Walk the line: station context, corridor type and bus rapid
transit walk access in Jinan, China
The MIT Faculty has made this article openly available. Please share
how this access benefits you. Your story matters.
Citation
Jiang, Yang, P. Christopher Zegras, and Shomik Mehndiratta.
“Walk the Line: Station Context, Corridor Type and Bus Rapid
Transit Walk Access in Jinan, China.” Journal of Transport
Geography 20, no. 1 (January 2012): 1–14.
As Published
http://dx.doi.org/10.1016/j.jtrangeo.2011.09.007
Publisher
Elsevier
Version
Author's final manuscript
Accessed
Fri May 27 00:13:12 EDT 2016
Citable Link
http://hdl.handle.net/1721.1/100709
Terms of Use
Creative Commons Attribution-NonCommercial-NoDerivs
License
Detailed Terms
http://creativecommons.org/licenses/by-nc-nd/4.0/
Title Page (WITH Author Details)
Click here to view linked References
Walk the Line:
Station Context, Corridor Type and Bus Rapid Transit Walk Access in
Jinan, China
Yang Jiang a,1, P. Christopher Zegras b, Shomik Mehndiratta c
a
China Sustainable Transportation Center. CITIC Building, Room 1903, No.19 Jianguomenwai Daijie, Beijing P. R. CHINA,
100004
b
Department of Urban Studies and Planning, Massachusetts Institute of Technology. 77 Massachusetts Avenue, Room 10-403,
Cambridge, MA 02139, United States
c
The World Bank, 1818 H Street, NW Washington, DC 20433, United States
1
Corresponding author. Tel: +86 10 85261955 ext. 103. Fax: +86 10 85262200.
E-mail addresses: yangjiang@chinastc.org (Yang Jiang), czegras@mit.edu (P. Christopher Zegras),
smehndiratta@worldbank.org (Shomik Mehndiratta).
*Blinded Manuscript (WITHOUT Author Details)
1
Introduction
Bus rapid transit (BRT) represents likely one of the most widespread urban public transportation
“revolutions” of recent decades. Although plans (largely unrealized) for BRT-type systems can be found
in United States urban contexts as far back as the late 1930s(Levinson, et al., 2003), BRT first became
widely recognized in the transportation community with the pioneering “southern” case of Curitiba
(Brazil) (Zegras & Birk, 1994). Essentially, what has generally become known as BRT aims to emulate
more up-front-capital-intensive rail-based systems on key performance characteristics – including
reliability, comfort, and speed – by utilizing measures like segregated and dedicated rights of way, paybefore-boarding at dedicated stations/stops, advanced traffic control and management measures for bus
priority, and enhanced system marketing and branding. BRT’s popularity has increased globally, due to
its promise for delivering a relatively low-cost, rapidly implemented, flexible, and high service quality
solution to developing cities’ transportation needs(Wright & Hook, 2007).
Public transportation will almost certainly play an important role in moving towards more sustainable
urban futures for our planet. In that sense, BRT offers various opportunities and challenges, including
several related to integrating BRT with the built urban environment. Can BRT deliver on possibilities for
transit-oriented development? How can BRT corridors and stations be integrated into the urban fabric to
induce ridership? Does the urban environment have an effect users’ willingness to walk for system
access?
In this paper we endeavor to shed light on answers to these questions. Specifically, using a recently
implemented BRT system in the city of Jinan, China, we examine built environment factors which
apparently influence station walk access distances. The remainder of this paper includes five additional
sections. The following section introduces the concept of transit station area catchment and examines the
possible role of the built environment. Section 3 describes the research context and approach, including a
description of the user survey. Section 4 presents the results of the survey analysis, including users’
perceptions of walking conditions, descriptive statistics, and multivariate regression attempting to identify
factors influencing station walk access distances. Section 5 discusses some planning implications of, and
limitations to, the analysis. Section 6 concludes.
2
Backdrop: The transit station catchment area and the role of the built environment
Most basically, a station “catchment area” represents the geographical area served by a particular transit
station (hereafter, we use “transit” as shorthand for “public transportation”) within a particular distance or
time, often some empirically determined “maximum” walk distance or area within which a majority of
users arrive by foot (Chalermpong & Wibowo, 2007). In fact, different catchment areas may be defined
for walk access/egress, feeder modes, automobile park-and-ride, and so forth. For urban transit systems
(as opposed to, say, suburban systems like commuter rail), the walking catchment area tends to be
particularly important, since walking is typically the primary access/egress mode for urban stations (e.g.,
Hsiao, et al., 1997). As a continuously growing base of research consistently reveals associations between
walking behavior and the built environment(Ewing & Cervero, 2001; Greenwald & Boarnet, 2001; Guo,
et al., 2007; Handy, et al., 2006), we would thus intuitively expect the built environment to exert some
influence on a transit station’s walk-based catchment area.
Theoretically, people make transit use decisions based on the expected benefit of the trip (i.e., whether the
system gets a user to a desired destination) and the relative (to other available modes) disutility of
realizing the trip, which includes price, in-vehicle time, wait times, transfer times, and access/egress times
– all of which may vary by the users’ socioeconomic and demographic characteristics. Each different time
component of a trip also includes two elements: actual and perceived (subjective) times, with the latter
influenced by comfort levels (e.g., vehicle crowding, wait time uncertainty, amenities). In this way, we
can formally understand the role that the station area built environment might play in determining the
effective catchment area for walk/access egress. The surrounding street and path networks impact actual
distances and times – by determining, for example, directness of routes and number of stops, crossings,
and other interferences – while these same networks’ conditions and other elements of the built
environment, such as density and diversity of different land uses, impact perceived times – by affecting,
for example, the overall walking experience.
Knowing the size of a station catchment area is important to transit system planning and operations since
the station area largely determines the number of final origins and destinations, thus potential demand, the
system will serve. However, even in a single city, actual station catchment areas may vary due in part to
physical characteristics: highly walkable station area surroundings will, all else equal, increase the
expected distance that people will walk to access/egress the system by reducing the real and perceived
time (disutility) of the walk. Therefore, knowing the relationship between a station’s effective catchment
area and its built environment could improve ridership forecasts and also inform how changes in the built
environment might increase the catchment area. This suggests a potentially mutually reinforcing
relationship between urban design and transit demand via transit-oriented development (TOD).
Unsurprisingly, no consensus exists among practitioners or researchers regarding a uniform standard for,
nor uniform approach to estimating, catchment area size. For light rail transit (LRT) system planning and
rail-based TOD, the walking distance guidelines range from 300-900m in Canada with variation across
cities, compared to 400-800m in the USA (Canepa, 2007; Ewing, 1999; O'Sullivan &Morrall, 1996).For
bus stops, 400m walking distance is usually considered(Ammons, 2001; Levinson, 1992). To empirically
test the relationship between the local built environment and transit ridership/ transit choice, (Ryan &
Frank, 2009) use 800m radii in calculating catchment-related attributes for San Diego’s Metropolitan
Transit (bus) system. In similar studies focusing on BRT systems, Estupinan& Rodriguez (2008) use
250m radii for Bogota’s BRT (Colombia), whereas Cervero, et al. (2009) use 800m radii for Los Angeles
County’s BRT (USA).
