WATER QUALITY MANAGEMENT GUIDANCE
WPD 3-76-02
LAND USE - WATER QUALITY RELATIONSHIP
MARCH 1976
ENVIRONMENTAL PROTECTION AGENCY
WATER PLANNING DIVISION
WASHINGTON, D.C. 20460
-------
Land Use-Water Quality Relationship
prepared under
Contract No. 68-01-2622
for the
Environmental Protection Agency
-------
EPA REVIEW NOTICE
This report has been reviewed by the Environmental
Protection Agency and approved as satisfying the
terms of the subject contract. Approval does not
signify that the contents necessarily reflect the
views and policies of the Environmental Protection
Agency, nor does mention of trademarks or commercial
products constitute endorsement or recommendation
for use.
-------
Table of Contents
Page
Section i Introduction 1-1
Section 2 Summary and Conclusions 2-1
Section 3 RecommendatiOns 3-1
Section 4 Toward the Analysis of the Land
Use/Environmental Quality
RelationshiP 4-1
A Conceptual Framework 4-].
Criteria for Evaluation of Models
of SubsystemS 4—4
Section 5 Land Use/Water Quality: Physical
Impact 5—1
Subsystems of the Urban Land Use/
Water Quality RelationshiP by
Definition 5—1
A Brief Review of Existing Models 5—4
Estimating Point-Generation
Emissions of Liquid Wastes 5—5
Estimating Areal Generation
and Emissions of Liquid Wastes 5—il
Models for Quantity and Quality
of Urban Runoff 5-20
Other EmissiOns and Interfaces
with Water Quality 5-32
Solid Waste Aspects 5-32
Aspects of On-Site WasteWater
Disposal Sy8tezns 537
Other Aspects 5-40
-------
Table of Contents (Continued)
Page
Models for Transport, Dispersion
and Assimilation 5—40
Models Developed and Selected 5—46
Sanitary Sewer and Wastewater
Treatment Plant Capacity
Evaluation Module 5—49
Linkage of Runoff and Water
Quality Model 5-55
Summary 5—81
Section 6 Land Use/Water Quality: Fiscal
Impact 61
Introduction 6—1
A Brief Review of Existing Models 6-2
Models for Wastewater Treatment
Costs 6-2
Modeling Collection Costs 6-3
Development Level Costs 6-5
Interceptor Level Costs 6—13
Modeling Stormwater Collection
Costs 6—13
Cost Evaluation Module: Models
Developed and Selected 6-16
Introduction 6-16
Sanitary Sewer System Cost
Estimates 6—18
Review of Cost Factors 6-18
-------
Table of Contents (continued)
Page
Sewer Connection Cost
Estimates (CC 3 ) 6-26
Lateral Sewer Costs
Estimates (CC 2 ) 6-28
Stormwater Laterals (CC 5 ) 6-33
Other Cost Functions 6-34
Details of the Cost Evaluation
Module 6—41
Summary 6—45
Section 7 Land Use Air Pollution Relationships 7—1
Introduction 7—1
Air Pollution from Stationary
Sources 7—1
Residential Emission 7—3
Commercial Emission 7—16
Industrial Emission 7—19
Summary 7—20
Appendix A Data Collection for STORM and SWMM A-i
Appendix B Results from Experimental Runs
with STORM B-i
Appendix C Review of Control Options for
Stormwater Management by STORM C-i
Appendix D Lateral Sewer Cost Estimates D-i
Appendix E Glossary E-1
-------
Section 1
Introduction
As more attention is given to the environment and
land use, quantitative descriptions of the most important
relationships between pollution and land use are needed.
The quantification of these relationships is necessary
to support environmental quality planning and management
as, for example, required by the 1972 Amendments to the
Federal Water Pollution Control Act* and the 1970 Clean
Air Act.** Varying degrees of comprehensive planning are
suggested in these laws.***
Federal, state, regional and local environmental and
land use policies, controls, and decisions have impacts on
the physical environment and on local governments’ economic
and fiscal conditions. The impacts have been recognized
in a qualitative manner, but there are few appropriate
tools to quantify the impacts. If land use and emission
controls are to be used effectively in implementing en-
vironmental policies it is necessary to know something
about the dynamics of the local environment in which these
controls are being imposed. Several interacting modeist
are needed which can be used at the local or regional
level to facilitate understanding of these relationships.
Such a set of models would permit the planner to look at
the interactions between land use, socio-econOmic and
fiscal conditions and environmental quality. The approach
* Public Law 92-500, October 1972.
** Public Law 91—604, December 1970.
*** The interdependency of planning processes is docu-
mented, for example, in Public Law 92—500 by Sections 201
(facilities plan), 208 (areawide plan), and 303 (basin
plan).
± We use the word “models” to refer to relationships of
various elements of the physical, economic, social system
that are part of the land use/water qualitY system. We
also use model more comprehensively to mean a model of the
whole system. Thus we will frequently use the word model
to refer to relationships among variables as well as to
refer to an integrated set of relationships among varia-
bles and ultimately to describe an impact model which
integrates a series of subsystem models.
-------
should consider all receiving media (air, water, land) and
take into account the processes which transfer residuals
from one medium to another.*
This project emphasizes urban land use/environmental
quality relationships. While a general framework for the
analysis of these relationships has been developed, and
much literature reviewed, the project has concentrated
on two types of models: (1) physical impact models for
assessing the environmental impact of urban stormwater
runoff, and for evaluating the capacities of sewers and
wastewater treatment plants; and (2) a cost distribution
model for assessing the cost to be covered by different
groups in response to new urban and suburban development.
The schematic in Figure 1-1 outlines the context within
which this project’s relationships have been considered.**
In this report, Section 2 summarizes the findings
of this study, Section 3 gives various recommendations as
results of our findings, Section 4 provides the introduc-
tion to our analysis of land use/environmental quality
relationships, and Sections 5 through 7 describe specific
parts of these relationships and the formulation of the
models. In Section 5 models for describing the path of
waterborne residuals, from generation to assimilation in
the environment, are reviewed and analyzed. There are
two major paths whereby wastes from urban land uses reach
receiving waters: through the discharge of sanitary
sewerage, and through the washoff of dust and dirt (in-
cluding nutrients) from pervious and impervious areas
* Incineration of sludge from a wastewater treatment
plant is an example of such a residual transfer.
** Note: In this scheme, landborne residuals, such as
refuse, are not treated separately, but are related to
either water or airborne residuals, dependent upon
whether their handling affects air or water quality.
Thus costs associated with handling are neglected here.
1—2
-------
INPUT
DATA BASE:
Socio Economic &
emographiC
Land Uses
Development
ChacacteriStiC s
(Residential 1 —
Commercial
Industrial)
Processing
CtmrocteriSti
Infrastructure
Hydrology / Hydraulic
Meteorology
S
.
OUTPUT
r— Emissions
— Ambient Quality
Costs to
— Federal Government
— State Government
— Community
— Household
— Private Enterprises
Cost Incidence
Figure 1—1: Urban Land Use/Environmental Quality Relationships
-
Cost of 1 FinOfl- Incidenc
—Collection Icing oft of
_Treatmeflt C05tS Costs
Generation
] sperSi j
of Airborne [ NOn-StationC Sou eS
L LResiduaIs __________________
-------
(Figure l_l).* The first ty,pe has been analyzed within
the sewer and wastewater treatment plant capacity evalua-
tion model. The second type is analyzed by linking an
existing runoff model and an existing water quality model.
Examples of their uses are presented.
Section 6 describes a model which is developed to
predict the infrastructure cost impact of land uses or
environmental standard changes (Figure 1-1). Additional
development, or higher environmental standards, generally
require an investment of resources to obtain the desired
level of control. Such investments are borne by different
sectors Within the economy, depending U Ofl institutional
and financial arrangements. Three models have been de-
veloped for the analysis of infrastructure cost impacts:
(1) a model that predicts the costs of providing required
infrastructure; (2) a model that maps these costs into a
timestream of required payments, taking into account
financing arrangements, interest rates, and so forth; and
(3) a model that determines the actual costs paid by
different public and private sector groups. The impact
model presented in Section 6 consists of these three
interacting models which together accomplish these three
tasks for sanitary sewage collection and treatment facili-
ties. The cost impacts are divided among three public
sectors (federal, state and local) and the private sector
as a whole.
The third component of Figure 1-1, that of airborne
residuals, is partially treated in this report. Section
7 discusses the generation of residuals from stationary
sources, placing emphasis on air pollutants generated by
residential activities. These activities generate air
* Note: Since the project deals largely with the urban
land use/environmental quality relationship, we do not
talk about point sources versus non-point sources, but
about discharge of sanitary wastewater and about urban
stormwater runoff, referring to the latter also as aereaj.
emissions. Discrimination between point and non-point
sources is currently an unresolved issue due to the recent
suit brought by the Natural Resources Defense Council
( Natural Resources Defense Council, Inc. V. Russell E.
Train and Environmental Protection Agency, et al . , U. S.
District Court for the District of Columbia, Civil Action
No. 1629—73).
1—4
-------
pollutants chiefly through the combustiOn of fuel for
heating purposes. Thus the impacts of different heating
fuel types and insulation practices are examined. No
attempt has been made to deal with non-stationary sources
or with the subsequent transport and dispersion of air
pollutants in the environment.
The linkages among the three components have not
been programmed, thus the complete system conceptualized
in Figure 1-1 has not been achieved during the course of
this study. However, the models developed are compatible
with each other. Thus they can be linked together when
resources are available to develop .a central control
program.* In the water-related system a partial linkage
between subsystems has been accomplished and has been
applied to the Town of Hamden/COflneCtiC ut and the Mill
River System (see Figure 1-2) within the limit of avail-
able data.
* Environmental protection Agency’s Water planning Division
has recently funded the documentation of the models coded
and developed in this project.
1—5
-------
Mile
Figure 1-2
Map of flainden/Connectjcut and the Mill River Basin
H omde n
NORTH
HAVEN
NEW HAVEN
1—6
-------
Section 2
Summary and Conclusions
This study on land use/environmental quality rela-
tionships has developed an approach to facilitate effective
implementation of recent environmental quality legislation.
This legislation requires a quantitative understanding of
the relationships among various environmental and land use
controls, measures of environmental quality, and regional
and local fiscal and economic conditions. For example,
areawide planning such as planning under Section 208 of
Public Law 92—500, calls for an evaluation of environmental
capacity as well as of the cost effectiveness of alterna-
tive physical and land use controls for water quality
management.
In particular, t.t.he study examines the physical
and fiscal aspects of land use and water quality relation-
ships. The physical relationships were developed to des-
cribe the flow path of residuals from points of gener tion
in various land uses to resulting instream qualities J A
literature review of existing models, describing subsystems
of these paths, was conducted and criteria were applied in
order to select appropriate models.t Models were selected
for use in urban runoff and water iality evaluations.
STORM, a continuous model, computes runoff, washoff and
erosion, without computationally burdensome sewer routing.
This model was reprogrammed and linked with the dynamic
receiving water body module of EPA ’s SWMM. The reprogram-
ming required considerable effort due to inadequate docu-
mentation and lack of representative examples, particularly
for STORM. The combination of these two models permits
evaluation of the impact of hydrographs and pollutographs
(suspended solids, settleable solids, BOD 5 , coliforins,
nitrogen, P0 4 ), as they are generated hourly for each sub-
basin, in the water quality module; each rain event has to
be considered by itself in the water quality module.
A model for evaluating sewer and wastewater treat-
ment plant capacity was developed and coded during the
study because the literature review did not reveal any
model appropriate for the purposes of this analysis. The
model permits assessment of existing and-f Liture sanitary
flow, determination of the capacity utilization at every
point in the system, and computation of points of potential
overflow.
Some existing cost functions were incorporated in the
cost model; others were newly developed. In particular,
-------
cost functions of sanitary laterals at the residential
development level and their house connections were estima-
ted, based on layouts of development as well as on actual
sewer system design. The literature review had shown that
existing functions either neglected a realistic layout or
an actual engineering design. The cost model permits
evaluation of costs incurred by new development or redevel-
opment and the distribution to different groups resulting
from the cost-sharing and financing mechanisms employed
by federal, state and local governments.
These models were coded and organized so that planners
or other users can use them together or individually to
predict the impacts of selected land development patterns
and environmental control strategies.
Literature on air pollution was reviewed with the
objective of developing impact models similar to those
for water impacts. The review Concentrated on stationary
sources. Methods and estimates of emissions were compiled
for residential, commercial, and industrial developments.
Lack of resources precluded the development of a computer-
ized evaluation model for emissions from stationary sources
and of costs associated with their control. However, this
section summarizes the state of the art on predicting
air—borne emissions.
The potential of the model combination of STORM and
SWMM to evaluate urban runoff has been illustrated in the
case study of Hamden, Connecticut (Mill River Basin). The
drawback of the area was the lack of data for detailed
calibration -- a common problem for planners. Various
conditions of development were selected for one of the
sub-basins. Initially, STORM was used to generate emission
rates, dependent upon dry antecedent and meteorological
conditions. Changes of the emission rates clearly ref lec—
ted different degrees of development as well as of other
factors, such as different Street cleaning strategies.
For some of the development Conditions the impacts of the
resulting loads on the receiving stream were computed by
SWNM. The influences of new development in only one sub-
basin, as well as in all sub-basins (for example, according
to the 1985 land use plan), were recognized by comparison
to results for the base case, the land use configuration
of 1974. The impacts were evaluated for different rain
intervals over the year and for various base flow condi-
tions in the Mu]. River. For these different assumptions
the resulting impacts differed by orders of magnitude.
While no genera1izatjon of our results are appropriate
2—2
-------
because every river system behaves uniquely, the exercise
did demonstrate the ability of the models to analyze these
systems. These results have led us to conclude that a
modeling package such as this could be used by various
planners concerned with the impact of land use changes
and/or environmental control strategies. We anticipate
that these models could be useful in three situations:
1. regional planning groups formulating or modifying
a regional plan, to assess environmental impacts and en-
vironjuental control-related fiscal effects of alternative
regional development patterns; planning agencies, such as
those operating under Section 208 of Public Law 92-500,
are a typical example of this category of potential users.
2. town (or local) planning groups formulating or
modifying a local plan and local government policy in-
struznents, to assess water quality impacts and environ-
mental control-related fiscal effects of alternative local
development patterns; and
3. town planners assessing the relative water quali-
ty impacts of, and the long-range water pollution control
capacity investments required by a proposed development
or series of developments.
We chose examples on the level of the town, but the
models can be applied in the same fashion to regional
analysis, such as required in 208-studies. When a town
is faced with a proposed subdivision or commercial!
industrial development, decisions of town authorities
as to whether or not to permit the development and as
to the types of fiscal and environmental requirements
that should be made of the developer (such as payment of
sewerage system expansion or provision of a certain mini-
mum level of heating insulation) depend upon the types of
impact information derived from this approach. At the
town level, it is possible to perform the fiscal analysis
without considering the finances of other communities
(except when a regional authority operates wastewater
treatment works). In some instances, this is also true
of water quality impacts although normally a situation
such as in our example, will exist: localized effects of
local land use changes on the ‘ocal receiving stream can
be assessed, but downstream (regional) effects in the next
higher order stream require nderstaflding of changes in
other upstream communities. The models could be coupled
in different ways and then be used in regional planning.
For most air pollution impacts, consideration of the full
2—3
-------
airshed is a necessity, with one town’s air pollutant
output being an input to a regional analysis. No attempt
was made to devise models appropriate for these problems.
The results of this study point to a number of con-
clusions. First, the definition of a land use classifica-
tion scheme and the creation of well—defined mechanisms
for estimating generation of the residuals associated
with the respective land use activities are unresolved
issues. Such a scheme should, of course, facilitate the
planning of pollution control, which includes controlling
the mix, area/location and temporal distribution of the
residual generating activities themselves. Most land use
(information) schemes are not sensitive to the issues
appropriate for planning environmental quality management.
An appropriate classification scheme should allow for esti-
mating residual coefficients of different degrees of re-
finement for a specific area. That is, it should be
possible to realize a classification scheme that would
allow the use of national or regional emission coeff i-
cients. Alternatively, area specific estimates of residual
generation mechanisms can be performed.
Second, some tools exist which allow for quantifica-
tion of some of the environmental, fiscal, and socio-
economic impacts of land use and environmental control,
but there are no operational computer programs available,
appropriate for the task, nor are most of the modeling
formulations (presented in the literature) organized in
such a way that they can be coupled into a package for
interactive use by the analyst.
Third, inadequate documentation of existing, publicly
available models and lack of examples representative of
problems in land use/environmental quality analyses appar-
ently tend to discourage potential users who are not well
trained in modeling -- this category includes the majority
of staff in public agencies who might apply these models.
Fourth, the analysis of a small sized watershed illus-
trated quantitatively the influence of additional develop-
ment on water quality during storm events. Generally well
documented qualitative knowledge that development affects
water quality, was quantified within the range of accuracy
required by a planning department for their impact assess-
ments.
2—4
-------
Section 3
Recommendations
The following recommendations point to areas where
more research and better quantitative data would be useful
as well as where more resources should be made available
to improve the use of existing models and concepts.
1. Land use classification scheme. An appropriate
set of land uses and their activities would be useful for
evaluation of land use and environmental controls. This
would facilitate analysis of mix, area/location and tem-
poral distribution of residual generating activities which
are used as inputs to evaluate the impact of urban pollu-
tion.
2. Dust and dirt accumulation rates and their com-
position. The dust and dirt accumulated on impervious
areas represents the main contribution to the polluting
effects of urban runoffs. No adequate data exists on the
accumulation as a function of traffic, surrounding land
use, air pollutant fallout, solid waste spillover, etc.
Right now, planners have to rely on accumulation rates per
type of land use, while rates functionally related to both
individual activities and land use type are more desirable.
3. Non—urban land use. In areas of significant non-
urban land use activities, these non-urban land uses
should be dealt with adequately by subdividing them into
an appropriate number of non-urban land uses, because each
one is made up of significantly different activities and
pollutant sources. Existing computer models such as STORM,
should be modified to the extent that prevailing knowledge
permits, to accommodate this greater degree of detail.
4. Spatial aggregation of large sub-basins. The
model STORM can be used for generating urban and non-urban
runoff and urban washoff from areas smaller than ten square
miles, as suggested by its developers. Due to this limi-
tation, STORM should be recoded so that multiple runs can
be executed for one sub-basin without water quality rout-
ing mechanisms. This requires that the model should ac-
cept simultaneously rain hyetographs as well as time
offset upstream hydrograPhs. The recoding would permit
the application of the framework developed in this study
to large regional areas without exhaustive emphasis on
water quantity and quality routing in the receiving water
bodies. It would also permit a stepwiSe aggregation from
-------
local to regional analysis. The main basin for the local
analysis would become a sub-basin of the basin of regional
interest.
5. Functional relationships. New relationships have
to be developed for determining temperature of runoff
(mainly in the summer) as a function of antecedent condi-
tions, and salt content of runoff as a function of salt
spreading strategies of communities, salt accumulation
behavior, etc.
6. Aspects of groundwater pollution. When a model
becomes available which permits evaluation of pollutant
behavior in groundwater, it should be added to the package.
Impact of on-site sewage disposal (such as septic tanks)
and of solid waste disposal sites on groundwater quality
could then be included in the analysis.
7. The extension of computer packages. A number of
submodules, covering aspects such as air and associated
control costs, could be added to the current programs of
models on the basis of existing knowledge.
8. Development of a master program. Currently the
package of submodules is computationally only loosely
interrelated; only STORM and SWMM are directly linked
(though the linkage should be computationally improved to
become more flexible). It is necessary to design a master
(control) program which permits the user to operate and
coordinate the models in an efficient and flexible way.
9. Cost functions, cost distribution, and equity
aspects. Better cost functions have to be developed to
estimate costs incurred by necessary construction of addi-
tional infrastructure. Costs of structures associated with
the large number of potential stormwater management and
relief strategies are a prime example. Impact of financing
the infrastructure costs on various socio-economic classes
and the resulting equity considerations have also not been
treated adequately and need further investigation.
10. Information banking and technical assistance to
environmental planning agencies. This study has suffered
somewhat from the inadequate documentation of publicly
available computer programs and the inadequate presenta-
tion of problems related to execution of programs under
specific conditions. EPA’S user assistance program (such
as for SWMM) should be extended to provide agencies with
advice and information on all relevant models. Only with
3—2
-------
such assistance in addition to documentation and transfer
of experience, will the available analyses and computer
packages be as useful as the sponsoring agencies had
originally anticipated.
11. Application. It is recommended that at least
two 208 agencies with very different problems be selected
as target areas to test the approach developed, to augment
and refine the computer package if necessary, and to
suggest additional improvements of the analyses and their
performance.
3—3
-------
Section 4
Toward the Analysis of the
Land Use/Environmental Quality Relationship
A Conceptual Framework
Our approach is based on the development of classes of
relationships. A list of the independent and dependent
variables is presented in Table 4-1. A class of independent
variables titled is suggested for the social, economic
and demographic characteristics of a community or region.
Such variables include population, population rate changes
by cohort distribution, and property values by types of
communities; each type may be further disaggregated into
categories such as residential, commercial and industrial.
The second class of independent variables, Class A,
refers to land use types. In this category traditional
physical/geomorphOlOgiCal variables such as soil type, slope
and cover are mixed with a set of land use variables which
characterize the use to which the land is put. Examples of
the latter are residential, commercial and industrial land
uses.
The third class of independent variables, Class B,
describes development and infrastructure characteristics.
These variables indicate, for example, the extent to which
a community is sewered or unsewered, the type of sewerage
system, the nature of the road network and drainage systems,
the type of utilities available, etc.
The dependent variables are divided into two classes.
The first class, Class C, includes measures of the costs of
land use development and the distribution of the costs
among communities and socio-eCoflomiC groups (which we refer
to as the cost incidence). Some of the cost measures
suggested are capital and operation and maintenance (Q&M)
costs expressed on a per capita and per acre basis. There
is the further potential breakdown by income and measures
of tax rate changes.
The second class of dependent variables, Class D, con-
sists of variables which represent environmental quality
indicators. These variables are divided into two sub-
classes, emission factors (D 1 ) and quality indicators (D 2 ).
Table 4-1 indicates a list of potential emission factors
measured by media on a per acre basis and a series of
indicators characterizing the environmental quality in time
and space.
4—1
-------
Table 4-’1
Summary of Possible Independent and Dependent Variables
Soc jo—economic
and
Demographic
Land
Use
Type
A
Cohort
distribution
Community
profile
- residential
— commercial
- industrial
Single family
— 1—3 lot sizes
— 3 soil types
— 2 depths of
bedrock
— 2 groundwater
conditions
- slope
Multi—family
- 4 densities
- soil type
— groundwater
- slope
K* */capjta
public
federal
state
local
K/capita
private
K/acre
O&M* * *
(See K)
Imp ii cit
tax rate
Emission by Media
(pollutant/space unit!
BOD time unit)
SS
N
P
ccli form
So 2
NO
particulate s
etc.
ambient Indicators
by Media (pollutant-
unit or concentra-
tion/time-unit
color
turbidity
coliforms
Independent Variables
Development
Characteristics
and
Infrastructure
Dependent Variables
Cost
and
Incidence*
B C
Emission Factors
Quality Indicators D 2
D
Un sewe red
Separated sewers
Combined sewers
Storage
- on line
- off line
Utilities
— heating systems
oil
gas
electric
* Distribution of costs among communities and socio-
economic classes.
** K is capital costs.
O&M is operation and maintenance costs.
-------
Table 4-1 (continued)
Socio—economic
and
Demographic
I)
Land
Use
Type
A
Couunercial
- strip
2 densities
- shopping
center
2 scales
— soil
- ground-
water
— slope
Industrial
Agriculture
Woodland
Development
Characteristics
and
Infrastructure
B
Water
- wells
- public
• surface
• groundwater
Road
Percent impervious
cover
Water using
appliances
- irrigation of
lawn
and
Incidence*
C
Time
streams
Emissions factors D 1
Quality Indicators D 2
D
Ambient Indicators by
Media (Continued)
± algae
DO, BOD
SO 2
particulates
CO
NO
etc.
Solid waste
collection
Cost
- kitchen
Open Space
- type
- quality
— amount
Meadow
-------
Relationships among the classes of variables can be
characterized by three categories. The first expresses the
cost and incidence variables (Class C) as a function of the
variable classes ‘1’, A and B: (C=f( , A, B)]. Thus we are
suggesting that cost, fiscal impact and incidence variables
are a function of all three classes of independent variables.
The second category of relationships links emission factors
and independent variable Classes A, ‘I’, and B; that is,
emission factors can be expressed as a function of land use
type variables, socio—economic variables and variables
describing the development and its infrastructure character-
istic: [ D 1 =f(A, B, [ 1)] The third class of relationships
relates quality indicators to the independent variables:
[ D 2 =g(A, B, ‘1’)]. A subset of these relationships relates
ambient quality indicators explicitly to emission factors
(and hence implicitly to variables of Class A, B, and )
using an additional class of functional relationships T,
which we shall call transfer functions. The transfer
function describes the transport, dispersion and assimila-
tion of the emissions. Thus we can calculate ambient
quality in time and space [ D 2 =f(D 1 , T)].*
In order to analyze residuals and economic impacts in
the overall system characterized by the relationships
postulated among the variable classes, the physical and
economic subsystems have to be identified which when linked
together contain the above relationships. Relevant sub-
systems were investigated on both the water and air side;
but the water side was emphasized in the review and analysis
of models.
Criteria for Evaluation of Models of Subsystems
A large number of models describing various aspects of
the land use/environmental quality relationship can be found
in the literature. There exists** for various subsystems
* In the cases where a specific land use, such as wet-
lands, explicitly affects the transport dispersion and
assimilation, then the function becomes (D 2 =f(D 1 , A, T)].
** For example, on the water side see E. A. McBean and
D. P. Loucks, “Planning and Analysis of Metropolitan Water
Resources System,” Technical Report No. 84,. Cornell Univer-
sity Water Resources and Marine Sciences Center, Ithaca,
New York, June 1974.
4—4
-------
a number of summaries that cover much of the material on the
subject written in the last decade. The complexity of the
models, the necessary data requirements and the resources
needed to solve the models vary widely. Our literature
review is aimed toward determining what earlier efforts
would be useful in synthesizing a chain of models which
adequately describes the land use/environmental quality
relationship. We have thus evaluated each model in terms of
its potential within a nested model.*
We have established a number of criteria for model
evaluation; a thorough evaluation should:
- provide sufficient justification for the use of
a particular model (i.e., quantitative reflection
of the mechanism under investigation) within the
spatial and temporal context of the study;**
- consider the details of input/output and of internal
relationships of each model within the total frame-
work;
- ensure the homology of the directly connected models,
i.e., the output of a model must be structurally
equivalent to the inputs of the models with which
it integrates;
- estimate the sensitivity of the total framework to
a particular model; a high sensitivity would require
* We speak of nested models when a set of models is
nested in such a fashion that outputs of one model are
inputs of other models (see Meta Systems mc,, “Systems
Analysis in Water Resource Planning,” prepared for National
Water Commission, July, 1971; published by Water Information
Center, Inc., Port Washington, N.Y., 1975).
** Generally many assumptions and si.mplificatiofls are made
which frequently prohibit judging a model adequate a priori
for other purposes than the one which caused its development.
It is worth noting the argument presented by various
operations research investigators (e.g., Hillier and Lieber-
man, 1967) that the appropriate criterion for judging
validity of a model is whether. within the scope of the
problem the model predicts the relative effects of alterna-
tive courses of action with suf? I itacCuraCy to permit
adequate analysis of that action and decision on it; see
Hillier F. and G. Lieberman, Introduction to Operations
Research , Holden Day, Inc., San Francisco, l96 .
4—5
-------
a model which is well detailed and reliable, but
would also imply that if such a refined model were
not available the connected models at the next
levels might be less rigorous;
- consider the temporal and spatial resolution of the
model in the context of the temporal and spatial
framework of the total analysis;
- determine the availability of data to calibrate and,
if possible, to verify the model; a complicated
model can be used to its full potential only when
an adequate data base is secured. Otherwise a less
complicated model with lesser data requirements
might produce superior results. It is possible to
fit every model to a limited data base; but the
quality of the model’s performance decreases with
the number of parameters which must be set arbitarily;
— estimate the computational resources needed to apply
the model, in the light of its proposed results and
of its position in the macromodel;
- evaluate if the type of model -- both the approach
and its generated output (in time and space) -- is
useful for the planning methodology. It is often
desirable to use a model which has already been
prepared to reduce the costs of coding; but such a
model might be irrelevant to a planner or to a
designer or both.
We have tried to apply these criteria carefully in the
review of the literature pertinent to the modeling of the
land use/environmental quality interrelations and in our
effort toward the synthesis of a land use/environmental
quality macromodel. In the following sections of the report
the particular relationship being focused upon is defined,
a review of pertinent analytical factors and existing models
or modeling approaches is made, and then features to handle
the relationship are synthesized.
4—6
-------
Section 5
Land Use/Water Quality: Physical Impact
Subsystems of the Urban Land Use/Water
Quality Relationship by Definition
In order to develop a chain of mathematical models
reflecting our framework and in order to review pertinent
literature, a scheme was devised to.display the connections
between types of land use and water quality via relevant
activities and events (Figure 5-1). Specific land use types
(commercial, industrial, and residential) imply aspects of
the classes of independent variables, i , A, and B (see
Table 4-1): density of residences, population, and comzner-
cia]. and industrial activity centers; cohort distribution
of residences, type of activities; percent open space; and
percent imperviousness and infrastructures such as roads,
utilities, etc. Functions inherent in land use types (such
as living, manufacturing, marketing, expanding, etc.) imply
use of resources and their subsequent transfer into a
different state. These functions are reflected in water
consumption, in generation of residuals (SS, BOD, P , N,
etc.) in traffic, in activities (such as loading of trucks,
repairs, recreation, etc.) on impervious and pervious areas,
and in construction of activity centers. The next level of
the chart is comprised of the emission of wastewater and of
solid wastes as well as the accumulation of pollutants on
impervious and pervious areas and the erosion potential.
The emissions are strongly related to the land use types
(see above). Accumulation of dust and dirt (being composed
of pollutants such as dissolved, suspended, and settleable
solids, BOD, N, P0 4 , coliform, etc.) is caused by traffic,
air pollution fallout, and other incidents as well as by
construction activities. The erosion potential is dependent
not only on temporary construction events but also on the
physiography and topography of an area. The next level in
the chart refers to collection and transport systems from
the point of generation (or point of contact in the case
of rainfall) to the point of discharge into the aquatic
environment, or littoral environment in cases of where
wetlands are the receiving body. Wastewater can be removed
from the point of generation by public collection systems
(separate or combined sewers) or by private on—site disposal
systems if sewers are not available (see also infrastructure
variable Class B, Table 4-1). Solid wastes will be either
collected by collection vehicles (private contractor,
publicly contracted collector, or public collection) or. will
be taken to the handling facilities by each individual
5—1
-------
PoHution by Pollut nti and Temperature, Stream
Enlargement: Erosion
Figure 5-1: Flow Chart of Connections Between Urban Land Use Types and Water Quality
-------
household.* It should be noted that spills also contribute
to the accumulation of pollutants. Stormwater and the
pollutants which it carries will be removed through drainage
facilities (combined sewers or stormwater conduits, includ-
ing detention facilities) or will run off directly to a
receiving water body.
The next level is concerned with treatment and then
discharges of the residuals of all activities into the
aquatic and/or littoral environment. It should be noted
that all discharges are divided into “treatment and dis-
charge” and “untreated discharge” such that the category
“treatment and discharge” covers all those discharges which
are currently subjected to permit application under the
National Pollution Discharge Elimination System (NPDES).
We are concerned with the quantity as well as the
quality of each discharge. Quality can be controlled by
treatment plants and other means, such as street cleaning
or sewer flushing. The amount of pollution entering the
receiving water body depends on the portion of untreated
to treated discharges. All discharges from the sanitary
sewers of a separate sewer system are categorized as
treated discharges. The quality of the latter depends on
the degree of treatment in the treatment plant. If the
combined sewer system works properly during dry weather
flow, the sanitary flow in combined sewer systems also falls
into this category. If the overflow devices work improperly,
the malfunctioning has to be recognized and recorded in
order to calculate the sanitary flow which is discharged
without treatment. The stormwater flow in combined systems
will be discharged largely untreated. The quality of
overflowing and bypassing waters of the combined flows has
to be adequately adjusted due to mixing with sanitary flows.
Discharges from storm drainages in a separate sewer system
are largely untreated.* The quality of these discharges
depends mainly on the washoff rate of accumulated dust and
dirt from urban areas and on the deposi s in the system.
Generally all unknown, unmeasured, or uncontrolled flows
from combined or separate storm drainages as well as other
systems will be dealt with as untreated discharges to
surface and groundwater. This includes stormwater which
may enter the receiving water body by seepage from conduit
* Note that some treatment facilities for stormwater
have been built in the last decades (see, eor example)
Lager, J. A., Stormwater Treatment: Four Case Histories,
Civil Engineering-ASCE , December 1974).
5—3
-------
system, overland flow, gullies, and minor tributaries.
The impact of private wastewater collection systems is
of interest only in terms of quality. Infiltration from
poorly operating on-site wastewater disposal systems affects
the groundwater quality; runoff from inadequately operated
leaching fields might affect the quality of surface and
groundwater (which in turn might further Lflfluence surface
water quality). The effects of solid waste disposal areas
are also limited to quality aspects. There exists evidence
that infiltration of rain water into solid Waste landfills
(or any other land disposal area) produces leachate, which
in the case of inadequate subsoil greatly affects the ground-
water quality.
The total quantity as well as the peak discharge of
urban stormwater runoff can have significant effects. The
fast runoffs from urban areas have a high Potential to cause
flooding in urban conduits as well as in receiving streams.
This can lead to stream enlargement by erosion of the banks
and to other channel effects. Depending On the physio-
graphic and topographic characteristics as Well as on man’s
land usages and construction activities, rainfall and sub-
sequent runoff might cause heavy erosion, Which, in turn,
could lead to sedimentation in artificial and natural
conduits. A high degree of imperviousness over large areas
contributes to the reduction of the base flow in streams,
mainly in dry seasons. Thus heavy runoff during these
seasons might influence the temperature in the stream.
In this framework all discharges will be ultimately
transferred into temporally and spatially distributed
determinants of water quality of ground and surfacewater.
The transfer depends on the characteristic of the receiving
water body and the emission components.
A Brief Review of Exisjing Models
The network of the land use/water quality relationship
typified in Figure 5-1 will be used as a basis for discuss-
ing mathematical models found relevant along the lines of
the established criteria. Estimates of, and approaches to
estimating emissions are considered first for discharges
consisting primarily of sanitary wastes generated at a point
of activity (point-generation), then for discharges largely
caused by precipitation on urban areas (areal generation).
The latter includes both surface accumulations of pollutants
and their washoff. Next, some existing approaches to
modeling urban runoff are reviewed. Models for other
5—4
-------
emissions which interface with water quality including
solid waste and on—site wastewater disposal follow, and
finally, models for transport/dispersion and assimilation.
are examined.
Estimating Point—Generation E nissions of Liquid Wastes
Unfortunately most of the emission data collected on
water use and wastewater characteristics are highly aggre-
gated. With water use, for example, data are generally
available on total community usage, while for wastewater,
most flow and quality characteristics are measured at the
“end of the pipe” and therefore represent the effects of
a combination of land uses. As a result, engineers who
develop design factors to apply to specific land use types,
whether for a particular project or for general reference,
have to rely to a considerable degree on their judgment and
experience in disaggregating the available information.
Over the years this process has evolved a set of more or
less generally accepted “rules of thumb” for use in waste-
water management planning. The quantities 0.17 lb. BOD 5 /
capita/day, 0.20 lb. SS/capita/day, and 100 gpcd appear
extensively in the wastewater engineering literature.*
* Investigators increasingly are turning to different
methodologies in attempting to develop reasonably accurate
emissions factors. Recently Meta Systems Inc has formulated
a framework for the evaluation of generation factors and has
shown how recent studies fit within that framework. Mechan-
istic and econometric approaches, individually or combined,
were suggested for determining the residuals. The econo-
metric approach seeks to avoid the detailed mapping of
consumer selections into residuals by neglecting the choice
of technology and its factor inputs: It assumes a fixed
technology and attempts to map directly from certain key
factors in the consumer’s decision, particularly income, to
residuals. Thus the approach follows in the spirit of much
of the traditional econometric literature on consumer
demands. The mechanistic approach, on the other hand,
ignores the consumer’s behavior and socio-economic basis
and concentrates on obtaining the appropriate technology
coefficients. This approach, for instance, should allow the
planner to estimate the desired residual coefficients from
a detailed analysis of the appliances and their uses in each
water-consuming and waterborne generating part of the house-
hold and the subsequent summation of all usages.
In both the econometric and mechanistic approaches
5—5
-------
Residential wastewater generation is Correlated to
water use, so that water use estimates can be applied for
estimating wastewater quantities. The average per capita
water use in the U.S. was estimated to be 147 gallons per
capita per day (gpcd) in 1954.** With the assumption that
40 percent of this represented domestic usage, it was
concluded that the average residential use was about 60
gpcd. Assuming that 60 to 80 percent of this use is non-
consumptive,*** we arrive at an estimated range of 36 to 48
gpcd for wastewater generation. This is considerably less
than the 100 gpcd figure commonly cited but is in general
agreement with values cited, for example, by Ligman et al .
There exist a number of models, aimed predicting
residential water consumption. Linaweaver,VT for instance,
used an econometric approach for his analysis. He found
that the principal factor influencing total annual water use
is the total number of homes. Three additional factors of
significance are the economic level of the consumer, climate,
* (footnote continued from previous page)
different levels of aggregation may be employed, depending
upon the purpose of the analysis, the data, and time avail-
able for the study. See for details R.J. deLucia, J.
Kuhner, and M. Shapiro, “Models for Land Use/Water Quality:
Some Observations on What Exists and What is Needed,”
presented at the 47th National ORSA/TIMS Meeting, Chicago,
April 30, 1975.
** Select Committee on National Water Resources, U.S.
Senate, Water Resources Activities in the United States,
Government Printing Office, Washington, D.C., 1960.
*** Metcalf and Eddy, Inc., Wastewater Engineering , New
York, McGraw-Hill, 1972.
K. Ligman, N. Hutzler, and W.C. Boyle, “Household
Wastewater Characterization,” J. Environmental Engineering
Div. , ASCE, vol. 100, EEl, February, 1974, pp. 201—215;
Ligman, et al. , tried to develop household quantity and
quality emission data by combining an empirical study of
the biochemical composition and usage patterns of common
household appliances with additional flow and composition
data available in the literature.
tt F. Linaweaver, “A Study of Residential Water Use,”
report prepared for F14A, Departm nt of Housing and Urban
Development, Washington, D.C., 1967.
5—6
-------
and charge mechanisms Linaweaver tested his hypothesis
regarding the influence of the “economic level of the con-
suraer” on water usage by analyzing water use data from 29
areas and by fitting a regression equation which contains
average domestic use per dwelling unit (gpd) as the depen-
dent and average assessed valuation of property (thousands
of dollars) as the independent variable. Additional
analysis of residential water demand by Howe and LinaweaVer*
resulted in an equation containing age of dwelling unit,
number of persons per dwelling unit, and average water
pressure in addition to the initial independent variables.
Based on extensive analysis of water use data from 21
eastern and western drainage basins, this study indicated
that only the market value of the dwelling unit and the
charge mechanism (metered versus flat rate) were statistically
significant in formulating regressions with average daily
water use as the dependent variable.
Determining wastewater generation from particular
industrial or commercial developments is more difficult than
from residential areas because types of non-residential
activities influence significantly the amount of wastewater
and emissions. However, land use and sewer planners who
must base their planning on projections 20 years or more
into the future are seldom in a position to predict the
types of activities which will locate in a particular
industrial or commercial zone, even less so the intensity
of the ultimate use. At this level planners are forced to
work with emission factors on a per acre basis. Not sur-
prisingly, the result is that projections vary widely
across the, country, as indicated in Table 5—1.
It seems probable that commercial development will
continue to be of either the strip commercial or shopping
center variety for most suburbanizing areas, and industry
will continue to select land-intensive production combifla
tions. Based upon design values employed in areas where
these types of activities are prevalent, we have found that
commercial and light industrial wastewater flows can be
expected to range from 2,000 to 10,000 gal/acre/day.
Estimates of per capita pollutant generation have been
developed mainly from combining estimated flow rates with
* C.W. Howe and F.P. Linaweaver, “The Impact of Price
on Residential Water Demand and Its Relation to System
Design and Price Structure,” Water Resources Research ,
vol. 3, no. 1, 1967.
5—7
-------
Table 5-1
Wastewater Generation of Non-Residential Land Uses
Location
Commercial
( gpd/acre )
Industrial.
( gpd/acre )
(unreported) *
Cincinnati, Ohio* *
Los Angeles, California**
Kansas City, Missouri**
I Tennessee**
Buffalo, New York**
Santa Monica, California**
15—60,0 00
40,000
11,700
5,000
2,000
50,000
9,700
* Source: Sigurd Grava, Urban Planning Aspects of Water
Pollution Control , New York, Columbia University Press, 1969.
** Source: Design and Construction of Sanitary and Storm
Sewers , American Society of Civil Engineers and the Water
Pollution Control Federation, 1969.
14,000
15,500
1o,000
2,000
13,600
5—8
-------
available data on wastewater composition. Using this
approach, we have obtained the table of emission factors
(Table 5-2), which is calculated by multiplying the concen-
tration for a medium-strength sewage (Metcalf and Eddy*)
by estimates of 50, 100, and 150 gpcd to obtain a reasonable
range of emission rates.
