Energy-Efficient Urban Form

Reducing urban sprawl could play an important role in addressing climate change.

Improving city layouts and transportation networks could reduce carbon emissions more than replacing all gasoline with corn ethanol (1). Although much attention on mitigating climate change has focused on alternative fuels, vehicles, and electricity generation, better urban design represents an important yet undervalued opportunity. Fortunately, such decisions are well within the reach of local governments and leaders and can reduce long-term carbon emissions.

The impact of cities—and urban design—on the global climate is becoming increasingly important. In 2008, urbanites will outnumber rural dwellers globally for the first time in human history (2). China’s population doubled between 1952 and 2003, but its urban population increased 7-fold; today, 170 Chinese cities have at least 1 million residents (3). The U.S. has 39 such cities (4). In coming decades, urban populations are expected to double while rural populations level off or decline.

Vehicle use is rising rapidly. From 1970 to 2005, U.S. total vehicle-kilometers increased 3× faster than the population (annual increases: 3.0% vs 1.0%) (5). Similar trends occurred in China (8.3% vs 1.7%, a 5-fold difference) and the world (4.3% vs 1.8%) during 1970–1990 (6). If current trends in total vehicle-kilometers continue, vehicle CO₂ emissions may increase even if emissions per mile decline (7).

In an influential paper in Science, Socolow and Pacala (8) argue that climate stabilization during the next half century means reducing CO₂ emissions by 175 GtC (33%) relative to a business-as-usual (BAU) scenario. They propose seven strategies, with each “stabilization wedge” representing emission reductions of 25 GtC during 2005–2054 (each wedge grows from no reduction in 2005 to 1 GtC per year [yr] reduction in 2054).

The race is now on to figure out ways to design and implement these wedges. Often neglected in the debate is the role of urban form (e.g., land-use patterns and the layout of transportation infrastructure) in meeting climate objectives. My estimates suggest that reducing urban sprawl in the U.S. alone could represent half or more of a stabilization wedge.

Impacts of urban form on transportation CO₂

Compact urban form can cut on-road gasoline emissions, the largest segment (62%) of transportation CO₂ in the U.S. The transportation sector is the largest emitter (33%) of CO₂, outpacing the residential, industrial, and commercial sectors. (Electricity generation, when totaled for all sectors, accounts for 41% of CO₂ emissions.) Records of automobile usage (Figure 1) show an inverse relationship between population density and per capita daily vehicle-kilometers traveled (VKT) (4, 9). Evidence suggests that VKT is causally related to population density and other urban form attributes, and therefore, sprawl reduction policies may curtail VKT (10–14). In denser urban areas, trip origins and destinations (e.g., home, work, shopping) are closer; driving disincentives (e.g., congestion, parking costs) are greater; and alternative modes of travel (e.g., walking, bicycling, mass transit) are more common (15).

Sprawl encompasses many aspects of land use, including leapfrog development, segregated land use, and automobile dependence (16, 17). Calculations presented next focus on one aspect of sprawl—declining urban average population density.

To estimate the potential carbon benefits of reducing urban sprawl, I considered five urban growth scenarios for the U.S., 2005–2054: high sprawl, BAU, reduced sprawl, no sprawl, and infill. Changes per decade in average urban population density are –47, –39, –13, 0, and +11%, respectively, for the five scenarios. In comparison, changes per decade in average urban population density were –13% during 1960–1990 and –34% during 1990–2000 (18). For the no-sprawl scenario, average density is constant, so urban area increases at the population growth rate (1% annually). The infill scenario approximates a strict urban growth boundary, with total urban area roughly constant (unrealistic, but a useful bounding estimate).

For each urban growth scenario, I considered no technology innovation and innovation (1% annual reduction in fleet-average gasoline CO₂ emissions per kilometer). As a sensitivity analysis, I evaluated the impact of rapid innovation (3% annually). Innovation could include more-efficient vehicles or reduced-carbon fuels. Historically, annual improvements >1% have been achieved, but not maintained, for 50 yr. Annual changes in U.S. passenger vehicle fleet-average fuel economy were –0.4% during 1936–1973, +2.0% during 1973–1995 (23 yr), and near-zero (+0.03%) during 1995–2005 (19–21). Year 2005 passenger-vehicle average fuel economy was 20 miles per gallon (mpg; 21).