A number of empirical studies on transit access walking distances exists (Alshalalfah&Shalaby, 2007;
Ker & Ginn, 2003; Loutzenheiser, 1997; O'Sullivan & Morrall, 1996; Olszewski & Wibowo, 2005;
Rastogi & Rao, 2003). Results show that actual walk access distances vary across cities and countries,
and with respect to a series of factors (see Table 1). Without accounting for these factors, uniform
standards or calculated catchment-related indexes are simplistic and possibly misleading. Recently,
scholars have questioned the legitimacy of a priori catchment standards, instead calling for more
advanced analytical techniques (Biba, et al., 2010; Canepa, 2007; Cheng & Agrawal, 2010; Foda &
Osman, 2010; Landex & Hansen, 2006; Maghelal, 2011; Zhao, et al., 2003).
Table 1.
Factors Influencing Transit Walk Access Distance in the Empirical Literature
Note: √ refers to significant factors; × refers to not significant factors; V.O. refers to vehicle ownership; * refers to
studies focusing on the walk share within a certain catchment rather than the walk access distance
Much of the existing catchment area research focuses on rail systems and factors of influence such as
station function or levels of service, the relative location of the station, and trip maker characteristics
(Table 1). Although some studies have focused specifically on the role of built environment, none has
explicitly modeled walk distance. Examining the walk access share of the BART rail stations in San
Francisco (USA), Loutzenheiser (1997) found that walking trips are influenced primarily by individual
characteristics, with urban design and station area characteristics (retail-oriented environment) playing a
modest role. Of more direct regional relevance to us, Olszewksi&Wibowo (2005) study mass rail transit
station access in Singapore and find that, besides walking distance, the number of ascending steps,
number of road crossings, and number of traffic conflicts around stations reduce the likelihood of walk
access. Chalermpong&Wibowo (2007) examine walk access choice (within a pre-defined 2 km catchment
area) to rail transit stations in Bangkok and find varying effects across different stations in the system;
while they do not attempt to quantify the built environment characteristics related to these station-specific
effects, they qualitatively describe some station-specific conditions that might be influencing the walk
access choice. Finally, Maghelal(2011)models the walking percentage to light rail stations in Dallas (USA)
for quarter-mile and half-mile distances and finds that sidewalk density is positively associated with
percent walking to transit at both distances. Besides the absence of directly measured walk access
distances, these studies also do not account for confounding factors such as transit station function, station
location, and trip-specific characteristics. In addition, we have found no studies focusing on BRT walk
access and catchment area identification.
In the China context, although transit infrastructure development has boomed in recent years and transit
tends to have a significant mode share in most cities(Darido, et al., 2010), walk access distances to transit
– especially to the relatively new BRT systems –have rarely been explored. Cervero& Day (2008)and
Pan, et al. (2009) each use a 1-km threshold in models to estimate (rail) transit proximity effects on travel
behavior and accessibility of residents in specific neighborhoods, yet neither study offers evidence to
support the selection of this threshold.
We have reason to suspect that station catchment areas may be very context- and culture-specific. Data
comparing average walking distance and times across countries suggest that residents of some Chinese
cities may have a higher tolerance for walking(Rastogi & Rao, 2003), although comparing results across
contexts should be done with caution due to the need to control for how trips are defined in the original
data sources, trip purposes studied, socio-demographic and economic variables of influence, the transport
system modes and performance, broader urban context, climate, etc. Mateo-Babiano&Ieda (2007), for
example, present some evidence that average pedestrian speeds in countries of developing Asia are lower
than in the West (and Japan). Those authors emphasize the socio-cultural underpinnings of pedestrian
behavior and the important role of street space use. The willingness to walk to/from transit stations and
the average walk access/egress distance are likely influenced by such factors. Furthermore, particular
urban characteristics of Chinese cities may have relatively unique transit access effects. For example,
China has a somewhat special urban form due to a clearly hierarchical street network. Particularly for
BRT systems, different corridor types (arising in part from the street network hierarchy), may impose
varying influences on the individual stations along a corridor due to trunk road design, station location
setting, associated surrounding urban fabric (e.g., side streets), intersection design, and overall scale.
In this paper, we examine the relationships between corridor type and station context on the effective
walk catchment areas of BRT stations in a specific Chinese city, with the objective of answering the
following questions:


What BRT corridor types exist in China and are they perceived differently, in terms of
walkability, by users?
What are the walk access/egress distances to BRT stations and do they vary according to the
station area built environment and BRT corridor type?
We attempt to answer these questions by looking at Jinan’s BRT system which, in summer 2009,
operated on three corridors in the city. We expect the findings to improve understanding of how the
functions and forms of different types of urban roads affect people’s accessibility to public transport,
regardless of the quality of public transport services themselves. We also hope to inform the design of
BRT infrastructure (stations and corridors) to increase both walking to/from stations and overall system
patronage.
3
Research context and design
3.1 National context
China’s ongoing urbanization, economic growth, and motorization have transformed the nation’s urban
landscape over the past decade. Transportation infrastructures have undergone rapid and massive change,
including through new and expanded arterial roads, boulevards, ring roads, and access-controlled
expressways. In recent years, authorities in many Chinese cities have increasingly recognized the
importance of improving public transport conditions, including via investments in rail-based mass transit
and, increasingly, in bus rapid transit (BRT). Since at least 1999, Chinese cities have started providing (or
planning to provide) “advanced” bus rapid transit (BRT) services (Fig. 1). However, the lack of upfront
integration of road design, public transportation planning, land-use planning and early-stage public
consultation has created challenges to providing high quality public transport services on many new urban
corridors.
Fig.1.BRT development in China 1999-2008.
Sources: data extracted from ITDP (2009) and ChinaBus.Info (2008).
3.2 Jinan context
Jinan is the capital of Shandong Province. The city’s urban area expanded from 117 km2 in 1986 to 295
km2 in 2009(Jinan Statistics Bureau, 2009) and, according to the Jinan city master plan, the urban district
will expand to a built-area of 410 km2 by 2020. The city’s population of 3.5 million people is expected to
increase by an additional 1 million people by 2020(Jinan Urban Planning Bureau, 2005).
To cope with increasing travel demand and urban growth, Jinan has been aggressively developing a BRT
system since 2005.The Chinese central government named Jinan a “BRT Demonstration City”(SDUTC,
2008). As of summer 2009, the city had 3 BRT lines on 3 corridors –Jingshi Road, Beiyuan Road, and
Lishan Road – running a total of 34kmswith 34 stations (seeFig.2).
According to plans, by the end of 2015, Jinan will have over 120 kms of a BRT network(SDUTC, 2010).
This ambitious plan carries important policy and planning relevance for the city. For corridors still in the
early planning stages, analysis of the existing system could inform corridor and station planning and
design, enhancing demand forecasts and potentially increasing catchment area size and system utilization.
For predetermined corridor types (i.e., where the corridor type is already fixed), analysis of the existing
system and its stations could still help improve ridership forecasts and understanding of the catchment
area.
Fig.2.JinanBRT system at the end of 2009.
Source: Wang, et al. (2010, p.4).
