JOURNAL OF GEOPHYSICAL
RESEARCH, VOL. 105, NO. D5, PAGES 6845-6853, MARCH 16, 2000
Role of lee waves in the formation of solid polar stratospheric
clouds' Case studies from February 1997
E.D. Rivi•re,• N. Huret,
•'2F.G.-Taupin,
• J.-B.Renard,
1M. Pirre,1'2S.D. Eckermann,
N. Larsen,
4T. Deshler,
5F.Lef•vre,
6S.Payan,
7andC.Camy-Peyret
7
Abstract. Recenttheoriesof solidpolar stratospheric
clouds(PSCs) formationhave shownthat
particlescouldremainliquiddownto 3 K or 4 K belowthe ice frostpoint.Suchtemperatures
are
rarelyreachedin the Arctic stratosphere
at synopticscale,but nevertheless,
solidPSCsare
frequentlyobserved.Mesoscaleprocesses
suchasmountain-induced
gravitywavescouldbe
responsible
for their formation.In thispaper,a microphysical-chemical
Lagrangianmodel
(MiPLaSMO) and a mountainwave model(NRL/MWFM) are usedto interpretballoon-borne
measurements
madeby an opticalparticlecounter(OPC) andby the Absorptionpar Minoritaires
Ozoneet NOr (AMON) instrumentaboveKiruna on February25 and 26, 1997, respectively.The
model resultsshowgoodagreementwith the particle size distributionsobtainedby the OPC in a
layer of largeparticles,andallow usto interpretthis layer asan evaporatingmesoscaletype Ia
PSC (nitric acid trihydrate)mixed with liquid particles.The detectionof a layer of solid particles
by AMON is alsoqualitativelyreproducedby the modeland is interpretedto be frozensulfateacid
aerosols(SAT). In this situation,the impactof mountainwaveson chlorineactivationis studied.
It appearsthatmesoscale
perturbations
amplifysignificantlythe amountof computedC10, as
compared
to synopticruns.Moreover,MiPLaSMO chemicalresultsconcerningHNO3 andHC1
agreewith measurements
madeby the Limb ProfileMonitorof the Atmosphere(LPMA)
instrument
on February26 at a verycloselocationto AMON, andexplainpartof the differences
betweenLPMA measurement
andReactiveProcesses
Ruling the OzoneBudgetin the
Stratosphere
(REPROBUS) modeloutputs.
1. Introduction
In polar regions,heterogeneousreactionswhich occur on
stratosphericaerosols and on polar stratosphericclouds
(PSCs)are a crucialstepin chlorineactivationresponsiblefor
ozone loss [Solomon et al., 1986]. The efficiency of those
processes
dependson threeparameters
stronglylinked to the
temperature: particle type, surface area, and chemical
compositionin the caseof liquid particles.Thus the way in
whichPSCsform is of primary importance.Our knowledge of
(STS), andtype II PSC madeof ice particles.Somepathways of
solid PSCs formationremainuncertain,especially concerning
PSC Ia [Tolbert, 1996]. Tabazadeh et al. [1994] and
MacKenzie et al. [1995] suggestthat PSC Ia can be formedby
freezing of STS. Zhang et al. [1996] show that NAT could form
by nucleation on preactivated SAT. Tabazadeh and Toon
[1996] suggest that solid type PSC could form through a
metastabledilute HNO3/H20 solid phase.Other works point
out that PSC Ia could be composedfirst of metastablenitric
acid dihydrate (NAD) particles which could formin a strong
PSCs formation has evolved over the past few years,
warming event [Tsias et al., 1997], and convert later into NAT
particularlyconcerningsolid particle formation.Five types of
[Middlebrook et al., 1992]. More recent studies suggest that
particles are usually considered:liquid binary aerosols,
NAT could form by nucleationon ice nuclei [Biermann et al.,
sulfuricacid tetrahydrate(SAT) particles,type Ia PSC madeof
1998]. Type II PSC would thereforebe the first step in the
nitric acid trihydrate (NAT), liquid type Ib PSC
formation of solid PSCs [Koop et al., 1997]. SAT particle
(H2SO4/HNO3/H20)alsonamedsupercooledternary solution
formation was studied fi'om binary solution freezing
experiments,although this mechanismdoes not occur under
•Laboratoire
de Physique
et Chimie
de l'Environnement,
CNRS, polar stratosphericconditions [Clappet al., 1997; Carleton
Orl6ans, France.
et al., 1997], since the droplets are STS at temperatures at
2Also
atUniversit6
d'Or16ans,
Orl6ans,
France.
