A REVIEW OF LAND USE M3036 AS DEVELGPED
F091 ‘ERAE‘ESEz’QR‘FATIQi’é STUDIES -
'FHEER. USE AND IMPLECAWONS T0 URMIN PLANMNG

‘fiwsis Far We {Emma 0f M. U. P.
MiG-{EGAN STATE UNWERSETY
Thomas Maw-(am Heilbron

$96.5

   
     

THESIS

  

LIBRARY

Michigan State
Universny

 

____——
”—
__.—-—-

R0051} USE QRLY

 

ABSTRACT

A REVIEW OF LAND USE MODELS AS
DEVELOPED FOR TRANSPORTATION STUDIES -
THEIR USE AND IMPLICATIONS TO URBAN PLANNING

by Thomas Howard Heilbron

The determination of future land use distribution
is a primary problem to both urban and transportation
planners. Urban transportation studies during the past
decade have sought to increase the efficiency and effec-
tivness of land use forecasts in order mainly to provide
a basis for estimating future traffic demand. A major
advancement toward this end was the development of mathe-
matical models designed to forecast and allocate urban
growth to small geographical areas. As individual trans-
portation studies were initiated, the design of these
models became more refined and their application
broadened. Chapter One traces this development in a
general way, pointing out the significant advances and
limitations of each study's approach to determining future
land use distribution.

From the analysis of past and present land use
models, an investigation is made of the uses and im-
plications of these models to the field of urban planning.
Later, the limitations of forecasting with the use of

mathematical models and problems of forecasting in general

Thomas Howard Heilbron

are enumerated in order to point out the research that is
still needed in this field of study.

From the analysis contained herein, it is concluded
that there is good reason why the future can not be
predicted far enough ahead to be particularly significant.
This uncertainty about the future is simply the corollary
of the freedom of human beings to make their own decisions,
and until a means is devised to predict human behavior
fairly accurately, it may never be possible to predict
the future state of an urban area. However, the results
need not be perfect in order for land use models to be
useful. If they can be used to measure the relative
merits and effects of policy decisions and development
alternatives, then a significant tool for planning will

be available.

A REVIEW OF LAND USE MODELS AS
DEVELOPED FOR TRANSPORTATION STUDIES -

THEIR USE AND IMPLICATIONS TO URBAN PLANNING

BY

Thomas Howard Heilbron

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER IN URBAN PLANNING

School of Urban Planning and Landscape Architecture

1965

ACKNOWLEDGMENTS

The author wishes to express his sincere apprecia-
tion to Professor Charles W. Barr whose advice and
supervision aided in the preparation of this study.

Acknowledgment is also given to Professors Stewart
D. Marquis, Sanford S. Farness, and Joseph M. Prochaska
for their review of this thesis and their valuable
comments.

Appreciation is also extended to my wife who aided

immeasurably in the preparation of this document.

ii

TABLE OF CONTENTS

ABSTRACT
A CKNOWLEDGWNTS o o o o o o o o o o o o 0
INTRODUCTION . o . . . o o o o . o o o o 0

Chapter
I HISTORICAL REVIEW OF MAJOR
TRANSPORTATION STUDIES . .

II USES AND IMPLICATIONS OF
LAND USE MODELS . . . o . .

Uses of Models in Planning
Implications of the Use of
Models to Planning

III LIMITATIONS OF EXISTING LAND
USE MODELS O O O O O O O 0

Conceptual Problems
Unknown Variables and

Informational Gaps
Model Structure

IV CONCLUSIONS AND RECOMMENDATIONS
FOR FURTHER RESEARCH . . o

BIBLIOGRAPIH O O O O O O O O O O O O O O O

Page

ii

'37

51

68

76

INTRODUCTION

Early attempts to alleviate transportation problems in
urban areas either dealt only with existing traffic demands
and made no forecasts, or expanded all traffic movements
uniformly according to past trends. A major breakthrough
in this limited methodology occured about 1952 when urban
travel patterns were related to the intensity of use and the
spatial separation of different types of land use. 1‘ It
followed that in order to determine future transportation
needs, it was necessary to determine future land use
distribution.

Urban planners have always been concerned with making
accurate land use plans, but the almost sudden need and vast
amounts of available money for urban transportation studies
accelerated the demand for better methods of determining
future land use needs and arrangements, especially in the
larger metropolitan areas. To fulfill this demand, urban
and transportation planners tried to find regularities in
urban development processes which would lend themselves to
mathematical relationships that could be helpful in the
preparation and testing of development proposals. These
mathematical relationships are commonly called models.

A model can be defined in broad terms as an imitation or
a design for action. There are several types of models in
common usage today. They are: abstract, symbolic, physical

and mathematical. 2 An abstract model is perhaps the most

2
versatile and most commonly used of the four. It can be as
simple as a representation of an atom for instructional
purposes, or be as complicated and ill-defined as a
thought process. The latter is used unconsciously every—
day by everyone. Simple acts such as tying a shoe or
more complicated processes such as weighing alternatives
in an emergency situation are both examples of abstract

models. A symbolic model consists of replacing actual

'T - plysical lengthfl: locations, and angles with mathematics.

A common use would be the prediction of eclipses by
scientists. The physical models are used in engineering
and science and usually represent, to scale, a structural
project or an experiment. A physical or scale model of a
drainage basin, for example, is very useful in predicting
potential flooding as a result of precipitation somewhere
in the drainage basin.

Mathematical models, with which this thesis is basically
concerned, consist of formulations expressing observed or
hypothesized relationships in mathematical terminology. A
formula such as X + Y = Z is actually a simple mathematical
model. However, the mathematical models discussed in this
thesis are much more complex and are usually solved with
the aid of an electronic computer. Mathematical models have
many uses, but the models discussed throughout this thesis
were all designed to describe the processes of urban develop-
ment so that future land use arrangements of urban areas

could be predicted. It is for this reason that they are

3
often referred to as land use models. All future use of
the word "model" in this paper will mean mathematical land
use models, unless otherwise stated.

At this point one might wonder why models are developed.
In answer to this common query the following paragraphs
briefly point out the more obvious advantages of models.

In the first place urban areas can not be designed in a
straightforward and controlled manner as a bridge or section
off highway. Urban areas are constantly changing as a
result of a myraid of unrelated decisions by individuals,
business firms, and government. To cope with the ever
broadening problem, the urban planner's field of interest
has expanded to include many disciplines other than the
traditional physical sciences. Because of the vast subject
matter contained in these various disciplines, and the
interrelated effects of changes in one or the other, the
planner (urban or transportation) needs help in assimilating
and evaluating data. Herein lies the basic advantage of
models. That is, a model is a device for systemizing
knowledge in a useful and managable form to increase its
user's awareness. The ramifications of development
proposals are so widespread that awareness of effects can
not be left solely to intuition. 3

Another major advantage of a model is that while being
formulated, or upon completion, it provides a frame of
reference for consideration of the problem. In other words,

before one can adequately mathematize a phenomenon such as

h

the urban environment or a scientific experiment, he must
have a good understanding 95 the contributing factors and
their interrelationships.

Immediately one can say of the above that it is not
known how all factors behave or are related. This is true,
but it also brings up another advantage of using models --
they suggest information gaps. Thus, when a model is being
formulated it brings out the need for more information, and
it also makes for a better understanding of the factors that
are used.

Another advantage, which is somewhat related to the
above, is that a model, when tested, may show its deficiencies.
These deficiencies could be either in the data or mathe-
matical expressions themselves. This suggests that
perhaps models and data inputs can be more and more refined.

Once a problem or situation is defined in symbolic
language, the various factors and relationships can be
manipulated at great speeds resulting in a high degree of
efficiency. Efficiency is achieved in another way by the
use of models. They provide a cheap way to predict the
outcome of a particular set of circumstances, and one does
not need to have the actual operation or physical facilities
at his disposal to see what will happen. h

The more technical bodies of knowledge, such as
astronomy, physics, and mathematics have long since
recognized the value of formulizing concepts, and as a

result, have made many significant advances in their fields.

5

It is relatively recent that models have been used in
disciplines such as economics, psychology, and now urban
planning. The advantages of models which were briefly
discussed above reflect why there has been such enthusiasm.
and interest in the development of mathematical land use
models. Throughout the following chapters, models will

be analyzed as to their uses, implications to urban

planning, limitations, and needs.

l.

2.

3.

FOOTNOTES

 

Robert B. Mitchell, Urban Traffic, A Function of
Land Use (New York: Columbia University Press,
I9SH), p. 2.

 

w. T. Bross, Design for Decision (New York:
Macmillan Co., 1953). pp. 162—167.

 

Arthur T. Bow and Ernest Jurkat, "The Economic
Forces Shaping Land Use Patterns," Journal of

the American Institute of Planners,*XXV (May,

1959), 79.

Bross, loc.cit., p. 169.

 

CHAPTER I
HISTORICAL REVIEW OF MAJOR TRANSPORTATION STUDIES

During the last ten years transportation studies have
been initiated in scores of urban areas ranging in size
from a few thousand persons to nearly fifteen million. As
will be brought out in this chapter, the larger of these
studies have become progressively more sophisticated in
their approach to determining future land use distribution.
The review and analysis emphasizing this fact include the
following features for each study:

1. General organization - including when the study

started, population, and area covered.

2. Objectives of the studies - in terms of land use
relationships and their function in transporta-
tion planning.

3. Study methodology - including assumptions and
concepts used, general procedure or functioning
of the models, and types of land uses projected.

h. Particular significance of the study - in regard
to its contribution to the development of the
transportation planning process.

In selecting studies to be analyzed, data on all major

urban areas that have completed or initiated transportation
studies were grouped in a chronological order as to their

initiation date. Upon review of this data, it was found

8
that several studies were very similar and others had not
progressed enough to afford sufficient analysis and were
consequently omitted from this chapter. As a result, the
studies which are included in the following pages are felt
to represent a non-repetitious chronological development of

land use projection methods as used by major transportation

studies.

Detroit Metropolitan Area Transportation Study

In 1953 the Michigan State Highway Department, county
of Wayne and City of Detroit established a policy committee
which was responsible for the conduct of the Transportation
Study. J. D. Carroll, Jr. was the director of the study
which covered an area of 709 square miles and contained
just under three million people in 1953. S

The purpose of the Detroit study was to develop traffic
projections and highway improvement plans. 6 To accomplish
this purpose a significant assumption was made as to the
importance of land use in determining traffic generation, and
was stated in the following manner: "Separation of places
in space and the need for travel to overcome this separation
is best reflected in the usage of land. Specification of
land uses and their arrangement in space are basic causal
factors creating traffic." 7 Thus, this linkage of land
use to traffic generation marked a new era of transportation
planning.

