Printer Friendly

The impacts of alternative urban development patterns on highway system performance.


Urban planners have long recognized that development patterns influence travel demand and congestion levels and that transportation system supply and performance characteristics in turn influence development patterns. As traffic congestion worsens, an important question is being asked with regard to development: What is the potential contribution of development patterns to the moderation of highway travel demand and traffic congestion levels in urban areas?

The Federal Highway Administration (FHWA) sought to answer this question through studies of the impacts of alternative development patterns in U.S. cities. The studies were done under FHWA sponsorship by the Metropolitan Planning Organizations (MPOs) in four urban areas -- Baltimore, Dallas, Washington, DC, and Seattle. This article documents the results of these studies and draws some general inferences from their analyses.

A second question of interest, particularly to the environmental community, is: What are the development effects of transportation system (particularly highway) supply and performance characteristics? The FHWA is currently pursuing a multiyear study on the affects of transportation system performance on development patterns.


The FHWA studies of four U.S. cities were macro level studies that sought to investigate the potential of alternative patterns of future growth to affect highway system performance. The studies used computerized simulation to determine the travel consequences of growth in a target year in the long-range future (2010 or 2020). Each MPO developed its own set of alternative long-range urban development patterns for analysis. They compared the alternatives to development forecasts previously adopted by their policy bodies in order to draw conclusions about the relative impacts of the alternatives on travel demand and highway system performance.

The studies were done using the traditional four-step travel demand modeling process to simulate travel demand and congestion impacts. Baltimore, Dallas, and Washington, DC, varied only land use inputs into the modeling process, keeping transportation system characteristics the same as the base case. Seattle tested combinations of land use and transportation system strategies. The results of all four studies are summarized in the following sections.


The Baltimore study looked at the effects of three alternative patterns of future residential development, keeping future growth in employment concentrated in a few activity centers. The alternatives were compared to a base scenario which essentially combined forecasts of household and employment location distributions provided by Baltimore's six local jurisdictions.

The first alternative concentrated regional household growth anticipated between 1990 and 2010. A significant portion of anticipated growth was allocated to areas within the region's "development envelope." The second alternative decentralized household growth, assigning a significant portion of the growth to areas outside the development envelope. The third alternative allocated regional household growth only to those areas with a high level of transit accessibility. In all three alternatives, anticipated regional employment growth was assigned to existing activity centers, creating more intense employment clusters. The programmed future transportation network was assumed for each alternative and the base. Table 1 shows the impacts of the alternatives.

The redistribution of the 1990-2010 growth changed the proportions of total households and employment within the development envelope by less than 3 percent. Over the 20-year period, total households were projected to grow by 24 percent over the 868,000 households existing in 1990, and total employment was projected to grow by 17 percent over the 1,357,000 jobs existing in 1990. These increases were not high enough relative to the 1990 base to significantly affect existing patterns of concentration.

The analysis indicated that the regionwide transportation impacts of the three alternatives would be relatively small. Person trips do not vary much between alternatives; the variation in vehicle trips reflects the greater propensity to use transit and ridesharing modes as urban development is concentrated.

The impacts with respect to regionwide vehicle miles of travel (VMT) and congestion indicators were also relatively small, although larger than vehicle trip impacts. The highway system performed best under the centralized alternative-- severely congested VMT and severely congested lane miles were both reduced by more than 1.5 percent relative to the base. This improvement was primarily due to the reduction in total regionwide highway travel demand, an almost 1-percent reduction in regionwide VMT.

Highway system performance deteriorated relative to the base under the other two alternatives. Under the decentralized alternative, severely congested VMT and lane miles increased by 1.6 and 3.0 percent, respectively, due to the increase in total VMT of almost 2 percent. Under the transit-oriented alternative, severely congested VMT increased by over 2 percent--even more than under the decentralized alternative. However, the congestion was more localized under this alternative, occurring on fewer lane miles.

The results of the Baltimore study suggest that concentrations of residential development can benefit highway system performance and reduce new highway capacity needs. Concentrations of residential development within areas with good transit access may not reduce highway capacity needs relative to base policies, but they could reduce new highway capacity needs relative to a decentralized pattern of development. Apparently, the transit-oriented alternative does not appear to reduce new highway capacity needs because vehicular travel demand is channelled into locations with little spare highway capacity.


