the seasonality of tropospheric ozone and reactive …...the seasonality of tropospheric ozone and...

The Seasonality of Tropospheric Ozone and Reactive Nitrogen: From Measurements to Models

3rd Year Analytical Seminar March 14, 2016

Outline:1) Project #1 – Colorado Emissions & Ozone Photochemistry2) Project #2 – Wintertime Reactive Nitrogen Partitioning

Erin E. McDuffie1,2,3, Steven S. Brown1,3, Jessica B. Gilman2,3, Brian M.Lerner2,3, Peter M. Edwards4, Dorothy Fibiger5, William P. Dubé2,3,Michael Trainer3, Wayne M. Angevine2,3, Stuart McKeen2,3, Daniel E. Wolfe6,Alex G. Tevlin7, Jennifer Murphy7, Emily V. Fischer8, Felipe Lopez-Hilfiker8,Joel A. Thornton8, Jeff Peischl2,3, Andrew O. Langford2,3, Christoph J. Senff2,3,John S. Holloway2,3, Eric J. Williams3, Joost deGouw1,2,3, Thomas B. Ryerson3,Kenneth Aiken2,3, Samuel Hall10, Kirk Ullmann10, Kathy O. Lantz10, BillKuster2,3, Rebecca Hornbrook10, Eric Apel10, Allan Hills10

1Department of Chemistry, University of Colorado, 2Cooperative Institute for Research in Environmental Sciences, University of Colorado, 3NOAA, Chemical Sciences Division, 4Department of Chemistry, University of York, United Kingdom, 5Geospace Postdoctoral Fellow, 6NOAA, Physical Sciences Division, 7Department of Chemistry, University of Toronto, Canada, 8Department of Atmospheric Science, Colorado State University, 9Department of Atmospheric Sciences, University of Washington, 10Atmosperic Chemistry Observations and Modeling Laboratory, NCAR, 11NOAA, Global Monitoring Division

Tropospheric Ozone: Motivation

‘Oxidized VOCs’ + NOx à O3

Greenhouse Gas + Health Hazard à Regulated by US EPA

Photochemical Source:

Tropospheric Ozone: Photochemistry

NO NO2

O3

NOzOH

Sunlight, O2

O3

NOxEmissions

NOxEmissions NO NO2

OH

Sunlight, O2

VOCOHO2

RO2 RO

O3

VOC Emissions


NOz

NO NO2 NOzOH

Sunlight, O2

VOCOHO2

RO2Organic nitrate

O3

RO


NOxEmissions

VOC Emissions

NO NO2 NOzOH

Sunlight, O2

VOCOHO2

RO2 RO

O3

Radical Recycling =

More Efficient O3 Production

O2


NOxEmissions

VOC Emissions

Local/Regional Emission Sectors

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 (

• Increased Western Wildfire Activity• Pollution Transport from Asia• Changes in Regional Urban and Oil

and Natural Gas (O&NG) Emissions

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 (

4000

3000

2000

1000

0 Surf

ace

Elev

atio

n (m

)

25 km

Urban Boundaries Active Oil&Gas Wells Agriculture Power Plants Measurement Site

Denver

Boulder

GreeleyFort

Collins

BAO

Ozone Nonattainment Area

Urban NOx + VOC

> 27,000 Active Oil and Natural Gas Wells

Colorado

Location Field Measurements OH Reactivity O3 Production Efficiency Box-Model

Colorado Northern Front Range (NFR)

4000

3000

2000

1000

0 Surf

ace

Elev

atio

n (m

)

25 km


Denver

Boulder

GreeleyFort

Collins

BAO


> 27,000 Active Oil and Natural Gas Wells

Colorado


Colorado Northern Front Range (NFR)

Urban NOx + VOC

300m Tower


Data: NOx, NOy, O3, Speciated VOCs, CH4, CO, NH3, j(NO2), albedo, meteorological data

