1

Garvin Heath, PhD A&WMA / AIChE Web Forum

May 29, 2013

Natural Gas and the Transformation of the U.S. Energy Sector: Electricity

Life Cycle GHG Emissions:

Original Research and Meta-Analysis of Previous Results

2

Outline

• About NREL and JISEA • Background: Study Goals • JISEA contributions

– Harmonization of previously published estimates for conventional and unconventional gas

– Original analysis for Barnet Shale

3

NREL Snapshot

• Leading clean-energy innovation for 35 years

• ~2000 employees with world-class facilities

• Campus is a living model of sustainable energy

• Owned by the Department of Energy

• Operated by the Alliance for Sustainable Energy

Only US national laboratory dedicated solely to energy efficiency and renewable energy

4

About JISEA

JISEA seeks to guide the transformation of the global energy economy through comprehensive, transdisciplinary research focused on the nexus of energy, finance, and society.

JISEA Founding Institutions

NREL

University of Colorado-Boulder

Colorado School of Mines

Colorado State University

Massachusetts Institute of Technology

Stanford University

5

Background: Study Goals and Sponsors

• JISEA designed this study to address four related topics : – Life cycle greenhouse gas (GHG) emissions associated with

shale gas compared to conventional natural gas and other fuels used to generate electricity

– Legal and regulatory frameworks governing unconventional gas development at federal, state, and local levels, and how they are changing in response to industry growth and public concerns

– Water-related practices of natural gas production companies – Electric power futures – Forecasts of demand for natural gas –

and other generating options – in the electric sector under a variety of policy and technology developments over the next 20 to 40 years

• Study sponsors: Multi-client sponsor group composed of natural gas producers, utilities, transmission companies, investors, researchers, and environmental NGO

6

Key Finding: GHG Emissions

• Life cycle greenhouse gas emissions associated with electricity generated from Barnett Shale gas extracted in 2009 were found to be very similar to conventional natural gas and less than half those of coal-fired electricity generation. – Comparison based on methodological harmonization of 200

previously published estimates for conventional gas and coal electricity generation

• Finding based on methods and data sources independent of previous estimates.

7

NG in Context (pre-2011 published estimates)

IPCC SRREN SPM Fig 8

conventional

8

EPA NG Upstream Emissions Estimates

Source: Larsen, 2013

US GHG Inv. 2011 vs. 2010

>2x production segment emissions

+10x liquids unloading EF (conv)

EFs for unconv.: • Completions • Workovers

Centrifugal

compressors

9

Reduction in Liquids Unloading Largest Change from 2012 to 2013 Inventory

Source: Larsen, 2013

10

Shale Gas LCAs Emerging…

• Howarth et al 2011 • Headline: NG worse than coal

• Jiang et al. 2011 • Marcellus < domestic gas (conv + unconv), both < coal

• Skone et al. 2011 (since updated, adding Marcellus)

• Barnett Shale > conventional (except onshore), both < coal • Hultman et al. 2011

• Unconv. < conv., both < coal • Stephenson et al. 2011

• Shale > conventional, both < coal • Burnham et al. 2012

• Conv. > shale, both < coal • Lorenzi and Jersey, 2013

• ExxonMobil data Marcellus shale gas < coal • Others: Tyndall Centre, Deutsch Bank, NYDEC,

Weber and Clavin

11

Challenges for the Field and JISEA Contributions

Challenges 1. Large number of LCAs that

could be leveraged, but almost all need updating to reflect latest evidence

2. LCAs not strictly comparable given differences in scope and assumptions which hampers collective conclusions

3. Unconventional gas LCAs rely on very limited set of data sources which are old and have limited sample sizes

Contributions

Harmonization

New research using alternative data set

12

Importance of Harmonization

Shale Conventional Diesel Coal Underground Surface High Low Low High

Gra

m C

arbo

n pe

r MJ

20-year time horizon Methane Indirect CO2 Direct CO2

Source: Howarth et al. 2011

Also: - Include liquids unloading in High case, not low - Not consider workovers

13

LCA Harmonization Project

Six articles in Special Issue of Journal of Industrial Ecology on Meta-Analysis of LCA (May, 2012)

