Impact of mixing and chemical change on ozone

Atmos. Chem. Phys., 5, 3139–3151, 2005
www.atmos-chem-phys.org/acp/5/3139/
SRef-ID: 1680-7324/acp/2005-5-3139
European Geosciences Union
Atmospheric
Chemistry
and Physics
Impact of mixing and chemical change on ozone-tracer relations in
the polar vortex
R. Müller1 , S. Tilmes1,* , P. Konopka1 , J.-U. Grooß1 , and H.-J. Jost2
1 ICG
I, Forschungszentrum Jülich, 52425 Jülich, Germany
Area Environmental Research Institute, Sonoma, CA, USA
* now at: ACD, NCAR, Boulder, CO, USA
2 Bay
Received: 8 June 2005 – Published in Atmos. Chem. Phys. Discuss.: 10 August 2005
Revised: 4 November 2005 – Accepted: 7 November 2005 – Published: 23 November 2005
Abstract. Tracer-tracer relations have been used for a
long time to separate physico-chemical change from change
caused by transport processes. In particular, for more than
a decade, ozone-tracer relations have been used to quantify
chemical ozone loss in the polar vortex. The application of
ozone-tracer relations for quantifying ozone loss relies on
two hypotheses: that a compact ozone-tracer relation is established in the ‘early’ polar vortex and that any change of
the ozone-tracer relation in the vortex over the course of winter is caused predominantly by chemical ozone loss. Here,
we revisit this issue by analysing various sets of measurements and the results from several models. We find that mixing across the polar vortex edge impacts ozone-tracer relations in a way that may solely lead to an ‘underestimation’
of chemical ozone loss and not to an overestimation. Further, differential descent in the vortex and internal mixing
has only a negligible impact on ozone loss estimates. Moreover, the representation of mixing in three-dimensional atmospheric models can have a substantial impact on the development of tracer relations in the model. Rather compact ozonetracer relations develop – in agreement with observations –
in the vortex of a Lagrangian model (CLaMS) where mixing is anisotropic and driven by the deformation of the flow.
We conclude that, if a reliable ‘early vortex’ reference can
be obtained and if vortex measurements are separated from
mid-latitude measurements, ozone-tracer relations constitute
a reliable tool for the quantitative determination of chemical
ozone loss in the polar vortex.
1 Introduction
In the stratosphere, compact relations are expected between
tracers for which quasi-horizontal mixing along isentropes
Correspondence to: R. Müller
([email protected])
is fast compared to their local chemical lifetimes (Plumb and
Ko, 1992). This fact has been exploited in a variety of studies
addressing a wide range of stratospheric research issues (e.g.,
Roach, 1962; Fahey et al., 1990; Volk et al., 1996; Sankey
and Shepherd, 2003, and references therein). Further, it has
been suggested that observed tracer-tracer relations can be
used as a tool for the validation of chemistry climate models
(Proffitt et al., 2003; Khosrawi et al., 2004) since the results
of such models are otherwise difficult to compare with observations.
As long as a strong polar vortex exists, compact relations
are also expected for the air mass inside the polar vortex.
In particular, as (in the absence of halogen-catalysed chemical loss) ozone in the winter polar vortex can be considered
long-lived (e.g., Proffitt et al., 1992; Sankey and Shepherd,
2003), compact ozone-tracer relations are expected for the
polar vortex (e.g., Proffitt et al., 1993; Müller et al., 2001;
Tilmes et al., 2004).
Therefore, tracer-tracer correlations (TRAC) have been
used, for more than a decade now, to quantify chemical ozone
loss in the polar vortices. Measurements from aircraft (Proffitt et al., 1990, 1993; Richard et al., 2001; Ross et al., 2004;
Ulanovskii et al., 2004), satellites (Müller et al., 1996; Tilmes
et al., 2003b, 2004), balloons (Müller et al., 2001; Salawitch
et al., 2002; Vogel et al., 2003; Robinson et al., 2005), and
from the Space Shuttle (Michelsen et al., 1998) were used
for such analyses.
The TRAC methodology has been described in detail elsewhere (e.g., Proffitt et al., 1990; Müller et al., 1996, 2001;
Salawitch et al., 2002; Tilmes, 2004; Tilmes et al., 2004).
Briefly, an ozone-tracer relation that is established in the
‘early’ polar vortex will be unaltered (‘frozen in’) throughout the existence of the polar vortex in the absence of mixing
and chemical ozone loss. In particular, diabatic descent inside the vortex alone cannot change an ozone-tracer relation.
Considering the change of ozone-tracer relations within
the vortex for quantifying chemical ozone loss in the polar
© 2005 Author(s). This work is licensed under a Creative Commons License.
outside vortex air is characterised by different ozone-tracer
relationships, with the outside vortex relationships showing
greater ozone mixing ratios (and a stronger variability) for
equal values
of theettracer
than inside relationships
R. Müller
al.: Ozone/tracer
relations in(e.g.,
the Profpolar vortex
fitt et al., 1990; Müller et al., 1999, 2002; Jost et al., 2002;
Tilmes et al., 2003b). Therefore, if mixing occurs between
rrrrrr
ozone and
mixing
ratios
(andvortex,
a stronger
for
rr
air greater
masses inside
outside
of the
it willvariability)
lead to
rrrrrr
rrr
the
same
values
of
the
tracer
than
inside
relationships
(e.g.,
r
r
r
r
r
rrrrr
points lying along the dotted mixing lines and thus above the
rr
rrrrrr
rr
Proffitt
et al., 1990;
Müller
al., 1999,
Jost et al.,
rr
rr rrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrr
original
ozone-tracer
relation
in theetvortex
(Fig.1, 2002;
top panel).
rrrrrrrrrrrrrrrrrrrrr r r r r r r rrrrr
rrr
rrrrrrrrrrrrrr
rrrrrrrrrr
2002; Tilmes
etbetween
al., 2003b).
Therefore,
if mixing
occurs beIn
contrast,
mixing
two
air
masses
that
correspond
rrrrrrrrrr rrrrrrrr
rrrrrrrrr r
rrrrrrrrrr r rrrrrr
rrrrrrrr r rrrr
tween
air
masses
inside
and
outside
the
vortex,
it
will
lead to
to
different
points
on
the
same
non-linear,
compact
ozonerrrrrrrr rrrr
rrrrrrrr rrrrr
rrrrrrr
tracer
relation,
results
in
points
that
fall
below
the
compact
points
lying
along
the
dotted
mixing
lines
and
thus
above
the
rrrrrrrr rrrrrr
rrrrrrrr rr
rrrrrrr rrrr
relation
(Fig.1,
bottom panel).
rrrrrrrrrrr
original
ozone-tracer
relation in the vortex (Fig. 1, top panel).
rrrrrrr rr
rrrrrrrrrrrr
rrrrrr
Moreover,
descent two
withair
subsequent
mixing
In contrast,differential
mixing between
masses that
correspond
rrrrrr rrrrrr
rrrrrr rr
r
within
the
vortex
impacts
tracer-tracer
relations.
Although
to different points on the same non-linear, compact ozoneRaytracer
et al. (2002)
andresults
Salawitch
et al. (2002)
conclude
relation
in points
that fall
belowthat
thethecompact
effect
of
differential
descent
increased
the
uncertainty
of
colrelation (Fig. 1, bottom panel).
umn ozone loss estimates by not more than ≈3-4% in Arctic
Moreover, differential descent with subsequent mixing
rr rrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrr
winter 1999-2000, this mechanism could potentially have an
r r rrrrrrrrrrrrrrrrrrrrr
rrrrrrrrrrrrrr
rr
within
vortex
Although
rrrrrrrrrrr
rr
influence onthe
ozone
lossimpacts
estimatestracer-tracer
deduced usingrelations.
tracer-tracer
rrrrrrrrr
rr
rrrrrrrrr
rr
Ray
et
al.
(2002)
and
Salawitch
et
al.
(2002)
conclude
that the
rrrrrrrrr
rr
rrrrrrrr
relations.
rr
rrrrr
rr
r r rrrrrrrrrrrrrrrr
effect
of
differential
descent
increases
the
uncertainty
Here, we revisit the question of the validity of the two ma- of colrr
r
r r rrrrrrrrrrrrrrr
r r rrrrrrr
ozone that
lossthe
estimates
by no more than
≈3–4%
in Arctic
r r rrrrrrrr
jor umn
hypotheses
TRAC methodology
is based
on. We
r r rrrrrr
r r rrrrrrr
r rrrrrrrr
winter
1999–2000,
this
mechanism
could
potentially
have an
consider
the
question
of
the
compactness
of
the
ozone-tracer
rrrrrr
rrrrrr
influence
ozonepolar
loss vortex.
estimates
deduced
tracer-tracer
relation
in theon‘early’
Further,
we using
discuss
the
impact
of mixing across polar vortex edge on ozone loss esrelations.
timates
deduced
using tracer-tracer
and extend
thetwo maHere,
we revisit
the questionrelations
of the validity
of the
discussion
to the related
issue of
descent within
jor hypotheses
on which
thedifferential
TRAC methodology
is based.
the We
vortex.
Finally,
wequestion
examineof
thethe
representation
of ozoneconsider
the
compactness
of
the
ozoneTracer
tracer relations in a variety of models. We conclude that, if
tracer relation in the ‘early’ polar vortex. Further, we discuss
applied with care, ozone-tracer relations constitute a reliable
impact of mixing across the polar vortex edge on ozone
Fig. 1. Schematic view of the impact of mixing on high-latitude toolthe
for the quantitative determination of chemical ozone loss
loss
estimates
deduced using tracer-tracer relations and exFig.
1.
Schematic
view
of
the
impact
of
mixing
on
high-latitude
ozone-tracer relationships. The solid line indicates a vortex ozone- in the polar
vortex.
ozone-tracer
Theasolid
line indicates
a vortex ozonetend the discussion to the related issue of differential descent
tracer
relation,relationships.
the dashed line
mid-latitude
ozone-tracer
relation,
relation,
dashed
line amixing
mid-latitude
thetracer
dotted
lines the
show
possible
lines.ozone-tracer
Top panel relation,
shows the
within the vortex. Finally, we examine the representation of
the
dotted
lines
show
possible
mixing
lines.
Top
shows the
conditions for mixing across the vortex edge onpanel
isentropes.
Note 2 ozone-tracer
relations inrelations
a varietyinofthe
models.
conclude
Reference ozone-tracer
‘early’ We
polar
conditions for mixing across the vortex edge on isentropes. Note
that there is a substantial difference in tracer mixing ratio between
that,
if
applied
with
care,
ozone-tracer
relations
constitute
a
vortex
that there is a substantial difference in tracer mixing ratio between
vortex and mid-latitude air masses on the same potential temperareliable tool for the quantitative determination of chemical
vortex and mid-latitude air masses on the same potential temperature
in the
thepolar
polarvortex.
vortex.Bottom
Bottom
panel 2.1ozone
Reference
constructed
turelevel
levelcaused
causedby
by the
the descent
descent in
panel
loss inrelations
the polar
vortex. from mixing lines
shows
betweentwo
twoairairparcels
parcels
characshowsthe
theconditions
conditions for
for mixing
mixing between
characterised
relation.
terisedbybythe
thesame
sameozone-tracer
ozone-tracer relation.
Michelsen et al. (1998) analysed ATMOS observations of
ozone
N2 O inozone-tracer
the Arctic in relations
March/April
1993.
By polar
2 and
Reference
in the
‘early’
considering the O3 /N2 O relationship they find that, on the
vortex
Thehinges
validityonof two
thesehypotheses:
hypotheses has
beenthat
debated
and
same N2 O level, much lower (≈50-60%) ozone mixing ravortex
First,
a compact
the questionrelation
was raised
whether mixing
thepolar
vortex
are observed inside compared to outside of the polar vorozone-tracer
is established
in theacross
‘early’
vor- tios2.1
Reference
constructed
from mixing
lines
edge
changes
the
ozone-tracer
relation
in
a
way
that
leads
tex.
Michelsen
et al. relations
(1998) interpret
this difference
between
tex, i.e., before the onset of chemical ozone loss. Second, that
to an overestimate or an underestimate of ozone loss (e.g.,
the out-of-vortex reference (blue line in Fig. 2) and the meaany change of the ozone-tracer relation in the vortex over the
Michelsen
al.vortex
(1998)
observations
of
Michelsen et al., 1998; Plumb et al., 2000; Müller et al.,
surements
insideetthe
as aanalysed
signatureATMOS
of chemical
ozone
course
of winter is caused predominantly by chemical ozone loss,
ozone
and N2 Othat
in chemical
the Arctic
in March/April
By
2001; Salawitch et al., 2002; Rex et al., 2002; Sankey and
but emphasise
effects
are unlikely to1993.
acloss.
considering
the difference.
