National Academies Press: OpenBook
« Previous: Underlying Traveler Response Factors
Page 101
Suggested Citation:"Related Information and Impacts." National Academies of Sciences, Engineering, and Medicine. 2003. Traveler Response to Transportation System Changes Handbook, Third Edition: Chapter 15, Land Use and Site Design. Washington, DC: The National Academies Press. doi: 10.17226/24727.
×
Page 101
Page 102
Suggested Citation:"Related Information and Impacts." National Academies of Sciences, Engineering, and Medicine. 2003. Traveler Response to Transportation System Changes Handbook, Third Edition: Chapter 15, Land Use and Site Design. Washington, DC: The National Academies Press. doi: 10.17226/24727.
×
Page 102
Page 103
Suggested Citation:"Related Information and Impacts." National Academies of Sciences, Engineering, and Medicine. 2003. Traveler Response to Transportation System Changes Handbook, Third Edition: Chapter 15, Land Use and Site Design. Washington, DC: The National Academies Press. doi: 10.17226/24727.
×
Page 103
Page 104
Suggested Citation:"Related Information and Impacts." National Academies of Sciences, Engineering, and Medicine. 2003. Traveler Response to Transportation System Changes Handbook, Third Edition: Chapter 15, Land Use and Site Design. Washington, DC: The National Academies Press. doi: 10.17226/24727.
×
Page 104
Page 105
Suggested Citation:"Related Information and Impacts." National Academies of Sciences, Engineering, and Medicine. 2003. Traveler Response to Transportation System Changes Handbook, Third Edition: Chapter 15, Land Use and Site Design. Washington, DC: The National Academies Press. doi: 10.17226/24727.
×
Page 105
Page 106
Suggested Citation:"Related Information and Impacts." National Academies of Sciences, Engineering, and Medicine. 2003. Traveler Response to Transportation System Changes Handbook, Third Edition: Chapter 15, Land Use and Site Design. Washington, DC: The National Academies Press. doi: 10.17226/24727.
×
Page 106
Page 107
Suggested Citation:"Related Information and Impacts." National Academies of Sciences, Engineering, and Medicine. 2003. Traveler Response to Transportation System Changes Handbook, Third Edition: Chapter 15, Land Use and Site Design. Washington, DC: The National Academies Press. doi: 10.17226/24727.
×
Page 107
Page 108
Suggested Citation:"Related Information and Impacts." National Academies of Sciences, Engineering, and Medicine. 2003. Traveler Response to Transportation System Changes Handbook, Third Edition: Chapter 15, Land Use and Site Design. Washington, DC: The National Academies Press. doi: 10.17226/24727.
×
Page 108
Page 109
Suggested Citation:"Related Information and Impacts." National Academies of Sciences, Engineering, and Medicine. 2003. Traveler Response to Transportation System Changes Handbook, Third Edition: Chapter 15, Land Use and Site Design. Washington, DC: The National Academies Press. doi: 10.17226/24727.
×
Page 109
Page 110
Suggested Citation:"Related Information and Impacts." National Academies of Sciences, Engineering, and Medicine. 2003. Traveler Response to Transportation System Changes Handbook, Third Edition: Chapter 15, Land Use and Site Design. Washington, DC: The National Academies Press. doi: 10.17226/24727.
×
Page 110
Page 111
Suggested Citation:"Related Information and Impacts." National Academies of Sciences, Engineering, and Medicine. 2003. Traveler Response to Transportation System Changes Handbook, Third Edition: Chapter 15, Land Use and Site Design. Washington, DC: The National Academies Press. doi: 10.17226/24727.
×
Page 111
Page 112
Suggested Citation:"Related Information and Impacts." National Academies of Sciences, Engineering, and Medicine. 2003. Traveler Response to Transportation System Changes Handbook, Third Edition: Chapter 15, Land Use and Site Design. Washington, DC: The National Academies Press. doi: 10.17226/24727.
×
Page 112
Page 113
Suggested Citation:"Related Information and Impacts." National Academies of Sciences, Engineering, and Medicine. 2003. Traveler Response to Transportation System Changes Handbook, Third Edition: Chapter 15, Land Use and Site Design. Washington, DC: The National Academies Press. doi: 10.17226/24727.
×
Page 113
Page 114
Suggested Citation:"Related Information and Impacts." National Academies of Sciences, Engineering, and Medicine. 2003. Traveler Response to Transportation System Changes Handbook, Third Edition: Chapter 15, Land Use and Site Design. Washington, DC: The National Academies Press. doi: 10.17226/24727.
×
Page 114
Page 115
Suggested Citation:"Related Information and Impacts." National Academies of Sciences, Engineering, and Medicine. 2003. Traveler Response to Transportation System Changes Handbook, Third Edition: Chapter 15, Land Use and Site Design. Washington, DC: The National Academies Press. doi: 10.17226/24727.
×
Page 115
Page 116
Suggested Citation:"Related Information and Impacts." National Academies of Sciences, Engineering, and Medicine. 2003. Traveler Response to Transportation System Changes Handbook, Third Edition: Chapter 15, Land Use and Site Design. Washington, DC: The National Academies Press. doi: 10.17226/24727.
×
Page 116
Page 117
Suggested Citation:"Related Information and Impacts." National Academies of Sciences, Engineering, and Medicine. 2003. Traveler Response to Transportation System Changes Handbook, Third Edition: Chapter 15, Land Use and Site Design. Washington, DC: The National Academies Press. doi: 10.17226/24727.
×
Page 117
Page 118
Suggested Citation:"Related Information and Impacts." National Academies of Sciences, Engineering, and Medicine. 2003. Traveler Response to Transportation System Changes Handbook, Third Edition: Chapter 15, Land Use and Site Design. Washington, DC: The National Academies Press. doi: 10.17226/24727.
×
Page 118
Page 119
Suggested Citation:"Related Information and Impacts." National Academies of Sciences, Engineering, and Medicine. 2003. Traveler Response to Transportation System Changes Handbook, Third Edition: Chapter 15, Land Use and Site Design. Washington, DC: The National Academies Press. doi: 10.17226/24727.
×
Page 119
Page 120
Suggested Citation:"Related Information and Impacts." National Academies of Sciences, Engineering, and Medicine. 2003. Traveler Response to Transportation System Changes Handbook, Third Edition: Chapter 15, Land Use and Site Design. Washington, DC: The National Academies Press. doi: 10.17226/24727.
×
Page 120
Page 121
Suggested Citation:"Related Information and Impacts." National Academies of Sciences, Engineering, and Medicine. 2003. Traveler Response to Transportation System Changes Handbook, Third Edition: Chapter 15, Land Use and Site Design. Washington, DC: The National Academies Press. doi: 10.17226/24727.
×
Page 121

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

15-101 of Strategy” — “Site Design” — “Transit Supportive Design and Travel Behavior.” Also, the authors acknowledge that key neighborhood descriptors were measured in the aggregate (one value per neighborhood). The attitudinal variables may have attracted an unknown proportion of their apparent power by virtue of having been the best measured variables. Another unknown is the extent to which attitudes are shaped by the environment, and the options it provides, as compared to the thesis that attitudes are externally derived. While the relative significance of individual attitudes as illustrated in the rows of Table 15-45 may be used with normal caution, it would seem that extra care should be applied in considering the broader findings of relative importance among overall categories of variables, the attitudinal variable category included. An issue related to attitudes is the possible or even likely existence of a predisposition on the part of some people or families to behave in ways observed to be more predominant in transit and pedestrian friendly communities, namely, to walk and take transit more, to shop locally, and to drive less. It has been postulated that significant percentages of families who choose to live in TODs, other urban high density areas, and pedestrian friendly communities are comprised of persons so inclined, and that this is perhaps why such communities exhibit the travel characteristics they do. This relationship is a possibility not well explored. Even if the postulate were to be proven, it might not have negative significance for urban travel, since just because one has a predisposition toward a particular lifestyle and set of travel behaviors does not mean it would be practicable to act that way in an unsupportive land use and transportation environment. This issue area is examined further in Chapter 17, “Transit Oriented Development.” RELATED INFORMATION AND IMPACTS Examples of Residential Densities Within this chapter and throughout related literature, transportation and land use findings are often presented in terms of residential densities. To help provide context, residential density examples and density conversion factors are provided here. Table 15-46 gives average residential densities for selected U.S. cities and suburbs. Some of the example densities may differ from seemingly corresponding statistics presented elsewhere in this chapter, as a result of being for different years or geographic area delineations. Averages like those presented in Table 15-46 present only a partial picture, because the averaging washes out variations among neighborhoods. The data in Table 15-47 give an idea of these variations by presenting the percentage distribution of residents among ranges of gross residential densities, calculated including all land uses in the area measure. Both Tables 15-46 and 15-47 are presented in terms of gross residential densities, calculated including all land uses in the areal measure. To convert such values to residential densities expressed with only residential uses in the areal measure obviously requires areal data broken out by residential and non-residential.

