OVERVIEW

The key goal of transit-oriented developments (TOD) is to provide an environment in which transit, walking, and some bicycling are the primary travel modes to reach a significant amount of one's daily needs and destinations. Within TOD (and smart growth more generally), there are three core elements to consider: density, land use mixture, and mobility infrastructure. The core theoretical image of this urban form is of a transit stop surrounded by quarter-mile concentric rings. Within the quarter- and half-mile rings, development is relatively dense, land uses are mixed, and there exists a mobility infrastructure that supports pedestrian movement. This research is predominantly focused on the third element-the pedestrian mobility infrastructure-and how the theory of concentric rings of walkability is translated into practice. Secondarily, this research also focuses on the transportation modal split around light rail transit stops between 1990 and 2000.

In translating the hypothetical concentric circles of walkability into an analysis of existing urban form around transit stops, the theory becomes compromised in two key ways. First, basing walking distances within concentric circles ignores the fact that people are not free to travel in any direction, but must travel along pathways. A quarter-mile zone from a transit stop based on walking would therefore not be a perfect circle. Understanding the actual shape of a walkable quarter-mile zone can give insight into the general pedestrian-friendliness of the urban form surrounding transit stops. Second, not all potential pedestrian pathways are of equal accessibility. If using a street network as a proxy for pedestrian mobility, it is clear that the existing hierarchy of street types (minor roads, arterials, major roads) is also relevant for pedestrians, and likely with an inverse relationship. That is, roads designated as appropriate for heavy volumes of automobiles may simultaneously be less desirable for pedestrian travel.

This research and report focuses on these two key elements (pattern and connectedness of the street network and the hierarchy of street types) in looking at eight TOD areas (four in Portland and four in Silicon Valley). Moreover, the change in urban form over time is incorporated by looking at the urban form before and after the development of the light rail systems in each region. An analysis of transit utilization mode of travel to work over time complements the urban form analysis. Finally, a spatial-temporal analysis of basic socio-demographic characteristics is presented for each region.

INTRODUCTION

The basic premise of transit-oriented developments (and smart growth efforts in general) is that a variety of land use factors affect travel patterns including density, land use mix, roadway connectivity and design, parking facilities design and building design.1 Three key elements of TOD areas are appropriate density, diversity, and design of a community-the three D's of the built environment.2 That is, the form, spatial location, and concentration of activities within an urban environment can influence transit ridership.

It is believed that good urban form can lead to a reduction of total transportation costs and auto usage, resulting in more livable communities.3 For example, Bernick and Cervero found that the residents of more pedestrian-friendly neighborhoods were more likely to go by foot to the market.4 Handy found that residents living in "traditional neighborhoods" made two-to-four more walk/bicycle trips per week to neighborhood stores than those living in nearby areas that were served mainly by automobile-oriented strip retail establishments, although there were similar rates of auto travel to regional shopping malls. A good walkable urban form, therefore, can be a key contributor to local mobility.5 And because TOD areas represent both local and regional mobility, the streets and character of the immediate surroundings, the neighborhood linkage with the transit stop, as well as the location of the neighborhood within the larger region may influence regional household travel behavior for neighborhood residents. Thus, the urban form at a neighborhood scale is an important variable that will allow a resident to exercise a non-automotive transportation choice, if such options are available.

THE MOBILITY INFRASTRUCTURE AND WALKABILITY

Because transit riders begin and end their trips as pedestrians, the environment where people walk to and from transit facilities is a significant part of the overall transit experience. Common sense suggests that an unattractive or unsafe walking environment discourages people from using transit. Conversely, a safer and more appealing pedestrian environment may increase transit ridership.6

Often lost in the debate about transit-oriented development is the walking environment surrounding transit stops that allows users (and potential users) to access the transit system and local amenities surrounding individual stops. The larger debate tends to concentrate on the right mix of uses and density surrounding transit stops with the ultimate goal of understanding the impacts of land use on transit ridership. Yet, the capacity for transit users to walk to and from their transit point of entry is a critical component of the overall TOD concept. Pedestrian impediments to reaching a transit station become significant impediments to transit usage. That is, "Since all transit trips involve some degree of walking, it follows that transit-friendly environments must also be pedestrian friendly." 7

There are many potential pedestrian conditions that enhance or impede one's ability or desire to reach a transit stop, including safety issues, the existence of appropriate paths, and an interesting viewscape at pedestrian scale.8 One central component of a transit stop's walkable service area, and one of the foci of this research, may be the most basic and central component of a walkable environment: the quality, connectivity, and accessibility of the road network. The road network represents the basic skeleton of the urban form, creating the range of opportunities and path choice that can make walking more or less desirable. In addition, the pattern and form of the street network defines the structure in which infill of the physical environment can take place.

