&EPA
United States
Environmental Protection
Agency
Office of Air Quality
Planning and Standards
Research Triangle Park NC 27711
EPA-450/3-78-014b
May 1978
Air
Growth Effects of Major
Land Use Projects
(Wastewater Facilities)
Volume II: Summary,
Predictive Equations
and Worksheets
-------
EPA-450/3-78-014b
Growth Effects of Major Land Use Projects
(Wastewater Facilities)
Volume II: Summary, Predictive Equations
and Worksheets
by
Peter H. Guldberg and Ralph B. D'Agostino
Walden Division of Abcor, Inc.
850 Main Street
Wilmington, MA 01887
Contract No. 68-02-2594
EPA Project Officer: Thomas McCurdy
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Air, Noise, and Radiation
Office of Air Quality Planning and Standards
Research Triangle Park, North Carolina 27711
May 1978
-------
This report is issued by the Environmental Protection Agency to report
technical data of interest to a limited number of readers. Copies are
available free of charge to Federal employees, current contractors and
grantees, and nonprofit organizations - as supplies permit - from the
Land Use Planning Office, Office of Air Quality Planning and Standards,
Environmental Protection Agency, Research Triangle Park, North Carolina
27711; or, for a nominal fee, from the National Technical Information
Service, 5285 Port Royal Road, Springfield, Virginia 22151.
This report was furnished to the Environmental Protection Agency by
Wai den Division of Abcor, Inc., Wilmington, Massachusetts 01887, in
fulfillment of Contract No. 68-02-2594. This report has been reviewed
by the Land Use Planning Office, EPA and approved for publication. Ap-
proval does not signify that the contents necessarily reflect the views
and policies of the Environmental Protection Agency, nor does mention of
trade names or commercial products constitute endorsement or recommendation
for use.
Publication No. EPA-450/3-78-014b
-------
ACKNOWLEDGEMENTS
Special appreciation goes to Mr. Thomas McCurdy, the U.S. Environmental
Protection Agency Project Officer for this study, whose extensive assistence
and advice was indispensable in the conduct at the study. Appreciation is
also extended to all EPA technical committee members for guidance given
throughout the study. The technical committee included representatives
of the Office of Transportation and Land Use Policy, the Municipal Con-
struction Division (Facilities Requirements Branch) and the Control Program
Development Division.
In addition, we wish to thank the more than one-hundred individuals in
city/county/regional planning agencies and transportation departments
nationwide who cooperated with us during the data collection task and helped
provide the data on which this study is based. Their time and cooperation
were invaluable.
Urban Systems Research & Engineering, Inc. (USRE) of Cambridge, MA were
subcontractors to Walden on this study, assisting in the definition of basic
model concepts, infrastructure relationships and exogenous variables, and
in the testing and refinement of the causal and predictive models. The
cooperation of Dr. James F. Hudson, who guided the USRE technical effort, is
deeply appreciated.
Finally, recognition is due the Walden staff who performed the long and
difficult task of field data collection - Mr. Mahesh Shah, Ms. Diane Gilbert,
Mr. Michael Geraghty and Mr. Keith Kennedy.
m
-------
TABLE OF CONTENTS
Section
Title
Page
II
III
INTRODUCTION 1-1
A. Study Overview 1-1
B. General Approach 1-3
C. Summary of Study Results and Conclusions ... 1-6
1. Volume I - Model Specification and
Causal Analysis 1-6
2. Volume II - Predictive Equations and
Worksheets 1-7
PHASE IV '- DEVELOPMENT OF PREDICTIVE MODEL .... 2-1
A. Land Use Predictive Equations 2-1
1. Approach 2-1
2. Discussion of Results 2-4
B. Model Validation 2-9
1. Cross-Validation 2-9
2. Weight Validity Index 2-10
C. Coefficient Stability Analysis 2-12
D. Emission Projection Worksheets 2-17
1. Definitions 2-19
2. Input Data Requirements 2-19
3. Instructions 2-25
4. English Unit Worksheets 2-41
5. Metric Unit Worksheets 2-59
6. Example on Using Worksheets 2-76
REFERENCES 3-1
APPENDIX A - Complete Statistical Output of the
Predictive Equations A-l
APPENDIX B - Glossary of Terms B-l
APPENDIX C - Definition of Model Variables .... C-l
APPENDIX D - Graphs of Actual Versus Predicted
Land Use for the Cross Validation
Analysis
APPENDIX E - Supplementary Information
D-l
E-l
IV
-------
LIST OF TABLES
Table Number Title Page
2-1 a Summary Statistics of the Land Use
Predictive Equations ............... 2-5
2-lb Summary Statistics of the Disaggregation
Equations .................... 2-5
2-2 Coefficients of Validity Between Actual and
Predicted Land Use from the Cross-Validation
Analysis ..................... 2-11
2-3 Weight Validity Indices for the Land Use
Predictive Equations ............... 2-13
2-% Summary Statistics of the Predictive Equations
Coefficient Stability Analysis .......... 2-14
-
Input Data Requirements for Impact Assessment
Model ...................... 2-20
2-6 Default Trip Generation Rates .......... 2-27
2-7 Parti cul ate and Sulfur Oxide Emission Factors . . 2-29
2-8 Single Family Detached or Attached Land Use
Based Emission Factors .............. 2-31
2-9 Multiple Family Land Use Based Emission Factors . 2-32
2-10 Commercial and Wholesale Land Use Based
Emission Factors ................. 2-33
2-11 Office-Professional Land Use Based Emission
Factors ..................... 2-34
2-12 Education Land Use Based Emission Factors .... 2-35
2-13 Other Land Use Based Emission Factors ...... 2-36
2-14 Estimated National Industrial Land Use Based
Emission Factors by Two Digit 1967 Standard
Industrial Classification Code .......... 2-37
2-15 Typical Uncontrolled Emission Factors for
Electric Utilities ............... 2-42
C-l Definition of Endogenous Model Variables ..... C-3
C-2 Definition of Exogenous Model Variables ..... C-5
C-3 Metric Units of Variables Used In the Predictive
Equations and Worksheets ............. C-16
-------
Figure Number
LIST OF FIGURES AND WORKSHEETS
Title
Page
1-1 Technical Approach to the Development of a
Statistical Model for Predicting the Growth
Effects of Major Wastewater Projects 1-4
2-1 Overview of the Land Use and Air Quality
Impact Assessment Procedure 2-18
2-2 Normal Seasonal Heating Degree Days 2-38
2-3 Normal Seasonal Cooling Degree Days 2-39
2-4 Annual Air Conditioner Compressor Hours 2-40
Worksheet Number Title Page
1/1M* Input Data Record 2-43/2-60
2/2M Land Use Projections 2-46/2-63
3/3M Final Land Use Projections 2-47/2-64
4/4M Default Disaggregation Equations 2-48/2-65
5/5M Confidence Intervals 2-49/2-66
6/6M Motor Vehicle Trips 2-51/2-68
7 Vehicle Miles Traveled (VMT) 2-52
7M Vehicle Kilometers Traveled (VKT) 2-69
8 Categorized VMT 2-53
8M Categorized VKT 2-70
9/9M Motor Vehicle Emission Factors 2-54/2-71
10/10M Composite Motor Vehicle Emission Factors 2-55/2-72
11/11M Total Motor Vehicle Emissions 2-56/2-73
12/12M Stationary Source Emissions 2-57/2-74
13/13M Emissions Summary 2-58/2-75
*The first set of numbers refers to worksheets in English units,
while the second set refers to metric units.
VI
-------
I. INTRODUCTION
A. STUDY OVERVIEW
Pursuant to 40 CFR 51.12(d)-(h), State Implementation Plans must
contain provisions to prevent any national ambient air quality standards
from being exceeded. These provisions are called Air Quality Maintenance
Area (AQMA) plans, and estimating the air quality impact of major land
use and urban development projects is a necessary part of AQMA planning
[1-6]. In addition, the National Environmental Policy Act [7] and the
Council on Environmental Quality (CEQ) "Guidelines on the Preparation of
Environmental Impact Statements" [8] require the consideration of secondary
impacts from major projects. CEQ states that:
"Many federal actions, in particular those that involve the
construction or licensing of infrastructure investments (e.g.,
highways, airports, sewer systems, water resource projects,
etc.), stimulate or induce secondary effects in the form of
associated investments and changed patterns of social and
economic activities. Such secondary effects through their
impacts on existing community facilities and activities, or
through changes in natural conditions, may often be more
substantial than the primary effects of the original action
itself [8]."
This has been a particular concern for wastewater systems, since their
primary impacts tend to be small, i.e., sewers and treatment plants generally
improve water quality, but they may lead to significant negative secondary
impacts. The probable large indirect impacts (redirecting growth and in-
ducing development) of a new or expanded regional sewage treatment facility
on ambient air quality, and the need for some procedure to ascertain what
its impact will be, is recognized in the AQMA Guideline series([1 ]:A-7ff,
[4]:21ff). To date, EPA has not developed a model to estimate what the
ambient impacts will be for use in AQMA planning. Thus, it was the purpose
of this study to develop such a model.
This study is entitled the Growth Effects of Major Land Use Projects
(GEMLUP-II), and it addresses the induced growth effects of wastewater major
1-1
-------
projects. Similar research on two other major project types, large residential
developments and large industrial/office parks, has been performed previously
and is reported on in the three volume set of GEMLUP-I final reports [9-11].
The secondary air quality impacts of wastewater projects are determined by
the land use growth induced by such projects. That is, the air impacts are
emissions from (1) residential complexes that appear at the end of and along
new sewer lines, (2) other service-oriented land uses (commercial, industrial,
office, government) that relate to residential development, and (3) motor
vehicles used as transportation between development areas. Thus, the key to
understanding secondary air quality impacts is to first understand the growth
effects of a wastewater facility on land use in a region. The objectives of
the study effort were: (1) to develop and test a path-analytic causal
model that represents the induced land use ten years after construction of a
wastewater major project, (2) to develop and validate a simplified predictive
model of induced development, (3) to test and correct the GEMLUP-I VMT model
[11], and (4) to develop worksheets that can be used to predict induced land
use and associated air pollution emissions. For these purposes, data were
collected from forty (40) case study wastewater projects nationwide.
The study project was divided into four major phases:
I. Definition of basic concepts and initial model specification,
II. Data collection,
III.Causal analysis of the land use model using path analysis,
IV. Development of predictive equations for the land use model
and worksheet procedures.
Two separate technical reports are to be prepared. This is the second volume
report and covers Phase IV of the study. The remainder of this chapter gives
an outline of the general study approach, and a summary of results and con-
clusions for the entire study. Chapters II and Appendix A summarize the
technical performance on Phase IV, including the development of the GEMLUP-II
model worksheets. The technical performance on Phases I-III is summarized
in the Volume I final report [12]. Appendix B gives a.glossary of terms used
in this report, and Appendix C summarizes definitions for the variables used
in the GEMLUP-II model.
1-2
-------
B. GENERAL APPROACH
The principal objectives of this study were to formulate a statis-
tical methodology to predict air pollutant emissions from:
• Induced land development associated with the construction and
operation of wastewater facilities in a community.*
• Motor vehicular traffic associated with the induced land
development.
The ability to accurately predict the secondary development induced by major
wastewater collection and treatment projects is dependent on understanding
the complex interrelationships inherent in such a model. Thus, an important
objective was to formulate and test a causal theory of induced development
using path analysis. Additional objectives were to test and correct the
motor vehicle traffic model developed as part of the previous GEMLUP-I
study [11] for use in the air pollutant emissions projection procedure, and
to integrate the predictive land use model, the traffic model, and land use
based emission factors in a set of easy-to-use worksheets.
The approach to fulfilling these objectives involved the sequential
execution of four separate project phases shown schematically in Figure 1-1
and summarized below.
In Phase I, our primary interest was to define the basic concepts
on which this study is based. The first step in this process was to
determine the modeling approach. Next, the infrastructure causal relation-
ships of wastewater facilities in communities were studied and the knowledge
used to define "induced development" in the model. The concepts of a
"wastewater major project" and the "area of analysis" for induced development
were also studied and defined. Knowing the important causal relationships
enabled us to develop a list of model variables representative of the relevant
factors involved (e.g., the major project, land use, regional growth).
*
The causal model for induced development produced by this study does not
take into account the effects of mitigating measures. Variables measuring
restrictions on on-site disposal and hookups to existing interceptor
lines were included in the development of the model, however, see Volume I,
Section II [12]. '
1-3
-------
PHASE I
INITIAL MODEL SPECIFICATION
PHASE II
DATA COLLECTION
PHASE III
CAUSAL ANALYSIS
PHASE IV
DEVELOPMENT OF PREDICTIVE EQUATIONS
DEFINE
MODELING
APPROACH
IDENTIFY
INFRASTRUCTURE
RELATIONSHIPS
_1
DEFINE MAJOR
PROJECT ADD
AREA OF
ANALYSIS
INITIAL CHOICE
OF VARIABLES
COEFFICIENT
STABILITY
ANALYSIS
DETERMINE
.NET CAUSAL
EFFECTS
PRELIMINARY
CASE STUDY
SURVEY
LIST OF
POTENTIAL
CASE STUDIES
MODEL
VALIDATION
PREDICTIVE
LAND USE MODEL
EQUATIONS
FINAL PATH
ANALYTIC LAND
USE MODEL
CAUSAL
ANALYSIS REPORT
(VOLUME I)
LIST OF
DATA
REQUIREMENTS
SL'I-"!ARY RE:
E :::
CASE STUDY
SELECTION
DEVELC?
E'-'ISSIO'-S
PROJECTION
VIORKSHEETS
SPECIFY
PRELIMINARY
MODEL
STEPWISE
REGRESSION
ANALYSIS
DATA
COLLECTION
THEORY
TRIMMING
VALIDATE
GEMLUP
TRAFFIC
MODEL
INITIAL
ANALYTIC
E MODEL
COMPUTER
DATA FILE
FIGURE 1-1
TECHNICAL APPROACH TO THE DEVELOPMENT OF A STATISTICAL MODEL
FOR PREDICTING THE GROWTH EFFECTS OF MAJOR WASTEWATER PROJECTS
-------
Finally, specific causal relationships between the variables were hypothesized
and formalized in an initial causal model.
The principal objective of Phase II was to collect a sufficiently
large, diverse, and thereby representative, cross-sectional data base on which
to develop the model. Due to the critical importance of the quality of this
data base and the large manpower effort required to collect such data, the
first task required a careful selection of 40 case study wastewater major
projects distributed on a nationwide basis, which had the potential, upon
construction, to induce a significant quantity of land development in their
communities and for which all the requisite data were available. To this end,
a case study mail survey based upon the screening of over 15,000 federally
funded wastewater collection and treatment projects nationwide was performed.
Final selection of the 40 case study projects was made, a data collection
training course was conducted for field personnel, followed by site visits
to the case study areas to collect the required data.
The objectives of Phase III were to develop the final causal model
and validate the GEMLUP-I VMT model [11]. The initial causal model was
tested using the set of case study data and the statistical techniques of
path analysis. This approach verified which of the hypothesized causal
relationships were significant, trimmed those that were not, and determined
model parameters for the final causal model. Tasks were also performed to
trace the direct and indirect effects the model variables have on one another
and to validate and correct the GEMLUP-I VMT model.
In Phase IV, predictive equations of induced land use associated
with wastewater major projects were developed and validated. These predic-
tive equations, the GEMLUP traffic model, and GEMLUP land use based emission
factors [10] were then used as the basis for an emissions projection procedure,
The procedure was developed in worksheet form to serve as a generalized
analytical tool for use by planners and environmental engineers in predicting
the induced land use and air pollutant impacts associated with major waste-
water collection and treatment systems.
1-5
-------
C. SUMMARY OF STUDY RESULTS AND CONCLUSIONS
1. Volume I - Model Specification and Causal Analysis [12]
A theoretical model of the Growth Effects of Major Land Use
Projects has been developed. This model represents the total land use in
the drainage basin of a wastewater collection and treatment system (the
major project), ten years after its construction. The model represents the
process of induced land use growth in the following 9 land use categories:
Residential Education
Commercial Recreation
Office-Professional Wholesale/Warehouse
Manufacturing Other
Highways (Non-expressway)
The assumption of a single basic causal structure for induced development,
and the use of cross-sectional data from 40 diverse case study major projects
throughout the United States, allowed a static approach to the testing of
the theoretical model, using path analysis.
Path analysis is a set of statistical techniques useful in test-
ing theories and studying the logical consequences of various hypotheses
involving causal relations. It is not capable of deducing or generating
causal relations, only testing them. The causal analysis of induced land use
development in the current study involved the use of two basic statistical
techniques: two-stage least squares and stepwise ordinary least squares
(multiple regression). The first technique was required to produce consis-
tent estimates of the path coefficients in a system of simultaneous equations
involving feedback loops (or reciprocal causation) in the models. The second
technique was used to solve the remaining recursive portions of the models.
The dependent variables in these regression analyses represented the total
land use in the previously noted 9 categories. Both linear and non-linear
forms were tested and the linear form was found to produce the best fit.
Specific statistical criteria were developed to identify model paths that
were insignificant or redundant, and these criteria were used in an itera-
tive process to trim unneeded and undesirable paths from the models. The
1-6
-------
trimming process eliminated almost half of the paths in the models as
originally specified. The statistical problems of multicollinearity,
suppressor variables and identification were eliminated through the approach
used to trim the initial model.
The final models of land use development show that strong statis-
tical relationships exist between the variables representing the 9 categories
of total land use and the other model variables representing induced and
non-induced land use growth processes. The results indicate that the final
2
model explains more than half of the variance in the case study data with R
values ranging from 0.27 to 0.82 and averaging 0.54. The residuals of the
final regressions do not exhibit any trends or patterns, indicating the
2
remaining unexplained variance (1-R ) is not due to poor specification of
the model, but rather due to wide variance in the case study data (i.e., the
problem of trying to develop one generalized model for a broad range of
situations). An analysis of the stability of the model coefficients deter-
mined that, in general, the coefficient values have low variance (± 15%)
and exhibit no extreme instabilities. An analysis of the net causal effects
in the land use model indicates that reserve collection system capacity of a
wastewater major project is a significant causal factor for induced land use
growth, principally in the residential, manufacturing, education and highways
categories. By contrast, treatment plant capacity was not found to be an
important causal factor.
The GEMLUP-I VMT model [11] was validated using transportation
data from 11 of the 40 case study major projects. Based on the validation
results, it was concluded that revisions to both the default predictive
equations for trip length and the default values for trip generation rates,
were necessary. The revised VMT model was validated and found to have an
average error (imprecision) of 23%, with no statistically significant bias.
2. Volume II - Predictive Equations and Worksheets
The development of predictive equations for land use development,
separate from the model equations obtained in the causal analysis, was
necessitated by the simultaneity of the causal relationships, i.e, the
causal equations include independent variables whose values will not be
1-7
-------
known in the future. Therefore, it was necessary to develop predictive
equations in which the endogenous variables appeared only as the dependent
variables. Such an assumption defined a system of equations which was
solved with ordinary least squares analysis.
Predictive equations were developed for the 9 land use cate-
gories used in the causal analysis. In addition, predictive equations were
developed for disaggregating residential and commercial land use totals into
3 subcategories each. Logit analysis [14] was used with ordinary least squares
2
regression to develop the disaggregation equations. The R statistics
averaged 0.67 and 0.76 for the land use and disaggregation equations, re-
spectively. Thus, the predictive equations explained the majority of the
variance in the dependent variables. Overall F statistics indicate all
equations are significant at or below the 1 percent level. The average error
associated with use of the predictive land use equations is ±63 percent.
These errors are reasonable for a generalized tool, applicable nationwide,
and are less than those introduced subsequently in the impact assessment
procedure [16].
A model validation using the techniques of cross-validation and
weight validity index [15] concluded that 8 of the 9 land use predictive
equations are generalized enough to produce good predictions on an independent
sample. Conflicting results were obtained for the OTHER equation, which may
not be well generalized.
A coefficient stability analysis testing for worst case in-
stabilities concluded that no significant instabilities occur in 7 of the 9
land use predictive equations. The exceptions, COMM and OTHER, contain
instabilities which indicate that the true functional form of these equations
changes with the magnitude of the data.
The predictive land use equations were included in an impact
assessment procedure that estimates the total air pollutant emissions
associated with the induced development from a wastewater major project.
This impact assessment procedure was formalized in a set of easy-to-use
worksheets, presented in Section II.D. These worksheets serve as an
operational tool for environmental engineers or planners to assess the
secondary air quality impacts associated with new or expanded wastewater
facilities. _
I -o
-------
II. PHASE IV - DEVELOPMENT OF PREDICTIVE MODEL
The objective of the fourth and final phase of this study was to
create and validate a predictive model to estimate the amount of induced
land use (by category type) associated with major wastewater projects.
These predictive equations were then incorporated along with the revised
VMT model [12] and GEMLUP land use based emission factors [10] into an emis-
sions projection procedure in worksheet form.
A. LAND USE PREDICTIVE EQUATIONS
1. Approach
The development of predictive equations for land use de-
velopment, separate from the model equations obtained in the causal analysis
(see Volume I [12]), was necessitated by the simultaneity of the causal
relationships, i.e., causal equations include independent variables whose
values will not be known in the future. Therefore* it was necessary to de-
velop predictive equations in which the endogenous variables* appeared only
as the dependent variables. Such an assumption defines a system of equations
which can be solved with ordinary least squares analysis.
When predictive equations are developed using ordinary least
squares for variables which are known to be effected by simultaneity, the
individual regression coefficients are biased. However, the final prediction
of the equation is an unbiased estimate of the dependent variable. Since
these equations are to be used for predictive and not analytical purposes,
the fact that the individual coefficients are biased estimates is not of great
concern. It must be emphasized, therefore, that the regression coefficients
obtained for the predictive equations should not be examined to judge the
effects of independent variables on the dependent variable, nor should these
coefficients be compared with those obtained in the causal analysis. Rather
the appropriateness of a predictor can only be determined by examining its
performance with regard to a set of objective statistical criteria.
See Appendix B for a definition of terms. See Appendix C for a definition
of model variables.
2-1
-------
In order to systematically decide which variables to include
in the predictive equations, stepwise regression techniques were employed.
The dependent variables in this stepwise regression analyses represented
the 9 categories of total land use analyzed in the causal analysis and 6
percentage disaggregation categories for residential and commercial land use
• RES = Residential
• COMM = Commercial
• OFFICE= Office-professional
• MANF = Manufacturing
• HIWAYS= Non-expressway highways
• EDUC = Education
• REC = Recreation
• WHOLE = Wholesale/warehousing
• OTHER = Hotel/Motel, culture, religion
Residential
• SFDET = Single family detached
• SFATT = Single family attached
• MF = Multiple family
Commercial
• PC.OMM1= Buildings with <50,000 ft GLA*
• PCOMM2= Buildings with 50-100,000 ft2 GLA
• PCOMM3= Buildings with >100,000 ft2 GLA
Although it would have been possible to develop predictive equations directly
for the land use subcategories, it is advantageous to predict total, aggre-
gate residential and commercial land use and then disaggregate the totals.
There are two reasons for such an approach. First, results of the GEMLUP-I
study indicate that average errors associated with aggregated predictive
equations tend to be smaller. Second, by not directly incorporating the
disaggregation of residential and commercial land uses into density classi-
fications in the predictive equations, we allow the user of the predictive
equations to enter in a land use density distribution applicable to some
future time period of interest, different from that of 1960-1970. This
*
Gross Leasable Area
2-2
-------
second concern is motivated by the shift in density of new housing construction
in the past few years away from single family detached developments, which
predominated during the 1960s. Thus, we believe that this approach ensuired a
more generalized, predictive model.
The next step was to decide which independent variables to
include in the predictive equations. A maximum of 6 independent variables were
included in each equation to keep the predictive model simple and easy-to-use,
to avoid possible degrees of freedom problems, and to keep confidence intervals
small. The set of potential independent variables included all exogenous and
instrumental variables*from the initial causal model. The choice of independent
variables was dictated by the stepwise regression analysis, i.e., at each step
the analysis chose from the set of all independent variables the one which
explained the most additional variance in the dependent variable. Since the
analysis can produce several different predictive equation forms depending upon
the independent variables included in the analysis, and their priority, a set of
statistical criteria were developed for application to each variable before its
final acceptance in the model equation. These criteria were the following:
• A test to ensure an independent variable was significant,
e.g., a minimum F statistic value of 1.7, corresponding to
the 10% significance level.**
• A test to ensure the choice of the model form which explains
the greatest amount of variance in the dependent variable,
viz. the requirement that the chosen final form have the
highest adjusted R2 value.
• A test to ensure the absence of multi col linearity, viz.
3 < 1.0.
The stepwise regression analysis was applied, as described
above, using data from all 40 case studies and programs from the SPSS soft-
ware package [13]. It was decided that it was preferable to use the full
set of 40 samples in developing the equations since this would produce the
* See Appendix B for a definition of terms. See Appendix C for a
definition of model variables, in both English and metric units.
** That is, there was at most a 10 percent chance of accepting a variable
as significant in the regression when it actually was not.
2-3
-------
most valid, generalized model. All variables were used in unstandardized
form. The results were predictive, linear equations in both English and
metric units, discussed in the next section, with full statistical output
summarized in Appendix A.
2. Discussion of Results
a. Land Use Equations
The land use predictive equations obtained by applying the
previously discussed objective criteria to the stepwise regression analyses
are summarized below. Summary statistics for these equations are shown in
2
Table 2-la. R values indicate the predictive equations are explaining
the majority of variance in the dependent variables, i.e., the mean value
for this statistic was 0.67. The overall F statistics indicate all of the
predictive equations are significant at or below the one percent level. The
coefficient of variation is defined as the ratio of the standard error of
estimate of the regression to the mean value of the dependent variable. The
mean of the coefficients of variation for the land use equations is 0.63,
indicating that the average error encountered in the use of these predictive
equations will be ±63 percent. These errors are reasonable for a generalized
tool, applicable nationwide, and are less than those introduced subsequently
in the impact assessment procedure [16].
The predictive equations shown below constitute a set of
equations applicable to an area of analysis where a wastewater major project
of a certain size range will be or already has been built. They do not
constitute a general land use predictive model.
A major project is defined as the construction or extension
of interceptor or collector sewer lines in a community in
the United States that affected an increase in absolute
system collection capacity of 1.0 million gallons per day
(MGD) or more. Based on the distribution of data used in
developing the model, the increase in collection capacity
should not exceed 100 MGD.
The area of analysis is defined as the legal service area
of the major project in the base year, and it must contain
a significant amount of vacant, developable land. Based on
the distribution of data used in developing, the model,
the size of the area of analysis should be 5,000 to 75,000
acres.
2-4
-------
TABLE 2-1 a
SUMMARY STATISTICS OF THE LAND USE PREDICTIVE EQUATIONS
Dependent
Variable
RES
COMM
OFFICE
MANF
WHOLE
HI WAYS
EDUC
REC
OTHER
Dependent
Variable
RP1
RP2
CP1
CP2
Number of
Predictors
6
6
6
6
6
5
6
6
6
TABLE
SUMMARY STATISTICS OF THE
Number of
Predictors
6
5
5
5
R2
0.74
0.57
0.70
0.63
0.76
0.52
0.75
0.65
0.67
2-1 b
2 Coefficient
R a of Variation
0.69
0.50
0.65
0.56
0.72
0.45
0.71
0.58
0.61
0.38
0.81
0.59
0.89
0.67
0.53
0.36
0.59
0.88
DISAGGREGATION EQUATIONS
R2
0.82
0.84
0-.73
0.65
R2
R a
0.76
0.80
0.69
0.60
2-5
-------
The use of these equations should be limited to situations in which the major
project and areas of analysis are in these size ranges. Also, any application
of the predictive equations should be qualified by the error range indicated
by the coefficients of variation shown in Table 2^1.
(1) English Units.
RES = - 8,692 VACANT + 5,347 LAND + 16.42 RECAP1 + 11.24 MANJOB - 18,804 STAY
+ 487.2 DRIVE + 13,466
COMM = 4,302 SERVED - 4,515 VACANT - 28.84 KIDS + 6,249 ACCESS - 17.04 JOBCHG
- 2.626 RECAP! + 5,376
OFFICE = 719.5 RRMILE + 33,478 OZONED + 36.70 PEAK + 0.3334 DUACRE - 1,791
IZONED + 572.0 POPDIF - 459.9
MANF = 2,591 RRMILE - 3,890 VACANT + 2.978 MANJOB + 85,923 OZONED + 234.7
VACOFF +22.7 AIRPRT + 611.2
WHOLE = 166.5 DRIVE + 86.28 VACOFF + 3.340 OFFJOB + 6,043 EMPOP + 6,949 UNEMP
- 0.1065 PERCAP -1,911
HIWAYS = 35.22 RRMILE + 22.98 LAND - 0.1167 GOVT + 0.5379 RECAP + 25.94 CROSS
+ 11.35
EDUC = -3,186 POOR + 723.0 SERVED + 437.5 LAND + 95.17 LIMITS - 405.2
PERCHG - 1.415 COST + 428.9
REC = 270.4 RRMILE + 2.631 CAPAC2 + 322.1 PERCHG + 0.9742 OFFJOB 3,226
VACHSE - 557.8 IZONED + 59.52
OTHER = 1,134 RRMILE - 0.3020 PERCAP - 12,175 VACHSE + 649.0 PERCHG -
295.4 INTDEN + 403.8 RZONED + 3,087
(2) Metric Units. The predictive model equations in this
section are reported in metric units. The conversion of units involved only
the dimension of distance. The following conversion factors were used:
mile2 -»• km2, multiply by 2.589
1,000 ft2 -> m2, multiply by 92.90
2
acres -> m , multiply by 4,047
miles -* km, divide by 1.609
-1 -?
acre + m , divide by 4,047
mile" ->• km , divide by 1.609
2-6
-------
RES = -2 148 VACANT + 1.321 LAND + 0.0041 RECAP1 + 0.0072 MANJOB - 4.646
STAY + 0.3117 DRIVE + 3.327
COMM = 987.5 SERVED 1,036 VACANT - 6.621 KIDS + 3,714 ACCESS - 10.13
JOBCHG - 0.6029 RECAP1 + 1,234
OFFICE = 102.7 RRMILE + 7,685 OZONED + 8.424 PEAK + 0.1981 DUACRE - 411.0
IZONED + 131.3 POPDIF - 105.6
MANF = 369.6 RRMILE - 893.0 VACANT + 1.770 MANJOB + 19,724 OZONED + 53.87
VACOFF + 3.239 AIRPRT + 140.3
WHOLE = 98.23 DRIVE + 19.80 VACOFF + 1.985 OFFJOB + 1,387 EMPOP + 1,595
UNEMP - 0.0244 PERCAP - 438.7
HIWAYS = 0.0087 RRMILE + 0.0091 LAND - 0.00005 GOVT + 0.0002 RECAP + 25.94
CROSS + 0.0045
EDUC = -731.3 POOR + 170.0 SERVED + 100.4 LAND + 21.85 LIMITS - 93.02
PERCHG - 0.3248 COST + 98.45
REC = 168.1 RRMILE + 2.631 CAPAC2 + 322.1 PERCHG + 2.522 OFFJOB - 3,226
VACHSE - 557.8 IZONED + 122.9
OTHER - 161.8 RRMILE - 0.0693 PERCAP - 2,795 VACHSE + 150.0 PERCHG - 67.81
INTDEN + 92.70 RZONED + 708.5
b. Disaggregation Equations
Predictive equations for disaggregating residential and
commercial land uses into subcategories were developed using logit analysis
[14]. This approach fits the percentage data (P) to a logistic S curve.
The result is a non-linear structural equation that relates the range 0 to 1
with a set of independent variables (Xn.Xp, X ), viz:
(b1x1+b9x9+ +b x ) ,
P = (1 + me ] ] 2 2 n n T1 (1)
P is a general term for the previously defined percentage variables SFDET,
SFATT, MF, PCOMM1, PCOMM2, and PCOMM3. This approach avoids two problems
which would arise if linear, stepwise regression analysis was used, namely
(1) the value for P would not be constrained between 0 and 1 (i.e., 0% and
100%), and (2) the sum of the percentages predicted for all subcategories
of a given land use type would not sum to 1 (i.e., 100%).
2-7
-------
To translate equation (1) to a form that can be solved
with linear, stepwise regression, the following transformation was made:
In (P/l-P) = ln(m) -Vrb2x2~' ' • ''Vn (2)
This is equivalent to:
Y = a0 + alXl + a2x2+....+anxn (3)
where: m = e
bn ' -an
To constrain the total of predicted subcategory percentages
to 1.0, it was necessary to apply logit analysis in an iterative manner.
First, the percent of single family detached and large commercial develop-
ments were predicted. Then the residual percentages were predicted. Thus,
Residential
% Single family detached = RP1* = SFDET
% Multiple family = RP2 (1-RP1), where RP2 = MF/(SFATT + MF)
% Single family attached = 1-RP2 -RP1
Commercial
r) •fc^f
% >100,000 f1T GLA = CP1 = COMM3
% 50-100,000 ft2 GLA = CP2 (1-CP1), where CP2 = PCOMM2/(PCOMM1+PCOMM2)
% <50,000 ft2 GLA = 1-CP2 - CP1
p
The R statistics for the logit analysis, shown in Table 4-lb,
indicate excellent predictive ability for the disaggregation equations, with
value ranging from 0.65 to 0.84, and averaging 0.76. The complete predictive
equations for RP1, RP2, CP1, and CP2 are summarized below.
*
The statistical output listed in Appendix A refers to RY1 which is simply
the transform of RP1 used to solve for the coefficient values, i.e., RY1 =
(RP1/1-RP1). This applies to RP2, CP1 and CP2 in a similar manner.
**
Gross Leasable Area.
2-8
-------
(1) English Units.
RP1 = (1 + 92.8EXP?- 0.0137 KIDS + 0.000918 CAPCHG - 0.784 SERVED
- 22.7 OZONED + 2.98 STAY + 0.181 HOSCHG)H
RP2 = (1 + 1.24 EXP(- 4.38"HSECHG + 5.48 POOR + 34.1 OZONED
- 0.732 RRMILE - 0.0365 PEAK))-!
CP1 = (1 + 0.104 EXP(0.215 VACOFF - 0.507 LIMITS + 0.0232
CAPAC2 - 0.0159 KIDS + 0.0603 TLIMIT))-'
CP2 = (1 + 0.585 EXP(1.54 PHASE + 0.00459 TRANS + 0.0178
DISCED + 0.00417 GOVT - 1.48 VACANT)H
(2) Metric Units.
The disaggregation equations in metric units are identical
to those for English units with two exceptions: (1) the coefficient for RRMILE
in RP2 changes to -0.454, and (2) the coefficient for DISCED in CP2 changes to 0.0111.
B. MODEL VALIDATION
The usefulness of any set of predictive equations depends upon their
generality. Only if a regression equation is based upon a data sample which
is representative of the general data population can it provide accurate pre-
dictions in different situations. In the current study, where a fairly small
data sample (40) was available for developing predictive land use equations,
the question of validity was important. The preferred test of an equation's
validity is an external validation, viz., applying it on a test case basis to
an independent sample of data (i.e., independent of the sample on which the
coefficient values were based) and observing its predictive ability. In the
current study a separate, independent sample was not available. Therefore,
the analytical technique of cross-validation was used to simulate the existence
of such a test sample. In addition, a more precise weight validity index was
also computed.
1. Cross-Validation
Cross-validation provides a measure of the overall validity of
the predictive equations. The procedure involved splitting the case study
data sample of 40 into two random samples of 20 each. The first sample of
20 consisted of case numbers* 1,2,3,5,6,7,10,12,13,14,18,21,24,26,27,29,30,
32,38, and 40; the second sample of 20 consisted of the remaining case num-
bers. The coefficient values in the predictive equations were recomputed
using the first data sample of 20. These coefficient values, specifying
*
See Table 3-3 in Volume I [12] for a definition of case study numbers.
** Y
EXP is the exponential function (e). Thus, EXP(x) = e .
2-9
-------
9 land use predictive equations, were used in conjunction with data from the
second sample of 20 to predict the values of the dependent variables in the
second sample. Statistical comparisons were then made between actual and
predicted values for the dependent variables in the second sample.
The correlation coefficients (R) between actual and predicted
values are shown in Table 2-2, and graphs of the actual versus fitted
(predicted) values are summarized in Appendix D. In this application, the R
statistic represents a coefficient of validity of the predictive ability of
each equation. Values in Table 2-2 range from -0.04 to 0.68'. All R values
but one are statistically significant at the 2% level or better. The ex-
ception is the OTHER equation, where R is not significantly different from
zero. These results can be interpreted that 8 of the 9 equations are generalized
enough to produce good predictions using an independent sample. Only the OTHER
equation may not be well generalized. The poor performance of this equation
could, however, be attributed to non-homogeneties in the data sample.s caused
by the small sample size.
2. Weight Validity Index
The weight validity index is a new statistical technique [15]
which essentially performs a validation without the need for a separate
independent sample or having to split the original sample, as in cross-
validation. Since the entire data sample of 40 is used in the test, it gives
more precise estimates of validity for the equations being tested.
