Toward a Quantitative Characterization of the Influence of … · 2016. 2. 4. · OZONE --- ATMOSPHERIC SIGNIFICANCE AND SOURCES MEASUREMENTS: LOCATION, CAMPAIGNS, AND INSTRUMENTS

OZONE --- ATMOSPHERIC SIGNIFICANCE AND SOURCES

MEASUREMENTS: LOCATION, CAMPAIGNS, AND INSTRUMENTS

OBSERVATION ANALYSIS

SUMMARY, CONCLUSIONS, AND ACKNOWLEDGEMENTS

BAO lies between NOx and VOC emission sources in the Denver-Colorado Front Range (urban to the south, O&NG to the north east) BAO measurement platforms: multiple ground sites and 300m tall tower with moveable carriage for vertically resolved measurements

Field Campaigns •  Summer 2014 – FRAPPE, Front Range Air

Pollution and Photochemistry Experiment

•  Summer 2012 – SONNE, Summer Ozone Near Natural Gas Emissions

1. Field measurements from the Boulder Atmospheric Observatory have been used with a photochemical box model to understand air pollution and photochemistry in a region currently out of compliance with national air quality standards for O3

2. Observed Ozone Production Efficiencies are ~5, but do not distinguish urban from Oil and Natural Gas O3

3. Photochemical box model indicates a 16% contribution of Oil and Natural Gas VOC Emissions to local O3

Methods to Differentiate OPEs from Distinct Regional Emission Sources…

FRAPPE (2014) Results– • Afternoon O3 always correlated with photochemically

oxidized reactive nitrogen (NOz) • OPE determined by Ox/NOz correlations over short

time intervals

FLEXPART backward wind trajectory from BAO at 6pm MDT (8/8/14),

illustrates regional mixing

Summertime daytime trends in 95th percentile O3 (1990-2010) at 52 rural locations.

Cooper, O., R. et al. 2012, J. Geophys. Res. 117, D22307

Instruments listed by Field Campaign

increased baseline ozone impacting some regions of the west-ern U.S.[39] Global average surface temperature increased by

0.074!C " 0.18!C per decade when estimated by a lineartrend for the 100 year period,1906–2005 [IntergovernmentalPanel on Climate Change (IPCC), 2007], while the rate ofincrease since the late 1970s is 0.15!C–0.20!C per decade[Hansen et al., 2010]. As global temperatures have increasedsince the 19th century so too has the global troposphericozone burden, primarily due to rising anthropogenic emis-sions of ozone precursors [Lamarque et al., 2005]. Based onthe model studies of ozone response to future climatechange, one might assume that past ozone changes have alsobeen influenced by climate change since the 19th century. Arecent intercomparison of 10 atmospheric chemistry models

run with year 2000 emissions but with 2000s and 1850sclimate shows a range of responses of tropospheric ozone tothe observed temperature increase [Stevenson et al., 2012].Six out of 10 models indicate ozone decreases at the surfaceof the northern hemisphere midlatitudes due to observedclimate change, but the decreases are small (<2 ppbv). Fourout of 10 models indicate regions of both positive and neg-ative surface ozone changes, but these changes are also small("2 ppbv). In the free troposphere of northern midlatitudesthe models indicate a range of ozone changes that are alsosmall ("2 ppbv).[40] Given the small and variable response of modeled

ozone to observed climate change from the 1850s to 2000swe do not expect a strong impact from climate change overthe much shorter time periods of 1990–2010 (surface ozone

Figure 8. As in Figure 7 but for 53 sites in summer.

COOPER ET AL.: RURAL U.S. OZONE TRENDS, 1990-2010 D22307D22307

14 of 24

Carriage Measurements Instruments: ARNOLD, PICARRO, QC-TILDAS

300m

Ground Measurements Instruments: GC-MS, Filter

Radiometers

b) Chemical Tracers • Tracers: Methane (CH4), CO, NOx, Ammonia (NH3) (see Measurements Section) • Advantages: unique to emission source, relatively long lived (τ > day), independent of wind history and speed

• Advantages: simple; urban influenced air flows from the south and O&NG from the north • Challenges: Front Range air masses are highly mixed

a) Wind Direction, Speed, and History

Chemical Tracers of Regional Emission Sources observed at BAO (2014), plotted by wind direction: 90th, 10th percentile 75thth, 25th percentile Average

• Limitations: sensitive to nitric acid deposition (i.e. artificially low NOz will raise Ox/NOz slope), difficult to extract slope in mixed air masses

At the BAO Tower: NOx and CO peaked slightly to the south (Denver), CH4 peaked slightly to the north (O&NG activity), but all three are regionally mixed

