Satellite observations of NO 2 and methane over U.S. oil and gas - - PDF document

▶
satellite observations of no 2 and methane over u s oil
SMART_READER_LITE
LIVE PREVIEW

Satellite observations of NO 2 and methane over U.S. oil and gas - - PDF document

May 17, 2023 120 likes •365 views

7/15/2020 Satellite observations of NO 2 and methane over U.S. oil and gas production areas Barb Barbara Dix Dix 1 , Joep Joep de de Bru Bruin 1, 1,2 , Es Esthe ther Roos oosen enbrand 1,2 1,2 , Ti Tim Vlem Vlemmix 3 , Colb Colby Fr Fran

slide-1
SLIDE 1

7/15/2020 1

Barb Barbara Dix Dix1, Joep Joep de de Bru Bruin1,

1,2, Es

Esthe ther Roos

  • osen

enbrand1,2

1,2, Ti

Tim Vlem Vlemmix3, Colb Colby Fr Fran ancoeu eur1,

1,4, Al

Alan an Gor Gorchov‐Negr Negron

  • n5,

Bria Brian McDonal McDonald1,4

1,4, Mi

Mikhai khail Zhizhin Zhizhin1,4

1,4,6, Chris

Christoph

  • pher Elvidg

Elvidge6, Pe Pepijn Veefkind2,

2,3, Pi

Piete eternel nel Lev evelt2,

2,3, Joo

Joost de Gouw Gouw1

Acknowledgements: Ronald van der A, Henk Eskes, Bud Pope NASA ACMAP program, Colorado Energy Research Collaboratory NOAA Cooperative Institute Agreement, Rocky Mountain Institute

Satellite observations of NO2 and methane

  • ver U.S. oil and gas production areas

1Univ

Univer ersity sity of

  • f Col

Colorado

2Univ

Univer ersity sity of

  • f De

Delft lft, the the Ne Netherla lands

3KNM

KNMI, the the Ne Netherla lands

4NO

NOAA AA

5Univ

Univer ersity sity of

  • f Mich

Michig igan an

6Co

Color lorad ado Scho School

  • l of
  • f Mi

Mines

Con Contact In Inform rmation fo for Pr Presente ters:

  • Dr. Barbara Dix, CIRES, Dept. of Chemistry, CU Boulder

Barbara.Dix@Colorado.EDU

  • Dr. Joost de Gouw, CIRES, Dept. of Chemistry, CU Boulder

Joost.deGouw@Colorado.EDU Barb Barbara Dix Dix1, Joep Joep de de Bru Bruin1,

1,2, Es

Esthe ther Roos

  • osen

enbrand1,2

1,2, Ti

Tim Vlem Vlemmix3, Colb Colby Fr Fran ancoeu eur1,

1,4, Al

Alan an Gor Gorchov‐Negr Negron

  • n5,

Bria Brian McDonal McDonald1,4

1,4, Mi

Mikhai khail Zhizhin Zhizhin1,4

1,4,6, Chris

Christoph

  • pher Elvidg

Elvidge6, Pe Pepijn Veefkind2,

2,3, Pi

Piete eternel nel Lev evelt2,

2,3, Joo

Joost de Gouw Gouw1

Acknowledgements: Ronald van der A, Henk Eskes, Bud Pope NASA ACMAP program, Colorado Energy Research Collaboratory NOAA Cooperative Institute Agreement, Rocky Mountain Institute

1. 1. Mo Motiv tivation

  • n and

and back backgr grou

  • und

2. Sa Satell llit ite da data used used in in th this is wo work 3. 3. Emissions:

  • ns: ob
  • bse

servatio ion and and sou source ce att attribut ution 4. 4. Met Methane ane monit monitori ring ng metrics metrics 5. 5. Summar mmary and and outlo

  • utlook

Satellite observations of NO2 and methane

  • ver U.S. oil and gas production areas

1Univ

Univer ersity sity of

  • f Col

Colorado

2Univ

Univer ersity sity of

  • f De

Delft lft, the the Ne Netherla lands

3KNM

KNMI, the the Ne Netherla lands

4NO

NOAA AA

5Univ

Univer ersity sity of

  • f Mich

Michig igan an

6Co

Color lorad ado Scho School

  • l of
  • f Mi

Mines

1 2

slide-2
SLIDE 2

7/15/2020 2

1 Moti tivation tion and and backgr background:

  • und: U.S. oil and natural gas production …

… is is at at an an all all‐tim time high. high.

