P 14 -7201 Atmospheric Measurement of Regional Methane

P 14 -7201 Atmospheric Measurement of Regional Methane Emissions Kenneth J. Davis 1, Thomas Lauvaux 1 and Colm Sweeney 2 1 The Pennsylvania State University, 2 NOAA ESRL/U. Colorado Current Developments and Impacts of Natural Gas in Transportation Research Board 93 rd Annual Meeting Washington, D. C. , 12 January, 2014

Outline • Global context • Need for “regional” emissions measurements • Overview of atmospheric methods – Local, regional, global • Regional emissions estimates – – Aircraft based Tower based “top-down” vs. “bottom-up” studies Role of satellite sensors • Conclusions

Global context • Human activity is driving the greenhouse gas (GHG) content of the atmosphere far beyond anything seen for at least 600, 000 years.

IPCC, Third Assessment Report, 2001 Ambient and ice core observations: Ultimate “top down” GHG assessment Last 2000 years

Global context • Human activity is driving the greenhouse gas (GHG) content of the atmosphere far beyond anything seen for at least 600, 000 years. • Large reductions in GHG emissions will be needed to stabilize climate, even at 2 K global warming (current <1 K). – What level of mobile source GHG emissions is tolerable?

B 1 scenario limits global warming to ~2 K by 2100 IPCC Assessment Report 4, 2007

IPCC Assessment Report 4, 2007 B 1 scenario requires large emissions reductions – peak only 25% above 2000 levels

Need for “regional” measurements of GHG emissions • Emissions mitigation (e. g. reduction of emissions by adoption of new natural gas – based technologies) will happen at “regional” scales. • Validation of emissions mitigation (e. g. current debate over methane leakage rates) will require independent measurements • Atmospheric GHG measurements have the potential to provide such independent emissions estimates. “Regional” = counties to continents

Atmospheric methods: A very brief summary

Summary of Atmospheric Emissions Measurement Methods: Gaps between Chamber/Turbulent Flux and Inversion Methods Eddy covariance or Plume dispersion year month day Chamber flux Time Scale Atmospheric Inversions Regional Global GAP Airborne flux hour 2 (1 m) -4 = 10 ha 2 (1 km) 2 = 10 ha 2 (10 km) 4 = 10 ha 2 (100 km) 6 = 10 ha 2 (1000 km) Rearth 8 = 10 ha Spatial Scale Bridging the gap between atmospheric inversions and turbulent flux measurements is my research expertise. 6

“Handshakes” needed • Combine regional estimates to match global budget • Combine chamber and/or turbulent flux estimates to match regional budget (e. g. CH 4 bottom up / top down comparison) • Must consider not just total emissions, but also emissions source (e. g. agriculture, gas production, wetlands, gas seeps)

Methodological clarification • Turbulent flux measurement – Eddy covariance or plume dispersion – Uses observations of the dispersion of a gas within the Atmospheric Boundary Layer (ABL) to infer source. – Typically used at 1 km 2 domain or smaller • Regional atmospheric budget/inversion – Uses changes in gas concentration over space and time in a well mixed ABL to infer sources. – More methodologically challenging – Suitable for “counties to continents”

Regional atmospheric budgets: a part of the needed toolbox of methods • Aircraft budgets – Excellent spatial coverage – Limited temporal coverage • Tower (or satellite) based atmospheric inversions – Excellent temporal coverage – Spatial coverage (domain, resolution) limited by density of long-term measurement network

Regional atmospheric measurements of CH 4 (and CO 2) sources and sinks

Regions of NOAA aircraft emissions estimates for natural gas production Denver Julesburg, CO Petron et al. 2012 (4%) Petron et al. submitted Marcellus, PA/NY Ongoing work Uinta, UT Karion et al. 2013 (8. 9%) Fayetteville, OK Peischl et al. in pr % values are estimated leakage rates as a fraction of production Barnett, TX Karion et al. in prep Haynesville, LA/TX Peischl et al. in prep

Penn State regional tower-based measurement campaigns N. Marcellus. 2014 -2016 Deployment stage INFLUX, 2010 -201? Pubs in prep Midcontinent intensive, 2007 -2009. Richardson et al (2012) Miles et al (2012) Lauvaux et al (2012 a, b) Schuh et al (2013) Lauvaux and Davis, 2013 Diaz et al, in review N. American tower GHG network circa 2008 Gulf coast intensive(? ), 2014 -2016. Funds pending.