The Jinan BRT case provides the possibility to examine the effect of three different corridor types,
representing different walking environments which are common in Chinese cities, while controlling for
city-specific variation. As summarized in Figure 3, in the Jinan case we observe:
1) The “arterial-edge” type corridor (Jingshi Rd), with mid-block curb-side BRT stations and
dedicated lanes on a ten-lane arterial. This represents a popular corridor type in China, with the
most famous one likely being Chang’an Avenue passing Tian’anmen Square in Beijing. Corridors
of this type tend to have super-wide arterials along with superblock development. Cities
sometimes create these high profile corridors to enhance city image, providing a window into
efforts to convey modernity(Tao, et al., 2010). Nonetheless, the public life along these roads
(including side streets) tends to be lacking, and people must walk additional distances from
intersections to the typically mid-block located stations.
2) The “integrated-boulevard” type corridor (Lishan Rd), with median BRT lanes and large treeshaded sidewalks, which usually have small setbacks and an active street-edge with retail. Side
streets tend to have similar human-scale built environment conditions as the corridor due to small
street blocks and ground-floor retail concentrations.
3) The “below-expressway” type corridor (Beiyuan Rd), with median-lane BRT right-of-way under
the viaduct and stations at major intersections. Many cities have elevated ring roads of the type
running above this BRT corridor. This corridor type and its side-streets tend to be car-oriented,
lacking a human-scale and being chaotic with little landscaping, since these often run along
existing poor settlements at the urban fringe which are partly being replaced by newly developed
superblock projects.
The unified BRT service planning and operation on three corridors in a single city provide a possibility to
control for other factors influencing system performance. Currently, the three BRT lines operationally
overlap each other on the three corridors, with similar speeds and station spacing. The system allows free
transfers between BRT lines at transfer stations. Therefore, even with some corridor-specific
characteristics within system/service operations, BRT riders tend to experience different corridors and the
associated system/service effect in one single BRT trip. Thus, Jinan’s BRT provides a unique opportunity
to somewhat control for overall quality of transit services when exploring how the functions and forms of
different types of urban roads affect people’s accessibility to the system.
Fig. 3. Major typologies of BRT corridor walkaccess environments in China and examples from Jinan
3.3 Research design
We collected BRT walk access information by interviewing people at BRT stations over a 4-day period in
late August 2009. Table 2shows the variation in the characteristics comprising the station-based sampling
frame. Although the initial sample was randomly chosen, approximately one third of respondents was in a
hurry and refused the survey, leading to a possible response bias towards people with a lower value of
time. The survey was conducted from 7 to 10 AM and 4 to7 PM each day from Wednesday to Saturday,
in an attempt to cover peak, non-peak, weekday and weekend periods. Nineteen BRT stations along the
three BRT corridors constituted our sampling frame (Fig. 4, Table 2).
Fig.4. Surveyed BRT stations and their contexts.
In the survey, those BRT users who walked to or from a station were asked to point out on a map their
approximate origins or destinations as well as their walking routes. Users also reported socioeconomic,
demographic, trip-specific (e.g., purpose), and other travel (e.g., availability of other travel modes)
information. Finally, respondents were asked to rate their walk access with respect to a series of related
statements. In total 2,155BRT users were surveyed, from which we obtained1406 observations with valid
walk access records, among which 1,233 remained valid after excluding responses with incomplete
information.
We recorded the reported walk paths and geo-coded them in a geographic information system (GIS). We
also geo-coded each BRT station’s side street network and calculated relevant distances (e.g., path
distances, straight-line distances, total length of side streets within a 600m-radius station area buffer.
Table 3 presents the data derived on the physical characteristics related to the corridors and the stations,
including the method and units of measurement.
Table 2
Station Sampling Frame
a. Income and age was collected as category info. Mid-point value for each category is used for average income and
age estimation.
Table 3
Data Captured on Corridor Type, Station Context and Walkability
4
Research analysis and findings
Among valid observations, approximately half were female and almost 92% were between 20 and 60
years old (see Fig. 5, left); 82% reported a household monthly income between 1000RMB and 5000RMB
(see Fig. 5, right). This number is consistent with the BRT user profile revealed from another
conventional on-board survey, in which 80.8% of respondents fall into the 1000-5000RMB income
category (SDUTC, 2010), and with Jinan’s officially reported average monthly income for urban
households: 1835 RMB (Jinan Statistics Bureau, 2009). About 85% of surveyed BRT users were
employed, 11% were students (see Fig. 6, left), and 43% were reporting on a work trip (see Fig. 6, right)
(the low work trip share reflects peak/off-peak reporting periods). The relatively small share of school
trips probably results from the survey period – late August, which coincides with China’s school holidays.
The degree of BRT “dependency” for the particular trip (i.e., no alternative mode available) attempts to
control for the increased likelihood of walking a greater distance when BRT is the only option available:
87% of respondents mentioned they would take conventional bus to make the same trip absent BRT; only
1% of respondents reported having no other choice than BRT. This supports the idea that, in Chinese
cities like Jinan, the BRT system mainly attracts former conventional bus users.
Fig.5. Distribution of BRT survey respondents by age and income.
Fig.6. Distribution of BRT survey respondents by occupation and trip purpose.
4.1 Corridor access walkability
We focus on four dimensions of walkability across the BRT corridors: protection, comfort, enjoyment and
directness. We evaluate the first three aspects based on survey respondents’ rating of their walk access
experiences (see Figure 7). We derive a proxy for the last measure, directness (ds), for station s as:
(1),
where:
is the walking distance from the reported origin/destination of user n to station s; is the
associated straight-line distance from the origin/destination of user n to station s; and N is the total
number of survey respondents at the station. Note that, technically, the measures are carried out for
stations on corridors; attributing them to the corridor presumes transitivity.
Protection refers to security against traffic safety risk and against crime. None of the three corridors rated
satisfactorily on this point. No more than a third of surveyed BRT users agreed that crossing and walking
on sidewalks was safe and easy, although, interestingly, the below-expressway is perceived as safer than
the other two corridor types. The low protection rating is not surprising, given our own observations of
the corridor conditions. The Jingshi corridor is often quite wide, up to 10 lanes at some crossings, with
people having to cross busy traffic with relatively short green-light cycles or using footbridges (Tao et al.
(2010) report similar problems with general-purpose arterials in the city of Fushun). The Lishan corridor
is narrower in the middle section, but also expands at the crossings. Finally, the Beiyuan corridor has
poorly designed or managed light signals at some crossings and presents such serious drainage problems
that people often cannot cross the street after a rain; this corridor rates significantly lower than the other
two on crossing satisfaction.
Comfort refers to the ease of walking (fewer obstacles), including sidewalk quality and street cleanliness.
The arterial-edge Jingshi corridor, representing a “city image” project as discussed above, ranks
significantly more favorably on comfort indicators than the other two corridors. About 67% of
respondents on the Jingshi corridor think the pavement is good, significantly higher than Lishan(50%)
and Beiyuan(38%). Similarly, more respondents perceive Jingshi as cleaner and with fewer sidewalk
blockages.
