ACPAtmospheric Chemistry and PhysicsACPAtmos. Chem. Phys.1680-7324Copernicus PublicationsGöttingen, Germany10.5194/acp-17-9081-2017Global atmospheric chemistry – which air mattersPratherMichael J.mprather@uci.eduhttps://orcid.org/0000-0002-9442-8109ZhuXinFlynnClare M.StrodeSarah A.https://orcid.org/0000-0002-8103-1663RodriguezJose M.SteenrodStephen D.LiuJunhuaLamarqueJean-Francoishttps://orcid.org/0000-0002-4225-5074FioreArlene M.https://orcid.org/0000-0003-0221-2122HorowitzLarry W.MaoJingqiuhttps://orcid.org/0000-0002-4774-9751MurrayLee T.https://orcid.org/0000-0002-3447-3952ShindellDrew T.https://orcid.org/0000-0003-1552-4715WofsySteven C.Department of Earth System Science, University of California,
Irvine, CA 92697-3100, USANASA Goddard Space Flight Center,
Greenbelt, MD, USAUniversities Space Research Association (USRA),
GESTAR, Columbia, MD, USAAtmospheric Chemistry, Observations and
Modeling Laboratory, National Center for Atmospheric Research, Boulder, CO
80301, USADepartment of Earth and Environmental Sciences and
Lamont-Doherty Earth Observatory of Columbia University, Palisades, NY,
USAGeophysical Fluid Dynamics Laboratory, National Oceanic and
Atmospheric Administration, Princeton, NJ, USAGeophysical
Institute and Department of Chemistry, University of Alaska Fairbanks,
Fairbanks, AK, USADepartment of Earth and Environmental
Sciences, University of Rochester, Rochester, NY 14627-0221, USANicholas School of the Environment, Duke University, Durham, NC,
USASchool of Engineering and Applied Sciences, Harvard
University, Cambridge, MA 02138, USAMichael J. Prather (mprather@uci.edu)27July20171714908191027December201616January201725May20173July2017This work is licensed under the Creative Commons Attribution 3.0 Unported License. To view a copy of this licence, visit https://creativecommons.org/licenses/by/3.0/This article is available from https://acp.copernicus.org/articles/17/9081/2017/acp-17-9081-2017.htmlThe full text article is available as a PDF file from https://acp.copernicus.org/articles/17/9081/2017/acp-17-9081-2017.pdf
An approach for analysis and modeling of global atmospheric
chemistry is developed for application to measurements that provide a
tropospheric climatology of those heterogeneously distributed, reactive
species that control the loss of methane and the production and loss of
ozone. We identify key species (e.g., O3, NOx, HNO3, HNO4,
C2H3NO5, H2O, HOOH, CH3OOH, HCHO, CO, CH4,
C2H6, acetaldehyde, acetone) and presume that they can be measured
simultaneously in air parcels on the scale of a few km horizontally and a few
tenths of a km vertically. As a first step, six global models have prepared
such climatologies sampled at the modeled resolution for August with emphasis
on the vast central Pacific Ocean basin. Objectives of this paper are to
identify and characterize differences in model-generated reactivities as well
as species covariances that could readily be discriminated with an unbiased
climatology. A primary tool is comparison of multidimensional probability
densities of key species weighted by the mass of such parcels or frequency of
occurrence as well as by the reactivity of the parcels with respect to
methane and ozone. The reactivity-weighted probabilities tell us which
parcels matter in this case, and this method shows skill in differentiating
among the models' chemistry. Testing 100 km scale models with 2 km
measurements using these tools also addresses a core question about model
resolution and whether fine-scale atmospheric structures matter to the
overall ozone and methane budget. A new method enabling these six global
chemistry–climate models to ingest an externally sourced climatology and then
compute air parcel reactivity is demonstrated. Such an objective climatology
containing these key species is anticipated from the NASA Atmospheric
Tomography (ATom) aircraft mission (2015–2020), executing profiles over the
Pacific and Atlantic Ocean basins. This modeling study addresses a core part
of the design of ATom.
Introduction
To understand global atmospheric chemistry is to understand the chemical
heterogeneity of air parcels across the vastness of the troposphere (e.g.,
Fishman et al., 1996; Ehhalt et al., 1997; Marenco et al., 1998; Jacob et
al., 2003, 2010; Olson et al., 2004; Kunz et al., 2008; Nicely
et al., 2016). These air parcels are ephemeral, being continually created,
evolving, and mixed with others. Even the concept of discrete as opposed to a
continuum of air parcels is a conceit based in part on our modeling of the
atmosphere in quantized units such as gridded cells or 1 s averages. Yet,
the concept of distinct air parcels remains useful for parsing in situ
aircraft measurements and for the analysis presented here in which we ask
which air is more important for the chemical evolution of global tropospheric
pollution.
To understand the mix of chemicals in the atmosphere and where they come
from is to recognize how humans have perturbed the common air we breathe. We
seek knowledge of the photochemical evolution in each air parcel to
understand the overall impact of this heterogeneity and to interpret human
impact on past changes and predict future ones. The atmosphere's integration
of this chemical reactivity over the ensemble of such heterogeneous air
masses controls the evolution of air pollutants and reactive greenhouse
gases, particularly methane and ozone. Hence, it allows us to evaluate the
consequences of many atmospheric pollutants as regards global air quality
and climate.
We have a tendency to simplify this heterogeneity as global, hemispheric, or
even regional averages that can be represented with an average chemical
composition. This holds true especially when diagnosing the sources and sinks
of critical pollutants or when comparing models with atmospheric
measurements. Yet, chemistry inherently involves quadratic reactions of two
or more species and hence is non-linear – viz. the chemistry integrated over
a mix of parcels is not necessarily the same as that over the average of the
mix (e.g., Chatfield and Delaney, 1990). We have progressed in modeling
atmospheric chemistry over the past 4 decades from a few boxes (e.g.,
stratosphere and troposphere, northern and southern hemispheres) to
high-resolution gridded models with many millions of cells. These models
simulate myriads of air parcels that at times represent the observed
atmospheric heterogeneity of species composition. For example, Fig. 1a
presents a single-day snapshot of the column loss of methane as simulated by
the UC Irvine chemistry-transport model (CTM) at a resolution of 1∘ in
latitude and longitude. Even column averages over 24 h show a filamentary
structure with most of the tropospheric loss of methane occurring in sharp
synoptic patterns. These chemical patterns have similarities with the
atmospheric rivers of column water vapor (Newell et al., 1992; Dacre et al.,
2015; Mundhenk et al., 2016) in terms of filamentary appearance and being
dominated by the lower half of the troposphere. Nevertheless, the
methane-loss filaments do not coincide with atmospheric rivers (Fig. 1a vs.
1b), indicating that chemical-specific heterogeneity other than tropical
water vapor plays a role in these fine-scale structures (e.g., Ehhalt et
al., 1997; Browell et al., 2003; Charlton-Perez et al., 2009).
This heterogeneity of species and chemical reactivity (e.g., methane loss) is
clearly structured and not simply Gaussian. Its structure reflects the
combined influence of meteorological transport and mixing as well as the
patterns that different species are co-emitted and transformed around the
globe. For example, large plumes from industrial regions or biomass burning
when lofted into the free troposphere by deep convection or frontal systems
will naturally be sheared into laminae, travel long distances, and appear
ubiquitously (Newell et al., 1999; Stoller et al., 1999; Singh et al., 2000;
Blake et al., 2003; Heald et al., 2003; Damoah et al., 2004; Hecobian et al.,
2011; Wofsy et al., 2011). This shear or random strain in the atmosphere acts
to maintain the pollution concentrated within the layer and preserves the
sharp gradients relative to the neighboring atmosphere before they dissolve
into the surrounding atmosphere, e.g., Prather and Jaffe (1990), Thuburn and
Tan (1997), Esler (2003), and Pisso et al.2009). Characterizing chemical species
in the atmosphere as having mean abundances, or even mean vertical profiles,
with a standard deviation to represent the observed variability, does not
really describe how these models generate heterogeneity and how the different
species co-vary. Assuredly, the atmosphere has more processes and structures
than are in our current, high-resolution models as seen in Fig. 1, but the
extent to which these models represent the key processes shaping the observed
patterns is understudied.
Characterizing atmospheric measurements of this chemical heterogeneity
specifically for testing models is problematic. Simple direct comparisons of
atmospheric rivers, pollution or biomass burning plumes, and other
structures in the troposphere or stratosphere are difficult, even with
models using the historical meteorology and chemical emissions, because of
slight phase errors in the location of large-scale gradients or laminae
(e.g., Reid et al., 1998; Manney et al., 1998; Wild et al., 2003; Kiley et
al., 2003; Allen et al., 2004; Schoeberl et al., 2007; Elguindi et al.,
2010). The other type of chemistry models, the chemistry–climate models
(CCMs), are our means of understanding future air pollution (Prather et al.,
2003; Mickley et al., 2004; Jacob and Winner, 2009; Fiore et al., 2012;
Barnes and Fiore, 2013; Turner et al., 2013; Fang et al., 2013; Schnell et
al., 2015), but CCMs describe the chemical climate and not the hindcast of
specific chemical measurements. Most large CCM groups have parallel CTM
versions, but these forced-meteorology versions will likely have different
clouds, convection, and transport, changing the chemical climatology.
Aircraft campaigns often use photochemical box models to provide an
observationally constrained check on reactive species (Olson et al., 2004, 2012;
Apel et al., 2012; Stone et al., 2012), and more
recently these have extended the box model as a transfer standard across
CCMs–CTMs (Nicely et al., 2017) that can integrate reactive chemistry over 24 h.
The problem remains that the 24 h integration requires a global
model's diagnostics for the diurnal cycle of cloud cover and ozone–aerosol
influence on photolysis.
We describe a new approach for developing chemical climatologies from
atmospheric chemistry measurements and for using the major global 3-D
CTMs–CCMs as box-models to integrate the 24 h rates for important species
like methane and ozone. Our goal is to provide climatologies that can point
to specific patterns of the key chemical species whose initial values
control the chemical evolution of the air parcels. Knowing the correct
multi-species patterns, and how different models succeed or fail in
reproducing them, will give developers the largest leverage in improving the
chemical and physical processes in the models. A critical issue in preparing
such a chemical climatology is representativeness, i.e., just how well do
the observations represent the region in which they were made and how well
should the models match the space–time frequency of the observations. There
is growing literature on the issues of representativeness of atmospheric
measurements (Nappo et al., 1982; Crawford et al., 2003; Hsu et al., 2004;
Ramsey and Hewitt, 2005; Larsen et al., 2014; Eckstein et al., 2017)
including defining the chemical patterns through cluster analysis (Köppe et
al., 2009).
There have been many aircraft missions designed to provide a wealth of in
situ, high-resolution atmospheric chemistry data, including some with a
nearly complete package of key species needed to calculate reactivities
(Jacob et al., 2003, 2017; Engel et al., 2006; Pan et al.,
2017). Unfortunately, many of these produced a biased, non-climatological
sampling, for example, by chasing pollution plumes (Hsu et al., 2004) or by
measuring only in clear skies (Nicely et al., 2016). The Pacific Exploratory
Missions, PEM-Tropics and PEM-West, were notable in providing a mostly
unbiased, exploratory sampling of specific regions in the remote Pacific
with a full chemical payload measuring most of the key species (Hoell et
al., 1996, 1999; Raper et al., 2001; Davis et al., 2003). The
MOZAIC-IAGOS program uses in-service aircraft and has provided a unique
multi-year, objective climatology of some key species (O3, CO,
H2O) but only along major flight routes at cruise altitudes (8–12 km)
and at profiles above airports (Marenco et al., 1998; Thouret et al., 1998;
Kunz et al., 2008; Elguindi et al., 2010; Logan et al., 2012; Gaudel et al.,
2015).
We examine below some aspects of making objective climatologies of chemical
observations, in particular the representativeness of atmospheric transects
over the remote ocean basins. Our approach was designed specifically as part
of the current NASA Atmospheric Tomography (ATom) aircraft mission in which
the DC-8 is instrumented to make high-frequency in situ measurements of the
most important reactive species and flies down the middle of the Pacific and
Atlantic oceans, profiling as frequently as possible. The resulting
climatology represents the heterogeneity of the atmosphere, including the
covariance of key reactive species.
This approach is tested here using six CTMs–CCMs described in Table 1. It
allows us to identify models that look alike in reactivity statistics and
those that are distinctly different. We have seen large uniform anomalies in
a specific species as well as different patterns or locations of the most
reactive parcels. For example, we list the models' average reactivities for
the tropical Pacific and the globe in Table 1c. The tropical Pacific average
P-O3 is similar across models and is about 1/2 that of the global
average, which is dominated polluted, near-surface parcels over land. The
L-O3 is typically the same for the Pacific and the globe; and the L-CH4 is
greater over the Pacific than over the globe. Model D stands out in reversing
or exaggerating these typical Pacific vs. globe differences, indicating very
different locations for the reactivity. We use these models to demonstrate
the methodology and the ability to discriminate among them with ATom-like
measurements. Model versions used here should be considered snapshots in the
development cycle. No model tuning or development occurred as part of this
work, except to correct where quantities were missed or misdiagnosed. These
diagnostics need to be revisited for the model versions used in upcoming
assessments (e.g., Lamarque et al., 2013; Collins et al., 2016)
Typically, the probability of occurrence of a species' abundance is weighted
by the air mass of the parcel, but, if we are interested in the chemical
reactivity, then the parcel should be weighted by the chemical rates in the
parcel (e.g., moles per day). Such weighting is an obvious choice in that it
tells us which air parcels matter for chemical budgets, including, for
example, whether infrequently observed pollution plumes are responsible for
a large fraction of ozone production.