Chemical tracers do not identify regional emission sources with distinct ozone

production efficiencies

Ammonia (NH3) agriculture (livestock feedlots)

CO and NOx urban activity (combustion)

Methane (CH4) O&NG production

5

10

15

20

5 10 15 20ppbv

N

S

90th, 10th Percentile 75th, 25th Percentile AverageEW

NH3

ppbv

3691215

3 6 9 12 15

N

E

S

W

NOy 90th, 10th Percentile75th, 25th Percentile Average

ppbv

ppbv

1.91.95

22.052.1

1.95 2 2.052.1ppmv

N

Methane 90th, 10th Percentile 75th, 25th Percentile Average

E

S

W

ppmv

6090

120150180

60 90120150180

CO 90th, 10th Percentile 75th, 25th Percentile Average

N

E

S

W ppbv

ppbv

FRAPPE Case Study – Chemical tracers can be used to identify distinct air masses in this unique instance

Toward a Quantitative Characterization of the Influence of Regional Emission Sources on Ozone Production in the Colorado Front Range

Erin E. McDuffie1,2,3, Peter M. Edwards4, Jessica B. Gilman1,3 , Brian M. Lerner1,3 , William P. Dubé1,3, Michael Trainer1, Daniel E. Wolfe3,5, Wayne M. Angevine1,3, Joost A. de Gouw1,2,3, Eric J. Williams1, Alex G. Tevlin6, Jennifer Murphy6, Emily V. Fischer7, and Steven S. Brown1,2

1NOAA ESRL Chemical Sciences Division, 2Department of Chemistry- University of Colorado, 3Cooperative Institute for Research in Environmental Sciences – University of Colorado, 4Department of Chemistry – University of York, 5NOAA ESRL Physical Sciences Division, 6Department of Chemistry - University of Toronto, 7Department of Atmospheric Science - Colorado State University

NO NO2

O3 Sunlight, O2

VOC RO2 RO OH

O2

OH NOz

The Boulder Atmospheric Observatory (BAO)

Denver

Greeley

Boulder

Fort Collins

BAO

Observed 2012 Diurnal Profiles of Non-Methane Hydrocarbon (NMHC) OH Reactivity, and NOx and O3

Why Ozone? •  Tropospheric ozone (O3) is a health hazard and greenhouse gas

that alters the radiative balance of Earth’s atmosphere •  Emissions and subsequent oxidation of NOx (= NO + NO2) in the

presence of volatile organic compounds (VOCs) is the only known mechanism for tropospheric O3 production

Front Range Air Pollution and Photochemistry Experiment – FRAPPE, summer 2014

Summer Ozone Near Natural Gas Emissions – SONNE, summer 2012

Total Reactive Nitrogen (NOy)

Atmos. Ring-down Nitrogen Oxide Laser Detector – ARNOLD (NOAA) --

NOx ARNOLD (NOAA) ARNOLD (NOAA)Ozone ARNOLD (NOAA) TECO (NOAA)

Speciated VOCs -- GC-MS (NOAA)Ammonia QC-TILDAS (U Toronto) --

CH4 PICARRO (CSU) PICARRO (NOAA)CO PICARRO (CSU) CO Analyzer (NOAA)

Meteorological Data Tower Met. Stations (NOAA) Tower Met. Stations (NOAA)jNO2 jNO2 Filter Radiometer (NCAR) --

Surface Albedo Multi-Filter Radiometer (NOAA) --

ANALYSIS METHOD #1: OBSERVED OZONE PRODUCTION EFFICIENCY ANALYSIS METHOD #2: OBSERVATIONS + PHOTOCHEMICAL BOX MODEL

O3 Production Efficiency (OPE): Slope of Ox (= O3 + NO2) against NOz ( = NOy - NOx) as a measure of O3 molecules produced per NOx molecules emitted and oxidized

• Afternoon winds dominantly flow from the east •  Individual OPEs indicate that OPEs from the

south east (urban) and north east (O&NG and agriculture) are not distinctly different

FRAPPE Afternoon (12-6pm) Individual OPEs (15-minute resolution) plotted by wind direction, filtered for Ox/NOz correlation coefficients (r2) > 0.5

152025

15 20 25OP

E

N

OPE

S

EW

12

8

4

0

OPE

North East

12

8

4

0

OPE

South East

Average = 5.76 ± 3.69

Average = 5.93 ± 4.50

152025

152025

OPE

N

OPE

S

EW

OPE (r2> 0.50)

Western Wind Direction South East Wind Direction North East Wind Direction

E

N

W

S

OPE

Average OPEs in each wind sector may reflect the regional average, but do not distinguish individual sources

FRAPPE, Entire Campaign – •  Individual OPEs show no correlation with the ratio of

enhanced CH4/CO, NOx, or NOz

∂x∂t

Photolysis

O3 OVOC

Nitrates

Physical Loss (dilution)

VOCs NOx Constrained Inputs

Temp.