Data: Baker Hughes

Production

  • duction and

and drillin drilling activ activities ties vary vary on

  • n diff

differen erent tim time sc scales. ales. … is is was was at at an an all all‐ti time high. high.

COVID-19 economic impact

1 Moti tivation tion and and backgr background:

  • und: U.S. oil and natural gas production

https://www.washingtonpost.com/graphics/national/united‐states‐of‐oil/

2019: 2019: U. U.S. is is the the larg larges est oil

  • il pr

produc

  • ducer

er worldwide. worldwide.

3 4

slide-3
SLIDE 3

7/15/2020 3

consumerenergyalliance.org

NOx NOx CH CH4

What What are are the the at atmospheric spheric im impa pacts of

  • f oil
  • il and

and gas gas pr production?

  • duction?

VO VOCs

1 Moti tivation tion and and backgr background:

  • und: Atmospheric impacts
  • NO

NOx:

nitrogen oxide = NO (nitric oxide) + NO2 (nitrogen dioxide) Source: combustion  on‐site motors/transportation Impacts: toxic (NO2), air pollution  ozone/particle formation Lifetime: ~ hours

  • VO

VOCs:

Volatile Organic Compounds/hydrocarbons Sources: oil/gas/fracking chemicals Impacts: harmful to toxic, air pollution  ozone/particle formation Lifetime: ~ hours ‐ days

  • CH

CH4:

methane Sources: infrastructure leaks, venting, flaring Impact: greenhouse gas Lifetime: ~ 9 years

NO NOx  Ai Air quality quality CH CH4  Clim Climate

Trivia question #1

5 6

slide-4
SLIDE 4

7/15/2020 4

  • EPA sites with NO2 and O3 sensors

1 Moti tivation tion and and backgr background:

  • und: NOx

4 3

several

(but mostly in Dallas)

Surf Surface air air quality quality mo monit nitoring ring in in O&G O&G production production reg regions is is very very lim limit ited ed.

3 1

1 Moti tivation tion and and backgr background:

  • und: Methane

Wh Why are are levels levels of

  • f methane

methane risin rising gl global

  • bally?

ly?

www.eurekalert.orgmultimediapub128882.php

What What is is the the co contri ribution bution of

  • f

O&G O&G sy syste stem emissions? sions?

7 8

slide-5
SLIDE 5

7/15/2020 5

1 Moti tivation tion and and backgr background:

  • und: Methane in the U.S.

www.epa.gov

*

* Some Some studies udies sugges uggest EP EPA emissions ssions are are unde undere restim imat ated

72% 72% of

  • f O&G

O&G sy syst stem met methane hane em emiss issions co come fro from pr production.

  • duction.

Oz Ozone

  • ne Mo

Monitori nitoring ng In Instrument (O (OMI) NASA AURA satellite July 2004 ‐ present Tropospheric

  • pospheric Mo

Monitori nitoring ng In Instrument (T (TROPOMI) ESA Sentinel 5P satellite October 2017 ‐ present Resolution: 24 x 13 km2 UV‐vis channel: NO2, HCHO Global coverage: 1 day Resolution: 7 x 3.5/7 km2 UV‐vis + SWIR channel: NO2, HCHO, CH4 Global coverage: 1 day

2 Sate tellite te da data used used in in this this wo work

9 10

slide-6
SLIDE 6

7/15/2020 6

Trivia question #2

SUN NADIR MEASUREMENTS SATELLITE backscattered Solar Radiation EARTH

2 Sate tellite te da data used used in in this this wo work

Ho How does does a rem remote sen sensin ing sat satellit llite work work? Satel tellite te obser servations ations of

  • f em

emiss issions are are aff affected ed by by surf surface refl flect ectanc ance and and cl clouds. uds.