15-102 Table 15-46 Gross Residential Densities of Selected U.S. Cities and Suburbs, 1989-1990 City Persons per Sq. Mile Corresponding Suburbs Persons per Sq. Mile Other Locales Persons per Sq. Mile New York City 23,699 (within New York State) 2,558 City of Philadelphia 11,739 San Francisco 15,503 (East Bay suburbs in Alameda County) 3,527 City of Minneapolis 6,703 Chicago 12,254 (within Illinois) 3,483 City of Portland, OR 3,504 Miami 10,084 (within Dade County) 4,884 City of Tulsa, OK 2,000 Los Angeles 7,426 (within Los Angeles County) 5,884 So. CA Cities in San Bernardino County 1,934 Notes: Individual city densities based on 1990 U.S. Census. Suburban and San Bernardino County city densities derived from unpublished 1989 population density data from Aetna. Source: Downs (1994). Table 15-47 Distribution of Residents by Gross Residential Densities for Selected U.S. Suburban Areas, 1989 Residents per Gross Sq. Mile New York Suburbs of N.Y. City Illinois Suburbs of Chicago Dade County, FL Suburbs Orlando Area Suburbs Los Angeles County Suburbs Cities of San Bernardino County, CA 2,500 or fewer 24.6% 13.6% 7.4% 36.9% 4.9% 28.7% 2,500 – 4,999 27.6 55.4 29.3 60.9 11.0 65.8 5,000 – 7,499 20.5 20.6 33.0 2.1 29.8 5.6 7,500 – 9,999 16.7 3.9 19.9 0 22.4 0 10,000 or more 10.5 6.5 10.3 0 31.9 0 Avg. Density 2,558 3,483 4,844 1,422 5,884 1,934 Notes: Derived from unpublished 1989 population density data from Aetna. Source: Downs (1994). One source suggests that the range for proportion of land devoted to residential use is from roughly 25 percent for high gross densities (based on 1970s Manhattan data) to 50 percent for relatively low densities (based on comparable Sussex County, New Jersey, data). These particular percentages, however, exclude even local parks and streets from being counted as residential use, producing a net measure (Downs, 1992). Applying conversions given below suggests that the comparable range expressed entirely in gross densities would be very roughly 35 to 65 percent residential. This matches with an American Planning Association overall estimate for the 1990s of 48 percent residential, although not as well with another estimate of 55 to 65 percent (Eager, 2002).

15-103 Residential densities calculated with non-residential uses excluded are found presented both as population densities per gross residential square mile, and as dwelling units (DUs) per gross residential acre, or per net residential acre. At the 1998 national average household size of 2.6 persons per household, persons per square mile may be divided by 1,664 to obtain DUs/gross acre.10 Thus 5,000 persons per square mile equals 3 DUs/gross acre, 15,000 persons per square mile equals 9 DUs/gross acre, etc. The conversion between gross and net residential density (which excludes land area for streets, etc.) is not a constant. One set of conversion factors (derived from 4 earlier studies) may be individually divided into net residential density to obtain gross residential density, or multiplied by gross residential density to obtain net residential density. The conversion factors are 1.25 in the range of 2 to 4 DUs/net acre, 1.30 at 5 to 7 DUs/net acre, 1.33 at 8 to 12 DUs/net acre, 1.40 at 13 to 18 DUs/net acre, and 1.50 at 19 DUs/net acre and above (adapted from Nelson, 2002). Unfortunately not all sources are clear as to units of measure, such as gross overall acreage including other uses, versus gross residential acreage, versus net residential acreage, or equivalent calculations at the square mile level of measure. Transit Service Feasibility Guidelines The complex interactions between transit ridership and development density have been discussed in the preceding “Underlying Traveler Response Factors” section, under “Transportation Service Levels.” Research findings on how transit ridership varies with density were presented earlier, in “Response by Type of Strategy” — “Density” — “Density Related to Transit Use.” Related studies have taken the complementary step of estimating at what densities different types of transit service are most likely to be appropriate. These studies take care to warn that such values are only suitable for use as general guidelines. Differing degrees of willingness to fund service from non-farebox sources will markedly affect what is “feasible.” Representative estimates of density thresholds for transit service of different types are presented here, followed by an identification of development design features that also have significant impact on practicality of bus operation. Density Thresho lds for Transit Service Research by Pushkarev and Zupan in the mid 1970s confirmed that density and distance to a city’s downtown were critical factors in justifying transit capital investments and operating cost commitments. The authors crafted a series of carefully qualified conclusions on the relationship between downtown size and different types of transit services to consider, and between residential densities and practicality of service. This seminal work was incorporated into a set of transit policy guidelines, presented here in consolidated form in Table 15-48. The researchers stress the conclusion, “High residential density by itself will do little for transit if there is no dominant place to go.” (Cervero and Radisch, 1995; Parsons Brinckerhoff, 1996a; Pushkarev and Zupan, 1982). Similarly, the Institute of Transportation Engineers in 1989 recommended the minimums summarized in Table 15-49 for correspondence between levels of transit service, and both residential density and employment center size (Parsons Brinckerhoff, 1996a). 10 One square mile = 640 acres = 259 hectares (metric).

15-104 Table 15-48 Transit Modes Related to Potentially Suitable Downtown Size Ranges and Minimum Appropriate Residential Densities Mode Service Levels Downtown Size Range Minimum Residential Density Local Bus (minimum) 1/2 mile between routes 20 buses/day 5-8 million square feet of non-residential floorspace 4 dwelling units (DU)/ residential acre Local Bus (intermediate) 1/2 mile between routes 40 buses/day 7-18 million square feet of non-residential floorspace 7 DU/residential acre ± (depends on downtown size and distance away) Local Bus (frequent) 1/2 mile between routes 120 buses/day 18-70 million sq. ft. of non- residential floorspace 15 DU/residential acre Express Bus (walk-on) 5 buses/2-hour peak period 50 million sq. ft. & up of non-residential floorspace 15/DU/res. acre average, 2 sq. mi. tributary area Express Bus (park-ride) 5-10 buses/2-hour peak period 20 million sq. ft. & up of non-residential floorspace 3/DU/res. acre average, 20 sq. mi. tributary area Light Rail 5-min peak headways or better 35-200 million sq. ft. of non-residential floorspace 9 DU/res. acre average, 25-100 sq. mi. corridor Rapid Rail (Metro) 5-min peak headways or better 70 million sq. ft. & up of non-residential floorspace 12 DU/res. acre average, 100-150 sq. mi. corridor Commuter Rail 20 trains/day 70 million sq. ft. & up of non-residential floorspace 1-2 DU/residential acre along an existing RR track Notes: Downtown size ranges are scaled from Figure 1 in the source document, omitting extensions of the ranges “feasib[le] only under unusual conditions.” Downtown is defined as a contiguous agglomeration of non-residential use (larger than the CBD as typically specified). Source: Pushkarev and Zupan (1982). Table 15-49 ITE Recommended Minimums — Transit Service Versus Residential Densities and Employment Center Size Mode & Service Level Residential Density Employment Center Size 1 bus/hour 4 to 6 dwelling units/residential acre 5 to 8 million square feet of commercial/office space 1 bus/30 min. 7 to 8 dwelling units/residential acre 8 to 20 million square feet of commercial/office space Light Rail/Feeder Buses 9 dwelling units/residential acre 35 to 50 million square feet of commercial/office space Source: Holtzclaw (1994) as presented in Parsons Brinckerhoff (1996a).

