The Impacts of Urbanization on Emissions and Air Quality

Center for Energy and Environmental Resources, The University of Texas at Austin, Austin, Texas, and Tufts University, Medford, Massachusetts. Environ...
0 downloads 0 Views 3MB Size
Download PDF
Environ. Sci. Technol. 2008, 42, 7294–7300

The Impacts of Urbanization on Emissions and Air Quality: Comparison of Four Visions of Austin, Texas JIHEE SONG,† ALBA WEBB,† BARBARA PARMENTER,‡ DAVID T. ALLEN,† AND E L E N A M C D O N A L D - B U L L E R * ,† Center for Energy and Environmental Resources, The University of Texas at Austin, Austin, Texas, and Tufts University, Medford, Massachusetts

Received March 4, 2008. Revised manuscript received July 13, 2008. Accepted August 5, 2008.

The impacts of alternative regional development patterns on emissions, dry deposition, and air quality were examined using four visions of future land use in Austin, Texas associated with a doubling of the population in 20-40 years from 2001. Emissions and their spatial allocation were determined based on the development pattern and used to predict hourly ozone concentrations. Differences in hourly ozone concentrations due to changes in anthropogenic emissions between the future case scenarios and a 2007 base case ranged from -14 to 22 ppb and were primarily associated with the implementation of federal mobile source standards; differences due to biogenic emissions and dry deposition due to urbanization ranged from only -1.4 to 0.7 ppb. These differences in the magnitude of emissions produced greater changes in air quality than differences in regional development patterns between the four scenarios. Differences in hourly ozone concentrations between the future development scenarios and a 2007 base case ranged from -14 to 22 ppb, in contrast to differences of -3 to 5 ppb between the future scenarios. The results imply thatalthoughtheeffectsofurbanizationpatternsarenon-negligible, the pattern of urban development is not as significant as reductions in emissions per capita.

Introduction Integrated land use-transportation scenario planning has become an increasingly common component of regional and urban planning processes over the past 15 years (1); an assessment by Bartholomew (1) for the Federal Highway Administration in 2005 identified 80 scenario planning projects in more than 50 metropolitan areas throughout the United States. In a typical metropolitan visioning process, stakeholders engage in a collaborative process to develop alternative future urban development patterns that reflect different policy options. These patterns are in turn evaluated based on a number of indicators, usually related to transportation values (1). Typically, however, these alternative scenarios are not directly incorporated into decadal-scale * Corresponding author phone: 512-471-2891; fax: 512-471-1720; e-mail: [email protected]. † The University of Texas at Austin. ‡ Tufts University. 7294

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 19, 2008

emission forecasts prepared for air quality regulatory requirements. This study examines the impacts of four visions of regional development on biogenic and anthropogenic emissions, dry deposition, and air quality using Austin, Texas as a case study. The five-county Austin-Round Rock Metropolitan Statistical Area (MSA) has been among the most rapidly growing urban areas in the United States. The population of the MSA is approximately 1.4 million, and it is concentrated in Travis County, which includes the city of Austin, and Williamson County. 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. Four growth scenarios for the region have been developed through a community-driven regional visioning process known as Envision Central Texas (ECT) that began in 2001 (2). The scenarios are all based on a doubling of population in 20-40 years from 2001, but assume different types of growth. ECT Scenario A assumes low-density, segregateduse development based on extensive highway provision; ECT Scenario B assumes concentrated, contiguous regional growth within 1-mile of transportation corridors; ECT Scenario C concentrates growth in existing and new communities with distinct boundaries; ECT Scenario D assumes high-density development in existing towns and cities with balanced-use zoning. Section S-1 of the Supporting Information (SI), shows current land cover types in the five-county region (3), and Figure 1 shows land use development patterns for each of the four scenarios. While implementation practices and particular details will differ to accommodate regional differences, visioning strategies share common attributes across regions with similar goals (4). Bartholomew (1) describes that the majority of scenario planning projects undertaken in 50 metropolitan areas in the United States focused on spatial patterns and urban form and included three to four scenarios chosen from five different archetypes: (1) center, cluster, or satellite; (2) compact, (3) dispersed, fringe, or highway-oriented, (4) corridor, and (5) infill or redevelopment. Urbanization has the potential to impact air pollutant concentrations through a number of mechanisms (5, 6). For example, Civerolo et al. (5) investigated the influence of land use change on surface meteorology and ozone concentrations in the New York City metropolitan area using the narrative of the “A2” Scenario of the Intergovernmental Panel on Climate Change. Urban growth produced higher near-surface temperatures across the New York area, but had complex, spatially heterogeneous impacts on ozone concentrations in outlying counties experiencing growth (5). As part of the scenario planning process, future vehicle miles traveled (VMT) are typically estimated and often onroad mobile source emissions, but differences in land cover and biogenic and other anthropogenic emissions, as well as predictions of air quality, are usually not considered as part of selecting a preferred community vision. Similarly, national and even state-level future emission scenarios may not incorporate community visions of development. The current study compares the magnitude of emission and air quality differences between four future visions of the Austin area using the Comprehensive Air Quality Model with extensions (CAMx (7);). Impacts on daily maximum 1 h ozone concentrations and hourly episodic concentrations of changes in biogenic emissions and dry deposition are also contrasted with those due to anthropogenic emissions from area and 10.1021/es800645j CCC: $40.75

