Click for next page ( 11


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 10
Superior walkability with small blocks and pedestrian traffic management priority. Extended hours of highly-reliable transit service at 5- to 15-minute intervals. Land use mix to meet daily needs paired with good transit connectivity to other activities. Density sufficient to support cost-effective transit, retail services, and infrastructure. Managed parking with reduced supply relative to standard development. RESPONSE BY TOD DIMENSION AND STRATEGY The focus of this chapter section is on the transit usage and other traveler response outcomes found to be associated with the presence or implementation of transit oriented developments (TODs). The effects along each of three definitional dimensions of TOD, highlighted in the "Types of Transit Oriented Development" subsection, are examined: regional context, land use mix, and primary transit mode. There is necessarily substantial overlap among these discussions as it is impossible to fully isolate the contribution of each dimension. Three of the travel demand measures often employed to describe outcomes involve different aspects of traveler choice of travel mode. These measures are mode share, mode of access share, and sub-mode share. Mode share (unless explicitly otherwise defined) is a fundamental measure referring to choice of primary travel mode between a trip's origin and its final destination. For example, a trip starting with driving alone to a commuter rail (CRR) station, followed by a train ride terminating 1/2-mile from the final destination, and concluding with walking to get there, would be classed as a commuter rail trip for purposes of mode share calculation. A mode share proportion is expressed as a percentage of all travel by all modes (including walking and bicycling) in the travel category of interest or, mainly in older studies, all travel by motorized modes only. Also of frequent interest, especially to local area traffic, parking, and environmental concerns, is mode of access share. This share describes the proportions among means of getting to and from the primary mode. The access and egress modes in the example just given would be the drive-alone (to the station) mode of access and the walk (from the station) mode of egress. Less often encoun- tered is sub-mode share, the proportion of transit trips using a particular form of transit, such as local bus or heavy rail transit (HRT). A true sub-mode share is expressed as a percentage of all transit travel. Note that the "rail" and "bus" mode shares within mode share tables in this chapter are cal- culated in the manner of primary mode shares, not sub-mode shares, because they are expressed as percentages of travel by all means and presented in context with modes of all types. These tab- ulated rail and bus mode shares may be summed to obtain transit mode shares, a calculation not supported by true sub-mode shares. Response to TOD by Regional Context Many articles discuss TOD as a suburban strategy. Possibly this is because it is harder to apply the "TOD" label to city center development that fits into an existing pattern. Also, perhaps the subur- ban interest reflects growing developer awareness of the market potential for suburban TOD and the contrast that TOD represents in terms of a departure from the status quo of suburban sprawl. While city center TOD may attract more transit riders in total, suburban TOD may represent greater potential change. 17-10

OCR for page 10
Attributes associated with the regional context of a TOD strongly influence the travel behavior response. City center TOD is more likely to be accessible to a larger potential transit market than is suburban TOD. City center TODs may have transit services radiating in multiple directions. Suburban TODs tend to be located along radial transit services and thus only have high quality service available in one or two directions. In addition, it is more likely that activities can be accom- plished by walking or transit in the urban context than in the suburban context, especially in early- stage TODs. Table 17-1 highlights contrasts between the two regional contexts. Table 17-1 Perspectives on TOD as Differentiated by Regional Context City Center Context Suburban Context Transit Urban sites are often directly accessible Markets served by high-quality transit Markets to and from multiple transit markets. service may be limited. For example, For example, Gallery Place Metro Ballston Metro station in Arlington, station in Washington, DC, is fed by Virginia, is fed by a single east-west rail three rail lines originating from five line originating from two suburban areas. different suburban areas and passing Other transit riders from around the through different downtown areas, all metropolitan area must transfer to arrive offering one-seat rides to the station. at the station and/or use bus service. Drive Highway accessibility remains Mode of access to suburban transit station Markets important to the urban real-estate developments tends to remain dominated market. Automobile-oriented by the automobile and therefore