ACPAtmospheric Chemistry and PhysicsACPAtmos. Chem. Phys.1680-7324Copernicus PublicationsGöttingen, Germany10.5194/acp-17-10269-2017Stratospheric ozone intrusion events and their impacts on tropospheric ozone in the Southern HemisphereGreensladeJesse W.jwg366@uowmail.edu.auAlexanderSimon P.https://orcid.org/0000-0001-6823-8857SchofieldRobynhttps://orcid.org/0000-0002-4230-717XFisherJenny A.https://orcid.org/0000-0002-2921-1691KlekociukAndrew K.https://orcid.org/0000-0003-3335-0034Centre for Atmospheric Chemistry, School of Chemistry, University of Wollongong, AustraliaAustralian Antarctic Division, Hobart, AustraliaAntarctic Climate and Ecosystems Co-operative Research Centre, Hobart, AustraliaSchool of Earth Sciences, University of Melbourne, AustraliaARC Centre of Excellence for Climate System Science, University of New South Wales, AustraliaSchool of Earth & Environmental Sciences, University of Wollongong, AustraliaJesse W. Greenslade (jwg366@uowmail.edu.au)1September20171717102691029016December20165January201726May201727July2017This 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/17/10269/2017/acp-17-10269-2017.htmlThe full text article is available as a PDF file from https://acp.copernicus.org/articles/17/10269/2017/acp-17-10269-2017.pdf
Stratosphere-to-troposphere transport (STT) provides an important natural
source of ozone to the upper troposphere, but the characteristics of STT
events in the Southern Hemisphere extratropics and their contribution to the
regional tropospheric ozone budget remain poorly constrained. Here, we
develop a quantitative method to identify STT events from ozonesonde
profiles. Using this method we estimate the seasonality of STT events and
quantify the ozone transported across the tropopause over Davis
(69∘ S, 2006–2013), Macquarie Island (54∘ S,
2004–2013), and Melbourne (38∘ S, 2004–2013). STT seasonality is
determined by two distinct methods: a Fourier bandpass filter of the vertical
ozone profile and an analysis of the Brunt–Väisälä frequency.
Using a bandpass filter on 7–9 years of ozone profiles from each site
provides clear detection of STT events, with maximum occurrences during
summer and minimum during winter for all three sites. The majority of
tropospheric ozone enhancements owing to STT events occur within 2.5 and
3 km of the tropopause at Davis and Macquarie Island respectively. Events
are more spread out at Melbourne, occurring frequently up to 6 km from the
tropopause. The mean fraction of total tropospheric ozone attributed to STT
during STT events is ∼1.0–3.5% at each site; however, during
individual events, over 10 % of tropospheric ozone may be directly
transported from the stratosphere. The cause of STTs is determined to be
largely due to synoptic low-pressure frontal systems, determined using
coincident ERA-Interim reanalysis meteorological data. Ozone enhancements can
also be caused by biomass burning plumes transported from Africa and South
America, which are apparent during austral winter and spring and are
determined using satellite measurements of CO. To provide regional context
for the ozonesonde observations, we use the GEOS-Chem chemical transport
model, which is too coarsely resolved to distinguish STT events but is able
to accurately simulate the seasonal cycle of tropospheric ozone columns over
the three southern hemispheric sites. Combining the ozonesonde-derived STT
event characteristics with the simulated tropospheric ozone columns from
GEOS-Chem, we estimate STT ozone flux near the three sites and see austral
summer dominated yearly amounts of between 5.7 and 8.7×1017 molecules cm-2 a-1.
Introduction
Tropospheric ozone constitutes only 10 % of the total ozone column but is
an important oxidant and greenhouse gas which is toxic to life, harming
natural ecosystems and reducing agricultural productivity. Over the
industrial period, increasing tropospheric ozone has been estimated to exert
a radiative forcing of 365 mW m-2,
equivalent to a quarter of the CO2 forcing . While
much tropospheric ozone is produced photochemically from anthropogenic and
natural precursors, downward transport from the ozone-rich stratosphere
provides an additional natural source of ozone that is particularly important
in the upper troposphere and references therein. The
contribution of this source to overall tropospheric ozone budgets remains
uncertain , especially in the Southern Hemisphere (SH).
Models show that stratospheric ozone depletion has propagated to the upper
troposphere . However, work based on the Southern
Hemisphere Additional OZonesonde (SHADOZ) network suggests stratospheric
mixing may be increasing upper-tropospheric ozone near southern Africa
. Uncertainties in the various processes which
produce tropospheric ozone limit predictions of future ozone-induced
radiative forcing. Here we use a multi-year record of ozonesonde observations
from sites in the SH extratropics, combined with a global
model, to better characterise the impact of stratospheric ozone on the
tropospheric ozone budget in the SH.
Stratosphere-to-troposphere transport (STT) primarily impacts the ozone
budget in the upper troposphere but can also increase regional surface ozone
levels above the legal thresholds set by air quality standards
. In the western US, for example, deep STT events during spring can
add 20–40 ppbv of ozone to the ground-level ozone concentration, which can
provide over half the ozone needed to exceed the standard set by the US
Environmental Protection Agency . Another hotspot for
STT is the Middle East, where surface ozone exceeds values of 80 ppbv in
summer, with a stratospheric contribution of 10 ppb .
Estimates of the overall contribution of STT to tropospheric ozone vary
widely e.g..
Early work based on two photochemical models showed that 25–50 % of the
tropospheric ozone column can be attributed to STT events globally, with most
contribution in the upper troposphere . In contrast, a more
recent analysis of the Atmospheric Chemistry and Climate Model
Inter-comparison Project (ACCMIP) simulations by found that
STT is responsible for 540±140 Tg yr-1, equivalent to
∼ 11 % of the tropospheric ozone column, with the remainder
produced photochemically . This wide range in model
estimates exists in part because STT is challenging to be accurately
represented, and finer model resolution is necessary to simulate small-scale
turbulence. Observation-based process studies are therefore key in
determining the relative frequency of STT events, with models then able to
quantify STT impact over large regions. Ozonesondes are particularly valuable
for this purpose as they provide multi-year datasets with high vertical
resolution.
Ozonesonde release sites and the regions used to examine STT effect on tropospheric ozone levels.
Lower-stratospheric and upper-tropospheric ozone concentrations are highly
correlated, suggesting mixing across the tropopause mainly associated with
the jet streams over the Atlantic and Pacific oceans .
Extratropical STT events most commonly occur during synoptic-scale
tropopause folds and are
characterised by tongues of high potential vorticity (PV) air descending to
lower altitudes. As these tongues become elongated, filaments disperse away
from the tongue and mix irreversibly into the troposphere. STT can also be
induced by deep overshooting convection , tropical cyclones
, and mid-latitude synoptic-scale disturbances
e.g.. STT events have been observed in
tropopause folds around both the polar front jet and the subtropical jet . The summertime pool
of high tropospheric ozone over the eastern Mediterranean is mainly
attributed to the downward ozone transport, as a result of the enhanced
subsidence and the tropopause fold activity
over the region. The eastern Mediterranean exhibits a summer maximum of
subsidence, which sits between 20 and 35∘ E and between 31 and
39∘ N, while zonally most subtropical tropopause folds occur during
winter and references therein. They are also observed
near cut-off lows , so both regional weather
patterns and stratospheric mixing are important to understand for STT
analysis.
Stratospheric ozone intrusions undergo transport and mixing, with up to half
of the ozone diffusing within 12 h following descent from the upper
troposphere . The long-range transport of enhanced ozone
can be facilitated by upper-tropospheric winds, with remarkably little
convective mixing, as shown by who measured STT air masses
2 days and thousands of kilometres from their source.
also showed how STT advection can transport stratospheric air over long
distances, with a modelled STT event spreading from the northern Pacific to
the east coast of the USA over a few days.
The strength (ozone enhancement above background levels), horizontal scale,
vertical depth, and longevity of these intruding ozone tongues vary with wind
direction and strength, topography, and season. While the frequency,
seasonality, and impacts of STT events have been well described in the
tropics and Northern Hemisphere (NH), observational estimates from the SH
extratropics are noticeably absent in the literature. The role of STT in the
SH remains highly uncertain due to the more limited data availability
compared to the NH and the temporal sparsity of these datasets
.
Here, we characterise the seasonal cycle of STT events and quantify their
contribution to the SH extratropical tropospheric ozone budget using nearly
a decade of ozonesonde observations from three locations around the Southern
Ocean spanning latitudes from 38 to 69∘S. In Sect. we describe the observations and methods used to
identify STT events and to relate STT occurrence to meteorological events. We
show how possible biomass burning smoke plume influence is detected and
handled, and we introduce the GEOS-Chem model which is used for ozone flux
estimation. Within Sect. we show the
seasonality, altitude, depth, and frequency of detected STT events, along
with a comparison of our findings to other literature where possible. In
Sect. we analyse how well GEOS-Chem captures the
tropospheric ozone seasonality and quantity near our three sites. In Sect. an extrapolation of the STT detection frequencies along
with GEOS-Chem monthly tropospheric ozone columns is used to estimate STT
ozone flux near our three sites. We also compare and contrast our results
against relevant literature. Finally, in Sect. we
examine in detail the uncertainties involved in our STT event detection
technique and ozone flux estimations.
