Click to edit Master subtitle style

2/6/12

Quantification of anthropogenic emissions from an urban region:

First results of time-integrated flask samples from the Indianapolis Flux

Project (INFLUX)Jocelyn Turnbull, Colm Sweeney, Doug Guenther, Anna Karion, Tim Newberger, Pieter Tans, NOAA/ESRL and CIRES, University of Colorado

Natasha Miles, Scott Richardson, Thomas Lauvaux, Kenneth Davis, Pennsylvania State University

Paul Shepson, Obie Cambaliza, Purdue UniversityKevin Gurney, Arizona State University

Scott Lehman, INSTAAR, University of ColoradoNOAA/ESRL Carbon Cycle Group, INSTAAR Isotope and 14C Lab Staff

2/6/12

Why use time-integrated flask samples?

In urban regions, can have substantial variability in mixing ratios – integrated samples reduce some of this noise

Match atmospheric sampling with the ~1 hour time step of regional models and meteorological data

Keep integrating time short enough that meteorology is consistent

0 10 20 30 40 50 6002468

1012

Minute of hour

In s

itu C

O2

(p

pm

)

INFLUX tower 1, 2011-04-09 20:00-21:00 UTC

Click to edit Master subtitle style

2/6/12

integrating volume (15L)

filter

valve

water

trap

Dryer/integrater package Integrating compressor package Portable flask package

filter

fast flushvalve

ventvalve

0-1L flow controller

pump

pressure reliefvalve pressure

sensor

pump

• Sample collected over 1 hour• Mix 1 hour of air by flowing through

a large mixing volume• Vary flow rate during filling to obtain

close to linear mixture• Collect air into 2 PFP flasks for

measurement

Integrated flask sampling

0 12 24 36 48 6002468

1012

Weighting function

Time (minutes)

Percentage of air in flask at

end of 60 minutes

0 12 24 36 48 600

0.0050.01

0.0150.02

0.0250.03

0.0350.04

0.0450.05

Weighting function

Time (minutes)

Percentage of air in flask at

end of 60 minutes

2/6/12

Integrated flask sampling validation

in situ – flask difference 0.01±0.09 ppm

Hourly averaged continuous – integrated flask CO2

For comparison: LEF 396 m in situ vs grab flask samples (2006-2011) in situ – flask difference 0.03±0.38 ppm

2/6/12

INFLUX: Tower flask results

Flask samples collected only when Tower 2 is downwind

Tower 1 is always upwind, background station

(Criteria met 25% of days)

Towers are 120-150m tall cell phone towers

Prevailing wind

2/6/12

CO2 enhancements across Indianapolis

Tower 2 (downwind) - Tower 1 (upwind)

2/6/12

Enhancements across Indianapolis

Strong enhancements in most anthropogenic species

2/6/12

Methane not always enhanced!

2/6/12

D14CO2 and fossil fuel

CO2D14CO2 is lower at

Tower 2 – 14C-free CO2ff

decreases D14CO2

D14CO2

CO2ff

CO2ff =

CO2obs (Δobs -Δbg)

Δff -Δbg

CO2r (Δr -Δbg)

Δff -Δbg

CO2ff enhanced at Tower 2

2/6/12

Contributions to CO2 enhancement

Flask samples from Dec 2010 – May 2011Slope = 1.1 ± 0.2 ppm/ppm

1:1 line (if DCO2 all due to CO2ff)

2/6/12

Source apportionment from CO2 isotopes: D14CO2

D14Csource = slope = -900 ± 150‰

DobsCobs=DbgCbg + DsCsCobs=Cbg + Cs

DobsCobs - DbgCbg=Ds(Cobs - Cbg)

2/6/12

D14C DCO2 fraction

Background atmosphere 40 ‰

Urban source -900 ± 150 ‰

Fossil fuels -1000 ‰(zero 14C content)

0.9 ± 0.15

Other sources (biofuel/respiration/photosynthesis)

40 ‰ 0.1 ± 0.15

DsCs=DffCff + DotherCother Cff/Cs=(Ds – Dother)/(Dff – Dother)

Source apportionment from CO2 isotopes: D14CO2

Indiana mandates 10% bio-ethanol in gasoline In winter, Indianapolis urban CO2 emissions are 90% CO2ff,

10% biofuelNo significant respiration/photosynthesis contribution, even in

summer!

2/6/12

Source apportionment from CO2 isotopes: d13CO2

d13Csource = slope = -29.1 ± 1.1‰

2/6/12

d13C D14C DCO2 fraction

Background atmosphere -8.0 ‰ 40 ‰

Winter urban source -29.1 ± 1.1 ‰ -900 ± 150 ‰

Fossil fuels -1000 ‰ 0.9 ± 0.15

Natural gas -40 ‰ 0.33

Petroleum -28 ‰ 0.30

Coal -24 ‰ 0.27

Other sources 40 ‰ 0.1 ± 0.15

Bioethanol (from corn) -10.4 ‰ 0.1

Respiration/photosynthesis

-26.0 ‰ 0.0

dsCs=dngCng + dpetCpet + dcoalCcoal + dbfCbf + dbioCbio

Source apportionment from CO2 isotopes: D14CO2 and d13CO2

Split CO2ff into different fuel types using d13CO2Compare with inventory estimates of fuel types

2/6/12

Correlation with CO and benzene

Emission ratios for CO and benzene are ~20% lower than observed in other US cities Large power plant contributes 30% of Indy CO2ff = lower emission ratios?

CO:CO2ff12±3 ppb/ppm

Benzene:CO2ff14±3 ppt/ppm

2/6/12

Summary INFLUX towers use a new time-integrated flask sampling

system to collect samples averaged over a 1 hour period

Tower sampling is designed to capture both upwind (background) and downwind signals

Strong enhancements in anthropogenic species across the urban region

In winter and in spring in Indianapolis, the CO2 enhancement appears to be entirely due to CO2ff and bio-ethanol (there is no apparent photosynthesis/respiration signal)

Reasonable correlations of CO2ff with combustion tracers such as CO and benzene, with emission ratios slightly lower than Western US cities, likely due to influence of power plant