The planning process developed by the Detroit study

9
was fairly simple, but unique for transportation planning at
that time. The first step in this process was a population
and economic forecast of the total study area. The second
task was a forecast 6f the total amount of land for various
uses which was based on the population and economic fore-
casts. The third step was the distribution of the future
population and land use. 8 This latter step, for the
Detroit study, was greatly simplified by the fact that
the City Planning Commission had prepared a master plan
in 1951, and the Metropolitan Area Regional Planning
Commission also had a plan for the metropolitan area
outside the city of Detroit. Carroll, the study director,
recognized the significance of these plans when he stated:
"The land use distribution must reflect local master plans
and must account for redevelopment plans. Close coopera-
tion with the public planning agency must be maintained as
the future land use distribution is the most critical step
in the transportation planning process, because the future
origins and destinations are based on it."

Having future land use plans available was fortunate,
but in order to adequately plan and stage transportation
facilities it was necessary for the staff to determine
when land would be developed. Therefore, priority ratings
were applied to available land, designating when it would
probably be developed. 9 However, there was no mention
as to how these ratings were established.

Several important assumptions and procedures were

lO

utilized in this study. The assignment of trip generation
rates for land uses was done manually on a zone by zone
basis - observing past trends and patterns. In this way the
established tendency for trips per acre to decline with
distance from central business district was taken care of
automatically. On the surface it appears that this assump-
tion anticipated no change from existing conditions.

Another important assumption was that improvements in
traffic facilities would be evenly distributed in any future
plan. Therefore, the friction between zones would remain
relatively the same in 1980 as in 1953. This assumption
seems rather poorly founded since so much is being done
with future accessibility in the more recent transportation
studies.

In summary, it can be said that the Detroit study was
a pioneer in several areas of transportation planning. It
was the first large comprehensive study that crossed many
political subdivision boundaries.‘ It was also the first
study to use electronic data processing to simulate traffic
flows from land use projections, and the first to use land
use projections as a basis for total metropolitan traffic

flow analysis. 10

Washiggton D. C. - Mass Transportation Study
NatiOnal Capital Rggion

The Washington D. C. study was organized in 1955 under
authorization of the 8hth Congress which stated that the

11
National Capital Planning Commission and the National
Capital Regional Planning Council were to jointly conduct
a survey of the present and future mass transportation
needs of the region. Kenneth M. Hoover directed the
study which contained approximately two million people.

The planning process utilized in this study was very
similar to the Detroit method in that future land use
distribution was the result of intuitive or professional
judgement based on the traditional planning approach. More
specifically, the future population and economic growth in
the region was estimated for 1980 and then distributed to
local areas on the basis of existing local trends. Two
basic assumptions were made in regard to this distribution.
First, it was assumed "... that local governments would not
be willing or able to significantly change this trend in
growth. It was further assumed that the local governments
would be able to exercise only a limited influence on the
pattern of development and growth." 11 These two assump-
tions were significant because they precluded any variation
in present trends and the introduction of community policies
as to how growth should occur. Of course the federal
government is the greatest factor affecting growth in
Washington, and it is not subject to local control.
However, this excuse is not available in other cities
where similar conservative or short-sighted assumptions are
made.

The projections in the Washington D. C. study were

12

defended as being realistic on the grounds that the approach
took into account the variety and diversity among communities
within the region in a way that could not be done through
reliance on quantitative relationships. 12

The unique feature of the Washington study which
contributed to the evolution and development of transpor-
tation planning was the application of future full-scale
simulated traffic flow methods to several radically different
tranSportation system alternatives. This study, unlike
the Detroit study which developed only highway plans,
analyzed several systems which varied mainly in respect to
the mixtures of modes of transportation used in each. One
was rail dominant, another bus dominant, and a third
automobile dominant. However, all three simulations were
calculated from a single land use projection. Thus, the
Washington study fully demonstrated the value of designing
and analyzing a range of major transportation alternatives,
even though the validity of the actual projection appears
limited by the overly simple assumptions about future
land use. 13

Chicago Area Transportation Stqu

The Chicago study was organized in 1955 by a special
agency which functioned in an advisory capacity to the
participating governmental units. J. D. Carroll, Jr.
directed the study which covered 1,326 square miles and
included 6.8 million people. In

13
The basic objectives of this study were enlarged over
previous studies in that it proposed to devise a long
range plan for needed highways and for mass transportation
facilities, to carefully design information so that it could
be re—examined and used for updating, and to develop a
future land use projection on which to base future travel
patterns.

Unlike the previous two studies which used conven-
tional planning techniques to determine future land use
distribution, the Chicago study staff used a more deter-
ministic method which could be classified as a semi-
mechanical model. One of the main reasons for this approach
was that there was no metropolitan plan available, nor a
planning agency that could prepare one. Because very few
metropolitan areas have master plans, transportation planners
try to develop a means (mechanical or otherwise) of
determining future land use patterns. Therefore, the Chicago
Area Transportation Study staff made land use projections
which were based on the assumption that the forces operating
presently will determine largely the land development
pattern expected by 1980. 15 The C.A.T.S. report makes
no pretentions, however, as to the projections being a
plan. The study staff clearly distinguishes between the
two in the following statement. "A land use forecast based
upon historical trends and current relationships, is not
the same as a land use plan. A land use forecast makes

estimates as to how land is most likely to be used, e.g.,

11L

that the trend toward low density, single family home
construction will continue and will have a spatial
pattern quite similar to the pattern of densities observed
in 1956. It does not consider whether this is good or
bad or ought to be encouraged. It is an attempt to state
what the future will be like if present trends continue." 16

Following this line of thinking, the Chicago study staff
found significant relationships between several variables
and the distance from the central business district. In
other words, their basic thesis was that the distribution
of activities and the intensity of land usage are highly
related to accessibility. It is important to note that the
accessibility of a parcel of land in the Chicago study
was measured as being directly proportional to the distance
of that parcel from the central business district. Another
relationship showed the peak population density was moving
outward at a rate of 1.6 miles per decade. These and
similarly complex relationships were used to project
future land usage based on past trends.

The distribution of each type of land use was determined
differently. "For residential land use, the capacity of
each zone was determined initially on the basis of existing
densities and vacant land. The change in population
density, population density per developed acre, and non-
white population were analyzed as a function of time and
distance from the central business district. Certain

zones were analyzed separately because of special

15
considerations. Commercial forecasts were obtained by
considering both planned projects and a fixed ratio of
commercial acreage to residential acreage. Public land
uses were broken into local uses such as grade schools and
into non—local used such as federal or state buildings. A
fixed ratio of local public land use acreage to residential
acreage was used, but no orderly relationship for non-
local usage was found. The acreage in streets remained
constant, except in zones where more than 10% of the
land was undeveloped. In these cases, a percentage of
land was set aside for streets, depending on the acres
of undeveloped land and the type of land use in the zone.
Open spaces such as parks and golf courses were forecasted
on the basis of past trends, community goals, and likely
available acreage." 17

In summary, this technique was based on the assumption
that urban growth is essentially a movement of densities
in a concentric pattern, declining as distance from the
center increases. As a model, the Chicago method does very
little in describing how the processes which shape the urban
18

pattern actually work. It neglects to envision changes
in trends and, most important, that the transportation
systems will change the projected growth pattern.

Carroll recognized this latter fact, but rationalized his
method in the following way. "The fact the accessibility

will be altered may not be significant on a regional scale.

Besides there are legal, geographic, financial and other

l6

factors that influence land use, so changes in the trans-
portation network may not have significant effect on the
ultimate development." 19

The major importance of C.A.T.S. lies in the fact that
it was a pioneer effort in describing urban growth in
mathematical language. It was also the first to relate
development potential to accessibility in a comprehensive

way.

Hansen Model

Following the Chicago Area Transportation Study there
was a rash of model developing and initiating of trans-
portation studies. The Pittsburg Area Transportation
Study was one such major study, but it followed C.A.T.S.
procedure so closely it contributed little to the evolution
of model development. However, Walter Hansen hypothesized
a model for distributing urban population growth (residential
only) based on growth data of the Washington, D. C.
metropolitan area between l9h8 and 1955. Hansen's model
is termed a potential or gravity model because it is an
extrapolation of the gravity concept of physical science.
The basic thesis of this model is that forecasted population
can be distributed to small zones using the accessibility
of these zones to other land uses and the amount of
vacant land in the zone. The accessibility in this model
differs extremely from that used in the Chicago study

where accessibility was assumed to be directly proportional

17
to the distance from the C- B. D. Here accessibility is

calculated by the following formula:

32 s3 sn
0 e e 4’ “'- x

 

 

 

A1:

where A accessibility

size of the activity ie. employment in a zone
time for travel between zones

exponent - the value of this is very difficult

to determine and creates many problems.

>480)

Hansen calculated the accessibility for all zones and by
applying a couple of constants and simple ratios of vacant
land to total forecasted growth could estimate future
residential development in a zone. 20
The model obviously has several weaknesses which
are acknowledged by Hansen, but it also has several impor-
tant advantages over previous types of projections. First,
it has greater behavioral content because how families
choose their locations is related to access to employment
and to the locational opportunities (vacant land) available
in various parts of the city. This measure of access also
takes into account the economies and diseconomies of a
transportation system to a limited degree. Secondly, this
method takes into account the distributive effects of the
transportation system and the Spatial array of employment.
Third, the output of this model is not as preconceived as
the Chicago model which was based on what someone thought

the area should look like. 21

18
The results of this research were fairly accurate
considering accessibility and vacant land were the only
two variables used. The significance of this study is that
it paved the way for research to isolate more factors or
variables and measure their effect on land development.
It also further emphasized the importance of accessibility

on land develOpment.

Baltimore-Washington Interregional Study

 

The Baltimore-Washington study was a continuation of
previous studies conducted independently by the two
metropolitan areas. It was organized as a joint project in
1957 under the technical advisorship of Alan M. Voorhees,
and the region studies contained nearly four million
people.