Data and methodsOzonesonde record in the Southern Ocean
Ozonesondes provide a high vertical resolution profile of ozone, temperature,
pressure, and humidity from the surface and up to 35 km. In the troposphere,
the ozonesondes generally perform 150–300 measurements. Ozone mixing ratio is
quantified with an electrochemical concentration cell, using standardised
procedures when constructing, transporting, and releasing the ozonesondes
(http://www.ndsc.ncep.noaa.gov/organize/protocols/appendix5/).
Ozonesondes are estimated to provide around 2 % precision in the
stratosphere, which improves at lower altitudes, and ozonesondes have been
shown to be accurate to within 5 % when the correct procedures are
followed .
Ozonesondes are launched approximately weekly from Melbourne (38∘ S,
145∘ E), Macquarie Island (55∘ S, 159∘ E), and
Davis (69∘ S, 78∘ E), as shown in
Fig. . Melbourne is a major city in the south-east
of Australia and may be affected by anthropogenic pollution in the lower
troposphere. Macquarie Island is isolated from the Australian mainland,
situated in the remote Southern Ocean, and unlikely to be affected by any
local pollution events. Davis is on the coast of Antarctica and also unlikely
to experience the effects of anthropogenic pollution.
For this study, we use the 2004–2013 data for Melbourne and Macquarie Island
and the 2006–2013 data for Davis because both ozone and geopotential height
(GPH) are available from the World Ozone and Ultraviolet Data Centre archived
data in these periods.
At Davis, ozonesondes are launched twice as frequently during the ozone hole season and preceding months (June–October) as at other times of year .
A summary of ozonesonde releases at each site can be seen in Table .
Number of sonde releases at each site over the period of analysis.
Characterisation of STT events requires a clear definition of the tropopause.
Two common tropopause height definitions are the standard lapse rate
tropopause and the ozone tropopause . The
lapse rate tropopause is defined as the lowest altitude where the lapse rate
(vertical gradient of temperature) is less than 2 ∘C km-1,
provided the lapse rate averaged between this altitude and 2 km above is
also below 2 ∘C km-1. The ozone tropopause is defined as the
lowest altitude satisfying the following three conditions for the ozone
mixing ratio (OMR) :
vertical gradient of OMR is greater than 60 ppb km-1;
OMR is greater than 80 ppb; and
OMR exceeds 110 ppb between 500 and 2000 m above the altitude under inspection (modified to between 500 and 1500 m in the Antarctic, including the site at Davis).
The ozone tropopause may misdiagnose the real tropopause altitude during
stratosphere–troposphere exchange; however, it is useful at polar latitudes
in winter, where the lapse rate definition may result in artificially high
values for tropopause height .
We require lapse-rate-defined tropopauses to be at a minimum of 4 km
altitude. Another commonly used tropopause definition is determined with the
use of PV (dynamical tropopause). In the extratropics the isosurface where
PV =2 PVU (1 PVU=10-6 m2 s-1 K kg-1) is often
used to define the tropopause, allowing the 3-D representation of tropopause
folds and other tropopause features in a sufficiently resolved model
. The PV is not calculable using the ozonesonde
measurements alone, so in this work the ozone tropopause is used when
determining STT events or measured tropopause altitude.
Figure shows the monthly median ozone tropopause
altitudes at each location (solid lines). At Melbourne, the tropopause
altitude displays a seasonal cycle with maximum in summer and minimum is
winter. This seasonality is missing at Macquarie Island and almost reversed
at Davis, which has a minimum during autumn and maximum from winter to
spring. Tropopause altitude decreases with latitude from 9–14 km at
Melbourne (38∘ S) to 7–9 km at Davis (69∘ S).
Multi-year monthly median tropopause altitude (using the ozone-defined tropopause) determined from
ozonesondes measurements at Davis (2006–2013), Macquarie Island (2004–2013), and Melbourne (2004–2013) (solid lines).
Dashed lines show the 10th to the 90th percentile of tropopause altitude for each site.
Figure shows multi-year averaged ozone mixing
ratios measured by ozonesonde over the three stations. Over Melbourne,
increased ozone extending down through the troposphere is apparent from
December to March and from September to November. The increased tropospheric
ozone in these months is due to STT (in summer) and possible biomass burning
influence (in spring), both discussed in more detail in the following
sections. Over Davis and Macquarie Island, tropospheric ozone is higher
between March and October, although the seasonal differences are small
compared to those at Melbourne. The seasonality shown in
Fig. for Davis is consistent with remote free-tropospheric photochemistry determined by solar radiation availability and
temperature, resulting in higher ozone in winter .
NO2 stratospheric observations have been conducted in the Southern
Hemisphere at Lauder, Macquarie Island, and Arrival Heights
i.e., which displays a winter minima in seasonality
consistent with an ozone maxima. Influence from the ozone hole can be seen
over Davis in October, with relatively low ozone levels extending up 5–6 km
into the stratosphere.
Multi-year mean seasonal cycle of ozone mixing ratio over Davis,
Macquarie Island, and Melbourne as measured by ozonesondes.
Measurements were interpolated to every 100 m and then binned monthly.
Black and red solid lines show median ozone and lapse-rate-defined tropopause altitudes (respectively), as defined in the text.
Model description
To provide regional and global context to the ozonesonde observations, we use
the GEOS-Chem version 10-01 global chemical transport model ,
which simulates ozone along with more than 100 other trace gases throughout
the troposphere and stratosphere. Stratosphere–troposphere coupling is
calculated using the stratospheric unified chemistry extension (UCX)
. Transport is driven by assimilated meteorological fields
from the Goddard Earth Observing System (GEOS-5) maintained by the Global
Modeling and Assimilation Office (GMAO) at NASA. Ozone precursor emissions
are from the Model of Emissions of Gases and Aerosols from Nature (MEGAN)
version 2.1 for biogenic emissions, the Emissions
Database for Global Atmospheric Research (EDGAR) version 4.2 for
anthropogenic emissions, and the Global Fire Emissions Database (GFED4)
inventory for biomass burning emissions. Our simulation
was modified from the standard v10-01 to fix an error in the treatment of
ozone data from the Total Ozone Mapping Spectrometer (TOMS) satellite used to
calculate photolysis (see
http://wiki.seas.harvard.edu/geos-chem/index.php/FAST-JX_v7.0_photolysis_mechanism#Fix_for_TOMS_to_address_strange_cycle_in_OH_output.).
Our simulations span 2005–2012 (following a 1-year spin-up) with horizontal
resolution of 2∘ latitude by 2.5∘ longitude and 72 vertical
levels from the surface to 0.01 hPa. The vertical resolution is finer near
the surface at ∼60 m between levels, spreading out to ∼500 m
near 10 km altitude. For comparison to the ozonesonde observations, the
model state was saved every 6 h within the grid boxes containing each
site. When comparing against ozonesondes, GEOS-Chem UTC+0 time samples are
used for all sites. This means that the simulated ozone profiles are analysed
at local times of 07.00 for Davis and 11.00 for Macquarie Island and
Melbourne. GEOS-Chem uses the tropopause height provided by GEOS-5
meteorological fields, which are calculated using a lapse rate tropopause
definition using the first minimum above the surface in the function 0.03×T(p)-log(p), with p in hPa .
Characterisation of STT events and associated fluxes
We characterise STT events using the ozonesonde vertical profiles to identify
tropospheric ozone enhancements above a local background (in moles per
billion moles of dry air, referred to here as ppb). The process is
illustrated in Fig. on an example ozone profile. First,
the ozone vertical profiles are linearly interpolated to a regular grid with
20 m resolution from the surface to 14 km altitude. Small vertical-scale
fluctuations in ozone, which are captured by the high-resolution ozonesondes,
can be regarded as sinusoidal waves superimposed on the large vertical-scale
background tropospheric ozone. As such, the interpolated profiles are
bandpass-filtered using a fast Fourier transform to retain
these small vertical scales, between 0.5 and 5 km (removing low- and high-frequency perturbations). The high-frequency perturbations are removed as
they may represent noise in the measurements. The perturbations with scales
longer than 5 km represent the vertical gradient of ozone concentration from
the surface to the stratosphere. In what follows, these filtered vertical
profiles are referred to as perturbation profiles.
For an event to qualify as STT, a clear increase above the background ozone
level is needed, as a bandpass filter leaves us with enhancements minus any
noise or seasonal-scale vertical profile effects. We next use all the
perturbation profiles at each site to calculate the 95th percentile
perturbation value for the site. The threshold is calculated from all the
interpolated filtered values between 2 km above the surface and 1 km below
the tropopause. This is our threshold for tropospheric ozone perturbations,
and any profiles with perturbations exceeding this value in individual
ozonesondes are classified as STT events. STT events at altitudes below 4 km
are removed to avoid surface pollution, and events within 0.5 km of the
tropopause are removed to avoid false positives induced by the sharp
transition to stratospheric air. We note that this ozone detection
methodology detailed above does not allow us to resolve STT events where the
ozone flux is spread diffusely across the troposphere without a peak-like
structure in the ozonesonde profile. In other words, STT events which might
have occurred some distance and time away from the location of the ozonesonde
profiles may not be readily detected using the high-vertical-resolution, but
infrequent, ozonesonde launches.
