  activated carbon is usually incomplete and
variable, and  changes in the characteristics of the sample may occur,due to
the carbon acting as  a catalyst.   It should be recognized that the procedure
of evaporating the sample to dryness will usually eliminate the lower boiling
compounds.  Also, activated carbon may only recover a low percentage of
organic substances from the water (164).

An excellent review of different concentrating procedures was recently made
(166).  These techniques included a) activated carbon, b) liquid-liquid
                                     216

-------
partition, c) carbowax k,000 d) sllicones chemically  bonded  to  diatomaceous
earth, e) aromatic and alkyl chlorosllanes or  celite,  f)  porous polyure-
thane plugs, g) macroretlculated  resins and  several solvent  extraction  pro-
cedures.  In another report three techniques were  described  for sample
concentration:  liquid-liquid extraction, head space  sampling and  a  column
packed with porous polymer heads  (164).  Methylene chloride  was used as the
solvent.In the liquid to liquid extraction with an Increase  In  trace organic
concentration achelved using 15 ml methylene chloride  In  500 ml  of the  water
sample and removing a 10 ml portion  for analyses.

The headspace- procedure  is achieved  by withdrawing gases  In  the region  above
a water  sample through a collection  columncontain ing  porous  polymer  head
packing.  The volatile organlcs  In the sample  are  retained by the  packing,
but water Is not.  The sample cell described  Is shown  In  Figure VI-7 (165) •
                        VACUUM
                               COLLECTION COLUMN
                                          CHARCOAL TRAP
                                      \\\\\N<—  AIR
                                           SAMPLE
       Figure VI-?.  Schematic diagram of a cell used to collect
                 headspace vapors from solids and liquids

                            From Reference 165
                                      217

-------
Advantages of headspace techniques for trace organic concentration were
given as:

     1.  Since no measurements are made on the water itself, many
         types of handling problems such as filtering of particu-
         lates or emulsion formation are not  encountered.

     2.  This type of measurement would be readily automated for
         semi-continuous monitoring.

     3.  This method is ideally suited to low-level  determination
         of very volatile compounds.  These are usually masked in
         a gas chromatogram when using common extraction solvents.

However, organics that are insoluble, complexed or adsorbed on solid
material will not be efficiently concentrated.  Surface film may produce
higher than expected concentrations.

The third concentration method described the use of a packed column with
porous polymer beads through which a water sample was passed.   The ad-
sorbed organic was then forced into a gas chromatograph which  was temper-
ature programmed.  Low molecular weight molecules did not show a high
recovery efficiency but larger molecules that tend to be less  soluble
in water exhibited a high recovery, as shown in Table VI-19-
                Table VI-19-   RECOVERY OF ORGANICS FROM
                  WATER USING POROUS POLYMER PACKINGS
Component
Methanol

Acetone

Chloroform

Benzene

Pyr id i ne

Phenol

Methyl Isobutyl Ketone
m-Cresol
o-Ethy 1 phenol
p-Ethyl phenol
Concentration ,
ppm
0.79
6.30
0.79
6.30
1.50
12.00
0.88
7.00
0.98
7.80
0.22
1 .90
2.70
0.15
0.06
0.35
Percent
Recovered
(one pass)
<5
<5
21
42
93
85
100
100
46
79
25
61
100
75
97
89
     Reference  165.
                                  218

-------
When liquid-liquid extractions are used, solvents may be either heavier
or ligher than water.  Extraction vessels have been described to utilize
either type of solvent (166).  Figure VI-8 shows the extractor used for
solvents ligher than water.  The operation of this  is as follows:

     The extractor design  is a two-cycle system.  The water cycle
     is continuous flow.  Water enters at A and exits at B.  In
     so doing it passes through chamber C which is half-filled
     with solvent.  A stopcock, D, can be provided to regulate the
     water flow rate.

     The second cycle is a solvent cycle.  This system  is closed
     in that the solvent cycles exclusively in the extractor.
     The 500-ml bulb E contains pure, nonmiscible, organic sol-
     vent.  This solvent is gently boiled and vapor rises  in area
     F up through the upper extractor tube G into reflux condenser
     H.  At this point it  is liquified and falls off of drip tang
     1 into funnel J.  The long funnel stem sets up a hydraulic
     head sufficient to drive the solvent through a porous glass
     frit at K.  The frit  homogenizes the solvent resulting in
     fine beadlike particles which form as an emulsion as they
     rise through the water  in chamber C.  This emulsion extracts
     organic solutes during the period of water-solvent contact.
     The emulsion separates  in the extractor's lower neck L and
     the solvent-solute mixture spills over connection tube F into
     boi1 ing flask E.

     The closed solvent system cycles fresh solvent from the
     boiling flask into the extractor.  After extracting the
     organic solutes, the  "loaded" solvent is returned to the
     boiling flask thus collecting and concentrating the extracted
     solutes in flask E but always supplying fresh solvent to the
     extraction chamber at C.  (167).

For solvents heavier than  water the extractor shown in Figure VI-9 was
used.  It operated as follows:

     Water enters at M and exits at N.  The stopcock 0  (optional)
     regulates flow  rate.  While in the lower nick of the extractor
     P, the water flows through the extraction solvent.

     The solvent cycle, which  is closed to the extractor, starts
     by the solvent  being  vaporized in bulb Q.  The vapor  rises
      in arm  R to condenser S where the vapor liquifies and drops
     from drip tang  T to funnel U.  The solvent under a hydraulic
     head is forced  through the upper extractor nick W.  Extraction
     of the water takes place at the  interface between the emulsi-
     fied solvent and the  water.  A stirring bar at X  (optional)
     stirs the solvent-water mix.  The solvent separates in the
     lower half of the extractor and  flows through tube Y  into bulb
     Q.
                                   219

-------
Figure VI-8.   Extractor design  for
   solvent lighter  than water

      From Reference  167
Figure VI-9-  Extractor design
for solvent heavier than water

      From Reference 167
                                220

-------
     Organic solutes extracted from water then concentrate in
     bulb Q (161»).

After the trace organics were removed from the water phase by concentra-
tion in the solvent, they were concentrated further by distilling the
solvent.  Freeze concentration and freeze drying were also considered
for the removal of excess solvent (166).

The above discussion of concentration and analysis of trace organics
will provide guidelines for analyzing these materials in storm generated
discharges when these analyses are necessary.  The routine determination
of trace organics in storm generated discharges  is not recommended unless
there Is a good possibility that these waters contain industrial contami-
nants or agricultural and other materials from land runoff.  The optimum
techniques for analyzing a given class of trace organics will vary de-
pending on their characteristics.  Other methods for trace organic
analyses are described  In the section on pesticides.

Recommended Analyses

Many of the parameters discussed above may be important for specific
storm flow studies.  For example, storm generated discharges may be
contaminated from industrial wastes that will greatly affect the waste-
water characteristics.  The best method to determine the proper analytical
program will include a careful assessment of the surrounding industries
and their potential contamination.  Some industries may have process
overflow connnections^to^sewer lines and surface runoff from industrial
areas may be diverted to sanitary, combined or storm sewers.  Parameters
such as oils and grease, asbestos, residual organics, or pH may become
the critical analytical parameter.  However, for routine analytical pro-
grams In storm flow studies most of the tests discussed in the previous
section will not be required.

The observations of unusual items by operating personnel at stormwater
treatment plants should be recorded In a log book.  The parameters to be
entered  in the log book will  Include such items as odor, color, presence
of foaming, presence of large debris, etc.  The use of tables similar to
those in  Figure VI-5 Is recommended and will facilitate the operator In
his log entries.

pH is the only parameter In the "other" category that should be regularly
monitored In storm flow studies.  This test will provide an  indication of
periods when chemical treatment can be expected to be inefficient as pre-
viously discussed.  Unusually low or high pH wastewaters greatly inhibit
the formation of floe particles  in chemical coagulation.  The pH of a
wastewater also determines, to a large extent,.the form of chlorine pre-
sent and the oxidizing power for a given chlorination dose.  This can con-
trol the necessary contact time for adequate bacterial kill or chemical
ox I dat ion.
                                    221

-------
pH should be measured as described in Section H4, page 276 of Standard
Methods (39).  Great care should be taken that the electrodes are main-
tained in a clean condition, free of grease, oil, or other surface
materials.  pH samples should be taken in situ if at all possible.  When
in situ pH measurements are made the probes or electrodes must be pro-
tected from breakage due to debris in the wastewater.  When in situ
sampling  is not possible the pH of individual or composite samples should
be made as soon as possible after the samples have been taken.  pH measure-
ments should always be conducted prior to addition of acid or other
chemicals used for sample preservation.  pH data from permanently instal-
led units should be periodically checked, preferably with a carefully
calibrated portable unit.

Other tests described above may be important in specific storm flows and
may become routine tests for these studies.  This discussion will serve
as an introduction to the analyst when he is required to test for unusual
contaminants.

ANALYSIS OF SLUDGES

Most, if not all, stormwater treatment processes produce a secondary flow
containing the solids removed from the sewer discharge.  Although this
secondary flow may vary in suspended solids concentration from 1200 mg/1
to 110,000" mg/1, it is usually referred to as sludge.  At the present time,
it is common practice to collect the sludge in a holding tank and pump it
back Into the sewerage system when the storm is over.  It is probable,
however, that disposal in this manner may not be the best method of sludge
disposal  in all cases.  Research is presently being conducted on alternate
methods of sludge treatment and disposal.  Because of the dilute nature of
the sludge, thickening is usually necessary before ultimate disposal.
Possible alternate forms of final disposal include incineration and
burial  in  a landfi11  site.

The analytical parameters to be measured depend, of course, on the sub-
sequent sludge treatment process and the method of final  disposal.  In
all cases, however, the sludge solids concentrations should be measured.
For dilute sludges, suspended solids should be measured.   Thick sludges
should  be analyzed for total solids.   Sludges discharged to a minicipal
sewerage system might also be analyzed for heavy metals,  oxygen demand
and nutrient analyses to indicate the effect on the sewage treatment
process.   Sludges which are incinerated might be analyzed for heat value.
Sludges transported to a landfill site might be analyzed  for pH and
heavy metals.

Sludge  thickening processes produce a clarified flow that must be dis-
charged.   If this flow is discharged to a receiving stream, the same
analyses should be performed as are done for the treated effluent.  In
addition,  occassional samples (say k per year) might be analyzed for
heavy metal  concentrations.  Since it is possible that the soluble con-
taminant concentration may make disposal  to a receiving stream undesir*
                                    222

-------
able, this flow may have to be discharged elsewhere.  The parameters to
be analyzed will then depend on the conditions to be met for disposal.

The specific analytical procedures to be used  in the analysis of the
clarified flow from a thickening process are no different than the pro-
cedures to be used for analysis of the storm generated discharge.
Procedures to be used  in the analysis of sludges, however, may have to
be modified.  The parameters that might be measured along with recommended
modifications for analysis of sludges are listed below:

     Total Residue - This  is the recommended method of residue
           analysis for thick sludges.  The procedure described
           on page 280, of the EPA Manual (1971) should be used.

     Suspended Residue - This is the  recommended method of re-
           sidue analysis  for sludges that are low  in solids
           content.  The procedures recommended for storm dis-
           charge should be used (page 167)-

     Volatile Residue - (Optional) - The procedure described
           on page 282 of  the EPA Manual (1971) should be used.

     Total Phosphate - (Optional) - High solids sludges may
           require high dilution prior to digestion.  Blending
           of the sample may facilitate getting a representative
           sample for dilution.  The diluted sample should be
           digested and analyzed according to  the procedure
           described on page 239 of the 1971 EPA Manual.

     Kjeldahl Nitrogen - (Optional) - The procedure described
           on page 149 of  the 1971 EPA Manual  should be used.
           Sample size should be reduced if large amounts of
           organic material are present.

     Oxidized Nitrogen - (Nitrate and Nitrite) (Optional) -
           Sludges must be clarified before the oxidized nitrogen
           analysis can be run.  Centrifugation or filtration
           through a 0.45 u membrane  filter is usually sufficient.
           When large amounts of colloidal material remain, however,
           it may be necessary to add zinc sulfate and form a floe
           by raising the  pH.  The floe can then be removed by
           centrifuge.  The clarified sample should then be
           analyzed by the cadmium reduction method described on
           page ^58 of Standard Methods, 13th  Edition.

     pH -  (Optional) - Because of chemicals used in the treatment
           process, some sludges may have a pH so low that storage
           in unprotected containers may cause problems.  The method
           described on page 230 of the 1971 EPA Manual, should be
           used.
                                    223

-------
BOD - (Optional) - Large dilutions are required when
      analyzing sludges containing high solids concen-
      trations.  After dilution the procedure described
      on page 1M of this report should be followed.

TOC, TOD or COD - (Optional) - Large dilutions are re-
      quired when analyzing sludges containing high
      solids concentrations.  Homogenization by mixing
      in a Waring Blender is recommended after dilution.
      Additional homogenization using a tissue grinder
      may be necessary for TOC and TOD analyses.  The
      recommended procedures for analysis of the diluted
      samples are:
TOC -
TOD -
COD -
page
page
page
146,
1*3,
U6,
this
this
this
report
report
report
Heavy Metals - Zn,  Pb, Cu, Ni, Cr - (Optional) - After
      dilution, most sludges can be digested by the
      nitric and hydrochloric acid procedure described
      on page 209 of this report.  In special cases,
      perchloric acid digestion may be necessary.  Analysis
      of the digestate by atomic absorption spectrophotometry
      is recommended (page 186 of this report)-

Mercury - (Optional) - After dilution, the sample should be
      digested by the nitric acid reflux procedure described
      on page 188 of this report.  Mercury analysis should be
      performed by the flameless atomic absorption method
      described on  page 187 of this report.

Density - (Optional) - The density of thickened sludge may
      be needed to compute hauling costs.  The recommended
      method of measuring density is by means of a wide
      mouth pycnometer such as the Hubbard - Carmick Specific
      Gravity Bottle (Corning #1620).

Heat Value - (Optional) - This parameter may be useful when
      final  disposal of sludge is by  incineration.  The
      sludge should be air dried and pulverized by mortar
      and pestle.  This material is then pellatized and the
      heat value measured with an adiabatic calorimeter.
      The instrument manufacturers instructions should be
      followed.

Pesticides and PCB's - (Optional) - The procedure described
      on page 189 has provision for the analysis of sludge
      samples and is recommended.
                             224

-------
ANALYSIS OF ACCUMULATED ROADWAY MATERIAL

In all cases acceptable storage and analytical methods must be used when
studying accumulated roadway materials.  The samples should be stored
under conditions as required for each specific analysis.  The time be-
tween collection and analysis must be kept to a minimum.  All samples
should be kept  in a refrigerated or ice cooled containers during and
after sampling.  This cooling will slow the biological reactions and
prevent significant changes  in the nature of the sample.

The only analysis that should be conducted at the  test site are volu-
metric (measuring the volume of flushed material).  Gravimetric analyses
should be conducted under suitable controlled laboratory conditions  (168),
Refer to specified sample storage and analytical procedures for each
pollutant that  is to be analyzed.

Chemical tests  can be determined either on wet and dry sample fractions
separately, or  the sample fractions may be combined proportionately by
study area for  analyses.  Bacterial analyses, if required, should be con-
ducted as soon  after sample  collection as possible.  Portable membrane
filter techniques using field  incubators are recommended.

Classical BODj.  analyses are  not believed accurate  for typical street
surface contaminants.  The reasons for this include the  low dilution
ratios needed and the toxic  effects of associated  pollutants on the
organisms (168).  One method that attempts to eliminate  these problems
is described by  Colston (169).  This method measures the COD of an aerated
sample with time.  The rate  of COD reduction  is assumed  to be proportional
to the BOD of the sample.

During the Washington, D.C.  study (170) procedures  in Standard Methods
(39) were followed in most cases.  However, numerous modifications were
made as these procedures were  intended primarily for use with liquid
samples and no  standard methods exist for the analysis of street surface
contaminants.   Investigators have used a diversity Of methods, some of
which need improvement and standardization so that  results of different
studies can be  compared.   Methods for grease and  for characterization
of grease into  hydrocarbon and normal paraffin fractions were pieced
together from a  number of exisiting procedures.  In some cases, no
satisfactory methods existed prior to this project  for measurement of
the parameters  of interest.  Therefore, methods for the  estimation of
asbestos and rubber had to be developed for the analysis of roadway
samples.  Development of these analytical methods  and their limitations
are discussed  in the following below.

Determination of Rubber

The technique of pyrolysis-gas chromatography was  used to develop a
method capable  of detecting  0.005% rubber  in  roadway dust and dirt
samples.  Pyrolysis-gas chromatography was first applied to the
                                   225

-------
Identification of vehicle tire rubber in roadway dust by Thompson, et a
in 1966 (171)-  More recently, this approach was used for the quantitati
estimation of rubbers in compound cured stocks (172).  Styrene-butadiene
rubber (SBR)  is converted to styrene and other low molecular weight com-
pounds by pyrolysis in a nitrogen atmosphere.  The styrene  is then
separated and measured via gas chromatography using a flame ionization
detector.  Briefly, the method entailed pyrolysis of 20 to 25 mg of
extracted sample for 20 seconds at 6^0°C in an inert nitrogen atmosphere.
Dust and dirt samples were first extracted with aqueous acid to remove
soluble materials and carbonates and then with hexane to remove inter-
ferring organics.  Next, the gaseous pyrolysis products were chromato-
graphed and the styrene peak measured.

SBR is the most commonly used synthetic rubber for vehicle tires manu-
factured in the United States.  Passenger car tires contain 70 to 80%
SBR, small truck tires 60 to 70% and large truck tires only 10 to 20%
SBR.  Since the total traffic at the roadway sites consisted largely
of passenger cars, estimation of SBR in dust and dirt will give a stais-
factory estimate of tire material in roadway samples.  The standard
curve shown in Figure VI-10 was generated by measuring styrene produced
upon pyrolysis of known amounts of passenger car tire rubber.   No rubber
was detedcted in several of the roadway samples initially examined be-
cause of large amounts of interfering compounds produced during pyrolysis.
These compounds obscured the styrene peak.  A preliminary extraction of
the acidified dust and dirt samples with hexane reduced the background
interferences to a satisfactory level.

Determination of Asbestos
The method for the determination of asbestos in dust and dirt and flush
fractions of roadway samples was based upon an industrial hygiene pro-
cedure recommended for airborne asbestos by the National Institute for
Occupational Safety and Health (NIOSH) (173).  In this procedure the
flush water or aqueous suspension of the dust and dirt was sonicated
briefly to disperse particulates and then membrane filtered.  The filters
were rendered transparent by the action of a mixed organic solvent and
the asbestos fibers enumerated using phase contrast optical microscopy.
Only fibers between 5 and 100 microns in length and having an aspect
ratio (length to breadth) of 3 or greater were counted.

During development of this procedure, a "standard" suspension containing
10 mg/1  of chrysotile asbestos was prepared and analyzed repetitively
for use in estimating precision and recovery levels.  Chrysotile was
selected as it is the variety of asbestos most commonly used in the
United States.  The "standard" suspension was found to contain 10.6 x
10^ fibers/ml with a standard deviation of 2.8 x 10^ fibers/ml.
Recoveries of asbestos fibers added to three dust and dirt samples
were 38%, 85%, and 65%, respectively.
                                   226

-------
 ^^. o


  x
o c
— 4)
li J.Q
Q.O
Ul X
z
LJ Ul
CC C 1
i- n I
                 0.5
.0         1.5

 RUBBER, mg
2.0
                Figure VI-10.  Standard curve  -  rubber in dust and dirt

                                   From Reference  168
                                   227

-------
Increasing the sonication time from one minute to five minutes did not
increase the yield from dust and dirt or from the asbestos "standard"
suspension.  This indicated that sonication was not fracturing fibers
in the samples.  Tap water was examined along with subsurface soil
samples thought to contain no asbestos fibers in an attempt to check
for naturally occurring inferences.  No asbestos was found in the tap
water (the detection limit in this analysis was about 10* fibers/1).
Values of less than 3 x 10^ fibers/g were found in the two soils
examined.  The levels found in the soils were at the limit of detection
for these particular samples and represent less than one fiber from
each soil in over 50 fields counted under the microscope.  Detection
limits on actual  roadway samples were generally over one order of
magnitude better than with soils.

The toxicology of asbestos fibers has not been well defined and the
NIOSH method is based upon expediency and precedents set by earlier
investigators.    Further,  it was not intended for environmental samples
but rather for industrial  hygiene purposes at mining operations or plant
areas where asbestos products are fabricated.  Presently, asbestos
analytical  methodology is  trending toward the use of techniques requiring
more sophisticated equipment and considerably more man hours per deter-
mination.  Transmission and scanning electron microscopy are being used
for the most critical analyses of environmental  samples to measure fibers
below the range of optical methods.  Particle size distribution and weight
of asbestos found are frequently required in addition to numbers of
fibers.

Review of Sampling and Analyzing Roadway Material

Listed below are salient points which summarize the procedures to be used
when studying accumulated  roadway material.

     1.   In order to ensure characteristic  samples of street
         surface contaminants samples should be collected by
         the use of hand sweeping, dry vacuuming, and water
         flushing of the study area.  If the data collection
         program is concerned with only the weight of solids
         accumulated,' however, hand sweeping of the study area
         should be sufficient.  Particle size distribution studies
         will require hand sweeping followed by dry vacuuming.
         Any pollutant considerations in addition to solids should
         include a water flush.

     2.   Each sample should consist of three fractions:  litter,
         dust and dirt, and water flush.  The particulate  materials
         collected by sweeping and vacuuming are separated on the
         basis of particle size into the litter fraction and dust
         and dirt fraction.  The litter fraction consists of that
         portion of the particulates retained by a U.S.A. No. 6
         sieve (greater than 3.35 mm in diameter).  The dust and
                                   228

-------
    dirt fraction will contain particulates smaller than
    3-35 mm in diameter.  The water flush fraction con-
    tains those components of the dust and dirt fraction
    which were not picked up at high efficiencies by the
    sweeping and vacuuming techniques.

3.  Sample site selection should consider the effects of
    area land use, average daily traffic, the roadway
    surface material and the condition'of the roadway,
    and the speed limit and other traffic controls.  Curb
    should be available at the sampling site for the
    collection of water flush samples.  The safety of the
    area during sampling should also be considered.  At a
    minimum, flagmen and traffic cones should be employed
    during all sampling operations.

4.  For a typical secondary street a single test area
    should cover 93 sq m, 7.6 x 12.2 m (1000 ft2)(25 x l»0
    ft2)-  Larger paved surfaces can be better sampled
    by using smaller test areas, 0.9 sq m (10 ft2), and
    averaging the results.

5.  The frequency of sampling will be largely dependent on
    the objectives of  individual sampling programs; however,
    any extended sampling program should be developed to
    cover seasonal variations and weekly variations (week-
    days versus weekends) in the accumulation of street
    surface pollutants.

6.  The following parameters have been used to characterize
    street surface pollutants:

         Total solids                 TKN
         Total volatile solids        Chlorides
         BOD                          Asbestos
         COD                          Fecal coliform
         Grease                       Fecal strep
         Petroleum                    Lead
         n-Paraffins                  Chromium
         Total phosphorus             Copper
         Nitrite-nitrogen             Nickel
         Nitrate-nitrogen             Zinc

    Most of these parameters have been found to be mainly
    present in the dust and dirt fraction of samples; however,
    BOD, TKN, and microorganisms are  largely present  in the
    water flush fraction.
                           229

-------
         The BODr is not recommended because it may not be
         accurate for typical street surface contaminants.
         The reasons include the low dilution ratios needed
         and the toxic effects of associated pollutants on
         the organisms.  One method that may eliminate these
         problems has been described by Colston (169).

         Development of standard methods for analyzing street
         surface contaminants are needed.  During the Washington,
         D.C., study (170), numerous modifications of Standard
         Methods (39) procedures were occassioned as these pro-
         cedures were intended primarily for use with liquid
         samples and no standard methods exist for the analysis
         of street surface contaminants.  The procedures
         developed by this study can be used as the basis for
         uniform analysis of street surface contaminant samples.
SUMMARY
This section has included a discussion of the selection of those para-
meters to be analyzed routinely in storm generated discharge projects
and the recommended analytical  methods.  The following analyses are
recommended:


         POTENTIAL OXYGEN DEMAND:   TOD and BOD5

         PARTICULATE CONCENTRATION:  Suspneded Solids

         PATHOGENIC INDICATOR:   Fecal  Coliform

         EUTROPHICATION POTENTIAL:  Total Oxidized Nitrogen,
             Total  Reduced Nitrogen (Kjeldahl) and Total
             Phosphorus

         OTHER:   pH

These parameters are to be analyzed routinely for the evaluation of storm
flows and the evaluation of treatment processes treating such flows.  In
addition, periodic analyses for certain metals, pesticides and oil and
grease along with a few other parameters that may be peculiar to a specific
discharge are recommended.  If, during these periodic analyses, significant
concentrations  are found, then  a more routine analysis program for these
other parameters should be instigated.

Preservation of  the entire sample begins with refrigeration or ice cooling
during the sampling period and  during the transportation period back to
the laboratory.   When compositing is performed it may be done either on
site or in the  laboratory, whichever is more convenient.  However, in most
cases it will be necessary to have a record of the flow available when
compositing is  being performed.  Once the sample has reached the labora-
tory it should  immediately be broken Into separate portions for each
                                    230

-------
analysis.  It is understood that  if the  laboratory  is more than a maximum
of 6 hours travel from the point  of sampling, then  sample splitting and
preservation should be done at the point of sampling.

For BOD samples analysis must be  started as soon as possible  (limit 6-12
hours) with no preservation other than refrigeration.  TOD samples should
be homogenized and then preserved with sulfuric acid and refrigerated.
This will allow sample storage for up to 7 days.  Samples for fecal coli-
form analysis cannot be preserved and must be analyzed immediately.
Chlorinated effluent must be dechlorinated at the site of sampling.  The
nitrite/nitrate samples should be preserved with ^0 mg/1 of HgCl2 and
refrigerated for up to seven days if they are not going to be analyzed
immediately.  Total Kjeldahl nitrogen can be preserved also by the addi-
tionof ^0 mg/1 of HgCl2 and refrigerated at 4°C, however, analysis as
soon as possible is recommended.  Total  phosphorus  and suspended solids
may be preserved by refrigeration for up to a week with no chemical
add!tion.

