Chapter 9: Tropospheric oxidant chemistry

Historical interest in tropospheric chemistry focused on air pollution�…but did not actually involve chemistry

Pittsburgh in the 1940s

London fog

Air pollutants (smoke) were considered to be directly emitted, transported by wind

Los Angeles smog revealed a different beast

Smog became worse as air drifted downwind of LA – chemical production?

Respiratory effects, eye irritation, damage to crops pointed to something else in air than just smoke

First recorded photochemical smog event: Los Angeles on July 26, 1943

• Visibility down to three city blocks
• Air smells like bleach
• Is this Japanese chemical warfare?

Los Angeles Times, July 27, 1943

Haagen-Smit (1953): identifies ozone produced in air from car exhaust�as the harmful component in Los Angeles smog

VOCs

NO_x

Postulated mechanism:

Shows that ozone is produced by oxidation of volatile organic compounds (VOCs)

in presence of NO_x and sunlight

Arie Haagen-Smit

sunlight

Ozone

Nitrogen dioxide (NO₂)

Early action to control VOCs and NO_x emissions �to fight this urban ‘photochemical smog’

Creation of California Air Resources Board (1967)

US EPA (1970):

ozone is “good up high, bad nearby”

O₂+hv

VOCs, NO_x + hv

Discovery of ‘acid rain’ in the 1960s:�long-range transport and atmospheric oxidation

• Sulfur dioxide (SO₂)

from coal power plants

• NO_x from fuel combustion

Acidification of remote lakes and streams (Scandinavia, Canada, US)

SO₂

NO_x

H₂SO₄

HNO₃

Oxidation…but how?

by oxygen in clouds?

The atmosphere as an oxidizing medium �

EARTH

SURFACE

Emission

Reduced gas

Oxidized gas/

aerosol

Oxidation

Deposition

Reduction

Atmospheric oxidation is critical for removal of many pollutants, e.g.

• methane (major greenhouse gas)
• toxic gases such as CO, benzene, mercury…
• gases affecting the stratosphere
• SO₂ and NO_x (forming acid rain)

But oxidation by O₂ is very slow because of stability of O₂ molecule

Urban smog aside, troposphere until 1970s was thought to be inert �except for slow oxidation by O₂

“The chemistry of the troposphere is mainly that of of a large number of atmospheric constituents and of their reactions with molecular oxygen…Methane and CO are chemically quite inert in the troposphere”

[Cadle and Allen, Atmospheric Photochemistry, Science, 1970]

Carbon monoxide (CO):

product of incomplete fuel combustion

Will CO from increasing vehicles accumulate to produce a global pollution cloud?

Discovery of OH radical production in troposphere (Levy, 1971)

solar flux I

at sea level

ozone absorption

cross-section σ

O(¹D)

quantum

yield q

qσI

Photolysis at sea level

is controlled by photons in 300-320 nm range

Chip Levy

surface

Ozone layer

310 nm flux

some flux remains

Tropospheric OH enables fast oxidation of non-radicals

Example: oxidation of CO to CO₂

Endothermic (requires heat): will not happen in atmosphere

Ozone is a stronger oxidant but is not a radical so reaction is too slow

Exothermic (releases heat)

but still negligibly slow in atmosphere

But oxidation of CO by OH is fast:

Oxidation by OH is further promoted if H abstraction can take place

hydrocarbon

OH is the PacMan of the atmosphere: it oxidizes anything that can be oxidized, with particular fondness for abstracting H atoms

OH

OH radical budget in the troposphere

where X is any non-radical reduced species

Source:

​

Sink:

Lifetime:

concentrations of OH are low, variable, very difficult to measure

Need to know mean [OH] to infer lifetime of X_i

Inferring global tropospheric OH from methylchloroform (Singh, 1977)

Methylchloroform (CH₃CCl₃):

• Uniquely anthropogenic (industrial solvent), banned by Montreal Protocol
• Removed from troposphere by oxidation by OH

Mass balance equation for troposphere (T):

loss rate from

CH₃CCl₃ + OH

emission

exponential decay

Lifetime =1/k[OH]_T

= 5 years

CO lifetime: 2 months

Methane lifetime: 10 years

Hanwant Singh

Global OH sink is controlled by CO and methane

• Methane (CH₄)

observed from space: 1800-1900 ppb

• Carbon monoxide (CO) observed from space: 50-200 ppb

Sources: fuel combustion, open fires, VOC oxidation

Sources: wetlands, livestock,

oil/gas production, landfills, coal mines, rice…

2018-2019 TROPOMI satellite data

Lifetime against oxidation by OH: 2 months

Lifetime against oxidation by OH: 10 years

…at the same time, CO and methane sinks are controlled by supply of OH

Until the 1980s, it was thought that tropospheric ozone was mainly of stratospheric origin – but that doesn’t work