An example of the STT identification and flux estimation methods
used in this work. The left panel shows an ozone profile from Melbourne on 8
January 2004 from 2 km to the tropopause (blue dashed horizontal line). The
right panel shows the perturbation profile created from bandpass filtering of
the mixing ratio profile. The STT occurrence threshold calculated from the
95th percentile of all perturbation profiles is shown as the orange dashed
line, and the vertical extent of the event is shown with the purple dashed
lines (see details in text). The ozone flux associated with the STT event is
calculated using the area outlined with the orange dashed line in the left
panel.
Seasonal cycle of STT event frequency at Davis (a),
Macquarie Island (b), and Melbourne (c). Events are
categorised by associated meteorological conditions as described in the text,
with low-pressure fronts (“frontal”) in dark blue, cut-off low-pressure
systems (“cut-off”) in teal, and indeterminate meteorology (“misc”) in
cyan. Events that may have been influenced by transported smoke plumes are
shown in red (see text for details).
We define the ozone peak as the altitude where the perturbation profile is
greatest between 2 km from the surface and 0.5 km below the tropopause. The
STT event is confirmed if the perturbation profile drops below zero between
the ozone peak and the tropopause, as this represents a return to
non-enhanced ozone concentrations. Alternatively, the STT event is also
confirmed if the OMR between the ozone peak and the tropopause drops below
80 ppb and is at least 20 ppb lower than the OMR at the ozone peak. If
neither of these conditions are met, the profile is rejected as a non-event.
This final step removes near-tropopause anomalies for which there is
insufficient evidence of detachment from the stratosphere. Vertical ozone
profiles recorded by ozonesondes are highly dependent on the time of launch
, and it cannot be guaranteed that detected ozone
enhancements are fully separated from the stratosphere, although this method
minimises that risk by removing detected events too near the tropopause.
We estimate the ozone flux into the troposphere associated with each event by
integrating the ozone concentration enhancement vertically over the altitude
range for which an STT event is identified (i.e. enhancement near the ozone
peak over which the perturbation profile is greater than zero). This estimate
is conservative because it does not take into account any ozone enhancements
outside of the detected peak that may have been caused by the STT and also
ignores any enhanced ozone background amounts from synoptic-scale
stratospheric mixing into the troposphere.
Our method differs somewhat from that used by to detect STT
events from ozonesonde measurements. Their definition is based on subjective
analysis of sondes released from 20 stations ranging in latitude from
35∘ S to 40∘ N. They identify an STT event if, starting
from 5 km altitude, ozone exceeds 80 ppb and then within 3 km decreases by
20 ppb or more to a value less than 120 ppb. Their technique would miss
many events due to the lower ozone concentrations found in the cleaner
SH.
Biomass burning influence
The STT detection algorithm described in
Sect. assumes all ozone enhancements are
caused by stratospheric intrusions. In some cases, however, these
perturbations may in fact reflect ozone production in lofted smoke plumes.
Biomass burning in southern Africa and South America has previously been
shown to have a major influence on atmospheric composition in the vicinity of
our measurement sites ,
particularly from July to December . On occasion,
smoke plumes from Australian and Indonesian fires can also reach the mid–high
southern latitudes, as seen from satellite measurements of carbon monoxide
(CO) discussed below.
Seasonal distribution of STT events using the alternative STT proxy,
obtained from consideration of the static stability at the ozone and lapse
rate tropopauses, for Davis (2006–2013), Macquarie Island (2004–2013), and
Melbourne (2004–2013).
The distribution of STT events' altitudes at Davis (a),
Macquarie Island (b), and Melbourne (c), determined as
described in the text. Events are coloured as described in Fig.
.
The distribution of STT events' depths, defined as the distance from
the event to the tropopause, at Davis (a), Macquarie Island
(b), and Melbourne (c), determined as described in the
text.
Events are coloured as described in Fig. .
Comparison between observed (black) and simulated (pink, red)
tropospheric ozone columns (ΩO3, in molecules cm-2)
from 1 January 2004 to 30 April 2013.
For the model, daily output is shown in pink, while output from days with ozonesonde measurements are shown in red.
For each site, the model has been sampled in the relevant grid square.
Large biomass burning events emit substantial quantities of ozone precursors,
some of which are capable of being transported over long distances and
driving ozone production far from the fire source . Ozone
production from biomass burning is complex and affected by photochemistry,
fuel nitrogen load, and time since emission, among other factors. While ozone
production occurs in some biomass burning plumes, this is not always the
case; therefore ozone perturbations detected during transported smoke events
may or may not be caused by the plume. For this reason all detected STT
events which could be caused by smoke plumes are flagged, following the
procedure outlined below. Calculations of seasonality, and ozone flux do not
include flagged events, but they are included in summary plots in this
work.
Possible biomass burning influence is identified using satellite observations
of CO from the AIRS (Atmospheric Infrared Sounder) instrument on board the
Aqua satellite . CO is emitted during incomplete combustion
and is an effective tracer of long-range transport due to its long lifetime
. In the SH, biomass burning
is the primary source of CO, making CO a good proxy for fire plumes
e.g.. To identify possible biomass burning
influence, AIRS vertical column CO is visually inspected for all dates with
detected STT events. Smoke plumes are diagnosed over areas with elevated CO
columns (∼2×1018 molecules cm-2 or higher), and any
sonde-detected STT event that occurs near (within ∼150 km of) a smoke
plume is flagged. Removal of these detections reduces the yearly estimated
ozone flux by ∼15% at Macquarie Island and ∼20% at
Melbourne.
All days with detected STT events were screened, with the exception of one
event during which there were no available AIRS data (January 2010). We find
that biomass burning may have influenced 27 events over Melbourne and
21 events over Macquarie Island. These events are flagged in the following
sections and are not used in our calculation of total STT flux. All of the
flagged events except for two occurred during the SH burning season (July to
December). No events at Davis were seen to be influenced by smoke transport.
Observed and simulated tropospheric ozone profiles over Davis,
Macquarie Island, and Melbourne, averaged seasonally.
Model medians (2005–2013 average) are shown as red solid lines, with red dashed lines showing the 10th and 90th percentiles.
Ozonesonde medians (over each season, for all years) are shown as black solid lines, with coloured shaded areas showing the 10th and 90th percentiles.
The horizontal dashed lines show the median tropopause heights from the model (red) and the observations (black).
Example comparisons of ozone profiles from ozonesondes (black) and
GEOS-Chem (red) from three different dates during which STT events were
detected from the measurements.
The dates were picked based on subjective visual analysis.
The examples show the best match between model and observations for each site.
GEOS-Chem and ozonesonde pressure levels are marked with red and black dashes respectively.
(a) Tropospheric ozone attributed to STT events.
(b) Percent of total tropospheric column ozone attributed to STT events.
Boxes show the interquartile range (IQR), with the centre line being the median, whiskers show the minimum and maximum, and circles show values which lie more than 1.5 IQR from the median.
Values calculated from ozonesonde measurements as described in the text.
Classifying synoptic conditions during STT events
Synoptic-scale weather patterns are examined using data from the European
Centre for Medium-range Weather Forecasts (ECMWF) Interim Reanalysis (ERA-I)
. This is done using the ERA-I data products over the three
sites on dates matching the detected STT events. We use the ERA-I 500 hPa
data to subjectively classify the events based on their likely meteorological
cause by visually examining each date where an event was detected. During
STT occurrence, the upper troposphere is typically characterised by nearby
cyclones, cut-off lows, or cold fronts. Over Melbourne and Macquarie Island,
we find that frontal and low-pressure activity are prevalent during STT
events (see Sect. ). Over Davis, the weather
systems are often less clear, but we see a higher portion of probable
cut-off lows. The stratospheric polar vortex may create tropopause folds
without other sources of upper-tropospheric turbulence such as low-pressure
fronts or cyclones e.g.. Cut-off
low-pressure systems can be seen clearly in synoptic-scale weather maps as
regions with lowered pressure and cyclonic winds. Low-pressure fronts in the
higher southern latitudes travel from west to east and lower the tropopause
height. We examine two cases in detail to illustrate the relationship between
synoptic-scale conditions and STT events over Melbourne. These are included
in the Supplement (Figs. S2 and S3) and show an
archetypal cut-off low and low-pressure front. To detect cut-off low-pressure
systems we look for cyclonic winds and a detached area of low pressure within
∼500 km of a site on days of event detection. For low-pressure fronts
we look for low-pressure troughs within ∼500 km. Frontal passage is a
known cause of STT as stratospheric air descends and streamers of ozone-rich
air break off and mix into the troposphere .
STT event climatologies
Figure shows the seasonal cycles of STT
frequency at Davis, Macquarie Island, and Melbourne. Frequency is determined
as detected event count divided by total launched ozonesondes for each month.
STT events in Figs. –
are coloured based on the meteorological classification described in
Sect. , with events classified as either
low-pressure fronts (“frontal”, dark blue), cut-off low-pressure systems
(“cut-off”, teal), or indeterminate (“misc”, cyan).
Events that may have been influenced by transported smoke plumes (Sect. ) are shown in red.