 2008 American Chemical Society

Published on Web 09/04/2008

FIGURE 1. ECT Scenarios: maps indicating land use changes that will occur for each of the growth scenarios. 2001 developed land and counties within the Austin-Round Rock MSA are identified for reference. nonroad mobile sources and on-road mobile sources. Because the assessment of future energy needs has been ongoing in Texas and because forecasts of the growth of point source emissions have been small relative to changes in mobile source emissions, point source emissions were assumed to be constant.

Estimation of Biogenic Emissions and Dry Deposition Biogenic Emissions Model. The Global Biogenic Emissions and Interactions System (GloBEIS) version 3.1 was used to develop biogenic emission inventories for a September 13-20th, 1999 modeling episode for the current study (7, 8). The photochemical modeling domain was a nested regional/ urban scale 36 km/12 km/4 km grid with 12 vertical layers from the surface to 3.9 km; the five-county Austin area was included within the 4 km domain. Data on temperature, photosynthetically active radiation (PAR), wind speed, and humidity used to obtain biogenic emission estimates from GloBEIS are described in Section S-1 of the SI. Although it is typically presumed that land cover remains constant in future predictions of air quality to support attainment demonstrations, changes in land cover can affect predicted air pollutant concentrations by influencing biogenic emissions, deposition velocities, and other physical parameters (9, 10). Land use/land cover (LULC) data for the current study were derived from several databases. The first database, which will be referred to as the 2007 Base Case, is described in detail in SI Section S-1. This database was developed by Wiedinmyer et al. (3, 11) in order to improve the characterization of land cover, leaf biomass, and leaf density distributions in Texas. The other LULC databases

were derived from each of the ECT scenarios. These scenarios used only 10-16 land use classes (10 land use types for Scenario A and 16 for the other scenarios). In contrast to the Wiedinmyer et al. database (3, 11), the ECT land use classifications included assumptions concerning impervious ground cover but no information on vegetation types. Therefore, the ECT land use scenarios were overlaid on the original land cover data from Wiedinmyer et al. (3, 11) and used to modify the original vegetation density. ECT planners estimated the fraction of impervious cover for each ECT land use type, which was used in this study to adjust the fraction of original vegetation expected to exist in that land use category. Section S-1 of the SI describes the assumed fraction of original vegetation remaining for each land use type for the ECT scenarios. Dry Deposition Model. Dry deposition estimation methods in CAMx are based on the work of Wesely et al. (12) and Walmsley and Wesely (13). This algorithm is the most commonly used approach in urban and regional-scale photochemical models, but has not undergone field validation in Texas. In this algorithm, dry deposition rates are influenced by resistances due to three mechanisms; aerodynamic transport, diffusion across a quasi-laminar sublayer, and surface uptake (12). The dry deposition flux is calculated as follows: Fc ) Vd · Cz

(1)

where Fc is the dry deposition flux of the gas of interest, Vd is the dry deposition velocity, and Cz is the concentration or mixing ratio at the midpoint of first vertical layer height in VOL. 42, NO. 19, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

