United States
                     Environmental Protection
                     Agency
Municipal Environmental Research
Laboratory
Cincinnati OH 45268
                     Research and Development
EPA-600/S2-83-124  Jan. 1984
SEPA          Project  Summary
                     Guidelines for the  Monitoring  of
                     Urban  Runoff  Quality
                     Michael B. Sonnen
                       This study was undertaken to define
                     adequate monitoring of urban runoff for
                     various  objectives and to develop a
                     guidebook for those contemplating a
                     program of measuring urban runoff in
                     the field. Emphasis is placed on measur-
                     ing urban runoff while it is still flowing
                     on land and street surfaces rather than
                     after it has been integrated with other
                     waters in storm or combined sewers.
                       The study stresses the need to state
                     the exact objective of any field monitor-
                     ing program, since costs and require-
                     ments vary greatly depending on the
                     purpose of the monitoring. Six possible
                     objectives are identified for urban runoff
                     monitoring programs (scientific re-
                     search, problem identification, alterna-
                     tive solution monitoring, support of the
                     final design,  regulatory compliance
                     monitoring, and operational perform-
                     ance monitoring).
                       The study also reviews current and
                     recent monitoring programs, literature
                     on monitoring program design, and
                     monitoring strategies for receiving
                     waters and best management practices.
                     Monitoring and protocols are developed
                     for satisfying various objectives, and a
                     set of guidelines is developed to assist
                     designers and planners of  monitoring
                     programs for urban runoff.
                       This Project Summary was developed .
                     by EPA's Municipal Environmental Re-
                     search Laboratory, Cincinnati,  OH, to
                     announce key findings of the research
                     project that is fully documented in a
                     separate reoort of the same title (see
                     Project Report ordering information at
                     back).

                     Introduction
                       Storm  and snowmelt runoff from urban
                     lands contains constituents that  can,
under certain circumstances, occur  at
damaging, polluting levels. These circum-
stances  and  the phenomena and pro-
cesses that produce the runoff  quality
levels are so poorly understood  and  so
inherently random and variable that only
field measurements can reliably charac-
terize them.
  Sometimes it is necessary or desirable
to have  extremely detailed information
about the processes that cause deterior-
ation in the quality of urban runoff. To get
such information, an expensive, compre-
hensive  field monitoring program must
be launched. By contrast, a city planner or
engineer, for example, may simply need
information about the  levels of  one  or
more constituents in local  storm runoff.
Such a need can be met by a fairly cursory
or modest monitoring program. This study
defines what adequate  monitoring is for
specific objectives and develops a guide-
book for those contemplating  a field
measurement program.
  Historically, runoff has generally been
sampled after it  has entered storm  or
combined sewers and  been  integrated
with waters  from various sources. By
contrast, this study emphasizes the qual-
ity of runoff still flowing on the land and
street surfaces. Some discussion is also
presented of flows and qualities of runoff
in storm and combined sewers or  at
points of combined sewer overflows.
  This study especially concentrates on
statistical requirements and  tests for
sufficiency or significance of data sets,
because  these statistical considerations
affect  the success of each monitoring
objective. Though insights to monitoring
program  design can be gained through
the use  of statistical  inferences and
hypotheses, considerable art is required
in specifying sample numbers, levels  of

-------
confidence, etc  In particular, the investi-
gator should be able to anticipate the
general behavior of the phenomena being
sampled, or considerable sampling effort,
time, and money may be wasted. These
concepts are developed in detail through-
out the report.

Procedures

Developing Monitoring
Objectives
  Since the requirements and costs of
monitoring  urban  runoff vary greatly
depending on the purpose of the monitor-
ing, this study stresses the need to state
the exact objective of the field monitoring
in each case. Six possible objectives were
identified: (1) Scientific research, (2) prob-
lem identification, (3) alternative solution
monitoring, (4) support of the final design,
(5) regulatory compliance monitoring, and
(6) operational performance monitoring.

Reviewing the Objectives and
Successes of Current and
Recent Monitoring Programs

  Past  monitoring programs reported in
the literature were reviewed, and  the
results and conclusions were compared
with the originally stated objectives.

Reviewing Literature on
Monitoring Program Design
  Past  works on  monitoring  of  urban
runoff  were  reviewed to  indicate  the
state-of-the-art  in monitoring protocol
design.

