United States
                    Environmental Protection
                    Agency
 Municipal Environmental Research
 Laboratory
 Cincinnati OH 45268
                   Research and Development
 EPA-600/S2-84-020 Mar. 1984
<>ERA         Project  Summary
                   Procedures  for Estimating
                   Dry  Weather  Sewage  In-Line
                   Pollutant  Deposition
                   Phase  II
                   William C. Pisano and Celso S. Queiroz
                                                                                       .\»
                     Planners,  engineers, and  municipal
                    managers  are  given generalized
                    procedures/equations to estimate the
                    amount of pollutants deposited  in
                    combined sewer systems during dry
                    weather so  they can  make  intelligent
                    decisions  about sewer  flushing
                    programs and other combined sewer
                    management controls.
                     The predictive equations  relate the
                    total  daily  mass of  accumulated
                    pollutants deposited within a collection
                    system to the physical characteristics
                    of collection systems such as per capita
                    waste  rate, service area,  total  pipe
                    length, average  pipe slope, average
                    diameter, and other more complicated
                    parameters that derive from analysis of
                    pipe  slope characteristics. Several
                    other predictive equations that can be
                    used with different available data and
                    user resources are given.  Pollutant
                    parameters include suspended solids,
                    volatile suspended solids, biochemical
                    oxygen  demand,  chemical oxygen
                    demand, total organic nitrogen, and
                    total phosphorous.
                     The equations were developed from
                    data  assembled  from  three major
                    sewage systems in eastern Massachu-
                    setts  and from  a  portion of  the
                    combined sewer system in the eastern
                    district of Cleveland. This study was an
                    extension of earlier work; broader data
                    was used here to prepare the  predictive
                    relationships.
                     This Project Summary was developed
                   by  EPA's  Municipal Environmental
                   Research Laboratory, Cincinnati,  OH,
to  announce  key findings of  the
research  project that  is  fully
documented in a separate report of the
same title (see Project Report ordering
information at back).

Summary of Results
  Results using the augmented data base
are summarized.  Various  predictive
models are described,   relating total
suspended  solids deposition within a
collection   system  with  independent
variables under the assumption of clean
pipe conditions.  These relationships
therefore apply to situations in which the
sewer  piping system is  properly
maintained.

Statistical Summary of
Regression Data
  In Table 1 are found the means and
standard deviations of the independent
variables used in this regression analysis.
  L, A, and D measurements for the
augmented  data base  increased over
those of the prior data base. By including
data from the  relatively flat  Cleveland
collection systems, the  average slope
parameters (§, SPD, and   SPD/4)  all
decreased.
  The average  collection-system
deposition rate computed over all four per
capita waste discharge levels (260, 190,
110, 40 gpcd, respectively) is 1.94
Ib/day/acre of service area. The average
and standard  deviation  of  the rates
computed for a per capita waste rate of
260 gpcd are 1.07 and 1.64 Ib/day/acre
of service area, respectively. The average

-------
Table 1.    Means and Standard Deviations of the Independent Variables Used in the Regression

                               A*                              fit
Variable}.
L(ft)
A (acre)
S(ft/ft)
5 (in.)
SPD (ft/ft)
SpD,4 (ft/ ft)
Mean
13702.
76.
0.0210
11.5
0.0101
0.0037
Standard
Deviation
22867.
102.
00126
2.0
0.0093
0.0033
Mean
14720.
106.9
0.0166
16.9
0.0067
0.0034
Standard
Deviation
20044.
110.8
0.0130
6.1
0.0072
0.0029
*A = the prior data base, i.e., 75 collection systems from the three major sewerage systems in
    eastern Massachusetts;
ffi = all data, the augmented data base, i.e., the previous 75 plus 28 collection systems located
    within the eastern district of Cleveland combined sewer system;
Ji = total length of all pipe in the sewer shed, in ft;
 A = the collection system area, in acres;
 B - average diameter of pipe in the collection system, in in.,
 S = the average collection-system pipe slope, in ft/ft;
 S PD = t he slope corresponding toPLD, in ft/ft (PL D= the percentage of pipe length corresponding to
      where 80 percent of the solids deposit in the collection system);
 Spo/4 = the slope corresponding to PL 0/4 .in ft/ft (PL D,t = one-fourth of the percentage of pipe length
       where 80 percent of the solids deposit).
deposition rates for per capita waste rates
of 190, 110, and 40 gpcd are 1.35, 1.91,
and  3.42 Ib/day/acre of  service area,
respectively.