Plastic sample containers can be  used for all of the routinely analyzed
parameters other than fecal coliform which should use separate glass
containers when a high degree of  accuracy is required for this analysis.
It  is estimated that a minimum volume of 2000 ml is needed to perform all
the routine analyses.

In general, those analytical procedures  as described in Standard Methods
(39) are recommended.

The recommended method for the BOD analysis addresses the importance of
insuring that a sample representative of the suspended solids present is
withdrawn from the sample jar.  Suspended solids analyses should be done
with glass fiber filters, again due to the large particulate concentration
Bacteria analysis by membrane filter technique  is recommended for fecal
coliform analysis unless extremely high  quality waters are involved.
Kjeldahl nitrogen and nitrate/nitrite analyses  are  the same as those re-
commended in Standard Methods (39).  Phosphorus should be digested with
persulfate and colorimetrical1y analyzed using  ascorbic acid.

Table VI-20 contains a summary of the type of container, sample volume,
preservation methods, holding times, and analytical techniques for the
recommended quality characteristics.
                                   231

-------
                             Table VI-20.  SUMMARY OF CRITERIA
                         FOR ANALYZING THE RECOMMENDED CHARACTERISTICS
Recommended
indicator
BOD
TOD
SS
TKN
Phosphorus
(total)
N02/N0
Fecal co li form

Oils & grease
Pesticides
Metals
Container
plasti c
plastic
plastic
plastic
plastic
plasti c
glass
(sterile)

glass
glass
glass/plasti
Volume
Routine:
600
100
500
100
200
100
100
Periodic:
1000
1000
ca 200
Preservation
Refrig. at k C
H SO^ addQ- Refrig.
Refrig. @ k°C
HgCl2 and/or
Refrig. @ A°C
Refrig. at k°C
HgCl. and/or
Refrig. @ ^°C
None

H SO. + Refrig.
@ A C
Refrig. @ ^°C or
freezing
HNO. add to pH 3
+ refrig.
Holding
time
6 hr
(2k max)

ASAP
7 days
7 days
7 days
ASAP

2k hours
ASAP
6 months
Analytical technique
Standard Methods +
mod.
Manuf. Instructions
Glass fiber f i I ter
Standard Methods
Persulfate
Ascorbic acid
Standard Methods
(Total reduced to No
Membrane f i 1 ter
technique

Standard Methods





2>



Soxhlet Extraction
modification described
Standard Methods

a.  Depends on metal type.

-------
             SECTION VII  - DESCRIBING STORM GENERATED FLOWS
The periodic nature and extreme variability of storm generated discharges
requires somewhat different procedures for describing their effects  and
impacts.  This section discusses the various methods of describing
storm generated discharges as well as methods of constructing hyetographs,
hydrographs and pollutographs.
EXPLANATION AND DISCUSSION OF TERMS AND UNITS
Pollutant concentrations are conventionally measured and reported in
units of mg/1.  These units are a measure of the concentration of
pollutants in weight per volume in the sample that was collected,
regardless of the time interval over which the sample was collected or
the method of compositing the sample.
mass/unit time

The most elementary method for reporting the mass of pollutants in a
wastewater stream  is to  integrate the pollutant concentration and the
flow rate.  This unit of measure enables the quantification of the
relative impact of discharges.  The selection of a specific unit of time
is a function of the variability of flows and pollutant concentration,
the frequency and  type of sampling and the availability of a continuous
record of flow.

Dry-weather municipal wastewater treatment plants usually composite
samples over a 2k  hour period and calculate mass pollutants on a kg/day
 (Ibs/day) basis.   This is calculated according to the following
relationship:

                                P - V x C

     where:  P  =  kilograms per day of pollutant
             V =» hydraulic flow in cubic meters per day
             C = concentration of a pollutant in mg/1
                                    233

-------
This parameter is useful for reporting the overall efficiency of a dry-
weather wastewater treatment plant and is readily understandable by
both technicians and laymen.  Because of the equalizing capabilities
of municipal sewerage systems, the variation in mass loadings  (kg)
is generally not greater than one to two times (perhaps more in smaller
plants)  the average.  On the other hand, combined sewage overflow and
stormwater  runoff are extremely variable both in hydraulic flow rate and
pollutant concentration.  In addition, storm events do not occur
regularly nor is the overflow necessarily 2k hours in duration.  As
such, the establishment of mass loading rates is a somewhat more diffi-
cult task.

To facilitate understanding of the importance of evaluating mass
loading rate for storm generated discharges, a somewhat simplified
example is  presented in Figures VI1-1 and VI1-2.  Twelve discrete
samples (two hour composites) were collected over a twenty-four hour
discharge period.  Average flow rates for each two hour period were
also determined.  This discrete data was then plotted In continuous
form In Figure VI1-1 .

The mean mass loading for the 2k hour period as well  as the continuous
mass loading rate for each two hour period is shown in Figure VI1-2.  It
is important to note that the peak mass loading rate is approximately
15 times the minimum rate, a factor of extreme importance when
considering treatment processes such as screening and filtration.
Calculation of the mean mass loading does not consider the magnitude
of these variations which occur during a storm event.  In the case
where treatment processes or quality models which are sensitive to
variation in mass loading are being considered, it is necessary to
evaluate loading rates over shorter time periods than 2k hours.


kg/ha (Ib/acre), kg/ha/day (1b/acre/day), kg/ha/year (1 b/acre/year),
   kg/cm/ha (1b/in./acre)

Once the total emissions from a storm generated discharge have been
established, it is desirable to relate this to the drainage basin.
A convenient relationship is kg/ha and can be established by dividing
the kg discharged by the area of the drainage basin.   The parameter
kg/ha can be compared to the dry^weather sewage contribution on this
basis.

Efforts have been made to relate unit emissions to some time scale by
the parameters kg/ha/day and kg/ha/year.  These parameters can be  quite
-useful  as "rule of thumb" numbers for comparing the magnitude of
storm related discharges with dry-weather sewage and other sources of
environmental  pollutants.  The shock impact of storm related discharges
tends to be masked in this parameter since the many dry days each  year
have zero contributions.  Pollutant loadings are also quite often
                                   234

-------
                    '01*00  	1200'

                       TIME  (HOURS)
                                     2000
                        I I I 1L.  \l IWUI\«J /


         Figure  VI 1-1.   Average  flow rates for

       twelve discrete  samples  (2  hour composite)
£-3^   9080 (20000
  •-^ i/i
O W1 4-»


ill 68'o (isooo
< 4-» 3
O 3 —
CO  .
                                             MEAN
        (10000



£ 2270  (5000
                0000 oi|00  0800 ,200 1600 20QO


                               TIME (HOURS)


    Figure  VI1-2.   Pollutant  mass loading curve
                           235

-------
expressed relative to the amount of  rainfall  over an area,  kg/cm/ha
(lb/in./acre).   In this way it  is possible to estimate the  pollutant
loading from given areas for different rainfall  amounts and to determine
the relationship of various amounts  of rainfall  on the amount of
pollutants generated.


kg/curb meter (Ib/curb mile)

The parameter kg of pollutant per curb meter  was coined by  the American
Public Works Association as a convenient  method  of expressing the
accumulation of pollutants on street surfaces.  This parameter appears
to have significant merit particularly when it is applied to land areas
from which the  major source of  pollutants may originate from street
surfaces such as central business districts and  other commercial  areas.
The parameter can have significant error  when utilized in areas where
there is a low  percentage of roadway in proportion to total drainage
area such as in parklands and low housing density subdivisions.  For
this reason, the parameter should be utilized and interpreted with
caution.  Also, the parameter should always be related to some unit
of time such as kg/curb meter/year (Ib/curb mile/year) or kg/curb meter/
days between rainfall (1bs/curb mi 1e/days between rainfall).
DEVELOPING BACKGROUND DATA

Contri but ing Area

The land area contributing to a storm generated discharge should be
described in as much detail as possible to permit future extrapolation
of results to other areas.  In the case of impact evaluation studies,
a rigorous definition of the drainage area may permit preliminary
estimates by the loading factor technique (1).  The detail  to be
utilized in describing the areas will be somewhat related to the
project objectives but engineers and scientists are urged to exercise
every effort to present as much detail as possible in reports and
technical papers.  Maximum benefits are gained from storm generated
discharge projects only when the data is complete enough so that it
can be evaluated and extrapolated by other interested parties.

As a minimum the following parameters should be presented and quantified,

     1.  Drainage area and subareas
     2.  Land usage
     3.  Population density
     ^.  Median incomes (residential)
     5.  Climate
     6.  Topography
     7-  Pervious and impervious areas
     8.  Street and curb miles length
     9.  Average daily traffic
    10.  Methods and frequency of street cleaning
    11.  Miles, size and slopes of sewers

                                  236

-------
Drinage Area and Subareas - The drainage area under Investigation
should be clearly identified on a map of suitable scale.  Drainage
areas may be further broken down Into subareas to more precisely
identify major land use categories or topographic characteristics.
Sewer maps, topographic maps and aerial photographs are generally
available from local authorities and should be utilized wherever possible.
Advanced techniques for plotting and reproduction may be extremely
valuable.  Where possible, It would be useful to identify the source of
maps.
Land Usage - The majority of prior and ongoing studies have broken down
subareas according to land usage.  The most common subdivisions are
residential  (single and multl-family), commercial, industrial  and open
lands and parks.  A typical breakdown of subareas according to land
usage is shown  in Figure VI 1-3.

The U.S. Bureau of the Census data tract can be of significant value in
assembling this information.  Information such as land use, density, etc.
is readily available from this source.   It is generally possible to
develop satisfactory land use information by the use of district sewer
maps, aerial photographs and census  tracts, or from private firms such
as the National Planning Data Corporation, Ithica, New York.


Population Density and  Income - Population density information (people
per ha) can  be  utilized for evaluating residential areas.  This
information  can be readily obtained  from the U.S. Census Bureau on a
block by block  basis.   In areas which are mixed  (i.e. commercial/
residential) population density  information Is of less value since the
non-residential activity may obscure the residential  impact.

Census bureau data on median income  from residential areas  is also
available and can be useful  in further breaking down pollutlonal Impact
from residential areas.  SWMM utilizes this type of information to
estimate water  usage and sanitary sewage contributions.


Cl'mate -  Because of the nature of urban runoff studies, climatic
conditions  is an extremely important parameter.  Although detailed
information  on  storm characteristics is discussed in other  sections
of this report, It Is appropriate to point out the desirability of
including a  summary discussion on climate for those unfamiliar with
a  particular area.

An excellent source of  this  information  is the Local Climatological
Data - Annual Summary with Comparative Dat_a_, published annually by
the U.S. Department of  Commerce, Environmental Science Services
Administration, Environmental Data Service, National Climatic Center,
Asheville, N.C. 28801  (Table VI 1-1).
                                    237

-------
to
CO
CO

Char.
0-1
0-2
0-3
0-5
0-6
0-7
0-8
RS-I
RS-2
R5-3
RS-4
RS-5
RS-6
RS-7
RM-I
RM-2
RM-3
RM-4
C-l
C-2
C-3

l-l
1-2
S-l
5-2
Area
(acres)
38.5
18.8
275.0
9.2
34.5
34.5
138.0
32.2
27.6
113.0
160.0
396.0
216.0
78.0
137.5
27.6
85.5
57.5
11.5
39.2
38.0
160.0
46.0
9.2
79.5
34.5
40.0

Land Use



Open Land and Parks





Residential
Single Family Housing



Residential
Multi-Fami ly Housing


Corronercial

Industrial
Schools
                                                                                          \-— '7
1
' *"!
}(r
J rji
x '/~~'c-L
/--U/ ! '
lo-j ; ',

i * ~j
L ' °"S
' ""'i 	 i
; i
"p 'T — "*
«-j j v X *-i^
	 -T~T i"1- 1
L c.! /..$ ! c-> r"1" ^
" \ 1 \ I
i e-« ) ! 0-4 1 ' — N '»« "^_ «s-7
•-4 / ', ,,j f } o 	 -\
tN "*' \ ' i ! \^'
I-~'V, \ u-i 	 -— ~-l-~ — *^
' \ ,'e.,, \ , z * -x S-*
\ V v \ " ! «-• y




,f
s
f
s



T\
\\
1 1
' 1
J



1
                                                                   f  '   I      ^.        I
                                                                   '^<3^  \	-/

                           39.0
                                              Hospital
                                                                                                                    l.....0p«ti Itntf mnd P«rk«

                                                                                                                       F*»Ily Housing
                                                                                                                   M	RcsldcntUt HultU
                                                                                                                       Faally Housing
                                                                                                                    C,.,..CoTmcrcl»l
                                                                                                                    I	Industrie!
                                                                                                                    S	Schoolt
                                                                                                                    H	Hetplul
                  Figure  VI1-3.   Division of Bloody Run sewer watershed  into different  land uses

-------
                      Table  VI 1-1.    LOCAL  CLIMATOLOGICAL  DATA,

                      ANNUAL   SUMMARY  WITH  COMPARATIVE   DATA,   1968

                                          MILWAUKEE,  WISCONSIN


NARRATIVE  CLIMATOLOGICAL  SUMMARY
The climate  of  Milwaukee is Influenced  by  the general storms which
move  eastward  along the northern  border  of  the  United States, and
by those  which  move northeastward  from the southwestern part of the
country to the Great Lakes; also by the high barometric pressure systems
that move eastward  or  southeastward  across the  country.  Far these
reason* the weather  changes  frequently.  During the winter and spring
months there  is seldom a period of  more than 2 or 3 days without a
distinct change in the character of the weather.

Milwaukee's  climate  is   also influenced  to  a considerable  extent by
Lake  Michigan.  This Is  especially true in the spring,  summer, and
fall months when the  temperature oi the Lake water varies considerably
from  the air temperature.  During  the  spring and early summer, a
shift of wind  from a  westerly to an easterly direction frequently causes
• sudden  10"  to 15" drop  in daytime temperatures. In the  autumn, the re-
latively warm^water  of the Lake prevents nighttime temperatures from
falling as low as  they do a few  miles inland  from the shoreline.

The following averages   and  extremes  are  based  upon  the combined
weather records  made at tne  former  city  office in downtown Milwaukee.
ind those made  at General Mitchell  Field, covering a period generally
from 1871 through 1964. *

Milwaukee's  annual  averaee   temperature for  the  period of record.
1871 through 1961, was  4o.~6*. It r-inres from 21.2° tn January to 70.8°
In July. The highest temperature ever  recorded in the City by the Weather
Bureau was 105' on  July 24, 1C34,  and the lowest was -25' on January Q,
1875.  The City  has an averare of 13 ciays per year when the temperature
reaches zero  or  lower, and 123 days when It  reaches 32° or lower. Minima
of 0" have been  recorded as late as March 25, and 32" as late  as May 27
In the spring. In the au:umn, a low of 32'  has been recorded as early as
September 20, and 0s as  early as  November 21.  The average  number of
days  per  year  In which  the temperature  reaches W or higher is 8,
Consecutive da>s with readings of 90° or higher seldom exceed 3, although
there have been as many as 10.

The average  annual  precipitation  is about 30 inches.  About two-thirds
Of the annual amount occurs  during the growing season. Since 1541,  the
wettest year was 1876  with  50.36 Inches, and the driest year was 1901
with 13.69 inches.

                        TEMPERATURE
                    (From 1871 through 1964)

           The Greatest Number of Consecutive Days With:

Max  90'  or above    10 days,  from  Aug. 25, 1953  to  Sept. 3, 1953
MIn 32° or lower    110 days,  from  Dec. 6, 1374 to Mar. 25. 1875
MIn  20*  or lower     46 days,  from  Dec. 31. 1°11 to  reb. 14, 1912
Mln  10*  or lower   -   24  days,  from  Jan. 12, 1Q63  to  Feb. 4, 1963
MIn  0'  or lower   17 days,'  from  Jan,  27.1895 to Feb. 12,1895
The long-term  average annual snowfall Is about 4ft inches, but It varies
considerably  from  season  to  season.  During :he winter of 1885-86,
110 Inches were measured,  while In 1884-85, the snowfall totaled only
11 Inches.

Thunderstorms  occur  less  frequently and  with less severity In the
Milwaukee area than in areas to the south and west. Hall size Is generally
1/2 Inch or  less,  although It has been noted as large as 2'lnches in
diameter with  unusually severe storms.  The maximum rainfall which
has occurred  In a 24-hour period  is 5.76 Inches In June 1917.  As much
as 0.79 Inch  haa fallen in  5  minutes, 1.11 inches in 10 minutes, 1.34
Irenes In 15 minutes, 1.86 inches In 30 minutes, and 2.25 Inches In 1 hour.

There are about twice  as many cloudy days during the winter as there
are during the  summer. The average percentage of possible sunshine
ranges from 40 percent in December to 70 percent In July,

The city office  of the Weather Bureau was located on the 7th floor of
the Federal B'ulldlng,  517  East  Wisconsin  Avenue, until It was closed
on May  1, 1954. The thermometers  and rain pages were located on the
roof". The  terrain Immediately around the  station Is fairly level; the
Federal  Building being  located on a  low ridge  between Lake Michigan
on the  east and  the Milwaukee  River on the west, about 1/2 mile from
the lake shore  and 1/4 mile from  the river. A few bloclts west of the
river  the  ground becomes  gently rolling  toward the west and north.

South  and  southwest  of the station  the  ground la level for a distance
of 2 to 4  miles,  being  a  fllled-ln swamp. "Between IS70 and May 1.
1954,  the  various locations  ot  the  dry  office were within A blocks of
each  other.  Temperatures were  Influenced to  a  conquerable  extent
by Lake  Michigan,  especially  during  the  spring and early summer.

The airport office  Is located on  the third  floor of the Administration
Building, General Mitchell  Field, which is 6.6 miles south of the Federal
Building.  The  field Is located In  the NNE sector of a vi?ry shallow cir-
cular depression about 4 miles In diameter.  The station Is about 3 miles
west of the Lake Mtchiean shore. Temperatures are Imluenced somewhat
by breezes from Lake Mlchli^an  during the spring and early summer.
Cold  air drainage from  the surrounding higher terrain results lr. lower
nighttime temperatures at the station  than at  other points In that vicinity.

                   FREEZE DATA (1871-1964)
Shortest growing season of record --JVD.days, 189-
Longest growing season of record - 223 daye, 1915
                  U.S.   DEPARTMENT   OF  COMMERCE
                  MAURICE H.  STANS,  Secretory
                  ENVIRONMENTAL  SCIENCE  SERVICES ADMINISTRATION
                  ROBERT M. WHITE, Adminiitiolor
                  ENVIRONMENTAL DATA SERVICE
                                                               239

-------
                            Table  VI 1-1   (continued).  LOCAL  CLIMATOLOGICAL DATA
to
tfi.
o
        LATITUDE     42- 57. „


        LON31TUDE     87" M' w

        ELEVATION (ojoaod) 672 Feet
METEOROLOGICAL DATA FOR THE CURRENT YEAR
 murxoxEE, «iscoKsrn
GENERAL UITCBELL FIELD

         19«8
Month
Ft
P
L.
NO
Ob
Temperature
Average*
E
3
J E
3 \
2a <. 1 i. 9
5: »
75 3
45 w

<.4 3
57 2
60 1
31 2

c

53 2
69 5
38 1

Extreme*
,c
o>
X

0
1
1
e
e

t>
l

5.
i *
AUG-

j

2
3
k
20

1

3V
6+
6
3
28
20
JAN.

1

75«
521
363
31
92
799

Piecipilillon
1

0.31
2.90
3.2«
3.9
3.3
2.S

Gr**!«ttn
24 bn.

0.09
0.90
1.09
1.97
1.22
0.98

1

6-17
23-2*
23-2«
28
JUN.

Snow. Sleet
I

1 2
0 4
0 0
0 0
0 U
0 3

.S
6 £

0 7
0 *
0 0
0 0
0 0
0 3

Q

S2-23
2*
28
DEC.

Rcl
hum
S
t
C

*
1
6
a


tnnd
me
ENTRJ





live
dity
ye,
rd
JRCd
\L





9
I
I

21
Temperatures
M»imum
ll

0
0
u
2
7
0
10
a '.



Minimum
K J3

20
9
0
0
0
0

"s J

9
0
0
0
0
0
0
u

-------
                                                 Table  Vll-l      (continued).     LOCAL  CLIMATOLOGICAL  DATA
                                                                              NORMALS,  MEANS,  AND  EXTREMES
•5
"w
TR
Temperature
Normal
|
>. |
• J
c "i
12. a
(.2.9
52.6
57.8

I
(b)
63 3
67 h
6- 3

Extremes
cc .c
9
69
95
93

1

62
64
61 +
UL.
Is
II
9
2
3
3

i

66
66
65+
61
N.
Normal degree days
[b)
372
135
47

Precipitation
Norma] lolal
(b)
3.16
3.44
3.06
2.72

0
3 >•
J •£
2B
5.27
B.28
7.07
9.87

1

5
0
1
S P.
l!
19
1 25
0 B5
b 46
u 30

1

1 61
1 65
1 *«
1 56
F e.
I J
il
2>
2 0&
5 13
* 05
5 2U

£

48
50
13
41
P.
Snow, Sleet
Mean total
26
T
0.0
0.0
T

Maximum
monthly
26
0.4
0.0
0.0
0.0
r

1

1960
FCB.
1 Maximum
1 1*24 tin.
28
0.4
0.0
0.0
0.0

1


FEB.
Relative
humidity
3
X
3
t
CE

74
76
BO
84

(D
am
HTfl
9
78
61
83
87

1
arc
use
.AL
0
60
60
60
62

*
d:
6
60
62
63
68

Wind
||
26
1
1

|l
14
HUE
NNE
SW
SW
55W

Faste* mU«
1
26


1
26
Sw
S
u
w
5

i


HA .
Pel. ol possible sunahln* |

e

Mean sky cover 1
sunrise to sunael |

6.3
5.9
5.1
i.3
5.3

Mean number of day*
SuruiM
' lo
•unaet
0

1
11
10
11
If
6
4
9
10
10
11
11
10
9

I
IB
M
15
14
12
9
9
10
11

I Precipitation
| -01 inch ormora
1
1
1
1

!
||
1*1
3
1
*
0
0
0
0
0
0

TbundeiBlorma



j?
ac
3
3
3
3
1
2
1
2

Temp*
Max.
|s
0
0
0
0
3
J
1
0

-a
C J
si
20
17
C
0
0
0
0
0
0

rahire*
Win.
J|
30
20
11
2
0
0
0

1
1
0
3
0
0
0
0
0
0
0
0
3
to
                      Mams aad extremes are  Iron  toe existing location.  Annual extremes  have been exceeded at other locations as follows:
                      Highest teriperatura 105 in July 1934; lowest temperature -25 In January 1875; maximum monthly precipitation 10.03 In June  1917; nlalmua monthly precipitation 0.05 in
                      Uxrcb 1910;  najcinun precipitation in 24 hours 5.76 in June 1917;  naxidun monthly snowfall 52.6 In January 1918;  maximum snowfall In 24 hours 20.3 in February 1924.
                                                           llnleia orhervlte Indicated, dlmer.Blonal  ur.lrs uied in thin bulletin are-. lemprnmre In degree* F.-     Figures Inutead or letter* ID t direction column Indlcjt* dlrcclloo In t*tu of dvcrMi from tnM North-
                                             (1931.1940).      prrcipka-ion.  Including anowliil, In Inchea; wind movemem In mllri p«r hcur; a'd rrlmlve humidity     Lc.. 09 - Ei.t,  18-Souih,  77-We«t, M, - Nonh. amj 00 - C»lm.  Rc.uUim wind li (tw vwxar «um 
-------
                     Table VI 1-1   (continued).   LOCAL CLIMATOLOGICAL DATA
            AVERAGETEMPERATURE
                                                     TOTAL DEGREE DAYS
to
*.
co
Year
1931
IV 38
1939
Jan
Feb.
Mar.
29.6 33.4' 33.7
35.4 31.1^TT.-f
i
!»-»
1946
19*7
19*8
1949
1950
1951
1952
1956
1957
1956
1959
1S60
1961
1964
1965
1966
196T
BECORD

iS.fc
25.6
21.9
13.-
26. 1
13.9
21. B
21.1

23.2
*!.,'
Apr.
!_*.,*
*J. 7
May
-£f
57.9
June
July
^•^4
66.0 3.2
[A-*
Sept.
^^^
73.4 t,i.B
72.0' 67.4
I • '
17.6
23.5
33 4
1-V.3 33.9







-5.C














"•'







Oct.
Nov.

5J.2








»..


40.9






Dec.
Annual
HSHf-JfcJ
27.4
JJ.B

ll:l
j ,,




*4.a

,..,
.8.2
«..