commuting is prominent even in the automobile accessibility is of substantial most transit accessible locations. importance. Parking It may be more acceptable to constrain It may be difficult to manage parking; the Management and manage parking in downtown suburban real estate market may dictate areas, especially by using pricing. parking space ratios that are higher-than- Constrained parking leads to higher optimal for transit. Examples abound transit attractiveness. People may own where developers build more parking fewer cars in central areas due in part to than is required. Also, higher rates of good transit service availability and automobile ownership among residents easy walking access to utility retail. are present. Phasing Existing nearby land uses may support Neighborhood services supportive of non- Effects a TOD project in reducing single- automobile, non-work travel may not pre- occupant vehicle usage for midday exist. Thus, until such uses are part of the trips. Alternatively, nearby legacy TOD, the early phases of a new TOD may development may retain automobile exhibit higher automobile mode share orientation and dampen the behavior than the later phases of a more mature impacts of adjacent TOD. TOD. City Center TODs City center TOD adjusts the focus of transit-accessible urban development to increase tran- sit ridership and to encourage pedestrian activity. Some aspects of the city center, such as grid street patterns and ground-level retail uses, are attributes usually shared with TOD. 17-11

OCR for page 10
However, city center TOD often addresses the office-domination represented in the existing land use mix by introducing housing and other non-daytime uses. It is also usually constructed with less parking than traditional development to better encourage transit use. City center TOD tends to be less discrete than suburban TOD in that it may comprise several projects over multiple blocks. While there are many examples of city center TOD, virtually none provide a case where comprehensive and specific development, transit service, and traveler response information has been documented. In general terms, it is reported that these projects have increased transit ridership, encouraged pedestrian activity, and have required less parking than more traditional projects (Costello, 2003). Two of the better-documented examples follow. Washington, DC. A transit-oriented downtown sports arena was constructed at the Gallery Place Metrorail HRT station in Washington, DC, opening in December 1997. The 20,000 seat venue was built to allow the professional hockey and basketball teams to transfer from their suburban location to downtown. Whereas at the suburban location virtually everyone drove to the facility, the new arena has a transit mode share of over 50 percent for events. Transit ridership was encouraged by constructing an enhanced station entrance and by building virtually no patron parking on-site. Comparing average 1997 weekday boardings at the Gallery Place station to 1998 boardings shows a 56 percent increase (6,525 to 10,179). During the same period, systemwide weekday ridership increased by just under 5 percent, including the effect of one new station. The area around the arena, part of the old downtown, has since become a revitalized retail and entertainment district. In addition, several residential and office projects have been constructed nearby. Correspondingly, from 1998 to 2002, average weekday boardings at Gallery Place grew 36 percent above the 1998 level (to 13,833) versus just under 20 percent for the system as a whole, including the effects of adding 8 more new Metrorail stations (Costello, 2003; WMATA, 2002). In the subsequent 4 year period, from 2002 to 2006, the Gallery Place weekday ridership growth was almost 50 percent (to 20,673) compared to 13 percent for the system (to 714,953 with 3 more stations) (WMATA, 2006b). Denver, Colorado. In Denver's city center, ridership growth has occurred in parallel with TOD and transit system expansions, including the introduction of light rail transit (LRT) to downtown. Much of the development has involved converting former office and industrial buildings to mixed use or residential use, with more than 50 formerly derelict structures put back into service as of 2003. In addition, transit-supportive policies such as parking maximums and encouragement of ground-level retail have been employed. The population in the downtown has grown from about 1,000 to 2,500 with an accompanying increase in median income. Average weekday transit ridership on the 16th Street Transit Mall free bus shuttle was reported in 1997 to be 45,000 daily riders.2 In 2004, following both the introduction of riders from a second LRT line and an extension of the transit mall shuttle service to reach residents of the Central Platte Valley redevelopment, mall bus ridership was reported as 60,000 daily riders. Transit mode share for work trips to the downtown core is reported to have increased from 2025 percent up to 35 percent during this critical period of downtown Denver's revitalization. More than 10 percent of the region's workforce is employed in downtown and almost half of these workers commute either via bus or rail transit or by ridesharing. The separate contributions to the ridership and travel behavior 2 A circa 1997 traveler response review of the Denver transit mall bus shuttle is provided in Chapter 10, "Bus Routing and Coverage," under "Response by Type of Service and Strategy"--"Circulator/Distributor Routes"--"Transit Terminal and Parking Distributors." 17-12