7295

TABLE 1. Transportation Characteristics for the ECT Scenarios scenario

daily VMT per capita

auto trips (%)

transit trips (%)

bike and walk trips (%)

commuter rail and toll roads

light rail

bus rapid transit

2000 ECT A ECT B ECT C ECT D

26.4 34.3 30.1 29.0 27.4

94 92 90 88 85

3 4 6 4 6

3 4 4 8 9

no yes yes yes yes

no no yes no yes

no yes no yes no

CAMx. For gases, the dry deposition velocity is calculated as follows: Vd )

1 ra + r d + r s

(2)

where ra is the aerodynamic resistance above the surface, rd is the deposition layer (or quasi-laminar sublayer) resistance and rs is the bulk surface (or canopy) resistance (12). Eleven land use/land cover categories are used in CAMx which are urban land, agricultural land, range land, deciduous forest, coniferous forest, mixed forest including wetland, water, barren land, nonforested wetland, mixed agricultural/ range land, and rocky open areas with low-growing shrubs. CAMx land use files assign the areal fractional distribution (0-1) of 11 land use categories in each individual grid cell. For the 2007 Base Case, land cover data from Wiedinmyer et al. (3, 11) were mapped to one of the 11 land use/land cover categories used by the dry deposition module in CAMx (14). For the ECT scenarios, the remaining vegetation for each development type (from Table S-1 in the SI) was classified as the original land cover, and the area fraction of newly developed land was classified as urban. SI Table S-2 shows the fraction of each of the eleven land use/land cover categories for the 2007 Base Case and the ECT scenarios as well as the percent of vegetative cover converted to urban land use. The difference in vegetative cover between the ECT scenarios and the 2007 Base Case ranged from 4 to 17% with the highest percentage consumed by ECT A that continues the current pattern of low-density development and the lowest percentage by ECT D that has infill and redevelopment in existing areas. Much of the new development associated with ECT A resulted in transitions from mixed agricultural and rangeland or mixed forest to urban land use in Travis and Williamson Counties.

Anthropogenic Emission Inventory Development The modeling episode, September 13-20th, 1999, selected for this work was used for the development of Austin’s Early Action Compact (15). Emissions inventories for a 1999 base year and 2007 projected attainment demonstration were developed for the regulatory analyses (16, 17). With the exception of emissions from nonroad sources that are described below, the 2007 future year anthropogenic emission inventories developed for Austin’s Early Action Compact served as the Base Case for this study. Emissions from stationary point sources for 2007 were developed by the TCEQ for the Houston-Galveston area Mid-Course Review and were supplemented with local activity projections from CAPCOG (17); these emissions remained constant between the 2007 Base Case and ECT future development scenarios. The methodology used to develop emission estimates for mobile and area sources for the ECT scenarios is described below and in the Supporting Information. On-Road Mobile Source Emissions. On-road mobile source emission inventories were developed for each of the ECT scenarios by combining travel demand model output from the Envision Central Texas Transportation Model (ECTTM) for the five-county Austin area link network with 7296

