JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 108, NO. D8, 8333, doi:10.1029/2002JD002839, 2003
In situ observations of ClO near the winter polar tropopause
Brett F. Thornton,1 Darin W. Toohey,2 Linnea M. Avallone,2 Hartwig Harder,3,4
Monica Martinez,3,4 James B. Simpas,3 William H. Brune,3 and Melody A. Avery5
Received 12 August 2002; revised 20 October 2002; accepted 20 October 2002; published 19 April 2003.
[1] Significant abundances of chlorine oxide (ClO) were observed throughout the
lowermost stratosphere at high latitudes during winter from the NASA DC-8 aircraft
during the SAGE III—Ozone Loss and Validation Experiment and Third European
Stratospheric Experiment on Ozone 2000 (SOLVE/THESEO 2000) campaign. Mixing
ratios of ClO averaging 15–20 parts per trillion by volume (pptv) were observed near the
tropopause, a region where ClO abundances are usually only a pptv or less at lower
latitudes. The ratio of ClO to inferred inorganic chlorine ([ClO]/[Cly]) was found to be
largest (7%) in air characterized by low abundances of ozone (100–250 parts per
billion by volume (ppbv)). This was the region where cirrus clouds were also observed
occasionally during the measurement period, although abundances of ClO directly within
cirrus clouds were not significantly different than background abundances. Nonzero
instrument signals during darkness are attributed to detection of 5–15 pptv of OClO.
BrO mixing ratios between 2 and 4 pptv are sufficient to produce these amounts of OClO,
assuming daytime mixing ratios of ClO between 15 and 20 pptv. At these levels of ClO
and BrO, approximately 10% of the ozone at these altitudes is chemically destroyed
per month in springtime by reactions of ClO and BrO, representing an effective loss
INDEX TERMS: 0305 Atmospheric
process for ozone near the high-latitude tropopause.
Composition and Structure: Aerosols and particles (0345, 4801); 0322 Atmospheric Composition and
Structure: Constituent sources and sinks; 0320 Atmospheric Composition and Structure: Cloud physics and
chemistry; 0340 Atmospheric Composition and Structure: Middle atmosphere—composition and chemistry;
0394 Atmospheric Composition and Structure: Instruments and techniques; KEYWORDS: Arctic, ozone,
chlorine, bromine, tropopause, cirrus
Citation: Thornton, B. F., D. W. Toohey, L. M. Avallone, H. Harder, M. Martinez, J. B. Simpas, W. H. Brune, and M. A. Avery, In
situ observations of ClO near the winter polar tropopause, J. Geophys. Res., 108(D8), 8333, doi:10.1029/2002JD002839, 2003.
1. Introduction
[2] Ozone changes in the lowermost stratosphere can
impact the radiative balance of the atmosphere, with consequences for global climate. Recent indications of temperature decreases in the stratosphere [Ramaswamy et al.,
2001] and their attendant implications for global warming
further highlight the importance of ozone changes near the
tropopause. However, the cause and severity of ozone
trends in this region are uncertain [Logan et al., 1999,
World Meteorological Organization (WMO), 1998]. In
particular, it remains unclear to what extent these trends
may be due to halogen chemistry or to changes in transport
in the lowermost stratosphere [WMO, 2002].
[3] Chlorine oxide (ClO), a species that catalytically
destroys ozone, is produced in the stratosphere by photolysis and oxidation of long-lived organic chlorine source
gases (e.g., CFCs and methyl chloride). In the midlatitude
lower stratosphere, inorganic chlorine is composed primarily of the relatively stable compounds HCl and ClNO3, such
that ClO usually represents 5% of the total inorganic
chlorine budget [Brune and Anderson, 1986]. However, in
the winter polar regions, reactions on the surfaces of
particles convert these relatively stable reservoirs of chlorine into photolabile forms such as Cl2. In particular, the
following reactions are believed to be responsible for the
large enhancements of ClO observed in the polar regions in
all cold winters:
1
Department of Chemistry and Biochemistry, University of Colorado,
Boulder, Colorado, USA.
2
Program in Atmospheric and Oceanic Sciences, University of Colorado, Boulder, Colorado, USA.
3
Department of Meteorology, Pennsylvania State University, University
Park, Pennsylvania, USA.
4
Now at MPI fuer Chemie, Luftchemie, Mainz, Germany.
5
NASA Langley Research Center, Hampton, Virginia, USA.
HCl þ ClNO3 ! Cl2 þ HNO3
ð1Þ
H2 O þ ClNO3 ! HOCl þ HNO3
ð2Þ
HCl þ HOCl ! Cl2 þ H2 O:
ð3Þ
[4] In the winter polar stratosphere at altitudes of approximately 16– 24 km, reactions (1) – (3) are extremely effec-
Copyright 2003 by the American Geophysical Union.
0148-0227/03/2002JD002839$09.00
SOL
76 - 1
SOL
76 - 2
THORNTON ET AL.: ClO OBSERVATIONS NEAR THE TROPOPAUSE
tive at activating chlorine (i.e., converting reservoir forms
into reactive forms) within a few weeks in the presence of
polar stratospheric clouds [Newman et al., 1993]. By
analogy to this widely recognized ozone hole chemistry, it
has been proposed that ozone losses could occur in regions
of the lowermost stratosphere where ice and sulfate particles
might similarly enhance reactive chlorine (ClO + OClO +
Cl2O2 + Cl) [Borrmann et al., 1996; Solomon et al., 1997;
Bregman et al., 2002]. Observed concentrations of ClO in
this near-tropopause region generally have been small outside of the perturbed polar vortices [King et al., 1991;
Avallone et al., 1993b; Smith et al., 2001]. However, there
are few observations in this region, primarily because most
measurement campaigns have focused on the higher altitudes where photochemical destruction of ozone is more
rapid.