OCR for page 10
changes attributable to LRT expansion, transit mall shuttle extension, and TOD and other development have not been isolated (Cervero et al., 2004; Costello, 2003; Ohland, 2001; Project for Public Spaces, 1997). Suburban TODs Some suburban TODs are built on undeveloped land at new or existing transit stations. Other such TODs are built on land previously used for surface parking lots. TOD has become viable on these suburban park-and-ride lot sites in part because metropolitan areas have expanded outward beyond the ends of the rail transit lines. However, there can be conflict between the use of the station area as a destination for park-and-ride transit users and its use as a place for people to live or work in the replacement development. Complicating matters are the requirements of suburban TOD for better highway access and greater accommodation of parking for tenants than needed for city cen- ter TOD, along with the mandates of many transit agencies for replacement of any park-and-ride parking removed for development purposes. Providing large amounts of parking has implications for both development economics and traveler behavior. Indeed, balancing the provision of parking is among the greatest challenges in planning suburban TOD (Bernstein, 2004). (For further background see the "Underlying Traveler Response Factors" discussions titled "Parking Supply" and "Parking Pricing and Transit Support.") Transit mode share for suburban TOD is higher than for traditional suburban development, but the automobile still plays a predominant role in providing mobility for TOD tenants. The "Underlying Traveler Response Factors" section discusses the attributes that distinguish the experiences of different suburban TODs in this regard, including automobile ownership rates, transit service characteristics, and parking policies. Suburban TOD has the challenge of generally being located along the tentacles of a radial transit system, automatically limiting the extent of the travel market having direct transit connections. Suburban TOD also has the challenge of siting within a regional context of higher automobile availability rates than the city center. Pleasant Hill Station, California. The Bay Area Rapid Transit (BART) Pleasant Hill HRT station was opened in 1973 as a next-to-the-end-of-the-line station with 3,245 spaces of surface parking. From 1986 forward, about 2,400 housing units, two hotels, offices with over 4,000 employees, and other improvements were built in the 140-acre area surrounding the station. However, the surface parking lots were retained, illustrating the challenge of reconciling the dual role of outlying transit stations as a collection point for park-and-ride commuters and a focus of rider origin or destination in their own right. In 1995, BART advanced an effort to redevelop much of the surface parking into TOD, including pedestrian-friendly connections and mixed-use development with office, residential, and retail tenants. To restore the nearly 1,500 commuter spaces on which the develop- ment was constructed, a six-story parking garage for transit users was incorporated. Parking for the new development is housed within the TOD's buildings, but at reduced ratios compared to those 17-13

OCR for page 10
used for traditional development in the county (Arrington et al., 2002; Tumlin and Millard-Ball, 2003). Table 17-2 highlights the parking space ratio reductions allowed.3 It has been reported that the Pleasant Hill Station TOD residential development generates 52 per- cent fewer peak period auto trips than the ITE Trip Generation Manual's observations for typical free-standing residential development. The station-area office development generates 25 percent fewer vehicle trips than typical stand-alone office buildings. For both development types, the trips that are made by automobile are said to be shorter because of the mix of uses in the vicinity. Together, these effects lead to fewer vehicle-miles of travel as compared with more typical subur- ban development (Belzer and Autler, 2002). Table 17-2 Standard Contra Costa Country Versus Pleasant Hill Station Parking Ratios Development General TOD Requirement Type Ratio Ratio Parking Supply Units of Measure Reduction Office 5.0 3.3 Spaces per 1,000 square feet of interior space 34% Residential 1.75 1. 