For comparison, the results of Ligman et al.,** are
presented alongside the computed values. It can be seen
that the Ligman study, which estimates loadings from
individual water uses within the house, agrees well with
the medium estimated value.
Emission factors for residential wastewater generation
have been reported not only,on a per capita, but also on a
per acre*** and per bedroom basis, From the standpoint
of land use planning applications, one of the latter types
of coefficients, particularly the per acre basis, might seem
more attractive. However, such figures have assumed certain
values for number of persons per dwelling unit and dwelling
units per acre and these assumptions often are not stated
explicitly. As a result, use of these coefficients can lead
to significant inaccuracies. It is more appropriate to
start with a per capita flow and’ then build up acreage
values based on explicit assumptions as to dwelling unit
composition and density. For example, Gravatt list.s
8,000 gal/acre/day as a figure for medium—cost residential
housing. At 100 gpcd and 3.5 persons/dwelling unit, this
implies a density of 23 units/acre, over five times the
figure for typical medium—cost single—family homes in
suburban settings.
For most light industrial and commercial land uses,
pollutants are, derived primarily from the sanitary wastes
of employees and shoppers. As a result, assumptions regard-
ing residential waste strengths can be applied to those
wastes as well. For certain commercial establishments such
* Metcalf and Eddy, Inc., op. cit .
** Ligman, et al., op. cit .
Grava, Sigurd, Urban Planning Aspects of Water pollu-
tion Control , New York, Columbia University Press, 1969 .
t Salvato, Joseph A., ., Environmental Engineering and
Sanitation , New York, John Wiley & Sons, Inc., 1972.
tt Grava, S., op. cit .
5—9
-------
Table 5-2
Residential Emission Factors for Pollutants
emission (lb/capita/day)
Ligman
Pollutant low medium high (et al.)*
Total Solids .292 .584 .876 .505
Suspended Solids .083 .167 .25 .198
BOD 5 .083 .167 .25 .174
Chemical Oxygen
Demand .209 .417 .625
Total Nitrogen .017 .033 .050
Total Phosphorous .004 .008 .013 .009
* Ligman, et al. , op. cit .
5—10
-------
as restaurants and laundries and for most heavy industries,
this generalization is not valid, In the case of commercial
land uses, the effect of a few individual establishments is
likely to be averaged - which means lost in the large
number of economical uses associated with any reasonably
sized planning area. For heavy industries virtually no
useful generalizations can be made. Salvato* presents an
extensive table of BOD and SS loadings per unit of product
for major polluting industries. Again, however, unless the
planner has some reasons to believe that a particular
industry will be likely to locate in a specific area, these
tables are of little use.
Assuming continuation of the present pattern of suburban
commercial and light industrial development, emission ranges
for these land uses can be estimated from the flow range
cited previously and the medium—strength sewage concentrations
employed for residential emissions (Table 5—3).
Estimating Areal Generation and EmissionS
of Liquid Wastes
Quantitative knowledge about areal pollution is still
at a very crude stage. Available data are of a very aggre-
gated nature and at best can give order-of-magnitude esti-
mates of potential loadings from different land uses. In
general, these sources can be divided into two categories
-- those associated with stable land uses and those associ-
ated with transient activities. Included in the former
category is runoff from the continuous use of land for some
particular urban or non—urban activity, such as forestry,
pasture, farming, and residential dwellings after land-
scaping is completed and new drainage patterns established.**
The range of constituents emitted from these sources, par-
ticularly urban land and active agricultural land, covers
virtually the entire range of identif led pollutants, from
oxygen demanding organics to heavy metals and complex
organic chemicals.
From a long-range standpoint pollution from stable land
uses are of most concern to planners. In the short term
however, transient construction or silviculture activities
may play a much larger role in determining water quality.
* Salvato, op. cit .
** Note, areal sources based on non-urban activities are
called “Non-Point Sources”.
5—11
-------
Table 5-3
Emission Factors: Commercial and Light Industry
lb/acre/day
Total Solids 11.7 — 58.4
SS 3.3 — 16.7
BOD 3.3 — 16.7
COD 8.3 — 41.7
N 0.67 — 3.3
P 0.17 — 0.83
5—12
-------
The problem of erosion and sedimentation due to construction
activities is well known. However, a variety of other
pollutants are also generated, including fertilizers,
synthetic organics, metals and a variety of inorganic salts
used for soil stabilization. Many of these are transported
by sediment movement after adsorption onto the fine soil
particles. Measures to control erosion such as early
planting of grass at construction sites can lead to other
types of pollution, notably runoff of fertilizers employed
to speed growth. -
In recent years attempts have been made to improve the
quantitative knowledge about the magnitude of areal source
pollution by investigating the activities of land use types
in detail. This has been attempted with stable urban land
use types by concentrating on deriving accumulation rates
of dust and dirt on surface areas (mainly impervious areas)
and then by computing washoff load and quality of runoff
based on these rates and the magnitude of precipitation.
These attempts are described before some of the aggregated
emission data are presented and discussed.
1. Surface Accumulation of Pollutants . Data show that
contaminants from impervious surfaces are by far the pre-
dominant pollutants of urban stormwater runoff. Therefore,
it is desirable to understand the pollutant accumulation
process in order to predict the rates of solid accumulation
and the composition of their pollutants dependent upon land
use activities (traffic, construction, littering, etc.). A
few published data on this subject exist;* however, use of
these data for general application is questionable. The
data are not disaggregated into accumulation rates per
activity and activities per land use type but are aggre-
gated merely on the basis of land use types. Each land use
type exhibits a different rate of dust and dirt accumula-
tion per dry day per ioo feet—curb (Table 5-4). There are
also differences among land uses in the pollutant composi-
tion per gram of accumulated dust and dirt.** The latter
is obvious because land use types differ in their prevailing
activities, which in turn cause different pollutant genera-
tion patterns (see Table 5-5). The number of dry days prior
to the storm event is important in terms of total mass built
* For example, American Public Works Association (APWA),
“Water Pollution Aspects of Urban Runoff,” Federal Water
Pollution Control Administration, Contract WP—20-15, 1969.
** APWA, op. cit .
5—13
-------
Table 5-4
Dust and Dirt Accumulatjon*
Land Use Pounds DD**/dry day/l00 ft-ci
1. Single family residential 0.7
2. Multi—family residential 2.3
3. Commercial 3,3
4. Industrial 4.6
5. Undeveloped or park 1.5
* Based on 1969 APWA’report for Chicago, op. cit. , p. 56.
** DD = dust and dirt.
5—14
-------
Table 5-5
MG Pollutant Per Gram of Dust and Dirta for Each Land Use Type
I
I-I
U I
Parameter
1
2
3
1000.0
4
1000.0
5 b
1000.0
SS
1000.0
3.6
7.7
3.0
5.0
BOD
5.0
40.0
39.0
40.0
20.0
COD
ColifOrmsC
Settleab].e
N
Solids
40.0
l.3x10 6
100.0
0.48
2.7x10 6
100.0
0.61
0.05
1.7x10 6
100.0
0.41
0.07
l.0x10 6
100.0
0.43
0.03
0.0
100.0
0.05
0.01
P0
0.05
1.00
1.00
1.00
100
Grease
1.00
a values are based on 1969 APWA report, op. cit .
b Values for undeveloped and park lands are assumed (coliform estimation unreasonable).
C UnitS for coliforms are MPN/gralfl.
d All values are assumed.
tLand use types 1 to 5 are defined in preceeding Table 5-4.
Source: SWI1M Interim Revised User’s Manual, Office of Research and Development,
EPA, August 1974, (Draft).
-------
Up or the potential washoff load, The frequency of Street
cleaning also plays a role, If a cleaning takes place
before the next rain, the potential washoff load IS reduced
in relation to the cleaning equipment’s efficiency. If the
interval between storms is long and the cleaning frequency
is low, significant loadings of suspended and settleable
Solids and of organic demands (BaD, COD) are imposed on the
system at the beginning of the storm. The pollutant
accumulation rates of APWA’s Chicago Study* are the basic
data used in most of today’s modeling efforts.
APWA’s study supplied many new data and Compiled and
analyzed others, but a number of presentations are mislead-
ingly definitive in appearance. For example, in that study,
Table 4 (page 30)** contains estimates of monthly Street
litter components from a lO—acre residential area in Chicago
(tons/month). The data are presented not only by month but
also by component. The only component that varies by month,
however, is “vegetation.” The important component “dust
and dirt” accumulated identically in spite of the fact that
several plausible hypotheses could be suggested that would
anticipate seasonal changes due to hydrology and climate as
well as to man’s activities. It should also be pointed out
that in the Chicago data approximately 49 percent of the
“dirt and dust” fraction was insoluble—-essentially all of
it is potentially suspended solids.*** Table S-6strongly
suggests that the Cincinnati overflow datat used in the
APWA study do not resemble the (total) “dirt and dust”
fraction from Chicago. Clearly these are different mate-
rials. On the other hand, a U.R.S. Research, Inc.,Tt
study examined the average composition of dirt and dust
from ten cities. With some variation the total dust and
dirt was similar to that of Chicago (after all Chicago data
contributed to the average). Table 5-7, however, shows
that when corrected for particle size, the dirt and dust
* APWA, op. cit .
** APWA, op. cit .
T. Flaherty, Process Research, Inc., persona], communi-
cation, 1975.
S. R. Weibe]. et a].., “Urban Land Runoff as a Factor in
Stream Pollution,” J. WPCF 37, July, 1964.
tt J. N. Sartor, and G. R. Boyd, “Water Pollution Aspects
of Street Surface Contaminants,” U.R.S. Research Company,
San Mateo, California, for U.S. E.P.A., E.P.A.—R2—72-081,
November, 1972.
5—16
-------
Table 5-6
Comparison of Cincinnati Overflow Rate
and Chicago’s Dust and Dirt Data
Cincinnati
Storm RunOff
Chicago
Dirt and Dust
BO lD 5
COD
VSS
P0 4
Total-N
4. 5—12%
3 3—5 8%
20—30%
0.25—0.48%
0.9 —1.6%
0.36—0.77%
3.9 —4.0%
0.38%
0.005—0.007%
0.51 —0.061%
Table 5-7
U.R.S. Data Corrected for Particle Size
Less Than 43it
4
URS Research, Inc.
Dirt and Dust < 43p
5.8%
33.0%
31.0%
0.88%
0.77%
BOO 5
COD
vSS
P0 4
Total-N
5—17
-------
component less than 43ji was reasonably similar to the
Cincinnati runoff. This explanation is intuitively
attractive because the finer materials are normally more
Susceptible to being washed off.
The difficulties involved in improving the data base
for modeling the quality of runoff are underlined by the
fact that very few additional data have been published.
Table 5—8 summarizes some of these data.*
A simple model has been formulated for the accumulation
process by taking into account for each land use type total
curb length, dry days, Street cleaning, and the correspond-
ing dust and dirt accumulation rate.** The mass of dust
and dirt is then separated into various pollutants by using
empirical fractions dependent on the land use (see above).
Regression equations have been published recently which
calculate accumulation for each land use as function of
traffic volume, pavement conditions, and time since rainfall
or sweeping.*** Verification of the results has not been
reported.
2. Areal Emissions . Some of the information available
in the literature on urban areal and non-point source
emission rates for BOD, CO , N0 3 -N, Total N, Total P, and
sediment has been summarized and is presented in Table 5-9.
The first seven columns represent emissions from stable land
uses. The tenth column is for construction activity; for
comparative purposes the emissions from untreated and treated
residential wastes at 3 dwelling units/acre are listed in
columns 8 and 9. An average of 3.5 persons per dwelling
unit is assumed for the latter estimates. Table 5—9 is
* The derivation of these rates is described in “Progress
Toward Synthesis and Integration of Land Use and Environ-
mental Quality,” Working Paper no. 4, prepared for E.P.A. by
Meta Systems mc, Contract no. 68-01-2622, June, 1975.
** “Stormwater Management Model, vol. 1, Final Report,”
11024 DOC 07/71, Metcalf and Eddy, Inc., University of
Florida, Gainesville, and Water Resources Engineers, Inc.,
Report for E.P.A, July, 1971.
R. Sutherland and R. McCuen, “A Mathematical Model for
Estimating Pollution Loadings in Runoff from Urban Streets,”
Preprint from Proceedings of the International Conference on
Mathematical Models of Environmental Problems, Southhantpton
(U.K.), September, 1975.
5—18
-------
?able 5 —8
1
Co sri,on of *cc i1atioit Rates (lb/dry day/1O.00 0 ft) in Literature
SUS
srr
—
B0D
— —
- —
3** 4 5
1
2
3 4
S
1
2
3
4
S
1
2
3*
4
5 3.
2
4
1
2
— — —
—
—
—
—
— —
.0035
—
.028
17
single
7.0
l5.5
6.63
,7
1.551
.66
.35
.83
2.8
6.62
1.88
3.78
4.2
.034
.328
.28
1.104
.13
.8
.0115
.092
.4
pailti
18.4
36.8
49.5
1.84
3.6(
4.95
.22
.14
.0231
.198 .
23
112.2
90.6
15.02
5.61 1.1.2
9.06
1.5
2.54
Inaustria
31
117.66
3.22
102.L
1.65 10.2
‘Literature values had to be converted to this for3aat of acculsu-
latica rates by making various assumptions.
1: AmeriCan Public Works Association (APWA) • “Water Pollu-
tion AspectS of Urban Runoff. Federal Water Pollution
Control Administration, Contract WP—20—15 (1969).
2: Moesner, stat. • “A Model for Evaluating Runoff-Quality
in Metropolitan Nester Planning.” ASCE Urban Water
Resources Research Program, Technical M orandum No. 23.
.pri1 1974, p. 62.
3: Storut Meter Pollution from Urban Land Activity.” AVCO,
Economic Systems Corporation. Water Pollution Control
Research Series. Federal Water Quality Administration,”
Report Wa. 11034, FRI. 07/70 (1970).
1.38
20.33
18.74
1.2 .135 1.06
.072
— —r —
4
.1751 1.
.0138
1.84 — .0075 .184j
.12
‘4
l.2
4: SartOr, . U., and G. B. Boyd. “Water Pollution Aspects of Street
Surface Contaminants.” URS Research Conpany. San Matec, California.
for U.S. Environmental Protection Agency, EPA—R2-72—OBl, Noveitber
1972 (Table 45, p. 144).
5: Graham. Ph. H.. L. s. Costello, and H. Mallon, “Estimation of Iiiper-
viousness and Specific Curb Length for Forecasting Storiswater Quality
and Quantity.” J. Water Pollution Control Federation . Vol. 46. No. 4.
April 1974.
* Organic Jtjeldahl Nitrogen.
** Soluble Orthophosphates.
c ercia 56
U I
1
I -a
1
OS
14
2.89
.75
.198
-------
Table 5-9
NOfl-Poiflt Source and Urban Stormwater nissiOfl Rates (lb/acre/Yr)
-4
4.,
irrigation
in Western U.S.
Cropland 3 Du/Acre
Active Surface Sub—surface Tile sewage o
FQrest Range Cropland Drainage Drainage Drainage Urban Treated* untreated
COD
(].b/A/Yr) — — — — — — 196—276 240. 1598.1 —
BOD
( lb/A,’Yr) — — — — — — 27—45 95.9 639.3
NOfN
(lb/A/Yr) .62—7.8 .62 — — 74. — - 63.4 0.0
‘ It
Total N
(lb/A/Yr) 2.741.6 — .09-11.6 2.7—24 37.4166 .2711.6 6.2—8 63.4 127.8
Total P
(lb/A ./Yr) .027—.27 .07 .05—2.6 .89-3.92 2.7—8.9 .009—.26 .98—5.0 12.8 31.9
Sediment
(Th/A/Yr) — 75 —750 — 15,000 — — — — 16,000 31.9 213.1 3,000 —
375,000
Source: Loehr, Raymond C., “Characteristics and Comparative Magnitude of Non Point Sources,”
JWPCF , 46(8), AuguSt 1974, pp. 1849—72.
Methods for Identifying and Evaluating the Nature and Extent of Non Point SourceS
Pollutan , EPA, Washington, D.C., 1973.
Bryan, Edward H., “Quality of Storm Water Drainage from Urban Land,” WRB, 8(3),
June, 1972.
Hyphen (-) means data unavailable.
* Standard Secondary Treateent, 60% P- Removal, 85% SOD Removals 50% NjtrifiCatiOfl.
-------
reported in pounds of pollutant per acre per year. Data
are commonly reported in the literature in terms of con-
centrations of pollutants. These data are difficult to
interpret meaningfully without knowledge of the runoff
quantities involved; thus we have chosen to work with total
loadings rather than concentrations. Where entries in
Table 5-9 are blank, it should not be interpreted to mean
that the particular pollutant is not produced by the
associated land use but rather that the data were not
available.
It can be seen that urban runoff produces about the
same order of magnitude of pollutant as secondary effluent
from separately treated sanitary sewage, with the exception
that it is somewhat lower in total nitrogen and higher in
sediment. The implication of this result is that the
attainment of ambient water quality standards in many areas
will require a combination of land use management and higher
treatment levels for point sources.
As indicated in Table 5-9, sediment loadings from con-
struction activities may be as much as one order of magni-
tude greater than from other land uses. Thus even if this
activity is confined to a small area of a watershed, it can
significantly alter sediment yield.
Although the data in Table 5-9 are appropriate for
order of magnitude analysis, they should not be regarded as
useful emission factors. To qualify for use, emission
factors for any particular region must be based upon know-
ledge of local soil characteristics, topography, climate
and other features.
Models for Quantity and Quality of Urban Runoff
Urban runoff models range from those that demand simple
calculations and very aggregated data to those that demand
extensive computer time and information. The information
requirement increases as the models reflect more of the
main processes which govern runoff: interception, evapora-
tion, transpiration, infiltration, surface detention,
overland flow, gutter flow and pipe flow. Empirical and
mechanistic formulations have been used to simulate indivi-
dual subprocesses. Only the computer revolution has made it
possible explicitly to analyze such large and complicated
systems in the level of detail and resolution suggested
here.
Before we discuss our review of runoff models, we
examine the few recurring formulations which have been used
5—20
-------
for calculation of runoff quality. These methods, which
are related to (1) washoff and (2) erosion from impervious
surface areas, are much less advanced than those used to
calculate runoff quantity.
1. Washoff from Impervious Areas . Runoff is the
crucial mechanism for the proliferation of stormwater
pollution. The accuracy of prediction of three processes
influences the quality of pollutographS and loadographs;*
the processes are:
a. peak rate and total volume of runoff;
b. accumulation of pollutants on pervious and
impervious areas; and
c. washoff and transport of pollutants by the
runoff.
The following paragraphs describe formulations relevant for
runoff quality. The problem in modeling the accumulation
process of dust and dirt on impervious areas was identified
when we described the existing data base for this process
(see above). A simple model is used to map the accumulation
of dirt and dust to runoff quality through the medium of
runoff quantity.
Washoff and transport of pollutants by runoff are
represented by a linear diffLrential equatIon. It is
assumed that the amount of pollutants washed off a street
surface in any time interval is proportional to the amount
of pollutants remaining on the street surface. EPA’s SWNM
was the ;Eirst program to include this model. In order to
fit the equation to these purposes some assumptions are made;
these are described below.
Because R, the runoff rate, effects the rate of pollu-
tant removal, the so—called “runoff constant, K, in the
differential equation must be at least loosely dependent on
R. However, given two watersheds identical except for their
areas, a higher runoff rate would occur from the larger
* A loadograph is a graph of p01 lutant load as a function
of time while a pollutograph is a graph of pollutant Ofl-
centration as a function of time.
** Metcalf and Eddy,IflC., et al. , Vol. l to 4, Final
Report, Op. cit .
5—21
-------
watershed for the same rainfall rate on both. The areal
effect can be lessened by dividing the runoff by the
impervious area of the watershed. The impervious area is
used because it is assumed that the amount of the runoff
from the pervious area is negligible -- which might be quite
incorrect for some soils under certain antecedent conditions.
Since cfs per acre are equivalent to inches per hour,* it
has been stated that the “runoff Constant” is functionally
dependent on the runoff rate R 1 (inch/hour) from the
impervious area. Finally, assuming that the constant is
directly proportional to R 1 , the functional form becomes
K = EuRI where Eu = urban washoff decay coefficient. The
original formulation can be integrated over a time interval
E t during which R 1 is constant to yield the rate of removal
of mass from the watershed. The value of 4.6 for Eu is
frequently cited and implies that a uniform rainfall of
1/2 inch per hour would wash away 90 percent of the pollu-
tant in one hour. But not all the dust and dirt on the
watershed can be washed off by the runoff at a given time
t. That implies that all the pollutants tied to the dust
and dirt are not available either. Again the data of EPA’s
SWMM** are generally cited as reasonable values to compute
the available fraction, at any given time, for suspended
solids, settleable solids, BOD, total nitrogen, ortho-
phosphates and coliforms.
The simple simulation model of the washoff process
described above appears to be the only presently existing
predictor of water quality in overland flow.*** Roesner,
et al .,l ’ claim to have computed good results with this
predictor while CO].StOfltt claims the contrary.
* ft 3 acre 12 in 3,600 sec = 9917 in
sec—acre * 43,560 ft 2 * ft * hr H P
** Metcalf and Eddy, et al .
Sutherland and McCruen, op. cit. , have recently presented
a new formulation, predicting the removal percentage of total
solids by different volumes of rain.
- L. A. Roesner, D. F. Kibler, and J. R. Monser, “Use of
Storm Drainage Models in Urban Planning,” presented at the
AWRA, Urbana, Illinois, pp. 400—405 (1973).
1t N.y. Coiston, “Pollution from Urban Land Runoff, “Water
Resources Research Institute, U. of North Carolina, March
1974.
5—22
-------
Unsatisfactory results could be calculated due to the
inadequacy of the formulation as well as due to poor model
calibration
2. Erosion . The universal soil loss equation* is
generally used to compute erosion as a function of factors
of rainfall-erosion, soil erodibility slope length, slope
gradient, cropping-management, and conservation practice.
This equation has been considered the most useful because
it reflects considerable data, especially east of the Rock-
ies. However, one problem occuring mainly in the west has
been in calculating the Wischnieier and Smith** rainfall
erosion indices. This index is calculated by multiplying
total kinetic energy of a given storm (foot-tons per acre)
with maximal 30-minute rainfall intensity of the storm
(inches/hour).
The kinetic energy of rainfall has been given empiri-
cally by Wischmeier and Smith*** and is dependent on the
rainfall intensity. Therefore, in order to compute the
total kinetic energy of a storm, information from recording
raingage charts must be utilized. The kinetic energy for
each intensity increment can be calculated and the result
multiplied by the depth of rainfall at that rate. These
partial products can be summed to yield the total kinetic
energy for the storm (E). The annual rainfall-erosion
factors for approximately 2,000 locations in the U.S. have
been summarized and published in the form of “iso-erodent”
maps.t Although the USDA Handbook was prepared for use in
agricultural areas, the methodology and data can be used
for estimating erosion in urban areas and at construction
* “Rainfall-Erosion Losses from Cropland East of the Rocky
Mountains,” Agriculture Handbook No. 282, Government Print-
ing Office, Washington, D.C., 1965.
** W. H. Wischxneier and D. D. Smith, “Rainfall Energy and
Its Relationship to Soil Loss,” Transactions of the American
Geophysical Union, Vo. 39, No. 2, April, 1958, pp. 285—291.
1’ ’ W. H. Wischmeier and D. D. Smith, 22.• cit .
t U. S. Department of Agriculture, Agticulture Research
Service, “Rainfall-Erosion Losses from Cropland East of the
Rocky Mountains,” U.S. Department of Agriculture, Agricul-
ture Handbook No. 282, 1965.
5—23
-------
sites. If desired, rainfall-erosion factors can be applied
which represent rainfall intensity for a specific storm such
as the maximum intensity storm of record. Thus the factors
are not limited to yearly averages. Simple empirical for-
mulas have been attempted for estimating the rainfall-
erosion index in the western states.*
The soil erodibility factor has been determined for
most of the soils found in the U.S. and are available from
the Soil Conservation Service. Way and Long,** for in-
stance, reproduce a USDA graph of curves which relate the
soil-loss ratio to slope length for gradients from 2 to
20 percent.
The additional necessary information can be found in
SCS maps and through contact with SCS personnel.
Although the universal soil loss equation fits very
well into a planning model, one must recognize that there
are a number of difficulties in applying this equation in
generalized land use planning. These difficulties include:
- finding good data for the rainfall-erosion index as
well as for the erodibility index of certain soils;
- averaging of length and slope factors in complex
topography;
- the equation has been established only to predict
soil loss resulting from splash, sheet and nil
erosion -- it is not supposed to deal with gully or
streambank erosion, or with erosion from non-
arable land;
- according to Williams and Berndt,*** the soil loss
equation will overstate soil loss from land where
* J. K. H. Ateshain, “Estimation of Rainfall Erosion Index,”
J. Irrigation and Drainage Div. , ASCE, IR 3, September,
1974, pp. 293—307.
** D. Way and S. Long, “Soils: Appendix to Report on Year
2,” National Science Foundation Study, Harvard Graduate
School of Design, December, 1973.
J. R. Williams and H. D. Berndt, “Sediment Yield Com-
puted With Universal Equation,” J. Hydraulics Division ,
ASCE, 12/72, Vol. 98, No. HY12, pp. 2087—2098.
5—24
-------
the slope is concave up, and understate it from
land where the slope is concave down; a convex
slope, if it exists, is an unstable condition and
the slope of the side facing the stream or outfall
should be used;
— if the grid system is too coarse, the interrela-
tionship between factors might not be realized
sufficiently. For instance, if the most erodible
soil is not on the steeply sloped land, results
might be different than if it is.
But many successful applications of the formula have
been reported. Way and Long,* for instance, adapted the
model to the Boston Southeast Sector. The rainfall-erosion
factor, according to the USDA, is 130. They found the
erodibility factor for three kinds of soils. They used
cells of one hectare to calculate the slope—length factor.**
Both the cover index factor, C, and the erosion control
practice factor, P, were assumed equal to 1 (construction
areas)
The sediment delivery ratio (SDR) is an additional
factor to be estimated in areawide modeling, in order to
calculate the amount of sediment which finds its way into
the receiving water body. This is an empirical coefficient
and takes account of particle size, flow depth, land use,
etc.
Considerable research is currently being performed in
order to refine and extend the usefulness of the universal
soil loss equation. Emphasis is being given (1) to as-
sessing the effects of vegetative cover and management
* D. Way and S. Long, cit .
** One must remember that because of non-linearity, the
number of segments into which the area is divided will in-
fluence the results.
P and C are the factors to be varied in order to
evaluate the impact of construction.
5—25
-------
variables for largely undisturbed areas, such as forest and
rangeland* and (2) to estimating the erodibility of soils
and subsoils based on fundamental physical and chemical
soil characteristics.** Williams*** modified the universal
soil loss equation in order to improve its application for
predicting storm sediment yields. He substituted the
product of storm runoff volume and peak runoff rate for
the rainfall energy factor. This modification should over-
come the problem that there is no single-valued relation
between sediment yield and a rainfall energy factor; thus
it should be possible to obtain for identical rainfall
amounts and intensities varying sediment yields if antecedent
moisture conditions are not identical. Williams used 18
watersheds in Texas and Nebraska to test the modified
equation. It explained 92% of the variation in sediment
yields. But due to the limited application of this equation
only the original equation has been introduced into the
available computer packages up to now.
These two mechanisms, washoff from impervious areas
and erosion, were found in various runoff models which we
* W. H. Wischmeier, “Estimating the Cover and Management
Factor for Undisturbed Areas, Purdue J. , pap. 4916, 1972.
** See, for example, W. H. Wischmeier and J. V. Mannering,
“Relation of Soil Properties to Its Erodibi].ity,” Soil Sd.
Soc. iner. , 33, pp. 131—137, 1969, and W. H. Wischmeier,
C. B. Johnson, and B. V. Cross, “A Soil Erodibility Nomo-
graph for Farmland and Construction Sites,” J. Soil Water
Conservation , Vol. 26, pp. 189—193, 1971.
*** J. R. Williams, “Sediment Yield Prediction with tjni-
versal Equation Using Runoff Energy Factor,” paper presented
at Sediment Yield Workshop, USDA Agriculture Resource Ser-
vice, Oxford, Mississippi, Nov. 28—30, 1972.
5—26
-------
investigated in some detail.* Table 5-10 characterizes
some of these models:**
- Model of URS Research Company (URS).***
- Model of Avco (Tulsa).t
- Model of University of Cincinnati (CURM).tt
- Model of EPA (SWMM).ttt
- Model of Road Research Laboratory Method (RRL).
* Working Paper No. 2, “Review and Analysis of Land Use
and Water Quality: Concepts and Models of the Relationships,”
prepared for the Environmental Protection Agency under Con-
tract No. 68-01-2622 by Meta Systems mc, January, 1975.
Only after most of our limited literature review was f in-
ished, an extensive review of conceptual models became known
to us. A. Brandstetter, “Comparative Analysis of Urban
Stormwater Models,” Batelle Pacific Northwest Laboratories,
Environmental Management Section, for U. S. Environmental
Protection Agency, 1974.
** Abbreviations in parenthesis will be used for further
characterization of the models.
*** “Water Quality Management Planning for Urban Runoff,”
URS Research Company for U. S. Environmental Protection
Agency, Office of Water Planning and Standards, Contract
No. 68—01—1846, August, 1974 (Draft).
t “Storm Water Pollution from Urban Land Activity,” AVCO
Economic Systems Corp., Final Report, Contract 14-12-187,
FWQA, 1970.
ti. C. Papadakis and H. C. Preul, University of Cincinnati
Urban Runoff Model, J. Hydraulics Division , ASCE, Vol. 98
(HY1O), October, 1972, pp. 1789—1804.
ttt Metcalf and Eddy, et al . ( . cit.)
§ M. L. Terstriep and J. P. Stall, “Urban Runoff by Road
Research Laboratory Method,” J. Hydraulics Division , ASCE,
95, November, 1969, pp. 1809—1834.
5—27
-------
Table 5-10
Comparison of Comprehensive Storm Water Management Models
Model
tJRS
TULSA
CURM
SWMM
RRL LECLERC
STORM
Temporal
Description
of Event
Discrete
Discrete
Discrete
Discrete
Discrete
Continuous
Interception
Evaporation
Transpiration
Neglect
Neglect
Neglect
Neglect
Discrete
Neglect Neglect
(Discrete Poss)
Neglect
Depression
Storage
-
Neglect
Neglect
Exponential
Filling
Rate
Fills before
overland flo
empirical
assumptions
Neglect Neglect
npirical
assumptions
Infiltration
Neglect
(impervious)
loss
factor
Horton
Horton
Neglect Horton w/constan
rvious) parameters
(impe
implicit in
runoff
coefficient
Overland Flo
•
Time
nomo-
graph
lag
Neglect
Profile with
depth (empiri*
WINNING
cal rel.)
increasing
ematic
ec uation
wave
{
Linear Tizne
Area Routinc
‘
wave
equation
Kinematic
.
%Rational
ala
{orm
(Mod.
flow
Gutter
Neglect
Outflow =
Z inflows
Manning-
Izzard
Neglect
No storage
using average
routing/lagged;
weighted
velocity
storage
Manning
routing
storage
lagged using
routing;
full bore
velocity
Pipe
Flow
Surcharge
Neglect
Neglect
Neglect
(gives mdi-
cation)
stores
(volume
contin.)
increases Neglect
pipe
diameter
Neglect
-------
Table 5-1.0 (continued).
U’
1
0
Coaparison of .Coaprehensiie Storm Water Management Models
Model
URS
TULSA
CURM
S l11
RRL
LECLE1
STORM
Data
eeded
I
Land use type
climatic
antecedaiit
slope conn-
ected Im fer,
vious main
drainage -,
lengtb Storm
event; pollii-
tion loading
land use
antecedant
street area
environmental
index (house,
lot, parcel)
land use
antecedent
physiographic
% Impervious
event; sewer
network
Manning’s n
land use
antecedant
physiographic
C impervious
sewer network
event; pollution
loading
Manning’s n
.
land use
terrain
connected
impervious
length of
main drain-
age; event
)Ianning’ S
n
See SWZV(
(also detailed
map)
land use
terrain
runoff coeff.
antecedant
pollutant load-
ing erosion/rain
record avail-
able storage
and treatment
Pollutant
Transfor—
mation
all of them:
five cate—
gories
:
—
.
.
12 pollutants
SOD, calif.,
SS, etc.
—
—
SS, settleable
solids, SOD,
N, P0 4 sedi-
ments
Output
hydrograph
loadograph
pollutograph
-
pollutant
load
runoff
(regres—,
sion)
h ra h
ag p
hydrograph
pollutograph
storage!
treatment
requirethents
erosion
h dr h
y ograp
h - r h
rag ap
hydrographs
pollutographs
storage/treat-
inent require-
inents erosion
Calibration
Verification
,:
None
-
None -
Yes!
Question-
able
(Yes/quantity
{okay; quali
(tY question—
able
(Yes!
. limited
(
Yea/question—
able limited
application
(Yes/in
process
-------
- Portion of MIT Catchment Model (Leclerc).*
- Model of Hydrologic Engineering Center, U. S. Corps
of Engineers (STORM).**
The table is self-explanatory, so only a limited
number of comments are made.
The first two studies (URS and Tulsa) are empirical and
employ simple formulations. The three models which have
been cited frequently in recent years have been CURM, SWMM
and RRL. Literature discussions have presented comparisons
among these models in order to establish the recognized
superiority of one or another. Heeps and Mein*** presented
an instructive comparison of the three models. It discusses
a number of important factors such as different levels of
aggregation, which influence model results, and is the only
study of its type currently available. Another recent com-
parative study of the same models became known to us; it
confirmed Heeps and Mein’s results.t In general, SWMM is
recognized as the model which gives the best results for
small catchments (less than five square miles).
The model of Leclerc was developed as part of the
overall MIT—catchment model. This part was selected for
review because it was the most recent in a series of
documents which refer to the MIT—catchment model, even
* G. Leclerc, “Methodology for Assessing the Potential
Impact of Urban Development on Urban Runoff and the Rela-
tive Efficiency of Runoff Control Alternatives,” Ph.D.
Thesis, MIT, June, 1973.
** U. S. Corps of Engineers, “Urban Runoff: Storage,
Treatment and Overflow Model, STORM,” U. S. Army, Davis,
California, Hydrologic Engineering Center, Computer Pro-
gram 723—S8—L2520, May, 1974.
D. P. Heeps and R. G. Mein, “Independent Comparison of
Three Urban Runoff Models,” J. Hydraulics Division , ASCE,
Vol. 100, HY7, July, 1974, pp. 995—1009.
t J. Marsalek, T. M. Dick, P. E. Wisner and W. G. Clarke,
“Comparative Evaluation of Three Urban Runoff Models,”
Water Resources Bulletin , Vol. II, April, 1975 (No. 2),
p. 306.
5—30
-------
though it is very complicated. In general, the MIT—catch-
ment model is viewed as a good predictor of the runoff
from large watersheds.*
The last model examined in detail is the STORM model
designed at the Hydrologic Engineering Center in California.
Of the models we reviewed in detail, this is the only con-
tinuous one. It retains conditions due to previous storms
and uses them as input for antecedent conditions of the
present one. It uses an adjusted rational formula con-
taining a composite runoff coefficient. In addition to a
series of runoff hydrographs, it also produces a series of
pollutographs for suspended solids, settleable solids,
BOD, N, and P0 4 , using formulations described above. STORM
models storage—treatment requirements, if desired, and
takes into account erosion by applying the universal soil
loss equation. STORM as recently been recoded by Meta
Systems such that coliforms (MPN) can also be used for
characterization of runoff quality. It does not have any
routing routines so that the application is limited to grid
spacings of approximately 10 square miles. But the model
could be modified through introduction of a simple time-
area routing mechanism,** so that larger sub-basins could
be simulated without a separate water quality routing
model.
The necessary data consist of land use categories,
terrain description, pollutant loading, runoff coefficients,
antecedent conditions, available storage and treatment,
erosion and a precipitation record. Non-urban and urban
runoff are included; the resulting equations are basically
the same. Non-urban runoff can be interpreted as urban
runoff from a land-use category with negligible impervious
area. Various types of outputs may be generated. The
time interval, in the current version of the model, is one
hour.
* Brendan Harley, Presentation on Urban Runoff Models,
BSCE/ASCE -- Meeting, April 8, 1975.
** See, for example, the method applied by 0. Lindholm,
Modeling of Wastewater Disposal Systems , in: R. A.
Deininger, (ed.), Models for Environmental Pollution
Control, Ann Arbor Science, 1973.
5—31
-------
Other Emissions and Interfaces with Water Quality
Beyond the point and non—point source emissions de-
scribed above there are other sources of emissions which
have a bearing on water quality. The following pages cover
solid waste disposal, on-site wastewater disposal and other
types of land use effects which impact water quality.
Solid Waste Aspects . Since solid waste has various
implications for water quality, many of these implications
are qualitatively recognized, but unfortunately have not
been formulated in a functional way.
There is no consensus in the literature on the amount
of solid waste generated in urban areas. This results from
the scarcity of good samples as well as from the problem
of discriminating between waste generation and waste col-
lection rates. The collection rate from regular pick-ups
is usually substituted for the generation rate because the
sampling is done only on the collection rate. Communities
have different regulations regarding the types of resi-
dential (and commercial) solid wastes eligible for regular
collection vehicle pick-up. Actual practice depends
heavily on the attitude of the collection crew. Five
pounds per capita per day in an urban area has frequently
been cited as a representative figure of the pick—up rate.
But data recently collected by the Data Acquisition and
Analysis System, Inc., for the Environmental Protection
Agency Solid Waste Management Office indicate that five
pounds per capita per day is far too high for residential
areas.* Its figures, based on the per capita pick-up
rates of 23 collection routes in five metropolitan areas
during January and February, 1972, indicate 2.0 to 2.5
pounds per capita per day. But reliance on this average
value might lead to inadequate policies due to high vari-
ances. For example, in a recent study in Springfield,
Massachusetts,** the average value was found to be
* “Data Acquisition and Analysis System, Quarterly Report
January-March, 1972,” report by ACT Systems for Environ-
mental Protection Agency, Solid Waste Management Office,
April, 1972, Contract Nos. 69—03—0034 and 68—03—0097.
** P. M. Meier, J. Kühner, and R. Bolton, “Wet Systems for
Residential Refuse Collection: A Case Study for Springfield,
Massachusetts,” Curran Associates, Inc. for Environmental
Protection Agency, 1974, NTISB-PB—234-499.
5—32
-------
approximately 2.5 pounds per capita per day, but the data
showed significant seasonal and spatial distribution with
high coefficients of variation. The composition of the
wastes also varied seasonally.
These facts indicate the importance of considering a
variety of events instead of a single event as the “planninç
event” in order to accommodate the spatial and temporal
differences. Lower per capita collection rates in core
cities, but greater potential for spillover* and thus
greater pollution potential from heavy runoff (due to high
degrees of imperviousness) have to be contrasted (or
balanced) with higher collection rates per capita in sub-
urban areas, but lower potential for spillover and thus
lower peak pollution potential from runoff (due to higher
perviousness). The seasonal significance of organic and
nutrient impact on the suburban runoff quality from garden-
ing and its wastes must also be considered in this light.
Some recent studies have attempted to derive aggregated
estimation of the solid waste generation rate (i.e., demand
rate for collection services) as a function of socio-econo—
mic factors.** The use of input-output analysis has been
suggested to predict generation rates of industrial wastes
on a regional basis.*** The number of employees, location,
type of commercial firm, etc., have been applied as in-
dependent variables, to estimate rates of commercial solid
* Inadequate storage of wastes for pick-up in high-density
living areas, stray dogs, rats, etc., lead to a high spill-
ing rate in core cities; see, for example, Alan M. Beck, The
Ecology of Stray Dogs, A Study of Free Ranging Urban Animals ,
Baltimore; York Press, 1973.
** See, for example, D. Grossman, J. F. Hudson, and D. H.
Marks, “Waste Generation Models for Solid Waste Collection,”
Journal of Environmental Engineering Division , ASCE Vol.
100 (EE6), December, 1974, pp. 1219-1230. R . Bolton, in
P. M. Meier, et al., 22 • cit .
H. I. Stern, “Regional Interindustry Solid Waste Fore-
casting Model,” Journal of Environmental Engineering Divi-
sion , ASCE, Vol. 99, December, 1973, pp. 851—872.
5—33
-------
waste generation.* It is difficult to find good approxi-
mation for the spillover rates from the collection of the
various solid wastes, even though these rates would be of
interest in determining the dust and dirt accumulation
and its composition in impervious urban areas (see above).
Relatively few studies have been done on the simul-
taneous estimation of composition and magnitude of solid
waste as functions of socio—econornic, seasonal and loca-
tional factors.** National averages are usually available.
Some attempts have been made to estimate composition on
a regional basis. These data are presented for comparison
without considering yard waste.***
Further data collection will be necessary before all
of the factors influencing waste generation are known and
their interrelationships are specified. Unfortunately,
data on seasonal variation of yard wastes (as well as of
other components in cases of spills) are required to evalu-
ate the impact on the quality of runoff. But seasonal
fluctuations of solid waste composition are very seldom
investigated and reported. The American Public Works
Association cites some data from New York and Chicago
(1939 and 1956, 1957 and 1958), but these seem to be out of
date. Hence, the necessity of making reasonable assump-
tions about the composition of solid waste and of its
spilled portion remains with the analyst.