Tropopause

Transport from stratosphere

10 Tmol O₃ a^-1

O₃ + hv → O₂ + O(¹D)

O(¹D) + H₂O → 2OH

46 Tmol O₃ a^-1

Chemical loss computed from observed tropospheric ozone

• A large tropospheric source of ozone is needed:
1. to balance the chemical loss of ozone
2. to avoid titration of OH by CO and CH₄ emissions

30 Tmol CO a^-1

30 Tmol CH₄ a^-1

00

In meantime, Demerjian et al. [1973[ used discovery of OH �to explain the formation of ozone in urban smog

VOCs

NO_x

Ken Demerjian

Chameides and Walker [1973] and Crutzen [1973] suggest that smog-like ozone production driven by CO and methane�could happen throughout troposphere

Bill Chameides

But where would the NO_x come from?

NO_x

OH

HNO₃

deposition

NO_x has a lifetime of a few hours against oxidation to HNO₃, so should only be present near combustion source regions

Discovery of NO_x in remote troposphere by aircraft in late 1980s

lightning

• Implies that chemical production is main source of ozone in global troposphere, stratospheric input is small in comparison
• Lightning NO_x drives ozone maximum in upper troposphere

Jacob et al. [1996]

Observations over South Atlantic (NASA TRACE-A): sufficient ozone production to offset loss

net production

net loss

We now know that NO_x is present throughout troposphere�and fuels ozone production

Sources: fuel combustion, open fires, lightning, soils

Tropospheric NO₂ columns measured by TROPOMI satellite (Apr-Sep 2018)

NO_x

HNO₃

OH

hours

Emission Deposition

organic nitrate reservoirs

Organic nitrate reservoirs allow transport of NO_x on global scale

Mechanism for ozone production from oxidation of CO�catalyzed by HO_x and NO_x radicals

Initiation: production of HO_x radicals

Propagation: oxidation of CO in presence of NO_x

Termination: loss of HO_x radicals

followed by H₂O₂ and HNO₃ deposition

H₂O₂ ≡ hydrogen peroxide

Mechanism drives ozone loss if NO is low (chain length <1)

HO_x and NO_x radicals catalyze ozone production in troposphere,�but loss in the stratosphere – why the difference?

In stratosphere,

In troposphere,

• CO/O₃ and NO/O₃ ratios are much higher in troposphere than stratosphere
• O concentrations are very low in troposphere

Methane oxidation cascade follows same schematic as CO

• Expanded HO_x family: HO_x ≡ OH + HO₂ + CH₃O₂ + CH₃O
• HO_x cycle: OH → CH₃O₂ → CH₃O → HO₂ → OH
• Oxidation of methane by above mechanism produces four ozone molecules
• If low NO_x, then CH₃O₂ + HO₂ → O₂ + CH₃OOH (methyl hydroperoxide)

​

CH₄ -4

CH₂O 0

CO +2

CO₂ +4

C oxidation #

methylperoxy radical

methoxy radical

formaldehyde

Formaldehyde can also photolyze:

• Five ozone molecules produced per molecule of methane
• New HO_x radicals produced in propagation steps to accelerate the chain;

CH₂O + hv is called a branching reaction.

​

CH₄ -4

CH₂O 0

CO +2

CO₂ +4

C oxidation #

methylperoxy radical

methoxy radical

formaldehyde

At low NO_x, methane oxidation does not produce ozone:

• Oxidation of methane is a sink of HO_x radicals and therefore an ozone sink, since ozone was consumed in the production of these HO_x radicals

CH₄ -4

CH₂O 0

CO +2

CO₂ +4

C oxidation #

methylperoxy radical

formaldehyde

methylhydroperoxide

CH₃OOH -2

Many VOCs besides methane are emitted to atmosphere

Formaldehyde observed from satellite

by solar backsatter: proxy for reactive VOCs

July 2006

10¹⁶ cm^-2

anthropogenic biogenic pyrogenic

• Anthropogenic VOCs:
• Combustion: alkanes, alkenes, aromatics
• Industry: aromatics, oxygenates
• Domestic products: oxygenates
• Biogenic VOCs: isoprene, terpenes
• Pyrogenic VOCs (open fires): alkenes, aromatics, oxygenates