35 Spaces per housing unit 23% Retail 5.0 4.0 Spaces per 1,000 square feet of interior space 20% Source: Arrington et al. (2002). The Pleasant Hill station was among the locations reported on in the collaborative "Travel Characteristics of Transit Oriented Development in California" study conducted in 2003 with publication in 2004. (This research will be referred to hereinafter as the "2003 California TOD travel characteristics study.") As part of the study, self-administered surveys were distributed to residents in four projects in the Pleasant Hill station area (identified later in Table 17-19). Overall, a survey response rate of 12.8 percent was achieved with a returned sample of 125 surveys. The 2000 U.S. Census data from nearby Walnut Creek was used for comparison (Lund, Cervero, and Willson, 2004a). Station-area residents reported for work trips about triple the transit mode share of their suburban Walnut Creek counterparts and about 1.4 times the transit mode share of residents over- all in the region's Central City, i.e., San Francisco.4 For walk trips, however, the low station area share is more typical of suburbia than San Francisco. Details are provided in Table 17-3. 3 The allowed parking space ratio reductions reflect possibly two considerations. It was undoubtedly the intention that the reduced TOD parking ratios for the Pleasant Hill station should reflect the lower automo- bile mode share and trip generation of TOD relative to standard suburban development. It should also be taken into account that the overall county requirements which were reduced from may have been on the high side even for average non-TOD development. See "Underlying Traveler Response Factors"--"Parking Supply"--"Development Parking," including Table 17-31 and footnote 15, for more. 4 The Central City (Census definition) of San Francisco is herein referred to, depending on context, as either the city of San Francisco or San Francisco County. They are geographically identical areas. 17-14

OCR for page 10
Table 17-3 Commute Mode Share of Surveyed Residents Versus Census in Nearby Cities Mode Pleasant Hill Station Area City of Walnut Creek City of San Francisco Drove Alone 48.9% 73.8% 43.5% Carpool 4.0 8.2 11.3 Rail Transit 44.3 13.5 9.8 Bus Transit 0.6 1.0 22.4 Walked 2.3 2.1 9.8 Other 0.0 1.3 3.2 Notes:Denominator used in Census mode share calculation excludes workers working at home. The city of San Francisco is included to provide a range for comparison (Walnut Creek, suburban; San Francisco, urban Central City). Source:Based on Lund, Cervero, and Willson (2004a) and 2000 U.S. Census SF3 data. For non-work trips, transit--and particularly rail transit--was found not to play as prominent a role for Pleasant Hill station-area residents as it does for commuting. As can be derived from Table 17-4, the not insubstantial 15.1 percent transit mode share for non-work trips is 1/3 the 44.9 percent work commute transit share. In terms of sub-mode share, a reported 38.4 percent of transit trips made for non-work purposes utilized bus rather than rail service. In contrast, only 1.3 percent of work pur- pose transit trips were made by bus. These contrasts likely reflect a more dispersed distribution of non-work destinations for station-area residents as compared to work destinations and the fact that, in contrast to broader bus coverage, the rail line only goes in two directions from the station (Lund, Cervero, and Willson, 2004a). In this case as in others, the rail service is not as relevant to non-work destinations as it is to work destinations, though the presence of the rail station is assuredly a major factor in the concentration of the bus service being used for non-work travel. Table 17-4 Work and Non-Work Mode Shares of Pleasant Hill Station TOD Residents Mode Work Mode Share Non-Work Mode Share Drove Alone 48.9% 70.9% Carpool 4.0 10.5 Rail Transit 44.3 9.3 Bus 0.6 5.8 Other 2.3 3.5 Trip Sample Size 176 86 Source: Lund, Cervero, and Willson (2004a). Over 96 percent of Pleasant Hill station-area residents using the BART station accessed it by walking. Some 80 percent of these same rail transit commuters reported walking to their workplace from the destination station while 16 percent reported using a bus as their station egress mode. These access and egress mode choice findings would suggest that having both a residence and workplace within walking distance of the high-quality BART transit service is 17-15

OCR for page 10
the most common situation among surveyed Pleasant Hill resident transit users (Lund, Cervero, and Willson, 2004a). Downtown Plano Station, Texas. The typically closer station spacing of LRT and bus rapid tran- sit (BRT) affords opportunity to engage in station function differentiation to an extent that may not be practical for HRT. While BART's Pleasant Hill HRT next-to-end-of-line station was designed with extensive park-and-ride parking, the Downtown Plano next-to-end-of-line LRT station north of Dallas on the Dallas Area Rapid Transit (DART) system was planned to function with only pas- senger drop-off/pickup (kiss-and-ride) parking. Park-and-ride parking was assigned to the adja- cent stations. The Plano station serves the old downtown's well-preserved farming-center main street and also two new downtown TOD components. Of these, Eastside Village I is immediately adjacent to the station and Eastside Village II is about two blocks away. TOD parking requirements were set at 75 percent of normal city of Plano standards. Densities are modest in this suburban