The need to have deeper understanding of urban form and its impacts on local accessibility is crucial in land use planning. The mobility infrastructure serves as the "skeleton" of the community as it creates the routes for accessibility, places for physical structures, and the forums for community interaction; in sum, this skeleton is a key to understanding urban form. Southworth and Ben-Joseph observe that residential streets provide the public framework that shapes urban form and guides neighborhood life. From this perspective, then, Southworth and Ben-Joseph argue that the significant contemporary urban issues of today-congestion, pollution and community isolation-are inextricably linked to residential roads patterns. 9

Calthorpe and Poticha state that a reduced or non-existent hierarchy of internal streets is the desired internal network type within an authentic TOD. They describe how streets designed for high automobile speeds are inappropriate in a mixed-use pedestrian-focused zone. In essence, they argue that the automobile and the pedestrian should be equals on the network, each having a place to traverse. 10

Certain auto-oriented roads (freeways and major arterials) present impedances to pedestrians because the scale and feel of such roads negatively impact one's ability or desire to cross or travel along them. By including the concept of impedance into the GIS-based qualitative visualization and quantitative analyses, the road network, route choice, intersection concentrations, and pedestrian-scaled environments can be more accurately identified and measured. Measuring the walkable environment around TOD areas can lead to an intra-urban level of analysis that allows one to capture the spatial qualities of the Elemental City perspective.11

The safest environment for pedestrians also should combine short blocks and frequent cross streets in order to create the maximum number of options for travel route and the most direct routes that have little or no out-of-direction travel.12 Ewing suggests that a greater number of intersections give pedestrians an enhanced sense of freedom and control as they are not forced to take the same path to a given destination time after time. He also states that more intersections make a walk seem more eventful, since it is punctuated by frequent crossing of streets and that additional intersections may shorten the sense of elapsed time on walk trips since progress is judged to some extent against the milestones of reaching the next intersection. Such measures have been recently used to assess the level of sprawl across metropolitan areas.13 At a very fine grain, and in an effort to create distinctions between types of development patterns, Jacobs provides comparative measurements of such things as numbers of intersections and cul-de-sacs across a small geographical area.14 Krizek, in looking at more of a neighborhood scale, found that people who live in more walkable areas, referred to as areas with good "neighborhood accessibility," are more likely to walk and use transit than those who live in more traditional auto-oriented environments.15 And increasingly, such concepts are being used to understand the connection between the built environment and physical activity-a connection largely dependant on the walkable nature of local neighborhoods.16 Thus, measuring the walking infrastructure-the routes and choices available to pedestrians-at a fine grain is an important component in identifying or evaluating the likely potential and range of local, destination-oriented walking.

SPATIAL INDICATORS OF URBAN FORM

The urban form around a TOD is of key importance and the street network often acts as the skeleton for this urban form. The work on quantitatively analyzing the walkable urban skeleton has recently been pursued by a variety of scholars. Table 1 lists a series of spatial measures used to understand connectivity at a variety of spatial scales.17

Table 1 Measures of Connectivity Used in Research Literature

Table Source: Dill, J. "Measuring Network Connectivity for Bicycling and Walking." Paper presented at the ACSP-AESOP, Leuven, Belgium. July 9, 2003. Used with author's permission.

Many of these street network-based analyses tend to treat all streets as equals, despite their different uses, qualities, and traffic volumes. For example, Krizek's innovative analysis of travel behavior using local measures of neighborhood accessibility looks only to the presence, absence, or concentration of certain street network characteristics, assuming that all streets and intersections are of equal quality and use. 18

The methods presented in the following chapter begin to make some distinction in street type, thereby influencing other measures of walkability such as intersection and dead-end densities. Refining how these basic components of the street network are modeled is needed for better planning (or, more likely, evaluation of past planning) of TOD (or smart growth) principles.

VISUALIZING TOD AREAS

Visualizing this urban skeleton is also an important component of understanding walkability. Lynch identified five basic components of urban form-paths, edges, districts, nodes, landmarks-each of which can be visualized in terms of a walkable urban network. Paths can be thought of as minor roads; edges equate to freeways or other large roads (e.g., arterials) that impede pedestrian movement; districts can represent concentrated zones of walkable urban form; nodes represent street intersections; and landmarks represent key origins or destinations, such as a transit stop. Each of these elements can be measured and viewed spatially to present a qualitative opportunity to assess local environments in terms of walkability.19

In terms of pure visualization, Jacobs presents a unique method of visualizing the urban form by using a figure-ground technique of displaying the road skeleton that makes up different urban environments. Using the same scale and same visualization techniques, Jacobs shows the importance of the street network in framing and supporting walkable urban forms.20 Southworth, et al. extends Jacobs' work by incorporating visual examinations of intersection patterns and quantifying several elements of the street network, leading to a spectrum of identifiable development types, based solely on the nature of the road network.21 Bossard takes a different approach, focusing on visualizing neighborhoods with TOD potential using a series of schema to conduct visual, spatial analysis and comparative socio-demographic analyses. He focuses on using small multiple images to enhance the simultaneous visual analysis of multiple variables.22

Thus, in analyzing TOD areas from a spatial approach, this research focuses on the mobility infrastructure of TOD areas, utilizing a variety of spatial indicators to assess that infrastructure from a pedestrian perspective, and utilizes key visualization techniques to evaluate the performance of TOD areas statically and temporally over time.

study design

This study's main objectives are to develop and utilize spatial indicators to measure the local walkable form around TOD areas, the change of this form over time, and the linkage between TOD development and transit usage. Specifically, this project was commissioned to conduct five main types of analyses on a total of four TOD areas (see Figure 2 for a visual diagram of the research design):

1. Street network analysis: The street network within the TOD areas was to be analyzed using a variety of spatial variables including the density of intersections and the length of road network per square mile. These measures provide indicators of accessibility-places with higher intersection densities and higher road lengths per square mile can be considered more walkable and transit-friendly because they are characteristics of places with more path choice. This analysis was to be conducted for the TOD sites prior to and after construction/designation.

