2007 Progress Report: Predicting the Relative Impacts of Urban Development Policies and On-Road Vehicle Technologies on Air Quality in the United States: Modeling and Analysis of a Case Study in Austin, Texas

EPA Grant Number: R831839
Title: Predicting the Relative Impacts of Urban Development Policies and On-Road Vehicle Technologies on Air Quality in the United States: Modeling and Analysis of a Case Study in Austin, Texas
Investigators: McDonald-Buller, Elena , Allen, David T. , Kockelman, Kara , Parmenter, Barbara
Institution: The University of Texas at Austin
EPA Project Officer: Chung, Serena
Project Period: December 20, 2004 through December 19, 2007 (Extended to December 19, 2008)
Project Period Covered by this Report: December 20, 2006 through December 19, 2007
Project Amount: $650,000
RFA: Regional Development, Population Trend, and Technology Change Impacts on Future Air Pollution Emissions (2004) RFA Text |  Recipients Lists
Research Category: Global Climate Change , Climate Change , Air

Objective:

Robust forecasts of land use, emissions, and activity patterns are essential for air quality model predictions. The objectives of this project are to:

  1. Apply an integrated transportation-land use model (ITLUM) to investigate the impacts of regional development scenarios on the magnitude and spatial distribution of emissions of ozone precursors using the Austin, Texas Metropolitan Statistical Area as a case study. ITLUM-based forecasts will be compared with four pre-determined metropolitan development scenarios developed through a regional “visioning” initiative known as Envision Central Texas (ECT): i) ECT A: low-density, segregated-use development based on extensive highway provision; ii) ECT B: concentrated, contiguous regional growth within 1-mile of transportation corridors; iii) ECT C: concentrated growth in existing and new communities with distinct boundaries; iv) ECT D: high-density development and balanced-use zoning.
  2. Compare the air quality impacts of regional development scenarios on predicted ozone concentrations and human exposure patterns using a photochemical grid model.
  3. Test the hypothesis that predicted human exposure patterns based on ITLUM emission forecasts will differ from those based on the U.S. EPA’s post-Clean Air Act Amendment emission scenario projections.
  4. Test the hypothesis that changes in land use and dry deposition patterns have at least as significant an impact on future air quality as changes in on-road vehicle emission control technologies.

The objectives of the project have not changed from the original application.

Progress Summary:

Urban development results in changes to land use and land cover and, consequently, to biogenic and anthropogenic emissions, meteorological processes, and processes such as dry deposition. The hypotheses described above are being investigated using a case study in the five-county Austin, Texas Metropolitan Statistical Area (MSA). Austin was among the first areas to prepare an Early Action Compact or voluntary State Implementation Plan (SIP) under the National Ambient Air Quality Standard (NAAQS) for ozone concentrations averaged over 8-hours. The Austin Metropolitan Statistical Area (MSA), which includes Travis, Williamson, Hays, Bastrop and Caldwell Counties, is located in Central Texas with a population of approximately 1.25 million people and an economic sector based on computer hardware and semiconductor manufacturing, software development, education, and state government. The area is growing rapidly; Williamson (5th), Hays (26th), Bastrop (30th), and Caldwell (51st) counties were among the 100 fastest growing counties by percent change in the country, while Travis (32nd) County was one of 100 fastest growing counties by numeric change in the country between 2000 and 2001.

This on-going project contrasts the emissions and air quality impacts of four urban growth scenarios developed through a regional “visioning” initiative known as Envision Central Texas with a scenario based on an integrated transportation-land use model. Visioning is a highly community-oriented planning technique used to create regional land use and transportation goals (FHWA 1996). In contrast, land use models are based on historical trends and attempt to forecast or predict what future land use patterns will look like based on those trends (along with changes in any policy, land use, travel cost or other variable that the analyst has incorporated into the model). While fundamentally different, both the visioning and modeling processes carry benefits (Lemp et al. 2007). Results to date focus on changes in air quality and population exposure due to changes in land use, and consequently emissions and dry deposition, for the four ECT scenarios. Development and application of the ITLUM are also described along with the land use and transportation effects of three policy scenarios, including roadway pricing, a density floor, and an urban growth boundary.