* T. V. de Gueare and J. E. Ongerth, “Empirical Analysis
of Commercial Solid Waste Generation,” Journal of Sanitary
Engineering Division , ASCE, Vol. 97, SA6, December, 1971,
pp. 843—850.
** One study by the Environiu ntal Protection Agency indi-
cates that at least among low-income families, areas of
multi—family dwellings may contribute less waste per capita
(and, of different composition) than families in less dense,
single—family structures: George R. Davidson, Jr., “Resi-
dential Solid Waste Generated in Low Income Areas,” U. S.
Environmental Protection Agency, Washington, D. C., 1972.
W. R. Niessen and J. Chansky, “The Nature of Municipal
Solid Waste,” ASNE, Incinerator Division , 1969.
t See American Public Works Association, Refuse Collection
Practice, Inter—State Printers and Publishers, Danville,
1967, p. 38.
5—34
-------
In summary, the number and precision of formulations
available for the generation of solid waste, including its
magnitude and composition, are very limited. The relation-
ship between generation and pick-up of solid waste is not
well documented, nor is the influence of pick—up rates and
methods on the accumulation of. dust and dirt in urban areas
However, it is necessary for the analyst to consider the
quality of solid waste management in his estimation of dust
and dirt accumulation.
Another important aspect is the potential danger of
leaching from disposal areas. Recent reports, for example,
from the Llangollen land disposal site in New Castle County,
Delaware,* have confirmed the dangerous impact leaching
might have on the quality of groundwater. The de9ree of
pollution from solid waste disposal areas has been a focus
of controversy for a long time. In recent years, several
attempts have been made to provide quantitative information
on the behavior of areas for land disposal of solid waste,
to quantify the amount and composition of leachate,** and
to develop specifications to control and/or prevent pollu-
tion. It is clear that as long as there is positive net
infiltration (net infiltration = rain - runoff - evapo-
transpiration), leachate will eventually be produced by the
solid waste disposal areas. The magnitude of infiltration
depends on surface grading and drainage characteristics,
surface treatment and planting of vegetation. The travel
time to the bottom layer of the landfill depends on the
water absorption capacity of the components in the waste.
The type of subsurface then determines how much of the
leachate finally reaches the groundwater. A bottom layer
of clay might essentially seal the’ disposal area from the
groundwater. The use of plastic liners as bottom layers
would have a similar effect.*** This technique requires
that the bottom layer be above groundwater, in contrast to
* See comments of Environmental Protection Agency officials
before a subcommittee of the Committee on the Environmental
Pollution of the Committee on Public Works, U. S. Senate,
July 17 and 18, 1974.
** Leachate is defined as a solution of the dissolved and
finally suspended solid and microbial wa te matter produced
by the contact of water with decomposing solid wastes.
*** See: Anonymous, “Asphalt-Lined Gravel Pit Solves Dis-
posal Problem,” The nterican City, 89(6): 68, June, 1974.
5—35
-------
the fzequent practice of uncontrolled disposal into areas
with groundwater contact.
The composition of the leachate that might reach the
groundwater depends on the composition of the solid waste,
on the soil type of the bottom layer, and on the depth to
the groundwater. The rate of solution varies for different
components of the waste.
Remson, et al.,* developed a moisture—routing procedure
for unsaturat T ndf ills. The model is for the one—dimen-
siona]. downward vertical flow system. The equation of con-
tinuity is used for predicting the hydrologic performance of
any soil and refuse layer. Thus the physical system is an
initial and boundary value problem. The uppermost or cover
soil layer obtains moisture by precipitation and loses mois-
ture by• evapotranspiration and downward drainage. In under-
lying layers, moisture is added by drainage from overlying
layers. Prom there it moves to still lower layers or is
reduced by evapotranspiration if roots penetrate to, or
almost to, the layer. The landfill’s hydraulic characteris-
tics have to be determined by experiments or comparison to
similar materials whose characteristics are tabulated. Thus
the model pays attention to the content, spatial distribution
and time variation of moisture by calculating the available
storage capacity** of each soil and refuse layer. The rout-
ing model has not been extended to include the quality as-
pects of leaching.
In order to evaluate the potential of pollution from
landfills, various studies have attempted to simulate land-
fills under laboratory conditions by using different soil
types, infiltration rates and waste compositions.*** Only a
* I. Remson, A. A. Fungaroli, and A. W. Lawrence, “Water
Movement in an Unsaturated Sanitary Landfill,” J. Sanitary
Engineering Division , ASCE, SA2, April, 1968.
** Available water is defined as a moisture range between
field capacity and the permanent wilting percentage, or
initial moisture (whichever is greater).
A. A. Fungaroli and R. L. Steiner, “Laboratory Study of
the Behavior of the Sanitary Landfill,” J. Water Pollution
Control Federation , February, 1971; S. R. Quasim and J. C.
Burchinal, “Leaching from Simulated Landfills,’! 3.. Water Pol-
lution Control Federation , January, 1970, pp. 371 379; and
R . Helmer, “Menge und Zusammensetzung von Sickerwä$sern aus
Deponien verschiedener Abfallstoffe,” Mull. und Abfaij , 3, 1974
5—36
-------
very few field data exist. Rovers and Farquhar* tried in a
semi-natural environment (cylindrical cells buried in a
landfill) to investigate the impact of variations in infil-
tration (such as seasonal changes) on composition and pro-
duction rates of leachate and gas. They observed a linear
relationship between the amount of material extracted per
ton of residential solid waste and the specific volume of
liquid movement through the refuse in inches per foot in
similar densities. The amount of leachate produced, and
its chemical and physical characteristics, varied seasonally
with changes in the amount of infiltration. Leachate
strength (as reflected by COD, N-NH 3 , N-ORG, Ca, Mg, Fe
and Cl) increased rapidly with an abrupt increase in inf ii-
tration. Alkalinity and pH decreased. These data are not
conclusive, but it is possible to utilize them as an esti-
mate of the most significant impact in cases of similar
conditions. But it seems impossible at this time to build
a general model to quantify the impact on the aquatic en-
vironment of existing or phased-out landfills due to the
lack of data. No leachate should reach the groundwater (in
case of impermeable bottom layers) from new landfills, or
the quality of the leachate should be improved by self-
purification during the vertical movement to the ground-
water, such that its impact is negligible. The majority of
states has introduced regulations which should assure proper
construction and operation of new landfills. We suggest
neglecting the direct impact of solid waste disposal areas
on the aquatic environment in the set-up of the model chain,
because the main pollution potential, namely leachate from
existing and phased-out landfills, cannot be assessed due
to the present lack of data on the spatial and temporal
distribution of the quantity and quality of leachate and
because modeling of the impact of pollution sources such as
leachate, on groundwater quality seems rather impossible
(see below) at present.
Aspects of On-Site WasteWater Disposal Systems . Most
of the modeling effort for on-site wastewater disposal has
concerned the engineered, economic design of septic tanks
as pre-treatment devices and subsurface soil—absorption
* F. A. Rovers and G. Farquhar, “Infiltration and Landfill
Behavior,” J. Environmental Engineering Division , ASCE, Vol.
99, October, 1973, pp. 671—690.
537
-------
fields* as disposal devices. Over many decades there have
evolved design rules based principally on balancing the
objective of returning the transporting water to the en-
vironment (principally by dispersal to the groundwater and
evapotranspiration)** and that of containing potentially
harmful organisms. The poorly understood mechanisms in-
volved have led to imposing minimal distances between soil-
absorption fields and wells used for drinking water supply.
Such design rules and constraints are the consequences of
empirical observation; at the present time they cannot be
related to measures of environmental quality.
Patterson,et al.*** have prepared a useful literature
review pertaining to septic tanks and soil absorption
fields. Kaufmant reviewed the chemical pollution of ground-
water, including that due to septic tanks and soil absorption
fields. Increasingly data are being accumulated on a case—
by-case.basis. For the most part such studies are rather
anecdotal in nature. Little in the way of generalization
appears possible.
* Usually in the form of shallow trenches or beds, or
deeper pits.
** Designs for long-term acceptance of septic-tank effluents
are usually based on short-term fresh water percolation
tests; see Manual of Septic-Tank Practice , U. S. Public
Health Service, 1957.
*** J. W. Patterson, R. A. Minear, and T. Nedved, “Septic
Tanks and the Environment,” Illinois Institute for Environ-
mental Quality, June, 1971, National Technical Information
Service, No. PB-204 519.
t W. J. Kaufman, “Chemical Pollution of Ground Waters,”
J.. American Water Works , Vol. 66, pp. 152-159, 1974.
5—38
-------
Specific pollutants are being studied.* However,
the literature often contains simplistic remarks of doubtful
generality such as, “Where excessive number of septic—tank
tile fields are found, especially in water—logged soils,
detergents may still persist over considerable distances.”
Freundlich isotherms have been applied** for prediction of
ammonia nitrogen, phosphate, and ABS. Unfortunately, for
the most part these isotherms have concerned laboratory soil
columns, containing such soils as Zimmerman Sand, Hayden
Silt, or Milaca Clay. Therefore, it is not immediately
clear that these formulations will be useful in the near
future.
For some regions the SCS publishes maps showing where
on-site wastewater disposal should not be attempted. These
restrictions are based on judgmental factors developed by
* S. A. Klein, “The Fate of Detergents in Septic Tank
Systems and Oxidation Ponds,” Sanitary Engineering Research
Laboratory, SERL Report No. 64-1, University of California,
Berkeley, 1964.
S. A. Klein and P. H. McGauhey, “Effects of LAS on the
Quality of Wastewater Effluents,” Sanitary Engineering Re-
search Laboratory, SERL Report No. 66-5, University of Cali-
fornia, Berkeley, 1966.
S. A. Klein, “The Fate of Carboxymethyloxysuccinate in
Septic-Tank and Oxidation Pond Systems,” Sanitary Engineering
Research Laboratory, SERL Report No. 72-10, University of
California, Berkeley, 1972.
S. A. Klein, “The Fate of NTA in Septic-Tank and Oxida-
tion Pond Systems,” Sanitary Engineering Research Laboratory,
SERL Report No. 71-4, University of California, Berkeley,
1971.
H. C. Preul, “Contaminants in Ground Water near Waste
Stabilization Ponds,” J. Water Pollution Control Federation ,
Vol. 40, pp. 659—669, 1968.
H. C. Preul and G. 1. Schroepfer, “Travel of Nitrogen in
Soils,” J. Water Pollution Control Federation , Vol. 40,
pp. 30—48, 1968.
** H. C. Preul and G. J. Schroepfer, ibid .
5—39
-------
experts who take into account soil type, drainage conditions,
topography, depth to bedrock and groundwater, etc. SCS
studies in Connecticut* have been made, intending to relate
such judgmental factors with the anticipated longevity of
a system.
Clearly, such results are not transferable to other
areas. Thus the lack of a sound quantitative description
of the potential impact of subsurface disposal systems on
the groundwater quality becomes obvious.
Other Aspects . Today’s knowledge of the way land use
influences algae production, temperature, and toxic materi-
als (including pesticides) in the receiving water body is
quite vague. Recent research has shed some light on these
areas** but has not advanced to such a state that a discus-
sion on modeling within the overall land use/water quality
system is justified here.***
Models for Transport, Dispersion and Assimilation
The final class of models reviewed in this section are
those which handle the transport, dispersion and assimila-
tion of waste in surface and groundwater.
Residuals of land use activities impact both surface
water quality and groundwater quality. The progress toward
development of groundwater related models for transport!
* D. E. Hill and C. R. Frink, “Longevity of Septic Systems
in Connecticut Soils,” Connecticut Agricultural Station,
Bulletin 747, June, 1974.
** For example, E. J. Pluhowski, “Urbanization and Its
Effect on the Temperature of the Streams on Long Island,
N. Y.,” U. S. Geological Survey Professional Paper 627-D,
1970; or “A Conceptual Model for the Movement of Pesticides
Through the Environment,” Ecological Research Series, En—
vironmental Protection Agency 660/3-74—024, December, 1974.
Some of these aspects were described in detail in
Working Paper No. 2, January, 1975.
5—40
-------
dispersion and assimilation of pollutants has been lirnited.*
The impact on groundwater quality cannot be easily evalua-
ted, because, first, the aquifer is characterized by a slow
response to input changes, and second, today’s knowledge
of groundwater contamination from septic tanks, soil ab-
sorption fields, etc., is rather anecdotal in nature (see
above). In the last years attempts, mainly of the U. S.
Geological Survey, have been made to develop models re-
flecting behavior of pollutants in the groundwater. Brede-
hoeft and Pinder** presented a physical-chemical digital
computer model for predicting mass-transport and dispersion
of contaminants under the limiting assumptions that no
chemical reactions take place. The same assumption is made
in Konikow and Bredehoeft’s model,*** which was developed
to predict changes in dissolved solid concentrations in
response to spatially and temporally varying hydrologic
stresses. These models are complicated and numerically and
computationally very burdensome. Other modeling efforts
have been concentrating on various chemical reactions in
groundwater.1 In contrast to these progresses, no success-
ful modeling of decay of organic matter has been reported.
Within the framework developed here, the only models of
interest are those capable of coupling mass-transport and
dispersion, and chemical reactions. No such model has been
successfully devised, calibrated, and verified up to now.
* See B. W. Adrian et al., Groundwater Pollution , Indepen-
dent Work Study, Department of Civil Engineering, MIT,
January, 1973.
** J. D. BredehOeft and G. F. Pinder, “Mass Transport in
Flowing Groundwater,” Water Resources Research , Vol. 9,
No. 1, p. 134—210, February, 1973.
L. Konikow and J. D. Bredehoeft, “Modeling Flow and
Chemical Quality Changes in An Irrigated Stream-Aquifer
System,” Water Resources Research , Vol. 10, No. 3,
pp. 546—562, June, 1974.
f See, for example, L. N. Pluinmer, “Mixing of Sea Water
with Calcium Carbonate Ground Water,” Geological Society of
America , Memoir 142, pp. 219—236, 1975; other research has
been concentrated on describing and programming the reactions
of dissolution and precipitation.
5—41
-------
Therefore we limit our discussion to surface water quality.
The conceptual discussion above introduced the notion of a
transfer function, a set of descriptive relationships which
relate measures of emissions from point and non-point
sources to measures of ambient quality. A set of transfer
functions constitutes a water quality model.
It is beyond the scope of this paper to review the
plethora of water quality models.* Rather, this section
briefly reviews approaches and mathematical formulations
that can be utilized as the descriptive transfer function
for water quality measures.
Choice is often narrowed to two practical approaches
to transfer functions for the development of a descriptive
model of water quality management analysis. On the basis
of careful examination of the characteristics of, and data
availability within, the basin, the flow system is divided
into a series of compartments. Each compartment is regarded
as a continuous flow reactor within which biological culture
is promoted by incessant input of energy and materials. The
output (or more precisely, the state) is a consequence of
complex physical, chemical and biological reaction on the
input. Values of parameters and variables describing the
compartment are approximately uniform and can be well rep-
resented by unique values. In other words, a parameter
value everywhere in the reactor is about the same and can
be lumped into a single, unique value. As a result of this
space—wise lumping the dynamics of the state can be repre-
sented by a system of ordinary differential equations. If
equations so obtained for all compartments are solved simul-
taneously (or sequentially in a straight—run river), a
complete picture of the state (or water quality condition)
is obtained. In most cases, a computer is used in deriving
solutions. Figure 5—2 gives an example of simulated re-
suits for a simple stream system. The state is given as a
scalar. Notice on the left, the value of the state variable
in each compartment is uniform for any time t and that the
profile consists of rectangular steps. The state of com-
partment i as a function of time is shown on the right; it
* For a discussion of the evolution of sanitary engineer-
ing analysis leading to these models, see: H. A. Thomas,
Jr., “Waste Disposal in Natural Waterways,” in Models for
Managing Regional Water Quality , R. Dorfman, H. Jacoby and
H. A. Thomas, Jr., eds., Harvard University Press, Cam-
bridge, Massachusetts, 1972.
5—42
-------
Figure 5-2
Results of Lumped Parameter Approach
for Simple Stream System
X Computed value
t: Time step
I $
I I
I I
Time t— ’
Figure 5-3
Results of Distributed Parameter Approach
for a Simple Stream System
Grid size
x Computed value
—1
I I
I I
I I
I i
_________ ___ I __
I __ __
Distance Downstream
4-
C
0
0.
V
0
a
4-
x: Computed value
&: Time step
4 L
I I
I I
I I
0 TImet
U
a
a,
4-.
2
C,)
4
Comportment (reach)
Down stream
0
4 - I
a,
E
4-
4-
a
a,
4-
a
4-
C,)
5—43
-------
consists of continuous line segments linking computed
points. The contrast between the discontinuous steps on
the left and continuous lines on the right is caused by
lumping performance in space but not in time. In model—
building an important accuracy problem often develops with
regard to lumping: the size (length) of reach (compartment)
is important. It is a compromise between the degree of
resolution required to fulfill the model’s purpose and its
computational feasibility, in addition to considerations
of data availability and variability of parameters. This
kind of lumped approach has been applied in many previous
studies. In particular, the well—known Delaware Estuary
Comprehensive Study has demonstrated its applicability*
as well as Meta Systems’ Study of the Tiber River.**
The second approach often used is that of the distri-
buted parameter system. Although this lies in the domain
of Eulean analysis, as does the previous one, they are
logically different. In the latter approach, a set of
partial differential equations represents the system dyna-
mics. These equations are obtained from a limiting process
by allowing the size of compartments to become infinitesi-
mal. The equations so obtained represent the state at
points but not in reaches (or at nodes, but not along arcs).
Therefore, solutions of the equations must be interpreted
accordingly. In deriving solutions, a river system is
discretized and numerical approximations used to replace
derivatives. Convergence and stability criteria play an
important role in determination of grid size. Figure 5-3
shows an example of solution for a single stream system.
The diagrams on the left of Figures 5-2 and 5-3 show the
distinction between the two approaches. For example, the
Saint John’s model has been built by following the distri-
buted parameter system approach with some modification ***
* R. V. Thomann, “Mathematical Model for Dissolved Oxygen,”
7. Sanitary Engineering Division , ASCE, 85, No. SA 5,
October, 1963.
** Meta Systems mc, “The Tiber River Basin, A Systems
Study: A Water Quality Management Model for the Tiber
Basin,” Consiglio Nazionale delle Ricerche, Institute di
Ricerca sulle Acque, Report No. 5, July, 1973.
R. J. deLucia, E. A. McBean and 7. J. Harrington, “Sys-
tem Optimization in the Saint John River,” presented at the
International Symposium on Applications of Computers and
erations Research to Problems of World Concern , Washington,
D. C., August 20—22, 1973.
5—44
-------
Some of the models developed limit their focus to the
BOD/DO, residual/indicator couple while others examine a
variety of conservative and non—conservative residual/
indicator couples.* But the characteristics and capabili-
ties of available models differ also in a number of other
areas. These include the specificity with which they treat
point versus non-point sources, the time frame of analysis,
whether they use a steady-state or a non-steady-state
approach, etc.**
In general, there are many well-developed physical!
chemical water quality models available to predict the
transformation of loads to instream quality. These are
often of a higher degree of resolution and detail than are
the models to estimate the loads as a function of land
use. Aquatic ecological models are less developed than
the physical/chemical water quality models; and ecological
models which include both terrestrial and littoral grass
systems as well as aquatic systems have only very limited
development. The latter type of model is. necessary to
examine questions concerning the primary productivity!
secondary productivity system in systems with significant
wetlands. And the presence of wetlands can markedly effect
the nutrient pathways in the food web and ultimately the
biological indicators of water quality and level of secon-
dary productivitY.***
* W. Chen and G. T. Or].ob, “Ecologic Simulation for Aquatic
Environments,” Final Report to the Office of Water Resources
Research, Water Resources Engineers, December, 1972.
** The most known non-steady-state model is the receiving
water body module of the Environmental Protection Agency’s
SWMM; it was developed from those models which were capable
of investigating tidaiwaters. The software of the module
was recently revised and is available from EPA; see:
W. C. Huber et al., Storm Water Management Model; User’s
Manual, Versi n II, University of Florida for Environmental
Protection Agency, Project No. R—802411, March, 1975.
*** A “preliminary model” considering such relationships
is presented in AppendiX 3, “Ecological and Water Quality
Constraints” in An Operational Framework for Coastal Zone
Management Planning , Meta Systems mc, . cit .
5—45
-------
Models Developed and Selected
Based on the review of literature and existing pro-
grammed models, we decided to develop models in some areas
for which no appropriate models were available and to select
several existing, programmed models to be used conjunctively
to fill other needs. The review showed that the lack of
models describing all the small subsystems of the urban land
use/water quality relationship does not permit disaggrega-
tion to the degree of detail described in Figure 5-1. It
rather seems appropriate to choose a coarser degree of
aggregation and to group the existing models into the fol-
lowing subsystems: residual generation, collection, emis-
sion, transport/dispersion and assimilation (Figure 5-4).
There are models appropriate for our purposes for some
of these subsystems. For other subsystems such models are
non-existent. Models of residual generation, for instance,
are not very well understood; they are frequently combined
with models describing emission. Sometimes residual gen-
eration mechanisms are replaced by constant values; for
example, models to be used for computing the washoff load
contain dust and dirt accumulation rates which are treated
as constant values per land use type instead of being
functionally related to the land use dependent mechanisms
which cause the dust and dirt accumulation. Estimates of
residual generation of stationary land use sources are.algo
largely based on “rules of thumb” rather than actual models.
We did not find for our purposes any appropriate models
which describe the collection of wastewater or stormwater.
There exist some models for estimating runoff and washoff
rates from storm events over a certain time period which
fit our purposes, even though certain assumptions such as
constant accumulation rates prevent a logic structure which
spans the total land use/stormwater runoff/water quality
relationship (see Figure 5-1). Models describing transport,
dispersion, and assimilation are frequently used and well
developed. Steady-state models are the most common, but
there are also models which are capable of simulating the
non—steady-state conditions due to time dependent variable
input (such as stormwater runoff).
In order to develop efficiently the necessary computer
software —- within available resources —- we have devised
5—46
-------
Figure 5-4: Scheme of Land Use Water Quality Relationship
Ii , ’
-4
-------
two separate computing modules:* first, a module consisting
of the subsystems of generation, collection, treatment, and
emission of sanitary wastewater; and second, a module con-
sisting of the subsystems of (aereal) emission (runoff,
washoff and erosion) and of transport/dispersion and assi-
milation. The former represents our framework for the
analysis of “treatment and discharge” of sanitary waste-
water. The latter makes up our framework of analysis of
the impact of intermittent urban stormwater discharges on
water quality; it is composed of a modified version of
STORM, a continuous model of the rainfall/land use/runoff
relationship (see above), which was linked to the receiving
water body module of SWMM, a dynamic water quality model.**
Our first framework is mainly designed to evaluate the
hydraulic load of sewer and treatment systems, to compute
the rates of utilization, and to indicate potential points
of overutilization where relief strategies are needed. In
neglecting the pollutants in the effluent, we have impli-
citly assumed that effluent standards are met.*** The
actual impact of these discharges on the water quality of
the receiving water body is considered by superimposing a
constant effluent and its assumed quality from the treatment
plants with the time dependent flow of stormwater. SWMM
could have been used for an independent evaluation of
wastewater discharge.t In the second framework emphasis
was put on the impact of stormwater pollution on the
receiving water quality rather than on evaluating the
hydraulic capacity of stormwater related infrastructure.
It has been implied for now that construction of additional
* We discriminate between conceptual subsystems we have
talked about up to now, and computing modules which make up
the software for one or more of the subsystems at a time.
** See W. C. Huber, et al., . cit .
It is feasible to devise a module which computes the
quality of the treatment effluent, but it would not have
added anything to the analysis at the time.
t Clearly the use of a steady—state model for this type
o’f analysis would be more appropriate than a non-steady-
state model, which uses up considerable computing time to
reach a point of stabilization. We have neglected the
steady-state analysis in our project, since it would have
required considerable resources without adding anything new.
5—48
-------
large—scale downstream infrastructure to accommodate in-
creased upstream runoff can be avoided; this means that
upstream measures are applied which insure that the peak
flow in the system does not increase beyond the present
estimation. Developing a module to evaluate stormwater
drainage capacity in a manner similar to the sanitary
sewer capacity evaluation module is very complicated due
to the large number of potential relief strategies (e.g.,
detention and retention ponds, surcharging, reuse, etc.)
and their combinations. Therefore, it was decided to con-
centrate solely on the question of impact of stormwater
on the receiving water body.
The sanitary sewer and wastewater treatment plant
capacity evaluation module has been tested only for an
artificial example, while testing of the linkage module
has been tuned to the Harnden/ConneCticUt area and the Mill
River Basin, in so far as appropriate data were available.
This area and the existing data are discussed in Appendix
A.
Sanitary Sewer and Wastewater Treatment Plant Capacity
Evaluation Module .
At present, in order for the planner to obtain data
concerning the overall performance of an area’s sewer sys-
tem assuming various population projections and development
patterns, detailed studies must be performed for each projec-
tion, usually at a high cost of time and money. These costs
significantly limit the capability of the planner to fully
examine the impacts of the projections, and thus impair his
ability to make more complete final recommendations.
To at least partially fill this gap, a capacity evalua-
tion module of the relevant infrastructure was designed to
provide the means with which a planner can obtain an esti-
mate of capacity utilization of the larger sewer pipes at
points throughout the system by using existing and projected
land use patterns and population data. The module is de-
signed so as to allow for a performance comparison of the
existing sewer system against the present (time T), and a
range of scenarios that may be projected for 10, 25, and 50
years into the future. The final output of the model
consistS of the actual flow, percent utilization and cumu-
lative overflows for each sewer pipe as well as flow and
utilization of the treatment plant. Once these results are
examined, hypothetical new links and relief interceptors
may be added to the existing system and the system’s
performance can again be checked against the projected
scenarios. This process can be repeated until the system
5—49
-------
reaches a desired state of flow design and capacity utili-
zation, or until it becomes clear that a feasible solution
does not exist. It is hoped that through these capacity
estimates the planner would be alerted earlier and more
thoroughly to potential wastewater drainage problems under
a wide range of projected or anticipated conditions.
The approach incorporated in the module enables the
planner, with few constraints, to:
a. Define only those pipes of the sewer system —-
to be called links -- he wishes included in
the analysis, and
b. Take advantage of existing data by defining and
dividing the area under study into subdivisions
-— to be called cells —— that conform to the
simplest form of data collection.
In the paragraphs below we briefly outline the model
structure.* A tree network of links which are to be moni-
tored has to be sketched on a base map of the area under
study. This network will generally consist of sub- and
feeder mains, trunk lines and final interceptors. Then a
grid system of cells must be overlaid on this map. The
cells need not be of any uniform or fixed geometry, but
should be defined to facilitate the collection of data
concerning the population and population equivalents
which are required for each grid cell. It is assumed that
the total wastewater generated within a cell is uniformly
distributed throughout that cell. Each link in the system
is assigned a percentage of the flow generated in the cell
in which the link is located. The assigned flow drains
into the link in a continuous, but not necessarily uniform,
fashion, for the length of the link. On the basis of the
above assumption, capacity checks will occur only at the
downstream node of each link. Nothing is suggested con-
cerning detailed layout and hydraulic design of relief
sewers. This omission is based on the multiplicity of
* Details of the development of the module are described in
Working Paper No. 4, “Progress Toward Synthesis and Inte-
gration of Land Use and Environmental Quality,” prepared
for the Environmental Protection Agency under Contract No.
68—01—2622 by Meta Systems mc, June, 1975.
5—50
-------
technical factors, many external to the module, which
would play an important role in any such statement. When
an overcapacity flow occurs, the planner, among the many
choices, may select an independent branch of links, or a
relief interceptor. In either case the technical assis-
tance of a sanitary wastewater engineer will be needed to
help determine exact location, pipe diameters, minimal
velocity requirements, and other characteristics.
The following data should be collected:
1. Link characteristics, such as identification number,
diameter in inches, length in feet, slope (feet
per 1,000 feet), roughness coefficient, and an
infiltration coefficient -— in gpd/inch diameter/
mile.
2. The total population or population equivalents for
the land use types under consideration.*
3. The expected wastewater generation in gallons per
capita per day by land use.
4. Percentage of cell wastewater allocated to each
link.
5. The type of unit, treatment plant or pumping
station and its capacity in CFS and receiving
flows from the terminal link.
A flow chart (Figure 5-5) indicates the logic of the program.
A sample output of results to be anticipated from the
module is found in Table 5_ll.** The maximum capacity of
the sewer network and of the waste treatment plant are com-
pared to the actual utilization at various future points in
* We consider six land uses: single family —- low density,
single family -- high density, multi-family, commercial, in-
dustrial, and open-space and recreational. The population
equivalents for commercial, industrial, and open-space, rec-
reational land uses should be based upon the wastewater
flow value (gpc/d) assigned to one of the land uses; for
example, single family low density.
** In order to reduce the number of tables, only three of
the six land uses were included in Table 5-11.
5—51
-------
Figure 5—5
C!
Flow Chart
Sewer Routing Module
+
Reconstruct Network Geometry
Caic
link
ulate maximum flow capacity of network
(using Manning’s equation)
by
4 -
Input year (T, T+lO, T+25, T+50)
against which system will be checked
+
Route sewage through system
recording quantity of flow
-4-
Print location and amount of overcapacity flows
+
A
5—52
Input:
1. Sewerage network characteristics
by link
2. Land use population or popula-
tion equivalents for present
(Time=T), T+lO, T+25, T+50
(single family — low density
single family - high density
multi family
commercial
industrial
open-space - recreation)
3. Expected gallons per capita/day
wastewater generated by land use
4. Percent allocation of cell waste—
water to links
5. Peak flow adjustment factors
+
B
4-
Print, by link, maximum flow capacity, actual
flows, percent utilization, cumulative over-
flows
-------
Figure 5- 5 (continued)
Flow Chart
Sewer Routing Module
Print percent capacity
utilization of treatu ent
facilities or pumping station
Decision to add new or
relief links to system
yes
End 1
5—53
[ Change and/or add
appropriate input
data
+
C,
+
Inp:t another
End of
check Year] +
run
f
4.
-
no ÷
4
I
I
+ yes
-------
Table 5-il
LINK CAPACITIES ND PLOWS
....CMECJUNGSYSTEM CAPACITY FCR. PROJECTED PCPULAT ION 0 YEARS...FROM PRESENL
FULL AND/OR OVERCApACITy FLC%iS
LINK 6 FLOW EXCEEDED TIE PAXIMU ’ CAPACITY BY 12.69 PERCENT 1 1.577 C S)
THERE WERE I FULL AND/CR CVERC*PACITY FLowr
(FLOW VALUES IN CFS)
LINK IC MAXIMLM ACTUAL FLOW PERCENT CUMULATIVE
FLOW CAPACITY UTILIZATION OVER FLOWS
1 1.203 1.114 92.61 0.0
.fl 2 1.203 0.101 0.0
1.957 1.533 78.35.._ _____ 0.0..
4 1.203 0.193 16.02 0.0
5 1.2C3 1.023 85.00 0.0
6 12.424 12.424 100.00 1.577
7 5.769 1.354 23.47 0.0
8 12,424 3.731 30.03 0.0
- 9 1.203 0.514 _ - 42.73
10 5.7 9 0.762 13.21 0.0
11 3.548 0.761. 21.44 0.0
12 1.957 1.501 76.69 0.0
TOTAL FLOW ENTERING TREATMENT PLANT (CFS) = 12.424
TREATMENT PLANT CAPACITY
STAGE MAXIMUM PERCENT
CAPACITY ICES) UTILIZATION
PRIMARy 30.00 41.41
SECONDARY 20.00 62.12
TERTIM Y 10.00 124.24
-------
time; and overflow per network link as well as overutili—
zation of the treatment plant are indicated. This mech-
anism permits the planner to assess the impact of an
anticipated development pattern on specific cells (sub-
areas) as well as on the total system. Changing the
development scenario and then computing the associated
utilization rates provides the planner with information
which can be transformed to impact indicators relevant
to the planning process (such as costs to be covered by
local taxes). Overutilization of the capacity of the
sewer network calls for relief sewers which can be built
in various ways, while hydraulic overload of the treatment
plant requires either capacity expansion or construction
of an additional plant at a different location. The
latter approach would be based on a concentration of
overutilized sewers in a certain area and on a relief
strategy which would advocate a separation of the sewer
network into two independent parts.
In summary, this module allows for evaluation of the
sewer and wastewater treatment capacity dependent upon the
prevailing and future land use scenario and permits the
planner to assess potential relief strategies. The di-
mensions of the necessary structures should then be passed
on to the cost evaluation module (described in Section 6)
to compute the associated costs and so to assess the
physical as well as the financial and socio-economic impact
of new development.
Linkage of Runoff and Water Quality Model
The second major module developed in the course of
this work was the linkage of the urban runoff and water
quality models. In contrast to the analysis of the impact
of sanitary wastewater discharges from treatment plants,
no generally accepted planning events such as a minimum
average seven-consecutive-day flow once in ten years have
evolved in the field of stormwater management. There does
not exist any consensus as to which criterion to use to
set the basic conditions: for example, a criterion in-
volving both receiving water body flow and runoff or a
joint recurrence interval for dry antecedent conditions,
runoff, and receiving water body.* This is true for all
* See J. Kühner and M. Shapiro, “Discussion of ‘Urban
Runoff Pollution Control State—of-the-Art’” (R. Field and
J. A. Lager), J. Environmental Engineering Division , ASCE,
1976 (forthcoming).
5—55
-------
intermittent discharges, for urban runoff as well as for
discharges from non—point sources, such as agricultural ot
forestry areas. In our further discussions, however,
we will emphasize urban runoff.
A runoff model seems to be desirable which allows the
planner to examine continuous records of storm and ante-
cedent dry weather conditions, thus enabling him to iden-
tify critical events from these data, instead of using a
model which can only handle a single rainfall event and
therefore requires an a priori setting of the conditions
to be evaluated. Another point of importance is the
degree of temporal and spatial resolution of the model.
Frequently the difference between the purposes of the
model uses is poorly articulated, and there are suggestions
that runoff models, for example, appropriate for detailed
design problems be used in planning.
STORM, a continuous model, is attractive because of
its relative simplicity of use. The data for which the
model calls are not very detailed and seem to be available
in most areas. The major drawback of the model is its
simplified approach to the runoff coefficient, a coeffi-
cient on which much depends. This aspect of the model
may be modified by choosing separate coefficients for each
area; but ultimately, as in the cases of the other models,
much reduces to subjective judgment in the choice of
parameters. Appendix B summarizes some of our experimen-
tal runs concerning the sensitivity of the model to various
parameters and assumptions.*
The model is designed to compute non-urban runoff and
washoff in addition to urban runoff and washoff. The
formulations pertinent to non—urban runoff are set up in
the same way as for urban runoff, while washoff is based
on emission rates rather than pollutant accuxnmulation
rates. Hydrographs and pollutographs are generated for
every hour of runoff. No provision is made for variable
time intervals of runoff generation.
* Details of the model can be found in its manual, avail-
able from the Hydrologic Engineering Center, Davis,
California; see U. S. Corps of Engineers, . cit .
5—56
-------
The hydrographs and pollutographs, generated by STORM
for each sub—basin for certain rainfall events, have to be
transferred to the non—steady-state water quality model,*
in order to evaluate flow and pollution propagation in
the receiving water. A computational link is required
(see Figure 5-6).
The selected receiving water body module has two
distinct phases (hydrodynamic and quality) which may be
simulated together or separately. In the first phase the
time history of stage, velocity, and flow is generated in
the total system, while in the second phase these results
are used to compute the concentration of conservative and
nonconservative quality constituents.** The receiving
water is divided into a series of discrete one- and two-
dimensional elements which are the connections among nodal
points. One—dimensional elements represent rivers and
specific channels, and two—dimensional elements represent
areas of continuous water surface. The velocity of flow
is assumed constant with depth for each element. For
each time-step, the equations of motion and continuity are
applied to all nodes to derive the hydrodynamics for the
system. The results are then used with equations for con-
servation of mass.
Various points have to be considered in the idealiza-
tion of the physical system into one-dimensional (or
channel) and two-dimensional (or area) discrete elements.
The nodes are the points of constant inflow and/or outflow
and of time dependent inflow and/or outflow. Being a non-
steady-state model, SWMN does not have the capability to
input runoff as a line source. Nodal points must have a
minimum distance from each other in order to keep the
numerical scheme stable. The dimensions of rectangular
channels idealizing the river have to be chosen such that
they reflect the real river dimensions to a large extent.
* W. C. Huber, et al. , op, cit .
** See Huber et al. , op. cit. , p. 269. Actually it turned
out that we had to make some changes of the software, in
order to run the phases separately. It isdesirable to run
these phases separately; the repeated use of the hydro-
dynamic routine would only use up enormous computing re-
sources for the computation of the quality, where the hydro-
dynamic conditions are constant.
5,-57
-------
________ ____ ____ ___ ____ Modified_STORM
1— ________ SUB BASINS _______ 1
__ __ I
RAIN I
SIMULATION PERIOD
NORAIN I
__ ________ I
LANDUSE
URBAN; NONURBAN
DUST 8 DIRT
ACCUMULATION
NON URBAN-POLL LOADING r I
J RUNOFF _______
1 WASH OFF
___________ I
STORAGE
Selected TREATMENT
Interval I
‘j RUNOFF; WASHOFFf I
I Pollutograph Hydrograph I
SORT ROUTINE Single Hydrographs and
____ _____ _____ Pollutographs for all subbosins
4 -I
WATER QUANTITY AND QUALITY ROUTING
DO, BOD, SS, MPN, N, p
SWMM
Figure 5-6
Analysis Framework for Stormwater Runoff
5—58
-------
This becomes particularly difficult when flow and quality
of shallow rivers with wide floodplains have to be simula-
ted. The introduction of some “storage nodes” or a
“parallel channel” option might be of some help. In an
urban setting, much runoff flows into the receiving water
without being channeled through drainage systems; that
means, that line-source type runoff has to be idealized
into point discharges, entering at one of the nodes. Six
conservative and nonconservative quality constituents can
be handled by SWMM. The data requirements for this model
do not exceed significantly the requirements that steady-
state models of the same complexity (number of quality
constituents to be modeled) show.*
When we attempted to link these two models (STORM and
SWMM), whose computer programs (deck, listing and docu-
mentation) were available, various difficulties arose:**
1. STORM is a continuous model, while SWMM handles
single events only. SWMM accepts the hydrographs and
pollutographs of all discharge points for, one rainfall event.
Simulation of other intervals for which hydrographs and
pollutographs are generated, should be performed subse-
quently. Thus, the hydrographs and poUutographs genera-
ted by STORM have to be stored on a discs and then handled
individually by SWMM.*** The definition of events pre-
sented some difficulty because STORM defines internal
events only when the rain event generates runoff. This is
a major disadvantage because STORM, can easily generate
runoff in sub-basin 1 while not generating runoff in
* How a calibration/verification exercise of the quality
portion of such models could be done, has been discussed by
Meta Systems at various occasions; see, for example,
R. J. deLucia and T. Chi, “Water Quality Management Models:
Specific Cases and Some Broader Observations,”, invited paper
prepared for presentation at the World Health Organization,
Government of Hungary Seminar on Systems Analysis in Water
Quality Management, Budapest, Hungary, “February, 1975.
** We discuss some details because a potential user, not
well practiced in the application and linkage of large-scale
models, frequently does not recogüize the diversity of the
problems encountered in linking such models.
*** Actually a routine could be developed to include the
“receiving water body module of SWMM in a loop of events.
5-.59
-------
sub—basin 2 for the same rain; the percentage imperviousness
of the sub-basin decides the actual generation of ru off
from rain. Thus, the notion of time intervals was i itro—
duced to control exogenously the rainfall runoff events of
interest.
2. STORM’S hierarchy of loops is watershed, event
(time interval), timestep (of each single event), and
pollutant, while SWMM’s input is organized in the order of
time interval (event) ,* timestep, pollutant, and water-
shed. Thus, a SORT-routine has to be used between the two
modules to make the I/O routines compatible. A number of
significant changes of input/output options had to be
accomplished to make the linked model as flexible as
possible.
Due to these arrangements, there are two points of
interaction for the planner. First, he can choose the rain
intervals (n) within the continuous simulation period of
STORM which he wants to investigate; experience will permit
the elimination of events of antecedent conditions and
rainfall, which are of inferior significance. Then he can
select those intervals (m) from STORM’S output of polluto-
graphs and hydrographs (m n), which seem to have a major
impact on the whole basin. This option permits him to
reduce significantly the computing costs for analyzing only
those intervals of interest to him. But it also allows the
planner to simulate all the events if he desires to compute
a frequency distribution of conditions which are of interes
to his agency. Since the quality module of SWMM can be run
without rerunning the quantity module, water quality com-
putations can be done by varying pollutograph inputs gen-
erated for each point discharge of runoff. This fact also
permits intensive testing of quality related parameters and
thereby calibration of the quality module. The code has
also been altered so that only quality constituents of
interest have to be read from discs filled by STORM with
six constituents.
* Since SWMM considers only one event in its current
version, this step would not have to be introduced in the
hierarchy of loops in SW .
5—60
-------
3. In calculating the quality of the runoff, SWMM’s
quality routine adds up the sediment load, calculated by
the universal soil loss equation, and the amount of suspended
solids, washed off the streets. The program performs this
summation at every hour. In contrast, STORM does not link
the two routines and hence does not sum up the two values.