Enjoyment refers to aesthetic and utilitarian aspects related to the presence of activities and relief from the
elements (e.g., shade from sun). On this dimension, the integrated-boulevard corridor, with 70% of
respondents agreeing the corridor’s trees on sidewalks make walking pleasant, ranks better than the other
two, where less than 50% responded similarly. The scale of the arterial-edge corridor, on the other hand,
manifests wide streets and deep building set-backs (see Fig. 8). We dare say that the Jingshi corridor’s
“city image” function provides a view more pleasing to drivers than pedestrians. For example, big trees
along sections of the corridor, set back from the sidewalk, serve more as a backdrop to the vehicle lanes,
impeding on-the-ground store development and interaction between pedestrians and buildings. The
below-expressway corridor provides poor pedestrian scale, overwhelmed by the overhead mega-structure.
Fig.7.BRT user perceptions of walkability experience on Jinan’s 3 BRT corridors.
Note: [ ] refers to “walkability” aspect (see text). Percentage refers to share of surveyed BRT users that agree with
walkability-related statements (i.e., scale rating as 4 or 5). All proportions within each aspect significantly different
from each other (p<0.05), except ^ (p<0.10), and * (p>0.10).
Fig.8.Jingshi corridor’s set-back with trees.
One potentially confounding factor in the differences described above is that users’ ratings of the local
environment might vary by income levels. A chi-square test revealed no significant relationship between
income category and reported perceptions, except for the comfort-related aspect, “pavement is good.”1
Finally, directness refers to a station’s relative “detour” factor, as measured by equation (1), above.
Averaging station-specific detour factors for each corridor type shows the arterial-edge corridor has an
average detour factor of 1.59, indicating less directness, whereas the integrated boulevard and belowexpressway corridors have lower values of 1.36 and 1.33, respectively. This reflects the arterial-edge
corridor’s access disadvantage due to the stations’ distance from major intersections. Pedestrians have to
walk about 17-20% longer to access stations than they would if the stations were closer to crossings.
As summarized in Table 4, the three corridors’ distinct physical arrangements, landscape, and street
facilities lead to different levels of walkability perceived by BRT users. Overall, the integrated-boulevard
corridor seems more walkable than the other two. The next section explores the relationship between
corridor type and BRT user walk access distance.
Table 4
Qualitative Assessment of Comparative Walkability on 3 BRT Corridors in Jinan
4.2
BRT walk access patterns
4.2.1 Descriptive analysis
Figure 9 shows the recorded BRT walk access routes. Purple lines represent walking paths. Blue circles
are 600-meter buffers at each station, represented by green dots. The map reveals that the majority of
origins/destinations around the three terminal stations extend well beyond the 600m buffer area,
suggesting a longer average walk access distance at the terminals. Second, the actual catchment areas of
stations on the arterial-edge corridor (the southern E-W route) look smaller than those of stations on the
other two corridors.
1
The critical χ² value is 11.07 (5 d.f.) at a p<0.05; the calculated χ² values are: Walking on sidewalks is safe, 7.897;
Crossing is safe and easy, 3.027; Pavement is good, 11.656; Streets are clean, 6.429; Few blockages are on
sidewalks, 3.313; Trees on sidewalks make walking pleasant, 4.823; Facilities along streets meet my demand, 9.770.
Thus we fail to reject the null hypothesis of no association between perceptions and income in all cases but
“Pavement is good.”
Scrutiny confirms our first impressions. Fig. 10(top) shows the cumulative distributions of walking
distance to BRT terminal stations, transfer stations, and typical stations. The distance walked clearly
relates to station function, with terminal stations having longer walk distances; 80% of respondents
walked farther than 600 m to a terminal station. The average walking distance to a terminal station, 1392
meters, is more than double that to a non-terminal station (Table 5).This might be partly attributable to the
station function – a terminal, being the end of a transit line, might deliver higher relative levels of
accessibility to the urban system and thus increase walk attractiveness to that station type – and/or to
potentially higher concentrations of trip destinations (e.g., shopping) around these station types.
Regarding corridor type, Fig.10 (bottom) shows the cumulative distributions of walking distance to BRT
stations by corridor. The integrated-boulevard appears to have longer average walk distances (see, also,
Table 5). Integrated-boulevard stations have an average walk access distance of 649 meters, compared to
475 meters and 580 meters on the arterial-edge and below-expressway corridors, respectively. The
integrated-boulevard may be somehow incentivizing people to walk farther to use the system (despite
roughly comparable levels of BRT service quality across the three corridors).
The walk route data also enable a comparison of the walk access route patterns. The integrated-boulevard
corridor presents a distinct hierarchical walk path pattern (i.e., access flows merge onto a few routes
connecting to the stations), whereas the other two corridors display relatively random patterns (Fig. 11).
This visual survey suggests that the integrated corridor may have a few walkable side streets, which
provide sufficient access means.
Station context might also influence walk access distance to the BRT. Beyond the apparent role of station
function (i.e., terminal or not), the station’s density gradient may also influence catchment area, although
it has been rarely examined. Statistically, controlling for the density gradient is necessary due to its
influence on the distribution of the full station-catchment population, from which we randomly drew the
survey. A station with a downward sloping density gradient, all else equal, will likely have a shorter
average walking distance observed from the survey, not because BRT users around the station are
necessarily less willing to walk, but because they do not have to walk as far on average when going to
stations. Those individuals are more likely to appear in the survey than people accessing farther away
locations. Unfortunately, Jinan has no publicly available geo-coded data on relevant land uses (e.g.,
business activities) and demographics (e.g., population density). We somewhat crudely identified stylized
density gradient patterns, using a three-dimensional map (EDUSHI, 2009): hill pattern (intensive
development adjacent to station), flat pattern (constant density emanating out from the station) and valley
pattern (low density or vacant land adjacent to station). Figure 12 shows examples of the “hill” and
“valley” density patterns.
Fig.9. Jinan BRT station’s access/egress walking routes
Fig.10.Cumulative frequency diagrams of walking distances at terminal, transfer and typical BRT stations
(top) and three corridor types (bottom).
Table 5
Walking Distance to BRT Stations by Corridor Type and Station Function
a. Based on walk distance to non-terminal stations, since Jinan’s integrated-boulevard corridor has no terminal
stations.
Fig. 11.Walking path patterns on the three corridors.
Fig. 12.Density gradient patterns of Jinan BRT stations.
Fig. 13.Average walking distance by household income (top) and alternative mode availability (bottom).
We should be cautious about drawing conclusions by directly comparing walk distances across station
functions and corridor types. Other potentially confounding factors exist. For example, income may affect
people’s willingness to walk, all else equal. Richer people tend to walk less to access BRT stations in
Jinan (Fig. 13, top), consistent with the theory of wealthier people having a higher value of time. Also,
access to alternative modes for the same trip may also have an influence – people with no choice other
than taking BRT walk longer than those with an alternative (Fig. 13, bottom). Those with an alternative
available for the trip reveal less variation in walk distances. Other potentially confounding sociodemographic and trip-specific factors include gender, age, occupation, trip purpose, and trip timing.
4.2.2 Regression analysis
To isolate the influence of BRT corridor type and station walking conditions on the measured actual
station walking distances, we specify an ordinary least squares (OLS) regression of the basic form:
(2),
where:
=
=
=
=
=
=
BRT station walk access distance of trip maker i,
a vector of socio-economic status variables of trip maker i,
a vector of trip-specific variables of trip maker i,
a vector of station context variables associated with trip maker i,
the BRT corridor type (dummy) on which trip maker i is interviewed, and
a random error term.
