Bus Transit and Land Use: Illuminating the Interaction

Attracting people to public transit in urban areas has proven to be a difficult task indeed. Recent research on the transportation–land use connection has suggested that transit use can be increased through transit-friendly land use planning. While significant evidence exists that a relationship between land use and transit is apparent, the exact nature of the relationship remains ambiguous. Despite the murky nature of the relationship, many practitioners and researchers have asserted claims regarding land use policy, namely TOD, and its effect on travel. This article examines the effect of land use, socioeconomics, and bus transit service on transit demand in the Twin Cities. The findings suggest that vertical mixed-use is important close to transit access and retail plays an important role up to a quarter mile from transit service. Population density is more important at a block-group level than block level, suggesting that density adjacent to the line may not play as critical a role as density in the larger surrounding area. Introduction Why do some intraurban areas attract more transit riders than others? What types of neighborhoods may induce greater transit demand? The pressure to find an answer has increased as a result of population growth, congestion, and discontent with existing transportation options. Journal of Public Transportation, Vol. 6, No. 4, 2003 2 2 Despite the growing disenchantment with urban transportation, people have shown little interest in changing their ways; transit’s share of work trips is still only about 4.5 percent nationally (Bureau of the Census 2000). An increase in travel times and the stability of the auto modal split suggests that people remain willing to sacrifice transportation convenience for perceived housing and neighborhood amenities. The relative low cost of auto ownership, existing cultural preferences, and transit-inefficient land use patterns only reinforce the current auto-oriented transportation situation. The relationship between transportation and land use has received increased attention in recent years, however, the exact nature of the relationship relative to other causes remains somewhat ambiguous. Despite the ambiguous nature, proponents of transit use have focused much attention on regulating development in a manner that is more supportive of transit use, which has been coined transitoriented development (TOD). TOD proponents have blamed much of today’s transportation woes on inefficient development patterns, and propose TOD as one of many contributors to a solution. In response to this problem and policy response, this article seeks to illuminate the complex relationship between transit demand and its influences, including density, land use, socioeconomic characteristics, and transit service. This analysis seeks to answer the question: What intraurban qualities make one area generate more or less demand for transit services? This article will first summarize the current state of transportation land use and transit literature; secondly, describe the methodology employed: next, present findings of this research; and finally expand on the findings to suggest directions of future research and public policy. State of the Literature Transportation Land Use and Travel Behavior The interaction between land use and travel behavior has been studied heavily in recent years; one need look no further than the most recent studies eloquently compiled by Ewing and Cervero (2001), and Seskin and Cervero (1996). The surveyed research typically measured one of six different outcome variables: trip frequency, trip length, mode choice, cumulative person miles traveled (PMTs), vehicle miles traveled (VMTs), or vehicle hours traveled (VHTs) (Table 1). The latter three variables are different measures representing the same phenomenon—agBus Rapid Transit and Land Use 2 3 gregate travel. Research to date has found the primary determinant of the various outputs to vary, although these concepts are interconnected. Table 1. Output Variables from Travel and Land Use Studies Generally, mode choice is affected primarily by density and land use (Table 1). This is particularly important given that local-level public policy has little direct effect on neighborhood socioeconomics or regional accessibility in the short term, while local land use regulations and neighborhood-level policy directly affect the land use and density. In addition to density, transit ridership appears to be a function of size of the central business district (CBD) and the distance from downtown (Puskarev and Zupan 1977), as well as parking supply and price, transit service quality, pedestrian accessibility, and land use mix (Figure 1). The size of the CBD and distance from the CBD of a given stop is important because, due to the radial nature of most public transit systems, a larger CBD equates to a more accessible transit system. In addition, a larger CBD often means fewer parking spaces per person or job, which decreases the incentive to drive. While those characteristics cited by Pushkarev and Zupan are important, parking supply, accessibility, and land use mix are also important. Recent research suggests a