2. Ped-shed analysis: Ped-sheds (re-named in this analysis as pedestrian catchment areas (PCA)) measure the accessibility of a given location based on a ratio of Euclidean distance to street network distance. This analysis calculates a number that represents how walkable a space is. This analysis was to be conducted for the TOD areas prior to and after construction/designation.

3. Transit ridership analysis: A key element of TOD areas is the utilization of public transit. Data on passenger loading and unloading for specific transit stops within TOD sites were to be analyzed (subject to availability) along the life of their existence. Using temporal data of this type can help one understand how transit usage has changed with the adoption of specific TOD sites. Census-based transit utilization at a variety of spatial scales was also to be used in a pre/post construction manner to understand the changes in travel behavior over time.

4. Street speeds analysis: Speeds (using road type as a proxy for actual speeds) along the road network within the case study TOD areas were to be analyzed to determine their consistency with walkability. Places with high automobile speeds are characteristic of locations more hostile to pedestrians. Pedestrian scale is important for transit ridership because TOD principles suggest that people walk to and from transit stops within TOD areas.

5. Socio-demographic analysis: A socio-demographic analysis of the TOD case study sites was to be conducted using 1990 and 2000 census data in order to compare the population and housing characteristics of the TOD sites. Examples of some of the socio-demographic variables include average age and age distributions, racial mix, and average income and income spread, among others.

Figure 1 Research Figure Schematic (note that this research includes a comparison of eight cases, although only 3 are depicted in the image above.)

Walkable Urban Form

The street network can provide an indicator of a variety of pedestrian-related conditions, including areas with more path choice, fewer miles of auto-centric roads, and greater connectivity. In addition to refining the measurements to more accurately capture the urban feel a pedestrian would experience in a given environment, it is important that analyses be conducted in spatially explicit means and that results be analyzed at both a visual and a quantitative level. The patterns, forms, concentrations, and absences that result from a street and intersection analysis can be clearly understood from a visual analysis of the data. Understanding the spatial relationships between transit stops and the surrounding urban form on a map can provide clear insight into the pedestrian appropriateness of the transit environment. Moreover, comparisons across TOD areas are easily conducted, especially when spatial images are created with similar reference scales and symbology.

Quantitative analysis provides another means by which TOD areas can be evaluated, both individually and comparatively to other locations. Quantitative measures can lead to the development of acceptable thresholds of certain criteria, for example the minimum density of intersections per square mile that results in good pedestrian urban form. Quantitative analysis also provides a means for comparing sites across space and time, to consistently rank and compare performance without the bias that may result from visual, qualitative inspections. Developing good visual and quantitative measures of walkability can be a key component in planning and evaluating a variety of smart growth concepts.

In this light, six specific measures have been developed to quantitatively and visually examine the quality, proximity, and connectivity of the underlying urban skeleton in terms important to the principles articulated for smart growth TOD communities (see Table 4). In terms of quality, a street classification analysis looks at the quantity and location of pedestrian-friendly and pedestrian-hostile street types. In terms of proximity, pedestrian and impeded pedestrian catchment areas have been identified, giving insight into the likely walkable zone surrounding a transit stop given the existing street network and the street types in close proximity to the transit stop. For connectivity, intersection density analysis, impeded intersection density analysis, and intersection surfaces have been developed that give insight into the areas of good and poor pedestrian environments.

Table 2 Measurement Domains and Techniques

Each of these measures are more deeply defined and illustrated in the section titled "Measures and Methods."

Change Over Time

Often lacking in TOD analyses is the longitudinal component of change over time. While the theory of TOD areas speaks to sophisticated integration of a variety of land use, commercial, transportation, and social goals, the reality is that retrofitting existing urban or suburban spaces or even developing anew within a greenfield, the process to realize the full TOD potential takes time. Land uses change slowly, commercial investment takes time to occur and adjust to local conditions, and local populations grow over time. Thus, while taking a static reading of current conditions is informative and allows us to evaluate reality against theory, looking at change over time allows us to see if things are moving toward or away from TOD goals.

For this study, the 1990 and 2000 decennial censuses provide logical anchor points for the temporal analysis because the light rail systems in both Portland and Silicon Valley were built out in the interim years. Using these two census points, socio-demographic and transportation-related data can be analyzed relating to pre-construction and post-construction periods in terms of when the light rail and corresponding TOD areas were built. In terms of the walkability analyses, the TIGER 23 street centerline data for 1992 and 2002 were used for the pre- and post-construction analyses. 24

Case Study Sites

Eight case study sites were chosen for analysis, four in the Portland region and four in Silicon Valley. The specific locations were chosen because they represented multiple desirable characteristics; most importantly that they were specifically designated as TOD areas, and they represent a range of development types. Table 3 lists the specific case study sites and the corresponding type of development they represent. Figure 2 and Figure 3 show overviews of the Portland and Silicon Valley regions, and the location of the case study TOD areas. These sites were chosen partly to reflect differing environments within which smart growth concepts are being implemented and partly to reflect the reliability of the measures across unique environments. Beaverton (OR) and Mountain View (CA) represent an in-fill TOD located within a very auto-centric, commercial shopping district. Orenco Station (OR) and Whisman (CA) are greenfield TOD areas master planned and implemented through the conversion of open space to mixed land uses. The Lloyd Center (OR) and Bonaventura (CA) are more urban commercial TOD areas, relatively close to downtown areas, and Gresham (OR) and Japantown/Ayer (CA) are TOD areas located in more traditional, pre-WWII gridded street neighborhoods.