The procedure involved first obtaining the R statistic for an
equation estimated on the full sample of 40 (see Table 2-la). Next, the
quantity p is computed as:
f I 1/2
P = 1 - (N-4)(l-R2)("l + 20-R2)^
N-n-1 \ N-n-1 /J
where: N = sample size = 40
n = number of independent variables (see Table 2-la)
/x
Then the weight validity index p is estimated by:
/^ /\
Pc = 2 p- R
2-10
-------
TABLE 2-2
COEFFICIENTS OF VALIDITY BETWEEN ACTUAL AND
PREDICTED LAND USE FROM THE CROSS-VALIDATION ANALYSIS
Dependent
Variable
RES
COMM
OFFICE
MANF
HI WAYS
EDUC
REC
WHOLE
OTHER
Coefficient of
Validity (R)
0.64
0.56
0.46
0.39
0.42
0.68
0.68
0.63
-0.04
Statistical
Significance Level
1%
1%
1%
2%
1%
1%
1%
1%
— ^
2-n
-------
The values for pc are summarized in Table 2-3. To interpret the results the
values of R and pc are compared for a given equation. If there is a signifi-
cant drop from R to pr, then the equations are not well generalized. In
U /v
the current application, p is only 3-9% less than R. This small drop in
value indicates that aV\_ the predictive equations can be used confidently
on a new sample and that they are all well generalized for predictive appli-
cations.
C. COEFFICIENT STABILITY ANALYSIS
To test the stability of the coefficient values in the land use
predictive equations, a technique was devised which is very sensitive to
instabilities. The analysis simulated a "worst case" in terms of instabili-
ties, and thus is more stringent than the jackknifing approach used in the
causal analysis [12]. The procedure involved reestimating the model coef-
ficients when excluding a cluster of 2 or 3 samples that had the greatest
potential for instabilities. For each equation, the independent variables
with the two largest 3 weights were identified and, together with the de-
pendent variable, examined for skewness in their distributions. Where extreme
values occurred, the case study number was noted. All such numbers were
combined to see if 2 or 3 case studies dominated. These were then the data
samples excluded from the stability analysis regressions (see Table 2-4).
The percentage change in coefficient values and the standard error of the
*
regression were then computed. These values are listed in Table 2-4 as well.
Note that the excluded values may or may not be associated with large residuals,
In fact, many times a small residual on an extreme value indicates the ex-
treme value has a significant impact on the regression line and thus more
potential for instabilities. Because of this, residuals were ignored in
choosing the case studies to exclude.
The results in Table 2-4 reveal significant changes in the model
coefficient values. The range in percent change is 1% to 111%, with an
average of 38% overall. However, since there is a mixture of both + and -
signs in the percentages, there may be compensating effects, i.e., the final
*
On a common degrees of freedom basis.
2-12
-------
TABLE 2-3
WEIGHT VALIDITY INDICES FOR THE LAND USE
PREDICTIVE EQUATIONS
Dependent
Variable
RES
COMM
OFFICE
MANF
WHOLE
HIWAYS
EDUC
REC
OTHER
Weight Validity
Index
0.83
0.69
0.79
0.74
0.84
0.66
0.83
0.75
0.77
Regression
R Value*
0.86
0.75
0.84
0.79
0.87
0.72
0.87
0.81
0.82
Coefficient
of Validity**
0.64
0.56
0.46
0.39
0.42
0.68
0.68
0.63
-0.04
**
From Table 2-1 a.
r
From Table 2-2.
2-13
-------
TABLE 2-4
SUMMARY STATISTICS OF THE PREDICTIVE EQUATIONS
COEFFICIENT STABILITY ANALYSIS
Dependent Case Studies
Variable Excluded
RES 7,11,13
COMM 18,35
OFFICE 3,18,35
MANF 18,35
Percent Change in Independent
Standard Error of Variables
Regression
- 9.9% VACANT
LAND
RE CAP 1
MANJOB
STAY
DRIVE
-36% SERVED
VACANT
KIDS
ACCESS
JOBCHG
RE C API
-15% RRMILE
OZONED
PEAK
DUACRE
IZONED
POPDIF
-20% RRMILE
VACANT
MANJOB
OZONED
VACOFF
AIRPRT
Percent Change in Model
Coefficient Values
+ 9
-11
-14
-47
+61
-73
-60
+59
+83
-86
*
+81
-54
-55
-34
-13
+72
-20
-64
+37
- 1
-46
-55
-81
*Variable not significant in regression excluding skewed cases
2-14
-------
TABLE 2-4 (CONTINUED)
SUMMARY STATISTICS OF THE PREDICTIVE EQUATIONS
COEFFICIENT STABILITY ANALYSIS
Dependent Case Studies Percent Change in
Variable Excluded Standard Error of
Regression
HIWAYS 6,13,18 - 2.9%
EDUC 3,13 - 3.3%
REC 11,18,27 - 5.6%
WHOLE 7,14 - 9.5%
Independent
Variables
RRMILE
LAND
GOVT
RECAP
CROSS
POOR
SERVED
LAND
LIMITS
PERCHG
COST
RRMILE
CAPAC2
PERCHG
OFFJOB
VACHSE
I ZONED
DRIVE
VACOFF
OFFJOB
EMPOP
UNEMP
RERCAP
Percent Change in Model
Coefficient Values
-81
-20
+28
-16
-49
+ 7
- 4
- 5
- 1
- 4
-Hll
-18
+15
-54
-12
+28
+ 35
-25
-13
-19
-31
-69
+25
2-15
-------
TABLE 2-4 (CONTINUED)
SUMMARY STATISTICS OF THE PREDICTIVE EQUATIONS
COEFFICIENT STABILITY ANALYSIS
Dependent Case Studies Percent Change in Independent Percent Change in Model
Variable Excluded Standard Error of Variables Coefficient Values
Regression
OTHER 12,18 -44% RRMILE -73
PERCAP -68
VACHSE -81
PERCHG -75
INTDEN -79
RZONED +40
2-16
-------
predicted values may or may not be changing. The standard error of the re-
gression is an indicator of the precision of the predicted values. Thus,
changes in the standard error, when excluding certain samples, provide infor-
mation about the stability of an equation. The percent changes in the standard
error listed in Table 2-4 range from - 2.9% to -44%. Note that all values are
negative, indicating that, as would be expected, the regression does-a better
job on a narrower range of data (i.e., that which excludes hvghly skewed
samples.) For 7 of the 9 equations, the percent change under worst case
conditions is -20%. This indicates that no significant instabilities occur
in these equations. The other two equations, COMM and OTHER, have percent
changes of 36% and 44%, respectively. This indicates in both cases that an
equation different from the true predictive equation does significantly better
on a narrower range of data. Thus, the excluded points are causing instabili-
ties in the COMM and OTHER equations. What these worst case instabilities
represent are the fact that the functional form of the land use relationships
change with the magnitude of the data. Thus, a single generalized equation
(i.e., that used in the GEMLUP-II model) cannot be as precise as one developed
for only a portion of the range of data.
D. EMISSION PROJECTION WORKSHEETS
The predictive land use equations discussed in the last section, serve
as the input to a two-part air quality assessment procedure outlined in
Figure 2-1. First, the predictions of future land use in the area of analysis
of a major project are translated into stationary source emissions (for all
criteria pollutants) using a set of land use based emission factors [10].
Second, the revised VMT model [11,12] estimates vehicle miles traveled, and
hence mobile source emissions, generated by transportation actively servicing
the land use types. These two emission components are then summed to give
the user the total air pollutant emissions associated with the induced develop-
ment from a wastewater major project.
This .impact assessment procedure has been formalized in a set of
worksheets, presented in this section. These worksheets serve as an opera-
tional tool that can be easily used by environmental engineers or planners
to assess the secondary air quality impacts associated with new or expanded
wastewater facilities.
2-17
-------
MAJOR PROJECT
CHARACTERISTICS
BASE YEAR
CONDITION
PROJECTED
REGIONAL GROWTH
PREDICTIVE
EQUATIONS
PROJECTED
FUTURE LAND
USE
LAND USE BASED
EMISSION
FACTORS
TRAFFIC
MODEL
STATIONARY
SOURCE
EMISSIONS
MOBILE \
SOURCE ]
EMISSIONS /
TOTAL AIR
POLLUTANT
EMISSIONS
FIGURE 2-1
OVERVIEW OF THE LAND USE AND AIR QUALITY
IMPACT ASSESSMENT PROCEDURE
2-18
-------
1. Definitions
The predictive equations used in the worksheets constitute a
set of equations applicable to an area of analysis where a wastewater major
project of a certain size range will be, or already has been, built. They
do not constitute a general land use model. They produce projections of
total land use in the area of analysis ten years after the initiation of
the major project. In this regard, the following definitions apply:
A major project is defined as the construction or extension
of interceptor or collector sewer lines in a community in
the United States that affected an increase in absolute
system collection capacity of between 1.0 and 100.0 million
gallons per day.
The area of analysis is defined as the legal service area
of the major project in the base year, and it must contain
a significant amount of vacant, developable land. It should
be between 5,000 and 75,000 areas in area.
The year t, or year of initiation, corresponds to the year
the major project's new or expanded collection system first
carried wastewater flows. If completion of the project was
(or is to be) phased, it should be completed before the
year t + 5.
The use of the worksheets should be limited to situations in which the major
project and areas of analysis are in the size and time ranges noted above.
2. Input Data Requirements
The data required for application of the worksheets are sum-
marized in Table 2-5. Here the variable names used in the worksheets are
defined (in English units) and data sources are listed. The data required
are generally available from regional planning agencies and the facility
plans, and the total data collection task will require approximately two
to three mandays of effort. Appendix C contains corresponding metric units
for each model variable, as well as a cross-index relating worksheet variable
names and predictive equation variable names. Photocopies of worksheet 1
(English units, starting on page 2-43) or 1M (metric units, starting on
page 2-60) should be used to record actual values when collecting data.
These are contained in Sections 4 and 5, respectively, along with all other
worksheets.
2-19
-------
TABLE 2-5
INPUT DATA REQUIREMENTS FOR IMPACT ASSESSMENT MODEL
Variable Name
Description (English Units1)
Data Source
County Area
Vacant Developable
Vacant Undevelopable
Zoned Residential
Zoned Office
Zoned Industrial
Onsite Restrictions
County-City Data
Book
Planning Agency
Planning Agency
Planning Agency
Planning Agency
Planning Agency
Planning Agency
or Local
Government
Area of county in square miles
Vacant developable acreage expected
in area of analysis in year t2
Vacant undevelopable acreage ex-
pected in area of analysis in
year t
Acres of land expected to be zoned
for residential use in area of
analysis in year t
Acres of land expected to be zoned
for office use in area of analysis
in year t
Acres of land expected to be zoned
for industrial use in area of
analysis in year t
Categorical variable to indicate the
severity of existing or expected
governmental restrictions on on-lot
sewage disposal during the years t
to t+10. Coded3 as follows:
4 = on-lot disposal prohibited
3 = prohibited except on large
lots
2 = permitted but percolation test
required
1 = permitted but package plants
prohibited
0 = no restrictions
The number of years between year t
and year t+10 that it is expected
that any on-site sewage disposal
restrictions will be in effect
(i.e., a value of 0 to 10)
Number of limited access interchanges Planning Agency
expected in county in year t+5
Number of limited access interchanges Planning Agency
expected in the area of analysis in
year t+5
I3ee Table C-3 in Appendix C for corresponding metric units.
2t is the year of major project initiation.
3When restrictions are a mix of the five codes given, use the most restrictive
condition (i.e., the highest value).
Restriction Years
Planning Agency
or Local
Government
County Interchanges
Limited Access
2-20
-------
TABLE 2-5 (CONTINUED)
INPUT DATA REQUIREMENTS FOR IMPACT ASSESSMENT
Variable Name
Description (English Units)
Data Source
Transit Stops
County Growth
Future Population
Population Growth
Office Vacancy
Future Houses
Median Price
Future Income
Future Employment
Future Medicals
CBD Distance
Airport Distance
Track
Number of transit stops (bus and
commuter rail) expected in area of
analysis in year t
Percent1 change in county population
projected for the years t to t+10
Projected SMSA2 population in year
t+10
Percent1 change in SMSA2 population
projected between years t and t+10
Percent3 of office buildings in area
of analysis in year t that are ex-
pected to be vacant
Projected housing units in SMSA2 in
year t+10
Projected median price of 1 acre of
vacant residential land ($) in area
of analysis in year t
Projected median family income in
SMSA2 in year t+10
Projected SMSA2 employment in year
t+10
Projected hospital employment in
SMSA2 in year t+10
Distance in miles'* from centroid of
area of analysis to centroid of
nearest central business district
in year t
Miles from centroid of area of
analysis to centroid of nearest
commercial airport in year t
Miles of railroad track in area of
analysis in year t
Planning Agency
Planning Agency
Planning Agency
or BEA Projec-
tions [28]
Planning Agency
or OBERS
Planning Agency
or Realtor
Planning Agency
or OBERS
Planning Agency
or Realtor
Planning Agency
or OBERS
Planning Agency
or OBERS Pro-
jections [18]
Planning Agency
Highway Road Map
Highway Road Map
USGS Topo-
graphical Map
:A value of 10% is coded as 0.1.
2If the area of analysis is not in an SMSA, use county data.
instances two different variables may end up with the same
3A value of 10% is coded as 10.
4Air miles, not road miles. Central business district used is that in the
In these
value.
nearest city with a population of 100,000 or more.
2-21
-------
TABLE 2-5 (CONTINUED)
INPUT DATA REQUIREMENTS FOR IMPACT ASSESSMENT MODEL
Variable Name
Description (English Units)
Data Source
Area of Analysis
Sewered Land
Interceptors
Collection Capacity
Peak Flow
Treatment Capacity
Population Served
Project Cost
Federal Funds
Phasing
SMSA Area
Area of analysis in acres
Acres of land within 5,000 ft of
the major project interceptor sewer
in the area of analysis in year t
Running length of interceptor sewer
lines in miles going through rela-
tively undeveloped land1 in area of
analysis in year t
Total hydraulic design capacity of
wastewater major project collection
system in million gallons per day
(mgd) in year t2
Anticipated peak flow in the waste-
water major project collection
system in mgd in year t
Total hydraulic design capacity of
the major project wastewater treat-
ment plant in mgd in year t
Population served by the major
project facility in year t
Total major project construction
cost in thousands of $
Federally funded share of major
project cost in thousands of $
Categorical variable to indicate
whether the completion of the col-
lection network will be phased over
several years
1 = phasing will occur
0 = no phasing
Area of SMSA3 in square miles
Facility Plan or
Water Resources
Plan
Facility Plan
Facility Plan
Facility Plan
Facility Plan
Facility Plan
Facility Plan
Facility Plan
Facility Plan
Facility Plan
County-City Data Bk.
See Appendix E for directions on how to gather these data.
^ess than one dwelling unit per acre.
2Year t or up to 5 years after year t if it is a phased project. Note this
model should not be applied to major projects which are not completed by year
t+5 (see Definitons on page 2-19).
3If the area of analysis is not in an SMSA, use county data. In these in-
stances, two different variables may end up with the same value.
2-22
-------
TABLE 2-5 (CONTINUED)
INPUT DATA REQUIREMENTS FOR IMPACT ASSESSMENT MODEL
Variable Name
Description (English Units)
Data Source
Tract Area
Dwelling Units
Current Houses
School Kids
Vacant Houses
Nonmobility
Median Income
Current Income
Poverty
Index One
Index Two
Government
Current Employment
Unemployment
Area of census tracts1 in sq. miles
Census tract housing units in 100s
in year t (round off to nearest
100 units)
Total housing units in SMSA2 in
year t
Population 0-14 years of age in
census tracts in year t
Percent3 vacant available dwelling
units in
Percent3 of families in year t who
were in the same house in year (t-5)
Median income of familes ($) in
county** in year t
Median family income ($) in SMSA2 in
year t
Percent3 of total families with
income below the poverty level in
Consumer Price Index5 for the
year t
Consumer Price Index5 for the year
of federal funding
Total county expenditures in
millions of $ in year t
Total SMSA2 employment in year t
Percent3 unemployment in
Census Tracts Map
U.S. Census
U.S. Census
U.S. Census
U.S. Census
U.S. Census
County-City Data
Book
U.S. Census or
OBERS
U.S. Census
(See footnote 5)
(See footnote 5)
County Government
U.S. Census or OBERS
U.S. Census
See Appendix E for directions on how to gather these data.
Census tracts which most closely approximate the area of analysis. If area
is untracted, use data for the municipality. This applies to all tracted
census variables.
2If the area of analysis is not in an SMSA, use county data. In these in-
stances, two different variables may end up with the same value.
3A value of 10% is-coded as 0.1.
^County containing most of the area of analysis.
5Use U.S. Department of Labor statistics for the nearest major city, based
on 1947-49 average prices being =100.0. If t or the year of federal
funding are not the current year, use consumer price index data for the
current year, adjusted by the expected annual inflation rate.
2-23
-------
TABLE 2-5 (CONTINUED)
INPUT DATA REQUIREMENTS FOR IMPACT ASSESSMENT MODEL*
Variable Name
Description (English Units)
Data Source
Office Workers
Manufacturing
Workers
Current Medicals
Drivers
Office employment in census tracts
in year t
Manufacturing employment in census
tracts in year t
Hospital Employment in SMSA1 in
year t
100s of workers who drive to work
in year t in the county (round off
to nearest 100 workers)
U'.S. Census
U.S. Census
U.S. Census or
OBERS
U.S. Census
See Appendix E for directions on how to gather these data.
If the area of analysis is not in an SMSA, use county data. In these in-
stances, two different variables may end up with the same value.
2-24
-------
3. Instructions
First, decide whether you will be using English or metric units, and
then photocopy a set of the appropriate set of worksheets. Next, collect all the
required data (as described in the previous section), enter the values on Worksheet
1*, perform the indicated operations, and record the answers in the boxes labeled
"Inputs to Worksheets." Finally, fill in the photocopies of Worksheets 2 through
13 using the instructions below.
a. Projection of Future Land Use
Worksheet 2 computes the values of the 9 land use variables. Fill
in all the blanks on the right-hand side of the equations with the appropriate data
from Worksheet 1. Perform the arithmetic operations and record the results in the
blanks on the left-hand side of the equations.
Worksheet 3 computes total land use in the area of analysis. First,
copy the land use projection data from Worksheet 2 boxes into column (1). Next,
enter the projected percentages for disaggregation of residential and commercial
land use in column (2). These values must be between 0.0 and 1.0, and the three
percentages for residential and commercial uses must each sum to 1.0. The per-
centages should take account of local factors which will affect the density of
development in the area of analysis. In the absence of local data, default values
can be estimated using the equations in Worksheet 4 (the input data for Worksheet
4 all come from boxes in Worksheet 1). Next, divide the Area of Analysis (first
box on Worksheet 1) by 10,000 and enter the value in column (3). Final land use
projections (4) are now obtained by taking the product of columns (1), (2) and (3).
Worksheet 5 is another optional worksheet which can be used to
estimate 90% confidence intervals for any one of the land use projections in Work-
sheet 3. First, enter the name of the land use category at the top of the worksheet.
Next, under Predictor Variable Name, list the name of the 5 or 6 predictors from
the appropriate equation in Worksheet 2. Next, enter the values for these predictors
from Worksheet 2 in the spaces to the right of the names. If only 5 predictors are
used, enter 0.0 on line 6. The co-variance data required is summarized in the sta-
tistical output in Appendix A, and Page A-l gives you an index to where to look for
these data. In order to read the statistical output, it is first necessary to
translate predictor names into the acronyms used in the model, and Table C-3 pro-
vides a cross-index to do this. Record the covariance data on Worksheet 5 and per-
form the indicated arithmetic operations. An example of this procedure is shown on
page 2-83 and discussed below. In the example on Page 2-83, a confidence interval
is being estimated for the "Single Family Detached" land use category so this is
recorded at the top of the Worksheet. Since this category evolved from the "Resi-
dential" equation on Worksheet 2 (Page 2-80), we list the names and values of the
6 predictors from the "Residential" equation in the spaces at the top of Worksheet
5. Then we go to Table C-3 (pp. C-16 thru C-18), find the acronyms for the pre-
dictor names, and record these at the top of Worksheet 5 as well. Next, we go to
Page A-l and in the index find that page A-2 contains the covariance data we
need. At the bottom of Page A-2 is a matrix from which we extract values corres-
ponding to a pair of variables. For example, the first covariance value needed in
Worksheet 5 is for ."(D x CD", that is the covariance of variable (T) (which is
VACANT) with itself. From the matrix we find "0.716E+07" at the intersection of
VACANT and VACANT which is then copied onto Worksheet 5, and so forth.
*The discussion in this section will use worksheets for English units. The in-
structions for metric worksheets are identical, except each worksheet number
has a suffix M, i.e., 1 becomes 1M, etc.
2-25
-------
b. Calculation of Motor Vehicle Traffic
The worksheets presented in this and following sections are based
on the revised GEMLUP-I traffic model [11,12]. Due to its simplistic approach,
the appropriateness of transportation parameters used in the model is extremely
important. Thus, although default values are provided for most parameters, it is
considered desirable to use local sources of transportation data whenever possible.
Worksheet 6 is used to compute total vehicle trips in the area of
analysis for two trip purposes - Work and Other. First, compute a value for the
Effective Radius by completing the calculation shown at the top of Worksheet 6.
This definition uses a value based on a circle with an area equivalent to that
of the actual area of analysis. That is,
Effective Radius = ((Area of Analysis/640)/Tt)1/2
Next, enter the values for Projected Land Use in column (1). Enter values for Work
and Other trip generation rates in columns (2) and (4), respectively. If local
data are not available, default values are given in Table 2-6. Compute Work trips
by multiplying columns (1) and (2); compute Other trips by multiplying columns
(1) and (4). Compute Residential Work Trips and Residential Other Trips by sum-
ming the first three rows of columns (3) and (5), respectively. Compute total
Work Trips and Total Other Trips by summing all items in columns (3) and (5).
Worksheet 7 is used to compute vehicle miles traveled (VMT) in
the Area of Analysis and in the "Impact Area". VMT in the Impact Area are defined
to include both the VMT within the Area of Analysis, as well as VMT outside the
Area of Analysis but occurring because of the presence of the major project and
its induced land uses. First, enter values for the Work and Other trip lengths.
If local data are not available, use the following default equations:
Work Trip Length = 0.00447* p°'22 S^'49 (in miles*)
fl
}0.18 s 1.40 + p0.26 s 1.251 ^-n m11esj
J
where: p = Future Population
S-, = Average network vehicle speed for work
trips in miles per hour
Sp = Average network vehicle speed for other
trips in miles per hour
Next, determine a value for the proportion of Work and Other Trip Lengths
that are less than the Effective Radius (default = 0.40). Enter data on the
next four lines from Worksheets 6 and 7 and compute values for the VMT
variables. Next, enter the proportion of VMT occurring in the peak hour of
the day (default = 0.10). The proportion of off-peak hour VMT is 1 minus the
peak hour proportion (default = 0.90). Next, enter the proportions of VMT
traveled on three facility types (exoressways. arterials, local streets) for
*To convert to kilometers for Worksheet 7M, multiply by 1,609,
2-26
-------
TABLE 2-6
DEFAULT TRIP GENERATION RATES*
Land Use Type
Single Family Detached
Single Family Attached
Multiple Family
Large Commercial
Medium Commercial
Small Commercial
Office - Professional
Manufacturing
Wholesale
Education
Other
Recreation
Trips Per Day
Per Measure
Dwelling Units
Dwelling Units
Dwelling Units
1 ,000 Square Feet
(Square Meters)
1 ,000 Square Feet
(Square Meters)
1 ,000 Square Feet
(Square Meters)
1 ,000 Square Feet
(Square Meters)
1,000 Square Feet
(Square Meters)
1,000 Square Feet
(Square Meters)
1,000 Square Feet
(Square Meters)
1,000 Square Feet
(Square Meters)
Acres
(Square Meters)
Work Trip
Rates
1.8
1.5
1.0
0
.0
0
16
(0.17)
5.0
(0.05)
4.0
(0.04)
0
0.13
(0.0013)
0
Other Trip
Rates
9.0
7.0
5.0
40
(0.43)
64
(0.69)
67
(7.2)
0
0
0
4.0
(0.04)
5.0
(0.05)
42
(169,970)
*Values in parentheses are in metric units.
2-27
-------
Work and Other type trips. The six proportions must sum to 1.0, If local trans-
portation data are unavailable, estimates can be made using the procedure described
in Volume III of the GEMLUP-I final report [11], or by using methods presented in
other standard publications [20,21], Next, enter average speeds for peak and off-
peak hours, and the three facility types. If local data are not available, de-
fault values are provided. Next, copy the indicated trip data from Worksheet 6
and compute Sum A and Sum B. These are used to calculate the Heavy Duty Correc-
tion factor, and finally, the proportion of vehicles in various classes1.
Worksheet 8 is used to compute VMT by various categories of area,
time of day and facility type. Enter VMT data in the columns indicated with arrows,
enter proportion data on the appropriate lines from boxes on Worksheet 7, and per-
form the indicated arithmetic operations.
c. Calculation of Mobile Source Emissions
The methodology for computing vehicular emissions is derived from
the EPA publication Mobile Source Emission Factors [19]. Since the worksheets
in this section rely heavily upon the data in that publication, it is essential
that a copy be available before beginning the calculations. A computer program
is available that calculates composite emission factors using data from reference
19, and so could be substituted for Worksheets 9 and 10. A copy of this program
can be obtained from the Office of Transportation and Land Use Policy, U.S.
Environmental Protection Agency, AW-445, 401 M Street S.W., Washington, D.C.,
20460 (202-755-0603).
Worksheet 9 is used to compute an average emission factor (in
grams per mile, or grams per kilometer) for a given set conditions involving ve-
hicle type, speed and pollutant. Since there are 42 vehicle types, 6 vehicle
speeds (reflecting peak or off peak hours, in conjunction with 3 facility types),
and 3 pollutants3, it is necessary to fill out 72 separate copies of Worksheet 9
for a given area of analysis. If one representative speed for the entire area is
used, the total number will be reduced to only 12, but the results will necessarily
be less accurate. The first step in completing Worksheet 9 is to record carefully
all of the descriptive data requested at the top of the worksheet. Next, fill in
the model years in column (2), starting with t + 10 and decreasing down to t - 2.
Values for columns (3), (4), (6) and (7)1* should be transcribed from the EPA
publication [19]5. Next, compute the Total Emission Rate in column (5) by sum-
ming columns (3) and (4). The Model Year Total Emissions in column (8) are then
obtained by taking the product of columns (5), (6), and (7). The average emis-
sion factor is the sum of column (8).
i Note a proportion value for Light Duty Truck/Gas is not calculated since a fixed
value of 0.118 is used.
2 Emission factors for motorcycles are not included in the worksheets since these
are usually negligible in an area total. If a particular case study contains
relatively large amounts of motocycle traffic, emissions can be calculated
using the procedures in reference [19].
3 There are really five criteria pollutants which motor vehicles emit. However,
average emission factors for particulates and sulfur oxides are relatively
invariable. Appropriate values are listed in Table 2-7.
it There are other correction factors for such things as air conditioning (A. ),
vehicle load (L ), trailer towing (U..-DW), and humidity (H-D). These correction
factors make little difference in composite emissions, thus we have not included
them in the worksheets. If deemed significant, the emissions can be corrected
using the procedures in reference [19].
5 When using Worksheet 9M, use the metric value for Ci , in grams per kilometer.
2-28
-------
TABLE 2-7
PARTICULATE AND SULFUR OXIDE EMISSION FACTORS
Particulates Sulfur Oxides
Vehicle Type gr/mi gr/km gr/mi gr/km
Automobile/Gas 0.54 0.33 0.13 0.08
(0.25)* (0.15)*
Light Duty Truck/Gas 0.54 0.33 0.18 0.11
(0.25)* (0.15)*
Heavy Duty/Gas 1.31+ 0.81+ 0.36 0.22
Heavy Duty/Diesel 1.31 0.81 2.80 1.70
*Unleaded gasoline
+Assumes an average of 8 tires per vehicle
Source: [22].
Worksheet 10 is used to calculate a composite emission factor
for a given combination of vehicle speed and pollutant. Since there are 6
vehicle speeds and 5 pollutants, it is necessary to fill out 30 separate
copies of Worksheet 10. The first step in completing Worksheet 10 is to
again carefully record the descriptive data requested at the top. Next,
enter the appropriate Average Emission Factors from Worksheets 9 and the
vehicle class proportions from the bottom of Worksheet 7. Emission factors
for sulfur oxides and particulates are listed in Table 2-7. Column (3) is
the product of columns (1) and (2). Compute the Composite Emission Factor
by summing column (3).
Worksheet 11 is used to calculate Total Motor Vehicle Emis-
sions in both the Area of Analysis and the Impact Area for a given pollutant.
Since there are 5 pollutants, it is necessary to fill out 5 separate copies of
Worksheet 11. The first step is to record the pollutant at the top of the
worksheet. Next, Average Route Speeds from Worksheet 7, corresponding to the
various conditions, should be listed in column (1) and VMT data from Worksheet
8 should be listed in column (2). Next, record in column C3) the values for
Composite Emission Factors from Worksheets 10. Total Emissions in column (4)
are computed by taking the product of columns (2) and (3). Area of Analysis
and Impact Area totals are then obtained by summing column C4) in two parts
and multiplying by the units correction factor of 0,805.
2-29
-------
d. Emission Summary
The emission estimates provided by the worksheets in this
section are totals for an entire year (in pounds or kilograms of pollutant)
in both the Area of Analysis and the Impact Area. These emission projections
correspond to conditions in year t + 10.
Worksheet 12 is used to calculate Stationary Source Emissions
for a given combination of fuel type (oil, gas, or electricity) and pollutant.
Since the first two fuel types each generate 5 types of pollutants, 10 combina-
tions result. When electricity is the fuel, the only "pollutant" is kilowatt-
hours (which are later converted to air pollutant emissions at an off-site
generating plant). Thus, it is necessary to fill out 11 separate copies of
Worksheet 12. The first step in completing Worksheet 12 is to record the fuel
type and pollutant at the top of the worksheet. Next, enter the amounts of
total land use in column (1) from Worksheet 6 (note, Total Manufacturing Land
Use is recorded at the 'bottom). Next, determine for each land use type, the
proportion of the total that will be using the fuel which is recorded at the
top of the worksheet (i.e., this should be a value between 0.0 and 1.0) and
enter these values in column (2). Note that for a given land use type, the
proportion values used in the 11 copies of this worksheet must sum to 1.0.
If local data are not available, national statistics from the U.S. Census
[23,24] may be used instead. Next, record the process, space heating, and
space cooling emission factors in columns (3), (5), and (7), respectively.
These values can be obtained from Tables 2-8 through 2-13*. The space heating
and space cooling emission factors must first be multiplied by the appropriate
degree-day or operating-hour statistics, displayed in Figures 2-2 through 2-4.
Next, a composite Industrial Emission Factor, based on the expected mix of SIC
codes in manufacturing development, should be computed from Table 2-14 and re-
corded at the bottom of the worksheet. Process Emissions, in column (4), are
calculated by taking the product of columns (1), (2), and (3). Space Heating
Emissions, in column (6), are the product of columns (1), (2) and (5), while
Space Cooling Emissions in column (8) are obtained by multiplying columns (1),
(2), and (7). Columns (4), (6), and (8) are then separately summed to obtain
Total Process, Space Heating and Space Cooling Emissions, respectively. Total
Industrial Emissions are obtained by performing the indicated multiplication.
Worksheet 13 is used to calculate total emissions (in
pounds/year or kilograms/year) for each of the 5 criteria pollutants in both
the Area of Analysis and the Impact Area. First, summarize the 11 copies of
Worksheet 12 in rows (1) through (4) and sum to obtain Total Stationary Source
Emissions in row (5). Next, record the values of Total Motor Vehicle Emissions
from Worksheet 11 in rows (6) and (11). Compute the Total Emissions, Area of
Analysis, in row (7), as the sum of rows (5) and (6). Next, enter values
for the Electric Utility Emission Factors in row (8), from either local data
or Table 2-15 (assuming either coal, oil, or gas powered generators). Then,
bring the value for Total Kilowatt-Hours down and record it in each blank in
row (9). The Electric Utility Emissions, in row (10), is then just the product
of rows (8) and (9). Compute the Total Emissions, Impact Area, in row (12),
as the sum of rows (5), (10), and (11).
*If Worksheet 12M is used, all of these emission factors must be converted
from pounds to kilograms by multiplying by 0.454.
2-30
-------
TABLE 2-8
SINGLE FAMILY DETACHED OR ATTACHED LAND USE BASE EMISSION FACTORS
pound of pollutant (or kilowatt-hours) per measure
PM
SO.
CO
HC
NO.
kWh
Measure
ro
i
GO
Space Heating
Electricity
Gas
Oil
Air Conditioning
Central
Electricity
Gas
Room
Electricity
Process
Hot Water
Electricity
Gas
Oil
Cooking
Electricity
Gas
Miscellaneous
2.6 x 10"4 1.5 x 10~5 5.1 x 10"4
2.2 x 10"3 3.2 x 10~2S 1.1 x 10"3
1.8 x 10~4 1.1 x 10~5 3.5 x 10"4
3.0 x 10"1 1.8 x 10"2 6.0 x 10"1
2.5
3.7 x 10"]S 1.2
1.1 x 10"1 6.6 x 10"3 2.2 x 10"1
2.0 x 10"4 2.6 x 10"3
6.6 x 10"4 2.6 x 10"3
1.4 x 10"4 1.8 x 10"3
2.4 x 10"1 3.0
7.5 x 10"1 3.0
8.8 x 10"2 1.1
3.8 dwelling unit-ht.d.d.
dwelling unit-ht.d.d.
dwelling unit-ht.d.d.
4.7 dwelling unit-op.hr.
dwelling unit-op.hr.
a.c. unit-operating hour
1.4x10 dwelling unit-year
dwell ing unit. year
dwelling unit -year
3.5x10 dwelling unit-year
dwel ling unit-year
7.9x10 dwelling unit-year
Note: A 1600 square foot dwelling unit is assumed.
'S1 represents the percent sulfur in oil, by weight.
-------
TABLE 2-9
MULTIPLE FAMILY LAND USE BASED EMISSION FACTORS
Activity
PM
pound
S0x
of
pol
lutant
CO
(kilowatt-hours)
HC
per
measure
NOX
kWh
Measure
ro
co •.
ro
Space Heating
Electricity
Gas
Oil
Air Conditioning
Central
Electricity
Gas
Oil
Room
Electricity
Process
Hot Water
Electricity
1.3
1.2 x 10
1.1 x 10
-4
-3
7.3 x 10"6
1.7 x 10"2S
2.4 x 10
5.7 x 10
-4
-4
9.7 x 10
3.4 x 10
-5
-4
1.2 x 10
1.4 x 10
-3
-3
1.5
6.2 x 10
4.5 x 10
-5
-4
3.7 x 10
6.4 x 10
-6
-3
1.2 x 10
2.2 x 10
-4
-4
5.0 x 10
1.3 x 10
-5
-4
6.2 x 10
5.3 x 10
-4
-4
5.1x10
-1
dwelling unit*ht.d.d.
dwell ing unit- ht.d.d.
dwelling unit-ht.d.d.
dwell ing unit-op.hr.
dwell ing unit«op.hr.
dwell ing unit-op.hr.
a.c. unit-op.hr.
1.1 x 10 dwelling unit*year
Gas
Oil
Cooking & Dryer
Electricity
Gas
Mi seel laneous
2
2
_
1
-
.4 x 10'1
.0
.2 x 10"1
1
2
_
7
-
.4 x ID''
.9 x 10+1S
.2 x 10'3
4
1
_
2
-
.8 x 10
.0
.4 x 10'1
1.9
6.0
_
9.6
-
x 10"'
x 10"1
x 10"2
2
2
_
1
-
.4
.4
.2
dwelling un
dwelling un
+3
3.8 x 10 dwelling un
dwelling un
+3
4.4 x 10 dwell ing un
Note: A 900 square foot dwelling unit is assumed.
'S' represents the percent sulfur in oil, by weight.
-------
TABLE 2-10
COMMERCIAL AND WHOLESALE LAND USE BASED EMISSION FACTORS
pound of pollutant (or kilowatt-hours) per measure
Activity PM SOV CO HC N0¥ kWh
X X
Space Heating
Electricity - - - - - 1.3
Gas 9.8 x 10"5 5.9 x 10"6 2.0 x 10~4 7.8 x 10"5 9.8 x 10"4 -
Oil 1.7 x 10"3 1.2xlO~2S 2.9 x 10"4 3.3 x 10"2 4.4 x 10"3 -
Air Conditioning
Electricity - - - - - 5.2
Process
Hot Water
Electricity - - - - - 5.0xl02
Gas 2.4 x 10"2 1.4xlO"3 4.8 x 10"2 1.9xlO~2 2.4 x 10"1
Oil 5.2 x 10"1 3.65 9.1 x 10"2 l.OxlO*1 1.4
Lighting - - - - - 8.0 x TO3
Auxiliary - - - - - 3.6 x 103
Equipment
Appliances - - - - - 2.0 x 10
Refrigeration - - - - - 8.9 x 10
Measure
3 2
NT ft
103ft2
103ft2
3 2
itrfr
103ft2
3 2
3 2
i(rfr
3 ?
lo^fr
103ft2
103ft2
3 2
l(Tft
•ht.d.d.