The three alternative development patterns studied in the Dallas urban area closely paralleled Baltimore's alternatives. Redistribution of growth in the Dallas alternatives was restricted to growth projected within the service area of the Dallas Area Rapid Transit (DART) system. As in Baltimore, the first alternative concentrated employment growth within predefined activity centers and distributed anticipated residential growth to zones within a specified distance from each activity center. A second alternative allocated employment and residential growth to currently uncongested areas, which are in dispersed locations. This alternative was designed to assess the impact of confining new development to areas with underused roadways. The third alternative was transit-oriented; it concentrated growth--both employment and residential--within a 3.5-mi radius of DART rail stations.

The alternatives were compared to forecasts from the region's transportation plan for the year 2010. Table 2 summarizes the results of the comparisons. Employment growth over a 25-year period (1986-2010) in the DART service area amounted to about 0.6 million or about 30 percent of total year 2010 employment in the service area. For the second alternative, the analysis area was expanded slightly because of a lack of sufficient uncongested areas within the DART service area. Significant redistribution of employment growth occurred with all three alternatives.

Population growth reallocated within the analysis area amounted to about 20 percent of total year 2010 population. For the first two alternatives, the change in the distribution of this population growth was less significant than for employment.

With respect to travel demand and congestion indicators, the alternative that dispersed growth in uncongested areas was the only one to reduce VMT significantly (by about 5 percent) and to produce significant improvements in level of service on the highway system. Average speed increased by 2 percent; travel time spent in delay was reduced by 4 percent; and roadway congestion was reduced by 10 percent.

The transit-oriented alternative failed to reduce VMT or improve highway levels of service. Reductions in VMT were insignificant due to relatively insignificant increases in transit ridership (about 35,000 additional linked transit trips regionwide daily). Higher congestion levels were the result of higher employment concentrations around the central business district.

The activity center-oriented alternative did not change either VMT or levels of congestion significantly. The somewhat slower average speed and increase in delay time are probably the effects of higher volumes of traffic being concentrated in the vicinity of activity centers.

Washington, DC

A somewhat different approach was taken in the selection of alternative urban development patterns for the Washington, DC, area. Two alternatives were selected. The first sought to promote a closer balance between employment and housing growth within the region. The second built upon the first and additionally sought to promote transit use by concentrating employment growth in areas of high transit accessibility.

The alternatives were compared to a base reflecting adopted 2010 forecasts, which were based on "pipeline" development proposals, zoning, available land, and other factors developed in cooperation with local jurisdictions. The results of the comparisons are presented in table 3.

Employment growth projected over a 15-year period (1995-2010) amounted to 713,000, a 27-percent increase over the 1995 base of 2,605,000. The distribution of this growth was unchanged for the first alternative. For the transit-oriented alternative, seven high employment growth areas that have superior transit accessibility were allocated twice the number of new jobs previously allocated.

The adopted 2010 forecasts for the region indicated an 18-percent increase in households. The 1995 base is 1,489,000 households. This increase of 268,000 households, assuming 1.5 workers per household, is about 200,000 fewer households than needed to balance the projected 713,000 new jobs projected. To minimize commuting into the study area from external counties, an additional 200,000 households were added to the two alternatives, resulting in an increase of 75 percent above the household growth in the base. [1]

Jobs and housing were also balanced within subareas. To balance the 604,000 new jobs projected within 29 employment growth areas, about 402,000 new households were needed, but only 143,000 were allocated to these areas in the base. The two alternatives sought to achieve a balance in those areas by drawing 200,000 households from outside the study area (as previously discussed) and the balance of 59,000 from other areas within the study area.

The additional 200,000 households represented an increase of 11 percent above the base total households of 1.75 million. In spite of this increase, regionwide VMT in the alternatives exceeded base VMT by only about 1 percent (about 1 million VMT). This was the result of reduced vehicle trips per household due to greater transit use and to shorter average trip lengths because of greater proximity of housing to jobs. The 200,000 new households, had they located outside the area, would have generated about 10 million VMT of which about 5 million would have been outside the study area. This 5 million VMT was saved outside the area with the two alternatives, at a cost of about a 1 million increase in VMT inside the study area.

The transit-oriented alternative did not appear to increase the effectiveness of the jobs/housing balance alternative with respect to VMT reduction.


Seattle's base case consisted of a composite of local growth patterns determined by each city and county in the region. In the base, new employment was scattered in office parks, shopping malls, and strip centers, with some new employment in major downtown areas. Most new housing development occurred in suburban areas. Transportation system improvements included a regional rapid transit system and modest expansion of highway capacity.