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 19, 2008

emission factors from the EPA’s MOBILE6.2 model as described in SI Section S-3 (18). The ECTTM was developed by Smart Mobility Inc., with support from the Capital Area Metropolitan Planning Organization (CAMPO); it follows the general four-step modeling framework of traditional travel demand models (trip generation, trip distribution, mode split, and traffic assignment), but includes a number of enhancements that make it sensitive to transportation infrastructure and land use (19). These include (a) an auto availability model that is sensitive to residential density and transit service, (b) a walk/bike trip model that is sensitive to residential density, employment density, and the balance between jobs and housing, (c) a mode choice model that is sensitive to land use, and (d) feedback of congested travel times to affect traveler behavior. Transportation characteristics for each of the ECT scenarios are summarized in Table 1 (2, 20); characteristics for 2000 are shown for comparison. MOBILE6.2 was used to calculate emission factors (grams mile-1) for volatile organic compounds (VOC), carbon monoxide (CO), and nitrogen oxides (NOx) by hour of day, by vehicle type, and by road type (or drive cycle) (18). Since the ECT scenarios are based on a projected doubling of population within 20-40 years from 2001, emission factors which include default federal motor vehicle control programs (FMVCP) were developed for the year 2030. The EPA’s FMVCP rules regulate fuel characteristics and require increasingly lower exhaust and evaporative standards for new vehicles. The most recently adopted FMVCP rules modeled in MOBILE6.2 include the Tier 2 and the heavy-duty 2007 rules (21). Tier 2 requires more stringent exhaust emission standards for all passenger vehicles, includes more stringent evaporative emission standards for light-duty classes, and reduces the average gasoline fuel sulfur content. The heavyduty 2007 rules introduce more stringent emission standards for gasoline and diesel heavy-duty engines as well as reducing the average diesel fuel sulfur content. As a consequence of fleet turnover which occurs over the span of decades, the number of older vehicles with less effective pollution controls which are still on the road will decrease resulting in an overall cleaner fleet for the ECT scenarios as compared to the 2007 Base Case. Non-Road Mobile Source Emissions. The 2007 nonroad mobile source emission inventory that was developed for Austin’s Early Action Compact analysis was not used for this study. Instead, population and households from 2001 that served as the base year for the visioning process by Envision Central Texas and land use and parcels data, that had become available from the City of Austin and CAPCOG since the development of the original 2007 nonroad inventory, were used in conjunction with the EPA’s NONROAD model version 2005 (NONROAD 2005) to project a new 2007 inventory and four ECT 2030 nonroad inventories. The exceptions were emissions from aircraft, military service operations, locomotives, and residential and commercial gas cans, which were developed for the 2007 Base Case inventory from local surveys (16, 17). Development of the emission estimates for all nonroad mobile sources is described in detail in SI Section S-4, including assumptions about equipment populations

TABLE 2. Weekday VMT and Total on-Road Mobile, Non-Road Mobile, And Area Source Emissions (tpd) of VOC and NOx for the 2007 Base Case and Four ECT Scenariosa 2007 Base Case VMT ) 44.5b

ECT A VMT ) 82.4b

ECT B VMT ) 72.2b

ECT C VMT ) 69.5b

ECT D VMT ) 65.9b

categories

VOC

NOx

VOC

NOx

VOC

NOx

VOC

NOx

VOC

NOx

on-road mobile nonroad mobile area

33.8 22.2 110.7

62.1 21.7 10.2

22.0 23.2 214.3

18.4 9.5 20.6

19.2 24.0 237.7

16.0 9.6 22.1

18.8 24.0 261.6

15.6 9.6 23.6

17.0 23.2 236.2

14.3 9.5 22.1

a Note: ECT scenario emissions are calculated for a future year of 2030. the 5-county Austin area.

and activity, emission controls, and in particular, the development of spatial allocation factors and surrogates. Typically, in air quality modeling conducted for regulatory analyses, changes in the magnitude of nonroad and area source emissions are accounted for in future year forecasts but changes in their spatial allocation are not. Thus, for Texas and other states that routinely prepare and update State Implementation Plans, spatial allocation factors and surrogates remain the same between the base year and attainment year. This study investigated how the magnitude of the emissions would change with a doubling of population and implementation of new federal standards, as well as account for spatial differences in emissions from different regional development patterns. Two levels of spatial allocation were developed: first, spatial allocation factors incorporated within the NONROAD model used to allocate state-level equipment populations to county-level equipment populations, and second, spatial surrogates for allocating county-level emissions to each grid cell in the CAMx modeling domain for the five-county Austin area. Area Source Emissions. Area sources include stationary point sources that are too small or numerous to be surveyed and characterized individually. Emissions from these sources are estimated collectively and spatially allocated according to surrogates such as population or income. Area source emission inventories for each ECT scenario were developed by projecting 2007 base year area emissions using human population growth and applying federal and state emission standards (17, 22). Spatial allocation of county-level area source emissions was conducted using the same approach as for the nonroad emissions.

Prediction of Air Pollutant Concentrations Hourly ozone concentrations for each scenario were predicted using CAMx with meteorological conditions representing those during September 13-20, 1999 (15). Meteorological conditions during this episode are typical of multiday high ozone events in Central Texas. The chemical mechanism and meteorological inputs used in CAMx as well as the 29 modeling simulations conducted are summarized in Section S-5 of the SI.

