1. On network design for the detection of urban greenhouse gas emissions: Results from the Indianapolis Flux Experiment (INFLUX) Natasha Miles 1, Thomas Lauvaux 1, Kenneth Davis 1, Scott Richardson 1, Laura McGowan 1, Daniel Sarmiento 1, Colm Sweeney 2, Anna Karion 2, Michael Hardesty 2, Jocelyn Turnbull 2,3, Laura Iraci 4, Kevin Gurney 5, Igor Razlivanov 5, Paul Shepson 6, M. Obiminda Cambaliza 6, James Whetstone 7 1. The Pennsylvania State University, 2. NOAA/ESRL, 3. GNS Science 4. NASA/JPL, 5. Arizona State University, 6. Purdue University, 7. NIST AGU Fall Meeting A44F-01 San Francisco, CA 12 December 2013

2. INFLUX motivation Emissions mitigation will happen at local and regional scales. Validation of emissions mitigation will(?) require (independent) measurements Atmospheric GHG measurements have the potential to provide such independent emissions estimates.

3. Develop improved methods for determination of urban area- wide emissions, and spatially and temporally-resolved fluxes of greenhouse gases, specifically, CO 2 and CH 4. Determine and minimize the uncertainty in the emissions estimate methods. INFLUX objectives

4. Inventory estimates of sector-by-sector emissions at high spatial resolution Periodic aircraft flights with CO2, CH4, and flask samples Periodic automobile surveys of CH4 12 surface towers measuring CO2, 5 with CH4, and 5 with CO & Mesoscale atmospheric inversion 6 automated flask samplers from NOAA –Identify sectoral emissions –Intensive study of diurnal cycle planned TCCON-FTS for 4 months (Sept - Dec 2012) 4 eddy-flux towers – model assessment (installed Nov 2013) Doppler lidar (installed Apr 2013) Tracer release experiment (?) INFLUX methodology: Simultaneous application of multiple methods

5. Vulcan and Hestia Emission Data Products Vulcan – hourly, 10km resolution for USA Hestia : high resolution emission data for the residential, commercial, industrial, transportation and electricity production sectors. http://hestia.project.asu.edu/ A53E-0216. Gurney et al, Friday afternoon 250m res - Indy

6. Aircraft mass balance approach: 1 June 2011 Flight path Cambaliza et al 2013 (ACPD)

7. 8 ppm CO2 26,000 moles s -1 50 ppb CH4 197 moles s -1 1 June 2011 Results Cambaliza et al 2013 (ACPD) A53E-0225. Quantification of the methane emission flux from the city of Indianapolis, IN: identification and contribution of sources, Cambaliza et al., Friday afternoon

8. NG leak CH4 enhancement (ppb) North Observed CH 4 enhancements directly downwind of SSLF landfill and a natural gas TRS on Harding St. obtained during a surface mobile measurement on Jan. 21, 2013. Note that a NG leak was also observed on Oliver Ave. Bridge just outside the city center. Courtesy of M. O. Cambaliza (Purdue Univ). Drive-arounds: Separation/quantification of CH 4 sources Instrumented vehicles used to identify and quantify individual sources.

9. Marion County Key: Drive Path Data shown: Total INFLUX drive paths and methane enhancements (CSU 2013, Purdue 2012- 2013) ***Threshold set to show CH 4 data > 3 stdev above average background Courtesy of M. O. Cambaliza (Purdue Univ)

10. Marion County Key: Drive Path x40 x9 x29 Data shown: Total INFLUX drive paths and methane enhancements (CSU 2013, Purdue 2012- 2013) Panhandle TRS #2 Oliver Ave Bridge SSLF Residential Leak See also: A53E-0213. Quantification of Methane Emissions From Street Level Data, Prasad et al, Friday afternoon; Courtesy of M. O. Cambaliza (Purdue Univ)

11. Picarro, CRDS sensors; NOAA automated flask samplers; Communications towers ~ 100 m AGL 10 km Results to date: Tower flask and in-situ Mesoscale atmospheric inversion

12. CO2 WMO recommendation: 0.1 ppm CADS (first generation) systems (circled): -0.18 to 0.1 ppm average site error Checks of network intercalibration: In-situ – flask comparison at 5 INFLUX sites (ongoing) NOAA 1 hour integrated flask samples Mean value in-situ - flask: CO2: 0.09 ppm CH4: 0.6 ppb CO: -4.1 ppb Within WMO recommendations (urban) Round-robin testing using 3 NOAA-calibrated tanks (Nov 2013)