Table 6 presents the regression results. The “control model” includes only the trip maker (
and trip2
related (
variables. Note the very low goodness-of-fit as evidenced by the adjusted R (0.012); these
variables alone account for very little variation in individual BRT station walk access. The “full model”
includes the corridor type dummies and several station context variables and shows a large improvement
in goodness-of-fit, as measured by the adjusted R2(0.223), relative to the “control model.”The coefficient
on the integrated-boulevard corridor variable is positive and significant. This suggests that, all else equal,
BRT users on the integrated-boulevard corridor walk an average 158 meters longer to the BRT stations
than those on the arterial-edge and below-expressway corridors. This may reflect the superiority of the
integrated-boulevard corridor in terms of its walkability perceived by BRT users, as discussed earlier.
Most station context variables are also significant. Compared to those at typical stations, walking access
distances at terminal stations are 373 meters longer. Transfer stations have the opposite effect, with walk
access distances 126 meters shorter than typical stations. One explanation for this result may be that the
BRT transfer stations are also usually well connected with conventional bus routes/stops, reducing the
number of longer walk access trips. In terms of the stylized density gradients, the “hill pattern” has a
lower walk distance relative to the “flat pattern,” as expected, while the “valley pattern” has a higher walk
distance, also as expected. The relative location of the station in the city also has a significant effect:
stations farther from the city center have longer walk access distance (for each kilometer from the city
center, a station has approximately 75 meters of additional walk access distance), all else equal. This is a
reasonable result: presuming the city center represents the point of highest accessibility in the city, the
farther away from the city an individual is (and subsequently the lower her relative accessibility at that
location), the more likely she will be willing to walk farther to access the BRT system.
Trip maker and trip-specific characteristics remain mostly insignificant in explaining walking access
distance (consistent with Chalermpong and Wibowo (2007)’s findings for walk choice access to rail
transit in Bangkok). In the “full model,” low income people walk 165 meters longer on average compared
to median income people, yet high income people walk as long as the median income people. Occupation
and gender have no significant relationship with walk access distance; nor do car ownership, frequency of
BRT use, weekend trip-making, or group trip-making.
Table 6
OLS Regression Models Predicting BRT Walk Access Distance
*p<.10, ** p<.05
The age effect remains interesting. On the one hand, people aged 40 to 60 walk less than people from the
other age groups surveyed. On the other hand, the other age groups show no statistically significant
differences in their walk distances, meaning older adults (over 60) seem to walk as far people aged under20 to 40. This result should be interpreted with some caution. For example, older adults may have a lower
value of time, seat privileges on the system, and enjoy benefits of a free ticket policy which give them an
incentive to walk more; but, older adults’ presence in the sample may be biased. The older adults
surveyed were ambulatory – if they had trouble walking, they would have a lower likelihood of being in
our sample. Anecdotally, from the survey implementation, we observed that some “super-healthy” older
adults even regarded walking to the BRT as exercise.
5
Implications and limitations
5.1 Research implications
This research has several implications. First, cities pursuing transit-oriented development should
recognize a unique opportunity around terminal BRT stations. Our analysis finds that the average radial
distance of the walk catchment area for such stations may be as large as 1350 m, meaning that the walk
catchment area, or pedestrian zone, of a BRT terminal station maybe up to five times larger than the
coverage area with the conventionally assumed 600m radial distance. However, the location of BRT
stations, if not close to road intersections, can reduce catchment area size, due to walk “detour” effects.
Second, for travel demand analysis, accounting for corridor type and station context may improve BRT
ridership forecasts in the China context. Conventional travel demand models assume that transit riders,
particularly choice users, are mostly sensitive to a bus stop’s service quality and fare levels. Given the
likelihood of a varying catchment area sizes, demand analysis should explicitly incorporate station type
and context, including walk-route connectivity, street crossing, and the like.
Third, for urban planning more generally, flexible catchment area definitions are needed, reflecting
corridor type and station context. Based on the model results, we develop initial guidelines for radial
distance walk catchment areas around BRT stations. Applying the relevant model coefficients yields the
expected average walk distance:
(3)
The final radial distance should be further discounted by a detour factor which is station-location
dependent. For example, the arterial-edge corridor in Jinan suggests a “discount” of up to 30% for this
corridor type. Table 7 presents rough estimates of radial distance catchment areas by corridor type and
station function.
Finally, from a public finance perspective, our analysis suggests investment should prioritize a few
pedestrian access routes to cost-effectively enlarge BRT station catchment areas. Evidently, the station
context along the Jinan integrated-boulevard corridor includes a few walking-friendly routes intensively
used by the majority of surveyed BRT riders accessing those stations. In addition, when investing in an
access route, particular attention should be paid to improving the “enjoyment” factor. On Jinan’s arterial
corridor, the handsome and apparently comfortable conditions do not incentivize walking farther, perhaps
due to the lack of attractions and/or pedestrian-oriented functionality.
Table 7
Estimated Variations in Catchment Area Distance Guidelines
The lessons and implications above, while specifically applicable to Jinan, can guide similar analyses in
other cities and help them develop more public transport-friendly urban road infrastructures, including by
helping to prioritize bus-priority measures based on corridor type, identify walk-friendly station-area
characteristics, and highlight possibilities for retrofitting facilities.
5.2 Research limitations
This research has a number of limitations. First, different aspects of corridor type may affect walk access
distances in different ways. Unfortunately we could not examine them due to the limited number of BRT
corridors (only 3) operating at the time in Jinan and the lack of walk-route-specific built environment
characteristics. The latter could be overcome via detailed field work characterizing specific walk route
attributes. The findings regarding variations by corridor type should thus be viewed tentatively because
corridor characteristics are “bundled” with local access conditions which may confound effects. Second,
the other end of each BRT trip sampled was not taken into account in the statistical analysis; presumably
riders consider both access and egress together when making trip decisions. Third, the model does not
include more accurate station context factors (e.g., actual density, land use mix), due to lack of data.
Fourth, we do not know whether walk access distances vary by season because the survey was conducted
in the summer; people may walk less in the cold winter. Finally, the survey implementation technique
certainly resulted in some biases, perhaps over- (or under-) estimating actual walk catchment areas.
6 Conclusions
We examined BRT station walk access patterns in rapidly urbanizing China, where BRT implementation
has been on the rise. Urban form features and station context and right-of-way configurations may
influence users’ willingness to walk to BRT and, thus, the walk access catchment area. We tested this
hypothesis with data from a user survey, conducted at 19 BRT stations along the city of Jinan’s three
existing (as of summer 2009) BRT corridors. We applied ordinary least squares regression to estimate the
relationship between walk access distances and aggregate station- and corridor-area characteristics,
controlling for individual- and trip-specific attributes.
The results suggest that people walk farther to BRT stations when the walking environment has certain
features (median transit-way station location, shaded corridors, busy and interesting). Among Jinan’s
three BRT corridors, stations on the integrated-boulevard corridor (Li-shan Road) have a 160 meter
longer average walk access distance than those on the arterial-edge and below-expressway corridors.BRT
users also perceive the integrated-boulevard corridor as the most walkable environment, although it shares
a safety problem with the other two corridor types.
