Arvind P Ravikumar
The Role of Natural Gas in a
Carbon-Constrained
World
@arvindpawan1
Stanford University | The Payne Institute at Mines
2
World share of primary energy
BP Statistical Review of Energy (2017)
• Globally, natural gas is expected to grow in both OECD and non-OECD countries
• Increased supply (US, Australia) and demand (SE Asia, China, India, EU)
Natural Gas: Future Growth in All Regions
• Assists grid integration of large % renewables
• Reduce PM2.5 pollution → improved air quality →health outcomes
IEA World Energy Outlook 2017
Energy-related emissions of pollutants and CO2
IPCC 5th Assessment Report (2014)
NOAA Global Air Sampling Network
https://www.esrl.noaa.gov/gmd/aggi/aggi.html 3
• Methane – second most abundant GHG globally.
• 16% of global emissions, 11% of US GHG emissions (EPA GHGI 2016)
• Significantly higher global warming potential than CO2 (GWP: 36 IPCC AR5)
• Oil and Gas activity major industrial source of emissions
• Easier to tackle methane from O&G due to readily available path to market
Methane as a Greenhouse Gas
4
Expectation vs. Reality
> 55% believe US should use < 10 Tcf of
gas by 2050
0
20
40
60
80
100
120
140
160
0
10
20
30
40
50
60
2000 2010 2020 2030 2040 2050
Natural gas productiontrillion cubic feet
2017history projections
High Oil and Gas Resource and TechnologyHigh Oil PriceHigh Economic GrowthReferenceLow EconomicGrowthLow Oil PriceLow Oil and Gas Resource and Technology
billion cubic feet per day
But, EIA projects record high production >
40 Tcf by 2040
5
Climate Implications of Cheap Natural Gas
• Total U.S. natural gas production = 27 Tcf (2016)
• Production/consumption estimates through 2050 diametrically oppose to Paris targets
• Fossil NG emissions account for 60 – 85% of total C-budget
Ravikumar et al. 2018 (in preparation)
6
Overview of Natural Gas Work
2. Technology Assessments 4. Spillover Impacts
3. GHG Mitigation Policy1. System Modelling
• Modeling and empirical studies of emissions
detection technology
• Stanford/EDF Mobile Monitoring Challenge –
testing new leak detection platforms
• FEAST platform – Fugitive Emissions Abatement
Simulation Toolkit
• Life-cycle Assessments of LNG trade, emissions,
and economic impact
• Multi-year effort on measuring efficacy of
emissions mitigation policy (Alberta, Canada)
• Advanced mitigation frameworks (multi-tiered
approach, quantification)
• Distributional impact of methane emissions
on electricity and other sectors
• Incorporating methane into broader climate
policies (carbon tax, cap and trade, etc.)
‘Micro Analysis’ ‘Macro Strategies’
Methane Leakage – Superemitters
1 – Leakage is dominated by ‘super-emitters’
• ‘5 – 50’ rule: small number of leaks responsible for large fraction of emissions
Ravikumar et al. Environ. Res. Lett. 12 044023 (2017)7
Methane Leakage – Difficult to Find Leaks
2 – Emissions are difficult to detect and expensive
• Millions of potential sources (production, gas plants, pipelines, etc.)
• Leaks are (mostly) stochastic – wear & tear, operator error, equipment malfunction
→ little predictive capability
• Technologies to detect methane are slow/expensive (this is changing)
8
Methane Leakage – Prescriptive Policies
3 – Emissions mitigation policy is prescriptive
• Mandate specific actions instead of mitigation targets
• Leak detection and repair (LDAR) programs most common approach
• ‘Allowed’ technologies include infrared cameras and handheld sensors
• Push to incorporate innovation into policy (drones, planes, trucks, etc.)
0.4%0.1%0.85%
1.4% 9
Overview of Natural Gas Work
2. Technology Assessments 4. Spillover Impacts
3. GHG Mitigation Policy1. System Modelling
• Modeling and empirical studies of
emissions detection technology
• Stanford/EDF Mobile Monitoring
Challenge – testing new tech platforms
• FEAST platform – Fugitive Emissions
Abatement Simulation Toolkit
• Life-cycle Assessments of LNG trade,
emissions, and economic impact
• Measuring efficacy of emissions
mitigation policy (Alberta, Canada)
• Advanced mitigation frameworks
(multi-tiered approach, quantification)
• Distributional impact of methane
emissions on electricity
• Impact of climate policies on future
use of natural gas
‘Micro Analysis’ ‘Macro Strategies’
10
Technology Innovation
~45 min flying time
Revisit time ~ 1 week
‘Fast screening’
Fox et al. In review (2018) 11
Stanford/EDF Mobile Monitoring Challenge
• Stanford/EDF Mobile Monitoring Challenge (MMC)
• Platforms – drones, trucks, and plane-based systems
• 3 weeks of blind controlled release tests: April – May 2018
• 10 technologies participated
• METEC test site (CSU, Fort Collins) Visit: methane.stanford.edu
12
Example Technology Testing - Drone
13
Technology A (Drone)
Total number of leaks 63Number of zeros 41
Yes No TotalLeak 59 4 63No Leak 0 41 41
Total number of leaks 63Number detected 59Number location identified 50% location identified correctly 0.85
Yes NoLeak True + False -No Leak False + True -
0.94 0.060.00 1.00
Locational Accuracy
Leak identification (overall)
• Best in class performance (detection & quantification)
• Professionally managed operations (standard protocols)
• Real time data including quantification (initial estimate)
14
Technology A - Quantification
• Most leak estimates within 2x of actual leak rates
• (Quantification, in general, is very difficult. Within 2x is exceptional performance for
sensors that don’t directly measure flow rates)
15Ravikumar et al. In review (2018)
16
How can technology inform mitigation
policy?