Dynamically Simple Model for Atmospheric Chemical Complexity (DSMACC) Emmerson, K. M., Evans, M. J., Atmos. Chem. Phys, 2009 • Chemistry described by Master Chemical Mechanism (MCM 3.3.1) ~4000 chemical species, ~15,000 chemical reactions • Photolysis rates calculated by TUV 5.2, then scaled to diurnal average observations

• Background levels of O3 plus 8 secondary products are entrained into box at the dilution rate o O3 – background from FRAPPE LIDAR data o Acetaldehyde, Acetone, Propanal, Butanal, MEK, Ethyl&Propyl Nitrates – backgrounds from 12am-3am averages

Observational Inputs and Constraints: • 24-hr simulations, initialized with observations at 6am local time (MDT) and

chemically constrained to SONNE (2012) diurnal average observations of primary VOCs (~40 compounds), NOx, and meteorological parameters (i.e. temperature)

0.8

0.6

0.4

0.2

0.0NM

HC V

OC

OH

Reac

tivity

(s-1

)

00:00 06:00 12:00 18:00 00:00Time of Day (MDT)

12

8

4

0

NO

x (ppbv)

1.0

0.8

0.6

0.4

0.2

0.0

12:00 AM8/4/12

6:00 AM 12:00 PM 6:00 PM

dat

80

60

40

20

0

O3 (ppbv)

Observed NOx Observed O3 Solar Profile

Mean OH Reactivity Alkanes - 49% Alkene+Alkynes - 4%

Aromatics - 3% Aldehydes+Ketones - 24%

Alcohols - 13% Biogenics - 7%

0.8

0.6

0.4

0.2

0.0NM

HC V

OC

OH

Reac

tivity

(s-1

)

00:00 06:00 12:00 18:00 00:00Time of Day (MDT)

12

8

4

0

NO

x (ppbv)

1.0

0.8

0.6

0.4

0.2

0.0

12:00 AM8/4/12

6:00 AM 12:00 PM 6:00 PM

dat

80

60

40

20

0

O3 (ppbv)





0.8

0.6

0.4

0.2

0.0NM

HC V

OC

OH

Reac

tivity

(s-1

)

00:00 06:00 12:00 18:00 00:00Time of Day (MDT)

12

8

4

0

NO

x (ppbv)

1.0

0.8

0.6

0.4

0.2

0.0

12:00 AM8/4/12

6:00 AM 12:00 PM 6:00 PM

dat

80

60

40

20

0

O3 (ppbv)





Model Results

1.00.0Normalized OH Reactivity (s-1)

Alkanes

Alkenes+Alkynes

Aromatics

Aldehydes+Ketones

Alcohols

Biogenics

VOC Class - OH Reactivity O&NG Fraction Non O&NG Fraction


Alkanes

Alkenes+Alkynes

Aromatics

Aldehydes+Ketones

Alcohols

Biogenics

VOC Class - OH Reactivity O&NG Fraction Non O&NG Fraction

O&NG - 87%

O&NG - 5%

O&NG - 18%

O&NG - 0%

O&NG - 0%

O&NG - 0%


Alkanes

Alkenes+Alkynes

Aromatics

Aldehydes+Ketones

Alcohols

Biogenics

VOC Class - OH Reactivity O&NG Fraction Non O&NG FractionO&NG VOC Fractions – Gilman, J., et al., ES&T 2013

• VOCs emitted from O&NG activity are predominantly alkanes. Remaining VOCs predominantly from surrounding urban and agricultural activities • O&NG VOC fraction removed from the observed Base

Case decreases the absolute amount of daily photochemical ozone: -5 to -25% (-16%, observed SONNE NOx) • Doubled O&NG VOC fraction from the observed Base

Case increases the absolute amount of daily photochemical ozone: +2 to +18% (+9%, observed SONNE NOx)

25

20

15

10

Phot

oche

mic

al O

zone

(ppb

v)

2520151050NOx (ppbv) - 24hr-Average

NMHC OH Reactivity Distributions (s-1)

Observed VOC Diurnal Average

(Base Case)

RemovedO&NG VOC

Fraction

DoubledO&NG VOC

Fraction

Modeled Photochemical Ozone

Observed SONNE NOx = 5.1 ppbv

*

The Northern Front Range Metropolitan Region of Colorado is in a NOx sensitive regime

On Average, Oil and Natural Gas (O&NG) VOC Emissions Contribute ~16%* to Photochemically Produced Ozone at BAO

