The Atlanta Roadside Emissions Exposure Study

Transcription

1 The Atlanta Roadside Emissions Exposure Study David D Onofrio, Atlanta Regional Commission, 40 Courtland Street NE, Atlanta, GA ddonofrio@atlantaregional.com Byeong-Uk Kim, Georgia Environmental Protection Division, 4244 International Parkway #120, Atlanta, GA 30354, Byeong.kim@dnr.state.ga.us Yunhee Kim, Georgia Environmental Protection Division, 4244 International Parkway #120, Atlanta, GA 30354, Yunhee.kim@dnr.state.ga.us Project Overview The transportation system is a significant source of unhealthy atmospheric pollutants in the Atlanta region. Research indicates that elevated air pollution is measurable up to several hundred meters from traffic. Factors like traffic volume, topography and weather all impact where that pollution spreads and how it affects neighborhoods in a region. The growth of vehicle travel in metropolitan Atlanta and increasing population density make it more important than ever to understand how air pollution disperses. By understanding how pollution disperses, citizens, planners and elected officials can take steps to reduce exposure and improve public health. The principle goal of the Atlanta Roadside Emissions Exposure Study (AREES) project is to understand how local-scale air quality is impacted by changes to the transportation system. By focusing on local emissions exposure at a human scale, instead of concentrating on a regional value as is the focus in transportation conformity, planners can make more informed decisions about how roadway projects will impact health in regional communities. To accomplish this goal, the Atlanta Regional Commission (ARC) established a research team consisting of staff from ARC and the Georgia Environmental Protection Division (EPD). This team worked to determine the best tools and methodologies to develop a regionally applicable dispersion modeling methodology. The development of the AREES tool took the better part of three years and involved partnership and communication with staff from Georgia Tech, US EPA, CDC, the University of New Mexico and the University of North Carolina. To determine the concentration of transportation-caused pollution in the atmosphere, the research team tied the ARC regional travel model to both an emissions model and a dispersion model. For the purpose of this study, the research team focused on fine particulate matter (PM 2.5) due to its well documented impact on human health and its relative stability in the atmosphere. The resulting output data was then crafted into an emissions exposure performance measure for use in transportation project evaluation, scenario planning and decision-making. Tying census and land-use data into the AREES outputs will allow planners to evaluate a variety of additional topics such as: health risks among vulnerable populations, environmental justice, and siting of sensitive land uses. Atlanta Roadside Emissions Exposure Study 1

2 The methodology used for AREES is replicable in other regions with the ability to model transportation and emissions. The research team selected freely available and open-source tools and methodologies to calculate near-road emission concentrations. Methodology Dispersion models are typically run to understand emissions within a few kilometer area around a single project. These projects are modeled over a small domain and include few sources. The AREES project expands this methodology to a 20-county regional geography of 16,585 km 2 with approximately 25,000 roadway segments. This distinction provides the principle challenge in developing a methodology for the AREES project that balances computation time and accuracy. Figure 1 outlines the methodology established to meet the modeling needs for the AREES project. The process can be divided into three separate steps. First, regional traffic modeled by ARC s travel demand model serves as the ultimate source of emissions by assigning traffic volumes and speeds to regional roadways. These data, along with information about the Atlanta region s vehicle fleet and composition are then fed into the US EPA-developed MOVES emissions model which, when post-processed, produces a mass rate of emissions from each roadway in the region. Finally, the geospatial emission data is fed into the R-LINE 1 dispersion model to determine the concentration of annual average PM 2.5 (in µg/m 3 ) around the region. Figure 1 AREES Modeling Flow Chart 1 Atlanta Roadside Emissions Exposure Study 2