FEAST
• Fugitive Emissions Abatement Simulation Toolkit
• Open-source
• Updated with new technologies, emissions data, policy scenarios
• Web-based version in development
• Tool simulates evolution of natural gas leaks over time
• Assess leak-detection technology
• Assess mitigation policy
• Assess long-term impact of business practices
17
Technology/Platform
Models
Methane Emissions
Data
• OGI, Method-21, Drones,
Planes, Trucks
• Economic data
• Activity/Inventory counts
• Published emissions data
(facility or component level)
Kemp et al. Environ. Sci. Technol. 50 4546 (2016)
FEAST Platform Capabilities
Policy Scenarios
Business Practices
1.Technology inter-comparison
studies
2.Long-term mitigation trends
(changes to baseline)
3.Cost-effectiveness of
alternative approaches
4.Trade-offs in survey
frequency vs. mitigation
18
• Markov process to ‘generate’ leaks
Emissions Mitigation = ‘Natural repair rate’ + Mitigation Policy
0.5 g/s = 5 tons per year
Various mitigation
scenarios
Random leak
generation
FEAST: Dynamic Simulation
19
• Existing policies require use of infrared cameras to detect leaks
• But regs do not specific how leak surveys should be designed
• Results in mitigation uncertainty
Effectiveness reduces
with increasing distance
Ravikumar et al. Environ. Sci. Technol. 51 718 (2016)
Analyze Policy Blind-Spots
20
FEAST Capabilities: Cost-Benefit Analysis
• EPA over-estimates cost of mitigation policy
Mitigation benefits
variable
Requires regional
focus in policy
Ravikumar et al. Environ. Res. Lett. 12 044023 (2017)
Uncertainty and Cost-effectiveness
21
Field Campaign To Assess Policy Effectiveness
• Campaign sponsored by consortium of Canadian O&G industry and regulators
• 50 x 50 km area NW of Calgary
• ~ 200 sites selected for leak detection
and repair surveys
• 3 different survey schedule (1, 2 or 3
times per year) and 1 control group
• Goals: Determine time evolution of
emissions mitigation – ‘sunset policy’
22
Bringing Policy Makers And Scientists Together
• Workshop to develop future mitigation frameworks for regulators
• Invite-only workshop with academics, regulatory agencies, and industry
• Environment Canada, Alberta Energy Regulator
• U.S. EPA, Colorado DPHE, California ARB
23
24
Why does this matter?
Impact on Broader Energy Systems
Barnett
Fayetteville
Upper
Green
River
Uintah
San Juan
25
Major Shale Plays and Emissions Studies
Peischl (2015)
Ren (2017)
Omara (2016)
Caulton (2014)
Peischl (2015)
Ren (2017)
Kort (2016)
Schneising (2014)
Petron (2012, 2014)
Brantley (2014)
Robertson (2017)
Smith (2017)
Karion (2013)
Robertson (2017)
Araiza (2015)
Lyon (2015)
Lan (2015)
Karion (2015)
Brantley (2014)
Lavoie (2017)
Schneising (2014)
Roest (2016)
Peischl (2015)
Peischl (2015)
Robertson (2017)
Schweitzke (2017)
• Large variation among basins
• Estimates have high uncertainty – including single point estimates
M. Omara (2016), X. Ren (2017), J. Peischl (2015), D. Caulton (2014), G. Roest (2016), O.
Schneising (2014), A. Robertson (2017), S. Schwietzke (2017), J. Peischl (2016), X. Lan
(2015), D. Lyon (2015), D. Zavala-Araiza (2015), A. Karion (2015), H. Brantley (2014), M.
Smith (2017), EPA (2016)
Production Normalized Leakage Rates
Large variation → distributional
impacts critical to understanding
benefits of natural gas use
26
• Avg. Emissions intensity: 430 g CO2/kWh, about 50% lower than coal
• Varies from 391 g CO2/kWh (GA) to 588 g CO2/kWh (IN)
GHG Emissions from Gas – No Leakage
Single cycle
natural gas
plants
Ravikumar et al. In review (2018)27
• Avg. Emissions intensity ↑ from 430 g CO2/kWh to 542 g CO2e/kWh
• Western states (WA, OR, CA, AZ, NV) typically do worse due to originating gas basins
GHG Emissions from Natural Gas – 100 yr GWP
28
A Cautionary Tale
Are we repeating mistakes of the past?
29