The major objective of the study was to project and
analyze future distribution patterns of population, em-
ployment, and automobile traffic growth in the Baltimore-
Washington region. While previous studies of these cities
had dealt with this objective, several changes were made
in this one. This study was to treat the Baltimore-
Washington Area, not as two independent areas, but as one
interrelated region. The previous studies relied upon
more or less conventional techniques for estimating popu-
lation and employment distribution. This study introduced
more recently developed methods of population and employ-

ment projections for small areas in the region. It also

19
considered alternative patterns of regional growth,
lending equal emphasis in analysis to: a) potential form
and extent of land development; and b) requirements for
transportation systems. 22

The actual process employed by this study expanded
the gravity model concepts developed by Hansen, and used
separate models for distributing population growth and
manufacturing employment. Where Hansen used only accessi-
bility and vacant land as variables in his model, the
Baltimore-Washington study staff added a "K" factor which
was to take into account socio-economic factors.

More specifically, the population growth model included
three factors: 1) available population capacity; 2) ac-
cessibility to employment; and 3) a group of socio—economic
and development factors reflecting the character of the
land, utilities, and established developments. The growth
index representing the actual or predicted growth for any
particular zone was determined by the following equation:

Growth Index (G) = Capacity (C) x Accessibility (A) x
other growth factors (K)

This study area was divided into zones, and the "K" factor
for each zone was determined empirically from data on

growth between l9h8 and 1957 by the following formula: 23

G actual growth (19h8-l957)

 

 

CxA theoretical growth

The unique feature of this study was in the analysis

20

of the "K" factors for each zone. Several component factors
were found to be contributing to the "K" values. They
were grouped into four categories, and in each category a
number of individual factors were listed as contributing
to the major category. These individual factors were
assigned weights and their sum made up a total weight for
each major group. The values of the four major groups were
then multiplied together, and the result was said to be
equivalent to the "K" value for each zone.

The four major groups and their individual factors
are listed below to give an indication of the analysis
procedure used at the time of this study.

1. Utilities - deficiencies in water or sanitary
sewers reduced the K factor.

2. Land and Development Factors - low land cost,
attractive lend, poor subdivision
practice, policies of land owners.

3. Community Factors - prestige, established communities,
quality of local government services.

A. Special Factors - not really known. 2h
After these factors were indentified deductively or sub-
jectively, they were analyzed in regards to their effect
on future growth. At times they were changed because of
significant variables pertinent to a zone. This was an
important aSpect because it showed a tendency to get away
from blindly following past trends.

In order to establish population capacities, and
therefore make the model work, future densities had to be

assumed. By varying the density groupings within the

21

region it was possible to derive alternative growth patterns
for analysis.

Manufacturing land use was projected by using the same
model with different variables comprising the "K" factor
and accessibility to population was used instead of
accessibility to employment. Commercial employment as
well as other land uses were more or less distributed in
direct proportion to the population distribution.

The significance of the Baltimore-Washington study
rests in the fact that it further developed mathematical
models to include a variety of factors, and that it briefly
examined alternative land use patterns as well as alternative
transportation systems. The interesting conclusion of the
study was: "TranSportation is not the dominant or controlling
factor in shaping our cities. With the mobility provided
by the automobile, the urban dweller has for all practical
purposes been freed of distance limitations in his choice of
a place to live." 25 This conclusion may not be applicable
to other areas because it seems that accessibility was

almost equal in all parts of this study area.

Los Angeles Regional Transportation Study

The L.A.R.T.S. was organized in 1959 under the general
coordination of the California Division of Highways and
was directed by Edward T. Telford. The study, the largest
in the world at that time, contained 7.5 million people in

a 9,000 square mile area. This study has yet to be completed.

22
Therefore, the following analysis is based on reports
which describe what is hoped to be accomplished.

The purpose of the study is to develop the travel
patterns of people and goods for use in determining the
present and future transportation needs of the five county
area. These patterns will be related to present and
projected land use, population, employment and other socio-
economic factors affecting traffic. Furthermore, the study
staff proposes to test alternative freeway systems, different
concepts of mass transit usage, alternative land use plans,
and various rates of economic growth. 26 Besides this
objective, which is obviously more inclusive that that of
any previous study's, the L.A.R.T.S. staff further stressed
the following: "The land use models developed in this study
will provide powerful tools to the local communities for
testing and evaluating alternatives of existing and future
land use and transportation plans. The study in itself
will not make any recommendations as to street and highway
improvements, freeway locations, or public mass tranSporta-
tion. Its prime responsibility is to provide information
which the various agencies and their legislative bodies,
in the study area, can use in their planning and decision
making." 27 This is a new approach used in a transpor-
tation study and may have significant effects in allowing
for local policies and objectives to be implemented to
a greater degree than previously.

The general procedure in regards to land use distribution

23
is very similar to that used in the Baltimore-Washington
study, in that it proposes a gravity type land use model
which will distribute population change throughout the
area. Industrial and other uses will each be distributed
by separate models. The population-land use model is
based on competition for projected growth. This compe—
tition is based on a multiple correlation analysis of many
factors which include water, sewers, highways, public
transit, land suitability, average vacant land cost per
acre, average topo conditions, accessibility to jobs, and
types of dwelling units. These factors are the same as
those usually used throughout the United States in similar
studies. However, their individual weights or importance
will vary with city and regional differences. Therefore,
each city which undertakes this type of study will have to
determine the relative importance of the factors which
affect their growth.

An observation made of this study and previous studies
brings out a problem.much discussed today; that is, how
can you determine the interrelated effects of land use and
transportation systems? Previous models used existing
accessibility to determine future land use distribution,
but the L.A.R.T.S. staff proposes to lay out a future highway
net to determine future travel times, then use them in the
model to predict land use distribution. Neither method
seems completely satisfactory because they are basing land

use distribution on transportation networks instead of

2A
vice versa or some combination of the two.

The Los Angeles Regional Transportation Study's
unique features are the broader objectives and the utili-
zation of the services of various planning jurisdictions
within the study area. Questionnaires were sent out
requesting various information on land development and
land use plans. Therefore, the final result of the Study
will represent a culmination of many ideas and will give

the local areas a basis on which to make decisions.

Penn-Jersey Transportation Study,

 

The Penn-Jersey study was initiated in 1959 by joint
agreement of the highway departments of Pennsylvania and
New Jersey and the City of Philadelphia. The study was
designed to include an area which contained over u.3
million persons. By far the most complex to date, this
study, directed by Henry Fagin, produced several innovations
to tranSportation planning.

The scope of the objective exemplifies the deviation
from previous studies to some degree. It was stated that
the study intended not only for the recommended transpor-
tation system to provide convenience and economy of travel,
but also that its influence on the development of the area
should tend toward facilitating a desired pattern of re-
gional development. Stress is laid on the development of
alternative patterns, both of possible tranSportation systems

and of the future regional development likely to be

25
associated with each system. 28

Previously, transportation plans have been based on
assumptions as to the future distribution of people and
activities. Some consisted of informed judgment while others
consisted of projecting past trends. In both these methods
one single projection of land use had been made, and various
transportation schemes were evaluated in relation to serving
that pattern. The Penn-Jersey staff developed a regional
growth model whereby land use projections can be made for
varying sets of assumptions and policies. These include
influences of land use change on alternative transportation
schemes.

Another basic difference in the Penn-Jersey study is
that an effort has been made to identify the behavioral
process which results in the distribution of new households
and employment opportunities. Previous models have inferred
that the factors influencing new development will be similar
to: those of the past. However, the elements which generate
new development tend to be only partially identified, and
the assumptions identifying these variables are rarely
tested. Because of these reasons and the inherent stress
on quantitative relationships, economic models have been
introduced by the Penn-Jersey staff to explain the "maxi-
mizing" behavior of individuals and business firms within
a region. 29 The basic thesis behind the model is that
if there is a consistency, it is not in the pattern per se,

but in economic and social behavior of the individuals who

26
locate their houses and business places.

The behavioral content of the model depends upon the
propensity of household locators to maximize the difference
between what they annually allocate to the purchase of their
residences (and all its attributes i.e. site size, public
services, age of area, views etc.) and what they actually
have to pay exclusive of site cost. This difference is
defined as the households rent paying ability, and it is
unique for each type of household, the particular parcel
of land associated with that household, and the area where
the site is located. The exact form of the model is a
linear program.

The interaction between this behavioral model and the
traffic model in the Penn-Jersey scheme depends upon how
"accessibility" as a variable influences household
location decisions. The accessibility of each household
to other land uses is defined in terms of trip making costs
instead of distance or time. Each location in the area
has a particular accessibility or cost of traveling from
that location to other locations. Thus, a resident must
consider this cost along with the housing and site costs.
Therefore, decisions to spend money for housing as
compared with other items in the family budget.

The actual regional growth model is made up of several
sub-models, each of which is complex and beyond scope of
this paper to explain. Let it suffice to generally list

four basic inputs to the regional model:

l.
2.

3.

27
Initial state of the system - all present artifacts.
Projections of external factors i.e. population
and economy.
Alternative policies in regard to transpor-
tation systems or alternative levels of service.
This may include public policies as to land
development, zoning, taxation and provision
of services.
Assumed relationships of the model itself which
consist of functions and parameters. (The
functions describe the form of relationships -
the way the real world is assumed to work. The
parameters are constants of the functions which
define the quantitative nature of the relation-

ShipSe) 30

The function of the model is to estimate changes in

the present condition which result from changed factors

in the above and other variables. The model begins with

existing l96O conditions as the given, and proceeds to

simulate changes in five year iterations. The pattern at

the end of the first five years is the given for the next

iteration.

The Penn-Jersey staff states that the mere passage

of time does not in itself lead to changes in patterns.

These changes result from some prior cause which ought to

be identifiable. If changes in pattern were related only to

the passage of time, or growth of population, then they

28

would be the same no matter what public policies were
pursued. (This is what Washington assumed.) The Penn-
Jersey staff proposes that the tastes and desires of
people for residential Space and the economic behavior
of individuals and commercial locators remains unchanged
in its essential character. However, changes in trip
making behavior would result from 1) changes in the en—
vironment which affect the opportunities open to locators
(trip makers); 2) from changes in the capacity of people
to satisfy their desires; or 3) from changes in the composi-
tion of the population seeking such satisfaction. The
changes in environment result from growth processes and
public policy decisions, and changes in the capacity of
people result from changes in income. 31

The foregoing discussion brings out some of the
significant features of the Penn-Jersey Study. It is the
most complex study of behavior and the variables affecting
it, and the inclusion of policy variations and their re-
sulting effects on development patterns produced significant
advances in this phase of planning. The involvement in
specifying the interrelated effects of land use and trans-
portation on one another is also a significant step for-
ward in transportation planning. The process of "watching
the region grow" by five year periods was advanced more

thoroughly than previously.