2014FRAPPE/DISCOVER-AQ(NOy, no speciated VOCs)

2012 – SONNE (Speciated VOCs, no NOy)

300m

10m

2014 Chemical Tracers

6090

120150180

60 90120150180

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

N

E

S

Wppbv

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

Composition at BAO characteristic of Urban and O&NG Emission Sectors

Oil and Natural Gas: Enhanced CH4

Urban: NOx and Enhanced CO


4000

3000

2000

1000

0 Surf

ace

Elev

atio

n (m

)

25 km


Denver

Boulder

GreeleyFort

Collins

BAO


Average

90th, 10th Percentiles

75th, 25th Percentiles

2012 VOCs

Composition at BAO characteristic of Urban and O&NG Emission Sectors

VOC Distribution (ppbC)SONNE Diurnal Average6am-8pm Avg

82%

1%

3%6%

1%7%

Average Fractional Contribution

O&NG VOC Contribution


Gilman, J.B. et al. Environ Sci

Technol, 2013, 43(3), 1297

Alkanes Alkenes/Alkynes

Aromatics Aldehydes/Ketones Alcohols Biogenic VOC

Bar Height: Mean Carbon Mixing Ratio

Alkanes


OH Reactivity (OHR):

Σ(kOH+VOC *[VOC])

Previously used in O&NG focused O3 studies



Σ(kOH+VOC *[VOC])

Previously used in O&NG focused O3 studies

Commonly used as a metric for ‘O3 forming potential’


Marcellus Basin, PA• O&NG: 7%• Biogenic: 47%

Barnett Basin, TX• O&NG: 13-20%• Biogenic: 70%

Comparison to other US O&NG Basins:

5048

4644

4240

3836

3432

3028

26

-128 -124 -120 -116 -112 -108 -104 -100 -96 -92 -88 -84 -80 -76 -72 -68 -64

Rutter, A. P., et al. Atmospheric Environment, 2015, 120, 89

O&NG VOCs contribute ~ 50% to OH Reactivity


Swarthout, R. F., et al. Environ. Sci. Technol., 2015, 49(5), 3175

OPE: Slope of Ox ( = O3 + NO2) against NOz (= NOy- NOx ) as a measure of the number of O3 molecules produced per NOx molecules oxidized


NOT previously used in O&NG focused O3 studies

Local/Regional Emission Sectors

OPE: Slope of Ox ( = O3 + NO2) against NOz (= NOy- NOx ) as a measure of the number of O3 molecules produced per NOx oxidized


Average of 80 Individual OPEs = 5.3 ± 3.6 ppbv/ppbv

100

90

80

70

60

50

40

Ox (

ppbv

)

6543210NOz = (NOy - NOx) (ppbv)

100

90

80

70

60

50

40

Ox (

ppbv

)


Individual 15-min OPE

2014: 12pm-6pm (MDT) 16 July-15 August

4000

3000

2000

1000

0 Surf

ace

Elev

atio

n (m

)

25 km


Denver

Boulder

GreeleyFort

Collins

BAO


100

90

80

70

60

50

40

Ox (

ppbv

)


North East South West

Wind Diretion

OPE vs. Wind Direction


Urban and O&NG OPEs not distinguished by wind direction

OPE Differences:1) Obscured by air mixing 2) < 1.4 ppbv/ppbv

100

90

80

70

60

50

40

Ox (

ppbv

)


North East South West

Wind Diretion

Individual 15-min OPE

Combined Wind Direction & Chemical Tracer Analysis Results:

1) Mixed air masses at BAO2) O&NG and Urban sectors do not uniquely influence OPE by > 1.4 ppbv/ppbv

OPE vs. Chemical Tracers


Urban and O&NG OPEs are not distinguished by simple chemical tracers

• O&NG: Enhanced CH4

• Urban: Enhanced CO and NOx

r2 + N àp-values > 0.05


Photochemical Box-Model: Dynamically Simple Model of Atmospheric Chemical Complexity (DSMACC) - Emmerson, K. M., Evans, M. J., Atmos. Chem. Phys., 2009

