GEOPHYSICAL RESEARCH LETTERS, VOL. 21, NO. 13, PAGES 1419-1422, JUNE 22, 1994
Trajectory model studies of CIOx activation during
the 1991/92 northern hemispheric winter
E. R. Lutman, J. A. Pyle, R. L. Jones, D. J. Lary,
A. R. MacKenzie, and I. Kilbane-Dawe
Centre for Atmospheric Science, Department of Chemistry, University of Cambridge, UK
N. Larsen and B. Knudsen
Danish Meteorological Institute, Copenhagen, Denmark
Abstract. Calculations with a chemical box model along air
parcel trajectories between November 1991 to January 1992 show a
build up of reactive chlorine in late December and early January,
culminating in values of ClOx greater than 2.0 ppbv widespread in
the vortex on 9 January 1992. These values are quantitatively
comparable to the MLS satellite measurements of CIO. We discuss
the chemistry occurring within the vortex, around the vortex edge
and outside the vortex.
Introduction.
Substantial declines in ozone have been observed in the northern
hemisphere in the last decade (Stolarski et al., 1991). The
prerequisite for 03 destruction is the production of levels of CIO
which are enhanced significantly above the usual lower
stratospheric values. This release of ppb-levels of reactive chlorine
compounds is referred to as 'chlorine activation'.
Daily meteorological nalyses of wind and temperature obtained
from the European Centre for Medium Range Weather Forecasts
(ECMWF) were used to calculate air parcel trajectories which have
been used in this study with a photochemical box model. The use of
trajectories allows the recent photochemical history of air masses,
including the exposure to polar stratospheric louds (PSCs), to be
evaluated. By coupling trajectories with a photochemical model,
estimates may be made of the photochemical evolution and, in
particular, the chlorine activation in and around the polar vortex. In
this study the chlorine activation is calculated. Two sons of
trajectories are used, ten-day isentropic trajectories and long
duration three-dimensional trajectories, running from 26 November
1991 to 9 January 1992.
Method
The box model used includes a detailed photochemical and
microphysical scheme. The photochemical and radiation scheme is
an extension of that used by Lary and Pyle (1992). The usual four
heterogeneous reactions which occur on PSC surfaces were
included in the model using sticking coefficients taken from WMO
(1990).
The reaction HCI + HOCI --> CI 2 + H20 (y=0.07) is included
(Hanson and Ravishankara 1991) and given the socking coefficient
suggested by Abbatt and Molina (1992) for "H20-rich NAT". Two
reactions also occur on sulphate aerosol in the model, with a
sticking probability of 0.1 for the reaction N205 + H20 --->2
HNO3. The sticking probability is calculated as a function of
temperature for the reaction CIONO2 + H20 ---> HOCI + HNO 3
and reaches a maximum value of 0.1 at around 195K (Hanson and
Ravishankara 1991).
A microphysical scheme is included for PSCs (Larsen, 1991).
Sulphate aerosol surface areas available for heterogeneous
chemical reactions are also calculated. Firsfly, inside the vortex an
Copyright 1994 by the American Geophysical Union.
Paper number 93GL03045
0094-853 4/94/93 GL-03 04 5503.00
"ordinary" background surface area (~ 0.5-0.75 gm2cm '3)was used
which compares well with in-vortex measurements of aerosol over
Kiruna on January 18 (Deshler tal. 1993), when ~ 0.7gm2cm '3
was measured at 50mb. Secondly, outside the vortex "volcanic"
(i.e. Pinatubo influenced) conditions are calculated (~ 10-15
ga'n2cm'3). These surface areas are somewhat lower than
measurements by Deshler et al. (1992) at Laramie, Wy. oming
between 26July and 29 August 1991 of between 9 and 841am2cm '3.
The recently re-evaluated temperature dependence of the HNO 3
photolysis cross section (Rattigan et al., 1992) is included.