The Energy Independence and Security Act, signed by President Bush in December 2007, will increase fuel economy by ~1.5% annually for the next 23 years (assumptions: 10-yr vehicle turnover rate; continuation of the ~20% gap between actual passenger-vehicle fuel consumption [20 mpg; 21] and Corporate Average Fuel Economy [CAFE]-rated consumption [25 mpg; 22, 23]). The act will increase CAFE standards (currently 27.5 mpg for automobiles and 22.2 mpg for light-duty trucks) to 35 mpg for both vehicle classes in 2020; this is below today’s CAFE-type standards in China (37 mpg) and the EU (44 mpg) (23).

For each urban growth and technology scenario, I calculated U.S. total VKT and the resulting on-road gasoline CO₂ emissions. See the Supporting Information (SI) for details.

The results indicate that sprawl reduction could play an important role in addressing climate change. Reduced sprawl, without technology innovation, decreases emissions by 10 GtC during 2005–2054 (by 0.5 GtC/yr in 2055) compared with BAU. These savings represent 41% of a wedge. The no-sprawl and infill scenarios offer 53% and 60% of a wedge, respectively, compared with BAU, with no innovation.

Improvements in emissions per kilometer would reduce emissions for all urban growth scenarios and also would reduce the differences between them. For BAU growth, innovation reduces emissions by 38% of a wedge compared with no innovation—almost the same reduction as converting, without innovation, from BAU to reduced sprawl. Converting from BAU without innovation to reduced- or no-sprawl growth with innovation would offer 66% or 74% of a wedge, respectively. These improvements are impressive, especially given that they involve action by the U.S. only (5% of the global population). Disruptive technologies such as low-carbon fuels or rechargeable hybrid-electric vehicles that use low-carbon electricity could deliver environmental innovation rates significantly higher than historic levels. Increasing innovation from 0% to 3% annually would reduce U.S. emissions by 83% of a wedge under BAU growth.

Analyses above use the extant (Figure 1) relationship between population density and VKT (specifically, a density-VKT elasticity of –30%; 9, 24). Greater attention and commitment to energy-efficient urban form (see below for example policies) could increase the carbon benefits from sprawl reduction. As an illustration, I repeated the analyses above, but assumed aggressive support for energy-efficient urban form (represented by doubling the density-VKT elasticity to –60%). I found that, without innovation, reduced- and no-sprawl scenarios offer 125% and 149% of a wedge, respectively, compared with BAU. Sprawl reduction would have a greater CO₂ impact if aimed at high-elasticity situations (9, 12) rather than low-elasticity ones.

Overall, these findings suggest that long-term climate impacts from shifts in urban form could be comparable to those from technological innovation, and that climate-mitigation strategies would have greater impact by addressing urban form and technological innovation, rather than only one of those two.

Reducing sprawl could help minimize U.S. dependence on foreign oil and could ease economic impacts of price shocks while providing financial savings to help offset the costs of low-carbon energy technologies. Conversely, if sprawl accelerates, it becomes another source of emissions to reduce or offset. In that case, additional mitigation wedges will be required. For example, compared with BAU, the high-sprawl scenario increases emissions by 14% of a wedge with innovation or by 21% without.

The values presented here are rough estimates only, involving several simplifications (see SI; even a highly detailed analysis would involve significant uncertainty). Still, the results highlight the potential importance of urban design for reducing CO₂ emissions and suggest that further investigation is warranted. One source of uncertainty is the influence of urban form on vehicle choice. The perceived convenience of a small vehicle is greater at high than at low density. The evidence suggests that the relationship is causal: lower densities and longer commutes increase preferences for light-duty trucks instead of automobiles (11). Including those interactions would increase the estimated benefit of sprawl reduction on CO₂ levels. Perhaps the largest source of uncertainty is the density-VKT elasticity. If elasticity were between –4% and –15% (9), the impact of reduced sprawl compared with BAU would shrink from 41% of a wedge to 4–17%. (If elasticity were zero, the CO₂ impact of urban form shifts would be zero.) Using the elasticity of –30% to –50% reported by modeling (12) and empirical (24) studies would increase the estimated CO₂ benefits of sprawl reduction.