•ht.d.d.
•ht.d.d.
•cl.d.d.
•year
•year
•year
•year
•year
•year
•year
Note: 'S' represents the percent sulfur in oil, by weight.
-------
TABLE 2-11
OFFICE-PROFESSIONAL LAND USE BASED EMISSION FACTORS
ro
u>
Activity
Space Heating
Electricity
Gas
Oil
Air Conditioning
Electricity
Gas
Oil
Process
PM
_
9.4 x 10"5
1.7 x 10"3
_
7.4 x 10"5
1.3 x 10"3
-
pound of pollutant (or kilowatt-hours) per measure
SO CO HC NO kWh
X A
_
5.6 x 10"6
1.2 x 10"2S
_
4.4 x 10"6
9.1 x 10"3S
-
_
1.9 x 10"4
2.9 x 10"4
_
1.5 x 10"4
2.3 x 10"4
-
_
7.5 x 10"5
3.3 x 10"2
..
5.9 x 10"5
2.6 x 10"2
-
1.9
9.4 x 10"4 -
4.4 x 10"3 -
1.5
7.4 x 10"4 -
3.4 x 10"3 -
2.8 x 10+4
Measure
103ft2-ht.d.d.
103ft2-ht.d.d.
103ft2-ht.d.d.
103ft2«cl.d.d.
103ft2-cl.d.d.
103ft2* cl .d.d.
103ft2-year
Note: 'S1 represents the percent sulfur in oil, by weight.
-------
ro
i
CO
en
TABLE 2-12
EDUCATION LAND USE BASED EMISSION FACTORS
Activity
Space Heating
Electricity
Gas
Oil
Air Conditioning
Electricity
Gas
Oil
Process
_
8.
1.
_
2.
4.
-
PM
0 x IO"5
2 x IO"3
3 x 10"5
1 x 10
pound of pollutant (or kilowatt-hours) per measure
SO CO HC NO kWh Measure
X X
_
4.8
8.5
_
1.4
2.8
-
x 10"6
x IO"3
x 10"6
x IO"3
_
1.6 x
2.1 x
_
4.6 x
7.1 x
-
ID'4
io-4
io-5
ID'5
_
6.4
2.4
_
1.8
8.0
-
x IO"5
x IO"2
x IO"5
x IO"3
-
8.
3.
-
2.
1.
-
0x10
2 x IO"3
3 x IO"4
1 x IO"3
1.7 103ft2 • ht.d.d.
103ft2 • ht.d.d.
103ft2 • ht.d.d.
4.7 x IO"1 103ft2 ' cl.d.d.
103ft2 * cl.d.d.
103ft2 • cl.d.d.
7.1 x 103 103ft2 'year
Note: 'S' represents the percent sulfur in oil, by weight,
-------
ro
u>
01
TABLE 2-13
OTHER LAND USE BASED EMISSION FACTORS
Activity
PM
pound of pollutant (or kilowatt-hours) per measure
SO CO HC NO kWh
X *»
Measure
Space Heating
Air
Electricity
Gas
Oil
Conditioning
Electricity
Gas
Oil
_
9.4 x 10"5
1.4 x 10"3
_
2.3 x 10"5
4.1 x 10"4
mm
5.6 x 10"6
9.9 x 10"3S
-
1.4 x 10"6
2.8 x 10"3S
_
1.9 x 10'4
2.5 x 10"4
-
4.6 x 10"5
7.1 x 10"5
_
7.5 x 10"5
2.8 x 10"2
-
1.8 x 10"5
8.0 x 10"3
_
9.4 x 10"4
2.8 x 10"2
-
2.3 x 10"4
1.1 x 10"3
Process _____
1.7 1
1
1
_1
4.7 x 10 ' 1
1
1
1.2 x 10+4 1
3 2
0 ft
03ft2
03ft2
•3 0
vn
n3 2
03ft2
3 2
•ht
•ht
•ht
•cl
•cl
-------
TABLE 2-14
ESTIMATED NATIONAL INDUSTRIAL LAND USE BASED EMISSION
FACTORS BY TWO DIGIT 1967 STANDARD INDUSTRIAL CLASSIFICATION CODE
Pounds of pollutant (or
SIC Partic-
Code ulates SO
J\
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
10 &
640
1220
580
60
60
110
3120
10
100
1060
510
170
4030
3060
140
220
220
680
950
39 80
500
1020
540
40
70
80
3090
20
460
2780
380
170
2670
2380
120
180
200
480
700
130
3 2
KWH of electricity) per 10 ft floor area per year
CO HC NO KwH
X
13
25
14
1.4
3.4
2.2
69
0.68
11
55
10
4.7
720
61
3.5
4.7
5.3
13
18
3.5
3.3
14
8.1
0.84
2.3
1.2
40
0.48
8.1
38
5.8
2.9
38
34
2.1
2.7
3.2
6.8
9.5
2.4
130
230
140
15
45
21
690
9.5
160
730
97
52
610
570
36
46
56
110
150
44
38,000
48,000
68,000
16,000
22,000
14,000
85,000
25,000
181,000
426,000
50,000
18,000
78,000
297,000
33,000
31,000
56,000
54,000
38,000
31 ,000
Note: The following is assumed: 2% sulfur in coal
10% ash in coal
0.2% sulfur in distillate oil
1.75% sulfur in residual oil
1967 SIC codes are used because of data availability. The
1972 SIC code manual provides conversions between 1967 and
1972 codes [25].
2-37
-------
Figure 2-2: NORMAL SEASONAL HEATING DEGREE DAYS ( BASE 65°F ) 1941-1970
ro
i
CO
co
-------
ro
GO
Figure 2-3: NORMAL SEASONAL COOLING DEGREE DAYS ( BASE 65°F ) 1941-1970
-------
ro
-fa.
o
FIGURE
2-4: ANNUAL AIR CONDITIONER COMPRESSOR-OPERATING HOURS FOR HOMES THAT ARE NOT
NATURALLY VENTILATED.
-------
TABLE 2-15
TYPICAL UNCONTROLLED EMISSION FACTORS FOR ELECTRIC UTILITIES
coal
oil
gas
PM
5.23
6.34
1.19
Ibs. emissions per kWh sold to customer
SO.. CO HC
•3
x 10' JA
x 10"4
x 10"4
x ,,
1.53 x
1.26 x
7.13 x
10 "S
10"2S
io-6
4.03 x
2.38 x
2.02 x
-4
10 '
io-4
io-4
1
1
1
.21 x
.58 x
.19 x
_4.
10 4
ID'4
io-5
NO
x 9
2.21 x
8.32 x
8.32 x
10 u
io-3
io-3
ro
Note: A 33.3% overall plant efficiency is assumed for coal-fired plants [26].
A 31.6% overall plant efficiency is assumed for oil- and gas-fired plants [26]
A ]Q% transmission loss 1s assumed [27].
'S' and 'A1 represent, respectively, the percent sulfur and ash in the fuel
by weight.
-------
4. English Unit Worksheets
Summarized in this section are all English unit worksheets,
Metric unit worksheets begin on Page 2-60.
2-42
-------
WORKSHEET 1
•INPUT DATA RECORD IN ENGLISH UNITS
Variable Name
Value/Computation
Inputs To Worksheets Units
Area of Analysis = |_
Vacant Developable
Vacant Undevelopable
Vacant Land = Vacant Developable T (Area of Analysis - Vacant Undevelopable) = |_
Median Price
Median Income
Land Cost = Median Price * Median Income = |_
Collection Capacity
Peak Flow = |_
% Collection Reserve = 100* (Collection Capacity - Peak Flow) v Peak Flow = f
Manufacturing Workers
Tract Area =
Manufacturing Density = Manufacturing Workers * Tract Area = [_
Normobility = |~
Drivers
T3 County Area
00 Driver Density
Sewered Land
Sewer Service
School Kids
Dwelling Units
Kid Density
Limited Access
Interchanges
Current Employment
Future Employment
SMSA Area
Employment Growth
Track
Railroads
= Drivers -f County Area
= Sewered Land - Area of Analysis
Schools Kids T Dwelling Units
= 640* Limited Access * Area of Analysis
= (Future Employment - Current Employment) - SMSA Area
Acres
Acres
Acres
= 640* Track •? Area of Analysis
Zoned Office
Million Gallons Per Day
Million Gallons Per Day
*
Employees
Square Miles
Employees Per Square Mile
*
100s of Drivers
Square Miles
100s of Drivers Per Square Mile
Acres
*
Children
100s of Dwell ing Units
Children Per 100 Dwelling Units
Interchanges
Interchanges Per Square Mile
Employees
Employees
Square Miles
Employees Per Square Mile
Miles
Railroad Miles Per Square Mile Land
Area
Acres
Unitless
ENGLISH
-------
WORKSHEET 1 (CONTINUED)
INPUT DATA RECORD IN ENGLISH UNITS
Variable Name
Value/Computation
Inputs to Worksheets Units
Office Zoning
House Density
Zoned Industrial
Industrial Zoning
Population Growth
Office Vacancy
Airport Distance
Office Workers
Office Employees
Future Population
Employee Ratio
Unemployment
Future Income
Government
Collection Reserve
Interceptors
Interceptor Density
Poverty
Onsite Restrictions
County Growth
Project Cost
Federal Funds
Index One
Index Two
Population Served
= Zoned Office r Area of Analysis
= 100* Dwelling Units T Tract Area
= Zoned Industrial * Area of Analysis
Office Workers •=• Tract Area
= Future Employment * Future Population
= Collection Capacity - Peak Flow
640* Interceptors T Area of Analysis
Dwelling Units Per Square Mile
Acres
Miles
Employees
Employees Per Square Mile
People
J Millions of $
Million Gallons Per Day
Miles
Miles of Interceptor Pipe Per Square
Mile Land Area
Thousand
Thousand
People
Unitless
ENGLISH
-------
WORKSHEET 1 (CONTINUED)
INPUT DATA RECORD IN ENGLISH UNITS
Variable Name Value/Computation
S ewe i" Costs = 1,000* ((Project Cost T Index One) - (Federal Funds - Index Two)) T Population Served =
Treatment Capacity
Vacant Houses
County Interchanges
Interchange Density = Interchanges r (640* County Interchanges -f County Area) =
Zoned Residential
Current Income
Income Growth = Future Income - Current Income =
Current Medicals
Future Medicals
Current Houses
Future Houses
Housing Growth = (Future Houses - Current Houses) T Current Houses =
Restriction Years
Phasing =
Transit Stops =
CBD Distance
Inputs to Worksheets
Units
$ Per Person
Million Gallons
Per Day
*
Interchanges
*
Acres
*
$
$
Employees
Employees
*
Dwelling Units
Dwelling Units
*
Years
it
it
Miles
Unitless
ENGLISH
-------
WORKSHEET 2
LAND USE PROJECTIONS
ro
i
Residential
Commercial
Office-
Professional
Manufacturing
Wholesale
Highways
Education
Recreation
Other
= (-8692 *
- (18804 *
= (4302 *
- (17.04 *
= (719.5 *
- (1791 *
= (2591 *
+ (234.7 *
= (166.5 *
+ (6949 *
= (35.22 *
+ (25.94 *
= (-3.186 *
- 405.2 *
= (270.4 *
- (3226 *
= (1134 *
- (245.4 *
Vacant
Land )
Nonmobility)
Sewer
Service)
Employment
Growth )
Railroads)
Industrial
Zoning )
Railroads)
Office
Vacancy)
Driver
Density)
Unemployment)
Railroads)
Interceptor
Density )
Poverty)
County
Growth)
Railroads)
Vacant
Houses)
Railroads)
Interchange
Density )
+ (5347 *
+ (487.2 *
- (4515 *
- (2.626 *
+ (33478 *
+ (572.0 *
- (3890 *
+ (22.7 *
+ (86.28 *
- (0.1065
+ (22.98 *
+ 11.35 =
+ (723.0 *
- (1.415 *
+ (2.631 *
- (557.8 *
- (0.3020
+ (403.8 *
Land
Cost)
Driver
Density)
Vacant
Land )
%Col lection
Reserve
Office
Zoning)
Population
Growth )
Vacant
Land )
Airport
Distance)
Office
Vacancy)
Future
* Income)
Land
Cost)
Mil f
10,000
Sewer
Service)
Sewer
Costs)
Treatment
Capacity )
Industrial
Zoning )
Future
* Income)
Residential
Zoning )
+ (16.42 *
+ 134P6 =
- (28.84 *
+ 5376 =
+ (36.70 *
- 459.9 =
+ (2.978 *
+ 611.2 =
(
+ (3.340 *
- 1911 =
- (0.1167
Per
Acres
+ (437.5 *
+ 428.9 =
+ (322.1 *
+ 59.52 =
- (12175 *
+ 3087 =
% Collection
Reserve )+ (11.24 *
- • Dwnl linn Unite Por
10,000 Acres*
Kid
Density) + (6249 *
Per 10,000 Acres
Peak Flow) + (0.3324 *
Per 10,000 Acres
Manufacturing
Density ) + (85923 *
Per 10,000 Acres
Office
Employees) + (6043 *
Per 10,000 Acres
Government) + (0.5379 *
Land
Cost) + (95.17 *
Per 10,000 Acres
County
Growth) + (0.9742 *
10,000 Acres
Vacant
Houses) + (649.0 *
Per 10,000 Acres
Manufacturing
Density
Interchanges)
House
Density)
Office
Zoning)
Employee
Ratio )
Collection
Reserve )
On-Site
Restrictions)
Office
Employees)
County
Growth)
10,000 Acres refers to Area of Analysis
ENGLISH
-------
WORKSHEET 3
FINAL LAND USE PROJECTIONS
(1)
Land Use Projections
From Worksheet 2
(2)
Disaggregation
Percentages
(3)
Area of Analysis
T By 10,000
(4)
Final Land Use Projection
(1) * (2) * (3)
(Per Area of Analysis^
Residential
% Single Family
Total Single Family Detached =
Dwelling Units
Residential
% Two Family
Total Single Family Attached =
Dwelling Units
•Residential
% Multi Family
Total Multiple Family =
Dwelling Units
Commercial
% Large Commercial '
Total Large Commercial =
1,000 Square Feet
Commercial
% Medium Commercial 2
Total Medium Commercial =
1,000 Square Feet
Commercial
% Small Commercial 3
Total Small Commercial =
1,000 Square Feet
Office-Professional
Manufacturing
Wholesale
Education
Other
Recreation
Highways
I
Total Office-Professional =
Total Manufacturing =
Total Wholesale =
Total Education =
Total Other =
Total Recreation =
Total Highways =
1,000 Square Feet
1,000 Square Feet
1,000 Square Feet
1,000 Square Feet
1,000 Square Feet
Acres
Lane Miles
i
Large commercial = commercial development with floor area > 100,000 ft .
7 2
Medium commercial = commercial development with floor area between 50,000 and 100,000 ft
o 2
Small commercial = commercial development with floor area < 50,000 ft .
ENGLISH
-------
WORKSHEET 4 (Optional)
DEFAULT DISAGGREGATION EQUATIONS
% Single Kid
Family =!-(!+ 92.8 * EXP (-0.0137 * Density
Income
- 0.000918 * Growth
Sewer
- 0.784 * Service
Office
- 22.7 * Zoning
2.98 * Nonmobility
Hospital
0.181 * Growth
SMulti
Family
Single
= (1-% Family
Office
+ 34.1 * Zoning
Housing
) * (1 4 (1 + 1.24* EXP (-438 * Growth
+ 5.48 * Poverty
- 0.732 * Railroads
- 0.0365 * Peak Flow
% Two
Family = (1-X Single Family
- % Multiple Family
% Large Office
Commercial =1*0+ 0.104 * EXP (0.215 * Vacancy
Onsite
- 0.507 * Restrictions
Kid
- 0.0159 * Density
Restriction
0.0603 * Years ))
Treatment
+ 0.0232 * Capacity
% Medium Large
Commercial = (1-% Commercial
CBD
+ 0.0178 * Distance
_)*(!*(!+ 0.585 * EXP (1.54 * Phasing
+ 0.00417 * Government - 1.48 * Vacant Land
Transit
+ 0.00459 * Stops
% Small
Commircial = (1-3J Large Commercial
- % Medium Commercial
) =
Note: EXP is the exponential function. Thus EXP(X) = e .
ENGLISH
-------
WORKSHEET 5 (OPTIONAL)
CONFIDENCE INTERVALS FOR PREDICTED LAND USE
Final Projected Land Use Category
PREDICTOR VARIABLE
Name
1.
2.
3.
4.
5.
6.
Constant Term
COMPUTING VARIANCE OF DEPENDENT
Covariance
Predictor Term from
Variables Appendix A
0 x
-------
Covariance
Predictor Term from
Variables Appendix A
© x ® x
(D x (D x
® x © x
(D x (5) x
(D x © x
-------
WORKSHEET 6
MOTOR VEHICLE TRIPS
Effective Radius (in miles) = 0.0223 *-^(Area of Analysis, in acres)
Miles
(1)
Total Land Use
Single Family
Detached
Single Family
Attached
Multiple
Family
(2)
Work Trip
Rates
(3)
Work Trips
(1)*(2)
(4)
Other Trip
Rates
(5)
Other Trips
(1)*(4)
Residential
Work Trips
Residential
Other Trips
Large
Commercial
Medium
Commercial
Small
'Commercial
Office-
Professional
Manufacturing
Wholesale
Education
Other
Recreation
i
Total Work
Trips
Total Other
Trips
*See Page 2-26 for a definition of Effective Radius.
2-51
ENGLISH
-------
WORKSHEET 7
VEHICLE MILES TRAVELED (VMT)
Work Trip
Length Miles
Other Trip Work (Default Other
Length Miles Proportion = 0.40) Proportion
VMT?,, [= (Total Work * Work Trip ) - (Residential
IU Trips
Length Work Trips
VMTflu | = (Total Work ^ Effective ) - (Residential
AW Trips
Radius Work Trips
VMTT(V |= (Total Other * Other Trip ) - (Residential
10 Trips
Length Other Trips
VMTnn |= (Total Other * Effective ) - (Residential
AU Trips
Peak Hour Proportion
Facility
Work
Expressways
Arterial s
Local Streets
Trip Sum A I = Manul
Radius Other Trips
(Default = 0.10) Off Peak Hour Pro
* Work
Proportion
* Work
Proportion
* Other
Proportion
* Other
Proportion
portion (
(Default
= 0.40)
* Effective )
Radius
* Effective )
Radius
* Effective )
Radius
* Effective )
Radius
3efault = 0.90)
Proportions Average Route Speeds (miles per hour)
Other Peak
(
(
(
"acturing Work + Manufacturing Other
Off Peak
Default = 37)
Default = 20)
Default = 15)
+ Wholesale Work
(Default = 45)
(Default = 28)
(Default = 18)
+ Wholesale Other
Trip Sum B [ |= Total Work Trips + Total Other Trips
HD Correction | = (!
Automobile/Gas Proportion 1
Heavy Duty/Gas Proportion |
Heavy Duty/ Diesel Proportion
>um A - Sum B) - 0.05, if less than 0, set equal to 0.
1 = 0.804 - HD Correction
1 = 0.346 + 0.3 * KD Correction
[ |= 0.032 + n.2 * HD Correction
*Subscripts for VMT variables are defined as follows: I = impact area, A = area of analysis, W = work trips, 0 = other trips.
ENGLISH
-------
AREA OF ANALYSIS
VMTAp, | 1 = Peak
VMTAp/,.| |= Peak
VMTAp. | |= Peak
VMTAdEl |= Off Peak
VMTAtfifl 1= Off Peak
VMTA(f| ] I = Off Peak
IMPACT AREA
VMTIprl |= Peak
VMT.pA.| |= Peak
VMT.p. | |= Peak
VHTTny | |= Off Peak
VMTI0v| |= Off Peak
VMT7n-, | |= Off Peak
ilOKK-'lE'/T J
CAiilCORI~ r
U"irA,,i
* ( * Expressways
Work
* ( * Arterials
Work
* ( * Local
Work
* ( * Expressways
Work
* ( * Arterials
Work
* ( * Local
Work
* ( * Expressways
Work
* ( * Arterials
Work
* ( * Local
Work
* ( * Expressways
Work
* ( * Arterials
Work
* ( * Local
Work
4 v .
-------
Vehicle Type =
Pollutant =
Speed =
WORKSHEET 9
MOTOR VEHICLE EMISSION FACTORS
(Automobile Gas, Light Duty Truck Gas, Heavy Duty Gas, or Heavy Duty Diesel)
Region = (Low alt., High alt., or Calif.)
°F
(CO, NO. or HC)
A
miles per hour
Year of Impact Assessment (t+10) =
Ambient Temperature =
Cold Starts =
Hot Starts =
(1)
Vehicle
Age
(Years)
1
2
3
4
5
6
7
8
9
10
11
12
>13
(2)
Model
Year
(3)
Base Emission
Rate
«W
(4)*
Hydrocarbon Crankcase/
Evaporate Emission Rate
(H^
(5)
Total Emission
Rate
(3)+(4)
(6)
Fraction of
Travel
(7)
Speed/Temp. /Cold-Hot
Starts Correction Factor
'ripstwx'
(8)
Model Year
Total Emissions
(5)*(6)*(7)
IS}
en
Note Hi = 0 for CO and NOX; for these pollutants,
enter the values for C- directly in column (5).
Average Emission Factor
-------
WORKSHEET 10
COMPOSITE MOTOR VEHICLE EMISSION FACTORS
Speed
(miles per hour)
Pollutant
(CO, NOV, HC, SOV or Particulates)
X X
Vehicle Class
(1)
Average Emission
Factor
(2)
Vehicle Class
Proportion
(3)
Product
Automobile/Gas
Light Duty Truck/Gas
Heavy Duty/Gas
Heavy Duty/Diesel
0.118
Composite Emission Factor
2-55
ENGLISH
-------
WORKSHEET 11
TOTAL MOTOR VEHICLE EMISSIONS
Pollutant
Condition
Peak, Expressways
Peak, Arterial s
Peak, Local Streets
Off Peak, Expressways
Off Peak, Arterial s
Off Peak, Local Streets
(1)
Average
Speed
i
(2)
VMT
Data
VMTAp£
VMTAPA-
VMTAPL
VMTA(JE
VMTAO*
VMTA(JL
Area of Analysis, Total Motor Vehicle Emissions
Peak, Expressways
Peak, Arterial s
Peak, Local Streets
Off Peak, Expressways
Off Peak, Arterial s
Off Peak, Local Streets
VWIPE
VHTIpA.
VHTIpL
VMTIO'E
VMTIOA-
VHTIOL
(3)
Emission
Factors
Sum
(4)
Total Emissions
(2)*(3)
!
* 0.805
=
(in pounds/year)
Sum
*
Impact Area, Total Motor Vehicle Emissions -
C1n p<
|
0.805
•
)unds/year)
2-56
ENGLISH
-------
WORKSHEET 12
STATIONARY SOURCE EMISSIONS
r-o
i
en
Fuel Type
(Gas, Oil, or Electricity)
Pollutant
(CO, HC, NO. SOV, Particulates, or Kwh)
X X
TOTAL EMISSIONS
(1)
Total Land Use
SF Detached
SF Attached
Mult. Family
Large Commercial
Med. Commercial
Small Commercial
Office-Professional
Wholesale
Education
Other
(2)
Fuel
Proportion
(3)
Process
Emission
Factor
(4)
Process
Emissions
(D*(2)*(3)
(5)
Space Heating
Emission
Factor
(6)
Space Heating
Emissions
(D*(2)*(5)
(7)
Space Cooling
Emission
Factor
(8)
Space Cooling
Emissions
(D*(2)*(7)
Process
Space Heating
Space Cooling
Industrial
= Total Manufacturing Land Use
* Fuel Proportion
* Industrial Emission Factor
ENGLISH
-------
5. Metric Unit Worksheets
Summarized in this section are all metric unit worksheets.
2-59
-------
WORKSHEET 1M
INPUT DATA RECORD IN METRIC UNITS
Variable Name
Value/Computation
Inputs to Worksheets Units
CTl
O
Area of Analysis
Vacant Developable
Vacant Undevelopable
Vacant Land
Median Price
Median Income
Land Cost
Collection Capacity
Peak Flow
% Collection Reserve
Manufacturing Workers
Tract Area
Manufacturing Density
Nonmobility
Drivers
County Area
Driver Density
Sewered Land
Sewer Service
School Kids
Dwelling Units
Kid Density
Limited Access
Interchanges
Current Employment
Future Employment
SMSA Area
Employment Growth
Track
Railroads
Zoned Office
= Vacant Developable * (Area of Analysis - Vacant Undevelopable)
= Median Price T Median Income
= 100* (Collection Capacity - Peak Flow) -. Peak Flow
Manufacturing Workers * Tract Area
= Drivers •=• County Area
= Sewered Land T Area of Analysis
= School Kids v Dwelling Units
= 1,000,000* Limited Access T Area of Analysis
= (Future Employment - Current Employment) * SMSA Area
= 1,000,000* Track r Area of Analysis
Square Meters
Square Meters
Square Meters
Million Gallons Per Day
Million Gallons Per Day
*
Employees
Square Kilometers
Employees Per Square Kilometers
*
100s of Drivers
Square Kilometers
100s of Drivers Per Square Kilometer
Square Meters
*
Children
100s of Dwelling Units
Children Per 100 Dwelling Units
Interchanges
Interchanges Per Square Kilometer
Employees
Employees
Square Kilometers
Employees Per Square Kilometer
Kilometers
Railroad Kilometers Per Square
Kilometer Land Area
Square Meters
Unitless
METRIC
-------
WORKSHEET 1M (CONTINUED)
INPUT DATA RECORD IN METRIC UNITS
Variable Name
Value/Computation
Inputs to Worksheets Units
Office Zoning
House Density
Zoned Industrial
Industrial Zoning
Population Growth
Office Vacancy
Airport Distance
Office Workers
Office Employees
Future Population
Employee Ratio
Unemployment
Future Income
Government
Collection Reserve
Interceptors
Interceptor Density
Poverty
Onsite Restrictions
County Growth
Project Cost
Federal Funds
Index One
Index Two
Population Served
= Zoned Office * Area of Analysis
= 100* Dwelling Units v Tract Area
= Zoned Industrial T Area of Analysis
Office Workers -f Tract Area
= Future Employment * Future Population
= Collection Capacity - Peak Flow
1,000,000* Interceptors - Area of Analysis
Dwelling Units Per Square Kilometer
Square Meters
Kilometers
Employees
Employees Per Square Kilometer
People
Mill ions of $
Mi 11 on Gallons Per Day
Kilometers
Kilometers of Interceptor Pipe Per
Square Kilometer Land Area
Thousand $
Thousand $
People
Unitless
METRIC
-------
WORKSHEET 1M (CONTINUED)
INPUT DATA RECORD IN ENGLISH UNITS
Variable Name
Value/Computation
Inputs to Worksheets Units
Sewer Costs
Treatment Capacity
= 1,000* {(Project Cost * Index One) - (Federal Funds * Index Two)) * Population Served =
Vacant Houses
County Interchanges
Interchange Density = Interchanges -f (640* County Interchanges * County Area)
Zoned Residential
Residential Zoning = Zoned Residential * Area of Analysis
Current Income
Income Growth
Current Medicals
Future Medicals
Hospital Growth
Current Houses
Future Houses
Housing Growth
Restriction Years
Phasing
Transit Stops
CBD Distance
Future Income-Current Income
= (Future Medicals - Current Medicals) * Current Medicals
= (Future Houses - Current Houses) * Current Houses
$ Per Person
Million Gallons
Per Day
Interchanges
*
Square Meters
Employees
Employees
*
Dwell ing Units
Dwell ing Units
Years
*
Kilometers
Unitless
METRIC
-------
WORKSHEET 2M
LAND USE PROJECTIONS
CTl
CO
Residential
Commercial
Office-
Professional
Manufacturing
Wholesale
Highways
Education
Recreation
Other
= (-2.148 *
- (4.646 *
= (987.5 *
- (10.13 *
= (102.7 *
- (411.0 *
= (369.6 *
+ (53.87 *
= (98.23 *
+ (1595 *
= (0.0087 *
+ (25.94 *
= (-731.3 *
- (93.02 *
= (168.1 *
- (3226 *
= (161.8 *
- (67.81 *
Vacant
Land )
Nonmobil ity)
Sewer
Service)
Employment
Growth )
Railroads)
Industrial
Zoning )
Railroads)
Office
Vacancy)
Driver
Density)
Unemployment)
Railroads)
Interceptor
Density )
Poverty)
County
Growth)
Railroads)
Vacant
Houses)
Railroads)
Interchange
Density )
+ (1.321 *
Land
Cost)
Driver
+ (0.3117 * Density)
- (1036 *
Vacant
Land )
%Col lection
- (0.6029 * Reserve )
+ (7685 *
+ (131.3 *
- (893.0 *
+ (3.239 *
+ (19.80 *
- (0.0244 *
+ (0.0091 *
+ 0.0045 =
+ (170.0 *
- (0.3248 *
+ (2.631 *
- (557.8 *
- (0.0693 *
+ (92.70 *
Office
Zoning)
Population
Growth )
Vacant
Land )
Airport
Distance)
Office
Vacancy)
Future
Income)
Land
Cost)
10,000
Sewer
Service)
Sewer
Costs)
Treatment
Capacity )
Industrial
Zoning )
Future
Income)
Residential
Zoning )
+ (0.0041 *
+ 3.327 =
- (6.621 *
+ 1234 =
+ (8.424 *
- 105.6 =
+ (1.770 *
+ 140.3 =
+ (1.985 *
- 438.7 =
- (0.00005
lometers Per
Square Meter
*• (100.4 *
+ 98.45 =
+ (322.1 *
+ 122.1 =
- (2795 *
+ 708. r> =
%Collection Manufacturin
Reserve ) + (0.0072 * Density
10,000 Acres*
Kid
Density) + (3714 * Interchanges)
Per 10,000 Square Meters
House
Peak Flow) + (0.1981 * Density)
Square Meters Moor Area
Manufacturing Office
Density ) + (19724 * Zoning)
Per 10,000 Square Meters
Office Employee
Employees) + (1387 * Ratio )
Per 10,000 Square Meters
Collection
* Government) + (0.0002 * Reserve )
5
Land Ons ite
Cost) + (21.85* Restrictions)
«•
Per 10,000 Square Meters
County Office
Growth) + (2.522 * Employees)
Per 10,000 Square Meters
Vacant County
Houses) + (150.0 * Growth)
c>> ,., Uni. rl „ ,
Per 10,000 Square Meters
10,000 Square Meters refers to Area of Analysis
METRIC
-------
WORKSHEET 3M
FINAL LAND USE PROJECTIONS
(1)
Land Use Projections
From Worksheet 2M
(2)
Dlsaggregation
Percentages
(3)
Area of Analysis
4 By 10,000
(4)
Final Land Use Projection
(1) * (2) * (3)
(Per Area of Analysis
Residential
% Single Family
Total Single Family Detached =
Dwelling Units
Residential
% Two Family
Total Single Family Attached =
Dwelling Units
Residential
% Multi Family
Total Multiple Family =
Dwelling Units
Commercial
% Large Commercial
1
Total Large Commercial
Square Meters
Commercial
% Medium Commercial
Total Medium Commercial =
Square Meters
Commercial
% Small Commercial
Total Small Commercial =
Square Meters
IS)
I
en
Office-Professional
Manufacturing
Wholesale
Education
Other
Recreation
Highways
Total Office-Professional =
Total Manufacturing
Total Wholesale =
Total Education =
Total Other =
Total Recreation
Total Highways =
Square Meters
Square Meters
Square Meters
Square Meters
Square Meters
Square Meters
Lane Kilometers
1 2
Large commercial = commercial development with floor area> 9,290 m .
2 2
Medium commercial = commercial development with floor area between 4,645 and 9,290 m .
3 2
Small commercial = commercial development with floor area< 4,645 m .
METRIC
-------
WORKSHEET 4M (Optional )
DEFAULT DISAGGREGATION EQUATIONS
% Single
Family
Kid
= 1 T (1 + 92.8 * EXP (-0.0137 * Density
Income
- 0.000918 * Growth
Office
- 22.7 * Zoning
+ 2.98 * Nonmobility
Hospital
+ 0.181 * Growth
Sewer
- 0.784 * Service
J)
% Multi
Family
Single
(1-% Family
Office
+ 34.1 * Zoning
Housing
J * (1 T (1 + 1.24* EXP (-438 * Growth
+ 5.48 * Poverty
- 0.454 * Railroads
- 0.0365 * Peak Flow
% Two
Family
= (1-% Single Family
- % Multiple Family
CTi
% Large
Commercial
Office
= 1 T (1 + 0.104 * EXP (0.215 * Vacancy
Onsite
- 0.507 * Restrictions
Kid
- 0.0159 * Density
Restriction
+ 0.0603 * Years
J)
Treatment
+ 0.0232 * Capacity
% Medium
Commercial
Large
(]-% Commercial
CBD
+ 0.0111 * Distance
J * (1 T (1 + 0.585 * EXP (1.54 * Phasing
+ 0.00417 * Government , - 1.48 * Vacant Land
Transit
+ 0.00459 * Stops
J))
% Small
Commercial
(1-% Large Commercial
- % Medium Commercial
Note: EXP 1s the exponential function. Thus EXP(X) = e .
METRIC
-------
WORKSHEET 5M,(OPTIONAL)
CONFIDENCE INTERVALS FOR PREDICTED LAND USE
Final Projected Land Use Category
PREDICTOR VARIABLE
Name
2.
3.
4.
5.
6.
Constant Term
7. 1.0
COMPUTING VARIANCE OF DEPENDENT VARIABLE
(T) x (f) x Covariance
(T) x (?) x Covariance
(l) x (§) x Covariance
(I) x (4) x Covariance
(T) x (5) x Covariance
(T) x (e) x Covariance
(V) x ^7/ x Covariance
(?) x (T) x Covariance
(?) x (2) x Covariance
(?) x (3) x Covariance
(2) .x (T) x Covariance
(?) x (£) x Covariance
(?) x (6) x Covariance
(?) x (T) x Covariance
(T) x CD x Covariance
(§) x (?) x Covariance
(D x (§) x Covariance
(5) x (£) x Covariance
(3) x (§) x Covariance
(5) x (6) x Covariance
(f) x (?) x Covariance
(4) x (T) x Covariance
(4) x (2) x Covariance
(4) x (§) x Covariance
(4) x (J) x Covariance
(4) x © x Covariance
(7) x © x Covariance
(4^ x (z) x Covariance
a.
= b.
= c.
= d.
= e.
= f.
= £:
h.
= i.
= k.
= 1.
= m.
- n.
= o.
= £1.
- r.
- s.
- t.
•= u.
= V.
w.
X.
- z.
aa.
= bb.
2-66
METRIC
-------
WORKSHEET 5M (OPTIONAL)
CONFIDENCE INTERVALS FOR PREDICTED LAND USE
£) x (2) x Covariance
Ji) x CD x Covariance
&) x (T) x Covariance
[6) x (§) x Covariance
*&) x (§) x Covariance
[D x CD x Covariance
5) x (T) x Covariance
5) x (D x Covariance
5) x © x Covariance
5) x (4) x Covariance
2) x (?) x Covariance
5) x © x Covariance
f?) x (7) x Covariance
Sum (a) through @) =
Standard Deviation of Dependent
Variable ="/ @
t - statistic of predictive equation
If Final Projected Land Use is Resi-
dential or Commercial, set equal
to the disaggregation percentage
± Confidence Interval
- $6) x $4) x © T 100
ee.
COMPUTING VARIANCE OF DEPENDENT VARIABLE
(?) x CD x Covariance = cc.
(?) x (D x Covariance - dd.
(?) x (§) x Covariance
(5) x (4) x Covariance
(?) x (?) x Covariance
© x (6) x Covariance
© x (?) x Covariance
® x CD x Covariance
ff.
gg.
= hh.
= 11.
- JJ.
= kk.
= 11.
mm.
= nn.
= oo.
PP.
go.
- rr.
= ss.
= tt.
= uu.
vv.
- ww.
43.
44.
46.
48.
1.69
47. 100
2-67
METRIC
-------
WORKSHEET 6M
MOTOR VEHICLE TRIPS
Effective Radius (in km) = 0.000564 * yj (Area of Analysis, in meters2)
Kilometers
(1)
Total Land Use
Single Family
Detached
Single Family
Attached
Multiple
Family
(2)
Work Trip
Rates
(3)
Work Trips
0)*(2)
(4)
Other Trip
Rates
'(5)
Other Trips
(D*(4)
Residential
Work Trips
Residential
Other Trips
Large
Commercial
Medium
Commercial
Small
Commercial
Off ice -
Professional
Manufacturing
Wholesale
Education
Other
Recreation
•
Total Work
Trips
Total Other
Trips
*See Page 2-26 for a definition of Effective Radius.