Table 3 Case Study Sites

Figure 2 Portland Locator Map

Figure 3 Silicon Valley Locator Map

Finally, some have called for additional TOD case studies to build our collective knowledge of these development types. This research, in part, addresses that desire.25 Thus, the discussion below is meant to promote, visualize, and quantify a series of measures that can be used to plan or evaluate development patterns or urban form that are supposedly based, at least partially, on pedestrian-oriented principles. Moreover, Talen suggests that the ideas and plans that characterize smart growth have outpaced the ability of planners and designers to measure and quantify them.26 This paper presents techniques designed to address some of the measurement challenges of smart growth, and to do it in a way that is generalizable, accessible, and useful to scholars and practitioners who are seriously engaging in these new development principles. The cases presented below are not meant to be used as critiques of specific places, although clearly, part of the process of understanding the measures is to relate their results to the places they measure in an evaluative fashion.

measures and methods

A total of eight case study sites have been analyzed: four in the Portland region and four in Silicon Valley. Ten unique analyses were conducted to understand the demographics, ridership performance, and walkability of the urban form surrounding the case study transit stops (see Table 4). All analyses were conducted over two distinct points of time, using the time frame of pre- and post-TOD construction as the central determinant of the selected timeframes. Eight of the ten analyses apply to urban form and were conducted at two geographical scales (quarter- and half-mile) to understand how theoretical conceptions of TOD play out at actual case study sites. Thus, for the urban form analysis, a total of 256 individual data points were derived (eight TOD areas x 2 time periods x 2 geographic scales x 8 variables). A description of each of these variables and analysis methods is presented in more detail below.

Table 4 Primary Analysis Categories

1990 AND 2000 CENSUS ANALYSIS

Spatial Data

Census block groups for both 1990 and 2000 formed the basis for the census analyses. The data was divided into nine spatial units of analysis in the Portland area, including:

1. Orenco Station-Block groups that intersected a 1/4 mile zone of the transit stop were selected.

2. Beaverton-Block groups that intersected a 1/4 mile zone of the transit stop were selected.

3. Lloyd-Block groups that intersected a 1/4 mile zone of the transit stop were selected.

4. Gresham-Block groups that intersected a 1/4 mile zone of the transit stop were selected.

5. LRT-All block groups within 1/4 mile of the entire light rail line were selected; this gives an indication of the general ridership figures for all people living in close proximity to the rail line.

6. Non-LRT-All block groups within the urban growth boundary (UGB), but more than a 1/4 mile from the light rail line were selected; this gives the breakdown on people who do not live within easy walking distance of the light rail.

7. UGB-All block groups within or intersecting the UGB.

8. Non-UGB-All block groups outside of the UGB but within the Tri-County area.

9. Tri-County-All block groups within the three-county Portland area.

Each of these nine units of analysis was spatially created via GIS. Block groups that intersected or were within any of the zones of interest were selected for analysis. Figure 4 illustrates these units of analysis for the Portland area. Summary Tape File 3 (STF3) was used for 1990 data analysis and Summary File 3 (SF3) was used for 2000 census analyses. Figure 5 illustrates the Silicon Valley units of analysis.

Figure 4 Portland Census-Based Spatial Units of Analysis

socio-Demographic Attribute Data

A socio-demographic analysis of the TOD case study sites was conducted using 1990 and 2000 census data in order to compare the population and housing characteristics of the TOD sites. The following specific variables were used:

Table 5 Census Variables Used

*derived through a GIS-based spatial calculation

**derived by subtracting the white population from the total population

Transit Ridership

Transit ridership was derived in two ways: 1) via the 1990 and 2000 census and 2) from count data provided by the local transit authority. 27 The census data variable "Means to Work-Workers 16+" was used to compare journey to work modes of travel between the two censuses.28 These two data sets are used as separate entities (rather than being integrated into a single analysis variable), meant to provide two different understandings of transit usage in the TOD areas. The same spatial units of analysis as described above were used.

WALKABLE URBAN FORM

Several types of urban form analyses were conducted in order to primarily understand the walkability of each TOD. The focus was on the street network as the primary means of pedestrian accessibility and special attention was given to the hierarchy of roads within a given TOD to test their consistency with TOD walkability principles. The three major categories for these analyses are street classification analysis, intersection analysis, and catchment area analysis (see Table 6). Each is described in more detail below.

Table 6 Walkability Variables

STREET CLASSIFICATION ANALYSIS

Overview

Street Classification Analysis is an evaluation and categorization of street type and purpose along the road network within TOD areas. This analysis provides insight into the basic quality of certain paths and reflects the hierarchy of road types within the study zones.