positive relationship between parking price and transit use (Hess 2001). This is particularly troubling to transit supporters due to the finding that free parking is enjoyed at the end of 99 percent of all trips (Cervero 1998). In addition to the economic influences of parking, parking lots, and ramps are poor land uses for inducing transit ridership. Although accessibility of transit systems has been shrinking relative to automobile accessibility for decades due to increased growth at the suburban fringe, it remains an important aspect of transit service. In addition to Output Variable Primary Determinants Trip frequency Socioeconomic characteristics Trip length Regional accessibility Mode choice (1) Density/(2) land use Cumulative PMTs/VMTs/VHTs Regional accessibility Source: Ewing and Cervero, 2001. Journal of Public Transportation, Vol. 6, No. 4, 2003 2 4 Fi gu re 1 . In fl ue nc es o f Tr an si t D em an d Bus Rapid Transit and Land Use 2 5 regional accessibility by means of transit, accessibility to transit is a critical factor in willingness to use transit. Due to safety concerns, perceived comfort and the effect of climate, the design of transit stops and station area amenities play an integral role in transit patronage. The importance of climate and comfort is particularly important in areas such as Minneapolis-St. Paul that often endure harsh winter weather conditions. The effect of land use on transit is murky, although it is believed that a positive feedback loop between transit and land use exists. Transit availability increases aggregate accessibility to a given location and the attributes of the specific location determine whether people visit the location. However, the precise effects of different land uses on transit use are unclear, in part, due to the degree of interconnectedness with density and socioeconomic influences. What is clear is that the greater the intensity of land use, the greater demand for transit. The general applicability of this research is unknown because most research has focused around rail transit, despite the prevalence of bus transit. Rail transit has become increasingly en vogue with policy-makers, the media, and researchers alike due to nostalgia (e.g., “new urbanism” or “rail revival”), potential environmental efficiency, the ease in the provision of high-frequency service, and the attractiveness of guaranteed service provision to potential developers and investors. TOD has received increased attention in recent years. TOD’s bark is perhaps bigger than its bite; it has been rarely practiced due to reluctance in the private land market and institutional barriers (Boarnet and Crane 1998). Also, there is little empirical evidence to support that individual TOD projects in a sea of single-family homes can actually sustain transit and lower auto reliance (Cervero 1998). A similar affinity toward TOD near rail transit has persisted, leaving the relationship between bus transit and TOD unclear at best. Minneapolis-St. Paul: Transit, Land Use, and History The Minneapolis-St. Paul metropolitan region has enjoyed significant economic vitality in recent years, and as a result significant population growth. The growth has manifest as primarily moderateto low-density development on the urban fringe. The Minneapolis-St. Paul region has long befriended the automobile and auto-oriented development. The Metropolitan Council, the regional transit operator and land use planning agency, has responded with several public policy and marketing programs aimed at limiting geographical dispersion of residential Journal of Public Transportation, Vol. 6, No. 4, 2003 2 6 growth and concentrating development along transit corridors. The Livable Communities Act was created to achieve these goals by dedicating a pool of public money that is awarded on a competitive basis for “Smart Growth” developments. The goal of the policy is to encourage developments that could be more easily served by transit in hopes of avoiding the high cost of constructing or expanding the highway system. The regional transit system, as of 2002 exclusively bus–transit, carries about 250,000 riders per weekday. The bus system will soon be joined by an 11-mile, $750 million light-rail transit line connecting the Minneapolis CBD, the Minneapolis-St. Paul International airport, and the Mall of America, the largest enclosed shopping mall in the United States. Metro Transit, the regional transit operator, currently operates the annual 73 million bus trips offered, primarily in the two central cities and inner-ring suburbs. Methodology This analysis uses the Sector 5 restructuring data obtained from the Metropolitan Council, the regional planning agency for the Minneapolis-St. Paul metropolitan area. Sector 5 is the transit planning subregion that consists of downtown Minneapolis and a radial slice running due south and southwest (Figure 2). Sector 5 contains four of the primary trip generators in the entire metro region: the Minneapolis CBD, Mall of America, International Airport, and part of the University of Minnesota Twin Cities campus. These data count only the downtown boardings onto buses that serve Sector 5. The attractiveness of Sector 5 in this analysis primarily lies in its