15-105 With regard to both sets of minimums presented in Tables 15-48 and 15-49, it bears reiteration that feasibility decisions — especially for low intensity bus service — are also heavily dependent on local funding and service policies. An agency willing to accept coverage of 15 percent of costs from fares will have a quite different standard for what transit service area coverage and frequencies are appropriate and feasible than one requiring, say, 50 percent cost coverage from fares. Sketch planning models for evaluating transit proposals take the next step beyond land use density thresholds by using smaller area population and employment information, along with other distance and/or service level parameters. Sketch planning procedures for estimating light rail transit and commuter rail ridership are covered in the “Related Information and Impacts” sections of Chapter 7, “Light Rail Transit,” and Chapter 8, “Commuter Rail.” Most major investment studies and some service design evaluations take the further step of demand estimation and analysis, utilizing full-scale regional transportation models and accompanying detailed network and land use representations. Design Features Supportive o f Transit Service Design as well as density is a significant factor in transit service practicality. Roadway layout within a development is a determinant of the cost effectiveness with which bus service can be provided, and affects likely success in garnering riders. Bus routes that must double back or take indirect routes to exit an area are both inefficient, wasting bus vehicle-miles and driver hours, and disliked by through riders. Dendritic roadway layouts have few through segments and many dead-ends. Any roadway pattern based on a “closed” system with only a single major entrance may not receive bus service, even if there is ridership potential, because added costs of service would not be matched by commensurate benefits. A mitigating feature, all too rarely used, is to invest the capital cost to provide a short bus-only connection allowing through service (see Chapter 4, “Busways, BRT, and Express Bus”). In contrast, when an area roadway system permits reasonably direct, through operation, the operating and capital costs of adding service may be low. Routes can then readily be established that operate in proximity to concentrations of activity. With low service costs, a route can often be provided even if initial ridership is light, creating the initial market. Additional service may then be added as demand grows. The location of activities and pedestrian access relative to streets on which buses can operate is another design feature that can effect ridership. Automobile age residential neighborhood design practice in many jurisdictions dictates that single family units should not front on arterial roadways. The separation created has — in many parts of the country — become even stronger in recent years, with barrier walls erected along busy arterials to buffer housing. Such walls can, if unbroken, force transit riders to use lengthy and indirect paths when walking to and from bus stops. In regions where walls are prevalent, transit agencies have had some success in encouraging developers to at least provide both access points and pedestrian connections. In some larger developments the practice of separating buildings from roadways has been extended to the major interior roadways — the collector system. Residences are located facing away from the collector, with access from a local street. Commercial buildings are set back from the roadway, typically behind unfriendly parking areas. These practices also increase the distance of transit service from the potential customers.

15-106 In contrast, clustering activities near major streets increases the effective density of travelers adjacent to transit service provided. Entire TOD communities may be designed around major bus routes, or stations of fixed route rail or bus rapid transit. Supportive or negative community design features in combination can make or break the feasibility of transit service, or affect the level that can be provided, with corresponding impact on traveler options and choice of mode. Consumption of Land A concern of many with low-density suburbanization or sprawl is the consumption of land involved. Undue loss of agricultural or natural landscape may result directly from excessive large-lot single-family and other low density zoning and land development, or indirectly (for all practical purposes) from discontiguous, leapfrog development. One quantitative indicator of sprawl is where land area is being consumed at a faster rate than the growth in population. In most U.S. metropolitan areas, urban land area has grown much faster than the population, as shown in Table 15-50 for some sample areas (NTI, 2000): Table 15-50 Expansion of Population Versus Land Area in Selected Metropolitan Areas, 1970 to 1990 Urban Region Population Change Land Area Change Chicago 4% 46% Los Angeles 45% 300% New York City 8% 65% Seattle 38% 87% Source: Diamond, H. L., and Noonan, P. F., Land Use in America, Island Press, Washington, DC (1996); graphic presented in NTI [2000]. A Washington State study adds detail to this aggregate data. In urban counties of the state, between 1970 and 1990, housing densities declined in cities and Census-designated places with more than 2,500 persons. The percentage of metro area population living in places with densities greater than 5,000 persons per square mile was halved, from about 32 percent to some 16 percent, even as older areas grew more dense. Employment densities averaged on the basis of area increased, but decreased on the basis of employee-weighted averages, reflecting conditions encountered by the typical worker. However, the proportion of jobs in the Central Puget Sound region located in districts deemed transit-oriented (at least 50 employees per gross acre and 15,000 or more total jobs) went from 0 in 1970 to 11 percent in 1980 and 13 percent in 1990 (Pivo, Hess and Thatte, 1995). For additional trend information from this study see “Response by Type of Strategy” — “Diversity (Land Use Mix)” — “Jobs/Housing Balance” — “Trip Containment Within Communities.” One metropolitan area that has not exhibited land area growth outpacing population is Portland, Oregon. Portland completed adoption of an urban growth boundary (UGB) in 1980, at the behest of Oregon’s Land Conservation and Development District, as part of a statewide effort to protect land resources and contain sprawl. The boundary encloses roughly 238,000 acres or 365 square miles (sic), with 19,000 additional acres designated for future urban expansion. The perimeter respects the natural and built environment, producing an irregular border some 200 miles in extent. From 1979 through 1997, only about

15-107 7,000 acres of expansion were allowed, a 3 percent land area growth (NTI, 2000). This is with a 1980 to 1997 population growth of 40 percent for Portland/Salem. The UGB has obviously not stopped Portland from experiencing strong growth throughout the 1980s and 1990s. The set of statistics in Table 15-51 suggests this growth has been without many of the attendant transportation-related problems experienced by other areas. Table 15-51 offers a comparison of Portland with its polar opposite in terms of growth management, Atlanta. Despite growth rates of the same order of magnitude, Portland’s growth appears to have come with generally better economic results and less transportation impact: SOV use is shown in Table 15-51 to have declined by 13 percent in a decade compared to a 15 percent increase in Atlanta, and VMT to have increased by only 2 percent, compared to 17 percent in Atlanta (Nelson, 2000). Table 15-51 Comparing Regulatory Regimes for Portland (Urban Containment) Versus Atlanta (Business as Usual) — Change Between Mid-1980s and Mid-1990s Measure Portland Atlanta Population Growth +26% +32% Preferred Direction of Movement Job Growth +43% +37% + Income +72% +60% + Housing Costs as Percent of Income +4% +5% – Government Revenue +34% +56% + Property Tax -29% +32% – Vehicle Miles Traveled +2% +17% – Single Occupant Vehicle -13% +15% – Commute Time -9% +1% – Air Quality in Ozone Days -86% +5% – Energy Consumption per Capita -8% +11% – Home Ownership +8% +1% + Persons per Room -2% -25% – Neighborhood Quality +19% -11% + Notes: Housing/Home data are from a different original source and may not be fully compatible. Over the 5-year period from 1991 to 1996, housing prices in Portland rose 61%, versus 19% in Atlanta, moving Portland’s prices up from 89% of Atlanta prices to 120% of Atlanta prices. Source: Nelson (2000), preferred direction of movement added by Handbook authors. Unfortunately, the jury is still out on Portland’s VMT, at least in view of other sets of statistics. Table 15-52 in the following “Trip Making and VMT” section suggests that the 1982-96 VMT growth on the freeways and principal arterials of greater Portland, including Vancouver, Washington, was 98 percent as compared to 119 percent in greater Atlanta, shown to have twice the percentage population growth in the 14-year period (U.S. Environmental Protection Agency, 2001). Another source, for Metro Portland only, suggests that with decentralization within the urban growth boundary, VMT rose 40 percent between 1980 and 1990, with 15 percent population growth (See Chapter 18, “Parking Management and Supply,” in the case study “CBD Parking Supply Management in Portland, Oregon.”)