Actually, STORM only calculates the total value for the
sediment load from each event. Thus, STORM’s subroutine
ERODE was reprogrammed in order to compute the sediment
load for every hour of the rainfall-runoff event. Clearly,
the total values per event are different for the two types
of computations, because the rainfall-energy-relation is
norl_linear.* In case of relatively uniform rainfall inten-
sities over the rain’s duration, the values are quite simi-
lar; but if the rain shows strong peaks of intensity, the
method of adding up hourly values yields higher total
values. Since EPA’S SWMM uses the niversal soil loss
equation on an hourly basis, we recoded STORM in the same
way.
• We have summarized most of the major changes made in
the individual programs. This indicates the efforts which
are necessary to link two models, whose programs are easily
available and whose combined execution seems relatively
àasy (Table 5-12).
The value of this package is determined by the results
it generates and the computations it permits (on the con-
dition that the accuracy of the computations is satisfactory
for the desired purposes). The combined models can be used
jn three basic ways depending upon the objectives of
the analysis:
- Emphasis is on emissions: STORM is repeatedly
executed without subsequent execution of SWMM;
- The impact of emissio a (generated in each selected
r inf all interval) on water quality of the receiving stream
j . of interest: STORM and SWMM areexec ted in series;
- Qtiality analysis under changing hydraulic and
y4rologic conditions is o primary Cop cen : SWMM a.
repeatedly executed without prior execution O STORM.
i • • - - •
Wischmeyer, op. cit .
5-6].
-------
Table 5-12
Changes for an Efficient Linkage between STORM and SWMN
A. Changes in STORM
-- Creation of GPH* files to pass results from
STORM to SWMM
—— Use of rain interval instead of rain event for
file generation
-- Calculate erosion on hourly basis
—— Accumulate erosion over the rain interval
-- Add eroded material to suspended solids
-- Add coliforms as sixth pollutant (in such way as
it was done in SWMM’s runoff and washoff module)•
-- Change output format for pollutants
-- Calculate amount of dust and dirt, accumulated at
the beginning of rain interval, as well as amount
left over after rain event
-- Modify input of land use data (acres instead of
percentage)
-- Frequent debugging of original program (logic,
core, program, files, default)
B. Changes in SWMM
-- Modification to accept GPH as sequential input
-- Write SORT routine to sequence GPH input to be
acceptable for SWMM
—— Selection of one or more out of the six pollutants
supplied in pollutograph
-- Mbdify sw i to run quality phase independently of
hydrodynamics phase
-- Introduction of weighting factors for GPHs
GPH means hydrograph and pollutograph.
5—62
-------
For each use there exists a large number of possible
combinations of input parameters and desired output.
Appendix B provides some examples of possible computation
with STORM, such as the evaluation of impact of land use
patterns on emissions rates or different street.sweeping
strategies under use of different sweeping equipment.
Exploring the ways that different future land uses in river
sub—basins will affect the river’s water quality requires
a close interaction of STORM and SWMM. Appendix C contains
a short summary of land use controls which can be compared
by use of STORM and, subsequently, SWMM. STORM permits a
comparison of the induced emission rates, while the hydro-
graph and pollutograph resulting from these land use controls
allow for the evaluation of the impact of land use and
environmental controls on the quality of the receiving water
body. Altering dimensions of channels to reflect develop-
ment of flow plains or conditions of constant flow and
water quality, for example, gives room for an intensive
study of the impact of time-dependent runoff. Problems such
as the determination of various roughness coefficients for
different flow situations can only be understood and solved
by repeated runs with the hydrodynamic and quality portions
of SWMM.
The following types of output ‘can be generated by
STORM and SWMM:
- the amount of every pollutant (total/year/sub-basin);
- the amount of every pollutant (total/rain interval/
sub—basin);
- the hourly amount of every pollutant per sub-basin
for specified rainfall intervals ;
- the amount of dust and’ dirt on impervious areas at
the beginning and end of rain intervals;,
- total erosion for selected rainfall events and the
amount finally reaching channels ‘(stream)’ after application
of a:,sediment delivery ratio;
hydrograph (total/year/Sub-basin) :
— hydrograph (total/interval/sub-basin);
- hydrograph (total/hour/sub—basin);
stage of water level. at each sele ked node. of the
river system for every rainfall eventi,
56
-------
- water level at every node of the river system for
each day;
- velocity and flow in every channel of the river
system for each day; and
— hourly concentration of selected pollutants at
every selected node.
we have carried out some of the modeling combinations
mentioned above and and have generated most of the possible
output relevant to analysts; but we limit our presentation
to a few examples. Appendix B contains a table of five
rainfall events (Table B—l) which have been used for our
test runs with STORM and which are also the basis for our
computations here. In the following paragraphs we compare
runs and their results in two ways: first, we compare the
effects of various rainfall levels and base flows in the
river on the quality of the receiving water body given the
same land use; second, we compare the impacts of changing
land uses under varying precipitation and base flow con-
ditions. Factors considered in the comparison are pol-
lutant and wave propagation in the receiving stream; super-
imposition of upstream waves and pollutants; accumulation
and behavIor of the pollutant in the lake; and actual
differences resulting from different rainfall and flow
conditions and changing land use patterns.
Results, discussed first, are based on the land use
configuration of 1974 in the Mill River Basin; and later we
will present results derived from the 1985 land use. Table
5-13 contains hydrographs of the first part of the rain
interval 1 (January 21, 1974, 8 a.m. to 6 p.m.), as they are
generated by STORM for the 11 sub-basins in the Mill River
Basin for the 1974 land use conditions (Figure A-l and
Table A—6). Assuming that the precipitation is uniform over
the whole river basin, the differences among the hydrographs
indicate the varying sizes and degrees of urbanization of
the sub-basins. Due to frequent precipitation in the time
of the year represented in interval 1, a relatively high
base flow was assumed in the Mill River; it is characterized
by an accumulation to 215 cfs at the dam, the downstream
point of the river system. Figure 5_7a* shows the change of
the water level at the dam (junction 1) and Figure 5-8 at
the point of inflow from the sub-basin Eaton Brook (junction
* The ordinate of all Figures is “Depth in Feet” according
to the definition in SWMM; it means distance from datum plane
to water level.
5—64
-------
Table 5—13
Hydrographs for All Junctions of the Mill River System
(First Part of Rain Interval 1; Land Use 1974)
INPUT JUNCT! )N NUMBERS FOR FIChS FRCM C*RDS ARE:
2 4 5 6 7 9 10 11 12
13
TIME
HOURS DAYS VOLUME (CFS)
U’
8.OC 0.33 D.C 0.0 0.0 0.0 0.0 0.0 0.0 0.0
UI. c.0
9.00 0.38 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
c.0 . ——
10.00 0.42 34.6 13.0 18.1 21.9 5.5 11.9 2.8 1.2 6 8 14.9
1 .8 .__ -.. --
11.00 0.46 41.5 16.9 28.6 29.3 10.6 20.3 4.3 10.7 177 36.5
50.4
12.00 0.50 36.3 15.7 29.3 27.5 .11.7 21.4 4.3 15.1 21.4 43.5
65.2 . --—--.- —
13.OC 0 54 48.4 20.9 39.1 36.6 15.6 28.6 5.8 20.1 28.6 58.1
87.0 -
14.00 0.58 187.7 80.9 151.7 141.9 60.6 110.8 22.5 78.0 110.8 225.0
33 7. C
15.00 0.63 7.7 31.3 SE.7 54.9 23.4 42.9 8.7 30.2 42.9 87.1
130.5 .. ______ ____
16.00 0.67 43C.0 185.3 347.4 325.1 138.8 253.7 51.4 178.8 253.7 - 515.3
771.9 . ....
17.OC 0.71 54.5 23.5 44.0 41.2 17.6 32.2 6.5 22.7 32.2 65.3
97.8 --—-- —
18.00 0.75 42.4 18.3 34.3 32.0 13.7 25.0 5.1 17.6 25.0 50.8
76.1 . .. - .... .
19.00 0.79 C.C 0.0 C.C 0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0
-------
A. Land Use 1974
Figure 5—7
Stage Graph at Dam
(Rain Interval 1; High Base Flow)
B. Land Use 1985
V
V
I
—
V
I
I
—
01
.1
0 i
C’
V
I
37.fl(U) —
I
V
I
I
V
V
V
—
I
I
I
I
I
I
I
V
I
0.0
as.
a...
* *
* $
* *
• *
* *
* . *
* *
* *
5. - — . -——
• *
* *
* . a
* *
* *
* *
* *
*
* *
* *
* *
* a.
** a
* *
* a.
0* *
* *
* *
* *
* *
*
*
a
*
S
S.
*
* **
* *
a
*
*
**
**
* *
* *
* *
* *
* *
* *
* *
* *
* *
* *
* * 00•
* *
• * *
* a
* . *
* *
* *
* *
* *
* a
* *
** a
* a
A *
a *
* *
* a
* a
a *
* a
a a
* a
* *
a a
a a
a a
at.. -
.
**
aa* aa*
a. - **
a.
t—-.---
— t——X
9.0 12.0 15. 21.0 24.0
I TM j?4 HOURS
.1.I
14
.1-I.
U).
04
S.. S.
I ** . a.
——1
9.0 12.0 15.0 ie.o 21.0 24.0
-------
Figure 5-8
—
V
I V
.7
V
1974)
Stage Graph at Junction 8
.(R&tn Xnt rva1 1; High Base Flow; Land Use
U’
I
$c.,N
V
I
104.000 — .
I.
..,
IN 7
.
....... .- ———— — —* —.—.—- —.________
-- - . .-—-. -
—. — —
.
—.- — .
e ey
1
. *
—
—_________
.1
. 4
. ** . ...
**
.
.
. -t
.7
**
. **
.
.
—
V .
a *
* a.
.
V.
.
* * -
* *
‘.
V
. a * *
V -
sa a. *
I
I
..*‘
* * *
V
* *
,.
,,
.
,..an_
I
* *
* a
. * a
.1
,
S
a.. 5* *
.._._.:.
__
.
.
,
***
a
V ** **
a *. *s...a.a a.. *.. ..*.*.....* .*.* — a...
V -
.
-
V
.
2.4IO0 I—- - 1——————! ————————!——————— I——————f————————1 .————————1
9.0 . 3.0 6.0 9.0 12.0 15.0 13.0 21.0 24.3 3.0 6.0
TIME 4 NC S S
-------
8). The stage graph for the dam shows lump effects from all
upstream input while the graph at junction 8 still reveals
characteristics of the hydrograph at junction 8: a first
peak after 3 p.m., and then another peak after 5 p.m. The
fact that the stage graph at the dam tails off slowly
reflects the upstream inflow, though due to fast travel
times some peak effects of the upstream portion get absorbed
in the peaks of this junction. Here we encounter problems
with rectangular channels as indicated in above sections.
An intermittent river like the Mill River has a very small
bed, but a relatively large flood plain area. It is possible
to compensate for this characteristic by assuming a rela-
tively wide profile for the base flow so that the dampening
effect of the flood plain can be accounted for during higher
flows. However, such an assumption creates numerical prob-
lems because of the shallowness of the base flow through
a wide river bed. If a small profile is set for base flow
according to the actual bed size, the increased flow during
a rain input greatly increases the velocity and produces an
unrealistic water stage level; Figure 5-8 documents this
problem. The analyst has to find some compromise, such as
using storage nodes or parallel channels, to absorb some of
the flood plain effects.* Another problem occurred when a
stage graph at the dam (for the same high base flow and
runoff from rain interval 1) revealed a very rapidly rising
and falling crest. This had been caused by rapid travel
times in the river. The rapid travel times were induced by
low values of the Manning coefficient. Since no travel time
analysis had been done for the Mill River, Manning coeffici-
ents close to the default values in SWMM were used in the
first runs, and resulted in unrealistic travel times.**
Table 5-14 shows the behavior of suspended solids and
Table 5-15 the behavior of coliforms for rain interval 1
(land use 1974) where the first rain period is from 10 a.m.
to 6 p.m. with 1.63 inches of precipitation and the second
period of precipitation is during the next day (9 a.m. to
4 p.m.) with an accumulation of only .56 inches. Due to
* Budget and time did not permit an exploration of these
options. But we feel that these problems frequently encoun-
tered have not been adequately documented in the literature.
** For example, Whipple’s description of selecting the
right Manning coefficient also reflects how inadequate,
simple assumptions such as default values of .018 (as given
inSWMM’s manual) can lead to unrealistic results (“BOD
Mass Balance and Water Qualii,y Standards,” Water Resources
Research, Vol. 6, No. 3, June, 1970).
5 -68
-------
TbI. 5-14
Suspended Solids (ag/i) - Rain Interval
Land Use 1974
I - sigh Base Flow
a’
e
118
13
8 :
86
86
9 1
9.l
92
92
:5
9
19
10
11.7
16.9
8.9
8.9
11.2
11.6
10.1.
11.3
10.2
9.4
10.5
13.3
21.5
12
11.6
22.3
9.3
.13.2
13.2
13.7
12.3
11.
10.8
14.
12.
10.
8.3
14
11.4
- 63.2
1.1.1
1.7.3
27.5
27.5
17.1
24.7
15.5
9.7
15.9
24.7
19.4
16
11.7
102.9
2]..
27.7
31.5
‘39.6
20.4
26.2
17.5
12.9
14.7
27.1
19.4
1
154
263
29
289
159
231
11.6
10
85
105
101
57
34
20
37.9
246
247
214
114
10.3
8
10
95
64
57
89
113
22
185
24.6
223.
10
87
95
9.4
69
74
99
75
89
102
34
18S6
24.
394
8.9
89
8.1
77
97
86
97
78
89
96
3
1 6
24.6
3.7.4
80
77
8.8
88
95
88
96
80
89
92
Z8.
24.6
132
66
86
92
89
93
90
93
83
89_
84
j
18,Q
fl6
12.4
86
84
9
88
89
88
86
73
77
61
12
17.7
182
116
80
78
83
81
8
7.4
69
53
66
47
14
1.7.1
14.9
10.4
7.7
7.4
7.6
7.
6.7
6.1
6.2
5.2
6.5
4.6
16
3.65
123
93
71
66
6.4
61
61
6
61
5
64
44
$
159
115
84
62
6
6
61
65
62
69
57
89
112
24
14.6
r 1
US
72
6.8
7.2.
87
88
95
87
96
80
89
92
Runoff occurred’ frcn 10*00 to 18*00 hours and on the next day fron 9:00 to 16*00 hour..
-------
Junc—
1
2
3
4
5
6
7
8
9
10
11
12
13
8
.06
.08
.007
.001
.001
.001
.001
.001
.001
.001
.001
• 001
.001
10
.06
27.0
.6
1.6
15.0
16.0
4.8
12.0
5.5
1.5
11.0
25.0
76.0
12
.07
57.0
2.6
25.0
24.0
24.0
16.0
9.7
0.0
30.0
22.0
11.0
12.0
14
.4
70.0
11.0
26.0
26.0
22.0
14.0
27.0
6.0
8.6
11.0
11.0
8.6
16
2.6
16.0
22.0
16.0
10.0
21.0
11.0
5.9
6.4
4.1
1.6
1.6
1.1
18
7.5
4.2
12.0
15.0
8.2
5.1
4.0
2.8
0.75
0.5
.57
.05
.05
20
8.5
3.9
12.0
4.7
3.6
3.0
0.6
.37
.36
.03
.001
.001
.001
22
8.7
3.9
9.4
2.7
1.4
.72
.3
.022
.001
.001
.001
.00].
.001
24
8.7
3.9
7.8
.74
.44
.21
.01
.001
.001
.001
.001
.001
.001
4
8.7
3.9
6.5
.22
.1
.13
.001
.001
.001
.001
.001
.001
.001
8
8.3
3.9
3.5
.001
.001
.001
.001
.001
.001
.001
.00].
.001
.001
10
8.].
6.1
2.9
6.1
1.4
1.3
.66
.85
.35
.28
1.7
1.3
1.4
12
7.8
6.8
2.5
1.7
1.6
1.5
.8
.7
.12
.97
1.0
.67
.67
14
7.3
6.3
2.3
1.5
1.2
.98
1.2
.97
.71
.50
.58
.35
39
16
6.8
5.5
1.9
.98
1.2
1.1
.69
.51
.41
.29
.33
.23
.22
18
24
6.4
5.7
5.1
1.6
.95
.65
.54
.32.
.23
.22
.13
.001
.001
.001
5.1
.99
.14
.07
.03
.001
.001
.001
.001
.001
.001
.001
Table 5—15
Coliforai (10 MPN/100 ml) — Rain Interval 1 — High Base Flow
Land Use 1974
0
RunoLf occurred from 10:00 to 18:00 houra and on the next day frc. 9:00 to 16 OO how-
-------
the high base flow associated with relatively low suspended
solids concentration (which might be a doubtful assumption)
the suspended solid concentration in the river due to runoff
remains quite insignificant, except for the one peak inflow
from the most urbanized area, of the basin, the western part
of Lake Whitney (junction 2). Generally, the concentrations
ranged from 40 mg/i. to figures as low as 3.4 mg/2 at the
end of the runoff period when runoff contains almost no wash-
off. The second rain period does not contribute to an in-
crease in concentration except at junction 2. Small storms
following a large storm do not have much material to wash
off in impervious areas. For instance, in this case, fig-
ures for Shepard Brook indicate that in the first ten hours,
51,800 pounds of dust and dirt were washed off from the sub-
basin by 1.56 inches of rain, *hile during the second event
only 1,599 pounds were washed o Cf by .56 inches of rain.
it should be noted that combine input of erosion generated
by the universal soil loss equation and washoff from the
urbanized areas would have significantly increased the con-
centration figures of suspended solids. But intensive
testing of a combination of these two options has not been
done. The accumulation of suspended solids and then the
slow drop-off of the concentration in the downstream portion
of the lake can be easily recognized. The time of delay
from the peak flow at around 4:00 p.m. to the peak concen-
tration is approximately ten hours The trend of coliform
is similar to that of suspended solids. A similar delay of
the peak concentration can be found; contribution of the
second storm is minor except at junction 2; and concentrations
are generally not very high.
The second rain interval (June 16, 7 p.m. to 9 p.m.,
precipitation of .51 inches after a dry antecedent period
of 367 hours) has been investigated for two flow conditions:
one accumulates to 105 cfs at the dam (called medium flow)
and the othe± to 69 cfs (called low flow). Figure 5-9a,
the stage graph at the dam for low flow condition and 1974
land use, reveals two peaks: one caused by the nearby heavy
runoff and the second by accumulated upstream effects, In
contrast to this graph, the stage graph o f the first rain
interval (Figure 5-7a), which started from a much higher
stage due to the high base flow, did not reveal these peaks
because the ten hour interval of rain at that specific
ntens1ty distribution dampened the upstream peak effects.
The figures show that the peak caused by rain interval 2 is
smaller than that of interval 1 by approximately .65 feet
Ø age level. The stage graph based Ofl the medium flow
dition has the same s) ape but the falling crest is con-
tracted by approximately two hours due to the shorter travel
times.
5—71
-------
Figure 5—9
Stage Graph at Darn
(Rain Interval 2; Low Base Flow)
B. Land Use 1985
ass
* 0
a.
36. 00
• 36.700
DEPTH IN
Ui FEET
-J
36.600
36.500
36.400
S
S
••—*
*
*
*
:
*
*
*
* .*
a * *
* * *
** *
* *
*
A. Land Use 1974
*
5*4* *5*4 * 5 * 5 *
lj . ‘V.
I —
I —--——-—•— —..- —•.—
a
S
a
a
:
- *
a
-
*
a
a
*
*
a • a
S *
*
a
a
a
*
*
*
*
a
a
a
a
a
a
a
S
*
5
as
a
•
-
as
a
*5*
--
* 5*5
aaa*****s*
.0 le.0 .. 21.0 24.0 . .0 9.0
L.
—I---------I—----—-I——---I I----
21.0 24.0 3.C 6.0 9 . .. . . . ..
-------
Table 5_16a* presents the behavior of the coliforms
during interval 2 for low flow and the 1974 land use condi-
tions. Relatively high concentrations prevail compared to
the coliform concentration in the river system during the
first rain interval. The final concentration at the dam
(junction 1), however, is in the same order of magnitude.
Due to the reduced base flow, the time of delay between peak
flow and highest concentration increases to at least one day.
The slow movement of the coliforxn concentration down the
river can be easily recognized. The coliform concentrations
for the medium base flow conditions during the same rainfall
are different from the concentration under low flow conditions
only in so far that they are approximately 1-1/2 times lower
than the concentration during low flow. That is consistent
with the same input of constant as well as time dependent
coliform load at all the junctions, while the flow for the
medium condition is approximately 1-1/2 times as high as for
the low flow. The travel times in the low flow case are a
little slower and therefore slightly change the path con-
figuration of the coliforin concentration in the system.
Clearly, a much higher flow and a longer period of precipita-
tion during interval 1 induceg a very different coliform
distribution in the river system. But again, the final
concentration at the dam is in the same order of magnitude,
which is consistent with the fact that the total input of
coliform is in the same order of magnitude for both amounts
of runoff. The trend of suspended solids (Tables 5-15 and
l7a) during intervals 1 and 2 is similar to the trend of
coliforins during those intervals; there are much higher
concentrations in the system during the immediate time of
precipitation in interval 2 and for some time afterwards,
but final concentrations at the dam are in the same order of
magnitude; again the rate of increase to the peak concentra-
tion at the dam is much slower in the second interval due
to reduced trave1 times.
After we compared the impact of the 1974 land use
configuration on the water body during different rain
intervals, we turned to the comparison of the impacts due
to different land use patterns. The first condition we
considered was the intensive development in Shepard Brook
as described in Appendix B, Table B-6. The intensive
development is characterized by addition of residential,
commercial and industrial subdivisions to the 1974 develop-
ment of the sub-basing. For rain interval 1, the peak of
the stage graph at the dam (Fig. 510) is slightly higher
* In order to make possible comparisons between 1974 and
1985 land uses, not all of the junctions are presented.
5 .p-73
-------
Table 5-16
Coliforms (10 MPN/100 ml. at Junctions of Mill River System
Rain Interval 2 - Low Base Flow
U’
A. Land Use 1974
Runoff occurred frcin 19:00 - 21:00 hours,
B. Land Use 1985
•on
Time
1
3
6
8
11-
12
1
3
6
8
11
12
18
.08
.03
.00
.002
.00
.O0
.08
.03
.002
.002
.002
.003
19
.08
.03
68.0
59.0
47.0
350.0
.08
.03
94.0
86.0
81.0
180.0
20
.08
9.2
92.0
55.0
210.0
49.0
.J7
8.7
130.0
120.0
180.0
49.0
21
.36
25.0
62.0
17.0
34.0
49.0
.37
32.0
99.0
69.0
33.0
49.0
22
1.0
34.0
53.0
30.0
2.6
49.0
1.4
47.0
100.0
57.0
2.6
49.0
23
2.0
390
25.0
110.0
7.7
.003
3.0
56.0
76.0
89.0
7.1
.003
24
3.3
43.0
27.0
18.0
2.6
.003
5.1
63.0
66.0
18.0
2.3
.003
2
5.4
42.0
49.0
3.8
.21
.003
8.5
70.0
62.0
3.6
.2
.003
6
7.0
41.0
17.0
.8
.004
.003
10.0
69.0
38.0
.78
.02
.003
14
8.8
32.0
1.7
.002
.002
.003
14.0
53.0
.33
.002
.002
.002
24
9.8
25.0
.002
.002
.002
.003
16.0
41.0
.01
.002
.002
.003
-------
Table 5-17
Suspended Solids (mg/U Rain Interval 2 -
Low Base Flow at Junctions of Mill River System
A. Land Use 1974
B. Land Use 1985
Junc-
ion
Time
1.
.
3
6
8
11
12
1
3
6
8
11
12
18
13.0
13.1
14.2
14.2
15.4
25.7
13.0
13.1
14.2
14.2
15.4
25.7
19
13 0
13.1
33.3
37.2
41.3
190.7
13.0
13.1
37.8
42.7
48.6
109.4
20
13.O
15.6
77.8
75.8
158.5
125.6
13.0
15.5
90.5
88.2
139.9
125.6
21
13.1
23.].
59.0
37.5
88.9
125.6
13.1
24 9
70.5
55 0
83.4
125.6
22
13.3
32.7
66.9
50.9
10.8
125.6
13.5
38 5
78.1
58 5
9.1
125 6
23
13.9
36.5
44.0
87.1
23.9
27.7
14.4
43 5
59 8
71.6
21 6
25 7
24
14.7
40.0
47.2
51.8
18.4
25.7
15.5
48.3
58 7
50.5
17 7
25.7
2
16.1
43.7
60.1
17.4
16.0
25.7
17.5
53.1
60.6
16.3
15.9
25 7
6
17.35
44.7
30.8
13.9
15.1
25.7
18.4
53.6
44.1
14.4
15.2
25.7
14
18.6
40.5
14.7
14.2
15.3
25.7
20.9
45.3
14.4
14.1
15.4
25.7
24
19.7
33.1
14.2
14.2
15.3
25.7
21.9
38.1
14.2
14.2
15.4
25.7
(1
I
Runoff occurred from 19:00 —
21:00 hours.
-------
I - ..
H
CD
II
P 1
I- .
H
H
1 5j
H
0 U)
H CD ITJ
H
r G
(DPI
‘ ,1
CD
H
C M
(DPI I
r1 H
0
CD
CD
I
3 .4O ,) —
**
-J
0
I
*
** _peak for
1974
I
•I
I
*
*
*
*
*
I yv1
.
——
*
I
I
*
*
*
*
I
.
*
*
I
.
*
*
I -
31.200 —
*
*
‘1
*
*
*
I
I
*.
*
*
- I
P!Pt .
-
*
*
*
in
feet-
*
*
‘p
:
I
*
7.J jo —
*
.
-
I
*
*
—________________
*
*
I
I
*
*
*
*
I
*
*
I
*
*
I
I
*
*
*
I
*
36.JJO —
*
*
!
*
*
I
*
*
I
I
*
*
**
*
I
I
***.
*
**u
*
I
***
***
*
**
**
I _____________ * _______________ **
3b.bUu I I 1———
0.0 3.u 6.0 9.J 12.0 - 15.0 ld.O 21.0 24.0
TI’4E U, I-ChUMS
-------
than in the 1974 configuration, while the shape is identical.
The concentration of coliforms and of suspended solids for
this scenario are slightly increased compared to the 1974
land use, but the increase is in a range which is not of
appreciable significance for the public health official or
the concerned public. Due to the minor changes, the coinpari—
son of the land use scenarios was limited to the rain interval
1.
The next land use scenario we investigated is based on
the land use plan of 1985 for Hamden. Using this plan, we havE
specified input parameters to the STORM model as accurately
as possible given the quality of information. Population
increase over today’s figures is estimated at approximately
13,000, distributed in the upper part of the basin. The
population in the area of Lake Whitney was kept constant.
Commercial and industrial areas were added as far as they
could be recognized from the 1985 land use map. Figure 5-7b
shows the stage graph of rain interval 1 at the dam. Its
shape is clearly comparable to the base case (Figure 5-7a),
but the peak is increased by approxitnate1y 1 foot of water
level. Increased travel times due to the higher flow from
the increased runoff causes a slight contraction of the
graph as can be seen, for instance, at the 37 O feet mark.
Comparing the stage graphs of rain interval 2 for the low
flow case yields similar results (Figure 5—9b). The stage
graph shows the same shape but both peaks are slightly
iigher, in a range of .15 feet. Again the faIling crest
of the 1985 scenario is a littlebit steeper than that of
the 1974 graph. Table 5-18 shows the suspended solids
concentration, and Table 5-19 the coliform concentration for
the 1985 land use scenario during rainfall interval 1. The
base flow conditions were kept the same as above. Significant
differences can be extracted by comparing the Tables 5-14
and 5—15 (reflecting the 1974 land use condition) with the
‘ ab1es of the 1985 land use plan. The 1985 concentrations
in the system are approximately 1-1/2 times as high for the
coliform, while only approximately 30 percent higher for
the suspended solids. The peak of concentration is approxi-
mutely at the same time, about 4 to 6 p,m., which is eight
to ten hours after the rain has finished. Similar obser-
vations can be made for rain interval 2 as indicated by
comparison of Tables 5-16a and b and 5-].7a and b for coliform
and suspended solids concent atJ . n, ) espectively. !1 he peak
concentration of suspended solid. at the dam increases only
slightly for the 1985 development pattern, from 19.7 to
21.9 mgL, while the highest colitorm concentration increases
from 9,800 to 16,000 MPN/100 ntg/ . - The peak bnóentration
itself does not seem as important as the trend revealed.
mt , even though the concentration increase is not relevant
in terms of public health considerations for the Mill River,
5—77
-------
Table 5—18
Suspended Solids (ag/I) - Rain Interval 1 — Righ Flow
Land Use 1985
%% Junc-
1
2
3
4
5
6
7
8
9
10
11
12
13
8
11.8
13.0
8.9
8.6
8.6
9.1
9.1
9.1
9.2
9.2
8.5
10
11.7
16.9
8.9
8.9
12.5
12.3
11.2
12.2
12.2
11.1
12.1
8.9
7.9
12
11.6
22.3
9.5
15.0
15.2
15.7
14.9
14.0
14.2
16.7
13.7
13.3
27.4 -
14
11.4
63.1
11.8
19.9
35.9
33.9
25.9
31.7
27.2
19.5
22.0.
9.7
16
11.9
102.5
25.2
32.5
44.2
48.5
35.0
37.2
33.3
25.4
23.4
26.3
18
17.3
26.8
39.0
34.6
26.4
31.5
20.2
17.1
10.4
8.9
9.7
27.0
5.7
26.5
20
21.4
25.1
33.2
29.0
19.0
17.3
9.5
8.7
9.0
6.3
5.6
8.9
3.4
22
22.3
25.1
30.2
16.3
11.9
9.8
8.9
6.9
7.4
9.9
7.5
8.9
10.2
24
22.6
25.1
26.8
9.7
8.9
8.0
7.6
9.7
8.6
9.7
7.8
8.9
9.6
4
22.8
25.1
20.5
8.2
8.3
9.2
8.9
9.4
8.9
9.5
8.1
8.9
8
22.5
25.1
16.4
8.6
8.6
9.2
9.0
9.3
9.0
9.3
8.3
8.9
8.8
8.4
I
RunoftoccuZt.d frca lOzOO to 18:00 hours.
-------
U I
I
%0
Table 5—19
Coliform (lOs MPN/l00 ml) — Rain Interval 1 — High Flow
Land Use 1985
tion
1
.06
2
.08
3
.007
4
.oo:
5
.001
6
.001
7
.001
13.0
8
.001
18.0
9
.001
22.0
10
.001
11.0
11
.001
20.0
12
. .001
25.0
13
.001
110.0
10
.06
27.0
.006
.1.6
23.0
22.0
28.0
11.0
17.0
12
14,
.07
58
57.0
70.0
3.6
15.0
35.0
39.0
36.0
41.0
37.0
41.0
300
34.0
32.0
190
28.0
41.0
10.0
28.0
30.0
94
16.0
51
16.0
23
11.0
16
12.0
16
16
18
4.0
12.0
170
4.3
320
22 0
230
21 0
14.0
8.5
5 5
3.2
.32
.89
35
39
03
56
05
05
07
001
20
14.0
4 1
3.9 0
7 7
5 0
3 6
—_03
001
.001
001
22.
14.0
4 1
15.0
3 3
1 6
78
29
03
.001
24
14.0
4.1
12.0
83
47
22
01
001
001
001
001
001
001
.00
.001
001
001
4
3.4.0
13.0
4.1
4 1
8 2
5.4
4
00
02
.001
.01
.001
001
001
001
001
001
.001
001
001
001
ff occurs frc 10:00 to 18:00 hours.
-------
an increase of this size might be relevant in another river
basin and result in potential danger to the public health.
These runs have demonstrated that, given similar base
conditions in a river system, the impact of different devel-
opment scenarios can be identified. Existing qualitative
knowledge can be quantified with this combination of models.
How useful are the results of these runs? Results of
STORM, reported in Appendix B, seem to be in the same order
of magnitude as results reported in the literature. SWMM
has been tested by its developers and its performance has
been found satisfactory. Clearly, significant problems of
calibration and verification occur when such a chain of models
is to run for many different situations. Are our results
useful despite the lack of detailed calibration? The con-
centration figures computed by SWMM at the dam site seem to
be in the right order of magnitude and to reveal the right
tendency, if we assume that it is possible to compare SWMM’s
hourly results to daily samples from the New Haven Water
Company. The samples of the water company show that during
periods of heavy precipitation the samples of coliform are
in he range of the results generated by SWMM. It is impos-
siba.e to be more specific in the interpretation because we
do ot know at what time of the day the samples were taken
and to what degree the distribution of precipitation in
Hamden is reflected in the precipitation record of Bridgeport
airport. Cross—series analysis of the Water Company’s coli-
form and water-stage records (monitored at the dam) yield a
peak of the cross—variance for a time lag of one day and a
peak of roughly .55 days in the cross—spectrum analysis.*
The small discrepancy is caused by the different degree of
discretization in the cross—series analysis package. But
clearly, a lag effect is existent; and these lag effects are
also indicated in our simulation of runoff and washoff.
During interval 1 the peak of coliform concentration lags
behind the peak of the water stage by about 1/2 day. The
concentration decreases then quite slowly over the period of
12 to 24 hours after the stage peak. During the rain
interval 2 the lag increased to at least one day, confirming
the above results. Disagreement between calculations and
sampling could arise because samples are taken in the middle
of the rain interval when the stage is already high but the
coliforms still low.
In summary, the linkage of STORM and SWMM is of great
* That means that this frequency contains the most common
variance of the two series.
5-80
-------
value to our land use water quality analysis (Table 5-20
summarizes the steps of the analysis). Results permit
identification of the impact of different land use patterns
on the magnitude of pollution from urban storm water runoff.
In the case of the Mill River it revealed that a quite
large increase of pollutant emissions has to be expected
within ten years. When standards are violated, the costs
of alternative methods of control and their impact have to
be evaluated (see section 6).
Summary
The computer models developed in this section provide
local and regional planners with a flexible tool for evalu-
ating the effects of land use patterns on the utilization of
wastewatërrelated infrastructure (see the sewer and treat—
nient plant capacity evaluation module), the magnitude of
waterborne emissions from urban runoff, and the resulting
water quality of receiving streams in response to runoff,
washoff, and erosion. The models can be employed to analyze
both the impacts of individual developments and of zoning
policies applied to an entire sub-basin. In addition, the
impacts of alternative land uses and emission controls can
béassessed with this module.
In our introduction to this report, we stressed the
necessity of considering the interaction between environ-
mental, social, and economic impacts in the land use plan-
ing process. Within this framework, the environmental
j mpact and capacity evaluation models such as the ones
developed here perform several important functions:
1. 208-areawide planning calls for an evaluation of
the environitiental impact of alternative physical and land
use controls for water quality management. Areal sources
as well as discharges from treatment plants have to be
considered. Previously, facility planning under Section
201 of Public Law 92-500 had been concerned primarily with
treatment plants and major interceptors. Their utilization
rate can be estimated by the capacity evaluation module.
Demonstration runs with the model combination of STORI4-SWMM
have shown its applicability for investigating land use
control alternatives and pollution from urban runoff.
Discr 1nat10n between quality impacts from different land
use configurations is possible. Thus the model enables
planners to consider and compare relevant controls.
2. Local land use decision-making takes place in a
context where certain groups show strong concern for the
0 vironment impact and for the potential overutilization
of infrastructure. By making it possible to predict this
5—81
-------
Steps Wet Flow
Level of Aggregation Define receiving waterbodies and sub-basins (idealize in
junctions and channels).
Land Use Analysis for Specified Year Define number of land use catories per sub-basin and land
use controls.
Rate of imperviousness; runoff coefficients; rate of dust and
dirt accumulation; select pollutants of interest; dust and
dirt breakdown into pollutants; pollutant emission (specify
seasonal or yearly average parameter).
Select rain invervals of interest; run STORM on all sub-basins
for the specified simulation period; calculate total runoff
and total load for simulation period; compute load and runoff
of specified intervals; save on disk.
Calibration of STORM In case of runoff gauging and washoff monitoring, calibrate
(and verify) quantity and quality modules.
Universal Soil Loss Function Use for erosion and sediment load, if desired.
Estimating Receiving Waterbody Parameter Hydraulic, hydrologic and quality related parameters in SWMM;
Parameters select integration step.
Calibration,’ Vrification Exercise Use steady-state flow; point source input (see calibration
exercise as described), if specifically distributed data are
available.
Route Quantity and Quality in Waterbody Select runoff event(s) of interest; select pollutants; select
background hydrograph and pollutograph; route; print stagetime
relationship and pollutant change at specified junction.
Adjust Calibration In case of available high-flow data adjust quality related
parameters; multiple runs of quality module.
Rerun and Analyze Make all desired runs; compare results; evaluate frequencies
of ter quality conditions of interest.
Table 5-20
Suninary of Analysis
Land Use Parameter Estimate per Land
Use Cagegory (urban; non—urban)
Generation of Input to Waterbody
(subject to storage specifications)
-------
impact to a certain extent, these models will serve to
facilitate local decisionmaking.
3. By indicating the impact of areawide as well as
local development given for existing pollution control
standards, the model will enable planners to determine if
these impacts are undesirable by existing standards.
Policies can then be designed to alleviate any undesirable
impacts which may be revealed.
4. The output of the capacity evaluation module can
be used as input to the fiscal input model.
5. The models developed in this section presently
consider only a restricted set of land uses and land char-
acteristics, and employ functions which may need consider-
able refinement and re-evaluation. Nevertheless, we feel
that the models, if used properly within their limitations
will give results which are valuable for many planning
purposes.
5—83
-------
Section 6
Land Use/Water Quality: Fiscal Impact
Introduction
Our concern in this section is the direct fiscal impact
of pollution control and, in particular, of water pollution
control. These impacts include the costs of waste transport
and treatment, the additional costs of employing alternative
activities or processes, and the manner in which these costs
are allocated to different governmental levels and socio-
economic classes. We are not concerned here with indirect
economic impacts, either positive or negative, resulting.
from pollution control activities, such as the creation of
jobs to build treatment plants or the loss of jobs resulting
from closing an antiquated, polluting plant. These, of
course, can be significant in some instances. Nor are we
concerned with economic measures of the damage (or benefit)
resulting from various ambient pollutant concentrations,
although there has been considerable interest in obtaining
such measures, particularly for air pollutants.*
Our discussion here follows the patterns set in our
treatment of generation and emission models in that we
separate the analysis into two categories: wastewater
and storrnWater controls. This approach also follows the
conventional engineering practice of separating sanitary
sewerage from stormwater. This practice may change in the
future because it is becoming increasingly evident that
control of stormwater will be required to meet water quality
goals. An attractive alternative tO accomplish this is to
revert to combined wastewater and stormwater treatment.
However, other options also are ava ilable and as yet no
definite commitment to any of these alternatives has been
made on a national scale. Thus in our review, here we do
not discuss combined sewerage systems or stormwater treat-
ment cost (other than for simple storage for sediment and
runoff control).
*See for example, T. Waddell, “The Economi.c Damages of Ai
pollution,” Environmental. Protection’. Agency) EPA 6OO/S—74-Oi ,
)974; also D. P. Tihanaky, Economic ’Damage$ to Household
items from Water SupplyUSe,’ Environmental Protection’
Agency, EPA—600/S73001, 1973.
6—1
-------
A Brief Review of Existing Models
There are three basic models or modeling approaches
which are covered in this section:
1. wastewater treatment;
2. collection and transport of wastewater at the devel-
opment and interceptor level; and
3. collection and transport of stormwater at the
development and interceptor level.
Models for Wastewater Treatment Costs
A substantial literature exists on the estimation of
capital and operating and maintenance costs for wastewater
treatment plants. We therefore will cover this topic only
briefly here. Smith* has reviewed and sununarized much of
the early literature in this field. Subsequent work for
the EPA** has further extended the literature on treatment
plant costs and updated cost estimates to account for the
requirements of the 1972 amendments to the Water Pollution
Control Act and to EPA’S interpretations of these amendments.
Techniques employed in these studies have included costing
of hypothetical treatment plant designs as well as empirical
analysis of actual plants built under various federal and
local programs.
Traditionally, the costs of wastewater treatment have
been modeled as simple power functions of a single variable,
usually total flow, population equivalents served, or BOD
load. This formulation is deficient in two major respects.
First, costs are affected both by hydraulic and organic
loadings and thus a cost estimating procedure based upon
only one of these variables may be misleading. Shah and
Reid*** have devised cost estimating equations from a re-
gression analysis of a sample of actual treatment plant
* R. Smith, “Cost of Conventional and Advanced Treatment of
Wastewater,” J. Water Pollution Control Federation , V. 40,
Septeither, 1968, pp. 1546—74.
** For example, Black and Veatch Consulting Engineers, “Es-
timating Cost and Manpower Requirements for Conventional Waste—
water Treatment Facilities,” Environmental Protection Agency,
1971.
D. L. Shah and W. Reid, “Techniques for Estimating Con-
struction Costs of Waste Treatment Plants,” J. Water Pollution
Control Federation , V. 42, May 1970, pp. 776-93.
6—2
-------
projects. Cost was modeled as a power function of both
organic loading and total flow. Both variables were found
to be significant in the regression.
A second drawback of the conventional formulation lb
that it implies a constant elasticity of cost with respect
to plant size over all capacities. Most of the data which
have been accumulated are for relatively small, plants (1-10
mgd) which exhibit significant economies of scale. Thus
extrapolation of these economies to larger size plants might
be inappropriate and could lead to excessively large region-
al plants and over—expansion of facilities. In fact, more
recent evidence* indicates that the economies of scale in
larger plants are not as great as in the smaller facilities.