Air Quality Predictions for the ECT Scenarios
All four of the ECT land use “visions” are based on the assumption of a doubling of population within 20 to 40 years in the five-county Austin area and introduction of the same number of new jobs. However, the scenarios assume very different types of growth. Biogenic and anthropogenic emission inventories for the four scenarios were developed for use in the Comprehensive Air Quality Model with extensions (CAMx), which is currently the photochemical model used by the State of Texas for attainment demonstrations. CAMx was used to predict the spatial and temporal patterns of ozone concentrations for each scenario. The results for the ECT scenarios were contrasted with a Base Case that included 2007 emission inventories for anthropogenic sources, emission controls adopted for Austin’s Early Action Compact, and a biogenic emissions inventory and dry deposition estimates based on a land cover/land use database developed by Wiedinmyer et al. (2000, 2001). As compared to the Base Case, total NOx emissions for the ECT scenarios decreased significantly primarily due to the phase-in of new emission standards for mobile sources, while total VOC emissions were predicted to increase due to increases in area source emissions which were projected using human population. Although the emission estimates varied between the scenarios, all of the model simulations were conducted using meteorology from a September 13-20, 1999 historical ozone episode that was originally developed for Austin’s Early Action Compact or voluntary State Implementation Plan under the National Ambient Air Quality Standard (NAAQS) for ozone concentrations averaged over 8-hours.

Inter-comparison of air quality predictions for the scenarios is complex and multifaceted, i.e., differences can be evaluated from a number of perspectives including their temporal impacts, spatial impacts, differences in magnitude across the five-county region versus only in areas with high ozone concentrations, and influence on population exposure.

Predicted 1-hour averaged daily maximum ozone concentrations for the 2007 Base Case ranged from 72 ppb to 90 ppb across the episode. Differences in daily maximum 1-hour ozone concentrations between the ECT scenarios and the 2007 Base Case due to the combined impacts of changes in biogenic emissions and dry deposition ranged from -0.9 ppb to 0.1 ppb with typical values of -0.2 ppb for 5-county Austin area. Differences in daily maximum 1-hour ozone concentrations due to the combined changes in anthropogenic emissions from on-road mobile, non-road mobile and area sources ranged from -11 ppb to -2 ppb with typical values of -6 ppb. Area-wide daily maximum ozone concentrations in the Austin region have historically been predicted to be most responsive to NOx emissions reductions, and reductions in ozone concentrations for the ECT scenarios are consistent with decreases in NOx emissions from the phase-in of new standards for mobile sources.

Maximum and minimum differences in 1-hour ozone concentrations that occurred across the region regardless of time of day or magnitude were also investigated. Maximum differences in hourly ozone concentrations between the ECT scenarios and the 2007 Base Case due to the combined impacts of changes in biogenic emissions and dry deposition ranged from -1.4 ppb to 0.7 ppb. Maximum differences in hourly ozone concentrations due to the combined changes in anthropogenic emissions from on-road mobile, non-road mobile and area sources ranged from -14 to 22 ppb. In contrast, maximum differences in hourly ozone concentrations between ECT scenarios ranged from -3 to 5 ppb. Overall the doubling of population and implementation of new federal mobile source standards produced greater changes in emissions and air quality than differences in spatial patterns due to different types of regional development implying that controlling the environmental impacts of urbanization involves multi-faceted strategies.

In addition to changes in ozone concentrations, a characterization of air pollutant exposure was examined. Total daily population-weighted exposure above a threshold ozone concentration (ppb) (Wang 2006) was characterized using the following metric:

Equation 1.

where pt is the total population in the five-county Austin area for the scenario, pg is population in each grid cell g and sg,h is the ozone concentration (ppb) over the threshold cthresh for each grid cell g at hour h, Equation 2. . Note that the population doubles from approximately 1,250,000 people in the Base Case to 2,500,000 people in the ECT scenarios. This metric was evaluated for various threshold values and estimated for each grid cell, summed over the Austin area modeling domain, and over all hours of the day.

Total daily population-weighted exposure was estimated for the Base Case and the two ECT scenarios that represent the most extreme differences in development patterns: (1) ECTA, which is consistent with Austin’s historical pattern of low-density, segregated-use development based on extensive highway provision, and (2) ECT D, which is high-density development and balanced-use zoning. For a threshold value of 40 ppb (which is a value typical of clean background conditions), all ECT scenarios show greater exposure than the Base Case due to additional increases in ozone and population in newly developed areas. For a threshold value of 80 ppb, the Base Case shows greater exposure than the ECT scenarios since daily maximum ozone concentrations were lower for the ECT scenarios. Concentrated high-density development in existing towns with balanced-use zoning produced lower exposure to high ozone concentrations than a more typical pattern of urban sprawl. Evaluating daily population exposure can provide additional information about the magnitude and spatial distribution of changes in ozone due to urban development and can be particularly relevant in the context of environmental equity.