The alternative strategies consisted of changes to both the land use patterns as well as to transportation system characteristics. The first, a major centers alternative, concentrated new employment growth in a few major centers and encouraged higher density residential development within walking distance of major transit access points. Transit investments were emphasized, including high-occupancy vehicle (HOV) lanes. Highway capacity expansion was restricted to critical links, and transportation demand management (TDM) programs were supported.

The second alternative focused on multiple centers. It concentrated new employment and housing growth in a relatively large number of centers with a balance of jobs and housing within each center's area of influence. Transit emphasis was high although less than the major centers alternative. Highway capacity expansions involved 60 percent more new lane miles than in major centers, while TDM programs were similar.

The third alternative was a dispersed growth scenario. It dispersed employment and housing into newly developing areas where new highways or major highway widening could be provided or where existing highways have spare capacity. This was similar to the dispersed alternative in Dallas. Only moderate investments in transit, sufficient to maintain present levels of service, were included. Highway capacity expansions included extensive radial and circumferential highways to serve the newly developing areas. TDM measures were supported.

The comparisons of the alternatives with the base are summarized in table 4. Over the 30-year analysis period (1990-2020), population was projected to grow by 52 percent over the 1990 population of 2.7 million; jobs were projected to grow by 66 percent over the 1990 employment of 1.3 million. None of the alternatives were expected to materially affect the rate or amount of regionwide growth. The relative variation among the alternatives with respect to the distribution of growth is indicated by the changes in the growth rate for King County (the central county) as shown in table 4.

Variations in regionwide travel demand are relatively small, ranging from a 4-percent reduction in VMT under the major centers alternative to a 3-percent increase with dispersed growth. As in the Dallas study, concentration of growth in a few centers (major centers alternative) was found to increase congestion levels, especially in the vicinity of the centers, and to reduce average speeds. Concentration of growth in many centers (multiple centers alternative), on the other hand, was found to reduce overall delay and congestion somewhat. But there was no change in average regional speeds, and a more detailed review of the network indicated that congestion in critical travel corridors was significantly higher, particularly in suburban and rural areas. The dispersed alternative was less effective in reducing congestion than multiple centers, probably due to higher VMT.

Inferences From the Studies

The studies suggest that concentrating urban development may reduce vehicular travel demand and congestion, as indicated in the Baltimore study. However, when there is excessive concentration at a few high-density centers, high congestion levels may be expected in the vicinity of the high-density activity centers, as indicated in the Dallas and Seattle studies.

Concentrating development in areas with superior transit access does not appear to shift sufficient travel to transit modes to reduce congestion levels although some reductions in regionwide VMT may be achieved. However, vehicle trips may have been overestimated in high-density zones due to limitations in the study methodology. Trip production rates in regional models are generally not sensitive to zonal density characteristics, Also, the shift to transit may have been underestimated by the modeling approach used in three of the four studies. Further analysis is needed to determine whether improvements to transit service and increases in parking costs could shift sufficient numbers of peak period highway users to transit or ride-sharing modes to significantly affect peak period congestion levels and new highway capacity needs in these high-density areas.

Dispersed growth patterns may reduce congestion levels if growth is directed to areas where spare highway capacity exists. However, total travel demand (VMT) will generally rise. If growth is not properly directed, both VMT and congestion levels will rise.

Providing affordable housing within current urban boundaries and in proximity to employment growth areas can significantly reduce highway travel demand (VMT) in the broader region including surrounding counties and exurbs. The Washington, DC, study indicates that VMT per household can be reduced by as much as 10 percent. However, VMT and congestion levels could rise within the urban boundaries due to the accommodation of housing units which would otherwise be outside the urban boundary.

Macro level, land use strategies that simply relocate future growth appear to have relatively little impact on highway travel demand. Regional VMT did not change by more than 2 percent in Baltimore, 5 percent in Dallas, and 4 percent in Washington, DC, and Seattle. While regional VMT may not be greatly affected, congestion levels can be influenced to a much greater extent because new development can be either forced into existing dense areas with little spare highway capacity, or it can be spread out to developing areas where spare capacity exists. Changes in average congestion measures of as much as 10 percent were observed in the studies.

The relatively small impacts on travel demand may be explained by the fact that three of the four studies did not look at transportation infrastructure investments and TDM measures in combination with land use alternatives, and none of the studies looked at further changes that might be induced by micro level, land use strategies in combination with TDM measures and infrastructure investments. Currently, regional models generally cannot be used to analyze urban design options. An FHWA research project is currently under way to improve the ability to model the impacts of micro level strategies and combined strategies.