Season
930- 1
937- B
938- 9


9&4-J5

967-68
5?
July
U-l
1
2
42
9
0
h~ IT

46
Aug.



i
i
i
~T


Sept
Oct
	 9*9 * T i
89
1*5
113

116
*73
290


104 ,373




Nov.
7*5
685
975
73;
,„

I ~6<.0'
615


Dec.
Jan.
1150 100
H-n 077
10*8
1168

1126
1137
1303


300
12"


33*>
561


F«b.
880
172
M*,
1TI
u:;
•

209




1C06
1067


Apr.
6M
668
5*3
648

500
516
*21
623


May
306
327
210
313
130
266

332
?2«
25*


June! Total
]
.»i
13S
103
42

1 	
60
,!!
73
119
68

6*64
~~V>OB
6850
7662
7081
6T17
6501

7038
62*3
659}
7152
7503
7361


-------
                 Table VI 1-1   (continued),   LOCAL CLIMATOLOGICAL  DATA
           TOTAL PRECIPITATION
TOTAL SNOWFALL
CO
^
CO
Year
1931
1932
1933
1935
1936
1937
19JB
1939
1941
1942
1944
1945
1946
19a7
.948
19' 9
1951
1 9 -12
1953
1954
1955
1956
1957
1958
1959
I960
1962
1963
1964
1966
1967
1966
PCCOPO
„ ,
Jan. reb.
1.0-' 0.45
TF5 .4.-'
.CO. .55
::: :;;
.12 .72
.60, .33
.63; .24
Mar.
4.76
3.93
1.97
1,98
0.67
1.74
3.29
1.54
Apr.
May
3.9->' 1.7n
2.97| 9.56
1.5? 2.73
3. 05' 2.29
2-30 2.55
4.80 2.70
0.97' 3.73
2. 81' 1.40
June
July
Aug.
Sept.
Oct. | Nov.
Dec
Annual
3.5F 2.25; 1.96 3.15' 2 . 1 C • 4.tV 1.65' 29. J9
2. 47! 4.55 1.75
2-32. 1.10 1.43
4.34 3.59 3.06
1.93J 0.26 5.92
2.64 3.36 0.80
6-93 2.7C 6.47
3.5C ' 0.51 S.03
2.57
<-.33
1. 12
5.59
6^12
1.53
.50 .63! 1.82; 1.93' 3.03! 3.42J 2.93 ! .?9 9.87
.4.J, .69
.31 .40
.97 ,ae
.26 .29
.07 .68
;»' -.7
.57 .43
-Jl .22
.481 .04
/JJ, .04
06 27
(.9 31
98 56
2.48
2.46
1.40
2.B6
1.7J
J.59
3.33
3.67



J.Ob
3.61
3.61
1.35
0.31
0.99 2.88
3.74' 2.33
0.94
3.66
1 .91
-.35
fc.OS
4.9l| 3.87
2.95, 2.86



3.81
3.47
2.67
2.70
2.90


2.17
2.57
2.12
2.00
1.80
3.28
2. i>' 1.5* 2.31
3.42 2-771 1.54
2. SI1 0*95
3.96' 2.17
3.19' 2.16
2.971 3.12
4.03 6.69




1.33 I i.T4
1.7C
0.85
1.68
7.38
7.7»
7.frfr
2.6*
3.32
1.63
1.58
0.46
2.56
3.59


1.98
6.15
3.2T
1.13
2.59
0.37
1. 26
6.03
2.75
O.Jt


1.49
l.T*
0.41
1.69
3.36
2. S3 1.03
*.;a e.56
1.37 3.43
3.77J 0.34
1.63 0.85
Q. 76] 1.66
2.1.3' 0.33

0-B3
0.29
1.79
1.65
0.33
0.17


2.1*
0.17
2.68
1.76
J.70
0.94
J.15
1.54
2.08
2.82
2.15
1.13
1-22
l.*2
2.1&
1.42
1-10
0.46
22. >•>
35.52
29.02
30. M
30-35
25.82
41.86
23. JB

0.99
1.14
1.54
1.72
2.50
1.99 2.^6
3.371 2.10
^T

0.81
2.02
2.70
1.52
2.56


0.55
0.98
3.73
2.31
1.33
<:o.7a
25. J7
2C.89
2*. 62
36.43
32,69


21.91
28,18
38.49
27. 1>
25.65
31.51
Season
93:1-31
931-32
93*.-35
935-36
936-37
937-38
728-39

942-tJ
91, 3-4 4
9*5-*6
946-67
950-51
951-52


961- 2
963- •
964- 5
965-66
966-67
967-68

Julv
1
Aug.jSept
Oct.
Nov.l Dec.
1
Jan.
O.ol D.O 3.0J T 4,Z| 9.0 10.1
:.;' • .; 2,: c.u' T i ;. ? 7.4
0.3
0.3
0.0
0.3
C.3

C.J
0.0
0.0
0.0
0.0
0.0


c.o
0.0
c.o
0.3

c.o
0.0
0.0
O.O
0.0
0.0


0.3 O.C
0.0
0.0
0.0

0.0
0.0
0.0
0.0
0.0

3.;
0.0
0.0

T
O.C
J.O
0.4
3.0
0.0
0.0
J-.0

0.0
J.O
o.u
0.0
0.0

I
;
T

T
c.o
O.C
T
c.o
o.u
T
T

T
1
T
0.0
0.8

3.5
1.0
0.9

3.1
1.6
1.5
0.1
9.0
5.8
12.4
I
7.6


T
1.*
T
2.0
U.4

6.3
8.1
1:5
4.2

a.i
i
11. «
5.2
20.1
30.7
6.3


7.
12.
14.
9.
1.

10,2
24.3
0.7
7.6
15.7

3.0
lb*i
16.2
15^7
K.5


23.1
3.8
23.6
24.6
13.1
Feb
3.8
2.."
n'.c
23.8
1.9
2.9
5.8

J.5
9.3
7. a
10.-
8.2
1.7


32. i
5.7
10.1
7.7
27.1
3.5
Mar j Apr. May [Junej Total
3 .3 T
1 .'••
1 .9
.2
,9
1 .1
.8
1 .1

;t
1 .6
i .0


11.4
19.8
26.7
2.B
7.4
1.2
c.e
0>2
0.4
13.0
T
0.3
0.6

T
0.0
0.9
T
0.8
1.7


4.0
I
r
1.3
T
0.4
T O.C
O.C
2.0
3.2
0.0
C.O
0.0
U.O

0.0
0.0
0.0
T
0.0
c.o


0.0
0.0
0.0
T
U.O
O.C
0.3
0.0
0.0
0.0
0.0
O.C

0.0
0.0
0.0
0.0
o.t
O.C
0.0
0.0


O.C
0.0
0.0
0.0
63.4
37^5
29.1
49.1
74.2
20.2
21.3
37.3

54.*,
21.0
30.1
48.7
49.1
79.3
90.8
25.0
37.1
32.9
69.8
74.0
M.O
12^1

-------
                        Table  VI 1-1   (continued).    LOCAL  CLIMATOLOGICAL  DATA
                                                                 STATION  LOCATION
  MILWAUKEE, WISCONSIN
GENERAL MITCHELL FIELD



Location

CITY OFFICE

Broadway and Michigan
Insurance Building,
Kitchell Building, North
Water and Michigan
Federal Bldg., 4th Floor
Federal Bldg., 5th Floor
Federal Bldg., 7th Floor
517 £. Wisconsin Avenue
AIRPORT STATION
Old Adnm. Bldg., 1011
Mitchell Field, 5.75 mi.
S of Post Office
New Adnm. Bldg., 1011
E. La'yton Ave., General
Mitchell Field
New Terminal Building
5300 S. Howell Avenue






g
S



12-10-70
3-23-78

4-22-99
3-10-32
3-28-41


4-21-27


8- 5-40


6-20-55








Occupied to

12 10 70

3-23-78
4-22-99

3-10-32
3-28-41
5- 1-54


8- 5-40


6-20-55


Present







J J
d a d
ll
111



50 ft. 11
300 ft. <

1000ft. ENE
20 ft. 3






100 ft. S


4750 feet
ssw






Latitude
North

.

43° 02'
43" 02'

43° 02 '
43" 02'
43° 02'


42° 57'


42" 57'


42° 57'







Longitude
West

7°

87° 54'
87° 54'

87° 54'
87° 54'
87" 54'


87° 54'


87° 54'


87" 54'




Elevation abov«
S«a
level
a
!
?
I
o

608

608
600

620
620



674


674


670
670
b672



Ground

J!
3
Wind Lnitnic



114
149

b221
C221



45


66


88
88
a20



m
1
a
o
B
5
•S
e
B
e
1




106

126
98
125


32


33


35
5





Psycbromete




106

125
97
124


32


33


33
33




1
Telepaychroi


















5
°



%
Tipping bud
rain gage




a 100

117
C88






31


33
33
33




1
tf
.3
3
o»
J








115





4


35
4
c-5





1
a
0}




100

117
88
115


28


31


34
3
d4




«
E
0
!>,
m






















Sea
level
1*
a>
I
a
i,
a

























Remark*





a - Added 10-14-91.

b - 139 ft. to 4-29-27.
c - Removed 3-1-41.










Elevations as of 6-23-56.
b - Effective 3-1-61.
c - Removed 3-1-61 ,
d - Effective 10-26-61.
to
                              Requests  for additional information should be directed to the Weather Bureau Office  for which this summary was issued.
               Sale Price-  15 cents per copy.  Checks and money orders should be made payable to the Superintendent  of Documents.  Remittances and correspondence regarding
               this publication should be sent  to the Superintendent of Documents, Government Printing Office,  Washington, D. C.  20402
                                                                                                              USCOHM-ESSA-ASHEVILLE -  900

-------
Topography - Topographic Information is most difficult to present in
final report format since typically topographic maps are prepared for
rather large areas and are not easily reduced to a satisfactory scale.
Where possible, reduction of topographic maps should be made.  If this
is not feasible, it is recommended that sources of maps be given in the
report to facilitate future work.
Pervious and Impervious Areas - Pervious and impervious areas play an
extremely important role in the runoff rates from various areas.  This
information is typically gathered from aerial photographs and can be
presented in summary form  in either tabular format or through the use
of subarea maps.  Street areas and length of curb can be handled in the
same manner.
Average Daily Traffic -  In certain areas of a metropolitan community,
heavy vehicular traffic can generate a significant buildup of dirt and
various other pollutants on streets and the adjacent right of ways.
 In most metropolitan areas average daily traffic  (ADT) counts can be
obtained from the traffic engineering department or city engineer's
office.  It  is useful to include this information in final reports.
The relationship between roadway usage and water pollution has been
reported (2).  The accumulation of pollutants on streets and highways
 is influenced by the frequency and method of street cleaning.  For studies
which are evaluating pollutants in urban runoff,  it is desirable to
 initiate a program for accumulating data on street cleaning at the
onset of the study.  This program should include a notation of the
dates of all cleaning activity, the method of cleaning employed and the
volume and weight of collected materials if possible  (3).


Miles, Size and Slopes of Sewers - The sewerage systems of most cities
are mazes of pipes, culverts and ditches of all sizes, materials of
construction and slope.  Oftentimes the existing maps are not current
and may require significant field inspection to verify as-built
conditions.

Sewer  information can be presented in tabular form which is referenced
to a sewer system schematic  (Figure VI !-*»).
Dry-Weather Flow and Quality

Dry-weather flow rates and quality characteristics are important
parameters for evaluating combined sewer overflows.  Hydraulic flow
rates determine the excess capacity available for conveying stormwaters
and/or  the capacity available for bleeding back flows from off line
storage reservoirs.
                                    245

-------
to
                                    .—_t>
                                          n	'i
                                                                                          r-	-if
                                                                                          \        l
                                                                                                   •
                                                                                                   •
                                                                                     o tonhotei

                                                                                     • Input Manholes
                        Figure VI 1-4.  Division of Bloody Run sewer  basin into subareas
                                   and numbering of the sewer system's  elements

-------
E 1 emen t
Number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
2k
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
Location
or Diameter
input manhole
54"
manhol e
54"
manhole
48"
manhole
48"
input manhole
66"
manhole
72"
rnanhol e
75"
input manhole
72"
input manhole
72"
input manhole
78"
input manhole
84"
manhole
60"
input manhole
90"
input manhole
60"
manhol e
60"
input manhole
10' x 8'
input manhole
10' x 8'
input manhole
10' x 8'
input manhole
11 ' x 9'
input manhole
11 ' x 9'
input manhole
12' x 9'
input manhole
12' x 9'
input manhole
96"
Slope

2.60%

2.30%

2.20%

2.60%

1.40%

0.90%

1.00%

1.40%

1.80%

1.25%

1 . 1 0%

1 . 0 0%

1.30%

1.20%

1 . 0 0%

0.70%

0.70%

0.70%

0 . 4 0%

0.40%

0.30%

0.28%

0.60%
Contributing Subarea
or Length (feet)
Drainage Subarea 1
394.0

689.3

380.4

467.0
Drainage Subarea 2
319.4

345.3

577.0
Drainage Subarea 3
462.7
Drainage Subarea 4
1 428 . 5
Drainage Subarea 5
1432.0
Drainage Subarea 6
664.0

149.7
Drainage Subarea 7
1238.5
Drainage Subarea 8
750.0

700 .0
Drainage Subarea 9
700 .0
Drainage Subarea 10
900 .0
Drainage Subarea 11
438.3
Drainage Subarea 12
1473.6
Drainage Subarea 13
350.0
Drainage Subarea 14
1673.1
Drainage Subareas 15&16
1433.4
Drainage Subarea 17
595.4
Figure VI1-4 (continued).
      subareas and number
      U . O U/o         !>??• T
  Division  of  Bloody Run  sewer  basin  into
ing of the  sewer  system's elements
                               247

-------
The dry-weather sewage characteristics can be used to predict quality
and quantity of combined sewer discharges and to key into certain
possible treatment schemes.   For example, interceptor sewers containing
high concentrations of dissolved BOD may alert design engineers to
potential difficulties in meeting discharge requirements with treatment
schemes which remove only particular contaminants (screening, flotation,
sedimentation).

Dry-weather flow rates and characteristics are also necessary input to
certain predictive models.  The EPA Stormwater Management Model
utilizes dry'weather characteristics to estimate solids accumulation in
sewers, velocities required  for scour and the determination of the
resultant sewage/stormwater  mixture to estimate combined sewage
characteristics.  Oftentimes flow characteristic information Is available
from various locations in the collection system such as lift stations,
or metering stations.  Pump station records can provide long term records
which are invaluable sources of information.

When this information is not readily available it may be necessary to
install sampling/gaging stations to collect the necessary data.   The
need for this data dictates  that time series sampling be conducted as
hourly variations (or as low as five minutes depending upon the
drainage area and study needs) if flow and characteristics are desired.
The parameters to be analyzed would generally be BOD (total and dissolved),
suspended solids and fecal coliform as a minimum.  A data record of at
least one week is desirable.  Quality characteristic data is sometimes
available from past studies  or sewage treatment plant records.

Prior to initiating this program, it is important to consider the
potential impact of industrial wastes and to make sure that various
contributing industries are  operating at normal conditions if they
have any significant impact  on wastewater characteristics.

Establishment of satisfactory raingage networks must be tailored to
the objective of the specific project.  Criteria for raingage networks
have been defined in Section V.  Because of the need for historical
records it is often not feasible to install  raingages for projects of
only 1  to 2 years duration.   However, where a project is attempting to
correlate predictive model results to observed runoff hydrographs it
may be necessary to install  one or more gages to supplement existing
raingage information.  In this case it is recommended that the following
criteria be utilized.

     A.  One centrally located gage for watersheds up to A8.5 ha or
         O.A8 sq km (120 acres or 0.18? sq mi).

     B.  Three evenly spaced gages for 260 ha or 2.6 sq km (6^0 acres
         or 1 sq mi) watershed with a length to width ratio of A.
                                    248

-------
     C.   Five evenly spaced gages  for 2600 ha  or  26  sq  km  (6400 acres
         or 10 mi)  watershed with  gages spaced about every 2.42 km
         (1.5 mi).

     D.   All  gages  should be selected, installed  and operated  according
         to Weather Bureau Observing Handbook No. 2, Substation
         Observations, Unites States Department of Commerce,  Environmental
         Science Services Administration Weather  Bureau.
RECOMMENDED METHOD OF EVALUATING STORM GENERATED DISCHARGES

Method of Constructing Hyetographs

Hyetographs are graphical representations of the variation of rainfall
intensity with time.  They can be constructed from recording raingages
with relative ease.

In order to prepare a hyetograph it is necessary to select a time over
which to calculate the rainfall intensities.  For hydrologic models
it may be of interest to calculate rainfall Intensities for periods as
short as 1-5 minutes.  For other purposes, less frequent intensities
may be desired.

The accuracy of calculated rainfall intensities is subject to the chart
speed and the recording scale of the raingage.  Typically recording
raingages have a horizontal  (time) scale of one hour - 9.65 cm (3.8 in.)
and a vertical (rainfall) scale of 2.54 cm  (1 in.) of rainfall - 2.06 cm
(13/16 In.).  For this scale the minimum readable data is approximately
0.64 cm  (0.025 in.) of rain and 10-20 minutes or an intensity of

      0.064 cm (0.025 In.)   60 minutes ,    w   /hr ((JJg ,n>/hour)
            10 minutes           hour


The sensitivity of raingages can be Increased by utilizing a faster
chart speed or a digital printer which prints rainfall depth in a
binary-decimal code suitable for automatic data processing procedures.

Hyetographs can be produced by taking the data from a recording
ratngage  (Figure VI1-5), transforming the data  Into proper format
(Table VI1-2), and then plotting the transformed data on rectilinear
graph paper  (Figure VI1-6).


Method of Constructing Pollutographs

The term pollutograph was coined by the authors of SWMM as a means of
defining the mass per unit time discharge rate of pollutants.  As such,
it can be applied to any wastewater constituent calculated as kg  (Ibs)
emitted as a function of time.  The pollutograph  is extremely useful In
                                   249

-------
                                                     OSS
                                                             CHART NO.  4043-B

                                                              DUAL TRAVERSE
                                                          12" CAPACITY - 7.A, HOURS
                                                                 RAIN GAGE
                                                          THE OENDIX CORPORATION
(Q
 C
  I
U1
 30
 (ft
 s
 01

 3
to
 01

-------
                                     Table VI 1-2.   RAINFALL HYETOGRAPH
                                               November 20, 1974
          Date of Storm:
          Gage Location:
April 26, 1974
Fire Station
to
01
Time
0000
0200
oo4o
0100
0120
OHO
0300
0200
0240
0300
0320
03^0
0400
0420
044 0
0500
0520
0540
0600
0620
0640
0700
0720
0740
0800
0820
0840
Gage
reading
—
--
0.00
0.05
0.05
0.05
0.06
0.09
0.10
0.10
0.10
0.10
0.10
0.10
0.10
0.10
0.10
0.10
0.10
0.10
Increment,
in.


0.00
0.05
0.00
0.00
0.01
0.03
0.01
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
cm


0.00
.127
0.00
0.00
.025
.076
.025
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
Intensity,
in. (cm) /hour


0.00
0.15
0.00
0.00
0.03
0.09
0.03
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00


0.00
.381
0.00
0.00
.076
.229
.076
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
Time
1100
1120
1140
1200
1220
1240
1300
1320
1340
1400
1420
1440
1500
1520
1540
1600
1620
1640
1700
1720
1740
1800
1820
1840
1900
1920
1940
Gage
reading
0.14
0.16
0.24
0.32
0.38
0.47
.05
.50
.57
.65
.72
.74
-75
1.75
1.75
—
—
—









Increment,
in .
0.04
0.02
0.08
0.08
0.06
0.09
0.58
0.45
0.07
0.08
0.07
0.02
0.01
0.00
0.00












cm
.102
.051
.203
.203
.152
.229
1.47
1.14
.178
.203
.178
.051
.025
0.00
0.00












Intensi ty ,
in. (cm) /hour
0.12
0.06
0.24
0.24
0.18
0.27
1.74
1.35
0.21
0.24
0.21
0.06
0.03
0.00
0.00












.305
.152
.610
.610
.457
.686
4.420
3.429
.533
.610
.533
.152
.076
0.00
0.00













-------
                               Table VII-2  (continued).  RAINFALL  HYETOGRAPH
Time
0900
0920
09^0
1000
1020
1040
Gage
reading
0.10
0.10
0.10
0.10
0.10
0.10
Increment ,
i n.
0.00
0.00
0.00
0.00
0.00
0.00
cm
0.00
0.00
0.00
0.00
0.00
0.00
Intensi ty,
in. (cm)/hour
0.00
0,00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
Gage Increment, Intensity,
Time reading in. cm in. (cm)/hour
2000
2020
201*0
2100
2120
2UO
to
en
to

-------
to
01
CO
                    5.08C

                   (2.00)
                                                  TOTAL RAINFALL - A.M»5 cm (1.75  inches)
± c   2.5k
j~  (1.00

U. 3

14   '•"
tt 5  (0.50
                              n
                                                             n-
                      0200  0300  0400  0500   0600   0700  0800  0900 J000   1100  ]20Q   1300


                                                              TIME
                    Figure VI 1-6.  Rainfall  hyetograph - Racine, Wisconsin, April  26,  197^

-------
storm generated discharge studies  because it indicates the rate at
which various pollutants are discharged.

Pollutographs are developed in SWMM by multiplying the hydraulic flow
rate times the pollutant concentration at various points in time.
Confirmation of the accuracy of calculated pollutographs requires a
continuous record of flow and a series of discrete analyses of the
pollutant(s) in question.

The pollutograph in Figure VI1-7 was constructed by calculating the
mass emission rate from the flow rate and suspended solids concentrations
in Figure VI1-8.  Pollutograph information is particularly valuable as
input to time variable water quality models and  as a means of
evaluating the loading rate on various treatment devices.
Method of Constructing Hydrographs

A hydrograph is a graph of stage or discharge against time.  Many methods
of preparing hydrographs may be utilized.   Hydrologists for example,
may be interested in monthly and annual  mean flow to  display a record of
runoff at a stream flow station.  Engineers evaluating storm generated
discharge sites are more concerned with  hydrographs from specific storm
events.

Oftentimes the objectives of a particular  project will dictate the level
of sophistication of the hydrographs which will  be generated.  Mathematical
models such as SWMM have the capability  to generate surface stormwater
inlet hydrographs given the proper hyetograph and land characteristics,
to route these hydrographs through a conveyance system and ultimately to
generate an outlet hydrograph at an overflow point in the system.  SWMM
is capable of generating reasonably accurate hydrographs for predictive
purposes (4).  Because of the complexity of the model, readers are
referred to the users manual  for operational  details.  The latest
edition of the users manual was published  in March, 1975 and is EPA
report No. EPA-670/2-75-01.  A somewhat  less  complicated method for
generating hydrographs has been prepared by URS Research Company (5).
Utilizing the unit hydrograph concept (the hydrograph of 2.5^ cm (1.0 in)of
direct runoff from a storm of specified  duration) and the empirical
equations of Epsey, one may synthesize a unit hydrograph to describe  a
specific study area.  The unit hydrograph  can then be modified to
reflect any runoff rate or runoff duration for the study area.  Readers
interested in utilizing the unit hydrograph procedure are urged to
remember that the procedure is a statistical  representation.  A rather
detailed discussion and attempt to quantify estimating error in this
procedure has been made.
                                   254

-------
          11.35
      ^  (25.0)
      ^   ,9.08
      £   (20.0)


      c

      5   6.81

      2   (15.0)

       *
      C/>

      |   k.Sk
           (10,0)
      o
      CO
      OL
      CO
      CO
           2.27
          (5.0)
               o4oo    0500   0600     0700     0800
                                     TIME,  hours


               Figure VII-7.  Pollutograph
                                                        0900
   300  _
 O)
CO
o
o
CO
        I 946,5
           (15)
        0 631.0
   200  2 (10)
S        8

|  100  ~315-5
CO
=>
to
        ,

        O
            (5)
                                              SUSPENDED SOLIDS
               4000    0500   0600    0700    0800     0900

                                 TIME, hours

Figure VI1-8.  Suspended solfds and flow variation wJth  time
                               255

-------
Synchronization of Hyetographs,  Hydrographs  & Pollutographs

Stormwater studies and operational  facilities require data input from
numerous ralngages, flow measuring  devices and automatic  samplers.
Although it is extremely difficult  to  synchronize timing  of these
devices it is often vitally Important  to  the success  of a project that
this control  be exerted.

Experience in Racine,  Wisconsin  has indicated that a  significant effort
must be exerted to assure that data collection Is synchronized.   The
Racine project involved the following  monitoring  equipment:

     A.  Three - remote recording  raingages  (hand wound clock).

     B.  Three - remote Honeywell Automatic  Water Quality Data
                 Acquisition Systems (continuous  chart recorder).

     C.  Three - remote Sigmamotor  sequential  samplers (electrically
                 operated timers).

     D.  Two  - inplant  Parshall flumes and  two overflow  weir  liquid
                 level  recorders  (electrically operated chart  recorders).

     E.  Eight - inplant sequential samplers (electrically operated
                 timers).

In spite of almost daily checking of chart speeds and time verifications,
electrical failures and other mechanical/electrical malfunctions caused
significant problems with data synchronization.  The  following
suggestions may be useful in minimizing  the  problems  of data
synchronization:

     1.  Perform dally checks on all clocks, recorders, and charts
         for  time accuracy.

     2.  Where feasible use event  recorders  to document activities
         at remote sites, i.e. event recorders can be utilized  to
         document sampling events on remote  samplers.

     3.  Utilize multiple recorders for  inplant monitoring to minimize
         the  possibility that charts are  recording at different  speeds.

     k.  Verify chart  times and  mark the  actual time  on the chart.
         Adjust chart  position  if  necessary.

     5.  Utilize lights or buzzer alarms  to  notify personnel  in  event
         of power failures.

     6.  Where real time data logging  from remote sites is contemplated,
         provide system power checks to  insure that the monitoring
         station is on line and  monitoring the proper data.
                                   256

-------
Even with these safeguards In practice it will be necessary to
constantly alert operating personnel to the Importance of Instrument
synchronization in order to assure accurate results.


EVALUATION OF STORM GENERATED DISCHARGES

Storm generated discharges are now recognized as having a deleterious
impact on the surface waters of the United States.  The problem is
complex and not easily comprehend!ble.  The uncertainty and unpredictable
nature of these discharges presents engineers a most difficult
challenge in providing cost effective sysems to control the discharge
of pollutants.

Storm flow monitoring and evaluation programs (other than those just
studying stormwater characteristics alone) generally fall Into one of
the following categories:

     - Evaluation of impact on receiving water quality

     - Facilities design studies

     - Process evaluation

     - Compliance with regulatory requirements

It  is anticipated that monitoring and evaluation programs for each of
these four categories will vary substantially.  Accordingly a model
program has been developed for each category.  These model programs
reflect the authors' evaluation of the state-of-the-art and should
be considered as a  guideline  for data collection programs.  Local
conditions may dictate modifications and engineers are urged to
seriously evaluate  data collection programs before  Initiating work.
Evaluation of  Impact on Receiving Water Quality

It  is extremely rare that existing quality data can be used in evaluating
the Impact of  storm generated discharges because receiving water data
is generally collected with no regard to the occurrence of storm
events.  Thus  there is generally no way of isolating the effect of the
storm.

