Ozone loss in the 2002-2003 Arctic vortex deduced from the assimilation of ODIN/SMR O3 and N2O measurements : N2O as a dynamical tracer

Abstract

In this paper we investigate the evolution of the northern polar vortex during the winter 2002–2003 in the lower stratosphere by using assimilated fields of ozone (O3) and nitrous oxide (N2O). Both O3 and N2O used in this study are obtained from the Sub-Millimetre Radiometer (SMR) aboard the Odin satellite and are assimilated into the global three-dimensional chemistry transport model of Météo-France, MOCAGE. O3 is assimilated into the ‘full’ model including both advection and chemistry whereas N2O is only assimilated with advection since it is characterized by good chemical stability in the lower stratosphere. We show the ability of the assimilated N2O field to localize the edge of the polar vortex. The results are compared to the use of the maximum gradient of modified potential vorticity as a vortex edge criterion. The O3 assimilated field serves to evaluate the ozone evolution and to deduce the ozone depletion inside the vortex. The chemical ozone loss is estimated using the vortex-average technique. The N2O assimilated field is also used to substract out the effect of subsidence in order to extract the actual chemical ozone loss. Results show that the chemical ozone loss is 1.1 ± 0.3 ppmv on the 25 ppbv N2O level between mid-November and mid-January, and 0.9 ± 0.2 ppmv on the 50 ppbv N2O level between mid-November and the end of January. A linear fit over the same periods gives a chemical ozone loss rate of ∼18 ppbv day−1 and ∼9.3 ppbv day−1 on the 25 ppbv and 50 ppbv N2O levels, respectively. The vortex-averaged ozone loss profile from the O3 assimilated field shows a maximum of 0.98 ppmv at 475 K. Comparisons to other results reported by different authors using different techniques and different observations give satisfactory results. Copyright © 2008 Royal Meteorological Society

Full text

QUARTERLY JOURNAL OF THE ROYAL METEOROLOGICAL SOCIETY
Q. J. R. Meteorol. Soc. 134: 217– 228 (2008)
Published online 8 January 2008 in Wiley InterScience
(www.interscience.wiley.com) DOI: 10.1002/qj.191
Ozone loss in the 2002–2003 Arctic vortex deduced from the
assimilation of Odin/SMR O3and N2O measurements: N2O
as a dynamical tracer
L. El Amraoui,a* V.-H. Peuch,aP. Ricaud,bS. Massart,cN. Semane,a,dH. Teyss`
edre,a
D. Cariollecand F. Karchera
aCNRM-GAME, M´et´eo-France and CNRS URA 1357, Toulouse, France
bUniversit´e de Toulouse, Laboratoire d’A´erologie, CNRS UMR 5560, Toulouse, France
cCERFACS, Toulouse, France
dCentre Nationale de Recherches M´et´eorologiques - DMN, Casablanca, Morocco
ABSTRACT: In this paper we investigate the evolution of the northern polar vortex during the winter 2002–2003 in the
lower stratosphere by using assimilated fields of ozone (O3) and nitrous oxide (N2O). Both O3and N2O used in this study
are obtained from the Sub-Millimetre Radiometer (SMR) aboard the Odin satellite and are assimilated into the global
three-dimensional chemistry transport model of M´
et´
eo-France, MOCAGE. O3is assimilated into the ‘full’ model including
both advection and chemistry whereas N2O is only assimilated with advection since it is characterized by good chemical
stability in the lower stratosphere. We show the ability of the assimilated N2O field to localize the edge of the polar vortex.
The results are compared to the use of the maximum gradient of modified potential vorticity as a vortex edge criterion.
The O3assimilated field serves to evaluate the ozone evolution and to deduce the ozone depletion inside the vortex. The
chemical ozone loss is estimated using the vortex-average technique. The N2O assimilated field is also used to substract
out the effect of subsidence in order to extract the actual chemical ozone loss. Results show that the chemical ozone loss is
1.1 ±0.3 ppmv on the 25 ppbv N2O level between mid-November and mid-January, and 0.9 ±0.2 ppmv on the 50 ppbv
N2O level between mid-November and the end of January. A linear fit over the same periods gives a chemical ozone loss
rate of ∼18 ppbv day−1and ∼9.3 ppbv day−1on the 25 ppbv and 50 ppbv N2O levels, respectively. The vortex-averaged
ozone loss profile from the O3assimilated field shows a maximum of 0.98 ppmv at 475 K. Comparisons to other results
reported by different authors using different techniques and different observations give satisfactory results. Copyright 
2008 Royal Meteorological Society
KEY WORDS chemical data assimilation; diabatic descent; chemical ozone loss
Received 14 September 2006; Revised 31 October 2007; Accepted 2 November 2007
1. Introduction
The evolution of ozone in the Arctic vortex depends
on both dynamical and chemical processes. The under-
standing of the ozone loss processes in the Northern
Hemisphere (NH) stratospheric vortex is essential for a
reliable prediction of the future evolution of the polar
ozone layer (Streibel et al., 2006). The determination of
the chemical ozone loss over the Arctic in the strato-
sphere is highly difficult due to the dynamical variability
caused by vertical and horizontal transport of air masses.
Different approaches within many sets of observations
have been proposed in order to remove the contribution
of transport and thus to quantify the ozone loss in the
winter stratospheric vortex. They have been described
and compared by Harris et al. (2002). The ‘tracer cor-
relation technique’ consists of removing the effect of
* Correspondence to: L. El Amraoui, M´
et´
eo-France (CNRM/GMGEC/
CATS), 42 Avenue G. Coriolis, 31057 Toulouse Cedex 01, France.
E-mail: elamraoui@cnrm.meteo.fr
transport by comparing the pre-winter and post-winter
relations between ozone volume mixing ratios and an
inert tracer (Singleton et al., 2005). Proffitt et al. (1990)
was the first to use in situ high-altitude aircraft mea-
surements to deduce the ozone loss in the Arctic polar
vortex by using this technique. The ‘vortex-average tech-
nique’ calculates the average of all vortex ozone data on
an isentropic surface as a function of time (Grooss and
M¨
uller, 2003). It assumes that the dynamical contribu-
tion to ozone inside the vortex is dominated by diabatic
descent. The ‘Match’ method is a pseudo-Lagrangian
technique, which permits us to quantify the chemical
ozone loss by measuring the difference in ozone in an air
parcel sampled at different times (e.g. von der Gathen
et al., 1995; Rex et al., 2002; Harris et al., 2002). The
‘passive tracer method’ consists of comparing modelled
ozone (run with and without chemistry) to observations. It
has been widely used in order to quantify the Arctic ozone
loss (Goutail et al., 1999; Sinnhuber et al., 2000; Hop-
pel et al., 2002). Although this method is probably the
Copyright 2008 Royal Meteorological Society