Same basic mechanism for higher VOCs but many branches

OH can also add to double bonds of unsaturated VOCs, producing hydroxyorganics

RO can also decompose or isomerize to produce a range of aldehydes, ketones, dicarbonyls…

NO can also add to produce organic nitrates

Oxidation product goes on to react with OH, adding functionality and making more ozone

RO₂ can also

isomerize, or react with HO₂ or RO₂,

producing peroxides, epoxides,

alcohols,

carboxylic acids…

A brief cheat sheet for organic functions

aldehyde

​

ketone

​

dicarbonyl

​

peroxide

alcohol

​

epoxide

​

carboxylic acid

​

organic nitrate

carbonyl

O₃

HO₂

H₂O

NO

H₂O₂

VOC

NO₂

hν

O(¹D)

hν

M

OH

HNO₃

Dependence of ozone and OH on NO_x and VOCs

• HO₂ in this diagram denotes the ensemble of peroxy radicals including RO₂
• View CO as just another VOC

Dependence of ozone and OH on NO_x and VOCs

O₃

HO₂

H₂O

NO

H₂O₂

VOC

NO₂

hν

O(¹D)

hν

M

OH

NO_x-limited regime

NO_x-limited regime:

O₃ ↑ as NO_x↑

OH↑ as NO_x↑ VOCs↓

Is the production of ozone limited by supply of emitted NO_x or emitted VOCs?

Dependence of ozone and OH on NO_x and VOCs

O₃

HO₂

H₂O

NO

VOC

NO₂

hν

O(¹D)

hν

M

OH

HNO₃

NO_x-saturated

regime

VOC-limited regime:

O₃ ↑ as NO_x↓ VOCs↑

OH↑ as NO_x↓

a.k.a. NO_x-saturated regime

Dependence of ozone and OH on NO_x and VOCs

O₃

HO₂

H₂O

NO

H₂O₂

VOC

NO₂

hν

O(¹D)

hν

M

OH

HNO₃

NO_x-limited regime

NO_x-saturated

regime

O₃

NO_x

NO_x-limited

NO_x-saturated

O₃

NO_x-limited

NO_x-saturated

VOCs

OH

NO_x

NO_x-limited

NO_x-saturated

OH

NO_x-limited

NO_x-saturated

VOCs

NO_x-limited regime:

O₃ ↑ as NO_x↑

OH↑ as NO_x↑ VOC↓

VOC-limited regime:

O₃ ↑ as NO_x↓ VOC↑

OH↑ as NO_x↓

a.k.a. NO_x-saturated regime

Ozone (ppb) vs. NO_x and VOC emissions�Generic box model calculation

NO_x-

saturated

regime;

urban, winter

conditions

NO_x-limited regime; most common

Ozone production efficiency (OPE) per unit NO_x emitted

View chain reaction for ozone production as catalyzed by NO_x radicals

OPE↑ as VOC ↑ and NO_x ↓

NO_x at low concentration is very efficient at making ozone

Hudman et al. [2004]

GEOS-Chem ozone production efficiency (OPE) per unit NO_x

at 2-4 km altitude

Peroxyacetylnitrate (PAN) enables NO_x transport to remote troposphere

PAN decomposes as air warms

Evidence of PAN as reservoir for NO_x

Aircraft observations of Asian plumes transported over California

CO

O₃

PAN

HNO₃

May 5 plume at 6 km:

High CO and PAN,

no O₃ enhancement

May 17 subsiding

plume at 2.5 km:

High CO and O₃,

PAN →NO_x→HNO₃

Hudman et al. [2004]

NO_x

NO_x

HNO₃

PAN

O₃

CO

Rynda Hudman

� Addressing complex problems in atmospheric chemistry�requires 3-D models coupling chemistry and transport �(chemical transport models)

Foundation of chemical transport models: the continuity equation

X

Y

emission

Transport (wind vector U)

chemistry

aerosol microphysics

deposition

(wet and dry)

Solve continuity equations for number densities n = (n₁,…n_K) of ensemble of K species:

local trend in

number density

transport

(flux divergence)

emissions, deposition,

chemical and aerosol processes

Flux F_i = n_iU

internal

P_i, L_i

elemental volume

Brasseur and Jacob, chap. 4

number density n_i

[molecules cm^-3]

System of K coupled 4-D PDEs

Solve in individual grid cells of 3-D domain – grand computational challenge!