Column density measurement in units of:

  • molecules per cm2 or
  • volume mixing ratio, e.g., ppb (parts per billion)

Surface Top of Atmosphere CH4 CH4 CH4 NO2 NO2 NO2 CH4 NO2 air air air air air air air air air NO2 CH4 CH4

11 12

slide-7
SLIDE 7

7/15/2020 7

Surface Top of Atmosphere CH4 CH4 CH4 NO2 NO2 air air air air air CH4 NO2 NO2 CH4 air air air air CH4 CH4 CH4 NO2 CH4 CH4 air air

Total column density = Emission + Background + Atmospheric chemistry + Atmospheric transport

NO2 NO2 + OH  HNO3

2 Sate tellite te da data used used in in this this wo work

What What does does a rem remote sensing sensing measurem measuremen ent me mean? an?

Satel tellite te me measur uremen ements of

  • f em

emiss issions are are aff affected by by back backgro round und co conc ncen entrations trations and and at atmospheri spheric chem chemis istr try and and transport. ansport.

2 Sate tellite te da data used used in in this this wo work

2018

OMI NO2

NO NO2 and and CH CH4 over

  • ver O&G

O&G pr prod

  • duction

tion area areas can can be be ob

  • bserv

rved ed fro

from space. space.

2018/05- 2020/04

TROPOMI NO2 TROPOMI CH4

2018/05- 2020/04 QA4ECV version 1.1; www.temis.nl Level 2, offline; https://scihub.copernicus.eu/ Level 2, offline; https://scihub.copernicus.eu/

13 14

slide-8
SLIDE 8

7/15/2020 8

2016: 2017: 2018:

Dix et al., GRL, 2020

3 Emi Emissions: ssions: NOx source attribution

Bakken Denver- Julesberg San Juan Permian Eagle Ford

NO NO2 ser serves es as as pro proxy fo for NO NOx emissions. sions.

NO2 = c1 ∙ background + c2 ∙ rig count + c3 ∙ oil volume

NO NOx em emiss issions fro from pr production

  • duction and

and drillin drilling ca can be be attri attribut buted ed separat parately ely. NO NOx em emiss issions are are do dominat nated ed by by tho those fro from drillin drilling of

  • f new

new wells wells.

NO NO2 signal signal ca caused ed by by: NO NOx fr fraction fr from pr production

  • duction

NO NOx fr fraction fr from dril illi ling NO NOx back backgr grou

  • und

Permian Basin

3 Emi Emissions: ssions: NOx source attribution by multivariate regression

Dix et al., GRL, 2020

15 16

slide-9
SLIDE 9

7/15/2020 9

Dix et al., GRL,2020; Francoeuret al., in prep.,2020 GorchovNegron et al., Environ. Sci. Technol., 2018

To Top‐do down wn (sa (satelli lite) and and bot bottom‐up up (in (inventory ry) sou source attributions attributions agree. agree. Dom Dominan nant but but declin declinin ing fra fracti tion of

  • f NO

NOx em emiss issions co come fro from drillin drilling.

3 Emi Emissions: ssions: NOx source attribution comparison with inventory

Permian Basin Bakken

Comparison with Fuel based Oil and Gas NOx inventory (FOG)

Dix et al., GRL,2020

NO NOx fro from flarin flaring co contribut ributes es ~5 ~5‐10% 10% to to tot total O&G O&G NO NOx emissions. sions.

(NO (NOx fro from flari flaring is is not not (ye (yet) di discernabl ernable fro from space. space.)

Flaring volumes derived from VIIRS satellite night‐time images

Elvidgeet al., Energies,2015

3 Emi Emissions: ssions: NOx from flaring

17 18

slide-10
SLIDE 10

7/15/2020 10

NO NO2 sc scales ales wit with oil

  • il and

and gas gas production production vo volumes. s.