Results and Discussion ECT Emissions Estimates. Isoprene emissions for the 2007 Base Case ranged from 1.5 to 3 Mmoles day-1 across the episode with an average of 2 Mmoles day-1. Predictions of isoprene emissions for the ECT scenarios were compared to predictions from the 2007 Base Case. Since the first two days of the modeled episode, September 13th and 14th, were used for model “spin-up”, results of these days are not included. Differences in land use/land cover led to 2-6% reductions in daily biogenic emissions across the five-county Austin MSA. If the percentage change in biogenic emissions is restricted to grid cells that experienced land cover changes, the percentage reductions are larger, ranging from 5 to 11%. ECT Scenario A, which assumes a typical urban sprawl pattern with the largest consumption of vegetative cover, shows the

b

VMT is given in units of 106 miles per day in

largest reductions in isoprene emissions. The reductions occur primarily in Travis and Williamson Counties where much of the transition from mixed agricultural and rangeland or mixed forest to developed land occurs. Estimates of total on-road mobile source emissions, nonroad mobile source emissions, and area source emissions for the four ECT scenarios are compared to those for the 2007 Base Case in Table 2. Disaggregation of these estimates by individual source categories is described in SI Section S-6. Differences in the spatially allocated on-road mobile source, nonroad mobile source, and area source emissions for two ECT scenarios, ECT A and ECT D, are also described in SI Section S-6 for one episode day (September 20 at 1400) as an example. These figures compare emissions from ECT A to the 2007 Base Case which represents changes due to the doubling of population with a continuation of current development patterns, and they also compare emissions from ECT D and ECT A, which represent the two most extreme development scenarios. With a continuation of current development patterns, VOC and NOx emissions from on-road mobile sources decreased substantially for ECT A even with a doubling of population relative to the 2007 Base Case. These reductions are due to more stringent federal motor vehicle emission control programs, including the EPA’s Tier 2 and heavy-duty 2007 rules. Emission reductions are concentrated in the urban core and along the main interstate highway (IH-35) while emissions increase relative to the 2007 Base Case in newly developed areas. Decreases in NOx emissions from nonroad sources are primarily due to reductions in emissions from industrial and construction/mining equipment, which are subject to the EPA’s Tier 4 engine standards (23). Emissions from these two source categories are allocated throughout the urban area. In contrast, increases in VOC emissions from nonroad mobile sources are primarily attributed to lawn and garden equipment and therefore concentrated in areas with new population growth. VOC and NOx emissions from lawn and garden equipment increased because emissions from gas cans, included in this category, were projected based on growth in the number of households as described in Section S-4 of the SI. Although total NOx emissions from nonroad sources decreased between the ECT Scenarios and the 2007 Base Case, NOx emissions from lawn and garden equipment nearly doubled relative to the 2007 Base Case because the NONROAD model does not incorporate new emission controls between 2007 and 2030 for most 2-stroke and some 4-stroke lawn and garden equipment. VOC and NOx emissions from area sources increased substantially because they were projected according to human population. Regions with large increases in NOx emissions were located in areas with agriculture, rangeland, deciduous forest, coniferous forest, and mixed agriculture/ rangeland. Relative to other counties, Travis County has very little rural area. When spatially allocating emissions for the ECT scenarios, NOx emission increases from agricultural production were concentrated in small areas, which led to VOL. 42, NO. 19, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