Background

Locales with high automobile speeds or large volumes of traffic are characteristic of locations hostile to pedestrians. Peter Calthorpe has recently called for a change in how we classify roads, from an auto-centric design focus (minor, feeder, and arterial) to one that reflects accessibility principals. By identifying and classifying road types with relevant typology-ones that reflect accessibility design principles-researchers can make a more accurate assessment of road functionality. 29

The Street Classification Analysis addresses this request by defining and exploring the relationship of "Impedance Roads," or hostile roads, and "Accessible Roads," or pedestrian-friendly roadways. An impedance road may spatially divide a community, splitting it into segments via a road that acts as a barrier. Identifying where these roads are reveals the spatial externality of the road placement. By spatially displaying where these roads are in map form, with accompanied metrics on quantity or share of road types, it is possible to create an accessibility profile base for impedance values.30

Figure 6 illustrates the process and results of identifying the variety of available paths. The top image represents a complete street network, the middle image highlights the location of impedance roads (roads classified as freeways or major arterials), and the bottom image illustrates the impact of removing impedance roads on using the road network to represent a pedestrian network. These images demonstrate the direct impact of auto-centric roads on pedestrian mobility, particularly evident by the increased number of dangling road segments, disconnected paths, and longer block faces.

Unclassified Network
Network Classified Identifying Impedance Roads
Impact of Classification on Pedestrian Network

Figure 6 Illustration of the Impact of Impedance Roads

INTERSECTION INTENSITY

Overview

The intersection intensity analysis examines the street network within the TOD sites based on the spatial location of certain types of intersections in order to capture the grain (density of intersections) and the interconnectedness (types of intersection) of a neighborhood.

Background

Intersections are a core set of data because they represent the number of choices available to a pedestrian and, from a spacial perspective, how these choices are arranged throughout the study zones. These measures provide indicators of accessibility. Areas with higher intersection densities, and/or more desirable intersection types (three- or four-way), can be considered more walkable because they are characteristic of places with a greater number of path choices for the pedestrians.

Theoretically, there should be a match between the location of the optimal pedestrian form and where the transit stop is located in order to maximize the pedestrian element of transit usage. Statistics such as intersection density (intersections per square mile) are important and increasingly used as a variable or urban analysis, although such statistics are often not used in spatially explicit ways. In urban form analysis it is valuable to know not only how many intersections there are per square mile, but also where the density of intersections fall within the study area. To understand more fully how these elements are related, two different types of analyses are presented below. The first method analyzes the concentration and location of "good" intersections (three- and four-way) and the location of dead-ends; the former represents environments with good pedestrian path choice and the latter representing a lack of mobility options. This analysis has two components as well, which are represented in the images on the left and right in Figure 7. The left image shows the location of intersections and dead-ends based on the assumption that all roads are equal. The image on the right is what results when the impedance roads are removed. Removing the impedance roads has two key effects. First, "good" intersections are reduced because a crossing of an impedance road no longer counts as an intersection: From a pedestrian point of view, reaching a major auto-centric road usually does not imply a full path choice-a pedestrian may choose to cross such a road, but is unlikely to travel along it. Second, dead-ends are increased; when a pedestrian road terminates at an impedance road, it can be considered a dead-end from the pedestrian point of view.

All Intersections	Impedance-Based Intersections

Figure 7 Intersection Comparison

The second intersection analysis method is purely visual in nature and displays density surfaces of desirable intersection types. A density surface map resembles a national weather map, but instead of showing areas of hot and cold weather, it shows areas of high and low intensities of intersections (see Figure 8). By creating density surfaces of the intersections, one can build on the intersection analysis by creating a qualitative, visual rendering of where the optimal intersections are found and how they relate to the spatial layout of the community. In such a manner, the grain of the community is visually apparent and available to assist in determining the level of connectivity and adjacency of the transit station to the larger community.

Figure 8 Intersection Surface Map (wth a local street network on top)

PEDESTRIAN CATCHMENT AND IMPEDED PEDESTRIAN CATCHMENT AREAS

"Pedestrian Catchment Areas," (also known as Ped-Sheds) are theoretical walkable zones that can be mapped to show the actual area and network within a five-minute (quarter-mile) or ten-minute (half-mile) walking distance from a transit stop. The data is presented as a ratio between the Euclidean distance and the network distance from a given point (e.g., transit station). The resulting maps are also highly visual estimates of an area's walkability. 31

The Pedestrian Catchment Area (PCA) methodology focuses on capturing the coverage of a street network within the designated TOD and determining how accommodating that network is for pedestrian movement. The basic calculation of a PCA is to divide the area of a quarter-mile or half-mile circle by the area of the polygon that results by traveling a quarter- or half-mile from a transit stop along the street network (see Figure 9).

Figure 9 Diagram of Pedestrian Catchment Area Ratio Calculation

A pedestrian catchment area score of .60 or higher has been presented as reflecting a walkable network, meaning that a majority of the theoretical walkable area can actually be reached by moving along the actual street network.32 A score of .80 would indicate a comprehensive grid network that encompassed an entire study zone and a score less than 0.30 reflects an inaccessible walking environment. Pedestrian catchment area ratios have been calculated in both the standard way (as described above) and by using the refined street network where "Impedance Roads" have been removed. This impeded pedestrian catchment area (IPCA) represents a new way to calculate and visualize an area that a pedestrian is potentially able to travel. That is, removing the impedance roads from the pedestrian catchment area analysis, it is possible to reflect the ability of a pedestrian to cross (assuming that there are crossing amenities) an impedance road, but not necessarily travel along it. By removing roads that are hostile to pedestrians to either cross or walk along, the IPCA provides an improved capacity to capture the pedestrian zone of a transit stop.