Another aspect is advanced wastewater treatment, asso-
ciated with higher treatment requirements. The costs of
two treatment plants should be compared only when the same
treatment levels are required. Technology significantly
influences the total capital and operating costs. For
example, certain cheap treatment technologies such as lagoons
are available only for small plants, not for large plants.**
Modeling Collection Costs
While a considerable amount of effort has been spent
on obtaining accurate generalized cost estimation procedures
for wastewater treatment processes, much less attention has
been given to developing suitable cost estimation procedures
for wastewater collection. This relative neglect is some-
what surprising in light of the fact that collection costs
account for 50 to 70 percent of the capital costs of .a
wastewater disposal system.
In a model of wastewater collection costs for both
new development areas and the agc.ompanying interceptor.
systems, there are a considerable number of: factors which.
must be taken into account since they. markedly affect costs.
To set the background for our discussion, these factors are
discussed briefly below under the.categories f engineering
standards and natural factors.
* R. Smith and R. G. Eilers, “The Economics of Consolidating
sewage Treatment Plants by Means of ‘Intercepting’ Sewers and
Porce Mains,” EnvironmentaL Protection. Agency, 197.1..
** “An Identification of the Municipal Choice of Options for
eeting the Effluent Limitations and Goals Specified in Public
Law 92-500,” in preparation by Meta Systems Inc for the Na-
tional Commission on Water Quality, Contract No. WQ5ACO47, 1975.
6—3
-------
As in any engineering project, the costs of a sanitary
sewer system are heavily influenced by prevalent engineer-
ing design standards. Such standards include factors of
safety and implicitly reflect the best judgments of the
profession regarding the tradeoffs between initial capital,
operating and maintenance, and replacement costs and between
cost and reliability or other measures of performance.
Fortunately, from the standpoint of devising generally
applicable cost estimates, there appears to be widespread
agreement on many of the important standards for sanitary
sewer design.* Indeed, the design of sanitary sewer systems
has become so standardized that Newville was moved to
comment:
The basic parameters for sanitary sewer design
were set at the turn of the century and, for
the most part have remained unquestioned since
that time. Sewerage collection systems today
are designed almost by rote, picking values off
charts and conforming to standards which were
in existence before the present generation of
engineers were born.**
Although many aspects of sanitary sewer design are well
established, in practice some design standards are subject
to state and local variation. These standards include
minimum pipe size, infiltration allowance, per capita flow,
ratio of peak to average flow, and minimal depth of place-
ment.
Besides engineering standards, several natural factors
also have a bearing on sewer designs. They are as follows:
* This does not imply that these standards are necessarily
rational or that they are not likely to change in the f u-
ture, however.
** Jack Newville, New Engineering Concepts in Community
Development , Urban Land Institute Technical Bulletin 59,
Washington, D. C., 1967.
For a complete discussion of how these standards in-
fluence costs, see “Data Collection and Review and Analysis
of Infrastructure Cost Relationships,” Working Paper No. 3,
prepared for the Environmental Protection Agency by Meta
Systems Inc under Contract No. 68-01-2622, February 15, 1975.
6—4
-------
1. Soil Type . The soil type determines the ease of
excavation and therefore is a major factor in determining
costs. For example, according to the Dod e Guide,* the unit
cost of excavation in clay is about 1.7 times that of com-
mon earth, and rock excavation costs are five times that of
common earth. In areas with extensive bedrock near the
surface, it may be infeasible to install traditional sewer-
age systems at typical suburban densities.
2. Depth to Water Table . A high water table increases
excavation and installation costs by adding requirements for
dewatering and sheeting of the excavation. A variety of
methods are available for this purpose, and the selection
of a particular technique depends upon the severity of the
problem. In extreme cases, the additional costs can be as
great as those imposed by the excavation of rock. ** The
depth to water table also can affect the infiltration allow-
ances employed in sewer design.
3. To ography . Because of the high cost associated
with trenching, particularly the sheeting and bracing cost,
the most economical sewer system designs are generally
those which minimize the total amount of excavation. Thus
the most economical design would be obtained where all col-
lectors could be placed at minimum depth, that is, parallel
to the surface slope. Unfortunately, there are upper and
lower limits on the slope conditions under which these
ideal conditions could be met. These limitations are de-
rived from the minimum and maximum velocity requirements
which have been adopted as part of engineering design
practice.
Development Level Costs . Taking into consideration
these factors, we have reviewed a number of studies of
wasteWater collection costs at the development level, and
summarize below their approaches and major results. The
studies reviewed were conducted by the Urban Land Institute
* 1975 Dodge Guide for Estimating Public Works Construction
cosEs , New York: McGraw Hill, 1974.
** R. S. Howe, “Planning Sewerage Service for New Towns,”
Ph.D. Thesis, University of Wisconsin, 1971.
6—5
-------
(ULI),* Downing,** Dajani and Gemmell (DG),*** the Real
Estate Research Corporation (RERC),t and HOwe.tt A detailed
comparison of the design methodologies employed in these
studies is presented in Table 6—l.ttt
All these studies represent synthetic costing efforts.
That is, the authors laid out what they felt were repre-
sentative land use patterns and assembled the costs for
sewering these patterns from the unit costs of pipes, ex-
cavation, and appurtenances.
All of the studies began with a basic geographic unit
of analysis. In the case of the ULI study, the basic unit
was a 1000-acre community. Downing, DG, and Howe employed
a 160-acre (1/4 square mile) subdivision as their basic
unit while RERC defined a basic unit in terms of a constant
number of dwellingunits (1000). Each study then proceeded
to design and estimate sewer system costs for different
densities and housing types on their basic unit. Here,
however, the paths of the studies diverged. The ULI and
RERC studies paid a great deal of attention to the layout
of the system. The RERC study assumed modern curvilinear
* Urban Land Institute, “The Effects of Large Lot Size on
Residential Development,” Technical Bulletin 32, Washington,
D.C., 1958.
** P. B. Downing, The Economics of Urban Sewage Disposal ,
New York: Praeger, 1969.
** J. S. Dajani and R. S. Gemmêll, “Economics of Waste—
water Collection Networks,” University of Illinois Water
Resources Center, Research Report No. 43, June, 1971.
t Real Estate Research Corporation, The Costs of Sprawl:
Detailed Cost Analysis , prepared for the Council on Environ-
mental Quality, Department of Housing and Urban Development,
and the Environmental Protection Agency, Washington, D.C.,
U.S. Government Printing Office, April, 1974.
tl R. S. Howe, 22 • cit .
t ft For purposes of comparison we have already introduced
our approach in this table., even though. it is not described
until Section 6, page 16.
6—6
-------
able 6—1
Sumaxy of Sewerage Studies
Urban Land
Institute
Downing
Dajani and
Gemmell
Howe
0
-.4
1)
S
- .4
14
U
4)
(3
S
II
S
0
14
0
41
V
z
Layouts
Basic Unit
Size -
1000 acres
160 acres
160 acres
160 acres
1000 dwelling
Units
160—200 acres
Gross Densi—
ties (PPA)
3.4,6.512.2
0.4,1,4,16,64,
128,256,512
10,25,50,100,
150,250,750,
1000
3,12,15,30,75
90,120,150,
300,450,600,
750,900,1200
7.0,8.75,11.0,
16.5,28
10.5,26.47,87.5
Meta
1’
-3
Rectangular
lots
Frontage/depti
Frontage
64/100
Block sises
vary with lot
size.
a) 8—20 acre
blocks
b) 16—10 acre
blocks
Five lateral.
one sub uain
per cell
Slope
Flat
Single family
conventional
Single family
clustered
Town house
Low rise apart-
ments
High rise
apartments
Planned mix
Flat
Conventional
single family
medium density
planned unit
development
High density PUI
High rise
apartments
Flat
Soil
Conditions
II
U
- .4
4,
S
-.4
I I
V
4 )
U
S
14
S
U
z
-4
U)
Average
Flat
a) Flat
b) Minimum
for flow
at 2 ft/
sec.
TM sasy Trench- Cameon
j gW Earth
Flat
moderate (2—12%)
steep (> 12%)
Co n earth
Common earth
Sand
Grave].
Silt
Clay
Shale
Bedrock
High Water
Table
Loam
Sand
Loose gravel
Compacted grave:
and till
Hard ciay and
shales
-------
Table 6-1 (continued)
Urban Land
Institute
Base
Downing
Boston,
Aug. 1955
Dajani and
Gessnell
1957—1959
Howe
(1957—1959)?
41.
S
a
41
S
0
U
1970
RERC
1973
Meta
Manhole
Costs
Pip. Costs
8” $8/ft
12”—15”
$12/ft
Prom Greely &
Eanaon 5
$/ft=l.403
1.49902 +
.0 19X 2
D - Dia.
X - Depth
+
Dodge Estilnat—
ing Guide**
Calculated
pipe and ex-
cavation
costs sepa—
rately
“
.5”
30 ”
$8.22/ft
$14.52/ft
$26.69/ft
.
Dodge Estimat-
ing GLide**
National aver-
age, aid 1975
Included in
pipe costs
$230
Per Capita
Flow (GPcD )
Not included
100
75.
$500
$500 for sew-
ers < 36”
$1500 for sew-
ers > 36”
100
Dodge Estimat-
ing Guide**
100
S
4 1
S
a
S
V
a
100 +
100
Peak Flow,
Average
Flow
—
3.0
5
1
-------
Table 6-1 (continued)
220’
300’
350’
* Greely and Hansen Engineers, Madison, Wisconsin Sewer District Report o sewerage
and Sewage Treatmej , Chicago, 1961:.
** Dodge E tiaating Guide for Public Works Cpnstruction - 1970 Annual Edition No. 2 .
New York, McGraw Hill Inforaation Systen CoMpany, 1970.
*** Cäà ittee of the Great Lakes - Upper Mississippi River Board of State Sanitary
Engineers, Reco nded Standards for Sewage Works - 1968 Edition , Albany, Esaith
Ebeetion Service, 1968.
Urban Land
Institute
Doiminq
Dajani and
Ge el1
Howe
REIC
Meta
C
0
I
I
t
Manning N
—
.012
.013
.013
—
.013
Miniau
Velocity
(tt/sec)
—
2
2
2
—
2
Maxiaua
Velocity
(ftfáec)
—
10
10
10
10
Slop.
aequirensnt$
—
—
Ten states
standardsa**
—
Metcalf and
Eddy’
Piaping
NO
No
No
When depth
exceeded 30’
—
When depth
exceeds 20’
Manhole
Spacing
300’
+ Met.onl.f and Eddy, Wastewater Engineering , New York: McGraw Hill, 1972.
-------
Street patterns and clustering of units, where appropriate.
The ULI study, which was done in 1957, employed a rectangu-
lar grid system but adjusted street spacing and lot geometry
with density in a realistic manner. Dajani and Gernmell and
Howe, on the other hand, ignored issues of layout and em-
ployed fixed collection grids. Dajani and Gemmell calcula-
ted costs for two sets of grids, one based on 10-acre
rectangular blocks and a second using 20-acre blocks.
Although there were significant cost differentials between
the two sets of designs, the authors did not discuss the
circumstances under which one or the other of the patterns
would be appropriate. Downing’s presentation does not
allow any conclusions about his treatment of the layout
issue except that he, too, used a rectangular Street grid.
In the actual design and cost estimation of the sewer
network the situation is reversed. The ULI and RERC
studies paid no attention to the actual system design.
They merely estimated the total length of pipe in the system
and applied average coefficients of cost per foot to calcu-
late total costs. Both Dajani and Gernmell and Howe pro-
duced detailed hydraulic designs for each sewer system
considered and used these designs as a basis for cost
estimation. Downing’s analysis was somewhere between these
two approaches in degree of detail.
Most of the studies were concerned with systems where
trenching conditions are favorable and topography is
either flat or gently sloping. Only Howe considered a
variety of soil conditions in his analysis. He also
employed two different types of topography, flat and
favorable (i.e., slope which requires the minimum of ex-
cavation). However, Howe’s layout probably was the least
satisfactory of all the designs.
Figure 6-1 compares data from all five studies which
we reviewed. We have attempted to make these data as
comparable as possible by adjusting all costs to December,
1973 and adjusting the data, where necessary, to add or
exclude cost components so that all data refer to a common
set of assumptions. We have excluded engineering fees,
legal fees, contingency costs and house connection costs.
All data refer to common soil conditions and flat terrain.
In one case, the Dajani and Gemmell study, the original
cost data were not available. Instead we used costs gen-
erated by equations fit to their data. These costs have
been estimated for densities corresponding to the original
networks employed in the study. Also, the Dajani and
6—10
-------
o HOWE
£ S DOWNING
o JANI AND GEMMELL- 10 ACRE
A “ II N -2OACRE
a
LULl
A,BC,D,EF RERC
—— LEAST SQUARES FIT 1-3OPPA
0
A
I I I I I 1111111
10 100
POPULATION DENSITY (PERSONS/ACRE)
WASTEWATER COLLECTION COSTS- SUBDIVISIONS
I 1i I I
; ‘ I liii
p ..
N
.
N
C
N
O N
0 0
- ENR CC INDEXZ 1942 (DEC. 1973)
- COMMON SOIL CONDITIONS, FLAT E
- TERRAIN; COSTS DO NOT INCLUDE
ENGINEERING, LEGAL FEES OR
CONTINGENCIES
- HOUSE CONNECTIONS NOT INCLUDED
• IT
1r
•‘
.
00
A
I
0
S
0
0
0
S
Figure 6-1
-------
Gemmell study did not include manhole costs. We have
added 15 percent to their costs to account for this item.*
In absolute terms the tJLI cost estimate data appear
to be the highest and the RERC data the lowest over their
applicable ranges of densities. In the case of the ULI
data the pipe costs used were specific to the Boston area.
One explanation for the high estimates would be a systema-
tic difference in excavation costs for this area which are
not incorporated by the cost index employed. The RERC
data agree very well with the Howe, Downing, and Dajani
and Gemmell 20-acre block data for its single family-
conventional (A) category, but drop off more rapidly than
the other data as density increases. As discussed above,
this effect is due to the fact that more efficient layouts
are employed for the higher density land uses in the RERC
study.
We employed data in the range of 1 to 30 persons/acre
to develop a general equation for estimating subdivision
collection costs. The mathematical model employed was the
simple power relationship
c = adb (6-1)
where c = capital cost per capita, in dollars, and
d = density, in persons per acre.
This model assumes that the size of the area serviced
does not influence costs. This assumption is strictly
appropriate only within a narrow range of subdivision
sizes. Thus, extrapolation to subdivisions significantly
larger than those in the studies reviewed should be avoided.
The parameter b can be interpreted as the percent change in
costs which will result from a one percent increase in
density.
The parameters a and b were estimated by applying
least squares to the logarithmic version of the equation,
giving
C = 1050 d 827 , 1 < d < 30. (6—2)
This equation may be employed to predict collection costs
for small subdivisions (100-200 acres)-with relatively
* J. Baffa, “Lateral Sewer Construction Costs,” Public
Works , 86, No. 11, November, 1955, pp. 71—77.
6—12
-------
favorable trenching conditions and level topography. It
also may be possible to employ the Howe data as a basis for
adjusting this equation to handle other soil conditions.
In summary, of the five studies reviewed, none appears
to be completely adequate to serve either as a sole source
of data or a prototype for a more general model. Two of
the studies (ULI and RERC) were conducted by planners and
emphasize the development layout but treat the engineering
and the cost estimations of collection networks rather
casually. Two other studies, conducted by engineers (Dajani
and Gemmell, Howe), use development layouts which are un-
realistic for current suburban subdivisions but take great
care with system design. The fifth study, by an economist
(Downing), falls between the two extremes. Only one study,
Howe’s, considers a realistic variety of soil conditions,
but this work has one of the least representative layouts.
All the studies focus on the development (subdivision)
level of costs.
Interceptor Level Costs . Interceptors link new develop-
ment with the existing sewer network. Their costs, requiring
a different set of calculations, have traditionally been com-
puted on the basis of relationships such as: dollar/mgd/mi.
= f (average flow at ultimate capacity). Various cost func-
tions such as Spencer’s are common.* The question of ultimate
capacity has to be determined by the planner by investigating
various projections of additional development in the same
area. The appropriateness of cost functions such as Spen-
cer’s, which is a simple 1958-based power function updated
by the ENR Index, might be questioned if it happens that
not only inflation but also changes of practice influence
the actual cost. But the function seems to be satisfactory
as a first approximation.
Modeling Stormwater Collection Costs
Storinwater collection costs are much more difficult to
characterize than those for wastewater. Unlike the latter
case, there are few generally accepted design rules for.
handling stormwater design problems. Ardis, Denker, and
Lenz have documented the divergence of design practices in
* C. C. Spencer, “Metropolitan Planning for Sewers ona
County Basis,” Public Works , 89 (8), August, 1958, p. 83:
Cost/mi. = 46,000 (mgd) 45 ; the equation is based on an
ENR-constructiOn index of about 1,000.
6—13
-------
their study of 32 cities in Wisconsin.* Despite the
geographically limited data base, the authors found great
variation in procedures employed by local public works
departments. For example, consider the frequency of the
design storm. The design storm is the storm which the
system would just barely handle. Thus a five-year design
storm implies that, on the average, once in five years
gutters will overflow, storm sewers will back up and
localized flooding will occur in basements and other
structures. Obviously the selected frequency will have an
important effect on overall system costs. Since loss of
life is generally not an issue in the type of minor flooding
we consider here, the frequency selected represents a
balance between cost and the potential property damage and
inconvenience which would be caused by minor flooding. The
frequencies employed by the 32 cities reporting in the
study ranged from one to 25 years.
In another part of the study, respondents were asked
to design a stormwater collection system for a hypothetical
15-acre area located in their jurisdiction. The results
of this exercise again demonstrated the great variability
in design procedures. Although all designs were for an
identical tract of land, the most expensive design differed
from the least expensive by a factor of about six.
Two of the cost studies described in the previous
section, the ULI study and the RERC study, also considered
storm sewer costs. However, these studies used “average”
coefficients of cost per foot in developing their estimates
and made no effort to indicate the effect on these costs
of soil, climate, topography or design standards. In fact,
they did not even fully describe the conditions under which
the estimates would be applicable. As a result, these
studies are not useful models for estimating stormwater
collection costs.
Rawls and Knapp** have made an empirical study of
stormwater collection costs. Their sample included designs
f or tracts of land ranging from 11 to 1,485 acres. Based
* C. V. Ardis, J. Denker, and A. T. Lenz, “Storm Drainage
Practices of 32 Cities,” J. Hydraulics Division , American
Society of Civil Engineers (ASCE), Vol. 95, January, 1969,
pp. 383—408.
** W. J. Rawls and J. W. Knapp, “Methods for Predicting
Urban Drainage Costs,” J. Hydraulics Division , ASCE, Vol. 38,
September, 1972, pp. 1575—85.
6—14
-------
on a sample of 70 projects they developed the following
equation for estimating stormwater costs in 1963 national
average prices:
CT= 58,273.0 + 8.73(F SG R DB Q AD
(6—3)
where:
CT = Project Cost, in 1963 dollars,
F = Recurrence Interval, in years,
SG = Average Ground Slope, in feet per 100 feet,
R = Runoff Coefficient, C from rational method,
DB = Smallest Pipe Size, in inches,
Q = Total Capacity, in cfs,
AD = Developed Area, in acres.
From a planning standpoint it would have been better
if the total capacity had not been included in the estima-
tion, since it would then have been possible to estimate
COStS from engineering practices and natural features
without resorting to detailed calculations.
Another factor to consider in stormwater management is
the use of on-site storage or infiltration systems. Tradi-
tional approaches to stormwater management have emphasized
removal of the water from the site as rapidly as possible.
This approach tends to make matters progressively worse
downstream since it reduces the times of concentration,
resulting in higher downstream peak flows. Recently, more
emphasis has been given to on-site storage as a method of
reducing peak flows. For many areas a combination of
storage and smaller storm sewers will be economically
superior to the sole use of large storm sewers. A variety
of methods for providing the necessary storage currently
are being tried. These include roof storage, parking lot
* The R 2 is equal to 0.785.
6—15
-------
storage, and permanent or temporary pondirig. The latter
method appears to be favored in suburban residential areas.
Since the trend toward on—site storage is relatively
recent, there has appeared no systematic cost analysis
suitable for planners. For suburban residential areas,
some guidance may be obtained from Maryland’s experience
with sedimentation basins. These basins are similar to
storage basins except they are designed to be temporary
facilities for use during construction (many silt basins
are, in fact, converted for stormwater storage after con-
struction is completed). The APWA* reports that sedimenta-
tion ponds in Maryland cost about $50-70 per house lot.
The APWA also reports data for an Illinois development
where detention ponds cost about $lOO-300 per residential
lot for lot sizes up to 1/2 acre.
Cost Evaluation Module: Models Developed and Selected
Introduction
This section deals with the cost impact module derived
from the foregoing review of the water-related system
models. When treating development and cost impact, we arrive
automatically at “intertemporal” aspects. Limited resources
have not permitted a truly intertemporal model, so we offer
a comparative model instead. We assume that the planner
establishes a base year for data -- say 1974. This data
base is modified according to projections the planner
wishes to examine. The model provides estimates of impacts,
costs, cost breakdowns, and thereby establishes incidences
for comparative analysis of the base and the projected
year. Decisions are predicated on a series of such com-
parisons.
There exists a difference between comparative physical
and economic analyses. Physical analysis simulates total
residuals and water quality impacts for events drawn from
the base and projected years. Economic/financial analysis
estimates additional costs (and their breakdown) due to
changes between the base and projected year.
* 1 merican Public Works Association (APWA), “Practices in
Detention of Urban Stormwater Runoff,” Special Report No.
43, 1974.
6—16
-------
The financial/economic module presented operates
only on single communities, but it should also be capable
of analyzing parts of the community. It concentrates on
residential development in the urban fringe area. The
following costs are considered:
1. Sanitary sewer laterals in the development and
house connections;
2. Drainage system for the development (including
temporary and permanent retention basins);
3. Interceptors to link the new development with
the existing sanitary sewer system;
4. Construction of relief sanitary sewers in case
of inadequate capacity of the existing system;
5. Expansion of the wastewater treatment plant,
required when the additional flow and load over-
load the treatment plant capacity;
6. Adjustment for additional runoff of man-made
and/or natural stormwater drainage system; and
7. Treatment of additional runoff to improve runoff
quality.
Based on these cost categories, the following eight
cost types j can be estimated by the cost evaluation module
for each proposed development:
j = 1 on-site disposal
j = 2 sanitary sewer laterals
j = 3 sanitary sewer building connections
j = 4 sanitary sewer mains/trunks
j = 5 storm sewer laterals
j = 6 stormwater detention ponds
j = 7 storm sewer mains/trunks
j = 8 sewage treatment plant
6—17
-------
Each cost type j is divided into a cost function for
capital costs (CCj) and operating and maintenance costs
(OMJ).
In the following sections the cost functions employed
will be described. Capital costs for sanitary sewer
laterals have been estimated in the course of the project,
and their development will first be outlined in some detail.
Cost functions related to cost types j = 1, 4, 5, 6, 7 and
8 are described later in the section.
Sanitary Sewer System Cost Estimates
Up to the present time there has been no generally
adequate, simple model available for estimating sanitary
sewer system costs at the level of detail required by the
planner. Such a model, which would be applicable for a
variety of common soil types, slope conditions, and develop-
ment patterns, has been developed in two major stages.
First, ten hypothetical subdivisions and associated lateral
sewer systems were designed or adapted from available
sources. Sewer costs were then estimated and synthesized
into a nwnber of simple equations for producing generalized
cost estimates.
Review of Costs Factors . Four key factors were taken
into account in the analysis: (1) the types of housing
and subdivision design; (2) physical design of the sewer
system; (3) soil types; and (4) topography. Each factor
is discussed below.
( 1) Housing Types and Subdivision Design . Six basic
types of housing units were included in the analysis. These
were selected to be representative of the types of housing
found in new developments in urban fringe areas. A short
description of each of these types follows:
A. Single—family conventional -- traditional single
family homes in moderate sized lots which vary in size
from 1/4 to 1/3 acre.
B. Single—family compact —- single family homes on
small lots of about 1/8 acre.
C. Townhouses -- attached two-story units in con-
figurations of four to eight units pe structure. Each
townhouse unit has a separate sewer connection.
D. Garden apartments -- three-story walk—up apartment
buildings with an average of 30 units per building.
6—18
-------
E. Medium—rise apartments —- five—story elevator
apartment buildings with an average of 100 units per
structure.
F. High-rise apartments —- 20-story apartment build-
ings, 500-900 dwelling units per building.
Various combinations of these six types of housing
were employed to develop a set of ten subdivision layouts
for detailed sewer design and cost evaluation. For each
subdivision, designs were evaluated for three different
soil categories: (1) loose sand, loam and gravel; (2) com-
pacted gravel and till; and (3) hard clay and shale. Thus
a total of 30 cost evaluations were completed. The soil
categories were based upon the categories of the 1975 Dodge
Guide,* which was the basis for all cost estimates.
Table 6—2 summarizes the ten subdivision designs
employed in our study. Each design is characterized by its
mix of dwelling unit types and average slope conditions.
The basis for the three slope classifications (flat, moder-
ate, steep) will be covered fully in a later subsection on
physical design.
It has been our intention, in selecting layout patterns
for detailed analysis, to adequately reflect the diversity
of the suburban housing market. Our designs therefore
include subdivisions consisting entirely of conventional
single-family homes at one extreme and entirely of 20-story
apartments at the other, while the majority of cases include
various mixes of dwelling units. In all cases layouts have
been selected to be representative of current practice in
the industry. The subdivisions ranged from 160 to 205 acres
in size, this scale being selected for compatibility with
our 160—acre grid employed as a data base. This size also
was judged to be large enough to allow for the evaluation
of complete developments while at the same time small enough
to allow for a reasonable number of designs to be examined
with our limited resources.
A variety of sources were employed to select appropriate
layouts. Design 1 is an actual development constructed in
the town of Clay, New York, with some minor modifications.
Designs 5 and 7 are based on design studies by the New York
* . cit .
6—19
-------
T
0
Table 6-2
Summary of Subdivision Designs
Design
Area
(acres)
Slope
Number of Dwelling Units
Gross
Density
DU/acre
Single Single Garden Medium-Rise High-Rise
Family Family Town Apartments Apartments Apartments
Conventional Compact Houses (3—story) (5-story) (20-story)
1
2
3
4
5
6
7
8
9
10
180
160
160
160
205
160
160
205
160
160
Flat
Moderate
Steep
Flat
Moderate
Steep
Flat
Moderate
Moderate
Flat
540 ——— ——— ——— ——— ———
471 —__ -——
471 ——— ———
359 ——— 652 207 ——— ———
——— 553 619 396 ——— ———
367 ——— 590 300 ——— ———
——— 300 276 540 1040 ———
——— 359 417 696 1290 ———
——— 227 233 576 1120 ———
——— ——— ——— ——— ——— 4000
3.0
2.9
2.9
7.6
7.6
7.8
13.5
13.5
13.5
25.0
-------
City Planning Department.* The remainder of the designs
were developed by our own staff and represent extrapolations
of patterns found in Massachusetts, New York and Connecticut.
A representative layout (design 1) is depicted in Figure
6—2 with a ten-acre grid overlaid to indicate the scale.
( 2) Physical Design . The procedures employed in the
design and cost estimation of the sanitary sewer systems
were much the same as those which would be used by a design
engineer in the preliminary analysis of such systems. We
began with a tentative network configuration, including
locations of pipes and manholes. Detailed hydraulic design
was then undertaken for each link in the network. The end
result of this procedure was a detailed specification for
each link, including upstream and downstream invert eleva-
tions, peak flow and velocity, and total amount of excava-
tion and sheeting required. These specifications then formed
the basis for calculating total system costs.
No excess capacity was deliberately designed into the
system, although some pipes in the system do have such ca-
pacity as a result of standards on minimum pipe size and
other restrictions. This approach implies that there is no
allowance for the “linking” of subdivisions. We assume
that a series of interceptors and mains will be provided
for this purpose. Such “higher order” collectors are gen-
erally provided by the local or county government and the
costs for these must be added to the collection costs
computed for one subdivision. Also, our approach assumes
no significant change in land use. At one time it was
common for engineers to design collection systems for the
highest density which they thought might develop in the life
of the system, usually 50 to 100 years. With the implemen-
tation of effective residential zoning policies, engineers
recognized that such conservative design procedures no
longer were required. Today the capacities of most sewer
systems in suburban areas are based upon the existing zoning
policies.
The influence of the various engineering standards on
system costs have been discussed previously.** Here, we
* New York City Planning Department, Planned Unit Develop-
ment , May, 1968.
** “Data Collection and Review and Analysis of Infrastructure
Cost Relationships,” Working Paper No. 3, prepared for the
Environmental Protection Agency by Meta Systems Inc under
contract No. 68—01—2622, February 15, 1975.
6-21
-------
Figure 6—2 Design 1
Each grid cell — 10 acres.
6—22
-------
shall only summarize the standards employed in this design
study. The standards selected were based upon recomznenda—
tions in the American Society of Civil Engineers/Water
Pollution Control Federation joint committee manual,* and
textbooks by Metcalf and Eddy** and Babbitt,*** unless
otherwise specified.
1) Number of persons per dwelling unit = 3,5 (from
Costs of Sprawlt assumption for single family homes).
2) Flow per capita 100 gallons per day.tt
Q
3) Ratio of peak to average flow,
___ 5.0
___ — p 0 . 2 (6—4)
where p = population served, in thousands.
4) Equation for flow -- Manning’s Equation
= 1.43 R 2 3 S° 2 (6—5)
* Design and Construction of Sanitary and Storm Sewers,
. Ajnerican Society of Civil Engineers and the Water Pollution
Control Federation, New York, 1970.
** Metcalf and Eddy, . cit .
H. E. Babbitt and E. R. Baumann, Sewerage and Sewage
Treatment , 8th edition, New York: John Wiley and Sons,
T 38.
Real Estate Research Corporation, . cit .
j In the past, the Environmental Protection Agency has
accepted 10Q g/ca p/d as a standard design value (see
“Interceptor Sewers and Suburban Sprawl,” prepared ior
Council on Environmental Quality by Urban Systems Research
and Engineering, Inc., Cambridge, Massachusetts, September,
1974, under contract EQ4ACO27); this value is currently
being reconsidered.
6—23
-------
n = coefficient of roughness = .013 by assumption
R = hydraulic radius, in feet
S = slope in ft/ft.
5) Minimum velocity for pipe flowing full 2 ft/sec.
Maximum velocity for pipe flowing full 10 ft/sec.
6) Minimum slopes :
diameter (inches) slope (ft/ft)
8 .004
10 .003
12 .0022
15 .0015
18 .0012
21 .001
24 .0009
>27 .0008
7) Minimum pipe diameter : street laterals = 8”;
House connectors = 6’s.
8) Minimum depth to invert = 5 feet.
9) Manhole spacing . Maximum spacing = 350’, on
curved sections wherever pipes join at an angle >2°.
10) Sheeting, bracing and trenchi . Trench designs
were developed to conform to the xequirements of the Federal
Occupational Health and Safety Act. This legislation has
had a significant impact on sewer costs, since it requires
a greater use of sheeting and bracing than had been common
previously. Design varied with soil types, which were
divided into the general categories cohesive and non-co-
hesive, as specified below:
6—24
-------
max. trench depth cohesive non—cohesive
0-6’ vertical trench sides sloped at
2:1 ratio
6-10’ vertical trench; full sheeting and
skeleton bracing bracing
>10’ vertical trench; full sheeting and
full sheeting and bracing
bracing
Minimum trench widths were 3’.
11) Pipe materials . Vitrified clay pipes were
employed in all sections of the system.
(3) Soil Types. All subdivisions are designed for three
soil classifications. These classifications were taken from
the Dodge Guide,* which established the categories as a
basis for cost analysis. The three categories were: loose
sand, loam, and gravel; compacted gravel and till; and hard
clay and shales. The first two were considered to be non—
cohesive, the third cohesive for trench design per (10)
above.
( 4) Topo raphy . There is a relatively broad range of
slopes over which pipes can be placed at minimum depth. As
a result, we have not considered a large number of slope
conditions in this study. Rather, we have dealt with three
categories of topography. Our flat topography category
includes land where average slopes are too small to allow
sewer lines to be laid at minimum depth. These include
slopes in the range 0-0.5 percent. The moderate slope
category includes slopes which allow most of the collectors
to be placed at minimum depth and is therefore the most
favorable condition. This category includes slopes from
0.5 to 11 percent. Finally, the steep slope category
includes slopes which are too severe to allow placement of
lines at minimum depth and includes slopes from 11 to 16
percent. The divisions between slope categories should not
be considered as exact. They represent our best estimates
based Ofl the experience gained in evaluating our ten project
designs, but some fuzziness is inevitable in any such
classification scheme. Depending on particular conditions,
* cit .
6—25
-------
the range of favorable slopes could be somewhat greater or
smaller. however, we feel that in most instances the
classification of a particular area should be straight-
forward.
All the designs which were evaluated had either uniform
or only slightly varying slopes over the extent of the de-
velopment. As a result, our cost estimates are not strictly
applicable to cases where the terrain is highly irregular.
We hypothesize that our estimates for the flat or steep
slope categories (the two sets of estimates are quite close)
could be employed as a first approximation to estimate the
costs for irregular topography, which is also an unfavorable
condition from the standpoint of sanitary sewer costs.
Costs for the systems based on the combinations of
layout and engineering designs, soil and topography were
synthesized from unit costs for pipe sections, connectors,
manhole sections and covers, excavation, fill and sheeting.
All unit costs were obtained from the Dodge Guide* and are
expressed in national average prices, mid-1975. A detailed
presentation of these cost estimates, including the propor-
tions of total cost allocated to each major component (exca-
vation, pipe, etc) is found in Appendix F (Lateral Sewer
Cost Estimates) to this report. In the paragraphs below we
discuss our efforts to synthesize these results into a set
of cost estimation equations for both connections and later-
al sewer lines.
Sewer Connection Cost Estimates (CC 3 ) . The setback,
or distance from the dwelling unit struc€ure to the curb
is a primary factor in determining the overall cost of
house connections. The setback, in turn, is generally
determined by the minimum standards set by a local community
through its subdivision ordinances or building codes. Since
the minimum requirement can vary widely from community to
community we have decided to develop cost estimation tech-
niques separately for the house connections and the re-
mainder of the collection system.
The general form of the equation for estimating the
costs of house connections is:
C.. =a.. +b.. x (6-6)
ijk ijk ijk
* E • Cit .
6—26
-------
where:
Cik = total cost per connection, in dollars
a. constant term
ijk
b, = coefficient which reflects effect of
uk distance with cost
x = setback distance, in feet
i = index of dwelling unit type
j = index of slope type
k = index of soil type
Thus to determine the total connection costs for an
entire subdivision, one calculates the individual con-
nection costs for each dwelling unit type in the sub-
division and multiply each such cost figure by the appro-
priate number of connections. Note that the number of
connections are identical with the number of dwelling units
only in the cases of single family homes and townhouses.
For apartments, we assume that each building has one con-
nection, thus the number of connections equals the number
of dwelling units divided by the average number of dwelling
units per building.
It was possible to reduce the number of equations
required by aggregating several of the housing type cate-
gories. The type of connection employed for conventional
single family homes, single family compact, and townhouses
is identical. Thus one set of coefficients can be applied
to all three housing unit types. Similar comments apply
to the garden apartment and medium-rise apartment categories.
Also we have found that costs in the “loose sand, loam and
gravel” and “compacted gravel and till” categories were
virtually identical. As a result, we have replaced them
with a single category, which we term “loose soils.” “Hard
clay and shales” remains a separate category and will be
abbreviated as “clay” soils.
The coefficients ajk and bjjk were computed by taking
average costs for connections of each type for each sub-
division where such connections were employed, then averaging
over the individual subdivision estimates. The constant
6—27
-------
coefficient a k represents the cost of the COnnection to
the lateral plus any associated special piping and elbow
sections. The coefficient bjjk was calculated by dividing
the remaining connection costs by the setback. Table 6-3
presents the coefficients for equation (6—10) for each
appropriate combination of dwelling unit type, soil con-
dition, and slope.
The slopes given in Table 6—3 apply only to situations
where the sewer connections are installed as part of the
original development, in conjunction with the system of
laterals and mains. This method produces considerable
economies, since connecting sections (t and y sections) can
be installed as part of the original network, re—excavation
of the laterals can be avoided, and it is not necessary to
break through and replace sidewalks and road surfaces. If
the connections are made after development, experience
indicates that they will be considerably more costly and
the data in Table 6-3 would not be applicable.
Lateral Sewer Costs Estimates (CC2 Table 6—4 sum-
marizes the results of our cost estimation procedures for
the designs listed in Table 6-2. The costs presented in
Table 6—4 include all collection system costs except
building connections, which have been dealt with in the
previous section. We shall refer to these estimates as
“net costs.” As in the case of house connections, there
was virtually no difference between the two soil categories,
“loose sand, loam and gravel,” and “compacted gravel and
till.” Again, these have been combined into a single
“loose soils” category in all analysis described below.
We experimented with a number of functional forms
in attempting to develop an equation for predicting net
costs as a function of the dwelling unit mix, soil type,
and average slope for a development. The form which we
finally settled upon is:
5
NET COST = a b .N., (6-7)
k9. i=l ki 1
where:
ak 9 = correction factor for slope type Q in soil
type k,
bk. = cost per unit of dwelling type i in
1 soil k,
6—28
-------
Dwelling Unit
single family
conventional,
single family
compact
and townhouse
garden apartments
and medium rise
apartments
high rise
apartments
Table 6—3
House Connection Cost Coefficient
a..
ijk
Slope
Flat Moderate
b..
ijk
_____ Slope
Steep Flat Moderate
52.66 3.45
Soil Type
loose
clay
73.73
73.73
37.59
37.59
52.66
3.15
3.45
3.15
Steep
3.45
3.15
loose
82.34
37.59
67.09
4.13
4.13
4.13
clay
82.34
37.59
67.09
3.83
3.83
3.83
loose
568.0*
437 90*
568.0*
7.19
7.19
7.19
clay
568.0*
437.90*
568.0*
6.80
6.80
6.80
* High rise apartments assumed to be connected at manholes.
-------
Table 6-4
Summary of Collection System Capital Costs*
* Costs do not include house connections, engineering and
legal fees, or contingencies. Costs are in mid 1975
dollars, national average.
Design
1
Slope
Flat
Loose Sand,
Loam and
Gravel
Compacted
Gravel
and Till
Hard Clay
and
Shales —
701,500
970,000
969,200
2
Moderate
205,700
206,600
165,600
3
Steep
467,900
469,100
371,900
4
Flat
963,200
964,300
658,800
5
Moderate
334,400
335,500
232,900
6
Steep
988,300
989,800
812,200
7
Flat
836,100
837,100
610,100
8
Moderate
302,800
303,700
213,000
9
Moderate
281,800
282,500
209,300
6—30
-------
N 1 = number of units of conventional single family
homes,
N 2 number of units of clustered single family
homes,
N 3 = number of units of townhouses,
N 4 = number of garden apartment buildings,
N 5 = number of medium-rise apartment buildings,
= 1—flat; 2-moderate; 3-steep,
k = 1-clay; 2-loose.
We omit the sixth dwelling unit type, high-rise apartments,
because it was only used in one design and therefore was
not included in our parameter estimation procedures.
The mathematical model assumes that the effect on
costs of an additional dwelling unit is linearly additive.
That is, a dwelling unit of type i will add o k bkj to the
total system costs regardless of the number and type of
other units in the subdivision and the relative location
of different types of units. A little reflection quickly
reveals that this assumption is not strictly valid. For
example, an apartment complex located at the upper reaches
of a collection network might require large size pipes in
the downstream links in the subdivision, while the same
complex located at the end of the system will only affect
the size of the final link. The total costs in the two
cases will generally not be the same. Thus the spatial
allocation of dwelling unit types within a development does
affect system costs, contrary to the model assumption.
However, we did not have enough data to include such
detailed effects in the model. In addition, in many situ-
ations the planner would not have information available on
the detailed location of different housing types when he is
to make his evaluation, Consequently, the model developed
here is intended to reflect average costs from typical
layout patterns.
An additional note of caution is warranted with respect
to subdivision size. As Table 6—2 indicates, all of our
designs have been for 160—205-acre developments, Sewerage
costs increase nonlinearly with area: for any given den-
sity, the larger the area serviced the greater will be the
per unit costs. Thus our model will tend to overpredict
6—31
-------
costs for developments significantly smaller than those
considered in this study and underpredict costs for signi—
ficantly larger developments.
Since equation (6-7) has seven parameters and the
sample includes nine designs, the parameters could be
evaluated directly by applying nonlinear estimation pro-
cedures using a criterion such as minimization of the sum
of squared differences between actual and predicted project
costs. However, because of the small sample size we have
chosen to employ a more robust, heuristic approach in our
initial analyses. This procedure was based upon the cri-
terion of minimizing the sum of absolute deviations between
observed and predicted costs.