Alternative Equation  Selections
  Two regression equations are present-
ed and recommended for  user applica-
tion;  the  alternative forms reflect  the
availability of data, or user resources, or
both. The simple form requires few data
and  has  the least predictive reliability;
whereas  the more  elaborate  equation,
requiring greater user resources and data
availability,   provides  estimates  with
extremely high reliability.

The Elaborate Equation
  The highest multiple correlation coeffi-
cient, R = 0.970 (variance explained, R2 =
0.940) was obtained using this equation:
TS = 0.00108
 <;    -0.148 n-0
 SPD/4     1
                          PD
 where:
  TS = deposited solids loading in Ib/day
    q = per capita waste rate, in  gpcd
  The value PLD (the percentage of pipe
 length corresponding to 80% of the loads
 depositing in  the collection  system) is
 derived  from  the  extensive  computer
 analyses.  A  discussion  involving
 estimation of SPD and SPD/4are contained
 in USEPA report numbers EPA-600/2-
77-120 (NTIS No. PB 270 695) and EPA-
600/2-79-133  (NTIS No.  PB 80-118-
524), respectively. The probability of the
pipe slopes can either be derived from
histograms  computed from  local pipe
slope data, or  it can be defined with
reasonable approximation from the mean
pipe slope, S only. If the pipe slopes are
not available, a regression relationship of
mean ground slope and mean pipe slope
can be used to estimate mean pipe slope.

The Simple Equation
  The highest   R2  value  that  can  be
obtained  with   the   least  number  of
independent variables is  given  by the
relatively simple regression equation:

TS = 0.0088 L1-066 (S)-°-433  q-o.539  (R2
      =0.880)

  The exponents of the  independent
variables and the multiplicative constant
of the equation  changed only slightly in
comparison with the equation  derived
from eastern Massachusetts collection
system data. The degree of fit  is high
(R2=0.0880) and superior to that of the
prior fit for eastern Massachusetts data
(R2=0.845).

Estimation of Total Pipe Length
  The total pipe length of the system, L,
and its corresponding collection area. A,
are  generally assumed to be known. In
cases where this  information   is  not
known and where crude  estimates will
suffice, the total pipe length  can be
estimated from the total basin area, using
the expressions:                       A
L = 220.9 A"-84? (R2=0.804) - low popula- \
    tion density (10-20 people/acre)

L = 238.0 A°-8*7 (R2=0.804) - moderate-
    high population density (30-60 people/
    acre)

These equations were developed from a
least squares analysis of the augmented
data base.

Estimation of Pipe  Slope
  If data on pipe slope are not available,
an  approximate  estimate  of  average
collection-system pipe  slope  can  be
obtained by computing a mean value for
the ground slope and then  using this
equation:

     § = 0.320 (§G)0-79°  (R2=0.850,

where:

   SG = mean ground slope, in  ft/ft.

  The above equation  resulted from a
regression  of mean ground and  mean
pipe slope for all 103 collection systems.
The procedure used  in estimating ground
slope data for the  Cleveland collection
systems was similar to the method used *
in the earlier study of eastern Massachu- \
setts collection systems and is included in
the project report.