Urban form is important outside of the U.S., too. Globally, the number of megacities (population >10 million) increased from 3 in 1975 to 20 in 2005 (25). However, most urbanites live in cities with <1 million people (2). Cities’ ecological footprint—accounting for inputs and outputs such as food, consumer goods, energy consumption, and waste products—is many times larger than that of urbanized land areas; the difference is a factor of 60 in the U.S. (95 ecological footprint vs 1.5 urban area; units: 1000 m²/person) and a factor of 80 in India (~8 vs ~0.1) (3, 18, 26). Most (~90%) global population growth during 2000–2025 is expected to occur in urban areas in less-developed countries (2). The size and shape of these cities will have long-term impacts on transportation CO₂ emissions. Although quantification is difficult because of a lack of relevant global databases, if the above results were to hold globally, it is feasible that the difference in global average VKT between the reduced- and non-reduced-sprawl scenarios would be a factor of 2 (roughly one stabilization wedge; 8).

The comments above suggest that urban form could offer a useful approach to climate mitigation, but important historic, economic, and social issues have not been explored here. The political challenges are daunting. Sprawl reduction may or may not be feasible and desirable in all cities.

Many disciplines, including planning and policy, sociology, economics, and engineering, explore the causes and consequences of transportation demand. Hundreds of empirical studies connect urban form and daily travel (10). Estimates presented here are broadly consistent with limited prior research regarding CO₂ emissions and urban form. Estimates have suggested that in 45–50 yr, compact development would reduce annual gasoline CO₂ emissions by between 5% (12) and 9–16% (10). Analogous estimates for comprehensive smart-growth policies include 3–14% in 15–20 yr and 10–25% in 45 yr (11). A life-cycle analysis found that CO₂-equivalent emissions are 60% less for high-density than for low-density development (27). Annual emission reductions estimated here vary by scenarios and by elasticity; for low sprawl versus BAU, with 30% elasticity, reductions are 15–20% for 15–20 yr and 39% for 45 yr. As discussed next, achieving those deep emission cuts via energy-efficient urban form would involve policies (e.g., mixed-use development) that supplement and support the increased density (10, 11).

Though not quantified here, urban form can influence nontransportation CO₂ emissions, for example, via district-scale building thermal management (heating and cooling); combined heat and power generation; availability of local food and products; and embodied CO₂ in roads, buildings, and other infrastructure.

Strategies for improving urban energy efficiency

Patterns of urban growth (e.g., sprawl vs infill) result from thousands of independent decisions about where to live, work, and shop on the basis of attributes like crime rates, housing affordability, and school quality. Factors supporting automobile-dependent sprawl in the U.S. include comparatively inexpensive gasoline, extensive highways (built mostly with tax dollars), building-height limits, and parking requirements (e.g., requiring new stores to build parking spaces before opening; 28).

Many important zoning and other land-use decisions occur at the local level, which offers both a challenge (e.g., lack of local political will regarding global issues) and an opportunity (e.g., for cities wishing to reduce emissions yet frustrated by lack of direct control over vehicle fuel economy and other factors). Aligning sprawl reduction with other goals (e.g., public health and economic development) will increase public acceptance.

If done well, reducing sprawl can improve quality of life while reducing emissions. Successful approaches likely differ among cities, especially between developing versus developed countries. In some cases, improving urban schools or reducing crime rates would decrease migration to suburbs and exurbs. Other cities may need to increase the supply of affordable, attractive medium- and high-density housing. Pedestrian- and bicycle-friendly neighborhoods, convenient mass transit, and land-use mixing (e.g., allowing retail near residences) can allow people to drive less each day if they wish (potentially increasing the density-VKT elasticity magnitude).

As one example, the “Living First” campaign in Vancouver required high-density neighborhoods to be aesthetically pleasant and full of amenities (e.g., easy access to parks, child-care facilities, and grocery stores; streetscapes with shops and row housing rather than blank high-rise walls; and safe, convenient mass-transit and pedestrian facilities; 29). The resulting neighborhoods benefit not only CO₂ emissions (two-thirds of trips are by mass transit, bicycle, or walking) but also public health: by reducing automobile usage, compact development also reduces traffic fatalities and obesity (29).