7297

FIGURE 2. Range of changes in hourly ozone concentrations (ppb) between the ECT Scenarios and the 2007 Base Case across the five-county Austin area due to changes in (a) biogenic emissions and dry deposition only, (b) nonroad mobile and area source emissions only, (c) on-road mobile source emissions only, (d) anthropogenic emissions only, and (e) biogenic emissions, dry deposition and anthropogenic emissions. large differences in NOx emissions between the ECT scenarios and the 2007 Base Case in these grid cells. Differences between ECT D and ECT A were smaller than the differences between these two scenarios and the 2007 Base Case. In general, ECT D resulted in lower anthropogenic emissions than ECT A. These differences were concentrated in the urban core of Austin and along major transportation corridors. Relative Air Quality Impacts of Urbanization due to Changes in Anthropogenic Emissions versus Changes in Biogenic Emissions and Dry Deposition. Predicted 1 h averaged daily maximum ozone concentrations for the 2007 Base Case ranged from 72 to 90 ppb across the episode. Changes in daily maximum hourly ozone concentrations due to decreases in biogenic emissions associated with increased urbanization (O3(ECT scenarios)-O3(2007 Base Case)), ranged from -0.02 to -1 ppb, with a typical value of -0.3 ppb for the Austin area. Differences in dry deposition velocities alone due to development increased daily maximum 1 h ozone concentrations relative to the 2007 Base Case by 0-0.3 ppb. In Wesely’s model, dry deposition velocities for mixed agricultural/range land or forests are higher than for urban areas during the daytime, but lower at night during midsummer conditions (24). Consequently, loss of vegetative cover due to urbanization leads to less removal of ozone during the afternoon and higher maximum daily ozone concentrations, but decreases in ozone concentrations during the night. Because of large NOx emissions reductions from nonroad mobile sources, changes in daily maximum 1 h ozone concentrations due to changes in area and nonroad mobile source emissions only ranged from -0.1 to 0.8 ppb, with typical values of 0.2 ppb for the Austin area. The impacts of changes due to on-road mobile sources alone were substantially larger. Reductions in on-road mobile source emissions, resulting from implementation of federal motor vehicle control programs, led to changes in area-wide daily maximum hourly ozone concentrations for the ECT scenarios of up to -10 ppb. Differences in daily maximum 1 h ozone concentrations due to the combined impacts of changes in biogenic emissions and dry deposition ranged from -0.9 to 0.1 ppb with typical values of -0.2 ppb for the five-county Austin area. Differences in daily maximum 1 h ozone concentrations due to the combined changes in anthropogenic emissions 7298

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 19, 2008

from on-road mobile, nonroad mobile and area sources ranged from -11 to -2 ppb with typical values of -6 ppb. In addition to differences in area-wide daily maximum 1 h ozone concentrations between the ECT scenarios and the 2007 Base Case, maximum and minimum differences in 1 h ozone concentrations that occurred across the region regardless of time of day or magnitude were investigated. This analysis could be relevant for assessments of population exposure, which may not be spatially or temporally correlated with area-wide daily maximum ozone concentrations. Figure 2 shows the range of changes in 1 h ozone concentrations between the ECT scenarios and the 2007 Base Case due to changes in (a) biogenic emissions and dry deposition, (b) nonroad mobile and area source emissions, (c) on-road mobile source emissions, (d) anthropogenic emissions, and (e) biogenic emissions, dry deposition, and anthropogenic emissions. Changes in ozone concentrations due only to changes in biogenic emissions and dry deposition are relatively smaller than the changes due to anthropogenic emissions. Both significant increases and decreases in ozone concentrations were associated with changes in anthropogenic emissions, and similar to the results of Civerolo et al. (5) for the New York City metropolitan region, the spatial patterns of ozone changes with urbanization were heterogeneous. Maximum differences in hourly ozone concentrations were predicted between ECT A and the 2007 Base Case for the changes due to biogenic emissions and dry deposition alone (+0.7 to -1.4 ppb), anthropogenic emissions alone (+22 to -14 ppb), and both in tandem (+22 to -14 ppb). As shown in SI Section S-7, decreases occurred in the afternoon (1400) in eastern Travis County and western Bastrop County. Increases were primarily due to reductions in on-road mobile source emissions along transportation corridors in the Austin urban core which resulted in less titration of ozone by NOx emissions during the early morning hours (0600). The range of differences in hourly ozone concentrations between the other ECT scenarios and the 2007 Base Case were generally within 1-2 ppb of the range of differences between ECT A and the 2007 Base Case. These results are consistent with previous studies that have shown that ozone formation in the Austin area is generally NOx-limited with VOC-limited conditions near the I-35 corridor in central Travis County (15). Consequently, as part of its Early Action Compact, the Austin area has pursued