15-108 Most planning practitioners view reduced consumption of land as beneficial for reasons including protection of agricultural and recreational lands, water pollution control and groundwater recharge, and habitat and biodiversity preservation (Ewing, 1997). Some, however, view such arguments as a red herring, and maintain that America is not running out of prime farmland and open space (Gordon and Richardson, 1997; O’Toole, 2001). Induced Development and Travel Many factors contribute to decentralization, including rising incomes, auto ownership, and technological innovation. A frequently debated issue is the extent to which expansion of the highway system facilitates decentralization by increasing access to cheaper, undeveloped land at the urban periphery, and hence induces travel. A Transportation Research Board committee studied the issue, and ended in a first-ever split decision documented in Special Report 245, published in 1995. A majority concluded that induced travel effects exist but are small and not large enough to be estimated with current models. The minority opinion felt they are larger (NTI, 2000). Individual studies to determine whether roads do, in fact, create their own demand, have found elasticities of VMT with respect to road supply of +0.1 to +0.9. Some studies specifically differentiated between short-term and long-term effects. One mid-1990s analysis used evidence from several subject areas including travel time budget theory, value of time, and response to fuel price changes, to derive traffic volume elasticities with respect to travel time. The estimated elasticities were -0.5 in the short term and -1.0 in the long term (an elastic, one-for-one response). A more recently published statistical analysis of VMT growth in response to highway lane mile additions arrived at VMT/lane-miles elasticities of +0.5 in the short term and +0.8 in the long term (U.S. Environmental Protection Agency, 2001). Peer-reviewed University of California at Berkeley studies by Hansen involving regression analysis of 1973-1990 time series panel data from 14 metropolitan areas in California, later expanded to 30 urban counties, found a very short-term elasticity of VMT to state highway lane miles of +0.2, building in two to five years to an average elasticity of +0.9 at the metropolitan level, or +0.6 to +0.7 at the county level. Residential development was shown to accelerate in corridors with new highway capacity. The VMT growths per lane mile were substantially higher in the larger metropolitan areas of San Francisco, Los Angeles, and San Diego (NTI, 2000; U.S. Environmental Protection Agency, 2001), which suggests more than a one-for-one response of VMT to lane miles in those areas. There is no indication that any of the various available VMT elasticities are normalized for population or employment growth. Trip Making and VMT VMT Growth Trends Federal Highway Administration statistics indicate that VMT growth in the United States averaged 3.1 percent annually between 1980 and 1996, as compared to an average population growth of 1.0 percent annually. Examples of individual metropolitan area VMT growth on freeways and principal arterials between 1982 and 1996, in comparison to population growth, are listed in Table 15-52 (U.S. Environmental Protection Agency, 2001).

15-109 Table 15-52 1982-1996 Urbanized Area Population and Major Highway VMT Growth Urbanized Area Fourteen-year Population Growth in Percent Freeway and Principal Arterial VMT Growth in Percent Atlanta, GA 53% 119% Boston, MA 6 31 Charlotte, NC 63 105 Chicago, IL-IN 11 79 Houston, TX 28 54 Kansas City, MO-KS 23 79 Miami-Hialeah, FL 18 61 Nashville, TN 25 120 New York, NY-NJ 3 40 Pittsburgh, PA 7 54 Portland-Vancouver, OR-WA 26 98 Salt Lake City, UT 32 129 San Antonio, TX 29 77 Seattle-Everett, WA 35 59 Washington, DC-MD-VA 28 78 Source: “Urban Roadway Congestion, Annual Report 1998,” Texas Transportation Institute, as presented in U.S. Environmental Protection Agency (2001). Many factors contributed to and accompanied this high rate of increase, as illustrated in Table 15-53, which corresponds with estimates of U.S. VMT growth six times population growth for the 1983 to 1990 period. Included are such major demographic trends as increases in workforce participation rates, downsizing of households, and increased real incomes, vehicle ownership and drivers’ licensing rates. Exactly what proportion should be attributed to continuing rapid rates of decentralization of population and jobs during this period is difficult to say with precision. A 1992 study for the Federal Highway Administration by Pisarski attributed the 1983-90 growth in VMT to population growth (13 percent), vehicle occupancy decline (17 percent), increase in person trips per capita (18 percent), mode shifts (17 percent), and increased vehicle trip length (35 percent). Based on review of this attribution and other data, it has been suggested that sprawl could be judged responsible for most of the trip length effect, and some of the mode shift effect, since auto dependency results from dispersed development. This approach leads to an estimate of the proportionate effect of sprawl on VMT growth of 25 to 50 percent (FHWA, 1991; NTI, 2000). Several economists and demographers projected that growth in VMT would level off somewhat in the post-1990 period as a result of saturation in auto ownership rates and various demographic factors, most notably workforce participation and drivers’ licensing. However, VMT has still increased at a rate of 2.5 to 3.4 percent per year — depending on source — with part of the trend attributed to continued trip-length growth associated with ongoing urban area expansion (NTI, 2000).

15-110 Table 15-53 Percentage Changes in Key Factors Accompanying Growth in VMT, 1983- 1990 Contributing Factors Growth Travel Measures Growth Population 6.0% Average Person Trip Length 8.0% Number of Households 9.3 Number of Person Trips 12.6 Number of Workers 14.6 Total Person Miles 13.9 Number of Licensed Drivers 10.9 Average Vehicle Trip Length 13.9 Number of Vehicles 15.0 Number of Vehicle Trips 25.2 Total Vehicle Miles of Travel 40.6 Source: FHWA (1991). Work Versus Non-Work Travel Work Purpose Travel. Results for the five NPTS surveys from 1969 through 1995 indicate that work purpose travel per household grew by 16 percent from 1969 to 1990, an average of 3/4 of 1 percent per year, but with the actual growth focused on the 1983 to 1990 period. It then jumped 34 percent in the first 5 years of the 1990s, an average of almost 7 percent per year. The 1990-1995 jump in large part reflects an increase in the number of workers per household, but increased trip length was also a factor (Hu and Young, 1999). Even discounting the effect of a booming economy in 1995, the quarter-century trend is clearly upward. Authors who have examined or reviewed comparisons between commuting into close-in versus outer areas seem to be in accord that decentralized job locations, greater distances, and displacement from core transit services — all of which accompany land-use sprawl — beget more commute VMT (Miller and Ibrahim, 1998; Rutherford, McCormack and Wilkinson, 1997; Prevedouros and Schofer, 1991). A study by Cervero and Wu combines comparison among San Francisco Bay Area job centers with examination of change over time. To facilitate comparison, the researchers classified employment areas into four groups (NTI, 2000): • Regional CBD: Downtown San Francisco • Mature Suburban Centers (mixed-use): East Bay core (central Oakland and Berkeley) • New Economy Center (large scale office park development): Silicon Valley • Outer Suburban Centers (smaller scale, mainly office development): Suburban Centers Table 15-54 summarizes results. The daily commute VMT per worker in 1990 was 37 percent more for Outer Suburban Centers (10.1 miles) than downtown San Francisco (7.4 miles). The difference reflected mainly higher transit shares for travel to the core (NTI, 2000). Transit share was undoubtedly also a factor in other differentials, along with commute distances. The most striking finding, however, is that the job centers with the higher VMT averages in 1980 are the same ones with the greatest VMT per employee growth between 1980 and 1990. The 1980 to 1990 growth rate in VMT per employee was 62 percent higher in the Outer Suburban Centers (27.7 percent) than in the regional CBD (17.1 percent), for example.