Compared to the station and corridor contexts, trip and trip maker characteristics play a relatively minor
role in defining BRT walk access distance. Only low income and BRT-captive people walk more than
average. Older adults do not necessarily walk less to access BRT stations, although this result may be
affected by sampling bias. Women are willing to walk as long as men. Occupation, car ownership, trip
purpose and time do not affect walk access distance among the BRT users surveyed.
Our findings have several implications. For urban planners, BRT stations offer an important opportunity
for transit-oriented development; our analysis suggests a need, and method, to account for and
subsequently influence walk access catchment area. For travel demand modelers, considering corridor
type and station context may improve BRT ridership demand analysis. Transit planners should use
flexible catchment area definitions, reflecting the corridor type and location station context. Finally, for
transit investment decision-makers, providing a few critical pedestrian access routes to stations may be a
cost-effective way to enlarge catchment areas.
Acknowledgements
The funding for this project was provided by the Transport Unit of the World Bank, Beijing Office. We
acknowledge the contribution of Asst. Prof. Zhang Ruhua and Mr. Wu Xiangguo from Shandong
University, who helped in survey organization and data collection, and Mr. Zhang Tao from Beijing
Normal University, who helped in GIS digitalization of the survey data. In addition, we thank Sam
Zimmerman, Ke Fang and several anonymous reviewers for their valuable comments and suggestions.
References
Alshalalfah, B. W., & Shalaby, A. S. (2007). Case study: relationship of walk access distance to transit
with service, travel, and personal characteristics. Journal of Urban Planning and Development,
133, 114.
Ammons, D. N. (2001). Municipal Benchmarks: Assessing Local Performance and Establishing
Community Standards, Second Edition. Thousand Oaks, CA: Sage.
Biba, S., Curtin, K. M., & Manca, G. (2010). A new method of determining the population with walking
access to transit. International Journal of Geographical Information Science, 24(3), 347-364.
Canepa, B. (2007). Bursting the Bubble: Determining the Transit-Oriented Development's Walkable
Limits (Vol. 1992, pp. 28-34): Transportation Research Board.
Cervero, R., & Day, J. (2008). Suburbanization and transit-oriented development in China. Transport
Policy, 15(5), 315-323.
Cervero, R., Murakami, J., & Miller, M. (2009). Direct Ridership Model of Bus Rapid Transit in Los
Angeles County: UC Berkeley: UC Berkeley Center for Future Urban Transport: A Volvo Center
of Excellence.
Chalermpong, S., & Wibowo, S. S. (2007). Transit station access trips and factors affecting propensity to
walk to transit stations in Bangkok, Thailand. Journal of the Eastern Asia Society for
Transportation Studies, 7, 1806-1819.
Cheng, C. L., & Agrawal, A. W. (2010). TTSAT: A New Approach to Mapping Transit Accessibility.
Journal of Public Transportation, 13(1), 55-72.
ChinaBus.Info. (2008). Development of BRT in China (Publication. Retrieved October 10, 2009:
http://www.chinabus.info/topic2008/brt2008/
Darido, G., Torres, M., & Mehndiratta, S. (2010). Urban Transport and CO2 Emissions: Some Evidence
from Chinese Cities. Paper presented at the Transportation Research Board Annual Meeting
2010.
EDUSHI. (2009). Jinan 3D Map. Accessed September 16, 2009, from http://jinan.edushi.com/
Estupinan, N., & Rodriguez, D. A. (2008). The Relationship between Urban Form and Station Boardings
for Bogota's BRT. Transportation Research Part A, 42, 296-306.
Ewing, R. (1999). Pedestrian-and Transit-Friendly Design: A Primer for Smart Growth: Washington, DC:
Smart Growth Network.
Ewing, R., & Cervero, R. (2001). Travel and the built environment: a synthesis. Transportation Research
Record: Journal of the Transportation Research Board, 1780, 87-114.
Foda, M. A., & Osman, A. O. (2010). Using GIS for Measuring Transit Stop Accessibility Considering
Actual Pedestrian Road Network. Journal of Public Transportation, 13(4), 23-40.
Greenwald, M. J., & Boarnet, M. G. (2001). Built environment as determinant of walking behavior:
analyzing nonwork pedestrian travel in Portland, Oregon. Transportation Research Record, 1480,
33-42.
Guo, J. Y., Bhat, C. R., & Copperman, R. B. (2007). Effect of the Built Environment on Motorized and
Nonmotorized Trip Making: Substitutive, Complementary, or Synergistic? Transportation
Research Record: Journal of the Transportation Research Board, 2010, 1-11.
Handy, S. L., Cao, X., & Mokhtarian, P. L. (2006). Self-selection in the relationship between the built
environment and walking: evidence from Northern California. Journal of American Planning,
72(1), 55-74.
Hsiao, S., Lu, J., Sterling, J., & Weatherford, M. (1997). Use of geographic information system for
analysis of transit pedestrian access. Transportation Research Record, 1604(1), 50-59.
ITDP. (2009). China Bus Rapid Transit (Publication. Retrieved October 10, 2009, from Institute for
Transportation & Development Policy (ITDP): http://www.chinabrt.org
Jinan Statistics Bureau. (2009). Jinan Statistics Yearbook. Beijing: China Statistics Press.
Jinan Urban Planning Bureau. (2005). Jinan City Master Plan 2005-2020. Jinan, Shandong Province,
PRC: Jinan Urban Planning Bureau. (in Chinese).
Ker, I., & Ginn, S. (2003, May 18-23). Myths and Realities in Walkable Catchments: The Case of
Walking and Transit. Paper presented at the 21st ARRB and 11th REAAA Conference, Cairns,
Australia
Landex, A., & Hansen, S. (2006). Examining the Potential Travellers in Catchment Areas for Public
Transport. Paper presented at the 26th Annual ESRI International User Conference.
Levinson, H. (1992). Urban Mass Transit Systems. In J. D. Edwards (Ed.), Transportation Planning
Handbook. New Jersey: Prentice Hall.
Levinson, H., Zimmerman, S., Clinger, J., Rutherford, S., Smith, R. L., Cracknell, J., et al. (2003). TCRP
Report 90: Bus Rapid Transit, Volume 1: Case Studies in Bus Rapid Transit. Washington, DC:
Transportation Research Board.
Loutzenheiser, D. R. (1997). Pedestrian access to transit: model of walk trips and their design and urban
form determinants around Bay area rapid transit stations. Journal of the Transportation Research
Board, 1604(1997), 40-49.
Maghelal, P. K. (2011). Walking to transit: influence of built environment at varying distances. Insititue
of Transportation Engineers, 81(2), 38-43.
Mateo-Babiano, I., & Ieda, H. (2007). Street space sustainability in Asia: the role of the asian pedestrian
and street culture. Journal of the Eastern Asia Society for Transportation Studies, 7, 1915-1930.
O'Sullivan, S., & Morrall, J. (1996). Walking distances to and from light-rail transit stations.
Transportation Research Record, 1538(1), 19-26.
Olszewski, P., & Wibowo, S. S. (2005). Using Equivalent Walking Distance to Assess Pedestrian
Accessibility to Transit Stations in Singapore. Transportation Research Record, 1927(2005), 3845.
Pan, H., Shen, Q., & Zhang, M. (2009). Influence of Urban Form on Travel Behaviour in Four
Neighbourhoods of Shanghai. Urban Studies, 46(2), 275-294.