3E. O. HulburtCenterfor SpaceResearch,
NavalResearchwhich they are supposedto freeze.SAT formationmechanism
is thereforestill unclear. However, SAT particles could play
Laboratory,Washington,D.C.
4Department
of Research,
DanishMeteorological
Institute,an importantrole in PSCs cycle formation:Koop and Carslaw
Copenhagen,
Denmark.
5
ß
[1996] have shown that SAT deliquescencecould convert
Departmentof AtmosphericScience, University of Wyoming,
SAT aerosolsinto STS droplets, which could freeze later to
Laramie.
producesolid PSCs [Iraci et al., 1998].
6Service
d'A6ronomie
duCNRS,Paris,
France.
7Laboratoire
dePhysique
Mo16culaire
etApplications,
CNRS,
Paris, Our knowledge about the temperatureat which solid
France.
particlesare formedhas alsodevelopedover the last few years.
Koop
et al. [1995] have shown that liquid particles would
Copyright2000 by theAmericanGeophysical
Union.
not freezeabovethe ice frost point (hereafterTiGe).Koop et al.
[1997] suggestthat liquid particlescould remain liquid down
Papernumber1999JD900908.
0148-0227/00/1999 JD900908509.00
to 3 K below TiGe,while Carslaw et al. [1998a] suggest 4 K
6845
6846
RIVII2RE ET AL.' LEE WAVES AND FORMATION OF SOLID PSCs
below Tice.This imposesvery restrictive conditions for solid
particle formation,since at synoptic scale, such temperatures
are not usually reachedin the Arctic stratosphere.However,
mesoscaleprocessessuchas mountain-inducedgravity waves
can generate vertical air motions that are large enough to
adiabaticallycool the air parcelsto 3-4 K below Tice[Carslaw
et al., 1998a]. The role of mountain-inducedwaves(lee waves)
on PSC microphysics and chlorine activation has been
investigated recently [Meilinger et al., 1995; Tsias et al.,
1997; Carslaw et al. 1998b, 1999].
In this context,the aim of this paper is to studytwo casesof
particle observation in February 1997 over Kiruna (67.9øN/
22.1øE), Sweden, which cannot be explained by synoptic
scaletemperaturehistories. On February 25, 1997, a balloonborne optical particle counter (OPC) [Deshler et al., 1993]
measureda layer of large particles, while the balloon-borne
Absorptionpar Minoritaires Ozone et NOx (AMON) [Renard
et al., 1996] instrumentdetecteda layer of solid particles on
February 26. The Microphysical and Photochemical Lagrangian StratosphericModel of Ozone (MiPLaSMO), which has
been developed to study the chemical composition of the
stratosphere(taking into account a detailed PSC model,
stratospheric chemistry, and heterogeneous reactions on
PSCs), is used to tentatively explain the measurements.
The
orographicgravity wave model ofBacmeister et al. [1994] is
also used in this work to investigate the role of mountain
waves in the formation of thesedisturbedlayers.
21 5
21 0
20 5
CONDENSATION
(H20,
HNO3)
200 CONDENSAT
'.
(H
20,
HNO
3) HN::0
•
195
.....
?::::::::::::::??:?:
..............
::::::::::
::
•:::•?itA)
•:'•5••
iN '""%•71
I FREEZING
The models used here are described in detail in section 2,
then the instruments OPC, AMON, and Limb Profile Monitor
of the Atmosphere(LPMA) are presentedin section 3. For the
two casesconsideredin this paper,the role of lee waves in the
formationof disturbedlayersis investigatedin sections4 and
5. In the caseof February26, 1997, the role of mountain waves
in chlorine activation is discussedby comparingMiPLaSMO
chemicaloutputs with Reactive ProcessesRuling the Ozone
Budget in the Stratosphere(REPROBUS) Chemistry Transport Model (CTM)calculations and LPMA measurements
of
HNO3 and HC1 [Payan et al., 1999].