218 L. EL AMRAOUI ET AL.
most popular technique used to determine the ozone loss
in the polar vortex, it can present some defects. Indeed,
limitations arise from uncertainties in dynamical forc-
ings and in resolved-scale representation of subgrid-scale
transport and of homogenous and heterogenous chem-
istry, as well as in the vertical and horizontal resolution
of the chemistry and transport model (CTM). Previous
studies also indicate that current CTMs cannot give a
satisfactory observed partial column ozone loss (Feng
et al., 2005). Many models underestimate ozone loss in
cold winters when compared to ozone loss inferred from
observations (e.g. Goutail et al., 2005).
Data assimilation consists of combining in an opti-
mal way observations provided by instruments with an
apriori knowledge about a physical system such as a
model output. It has the advantage that observational
and model errors are accounted for and can be verified
a posteriori by considering e.g. observation minus fore-
cast (OMF) statistics or a χ2test (e.g. El Amraoui et al.,
2004). Thus, it is a powerful tool for improving model
outputs by constraining the apriori knowledge of the
model with the observations in order to better estimate
the state of the atmosphere.
Data assimilation has proven to be useful in vari-
ous fields of applications and was successfully applied
for studies of atmospheric chemistry and trace gas
distribution. Siegmund et al. (2005) demonstrated the
performance of the assimilation to study the ozone bud-
gets related to transport and chemistry in the Antarc-
tic region during the stratospheric warming of 2002. A
notable example of the power of data assimilation is
the successful prediction of the split Antarctic ozone
hole in September 2002 by Eskes et al. (2003) using
a tracer transport and assimilation model. Geer et al.
(2006) have assessed the assimilation of ozone data using
various technique and models in the Assimilation of
ENVISAT Data (ASSET) ozone intercomparison project.
They validate their analysis by comparing them with
independent data from ozonesondes and HALOE (the
Halogen Occultation Experiment on the Upper Atmo-
sphere Research Satellite), and also with the assimilated
fields of observations from the Michelson Interferome-
ter for Passive Atmospheric Sounding (MIPAS) aboard
the ENVISAT satellite. Recently, Bencherif et al. (2007)
demonstrated the ability of Odin/SMR N2O assimilated
fields to describe the tropic–midlatitude exchanges dur-
ing the 2002 major warming in the Southern Hemisphere.
This paper aims to:
•show the capability of assimilated fields to describe the
large-scale evolution of the NH stratospheric vortex
during the 2002–2003 winter, and
•evaluate the vortex-average technique with N2Oasa
dynamical tracer.
In this paper, we assimilate O3and N2Omeasure-
ments from the Odin/SMR instrument in order to describe
the evolution of the NH stratospheric vortex during
2002–2003 Arctic winter. The N2O assimilated fields are
used as a dynamical tracer in order to determine the edge
of the vortex using their maximum gradient following
the method suggested by Nash et al. (1996). They are
also used to quantify the diabatic descent inside the vor-
tex. The O3assimilated fields serve to evaluate the ozone
evolution as well as the chemical ozone loss inside the
vortex.
The outline is as follows. Section 2 presents the
measurements of O3and N2O as well as the assimilation
system used in this study. In Section 3, we present the
meteorological situation of the 2002–2003 Arctic vortex.
Main results including the validation of assimilated fields,
the diabatic descent, the ozone evolution as well as the
chemical ozone depletion inside the vortex are presented
in Section 4. Some comparisons with other results are
done in Section 5. Conclusions are given in Section 6.
2. Data analysis
2.1. O3and N2O measurements
The Odin satellite was launched on 20 February 2001 into
a polar sun-synchronous circular orbit with the ascending
node at 18 hours and an inclination of about 97.8°.The
satellite makes about 14 orbits per day and provides
a global coverage in the sense that all longitudinal
ranges are sampled. Each orbit contains about 60 limb
profiles of each measured species. Measurements are
made in the plane of the orbit covering the latitude range
from 83 °Sto83
°N. Odin has the unique mission to
make both astronomy and aeronomy measurements. It
includes two instruments: the Optical Spectrograph and
Infrared Imager System (OSIRIS) and the Sub-Millimetre
Radiometer (SMR). For the aeronomy part of the Odin
mission, SMR in the main stratospheric mode is able
to scan the limb of the atmosphere from 7 to 70 km
with a vertical resolution of about 1.5 km. Both O3and
N2O used in this study are retrieved in the 501.8 GHz
band. Typically N2O is retrieved in the stratosphere above
15 km with a single-scan precision of the order of 5%
(10–20 ppbv), and O3between ∼19 and ∼50 km with
a single-scan precision of about 25% (0.5–1.5 ppmv)
(Urban et al., 2005).
2.2. Assimilation system
The assimilation system used in this study is MOCAGE–
PALM developed jointly between M´
et´
eo-France and
CERFACS (Centre Europ´
een de Recherche et de Forma-
tion Avanc´
ee en Calcul Scientifique) in the framework of
the ASSET European project (Geer et al., 2006).
MOCAGE (Mod`
eledeChimieAtmosph
´
erique `
a
Grande Echelle; Peuch et al., 1999) is a three-dimensional
CTM of the troposphere and stratosphere forced by
external wind and temperature fields from the ARPEGE
model, the operational meteorological model of M´
et´
eo-
France (Courtier et al., 1991). The MOCAGE horizon-
tal resolution is 2°both in latitude and longitude and
the transport scheme is semi-Lagrangian. It includes 47
Copyright 2008 Royal Meteorological Society Q. J. R. Meteorol. Soc. 134: 217– 228 (2008)
DOI: 10.1002/qj

OZONE LOSS IN THE 2002–2003 ARCTIC VORTEX 219
hybrid (σ,P) levels from the surface up to 5 hPa (Cathala
et al., 2003), where σ=P/P
s;Pand Psare the pres-
sure and the surface pressure, respectively. MOCAGE
has a vertical resolution of about 800 m in the vicinity
of the tropopause and in the lower stratosphere. It has
the flexibility that it can be used for stratospheric as well
as for tropospheric studies, including detailed chemical
schemes (e.g. Dufour et al., 2004; Pradier et al., 2006).
The assimilation module is PALM (Projet d’Assimil-
ation par Logiciel Multim´
ethode), a modular and flexible
software developed at CERFACS, which consists of
elementary components that exchange data (Lagarde
et al., 2001). It manages the dynamic launching of the
coupled components (forecast model, algebra operators
and input/output of observational data) and the parallel
data exchanges.
The technique implemented within PALM and used for
the assimilation of O3and N2O profiles from Odin/SMR
is the 3D-FGAT method (First Guess at Appropriate
Time). This method is a cheap compromise between
the well-known 3D-Var and 4D-Var techniques (Fisher
and Andersson, 2001). It compares the observation and
background at the correct time and assumes that the
increment to be added to the background state is constant
over all the assimilation window, instead of propagating
it with a linear model. The choice of this assimilation
technique limits the size of the assimilation window,
since it has to be short enough compared to chemistry
and transport time-scales. Using ozone profiles from the
MIPAS instrument, this technique has produced good-
quality results compared to independent data and many
other assimilation systems (Geer et al., 2006).
3. 2002–2003 Arctic vortex
During the Arctic winter, the vortex is often affected by
stratospheric sudden warmings associated with planetary-
scale wave perturbations that originate in the troposphere
(Harris et al., 2002). The Arctic vortex is less stable than
its Antarctic counterpart and is often displaced off the
pole. The area and the position of the Arctic vortex
vary enormously from year to year. At temperatures
below 195 K, heterogenous reactions can occur on the
surface of polar stratospheric clouds (PSCs). Figure 1
shows the minimum temperatures north of 40 °N (from
ARPEGE analyses) at four different potential temperature
levels from 475 K to 625 K corresponding to the lower
stratosphere from November 2002 to March 2003. The
2002–2003 Arctic winter was dominated by a very
cold vortex in December and during the first half of
January. ARPEGE analyses show an exceptionally low
minimum temperature especially from early December to
mid-January in the lower stratosphere. The temperature
during this period is below the formation temperature
of PSCs. This is in agrement with other results using
ECMWF analysis (Raffalski et al., 2005; Streibel et al.,
2006), Met Office analysis (Singleton et al., 2005) and
lidar PSC observations (EORCU, 2003). By the end of
December, a minor warming developed in the upper-
middle stratosphere. This had practically no effect on the
lower-stratospheric polar vortex since it remained strong
and stable (EORCU, 2003). In mid-January, a major
warming appeared perturbing the vortex strongly. Around
20 January the vortex split into two small vortices. Two
other minor warmings appeared around the middle of
February and the middle of March. A final warming
began at the end of March and the vortex broke down
by mid-April. Our analysis indicate that the Arctic polar
vortex formed in November 2002 at 625 K, 575 K and
525 K, but at 475 K the vortex was established by
5 December 2002.
4. Results
4.1. Validation of assimilated fields
To assimilate ozone in the model, we have used the
linear ozone parametrization developed by Cariolle and
Tey s s `
edre (2007). The scheme is implemented with
activation of the ozone destruction due to heterogeneous
chemistry with the simplest formulation that uses only
local temperature. On the other hand, N2O is assimilated
in the same model but with advection only (i.e. all
chemical reactions are switched off).
The initialization field for both the chemically inte-
grated and passive fields is obtained by one month of
ozone assimilation. Thus, for the 3 November initializa-
tion date, ozone was assimilated between 3 October and
3 November. Separately, the same exercise was done for
N2O within the same period. O3is assimilated between
∼60 and ∼10 hPa (∼19–28 km), whereas N2O is assim-
ilated in the pressure range ∼100–10 hPa (∼16–28 km).
At this stage, it is essential to evaluate the consistency
of the assimilated fields. We then give a brief overview
Figure 1. Minimum temperatures in the Northern Hemisphere (north
of 40 °N) deduced from ARPEGE analyses from November 2002 to
March 2003 at 475, 500, 550 and 625 K potential temperature levels
(corresponding to the lower and the middle stratosphere).
Copyright 2008 Royal Meteorological Society Q. J. R. Meteorol. Soc. 134: 217– 228 (2008)
DOI: 10.1002/qj