Aircraft Mass Balance Method Downwind CH 4 Wind emissions CH 4 flux mixing height (PBL) Background CH 4 Molar CH 4 enhancement in PBL Perpendicular wind speed References: White et al. , 1976; Ryerson et al. , 2001; Mays et al. , 2009

June 1, 2011 Flight path INFLUX, Purdue/Shepson group Cambaliza et al, in review

Vertical structure of the atmospheric boundary layer (ABL) Vertical Profiles of Potential Temperature and H 2 O (~ 1: 00 to 1: 30 p. m. EDT) 6 June, 2012 Stably stratified “free troposphere” ABL top, entrainment fluxes Turbulent ABL Heat, water fluxes INFLUX, Purdue/Shepson group

Vertical Profiles of CO 2 and CH 4 (~ 1: 00 to 1: 30 p. m. EDT), 6 June, 2012 INFLUX, Purdue/Shepson group

June 1, 2011 Results 22, 000 moles s-1 INFLUX, Purdue/Shepson group 203 moles s-1 Cambaliza et al, in review

Methane Flux Matrix (June 1, 2011) City CH 4 Flux: 79. 9 mol s-1 Landfill Flux: 38. 3 mol s-1 INFLUX, Purdue/Shepson group

CH 4 Emission Flux from Indianapolis and contributions from Southside Landfill (SSLF) and Wastewater Treatment Plant (WWTP) INFLUX, Purdue/Shepson group

Utah, 2012 downwind CH 4 (ppb) HRDL upwind Distance perpendicular to wind (km) NOAA/Sweeney group Karion et al. 2013

NOAA/Sweeney group Uncertainty Parameter Mean Value Variability (1 s) Relative Uncertainty wind speed (V) 5. 2 m/s 1. 2 m/s 24% wind direction 55. 2° 10. 1° Vcosq 3. 8 m/s 0. 7 m/s 24% DXCH 4 56. 3 ppb 5. 6 ppb 10% BL depth 1700 m 125 m 7% CH 4 Flux 56 tonnes/hr 15 tonnes/hr 28% 8. 9% of production is leaking in Uintah Basin

Trace gases for attribution • Ethane, propane, butane – associated with natural gas sources, but not agriculture or wetland sources • 13 CH 4 – different sources have different isotopic ratios • CO – associated with combustion NOAA/Sweeney group

Conclusions: Airborne budgets • Powerful regional "snapshots" of total emissions - moderate levels of uncertainty. • CH 4 source types can often be disaggregated with trace gases or source location data. • Estimates to date suggest bottom-up methods underestimate total CH 4 emissions. Why? Not clear at this point. • Temporal variability difficult to capture. But if you sample a large enough area, perhaps statistics are in your favor. • Little info on spatial distribution of emissions within the "box. "

Tower-based “atmospheric inversion” CO 2 (and CH 4) source and sink estimates

Penn State regional tower-based measurement campaigns N. Marcellus. 2014 -2016 Deployment stage INFLUX, 2010 -201? Pubs in prep Midcontinent intensive, 2007 -2009. Richardson et al (2012) Miles et al (2012) Lauvaux et al (2012 a, b) Schuh et al (2013) Lauvaux and Davis, 2013 Diaz et al, in review N. American tower GHG network circa 2008 Gulf coast intensive(? ), 2014 -2016. Funds pending.