The resulting equations were:
Hard Clays and Shale :
Cost = (352N 1 + 2l6N 2 + lOiN 3 + 3620N 4 + 6656N 5 )
(6—8)
a 11 = 2,8
a 12 = 1.0
a 13 = 2.7 R 2 = 0.82
Loose Soil :
Cost = a 2 , (437N 1 + 259N 2 + l9lN 3 + 5190N 4 + 8290N 5 )
(6—9)
a 21 = 2.8
a 22 = 1.0
a 23 = 2.7 R 2 = 0.85
The value indicates the proportion of the variance
which is explained by the model. Since we minimized abso-
lute rather than squared deviations the parameter estimates
did not optimize upon the value of R 2 . Nevertheless, the
value of R 2 is useful in judging the explanatory power of
6—32
-------
the model. In both cases it signifies that substantial
amounts of unexplained variation remain unaccounted for.,
This is due to the variety of different layouts employed
in our design studies.
Another way of evaluating the model results is to
check the consistency of the parameter estimates with
commonly accepted rules of thumb. For example, rough
estimates of sewerage costs often are based on lot frontage
because for small pipe sizes, the most significant costs
are associated with total system length (pipe length,
sheeting and excavation, number of manholes) rather than
capacity. By this criterion the relative costs of sewering
single family conventional, single family clustered and
townhouses should be approximately in the ratios 80/50/25
respectively, which are their average frontage, in feet.
Simplified, these ratios become 1/0.625/0.3125. From
equation (6—17), the associated ratios are 1/0.61/0.29
which agree closely with the hypothesized relation. Equa-
tion (6-9) gives ratios of 1.0/0.59/0.43; the townhouse
costs are somewhat more expensive relative to single family
conventional units than for equation (6—8).
As with the case of house connections, the costs cited
here apply only where sewers are installed with the origi-
nal components which must be added to establish a realistic
cost estimate. These components are generally listed as
“engineering, legal fees and contingencies” in consultants’
estimates and account for approximately a 25% additional
charge to the base estimate. Finally, in areas with ex-
tensive bedrock formations close to the surface or a high
water table, costs would be significantly greater than
predicted by equations (6-8) or (6-9).
Stormwater Laterals (CC 5 )
Because of the lack of unified design standards for
stormwater management, it was not possible to develop a
useful model for estimating these costs along the lines of
our sanitary sewer Cost evaluation. For comparative pur-
poses, we have evaluated the costs for designing a storm-
water collection system for subdivisions 2, 5, and 8 of
Table 6-2, employing the 25 year design storm recommended
by the engineering department of the Town of Hamden, Con-
necticut.* The resulting costs were:
* Drainage Manual, Town of Hamden, Connecticut , Hamden
Engineering Department, July, 1971.
6—33
-------
Subdivision Loose Clay
2 $529,000 $556,000
5 $954,000 $897,000
8 $863,240 $797,000
Thus the costs of the stormwater collection system is about
2.5 - 3 times that of the sanitary sewers. However, the 25
year storm is an unusually high design standard, and no
general conclusions could be made on the basis of the re-
sults. In our model we have employed the equation estimated
by Rawls and Knapp (see above) to estimate stormwater col-
lection cost.
Other Cost Functions
The capital cost functions for cost types j = 2, 3, 5
were discussed in preceding sections. The others will be
input as follows:
On—site disposal:
CC 1 : C = 125. + .15 (V—500) + CLF* (6—10)
where:
CC 1 = delivered tank cost
V = net volume of the tank in gallons
CLF** cost of associated leaching field.
V is calculated from:***
* “On-site Household Wastewater Treatment Alternatives --
Laboratory and Field Studies,” R. J. Otis, N. 3. Hutzler,
W. C. Boyle, in Water Pollution Control in Low Density
Areas, edited by W. 3. Jewell and R. Swan.
** Because of regional variations in labor and material
costs, this variable which describes mainly the excavation
of trenches to be filled with gravel, has to be estimated
and input by the planner on a case by case basis.
Manual of Septic Tank Practice, U.S. Department of
Health, Education and Welfare, Public Health Service, Publi-
cation No. 526, Reprinted 1969.
6—34
-------
V = 1.5 Q for Q < 1500 gallons wastewater per day
= 1125. ÷ .75 Q for Q > 1500 gallons wastewater
per day
(6—li)
an installation cost of $275 is assumed.
OM: assumes one pumping every three years at a cost
of $30 per pumping or $l0/year.*
Sanitary sewer iaterals:**
0M 2 : CH = .2362*** (6—12)
where:
CH = annual per household cost
= feet of sewer per capita and,
= 54X .65 (6—13)
where X = population density in persons per acre.
Sanitary sewer building connections:
0M 3 are considered negligible in this context.
Sanitary sewer mains/trunks:
* Ibid .
** CC 2 and CC3, as previously noted, do not take into
account the effect of industrial or commercial establish-
ments on the collection systems costs. Similarly, ON 2 also
does not.
*** P. M. Meier, 3. Kiihner and J. C. Martell, “A Prelim-
inary Assessment of Wet Systems for Residential Refuse
Collection,” Curran Associates, Inc. (for the Environ-
mental Protection Agency), July, 1973, Contract No.
68—03—0183, p. 203 (NITS—PB 234436).
6—35
-------
CC 4 : n C 4 = .07167 + (1.04284 — .006114Z) 9,n D
+ .06147Z* (6—14)
C 4 = construction cost, pipe in place, in $/ft.
based on an ENR Construction Cost Index of
1975.
z = average depth of trench, in feet
D = diameter of sewers, in inches
0M 4 : average number of repairs per mile of sewer is
assumed to be 2.2 per year with an average cost
of $162 per repair;** only the costs of the new
system are considered.
Storm sewer laterals:
0M 5 : Same as 0M 2 .
Stormwater detention ponds:
CC 6 : $200 - $600 per residential acre***
0M 6 : no realistic estimates available.
Storm sewer mains/trunks:
CC 7 : Same as CC 4
0M 7 : Same as 0M 4
* H. A. Thomas, M. Shapiro, J. Houghton, “Paretian Analysis
of Regional Systems for Sewage Disposal,” Discussion Paper
74-2, August, 1974, Environmental Systems Program, p. 17.
** P. M. Meier et a l. , 2 • cit., p. 203.
American Public Works Association, “Practices in
Detention of U.S. Stormwater Runoff,” Special Report No.
43, 1974.
6—36
-------
Sewage treatment plant (new plant and/or capacity expan—
sion) .*
b b
cc 8 = a 1 x 1 1 x 2 2 [ 1 + f(ip)] (6-’ 15)
b
OM = a x (6—16)
8 23
where:
x 1 = design population equivalent (see below)
X 2 = design flow, mgd
X 3 = actual flow, mgd
= design population (in persons)
f = ancillary works factor (dimensionless).
Values for the parameters are given in Tables 6-5, 6—6, and
6-7. The population equivalent, X 1 , is obtained using the
following expression:**
8.33X C
x l = 0.17 (6—17)
where:
8.33 = conversion constant (8.33 lbs/million gallons
per mg/94
0.17 = lbs of five-day biochemical oxygen demand (BOD)
per capita per day
C = BOD 5 of wastewater in mg/2 .
The detailed breakdown of the treatment plant into
subprocess and the availability of the corresponding
* For detailed description of variables refer to: Meta
Systems mc, “Evaluation of Alternative Methods for Finan-
cing Municipal Waste Treatment Works,” Socioeconomic En-
vironmental Studies Series, Environmental Protection Agency,
600/5—75—001, February, 1975.
** Meta Systems mc, Ibid .
6—37
-------
Table 6-5
Parameters of Capital Cost Functions
Biological Treatment
a b
Type 1 2
Activated sludge .00812 .461 .262
Filtration .122 0 .656
Sludge Handling:
Sludge Pump .0125 0 .480
Sludge Digester .0575 0 .650
Sludge Holding Tank .0224 0 .590
Vacuum Filtration .326 0 .560
Incineration .0150 0 .560
Physical-Chemical Treatment
Coagulation and
Sedimentation .067 0 .890
Filtration .122 0 .656
Carbon Adsorption
X 2 < 10 mgd .546 0 .613
x 2 > 10 mgd .293 0 .983
Chlorination .0202 0 .664
Sludge Handling:
(as above)
SOURCE: Meta Systems mc, Ibid .
6—38
-------
Table 6-6
Parameters of O&M Cost Functions
Biological Treatment
Type
Activated Sludge .042 .876
Filtration .022 .650
Sludge Handling:
Sludge Pump .0021 .452
Sludge Digester .0095 .712
Sludge Holding Tank .0015 .530
Vacuum Filtration .063 .706
Incineration .0089 .570
Physical-Chemical Treatment
Coagulation and
Sedimentation .0147 .986
Filtration .0136 .638
Carbon Adsorption
X 3 < 10 mgd .1058 .483
x 3 > 10 mgd .0502 .808
Chlorination .0043 .905
Sludge Handling:
(as above:
SOURCE: Meta Systems mc, Ibid .
6—39
-------
Table 6-7
Ancillary Works Factor
Dependent Upon opu1ation
Size
500
999
2,499
4,999
9,999
24,999
49,999
99,999
249,999
‘P > 250,000
0
500
999
2,499
4,999
9,999
24,999
49 999
99,999
f = 0.373
0.384
0.498
0.722
0.790
1.060
1.487
1.533
1.763
2.473
SOURCE: Meta Systems mc, Ibid .
but >
6—40
-------
parameters to fit the above type of equation for each
subprocess, allows the planner to use this equation for
estimating costs of a new plant as well as of a capacity
expansion of a plant.
The parameters of the cost functions are taken from
the literature and are either standard engineering esti-
mates* or results of statistical analyses of survey—
obtained data. **
Details of the Cost Evaluation Module
The cost evaluation module estimates the costs incurred
due to added development within the community and provides
an allocation of costs to those groups sharing the project’s
costs. By varying the cost allocation schemes, the planner
is able to generate a range of financial impacts, over time
on the community. Thus, the module complements the sewer
capacity evaluation module by analyzing the financial impacts
of community development. A general logic diagram of the
module is presented in Figure 6-3.
Regardless of the cost types j to be estimated, a set
of community and development characteristics is required
as input:***
a. Projected population/population equivalents
of the community for six land uses, for four
time periods -- the present (time T) and 10,
25 and 50 years into the future:
(1) single family —- low density;
(2) single family -- high density;
* K. Shah and G. Reid, “Techniques for Estimating Construc-
tion Costs of Waste Treatment Plants,” J. Water Pollution
Control Federation 42 , No. 5, Part I, May, 1970.
** Black and Veatch, “Estimating Costs and Manpower Re-
quirements for Conventional Wastewater-Treatment Facilities,”
for Environmental Protection Agency, October 1971, Project
#17090 DAN.
*** It should be noted that some of the data is deliber-
ately in formats compatible to the sewer capacity evaluation
module.
6—41
-------
Figure 6—3
Logic of Cost Evaluation Module
‘ iSITE 2 SANITARY SANITARY SANITARY STORM 6bTORM ATER STORM B 1 SEWAGE
DISPOSAL SEWER I ISEWER I I SEWER SEWER ETENTIO I ISEWER I
_________ jLATERALSJ SUILO NG IMAINS/TRUNNSI 1 LATLR LS 1 rONDS i MAINSIT 1IPC3 PLANT
COST ICONNECT IONSI I
•TYPEJ
6—42
-------
(3) multi—family;
(4) commercial;
(5) industrial;
(6) open space -- recreational;
b. the expected gallons per capita per day of
wastewater generated, by land use;
c. present and projected assessed property values
of the community, by land use for the four time
periods T, T + 10, T + 25, T + 50;
d. projected assessed property values of the proposed
development, by land use for the four time periods
T, T + 10, T + 25, T + 50.
e. the numbers and kinds of residential structures
to beconstructedin the proposed development;
f. expected number of persons per household for the
various residential dwellings within the proposed
development;
g. interest and discount rates to be used in calcu-
lating the cost streams.
In addition to this data, certain development charac-
teristics will be required by the individual cost types j
to satisfy the cost functions. The details of such input
should be included in a program documentation.
For each cost type j there will be a maximum of up to
four groups 2 sharing the costs:
= 1 developer
£ = 2 local government
= 3 state government
= 4 federal government
Having defined the cost types to be studied, the plan-
ner decides on a cost allocation scheme based upon local
practices, and state and federal cost sharing programs.
Thus, for each cost type j input fectors will be required
6—43
-------
indicating the percent share of the capital costs (k=l) and
the operation and maintenance costs (k=2) to be allocated
to each group 2.
From this point a k,Q the share of cost in dollars of
cost component k, for cost type j, to be allocated to group
9.,, is calculated. The assumptions made concerning the cost
calculations and allocations are:
1. all operation and maintenance costs are in average
annual cost except for sewage treatment plants
(j=8), for which annual operation and maintenance
costs vary in accordance with the capacity utili-
zation of the plant and so are calculated on
yearly basis until full capacity is reached.
2. any financing of capital costs by the federal or
state governments will be realized in one payment
at the beginning of the construction period,
3. any financing of operation and maintenance costs
from the federal or state government will be
realized as an annual fixed percentage of the
costs,
4. within the local government costs may be further
broken down by the method in which funds are
raised to finance the project:
a. special assessment -- a one-time charge is
assessed against the property owners of the
development in dollars per $1000 assessed
value*
b. bond issue -- the community may decide to
float a bond issue in order to finance the
cost type. The revenues required to then pay
off the bond issue will be raised in two ways:
(1) property taxes -- the additional tax per
$1000 assessed property values that will
be paid by each land use over the bond
issue payback period.
(2) user charge —— given the total amount of
revenues to be raised by user charges,
the equivalent dollars per capita and
* Note: assessed values are obtained from straight line
interpolation between time periods.
6—44
-------
dollars per 100 gallons wastewater
generated are computed by land use.
Required input for the bond issue
mechanism includes the payback period,
the percentage of the bond issue to be
financed by property taxes and user
charges, the percentage of each to be
paid by the six land uses and, for the
capital costs financed with this mechan-
ism, the annual rate of increase in the
amount paid back each year.*
Table 6-8 lists the mechanisms which may be
employed to finance the individual cost
categories.
5. costs borne by the developer will generally be
passed on to the consumers within the development
in the same fashion as a special assessment by the
local government. **
Table 6—9 represents a sample of the output which can
be generated by the module. This is for a residential
development which requires sanitary and stormwater lateral
and interceptor sewers. Capital and operating and mainte-
nance costs for each infrastructure component are listed
separately and allocated among different financing methods.
The program can also produce more detailed tables indicating
the temporal allocation of costs, effects upon property
taxes, and the required user charges of each cost type
among the various land uses. After examination of the costs
the planner may choose to rerun the program with a different
local government financing mechanism.
Summary
The fiscal impact module developed in this study
provides local and regional planners with an inexpensive
and flexible tool for evaluating the economic impacts of
* If the annual rate of increase in the amount to be paid
back is zero, the annual payback is constant.
** If the housing market is highly competitive, or develop-
ers in nearby locations do not have to pay as much, the
developer may absorb part of the costs to remair ’ competitive.
6—45
-------
Table 6—8
Financing Mechanisms for Different Cost Categories
Cost On-site Sani- Sani- Sani- Storm Storm- Storm Sewage
Categories disposal tary tary tary sewer water sewer treat-
___________ — sewer sewer sewer later— deten— trunks! ment
later— building trunks! als tion mains plant
als con— mains ponds
nections
Mechanisms Cap OM ap OM Cap OM Cap OM Cap OM Cap OM Cap OM Cap OM
9’ Special
assessment x x x x X X X X X X
0•
Bond Issue
property
tax X X X X X X X X X X X X
user
charge x x x x x x x x x x x x
-------
RESIDENTIAL DEVELOPMENT — TEST DATA — .iuiv 29. t91)
Table 6—9
— — -- . - SL MARY TA8LE i _______________
TOTAL COSTS (.) ____ ____________________
- LOCAl. GOVERNMENT
- _ . . . . . SPECIAL -— .. _._PAOPERTV ._. SlATE EDE RAL_
COST TVPt S..) DEVELOPER A SESSMENf USER CHA E TAX GOVERNMENT GOVERNMENT
SANiTARY SEWER lATERALS
_CAPI .TAL. .. _______ .i037L4. ________ 0.0 êO7 2.v? 162228.8$ & 0742.95_________________
1474.75 0.0 368.69 0.0
SANITARY SEWER INTERCEPTORS -
CAPITAL . . . . . 218944.31. . . ._ 22243 5.44__._._ . . 0.41 . ______— 0.0 ________ _________________
U. N . . . 162.00 648.00 202.50 0.0
STORM SEWER LATERALS
— - CAPITAL . 280.9 * - ° —- —.- 0.0 10592,39 . •j _
O.M 0.0 800.76 228.79— U4.39
STORM SEWER iNTERCEPTORS
_CAP!TAL _ . . -- 290 22. 56. — ._ . ... .O.o.....L. • 90694.56 272083.75 7255.6’ fl-n
0.0 810.00 202.50
S.) ALL VALUES ARE IN oASE YEAR DOLLARS
A LA14lç ENTMY INDICATES Tb GROUP UR MECHANI$4 iS NOT USED FOR FU IAT4CIG THE COST TYPE ________________________________
As. El ,TRY OF ZERO INOICATES THE GROUP OR NECHASISM HAS NOT CHOOSE$ FOR FINANCING THE COST TYPE
-- I . .) CAPITAL COSTS ARE IN DOLLARS PER GROUP OR MECHANISM
- - 0.11 COSTS ARE IN OIIL I.A&S PER YEAR PER CRIMP 04 MECHANISM _.._. _. . ._ __________________________
-4
-------
providing environmental control facilities for new develop-
ments. The model can be employed for analyzing the
impacts of individual developments or of zoning policies
applied to an entire sub-basin. In addition, the impacts
of alternative cost allocation schemes can be assessed with
this model.
In our introduction to this report, we stressed the
importance of considering the interaction between environ-
mental, social, and economic impacts in any land use
planning process. Within this framework, a fiscal impact
model such as the one developed here performs several
important functions:
1. 208 areawide planning calls for an evaluation of
the cost effectiveness of alternative physical and land use
controls for water quality management. Previously, facility
planning under Section 201 of Public Law 92-500 had been
concerned primarily with treatment plants and major inter-
ceptors. The fiscal impact model makes it possible to
include land use control alternatives because it enables
planners to consider all relevant control costs, including
those of on-lot disposal and local collection costs.
2. Local decision-making on land use takes place in
a political context where an overriding concern is often
that of fiscal impact and, in particular, residential
property taxes. By making it possible to predict at least
one class of costs, the fiscal impact model will serve to
facilitate local decision-making.
3. The fiscal impact model will enable planners to
determine the cost allocations implied by existing pollution
control standards and financing policies. Plans and policies
can then be designed to alleviate any undesirable distribu-
tional impacts which may be revealed. These may involve a
change in financing policies or other fiscal measures.
The fiscal impact model which we have developed pre-
sently considers only a restricted set of land uses and
land characteristics, and employs several cost functions
which are in need of considerable refinement and re-evalua-
tion. Nevertheless, we feel that the model, if used
properly, will give results which are adequate for many
planning purposes. Moreover, we are sàtisfied that the
existing model structure provides an excellent basis for
development of a more complete impact model, should that
need arise.
6—48
-------
Section 7
Land Use Air Pollution Relationships
Introduction
The connections between land use types and air quality
are shown in Figure 7-1. The relationships described do
not extend beyond the level of emissions. There is little
concensus about the present state—of-the-art for models to
forecast short—term relationships between emissions and
ambient air quality for various pollutants. This is parti-
cularly true in modeling non-stationary sources. Therefore,
our attempt to show a way of integrating consideration of
air, water, and land use planning into one framework would
be jeopardized by the potential problems of ambient air
quality modeling. If we intend to have congruent efforts
with water and air, then short-term air quality modeling
rather than modeling of average annual values is desired.
The framework outlined should provide the necessary
guidelines for a systematic search through pertinent litera-
ture and work-in-progress to determine the state-of-the—art
and what models or guidelines are now available for use by
land use planners. For example, it should tell the planner
where to find industrial emission factors and how to esti-
mate home heating system emissions. We have limited,
iowever, the discussion in this section to stationary sources.
Air Pollution from Stationary Sources
In our evaluation of air pollution, we divide stationary
sources into three general types: residential, commercial,
and industrial. In the two former types, air pollution
results from emissions due to fuel consumption for both
space heating and appliance use. In the latter, it results
from emiSSiOnS due to fuel use for space heating and process
heating and from emissions due to the processing itself.
The literature on assessing fuel usage for space heat-
ing and appliances focuses on four basic methods. The most
-------
DU Dwellinq Unit
EF • Effluent
Figure 7 1: LAND USE AIR POLLUTANT EMISSION RELATIONSHIP
-------
common is the ASHRAE method.* It calculates the loss from
each component of a structure and then, along with certain
other factors, adds them to find total hourly heat loss.
The next method of calculating heat loss is based on a com-
puter program devised by Hittman Associates. It is called
the Time Response Method.** It describes the structure as
a thermal system acted upon by internal and external loads.
Direct measurement is the third method. Finally, in the
case of appliances, utility companies have done saturation
surveys to determine how many houses have what types of
appliances.
The general approach for converting space heating loads
to yearly emissions consists of four basic steps: (1) esti-
mates of hourly heat losses are made; (2) these hourly
estimates are converted to yearly values using formulas dis-
cussed below; (3) efficiencies in fuel requirements are
determined and (4) finally, emission factors such as those
developed by the EPA,*** are applied to the fuel figures to
determine the resulting emissions.
Emissions from industrial processes are found using
EPA emissions factors for the SIC processes. Some litera-
ture also exists on fuel usage in manufacturing. These data,
along with EPA emission factors for fuel, can be used to
give air emissions only for the fuel-using portion of the
process.
The three types of stationary sources are discussed
below.
Residential Emission
We have computed emissions from a 1,500 square foot
house and also from a development of 540 such units by using
a variety of literature sources and computing methods. (See
values of fuel usage and emissions in Table 7—1.)
* American Society of Heating, Refrigerating, and Air Con-
ditioning Engineers, Handbook of Fundamentals , 1972.
** Hittman Associates, Residential Energy Consumption:
Verification of the Time-Response Method for Heat-Load
Calculation, HUD-HAI, June, 1973.
U.S. Environmental Protection Agency, Compilation of
Air Pollutant Emission Factors, Second Edition, April, 1973
AP—42; relevant emission rates for gas (pounds/b 6 ft. 3 )
and oil (pounds/10 3 gallon) are summarized in Table 7—1.
7—3
-------
Table 7-1
Emission Rates*
Gasa 011 b
Particulates 10 10
so 2 0.6 142SC
so 3
Co 20 5
Hydrocarbons 8 3
NO 2 80 12
HCHO (aldehydes) 2
a Emission rate in lbs/10 6 ft. 3
b Emission rate in lbs/10 6 gal.
C S is equal to percent by weight of sulphur in the oil.
* U.S. Environmental Protection Agency,”Compilation of Air
Pollutant Emission Factors, Second Edition, April 1973,
AP-42; U.S. Environmental Protection Agency, Supplements
#1, 2, and 3 to the above.
7—4
-------
The ASHRAE heat loss method is used by fuel companies
when calculating the size of unit and amount of fuel needed
for a structure. In the ASHR E method square footage of
exposed area is estimated and coefficients of heat transmis-
sion are calculated for the different boundaries (masonry
walls, windows, etc.). From these values, transmission*
heat loss is calculated. From estimates of air exchange and
the volume of air in the structure, heat loss from infiltra-
tion is computed. The sum of transmission and infiltration
heat losses gives the total hourly heat loss. Hourly heat
loss is converted to yearly loss by employing the “degree
day” method which is based on the lowest outdoor temperature
and number of degree days in an area. Then, assuming a fuel
use efficiency (e.g., ASHRAE recommends 70 percent for
natural gas), the yearly heat loss is converted to its equi-
valent in fuel quantity. (See summary of method in Table
7—2.)
The New England Fuel Institute (NEFI) makes a slide
rule type device called the Heat Cost Calculator.** Yearly
fuel usage can be calculated for residences of five differ-
ent kinds of construction. This gives good results and can
be used if the investigator does not wish to calculate
transmission coefficients for every type of boundary in the
structure or if he only knows the approximate kind of
insulation or construction. This calculator is a shortened
version of the ASHRAE method combined with the degree day
method (see summary of method in Table 7-3). We include in
Table 7-4 calculations done with the Heat Cost Calculator
for three different kinds of construction.
The Electricity Council Research Center (ECRC) in
England has developed a mathematical model of the thermal
behavior of buildings.*** It treats the external boundaries
as networks of resistances and capacitances, as derived from
the thermal properties of the boundaries. The rooms in a
* By transmission we mean the combined effects of three
modes of heat transfer: conduction, convection and radia-
tion.
** New England Fuel Institute, Heat Cost Calculator (Manual
Device).
P. Basnett, “Modeling the Effects of Weather, Heating
and Occupancy on the Thermal Environment Inside Houses,”
Conference: Mathematical Models for Environmental Problems,
Southampton/U.K., September, 1975.
7—5
-------
Table 7-2
ASHRAE Method
(Not Room by Room but Over the Whole Structure)
1. To calculate hourly heat loss (BTUH)
A. Find exterior wall area, floor area, roof area,
door area, and window area. If oniy the square
footage of the structure is known, it will be
necessary to assume a shape.
B. Calculate transmission coefficients (u) for
each (wall, roof, etc.) by using ASHRAE tables
in their handbook. It is necessary to know
each kind of construction material used, 1/2”
plywood, 2-1/4” fiberglass insulation, single
glazed windows, etc.
C. Assume an outdoor and indoor temperature to
get temperature difference across each boundary.
D. A, B, and C are multiplied together for each
component and infiltration is added to the sum.
E. To get infiltration, cubic feet of the structure
must be known along with an infiltration coeffi-
cient (ASHRAE tables). This coefficient will vary
between residential and commercial, etc. Tem-
perature difference between inside and outside
must also be known.
2. To calculate fuel usage
A. Assume number of degree days.
B. Assume a correction factor (ASHRAE tables) based
on outdoor temperature.
C. Determine a fuel consumption constant (ASHRAE
tables) based on an assumed efficiency of
utilization (one could do step 2 on the heat
cost calculator if one wished).
3. To calculate emissions: multiply fuel usage by EPA
emission factor.
7—6
-------
Table 7-3
Heat Cost Calculator Method
1. To calculate Hourly Heat Load
A. Know or assume number of cubic feet (square
feet x ceiling height) in the structure.
B. Assume one of five kinds of construction from
not insulated to very well insulated.
C. Know outside design temperature.
2. To calculate fuel usage
A. Assume number of degree days.
B. Assume design temperature difference
(difference between indoor and outdoor).
C. Assume an efficiency of utilization.
3. To calculate emissions: multiply fuel usage by
EPA emission factor for that type of fuel.
7—7
-------
TaM. 7-4
kir i. .sians Du. to Pu ,1 Us. for Hasidential Heating
$ing l. Unit’
540 lAUts
Singl.
540 Unit.
Single Unit?
540 Units
Single Unit’ 1
540 Units
8 untts
H ttman
540 Un ta
Hittaan
Il tit e en
PItttnn
Tow .,
low,,
A
A
C
C
S
S
House
House
House
House
Ss*t LoeS IIU/y..r
13 s
9.86 x 10
5.78 8 10
8.53 x 101
58 a
r i u.. t gas (ft 9 zlO 6 )
oil (g .1 10 )
.179
1.16
96.7
626.0
.135
.880
72.9
475.0
.0789
.516
42.6
279.0
.117
.750
63.2
405.0
.83
5.18
56.0
350.0
gas
p .xticulats . oil
1.79
11.6
970.0
6,264.0
1.35
8.8
729.0
4,752.0
.79
5.16
427.0
2,786.0
1.17
7.5
632.0
4,050.0
8.3
51.8
560.0
3,500.0
gas
SO 2 oil
.107
164.7$
58.0
88,938$
.081
1255
44.0
67,500 $
.047
73.3S
25.0
39,582S
.070
106.5S
38.0
57,510S
.498
735S
33.6
49, ’ .12S
S0
oil
C
-
2.32S
-
1,2533
-
1.763
-
950S
-
1.03$
-
556S
-
1.5$
-
Hios
-
10.4S
-
70i 3
. 8
gas
co
• oil
C
3.58
5.8
1,933.0
3,132.0
2.7
4.4
1,458.0
2,376.0
1.58
2.58
853.0
1,393.0
2.35
3.75
1,269.0
2,025.0
16.6
25.9
1,120.0
1,748.0
0
gas
oil
1.43
3.46
772.0
1,879.0
1.08
2.64
583.0
1,426.0
.63
1.55
340.0
837.0
.94
2.25
508.0
1,215.0
6.64
15.5
446.0
1,04’ .0
A—
gas
109 ‘ oil
14.3
13.9
7,722.0
7,506.0
10.8
10.6
5,830.0
5,724.0
6.3
6.19
3,400.0
3,343.0
9.4
9.0
5,080.0
4,8(0.0
66.4
£2.0
4,462.0
4,12 .0
8610 gas
(a2 s) oil
-
2.32
-
1,253.0
-
1.76
-
950.0
-
1.03
-
556.0
-
1.5s
-
810.0
-
-
Pootnotest S.. Table 7—4 (continued) attached.
-------
Footnotes to Table 7-4 :
a. House Type A: no insulation, 1500 ft. 2 , 8 ft. ceiling,
700 indoor, 00 outdoor, 4500 degree days, calculation
done with Heat Cost Calculator.
b. House Type C: 2” or more of insulation on walls and
ceiling, 1500 ft. 2 , 8 ft. ceiling, 700 indoor, 0°
outdoor, 4500 degree days, calculation done with
Heat Cost Calculator.
c. House Type E: 6” insulation on ceiling, 2” or more of
insulation on sidewalls, storm sash, 1500 ft. 2 , 8 ft.
ceiling, 70° indoor, 0° outdoor, 4500 degree days,
calculation done with Heat Cost Calculator.
d. Hittman House: 5” c i1ing insulation, 2—1/4” sidewall
insulation, 1500 ft.’, 8 ft. ceiling, 75° indoor, 100
outdoor, 4500 degree days, calculation done using the
aggregate form of the ASHRAE method.
e. Hittman Townhouse: 5” ceiling insulation, 2 air spaces
in sidewall, each unit 1300 ft. 2 , 8 ft. ceilings, town-
houses in groups of 8 units -- 4 upstairs and 4 down-
stairs, 750 indoor, 100 outdoor, 4500 degree days,
calculation done using the aggregate form of the ASHRAE
method.
f. Gas efficiency assumed 70%; oil efficiency assumed 80%.
g. For emission factors, see Table 7-]..
7—9
-------
building are linked to each other and to the outside air by
these networks. The rooms themselves are treated as indi-
vidual volumes of air at uniform temperature. The results
of the ECRC model have been for the most part good, in that
on overcast days calculated consumption has been within 5
percent of measured consumption. The model has not worked
so well, though, on sunny days when its values are up to
25 percent away from the measured ones. This deviation has
been attributed to the effects of solar radiation, both
that which goes directly into the room and that which is
stored in the walls, the latter reducing the heat loss
during the night. Effort is under way to incorporate these
effects into the model.
When ECRC was constructing the model of a wall using
the values for conductivity, diffusivity and other factors
found in the Institute for Heating and Ventilating Engineers
(IHVE) Guide (comparable to ASHRAE), the calculated tempera-
ture values were very different from actual temperature
measurements made in the wall. ECRC claims this is due to
the fact that the IHVE conducts their measurements when the
material is quite hot and therefore possesses a lower mois-
ture content. Therefore, values used in the model are
slightly different from those in the IHVE Guide. This is
interesting in that the loss due to transmission, that part
which is calculated using these values, is usually thought
to be the most exact part of the calculation.
Before a detailed calculation is presented, another set
of heat loss values is compared. Hittinan Associates have
done a series of residential heating studies for HUD.* As
part of their studies, they have developed the Time Response
Method, which simulates heat loss in a structure.** Table
7-5 shows values calculated with this method for a 1,646
square foot house. Also included in the series of Hittman
reports and Table 7-5 are estimates from fuel executives,
servicemen, and measured values. Also listed in Table 7—5
are estimates by Environmental Research and Technology, Inc.
* Hittman Associates, “Residential Energy Consumption:
Single Family Housing,” Final Report, March, 1973, HUD-HAI-2;
Hittman Associates, “Phase 1 Report, March, 1973, HUD-HAI-l;
Hittman Associates, “Evaluation of Heating Loads in Old
Residential StructureS,” January, 1974, HUD-HAI-7.
** Hittrnan Associates, “Residential Energy Consumption:
Verification. 99.”
7—10
-------
Table 7—5
Yearly Heat Load
ERT High Density 4.7 x 1O BTU
ERT Low Density 10.4 x BTU
Hittmafl use of ASHRAE 7.1 x 1O 7 BTU
Method - 1500 sq. ft. house
Fuel Oil Distributor A 7.68 x BTU
Hittman 1500 sq. ft. house The same
7 exact
Fuel Oil Distributor B 11.9 x 10 BTU house
Hittman 1500 sq. ft. house
Utility A 7.60 x BTU
Hittmafl 1500 sq. ft. house
Utility B 11.0 x 1O 7 BTU
Hittman 1500 sq. ft. house
Hittman Chart in Phase 4]. 10 ÷-‘ 7.51 io 7
One Report - Range of BTU
Values for 1500 sq. ft. house -
(Utility Company’s Chart)
Hittman Calculation using 8.0 x 10 BTU
time response method of heat
load for their characteristic
house
Hittman Multi-family Report - 11.6 X BTU
1500 sq. ft. single family house
Average and range of measured average: 6.9 x STU
loads in twenty-three 1646 sq.
ft. houses in Twin Rivers, range: 4.7 x l0 BTU to
New Jersey - Hittman Report 8.8 x 10 BTU
7—1].
-------
(ERT) for homes in low density and high density develop-
ments.* ERT has not, however, told us how these figures
were derived.
The question naturally arises as to what extent the
values for heat load or fuel usage for a certain square
foot house in a certain area of the country and of a certain
type of construction, can be adapted to a house of a differ-
ent size, in a different area, or of a different. kind of
construction.
We reproduced a chart from the Hittman report showing
ranges of heat load for houses of different sizes (Figure
7-2). The relation is not linear since the addition of a
room, for example, includes windows and other components
of differing transmission coefficients. Thus it is at best
an approximation to multiply a value obtained for a 1,500
square foot house by 4/3 to obtain the load for a 2,000
square foot house. Neither are there easily usable rela-
tions between different kinds of construction, short of
those obtained through the ASHRAE transmission coefficient
tables. Adjustment according to region, however, is an
easier matter. The two coefficients used are (1) an adjust-
ment factor for lowest outdoor temperature and (2) the
number of degree days. These are purely multiplicative
factors.
To illustrate how heat loss computations are made, a
sample calculation is presented. In this sample we apply
an aggregated form of the ASHRAE—degree day method (see
Table 7-2) to a house characteristics of the design employed
in the Hittman Associates reports. The home is assumed to
be located in Maryland.**
The transmission coefficient, U, is equal to the
inverse of the sum of the respective resistances across
each layer of the boundary and also those of the air on each
side. These resistances are found in ASHRAE tables. U has
the dimensions BTUH/ft. 2 -degree.
* Environmental Research and Technology, Inc., “Methodology
for Determining Emissions from Land Use Planning Data,” June,
1972.
** Our aggregated form of the ASHRAE-degree day method
applied to another housing structure resulted in a heat
loss of about 10 percent less than the one calculated by
the room to room method used by ASHRAE. This is an accept-
able deviation in the light of all embedded uncertainties.
7—12
-------
Figure 7—2
Annual Electrical Heating Use Pattern Versus Floor Area
(Reference 2)
3000- I 1±20%
Variations attributable primarily
to house construction and style.
I-
2500-
0
C,,
4
w
-J 4
2O00- F
(A) Q
U-
U
£
u O e.
1500 — Heating Only
1000 1±30%
0 4 8 12 16 20 24 28 32 36 40
HEATING REQUIREMENT, THOUSAND KWHR
-------
1/Uwalls = .68 (still exterior air) + .81 (wood shiplap) +
.62 (1/2” plywood) +.5 (1/2” drywall) ÷
.68 (still interior air) = 3.29 -* U = 0.304.
But there is R - 7 Batting (insulation) so, according to the
ASHRAE tables, U is adjusted to .1.
1/Uroof: There is an unheated, ventilated attic, but
in this aggregated method we assume the roof
adjoins the 2nd story ceiling.
1/Uroof and ceiling = .68 (outside air) + .44 (asphalt
shingles) + 3.70 for each inch of 5 inch
loose fill blown in ceiling insulation and
.68 (inside air) = 20.3, so U = 0.05.
U windows = 1.13
U wood doors = .49
U patio doors = 1.
U floor = .2 (this is a very loose approximation
based on an ASHRAE example)
If only the area of the house is known, then one must assume
a shape in order to find exterior wall area. We assume that
it is square. Table 7-6 shows the resulting heat loss in
BTUH. To compute yearly heat loss and fuel usage, we apply
the Degree-Day Method:
Fuel Consumption = (quantity of fuel used per degree day
x 1000 BTUH x(BT g 00 ) x (number of)
degree days) x (temperature correction
factor)
The first factor is a unit fuel consumption constant given
in the ASHRAE Handbook.* BTUH are calculated. The number
of degree days (dd) for the area can be found from industry
sources or weather bureau data, and the temperature correc-
tion factor is found in the ASHRAE Handbook for the lowest
temperature in the area (+ 100 in this case).
Thus, the consumption of gas and oil for the same
assumptions are as follows:
* It depends upon utilization efficiency. ASRAE suggests
a 70% efficiency for gas, which is what we use -— and an 80%
efficiency for oil which we also use.
7—14
-------
Table 7-6
Heat Loss Calculation
Exterior Temp. Difference Heat Loss
U Area (sq. ft.) Across Boundary BTUH
Walls .10 1856 65 12,064
Roof .05 850 65 2,762
Wood Doors .49 60 65 1,911
Patio Doors 1.0 40 65 2,600
Windows 1.13 180 65 13,221
Floor .2 850 25 4,250
Infiltration 1 15,865
TOTAL 52,673
BTUH
1 lmplies infiltration coefficient of .018 and one air
exchange per hour for the house volume of 13,560 ft. 3 , and
a temperature difference of 65° across boundaries.
7—15
-------
Gas Consumption (therms) = (.0049/dd/l000 BTUH)
x 52,673/1000 x 1000 BTUH x 4500 dd x 1.167
= 1350 therms.
Oil Consumption (gallons) = (.00304/dd/l000 BTUH)
x 52,673/1000 x 1000 BTUH x 4500 dd x 1.167
= 840 gallons.
The summary of applications of our method and of the
heat loss slide rule in Table 7-4 shows that the type of
construction will greatly affect fuel usage. The House of
Type A uses more than twice as much as the House of Type E.
Our aggregated ASHRAE method calculated fuel usage for a
house with 5” insulation and no storm windows. Thus,
structurally it falls somewhere between Type C and Type E.
And our calculation showed that fuel usage also falls in
between these two. Moreover, according to our calculations,
if the insulation were removed from this house, making its
construction comparable to A, then fuel usage would increase
approximately 75 percent, a value close to that of House A.
Finally, space heating for residential structures
comprises only about 70 percent of total fuel use. The
other major fuel using activity is hot water heating,
followed in importance by cooking.*
Commercial Emission
Unlike the case for residential structures, there
exists no reliable method for calculating heat loss from
commercial structures. The heat loss for the same kind of
operation in the same type of building can vary by 100 or
200 percent or more. A heating engineer for a local gas
company answered us in an interview: “The only thing that’s
consistent is the inconsistency.”** This gas company
attempted to relate heat loss from restaurants, for example,
to many parameters including size, occupancy, type of
establishment and prices on the menu; but the were unsuc-
cessful.
Because of the difficulty of prediction, there exists
virtually no literature on heat loss from commercial struc-
tures. Almost all of our information was supplied by the
* Hittman Associates, “Residential Energy Consumption:
Single Family Housing.
** Commonwealth Gas Company, Southboro, Massachusetts,
June 11, 1975.
7—16
-------
aforementioned gas company. In our Table 7—7 we present
figures for heat loss for different kinds of commercial
establishments which were selected to be representative of
their respective categories.
We do not have figures for all types of commercial
establishments since either they were not on hand at the gas
company or because they varied to such an extent that they
would be meaningless.
When a heating engineer in the Boston area (5,600 degree
days/year) makes a heat loss estimate for a building, to be
on the safe side he usually begins with a figure like 50
BTU/ft. 2 . This assumes about 4-inch wall insulation and 6-
inch ceiling insulation. The actual loss is more like 35
BTU/ft. 2 . This will vary greatly, of course. The gas corn-
pany’S building is new and well—designed so heat loss is
about 20 BTU/ft. 2 . An old building will have a heat loss
greater than 50 BTU/ft. 2 . But as we have already noted, this
will vary to a great extent according not only to the type
of establishment but to which particular establishment one is
studying.
It should be noted that one of the differences between
commercial or industrial and residential is that the first
two have specified hours of operation during which the tem-
perature is kept higher than during the off-hours. The
? merican Gas Association, therefore, recommends the follow-
ing method of finding an average indoor temperature and of
proceeding to an annual fuel requirement:
Degree hrs./week = temperature x (hrs. per day) x (days/week)
(Degree hrs./week)/(168 hrs./week) = average temperature indoor
(BTU/hOUr heat loss) x 24/(Design temperature difference)
= BTU/Degree Day
(BTU loss per Degree day) x (Degree Day/season)/heating
efficiency fuel required per season.