relative importance and potential for increased service. It currently serves 55 percent of all transit riders, offers 38 percent of all routes, and almost 20 percent of the jobs and residents in the entire region. This is particularly important given the area only comprises about 10 percent of the geographic area. The Sector 5 data used for this analysis consist of weekday transit boardings at bus stops in Sector 5 south of the Minneapolis CBD and west of the Mississippi River, known as areas B and C of Sector 5 (approximately 95% of all stops in the sector). Bus stops of different routes at the same location have unique ID numbers, which allowed control of route orientation and service. Only boardings were used due to the correlation between boardings and alightings; in other words, people start their return trip the same place they ended the beginning trip. This assumption was affirmed via visual confirmation of boarding and alighting maps and tables. Bus Rapid Transit and Land Use 2 7 Figure 2. Sector 5 Reference Map Journal of Public Transportation, Vol. 6, No. 4, 2003 2 8 The data were compiled using GIS to select and join the relevant census and land use information with the exact location of the bus stop. The data were entered into a demand model and analyzed using a linear regression model. The land use data were simplified by combining open space, roads, and other categories and omitted from the model to allow a comparison of other land uses to relatively “dead” transit uses. The cross-town routes are controlled for, while the radial routes feature approximately the same levels of parking availability and price. Generally all parking for all destinations on routes going away from the CBD is free, and all radial routes going toward the CBD terminate there. The use of transit stops as data collection points, as opposed to individual data via travel behavior surveys, is useful for several reasons. First, transit agencies plan routes based primarily on area statistics. Similarly, land use planning can more easily create types of environments that are more conducive to transit ridership, than it can cause people to use transit. Although ideally both would be used, the small area level analysis is often overlooked, and potentially more useful to local planning agencies. The transit demand model was created to illuminate the intraurban differences, and so many causes of demand were eliminated. For example, macro-level predictors on transit certainly affect transit use. Recent evidence shows that much of the 12 percent decline in transit ridership in the first half of the 1990s can be attributed to a sluggish economy and low gas prices, while the increased gas prices and burgeoning economy resulted in a 21 percent increase in ridership in the later half of the 1990s (Pucher 2001). Density, land use, and transit service provide stronger explanatory power given the complex decision-making process associated with mode choice within a metropolitan region. Because this analysis only looked at one market at one point in time the macro-level predictors were eliminated from the analysis. Similarly, parking prices and size of the CBD were left out because these aspects are relatively constant in the area of analysis. The land use data were classified into basic categories: single-family, multifamily, retail–commercial, office, industrial–-utility, mixed-use, and other. The other category includes open space, roads, and unused/vacant lands or spaces. Interaction variables were entered to tease out the influence that various mixes have on transit demand. Land uses were categorized into groups based whether they are primary Bus Rapid Transit and Land Use 2 9 job-based (office, industrial–utility), shopping-based (retail–commercial), and housing-based (single-family, multifamily). Findings Geography A majority of the weekday transit demand in Sector 5 is currently located inside the City of Minneapolis, and more than 12 percent of total boardings in Sector 5 occurred in downtown Minneapolis (Table 2). Demand is clustered along Lake Street in South Minneapolis and peaks at the confluence of other transit routes (Figure 3). The ridership clustered along the Lake Street corridor, featuring crosstown service, is the area of maximum transit accessibility in Sector 5. This area is at maximum accessibility because route 21 runs the distance of Lake Street, connects to nearly every radial route in sector 5, and offers very high-frequency service. Table 2. Weekday Boardings by Location in Sector 5 Transit Service The type of transit service plays a significant role in demand of a given stop (Table 3). Compared to the Urban Local service, the most prevalent service in the core area, Urban Local-Limited Stop was negatively associated with demand (Table 3). Surprisingly the level of weekday service was not a significant determinant of transit demand. To better understand the relationship between transit service and ridership, a longitudinal analysis is warranted. Location Boardings % of Total Minneapolis City 46964 76.1% St Paul City 6551 10.6% Minneapolis Suburbs 7571 13.3% Total Boardings in Analysis 61697 100.0% Major Trip Generators Minneapolis CBD 8553 13.9% Mall of America 838 1.3%