29
Tri-State Transportation Study_

The Tri-State study was initiated by the governors of
Connecticut, New Jersey and New York in August of 1961.
The study area was, for the first time, to envelop three
states totaling 11,000 square miles and containing lh.7
million people. Roger H. Gilman, the director of the
study, was given a twofold objective of formulating a
long range transportation plan and program that will
promote the orderly and desirable development of the
region during the next two decades, and secondly, to plan
solutions for the most critical immediate problems. In

regards to land use projections the Prosgectus stated that

 

a specific method had not been settled upon as yet, but
that they intended to study previous methods and then
devise a model suitable for the Tri-State region. 32

The general procedure to be followed encompassed most
of the advancements of transportation planning to date.
In the first place, the inter-relationship between the
region's land development and transportation system is
to be studied so that recommendations can be made for
achieving the optimum development of the region. Further-
more, it was proposed to study the area in a series of
incremental growth periods - much like the Penn-Jersey
study. The local areas are to work out their own plans
and policies of a subregional nature in order to involve

local objectives. 33

30
Southeastern Wisconsin Regional Transportation Study

This study was initiated in 1962 and is set up as a
project of the S.E.W.R.P.C. (Southeastern Wisconsin Regional
Planning Commission) which is charged with the reSponsi-
bility of developing a regional master plan. Sanford
Farness originally directed this study whose organization
is ideal for the true integration of land use and transpor-
tation planning. The primary objective of the agency is
to develop alternative regional transportation plans and
development patterns. In order to do this, the region
will be described in a systems approach and land use
models will be developed.

The land use model, like the Penn-Jersey model, is
a process model. "It will be used to test the effectiveness
of various ways of achieving desired land use plans. The
model contains variables such as water and sewer location,
zoning, etc., which are under governmental control. By
adjusting these controlled variables in the model and
observing the resulting forecasts of land use, the best
way to achieve a desired plan can be determined.

"The land use model is divided into five nearly inde-
pendent simulations, each one dealing with a different
category of land use, namely, residential, industrial,
commercial, agricultural, and Special (such as park and
open spaces). The forecasts of agricultural and special
land uses in each zone is made in an 'ad hoc' fashion.

These forecasts are then used with other data to forecast

31
residential, industrial, and commercial land uses.

"As an example, consider the operation of the residential
model. Initially, households are divided into categories
according to structure, market value, and owner versus
renter status. The actual categories are made on the basis
of empirical data so that all the households in a given
category live mostly in a few categories of housing.

In addition, zones are categorized according to taxes,
water, sewers, and accessibility to work, shopping, and
population. Then the desirability of a particular housing
type for a zone type is determined from empirical data by
assumming that current housing types are located in desir-
able zones.

"Growth occurs then as follows. First, a yearly turnover
rate is empirically determined for each category of house-
holds. These 'turning-over' households move into vacant
housing types desirable to their category of household.

If more than one zone contains such vacant housing, the
zone with the highest accessibility to facilities is moved
into by the turning over household. Land developers

and builders produce new housing types. The quantity of
each housing type is determined by past sales. The
location is determined by finding zones that are desirable
for these housing types, subject to constraints of avail-
able land and zoning restrictions. The model is then
iterated from one year to the next to simulate the growth

of the region." 3h

32

There are several significant features of the S.E.W.R.P.C.
study. Besides including many features of other studies,
it stresses the fact that this will be a regional plan
instead of metropolitan in scope. By this they mean to
include the rural fringe in the development analysis.
Furthermore, the transportation study was launched by an
existing planning agency which can make plans, integrate
land use and transportation plans, and implement them on
a continuing basis. Another significant feature is that
this is the first time alternative transportation plans and
regional development plans will be evaluated in light of
a comprehensive assessment of effect on the natural re-
sources, the ecological balance, and on the resulting total
environment, as well as the overall costs of developing
and operating the combined regional land use-transportation

system. 35

Summary

The preceding analysis has necessarily been limited to
only an overview of transportationmstudies and their
land use models, as dwelling on details would have length-
ened this chapter without adding substantially to its
basic purpose. A review of the studies contained herein
illustrates the general development and sophistication of

land use models and the broadening of study objectives.

33
These objectives reflect to a great degree upon the
models' form and use.

The Detroit Area Transportation Study was the first
to use future land use as a basis for planning transportation
facilities. Its objective was to develop highway plans
in accordance with an existing master plan. As other
studies were initiated, broader objectives, new concepts,
and new procedures were introduced. Today, many trans-
portation studies utilize mathematical models that are
designed for use in simulating urban development, testing
alternative policies, and describing the interrelated

effects of land use and transportation facilities.

10.

11.

12.

13.

1h.

15.
16.

17.

FOOTNOTES

 

Detroit Metropolitan Area Traffic Study4_Part I:
Data Summary and“Interpretationi(Lansing: Speaker -
Hines and Thomas, Inc., 1955). P. 13.

American Institute of Planners, Proceedings of the
1963 Annual Conference (Milwaukee,Wis., 1963), p. 107.

 

Detroit, 100. cit.

 

Highway Research Board, Proceedings of 36th Annual
Meeting (Washington, D. 0., 1957), pp. 1-7.

 

Detroit Metropolitan Area Traffic Study, Part II,
FutureIFrafTic and a’Long Range Expressway Plan
(Lansing: SpeakereHines and“Thomas,’Inc., 1956),
pp. 15-270

Henry Fagin, "The Penn Jersey Transportation Study",
Journal of the American Institute of Planners, XIX

(Feb. 1963); 10.

Lowdon Wingo, Jr. and Harvey S. Perloff, The Washington
TranSportation Plan: Technics or Politics (washington,
D.C.: Resources for the Future, Inc., 1960), cited by
S. D. Marquis, Excerpt No. 77 (Michigan State University,
1962) p. 13.

 

 

 

 

Morris Trevithick, Land Use Analysis In Metropolitan
Transportation Studies - A Summary ofMethdds,
IEStitfiteSofiTransportation andTraffic_Engineering,
Paper No. S, University of California (Berkeley, 1962),
p. 7.

Wingo and Perloff, loc. cit.

Chicago Area Transportation Study, Volume I, Study
Findings (ChICago: Western Engraving and‘Embossing
00.. I959). pp. 1-3.

Ibid., II, 20.
John R. Hamburg and Robert H. Sharkey, "Land Use

Forecast" (Chicago Area TranSportation Study), 1961.
Cited by Trevithick, loc. cit., p. 9.

Bart Burkhatter, "Summaries of Growth Models", A

memorandum to Tri County Planning Commission,
Lansing Michigan, Mbrch 12, 196a, p. 5.

3h

18.

19.

20.

21.
22.

23.
2h.
25.
26.

27.
28.

29.
30.

32.

33.
3h.

35

Lowden Wingo, Jr., Transportation and Urban Land
(Washington, D. C.: Resources for the Future, Inc.,
1961). p. In.

 

J. D. Carroll, Jr., "Fitting Transportation Systems
Plans to Urban Land - Use Projections", The Dynamics
of Urban Transportation (Detroit: Proceedings of A
National—Symposium Sponsored by Automobile Manufac-
turers Association, Inc., Oct. 23-2h, 1962), p. 10-11.

 

 

Walter G. Hansen, "How Accessibility Shapes Land Use,"
Jouznal of the American Institute of Planners, XXV

 

Wingo, loc. cit., p. 15.

Baltimore - Washington Interregional Study; Land Use
and Transportation (Baltimore: Baltimore‘RegionaI
Planning Council, 1960).

 

Ibid., p. 19.
Ibid., p. 25.
Ibid., p. 10

A Prospectus; LABTS (Los Angeles: Los Angeles Regional
TranSportationPSgudy, 1960).

Ibid.

Prospectus (Philadelphia; Penn-Jersey Transportation
uy! 9), p. 20

Trevithick, loc.cit., p. 15.

Britton Harris, PlanninggMetrppolitan Transportation
Studies, P.J. Paper No.91 (Philadelphia: Penn Jersey
TransportationgStudy, 196I), p. 2.

John D. Herbert and Benjamin H. Stevens, A Model for
the Distribution of Residential Activity, PZJ. Paper
No. 2 (Philadelphia; Penn Jersey TranSportation
Study, 1960).

 

Prospectus (New York: Tri-State Transportation
Committee, 1962), p. 9.

 

Ibid.

Burkhalter, loc. cit., p. 3.

36

35. Regional Planning Program Prospectus (Waukesha:
Scutheastern Wisconsin Regional Planning Commission,
1962). pp. 30-37.

 

CHAPTER II
USES AND IMPLICATIONS OF LAND USE MODELS

As was shown in Chapter One, the use of models has
spread from strictly transportation planning to urban
and regional planning as well. It is the purpose of this
chapter to define the potential use of models in the
permanent planning process, and to point out the implica-

tions which the use of models may have to planning.

Uses of Models in Planning

 

It has been said that the validity or usefulness of

a model should be judged by its suitability for a particular
purpose. 36 The purpose of land use models to date has
been to aid in predicting future land use arrangement.
Assuming that this purpose can be accomplished with an
existing model, many uses related to the planning process

can be visualized. The planning process, diagrammed on the
following page, consists of the following steps: 1) concept
and theory formulation; 2) goal forming; 3) survey and fact
finding; u) analysis; 5) plan making; and, 6) programming. 37
The remainder of this section describes the uses that

present day models could perform in each step of the

planning process.

37

PLANNING PROCESS

 

 

 

 

 

6 . Programming
5 Plan Making
A Analysis

 

 

Survey & Fact
3 Finding

 

 

2 Goal Forming

 

 

 

Concept & Theory
Formulation

 

38

39

Concept and Theory Formulation

 

This initial step of the planning process is not
subject to mathematical formulization because it is here
that basic ideas are born. It could be said that this is

where mathematical land use models are developed.

Goal Formigg

 

This is one of the most important steps in the planning
process. It is here that the framework for remaining actions
is established. A model could be of significant value at
this stage if it were well defined, and the type that could
be used to test alternative policies and their resultant
effects. If models were well defined as to the inputs
needed and outputs that could be expected, planners (or
decision makers) would be more encouraged to form goals
and objectives. For example, consider a model that has as
an input some measure of sewer extension possibilities.

This would necessitate a policy decision which presumably
would be based on goals and objectives. Following this line
of reasoning, the use of models would provide for a better
consensus of area goals and policies. In fact, alternative

policies have been inputs to several models to date.

Survey and Fact Finding

 

Once a model has been developed for a particular area,
it seems obvious that the types and form of data to be

collected would be well defined. Also, since the data

he
would usually be used on a computer, it could be collected
in a manner which would facilitate its being transferred
to IBM cards or magnetic tapes for easy storage. Thus,
efficiency could be achieved in the handling of data, and in
data collection by knowing beforehand precisely the kind
of data needed.