Similarly, Singapore’s convenient mass transit, mixed-use development (shops near residences), and disincentives for driving (e.g., vehicle taxes, congestion pricing) result in a commuting mode split of 53% of the population using mass transit (subway or bus), 31% using a private vehicle or taxi, and 15% not commuting or using “other” modes (e.g., bicycle) (30). Mass-transit use is 47% and 56% for professionals and associate professionals, respectively (30). Globally, mode share for transit/walking/biking varies from >55% (Berlin, Madrid, Vienna) to <6% (Houston, Atlanta) (31).

London’s congestion toll reduced vehicle CO₂ emissions by 20% in the charging zone; emissions on the surrounding ring road did not increase because lower per-kilometer emissions from less congestion offset minor increases in vehicle usage of ring roads (32). Lower VKT and reduced congestion contributed equally to the inside-zone emission reduction (32). These examples further emphasize that urban design and transportation policy can influence transportation CO₂ emissions.

Calls for transportation- or energy-efficient urban form are not new—the 1973 oil crisis motivated multiple studies (33)—but recent concern about climate change has reinvigorated discussion and provided political motivation for near-term actions yielding short-, medium-, and long-term benefits. For example, Washington State’s greenhouse gas legislation (Bill 2815), enacted in March 2008, seeks to reduce annual per capita vehicle-miles traveled by 18% by 2020, 30% by 2035, and 50% by 2050.

Of interest to policy makers, sprawl reduction might require less government intervention, not more. A common myth in the U.S. is that urban development patterns (e.g., sprawl) are a “natural” reflection of free markets. Land and transportation are not free markets; subsidies, fees, regulations, and externalities abound. The largest barrier to infill development cited by U.S. developers is regulation (78%), not insufficient market interest (34).

Climate change mitigation will require a combination of approaches. Fuel efficiency deserves attention, and current research could generate a viable low-carbon fuel. But not all climate solutions require new technologies. Energy-efficient urban design offers an important tool that local governments can use to begin tackling this global problem.

Julian D. Marshall is an assistant professor of environmental engineering in the department of civil engineering at the University of Minnesota. Address correspondence about this article to Marshall at julian@umn.edu.

Acknowledgments

Thoughtful comments by Michael Brauer, Hadi Dowlatabadi, Kevin Krizek, David Levinson, Eric Mazzi, Conor Reynolds, and three anonymous reviewers are gratefully acknowledged.

Supporting Information

An equation, input data, results, and discussion are presented free of charge in SI at http://pubs.acs.org.

References

(1) Farrell, A. E.; et al. Ethanol Can Contribute to Energy and Environmental Goals. Science 2006, 311, 506–508.

(2) UN. World Urbanization Prospects: The 2005 Revision; 2006; www.un.org/esa/population/publications/WUP2005/2005wup.htm.

(3) Liu, J.; Diamond, J. China’s Environment in a Globalizing World. Nature 2005, 435, 1179–1186.

(4) U.S. Department of Transportation. Our Nation’s Highways—2000, Selected Facts and Figures; 2003; www.fhwa.dot.gov/ohim/onh00.

(5) U.S. EPA. Air Emissions Summary through 2005; 2006; www.epa.gov/airtrends/2006/emissions_summary_2005.html.

(6) Schafer, A. The Global Demand for Motorized Mobility. Transport. Res. A 1998, 32, 455–477.

(7) Winkelman, S. Comment on Stone: Could the Worst of Times for the Planet Be the Best of Times for Planning? J. Am. Plan. Assoc. 2007, 73, 418—420.

(8) Pacala, S.; Socolow, R. Stabilization Wedges: Solving the Climate Problem. Science 2004, 305, 968–972.

(9) Marshall, J. D.; et al. Inhalation of Motor Vehicle Emissions: Effects of Urban Population and Land Area. Atmos. Environ. 2005, 39, 283–295.

(10) Ewing, R.; et al. Growing Cooler: The Evidence on Urban Development and Climate Change; Urban Land Institute: Chicago, 2007.