In Sect. 2 we define our use of reactivity in this paper (i.e., the
production and loss of ozone, the loss of methane) and identify about a
dozen key chemical species and other variables that once initialized
determine the chemical evolution of an air parcel. In Sect. 3 we show how
the CTMs–CCMs can be altered slightly to calculate the reactivity of air
parcels using the native grid cells of the model and a prescribed
initialization of the key chemical species. This approach allows the
CTMs–CCMs to be run using either model data or observations, or a mixture of
both. In Sect. 4 we derive multi-dimension probability distributions for
these key variables over a suitable latitude–longitude–pressure domain using
grid-cell values from several CTMs–CCMs. Tables of simplified statistics
describing these probability distributions are presented and discussed in
the Supplement to this paper. The full distributions and simple statistics
clearly show the basic differences in chemical heterogeneity and reactivity
across the six models. We conclude in Sect. 5 with a summary of the model
comparisons and what is learned from the new diagnostics. We also discuss
the ongoing NASA ATom mission (2015–2020), which will provide the air parcel
measurements of key species to initialize the models' calculation of
reactivity in each parcel and thus provide an observed climatology of the
chemical reactivity of the troposphere. This approach moves us towards an
understanding of which species exert the largest influence on the
atmosphere and thus which are most crucial for us to establish a
global climatology.
Chemistry-Transport and Chemistry-Climate models used in this work.
(a) Participating models ModelTypeDriving meteorologyYearModel gridEffective resol. 500 hPaCAM4-ChemCCMSSTs2000s0.47∘× 0.625∘× 52 L0.47∘× 0.625∘× 38 hPaGEOS-ChemCTMGEOS5-FP20132∘× 2.5∘× 72 L2∘× 2.5∘× 38 hPaGFDL-AM3CCMNCEP (nudged)2013C180L480.5∘× 0.5∘× 71 hPaGISS-E2CCMDaily SSTs prescribed, winds nudged to MERRA20132∘× 2.5∘× 40 L2∘× 2.5∘× 50 hPaGMI-CTMCTMMERRA20011∘× 1.25∘× 72 L1∘× 1.25∘× 38 hPaUCI-CTMCTMECMWF IFS Cy38r12005T159N80L601.1∘× 1.1∘× 38 hPa(b) Points of contact and model url's ModelPOCEmailModel urlCAM4-ChemJean-Francois Lamarquelamar@ucar.eduhttp://www.cesm.ucar.edu/models/current.htmlGEOS-ChemLee Murraylee.murray@rochester.eduhttp://wiki.seas.harvard.edu/GFDL-AM3Arlene Fioreamfiore@ldeo.columbia.eduhttps://www.gfdl.noaa.gov/am3-model/GISS-E2Lee Murraylee.murray@rochester.eduhttp://www.giss.nasa.gov/tools/modelE/GMI-CTMSarah Strodesarah.a.strode@nasa.govhttp://gmi.gsfc.nasa.govUCI-CTMMichael Prathermprather@uci.eduftp://halo.ess.uci.edu/public/xzhu/qcode_72c(c) Model average reactivities (ppb day-1) for P-O3, L-O3, and L-CH4 ModelCodeP-O3 L-O3 L-CH4 Tr. Pac.GlobalTr. Pac.GlobalTr. Pac.GlobalCAM4-ChemA0.9792.0701.9631.8021.0170.745GEOS-ChemB0.7912.2521.6161.8370.7650.738GFDL-AM3C0.8602.0361.5501.5350.7260.599GISS-E2D1.0923.7152.5893.4090.4530.693GMI-CTME0.7781.5131.8341.6900.8480.674UCI-CTMF1.0882.1001.7881.9900.8540.702All results are mass-weighted by tropospheric parcels up to 200 hPa from the model C-runs for 16 August. (d) Model references ModelCodeRelevant referencesCAM4-ChemALamarque et al. (2012); Tilmes et al. (2016)GEOS-ChemBBey et al. (2001); Eastham et al. (2014)GFDL-AM3CDonner et al. (2011); Naik et al. (2013a); Li et al. (2016)GISS-E2DSchmidt et al. (2014); Shindell et al. (2013)GMI-CTMEStrahan et al. (2007); Duncan et al. (2007)UCI-CTMFHolmes et al. (2013, 2014); Prather (2015); Sovde et al. (2012)
(a) Column tropospheric loss frequency (1 yr-1) for
CH4 and (b) column average H2O abundance
(g-H2O kg-air-1) taken from a 1-day integration
(16 January 2005) using the University of California, Irvine (UCI)
chemistry-transport model (CTM) run at T319N80L57 resolution
(∼ 1∘ horizontal) using forecast meteorology from the European
Centre for Medium-Range Weather Forecasts, see Sovde et al. (2012). As
expected, the northern winter shows very little CH4 loss above
40∘ N.
Key chemical species for tropospheric reactivity
The reactivity of an air parcel is defined here as a daily average of the
rates affecting critical species, in this case, ozone (O3); a greenhouse
gas and air quality threat; and methane (CH4), the second most important
emitted greenhouse gas after CO2. Methane is emitted mostly through
human activities but also naturally; and it is lost primarily
(> 80 %) through reaction with the hydroxyl radical (OH) in the
troposphere (Reaction 1). Other atmospheric losses in decreasing order of
magnitude and certainty are reaction with stratospheric OH, surface uptake by
biota, and reaction with Cl atoms (Prather et al., 2012; Ciais et al., 2013).
CH4+OH→CH3+H2O
The CH4 abundance varies little throughout the troposphere (∼ 10 %), and the destruction of CH4 occurs with a mean loss frequency of
∼ 0.1 yr-1 (see Fig. 1a). Here we focus on calculating the
tropospheric loss of CH4 by OH over 24 h (Reaction 1, designated
L-CH4) in units of ppb (nanomoles mol-air-1) per day. L-CH4 is highly
variable across parcels, and the integral of L-CH4 determines the
atmospheric lifetime of CH4 and the buildup of its emissions in the
atmosphere.
Tropospheric O3 has stratospheric sources and surface sinks, which
average to about 0.2–0.3 ppb per day, and much larger in situ photochemical
production and losses that average about 1.1–1.5 ppb per day (Stevenson et
al., 2006, 2013; Young et al., 2013; Hardacre et al.,
2015). The O3 abundance varies greatly throughout the troposphere, by a
factor of 10 or more, and its mean residence time is about a month (Stevenson
et al., 2006; Wu et al., 2007; Hsu and Prather, 2009). O3 is an
intermediate source of atomic O in many tropospheric reactions, and its net
production and loss is determined in the long term by the breaking and
reforming of the O–O bond originating with molecular oxygen. Chemical
reactions are traditionally grouped into production (P-O3,
ppb day-1),
HO2+NO→NO2+OH,RO2+NO→NO2+RO,whereNO2+hν→NO+OandO+O2→O3,O2+hν→O+O(×2),
and loss (L-O3, ppb day-1).
O3+OH→O2+HO2,O3+HO2→HO+O2+O2,O(1D)+H2O→OH+OH,whereO3+hν→O(1D)+O2.
In the troposphere, Reaction (2d) is important only in the tropics above
12 km (Prather, 2009). The true P minus L of O3 includes a large number
of other reactions, particularly involving oxides of nitrogen and
hydrocarbons; but throughout the remote troposphere (i.e., away from fresh
pollution sources), Reactions (2) minus Reactions (3) accurately approximate
the true P–L that the models calculate using the full set of reactions. One
reason for separating P and L in this way is to think of P as independent of
O3 and L as being linearly proportional. Unfortunately, while the P
Reactions (2) have no obvious O3 terms, both these reactions and the OH
and HO2 abundances in Reactions (3) depend indirectly on O3; and
thus with a true linearization of P–L, the lifetime of O3 is much
shorter than inferred from L (Prather and Holmes, 2013). A similar chemical
feedback with opposite sign occurs for CH4 whereby the lifetime of a
CH4 addition is longer than inferred from the linear relationship of
Reaction (1) (Prather, 1996). We retain these definitions of P-O3,
L-O3, and L-CH4 because they still represent the reactivity in remote
regions and the reaction rates, rather than a linearization, are
straightforward CTM–CCM diagnostics.
We define the reactivity of an air parcel (Reactions 1–3) in terms of 24 h
average rates and hence the units of ppb per day. Reactivity defined here
requires sunlight; nighttime sources of OH from alkenes and isoprene via
ozonolysis or nitrate radicals (Paulson and Orlando, 1996) are important
primarily in continental air over emission sources. This calculation
integrates over the diurnal cycle of photolysis rates driven by changing
solar zenith angle, clouds, O3, and aerosol profiles, all of which are
simulated in CTM–CCMs.
What key constituents are needed for modeling reactivity? Models simulate
many tens to hundreds of chemical species. While many are important for
calculating the instantaneous reaction rates, e.g., O(1D), they are not
the key species. Key is defined here as a constituent whose initial value
significantly affects the 24 h reactivity, whereas other species can be
initialized to any reasonable value and not affect it. For example, OH and
HO2 are radical HOx species whose
abundances directly determine the rates of Reactions (1)–(3). Nevertheless,
these are not key species as their abundances can be initialized to zero and
are rapidly reset in seconds to a temporary steady state with first sunlight
or changing clouds through Reactions (3c, d) among others (Rohrer and
Berresheim, 2006). This argument applies to similar radical species such as
CH3OO, but not to HOx sources like CH3OOH and HOOH whose initial
values will control the abundance of OH and the reactivities over the day.
A similar situation applies to NO and NO2 (collectively designated
NOx), whereby total NOx changes over the day as it is exchanged with
higher oxides of nitrogen, but the fraction of NOx in the form of NO is
determined rapidly in sunlight by Reactions (2a), (2b), (2c), and (4).
NO+O3→NO2+O2
In the dark, NOx is almost entirely NO2, and it is critical to
initialize NOx, but not NO and NO2 separately.
Based on sensitivity tests with the UCI-CTM, our list of 18 key species
includes O3, NOx,
HNO3, HNO4, PAN (C2H3NO5= peroxyacetyl nitrate), RNO3 (CH3NO3 and all
alkyl nitrates), HOOH, ROOH (CH3OOH and smaller contribution from
C2H5OOH), HCHO, CH3CHO (acetaldehyde), C3H6O
(acetone), CO, CH4, C2H6, alkanes (all C3H8 and
higher), alkenes (all C2H4 and higher), aromatics (benzene +
toluene + xylene), and C5H8 (isoprene + terpenes).
We also add p (hPa), T (K), q (g-H2O kg-air-1), and latitude and
longitude to make up the 23 key variables in each air parcel. Some
collectives like alkanes may be treated as multiple, separate species in some
models or may be lumped according to their reaction rates.
The abovementioned list tends to
be inclusive because, for much of the troposphere, a smaller list can apply.
For some species (e.g., C5H8), their role is key only if they are
present in large enough abundances, but even when sampling across the Pacific
Ocean basin one may find plumes with recent biospheric sources.
This simplification of the chemical system fails in regions of intense
emissions of short-lived species or in highly polluted environments such as
urban, industrial, or open fires. After pollution plumes have been separated
from sources and aged a few days, our key variables should define the
reactivity. Such conditions apply to most of the troposphere, particularly
the air over the vast Pacific and Atlantic Ocean basins. With aged pollution
plumes, we expect that some key species (e.g., alkenes, isoprene, aromatics,
and higher alkanes) will drop off the list because their abundances in much
of the remote mid-ocean regions will have fallen below the relevance
threshold.
Modeling the reactivity of air parcels
Why use the global models instead of single-box models to calculate
reactivity statistics? There are several reasons. For one, these CTMs–CCMs
simulate the full meteorology including cloud cover and its variation over
large regions, which is a critical component of reactivity. Second, they
usually include self-evaluated ozone and aerosol profiles also needed for
the photolysis rates. Third, these models automatically simulate the diurnal
cycle in radiation at all seasons, latitudes, and longitudes. And fourth,
most importantly, these models have built-in chemistry modules that already
calculate reactivities, and they are the ones we rely on for climate and air
quality assessments. The goal here is to test their simulated chemical
heterogeneity. While a box model could be designed (using 3-D meteorology) to
address the first three needs (e.g., Nicely et al., 2016), it cannot address
the last. More simply, all the necessary Earth system components are already
built in to the CTMs–CCMs, and our approach of testing the modeled
climatologies includes that of testing the Earth system components (e.g.,
emissions, transport, chemistry, scavenging, air–sea exchange, and
land–surface interactions).
In a standard CTM or CCM simulation (defined here as a C-run), we calculate
the reactivity at a given grid cell, but not that of a parcel. Air parcels
move, change location, and mix with neighboring parcels: i.e., there is no
way to track quantitatively what might be considered the original parcel.
Effectively, we keep integrating the rates in that grid cell as different
parcels travel through it and are mixed within it. Let us take a large enough
domain of grid cells (e.g., tropical Pacific, 150–210∘ E, from
surface to 200 hPa) and calculate the statistical distribution of
reactivities of all those grid cells. We take these statistics to be
equivalent to those we would get from integrating the reactivity over
isolated air parcels with the same initialization. Of course the latter is
only a thought experiment since the parcels do not remain isolated. In C-runs
new air parcels are entering the domain and others are exiting. In a single
cell we can start with a polluted lamina and end with clean air convected
from the marine boundary layer, but much of the polluted lamina remains in
the larger domain. As long as the domain retains a statistical mix of the key
chemical species similar to the initialization, then the reactivity
statistics of the C-run should represent the hypothetical reactivity of those
initialized parcels.