Two sets of trajectories were used. First we used the ensembles
of 10 day, backward isentropic trajectories (Knudsen and Carver,
this issue) which were run over the 475K potential temperature
surface. End points include all the EASOE measurement stations as
well as points along the 40øW longitude line between 84øN and
42øN and along 40øE from 84øN to 30øN (see figure 1, Knudsen
and Carver, this issue). Given some initialisation, the chemical
evolution along the trajectories can be calculated. By running a
whole series of trajectories on, say, the 475K surface, a time series
of the evolution of the two dimensional (latitude, longitude)
chemical behaviour is built up. To allow crudely for transport
chemical data sets, as shown in Table 1, are used to initialize the in-
and out-of-vortex cases (judged by the steepest gradients in
potential vorticity, (Braathen et al. 1992)).
The initialisation gives a total inorganic hlorine (ClOy)of
3.5ppbv in the vortex and 2.9 outside. These values could be a-little
high. For example, Schmidt et al. (this issue) calculate, using
measurements of the CFCs, only a little over 3ppbv within the
lower stratospheric polar vortex during EASOE, while Oelhaf et al.
(this issue) measured 3.3 ppbv of CIONO 2 deep inside the vortex in
mid March 1992. This approach, in which only in- or out-of- vortex
cases are considered, will also oversimplify the structure on any
particular surface. Nonetheless, it should allow us to identify the
impact of processing by PSCs and aerosol, the main objective of
this study.
This approach does not take into account he effect of downward
motion inside and around the polar vortex further changing the
latitudinal gradients of C1 and other species as the winter proceeds y ß
Nor do the 10 day trajectories have any "memory" of previously
calculated C10 x from one run to the next; the chlorine calculated is
that which is activated during each 10 day run. To overcome these
problems we have also run three-dimensional forward trajectories
coveting 48 days from 26 November 1991 to 9 January 1992, again
based on the ECMWF analyses. Initialisation is still a problem; for
simplicity we have again used the in / out-vortex values in Table 1,
except for CIONO 2 and HC1 as described later. Trajectories were
studied which ended between 45 and 60mb on 09.1.92. In the
following we present results at the trajectory end points showing
the build up of CIO x (Cl + CIO + 2 x C1202).
Results from 10 day Trajectories.
By late December the vortex had spun up considerably.
Temperatures from 15 December to 25 December were around
198K in polar regions. However from 19.12.91 to 21.12.91 small
pools of air around 193K were formed near the pole. This was
reflected in the CIO x field calculated for 25.12.91. Several air
1419

1420 Lutman et al.' Trajectory Model Studies
TABLE 1. Initialisations/ppbv for position relative to vortex.
HC1 NO x N205 HNO 3 CH 4 N20
In 2.7 2.4E- 1 0.3 10.7 5.0E+2 9.0E+I
Out 2.1 5.0E-1 1.1 3.7 1.0E+3 1.9E+2
(and at all latitudes HOC1 = 0.2, C10 x = 0.2, C1ONO 2 = 0.4, O x =
3.5E3, HNO 4 = 0.2, Br x = 8.0E-3, CO = 1.8E+l, H202 = 1.0E-3,
H20 = 5.0E+3 ppbv).
parcels experienced low enough temperatures for PSC formation
and patches of C10 x were observed with values of around 0.4ppbv,
for example, over northern Scandinavia and Greenland. However
the cold temperatures were not widely prevalent throughout the
vortex. Many air parcels had not been chemically activated, having
missed the patches of cold air during the previous ten days.
Nonetheless, the picture is clear with local activation having
occurred, but with a rather patchy geographical distribution; the
areas of high CIO x were larger than, and not necessarily collocated
with, the cold patches.
During late December and early January the vortex moved
across to the European sector. Low temperatures were widespread
inside the vortex. Figure 1 shows considerable CIO x activation on
09.1.92 of over l ppbv, which appears to be contained within the
vortex (the 30PV unit contour, a rough indication of the vortex
edge, is shown in the figure). The highest C10 x values of over
1.5ppbv are located over Iceland and Scandinavia.
The C10 x values calculated on 09.01.1992 are in qualitative
agreement with the satellite MLS C10 observations (Waters et al.,
1993). In addition, the modelled chlorine activation agrees well
with the large reduction in HCI within the vortex reported by Bell et
al. (this issue). However absolute values of C10 (with a maximum
of between 1-1.6ppbv) are somewhat lower than those measured by
MLS, where a maximum of more than 2ppbv is widely observed at
45mb. The difference reflects in part the problem of chemical
initialisation of our calculations. The high C10 x values had been
activated within the previous 10 days only and C10 x produced
earlier in December, for example, is not carded forwards. Also the
low CIONO 2 inifialisation limits the activation. Another problem is
the non-uniform nature of the trajectory grid, chosen to coincide
mainly with the EASOE measurement stations, which limits the
geographical extent oœ our chlorine map, e.g. very few trajectories
were produced which ended between 40E and 40W where some
high C10 was seen by MLS.
