Date: 	2/1/10	

To: 	EPA-HQ-OAR-2005-0161

Subject:	Petroleum Indirect Impacts Analysis

I. Overview

Per the statutory requirements of Section 201 of the Energy Independence
and Security Act (EISA), EPA was required to establish a baseline based
on the “average lifecycle greenhouse gas emissions, as determined by
the Administrator, after notice and opportunity for comment, for
gasoline or diesel (whichever is being replaced by the renewable fuel)
sold or distributed as transportation fuel in 2005.” This baseline is
used to determine whether certain renewable fuel pathways are able to
meet the GHG reduction thresholds also mandated in Section 201 of EISA.

In the preamble to the proposed Renewable Fuel Standard (RFS2)
regulation, we acknowledged that it was possible that renewable fuels
may actually displace fuels which on the margin had higher lifecycle GHG
emissions than the average petroleum fuel baseline.  Furthermore, we
took comment “on whether-strictly for purposes of assessing the
benefits of the rule (and not for purposes of determining whether
certain renewable fuel pathways meet the GHG reduction thresholds set
forth in EISA), we should assess benefits based on a marginal
displacement approach and, if so, what assumptions we should use for the
marginal displacements.”  We also received numerous comments that
urged us to assess the indirect impacts of petroleum use.

   

For this final RFS2 rulemaking, we carried out an analysis of the
indirect impacts of petroleum, employing the Department of Energy’s
(DOE) Energy Technology Perspectives (ETP) model.  The ETP model is a
MARKAL-based model that was developed by the International Energy Agency
and was modified and updated by DOE.   It is a partial equilibrium model
that incorporates a representation of the physical energy system and
represents the flow of energy carriers through the physical
infrastructure from the resource base through the various energy
conversion technologies to the end user.  The MARKAL modeling framework
allows for a full “well-to-wheels” comparison.   

The main objective of our study was to understand the sources of
marginal crude oil that would be displaced by renewable fuels in 2022. 
This would give us insight into whether the lifecycle GHG emissions for
marginal crude in 2022 would differ significantly from the 2005 average
petroleum baseline established by this rule.

II. Model Background

The ETP model consists of fifteen world regions. These are broken out
into: the United States; Canada; Mexico; IEA Europe; Japan; South Korea;
Australia/New Zealand; Central and South America; Eastern Europe; Former
Soviet Union; Middle East; China; India; Other Developing Asia; and
Africa.  All major energy sources are covered, including coal, oil,
natural gas, nuclear power, and renewable energy. Energy conversion
processes include power generation and refinery operations and other
non-conventional processes such as coal-to-liquids and
biomass-to-liquids. Demand is represented through industrial,
commercial, residential, and transportation sectors. The model runs
through 2050 in five year increments.  For this study, ETP has been
benchmarked and updated using the Energy Information Administration’s
(EIA) Annual Energy Outlook (AEO).

III. Scenario Description

The purpose of this study was to isolate the impact of petroleum use on
total energy sector greenhouse gas emissions.  Thus, we created a
“demand boost case”, which was a domestic gasoline demand increase
of one million barrels per day.  The boost was phased in beginning in
2012 with annual 200,000 barrel per day increments over a five year
period. The increase in demand could be met by both foreign and domestic
refined products. 

IV. Assumptions and Limitations

There are a number of economic assumptions in the ETP model.  For this
scenario, the ETP projected demand-driven perturbations from the AEO oil
price path.  Thus, alternative oil price paths could affect the market
share and energy consumption results anticipated by the model.  

In addition, the global oil market is not modeled as fully competitive.
It is assumed that the Middle East members of OPEC have significant
market power and act as a dominant firm. In this representation, the
Middle East OPEC countries set world oil prices, and other countries and
regions produce up until the point where their marginal cost of
production equals the world oil price. 

In ETP, the oil price consists of five different components: finding
costs; production costs; transportation costs; oil company profits; and
government revenues. Government revenues in the Middle East are the main
lever used to set oil prices. A high oil price case is thus one where
Middle East producers restrict supply, which pushes prices higher and
allows them to extract a higher economic rent on each barrel of oil
produced. In ETP, this “rent seeking” is the variable used to
calibrate to a given oil price path. For the purposes of this study, the
Middle East rent seeking is set to the level that is implicit in the AEO
2009 oil price path, given the production cost of Middle East producers.
This level of rent seeking is held constant between scenarios.

The ETP model is driven by exogenously specified demands for energy
services (e.g. light-duty vehicle miles traveled, ton-miles of rail
freight, etc.). The U.S. petroleum demand base case was calibrated to
the AEO 2009 projections by adjusting these service demands to the
levels forecasted by the EIA.  Demands were derived from AEO trends for
the model years after the AEO forecast period ends.  

Overall oil resources are based on IEA (2008) ETP data for conventional
and unconventional resources. Unconventional resources include
ultra-deepwater crude, Canadian oil sands, Venezuelan extra-heavy crude,
and coal-to-liquids. Canadian oil sands production rates were
endogenously determined based on the price of other competing resources.
Within ETP, data are limited for other unconventional resources;
however, the completed scenario suggests that other unconventional
production would not play a large role in meeting the boosted demand.