[5] Following proposals by Borrmann et al. [1996] and
Solomon et al. [1997], there has been a renewed interest in
halogen chemistry in the lowermost stratosphere. Near the
midlatitude tropopause, ozone levels appear to be decreasing [Logan et al., 1999]. Such a change could be due to
enhancements of chlorine by particles and increasing trends
in chlorine source gases [Solomon et al., 1997], or to
changes in transport [Fusco and Salby, 1999]. Although
ozone mixing ratios are small in this region of the atmosphere, the trends near the tropopause could represent a
sizable fraction of the overall trends in column ozone.
[6] In the lowermost stratosphere, cirrus clouds may
provide important surfaces for heterogeneous reactions that
convert reservoir forms of chlorine into reactive forms.
Unlike background sulfate aerosols, these clouds occur
rarely and have short lifetimes. The lowermost stratosphere
is a region where NOx mixing ratios are generally quite low
due to heterogeneous hydrolysis of N2O5 to form nitric acid.
Because the photochemical lifetime of nitric acid is long
and perturbations to reactive chlorine can persist for weeks
to months before recovering, cirrus induced enhancements
can rapidly destroy ozone at high solar zenith angles.
Models have shown activation of up to 50% of reservoir
chlorine, depending on latitude and season [Solomon et al.,
1997; Meilinger et al., 2001]. Solomon et al. [1997] showed
that, if they occur frequently, such levels of chlorine
activation could help explain observed ozone trends at
midlatitudes.
[7] Observations of enhancements of ClO are necessary
to establish the link between the hypothesis of chlorine
activation on cirrus and ozone trends observed in the lowermost stratosphere. Avallone et al. [1993a] and Borrmann
et al. [1997] found that ClO abundances in the first few km
above the midlatitude tropopause were significantly higher
than expected, such that ozone loss rates were consequently
larger by a factor of two than could be explained by a model
that uses standard photochemistry [Avallone et al., 1993a].
However, Smith et al. [2001] recently showed that ClO
abundances near the tropopause throughout the lowermost
stratosphere at middle latitudes were at or below the
detection limit (a few pptv) as a consequence of H2O
subsaturation conditions that prevail in this region. The
models of chlorine activation rely on the presence of cirrus
at altitudes where there are significant quantities of inorganic chlorine [Solomon et al., 1997]. Generally, these two
conditions (water vapor saturation and photochemical
release of inorganic chlorine) are mutually exclusive in
the stratosphere. That is, ‘‘wet’’ airmasses are typically
photochemically ‘‘young.’’ Consequently, the different conclusions reached by Borrmann et al. [1997] and Smith et al.
[2001] may reflect differences in conditions, rather than
conflicting results. To examine this discrepancy in more
detail, here we report new observations of ClO obtained
from the NASA DC-8 aircraft in the lowermost stratosphere
during winter and spring of 2000.
2. Measurements
[8 ] The measurements reported here were obtained
between December 1999 and March 2000 from the NASA
DC-8 aircraft during SOLVE/THESEO 2000 (SAGE III—
Ozone Loss and Validation Experiment and Third European
Stratospheric Experiment on Ozone). This campaign was
designed to examine the processes responsible for ozone
changes in the Arctic polar region. Here we focus on in situ
observations of ClO, ozone, temperature, and several chlorine source gases in an effort to examine the photochemical
environment of the lowermost polar stratosphere.
[9] ClO abundances were determined by resonance fluorescence of chlorine atoms produced by rapid titration of
ClO by nitric oxide. Although this technique has been
employed for such measurements in the stratosphere for
nearly three decades [e.g., Anderson et al., 1977; Brune et
al., 1988; Toohey et al., 1993], the instrument flown on the
DC-8 was new for SOLVE/THESEO 2000, and considerably different than those flown on high-altitude aircraft and
balloons. In order to optimize space and weight to limit
costs, the integration process was simplified by adding a
chlorine atom detection system into the Penn State Airborne
Tropospheric Hydrogen Oxides Sensor (ATHOS) [Brune et
al., 1995, 1998], rather than developing a stand-alone
instrument. This was done, necessarily, at the expense of
the second, dedicated, detection axis for HO2, such that OH
and HO2 were measured alternately in the first ATHOS
detection axis and ClO was detected downstream during
periods when OH was monitored.
[10] For these measurements, two sets of NO injectors
were employed. The first was intended for detection of HO2
in the upstream OH/HO2 detection axis with a 10 s duty
cycle for HO2 followed by 30 s of background for OH (and
laser diagnostics). The second injector, positioned midway
between the HOx and ClO detectors, operated with a 15 s/25
s duty cycle, coordinated with the addition of NO for HO2
measurements. Thus, in any given 40 s major cycle,
chlorine signals were obtained for 25 s with 15 s of background. Positive ClO signals were observed in both NOaddition cycles. To maximize the signal-to-noise ratio for
the results reported here we include signals from both
channels. Data were collected continuously during flight
except for short calibration periods for the HOx laser and
periods of rapid ascent or descent, generally at the beginning and end of a flight. The final calibration factors were
determined by postmission calibrations at instrumental
conditions that prevailed during flights.