Ozonesonde releases are summarised in Table and detected event counts are summarised in Table .
Total number of ozonesonde detected STT events, along with the
number of events in each category (see text).
(a) Tropospheric ozone, impact per event, and
probability of event detection per sonde launch, averaged over the region
above Davis. The tropospheric ozone column ΩO3 (black, left
axis) is from GEOS-Chem, while the STT probability P (magenta, right axis)
and impact I (teal, right axis) are from the ozonesonde measurements. The
STT impact is multiplied by 10 to better show the seasonality. (b)
Estimated contribution of STT to tropospheric ozone columns over the region,
with uncertainty (shaded area) estimated as outlined in
Sect. . The black line shows STT ozone flux if event
lifetime is assumed to be 2 days, with dashed lines showing the range of
flux estimation if we assumed events lasted from 1 day to 1 week.
As described in Fig. , for the region
containing Macquarie Island.
As described in Fig. , for the region
containing Melbourne.
There is an annual cycle in the frequency of STT events
(Fig. ) with a summertime peak at all three
sites. This summertime peak is due to a prevalence of summer low-pressure
storms and fronts, which increase turbulence and lower the tropopause
. At Davis, there are more STT detections during winter
relative to our other sites, which may be due to the polar vortex and its
associated lowered tropopause and increased turbulence. STT events associated
with cut-off low-pressure systems are more prevalent during summer, while STT
events associated with frontal passage occur throughout the year. The high
frequency of STT ozone enhancements is comparable to the >25%
frequencies seen over Turkey and east of the Caspian sea in an ERA-I analysis
performed by .
The SH summer maximum we see for STT ozone flux can also be seen in Fig. 16
of , which shows seasonal flux over the Southern
Ocean,
although this is less clear over Melbourne. This seasonality is not clear in
the recent ERA-I tropopause fold analysis performed by
, where a winter maximum of tropopause fold frequency
(∼0.5% more folds in winter) over Australia can be seen to the
north of Melbourne. Their work seems to show slightly higher fold frequencies
over Melbourne in summer Fig. 5 but not to the
same extent that our summer peak suggests. Their winter maximum is in the
subtropics only – from around 20 to 40∘ S, which can be seen as the
prevalent feature over Australia in their Fig. 5. look at
modelled (CTM driven by ECMWF output) and measured ozone distributions and
find more SH ozone in the lower troposphere during austral winter, but
they note that the ECMWF fields are uncertain here again due to lack of
measurements. Their work shows a generally cleaner lower troposphere in the
SH summer but this cannot be construed to suggest more or less STT folds in
either season. examine modelled STT folds using ECMWF
output over March 2000–April 2001, and show that for this year there is a
clear austral winter maximum, again over the 20 to 40∘ S band. The
winter maximum does not include Melbourne, or the Southern
Ocean, which
explains why we see a seasonality not readily evident in these global-scale
studies.
The measurement sites are not in the regions which have a clear winter
maximum seen in Fig. 1 of , and the large-scale winter
maximum shown by all three studies seems to be dominated by the system in
that region. The seasonality of our three sites is not driven by the larger
STT system seen over the southern Indian Ocean and middle Australia which
dominates prior analysis near or over Australia.
To examine the robustness of the distributions shown in
Fig. , we developed an alternative assessment of
the seasonal occurrence of STT events, with results shown in
Fig. . Here STT occurrence is evaluated by
consideration of the square of the dry Brunt–Väisälä frequency
(N2) at the heights of the ozone tropopause (zOT) and lapse
rate tropopause (zLRT) in each ozonesonde profile that has been
binned to 100 m resolution. We use N2 to assess atmospheric stability,
which is normally distinctly higher in the stratosphere than in the
troposphere, and assume that the vertical temperature gradients within the
intrusion respond most rapidly to transported heat, which is an additional
characteristic of stratospheric air. N2 is evaluated using 250 m
resolution data (to smooth variability in the vertical gradient of potential
temperature that is due to small temperature fluctuations likely associated
with gravity waves). The altitude binning chosen is a compromise between
vertical resolution and the level of variability in N2 introduced by
temperature gradients associated with perturbations from gravity waves and
changes near the lapse rate tropopause and is the minimum that produces a
robust seasonal distribution. We define STT as having taken place if
N2(zOT) > N2(zLRT) and
zOT<zLRT; in this way the characteristically
higher static stability and ozone concentration of stratospheric intrusion is
regarded as being retained as it penetrates below the lapse rate tropopause.
The seasonal distributions shown for the three stations in
Fig. are generally similar to those shown in
Fig. (although detected events are less
frequent), with the main exception that very few events are identified with
the alternative method at Davis in the first half of the year. For our STT
proxy, we only detect intrusions where the lowest altitude of the intrusion
satisfies the ozone tropopause definition. During summer and autumn, the
vertical ozone gradients at Davis are weaker compared with the other seasons,
and the detected ozone tropopause tends to lie above the lapse rate
tropopause, potentially reducing the ability to identify STT events based on
the definition of our proxy.
Figure shows the altitudes of detected events,
based on the altitude of peak tropospheric ozone (local maximum ozone within
enhancement altitude) in the ozonesonde profile. STT event peaks most
commonly occur at 6–11 km above Melbourne and anywhere from 4 to 9 km at
Davis and Macquarie Island. There is no clear relationship between
meteorological conditions and event altitude, which may reflect the fact that
the ozonesondes observe a snapshot of an event at different stages of its
life cycle.
Figure shows the distance from the event peak to
the ozone-defined tropopause, referred to as event depth. The majority of STT
events occur within 2.5 km of the tropopause at Davis and Macquarie Island.
Over Melbourne, the event depth is more spread out, with peak ozone
enhancement generally occurring up to 6 km below the tropopause. Again,
there is no clear relationship between meteorological conditions and event
depth.
Simulated ozone columns
Figure compares the time series of
tropospheric ozone columns (ΩO3) in molecules cm-2
simulated by GEOS-Chem (red) to the measured tropospheric ozone columns
(black). GEOS-Chem outputs ozone density (molecules cm-3) and height
of each simulated box, as well as which level contains the tropopause,
allowing modelled ΩO3 to be calculated as the product of
density and height summed up to the box below the tropopause level. In both
observations and model, the maximum ozone column at Melbourne occurs in
austral summer and the minimum in winter, while Macquarie Island and Davis
show the opposite seasonality.
GEOS-Chem provides a reasonable simulation of the observed seasonality and
magnitude of ΩO3. Reduced major axis regression of observed
versus simulated ΩO3 gives a line of best fit with slopes
of 1.08 for Davis, 0.99 for Macquarie Island, and 1.34 for Melbourne. The
model is only partially able to reproduce the variability in the
observations, with paired r2 values of 0.38 for Davis, 0.18 for Macquarie
Island, and 0.37 for Melbourne. Much of the variability is driven by the
seasonal cycle, and after removing this effect (by subtracting the multi-year
monthly means), the r2 values decrease to 0.07, 0.11, and 0.30,
respectively, although the slope improves at Melbourne to 1.08.
Figure shows the observed and simulated
ozone profiles at all sites, averaged seasonally. The model generally
underestimates ozone in the lower troposphere (up to 6 km) over Davis,
although this bias is less pronounced during summer. Over Melbourne, ozone in
the lower troposphere is well represented, but the model overestimates ozone
from around 4 km to the tropopause. Over Macquarie Island we see model
overestimation of ozone above 4 km as well as underestimated ozone in the
lower troposphere, suggesting that this region is influenced by processes
seen at both of our other sites. Also shown is the mean tropopause height
simulated by the model (horizontal dashed red line), which is always higher
than the observed average, although this difference is not statistically
significant. The effect of local pollution over Melbourne during austral
summer (DJF) can be seen from the increased mean mixing ratios and enhanced
variance near the surface. The gradient of the O3 profiles is steeper in
the measurements than the model, at all sites during all seasons. Recently
examined GEOS-Chem ozone simulations and found a similar
overestimation of upper troposphere ozone in the mid-southern latitudes when
using the GEOS5 meteorological fields.
Figure compares modelled (red) and observed
(black) ozone profiles on 3 example days when STT events were detected
using the ozonesondes. The figures show the profile for each site with the
closest (qualitative) match between model and observations. The resolution
(both vertical and horizontal) of GEOS-Chem in the upper troposphere is too
low to consistently allow detection of STTs, although in a few cases (e.g.,
Melbourne in Fig. ) it appears that the
event was large enough to be visible in the model output.
Stratosphere-to-troposphere ozone flux from STT eventsMethod
We quantify the mean stratosphere-to-troposphere ozone flux due to STT events
at each site based on the integrated ozone amount associated with each STT
event (see Sect. ). Events that may have been
influenced by transported biomass burning are excluded from this calculation.
Our estimate provides a preliminary estimate of how much ozone is transported
from the stratosphere by the events detected by our method. The estimate is
conservative for several reasons: it ignores secondary ozone peaks which may
also be transported from the stratosphere, potential ozone
enhancements which may have dispersed and increased the local background
mixing ratio, and any influence from STT events nearby which may also
increase the local background ozone.