This is a commendable use for models, assuming that
one has been developed for the planning area in question.
This assumption is necessary because at the present state
of knowledge it does not seem possible to use a model which
was developed for a city in one section of the country for
a city eISewhere, due to the fact that each city is unique
in its particular patterns of growth, etc. As a result,
much data is necessarily collected for use in constructing
the model -— sometimes in a trial and error fashion.
Therefore, much of the efficiency is lost in the initial
planning efforts. Consequently today's models can offer
only a rough guide to the survey and fact finding phase of
planning. The real advantage comes in the continuing and

updating processes which follow the initial plan making.

Analysis

This phase of the planning process is often nebulous
and hard to define. In many respects it overlaps with the
plan making phase. However, for definition reason, analysis
can be said to consist of evaluating and synthesizing

data in order to identify problems. The models today

hl
that are used for predicting are of little value in
identifying existing problems. Their usefulness lies in
training the observer to examine the city as a whole by
requiring, as inputs, total area effects of individual
developments. Also, the analysis stage could afford the
opportunity for the model and its data to be tested on

known past growth.

Plan Making

 

It has been said that the ultimate purpose of all
planning work is to recommend decisions. A decision involves
a choice, and the probability of choosing the correct
alternative depends on the manner in which the alternative
is stated, and the amount of data that is taken into account. 38
If through the use of models much data and many alternative
solutions can be analyzed, then the probability of making the
right decisions is increased. To digress for a moment,
consider an investment of one billion dollars, which is
not too uncommon in a complex urban area. If the probability
of choosing the correct alternative, concerning this in-
vestment, is increased by as little as one per cent, a
savings of ten million dollars could result. Herein
lies the justification for the expenditure of time and
money on the development of computer models. 39

In regard to the plan making phase of the planning
process, perhaps the first use to which models can be put is

projecting existing trends and policies into the future.

lie
This enables the identification of future problem areas if

existing trends are followed. Also, in the initial stages
of plan formulation, models can be used as a substitute for
actual experiment. Growth can be tested under various
assumptions, and the effects of various policies on devel-
opment can be shown. Environmental changes can be hypothe-
sized and their effects demonstrated. All of these can
result in the identification of problems, pressures, and

a future range of choice.

With the use of models, planners can better weigh
alternatives in terms of goals or their achievability, and
the government investments, policies and other actions,
including transportation facilities, needed to implement
them. In the various alternatives that may be presented,
the degree to which objectives are achieved in each can
be measured. Besides testing the achievability of certain
objectives, more detailed needs such as schools, recreation
areas, and transportation facilities could be identified
by comparison to specific standards.

In summary it can be said that planners can not make
sound recommendations, and public officials can not intelli-
gently decide issues unless a clear understanding exists
of the nature and magnitude of the issues involved in
each case. The purpose of models is to enhance the
decision making ability by providing a means of studying the
implications and consequences of various alternatives. "0

In this respect it is obvious that models, contrary to some

LLB

suspicions, supplement not supplant the planner and his

functions.

Prggramming

 

Once a plan is decided upon, a series of steps and
actions are necessary for its implementation and continual
updating. The use of a model can help formulate such a
series of steps or program. With problems and needs of the
future identified in quantative terms, the planner can
program both community and professional resources to solve
these problems. One could argue that conventional planning
techniques offer this advantage. The difference, however,
occurs in the way in which the future and its problems are
predicted. Through the use of models it is possible to
age the city in a series of iterations under various
hypotheses. Thus a much better idea of when, and under
'what circumstances a particular action or expenditure will
be needed. This has great potential when it comes to
effecting policies, programming capital improvements, and
programming staff operations. In light of the above uses,
it follows that as a result of using models, planning can

more easily be turned into a program of action.

In summarizing the uses of models in planning, several
conclusions can be drawn. Land use models today are
designed for the purpose of predicting the future dis-

tribution of urban activities. It is generally conceded

At
that this purpose has not yet been demonstratably achieved,

and consequently models are considered by many as being
useless. However, the sole criterion for usefulness should
not rest on this near-impossible to achieve objective.

The final results need not be perfect for a model to be
useful because it can serve a significant purpose if it
enhances our understanding of the city and how it functions.
For example, considertshe use in the plan making stage
where the results of specific alternative policies were
simulated. The accuracy of the forecast in this instance
would not be as important as the identification of the
effects of various policies. Another example is evident

in the data collection phase. The ability to predict with
a model does not enter into the determination of the
processes and data that shape the urban form. Only the
structure of the model and its necessary inputs need to

be known. This line of reasoning can be applied to many

of the uses described in this section.

Therefore, it can be said of present models that even
though they are not yet able to predict, they are performing
a useful purpose in helping to identify urban functioning,
and defining the scope of planning, as well as serving the
other purposes listed in this section. These uses will
become more apparent and acceptable as models are more
refined - as they inevitably will be. Perhaps, as is
suggested in a following section, models will be developed

which will not be aimed solely at predicting the future

LLS

state of development.

Implications of the Use of Models to Planning

Reflection on the development and use of land use
models to date suggests several implications regarding
urban planning. In this section, the implications felt
'to be of significant consequence are discussed.

As shown in Chapter One, the early transportation
studies were organized on an ad hoc or independent basis.
The only relation to the urban planning function occurred
through the basing of transportation plans on land use
projections, which may or may not have been prepared by
the local planning agency. During the evolution of these
transportation studies, technical and organizational
sophistication has taken place. Today, it is not uncommon
for a transportation study to be organized on a continuing
basis, or as part of a permanent regional planning operation.
This advancement was due partly to the recognition that
transportation facilities not only serve land use but are
a major force in shaping its distribution.

Assuming that a model were operational for a particular
urban area, several implications can be visualized resulting
from an integrated organizational structure. For instance,
at the outset of the planning processlong range data
needs could be established by knowing the necessary inputs

to the model. Consequently, other governmental departments

A6

could be more efficiently brought into the planning process
and better cooperation obtained. The most important
implication, in regard to organization, is that transporta-
tion facilities can be more effectively used as a means for
implementing desired plans or goals because transnortation
facilities do shape the urban environment more that any
other single factor.

Aside from organizational implications, the size of
the area for which models are effective indicates another
implication. From past experience models seem to be
more applicable to larger areas because behavioral actions
and other inputs to the models must be handled on a gross
basis to be statistically reliable. Therefore, metropolitan
and regional planning tend to benefit from the use of
models more than the small community or neighborhood type
of planning. The advantage of this situation is that a
broader, more encompassing regional viewpoint is established,
whereas the disadvantage lies in the possibility that the
more detailed neighborhood interests may be obscured.

Because of the vast complexity of metropolitan planning,
and the advent of data processing and computer models, it
seems inevitable that many steps in the planning process
will become more computerized. The trend in this direction
is obvious today. Established data banks can be of
significant use to model development, and the use of
models can help establish and supplement these efforts

toward more efficiency. As a result of this increased

LL?

efficiency in data handling, the planner will have more
time to devote to analysis and the more human aSpects of
planning. Another implication in relation to data and the
use of models is that a greater stress will be put on
research to determine the factors that affect growth. This
data wbuld be necessary before a model could be made
operational. Also, the kind of research or data to be
collected in the future for evaluating methods on a
continum can be identified by the use of models.

With the vast amount of data that could be available,
many coordinated functions of planning, besides transporta-
tion, would benefit through the use of these data and models
One such function would be urban renewal. Models could
simulate how the market operates under given assumptions
and pressures. Thus, more effective housing and reuse
of cleared areas could be accomplished.

Perhaps the most important implication of model usage
is in the area of policy making.' Since models are now
designed to simulate area growth under various alternative
assumptions and policies, it is imperative that the decision
makers in the region, or community adopt a set of policies.
This implies that policy planning may become more important.
There is certainly a need for such a step because, tradi-
tionally, policies are seldom mentioned, seldom identified
and even less often adopted. Policy planning and achieving
community objectives could be made to work hand in hand

through the use of models. Different policies regarding

AB
zoning, taxes, the extension of services, and others
could be tested with the use of models to see which combin-
ation best implements community objectives.

Another implication of using models in planning can
be visualized in planner education.' Specialization in
model development and research in the behavioral aspects of
urban processes is a must if better models are to be
developed. Perhaps metropolitan and regional planners
could forego detailed studies of such aspects as zoning
and concentrate on studies involved with model develop-
ment. 0n the other hand, it may be desirable to have a
detailed knowledge and then study model development as an
advanced interest or degree.

Finally, it must be mentioned that models do not seem
able to predict industrial land use patterns to the same
degree of accuracy as residential land uses. "1
(Commercial uses are directly related to residential,
therefore they can be derived from or after residential
uses have been distributed.) Major industrial uses are
difficult to locate, and because they are usually the
basis on which residential uses are distributed (via
accessibility to work areas), they become the foundation
for future land use pattern. The implication here is that
conventional planning methods will be necessary to locate
large industrial uses at least until a better method of

specifying their location is developed.

’49

In conclusion, many implications exist in the present
state of model development. A more powerful planning
operation can be undertaken through the integration of
transportation and urban planning and through the use of
models. If development factors and actions needed are
identified along with their results, then perhaps more
imaginative and broader planning can be achieved, as well
as some degree of implementation of plans through transpor-

tation facility layouts.

FOOTNOTES

 

36. Jay W. Forrester, Industrial Dynamics (Cambridge:
MIT - Wiley, 1961), p. 115.

 

37. Sanford S. Farness, The Planning Process and Environ-
mental Health, paper presented at the Conference on
EnvironmentaI Health Planning for Metropolitan Areas,
University of Michigan, Ann Arbor, March 23, 196A,
mimeo.

 

 

38. Hans Blumenfeld, "Limitations of Simulation of
Futgre Behavior", Plan Canada, Vol. V, No. 1, 196A,
p. o

39. Ibid.

 

 

ho. Anthony R. Tomazinis, An Introduction to Mathematical
Models, P.J. Paper No. 8"(Philadelphia; PennfiJerseyfi_
TranSportation Study,I960), p. 17.

 

A1. Review of Existing Land Use Forecasting Technigues
(NewIYork: Ttaffic’Research CorporatiOn, 1963), p. 76.

SO

CHAPTER III
LIMITATIONS OF EXISTING LAND USE MODELS

In Chapter One it was shown how transportation studies
have progressed from merely designing transportation
facilities to serve a projected land use to the present
stage, where it is realized that transportation
facilities, to a large extent, determine urban form.
Consequently, transportation planners are putting more
emphasis on the study of factors which shape present, and
especially, future urban patterns.