[11] The inlet is mounted beneath the forward cargo bay,
aft of the front landing gear of the DC-8, with flow system,
NO injection system, and vacuum pump all installed in the
forward cargo bay [Brune et al., 1998]. In this design, air is
THORNTON ET AL.: ClO OBSERVATIONS NEAR THE TROPOPAUSE
first slowed from true airspeed of approximately 240 m s1
to 10– 40 m s1, and then drawn through a 1.5-mm diameter
orifice and through a 5-cm sampling tube into the detection
chamber by an Eaton roots blower backed by a rotary vane
vacuum pump. To maintain optimal flow for detection of
OH and HO2, the instrument was operated at 4.5– 5.5 hPa,
pressures that are much lower than normally employed for
optimum measurements of ClO. The main limiting factor
for measurements of ClO is the background signal in the
absence of nitric oxide, a term that depends on scatter from
air molecules (i.e., Rayleigh scatter) and instrument surfaces
(chamber scatter). At operating pressures higher than about
20 hPa, the magnitudes of these two terms are comparable,
and the detection limit for a 1-min integration is about
2 pptv. However, in the configuration flown on the DC-8,
chamber scatter dominates over Rayleigh scatter, necessitating a reduction in the solid angle of the detection optics,
significantly limiting the throughput of light. Thus the
measurement response was dramatically reduced, such that
significant averages (typically 15 min and longer) were
required to achieve the desired pptv precision. At these
low pressures, although the detection limit in terms of
concentration remains high (106 molecules cm3), the
minimum detectable mixing ratio is quite large as a consequence of the low density. Atomic chlorine was detected
within several ms (i.e., several cm of distance) of the
downstream injector, whereas it was detected within 20
ms when NO was added through the upstream injector. As
in previous in situ measurements, the abundance of ClO is
calculated from the difference of signals with and without
NO, on the basis of laboratory calibrations as described by
Brune et al. [1988] and Toohey et al. [1993].
[12] For abundances of NO that were added to detect ClO
during this campaign, the instrument is also sensitive to
chlorine atoms produced by OClO. However, it is relatively
insensitive to Cl2O2 because thermal decomposition of this
species within the ATHOS flow system and the reaction of
Cl2O2 with NO are very slow (less than 1 s1) under the
conditions employed to measure ClO. The sensitivity to
OClO has been a minor concern during previous campaigns
because much smaller concentrations of NO were employed
to detect ClO, such that only small fractions of OClO could
have been converted to chlorine atoms, and, at the higher
altitudes typically sampled, detectable OClO abundances
are generally accompanied by large enhancements of ClO,
such that signals from the latter dominate those from OClO.
As we will show later, apparently this was not the case for
the lower-altitude DC-8 observations in darkness.
[13] Cly was determined by the method described by
Engel et al. [1997], on the basis of gas chromatographic
(GC) measurements of CFC-11, CFC-12, and Halon-1211
from the Tropospheric Ozone and Tracers from Commercial
Aircraft Platforms (TOTCAP) instrument. These species
were measured approximately every 3.5 min with a precision of 1% and an accuracy of 2%. Ozone was measured
by observing near-infrared chemiluminescence from
excited-state nitrogen dioxide formed by the reaction of
pure reagent nitric oxide with ozone in sampled air. This
well-established technique is described by Gregory et al.
[1987], as adapted for use on aircraft. These measurements
are inherently faster than those of ClO for these studies and
were averaged to match the rate of the ClO observations
SOL
76 - 3
Figure 1. In situ observations of ClO (pptv), ozone
(ppbv), and local solar zenith angles (SZA in degrees)
plotted versus UT (s) for the flight of 27 January 2000.
Vertical error bars on the ClO measurements represent 1s
precision, whereas the horizontal bars indicate the time
period over which the measurements were averaged. Ozone
is plotted at the native 1 Hz sampling frequency. SZAs are
calculated from DC-8 navigational data.
(minutes per data point). However, in Figure 1 we show the
1-Hz ozone data to emphasize the variability of this species
on finer scales.
3. Results and Discussion
[14] Temporal profiles of ClO, ozone, and solar zenith
angle (SZA) are shown in Figure 1 for the flight of 27
January 2000. This flight in highlighted because it represents one of the few in which the DC-8 spent most of the
time in sunlight (or twilight), where we might expect to
detect ClO. There is some correlation between the abundances of ClO and ozone in sunlight, as has been reported
previously for measurements at higher altitudes [King et al.,
1991; Avallone et al., 1993b]. This has been interpreted to
be the result of increased production of inorganic chlorine,
where the mixing ratio of ozone is a reasonable indicator of
photochemical age in the lowermost stratosphere because it
is produced in regions of the stratosphere where source
gases of chlorine are photolyzed. Although there were other
flights that exhibited similar correlations between ClO and
O3, the correspondence was not as strong as has been
observed previously at midlatitudes [e.g., Avallone et al.,
1993b]. Rather, it was often the case that ClO mixing ratios
were typically between 10 pptv and 20 pptv, even though
ozone abundances varied over the range 200– 500 ppbv.
These ClO mixing ratios are considerably higher (as much
as a factor of 10) than those observed previously at similar
abundances of ozone. Such an enhancement could be due to
activation of chlorine for a given value of available inorganic chlorine (Cly) or to a substantial increase in Cly.
[15] To further investigate the source of enhanced ClO, in
Figure 2 we plot averages of ClO and Cly against ozone in
50 ppbv bins. Although the results of this binning do not
necessarily cover the same segments of flights for ClO and
Cly, we do not expect Cly to vary significantly with respect
SOL
76 - 4
THORNTON ET AL.: ClO OBSERVATIONS NEAR THE TROPOPAUSE
Figure 2. ClO and Cly mixing ratios (in pptv), averaged
with respect to ozone in 50 ppbv-wide bins, plotted versus
ozone (in ppbv). ClO error bars represent 1s precision.
to ozone at these latitudes because both quantities have long
lifetimes. As expected, Cly increases nearly monotonically
with increasing ozone whereas ClO increases sharply from
the upper troposphere (e.g., O3 < 100 ppbv) into the lowermost stratosphere (O3 > 200 ppbv). ClO abundances level
off at intermediate ozone values until ozone exceeds 500
ppbv. The possible increase (at the 1s uncertainty level) in
ClO at the highest ozone values occurred when the DC-8
ascended to its highest cruise altitudes near the ends of
several flights. This is a region where ClO has been
observed to increase rapidly with altitude [see, e.g., Pierson
et al., 1999], and it is likely that the DC-8 was penetrating
the lowest part of the perturbed polar vortex during these
few segments.