How can we design a calculation using the CTMs–CCMs that allows us to
initialize a subset of grid cells with observed air parcels and then
calculate a reactivity for those parcels? The goal here is to be able to use
the NASA ATom aircraft mission (2015–2020), which was designed to measure
those 23 key variables in air parcels profiling from near-surface to 12 km
altitude, flying ascents and descents down the middle of the Pacific and
Atlantic Ocean basins. Thus ATom data will not fill the global 3-D model
grid, and thus many cells will be initialized with the model's original
chemistry values. The critical design requirement is that we let the model
integrate for 24 h as it normally does in a way that the chemistry in each
grid cell depends minimally on any of the grid cells around it.
We thus propose an A-run mode (named after the ATom mission) for the CTMs–CCMs
in which individual parcel reactivities can be calculated, albeit with some
simplifying approximations. Consistent with our definition of reactivity, we
consider only ATom parcels that are tropospheric. The A-runs disable
processes that connect and mix air parcels. First we drop all calls to the
tracer transport sections (advection, convection, diffusion, boundary layer
mixing). Second, we must cut all emissions, including lightning and aircraft
NOx, because without transport the emissions would build up
unrealistically in the source cells. Third, all tracer scavenging modules
must be turned off because in many models the scavenging depends on the
vertical distribution of the species.
In this A-mode, the remaining connection of the reactivity calculation with
neighboring grid cells is through the photolysis rates, which require
profiles of clouds, aerosol layers, and ozone. It is impossible to prescribe
all these data over the diurnal cycle for each parcel from observations, and
thus we must rely on the CTM–CCM to generate a suitably realistic, diurnal,
regional, seasonal climatology for these and hence the photolysis rates. To
better average the reactivity over synoptic variations in clouds, we expect
to repeat the same initialization of the A-runs for a range of days over a
month containing the observations.
Each ATom parcel (2–8 km along the flight path) will be assigned a unique
model grid cell to best match the observation: latitude and pressure grids
containing the measurement, and longitude chosen as close as possible but
maintaining a unique cell for each parcel. ATom parcels in adjacent grid
cells may represent air masses separated by a few km instead of the
grid-cell size of order 100 km. A high density of ATom parcels in a region
will be placed in the correct latitude and pressure cells but may be strung
out in longitude cells. The parcel will use the mole fraction of key species,
water vapor (q) and temperature (T) as measured, but will adopt the mean
pressure of the grid cell. The model may need to maintain separate storage
for the hourly T and q used in the CCM dynamics because it is important to
maintain the clouds as they would be done in the C-run, and thus the
main-code values of T and q cannot be overwritten with ATom values. The A-run
treatment of stratospheric O3 (i.e., fixed) is unlikely to be identical
to the C-run, but it does not appear to drive major changes in the average
photolysis rates over a region (see below).
In defining the A-runs, we have thus created some biases in the reactivities
relative to the C-runs. Examination of the NOx and HOx budgets of parallel
A- and C-runs shows two obvious differences. The A-runs lack emissions.
Over the remote ocean basins, the most important emission is NOx
(lightning, shipping, aviation). Thus A-runs show a 24 h decline in NOx
abundances compared with the C-runs, resulting in generally lower P-O3.
The A-runs also lack scavenging and thus accumulate more HNO3 and
HOx precursors like HOOH, affecting L-CH4. No
other simple objective approach has been found, and we must accept and
document these biases in the A-runs.
An examination of how the A- and C-runs differ is shown in Fig. 2 using the
UCI-CTM's 1-D probability distributions of six key species (NOx,
HNO3, HNO4, PAN, HCHO, HOOH) for the central tropical Pacific. The
initial distribution for both runs (12 h local solar time at
180∘ E, black solid) can be compared with that for 24 h later (36 h)
for the C- (black dashed) and A-runs (cyan squares, only for four species). The
number of moles at the beginning and end of the 24 h in the C-run (see
legend) is a measure of the daily changes in the air parcels entering and
leaving the domain. It varies from 0 to 4 %, which is well within the expected
representativeness of a given day. With the A-run, however, we see large
systematic shifts due to the lack of emissions (NOx) and scavenging
(HNO3, HOOH). For HNO3 the content increases overall by 9 %,
with the high-end (> 100 ppt) distribution not changing, but the
low-end (< 20 ppt) air gains HNO3, increasing the middle
section (20–100 ppt). This is logical because the low-HNO3 regions
have the most scavenging. This change in distribution over the 24 h
integration of the A-runs is unlikely to change the reactivities as the
release of NOx from HNO3 will be more important in the high-HNO3
regions. For NOx the content decreases overall by 18 %, with most air
parcels (4–100 ppt) becoming less frequent and an increase in frequency
only for parcels with very low NOx, < 4 ppt. The 1-D
distribution of HCHO shifts lightly but with little overall change in
content. The lack of scavenging is even more important for HOOH with an
overall increase of 41 % and a dramatic shift in the distribution:
decreases in 0.3–1.0 ppb appear as very large increases from 1.0 to
2.5 ppb. The implications for using the A-run bias in computing the
reactivities are examined with all six models below.
An important assumption in using key species to initialize the reactivity
simulations is that the diurnal cycle is not critical, and ATom
measurements can be used without trying to make corrections for the time of
measurement. In running these global models, it is not practical to
initialize parcels at other than a standard day (i.e., beginning at 00:00 UT).
For some species like HCHO, the daytime loss frequency in the tropics is
about 1/2 h-1 (see for example loss photolysis rates for various
oxygenated hydrocarbons in Prather, 2015), and thus one might expect it to
vary greatly over the sunlight day or with cloud variations. The diurnal
change in 1-D distributions of the 6 key species is also shown in Fig. 2 for
the C-runs at 18 h (local solar time, red dashed), 24 h (dark blue dashed),
and 30 h (green dashed). The C-runs are in approximate steady-state over the
tropical Pacific domain as seen by comparing 12 h with 36 h, and thus these
sunset–midnight–sunrise times show the daily variations. The diurnal cycle
does produce visible shifts in the 1-D distributions, particularly at the end
of the night (30 h). The shifts in HCHO are small considering its high loss
frequency, primarily because both sources and sinks respond similarly to
photolysis rates. The seemingly longer-lived HOOH shows larger shifts because
production occurs in sunlight but scavenging occurs day and night. PAN and
HNO4 show small diurnal cycles at the high abundance end of their
distributions where they can be important NOx sources, and initialization
errors caused by the diurnal cycle at the low abundances will have smaller
impacts on reactivity.
A test of A- vs. C-runs for all six CTMs–CCMs is shown Fig. 3. All models
were spun up for a year and stopped at 00:00 UT on 16 August, with the
chemical abundances at this time being used to initialize each model's own C-
and A-runs. In this case all species in the model were initialized and not
just the 18 key species. Each model ran their own chemistry and meteorology
intended to simulate a specific historical year or a typical climate year.
All were intended to be typical of the last decade. The models were then run
for 24 h and the rates and reactivities diagnosed for both C-runs and
A-runs. All models have different resolutions, ranging from 0.5 to
2∘. All model statistics (key variables, reactivities, plus 24 h
average photolysis rates) were stored globally. This analysis examines a
north–south transect flight over the Pacific Ocean basin as in the NASA ATom
flights but greatly expands the region to include more grid cells: six
domains with latitude boundaries at 60–40∘ S, 40–20∘ S,
20∘ S–0∘, 0–20∘ N, 20–40∘ N, and
40–60∘ N
(each region is
color keyed in Fig. 3); longitude, in a single broad domain,
150–210∘ E. Vertical profiles (200–1000 hPa) on the models' native
grid are shown for the six domains as different colors. The standard C-runs
with all transport and emissions included are solid lines, while the
ATom-like A-runs are dashed.
For L-CH4, the only general agreement is the lesser importance of parcels
at altitudes above 500 hPa. For this August test, most models find that the
20–40∘ N dominates (note that plots are ppb day-1 and not area
weighted), and the 60–40 and 40–20∘ S domains are
least important (similar to OH structures in Spivakovsky et al., 2000;
Lawrence et al., 2001). Most models show increasing L-CH4 in the first few
km above the ocean because of low-level clouds shifting photolysis to the
middle troposphere The results for L-O3 show similar patterns of agreement
and disagreement among models but emphasize the dominant role of the middle
troposphere (500–800 hPa) for O3 loss. P-O3 has distinct patterns,
demonstrating the importance of larger NOx values in the upper
(200–500 hPa) and lower troposphere (800–1000 hPa), presumably from
lightning NOx. Only GMI-CTM lacks lower troposphere sources of O3 at
about 180∘ E. Overall the models show modest, similar amplitudes (but not
always sign) in the bias of A-runs relative to C-runs. Thus we can use the
model A-runs to tag each parcel in the ATom measured climatology by its
reactivity in the absence of emissions and transport. Clearly these models
have largely different chemical climatologies for the middle of the Pacific,
and, with the ATom climatology to initialize all six models, we will be able
to test whether these differences reflect the initial key species and/or the
photochemical components.
Photolysis rates (J-values) are the driving force for reactivity, and we
include also a comparison of the 24 h average J's (Reactions 2d and 3c) in
Fig. 4. The model spread in J-NO2 is 20 % and likely due to
differences in cloud cover as well as the photolysis module in the model. The
wide, factor-of-2 range in J-O3(1D) cannot be simply explained
through differences in clouds or ozone; for example, a 20 % reduction in
column O3 gives only a 33 % increase. Such differences will drive a
large part of the model differences seen in Fig. 3. For example, the large
J-O3 for GISS, and hence large production of OH, can explain in part
why GISS has very large L-O3 and P-O3, but not why the L-CH4 (also
dependent on OH) matches the other models. Surprisingly GEOS-Chem has an even
larger J-O3 but its reactivities are within the range of the other
four models. A comparison between the A- and C-runs (not shown) confirms that
these two runs have almost identical J's as expected since these changes in
ozone and aerosols over 24 h between these two simulations will have a small
impact on regional average J's.
While the A-run is clearly asking the modeling groups to make some rather
uncomfortable code modifications, these tend to be at the very high level of
disabling entire components. Other approaches for indirectly comparing
chemical models without transport have been developed (e.g., neural networks
in Nicely et al., 2017). We choose the A-run approach as it will allow us to
more directly compare modeled reactivities based on the primary CTM–CCM
coding and still allows for all models to be initialized with the same
chemical composition.
One-dimensional probability distributions for HNO3, NOx, HNO4, PAN,
HOOH, and HCHO from the UCI-CTM. The domain sampled is the tropical Pacific:
20∘ S–20∘ N, 150–210∘ E, 0–12∘ km, on
16 August. The units are moles of air per log-scale bin (20 bins per factor
of 10). The area under the curve in the log plot is the air mass of the
domain, except for HNO4 and PAN for which there are numerous
observations below the cutoff at 0.1 ppt. Five different times are shown for
the C-run: local noon (12 h), sunset (18 h), midnight (24 h), sunrise
(30 h), and the following noon (36 h). Also shown is the A-run at noon
(12 h, same as C-run) and the following noon (A 36 h). The numbers of moles
of the species in the domain are given in the legend.
Profiles of reactivity (ppb day-1) for loss of CH4
(L-CH4, top panel), loss of O3 (L-O3, middle panel), and production of O3
(P-O3, bottom
panel) from six global models (Table 1). Cells from each model grid are
averaged over 20∘ latitude domains (different colors, see legend),
longitudes from 150 to 210∘ E, and for the single day of 16 August.
Years vary by model, see text. Solid lines are standard model simulations
(C-runs) with the values representing air that passed through the cell over
24 h. Dashed lines are the no-transport, no-emissions A-runs that keep the
initialized chemical values in the same cell over 24 h.
Modeled 24 h average J-values for
O3+hv= > O(1D) + O2 (a, s-1)
and NO2+hv= > NO + O (b, s-1) for the
tropical Pacific (20∘ S–20∘ N, 150–210∘ E). See
Fig. 3 and Table 1 for model codes.
Probability distributions of species and reactivities
We characterize the heterogeneity in tropospheric chemistry through the
joint-probability distributions of the frequency of occurrence of chemical
species in air parcels for the six models here. These diagnostics are readily
suited to high-frequency in situ observations from an extensive aircraft
mission such as ATom, for example see Köppe et al. (2009). This paper then
takes a novel approach by focusing on the chemical budgets for tropospheric
ozone and methane. In addition to weighting a parcel according to its
occurrence or parcel mass, we include a factor that accounts for the
model-calculated reactivities of that parcel. For example, the basic weight
of a parcel (moles of air) can be scaled by P-O3 (ppb day-1), and the final
weight is the moles-O3 day-1. In this case the sum of weighted parcels in
a region gives the moles of O3 produced per day in that region. These
reactivities can be calculated with A-runs for both models and measurements.
Thus, the modeled and measured probability distributions reflect the parcels
most important in determining the chemical budgets in these models and hence
the evolution of the atmosphere.
Given the number of key species, the joint-probability distributions are
multidimensional, but for the most part we view them in 1-D or 2-D graphs. There
is a history of comparing models and measurements using such graphs (Hoor et
al., 2002; Hsu et al., 2004; Engel et al., 2006; Pan et al., 2007; Strahan
et al., 2007; Parrington et al., 2013; Gaudel et al., 2015). Often the goal
is simply to define a linear correlation, but in many cases a line fit
simply does not describe the heterogeneity (Köppe et al., 2009).