Urban planners have been confronted with the problem
of determining future urban form for many decades, but are
still struggling with many of the original problems and
several new, more complex ones. With the relatively recent
advent of large scale transportation planning and its ob-
vious far reaching effects, the problems of forecasting or
"planning for the future" have been brought to a head.

The design of land use models has been an important step

in trying to solve these problems by predicting future urban
form on both a comprehensive and skillful level. However,
there are still many basic limitations to forecasting in
general, as well as limitations to the more refined level of
model structure. This chapter will point out these limita-
tions which must be considered before we can honestly plan

for the future. The purpose here is not to debate the

51

52

issues which will be raised, but merely to point them
out for consideration.

Upon analyzing the many limitations, it was felt that
they could be categorized in the following manner:
conceptual problems; unknown variables or informational gaps;
and problems of model structure. In the following pages
these categories are further broken down into sub topics

for discussion purposes.

Conceptual Problems

 

One might consider conceptual problems as inappropriate
to models per se, but the understanding of the fundamental
issues raised here is a prerequisite to any type of long
range planning. It is felt that until this understanding is
accomplished and to some extent incorporated in models,

model building is little more than a mathematical exercise.

Values - Goals - Policies

 

The formulation of goals and objectives is the most
fundamental aspect of any planning process, yet it is often
ignored or overlooked. If one accepts the fact that trans-
portation facilities do affect the structure or pattern of
the city, then it is evident that transportation planners
should consider what the desirable pattern should be. It
is felt that this most fundamental question - "what

objectives do we want to accomplish and how?" - has not

53
been answered to date. If this is true, then billions of

dollars as well as less tangible things are being wasted
and lost by blindly following past trends when planning
for the future. Is the general population too entrenched
in past and present cultural traits to foresee or envoke
change? Just what do people really want? These questions
are the ones that must be answered, however difficult.
One might say that it is impossible to determine the
values and goals of such a pluralistic society as ours,
but in response it can be said that this only emphasizes
the basic contention that we do not know what we are
planning for.

One of the reasons for this uncertainty lies in the
fact that metropolitan areas are fairly new formations. At
first, small scale planning principles (i.e. neighborhood
theory) were applied to the metropolis, but it was soon
realized that a new concept was needed. Many concepts are
presently in existence or are being formulated, but none are
widely accepted as yet. Several questions arise regarding
basic issues and concepts which have limitless effect on
the urban structure. A few of these are given as examples:
1) Do we need a single central business district, or are
several subcenters the answer? 2) Should families with
higher incomes locate nearer the central business district?
They are the most productive, and therefore should they
not live near the most productive area in order to minimize

travel distance? A2 3) Should we plan for an "efficient"

Sh
metropolis? If so, what is "efficiency", and who determines
it? A) Do people actually like living in the city or is
it a necessity? If they do like it, why are thetapproaches
to big cities always crammed with people on Sunday evenings
who are trying to get back home after a weekend "away"?

The above questions are intended to illustrate that
before we dash headlong into planning transportation faci-
lities, values, goals, and objectives need to be formulated.
This is no easy task if done thoroughly, but the result
may well be worth the effort. Consider a new type of
machinery that is first conceived in the minds of men, then
designed and built. Could this not apply to metropolitan
structure? "3

A corollary to the need for formulating objectives is
the statement of policies by the governing bodies. Tradi-
tionally this phase of planning has been left out, and as a
result, there has been little implementation of plans. The
early transportation studies did little with "policy planning."
Recently, however, several studies have made this phase very
important and are to be commended. They have realized that
the assumption of weak policies means poor long range
planning. Finally, if there can not be a theory of city
structure, because of constant changes in our society, then
the formulation of long range goals and policies can

provide the necessary guide to follow.

SS

Assumptions and Bases of Forecasting

 

Assumptions play a major role in forming the basis of
forecasting and model construction. In the first chapter,
several limitations concerning assumptions were brought out
while discussing the evolution of transportation studies.
For illustration purposes a few of these will be reiterated
here: 1) Detroit assumed the travel friction between
zones would be equal in the future because future transpor-
tation facilities would be evenly spaced; 2) Washington D.C.
assumed that local agencies would effectuate no policies to
induce change in the existing trends; 3) Hansen assumed
that people would react as molecules do in the gravitational
theory; and u) Penn Jersey had to make many assumptions
about human behavior before their model could be formulated.
With a little reflection, it should be evident that assump-
tions have far reaching effects, and an error in this stage
could be disasterous.

It is conceded that there could be little activity,
be it planning or other, without making assumptions.
However, in examining the bases for these assumptions
many limitations become apparent. First, present models
and other forecasting methods are based on data derived
from metropolitan areas in the United States, many of which
have had roughly the same population and economic composi-
tions, the same institutions and the same technology. The
similarity in land use patterns is a result of these cultural

traits. What would happen if one or more of these traits

d-..

 

56

were changed? In South American cities the high income
families live near the center of the city and the low
income groups on the periphery -- exactly opposite of the
pattern in this country. This is a result of different
cultures. The main theme of this discussion is to point
out that there is no law of urban form, but only a con-
sistency in the behavior of culturally conditioned peoples.
Therefore, any forecast which is based on past trends and
does not examine the differences in cultural characteristics
existing at that time and the ones most likely to exist in
the future, is worthless.

Another weakness of present forecasting is the under-
estimating of society and individuals and their ability to
effectuate change. As an example, consider the models
that are based to some extent on land values. Can not the
local governments manipulate these land values since they
are merely manifestations of our culture? Also, the models
based on the gravity theory underestimate the mobility of
industry and residences. They could, conceivable, turn the
city inside out. "5

Another limitation can be seen in models which take
the existing city and add increments of growth onto it.
These models overlook, except for immediate projects,
the possibility for drastic change in the existing structure

from urban renewal, or more basic cultural changes.

57

Unknown Variables and Informational Gaps

 

In the preceding section it was pointed out that values,
goals and the basis for several assumptions are relatively
unknown. Proceeding in a similar vein, this section is
designed to point out types of information that are very
important in shaping future urban patterns, but are for
all practical purposes, unknown. It was felt that these
unknowns could be grouped into the following four areas
for discussion: 1) behavioral; 2) technological;

3) general information; and A) model structure.

Behavioral Unknowns

 

Data on human behavior in many respects, particularly
in regards to tranSportation planning needs, is very limited.
In observing time trends where data is available, it is
usually impossible to hold constant the environmental
forces over a period of years, and thus impossible to
derive a true picture of the behavior itself.

Since behavioral data is difficult to obtain and fore-
cast, it has been necessary to delineate behavior patterns
with reference to cross sectional data. The models which
simulate behavior are limited in their accuracy to what can
be scientifically foreseen from the study of the inter-
relationships of this cross sectional data, and to what can
be measured or inferred from urban statistics. AC Since

behavior itself can not actually be determined, and only

58
a cursory relationship between it and certain other variables
can be made, a definite limitation exists, and a problem
arises as to which variables to project and how.

It can be said that little is known about the funda-
mental elements expressed in the choices people now make.
Neither theory nor data is available to help anticipate
human behavior under various hypothesized conditions of
the future. Therefore, before more planning is done, or
an adequate theory of metropolitan growth and structure
is developed, assimilation of considerably more knowledge
as to the behavior of locational decision makers and the

effects of various policies on their behavior is needed.

Technological Unknowns

 

It is generally agreed that advances in technology are
mighty forces in shaping our living patterns. (Witness
the automobile for a much used example.) There exists
a great myth that technological advances come about as if
by magic, under no control of humans. Most people do not
realize that such advances are the product of men and our
culture. The problem or limitation of most forecasting
today is the complete lack of trying to determine these
advances. Planners in general sit back and make plans
knowing that changes in technology will occur, but falling
back on the excuse that no one can foresee technological
change. Thus when the future arrives and plans are com-

pletely antiquated, there is no remorse because "technology"

59

caused the failure.

It is conceded that technological innovations are very
difficult to foresee, and often impossible to predict.
This in itself can be viewed as a limitation, and will
continue to be one until man learns to either control or
predict these advances. However, a more serious limitation
of most long range planning is the ommission of technological
advances which are figuratively, "just around the corner."
Inclusion of these and their alternative effects would not
require too much foresight since many technical advances are
in the process of development now. As an example, consider
the emphasis by several studies on the requirement that
sewer service be available before development occurs.
This is valid now, but how will chemical toilets, that burn
the sewage on the spot, affect development? Another
example is the development of very small individual cars
that would require less than half the lane space now
needed and therefore double existing street capacities.
One area which has great potential and far reaching influence
is that of communications. The substituting of communica-
tions for tranSportation could drastically reduce travel
and alter development patterns. "7

Therefore, the need in planning and model building is
to anticipate technological advances, and to explore their
implications so that projections are not in considerable
error when the future arrives. Planning today should be of

dynamic, revolutionary methods - not based on the extension

60

of past trends and outdated empirical methods.

General Informational Ggps

 

It is acknowledged that models are no more accurate
than the data which goes into them. This data, in many
instances, is relatively unknown and therefore assumed, or
it needs to be forecasted, and is consequently subject to
all the conventional problems of present day forecasting
methods. Without belaboring this latter point, which has
been discussed to some degree previously, the remainder of
this section will be devoted to listing several areas in
which more data and research are needed before adequate
projections can be made.

Research needs to be focused on the metropolis as a
functioning, evolving process in order to answer the follow-
ing questions:

1. What are the effects of size, dispersion, or
areal Specialization on the functioning of the
metropolis?

2. How does the metropolis work at any point in time?
How does it change? Can this change be controlled
and directed?

3. What are the significant differences in various
types of urban form?

A. Exactly how does the city respond to transpor-
tation? Will the future metropolis take one

form if certain arrangements are made, and

61

another if not?

Listed below are many other general questions that

need to

future:

1.

be answered in order to prOperly anticipate the

How will people behave with larger levels of
income?

How would modern business react to increased or
decreased mobility?

How would home buyers react to a more liberal,
mortgage policy on used housing?