[16] The ratio ClO/Cly is a measure of the extent of
activation of chlorine, and as evident in Figure 3, for ozone
abundances below 600 ppbv this ratio exhibits a maximum
in a narrow region just at and above the tropopause where
ClO mixing ratios increased sharply. Assuming that inorganic chlorine abundances are not locally enhanced in this
region by a process such as sedimentation of chlorinecontaining particles, this narrow maximum in ClO/Cly ratio
is indicative of repartitioning of inorganic chlorine from
reservoirs (e.g., HCl and ClNO3) to reactive forms (e.g.,
ClO either by substantial reductions in NOx or by heterogeneous reactions on particle surfaces), whereas typical
values for ClO/Cly observed at lower latitudes are 1 –
2%. Significant masses of particles were frequently detected
in the lowermost stratosphere by the TOTCAP total water
instrument during SOLVE/THESEO 2000 [Kondo et al.,
2003; A. G. Hallar et al., manuscript in preparation, 2003].
However, we cannot rule out the possibility that the
enhancements of ClO result from a gas-phase repartitioning
of Cly due to loss of NOx. We have examined NOx
measurements from the DC-8 during SOLVE/THESEO
2000 for signs of variability that might correlate with the
observed enhancements in the ClO/Cly ratio. Unfortunately,
throughout the campaign the mixing ratios of NOx were
very small, at or below the limit at which chemiluminescent
measurements are considered reliable.
[17] On a number of occasions during SOLVE/THESEO
2000, the DC-8 was specifically directed into regions where
cirrus clouds (or their influences) might be expected. Based
on the proposals by Borrmann et al. [1996] and Solomon et al.
[1997] one might expect to see fairly large enhancements in
ClO at these latitudes, provided cirrus clouds are present
during this season. This is because deactivation of reactive
chlorine at high solar zenith angles is slow and the models
predict upward of 50% conversion of reservoir forms to active
forms within cirrus. Based on abundances of Cly inferred
from the gas-chromatograph measurements (Figure 2), we
might expect regions of the atmosphere with 100 pptv or
greater ClO [see, e.g., Solomon et al., 1997, Figure 7]. In
fact, the DC-8 flew through thin cirrus on several occasions, as indicated by measurements of large ice particles
by the TOTCAP H2O instrument (for example, on 23 and
25 January and 8, 9, and 11 March) (A. G. Hallar et al.,
manuscript in preparation, 2002). In most cases the abundances of ozone were quite low, indicating that the cirrus
had formed very near the tropopause, as inferred previously
for this region [Murphy et al., 1990]. However, there were
occasions where cirrus occurred in air with significant
stratospheric character (e.g., ozone abundances greater than
200 ppbv). These air masses also exhibited significant
tropospheric character (e.g., they were ‘‘wet,’’ with total
water in excess of 50 parts per million by volume (ppmv)).
These airmasses were likely produced by exchange and
mixing of stratospheric and tropospheric air.
[18] The distribution of ClO within cirrus did not look
significantly different from that observed in background air
(i.e., outside of cirrus). On many occasions, ClO abundances were at or below the detection limit, whereas on several
occasions values in excess of 75 pptv were observed. For
those occasions when the ClO was low, ozone abundances
were typically representative of upper tropospheric air (i.e.,
Figure 3. ClO/Cly ratio as a function of ozone in 50 ppbv
bins. ClO values are from all local (polar) flights, whereas
Cly was determined from TOTCAP gas chromatograph
measurements on four local flights in January and March.
ClO/Cly represents the extent of activation of chlorine from
longer-lived reservoir species.
THORNTON ET AL.: ClO OBSERVATIONS NEAR THE TROPOPAUSE
the clouds were tropospheric in character). On occasions
when ClO exceeded 10 pptv, the airmasses had stratospheric
character. It is tempting to try to look for a direct correlation
between ClO abundances and particle surface areas in
cirrus. Unfortunately, the character of the air varied significantly during cloud encounters, and with the reduced
sensitivity of the instrument, fairly long averages are
required to derive meaningful results. However, histograms
of ClO measured in and out of regions with enhanced
particle surface areas are remarkably similar, showing no
clear systematic increase in the number of occurrences of
high ClO (>50 pptv). This does not necessarily imply that
cirrus clouds are ineffective at activation of chlorine. Rather,
we believe that at these very high latitudes the deactivation
of chlorine occurs too slowly to compete with reactivation
processes. Therefore relatively high abundances of ClO
observed at these latitudes (for example, see Figure 5)
reflect the combination of enhanced particle surface area
and reduced rate of production of NOx.
[19] Figure 4 is a summary of ClO as a function of ozone
for all SOLVE DC-8 flights except for the transits between
Kiruna and NASA Dryden (38N). The transits have been
excluded because they typically sampled midlatitude air of
tropospheric character. The day bin was defined as solar
zenith angles (SZA) less than 88, and the night bin was
solar zenith angles greater than 95. Sunrises and sunsets
bridge these two regimes, and were differentiated by the
local time at the measurement location (i.e., rising Sun and
setting Sun before and after local noon, respectively). The
maximum value of the daytime ClO average (18– 20 pptv)
occurred between 100 and 400 ppbv of ozone, whereas
below 100 ppbv of ozone ClO was essentially zero, as
might be expected for air of primarily tropospheric origin.
ClO mixing ratios of 20 pptv and larger at the higher levels
of ozone (>600 ppbv) are consistent with previous observations at these latitudes in winter [Avallone et al., 1993a,
1993b].