Observed tropospheric columns are calculated from the ozonesondes by
calculating the ozone number density (molecules cm-3) using measured
ozone partial pressure (PO3) and integrating vertically up to
the tropopause:
ΩO3=∫0TPPO3(z)kB×T(z)dz,
where z is the altitude (GPH), TP is the altitude at the tropopause, T is
the temperature, and kB is the Boltzmann constant.
Three regions are used to examine possible STT flux over a larger area using
modelled tropospheric ozone concentrations. The regions are shown in
Fig. . The regions are centred at each site, plus
or minus 10∘ latitude, and plus or minus 25, 16, and 11∘
longitude for Davis, Macquarie Island, and Melbourne respectively. These
boundaries approximate a rectangle centred at each site with ∼2000 km
side lengths, covering ∼4.4, 4.6, and 4.8 million square km, for
Davis, Macquarie Island, and Melbourne respectively.
To determine the ozone column attributable to STT, we determine monthly
averaged STT impact (I; fraction of tropospheric ozone sourced from the
stratosphere as shown above) and the monthly mean tropospheric ozone column
(from the GEOS-Chem multi-year mean, ΩO3) over the regions
described above. This can be expressed simply as the STT flux per event
(fluxi in each month: fluxi=ΩO3×I). Next we
determine how many events are occurring per month by assuming that only one event
can occur at one time and no event is measured twice. These assumptions
allow a simple estimate of events per month from the relatively sparse
dataset and should hold true as long as our regions of extrapolation are not
too large. The probability of any sonde launch detecting an event is
calculated as the fraction of ozonesonde releases for which an STT event was
detected, calculated for each month. We assume events last N days, then
find how many events per month we expect by multiplying the days in a month
by P and dividing by this assumed event lifetime. For example, if we expect
to see an event 25 % of the time in a month, and events last 1 day, then we
expect one event every 4 days (∼7.5 events in that month); however, if
we expect events to last a week then we would expect ∼ one event in that
month. This leads us to multiply our flux i by P and then by the term
M (M=dayspermonthN) determined by our assumed
event lifetime in order to determine monthly STT ozone flux.
The longevity of ozone events is very difficult to determine, and we have
chosen 2 days as a representative number based on several examples in
where intrusions were seen to last from 1 to 3 days
(occasionally longer) and an analysis of one large event by
showing that most of the ozone had dispersed after 48 h. Worth noting is the
recent work of , where intrusions are detected >2 days and
thousands of kilometres away from their initial descent into the troposphere
over Greenland or the Arctic. In those regions with high wind shear, mixing
appears to be slower, which allows ozone intrusions to be transported further
without dissipating into the troposphere. Relative uncertainty in our M
term is set to 50 %, as we assume these synoptic events to generally last
from 1 to 3 days.
Results
The top panel of Fig. shows the STT ozone enhancements,
based on a vertical integration of the ozone above baseline levels for each
ozonesonde where an event was detected. The area considered to be
“enhanced” ozone is outlined with yellow dashes on the left panel of
Fig. . We find that the mean ozone flux associated with STT
events is ∼0.5-2.0×1016 molecules cm-2. The bottom
panel shows the mean fraction of total tropospheric column ozone (calculated
from ozonesonde profiles) attributed to stratospheric ozone intrusions at
each site for days when an STT event occurred. First the tropospheric ozone
column is calculated, and then the enhanced ozone column amount is used to
determine the relative increase. At all sites, the mean fraction of
tropospheric ozone attributed to STT events is ∼1.0–3.5%. On
3 separate days over Macquarie and Melbourne, this value exceeds
10 %.
The upper panels in
Figs. – show the
factors I, P, and ΩO3, which are used along with the
assumed event lifetime to estimate the STT flux. The tropospheric ozone and
area of our region is calculated using the output and surface area from
GEOS-Chem over our three regions. The lower panel of these figures show the
results of the calculation when we choose 2 days for our flux estimation,
with dotted lines showing the range of flux calculated if we assume events
last from 1 day to 1 week. The seasonal cycle of ozone flux is mostly
driven by the P term, which peaks in the SH summer over all three sites.
Total uncertainty (shaded) is on the order of 100% (see
Sect. ). We calculate the annual amount based on the
sum of the monthly values. The regions over Davis, Macquarie Island, and
Melbourne have estimated STT ozone contributions of ∼5.7×1017, ∼5.7×1017, and ∼8.7×1017
molecules cm-2 a-1, respectively, or equivalently ∼2.0, 2.1,
and 3.3 Tg a-1.
Comparison to literature
show an estimate of roughly 40 to 150 kg km-2
month-1 in these regions, over all seasons (see Figs. 16, 17 in their
publication), while we estimate from 0 to 180 kg km-2 month-1 STT
impact, following a seasonal cycle with the maximum in austral summer. We
estimate higher maximum flux over Melbourne (178; 150 kg km-2
month-1 in January and February) than in either Davis (89 kg km-2
month-1 in March) or Macquarie Island (68 kg km-2 month-1 in
January). Our calculated seasonal contributions, along with total uncertainty
are shown in Table .
Seasonal STT ozone contribution in the regions near each site, in
kg km-2 month-1. In parentheses are the relative uncertainties.
This result disagrees with several other studies which have found STT ozone
fluxes in the SH extratropics are largest from autumn or winter to early
spring. used a model carrying a tracer for stratospheric
ozone to estimate STT impacts. They see higher SH tropospheric ozone
concentrations, as well as STT flux, in the SH winter. Our model also shows
ozone column amounts peaking in winter, but flux is maximised in summer
due to our detected event frequencies. examine STT using
ECMWF data for prior to 1996, using PV and Q vectors to determine STT
frequency and strength, and suggest fewer fold events in the SH occur from
December to February. used PV and winds from the GEOS
reanalysis combined with ozone measurements from the TOMS satellite to
estimate that the ozone flux between 30 and 60∘ S is
210 Tg yr-1, with the maximum occurring over SH winter.
model the upper-tropospheric ozone and its source
(emissions/lightning/stratospheric) over the Atlantic Ocean between 30 and
45∘ S and suggest that most of this is transported from the
stratosphere from March to September, which is when the subtropical jet
system is strongest.
The disagreements largely reflect the difference between point-source-based
estimates and zonally averaged estimates, as the meteorological behaviour at
our three sites is not the same as the system that dominates the SH in general. As detailed in Sect. , the
maximum STT influx which occurs during SH winter is almost entirely due to
the dominant STT system which occurs annually over the southern Indian Ocean
and middle of Australia. It is difficult to compare remote ozonesonde
datasets with area averaged models or reanalyses based on non-co-located
measurements (such as ERA).
Sensitivities and limitationsEvent detection
Our method uses several subjectively defined quantities in the process of STT
event detection. Here we briefly discuss these quantities and the sensitivity
of the method to each. Using the algorithm discussed in
Sect. , we detect 80 events at Davis, 105
(21 fire influenced) events at Macquarie Island, and 127 (27 fire influenced)
events at Melbourne.
The cut-off threshold (defined separately for each site) is determined from
the 95th percentile of the ozone perturbation profiles between 2 km above
the earth's surface and 1 km below the tropopause. We use the 95th
percentile because at this point the filter locates clear events with fewer
than 5 % obvious false positive detections. Event detection is sensitive
to this choice; for example, using the 96th, and 97th percentile instead
decreased detected events by 2, 9 (2, 10 %) at Davis, 13, 31 (11,
28 %) at Macquarie Island, and 8, 24 (6, 18 %) at Melbourne. Event
detection is therefore also sensitive to the range over which the percentile
is calculated. This range was chosen to remove anomalous edge effects of the
Fourier bandpass filter and to discount the highly variable ozone
concentration which occurs near the tropopause.
Ozone enhancements are only considered STT events if they occur from 4 km
altitude up to 500 m below the tropopause. This range removes possible
ground pollution and events not sufficiently separated from the stratosphere,
while still capturing many well-defined events that occur within 1 km of the
tropopause. An example of a well-defined event that occurs within 1 km of
the tropopause is shown in the Supplement (Fig. S2). However, STT events
which reach below 4 km are physically possible and we may have some false
negative detections due to the altitude-restricted detections.
Flux calculations
Flux is calculated as I×P×M×ΩO3, with
each term calculated as described in Sect. . The
uncertainty is determined using the standard deviation of the product, with
variance calculated using the variance of a product formula, assuming that
each of our terms is independent:
var(ΠiXi)=Πi(var(Xi)+E(Xi)2)-(ΠiE(Xi))2.
The standard deviations for the I and ΩO3 terms are
calculated over the entire dataset. These terms are considered to be
homoscedastic (unchanging variance over time). Uncertainty in assumed event
lifetime is set at 50 %, as we believe it is reasonable to expect events
to last 1–3 days. P is the probability of any ozonesonde detecting an
event and is assumed to be constant (for any month). The overall uncertainty
as a percentage is shown in parentheses in
Table ; these values are on the order of
100 %, largely due to relative uncertainty in the I factor which ranges
from 50 to 120 % for each month.
Small changes in the region do not have a large affect on the per area flux
calculations: increasing or decreasing the regions by 1∘ on each side
(∼10% change in area) changes the resulting flux by ∼1%.