DELIQUES•ENCE
EVAPORAT
Figure 1. Schemeof PSC formationused in the model as a
function of the temperature.Typical values of TN^T,Ti• and
rrneltare indicated.
PSCII, solidsulfuric
aerosols
supposed
to beSAT,andliqdid
particles.All the particlesare distributedin 50 sizebins from
0.001gm to 100 pro.Figure 1 presentsthe pathwaysusedin
this codefor particleevolution,as a functionof temperature.
As the temperaturedecreases,
the liquid binary aerosols
grow
as
STS
by
uptake
of
HNO3
and
H20 fromthe gas phase.
2.1. MiPLaSMO
PSC II particles are the first solid particles to form from
MiPLaSMO is a numericalmodel, which was developedto
freezingof liquid droplets.The processbegins at temperatures
study in detail microphysical and chemical processes lessthan Tice-1 K. Accordingto Koop et al. [1997], particles
associated with PSCs. This model describes the evolution
of
can remainliquid until the Tice-3 K thresholdis reached.
the microphysical,thermodynamical,and chemicalproperties Duringthis freezingprocess,the condensed
HNO3 and H2SO4
of an air parcel following an isentropic trajectory. The PSC are conservedthrough the formationof an inner NAT shell
code and the photochemicalcode are coupled via heterogene- surroundinga SAT core in the ice particle. All the sulfuric
ous chemistry.In this section, the differentpackagesof the acid is in the condensedphase.As the temperaturewarmsup,
2. Model Description
model are described.
2.1.1. Trajectory model. MiPLaSMO uses isentropic
trajectoriesas input.Trajectoriesare calculated[Knudsenand
Carver, 1994] using European Centre for Medium-Range
Weather Forecast (ECMWF) meteorologicaldata with a
spatialresolutionof 1.125ø x 1.125ø. They areavailableevery
6 hours. Trajectory data are calculated every 2 hours.
Mesoscaleperturbationsof temperatureand pressurecan be
prescribedalong the trajectory.
2.1.2. Microphysical code. The microphysicalcode was
initially developedat the DanishMeteorologicalInstituteby
N. Larsen [Larsen et al., 1997]. It describesthe time evolution
of severalkinds of particles:PSC Ia (supposedto be NAT),
the ice particlesevaporate,revealingthe NAT coreandNAT
can evaporatesto leave SAT particles. Size evolution of
liquid, NAT,
and ice particles due to
the
condensation/evaporation
mechanismis taken into account.
Heterogeneousnucleationof PSC II on PSC Ia is also
consideredwhen Ti• is reached.In the model we assumethat
SAT can return to liquid particles in two cases: when
temperatures
passbeyondthe meltingtemperature
of SAT, or
when temperaturesgo below the temperatureof SAT
deliquescence[Koop and Carslaw, 1996] if no other solid
particlesexist.
2.1.3. Photochemical
code.
The chemical code describes
the evolutionof 41 speciesof the NOy, HOx, Cly,O,•,andBry
RIVII•RE ET AL.' LEE WAVES AND FORMATION OF SOLID PSCs
6847
families.It takesinto account112 homogeneous
reactionsand
five heterogeneousreactionson clouds or aerosols.Those
heterogeneousreactionsare
condensation
nuclei(r > 0.01 pm)andof opticallydetectable
aerosol
(0.15< r < 10.0[tm).Theaerosol
counter,
followingan
(R1)
(R2)
(R3)
(R4)
(R5)
determinedfrommeasurements
of the intensity of scattered
HC1+CIONO2 -• Cl2 + HNO 3
H20 +C1ONO2 -• HOCI+ HNO3
N 2¸ 5+ HC1-->CINO2 + HNO3
N 2¸ 5+ H 2¸ -->2 HNO3
HOCI + HCI -->C12+ H 20
originaldesignby Rosen[1964],countsandsizesaerosol
particlesdrawn into a scatteringregion. The size is
whitelightat 40ø fromtheforwarddirection
usingMie theory
andassuming
spherical
particles
withan indexof refraction
of
1.45 [Hofmannand Deshler,1991].
First-order
rateconstants
k (s'l) for heterogeneous
reactions
dependon type andsurfaceof particles.They are calculatedas
follows:
k=TSF
(1)
where S is the total areaof a particle type consideredfor the
consideredparticle type, F the meanthermalvelocity, and ),
the reaction probability. Reactionprobability are taken from
De More et al. [1997], Carslaw and Peter [1997], and
Ravishankara and Hanson [1996].