O3 /N2 O relationship
they findanthat,
Shepherd, 2003; Tilmes et al., 2003b, 2004). Indeed, the imcount
for the entire
They also construct
al- on the
The
validity
of
these
hypotheses
has
been
debated
and
same
N
O
level,
much
lower
(≈50–60%)
ozone
mixing
rapact of mixing on tracer-tracer relationships was neglected
ternative reference
for the O3 /N2 O relation (the yellow line
2
early studies
thewhether
TRAC methodology
(e.g.,the
Müller
plateobserved
4a, also shown
Fig. 2). This
reference
thein question
wasusing
raised
mixing across
vortex in their
tios are
inside incompared
to outside
therelapolar voret al.,
1996, 1997).
However, over
the course
winter
is constructed
mixing
line
for a single
eventbetween
edge
changes
the ozone-tracer
relation
in a of
way
that and
leads tiontex.
Michelsenasetaal.
(1998)
interpret
this mixing
difference
the vortex air
will be influenced
to a certain
bottom panel)reference
and is derived
tospring,
an overestimate
ormass
an underestimate
of ozone
loss ex(e.g., (Fig.
the1,out-of-vortex
(blueconsidering
line in Fig.mixing
2) and in
the mea-
Ozone
Ozone
tex, i.e., before the onset of chemical ozone loss. Second, that
any change of the ozone-tracer relation in the vortex over the
course of winter is caused predominantly by chemical ozone
loss.
3140
Michelsen et al., 1998; Plumb et al., 2000; Müller et al.,
2001;
Salawitch
et al.,0000,
2002;
Rex et2005
al., 2002; Sankey and
Atmos.
Chem. Phys.,
0001–15,
Shepherd, 2003; Tilmes et al., 2003b, 2004). Indeed, the impact of mixing on tracer-tracer relationships was neglected
in early studies using the TRAC methodology (e.g., Müller
et al., 1996, 1997). However, over the course of winter and
spring, the vortex air mass will be influenced to a certain extent by mixing in of air from outside the vortex. Inside and
outside vortex air is characterised by different ozone-tracer
relationships, with the outside vortex relationships showing
Atmos. Chem. Phys., 5, 3139–3151, 2005
surements inside the vortex as a signature of chemical ozone
loss, but emphasise
that chemical effects are unlikely to acwww.atmos-chem-phys.org/acp/0000/0001/
count for the entire difference. They also construct an alternative reference for the O3 /N2 O relation (the yellow line
in their plate 4a, also shown in Fig. 2). This reference relation is constructed as a mixing line for a single mixing event
(Fig. 1, bottom panel) and is derived by considering mixing
in the N2 O/CH4 tracer space. However, it is unlikely that
mixing of air between the two end-points of a mixing line
that are separated by ≈25 km in altitude for an out-of-vortex
www.atmos-chem-phys.org/acp/5/3139/
R. Müller et al.: Ozone/tracer relations in the polar vortex
3141
R. Müller: Ozone/tracer relations in the polar vortex
3
profile (Michelsen et al., 1998, Plate 3) occurred in reality.
Thus, this line should only be seen as an approximation to a
reference relation that would develop by continuous mixing.
Comparing this reference with the March/April 1993 vortex observations from ATMOS, Michelsen et al. (1998) estimate only half of the ozone deficit that they deduce from
an out-of-vortex reference. This finding is sometimes erroneously generalised to cases (e.g., Müller et al., 1996; Salawitch et al., 2002; Tilmes et al., 2004) where ozone-tracer relationships in the late vortex are compared with an ‘early vortex’ reference (that is, a reference derived from observations
made inside the vortex at a time before chemical ozone loss
occurs) to deduce chemical ozone loss. Out-of-vortex ozonetracer relations are characterised by greater ozone mixing
ratios for a given tracer value than vortex relations (Proffitt
2. Reference relations for the ozone-methane relation in the early vortex in winter 1992/93 from HALOE (Russell et al., 1993) and
et al., 1993; Müller et al., 2001; Tilmes et al., 2004). There-Fig.
ATMOS (Michelsen
et al.,Reference
1998). The black line
shows the reference
relationozone-methane
deduced from HALOE measurements
Fig. 2.
relations
for the
relationin November
in the1992 in the
early
vortex, the dotted lines give the range of uncertainty (Tilmes et al., 2004). The yellow line shows the ‘mixing line’ reference estimated
fore, it is incorrect to interpret the finding of Michelsen et al.by Michelsen
◦ N-69◦ N), reference
early
vortex
inATMOS
winter
1992/93
from
HALOE
(Russell
al.,(631993)
et al. (1998)
based on
measurements,
the solid
blue line
shows the out-of-vortex,
highet
latitude
measurements from 8.-16. 4. 1993 (Michelsen et al., 1998). The ATMOS relations were converted from a N2 O/O3 relation
(1998) as ‘mixing could produce about half the changes inbased on ATMOS
ATMOS (Michelsen et al., 1998). The black line shows the refto a CH4 /Oand
3 relation using the reported ATMOS CH4 /N2 O relation (Michelsen et al., 1998). Further shown are HALOE measurements
inside the vortex
in late relation
March (purplededuced
diamonds), early
April (orange
diamonds),
and late April 1993 (red
The blue diamonds
the O3 /N2 O relation’ (Plumb et al., 2000).
erence
from
HALOE
measurements
indiamonds).
November
indicate HALOE measurements in April 1993 in high latitudes outside the vortex.
In Fig. 2, we compare the out-of-vortex measurements
1992 in the early vortex, the dotted lines give the range of uncertainty (Tilmes et al., 2004). The yellow line shows the ‘mixing line’
from ATMOS in April 1993 (blue line), the Michelsen et al.
the N2 O/CH4 tracer space. However, it is unlikely that mix1993; Müller et al., 2001; Tilmes et al., 2004). Therereference
estimated by Michelsen etfore,
etal.,al.
(1998) based on ATMOS
(1998) ‘mixing reference’ (the yellow line in their plateing of air between
the two end-points of a mixing line that are
it is incorrect to interpret the finding of Michelsen et al.
separated measurements,
by ≈25 km in altitude forthe
an out-of-vortex
profile
(1998)
as “mixing
could produce about half
the changes in
solid
blue
line
shows
the out-of-vortex,
high4a shown here again as a yellow line) and the early vor-(Michelsen et al., 1998, Plate◦ 3) occurred
in reality. Thus,
the O3 /N2 O relation” (Plumb et al., 2000).
◦ N),
latitude
(63
N–69
reference
based
on
ATMOS
measurements
this
line
should
only
be
seen
as
an
approximation
to
a
refertex reference derived from HALOE early vortex measure-ence relation that would develop by continuous mixing.
Inal.,
Fig.1998).
2, we compare
out-of-vortex
measurements
from 8–16 April 1993 (Michelsen et
ThetheATMOS
relafrom ATMOS in April 1993 (blue line), the Michelsen et al.
ments in 1992 (Tilmes et al., 2004, black line). Obviously,
tions
were converted
from1993
a Nvorto a CH
Comparing
this reference
with the March/April
‘mixing reference’
(the
yellow
line in their plate 4a
2 O/O(1998)
3 relation
4 /O
3 relation
the latter reference and the best reference (yellow line) sug-tex observations from ATMOS, Michelsen et al. (1998) deshown here again as a yellow line) and the early vortex reftheozone
reported
CH4 /N
relation
(Michelsen
etmeasurements
al.,
2 O derived
duce onlyusing
half of the
deficit thatATMOS
they deduce from
erence
from HALOE
early vortex
in
gested by Michelsen et al. (1998) are in reasonable agree-an out-of-vortex
reference.
This finding
is sometimes
erro1992
(Tilmes et al., 2004,
black line).
1998).
Further
shown
are HALOE
measurements
inside
the Obviously,
vortex the latter
neously
generalised
to
cases
(e.g.,
Müller
et
al.,
1996;
Salawreference
and
the
best
reference
(yellow
line)
suggested
by
ment. Certainly, they are closer to each other than any ofitch et al.,in
late
March
(purple
diamonds),
early April
(orange
diamonds),
2002;
Tilmes
et al., 2004)
where ozone-tracer
reMichelsen
et al. (1998)
are in reasonable
agreement. Cerin the late
late vortex
are compared
vortainly,The
they are
closerdiamonds
to each other than
any of them to the
the lines to the out-of-vortex ATMOS measurements. There-lationshipsand
April
1993 with
(redan early
diamonds).
blue
indicate
tex reference (that is, a reference derived from observations
out-of-vortex ATMOS measurements. Therefore, ozone loss
fore, ozone loss estimates deduced from these two referencesmade insideHALOE
the vortex, atmeasurements
a time before chemicalin
ozone
loss 1993
estimates
from these two
referencesthe
will be rather
April
in deduced
high latitudes
outside
to deduce chemical ozone loss. Out-of-vortex ozonesimilar. Using the early vortex reference from Tilmes et al.
will be rather similar. Using the early vortex reference fromoccurs)
vortex.
tracer relations are characterised by greater ozone mixing
(2004) yields an ozone loss of 117 ± 17 DU for April 1993
Tilmes et al. (2004) yields an ozone loss of 117±17 DU forratios for a given tracer value than vortex relations (Proffitt in the altitude range 380-550 K (corresponding to 1.4 ppm <
April 1993 in the altitude range 380–550 K (correspondingwww.atmos-chem-phys.org/acp/0000/0001/
Atmos. Chem. Phys., 0000, 0001–15, 2005
tion of different compact tracer-tracer relations in the stratoto 1.4 ppm<CH4 <0.5 ppm), whereas employing the ‘mixing
sphere (e.g., Michelsen et al., 1998; Plumb, 2002; Proffitt
line reference’ from Michelsen et al. (1998) yields an ozone
et
al., 2003).
loss of 98 DU, that is, only 16% less ozone loss than obNevertheless, even state-of-the-art three-dimensional
tained from the Tilmes et al. (2004) reference. Nonetheless,
models will misrepresent transport, and thus tracer-tracer relong-range (in tracer space) mixing lines can only yield an
lations, in some way. Depending on the numerical advection
approximation to the atmospheric conditions. Therefore, if
scheme employed, transport errors are introduced through
used as reference relations for ozone loss estimates, mixing
numerical diffusion and dispersion (e.g., Rood, 1987; Müller,
line references will be associated with a greater uncertainty
1992). Dispersion errors that arise particularly strongly in
than an early vortex reference deduced from measurements.
centred difference and in spectral schemes, may lead to spurious scatter in simulations of nonlinear tracer-tracer relation2.2 Representation of ozone-tracer reference relations in
ships (Thuburn and McIntyre, 1997).
models
While classical two-dimensional models have been used in
the past to study tracer-tracer relations (Plumb and Ko, 1992;
Proffitt et al., 1992), today three-dimensional models are
available for the simulation of transport and chemistry in the
stratosphere. A particular advantage of three-dimensional
models is that transport barriers in the stratosphere can be
better represented than in two-dimensional models. The most
important stratospheric transport barriers, the polar vortex
edge and the subtropical barrier, are essential for the formawww.atmos-chem-phys.org/acp/5/3139/
2.2.1
Three-dimensional Chemistry Climate Models
Recently, Sankey and Shepherd (2003) presented a thorough
investigation of tracer relations in the stratosphere using the
Canadian Middle Atmosphere Model (CMAM); they used
a number of observed tracer-tracer relations to validate the
model results. The model uses a (T32) spectral transport
scheme with 50 layers in the vertical and a ≈3 km vertical
resolution in the middle atmosphere. The vertical diffusion
Atmos. Chem. Phys., 5, 3139–3151, 2005
3142
R. Müller: Ozone/tracer relations in the polar vortex
R. Müller et al.: Ozone/tracer relations in the polar vortex
5
ozone loss in the Arctic using CMAM results (Sankey and
Shepherd, 2003). In the CMAM Antarctic vortex, for which
a transport barrier is simulated, a compact ozone-tracer relation develops in the model. In CMAM it takes months for the
compact correlation to develop; a truly compact correlation is
established in August. A recent analysis of ILAS II measurements in Antarctic winter 2003 indicates that a compact relation had developed by the end of June (Tilmes et al., 2005b)
(Tilmes et al., 2005a1 ). The fact that the simulated ozonetracer relation in the Antarctic vortex in CMAM does not
resemble measurements by HALOE, ATMOS, and ILAS II
(Müller et al., 1996; Michelsen et al., 1998; Tilmes et al.,
2005b) is most likely due to an underestimate of halogencatalysed ozone depletion in the CMAM Antarctic vortex.