220 L. EL AMRAOUI ET AL.
about the validation of assimilated measurements from
Odin/SMR within MOCAGE–PALM. The comparison
of the total ozone derived from O3assimilated field and
the Total Ozone Mapping Spectrometer (TOMS) shows
a difference of 5– 10% at high latitudes in the NH. The
comparison with ozonesondes in the NH in terms of ver-
tical profiles indicates that the difference is below 2%
for the whole pressure range (60–10 hPa). The devia-
tion between O3analysis and HALOE is less than 12%
between 60 and 10 hPa. The comparison with MIPAS
shows that the largest discrepancies are found at low
latitudes around 15 hPa where analysis overestimates
the MIPAS ozone by 10%. In the NH, comparison of
N2O assimilated field with data obtained from MIPAS
yields a good agreement within 5% for the altitude range
100–10 hPa. For more details about the validation of
assimilated fields from Odin/SMR, the reader is referred
to Massart et al. (2007).
In conclusion, the assimilated fields in the lower
stratosphere generally have a precision of about 12% and
5% for O3and N2O, respectively. Note that all the above
differences are expressed in terms of standard deviation.
Figure 2 presents the evolution of ozonesonde mea-
surements at Ny- ˚
Alesund (79 °N, 12 °E) at the 57.2 hPa
level (∼19.7 km) compared to the O3assimilated field,
the modelled field with the ‘full’ model and the ozone
passive tracer. The comparison is made between the
beginning of November 2002 and the end of March
2003. The ‘full’ model (including advection +linear
ozone chemistry) and modelled passive ozone tracer are
in reasonable agreement with ozonesonde observations
from the beginning of November to practically mid-
December 2002. From mid-December 2002, the agree-
ment between the chemistry/no chemistry runs disappears
due to the chemical depletion of ozone. In comparison
to ozonesonde observations, the model with chemistry
behaves relatively well, but sometimes overestimates the
ozone at this altitude. However, the use of the assimila-
tion of Odin/SMR ozone in the same model gives satis-
factory results; the maximum absolute difference between
the two fields does not exceed 0.7 ppmv over the whole
period, except on one day (Julian day =52) where the
difference reaches 1.6 ppmv. The correlation between the
two fields is about 0.82. The ozone assimilated field and
the independent ozonesonde observations are generally
in good agreement over the whole assimilation period
(November 2002–March 2003).
4.2. Assimilation versus model
In this paragraph, we quantify the added value of the
assimilation process on the transport as well as on the
chemistry modelled fields. A correlation plot of the
averaged values inside the vortex of modelled N2O
versus assimilated N2O at 625 K potential temperature
level is shown in Figure 3(a). A linear regression slope
of 0.90 is observed, which is very satisfactory. Since
N2O is assimilated without chemistry, this significant
agreement suggests that the transport is well modelled
within MOCAGE.
In order to have an idea about how the assimilation
improves the chemistry parametrization used in the
model, we present in Figure 3(b) the time evolution of
the averaged values inside the vortex of O3(modelled
and assimilated) at the same isentropic level 625 K. From
early November to mid-January, the difference between
both fields is small (0.2– 0.45 ppmv). From mid-January
until the end of March, the difference is much more
Figure 2. Ozonesonde observations (open circles) at Ny- ˚
Alesund (79 °N, 12 °E) at the 57.2 hPa level compared to MOCAGE ozone passive
tracer (dash-dotted line), the modelled field with the ‘full’ model (thin solid line) and the Odin/SMR ozone assimilated fields (bold solid line).
Copyright 2008 Royal Meteorological Society Q. J. R. Meteorol. Soc. 134: 217– 228 (2008)
DOI: 10.1002/qj