The decision of what is or is not an impedance road can be a difficult task to accomplish using existing spatially-referenced data, as a variety of facts, beyond the volume of automobiles, can effect road impedance. These include:

1. Presence or absence of sidewalks

2. Width of sidewalks

3. Nature of separation of sidewalk from moving traffic

a. Sidewalk setback from road

b. Trees in sidewalk setback from road

c. Presence of parking lane at edge of road

d. Presence of guard rails/barriers/fences along edge of roads.

4. Ease of crossing road

a. Traffic light or stop sign and clear pedestrian walkway

b. Presence of safety islands and bulb outs to reduce length of exposed roadway crossings.

Nonetheless, the methods described in this chapter represent a series of techniques that can be readily and easily applied to any area in the country because the street network, as well as an embedded classification of that network, exists for every municipality in the United States and is available at no cost over the Internet. These techniques can eventually be more accurately applied when more detailed data on pedestrian networks exist, but in the interim, the approaches covered in this chapter represent a useful method for evaluating and planning how well the theory of TOD area development matches with the implementation in practice.

CASE STUDIES

PORTLAND METROPOLITAN AREA

The rationale for selecting Portland as the case study region stems from their planning approach to regional growth management and adherence to TOD development concepts. In 1990, voters gave Metro-the regional managing agency of Portland, Oregon-the authority to adopt a regional planning framework in accordance with the broad principals of smart growth, and specific detail of a TOD. This regional planning framework is called the Metro 2040 Growth Concept. A key component of the Metro 2040 Growth Concept focuses on creating compact communities around transit and redeveloping around existing station communities. The plans calls for an aggressive expansion of regional light rail (MAX Line) and bus service, with expected mode share splits for regional transit use to grow by over 300 percent. 33

A specific feature of the regional planning effort is the Transit Station Area Planning (TSAP) program, which is a collaborative effort between Tri-Met (the regional transportation authority), Metro (the regional growth management body), the cities of Portland and Gresham, and Multnomah County (including affected cities contained within it). The goals of TSAP are to build support for TOD areas along the rail line and to promote opportunities for increasing the system's ridership. To date, the TSAP program has included market studies, coordination with other regional planning efforts, detailed station area plans, and design guidelines.34 Included in the suite of objectives are:

Rezoning station areas to transit supportive uses,

Setting of minimum residential and commercial densities, and

Application of a design overlay that requires pedestrian orientation.

The first segment, Eastside MAX, runs 15 miles east from downtown Portland to Gresham; it was completed in 1986. The second segment, Westside MAX, was built through wide stretches of undeveloped land from Portland city center to Hillsboro; it opened in 1998. Currently the entire system has 50 stations. As of September 2003, Tri-Met ridership has increased for 15 consecutive years and the MAX now has an average daily ridership of 79,600 boardings. 35

SILICON VALLEY AREA

Paying service began on the Santa Clara Valley Transportation Authority's (VTA) light rail system in December 1987. The original nine-mile segment from Baypoint Station, at the extreme north of the city of Santa Clara, south through downtown San Jose was completed in June 1988. Service to the Tamien Station two miles south of downtown San Jose began in August 1990. The entire 20.8-mile line was completed in April 1991. The Tasman West line, which connects Mountain View to the existing light rail service, was completed in December 1999. Construction is already underway on a new line, the Tasman East/Capitol extension from Baypoint station north to the city of Milpitas. Currently, 46 stations make up the combined lines of the VTA light rail system.

When compared with other light rail lines in the U.S., San Jose's light rail vehicles appear quite underutilized. On average, San Jose light rail vehicles carry an average of 14.8 people, which is less than 57 percent of the national average. In addition, in 2000, the VTA carried 1,750 passengers per mile, less than half the national average (4,400) and only about a third of Portland's level (5,937).

READING THE WALKABILITY MAP

Each map below, unless otherwise specified, reflects post-construction data from the year 2002.36 In order to streamline the presentation, maps for the pre-construction period are not shown, except to demonstrate specific issues, in which case they will appear in the next chapter. The spatial map images are designed more to orient the reader to the analysis technique, than to be used as a visual analysis tool. Below the maps are tables of data that resulted from the spatial analyses; these tables do contain data of pre- and post-construction time periods. Each table further delineates the data into quarter-mile (0.00-0.25) and half-mile (0.00-0.50) distances. Data at the half-mile distance is inclusive of everything inside of that circle (i.e., it does not represent the unique band of space between a quarter- and a half-mile from the transit stop).