In order to collect valid Information on receiving waters  it Is almost
mandatory that a special monitoring program be established.  If data
collection is  to be used for verifying model  predictions it will be
necessary to develop a program to yield the necessary data.

It  is recommended that unless special circumstances warrant continuous
monitoring, the data collection be accomplished manually.  It Is
difficult to staff a project to respond to storm monitoring, however,
reviews of the literature as well as the authors' personal  experience
                                   257

-------
     k.  Does the waste contain toxic materials (heavy metals primarily)
         that would adversely affect the BOD test  and perhaps biological
         treatment?

     5.  Does the waste contain a significant amount of high specific
         gravity solids (sand, etc.) which might adversely affect plant
         operations?  Are special considerations to handle this material
         warranted?
     6.
Does the waste contain significant quantities of grease which
might adversely affect screening or filtration?
It is important to recognize that storm generated wastewaters are
significantly different than dry-weather municipal  wastewaters.   Engineers
are urged to carefully evaluate the characteristics to insure that the
selected treatment process will perform as expected.   Variations of
characteristics within and between cities must be considered when
selecting treatment processes.   Sampling and analytical  procedures must
be carefully selected to Insure the validity of collected data.


Process Evaluation
Detailed procedures for evaluating treatment processes are included in
Section VIII.
Compliance With Regulatory Requirements

At the present time there are no known nationwide standards or require-
ments for monitoring storm generated discharges or treatment facilities
for these wastes.   However, it is anticipated that these requirements
may be promulgated in the near future.  The cost of monitoring will  be
significant since some areas of the country have as many as 60-70 overflow
events per year and in some cases hundreds of overflow points.  Should
regulations require monitoring of all  overflow points, it will be
necessary to carefully consider the analysis requirements and the
overall feasibility of such action.

Where overflow monitoring is required it is recommended that flow
measurement equipment and automatic sampler systems be Installed which
are capable of providing samples proportional to flow.  A single
composite sample in many cases will be adequate to insure compliance
with regulations.   The following parameters should in most cases
adequately characterize the discharge.

               - TOD                         - Fecal  Coliform
               - BOD (including dissolved)   - Suspended Solids
               - Total Kjeldahl Nitrogen     - pH
               - Total Phosphorus
                                   258

-------
indicate that a few thorough and complete water quality surveys are more
useful  than a year of data which, because of cost, must necessarily be
less intense.

Parameters to be monitored during water quality surveys are of course
specific to a particular project.  It is recommended, however, that
the following parameters be given serious consideration in water quality
surveys:

     BOD5 - total and dissolved
     TOD
     Suspended Sol ids
     Total Kjeldahl Nitrogen
     NOo and N02 by N02 Method
     Total Phosphorus
     Fecal Coliform
     pH
     Settleable Solids - occasionally
     Dissolved Oxygen
     Temperature  (when required  by models)

These parameters are viewed as representing a thorough definition of
water quality for most situations.  The participate nature of storm
generated pollutants does require the careful consideration which this
list of parameters provides.
 Facilities Design Studies

 Studies conducted to develop design  information for treatment systems
 for storm generated discharges must  be directed toward the specific
 problem.  Since these studies are directed toward developing suitable
 unit processes to detain and/or treat potential overflows, sampling and
 analysis procedures must be selected to yield  the data necessary to
 select suitable treatment processes.  Some important points to consider
 are:

     1.  If screening is to be employed, discrete suspended solids
         analyses may be required to determine if solids loading may be
         the limiting factor for screen design during first flush
         conditions.

     2.  What  is the dissolved organic concentration and does it change
         significantly during a storm?  Is dissolved organic removal
         required or can discharge requirements be met with a physical
         process?

     3.  Oxygen demand should be measured by methods other than just the
         BOD^  test since the literature indicates that the BODg of
         storm generated discharges  is only a fraction of the potential
         ultimate oxygen demand.
                                   259

-------
                  SECTION  VIII  -  METHODS  FOR  EVALUATION
                     OF STORMWATER TREATMENT  PROCESSES
Prevention or reduction  of  storm generated  discharges on  a  volumetric or
constituent basis can be accomplished  in  a  myriad of ways.   Included are:

     1.  Land use planning  and  city  planning  to  provide balance between
         pervious and impervious areas.

     2.  Installation of porous pavement.

     3.  Allowances  for  surface storage.

     4.  Street cleaning programs.

     5.  Sewer and catchbasin cleaning programs.

     6.  Street maintenance programs (summer  & winter).

     7.  Use of polymers to increase flow in  the sewers or  to  prevent
         surcharging.

     8.  Flushing of sewers.

     9.  Sewer maintenance  and  rehabilitation programs  to minimize
         infiItration.

    10.  Pumping down of sewers in anticipation  of  a storm  event.

    11.  Separation  of sewers.

    12.  Construction site  erosion control.

    13.  Limiting surface chemical pollutants such  as fertilizers,
         pesticides, herbicides, deicers  and  toxic  compounds (e.g.
         lead i n gasoline).

Each of the above measures  Is  intended to minimize  the  pollutional  aspects
of storm runoff and/or combined sewer  overflows.  These preventative
measures are intended for use  in dry-weather  and as such  are not of  concern
                                   260

-------
in the evaluation of wet weather control  or treatment  of  wet  weather
discharges.  Only those methods which are used to retain  or  treat
potential storm generated discharges wtl1 be considered for  evaluation.
The reader is referenced to the U.S. EPA  report entitled  "Urban
Stormwater Management and Technology:  An Assessment"  (1) for a  thorough
description and application of the following treatment processes and  for
examples of pilot or prototype installations of the processes.


BRIEF INVENTORY OF CONTROL AND TREATMENT  METHODS

In order to develop methods for evaluating treatment processes for
storm generated discharges it is first necessary to understand the
principle of operation and operating variables associated with each  type
of system.  Listed below are many of the  treatment processes  studied  for
storm generated discharge control:

     1 .  Storage

         a.  ln-1ine
         b.  Off-line  (with and without settling)

     2.  Physical and Physical/Chemical

         a.  Coarse screening
         b.  Fine screening
         c.  Gravity sedimentation
         d.  Swirl concentrator
         e.  Tube settling
         f.  Sedimentation-fJ1tratton-adsorption
         g.  Screening/high rate filtration
         h.  Adsorption-coagulation-tube settling-multi media filtration
         i.  Coagulation-tube settling
         j.  Series high rate filtration
         k.  Screening/dissolved-air flotation

     3.  Biological

         a.  Contact stabilization
         b.  Trickling  filter
         c.  Biological contactors
         d.  Lagoons

     k.  Disinfection

         a.  Chlorinatlon
         b.  Sodium and calcium hypochlorite
         c.  Chlorine dioxide
         d.  Ozonation
                                   261

-------
Following is a general  discussion of  how these basic categories are
usually uti1ized.
Storage

Two types of storage facilities  are generally considered for storm
generated discharges.   In-line storage is  an attempt to utilize existing
unused sewer capacity to retain  combined  sewer flow during wet weather.
This is accomplished by controlling the levels in the sewers through a
series of dams, gates,  or flow restrictions, thereby providing storage
by creating backwater.

An extensive monitoring and control  system is required to prevent
upstream flooding conditions.   Depth of flow must be monitored In major
trunk sewers in the system and manual  overrides of the control system
must be maintained.  This type of  system  is  best used where the sewer
grades in the vicinity  of the  interceptor  are relatively flat and where
the interceptor capacity is great.   The initial effectiveness of this
type of system is limited only by  the total  unused volume of the
sewerage system which is governed  by the  elevation profile.  Once the
sewerage system is full and flowing at maximum capacity its effectiveness
is limited by the hydraulic characteristics  of the sewers, the dry-weather
treatment plant capacity and the intensity-duration of the runoff.

The second type of storage facility is "off-line" and includes lagoons,
lakelets, reservoirs, underground  silos or tanks, underwater bags, deep
tunnels, mined labyrinths and  conventional rectangular concrete tanks.
These devices are used  to retain storm generated flows.  A general
flow diagram for storage facilities is shown in Figure Vlll-l.  At a
specified point in the  interceptor  a regulating device is set up so that
flows in excess of dry-weather treatment  plant capacity or in excess of
the downstream interceptor capacity are diverted to the storage
facility.  The diverted flow may be pretreated by screening, a swirl
concentrator or other means to remove gross  solids and by chlorination
to eliminate odor* and  to disinfect the discharge.  Flows in excess of
the storage capacity are bypassed  either  around the facility or through
the tank to take advantage of  the  chlorine contact time and the
sedimentation which may occur  in the tank.  At the end of the overflow
condition when flow subsides in  the downstream interceptor, the contents
of the tanks are bled back to  the  dry-weather treatment plant or to a
separate facility for treatment.  Solids  retained in the storage tank
may be removed at the site or  resuspended  through the use of mechanical
or diffused air mixers  and also  be bled back to the treatment plant.
These facilities are generally designed to accept the maximum hydraulic
flow expected at the overflow  point.  They are limited by the total
available storage capacity at  the  onset of an overflow event.
                                   262

-------
to
O5
CO
         TO DRY WEATHER

              PLANT
                                                                       STORAGE
                                                                                       BYPASS
                                                                                       SCREENED OR SETTLED
                                                                                       'SOLIDS TO DISPOSAL
                                                                                       OVERFLOW
                                   Figure VI I 1-1.   Storage process diagram

-------
The available storage capacity Is determined by the rate at which  the
storage facility can be emptied between storm events.  Pumpback or bleed-
back rates must be as high as possible to prepare the facility for sub-
sequent storms.  However, the pumpback/bleedback rate must be low enough
to prevent downstream overflows and to prevent overloading the dry-weather
treatment plant either hydraulically or on a solids basis.  Pumpback/
bleedback rates on existing or planned storage facilities range from
12 hours to 50 days.


Physical  and Physical/Chemical

Screening -  Screens are used  for  removal  of  gross  solids  and  particulate
matter from storm generated  discharges.   Basically,  screens operate
on the principle that all  solids  with  an  effective diameter  larger
than that of the screen opening will  be  retained on  the  screen and
removed from the flow.   In practice,  some solids  smaller  than the
screen openings are also retained because the mat  of solids  formed  on
the screen surface acts as a  "precoat" and lessens  the  dimensions  of
the screen openings.  Because of  the  formation of  this solids mat,
the screens  must be cleaned  either mechanically or by use of  a backwash
system when hydraulic headlosses  become  excessive.

Screens are classified by the sizes of the openings.  Those with relatively
large openings are used as pretreatment  devices.   These  include bar
screens, coarse screens, and  fine screens.  Their  major  use  in combined
sewer overflow treatment is  for  the removal of large solids and grit for
the protection of downstream equipment.


Types of screens used as the major process for treating storm generated
discharges include fine screens,  and  rotary fine screens.  The fine
screens and the microscreens  are essentially the same device.  The only
difference is the size of the screen  opening.  Fine screens  have openings
that range between 66 and 3323 microns while openings on microscreens
range from 15 to 65 microns  (1).   Screens of this  type are relatively
simple devices.  The screening medium, usually a tightly woven wire mesh,
is fitted on the periphery of a  drum.   The horizontal  drum,  usually 0.9
to 3 meters  (3-10 ft) in diameter and 0.6 to 5 meters (2-15 ft) long (2)
(3) W ,  rotates continuously at  4 to 7  revolutions per mintue.  Wastewater
enters one end of the drum and flows  radially through the rotating screen.
The driving  force through the screen  is  the head between the  inside of
the screen,  and the screened  water chamber,  generally 0.15 to 0.6 meters
(0.5 to 2.0  ft) of water.  Seals  at each end of the drum prevent the
wastewater from escaping around  the ends of the drum into the screened
water.  As the drum rotates,  filtered solids trapped on the screen are
lifted above the water surface.   As they reach the top of the drum the
solids are backwashed from the screen into a sludge trough by high pressure
spray jets.   Operating variables  which may affect efficiency of treatment
by screens of this type include  mesh  size, rotational speed,  backwash  rate,
raw flow rate, suspended solids  mass  loading rate, and particulate charac-
teristics.
                                    264

-------
Rotary fine screens are similar to the microscreen in that tightly
woven wire mesh fabric fitted around a drum is used to strain the waste.
The drum rotates about a vertical axis at high speeds (30 to 65 rpm).
The influent   introduced into the center of the rotating drum, is
directed along horizontal baffles that distribute the flow evenly to
the entire surface of the screen.  Flow passing through the screen is
discharged to  the receiving water or routed for further treatment.  The
reported screen opening sizes ranged from Ik to 230 microns (!)•
Backwashlng to remove the retained solids is done at discrete intervals
by high pressure spray jets on both sides of the screen.


Physical-Chemical Treatment - Physleal-chemical treatment elements
considered for treatment of storm generated discharges  Include clarifi-
cation with or without chemicals, filtration and carbon adsorption.
These three process units are generally operated In series to obtain
the best possible effluent quality.  A schematic diagram of this type
of system  Is contained  In Figure VI I 1-2.  These processes are well
suited to  treatment of storm generated discharges because of their
adaptability to automatic operation,  Instantaneous startup and shutdown,
and their  resistance to shock concentrations or toxic substances.

The bulk of the flotable  and settlcable solids are removed In the
clarification  process.  With the use of chemical aids such as alum,
lime, or ferric chloride, colloidal solids and some dissolved solids
may also be removed  in  this step.  The clarlfier may be one of many
types  Including gravity settling basins, tube settlers or flotation
clarlfiers.  Following clarification, a filter media is used for polish-
ing the effluent by removal of residual suspended solids.  Sand
filters which  may be used  include both gravity filters  and pressure
filters.   Soluble organics remaining  after clarification and filtration
can then be reduced by activated carbon adsorption  In a gravity flow,
pressure flow  or upflow carbon column.

Effluent quality from a system as described  is comparable or possibly
better  than the effluent from a  secondary treatment plant.  Efficiency
of the  system  is not subject to  varying quality of  the  influent flow.
 It  Is  subject, however,  to variability  In hydraulic flow.

 In addition to the multiple unit processes just described, filtration
may be  used alone to treat storm generated discharges.  To obtain  the
best  possible  effluent  from filtration a multimedia filter is preferred
where  the  diameter of the media  decreases with layers  (depth) In  the
filter.  This  Is generally accomplished through the use of coal
 (anthracite) as the  less dense larger diameter particles  for  the  top
layer and  sand as the more dense smaller diameter particles for  the
bottom  layer.  With  this type filter, solids capture  is distributed
through the bed rather  than just at  the surface.  For use on  storm
generated  discharges the filter  should  be preceded  by a fine  screening
                                   265

-------


CHEMICAL
ADDITION


COAGULATION/
FLOCCULATION


                                               BACKWASH
to
OS
OS
EFFLUENT
                                                                                           FILTER
                                                                                          BACKWASH
                                                     SPENT
                                                     CARBON
                                     SLUDGE
                                    DISPOSAL
BACKWASH
  WATER
                                       Figure VI I 1-2.   Physical-chemical process diagram

-------
device to prolong filter runs.  The filter which is operated on a
gravity basis should be equipped with a backwash system utilizing air
scour to break up mud balls and to redistribute the filter media.  The
system must also be provided with either storage or treatment for the
backwash.  The backwash system  is Initiated when the headless through
the filter bed Increases above a predetermined level or when breakthrough
occurs.
Dissolved Air Flotation - The process of dissolved air flotation is
based on the principle that air  is soluble in water proportional to
the total pressure of the air in contact with the water.  When the
pressure on water, saturated with air at an elevated pressure, is
suddenly  reduced to atmospheric pressure, the dissolved air is
released from solution In the form of minute air bubbles.  These air
bubbles become attached to particulate matter or are enmeshed within
a floe structure; the net result being that the bulk density differen-
tial between the air-solid structure and the water medium is greater
than between the initial solid and water, but in a reverse direction.
The solids are allowed to float  to the surface of the flotation tank
where they are removed by overhead scrapers.

The amount of air introduced  into the system is a function of the
efficiency of saturation, the amount of flow pressurized and the
operating pressure.  Pressurization flow with combined sewer overflow
is generally from one of two sources.  Either a portion of the raw flow
is pressurized  (split flow) or a portion of the effluent Is pressurized
(recycle flow).  The minimum amount of pressurized flow which is
effective is 20% when calculated as the ratio of pressurized flow to
unpressurIzed flow.  Most flotation units operate at gage pressures
of 2.8 to k.2 kg/sq cm  (AO-60 psi).  Operating parameters which may
affect operation of the units include chemicals used, chemical dosage,
percent pressurized flow, surface loading rate  (both hydraulic and
solids) and operating pressure.

Dissolved air flotation is effective for removals of suspended solids,
fine flotables, oils and grease, and, when used with chemicals,
colloidal solids.  A schematic diagram of the flotation process
utilizing split flow pressurIzatIon is shown in Figure VI 11-3.  Screening
Is usually prerequisite to prevent materials of high specific gravities
from settling to the bottom of the flotation tank.
 Biological  Systems

 Biological  systems  by  their very  nature must  be operated continuously
 to maintain a viable biomass or be able to  borrow  the biomass from a
 system which does operate  continuously.   For  this  reason they are
 generally  located at or  near existing  dry-weather  treatment plants.
                                    267

-------
                            SCREEN
                           BACKWASH
to
O5
00
                                          STORAGE
                                         CHEMICAL
                                         ADDITION
                                        FLOCCULATION
                                        PRESSURIZED
                                           FLOW
                   SCREENED SOLIDS
     DISSOLVED AIR

    FLOTATION UNITS
                                                                                             EFFLUENT
TOP FLOAT AND
BOTTOM SOLIDS
                                                                               -*> DISPOSAL
                                  Figure VI I 1-3.  Dissolved air flotation process diagram

-------
Two unit operations are common to all systems - a.unit where the waste
is contacted with the biological solids for assimilation of the organics
and a sedimentation basin for separation of suspended solids, including
the biological sol ids.
Contact Stabilization - A schematic diagram of the contact stabilization
process integrated with an existing dry-weather treatment plant is
shown in Figure VI I 1-^.  The process basically consists of two
aeration chambers and a clarifier.  In the first aeration chamber
the biomass or aerated return sludge is contacted with the combined sewer
overflow to be treated.  The particulate organics and a portion of the
soluble organics are adsorbed on the biological solids.  After a short
period of aeration, 15 to 30 minutes, the solids are separated from
the liquid in the clarifier.  Solids separated in the clarifier are
returned to the second aeration tank for reaeration or stabilization.
This reaeration time usually lasts from one hour up to many days.

Because this system requires a ready supply of activated sludge, It can
only be operated at an existing treatment plant utilizing a biological
treatment process.  Operating parameters which may affect the efficiency
of the process Include sludge stabilization time, food to microorganism
ratio, contact time, and sludge reaeration time  (5)-
 Trlckling Filter - In  this  process, a biological growth is supported
 on a  stationary medium and  the wastewater  is distributed over the surface
 and allowed  to flow through  the media.  Organic matter is absorbed
 from  the waste in  this process and converted into new biomass, gases and
 energy.  Portions  of  this  biological growth are constantly sloughed
 off of  the medium  by  hydraulic scouring.   For  this reason the discharge
 from  a  trickling filter  must  be settled  in a clarifier.  For dry-weather
 operation, a  once-through  system  is generally  not sufficient to accom-
 plish the desired  levels of  removal so most systems are provided with
 reci rculation.

 Operating parameters  which  may affect the  efficiency of these units
 include surface area, recirculation rate and surface loading  (both
 hydraulic and organic).


 Rotating Biological Contactors -  The operating principle of these units
 is  identical  to that  of  trickling filters. But  instead of the biological
 growth  supported on a stationary  medium,  it is supported on large
 diameter  (3.1 meter  [10  ft]), closely spaced rotating discs which are
 partially submerged  (50%)  and rotated at  low speed  in a tank containing
 the wastewater  (6).   Excess  biological growth  sloughs off of the discs
 and must be  separated in a  clarifier.  Both trickling filter and
 rotating biological contactors must be in  use  continuously to maintain
 a viable biological growth.
                                    269

-------
to
-d
o
                                                      CONTACT
                                                       TANK
                        GRIT REMOVAL
  SLUDGE
STABILIZA-
 - TION
                                                                              SLUDGE FROM DRY-
                                                                               WEATHER PLANT
                                                                                      RETURN
                                                                                      SLUDGE
                                                                1
                                                              WASTE
                                                             SLUDGE
                                  Figure VI 11-4.  Contact stabilization process diagram

-------
Operation variables which may affect the efficiency of this system
include  surface area submerged, rotational speed, loading rates
(hydraulic and organic) and the number of units in series.


Lagoons - Three types of treatment lagoons have been used for storm
discharges and combined sewer overflows.  These are oxidation ponds,
aerated lagoons and facultative lagoons  (1).  All  are, in effect,
biological treatment systems.  Oxidation ponds and aerated lagoons are
aerobic processes with the difference being the source of oxygen.
Oxidation ponds, which are generally very shallow, rely on surface
reaeration and algal oxygen production to maintain aerobic conditions
for biological degradation of organics.  Sedimentation also occurs in
the oxidation ponds.

Aerated lagoons, being deeper than oxidation ponds, rely on artificial
means  to maintain aerobic conditions  in addition to surface reaeration.
These  artlcifical means may consist of mechanical  surface aerators,
turbine aerators or diffused air systems.

Facultative  ponds are  the deepest of  the three  (or as deep as aerated
lagoons) but  do not have artificial means of providing oxygen.  In this
way the ponds are allowed to develop  two distinct types of treatment,
aerobic near  the surface due to algae and surface oxygen transfer, and
an anaerobic  zone near the bottom sludge deposits.

Generally all types of lagoons  require some form of post treatment for
removal of suspended solids and/or algae.  Loadings to the lagoons are
generally reported  in  terms of  g BOD/day/sq m  (Ib BOD/day/acre).
 Disinfection

 Almost  all  storm  generated  discharge  treatment or storage systems
 provide disinfection of  some  sort.   In addition, the chemical used for
 disinfection may  also  be used for odor control.  The various types of
 disinfection  involve the use  of chlorine gas, sodium or calcium
 hypochlorlte,  chlorine dioxide or ozone.
 EVALUATION  OF  TREATMENT  PROCESSES

 Process  Efficiency

 Evaluation  of  a  treatment  process will  depend on whether the treatment
 system is a prototype  installation or a  demonstration  (pilot) unit.
 When  evaluating  a  prototype  system,  the  total pollutlonal reduction at
 the overflow site must be  determined.  This  includes evaluation of the
 effluent from  the  treatment  system as well as the  bypass around the
 unit  and the discharge at  upstream and downstream  points.  Also included
                                    271

-------
must be the volume and quality of residuals which are pumped back to
the dry-weather treatment plant and their possible effect on the
efficiency of that plant.  When evaluating demonstration or pilot
units the efficiency of the treatment process is of specific concern.
The purpose of a pilot unit is to determine whetheror not the process
will work on the particular waste and to optimize the process variables
for most efficient operation.

A generalized flow diagram has been developed that is applicable to all
storage/treatment schemes.  It is shown In Figure VI11-5.  Sewage
generated during a storm event flows through the combined sewer 1 (all
numbers refer to Figure VI11-5) until it reaches a flow regulating
device.  Flow In excess of that which can be handled  hydraulleally in
the downstream sewer 2 or interceptor Is allowed to enter an outfall
sewer 3.  The storage/treatment system located at this overflow Is
generally designed to meet some maximum design flow rate or volume.
Any flow in excess of the design rate or volume Is bypassed through
another sewer k to the receiving stream.  Flow to the treatment system
5 will result in two separate flows; the treated effluent 6 and the
residual sludge 7 and/or stored flow 9 which is pumped back to the dry-
weather treatment plant after a storm event.  Flow which goes to
the receiving stream may be combined Into a single stream 8 consisting
of treated effluent and treatment bypass.

When evaluating a treatment system various types of efficiencies may
be determined.  These Include volumetric efficiency (how much of the
total overflow is treated or retained?), overall efficiency (how much
of the contaminant (s) is prevented from overflowing to the receiving
stream?) and treatment efficiency (what portion of the contaminant(s) is
removed in each unit process of the treatment scheme?).  Both volumetric
efficiency and overall efficiency must take Into account all flows
bypassing the storage/treatment unit for a study approaching the
overall drainage area.