location. The year 2000 population density within 1/4 mile of the center of downtown was 5 persons per acre, essentially the same as the 2003 population density for Plano as a whole. The estimated central area resident population density will reach 9 persons per acre at 98 percent occupancy of the 463 dwelling units added by Eastside Village I and II along with 40,000 square feet of non-residential space. The planning goal is 13 persons per acre (Turner, 2003). Table 17-5 compares average weekday passenger boardings for the Downtown Plano station and the two adjacent LRT stations, Parker Road (the end-of-line station) and Bush Turnpike (nearer Dallas). The Downtown Plano station, at one point planned as a stop for special events only, recorded 773 average weekday boardings during May 2006. The boardings were up 16 percent in a year, during a period that saw neighboring station increases of 17 percent (Parker Road) and 6 percent (Bush Turnpike) in the context of increasing gasoline prices and a systemwide 8 percent ridership increase. Longer-term growth statistics are provided in Table 17-5. Downtown Plano station ranks 21st in an array of 24 non-downtown, non-terminal DART LRT stations that ranged in average May 2006 weekday ridership from 3,291 (Mockingbird Station in Dallas with TOD and park-and-ride) down to 334 boardings (Turner, 2003; Dallas Area Rapid Transit, 2006b; Hufstedler, 2006). 17-16

OCR for page 10
Table 17-5 Parking Spaces and Ridership for Downtown Plano and Adjacent Stations DART LRT Park-and- Average Weekday Boardings 3-Year Station Ride Spaces May 2003 May 2004 May 2005 May 2006 Growth Parker Road 1,555 2,766 3,076 2,872 3,349 21% Downtown Plano 0a 591 637 668 773 31% Bush Turnpike 778 846 997 1,097 1,163 38% DART LRT System 11,587 60,789 61,496 58,179 62,725 3% b Notes: a With park-and-ride spaces at Parker Road and Bush Turnpike at or near capacity as of 2006, DART Webpages are suggesting free all-day parking in the small downtown Plano public lots as an option. b The majority of the DART LRT system stations opened in 1996-97, more than sufficient time for ridership to fully mature prior to 2003-2006. In contrast, service to Downtown Plano and adjacent stations began in December 2002. Sources: Dallas Area Rapid Transit (2006a and b), Hufstedler (2006). The available Plano information does not support quantitative TOD-specific travel demand computations, but comparing the non-end-of-line station ridership volume and growth figures is nonetheless instructive. The comparison suggests that both the park-and-ride-oriented approach at Bush Turnpike Station and the moderate-density, TOD-oriented approach without park-and-ride parking at the Downtown Plano Station are viable suburban ridership development options and presumably a good pairing for adjacent stations. City Center Versus Suburban TOD Comparisons Generally available examples of TODs, such as those presented above, provide individualized snapshots in varied contexts. They do not offer the consistency needed to support direct compar- ison of the travel characteristics exhibited by city center TODs relative to outlying TODs. The special research described in the "Portland, Oregon, Metro Region TOD Travel Effects Investigation" case study is a rare instance where direct comparison is facilitated, at least as far as non-work travel (the focus of the research) is concerned. In addition, survey data from the Washington, DC, region provides reasonably consistent mode share comparisons inclusive of commute trips. Portland, Oregon, Non-Work Travel. Table 17-6 provides summary comparisons of Portland area non-work travel characteristics for small areas (traffic analysis zones) identified as "Central Area TOD" versus areas identified as "Outlying TOD." The summary also provides comparisons with non-TOD areas. The non-TOD observations are not stratified into central area and outlying given the small number of zones classified as non-TOD within the cen- tral area. The judgmental process used for traffic analysis zone classification is described in the case study. The TOD definition employed includes under the TOD umbrella any neighborhood, whatever the era of its development, that exhibits a satisfactory array of TOD- like physical and transit service characteristics. The summarized regional survey data were obtained at a time that only Portland's Eastside Blue Line LRT was in operation. The outlying 17-17

OCR for page 10