INFLUX objectives • Develop improved methods for determination of urban area-wide, and spatially and temporally-resolved (e. g. monthly, 1 km 2 resolution) fluxes of greenhouse gases, specifically, CO 2 and CH 4. • Determine and minimize the uncertainty in the emissions estimate methods.

Observational system • 12 surface towers measuring CO 2 mixing ratios, 5 with CH 4, and 5 with CO. (Penn State) • 4 eddy-flux towers from natural to dense urban landscapes. (Penn State) • 5 automated flask samplers. (NOAA/CU) • Periodic aircraft flights (~monthly) with CO 2, CH 4, and flask samples. (Purdue / NOAA) • Periodic automobile surveys of CO 2 and CH 4. (Purdue) • Doppler lidar. (NOAA/CU) • TCCON-FTS for 4 months (Sept-Dec 2012). (NASA Ames)

INFLUX ground-based instrumentation

Atmospheric inversions 101 • Take a first guess at emissions • Transport these through the atmosphere using an atmospheric model (reanalysis) • Compute CH 4 at measurement points • Compare modeled and observed CH 4 • Adjust first guess of emissions to minimize the difference between observed and modeled CH 4.

Comparison of [CO 2] at INFLUX sites Afternoon daily [CO 2] 2011 2012 2013

Comparison of [CO 2] at INFLUX sites • Afternoon [CO 2] with 21 -day smoothing • Site 03 (downtown): high [CO 2] • Site 01 (background): low [CO 2] • Seasonal and synoptic cycles are evident

Spatial Structure of Urban CO 2 Average [CO 2] above background site • Site 09 measures 0. 3 ppm larger than Site 01 • Site 03 (downtown site) measures larger [CO 2] by 3 ppm Afternoon daily values, 1 Jan – 1 April 2013

Vulcan and Hestia Emission Inventories / Models Vulcan – hourly, 10 km resolution for USA • See: Kevin Gurney/ • http: //hestia. project. asu. edu/ 250 m res - Indy. Hestia: high resolution emission data for the residential, commercial and industrial sectors, in addition to the transportation and electricity production sectors.

Combined sector temporally and spatially resolved Hestia emissions

Spatial structure: Model-data comparison • Backward model results using footprints and Hestia 2002 fluxes • Agreement in terms of the ordering of the sites • Observations are 25% higher than modeled values, on average Average [CO 2] above background site Miles et al, in prep

Lauvaux et al, in prep

CH 4 Enhancement (Site 02 – Site 01) as a Function of Wind Direction April – November 2011 (Afternoon hours only)

Urban enhancement (Site 02 – Site 01): 100+ m AGL tower: CH 4 7 ppb • Green arrows point to the sources of enhanced CH 4 measured at Site 02, compared to Site 01 • Large source to the southeast of Site 02, as well as to the west (urban center) • Maximum enhancements: ~ 10 ppb CH 4

Conclusions: Atm inversions • Capture total emissions, like airborne mass balance. strength and weakness. • Can be used to quantify temporal variability in emissions over years • Sources can be disaggregated via 1) trace gases or 2) prior knowledge of location and time of emissions. • Can provide spatially resolve emissions - given sufficient atmospheric data density. • “Footprints" are still relatively small and dependent on wind direction. • Uncertainty assessment is complex, but MCI results show inversion uncertainties equal to those of agricultural inventories (Schuh et al. , 2013).

Conclusions: Atm inversions • INFLUX and N. Marcellus experiments are attempting to bring together (relatively) rigorous top-down and bottom-up assessments, and include both airborne and tower-based atmospheric inversions.

Overall summary • Comparisons across methods / scales are important to gather understanding of the true net impacts on the global budget. • Constructing a rigorous comparison is challenging. • Aircraft are great for covering space, and for simple methods/uncertainty assessment, but are poor for temporal sampling. • Towers are great for long-term monitoring and temporal variability, but the methodology is complex. • In both cases, high measurement accuracy is required and sensors are costly. • The effort is important - reducing GHG emissions is critical – we need to achieve (and verify) large reductions in C emissions.