United States
                                Environmental Protection
                                Agency
                Risk Reduction Engineering
                Laboratory
                Cincinnati OH 45268
                                Research and Development
                EPA/600/M-89/014  Jan.1990
         &EPA      ENVIRONMENTAL
                                 RESEARCH
                 Methods of Analysis for Waste Load Allocation

                    J. Wayland Eheart,1 Jon C. Liebman,1 and E. Downey Brill, Jr.2
 Abstract

 This research has addressed several unresolved questions
 concerning the allocation of allowable waste loads among
 multiple wastewater  dischargers within  a  water quality
 limited stream segment. First, the traditional assumptions
 about critical design conditions for waste load allocation
 were shown to be true (except for some highly uncommon
 situations) in multi-discharger settings—namely, that lower
 streamflows and higher temperatures lead to  more
 stringent allowable  loads. Second, a  method was
 developed  for  aggregating dischargers together into
 discrete groupings so that the  water  quality  interactions
 between groups was minimized. This allows waste load
 allocations to  be made separately for each grouping,
 thereby simplifying the overall computational  process.
 Third, the issue of setting aside unallocated waste load
 capacity as a reserve against future growth or modeling
 uncertainty  was  examined. A case study  illustrated the
 unique relationships that might exist between this reserve
 capacity, the frequency of water quality excursions, and
 the cost of  wastewater treatment. Finally, a method was
 developed for designing multi-discharger seasonal  waste
 load allocations. It determines seasonal discharge limits for
 each discharger that minimizes the degree of treatment
 necessary  to provide  the same risk of  water quality
-- excursion as would exist under a nonseasonal waste load
 allocation.

 This  Research  Brief was developed by  the  principal
 investigators and EPA's  Risk  Reduction  Engineering
  1 University of Illinois, Urbana, IL 61801
  2North Carolina State University, Raleigh, NC 27695
Laboratory, Cincinnati, OH, to announce key findings of the
research project that is fully documented in the  reports
and publications listed at the end.

Introduction

Waste load  allocation  is the  process  of determining
allowable levels of pollutant discharges that will maintain
acceptable receiving water quality.  Several factors can
complicate this task. The presence  of multiple pollutant
sources, including the need to find an  economic and
equitable  allocation of  allowable  load among these
sources, is one such factor. Another is the  need to account
for natural variations in receiving water conditions such as
streamflow and temperature when determining  the effects
of a proposed load allocation on water quality. Additional
concerns  may involve accommodating  possible future
discharge sources and uncertainty in water quality
modeling  predictions when conducting a waste  load
allocation.

This project has addressed several previously  unresolved
questions related  to  allocation of  waste  loads to
dischargers along  a stream segment.  The questions are:

   1.  Parameters that affect water  quality, such as
       reaeration coefficient, and^traveMime^hange with
       streamflow  and  temperature in  a  manner that
       sheds doubt on the traditional assumptions of low
       flow and  high temperature  as the "worst-case"
       conditions for decaying pollutants. What are the
       limitations for the validity of these  assumptions?

   2.  On a large river there is usually an unmanageably
       large number of dischargers. Is  there a way to

-------
        split them  into groups that can  be analyzed
        separately  thereby reducing the  computational
        burden?

    3.  How should part of the allowable load capacity be
        reserved  and how will this  affect the goals of  a
        waste load allocation program?

    4.  Seasonal (or periodic) discharge limits are used
        by  many states  to  reduce  the costs of  water
        quality control  programs. How can this approach
        be   applied   in  an  economically   and
        environmentally  sound  manner  to multiple
        discharger settings?

A brief discussion of the research findings related to each
of these questions follows.
Critical Water Quality Conditions	

The assumption that the worst water quality occurs at the
lowest streamflow may  not  always hold  in  instances
involving multiple  discharges of  nonconservative
pollutants. The additional dilution resulting from increased
streamflow may be  offset  by adverse  changes  in  the
parameters that govern water quality and  in decreased
travel time, which allows the stream less time to recover
from the effect of one discharge before receiving another.