However, due to the large portion of winter STT events being flagged due to
potential smoke plume influence, a significant change in the yearly flux is
seen when we do not remove these events. Without removing smoke flagged events
we see an increase in estimated yearly flux of ∼1.1,2.1×1017 molecules cm-2 yr-1 (which is a change of ∼15,20%)
over Macquarie Island and Melbourne respectively.
Considering the I factor, as discussed in here and in
Sect. , there are several uncertainties in our method
that are likely to lead to a low bias, such as the conservative estimate of
flux within each event. Although there are little available data on SH ozone
events for us to compare against, consider , who estimated
that up to 30–40 % of the ozone at 500 hPa was transported from the
stratosphere in the NH.
Our STT event impact estimates have some sensitivity to our biomass burning
filter: including smoke-influenced days increases the mean per area flux by
15–20 %. Although events which are detected near fire smoke plumes are
removed, some portion of these could be actual STTs. The change in our P
parameter when we include potentially smoke influenced events leads to a
yearly estimated STT of 11×1017 molecules cm-2 yr-1
over Melbourne, which suggests that up to 2.1×1017
molecules cm-2 yr-1 ozone enhancement could be caused by smoke
plume transported precursors. This is a potential area for improvement, as a
better method of determining smoke influenced columns would improve
confidence in our estimate.
Other possibly important uncertainties in our calculation of STT flux which
we do not cover are listed here. Filtering events which occur within 500 m
of the tropopause may also lead to more false negatives. This could also
cause lower impact estimates due to only measuring ozone enhancements which
have descended and potentially slightly dissipated. However, we have no
measure of how often the detached ozone intrusion reascends into the
stratosphere, which would lead to a reduced stratospheric impact. The
estimated tropospheric ozone columns modelled by GEOS-Chem may be biased; for
instance, suggest that in general GEOS-Chem (with GEOS-5
meteorological fields) underestimates STT, with ∼360 Tg a-1
simulated globally compared to ∼550 Tg a-1 observationally
constrained. Transport uncertainty is very difficult to estimate with the
disparate point measurements; it is possible that detected events are (at
least partially) advected out of the analysis regions, which would mean we
overestimate the influx into the region, and it is also possible that we are
influenced by STT events outside the regions of analysis. Uncertainty in
event longevity is set to 50 %, but this implies a very simplistic model
of event lifetimes. A great deal of work could be done to properly model the
regional event lifetimes, but this is beyond the scope of our work.
Uncertainties in STT ozone flux detection are ∼100% and could
be directly improved with larger or longer datasets. Possible
parameterisations and an improved model of event lifetime could also improve
the confidence in our estimate of event impacts, as well as allowing fewer
assumptions.
Conclusions
Stratosphere-to-troposphere transport can be a major source of ozone to
the remote free troposphere, but the occurrence and influence of STT events
remains poorly quantified in the southern extratropics. Ozonesonde
observations in the SH provide a satellite-independent quantification of the
frequency of STT events, as well as an estimate of their impact and source.
Using almost 10 years of ozonesonde profiles over the southern high
latitudes, we have quantified the frequency, seasonality, and altitude
distributions of STT events in the SH extratropics. By combining this
information with ozone column estimates from a global chemical transport
model, we provided a conservative first estimate of the influence of STT
events on tropospheric ozone over the Southern Ocean.
Our method involved applying a bandpass filter to the measured ozone profiles
to determine STT event occurrence and strength. The filter removed seasonal
influences and allowed clear detection of ozone-enhanced tongues of air in
the troposphere. By setting empirically derived thresholds, this method
clearly distinguished tropospheric ozone enhancements that are separated from
the stratosphere. Our method is sensitive to various parameters involved in
the calculation; however, for our sites we saw few false positive detections
of STT events.
Detected STT events at three sites spanning the SH extratropics (38, 55, and
69∘ S) showed a distinct seasonal cycle. All three sites displayed a
summer (DJF) maximum and an autumn to winter (AMJJA) minimum, although the
seasonal amplitude was less apparent at the Antarctic site (Davis) as events
were also detected regularly in winter and spring (likely due to polar-jet-stream-caused turbulence). Analysis of ERA-I reanalysis data suggested
the majority of events were caused by turbulent weather in the upper
troposphere due to low-pressure fronts, followed by cut-off low-pressure
systems. Comparison of ozonesonde-measured ozone profiles against those
simulated by the GEOS-Chem global chemical transport model showed the model
is able to reproduce seasonal features but does not have sufficient vertical
resolution to distinguish STT events.
By combining the simulated tropospheric column ozone from GEOS-Chem with
ozonesonde-derived STT estimates, we provide a first estimate of the total
contribution of STT events to tropospheric ozone in these southern
extratropical regions. We estimate that the ozone enhancement due to STT
events near our three sites ranges from 300 to 570 kg km-2 a-1,
with seasonal maximum in SH summer.
Estimating STT flux using ozonesonde data alone remains challenging; however,
the very high vertical resolution provided by ozonesondes suggests they are
capable of detecting STT events that models, reanalyses, and satellites may
not. Further work is needed to more accurately translate these ozonesonde
measurements into STT ozone fluxes, particularly in the SH where data are
sparse and STT is likely to be a major contributor to upper-tropospheric
ozone in some regions. More frequent ozonesonde releases at SH sites could
facilitate development of better STT flux estimates for this region.
All GEOS-Chem model output is available from the authors
upon request. GEOS-Chem model code is publicly available, with download and
run instructions accessible at
http://acmg.seas.harvard.edu/geos/doc/archive/man.v10-01/index.html.
Ozonesonde data come from the World Ozone and Ultraviolet Data Centre at
https://doi.org/10.14287/10000008, and are available from
http://woudc.org/data/explore.php. The ERA-Interim data were downloaded
from the ECMWF website
(http://apps.ecmwf.int/datasets/data/interim-full-daily/levtype=pl/)
following registration.
The Supplement related to this article is available online at https://doi.org/10.5194/acp-17-10269-2017-supplement.
JWG wrote the algorithms, ran the GEOS-Chem simulations, performed the analysis, and led the writing
of the paper under the supervision and guidance of SPA, RS, and JAF. AK contributed the Davis ozonesonde data and performed the analysis of
the alternate STT proxy. All authors contributed to editing and revising the manuscript.
The authors declare that they have no conflict of
interest.
Acknowledgements
We thank Sandy Burden for help clarifying some of the uncertainties involved
in methods within this work. We also thank Clare Paton-Walsh, who identified
the need to account for smoke-influenced events and provided discussions on
how to go about doing so. This research was undertaken with
the assistance of resources provided at the NCI National Facility systems at
the Australian National University through the National Computational Merit
Allocation Scheme supported by the Australian Government. This work was
supported through funding by the Australian Government's Australian Antarctic
science grant program (FoRCES 4012), the Australian Research Council's Centre
of Excellence for Climate System Science (CE110001028), and the Commonwealth
Department of the Environment ozone summer scholar program. This research is
supported by an Australian Government Research Training Program (RTP)
Scholarship.
Edited by: Andrea Pozzer
Reviewed by: three anonymous referees
ReferencesAkritidis, D., Pozzer, A., Zanis, P., Tyrlis, E., Škerlak, B., Sprenger,
M., and Lelieveld, J.: On the role of tropopause folds in summertime
tropospheric ozone over the eastern Mediterranean and the Middle East,
Atmos. Chem. Phys., 16, 14025–14039,
10.5194/acp-16-14025-2016, 2016.Alexander, S. P., Murphy, D. J., and Klekociuk, A. R.: High resolution VHF
radar measurements of tropopause structure and variability at Davis,
Antarctica (69S, 78E), Atmos. Chem. Phys., 13, 3121–3132,
10.5194/acp-13-3121-2013, 2013.Baray, J. L., Daniel, V., Ancellet, G., and Legras, B.: Planetary-scale
tropopause folds in the southern subtropics, Geophys. Res. Lett.,
27, 353–356, 10.1029/1999GL010788, 2000.Beekmann, M., Ancellet, G., Blonsky, S., De Muer, D., Ebel, A., Elbern, H.,
Hendricks, J., Kowol, J., Mancier, C., Sladkovic, R., Smit, H. G. J., Speth,
P., Trickl, T., and Van Haver, P.: Regional and global tropopause fold
occurrence and related ozone flux across the tropopause, J.
Atmospheric Chemistry, 28, 29–44, 10.1023/A:1005897314623, 1997.Bethan, S., Vaughan, G., and Reid, S. J.: A comparison of ozone and thermal
tropopause heights and the impact of tropopause definition on quantifying the
ozone content of the troposphere, Q. J. Roy.