FIGURE 3. Range of changes in hourly ozone concentrations (ppb) between ECT D and ECT A across the five-county Austin area due to changes in (a) biogenic emissions and dry deposition only, (b) nonroad mobile and area source emissions only, (c) on-road mobile source emissions only, (d) anthropogenic emissions only, and (e) biogenic emissions, dry deposition and anthropogenic emissions. NOx control strategies as more effective than VOC strategies for reducing ozone levels (15). The responsiveness of hourly peak ozone concentrations to anthropogenic NOx reductions in the Austin area is predicted to continue with future patterns of urbanization. SI Section S-8 shows the differences in hourly ozone concentrations between ECT A and the 2007 Base Case versus the 2007 Base Case across all grid cells and episode days due to the combined changes in biogenic emissions, dry deposition, and anthropogenic emissions. Decreases in ozone concentrations are primarily associated with high ozone concentrations. Plots for the other ECT Scenarios showed similar trends. Both reductions in high ozone concentrations and increases in lower ozone concentrations were due to reductions of emissions from on-road mobile sources in the future scenarios. Differences in hourly ozone concentrations between the ECT scenarios and the 2007 Base Case were much greater than differences between the ECT scenarios due to the large changes in emissions between the 2007 Base Case and future year projections. For all of the ECT scenarios, changes due to biogenic emissions and dry deposition were relatively smaller than changes due to anthropogenic emissions. Maximum differences in hourly ozone concentrations between ECT D and ECT A, shown in Figure 3, ranged from -3.0 to 4.5 ppb in contrast to differences between the future scenarios and the 2007 base case which ranged from -14 to 22 ppb. SI Figure S-6 shows an example of the spatial differences in predicted ozone concentrations between ECT D and ECT A. ECT A assumes a typical urban sprawl pattern which concentrates growth in Travis and Williamson counties, whereas ECT D assumes new development in existing towns throughout the five-county Austin region resulting in decreases in ozone concentrations near the urban core and increases in the surrounding areas. Decreases in hourly ozone concentrations with ECT D were associated with high ozone levels, as described in SI Section S-8 and were primarily due to reductions in on-road mobile source NOx emissions in the urban core of Travis County. Overall differences in the magnitude of emissions produced greater changes in hourly ozone concentrations than differences in regional development patterns between the four scenarios. The results imply that although the effects of urbanization patterns are non-negligible, the pattern of urban development is not as significant as reductions in emissions

per capita and controlling the environmental impacts of urbanization involves multifaceted strategies.

Acknowledgments Although the research described in this article has been funded by the EPA (STAR Grant No. RD83183901), it has not been subjected to the Agency’s peer and policy review and therefore does not necessarily reflect the views of the Agency and no official endorsement should be inferred. We express our appreciation to Kara Kockelman and Shashank Gadda for conducting the travel demand modeling. We thank Christine Wiedinmyer and Yosuke Kimura for their insights.

Supporting Information Available Emission inventory development, the matrix of modeling simulations conducted as well as differences in emissions and ozone concentrations between the ECT scenarios and the 2007 Base Case is decribed. This material is available free of charge via the Internet at http://pubs.acs.org.

Literature Cited (1) Bartholomew, K. Integrating Land Use Issues into Transportation Planning: Scenario Planning. Summary report for the Federal Highway Administration under Cooperative Agreement No. DTFH61-03-H-00134, 2005. http://www.fhwa.dot.gov/planning/scenplan/practices.htm. (2) Envision Central Texas (ECT). http://www.envisioncentraltexas.org. (3) Wiedinmyer, C.; Guenther, A.; Estes, M.; Strange, I. W.; Yarwood, G.; Allen, D. T. A land use database and examples of biogenic isoprene emission estimates for the state of Texas, USA. Atmos. Environ. 2001, 35 (36), 6465–6477. (4) Lemp, J. D.; Zhou, B.; Kockelman, K.; Parmenter, B. Visioning Vs. Modeling: Analyzing the Land Use-Transportation Futures of Urban Regions. Proceedings of the 86th Annual Meeting of the Transportation Research Board, Washington DC, 2007. (5) Civerolo, K.; Hogrefe, C.; Lynn, B.; Rosenthal, J.; Ku, J. Y.; Solecki, W.; Cox, J.; Small, C.; Rosenzweig, C.; Goldberg, R.; Knowlton, K.; Kinney, P. Estimating the effects of increased urbanization on surface meteorology and ozone concentrations in the New York city metropolitan region. Atmos. Environ. 2007, 41 (9), 1803–1818. (6) Wang, X. M.; Lin, W. S.; Yang, L. M.; Deng, R. R.; Lin, H. A numerical study of influences of urban land-use changes on ozone distribution over the Pearl River Delta region, China. Tellus B 2007, 59 (3), 633–641. VOL. 42, NO. 19, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