Some experts have said there is a lack of
information on non-residential land use and

its distribution. "8 Another has said there exists
adequate information on industry and commercial
uses, and what is needed is more information on
residential distribution. "9 Contradictions

like this emphasize the need for more research.
More information is needed on the effects of
aesthetic and social factors, and a meaningful
measuring device for these needs to be developed.
Locators may wait varying lengths of time before
incurring the expense of moving, even though other
factors are present which make relocation optimal
at that time. Therefore, there is a need for long
term information to determine how to anticipate
this behavioral lag. 50

What actually are the formative factors that shape

62
the metropolitan region? Assuming that transpor-
tation is one of them, where does it belong in
the hierarchy of other forces?
Many more data needs could be listed, but the above
examples are enough to indicate that there is a need for
much inter-disciplinary research before forecasts can be

made with a substantial degree of accuracy.

Model Structure

 

Most of the previous limitations discussed were
applicable to any type of forecasting. This section is
devoted to pointing out those weaknesses which are unique
to models. The following limitations are of a general
nature, and are not detailed in regards to mathematical
and statistical methodology for the following two reasons:
1) the author's knowledge and experience in this Special-
ized field does not qualify him for such an undertaking;
and 2) the model builders themselves do not really know
the reliability of their formulations until the results are
evaluated at the end of the forecast period.

Generally, there are several disadvantages which are
applicable to all types of models. In the first place,
land use models often oversimplify the problem and solution.
Akin to this is the fact that the symbolic language which
computers use is limited. Another type of problem arises

when one becomes too engrossed in the mathematical model

63

and feels that it is actually the real world. When this
happens the model loses its usefulness as a tool and its
user loses Sight of reality. 51 Also, present day

models that are designed for use in prediction have not
had sufficient time to test their results, even though
some models were tested on past data and results compared
to the present Situation. The results of this type of
procedure are usually accurate, but they should be because
the models were designed using this data.

In order to more fully relate the limitations of
models it is necessary to classify the various types of
models and identify their individual limitations.

Anthony Tomazinis states in Penn Jersey Paper # 8, "An
Introduction to Mathematical Models", that models can be
classified in three general types according to the nature
of the resultant product. 52 These are deterministic,
stochastic and optimal. Subsequently, each of the above
can be divided into two kinds according to the method in
which the model functions. Thus, for deterministic and
stochastic, there are analytic and numerical models, and
for optimal there are plain optimal and twin goal optimal

models. A Brief discussion of each follows.

Deterministic - Numerical Models
Models of this type express graphically instead of
mathematically the relationship between variables. This

is valuable when the behavior of variables is not well

6A

known and can not be formulated. The Chicago Area Trans-
portation Study used this type in defining the relationship
between distance from the central business district and
density of development. These models have by nature limited
predictive ability, but are valuable in tracing historical
data. The technical limitation lies in its inability to
combine relationships involving more that two or three
variables in a satisfactory manner.

Neither deterministic model can be used to describe
the processes which shape the urban pattern. Their
users can not be concerned with processes of change or the
willful redesign of land use patterns, and they essentially

assume development takes place outside human control.

Stochastic - Analytic Model

This type of model indicates both a non-linear
relationship and a degree of probability of the occurrence
of a certain event. The Chicago Area Transportation Study
staff applied a form of this in determining trip distri-
bution. Because of the probabilistic and statistical
problems of this type of model, none have teen used for
land use projections. This holds true for stochastic

numerical models also.

Plain Optimal Model

This type of model is a Simple form of linear programming

which requires its inputs to be linear in nature. No

65
model of this type has been used in urban planning for
several reasons; one being the inability of the social
sciences to provide clear cut objective functions, input
coefficients, and a list of the important variables and

constraints needed.

Twin - Goal Optimal Model

 

This type of model is also part of the linear program-
ming concept, but is more complex than the previous type.
Penn-Jersey is pioneering the use of this type of model
and it looks promising because of comprehensiveness of
the concept. Where the previous type of model was
concerned with only one process of achieving a result,
this model type can maximize or minimize a function in
response to a second consideration which could include
an allocation of resources according to both the total
output and a preference schedule according to the availa-
bility of resources in the area. However, the mathematical
complexities and the newness of this concept are limiting
factors. Also this "...linear programming approach is
questionable as a Simulating technique because of its
property of maximizing or minimizing some aggregate function
(such as total rent paying ability of all residential
locators). Economic theory holds that each individual is
trying to maximize his economic position so that a linear
program can not be said to simulate locational behavior

unless it can be shown that the sum of all actions to

66

maximize individual economic position is synonymous

with maximization of the aggregate economic position.

Such a relationship does not seem to be readily demonstra-
table, either for residential or industrial locators.

Land use models based on linear programs may also be
questioned because of their reliance on economic motiva-
tions and their consequent exclusion of all input variables
which can not meaningfully be translated into economic

terms." 53

To summarize, many limitations have been expressed
in this chapter. All of them have not dealt with models
particularly, but with diverse areas of the social and
physical sciences. However, the issues raised herein
can neither be forgotten nor excluded from consideration
when developing a model, as they are basic to determining

possible future development patterns.

u2.

A3.

A5.
A6.

li7.

1180
A9.

50.
51.
S2.
S3.

FOOTNOTES

 

Richard Weistmantal, "Dantes Inferno - The First
Land Use Model," Journal of the American Institute
of Planners, XXV(Nev. 1, 1959).

Guthiem, "Integration of Transportation and Land Use
Planning Through Quantitative Models", Transportation
Design Considerations (Woods Hole Conference: NAS -

 

 

NRC,8196l).

Melvin M. Webber, "Transportation Plannink Models",
Traffic Quarterly, (July, 1961) pp. 373-3 9.

Weistmantal, loc. cit.
R. M. Zettel and R. R. Carll, Summary Review of

Metropolitan Area Transportation Studies in the U.S.
(University of CaIIfornia,l962).

 

 

 

Frederick W. Memmott III, The Substitutability of
Communications for TranSportation, TechnicaI’Paper,
Niagara Frontier TranSportati0n_Study, pp. 1-12.

 

Wingo, loc. cit.

 

Review of Existing Land Use Forecasting Techniques,

Iod. cit.

Zettel, loc. cit.
Bross, loc. cit.
Tomazinis, loc.cit., pp. 3-16.

Review of Existing Land Use Forecasting Techniques ,

loc. cit., p. 75.

67

CHAPTER IV

CONCLUSIONS AND RECOMMENDATIONS

FOR FURTHER RESEARCH

The development of land use models during the last
decade has been viewed with both enthusiasm and skepticism
by planners as well as members of other professions. As
brought out in this thesis, there are justifications for
these opposite views, depending on the manner of analysis.
If one approaches present day models with the thought
that more insight can be obtained on the sensitivity of
urban areas to changes in policy and behavior and that a
better overall understanding of urban development can be
achieved, then there is a valid purpose to which existing
models may be put. However, if one expects to successfully
predict a future urban development pattern, then there is
much room for skepticism. In either case the progress of
this particular subject to date, and the great potential it
has for the future Should be considered.

The Chicago Area Transportation Study isolated
various indices identifying land development. However,
these indices were not the actual forces causing develop-
ment patterns, but only the results of certain forces.
More recent studies have tended to identify these forces
or causal factors which have created certain development

patterns. However, there are still vast areas of

69

research needed to better substantiate present concepts and
to provide more reliable data as inputs to models. Some of
these research and data needs are: l) Forthcoming policy
decisions that are likely to affect the pattern of growth
for the region and smaller areas need to be identified,

as well as the nature and costs of various alternative
policies, plans, and programs that are to be tested in

the model; 2) Also needed is a better identification of
factors expressing the interrelationships among the elements
of the metropolis. Since each region has different climate,
topography, resources, and historical period in which it
was developed, it is necessary to evaluate each metropolis
independently; 3) The effect of tranSportation on land
development and land development on transportation has

been researched a great deal, but the interrelationships

are still not well known. A) Consideration must be

given to the respective needs and programs of the various
industrial, commercial, and social organizations in the
region since they constitute a major force in development;
5) The whole gamut of behavioral characteristics and

their potential change is poorly defined and needs research;
6) Technological changes which may rapidly affect the
economy or efficiency of transportation should be investi-
gated; 7) Analysis of the political structure of regions
is necessary to determine coordination and implementation
possibilities.

In reviewing existing models, there are a number of

7O

concepts which should be included or at least considered,
for use in all models. These include: 1) a modelwhich can
be used to forecast changes rather than absolute values;
2) the use, whenever possible, of exogenously determined
region-wide control totals; 3) a model which can deal
with all variables simultaneously; h) a model which
contains as many behavioral relationships as may feasibly
be utilized without an inordinate amount of mathematical
complexity; 5) a model which is sensitive to accessibili-
ties as measured by travel times and/or costs by all major
modes; and 6) models should be flexible so that they can
accomodate changes in information and technology. 5"
To date no model has all the desired properties listed
above, but perhaps as more experience is gained, a truly
comprehensive model may be constructed.

To date, many models have been developed for use in
transportation and land use planning. Most of these
are quite diverse in their methodology and concepts, but
they are all aimed at predicting some future state of land
use development in the metropolitan area. This objective is
very commendable, and if realized, could be most useful.
But the big question which arises is whether or not this
objective of predicting a future state of development can
be achieved by using models, or any other means.

The validity of any type of model depends on its
suitability for a particular purpose. If predicting the

future state of a system is the purpose, then the behavior

71
of that system must be fairly constant and well defined.
A good example is the solar system whose future state can
be described rather accurately because its behavior is not
affected to any extent by outside forces, and is not likely
to change on its own accord. The city and its functioning
are not as stable as the solar system. Even if it were, it
would be subject to changes of the type Forrester expresses:
"If a model could predict specific events far enough
ahead to allow action to be taken as a result of the
prediction, there may be sufficient time intervening
for random events to alter the predicted values.
If random disturbances do not exist, if the model is
sufficiently near to perfect, and if the momentum of
the system is not strong enough to inhibit action in
anticipation of the predicted events, then the action
taken is itself highly likely to affect the predicted
outcome. The prediction procedure would then need to
be a system decision function incorporated so that the
model recognized the effect of the prediction on other
decisions." 55
Since the city is not stable, but subject to many
changes, it can be concluded that a final predicted state
of development is not possible. However, this does not
rule out the use of models in urban planning. It merely
suggests that the basic purpose or objective of models be
changed to that of predicting the behavior of thetzity as
a system. In other words, models should be developed with
the express purpose of enhancing the understanding of the
forces and their results which ultimately shape the urban
environment.

A model which could achieve the above purposes would

indeed be complex, but cities themselves are no less complex.