TOD examples do not, therefore, include today's Westside Line examples and only a few were dominantly LRT-served. Most were entirely or primarily served by buses, including bus routes utilizing Portland's downtown bus mall. Table 17-6 Portland, Oregon, 1995 Observed Non-Work-Trip Mode Share Percentages for Central Area TOD Versus Outlying TOD and for Non-TOD Walk or Bike Public Transit Auto Driver/Passenger P's A's P's A's P's A's Central Area TOD 33% 18% 8% 7% 59% 75% Outlying TOD 14% 18% 2% 2% 84% 80% Non-TOD (Central & Outlying) 8% 8% 1% 1% 91% 91% Central Area TOD Share as a Per- 246% 98% 346% 472% 70% 93% centage of Outlying TOD Share Central Area TOD Share as a Per- 446% 237% 646% 828% 64% 82% centage of Non-TOD Share a Outlying TOD Share as a Percen- 181% 241% 187% 167% 92% 88% tage of Non-TOD Share Notes: "P's" = Non-Work Trip Productions (trips observed at the home end of a trip). "A's" = Non-Work Trip Attractions (trips observed at a non-home end of a trip). Mode share percentages in the first three data rows of this table cannot, because of rounding, be used to precisely compute the share comparison percentages provided in the last three data rows. The values provided are computed from observations, not from rounded percentages. a As most of the zones classified as non-TOD lie in the outlying area, with few in the central area, this statistic potentially lends itself to "apples and oranges" comparisons. It should be interpreted and used only in a central area TOD versus outlying area non-TOD context. Source: Derived from Table 17-47, the source for which is Evans and Stryker (2005). Table 17-6 demonstrates that in 1995 central area TOD residents exhibited about 2-1/2 times the likelihood of choosing walk or bike for their non-work travel mode as outlying TOD residents. Persons attracted to non-work activities in central area and outlying TODs had roughly the same propensity to choose a non-motorized travel mode, at 18 percent of trips in either case. Whether made by residents or not, non-work trips to or from central area TODs were roughly four times as likely to be made on public transit as trips to or from outlying TODs. Conversely, Portland central area TOD likelihood of using an auto for non-work travel was 70 percent (trips by residents) to 93 percent (trips attracted) of the likelihood for trips from and to outlying TODs. Lest outlying TODs be dismissed as inconsequential, however, note that compared to non-TOD areas Portland's 1995 outlying TODs were associated with roughly twice the propensity to choose a non-motorized non-work travel mode and almost twice the likelihood of choosing public transit. Outlying TOD auto use for non-work travel was on the order of 90 per- cent of the proportion associated with non-TOD areas. 17-18

OCR for page 10
Relative mode shares such as these do not speak directly to causality. Socio-demographic characteristics of the residents probably play a role along with TOD design and location. To examine the likely importance of this possibility, the Portland research data were paired with socio-demographic data. Average household size was found to be 1.7 for central area TODs, 2.5 for outlying TODs, and 2.6 for non-TOD areas. The smaller household sizes for TOD dwellers, notably in the central area, would have a largely unknown effect on mode choice. Smaller households would logically result in generation of fewer trips, but that is an issue beyond the scope of Table 17-6, which focuses on mode shares. Average reported 1995 household annual income was found to be approximately $25,000 for central area TODs, $31,000 for outlying TODs, and $37,000 for non-TODs. This is a noticeable difference, and likely contributes to the travel demand outcomes, even though the income differences are diluted at the individual level by household size relationships. Average auto ownership was found to be 1.09 vehicles per household for central area TODs (including the 26 percent of households with no vehicle), 1.67 for outlying TODs (8 percent no vehi- cle), and 1.91 for non-TODs (4 percent no vehicle). The auto ownership rate differences surely relate in significant measure to household size and income, and possibly to the lesser need to make auto trips in TODs. Interestingly, the average number of vehicles per person is essentially the same for central area TODs (0.64) as for outlying TODs (0.66). Overall the socio-demographic differences found are sufficient that the importance of development layout and location in producing the observed travel choice differences among central area TODs, outlying TODs, and non-TOD areas in Portland cannot be ascertained on the basis of data tabulations alone. Washington Region Station-Area Mode Shares. Table 17-7 provides a comparison of mode shares (1) in the Washington, DC, central business district (CBD); (2) outside the CBD but inside the Capital Beltway (I-495); and (3) outside the Beltway, for office and residential sites within or close to a 1/2-mile radius of Metrorail HRT stations. These shares are from the "2005 Development- Related Ridership Survey" by the Washington Metropolitan Area Transit Authority (WMATA). About one-half of the sites were judged to be TOD in nature. The Capital Beltway serves as a handy, albeit imperfect, line of demarcation between Washington and its older suburbs (such as Arlington and Bethesda) and newer suburbs and edge cities (such as Bowie, Gaithersburg, and Tysons Corner). The study cautions that the mode share averages obtained are not area averages but rather averages of the specific sites surveyed, which included 17 office sites and 18 residential sites in total. Only two of each were in the CBD. Survey response rates ranged from 4 to 51 percent at individual office sites, averaging 15 percent, and 6 to 28 percent at individual residential sites, averaging 12 percent. 17-19