[20] A curious result is that mixing ratios of ClO of 5 pptv
and larger were often observed in darkness. As clearly seen
in Figure 1, in the region near the end of the flight when the
DC-8 ascended to highest flight altitudes, apparent ClO
abundances in the dark even exceeded those in daylight
earlier in the flight. Previous observations in the lower
stratosphere have found ClO to decrease to sub-pptv abundances following sunset [Brune et al., 1990; Toohey et al.,
1991], consistent with production of nighttime reservoirs
such as HOCl and ClNO3, except in regions of large
enhancements of ClO [Avallone and Toohey, 2001].
Figure 4 illustrates day/night differences in ClO abundances
binned with respect to ozone. In this case, the observations
of ClO and Cly are not necessarily in the same airmasses.
Except very near the tropopause (O3 < 100 ppbv), ClO
values in daylight are equal to or greater than those
observed in darkness, consistent with photochemical production of ClO. Normally, ClO would be expected to react
with BrO and NO2 within several hours after sunset to form
stable reservoirs. The most likely explanation for the
surprisingly high signals observed in darkness is that the
ClO instrument was detecting the presence of OClO in these
air masses. Assuming a rate constant of (3 – 4) 1013
cm3 molecule1 s1 for OClO + NO ! ClO + NO2 at
temperatures within the flow tube [Li et al., 2002] we
SOL
76 - 5
Figure 4. ClO mixing ratios as a function of ozone in 100
ppbv bins. The results are segregated by day (SZA < 88)
and night (SZA > 95). Error bars represent 1s precision.
estimate that 75% or more of the OClO present would
be converted into chlorine atoms between the NO injector
and chlorine atom detection volume (because ClO is rapidly
converted into Cl). On this basis, we conclude that a
majority of the signal detected in darkness was likely the
result of titration of most of the existing OClO to chlorine
atoms. It is worth considering the possibility that nonzero
signals for ClO in darkness could be due to an unknown
artifact in the ClO detection region. However, extensive
laboratory tests of synthetic air (including ozone) failed to
reveal any positive signals in the NO addition cycle. In fact,
at ambient pressures (e.g., in measurements from the ER-2),
a small negative signal results from absorption of the
excitation light by NO.
[21] Because tropopause height increases with decreasing
latitude, and the DC-8 did not fly south of NASA Dryden,
the amount of lowermost stratosphere ClO data available
from SOLVE/THESEO 2000 decreases with decreasing
latitude. There is no data south of 38N; however, for these
lower latitudes we have extensive observations of ClO in the
lowermost stratosphere from the 1998 WAM (WB-57 Aerosol Mission) and 1999 ACCENT (Atmospheric Chemistry of
Combustion Emissions Near the Tropopause) campaigns
[Ross et al., 2000]. Figure 5 shows a composite of WB-57
and DC-8 ClO mixing ratios in daylight for a fixed ozone bin
in the lowermost stratosphere versus latitude. The highest
average abundances of ClO, between 15 and 20 pptv, were
observed at higher latitudes (55 N and above), with values
between 5 and 10 pptv for the range 35 and 55 N. Below
35 N, ClO mixing ratios are less than a few pptv, and not
statistically different from zero, consistent with earlier
reports [Avallone et al., 1993b; Smith et al., 2001]. Although
not shown here, throughout the WAM and ACCENT campaigns the lowermost stratosphere was subsaturated with
respect to water. In addition, lower latitudes experience
higher insolation, resulting in more abundant NOx to sequester active chlorine. Thus it is not surprising that the largest
mixing ratios of ClO are observed in the more wet and dark
SOL
76 - 6
THORNTON ET AL.: ClO OBSERVATIONS NEAR THE TROPOPAUSE
Figure 5. ClO mixing ratios averaged in 10 latitude bins.
Only observations obtained at solar zenith angles less than
88 and with ozone ranging from 150– 400 ppbv are
included. This selects sunlit air that is in the lowermost
stratosphere. Measurements at latitudes below 35N were
obtained from the NASA WB-57 aircraft during the WAM
and ACCENT missions. Error bars represent 1s precision.
regions of the stratosphere. However, we refrain from
drawing stronger conclusions regarding the link between
elevated ClO and particles for several reasons. First, the WB57 observations were obtained in April, May, and September, whereas the DC-8 observations were made in winter
months (December – March). Second, the WB-57 and DC-8
instruments employ different detection strategies, and it is
possible that there is an unknown systematic bias between
the two instruments. In regards to this second point, we note
that the DC-8 instrument very often saw no detectable ClO at
low latitudes (e.g., on the southernmost ends of transit flights
to and from Sweden). These observations were primarily in
the uppermost troposphere where very little active chlorine
is expected.
[22] The source of nighttime OClO is very likely the
reaction of BrO + ClO ! OClO + Br. Although we
attempted to measure BrO on the DC-8 during SOLVE/
THESEO 2000, abundances were below the detection limit
(>5 pptv) of the instrument under the low operational
pressures in the ATHOS flow system. Using a 0-dimensional photochemical model, described elsewhere [Pierson
et al., 1999], we can estimate the mixing ratios of BrO that
would be necessary to produce the chlorine atom signals
that we observed in darkness, with the explicit assumption
that those signals represented the sum of ClO and OClO
with nearly equal conversion efficiency. The key species for
simulating OClO are ClO, BrO, and NO2, the last being
important for limiting abundances of ClO and BrO through
formation of long-lived reservoirs ClNO3 and BrNO3.