Reviewing and Developing
Monitoring Strategies for
Receiving Waters and Best
Management Practices

  Reviews were made of special-purpose
monitoring for characterizing impacts of
urban runoff on receiving waters and for
planning or monitoring performance of
certain runoff pollution control measures.
Guidance is given where weaknesses are
apparent.

Develop Monitoring
Protocols for Satisfying
Objectives

  The  study develops statistical infer-
ences  of reliable  numbers of sampling
sites and sample numbers for each of the
six monitoring objectives selected.
Designing and Producing a
Guidelines Report
  A guidebook was developed to describe
all the foregoing material for those con-
templating a field measurement program.
Minimum measurements were described
for satisfying various objectives.

Summary Discussion

  Storm and snowmelt runoff from urban
lands contains quality constituents which
under  certain  circumstances occur in
damaging, polluting levels. These circum-
stances, the phenomena and processes
making runoff quality levels what they
are, remain so poorly understood and are
so inherently random  and variable  that
only field  measurements  can be relied
upon to characterize their  behavior.
  Sometimes it is necessary or desirable
to learn in great spatial  and temporal
detail what the processes  are that cause
deterioration in the  quality  of  urban
runoff. In  these  cases an  expensive,
comprehensive field monitoring program
must be launched to learn the required
information. By contrast there are other
instances  in which  a planner or  city
engineer, for example, may wish to learn
whether or not a particular constituent or
several constituents in storm runoff attain
potentially damaging levels in his or her
community. In this case, a fairly cursory
and  modest  monitoring  program  will
suffice to satisfy this objective.
  Statistical theory provides some guid-
ance to the numbers of samples required
for satisfying various objectives for moni-
toring. Buta number of arbitrary variables,
such as  the degree  of  confidence or
acceptable error must be specified by the
monitoring program designer. Moreover,
many statistical tests require that charac-
teristics of some sampled data already be
available,  such as the sample  standard
deviation in the t-test, before they can be
used to give guidance on the acceptability
of sample numbers for still further moni-
toring.
  Numbers of samples for example pro-
grams  specified for  satisfaction of six
different  monitoring  objectives ranged
from 24 for a design objective to 54,000
for a comprehensive scientific, process-
monitoring objective. Costs for monitoring
varied proportionately,  from  about
$11,000  to $9,300,000, for  the  two
respective objectives.
   Detailed sampling and laboratory anal-
yses are expensive. Sometimes the ex-
pense of sampling and analysis will prove
prohibitive. The major point of this work.
however, is that curtailment of monitoring
for reasons of expense or for any other
reason can mean that the original moni-
toring objective cannot be met. If one sets
out to elucidate runoff quality processes
or to build and verify a model of these
causative processes and finds that he can
afford only $75,000 per year to do the
indicated  monitoring,  then  in  fact  he
cannot adequately define processes or
verify his model. The best he can hope for
is to monitor operational performance, to
gather information on what happened,
what  the  effects  were, and not what
caused those effects to be realized. The
information gained  is likely not to be
sufficient to allow prediction of what will
happen in  other places or under changed
conditions.
  Urban runoff quality in the early 1980's
is not considered a high priority  item of
national environmental interest, except
in those instances where urban runoff is
conveyed  and discharged through sys-
tems. Runoff quality  control remains a
new field  of engineering endeavor. It is
natural, therefore, that design  criteria,
discharge  standards and receiving water
criteria, and scientific knowledge about
this field are still preliminary, incomplete,
and rudimentary. For these same reasons,
however,  it  is virtually axiomatic that
mistakes may be made, monitoring and
construction monies may occasionally be
misspent on unnecessary or unsupported
or inadequate facilities or monitoring
programs. It  is to be  hoped  that  this
guidebook will assist in keeping straight
what is adequate monitoring for separate
objectives and what is not.

Conclusions

1. Urban runoff may be sampled for a
   variety of reasons, i.e., to satisfy any
   of a number of objectives. The ob-
   jective in each case should be stated
   explicitly and restated as necessary
   in concrete,  numerical  terms  so a
   program of sampling can be devised
   that will clearly satisfy that objective.