15-111 Table 15-54 Weekday VMT per Employee over Time in Four San Francisco Bay Area Job Centers VMT per Employee Job Center Classification 1980 1990 Percent Change 1980-90 San Francisco Regional CBD 6.30 7.40 17.1% East Bay Core Mature Suburban Centers 7.26 8.56 19.9% Silicon Valley New Economy Center 7.09 8.81 24.1% Suburban Centers Outer Suburban Centers 8.04 10.13 27.7% Total All 7.11 8.74 22.8% Source: Cervero, R., and Wu, I., “Subcentering and Commuting: Evidence from the San Francisco Bay Area, 1980 to 1990,” Urban Studies, Vol. 35 (1998), as graphed in NTI [2000]. Non-Work Purpose Travel. Commute travel growth is unquestionably a significant concern. The work purpose trip is concentrated into a fairly narrow time window, has the longest average trip length aside from social/recreational trips, has perhaps the lowest auto occupancy of all trip purposes, and is most associated with urban freeway congestion and air quality problems. However, the biggest absolute VMT total is for non-work travel. Moreover, non-work travel is a major — or the major — contributor to VMT throughout the day, even dominant in terms of peak travel period trip starts. The quarter century’s worth of NPTS survey results suggests that the proportion of household VMT accounted for by work-related travel has remained fairly constant at around 30 percent. This relative stability means that while household commute trip VMT has grown along with the expansion of metropolitan areas, so has non-work VMT. This phenomenon is evident from the NPTS trend data in Table 15-55. Non-work travel per household expanded by 22 percent from 1969 to 1990, an average of 1 percent per year, although the growth was in the 1980s. It then rose 8 percent from 1990 to 1995, an average of over 1-1/2 percent per year. This sustained non-work VMT growth occurred despite a decrease of 17 percent in average household size, from 3.16 to 2.63, during the 1969 to 1995 period. The biggest VMT increases were for shopping and family/personal business travel (Hu and Young, 1999). Since many non-work travel needs are not highly differentiated as to which opportunity site will accommodate them — one store in a chain should do about as well as another — there must be other reasons for non-work VMT growth besides area expansion per se. There are actually three likely explanations, all of which probably pertain: • An explanation for non-work VMT growth popularized for years is that rising incomes have led to more discretionary travel. • Methodological/definitional survey changes have tended to recategorize some work trips into non-work trips and/or to detect more non-work trips (Research Triangle Institute and FHWA, 1997). The seemingly small share of work trips in the 1995 NPTS starting during rush periods that are actually trips directly to or from work (less than 1 of 3 person trips) was judged to probably reflect trip chaining, where stops are made on the way to or from another destination (FHWA, 1997).

15-112 Table 15-55 Trends in Annual Household VMT by Trip Purpose Trip Purpose 1969 1977 1983 1990 1990 Adjusted a 1995 To/From Work 4,183 3,815 3,538 4,853 4,853 6,492 Work Percentage 34% 32% 30% 32% 27% 31% Shopping 929 1,336 1,567 1,743 2,178 2,807 Family/Personal Business 1,270 1,444 1,816 3,014 4,250 4,307 Social/Recreational 4,094 3,286 3,534 4,060 5,359 4,764 Miscellaneous b 1,947 2,155 1,284 1,430 1,521 2,525 Subtotal, Non-Work 8,240 8,221 8,201 10,247 13,308 14,403 Non-Work Percentage 66% 68% 70% 68% 73% 69% TOTAL 12,423 12,036 11,739 15,100 18,161 20,895 Persons per Household 3.16 2.83 2.69 2.56 2.56 2.63 Notes: a Adjusted for better comparability with 1995 data. Shaded 1990s travel data should not be compared directly with earlier data. b Includes other purposes such as school, church, medical, and work-related. Source: Hu and Young (1999); “Miscellaneous” (determined by subtraction) and percentages by the Handbook authors. • The third explanation for non-work VMT growth is supported by research on travel in traditional neighborhoods as compared to conventional suburban development. (For selected research see “Response by Type of Strategy” — “Site Design” — “Community Design and Travel Behavior”). Most suburban residents today live in conventional suburban development (CSD) housing subdivisions largely separate and often distant from routine shopping, personal services, entertainment, recreational facilities, and even schools. At the extreme, dispersed and segregated land use requires that all of a household’s non-work travel needs be contingent on auto travel. If it’s not practical to walk to a store, to school, to a bank or barber shop, and if children must be chauffeured by parents to school functions, sports, and other activities, higher non-work travel VMT will result. This lack of choice but to drive when satisfying household needs in CSD-type suburbs appears to be one of the key areas where the impact of suburbanized land use patterns is most felt. Not only do non-work trips account for the majority of household VMT, a high percentage occur during time periods most commonly assumed to be dominated by commute travel. NPTS tabulations indicate that, in 1995, 37 percent of all daily person trips occurred in the morning (6:00 to 9:00 AM) or evening (4:00 to 7:00 PM) peak travel periods. About 64 percent of the trips starting in the morning peak period and 78 percent of those starting in the evening peak period were non-work in purpose (FHWA, 1997). Even though non-work trips tend to be shorter and generate less VMT per trip, this finding still suggests that non- work travel may contribute significantly to what is perceived to be commuter-related congestion.

15-113 Research results indicate there is potential to satisfy a higher percentage of non-work trips locally and with less VMT — partly by promoting shorter trip lengths, partly by containing trips on the local, non-arterial network, and some by providing opportunities requiring no auto use at all. This potential is afforded by more compact, mixed use, walkable settings, as uncovered in detailed analysis of places like Rockridge (Cervero and Radisch, 1995) and Queen Anne and Wallingford (Rutherford, McCormack and Wilkinson, 1997). Designs that are responsive to this aspect of traveler behavior may offer a partial but important response to the difficult problem of growing suburban congestion. Trip Making and VMT Differentia ls There is a substantial body of evidence and agreement that households located in denser environments and in traditional, mixed-use, pedestrian-friendly neighborhoods generate less VMT — even after correcting for socio-demographic characteristics — because they make fewer trips as auto drivers and shorter vehicle trips than do households in conventional suburban developments (Burchell, 1998). An important related question is whether these households travel less — or more — overall. Here it is very important to distinguish between person trips by any mode, including walking, and vehicle trips, which for personal travel equate to auto driver trips. Details and important caveats underlying the exposition of trip making and VMT differentials that follows are found in the earlier “Response by Type of Strategy” section. Person Trip Generation. U.S. national data do show variation in the person trip count as densities increase, amounting at most to the 15 percent difference found between the very highest and lowest densities (Dunphy and Fisher, 1996). There is, however, a preponderance of agreement that there is no causal relationship between population or employment densities and frequency of person trip making (Ewing and Cervero, 2001). For example, the comparative studies of “traditional” Rockridge and “conventional” Lafayette in the San Francisco East Bay found a fairly similar number of total daily non-work trips in both places, and concluded that walk trips substituted for auto trips rather than supplementing them (Cervero and Radisch, 1995). The substitution issue is still outstanding, however. At least three studies have identified somewhat elevated person trip generation where population density, employment density, and mixing of land uses are greater. This trip rate elevation has primarily taken the form of additional non-auto (mostly walking) trips and/or trip chaining, and appears to be coupled with vehicle trip generation, VMT, or person miles of travel rates that are lower (Frank, 1994; Kitamura, Mokhtarian and Laidet, 1994; Rutherford, McCormack and Wilkinson, 1997). There are also studies that have found slightly lower person trip rates in “traditional” neighborhoods even controlling for income (for example, McNally and Kulkarni, 1997). Vehicle Trip Generation. In any case, a preponderance of studies have found that vehicle trip generation is damped, to a limited degree, by density, local land use mix, and/or appropriate site design. The few dissenting studies see no effect, rather than opposite effect. The one major exception is that land use balance over broad areas apparently does not affect vehicle trip making; it only affects trip distance and thus VMT (see “Response by Type of Strategy” — “Diversity (Land Use Mix)” — “Jobs/Housing Balance”). Typical elasticities for vehicle trips relative to local population density, diversity (mix), and design have been estimated by meta-analysis (see “Consolidated Vehicle Trip and VMT Elasticities” below). The results of applying these individual elasticities may be summed (Ewing and Cervero, 2001). Assuming concurrent and equal enhancements to the “3 Ds,” the cumulative effect should be equivalent to