Models of Travel Demand
As part of our integrated transportation-land use modeling effort, travel demand models (TDMs) have been applied in coordination with land use models. To date, the TDM used for these purposes was calibrated using the 1996-1997 Austin Travel Survey (ATS) by Smart Mobility, Inc. (SMI) as part of the Envision Central Texas (ECT) visioning project. The model uses a variation of the traditional four-step travel demand modeling framework. Two other TDMs also have been recently calibrated for the region; and these serve as enhancements to SMI’s approach, while offering a valuable point of TDM comparison. The first is similar to SMI’s TDM in that it follows the traditional four-step framework. However, trip distribution is performed in a destination choice framework (instead of SMI’s simpler gravity-based approach) and feedback procedures incorporate the method of successive averages (MSA) (rather than SMI’s less efficient, naïve feedback approach). The second TDM uses an activity-based, microsimulation structure, which differs substantially. This model simulates the behavior of each individual in the region distinctly, and uses a tour- (rather than trip-) based approach, generally considered to be more behaviorally and theoretically sound. The activity-based, microsimulation TDM is more sensitive to changes land use and transportation system variables and will be integrated with the land use models in the coming year.

Development and Application of an Integrated Transportation-Land Use Model
As a critical point of comparison for the four Envision Central Texas (ECT) “visions” of Austin’s future, the five-county region’s year 2030 travel conditions and household and employment distributions were predicted using a variation of Steven Putman’s gravity-based land use model (LUM). Model calibration and applications were coded in MATLAB. This model system has three components: RESLOC (household allocation), EMPLOC (employment allocation), and LUDENSITY (land use density). A travel demand model (TDM) was externally linked to the LUM system in order to update travel conditions and provide a well-defined, series of related steps to all future household and employment forecasts (at five-year intervals). All together, the system of equations form what is called an “integrated transportation-land use model” (ITLUM). In this integrated modeling framework, the EMPLOC model runs before the RESLOC, followed by LUDENSITY and a TDM. The EMPLOC model output (employment by category by zone) serves as an input to the RESLOC. Predicted household and employment levels (by category/type) are LUDENSITY’s primary inputs. A TDM was applied immediately after allocating households and jobs (and estimating land consumption levels), in order to update travel times between zones and the relative attractiveness of each zone.

To be consistent with the ECT scenarios, the model system predicts the spatial distributions of six household types (categorized by number of workers [0, 1 and 2+] and presence of children) and three employment categories (basic, retail and service jobs). Three other employment types (namely Airport, K-12 Education and Higher Education) were assumed to follow ECT’s trend scenario because these three employment types vary significantly over space yet are relatively stable over time (at least for all zones that have non-zero employment counts). Non-linear least squares (NLLS) techniques were used in model calibration for the three components.

The land use and transportation effects of four distinctive policies were investigated: including a business-as-usual (base) scenario, roadway pricing (congestion pricing plus a per-mile-traveled carbon tax), a density floor (no new low-density development), and an urban growth boundary (prohibiting new development in presently peripheral, largely undeveloped zones).

Households and employment tend to remain concentrated in the urban areas and along regional freeways, assuming business as usual. The distribution patterns of implementing congestion pricing and a carbon tax are similar to the results of business as usual. This suggests that the combined policy of congestion pricing and carbon tax does not alter the location choices of households and firms in a significant way, but does affect how far people travel. As discussed later, such policies significantly reduce overall VMT. As compared to the base scenario, households are more “sprawling” in the northern part of the City of Austin (CoA) in the density floor policy. This suggests that this policy allows more households to locate in residentially preferred places. Employment allocation is not directly affected by this policy, and therefore differences in employment forecasts between the base scenario and this density floor policy scenario are less obvious than those in households. It seems employment tends to develop in the north part of the CoA and along regional freeway IH-35. As required by the UGB policy, all the new development (households, basic, retail and service employment) happens within the pre-defined zones; any households and basic, retail and service jobs outside of the boundary already existed in the year 2005.

The TDM results show that the road pricing and UGB policies are very effective, in terms of reducing VMT. These two policy scenarios are estimated to reduce regional VMT by 16.0% and 17.2%, respectively, resulting in reductions of 13.54 million and 14.58 million VMT per day, respectively. In contrast, the density floor policy is estimated to slightly increase regional VMT, by 0.2% (or 165,000 VMT). The added development at the City’s periphery may explain this VMT increase.

The business as usual, road pricing and density floor policies generate similar amounts of personal and commercial trips, and the UGB policy generates less of both types of trips. While the road pricing policy has the same amount of trips as the business as usual scenario, the UGB policy is estimated to decrease the total trips by 4.8%, or 0.57 million trips. The reductions in VMT and trips suggest that the VMT reduction of the road pricing policy basically comes from shorter trips, while VMT reduction of the UGB policy comes from both shorter trips and fewer trips. The external-local and external-through personal trips are 185,660 and 73,461 for the four policy scenarios.

