On-road transportation accounts for approximately 20% of United States energy consumption. Associated tailpipe emissions alone account for 40−60% of ground-level ozone (O₃) precursors, 6% of fine particulate matter (PM_2.5), and 22% of greenhouse gases (GHGs) emitted. Upstream processes involved in fuel production also contribute to overall environmental impacts. Life cycle assessment (LCA) has been used extensively to quantify the combined effects of fuel production and use, but descriptions of where and when emissions occur are typically not reported in life cycle inventories (LCI). Such information is generally not relevant for long-lived GHGs or for fossil fuel depletion, which together have received overwhelming attention among extant LCAs of transportation fuels. For many non-GHG pollutants, knowledge of spatial and temporal aspects of emissions is critical for understanding life cycle impacts; such information has been identified as a priority for inclusion in future analyses.

...Here we add process-specific spatial and temporal information to an existing attributional life cycle inventory (LCI) so as to reveal patterns in the geographic distribution and intra-annual timing of emissions. We focus on transportation fuels in the US and analyze three fuels pathways: gasoline, ethanol from corn grain, and cellulosic ethanol from corn stover. One goal of our work is to set the stage for future air quality modeling in the preparation of advanced life cycle impact assessments (LCIA). For example, our approach uses existing chemical speciation factors to describe pollutant emissions by chemical group. Another goal is to explore effects of model spatial resolution on the apparent distribution between urban and rural emissions.

—Tessum et al.

The team built on the Greenhouse Gases, Regulated Emissions, and Energy Use in Transportation (GREET) model, version 1.8d1, from Argonne National Laboratory. GREET models five groups of air pollutant emissions: NO_x; non-methane volatile organic compounds (VOCs); primary particulate matter less than 2.5 and 10 micrometers in diameter (PM_2.5 and PM₁₀); and sulfur oxides (SO_x). The researchers included a sixth pollutant, ammonia (NH₃).

They focused on adding to an existing LCI rather than refining the existing data and used default GREET settings with the following exceptions: they assumed that (1) corn ethanol plants use 100% natural gas process heat; (2) the ethanol produced is 100% ethanol without denaturant; (3) gasoline production is 100% conventional (i.e., not reformulated) gasoline; and (4) crude oil production is 100% conventional crude (most oil sands production occurs outside of our spatial modeling domain—i.e., the contiguous US). Vehicle energy-efficiency and emissions are the same for all fuels, except SO_x emissions, which are lower for ethanol vehicles. Efficiencies and emissions factors reported are for year 2010. Emissions are represented in 12 km resolution grids.

They excluded emissions occurring outside of the contiguous US and surrounding waters) from the results below. Excluding international emissions had the largest effect on the gasoline life cycle—96% of emissions from transportation of crude oil by ocean tanker occur outside their spatial modeling domain.

Among their findings were:

• In general, gasoline emissions tend to be correlated with vehicle use and so are distributed in or near urban centers.

• Ethanol emissions tend to be correlated with ethanol production and so are concentrated in the Midwest “Corn Belt”.

• The area along the Kentucky/Virginia border extending into West Virginia experiences a reduction in PM_2.5 emissions owing to reduced coal mining activity caused by excess electricity generation in biorefineries which is sold to the electrical grid and assumed to offset electricity produced elsewhere.

• For all three fuels, the greatest intensity of emissions per land area occurs in the Northeast for VOCs, owing to the large portion of total vehicle miles traveled per land area occurring there.

• The Midwest receives a large amount of emissions for both ethanol fuels, owing to ethanol fermentation plants and ammonia emissions from fertilizer nitrification.

• For the Midwest, emissions are lower for gasoline than for ethanol, with the exception that SO_x emissions are negative (i.e., reduced) for stover cellulosic ethanol owing to excess electricity generation at fermentation plants.

• SO_x emissions in the Southeast for corn ethanol are mainly attributable to Florida-based sulfuric acid production for phosphate fertilizer.

• The Southwest and West regions generally do not receive large proportions of pollutant emissions for any of the three fuels (exception: SO_x emissions for gasoline).

• Pollutant emissions from the gasoline life cycle do not vary appreciably by month. For the corn ethanol life cycle, however, there is a spike in NH₃ and NO_x emissions in the spring.

• All fuels show a slight decrease in emissions on weekends, with weekday emissions commonly showing a bimodal distribution around the morning and evening rush hours. For fuels and pollutants where farming activities are a major contributor, however, the weekday and weekend emissions are unimodally distributed around the daylight farming hours.

• Stover ethanol emits the lowest total amount of PM_2.5 of all three fuels, but emits more black carbon (15 mg mi⁻¹) than either of the other two fuels (gasoline: 4.5 mg mi⁻¹, corn ethanol: 10). Stover cellulosic ethanol also emits the lowest amount of sulfate aerosols (−1.3 mg mi⁻¹; gasoline: 1.9, corn ethanol: 3.8), which cause atmospheric cooling.

  Emissions of ethanol are 30,000−40,000 times higher for the ethanol fuels than for gasoline; ethanol in the atmosphere may be oxidized to form acetaldehyde (a carcinogen). However, emissions of benzene (another carcinogen) are higher for gasoline than for the ethanol fuels (the relative amounts by fuel depend on the ethanol feedstock and the blend level of the final fuel).

We have focused here on the air pollutant implications of the choice between ethanol and gasoline as a transportation fuel. In general, methods presented here can provide insight into any spatially or temporally inhomogeneous environmental impact categories, such as water quality and availability, soil properties, or wildlife habitats. They can also be expanded to study specific processes that affect those impact categories, such as agriculture and food production, building construction, or electricity generation; the possible applications are only limited by the availability of data.

—Tessum et al.

Resources

• Christopher W. Tessum, Julian D. Marshall, and Jason D. Hill (2012) A Spatially and Temporally Explicit Life Cycle Inventory of Air Pollutants from Gasoline and Ethanol in the United States. Environmental Science & Technology doi: 10.1021/es3010514