Each TOD is presented at a 1" = 8,000' scale (intersection surface maps are at 1" = 12,000'), so the spatial extent remains constant across images, enhancing the reliability of the measures across locations (and potentially across time). Maps should be viewed for the general patterns that emerge and to compare patterns across study sites in order to understand the impacts of the presence and location of impedance roads. Each map includes two circles, representing quarter-mile and half-mile radii from the transit stop. Outside of the circles, tax lots are shown in order to have some sense of the urban pattern beyond the traditional TOD walking distances. Within these circles, the background of the maps is a solid gray, enhancing the visualization of the key variable being represented, but sacrificing the underlying form communicated by tax lot size and location. Figure 10 displays the legend for the symbols used in the maps.

Figure 10 Map Symbols

PORTLAND: ORENCO STATION

Orenco Station is on the western portion of the MAX line and represents a greenfield development planned and built specifically with transit-oriented and walkability principles in mind. The existing built areas have won numerous design awards for their attention to neighborhood amenities, community form, and architectural style. Until 2003, most of what has been built begins about a quarter-mile north of the transit stop. Access to the stop from the built-out areas necessitates the crossing of a major east-west arterial road and then a walk through a series of undeveloped and generally unkempt lots along a road that will some day provide access to these lots. Figure 11 shows two photographs of the Orenco Station area. In 2003, the land immediately adjacent and south of the transit stop has begun to be aggressively developed at medium to high density, affording much better light rail access than the existing developed areas.

The view looking north from the transit stop to the existing Orenco Station community, which is quite far in the background. Until 2003, land adjacent to the stop to the south was vacant. Currently, it is being aggressively developed at high densities.		Medium- to high-density development typical within the Orenco Station neighborhood.

Figure 11 Orenco Station TOD

Demographics

In the transition from greenfield to TOD, Orenco Station underwent some fairly significant demographic changes (Table 7). Population density increased over 250 percent to 1,747 people per square mile. That population became more diverse as well, transforming from 97 percent white to 80 percent, including a five-percent increase in the Hispanic population. The median age over the decade dropped by three years with the predominant shift in age coming in an increase of people in the 18-44 age cohort. Over this period, average household income increased almost 40 percent to almost $62,000 in the year 2000.

Table 7 Socio-Demographic Characteristics of Orenco Station, 1990-2000

Transit Ridership-Census

In 1990, when Orenco Station was a mostly undeveloped greenfield site, the automobile accounted for 100 percent of all work trips (Table 8). In 2000, however, after the light rail line was extended outward and Orenco Station was developed as a TOD, automobile use decreased, accounting for only 86.5 percent of work trips. The new train accounted for about five percent of this change, while new bus service accounted for 2.5 percent, and biking and walking accounted for another 2.5 percent. In terms of TOD goals, the construction of Orenco Station changed the relative mobility choices of about seven-and-a-half percent of the population to either take the train, bike, or walk.

Table 8 Orenco Station, Means of Travel to Work, Workers 16+, 1990-2000

Transit Ridership-Metro

In the first four years of Orenco Station's existence, weekday boardings increased by 58 percent to an average of 735 boardings per day (Table 9). The land around the transit stop was still mostly undeveloped in 2002, and it is likely that, as this greenfield development matures, transit ridership will continue to increase.

Table 9 Orenco Station, Daily Weekday Boardings, 1990-2000

Street Classification

Figure 12 shows the 2002 street network for the Orenco Station area with one map showing all streets and the other map classifying the roads into minor and major categories.

Orenco Station
All Roads	Classified Roads

Figure 12 Orenco Station Street Classification

The Orenco Station network is irregular in pattern with clustered areas of grid and modified grid patterns. The impedance roads are sparse in coverage and are minimal in quantity, but the single pair bisects the study area. The walkable roads are abundant in the half-mile study area and are less prevalent in the quarter-mile study area. Table 10 lists the quantities of road types and the change over time.

Table 10 Orenco Station Street Classification

In 1993, Orenco Station was mostly open space, as reflected by the low number of roads. By 2002, however, the quantity of roads had increased quite dramatically. More importantly, the increase in roads was dominated by minor, walkable roads at both the quarter-mile (+312 percent) and half-mile distances (+218 percent) at the same time that auto-centric roads actually decreased slightly over the same time. As a master-planned greenfield development, special thought was given to the road network as space to accommodate multiple uses, and is characterized by the dominance of minor, pedestrian-friendly road segments.

Catchment Areas

Figure 13 shows the 2002 pedestrian catchment areas (PCA) and the impeded pedestrian catchment areas (IPCA) for Orenco Station.

Orenco Station
PCA-Quarter- & Half-mile	IPCA-Quarter- & Half-mile

Figure 13 Orenco Station Pedestrian and Impeded Pedestrian Catchment Areas

The pedestrian catchment area at Orenco Station is generally limited in its coverage and is confined mostly to the northern section of the study site. The quarter-mile zone is comprised of only three road segments and fails to provide much neighborhood level function. The half-mile PCA is larger, but not widespread. The Orenco Station IPCA is very small and extremely limited in its coverage. The service area is completely located on the northern side of the study areas, incorporating only a small share of the total potential of the theoretical service areas. This example demonstrates the severe impacts of the types of paths available to a pedestrian on their accessibility. Table 11 lists the PCA and IPCA ratios for Orenco Station.

Table 11 Orenco Station Pedestrian and Impeded Pedestrian Catchment Areas

* Change over time data for Orenco Station is not a useful measure due to methodological limitations of the spatial calculation. This problem is discussed more fully in the appendix.