Sankey and Shepherd (2003) raised the question whether,
given limited observations at a particular latitude, one could
Fig. 3. The ozone-methane relation in the Northern hemisphere
be under the false impression that a compact ozone-tracer rein January 1992. Shown are HALOE observations for the altitude
lation exists, when in reality this is not the case. To illustrate
range 10–100 hPa for the period 10 to 14 January 1992. The obtheir point, they consider (their Fig. 19) CMAM data for the
servations were made at equivalent latitudes of 25 –40 N (purple
O3 /CH4 relation inside the polar vortex for 1 January. (The
symbols), 40◦ –50◦ N (green symbols), 50◦ –60◦ N (blue symbols),
◦ –65
◦N
we consider
actual
HALOE
mid Janthe early springare
vortex
in years when HALOE
a
vortex
being defined as a column of air with potential vortic60the
(redmeasurements
symbols).in The
blackofdiamonds
observations
in- provides
uary 1992 (Fig. 3) colour-coded according to equivalent latigood coverage of the polar area (e.g., early April 1992, 1995,
side
the
polar
vortex
according
to
the
potential
vorticity
criterion
of
ity greater than 400 PVU at 10 hPa, D. Sankey, pers. comm.).
tude. Based on the selection criteria of Tilmes et al. (2004),
1996) show compact relations of both HCl and ozone with
three HALOE
measurements
(black
symbols
in Fig.
3) are area
long-lived
tracers
et al.,of
1996;
Tilmes et al., 2004).
Tilmes
(2004).
The
grey
shaded
shows
the(Müller
range
CMAM
The CMAM model vortex is then sampled at the actual lolocated in the polar vortex in January 1992. These profiles
model
results
for theclearly
vortex
in a simulated January (Sankey and
show a rather
compact
O3 /CH4 relation,
more comcations of HALOE measurements in January 1992. Due
pact than the CMAM relation at the location of the HALOE
Shepherd, 2003). (The vortex is defined for the model results as
to the HALOE measurement geometry, these measurement
measurements. Solely, the HALOE measurements at top altitudes show
characteristics
of outsidegreater
vortex air.than
The obpotential
vorticity
400 PVU at 10 hPa).
locations rarely intercept the CMAM vortex. The CMAM
served O3 /CH4 relation for outside vortex, mid-latitude air
shows considerable scatter, but always greater ozone values
O3 /CH4 values at the HALOE measurement locations form
for a given tracer mixing ratio than the vortex air.
a compact relation for CH4 lower than ≈1.3 ppm; for CH4
Further,coefficient
measurements of K
the ILAS
set to 1(Sasano
m2 s−1 . In the model simulations,
zz is instrument
greater
than ≈1.3 ppm, the CMAM O3 values at the HALOE
et al., 1999) which provides a good coverage of the polar area
In summary, the lack of a compact O3 /CH4 relationship
in
distinct
tracer-tracer
relations
in1 the
mid-to the concluthe CMAM
Arctictropics,
vortex couldthe
be leading
in winter 1996-97
demonstrate
that a compact
ozone-tracer develop
measurement
locations show some scatter, albeit less scatter
sion that apparently compact correlations in the vortex may
relation develops in the Arctic vortex which is clearly sepasurfrelation
zone,
and
in2003b).
the Antarctic
wintersampling
vortex
ac-air-mass when
result from insufficient
of thein
vortex
rated fromlatitude
the out-of-vortex
(Tilmes
et al.,
than
all
the
CMAM
vortex model points (Sankey and Shepreality no compact
exits (Sankey
The ILAScordance
Arctic vortex measurements
show no indication
with observations
(e.g.,in2003).
Michelsen
etrelation
al., 1998;
Rayand Shepherd,
of a variability of the ozone-tracer relation resembling that
However, there is no evidence from measurements
in
herd,
2003). That is, if the real atmosphere were as simulated
polarwith
regions the
that such
a problem occursarin reality.
in the CMAM
model2002;
vortex. Also,
HALOEet
measurements
et al.,
Proffitt
al., 2003)theand
theoretical
by CMAM, and had HALOE observed that model atmoguments of Plumb (2002).
sphere during January 1992 the apparent correlation would
www.atmos-chem-phys.org/acp/0000/0001/
Atmos. Chem. Phys., 0000, 0001–15, 2005
However, the model is incapable of producing a strong pohave been the result of insufficient sampling. This thought
lar vortex in the Arctic: ‘there is virtually no barrier at all
experiment indicates the possible existence of a problem
in the CMAM Arctic vortex’ (Sankey and Shepherd, 2003).
(Sankey and Shepherd, 2003).
A transport barrier at the Arctic vortex edge (together with
To address the question whether such a problem might exrapid mixing within the vortex) is the reason why compact
ist
for conditions in the early polar vortex in the Arctic, we
tracer-tracer relations develop inside the polar vortex that
consider
the actual HALOE measurements in mid-January
are distinct from the tracer-tracer relations outside the vor1992
(Fig.
3) colour-coded according to equivalent latitude.
tex (e.g., Plumb, 2002). The segregation between compact
Based
on
the
selection criteria of Tilmes et al. (2004), three
tracer-tracer relationships inside and outside the vortex has
HALOE
measurements
(black symbols in Fig. 3) are lobeen observed on many occasions (e.g., Proffitt et al., 1993;
cated
in
the
polar
vortex
in January 1992. These profiles
Michelsen et al., 1998; Müller et al., 1999; Ray et al., 2002;
show
a
rather
compact
O
/CH
3
4 relation, clearly more comTilmes et al., 2003b).
pact than the CMAM relation at the location of the HALOE
The formation of such a segregation of tracer-tracer relameasurements. Solely the HALOE measurements at top altionships and the existence of a transport barrier at the edge
titudes show characteristics of outside-vortex air. The obof the Arctic vortex are the central assumptions for applying
served O3 /CH4 relation for outside-vortex, mid-latitude air
tracer-tracer relations to deduce chemical ozone loss inside
the vortex (e.g., Müller et al., 1996; Tilmes, 2004). In the
1 Tilmes, S., Müller, R., Grooß, J.-U., Nakajima, H., and Sasano,
CMAM model, for Arctic conditions, these assumptions are
not valid due to the the lack of a transport barrier at the edge
Y.: Development of tracer relations and chemical ozone loss during
of the Arctic vortex in CMAM. Consequently, by means of
the setup phase of the polar vortex, J. Geophys. Res., submitted,
tracer-tracer relations it is not possible to deduce chemical
2005a.
Fig. 3. The ozone-methane relation in the Northern hemisphere in January 1992. Shown are HALOE observations for the altitude range
10-100 hPa for the period 10 to 14 January 1992. The observations were made at equivalent latitudes of 25◦ -40◦ N (purple symbols), 40◦ 50◦ N (green symbols), 50◦ -60◦ N (blue symbols), 60◦ -65◦ N (red symbols). The black diamonds are observations inside the polar vortex
according to the potential vorticity criterion of Tilmes (2004). The grey shaded area shows the range of CMAM model results for the vortex
in a simulated January (Sankey and Shepherd, 2003). (The vortex is defined for the model results as potential vorticity greater than 400 PVU
at 10 hPa).
◦
◦
Atmos. Chem. Phys., 5, 3139–3151, 2005
www.atmos-chem-phys.org/acp/5/3139/
shows considerable scatter, but always greater ozone values
for a given tracer mixing ratio than the vortex air.
Further, measurements of the ILAS instrument (Sasano
et al., 1999), which provides good coverage of the polar area
in winter 1996–1997, demonstrate that a compact ozonetracer relation develops in the Arctic vortex which is clearly
separated from the out-of-vortex relation (Tilmes et al.,
2003b). The ILAS Arctic vortex measurements show no indication of a variability of the ozone-tracer relation resembling that in the CMAM model vortex. Also, HALOE measurements of the early spring vortex in years when HALOE
provides good coverage of the polar area (e.g., early April
1992, 1995, 1996) show compact relations of both HCl and
ozone with long-lived tracers (Müller et al., 1996; Tilmes
et al., 2004).
In summary, the lack of a compact O3 /CH4 relationship in
the CMAM2 Arctic vortex could lead to the conclusion that
compact correlations in the vortex may be deluded by insufficient sampling of the vortex air mass when in reality no such
compact relation exits (Sankey and Shepherd, 2003). However, there is no evidence from measurements in the polar
regions that such a problem occurs in reality.
6
R. Müller: Ozone/tracer relations in the polar vortex
3143
PDF (%)
700 K
8
3.00
650 K
1.00
0.70
6
O3 (ppmv)
R. Müller et al.: Ozone/tracer relations in the polar vortex
0.50
600 K
0.30
4
0.10
550 K
0.07
2
450 K
11.01.2003
500 K
0.05
0.03
0
0.01
0.4
0.6
0.8
pass
1.0
1.2
CH4 (ppmv)
1.4
1.6
1.8
pass
Fig. 4. The relation of passive ozone (O3 ) and CH4 on 11 January 2003 from model simulations (Konopka et al., 2005a) of the Chemical
Lagrangian Model of the Stratosphere (CLaMS). Shown are all CLaMS model points poleward of an equivalent latitude of 40◦ N between
pass
the 350 K and 700 K isentropes. The colour scale shows the probability distribution function (PDF) of O3 and CH4 ; reference relations for
vortex air and midlatitude air are shown as thick grey lines. For comparison, an early vortex reference relation (Tilmes et al., 2003a) deduced
from MkIV measurements on 16 December 2002 is also shown (black dashed-dotted line). Also shown are approximate isolines of selected
pass
potential temperature levels. PDFs and potential temperature isolines were calculated for O3 and CH4 bin sizes of 0.05 ppmv and 0.25
ppmv, respectively.
Fig. 4. The relation of passive ozone (O3 ) and CH4 on 11 January 2003 from model simulations (Konopka et al., 2005c3 ) of the
Chemical Lagrangian Model of the Stratosphere (CLaMS). Shown
are all CLaMS model points poleward of an equivalent latitude
2.2.2 of
Three-dimensional
chemistry transport
transport
barriers can be represented
well
40◦ N between
the models
350 K andtures,
700
K isentropes.
Theparticularly
colour
in CLaMS.
pass
shows
theModel
probability
distribution
function
(PDF)
of Oet 3al. (2005a)
andcalIn thescale
Chemical
Lagrangian
of the Stratosphere
For Arctic
winter 2002/2003,
Konopka
(CLaMS), advection and mixing are represented in a way that
culated the development over the course of the winter of pasforinvortex
air and pass
midlatitude air are shown
is veryCH
different
from the transportrelations
schemes employed
Eu4 ; reference
sive ozone (O3 ) and CH4 , both treated as pure transport
lerian models. The mixing intensity in CLaMS can be conquantities.
The probability
distribution
function (PDF)
as
thick
grey
lines.
For
comparison,
an early
vortex
reference
rela-for
pass
trolled and is anisotropic in space and time (McKenna et al.,
O3 and CH4 on 11 January 2003 shows that a clear sepa2002b;
Konopka
et al., 2004,et2005c).
to these
fearationfrom
between MkIV
mid-latitudemeasurements
and vortex air masses develops
tion
(Tilmes
al., Owing
2003a)
deduced
on
in the model (Fig. 4); shown are all CLaMS model points
2.2.2 Three-dimensional chemistry transport models
◦ N between
Here
shortcomings of 2002
the CMAMis
model
are being
dis16several
December
also
shown
(blackof andashed-dotted
Also
poleward
equivalent latitude of 40line).
the 350
cussed that become apparent, when the CMAM model results are
K and 700 K isentropes.
shown
aremeasurements.
approximate
of
selected potential temperature
compared
in detail with
The aim of thisisolines
discussion
These CLaMS results suggest that in early January a transis not to point to any particular weakness of the CMAM model;
In the Chemical Lagrangian Model of the Stratosphere rather levels.
port barrier
has formed in
the model
at the edge of the
PDFs(2003)
and
potential
temperature
isolines
were
calculated
forpoSankey and Shepherd
should
be commended for
conlar vortex leading to a separation between inside and outside
such
a rather detailed comparison of results from a general
pass
(CLaMS), advection and mixing are represented in a way ducting
O3model with
andmeasurements.