Thus, if on weekdays a store is open for ten hours during
which the temperature is kept at 70°, closed for 14 hours
during which the temperature is kept at 60°, and then on
weekends closed for 24 hours during which the temperature
is kept at 500 then,
7—17
-------
Table 7—7
Fuel Usage and Waission. for $.l.ct.d C rcia1 *stabli.h.unt.
Distributing
Dupersarkat Discount Store Warehouse Warehouse Car Dealer Warehouse Large Office C1n
area (tt 2 xlO 3 ) 30. 55.04 41.6 54.4 16.72 400.0 201.0 sot available
voijae (tt 3 x10 3 ) 360. 770.6 832.0 1,360.0 200.6 10 ,400.0 4,820.0 sot availabi.
73-l374
Gas (tt 3 giO t ) 2.92 1.63 3.35 7.6 2.13 32.26 13.54 4.46
Oil (gal xlO ) 18.96 10.56 21.77 49.40 13.82 209.7 07.99 28.02
b
. .i i ,ns (lbs. )
gas 29.16 16.25 33.49 76.00 21.27 322.6 135.4 44.64
particulate. oil 189.6 105.6 217.7 494.0 138.2 2,097. 079.9 290.2
gas 1.75 .98 2.01 4.54 1.28 19.36 8.12 2.68
2 oil 2692$ 1500$ 309 1S 7015$ 19625 29,7778 12.4948 41218
$03 9 5 — — - — — — —
oil 37.9S 21.128 43.548 98.8$ 27.64$ 419.4$ 175.985 58.04$
gas 56.32 32.50 66.90 152.0 42.54 645.2 270.8 09.20
CO oil 94.8 52.8 108.4 247.0 69.1 1048.5 440. 145.1
gas 233.3 130. 268. 608. 170. 2581 1.083. 357.
803 oil 227.5 126.7 261.2 592.0 165.8 2516 1.056. 340.2
NCHO aid gas — - - - - -
e., as oil 37.92 21.12 43.54 90.8 27.64 419.4 175.95 50.04
9 5 5 21.31 13.0 26.0 60.0 17.0 258.1 108.3 35.7
/ oil 56.88 31.68 65.31 140.2 41.46 629.1 264.0 07.06
a. Gas use seasured by C nwea1th Ga.j to get oil use we first soltiplis4 a 98$sa by .7 (thu. a.sueinq 70% gsa efficiency). ibis gee. we
heat load. We divided this by .8 (thus aasueing 80% oil efficiency) t rsby .ttinq oil use.
b. For e.ission factors, see Table 7-1.
-------
70 x 10 x 5 = 3,500
60 x 14 x 5 = 4,200
50 x 24 x 2 = 2,400
10,100 degree hrs./week
10,100 = 600 average indoor temperature.
Industrial Emission
Air emissions due to industrial processing can be found
by referring directly to EPA’S compilation of emission fac-
tors.
Air emissions due to fuel use for space heating is much
more difficult, however. Environmental Research and Tech-
nology, Inc. (ERT) has done a study in which they present
figures for different industries; but the variances are so
huge that ERT expressed warnings in their report with respect
to the unreliability of the figures.*
Heat loads for industrial establishments cannot be cal-
culated on the basis of physical characteristics of the
building since machinery produces large heat gains. There is
therefore great interaction between process heat and space
heat. This is obvious in the case of a steel mill. And in
Hartford, for instance, an office building is supposedly
being heated with the heat given of f by the computers within
the building. Therefore, when machinery is involved, it is
futile to make estimates which disregard the large heat gain.
Moreover, the myriad number of ways of ordering activities
within a manufacturing establishment make estimation even
more difficult.
Thus, the method used to calculate heat loads for resi-
dential structures is for the most part not applicable to
industrial structures unless the large internal heat gains
are also taken into account.
* Environmental Research and Technology, Inc., Air Quality
for Urban and Industrial Planning, March, 1974; U.s. Environ-
mental Protection Agency, “A Guide for Considering Air
Quality in Urban Planning,” March, 1974.
7—19
-------
S urnma ry
In order to get a rough idea of the importance of emis-
sions due to space heating in comparison to total industrial
emissions, we give values of emissions for several paper
production plants, a group of open hearth furnaces, an elec-
tric arc furnace, and a coal—burning power plant in Table
7-8. As can be seen, emissions due to space heating for
developments such as we are considering seldom come close to
those from other activities.
This is not to say that emissions due to space heating
should not be taken into account. Space heating produces
8 million tons of the nation’s 133 million tons of air
pollution produced per year.* Thus while small, it is not
inconsequential. Moreover, local effects of space heating
may be significant in particular areas.
On the level at which we are speaking then, small dif-
ferences due to different methods of estimation do not make
much cumulative difference. In the case of residential
heating, the planner can go through the simple ASHRAE
method. In the case of commercial heating, he can make use
of the approximate values we give in the tables, a rough
factor such as 35 BTUH/ft. 2 , the ASHRAE method, or a combina-
tion of all three. Finally, in the case of industrial
sources, EPA emissions estimates, scaled down by the analyst
for internal heat gain, can provide the basis for emissions
estimates.
* Edmund K. Faltermeyer, “We Can Afford Clean Air,” in
Pollution: The Effluence of Affluence , Frank J. Taylor,
Philip G. Kettle, and Robert G. Putnam, Methuen, 1971.
7—20
-------
Table 7-8
Emissions from Selected Activities
1. Iron and Steela
Particulates (lbs/day )
13 open hearth furnaces 26,800
(175 ton is the rated (Best available tech-
capacity of each) nology -- 5,100)
Two electric arc furnaces 5,250
(50 ton is the rated capa- (Best available tech—
city of each) nology -- 600)
2. Paper and p i b
Particulates TRS*
( lbs/day) ( lbs/day )
1,000 57.7
Number of 4,000 240.
Selected —— 2,750 (SO 2 only)
Plants 6,136 7,000 — 9,000
7,625 2,500 — 3,000
a
3. Coal Fired Power Plant
Particulates (lbs/day )
300,000 kw with good dust
collector 11,000
300,000 kw with fair dust
collector 40,000
* TRS = total residual sulphur (Hydrogen Sulfide, Methyl
MerCaptan, Dimethyl Sulfide, Dimethyl Disulfide).
a Schueneman, J., et al., Air Pollution Aspects of the Iron
and Steel Industry , U.S. Department of Health, Education,
iid Welfare, 1963.
b Council on Economic Priorities, Paper Profits , MIT Press
1972.
7—21
-------
Appendix A
Data Collection for STORM and SWMM
The Mill River Drainage Basin*
The portion of the Mill River Basin investigated in our
study extends from Whitney Lake Dam to the headwaters. The
drainage basin above the lower limits of this study area is
37.7 square miles and is of triangular shape, about 13.5
miles long and about 5.5 miles wide at the upper end.
Elevations range from 880 feet (MSL) on top of Mt. Sanford
to 36.3 feet (MSL) at Whitney Lake Dam Spillway. The Mill
River flows in a southerly direction adjacent to Route 10
from the center of Cheshire in the vicinity of Route 70 to
New Haven Harbor and Long Island Sound. The average slope is
20 feet per mile from the headwaters in Cheshire to the
Hamden town line and 9.4 feet per mile from the Cheshire—
Hamden town line to Whitney Lake Darn.
The largest tributary to the Mill River is Willow Brook
with a drainage area of about 12.7 square miles. It joins
the river in Hamden near the Cheshire-Hamden town line.
Other contributors to the flow on the river are Butterworth
Brook, Jepp Brook, Eatons Brook and Shepard Brook.
There is one major dam in the Mill River System:
Whitney Lake Darn. The dam, constructed in 1860-61, was
modified in 1916 and now has a spiliway crest length of
approximately 250 feet. A flow rating curve of the darn was
drawn by Malcolm Perne as part of a study entitled “Report
on the Effects of Flood Flows on the Lake Whitney Darn”,
done for the New Haven Water Company in May 1956. The
curve is
Q = CLH 15 (A-l)
with C = 3.3, reflecting a Broad Crested weir;
L = 250 feet, the effective length of the weir; and
H = head on the weir (water level).**
* Most of the following information is drawn from a flood-
plain study made by the Corp of Engineers: Floodplain
Information, Mill River, Hamden, Connecticut , Prepared for the
town of Hamden by Department of the Army, New England Division,
Corps of Engineers, Waltham, Massachusetts, March 1968.
** Above a head of one foot a correction factor was applied
due to submergence of the side channel portion of the spiliway.
A-i
-------
Since the spiliway crest is long, the spiliway is capable of
passing high flows at relatively low heads. Lake Whitney
has a useable storage of 258 million gallons.*
Flow records on the Mill River are sparse; no official
gauging stations exist. The only recording has been done
by the New Haven Water Company. Daily levels of Lake Whitney
at the company’s water intake have been monitored since 1897.
During the 40-year period from 1918 to 1957 the average flow
at the dam was 42 mgd. Thus, the average detention time
is about 6.1 days. The minimal annual average flow for the
above period was 26 mgd in 1930 (i.e., about 8.2 days of
detention in the lake) and maximal was 77.2 mgd in 1953
(i.e., about 3.3 days of detention). No travel time estima-
tion has been performed for the Mill River. The travel time
from the Flamden—Cheshire town line to Waite Street (the
upper end of Lake Whitney) is about 12 hours for a flow
velocity of 1 foot per second. Some profiles of the Mill
River are kept in the files of the U.S. Army Corps of Engi-
neers originating from their floodplain study.** They were
used for developing the data necessary for the modeling
efforts.
The drainage was divided into sub—basins. Natural
drainage patterns determined the selection of the following
11 sub-basins (see Figure A—i)
Lake Whitney (West)
Lake Whitney (East)
Shepard Brook
Centerville and Hamden
Central Mill
Eaton Brook
Northern Country Club
North-Eastern Mill (Butterworth)
Willow Brook -- Hickory Jeep
Willow Brook North (including parts of Cheshire)
Mill River (Cheshire)
* S. Jacobson, E.L. McLeman and R.E. Speece, “Estimating
Reservoir Yields on a Digital Computer”, J. American Water—
works Association , January 1959, pp. 51—54.
** U.S. Army Corps of Engineers, op. cit .
A-2
-------
Willow Brook /
/
/
/
/
/
/
/
/
/
/
/
___J_- — — — —
-
\
I S.-
‘I
—S
/ —- 5-
/
/
/
/
/
/
/
/
/ Lake Whitney
West
/
/
/
Mill-Cheshire
N.E. Mill
(Butterworth)
Tuttle Street
— — — —
————
Axeishop Pond
Cs
Dixwell Ave
— —.
—- 1
I
Skiff StreetJ
Lake. Whitney / C1 12 Channels
East , 1—13 ,: Junctions
I Inflow of
I Runoff
I
Dam
Figure A-i: Conceptualization of Mill River for Analysis
C II
Cl
R.
Willow Br.
-Hickory Jeep
I0
C3
Eaton
Brook
C5
Central Mill
Woodruff Darn
4
— — — — — — — ——5-----
I
I
1
I
1
— — —
A-3
-------
Water Quality Data*
During the search for water quality data sources of
varying reliability, precision, and extent were found:
1. The data of the New Haven Water Company, derived
from grab samples at the filter intakes:
— daily coliform (5 days/week) since 1966, data are
given as MPN)
- daily temperature of intake water
- daily outside temperature
- daily precipitation
- daily intake of water (mg)
( note : all the above data exist since 1923
- daily data of color and turbidity (5 days/week)
(note: these data must be extracted from unofficial
files)
— published monthly data: average values of color,
turbidity, and precipitation from the daily records
- quarter—yearly data on odor, color, chlorides, free
ammonia, albuminoid ammonia, nitrites, nitrates,
total hardness, conductivity, pH, alkalinity, iron,
total solids, fixed solids, fluoride, calcium,
magnesium, copper and sodium.**
2. A few grab samples of physical and chemical data of
raw water at the filter intake point were taken by the State
Department of Health; they are available on magnetic tape
from 1966 to the present. Six samples of interest were
identified.
* Our report is limited to surface water because no
groundwater quality data could be found. Dr. D.E. Hill of
the Connecticut Agricultural Experiment Station (New Haven)
believes that no samples of groundwater quality have been
taken on any systematic basis.
** Some of the data exist only in recent years; but tur-
bidity, color, chlorine (from NaC1), nitrogen as ammonia and
as nitrites and nitrates, total hardness and alkalinity are
sampled since the twenties; total and fixed solids as well
as iron were added inthe thirties, while oxygen consumed
(COD) was dropped.
A-4
-------
3. For one year (July 1955 to June 1956), weekly samples
of MPN, nitrite nitrogen, nitrate nitrogen, chlorides,
soluble orthophosphates (P0 4 ), pH, alkalinity (CaCO3) and
acidity (as C0 2 ) are available for a network of six stations
(3 at Lake Whitney, 3 at the Mill River). Some sampling
stations are located at tributaries of the Mill RIver, such
as Shepard Brook. Most of the data have to be read from
graphs in the official publication.*
4. For the period of March to November 1973 (except
August), the Quinnipiac College took monthly samples at four
locations in Clarks Pond and at two close points upstream
and one downstream of the pond.** Various biological as well
as physical and chemical indicators were sampled from a row-
boat within the pond: water temperature; Do; free C0 2 ; pH;
H 2 S; NO 3 ; NO 2 ; NH 3 ; organic N; P0 4 ; metaphosphate; silica;
color; turbidity; SO 4 ; total hardness; Cl; Fe; BOD; oxygen
demand index; conductivity.
5. Some samples were taken by the Southern Connecticut
State College:*** total solids, dissolved oxygen, and tur-
bidity were sampled five times during a three-week period in
spring and five times during a three-week period in summer
over a period of three years; scattered measurements of
chlorides, nitrates, and phosphates were taken. Measurements
were made at Clark’s Pond, Sleeping Giant, and Dixwell Bridge.
More complete bi-weekly measurements at six points on the
river were taken since August, 1974; the data were promised
but have never been made available.
Without discussing details of the available data the
following points should be made with regard to the desirable
calibration of the water quality model:
- there exist long time series at the intake point
for all the major constituents except for soluble
ortho-phosphates (P0 4 ); but they are mainly
quarter-yearly data except for daily coliform,
temperature, color and turbidity data;
* S. Jacobson and E. L. MacLeman, “Lake Whitney Sanitary
Conditions in Relation to Certain Laboratory Examinations,”
J.NEWWA , 1959, pp. 62-88.
** The purpose of the study was to find out the reason for
incomplete decomposition of plant life (including poison
ivy) in the pond: Cohen G. Becker, et al. , “A Limnological
Study of Benthic Decomposition in Clark’s Pond, Hamden,
Conn.,” Senior Research Paper, Dept. of Biology, Quinnipiac
College, 1973.
Dr. Dwight Smith is the faculty member responsible for
this research.
A-5
-------
— there exist almost no cross-sectional data for
total solids;
- there exist almost no BOD-DO measurements;
- there exist no reaction rates of non-conservative
constitutents in the water body;
— there exist very few cross—sectional data,
compared to the number of time—series data at the
intake point to the filters; the few available
cross-sectional data do not contain any flow
measurements;
- missing data in the series of daily data had to be
estimated because no water quality samples are
taken on holidays.* Sampling is also not conducted
on weekends, so it was decided to work with a
five—day week.
Climatological Data
The New Haven Water Company and the Mt. Carmel Weather
Station report daily temperature and precipitation data.
The former are available from the company’s records, while
the latter can be received from the U. S. Commerce Depart-
ment’s climatological reports. The Bridgeport airport
weather station is the nearest station which has recorded
hourly precipitation data up to date as needed for the
runoff model (see Table A-i for analysis of 1974 precipita-
tion record). The direct transfer of information from this
station to the Hainden area might result in incorrect preci-
pitation input. But this problem is common for almost
every study so that some guiding assumptions have to be
made. Otherwise the application of the models would not be
justified at all.
Drainage in the Hainden Area
The following brief section describes storm water
drainage characteristics prevailing in the subbasins in the
* A linear smoothing function was used to compute the
missing data; i.e. the average of the preceding and of the
succeeding measurement were taken. See also: R.E. Wyzga,
“Note on a Method to Estimate Missing Air Pollution Data”,
J. Air Pollution Control Association , Vol. 23, No. 3,
March 1973, pp. 207—208.
A-6
-------
Table A-i
Number of Days per Precipitation-free Interval
1974: Bridgeport Airport
Interval
w/o Precipitation
(days)
No. of Days
W/O Precipitation
1/2
76
1
12
11/2
8
2
6
3
20
4
12
5 •
9
6
10
7
3
8
4
9
1
10
3
11
1
>15
2
A-7
-------
Hamden area. The sanitary sewer system is connected to the
New Haven System. No combined sewers exist which could
overflow in the Mill River. Therefore the sanitary sewer
system is not further discussed.
Lake Whitney West . This area has storm sewers. The
waters from the area to the south of the western fork of the
river run off across Putnam Avenue to an outfall on the fork,
and the stormwaters from the west of the western fork dis-
charge to three or four outfalls. The fork divides the
runoff, with half flowing to outfalls on the western fork
and half flowing into Lake Whitney itself. There is a
drainage problem on Augur Street.
Lake Whitney East . This area is storm sewered for the
most part, with outfalls at many points to the river. There
is a portion of residential area, however, that is not storm
sewered. Some of the runoff from the central portion of
this sub-basin flows into a brook and then to the river.
Shepard Brook . For the most part this area is not
storm sewered. Most of the runoff flows into Shepard Brook
and then into the river, though some finds its way into
ditches. The waters from the storm sewers on Shepard Avenue
flow into the Brook at several points.
Centerville (Hamden) . This is fer the most part storm
sewered with outfalls at various points on the river. There
are two drainage problems: one where Pardee Brook (which
also carries some of the area’s runoff) runs into an apart-
ment complex, and another on Forest Street off Whitney
Avenue.
Central Mill . Some of the runoff from this area runs
into a brook and then into the river. Some also runs over
Mt. Carmel Avenue and then to the river.
Eaton Brook . Much of this is not storm sewered. The
part which runs partly down Shepard Avenue into Shepard
Brook and partly over West Todds Road into the river.
Runoff also flows into Eaton Brook and then to the river.
Northeast Mill . This area is not storm sewered. Some
of the runoff flows into a small brook and then to the river.
Northern Country Club and Willow Brook . There exist
very few storm sewers here. The existing ones lead over
Still Hill into River Road and then down Whitney Avenue into
an outfall on the river. Some of the runoff flows into
Willow Brook, while some flows into a ditch.
A-8
-------
Data on Soil and Slope Characteristics
Maps from the Soil Conservation Service, U.S. Department
of Agriculture, showing soil and slopes of our sub-basing
and a booklet,* prepared by the Connecticut Soil Conservation
Service, were available. The preparation of data for all
sub-basins to be used in the universal, soil loss equation
took considerable time.
1. Cheshire silt loam, a well-drained soil; hydrologic**
group B; and K-factor = .43
2. Penwood loamy sand, an excessively well—drained
soil; hydrologic group A; K-factor = .17
3. Cheshire fine sandy loam, well-drained; hydrologic
group A; K-factor = .17
4. Wethersfield loam, well-drained but having a shallow
fragipan which restricts internal drainage; hydro-
logic group C; K-factor = .43, B Horizon = .17
5. Watchaug fine sandy loam, moderately well—drained;
hydrologic group B; K-factor = .43
6. Manchester gravely sandy loam; excessively drained;
hydrologic group B; K-factor = .43
7. Holyoke rocky silt loam (also 94L and 9414, which
are, respectively, very rocky and extremely rocky);
a very rocky poorly drained soil; hydrologic group
D; K—factor = .43
8. Urban land
9. Branford silt loam, well-drained; hydrologic group
B; K—factor; B Horizon = .64; C Horizon = .17.
* U. S. Dept. of Agriculture, Soil Conservation Service,
“Special Soils Report: New Haven County, Connecticut, Soil
InterPretations for Urban Uses”, 1972 and “Know Your Land,
Natural Soil Groups for Connecticut”, Cooperative Extension
Service, Storrs/Conn., 1972.
** The hydrologic groups A to D reflect the drainage
characteristics of the soils; a soil of group A has the
best drainage. The K-factor describes the erodibi].ity of
a goil.
A-9
-------
The following section describes the soils found in the
different sub—basins.
Lake Whitney West . This sub-basin consists primarily
of urban land and Penwood. Some Manchester is also found
along with smaller quantities of Cheshire and Wethersfield.
The land is for the most part flat.
Lake Whitney East . This consists of mostly urban with
a large amount of Branford. Some Manchester is also found
along the river. The land is fairly flat.
Centerville . In the southern part there is a good deal
of Branford, while in the northern part, some very rocky
Holyoke occurs which is moderately to steeply sloped. There
is a strip of Branford which stretches up the length of the
sub-basin as well. Other than that, there is mostly Cheshire,
both silt and fine-sandy.
Central Mill . This area consists mostly of lightly to
moderately sloped Cheshire and moderately to heavily sloped
Holyoke. The Holyoke is especially predominant in the
northern part.
Shepard Brook . This area consists mostly of moderately
to heavily sloped Cheshire and heavily sloped Holyoke. It
becomes flatter in the southern part.
Eaton Brook . This portion consists of a lot of Cheshire,
but some Branford, Holyoke, Wethersfield and Watchaug are
also found. The slopes vary.
Northeastern Mill . This section consists mostly of
very rocky to extremely rocky Holyoke, with steep slopes.
A mix of other soils including Cheshire, Manchester and
Branford is also present.
Willow Brook . A lot of Cheshire is found here, lightly
to moderately sloping. A mix of other soils is also found.
Northern Country Club . This is mostly Holyoke, with
some Cheshire and Manchester.
The rainfall factor “R” for the region is assumed to
be “150” according to the “iso-erodent” maps.*
* u.s. Department of Agriculture, Agriculture Research
Service, Rainfall-Erosion Losses from Cropland East of Rocky
Mountains, U.S. Dept. of Agriculture, Agricultural Handbook
No. 282, 1965.
A-lO
-------
Land Use Data
Present Availability
The land use data for the Mill River Basin is available
from five sources for the town of Hamden, the ‘core of the
basin, and available only for the most recent period for
North Haven and Cheshire, the eastern and northern edges of
the basin. The data which is available is of high quality
and reasonably consistent, permitting the flexibility of
creating, through interpolation and research in town records,
continuous land use data for that portion of the basin within
Hamden. It should be noted, however, that the quality of
data available for Hamden is not representative of that gen-
erally available in communities within the U.S. The implica-
tions of this will be discussed.
Hamden’s first land use survey was undertaken in 1941.*
The second land use survey, and the first available to Meta
Systems, was done by Technical Planning Associates of North
Haven in 1948-49, “A Pilot Study for a Town Plan.” The
results of this study are presented on an exceedingly large
scale map presented in the report. The study reported land
use by 13 categories, only three of which would be comparable
to more recent land use studies showing compact residential,
commercial, industrial and rural uses. While this map is
difficult to use for detail, it is a useful benchmark from
which to begin a study of those areas which were densely
developed in the immediate post-war period. Furthermore,
the street pattern on that map gives a clear indication of
the built-up areas. This map combined with analysis of sub-
division permits from the same period could be used as an
accurate starting point of a study of the residential,
industrial and commercial land uses for the town. To extract
agricultural land use is considerably more difficult both in
the early and later studies because the definition of types
of land uses is far less clear. It is generally not possible
to tell when a piece of land was taken out of active produc-
tion a d entered a speculative period prior to its development
for residential purposes. Further, many of the farms ceased
to be productive even though they were not immediately broken
up for development. It is obvious that the available data
on this sector are limited for non-point source/water quality
considerations.
* We report this part in detail because it gives the
potential user an idea of the efforts needed to utilize even
a relatively uncomplicated model such as STORM.
A-il
-------
In 1954-55 Maurice E.H. Rotival and Associates prepared
A Short Approach Plan for Hamden. This report contained an
existing land use map, again of only moderate detail, and
scale. Residential single family is the only family classi-
fication employed, along with community, industrial,
institutional and several classifications for utilities.
As with the 1948 survey the detail contained is not sufficient
for purposes of a good intertemporal land use analysis but
may offer a second view of community settlement patterns
(through the road system) with which to follow the pattern
of development occurring in Hamden.
In 1963 Goodkind and O’Dea carried out a comprehensive
land use study of the town of Hamden. This study has become
the basis for all subsequent planning and zoning in the
community. At the time of the study the town also had a set
of aerial photographs taken by the Sanborn company; these
are available in the Town Hall.
The land use material presented in the Goodkin and O’Dea
report is on 1” to 200’ scale maps which have become the
basic planning maps for the town of Hamden. The accompanying
report contains detailed land use information on six classi-
fications of residential land use, two major classifications
of industries, utilities, institutional and other land uses
(which includes agricultural uses).
In 1974 a land use map of Hamden was prepared by the
staff of the University of Southern Connecticut. This land
use effort parallels closely that of Goodkind and O’Dea in
1963, but at a scale of roughly 1:2,000. There is only one
copy of this map; it is available in the town planning office
of Hamden.
Land use information for the remainder of the basin is
far less complete than that for Hamden. Both the Towns of
Cheshire and North Haven are without major land use maps of
any point in time. North Haven maintains a file of informa-
tion concerning past zoning in the community. Cheshire
maintains only a current zoning map.
One additional and potentially significant land use map
for the region exists at the USGS headquarters in Reston,
Virginia and at the Department of Geography at Dartmouth, N.H.
This map of the New Haven metropolitan area was developed as
a prototype mapping system by Robert Simpson of Dartmouth
and the USGS. The map is at a large scale, 1:100,000, and
covers an area of roughly 16 by 20 miles. It has been
developed from high altitude photographs and verified from
low altitude. The scale of the map makes it less useful
for this analysis, and it is for only one point in time.
A-12
-------
Given these considerations, however, the work is of consider-
able assistance in defining present land uses in areas
peripheral to the basin. The system employed recognizes two
classifications of housing —- single family and multiple
family or mixed family, industrial, commercial, transport
and utilities, institutional, recreational, agricultural,
forest and vacant land.
In summary, the land use information available for the
Mill River basin is considerably better than that for most
communities within the United States. The data for Hamden
is of relatively consistent quality over a twenty year period
and of excellent quality over the past ten years. Effort is
required, however, to interpolate between the views provided
by the various types of data. The availability of a large
scale recent land use map for the entire New Haven Metropoli-
tan region allows for complete information on current land
use within the basin in the Communities of North Haven and
Cheshire where no town level information is available. Use
of the USGS map is unfortunately limited because at present
only two prototypes of these maps exist.
Collection of Land Use Data
There were two major considerations in our collection
and storage of land use information; first, the need for a
system sufficiently general to be applicable to a large
number of sites where similar analyses would be undertaken;
and, second, development of a structure for collection and
storage which mirrors the needs of the analytical tool being
developed. The smaller the unit for data collection, the
more flexibility will, be available to the researcher for
aggregating data to a number of political and physical
onfigUrati0flS. Efforts have been made to develop general
models using disaggregated data; the National Science Founda-
tion—sponsored research on land use and environmental modeling
at the Harvard Graduate School of Design is one such case.*
The question of the appropriate size of basic grid cells is
not an easy one to answer. There are as many answers as
there are problems which require a spatial framework for
* See: The Interaction between Urbanization and Land:
QualitY and Quantity in Environmental Planning and Design,
Lä nd8caPe Architecture Research Office, Graduate School of
Design, Harvard University, Cambridge/MaSS. Progress report
for years one to three, October, 1974. (NSF R NN grant No.
GI—3 2603 ) See also: Tabors, Richard and Michael Shapiro
“Land Values and Public Investment in Urban Fringe Areas:
A case study of Clay, York”. Environmental Systems Program.
Harvard UniversitY, Cambridge/Mass., Working Paper, January
1975.
A-13
-------
analysis. Clearly the choice of a grid size for land use
work must be determined by the specific analysis being under-
taken, by the level and quality of data available and by the
cost effectiveness of finer and finer levels of detail. In
many planning environments the level of detail required of
a data system such as that developed by the group at Harvard
Graduate School of Design is not available and cost con-
straints prevent the collection of such data.
The geographic grid structure which we have applied in
the Mill River Basin is, we believe, a sensible compromise
which retains the advantages of a grid structure for storage
and retrieval of information and aggregation to various
configurations, but also retains a level of simplicity both
in collection and storage that can be implemented within a
planning office. The unit of data collection for the study
was fixed at 1/4 square mile or 160 acres. This selection
was based on several general considerations. First a lower
bound on the grid size was set by the quality of data avail-
able. Given the data base described above, a substantially
smaller grid, such as the one hectare size employed by the
study at Harvard Graduate School of Design, was not warranted.
Moreover, this conclusion was reinforced by our a priori
beliefs about the sensitivity of the runoff model. It was
felt that, given the hydrologic characteristics of the region
and the nature of the model it would not be possible to
determine the impact of land use changes in individual cells
as small as 2.5 or even 20 acres. At the upper limit, the
grid size was constrained by our selection of sub—basins
as the fundamental unit of the impact analysis (see below).
Since these basins were as small as two square miles, a grid
cell much larger than 1/4 square mile, say 1/2 square mile,
would have provided little more information than analysis
on the sub-basin level alone. In addition, there would have
been excessive difficulties in trying to reconcile grid cells
which overlapped drainage basins.
Given these limitations, there still existed a substan-
tial range of potential grid sizes, from roughly 20 to 320
acres. The selection of the 160 acre grid was an arbitrary
but, we feel, reasonable choice from among these remaining
possibilities. In addition, a number of previous studies of
infrastructure costs (see Section 6) had employed a 160 acre
grid. Thus the use of this size permitted us to compare the
results of our cost analysis with other results.
One of the considerations in the organization of a data
system for the Mill River Basin was the structure of
information required by the STORM model; STORM is designed
to work on basins of less than 10 square miles.* Based on
* See Section 5.
-------
the natural drainage conditions, the Mill River Basin was
divided into a series of sub—basins, each in the range of
two to nine square miles.* In addition to this criterion,
the intertemporal development patterns of the community
became a major criterion for this dividing pattern. Thus,
in the entire Mill River Basin we have defined three sub-
basins that comprise most of the older portions of the town
of Hamden, two that cover the new development areas of Hamden,
and the remainder which cover the less developed regions of
the three towns.
Each of the sub-basins has a relatively different path
of development in both time and space. For instance, the
Shepard Brook area has seen development in the southern
portion since the nineteen-fifties. The development has
proceeded north in the basin constrained only by the availi-
bility of land for development and the existence of a large
industrial park in the center of the sub-basin. The zoning
of the industrial area prevented the use of that land for
any other purpose. Because the industrial development was
slow, the center of the basin has remained relatively open
until quite recently. The Eaton Brook, the sub-basin to
the north of Shepard Brook, has yet to see any major develop-
ment. Only in the last five years has there been any
construction activity and that only of single family housing
on relatively large and isolated lots. The character of each
of the 11 sub-basin areas is sufficiently different to allow
for comparison on temporal and spatial growth within the
total Mill River Basin.
A significant aspect of our analysis of the methods
applied and of the use of the STORM model requires the
analysis of the spatial sensitivity of the model. To
accomplish this we have chosen one sub—basin, that of the
Shepard Brook, for a study in such detail sufficient to
* All models require some type of regionalization of data.
Choice of the regions for analysis could have been along any
one of a number of criteria and could have been either nodal
or homogeneous. Our choice of regions for the Mill River
watershed were physical rather than political or economic,
and they were homogeneous rather than nodal. There is nothing
that makes our choice more legitimate than any other for
analysis of land use and water and air quality in the basin.
Dealing with water or sewer facilities and with run off leads
to the use of natural drainage patterns and logically to the
use of sub—basins as the unit of analysis. For a more
detailed discussion of the process of regional definition,
see Tabors, R., The Definition of Multifunctiona]. Planning
9 ions , Harvard University Center for Population. Studies,
1971.
A-15
-------
attempt to identify the impact of concentrated development
activities as small as 160 acres, the size of a moderate
housing development. Our analysis then aggregates these
small units, forming progressively larger geographic areas
of analysis, finally reaching the sub—basin (see Appendix B).
Land use grid values for the Shepard Brook sub-basin for
1964 and 1974 are contained in Tables A-2 and A-3.
Collection of data was done utilizing a relatively pri-
mitive yet readily and inexpensively available method, grid
cell counting. Overlay grids one—half mile to a side (160
acres) were prepared. Each grid cell was then divided into
100 cells which were counted according to the major land use
type within the cell. The orientation of the grid pattern
within the basin was arbitrary but consistent in following
the major map boundaries of the 1963 land use study of
Hamden. The grid cells within the Shepard Brook sub-basin
were then labeled by numb!er and counted. For use in the
STORM model information for the other sub—basins was not
required in the level of detail and location specificity of
the Shepard Brook area. Nonetheless, the system of storage
and ultimate use made this the most convenient method of
collection for the total basin.
The land use data for the basin was collected in six
categories. High density single family development (density
greater than two units per acre), low density single family
development (less than two units per acre), multi-family
housing, commercial and institutional uses, industrial uses,
urban open land and nonurban land. The categories chosen
were picked because of their significance in the analysis
of Hamden. The definitions might be altered for other
communities.
A short note on our data collection method, grid cell
counting, is required at this point. There are a number of
relatively sophisticated and more accurate methods of areal
data gathering. The most commonly used system is the use of
a planimeter to measure the area within a set boundary. We
have chosen the visual grid method because we believe that
it is of sufficient accuracy for our purposes and because it
is a technique that can readily be applied by a planning
office to a land use problem such as the one under investi-
gation. The equipment requirements are no more than a sheet
of acetate, a straight edge, a permanent fine point marker
and the patience of the planner to count and record accurately
and systematically the major contents of each cell. The
method is also useful in measuring permeability, feet of
gutter and other runoff characteristics of each land use
type (see below).
A-l6
-------
Table A-2
Shepard Brook: 1961/1962 Land Use
(Percent of Cell ) _____
Residential
Multiple
Commercial
and
Institutional
Industrial Open
single Single
1 2
50
38
7.1 14.2
5
3
Grid
No.
0
1
2
3
4
5
6
7
8
9
10
1].
12
13
14
15
16
17
18
19
20
Cell
Size *
.06
.4
.4
.4
.16
.28
1.0
1.0
1.0
.4
.13
1.0
1.0
.65
1.0
.64
.17
1.0
.5
.2
.94
Non - Urban
50
100
62
100
100
79
93
94
99
100
54
9].
88
62
92
58
76
45
60
50
38
3
4
25
5
6
8
1
1
1
6
2
6
1
26
1
4
50
14
46
6
6
8
3
34
18
50
32
4
2
1
1.
17
- : ;- Basil 5% 10% 3% 2% — 2% 78%
* A full cell 1/2 mile x 1/2 mile.
A-17
-------
Table A—3
Shepard Brook 1974 Land Use
(Percent of Cell)
mile x 1/2 mile.
Grid
Cell
Commercial
and
Single
Single
Multiple
No.
Size*
1
2
Institutional
Industrial
Open
Non Urban
42.5
20
10
22
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
0
.4
.4
.32
.06
.3
1.0
1.0
1.0
.33
.19
1.0
1.0
.73
.77
.88
.1
1.0
.4
.2
.94
.06
7.5
2
2
11
30
15
100
28
3
10
11
4
8
10
7
10
16
50
100
100
100
80
77
68
100
100
58
70
55
44
70
41
36
7
13
50
10
9
4
13
42
8
35
45
30
51
60
57
68
14
50
17
* A full cell = 1/2
Total Basin
4.6%
21.3%
1.7%
4.8%
2.9%
5.8 !
58.9%
A-18
-------
The STORM model employed in our analysis requires the
determination of a number of runoff characteristics of each
of the land use types e.g., the amount of impervious land
and the number of gutter per unit of land in each land use
type must be carefully estimated for STORM. The town of
Hamden again presented a convenient study area because it
allowed us access to aerial photographs of the town, taken
in 1963 at the time of th Goodkind and O’Dea land use plan.
These photographs are at the same scale as the land use maps,
1:200. We used these to select a set of areas within each
land use type which could be examined in detail to determine
the quantity of roof surface and road and driveway surface
per unit area. Once again the process of data extraction
was not highly sophisticated. We chose a sample of cells,
each equal to 1/100 of our original one—quarter square mile
grids. We then again reduced the size of the counting grid
by 1/100, roughly to 70 square feet. At this scale it is
possible to calculate the proportion of nonpermeable sur-
faces associated with each land use type and to measure the
feet of gutter associated with each type. Because we had
used a sample of grids within each land use type, our final
coefficients for impermeable area is an arithmetic average
of the values calculated for the individual cells. The
values derived are summarized in Tables A-4 and A-5. Like
the method of deriving the proportion of each of the sub-
basins occupied by each land use type, our method of data
manipulation for permeability was straightforward and easily
applied. Previous work by those associated with the project
has shown that while more sophisticated methods yield more
accurate measurements, the nature of most land use studies,
in which only the proportion of a land use type within a
given geographic area is of concern, cannot justify the use
of methods more sophisticated than those applied here. In a
number of instances we have found that to arrive at this
gross level of information the grid method was also more
accurate given the additive nature of the measurement error
inherent in many small measurements done with a planiifleter
or other devices.
A set of assumptions have been made about the coeff i-
cients for both perviousness and for feet of gutters. One
assumption, to date untested, is that there has been no
significant change in the relative proportion of pervious
and impervious surface by land use type over the past 20
years. We believe this assumption is viable because, for
the most part, the spatial relationships have not changed
dramatically in the past years. R4 zoning is very much the
same pattern today that it was 10 years ago with the possible
difference that there are now fewer straight streets.
A further assumption is that the relationships
A-l9
-------
Table A-4
Impermeable Areas (%)
Percent
Permeable
Land Use Type Driveway Roof Road Parking Lot Open Storage Areas
1. Residential 1 3.0% 5.5% 12.5% 79.0%
2. Residential 2 9.5% 16.0% 21.3% 53.2%
3. Multiple family —— 5.0% 17.0% 16.0% 63.0%
4. Coxnmercial*/
Institutional —— 18.0% 8.0% 40.0% 34.0%
5. Industrial —— 10.0% 10.0% 4.5% 34.0% 41.0%
* Note: Includes great variation between shopping center and school.
-------
Table A—5
Feet of Gutter per 69696 square feet
(1/100th of 1/4 sq. mile grid cell).
Land Use Category Feet of Gutter Road Width
1. Residential 0 20
2. Residential 2 1484 20
3. Multiple 600 40
4. Commercial!
• Institutional 275 40
5. Industrial 350 40
* Assumes gutters on both side of road where there are gutters.
** Large size lots are “rural” in nature and do not have gutters.
A-21
-------
discovered for a given geographic area can in fact be
treated as average values to be applied over considerably
larger areas. Our analysis of Hamden for 1963 indicates
that this assumption is robust, although considerably more
sampling for other areas would be required to discover the
precise limitations to scaling of the Hamden area coeffici-
ents.
The 1974 land use data of all sub-basins are summarized
in Table A-6.
Other Aspects
The impervious areas are the main locations for accumu-
lation of dust and dirt. The town of Hamden provides some
streetcleaning. Dixwell, Whitney and State are state
streets; they are cleaned once a year. The rest of the
streets in Hamden are cleaned with brush type cleaners about
2-1/2 to 3 times a year. The efficiency of this cleaning
equipment is approximately 30 to 50 percent.
A- 22
-------
Table A-6
Summary of Mill River Subbasins
Subbasin
Total Area
(acres)
Open Space
&
Non—Urban
Developed
% Land Use
(developed
Types
area)
1
2
3
4
5
Lake Whitney West
1286.0
190.0
1096.0
0.
46.7
10.2
28.5
14.6
Lake Whitney East
734.0
296.0
438.0
0.
92.0
0.
8.0
0.
Shepard Brook
1893.0
1245.0
648.0
13.0
60.0
4.9
14.0
8.0
Centerville
& Haniden
1336.0
610.0
726.0
0.
69.1
9.6
19.0
2.2
Central Mill
869.0
688.0
181.0
24.9
23.8
0.
48.6
2.8
Eaton Brook
1467.0
1048.0
419.0
10.3
82.1
0.
7.6
0.
Northern Country
Club
272.0
165.0
107.0
28.0
62.6
0.
9.3
0.
North Eastern Mill
(Butterworth)
1600.0
1534.0
66.0
100.0
0.
0.
0.
o.
Willow Brook —
Hickory Jepp
1936.0
1608.0
328.0
78.0
15.5
0.
5.8
0.6
Cheshire Mill
3712.0
3155.0
557.0
19.9
80.1
0.
0.
0.
Willow Brook North
6192.0
5668.0
524.0
24.0
76.0
0.
0.
0.
Land Uses: 1 = single family — low density; 2 = single family - high density; 3 = multi—family; 4 = commercial;
and 3 = industrial.
-------
Appendix B
Results from Experimental Runs with STORM
Since there are no data available for calibration and
verification of STORM, our testing of the model is limited
to a logical verification, i.e. checking the consistency and
logic of results in response to single or simultaneous
changes of parameters and input variables. Furthermore we
attempted to compare STORM’s results with literature values.
We limit our discussion to the runoff part of STORM
(quantity and quality). The universal soil loss equation
(USLE) and its results have been discussed frequently.*
Because the calculation of erosion is not influenced by
results generated in the runoff part, but only by the
rainfall input, the performance of the universal soil loss
equation is independent of the performance of the other
STORM segments.