A much more difficult problem is that of representativeness: i.e., how much
of the Pacific basin must one sample to get joint-probability distributions
similar to that of the whole basin? Can aircraft-measured heterogeneity be
compared with models that do not follow the exact flight route for the exact
period of measurements (e.g., Hsu et al., 2004)? This latter question is
critical if we are to use the ATom measurements to test such a wide variety
of CTMs–CCMs. Here, we consider an idealized test case for representativeness
where we sample a model as objectively as possible and then compare with
different sampling “paths”.
One test of representativeness looks at the reactivities sampled along a
single longitude and then integrated over latitude–pressure domains. For
example, Fig. 1a clearly shows that the instantaneous column integrated L-CH4
varies greatly along longitude transects in the mid-Pacific. The
point-to-point variance in 3-D will be very large, but, if we average over
regional domains, can we achieve a representative mean value for reactivity?
Based on the profiles of reactivity (Fig. 3), we take three pressure domains
(boundaries at the surface, 850, 500, and 200 hPa, but with stratospheric
values screened out by model-designated discriminators) and three latitude
domains (60–20∘ S, 20∘ S–20∘ N, and
20–60∘ N). The means (ppb day-1) and standard deviations
(ppb day-1) of single-longitude sampling across the mid-Pacific
(155–233∘ E) on 16 August are shown for the UCI-CTM in Table 2
along with the standard deviation (in %) over the 31 days of August of the
daily full-domain average. The standard deviations are a measure of the
representativeness of the sampling, by longitude or by day. For L-CH4, the
dominant mean loss, > 1 ppb day-1, is in the
surface–500 hPa in the tropics and summer (northern) mid-latitudes as seen
in Fig. 3. For these regions the standard deviation across the longitudinal
samples is of order 6–11 %, whereas outside of these, it is as large as
20 %, but the absolute values are small. A similar pattern holds for L-O3
with standard deviations in dominant regions of 6–14 %. Thus any single,
fully sampled longitudinal transect through this domain has a 68 %
likelihood of being within 6–14 % of the mid-Pacific average. The
variance of P-O3 is slightly larger, 8–17 %, in part because P-O3
depends on the less-frequent high-NOx regions. Assembling a representative
sampling of P-O3 at the same % level as L-O3 will be slightly more
difficult. Such single-transect representativeness is about as good as we can
expect. Thus, model–model differences comparing individual transects from
each model would not be significant unless they exceed these percentages.
Averaging over the basin and/or several days should resolve model differences
at finer scales. The day-to-day standard deviation for the mid-Pacific
averages in Table 2 is shown in percent; it is smaller than across individual
longitudinal transects for a given day; and in key regions (surface to
500 hPa, 20∘ S to 60∘ N) it ranges from 1–4 % for
L-CH4 to 2–8 % for P-O3. A remaining question (not resolved with the
datasets assembled here) is the year-to-year variance of basin-wide
reactivities perhaps associated with the El Niño–Southern Oscillation.
The six models' 1-D probability distributions for O3, CO, NOx, and
HCHO over the tropical mid-Pacific basin are shown in Fig. 5, and simple
statistics (mean ± SD) are presented in the Supplement
Table S5. Modeled data are sampled on the native grid of each model and not
interpolated. This approach readily allows us to compare different models.
Both 1-D and 2-D distributions presented here are sorted into 20 log-spaced
bins per each factor of 10 (decade) in abundance (ppb or ppt). The dashes in
the upper/lower rows of Fig. 5 indicate widths of these bins on each plot.
For example, NOx distributions cover more than 3 decades (very small
dashes), while the CO covers less than a decade (wide dashes). In the first
row labeled “AIR”, each grid cell is weighted by its size in moles, and
thus the plot shows petamoles per logarithmic bin. In each subsequent row,
the cells are weighted by the reactivity (L-CH4, L-O3, P-O3) in
moles day-1, plotting thus megamoles per day per bin.
The AIR plots show clear model differences. Models A and B have much greater
frequency of O3 occurrence from 50–150 ppb, and half the models (B, D,
E) show a reasonable frequency of O3 at 10 ppb and less, as might be
expected in the tropical Pacific boundary layer (Kley et al., 1996; Singh et
al., 1996; Nicely et al., 2016). For CO, model A shows unusually low
abundances. For NOx, models C and F lack the NOx below 2.5 ppt that
others have. The models are quite similar for HCHO, except for D, which has
an unusually symmetric distribution and much lower abundances. When
reactivity weighted, new features are found. Note that the area under the
AIR-weighted curve is the same for all models, but the area in
reactivity-weighted 1-D plots is each model's total reactivity
(moles day-1). Model D has lower values overall for L-CH4 compared
with the other models, but it is similar or even slightly higher for L-O3
and P-O3. The high-O3 abundances in A remain equally important when
weighted by any reactivity, but those in B become less important for L-CH4
and L-O3, but even more important for P-O3. This unusual feature adds a
new dimension to diagnosing and understanding model differences. The
reactivity weighting of the CO distribution does not show anything unusual.
The NOx 1-D plots show that L-CH4 is more heavily weighed to low NOx
values than is L-O3, but P-O3 is weighted strongly to the higher NOx
abundances (> 10 ppt) as expected. The HCHO reactivity weights
in the opposite direction with high abundances (> 200 ppt)
favoring L-CH4 and L-O3 but lower ones favoring P-O3, probably
because the lower ones are from the upper troposphere where colder
temperatures suppress both L-CH4 and L-O3 (Fig. 3). The results from
the full probability distribution (Fig. 5) are mostly represented in the
central statistics of Table S5. The reactivity weighting adds a new dimension
to the diagnostics, and after the ATom dataset becomes available it would be
productive to make a more detailed comparison that identifies the location
and other key species controlling these shifts in reactivity.
Representativeness of reactivities (L-CH4, L-O3, and P-O3 – all
in ppb day-1) averaged over three latitude and three pressure domains over the
central Pacific (155–233∘ E). The first standard deviation (ppb day-1) is
over the different longitudinal transects on mid-August; and the second
(%) is for the average across longitudes sampled over 31 days of August.
Results are from the UCI-CTM C-runs for 16 August and
1–31 August. The 155–233∘ E domain includes 69 longitudinal
transects. All tropospheric grid cells in the domain are sampled equally and
weighted by mass. The period 1–31 August shows trends in some domains as the
sun moves southward, and this was removed with a line fit to calculate the
standard deviation over the month. Results for the A-runs (not shown) differ
in mean and standard deviation by a few percent.
Six modeled 1-D probability distributions for O3, CO, NOx,
and HCHO, where the air parcels have been weighted by air mass (row 1),
L-CH4 (row 2), L-O3 (row 3), and P-O3 (row 4). The domain being
sampled in the models is the tropical Pacific:
20∘ S–20∘ N, 150–210∘ E, 0.5–12 km. Units for
the air weighting are petamoles per bin where the bins are set at 20 per
decade (sizes marked by dashed lines in upper or lower panels) and Mmoles per
bin per day for the reactivity-weighted plots (rows 2–4).
(a) Six modeled 2-D probability distributions for NOx
vs. HOOH as weighted by air mass. These are the initial chemical abundances
for each model and hence the same for A- and C-runs. All grid cells were
binned at 20 per decade in species abundance (mole fraction, ppt for NOx,
ppb for HOOH). The density value for each plot is scaled so that a uniform
distribution over exactly 1 decade in both species would give the
yellow-green color of 1.0. The domain being sampled in the models is the
tropical Pacific: 20∘ S–20∘ N, 150–210∘ E,
0.5–12 km. Model A = CAM4-Chem; B = GEOS-Chem; C = GFDL-AM3;
D = GISS-E2; E = GMI-CTM; F = UCI-CTM. (b) Six modeled
2-D probability distributions for O3 vs. H2O as weighted by air
mass. This color-bar scale differs slightly from other 2-D plots. See
(a).
Model E (GMI CTM) 2-D probability distributions from A-run for
NOx vs. HOOH as weighted by air mass, L-CH4, L-O3, and P-O3. The
domain being sampled is 20∘ S–20∘ N, 150–210∘ E,
0.5–12 km, see Fig. 6a.
AIR-weighted 2-D probability distributions for NOx vs. HOOH
averaged over tropical Pacific block (150–210∘ E,
20∘ S–20∘ N, 0–12 km) and for different single-longitude
transects from 150–210∘ E, shown for models (a) C
(GFDL-AM3) and (b) F (UCI-CTM). The fitted 2-D ellipses are shown
for the full block (thick black line) and five longitude transects (colored
lines) for models (c) C and (d) F. The block ellipse for
the other model is shown as a thin black dashed line.
ATom proposed flight tracks (a) and estimated sampling
frequency by 1 km altitude bins (b). The actual flights are
somewhat altered. The altitude sampling is based on the proposed
∼ 90 h of flight time, ∼ 180 profiles taking ∼ 35 min for
each pair of climb-descend, and 5 min spent each profile in the marine
boundary layer. For up-to-date information on the ATom mission and
deployments, see https://espo.nasa.gov/missions/atom/content/ATom.
These new diagnostics do not instantly identify the cause of model
differences, but they do add a new dimension. For example, if we seek to
understand why model D is different, we can look at global budgets: both
models A and D have P-O3 and L-O3 tropospheric means between 2.5 and
3.5 ppb day-1, whereas the other four models have values between 1.0 and
2.0. The global L-CH4 – 0.50 to 0.65 ppb day-1 – is similar for
all models, with D in the middle. So globally, models B and D are similar,
but,
in the mid-Pacific, they are distinct with model D having much lower L-CH4
values in the tropics and especially the lower tropics (Fig. 3, see also
Fig. S1 of Naik et al., 2013b). CH4 loss is a major source of HCHO in
the unpolluted atmosphere and this may partly explain D's lower values of
tropical HCHO compared with other models. Some of the reduced tropical
reactivity in D may be caused by more low clouds in the tropics, and this is
apparent in the more rapid fall off in J-O3(1D) compared with other
models (Fig. 4); yet models B and D (not A and D as found in L-O3 and
P-O3) have much higher values of J-O3(1D). With the ATom A-run
approach we will be able to remove differences caused by the widely ranging
chemical climatologies of species (e.g., seen in Figs. 5, 6, 8) and more
directly trace the range of results to the models' basic photolysis and
kinetics.
The 2-D distributions simply weighted by AIR show remarkable structures that
differ significantly across the models, as shown in Fig. 6, with summary
statistics in Table S6. All 2-D plots use the same 20-per-decade log scale as
in the 1-D analysis, and they are normalized such that if all parcels are
distributed uniformly within a 20 × 20 square (e.g., 0.1–1.0 ppb HOOH,
10–100 ppt NOx) the arbitrary density value would be 1 (a yellow-green
color in Figs. 6–7). Thus, the reactivity-weighted 2-D plots are renormalized
and do not reflect the individual model's total reactivity. In Fig. 6a the
AIR-weighted NOx–HOOH plots show a boomerang structure with greatly
varying degrees of concentration about some points in the center (reddish
regions). For example, models A and D show a very diffuse distribution with a
much wider spread in HOOH values at lower NOx. Even for the four models
with a central (NOx, HOOH)-line defining a peak frequency of occurrence,
this line occurs at different locations. The O3–H2O density plots
(Fig. 6b) show examples of highly standard and well-measured species with
extreme distributions: O3 fall within 1 decade throughout most of the
troposphere, but H2O easily spans 3. Several show the bimodality of many
parcels with low O3 with high H2O (marine boundary layer and above)
and a second peak at higher O3 and dry. For example, C and E look very
much alike, but B has these two peaks more separated, and E has a much
broader spread in upper tropospheric O3 abundances.
Simple statistics for the probability distributions in Fig. 6 are presented
in the Tables S6a, b, c, d. Comparisons of the 1-D distributions show that
the log-normal distribution in mole fraction (mean μ and standard
deviation σ) as represented by (μ-σ, μ, μ+σ) is for the most part very close to the equivalent percentile distribution
(16, 50, 84 %). For 2-D summary statistics, we introduce a fitted ellipse
centered at the mean value centroid (X0, Y0) with semimajor and
semiminor axes defined as the standard deviation in the two orthogonal axes
(σX, σY) rotated to find the flattest ellipse (i.e.,
maximum of σX/σY). The values of centroid,
semimajor and semiminor axes, and the degree of rotation are given in Tables S6 for
all plots in Fig. 6. An example showing a fitted ellipse on top of the 2-D
probability distribution is given in Fig. S1, and the ellipses for all six model distributions in Fig. 6a and b are plotted together in Fig. S2a and b,
respectively. These ellipses can provide a more direct and simple comparison
of the central distributions of the models and support the discussion of
Fig. 6 above.
These plots include all altitudes that can be sampled by the ATom flights.
When comparisons with ATom data are made, it will be useful to identify
discrepancies in the 2-D plots by separating altitude regions.
The 2-D plots can change the emphasis of certain regions when weighted by
reactivity. For example, we take the GMI-CTM modeled NOx–HOOH density
(Fig. 6a, panel E) and show the reactivity weightings in Fig. 7. With AIR
weighting, the quasi-boomerang has a strong central line with a negative
slope. With P-O3, a much broader range is seen and the peak occurrence
shifts to lower HOOH values and somewhat even to lower NOx values. With
L-CH4, the line disappears and a galaxy-like pattern widens the range of
parcels, picking up lower NOx values in two spiral arms. The L-O3
weighting is similar to L-CH4, and differences are discernable only in
small features. Clearly, species other than NOx and HOOH determine the
reactivity of parcels, and thus other 2-D plots will add new information. We
anticipate that ATom measurements will be plotted not only with AIR
weightings but also with reactivities calculated for that air parcel with
these models (Auvray et al., 2007).
