navigation bar

Department of Transportation

NTL Menu

Building Orientation - A Supplement to the Pedestrian Environment

Click HERE ffor graphic.

Table of Contents

Summary                                                            1
Overview                                                           3
Household Travel Relationships                                     5
Modeling Household Travel                                          9
Conclusions                                                       13

List of   Tables

Table 1A  Distribution of Zones and Households                     5
Table 2A  Travel Mode Choices                                      6
Table 3A  Household VMT Model Predicted Impacts                    9
Table 4A  Equivalent Variable Impacts on VMT                      11
Table 5A  Alternative Regression Model Results                    14
Table 6A  Land Use Variable Correlation Coefficients              14

List of Figures

Figure 1A Non-Auto Modal Shares                                    6
Figure 2A Average Daily VMT                                        7
Figure 3A Impacts of Land Use Variables on Household VMT          10



This report furnishes information on the role of the built
environment in affecting travel behavior.  Specifically, it focuses
on the setback and building orientation of commercial structures,
as these features influence household vehicle miles of travel

The report is a supplement to The Pedestrian Environment (December
1993), which provides a more detailed explanation of research
methods and data used in this analysis.  As was done in the earlier
report, researchers have examined actual travel behavior by
households in the Portland metropolitan area to analyze
transportation/land use relationships.  In this supplemental
report, researchers defined a new variable not,previously used in
the statistical analysis.  Data for the age of all commercial
structures in three Portland metropolitan area counties were
aggregated to establish an index for each traffic zone in the
region measuring the proportion of all commercial structures in the
zone built before 1951.  The assumption behind the use of this
variable is that commercial structures built before that date are
typically built to the front of the private lot line, rather than
set back to allow for surface parking on private property.  Thus,
the age of the commercial structure serves as an indicator of
building orientation.

Researchers used data from Metro's geographic information system to
develop the values for this variable.  While building age data was
incomplete for some zones, over 90% of the household observations
used in ne Pedestrian Environment report were available for use in
this analysis.

The principal finding of this research is that the indicator used
for building orientation is statistically significant in explaining
observed variations in vehicle miles traveled (VMT) per household
in the Portland metropolitan region.  Variation in building
orientation at the zonal level can account for changes of 10% or
more in VMT per household, over the observed range of values of
zonal building orientation (age of structure) in this database.

In addition, the equations used in this research included a
variable for employment density at the zonal level.  Like the
indicator variable for building orientation, this variable was not
previously used in the analysis included in The Pedestrian
Environment report.  This measure of "mixed use" at the zonal or
neighborhood level was also statistically significant in explaining
observed variations in automobile dependence.

Building Orientation: Volume 4B                                    1


In the real world, outside of the laboratory of statistics, this
research suggests that a number of aspects of the built environment
work together to influence vehicle miles of travel and automobile
dependency- Building orientation and pedestrian orientation are
closely correlated.  Ordinances and policies which are designed to
regulate the built environment need to be drafted in a manner that
reflects these lessons learned from Portland's "traditional"

2                                    Building Orientation: Volume 4B


In prior research done for the LUTRAQ project, including that
presented in The Pedestrian Environment, Volume 4A, there is
substantial evidence indicating the influence of land use on travel
behavior.  One aspect not examined in the research completed to
date is the role of building orientation and building setback in
influencing travel mode choice and thus, vehicle miles of travel.

This aspect of the built environment is the subject of substantial
discussion in Oregon and elsewhere, as planners draft and implement
ordinances which are designed to reduce automobile use.  Casual
observation of pedestrian and travel behavior at large commercial
developments, with substantial setbacks from the public right-of-
way, suggests that the effect of numerous buildings being set back
from front lot lines and from one another is to increase the use of
automobiles, even for relatively short trips.  However, it has thus
far been difficult to estimate the effects of "traditional"
building orientation and setback in quantitative terms.  We only
know by observation that development in the automobile era
(essentially that development which has occurred since the end of
World War II) looks very different from commercial development
prior to that date, and the travel behavior in auto oriented
developments may be partly explained by this fact.

To analyze this relationship more systematically data was gathered
on the proportion of buildings in each traffic analysis zone (the
neighborhood-level areas at which traffic behavior is analyzed in
Portland's travel demand forecasting model).  The key assumption in
this analysis is that those structures built during or before 1950
were built in an era in which walking and public transit played
important roles in urban mobility.  While the private automobile
had already begun to influence land use, the design of commercial
structures prior to 1951 appears largely not to have been
influenced by this trend. (E.g., the first shopping centers in
America were built in the early 1950's).