Results from 48 day trajectories.
Because of the difficulties with trajectory initialisation and the
non-uniform distribution of trajectory endpoints, complementary
three-dimensional trajectories were mn starting in late November
Fig. I C10 x (ppbv) for 09.01.92 at 475K. Calculated using 10 day
trajectories. The 30PV contour is shown.
and ending on 09.1.92. We note that photochemical trajectory
calculations in the stratosphere have not traditionally been run for
longer than about 10 days. However O'Neill (personal
communication) has shown that the features revealed by long-
duration, "domain filling" trajectories, calculated after data
assimilation of meteorological analyses, exhibit a coherent
structure which agrees well with other, independently derived
quantities, e.g. potential vorticity.
As our levels of CIO in figure I are lower than those seen by
MLS a new trajectory initialisation was used with CIONO 2 = HCI =
1.55ppbv. We discuss the implications of this later. Figure 2 shows
C10 calculated using these long 3-D trajectories.
The chlorine activation presented in figure 2 may be divided into
several sections.
High in-vortex ClO x. The entire vortex region is now full of
highly processed air. Calculated values of C10 x (figure 3) are
mainly between 2 and 3ppbv, higher than in figure 1. Both the
spatial extent of the chlorine activation and the magnitude of the
C10 mixing ratios shown in figure 2 agree well with MLS
observations for09 January 1992. The highest C10 values (1.5 - 2.0
ppbv) are found at around 60øN, stretching eastwards across
Europe. There is a strong gradient of C10 across the terminator. Our
CIO pattern is displaced slightly eastwards compared to MLS, but
generally the agreement is very good.
The levels of C10 x produced inside the vortex in our calculations
depend on several factors' total CI initial chlorine partitioning and ß y,
choice of heterogeneous reaction probability. As expected our new
initialisation (C1ONO 2 = HC1 = 1.55ppbv) produces higher levels
of CIO, between 1.5 and 2.3 ppbv, comparable to those measured
by MLS. Whether this higher level of C1ONO 2 is realistic depends
on either high levels of chlorine nitrate being present initially, or on
levels of NO 2 being sufficient to recycle some CIO x back into
C1ONO 2. Such levels of NO 2 are not produced in our model with
the fast reaction N20 5 + H20 occurring on aerosol and with slow
nitric acid photolysis rates. If this level of chlorine nitrate is
unrealistic then we must conclude that we do not yet fully
understand the processes controlling the magnitude of the C10 x
activation.
When a lower value was chosen for total C OwY, ilower C10 values were calculated. Our quantitative agreement th the MLS data
requires the high ClOy initialisation.
If CIONO 2 was depleted, for example due to macring with water
on aerosol, and thus unable to react with HC1, the reaction HC1 +
HOCI becomes potentially very important for activating chlorine.
2.00
1 .go
1.8O
1,70
1.60:
1.5o
1
1
1.2o
1.1o
1 .oo
0.90
0.80
0.70
0.60
0.50
Fig. 2 09.01.92 CIO(ppbv) between 45 and 60 mbar at local noon,
calculated using 48 day trajectories. The terminator determines the
sharp CIO edge inside the vortex. High activation, in the form of
C120 2, is found polewards of the terminator. Note that the out-of-
vortex activation is generally produced through reactions on
sulphate aerosol although there are also some incidents of high CIO
produced through activation in the tropics. These occur mainly at
the lower latitudes which are not shown in the figure.

Lutman et al.: Trajectory Model Studies 1421
Fig. 3 CIO x (ppbv) for 09.01.92 between 45 and 60 mbar at local
noon. Calculated using 48 day trajectories. Locations on the
trajectories below the PSC threshold are also shown (X) for
07.01.92, the day with the largest area of PSC temperature.