Meteor. Soc., 122, 929–944, 10.1002/qj.49712253207, 1996.Bey, I., Jacob, D. J., Yantosca, R. M., Logan, J. A., Field, B. D., Fiore,
A. M., Li, Q.-B., Liu, H.-Y., Mickley, L. J., and Schultz, M. G.: Global
Modeling of Tropospheric Chemistry with Assimilated Meteorology: Model
Description and Evaluation, J. Geophys. Res., 106, 73–95,
10.1029/2001JD000807, 2001.Cooper, O., Forster, C., Parrish, D., Dunlea, E., Hübler, G., Fehsenfeld,
F., Holloway, J., Oltmans, S., Johnson, B., Wimmers, A., and Horowitz, L.:
On the life cycle of a stratospheric intrusion and its dispersion into
polluted warm conveyor belts, J. Geophys. Res., 109, 1–18,
10.1029/2003JD004006, 2004.Danielsen, E. F.: Stratospheric-Tropospheric Exchange Based on Radioactivity,
Ozone and Potential Vorticity, 2, 1–28,
10.1175/1520-0469(1968)025<0502:STEBOR>2.0.CO;2, 1968.Das, S. S., Ratnam, M. V., Uma, K. N., Subrahmanyam, K. V., Girach, I. A.,
Patra, A. K., Aneesh, S., Suneeth, K. V., Kishore Kumar, K., Kesarkar,
A. P., Sijikumar, S., and Ramkumar, G.: Influence of tropical cyclones on
tropospheric ozone: possible implications, Atmos. Chem.
Phys., 16, 4837–4847, 10.5194/acp-16-4837-2016, 2016.Dee, D. P., Uppala, S. M., Simmons, A. J., Berrisford, P., Poli, P., Kobayashi,
S., Andrae, U., Balmaseda, M. A., Balsamo, G., Bauer, P., Bechtold, P.,
Beljaars, A. C. M., van de Berg, L., Bidlot, J., Bormann, N., Delsol, C.,
Dragani, R., Fuentes, M., Geer, A. J., Haimberger, L., Healy, S. B.,
Hersbach, H., HÃlm, E. V., Isaksen, L., KÃllberg, P.,
Köhler, M., Matricardi, M., McNally, A. P., Monge-Sanz, B. M.,
Morcrette, J.-J., Park, B.-K., Peubey, C., de Rosnay, P., Tavolato, C.,
ThÃpaut, J.-N., and Vitart, F.: The ERA-Interim
reanalysis: configuration and performance of the data assimilation system,
Q. J. Roy. Meteor. Soc., 137, 553–597,
10.1002/qj.828, 2011.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.Edwards, D. P.: Tropospheric ozone over the tropical Atlantic: A satellite
perspective, J. Geophys. Res., 108, 4237,
10.1029/2002JD002927, 2003.Edwards, D. P., Emmons, L. K., Gille, J. C., Chu, A., Attié, J. L.,
Giglio, L., Wood, S. W., Haywood, J., Deeter, M. N., Massie, S. T., Ziskin,
D. C., and Drummond, J. R.: Satellite-observed pollution from Southern
Hemisphere biomass burning, J. Geophys. Res., 111, 1–17,
10.1029/2005JD006655, 2006.Elbern, H., Hendricks, J., and Ebel, A.: A Climatology of Tropopause Folds by
Global Analyses, Theor. Appl. Climatol., 59, 181–200,
10.1007/s007040050023, 1998.Forster, P., Ramaswamy, V., Artaxo, P., Berntsen, T., Betts, R., Fahey, D.,
Haywood, J., Lean, J., Lowe, D., Myhre, G., Nganga, J., Prinn, R., Raga, G.,
Schulz, M., and Dorland, R. V.: Changes in Atmospheric Constituents and in
Radiative Forcing, in: Climate Change 2007: The Physical Science Basis,
Contribution of Working Group I to the Fourth Assessment Report of the
Intergovernmental Panel on Climate Change, edited by: Solomon, S., Qin, D., and Man, M.,
https://www.ipcc.ch/publications_and_data/ar4/wg1/en/ch2.html (last access: 14 January 2017), 2007.Frey, W., Schofield, R., Hoor, P., Kunkel, D., Ravegnani, F., Ulanovsky, a.,
Viciani, S., D'Amato, F., and Lane, T. P.: The impact of overshooting deep
convection on local transport and mixing in the tropical upper
troposphere/lower stratosphere (UTLS), Atmos. Chem. Phys.,
15, 6467–6486, 10.5194/acp-15-6467-2015, 2015.Galani, E.: Observations of stratosphere-to-troposphere transport events over
the eastern Mediterranean using a ground-based lidar system, J.
Geophys. Res., 108, 1–10, 10.1029/2002JD002596,
2003.Giglio, L., Randerson, J. T., and Van Der Werf, G. R.: Analysis of daily,
monthly, and annual burned area using the fourth-generation global fire
emissions database (GFED4), J. Geophys. Res., 118, 317–328,
10.1002/jgrg.20042, 2013.Gloudemans, A., De Laat, J., Krol, M., Meirink, J. F., Van Der Werf, G.,
Schrijver, H., and Aben, I.: Evidence for long-range transport of carbon
monoxide in the Southern Hemisphere from SCIAMACHY observations, European
Space Agency, Special Publication, 33, 1–5, 10.1029/2006GL026804,
2007.Guenther, A. B., Jiang, X., Heald, C. L., Sakulyanontvittaya, T., Duhl, T.,
Emmons, L. K., and Wang, X.: The model of emissions of gases and aerosols
from nature version 2.1 (MEGAN2.1): An extended and updated framework for
modeling biogenic emissions, Geosci. Model Dev., 5, 1471–1492,
10.5194/gmd-5-1471-2012, 2012.
Hu, L., Jacob, D. J., Liu, X., Zhang, Y., Zhang, L., and Kim, P. S.: Global
budget of tropospheric ozone: evaluating recent model advances with satellite
(OMI), aircraft (IAGOS), and ozonesonde observations, Atmos.
Environ., 1–36, 2017.Jacobson, M. C. and Hansson, H.: Organic atmospheric aerosols: Review and
state of the science, Rev. Geophys., 38, 267–294,
10.1029/1998RG000045, 2000.Jaffe, D. A. and Wigder, N. L.: Ozone production from wildfires: A critical
review, Atmos. Environ., 51, 1–10,
10.1016/j.atmosenv.2011.11.063, 2012.Langford, A. O., Brioude, J., Cooper, O. R., Senff, C. J., Alvarez, R. J.,
Hardesty, R. M., Johnson, B. J., and Oltmans, S. J.: Stratospheric influence
on surface ozone in the Los Angeles area during late spring and early summer
of 2010, J. Geophys. Res., 117, 1–17,
10.1029/2011JD016766, 2012.Lefohn, A. S., Wernli, H., Shadwick, D., Limbach, S., Oltmans, S. J., and
Shapiro, M.: The importance of stratospheric-tropospheric transport in
affecting surface ozone concentrations in the western and northern tier of
the United States, Atmos. Environ., 45, 4845–4857,
10.1016/j.atmosenv.2011.06.014, 2011.Lelieveld, J. and Dentener, F. J.: What controls tropospheric ozone?, J.
Geophys. Res., 105, 3531–3551, 10.1029/1999JD901011, 2000.Lelieveld, J., Hadjinicolaou, P., Cammas, J.-P., Pozzer, A., Hoor, P., and
Jöckel, P.: Severe ozone air pollution in the Persian Gulf region,
Atmos. Chem. Phys., 9, 1393–1406,
10.5194/acp-9-1393-2009, 2009.Lin, M., Fiore, A. M., Cooper, O. R., Horowitz, L. W., Langford, A. O., Levy,
H., Johnson, B. J., Naik, V., Oltmans, S. J., and Senff, C. J.: Springtime
high surface ozone events over the western United States: Quantifying the
role of stratospheric intrusions, J. Geophys. Res., 117,
1–20, 10.1029/2012JD018151, 2012.Lin, M., Fiore, A. M., Horowitz, L. W., Langford, A. O., Oltmans, S. J.,
Tarasick, D., and Rieder, H. E.: Climate variability modulates western US
ozone air quality in spring via deep stratospheric intrusions, Nat.
Commun., 6, 7105, 10.1038/ncomms8105, 2015.Liu, J., Rodriguez, J. M., Thompson, A. M., Logan, J. A., Douglass, A. R.,
Olsen, M. A., Steenrod, S. D., and Posny, F.: Origins of tropospheric ozone
interannual variation over Réunion: A model investigation, J.
Geophys. Res., 1–19, 10.1002/2015JD023981, 2015.Liu, J., Rodriguez, J. M., Steenrod, S. D., Douglass, A. R., Logan, J. A.,
Olsen, M. A., Wargan, K., and Ziemke, J. R.: Causes of interannual
variability over the southern hemispheric tropospheric ozone maximum,
Atmos. Chem. Phys., 17, 3279–3299,
10.5194/acp-17-3279-2017, 2017.Mari, C. H., Cailley, G., Corre, L., Saunois, M., E, A., L, J., Thouret, V.,
and Stohl, A.: Tracing biomass burning plumes from the Southern Hemisphere
during the AMMA 2006 wet season experiment, Atmos. Chem.
Phys., 8, 3951–3961, 10.5194/acpd-7-17339-2007, 2008.Mihalikova, M., Kirkwood, S., Arnault, J., and Mikhaylova, D.: Observation of
a tropopause fold by MARA VHF wind-profiler radar and ozonesonde at Wasa,
Antarctica: comparison with ECMWF analysis and a WRF model simulation,
Ann. Geophys., 30, 1411–1421, 10.5194/angeo-30-1411-2012, 2012.Monks, P. S., Archibald, A. T., Colette, A., Cooper, O., Coyle, M., Derwent,
R., Fowler, D., Granier, C., Law, K. S., Mills, G. E., Stevenson, D. S.,
Tarasova, O., Thouret, V., von Schneidemesser, E., Sommariva, R., Wild, O.,
and Williams, M. L.: Tropospheric ozone and its precursors from the urban to
the global scale from air quality to short-lived climate forcer, Atmos.