The numerical scheme used for the chemical code is semi-
3.2. AMON
Instrument
AMON
is a balloon-borne
UV-visible
spectrometer
devotedto nighttimemeasurements
of the stratospheric
species
03, NO2,andNO3[Naudet
et al., 1994;Renardet al.,
1996].Aerosolextinctioncanalsobe retrievedasa function
of the wavelength.Observations
are performed
using the
stellar occultationmethod,which consistsin analyzing the
modifications
of the spectrum
of a setting star inducedby
absorbingatmospheric
species.A reference
spectrumis
obtainedwhen the star is above the balloon horizon and the
balloon is afloat.The transmissionspectraare obtained by
implicitsymmetric(SIS). The nonlinearequationsof chemical dividingtheobserved
spectra
by thereference
spectrum.
The
species evolution are converted into a linear system of retrieval of the slant column densities is performedusing a
equations.A descriptionof this methodcanbe found in the leastsquares
fit betweenthe observed
transmission
spectra
andthespectra
calculated
usingtheknowncrosssections
of
2.2. Mountain-Wave
the absorbingspecies.The residual in the transmission
spectra,afterremovingthe contribution
of Rayleighscattering
Calculation
Coolings induced by mountain waves are calculated offline by the Naval Research Laboratory Mountain Wave
ForecastModel (NRL/MWFM). It is describedby Bacmeister
et al. [1994]. Extensionsof the model to calculatetemperature
variability are described by Eckermann and Bacmeister
[1998]. The model relies on a database of dominant
topographicmountainrangeswith estimationof their height,
width, and orientation. Vertical motions induced by the
mountainridge are calculatedat eachstandardpressurelevel
(including 50, 30, and 10 hPa) and every 6 hours by
NRL/MWFM. The temperatureperturbation is then deduced
using adiabatic parcel thermodynamics,since stratosphere
mountain wave advection is adiabatic to a good
approximation [Eckermann et al., 1998]. The temperature
perturbation introduced in MiPLaSMO is deduced from
cooling pathways previously calculated by NRL/MWFM.
Interpolationsare made to the pressuresand timesof interest.
As in Dias et al. [1997], the mesoscaletemperaturevariation
is assumedsinusoidal. The amplitude of the sinusoid varies
with respect to the maximumcooling computed by the
mountain wave model. The period of the oscillation is
estimated
from
the
mean
time
between
two
consecutive
and the contribution of molecularabsorption,is assignedto
aerosols.
The verticalprofilesof the speciesand of the aerosol
extinction are obtained after inversion performedusing the
onionpeelingtechnique[Renardet al., 1996].
3.3. LPMA
Instrument
The LPMA is a remotesensing infraredFourier transform
instrumentoperatingin absorptionagainstthe Sun [Camy-
Peyret, 1995]. With its high spectralresolutionand
sensitivity,the retrievalof verticalprofilesof tracespecies
havingstratospheric
mixingratiosas smallas 0.1 ppbv is
possible.A global fit algorithm[Carlotti, 1988] associated
with an efficientminimizationalgorithm of the Levenberg-
Marquardttype [Presset al., 1992] is usedfor the retrieval.
This retrievalalgorithm[Payan et al., 1998] allowedus to
retrievevertical profiles of HC1, HNO3, and severalother
stratospheric
speciesfrom occultation
data.
4. Study of February 25, 1997
4.1. OPC
Measurements
cooling eventsas forecastby the NRL/MWFM model.
Observations
wereperformedfromKiruna (67.9øN/22.1øE)
on February25, 1997.TheOPCflight startedat 0918 UT and
3.
endedat 1051 UT. Data were recordedfromthe ground to 30
km. The aerosolconcentration
profilesfor differentsize classes
Instruments
In the flame of the validation campaignof the Improved
Limb Atmospheric Spectrometer(ILAS) instrument on board
the ADEOS satellite [Sasano et al., 1999], chemicalspecies
and aerosol measurements
were performedin early 1997 with
balloon-borne instruments from Kiruna, Sweden. Here we
presentthe OPC, AMON, and LPMA instrumentswhich have
performedmeasurements
on February25 and 26.
used for in situ measurements
4.2. Initialization
of MiPLaSMO
For this case,the initial distributionof particles,typical of
background
aerosol,is deducedfromOPC measurements.
