Researchers at the University of Iowa have developed a methodology to assess the environmental justice impacts of air quality and noise from transportation projects. The methodology uses a combination of census data analysis, vehicle emission models, dispersion models, and noise models in a GIS environment to determine the impact of transportation projects on low-income and minority populations. The authors illustrate these techniques using an example in Waterloo, Iowa.
Figure 1. Waterloo, Iowa
The goal of authors David Forkenbrock and Lisa Schweitzer was to develop an approach to estimating a variety of impacts relevant to environmental justice. After consulting with the Iowa and Minnesota Departments of Transportation, they chose to test their approach in Waterloo, Iowa. Waterloo is a city in northeastern Iowa with a 1990 population of about 66,000. It is relatively diverse (13.2 percent minority population) and has a relatively low-median income compared to typical metropolitan areas in Iowa and Minnesota.
Running north-south through the center of Waterloo is U.S. Highway 63 (Figure 1). North of the city center, the route runs through racially mixed and low-income neighborhoods. The authors modeled air pollution and noise at a typical major intersection in this corridor, using existing conditions as the "base case" of a hypothetical alternatives analysis. While no changes are currently under consideration to the corridor, air pollution and noise impacts could be estimated for alternative scenarios based on projected changes in traffic volumes, speeds, and composition.
The first step is to identify the roadways or other transportation facilities that will experience significant changes in traffic volumes or speeds, and therefore in emissions or noise, as a result of the project. These may include facilities that are part of the project as well as other facilities that will experience changes in traffic. The project study area should encompass these facilities.
For a regional transportation program or set of projects, the best approach may be to screen individual projects or areas to identify the cases of potential highest noise and emission impacts (e.g., congested intersections or high-volume roadways) and to select these for more detailed analysis.
The next step is to estimate the concentration of low-income and minority populations at the census block level. This allows the impacts of emissions and noise to be examined at a high level of spatial detail. Data on race can be obtained at the block level directly from the census. The number of low-income households can be estimated using the methodology described in Box 1.
The data can then be mapped to show either total affected populations or concentrations of minority and low-income populations by block level, in the area that will be affected by the proposed transportation projects. Figure 2 shows the minority population by census block in the study area.
Figure 2. Percent Minority Population
Source: Forkenbrock and Schweitzer (1997).
|Box 1. Estimating Low-Income Population at the Block Level
The census only reports the number of low income households at the block group level or higher. This figure, however, can be estimated at the block level based on other known block-level characteristics. For the Waterloo, Iowan MSA, Forkenbrock and Scheizer (1997) construct a regression equation that predicts population in poverty at the census block level, as follows (where A, B, C, and D are coefficients):
This equation is estimated from block group data for the Waterloo MSA and predicts 65 percent of the variance in poverty levels. The authors stress that the specific equation is only valid for the Waterloo area. For other areas, different variables may be appropriate, and coefficients should be estimated based on local census data.
An emissions dispersion model such as CAL3QHC requires the following information:
Noise models require similar data on traffic volumes, speeds, and mix of vehicles by link. Optional data to refine noise estimates include the height of adjacent structures, elevation of surrounding terrain, and traffic mix, speed, and volume of intersecting streets.
Forecast traffic data should be obtained with and without the proposed project. The regional MPO may have traffic volume and speed data for major facilities. Data may also be available from traffic modeling for the proposed project.
The next step is to estimate the emissions and the spatial extent of these emissions from existing or planned roadway facilities. This involves three primary substeps:
CAL3QHC provides outputs in the form of pollutant concentrations at various user-specified X-Y coordinates. These concentrations can be plotted in GIS using a grid-based modeling package such as is included in ArcView Spatial Analyst or MapInfo Professional. The GIS software contains built-in interpolation algorithms that can be used to plot contours of emission concentration, based on the X-Y concentration data. (The X-Y coordinates used to identify link locations and receptor endpoints in CAL3QHC should be consistent with the coordinates used to map the road network in the GIS software.)
Noise contours can be estimated, similar to the air pollution contours developed above. A noise model such as FHWA's Traffic Noise Model (known as STAMINA in its previous version), can be used to determine noise levels as a function of traffic volume, mix, and speeds. In the Waterloo case study, the authors use MINNOISE, an updated version of STAMINA developed by the Minnesota Department of Transportation. While additional data such as height of adjacent barriers or local topography can be used to refine noise estimates, a simplified set of inputs based only on traffic were utilized in this analysis.
Noise models typically estimate noise levels as a function of distance from the roadway centerline. Contours can be plotted in GIS by creating a new shapefile and drawing buffers around a roadway link. Noise levels are then associated with each buffer based on the distance of the buffer from the link. To model roadways with intersections, the road may need to be broken into multiple segments, for which vehicles are designated as accelerating, decelerating, or cruising.
Figure 3. Noise Contour Overlaid on Minority Population
Source: Forkenbrock and Schweitzer (1997).
The results of these steps can be combined in a number of ways, using GIS tools, to assess the social impacts of the transportation project:
Maps of air pollution or noise contours can be overlaid graphically with maps showing total populations and populations of interest by census block (Figure 4).