2-68
METRIC
-------
WORKSHEET 7M
VEHICLE KILOMETERS TRAVELED (VKT)
Work Trip
Length
VCT!H
VKTflM
VKTIO ;
VKT.Q
t>o Peak Hour Proporf
i
CTl
-------
WORKSHEET 8M
CATEGORIZED VKT
AREA OF ANALYSIS
VKTAPF | 1 = Peak
VKTAPA'l 1= Peak
1= Peak
VKTAOF 1 1= Off Peak
VKW 1 1 = of f Peak
VKTA1t. | 1= Off Peak
ro
^ IMPACT AREA
VKTTPF | |= Peak
VKTIPA'I l= Peak
VKTTPI | |= Peak
It'L ' • •
VKT-mV | | = Off Peak
VKTirfA* 1 1 = Off Peak
m
70
0 VKTT(J, 1 1 = Off Peak
1UL ' ' '
* (
* (
* (
* (
* (
* (
* (
* (
* (
* (
* (
* (
* Expressways
Work
* Arterials
Work
* Local
Work
* Expressways
Work
* Arterials
Work
* Local
Work
* Expressways
Work
* Arterials
Work
* Local
Work
* Expressways
Work
* Arterials
Work
* Local
Work
+ * Expressways ]
Other
+ * Arterials !
Other
+ * Local 1
Other
+ * Expressways !
Other
+ * Arterials \
Other
+ * Local ]
Other
+ * Expressways )
Other
+ * Arterials )
Other
+ * Local )
Other
+ * Expressways }
Other
+ * Arterials )
Other
+ * Local )
Work
*Subscripts for VMT variables are defined as follows: I = impact area, A = area of analysis, F = peak hour
0 = off-peak hour, E = expressways, A = arterials, L = local streets.
-------
WORKSHEET 9M
' MOTOR VEHICLE EMISSION FACTORS
Vehicle Type
Pollutant = _
Speed =
(Automobile Gas, Light Duty Truck Gas, Heavy Duty Gas, or Heavy Duty Diesel)
Region = (Low alt., High alt., or Calif.)
°F
(CO, N0x, or HC)
miles per hour
Year of Impact Assessment (t+10) =
Ambient Temperature =
Cold Starts =
Hot Starts =
(1) .
Vehicle
Age
(Years)
1
2
3
4
5
6
7
8
9
10
n
12
£13
(2)
Model
Year
(3)
Base Emission
Rate
(4)*
Hydrocarbon Crankcase/
Evaporate Emission Rate
(H^
(5)
Total Emission
Rate
(3)+(4>
|
1
i
T
•
(6)
Fraction of
Travel
(7)
Speed/Temp. /Cold-Hot
Starts Correction Factor
^ripstwx'
(8)
Model Year
Total Emissions
(5)*(6)*(7)
ro
•yo
i—i
o
Note H-J = 0 for CO and NOX; for these pollutants,
enter the values for C. directly in column (5).
Average Emission Factor
-------
WORKSHEET 10M
COMPOSITE MOTOR VEHICLE EMISSION FACTORS
Speed
(miles per hour)
Pollutant
(CO, NO. HC, SOV or Particulates)
A A
Vehicle Class
(1)
Average Emission
Factor
(2)
Vehicle Class
Proportion
(3)
Product
0)*(2)
Automobile/Gas
Light Duty Truck/Gas
Heavy Duty/Gas
Heavy Duty/Diesel
0.118
Composite Emission Factor
2-72
METRIC
-------
TOTAL MOTOR VEHICLE EMISSIONS
Pollutant
Condition
Peak, Expressways
Peak, Arterial s
Peak, Local Streets
Off Peak, Expressways
Off Peak, Arterial s
Off Peak, Local Streets
(1)
Average
Speed
(2)
VKT
Data
VKTApE
VKTApft.
VKTApL
VKTAO'E
VKTArfA'
VKTAOl
Area of Analysis, Total Motor Vehicle Emissions
Peak, Expressways
Peak, Arterials
Peak, Local Streets
Off Peak, Expressways
Off Peak, Arterials
Off Peak, Local Streets
-
VKT,pE
VKT,pA.
VKT,pL
VKTIO'E
VKTIOV
VKTIOL
(3)
Emission
Factors
Sum
(4)
Total Emissions
(2)*(3)
|
.
* . 0.365
(in ki
Sum
*
Impact Area, Total Motor Vehicle Emissions
(in k-
lo grams /year)
.
0.365
ilograms/year)
2--73
METRIC
-------
WORKSHEET 12M
STATIONARY SOURCE EMISSIONS
Fuel Type
(Gas, 011, or Electricity)
Pollutant
(CO, HC, NOX, SOX, Partlculates, or Kwh)
(1)
Total Land Use
SF Detached
SF Attached
Mult. Family
Large Commercial
Med. Commercial
Small Commercial
Office-Professional
Wholesale
Education
Other
(2)
Fuel
Proportion
(3)
Process
Emission
Factor
(4)
Process
Emissions
0)*(2)*(3)
(5)
Space Heating
Emission
Factor
(6)
Space Heating
Emissions
d)*(2)*(5)
(7)
Space Cooling
Emission
Factor
(8)
Space Cooling
Emissions
0)*(2)*(7)
TOTAL EMISSIONS
Process
Space 'Heating
Space Cooling
Industrial
= Total Manufacturing Land Use
* Fuel Proportion
* Industrial Emission Factor
O
-------
WORKSHEET 1 3M
EMISSIONS SUMMARY
Total Emissions
(1) Process
(2) Space Heating
(3) Space Cooling
(4) Industrial
(5) Total Stationary
Source Emissions
(1H2H3H4)
(6) Area of Analysis,
Total Motor
Vehicle Emissions
(7) Total Emissions
Area of Analysis
(5)+(6)
(8) Electric Utility
Emission Factors
(9) Total Kilowatt-
Hours
(10) Electric Utility
Emissions
(8)*(9)
(11) Impact Area,
Total Motor
Vehicle Emissions
[12)Total Emissions,
Impact Area
(5H10H11)
CO
HC
NOX
S°x
Particulates
Kilowatt-Hours
V
-*".
ro
en
-------
6. Example on Using Worksheets
This section presents an example of the use of the computation
worksheets, in English units. One example of each worksheet is filled out,
even though in an actual application, several copies of some of the work-
sheets would be required. The data used in this example were collected
from case study #1, Willimantic, CT.
2-76
-------
ro
WORKSHEET 1
INPUT DATA RECORD IN ENGLISH UNITS
Variable Name
Value/Computation
Inputs To '..'o"''sheets Units
Area of Analysis
Vacant Developable
Vacant Undevelopable
Vacant La^:'
Median Price
Median Income
Land Cost
Collection Capacity
Peak Flow
% Collection Reserve
Manufacturing Workers
Tract Area
Manufacturing Density
No mobility
Drivers
County Area
Criver Density
Sewered Land
Sewer Service
School Kids
Dwelling Units
Kid Density
Limited Access
Interchanges
Current Eroployrest
Future Employment
SM3A Area
Employment Growth
Track
Railroads
Zoned Office
6456
Acres
Acres
/V.rcs
'ji D
3
Developable.^ (Area cf Analysis - Vacant Undevelopable)
/rOQ
= Median Price v Median Income
9.0
= 100* (Collection Capacity - Peak Flow) * Peak Flow
9.0
76
4.5*
Manufacturing Workers •=• Tract Area
O.S7Z
Drivers * County Area
3 ce^
Sewered Land T Area of Analysis
37
= Schools Kids 4 Dwelling Units
o
= 640* Limited-Access 4 Area of Analysis
5/4
- (Future F-.-loyiv-rt - Current Employment) r SMSA Area
= 640* Track T Area of Analysis
0.248
Million C.i "lens Per Cay
Million Gallons Per Oey
Employees
Square Miles
Employees Per Square "ile
*
100s of Drivers
Square Miles
100s of Drivers Per Square Mile
Acres
*
Children
100s of Dwelling Units
Children Per ICO :v:cll ir; Jrits
Interchanges
Interchanges Per Square '^le
Employees
Employees
Square Miles
Employees Der Square Mile
"'les
Railroad '-'iles Per Sq-jare "ile Land
Area
-------
ro
^J
co
WORKSHEET 1 (CONTINUED)
INPUT DATA RECORD IN ENGLISH UNITS
Variable Name
Value/Computation
Office Zoning
House Density
Zoned Industrial
Industrial Zorirj
Population Growth
Office Vacancy
Airport Distance
Office Workers
Office Employees
Future Population
Employee Ratio
Unemployment
Future Income
Government
Collection Reserve
Interceptors
Interceptor Density
Poverty
Onsite Restrictions
County Growth
Project Cost
Federal Funds
Index One
Index Two
Population Served
= Zoned Office v Area of Analysi*
= 100* Dwelling Units 4- Tract Area
770
= Zoned Industrial T Area of Analysis
= Office Workers v Tract Area
Future Employment 4- Future Population
0.075-
10067
10.9
= Collection Capacity - Peak Flow
1.2.
= 640* Interceptors v Area of Analysis
690
Inputs to Worksheets Units
Dwelling Units Per Square f-'i's
Acres
Miles
Employees
Employees Per Square '-'He
People
."•ill ions of 3
Million Gallons Per Day
Miles
Miles of Interceptor Pipe Per Square
Kile Land Area
Ife 973
Thousand S
Thousand $
People
Unit!ess
-------
WORKSHEET 1 (CONTIiiUED)
INPUT DATA RECORD IN ENGLISH UNITS
ro
i
Variable Name
Value/Computation
Inputs to Worksheets Units
= 1,000* ((Project Cost •=• Index One) - (Federal Funds -•- Index Two)) * Population Served = f Stf • 374- j $ per Person
o.ozs
= Zoned Residential - Area of Analysis
Sewer Costs
Treatment Capacity
Vacant Houses
County Interchanges ^» f
Interchange Density = Interchanges T (640* County Interchanges v County Area)
Zoned Residential
Residential Zoning
Current Income
Income Growth
Current Medicals
Future Medicals
Hospital Growth
Current Houses
Future Houses
Housing Growth
Restriction Years
Phasing
Transit Stops
CBD Distance
= Future Income - Current Income
fe
10
= (Future Medicals - Current Medicals) r Current Medicals
72
a-76
= (Future Houses - Current Houses) T Current Houses
nil! ion Gal Ions
o.tt?
Interchanges
*
Acres
S
$
Employees
Employees
Dwelling Units
Dwell incj Units
*
Years
Miles
Unit! ess
-------
WORKSHEET 2
LAND USE PROJECTIONS
Residential
(-8692 * Q*7S>L
• (18804 * C
acant
Land )
Commercial = (4302 *
Sewer
Service)
+ (5347 *
+ (487.2 *
Driver
Density)
+ (16.42 *
+ 13466 =
% Collection
_Reserve ) + (11.24 *
Dwelling Units Per
10,000 Acres*
Bx,/i Employment
rjQ Growth )
--.^Vacant
- (4515 * 0* 7/5 Land )
n %Collection
- (2.626 * /O reserve ) + 5376
'••nu'Ecturing
- (28.84 * C^7 Density) + (6249 * D Interchanges)
4 * ^tf^M jw« ~^*.«
/ T/13 Per 10,000
Office-
Professional
= (719.5 * 0*245Rail roads) + (33478 -
., Industrial
- (1791 * OJIT3 Zoning ) + (572.0 *
Office
_Zoning)
Population
House
x% nouse
+ (36.70 * 7*0 Peak Flow) + (0.3324 * IQZi- Density)
Growth ) -459.9 = _/_ O
—/—-^-^- 1,000 Square Feet
~ Per 10,000 Acres
oo
o
Manufacturing = (2591
Wholesale
+ (234.7 *
Railroads)
Office
Vacancy)
- (3890 *
+ (22.7 *
Qf?7j Land )
. Airport
/ Distance)
._ 5/55Driver
(166.5 * W»Jl3Density)
n i Office
+ (86.28 * ^» *- Vacancy)
+ (5949
* d)<.Q7v5
Unemp1oynient) _ (0.1065 *
Future
Income)
/• oo o Manufacturing ^,^ Office
+ (2.978 * be*?lgt Density )+ (85923 * CM?^Z Zoning)
+ 611.2 =
_^^» ^ - I ) WVJ W ^>^U'_lt IUC
93/ Per 10,000 Acres
Employees) + (6043 *
Per 10,000 Acres
Employee
Ratio )
Highways = (35.22 * 0*24% Railroads)
A crtf-jLand
+ (22.98 * OOTlCost)
- (0.1167
+ (25.94 *
Interceptor
Density )
+ 11.35 =
Miles Per
10,000 Acres
f <* Collection
Government) + (0.5379 * fe>*o Reserve )
01 in
*l£p Poverty)
n 01 2 C°Unty
- 405.2 * U» MT> Growth )
A £T7/^ewer
(723.0 * ^^ // Service)
- (1.415 *
Sewer
Costs)
+ (437.5 *
+ 428.9 =
+ (95.17 *
1,000 Square Feet
Per 10,000 Acres
On-S-'te
Restrictions)
Recreation = (270.4
Railroads)
f\ A?iTVacan'l:
(3226 * 0*065 Houses)
<2 *> Treatment .. -7C. County
+ (2.631 * *'** Capacity ) +(322.1 * Ut^J)Growth)
[Industrial
_ Office
+ (0.9742 * O'Z~ Employees)
t\ ;;/y-2lndustrial . —, ^ Acres Per
- (557.8 * ^"^^Zoning ) + 59.52 = I O7»JLI 10,000 Acres
Other
= (1134 * 0>Z4& Rail roads) - (0.3020 *10C?67 Incoml) - (12175
0 Interchange
v ._-,„.-, _ Density )
_ ^..i Residential i
+ (403.3 * >?lf?rTZcrin9 ) + 3087 = L
+ (649.0 *
1 ,000 Square Feet
Per 10,000 Acres
County
-------
I
00
'.•.'O'K'-.KECT 3
FINAL LAND I'SF PROJEC
i n:'
I Lani Use Projections i
| From V.'orksheet 2 i
! I
(2)
Disaggregation
Percintanes
of Analysis :
3y 10.000
Final Land Use Projection
0) * (2) * (3)
! Residential
% Single Family ,39$ « 6^5*6 j Total SinSle Far":i^ C-t-:chc'J =
Tf?1
resicer-f'al
Co-rercial
Commercial
Conercial
74*4
1473
14-73
( 473
% ;';.;! ti Family
% Large Commercial
',' Kedium Commercial
% Snail Medium
j
602 i -6454 Totel -Uipic Family = ^907 C -.-I-.; J«;ts '
! i
O&7 ' ^4^^ ! "r°':a^ Large Co~i~:3rcia'i = 9^9 1 -CGO Squ:*''- Fset '
All j £4££ ( Total r'edium Commercial = /A j^ 1 .COO Sc^are Feet '
• V 1 1 1 ** -f ^"W «^
1 1
rtA^ ' fiA^tm. ' Total Snail Co^aercial = / „ O^ 1 ,C".~ Squc^e Feet 1
• W«B : • ^™^^P . • • TV '
Office-Professional |Q QO
n
iota" Of"ice-Professi~r.a
-------
'/.'OP.KS-iEET 4 (Optional)
DEFAULT C!SAf?TREGATIO\ CQUATIONS
:i Single
Family
Y + 0.00417 * Government '
Transit -.^
+ 0.00459 * Stops O
/ - 1.48 * Vacant Land ^' __'•))
Commercial j Q.Q03» = (1-- Large Commercial Ot°vl - « ;-edium
Cornnercial
Note: EXD is the exponential function. Thus EXP(X) = e .
-------
WORKSHEET 5 (OPTIONAL)
CONFIDENCE INTERVALS FOR PREDICTED SAND USE
Final Projected Land Use Category
PREDICTOR VARIABLE
Name
VA£d/vd UjMrd
Ld/nuL tost
%£oMwM,ty.w
MdM- D&wtu
Nfliaw)bittU|
Drwt D^fy
Constant Term
= VACANT
= UND
PV^r RE^APl
«= MAJOJOb
^STAV
- bUWE
1.
2.
3.
4. j
5.
6.
7.
COMPUTING VARIANCE OF DEPENDENT VARIABLE
Covariance
Predictor Term from
Variables Appendix A Multiplier
0 x © x .Ilk
0 x © x -Ml
0 x © x .&VL
0 x © x .103
0 x © x --^
0 x © x -414
0 x © x -.?#£
(D x © x . \0<\
© x G) x -102.
© x © x -.46!
© x (5) x -. i&q
(2) x © x ~ • *3lr'
El x
*- ~*^ X
El x
E4 x
(•• / y
E^? x
> E! x
&~7 x
E4 x
E2 x
EG x
2. &5 x
1.0
2.0
2.0
2.0
2.0
2.0
2.0
1.0
2.0
2.0
2.0
2.0
0- fH4-
10.0
win
O.'jll
Q.^o&
1.0
Resultant
, a. 4,^30,415
— * Q ^^ T 1 i ^^ X^ ^
^J X ""^ ^3
^ U • / / ^"^ "^
= e. - 2,243,096
= fl 1*5^404
— y . / /
= h. 554X5^|
= i. 04,823
= j. - 34,5iq
= k. - 116,451
k, . a4 -311
2-83
-------
Covariance
Predictor Term from
Variables Appendix A Multiplier
\D x (Z) x "-423 £k x 2.0
© x © x . 141 £1 x i.o
® x © x -. U4 El x 2.0
© x © x -^46? ^Ar x 2.0
© x © x ""• 113 E3 x 2.0
© x © x — •5'2j3 E4- x 2.0
© x © x .U£> &1 x 1.0
© x © x -.13^ £^5 x 2.0
s~*k. ^*+. "^ K^ *^7 f™" /**)
© x © x •o2-/ t-Z- x 2.0
© x © x .2.60 E4 x 2.0
© x © x -4-31 E8 x i.o
s\ r\f} *-/*
© x © x -. "£ Eb x 2.0
© x © x ~ • \c^b E^ x 2.0
© x © x .4^4 £^ x i.o
© x © x -.1)3 E1^ x 2.0
© x © x - \O^> C-C3 x 1.0 =
Sum (a) through (a^ :
y
Standard Deviation of Dependent Variable =v(8) :
Disaggregation proportion:
(This is 1.0 for all land uses except
residential and commercial. The pro-
portion is the decimal equivalent of
% Single Family, etc. The numbers
come from Col. 2 of Worksheet 3.)
Confidence Interval ( + ) = (9) x (fcD x 1 .69 :
Resultant
1.- W*,5-24
m. 71,0^0
n. - 103, &5^
n 45£>. €>56
w • f
P • '
q.- "74^,6^0
r. 4(e>6&, J6&
s.-q,g)^^ifo
u • /
U • / *
v • •
w • •
x. -10,075,00?
y. 7, 166
56,041
d d • / /
1,C
-------
WORKSHEET 6
MOTOR VEHICLE TRIPS
Effective Radius*= (0.0004973* Area of Analysis
= /*79 Miles
1/2
(1)
Total Land Use
Single Family fl
Detached «"Zi
Single Family _
Attached °
Multiple 29 9
Family x.Twy
(2)
Work Trip
Rates
Ii8
l.sr
l.O
(3)
Work Trips
(1)*(2)
2461
0
2*09
(4)
Other Trip
Rates
9.0
7.0
5.0
(5)
Other Trips
(D*(4)
/7307
o
14 543*
Residential
Work Trips
5370
Residential
Other Trips
Large ^ -^
Commercial ' 3f
Medium .^ ^
Commercial '*'••&
Small . ^
Commercial f*/Q
Office- ^.y
Professional «*//
Manufacturing /£34-
Wholesale ^8f
Education 75"^
Other //^
Recreation 3£.?
0
O
-------
WORKSHEET 7
VEHICLE MILES TRAVELED (VMT)
Work Trip / ~
Length » • *
VMT?,., I70771&
VMT. 4/7ZO
VMTin 43/ 7?£
VMTnn 1 //£ f#/
Peak Hour Proport
/ Other Trip
* Miles Length
= (Total Work
Trips
= (Total Work
Trips
25455
= (Total Other
Trips 756* 1
= (Total Other
Trips ^ZSfc'T/
ion 0
«/0 (
Facility Proportions
ro
co Expressways
Arterial s
Local Streets
Trip Sum A 1 93 «
Work
O./tT
£.£/
O.i3
Other
0-10
O.ZI
O-tto
5-61 = Manufacturing Work
Trip Sum B \IOf QQ6\ = Total
HD Correction | O,
Work Trips 4
•O42.1 = (Sum A T Sum B)
Automobile/Gas Proportion 0»*
Heavy Duty/Gas Proportion 0.
Heavy Duty/Diesel
Proportion j
/, -». Work ^ ^^ (Default Other A ^^ (Default
V.Vf Miles Proportion CJ -*IU =0.40) Proportion t*«~V =0.40)
^ Work Trip x1 O/ ) - (Residential ^ --*- * Work ^ * * Effective . «^ )
Length fc»ob Work Trips 5 Of" Proportion O & Radius *''Y
^ Effective • ^ /7 ) - (Residential -.^ — _ * Work ^ * Effective * — ^ )
Radius /» / j Work Trips ^ O/v Proportion 0 »^r Radius '**i
+ Other Trip - ) - (Residential 7/0£« * Other ^ ^ Effective y — ^ )
Length £»O/ Other Trips -J't^i Proportion G *^T Radius 9 * rj
i, Effective . ^^ ) - (Residential (Default = 15) / W (Default = 18)
ol 7 0 + Manufacturing Other O + Wholesale Work //S^ + Wholesale Other O
2S4S& + Total Other Trips 75"£4/
- 0.05, if less than 0, set equal to 0.
762* = 0.804 - HD Correction
DfcOl = 0.046 + 0.8 * HD Correction
o.O4d =
0.032 + 0.2 * HD Correction
''Subscripts for VMT variables are defined as follows: I = impact area, A = area of analysis, W = work trips, 0 = other trips.
-------
WORKSHEET 8
CATEGORIZED VMT
AREA OF ANALYSIS +VrTTfl,/ iVMTAn4'
HH _ r\-J
VXTW I75Z = Peak O.I * (4l fZO * Expressways C/.'»
APE Work
. * f». tf 1
VMTBnl, | 324 / |= Peak O. 1 * (4C 7/C * Arterials V- CrJ
APA Work
V>'7nn, | 1794-1= Peak CP.| * (4( 7/6 * Local U« 1,3
Ml L Work
VMT.Xr 1 757*51= Off Peak O .7 * ( 41 7ZO* Expressways V.f^
AUt Work
VMT,,,,-,, I 29 /$5[= Off Peak v»7 * (4*1 /IP * Arterials D,ZJ
AOA Work r
VMT,«', 123" /*A - Off Peak C/.T *(4l ItO * Local £/»Ld
""u V/ork
INS IMPACT AREA ^VMTj^
CO
VMTjpr 1 45 WO 1= Peak C»\ *( /TO 7/(i* Expressways Q.l^
...,,„,, .^^^-,, .v.-., •/ M/70 7/6* Arterials O-t-f
IPAL^ ' Work
Of 1*0**. 10 )
Other
+ f|Z59/* Arterials 0.2f )
Other
+ IIZS1I * Local 0,2O )
Other
+ I IC^V / * Expressways 0«IO )
+ I l£5"?f * Arterials 0 • 2| )
Other
+ WZ5"f|* Local 0*20 )
Other
+ nil rj» * Expressways 0«IO )
+ 4317% * Arterials 0 -2f )
Other
+ 431 71K * Local 0»ZO )
+ T*«7ffc * Expressways 0»|0)
Other
+ 431 7/6** Arterials 0 . t( )
Other
+ 43/7%.Loca1 0.10)
Work
Subscripts for VMT variables are defined as follows: I = impact area, A = area of analysis P = peak hour
0 = off-peak hour, E - expressways, A = artcrials, L = local streets. ' '
-------
Vehicle Type =
Pollutant =
Speed =
WORKSHEET 9
MOTOR VEHICLE EMISSION FACTORS
(Automobile Gas, Light Duty Truck Gas, Heavy Duty Gas, or Heavy Duty Diesel)
(CO, NO , or HC) Region = LOvu fl>U (Low alt., High alt., or Calif.)
A ~~
miles per'hour Ambient Temoerature = 15 °F
Year of Impact Assessment (t+10) = I9°O
CO
30
Ambient Temperature =
Cold Starts =
Hot Starts =
(1)
Vehicle
Age
(Years)
1
2
3
4
5
6
7
8
9
10
11
12
>13
(2)
Model
Year
I
0.107
o.l(£
O.IOZ
0.09i
o, off
0.077
0.044
0.o4?
0.0-S3
0.0^3
D.cM-
(7)
Speed/Temp. /Cold-Hot
Starts Correction Factor
(ripstwx^
0*
0.63
0.6*3
^63
0.^3
o.GS
0.63
0.63
0.63
0.6?
0.63
0.63
0.63
(8)
Model Year
Total Emissions
(5)*(6)*(7)
0*143
o.z/s~
O.ZZ-?
0.2+0
0.25V
0.254-
0.2-f?
C.213
0.202
0.144
£.//£
^.or/
0. Z26
ro
co
CO
Note Hi = 0 for CO and NOX; for these pollutants,
enter the values for C, directly in column (5).
Average Emission Factor
-------
WORKSHEET 10
COMPOSITE MOTOR VEHICLE EMISSION FACTORS
Speed
Pollutant CO
(miles per hour)
(CO, NO. HC, S0v or Particulates)
X X
Vehicle Class
Automobile/Gas
Light Duty Truck/Gas
Heavy Duty/Gas
Heavy Duty/ Diesel
(1)
Average Emission
Factor
Z.fco
9.*0
M7
28.7
(2}
Vehicle Class
Proportion
0. 762-
0.118
0.0*0
0.040
(3)
Product
|.
-------
TOTAL MOTOR VEHICLE EMISSIONS
Pollutant CO
Condition
Peak, Expressways
Peak, Arterials
Peak, Local Streets
Off Peak, Expressways
Off Peak, Arterials
Off Peak, Local Streets
(1)
Average
Speed
37
2o
(6~
4f
30
)«
(2)
VHT
Data
VMTAPE
/ant
VMTAPA
32*1
VMTApL
H794
VMTAOE
/S7(&
VMTAO'A
tlltf
WTAdL
ZSI4&
Area of Analysis, Total Motor Vehicle Emissions
Peak, Expressways
Peak, Arterials
Peak, Local Streets
Off Peak, Expressways
Off Peak, Arterials
Off Peak, Local Streets
37
2o
/£-
45-
30
IB
VMT £
* ?ro
VMTJpA
/2<5*-
VHTIpu
IOVS6
mi6i
619/1
VMTIOA
111 886
VMTIO'L
f77 37/
0.805
869 Wi
)unds/year)
5^934-
2o4-f9S"
23o 147
4tl IZZ.
/ 5%0£S&
/74B 9oe
4 ml 338
0.805
^ 3ft> /a7
?unds/year)
2-90
ENGLISH
-------
WORKSHEET 12
STATIONARY SOURCE EMISSIONS
Fuel Type
Oil
(Gas, Oil, or Electricity)
Pollutant CO
(CO, HC, NO. SO. Particulates, or Kwh)
X A
(1)
Total Land Use
SF Detached |fl23
SF Attached Q
Mult. Family 2^0^
73?
Large Commercial
/0.5"
Med. Commercial
/ -90
Small Commercial
671
Office-Professional
Wholesale £»f
Education YOv
Other /76
(2)
Fuel
Proportion
0.6
0.6
0-6
0*4
0.6
€>-£»
0-t
0.6
0-6
O-&
(3)
Process
Emission
Factor
L£
1^
1.0
O.Of
o.o?
0.07
0
0.0^
0
0
(4)
Process
Emissions
(D*(2)*(3)
ise^
o
1745"
5^1
1
0.1
0
U
0
0
(5)
Space Heating
Emission
Factor
6.Z7
6.Z7
3-Z5"
/.6S"
/.65^
A 65"
/-65~
/.65"
i.LO
Af3
(6)
Space Heating
Emissions
(1)*(2)*(5)
7234-
C
5T673
^30
10
Z
664
£«
5Vo
151
(7)
Space Cooling
Emission
Factor
O
0
O.I/
o
o
o
O.IZ*
0
0.0*
0. c4
(8)
Space Cooling
Emissions
(D*(2)*(7)
O
o
/fa
o
0
o
4r
$
K
+
ro
10
TOTAL EMISSIONS
Process J I ~f 9 Space Heating ^
Industrial JT/OlO Total Manufacturing Land Use
* Fuel Proportion
Space Cooling
* Industrial Emission Factor
-------
WORKSHEET 13
EMISSIONS SUMMARY
Total Emissions
CO
HC
NO.
SO,
Participates
Kilowatt-Hours
(1) Process
3198
(2) Space Heating
/ SA10
(3) Space Cooling
ro
ro
m
•2.
CD
[—
1—4
CO
(4) Industrial
rofe
(5) Total Stationary
Source Emissions
33
(6) Area of Analysis,
Total Motor
Vehicle Emissions
86? 69?
(7) Total Emissions
Area of Analysis
903 3SS"
(8) Electric Utility
Emission Factors
10
(9) Total Kilowatt-
Hours
(10)Electric Utility
Emissions
(8)*(9)
(11)Impact Area,
Total Motor
Vehicle Emissions
(12)Total Emissions,
Impact Area
=F
4*70 707
^ i .
/v
-------
III. REFERENCES
1. Office of Air Quality Planning and Standards, Guidelines for Air
Quality Maintenance Planning and Analysis; Volume 2: Plan Prepa-
ration, EPA Publication No. EPA-450/4-74-Q027Research Triangle
Park, NC, 1974.
2. Office of Air Quality Planning and Standards, Guidelines for Air
Quality Maintenance Planning and Analysis; Volume 3: ControT
Strategies, EPA Publication No. EPA-450/4-74-003, Research
Triangle Park, NC, 1974.
3. Office of Air Quality Planning and Standards, Guidelines for Air
Quality Maintenance Planning and Analysis; Volume 4: Land Use
and Transportation Considerations. EPA Publication No. EPA-450/4-74-004,
Research Triangle Park, NC, 1974.
4. Office of Air Quality Planning and Standards, Guidelines for Air
Quality Maintenance Planning and Analysis; Volume 6: Overview of
Air Quality Maintenance Area Analysis, EPA Publication No.
EPA-450/4-74-007, Research Triangle Park, NC, 1974.
5. Office of Air Quality Planning and Standards, Guidelines for Air
Quality Maintenance Planning and Analysis; Volume 9: Evaluating
Indirect Sources, EPA Publication No. EPA-450/4-75-001, Research
Triangle Park, NC, 1975.
6. Office of Air Quality Planning and Standards, Guidelines for Air
Quality Maintenance Planning and Analysis; Volume 12: Applying
Atmospheric Simulation Models to Air Quality Maintenance Areas,
EPA Publication No.EPA-450/4-74-013,Research Triangle Park,
NC, 1974.
7. National Environmental Policy Act of 1969, 42 U.S.C., Section 4321
at seq.
8. Council on Environmental Quality, "Guidelines for Preparation of
Environmental Impact Statements," Federal Register, 38/147), Part II,
August 1, 1973.
9. Benesh, F., Guldberg, P., and D'Agostino, R., Growth Effects of Major
Land Use Projects: Volume I - Specification and Causal Analysis of
Model. EPA Publication No. EPA-450/3-76-012a, Research Triangle
Park, NC, May 1976.
10. Benesh, F., Growth Effects of Major Land Use Projects: Volume II -
Compilation of Land Use Based Emission Factors, EPA Publication No.
EPA-450/3-76-0126, Research Triangle Park, NC, September 1976.
3-1
-------
11. Benesh, F., Guldberg, P., and D'Agostino, R., Growth Effects of ^
Land Use Projects: Volume III - Summary, EPA Publication No.
EPA-450/3-76-012C, Research Triangle Park, NC, September 1976.
12. Guldberg, P., D'Agostino, R., and Cunningham, R., Growth Effects of
Major Land Use Projects (Wastewater Facilities); Volume i; Model
Specification and Causal Analysis, EPA Publication No. EPA-450/3-78-Ol4a,
Research Triangle Park, NC, March 1978.
13. Nie, N.H., Hull, C.H., Jenkins, J.G., Steinbrenner, K., and Bent, D.H.,
Statistical Package for the Social Sciences (2nd Edition), Mc(5raw-
Hill, New York, NY, 1975.
14. Finney, D.J., Statistical Methods in Biological Assay (2nd Edition),
Hafner Publishing Co., New York, NY, 1964.
15. Personal communication, Dr. Edward E. Cureton, University of Tennessee,
Knoxville, TN, 1977. The substance of the weight validity index
technique is contained in an article of Dr. Cureton's which has been
accepted for publication in a professional journal, but has not yet
been released.
16. Guldberg, P., Benesh, F. and McCurdy, T., "Secondary Impacts of
Major Land Use Projects," Journal of the American Institute of Planners,
43(3): 268-269, July 1977^
17. Heise, D., Causal Analysis, John Wiley & Sons, New York, NY, 1976.
18. U.S. Water Resources Council, 1972 Obers Projections, Volumes I-VII,
Washington, DC, April 1974.
19. Office of Transportation and Land Use Policy, Mobile Source Emission
Factors, Environmental Protection Agency, Washington, DC, January 1978.
20. Cerighton, "Estimating Efficient Spacing for Arterials and Expressways",
Highway Research Board Bulletin No. 253, Washington, DC, 1960.
21. National Research Council, Highway Capacity Manual, 1965, Highway
Research Board Special Report Number 87, Washington, DC, 1965.
22. Office of Air and Waste Management, Compilation of Air Pollutant
Emission Factors, Third Edition, Environmental Protection Agency
Publication No. AP-42, Research Triangle Park, NC, August 1977.
23. Bureau of the Census, 1970 Census of Housing, Washington, DC.
24. Bureau of the Census, 1972 Census of Manufacturers. Washington, DC.
25. Office of Management and Budget, Standard Industrial Classification
Code Manual, Washington, DC, 1972.
3-2
-------
26. National Coal Association, Steam Electric Plant Factors, Washington,
DC, 1973.
27. Edison Electric Institute, Statistical Yearbook of the Electric
Utility Industry, New York, NY, 1972.
28. Bureau of Economic Analysis, Population, Personal Income, and Earnings
By State, Projections to 2000, U.S. Department of Commerce, Washington,
DC, October 1977.
3-3
-------
APPENDIX A
COMPLETE STATISTICAL OUTPUT
OF THE PREDICTIVE EQUATIONS
Covariance data for the following Final Projected Land Use categories
can be found on the indicated page:
Final Projected Land Use Category Page
Single Family Detached A-2
Single Family Attached A-2
Multiple Family A-2
Large Commercial A-4
Medium Commercial A-4
Small Commercial A-4
Office-Professional A-6
Manufacturing A-8
Wholesale A-10
Education A-14
Other A-18
Recreation A-16
Highways A-12
A-l
-------
1. RES Equation
EQUATION 1
***********
SMPL VECTOR
1 40
ORDINARY LEAST SQUARES
VARIABLES...
RES
VACANT
LAND
RECAP1
MANJOB
STAY
DRIVE
C
MEAN OF DEPENDENT VARIABLE IS 8713.5469
£
INDEPENDENT
VARIABLE
VACANT
LAND
RECAPl
MANJOB
STAY
DRIVE
C
R-SQUARED =
ESTIMATED
COEFFICIENT
-8691.57422
5347.30078
16.4241486
11.2375050
-18803.8398
487.237305
13466.2617
0.7416
STANDARD
ERROR
2675.66455
1041.f2104
3.F32589U
3.43206120
6609.87891
215.450577
3235.0POP1
T-
STATISTIC
-3.24P37875
5.12264751
4. 28539181
3.27427197
-2.844807f 2
2.26148033
4.16257286
PEAN OF
VARIABLE
0.6f<694695
0.53672218
181.299986
504. 2f 3763
0.44994849
1.72454357
1.00000000
DURBIN-UATSON STATISTIC (ADJ. FOP 0 GAPS) = 2.0923
NU^SER OF OBSERVATIONS = 40
SUM OF SQUARED RESIDUALS = .356145E+09
STANDARD ERROR OF THE REGRESSION = 3285.16
ESTIMATF OF VARIANCE-COVARI A\CE MATRIX OF FfTIPATFC COEFFICIENTS
1/ACAUT-0.7T6E + 07 -0.262E + 06 0.892E+02 0.203E + 04 -0.253E+07 0.219E + 06 -C.388E + C7
i_/»/wo-0.262E*06 0.109E + 07 0.102E + 04 -0.467E*02 -0.189E+06 -0.522E + 05 -0.423E + 06
«.«92E+02 0.102E+04 0.147E*02 -0.119E*01 0.b4PF+04 -0.113E+03 -0.529E+04
>0<» -0.*»67E + 02 -0.119E + 01 0.118E»02 -0.139r.»05 0-327F + 02 0
>-07 — O»189E*O6 O*b**6E*OA —O«13^E»O5 O»^»37E*Ofl -—0-222L*O& — 0 • :
-------
LINF.
PRINCETO\ UNIVERSITY
* TSP
CF AlubST,
PAGE.
PLOT OF ACTUALC)
FITTED<+) VALUF.S
PLOT OF RESICUALS(O)
10
ACTUAL
FITTED
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
?2
23
24
25
26
27
28
29
30
31
32
33
35
3ft
37
38
39
4C
8209.
2782.