7299

(7) ENVIRON International Corporation. Users Guide to the Comprehensive Air Quality Model with Extensions (CAMx) version 4.03. http://www.camx.com. (8) Yarwood, G.; Wilson, G.; Shepard, S.; Guenther, A. User’s Guide to the Global Biosphere Emissions and Interactions System Version 3.1. http://www.globeis.com. (9) Civerolo, K.; Sistla, G.; Rao, S. T.; Nowak, D. J. The effects of land use in meteorological modeling: implications for assessment of future air quality scenarios. Atmos. Environ. 2000, 34 (10), 1615– 1621. (10) Bell, M.; Ellis, H. Sensitivity analysis of tropospheric ozone to modified biogenic emissions for the Mid-Atlantic region. Atmos. Environ. 2004, 38 (13), 1879–1889. (11) Wiedinmyer, C.; Strange, I. W.; Estes, M.; Yarwood, G.; Allen, D. T. Biogenic hydrocarbon emission estimates for north central Texas. Atmos. Environ. 2000, 34 (20), 3419–3435. (12) Wesely, M. L. Parameterization of surface resistances to gaseous dry deposition in regional-scale numerical models. Atmos. Environ. 1989, 23 (6), 1293–1304. (13) Walmsley, J. L.; Wesely, M. L. Modification of coded parametrizations of surface resistances to gaseous dry deposition. Atmos. Environ. 1996, 30 (7), 1181–1188. (14) McDonald-Buller, E.; Wiedinmyer, C.; Kimura, Y.; Allen, D. Effects of land use data on dry deposition in a regional photochemical model for eastern Texas. J. Air Waste Manage. Assoc. 2001, 51 (8), 1211–1218. (15) Capital Area Planning Council (CAPCO). Photochemical modeling for Austin’s Early Action Compact: Analysis of emission control strategies for ozone precursors, 2004. http://www. tceq.state.tx.us/implementation/air/sip/nov2004eac.html.

7300

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 19, 2008

(16) Capital Area Planning Council (CAPCO). Austin-Round Rock MSA 1999 Ozone Precursor Emissions Inventory, 2004. http:// www.tceq.state.tx.us/implementation/air/sip/nov2004eac.html. (17) Capital Area Planning Council (CAPCO). Austin-Round Rock MSA 2007 Future Year Ozone Precursor Modeling Emissions Inventory, 2004. http://www.tceq.state.tx.us/implementation/ air/sip/nov2004eac.html. (18) User’s Guide to MOBILE6.1 and MOBILE6.2: Mobile Source Emission Factor Model, EPA420-R-03-010; U.S. Environmental Protection Agency: Washington, DC, 2003. (19) Envision Central Texas Transportation Model: Technical Documentation; Smart Mobility Inc.: Norwich, Vermont, 2003. (20) Capital Metropolitan Planning Organization (CAMPO). Mobility 2030Plan.June2005.http://www.campotexas.org/pdfs/AdoptedMobility2030Plan.pdf. (21) Accounting for the Tier 2 and Heavy-Duty 2005/2007 Requirements in MOBILE6, EPA420-R-01-057, November 2001; U. S. Environmental Protection Agency: Washington, DC, 2001. (22) Capital Area Planning Council (CAPCO). Austin/Round Rock MSA Emissions Reduction Strategies, 2004. http://www.capcog.org/CAPCOairquality/CAAP_Apps/App5-2AustinRR%20Emission%20Reduction%20Strategies.pdf. (23) Program Update: Reducing Air Pollution from Nonroad Engines, EPA420-F-03-011; U.S. Environmental Protection Agency: Washington, DC, 2003. (24) Song, J. Land Use Forecasting in Regional Air Quality Models. Ph. D. Dissertation, The University of Texas at Austin, 2007.

ES800645J