CH4 bin sizes of 0.05 ppmv
and 0.25
circulation
tracer relations.
Further, ppmv,
by that time,respectively.
mixing within the vor1
that is very different from the transport schemes employed
in Eulerian models. The mixing intensity in CLaMS can be
controlled and is anisotropic in space and time (McKenna
et al., 2002b; Konopka et al., 2004, 2005b). Due to these features, transport barriers can be represented particularly well
in CLaMS.
For Arctic winter 2002/2003, Konopka et al. (2005c)3 calculated the development over the course of the winter of paspass
sive ozone (O3 ) and CH4 , both treated as purely transport
quantities. The probability distribution function (PDF) for
pass
O3 and CH4 on 11 January 2003 shows that a clear separation between mid-latitude and vortex air masses develops
in the model (Fig. 4). The figure shows all CLaMS model
points poleward of an equivalent latitude of 40◦ N between
the 350 K and 700 K isentropes.
2 Here shortcomings of the CMAM model are discussed that be-
come apparent when the CMAM model results are compared in detail with measurements. The aim of this discussion is not to point
to any particular weakness of the CMAM model; rather Sankey and
Shepherd (2003) should be commended for conducting such a detailed comparison of results from a general circulation model with
measurements.
3 Konopka, P., Engel, A., Funke, B., Müller, R., Grooß, J.-U.,
Günther, G., Wetter, T., Stiller, G., von Clarmann, T., Oelhaf, N.
G. H., Wetzel, G., López-Puertas, M., Pirre, M., Huret, N., and
Riese, M.: Ozone loss driven by nitrogen oxides and triggered by
stratospheric warmings may outweigh the effect of halogens, Geophys. Res. Lett., submitted, 2005c.
www.atmos-chem-phys.org/acp/5/3139/
Atmos. Chem. Phys., 0000, 0001–15, 2005
www.atmos-chem-phys.org/acp/0000/0001/
These CLaMS results suggest that in early January a transport barrier has formed in the model at the edge of the polar vortex leading to a separation between inside and outside
tracer relations. Further, by that time, mixing within the vortex has homogenised the vortex air mass resulting in a compact ozone-tracer relation in the vortex.
The ozone-tracer relation in Eulerian chemical transport
models has not been studied extensively. However, its has
been shown that compact tracer-tracer relations can be simulated for the polar vortex by Eulerian chemical transport
models (Plumb et al., 2000). Recently (Robinson et al., 2005)
reported simulations with the SLIMCAT model, showing
compact ozone/CFC-11 relations in agreement with balloonborne measurements.
3
Impact of mixing on ozone-tracer relationships in the
polar vortex
3.1 Impact of differential descent on tracer relations
Differential descent of air masses within an otherwise isolated polar vortex may impact non-linear tracer-tracer relations for the vortex (Ray et al., 2002; Salawitch et al., 2002).
Potentially, this mechanism could have an influence on ozone
loss estimates deduced using tracer-tracer relations.
Atmos. Chem. Phys., 5, 3139–3151, 2005
Differential descent of air-masses within an otherwise isolated polar vortex may impact non-linear tracer-tracer relations for0the vortex (Ray et al., 2002; Salawitch et al., 2002).
0 this mechanism
100
300 on ozone
400
Potentially,
could200
have an influence
3144
N
O
(ppb)
8
2
loss estimates deduced using tracer-tracer relations.
Time Loop Plot
4.5
4.0
O3 (ppm)
Ozone (ppm)
3.5
3.0
2.5
2.0
1.5
1.0
0
Ozone-tracer relations were used for calculating a proxy
ozone profile from a tracer profile in late winter, early spring.
If the R.
twoMüller
relations
shown
in Fig. 5 are
used toindeduce
proxy
et al.:
Ozone/tracer
relations
the polar
vortex
R. Müller: Ozone/tracer
relations
in the polar vortex
ozone profiles from a late winter vortex tracer profile characterised by vertical descent, the column ozone corresponding
3.5
to these profiles differs by only ≈ 2.5 DU. In a scenario that
is perhaps a bit more realistic, where ∆θ = 2 K, α = 0.33,
δ = 1.53.0
and the number of timesteps n = 50, the corresponding difference in column ozone is less than 1 DU; the maximum change
of ozone for a given tracer value due to differ2.5
ential descent and subsequent mixing under these conditions
is ≈ 50 ppb.
50
100
150
N2O (ppb)
200
2.0
This means that the contribution of the effect of differential descent to the overall uncertainty in the deduced chemical
1.5be small. Given the magnitude of other sources of
loss will
uncertainty, the contribution from differential descent is very
likely to
1.0be insignificant. This finding corroborates conclu50 et al. (2002)
100 and Salawitch
150
sions of Ray
et al.200
(2002) who250
N2O (ppb)
250 report, based on balloon measurements
in winter 1999-2000,
that the effect of differential descent increased the uncerFig. 6. The O3 /N2 O relation from measurements during winter 1999/2000. The large black squares show measurements of the MkIV
Fig. 6. on The
O3 /N
relation
from measurements
during winter
2O
instrument (Toon
et al., 1999)
3 December
1999.
The
coloured
symbols showby
measurements
of the NASA-ER2 aircraft on 23 January 2000
tainty
of column
ozone
loss
estimates
≈3-4%.
1999/2000.
The2000
large
squares were
show
theand McLaughlin,
(green),
(blue), and 11 March
(red). black
Ozone measurements
mademeasurements
with a UV-photometer of
(Proffitt
Fig.
5. 5.
The
impact
of of
differential
descent
on on
thethe
O3O
/N32/N
O2relation
in in7 March 2000
Fig.
The
impact
differential
descent
O relation
1983), the N2 O measurements with the ALIAS (Webster et al., 1994, 23. 1.) and ARGUS (Loewenstein et al., 2002, 3. and 11. 3.) TDL
MkIVforinstrument
(Toon et al., 1999) on 3 December 1999. The
thethe
vortex
from
anan
idealised
calculation.
TheThe
black
solid
lineline
shows
vortex
from
idealised
calculation.
black
solid
shows Overplotted
instruments.
comparison is the O3 /N2 O relation (small black symbols) based on N2 O measurements by an in-situ gas
chromatograph (Elkins
et al., 1996)
with a lessershow
temporalmeasurements
resolution.
coloured
symbols
of the NASA-ER2 aircraft
thethe
initial
O3O/N
relation,
thethe
redred
lineline
thethe
O3O
/N32/N
O2relation
after
initial
relation,
O relation
after
2 O2 O
3 /N
3.2
Impact
of
cross
vortex
edge
mixing
on (blue),
ozone-tracer
on
23
January
2000
(green),
7
March
2000
and 11reMarch
2020
iterations
of
differential
descent.
It
is
assumed
that
one
third
of
iterations of differential descent. It is assumed that one third
of mixing
thepolar
vortex
edge
tracer made
relation inside
vortex.
According to the arguments
lations
inOzone
the
vortex were
2000
(red).across
measurements
with the
a UV
photometer
theofvortex
air mass
descending
as as
thethe
restrest
of 3.2.1
the
the vortex
air is
mass
descendstwice
twiceasasmuch
much
of theObservations
above, one expects ozone to increase for a given tracer mixand
McLaughlin,
1983), the
N2 OThis
measurements
theby the ILAS invortex
airair
mass,
followed
byby
instantaneous
horizontal
mixing
(see
Rex(see
et al. (2002) (Proffitt
and Jost et al.
(2002),
based on high altitude
vortex
mass,
followed
instantaneous
horizontal
mixing
ing ratio.
effect was indeed with
observed
aircraft
in the(Webster
Arctic in early
2000,1994,
demon-23 January)
strument in winter
1996-1997 (Tilmes
et al., 2003b). In this
ALIAS
et al.,
and ARGUS
(Loewentext
forfor
details).
The
green
dotted
lines
show
thethe
gradual
change
of ofmeasurements
text
details).
The
green
dotted
lines
show
gradual
change
strated that mixing across the vortex edge leads to an increase
vortex was perturbed by a stratospheric warming
stein
et
al.,
3 and
March)winter,
TDLtheinstruments.
thethe
O3O/N
relation
with
each
timestep.
Mixing
across
the vortex
edge
non-local
in Overplotted
tracer2 O2 O
relation
with
each
timestep.
in the mixing
ratio
of ozone
for a2002,
given
value
of N11
2 O. Fur-is usually
in early winter
and therefore
the
vortex edge was permeable
3 /N
ther, Rex et tracer
al. (2002)
and Rayand
et al.thus
(2002)
emphasise
thatrelation
for comparison
is the
O3the
/N2potential
O
(small
black
symbols)
untilto
late
December.
ILAS
providedreabased
good coverage of the
space
has
alter
tracer-tracer
although in this winter mixing lines were observed, entrainArctic gas
polar chromatograph
region in winter 1996-1997
covering a latitude
on air
N2The
O measurements
by
an in-situ
(Elkins
◦
◦
Here, we conduct a rather simple calculation to assess the
lations.
relation
shows
greater
ozone
mixment of extra-vortex
did notozone-tracer
significantly alter the
comrange between 55 -77 N and measuring about fourteen proet
al.,
lower
temporal
resolution.
position of the vortex
air 1996)
between with
Januarya and
March
2000.
per day.
These measurements
impact of differential descent on ozone-tracer relationships
ing ratios outside of the vortex thanfiles
inside
(e.g.,
Proffitt et show
al., (see Fig. 7 in Tilmes
Here, we conduct a rather simple calculation to assess
the
Figure 6 shows an example of measurements of such mixet al., 2003b) that inside the vortex ozone increased at a given
ing lines between
vortex
air, depleted
in ozone,
and outsideinimpact
the polar
vortex. Wedescent
consideronidealised
conditions:
the
1993;
Müller
et al.,
2001;
Rex et al.,
Jost et (when
al., 2002;
of differential
ozone-tracer
relationships
N2 O2002;
level in December
the vortex was still relatively
vortex air. The vortex air is characterised on 23 January 2000
weak) and that ainside
compact and
O3 /N2 outO relation was established
vortex
described
by a We
single
columnidealised
of air; theconditions:
vertical coTilmes
et
al.,
2004),
mixing
lines
connecting
in theis polar
vortex.
consider
the
by ozone mixing ratios up to about 2.5 ppm (green symbols),
by early January (when a strong vortex had formed).
is mixing
perhaps
aupbit
more
realistic,
where
1θbetween
=2 K, α=0.33,
on 7 March 2000
byvortex
ozone
to about
ppm
ordinate
being
potential
θ . In
given
sidethat
airratios
masses
are1.7thus
expected
to lie
the
vortex is
described
by temperature
a single column
of aair
and timestep
the vertical
(blue symbols), and
on 11 March
2000
by
ozone mixing
raIn casesn=50,
when a comprehensive
set of tracer measurements
δ=1.5
and
the
number
of
timesteps
the
correspond∆tcoordinate
the main vortex
mass istemperature
descending θby
∆θa, given
with atimestep
fracinside
and
the
outside
vortex
ozone-tracer
relation
(Fig.
is the air
potential
. In
tios of about 1.2 ppm. Clearly noticeable are the mixing lines
is available, further insight on mixing1,
under weak vortex condifference
in column
is less
than
1 DU;
the maxitowards air top
withing
greater
ozone and greater
N2 O mixingozone
raditions
can be
obtained.
Considering
relations of two longtion
of main
the vortex
toby
be 1θ
descending
(unpanel).
1tαthe
vortexairairmass
massassumed
descends
, with a fraction
tios, that is air of mid-latitude characteristics. Clearly, there
lived tracers that are curved could be used to identify mixing.
mum change of ozone for a given
tracer value due to differα of the vortex air mass assumed to be descending (unmixed)
is no evidence from the measurements of mixing lines on the
If both tracers are chosen so that they are not subject to chemozone’ sideential
of the O3descent
/N2 O relations.
ical changeunder
over the timescale
in question, as it is the case e.g.
and subsequent mixing
these conditions
by δ1θ with δ>1. The vortex air mass is then assumed‘low
toWhen
be
the transport barrier at the edge
of the Chem.
vortex is Phys.,
for CFC-11/N
chemical change
will not interfere with the
www.atmos-chem-phys.org/acp/0000/0001/
Atmos.