8223.
5911.
6920.
6257.
.2891E+05
9706.
.1197E*05
5868.
.2357E*05
9987.
.1926E+05
7094.
6072.
3745.
9709.
.1517E*05
5100.
7540.
6771.
5017.
7372.
.1894E+05
5903.
582.0
6515.
4626.
6550.
.1498F. + 05
7772.
2112.
651f.
.1050T+05
369.0
.1322E+05
6313.
5375.
Tti39.
7484.
-12.56 +
.1046E+05
5511.
.1249E+05
7808.
.2351E+05
6300.
.1512E+05
. 1 155E+05
.1731E»05
9811.
.1978E*05
4616.
6941.
7860.
9520.
8073.
8030.
6669.
55*3.
2994.
4533.
.1769E+05
1451. *
14?7. *+
59CO.
5867.
8455.
. 1CC>6E + 05
463t.
?402.
6753.
.IS^SE+C^
37/0.
.1145E+05
7227.
6790.
bO 1 P.
RESIDUAL
7?5.
.279E+04
-.224E+04
400.
-.557E+04 0
-.155E*04
.540E+04
.341E+04
-.315E+04
-.568E»04 0
.625E*04
176.
-525.
.248E*04
-869.
-.412E+04
1B9.
.7C9E*04
871.
'.119E*0«
.202E*04
-675.
£15.
-.191E+04
-579.
-.329E»04
-237.
-.3tOE+04
.177E+0«
-.141E+01
0.0
.0
0.
0 .
0 .
0
c.
0 .
0 .
-------
2. CQMM Equation
ECUATION ?
SWPL VECTOR
1 40
ORDINARY LEAST SQUARES
VARIABLES...
COMM
SERVED
VACANT
KIDS
ACCESS
JOBCHG
RECAP1
C
MEAN OF DEPENDENT VARIABLE IS 1848.
INDEPENDENT
VARIABLE
SERVED
VACANT
KIDS
ACCESS
JOBCHG
RECAP1
C
ESTIMATED
COEFFICIENT
4301.81250
-4514.91016
-28.P41201B
6248.92187
-17.C374908
-2.62619495
5J75. 52344
STCNOARD
ERROR
1002.06226
1133.73901
11.1874228
2225.56396
7.56761165
1.75866601
1415.94165
T-
STATISTIC
4.29287338
-3.98231792
-2.57800198
2.80779171
-2.25136948
-1.49328709
3.79642963
KEAN OF
VARIABLE
0.60547256
0.60694695
99.0000000
0.09044975
36.6881866
181.299988
1.00000000
R-SGUARED = 0.5734
DURBIN-UATSON STATISTIC (ADJ. FOR C GAPS) = 2.1626
NUMBER OF OBSERVATIONS = 40
SUP OF SQUARED RESIDUALS = .749141E+C8
STANDARD ERROR OF THE REGRESSION = 1506.69
ESTIMATE OF VAPIANCE -COVAR IANCE MATRIX OF ESTIMATED COEFFICIENTS
-0.225E+06 -0.1P3F+04 -0.^68E+03 -0.207E+06
0.1P9E+P7 C.177r+04 -P.8f2E*0= n.l2U + 0« 0.145E + 0? -C.102E + 07
KiDS-0.211E*04 0.177E+04 0.i:5E+03 -0.731E+04 -0.332E+01 0.804E+00 -0.115E+05
*<•<•' ,-0.225E*06 -C.862C + 05 -0.731F + P4 0.495E+JT7 -t .435 T + 04 ^Oj.326E+03 0.683E+06
r?6>Ht-0.183E*04 C.121E+04 -0.3T?r + 01 -Q.435E + 04 0.573E + 02 C.177E + 01 -0.133E + 04
RF.t-'i -0 ,56fiE+07> 0.145E + C^ O.^flE + OO -0.326E + 071 0.177E + 01 0.309L + 01 -0.420E + 03
-------
LINE
PRINCETCf. UNIVERSITY
TSF
VERSlCf. OF AUGUST,
PLOT OF ACTUAL(*> AfuC FITTEO<+> VALUES
PLOT OF RESIDUALS(O)
in
ACTUAL
FITTED
RESIDUAL
Ul
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
*4
35
36
37
3 ft
39
1419.
419.4
879.1
892.4
1474.
2778.
2959.
1794.
1206.
388.8
3862.
2470.
2971.
2154.
1118.
1104.
2027.
8732.
701.9
2776.
1462.
1897.
1640.
2771.
77".0
145.4
729. «
722.8
174.?
22't.t.
1197.
171.6
F07.6
1 n71 .
.11T7L+05
245.7
1510.
1875.
609. ?
1472.
-1478.
1162.
890.9
1901.
2632.
2488.
2340.
775.5
744.9
1936.
1693.
3948.
1559.
651.9
2505.
1035.
5986.
3724.
3446.
2373.
83.45
1464.
1812.
1888.
95.07
-823.1
1904.
754.0
3674.
C772.
-2C7.9
2012.
1971.
f- f C 7 .
-28f .2
2443.
3317.
^rc. 9
18*0.
-53.3
.190E+04
-283.
1.51
-428.
146.
471.
-546.
430.
-356.
.193E+04
777.
-978.
596.
466.
-.140E+04
992.
,275E*04
-.302E+04 0
-670.
-911.
.181E+04
176.
958.
-.111E+04
50.3
.155E+04
-.118E+04
-580.
-.144E+04
-.157E+04
380 .
-.120E+04
.466E+04
532.
-533.
-.144E+04
-192.
-.137F. + 04
0.0
0
.0
0
0
. 0.
.0
. . 0 .
. 0.
. 0 .
. 0.
.0
. 0 .
.0 .
. 0 .
. 0 .
0 . .
. 0.
0
• • .
. 0 .
. 0 .
.0
. .0 .
. 0 .
.0 . .
.0
0
.0 .
. 0.
0
0
. 0 .
.0 .
. . .
. 0 .
. c .
0
0
0
-------
3. OFFICE Equation
LGUATICN
SMPL VECTOR
1 40
ORC/INtRY LEAST SQUARES
VARIABLES...
OFFICE
RRHILF.
OZONED
PEAK
DUACPE
IZONED
POPDIF
C
MEAN OF DEPENDENT VARIABLE IS 503.1760
INDEPENDENT
VARIABLE
RRHILE
OZPNED
PEAK
OUACRE
IZONED
POPDIF
C
ESTIMATED
COEFFICIENT
719.511719
33478.0352
36.6961060
0.33335096
-1790.61548
572.025391
-459.871338
STflNDARD
ERROR
137.541382
6491.65156
10.2877779
0.10663390
772.712891
271.222656
147.449326
T-
STATISTIC
5.23123741
5.15693092
3.56696129
3.12612534
-2.31730938
2.10906124
-3.11884308
MEAN OF
VARIABLE
0.46907347
0.00529748
5.88499546
692.524902
0.06514716
0.20634961
1.00000000
R-SQUARED = 0.7017
DUPBIN-UATSON STATISTIC (ADJ. FOR 0 GAPS) = 2.1360
NUMBER OF OBSERVATIONS = 40
SUM OF SOUAPtD RESIDUALS = .293272E+07
STANDARD ERROR OF THE REGRESSION = 29fl.lll
ESTIMATE OF VARIANCE-COVARIANCE MATRIX OF ESTIKATEC COEFFICIENTS
RRMTIE0.169E + 05 0.196E»06 0.148E + 03 -0.361E + 01 -0.40U + 05 0.182E + 04 -0.603E+0«
-^v£,>0.19tE + 06 0.'»21E + Ofl -0.903E + 04 -0.31CE + 02 -C.184F + 07 0.267E + 06 -0.175E + 06
Q.14PE+03 -O.S03E+04 0.106t*03 0.682E-01 -0.131E+04 0.170E+03 -0.fcl2E+03
-0.316E*02 0.f>82E-01 0.114E-P1 C.15CF+02 C.525E»01 -0.928E + 01
.^10 IE
O . 1 7 O f. » 0 3
Ci r .597'. »0fc -0.1S7r»OS -n.
O . T ^ E -r C 1 - O . 1 •=• 7 F » n e. ^O.7?eF.O5 -O.a3BE»Of>
-------
LINE
PRINCFTOM UNIVERSITY
TSP
OF AUGUST, 15(9
PAGE
12
PLOT OF ACTUAL<*> AND FITTEDC+) VALUES
PLOT OF RESIDUALS(O)
ID
ACTUAL
FITTED
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
3f.
37
38
39
40
552.7
411.8
1687.
86.40
337.3
339.7
1169.
288.5
555.1
147.9
687.5
589.3
390.4
382.4
935.9
403.4
822.3
1999.
995.9
230.9
199.0
177.8
34.00
567.3
94.70
65.10
251.9
144.6
40.60
P.87.5
577.2
33.90
T85.5
?96.f,
2111.
94.60
"H9.7
380.1
98.60
185.2
1046.
-23.02
1474.
168.3
298.5
404.3
855.9
759.1
939.3
261.7
553.9
609. f
711.8
451.1
673.5
578.3
234.2
1626.
923.7
479.9
64.30
18.59
215.1
325.3
428.6
50.92
345.6
441.4
-208.6
896.8
b45.7
115.5
41^.8
431.2
1446.
13?. 6
581.4
-19.17
12?. 3
447.5
* *
RESIDUAL
-493.
435.
212.
-ei.9
38.fl
-64.6
713.
-471.
-384.
-114.
134.
-20.3
-321.
-68.7
262.
-17b.
588.
373.
72.2
-249.
135.
159.
-181.
242.
-334.
14.2
-53.7
-297.
249.
-9.34
-269.
-81.6
-34.3
-135.
666.
-38.0
-91.7
399.
-?3.7
-262.
0.0
0.
.0
0.
.0
0.
.0
0.
0.
0.
0.
0.
-------
4. MANF Equation
EQUATION 4
SKPL VICTOR
1 4C
ORDINARY LEAST SQUARES
VARIABLES...
RRMILE
VACAM
P.ANJOB
OZONED
VACOFF
AIRPPT
C
MEAN OF DEPENDENT VARIABLE IS 1B05.8608
INDEPENDENT
VARIABLE
RRHILE
VACANT
MANJCB
OZONEP
VACOFF
AIPPRT
C
ESTIMATED
COEFFICIENT
2590.97559
-3890.12744
2.97811413
65922.3750
234.675339
22.7014465
611 .240967
STANDARD
ERROR
697.218994
1287. (5332
1.40823187
33632. S289
93.959^58
12.R664851
1068.74121
T-
STATISTIC
3.71615696
-3.02109814
2.11463928
2.55473232
2.49761772
1.76164722
O.E7192606
MEAN OF
VARIABLE
0.469D7347
0.60694695
204.263763
0.00529748
3.16749668
23.5024567
i.oooooooa
R-SQUARED = P.6301
DURBIN-UATSON STATISTIC (ADJ. FOR 0 GAPS) = 2.4249
NUMBER OF OBSERVATIONS = 4P
SUM OF SQUAPEC RESIDUALS = .8e0977E»08
STANDARD ERROR OF THE REGRESSION = 1605.Pt
ESTIMATE OF VAPIANCE-CCVARIA\CE MATRIX OF LSTIPATEC COEFFICIENTS
£p.HTLE \Jf\(.f\KJT AlAA/lOfi 0*<3A/£p VAtafp ftrkPRT
r-0.486E*06 0.1f.9E + 06 -0.112E + 03 0.213E + 07 -0.938E + 04 -0.267E + 03 -0.283E»06
It^r+Ob 0.16br+07 0.4?3£+03 -P.171E+07 0.660E+04 -0.'(75E*04 -0.107E+07
112E + 03 0.«33E' + 03 0.19PE + 01 -O.llH + Of 0.1[59F + 02 -0.332E + 01 -0.529E + 03
C.213E1+07 -G.171E+07 -0.111E+05 O.llJL+10 0.743E+06 0.498E+05 -0.721L+07
V-V-f •- -0 .
0.1C,BE»03 -D-
-------
LINE
'F AUGUST*
15
PLOT CF 4CTUAL(*> AI-.C FJTTt;0(+) VALUES
PLCT OF f
ID
ACTUAL
FITTED
to
1
2
7
1
5
6
7
8
9
10
11
12
13
11
15
16
n
18
19
20
21
22
23
21
25
26
?1
28
29
70
31
32
73
7<4
7.5
3^
77
Zf
39
10
2339.
191 .1
1209.
17«5.
1831.
20Q2.
6213.
1050.
511.2
267.7
736.8
672.7
1208.
715.^
1071.
3265.
553.1
.1011E+05
1126.
1859.
388.1
1271.
2365.
3106.
359.2
fO.fO
12».8
1011.
83.f.0
621. «•
1151.
121.8
1711.
517.7
.10f 6E+05
i7^.b
"512.
970.2
110.8
17.10
25M.
-1520.
1916.
1877.
1251.
2193.
5900.
2057.
1677.
976.5
525.2
82C.9
1392.
2 1 ? 7 .
96^.0
10f 2.
-326.2
671 1.
1 8 r> 0 .
2799.
531.1
&36.1
-28.29
2770.
871.7
-1271.
1126.
706.7
-53.61
?733.
330?.
-163.1
1791.
3219.
f 705.
-571.9
7f'4l.
1 1 f> 1 .
171.8
1 7 ?. » .
01
REJ1CUAL
-192.
.201ET
-707.
-P8.2
580.
-101.
313.
-.101E+01
-.116E+01
-709.
-118.
-183.
-.168E+01
106.
-S17.
£fc2.
.310E»01
-.312E+01
.206E+01
-116 .
125.
.239E+01
f.35.
-E15.
.173E+01
-S97.
338.
137.
-.211E+01
-.215E+01
288.
-180.
-.273E+01
.395E+01
907.
??P.
-221.
-71.0
-.128E+01
.0
0.0
0
0
.0
0.
.p
-------
5. WHOLE Equation
EQUATION ?.
***********
VECTOR
1 40
ORDINARY LEAST SQUARES
VARIABLES...
WHOLE
OKI VE
VACOFF
CFFJOP
EMPOP
UNEMP
PERCAP
C
MEAN OF DEPENDENT VARIABLE IS 477.0664
£
INDEPENDENT
VARIABLE
DRIVE
VACOFF
OFFJOB
EMPOP
ur:EMP
PERCAP
C
ESTIMATED
COEFFICIENT
166.454239
86.2756805
3.34034920
604?. 78906
6948.57422
-0.10651082
-1911.15161
STANDARD
ERROR
20.7087250
19.°433136
0.9781018?
1535.17920
2431.63574
0. "4552276
668. f 82324
T-
STATISTIC
8.03787899
4.32604504
3.41513348
3.93621063
2.85757160
-2.33972645
-2.85723114
MEAN OF
VARIABLE
1. 72454357
3.16749668
75.1841431
0.37607706
0.04694974
9726.22266
1.00000000
R-SGUARED = 0.7597
DURBIN-UATSON STATISTIC (ADJ. FOR 0 GAPS) = 2.2514
NUMBER OF OPSERVATIONS = 40
SUM OF SQUARED RESIDUALS = .33509f,E+07
STANDARD ERROR OF THE REGRESSION = Mfl.6£0
ESTIMATE OF VARIANCE-COVARIANCE MATRIX OF ESTIMATED COEFFICIENTS
Jflve >jhCoFF OFFjofi 2WPOP v/ufwP
H 0.429E + C3 0.?,29E. + 02 0.6R5E + 00 -0.709E + 02 0.73SE + 04
L C.685E + 00 0.4?4T*01 0.9b7E+00 0.22f>e+0.T -C.3?4f + 03
'-0.709E+02 -0.414T+04 0.226F+03 0.?3fc,L+07 0.111L+07
-0.209E + 00 C.709E + 03
0.3S4E-01 0.434E + 03
0.118E-01 -0.270E + 03
-C.24CE+02 -0.711E+06
» - i o ** r -*-o
.03 — o . zr
- .:> ,1 -n_viiK»n»-
0.1S2E-r2 -r.780E»06
-------
PPINCETO' UNIVERSITY
isr
VEPSKN CF AUGUST,
PfiGE
1R
PLOT OF aCTUAL(') AND FTTTEDC+) VALIES
PLOT OF RESIDUALS(O)
ID
ACTUAL
FITTED
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
??.
24
25
26
27
28
29
30
31
32
33
34
35
36
37
36
39
40
114.2
21.20
189.3
74.30
558.1
68.30
2535.
182.3
937.6
133.3
312.5
667.8
205.0
2364.
333.7
432.9
.0
551.5
425.0
1871.
346.8
1399.
300.6
760.6
646.0
14.50
218.3
361.4
20.50
228.4
570.4
27.40
671.0
1?6 .5
256.5
9.300
732.1
175.0
10.90
If 1.1
405.8
-215.0
233.9
36R.5
413.9
-109.7
2322.
28.37
1171.
545.6
-96.10
504.3
643.8
1993.
59.36
587.0
13P.5
797.1
669.4
1095.
402.6
77P.1
568.6
966.8
303.9
85.42
772.6
597.9
119.0
44^.5
14^.6
-5.30C
907.3
705.3
6 4 3 . fe
264.0
531.?
194.4
-340.2
237.5
RESIDUAL
236.
-44. 6
-294.
144.
198.
213.
154.
-233.
-412.
409.
164.
-439.
371 .
274.
-154.
-139.
-^46.
-244.
776.
-55. F.
620.
-26U.
-206.
342.
-70.9
-154.
-236.
-Sf.5
-221.
425.
33.3
-236.
-1*9.
-307.
-255.
201.
-19.4
351.
-56.4
0.0
C.
0 .
0
. 0
0.
0 .
0 .
.0
0 .
0.
•
0.
0 .
0.
0.
0 .
0.
-------
6. HIWAYS Equation
EQUATION
SMPL VECTOR
1 40
ORDINARY LEAST SQUARES
VARIABLES...
HlUAYS
RRMILE
LAND
GOVT
RECAP
CROSS
C
MEAN OF DEPENDENT VARIABLE IS
43. 3924
INDEPENDENT
VARIABLE
RRKILE
LAND
GOVT
RECAP
CROSS
C
ESTIMATED
COEFFICIENT
35.2186584
22.9822540
-0.11664993
0.53785098
25.9384308
11.3462658
STANDARD
ERROR
9.71726036
7.10037518
O.P4293521
0.23814219
16.6402893
8.29916477
T-
STATISTIC
3.6243*006
3.23676586
-2.71688175
2.25852776
1.55877209
1.36715698
MEAN OF
VARIABLE
0.469073*7
0.53672218
71.3199310
tO»2699714
0.23038459
1.00000000
R-SQUARED = 0.5226
DURBIN-UATSON STATISTIC (ADJ. FOR 0 GAPS) = 2.1316
NUMBER OF OBSERVATIONS = 40
SUM OF SQUARED RESIDUALS = 18421.3
STANDARD ERROR OF THE REGRESSION = 23.2767
ESTIMATE OF VARIANCE-COVARIANCE MATRIX OF E«Tlr"ATED COEFFICIENTS
\-f.HZLE. L/WD Govr RtcfiP CROSS (ionsr*irr)
0.944E*02 -0.163E+01 -0.966E-02 -0.552E-01 -0.205F*02 -0.374E*02
.Lft'JO -0.163E+01 0.504E+0? -0.633E-01 0.34flE-01 -0.889E+01 -0.201E + 02
iovr-0.966E-02 -0.633F-01 0.184E-02 0.1"1E-01 O.fc^r-02 -0.946f-01
Ht^P -0.552E-01 0.348E-01 0.141E-04 0.567E-01 0.593E + 00 -0.714E + 00
CFos$-0.2D?E*02 -0.869E+01 0.655E-02 0.593E»00 0.277E+03 -0.560E+02
-0.201E»02 -0.946E-01 -0.71tE«00 -0.560E + 02 0.689E»02
-------
LINT
PRINCCTON UNIVERSITY
TbP
VERSION OF AUGUST, 1969
>AGE 21
PLOT Of ACTUAL(*> AND
VALUES
PLCT OF RESlLlALS(C)
ID
ACTUAL
FITTED
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
lc
15
?0
21
22
23
24
25
?S
27
2P.
29
30
31
32
33
T A
-i c;
36
37
38
39
40
70.90
30.10
26.80
15.40
26.70
92.20
66.60
"7.00
84.00
55.00
57.50
f. 2 . 1 0
75.60
P3.50
C6.60
42.30
25.30
170.0
3.000
15.20
31.40
45. fO
13.40
80.20
30.30
4.1PC
4P.°0
30.80
32. TO
58.50
4A.30
5 . 3, 1 0
34.50
P * . B 0
48.20
14.00
36. 5C
27.70
1 5 . '. 0
P .4fO
38.94
25.83
55.48
-6.09?
33.33
62.04
68.37
41. P5
£2.55
36.26
32.46
61.43
hj.f'?.
49.46
53.88
54.73
37.48
111.9
42.40
46.81
53.30
2H.40
12.03
60.59
48.ft4
16.45
43.65
38.57
2.frM
43. P5
"3. 07
3?. 49
52.55
54.74
4P.59
2^ .59
37.54
42. P 4
-4.V24
37.61
RLSIOUAL
32.0
4.27
-28.7
21.5
-6.3T
30.2
-1.57
5.15
1 .01
18.7
25.0
-<>.33
-12.2
34.0
2.72
-12.4
-12.2
O.C
0 . • •
0
. 0 . .
.0
0.
. .0 .
0
. 0.
. . .0
. 0 . .
. 0 • .
0
Z1.6
21.9
17.2
1.37
19. t
18.5
12.3
= .21
7.77
30.1
14.7
5.23
?7.2
•18. C
25.9
.395
11.6
1.04
•15.1
20.8
•25.2
0 .
0
•
•
•
.0
. 0
•
0
•
•
•
0.
.0
0 .
•
. C
•
. 0
•
0.
•
•
•
0
•
•
•
.0
•
•
•
.0
•
•
•
0
•
0.
•
•
•
•
•
0.
•
n
•
•
•
•
. 0
0 .
•
•
•
•
•
•
•
•
0
•
-------
7. EDUC Equation
EQUATION 7
S"FL VECTO"
1 4P
ORDINARY LEAST SQUARES
VARIABLES...
EOUC
POOP
SERVED
LAND
LIMITS
PERCH6
COST
C
MEAN OF DEPENDENT VARIABLE IS 662.7135
INDEPENDENT
VARIABLE
POOR
SERVED
LAND
LI-ITS
PE"CHG
COST
C
ESTIMATED
COEFFICIENT
-3185.58545
722.979492
437. 52£ 123
95.1653137
-405.2014U,
-1.41497135
428.885010
STANDARD
EFRCR
670.197266
144.C78339
83.0681152
30.2513428
180.390137
0.77642828
168.597427
T-
STATISTIC
-4.75320530
5.01796055
5.26707649
3.14562062
-2.24625015
-1.82241058
2.54384041
KEAr, OF
VARIABLE
0.15904957
0.60547256
0.53672218
2.47499943
0.25504965
45.4195862
1.00000000
^-SQUARED = 0.7545
DURBIN-yATSON STATISTIC (ADJ. FOP 0 GAPS) - 2.0024
NUMBER OF OBSERVATIONS = 40
SUM OF SQUARED RESIDUALS = .192734E+07
STAMDAPD ERROR OF THE REGRESSION = 241.669
ESTIMATE OF V AP I Ah,CE-COVAR IANCE MATRIX OF ESTIMATED COEFFICIENTS
POOfiv SERVEQ LflWO LTMTTS PERCHt COST
0.445E + C6 0.163L + 0? Q.cr;56E + 04 -0.34°i: + 04 0.153E+Ob -0.178E + 03 -C.709E + 05
S£Rveo 0.153E + 05 0.208r + Cc; -C.933E*03 -C.108E + 03 -0.371E + 04 -0.717E + 01 -0.130E + 05
0.55f-E + 04 -0.9?3F+Ci 0.69C.E+C4 0.645E + 03 0.1?ff + 04 -C.254E + 02 -C.482E + 04
*94 -C.lCct+0: O.S45E + 03 0.915r*0.' -0.f-18E + 03 C.227E + 01 -0.194E*04
-------
LIME
PRINCETON UNIVERSITY
TSF
VERSICK OF AUGUST» Iv/t9
PAGE
24
PLOT CF ACTUALi*) AND FITTEC(+> VALUES
PLOT CF RESICUALS(O)
ACTUAL
FITTED
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
3C
31
32
33
34
3":
36
37
3P
39
4C
866.0
563.4
2135.
454.0
524.9
709.5
1554.
635.8
724.0
603.9
984.8
f67.fl
483.3
447.4
350.8
499.?
982.5
1517.
368.6
1279.
286.9
707.3
f55.0
1374.
359.2
69.70
346. C
715.5
147.4
11C6.
662.5
f: 1 . 1 0
£•43. 7
1037.
4 ri 7 . 6
95.50
615.5
607.9
11. 6P
160.1
751.0
443.0
1805.
529.0
816.1
965.0
1194.
787.6
725.6
425.9
485.5
760.6
f 15.8
214.9
576.0
505.3
llr.5.
1283.
424.3
1075.
4ie.5
411.5
407.2
1311.
580.5
-168.1
434.0
643.0
230.3
7BP.4
700.5
329.3
72?. 7
9f I .4
582.1
-50.90
7^4.7
11"6.
14C.1
3 c ' . 7
RESIDUAL
115.
120.
229.
-75.0
-291 .
-255.
759.
-152.
-1.65
178.
499.
-92.8
-132.
233.
-2?5.
-5.75
-172.
233.
-55.7
204.
-129.
29f. .
b3.0
-221.
238.
-f-b.O
-127.
-82.9
325.
-.'fl.O
-248.
-80.0
cO. 1
-125.
I4f. .
-149.
-C14
0.0
. 0 .
. 0 .
0 .
0 .
0.
.0
. 0 .
.0 .
. 0 .
• *
0.
).
. 0 .
• •
• 0 .
• •
.0
• •
. 0
.0
0 .
-------
8. REC Equation
EQUATION H
SMPL VECTOR
I 40
ORDINARY LEAST SQUARES
VARIABLES...
REC
RRWILE
CAPAC2
PERCHG
OFFJOB
VACHSE
IZONED
C
MEAN OF DEPENDENT VARIABLE IS 207.8000
INDEPENDENT
VARIABLE
RRWILE
CAPAC2
PERCH G
OFF JOB
VflCHSE
I ZONED
C
ESTIMATED
COEFFICIENT
270,401123
2.63108635
322.055176
0.97423780
-3225.57251
-557.780029
59.5172862
STANDARD
ERROR
56.?478333
1.06722641
91.3438721
0.35317051
1262.37156
294.C32471
71.8569489
T-
STATISTIC
4.8IJ7315B3
2.46534920
3,52574444
2.7SB54778
-2.55516815
-1.89700127
0.82B27461
I«E*N OF
VARIABLE
0,46907347
10.8374815
0*255114965
75.18*1431
0.03909981
4.06514716
1.00000000
R-SQUARED = 0.6471
DURBIN-VATSON STATISTIC (ADO. FOR 0 GAPS) = 1.7751
NUMBER OF OBSERVATIONS = 40
SUM OF SQUARED "ESIDUALS = 498049.
STANDARD ERROR OF THE REGRESSION = 122.851
ESTIMATE OF VARIAKCE-COVARIANCE MATRIX OF ESTIMATED COEFFICIENTS
RRNILE cfwca. PERCH & OFFJOB Vflcnse rio/veo
£ 0.316E+04 0.105E+01 0.628E+03 -0.475E+01 0.143E+05 -0.521E*04 -0.152E+04
0.105E + 01 O.lltF + 01 -0.390E*01 -0.970L-02 0.128E»03 -0.266E + 02 -0.144E + 02
i 0.628E + 03 -0.390E + 01 O.RT4E + 04 0.741E + 01 -O.^6?t + 04 -C.277E+04 -C.237E + 04
-0 .475E + C1 -0.970E-02 0.741E + C1 C.12r.E + CO -C.£t?E + 02 0.229E + 02 -0.791E + 01
-------
LUF
10
PRI\CtTCN UMVFRSITY
TSF-
1'jfc1?
PAGE 27
PLOT CF tCTUALC) £ND FTTTEH
VALUES
PLOT OF RESICUALS(O)
ID
ACTUAL
FITTED
RESIDUAL
1
?
3
-------
9. OTHER Equation
ECUATIOM
SMPL VECTOR
1 40
ORDINARY LEAST SGUAKES
VARIABLES...
CTHEK
RRMILE
PFRCAP
VACHSE
PERCHG
INTDtfJ
R Z 0', f 0
C
MEAN OF DEPENDENT VARIABLE. IS
151 .23136
3=.
I
oo
INDEPENDENT
VARIABLE
ESTIMATED
COF.KFICIf NT
1134.4^140
-0.30158252
-12175.2109
64A.Q79980
-2^5.404785
STANDARD
EPF.CR
180.363739
311 .26b602
PERCAF
VACHSE
PERChG
I \TOEf.'
RZONED
r. ?08f. 53931
R-SQUARED = 0.6721
DURBIi\-WATSON' STATISTIC (ATJ. FOR 0 GAPS) = 2.3041
NUMBER OF OBSLKVATIONS = 40
SUM OF SQUARE L RESIDUALS = .5244^1f.+07
STANDARD ERROR OF THE REGRESSION = 39p.(66
ESTIMATE OF V AR I ANCE-CO VAR I ANCE '•'ATRIX OF E^TIMATtC COt
T-
STiTISTIC
6.26568620
-4.77312660
-2.51952839
1.50167999
-1.4P177052
1.46485615
4.45826244
MEAN OF
VARIABLE
0.46907347
9726.22266
0.03909981
0.25504965
0.18093479
0.33849955
1.00000000
^ri-E 0.325E
-C.104E + 01 O.l^r'E + Ce 0.122r
0.400'"-02 0.114F+n3 -f.
C.158E. + 06 0.114li + i;; C.2itf. + 08 -C.
0.1??E*05 -C.lnlE*C2 -
C.H3F + 01
0.414t+ni
FFICIEf'TS
R 7 OWED f'ouim/"^
0.122E+05 -0.200E+05
-0.368E+01 -0.398E+0?
0.101L+06 -C.213E+07
O.filJE+03 C.789E+05
0.101E.-»-CE -C.t, 0 *T- + Q =
C_/. Pt •C1" — C . !:•*•» • ft »
-------
LIKE
11
PRINCETON UNIVERSITY
TSP
VERSION' OF AUGUST,
PAGE
30
PLOT OF ACTUAL(') AND FITTED(+) VALUES
PLOT CF RESIDUALS(O)
ID
ACTUAL
FITTED
to
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
85.70
50.00
353.2
486.1
83.70
548.6
578.2
308.8
603.7
403.4
76?. 6
1124.
304.4
477.8
866. 2
377.0
434.4
4044.
268.7
558.5
160.6
532.6
82.20
288.2
389.8
33.10
230.7
274.6
41.10
642.2
414.0
38.30
282.9
146.9
316.7
b8.30
1016.
309.1
45.40
146.3
279.8
-521.3
298.4
197.7
1.213
40.70
730.6
-26.88
811.8
869.2
64=. 2
853.9
453.3
1034.
627.2
720.1
625.0
2803.
761.4
889.6
370.6
257.4
-118.2
236.4
922.3
228.6
451.3
409.3
83.55
234.0
712.9
16.05
253.9
754.9
274.7
.8671
631.5
182.8
-321.9
474.^
* +
»*
RESIDUAL
-154.
571.
5«.8
288.
62.5
508.
-152.
336.
-208.
-486.
118.
270.
0.0
.0
•
.0
-557.
239.
-3t3.
-191.
.124E+04
-493.
-331.
-210.
?75.
200.
-532.
-195.
-221.
-I?1).
-42.4
108.
-299.
22.2
29.0
-608.
42.0
r.7.4
3H5.
126.
3(7.
-?.2fi.
.0
. 0
.0
C.
.0
-------
10. RY1 Equation
PREDICTIVE FQUATION REGRESSION
FILE PREDICT (CREATIO^ DATE = 02/17/78)
DEPENDENT VARIABLE.. PY1
VARIAELE(S) ENTERED ON STEP Nl'itER 6..
("L-/17/7P
U L T I
HOSCHG
MULTIPLE R
R SQUARE
ADJUSTED P SQUARE
STANDARD ERROR
0 .9052F.
0.81953
0.75581
0.28002
ANALYSIS
RCGRESSlO
RESIDUAL
VARIABLE
KIDS
CAPCHG
SERVED
OZONED
STAY
HOSCHG
(CONSTANT)
VARIABLES IK THE EQUATION
B BETA STL IRKCR b
0.13729120-01
-0.91820930-03
0 .7842406
22.70408
-2.980871
-0.1814667
4.531282
O.f.2662
-0.71293
0.36504
0.37C74
-0.44621
-0.27103
0.00246
0.00017
0.24114
7.17613
0.91678
o.oai-u
31.204
21.16u
10.57?
10.010
10.t72
".957
ro
o
l l '. l l r
b^ GL'<'
t .
17 .
VAKIABLt
'. U.'-C^E
V A C h S F
1NCOL
VACOFF
UM-'iF
OFFJOC
UUJOL
i1 A i1. o c b
oor s
i\ pf^nh
FOCR
VALUf
<*OC> £
UMV
KELJOB
RELUW
FELOFF
R E L C" A f-1
GCVT
LAf- ;•
POP JEN
DISC6D
R R M I L t"
Accrss
INTL.LN
TRANS
A I H F ' R T
VACANT
RZON'ED
CZONED
I70NFC
LIKITS
POLICY
TLI'-'IT
TPOLIC
CF r, TUAFtS
C . C 5 3 a 2
1 .33304
F.ETA I'
-P. 26276
-0.01227
-0.090?"
-0.11146
0.1167?
-0.0C793
0.0912P
0 .076?4
0 .15144
-0.02842
-0.03401
C . 0 i c (. Z
-P.0?777
0.11174
0.02776
C .1^166
0.13P"9
0 .1 J5 7>
r. . c r 5 ? P
0.06^93
0 .13291
-0.0234"
-0.05721
-0.2250-:
-n. 14391
0.12^7
0.0997"
0.00995
-O.OOOCb
-0.195^1
- 0 . C 9 c (• 0
0 .CPf 41
-O.lil60
C.C2211:
;• 1 1 r.
l .
c r r T r r 1 1-> i
. i 1 t* 1 i 1* 1 " •
c f. F T I f. L
0.09311-
-n «43oit
-O.r2i>12
-0.1F.171
-0.1^912
0 .210JJ3
-0.1^61?
0 .Ifc660
0 .17644
0 .^6022
-O.C4142
- 0 . ! 7 4 f, t
O.C614:.
-J.[HC33
C.2f>0i?0
0 . 0 4 d 6 5
0.2S305
0.25251
0 .244,04
n . p i c 6 1
0.12410
J. 21467
-0.04488
-0.1097fr
-0.17366
-0.30 OOP
0 .T5751
0 ..2201
0 . (.' 2 0 6 7
-O.OOC1C
-0.',-;752
- 1 . r r 4 1 r
0.0 19b3
-0 .cb743
r.cMOi
0"9,-
C . f «. i' 7 7
C . 4 '"' 4 5 0
C.VG4R3
0.73HQ6
0 .541- 15
0.5&404
0.8 C. 5 ? P
C .io'(l9
0 . 5 7 7 o 1
L .5.52!: 1
0 . 3 o 7, 3 7
0.86076
C . 7 0 c. 0 L'
0.1-075
0 .61096
0.55355
C. 60289
0.59996
C.f "631
C .7C2f 1
0 .56f"2f;
0. till 32
C.66 1"2
0 .f.£u<47
0 .67071
0 .76467
0'.70972
0.89115
0.77921
0.78691
0.6016 I
0 . 6 ; c 2 1
0.97200
0 . H b 8 3 0
0.9t776
L ! ' 1 1
L ! i T 1
p
1 ? . r t. L V
f
(.'.110
? . 6 <•;
0 . 0 I ?
0.5"^
0 . fa f. 1
0.7 5>
P . 2 r 5
O."r7
0.30J
1 . 1 h 2
0 . f ? 7
C.C^O
0 •C'' 7
C . C 3 3
1 .Ifec
0 . 0 3 P
1 . 3 ° I
1.09C
1.514
0 . 0 C 2
0.23C
l.Ol"
0.032
0 . 1 " 5
3.71C
1.TP3
1 . 1 7, 7
0.829
O.OC7
o.ccr
2.345
C .6?P
P .C P6
1.411
0.042
-------
11. RY2 Equation
PRCDICllVE EQUATION REGRESSION
FILE PREDICT (CREATION DATE = 02/17/7b)
****************** * * * I
DEPENDENT VARIABLE.. RY2
VARIABLE(S) ENTERED ON STEP MJCRER 5..
" U L T I F L b
PEAK
/1 7 / 7 -
L: ?T
L I', T
MULTIPLE R
R SQUARE
ADJUSTED R SQUARE
STANDARD ERRO-R
0.91390
0.83521
0.78943
0.33457
A'JALYSIS OF VARIA\CF
PEGRESSIOh
RESIDUAL
VARIABLE
HSECHG
PCOR
OZONED
RRMILE
PEAK
(CONSTANT)
B
4.376418
-5.479691
-34.05850
0.7322320
0.3645832D-01
0.2180154
VARIABLES IN THE EQUATION
BETA STD ERROR
0.70395
-0.48705
-0.43228
0.35431
0.26902
0.68429
1.25251
8.21178
0.20536
0.01399
40.903
19.140
17.202
12.714
6.795
OF SOI'
rj.