OZONE LOSS IN THE 2002–2003 ARCTIC VORTEX 221
Figure 3. (a) Correlation plot of the averaged values inside the vortex
of modelled N2O versus assimilated N2O at the 625 K potential
temperature level. The linear regression slope is 0.90. (b) Time
evolution of the averaged values inside the vortex of O3(modelled and
assimilated) at the 625 K potential temperature level. In this study, the
model uses the linear ozone parametrization of Cariolle and Teyss`
edre
(2007) to estimate the O3evolution (see text for details).
important. Note that the temperature inside the vortex
during this period at this potential temperature level is
above 195 K (Figure 1). The ozone destruction inside
the vortex due to PSC chemistry in the linear ozone
parametrization is based on the local temperature. From
mid-January, the model is thus unable to destroy ozone
through the heterogeneous reactions. This explains the
rapid increase of ozone inside the vortex from mid-
January in the modelled field, which does not appear in
the assimilated field. This example shows that the used
linear parametrization of ozone is unable to reproduce the
evolution of ozone inside the vortex during the whole
period (November 2002–March 2003). In conclusion,
the assimilation of Odin data within MOCAGE brings
a much more important correction to the chemistry part
of the model than to its transport part.
4.3. Vortex edge definition
The vortex edge is the term commonly used to refer
to the location of the barrier between the air masses
inside and outside the vortex (M¨
uller and G¨
unther,
2003). The determination of the ozone loss inside the
vortex with the vortex-average method requires that the
edge of the vortex is well localized. Many authors
define the vortex edge by using the potential vorticity
(PV) field according to Nash criterion (e.g. Rex et al.,
1999). However, Greenblatt et al. (2002) have shown that
small-scale vortex edge features might not be properly
represented in analysis of PV directly derived from
meteorological fields. Moreover, the degree of isolation
of the polar vortex as defined by the PV field is not
well known (Jost et al., 2002). The PV field is poorly
conserved in small-scale filaments due to radiative and
dynamical processes (Hauchecorne et al., 2002). Allen
and Nakamura (2003) also argue that the tracer-based
method has several advantages over the PV method in
determining the location of the vortex; in particular, PV
is not conserved isentropically beyond the time-scale in
which diabatic effects are negligible.
Another parameter used to diagnose the polar vortex
is the modified PV (MPV) introduced by Lait (1994) in
order to remove the altitude dependence of PV without
destroying its structure on a given isentropic surface. This
parameter has been recently used for polar vortex studies
(e.g. Christensen et al., 2005). It is defined as a function
of PV and potential temperature θas
MPV =PV θ
475 −9
2.
In this study we determine the vortex edge by using
the N2O assimilated field as a dynamical tracer since this
molecule is chemically inert in the lower stratosphere.
Moreover, the sharp gradient in N2O at the vortex edge
is likewise noticeable in satellite measurements (e.g.
Manney et al., 1999). In order to evaluate this method,
the results are compared to the use of the maximum
gradient of the MPV field as a vortex edge criterion.
Figure 4 (left-hand side) shows maps of the assimilated
N2O field at different potential temperature levels on
selected days. The same figure (right-hand side) gives
the corresponding MPV fields. In both cases, the edge
of the vortex determined by the maximum gradient
is denoted by a solid black contour. The comparison
between the two fields shows that in the NH the minimum
values of N2O correspond to the maximum values of
MPV. Both fields also show almost the same structure
during December and January at 475 K and 525 K
when the vortex was strong (first and second row).
By mid-February, the vortex was perturbed and began
to split, this being more marked at 575 K and 625 K
(third row). In this case the comparison between the
two fields shows that the MPV field is smooth, while
N2O assimilated fields show more detailed structures and
filaments. The equivalent latitude corresponding to the
maximum gradient of both fields during this period shows
some differences (e.g. Table I for 20 January 2003 for
575 K and 625 K). With the N2O method, we clearly
see midlatitude intrusions of air masses in the inner
vortex (Figure 4, third row). Qualitatively there is good
agreement between the two methods. However, when
calculating the equivalent latitude corresponding to their
maximum gradient (Table I), we note some differences
Copyright 2008 Royal Meteorological Society Q. J. R. Meteorol. Soc. 134: 217– 228 (2008)
DOI: 10.1002/qj

222 L. EL AMRAOUI ET AL.
Table I. Equivalent latitude (in degrees) corresponding to the maximum gradient of N2O and MPV fields at different potential
temperature levels for specific days.
475 K 525 K 575 K 625 K
N2OMPVN
2OMPVN
2OMPVN
2OMPV
14 December 2002 63.0 63.0 65.0 63.8 63.067.761.069.8
30 December 2002 64.0 65.0 63.0 63.4 63.170.858.262.5
20 January 2003 68.5 68.6 65.2 66.4 62.066.463.067.5
28 January 2003 67.7 68.1 67.0 67.3 67.2 68.3 65.1 66.0
10 February 2003 67.1 66.5 69.2 68.8 67.1 67.5 66.2 68.1
25 February 2003 69.0 69.1 67.7 69.4 65.069.665.369.6
5 March 2003 69.2 69.1 67.3 68.9 67.2 69.7 67.7 69.2
23 March 2003 69.0 70.2 69.0 69.7 69.4 69.8 69.0 70.2
Bold denotes that the values determined by the two methods differ by more than 4 degrees.
Figure 4. Odin/SMR assimilated N2O field (left-hand side) and the corresponding modified potential vorticity (MPV) field from ARPEGE
analyses (right-hand side) for potential temperature levels (475, 525, 575 and 625 K) for the selected days (24 December 2002, 12 January 2003,
20 January 2003 and 6 March 2003). The edge of the vortex determined by both methods is denoted by the solid black contour.
Copyright 2008 Royal Meteorological Society Q. J. R. Meteorol. Soc. 134: 217– 228 (2008)
DOI: 10.1002/qj

OZONE LOSS IN THE 2002–2003 ARCTIC VORTEX 223
especially at 575 K and 625 K during the minor warming
of December 2002, the major warming of January 2003
and the minor warming of February. An investigation of
the difference between the two methods in determining
the vortex edge is beyond the scope of this paper and is
the subject of ongoing work.
4.4. Diabatic descent
The evolution of stratospheric ozone in the winter Arctic
vortex is governed by both chemistry and transport.
Ozone variations due to transport are often of the same
magnitude as those due to chemical ozone destruction
(Tilmes et al., 2004). Subsidence always masks the
chemical ozone loss in the lower stratosphere. Descent
of air tends to increase the ozone mixing ratio at a
given altitude. Air with large mixing ratios of ozone is
transported downwards into the lower stratosphere, the
region where chemical ozone depletion occurs.
N2O has only tropospheric sources and is mainly
removed by photolysis at high altitudes. This results in a
lifetime of more than 1 year below 33 km altitude and of
100 years up to 22 km (WMO, 1986). Therefore N2O
is well suited to study the vertical motion of the air
masses inside the polar vortex. The evolution of N2O
inside the vortex is a combination of both diabatic descent
and mixing. Figure 5 shows the time evolution from 10
December 2002 to 20 March 2003 of analyzed N2O
mixing ratio averaged over ten-day intervals over the
vortex area at the 475, 525, 575 and 625 K isentropic
levels. The descent at 475 K and 525 K inside the vortex
was stronger, and extended over a longer period, than
the descent at 575 K and 625 K levels. Further, the N2O
values inside the vortex decreased rapidly in December,
and stabilized from mid-January to mid-February for
475 K and 525 K. The value of N2O inside the vortex
increased for the 575 K and 625 K levels during the same
period. This is associated with the mixing with air from
outside the vortex that occurred after the major warming
of mid-January as reported by Urban et al. (2004). N2O
values decreased during early March, and after that time
we see a final increase of N2O inside the vortex at all
the studied levels, which is in connection with the vortex
break-up.
Figure 6 shows mean potential temperature levels
corresponding to the vortex averages of the 25, 50, 75
and 100 ppbv levels of N2O. The descent of all these
levels is clearly seen from the beginning of November
until the beginning of February. The 25 ppbv N2O
level descended from a potential temperature level of
595 K to ∼525 K. The 50 ppbv level descended from
560 K to ∼495 K. During the beginning of February, the
potential temperature level of all N2O levels hardly varied
(∼530 K and ∼500 K for 25 and 50 ppbv N2O levels,
respectively). This indicates that the diabatic descent in
this period was weak. The 75 and 100 ppbv N2Olevels
descended below 500 K from the middle of December.
Figure 5. Time evolution of the assimilated N2O mixing ratios inside
the vortex averaged over ten-day periods at 475, 525, 575 and 625 K
for the whole assimilation period (November 2002–March 2003).
Figure 6. Diabatic descent of the averaged values of different N2O
levels (25, 50, 75 and 100 ppbv) for the whole assimilation period
(November 2002–March 2003). The 25 and 50 ppbv N2O level air
masses are above the 500 K potential temperature level for the whole
assimilation period, whereas the 75 and 100 ppbv N2O level air masses
descend below 500 K starting in mid-December 2002.
4.5. Evolution of ozone inside the vortex
In order to investigate the evolution of ozone inside
the polar vortex, the average values were calculated at
different potential temperature levels. Figure 7 represents
the evolution of the assimilated ozone averaged inside
the vortex from November 2002 to March 2003 at
475, 525, 575 and 625 K potential temperature levels.
The enhancement of the average ozone value at the
beginning of formation of the vortex can be attributed
to the diabatic subsidence, which brought ozone-rich air
masses from higher to lower levels. After a few days,
the ozone average value inside the vortex decreased
substantially. This is due to the chemical ozone depletion,
Copyright 2008 Royal Meteorological Society Q. J. R. Meteorol. Soc. 134: 217– 228 (2008)
DOI: 10.1002/qj