– Coal – Nuclear – Wind – Concentrating solar power – Crystalline silicon PV – Thin-film PV

• Forthcoming: Conventional Natural Gas

www.nrel.gov/harmonization

14

Harmonization Steps

Harmonization Step Conventional Unconventional

System Harmonization GWP to 100 yr IPCC X X Functional unit to kWh generated (not MJ or kWh delivered) X X Include power plant upstream/downstream X X Include embodied GHG in well pre-production X X Include liquids unloading X X Include recompletions and at NSPS frequency X

Technical Harmonization Thermal efficiency (51% NGCC) X X Power plant lifetime and capacity factor X NG heating value X

15

Preliminary Harmonization Results

Sources: Figure from JISEA, 2013. Conventional results – O’Donoughue, P.; Heath, G.; Dolan, S. Vorum M.; (2012). “Life Cycle Greenhouse Gas Emissions from Natural Gas–Fired Electricity Generation: Systematic Review and Harmonization.” Journal of Industrial Ecology (accepted); Unconventional results – Heath. G.; O’Donoughue, P.; Arent, D.; Bazilian, M. (submitted). Coal: Whitaker et al. 2012.

Subject to change during publication process

EUR = estimated ultimate recovery NGCC = natural gas combined cycle

16

Original Estimate of Life Cycle GHG Emissions from Barnett Shale Gas

17

Data Sources for Barnett Shale Estimate

Leveraged high-resolution, Texas Commission on Environmental Quality inventories of volatile organic compounds from NG production and processing stages

– 16,000+ sources – Very high industry participation rates – Year 2009

18

GHG Sources from Barnett Shale TCEQ Inventories

• In the Barnett Shale area, many GHG sources are potentially controllable

aPneumatics, from the area source inventory, have no count of individual sources

19

Methods for Barnett Shale Estimate – Original vs. Literature Sources

20

Methods Step 1: VOC to GHG Inventory

Key (Extended) Gas Composition

Translates VOC (non-methane) to methane plus CO2 by ratio Compiled a quality-screened data set of county-level, extended gas composition analyses of produced (raw) gas

– Demonstrates wide variability of methane and VOC content within the Barnett Shale play.

21

Barnett Shale Gas Composition Variability

22

Barnett Shale Gas Composition Variability

23

Methods Step 1: VOC to GHG Inventory

For original estimates: • Venting/fugitives: Methane and CO2 emissions by

ratio of inventory VOC and county-average composition analysis

• Combustion sources: Methane and CO2 emissions by emission factor using inventory activity data

For estimates taken from literature: • Sources: EPA GHG inventory and other LCAs • Sensitivity analysis on estimated ultimate recovery

of wells and emissions from liquids unloading, well completion and well workover

24

Methods for Barnett Shale Estimate – System Boundary and Co-Product Allocation

25

Co-Product Allocation – GHG Emissions

26

Co-Product Allocation – Effect

• Overall, 1% of the estimated GHG emissions are allocated to condensate instead of natural gas.

• For comparison, Stephenson (2011) report 16% co-product allocation to condensate, ethane, LPG based on N. America average gas composition

• This study omits about 25% of the individual sources reported in Special and Point Source inventories, which collectively represent 60% of the total reported VOC emissions, because they are associated only with the production and processing of the co-products of crude oil and condensate.

27

Methods Step 2: From Inventory to LCA

• TE = thermal efficiency • EFx = emission factor for all processes in basin (after application of

adjustment factors, the sum of estimated emissions for sources in each stage is divided by the production volume of gas associated with those emissions.