(11) Dierkers, G.; et al. CCAP Transportation Emissions Guidebook, Part One: Land Use, Transit and Travel Demand Management; Center for Clean Air Policy: Washington, DC, 2005.

(12) Stone, B.; et al. Is Compact Growth Good for Air Quality? J. Am. Plan. Assoc. 2007, 73, 404–418.

(13) Handy, S.; Cao, X.; Mokhtarian, P. Correlation or Causality between the Built Environment and Travel Behavior? Evidence from Northern California. Transport. Res. D 2005, 10, 427–444.

(14) Giuliano, G.; Narayan, D. Another Look at Travel Patterns and Urban Form: The U.S. and Great Britain. Urban Stud. 2003, 40, 2295–2312.

(15) Boarnet, M. G.; Crane, R. Travel by Design: The Influence of Urban Form on Travel; Oxford University Press: New York, 2001.

(16) Bruegmann, R. Sprawl: A Compact History; University of Chicago Press: Chicago, 2005.

(17) Hayden, D. Building Suburbia: Green Fields and Urban Growth, 1820–2000; Random House: New York, 2003.

(18) Marshall, J. D. Urban Land Area and Population Growth: A New Scaling Relationship for Metropolitan Expansion. Urban Stud. 2007, 44, 1889–1904.

(19) U.S. Department of Transportation, Highway Statistics Summary to 1995, Vehicle Miles of Travel; 1997, Table
VM-201a; www.fhwa.dot.gov/policy/ohpi/qftravel.htm.

(20) U.S. Department of Transportation, Highway Statistics 1995; 1996, Section V, Table VM-1; www.fhwa.dot.gov/policy/ohpi/hss/hsspubs.htm.

(21) U.S. Department of Transportation, Highway Statistics 2006; 2007, Section V, Table VM-1; www.fhwa.dot.gov/policy/ohpi/hss/hsspubs.htm.

(22) U.S. Department of Transportation. Annual Energy Outlook; 2007, Supplemental Tables, Table 35: Transportation Sector Energy Use by Mode and Type; www.eia.doe.gov/oiaf/aeo/supplement/supref.html.

(23) An, F.; Sauer, A. Comparison of Passenger Vehicle Fuel Economy and Greenhouse Gas Emission Standards around the World; Pew Center on Global Climate Change: Arlington, VA, 2004; www.pewclimate.org/docUploads/
Fuel%20Economy%20and%20GHG%20Standards_010605_110719.pdf.

(24) Holtzclaw, J.; et al. Location Efficiency: Neighborhood and Socioeconomic Characteristics Determine Auto Ownership and Use—Studies in Chicago, Los Angeles, and San Francisco. Transp. Plan. Tech. 2002, 25, 1–27.

(25) UN. Urban Agglomerations 2005; 2006; www.un.org/esa/population/publications/WUP2005/2005urban_agglo.htm.

(26) UN. Demographic Yearbook; 2004, Table 8; unstats.un.org/unsd/demographic/products/dyb/dyb2004.htm.

(27) Norman, J.; MacLean, H. L.; Kennedy, C. A. Comparing High and Low Residential Density: Life-Cycle Analysis of Energy Use and Greenhouse Gas Emissions. J. Urban Plan. Dev.—ASCE 2006, 132 (1), 10–21.

(28) Manville, M.; Shoup, D. Parking, People, and Cities. J. Urban Plan. Dev.—ASCE 2005, 131, 233–245.

(29) Montgomery, C. Futureville. Can. Geog. 2006, May/June, 44–60.

(30) Singapore Department of Statistics, General Household Survey 2005; 2006, Table 11; www.singstat.gov.sg/pubn/ghs.html.

(31) Sorensen, A.; Hess, P. Metropolitan Form, Density, Transportation; Neptis Foundation: Toronto, 2007.

(32) Beevers, S. D.; Carslaw, D. C. The Impact of Congestion Charging on Vehicle Emissions in London. Atmos. Environ. 2005, 39, 1–5.

(33) Keyes, D. L. Energy and Land Use: An Instrument of U.S. Conservation Policy? Energy Policy 1976, 4, 225–236.

(34) Levine, J. Zoned Out: Regulation, Markets, and Choices in Transportation and Metropolitan Land Use; RFF Press: Washington, DC, 2006.