Volumetric efficiency requires only measurement of total flow In the
discharge at the outfall 3 and In the bypass k.   It is a measure of the
percentage of overflow which has been treated In the  facility.   A
graphical  representation of the meaning of the volumetric efficiency
is illustrated in Figure VI I 1-6.  The curve Is a recording of flow
at 3-  The shaded portion of the area under the curve (area labeled C)
represents the bypass volume (all flow In excess of the design flow rate
On) as measured at k.  The ratio of area B to the total area under the
curve (B+C) Is the volumetric efficiency.  Additional  valuable Information
can be obtained from this plot.  The various ratios that can be
developed and the significance of each is listed In Table VI I 1-1.  These
ratios are useful when an evaluation of system design Is attempted or
when alternatives for Improving efficiency are investigated.  Among the
alternatives which can be studied are, for example:
                                   272

-------
                                                        TREATMENT RESIDUALS
                                                               (SLUDGE)
   COMBINED
    SEWER

         1
 FLOW
REGU-
LATOR
to
-a
CO
                                         SYSTEM
                                       PUMPOUT
                                                 10
                            3
                            Q,
                               OVERFLOW
                                     10
TREATMENT

 SYSTEM
-•i  SYSTEM
    EFFLUENT
                                                          SYSTEM BYPASS
                    TO  DRY WEATHER
                    TREATMENT PLANT
                                                               TO RECEIVING
                                                                  STREAM
                                            Figure  VI I 1-5.   Generalized combined  sewer
                                                     overflow treatment  system

-------
                     B
Figure VI I 1-6.   Graphical  illustration of a hypothetical
               flow rate  to a treatment system
     Q.L
   Figure VIII-7-   Graphical  illustration of a hypothetical
                 flow rate  to a  storage tank
                             274

-------
    Table Vlll-l.  RATIOS DEVELOPED FROM INFORMATION IN FIGURE VIII-6
     Ratio
                                    Significance
£
B
                    Total volume bypassed
bypassed volume
treated volume
Also equals the stor-
age capacity required
to provide treatment
for 100% of the
overflow
Ratio of untreated to
treated flow
A + B + D
  A + D
A + B + D
                    bypassed volume
                    total overflow
                    volume treated
                    total overflow
volume treated
total capacity
	unused capacity	
total treatment capacity
Equals the portion
of the total flow
which does not
receive treatment

Equals the portion
of the total flow
which receives
treatment

Portion of the amount
actually treated to
the maximum poten-
tial treatment
capacity
Portion of capacity
unuti1ized
  D_
  C
V
volume treated
unused capacity

unused capacity after a bypass
       volume bypassed
    time of bypass
total time of overflow
    - t,) +  (t2'ti)  total  time without  bypass
    —•	   total time of overflow
Ratio of used to
unused capacity

Indicates the need or
effectiveness of stor-
age and/or flow
attenuation

Portion of time that
design flow  (0_n.) Is
surpassed

Portion of t ime
operating at below
design rate
                                    275

-------
     1.  How much storage capacity is  required  to provide treatment of
         100% of the overflow?

     2.  What volumetric efficiency can  be achieved  by the addition of
         limited storage (in-line or off-1!ne)  or by increased treatment
         capacity?

For storage systems a similar  type of  volumetric efficiency can be
determined utilizing a curve such as that found in Figure VI I 1-7.
To obtain a plot similar to that shown in Figure VII 1-7,  only  one  flow
recorder is necessary - In the  total discharge  stream 3-   The  initial
storage capacity available must be known, so that only the flow in
excess of this volume is counted as bypass  or overflow.   The total area
under the curve (A+B) represents the total  volume of storm generated
discharge at the storage site.   If the first discharge from the tank
or bypass around the tank occurs at t],  the volume A represents the
total amount captured (or the  available space at the start of  the
overflow) and the volume B is  the total  tank overflow or  bypass.
Several informative relationships may  be obtained from a  plot  of  this
type.  For example, A/(A+B) Is  the portion of the total overflow detained
by the storage tank.  QT represents the  design  rate for any treatment
system which must be added for  100% capture and treatment.  B  represents
additional storage capacity needed for 100% storage.   Also, If the
treatment system Is started at  the beginning of the  discharge  the
effective storage volume is greater and  the required design treatment
rate, Q-y, may be less.

The development of a "design"  curve such as those found  in Figures
VI 11-6 and VII 1-7 is up to the discretion of the design engineer when
designing either a new facility or when  contemplating expansion of an
existing facility.  In cases where new facilities are being constructed,
empirical methods can be used  to develop a design hydrograph.   When an
existing facility has been In  operation  the actual operating hydrographs
can be utilized, or as in the  case of  new facilities, an  empirical
hydrograph can be developed for the desired recurrence interval.   The
choice of the design storm will be a function of the regulatory con-
straints and/or available funds and/or receiving water quality
objectives.

Another type  of efficiency to  be considered is the overall efficiency.
Unlike volumetric efficiency, which considers only the quantitative
data of how much flow was treated or stored, the overall  efficiency is
concerned with determining the  amount  of pollutants  which are  prevented
from being discharged to the receiving water.  An analysis of  this type
requires both measurement of the volume and sampling of the various
flows.  To determine the overall efficiency of  the system, a material
balance between the point of overflow  3  and the receiving stream 8 is
all that is required.  No information  is required concerning the
process parameters or other operating  variables of the equipment.   This
"black box" approach will yield all information necessary to determine
                                  276

-------
operating efficiency for either the individual treatment process or
for the entire system.

In order to determine the efficiency of a system it is recommended that
samples be taken at constant time  intervals and when composited, they
may be proportioned either according to the flow rate at the time the
sample was taken or according to the volume of flow since the last time
the sample was taken.  Either of these methods will provide a
sufficiently representative sample.  A third method of sampling and
compositing that is sufficient  is  to take samples of equal volume at
equal  intervals of flow volume.  However, this type of sampling will be
variable.  The ideal method of  sampling would be to draw a sample
continuously, with the flow of  sample being proportional to the flow
rate.  However, an automatic sampler of this type with proven field
reliability for storm generated discharge studies, may not be practical
to  require at this time.  A complete discussion of sampling can be
found  in Section IV and a discussion of compositing methods is
contained in Reference 7-

In  determining the Interval between samples, especially when using
automatic samplers, the duration of the average treatment run should be
estimated and divided by the number of samples available.  For example,
if  the average run is estimated to be k hours and 2k samples are
available, the sampling  interval will be 10 minutes.  For runs that last
longer than the average the samples can be removed and replaced with
new bottles and the sampler reprogrammed.  In the above determination it
is  assumed that the median run  will be less than the average.  Also,
samples should not be proportioned until after a run is complete, since
in  the case of short duration runs, a large amount of each sample will
have to be taken to provide a composite quantity large enough for
analysi s.

For compositing done at equal volumes of flow with a variable time
interval, the volume  interval chosen for taking the sample should be
equal  to the average total volume  treated divided by the number of
sample jars available.  For example, If the average volume of discharge
treated is 18,925 cu m  (5 million  gal.) and 2k samples are available,
one sample would be taken every 788 cu m (0.21 million gal.).  In this
case again, if it appeared that more than 2k samples would be taken,
the sampler could be reprogrammed.

It  should be noted that when the process efficiency variability of
certain treatment processes, specifically physical-chemical, is to be
determined, the constant volume variable time Increment sample is
recommended.  The reasoning for this type of sampling is that with
these  systems there is a more direct time dependent relationship between
the influent and effluent quality  than In a biological treatment process
for instance.  Figure VIM-8 contains hypothetical plots of an influent
flow and the resultant quality  from a physical or chemical treatment
system and from a biological treatment system.
                                   277

-------
  MASS
 INFLUENT
 LOAD ING
                                  FLOW
                            TIME
                                          -*-
•¥
 i
   MASS
 EFFLUENT
 LOAD ING
                  PHYSICAL OR CHEMICAL  SYSTEM
J
1
1
1
1
\
1
1
1
1

1
1
1

1
1
                            BIOLOGICAL  SYSTEM
MASS
EFFLUENT
LOADING
^, 	 -5K —
v 	 X — ^ ^**.^
^ 	 ^^ . v
Y f
i i
! !
                             TIME

Figure VI I 1-8.   Hypothetical  mass loading on two different types of
  treatment  processes and the resultant effluent quality pattern
                              278

-------
Although the area under the two effluent curves is identical  (same
mass In the effluent), the constant time interval  sampling will  miss
the effluent breakthrough from the physical or chemical  system,  but  the
biological system with more process mixing and damping effect will  show
the effects of the poorer effluent quality to a less degree but  for
a longer period of time.  Thus, although both effluents  would be of
equal quality the physical chemical system would result  in samples of
higher quality.  Therefore, it is recommended that when  sampling
programs for process efficiency are designed, it is important that a
great deal of thought be given to the type of effluent variability that
can be expected, and the sampling programs chosen to correspond  to
these needs.

For almost all types of systems, only two  sampling and flow measuring
locations are needed.  These are at the point of overflow 3 and  at the
effluent from treatment 6.  The flow rate  at A can be calculated as  the
difference between the flow rates at 3 and 6, and the quality parameter
concentration at k will be the same as at  3.

To determine the total mass of a contaminant removed, a  single composite
of the  Influent and effluent including bypass is necessary.  With this
type of sample the determination of overall efficiency is relatively
simple  for any parameter.

                          V6(c3-c6)
                    E  =  —_	  x  100
                             33           (subscripts refer to Figure  VI I 1-5)

          where:  En = overall efficiency  (%) for a specified parameter

                  V, » total volume of discharge at the overflow point

                  V^ = total volume of effluent from treatment
                   b
                  C, = parameter concentration at the overflow point

                  C,- = parameter concentration  in the treatment effluent


Any volumetric or concentration units may  be used as long as they are
consistent throughput.  The overall efficiency can be calculated using
the above equation for any parameter.  The recommended parameters for
study are discussed in Section VI.

The process efficiency for the treatment system alone can be calculated
as :
                                   (crc,)
                             E  » —£——  x 100
                              P     £3

The total loading on the receiving body of water can be simply calculated
as follows:
                                     279

-------
                       LT  *  V6C6  +  
                 J|t=tj        teit2              ts>tn
              (vc,
               y Jl
                      .
                      I
where t » t.    » the first time Increment

      t » t     « the last time increment
           n


However, the main purpose of the discrete analysis Is to determine the
mass loading on a treatment process, the variance in treatment efficiency
as a function of mass loading, and the effluent mass loading to a
receiving stream as a function of time.  It is recommended that for
each time increment that the actual  sample be taken in the middle of the
time increment.  For example, in the above example of efficiency
determination the samples are taken  at t r, t. r, t0 r, etc.
                                        0   I . p   i o
The following is an example of how the above type of analysis can be
performed.  Hypothetical flow and quality data for one storm event is
presented in Table VI 1 1-2.  The data Is analyzed for a dissolved air
flotation unit utilizing split flow, with a surface area of 185 sq m
(2,000 sq ft).  The samples were taken as described above.  The raw data
obtained during operation of the unit is contained In columns I to 4 of
Table VI 1 1-2.  A flow diagram is constructed by determining the average
flow rate during each time Increment, i.e.,
                                   280

-------
                    Table  VI11-2.  HYPOTHETICAL  DATA  ANALYSIS  (DISSOLVED-AIR FLOTATION SYSTEM)
to
CD
Time
No.
1
2
3
4
5
6
7
B
9
increment t
mm
(0-10)
(10-20)
(20-30)
(30-40)
ClO- 50)
(50-60)
(60-70)
(70-80)
(80-90)
Volume
In the Suspended solids, Average flow
increment, Raw,
cu m
94.6
302.8
246.0
151.4
159-0
132.5
131.6
37.8
30.3
(ga IT) mg/l
(25000) 200
(80000) 500
(65000) 400
(40000) 450
(42000) 200
(35000) 150
(30000) 75
(10000) 75
(8000) 75
Effluent, rate,
mg/l
75
300
250
200
100
70
50
40
40
I/sec
158
505
410
252
265
221
189
63
50
; (gp"0
(2500)
(8000)
(6500)
(4000)
(4200)
(3500)
(3000)
(1000)
(800)
Loading rate,
1/min/sq m
(gpn/sq ft)
51 (1.2)
163
132
81
85
73
61
20
16
(4.0)
(3-2)
(2.0)
(2.1)
(1.8)
(1.5)
(0.5)
(0.4)
kg/day/sq m
(Ib/day/sq ft)
14.6
117-1
76.1
52-7
24.4
15.6
6.8
2.2
1.8
.(3.0)
(24.0)
(15-6)
(10.8
(5.0)
(3-2)
(1.4)
(0.45)
(0.36)
(Influent) (Effluent) Treatment
kg/min
(ib/min)
1.9 (4.2)
15.1 (33-4)
9.8 (21.7)
6.8 (15.0)
3.2 (7.0)
2.0 (4.4)
0.86 (1.9)
0.29 (0.63)
0.23 (0.50)
kg/mi n Efficiency,
(Ib/min) %
0-72
9-1
6.2
3-0
1.6
0.91
0.54
0.15
0.12
(1.6)
(20.0)
(13.6)
(6.7)
(3.5)
(2.0)
(1.2)
(0.33)
(0.27)
62.5
40.0
37.5
55.5
50.0
53.7
33.3
46.6
46.6

-------
                    total  flow  during  time  increment
total  time
rng
duri
                                     ng  Increment
These values are shown  in  Column  5  and  plotted  in  Figure  VI 1 1-9.
1
(J
I/I
-v.
o
_1
LL.


631
(10000)
505
(8000)
379
(6000)
252
(4000)
126
12000;


~
-
-




L— -L_^_^
AMOUNT PRESSURIZED
1 	 ' 	
            0    10    20   30    40    50    60   70    80   96
                                   TIME, min
       Figure VIII-9.  Flow diagram (total flow to a treatment unit)
Loading diagrams for flotation units are calculated on the basis of
both flow and solids.  The hydraulic loading rate (1/min/sq m [gpm/sq ft])
is based on the flow to the units.  A tabulation is presented in column
6, Table VI I 1-2 and plotted in Figure VI I 1-10.  As can be seen, the
shape of this plot is identical to the flow diagram.  However, if the
percent recycle (ratio of pressurized flow to unpressurized flow x 100)
is plotted on the same graph, the varying conditions of operation will
be evident.  In the example it is assumed that 25% recycle is used based
on a design total  loading of 163 1/min/sq m (4.0 gpm/sq ft) (101 I/sec
[1600 gpm] pressurized, 405 1 sec [6400 gpm] unpressurized).  With
flotation units the quantity pressurized is generally held constant at
the design flow (i.e., 101 I/sec in the example).

Solids loading expressed as kg (Ibs) of suspended solids per day per
sq m (sq ft)  of surface area is calculated (column 7) and plotted in
Figure VII 1-11.  Pollutographs, or mass per unit time, can be constructed
for flow to the unit and effluent from the unit. The  effluent  pollutograph
                                   282

-------
203.5
(5.0)
162.8
0».o)
HYDRAULIC
LOADING,
i/ • / 122-]
1/min/sq m /, M
(gpm/sq ft) ^'°}
to
S 81 .A
(2.0)
kO.J
(1.0)

-
-
.. DESIGN RATE
TOTAL
^RESSURIZATION
I
_j
r
I, J H

r j
i 	 L J
i
i
, 	 J
~* DESIGN RATE ' 	
i 	 1 	 1 	 1 	 1 	 _ —
          10      20     30      AO      50
                             TIME, minutes
60     70     80
                                                                      i 150
                                                                      H 100
                                                                      -I  50
90
                              RECYCLE

                             PERCENTAGE
Figure VI11-10.   Hydraulic loading  rate  to a  flotation unit  (split flow)
                indicating change in  recycle  percentage

-------
O U_
z
—  O"
o  in
< >s.
o  >•
-I  <0
  •o
t/1 V,
0 -Q


O
(^  6

ui  cr
o  tn
< \
U.  X
CC.  ID
•^ -o
«^ '^
   O)
  -*
    122

    (25f


    98
    (20)


    73
    (15)
    (10)


    24
    (5)
         Figure VI11-11
                           TIME, minutes


                        Solids loading  as  a function of time
     13.6
     (30)


     11. A
     (25)
UJ \
o o>
   6.8
  (15)


   4.5
      2.3
     (5)
                            i	
             i
             i
       i	1
              10
20
30    AO     50    60
   TIME, minutes
                                                70    80     90


Figure VI 1 1-12.   Pollutograph for suspended solids  (loading vs  time)
                                  284

-------
is the same as the receiving water body loading.  The incremental loading
for influent and effluent streams are contained in columns 8 and 9
of Table VI 1 1-2 and plotted  In Figure VI 1  1-12.

Efficiencies for each time  increment are  calculated  (column 10) and
plotted in Figure VI 1 1-13,  As can be seen  in the plot, the efficiency
remains relatively constant over a wide range of operating conditions
even though loadings vary significantly.

With plots such as those shown In these examples comparisons of different
storm events or different treatment methods are simplified.

The efficiencies described  in the preceding discussion should be calcu-
lated for each individual storm event (i.e. the volumetric efficiency
and the overall treatment efficiency for  each constituent).  Sufficient
storm events should have the samples analyzed separately and the
incremental efficiencies calculated to determine the variability of the
parameter concentration and effect of treatment at various loading
conditions .

The efficiency for an entire year  (or longer  if possible) of operation
should also be determined.  This should be  expressed as both the
average efficiency, and as  the weighted efficiency.  The average
efficiency  is the sum of the individual efficiencies divided by the
number of values :                   "  _
                      *        i-
                      Average  EQ
This efficiency can  be calculated  for  both the volumetric efficiency and
the treatment efficiency.  This value  only indicates the average
efficiency of a series of  runs.   It does  not  indicate the amount removed
during  the same series of  runs because of the differences in volumes
treated and differences  in concentrations.  The weighted efficiency must
be used to determine the percentage of the total volume treated or the
percentage of each constituent removed on a yearly basis.  The weighted
average is calculated on both a total  volumetric basis and on a total
kg (lb) basis.

Volumetric efficiency can  be determined using the following relationship.
                                               n

                                              n-1 V6
                      Average Volumetric  EQ » ~	  x 100

                                              n-1  3

          where V, « total volume  to treatment or storage for each
                     overflow event

                V, - total volume  of discharge for each overflow event
                      (subscripts  refer to Figure VI I 1-5)
                                    285

-------
to
CO
                         100
                          80

                 EFFICIENCY
                 PERCENTAGE

                          60
                          20
                              0     10      20     30      AO     50     60      70     80     90

                                                          TIME, minutes
                                      Figure VI 11-13.  Overall efficiency at each time increment

-------
Similarly, the weighted overall efficiency for each waste constituent
is calculated using the following (bypass streams Included):
                   Weighted Average En - ~	  x 100
                                         E    V?C*
                                         n=l   * *


If only the treatment process efficiency is desired the following
equation applies:

                                          n

                                            ,6   36
                   Weighted Average E  - —	
                                     p      n
One aspect of the overall efficiency that has generally been ignored is
the additional  loading on the dry-weather treatment plant when it
receives residuals from the storm flow treatment or storage systems.
Off-line storage tanks retain large quantities of the storm generated
discharges.  During the first flush and possibly during extended over-
flow events the strength of the retained flow is greater than that of
normal sewage.  With some storage systems (flow-through types) additional
solids are captured by utilizing the tank as a sedimentation basin.
Following a storm event when flow in the interceptor has subsided, the
retained flows are pumped or bled back to the dry-weather treatment plant.

Unless 100% of  the retained contaminants are removed at the treatment
plant, they cannot be considered to have been completely removed from
the overflow.   Because of the strength of the waste to be pumped back
and the added load to the solids handling capacity of the treatment
plant, the efficiency of the plant may actually be reduced.  With a
small stormwater installation the effect on a dry-weather treatment
plant may be negligible.  However, if the storm generated discharge for
an entire city, drainage basin or sewerage district is retained, the
volumetric and  pollutional loading may be significant.  The problem may
be as severe with the disposal of sludges from stormwater generated
treatment systems.

In a well run treatment system ft is generally attempted to obtain as
concentrated a  residual as possible to minimize storage space for  the
residuals.  Unlike many of the pilot unit or demonstration units  in
existence where the residuals are put right back  in the sewer, any full
scale system must provide storage for the solids generated or removed.
These solids must be treated in one of two ways, either by on-site
dewatering and  disposal or by pumping or bleeding back to the dry-
                                   287

-------
weather treatment plant.   If the latter, the shock load on a treatment
plant by a slug of the high concentration waste may cause an upset or
decrease efficiency of the plant.   Although at the present time there fs
no simple way to predict  or measure the effect of treatment residuals
on the operation of a treatment system, both the volume and strength
of the residuals must be  recorded.   Thus, when evaluating a stormwater
treatment system, careful attention should be paid to the resultant
effect on the existing dry-weather  treatment plant.
Appl1 cat ion Data

The general purpose of storm flow treatment process projects at the
present time is to determine the applicability of various processes to
the treatment of these discharges and to determine their effectiveness
for removal of pollutants.   However, sufficient information should also
be provided about the system and Its application in one location to
enable engineering and economic analyses to be performed for application
in a different location.   A uniform method of reporting results should
be initiated.  General information, not necessarily required for deter-
mination of process efficiency for each Individual  project, but essential
when trying to determine process applicability to other sites,  should
be reported.  This data includes:

     1.  Rainfall data
     2.  Drainage area description
     3.  Sewerage system data
     4.  Physical description of the treatment system
     5.  System operation
     6.  System costs
     7>  Sludge handling facilities
     8.  Dry-weather treatment plant data
     9.  Combined sewage data (for combined systems only)

The general information required is the same for all  types of treatment
systems.
Rainfall Data - Each storm event treated should be described.   Rainfall
intensity and duration in the study area should be reported.   Differences
in the rainfall history may affect the quality and quantity of discharge.
Intensity should be calculated at 5 to 15 minute intervals so  that a
rainfall hyetograph may be constructed.  If more than one raingage is
used In the study area the location of the raingages and the method
of calculating the complete area rainfall  intensity should be  stated.
From this information and the rainfall duration the average intensity
can be calculated.  Antecedent rainfall should also be reported (i.e.,
dry-weather days prior to the event studied).   This may also affect the
quality of the overflow because of the buildup of solids In the
sewerage system.  Also, the rate and volume of runoff may be affected
                                   288

-------
by the antecedent rainfall,  in summary, the data should include the
following information:

     1.  Rainfall intensity  (cm/hr [in./hr] at 5-15 minute intervals).
     2.  Rainfall hyetograph (see Sect ion V!I).
     3.  Rainfall duration.
     k.  Average rainfall  intensity (cm/hr  [in./hr]).
     5.  Location of raingages.
     6.  Method of calculating intensity.
     7-  Antecedent dry-weather days.

This data will be useful  in  determining whether the rainfall  event is
typical for the study area and comparable to rainfall patterns in other
locations of  interest.  A complete discussion of rainfall  data can be
found  in Section V.
Drainage Area Description - The type of area contributing to the runoff
will affect both the quantity and quality of the storm generated
discharge.  Data to be reported includes:

     1.  Total area contributing to wet weather flow  (ha [acres]).
     2.  Topography.
     3.  Land use and population density.
     k.  Imperviousness.
     5.  Runoff coefficient.
     6.  Specific curb length  (m/ha [ft/acre])
 Sewerage System Data - The physical size and characteristics of the
 sewerage system can affect the amount and characteristics of the
 discharges.   In a combined sewer  the characteristics of the overflow
 will be dependent to some extent  on the time of day that the overflow
 occurs because of the diurnal variation of both the dry-weather flow and
 sewage constituents.  The diurnal variation in the sewer may take on
 more significance when consideration is given to the pumpback of stored
 flow or treatment residuals.   Information which should be reported
 i ncludes:

     1.  Physical size of the main trunk and  interceptor sewers.
     2.  Design flow capacity.
     3.  Number of discharge points in the contributing area.
     k.  Type of flow regulating  device.
     5.  Type of discharge conduits.
     6.  Dry-weather flow variation (for combined systems only) and
             infiItration etc.
     7.  Dry-weather sewage constituent variation  (for combined systems
            only).
                                    289

-------
Physical  Description of the Treatment System - A complete description of
the treatment and/or storage system should be reported.   For many
applications the physical  size of the system may be a determining factor
for selection.  The following information should be included:

     1.  A plan view of the system.
     2.  Physical  dimensions of each piece of equipment  and tanks.
     3.  A list of all  appurtenant equipment.
     k.  Design hydraulic  and constituent loading rate for each piece
            of equipment.
     5.  Chemical  feed  rate and storage capacity.
     6.  Total land requirements.
     7.  Laboratory equipment requirements at the site.
     8.  Description of flow measurement and recording equipment.
     9.  Description of sampling equipment and sampling  points.
    10.  Description of provision for sludge storage and/or handling.
System Operation - Information to be reported  for  system operation is
of two types.   General  information Including  schematic diagrams,
process and instrumentation diagrams, and maintenance programs  is  the
first type.  Specific information on each individual  type of system is
the second.  The former is summarized below and  the latter Is described
for each type of treatment system later In this  section.

The complexity of the system operation will become evident with  the
following information.

     1.  Schematic diagram of the system.
     2.  A proces's and  instrumentation diagram.
     3.  A narrative description of  the system operating including
            controls.
     k.  Startup and shutdown procedures.
     5.  Maintenance schedule.
     6.  Description of sampling procedures.
     7-  List  of sample analyses and methods.
System Costs - A complete breakdown of system costs,  as  installed,
should be included.   The capital  and construction  cost  breakdown should
include individual  costs for  the  following.

     1.  Engineering costs.
     2.  Land.
     3-  Si te preparation.
     k.  Purchased  equipment  (pumps, mixers,  etc.).
     5.  Installation and erection.
     6.  Concrete.
     7-  Piping.
     8.  Electrical
                                  290

-------
     9.  Instrumentation and controls.
    10.  Building 1ncluding utllItles.

This list may not be all-inclusive for every type of system.   Other
capital costs required to render the system operational  should also  be
reported.  These costs are reported at the total  dollar cost,  including
extras and any escalation clause costs, at the time of construction
completion, and dated such.  For example, if a bid were accepted  at
a cost of 1.0 million dollars on January 1, 1975 and construction was
completed on January 1, 1976 at a cost of 1.1  million dollars, the
later date and costs would be used.   For true comparison of capital
costs the dollar value must be adjusted for present worth.   This  can
only be done by using one of the cost  indices.  The Construction  Cost
Index of Engineering News Record  is recommended for use.

Annaul expenditures costs are much more difficult to report on
the same basis for comparative purposes.  The annual costs  consist of  some
charges which are expended whether or  not the system operates  (amortiza-
tion of capital costs, routine maintenance, insurance, etc.),  some are
dependent upon the duration of operation (labor and power), some  are
dependent upon the rate of treatment or loading to the units (chemicals
and power), and some are dependent upon the number of times that  the
system operated  (laboratory analysis,  residual handling, system cleanup,
data analysis).  Each of these categories should be reported separately
and their basis explained.  Actual costs experienced for the treatment/
storage system and a theoretical  basis for comparative purposes should
both be reported.

For the actual annual costs the following  information and bases should be
supplied on a yearly basis.