The road pricing and UGB policies are most effective in promoting transit usage and decreasing the number of auto trips at the same time. Transit trips when implementing these two policies are estimated to increase by 0.13 million and 0.07 million in the whole region, as compared to the business as usual scenario, and the auto trips decrease by 0.12 million and 0.70 million, respectively. The UGB policy enjoys the highest number of walk/bike trips, bettering all the other three policies by 0.1 million. The density floor policy seems to have similar mode share patterns with the business as usual scenario, with slightly fewer transit trips and slightly more auto trips.

Future Activities:

During 2008, the team will develop emissions estimates for the ITLUM forecasts and compare the emissions and air quality predictions from these scenarios. Human exposure patterns based on the ECT and ITLUM forecasts will be compared by coupling predicted ozone concentrations from CAMx with demographic forecasts to generate population exposure metrics. U.S. EPA is in the process of developing post-Clean Air Act Amendment emission scenario projections which, when available, can be used to test the hypothesis that predicted human exposure patterns based on these scenarios will differ from those based on the ITLUM or ECT emission forecasts.

Another set of land use and traffic forecasts will be generated using the combination of land use change models (Zhou and Kockelman, 2005) and land use intensity models (Zhou and Kockelman, 2006). Since two sets of land use data for the entire 5-county region are not available in the near term, the land use change models require modifications of parameters that were estimated using two GIS-encoded land use maps for the City of Austin.


Journal Articles on this Report : 5 Displayed | Download in RIS Format

Other project views: All 51 publications 14 publications in selected types All 13 journal articles
Type Citation Project Document Sources
Journal Article Lemp JD, McWethy LB, Kockelman KM. From aggregate methods to microsimulation: assessing benefits of microscopic activity-based models of travel demand. Transportation Research Record 2007;1994:80-88. R831839 (2007)
R831839 (Final)
  • Full-text: University of Texas-Prepublication Paper
    Exit
  • Abstract: TRB-Abstract
    Exit
  • Journal Article Wang X, Kockelman K. Tracking land cover change in a mixed logit model:recognizing temporal and spatial effects. Transportation Research Record 2006;1977:112-120. R831839 (2005)
    R831839 (2006)
    R831839 (2007)
    R831839 (Final)
  • Abstract: Transportation Research Board-Abstract
    Exit
  • Other: University of Texas-Prepublication Paper
    Exit
  • Journal Article Wang X, Kockelman KM. Specification and estimation of a spatially and temporally autocorrelated seemingly unrelated regression model: application to crash rates in China. Transportation 2007;34(3):281-300. R831839 (2006)
    R831839 (2007)
  • Abstract: SpringerLink-Abstract
    Exit
  • Journal Article Zhou BB, Kockelman KM. Predicting the distribution of households and employment: a seemingly unrelated regression model with two spatial processes. Journal of Transport Geography 2009;17(5):369-376. R831839 (2007)
  • Abstract: ScienceDirect-Abstract
    Exit
  • Journal Article Zhou B, Kockelman KM. Neighborhood impacts on land use change:a multinomial logit model of spatial relationships. The Annals of Regional Science 2008;42(2):321-340. R831839 (2007)
    R831839 (Final)
  • Abstract: SpringerLink-Abstract
    Exit
  • Other: University of Texas-Prepublication Paper PDF
    Exit
  • Supplemental Keywords:

    urban air quality, land use models, travel demand models, emission forecasts, ozone, air quality models, central Texas, air, isoprene, dry deposition,, RFA, Health, Scientific Discipline, PHYSICAL ASPECTS, Air, Ecosystem Protection/Environmental Exposure & Risk, RESEARCH, Health Risk Assessment, climate change, Air Pollution Effects, Risk Assessments, Monitoring/Modeling, Monitoring, Environmental Monitoring, Physical Processes, Ecological Risk Assessment, Atmosphere, ecosystem models, integrated assessments, particulate matter, air quality modeling, atmospheric measurements, model-based analysis, remote sensing, motor vehicle emissions, fine particles, automobile exhaust, exposure, global change, model assisted estimation, air pollution, green house gas concentrations, air quality model, modeling, human exposure, climate models, environmental stressors, human activity, landscape characterization, air quality assessments, airborne urban contaminants, human health risk, land use, air quality, ambient air pollution, public health effects, ozone concentrations, transportation, atmospheric chemistry

    Progress and Final Reports:

    Original Abstract
  • 2005 Progress Report
  • 2006 Progress Report
  • Final Report