The 2-D plots shown here intentionally included all air parcels over the
mid-Pacific to ensure that a robust distribution was obtained (see Table 2).
If we have only a single longitude slice as in ATom, then will these be so
clearly defined? We examine this representativeness test by subsampling two
models (C: GFDL-AM3 and F: UCI-CTM) at longitudes of 150, 165, 180, 195, and
210∘ E in Fig. 8 to compare with the average over the mid-Pacific
domain. The densities are renormalized and show similar peaks and patterns,
but of course there is more pixel-level noise and some differences. The
transect at 150∘ E is clearly less representative of the
mid-Pacific, which is understandable since that longitude includes Papua New
Guinea and eastern tropical Australia. Most importantly, the differences for
165–210∘ E are less than those across the six models (Fig. 7a). We
need to develop an objective measure for comparing 2-D plots between models
and ATom measurements and for judging if their differences are within the
range of representativeness. Fortunately, the fitted ellipses provide a
remarkably simple approach to evaluating the similarities and differences in
these different transects and are plotted in Fig. S8c, d. For both models C
and F we see that the central Pacific single transects (165, 180, 195,
210∘ E) with overlapping ellipses match the full block of data (150–210∘ E). In terms of overlapping area, the
single transects overlap the full block at the 86–94 % level, whereas
the 150∘ E transect is distinctly different with overlap of only 42
(F) and 63 % (C). The full-block ellipse from the other model is
plotted in Fig. 8c, d (dashed lines) to show that the models can be
distinguished from even single transects (overlap of 60 %).
Discussion and preparation for the ATom dataset
This paper is based on the underlying premise that high-frequency profiling
of the key species controlling the daily average reactivity of individual
air parcels throughout the remote ocean basins can provide a unique,
objectively sampled chemical climatology identifying those air parcels that
matter, i.e., are most important in controlling methane and tropospheric
ozone. Such data will further provide the most rigorous testing and
diagnosis of the global chemistry models, in particular the
chemistry–climate models, which require a climatology.
Here we present a six model comparison using this new approach. We outline
the model development (i.e., the A-runs) that enables global chemistry
models to readily use high-frequency measurements from aircraft campaigns
like the NASA ATom mission to calculate the chemical reactivity in
individual parcels and over chemical regimes. The multi-model comparison has
already identified some commonalities and highlighted several differences
among the models in their calculation of tropospheric ozone and methane
tendencies. For models that are outliers in particular diagnostics, it is a
challenge for them to identify the cause within their own model and perhaps
explain why the more common results are the ones in error. A test of these
models, isolating the photochemical module by using A-runs with the same
string of simulated measurements, is underway.
The multi-model comparison has provided a range of scientific results:
All six models show distinctly different reactivity profiles in the Pacific
basin, with model–model differences much larger than the A-run and C-run
differences; models that look similar in one reactivity can appear different
in another (e.g., L-CH4 in B and C vs. L-O3).
It is hard to find a consistent pattern in P-O3; we attribute these model
differences to wide variations in NOx abundances over the remote Pacific.
J-values in the tropics, particularly J-O(1-D) (Reaction 3d), differ widely
across the six models; this is unexpected considering the general agreement
with photolysis model comparisons (PhotoComp, 2010) and indicates that
implementation of the photolysis codes in different models may be
inconsistent.
Probability distributions for the tropics show robust differences with clear
outliers, different models are singled out for different species (model A for
CO and O3, D for HCHO and NOx), and surprisingly the water vapor shows
a large range across models.
Reactivity-weighted probability distributions show shifts that might be
expected, based on L-CH4 and L-O3 occurring primarily in the lower
mid-troposphere and P-O3 occurring near the surface and in the upper
troposphere; however, not all models show the same shift, implying a very
different distribution of reactivity and/or dependence on the key species.
Representativeness, specifically the ability of a few Pacific transects to
provide a chemical climatology for the entire basin, was tested extensively
in model F for average reactivities across different longitudes and days and
showed modest variability; when compared in terms of 2-D probability
densities and fitted ellipses, two models showed that longitudinal transects
from 165 to 210∘ E were nearly identical, yet distinct from the other model.
The 1-D and 2-D probability distributions of key species are sufficiently
diverse across the models so that climatology measurements, like those from
ATom, will easily be able to differentiate among them and likely identify
specific model discrepancies. For example, in Fig. 6a models A and E are
alone in identifying a population of parcels with low-HOOH that also have
low-NOx. If this is not found in the observations, then we have some clues
(also looking at other key variables like in Fig. 6b) that will identify
locations and processes. Further, by looking at the reactivity of these
parcels (Fig. 7), we can find that this region is important for methane and
ozone loss. Some work remains in establishing just how close is good enough
in matching 2-D (and multi-D) probability distributions of the key species,
although the overlap of the 2-D fitted ellipses begins to address this.
There are other ATom measurements beyond just key species that might prove
useful as climatological tests for the models. The OH loss frequency (L-OH,
Sinha et al., 2008; Mao et al., 2009) is primarily determined by the
longer-lived reactive species listed here, which can be derived from the key
species, but it is not really a product of the 3-D models. Effectively, L-OH
provides a climatology of a weighted basket of species. The models' predicted
L-OH using their own key species could be tested with the L-OH observations,
but then we are just testing the model's key species and our direct
comparisons are more useful. Actinic fluxes and thus J-values are being
measured by ATom and can be analyzed on a case-by-case basis (Palancar et
al., 2011) to assess the role of clouds in determining instantaneous
reactivity. To be useful as a climatology, the models would need to develop
statistics on how the observed J-values (with clouds) deviated from
clear-sky (modeled) values, thus checking if the photolysis effect of the
cloud statistics in the models is similar that observed. In this case
ATom is probably one of the only useful datasets because flight plans were
made independent of clear or cloudy conditions (except for aircraft safety).
At present there is no clear path to use either L-OH or J's to improve the
climatologies of L-CH4, L-O3, and P-O3.
ATom involves four deployments: ATom-1 completed in August 2016, ATom-2
completed in February 2017, ATom-3 scheduled for October 2017, and ATom-4
completes in May 2018. ATom was successful in completing all flights with
instruments working, acquiring well over 90 % of the proposed dataset, and measuring more than 30 000 10 s air parcels.
A quick look at the
pre-ATom planned flight tracks and sampling in Fig. 9 shows the coverage of
the ocean basins, the large numbers of profiles, and the sampling frequency
as a function of altitude. The expected release of ATom-1 data is mid-2017
and will include the global chemical model products discussed here. These
measurements and analysis will provide a new approach for understanding which
air matters.
The netcdf files of the model output that is analyzed here
are posted with the NASA ATom mission measurements. The publicly available
ATom data can be found at
https://espoarchive.nasa.gov/archive/browse/atom. The .nc files are
located at https://espoarchive.nasa.gov/archive/browse/atom/Model. A
DOI has just been registered for the ATom data including model data
(https://doi.org/10.5067/Aircraft/ATom/TraceGas_Aerosol_Global_Distribution),
and in time the data will be migrated to that DOI.
The Supplement related to this article is available online at https://doi.org/10.5194/acp-17-9081-2017-supplement.
The authors declare that they have no conflict of
interest.
Acknowledgements
This work was supported by NASA funding of the EVS2
Atmospheric Tomography (ATom) mission through a range of specific funding
mechanisms to UC Irvine (NNX15AG57A), NASAS GSFC, Columbia U, NCAR, and
Harvard U. Michael J. Prather thanks Daniel Cariolle (CERFACS) and Valérie Thouret
(Laboratoire d'Aerologie) for hosting his sabbatical, during which this paper
was written and submitted.
Edited by: Marc von Hobe
Reviewed by: two anonymous referees
ReferencesAllen, D., Pickering, K., and Fox-Rabinovitz, M.: Evaluation of pollutant
outflow and CO sources during TRACE-P using model-calculated, aircraft-based,
and Measurements of Pollution in the Troposphere (MOPITT)-derived CO
concentrations, J. Geophys. Res.-Atmos., 109, D15S03,
10.1029/2003jd004250, 2004.Apel, E. C., Olson, J. R., Crawford, J. H., Hornbrook, R. S., Hills, A. J.,
Cantrell, C. A., Emmons, L. K., Knapp, D. J., Hall, S., Mauldin, R. L.,
Weinheimer, A. J., Fried, A., Blake, D. R., Crounse, J. D., St Clair, J. M.,
Wennberg, P. O., Diskin, G. S., Fuelberg, H. E., Wisthaler, A., Mikoviny, T.,
Brune, W., and Riemer, D. D.: Impact of the deep convection of isoprene and
other reactive trace species on radicals and ozone in the upper troposphere,
Atmos. Chem. Phys., 12, 1135–1150, 10.5194/acp-12-1135-2012, 2012.Auvray, M., Bey, I., Llull, E., Schultz, M. G., and Rast, S.: A model
investigation of tropospheric ozone chemical tendencies in long-range
transported pollution plumes, J. Geophys. Res.-Atmos., 112, D05304,
10.1029/2006jd007137, 2007.Barnes, E. A. and Fiore, A. M.: Surface ozone variability and the jet
position: Implications for projecting future air quality, Geophys. Res.
Lett., 40, 2839–2844, 10.1002/Grl.50411, 2013.Bey, I., Jacob, D. J., Yantosca, R. M., Logan, J. A., Field, B. D., Fiore, A.
M., Li, Q. B., Liu, H. G. Y., Mickley, L. J., and Schultz, M. G.: Global
modeling of tropospheric chemistry with assimilated meteorology: Model
description and evaluation, J. Geophys. Res.-Atmos., 106, 23073–23095,
10.1029/2001jd000807, 2001.Blake, N. J., Blake, D. R., Simpson, I. J., Meinardi, S., Swanson, A. L.,
Lopez, J. P., Katzenstein, A. S., Barletta, B., Shirai, T., Atlas, E.,
Sachse, G., Avery, M., Vay, S., Fuelberg, H. E., Kiley, C. M., Kita, K., and
Rowland, F. S.: NMHCs and halocarbons in Asian continental outflow during the
Transport and Chemical Evolution over the Pacific (TRACE-P) Field Campaign:
Comparison with PEM-West B, J. Geophys. Res.-Atmos., 108, 8806,
10.1029/2002jd003367, 2003.Browell, E. V., Fenn, M. A., Butler, C. F., Grant, W. B., Brackett, V. G.,
Hair, J. W., Avery, M. A., Newell, R. E., Hu, Y. L., Fuelberg, H. E., Jacob,
D. J., Anderson, B. E., Atlas, E. L., Blake, D. R., Brune, W. H., Dibb, J.
E., Fried, A., Heikes, B. G., Sachse, G. W., Sandholm, S. T., Singh, H. B.,
Talbot, R. W., Vay, S. A., Weber, R. J., and Bartlett, K. B.: Large-scale
ozone and aerosol distributions, air mass characteristics, and ozone fluxes
over the western Pacific Ocean in late winter/early spring, J. Geophys.
Res.-Atmos., 108, 8805, 10.1029/2002jd003290, 2003.Charlton-Perez, C. L., Evans, M. J., Marsham, J. H., and Esler, J. G.: The
impact of resolution on ship plume simulations with NOx chemistry, Atmos.
Chem. Phys., 9, 7505–7518, 10.5194/acp-9-7505-2009, 2009.
Ciais, P., Sabine, C., Bala, G., Bopp, L., Brovkin, V., Canadell, J.,
Chhabra, A., DeFries, R., Galloway, J., Heimann, M., Jones, C., Le
Quéré, C., Myneni, R. B., Piao, S., and Thornton, P.: Carbon and
Other Biogeochemical Cycles, in: Climate Change 2013: The Physical Science
Basis, IPCC WGI Contribution to the Fifth Assessment Report, edited by:
Stocker, T. F., Qin, D., Plattner, G.-K., Tignor, M., Allen, S. K., Boschung,
J., Nauels, A., Xia, Y., Bex, V., and Midgley, P. M., Cambridge, UK,
46-5-570, 2013.Collins, W. J., Lamarque, J.-F., Schulz, M., Boucher, O., Eyring, V.,
Hegglin, M. I., Maycock, A., Myhre, G., Prather, M., Shindell, D., and Smith,
S. J.: AerChemMIP: quantifying the effects of chemistry and aerosols in
CMIP6, Geosci. Model Dev., 10, 585–607, 10.5194/gmd-10-585-2017, 2017.Crawford, J., Olson, J., Davis, D., Chen, G., Barrick, J., Shetter, R.,
Lefer, B., Jordan, C., Anderson, B., Clarke, A., Sachse, G., Blake, D.,
Singh, H., Sandolm, S., Tan, D., Kondo, Y., Avery, M., Flocke, F., Eisele,
F., Mauldin, L., Zondlo, M., Brune, W., Harder, H., Martinez, M., Talbot, R.,
Bandy, A., and Thornton, D.: Clouds and trace gas distributions during
TRACE-P, J. Geophys. Res.-Atmos., 108, 8818, 10.1029/2002jd003177, 2003.Dacre, H. F., Clark, P. A., Martinez-Alvarado, O., Stringer, M. A., and
Lavers, D. A.: How Do Atmospheric Rivers Form?, B. Am. Meteorol. Soc., 96,
1243–1255, 10.1175/Bams-D-14-00031.1, 2015.Damoah, R., Spichtinger, N., Forster, C., James, P., Mattis, I., Wandinger,
U., Beirle, S., Wagner, T., and Stohl, A.: Around the world in 17 days –
hemispheric-scale transport of forest fire smoke from Russia in May 2003,
Atmos. Chem. Phys., 4, 1311–1321, 10.5194/acp-4-1311-2004, 2004.Davis, D. D., Chen, G., Crawford, J. H., Liu, S., Tan, D., Sandholm, S. T.,
Jing, P., Cunnold, D. M., DiNunno, B., Browell, E. V., Grant, W. B., Fenn, M.