      1.  The number of storm events during which operation occurred.
     2.  The total volume stored  or treated during the year and the
            pounds of the appropriate contaminants removed.
      3.  Amortization costs and the basis for their calculation
             (percentage rate and  length of payment).
     4.   Insurance cost.
      5.  Total number of man-hours expended and the pay scale for
            a.  routine maintenance
            b.  operation of the  system
      6.  Total chemical cost including the type of chemical, chemical
            dosage and cost of chemicals per unit weight.
      7.  Total power cost  including the kw-hr used and cost/kw-hr.
      8.  Cost for laboratory analyses  and data analyses.
      9.  Cost of residual handling at  the treatment/storage site,
             (i.e. hauling cost or pumping cost).
    10.  Actual  sludge dewatering and  disposal cost at  the dry-weather
            facilities based on the prevailing cost for dry-weather
            sludge cost/ton of  solids.
                                    291

-------
The total  annual  operating  cost  calculated  using  the  sum of  the  above
individual costs  should  be  reported  for  each  year of  operation.   These
costs divided by  the total  treated flow  with  the  appropriate conversion
factors will  yield an actual  cost on a volumetric basis such as  C/cu  m
U/1000 gal.).

Costs reported  in the above manner are not  suitable for comparative
purposes because  of the  many  possible variations. For example,  with  the
same system,  equal volumes  of storm  flow may  be treated  in  two consecutive
years.  However,  this same  total volume  may be generated by  10 storm
events in one year and AO storm  events the  next year.   Identical  monies
will be expended  In both years for amortization,  insurance and chemicals.
However, labor  costs for operation and routine maintenance,  power costs,
laboratory costs, data analysis  costs and sludge  handling costs  may be
four times greater the second year.

Another example which shows the  possible variations would be  treating
the same number of storm events  but  with the  total volume treated the
second year being k times as  great as the first year.   In this example
amortization, insurance, labor,  laboratory  analyses,  data analyses  and
power cost may  be the same  with  only chemical  costs.and sludge handling
costs different.   The cost  per cubic meter  (1000  gal.) treated would
approach only one fourth the  cost  the second  year.

In order to compare total annual costs for  different  systems
or for the same system at different  locations, the method of reporting
these costs must  be standardized.

The following method which  standardizes  all of these  variables  is
proposed for  reporting annual  costs. Many  assumptions must  be made.  These
include the following:

   1.  Fifty  discharge events occur  per  calendar  year at one week
          intervals.
   2.  Flow rate  to the  systems  is  50% of design  flow for systems
          operating at variable  flow rates  (due to excess capacity
          at  the  onset and  near  the  end  of  a  discharge).
   3.  The duration of each discharge is six  hours.
   k.  Amortization rate is 7.5% for 25  years.
   5.  The insurance rate is  1 %  of  the capital equipment cost.
   6.  Labor  costs are $12.00/man-hour  including  overhead and supervision
          for all labor  and classifications.   Between operation
          maintenance and associated materials must also be  estimated.
   7.  Power  cost is $0.03/kw-hr.
   8.  Chemical  cost is a truck load  rate not  including delivery.
   9.  Sludge handling cost at the  dry-weather facility at  2.5c/cu m
          (IOC/1000 gal.).
                                   292

-------
Total annual costs calculated on the above basis can be
used for comparisons when computed as cost/cu m (/1000  gal).

Other cost ratios which should be reported includes the following:

     1.  Capital cost expressed as $/cu m/day ($/mgd).
     2.  Operation and maintenance costs for both the actual  and theore-
           tical cases, expressed as $/kg ($/lb) of constituent removed,
           and $/storm event.


Sludge Handling Facilities - A description of all  facilities for handling
residuals from the treatment or storage of stormwater runoff should be
provided.   If none are used, such as with a pilot plant where residuals
are  routed back to the sewer during operation, a discussion of possible
methods for  treatment with a complete analysis of the sludges should
be given.   If possible, bench scale tests for sludge treatment methods
should be performed.  The description of sludge handling facilities
should include the following:

     1.  Description of the physical equipment  (size, volume, etc.).
     2.  Procedure for operation or schedule for pumpback.
     3.  Complete analysis of the sludge generated.
     k.  Volume generated as a percentage of the treated flow.
     5.  Method for ultimate disposal of the residuals.
     6.  Source of the residuals.
 Dry-Weather Treatment  Plant  Data  -  Information concerning the design
 capacity and diurnal flow variation  to  the dry-weather treatment plant
 is  necessary for design of storage facilities for combined sewage or
 treatment  residuals  from storm  flow  treatment.  Efficiency of the dry-
 weather plant may  be affected by  the volume or pollutional aspects of
 the pumpback rate.   Any decrease  in  dry-weather plant treatment capacity
 efficiency must be attributed to  the pumpback if excessive.  The data
 required  is:

     1.  Design capacity.
     2.  Diurnal flow variation.
     3.  Diurnal constituent  parameter variation.
     A.  Efficiency of  treatment.
     5.  Solids handling capacity.
     6.  Estimate of  effects  of  pumpback on treatment efficiency.
 Combined  Sewage or  Urban  Runoff  Data  -  Flow of  combined  sewage  to  the
 treatment system should be  recorded for instantaneous  rate and  totalized,
 In addition to the  flow through  the treatment system samples  taken at
 constant  time intervals and composited, for both  raw discharge  and
 treatment effluent, should  be analyzed.  In addition,  frequent  discrete
                                    293

-------
samples should be obtained on selected storm events to characterize
the overflow on a time basis.


Specific Information for Each Type of Treatment System

In addition to the general information to be reported for all  types of
systems, each type of system has operating parameters that affect
efficiency of the units.  With most demonstration or pilot units these
equipment variables are closely regulated to determine the optimum
operating conditions and to determine efficiency at the various  operating
conditions.  When reporting results of operation of these units  all
pertinent information must be generated.   The following items  discuss
the minimum amount of information to be reported and the method  of
reporting such for each type of system.


In-Line Storage - None of the methods previously discussed for measuring
efficiency are applicable for this type of system.  The effectiveness
of in-line storage is measured by the difference between potential
overflow and actual overflow.  The important parameter that should be
measured or calculated is the total storage capacity before installation
of the control devices and the total  storage capacity after installation.
The amount of pollutants captured is determined by the composite quality
of the flow in the sewers.

Flow measurement is also an important parameter to determine for this
type of system.  The difference between normal  dry-weather flow  and the
flow to the dry-weather plant during and following storm events  repre-
sents the total amount of stormwater captured.   Both of these  flows
should be measured.  The total  amount of overflow from all discharge
points in the network should be measured and sampled to determine the
total flow and constituents to the receiving stream.  The time it
takes to restore the sewers back to dry-weather flow conditions  should
also be reported.

In summary the specific information that should be reported for  in-line
storage systems includes:

    I.  Total  storage capacity for wet-weather  flow before and after
           installation of the system.
    2.  Flow measurement of,

           a.   Dry-weather flow
           b.   Wet-weather flow to the dry-weather plant.
    3.  Composite samples of each overflow should be obtained.
    4.  Record of sewer recovery to dry-weather conditions.
    5.  Annual operating results.
                                   294

-------
Off-line Storage -  In addition to evaluation of each individual  storm
event (both volumetric and constituent parameter efficiency)  a  running
account of available storage capacity should be maintained as well  as
a record of the pumpback.  The records should be kept over an entire
overflow season.  The logic behind this type of record is that  with a
storage system the  frequency of rainfall will affect the volumetric
efficiency.  Two rainfalls occurring at close intervals will, in all
likelihood, yield a lesser efficiency'because of the available  storage
space than the same two rainfall events at a longer interval.  The
pumpback or dewatering rate will affect efficiency in the former case.
Pumpback or dewatering procedures must be described fully.

If chemicals are used to aid settling in the storage tank or for
disinfection, the chemical injection points and methods of proportioning
chemicals should be described.
Bar Screens or Coarse Screens - Because this equipment is used only for
pretreatment and the materials captured are generally not included in
analyses of the overflow  (rags, leaves, paper, etc.) only the general
data plus the total volume captured  (cu m  [ft^]) need be reported.  The
method and cost of disposal of the captured material should be reported.
 Screens  (other  types) -  In  the description of the physical equipment the
 following  information regarding screens should be provided.

     1 .  Mesh si ze.
     2.  Size of open!ngs.
     3.  Wi re diameter.
     k.  Type of backing material.
     5.  Total open area of the screen  (as % of screen area).
     6.  Material of construction.
     7-  Screen  diameter and  length.
     8.  Submergence.

 The  method of cleaning the  screens during operation and for maintenance
 should be  reported to include the following  information:

     1.  Type of backwash system.
     2.  Type of spray nozzles.
     3.  Source  of backwash  water.
     A.  Backwash rate in I/sec  (gpm)  and 1/sec/m  (gpm/ft).
     5.  Method  of backwash  system activation  if not continuous.
     6.  Method  and frequency of cleaning to maintain  integrity.

 During operation of the screen the following  data and  information should
 be reported.
                                   295

-------
    ].   Hydraulic  and  solids  loading with  respect  to  time on  selected
           runs.
    2.   Operating  variables  including:
           a.   Screen  submergence          d.   Backwash  rate
           b.   Screen  rotation  rate        e.   Headless
           c.   Backwash  frequency

    3.   Total  sludge volume  produced and  its concentration.
    4.   Volumetric and constituent parameter efficiency.
Physical-Chemical  Treatment  -  Each  element  in  the  processes described
require specific  information to  be  reported.


Clarification - The type of  clarifier  used  will  determine  the  information
to be reported.   For a  flotation separator  report  the  same information
as listed in the  following section  on.dissolved  air  flotation.

With a gravity clarifier, report the following:

    1.  Design overflow rate.
    2.  Design surface  loading rate.
    3.  Type of mechanism used for  removal  of  floated  scum (grease  & oil)
           and settled  sludge.
    4.  Provisions for  handling  or  storing  the residual  sludges.
    5.  Sludge volume and concentration.
    6.  Variability in  loading rates during operation.
    7.  Volumetric and  constituent  parameter efficiency.
    8.  Chemicals  and dosages  if used.

If tube settlers  are used information  to  be reported includes all of
the above plus the following:

    1.  Tube configuration and dimensions.
    2.  Material  of construction.
Fi1tration - The filter  and  all  operating  parameters  should  be  described
fully.   This includes:

    1.   The type of filter - pressure  or gravity.
    2.   Type of  filter media and depths of the  various  media.
    3.   Size or  gradation of filter  media.
    4.   Bulk density and particle size.
    5.   Design loading  rates (hydraulic and solids).
    6.   Source of backwash water and type  of backwash system.
    7-   Backwash rate and frequency.
                                   296

-------
    8.  Disposition of backwash water.
    9-  Constituent parameter efficiency.
Carbon adsorption - The following specific information  should  be
reported.

    1.  Type of system utilized - upflow, downflow,  gravity or
           pressure, parallel or series.
    2.  Type  of activated carbon used.
    3-  Particle size and variation.
    4.  Design rate (hydraulic and constituent).
    5*  Exhaustion rate.
    6.  Backwash rate If required and disposition and source of
           backwash waters.
    7-  Provisions for carbon regeneration or disposal.
    8.  Constituent efficiency.

Each of these individual elements should be analyzed for:

    1.  Actual loading rates as a function of time.
    2.  Total residual solids.
Dissolved Air Flotation - In addition to determination of efficiencies
the following information and data should be reported:

    1.  Design overflow rate (1/min/sq m [gpm/sq ft])  based on
           unpressur!zed flow and on total  flow.
    2.  Design surface loading rate (g/day/sq m [Ib/day/sq ft]).
    3.  Method of pressurization.
    4.  Degree of saturation (% saturation).
    5-  Operating pressure.
    6.  Source of pressurized flow (i.e. total pressurization,  split
           flow or recycle).
    7.  Design pressurized flow rate expressed as the  ratio of
           pressurized flow to unpressurized flow.
    8.  Data to be acquired:

           a.  Actual  overflow rates and solids loading rate  as a
                  function of time.
           b.  Total residual sludge volume and concentration.
           c.  Chemicals used and their dosage.
Contact Stabilization - Both volumetric and overall efficiencies should
be reported for this type of biological system.  In addition, the
following information and data should be reported:
                                  297

-------
     1.  Sludge age at the start of treatment.
     2.  Sludge return rate and transfer rate.
     3.  Contact time (as a function of time).
     k.  Reaeration time (as a function of time).
     5.  Sludge concentration in reaeratton tank.
     6.  Mixed liquor concentration.
     7.  Aeration rate (cu m/min [cfm]).
     8.  Hydraulic and solids loading rate to the final  clarifier
            as a function of time.
     9.  Total amount of solids produced and added to the dry-weather
            system.
    10.  Description of the dry-weather treatment plant  associated with
            the storm generated discharge system.


Trlckling Filter - For this type of biological system specific information
about the equipment and operation should include the following:

     1.  Type of filtering media and material and total  volume of media.
     2.  Void space (5)  and total surface area (sq m/cu  m [sq ft/cu ft]).
     3.  Type of distribution mechanism.
     k.  Depth of filter.
     5.  Design loading rate (hydraulic and organic loading rate) based
            on once-through and based on recirculation (If used).
     6.  Actual loading rate as a function of time (once through  and
            with recirculat Ion).
     7.  Design recirculation rate.
     8.  Dry-weather operation to maintain biological  growth.
     9.  Description of the dry-weather treatment plant  associated with
            the storm generated discharge system.

Because a clarifier must be used in conjunction which a  trickling  filter
to remove the sludge sloughed off of the filter all  variables  previously
reported for the clarifier in a physical-chemical system must  also be
reported for this system.


Biological  Contactors - In addition to a physical description  of  the
system, the following information should be reported.

     1.  The number and diameter of the discs and their  materials
            of construction.
     2.  Total surface area of the disc and surface area submerged.
     3.  Rotational speed.
     k.  Number of modules in series or in parallel.
     5.   Design loading rates.
     6.   Actual loading rates (hydraulic and constituent parameter)
            as a function of time.
     7.   Dry-weather operation to maintain a viable system including
            loading rates.
                                   298

-------
As with other biological systems, a clarlfier Is an Integral  part of the
system and all variables associated with the operation of a  clarlfler
must be reported.


La9°?ns " Lagoons are similar  in operation to storage tanks  except for
on-site treatment by aeration or oxidation.  Stored flow is  generally
not pumped back to a dry-weather treatment plant but to a receiving
stream.  For this type of system, only the efficiencies need to be
reported plus the quantity, concentration and final disposition of
residual sludges.  This  Is  in addition to the general Information.


Disinfection - Specific  information to be reported when disinfection
is used should include:

     1.  The specific type of disinfectant used, its strength and the
            method of addition.
     2.  Type of device  used to provide contact time.
     3-  Design and actual contact time.
     A.  If chlorine is  used, the chlorine residual in the effluent.
     5.  If ozone is used, describe completely the ozone generation
            equipment (air or oxygen feed gas) and cost per  unit of
            ozone generated ($/kg [$/lb] or kw-hr/kg[ kw-hr/Jb]).


ECONOMIC DECISION MAKING CONSIDERATIONS

 It  is  becoming  increasingly apparent  that  the  problems of discharges
 resulting from  rainfall  events will be  solved  on  individual  bases  rather
 than by all encompassing effluent guidelines.   In  other words,  the
solutions rendered will  be  determined by a combination of desired
 receiving water quality  and allowable economic investments.  Thus, each
situation will  require an  in-depth determination  of  what  impact  the
present storm generated  discharges have on the receiving waters  and
just what the change  in  water  quality would  be for various  degrees of
storm  generated discharge abatement.  This type of approach  has  become
 the obvious as a  result  of  various surveys and needs  studies which have
shown  the cost of  implementing complete storm  generated  discharge  abate-
ment to be an order of magnitude greater  in  cost  than any other  water
pollution control abatement undertaking.

 It  is  also  imperative that  any study  concerning storm generated  discharges
 take into account other  sources of water pollution affecting the receiv-
 ing body of water, most  notably the discharge  from the existing
municipal and  industrial sewage treatment  plants.   A solution  methodology
must be employed which will show the  most  cost-effective means of
 improving water quality. This methodology must take into consideration
 the receiving water  impact  and must be  site  specific looking at tertiary
 dry-weather treatment versus  wet weather  flow treatment, structural
                                    299

-------
solution and non-structural solutions, the possibility of integrated
dry-weather and wet-weather systems, the possibility of integrating
flood and erosion control technology with pollution control  technology
and the integration of land management and non-structural techniques.

In order to optimize the above type of Analyses, it will probably be
necessary to use computerized techniques, especially for assessing the
water quality impact.  Just as important is that the critical or limit-
ing parameters which affect the water quality be identified and the
long term quality of the receiving water be estimated.

A hypothetical example of this type of approach is presented in Figure
VIM-l^t.  These curves represent the percentage of time the receiving
water D.O.  level is greater than or equal to a D.O. level  on the
abscissa.   They should represent  at least one year's continuous flow of
data.   This case is for D.O., but actual  studies would use the parameter
or parameters found to be the most critical in that particular case.   By
this analysis we can make true cost-effectiveness comparisons based on
the total  of receiving water impacts and  associated abatement costs.

For example, if we desire 5 mg/1  D.O.  in  the receiving water 75% of the
time as a  standard, then we need'to go to an advanced form of wet-weather
treatment  or primary wet-weather treatment integrated with land manage-
ment.   The  latter is most effective at 3  million dollars in  cost.   This
or similar  methodologies can help set cost-effective standards as  well as
select alternatives.
                                  300

-------
   lOO-i
to
CO

CO
o

D

O
 CN
u
UJ
LJU
s
                                        WET-II (75.%  Rem) OR

                                    \ ("WET-I (PR!M)/LM

                                     ^^>    (75% Rem)
                                 I   TERTIARY  NN\
                            D.O. (mg/l)
CONTROL
ALTERNATIVES
EXISTING
TERTIARY
WET-I (PRIMARY)
WET-II (ADV)
WET-I/LAND MGMT.
% BOD REMOVAL
DRY WEATHER
85
95
85
85
85
WET WEATHER
0
0
25
75
75
COST
($xl06)
—
6
1
6
3
     Figure VI11-14.  Hypothetical example of  the

        economic solution methodology approach
                         301

-------
                      SECTl'ON  IX - REFERENCES
                            SECTION  IV

 1.   Cornell,  Rowland,  Hayes and  Merryfield  Engineers,  "Rotary
     Vibratory Fine Screening of  Combined Sewer Overflows",  USEPA
     Report No.  11023FDD03/70,  NTIS-PB 195 168,  March  1970.

 2.   University  of Cincinnati,  "Urban  Runoff Character fs tics",  USEPA
     Report No.  1 1024DOJJ 10/70,  NTIS-PB 202 865,  October 1970.

 3.   Underwater  Storage Inc. and  Silver Schwartz,  Ltd., "Control  of
     Pollution by  Underwater Storage", USEPA Report  No. 11020DWF12/69,
     NTIS-PB 191 217,  December 1969.
 k.   Shelley,  P.E., and Kirkpatrick,  G.W., "An  Assessment  of Automatic
     Sewer Flow  Samplers",  USEPA  Report No.  EPA-R2-73-261 , NTIS-PB
     223 355,  1973.
 5.   Data from EPA Grant No. S8007M,  Racine, Wisconsin,  Envirex  Inc.,
     Milwaukee,  Wisconsin.

 6.   Fair, J.R., Crocker, B.B., and Null, H.R.,  "Sampling  and Analyzing
     Trace Quantities", Chemical  Engineering, 79,  21,  IA6-154,  1972.

 7.   Lager, J.A.,  and  Smith, W.G., "Urban Stormwater Management and
     Technology: An Assessment",  USEPA Report No.  EPA-670/2-7^-040,
     NTIS-PB 240867/AS,
 8.   Crane Company,  "Mlcrostraining  and  Disinfection  of  Combined  Sewer
     Overflows",  USEPA Report  No.  1 1023EV006/70,  NTIS-PB 195  W,  June  1970.
 9.   Glover,  G.E.,  and Herbert,  G..R.,  "Microstralning  and  Disinfection
     of Combined Sewer Overflows  -  Phase  II",  USEPA Report No.
     EPA-R2-73-124,  NTIS-PB  219  879,  1973.

10.   Hayes,  Seay,  Mattern  and Mattern,  "Engineering Investigation  of
     Sewer Overflow Problems",  USEPA  Report No.  1 1024DMS05/70,
     NTIS-PB  195 201,  May  1970.

11.   AVCO Economic Systems Corp., "Storm  Water Pollution From Urban  Land
     Activity",  USEPA  Report No.  1 1034FKL07/70,  NTIS-PB  195 281, July  1970.
12.   Davis,  P.L.,  and  Borchardt,  F.A. ,  "Combined Sewer Overflow Abate-
     ment Plan,  Des  Moines,  Iowa",  USEPA  Report No. EPA-R2-73-170, 1973.

13.   Burgess  and Niple,  Ltd.,  "Stream Pollution and Abatement from
     Combined Sewer Overflows,  Bucyrus, Ohio", USEPA Report No.
     1102AFKN11/69,  NTIS-PB  195  162,  November  1969.

\k.   The City of Chippewa  Falls,  Wisconsin, "Storage and Treatment of
     Combined Sewer Overflows",  USEPA Report  No.  EPA-R2-72-070,
     NTIS-PB  2IA 106,   1972  .
                                   302

-------
15.  Data from EPA Grant No. S80CW, Racine,  Wisconsin,  Envlrex  Inc.,
     Milwaukee, Wisconsin.

16.  "Handbook In Monitoring Industrial Wastewater",  EPA  Technology
     Transfer, August  1973.

17.  "Methods for Collection and Analysis of Water Samples  for Dissolved
     Minerals and Gases", Chapter A1 , Book 5,  Laboratory  Analysis, by
     Brown, E., Skougstad, M.W., and Fishman,  M.J., U.S.  Dept  of  the
     Interior Geological Survey, 1970.

18«  1973 ASTM Standards. Part 23,  Water; Atmospheric Analysis, ASTM,
     Philadelphia, Pennsylvania, 1973.

19.  "Recommended Methods for Water-Data Acquisition", Preliminary
     report, U.S. Dept. of the Interior Geological Survey,  Office of
     Water Data Coordination, Washington, D.C., December  1972.

20.  "Methods for Chemical Analysis of Water and Wastes 1971,  "EPA,
     National Environmental Research Center Analytical Quality Control
     Laboratory,  Cincinnati, Ohio,  USEPA Report No. EPA 625/6-74-003,
     1974.

21.  Shelley, P.E., "Implementation of New Technologies in  the Design
     of an Automatic Liquid Sampling System",  paper presented  at
     International Seminar and Exposition on Water Resources Instrumenta-
     tion, Chicago, Illinois, June 4-6,
22.  Shaheen, D.G. , "Contributions of Urban Roadway Usage to Water
     Pollution", USEPA Report No. EPA-600/2-75-004, 1975.

23.  Street Surface Sampling and Analytical Techniques, Personal
     communication from Woodward-Clyde Consultants to USEPA, Edison,
     New Jersey, April 3, 1975.

24.  American Public Works Association, "Water Pollution Aspects  of
     Urban Runoff", USEPA Report No. 11030DNSO J/69, NTIS-PB 215 532,
     January 1969.

25.  Sartor, J.D., and Boyd, G.B., "Water Pollution Aspects of
     Street Surface Contaminants", USEPA Report No. EPA-R2-72-081 ,
     NTIS-PB 214 408, 1972.

26.  Pitt, R.E., and Amy, G. ,  "Toxic Materials Analysis of Street
     Surface Contaminants",  USEPA Report No. EPA-R2-73-283, NTIS-
     PB-224-677/AS, 1973.
                                   303

-------
                              SECTION V
 1.  Baker, D.  G., "Prediction of Spring Runoff",  Water Resources Research,
     8_, No. 4,  pp. 966-972,  1972.

 2.  Kinsley,  R.  K.,  Jr.,  Kohler, M.  A., and Paulhus,  H.  L.,  "Hydrology for
     Engineers",  McGraw Hill  Book Co.,  New York,  1958.

 3.  Viessraan,  W.  Jr.,  Harbaugh,  I.E.,  and Knapp,  J.W., "Introduction  to
     Hydrology",  in text,  Education  Publishers,  New York,  1972.

 4.  Davis, C.V.,  and Sorensen,  K.E.,  "Handbook  of Applied Hydraulics",
     3rd Edition,  McGraw Hill, New York, 1969.

 5.  Diskin,  M.H., "On  the Computer  Evaluation of  Thiessen Weights",
     Journal  of Hydrology, 11, No.  1,  pp.  69-78,  1970.

 6.  Hutchinson,  P.,  "A Contribution  to the Problem of Spacing Raingages in
     Rugged Terrain", Journal  of Hydrology, 12,  No. 1,  pp. 1-14,  1970.

 7.  Huff,  F.  A.,  "Time Distribution  of Rainfall  in Heavy  Storms",
     Water  Resources  Research, 3, No.  4, pp. 1007-1019, 1967.

 8.  Fogel, M.M.,  and Dickstein,  K.,  "Prediction of Convective Storm
     Runoff in  Semi-Air Regions", Int.  Assoc.  Science  Hydrology  Pub.,
     96, pp.  465-478, 1970.

 9.  Dickstein, L., Fogel, M.M.,  and  Kisiel, C.C., "A  Stochastic  Model  of
     Runoff Producing Rainfall for Summer Type Storms", Water Resources
     Research,  8,  No. 2, 1972.

10.  Pattison,  A., "Synthesis  of Hourly Rainfall Data", Water Resources
     Research,  4,  pp. 489-498, 1965.

11.  Babbitt,  H.E., Sewerage  and Sewage Treatment, 7th  Edition,  John Wiley
     6 Sons,  Inc., New  York,  1956.

12.  Hess,  J.G.,  and  Manning,  F.G., "A Rational  Determination of  Storm
     Overflows  from Intercepting Sewers",  Sewage and Industrial Wastes,
     22, No.  2, pp. 145-153,  1950.

13.  Clark, J.W.,  Viessman, W. Jr., and Hammer,  M.J.,  Water Supply  and
     Pollution  Control, 2nd Edition,  International Textbook Co.,  Scranton,
     Pennsylvania, 1971.

