AIR POLLUTION
Temperature-dependent emissions dominate aerosol
and ozone formation in Los Angeles
Eva Y. Pfannerstill
1
*
†
, Caleb Arata
1
, Qindan Zhu
2,3,4
‡
, Benjamin C. Schulze
4
, Ryan Ward
4
,
Roy Woods
5
, Colin Harkins
3,6
, Rebecca H. Schwantes
6
, John H. Seinfeld
4
, Anthony Bucholtz
5
,
Ronald C. Cohen
2,7
, Allen H. Goldstein
1,8
*
Despite declines in transportation emissions, urban North America and Europe still face unhealthy
air pollution levels. This has challenged conventional understanding of the sources of their volatile
organic compound (VOC) precursors. Using airborne flux measurements to map emissions of a wide
range of VOCs, we demonstrate that biogenic terpenoid emissions contribute ~60% of emitted VOC OH
reactivity, ozone, and secondary organic aerosol form
ation potential in summertime Los Angeles and that
this contribution strongly increases with temperatur
e. This implies that control of nitrogen oxides is key
to reducing ozone formation in Los Angeles. We also s
how some anthropogenic VOC emissions increase with
temperature, which is an effect not rep
resented in current inventories. Air pollution mitigation efforts must
consider that climate warming will strongly change emission amounts and composition.
A
mbient air pollution is the fourth-ranking
human health risk factor globally (
1
),
leading to an estimated 4.2 million pre-
mature deaths per year (
2
). Important
pollutants causing cardiovascular and
respiratory diseases are
fine particulate matter
(PM
2.5
) and tropospheric ozone (
2
). Volatile
organic compounds (VOCs) are precursors
to both: A large fraction of PM
2.5
is secondary
organic aerosol (SOA) that forms through the
oxidation of VOCs (
3
). In the presence of ni-
trogen oxides (NO
x
) and sunlight, VOC oxida-
tion leads to ozone formation.
Ninety-nine percent of the world
’
s popu-
lation lives in places where the World Health
Organization air quality guidelines are not
met (
2
). This includes the US megacity of Los
Angeles, where ozone and PM
2.5
are frequently
at unhealthy levels, especially in the summer
(
4
,
5
). As in many industrialized cities, tech-
nologies such as efficient three-way catalytic
convertersandeffortsspurredbyregulation
led to a steep decrease in automotive VOC emis-
sions and thus to a decades-long decrease of air
pollutant concentrations (
6
). However, con-
centrations of ozone and PM
2.5
particle pollu-
tion have stopped decreasing since ~2010 (
7
,
8
).
Recent studies indicate the increasing rela-
tive importance of volatile chemical products,
which now contribute as much as half of ur-
ban fossil fuel VOC emissions in industrialized
cities (
9
,
10
). The relative contribution of the
biogenic VOC fraction must also have increased
with declining transportation emissions. As
a result of these changes in the emitted VOC
mixture over the course of a few decades, the
sources of secondary air pollution have been
called into question. For example, model-
observation comparisons have raised doubts
on whether the models correctly reproduce
theemissionsourcemixturecontributingto
SOA (
11
,
12
). The origin of SOA in Los Angeles
is under debate, with some studies reporting a
predominantly vehicular source (
11
,
13
), where-
as a temperature-dependent analysis of PM
2.5
concentrations and isoprene concentrations
indicated a major biogenic origin (
8
). The
percentage of days on which PM
2.5
exceeds
12
m
g/m
3
is <10% at 20°C and >40% at 30°C
and reaches 70% at 40°C (
8
). Similarly, the
likelihood of ozone exceedances (
≥
70 parts
per billion) is close to 0% at 20°C and >70%
at temperatures
≥
30°C, with temperature-
dependent biogenic emissions of reactive ter-
penoids suggested as one of the driving factors
(
14
). However, concentration-based temper-
ature dependencies may be influenced by me-
teorology instead of emissions because hotter
days tend to be more stagnant (
15
). This shows
the need for direct emission observations. With
climate change, an increase in the number of
days with high temperatures is expected (
16
,
17
).
Thus, it is important to understand how in-
creasing temperatures affect VOC emission
amounts and mixture and what this means
for secondary air pollutant formation, as well
as regulation strategies.