A., Anderson, B. E., Barrick, J. D., Sachse, G. W., Vay, S. A., Hudgins, C.
H., Avery, M. A., Lefer, B., Shetter, R. E., Heikes, B. G., Blake, D. R.,
Blake, N., Kondo, Y., and Oltmans, S.: An assessment of western North Pacific
ozone photochemistry based on springtime observations from NASA's PEM-West B
(1994) and TRACE-P (2001) field studies, J. Geophys. Res.-Atmos., 108, 8829,
10.1029/2002jd003232, 2003.Donner, L. J., Wyman, B. L., Hemler, R. S., Horowitz, L. W., Ming, Y., Zhao,
M., Golaz, J. C., Ginoux, P., Lin, S. J., Schwarzkopf, M. D., Austin, J.,
Alaka, G., Cooke, W. F., Delworth, T. L., Freidenreich, S. M., Gordon, C. T.,
Griffies, S. M., Held, I. M., Hurlin, W. J., Klein, S. A., Knutson, T. R.,
Langenhorst, A. R., Lee, H. C., Lin, Y. L., Magi, B. I., Malyshev, S. L.,
Milly, P. C. D., Naik, V., Nath, M. J., Pincus, R., Ploshay, J. J.,
Ramaswamy, V., Seman, C. J., Shevliakova, E., Sirutis, J. J., Stern, W. F.,
Stouffer, R. J., Wilson, R. J., Winton, M., Wittenberg, A. T., and Zeng, F.
R.: The Dynamical Core, Physical Parameterizations, and Basic Simulation
Characteristics of the Atmospheric Component AM3 of the GFDL Global Coupled
Model CM3, J. Clim., 24, 3484–3519, 10.1175/2011jcli3955.1, 2011.Duncan, B. N., Strahan, S. E., Yoshida, Y., Steenrod, S. D., and Livesey, N.:
Model study of the cross-tropopause transport of biomass burning pollution,
Atmos. Chem. Phys., 7, 3713–3736, 10.5194/acp-7-3713-2007, 2007.Eastham, S. D., Weisenstein, D. K., and Barrett, S. R. H.: Development and
evaluation of the unified tropospheric-stratospheric chemistry extension
(UCX) for the global chemistry-transport model GEOS-Chem, Atmos. Environ.,
89, 52–63, 10.1016/j.atmosenv.2014.02.001, 2014.Eckstein, J., Ruhnke, R., Zahn, A., Neumaier, M., Kirner, O., and Braesicke,
P.: An assessment of the climatological representativeness of IAGOS-CARIBIC
trace gas measurements using EMAC model simulations, Atmos. Chem. Phys., 17,
2775–2794, 10.5194/acp-17-2775-2017, 2017.
Ehhalt, D. H., Rohrer, F., Kraus, A. B., Prather, M. J., Blake, D. R., and
Rowland, F. S.: On the significance of regional trace gas distributions as
derived from aircraft campaigns in PEM-West A and B, J. Geophys. Res.-Atmos.,
102, 28333–28351, 1997.Elguindi, N., Clark, H., Ordonez, C., Thouret, V., Flemming, J., Stein, O.,
Huijnen, V., Moinat, P., Inness, A., Peuch, V. H., Stohl, A., Turquety, S.,
Athier, G., Cammas, J. P., and Schultz, M.: Current status of the ability of
the GEMS/MACC models to reproduce the tropospheric CO vertical distribution
as measured by MOZAIC, Geosci. Model Dev., 3, 501–518,
10.5194/gmd-3-501-2010, 2010.Engel, A., Bönisch, H., Brunner, D., Fischer, H., Franke, H.,
Günther, G., Gurk, C., Hegglin, M., Hoor, P., Königstedt, R.,
Krebsbach, M., Maser, R., Parchatka, U., Peter, T., Schell, D., Schiller, C.,
Schmidt, U., Spelten, N., Szabo, T., Weers, U., Wernli, H., Wetter, T., and
Wirth, V.: Highly resolved observations of trace gases in the lowermost
stratosphere and upper troposphere from the Spurt project: an overview,
Atmos. Chem. Phys., 6, 283–301, 10.5194/acp-6-283-2006,
2006.Esler, J. G.: An integrated approach to mixing sensitivities in tropospheric
chemistry: A basis for the parameterization of subgrid-scale emissions for
chemistry transport models, J. Geophys. Res.-Atmos., 108, 4632,
10.1029/2003jd003627, 2003.Fang, Y. Y., Mauzerall, D. L., Liu, J. F., Fiore, A. M., and Horowitz, L. W.:
Impacts of 21st century climate change on global air pollution-related
premature mortality, Climatic Change, 121, 239–253,
10.1007/s10584-013-0847-8, 2013.Fiore, A. M., Naik, V., Spracklen, D. V., Steiner, A., Unger, N., Prather,
M., Bergmann, D., Cameron-Smith, P. J., Cionni, I., Collins, W. J., Dalsoren,
S., Eyring, V., Folberth, G. A., Ginoux, P., Horowitz, L. W., Josse, B.,
Lamarque, J. F., MacKenzie, I. A., Nagashima, T., O'Connor, F. M., Righi, M.,
Rumbold, S. T., Shindell, D. T., Skeie, R. B., Sudo, K., Szopa, S., Takemura,
T., and Zeng, G.: Global air quality and climate, Chem. Soc. Rev., 41,
6663–6683, 10.1039/C2cs35095e, 2012.Fishman, J., Hoell, J. M., Bendura, R. D., McNeil, R. J., and Kirchhoff, V.
W. J. H.: NASA GTE TRACE A experiment (September–October 1992), Overview, J.
Geophys. Res.-Atmos., 101, 23865–23879, 10.1029/96jd00123, 1996.Gaudel, A., Clark, H., Thouret, V., Jones, L., Inness, A., Flemming, J.,
Stein, O., Huijnen, V., Eskes, H., Nedelec, P., and Boulanger, D.: On the use
of MOZAIC-IAGOS data to assess the ability of the MACC reanalysis to
reproduce the distribution of ozone and CO in the UTLS over Europe, Tellus B,
67, 27955, 10.3402/Tellusb.V67.27955, 2015.Hardacre, C., Wild, O., and Emberson, L.: An evaluation of ozone dry
deposition in global scale chemistry climate models, Atmos. Chem. Phys., 15,
6419–6436, 10.5194/acp-15-6419-2015, 2015.Heald, C. L., Jacob, D. J., Fiore, A. M., Emmons, L. K., Gille, J. C.,
Deeter, M. N., Warner, J., Edwards, D. P., Crawford, J. H., Hamlin, A. J.,
Sachse, G. W., Browell, E. V., Avery, M. A., Vay, S. A., Westberg, D. J.,
Blake, D. R., Singh, H. B., Sandholm, S. T., Talbot, R. W., and Fuelberg, H.
E.: Asian outflow and trans-Pacific transport of carbon monoxide and ozone
pollution: An integrated satellite, aircraft, and model perspective, J.
Geophys. Res.-Atmos., 108, 4804, 10.1029/2003jd003507, 2003.Hecobian, A., Liu, Z., Hennigan, C. J., Huey, L. G., Jimenez, J. L., Cubison,
M. J., Vay, S., Diskin, G. S., Sachse, G. W., Wisthaler, A., Mikoviny, T.,
Weinheimer, A. J., Liao, J., Knapp, D. J., Wennberg, P. O., Kurten, A.,
Crounse, J. D., St Clair, J., Wang, Y., and Weber, R. J.: Comparison of
chemical characteristics of 495 biomass burning plumes intercepted by the
NASA DC-8 aircraft during the ARCTAS/CARB-2008 field campaign, Atmos. Chem.
Phys., 11, 13325–13337, 10.5194/acp-11-13325-2011, 2011.Hoell, J. M., Davis, D. D., Liu, S. C., Newell, R., Shipham, M., Akimoto, H.,
McNeal, R. J., Bendura, R. J., and Drewry, J. W.: Pacific exploratory
Mission-West A (PEM-West A): September–October 1991, J. Geophys.
Res.-Atmos., 101, 1641–1653, 10.1029/95jd00622, 1996.Hoell, J. M., Davis, D. D., Jacob, D. J., Rodgers, M. O., Newell, R. E.,
Fuelberg, H. E., McNeal, R. J., Raper, J. L., and Bendura, R. J.: Pacific
Exploratory Mission in the tropical Pacific: PEM-Tropics A, August–September
1996, J. Geophys. Res.-Atmos., 104, 5567–5583, 10.1029/1998jd100074,
1999.Holmes, C. D., Prather, M. J., Sovde, O. A., and Myhre, G.: Future methane,
hydroxyl, and their uncertainties: key climate and emission parameters for
future predictions, Atmos. Chem. Phys., 13, 285–302,
10.5194/acp-13-285-2013, 2013.Holmes, C. D., Prather, M. J., and Vinken, G. C. M.: The climate impact of
ship NOx emissions: an improved estimate accounting for plume chemistry,
Atmos. Chem. Phys., 14, 6801–6812, 10.5194/acp-14-6801-2014, 2014.Hoor, P., Fischer, H., Lange, L., Lelieveld, J., and Brunner, D.: Seasonal
variations of a mixing layer in the lowermost stratosphere as identified by
the CO-O-3 correlation from in situ measurements, J. Geophys. Res.-Atmos.,
107, 4044, 10.1029/2000jd000289, 2002.Hsu, J. and Prather, M. J.: Stratospheric variability and tropospheric ozone,
J. Geophys. Res.-Atmos., 114, D06102, 10.1029/2008jd010942, 2009.Hsu, J., Prather, M. J., Wild, O., Sundet, J. K., Isaksen, I. S. A., Browell,
E. V., Avery, M. A., and Sachse, G. W.: Are the TRACE-P measurements
representative of the western Pacific during March 2001?, J. Geophys.
Res.-Atmos., 109, D02314, 10.1029/2003jd004002, 2004.Jacob, D. J. and Winner, D. A.: Effect of climate change on air quality,
Atmos. Environ., 43, 51–63, 10.1016/j.atmosenv.2008.09.051, 2009.Jacob, D. J., Crawford, J. H., Kleb, M. M., Connors, V. S., Bendura, R. J.,
Raper, J. L., Sachse, G. W., Gille, J. C., Emmons, L., and Heald, C. L.:
Transport and Chemical Evolution over the Pacific (TRACE-P) aircraft mission:
Design, execution, and first results, J. Geophys. Res.-Atmos., 108, 1–19,
10.1029/2002jd003276, 2003.Jacob, D. J., Crawford, J. H., Maring, H., Clarke, A. D., Dibb, J. E.,
Emmons, L. K., Ferrare, R. A., Hostetler, C. A., Russell, P. B., Singh, H.
B., Thompson, A. M., Shaw, G. E., McCauley, E., Pederson, J. R., and Fisher,
J. A.: The Arctic Research of the Composition of the Troposphere from
Aircraft and Satellites (ARCTAS) mission: design, execution, and first
results, Atmos. Chem. Phys., 10, 5191–5212, 10.5194/acp-10-5191-2010,
2010.Kiley, C. M., Fuelberg, H. E., Palmer, P. I., Allen, D. J., Carmichael, G.