0000,2 O,0001–15,
2005
is
≈50
ppb.
weak, mixing is expected to noticeably influence the ozonediagnosis of mixing.
immediately homogenised by horizontal mixing. This proceThis means that the contribution of the effect of differendure is repeated n times with the intention of approximately
Atmos. Chem. Phys., 0000, 0001–15, 2005
www.atmos-chem-phys.org/acp/0000/0001/
tial descent to the overall uncertainty in the
deduced chemidescribing the development of the vortex over a winter seacal loss will be small. Given the magnitude of other sources
son.
of uncertainty, the contribution from differential descent is
For the vortex air mass typical vertical profiles of ozone
likely to be insignificant. This finding corroborates the conand N2 O are prescribed as initial conditions, resulting in an
clusions of Ray et al. (2002) and Salawitch et al. (2002),
initial O3 /N2 O relation in the polar vortex (solid black line in
who report, based on balloon measurements in winter 1999–
Fig. 5). The change of this initial O3 /N2 O relation due to the
2000, that the effect of differential descent increased the unaction of differential descent and subsequent mixing in the
certainty of column ozone loss estimates by ≈3–4%.
vortex can then be followed. In Fig. 5, this change is shown
for a rather extreme case, where 1θ =5 K, α=0.33, δ=2 (that
3.2 Impact of cross-vortex-edge mixing on ozone-tracer reis, one third of the vortex air mass descends twice as much as
lations in the polar vortex
the remaining two thirds) and the number of timesteps n=20.
Obviously, even for such conditions, the O3 /N2 O relation is
little affected by the effect of differential descent.
Mixing across the vortex edge is usually non-local in tracertracer space and thus has the potential to alter tracer-tracer reOzone-tracer relations are used for calculating a proxy
lations. The ozone-tracer relation shows greater ozone mixozone profile from a tracer profile in late winter/early spring.
ing ratios outside the vortex than inside (e.g., Proffitt et al.,
If the two relations shown in Fig. 5 are used to deduce proxy
1993; Müller et al., 2001; Rex et al., 2002; Jost et al., 2002;
ozone profiles from a late winter vortex tracer profile characterised by vertical descent, the column ozone correspondTilmes et al., 2004). Mixing lines connecting inside- and
ing to these profiles differs by only ≈2.5 DU. In a scenario
outside-vortex air masses are thus expected to lie between
Atmos. Chem. Phys., 5, 3139–3151, 2005
www.atmos-chem-phys.org/acp/5/3139/
R. Müller et al.: Ozone/tracer relations in the polar vortex
3145
January
February
17−Feb−1997 to 21−Feb−1997
6
5
5
O3 (ppmv)
O3 (ppmv)
1−Jan−1997 to 8−Jan−1997
6
4
3
2
4
3
2
1
0
50
100
150
200
1
0
250
50
N2O (ppbv)
/private/walter/hda6/simone/JUELICH/ILASV5/haloe_intern_set_plot.pro
Wed Nov 17 12:06:18 2004
/private/walter/hda6/simone/JUELICH/ILASV5/haloe_intern_set_plot.pro
N2O_SCATTER/ilas_jan1_n2o_in_out_folien.ps
200
250
6
6
5
5
4
3
2
Wed Nov 17 12:06:28 2004
N2O_SCATTER/ilas_feb3_n2o_in_out_folien.ps
May
1−May−1997
to 15−May−1997
O3 (ppmv)
O3 (ppmv)
150
N2O (ppbv)
March
21−Mar−1997
to 31−Mar−1997
1
0
100
4
3
2
50
100
150
200
250
1
0
N2O (ppbv)
/private/walter/hda6/simone/JUELICH/ILASV5/haloe_intern_set_plot.pro
Wed Nov 17 12:06:36 2004
50
100
150
200
250
N2O (ppbv)
/private/walter/hda6/simone/JUELICH/ILASV5/haloe_intern_set_plot.pro
N2O_SCATTER/ilas_mar3_n2o_in_out_folien.ps
Wed Nov 17 12:06:42 2004
N2O_SCATTER/ilas_may1_n2o_in_out_folien.ps
Fig. 7. The development of the O3 /N2 O relation in the Arctic in winter 1996/1997 based on ILAS measurements (Tilmes et al., 2003b).
Shown are four different time periods (1–15 January, 17–21 February, 21–31 March, 1–15 May 1997). Black lines indicate profiles measured
inside the polar vortex core, defined using the Nash et al. (1996) criterion. Small black dots indicate measurements outside the polar vortex.
7. The development of the O3 /N2 O relation in the Arctic in winter 1996/1997 based on ILAS measurements (Tilmes et al., 2003b).
Also shownFig.
is the
early vortex reference relation (thick black line, Tilmes et al., 2003b) and its uncertainty (grey shaded area). The reference
Shown are four different time periods (January 1-15, February 17-21, March 21-31, May 1-15, 1997). Black lines indicate profiles measured
is deduced inside
from the
the polar
earlyvortex
January
ILAS measurements.
core, defined using the Nash et al. (1996) criterion. Small black dots indicate measurements outside the polar vortex.
Also shown is the early vortex reference relation (thick black line, Tilmes et al., 2003b) and its uncertainty (grey shaded area). The reference
is deduced from the early January ILAS measurements.
the inside and the outside vortex ozone-tracer relation (Fig. 1,
top panel).
3.2.1
Observations of mixing across the vortex edge
Based on high altitude aircraft measurements in the Arctic in
early 2000, Rex et al. (2002) and Jost et al. (2002) demonstrated thatAtmos.
mixing
across
the 0000,
vortex0001–15,
edge leads
Chem.
Phys.,
2005to an increase
in the mixing ratio of ozone for a given value of N2 O. Further, Rex et al. (2002) and Ray et al. (2002) emphasise that
although in this winter mixing lines were observed, entrainment of extra-vortex air did not significantly alter the composition of the vortex air between January and March 2000.
Figure 6 shows an example of measurements of such mixing lines between vortex air, depleted in ozone, and outsidevortex air. The vortex air is characterised on 23 January 2000
by ozone mixing ratios up to about 2.5 ppm (green symbols),
on 7 March 2000 by ozone mixing ratios up to about 1.7 ppm
(blue symbols), and on 11 March 2000 by ozone mixing ratios of about 1.2 ppm. Clearly noticeable are the mixing lines
towards air with greater ozone and greater N2 O mixing ratios, that is air of mid-latitude characteristics. Clearly, there
www.atmos-chem-phys.org/acp/5/3139/
is no evidence from the measurements of mixing lines on the
‘low ozone’ side of the O3 /N2 O relations.
When the transport barrier at the edge of the vortex is
weak, mixing is expected to noticeably influence the ozonetracer relation inside the vortex. According to the arguments
above, one expects ozone to increase for a given tracer mixing ratio. This effect was indeed observed by the ILAS instrument in winter
1996–1997 (Tilmes et al., 2003b). In this
www.atmos-chem-phys.org/acp/0000/0001/
winter, the vortex was perturbed by a stratospheric warming
in early winter and therefore the vortex edge was permeable
until late December. ILAS provided good coverage of the
Arctic polar region in winter 1996–1997 covering a latitude
range between 55◦ –77◦ N and measuring about fourteen profiles per day. These measurements show (see Fig. 7 in Tilmes
et al., 2003b) that inside the vortex ozone increased at a given
N2 O level in December (when the vortex was still relatively
weak) and that a compact O3 /N2 O relation was established
by early January (when a strong vortex had formed).
In cases where a comprehensive set of tracer measurements is available, further insight on mixing under weak vortex conditions can be obtained. Considering relations of two
long-lived tracers that are curved could be used to identify
Atmos. Chem. Phys., 5, 3139–3151, 2005
3146
R. Müller et al.: Ozone/tracer relations in the polar vortex
mixing. If both tracers are chosen so that they are not subject
to chemical change over the timescale in question as is the
case for example with CFC-11/N2 O, chemical change will
not interfere with the diagnosis of mixing.
Mixing across the vortex edge is likewise important for establishing the ozone tracer relation in the very early, incipient
vortex. The air in high latitudes in autumn from which the
polar vortex is formed, is characterised by low ozone mixing
ratios, a remainder of summertime ozone loss. Because the
transport barrier of the incipient vortex, while established,
is still somewhat permeable, ozone-rich, mid-latitude air is
mixed into the vortex. In this way, the ozone-tracer relation
in the vortex recovers from summertime ozone loss. This
process was first noted by Proffitt et al. (1992) and recently
analysed in detail based on ILAS II measurements in the high
latitudes of both hemispheres (Tilmes et al., 2005a1 ).
3.2.2
Cross-vortex-edge mixing in conceptual models
Plumb et al. (2000) presented a conceptual model, a simple advective diffusive model in cylindrical geometry, to describe the impact of mixing on tracer-tracer relations. Their
model simulates the effect of mixing on the relationship of
two artificial tracers χ1 and χ2 that are designed to resemble
N2 O and NOy . The relation of χ1 and χ2 is the same on both
sides of a transport barrier that represents the vortex edge in
the model. The typical time scale T for the model is the lifetime of the vortex, say four months or ≈107 s, and the typical
horizontal length scale R is from the pole to the subtropical
transport barrier, say ≈7·106 m. The dimensionless diffusivity in the vortex edge region is chosen to be K=0.05 by
Plumb et al. (2000). If this dimensionless diffusivity is converted to a dimensioned diffusivity D, where D=(K·R 2 )/T ,
one obtains D≈3·105 m2 s−1 .
This value may be compared with effective atmospheric diffusivities deduced from aircraft measurements
Deff =5·103 m2 s−1 (Waugh et al., 1997) or diffusivities
(D=1.1·103 m2 s−1 ) employed in the chemical transport
model CLaMS for reproducing observed filamentary tracer
structures at the polar vortex edge (Konopka et al., 2003a,
2005b). The diffusivities in the CLaMS model are the same
for studies addressing rather different problems such as reproducing observed filamentary tracer structures (e.g., Khosrawi et al., 2005; Konopka et al., 2005b) or simulating the
stratospheric warming in the Antarctic in September 2002
(Konopka et al., 2005a). Note that the Eulerian parametrisation of mixing in the model by Plumb et al. (2000) is different
from the Lagrangian concept used in CLaMS, where the mixing intensity is proportional to the gradient of the wind rather
than to the wind itself as is the case in Eulerian models.
That means, the cross vortex edge diffusivity is likely to
be substantially greater in the Plumb et al. (2000) conceptual
model than in the stratosphere. Therefore, conclusions based
on the results of this conceptual model do not carry over directly to stratospheric conditions. A further caveat for the
Atmos. Chem. Phys., 5, 3139–3151, 2005
application of conclusions from this conceptual model to the
behaviour of ozone-tracer relations in the stratospheric polar
vortex is that Plumb et al. (2000) assume the same tracertracer relation inside and outside the vortex; an assumption
that is not valid for the ozone-tracer relation in the polar region (Müller et al., 2001, see also discussion below).
3.2.3 Cross-vortex mixing in a Lagrangian chemical transport model
Konopka et al. (2004, 2005c3 ) have shown that in the Lagrangian model CLaMS rather little exchange occurs across
the polar vortex edge as long as a stable vortex exists. If
exchange occurs, air masses will mix approximately along
isentropic surfaces leading to an increase of ozone mixing
ratios with respect to the vortex reference relation (Fig. 4).
However, the lower the potential temperature level at which
the mixing occurs, the less is the deviation of the isentropic
surfaces from the vortex reference relation in ozone-tracer
space (Fig. 4). That means, the lower down in the vortex
mixing occurs, the smaller the impact of mixing events is
on the ozone tracer relation (see Lemmen et al., 20054 for a
more detailed discussion of this issue).
4
Discussion
We have discussed here the validity of the two major hypotheses on which the use of ozone-tracer relations for deducing chemical ozone loss in the polar vortex is based,
namely that a compact ozone-tracer relation is established
in the ‘early’ polar vortex before the onset of chemical ozone
loss and that any change of the ozone-tracer relation in the
vortex over the course of winter is mainly caused by chemical ozone loss.