The  question  is whether,  with multiple  sources  of
nonconservative  pollutants,  water quality might  worsen
with  increasing streamflow. This  can be  examined by
finding the pattern of discharge that maximizes the rate of
change of critical  pollutant  concentration  (i.e.,  the
concentration at the location where the quality standard is
binding) with respect to streamflow. If that rate is positive,
the standard will  be  violated  with   an  increase  in
streamflow. If, regardless of the loading pattern, the rate is
negative, then it is impossible to have such a situation and
it may be assumed that the lowest streamflow is the worst
slrearnflow.

Under constant temperature  conditions,  a  linear
programming model showed that this discharge pattern is
a uniformly  distributed load along the  entire  length  of
stream.  This  suggests that  streams receiving a  large
number  of discharges may be more susceptible  to
concentration increasing  with decreasing flow than  are
streams receiving a small number of discharges [Eheart,
1988].

For substances that decay according to first-order kinetics,
such as biochemical oxygen demand (BOD), whether the
maximum  value of  the rate  of  change of  critical
concentration with  respect  to  streamflow  is positive  or
negative  depends on the value  of  the  exponent  in  the
power law that relates stream velocity to streamflow.  For
dissolved oxygen  (DO)  deficit,  the exponent relating
reaeration coefficient  to  streamflow is  also important.
Results indicate that the rate is negative for most natural
streams for both types of water quality parameters. Thus,
the traditional assumption that the lowest streamflow is the
worst from a water quality perspective will usually be valid.
Exceptions,  however,  could occur  in  highly  polluted
backwater (impounded)  stream reaches for which  depth
decreases as velocity increases.

For first-order pollutants,  low  temperature is the  most
pessimistic condition, since the pollutant decays  more
slowly and is hence present in larger concentration.  For
DO,  however, the question is more  complicated. The
classical assumption that  the  lowest  DO occurs at  the
highest temperature  may not always  hold. The  DO
saturation  concentration decreases monotonically with
increasing  temperature,  lowering the DO,  but  the
reaeration coefficient  increases  monotonically  with
increasing temperature, tending to raise it.  The  BOD decay
coefficient increases  monotonically  with  increasing
temperature, lowering the DO for single discharges but not
necessarily for  multiple  discharges. (Lower decay rates
attending lower temperatures could result in low DO  at the
point  where the effect  of  one discharge meets that of
another.)
The question  of""whether  Dt)  rmgTft"under'some
circumstances worsen  with decreasing temperature has
been addressed [Eheart and  Park,  1989]. A linear
programming model showed that, for a uniform stream at
constant streamflow, the discharge pattern that maximizes
the  rate of  change  of critical DO with  respect  to
temperature is  a uniformily  distributed  load along  the
entire  length  of stream. This  suggests that streams
receiving a  large number of discharges may be  more
susceptible to DO increasing with decreasing temperature
than are streams receiving a small number of discharges.

The maximum rate of change of critical DO with respect to
temperature depends  on temperature, the  DO  standard
(C*), and  the temperature  adjustment  factor  for the
reaeration coefficient (4>). It does  not depend on the BOD
decay coefficient or  its temperature adjustment factor. For
the maximum reported value of  of I.047, the assumption
that DO decreases  monotonically with  increasing
temperature is valid for C* greater than about 6 mg/L. This
assumption breaks down, however, for  values just above
the range reported in the literature and for C* values just
below the normally chosen range of 5 to 6 mg/L.

Discharger Grouping

ln_deyelgping programs for regulating waste discharges
into  wafer "'" bodies'"it" is- ofterHcbriveniehtand,- frvmariy
cases,  necessary,  to  subdivide the dischargers  into
independently administered  groups. If  each  group  of
dischargers is to be responsible for the water quality at a
particular group  of  checkpoints,  the groups  should  be
chosen so that the effect of the dischargers  of one group
on the checkpoints of another is minimal.

A heuristic method was developed that groups dischargers
on the basis of minimizing the effects of the dischargers
included in  a group  on checkpoints associated with other
groups of dischargers [Eheart  et al.,  1989].  The method
was  illustrated with the use of data for several river basins,
viz.,  the Lower Fox River in  Wisconsin,  the Willamette
River in Oregon, and the Mohawk River  in  New  York.
Several different grouping  criteria were  chosen  to
represent the effect of the excluded dischargers, and it
was  observed that, for the selected rivers, as examples,

-------
 the groupings are somewhat insensitive to the choice of
 this criterion when the excluded effect is low, but not when
 it is high.