15-114 a combined elasticity of about -0.13, excluding any regional accessibility effects. Individual elasticity estimates generally conform to this modest order of magnitude. In Portland, Oregon, both accessibility to jobs and especially quality of the pedestrian environment were found to be negatively related to vehicle trip making, with 0.6 fewer daily vehicle trips per household in areas with a fairly good pedestrian environment, compared to pedestrian-hostile areas (Cambridge Systematics, Putman Associates and Calthorpe Associates, 1992). Within suburban activity centers, employee vehicle trip rates exhibit an elasticity of -0.06 to land use mix (Parsons Brinckerhoff, 1996a). Edge city office buildings with retail had office employee vehicle trip rates 6 or 8 percent lower than without retail (NTI, 2000). Study of large-scale regional shopping centers found vehicle trips may be reduced 1 to 3 percent with improved pedestrian access (JHK and Associates and K.T. Analytics, 1993). Worksites with TDM programs and good availability of on-site services averaged vehicle trip rates 15 percent lower than other worksites with TDM programs (Comsis, 1994). Southern California studies confirmed income as a stronger determinant than community types, but nevertheless found auto trip rates (driver or passenger) in “traditional” TND communities to be 10 to 23 percent less within individual income groupings than for conventional PUDs (McNally and Kulkarni, 1997). The Rockridge versus Lafayette paired community analysis found 12 percent less auto use for non-work travel in the TND neighborhood as compared to the CSD area, and 20 percent less auto use for commuting (Cervero and Radisch, 1995). About half the commute trip differential (10 percent) seems reasonably attributable to neighborhood design. Vehicle Miles of Travel. It is with respect to VMT that all aspects of land use and site design come into play, including not only density, mix, and site design, but also land use balance and street network design. VMT is the product of vehicle trips and trip length. Despite the general agreement that more centralized, compact, diverse, walkable urban areas equate to lower VMT, there are two schools of thought about how much lower. One school is generally represented by studies that allow land use density to stand as a surrogate or marker for all commonly associated urban form characteristics, and the second- order effects that density historically brings with it. Density from mid-20th Century and before is linked with household characteristics associated with higher dependency on public transit, a wider array of choices for meeting daily travel needs ranging from better transit to walking-distance shopping opportunities, and driving and parking conditions and prices that make auto use less attractive (Dunphy and Fisher, 1996). Such density is also associated with central locations, associated trip distribution patterns, and high accessibility to jobs and other “opportunity sites” (see “Response by Type of Strategy” — “Density” — “Density and Other Indicators at the Behavioral Level”). Studies using this all-inclusive approach to density have estimated that double the density is associated with 15 to 30 percent less VMT (NTI, 2000; Cervero and Radisch, 1995; Dunphy and Fisher, 1996). These estimates equate to density elasticities in the -0.23 to -0.5 range. To actually approach such a reduction with “new” density, regional location, urban transportation alternatives, and built environment characteristics historically linked with the higher density would need to be provided (Parsons Brinckerhoff, 1996a; Handy, 1997). This would seem likely only in the case of well located and well designed infill development. Indeed, recent analyses of infill development proposals in Atlanta, San Diego, West Palm Beach, Florida, and Montgomery County, Maryland, done using traditional four-step

15-115 transportation planning models, projected 15 to 52 percent VMT savings for infill sites as compared to suburban greenfield development. A large measure of the variance had to do with which suburban site was assumed for comparison (U.S. Environmental Protection Agency, 2001). The second school of thought defines and analyzes density more precisely, excluding concomitant and second order effects. Typical research elasticities for VMT as a function of this more narrowly defined density are small. The meta-analysis summarized in the next subsection provides an elasticity of VMT to density of -0.05. The results of applying this elasticity estimate are deemed to be additive to results of applying other built environment elasticities derived so as to be isolated from overlapping characteristics and influences (Ewing and Cervero, 2001). Such a density elasticity could reasonably be construed to include effects channeled through auto ownership, but not other second-order effects. The meta-analysis VMT elasticity for density is accompanied by corresponding elasticities for local diversity (-0.05), local design (-0.03), and regional accessibility (-0.20). Again, results of applying the individual elasticities may be added (Ewing and Cervero, 2001). Assuming concurrent and equal enhancements to local density, diversity, and design — holding aside regional accessibility — the cumulative effect should be equivalent to a combined elasticity of about -0.13. Although not valueless by any means, this indicates “that dense, mixed-use developments in the middle of nowhere may offer only modest regional travel benefits” (Ewing and Cervero, 2001). If, however, concurrent and equal enhancements to regional accessibility are also assumed, the cumulative effect becomes equivalent to a combined elasticity on the order of -0.33. This is within the lower end of the range of the elasticities associated with “density” when it is allowed to act as a surrogate for any and all concomitant and second order effects. The flip side of this substantial elasticity is that not only must a fairly optimal mix of density, diversity, and design enhancements be applied, a higher accessibility location must also be obtained either through geographic placement or some combination of that and long range land use and transportation plan implementation. The more rigorous definition of density, along with identification of the benefits of land use mix, site design, and accessibility, has the advantage of disentangling itself from socio- economic influences. It also allows more meaningful policy-sensitive evaluations. It does not, however, automatically bring in the important second order effect of enhanced transit service feasibility that goes along with density and transit-friendly design. The impact of that highly beneficial effect must be calculated separately when using the more rigorous definitions of density and like measures along with their corresponding elasticities. Turning to case example research studies, national data indicates that U.S. daily per capita VMT declines with higher levels of density, differing by roughly a factor of 9 between the lowest and highest density group (Dunphy and Fisher, 1996). This does not in itself indicate a causal relationship for density. Lower household VMT has been found to be associated with locations nearer the central business district, and as identified in greater Toronto, nearer other major employment centers well served by transit (Prevedouros and Schofer, 1991; Miller and Ibrahim, 1998). Among six Florida communities studied, households in the one with the lowest density and accessibility produced 63 percent more vehicle hours of travel than in the one with the highest (Ewing, Haliyur and Page, 1994).