Based on both measurements in the polar stratosphere and
model results, we argue that a compact ozone-tracer relation develops within the polar vortex. Model results (Sankey
and Shepherd, 2003) showing no compact ozone-tracer relation in the Arctic vortex and a rather slow development of
compactness of the ozone-tracer relation in the Antarctic vortex are likely caused by an insufficient representation of the
transport barrier at the vortex edge in the model employed.
Recently, Konopka et al. (2004) analysed the development
of ozone-tracer relations in simulations of the Lagrangian
model CLaMS (McKenna et al., 2002b), a model in which a
strong transport barrier develops at the edge of the polar vortex (e.g. Konopka et al., 2003b, 2004; Steinhorst et al., 2005).
Compact ozone-tracer relations are found in the CLaMS results between simulated tracer (CH4 ) and passive ozone mixing ratios in the Antarctic vortex for September to November
4 Lemmen, C., Müller, R., Konopka, P., and Dameris, M.:
Tracer-tracer relations as a tool for analyzing polar chemical ozone
loss in chemistry-climate models, J. Geophys. Res., submitted,
2005.
www.atmos-chem-phys.org/acp/5/3139/
R. Müller et al.: Ozone/tracer relations in the polar vortex
2002 (Konopka et al., 2004). Compact correlations are likewise found for the ozone-tracer relation, when ozone is deduced from POAM satellite measurements interpolated to the
CLaMS model grid points. Similarly, CLaMS model simulations for the mid-winter Arctic vortex (Fig. 4, Konopka et al.,
2005c3 ) show both a clear separation between mid-latitude
and vortex air masses and a compact ozone-tracer relation
within the vortex.
Indeed, HALOE measurements in the Arctic vortex in January 1992 show a compact ozone-tracer relation (Fig. 3). The
same is true for ILAS measurements in the Arctic vortex in
January 1997 which provide a better coverage of the vortex
(Fig. 7, see also Tilmes et al., 2003b).
Reference relations for deducing ozone loss that are derived from mixing lines (Michelsen et al., 1998) yield results
for the deduced chemical ozone loss comparable to those obtained by using an ‘early vortex’ relation from within the vortex. Nonetheless, we consider an ‘early vortex’ relation as a
more reliable reference than one derived from mixing lines.
Inside- and outside-vortex air is characterised by different
ozone-tracer relationships, with the outside-vortex relationships showing greater ozone mixing ratios (and a stronger
variability) for equal values of the tracer than inside relationships (e.g., Proffitt et al., 1990; Müller et al., 1999, 2002;
Tilmes et al., 2003b). Consequently, mixing between air
masses inside and outside the vortex should lead to points
lying above the original ozone-tracer relation in the vortex
(Fig. 1, top panel), as has been indeed observed in Arctic
winter 2000 (Rex et al., 2002; Jost et al., 2002, see also
Fig. 6).
Therefore, mixing across the vortex edge, if it occurs during the period of chemical ozone loss, should lead to an ‘underestimation’ of ozone loss. This finding is in contrast to the
conclusions of Plumb et al. (2000), who state that ‘estimates
of ozone depletion inferred from ozone-tracer relations are
likely to be overestimates’. We argue here that this different
conclusion is reached by Plumb et al. (2000) because they
assume the same tracer-tracer relation inside and outside the
vortex and use a rather high diffusivity at the vortex edge in
their conceptual model.
Simulations with the Lagrangian model CLaMS (Konopka
et al., 2004, 2005c3 ; Lemmen et al., 20054 ) indicate that if
exchange across the transport barrier at the vortex edge occurs, it will lead to an increase of ozone mixing ratios in
the vortex. Further, as air masses will mix approximately
along isentropic surfaces crossing the vortex edge, mixing
will have an increasingly smaller impact on the ozone tracer
relation in the vortex, the lower down a mixing event occurs (Fig. 4). Indeed, Jost et al. (2002) report a case where
mixing across the vortex edge occurred at ≈380 K in early
March of Arctic winter 2000 that did not noticeably change
the O3 /N2 O relation.
Esler and Waugh (2002) proposed to construct an artificial
reference tracer from a linear combination of several longlived tracers so that a linear relation of a chemically active
www.atmos-chem-phys.org/acp/5/3139/
3147
tracer with the artificial reference tracer would result. In
this way, the impact of a mixing event of the type shown
in Fig. 1, bottom panel, can be removed from the analysis
since linear tracer-tracer relationships are unaffected by such
mixing events. However, since inside- and outside-vortex
air is characterised by different ozone-tracer relationships,
the construction of an artificial reference tracer as suggested
by Esler and Waugh (2002) (for estimating denitrification)
would not help to eliminate the impact of mixing between
inside- and outside-vortex air masses on ozone-tracer relations inside the polar vortex.
Differential descent within the vortex with subsequent horizontal mixing results in mixing events of the type shown
in Fig. 1, bottom panel. We have shown here that over the
lifetime of the polar vortex this mechanism can only have
a minor impact on ozone-tracer relations within the vortex
(Fig. 5), corroborating conclusions based on measurements
in winter 1999/2000 (Ray et al., 2002; Salawitch et al., 2002).
In any event, if chemical ozone loss in the polar vortex
occurs, it will reduce ozone mixing ratios in the vortex and
thus it will alter the reference ozone-tracer relation. In winter
1996/97 (Fig. 7), the beginning ozone depletion is clearly noticeable as a moderate deviation of the ozone-tracer relation
from the reference relation. In late March, after substantial
ozone loss, a strong deviation from the reference relation is
obvious. The ozone-tracer relation in March is ‘not’ compact
because of a chemical reason, namely substantially different chemical ozone loss in different parts of the polar vortex
(McKenna et al., 2002a; Tilmes et al., 2003b). In early May
1997, compact ozone-tracer relations were again observed in
the remaining vortex, consistent with the view that mixing in
the vortex had, by that time, homogenised the vortex air mass
again.
The chemical ozone loss deduced using the TRAC method
has been compared with the results of other methods (Harris
et al., 2002; Tilmes et al., 2004). The results deduced using tracer-tracer relations are in general agreement with those
obtained by other methods; however, the very strong ozone
loss deduced from SAOZ measurements for January in some
winters (e.g., Goutail et al., 1999) cannot be identified with
TRAC.
5
Conclusions
Ozone-tracer relations have been used for more than a decade
to quantify chemical ozone loss in the polar vortex. An issue
neglected in the early studies is that it is necessary to assess
the impact of mixing processes on tracer-tracer relationships
in the polar vortex for a reliable quantification of ozone loss.
Recently, also the question of the behaviour of ozone-tracer
correlations in the results of model simulations has attracted
interest.
Analysing various sets of measurements and the results
from several models we find here that mixing across the
Atmos. Chem. Phys., 5, 3139–3151, 2005
3148
6
Ozone (ppm)
Pre−Reference
4
2
0
6
Ozone (ppm)
Reference
4
2
0
6
Ozone (ppm)
Beginning Loss
4
2
0
6
Ozone (ppm)
Established Loss
4
2
0
0
50
100
150
N2O (ppb)
200
250
Fig. 8. Schematic view of the development of ozone-tracer relations
over the course of the winter inside the polar vortex. In all panels,
the thick solid line indicates the reference relation for ‘early vortex’
conditions and the thick dashed line an average mid-latitude relation. The grey shaded area indicates schematically the variability of
Fig. 8. Schematic
view of
the development
of ozone-tracer
the ozone-tracer
relations
within the vortex.
The dotted relations
line in the
panelofshows
the ozone-tracer
relation
the incipient
beover the top
course
the winter
inside the
polarinvortex.
In allvortex,
panels,
transport the
barrier
is established
at the
The
the thick fore
solida strong
line indicates
reference
relation
forvortex
‘earlyedge.
vortex’
panel
showsdashed
the reference
relation
for an
established vortex
conditions
andbelow
the thick
line an
average
mid-latitude
relaand one panel below the ozone-tracer relation for beginning chemtion. The grey shaded area indicates schematically the variability of
ical ozone loss. The bottom panel shows the conditions for estabthe ozone-tracer
relations
within
the vortex.
The dotted
line in
the
lished ozone
loss, the
dash-dotted
line indicates
a possible
mixing
top panelline
shows
the ozone-tracer
relation inair.
the incipient vortex, bebetween
vortex and out-of-vortex
fore a strong transport barrier is established at the vortex edge. The
panel below shows the reference relation for an established vortex
and one further panel below the ozone-tracer relation for beginning
vortex edge can only lead to an ‘underestimation’ of ozone
chemicalloss
ozone
The
panel shows
conditions
for Furesandloss.
not to
an bottom
overestimation
as hasthe
been
suggested.
tablishedther,
ozone
loss,
the
dash-dotted
line
indicates
a
possible
mixdifferential descent within the vortex and the subsequent
ing line between
vortex and
air. impact on ozone loss
internal mixing
has out-of-vortex
only a negligible
estimates.
Moreover, the representation of mixing in threedimensional atmospheric models can have a substantial imparticular importance is a realistic representation of transpact on the development of tracer relations in the model. Of
port barriers
(such as the polar vortex edge) in the model.
particular importance is a realistic representation of transport
Rather compact
ozone-tracer
develop
—in agreebarriers (such
as the polarrelations
vortex edge)
in the model.
Rather
ment with observations— in the vortex of a Lagrangian
Atmos. Chem. Phys., 5, 3139–3151, 2005
Atmos. Chem. Phys., 0000, 0001–15, 2005
the deformation of the flow.
Drawing together all pieces of information considered
here, the following picture of the development of ozoneR. Müller et al.: Ozone/tracer relations in the polar vortex
tracer relations in the polar vortex develops (Fig. 8): In the
incipient vortex the vortex air mass is characterised by low
compact ozone-tracer relations develop – in agreement with
ozone
mixing ratios; the result of polar ozone destruction
observations – in the vortex of a Lagrangian model (CLaMS)
due
to
NO
chemical
in the
summer
(e.g., of
Farman,
x driven
where mixing
is anisotropic
and cycles
driven by
deformation
1985)
and
autumn
(Kawa
et
al.,
2002;
Tilmes
et
al.,
2005a).
the flow.
Although
vortex
air pieces
mass ofisinformation
separated considered
to a certain exDrawing the
together
all the
here,from
the following
picture
emerges of air,
the development
tent
ozone-rich,
mid-latitude
the vortex isofnot yet
ozone-tracer isolated.
relations inThus,
the polar
vortexmixing,
developsozone
(Fig. 8).
completely
through
values in
In
the
incipient
vortex
the
vortex
air
mass
is
characterised
by2003b,
the vortex increase during this period (Tilmes et al.,
low ozone mixing ratios; the result of polar ozone destruction
2005b). When the vortex is well established, the transport
due to NOx -driven chemical cycles in summer (e.g., Farman,
barrier
at autumn
the vortex
edge
effectively
separates
the vortex
air
1 ).
1985) and
(Kawa
et al.,
2002; Tilmes
et al., 2005a
from
mid-latitude
air.
Mixing
within
the
vortex
air
mass
is
Although the vortex air mass is separated to a certain exproducing
a
compact
ozone-tracer
relation
in
the
vortex.
This
tent from ozone-rich, mid-latitude air, the vortex is not yet
completely
Thus, through
mixing,relation
ozone values
in TRAC
is
the pointisolated.
in time when
the reference
for the
the vortexshould
increase
duringbethis
period (Tilmes et al., 2003b,
method
ideally
derived.
2005b).
the vortex
is well chemical
established,ozone
the transport
In coldWhen
winters,
a beginning
loss becomes
barrier at the vortex edge effectively separates the vortex air
first noticeable at greater altitudes (≈500 K), typically in late
from mid-latitude air. Mixing within the vortex air mass proJanuary/early
in the
Arctic
(Tilmes
al., is2003b).
duces a compactFebruary
ozone-tracer
relation
in the
vortex.etThis
Towards
late
winter,
with
solar
illumination
strongly
increasthe point in time when the reference relation for the TRAC
ing,
the
ozone
loss
is
accelerating.
Thus,
in
early
spring,
the
method should ideally be derived.