Using data furnished by county assessors in Multnomah, Clackamas
and Washington counties, researchers established an index of the
proportion of buildings in each of the region's 400 traffic
analysis zones built on or before 1950.  This number, ranging from
0 to 100%, was used in a multiple regression model.

The results of the analysis are described below.

Building Orientation: Volume 4B                                    3


4                                    Building Orientation: Volume 4B

                                      Household Travel Relationships

Household Travel Relationships

Because commercial building data in certain zones was incomplete, a
small number of household observations were removed from the sample
used in The Pedestrian Environment report in order to develop the
regression model.  Of the 2421 households in the original sample,
2223 remained available for use in this research.  These households
reported a total of 13,788 trips, a decrease of 1,350 trips from
the number available in the previous regression analyses. 
Nevertheless, with over 90% of the households and over 90% of the
trips still available, the dataset was sound enough for analytic

                              Table 1A
          Distribution of Zones and Households by Share of
                    Pre-1951 Commercial Buildings

     ZONAL SHARE OF PRE-1951       NUMBER         NUMBER OF

          0%                        98             546
          1-20%                     31             263
          21-40%                    58             394
          41-60%                    63             504
          61-80%                    43             343
          81-100%                   26             178

          Totals*                  319            2,228

          16 of the 400 Transportation Analysis Zones are
considered external to tho Portland Metropolitan Area and
comprehensive building age data was not available for 65 of the
remaining 384 zones.

Table 2A exhibits travel mode choice data for those trips.  As
shown there, and in Figure 1A, the number of trips made by transit,
and on foot or by bicycle, appears to increase steadily as the
proportion of buildings in the neighborhood oriented toward the
street (i.e., built before 1951) increases.  In the neighborhoods
with the newest commercial development, fewer than 3% of the
reported trips are made by transit and fewer than 2% are made on
foot.  At the other extreme, in those analysis zones or
neighborhoods in which 81 to 100% of the buildings are oriented
toward the street (built before 1951) transit and nonmotorized
trips both exceed 10% of all reported trips.  Furthermore, this
relationship holds across each of the sets of neighborhoods

Building Orientation: Volume 4B                                   5

Click HERE ffor graphic.

Click HERE ffor graphic.

Figure 2A presents the vehicle miles traveled (VMT) of residents in
these zones.  As shown in the graphic, households in zones where
most or all commercial buildings are set back from the street,
typically drive over 50% more miles per day than households in
zones where most of the buildings are oriented toward the street.

6                                  Building Orientation: Volume 4B

                                      Household Travel Relationships

Click HERE ffor graphic.

The descriptive data presented in this graphic, of course, do not
include controls for the various social or economic attributes
which have been shown to influence travel behavior.  The results of
including controls of this kind will be discussed below.  Further,
the relationship between household VMT and building orientation is
indirect.  The effect of building orientation at the neighborhood
level would be felt most directly in the form of,vehicle trips
eliminated and replaced by nonmotorized trips or by transit trips. 
Also, the correlation between building age and several other
neighborhood land use variables, such as household density, clearly
effect the relationships displayed in this figure.  A multiple
regression (discussed below) was successful in sorting out these

Building Orientation: Volume 4B                                   7

Household Travel Relationships

8                                    Building Orientation: Volume 4B

                                           Modeling Household Travel

Modeling Household Travel

The researchers made use of a multiple regression model similar to
those used in The Pedestrian Environment report to measure the
individual effects of several land use and socioeconomic variables
on vehicle miles traveled by Portland area households.  The
variables included in this analysis are shown in Table 3A. 
Household variables include the number of persons per household,
the average household income, the number of cars available, and the
number of employed individuals, as well as the average age of
household members.  In addition, the equation includes four
zonal/neighborhood land use variables.  These are a measure of
residential density, a measure of employment density, a measure of
automobile accessibility to employment within the region, and the
indicator of building orientation (the proportion of commercial
buildings within the zone built on or before 1950).

Click HERE ffor graphic.

The building orientation indicator was used in place of the
"pedestrian environmental factor" (PEF)--the index of pedestrian
friendliness used in The Pedestrian Environment report regressions. 
Building orientation toward the street usually occurs in
conjunction with the indicators used to establish the PEF index
(i.e., street connectivity, sidewalk continuity, ease of street
crossings, and topography).  Thus, it is statistically correlated
with the PEF variable.