14.  Burgess  and  Niple, Ltd.,  "Stream Pollution  and Abatement from
     Combined  Sewer Overflows, Bucyrus, Ohio", USEPA Report No.
     11024FKN11/69, NTIS-PB  195  162,  November 1969.

                                    304

-------
15.  Mitchell, W.D., "Linear Analysis of Hydrographs", Water Resources
     Research. 3, No. 3, pp. 891-895, 196?.

'6.  Singh, R., "Water duality of Urban Storm Runoff", paper presented
     at International Seminar and Exposition on Water Resources
     Instrumentation, Chicago, Illinois, June 4-6, 1974.

17.  Eagleson, P.S., and Shaake, W.J., "Some Criteria for the Measurement
     of Rainfall and Runoff", Water Resources Research, 2, No. 3,
     pp. 427-426, 1966.

18.  Hicks, W.I., "A Method of Computing Urban Runoff", Proceedings
     ASCE, 109, No. 2230, pp. 1217-1320, 1944.

19.  Tholin, A.L., and Kiefer, C.J., "The Hydrology of Urban Runoff",
     Journal Sanitary Engineering Division, ASCE, 85, No. SA2, pp. 47-106,
     1959.

20.  Davis, P.L., and Borchardt, F.A., "Combined Sewer Overflow
     Abatement Plan, Des Moines, Iowa", USEPA Report No. EPA-R2-73-170,
     1973.

21.  Lager, J.A., and Smith, W.G., "Urban Stormwater Management and
     Technology: An Assessment", USEPA Report No. EPA-670/2-74-040,
     NTIS-PB 240867/AS,  1974.

22.  Ritter, F.G., and Warg, C., "Upgrading City Sewer Installations".
     Engineering Digest, Apri1 1971.

23.  Terstriep, M.L., and Stall, J.B., "Urban Runoff by the Road
     Research  Laboratory Method", ASCE Hyd raulie DIv i s i on, 95, No. HY6,
     pp. 1809-1934, November 1969.

24.  Stall, J.B., and Terstriep, M.L., "Storm Sewer Design - An Evaluation
     of the RRL Method", USEPA Report No. EPA-R2-72-068, NTIS-PB 214 134,
     1972.

25.  Crawford, N.H., and Linsley, R.K., "Digital Simulation in Hydro-
     logy: Stanford Watershed Model IV", Stanford University. Department
     of Civil  Engineering Technology Report 33,  1966.

26.  Metropolitan Sewer Board, St. Paul  Minnesota, "Dispatching
     System for Control of Combined Sewer Losses", USEPA Report
     No. 11020FAQ03/71, NTIS-PB 203 678, March 1971.

27.  "A Mathematical Model for Optimum Design and Control of Metro-
     politan Wastewater Management Systems", Battelle Pacific Northwest
     Laboratories, April 1973.
                                    305

-------
28.  Pianno, F.R.,  Anderson,  D.J.,  and Derbin,  S.,  "The Use of
     Mathematical  Modeling to Evaluate Sewer Design and Operation",
     City of Cleveland exhibit handout,  WPCF 46th Annual Conference,
     Cleveland,  Ohio,  September 30-October 5,  1973.

29.  Metcalf and Eddy  Engineers,  "Storm Water Management Model,  Vol.  1,
     Final Report", USEPA Report  No.  11024DOC07/71, NTIS-PB 203  289,
     July 1971.

30.  Metcalf and Eddy  Engineers,  "Storm Water Management Model,  Vol.  II,
     Verification  and  Testing ",  USEPA Report No. 11024DOC08/71,
     NTIS-PB 203 290,  August  1971.

31.  Metcalf and Eddy  Engineers,  "Storm Water Management Model,  Vol.  Ill,
     User's Manual",  USEPA Report No.  11024DOC09/71,  NTIS-PB 203 291,
     September 1971.

32.  Metcalf and Eddy  Engineers,  "Storm Water Management Model,  Vol.  IV,
     Program Listing", USEPA  Report No.  11024DOC10/71,  NTIS-PB 203 292,
     October 1971.

33.  Renard, K.G.,  and Osborn, H.B.,  "Rainfall  Intensity Comparison  from
     Adjacent 6-Hour  and 24-Hour  Recording Gages",  Water Resources
     Research, 2,  No.  1, pp.  145-146,  1966.

34.  Amorocho, J.,  Brandstetter,  A.,  and Morgan,  D.,  "The Effects of
     Density of Recording Rain Gauge  Networks  on  the  Description of
     Precipitation  Patterns:   Geochemistry,  Precipitation, Evaporation,
     Soi1-moisture, Hydrometry",  General Assembly of  Bern, September-
     October, 1967.

35.  Hendrick, R.L.,  Comer,  G.H.,l"Space Variations of  Precipitation  and
     Implications  for  Raingage Network Design",  Journal of Hydrology,  10,
     No.  2, pp.  151-163, 1970.

36.  Eagleson, P.S., "Optimum Density of Rainfall Networks", Water Resources
     Research, 3.,  No.  4, pp.  1021-1033,  1967.

37-  Osborn, H.B.,  Lane, J.,  and  Hundley, J.F.,  "Optimum Gaging  of
     Thunderstorm Rainfall in South Eastern  Arizona", Water Resources
     Research, &_,  No.  1, pp.  259-265,  1972.

38.  Tucker, L.S.,  "Sewered  Drainage  Catchments  in  Major Cities", ASCE
     Program Technical Memorandum No.  10, March  31, 1969.

39.  Basic Information Needs  in Urban Hydrology,  a study by American  Society
     of Civil Engineers sponsored by  Geological  Survey, Contract No.
     14-08-0001-11257, April, 1969.
                                    306

-------
*»0.  Nebolsine, R., Harvey, P.J,, and Fan, C., "High Rate Filtration of
     Combined Sewer Overflows", USEPA Report No. 11023EY104/72,  NTIS-PB
     211 144, April 1972.

41.  Metropolitan Sewer Board, St. Paul, Minnesota, "Dispatching
     System for Control of Combined Sewer Losses",  USEPA Report  No.
     11020FAQ03/71, NTIS-PB 203 678, March 1971.
42.  Municipality of Metropolitan Seattle, Washington, "Maximizing
     Storage in Combined Sewer Systems", USEPA Report No. 11022ELK12/71,
     NTIS-PB 209 861, December, 1971.

A3.  Havens and Emerson, "Feasibility of a Stabilization - Retention
     Basin in Lake Erie at Cleveland, Ohio", USEPA Report No.
     11020—05/68, NTIS-PB 195 083, May 1968.

44.  Rainfall Frequency Atlas of the United States, Technical  Paper
     No. 40, U.S. Department of Agriculture.

45.  Hayes, Seay, Mattern and Mattern, "Engineering Investigation  of
     Sewer Overflow Problems", USEPA Report No. 11024DMS05/70,
     NTIS-PB 195 201, May 1970.

46.  Crane Company, "Microstraining and Disinfection of Combined Sewer
     Overflows", USEPA Report No. 11023EV006/70, NTIS-PB 195 674,June 1970.

47<  "Develop and Evaluate Methods for Determining Cumulative  Stormwater
     Runoff Volume and Flowrates for Design and Operation of Combined Sewer
     Overflow and Stormwater Runoff Control and Treatment Facilities",
     Progress Report #1, 68-03-0302, September 10,  1974, Urbana  USGS Gaging
     Station, University of Illinois.

48.  Dow Chemical Company, "Chemical Treatment of Combined Sewer
     Overflows", USEPA Report No. 11023FDB09/70, NTIS-PB 199 070,
     September  1970.
49.  Data from  EPA Grant No. S800744, Racine, Wisconsin, Envirex Inc.,
     MiIwaukee, Wisconsin.
50.  "San Francisco Master Plan for Waste Water Management",
     Preliminary Comprehensive Report by Department of Public Works,
     September  15, 1971.
51.  Agnew, R.W., et al., "Biological Treatment of Combined Sewer
     Overflow at Kenosha, Wisconsin", USEPA Report No. E.PA-670/2-75-019,
     NTIS-PB 242126/AS,  1975.

52.  City of Milwaukee, Wisconsin Department of Public Works^MlIwaukee,
     Wisconsin  and Consoer, Townsend and Associates, Consulting
     Engineers," Detention Tank for Combined Sewer Overflow - Milwaukee,
     Wisconsin, Demonstration Project", USEPA  Report No. EPA-600/2-75-071,
     December  1975.
                                    307

-------
53.  Liskowitz,  J.W.,  et al.,  "Suspended Solids  Monitor",  USEPA Report
     No.  EPA-670/2-75-002,  NTIS-PB  241  581/AS,  1975.

5k.  Brater,  E.F.,  "Rainfall  - Runoff Relations  on  Urban  and  Rural
     Areas", USEPA Report No.  EPA-670/2-75-046, NTIS-PB  242830/AS,  1975.


                             SECTION  VI

 1.  Hittman  Associates, Inc., "The Beneficial Use  of Storm Water",
     USEPA Report No.  11030DNK08/68 NTIS-PB  195  160,  August 1968.

 2.  Sartor,  J.D.,  and Boyd,  G.B.,  "Water Pollution Aspects of
     Street Surface Contaminants".  USEPA Report  No. EPA-R2-72-081,
     NTIS-PB  211* 408,  1972.

 3.  Metcalf  and Eddy  Engineers,  "Storm Water Problems  and  Control  in
     Sanitary Sewers,  Oakland  and Berkeley,  California",  USEPA Report
     No.  11024EQG03/71,  NTIS-PB 203 680,  August  1970.

 4.  Black, Crow and Eidsness, Inc.,  "Storm  and  Combined  Sewer Pollution
     Sources  and Abatement,  Atlanta,  Georgia", USEPA  Report No.
     11024ELB01/71, NTIS-PB  201 725,  January 1971.

 5.  Hayes, Seay,  Mattern and  Mattern,  "Engineering Investigation of
     Sewer Overflow Problems", USEPA  Report  No.  11024DMS05/70,  NTIS-PB
     195  201, May 1970.

 6.  Simpson, G.D.  and  Curtis, L.W., "Treatment of Combined  Sewer  Over-
     flows and Surface Waters  at  Cleveland,  Ohio",  JWPCF,  41,  2,  151,  1969.

 7.  DeFilippi,  J.A.,  and Shih, C.S., "Characteristics  of Separated
     Storm and Combined Sewer  Flows", JWPCF, A3, 10,  2033 (1971).

 8.  Burn, R.J., Krawczyk,  D.F.,  and  Harlow, G.L.,  "Chemical  and
     Physical Comparison of  Combined  and Separate Sewer Discharges",
     JWPCF, 40,  1,  112,  1968.

 9.  Stegmaier,  R.B. Jr., "Storm  Water  Overflows",  Sewage Works Journal,
     14,  6, 1264,  1942.                             ""*

10.  Benzie,  W.J.,  and Courchalne,  R.J.,  "Discharges  from Separate  Storm
     Sewers and  Combined Sewers", JWPCF,  38, 3,  4lO,  1966.

11.  Colston, N.V., "Characterization and Treatment of  Urban  Land Runoff",
     USEPA Report No.  EPA-670/2-74-096,  NTIS-PB  240 987/AS, 1974.

12.  Metropolitan Sewer Board, St.  Paul,  Minnesota, "Dispatching
     System for  Control  of Combined Sewer Losses",  USEPA  Report
     No.  11020FAQ03/71,  NTIS-PB 203 678,  March  1971.
                                   308

-------
13.  Merrimack College, "Proposed Combined Sewer Control by Electrode
     Potential", USEPA Report No. 11024DOK02/70/, NTIS-PB 195 169,
     February 1970.

1**.  Envirogenics Company, "Urban Storm Runoff and Combined Sewer
     Overflow Pollution, Sacramento, California", USEPA Report No.
     11024FKM12/71, NTIS-PB 208 989, December 1971.

15.  University of Cincinnati, "Urban Runoff Characteristics", USEPA
     Report No. 1 1024DOJJ10/70, NTIS-PB 202 865,  October 1970.

16.  Nebolsine, R., Harvey, P.J., and Fan, C., "High Rate Filtration of
     Combined Sewer Overflows", USEPA Report No. 11023EY104/72,
     NTIS-PB 211 144, April 1972.

17.  Hansen, C.A., Gupta, M.K., and Agnew, R.W., "Two Wisconsin Cities
     Treat Combined Sewer Overflows", Water and Sewage Works,  120,
     8, 48, 1973.

18.  "Characterization and Control of Combined Sewer Overflows in
     San Francisco", Water Research, 3, 531 (1969).

19.  Rhodes Corporation, "Dissolved -Air Flotation Treatment  of
     Combined Sewer Overflows", USEPA Report No. 11020FK101/70,
     NTIS-PB 189 775, January  1970.

20.  Underwater Storage Inc. and Silver Schwartz, Ltd., "Control  of
     Pollution by Underwater Storage", USEPA Report No. 11020DWF12/69,
     NTIS-PB 191 217, December 1969.

21.  Burgess and Niple, Ltd., "Stream Pollution  and Abatement  from
     Combined Sewer Overflows, Bucyrus, Ohio", USEPA Report
     No. 11024FKN11/69, NTIS-PB 195 162, November 1969.

22.  Hercules, Inc., "Crazed Resin Filtration of  Combined Sewer Overflow",
     USEPA Report No. 11020—10/69,  NTIS-PB 187 867,  October  1969.

23.  Rand Development Corp., "Rapid-Flow Filter  for Sewer  Overflows",
     USEPA Report No. 11023DP108/69, NTIS-PB 194 032,  August 19&9.

24   Fram Corporation, "Strainer/Filter Treatment of Combined  Sewer
     Overflow", USEPA Report No.  11020EXV07/69,  NTIS-PB 185 949, July  1969.

25   American Process Equipment Corp., "Ultrasonic Filtration  of
     Combined Sewer Overflows", USEPA Report No. 11023DZF06/70,
     NTIS-PB 212 421, June 1970.

26   USEPA Storm and Combined Sewer Pollution  Control  Branch,  "Combined
     Sewer Overflow Abatement Technology", USEPA Report No. 11024—06/70,
     NTIS-PB 193 939, June 1'970.
                                    309

-------
27.  USEPA Storm and Combined Sewer Pollution Control  Branch, "Combined
     Sewer Overflow Abatement Technology", USEPA Report No. 1102^—06/70,
     NTIS-PB 193 939, June 1970.

28.  Crane Company, "Microstraining and Disinfection of Combined Sewer
     Overflows", USEPA Report No. 1 1023EV006/70, NTIS-PB 195 671*,
     June 1970.

29.  Springfield Sanitary District, "Retention Basin Control of
     Combined Sewer Overflows", USEPA Report No. 11023—08/70,
     NTIS-PB 200 828, August 1970.

30.  Dow Chemical Company, "Chemical  Treatment of Combined Sewer
     Overflows", USEPA Report No. 11023FDB09/70, NTIS-PB 199 070,
     September 1970.

31.  Envirogenics Company, "In-Sewer Fixed Screening of Combined
     Sewer Overflows", USEPA Report No.  1102*fFKJ 10/70,  NTIS-PB
     213 118, October 1970.

32.  Melpar, Falls Church, Virginia,  "Combined Sewer Temporary Under-
     water Storage Facility", USEPA Report No. 11022DPP10/70, NTIS-
     PB 197 669, October 1970.

33.  Dodson, Kinney and Lindblom,  "Evaluation of Storm Standby
     Tanks,  Columbus, Ohio",  USEPA  Report  No. 11020FAL03/71, NTIS-PB
     202 236, March 1971.

3^.  Department of Public Works, Portland, Oregon,  "Demonstration of
     Rotary  Screening for Combined  Sewer Overflows", USEPA Report
     No. 11023FDD07/71, NTIS-PB 206 8H, July 1971.

35.  FMC Corporation, "A Flushing System for Combined  Sewer Cleansing",
     USEPA Report No. 11020DN003/72,  NTIS-PB 210 858,  March 1972.

36.  The City of Chippewa Falls, Wisconsin,  "Storage and Treatment of
     Combined Sewer  Overflows", USEPA Report No.  EPA-R2-72-070,
     NTIS-PB 2\k 106, (1972).

37.  Municipality of Metropolitan Seattle, Washington,  "Maximizing
     Storage in Combined Sewer Systems", USEPA Report  No.  11022ELK12/71,
     NTIS-PB 209 861, December, 1971.

38.  Karl  Rohrer Associates,  Inc.,  "Underwater Storage of Combined
     Sewer Overflows", USEPA Report No.  11022ECV09/71,  NTIS-PB
     208 3^6, September 1971.

39.  Standard Methods for the Examination  of Waters  and Wastewaters,
     APHA,  AWWA, WPCF, 13th  Edition,  1971.
                                   310

-------
 *»0.  Methods for Chemical Analysis of Waters and Wastes  1971,  EPA
      National Environmental Research Center Analytical Water Quality
      Control Laboratory, Cincinnati, Ohio,  USEPA Report  No.
      EPA-625/6-74-003,  1974.

 ^1«  Annual Books of ASTM Standards, Water;  Atmospheric  Analysis,
      American Society for Testing and Materials, Philadelphia.

 42.  Strickland,  J.D.  and Parsons, T.R.,  A  Manual  of Sea Water  Analysis,
      Bulletin No. 125,  Fisheries  Research Board of Canada,  Ottawa,  1965.

 43.  "Methods of  Soils  Analysis,  Parts I  and II",  American  Society  of
      Agronomy,  Madison,  Wisconsin, 1965.

 44.   Rainwater, F.H.  and Thatcher, L. L. ,  Methods for Collection  and
      Analysis of  Water  Samples, Geological  Survey  Water  Supply  Paper
            I960.
45.  Handbook  for Analytical Quality  Control  in Water  and Wastewater
     Laboratories,  USEPA, June,  1972.

46.  Hnadbook  for Monitoring Industrial Wastewater, EPA Technology
     Transfer, August  1973.

47.  Hansen, C.A.,  Agnew, R.W.,  and Murray, D.L., "Quality Characteristics
     of Combined Sewer Overflows and  Urban Stormwater  Runoff", presented at
     Central States Water Pollution Control Association meeting, June 15,
     1972, Milwaukee, Wisconsin.

48.  Sawyer, C.N.,  and McCarty,  D.L.. Chemistry for Sanitary Engineers,
     McGraw Hill Book Co., 2nd Edition, pp. 384-400, 1967.

49.  Heukelekian, H., Gel Iman, I., "Effect of Certain Environmental
     Factors in the Biochemical  Oxidation of Waste", Sewage and Industrial
     Wastes, December 1951.

50.  Marske, D.M.,  and Polkowski, L.B., "Evaluation of Methods for
     Estimating Biochemical Oxygen Demand Parameters",  JWPCF,  44,
     No.  10, pp. 1987-2000, October 1972.

51.  Slette, 0., "Determining BOD Curve Parameter", Water and  Sewage
     Works, p. 133-138, April 1966.

52.  Fair, G., "The Log Difference Mehtod of Estimating the  Constants
     of the First-Stage Biochemical Oxygen Demand  Curve",  Sewage
     Works Journal, 8, 3, 430-434.

53.  Thomas, H.A. Jr., "The 'Slope' Method of Evaluating  the Constants
     of the First Stage Biochemical Oxygen-Demand  Curve",  Sewage Works
     Journal,  9, 3, 425-430.
                                   311

-------
54.  Revelle, C.S., Lynn, W.R., Rivera, M.A., "Bio-Oxidation Kinetics
     and a Second-Order Equation Describing the BOD Reaction", JWPCF,
     37, 12, 1679-1692, December 1965.

55.  Gaudy, A.F. Jr., and Gaudy, E.T., "ACOD Gets Nod Over BOD Test",
     Industrial Water Engineering, pp. 30-34, Aug.Sept. 1972.

56.  AH, H.I., and Bewtra, J.K., "Effect of Turbulence on BOD
     Testing", JWPCF, 44, 1798, 1972.

57.  Cripps, J.M., and Kenkins, D., "A COD Method Suitable for Analysis
     of Highly Saline Water", pp. 1240-1246, JWPCF, October 1964.

58.  Standard Method Test for "Chemical Oxygen Demand (Dichromate Oxygen
     Demand) of Wastewater", ASTM, pp. 1252-1267.

59.  Burns, E.R., Marshall, C., "Correction for Chlorine Interferences
     in the COD Test", JWPCF, pp. 1716-1721, December 1965.

60.  Analytical Quality Control Bulletin, July 1971, USEPA National
     Environmental Research Center, Cincinnati, Ohio.

61.  Takahaski, Y., Moore, R.T., Joyce, R.J., "Direct Determination of
     Organic Carbon in Water by Reductive Pyrolysis", American
     Laboratory, 4_, No. 7, pp. 31-38, July 1972.

62.  Anon., "Brave New World of TOC and TOD", Ind. Wa ter Eng i nee ring,
     p. 13, September/October 1973.

63.  Goldstein, A., Katz, W., Mailer, F., Murdoch, D., "Total Oxygen
     Demand.  A New Automated Instrumental Method for Measuring
     Pollution and Loading on Oxidation Processes", ACS, 1968.

64.  Christie,  T., Bergmann, W., "New Instrumental Approach to Water
     Pollution Problems", American Laboratory, July 1969.

65.  O'Herron,  R., "Literature Survey of Instrumental Measurements of
     Biochemical Oxygen Demand for Control Application, 1969-1973",
     National Environmental  Research Center, Cincinnati, Ohio.

66.  "TOD Analyzer Applications in Three Oxygen Demand Ranges", Yvasa
     Bettery Co., Ltd., February 1973.

67.  Zanoni, A.E., and Rutkowski, R.J., "Per Capita Loading of Domestic
     Wastewater", JWPCF, 44:1756-1762, September 1972.

68.  Romer, H.  and Klashman, L.M., "How Combined Sewers Affect Water
     Pollution", Public Works, p. 88, April 1963.
                                  312

-------
69.  Weibel, S., Anderson, R., and Woodward, R.I. , "Urban Land Runoff
     as a Factor in Stream Pollution", JWPCF, 36, 914-924, July 1964.

70.  Weston, R.F., Germain, J.E., and Fiore, M.E., "Solving the
     Combined Sewer Overflow Problem of a Major City", Public Works,
     pp. 106-108, May 1972.

71.  Engelbrecht, R.S., and McKinney, R.E., "Membrane Filter Method
     Applied to Activated Sludge Suspended Solids Determinations",
     Sewage and Industrial Wastes, 28, 1321-1325, November 1956.

72.  Wyckoff, B.M., "Rapid Solids Determination Using Glass Fiber
     Filters", Water and Sewage Works, 111, pp. 277-280, June
73.  AWWA Committee on Viruses, "Viruses in Water", JWPCF, 61, p. 491, 1969.

74.  Geldreich, E.E., "Water-Borne Pathogens", Water Pollution Micro-
     biology, Edited by R. Mitchell, Wiley-lnterscience,  1972.

75.  Taylor, S.W., Bacteria  in Relation to Water", Proc.  Inst. Civil
     Eng. , 2, p. 398, December 1963.

76.  Weibel, S.R., Dixon, F.R., Weidner, R.B., and McCabe, L.J.,
     "Waterborne-Disease Outbreaks,  1946-60", Journal AWWA, 56,
     p. 947, 1964.

77.  Kabler, P., "Removal of Pathogenic Microorganisms by Sewage Treatment
     Processes", Sewage and  Industrial Waste, 31, p. 1373, 1959.

78.  Geldreich, E.E. , "Buffalo Lake  Recreational Water Quality:  A
     Study in Bacteriological Data Interpretation", Water Research, 6_,
     p. 913, 1972.

79.  Gallagher, T.P., and Spino, D.G., "The Significance of Numbers of
     Col I form Bacteria as an Indicator of Enteric Pathogen",  Water
     Research, 2, p. 169, 1968.

80.  Clark, H.F., and Kabler, P.W.,  "ReevaluatSon of the Significance
     of the Coliform Bacteria", Journal AWWA, 56, p. 931, 1964.

81.  Geldreich, E.E., Bordner, R.H., Huff, C.B., Clark, H.F., and
     Kabler, P.W., "Type Distribution of Coll form Bacteria in the Feces
     of Warm-Blooded Animals", JWPCF, 348 p. 295, 1962.

82.  Wolf, H.W., "The Coliform Count as a Measure of Water Quality",
     Water Pollution Microbiology,, Edited by R. Mitchell, Wiley-
     lnterscience, New York, 1972.
                                    313

-------
83.  Bott, T.L., "Bacteria and the Assessment of Water Quality",
     Biological Methods for the Assessment of Water Quality, ASTM STP 528,
     American Society for Testing and Materials, 1973^

84.  Kabler, P.W., and Clark, H.F., "Coliform Group and Fecal Collform
     Organisms as  Indicators of Pollution in Drinking Water", Journal AWWA,
     52, p.  1578,  I960.

85.  Geldreich, E.E., Huff, C.B., Bordner, R.H., Kabler, P.W., and
     Clark,  H.F.,  "The Fecal Coli-Aerogenes Flora of Soils from Various
     Geographical  Areas", The Jqurna1 of App1. Bact., 25, 87, 1962.

86.  Eliassen, R., "Coif form Aftergrowths in Chlorinated Storm Overflows",
     Journal of the S.E.D. of ASCE, SA2, 371, 1968.

87.  Weidner, R.B., Christiansen, A.G., Weibel, S.R., and Robeck, G.G.,
     Rural Runoff  as a Factor in Stream Pollution", JWPCF, 41,
     P. 377, 1969.

88.  McFeters, G.A., and Stuart, D.G., "Survival of Coliform Bacteria in
     Natural Waters:  Field and Laboratory Studies  with Membrane-Filter
     Chambers", Appl. Hicrob,, 24, p.  805, 1972.

89.  Kittrell, F.W., and Furfari, S.A., "Observations of Coliform
     Bacteria in Streams", JWPCF, 35,  p. 1361, 1963.

90.  Geldreich, E.E., "Applying Bacteriological Parameters to Recreational
     Water Quality", Journal AWWA, 62, p. 113, 1970.