NOT previously used in O&NG focused SUMMER O3 studies

Photolysis – NCAR’s TUV à Scaled to j(NO2)/j(O1D) observations Chemistry – Master Chemical Mechanism (MCM) à 4002 species, 15555 reactions

Wintertime O3 Production in Utah’s Uintah O&NG BasinEdwards, P. M. et al., Nature, 2014, and Edwards, P. M. et al., Atmos. Chem. Phys., 2013

Base Case Simulated O3

70

60

50

40

30O

zone

(ppb

v)

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

Model

Observations

-2.5%


On Average, Oil and Natural Gas (O&NG) VOC Emissions Contribute ~19%* to Maximum O3 Photochemically Produced at BAO

The Northern Front Range of Colorado is in a NOx Sensitive Regime

Alkanes Alkenes/Alkynes Aromatics Aldehydes/Ketones Alcohols Biogenic VOC

10

8

6

4

2

0

OPE

(ppb

v/pp

bv)

frap_obs frap_modl son_mod

2014 2012


Modeled OPE: Dependent on NOx mixing ratio and temperature

Observed Modeled Modeled

Average OPE vs Base Case

1) Modeled OPEs within 1 standard deviation of Observed 2014 Average (accurate Base Case)

10

8

6

4

2

0O

PE (p

pbv/

ppbv

)fb fx sb sx

Base Case No O&NG VOC

2014 Modeled 2012 Modeled

Base Case vs No O&NG VOCs

2) O&NG Emissions influence OPE by ~1.0 ppbv/ppbv(consistent with OPE vs wind direction analysis)

6.2 vs 5.2 ppbv/ppbv

6.3 vs 5.3 ppbv/ppbv

Project #1: Summary and Conclusions

Regional O3 photochemistry is dependent on local NOx and VOC emission sources

BAO experiences both O&NG & urban emissions

4000

3000

2000

1000

0 Surf

ace

Elev

atio

n (m

)

25 km


Denver

Boulder

GreeleyFort

Collins

BAO


At BAO:1) O3 production is sensitive to NOx emissions

2) O&NG VOC emissions contribute:~ 80% to observed carbon mass (ppbC)~ 50% to O3 forming potential (OHR) ~ 19% to maximum photochemical O3~ 1.0 ppbv/ppbv OPE

Project #2: Influence of Reactive Nitrogen Chemistry on Wintertime Tropospheric O3

NOzNOx

Emissions NO NO2OH

Sunlight, O2

O3

VOCOH

O2RO2 RO

VOC Emissions

NOzNOx

Emissions NO NO2OH

Sunlight, O2

O3

VOCOH

O2RO2 RO

VOC Emissions

Project #2: Quantify the Influence of Heterogeneous Reactive Nitrogen Mechanisms on Winter O3

Tropospheric Ozone: Dark Chemistry

NOxEmissions NO NO2


O3

Sunlight, O2

O3

NO3NOx

Emissions NO NO2


O3

O3+ N2O5

HNO3

ClNO2Sunlight, O2

O3

Consequences for:

• Wintertime tropospheric O3 budget• Lifetime and distribution of NOx, O3, and additional tropospheric oxidants (i.e. OH)• Magnitude and mechanisms for sources of tropospheric halogens

Reactive Nitrogen Partitioning

NO3NOx

Emissions NO NO2 O3+ N2O5

HNO3

ClNO2

VOCOH

O2RO2 RO

VOC Emissions

O3 NOx Reservoir

NOx Sink

Consequences for:

• Wintertime tropospheric O3 budget• Lifetime and distribution of NOx, O3, and additional tropospheric oxidants (i.e. OH)• Magnitude and mechanisms for sources of tropospheric halogens