 2.  Sampling in the field is expensive.
    Laboratory analyses are both expen-
    sive and relatively inexpensive, de-
    pending  on which constituents are
    being analyzed.  But economies of
    scale in  laboratory analyses are
    available, and opportunities should
    be sought for learning more for the
    same total budget by having  more
    analyses  performed  and  sampling
    fewer events in the field.
i

-------
3.   Statistical theory that exists today is
    somewhat useful in guiding decisions
    about sampling program design. Cer-
    tain concepts from statistics, such as
    the existence of Type I and Type II
    errors,  should be appreciated since
    they are axiomatic and unavoidable.
    Statistical tests such as the t-test,
    the F-test, and  the  chi-square test
    are useful and instructive and worthy
    of  consideration, but they do  not
    function usefully in the absence of
    any prior sample information. That is
    very frustrating  but  inescapable.
    Statistical theory on the whole is not
    developed sufficiently to assist mean-
    ingfully in multiple-site,  multiple-
    phenomena, multiple-constituent,
    stochastic-process sampling.
4.  Mathematical models existing today
    for estimation of urban runoff quality
    loads do not have a solid scientific
    base. Some do  not  require or even
    give regard to  the satisfaction of
    mass continuity. Directly as a result,
    they are either difficult or impossible
    to  calibrate or they must be recali-
    brated  with practically  every use.
    Because they are so limited in trans-
    ferability and even  the  most basic
    level of scientific rigor, they cannot
    be relied upon  solely to  replace or
    obviate the  need for local  quality
    sampling, regardless of the objective.
5.  Research needs to be done to formu-
    late  the  physics and chemistry of
    urban  runoff  behavior at several
    levels of sophistication and approxi-
    mation. Just as  it is not necessary to
    monitor each drop of rainfall, each
    clay particle, or each liter of runoff to
    design a detention  basin, it  is not
    necessary to model each of these
    things  rigorously to design the basin
    without sample  results. But design
    "theory" does need to be developed
    that is consistent  with  scientific
    principles, that allows for variability
    in  behavior  both chemically  and
    physically, both  quality-wise  and
    hydraulically, and that approximates
    at  least the  continuity of mass  and
    the processes of scour and deposition
    and chemical equilibria and rates of
    change of concentration with time
    and with changes in chemistry of the
    incoming and  surrounding  fluid.
    Current  models in wide use  are
    predicated on an assumed first-order
    decay or wash-off or first-flush phe-
    nomenon, which data frprn Denver
    and other places have clearly indi-
    cated does not always occur. Concen-
6.
trations or masses of pollutants in
runoff do not always decrease expo-
nentially or at all with time after the
onset of rainfall. Current theory on
the behavior of urban runoff and its
quality constituents is  poor  at best
and  misleading or even wrong at
worst. Hydraulic behavior is much
better understood,  and  it is approxi-
mated adequately at various levels of
study detail, i.e., for use with varying
scientific, planning, design, or moni-
toring objectives.  Quality  theory
should be developed and advanced in
a commensurate fashion.

There  is no guarantee that water
quality data, no matter how carefully
collected, will be transferable to other
areas and other circumstances. How-
    ever,  because data  collection  and
    interpretation  is so  expensive  and
    time-consuming, EPA is encouraged
    to maintain a data base, such as that
    now started at  the University of
    Florida, wherein runoff quality data
    from across the nation may be stored
    and accessed by subsequent users.
    But it is recommended that data so
    stored  be characterized by the con-
    tributor as to their statistical proper-
    ties (numbers of samples, variance,
    and the like) and as to the conditions
    and purposes  under and for which
    they were collected.
  The full report was submitted in fulfill-
ment of Grant  No.  R-806704  by the
Denver Urban Drainage & Flood Control
District under the sponsorship of the U.S.
Environmental Protection Agency.
   Michael B. Sonnen is with the M. B. Sonnen Co.. Inc., Walnut Creek, CA 94598.
   Douglas Ammon is the EPA Project Officer {see below).
   The complete report, entitled "Guidelines for the Monitoring of Urban Runoff
     Quality." (Order No. PB 84-122 902; Cost: $14.50. subject to change) will be
     available only from:
          National Technical Information Service
          5285 Port Royal Road
          Springfield,  VA 22161
          Telephone: 703-487-4650
   The EPA Project Officer can be contacted at:
          Municipal Environmental Research Laboratory
          U.S. Environmental Protection Agency
          Cincinnati. OH 45268
                                                *US. GOVERNMENT PRINTING OFFICE 1984-759-015/7267

-------
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
  BULK RATE
U.S. POSTAGE
    PAID
Cincinnati, Ohio
Permit No. G35
Official Business
Penalty for Private Use $300
             MEĞLoOtt209l
            CHICAGO  1L  60605

-------