R., Jacob, D. J., Mari, C., Pierce, R. B., Pickering, K. E., Tang, Y. H.,
Wild, O., Fairlie, T. D., Logan, J. A., Sachse, G. W., Shaack, T. K., and
Streets, D. G.: An intercomparison and evaluation of aircraft-derived and
simulated CO from seven chemical transport models during the TRACE-P
experiment, J. Geophys. Res.-Atmos., 108, 8819, 10.1029/2002jd003089,
2003.Kley, D., Crutzen, P. J., Smit, H. G. J., Vomel, H., Oltmans, S. J., Grassl,
H., and Ramanathan, V.: Observations of near-zero ozone concentrations over
the convective Pacific: Effects on air chemistry, Science, 274, 230–233,
10.1126/science.274.5285.230, 1996.Köppe, M., Hermann, M., Brenninkmeijer, C. A. M., Heintzenberg, J., Schlager,
H., Schuck, T., Slemr, F., Sprung, D., van Velthoven, P. F. J., Wiedensohler,
A., Zahn, A., and Ziereis, H.: Origin of aerosol particles in the
mid-latitude and subtropical upper troposphere and lowermost stratosphere
from cluster analysis of CARIBIC data, Atmos. Chem. Phys., 9, 8413–8430,
10.5194/acp-9-8413-2009,
2009.Kunz, A., Schiller, C., Rohrer, F., Smit, H. G. J., Nedelec, P., and Spelten,
N.: Statistical analysis of water vapour and ozone in the UT / LS observed
during SPURT and MOZAIC, Atmos. Chem. Phys., 8, 6603–6615, 10.5194/acp-8-6603-2008, 2008.Lamarque, J. F., Emmons, L. K., Hess, P. G., Kinnison, D. E., Tilmes, S.,
Vitt, F., Heald, C. L., Holland, E. A., Lauritzen, P. H., Neu, J., Orlando,
J. J., Rasch, P. J., and Tyndall, G. K.: CAM-chem: description and evaluation
of interactive atmospheric chemistry in the Community Earth System Model,
Geosci. Model Dev., 5, 369–411, 10.5194/gmd-5-369-2012, 2012.Lamarque, J. F., Shindell, D. T., Josse, B., Young, P. J., Cionni, I.,
Eyring, V., Bergmann, D., Cameron-Smith, P., Collins, W. J., Doherty, R.,
Dalsoren, S., Faluvegi, G., Folberth, G., Ghan, S. J., Horowitz, L. W., Lee,
Y. H., MacKenzie, I. A., Nagashima, T., Naik, V., Plummer, D., Righi, M.,
Rumbold, S. T., Schulz, M., Skeie, R. B., Stevenson, D. S., Strode, S., Sudo,
K., Szopa, S., Voulgarakis, A., and Zeng, G.: The Atmospheric Chemistry and
Climate Model Intercomparison Project (ACCMIP): overview and description of
models, simulations and climate diagnostics, Geosci. Model Dev., 6, 179–206,
10.5194/gmd-6-179-2013, 2013.Larsen, M. L., Briner, C. A., and Boehner, P.: On the Recovery of 3D Spatial
Statistics of Particles from 1D Measurements: Implications for Airborne
Instruments, J. Atmos. Ocean Tech., 31, 2078–2087,
10.1175/Jtech-D-14-00004.1, 2014.Lawrence, M. G., Jökel, P., and von Kuhlmann, R.: What does the global
mean OH concentration tell us?, Atmos. Chem. Phys., 1, 37–49,
10.5194/acp-1-37-2001, 2001.Lelieveld, J. and Crutzen, P. J.: The Role of Clouds in Tropospheric
Photochemistry, J. Atmos. Chem., 12, 229–267, 10.1007/Bf00048075,
1991.Li, J. Y., Mao, J. Q., Min, K. E., Washenfelder, R. A., Brown, S. S., Kaiser,
J., Keutsch, F. N., Volkamer, R., Wolfe, G. M., Hanisco, T. F., Pollack, I.
B., Ryerson, T. B., Graus, M., Gilman, J. B., Lerner, B. M., Warneke, C., de
Gouw, J. A., Middlebrook, A. M., Liao, J., Welti, A., Henderson, B. H.,
McNeill, V. F., Hall, S. R., Ullmann, K., Donner, L. J., Paulot, F., and
Horowitz, L. W.: Observational constraints on glyoxal production from
isoprene oxidation and its contribution to organic aerosol over the Southeast
United States, J. Geophys. Res.-Atmos., 121, 9849–9861,
10.1002/2016JD025331, 2016.Manney, G. L., Bird, J. C., Donovan, D. P., Duck, T. J., Whiteway, J. A.,
Pal, S. R., and Carswell, A. I.: Modeling ozone laminae in ground-based
Arctic wintertime observations using trajectory calculations and satellite
data, J. Geophys. Res.-Atmos., 103, 5797–5814, 10.1029/97jd03449, 1998.Mao, J., Ren, X., Brune, W. H., Olson, J. R., Crawford, J. H., Fried, A.,
Huey, L. G., Cohen, R. C., Heikes, B., Singh, H. B., Blake, D. R., Sachse, G.
W., Diskin, G. S., Hall, S. R., and Shetter, R. E.: Airborne measurement of
OH reactivity during INTEX-B, Atmos. Chem. Phys., 9, 163–173,
10.5194/acp-9-163-2009, 2009.Marenco, A., Thouret, V., Nédélec, P., Smit, H., Helten, M., Kley,
D., Karcher, F., Simon, P., Law, K., Pyle, J., Poschmann, G., Von Wrede, R.,
Hume, C., and Cook, T.: Measurement of ozone and water vapor by Airbus
in-service aircraft: The MOZAIC airborne program, an overview, J. Geophys.
Res.-Atmos., 103, 25631–25642, 10.1029/98jd00977, 1998.Mickley, L. J., Jacob, D. J., Field, B. D., and Rind, D.: Effects of future
climate change on regional air pollution episodes in the United States,
Geophys. Res. Lett., 31, L24103, 10.1029/2004gl021216, 2004.Mundhenk, B. D., Barnes, E. A., and Maloney, E. D.: All-Season Climatology
and Variability of Atmospheric River Frequencies over the North Pacific, J.
Clim., 29, 4885–4903, 10.1175/Jcli-D-15-0655.1, 2016.Naik, V., Horowitz, L. W., Fiore, A. M., Ginoux, P., Mao, J. Q., Aghedo, A.
M., and Levy, H.: Impact of preindustrial to present-day changes in
short-lived pollutant emissions on atmospheric composition and climate
forcing, J. Geophys. Res.-Atmos., 118, 8086–8110, 10.1002/jgrd.50608,
2013a.Naik, V., Voulgarakis, A., Fiore, A. M., Horowitz, L. W., Lamarque, J. F.,
Lin, M., Prather, M. J., Young, P. J., Bergmann, D., Cameron-Smith, P. J.,
Cionni, I., Collins, W. J., Dalsoren, S. B., Doherty, R., Eyring, V.,
Faluvegi, G., Folberth, G. A., Josse, B., Lee, Y. H., MacKenzie, I. A.,
Nagashima, T., van Noije, T. P. C., Plummer, D. A., Righi, M., Rumbold, S.
T., Skeie, R., Shindell, D. T., Stevenson, D. S., Strode, S., Sudo, K.,
Szopa, S., and Zeng, G.: Preindustrial to present-day changes in tropospheric
hydroxyl radical and methane lifetime from the Atmospheric Chemistry and
Climate Model Intercomparison Project (ACCMIP), Atmos. Chem. Phys., 13,
5277–5298, 10.5194/acp-13-5277-2013, 2013b.
Nappo, C. J., Caneill, J. Y., Furman, R. W., Gifford, F. A., Kaimal, J. C.,
Kramer, M. L., Lockhart, T. J., Pendergast, M. M., Pielke, R. A., Randerson,
D., Shreffler, J. H., and Wyngaard, J. C.: The Workshop on the
Representativeness of Meteorological-Observations, June 1981, Boulder, Colorado,
B. Am. Meteorol. Soc., 63, 761–764, 1982.Newell, R. E., Newell, N. E., Zhu, Y., and Scott, C.: Tropospheric Rivers –
a Pilot-Study, Geophys. Res. Lett., 19, 2401–2404, 10.1029/92gl02916,
1992.Newell, R. E., V, T., Cho, J. Y. N., Stoller, P., Marenco, A., and Smit, H.
G.: Ubiquity of quasi-horizontal layers in the troposphere, Nature, 398,
316–319, 10.1038/18642, 1999.Nicely, J. M., Anderson, D. C., Canty, T. P., Salawitch, R. J., Wolfe, G. M.,
Apel, E. C., Arnold, S. R., Atlas, E. L., Blake, N. J., Bresch, J. F.,
Campos, T. L., Dickerson, R. R., Duncan, B., Emmons, L. K., Evans, M. J.,
Fernandez, R. P., Flemming, J., Hall, S. R., Hanisco, T. F., Honomichl, S.
B., Hornbrook, R. S., Huijnen, V., Kaser, L., Kinnison, D. E., Lamarque, J.
F., Mao, J. Q., Monks, S. A., Montzka, D. D., Pan, L. L., Riemer, D. D.,
Saiz-Lopez, A., Steenrod, S. D., Stell, M. H., Tilmes, S., Turquety, S.,
Ullmann, K., and Weinheimer, A. J.: An observationally constrained evaluation
of the oxidative capacity in the tropical western Pacific troposphere, J.
Geophys. Res.-Atmos., 121, 7461–7488, 10.1002/2016JD025067, 2016.Nicely, J. M., Salawitch, R. J., Canty, T., Anderson, D. C., Arnold, S. R.,
Chipperfield, M. P., Emmons, L. K., Flemming, J., Huijnen, V., Kinnison, D.
E., Lamarque, J. F., Mao, J. Q., Monks, S. A., Steenrod, S. D., Tilmes, S.,
and Turquety, S.: Quantifying the causes of differences in tropospheric OH
within global models, J. Geophys. Res.-Atmos., 122, 1983–2007,
10.1002/2016JD026239, 2017.Olson, J. R., Crawford, J. H., Chen, G., Fried, A., Evans, M. J., Jordan, C.
E., Sandholm, S. T., Davis, D. D., Anderson, B. E., Avery, M. A., Barrick, J.
D., Blake, D. R., Brune, W. H., Eisele, F. L., Flocke, F., Harder, H., Jacob,
D. J., Kondo, Y., Lefer, B. L., Martinez, M., Mauldin, R. L., Sachse, G. W.,
Shetter, R. E., Singh, H. B., Talbot, R. W., and Tan, D.: Testing fast
photochemical theory during TRACE-P based on measurements of OH, HO2, and
CH2O, J. Geophys. Res.-Atmos., 109, D15S10, 10.1029/2003jd004278,
2004.Olson, J. R., Crawford, J. H., Brune, W., Mao, J., Ren, X., Fried, A.,
Anderson, B., Apel, E., Beaver, M., Blake, D., Chen, G., Crounse, J., Dibb,
J., Diskin, G., Hall, S. R., Huey, L. G., Knapp, D., Richter, D., Riemer, D.,
Clair, J. S., Ullmann, K., Walega, J., Weibring, P., Weinheimer, A.,
Wennberg, P., and Wisthaler, A.: An analysis of fast photochemistry over high
northern latitudes during spring and summer using in-situ observations from
ARCTAS and TOPSE, Atmos. Chem. Phys., 12, 6799–6825,
10.5194/acp-12-6799-2012, 2012.Palancar, G. G., Shetter, R. E., Hall, S. R., Toselli, B. M., and Madronich,
S.: Ultraviolet actinic flux in clear and cloudy atmospheres: model
calculations and aircraft-based measurements, Atmos. Chem. Phys., 11,
5457–5469, 10.5194/acp-11-5457-2011, 2011.Pan, L., Atlas, E., Salawitch, R., Honomichl, S., Bresch, J., Randel, W., Apel, E.,
Hornbrook, R., Weinheimer, A., Anderson, D., Andrews, S., Baidar, S.,
Beaton, S., Carpenter, L. J., Chen, D., Dix, B., Donets, V., Hall, S.,
Hanisco, T.,
Homeyer, C., Huey, L., Jensen, J., Kaser, L., Kinnison, D., Koenig, T.,
Lamarque, J., Liu, C., Luo, J., Luo, Z., Montzka, D., Nicely, J., Pierce, R.,
Riemer, D., Robinson, T., Romashkin, P., Saiz-Lopez, A., Schauffler, S.,
Shieh, O., Stell, M., Ullmann, K., Vaughan, G., Volkamer, R., and Wolfe, G.:
The Convective Transport of Active Species in the Tropics (CONTRAST)
Experimen, B. Am. Meteor. Soc., 106–128, 10.1175/BAMS-D-14-00272.1,
2017.Pan, L. L., Wei, J. C., Kinnison, D. E., Garcia, R. R., Wuebbles, D. J., and
Brasseur, G. P.: A set of diagnostics for evaluating chemistry-climate models
in the extratropical tropopause region, J. Geophys. Res.-Atmos., 112, D09316,
10.1029/2006jd007792, 2007.Parrington, M., Palmer, P. I., Lewis, A. C., Lee, J. D., Rickard, A. R., Di
Carlo, P., Taylor, J. W., Hopkins, J. R., Punjabi, S., Oram, D. E., Forster,
G., Aruffo, E., Moller, S. J., Bauguitte, S. J. B., Allan, J. D., Coe, H.,
and Leigh, R. J.: Ozone photochemistry in boreal biomass burning plumes,
Atmos. Chem. Phys., 13, 7321–7341, 10.5194/acp-13-7321-2013, 2013.Paulson, S. E. and Orlando, J. J.: The reactions of ozone with alkenes: An
important source of HOx in the boundary layer, Geophys. Res. Lett., 23,
3727–3730, 10.1029/96gl03477, 1996.
PhotoComp: Stratospheric Chemistry SPARC Report No. 5 on the
Evaluation of Chemistry, Climate Models, 194–202, 2010.Pisso, I., Real, E., Law, K. S., Legras, B., Bousserez, N., Attie, J. L., and
Schlager, H.: Estimation of mixing in the troposphere from Lagrangian trace
gas reconstructions during long-range pollution plume transport, J. Geophys.
Res.-Atmos., 114, D19301, 10.1029/2008jd011289, 2009.
Prather, M. and Jaffe, A. H.: Global Impact of the Antarctic Ozone Hole –
Chemical Propagation, J. Geophys. Res.-Atmos., 95, 3473–3492, 1990.Prather, M., Gauss, M., Berntsen, T., Isaksen, I., Sundet, J., Bey, I.,
Brasseur, G., Dentener, F., Derwent, R., Stevenson, D., Grenfell, L.,
Hauglustaine, D., Horowitz, L., Jacob, D., Mickley, L., Lawrence, M., von
Kuhlmann, R., Muller, J.-F., Pitari, G., Rogers, H., Johnson, M., Pyle, J.,
Law, K., van Weele, M., and Wild, O.: Fresh air in the 21st century?,
Geophys. Res. Lett., 30, 1100, 10.1029/2002gl016285, 2003.Prather, M. J.: Time scales in atmospheric chemistry: Theory, GWPs for CH4
and CO, and runaway growth, Geophys. Res. Lett., 23, 2597–2600, 1996.Prather, M. J.: Tropospheric O3 from photolysis of O2, Geophys.