3 Results are post-processed to match transportation source apportionments from regional PM 2.5 monitors 2. Inputs in each of the travel and emission models is consistent with the methodology ARC, and other MPOs, utilize for transportation conformity, enhancing its replicability. Two principle deviations from standard dispersion modeling are incorporated to improve model runtime. First, the standard practice for dispersion modeling is to provide at least one full year of meteorology and emission inputs to determine annual average pollutant concentrations. When the domain of the project was considered, along with a full year of meteorology, an initial estimate of R-LINE model runtime for a 20-county area of metro Atlanta on a small ARC server was over one year. To shorten this runtime, the research team altered meteorology based on a methodology developed by Shih Ying Chang at UNC, called the Stability Array (STAR) approach. The STAR approach reduces the 8,760 hours of meteorology in a year down to 100 hours by applying weights to representative meteorological hours. When this technique was applied to the 2011 Atlanta AERMET data, the resulting file contained 77 hours of meteorology some combinations of parameters were not observed in Atlanta in A small domain comparison between running R-LINE with full meteorology or STAR meteorology found an R 2 value of 0.98 between the two methodologies. The second deviation from standard dispersion modeling practice deals with the domain. Typically, dispersion models are run with a single domain of sources and receptors. Running R-LINE in this fashion would calculate the contribution of every source to every receptor in the region, some of which are over 100km apart. When in reality, the contribution of a roadway to ambient PM 2.5 concentrations spikes nearby and quickly falls off over the distance of a few kilometers. To reduce runtime and remove extraneous calculations the research team decided to reduce the domain of each individual R-LINE run. Using 12km x 12km SEMAP grids, established for photochemical modeling by EPD as a base geometry, a series of tests determined the optimal receptor domain of each R-LINE run was a square grid 48km x 48km with source links included within a domain of 72 km x 72 km. This step reduces the number of calculations and greatly reduces runtime. This methodology also decouples the need to run the entire region with every test. New transportation projects can be evaluated by running the set of grids that are relevant to the project. Applying the STAR meteorology and limited domain reduced the total runtime from approximately 1 year to 231 hours (9.6 days) on one 8-core desktop PC. Results The output of the R-LINE runs is a set of 486,000 receptors with the annual average PM 2.5 concentrations. Figure 2 illustrates this output with the receptor as the centroid of a 200m x 200m grid i.e. with no interpolation. Looking at the map, it becomes very obvious where Interstates crisscross the region, as well as some of the limited-access state roadways and major arterials. 2 Task led by Georgia Tech professor Ted Russell and research students Poornima Sampath and Xixin Zhai Atlanta Roadside Emissions Exposure Study 3

4 Figure 2 Annual Average PM 2.5 Concentrations from Automobile Sources In addition to regional runs, the project team evaluated emissions from a small domain transportation project. This scenario was designed to test the sensitivity of the AREES methodology to evaluate annual average PM 2.5 concentration changes associated with roadway capacity increases. The project was not selected to be realistic, i.e. in an existing plan, and was meant to test the sensitivity of the model to planning decisions. In this scenario the research team widened I-75 from I-285 to the Downtown Connector interchange in northwest Atlanta. One additional lane was added to the transportation model in each direction of travel. This project functions as an upper limit to the type of capacity expansion found in the ARC s Regional Transportation Plan (The Atlanta Region s Plan 3 ). Figure 3 maps the changes in concentrations around the city of Atlanta due to the modeled widening of I-75. In Figure 3, within a 500m buffer of the project the average change in annual PM 2.5 concentration is 0.05 µg/m 3, which corresponds to a 1.1% increase over the base scenario. The maximum increase measured is 0.30 µg/m 3, or a 4% increase over the base scenario. The spatial distributions of emissions in the city varies due to the travel model reassigning trips due to the increased capacity on I-75. The redistribution of trips results in a reduction of modeled emissions in some areas and an increase in others not adjacent to the project limits. 3 Atlanta Roadside Emissions Exposure Study 4

5 Figure 3 Change in Annual Average PM 2.5 Concentrations from Widening I-75 Conclusion The AREES project provides local and regional governments with a reliable methodology to quickly assess the change in ambient transportation emissions due to infrastructure investment decisions. The methodology is adaptable and can be tailored to many regions of the country. ARC staff have begun to use the results of the AREES project to inform the transportation planning process. AREES results were incorporated into the Atlanta Region s Plan transportation project evaluation as a performance measure to aid in decision-making. The research team created an interactive website to share results with stakeholders and the community. 4 Moving forward, the research team plans on continuing to refine the source emissions by updating AREES results with the latest travel demand model and emission model output. The process to engage the public on the results and their meaning will begin in earnest in ARC staff are continually working on additional ways to incorporate AREES results into the transportation planning and sensitiveland use siting processes. 4 Atlanta Roadside Emissions Exposure Study 5