 Reserve Capacity

 To  act  as a  hedge  against any potential  errors  or
 inaccuracies in the predictions made with the water quality
 model  and/or  to  allow  for  future  growth  in waste
 discharges, some of the river's allowable load capacity can
 be  set  aside  and  made  unavailable  for use by  the
 dischargers. Depending  on how the reserve  capacity is
 actually incorporated into the  waste  load  allocation
 process, however, the effects on water quality could vary.

 To  include reserve capacity in such  a process, several
 criteria should  be considered and  evaluated.  Acceptable
 levels  of  water quality  should be  maintained at  an
 acceptable probability and an acceptable cost. In addition,
_the: methods for inciudirjg reserves  should be equitable in
 terms of the distribution of required treatment  levels and
 costs.

 Approaches have  been examined for  including a reserve
 capacity in the waste load allocation process, and methods
 have been developed for  analyzing  these approaches in a
 given application  [Michels, 1987].  Relationships among
 reserve  capacity,  frequency of water  quality  excursions,
 and  total  cost have  been  examined.  It was  found  that,
 depending  on  how reserve is incorporated into  a  waste
 load allocation, these relationships  may be quite different
 and  the effect  on overall  water quality may  vary. The
 relationships were illustrated quantitatively with  the use of
 linear programming  models  for  three load  allocation
 programs employing Lower Fox River data.

 Seasonal Waste Load Allocation

 Water quality management programs  that  allow  different
 waste discharge rates during different times of the year are
 an  innovative approach  for reducing  the  cost of  waste
 treatment. Under such programs, referred to as  periodic or
 seasonal discharge programs, the rate of waste discharge
 allowed at a given time  is based on  the  assimilative
 capacity of the receiving  water body during that time, the
 water quality  goals of the  river  basin authority,  and the
 acceptable risk o_L water guajjty, violation. The number and
 length of the seaso'ris"a'?e' typically chosen considering the
 physical conditions  of the stream, the waste treatment
 limitations of  the dischargers,  and the  estimated
 administrative  burden  and  treatment  cost savings
 associated with  each potential  combination  of season
 number and length.

 An important criterion of  seasonal  discharge programs is
 that the degree  of water quality protection achieved under
 such programs should be the same  as  that achieved under
 an  accepted or existing  nonseasonal  discharge program
 for the same river basin. One expression of this criterion is
 referred to as risk equivalency [Rossman, 1989] because it
 requires that  the risk of violating  a given water quality
 standard under the seasonal  discharge program be  no
 greater than that  allowed under an existing  nonseasonal
 discharge program.  Rossman  [1989]  describes  an
 approach for designing risk equivalent seasonal discharge
limits for single discharger stream segments.  Under this
approach, risk is defined as the probability of incurring one
or more water quality violations in any given year, and the
level of risk allowed may be related to the return period  of
the design streamflow used  in  an  existing nonseasonal
discharge program. For example, the 10-year return  period
associated with the annual 7-day, 10-year (7Q10) low flow
would correspond to a 10% risk of water quality violation  in
any  year. The  seasonal  discharge  limits for a  single
discharger under  this  approach are  designed  to  maintain
risk equivalency  with a nonseasonal discharge limit while
minimizing the waste treatment effort of that discharger.

In this research,  risk equivalent seasonal waste discharge
programs for stream  segments  with several  dischargers
were examined  [Lence  et  al., 1989].  Although this
approach is similar to that proposed by Rossman [1989], it
was  modified to accommodate river segments with several
dischargers. Two  management objectives  were proposed
as, substitutes for_-minimizing_waste-treatment-effort, 4he
minimum average uniform  treatment  (UT) and the
maximum total discharge  (TD)  objectives. Under the UT
objective, a seasonal set of uniform percent removal levels
is designed  such that the average percent removal over
the year is minimized and a limit on the frequency of water
quality violations is  satisfied.  Under  the TD objective, the
total  design waste load in each season (i.e., the sum of the
waste loads for the individual dischargers) is determined
such that the TD over all seasons is maximized and  a limit
on the frequency of water quality  violations is satisfied.