Discussion of Prior and
New Results
  Table 2  presents an  overview  of the
predictive equations prepared from the
original and the new combined data  set
appended to the original data. The R2 for
the  two elaborate  equations are very
similar. All three collection-system slope
parameters (S, S PD and SPD/4) entered the
regression  equation  at   a  high
significance  level  (Student's t  values
exceeding 4.0) for the combined data set;
whereas, only S PD and SPD/4 entered the
elaborate equation  for  the  original set.
This ifference is due, in part, to the limited
range qf_average collection-system  pipe
slopes,  S, for the three sewerage systems
comprising the  original data  set.  The
range   of   the  average  pipe  slopes
computed  for  the  collection  systems
within  each  of the  three sewerage
systems in eastern Massachusetts was
from 0.0175 to 0.0254 and for  the 28
systems in Cleveland, 0.0028 to 0.0178.
   The inclusion of flatter pipe slope data
from the  Cleveland  sewerage  system
increased  the overall range of average^

-------
collection system variable, S, by an order
of magnitude.  The_ average collection-
system pipe slope, S, for the combined set
of data assumed a more dominant role as
an explanatory variable in describing the
variance of estimated deposition  solids
loadings per collection system. A  visual
scanning  of  the  average  collection-
system pipe slope for the entire data set,
revealed a fairly reasonable spread of
observations  along  the entire  range,
thereby  minimizing  the  concern  of
spurious  correlation  created  by  data
clustering at opposite ends of the range.
The  results for the  simple  equations
shown in Table 2 are similar and confirm
the above discussion in that the R2 is 4.5
percent higher for the combined data set.
  As a side note, the original equations
(derived  from  eastern  Massachusetts
data) were used to estimate daily deposits
for the 28 Cleveland collection systems.
The resulting R2 values for the Cleveland
data, using the equations generated from
eastern Massachusetts  data, are  0.717
and 0.811  for the elaborate and simple
equations, respectively. The  modified
equations (derived for the combined data
set) were then  used  to estimate daily
deposits for the 75 eastern  Massachu-
setts collection systems. The R2 values for
the eastern Massachusetts data,  using
the  equations  generated  from  the
combined data base, are 0.821 and 0.867
for  elaborate   and simple  equations,
respectively.  In  these  numerical
sensitivity  experiments,  note  that  the
degree of fits was superior, in a predictive
sense,  for  the simple equations  in
comparison with those of the elaborate.
  The more favorable R2 results for the
simple  equations  suggest  that  the
elaborate equations are  too specific and
sensitive to changes  in data  input. In
addition, the simple equations require
comparatively little input data compared
with that  needed for the elaborate. The
user needs only prepare estimates of the
total collection-system pipe length, L; the
average collection-system pipe slope, S;
and an estimate of the per capita waste
rate, q, to use the simple equations.
  The  simple  equations are therefore
preferred. Since the exponents  of  the
independent   variables  and  the
multiplicative  constant  of the simple
equation  for  eastern  Massachusetts
differ only slightly  from those of  the
combined (eastern Massachusetts and
Cleveland)  data, the  equation derived
from  the  combined  data, based  on a
broader base, is recommended.
  The full report was submitted in partial
fulfillment of  Grant No. R-804578  by
Table 2.
Equation
Elaborate
Elaborate
Simple
Simple
Comparison of Deposition Predictive Equations
Data Source ft2 Explanatory Variables
E. Mass.
Combined
E. Mass.
Combined
.949
.940
.845
.880
L, Spg, SpD/4,
L, SPQ. SpD/4,
L. S, q
L'S.q
q
S.q
Northeastern  University  and  Environ-
mental Design and Planning, Inc., under
the sponsorship of the U.S. Environment-
al Protection Agency.
   William  C.  Pisano  and Celso S. Queiroz are with Environmental Design &
    Planning, Inc., Boston, MA 02134.
   Richard Field is the EPA Project Officer (see below).
   The complete report, entitled "Procedures for Estimating Dry Weather Sewage
    In-Line Pollutant Deposition:  Phase II," (Order No.  PB 84-141 480; Cost:
    $8.50, 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 1984-759-015/7320

-------
United States                    Center for Environmental Research
Environmental Protection            Information
Agency                        Cincinnati OH 45268
Official Business
Penalty for Private Use $300
                                    PS   0000329
                                    U  S ENVIR PROTECTION  AGENCY
                                    REGION 5  LIBRARY
                                    330 S  DEARBORN  STREET
                                    CHICAGO IL 60604

-------