For Orenco Station, the key quantitative figures to concentrate on are the 2002 PCA and IPCA ratios. In 1993, the Orenco Station was mostly undeveloped land and due to some methodological limitations of the spatial analysis method, the 1993 ratios do not offer a good base for a temporal look at the walkable environment. The limitation occurs in greenfield or large in-fill types of situations where the extent and coverage of the road network was minimal at the initial data point and where the road network is quite a distance from the eventual transit stop location. The result, in the case of Orenco Station, is that in the 1993 data analysis (where no real station existed), the PCA and IPCA ratios were based on starting one's trip from the nearest available road, even though it was some distance from the transit stop location. In 2002, when more roads had been constructed, including ones that were adjacent to the transit station, the total distance one could travel from the transit stop was reduced. That is, in 2002, one's distance traveled began at the transit stop, whereas in 1993, one's travel would have started at the nearest road. The discrepancy in starting places leads to the erroneous ratios of the initial data points in the Orenco Station area. That said, 2002 data does provide a good reflection of the local environment and can be used as a base for future temporal analyses.

In terms of the 2002 data, Orenco Station scores relatively poorly for the pedestrian catchment area (PCA) at the quarter-mile (0.26), but much better at the half-mile (0.39), although still under the level considered the minimum for good walkability (0.60). When looking at the 2002 IPCA, the zone of walkability when accounting for the presence of auto-centric arterials, the ratio between this walkable zone and a Euclidean-based area shrinks considerably at both the quarter- and half-mile distances. At the quarter-mile, the IPCA ratio of 0.16 is 39 percent smaller than the PCA of 0.26, while at the half-mile the IPCA ratio of 0.14 is 65 percent smaller than the PCA ratio of 0.39, meaning that the presence of arterials has a major impact on the likely zone of walkability to or from the transit stop.

Intersection Analysis

Figure 14 visualizes the intersection intensities for Orenco Station.

Orenco Station
Intersection Intensities	Impedance-Based Intersection Intensities

Figure 14 Orenco Street Intersection Intensities

Orenco Station in an interesting case in that its density of intersections is very high at both scales, across both data sets, and across space, implying good walkability across the TOD. Yet, dead-ends, often recognized as an impediment to walkability, are also very high. As a greenfield development, some of the dead-ends represent areas where existing streets terminate at vacant lots, which conceivably will be developed and integrated in the near future. Table 12 lists the quantitative figures for the intersection intensity analysis.

Table 12 Orenco Station Intersection Intensities

There are three key pieces of information contained in Table 12:

1. The high density of intersections in 2002. In 2002, Orenco station had a higher intersection density at both the quarter- and half-mile distances than all other Portland areas, except the half-mile Lloyd Center area. Moreover, the impedance-based intersection densities far outscore any of the other TOD areas, reflecting a tight network of internal neighborhood streets deliberately built within this greenfield development.

2. The relatively low drop off between "regular" intersection densities and impedance-based intersection densities. In 2002, there was a relatively low reduction in intersection density when intersections based on the presence of major roads were removed. This small drop-off reflects the relatively low impact and ratio the major arterials have in relation to the presence of minor roads.

3. The positive change over time in intersection densities. In 1993 the Orenco Station area had the lowest intersection densities of any of the Portland case study areas, yet within a decade it had the highest densities, doubling the overall intersection density at both geographic scales. This change over time reflects the explicit pedestrian-oriented planning that was a focus of the greenfield development. Also, although the number of dead-ends increased dramatically (they started with a low n, so the increase may be a bit distorted), many of these dead-ends are where streets temporarily terminate at vacant lots. A future analysis conducted after Orenco Station is fully built out will most likely see a reduction in these dead-ends as the vacant lots get converted to more mixed use or residential housing consistent with the other development in the area.

Intersection Surface Map

The Orenco Station area is characterized by three pockets that contain high levels of internal street connectivity defined by a high density of "good" intersections. The dark pocket to the north of the transit stop is the primary area that has been developed over the last decade and Figure 15 shows how deliberately the area was developed with walkability principles in mind. There is another pocket of relatively high internal connectivity just south of the transit stop and a third pocket that begins about a half-mile south of the transit stop. As of 2002, much of the area immediately south of the transit stop remained undeveloped, although it is currently experiencing rapid development. It is likely that future calculations would show a considerable increase in the islands of walkability.

Figure 15 Orenco Station Surface Map

Portland: Beaverton

Figure 16 Beaverton TOD Arterials

Demographics

In 2000, Beaverton had the highest population density of the four Portland TOD areas (4,065 people per square mile), which also represented a 26 percent increase in density compared to 1990 (Table 13). Similar to Orenco Station, the population became more diverse with a 14 percent increase of the non-white population share, including a big increase of 17 percent in the Hispanic share of the population. The median age decreased by a bit more than two years over the decade, with a small increase in population share of the 0-17 and 45-64 age cohorts. The over-65 age cohort saw the biggest share decline (down by 2.6 percent), while the 18-44 age cohort saw a small decline. Household size increased slightly, and the average household income increased by 28 percent to almost $37,000.

Table 13 Socio-Demographic Characteristics of Beaverton, 1990-2000

Transit Ridership-Census

Beaverton also experienced a change in travel mode to work with the introd