13. Flask results: C14 Turnbull et al., in prep Flask analysis: fossil fuel CO2

14. Sector-by-sector atmospheric CO2 mole fractions, in percentage contribution for each site Winter mean daytime [CO2] Hestia emissions combined with footprint analysis Commercial Industrial Mobile Residential Power Plant Commercial Industrial Mobile Residential Power Plant

15. Afternoon [CO2] with 21-day smoothing Site 03 (downtown): high [CO2] Site 01 (background): low [CO2] Seasonal and synoptic cycles are evident Comparison of [CO2] at INFLUX sites 201120122013

16. Afternoon [CO2] with 21-day smoothing Site 03 (downtown): high [CO2] Site 01 (background): low [CO2] Seasonal and synoptic cycles are evident Comparison of [CO2] at INFLUX sites

17. 201120122013 Afternoon daily [CO2]

18. Range = 10 ppm 3 ppm Comparison of [CO2] at INFLUX sites

19. CO2 range as a function of wind speed Observations: CO2 range amongst INFLUX sites

20. CO2 range as a function of wind speed Observations: CO2 range amongst INFLUX sites Increased residence time (at low winds) tends to increase the CO2 range 10% of ranges are > 10 ppm 29% of ranges are < 3 ppm

21. CO2 range as a function of wind speed Observations: CO2 range amongst INFLUX sites Model: Difference along domain- averaged wind direction Increased residence time (at low winds) tends to increase the CO2 range

22. Site 09 measures 0.3 ppm larger than Site 01 Site 03 (downtown site) measures larger [CO2] by 3 ppm Spatial Structure of Urban CO2 Average [CO2] above background site East of city Downtown Afternoon daily values, 1 Jan – 1 April 2013 Eastern edge of city

23. 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 [CO2] above background site Spatial structure: Model-data comparison

24. Jan – Apr 2013 AFTERNOON Site 02 – Site 01 CO2, ppm Urban [CO2] enhancement Eastern edge of city - background Average over all wind dir: 1.5 ppm Downwind of city: 2.4 ppm ** Arrows point to sources

25. Jan – Apr 2013 AFTERNOON Site 02 – Site 01 CO2, ppm Site 30 km east of edge of city - background site Average over all wind dir: 0.3 ppm Downwind of city: 1.1 ppm ** Arrows point to sources ** Black: Site 01 larger than Site 09 Urban [CO2] enhancement Jan – Apr 2013 AFTERNOON Site 09 – Site 01 CO2, ppm Eastern edge of city - background Average over all wind dir: 1.5 ppm Downwind of city: 2.4 ppm ** Arrows point to sources

26. SiteDescriptionSample height(s), m AGL Site 01Background10/40/121 Site 02East10/40/136 Site 03Downtown10/20/40/54 Site 04South60 Site 05NorthWest125 Site 06NorthEast39 Site 07West58 Site 08NorthEast (20 km from I-465) 41 Site 09East (30 km from I-465) 10/40/70/130 Site 10South40 Site 11North130 Site 13SouthEast (10 km from I-465) 87 How high do in-situ measurements need to be? What is the vertical structure of urban CO2 mole fractions?

27. Vertical profiles of daytime CO2, compared to top level -0.6 0.2 0.91.0 Background siteDowntown siteMixed site 3.41.3-1.0 0.3

28. Vertical CO2 (Daytime) Differences Site 02 – Mixed site CO2 at 10 m – CO2 at 136 m Arrows point to sources Mixed Site

29. Vertical CO2 (Daytime) Differences Site 02 – Mixed site CO2 at 10 m – CO2 at 136 m Arrows point to sources Source area varies for various heights Mixed Site

30. Vertical CO2 (Daytime) Differences Site 03 – Downtown site CO2 at 10 m – CO2 at 54 m Arrows point to sources Downtown Site

31. Vertical CO2 (Daytime) Differences Site 03 – Downtown site CO2 at 10 m – CO2 at 54 m Arrows point to sources “Spaghetti bowl” How well does the model do for this site? Downtown Site

32. Conclusions Whole city flux estimates achieved via aircraft mass balance. Drive- arounds used for source identification. Flask: Winter, CO2 = CO2ff. Summer, not true. Tower observations detect a clear urban signal in both CO2 (buried amid lots of synoptic “noise”). Differences vary greatly with weather conditions. Model-data comparisons show similar spatial structure. For more information, see http://influx.psu.edu