224 L. EL AMRAOUI ET AL.
Figure 7. Time evolution of the assimilated ozone averaged inside the vortex for different potential temperature levels (475, 525, 575 and 625 K)
from November 2002 to March 2003. The vortex edge is determined by the maximum gradient of the N2O assimilated field on each potential
temperature level.
which is more important than the diabatic subsidence.
The major ozone decrease inside the vortex is shown
in December 2002 and early January 2003. During
this period the vortex was somewhat centred off the
pole and moved towards lower latitudes (e.g. Figure 4,
upper three rows). Singleton et al. (2005) reported that
during December 2002 and January 2003 the vortex was
often elongated allowing in-vortex air to make frequent
excursions into the sunlight at lower latitudes. This
situation as well as the low temperatures allowed the early
advent of the ozone loss.
The evolution of O3inside the vortex in Figure 7
shows that on the 475 K level, ozone inside the vortex
decreased from mid-December until the beginning of
February when the assimilated ozone averaged inside the
vortex was a minimum. For this level, ozone inside the
vortex was reduced by 21.8 ±5% by the beginning
of February with respect to its average value at the
end of November. At 525 K, ozone was reduced by 1.0
±0.2 ppmv between the end of December and the middle
of January. At 575 K, at the beginning of formation
of the vortex by the middle of November, the average
value of ozone was 4.2 ±0.3 ppmv, and by the end of
December it was 3.0 ±0.3 ppmv. The ozone reduction on
this potential temperature level was 29 ±6%. At 625 K,
we find that ozone was reduced by ∼28 ±6% between
the beginning of November and the end of December.
The evolution of O3inside the vortex in Figure 7
includes both chemical loss and diabatic descent. In order
to extract the actual chemical ozone loss, we use the N2O
assimilated field to subtract out the effect of subsidence,
which masks the chemical ozone loss.
4.6. Chemical ozone loss
During winter, diabatic cooling in the vortex results in
subsidence. Diabatic descent must then be accounted
for in the calculation of chemical ozone loss. The time
evolution of N2O inside the vortex is used to remove the
contribution of subsidence. Figure 8 shows time series
of ozone inside the vortex at 25 ppbv and 50 ppbv N2O
levels. On the 25 ppbv N2O level, ozone was chemically
depleted by 28 ±6% (1.1 ±0.3 ppmv) by mid-January
with respect to its average value at the middle of
November. During this period, air masses on this level
descended from ∼581 K down to ∼531 K (Figure 6). On
the 50 ppbv level, the ozone was chemically depleted
by 26 ±5% (0.9 ±0.2 ppmv) between the middle of
November and the end of January. A linear fit over the
same periods as the chemical ozone loss gives a chemical
ozone loss rate of ∼18 ppbv day−1and ∼9.3 ppbv day−1
on the 25 ppbv and 50 ppbv N2O levels, respectively.
In order to deduce the chemical ozone loss profile,
the time evolution of N2O inside the vortex is used to
remove the contribution of subsidence using the same
Copyright 2008 Royal Meteorological Society Q. J. R. Meteorol. Soc. 134: 217– 228 (2008)
DOI: 10.1002/qj

OZONE LOSS IN THE 2002–2003 ARCTIC VORTEX 225
Figure 8. Time series of averaged ozone inside the vortex at 25 and
50 ppbv N2O levels from November 2002 to March 2003.
method as Rex et al. (2002). Figure 9 presents the vortex-
averaged ozone loss profile between mid-December and
early February as a function of potential temperature.
Theprofilestartsat475K(∼19 km); this altitude is
the lower limit of Odin/SMR O3measurements in the
retrieved band (501.8 GHz). The largest chemical ozone
loss deduced from the O3assimilated field is observed at
475 K at 0.98 ±0.2 ppmv.
Since the transport is well modelled within MOCAGE,
the modelled ozone without chemistry and the ozone
assimilated field are used through the passive tracer
method in order to infer the chemical ozone loss and
validate the results obtained with the vortex-average
technique. The advantage of the passive tracer method
is that no correction related to the diabatic descent is
required. The chemical ozone loss is directly inferred on
each selected isentropic level.
Figure 9. Estimated ozone loss mixing ratio between mid-December
and early February versus potential temperature after removing the
effects of diabatic descent. In the studied domain, the maximum ozone
loss of ∼1 ppmv is observed at 475 K.
Results indicate that the passive tracer loss is quite
similar to the vortex-average results above 500 K. Nev-
ertheless, the passive tracer gives slightly higher esti-
mates of the ozone loss at 475 (∼1.15 ppmv) and 500 K
(∼0.9 ppmv).
The objective of the next Section is to compare
our results with the findings from other independent
observations.
5. Comparison with other results
The ozone loss of the winter 2002–2003 was reported
by many authors using many datasets and a wide range
of methods.
Raffalski et al. (2005) used ozone measurements from
the millimetre wave radiometer at Kiruna, Sweden to
deduce the ozone decrease. They used N2Omeasure-
ments from the Odin/SMR instrument in order to remove
the effect of subsidence. They found that between mid-
December and mid-February, the chemical ozone loss
was 0.5 and 0.9 ppmv on the 25 ppbv and 50 ppbv levels,
respectively. In comparison to these results, we found that
between the middle of December and the middle of Jan-
uary, the chemical ozone loss was ∼0.7 and ∼0.9 ppmv
on the 25 and 50 ppbv N2O levels, respectively.
Using the passive tracer technique, Singleton et al.
(2005) used Polar Ozone and Aerosol Measurement
(POAM III) satellite observations and the SLIMCAT-
CTM to deduce the ozone loss. They found that the ozone
loss was approximately 0.7 ppmv between December and
mid-January at the 500 K isentropic level. Our results
indicate that on the same potential temperature level,
ozone is chemically depleted by 0.82 ppmv between mid-
December and early February.
Tilmes et al. (2003) reported a very early chlorine
activation and ozone loss from the analysis of the
HALOE data using the ozone-tracer correlation method.
They found a maximum local ozone loss of 1.5 ppmv
at ∼440 K between mid-December and February. Their
analysis shows that the ozone loss during the same
period was ∼1.1 and ∼0.8 ppmv at 475 K and 500 K,
respectively. Unfortunately, in the used band, Odin/SMR
data do not allow ozone loss calculation for lower
altitudes (below 475 K) due to the limited range of
O3measurements. Nevertheless, in comparison to these
results, we found that the chemical ozone loss obtained at
475 K and 500 K was 0.98 and 0.82 ppmv, respectively.
Christensen et al. (2005) have used the vortex-average
technique to infer the ozone loss in the Arctic vortex (as
defined by Nash et al., 1996) from ozone sonde data. We
find quite good agreement when we compare our results
to their results. In fact, they found that ozone loss is 1.0
±0.2 ppmv at 479 K, while we find 0.98 ±0.2 ppmv
at the 475 K potential temperature level.
Streibel et al. (2006) used the Match technique to
determine the chemical ozone loss inside the vortex dur-
ing winter 2002 –2003. Their results reveal that the ozone
destruction rates varied between 10 and 15 ppbv day−1.
Copyright 2008 Royal Meteorological Society Q. J. R. Meteorol. Soc. 134: 217– 228 (2008)
DOI: 10.1002/qj