TROPOMI NO2 Feb 2019 Feb 2019 Feb 2019

Esther Roosenbrand, CU Boulder, University of Delft, the Netherlands, master thesis, 2020 Industrial activity data from Enverus(formerly known as DrillingInfo)

3 Emi Emissions: ssions: NOx source attribution by spatial correlation

TROPOMI CH4 Feb 2019 Feb 2019 Feb 2019

CH CH4 sc scales ales wit with oil

  • il and

and gas gas production production vo volumes. s.

3 Emi Emissions: ssions: CH4 source attribution by spatial correlation

de Gouw et al., Scientific Reports, 2020 Industrial activity data from Enverus(formerly known as DrillingInfo)

19 20

slide-11
SLIDE 11

7/15/2020 11

3 Emi Emissions: ssions: CH4 source attribution by seasonal signal

Surf Surface CH CH4 em emiss issions are are “trapped” rapped” in in te temperature erature in invers versio ion during during win winter er.

2018/12- 2019/08

Uintah basin

TROP OPOMI OMI CH CH4 2018 2018/12 ‐ 2019 2019/03 TROP OPOMI OMI CH CH4 2018 2018/05 ‐ 2020 2020/04

de Gouw et al., Scientific Reports, 2020

Satel tellite te co column measurem measuremen ents ts are are consi consistent wit with surf surface mo monit nitori ring ng (U (Univ. of

  • f Ut

Utah). ah).

Emi Emissions: sions: CH4 source attribution by seasonal signal

de Gouw et al., Scientific Reports, 2020

21 22

slide-12
SLIDE 12

7/15/2020 12

3 Emi Emissions: ssions: CH4 and NO2 over the Denver‐Julesberg and San Juan basins

CH4 NO2 gas

  • il

gas

  • il

2018/05 ‐ 2019/12 Denver‐Julesberg San Juan

CH4 NO2

TROPOMI OPOMI NO NO2 spatial spatial di distributions stributions are are do dominat nated ed by by urban urban spra sprawl. TROPOMI OPOMI CH CH4 spatial spatial di distri stributions butions sho show som some co correl elation ation wit with O&G O&G production production areas. areas.

Industrial activity data from Enverus(formerly known as DrillingInfo)

3 Emi Emissions: ssions: CH4 over the Denver‐Julesberg basin: current work

Fort Collins Boulder Denver DIA Greely

*CH4 = CH4 – background CH4 background CH4 = monthly 10th percentile  more sensitive to boundary layer excess CH4

  • ngoing cooperation with CDPHE

 3 research flights in summer 2020 Denver‐Julesberg 2018/05 ‐ 2020/04

CH4*

Re Research fligh flights will will help help to to link link TROPOMI OPOMI CH CH4 obser servations ations to to loc local O&G O&G emissions. ssions.

suggested flight tracks

23 24

slide-13
SLIDE 13

7/15/2020 13

Trivia question #3

4 CH CH4 moni nitoring ring met metrics: s: trends in CH4 and CH4 vs production slopes

Note: 10 ppb  0.5% of total column

Ho How can can we we mo monit nitor the the indus ndustry’s perfor

  • rmance?

mance?

Slope = CH4 / prod. unit

Curren Currently: ly: average average CH CH4 em emiss issions per per oil

  • il and

and gas gas production production ~ co cons nstant Reduction duction in in CH CH4 emissions: sions:  slop slope will will decrease decrease

25 26

slide-14
SLIDE 14

7/15/2020 14

2019/01/31

4 CH CH4 moni nitoring ring met metrics: s: trends in CH4 – NO2 ratios

Curren Currently: ly: average average NO NO2 /CH CH4 rat ratios ~ co cons nstant Reduction duction in in CH CH4 emissions: sions:  rat ratio will will increa increase

Slope = NO2 /CH4 de Gouw et al., Scientific Reports, 2020

Di Distri ributions butions are are det determ rmine ined by by em emiss issions and and wind. wind.  Upset Upset em emiss issions are are captured captured in in 90 90th

th perc

percen entile. le.

emissions wind

4 CH CH4 moni nitoring ring met metrics: s: detection of upset emissions by pixel statistics

27 28

slide-15
SLIDE 15

7/15/2020 15

4 CH CH4 moni nitoring ring met metrics: s: detection of single events

2019/02/26 2019/02 2019/03 2019/03/17

Sing Single le day day (m (multi ti‐day?) day?) even events ts get get “lo “lost” in in mo monthly hly averages. averages.