Spatially resolved direct airborne
measurements of VOC emission fluxes
Previous efforts to understand the magni-
tude and composition of VOC emissions in
Los Angeles, as in other megacities, have re-
lied on indirect methods
—
either by using tra-
ditional bottom-up emission inventories (
18
)
or by inferring emissions top down from con-
centration measurements with chemical trans-
port models (
10
). Both approaches are indirect
andrelyonarangeofassumptionsandthus
are subject to large uncertainties. To overcome
these limitations, we performed airborne eddy
covariance measurements to provide the first
direct observations of spatially resolved VOC
emissions in Los Angeles. Emission and depo-
sition fluxes were calculated from 10-Hz concen-
tration and vertical wi
nd speed measurements
by using continuous wavelet transformation
(
19
). State-of-the-art instrumentation [pro-
ton transfer reaction (PTR)
–
time-of-flight
mass spectrometry
–
time-of-flight mass spec-
trometry] provided a comprehensive range of
VOC species for which spatially resolved urban
fluxes were observed, including source-specific
tracers for biogenics, vehicle emissions, per-
sonal care products, and solvents, among others.
Nine flights were conducted in June 2021 be-
tween 11:00 and 17:00, with flight days selected
to cover a temperature range as wide as possible
(for maps of flight tracks, see figs. S1 and S5).
Median flight temperatures ranged from 23° to
31°C, with minima and maxima stretching from
15° to 37°C, respectively.
Ozone formation in Los Angeles is still sen-
sitive to VOCs, with recent analyses suggest-
ing that current NO
x
emissions need to be
reduced substantially (>50%) to move to a
NO
x
-sensitive ozone formation regime (
7
,
14
).
The contribution of VOCs to ozone formation
depends on the reacti
on frequency of each
VOC species with the hydroxyl (OH) radical, the
primary oxidant in the daytime troposphere.
This reaction frequency is referred to as VOC
OH reactivity (hereafter: OH reactivity) and is
calculated for emissions as
OHR
F
¼
k
OH
;
VOC
F
VOC
ð
1
Þ
where OHR
F
is the OH reactivity of the flux
in meters per second squared,
k
OH,VOC
is the
reaction rate constant of a VOC with the OH
radical in meter cubed per molecule per sec-
ond, and
F
VOC
isthefluxoftheVOCinmol-
ecules per meter squared per second. Because
the daytime boundary layer over Los Angeles
can be considered as a box to which emissions
are continuously added, emitted OH reactivity
can be considered a proxy for in situ ozone for-
mation potential withi
n the Los Angeles basin.
The contribution of VOCs to SOA formation is
more complex to quantify. We estimated aerosol
yields using the statistical oxidation model
(
20
) combined with a one-dimensional volatility
RESEARCH
1
Department of Environmental Science, Policy and
Management, University of California at Berkeley, Berkeley,
CA, USA.
2
Department of Earth and Planetary Science,
University of California at Berkeley, Berkeley, CA, USA.
3
Cooperative Institute for Research in Environmental
Sciences, University of Colorado Boulder, Boulder, CO, USA.
4
NOAA Chemical Sciences Laboratory, Boulder, CO, USA.
5
Department of Environmental Science and Engineering,
California Institute of Technology, Pasadena, CA, USA.
6
Department of Meteorology, Naval Postgraduate School,
Monterey, CA, USA.
7
Department of Chemistry, University of
California at Berkeley, Berkeley, CA, USA.
8
Department of
Civil and Environmental Engineering, University of California
at Berkeley, Berkeley, CA, USA.
*Corresponding author. Email: ahg@berkeley.edu (A.H.G.);
e.pfannerstill@fz-juelich.de (E.Y.P.)
†
Present address: Institute for Energy and Climate Research 8:
Troposphere, Forschungszentrum Jülich, Jülich, Germany.
‡
Present address: Department of Earth, Atmospheric and Plane-
tary Sciences, Massachusetts Institute of Technology, Cambridge,
MA, USA.
Pfannerstill
et al
.,
Science
384
, 1324
–
1329 (2024)
21 June 2024
1of6
Downloaded from https://www.science.org at California Institute of Technology on June 21, 2024