226 L. EL AMRAOUI ET AL.
They also found that the accumulated ozone loss was 1.3
and 0.55 ppmv at the 475 K and 500 K potential temper-
ature levels, respectively. We found a chemical ozone loss
rates of the same magnitude (∼9.3 and ∼18 ppbv day−1
on the 50 ppbv and 25 ppbv levels, respectively). The
chemical ozone loss obtained by our analysis is 0.98 and
0.82 ppmv at 475 K and 500 K, respectively. The differ-
ences between our method and the Match technique may
be explained by the fact that the two methods do not
use the same vortex edge definition (Christensen et al.,
2005).
Ozone loss results inferred from the modelled ozone
without chemistry and the ozone assimilated field within
the passive tracer method are quite similar to the vortex-
average results above 500 K. However at 475 and 500 K,
results indicate that the ozone loss deduced by the passive
tracer is slightly higher. The differences between the two
methods is not the subject of this study. To understand
the discrepancies between the two methods, additional
investigations are required.
Generally, the comparison of our results regarding
the 2002–2003 Arctic vortex to those of other authors
using different datasets and different techniques generally
gives quite good agreement. However, Feng et al. (2005)
reported that the high-resolution model shows large
ozone loss in the lower stratosphere. The models still
fail to reproduce many aspects of polar chemistry and
transport (e.g. chemical ozone loss and diabatic descent).
Indeed, in this study, the model with the implemented
parametrization of ozone is unable to reproduce the
ozone evolution inside the vortex (Figures 2 and 3).
Consequently, it is unable to provide a better estimate of
the chemical ozone loss inside the vortex. In this case, the
use of chemical data assimilation can be a valuable tool
to overcome these possible deficiencies of the models.
In particular, chemical data assimilation is mainly used
here to correct the model heterogeneous ozone depletion
and reproduce a near-complete ozone destruction inside
the vortex. Consequently, the assimilated fields can better
describe the large-scale evolution of the NH stratospheric
vortex in comparison to the model alone.
6. Conclusions
We have used assimilated fields of O3and N2O from
Odin/SMR in order to describe the large-scale evolu-
tion of the polar stratospheric vortex during the winter
2002–2003.
The N2O assimilated fields were used to estimate the
location of the vortex edge using their maximum gradient
according to Nash et al. (1996). The results are compared
to the maximum gradient of modified potential vorticity
as a vortex edge criterion in terms of equivalent latitude.
Both methods agree well during the whole assimilation
period for all studied levels except at high levels (575 K
and 625 K) during the minor warming of December, the
major warming of January and after the minor warming
during February. A detailed comparison between the two
methods is not the subject of this paper. It is the subject
of an ongoing work.
The N2O assimilated fields were also used to remove
the contribution of subsidence. Results indicate that
diabatic descent was important from the beginning of
November to early February. In particular, during this
period, air masses on the 25 ppbv N2O level subside from
∼595 K down to ∼525 K. Those of 50 ppbv N2Olevel
descend from 560 K down to 495 K.
The assimilation of ozone enabled us to overcome
the imperfections of the MOCAGE model in connection
with the linear ozone parametrization used in this study.
Chemical data assimilation of ozone is mainly used
here to describe the ozone evolution and to correct the
model heterogeneous ozone depletion and reproduce a
near-complete ozone destruction in the vortex. Results
indicate that ozone is chemically depleted by 1.1 ±
0.2 ppmv on the 25 ppbv N2O level between the middle
of November and mid-January. On the 50 ppbv N2O
level, ozone is chemically depleted by 0.9 ±0.2 ppmv
between the middle of November and the end of January.
A linear fit over the same periods gives a chemical
ozone loss rate of ∼18 ppbv day−1and ∼9.3 ppbv day−1
on the 25 ppbv and 50 ppbv N2O levels, respectively.
The vortex-averaged ozone loss profile from the O3
assimilated field shows a maximum of 0.98 ppmv at
475 K. Additionally, the passive tracer method indicates
that the ozone loss results are quite similar to the vortex-
average results above 500 K. At 475 and 500 K, the
passive tracer method gives slightly higher values.
Comparisons of ozone loss at different isentropic levels
with the Match technique reveal that the Match technique
gives slightly higher estimates, particularly at 475 K.
This discrepancy may be explained by the fact that
the two methods do not use the same vortex edge
definition. Thus, great care should to be taken in order
to have a careful comparison with the Match technique
(Christensen et al., 2005).
Comparisons with other results from the winter
2002–2003 show quite good agreement with chemical
ozone loss inferred from HALOE data using the the
ozone-tracer correlation method (Tilmes et al., 2003).
Results derived from POAM III and the SLIMCAT-CTM
using the passive tracer method are slightly smaller than
our results, while we find very good agreement with the
results of Christensen et al. (2005) using ozonesonde data
within the vortex-average technique.
The comparison of our results to others using different
techniques and different measurements during the same
period shows generally good agreement. This demon-
strates the power of chemical data assimilation to over-
come the possible deficiencies of the model and to
describe the large-scale behaviour of the NH stratospheric
vortex. This is particularly interesting insofar as the spe-
cific objective of chemical data assimilation is to produce
a self-consistent picture of the atmosphere taking into
account both the available observations and our theoret-
ical understanding of the chemical and transport system.
Copyright 2008 Royal Meteorological Society Q. J. R. Meteorol. Soc. 134: 217– 228 (2008)
DOI: 10.1002/qj