The  seasonal waste discharge programs that  meet these
objectives  were  demonstrated  for  managing   BOD
discharges  on  the Willamette  River.  Streamflow  and
temperature data for  the  middle fork  of  the  Willamette
River, the stream  segment  on which  the  10  major
dischargers  are  located, were obtained from  the  United
States Geological  Survey. The dischargers' locations,
waste load  characteristics, design flows,  and treatment
cost  data were taken from Kilgore [1985].  Alternative risk
equivalent discharge programs for the dischargers on the
Willamette River were  compared  with each  other and were
evaluated with  respect  to   the resulting  total  waste
discharge,  water quality, and  total  treatment  cost.  A
nonseasonal  and  a two-season  discharge  schedule was
examined for both  the  UT and  TD  objectives.  The DO
standard was to be  achieved  with an annual risk of water
quality violation  of  10%,  and the alternative  standards
considered varied from 5.0 to 7.5 mg/L.

The  results of this analysis show that, when compared with
nonseasonal  discharge  schedules,  substantial potential
reductions in waste treatment effort and cost were possible
under a  seasonal discharge schedule for  the  Willamette
River. The two-season discharge programs, which allowed
between 0 and 66,500 Ib/day more  waste  discharge than
the nonseasonal  TD programs, resulted in treatment cost
savings of between 0.0 and $22.0 million/yr, (0% to 48%)
with  respect to the nonseasonal discharge schedule.

To evaluate the dependence of cost on the  relative lengths
of the seasons under the two-season discharge schedule,
the length of the critical season (i.e., the season  in which
the average streamflow is the lowest)  was varied. For a
given DO standard and critical  season length,  the total

-------
treatment  costs of both  the  UT  and TD  management
programs  were obtained.  For both of these  discharge
programs, the optimal critical season length ranges from 2
to 5 months, depending on the DO  standard. The total cost
of waste treatment is more sensitive to the critical season
length  for high values  of  the DO  standard than for  low
values. The results for the UT allocation follows the same
trend, but the UT allocation is  generally less  expensive
than the TD allocation.

This research brief  summarizes  work  done  under
Cooperative Agreement No.  CR812577-01 by  the
University of Illinois  under the  sponsorship of the U.S.
Environmental  Protection  Agency.  The  Principal
Investigator was J. Wayland Eheart and  the EPA Project
Officer was Lewis A. Rossman.
References      _  _ _  _    _,_,^,_	

"Eheart, J.W., "Effects  of streamflow variation  on critical
    water quality for  multiple discharges of decaying
    pollutants."  Water  Resources Research,  24(1 ):1-8,
    1988.
                        'Eheart, J. W.  and  H. Park, "Effects of temperature
                            variation on critical stream dissolved oxygen," Water
                            Resources Research, 25(2) :145-151, 1989.
                        'Eheart, J. W.,  E. D.  Brill, Jr., and J. C.  Liebman,
                            "Discharger  grouping  for water quality  control,"  in
                            review, 1989.
                        Kilgore, J. D.,  "Seasonal  static  transferable discharge
                            permits for the control of biochemical oxygen demand
                            in  the Willamette  River," Working  Paper No. 2,  NSF
                            Award PRA 8I-2I692, Department of Civil Engineering,
                            University of Illinois, Urbana, Illinois, 1985.
                        "Lence, B. J., J. W. Eheart, and E. Downey  Brill, Jr., "Risk
                            equivalent seasonal  discharge .programs for  river
                            basins with several dischargers," in review, 1989.
                        'Michels,  C.  M.,  "Incorporating reserve assimilative
                            capacity in  the waste  load allocation process,"  M.S.
                            Thesis,  Environmental  Engineering  and  Science
                            Program, Department of Civil Engineering, University
                            of  Illinois, 1987.
                       _B,ossrnan,_J	A.,_ ILBisk equivalent_seas.Qn.aL waste,  load,
                            allocation,"  accepted by Water Resources Research^
                            1989.
                         •Product of the current project
 United States
 Environmental Protection
 Agency
Center for Environmental Research
Information
Cincinnati OH 45268
 Official Business
 Penalty for Private Use $300

 EPA/600/M-89/014

-------