Trivia question #4

29 30

slide-16
SLIDE 16

7/15/2020 16

4 CH CH4 moni nitoring ring met metrics: s: discussion on emissions

Be Benefit nefit of

  • f nat

natural ural gas gas over

  • ver co

coal al rem remain ins only

  • nly if

if leak akage age is is belo below break break‐even even poin point.  Ho How do do we we define define break break‐even even poin point?

Natura Natural gas gas as as brid bridge ge fu fuel fo for a carb carbon

  • n‐fr

free ee fu future?

Permian: 2.7 ± 0.5 Mt/a

break‐even point based on gas system TROPOMI CH4 inversion Zhang et al., Science Advances, 2020 (CH4 emissions from atmospheric modeling) Schneising et al., ACPD, 2020 (CH4 emissions from mass balance) break‐even point based on oil and gas system

Permian: 3.16 ± 1.13 Mt/a de Gouw et al., Earth’s Future, 2014

Sum Summary ary

  • NO

NOx fr from pr production

  • duction and

and drillin drilling ca can be be att attributed ibuted separ separately ly using using NO NO2 tim time serie

  • series. To

Top‐down down and and bot bottom‐up up sou source att attribut butions ions agr agree on

  • n di

dist stri ribut bution ion and and trend. end.

  • In

Increased co columns of

  • f NO

NO2 and and CH CH4 consis consistently ly cor correla elate wit with oil

  • il and

and ga gas pr product

  • duction.
  • n. De

Develop loped metrics metrics ca can iden identify tify basin basin‐wi wide de trends nds and and loc local

  • utlie
  • utliers.
  • Daily

Daily glob global cover coverage ge is is suit suitab able le to to ob

  • bserve in

interm rmit ittent indus industrial rial activity activity.

  • Sa

Satellit llite da data ca can help help asse assess ss eff effect ctiv iveness eness of

  • f NO

NOx and and CH CH4 re regulations. Outloo Outlook

  • Co

Continued nued wo work on

  • n metrics

metrics dev developm lopmen ent. t.

  • Quan

Quantif tifying basin basin‐wi wide de NO NO2 and and CH CH4 emi emissions ssions and and pus pushi hing ng sub sub‐basi basin sc scale ales.

5 Sum Summary and nd outloo

  • utlook

31 32

slide-17
SLIDE 17

7/15/2020 17

Additional slides Natural gas as bridge fuel to a carbon free future ?

  • Benefit of natural gas only remains if leakage is

below breakeven point: ‐ 3.2% gas system (Alvarez et al., 2012) ‐ 2‐3% oil and gas system (this study)

  • Top‐down and bottom up estimates don’t agree

 heterogeneity of leakage  Need to understand technical equipment

4 CH CH4 mo monitori toring ng met metrics: s: Discussion on emissions

de Gouw et al., Earth’s Future, 2014

33 34

slide-18
SLIDE 18

7/15/2020 18

Epa.gov

EPA.gov

35 36

slide-19
SLIDE 19

7/15/2020 19

Ba Backgr ckground:

  • und: Ev

Evolution of

  • f UV

UV‐visible visible satel tellites tes measuring measuring NO NO2 (1996 (1996 – 2017) 2017)

SCIA SCIAMACHY

30 30 x 60 60 km km2

OM OMI

13 13 x 24 24 km km2

www.temis.nl

TROP OPOMI OMI

3.5 3.5 x 5.6 5.6 km km2

GO GOME

40 40 x 320 320 km km2

Jacob et al., ACP, 2016

37 38

slide-20
SLIDE 20

7/15/2020 20

conc concentr tratio ion Troposphere Stratosphere Product of DOAS analysis: Slant Column Density SCD = c(l)dl

Ba Backgr ckground:

  • und: th

the DO DOAS te techni chnique que

Vertical Column Density VCD = c(z)dz Models: Models: Ra Radiativ ive Tr Transfer Che Chemistry Tr Transport

  • = VCD

VCD

1E + 08 1E + 10 1E + 12 1E + 14 1E + 16

200 300 400 500 600 700 800 W avelength [nm ] Intensity [arbitrary units

O 3 UV O 3 vis H CHO O ClO O 4 O 2 H 2O SO 2 NO 2 BrO

Satellite group: http://giger.iup.uni-heidelberg.de/

Ba Backgr ckground:

  • und: Me

Measur uremen ement principle: principle: Differ eren ential ial Op Optic tical Ab Absor sorption tion Spe Spectroscopy (D (DOAS)

 Trace gases are identified by their narrow band absorption structures

39 40

slide-21
SLIDE 21

7/15/2020 21

Ar Area ea an and poi point so sour urce em emissio issions: s: Fi First re results

Beirleet al., Science Advances, 2019

2018/05 – 2020/03

Permian Basin

power plant 2018/05 – 2020/03

Permian Basin

power plant

po power pla plant Thi his wo work* EPA‐CEM CEMS NO NOx emissio emission [t [tons/ s/day] 4. 4.3 ± 2 ± 2.2 6.1

  • il
  • il and

and ga gas prod

  • duction

uction Thi This wo work** FOG

sc scaled fr from 201 2015

NO NOx emissio emission [t [tons/ s/day] 301 301 ± 200 200 312 312 ± 9 ± 93

** assumes 100 tons/day of other NOx sources * near detection limit; correction factor 1.5

NO2 tropospheric VCD NOx emissions

Following Beirle et al., 2020

NOx lifetime = 4h; NOx/NO2 ratio = 1.32

pixels with active well pads

Top‐down and bottom‐up NOx emissions agree reasonably well.  Promising method to quantify NOx from oil and gas production spatially resolved. Met Method:

  • d: Schne

Schneisin ising et al al., ., AC ACPD PD, 2020 2020

  • sele

select Re Region

  • dail

daily CH CH4

  • BL

BL wind wind

‐ AV AVG all all BL BL wi winds nds ‐ 11 11‐13 13h loc local

  • Coor

Coordina nate Trans ansforma

  • rmation

tion

(R (Rot

  • tated Spheric

pherical Coo Coordinates)

1.

  • 1. Cen

Center of

  • f re

regi gion:

  • n: 0°

0° lon/lat

almo almost re rect ctangular angular dis distance gri grid*

2.

  • 2. zo

zonal wind wind di direction

eas easy to to ca calculate in integrated flu flux eas easy to to av average da data

  • sub

subtract “s “suita table” backgr ground

  • und

* Illu Illustration (flu (flux ca calculatio ion is is bas based on

  • n ar

area ea) TROPOMI/operational product Permian 20181027 quality filtered all data center of rotation

XCH4 [ppb]

bg  0 region of interest hot spot

XCH4 = column-averaged dry air mixing ratio [ppb] XCH4 = XCH4 above background [ppb]

center of rotation

41 42

slide-22
SLIDE 22

7/15/2020 22

Understanding CH4 satellite observations: background levels

www.esrl.noaa.gov/gmd

CH4 is rising globally and has a seasonal cycle  Any local observation contains backgrounds

Average enhancement

  • ver Permian Basin

Understanding CH4 satellite observations: atmospheric transport

Turner et al., ACP, 2018

averaged CH4 is higher above sources magnitude and area is “smeared out” by wind

Turner et al., ACP, 2018

43 44

slide-23
SLIDE 23

7/15/2020 23

CH CH4 and and NO NO2 are are enhanc nhanced ed over

  • ver

the the sam same areas areas where where oil

  • il and

and gas gas are are pr produc

  • duced.

ed.

2018/1 2018/12 ‐ 2019/0 2019/08

CH4

Emi Emissions: sions: CH4 NO2 correlations

NO2 gas

  • il

de Gouw et al., Scientific Reports, 2020

45