OZONE LOSS IN THE 2002–2003 ARCTIC VORTEX 227
Acknowledgements
Odin is a Swedish-led satellite project funded jointly by
Sweden, Canada, Finland and France. This work was
funded in France by contracts from the Centre National
de Recherches M´
et´
eorologiques (CNRM) and the Centre
National de Recherches Scientifiques (CNRS).
References
Allen DR, Nakamura N. 2003. Tracer equivalent latitude: A diagnostic
tool for isentropic transport studies. J. Atmos. Sci. 60: 287–304.
Bencherif H, El Amraoui L, Semane N, Massart S, Charyulu DV,
Hauchecorne A, Peuch V-H. 2007. Examination of the 2002
major warming in the southern hemisphere using ground-based and
Odin/SMR assimilated data: Stratospheric ozone distributions and
tropic/mid-latitude exchange. Odin Special Issue in Canad. J. Phys.
85: 1287– 1300. DOI: 10.1139/P07-143.
Cariolle D, Teyss`
edre H. 2007. A revised linear ozone photochemistry
parameterization for use in transport and general circulation models.
Multi-annual simulations. Atmos. Chem. Phys. 7: 2183 –2196.
Cathala M-L, Pailleux J, Peuch V-H. 2003. Improving chemical
simulations of the upper troposphere –lower stratosphere with
sequential assimilation of MOZAIC data. Tellus 55B: 1– 10.
Christensen T, Knudsen BM, Streibel M, Andersen SB, Benesova A,
Braathen G, Claude H, Davies J, De Backer H, Dier H, Dorokhov V,
Gerding M, Gil M, Henchoz B, Kelder H, Kivi R, Kyr ¨
o E, Litynska
Z, Moore D, Peters G, Skrivankova P, St ¨
ubi R, Turunen T, Vaughan
G, Viatte P, Vik AF, von der Gathen P, Zaiteev I. 2005. Vortex-
averaged Arctic ozone depletion in the winter 2002/2003. Atmos.
Chem. Phys. 5: 131 –138.
Courtier P, Freydier C, Geleyn J-F, Rabier F, Rochas M. 1991. The
ARPEGE project at M´
et´
eo– France. In: Workshop on numerical
methods in atmospheric models, 2: 193 –231.
Dufour A, Amodei M, Ancellet G, Peuch V-H. 2004. Observed and
modelled ‘chemical weather’ during ESCOMPTE. Atmos. Res. 74:
161– 189.
El Amraoui L, Ricaud P, Urban J, Th´
eodore B, Hauchecorne A, Lauti´
e
N, de La No¨
e J, Guirlet M, Le Flochm¨
en E, Murtagh D, Dupuy E,
Frisk U, d’Andon OF. 2004. Assimilation of Odin/SMR O3and N2O
measurements in a three-dimensional chemistry transport model. J.
Geophys. Res. 109: D22304, DOI: 10.1029/2004JD004796.
EORCU. 2003. ‘The northern hemisphere stratosphere in the
2002/03 winter: Preliminary results from the first phase of
VINTERSOL’. Tech. report, European Ozone Research Coordinating
Unit, Department of Chemistry, University of Cambridge.
(http://www.ozone-sec.ch.cam.ac.uk/EORCU/Reports/wr0203.pdf).
Eskes HJ, Van Velthoven PFJ, Valks PJM, Kelder HM. 2003.
Assimilation of GOME total-ozone satellite observations in a three-
dimensional tracer-transport model. Q. J. R. Meteorol. Soc. 129:
1663– 1681.
Feng W, Chipperfield MP, Davies S, Sen B, Toon G, Blavier JF,
Webster CR, Volk CM, Ulanovsky A, Ravegnani F, von der Gathen
P, Jost H, Richard EC, Claude H. 2005. Three-dimensional model
study of the arctic ozone loss in 2002/2003 and comparison with
1999/2000 and 2003/2004. Atmos. Chem. Phys. 5: 139 –152.
Fisher M, Andersson E. 2001. ‘Developments in 4D-Var and Kalman
filtering’. In: Technical Memorandum 347, ECMWF: Reading, UK.
Geer AJ, Lahoz WA, Bekki S, Bormann N, Errera Q, Eskes HJ, Fonteyn
D, Jackson DR, Juckes MN, Massart S, Peuch V-H, Rharmili S,
Segers A. 2006. The ASSET intercomparison of ozone analyses :
method and first results. Atmos. Chem. Phys. 6: 5445 –5474.
Goutail F, Pommereau J-P, Phillips C, Deniel C, Sarkissian A, Lef`
evre
F, Kyr¨
o E, Rummukainen M, Ericksen A, Andersen SB, K˚
astad
Høiskar B-A, Braathen G, Dorokhov V, Khattatov VU. 1999.
Depletion of column ozone in the Arctic during the winters of
1993– 1994 and 1994–1995. J. Atmos. Chem. 32: 1 –34.
Goutail F, Pommereau J-P, Lef`
evre F, Van Roozendael M, Andersen
SB, K˚
astad Høiskar B-A, Dorokhov V, Kyr ¨
o E, Chipperfield MP,
Feng W. 2005. Early unusual ozone loss during the Arctic winter
2002/2003 compared to other winters. Atmos. Chem. Phys. 5:
665– 677.
Greenblatt JB, Jost H-J, Lowenstein M, Podolske JR, Hurst DF, Elkins
JW, Schauffler SM, Atlas EL, Hermann RL, Webster CR, Bui TP,
Moore FL, Ray EA, Oltmans S, V¨
omel S, Blavier J-F, Sen B,
Stachnik RA, Toon GC, Engel A, M¨
uller M, Schmidt U, Bremer H,
Pierce RB, Sinnhuber B-M, Chipperfield M, Lef`
evre F. 2002. Tracer-
based determination of vortex descent in the 1999/2000 arctic winter.
J. Geophys. Res. 107: (D20), 8297, DOI: 10.1029/2002JD000937.
Grooss J-U, M¨
uller R. 2003. The impact of midlatitude intrusions into
the polar vortex on ozone loss estimates. Atmos. Chem. Phys. 3:
395– 402.
Hauchecorne A, Godin S, Marchand M, Heese B, Souprayen C. 2002.
Quantification of the transport of chemical constituents from the
polar vortex to middle latitudes in the lower stratosphere using high-
resolution advection model MIMOSA and effective diffusivity. J.
Geophys. Res. 107: (D20), 8289, DOI: 10.1029/2001JD000491.
Harris NRP, Rex M, Goutail F, Knudsen BM, Manney GL, M ¨
uller
R, von der Gathen P. 2002. Comparison of empirically derived
ozone losses in the Arctic vortex. J. Geophys. Res. 107: (D20), DOI:
10.1029/2001JD000482.
Hoppel K, Bevilacqua R, Nedoluha G, Deniel C, Lef`
evre F, Lumpe
J, Fromm M, Randall C, Rosefield J, Rex M. 2002. POAM III
observations of arctic ozone loss for the 1999/2000 winter. J.
Geophys. Res. 107: (D20), DOI: 10.1029/2001JD000482.
Jost H-J, Loewenstein M, Greenblatt JB, Podolske JR, Bui TP,
Hurst DF, Elkins JW, Hermann RL, Webster CR, Schauffler SM,
Atlas EL, Newman PA, Lait LR, Wofsy SC. 2002. Mixing events
revealed by anomalous tracer relationships in the arctic vortex
during winter 1999/2000. J. Geophys. Res. 107: (D24), 4795, DOI:
10.1029/2002JD002380.
Lagarde T, Piacentini A, Thual O. 2001. A new representation of data-
assimilation methods: The PALM flow-charting approach. Q. J. R.
Meteorol. Soc. 127: 189– 207.
Lait LR. 1994. An alternative form of potential vorticity. J. Atmos. Sci.
51: 1754– 1759.
Manney GL, Michelsen H, Santee ML, Gunson M, Irion F, Roche
A, Livesey N. 1999. Polar vortex dynamics during spring and
fall diagnosed using trace gas observations from the Atmospheric
Trace Molecule Spectroscopy instrument. J. Geophys. Res. 104:
18841– 18866.
Massart S, Piacentini A, Cariolle D, El Amraoui L, Semane N. 2007.
Assessment of the quality of the ozone measurements from the
Odin/SMR instrument using model assimilation. Odin Special Issue
in Canad. J. Phys. 85: 1209 –1223, DOI: 10.1139/P07-143.
M¨
uller R, G¨
unther G. 2003. A generalized form of Lait’s modified
potential vorticity. J. Atmos. Sci. 60: 2229 –2237.
Nash ER, Newman PA, Rosenfield JE, Schoeberl MR. 1996. An
objective determination of the polar vortex using Ertel’s potential
vorticity. J. Geophys. Res. 101: 9471 –9478.
Peuch V-H, Amodei M, Barthet T, Cathala ML, Josse B, Michou
M, Simon P. 1999. MOCAGE, MOd`
ele de Chimie Atmosph´
erique
`
a Grande Echelle. Pp. 33 –36 in Proceedings of Workshop on
Atmospheric Modelling, M´
et´
eo– France: Toulouse.
Pradier S, Atti´
e J-L, Chong M, Escobar J, Peuch V-H, Lamarque J-
F, Khattatov B, Edwards D. 2006. Evaluation of 2001 springtime
CO transport over West Africa using MOPITT CO measurements
assimilated in a global chemistry transport model. Tellus 58:
163– 176.
Proffit M, Margitan J, Kelly K, Loewenstein M, Podolske J, Chan
K. 1990. Ozone loss in the Arctic polar vortex inferred from high-
altitude aircraft measurements. Nature 347: 31– 36.
Raffalski U, Hochschild G, Kopp G, Urban J. 2005. Evolution of
stratospheric ozone during winter 2002/2003 as observed by a
ground-based millimetre wave radiometer at Kiruna, Sweden. Atmos.
Chem. Phys. 5: 1399 –1407.
Rex M, von Der Gathen P, Braathen GO, Harris NRP, Reimer E, Beck
A, Alfier R, Kr¨
uger-Carstensen R, Chipperfield M, De Backer H,
Balis D, O’Connor F, Dier H, Dorokhov V, Fast H, Gamma A, Gil
M, Kyr¨
o E, Litynska Z, Mikkelsen IS, Molyneux M, Murphy G,
Reid SJ, Rummukainen M, Zerefos C. 1999. Chemical ozone loss
in the Arctic winter 1994/95 as determined by MATCH technique.
J. Atmos. Chem. 32: 35 –59.
Rex M, Salawitch RJ, Harris NRP, von Der Gathen P, Braathen
GO, Schulz A, Deckelmann H, Chipperfield M, Sinnhuber B-M,
Reimer E, Alfier R, Bevilacqua R, Hoppel K, Fromm M, Lumpe
J, K¨
ullmann H, Kleinb¨
ohl A, Bremer H, von K ¨
onig M, K¨
unzi K,
Toohey D, V¨
omel H, Richard E, Aikin K, Jost H-J, Greenblatt JB,
Loewenstein M, Podolske JR, Webster CR, Flesch GJ, Scott DC,
Hermann RL, Elkins JW, Ray EA, Moore FL, Hurst DF, Romashkin
P, Toon GC, Sen B, Margiatn JJ, Wennberg P, Neuber R, Allart M,
Bojkov BR, Claude H, Davies J, Davies W, De Backer H, Dier H,
Dorokhov V, Fast H, Kondo Y, Kyr ¨
o E, Litynska Z, Mikkelsen IS,
Molyneux M, Moran E, Nagai T, Nakane H, Parrondo C, Ravegnani
Copyright 2008 Royal Meteorological Society Q. J. R. Meteorol. Soc. 134: 217– 228 (2008)
DOI: 10.1002/qj