• Lx = a loss factor representing the portion of gas extracted that remains in the product flow to be used as an input for combustion, reflecting process-chain losses inclusive of this life cycle stage onward (kWh-equivalent extracted/kWh-equivalent input)

28

GHG Emissions from Production and Processing

• Most emissions from CO2

• Much of emissions is potentially controllable, including combustion (e.g., higher efficiency)

29

Effect of Alternative Gas Composition – Production and Processing

Denton Countya Johnson Countya

Tarrant Countya

Wise Countya 22-County Total

Barnett Shale average vs. main results 12% -5% -33% 29% 1%

National average vs. main results 15% -11% -36% 29% -3%

Denton Countya Johnson

Countya Tarrant Countya

Wise Countya

Barnett Shale play averageb

National averagec

Volatile organic compounds contentd 18% 19% 6% 23% 16% 18%

CO2 contentd 2% 2% 1% 3% 2% 2%

Methane contentd 63% 63% 80% 56% 66% 78% a Only the four top-producing counties in the Barnett Shale play are shown. b Production-weighted average across the 22 counties of the Barnett Shale play c As reported in EPA (2011) d Percentage by mass

These results have implications for developing more accurate GHG emission inventories at sub-national levels and any regulatory system that might seek to identify high emitters within plays.

30

Barnett Shale Life Cycle GHG Emissions by Life Cycle Phase and GHG

• Most life cycle GHG emissions attributable to generation of electricity from combustion at power plant

• Fuel cycle emissions of both CO2 and methane are important f Multiple estimates, in parentheses, pertain to high EUR (4.5), base case EUR (1.42), and low EUR (0.4), respectively. Single estimates pertain to stages without sensitivity to EUR. The error bar is plus or minus the total bar length (life cycle GHG emissions).

31

Loss and Leakage of Produced Gas

• 7.1% fuel cycle loss rate (incl consumed) NG volume per volume of NG produced (at average EUR = 1.42 bcf). – 1.5% NG loss per NG produced – 1.3% methane emitted per NG produced

• EUR sensitivity: – low EUR (0.45 bcf) 2.8% methane leakage rate; 8.9% loss, – high EUR (4.3 bcf) 0.8% leakage and 6.5% losses.

Completions and

Workoversb Production Processing Transmissionc Total Extracted from Ground 100.0% 100.0% Fugitive Losses – 0.1% 0.0% 0.5% 0.6%

Potentially Controllable Leakage 0.8% 0.1% 0.0% 0.0% 0.9%

Combusted as Fuel – 0.9% 3.9% 0.8% 5.6% Delivered to Power Plant 92.9% a reported as volume of natural gas consumed or lost per volume of natural gas produced. b See footnote to Figure 9 c From Skone et al. (2011)

32

Sources: Conventional – O’Donoughue, P.; Heath, G.; Dolan, S. Vorum M.; (2012). “Life Cycle Greenhouse Gas Emissions from Natural Gas–Fired Electricity Generation: Systematic Review and Harmonization.” Journal of Industrial Ecology (accepted); Unconventional – Heath. G.; O’Donoughue, P.; Arent, D.; Bazilian, M. (In Preparation)

Barnett Shale Gas Estimate Compared to Harmonized Literature Estimates for Coal, Conventional and Shale Gas

Harmonized results subject to change during publication process

EUR = estimated ultimate recovery NGCC = natural gas combined cycle

33

Key Finding: GHG Emissions

• Life cycle greenhouse gas emissions associated with electricity generated from Barnett Shale gas extracted in 2009 were found to be very similar to conventional natural gas and less than half those of coal-fired electricity generation. – Comparison based on methodological harmonization of 200

previously published estimates for conventional gas and coal electricity generation

• Finding based on methods and data sources independent of previous estimates.

34 34

Learn More

Download the full report at

nrel.gov/docs/fy13osti/55538.pdf

Learn more about JISEA’s natural gas work at JISEA.org/Natural_Gas.cfm