18.
VARIABLE
DUACPE
VACHSE
INCOME
VACOFF
UN'EMP
OFFJ06
UUJOf:
MA" JOB
JCLS
NONHH
VALUE
ROOMS
KIDS
UNIV
RELJOB
RELUU
RELOFF
RELHAN
GCVT
LAND
POPDEN
DISCED
ACCESS
INTPEN
TRANS
AIPFRT
VACANT
RZONED
CZONED
IZONED
LIMITS
POLICY
T L I * I T
TPCLIC
DRIVE
CF SQUARES
1 0 . _> 1 1 .; I
2.014P2
PETA I*
-0.12*56
-O.OM 43
-o.ieror
-0.01948
0.09182
-0.09540
-0.11675
- 0 . 1 1 B 1 8
-0.13060
0.0897
0.70678
T.87553
O.fc°123
0.59812
t.56073
0.52848
0.63635
0.81047
0.95514
0.94348
0.96716
0.90357
0.76183
0.82320
0.75030
0.65692
0.80816
0.h45:'5
0.83137
0.70522
0.6=524
0 .86b95
C .80256
D .80549
O.S'-Ull
0 . 8 1 1 " 5
0.72629
1.4F9
p. OPC
2.535
0 . 0 3 U
O.P64
1.031
1 .11C
1.530
0.50*
0.005
0.705
3.380
0.5] H
0.432
0.224
0.3&0
0.43C
1.009
O.lCfe
1.312
0.150
1.796
0.012
O.lPt
0.257
0.475
1.373
0.486
0.269
0.336
1.714
0.007
1 .014
-------
12. CY1 Equation
PREDICTIVE EQUATION REGRESSION
FILE PREDICT (CREATION DATE = r.2/1i/7t)
DEPENDENT VARIABLE.. CY1
VARIABLE(S) ENTERED ON STEP NUMBER 5.. TLIMT
U L T I P L
-.••••*. 1 '• I.L
(.,•• •Sill \
LIST 1
L U' T j
MULTIPLE R
R SQUARE
ADJUSTED R SQUARE
STANDARD ERROR
0.85281
0.72729
0.68718
0.57838
ANALYSIS OF VARIANCE
REGRESSION
RESIDUAL
VARIABLE
VACOFF
LIMITS
CAPAC2
KIDS
TL1HIT
(CONSTANT)
VARIABLES IN THE EQUATION
B BETA STD ERROR B
-0.215*105
0.5069719
-0.2320196D-01
0.15897710-01
-0.6026162D-01
-2.265219
-0.61369
0.67519
-0.11751
0.37027
-0.30139
0.03275
0.07750
0.00521
0.00415
0.02182
13.215
12.790
19.616
14.693
7.630
DF SLK
b.
31.
VARIABLE
OUACRE
VACHSE
INCOME
UNt (• f
OFF JOB
WUJOC
iwAf^joe
JOfS
NONHH
POOR
VALUE
ROOMS
UNIV
PEL006
RCLWW
RILOFF
RELMAN
OOVT
LANG
POPOEN
OISCBO
RRMILE
ACCESS
INTDEN
TRANS
AIRPf T
VACANT
RZONED
CZONEO
ozorf o
I70NED
POLICY
TPOLIC
OR IVE
R1CET-
CF SQUARE?
i 0 . 3 3 d : 0
11.373nb
\/ADTAkirc
BETA U
-0.11611
-0.02212
-0 .00681
-O.OlObP
-0.13188
0.0103?
-O.lf til
-o.ii7f •:
0.093b7
0.139t8
-0.09852
-0.00568
-0.09690
-0.06567
-0.09265
-0.Of.fcS1>
-0.08361
0.09110
-0.06013
-0.13875
-O.l9b71
0.020f 1
-0.11712
-0.06502
-0.01726
-C .nc"CCn
0.05175
-0.05557
-0.1170fc
O.llb13
0.0b372
c.oc;r27
O.Obc fjO
-0.10197
-fj.1'0138
r- E /i r S
(• .
r.
luf^T Tfv Tut
NCI I f\ I n t
P iKTI AL
-C ,21f-35
-0 .01lOf
-0 .01259
-C. 01763
-0.23770
O.C1973
-0.29380
-O.Jr:520
0.16931
0.^2271
-0 .16907
-0.009ie
-0.13tl1
-C. 12387
-0.17273
-C. 130.23
-0.15852
0.15883
-0.10776
-0.216H
-0.3151b
0.0372f.
-0.11*751
-0 .1099h
-0 .02'.'95
-T .CP7P2
0.092Cl>
-0.088f C
-0.2115L
0.25213
C.G8';? T
O.OR137
L'.1025f
-0 .lf.f-.2f
-r . co 197
,LA«!-
066ST
.'3153
TCLERA..CE.
0.7r?H2
0.91150
0. 9235:2
0.7723L
D.8if-9t.
0.9921P
O.HbOOV
C .-(.(> 5 4
0 . b 8 7 b 1
0.69328
O.HC 323
0.70715
0.55bf>2
0.9C13P
0.9l7Cb
0.97178
0.97=.57
0.77205
0.1.17-90
O.P5P25
0.819HL
0.891^8
0.77180
0.779Hf
0.f2C>-l
0.71tc2
C.7f-l93
0.6<3-'2b
0.85^'fcl
0 . 8 2 1 7 r
0.7^6.-".
o .7^1:?
0 .6r..bJ
fj .^:. -1 It
r . 5 ?. v « 2
^
1 f . l J "< r. 1
F
2.1'..^
O.Cbf
0.101:
D.01L
1.97E
0.013
3 . ] 1 »•'
2.2U
0.971
1 ,72i
0.971
O.OOi
0.61 =
O.bll
1.015
0.5C9
0.851
O.d51
0.3B;.
2 . 1 2 1
1.173
O.OIb
1.310
0.111
0.030
0 . ? 3 *
0.2 0 7
0.26,1
1.551
2.21^
o . : ••- <•
c . c ^ 7
0 . ? b ]
C. V.-'.C
o. c! . :•
-------
13. CY2 Equation
PREDICTIVE EQUATION REGRESSION
FILE PREDICT (CREATION DATE = 02/14/78)
DEPENDENT VARIABLE.. CY2
VARIABLE(S) ENTERED ON STEP NUMBER 5..
? /1 4 / 7 *
MULTIPLE
R C r- IV E S S I C I1
VACANT
V - f I 6 r L -. LIST
R E o * E 5 S t C !\ LIST
MULTIPLE R
R SQUARE
ADJUSTED R SQUARE
STANDARD ERROR
0.80663
0.65065
0.59928
0.80603
ANALYSIS OF VARIANCE
REGRESSION
RESIDUAL
VARIABLE
PHASE
TRANS
DISCBD
GOVT
VACANT
VARIABLES IN THE EQUATION
B BETA STD ERROR B
-1.538799
-0.4589325D-02
-0.1779279D-C1
-0.4169631D-02
1.481964
-0.6111''
-0.48208
-0.42150
-0.29076
0.25541
0.26515
0.00106
0.00461
0.00152
0.62051
33.60?
18.92C
14.872
7.491
5.727
(CONSTANT) -0.5360620
OF SU1-'
5.
34.
VARIABLE
DUACRE
VACHSE
INCOME
VACOFF
UNFMP
OFFJOB
UUJCB
MANJOB
JOBS
NONHH
POOR
VALUE
ROOMS
KIDS
UN IV
RELJOB
RELUU
RELCFF
RE L MAN
LAMP
POPDEN
RRWILE
ACCESS
INTDEN
AIRPRT
RZONED
CZCNED
OZONED
IZCNED
LIMITS
POLICY
TLI^'IT
TPOLIC
DRIVE
WIDET
CF SQUARES
41.11118
22.06952
BETA IN
P . 0 4 1 f • 0
-0.06327
-0.07297
-0.1C941
-0.17053
C. 14171
0.0614C
-0.12059
0.04676
0.13476
0.07962
0.02P92
-0.00608
-0.09082
0.1067?
-0.10005
0.03920
-0.144*6
-0.21026
0.03009
0.02348
0.11415
-0.04111
-0.05989
0.0632C
-0.02819
0.07065
-0.13881
-0.05335
0.20292
0.01421
o . o n 1 7 »
0.01234
-0. lifter
-o ,047e?
' FA1*, SCUARF.
t .2?r?t
C.64S65
NOT IN THE CliUATIO-
1 2 . 6 6 1' ,'
PARTIAL
0 .057?*-
-O.C9540
-0.11904
-0.17081
-0.26649
0.19066
0.09586
-0.19289
0.07013
0 .21902
0.11169
0.03291
-0.00907
-0.12926
0.12613
-0.12888
0.04531
-0 .19766
-0.30112
0.04512
0.03301
0.1869P
-0.06520
-0.08790
0.09367
-0.043U
0.11010
-0.21978
-0.07414
0.31177
C .02277
T.01 '77
0.01^82
-0.15150
-C .1)6650
TOLERANCE
I). 79421
0.92956
0.85151
0.85310
0.63238
0.85149
0.89IC6
0.72261
0.92272
0.68399
0.86453
0.77612
0.70770
0.48792
0.57968
0.46686
0.65319
0.71651
0.78574
0.69068
0. 9?f>66
0.87877
0.75259
C.76744
0.61662
0.84H37
0.87576
0.67450
0.82471
0.89691
0 •(•(•? 14
0.50118
0.6C742
0.71477
0.109
0.303
0.474
0.992
2.52?
1.245
0.306
1.275
0.16?
1.66?
0.417
0.036
0.003
0.561
0.533
0.557
0.068
1.342
?.291
0.067
0.036
0.141
0.2^7
0.29?
O.CC1
0.4C5
1.675
0.162
3.55?
O.C17
O.C13
0.77?
0 .1 •:•. (•
-------
APPENDIX B
GLOSSARY OF TERMS
B-l
-------
A dependent variable is the output variable of a mathematical
function, which has as its arguments one or more input or independent variable^.
Given a set of interrelated variables, an exogenous variable is one whose
value is determined by forces external to or outside the system of relationships
An endogenous variable is an internal variable of the system which is completely
determined (i.e., caused) by one or more exogenous and/or endogenous variables.
The relationships of any one endogenous variable can be translated into mathe-
matical form by designating it as the dependent variable of a function whose
independent variables are all related causal factors, whether they be endo-
genous or exogenous variables.
Instrumental variables are exogenous variables which are introduced into
nonrecursive systems of causal relationships (i.e., those involving inter-
actions between endogenous variables) to allow the estimation of structural
coefficients. A detailed list of conditions associated with instrumental vari-
ables is given in Heise [171.
A wastewater major project is defined as the construction or extension
of interceptor or collector sewer lines during the period 1958 to 1963 in a
community in the United States. If construction date information is not
available, then a grant funding date in the period 1956-1960 is acceptable.
The project had to affect an increase in absolute system collection capacity
of 1 MGD or more, and had to cost a minimum of $200,000 to construct. Phased
projects are considered if the first phase of construction on the collection
network was complete within 1958-1963 and the last phase of construction was
complete by 1965.
Induced development is defined as urban land use associated with, caused,
stimulated, or allowed by, or located because of the construction and operation
of a major project.
Secondary development is induced development which has a direct causal
relationship with a major project.
Tertiary development is induced development which has a direct causal
relationship with secondary development, and so is only indirectly related to
a major project.
The drainage basin of a wastewater project is defined as the land area
which drains, by gravity, to any point of the collection network of the major
project. In the case of essentially flat terrain (e.g., river deltas), this
area is restricted to the locus of points no greater than 1,000 feet from any
point along the interceptor line of the major project.
B-2
-------
The legal service area of a wastewater major project is defined as
the drainage basin of the major project plus any additional areas connected
to the collection network by pumping stations and force mains.
The area of analysis is defined as the legal service area of a wastewater
major project in the base year. It must be a minimum size of 5,000 acres (or
approximately 8 square miles), and contain significant amounts of vacant de-
velopable land, some of which must be more than 5,000 feet from the nearest
interceptor line in the base year.
"201" refers to Section 201 of the Federal Water Pollution Control Act
Amendments of 1972 (PL 92-500). Section 201 calls for detailed planning for
the wastewater treatment facilities needed to achieve the goals of the Act.
"208" refers to Section 208 of the Federal Water Pollution Control Act
Amendments of 1972 (PL 92-500). Section 208 provides for the designation of
state and areawide agencies for the purpose of developing effective water
quality management plans for areas that, because of "urban-industrial con-
centrations" or other factors, have "substantial water quality control problems."
The approach is aimed at integrating controls over municipal and industrial
wastewater, storm sewer runoff, nonpoint source pollutants and land use.
B-3
-------
APPENDIX C
DEFINITION OF MODEL VARIABLES
C-l
-------
This appendix summarizes the model variables used in the GEMLUP-II model,
A complete discussion of variable selection is given in Section II.6 of the
first volume report [12]. Table C-l lists the definitions of endogenous
model variables, while Table C-2 summarizes the definitions of exogenous
model variables, both in English units. The metric units of those variables
actually used in the predictive model and worksheets are listed in Table C-3.
C-2
-------
TABLE C-l
DEFINITION OF ENDOGENOUS MODEL VARIABLES
Name
Description
Land Use Codes [106]
RES
COMM
OFFICE
o
I
CO
MANF
WHOLE
HIWAYS
EDUC
Number of dwel 1 ing' units per 10,000 acres of area
of analysis in 1970.
Commercial land use per 10,000 acres of area of
analysis in 1970 in 1,000 square
feet.
Office-Professional services land
use per 10,000 acres of area of analysis
in 1970 in 1,000 square feet.
Manufacturing land use per 10,000 acres
of area of analysis in 1970 in '1,000 sq. ft.
Wholesale-warehousing land use per 10,000
acres of area of analysis in 1970 in
10,000 square feet.
Non-expressway highway lane miles per 10,000
acres of area of analysis in 1970.
Educational land use per 10,000 acres of area
of analysis in 1970 in 1,000 square feet.
11 Household units
12 Group quarters
13 Residential hotels
14 Mobile home parks or counts
19 Other residential
52-b9 Retail trade
62 Personal services
64 Repair services
66 Contract construction services
61 Finance ,insurance, and real
estate services
631-636 Business services (excludes
638,639 warehousing)
65 Professional services
692 Welfare and Charitable
services
699 Other services
2,3 Manufacturing
51 Wholesale trade
637 Warehousing and storage services
45 Highway and street right-of-way
681 Nursery, primary, and secondary
education
-------
TABLE C-l (CONTINUED)
DEFINITION OF ENDOGENOUS MODEL VARIABLES
Name
Description
Land Use Codes [106]
REC
OTHER
Active, outdoor recreational land use per 10,000
acres of area of analysis in 1970 in acres.
Other urban land uses per 10,000 acres of area of
analysis in 1970 in 1,000 square feet.
73 Amusements
74 Recreational activities
75 Resorts and group ca;nps
15 Transient lodgings
691 Religions activities
71 Cultural activities and
nature exhibitions
79 Other cultural, entertain-
ment, and recreational
o
-------
TACLE C-2
DEFINITION OF EXOGENOUS MODEL VARIABLES
Name
jescnpnon
Data Source
Service Area Base Year Characteristics: Socioeconomic Variables
DUACRE = Dwelling units per mile2 in area of analysis in 1960.
= (100*du60)/acre
where: du60 = 1960 census tracts* housing units in 100s Census
acre = 1960 area of census tracts in miles Census
VACHSE = Percent vacant available dwelling units in area of
analysis in 1960. Census
INCOME - Relative medium income of families and unrelated
individuals in area of analysis compared to county
income levels in 1960.
= (10*inc)/median
where: inc = 1960 median income for families in
census tracts in $10s
median = 1960 median ir.come for families in
the county**
VACOFF = Vacancy rate of office buildings in area of analysis
in 1960.
UNEMP = Unemployment rate in area of analysis (census tracts)
in 1960.
OFFJOB = Office employment per mile in area of analysis in 1960.
=^ (100*smoff )/acre
where: smoff = I960 office employment in census Census
tracts in 100s***.
WWJOB = Warehouse and wholesale employment per mile in area
of analysis in 1960.
= (100*wwemp)/acre
where: wwemp = 1960 employment in wholesale trade4" in Census
census tracts in 100s.
Census
Census
BOMA/PTanning
Agency
Census
* Census tracts refers to those tracts most closely approximating the
area of analysis in areal extent.
**
County refers to the county containing most of the legal service area.
*** Fire, Business Services, Public Administration, and Repair Services.
+ Trucking, Warehousing, and Wholesale Trade.
C-5
-------
TABLE C-2 (CONTINUED)
DEFINITION OF EXOGENOUS MODEL VARIACLES
Name - Description
Data Source
MANJOB = Manufacturing employment per mile in area of analysis
in 1960.
= (100*manemp)/acre
where: manemp = 1960 manufacturing employment in
census tracts in 100s.
JOBS = Total employment per mile in area of analysis in 1960,
(100*totemp)/acre
where: totemp = 1960 total civilian employment in
census tracts in 100s.
2
NONHH = Non-household population per mile in area of analysis
in 1960.
(100*nonh60)/acre
Census
Census
POOR
RENTS*
VALUE
ROOMS
KIDS
where: nonh60 = 1960 population in group quarters in
census tracts in 100s.
= Percent of total families with income below $3,000
in area of analysis (census tracts) in 1960.
= Percent of total housing units that are renter occu-
pied in area of analysis in 1960.
= Median value of housing units in area of analysis
(census tracts) in 1960.
=v Median number of rooms in housing units in area of
analysis (census tracts) in 1960.
= School age children per 100 households in area of
analysis in 1960.
= 100*sch60/du60
where: sch60 = 1960 population 0-14 years of age
in census tracts in 100s.
Census
Census
Census
Census
Census
Census
* Used only in the disaggregation analysis of RES.
C-6
-------
TABLE C-2 (CONTINUED)
DEFINITION OF EXOGENOUS MODEL VARIABLES
Name - Description
Data Source
GOVT = Total county government expenditures in 1962 in 10 $.
LAND = Price of vacant land in area of analysis in 1960
relative to median regional income.
= price/median
where: price = median price of one acre of residential
vacant land ($) in area of analysis
in 1960.
Planning
Agency
Census
Census
UNIV = Categorical variable to indicate the presence or
absence of a college or university in the area of
analysis in 1960.
where: 1 = a college or university existed in census
tracts in 1960.
0 - none existed.
RELJOB = Relative employment density of the area of analysis
in 1960.
= JOBS/(100*rempl/areal)
where: rempl = 1960 total employment in the SMSA
in 100s. 2
areal = 1960 SMSA land area in mile .
RELWW = Relative warehouse and wholesale employment density
of the area of analysis in 1960.
= WWJOB/(100*rwwemp/areal)
where: rwwemp = 1960 employment in warehouse and
wholesale trade in the SMSA in 100s.
RELOFF = Relative office employment density in the area of
analysis in 1960.
= OFFJOB/(100*roffl/areal)
where: roffl = 1960 SMSA office employment in 100s.
V
RELMAN = Relative manufacturing employment density of the area
of analysis in 1960.
= MANJOB/(100*rempl*manper/areal)
where: manper = percent of 1960 total SMSA employment
in manufacturing.
Census
Census
Census
Census
Planning
Agency
C-7
-------
TACLE C-2 (CONTINUED)
DEFIMTIOIJ OF EXOGENOUS MODEL VARIABLES
Name - Description
Data Source
POPDEN = Population density of the area of analysis in 1960
in persons per mile2.
= (100*pop)/acre
where: pop = total population of census tracts in
1960 in 100s.
Census
SFDET = Percent of housing that is
area of analysis in 1970.
= unitl/hse70
where: unit! =
single family detached in
number of single family detached
units in census tracts in 1970 in 100s.
hse70 = total number of housing units in
census tracts in 1970 in 100s.
Census
Census
MF
Percent of
analysis in
(unit34+uni
where: unit34 =
housing that is multifamily in area of
1970.
t5+unit50)/hse70
4 family housing
tracts in 1970 in
number of 3 and
units in census
100s.
units = number of 5-49 family housing units
in census tracts in 1970 in 100s.
unit50 = number of 50+ family housing units
in census tracts in 1970 in 100s.
Census
Census
SFATT =
((hse70-uni
Percent of
in area of
tl)/hse70)-MF
housing that is single family attached
analysis in 1970.
PCOMM1 = Percent of
50,000 ft2
commercial development less than
in floor area.
= comml/(comml
where: comml
comm2 + comm3)
total commercial floor space in
1,000 ft2 in area of analysis in
1970 for buildings with less than
50,000 ft2.
commZ - total commercial floor space in
1,000 ft2 in 1970 in area of analysis
for buildings with less than 100;000-ft:
but greater than 50,000 ft2.
comm3 = total commercial floor space in 1,000
ft2 in 1970 in area of analysis for
buildings with greater than 100,000 ft2.
Planning
Agency
C-8
-------
TABLE C-2 (CONTINUED)
DEFINITION OF EXOGENOUS ''ODEL VARIABLES
-;a,T;-5 - Description
Data Source
PCOMM2 = Percent of commercial development with
floor area between 50,000 and 100,000 ft2.
= comm2/(comml + comm2 + commS)
PCOf'M3 = Percent of commercial development with
floor area greater than 100,000 ft2.
= comm3/(comml + comm2 + commS)
Service Area Base Year Characteristics: Accessibility
DISCED = Distance in miles from centroid of area of analysis
to centroid of nearest CBD* in year (t + 0)**
RRMILE = Railroad mileage per mile in the area of analysis
analysis in year (t + 0)
= (rail*640)/AREA
where: rail = railroad mileage in area of analysis
in year (t + 0)
ACCESS = Limited-access highway interchanges per mile in
area of analysis in year (t + 5)
= (intchg*640)/AREA
where: intchg = number of limited access inter-
changes in area of analysis in
year (t + 5)
INTDEN = Relative limited-access highway interchange density
vof the area of analysis in year (t + 5)
= ACCESS/(ctyacc*640/county)
where: ctyacc = number of limited access inter-
changes in the county in year (t + 5)
county = area of county in mile2
TRANS = Number of transit stops (bus and commuter rail) in
the area of analysis in year (t + 0)
AIRPRT = Distance in miles from centroid of area of analysis
to centroid of nearest commercial airport in the
year (t + 0)
USGS topo-
graphic map
USGS topo-
graphic map
USGS topo-
graphic map/
Planning
Agency
USGS topo-
graphic map/
Planning
Agency
Planning
Agency
USGS topo-
graphic map
* Central Business District, defined as the center of the nearest urban
area with population exceeding 100,000.
**t U cue year tne-wastewater major project was completed and became
"operational.-
C-9
-------
TABLE C-2 (CONTINUED)
DEFINITION OF EXOGE.'i.'JS MODEL VARIABLES
liar.e - Description
Data Source
Service Area Base Year Characteristics: Land Use Constraints
AREA = Area of analysis in acres
VACANT = Percent vacant developable acreage in area of analysis
in year (t + 0)
= vacdev/(AREA-vacund)
where: vacdev = vacant developable acreage in area of
analysis in year (t + 0)
vacund = vacant undevelopable acreage in area
of analysis in year (t + 0)
RZONED = Percent of total acreage zoned for residential use in
the area of analysis in year (t + 0)
= rzone/AREA
where: rzone = acres of land zoned for residential use
in the area of analysis in year (t + 0)
CZONED = Percent of total acreage zoned for commercial use in the
area of analysis in year (t + 0)
= czone/AREA
where: czone = acres of land zoned for commercial use
in the area of analysis in year (t + 0)
OZONED = Percent of total acreage zoned for office use in the
area of analysis in year (t + 0)
= vozone/AREA
where: ozone = acres of land zoned for office use in
the area of analysis in year (t + 0)
IZONED = Percent of total acreage zoned for industrial use in
the area of analysis in year (t + 0)
= izone/AREA
where: izone = acres of land zoned for industrial use
in the area of ana-lysis in year (t + 0)
SOILS = Percentage of total area of analysis having a "severe"
soil type classification for urban development suita-
bility
= soil/AREA
where: soil = Acreage in area of analysis with a
"severe" soil type classification with
regard to suitability for urban devel-
opment.
Project Data
Planning
Agency
Planning
Agency
Planning
Agency
Planning
Agency
Planning
Agency
Planning
Agency
Planning
Agency/Soil
Conservation
Service (SCS)
£-10
-------
TABLE C-2 (CONTINUED)
DEFINITION OF EXOGENOUS HODEL VARIABLES
r;are - Description
Data Source
"severe"
ONLOT = Percentage of total area of analysis having a
soil type classification for on-lot sewage disposal
suitability.
= onsite/AREA
where: onsite = Acreage in area of analysis with a
"severe" soil type classification for
on-lot sewage disposal suitability.
LIMITS = Categorical variable to indicate the seventy of gov-
ernmental restrictions on on-lot sewage disposal during
the period (t + 0) to 1970.
is prohibited entirely.
is prohibited except on
where: 4 = on-lot disposal
3 = on-lot disposal
large lots.
2 - on-lot disposal is permitted but percola-
tion tests are required.
1 = on-lot disposal is permitted but package
plants are prohibited.
0 = no restrictions.
POLICY = Categorical variable to indicate the presence or
absence of governmental policies designed to limit the
number of hookups to the sewerage system in area of
analysis anytime during the period (t + 0) to 1970.
Examples are sewer moratoria and rationed connections.
where: 1 = policies on hookup limitations existed
0 = policies did not exist.
TLIMIT = The number of years during the period of analysis that
on-site sewage disposal was limited
= 1970-ylimit
where: ylimit = the year on-site disposal limitations
went into effect (or 1970 if no
restrictions).
TPOLIC = The number of years during the period of analysis that
sewer system hookup limitations were in effect
= 1970-ypolic
where: ypolic = the year limitations on sewage system
hookups went into effect (or 1970 if
no restrictions)
Planning
Agency/SCS
Planning
Agency/
Local
Government
Planning
Agency/
Local
Government
Planning
Agency/
Local
Government
Planning
Agency/
Local
Government
C-ll
-------
TABLE C-2 (COiiTi:;UED)
DEFINITION OF EXOGENOUS MODEL VARI-V,LES
Name - Description
Data Source
Census
Regional Growth Factors
ENERGY = Relative electrical energy cost factor in municipality
compared to average U.S. commercial rate in 1960.
= encost/$51.59
where: encost = cost of energy for commercial users
in 1960 in units of dollars per
1500 KWh.
DRIVE =
- 100s of workers who drive per mite2 in the county in 1960.
= (autoeoj/county
where: autoco = workers who drive to work in county
in 1960 in 100s.
RIDET = Workers who ride mass transit per mile in the county
in 1960.
= massco/county
where: massco = workers who use mass transit in county
in 1960 in 100s.
PERCHG = Percent change in county population 1960-1970
DENCHG = Population density change in the county 1960-1970.
= (copop2/county - copopl/county)
where: copopZ = county population in 1970.
copopl = county population in 1970.
JOBCHG -'Change in total regional employment per mile 1960-1970.
= 100*(remp2 - rempl)/SMArea
where: rempl = 1960 total employment in SMSA in 100s.
rempZ = 1970 total employment in SMSA in 100s.
SMArea= area of SMSA in miles .
HSECHG = Percent change in total regional housing units 1960-1970.
= (rhse2 - rhsel)/rhsel
where: rhsel = total housing units in SMSA in 1960
in 100s.
rhse2 = total housing units in SMSA in 1970
in 100s.
Census
Census
Census
Census
Census
Census
Census
Census
Census
Census
C-12
-------
TABLE C-2 (CONTINUED)
DEFINITION OF EXOGENOUS TCOEL VARIACLES
Name - Description
Data Source
COMCHG = Percent change in total regional retail trade* employment
1960-1970.
= (rcom2 - rcoml)/rcoml
where: rcoml = total retail trade employment in SMSA
in 1960 in 100s.
rcomZ = total retail trade employment in SMSA
in 1970 in ICOs.
POPDIF = Percent change in regional population 1960-1970.
MANCHG = Percent change in regional manufacturing employment
1960-1970.
= (perman*remp2-manper*rempl)/(manper*rempl)
where: perman = % of 1970 total SMSA employment in
manufacturing
manper = % of 1960 total SMSA employment in
manufacturing
SERCHG = Percent change in regional services** employment 1960-
1970.
= (rser2 - rserl)/rserl
where: rserl = 1960 SMSA service employment in 100s.
rser2 = 1970 SMSA service employment in 100s.
HOSCHG = Percent change in regional hospital employment 1960-1970.
= (rhosp2 - rhospl)/rhospl
where: rhospl = 1960 SMSA hospital employment in 100s.
v rhosp2 = 1970 SMSA hospital employment in 100s.
EDUCHG = Percent change in regional educational employment
1960-1970.
= (red2 - redl)/redl
where: red! = 1960 SMSA educational employment in 100s.
red2 = 1970 SMSA educational employment in 100s.
OFFCHG = Percent change in regional office*** employment 1960-
1970.
= (roffZ - roffl)/roffl
where: roffl = 1960 SMSA office employment in 100s.
roff2 - 1970 SMSA office employment in 100s.
Census
Census
Census
Census
Census
Census
Census
Census
Census
Census
Census
Census
Census
* Food and Dairy, Eating and Drinking, and Other Retail.
** Professional and Related Son/ices, Other Personal Services, Entertainment
and Welfare and Fraternal Organizations.
*** FIRE, (Finance, Insurance, and Real Estate) Business Services Public
Administration, Repair Services, and Public Administration.
C-13
-------
TABLE C-2 (CO;ri;;UED)
DEFINITION OF :x.occ;;oi'S "OCLL VARIABLES
Nane - LQScriotion
Data Source
PERCAP = Median far-rHy income in SMSA in 1970. Census
EMPOP = Regional (SMSA) employment to population ratio in 1970. Census
= remp2/rpop2
where: rpop2 = 1970 SMSA population in 100s. Census
CAPCHG = Change in median family SMSA income 1960-1970.
= PERCAP - per60
where: per60 = median family income in SMSA in 1960. Census
STAY Index of mobility - % of 1960 families who were in the Census
same house in 1955.
Wastewater Treatment Major Project Characteristics
TIME = Number of years available for secondary growth to occur.
= phyear-year
where: phyear = the year aerial photographs were taken Planning
from which land use data was extracted. Agency
year = the year construction was completed on Project
the major project on initial phase. Data
PHASE = Categorical variable to indicate the presence or Project
absence of phasing of the major project. Data
where: 1 = phasing has occurred.
0 = phasing has not occurred.
COST =>Normalized per capita local costs ($) of the major
project in area of analysis in year (t + 0).
= 1,000* (totest/pil - fedcst/pi2)/popcom
where: totcst = total major project construction Project
cost (1,000 $) Data
fedcst = federally funded share of major Project
project cost (1,000 $) Data
pi! = consumer price index for "year"
pi2 = consumer price index for the year of
federally funding of the major project
popcorn = population served by major project Project
facility in year (t + 0) Data
LENGTH = Running length of interceptor sewer lines in miles Project
going through relatively undeveloped land (<.1 du.per Data
acre) in area of analysis in year (t + 0).
-------
TABLE C-2 (CONTINUED)
DEFINITION OF EXOGENOUS MODEL VARIABLES
Name - Description
Data Source
CROSS = Index of available undeveloped land in the area of
analysis through which interceptor sewers go through
in t + 0.
= 640*LENGTH/AREA
SERVED = Percent of the area of analysis easily served by the
major project in year (t + 0)
= area5k/AREA
where: areaSk = acres of land within 5,000 feet of the
major project interceptor sewer in
area of analysis in year (t + 0)
CAPAC1 = Total hydraulic design capacity of the major project
wastewater collection system in million gallons per
day (mgd) in year (t + 0) (or in 1965 if a phased
project)
CAPAC2 = Total hydraulic design capacity of the major project
wastewater treatment plant in mgd in year (t + 0)
PEAK = Actual Peak flow in the major project wastewater system
in mgd in year (t + 0)
RECAP = Reserve capacity of the wastewater major project
collection system in year (t + 0)
= CAPAC1 PEAK
RECAP1 = Percent reserve capacity of the wastewater major
project collection system in year (t + 0)
= 1QO*RECAP/PEAK
RECAP2 = Percent reserve capacity of the wastewater major
project treatment plant in year (t + 0)
- 100*(CAPAC2 - PEAKJ/PEAK
Project
Data
Project
Data
Project
Data
Project
Data
C-15
-------
TABLE C-3
METRIC UNITS OF VARIABLES USED IN THE PREDICTIVE EQUATIONS AND WORKSHEETS
Variable
Predictive
Equations
Dependent
RES +
COMM+
OFFICE +
MANF +
WHOLE+
HIWAYS+
EDUC +
REC +
Names
Worksheets
Residential
Commercial
Office-
Professional
Manufacturing
Wholesale
Highways
Education
Recreation
Metric Units
Dwelling units per 10,000 square meters of
area of analysis
Square meters of floor area per 10
meters of area of analysis
Square meters of floor area per 10
meters of area of analysis
Square meters of floor area per 10
meters of area of analysis
Square meters of floor area per 10
meters of area of analysis
Lane kilometers per 10,000 square
area of analysis
Square meters of floor area per 10
meters of area of analysis
Square meters of recreational land
,000 square
,000 square
,000 square
,000 square
meters of
,000 square
use per
10,000 square meters of area of analysis
OTHER +
SFDET
SFATT
MF
PCOMM1
PCOMM2
PCOMM3
Independent
VACANT
LAND
RECAP!
MANJOB
STAY
DRIVE
SERVED
KIDS
ACCESS
Other
% Single Family
% Two Family
% Mul ti family
% Small Comm'l
% Med Comm'l
% Large Comm'l
Vacant Land
Land Cost
% Collection
Reserve
Manufacturing
Density
Nonmobility
Driver
Density
Sewer Service
Kid Density
Interchanges
Square meters of floor area per 10
meters of area of analysis
*
*
*
*
*
*
*
*
*
Employees per square kilometer of
analysis
*
,000 square
area of
100s of drivers per square kilometer of area
analysis
*
*
Interchanges per square kilometer
of analysis
of area
of
* Same as English units.
+ Units for Predictive Equation variables are per 10,000 acres (square meters) of
area of analysis, while units for Worksheet variables are totals for the entire
area of analysis.
C-16
-------
TABLE C-3 (CONTINUED)
METRIC UNITS OF VARIABLES USED IN THE PREDICTIVE EQUATIONS AND WORKSHEETS
Variable Names
Predictive
Equations
JOBCHG
RRMILE
OZONED
PEAK
DUACRE
IZONED
POPDIF
VACOFF
AIRPRT
OFFJOB
EMPOP
UNEMP
PERCAP
GOVT
RECAP
CROSS
POOR
LIMITS
PERCHG
COST
CAPAC2
VACHSE
INTDEN
RZONED
CAPCHG
HOSCHG
HSECHG
TLIMIT
PHASE
TRANS
DISCED
vacdev
vacund
AREA
price
median
Worksheets
Employment Growth
Railroads
Office Zoning
Peak Flow
House Density
Industrial Zoning
Population Growth
Office Vacancy
Airport Distance
Office Employees
Employee Ratio
Unemployment
Future Income
Government
Collection Reserve
Interceptor Density
Poverty
Onsite Restrictions
County Growth
Sewer Costs
Treatment Capacity
Vacant Houses
Interchange Density
Residential Zoning
Income Growth
Hospital Growth
Housing Growth
Restriction Years
Phasing
Transit Stops
CBD Distance
Vacant Developable
Vacant Undevelopable
Area of Analysis
Median Price
Median Income
Metric Units
Employees per square kilometer of area
of analysis
Kilometers of railroad track per
square kilometer of area of analysis
*
*
Dwelling units per square kilometer
of area of analysis
*
*
*
Kilometers
Employees per square kilometer of area
of analysis
*
*
*
*
*
Kilometers of interceptor pipe per
10,000 square meters of area of
analysis
*
*
*
*
*
*
*
*
*
*
*
*
*
*
Kilometers
Square meters
Square meters
Square meters
*
*
Same as English units.
C-17
-------
TABLE C-3 (CONTINUED)
METRIC UNITS OF VARIABLES USED IN THE PREDICTIVE EQUATIONS AND WORKSHEETS
Variable Names
Predictive
Equations Worksheets
Metric Units
CAPAC1
manemp+
acre
autoco
county
areask
sch60-t-
du60
intchg
rempl +
remp2+
SMArea
rail
ozone
izone
smoff+
rpop2+
LENGTH
to test
fedcst
pil
Pi 2
popcorn
ctyacc
rzone
per60
rhospl+
rhosp2+
rhsel+
rhse2+
Collection Capacity
Manufacturing Workers
Tract Area
Drivers
County Area
Sewered Land
School Kids
Dwelling Units
Limited -Access
Current Employment
Future Employment
SMSA Area
Track
Zoned Office
Zoned Industrial
Office Workers
Future Population
Interceptors
Project Cost
Federal Funds
Index One
Index Two
Population Served
County Interchanges
Zoned Residential
Current Income
Current Medicals
Future Medicals
Current Houses
Future Houses
Square kilometers
*
Square kilometers
Square meters
*
*
*
*
Square kilometers
Kilometers
Square meters
Square meters
*
*
Kilometers
*
*
*
*
Square meters
*
*
*
*
*
*Same as English units.