Simulations that successfully reproduce OClO in the range
observed in darkness (5 – 10 pptv) have the following
features. Exchange between ClO and OClO occurs within
a few hours of sunset under conditions encountered at
12 km, and abundances of OClO are strongly dependent
on the presence of NO2 and BrO. In particular, daylight
abundances of NO2 must be relatively low, below 40 pptv,
and those of BrO must be 2 pptv and greater. Finally, we
find that at high mixing ratios of BrO, the maximum
abundance of OClO depends strongly on available ClO,
but is insensitive to change in BrO (i.e., OClO saturates at
high BrO). We note here that significantly more OClO than
Cl2O2 is generated in the model at sunset. This is because
the rate for the ClO + BrO reaction is greater than that for
the ClO + ClO association. In addition, for temperatures
typical of the observations from the DC-8, thermal decomposition of Cl2O2 is significant. Thus we believe that OClO
is the dominant reactive chlorine species at night.
[23] In Figure 6, we show the sensitivity of nighttime
OClO production to daylight BrO abundances, the latter
being the independent variable for the model runs. We have
varied the daytime BrO abundances from 0.5 pptv to at least
10 pptv, above reasonable values at these altitudes. The
shaded region indicates the most likely (i.e., lower range)
values of BrO necessary to reproduce the observed values of
OClO in darkness, constrained by measurements of ClO in
daylight under similar conditions (e.g., based on ozone
mixing ratios). Although not evident on this plot, there is
a good correlation between the simulated abundances of
BrO and the abundances of ozone used as the indicator for
‘‘age of air’’ in this region of the stratosphere.
[24] In Figure 7 we plot the daytime BrO as inferred from
the model/observation relationship in Figure 6 versus altitude, overlaying previous measurements of BrO at high
latitudes using remote sensors on balloons. The abundances
inferred at DC-8 altitudes are in excellent agreement with
the earlier measurements at similar latitudes, further supporting the inference that OClO is present at levels necessary to explain our typical signals in darkness (5 – 10 pptv
of ClO + OClO). There were occasions when the observed
nighttime ClO + OClO exceeded 10 pptv, but these only
Figure 6. Plot of modeled nighttime OClO mixing ratios
produced by the reaction ClO + BrO as a function of
daytime BrO mixing ratios. Labels on the individual curves
refer to the daytime mixing ratios of ClO assumed as initial
conditions. The patterned region indicates the range of
values that correspond to the ClO observations, assuming
that signals measured in darkness are due to OClO.
THORNTON ET AL.: ClO OBSERVATIONS NEAR THE TROPOPAUSE
Figure 7. Comparison of the BrO profile determined from
the model runs shown in Figure 6 with previous profiles
obtained from 67N in winter with balloonborne DOAS
instruments [Harder et al., 1998; Pundt, 1997]. The two
highest-altitude points inferred from the DC-8 data
represent minima because the inferred mixing ratios of
OClO in darkness were at the asymptote of the surface
defined by ClO, OClO, and BrO in Figure 6.
occurred when ozone abundances were relatively high
(>600 ppbv). Such air has relatively large abundances of
Cly, and presumably high Bry, as supported by the observations of low Halon-1211 by the TOTCAP gas chromatograph during these times. Under such conditions, daytime
ClO mixing ratios were expected to be in excess of 20 pptv
[Avallone et al., 1993b], and BrO mixing ratios required to
explain the chlorine signals in darkness were larger than 4
pptv (e.g., Figure 7), values more typical of those at higher
altitudes. Model runs under these conditions indicate that a
significant amount of ClO remains in darkness together with
OClO because there is inadequate BrO to titrate out all the
ClO. That is, BrO is converted to BrCl before all the ClO is
converted to OClO because there is a large excess of ClO
over BrO. This situation is similar to that at higher altitudes
within the perturbed polar vortex. It is also necessary for
NO2 in darkness to be very low (a few pptv or less), well
below the detection limit for the NOx measurement on the
DC-8 during SOLVE/THESEO 2000 [Kondo et al., 1997].
This inference is in contrast to measurements of NO2
reported during the campaign [Gaines, 2000], which were
typically greater than 20 pptv in darkness. We are forced to
conclude that there is unknown chemistry that can sustain
high abundances of ClO and BrO in the presence of some
NO2 or that the measurements of NO2 in darkness are in
error. Recent observations of the coexistence of OClO and
NO2 at high latitudes in winter may shed light on this issue
[Rivière et al., 2002].
4. Conclusions
[25] These are the first observations of ClO focused on
the lowermost stratosphere, especially near the tropopause.
We have found mixing ratios of this species, an indicator of
SOL
76 - 7
ozone destruction by chlorofluorocarbons, to be significantly enhanced (a factor of ten or more) compared to
values that have been observed at lower latitudes [Avallone
et al., 1993b; Smith et al., 2001]. In addition, another
compound that produces chlorine atoms upon addition of
NO was detected in darkness. Abundances of this reactive
compound are consistent with OClO generated from reaction of ClO with BrO.
[26] On the basis of the recommended rate constants for
reactions of ClO and BrO with each other and with HO2
[Sander et al., 2000], chemical loss of ozone is estimated to
be 1 ppbv day1 at vernal equinox for the concentrations
of ClO and HO2 (2 pptv) that were measured and those of
BrO that were inferred. Thus 10– 12% of the ozone would
be destroyed by catalytic reactions of these species in about
a month at these latitudes and altitudes (i.e., poleward of 55
N at 12 km). This destruction is occurring over a region that
is considerably larger than that of the perturbed polar vortex
at higher altitudes. Consequently, it is very likely that the
trends that have been observed in the ozone-sonde data in
winter and spring [Logan et al., 1999] are the result of
halogen photochemistry and trends in inorganic chlorine
and bromine.