The sticking probability for this reaction is still uncertain (Abbatt
and Molina, 1992). If it is set to the lower value of 0.002 as
indicated for "HNO3-rich NAT" by Abbatt and Molina (1992),
C10 x levels are again reduced. In-vortex levels of C10 are then
calculated to be below 1.5ppbv, and indeed only a few points have
values of C10 greater than 1 ppbv. Thus quantitative agreement with
MLS data also requires use of a high sticking coefficient for the
reaction HC1 + HOC1.
The spatial area of perturbed C10 x is significantly larger than the
area of cold temperatures which is seen by the model. Figure 3 also
shows a snapshot in time of the area of PSC activation seen by the
trajectories (marked by X). The figure shows day 45.89 (07.01
1992), chosen as the model day with the largest area with
temperatures below the NAT point calculated in our model. Figure
3 again confirms the idea put forward by Jones et al. (1990), and
others, describing the northern hemispheric vortex as being
characterised by air flowing through relatively small patches of cold
temperatures, and processing the entire vortex.
Edge of vortex activation on PSCs. High values of C1Ox (over
2ppbv) are calculated for several trajectories ending over the
Mediterranean (figure 2). In order to study the chlorine activation in
detail, in figure 4 we consider the chemical evolution along a
typical trajectory ending at 45.03øN, 6.41øE on 09.01.1992. The
latitudinal movement of the air parcel (not shown) falls into two
distinct periods. The trajectory is confined to the vortex edge for
most of the run, undergoing considerable latitudinal excursions
between 45øN and 75øN. Zenith angles are always greater than 70 ø
but there is sufficient ultraviolet radiation to initiate
photochemistry.
The temperature of the air parcel (not shown) remains mostly
between 200K and 230K, too warm for heterogeneous reactions on
sulphate aerosol or PSC formation, apart from 3 short periods. On
days 24 and 45 (20.12.91 and 09.01.92) temperatures drop below
200K, cold enough for heterogeneous reactions on sulphate aerosol.
On day 36 (02.12.91) a PSC is encountered as the temperature
drops to 193K.
The temperature history is reflected in the behaviour of the
chlorine reservoirs. Throughout the first (warm) half of the run
C1ONO 2 values remain high, until day 24 (20.12.91) when the
temperature drops below 200K and CIONO 2 is converted into
HOC1. HC1 values drop slowly through gas phase chemistry until
day 36 (02.12.91) when the PSC is encountered. 1 1ppbv of CIO is
activated, mostly through C1ONO 2 reacting with HC1. The final
value of C10 x is 0.79ppbv, lower than levels inside the vortex but
still representing significant activation. More CIO x would be
produced if the trajectory was run on, due to photolysis of the HOC1
which was formed during the run. Appreciable ozone loss (7% in
48 days) occurs in this calculation, mainly when the air parcel
moves away from the vortex edge to somewhat lower latitudes,
during the last few days of the run.
0 5 10 15 20 25 30 35 40 45 50
2.0 I I I I I I I I I 2.0
ß CIONO2
1.0- - 1.0
0.5- -0.5
0.0 0.0
0 5 10 15 20 25 30 35 40 45 50
DAY OF RUN
Fig. 4 An "edge of vortex" trajectory ending at 45.03øN, 6.41øE.
C1ONO2, HC1 and HOC1 / ppbv versus day of run are shown. Day 0
= 26 November 1991.
Out-Vortex activation on aerosol. Waters et al. (1993) report
patches of elevated C10 outside the vortex. Modest chlorine
activation compared to that observed inside the vortex is seen in our
calculations for some air parcels between the vortex edge and
approximately 40øN, with C10 values between around 400 and 800
pptv. Some activation is evident in figure 2 while some occurs at
lower latitudes (not shown). There is some structure in the amount
of activation. Some trajectory endpoints in this region show very
small values of C1Ox interspersed with other points of higher ClOx;
there are some isolated cases of high ClOx at even lower latitudes.
Apart from the cases specifically mentioned in the above section,
these trajectories have not seen any PSC activation. Instead,
chlorine activation has occurred in the model on volcanic sulphate
aerosol. The levels of chlorine activation predicted by our model
are thus heavily dependent on the rates of these aerosol reactions.