72
The major difficulty of models today is that they do not,

and perhaps can not include adequate descriptions of the
changing metropolis. For the most part, planners abstract
a few variables and constants, and formulate a model. A
major breakthrough is needed in this field so that more
complexity, and therefore better results can be obtained.
As a proposal in this direction, it is conceivable that
the city could be described in terms of a dynamic open
system with many subsystems interacting to bring about

the urban development. It is important to realize that
one can talk about the behavior of systems as well as the
behavior of individuals. Therefore, "behavior models"
could be the basis for more complex and complete models.
It is not the intent here to present a lengthy dissertation
of systems, but in order to give a little insight on the
topic a brief explanation of the city asja dynamic open
system is presented in the following paragraph.

Within open systems, change takes place according to
the unfolding of developmental potentials which are
actually in the system at its inception. The inputs which
give rise to this change tend to be of a developmental,
non-recurring type, and as they continue to invade the
system, the total system continues to change, the result
being a sequence of systems which build upon and take from
its predecessor. As this process takes place, changes of
varying kinds and degrees of magnitude are continually

occuring so that there is never a particular, uniquely

73

identifiable system. The system can never be made to
stand still, but can only be represented as such in
abstraction for analytical purposes. 56

The developmental potential mentioned above could be
any given piece of land in a city. Presumably it has a
potential for use according to the state of the city
system at any one time, and it will obviously have a
different potential resulting from changes caused by inputs
into the system. AS long as the city remains an open
system, then model makers are going to have to view the
city as an evolving continuous process. The best tool
for this seems to be a systems approach where, hopefully,
the city system could be broken down into several subsystems
and their actions and interactions within and upon the
total complex could be identified.

If one could design a systems modelvvhich simulated
the urban process it would necessarily involve the interac-
tions between many subsystems in the community. Some of
these would be the household subsystem, developer subsystem,
builder subsystem, bank subsystem, real estate subsystem,
local government subsystem, transportation subsystem, and
others. In order to relate all of these a common denomina-
tor must be found. Gifford suggests that "These elements
are related by rules which have evolved for their inter-
action and coordination towards a maximization of individual
profit and the general welfare. And, in general, these

elements and rules may be conceived as having their basis

7A
in the optimization of the survival of the individual

personal systems of the inhabitants of the city, as this is
accomplished through a combination of individual and
socially organized actions." 57

Whether the common denominator which Gifford suggests
is the most valid is not the question here. The manner in
which the subsystems are Specifically related will depend
on results of much research. It does seem, however, that
the description of these systems will be based on their
behavior under varying circumstances, and under each of
these circumstances decisions will need to be made by
representatives of each subsystem. Since decisions are
usually based on previous decisions or their results, it
would seem that the sequence of decisions by subsystems
would be of primary importance. Therefore, a "sequential
decision model of the city system" is suggested as a
means of understanding the city, and as a result, provides

a better basis for planning.

FOOTNOTES

 

Sh. Review of Existing Land Use Forecasting Technigpes,
IOc. cit., p. I28.

55. Forrester, loc. cit., p. 128.

56. John V. Gifford, The Nature of the City as a System,
A paper presented—befbre the Bay Area Systems Group
of the Society for General Systems Research, San
Francisco, 1961, p. 7.

57. Ibid., p. 90

 

BIBLIOGRAPHY

Books

Bennis, W. G.; Benne, K. D.; and Chin, R.; "Conceptual
Tools for the Change Agent: Social Systems and
Change Models", in Bennis, W. G.; Benne, K. D.;
and Chin, R. (editors): The Planning of Change,
Holt, Rinehart and Winston, 1961, 187-198.

 

Bross, W. T. Design for Decision. New York:
Macmillan 00., 1933.

 

Forrester, Jay W. Industrial Dynamics. Cambridge:
PIIT - Wiley, 1961.

 

Isard, Walter, Location and Space Economy. Cambridge:
MIT Press, 1956.

 

Meir, Richard L. A Communications Theory of Urban
Growth. Cambridge: MIT Press, 1962.

Mitchell, Robert B. and Rapkin, Chester. Urban Traffic,
A Function of Land Use. New York: Cdlumbia
University Press, l95h.

 

 

Wingo, Lowden Jr. Transportation and Urban Land.
Washington, D. C.: ’Resources fhr the‘Future, 1961.

 

Articles and Periodicals

 

American Institute of Planners. Proceedings of the 1963
Annual Conference. Milwaukee, 1963, 103Ll3h.

 

 

Blumenfeld, Hans. "Limitations of Simulation of Future
Behavior", Plan Canada, V, No.1, 196M.

 

Blumenfeld, Hans. "Are Land Use Patterns Predictable?"
Journal of the American Institute of Planners, XXV

(May. 1959).
Dynamics of Urban Transportation. Proceedings of a National

symposium Sponsored by Automobile Manufacturers Associ-
ation, Inc., Detroit, Oct. 23-2h, 1962.

76

77

Fagin, Henry. "The Penn Jersey Transportation Study",
Journal of the American Institute of Planners,
XIX (FBI), 1963K

 

Guthiem. "Integration of Transportation and Land Use
Planning Through Quantitative Models," Trangportation
Design Considerations. National Academy of Science -
National—Research Council, Publication 8&1. Conference
Proceedings, Washington, D. C., 1961.

 

Hamburg, J. R., and Creighton, R. L. "Predicting Chicago's
Land Use Pattern", Journal of the American Institute
of Planners, XXV (May, 1959).

 

 

Hansen , Walter G. "How Accessibility Shapes Land Use,"
Journal of the American Institute of Planners, XXV
TMéy. 1959).

 

Harris, Britton, "Experiments in Projection of Transpor-
tation and Land Use," Traffic Quarterly (April, 1962).

 

Harris, Britton. "Plan or Projection", Journal of the
American Institute of Planners, XXVI (August, 1960).

 

 

Herbert, John D., and Stevens, Benjamin H., "A Model for
the Distribution of Residential Activity in Urban
Areas", Sournal of Regional Science, (Fall, 1960).

 

Highway Research Board. Proceedings of 36th Annual
Meeting. Nashington,7D. C., I957, 147.

 

Mitchell, Robert B. "The New Frontier in Metropolitan
Planning", Journal of the American Institute of
Planners, XXVII (August: 1961).

 

Row, Arthur T., and Jurkat, Ernest. "The Economic Forces
Shaping Land Use Patterns," Journal of the American
Institute of Planners, XXV (May,I9S9).

 

 

Voorhees, Alan M. "The Nature and Uses of Models in City
Planning", Journal of American Institute of Planners,
XXV (May, 1939).

 

Voorhees, Alan M. "Development Patterns in American
Cities", Highway Research Board Bulletin 293,
(Jan, 1961).

 

Webbep, Melvin M. "Transportation Planning Models",
Traffic Quarterly, (July, 1961)."

 

78

Weistmantel, Richard. "Dante's Inferno - The First Land
Use Model", Journal of the American Institute of
Planners, XXV—(Nov.,’l9597.

 

Reports

AASHO Committee on Urban Transportation Planning.
Organizational Procedures of 17 Urban Transpor-
tation Studies. New York: 1963.

Baltimore Regional Planning Council. Baltimore-
Washington Interregional Study; LandWUSe and Trans-
portation. ‘Baltimore: 1960.

 

 

Chicago Area Transportation Study. §EPdV Findings,
Volume I. Chicago: Western Engraving and
Embossing Co., 1959.

 

Chicago Area Transportation Study. Data Projections,
Volume II. Chicago: Western Engraving and
Embossing Co., 1960.

 

 

Detroit Metropolitan Area Traffic Study. Data Summary
and Interpretation, Part I. Lansing: ‘SpeaEer
Hines and Thomas,‘l9§5.

 

Detroit Metropolitan Area Traffic Study. Future Traffic
and a Long Range Expressway Plan, Part II. Lansing:
Speaker Hines and Thomas, 1956.

 

Harris, Britton. Planning Metropolitan TranSportation
Studies,_P.J._Paper No. 1. Philadelphia: ’Penn
Jersey Transportation Study, 1961.

Herbert, John D., and Stevens, Benjamin H. A Model for
the Distribution of Residential Activity,_P.J. Paper
"No. 2. Philadélphia: Penn JerseyfiTranSportation
assay, 1960.

 

Los Angeles Regional Transportation Study, Prospectus;
LARTS. Los Angeles: 1960.

 

Penn Jersey Transportation Study, PrOSpectus. Phila-
delphia: 1959.

 

Pittsburgh Area TranSportation Study. Final Report;
Vol. I StudygFindingg.

 

 

79

Regional Planning Commission, Regional Planning Program
PrOSpectus. Waukesha: 1962.

 

 

Tomazinis, Anthony R. An Introduction to Mathematical
Models, P.J. Paper—No. 31 Philadelphia: Penn
Jersey—TransportatiOn Study, 1960.

 

Traffic Research Corporation, Review of Existing Land
Use Forecasting Techniques. New‘York: 1963.

 

Trevithick, Morris. Land Use Analysis In Metropolitan
Transportation Studies - A Summary of‘Methods.
IEStitute ofITransportation andTrafficEngineering.
University of California, Berkeley, 1962.

 

 

Tri State Transportation Committee. Prospectus.
New York: 1962.

 

U.S. National Capital Planning Commission. Transportation
Plan; National Capital Region. Washington: 71959.

 

 

U.S. Department of Commerce, Bureau of Public Roads.
A Report on a Land Use Conference December 16-171 1963.
washington: “Department of Commerce,’l963.

 

Zettel, R.M., and Carll, R.R. Summary Review of Metro-
Politan Area Transportation Studies in the U.S.
UnIVersity of California, I962.

 

 

Uppublished Material

 

Burkhalter, Bart. "Summaries of Growth Models", A
memorandum to Tri County Planning Commission,
Lansing,Michigan, March 12, 196M.

Farness, Sanford S. "The Planning Process and Environ-
mental Health", Paper presented at the Conference
on Environmental Health, University of Michigan,
March 23, 196M. (Mimeographed.)

Gifford, John V.- "The Nature of the City as a System."
A paper presented before the Bay Area Systems
Group of the Society for General Systems Research.
San Francisco, Nov. 15, 1961. (Mimeographed.)

Marquis, S. D. "A Systems Approach to Communities and
Plannin .Areas." Michigan State University, March 7,
1963. Mimeographed.)

80

Marquis, S. D. The Washington Transportation Plan:
Technics or_Politics, Excerpt—No. 77. ISchboI
of‘Urban Planning, Michigan State University,
1962. (Mimeographed.)

 

 

Memmott, Frederick W. The Substitutability of
Communications for Transportation. A Technical
Paper forCNiagara*Frontier_Transportation Study.
(Mimeographed.)

 

 

III II "II III III I

74 3384