In cold
winters, a beginning
chemical
ozone loss
firstozonesignal
of accumulated
ozone loss
is pronounced
in the
becomes
noticeable
altitudes
K),such
typically
tracer
relation
(Fig.at8,greater
bottom
panel).(≈500
Under
conditions,
in late January/early
February
in is
theclearly
Arctic leading
(Tilmes to
et al.,
mixing
across the vortex
edge
an under2003b). Towards late winter, with solar illumination strongly
estimate of ozone loss.
increasing, the ozone loss is accelerated. Thus, in early
We conclude
that for using
ozone-tracer
relations
spring,
there is a pronounced
signal
of accumulated
ozone to deduce
chemical
ozone
loss,
it
is
crucial
that
a
reliable
loss in the ozone-tracer relation (Fig. 8, bottom panel). Un- ‘early
vortex’
be obtained
thatedge
vortex
measureder suchreference
conditions,can
mixing
across theand
vortex
clearly
leads toare
an underestimate
of ozone
ments
well separated
fromloss.
mid-latitude measurements.
We conclude
that when
usingozone-tracer
ozone-tracer relations
to constitute
deIf these
conditions
are met,
relations
chemical
loss,
it is crucialdetermination
that a reliable ‘early
aduce
reliable
toolozone
for the
quantitative
of chemical
vortex’ reference can be obtained and that vortex measureozone
loss in the polar vortex.
ments are well separated from mid-latitude measurements.
If these conditions are met, ozone-tracer relations constitute
Acknowledgements.
thank D.determination
Sankey, T. of
S.chemical
Shepherd, and
a reliable tool for the We
quantitative
C.
Lemmen
for
fruitful
discussions
and
for
helpful
comments on
ozone loss in the polar vortex.
the paper. Further, we thank J. W. Elkins for the ACATS data,
Acknowledgements.
We thank
D. Sankey,
T. G.
S. C.
Shepherd,
E.
Richard for the ozone
photometer
data,
Toon forand
the MkIV
C. Lemmen
for fruitfulfor
discussions
anddata,
for helpful
on III and
data,
C. R. Webster
the ALIAS
and J.comments
M. Russell
the HALOE
paper. Further,
we the
thank
J. W. Elkins
the ACATS data,
the
team for
HALOE
(v19) for
data.
E. Richard for the ozone photometer data, G. C. Toon for the MkIV
data, C. R. Webster for the ALIAS data, and J. M. Russell III and
the HALOE team for the HALOE (v19) data.
References
Edited by: M. Dameris
Elkins, J. W., Fahey, D. W., Gilligan, J. M., Dutton, G. S., Baring,
T. J., Volk, C. M., Dunn, R. E., Myers, R. C., Montzka, S. A.,
Wamsley, P. R., Hayden, A. H., Butler, J. H., Thompson, T. M.,
Swanson, T. H., Dlugokencky, E. J., Novelli, P. C., Hurst, D. F.,
Lobert, J. M., Ciciora, S. J., McLaughlin, R. J., Thompson, T. L.,
Winkler, R. H., Fraser, P. J., Steele, L. P., and Lucarelli, M. P.:
Airborne gas chromatograph for in situ measurements of longlived species in the upper troposphere and lower stratosphere,
Geophys. Res. Lett., 23(4), 347–350, doi:10.1029/96GL00244,
1996.
www.atmos-chem-phys.org/acp/5/3139/
www.atmos-chem-phys.org/acp/0000/0001/
R. Müller et al.: Ozone/tracer relations in the polar vortex
References
Elkins, J. W., Fahey, D. W., Gilligan, J. M., Dutton, G. S., Baring,
T. J., Volk, C. M., Dunn, R. E., Myers, R. C., Montzka, S. A.,
Wamsley, P. R., Hayden, A. H., Butler, J. H., Thompson, T. M.,
Swanson, T. H., Dlugokencky, E. J., Novelli, P. C., Hurst, D. F.,
Lobert, J. M., Ciciora, S. J., McLaughlin, R. J., Thompson, T. L.,
Winkler, R. H., Fraser, P. J., Steele, L. P., and Lucarelli, M. P.:
Airborne gas chromatograph for in situ measurements of longlived species in the upper troposphere and lower stratosphere,
Geophys. Res. Lett., 23(4), 347–350, doi:10.1029/96GL00244,
1996.
Esler, J. G. and Waugh, D. W.: A method for estimating
the extent of denitrification of Arctic polar vortex air from
tracer-tracer scatter plots, J. Geophys. Res., 107(6), 1–14,
doi:10.1029/2001JD001071, 2002.
Fahey, D. W., Kelly, K. K., Kawa, S. R., Tuck, A. F., Loewenstein,
M., Chan, K. R., and Heid, L. E.: Observations of denitrification
and dehydration in the winter polar stratosphere, Nature, 344,
321–324, 1990.
Farman, J. C.: Ozone photochemistry in the Antarctic stratosphere
in summer, Q. J. R. Meteorol. Soc., 111, 1013–1025, 1985.
Goutail, F., Pommereau, J.-P., Phillips, C., Deniel, C., Sarkssian,
A., Lefèvre, F., Kyrö, E., Rummukainen, M., Ericksen, P., Andersen, S. B., Kaastad-Hoiskar, B.-A., Braathen, G., Dorokhov,
V., and Khattatov, V. U.: Depletion of Column Ozone in the Arctic during the winter of 1993–94 and 1994–95, J. Atmos. Chem.,
32, 1–34, 1999.
Harris, N. R., Rex, M., Goutail, F., Knudsen, B. M., Manney, G. L.,
Müller, R., and von der Gathen, P.: Comparison of empirically
derived ozone loss rates in the Arctic vortex, J. Geophys. Res.,
107(D20), doi:10.1029/2001JD000482, 2002.
Jost, H.-J., Loewenstein, M., Greenblatt, J. G., Podolske, J. R., Bui,
T. P., Hurst, D. F., Elkins, J. W., Herman, R. L., Webster, C. R.,
Schauffler, S. M., Atlas, E. L., Newman, P. A., Lait, L., and
Wofsy, S. C.: Mixing events revealed by anomalous tracer relationships in the Arctic vortex during winter 1999/2000, J. Geophys. Res., 107, 4795, doi:10.1029/2002JD002380, 2002.
Kawa, S. R., Bevilacqua, R. M., Margitan, J. J., Douglass, A. R.,
Schoeberl, M. R., Hoppel, K. W., and Sen, B.: Interaction between dynamics and chemistry of ozone in the setup phase of the
Northern Hemisphere polar vortex, J. Geophys. Res., 108, 8310,
doi:10.1029/2001JD001527, 2002.
Khosrawi, F., Müller, R., Proffitt, M., and Nakajima, H.: Monthly
averaged ozone and nitrous oxide from the Improved Limb
Atmospheric Spectrometer (ILAS) in the Northern and Southern Hemisphere polar regions, J. Geophys. Res., 109, D10301,
doi:10.1029/2003JD004365, 2004.
Khosrawi, F., Grooß, J.-U., Müller, R., Konopka, P., Kouker, W.,
Ruhnke, R., Reddmann, T., and Riese, M.: Intercomparison
between Lagrangian and Eulerian simulations of the development of mid-latitude streamers as observed by CRISTA, Atmos.
Chem. Phys., 5, 85–95, 2005,
SRef-ID: 1680-7324/acp/2005-5-85.
Konopka, P., Grooß, J.-U., Bausch, S., Müller, R., McKenna, D. S.,
Morgenstern, O., and Orsolini, Y.: Dynamics and chemistry of
vortex remnants in late Arctic spring 1997 and 2000: Simulations with the Chemical Lagrangian Model of the Stratosphere
(CLaMS), Atmos. Chem. Phys., 3, 839–849, 2003a,
SRef-ID: 1680-7324/acp/2003-3-839.
www.atmos-chem-phys.org/acp/5/3139/
3149
Konopka, P., Grooß, J. U., Günther, G., McKenna, D. S., Müller, R.,
Elkins, J. W., Fahey, D., and Popp, P.: Weak impact of mixing on
chlorine deactivation during SOLVE/THESEO2000: Lagrangian
modeling (CLaMS) versus ER-2 in situ observations, J. Geophys.
Res., 108, 8324, doi:10.1029/2001JD000876, 2003b.
Konopka, P., Steinhorst, H.-M., Grooß, J.-U., Günther, G., Müller,
R., Elkins, J. W., Jost, H.-J., Richard, E., Schmidt, U., Toon, G.,
and McKenna, D. S.: Mixing and Ozone Loss in the 1999–2000
Arctic Vortex: Simulations with the 3-dimensional Chemical Lagrangian Model of the Stratosphere (CLaMS), J. Geophys. Res.,
109, D02315, doi:10.1029/2003JD003792, 2004.
Konopka, P., Grooß, J.-U., Hoppel, K., Steinhorst, H.-M., and
Müller, R.: Mixing and chemical ozone loss during and after
the Antarctic polar vortex major warming in September 2002,
J. Atmos. Sci., 62(3), 848–859, 2005a.
Konopka, P., Günther, G., McKenna, D. S., Müller, R., Offermann,
D., Spang, R., and Riese, M.: How homogeneous and isotropic
is stratospheric mixing? Comparison of CRISTA-1 observations
with transport studies based on the Chemical Lagrangian Model
of the Stratosphere (CLaMS), Q. J. R. Meteorol. Soc., 131(606),
565–579, doi:10.1256/qj.04.47, 2005b.
Loewenstein, M., Jost, H., Grose, J., Eilers, J., Lynch, D., Jensen,
S., and Marmie, J.: Argus: A new instrument for the measurement of the stratospheric dynamical tracers, N2 O and CH4 ,
Spectrochimica Acta Part A-Molecular and Biomolecular Spectroscopy, 58(11), 2329–2345, 2002.
McKenna, D. S., Grooß, J.-U., Günther, G., Konopka, P., Müller, R.,
Carver, G., and Sasano, Y.: A new Chemical Lagrangian Model
of the Stratosphere (CLaMS): Part II Formulation of chemistryscheme and initialisation, J. Geophys. Res., 107(D15), 4256,
doi:10.1029/2000JD000113, 2002a.
McKenna, D. S., Konopka, P., Grooß, J.-U., Günther, G., Müller,
R., Spang, R., Offermann, D., and Orsolini, Y.: A new Chemical
Lagrangian Model of the Stratosphere (CLaMS): Part I Formulation of advection and mixing, J. Geophys. Res., 107(D16), 4309,
doi:10.1029/2000JD000114, 2002b.
Michelsen, H. A., Manney, G. L., Gunson, M. R., and Zander,
R.: Correlations of stratospheric abundances of NOy , O3 , N2 O,
and CH4 derived from ATMOS measurements, J. Geophys. Res.,
103, 28 347–28 359, 1998.
Müller, R.: The performance of classical versus modern finitevolume advection schemes for atmospheric modelling in a onedimensional test-bed, Mon. Wea. Rev., 120, 1407–1415, 1992.
Müller, R., Crutzen, P. J., Grooß, J.-U., Brühl, C., Russel III, J. M.,
and Tuck, A. F.: Chlorine activation and ozone depletion in the
Arctic vortex: Observations by the Halogen Occultation Experiment on the Upper Atmosphere Research Satellite, J. Geophys.
Res., 101, 12 531–12 554, 1996.
Müller, R., Grooß, J.-U., McKenna, D. S., Crutzen, P. J., Brühl, C.,
Russell, J. M., and Tuck, A. F.: HALOE observations of the vertical structure of chemical ozone depletion in the Arctic vortex
during winter and early spring 1996–1997, Geophys. Res. Lett.,
24, 2717–2720, 1997.
Müller, R., Grooß, J.-U., McKenna, D. S., Crutzen, P. J., Brühl, C.,
Russell, J. M., Gordley, L. L., Burrows, J. P., and Tuck, A. F.:
Chemical ozone loss in the Arctic vortex in the winter 1995–
1996: HALOE measurements in conjunction with other observations, Ann. Geophys., 17, 101–114, 1999,
SRef-ID: 1432-0576/ag/1999-17-101.
Atmos. Chem. Phys., 5, 3139–3151, 2005
3150
Müller, R., Schmidt, U., Engel, A., McKenna, D. S., and Proffitt,
M. H.: The O3 /N2 O relationship from balloon-borne observations as a measure of Arctic ozone loss in 1991–1992, Q. J. R.
Meteorol. Soc., 127, 1389–1412, 2001.
Müller, R., Tilmes, S., Grooß, J.-U., McKenna, D. S., Müller, M.,
Schmidt, U., Toon, G. C., Stachnik, R. A., Margitan, J. J., Elkins,
J. W., Arvelius, J., and Russell III, J. M.: Chlorine Activation and
Chemical Ozone Loss Deduced from HALOE and Balloon Measurements in the Arctic during the Winter of 1999–2000, J. Geophys. Res., 107, 8302, doi:10.1029/2001JD001423, 2002.