Rastogi, R., & Rao, K. V. K. (2003). Travel characteristics of commuters accessing transit: Case study.
Journal of Urban Planning and Development, 129, 684.
Ryan, S., & Frank, L. F. (2009). Pedestrian Environments and Transit Ridership. Journal of Public
Transportation, 12(1), 39-57.
SDUTC. (2008). Technical Evaluation of Jinan Bus Rapid Transit System (BRT) Performance in Trial
Operations. Jinan: Transportation Planning & Design Research Center, Shandong University
(SDUTC) and Jinan Bus Company Research Institute. (in Chinese).
SDUTC. (2010). Jinan Bus Rapid Transit System (BRT) Evaluation (Draft). Jinan: Transportation
Planning & Design Research Center, Shandong Univerity (SDUTC). (in Chinese).
Tao, W., Mehndiratta, S., & Deakin, E. (2010). Compulsory Convenience? How Large Arterials and Land
Use Affect Midblock Crossing in Fushun, China. Journal of Transport and Land Use,
forthcoming.
Wang, T., Zhang, R., Zhu, X., Wu, X., & Zhang, R. (2010). Bus Rapid Transit in Jinan, China- Applying
Flexibility to Transit System. Paper presented at the TRB 2010 Annual Meeting.
Wright, L., & Hook, W. (2007). Bus Rapid Transit Planning Guide. New York: Institute for
Transportation and Development Policy.
Zegras, C., & Birk, M. (1994). Moving toward integrated transport planning: energy, environment, and
mobility in four Asian cities. Transportation Research Record: 1444, 84-92.
Zhao, F., Chow, L.-F., Li, M.-T., Gan, A., & Ubaka, I. (2003). Forecasting transit walk accessibility: a
regression model alternative to the buffer method. Paper presented at the Transportation Research
Board Annual Meeting 2003.
Figure1
20
18
16
14
# of BRT Lines
12
10
8
6
4
2
0
1999 2000 2001 2002 2003 2004 2005 2006 2007 2008
Fig.1.BRT development in China 1999-2008.
Sources: data extracted from ITDP (2009) and ChinaBus.Info (2008).
Figure2
Fig.1.JinanBRT system at the end of 2009.
Source: Wang, et al. (2010, p.4).
Figure3
BRT-Corrid
dor Typology
1. Arterial-Ed
dge
Jinan Example (Trunk Roads)
Jinan Exaample (Side Streeets)
Corridor
BRT
Station
Sidewalk
Corridor
Sidestreet
Green
Buffer
Jing
gshiRd.
2. Integrated--Boulevard Corrid
dor
Corridor
Trees on
sidewalk
LishanRd.
3. Below-Exp
pressway Corrido
or
Corridor
Beiy
yuan Rd.
Fig. 3
Major ty
ypologies of BRT
T corridor walkacccess environmentts in China and ex
xamples from Jinnan.
Figure4
Fig.4. Surveyed BRT stations and their contexts.
Figure5
Age (Years)
Household Income (RMB/Mo)
2% 2% 3%
5% 3%
5%
<20
8%
20‐30
26%
30‐40
22%
<600
11%
40‐50
57%
50‐60
1000‐2000
2000‐5000
56%
>60
Fig.5. Distribution of BRT survey respondents by age and income.
600‐1000
5000‐10000
>10000
Figure6
Occupation
2%
1%
4% 2%
6%
11%
Trip Purpose
Teacher
5%
Work
Student
12%
Worker
13%
9%
47%
Government official
Company employee
Service/self‐
employed
10%
43%
2%
Shopping
10%
8%
Recreation
12%
3%
Fig.6. Distribution of BRT survey respondents by occupation and trip purpose.
School
Figure7
[Protection] Walking on sidewalks is safe.
26%*
28%*
32%
[Protection] Crossing is safe and easy.
29%*
26%*^
21%^
69%
[Comfort] Pavement is good.
50%
38%
47%
[Comfort] Streets are clean.
33%*
35%*
45%
[Comfort] Few blockages are on
sidewalks.
24%*
27%*
48%
[Enjoyment] Trees on sidewalks make
walking pleasant.
70%
39%
33%
[Enjoyment] Facilities along streets meet
my demand.
58%
49%
0%
Arterial-edge
20%
Integrated-boulevard
40%
60%
80%
100%
Below-expressway
Fig.7.BRT user perceptions of walkability experience on Jinan’s 3 BRT corridors.
Note: [ ] refers to “walkability” aspect (see text). Percentage refers to share of surveyed BRT users that agree with
walkability-related statements (i.e., scale rating as 4 or 5). All proportions within each aspect significantly different
from each other (p<0.05), except ^ (p<0.10), and * (p>0.10).
Figure8
Fig.88.Jingshi corrridor’s set-back with trees..
Figure9
(surveyed only)
Fig.9. Jinan BRT station’s access/egress walking routes.
100.00%
90.00%
80.00%
70.00%
60.00%
50.00%
40.00%
30.00%
20.00%
10.00%
0.00%
Terminal Station
Transfer Station
Typical Station
0
150
300
450
600
750
900
1050
1200
1350
1500
1650
1800
1950
2100
2250
2400
2550
2700
2850
3000
3150
3300
3450
3600
3750
3900
Percentage of BRT riders
Figure10
100.00%
90.00%
80.00%
70.00%
60.00%
50.00%
40.00%
30.00%
20.00%
10.00%
0.00%
Arterial‐Edge
Integrated‐Boulevard
Below‐Expressway
0
150
300
450
600
750
900
1050
1200
1350
1500
1650
1800
1950
2100
2250
2400
2550
2700
2850
3000
3150
3300
3450
3600
3750
3900
Percentage of BRT riders
Access/Egress Walking Distance (m)
Access/Egress Walking Distance (m)
Fig.10.Cumulative frequency diagrams of walking distances at terminal, transfer and typical BRT stations
(top) and three corridor types (bottom).
Figure11
Arterial- Edge
(Jingshi St.)
Integrated- Boulevard
(Lishan Rd.)
Fig. 11.Walking path patterns on the three corridors.
Below- Expressway
(Beiyuan St.)
Figure12
Station 3
Station 8
BRT
BRT
Valley
Pattern (concave)
Valley
Pattern
(concave)
Pattern (convex)
HillHill
Pattern
(convex)
Fig. 12.Density gradient patterns of Jinan BRT stations.
Source: Adapted from (EDUSHI, 2009).
Figure13
900
Avg Walking Distance (m)
800
771
781
646
700
647
639
600
509
500
400
300
200
100
0
<600
600‐1000 1000‐2000 2000‐5000 5000‐10000 >10000
Household Income (RMB/Mon)
Avg Walking Distance (m)
1025
1000
800
600
591
667
616
679
733
615
554
400
200
0
Mode Alternative for the Same Trip
Fig. 13.Average walking distance by household income (top) and alternative mode availability (bottom).
Table1
Table 1.
Factors Influencing Transit Walk Access Distance in the Empirical Literature
Station
Station
Trip maker
Trip
Built
Transit
function
location
features
features environment
system
(level of
service)
Purpose
MultiTransit type√ CBD√
Gender×
×
×
mode
Route
Age
×
Length
√
√
systems
frequency
Transit Pass
Transfer√
(Toronto,
Dwell type×
Canada)
√
V.O.