Sunlight, O2

Reactive Nitrogen Partitioning

Heterogeneous N2O5 Processes: Key Parameters

γϕ

γ(N2O5) Reactive Uptake

Coefficient

ϕ(ClNO2)Production Yield

Range γ(N2O5):

10-3 – 0.04

Range ϕ(ClNO2):

0 – 1

NO3 N2O5

HNO3

ClNO2NOx

Reservoir

NOx Sink

NOxReservoir

Remaining Scientific Uncertainties: γ, ϕ in different ambient environments (e.g. temperature, RH, aerosol composition)

…Approach… field data to constrain values

February – March 2015• 6 weeks – 13 research flights

• Day and Night• Coastal and Interior Environments

Campaign: Wintertime Investigation of Transport, Emissions, and Reactivity (WINTER)

6-Channel CRD: N2O5, NO3, NOx, O3, NOy

5048

4644

4240

3836

3432

3028

26

-128 -124 -120 -116 -112 -108 -104 -100 -96 -92 -88 -84 -80 -76 -72 -68 -64

US East CoastNO3 N2O5

HNO3

ClNO2NOx

Reservoir

NOx Sink

NOxReservoir

Heterogeneous N2O5 Processes: Observed Variability

3 Examples from WINTER:

γϕ

2.01.51.00.5N2O5 (ppbv)

N2O5 as NOxreservoir

Feb. 24th

Washington DC

*HNO3 and ClNO2 from University of Washington CIMS

5048

4644

4240

3836

3432

3028

26

-128 -124 -120 -116 -112 -108 -104 -100 -96 -92 -88 -84 -80 -76 -72 -68 -64


HNO3

ClNO2NOx

Reservoir

NOx Sink

NOxReservoir


N2O5 HNO3 ClNO2


γϕ

2.01.51.00.5N2O5 (ppbv)


2.01.51.00.5N2O5 (ppbv)

N2O5conversion to

HNO3

Feb. 24th Feb. 23rd

Washington DC


5048

4644

4240

3836

3432

3028

26

-128 -124 -120 -116 -112 -108 -104 -100 -96 -92 -88 -84 -80 -76 -72 -68 -64


HNO3

ClNO2NOx

Reservoir

NOx Sink

NOxReservoir


N2O5 HNO3 ClNO2


γϕ

N2O5 HNO3 ClNO2

2.01.51.00.5N2O5 (ppbv)


2.01.51.00.5N2O5 (ppbv)

N2O5conversion to

HNO3

6.7%

39.6% 53.7%

2.01.51.00.5N2O5 (ppbv)

N2O5 conversion to ClNO2

Feb. 24th Feb. 23rd Feb. 8thWashington DC

New York City


5048

4644

4240

3836

3432

3028

26

-128 -124 -120 -116 -112 -108 -104 -100 -96 -92 -88 -84 -80 -76 -72 -68 -64


HNO3

ClNO2NOx

Reservoir

NOx Sink

NOxReservoir



γϕ

Reactive Nitrogen Chemistry not well quantified or parameterized, important to winter O3 budgetObservational Winter data collected during recent campaign

• Goals:– Constrain observed gamma– Model gamma under different observed conditions

Project #2: Summary and Future Work

Scientific Questions1) What drives Reactive Nitrogen Partitioning (i.e. differences in γ and ϕ)?2) Can we associate certain partitioning with certain ambient conditions?3) Can we use conditions/chemistry to predict γ and ϕ?

WINTER data show: γ and ϕ are highly variable

Methods1) Derive γ, ϕ from N2O5, ClNO2,

aerosol observations2) Box - Model γ

Model Output:

k(N2O5) = 14*c*SA *γ (N2O5 )

Questions?

*NYC from the C130 cockpit, photo courtesy of Joel Thornton

the seasonality of tropospheric ozone and reactive …...the seasonality of tropospheric ozone and...

Documents