Res. Lett., 36, L03811, 10.1029/2008gl036851, 2009.Prather, M. J.: Photolysis rates in correlated overlapping cloud fields:
Cloud-J 7.3c, Geosci. Model Dev., 8, 2587–2595, 10.5194/gmd-8-2587-2015,
2015.Prather, M. J. and Holmes, C. D.: A perspective on time: loss frequencies,
time scales and lifetimes, Environ. Chem., 10, 73–79, 10.1071/En13017,
2013.Prather, M. J., Holmes, C. D., and Hsu, J.: Reactive greenhouse gas
scenarios: Systematic exploration of uncertainties and the role of
atmospheric chemistry, Geophys. Res. Lett., 39, L09803,
10.1029/2012gl051440, 2012.Ramsey, C. A. and Hewitt, A. D.: A methodology for assessing sample
representativeness, Environ. Forensics, 6, 71–75,
10.1080/15275920590913877, 2005.Raper, J. L., Kleb, M. M., Jacob, D. J., Davis, D. D., Newell, R. E.,
Fuelberg, H. E., Bendura, R. J., Hoell, J. M., and McNeal, R. J.: Pacific
Exploratory Mission in the tropical Pacific: PEM-Tropics B, March–April
1999, J. Geophys. Res.-Atmos., 106, 32401–32425, 10.1029/2000jd900833,
2001.Reid, S. J., Rex, M., Von der Gathen, P., Floisand, I., Stordal, F., Carver,
G. D., Beck, A., Reimer, E., Kruger-Carstensen, R., De Haan, L. L., Braathen,
G., Dorokhov, V., Fast, H., Kyro, E., Gil, M., Litynska, Z., Molyneux, M.,
Murphy, G., O'Connor, F., Ravegnani, F., Varotsos, C., Wenger, J., and
Zerefos, C.: A study of ozone laminae using diabatic trajectories, contour
advection and photochemical trajectory model simulations, J. Atmos. Chem.,
30, 187–207, 10.1023/A:1005836212979, 1998.Rohrer, F. and Berresheim, H.: Strong correlation between levels of
tropospheric hydroxyl radicals and solar ultraviolet radiation, Nature, 442,
184–187, 10.1038/nature04924, 2006.Schmidt, G. A., Kelley, M., Nazarenko, L., Ruedy, R., Russell, G. L.,
Aleinov, I., Bauer, M., Bauer, S. E., Bhat, M. K., Bleck, R., Canuto, V.,
Chen, Y. H., Cheng, Y., Clune, T. L., Del Genio, A., de Fainchtein, R.,
Faluvegi, G., Hansen, J. E., Healy, R. J., Kiang, N. Y., Koch, D., Lacis, A.
A., LeGrande, A. N., Lerner, J., Lo, K. K., Matthews, E. E., Menon, S.,
Miller, R. L., Oinas, V., Oloso, A. O., Perlwitz, J. P., Puma, M. J., Putman,
W. M., Rind, D., Romanou, A., Sato, M., Shindell, D. T., Sun, S., Syed, R.
A., Tausnev, N., Tsigaridis, K., Unger, N., Voulgarakis, A., Yao, M. S., and
Zhang, J. L.: Configuration and assessment of the GISS ModelE2 contributions
to the CMIP5 archive, Journal of Advances in Modeling Earth Systems, 6,
141–184, 10.1002/2013MS000265, 2014.Schnell, J. L., Prather, M. J., Josse, B., Naik, V., Horowitz, L. W.,
Cameron-Smith, P., Bergmann, D., Zeng, G., Plummer, D. A., Sudo, K.,
Nagashima, T., Shindell, D. T., Faluvegi, G., and Strode, S. A.: Use of North
American and European air quality networks to evaluate global
chemistry-climate modeling of surface ozone, Atmos. Chem. Phys., 15,
10581–10596, 10.5194/acp-15-10581-2015, 2015.Schoeberl, M. R., Ziemke, J. R., Bojkov, B., Livesey, N., Duncan, B.,
Strahan, S., Froidevaux, L., Kulawik, S., Bhartia, P. K., Chandra, S.,
Levelt, P. F., Witte, J. C., Thompson, A. M., Cuevas, E., Redondas, A.,
Tarasick, D. W., Davies, J., Bodeker, G., Hansen, G., Johnson, B. J.,
Oltmans, S. J., Vomel, H., Allaart, M., Kelder, H., Newchurch, M.,
Godin-Beekmann, S., Ancellet, G., Claude, H., Andersen, S. B., Kyro, E.,
Parrondos, M., Yela, M., Zablocki, G., Moore, D., Dier, H., von der Gathen,
P., Viatte, P., Stubi, R., Calpini, B., Skrivankova, P., Dorokhov, V., de
Backer, H., Schmidlin, F. J., Coetzee, G., Fujiwara, M., Thouret, V., Posny,
F., Morris, G., Merrill, J., Leong, C. P., Koenig-Langlo, G., and Joseph, E.:
A trajectory-based estimate of the tropospheric ozone column using the
residual method, J. Geophys. Res.-Atmos., 112, D24S49,
10.1029/2007jd008773, 2007.Shindell, D. T., Pechony, O., Voulgarakis, A., Faluvegi, G., Nazarenko, L.,
Lamarque, J. F., Bowman, K., Milly, G., Kovari, B., Ruedy, R., and Schmidt,
G. A.: Interactive ozone and methane chemistry in GISS-E2 historical and
future climate simulations, Atmos. Chem. Phys., 13, 2653–2689,
10.5194/acp-13-2653-2013, 2013.Singh, H. B., Gregory, G. L., Anderson, B., Browell, E., Sachse, G. W.,
Davis, D. D., Crawford, J., Bradshaw, J. D., Talbot, R., Blake, D. R.,
Thornton, D., Newell, R., and Merrill, J.: Low ozone in the marine boundary
payer of the tropical Pacific Ocean: Photochemical loss, chlorine atoms, and
entrainment, J. Geophys. Res.-Atmos., 101, 1907–1917, 10.1029/95jd01028,
1996.Singh, H. B., Viezee, W., Chen, Y., Bradshaw, J., Sandholm, S., Blake, D.,
Blake, N., Heikes, B., Snow, J., Talbot, R., Browell, E., Gregory, G.,
Sachse, G., and Vay, S.: Biomass burning influences on the composition of the
remote South Pacific troposphere: analysis based on observations from
PEM-Tropics-A, Atmos. Environ., 34, 635–644,
10.1016/S1352-2310(99)00380-5, 2000.Sinha, V., Williams, J., Crowley, J. N., and Lelieveld, J.: The Comparative
Reactivity Method – a new tool to measure total OH Reactivity in ambient
air, Atmos. Chem. Phys., 8, 2213–2227, 10.5194/acp-8-2213-2008, 2008.Sovde, O. A., Prather, M. J., Isaksen, I. S. A., Berntsen, T. K., Stordal,
F., Zhu, X., Holmes, C. D., and Hsu, J.: The chemical transport model Oslo
CTM3, Geosci. Model Dev., 5, 1441–1469, 10.5194/gmd-5-1441-2012, 2012.
Spivakovsky, C. M., Logan, J. A., Montzka, S. A., Balkanski, Y. J.,
Foreman-Fowler, M., Jones, D. B. A., Horowitz, L. W., Fusco, A. C.,
Brenninkmeijer, C. A. M., Prather, M. J., Wofsy, S. C., and McElroy, M. B.:
Three-dimensional climatological distribution of tropospheric OH: Update and
evaluation, J. Geophys. Res.-Atmos., 105, 8931–8980, 2000.Stevenson, D. S., Dentener, F. J., Schultz, M. G., Ellingsen, K., van Noije,
T. P. C., Wild, O., Zeng, G., Amann, M., Atherton, C. S., Bell, N., Bergmann,
D. J., Bey, I., Butler, T., Cofala, J., Collins, W. J., Derwent, R. G.,
Doherty, R. M., Drevet, J., Eskes, H. J., Fiore, A. M., Gauss, M.,
Hauglustaine, D. A., Horowitz, L. W., Isaksen, I. S. A., Krol, M. C.,
Lamarque, J. F., Lawrence, M. G., Montanaro, V., Muller, J. F., Pitari, G.,
Prather, M. J., Pyle, J. A., Rast, S., Rodriguez, J. M., Sanderson, M. G.,
Savage, N. H., Shindell, D. T., Strahan, S. E., Sudo, K., and Szopa, S.:
Multimodel ensemble simulations of present-day and near-future tropospheric
ozone, J. Geophys. Res.-Atmos., 111, D08301, 10.1029/2005jd006338, 2006.Stevenson, D. S., Young, P. J., Naik, V., Lamarque, J. F., Shindell, D. T.,
Voulgarakis, A., Skeie, R. B., Dalsoren, S. B., Myhre, G., Berntsen, T. K.,
Folberth, G. A., Rumbold, S. T., Collins, W. J., MacKenzie, I. A., Doherty,
R. M., Zeng, G., van Noije, T. P. C., Strunk, A., Bergmann, D.,
Cameron-Smith, P., Plummer, D. A., Strode, S. A., Horowitz, L., Lee, Y. H.,
Szopa, S., Sudo, K., Nagashima, T., Josse, B., Cionni, I., Righi, M., Eyring,
V., Conley, A., Bowman, K. W., Wild, O., and Archibald, A.: Tropospheric
ozone changes, radiative forcing and attribution to emissions in the
Atmospheric Chemistry and Climate Model Intercomparison Project (ACCMIP),
Atmos. Chem. Phys., 13, 3063–3085, 10.5194/acp-13-3063-2013, 2013.Stoller, P., Cho, J. Y. N., Newell, R. E., Thouret, V., Zhu, Y., Carroll, M.
A., Albercook, G. M., Anderson, B. E., Barrick, J. D. W., Browell, E. V.,
Gregory, G. L., Sachse, G. W., Vay, S., Bradshaw, J. D., and Sandholm, S.:
Measurements of atmospheric layers from the NASA DC-8 and P-3B aircraft
during PEM-Tropics A, J. Geophys. Res.-Atmos., 104, 5745–5764,
10.1029/98jd02717, 1999.Stone, D., Whalley, L. K., and Heard, D. E.: Tropospheric OH and HO2
radicals: field measurements and model comparisons, Chem. Soc. Rev., 41,
6348–6404, 10.1039/c2cs35140d, 2012.Strahan, S. E., Duncan, B. N., and Hoor, P.: Observationally derived
transport diagnostics for the lowermost stratosphere and their application to
the GMI chemistry and transport model, Atmos. Chem. Phys., 7, 2435–2445,
10.5194/acp-7-2435-2007, 2007.Thuburn, J. and Tan, D. G. H.: A parameterization of mixdown time for
atmospheric chemicals, J. Geophys. Res.-Atmos., 102, 13037–13049,
10.1029/97jd00408, 1997.Tilmes, S., Lamarque, J. F., Emmons, L. K., Kinnison, D. E., Marsh, D.,
Garcia, R. R., Smith, A. K., Neely, R. R., Conley, A., Vitt, F., Martin, M.
V., Tanimoto, H., Simpson, I., Blake, D. R., and Blake, N.: Representation of
the Community Earth System Model (CESM1) CAM4-chem within the
Chemistry-Climate Model Initiative (CCMI), Geosci. Model Dev., 9, 1853–1890,
10.5194/gmd-9-1853-2016, 2016.Turner, A. J., Fiore, A. M., Horowitz, L. W., and Bauer, M.: Summertime
cyclones over the Great Lakes Storm Track from 1860–2100: variability,
trends, and association with ozone pollution, Atmos. Chem. Phys., 13,
565–578, 10.5194/acp-13-565-2013, 2013.Wild, O., Sundet, J. K., Prather, M. J., Isaksen, I. S. A., Akimoto, H.,
Browell, E. V., and Oltmans, S. J.: Chemical transport model ozone
simulations for spring 2001 over the western Pacific: Comparisons with
TRACE-P lidar, ozonesondes, and Total Ozone Mapping Spectrometer columns, J.
Geophys. Res.-Atmos., 108, 8826, 10.1029/2002jd003283, 2003.Wofsy, S. C., Team, H. S., Team, C. M., and Team, S.: HIAPER Pole-to-Pole
Observations (HIPPO): fine-grained, global-scale measurements of climatically
important atmospheric gases and aerosols, Philos. T. R. Soc. A, 369,
2073–2086, 10.1098/rsta.2010.0313, 2011.Wu, S. L., Mickley, L. J., Jacob, D. J., Logan, J. A., Yantosca, R. M., and
Rind, D.: Why are there large differences between models in global budgets of
tropospheric ozone?, J. Geophys. Res.-Atmos., 112, D05302,
10.1029/2006jd007801, 2007.Young, P. J., Archibald, A. T., Bowman, K. W., Lamarque, J. F., Naik, V.,
Stevenson, D. S., Tilmes, S., Voulgarakis, A., Wild, O., Bergmann, D.,
Cameron-Smith, P., Cionni, I., Collins, W. J., Dalsoren, S. B., Doherty, R.
M., Eyring, V., Faluvegi, G., Horowitz, L. W., Josse, B., Lee, Y. H.,
MacKenzie, I. A., Nagashima, T., Plummer, D. A., Righi, M., Rumbold, S. T.,
Skeie, R. B., Shindell, D. T., Strode, S. A., Sudo, K., Szopa, S., and Zeng,
G.: Pre-industrial to end 21st century projections of tropospheric ozone from
the Atmospheric Chemistry and Climate Model Intercomparison Project (ACCMIP),
Atmos. Chem. Phys., 13, 2063–2090, 10.5194/acp-13-2063-2013, 2013.