[27] Although we have observed appreciable levels of
active chlorine in the high-latitude lowermost stratosphere,
the mechanism responsible for maintaining these high levels
is unclear. It is likely that the enhanced particle surface areas
observed by the TOTCAP particulate water instrument are
sufficient to periodically activate chlorine and reduce NOx
throughout the winter. However, in this region of the
atmosphere it is difficult to separate the effects of chemistry
and transport. At these high latitudes in winter, chlorine
deactivation is slow. It may be the case that the definitive
signature of chlorine activation by cirrus will be found at
lower latitudes, where insolation is greater and recovery of
active chlorine more rapid.
[28] Acknowledgments. This work was funded by the NASA Upper
Atmosphere Research Program, grant number NAG2-1307, under the
direction of Mike Kurylo and Phil Decola. We would like to thank the
NASA Dryden DC-8 pilots, flight and ground crews, and mission support
personnel too numerous to mention by name. We thank Bob Lesher for
assistance with integration. In situ NOy measurements were provided by
Yutaka Kondo and colleagues at the Research Center for Advanced Science
and Technology, University of Tokyo. DWT acknowledges support from
the University of Colorado for a temporary leave to carry out this work.
References
Anderson, J. G., J. J. Margitan, and D. H. Stedman, Atomic chlorine and
the chlorine monoxide radical in the stratosphere: Three in situ observations, Science, 198, 501 – 503, 1977.
Avallone, L. M., and D. W. Toohey, Tests of halogen photochemistry using
in situ measurements of ClO and BrO in the lower polar stratosphere,
J. Geophys. Res., 106, 10,411 – 10,421, 2001.
Avallone, L. M., D. W. Toohey, W. H. Brune, R. J. Salawitch, A. E. Dessler,
and J. G. Anderson, Balloonborne in situ measurements of ClO and
ozone: Implications for heterogeneous chemistry and midlatitude ozone
loss, Geophys. Res. Lett., 20, 1795 – 1798, 1993a.
Avallone, L. M., D. W. Toohey, M. H. Proffitt, J. J. Margitan, K. R. Chan,
and J. G. Anderson, In situ measurements of ClO at midlatitudes: Is there
an effect from Mt. Pinatubo?, Geophys. Res. Lett., 20, 2519 – 2522,
1993b.
Borrmann, S., S. Solomon, J. Dye, and B. Luo, The potential of cirrus
clouds for heterogeneous chlorine activation, Geophys. Res. Lett., 23,
2133 – 2136, 1996.
Borrmann, S., S. Solomon, L. Avallone, D. Toohey, and D. Baumgardner,
On the occurrence of ClO in cirrus clouds and volcanic aerosol in the
tropopause region, Geophys. Res. Lett., 24, 2011 – 2014, 1997.
SOL
76 - 8
THORNTON ET AL.: ClO OBSERVATIONS NEAR THE TROPOPAUSE
Bregman, B., P.-H. Wang, and J. Lelieveld, Chemical ozone loss in the
tropopause region on subvisible ice clouds, calculated with a chemistrytransport model, J. Geophys. Res., 107(D3), 4032, doi:10.1029/
2001JD000761, 2002.
Brune, W. H., and J. G. Anderson, In situ observations of midlatitude
stratospheric ClO and BrO, Geophys. Res. Lett., 13, 1391 – 1394, 1986.
Brune, W. H., E. M. Weinstock, and J. G. Anderson, Midlatitude ClO below
22km altitude: Measurements with a new aircraft-borne instrument, Geophys. Res. Lett., 15, 144 – 147, 1988.
Brune, W. H., D. W. Toohey, S. A. Lloyd, and J. G. Anderson, The sunrise
and sunset variation of ClO in the lower stratosphere, Geophys. Res. Lett.,
17, 509 – 512, 1990.
Brune, W. H., P. S. Stevens, and J. H. Mather, Measuring OH and HO2 in
the troposphere by laser-induced fluorescence at low pressure, J. Atmos.
Sci., 52, 3328 – 3336, 1995.
Brune, W. H., et al., Airborne in-situ OH and HO2 observations in the
cloud-free troposphere and lower stratosphere during SUCCESS, Geophys Res. Lett., 25, 1701 – 1704, 1998.
Engel, A., U. Schmidt, and R. A. Stachnik, Partitioning between chlorine
reservoir species deduced from observations in the Arctic winter stratosphere, J. Atmos. Chem., 2, 107 – 126, 1997.
Fusco, A. C., and M. L. Salby, Interannual variations of total ozone and
their relationship to variations of planetary wave activity, J. Clim., 12,
1619 – 1629, 1999.
Gaines, S., (Ed.), SOLVE Mission Data set[CD-ROM], NASA/UARP-0009,
NASA Ames Res. Cent., Moffett Field, Calif., 2000.
Gregory, G. L., C. H. Hudgins, J. Ritter, and M. Lawrence, In situ ozone
instrumentation for 10-Hz measurements: Development and evaluation,
in Proceedings of 6th AMS Symposium on Instrumentation, 12 – 16 Jan.,
pp. 136 – 139, Am Meteorol. Soc., Boston, Mass., 1987.
Harder, H., et al., Stratospheric BrO profiles measured at different latitudes
and seasons: Atmospheric observations, Geophys. Res. Lett., 25, 3843 –
3846, 1998.
King, J. C., W. H. Brune, D. W. Toohey, J. M. Rodriguez, W. L. Starr, and
J. F. Vedder, Measurements of ClO and O3 from 21N to 61N in the
lower stratosphere during February 1988: Implications for heterogeneous
chemistry, Geophys. Res. Lett., 18, 2273 – 2276, 1991.
Kondo, Y., S. Kawakami, M. Koike, D. W. Fahey, H. Nakajima, Y. Zhao,
N. Toriyama, M. Kanada, G. W. Sachse, and G. L. Gregory, Performance
of an aircraft instrument for the measurement of NOy, J. Geophys. Res.,
102, 28,663 – 28,671, 1997.