V. Description of Results

In 2022, ETP predicted a 0.73 million barrels of oil equivalent (MMboe)
per day increase in conventional oil production and a 0.04 MMboe per day
increase in oil sands/bitumen production.   The increase in petroleum
production came mainly from the Middle East (0.46 MMboe per day, 59%),
with no other region accounting for more than 8% of the total increase
in production.  Oil sands/bitumen, where production was split between
Canada and Central/South America, accounted for less than 5% of the
total increase in production.  The ETP results show a considerably
different production distribution for the 2022 marginal crude mix versus
the 2005 average crude mix used for the final rule.

Comparison of 2022 marginal crude mix from ETP run to 2005 crude mix
used for RFS2 petroleum baseline calculation

VI. Discussion of Results

Using estimates of well-to-tank GHG emissions for crudes produced in
various countries, we calculated the well-to-tank GHG emissions for the
marginal crude mix.  As ETP did not have data for specific countries,
emissions for regions were estimated by taking the straight average
emissions rate for countries in that region.  For example, well-to-tank
values for Africa were estimated by taking the average of the values for
Nigeria, Angola, and Algeria.  However, there were some regions, such as
the “Former Soviet Union”, where well-to-tank estimates were not
available.  In these cases, the value for “Other Imports” was used
as a default value.  

Well-to-tank GHG emissions for domestic diesel production by crude
source

Country of Origin	Well-to-Tank Values

(kg CO2e/MMBtu LHV)

Canada Oil Sands	34.0

Venezuela Bitumen	30.8

Nigeria	29.7

Mexico	24.1

Angola	23.0

Kuwait	19.6

Iraq	18.7

Venezuela Conventional	18.6

Canada Conventional	18.0

Ecuador	17.8

Saudi Arabia	17.4

Other Imports	17.3

U.S.	13.5

Algeria	12.4



These calculations yielded a well-to-tank value for the marginal crude
mix of 19.0 kg/MMBtu.  This value was slightly higher than the
well-to-tank value used for the 2005 average petroleum baseline in the
final rule.  When combined with the tailpipe emissions, this yielded a
well-to-wheels value of 98.0 kg/MMBtu, about 0.6% larger than the
average petroleum baseline value.

One area we examined in more detail was the use of the “Other
Imports” value as a default.  Overall, the default value was used for
18% of the marginal crude mix.  To test the sensitivity of this value in
the overall calculations, we considered a range of default values for
conventional crude oil, from 12.4-29.7 kg/MMBtu.  Utilizing this range
of default values, we find that, for the marginal crude mix, this
yielded well-to-tank values from 18.1-21.2 kg/MMBtu and well-to-wheels
values from 97.1-100.2 kg/MMBtu.  The marginal well-to-wheels values
ranged from being 0.3% less to 2.9% greater than the average 2005
petroleum baseline value used in the final rule depending on the default
value used.

VII. Conclusions

We created a scenario with increased domestic demand for gasoline and
used DOE’s ETP model to analyze how the increased demand would be met
in 2022.  The model indicated that crude oil production would increase,
with close to 60% of the increase in production occurring in the Middle
East and less than 5% of the increase in production coming from oil
sands/bitumen.  Overall, well-to-wheels emissions for the marginal crude
oil were virtually the same as the average well-to-wheels emissions
calculated for the 2005 petroleum baseline. 

 The ETP model operates in energy units on a lower heating value basis.
The energy equivalent of 1 million barrels/day was obtained by using the
lower heating value energy content of conventional motor gasoline as
reported in EIA’s Annual Energy Review 2008 (approximately 4.90
MMBtu/barrel).  The 1 million barrels/day converts on an annual basis to
about 1.78 quads of gasoline.

 The marginal value is higher than the average because low-GHG domestic
crudes are replaced by Middle East crudes, and there is a slightly
higher amount of oil sands.

  Clean Air Act Section 211(o)(1)

 IEA, 2006. Energy Technology Perspectives 2006.  Organisation for
Economic Co-operation and Development (OECD)/IEA, Paris, France. 

 IEA, 2008. Energy Technology Perspectives 2008.  OECD/IEA, Paris,
France.

 Fishbone, L., Abilock, H., 1981. MARKAL, a linear-programming model for
energy systems analysis: Technical description of the BNL version.
Energy Research, 5, 353-375.

 Hamilton, L.D., Goldstein, G.A., Lee, J., et al., 1992. MARKAL-MACRO:
An overview. Report No. BNL-48377. Brookhaven National Laboratory,
Upton, NY.  

 Loulou, R., Goldstein, G., Noble, K., 2004. Documentation for the
MARKAL Family of Models.  Available at < HYPERLINK
"http://www.etsap.org" http://www.etsap.org > Accessed February 2009.

 EIA, 2009. An Updated Annual Energy Outlook 2009 Reference Case
Reflecting Provisions of the American Recovery and Reinvestment Act and
Recent Changes in the Economic Outlook. Report #: SR-OIAF/2009-03. 
Available at: < HYPERLINK
"http://www.eia.doe.gov/oiaf/servicerpt/stimulus/index.html"
http://www.eia.doe.gov/oiaf/servicerpt/stimulus/index.html > Accessed
December 2009.

 Department of Energy: National Energy Technology Laboratory. 2009.
NETL: Petroleum-Based Fuels Life Cycle Greenhouse Gas Analysis - 2005
Baseline Model. Available at <  HYPERLINK
"http://www.netl.doe.gov/energy-analyses" 
www.netl.doe.gov/energy-analyses > Accessed December 2009.