All 9 of the variables used in the regression analysis were
statistically significant in explaining observed variation in
household vehicle miles of travel (see

Building Orientation: Volume 4B                                  9

Modeling Household Travel

The Pedestrian Environment report for a full explanation of the
nature of statistical significance).  Both the household and the
land use coefficients had the expected signs.  The coefficients
were quite similar to those observed for the same variables in the
regressions included in The Pedestrian Environment report.

The equations can best be understood in the terms presented in
Table 3A and Figure 3A.  There, specific measures for each of the
variables are presented in terms of their effect on household VMT. 
Thus, for every $5,000 increase of household income, the model
suggests an increase of 0.8 miles per day in vehicle travel. 
Increases in household size, automobile ownership and workers per
household also had similar, predictable effects.

Click HERE ffor graphic.

Among the land use variables, an increase in residential density
from 3 to 4 households per zonal acre corresponded to a decrease of
0.8 miles in household vehicle travel.  A 20,000 increase in the
number of jobs accessible within 30 minutes travel by automobile
had a similar effect.

A variable not included in the previous regression models, a
measure of employment density, was statistically significant as
well.  This measure of zonal or neighborhood based employment can
be seen as an indicator of mixed use in the neighborhood.  The more
employees found within the zone of residence of a household, the
more opportunities for short trips or for changes in mode choice
from auto to other modes.

10                                    Building Orientation Volume 4B

                                           Modeling Household Travel

Lastly, the measure of building orientation (building age) also
made a statistically significant contribution to the equation.  An
increase of 30 percentage points in the proportion of commercial
buildings in the zone built prior to 1950 corresponded to a
decrease of 1.3 miles (approximately 5%) in the household daily

Building orientation (building age), in the context of the
regression equation, can explain a change of 10% in VMT over a 63
percentage point change in the proportion of buildings in the zone
built in 1950 or before.  A 63 percentage point swing in building
orientation represents a change from the very lowest to the very
highest quintile of the 400 traffic analysis zones included in this
analysis.  As is the case for the equations included in The
Pedestrian Environment report, the correlation between zonal land
use variables should be noted.

                              Table 4A

       Equivalent Variable Impacts on VMT Per Household Person


-    A 63 Point Increase in the Zonal Share of Pre-I 951 Commercial
     Buildings, or

-    A $17,500 Decrease in Household Income, or

-    A 1.5 Car Decrease in the Number of Cars per Household, or

-    An Increase from 2 to 5 Households per Zonal Acre, or

-    A 70,000 Increase in Employment Accessible by Auto in 30
     Minutes, or

-    An Increase from 1 to 50 Employees per Zonal Acre

Building Orientation: Volume 4B                                   11

Modeling Household Travel

12                                   Building Orientation: Volume 4B


In an equation in which a set of socioeconomic variables and a set
of land use variables have been combined, building orientation
(building age) has been shown to be a statistically significant
influence on household vehicle miles of travel.  The results of
this research are significant in the real world of public policy
for the following reasons:

1.   The research demonstrates that building orientation, as one of
     several land use variables which can be influenced by public
     policy, has a statistically significant impact on household
     vehicle miles of travel, an important measure of travel

2.   Employment density, household density, overall urban form,
     (expressed as ease of accessibility to employment), and
     building orientation (expressed as age of commercial
     structures), intermingle in the real world and in this
     statistical research.  While it is important to identify the
     significance of each attribute in affecting travel behavior,
     it is equally important to note the significance of their
     effect as a group.  The reader should examine other reports
     completed as part of the LUTRAQ Project for further
     information on the effect of land use on travel.

Building Orientation: Volume 4B                                   13


Click HERE ffor graphic.

Click HERE ffor graphic.

14                                    Building Orientation Volume 4B

About the Authors of This Volume

Parsons Brinckerhoff Quade & Douglas, Inc.

Parsons Brinckerhoff Quade & Douglas, Inc. is the leading provider
of transit p and design services in the United States.  The firm
has been involved in more than 75 percent of the nations light rail
transit systems in operation or under construction today.  The
firm's architects have developed concepts for or designed over 200
transit stations in the last ten years.

Brent Baker, of the firm's Seattle office, is the principal author
of this report.  Samuel Seskin, Cathy Strombom and Youssef Dehghani
contributed to the research.

Building Orientation: Volume 4B                                   15

16                                   Building Orientation: Volume 4B

Jump To Top