Hybrid InputOutput

1
Hybrid Input-Output Lifecycle Models
Heather L. MacLean University of
Toronto hmaclean_at_ecf.utoronto.ca
Oil Transition Workshop April 20, 2006
2
Outline

• Motivation for life cycle perspective in
  informing oil transition
• Life cycle approaches
• Conventional process model based
• Economic Input-Output (EIO) based
• U.S., Canadian, Binational
• Hybrid life cycle approaches
• EIO model limitations/advantages
• Future application of hybrid model to oil sands

3
Motivation

• Life cycle perspective required for informed
  decision-making for oil transition
• Multi-attribute nature of decision problem
• Requires broader analysis than conventional life
  cycle assessment
• Focus on energy use, greenhouse gas air
  pollutant emissions
• Fuel cycle and conversion system
• Limited set of supply chain activities
• Strength is analysis of existing systems

4
Analysis (optimally) should include

• Three components of sustainability
• Environmental (global, regional, local)
• Economic
• Social
• Also,
• Technological change
• Dynamic analysis
• Expanded boundary (activities, international)
• Explicit uncertainty analysis
• Transparency

5
Life Cycle Assessment (LCA)

• Conventional process-based (SETAC/EPA)
• Economic Input-Output based (EIO-LCA)
• Hybrid approaches

6
Circularity effects in economy must be accounted
for

• Cars are made from steel, steel is made with iron
  ore, coal, steel machinery, etc. Iron ore and
  coal are mined using steel machinery, energy,
  etc...

product
emissions
7
Economic Input-Output (EIO) LCA
IO tables quantify interrelationships between
industrial sectors in an economy (Leontief,
1941) Environmental IO extends IO analysis to
account for direct and indirect environmental
effects in an economy
100 Million Demand for Motor Vehicles
8
U.S. Economic Input-Output Life Cycle Assessment
(EIO-LCA) model

• Originally developed by Carnegie Mellon
  Universitys Green Design Institute (1995)
• 1 million users to date
• Standard augmented SNA framework
• 1997 US Benchmark Input-Output Table
• (491 industry sectors)
• Coupled to wide array of environmental data
• Energy, fuel use, greenhouse gases, air
  pollutants, toxic releases, social costs, OSHA
• Available at www.eiolca.net

9
Canadian EIO model

• Developed by University of Toronto
• (Bjorn et al., 2004)
• Augmented SNA framework
• 1997 Canadian IO table (117 sectors)
• Energy and fuel use, greenhouse gases, toxic
  releases (Statistics Canada)
• Next steps water use, air pollutants
• Both US and Canadian models have competitive
  import assumption

10
Comparing US and Canada
Significant international differences in fuel use
Agriculture
Mining, Oil Gas
Manufacturing
Canada
U.S.
Canada
U.S.
Canada
U.S.
Propane
Petroleum Products
Natural Gas
Coal
Electricity
11
Binational U.S./Canada EIO model

• Based on a Chenery-Moses Multiregional IO
  Framework (Chenery Clark, 1959 Moses, 1955)
• 76 sectors
• Separates out trade in the U.S./Canada IO tables
• Final demand domestic demand exports
• Economic output domestic output imports

12
US energy pull on Canada (top sectors)
Energy Use Resulting from 1M U.S. Final Demand
Petroleum mfg.
TJ/US Million of U.S. Final Demand
Oil gas extraction
13
Canadian energy pull on US (top sectors)
Energy Use Resulting from 1M Canadian Final
Demand
Pesticides, fertilizer mfg
Chemical mfg.
TJ/US Million of Cdn. Final Demand
14
Energy use final demand 25,000 motor vehicles
sector in Canada
15
EIO-LCA

• Limitations
• Simplified linear, static view
• Aggregation issues
• Cradle-to-gate
• Not sufficient for regional/local implications
• Monetary vs physical units basis
• Strengths
• Intersectoral detail
• System boundary is entire economy (US/Canada)
• Quick, transparent often needed for
  real-world policy (publicly available data)
• Good first-round estimation for a snapshot in time

16
Potential of hybrid LCA models

• Incorporate advantages of SETAC/EPA and EIO
  models, lessen disadvantages
• Include detailed, process-level data, as well as
  the economy-wide effects
• What questions could the model address?
• Effects of choice of resources/ conversion
  processes on environmental impacts
• Impacts on economic variables of a transition
  from conventional oil

17
Albertas oil sands
18
Oil sands prod. 2004 1 million bpd
WCSB Western Canada Sedimentary Basin
Source CAPP 2005
19
Schematic diagram of oil sands hybrid LCA
Process assoc. emissions
Regional emissions factors
National emissions/ econ. activity
Canada/US EIO model
Region and sector specific process model
Oil sands extraction and upgrading
Oil gas extraction
Natural gas
Bitumen
Refined products
Machinery manufacture
Construction
Non-res. construction
Electricity
Labor
20
Acknowledgments

• Auto-21 Network Centre of Excellence
• General Motors
• NSERC
• Jon Norman, Ph.D. student
• Thomas Ferguson, MASc. student
• Shiva Habibi, postdoctoral fellow
• Prof. Scott Matthews, Carnegie Mellon U.

21
Questions? Comments.
22
(No Transcript)
23
Linking Methodology

• Captures Biregional Feedbacks
• Demand Canada ?
• Production Canada
• Demand U.S. Imports ?Production U.S.
• ? Demand Can Imports ? Production Canada
• .etc.

24
Integration of Conventional LCA data into EIO-LCA
EIO-LCA
25
Integration of EIO-LCA data into conventional LCA
Conventional LCA
EIO-LCA
26
Economic Input-Output (IO) analysis

• Uses IO tables to quantify interrelationships
  between industrial sectors in an economy
  (Leontief, 1941)
• Systems of National Accounts (SNA)

Outputs
Final Demand Y
Interindustry Demand Z
Inputs
Value Added
Total Output, Xi Zij Yi
27
Environmental IO analysis

• Extends IO analysis to account for direct and
  indirect environmental effects in an economy
• Environmental effects are considered in a
  satellite account of impact coefficients
  (pollutant / 1 sector output)

Chg. in Final Demand
Leontief Inverse
Env. Coefficients
Total Impacts

P

E
k impacts

X
(I-A)-1
Y
n sectors
X
Aij Zij / Xj
?X Change in economic output by sector