As yet there is little field evidence, as opposed to laboratory
measurements, that CIONO 2 reacts with water in the atmosphere on
sulphate aerosol. Should this reaction occur more slowly than is
currently thought, or even not occur at all, the gradients of ClOx
would become much sharper across the vortex boundary than is
shown in figure 2. Rerunning these trajectories with the background
aerosol, instead of the volcanic aerosol, produces much smaller
values of C10 x, around 10-100pptv.
These out-of-vortex air parcels have moved widely in mid
latitudes, remaining in general between 40øN and 70øN. Due to the
large latitudinal excursions experienced, these moderate levels
(400-600pptv) of CIOx produce substantial amounts of ozone
destruction, in general around 10% during the 48 days.
Tropical Activation. As mentioned above, over the 48 day run,
many trajectories appear to move in the surf zone, repeatedly
experiencing large latitudinal excursions, the majority moving
between 40øN and 70øN. Some small number of the trajectories
seen in middle latitudes had spent time at considerably lower
(tropical) latitudes including some which started and finished in
middle latitudes. If this air had initially a chemical composition
characteristic of middle latitudes (e.g. some chlorine in the
inorganic forms, HCI and C1ONO2), then some tropical PSC
activation could have taken place. Certainly, in a few of the
trajectory runs, PSCs were activated and due to our chlorine
initialisation high amounts of CIO x (over l ppbv) were produced
and transported to mid-latitudes. One such trajectory moved
between 5øN and 60øN and the chlorine produced (over l ppbv) led
to over 25% ozone depletion during the run. Even though such air
parcels did not produce as much C1Ox as air inside the polar vortex,
large amounts of ozone destruction ensued due to the greater
insolation. Tropical activation may or may not be an explanation for
the small areas of high CIO observed by MLS in mid to low
latitudes. However it does seem to be a mechanism which should be
explored further, particularly at higher pressures than 50mb where
extra tropical air temperatures are, on average, colder.

1422 Lutman et al.: Trajectory Model Studies
No other evidence was found of the patches of high C10
observed in mid latitudes by MLS on 09.1.92. This could be
because our 350 trajectories missed air parcels which were
activated inside the vortex, before peeling away from it (as
described above). The trajectories may also have missed local small
scale temperature fluctuations which were cold enough to activate
chlorine. It is also possible that if we had included the reaction
between HOCI and HC1 on sulphate aerosol as proposed by Cox et
al., (this issue) higher amounts of C1Ox could have been produced
at the lower latitudes where some aerosol activation was calculated
(see above). We note also that the patches of C10 may not have
been real, but could arise from spurious instrument noise (Waters et
al., 1993).
Conclusions.
The activation of chlorine during EASOE in early January 1992
has been modelled using a trajectory model. The period appears to
have been well reproduced using three-dimensional trajectories.
The spatial extent of the largest chlorine activation is in good
agreement with satellite measurements. The small scale detail in
C10 depends critically on assumptions of chemical initialisafion
and details of heterogeneous chemistry. We can only reproduce the
magnitude of the C10 values seen by MLS using a high total
chlorine concentration of 3.5ppbv, and by using a high inifialisation
of CIONO 2. A high sticking coefficient for the heterogeneous
reaction HC1 + HOCI is also necessary to reproduce the
measurements of ClO.
Some PSC activation occurred in the model around the vortex
edge, activating high values of CIO. Moderate chlorine activation
was also seen outside the vortex edge; this was produced in the
model on sulphate aerosol and is very dependent on assumptions
made regarding the rates of heterogeneous reactions on sulphate
aerosol.
No other evidence was seen of the patches of ClO seen at mid
latitudes by MLS. If this high CIO was real, it has been missed by
our trajectories.
In a companion paper, the aleactivation of chlorine and ozone
loss during the 1991/92 northem hemisphere winter is discussed.
(Lutman et al., this issue.)
Acknowledgements. This work was funded by DG XII of the
CEC under contract STEP-CT91-0139 for the support of EASOE.
ERL was supported by a Gassiot Award from the Meteorological
Office and IK-D was funded by the Human Capital and Mobility
Programme (HCMP) Proposal Number ERB4001GT921327. The
modelling work described here is part of our NERC supported
UGAMP effort.