Nash, E. R., Newman, P. A., Rosenfield, J. E., and Schoeberl, M. R.:
An objective determination of the polar vortex using Ertel’s potential vorticity, J. Geophys. Res., 101, 9471–9478, 1996.
Plumb, R. A.: Stratospheric transport, J. Meteorol. Soc. Jpn.,
80(4B), 793–809, 2002.
Plumb, R. A. and Ko, M. K. W.: Interrelationships between mixing
ratios of long-lived stratospheric constituents, J. Geophys. Res.,
97, 10 145–10 156, 1992.
Plumb, R. A., Waugh, D. W., and Chipperfield, M. P.: The effect of
mixing on tracer relationships in the polar vortices, J. Geophys.
Res., 105, 10 047–10 062, 2000.
Proffitt, M. H. and McLaughlin, R. J.: Fast-response dual-beam UV
absorption ozone photometer suitable for use on stratospheric
balloons, Rev. Sci. Instr., 54, 1719–1728, 1983.
Proffitt, M. H., Margitan, J. J., Kelly, K. K., Loewenstein, M.,
Podolske, J. R., and Chan, K. R.: Ozone loss in the Arctic polar
vortex inferred from high altitude aircraft measurements, Nature,
347, 31–36, 1990.
Proffitt, M. H., Solomon, S., and Loewenstein, M.: Comparison of
2-D model simulations of ozone and nitrous oxide at high latitudes with stratospheric measurements, J. Geophys. Res., 97,
939–944, 1992.
Proffitt, M. H., Aikin, K., Margitan, J. J., Loewenstein, M.,
Podolske, J. R., Weaver, A., Chan, K. R., Fast, H., and Elkins,
J. W.: Ozone loss inside the northern polar vortex during the
1991–1992 winter, Science, 261, 1150–1154, 1993.
Proffitt, M. H., Aikin, K., Tuck, A. F., Margitan, J. J., Webster,
C. R., Toon, G. C., and Elkins, J. W.: Seasonally averaged ozone
and nitrous oxide in the northern hemisphere lower stratosphere,
J. Geophys. Res., 108, 4110, doi:10.1029/2002JD002657, 2003.
Ray, E. A., Moore, F. L., Elkins, J. W., Hurst, D. F., Romashkin, P. A., Dutton, G. S., and Fahey, D. W.: Descent
and mixing in the 1999–2000 northern polar vortex inferred
from in situ tracer measurements, J. Geophys. Res., 107, 8285,
doi:10.1029/2001JD000961, 2002.
Rex, M., Salawitch, R. J., Harris, N. R. P., von der Gathen, P.,
Schulz, G. O. B. A., Deckelman, H., Chipperfield, M., Sinnhuber, B.-M., Reimer, E., Alfier, R., Bevilacqua, R., Hoppel, K.,
Fromm, M., Lumpe, J., Küllmann, H., Kleinböhl, A., von König,
H. B. M., Künzi, K., Toohey, D., Vömel, H., Richard, E., Aiken,
K., Jost, H., Greenblatt, J. B., Loewenstein, M., Podolske, J. R.,
Webster, C. R., Flesch, G. J., Scott, D. C., Herman, R. L., Elkins,
J. W., Ray, E. A., Moore, F. L., Hurst, D. F., Romanshkin, P.,
Toon, G. C., Sen, B., Margitan, J. J., Wennberg, P., Neuber,
R., Allart, M., Bojkov, B. R., Claude, H., Davies, J., Davies,
W., De Backer, H., Dier, H., Dorokhov, V., Fast, H., Kondo,
Y., Kyrö, E., Litynska, Z., Mikkelsen, I. S., Molyneux, M. J.,
Moran, E., Nagai, T., Nakane, H., Parrondo, C., Ravegnani, F.,
Viatte, P. S. P., and Yushkov, V.: Chemical depletion of Arc-
Atmos. Chem. Phys., 5, 3139–3151, 2005
R. Müller et al.: Ozone/tracer relations in the polar vortex
tic ozone in winter 1999/2000, J. Geophys. Res., 107, 8276,
doi:10.1029/2001JD000533, 2002.
Richard, E. C., Aikin, K. C., Andrews, A. E., Daube, B. C., Gerbig, C., Wofsy, S. C., Romashkin, P. A., Hurst, D. F., Ray, E. A.,
Moore, F. L., Elkins, J. W., Deshler, T., and Toon, G. C.: Severe chemical ozone loss in the Arctic polar vortex during winter
1999–2000 inferred from in-situ airborne measurements, Geophys. Res. Lett., 28(11), 2197–2000, 2001.
Roach, W. T.: Aircraft observations in the lower sub-Arctic stratosphere in winter, Meteorol. Res. Comm. Pap., 121, 1962.
Robinson, A. D., Millard, G. A., Danis, F., Guirlet, M., Harris, N.
R. P., Lee, A. M., McIntyre, J. D., Pyle, J. A., Arvelius, J., Dagnesjo, S., Kirkwood, S., Nilsson, H., Toohey, D. W., Deshler,
T., Goutail, F., Pommerau, J.-P., Elkins, J. W., Moore, F., Ray,
E., Schmidt, U., Engel, A., and Müller, M.: Ozone loss derived
from balloon-borne tracer measurements in the 1999/2000Arctic
winter, Atmos. Chem. Phys., 5, 1423–1436, 2005,
SRef-ID: 1680-7324/acp/2005-5-1423.
Rood, R. B.: Numerical advection algorithms and their role in atmospheric transport and chemistry models, Rev. Geophys., 25(1),
71–100, 1987.
Ross, D., Pyle, J., Harris, N., McIntyre, J., Millard, G., Robinson, A., and Busen, R.: Investigation of Arctic ozone depletion sampled over midlatitudes during the Egrett campaign of
spring/summer 2000, Atmos. Chem. Phys., 4, 1407–1417, 2004,
SRef-ID: 1680-7324/acp/2004-4-1407.
Russell, J. M., Gordley, L. L., Park, J. H., Drayson, S. R., Tuck,
A. F., Harries, J. E., Cicerone, R. J., Crutzen, P. J., and Frederick,
J. E.: The Halogen Occultation Experiment, J. Geophys. Res.,
98, 10 777–10 797, 1993.
Salawitch, R. J., Margitan, J., Sen, B., Toon, G. C., Osterman, G. B.,
Rex, M., Elkins, J. W., Ray, E. A., Moore, F. L., Hurst, D. F.,
Romashkin, P. A., Bevilacqua, R. M., Hoppel, K., Richard, E. C.,
and Bui, T. P.: Chemical loss of ozone during the Arctic winter
of 1999–2000: an analysis based on balloon-borne observations,
J. Geophys. Res., 107(D20), 8269, doi:10.1029/2001JD000620,
2002.
Sankey, D. and Shepherd, T. G.: Correlations of long-lived chemical species in a middle atmosphere general circulation model,
J. Geophys. Res., 108(D16), 4494, doi:10.1029/2002JD002799,
2003.
Sasano, Y., Suzuki, M., Yokota, T., and Kanzawa, H.: Improved
Limb Atmospheric Spectrometer (ILAS) for stratospheric ozone
layer measurements by solar occultation technique, Geophys.
Res. Lett., 26, 197–200, doi:10.1029/1998GL900276, 1999.
Steinhorst, H.-M., Konopka, P., Günther, G., and Müller, R.:
How permeable is the edge of the Arctic vortex – Model studies of the winter 1999–2000, J. Geophys. Res., 110, D06105,
doi:10.1029/2004JD005268, 2005.
Thuburn, J. and McIntyre, M.: Numerical advection schemes,
cross-isentropic random walks, and correlations between chemical species, J. Geophys. Res., 102(D6), 6775–6797, 1997.
Tilmes, S.: Chemical ozone loss in the Arctic polar stratosphere,
Tech. rep., Institute for Chemistry and Dynamics of the Geosphere, Jülich, Germany, 2004.
Tilmes, S., Müller, R., Grooß, J.-U., Höpfner, M., Toon, G. C., and
Russell, J. M.: Very early chlorine activation and ozone loss in
the Arctic winter 2002–2003, Geophys. Res. Lett., 30(23), 2201,
doi:10.1029/2003GL018079, 2003a.
www.atmos-chem-phys.org/acp/5/3139/
R. Müller et al.: Ozone/tracer relations in the polar vortex
Tilmes, S., Müller, R., Grooß, J.-U., McKenna, D. S., Russell, J. M.,
and Sasano, Y.: Calculation of chemical ozone loss in the Arctic
winter 1996-1997 using ozone-tracer correlations: Comparison
of Improved Limb Atmospheric Spectrometer (ILAS) and Halogen Occultation Experiment (HALOE) results, J. Geophys. Res.,
108, 4045, doi:10.1029/2002JD002213, 2003b.
Tilmes, S., Müller, R., Grooß, J.-U., and Russell, J. M.: Ozone loss
and chlorine activation in the Arctic winters 1991–2003 derived
with the tracer-tracer correlations, Atmos. Chem. Phys., 4(8),
2181–2213, 2004,
SRef-ID: 1680-7324/acp/2004-4-2181.
Tilmes, S., Müller, R., Grooß, J.-U., Spang, R., Sugita, T., Nakajima, H., and Sasano, Y.: Chemical ozone loss and related processes in the Antarctic winter 2003 based on ILAS II observations, J. Geophys. Res., accepted, 2005b.
Toon, G. C., Blavier, J.-F., Sen, B., Margitan, J. J., Webster, C. R.,
May, R. D., Fahey, D., Gao, R., Del Negro, L., Proffitt, M.,
Elkins, J. W., Romashkin, P. A., Hurst, D. F., Oltmans, S., Atlas,
E., Schauffler, S., Flocke, F., Bui, T. P., Stimpfle, R. M., Bonne,
G. P., Voss, P. B., and Cohen, R. C.: Comparison of MkIV balloon and ER-2 aircraft measurements of atmospheric trace gases,
J. Geophys. Res., 104, 26 779–26 790, 1999.
Ulanovskii, A. E., Luk’yanov, A. N., Yushkov, V. A., Sitnikov,
N. M., Volk, M., Ivanova, E. V., and Ravegnani, F.: Estimation
of the chemical loss of ozone in the Antarctic stratosphere in the
1999 winter-spring season from direct measurements and simulations, Izv. Atmos. Ocean. Phys., 40(6), 695–703, 2004.
www.atmos-chem-phys.org/acp/5/3139/
3151
Vogel, B., Müller, R., Deshler, T., Grooß, J.-U., Karhu, J.,
McKenna, D. S., Müller, M., Toohey, D., Toon, G. C., and
Stroh, F.: Vertical profiles of activated ClO and ozone loss in
the Arctic vortex in January and March 2000: In situ observations and model simulations, J. Geophys. Res., 108(D22), 8334,
doi:10.1029/2002JD002564, 2003.
Volk, C. M., Elkins, J. W., Fahey, D. W., Salawitch, R. J., Dutton,
G. S., Gilligan, J. M., Proffitt, M. H., Loewenstein, M., Podolske,
J. R., Minschwaner, K., Margitan, J. J., and Chan, K. R.: Quantifying transport between the tropical and mid-latitude lower
stratosphere, Science, 272, 1763–1768, 1996.
Waugh, D. W., Plumb, R. A., Elkins, J. W., Fahey, D. W., Boering, K. A., Dutton, G. S., Volk, C. M., Keim, E., Gao, R.-S.,
Daube, B. C., Wofsy, S. C., Loewenstein, M., Podolske, J. R.,
Chan, K. R., Proffitt, M. H., Kelly, K. K., Newman, P. A., and
Lait, L. R.: Mixing of polar vortex air into middle latitudes
as revealed by tracer-tracer scatterplots, J. Geophys. Res., 102,
13 119–13 134, 1997.
Webster, C. R., May, R. D., Trimble, C. A., Chave, R. G.,
and Kendall, J.: Aircraft laser infrared absorption spectrometer
(ALIAS) for in situ atmospheric measurements of HCl, N2 O,
CH4 , NO2 , and HNO3 , Appl. Opt., 33, 454–472, 1994.
Atmos. Chem. Phys., 5, 3139–3151, 2005

Download Report

Impact of mixing and chemical change on ozone

Paperzz.com

Your Paperzz