Chem. Phys., 15, 8889–8973, 10.5194/acp-15-8889-2015, 2015.Mze, N., Hauchecorne, A., Bencherif, H., Dalaudier, F., and Bertaux, J. L.:
Climatology and comparison of ozone from ENVISAT/GOMOS and
SHADOZ/balloon-sonde observations in the southern tropics, Atmos.
Chem. Phys., 10, 8025–8035, 10.5194/acp-10-8025-2010, 2010.Olsen, M. A.: A comparison of Northern and Southern Hemisphere
cross-tropopause ozone flux, Geophys. Res. Lett., 30, 1412,
10.1029/2002GL016538, 2003.
Oltmans, J., Johnson, J., Harris, M., Bendura, J., Logan, A., and Tabuadravu,
J.: Ozone in the Pacific tropical troposphere from ozonesonde observations,
J. Geophys. Res., 106, 32503–32525, 2001.Pak, B., Langenfelds, R., Young, S., Francey, R., Meyer, C., Kivlighon, L.,
Cooper, L., Dunse, B., Allison, C., Steele, L., Galbally, I., and Weeks, I.:
Measurements of biomass burning influences in the troposphere over southeast
Australia during the SAFARI 2000 dry season campaign, J. Geophys.
Res., 108, 1–10, 10.1029/2002JD002343, 2003.
Press, W. H., Teukolsky, S. A., Vetterling, W. T., and Flannery, B. P.:
Numerical Recipes in C, 2nd Edn., The Art of Scientific Computing,
Cambridge University Press, New York, NY, USA, 1992.Price, J. D. and Vaughan, G.: The potential for stratosphere-troposphere
exchange in cut-off-low systems, Q. J. Roy.
Meteor. Soc., 119, 343–365, 10.1002/qj.49711951007, 1993.Reutter, P., Škerlak, B., Sprenger, M., and Wernli, H.:
Stratosphere-troposphere exchange (STE) in the vicinity of North Atlantic
cyclones, Atmos. Chem. Phys., 15, 10939–10953,
10.5194/acp-15-10939-2015, 2015.Rienecker, M.: File Specification for GEOS-5 DAS Gridded Output, 1–54,
https://gmao.gsfc.nasa.gov/products/documents/GEOS-5.1.0_File_Specification.pdf
(last access: 14 January 2017), 2007.Roelofs, G. J. and Lelieveld, J.: Model study of the influence of
cross-tropopause O3 transports on tropospheric O3 levels, 49,
38–55,
10.1034/j.1600-0889.49.issue1.3.x, 1997.Sinha, P., Jaeglé, L., Hobbs, P. V., and Liang, Q.: Transport of biomass
burning emissions from southern Africa, J. Geophys. Res.,
109, D20204, 10.1029/2004JD005044, 2004.Škerlak, B., Sprenger, M., and Wernli, H.: A global climatology of
stratosphere-troposphere exchange using the ERA-Interim data set from 1979 to
2011, Atmos. Chem. Phys., 14, 913–937,
10.5194/acp-14-913-2014, 2014.Škerlak, B., Sprenger, M., Pfahl, S., Tyrlis, E., and Wernli, H.:
Tropopause folds in ERA-Interim: Global climatology and relation to extreme
weather events, J. Geophys. Res., 120, 4860–4877,
10.1002/2014JD022787, 2015.Smit, H. G. J., Straeter, W., Johnson, B. J., Oltmans, S. J., Davies, J.,
Tarasick, D. W., Hoegger, B., Stubi, R., Schmidlin, F. J., Northam, T.,
Thompson, A. M., Witte, J. C., Boyd, I., and Posny, F.: Assessment of the
performance of ECC-ozonesondes under quasi-flight conditions in the
environmental simulation chamber: Insights from the Juelich Ozone Sonde
Intercomparison Experiment (JOSIE), J. Geophys. Res., 112,
1–18, 10.1029/2006JD007308, 2007.Sprenger, M., Croci Maspoli, M., and Wernli, H.: Tropopause folds and
cross-tropopause exchange: A global investigation based upon ECMWF analyses
for the time period March 2000 to February 2001, J. Geophys.
Res., 108, 10.1029/2002JD002587, 2003.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., Müller, 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., 111,
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.Stohl, A., Wernli, H., James, P., Bourqui, M., Forster, C., Liniger, M. A.,
Seibert, P., and Sprenger, M.: A new perspective of stratosphere-troposphere
exchange, B. Am. Meteorol. Soc., 84,
1565–1573, 10.1175/BAMS-84-11-1565, 2003.Struthers, H., Kreher, K., Austin, J., Schofield, R., Bodeker, G., Johnston,
P., Shiona, H., and Thomas, A.: Past and future simulations of
NO2 from a coupled chemistry-climate model in comparison with
observations, Atmos. Chem. Phys., 4, 2227–2239,
10.5194/acp-4-2227-2004, 2004.Tang, Q. and Prather, M. J.: Correlating tropospheric column ozone with
tropopause folds: The Aura-OMI satellite data, Atmos. Chem.
Phys., 10, 9681–9688, 10.5194/acp-10-9681-2010, 2010.Tang, Q. and Prather, M. J.: Five blind men and an elephant: can NASA Aura
measurements quantify the stratosphere-troposphere exchange of ozone flux?,
Atmos. Chem. Phys., 11, 2357–2380,
10.5194/acp-12-2357-2012, 2012.Terao, Y., Logan, J. A., Douglass, A. R., and Stolarski, R. S.: Contribution
of stratospheric ozone to the interannual variability of tropospheric ozone
in the northern extratropics, J. Geophys. Res., 113,
10.1029/2008jd009854, 2008.Texeira, J.: AIRS/Aqua L3 Daily Standard Physical Retrieval (AIRS-only)
1∘× 1∘ V006,
10.5067/AQUA/AIRS/DATA303 (last access: 2 December 2015), 2013.Thompson, A. M., Balashov, N. V., Witte, J. C., Coetzee, J. G. R., Thouret, V.,
and Posny, F.: Tropospheric ozone increases over the southern Africa region:
Bellwether for rapid growth in Southern Hemisphere pollution?, Atmos.
Chem. Phys., 14, 9855–9869, 10.5194/acp-14-9855-2014, 2014.Tomikawa, Y., Nishimura, Y., and Yamanouchi, T.: Characteristics of Tropopause
and Tropopause Inversion Layer in the Polar Region, SOLA, 5, 141–144,
10.2151/sola.2009-036, 2009.Trickl, T., Vogelmann, H., Giehl, H., Scheel, H. E., Sprenger, M., and Stohl,
A.: How stratospheric are deep stratospheric intrusions?, Atmos.
Chem. Phys., 14, 9941–9961, 10.5194/acp-14-9941-2014, 2014.Tyrlis, E., Škerlak, B., Sprenger, M., Wernli, H., Zittis, G., and
Lelieveld, J.: On the linkage between the Asian summer monsoon and
tropopause fold activity over the eastern Mediterranean and the Middle East,
J. Geophys. Res., 119, 3202–3221, 10.1002/2013JD021113, 2014.
Vaughan, G., Price, J. D., and Howells, A.: Transport into the troposphere in
a tropopause fold, Q. J. Roy. Meteor. Soc.,
120, 1085–1103, 10.1002/qj.49712051814, 1993.
Wauben, W. M. F., Fortuin, J. P. F., and Velthoven, P. F. J. V.: Comparison of
modeled ozone distributions observations, J. Geophys. Res.,
103, 3511–3530, 1998.Wirth, V.: Diabatic heating in an axisymmetric cut-off cyclone and related
stratosphere-troposphere exchange, Q. J. Roy.
Meteor. Soc., 121, 127–147, 10.1002/qj.49712152107, 1995.
WMO, W. M. O.: Meteorology A Three-Dimensional Science, Geneva, Second
Session of the Commission for Aerology, 4, 134–138, 1957.Young, P. J., Naik, V., Voulgarakis, A., Fiore, A. M., Horowitz, L. W.,
Lamarque, J. F., Lin, M., Prather, M. J., Bergmann, D., Cameron-Smith, P. J.,
Cionni, I., Collins, W. J., Dalsøren, 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, 2013.Zanis, P., Hadjinicolaou, P., Pozzer, A., Tyrlis, E., Dafka, S., Mihalopoulos,
N., and Lelieveld, J.: Summertime free-tropospheric ozone pool over the
eastern Mediterranean/middle east, Atmos. Chem. Phys., 14,
115–132, 10.5194/acp-14-115-2014, 2014.Zhang, L., Jacob, D. J., Yue, X., Downey, N. V., Wood, D. A., and Blewitt, D.:
Sources contributing to background surface ozone in the US Intermountain
West, Atmos. Chem. Phys., 14, 5295–5309,
10.5194/acp-14-5295-2014, 2014.