Input meteorological data from NASA GEOS system:

MERRA-2, 1980-present (0.5^ox0.625^o)

GEOS-FP, 2012-present (0.25^ox0.3125^o)0.5^ox0.625^o, 72 vertical levels

Modules

• transport
• emissions (HEMCO)
• chemistry
• aerosol microphysics
• deposition

Model solves 3-D chemical continuity equations

on global or nested Eulerian grid, native or coarser resolution

GEOS-Chem model

GEOS-Chem is used by hundreds of research groups around the world

10^th GEOS-Chem meeting (IGC10)

Washington U., June 2022

Late 1990s: first global 3-D models for tropospheric chemistry

Ozonesonde observations (solid), model (dashed)

Pressure, hPa

Model budget of tropospheric ozone: chemical production and loss dominate everywhere

Yuhang Wang

Wang et al. [1998]

Global budget of tropospheric ozone (Tg O₃ a^-1): �mostly determined by production and loss in troposphere�

O₃

O₂

hν

O₃

OH

HO₂

hν, H₂O

Deposition

NO

H₂O₂

CO, VOC

NO₂

hν

STRATOSPHERE

TROPOSPHERE

8-18 km

IPCC (2014) average across models

Tropospheric ozone lifetime: 24 ±4 days

HNO₃

2000s: first satellite observations of tropospheric ozone

OMI 2013 observations at 700-400 hPa

• Lifetime of 3 weeks allows transport around latitude bands

• Maximum values at northern mid-latitudes in spring-summer due to anthropogenic pollution

​

• High values in tropical regions affected by seasonal biomass burning

​

• Minimum values over tropical oceans due to chemical loss

Hu et al. [2017]

Ozone spatial and seasonal patterns are reproduced by models

• Partly natural: stratospheric influence, lightning, wildfires
• Partly anthropogenic: methane, intercontinental pollution, ag fires, ships, aircraft…

OMI satellite data

(X, Liu, SAO)

GEOS-Chem model

Mean 500 hPa ozone in JJA 2013

Hu et al. [2017]

But models have a much harder time with long-term ozone trends

NO_x emissions

historical ozone trend (1900-present)

trends from satellites (1996-present)

Unresolved questions:

1. Why does historical record show no increase until 1980?
2. Why is global tropospheric ozone presently increasing, when NO_x emissions are thought to have plateaued?

TOAR [2020]

Estimating the OH trend with the methylchloroform proxy

No emissions after 1996:

global mean [OH]

tropospheric mass of methylchloroform

methylchloroform+OH rate constant

Differentiate:

TRENDS IN GLOBALTROPOSPHERIC OH�inferred from methylchloroform and computed from models

There is presently no confidence or understanding of OH trends

But methane ice core record tells us that OH must have been relatively steady in past:

This is all still a mystery

models

methylchloroform

Stevenson et al., 2020

models

The weird growth curve of methane

Could this weirdness be driven by changes in OH? We don’t know.

My group’s recent work on this…

From inversion of satellite observations of atmospheric methane

• Changes in OH (methane lifetime) contribute to interannual variability of methane and to the recent methane surge
• Wetland inundation in Africa is the principal contributor to the surge

​

Qu et al. [2024]

Air pollution today decreases life expectancy�by 1.8 years on average worldwide

Decrease in life expectancy, years

data from https://www.stateofglobalair.org/

Air pollution

PM_2.5 surface ozone cancer smoking malaria water sanitation

PM_2.5 = particulate matter finer than 2.5 μm diameter

Exceedances of air quality standards in the US

https://www.epa.gov/air-trends/air-quality-national-summary

70 ppb (8-h average)

Latest ozone data from EPA

https://www.epa.gov/system/files/documents/2024-12/o3_irp-vol-1_final_1.pdf

Air quality standard (design value) is annual 4^th highest maximum daily 8-h average averaged over 3 years

0 20 40 60 80 100 120 ppb

Europe AQS

(seasonal)

U.S. AQS

(8-h avg.)

U.S. AQS

(1-h avg.)

Preindustrial

ozone

background

Present-day ozone background at northern mid-latitudes

Europe AQS

(8-h avg.)

Canadian AQS

(8-h avg.)

Air quality standards have been tightening with time…

…and now approach background tropospheric levels

2008

2014

1997

China AQS

(8-h avg.)