×
Occupation
√
Rail
Suburb
(Perth,Aust
rilia)
DownRail (Bay
Gender√
Density×
√
√
area, USA)
town√
Age
Retail
√
Ethnicity
Income×
√
Light rail
Terminal
Gender×
CBD√
√
(Calgary,
Transit type
Canada)
Road
Rail
Transit type√
Gender√
√
×
(Singapore)
crossings
Age
Traffic
conflicts√
Ascending
steps√
√
Rail
Income
(Mumbai,
V.O.√
India)
×
Specific-station
Rail
Gender
√
×
(Bangkok,
dummy
Age
Thailand)
V.O.×
×
Income
Occupation×
Sidewalk
Light rail
Income√
√
√
(Dallas,
density
Ethnic
USA)
Housing
density×
Land use mix×
Author(s)
Alshalalfah&
Shalaby(2007
)
Ker &Ginn,
(2003)
Loutzenheiser
(1997)*
O'Sullivan&
Morrall(1996
)
Olszewksi&
Wibowo
(2005)*
Rastogi&Rao
(2003)
Chalermpong
&Wibowo
(2007)*
Maghelal
(2011)*
Note: √ refers to significant factors; × refers to not significant factors; V.O. refers to vehicle ownership; * refers to
studies focusing on the walk share within a certain catchment rather than the walk access distance
Table2
Table 2
Station Sampling Frame Station
Station Distance to Road Length in # Feeder
Respondent Respondent
ID
Function City Center
500m
Bus
Average Income Average
(km)
Catchment (km) Routes (RMB/month)a
Age a
Typical
1.9
15.3
2
3733
36
1
Typical
1.5
16.7
6
4768
36
2
Typical
2.4
15.4
1
4050
31
3
Typical
2.8
18.2
3
3430
29
4
Terminal
4.5
13.3
10
3066
33
5
Transfer
1.6
15.9
6
3740
31
6
Transfer
2.0
15.4
12
2542
30
7
Transfer
2.4
13.2
1
3041
37
8
Transfer
3.1
15.5
10
3447
34
9
Transfer
4.2
16.1
6
4570
28
10
Transfer
3.6
18.6
6
3699
26
11
Transfer
3.2
15.8
0
2127
29
12
Transfer
2.7
12.5
3
3845
29
13
Typical
2.9
20.2
4
4011
29
14
Typical
3.2
22.2
0
3180
33
15
Typical
3.6
20.3
20
2479
27
16
Typical
4.3
18.6
3
2214
27
17
Terminal
6.2
13.6
5
2682
29
18
Terminal
5.9
13.6
4
3234
28
19
a. Income and age was collected as category info. Mid-point value for each category is used for average income and
age estimation.
Table3
Table 3
Data Captured on Corridor Type, Station Context and Walkability
Data Item Description
Method
Unit of Measure
Corridor type
Visual
Categorical
Walkability
Visual (pictures and
Ordinal (Likert-scale):1-stongly
videos); Rating on
disagree, 5-strongly agree
statements
Station density gradient
Read from local 3D map
Categorical
Station function
Read from local BRT map Categorical
Length of side streets
Measured in GIS
Kilometers
Distance to city center
Measured in GIS
Kilometers
Number of bus lines to station
Count from local bus map Numbers
Table4
Table 4
Qualitative Assessment of Comparative Walkability on 3 BRT Corridors in Jinan
Corridor Type (Name)
Walkability Perceptions from BRT Users
Protection
Comfort
Enjoyment
Arterial-Edge(Jingshi)
poor
good
poor
Directness
poor
Integrated-Boulevard(Lishan)
poor
average
good
good
Below-Expressway(Beiyuan)
poor
poor
average
good
Table5
Table 5
Walking Distance to BRT Stations by Corridor Type and Station Function
Corridor Typea
Statistics
Station Function
Arterial- IntegratedBelowEdge
Boulevard Expressway
Typical
Transfer
Terminal
Mean
475
649
580
549
586
1392
Median
412
520
458
435
458
1311
Maximum
1635
2023
2738
2738
2067
5114
Minimum
102
47
36
102
37
97
95% Confidence Interval
for Mean (Lower Bound)
444
599
546
516
555
1234
95% Confidence Interval
for Mean (Upper Bound)
505
699
613
578
619
1496
No. of Valid Observations
332
271
631
627
607
172
a. Based on walk distance to non-terminal stations, since Jinan’s integrated-boulevard corridor has no terminal
stations. Table6
Table 6
OLS Regression Models Predicting BRT Walk Access Distance
Control Model
Variable
Coefficient T-test
BRT Trip Maker & Trip Characteristics
Income <2000RMB
120.371*
1.69
Income 2000-10000RMB
ref
Income >10000RMB
-133.728
-1.08
Occupation: Professional
24.397
0.58
Occupation: Blue Collar
105.998
1.40
Occupation: Service/ Self-employed
15.386
0.28
Gender: Female
-29.701
-0.99
Age <20
2.552
0.04
Age 20-40
ref
Age 40-60
-36.600
-0.79
Age >60
200.407**
2.20
BRT-Dominant User
19.723
0.63
Car Ownership
-26.006
-0.62
Trip Purpose: Commuting/ Schooling
ref
Trip Purpose: Shopping
-68.515
-1.37
Trip Purpose: Recreation/ Social
53.361
1.21
Trip Purpose: Personal Business/ Other
-59.678
-1.55
No Alternative Mode Available
470.689**
2.55
Trip Time: Weekend
-7.556
-0.22
In Group
13.516
0.39
BRT Corridor Type
Integrated-Boulevard (Lishan Rd)
Below-Expressway (Beiyuan Rd)
Arterial-Edge (JingshiRd)
BRT Station Context
Terminal Station
Transfer Station
Typical Station
Density Gradient: Hill
Density Gradient: Flat
Density Gradient: Valley
Number of Feeder Bus Routes
Distance to City Center (km)
Feeder Road Length in 600m Catchment Area
(Constant)
No. Observations
(df)
F
Adjusted R2
*p<.10, ** p<.05
640.032**
12.27
1,233
(18,1214)
1.882*
0.012
Full Model
Coefficient
T-test
165.651**
ref
-54.418
-9.133
-43.635
-48.788
2.330
-72.527
ref
-73.832*
26.446
42.035
6.414
ref
-46.560
22.799
-21.551
415.598**
-26.062
28.053
2.60
-0.49
-0.24
-0.64
-1.00
0.08
-1.14
-1.75
0.32
1.47
0.17
-1.04
0.58
-0.62
2.53
-0.85
0.90
158.810**
-20.432
ref
2.60
-0.32
372.886**
-126.453**
ref
-156.771**
ref
153.714**
0.583
75.926**
-11.127
3.52
-2.34
597.833**
1,233
(27,1205)
14.576**
0.223
-4.15
3.52
0.18
2.40
-1.16
3.06
Table7
Table 7
Estimated Variations in Catchment Area Distance Guidelines
Radial Distance (meters)
Terminal
Station
Non-terminal
Station
Arterial-Edge
600-1000
300-600
Integrated-Boulevard
1000-1500
600-1000
Below-Expressway
800-1200
400-800
Corridor Type