These four studies were an important step in improving understanding of the general magnitude and direction of the impacts of macro level, land use strategies. Limited study resources precluded extending the scope of the studies to get more refined impact estimates based on recursive modeling approaches-- for example, by including the effects of transportation system supply and performance characteristics on trip generation, trip distribution, land use location decisions, and mode choice. Research is under way at the FHWA to develop such modeling enhancements for wider application in urban areas.

Study results suggest that urban areas should add macro level, land development decisions to the toolbox for congestion management. Regional land use planning can make a significant difference. However, if urban areas are to succeed in implementing such regional strategies, they will have to enhance their intergovernmental structures and processes to facilitate the key policy decisions needed to guide urban development into patterns that are more effective in reducing congestion.


(1) Impact of Land Use Alternatives on Transportation Demand, Baltimore Regional Council of Governments, Baltimore, MD, January 22, 1992.

(2) Urban Form/Transportation System Options for the Future, North Central Texas Council of Governments, Dallas, TX, January 1992.

(3) Transportation Demand Impacts of Alternative Land Use Scenarios, Metropolitan Washington Council of Governments, Washington, DC, May 31, 1991.

(4) Summary and Comparison Between Alternatives: Vision 2020, Puget Sound Council of Governments, Seattle, WA, September 1990.

Patrick DeCorla-Souza is a community planner in the Federal Highway Administration's (FHWA's) Office of Environment and Planning in Washington, DC. He has master's degrees in civil engineering (University of Toledo) and in planning (Florida State University), and he is a charter member of the American Institute of Certified Planners. Before joining the FHWA in 1987, Mr. DeCorla-Souza worked for a decade with the Toledo (Ohio) Metropolitan Area Council of Governments; for several years prior to that, he was a consultant in Florida. He is currently involved in FHWA research, technical assistance, and training in the areas of land use/transportation interactions, congestion pricing, air quality planning, and economic analysis in transportation.

[1] Italic numbers in parentheses identify references on page 78.

[Tabular Data Omitted]

Urban Development Studies

Urban development patterns can be studied at the micro level, macro level, or in combination. Micro level studies look at the impacts of localized strategies. For example, the physical layout of new developments can be designed to create circulation patterns and environments conducive to travel by transit, bicycles, and walking. In addition, developments can be designed with mixed land uses to spread travel to them throughout the day; different types of land uses have different travel peaking characteristics. Mixed use developments can increase use of carpools, vanpools, and transit since people will not need their cars during midday if service establishments are within walking distance. Mixing employment areas and residential land also provides opportunities for those who wish to live near their workplaces, encouraging bicycling and walking trips.

Macro level studies look at the regional impacts of urban form alternatives. For example, development in a region can be either concentrated or decentralized. It can be confined to transit accessible corridors, or it can be dispersed to the periphery of urban areas where spare highway capacity is available. Employment can either be concentrated in a few major activity centers, or it can be distributed to many small centers. Jobs and housing may be balanced within subareas, or disproportionate amounts of housing may be developed on the fringes of urban areas and in the exurbs, which is the region beyond the Suburbs, to take advantage of lower housing costs.
COPYRIGHT 1992 Superintendent of Documents
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1992 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:DeCorla-Souza, Patrick
Publication:Public Roads
Date:Sep 1, 1992
Previous Article:Transfer lengths in rectangular prestressed concrete concentric specimens.
Next Article:ATMS technology - what we know and what we don't know.

Related Articles
A 50-year plan for metropolitan Portland.
Paying for new development: the urban structure program of the City of Lancaster.
Strategic Work-Zone Analysis Tools.
Transportation Studies Institute: Herff College of Engineering. (Research Centers at The University of Memphis).
Reshaping the future of cities: unless significant changes are made to the way we shape and build our cities, the 21st century looks bleak for...
An alternative to costly highway construction: CLF develops innovative approach to easing traffic woes.
Geography, forestry, conservation, and planning.
Scenario planning.
QuickZone: modeling in the zone; With the revised Federal work zone safety and Mobility Rule now in effect, this modeling tool can help DOTs comply.
Cities and accessibility: the potential for carbon reductions and the need for national leadership.

Terms of use | Copyright © 2016 Farlex, Inc. | Feedback | For webmasters