B raathen G.O., Stordal E, Gunstrom T., Knudsen B., Kloster K.,
EASOE Meteorology report, Norwegian Institute for air
research, 1992.
Cox, R.A., et al., Activation of stratospheric hlorine by reactions
in liquid sulphuric acid, Geophys Res. Lett., this issue, 1993.
Deshler T., In situ measurements of the size distribution of the
Pinatubo aerosol over Kiruna, on four days between 18 January -
13 February 1992, Geophys Res. Lett., this issue, 1993.
Deshler T., D.J. Hoffman, B.J. Johnson, W.R. Rozier, 1992,
Balloonborne Measurements of the Pinatubo Aerosol Size
Distribution and Volatility at Laramie, Wyoming during the
summer of 1991, Geophys Res. Lett., 19, 199-202, 1992.
Hanson D.R. and A.R. Ravishankara The reaction probabilities of
ClONO 2 and N20 5 in 40 to 75% sulphuric acid solutions,
J. Geophys. Res., vol. 96, 17307-17314, 1991.
Jones R.L., D.S. McKenna, L.R. Poole, S.Solomon, On the
influence of Polar Stratospheric Cloud formation on chemical
composition during the 1988/89 Arctic winter, Geophys Res.
Lett., vol. 17, 545-548, 1990.
Knudsen B. and G. Carver, Accuracy of the isentropic trajectories
calculated for the EASOE campaign, Geophys Res. Lett., this
issue, 1993.
Larsen N. Polar Stratospheric Clouds: A Microphysical Simulation
Model, Danish Meteorological Institute Scientific Report, 1991-
92, Copenhagen, 83 pp, 199 I.
Lary D.J. and J.A. Pyle Diffuse Radiation, Twilight and
Photochemistry, J. Atmos. Chem., 13, 373-392, 1991.
Lutman E.R., R. Toumi, R.L. Jones, D.J. Lary, J.A.Pyle, Box Model
Studies of ClO x aleactivation a d ozone loss during the 1991/92
northern hemispheric winter, Geophys Res. Lett., this issue,
1993.
Rattigan O., E.R. Lutman, R.L. Jones, R.A. Cox, K. Clemitshaw,
and J. Williams, Corrections to 'Temperature dependent
absorption cross-sections of gaseous nitric acid and methyl
nitrate', d. Photochem. Photobiol., 69, 125-126, 1992.
Stolarski R.S., D. Bloomfield., R.D. McPeters, J.R. Herman, Total
ozone trends deduced from Nimbus 7 TOMS data, Geophys Res.
Lett., vol. 18, 1015-1018, 1991.
Waters J.W., L. Froidevaux, W.G. Read, G.L. Manney, L.S. Elson,
D.A. Flower, R.E Jamot, R.S. Harwood, Stratosphefic C10 and
ozone from the Microwave Limb sounder on the Upper
Atmosphere Research Satellite, Nature, 362,597-602, 1993.
WMO, Scientific assessment of stratospheric ozone: 1990, World
Meteorological Organisation, Global Ozone Research and
Monitoring Project Report No. 20.
References
Abbatt J.P.D. and M.J. Molina, The heterogeneous reaction of
HOC1 + HCl -• C12 + H20 on ice and nitric acid trihydrate:
reaction probabilities and stratospheric implications, Geophys
Res. Lett., 19, 461-464 1992.
Bell W., N.A. Martin, T.D. Gardiner, N.R. Swann, P.T. Woods, P.F.
Fogal, J.W. Waters, Groundbased measurements of stratospheric
constituents over Are, Sweden in the winter of 1991/92, Geophys
Res. Lett., this issue, 1993.
R.L. Jones, I.Kilbane-Dawe, D.J. Lary, E.R. Lutman, A.R.
MacKenzie, J.A. Pyle, Centre for Atmospheric Science, Dept. of
Chemistry, University of Cambridge, Lensfield Rd., Cambridge,
CB2 IEW, U.K.
B. Knudsen, N. Larsen, Danish Meteorological Institute, DK-
2100 Copenhagen 9, Denmark.
(Received: November 27, 1992 Revised: May 6, 1993
Accepted: October 29, 1993.)