Figure 4. Noise Contour Overlaid on Minority Population (detail)
GIS analysis capabilities can be used to estimate the number of total, minority, or low-income persons who live within unacceptable noise or air pollution contours ("unacceptable" levels may be based on federal, state, or local guidelines). This can be done by first identifying the census blocks falling within a given emissions or noise contour, and then identifying the number of people in the population groups of interest living within these blocks. As a further refinement, for blocks that fall only partially within the contour, the proportion of the block falling within the contour can be calculated using a GIS overlay. This proportion can be multiplied by the total population of the block to determine the number of people adversely affected.
Measures of affected population groups can be compared for conditions with and without a proposed transportation program, to evaluate the increase or decrease in affected persons as a result of the project. For example, to evaluate whether the program has a disproportionate impact on minority and low-income populations, the percent of total area-wide population adversely affected by the project could be compared by the percent of area-wide minority and low-income population that is adversely affected.
The U.S. 63 noise analysis showed a maximum of 64 dBA near the centerline of the highway, attenuating to 55 dBA by 1,000 feet away from the highway. For comparison, standards established by the Minnesota Pollution Control Agency suggest that in residential areas, noise levels should not exceed 65 dBA more than 10 percent of the time during the daytime, and should not exceed 55 dBA more than 10 percent of the time at night. Additional noise standards are shown in Table 1.
Table 1. Maximum Allowable Noise Level Outside Buildings, Minnesota
Source: Minnesota Department of Transportation as reported in Forkenbrock and Schweitzer (1997).
Note: L50 is the decibel level which should not be exceeded more than 50 percent of the time. L10 is the decibel level that should not be exceeded more than 10 percent of the time.
By using the GIS to overlay noise contours on population data, the authors estimated that the daytime Minnesota standard was exceeded for 189 low-income people (or 16 percent of the population in the study area) and 499 minority people (42 percent of the population).
The authors also tested a hypothetical scenario in which heavy truck traffic was added to U.S. 63 to approximate the effect of a significant amount of industrial development along the route. The model shows that even if there were a significant increase in truck traffic during the day, noise levels would remain below federal standards.
The air pollution analysis showed a maximum CO concentration of 31 ppm, compared to the Federal one-hour standard of 35 ppm. The analysis also showed a PM level of 12.5 ug/m3, well below the maximum allowable annual mean of 50 ug/m3.
This analysis demonstrates a number of useful techniques for assessing the impacts of transportation projects on different population groups. Specifically, it illustrates:
Disaggregation of population, network, and emissions data into grid cells is an alternative way of estimating exposure to noise and air pollution by socioeconomic group. This approach is illustrated in the SPARTACUS Case Study.
Some limitations also must be noted with respect to the technical methods utilized and any conclusions that can be drawn from this type of analysis:
The methods described here can be used to identify environmental impacts on local populations in a relatively precise manner. The authors do not, however, outline a complete methodology for comparing the distribution of burdens or draw specific conclusions with respect to environmental justice. Furthermore, local populations may experience other benefits or costs resulting from the project, such as increased accessibility or economic development, that need to be weighed against the environmental impacts. Identifying the environmental justice implications of any project or action raises difficult questions. Analysts adopting the methodology described here will need to develop a method for assessing the fairness of the various project impacts across population groups, based on local goals and objectives.
While the Waterloo case study focuses on the area surrounding a major highway intersection, this methodology could be extended to analyze the regional impacts of transportation projects and policies, as well as land use policies. Some of the policies that might be tested include:
Regional highway construction or expansion programs;
Pricing, land use, or other strategies that significantly affect regional VMT;
Truck traffic policies, such as exclusive truck lanes or routing restrictions, that change the composition of traffic on various roadways; and
Land use policies that affect the siting of residences in relation to transportation facilities.
Since the noise and air quality analyses must be carried out on a facility-level basis, doing this for every major facility in the region is not practical. An alternative approach would be to identify areas of potential greatest impact for more detailed analysis. Such an approach might be done as follows:
Traffic volume data for both future baseline and proposed conditions (e.g., a "no-build alternative" and a "build alternative 1") would be used to identify links of greatest potential impact as well as greatest changes in impacts. These data could be obtained from regional travel demand model output for each future transportation and/or land use scenario.
Maps of population data illustrating expected population density (for total, low-income, and/or minority population) at the census tract or TAZ level would be developed.
These data would be overlaid to identify areas of greatest potential population exposure to noise and emissions (baseline as well as increases). A few "high-impact" areas would be selected for more detailed analysis.
Exposure of total, low-income, and minority populations to unacceptable levels of emissions and noise would be estimated under the "baseline" and "alternatives" conditions, using the methodology described previously.
Forkenbrock, David J. and Lisa A. Schweitzer (1997). Environmental Justice and Transportation Investment Policy. Public Policy Center, University of Iowa. (Internet: http://www.uiowa.edu/~ppc/)
|Public Policy Center, University of Iowa||David Forkenbrock||(391) 335-6800|
|Lisa Schweitzer||(391) 335-6800|