15-116 Much as trip lengths are typically shorter in areas with jobs-housing balance, so are household VMT averages lower, although again causality is not fully understood. Balanced zones in San Diego were found to have 16 to 20 percent shorter work trip lengths than zones with excess jobs or housing (Ewing, 1997). In San Francisco Bay Area cities with a high surplus of jobs, workers selected the drive-alone mode 5 percent more often, encountered commute times 11 percent longer, and produced commute VMT per employee 7 percent higher than in other cities (NTI, 2000). Of studies using hypothetical network analyses to evaluate effect of local street design, the more conservative estimated that TND neighborhoods with the same level of vehicle trip generation as PUDs would have 10 percent less vehicle kilometers of travel (McNally and Ryan, 1993). Average daily travel mileage in Seattle area mixed use neighborhoods measures 10 to 24 percent less than in the corresponding subregional areas. It may be safely presumed that the VMT differential would be at least as large (Rutherford, McCormack and Wilkinson, 1997). Very pedestrian-friendly environments within Portland, Oregon, were found to produce 10 percent less VMT per individual than average environments, holding everything else constant (Cambridge Systematics, Putman Associates and Calthorpe Associates, 1992). Overall, in the San Francisco East Bay paired community evaluation, the average daily VMT per resident was found to be 10.8 miles in “TND” Rockridge versus 19.6 miles in the somewhat further-out “CSD” Lafayette (NTI, 2000). As already noted, further detail on individual research efforts is provided in the earlier “Response by Type of Strategy” section. Consolidated Vehicle Trip and VMT Elasticities In collaboration, researchers Ewing and Cervero have synthesized results from a large number of land use and site design studies, and have also developed a consolidated set of elasticities of travel with respect to the built environment. Development of these elasticities involved going back to original sources, re-analyzing the data in selected cases, and developing “typical” partial elasticities in a meta-analysis. The elasticity results are reproduced in Table 15-56. The researchers state that the results “are partial elasticities, which control for other built environment variables when estimating the effect of any given variable. Hence, the elasticities should be additive” (Ewing and Cervero, 2001).11 11 The density, diversity, design, and regional accessibility changes must be concurrent and equal for the elasticities to be truly mathematically additive, in other words, they must be applied together in equal percentages (for example, a concurrently applied 10 percent enhancement in each of density, diversity, design, and regional accessibility). Otherwise, it is only the results of application of the individual elasticities that are additive. In any case, thinking of the elasticities as additive is certainly appropriate for visualizing approximate scale.

15-117 Table 15-56 Typical Elasticities of Travel with Respect to the Built Environment Urban Form Characteristic Nature of Measure Vehicle Trips (VT) Vehicle Miles of Travel (VMT) Local Density Residents plus employees, ÷ by land area -0.05 -0.05 Local Diversity (Mix) Jobs/population balance -0.03 -0.05 Local Design Sidewalk completeness, route directness, street network density -0.05 -0.03 Regional Accessibility Index with a gravity model derivation — a -0.20 Notes: These values represent a refinement of those incorporated into the 2001 version of EPA’s Smart Growth Index® (SGI) model as presented in Criterion Planners/Engineers and Fehr & Peers Associates (2001). This model is intended to supplement underspecified four-step regional transportation planning models, adding responsiveness to the built environment. a Ewing and Cervero (2001) infer a nil elasticity value; in Fehr & Peers Associates (2002) a -0.05 value is offered in a Minneapolis – St. Paul context, based on a -0.036 estimate from Kockelman (1997) along with other findings suggesting a somewhat stronger relationship. Sources: Ewing and Cervero (2001 and 2002), Fehr & Peers Associates (2002), Walters (2002). It would appear that these elasticities probably encompass certain second order effects, such as the impact of density on vehicle trip generation that is channeled via auto ownership, but not others, such as the impacts of enhanced transit service made feasible by higher densities. The non-inclusion of transit service impacts is underscored by the fact that the elasticities were developed (in a slightly earlier version) for use with EPA’s Smart Growth Index® (SGI) model, intended to supplement four-step regional transportation planning models lacking sensitivity to one or more primary urban form parameters (Criterion Planners/Engineers and Fehr & Peers Associates, 2001). The regional models would be relied upon to introduce transit service level effects, and it would be up to the analyst to specify service levels appropriate to the land use densities being tested. Note that regional accessibility, a function of both density and land use mix, as well as the transportation system, is treated separately from local area indicators of urban form. The authors note (Ewing and Cervero, 2001): “Advocates of urban planning and design will be disappointed that the values are not larger. Those skeptical of public policy interventions will be equally disappointed, as the elasticity values are significantly different from zero in most cases and, when summed across regional accessibility, density, diversity (mix), and design, suggest fairly large cumulative effects.” Energy and Environmental Impacts Land use and urban form obviously are linked to a broad array of impacts ranging from consumption of land to water quality to societal effects. In context with this Handbook’s concentration on travel behavior-related information and issues, the examination here of resource and environmental linkages is focused on automotive energy consumption and the air quality effects of auto transportation emissions. Mass transit energy consumption and emissions are consequential, but not well studied in a land use context, a notable limitation even in an overview as offered here.

15-118 Energy Literature review indicates many researchers conclude sprawl development is associated with higher auto, and hence energy, use. However, there is a minority not convinced that a link has been established, or that energy conservation is all that important (Gordon and Richardson, 1997; O’Toole, 2001). The majority find that the relationship between energy consumption and urban form parallels that of travel, with compact development patterns consistently outperforming low density sprawl in minimization of auto use and transportation energy consumption. Though vehicles operate with less fuel efficiently in congested areas, per capita fuel consumption is substantially lower in central cities because people drive much less (Ewing, 1997). One study result, illustrated in Figure 15-8, shows travel to be faster in low-density than high- density areas, due to higher running speeds. However, the resulting fuel savings per vehicle mile are more than offset by longer trips and more motorized travel (NTI, 2000). Figure 15-8 The speed-energy tradeoff in outer, inner, and Central City areas Source: Newman, P. W. G., and Kenworthy, J. R., “The Transport Energy Trade-Off: Fuel-Efficient Traffic versus Fuel-Efficient Cities.” Transportation Research A, Vol. 22A (1988) as presented in NTI [2000].

15-119 These conclusions notwithstanding, in larger urban areas, the central city is relatively less accessible to development in the outer communities. At some size of metropolitan area, the emergence of other centers besides the center city CBD is beneficial from the standpoint of transportation and energy. When studies include polycentric development as an alternative, that emerges as the preferred energy-efficient settlement pattern. Thus in large metropolitan areas, energy efficiency is served by development concentration, but not to the extent of adhering to a single dominant center (Ewing, 1997). Air Quality Determinations relating to air quality are inherently more complex than in the case of energy consumption. The incidence of urban air pollution is influenced by many factors besides regional automotive travel totals, including placement of major VMT concentrations within the metropolitan area, location of mountain barriers, and general climate including wind speeds and direction and temperature inversions. Thus there is substantial disagreement on whether the benefits of reduced automotive travel obtained from compact development are or are not counterbalanced by problems of pollutant concentrations. New Jersey State Development and Redevelopment Plan impact assessment found, for example, that most emissions reduction would derive from more stringent and effective emissions controls, and that air quality enhancement would be correspondingly enhanced under either sprawl or compact development scenarios (Burchell et al., 1998). Others argue that growth of vehicle trips and VMT will wipe out emissions reductions gained through further vehicle emissions control improvements. As with fuel consumption, total vehicle emissions increase with increased VMT. They decrease — up to a point — as average operating speeds increase. Above 35 mph in the case of NOx, and 50 mph for carbon monoxide and hydrocarbons, emissions increase. Carbon dioxide emissions track fuel use at all speeds (Ewing, 1997). State of Washington sponsored research has used Puget Sound Transportation Panel travel survey data for 1996 in conjunction with Census tract level statistical analysis and modeling of housing and employment density, mix, travel, and pollutant emissions interactions to investigate emissions effects of land use differentials. Household size, auto ownership, and income were controlled for. Both household density and workplace employment density were estimated to be significantly linked with lower CO, VOC and NOx emissions. Street connectivity, and mixed uses in residential settings — as identified by density of residential area employment — were either associated with lower emissions or had negligible effect. Detracting from reduced emissions was a degree of positive correlation of connectivity and mix with vehicle trip generation and cold starts. Overwhelming this apparent disadvantage were lower VMT/capita rates associated with the same land use factors (Frank, Stone and Bachman, 2000). At the regional level, the comparison of Portland, Oregon, with its urban growth boundary, and Atlanta, which in the 20th Century grew under a “business as usual” approach, is instructive. The figures presented earlier in Table 15-51 indicate that between the mid-1980s and the mid-1990s, Portland experienced an increase in VMT of 2 percent compared to Atlanta’s 17 percent, had an 8 percent reduction in energy consumption per capita versus Atlanta’s 11 percent increase, and an 86 percent reduction in Ozone Days versus Atlanta’s 5 percent increase (Nelson, 2000). As noted in the discussion accompanying Table 15-51, in the subsection on “Consumption of Land,” potentially conflicting VMT data for Portland