+Units for Predictive Equation variable are in 100s while units for
Worksheet variables are in Is.
C-18
-------
APPENDIX D
GRAPHS OF ACTUAL VERSUS PREDICTED LAND USE
FOR THE CROSS-VALIDATION ANALYSIS
D-l
-------
1. RES Equation
TSP
flGUSN
P A ii L
PLOT OF ACTUAL<*> AMD FITTEC<+> VALUES
PLOT CF f-.ESIt UALS(O)
10
8
9
11
15
16
17
19
20
22
t-j 23
25
28
31
33
35
36
37
ACTUAL
5911.
9706.
.1197E+05
.2357E+OS
6072.
3745.
9709.
5100.
7540.
5017.
7372.
5903.
4626.
7772.
6516.
.1358E+05
.1050E+05
369.0
.1322E+05
FITTED
6919. * +
7983. + *
.1547E+05 * +
.1374E+05 +
6033. +
.1114E+05 * +
b938. + *
.1205E+05 * +
6407. + *
1461. + *
3406. + *
469.3 + *
3351. «• *
5713. + *
6803. *+
.1344E+05 +
.1412E*05 * +
4711. * +
.1461E+05 * +
RESIDUAL
0.0
-.101E+04 . 0.
.172t+04 . . 0 .
-.35CL+04 0
* .9B2E+04
39.4 . 0
-.740F+04 0 . .
771. . .0
-.6-95E + 04 0
.113E+04 . . C .
,3b6t+04 . . 0.
.397T+04 . . C
.54-+04 . . . 0
.127^04 . . 0 .
.206L+04 . . 0 .
-287. . 0
134. .0 .
-.362L+04 C .
-.434E+04 C.
-.1391+04 . 0 .
39
5375.
3670.
.171E+T4
-------
2. COMM Equation
LI'":
1
TC'
T~p
V!
1 f f CF 61 MS 1 •
PLOT Ct- A C T L A L ( * > A > D F I T T j. L ( * ) V A LIt S
'LC T fF .''
MM," )
TO
1
*
8
9
11
15
16
17
19
20
22
23
25
28
31
33
34
35
36
37
ACTUAL
892.4
1794.
1206.
3H62.
111?..
1104.
2027.
701. "
?776.
1897.
1640.
774. C
722.8
1197.
807.8
1071.
.1127E+05
245.7
1510.
FITT£D
1 438. * +
2fcS7.
250. P. + *
1 9 1 . t + *
llc.9. * +
2251. * +
251.4 + *
4336. * +
56^3. * +
178.2 + *
1870. *+
2373. * +
1379. * +
2596. * *
537.4 +*
-133. « + *
6266. *
-37.09 * *
1950. * +
f c c. lot A I
-tOL- .
-873.
c.i L 5 .
.3f,7r + C4
-71.2
-•119L+G4
.178!:+04
-.?«-.?E*C* C
-.2V2L+C4
.172E+01
-270.
-.1COL+04
-656.
-.140E+04
270.
.120F. + 04
* .bOOE+04
?P3.
-441 .
0.0
C.
• * • •
0
• • •
0
. c .
. . 0.
1
• • •
0 .
• • [. •
0. " .
.0 . .
. c .
.0 . .
.0 •
. I .
• • •
. .0 .
.0. .
39
609.2
-------
CD
J=-
3. OFFICE Equation
PRINCETON UNIVERSITY
ID
LI
4
8
9
11
15
16
17
19
20
22
23
25
28
31
33
34
35
36
37
NE 15
ACTUAL
86.40
288.5
555.1
687.5
935.9
403.4
822.3
995.9
230.9
177.8
34.00
94.70
144.6
577.2
385.5
296.6
2111.
94.60
489.7
PRII
FITTED
468.1
914.7
1009.
-214.4
648.9
647.7
161.0
1000.
844.7
287.5
210.7
655.8
393.9
681.8
515.2
1065.
1312.
64.15
681.7
TSP
PLOT OF ACTUAL(*) AND FITTEDC+) VALUES
OF fUOUSTi I?*-"?
HAGL
RESIDUAL
-7H2.
-62f .
-454.
S02.
287.
-244.
661.
-4.31
-614.
-110.
-177.
-561.
-249.
-105.
-130.
-768.
799.
30.4
-192.
PLOT Ch RESIt OALS«j )
0.0
0 .
0.
0.
0
. 0
39
98.60
110.7
-12.1
-------
UNIVERSITY
7 Si
VETS I CK OF i'cL l.'ST
' r-E. 1
PLC7 OF AC7UAL<*) t"D FITTLr<+> VALLLS
f:LOT OF F05
335.5
4542.
FITTER
1717. +*
1090. +
2irft. * +
lie;. * *
1205. **
3332. +
435.1 *
4755. * *
4363. * *
156.4 + *
177.2 •• *
2727. * +
344.9 * *
1313. +
1848. * +
5971. * *
3865. +
51.18 +*
2259. + *
Fu" S 1 C U A L
f-7.9
-39.4
-.159E+04
-366.
-134.
-6£.8
lit.
-.333F+04
496.
.111F+04
.219E+04
-.237E*04
700.
-158.
-r37.
-.545E*04 0
* .68CE+C4
284.
.228E+04
C .0
0
c
. c .
0.
0.
0
0
0 . .
.0
. 0
» •
0
.0
0.
0.
* •
• •
0
• *
•
•
•
•
•
•
•
•
*
•
c
.
•
•
•
•
•
•
0
39
140. P
-363.1
504.
.0
-------
5. HIWAYS Equation
LINE 17 PRINCETON1 UNIVERSITY
TSF
VEkSICr. Of- AUGUST,
21
PLOT OF ACTUAK*) AND FITTED(+) VALUES
PLOT CF FESICUALS(O)
ID
cr>
8
9
11
15
16
17
19
20
22
23
25
28
31
33
34
35
36
37
39
ACTUAL
15.40
47.00
84.00
57.50
56.60
42.3C
25.30
3.000
15.20
45.60
13.40
30.30
30.80
48.30
34.50
24.80
48.20
14.00
36.50
15.50
FITTED RESIDUAL
8.726 + * 6.67
48.93 *+ • -1.93
114.9 * + -30.9
38.66 + * 18.R
53.74 +* 2.8f
68.18 * •» -25.9
37.00 * * -11.7
62.94 * + -59.9
flO.OO * + -64.8
31.15 * * 14.4
19.50 * + -6.10
70.58 * + -40.3
40.50 * * -5.70
53.30 * + -5.00
76.19 * > -41.7
105.9 * + -Fl.l 0
64.83 * + -16.fc
21.67 * «• -7.67
41.14 * * -4.64
5.484 + * 10.4
c.r
. C
0
.0
0.
0.
0 .
0.
0 .
0.
0.
0 .
-------
6. EDUC Equation
L]
. ID
*
8
9
11
15
16
17
19
20
22
23
25
V
•Nl 28
31
33
34
35
36
37
It.? 18
ACTUAL
454.0
635.8
724.0
984.8
350.8
499.5
982.5
368.6
1279.
707.3
655.0
359.2
715.5
662.5
643.7
1037.
457.6
95.50
615.5
PRII
FITTED
488.2
694.8
747.4
21*. 8
652.5
408.0
1108.
440.9
896.0
489.7
320.7
456.4
863.8
690.1
630.7
895.5
460.7
-41. b6
804.4
PRINCETON UNIVERSITY
VEF-Sin- OF- AulUSTi
PLOT OF- ACTUALt*) AND FTTTLC<+) VALUE.S
PLCT CF
RESIDUAL
* + -34.2
* + -59.0
*+ -23.4
+ * 766.
* * -302.
+ * 91.5
* + -126.
* + -72.3
+ * 383.
* * 218.
+ * 334.
* * -97.2
* + -148.
*+ -27.6
+ 13.0
•» * 141 .
* -3.11
« 137.
* + -189.
c.c
C .
• 0 . .
. 0 . .
• • . <
0. . .
. . 0 .
. 0 . .
. 0 . .
. • .
0
• • .
. 0 • .
. 0 .
. 0 . .
C
. 0 .
0
. 0 .
. 0 . .
39
11.60
109.6
-98.0
0 .
-------
7. REG Equation
LINE 19
PRINCETON UNIVERSITY
TSP
CF AUGUST,
PLOT OF ACTUALC*) AND FITTEDC + ) VALUE'S
ID
4
8
9
11
15
16
17
19
20
22
V "
00 25
28
31
33
35
36
37
ACTUAL
28.00
37.00
437.0
500.0
13.00
103.0
106.0
444.0
289.0
65*00
23D.O
382.0
62.00
247.0
98.00
193.0
203.0
19.00
215.0
FITTED
108.1 * +
184.5 * +
930.6 *
547.9 * +
292.8 * +
263.1 * +
193.2 * +
362.9 *
401.6 * +
-35.56 + *
129.3 + *
432.1 * +
93.51 * +
292.9 * +
229.2 * +
314.9 * +
680.7 * +
140.4 * +
76.13 + *
RESIDUAL
-tO.l
-147.
+ -494.
-47.9
-2eo.
-160.
fcl.l
-113.
101.
101.
-50.1
-31.5
-45.9
-131.
-122.
-478.
-121.
139.
PLOT OF RESIDUAL^C)
0.0
.0
. 0
0.
0.
0.
. 0
. 0
•
. 0
0 .
39
129.0
60.81
f 8.2
-------
8. WHOLE Equation
CD
I
ID
4
8
9
11
15
16
17
19
20
22
23
25
28
31
33
34
35
36
37
IN"7 20
ACTUAL
74.30
182.3
937.6
312.5
333.7
432.9
.0
425.0
1871.
1399.
300.6
646.0
361.4
570.4
671.0
156. 5
256.5
9.300
732.1
I1 P. I!*
FITTED
389.2
17.65
1156.
-165.6
-11.39
563.2
108.5
652.7
1099.
672.5
576.9
322.7
511.5
53. P9
953.5
382.3
769.9
257.0
547.2
I1 P. I \CFTu;. UNIVERSITY
TiF
v L « S T CI f F A I: 'J I S T , 1'.
PLOT OF ACTUAL(*> Al>0 FITTf.C( + > VALUES
CF
+ * Ifa5.
«• -219.
+ * 478.
+ * 345.
* + -130.
* + -10V.
* + -226.
+ * 7 7 2 .
+ * 726.
* * -?7f .
* * 3?3.
* + -150.
+ * 517.
+ -282.
* + -22f..
* + -513.
+ -248.
* * 165. .
0
.
•
. 0
.
. 0
.
.
.
. 0
0
0 .
. 0
*
I .C
. C
0 .
. 0
39
10.90
-500.fi
-------
9. OTHER Equation
LINE 21
PRINCETON UNIVERSITY
TSP
VERSION1 OF AUGUST, 19b9
PLOT OF ACTUALC*) A\P FITTEC(*> VALULS
f-LOT Cf RFSILUALS(O)
ID
8
9
11
15
16
IT
19
20
22
23
25
28
31
33
34
35
36
37
39
ACTUAL
486.1
308.8
763.6
866.2
377.0
268.7
558.5
532.6
82.20
389.8
274.6
414.0
282.9
146.9
316.7
58.30
1016.
45.40
FITTED RESIDUAL
100.1 + * 3P6.
116.4 * * 152.
680.0 ** -7b.3
446.1 •» * 317.
588.9 + * ?77.
842.2 * * -465.
133.8 + * ?01.
1174. * + -905.
1510. * + -552.
159.9 + * 373.
117.1 *+ -34.9
1233. * + -M3.
510.2 * + -236.
940.8 * + -527.
605.6 * * -223.
2343. * + -.220E+-04 0
724.3 * * -tCtt.
-227.4 «• * 286.
220.0 + * 796.
61.86 *+ -16.5
.0
0.0
•
.0
0.
. 0
. 0
•
. 0
0 .
0.
-------
APPENDIX E
SUPPLEMENTARY INFORMATION
Thomas McCurdy
EPA Project Officer
E-l
-------
I. General
The EPA Technical Committee mentioned in the Acknowledgements on p. iii
included the following personnel:
Martha Burke Office of Transportation and Land Use Policy
J. David Foster Control Programs Development Division, OAQPS
Walter Issac Municipal Construction Division, OWPO
Thomas H. Pierce Office of Land Use Coordination
John L. Robson Strategies and Air Standards Division, OAQPS
David Sanchez Control Programs Development Division, OAQPS
Carol Wegrzynowicz Municipal Construction Division, OWPO
OAQPS stands for Office of Air Quality Planning and Standards and OWPO means
Office of Water Program Operations.
A number of Technical Committee members reviewed a draft of this report,
A consensus comment was that a number of items were not fully explained in
the contractor's text. Thus, the purpose of this appendix, written by the
EPA Project Officer, is to supply needed supplementary information. The
material is organized around the worksheets. Information is presented on
the following topics:
1. how to obtain input variables for the predictive model (Worksheet 1)
2. how to estimate the sensitivity of the model to changes in input
variables (related to Worksheet 1).
3. how to distribute predicted land use within the area of analysis
(related to Worksheet 2).
4. how to estimate residential and commercial breakdowns (Worksheets
3 and 4).
E - 2
-------
II. Simplifying the Analysis and Obtaining Input Data
To use the nine predictive (Worksheet 2) and six disaggregated (Worksheet 4)
equations requires the user to come up with values for 52 variables. (Additional
input data are required for VMT and emissions prediction equations). Some datum
will be easy to obtain because it is available from the wastewater treatment
facility owner or operator; this is particularly true of "major project"
variables, such as collection system capacity and peak flow. Other data may
be difficult to obtain, especially non-OBERS future year socio-economic data.
It is the purpose of this Section to provide some help on how a user can obtain
needed input data. A simplifying procedure is also provided to reduce user
effort to a minimum. This procedure, however, correspondingly reduces the
amount of confidence that can be placed in GEMLUP predictions. Using the
full-blown analysis allows the user to say that there is a 90 percent pro-
ability that the predicted land uses will be within ± (x) percent of their
true value (the percentages being obtained from optional Worksheet 5 as a
confidence interval). If any short-cut procedure is used, on the other
hand, nothing can be said about the confidence level of a prediction. The
tradeoff in using a simplified procedure is between time and rigor. Only a
user can decide which is more important in his or her situation.
Because many input variables are used in more than one predictive or
disaggregation equation, the number of input variables required varies with the
combination of predictive equations used. Although there are numerous ways
to combine the 9 land use predictive equations, there is one way that is
preferred. The preferred order of land uses is:
E - a
-------
1. residential 6. education
2. manufacturing 7- wholesale
3. commercial 8. other
4. office-professional 9. recreation
5. highways
Something should probably be said about highways being placed fifth in the
list. The causal analysis (Volume I) confirmed that Highways is an important
land use in inducing other development. A question may arise, then, in the
reader's mind concerning the reason why Highways is not ranked higher than
fifth. The reason is that the above listing strikes a balance between
theoretical importance and air pollution emissions importance -- for purposes
of GEMLUP, highways do not have any direct emissions. Motor vehicular
emissions due to vehicles using highways are accounted for in the VMT model
and are not tied directly to the amount of highway land use. Therefore, from
the viewpoint of an inventory for direct emissions, highways are an important
land use.
To use the residential predictive equation requires that 12 input vari-
ables be obtained; of these, nine are considered to be sensitive and the
remaining 3 variables are considered to be very sensitive. (The relative
sensitivity of a variable depends upon how much it affects predicted output
vis-a-vis other variables. This is discussed below under "sensitivity analysis."
Note that all residential variables are defined to be sensitive because of the
critical position that residential predictions play in subsequent analyses).
If manufacturing predictions are made after residential predictions, only four
additional variables are required. One of these is sensitive and another is
very sensitive. Five of the variables obtained for residential predictions
are also used in the manufacturing equation.
E - 4
-------
Marginal new information requirements for each equation, in the order,
listed above, are summarized in Table E-l. Also included are the disaggregated
equations, listed in the only order that is logical. The first five land
uses listed in E-l are considered to be "core" uses that should be predicted
for each application. The remaining land uses are considered to be of secondary
importance both from an emissions (practical) and theoretical point of view.
Given the extensive data requirements of the secondary and disaggregated
equations and the relative insensitiveness of their predictive land uses to air
pollution emissions, it probably is not worth the effort to obtain data needed
to use the equations if the analysis is being done solely to estimate area of
analysis emissions. In this instance, a user can save a lot of time by just
predicting core land uses and "scaling" secondary land uses from them by simple
ratios. The ratios, found below in Table E-2, are developed from raw data
obtained during GEMLUP case studies. They are based on mean values for the various
land uses. "Redundant" ratios are provided so that more than one core land
use can be used to estimate a particular secondary land use. The differing
estimates could then be averaged to obtain a representative value.
To provide an example, suppose the residential predictive equation was used
and it estimated that the area of analysis would have 10,000 dwelling units ten
years after the waterwater facility of interest is constructed. Using
2
the default ratios of Table E-2, approximately 710,000 ft of educational
building area might be expected (0.071 x 10,000 x 1,000). Suppose also that
2
the manufacturing equation was used and it predicted that 2,000,000 ft of
manufacturing would exist in the area of analysis by the end of the 10 year
period. Using the education-to-manufacturing default ratio of Table E-2
E-5
-------
Table E-l
MARGINAL INPUT INFORMATION REQUIREMENTS TO RUN THE PREDICTIVE
AND DISAGGRATED EQUATIONS
Land Use Type
CORE
Residential
Manufacturing
Commercial
Office - Professional
Highways
SECONDARY*
Education
Wholesale
Other
Recreation
DISAGGREGATED+
% Single Family
% Multi-Family
% Two-Family
% Large Commercial
% Medium Commercial
% Small Commercial
# of New
Input
Variables
Required
12
4
7
2
2
8
4
5
4
4
3
0
4
3
0
# of New
Sensitive or
Very Sensitive
Variables
12
2
3
0
0
1
3
4
3
1
0
0
0
0
0
# of
Previously Used
Input Variables
Required
0
5
5
6
6
4
5
3
4
6
4
0
2
4
0
Notes: *New variables are defined to be those not included in one or more
of the core land use equations, regardless of whether or not the
variable appears in a secondary land use equation.
+New variables are defined to be those not included either in one or
more of the core land use equations or in one or more of the preceding
disaggregated equations within the same class (i.e., residential and
commercial).
E-6
-------
Table E-2
DEFAULT RATIOS TO RELATE SECONDARY LAND USES TO CORE LAND USE TYPES
Variables Ratio
. . . with RESIDENTIAL
Education 0.071
Wholesale 0.054
Other 0.052
Recreation 0.024
Manufacturing 0.207
Commercial 0.212
Office - Professional 0.058
Highways 0.005
. . . with MANUFACTURING
Education 0.367
Wholesale 0.264
Other 0.252
Recreation 0.115
... with COMMERCIAL
Education 0.355
Wholesale 0.258
Other 0.246
Recreation 0.112
. . . with OFFICE - PROFESSIONAL
Education 1.317
Wholesale 0.948
Other 0.903
Recreation 0.413
Units1
1,000 fr/du
acres/du
1,000 ft2/du
lane miles/du
1,000 ft2/!,000 ft2
acres/1,000 ft'
1,000 ft2/!,000 ft2
acres/1,000 ft'
HOOO ft2/1,000 ft2
acres/1,000 ft'
Note: Abbreviations used in the units are:
p
ft = square feet
du = dwelling unit (housing unit)
Note that the unit given as 1,000 ft2/!,000 ft2 could be stated as
ft?/ft2. The first choice is shown because it has the same dimensions as
the land use prediction equations (Worksheet 3).
E-7
-------
p
results in an estimate of 734,000 ft of educational building area (0.367 x
2,000 x 1,000). Averaging the two estimates gives a "best-guess" prediction of
722,000 ft2 of educational building area.
Default value percentages for the two disaggregated land uses, residential
and commercial, are presented in Table E-3. While the estimates represent
mean values for the 40 cases used in GEMLUP, local data should be used
wherever possible. Since the data are relatively easy to obtain, particularly
for residential, the effort probably will be well worth it.
To use the disaggregated estimates, simply multiply the decimal equivalent
of the percentage value shown in Table E-3 by the predicted value "output"
by the commercial equation of Worksheet 2 multiplied by the area of analysis/
10,000 acres factor needed to convert density figures to totals. Suppose, for
•5 2
example, that the Worksheet 2 commercial prediction is 2,000 x 10 ft 710,000
acres and the area of analysis is 20,000 acres. The factor is therefore 20,000
acres divided by 10,000 or 2 (forget its units: the figure essentially
represents the number of 10,000 acres parcelsper area of analysis). Multiplying
2,000 x 103 ft by 2 gives 4,000,000 ft of commercial development building
area within the area of analysis as the predicted value. Using the default
percentages of Table E-3 results in the following breakdown by size of
2 2
commercial establishment. 428,000 ft of large commercial, 420,000 ft of
medium commercial, and 3,152,000 of small commercial (4 x 10 times 0.788).
If the user is going to predict with core equations and derive secondary
and disaggregated land uses from them, only the variables listed in Table E-4
will have to be obtained as inputs to Worksheet 1. The variables appear in the
order shown on Table 2-5 (p. 2-20). There are now 27 variables that have to be
obtained instead of the original 52 listed in Table 2-5. That table should be
consulted for a full description of the variables.
E-8
-------
Table E-3
DEFAULT PERCENTAGES FOR DISAGGRATED LAND USES
% of Total Land
Land Use Use in the Category
Single Family Detached 73.5
Multi-Family 10.6
Two Family^ 15.9
Subtotal 100.0
2
Large Commercial - 10.7
Medium Commercial 10.5
Small Commercial4 78.8
Subtotal 100.0
Notes: Included are single-family attached, single family quadplexes,
and the like
2 2
Commercial development with floor area > 100,000 ft
3 2
Commercial development with floor area between 50,000 and 100,000 ft
4 2
Commercial development with floor area < 50,000 ft
E-9
-------
Table E-4
LIST OF VARIABLES NEEDED TO PREDICT CORE LAND USES
Variable
Equation(s)
County Area
Vacant Developable
Vacant Undevelopable
Zoned Office
Zoned Industrial
Limited Access
Population Growth
Office Vacancy
Median Price
Future Employment
Airport Distance
Track
Area of Analysis
Sewered Land
Interceptors
Collection Capacity
Peak Flow
SMSA Area
Tract Area
Dwelling Units
School Kids
Nonmobility
Median Income
Government
Current Employment
Manufacturing Workers
Drivers
Residential
Residential, Manufacturing, Commercial
Residential, Manufacturing, Commercial
Manufacturing, Office-Professional
Office-Professional
Commercial
Office-Professional
Manufacturing
Residential, Highways
Commercial
Manufacturing
Manufacturing, Office, Highways
All Core Equations
Commercial
Hiahways
Residential, Commercial, Highways
Residential, Commercial, Office -
Professional, Highways
Commercial
Residential, Manufacturing, Commercial
Commercial, Office-Professional
Commercial
Residential
Residential, Highways
Highways
Commercial
Residential, Manufacturing
Residential
E-10
-------
III. Obtaining Needed Input Variables for Worksheet 1
Input data must be obtained for variables listed in Table 2-5 in order
to use all of the predictive equations. Basically, data are needed for three
time periods (time of wastewater facility completion, t, and five and ten years after
that ( t+5 & t+10 ) and four geographical area (area of anlysis, census tract(s)
most representative of the area of analysis, county, and SMSA). Because t
itself will most likely be in the future, there is a problem of getting values
for these supposedly current time period variables. The material to follow is
designed to help solve this problem. The areal issue will be discussed first.
The area of analysis is the legal service area of the wastewater treatment
facility. It generally includes only one watershed, but may include contiguous
areas whose sewage is conveyed to the natural drainage basin by some mechanical
means. Data needed for the area of analysis is entirely "physical" such as
geographical area, vacant undevelopable land, and amount of zoned residential
land. These data mostly come from maps.
Census tract socio-economic data are required. These data come from the
U.S. Bureau of Census's Census of Population and Housing;Census Tract series.
Updating this information will have to be done; some guidance on doing it appears
below. If the area of analysis is not tracted, municipality-level data are
substituted for census tract data. In that case, the data come from the Census
of Population and the Census of Housing:- Metropolitan Housing Statistics. Again,
the data will have to be updated.
The user will have to decide what census tract(s) are to be used. A rule-of-
thumb is to use census tracts that represents the area of analysis most closely.
This simple rule is difficult to apply in practice, as there are a large number
of circumstances possible, such as the area of analysis is:
E-ll
-------
1. fully tracted (in one or more municipalities).
2. partially tracted, with the remaining area in another municipality
(which itself may be tracted).
3. untracted in an incorporated area.
4. untracted in an unincorporated area.
By far the most common circumstance encountered during GEMLUP data gathering
is the first case -- the area of analysis was fully tracted. Often, however, a
number of municipalities were involved (or one municipality and tracted areas in
the county). This situation' is depicted in Figure E-l.
Required census tract socio-economic data would definitely be obtained
for the following census tracts: 4 to 6 and 8. Less than half of census tracts
1 and 9 are within the area of analysis, so they would not be used. Whether data
for census tract 2 would be used or not depends upon the situation. If most of
the people living in the tract are located within the area of analysis or_ if
population is distributed evenly throughout the census tract, the tract should
be used.
Note that just about all socio-economic variables are used on a per unit
area basis (or are relative variables to begin with; i.e., are percentage variables),
This is true for all census tract variables, which means that non-relative census
tract data are divided by tract area to derive the per unit area (density) variable
used in GEMLUP predictive equations. Examples are:
1. Manufacturing Density, or manufacturing employment divided by census
tract area.
2. House Density, or the number of housing units divided by census
tract area.
Census tract area will have to be obtained directly off of the maps included
in the SMSA Census Tract report (with a planimeter preferably) if the data are
E-12
-------
Figure E-l
HYPOTHETICAL EXAMPLE OF A FULLY-TRACTED AREA OF ANALYSIS
r
1
Municipality A
MMM» * •MVM* *» ••• • J ^ *M
$%
It;-':';:1
*.v;.;:'.:
..'.'.»
'•.V'v.v'.vV*
II
ill
<—t
Census Tract
^^_i^ ^b ^^i
^
Municipality B
Of
W$% Analysis
*i
County
Census
Tracts
E-13
-------
not already available. Many city and regional planning agencies have census
tract area data because of its use in transportation and housing plan develop-
ment. These agencies should be checked first for the area data.
Area information is also required for the municiaplity (but only if census
tracts are not used), the county, and the SMSA. Area information for these
jurisdictions are most easily found in the latest County and City Data Book.
The information is also contained in the Census of Population.
We turn now to the problem of obtaining estimates for variables in the
future. As mentioned in the opening paragraph of this section, basically three
time periods are involved: t, when the project "opens," t+5, five years after t,
and t+10, or ten years after the project opens. As we shall discuss below,
there probably will not be a good source of data (certcintly no single source) for all
variables for the needed time period, since t and t + 10 could be any year and
projected data are often only done for decennial or, at best, quinquennial periods.
Consequently, data available will have to be manipulated to obtain needed values.
Some assistance follows on how this can be done for each variable listed in Table 2-5.
V. County Area: area of the county in square miles for time t
(and t + 5). This variable does not change, except for minor,
infrequent boundary adjustments. Use the value in the latest
County and City Data Book or other Census material, or obtain
it from the county or regional planning agency. This variable
is required in the Residential equation, so it must be obtained
even if the simplified predictive approach is used. It is a
"sensitive" variable.
2. Vacant Developable: vacant developable acreage in the area of
analysis (area) in time t. This variable and the next, vacant
Undevelopable, are related by the following simple relation:
(1) (2) (3) (4)
Total Area = Developed + Vacant + Vacant
Acreages Area Developable Undevelopable
(Area of Acreaget Area Area
Analysis) Acreaget Acreaget
E-14
-------
Because it is probably easier to get estimates for variables
1,2 and 4, vacant developable can be obtained as a residual. The
fact that time t will be in the future and that variables 1 and 4
do not change over time, means that vacant developable is most easily
obtained by estimating developed area acreage (2) for time t and
substracting it from the quanity [(1) - (4)] . (How 1 and 4 are
obtained is explained later.)
Developed area acreage for time t is estimated from developed
area coverage for a current or past time period (t1) updated by some
proportionality factor. Development in time t1 is obtained from
aerial photographs (as was done in GEMLUP case studies), or from
existing land use maps or other planning studies. The portionality
factor applied to development in time t1 that intuitively makes a
lot of sense is a population/density ratio of the form:
population.
(estimated)
population^,
(known)
family size
(estimated)
housing unit density.
(estimated)
This ratio is called the "factor" in subsequent discussion,
simple formula for developed acreage becomes:
The
Developed
Area
Acreage^
Developed
Area
Acreage^,
Factor
+ (Acres)
The factor is trend related as it includes future residential
density and family size (in time t), which might change from the base
year (t'). However, since the time difference between t and t1 will
probably never be greater than 10 years, most likely there will not
be a big change in the variables in question for the t1—> t time period.
Population of the area of analysis in times t and t1 must be
obtained. Because the area of analysis will probably not be the
same as a census area, population data must be apportioned somehow.
(Population-^ will probably be provided in the facility plan documentation,
but is assumed not to be available for this discussion.) There are a
number of ways to apportion population data and a number of places
E-15
-------
where the information could come from. Census block and
tract data could be apportioned to the area of analysis, if
available, to estimate population in time t'. The past annual
population change rate could then be used to estimate population
in time t. Traffic zone data may also be used for this purpose,
as may watershed-based pouplation projections, if available.
Because there are a lot of possibilities, it is not possible
to cover them all here. Local and regional planners deal with this
issue a lot; they should be able to provide assistance in popula-
tion apportionment.
A numercial example using the factor should clarify the
procedure and tie loose ends together. Population in time t1 is
estimated to be 10,000. It was apportioned from census tract and
block data. Developed land associated with this population is
about 2,500 acres (@ 4 housing units per acre). Population^ is
estimated to be 20,000; this figure comes from the facility plan.
Family sizet is thought to be 3.1 persons per family (actually
3.1 persons per housing unit), which is the latest estimate of family
size available from the regional planning agency. It is not expected
to change between now and t. Housing unit densityt is expected to be
about 4.5 units per acre based on recent development trends,
Substituting the above into the formula for developed area
acreage^ gives:
Developed
Area
Acreage^
2,500 acres +
(20,000 - 10,000) people
3.1 people
unit
4.5 units
acre
= 2,500 acres + 717 acres
= 3,217 acres
This value would be subtracted from the known quantity of (area
of analysis) minus (vacant undevelopable area) to obtain Vacant
Developable.
A totally different tack could be taken to come up with estimates
of vacant developable land. This method is also based on estimating
developed land in some past or current time t1, but is more direct. It
simply entails determining how much land is being subdivided annually
in the area of analysis and multiplying it by the time period (t - t1).
E-16
-------
For example, say that 23 acres of land are being developed annually
in the area of analysis -- on average, and this rate is expected to
remain pretty constant. For an 8 year period, this means that 184
acres will be developed. This has to be added to the already
developed acreage to get total developed land in time t. The subtractions
mentioned above would be required to get Vacant Developable.
Vacant Developable is used in the Residential and two other "core"
predictive equations. It is a sensitive variable.
3. Vacant Undevelopable: vacant and undevelopable acreage in the area of
analysis in time t. The variable appears in the same 3 core equations as
Vacant Developable. It is a very sensitive variable, so care should be
used in estimating its value.
Vacant Undevelopable is a semi-fixed physical variable. It should
include flood plains, steep slopes, quarries, and publically owned land
that won't be permanently developed. Datum can be obtained from planning
agency studies of natural constraints (or physical features) or from U.S.
Soil Service county soil surveys. The data may have to be scaled off of
maps.
4. Zoned Residential: acres of land zoned for residential use (all types)
in the area of analysis in year t. It is a sensitive variable but does not
appear in any of the core equations.
The variable can be estimated from a current (time t1) zoning map, or
maps if more than one municipality is involved. The datum probably will
have to be scaled off from the map(s). The municipality comprehensive
plan(s) should be checked to determine if there is more land on the plan
designated as residential than currently zoned for. If so, the situation
should be discussed with local or regional planners to determine what the
probable trend in zoning will be between t1 and t.
5. Zoned Office; acres of land in the area of analysis zoned for office,
including professional office development in year t. The variable apepars
in two core predictive equations, and it is a sensitive variable. It is
obtained in the same manner as Zoned Residential.
6. Zoned Industrial: acres of land in the area of analysis zoned for
industrial development in year t. It appears in one core equation; it
is not a sensitive variable. Again, it is obtained in the same way as Zoned
Residential.
7. Onsite Restrictions: a coded variable to indicate what governmental
restrictions are expected to be placed on on-lot sewage disposal between
the time t and t + 10. This variable is not sensitive nor does it appear
in a core equation.
E-17
-------
The datum may be available in the 208 Plan (Areawide Wastewater
Management Plan). The director of a local or regional sewer authority
may hazard a guess as to what the future will bring in the way of
on-lot disposal restrictions. Knowledge of past restriction policies
may help to ascertain what the future will bring. For example, if
the policy is or was that all on-lot disposal is prohibited if housing
densities reach 2 per acre, and future densities are expected to be
greater, then the variable would be coded as 4.
A problem may arise if more than one policy is expected in an
area of analysis. The coding rule that should be used in this and
other ambiguous cases is: use the most restrictive category applicable
to most of the area of analysis most of the time.Also, severity of
restriction has precedence over time but not area. Suppose, for example,
that most ( >50%) of an area of analysis will not be allowed to have
on-lot disposal for six of the 10 years. This is a straightforward case,
and Onsite Restrictions is coded 4 (see Table 2-5). If the disposal
restriction will only apply for two years on most of the land, the
variable is still coded 4 (Restriction Years goes from 6 to 2, however.)
Finally, if most of the area will have no restrictions during the ten
year period, Onsite Restriction is coded 0 - even if it had more severe
restrictions for some of the area of analysis some or all of the time.
8. Restrictions Years: the number of years of time between time t and
t + 10 that the coded Onsite Restrictions condition is expected to exist.
The variable is not sensitive nor is it in one of the core equations.
It is obtained in the same manner as Onsite Restrictions.
9. County Interchanges: the number of limited access interchanges (regardless
of functional classification of the roadway) expected in t + 5 in the
county containing most of the area of analysis. (If the area of analysis
is split evenly among two or more counties, then all counties should be
used. Other county level variables must reflect the multi-county situa-
tion also, however.)
County Interchanges is not in a core land use predictive equation.
It is not a sensitive variable either.
Datum for the variable will usually come from the regional
planning agency. If the area of analysis is in an urbanized transportation
planning area having a so-called 3C planning process, the figure will be
available as part of the long-range programming element. If not, the state
transportation agency should be consulted. Municipality comprehensive
plans, or transportation planning programs, may also be of help in some
areas. The City Engineer will be of help in those cases.
10. Limited Access: the number of limited access interchanges (regardless
of functional classification of the roadway) expected in t + 5 in the
area of analysis.
E-18
-------
This variable is not sensitive, but does appear in the
Commercial core equation. It is obtained in similar fashion to
County Interchanges.
11. Transit Stops: the number of bus and commuter rail transit stops expected
in t + 5 in the area of analysis. The variable is not sensitive nor does
it appear in a core land use equation.
Transit Stops may be obtained from a regional planning agency, or
other agency, doing the 3C Plan because a transit element is part of
the planning process. Probably, however, the bus or rail company will
have to be contacted. This was the way the datum was obtained most often
in the case study phase of GEMLUP.
It will be difficult to get an estimate of transit stops for the
future. If the transit company or authority does not know what will
happen in t + 5, use data for the present adjusted by what the recent
trend in line closings has been in the area of analysis. If there is
no service in the area now, try to find out if any is being seriously
discussed. If existing housing densities are low, probably there would
not be any service by t + 5 even if a transit company is operating in near-
by locales.
12. County Growth: the percent change in county population projected for
the time period t to t + 10. This variable is sensitive but is not used
in a core equation.
Datum for the county containing most of the area of analysis is used.
If the area is fairly evenly split between two or more counties, all
should be used. (If percentage changes are used, a weighted average must
be derived. See a standard statistical textboook for information on how
to compute a weighted average. The problem can be avoided by adding
together the actual population estimates for the two years and computing
the percentage change for the total.)
County population estimates are available from many sources. The
best might be the regional planning agency, but county planning departments
and the state community affairs department are other possibilities. The
estimates should be those "officially recognized" for air quality maintenance
planning purposes or 208 planning. (If there are differences between these
two estimates, take an average.)
Probably decennial or quinquennial projections will have to be
manipulated to obtain population estimates for t and t + 10. A linear
interpolation between years is probably adequate, unless population levels
are expected to change dramatically in the counties. In those instances,
the projections should be plotted and a smooth curve drawn to connect the
point estimates. Year t and t + 10 projections can then be read off the curve,
E-19
-------
13. Future Population: projected SMSA population for the year t + 10.
If the area of analysis is not in a SMSA, use projected county popula-
tion (see County Growth). The variable is very sensitive, but does not
appear in any core equation.