91.  Kabler, P.W., Clark, H.F., and Geldreich, E.E., "Sanitary
     Significance  of Coliform and Fecal Coliform Organism in Surface
     Waters", Public Health Report, 79, p. 58, 1964.

92.  Geldreich, E.E., "Applying Bacteriological Parameters to
     Recreational Water Quality", Journal AWWA. 62, p. 113, 1970.

93.  Geldreich, E.E., Kenner, B.A., and Kabler, P.W., "Occurrence of
     Coliform,  Fecal Coliform, and Streptococci on  Vegetation and
     Insects",  Appl. Micro., 12, p. 63, 1961.

94.  Van Donsel,  D.J., Geldreich, E.E., and Clarke, N.A., "Seasonal
     Variations in Survival  of Indicator Bacteria in Soil and Their
     Contribution  to Storm Water Runoff", Appl. Micro., 15, p.  1362, 1967.

95.  Kenner, B.A., Clark, H.F., and Kabler,  P.W., "Fecal  Streptococci-I I
     Quantification of Streptococci in Feces", Am.  Journal of Public
     Health, 50,  p. 1553, I960.
                                   314

-------
 96.   Kenner, B.A., Clark, H.F., and Kabler, P.W., "Fecal  Streptococci -I .
      Cultivation and Enumeration of Streptococci  in Surface Waters",
      Appl.  Micro.. 9, p. 15, 1961.

 97-   Slanetz, L.W., and Bartley, C.H., "Detection and Sanitary Significance
      of Fecal Streptococci  in Water", Am. Jour, of Public Health,'
      5A,  p.  609,
 98.   Unz,  R.F., "Fecal Col i form and Fecal Streptococci  in the Bacteriology
      of Water Quality", Water and Sewage Works, 115, R-238, 1968.

 99.   Geldreich, E.E., and Kenner, B.A. ,  "Concepts of Fecal Streptococci
      in Stream Pollution", JWPCF, k] , R336, 1969.

100.   Geldreich, E.E., and Van Donsel, D.J., "Salmonel lae in Fresh  Water",
      Proceedings of the National Specialty Conference  on Disinfection,
      ASCE, p. 495, July 8-10, 1970.

101.   Van Donsel, D.J., and Geldreich, E.E., "Relationship of Salmonellae
      to Fecal Col i form in Bottom Sediments",  Water Research, 5,
      p. 1079, 1971.

102.   Smith, R.J., Twedt, R.M., and Flanigan,  L.K. , "Relationship of
      Indicator and Pathogenic Bacteria in Stream Waters", JWP C F ,
      *t5, P. 1736, 1963.

103.   Dutka, B.J., and Bell,  J.B., "Isolation  of Salmonellae from
      Moderately Polluted Waters", JWPCF, ^5,  p. 316, 1973.

10^4.   Shuval, H.I., Katzenelson, E. , "The Detection of  Enteric Viruses
      in the Water Environment", Water P ol 1 u t i on M i c ro b i o 1 ogy , Edited by
      R. Mitchell, Wi ley- In terse ience, New York, \S72~.

105.   Smith, J.E., and McVey, J.L., "Virus Inactivation  by Chlorine
      Dioxide and its Application to Storm Water Overflow",  presented
      before the Division of Water, Air and Waste Chemistry, American
      Chemical Society, Chicago,  Illinois, August 26-31,  1973-

106.   Moffa, P.E., Tifft, E.C., Richardson, S.L., and Field, R. ,
      "Disinfection Techniques for Point-Source Treatment of Combined
      Sewer Overflows", unpublished paper by O'Brien 6  Gere Engineers,  Inc.
      Syracuse, Mew York, September 1971*.

107.   Geldreich, E.E., Best,  L.D., Kenner, B.A., and Van  Donsel, D.J.,
      "The Bacteriological Aspects of Stormwater Pollution", JWPCF, kQ,
      p. 1861, 1968.

108.   Millipore Application Manual, AM 302, 1973.
                                     315

-------
109.  Cleveland, A.M.,  Reid,  G.W.,  and Walter,  P.R.,  "Storm Water
      Pollution from Urban Land Activity",  ASCE Environmental  Meeting,
      Chicago, Illinois,  October 13-17, 1969.

110.  Booth, F.T., Vardin, N.,  and  Ball,  G.L.,  "Toronto Confronts
      Outdated Sewers",  JWPCF,  38,  p.  1557,  1967.

111.  Evans, F.L., Geldreich,  E.E.,  Weibel,  S.R.,  and Robeck,  G.G.,
      "Treatment of Urban Stormwater Runoff",  JWPCF,  40,  R162, 1968.

112.  Walter, W.G., and Bottman,  R.P., "Microbiological and Chemical
      Studies of an Open  and  Closed Watershed",  Journal of Environmental
      Health, 30,  p. 157, 1968.

113.  Burns, R.J., Vaughan,  R.D., "Bacteriological  Comparison  Between
      Combined and Separate Sewer Discharges in  Southeastern Michigan",
      JWPCF, 30, p. 400,  1966.

114.  Benzie, W.J., and  Courchaine,  R.J., "Discharges  from Separate
      Storm Sewers and  Combined Sewers",  JWPCF,  38, p.  410,  1966.

115.  Mason, D.G., "Treatment of Combined Sewer  Overflows",  JWPCF,
      44,  p. 2239, 1972.

116.  Dunbar, D.D., and Henry,  J.G.,  "Pollution  Control Measures  for
      Stormwaters  and Combined  Sewer Overflows", JWPCF, 38,  p.  9,  1966.

117.  Hendricks,  C.W.,  "Enteric Bacterial Growth Rates  in  River Water",
      Appl.  Micro., 24, pp. 168-174,  August  1972.

118.  Klock, J.W., "Survival of Coliform Bacteria  In  Wastewater
      Treatment Lagoons", JWPCF, 43,  pp. 2071-2083, October  1971.

119.  Little, J.A., Carroll, B.J.,  and Gentry,  R.E.,  "Bacterial
      Removal in Oxidation Ponds",  2nd International  Symposium for
      Waste  Treatment Lagoons,  June  23-25,  1970, Kansas City,  Missouri.

120.  Bordner,  R.H., "Coliform  Recovery Studies", Newsletter,  p.  10,
      Analytical  Quality  Control USEPA, National Environmental  Research
      Center, Cincinnati, January 1974.

121.  "Eutrophication:  Causes,  Consequences, Correctives".  National
      Academy of Sciences,  Washington, D.C., 1969.

122.  Kuentze),  L.E., "Bacteria, Carbon Dioxide, and  Algae Blooms",
      JWPCF, 41,  pp. 1737-1747,  October 1969.
                                  316

-------
123.   Fruh, G.W., "Nutrients and Aquatic Vegetation Effects", Journal
      of the Environmental Engr. DIv. of ASCE, Vol. 100, No. EE2,
      PP. 2*9-4ytt, April \W. - ----
124.  Maler, W.J., and McDonnell, "Carbon Measurements In Water Quality
      Monitoring", JWPCF, 46, pp. 623-633, April 1971*.

125.  Johnson, M.G., and Owen, G.E., "Nutrients and Nutrient Budgets
      in the Bay of QuSnte", JWPCF, 43, pp. 836-853, May 1971.

126.  Bhagat, S.K., Proctor, D.E., and Funk, W.H.,  "Sampling and
      Measurement In the Aquatic Environment", Paper presented  at 25th
      Purdue Ind. Waste Conf . , Laf., Ind,, May 5-7, 1970.  22 pps.

127.  Stewart, K.M., and Rohlich, G.A., "Eutrophi cation - A Review",
      Publication No. 34, State Water Quality Board, State of California,
      1967,

128.  Malhotra, S.K. , and Zanoni, A.E., "Chloride Interferences in
      Nitrate-Nitrogen Determination", Journal AWWA, 62, pp.  568-571,
      September,  1970.

129.  Barth, E.F., "Nutrient Control Chemistry", Proceedings  of ASCE
      Specialty Conference, Rochester Institute of  Technology,  June  26-28,
      1971.

130.  Hetling, L.J., and Sykes, R.M., "Sources of Nutrients in
      Canadarango Lake", JWPCF, 45, pp. 146-156, January 1973.

131.  Sartor, J.D., Boyd, G.B., and Agardy, F.J., "Water Pollution
      Aspects of Street Surface Contaminants", .^£££^46, pp. 458-467,
      March 1974.

132.  Kluesener, J.W., and Lee, G.F., "Nutrient Loading From  a  Separate
      Storm Sewer in Madison, Wisconsin",  JWPCF, 46, pp.  920-936, May  1974.

133.  Whipple, W., Hunter, J.V., and Yu, S.L., "Unrecorded  Pollution
      From Urban Runoff", JWPCF, 46, pp. 873-885, May 1974.

134.  Leary, R.D., Ernest, L.A., Powell, R.S., and  Manthe,  R.M.,
      "200 MGD Activated Sludge Plant Removes Phosphorus by Pickle
      Liquor", USEPA Report No. EPA-67-12/73-050, September 1973.

135   Zanoni, A.E., "Phosphorus Removal by Trickling Filter Slimes",
      USEPA Report No. EPA-R2-73-279, July 1973.

136.  Unpublished data, USEPA Contract No. 68-03-0242, Envlrex  inc.,
      Milwaukee, Wisconsin.
                                    317

-------
 137.  Fernandez, F.J,, and Manning, D.G., "The Determination of Arsenic
      at Sub-Microgram Levels by Atomic Absorption Spectrophotometry",
      Atomic Absorption Newsletter, Vol. 10, No. 4, p. 86, July-August
      WT

 138.  Armour, J.A., and Burke, J.A., "Method for Separating Polychlor-
      inated Biphenyls from DDT and its Analogs", Journal of the ADAC,
      53, p. 674,  1970.

 139.  Young, S.J.V., and Burke, J.A.,  "Micro Scale Alkali Treatment
      for Use in Pesticide Residue Confirmation and Sample Cleanup",
      Bui let tn of _ Environmental Contamination and Toxicology, 7, p. 160,
 140.  Lamberton, J.G., Claeys, R.R., "Large, Inexpensive Oven Used to
      Decontaminate Glassware for Environmental  Pesticide Analysis",
      Journal of the ADAC, 55, p. 898, 1972.

 141.  Hall, E.T., "Variations of Florisil Activity:  Method to Increase
      Retention Properties and Improve Recovery  and Elution Patterns
      of Insecticides", Journal  of the ADAC, p.  1349,  1971.

 142.  "Method for Organochlorine Pesticides in  Industrial Effluents",
      EPA,  National Environmental Research Center,  Analytical Quality
      Control Laboratory,  Cincinnati,  Ohio, 1973.

 143.  "Method for Polychlorinated Biphenyls (PCB's) in Industrial
      Effluents", EPA, National  Environmental  Research Center,  Analytical
      Quality Control  Laboratory, Cincinnati,  Ohio, 1973.

 144.  McDermott, G.N., "Water Characterization Concepts  Predicting
      Potential  Damages and Treatability  of Waste Waters," 17th Annual
      Purdue Industrial Waste Conference, Series No.  112, pp. 803-805,
      May 1962.

 145.  American Public  Works Association,  "Water  Pollution Aspects  of
      Urban Runoff", USEPA Report No.  1 1030DNS01/69,  NTIS-PB 215 532,
      January 1969.

146.  Storm and  Combined Sewer Technology Branch, USEPA,  Edison,  N.J.,
      "Combined  Sewer  Overflow Seminar Papers",  USEPA Report No.
      EPA-670/2-73-077, NTIS-PB  231  836,  1973.

147.  McFarren,  E.F.,  "Determination of Asbestos by Optical  Microscopy",
      AWWA,  Water Quality  Technical  Conference,  Cincinnati,  Ohio,
      December 3-4, 1973.
                                  318

-------
148.   Private Communications with Dr.  EmMio Sturino,  USEPA,  Central
      Regional Laboratory, 1819 W. Pershing Drive,  Chicago,  Illinois,
      June 21,
149.   Micholson, W.J., Maggiore, C.J., and Selikoff,  I.J.,  "Asbestos
      Contamination of Parenteral Drugs", Science,  177,  No.  4044,
      PP. 171-173, July 14, 1974.

150.   Richards, A.L. ,  "Estimation of Submicrogram Quantities of Chryso-
      ti le Asbestos by Electron Microscopy", Analytical  Chemistry,  45, 4,
      pp. 809-811, 1973.

151.   Richards, A. L.,  "Estimation of Trace Amounts  of Chrysotile
      Asbestos by X-Ray Diffraction", Analytical  Chemistry,  44, 11,
      pp. 1872-1873, September  1972.

152.   Patersen, J.E.,  Barbour, R.V., Dorrence,  S.M.,  Barbour,  F.A. ,
      and Helm, R.V.,  "Molecular Interaction of Asphalt", Analytical
      Chemistry, 43, 11,  p. 1491, 1971.

153-   Racine, Wisconsin;  Log Book for Stormwater Treatment  Plant,
      USEPA Grant No.  S800744.

154.   Kenosha, Wisconsin; Log Book for Stormwater Treatment  Plant,
      USEPA Grant No.  11023 EKC.

155.   Glover, G.E., and Herbert, G.R., "MIcrostraining and  Disinfection
      of Combined Sewer Overflows - Phase II",  USEPA  Report  No.
      EPA-R2-73-124, NTIS-PB 219 879, 1973.

156.   Storm and Combined Sewer Technology Branch, USEPA,  Edison, N.J.,
      "Combined Sewer Overflow Seminar Papers", USEPA Report No.
      EPA-670/2-73-077, NTIS-PB 231 836, 1973.

157.   Carcich,  I.G., et al., "A Pressure Sewer System Demonstration",
      USEPA Report No. EPA-R2-72-091 , NTIS-PB 214 409, 1972.

158.   Diaper, E.J., and Glover, G.E., "Microstraining of Combined
      Sewer Overflows", JWPCF, 43, 10, pp. 2101-2113, 1971.

159.   Envi rex,  Inc., "Sequential Screening Treatment  of  Combined
      Sewer Overflows", USEPA Contract 14-12-40;  Unpublished interim
      summary report,  August  1972.

160.   Wadelin, C.W., and Morris, M.C., "Rubber",  Analytical  Chemistry,
      45, 5, PP.  333R-343R, 1972,
                                   319

-------
 161.  Krishen, A., "Quantitative Determination of Natural Rubber
      Styrene-Butadlene Rubber and Ethylene-Propylene-Terpolymer Rubber
      in Compounded Cured Stocks by Pyrolysis-Gas Chromatography",
      Analytical Chemistry, 44, 3, pp. 494-497, 1972.

 162.  Beeton, A.M., "Changes in the Environment and Biota of the Great
      Lakes", Eutrophication, National Academy of Sciences, pp. 150-187,
      1969.

 163.  Hites, R.A., "Analysis of Trace Organic Compounds in New England
      Rivers", Journal of Chromatographic Science,  11, p. 11, November
      1973.

 164.  Burnham, A.K., Calder, G.V., Fritz, J.S., Junk, G.A., Svec, H.J.,
      and Vick, R., "Trace Organics in Water:  Their Isolation and
      Identification", Journal American Water Works Association, 65, 11,
      pp. 722-725, November 1973.

 165.  Mieure, J.P., and Dietrich,  M.W., "Dtermination of Trace Organics
      in Air and Water", Journal of Chromatographic Science, 11, 11,
      Pp. 559-570, November 1973.~

 166.  Ahnoff, M. and Josefsson, B., "Simple Apparatus for On-Site
      Continuous Liquid-Liquid Extraction of Organic Compounds from
      Natural Waters", Analytical  Chemistry, 46,  6, pp. 658-663, May 1974.

 167.  Goldberg, M.C.,  and Delong,  L.,  "Extraction and Concentration of
      Organic Solutes  from Water", Analytical Chemistry, 45, 1, pp. 89-93,
      1973.

168.  Street Surface Sampling and  Analytical Techniques, Personal
      Communication from Woodward-Clyde Consultants to USEPA, Edison,
      New Jersey,  April  3,  1975.

169.  Colston,  N.V., "Characterization and Treatment of Urban Land
      Runoff",  USEPA Report No. EPA-670/2-74-096, NTIS-PB 240 987/AS,
      1974.

170.  Shaheen,  D.U., "Contributions of Urban Roadway Usage to Water
      Pollution",  USEPA Report No. EPA-600/2-75-004, 1975.

171.  Thompson, R.N.,  New,  C.A., and  Lawrence,  C.H., "Identification
      of Vehicle Tire  Rubber in Roadway Dust",  American Industrial
      Hygiene Association Journal, 27, p. 488,  1966.

172.  Kirshen,  A.,  "Quantitative Determination  of Natural Rubber,  Styrene-
      Butadiene Rubber,  and Ethylene-Propylene-Terpolymer Rubber in
      Compounded Cured Stocks by Pyrolysis - Gas  Chromatography",
      Analytical  Chemistry, 44, p. 494,  1972.
                                    320

-------
173.   "Criteria for a Recommended Standard...Occupational  Exposure to
      Asbestos", U.S. Department of Health, Education and  Welfare, Public
      Health Service, Health Services and Mental Health Administration,
      National  Institute for Occupational Safety and Health,  1972.
                             SECTION VII


  1.   Condon, Francis J., "Methods of Assessment of Non-Point Runoff
      Pollution", The Diplomat, December 1973.

  2.   Shaheen, D.G., "Contributions of Urban Roadway Usage to Water
      Pollution",   USEPA Report No.  EPA-600/2-75~00*»,  1975.

  3.   American Public Works Association, "Water Pollution Aspects of
      Urban Runoff", USEPA Report No. 11030QNS01/69, NTIS-PB 215 532,
      January 19&9.

  A.   Metcalf and Eddy Engineers, "Storm Water Management Model, Vol.  1,
      Final Report", USEPA Report No. 1102ADOC07/71, NTIS-PB 203 289,
      July 1971.

  5.   Amy. G., Pitt, R., et al., "Water Quality Management Planning
      for Urban Runoff", USEPA Report No. EPA-4W9-75-004, 1975.

  6.   Lager, J.A., and Smith, W.G., "Urban Stormwater Management and
      Technology;  An Assessment", USEPA Report No. EPA-670/2-7^-040,
      NTIS-PB 240867/AS, 1971*.
                             SECTION VI I I
  i.  Lager, J.A., and Smith, W.G., "Urban Stormwater Management and
      Technology:  An Assessment", USEPA Report No.  EPA-670/2-74-0^0,
      NTIS-PB 240867/AS, 1974.

  2.  "North Water Filters,  'Fine' Screens for Industrial  Use",
      Manufacturers Bulletin.

  3.  "Zurn Micro Matic Screening Systems", Manual  121,  Manufacturers
      Builetin.

  i».  "Beloit Passavant Micro-Sieve", Bulletin 1660, Manufacturers
      Builetin.
                                   321

-------
5.  Hansen, C.A., "Absorption/Adsorption in Contact Stabilization",
    Water and Sewage Works, February 1974.

6.  Antonie, R.L., Kluge, D.L.,  and Mielke, J.H.,  "Evaluation of
    a Rotating Disk Wastewater Treatment Plant",  JWPCF, 46  3
    March 1974.                                   	
                                  322

-------
                          SECTION X - GLOSSARY

APWA - American Public Works Association

ASTM - American Society for Testing and Materials

ATP - Adenosine tri-phosphate

Average Efficiency - The mean value of percent removals achieved by a
treatment facility for a series of events.

Bleedback - Dewatering a storage facility by means of gravity drainage
to a sewerage system.

BOD - Unless otherwise noted, this refers to the standard five day Bio-
chemical Oxygen Demand test.

BODx ~ This refers to a Biochemical Oxygen Demand test with an incubation
time of x days.

COD - Unless otherwise noted, this refers to the standard Chemical
Oxygen Demand test.

Composite Sample - Consists of more than one, and usually a number of
samples which have been mixed together for analytical purposes.

CSO - Combined Sewer Overflow - any discharge from a combined sewer,
usually associated with the addition of stormwater to the combined
sewer system.

DAF - Dissolved Air Flotation

Dewater - The removal of a stored amount of a storm or combined sewer
discharge from a storage facility to further treatment or ultimate
disposal.

Dry-Weather Flow - That flow in sanitary or combined sewers that contains
no stormwater.

EPA WQ.O Manual - Methods for Chemical Analysis of Water and Wastes 1971,
USEPA Renort No. 16020---07/71, July 1971.

FC - Fecal Coliform

First Flush - A term used to describe the early portion of a storm or
combined sewer discharge which oftentimeshas a significantly greater
contaminant concentration than the remainder of the discharge.
                                   323

-------
£S_ - Fecal Streptococci

Grab Sample - This is a single simple sample taken at any particular
instant and analyzed by itself.  A number of grab samples, either
taken randomly or according to a programmed schedule can be mixed
together to form a composite sample.

Hyetograph - A graph plotting rainfall intensities during various time
increments versus time.

MF_ - Membrane Filter

MPN - Most Probable Number

PCB - Polychlorinated Biphenyls

Pilot System - A facility used for testing a process on a scale much
smaller than would actually be required.   Subject to scale-up errors.

Pollutograph - A graph plotting pollutional loading in mass per unit
time versus time or pollutional concentration per unit time.

Prototype - A facility used for testing a process either at full scale
or at a scale such that no scale-up errors would be expected if full
scale application were to follow.

Pumpback - Dewatering a storage facility using pumps.

Storm Generated Pi scharge - Any discharge from a storm or combined sewer
resulting from a precipitation event.

Stormwater - The water resulting from a precipitation event which may
stay on the land surface, percolate into the ground, run off into a
body of water, enter a storm sewer, enter a combined sewer, infiltrate
a sanitary sewer, or evaporate.

Stormwater Runoff - That Stormwater flowing overland.

Storm Sewer Discharge - The discharge from a storm sewer resulting from
Stormwater runoff entering the storm sewer system.

SWMM - Abbreviation for the EPA Stormwater Management Model.

T£ - Total Coliform

TOC - Total Organic Carbon

TOD - Total Oxygen Demand
                                   324

-------
Weighted Efficiency - An expression of the percent removal achieved by
a treatment facility for a series of events determined by the total
mass removal divided by the total mass treated from all the events.

Wet-Weather Flow - The flow in sanitary or combined sewers that contains
some stormwater.
                                    325

-------
                                   TECHNICAL REPORT DATA
                            (Please read Instructions on the reverse before completing)
 1. REPORT NO.
  EPA-600/2-76-145
                                                          3. RECIPIENT'S ACCESSIOf*NO.
4. TITLE AND SUBTITLE
  METHODOLOGY FOR THE  STUDY  OF  URBAN STORM GENERATED
  POLLUTION AND CONTROL
             5. REPORT DATE
              August  1976 (Issuing Date)
             6. PERFORMING ORGANIZATION CODE
 7. AUTHOR(S)
  Richard E. Wul1schleger,  Alphonse E.  Zanoni,
  and Charles A. Hansen
                                                          8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
  Envi rex  Inc.
  A Rexnord Company
  Environmental  Sciences  Division
  Milwaukee, Wisconsin  53214
              10. PROGRAM ELEMENT NO.

               1BC6H
             11. CONTRACT/
              68-03-0335
                             NO.
 12. SPONSORING AGENCY NAME AND ADDRESS
  Municipal Environmental  Research Laboratory
  Office of Research  and  Development
  U.S. Environmental  Protection Agency
  Cincinnati, Ohio  45268
                                                           13. TYPE OF REPORT AND PERIOD COVERED
             14. SPONSORING AGENCY CODE

               EPA-ORD
 15. SUPPLEMENTARY NOTES
  Project Officer:   Richard  Field,  Phone 201/548-3347, ext.  503
 16. ABSTRACT
  This  report  contains  recommendations for standard precedures  to be followed  in the
  conduct of projects  dealing with pollution assessment  and  abatement of storm gen-
  erated discharges.   The  purpose of this project was  to develop standard procedures
  needed to  insure  that all  discharges and treatment processes  could be evaluated by
  the same means.   The  procedures chosen were those found to be the most applicable
  and optimum  for the  field  of storm and combined sewer  overflow pollution control.

  The project  efforts were devoted to the major areas  listed below.
        I.  Recommended  methods for sampling and sample preservation.
        2.  Appropriate  monitoring instrumentation available.
        3.  The choice of quality parameters to be utilized.
        4.  The analytical  procedures to be followed.
        5.  The methods  for evaluating storm generated  discharge pollution.
        6.  The standard procedures for evaluating treatment  processes treating storm
           generated flows.

  Choice of  the recommended  procedures was based upon  the U.S.  EPA research and demon-
  stration project  reports in this and associated fields, other published literature,
  ongoing U.S.  EPA  funded  projects,  and the contractor's experience  in the field of
  stormwater pollution  control.	
 7.
                               KEY WORDS AND DOCUMENT ANALYSIS
                  DESCRIPTORS
                                             b.IDENTIFIERS/OPEN ENDED TERMS
                           c. COS AT I Field/Group
Water Pol 1uti
Water Runoff,
Quali ty,  Cost
f1ows—Sewers
 hemical  Anal
riology,  Samp
(Meteorology)
on, Sewage Treatment,  Surface
 Sewage, Contaminants,  Water
 Analysis, Storm Sewers,  Over-
,  Combined Sewers,  Hydrology,
ysis, Water Chemistry,  Bacte-
 ing, Samplers, Precipitation
--Hydrology
.Water Pollution  Control/
Treatment,  Combined Sewer
Overflow,  Storm  Sewer Dis
charges,Treatment  Process
Evaluation, Sample  Preser-
vation,  Laboratory Pro-
cedures,  Precipitation
Monitoring
       13B
 3. DISTRIBUTION STATEMENT
     RELEASE TO  PUBLIC
                                             19. SECURITY CLASS (ThisReport)
                                                   UNCLASSIFIED
                           21. NO. OF PAGES
                                 342
20. SECURITY CLASS (This page)
     UNCLASSIFIED
22. PRICE
EPA Form 2220-1 (9-73)
                                           326

-------