15-120 exist. Conceivably the energy consumption information is likewise open to question, but the Ozone Days comparison should be independent of these concerns. Cost Effectiveness Observers of the fiscal requirements and impacts of sprawl versus compact forms of development are, to some degree, in agreement that decentralized growth is considerably more expensive to provide for in terms of capital cost outlays. The situation is not clear with respect to operating costs, as centralized areas tend to have higher levels of such services as public transit, making meaningful comparison difficult (Burchell et al., 1998). Investigations of the fiscal impacts associated with growth have been carried out both utilizing forecasts, and on the basis of examining actual experience. Assessments of alternative future plans necessarily involve the forward-looking estimation approach, as in the case of evaluating New Jersey’s State Development and Redevelopment Plan, done in 1992. Two alternative futures were compared — Trend versus Planned — each with the same assumed growth in population, households, and jobs. The impacts of each were estimated in economic and environmental terms, with results as summarized in Table 15-57. Table 15-57 New Jersey Impact Assessment: Trend Versus Planned Development Estimated Impact Trend Development Planned Development Percent Savings Planned vs. Trend Roads (1990 $ millions) $2,924 $2,225 23.9% Utilities (1990 $ millions) $7,424 $6,836 7.6% Schools (1990 $ millions) $5,296 $5,123 3.3% Land Consumption (acres) 292,079 117,607 59.7% Median Housing Cost (1990 $) $172,567 $162,162 6.1% Source: Burchell, R. W., et al., “Impact Assessment of the New Jersey Interim State Development and Redevelopment Plan. Report III: Supplemental AIPLAN Assessment.” April 30, 1992, as presented in Burchell and Listokin (1996). The study found that there would be significant savings under the “Planned” approach, the more compact alternative. Over the period of 1990 to 2010, the planned alternative was estimated to require $699 million less in roads investment, a 24 percent savings. Savings were also estimated for other capital costs, as shown in the table. Summing all capital costs, the planned alternative was estimated to save $1.4 billion over 20 years, about 10 percent. It was calculated to consume 174,500 fewer acres of land, 60 percent less, including 30,300 acres of environmentally sensitive land and 42,000 acres of agricultural land. The more compact alternative was adopted as the New Jersey State Plan (Burchell and Listokin, 1996; NTI, 2000). An analysis carried out in 1989 by Duncan for the Florida Department of Community Affairs provides data on actual capital costs incurred by several completed residential and non- residential developments. The five development patterns originally examined individually were subsequently grouped in an analysis by Burchell and Listokin into two: trend, encompassing “scattered,” “linear,” and “satellite” developments; and planned, covering the original “contiguous” and “compact” categories. The capital costs associated with each type

15-121 of development were compiled. As illustrated in Table 15-58, the total public capital cost for a detached dwelling unit built in trend type development in Florida approached $16,000, versus less than $11,000 for planned development, or about 53 percent more. As in the New Jersey estimates, the biggest cost differences (actual differences in the Florida study) were in relation to roads: 155 percent higher with trend development (Burchell and Listokin, 1996; Burchell et al., 1998). Table 15-58 Florida Growth Pattern Study: Capital Facility Costs per Dwelling Unit Under Trend Versus Planned Development (1990 Dollars) Capital Cost Category Average for Trend Development Average for Planned Development Percent Difference Trend vs. Planned Roads $7,104 $2,784 +155.2% Schools 6,079 5,625 +8.1% Utilities 2,187 1,320 +65.7% Other 661 672 -1.6% Total $15,941 [sic] $10,401 +53.3% Source: Duncan, J. E., et al., “The Search for Efficient Urban Growth Patterns. Tallahassee: Department of Community Affairs (1989), as presented in Burchell and Listokin (1996). A study for the Urban Land Institute (ULI) in 1989 by James Frank reviewed four decades of literature on fiscal impacts of alternative land development forms, and concluded that multiple factors affect development costs including density, contiguity of development, and distance to central public facilities such as sewage and water plants. It was found that capital costs were highest in situations of low-density sprawl and could be dramatically reduced in situations of higher density development that is centrally and contiguously located. The estimated total public capital cost per dwelling unit in a low density/sprawl area was about $35,000 in 1987 dollars, rising to $48,000 if that development were also located 10 miles from the central water source, sewage plant, and major concentration of employment. This ULI study estimated that the public cost could be reduced to less than $18,000 by choosing a central location, using a mix of housing types (30 percent single family/70 percent apartment), and by placement allowing contiguous development as compared to leapfrogging open land (Burchell and Listokin, 1996; Burchell et al., 1998). One criticism leveled at planned/controlled growth has been that it may lead to inequity, particularly if it does not offer affordable housing to lower- and moderate-income households. In the New Jersey study above, however, median housing cost was an estimated 6.1 percent lower under the planned development approach, and a separately-calculated index of housing affordability was 6.7 percent higher (more affordable) under the planned alternative (Burchell and Listokin, 1996). In Portland, Oregon, where development is subject to containment within an urban growth boundary, housing prices indeed rose by 62 percent between 1991 and 1996, in comparison to only 19 percent in Atlanta. However, home ownership rates have been very comparable (60 to 65 percent) though the 1986 to 1996 period. The share of income spent on housing in both areas was virtually the same (19 to 20 percent), suggesting that income grew faster in Portland than in Atlanta. Portland residents consistently rate the quality of housing and neighborhoods at higher levels than Atlantans (Nelson, 2000). The inference is that because Portland’s growth boundary stabilized land supplies, development opportunities actually increased due to upzoning, with the surge in

Next: Additional Resources »
Traveler Response to Transportation System Changes Handbook, Third Edition: Chapter 15, Land Use and Site Design Get This Book
×
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

TRB’s Transit Cooperative Research Program (TCRP) Report 95: Chapter 15 – Land Use and Site Design provides information on the relationships between land use/site design and travel behavior. Information in the report is drawn primarily from research studies that have attempted to measure and explain the effects.

The Traveler Response to Transportation System Changes Handbook consists of these Chapter 1 introductory materials and 15 stand-alone published topic area chapters. Each topic area chapter provides traveler response findings including supportive information and interpretation, and also includes case studies and a bibliography consisting of the references utilized as sources.

  1. ×

    Welcome to OpenBook!

    You're looking at OpenBook, NAP.edu's online reading room since 1999. Based on feedback from you, our users, we've made some improvements that make it easier than ever to read thousands of publications on our website.

    Do you want to take a quick tour of the OpenBook's features?

    No Thanks Take a Tour »
  2. ×

    Show this book's table of contents, where you can jump to any chapter by name.

    « Back Next »
  3. ×

    ...or use these buttons to go back to the previous chapter or skip to the next one.

    « Back Next »
  4. ×

    Jump up to the previous page or down to the next one. Also, you can type in a page number and press Enter to go directly to that page in the book.

    « Back Next »
  5. ×

    To search the entire text of this book, type in your search term here and press Enter.

    « Back Next »
  6. ×

    Share a link to this book page on your preferred social network or via email.

    « Back Next »
  7. ×

    View our suggested citation for this chapter.

    « Back Next »
  8. ×

    Ready to take your reading offline? Click here to buy this book in print or download it as a free PDF, if available.

    « Back Next »
Stay Connected!