Probably the best source of Future Population is OBERS. The
information is also often available from a regional planning agency.
If OBERS is used, the data will have to be interpolated to obtain year
t + 10 projections (see County Growth).
14. Population Growth: the projected percentage change in SMSA population
between t and t + 10. Population Growth is obviously related to Future
Population:
Population _ Future Population. lr, - SMSA Population.
Growth " l+lu r
SMSA Populationt
Probably the best place to obtain an estimate for SMSA Population^
is OBERS. Thus both input variables for Population Growth could come
from the same place, which has some advantages, such as: fewer sources
to locate, fewer assumptions that must be considered, and compatability
of inputs and format.
The variable is not sensitive, but does appear in one core equation.
15. Office Vacancy: percent of office space located in the area of analysis
that is expected to be vacant in year t. Office Vacancy is in one care land
use equation. It is not a sensitive variable.
Obtaining an estimate of relative office vacancy for a future time
period will be very difficult, unless there is a sophicated economic
planning effort underway in the region. Since this is uncommon, probably
the best that can be done is call a local commercial realator, or building
manager, and determine what the approximate vacancy rate is now and what
the recent trend has been. Because office vacancy is negatively correlated
with median income, perhaps a current vacancy rate could be combined with an
income proportionality factor to estimate time t office vacancy. However,
the effort to do so may not be worth the increase in precision of the
variable estimate.
16. Future Houses: the projected number of housing units expected in the SMSA
in time t+10. If the area of analysis is not in an SMSA, the projected
number of housing units expected in the county in t+10 will be used. (If
more than one county is involved, proceed as discussed in 12.^ County
Growth. If more than one SMSA is involved, an unlikely situation, use the
SMSA that most of the area of analysis is in.)
E-20
-------
Future Houses is not a sensitive variable and does not appear
in a core predictive equation.
Datum for the variable may be in a regional or local housing
plan. If so, interpolation will most certainly have to be done
for the year t+10. The datum could also be obtained by using the
OBERS projection of SMSA population and multiplying it by a housing
unit per population factor (the reciprocal of average family size).
(See also the discussion of family size in 2., Vacant Developable.)
17. Median Price: the projected median price of one acre of vacant, "raw"
residential land in the area of analysis in year t. This is a sensitive
variable and it appears in two core land use equations, including
Residential.
It will be very difficult to get an estimate for Median Price.
Possible sources are asking a local developer what the going price of
raw land is now and how he thinks it will change by time t. A current
land price estimate could be "inflated" by a fixed annual percentage also,
if the user can buy the argument that land values go along with (but do
not rise faster than or lag behind) the inflation rate. The projected
inflation rate itself could come from a general U.S. projection put out
by the Treasury Department or other federal government agency.
There is another possible solution. Median Price is not used in
isolation in any predictive equation -- it is part of the Land Cost
variable, which is Median Price divided by Median Income (see below).
If the user wants to make the assumption that these two items vary at
the same rate, which is almost identical to assuming that Median Price
increases as does inflation, then a current or past Median Price to
Median Income ratio could be used for time t also. In that case, the
user would first get Median Income estimate for a specific year and
then get a Median Price estimate for the same year. The ratio, which
was 0.537 on average and varied between 0.058 and 2.546 for the 40 GEMLUP
cases, would then be used as Land Cost for time t. If land cost is
changing differently than income in the area of analysis, however, this
approach should not be used.
18. Future Income: the projected median family income in the SMSA in year t+10,
Unrelated individuals are excluded from this definition. The variable is
not sensitive and does not appear in a core land use equation.
Future Income estimates may be available from a regional planning
agency. It can be obtained from OBERS's per capita income estimate by
multiplying the estimate by expected average family size (persons per
family). This is the preferred method. Interpolation of data for the
correct year (t+10) probably will have to be done.
E-21
-------
If the area of analysis is not in an SMSA, then a county-level
income estimate must be obtained. This may be available from a regional
or State planning agency.
19. Future Employment: projected total employment in the SMSA in t+10. If
the area of analysis is not in a SMSA use projected county employment
instead. The variable is a very sensitive one; it is used in one core
land use equation.
OBERS is a good source of total employment estimates. Interpolation
will probably have to be done.
20. Future Medicals: projected SMSA hospital employment in year t+10. The
variable does not appear in any core equation and is not particularly
sensitive.
Future Medicals should probably be obtained from OBERS estimates
of future employment or population. This is done by developing a hospital
workers-to-total employment or hospital workers-to-population ratio from
the most recent Census data, and applying it to OBERS projections. The
second ratio was relatively constant for GEMLUP case studies, and is
preferred to the first. Whichever ratio is used, the user must assume
that it will not change over time.
21. CBD Distance: the distance in air miles (straight-line) from the approximate
centroid of the area of analysis to the approximate centroid of the nearest
central business district in a city of 100,000 population or more in year
t. This variable is not a sensitive one, nor is it used in a core equation.
The nearest 100,000 city in time t must first be identified and its
CBD located. The distance between centroids is then scaled off of a map
with a known scale. A state highway map is generally suitable for this
purpose.
22. Airport Distance: the distance in air miles (straight-line) between the
approximate centroids of the area of analysis and nearest commercial
airport in year t. A commercial airport is an airport with at least one
in and one out regularly scheduled flight available to the public on a
fee basis. Sight-seeing flights are not included in this definition.
After identifying such an airport for time period t, scale off the
distance between centroids on a road map. (Perhaps the regional planning
agency should be checked to determine if a new commercial airport will be
built before time t that may be closer to the area of analysis than the
existing airport.)
The variable is in one core equation (manufacturing), but is not a
sensitive quantity.
E-22
-------
23. Track: miles of railroad track in the area of analysis in time t.
Note that parallel tracks are counted as two tracks. Track is a very
sensitive variable and appears in three of the five core land use
equations.
Track is easily obtained from USGS maps. A local or regional planning
agency should be consulted to determine what changes have occured since
the USGS maps were published and what might be expected by the year t.
Note that abandoned rail lines that go someplace (i.e., are connected to
a mainline somewhere) are counted even if they are not currently used.
24. Area of Analysis: the area of analysis is the legal service area of the
wastewater treatment project being studied. It includes the drainage basin
for the stream receiving wastes from the project plus other areas connected
to the treatment plant by mechanical means. Unit of the variable is acres.
Datum will be available in the facility plan provided by the project
developer or owner.
The variable appears in all land use predictive equations, both core
and secondary. It is a very sensitive variable and must be accurate.
25. Sewered Land: land area in acres within 5,000 linear feet of the major
project interceptor sewer(s), if any, in the area of analysis in year t.
The 5,000 feet distance is omni-directional, but is only within the area
of analysis. The variable is sensitive and appears in one core equation.
The location and extend of all new interceptor sewers will be in the
facility plan. If maps provided with the plan are large enough, Sewered
Land may be scaled directly from them. If not, the new sewers will have
to be plotted onto adequate maps and Sewered Land scaled off of those
new maps.
26. Interceptors: running length in miles of the major project interceptor
sewer(s), if any, going through relatively undeveloped land in the
area of analysis in year t. Relatively undeveloped is defined to be
less than one housing unit per acre, and the relatively undeveloped area
should be at least 1000 feet deep (i.e., from the interceptor sewer). In
other words, if only land adjacent to the pipe is relatively undeveloped
and the rest is subdivided and developed, there is no relatively undeveloped
land for purposes of computing Interceptors.
Datum is obtained by plotting the interceptor sewer onto aerial
photographs or land use maps and scaling off the running length of pipe
going through <1 housing unit/acre land. This variable is most easily
obtained in conjunction with Sewered Land.
Interceptors is in one core equation, but is not a sensitive variable.
E-23
-------
27. Collection Capacity: hydraulic design capacity of an interceptor
sewer(s) major project in million gallons per day (MGD) in time t.
If the interceptor sewer(s) is(are) phased, the design capacity
may be that for the year when the last phase is completed; however,
all phasing must occur before t+5. If phasing of interceptor sewers
occurs within the area of analysis beyond t+5, the GEMLUP model
should not be used.
The datum will be available in the facility plan. The variable
is sensitive and it appears in three core equations, including Residential.
If Collection Capacity is not available in the facility plan for
some reason, it can be estimated by summing hydraulic flow of the major
project interceptor sewer(s), using nomographs of Manning's formula
(based on pipe size and slope; see Metcalf and Eddy, Inc. Wastewater
Engineering. New York: 'McGraw Hill Book Company, 1972.). If slope is
not known, average values for hydraulic flow based on pipe diameter
alone can be used (see Facilities Requirements Branch. Guidelines for
1976 Update of Needs for Municipal Wastewater Facilities" Washington,
D.C.: U.S. Environmental Protection Agency,1976).
28. Peak Flow: anticipated peak flow in the interceptor sewer(s) major project
in MGD in year t. If phasing occurs, Peak Flow is for the last phase.
Again, all phasing must be complete by t+5.
Peak Flow should be available in the facility plan. If it is not,
it can be estimated by scaling known average flow by a population-dependent
peaking factor. Such a factor appears as Figure E-2.
The variable is sensitive and it appears in four core equations,
including Residential.
29. Treatment Capacity; hydraulic design capacity of the wastewater treatment
plant in MGD in year t. This information will be available in the facility
plan. The variable is not sensitive nor does it appear in a core equation.
30. Population Served: the population served by the major project facility in
year 5. This information will be in the facility plan. The variable is
not sensitive. It does not appear in any core land use equation.
31. Project Cost: the total cost of major project construction (in thousands
of dollars). This value will be in the facility plan. The variable is not
sensitive, and it is not in a core equation.
32. Federal Funds: the federal share of major project cost (in thousands of
dollars).TTTis non-sensitive, non-core equation variable will be part of
the facility plan.
E-24
-------
Figure E-2
RATIO OF MINIMUM AND PEAK FLOWS TO AVERAGE DAILY FLOWS
40
60
80
100
Population in Thousands
E-25
-------
33. Phasing: a nominal variable that indicates if multi-year construction
phasing of interceptor sewers will occur in the area of analysis. This
variable will be in the facility plan. It is not sensitive and is not
in a core equation.
34. SMSA Area: area of the SMSA in year t in square miles. If the area of
analysis is not in a SMSA, county area should be used. Data for the
variable is found in the County and City Data Book. A regional planning
agency would have the information also.
SMSA Area appears in one core equation, but it is not a particularly
sensitive variable.
35. Tract Area: area of the census tracts for which socio-economic data are
obtained. Units are in square miles. The data can be obtained by scaling
the census tract maps, but it may be available from the regional planning
agency also.
If the area of analysis is not tracted, then Tract Area becomes area
of the municipality or other civil division used for socio-economic data
gathering. Municipality area is found in the County and City Data Book.
Tract Area is a very sensitive variable because of its use in many
GEMLUP socio-economic variables (as a divisor). It is in three core
equations.
36. Dwelling Units: number of dwelling units in the census tracts in year t,
rounded off to the nearest 100 units. (I.e., if the value is 5,327 then
53 is used.) The variable is sensitive and appears in two core land use
equations.
Values for the variable for a current or past time period (t1) can be
obtained from census tract data. The time t1 variable can be updated to
time t using a population ratio of populationt/population^i. The number
of dwelling units in t would therefore be obtained thusly:
Dwelling = Dwelling Popu1ationt
Unitst UnitSf ' Populationt-
Since both t1 variables are known, only Population^ is required. A
regional or city planning agency would be the source for census tract
population projections.
This method assumes that the variable of interest (Dwelling Units)
is linearily related to population. If that is not true, then a cor-
rection to the ratio has to be made. This can simply be done by
applying another ratio to the population ratio to rectify the non-
linearity. For example, suppose that average family size has.changed
radically between t' and t. A family size ratio should then be used
E-26
-------
along with the population ratio to estimate Dwelling Units.
Dwelling . Dwelling . P°Pu1ationt . Family Sizet.
Unitst Units^1 Population^ Family Sizet
An example should elucidate both problem and solution. Input data
are as follows; DU^ would not be known in practice.
Time t]_ t. Units
DU 100 (150) Dwelling Units
Family Size 3.00 2.67 People/Dwelling Units
Population 300 400 People
If just the population ratio was used to estimate DUt, Dwelling Units
would be underestimated by about 17 units.
DU. » 100 . 400 = 133
r 300
Using the correction factor of family size produces the correct answer.
DU. = 100 . 400 . 3.0CL = 150
T 300
The principle of using a correction factor to adjust for non-
linearity in the dependent variable-to-population relation is used in
computing many of the variables that follow.
37. Current Houses: number of dwelling (housing) units in the SMSA in time
t: The variable is not sensitive and does not appear in any core equa-
tion. Current Houses should be obtained in the same manner as the pre-
vious variable. Time t1 values would come from the County and City
Data Book, while t values would be obtained from regional planning
agency projections (its housing element).
38. School Kids: the number of people less than 15 years of age living
1n the census tracts in year t. The variable is sensitive and
appears in one core equation.
The variable is obtained in the same manner as variable 36. Time
t variables will have to come from regional or local planning agency
E-27
-------
projections. The population ratio factor described earlier she-Id be used,
supplemented by other ratios of family size or the relative proportion of
children in the total population (or both) changes.
For example, if the age structure of the population is changing, a for-
mula for computing school kids might be:
School _ School Populationt Relative Proportion of Kidst
Kids. " Kids.,
Population^' Relative Proportion of Kidst.
Using the same numerical example used in 36 above, say the proportion of
school aged children (0-14) in the total population changed from 0.367
in t' to 0.263 in t. This means that the population is aging. Multiplying
these proportions by total population results in 110 Schools Kids in t' and
105 in t. Plugging numbers into the formula gives the correct answer.
School 110 400 0.263 = 105 children
' 0.367
The alert reader may question why the time period indices of the cor-
rection factors are different in 36 and 38. In 36 the time t1 correction
factor is in the numerator, while in 38 it is in the denominator. The
reason is that time indices must cancel as well as variable units. In 36,
the Family Size factor has units of:
Populationf Dwelling Unitt
Dwelling Unitf Population^
(The numerator is multiplied by the reciprocal of its denominator.) The
units now cancel throughout the original equation.
In 38, Relative Proportion of Kids has units of:
School Kidst Populationf
Populationt ' School Kidst1
These units also now cancel with the original equation.
39. Vacant Houses: the percentage of vacant available dwelling (housing)
units in the census tracts in time t. The percentage is coded as a
decimal; i.e., 10% is coded as 0.1 and 17.3% is coded as 0.173. Datum
for time t1 would come from census tract information. This could be
used for time t, or an appropriate planning agency could be contacted
for a more recent estimate. While the variable does not appear in a
core equation, it is a very sensitive variable in those equations
where it does appear.
E-28
-------
40. Nonmobility: the percentage of families in time t living in the same
house as they did in t-5. The geographical area is the census tracts,
and the datum comes from Census of Population and Housing: Census
Tract books. The variable appears in the residential core equation and
is sensitive. Percentages are coded as decimal equivalents.
The most recent census value could be used as a surrogate for
Nonmobility. If a comprehensive housing element planning effort exists
in the region, the variable could probably be estimated from it. The
variable will most likely fall between 45 and 65 percent and is thought
to be relatively constant in the short run.
41. Median Income: the median income of families living in the county or
counties containing most of the area of analysis. The time index is
year t.
A time t' value can be obtained from the County and City Data Book.
A more recent estimate will probably be available from a socio-economic
profile done by the regional planning agency.
Median Income is only used in conjunction with Median Price (17).
See the discussion of that variable for more information. Median Income
is sensitive and appears in the residential and one other core land use
equation.
42. Current Income: the median family income of people living in the SMSA
containing most of the area of analysis (in time t). If the area is
not located in a SMSA, county data are used, which means that Median
Income = Current Income. The variable is not sensitive nor does it
appear in a core equation; in fact, it is only used in disaggregating
residential predictions into the single- and multiple-family categories.
If the area of analysis is in a SMSA, then time t data can come
from OBERS by multiplying the per capita income estimate by average
family size. Because OBERS data are in terms of some specified base
year dollars, Current Income should be updated to account for infla-
tion.
Another source of information can be a regional planning agency
socio-economic profile.
43. Poverty: the percentage of total families in the census tracts with
income below the H. E. W..poverty level in time t. The percentage is
coded as a decimal. The variable is not sensitive and is not used in
a core equation.
Time t1 values can be gotten from census tract data. Because the
variable does not change rapidly, the time t1 figure should be adequate
for predictive, purposes. If a more recent estimate is desired, then a
socio-economic profile done by a public planning agency should be con-
sulted.
E-29
-------
44. Index One: the Consumer Price Index (CPI) for year t. The Index is
published for selected major cities in CPI-Detailed Report (Bureau of
Labor Statistics, U.S. Department of Labor).Use the value for the
nearest major city.
The variable is only used in combination with Index Two. An
estimate for time t if it is in the future will not be available.
The recommended way to proceed is to use the current CPI for whichever
Index (One or Two) applies to the later year and go backward in time
for the other index. There is no need to update the two indices with
an inflation rate, because each index would be changed by the same per-
centage and the net result would be the original difference.
The variable is not sensitive and is not used in any core land
use equation.
45. Index Two: the CPJ for the year of federal funding; the year should
be less than time t. The variable is not sensitive nor is it. in a
core equation. It should be obtained in the manner in 44, above.
46. Government: the total amount of county government expenditures in
millions of dollars in year t. (Round off $7,761,000 to 7.76, for
instance). Use the county or counties containing most of the area
of analysis. The variable appears in one core land use equation; it
is not sensitive.
It will be hard to obtain data for this variable unless the
county projects expenditures in advance. Perhaps the regional plan-
ning agency might be doing fiscal analyses of future expenditures.
Probably the only way to get an estimate is to obtain recent county expen-
diture audits, determine what the yearly trend is by plotting values
on a graph, and project ahead using the same rate-of-change. This
trend analysis will only be approximately valid, but the estimate
should be accurate enough for the use to which it is put.
47. Current Employment: total SMSA employment in year t. If the area of
analysis is not in a SMSA, use county employment data. The variable
appears in one core equation, but is not sensitive.
SMSA total employment projections are available in OBERS. Some
interperlation between OBERS projection years will have to be done,
however. If the area is not in a SMSA, perhaps the information can
come from a county-level socio-economic profile maintained by a county
or regional planning agency (such information will be available in a
3C transportation plan, for instance). If not, use past census estimates
projected ahead by a percentage change number used by state or regional
economic/industrial development organizations.
48. Unemployment: the percentage of unemployed workers in. the census tracts
in year t. (Percentage is coded as a decimal). This variable is sensitive,
but does not appear in a core equation.
E-30
-------
An estimate for Unemployment might be contained in a socio-economic
profile done by a county or regional planning agency. Since it is a
relative variable; the current value could be used by assuming no change.
A proportion could also be established between unemployment percentages
for the area of analysis and some (larger) municipality for which a time
t estimate is available. For example, say the census tracts containing
most of the area of analyses has an unemployment rate of 6% in 1980,
the latest census year. The county's unemployment rate was 10% in 1980.
Year t is 1984, and the state has projected an unemployment rate of 8%
for the county in 1984. An estimate of Unemployment, then is:
Unemployment = Unemployment^
County Unemployment
County
(0.06)
0.08
0.10
0.048
Unemployment,
0.05
49. Office Employment: the number of employed office workers in the census
tracts in time t. The variable is sensitive, but is not in a core
equation.
Values for time t will be difficult to come by. Office Employment
should be obtained for a current or recent time period t1 and adjusted
by a population ratio (see 36).
50. Manufacturing Workers: the number of manufacturing workers living in
the census tracts in time t. This variable appears in two core equa-
tions, including residential. It is also a sensitive variable. Prob-
ably the easiest way to obtain a value for this variable is to get the
most recent U. S. Census estimate and apply a population ratio to it
(see 36).
51. Current Medicals: the number of hospital workers living in the SMSA--
if there is one—in year t. If the area is not in a SMSA, use county
figures. Time t1 values come from the U. S. Census, and a projection
to time t can be made using a population ratio procedure. See item
36. The variable is not sensitive nor is it in a core equation.
52. Drivers: the number of people living in the county who drive to work,
either as a 'driver or a passenger, in year t. The value should be
rounded and coded to the nearest one hundred people; i.e., 1,371 is
coded as 14. The variable appears in the residential core equation;
it is not sensitive.
E-31
-------
Drivers may be available in a regional 3C plan. Probably a
population ratio will have to be used to update U.S. Census infor-
mation. Exactness is not required, as evidenced by rounding off of
the datum.
E - 32
-------
III. Sensitivity Analysis
The impact that a variable has on a particular equation depends
upon its contribution to total variance explained in the dependent vari-
able by the equation. I call this impact "sensitivity." A sensitivity
analysis was performed for all predictive equations.* The analysis took
the form of answering the question: "What is the percentage change in the
dependent variable when all other independent variables are set to their
mean value?" The higher the resultant change, the greater a particular
variable's impact is on the resultant, and the greater is its sensitivity.
There are other ways to do sensitivity analyses, and one alternative method
is used for the Residential equation for illustrative purposes.
The user should be careful to obtain good data for sensitive variables
because of their importance. Assuming that carefulness is related to time
spent in obtaining data and that user time is a scarce resource, a user should
spend most of his constrained time on important variables and relatively less
on the others.
Note that this sensitivity analysis has nothing to do with model vali-
dation, which is discussed in Section II.B of the Report. In particular, this
sensitivity analysis should not be confused with coefficient stability analy-
sis described in that section, which it superficially resembles.
Results of the sensitivity analyses follow. If the percentage change
value has a negative sign, it means that a positive change in the independent
variable causes a negative change in the dependent variable (an inverse re-
lationship, in other words).
Residential Equation
The Residential equation is most sensitive to changes in Nonmobility,
and is fairly sensitive to all other variables except Driver Density. The
*In English units only; results of the analysis in metric units would probably
be different.
E - 33
-------
percent change in Residential (in terms of housing units per acre) due
to a 10.0% increase in each independent variable, with all other indepen-
dent variables held constant at their mean value, is:
Nonmobility -9.7
Vacant Land -6.1
% Collection Reserve 3.3
Land Cost 3.3
Manufacturing Density 2.6
Driver Density 1.0
Another way to look at variable sensitivity is to set the variable at
its extremes (low, usually zero, and high, the maximum value recorded
during GEMLUP field data gathering effort) and see what happens to the de-
pendent variable. This was done for the Residential equation; results are
presented in Table E-5.
Office-Professional
The Office-Professional equation is most sensitive to changes in Office
Zoning, which may be a surprise to jaundiced observers of the local zoning
process in this Country. No other variable has nearly as important an im-
pact on Office-Professional as does Zoning. The next most important variable,
Railroads (or the density of railroad track in the area of analysis), is odd.
Its importance is undoubtedly due to its being correlated to a factor that is
also highly correlated with office space, perhaps size or age of city (them-
selves highly correlated), a high order of central place functions, or indus-
trialization.
The percentage change in Office-Professional due to a 10.0% increase in
each independent variable, with all other independent variables held con-
stant at their mean value, is:
Office Zoning 7.9
Railroads 2.1
House Density 1.5
Peak Flow 1.4
Population Growth 0.8
Industrial Zoning -0.7
E - 34
-------
Table E-5
PERCENTAGE CHANGE IN RESIDENTIAL DUE
TO EACH INDEPENDENT VARIABLE SET TO ITS EXTREME
VALUES & ALL OTHER INDEPENDENT VARIABLES
SET AT THEIR MEAN VALUE
Independent Range of Dependent Difference in the
Variable Variable (%) Range (%}
Land Cost 33.0 - 233.5 200.5
% Collection Reserve 65.8 - 201.1 135.3
Nonmobility 63.3 - 197.1 133.8
Vacant Land 63.8 - 160.5 96.7
Manufacturing Density 73.6 - 157.9 73.6
Driver Density 90.4 - 161.4 71.0
As can be seen by comparing data in the text with Table E-5, a variable's
rank order with respect to sensitivity varies from one method to another.
The reason for this is that the relationship between mean and extreme values
varies greatly among independent variables. Variables widely scattered
around their mean or having a skewed distribution will have a potentially
large impact on the dependent variable if they are misspecified.
Recreation
Recreation, in units of acres of land per area of analysis, is most
sensitive to Vacant Houses (negative relation) and the ubiquitous Railroads
The percentage change in Recreation because of a 10.0% increase in each
independent variable, with all other variables held constant at their mean
value, is:
Railroads 6.1
Vacant Houses -6.1
Office Density 4.6
County Growth 4.0
Industrial Zoning -1.7
Treatment Capacity 1.3
E - 35
-------
Manufacturing
Manufacturing, in units of floor area per area of analysis, is most
sensitive to Office Zoning and Vacant Land. The second relation is logical
but the first is not. Apparently there is a factor that affects both in-
dustrial and office land uses (but not industrial zoning). This factor
could probably be identified by analyzing causal relations found in Volume
I of GEMLUP (EPA-450/3-78-014a), but this has not been done because of time
constraints.
The percentage change in Manufacturing due to a 10.0% increase in each
independent variable with all other independent variables held constant at
their mean value is:
Office Zoning 7.0
Vacant Land -5.2
Railroads 2.7
Office Vacancy 1.6
Manufacturing Density 1.3
Airport Distance 1.2
Wholesale
Wholesale land use, in terms of floor area per area of analysis is most
sensitive to Employee Ratio and Income. Of all the independent variables
investigated so far, they are the only ones that produce a greater than one-
to-one relative change in a dependent variable. Compared with preceeding
equations, Wholesale is relatively sensitive to all its independent vari-
ables, but Employee Ratio and Income will have a much greater impact than any
of the others.
The percentage change in Wholesale due to a 10.0% positive change in
each independent variable, with all other variables held constant at their
mean value, is: Employee Ratio 47.6
Income -21.7
Unemployment 7.1
Driver Density 6.0
Office Vacancy 5.7
Office Workers 5.3
E - 36
-------
Highways
Highways, in units of non-expressway lane miles per area of analysis,
is the most sensitive to Railroads, which has to be a surrogate for some
other variable. The percentage change in Highways because of a 10.0% in-
crease in each independent variable, with all other variables held constant at
their mean value, is:
Railroads 3.9
Land Cost 2.9
Government -2.0
Collection Reserve 1.3
Interchange Density 1.1
Education
Education, in terms of floor area per area of analysis, is most sen-
sitive to Sewer Service. The percentage change in Education due to a 10.0%
increase in each independent variable, with all other variables held constant,
is:
Sewer Service 3.7
Land Cost 2.0
On-site Restrictions 2.0
County Growth -0.9
Sewer Costs -0.5
Poverty *
* Less than a -0.05% change.
Commercial Equation
The Commercial equation is equally sensitive to changes in Kid Density,
Vacant Land, and Sewer Service. The percentage increase in Commercial (in
terms of floor area per area of analysis) due to a 10.0% change in each
independent variable, with all other variables held constant (at their mean
value) is:
Kid Density -3.9
Vacant Land -3.7
Sewer Service 3.6
Employment Growth -0.9
Interchanges 0.8
% Collection Reserve -0.6
E - 37
-------
Other
Other land uses include transient lodgings (hotels, motels, and rooming
houses), churches and other religious structures, cultural activities (museums,
theaters, auditoria, and libraries), and indoor recreational facilities
(separate from schools). Other does not include hospitals, prisons, or
colleges. The units of Other are in square feet of floor area per area of
analysis. The percentage change in Other because of a 10.0% increase in each
independent variable, with all other variables held constant, is:
Income -116.2
•Vacant Houses - 18.2
Railroads 13.9
County Growth 6.6
Residential Zoning 5.4
Interchange Density -2.1
The highly sensitive negative relationship between Income and Other is
puzzling at first glance because, it is commonly said, cultural activities
are directly related to the income and educational level of residents. An
explanation for the inverse relation can be obtained from the zero-order
correlation coefficient. Other is significantly positively correlated with
population density, which implies that cultural activities are most often
found in highly developed areas. This is certainly true of many museums and
auditoria, although theaters are more evenly dispersed throughout a metropoli-
tan area (espcially movie theaters). Motels and hotels are also found down-
town, particularly in large cities. Next, there is a negative, significant
correlation between population density and per capita SMSA income, which
Income measures. These two factors taken together would result in the negative
relationship between Income and Other that is observed above.
E - 38
-------
Summary
The preceding analyses of sensitivity were used in developing a list
of very sensitive and sensitive variables. Variables not on the list were
labeled insensitive or not sensitive variables. The designations correspond
to those used in Section III of this Appendix.
Postscript
Peter Guldberg, principal author of the main report reviewed this
Appendix and made an important point regarding this section. He says that
really nothing can be said about causal relations in predictive equations
because individual regression coefficients are biased due to interaction
among the independent variables. This can be seen by changing the order
that variables enter a regression equation; variable coefficients change as
their order of entry changes. Thus, comments concerning the relative causal
effect of a variable vis-a-vis others cannot be made.
I make these types of comments in Office-Professional when discussing
the Office Zoning variable and in Other when describing the Income variable.
Please disregard any implications of causality in these discussions. What
I say about relative sensitivity (as defined here) of the variables and
the pragmatic reasons for their relative sensitivity still hold, however,
but only within the abstract context of the regression equation itself. See
Volume I for an extended discussion of "real-world" causal implications of
GEMLUP equations.
E - 39
-------
IV. Distribution of Predicted Land Uses Within the Area of Analysis
The quantities of land uses predicted to occur within ten years of
increased sewage capacity apply to the area of analysis as a whole. There
is no spatial resolution to the predictions, in other words. Nor was any
ever intended.
Spatial disaggregation was not attempted for three reasons: (1) it
would have complicated the modeling exercise immensely, (2) it would have
ignored local knowledge of an area, and (3) it was unnecessary from
an air pollution emissions estimation viewpoint.
Our desire to provide relatively simple, easy-to-use predictive equations
would have been obviated if disaggregated predictions were made. Doing this
would require that some sort of internal grid system be used, that land uses
by grid cell be obtained, and that cell-specific predictive equations be
generated. The user would then have to grid off his area of analysis and
use nine predictive equations for each cell. The user would also have to
allocate existing land uses to each cell to ascertain what new development
was predicted to occur within a particular cell.
We wanted not only to avoid doing cell-by-cell analysis but to use
local knowledge of existing and proposed development. Local and regional
planners know where things are and where and what kind of development is
about to go in next. They also know what the zoning is and where streets
and highways will be constructed or improved. It was felt that this local
knowledge would be more accurate in internally allocating development than
using a numerical technique based on case study information. Not only
that, but if all development followed a deterministic pattern there would
E - 40
-------
not be a need for planning at all. Phrased another way, the community-
represented by its planning agency—should be able to decide where develop-
ment should go given the fact that development is coming. Zoning is
obviously based on that premise, and we wanted to work within that framework.
Finally, for most air quality impact analyses, it is immaterial where
development is within a region as only total regional emissions are used in
the impact assessment. This is particularly true for most land use types,
whose individual emissions are so low that they are lumped together as
"area source" emissions anyway. In other words, air pollution modelers
do not need individual source emission estimates for their diffusion work
for most land uses in a community. They only require specific estimates for
large "stationary sources" emitting approximately 100 tons or more of
pollution per year; these sources are large ("heavy") industrial firms,
power plants, refineries, and some extractive manufacturers. And it should
be noted that specific estimates may not even be required for these large
sources for all pollutants. Hydrocarbon emissions, for instance, are usually
only needed for an urbanized area on SMSA, because the analytical methods
used to convert hydrocarbon emissions into ambient ozone concentrations are
aspatial. In fact, the air quality analyses are usually done on a multi-
county area larger than an SMSA (called an AQCR, or Air Quality Control
Region).
Distribution of predicted land uses within the area of analysis, then,
is done by the user using personal knowledge of the area, data in the
comprehensive plan, the zoning map, and information on development trends
E - 41
-------
obtained from personal contacts, subdivision proposals and building permits.
It is intuitive to some extent, obviously, but should be suitable for most
projection purposes. If disaggregated land use predictions are required
for some reason, the user will have to use another methodology; GEMLUP cannot
help.
E - 42
-------
V. Disaggregating Residential and Commercial Predictions
Worksheet 4 contains formula to disaggregate residential and commercial
land use predictions. (Worksheet 3 contains row entries for them, but data
for the rows come from Worksheet 4.) As explained in Section I.A.Z.b., the
disaggregation equations were developed using logit analysis, where percentage
breakdowns of the classes of residential and commercial land uses were fitted
to a logistic S-shaped curve. The equations were transformed into regression
equations for the user.
The disaggregation equations require that the following variables be
defined. (If a variable is derived from input variables—those a user must
obtain data for—the input variables are shown in parentheses.)
Residential Disaggregation Equation
Kid Density (school kids, dwelling units)
Income Growth (future income, current income)
Sewer Service (Sewered land, area of analysis)
Office Zoning (zoned office, area of analysis)
Nonmobility
Hospital Growth (future medicals, current medicals)
Housing Growth (future houses, current houses)
Poverty
Railroads (track, area of analysis)
Peak Flow
Commercial Disaggregation Equation
Office Vacancy
Onsite Restrictions
Treatment Capacity
Kid Density (school kids, dwelling units)
Restriction Years
Phasing
Transit Stops
CBD Distance
Government
Vacant Land (vacant developable, area of analysis)
E - 43
-------
Of the 24 different input variables needed to use the disaggregation equations,
11 are used in core equations and only 4 are used in the residential equation.
Consequently, if a user opts for the simplified approach discussed in
Section I of this Appendix, he or she would not be able to disaggregate
residential and commercial predictions without getting more information.
This defeats the intent of simplification.
How the user gets around this problem should depend upon what he or she
does with the predictions. For instance, if the user is primarily interested
in using GEMLUP results to estimate area emissions, the commercial disaggre-
gation need not be done since the per floor area emission rate is the same
for all commercial categories. A residential disaggregation should be done in
this case, however, because per unit area emission rates do vary for residential
categories for some, but not all, pollutants.
Two logical ways to disaggregate commercial and residential predictions
are (1) use the existing local breakdown for the future estimates, or (2) use
the breakdown found in the GEMLUP case studies. The existing local breakdown
for the residential categories can be obtained from U.S. Census information.
The existing local breakdown for commercial land uses may be available from
a local or regional planning agency, but probably will not be. In that case,
the only option is to use GEMLUP case study information.
The residential and commercial breakdowns found in the GEMLUP cases
appear in Table E-6. They could be entered directly onto Worksheet 3
(as decimal equivalents) and multiplied by the predicted quantities on
the left to come up with estimates of disaggregated land use. Again, local
knowledge should be used to determine if the GEMLUP breakdowns are reasonably
E - 44
-------
consistent with the existing development pattern of the area being inves-
tigated. The case studies obviously had a lot of single family detached
housing and small (usually strip commercial and neighborhood shopping center)
commercial development. If the user's area of analysis has a greatly different
development pattern, the disaggregated equations should be used instead of
case study breakdown.
Table E-6
RESIDENTIAL AND COMMERCIAL BREAKDOWN
FOR GEMLUP CASE STUDIES
Disaggregated
Category
Residential
Single Family
Two-Family
Multi-Family
Commercial
Large
Medium
Small
Mean Value
(Percent]
100.0
83.2
4.8
12.0
100.0
10.7
10.5
78.8
Range of Values
(Percent)
33.7 - 95.1
0.0 - 27.0
0.0 - 39.3
0.0 - 54.5
0.0 - 49.2
29.6 -100.0
E - 45
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT
3fpRAT-4%0/3-78-014b
2.
3. RECIPIENT'S ACCESSION-NO.
5. REPORT DA
'ORT DATE
May, 1978
4. TITLE AND SUBTITLE
Growth Effects of Major Land Use Projects (Wastewater
Facilities); Volume II: Summary, Predictive Equati or|s PERFORMING ORGANIZATION CODE
and Worksheets
7. AUTHOR(S)
Peter H. Guldberg, Ralph B. D'Agostino
8. PERFORMING ORGANIZATION REPORT NO.
C-921
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Walden Division of Abcor, Inc.
850 Main Street
Wilmington, MA 01887
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
68-02-2594
12. SPONSORING AGENCY NAME ANO ADDRESS
Environmental Protection Agency
Office of Air Quality Planning and Standards
Strategies and Air Standards Division MD-12
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Final
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT
Growth Effects of Major Land Use Projects in a research program whose goal is to
develop methodologies to predict the total air pollutant emissions resulting from
the construction and operation of major land use projects. Emissions are quanti-
fied from the major project, from land use induced by the major project, from
secondary activity occurring off-site (e.g., electrical generating stations), and
from motor vehicle traffic associated with both the major project and its induced
land uses.
This report documents the development of predictive equations for the induced land
use from wastewater major projects. The predictive equations are included in an
impact assessment procedure that estimates the total air pollutant emissions as-
sociated with induced development from a wastewater major project. This procedure
is formalized in a set of easy-to-use worksheets, which serve as an operational
tool for environmental engineers and planners.
17.
KEY WORDS AND DOCUMENT ANALYSIS
a.
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS C. COS AT I Field/Group
Land Use
Planning
Sewage Treatment
Plants
Path Analysis
Causal Analysis
Secondary Effects
Induced Land Use
18. DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
207
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
------- |