Days per year exceeding 70 ppb ozone standard, 2010-2014

TOAR [2020]

Very severe ozone pollution problem in China

Li et al. [2019a]

China

air quality standard

US

air quality standard

revealed by air quality network observations starting in 2013

Controlling ozone pollution requires understanding�of whether ozone production is NO_x- or VOC-limited

This is because anthropogenic sources are different:

NO_x is from fuel combustion

VOCs are from leaks, industry, domestic products

Until 1990s, ozone production in US was thought to be VOC-limited…

Typical

anthropogenic

emission inventory

estimate

…motivating strategy of VOC emission controls

Strong action to reduce VOC emissions

VOC concentration trends in Los Angeles

Warneke et al., 2012

1970-2003 trend of US emissions

Focus was on VOC controls,

but ozone was not decreasing except in

Los Angeles and New York. Why?

Fiore et al. [1998]

1980-1995 trend in summer ozone (ppb a^-1)

Arlene Fiore

Emission of biogenic VOCs first quantified in late 1980s

Many VOCs emitted as photosynthesis and metabolism by-products, insect attractors or repellents, response to stress and injury

Principal hydrocarbons are emitted as C₅H₈ units:

Isoprene (C₅H₈)

Terpenes (C₁₀H₁₆)

Limonene

Myrcene

α-pinene

β-pinene

Caryophyllene

Sesquiterpenes (C₁₅H₂₄)

50% of total biogenic VOCs

Atmospheric lifetime ~ 1h

Pat Zimmerman

Biogenic VOCs dominate over anthropogenic VOCs in summer �(ozone pollution season)

OMI satellite observations of formaldehyde (HCHO) columns, May-Aug 2005-2014

HCHO

hν , OH

oxidation

~ 1 hour

Zhu et al. [2017]

isoprene

~ 1 hour

Other VOCs

But what this really means is that ozone production is NO_x-limited

anthropogenic

emissions

…so emission controls should focus on NO_x

+ biogenic emissions

total

emissions

Rethinking the ozone problem (NRC, 1991)

Tells EPA that ozone is a regional pollution problem, advises NO_x emission controls

John Seinfeld

Jennifer Logan

US NO_x emissions have decreased steadily since 2000

US EPA, National Emission Inventory (NEI) 2011

Changes in Nitrogen Dioxide in the USA, 2005-2022 | Air Quality

Evidence of NO_x emission decrease from satellite NO₂ data

Russell et al. [2012]

Evidence of NO_x emission decrease from satellite NO₂ data

30% decrease in NO_x emissions from 2005 to 2011

Russell et al. [2012]

Take it from the communicator-in-chief…

US ozone trend

https://www.epa.gov/system/files/documents/2024-12/o3_irp-vol-1_final_1.pdf

Trend in #days/year with ozone > 70 ppb, �summer 2000-2014

Megacity trends

Parrish [2014], TOAR [2017], Wang et al. [2017]

Ozone in China has been increasing despite Clean Air Action starting in 2013

• Emission controls have been targeted (successfully!) at PM_2.5
• NO_x emissions have decreased by 30% since 2013 while VOC emissions have stayed flat.

Unlike in US, ozone production in China is VOC-limited

OMI annual mean tropospheric column data, 2006-2007

Glyoxal (CHOCHO) Formaldehyde (HCHO) NO₂

(440-460 nm) (340-360 nm) (420-450 nm)

NO_x emissions are very high and VOC emissions are largely anthropogenic

Chan Miller et al. [2016]

NO_x controls are needed to meet current ozone standards …�even if production is locally VOC-limited

start point

• NO_x controls are only way to get to current ozone standards and have side benefits

(NO₂ air quality, nitrogen deposition)

• VOC controls will only get you so far until you are limited by biogenic background

​

As ozone standard tightens, the nature of the problem changes

60 ppb exceedances are largest in Intermountain West

Ozone in Intermountain West originates out of N America

Background = ozone present in absence of anthropogenic sources in North America

• Domestic emissions have little influence on intermountain west
• Anthropogenic background contributes ~15 ppb with little day-to-day variability

Zhang et al. [2014]

downwelling

Mean profile over NE Pacific,

Apr-May 2006 (INTEX-B)

Pinedale, Wyoming

2.4 km altitude

Intercontinental transport of ozone pollution

2012 OMI NO₂ column, 10¹⁵ molecules cm^-2

Ozone pollution transport (GEOS-Chem model)

Strong westerlies

transport ozone

around mid-latitudes

N America

​

​

Europe

​

​

Asia