Kondo, Y., et al., Uptake of reactive nitrogen on cirrus cloud particles in the
upper troposphere and lowermost stratosphere, Geophys. Res. Lett.,
30(4), 1154, doi:10.1029/2002GL016539, 2003.
Li, Z. J., R. D. Wuebbles, and N. J. Pylawka, Rate constant measurement
for the reaction of OClO with NO at 220 – 367 K, Chem. Phys. Lett., 354,
491 – 497, 2002.
Logan, J. A., et al., Trends in the vertical distribution of ozone: A comparison of two analyses of ozonesonde data, J. Geophys. Res., 104, 26,373 –
26,399, 1999.
Meilinger, S. K., B. Karcher, R. von Kuhlmann, and T. Peter, On the impact
of heterogeneous chemistry on ozone in the tropopause region, Geophys.
Res. Lett., 28, 515 – 518, 2001.
Murphy, D. M., K. K. Kelly, A. F. Tuck, M. H. Proffitt, and S. Kinne, Ice
saturation at the tropopause observed from the ER-2 aircraft, Geophys.
Res. Lett., 17, 353 – 356, 1990.
Newman, P., L. R. Lait, M. Schoeberl, E. R. Nash, K. Kelly, D. W. Fahey,
R. Nagatani, D. Toohey, L. Avallone, and J. Anderson, Stratospheric
meteorological conditions in the arctic polar vortex, 1991 to 1992,
Science, 261, 1143 – 1146, 1993.
Pierson, J. M., K. A. McKinney, D. W. Toohey, J. Margitan, U. Schmidt,
A. Engel, and P. A. Newman, An investigation of ClO photochemistry
in the chemically perturbed arctic vortex, J. Atmos. Chem., 32, 61 – 81,
1999.
Pundt, I., Étude de la distribution verticale des oxydes d’iode et de bromine
dans la basse stratosphére à l’aide de sonde SAOZ-balloon, Ph.D. thesis,
Univ. de Paris VI, Paris, 1997.
Ramaswamy, V., et al., Stratospheric temperature trends: Observations and
model simulations, Rev. Geophys., 39, 71 – 122, 2001.
Rivière, E. D., M. Pirre, G. Berthet, J.-B. Renard, F. G. Taupin, N. Huret,
M. Chartier, B. Knudsen, and F. Lefèvre, On the interaction between
nitrogen and halogen species in the Arctic polar vortex during THESEO
and THESEO 2000, J. Geophys. Res., 107, 8311, doi:10.1029/
2002JD002087, 2002. [printed 108(D5), 2003]
Ross, M. N., et al., Observations of stratospheric ozone depletion associated
with Delta II rocket emissions, Geophys. Res. Lett., 27, 2209 – 2212,
2000.
Sander, S. P., et al., Chemical kinetics and photochemical data for use in
stratospheric modeling, Evaluation number 13, JPL Publ. 00-3, Jet Propulsion Lab., Calif. Inst. of Technol., Pasadena, Calif., 2000.
Smith, J. B., E. J. Hintsa, N. T. Allen, R. M. Stimpfle, and J. G. Anderson,
Mechanisms for midlatitude ozone loss: Heterogeneous chemistry in the
lowermost stratosphere?, J. Geophys. Res., 106, 1297 – 1309, 2001.
Solomon, S., S. Borrmann, R. R. Garcia, R. Portmann, L. Thomason, L. R.
Poole, D. Winker, and M. P. McCormick, Heterogeneous chlorine chemistry in the tropopause region, J. Geophys. Res., 102, 21,411 – 21,429,
1997.
Toohey, D. W., W. H. Brune, K. R. Chan, and J. G. Anderson, In situ
measurements of midlatitude ClO in winter, Geophys. Res. Lett., 18,
21 – 24, 1991.
Toohey, D. W., L. M. Avallone, N. T. Allen, J. N. Demusz, J. N. Hazen,
N. L. Hazen, and J. G. Anderson, The performance of a new instrument
for in situ measurements of ClO in the lower stratosphere, Geophys. Res.
Lett., 20, 1791 – 1794, 1993.
World Meteorological Organization (WMO), SPARC/IOC/GAW Assessment
of Trends in the Vertical Distribution of Ozone, SPARC Rep. 1, World
Meteorol. Org. Global Ozone Res. and Monit. Proj Rep. 43, edited by
N. Harris et al., World Meteorol. Org. Global Ozone Res. and Monit.
Proj., Geneva, 1998.
World Meteorological Organization (WMO), SPARC/IOC/GAW Assessment
of Trends in the Vertical Distribution of Ozone, SPARC Rep. 1, World
Meteorol. Org. Global Ozone Res. and Monit. Proj Rep. 44, edited by
N. Harris et al., World Meteorol. Org. Global Ozone Res. and Monit.
Proj., Geneva, 2002.
L. M. Avallone and D. W. Toohey, Program in Atmospheric and Oceanic
Sciences, University of Colorado, Campus Box 311, Boulder, CO 80309,
USA. (avallone@lasp.colorado.edu; Darin.Toohey@colorado.edu)
M. A. Avery, M/S 483, NASA Langley Research Center, Hampton, VA
23681-2199, USA. (m.a.avery@larc.nasa.gov)
W. H. Brune and J. B. Simpas, Department of Meteorology, 503 Walker
Building, Pennsylvania State University, University Park, PA 16802, USA.
(brune@essc.psu.edu; jbsimpas@essc.psu.edu)
H. Harder and M. Martinez, MPI fuer Chemie, Luftchemie, Mainz,
Germany. (harder@mpch-mainz.mpg.de; martinez@mpch-mainz.mpg.de)
B. F. Thornton, Department of Chemistry, University of Colorado,
Campus Box 215, Boulder, CO 80309, USA. (brett.thornton@colorado.
edu)