• Model indicates that OPE is highly sensitive to NOx. For a given VOC mixture, OPE decreases with increasing NOx (not shown) • Given the 24-hr diurnal average NOx mixing ratio observed during

FRAPPE (3.8 ppbv), the model reproduces the observed average of individual FRAPPE OPEs (r2 >0.5) within 1 standard deviation • Modeled SONNE OPE (Base Case) is within 10% of the observed

average of individual FRAPPE OPEs The author would like to thank Samuel Hall (NCAR) for jNO2 data, Kathleen Lantz (NOAA) for albedo measurements, and

Andrew Langford for O3 LIDAR data at BAO for FRAPPE 2014

•  O3, NOy, NH3, CH4, CO maxima correlate (above) •  CH4/CO ratio identifies different Ox/NOz regimes (right) •  High CH4/CO corresponds to different O&NG CH4/NH3 ratios (not shown) •  Distinct air masses do not have distinctly different OPEs (right)

30

20

10

0

NH 3

, NO y

(pp

bv)

12:00 PM8/8/14

2:00 PM 4:00 PM 6:00 PMLocal Time (MDT)

2.20

2.10

2.00

Total CH4

80

60

40

20

0

O 3 (

ppbv

)

200

160

120

80

CO (ppbv)August 8th - Time Series O3, NH3, CH4

NOy, CO

80

70

60O x

ppb

v

54321NOz (= NOy - NOx) ppbv

2.50.0 slope = 5.17±0.13 slope = 4.65±0.03

Total CH4 / CO

O3 Production Efficiency

20

15

105

0

-5

OPE

(r2 <

0.5)

12108642NOx, NOz, CH4/CO (ppbv)

OPE vs NOz OPE vs CH4/CO OPE vs NOx

OPE as a metric: • Advantages: independent of air mass age, dependent

on air mass composition (i.e. changes with emission source), not sensitive to local titration

RESEARCH GOAL: DETERMINE THE CONTRIBUTION OF REGIONAL EMISSION SOURCES TO OZONE IN THE COLORADO FRONT RANGE

Why the Front Range?

• The Denver urban area is currently out of compliance with national air quality standards for O3 in the summer months

• In contrast to the eastern US, many western locations have positive daytime summer O3 trends

Possible reasons include increases in: -  Western wild fire activity -  Stratospheric O3 intrusions -  Pollution Transport from Asia -  Changes in regional Urban and Oil and

Natural Gas (O&NG) Emissions

1.00.90.80.70.60.5Ac

etal

dehy

de (

ppbv

)

06:00 12:00 18:00Local Time (MDT)

7060

50

40

30Ozon

e (p

pbv)

06:00 12:00 18:00Local Time (MDT)

Acetaldehyde Average Relative Deviation: +19%

Ozone Average Relative Deviation: -11%

• Dilution is only adjustable model input parameter, set to reproduce observed values of 9 secondary products between 11am-3pm (see examples right)

1086420

Model SONNE

1086420

OPE

FRAPfit Avg Model FRAPObservedFRAPPE

ModeledFRAPPE

ModeledSONNE

1σ fit error Individual OPE Average

Model OPE Results

Correspondence & Questions: [email protected]

201510

50

-5

OPE

(r2 <

0.5)

6.05.04.03.02.01.00.0NOx, NOz, CH4/CO (ppbv)

OPE vs NOz OPE vs CH4/CO OPE vs NOx

>

0.8

0.6

0.4

0.2

0.0NMHC

OH

Reac

tivity

(s-1

)

00:00 06:00 12:00 18:00 00:00Local Time (MDT)

12

8

4

0

NOx (ppbv)

1.0

0.8

0.6

0.4

0.2

0.0

12:00 AM8/4/12

6:00 AM 12:00 PM 6:00 PM

dat

80

60

40

20

0

O3 (ppbv)





1.00.90.80.70.60.5Ac

etal

dehy

de (

ppbv

)

06:00 12:00 18:00Local Time (MDT)

Observed Mixing Ratio Modeled Mixing Ratio Time when Model-Observation

Deviations Minimized

~16%

100

80

60

40

O x (

ppbv

)

7654321NOz (= NOy - NOx) (ppbv)

Individual OPE (15 minute resolution)

Afternoon Ox - NOz CorrelationJuly 16th - August 15th 2014, Daily 12pm-6pm MDT

NE SW

Documents

Toward a Quantitative Characterization of the Influence of … · 2016. 2. 4. · OZONE --- ATMOSPHERIC SIGNIFICANCE AND SOURCES MEASUREMENTS: LOCATION, CAMPAIGNS, AND INSTRUMENTS