228 L. EL AMRAOUI ET AL.
F, Skrivankova P, Viatte P, Yushkov V. 2002. Chemical depletion
of Arctic ozone in winter 1999/2000. J. Geophys. Res. 107: (D20),
8276, DOI: 10.1029/2001JD00533.
Siegmund P, Eskes HJ, Van Velthoven P. 2005. Antarctic ozone
transport and depletion in austral spring 2002. J. Atmos. Sci. 62:
838– 847.
Singleton CS, Randall CE, Chipperfield MP, Davies S, Feng W,
Bevilacqua RM, Hoppel KW, Fromm MD, Manney GL, Harvey VL.
2005. 2002–2003 Arctic ozone loss deduced from POAM III satellite
observations and SLIMCAT chemical transport model. Atmos. Chem.
Phys. 5: 597– 609.
Sinnhuber B-M, Chipperfield MP, Davies S, Burrowset JP, Eichmann
K-U, Weber M, von der Gathen P, Guirlet M, Cahill GA, Lee AM,
Pyle JA. 2000. Large loss of total ozone during the Arctic winter of
1999/2000. Geophys. Res. Lett. 27: 3473 –3476.
Streibel M, Rex M, von der Gathen P, Lehmann R, Harris NRP,
Braathen GO, Reimer E, Deckelmann H, Chipperfield M, Millard
G, Allaart M, Andersen SB, Claude H, Davies J, De Backer H, Dier
H, Dorokhov V, Fast H, Gerding M, Kyr¨
o E, Litynska Z, Moore
D, Moran E, Nagai T, Nakane H, Parrondo C, Skrivankova P, St ¨
ubi
R, Vaughan G, Viatte P, Yushkov V. 2006. Chemical ozone loss in
the Arctic winter 2002/2003 determined with Match. Atmos. Chem.
Phys. 6: 2783– 2792.
Tilmes S, M¨
uller M, Grooss J-U, H¨
opfner M, Toon GC, Russel III
JM. 2003. Very early chlorine activation and ozone loss in the
Arctic winter 2002–2003. Geophys. Res. Lett. 30: (23), 2201, DOI:
10.1029/2003GL018079.
Tilmes S, M¨
uller M, Grooss J-U, Russel III JM. 2004. Ozone loss and
chlorine activation in the Arctic winters 1991 –2003 derived with the
tracer– tracer correlations. Atmos. Chem. Phys. 4: 2181 –2213.
Urban J, Lauti´
eN,LeFlochm
¨
en E, Murtagh D, Ricaud P, de La No¨
e
J, Dupuy E, Drouin A, El Amraoui L, Eriksson P, Frisk U, Jim´
enez
C, Kyr¨
ol¨
a E, Llewellyn E, M´
egie G, Nordh L, Olberg M. 2004. The
northern hemisphere stratospheric vortex during the 2002 –03 winter:
Subsidence, chlorine activation and ozone loss observed by the Odin
Sub- Millimetre Radiometer. Geophys. Res. Lett. 31: L07103, DOI:
10.1029/2003GL019089.
Urban J, Lauti´
eN,LeFlochm
¨
en E, Jim´
enez C, Eriksson P, de La No¨
e
J, Dupuy E, Ekstr ¨
om M, El Amraoui L, Frisk U, Murtagh D, Olberg
M, Ricaud P. 2005. Odin/SMR limb observations of stratospheric
trace gases: Level 2 processing of ClO, N2O, HNO3and O3.J.
Geophys. Res. 110: D14307, DOI: 10.1029/2004JD005741.
von der Gathen P, Rex M, Harris NRP, Lucic D, Knudsen BM,
Braathen GO, De Backer H, Fabian R, Fast H, Gil M, Kyr¨
oE,
Mikkelsen IS, Mummukainen M, St¨
ahelin J, Varotsos C. 1995.
Observational evidence for chemical ozone depletion over the Arctic
in winter 1991–1992. Nature 375: 131–134.
WMO 1986.. Atmospheric Ozone 1985. Global Ozone Research and
Monitoring Project Report No. 16, Volume 3, World Meteorological
Organization: Geneva.
Copyright 2008 Royal Meteorological Society Q. J. R. Meteorol. Soc. 134: 217– 228 (2008)
DOI: 10.1002/qj