The following parameters and input variables determine
the performance of STORM’s runoff computations:
1. Runoff coefficients for urban and non-urban areas;
2. Depression recovery coefficients;
3. Size of the area for which runoff is predicted;
important questions are concerned with the
resolution potential of the model when applied to
small areas, and with the areal size beyond which
the neglect of routing leads to obvious errors;
4. Number of land uses distinguished; data require-
ments, computer time, etc., are significantly
affected by this choice;
5. Length of gutter/acre for each land use;
6. Accumulation rate of dust and dirt and composition;
7. Street sweeping intervals and efficiency of
sweeping equipment;
8. Percent imperviousness of land devoted to each
land use;
* See, for example, the update of the User’s Manual of
SWMM, University of i 1 lorida, Gainesville, Feburary, 1975.
B-l
-------
9. Precipitation data, which is particularly important
when no rain gages exist near the area being
investigated.
Because quantity and quality of runoff are simultane-
ously considered, it must be noted that all issues and
parameters which affect quantity also affect quality in
STORM. Parameters (5), (6) and (7) above affect STORM’s
calculations of quality alone.
A number of computer runs have been made to explore the
sensitivity of the model to these variables and parameters.
The first runs were based on modification of the original
test data, as supplied by the Hydraulic Engineering Center,
Davis, California, while subsequent runs utilized actual
data: rain data from the airport rain gage at Bridgeport,
Connecticut and land use data of Shepard Brook, one of the
sub-basins of the Mill River. Detailed land use data of
the 2.5 square mile sub—basin were collected from available
land use maps.*
The following types of runs have been made:
1. Change of rain intensity;
2. Variations of runoff coefficients;
3. Change of percentage of imperviousness for each
land use;
4. Different aggregation of land uses for continuous
development;
5. Use of STORM on small individual developments; and
6. Change of accumulation rates of dust and dirt and
their composition.
For most of the runs the same meteorological condition
was employed, namely the 1974 precipitation record at the
Bridgeport Airport. We have evaluated in detail up to five
rain intervals and generated pollutographs on an hourly
basis, and we have tried to obtain a representative sample
of events, including brief, intensive rains with short and
long dry antecedent conditions, as well as long and less
intensive rains with long and short dry antecedent conditions
(see Table B—i).
* See Appendix A.
B-2
-------
Table B-].
Rain Interval of Interest
:nterval
Date
Dry
Antecedent
Conditions
(hours)
Length
of
Interval
(hours)
Intermediate
Dry Period
(hours)
Rain (inches)
Event*
1 2
1
2
3
4
5
Jan. 21, 74
June 16, 74
Sept. 3, 74
Oct. 31, 74
D cc. 1, 74
112
367
35
127
256
32
3
7
6
11
13
(after 10 hrs)
—
—
2
(after 2 hrs)
—
1.63
.51
1.11
.08
2.84
.56
.45
* When a rain interval was interrupted by a dry period, the part before
the dry period is called event 1, and the one afterward event 2.
Table B—2
Characteristics of Hypothetical Developments
area
(acres)
% imperviousness
gutter length/acre
residential
commercial
industrial
160
123.2
160
36.3
13.7
22.7
600
172
219
B- 3
-------
To simplify the presentation, two sets of test runs
were chosen for detailed discussion. The first set comprises
simulation of runoff from developments at a subdivisional
level, the second runoff from the Shepard Brook sub-basin
incorporated those developments. The second set of runs is
also employed to compare different aggregations of land—use
types.
In the first subset some relatively undeveloped areas
of Shepard Brook were picked and subdivisions for develop-
ment designed (see Table B-2 on p. B-3).
1. Residential single—family development (lot sizes
smaller than 1/2 acre) (res);
2. Commercial development of a shopping center (corn);
and
3. Light industrial development (md).
The base year is 1974. Table B-3 shows the resulting
runoff and washoff for the rain events of interest. The
calculations are at present limited to suspended solids
(SS), settleable solids, BOD, nitrogen, and P0 4 emissions,
because the recoding of STORM to calculate coliforms was
not finished when these runs were made. Table B—3 reveals
the importance of the dry antecedent conditions. Comparing
intervals 1 and 2 in terms of washoff and runoff, but
ignoring intensity of rain, we recognize that the washoff
of suspended solids* in interval I is only about 1.25 times
as high as in interval 2, even though the runoff of 2 is
less than a quarter of the runoff of 1. But the number of
dry hours preceding the rain interval is 112 hours in the
former case and 367 hours in the latter. Rain interval 3
confirms these findings. its resulting runoff of 0.27
inches generates only some 2,400 pounds compared to some
4,700 pounds generated by 0.11 inches of runoff (interval 2)
because interval 3 was preceded by a dry period of only 35
hours. Clearly, an important consideration is the length
and intensity distribution of the rain in each interval.
Interval 2 is only 3 hours long with 1 hour of high inten-
sity, while 1 is more evenly distributed over a much longer
period. The intensity of the rain is directly related to
the washoff. The 1-hour peak washoff of interval 2 com-
prises about three-quarters of the total washoff.
* Our discussions are largely limited to suspended solids
(SS), because discussing all pollutants at the same time
is more confusing than enlightening. The tendency is
basically the same.
B- 4
-------
( ‘I
* urban area: 648 acres; non-urban: 1245 acres.
Table B-3
Runoff and Washoff
Land Use 1974 of Shepard Brook*
Figures in parentheses are the hourly peak runoff (cfs) and maximum SS washoff, respectively of each rain-
fall interval.
Interval
Length
(hours)
Rain
(inches)
Runoff 1
(inches)
SS 1
(ibs)
Settleable Solids
(ibs)
BOD
(ibs)
N
(ibs)
P0 4
( lbs)
Total
39.4
8.73
131,922
16,368
19,599
6,786
916
1
32
2.19
.54
(355.1)
6,061
(3,231)
395
759
318
34
2
3
.52
.11
(150.0)
4,729
(3,658)
500
732
240
37
3
7
1.11
.27
(130.0)
2,376
(909.8
486
324
113
21
4
6
.53
.12
(208.0)
3,495
(3,355)
363
484
178
22
-------
Table B-4 summarizes the runoff and washoff of the 1974
total and of four of the five rain intervals in the three
hypothetical developments. The residential area has the
highest percentage of imperviousness and the greatest
gutter length per acre; thus even though its accumulation
rate of dust and dirt is quite low, the washoff from that
area is only about two-thirds of the washoff generated in
the commercial and industrial areas, while total runoff is
necessarily greater. The washoff rates from the commercial
and industrial areas are quite close. The comparability of
these two developments is based on the fact that even though
the industrial development has higher total dust and dirt
accumulation rates and a larger rate of imperviousness, the
commercial development has a higher proportion of the dust
and dirt accumulation in the components considered. The
runoff and washoff rates reveal that each single development
might have a significant impact on a sub-basin.
In order to make comparative runs between sub—basins
without additional development and sub-basins with addi-
tional development, the question of how to incorporate the
developments into the total scheme has to be solved. In
particular, the handling of the percentage of imperviousness
for each affected land use has to be considered. Since the
new developments do not have exactly the imperviousness of
the existing land use type in the sub-basin, either the land
use type’s overall imperviousness has to be adjusted, or the
new development’s rate of imperviousness has to be adjusted
to the figure of its land use type by taking away some of
the pervious area and adding it to the category of open
space, parks and non-urban area. These options have
different impacts on the generated results, since land use
types exhibit different accumulation rates. Imperviousness
affects the runoff and the washoff of each category. In
order to document the arising problems, the aggregation is
done both ways. Table B-5 shows the results for the
industrial area.
The first array of events is based on the original
industrial development except that 37 acres are deleted.
These have not been developed and are separated from the
core industrial development by a railroad; these acres are
added to the non-urban category. Since the imperviousness
of the industrial area does not correspond exactly to the
overall imperviousness of the industrial land use category
(Table A—4), the latter is adjusted, i.e., actually reduced.
In the second array, the size of the industrial development
is reduced so that the new development’s imperviousness is
equal to the overall industrial imperviousness. This
procedure cannot always be done because a new 160 acres
development might have such a high imperviousness that it
B-6
-------
Table B—4
p noff and Washoff
Individual Develoç*nents
Length ss Settleable Solids SOD N
Interval (hours) (inches) (inches) (lbs) (ibs) (ibs) (lbs) (ibs)
‘rotal 394 13.35 24,345 2,303 3,600 1,331 108
.4
1 32 2.19 .84 1,335 80 167 70 7
4 J
2 3 .52 .18 809 47 126 45 5
3 7 1.11 .42 355 64 46 19 2
4 6 .53 .2 701 59 105 39 4
5 1]. 2.78 1.12 2,435 445 295 130 13
w Total 39.4 8.13 38,784 3,709 5,759 2,066 179
1 32 2.19 .52 2,128 128 265 110 11
- ‘I
2 3 .52 .11 1,289 76 202 69 7
o 7 1.11 .26 566 103 73 30 3
4 6 .53 .12 1,117 95 168 59 6
5 11 2.78 .68 3,880 716 471 205 21
Total 39.4 10.21 35,233 3,513 5,217 2,024 165
‘-I
1 32 2.19 .68 1,933 121 240 104 10
‘4
2 3 .52 .14 1,171 72 183 68 7
3 7 1.07 .32 514 97 67 29 3
4 6 .53 .16 1,014 90 152 59 5
5 11 2.78 .86 3,525 680 427 192 19
-------
Table B-5
Runoff and Washoff
Comparison of Different Incorporation of Industrial Development
Length Rain SS Settleable Solids BOO N
Interval (hours) (inches) (ibs) (ibs) (ibs) (ibs) (lbs)
Total 4 39.4 158,517 18,703 23,524 8,377 1,039
1 32 2.19 7,637 491 954 403 42
2 3 0.52 5,579 531 865 293 41
3 7 1.11 2,719 546 367 134 22
4 6 .53 4,292 427 645 233 27
Tota1. 39.4 142,501 17,248 21,157 7,381 966
1 32 2.19 6,684 431 837 350 37
2 3 0.52 5,068 512 785 259 39
3 7 1.11 2,514 509 342 121 21
4 6 .53 3,811 •387 573 203 25
+ urban area: 694 acres; non—urban area: 1199 acres.
urban area: 771 acres; non-urban area: 1122 acres.
-------
could not be adjusted to the overall proportion. The first
arrangement yields higher washoffs; this is explainable from
the fact that even though the rate of imperviousness is
lower than in the second arrangement, the factor, gutter/acrc
which influences dust and dirt accumulation, is kept constant
Thus the second arrangement yields a lower washoff rate
because the reduction of acreage also reduces the total
gutter area. Thus if the largely pervious land use category
(i.e., parks, open space and non-urban) could be divided
into a number of different pervious land uses, the analysis
could be better tuned; but computer costs and manpower
resources (to collect the necessary data) increase steadily,
so that the non-urban land uses should only be subdivided
if size and diversity warrant such an extension of the
analysis. *
Both types of runs are also made after incorporation
of all the hypothetical developments in Shepard Brook
(Table B-6). The first set of results is obtained by keep-
ing the original development areas and adjusting the overall
imperviousness, while the second is obtained by adjusting
the development areas such that the assumed overall imper—
viousness rates are met. These runs are of interest because
the addition of commercial and industrial areas could be
easily envisioned after a large residential development
had been built, or vice versa . The planner might work
through these possibilities for additional interpretation
of the impact of each single development. Table B-6 shows
that total emission increases significantly beyond the
emission level of the base year 1974 (see Tables B-3 and
B-6). The actual difference in the washoff rate is influ-
enced by the method of aggregation applied; this is parti-
cularly true for SS and BOD. The results confirm the
tendency emerged in Table B_5.**
Now another type of comparison is presented. The
simulation is run separately for individual development,
for example, residential or commercial, and for the entire
sub-basin without the development. Another run is then
made for the entire sub-basin including the new development.
* Considering the accuracy of the model, we do not feel
the differences are very significant in this case.
** Runoff has not been considered because the program
yields just results in inch/area, but not the total amount,
adjusted to the ratio of urban to non-urban areas.
B— 9
-------
Table B—6
Runoff and Washoff
All Hypothetical Developnents Incorporated in Shepard Brook
Length Rain SS Settleable Solids BOD N P0 4
Inteiial (hours) (inches) (ibs) (ibs) (lbs) (ibs) (lbs)
Total 39.4 212,263 23,041 31,466 11,386 1,286
1 32 2.19 10,830 678 1,352 570 58
2 3 .52 7,293 587 1,134 390 48
3 7 1.11 3,409 660 453 174 24
4 6 .53 5,902 550 888 322 35
Total 39.4 167,835 19,153 24,900 8,868 1,072
32 2.19 8,250 522 1,031 434 45
2 3 .52 5,858 526 909 308 42
3 7 1.11 2,816 556 378 140 22
4 6 .53 4,586 444 690 248 28
+ xrban area: 1031 acres; non-urban area: 861 acres.
• urban area; 880 acres; non-urban area; 1013 acres.
-------
It has been felt that a desirable property of the model is
linear additivity. That is, the sum of emissions as well
as runoff calculated for the development and basin indivi-
dually should equal (approximately) the total emissions
and runoff, respectively predicted by aggregate analysis.
Table B-7 indicates that this is indeed the case.
There are also runs performed for the 1974 land use
utilizing the accumulation rates and composition derived
from the calibration exercises in Castro Valley.* Table
B-B reveals the necessity of good calibration or intuitive
derivation f the parameters because increased accumulation
rates significantly influence the runoff quality of an area
(compare Tables B-3 and B-8).
Table B-9 displays portions of pollutographs of the
intervals 1 and 2. It shows how the washoff is closely
related to the intensity of the rain and to the time elapsed
from the start. It also reveals that the total load washed
into the receiving water body is of a significant size,
while, with a few exceptions, the concentrations of all
pollutants except SS are not very high. The concentration
of BOD 5 in event 2 is close to the BOD 5 of a secondary
treatment plant’s effluent. These pollutants are based on
the Chicago values for dust and dirt accumulation and
composition; it is obvious that the pollutographs for these
events would be characterized by higher concentration if
the Castro Valley parameters are employed.
Finally, some runs are made to compare the impact of
different Street sweeping strategies. The frequency of
sweeping is increased to once a week in commercial and
industrial areas and to once every two weeks in residential
areas. In a second run the sweeping efficiency of the
equipment was improved in addition to the increased
frequency of sweeping. Comparing the washoff from Shepard
Brook for the prevailing situation** (Table B—3) with the
results in Table B—1O shows that these strategies cause a
* See section 5, Table 5-8; data are reported in Roesner,
et al . , “A Model for Evaluating Runoff-Quality in Metro-
politan Master Planning,” ASCE Urban Water Resources
Research Program, Technical Memorandum No. 23, April, 1974,
p. 62.
** See Appendix A.
B-li
-------
Table B-7
Runoff and Washoff 1
Check on Linear Additivity
Runoff SS Settleable Solids BOD N P0 4
Land Use (inches) (ibs) ( lbs) (ibs) ( lbs) Ciba)
Ri 8.56 128,802 15,590 19,119 6,676 867
R2 13.35 24,345 2,303 3,600 1,331 108
153,147 17,893 22,719 8,007 973
R3 8.9* 150,851 17,693 22,371 7,902 992
r.J
1
All results are totals for 1974 precipitation.
Ri 1974 land use (urban: 648 acres; non—urban: 1245 minus 160 acres).
R2 Individual residential development of 160 acres.
R3 1974 land use incorporating residential development (urban: 808 acres; non—urban: 1085 acres).
* Runoff is compared by weighting the runoff with the corresponding areas;
8.56 * 1733 + 13.35 * 160 — 8.9 * 1893
16985 — 16850
-------
Table 8-8
Runoff and Washoff 1
Shepard Brook 1974: Increased Accumulation Rate of Dust and Dirt
Length Rain Runoff SS Settleable Solids BOD N
Interval (hours) (inches) (inches) (lbs) (lbs) (ibs) (ibs) (ibs)
Total 39.4 8.73 302,56]. 44,195 69,969 16,693 2,921
1 32 2.19 .55 15,976 2,002 3,090 824 150
(.05) (1,566) (258) (649) (103) (22)
w 2 3 .51 .12 8,402 924 2,168 494 78
3 7 1.07 .28 5,907 1,008 1,037 293 56
(.1) (2,882) (473) (568) (149) (28)
4 6 .53 .13 7,070 867 1,731 417 62
(.01) (361) (66) (20) (29) (6)
5 11 2.84 .73 31.457 6,254 5.206 1,621 263
(.09) (5,329) (710) (1,543) (315) (58)
Figures in parentheses reflect runoff and washoff in the first four hours of the interval; interval 2
lasted less than four hours.
-------
Table B-9
Hourly Runoff and Washoff from New Residential
Develojinent Incorporated in Lande Use 1974*
(Urban: 808 acres Non-urban: 1,085 acres)
Hours of
Date runoff - Rain Runoff ss Settleable Solids BOD N 4
(begin) from- start (inches) (inches) (lbs) (lbs) (ibs) (ibs) Ciba)
January 3 .06 .02 154.9 6.3 33.3 9.4 1.0
21/74 (30.8) (22.4) (0.9) (4.8) (1.36) (0.15)
(9 a.m.) 4 .08 .02 228.1 8.7 41.0 13.1 1.4
(41.0) (24.7) (0.9) (4.4) (1.42) (0.15)
5 .31 .08 1655.5 64.9 201.8 86.3 8.9
(159.0) (46.3.) (1.8) (5.6 (2.42) (0.25)
6 .12 .03 294.7 12.8 35.5 15.2 1.6
(61.5) (21.3) (0.9) (2.6) (1.10) (0.12)
7 .71 .19 3420.8 265.6 357.5 173.9 17.8
(364.1) (41.8) (3.2) (4.4) (2.13) (.22)
8 .09 .02 37.2 5.1 4.1 1.9 0.2
(46.1) (3.6) (0.5) (0.4) (0.18) (0.02)
9 .07 .02 24.9 3.7 2.8 1.2 0.2
(35.9) (3.1) (0.5) (0.3) (0.15) (0.02)
June 1 .2 .03 1016 78 210 59 7.5
16/74 (50.1) (90.3) (6.9) (18.7) (5.25) (0.67)
(6 p.m.) 2 .3 .08 3800 411 535 188 29.5
(153.8) (109.3) (11.9) (15.5) (5.44) (0.9)
3 .02 .01 101.6 18.4 16.6 4.6 1.1
(10.3) (44.1) (8.0) (7.2) (2.01) (.48)
* Figures in parentheses describe the peak runoff (in of s) and the resulting concentration (ing/1),
respectively.
-------
Table B.-l0
Comparison of Street Cleaning Strategies
Shepard Brook: Land Use 1974*
SS Settleable BOD N p0
( lbs) Solids (ibs) (ibs) (ibs) (ibs)
Total+ 109,143 13,043 16,653 5,204 766
January 4,163 278 526 211 23
June 3,078 392 388 93 23
September 2,225 409 397 146 24
October 2,814 252 259 45 9
I-I
December 10,293 1,742 1,280 500 72
Total 96,024 11,516 14,799 4,524 688
January 2,923 191 390 149 17
June 2,090 336 284 66 20
September 2,147 374 380 137 23
October 2,426 203 243 41 9
December 9,227 1,473 1,153 444 67
* See Table B-3 for comparison of existing strategy.
+ Increased sweeping frequency: ceminercial and industrial areas: 1/wk; residential areas: 1/2 wka.
++ Additional increase of sweeping efficiency: 70%.
-------
significant reduction in washoff.*
Table B—il compares our 1974 Shepard Brook yields
(pounds/acre/year), based on the Chicago values of accumula-
tion and composition, to values reported for Cincinnati and
San Francisco. It is difficult to compare loading figures
of different cities for different years, when no details
are known about the way these figures were arrived at. Such
a comparison is to develop a feeling for the orders of
magnitude involved. Except for all SS yields and the BOD
yields from San Francisco, values from the urban part of
Hamden are comparable to the other cities in order of
magnitude.
In general, we believe that our limited number of runs
demonstrate the potential of STORM as a planning tool for
estimating the order of magnitude of pollution, but also
indicate that it does not come close to the accuracy
required of a design tool.
* Based on these results, an economic comparison of
various strategies for reducing the load of washoff to
receiving streams might be undertaken.
B- 16
-------
Table B-li
Comparison of Results from STORM
and Literature
Sett leable
SS Solids BOD N P0 4
Total (ibs)
Shepard Brook
131,922
16,368
19,589
6,786
916
Urban
Total (lbs)+
107,681
10,298
16,101
5,443
533
Urban Yield++
(ibs/acre/yr)
166
16
25
8.4
0.9
Cincinnati*
(ibs/acre/yr)
730
—
33
8.9
2.5
San Francisco**
(ibs/acre/yr)
632
—
101
10.6
2.4
540
—
136
15.6
3.2
+ The total yearly yields and the total yearly urban yields
from Shepard Brook are computed by STORM for 1974.
++ 648 acres are urban acres (Land Use 1974); dividing the
total urban yields by the urban area gives the urban yield
in lbs/acre/yr.
* Weibel, et al. , Urban Land Runoff as a Factor in Stream
Pollution, J. of Water Pollution Control Federation , Vol. 36,
1964.
** Eckhoff, D. W., A. 0. Friedland, and H. F. Ludwig,
Characterization and Control of Combined Sewer Overflows,
San Francisco, Water Research , Vol. 3, No. 7, July 1969;
note: figures for two sampling areas are presented in this
paper.
B-17
-------
Appendix C
Review of Control Options for
Storinwater Management by STORM*
Controls of watershed runoff and the pollutants con-
tained in that runoff take many forms and can be applied
over a wide range of drainage areas. For example, in the
control of storinwater runoff, one can apply strategies
ranging from ponding in individual house lots to the con-
struction of large basins serving many square miles. A
major problem, then, is to delineate those options which
are candidates for inclusion in the planning analysis. It
appears that the choice of control options must be based on
answers to the following questions:
1. Which options will produce a significant difference
in the runoff model, i.e., STORM? This question
relates in part to selection of the scale of devel-
opment. In any case, it is clear that it makes no
sense to include options to which the model is
relatively insensitive.
2. What are the major tradeoffs? Theoretically
the stormwater control decision represents a
tradeoff between investment in storage and
transport (conveyance systems), while sediment
control decisions involve tradeoffs between
investment in reducing erosion and removing sedi-
ment (resulting from erosion). But within these
broad categories there remains a large number of
options, many of which do not fit into simple
categorization. The availability of cost data
for all possible options is also a problem. Since
major interest in non-point sources is fairly
recent, the literature is still relatively sparse
on technical and cost details.
3. What are the design standards for stormwater
erosion and sediment control facilities? At one
point in the project, a design goal of maintaining
* This appendix is a summary of a lengthy memorandum on
non-point source controls (7 April, 1975).
C-i
-------
the natural unit hydrograph was discussed. The
EPAI* on the other hand, talks about maintaining
the peak of the natural unit hydrograph, but not
the total quantity of runoff. Also, there is a
growing tendency on the part of states and muni-
cipalities to place higher standards on the prac-
tices of developers with regard to non-point
pollution, particularly sediment.** Ideally an
analysis could be devised which indicates the
proper level of performance standard to place on
developments; realistically, however, only a few
combinations of control options can be considered
in an analysis, based on our suggested tools.
Thus, it will be important, for instance, for a
planning department to establish a reasonable
number of control options; one of which should
then be chosen as a performance standard for the
particular development.
We have briefly summarized a number of the available
control methods and suggest how they may be incorporated
within the model framework of STORM. In listing some of
the available physical control options, we have considered
only two objectives: control of sediment and control of
stormwater runoff rates. There are a number of other
pollutants associated with stormwater runoff which present
serious problems for water quality. However, at the present
time the methodologies exist for handling only the two
components mentioned. To some extent these methods also
help to control the other pollutants, but we do not expect
these problems to be eliminated simply because sediment and
runoff rates are adequately controlled.
In our summary we have identified three functional
areas of control within the two objectives: erosion con-
trol, sediment removal, and stormwater control. By erosion
control we mean measures designed to prevent the removal
* U. S. Environmental Protection Agency, “Processes, Pro-
cedures and Methods to Control Pollution Resulting from All
Construction Activity,” October, 1973.
** In Virginia, for example, developers must submit and
gain approval of a sediment control plan prior to initiating
development.
C- 2
-------
and suspension of soil particles by storm events. Sediment
removal processes are designed to remove sediment particles
after suspension has occurred. Stormwater control refers
to any method for altering the timing, quantity or rate of
runoff from a site. To some extent this division is arti-
ficial, since there are clearly interactions between the
different types of controls. Erosion and sediment control
techniques are complementarY; any comprehensive strategy
must consider and include measures from both sets of controls
in order to obtain acceptable performance. Similarly, many
erosion and sediment—control methods alter the quantity or
timing of runoff and therefore act as stormwater control
devices as well. Despite these interactions, we have found
it useful to adopt the present classification scheme from an
EPA* report on control of non-point source pollution.
Each of the tables presents a different set of control
methods: Table C—i for stormwater control, Table C—2 for
erosion control, and Table C-3 for sediment control. For
each method we have listed specific control mechanisms (by
mechanism we mean a physical effect such as detention,
infiltration, reduction of velocity, etc.), the secondary
impacts of the method, and some adjustments of STORM which
would reflect the particular control method.** Clearly,
in a crude model, such as STORM, various combinations of
adjustments would constitute a particular control method.
Therefore, a list can never be complete, but only suggestive
to an experienced planner.
* Ibid .
** Note, an S in parentheses indicates that parameters of
the runoff module are changed; a U indicates that parameters
of the erosion module, i.e., the Universal Soil Loss Equa-
tion, are adjusted.
C-3
-------
Table C—i
Stormwater Control Methods
Principal
Control Mechanisms Method of
Secondary Modeling in
Method Detention Infiltration Impaôts STORM
Rooftop Storage x increased costs of storage option(s)
roof construction
Parking Lot — X inconvenience to users storage option(s)
Storage of facility, sediment trapping ratio(u)
sediment control
Parking Lot — X sediment control adjust runoff coef (s) 1
Infiltration
sediment trapping ratio(u)
C) Detention Basins X sediment control, possi- storage option for dry de—
wet/dry ble recreation use, tention basins(s); also
multi/single aesthetic problems, for wet basins in case of
purpose potential dangers for similar control rules(s)
small children, high
maintenance costs
On lot ponding X X sediment control, storage option(s)
and seapage inconvenience to diversion option(s)
systems homeowner sediment trapping ratio(u)
Storage in sewers x greater complexity of storage option(s)
operating system, ground-
water pollution
Porous x x potential groundwater depression storage(s)
Pavements pollution diversion option(s) 2
adjust runoff coeff. (s)
Adjusting the runoff coefficient seems to be justifiable only when the runoff modeling is limited to
the commercial (or industrial area) with the parking lot as major portion of the total area.
2 Only justifiable when the porous pavement makes up a major part of the “impervious” area.
-------
Table C-2
Erosion Control Methods
Principal
Control Mechanisms
Method of
Soil Flow Secondary Modeling in
Method Stabilization Modification Diversion Impacts STORM
surface X increases infil— diversion factor(s)
roughening tration, provides soil erodibility factor(u)
base for mulching control practice factor(u)
Diversion X diversion option(s)
ditches control practice factor(u)
Chutes and x diversion option(s)
Flumes control practice factor(u)
Level X control practice factor Cu)
spreaders
Vegatative diversion option(s)
Stabiliza- X X increases inf ii— cover index factor(u)
tion tration runoff coeff.(s)
Non—vega— X x mulch increases diversion option(s)
tative sta— infiltration, cover index factor(u)
bilization inorganic binders cover index factor
decrease inf ii— runoff coeff.(s)
tration
G ade
stabiliza- X slope-lenth factor(u)
tion
-------
Table C-3
Sediment Control Methods
Principal
Control Mechanisms Methods of
Secondary Modeling in
Method Settling Filtration Impacts STORM
Storm drain
filters x x frequent cleaning sediment trapping ratio(u)
traps x required diversion option(s)
trav or x control of cover index factor(u)
ay bales peak runoff sediment trapping ratio(u)
storage option(s)
arth dikes X x control of sediment trapping ratio(u)
peak runoff storage option(s)
diversion opt ion (s)
;ediment Basin control of sediment trapping ratio Cu)
vet/dry x peak runoff, storage Optjon.(s)1
may be converted
to permanent storm—
water control basin
1 Largel.y for a dry basin
-------
Appendix D
Lateral Sewer Cost Estimates
Tables D-1 through D-lO detail the sanitary sewer cost
estimates for our ten hypothetical developments. These
costs are broken down into several major subcomponents:
the costs listed for street pipes and service lines include
purchase of the pipe, transportation costs, and costs of
installation in the trench; the house connection costs refer
to the tee and y sections and elbows used to connect the
service lines to the lateral and include both purchase and
installation; sheeting costs are the costs of installing
wood bracing and sheeting in the trenches; manhole costs
cover the purchase and installation of manhole sections
and cover; and excavation costs cover both the costs of
digging the trenches and backfilling them after the lines
are installed.
The high-density development (D-l0) has by far the
lowest costs of the developments analyzed; however, this
result somewhat exaggerates the economies of high-density
development. In the large 20-story apartments the horizon-
tal sewer collection system has been replaced by an internal,
vertical infrastructure. The costs of this extensive
plumbing network are not reflected in our estimates. Also
our estimates do not include the higher costs imposed on
trunk and interceptors by the larger loads generated; thus
the data on subdivisions should not be used directly for
economic comparisons among different development types.
This is the purpose of the fiscal impact model developed in
this study.
The proportion of total costs which is attributed to
each of the major components varies significantly depen-
ding upon the nature of the topography. This variation is
made clear in the summary graph of Figure D—l. Here we
have collapsed the cost breakdowns into four components:
pipes (including connecting sections), sheeting, manholes,
and excavation. The graphs are based upon the averages of
projects in the associated slope categroies.* For the flat
and steep slopes sheeting and bracing costs are by far the
most important, accounting for about 60 percent and 54
percent of the total costs, respectively, while pipe costs
are about 25 to 30 percent of the total. In the moderate
* Project 10 was not included in these calculations
because its design is so different from the others.
D-l
-------
slope case the situation is reversed, with pipe costs
accounting for over 60 percent of the total and sheeting
reduced to only 13 percent. The difference, of course, is
due to the shallow trenches possible with favorable topo-
graphy, eliminating the need for extensive bracing.
Manhole and excavation costs also account for a higher
proportion of the total costs in the moderate slope case
for the same reason (but in absolute terms excavation costs
are greater for the flat and steep topography; it is just
that they are overwhelmed by the sheeting component).
Overall, manholes account for 9-16 percent of the total,
excavation for 4-9 percent.
D- 2
-------
TABLE D1 - COST SU 4ARY - DESIGN 1
540 single Family,
180 Acres
F].at Slope
Conventional units
ITEI4
UNITS
QUANTITY
f
COST ($)
street
pipe
LF
28,745
99,752
service
lines
LF
16,200
43,902
house con
-nections
t
540
35,489
sheeting
SF
507,780
756,490
(492, 610)
anho1os
I
152
•
83,501
SUBTOTAL
1,019,134
(755,254) *
* Hard Clay and Shales.
Excavation
47,386 CY
(36,311)*
loam, sand,
gravel -
compacted gravel
till
TOTAL COSTS (5)
cOSTS/DU ($)
1,060,351
1,964
1,061,208
1,965
788,053
1,459
-------
ThBLE D—2 — COST SU? 1ARY - DESIGN 2
471 single Family, Conventional
160 Acres
Moderate Slope
ITEM
UNITS
QUANTITY
COST ($)
.
Street
pipe
LF
28,410
96,310
service
lines
LF
16,485
44,674
house con
-nections
#
471
17,705
sheeting
SF
22,661
41,373
(10,83l)
manholes
1
101
44,226
SUBTOTAL
244,288
(213,746)*
* Hard clay and shales.
Excavation = 43,151 CY
(24,405)*
loam, sand, loose -
gravel
compacted gravel and
till
hard clay and shales
TOTAL COSTS Cs)
0STS/DU ($)
280,103
595
281,398
597
235,222
499
-------
TABLE D—3 — COST SU 4ARY - DESIGN 3
ITEM
UNITS
QUANTITY
COST ($)
street
pipe
LF
28,410
96,310
service
lines
LF
16,485
44,674
house con
-nections
t
471
20,226
sheeting
manholes
SF
t
153,753
101
293,669
(2 05,316)*
51,800
SUBTOTAL
506,679
(418,326)*
loam, sand, loose
gravel
compacted gravel and
till
hard clay and shales
TOTAL COSTS Cs)
cOSTS/DIJ (5)
545,047
1,157
546,853
1,161
444,217
943
471 Single Family, Conventional
160 Acres
Steep Slope
V
U ’
* Hard Clay and Shales.
Excavation
45,139 CY
(28,768)*
-------
ITEM
UNITS
QUANTITY
COST ($)
Street
pipe
LF
31,710
111,076
service
lines
LF
19,115
52,046
house con
-nections
#
1,023
66,536
sheeting
manholes
SF
#
395,268
139
754,961
(442,545)
72,825
*
SUBTOTAL
1,057,444
(745,028)
*
loam, sand, loose
gravel
compacted gravel and
till
hard clay and shales
TOTAL COSTS ($)
COSTS/DU Cs)
1,096,131
900
1,097,687
901
785,876
645
TABLE D—4 — COST SU 1ARY - DESIGN 4
359 Single Family, Conventional
652 Townhouses
207 Garden Apartment units
Flat Slope, 160 Acres
* Hard Clay and Shales.
Excavation = 45,451 CY
-------
TABLE D-5 - COST SU? 1ARY - DESIGN 5
553 single Family, Compact
615 Townhouses
396 Garden Apartment units
Moderate Slope, 205 Acres
loam, sand, loose
gravel
compacted gravel and
till
hard clay and shales
TOTAL COSTS Cs)
COSTS/DU (5)
491,548
314
439,596
316
378,981
242
¶2
-4
ITEM
UNITS
QUANTITY
COST ($)
street
pipe
LF
33,975
120,334
service
lines
LF
38,415
105,152
house con
-nections
t
1,182
24,054
sheeting
SF
64,398
-
123,000
(32,199)*
manholes
•
I
138
62,421
SUBTOTAL
434,961
C344,160)*
* Hard Clay and
Shales.
Excavation = 68,181 CY
(39,570) *
-------
TABLE D—6 - COST SU?’NARY - DESIGN 6
367 Single Family, Conventional
590 Townhouses
300 Garden Apartment units
Steep Slope, 160 Acres
.
loam, sand, loose
gravel
compacted gravel and
till
hard clay and shales
TOTAL COSTS Cs)
cOSTS/DU ($)
1,143,472
:910
1,145,875
912
959,014
763
ITEM
UNITS
QUANTITY
COST ($)
Street
pipe
LF
30,740
104,209
service
lines
LF
27,300
74,051
house con
-nections
1
967
60,302
sheeting
manholes
SF.
395,308
146
755,038
(583, 372) *
97,617
SUBTOTAL
1,091,217
(919,551) *
1aLI.& ciay
an Shales.
Excavation —
60,603 CY
(43, 365) *
-------
TABLE D-7 - COST SU 1ARY - DESIGN 7
ITEM
UNITS
QUANTITY
COST ($)
street
pipe
LF
22,075
98,022
service
lines
LF
20,020
54,853
house con
I
•
585
48,169
-nections
sheeting
SF
350,552
77
.669,555
(445,301)*
45,676
manholes
SUBTOTAL
916,275
(692,021) *
loan, sand, loose
gravel
oinpacted gravel
till
TOTAL COSTS Cs)
cOSTS/DIJ Cs)
953,675
442
955,394
443
721,568
335
300 single Family, Compact
276 Townhouses
540 Garden Apartment units
1040 Medium Rise Apartment units
Flat Slope
160 Acres
w
* Hard Clay and Shales.
Excavation — 42,789 CY
(32,469)*
-------
Excavation • 52,058 CY
(30,390)*
loan, sand, loose
gravel
compacted gravel and
till
hard clay and shales
TOTAL cOSTS ($)
COSTS/DU ($)
422,267
153
423,828
153
325,164
118
TABLE D—8 — COST SUMMARY - DESIGN 8
359 Single Family, Compact Moderate Slope
417 Townhouses 205 Acres
696 Garden Apartment units
1290 Medium Rise Apartment units
0
ITEM
UNITS
QUANTITY
COST ($)
street
pipe
LF
28,980
108,067
service
lines
LF
25,630
70,426
house con
-nections
I
807
30,448
sheeting
s
61,190
116,872
(36,235)*
manholes
1
116
53,225
SUBTOTAL
379,038
(298,401)*
•* Hard Clay and Shales.
-------
TABLE D-9 — COST SUMMARY - DESIGN 9
227 single Family, Conventional
233 Townhouses
576 Garden Apartment units
1120 Medium Rise Apartment units
Excavation 38,890 CY
(23,225)*
loam, sand, loose
gravel
conipacted gravel and
till
hard clay
TOTAL COSTS (5)
cOSTS/DU (5)
356,683
165
357,850
166
279,389
130
I - .
I -i
lIEN
UNITS
QUANTITY
COST Cs)
Street
pipe
I.E
.
27,265
83,117
service
lines
I.E
16,009
44,637
house con
-nections
1
486
18,269
sheeting
manholes
SF
I
74,519
76
142,331
(77, 443) *
35,272
SUBTOTAL
323,626
(258, 738) *
iiara i ;
ay and Shales.
-------
TABLE D-l0 — COST SIJI’L’IARY - DESIGN ]
ITEM
UNITS
QUANTITY
COST ($)
Street
pipe
LF
2,380
28,441
service
lines
LF
440
2,763
house con
-nections
#
connected at
manholes
sheeting
SF
10,661
19,417
manholes
It
12
(5, 083) *
5,653
SUBTOTAL
56,274
(41,940) *
loa n, sand, loose
gravel
compacted gravel and
till
hard clay and shales
TOTAL COSTS ($)
COSTS/DU ($)
58,735
15
58,824
15
43,581
11
4000 High Rise Dwelling Units
160 Acres
Flat Slope
I -J
* Hard Clay and Shales.
Excavation = 2,964 CY
(1,844) *
-------
%OF
TOTAL
COSTS
5,
SL0PE FLAT MODERATE .STEEP
50
40
30
U)
2
C l) 0 _J
0 —o
•w —< >0
z UJ =
Z UJQ3 2
tO I
0_C.) Cl) 2 Ui
Cl )
2 2-i
o C !) Cl) 0-i
— :2 Ui
0—C.)
4W t—
Z UJ 2
‘±0 i: 4
0Cl) 2 uj
0=j
-j
1-4 0
Z (iJft I
a_0
Figure D-1: Breakdown of Sewer Costs
-------
Appendix E
Glossary
Combined Sewer : A sewer which carries both wastewater and
water.
Degree Days : Sum of negative departures of average daily
temperature from 65° F; used to determine demand for
fuel for heating purposes.
Detention : The temporary storage of stormwater in order to
regulate the amount of flow in the transport system.
Dependent Variable : Variable whose values are functionally
determined by the independent variable.
Elasticity : The elasticity of variable x with respect to
variable y is the percent change in the value of x
associated with a one percent change in the value of y.
Emissions : Effluents discharged into the environment, spe-
cified as weight per unit time for a given pollutant
from a given source.
House Connection : A pipe which conveys wastewater from a
structure to the sewer system.
Hydraulic Radius : A measure of the depths of flow in a conduit
or channel; more formally it is the cross-sectional
area of flow divided by the perimeter of the channel
in contact with the fluid.
Independent Variable : Variable which can be directly mani-
pulated.
Infiltration : The water entering the sewer system and
connections from the ground.
Intercepting Sewer (interceptor) : The final sewer line in
a collection system, leading to the treatment plant.
Invert : The lowest point on the inside of a sewer or other
closed conduit.
Land Use Mix : The portions of a region allocated to spe-
cific land use types.
B-i
-------
Lateral Sewer : A sewer which receives wastes only from the
house connections.
Loadograph : A graph of pollutant load as a function of time
over the runoff period.
Main Sewer : A sewer which receives wastes from several
other sewer lines.
Module : For the purpose of this report a computational
package which can be independently manipulated.
Non-Point Source : For the purpose of this report all runoff
sources from non—urban activities.
Non-Stationary Source : Mobile activity which produces air
pollutant emissions.
Overflow : Excess flow discharged from a sewer which is
overloaded.
Point Source : Source of liquid discharge, as defined in
Section 502(14), Public Law 92—500, October, 1972.
Pollutograph : A graph of pollutant concentration as a
function of time over the runoff period.
Residential Density : The number of persons per unit of
residential land area. Net density includes only
occupied land while gross density includes in the
computation unoccupied portions of residential areas,
such as roads and open space.
Residual : Byproduct of activity which is released into the
environment.
Retention : The storage of stormwater to prevent it from
entering the sewer system; may be temporary or perma-
nent.
Runoff : That portion of precipitation or irrigation which
is not absorbed by the deep strata but finds its way
into the stream.
Sanitary Sewer : A sewer which carries only wastewater and
low volumes of ground-, storm— and surface water,
which are not admitted intentionally.
E-2
-------
Slope : The inclination of the ground surface or structure
(such as a sewer line). Generally expressed in number
of units of rise (or fall) per unit of horizontal
distance.
Stationary Source : Activity at fixed location which pro-
duces air pollutant emissions.
Storm Sewer : A sewer that carries stormwater and surface
water, Street wash and other wash waters, but excludes
domestic wastewater and industrial wastes.
Storm Sewer Discharge : Flow from a storm sewer that is
discharged into a receiving water.
Stormwater Runoff : Precipitation that falls onto the sur-
faces of roofs, streets, grounds, etc., and is not
absorbed or retained by those surfaces, but collects
and runs off, eventually reaching a sewer, stream or
another body of water.
E-3
U.S.GOVERNMENT PRINTING OFFICE: 1977. .U1 37:8
------- |