The
distribution
at the
desired altitude
is
obtained by
interpolation between two typical distributions of
background
aerosols
observed
belowand abovethis altitude.
3.1. Optical Particle Counter
The OPC
areshownin Figure2. A layerof largeparticleswasobserved
between 20 and 22 km.
from the balloon-
borne gondola measures the number concentration of
The distribution measuredon February23, 1997 (T. Deshler,
unpublishedILAS database,1997) has been used. These
6848
RIVII•RE ET AL.' LEE WAVESAND FORMATIONOF SOLIDPSCs
'
' ' '''"1
'
' ' ' '"'1
I
I
Opt ical Part icle Count er
February 25, 1 997
3O
I I I lll[
radius
_
•>
0.15 lam
.....
> 0.25 lam
----->0.301am
•>0.501am
.....
> 0.75 lam
----->1.081am
..............
> 1.25 lam
•..>1.751am
....... > 2.50 lam
•>3.501am
.....
> 5.00 Izm
----->10.01zm
25
2O
-
_
•
15
lO
_
_
_
_
_
i
0.001
0.010
0.100
i i i i iill
1 .000
1 0.000
concentration(cm-3)
Figure2. Verticalprofilesof aerosolconcentration
forparticlesbetween0.15 gmand 10 gmfromthe in situ
OPC measurements
on February25, 1997.
measurements
have beenpreferredto those of February25
becausethe disturbed layer involving large particlesis
thinner.This allowsa betterprecisionin the interpolation
of
aerosolparameters.
At initializationof the model,particlesare
At 500 K the OPC measurements
were expressed
in termsof
concentrationas a functionof the radius.The back trajectory
for February25 crossedthe Scandinavianmountainranges
prior to its arrivalat the samplinglocation (not shown).Two
assumedto be liquid. Chemicalspeciesinitialization is taken
fromREPROBUS3D CTM [LeJ•vreet al., 1998] simulations
of the whole winter 1996-1997.
220
4.3. Synoptic Situation
A 20-day isentropic back trajectory was calculatedto
obtainthe temperature
historyat the 500 K level,
210
corresponding
to 21.5 km (35 hPa),wherethe largestparticles
are observed(see Figure3). At the locationof the
•'
measurement,
thesynoptictemperature
(192 K) is 5 K overthe
• 200
frostpoint,andthus,toowarmfor a PSC formation.
Moreover, •
thecoldest
temperature
reached
alongthis trajectory
was3 K
E
abovethe icefrostpoint(120hoursbefore
themeasurement).•
As a consequence,
the modeldoesnot calculateany solid
190
....Tice
PSCs, using synoptic-scale
temperatures.
180
4.4. MesoscaleInterpretation
i,,,,,
-500
, , , I ,,
-400
,1
, ,,
, , I,,,,
-300
, ,,,
...........................
,I
.........
-200
I , I i i i i i i i
-100
0
time (h)
A preliminarystudyof this casewasmadeby Huret et al.
[1998],considering
a theoreticalmesoscale
perturbation.
The
time evolution for the 20-day
authorshaveshownthat the hypothesisof a mountainwave Figure 3, Synoptictemperature
back trajectory (heavy solid line) at the 500 K level for
eventcouldbe an explanation
to interpretthe measurements
of February 25, 1997. Melting temperatureof SAT (Tmelt),
February
25, 1997.Herewe usemountain-induced
wave(lee temperatureof NAT condensation(TN^T),temperatureof SAT
wave) perturbationspredictedby the NRL/MWFM model deliquescence
(Tdeliq),
and ice frostpoint(Tic½)
are also shown;
describedaboveto tentativelyexplaintheseobservations.
t = 0 correspondsto the time of the measurements.
RIVII•RE ET AL.' LEE WAVES AND FORMATION OF SOLID PSCs
6849
215
very small contribution from SAT particles) and the second
mode centeredon 1 gm is due to NAT particles.
205
5. Study of February 26, 1997
5.1. Measurements by AMON and LPMA
Observationswere also performedfrom Kiruna, on February
195
26, 1997. The AMON
observations started at 2105 UT and
ended at 2211 UT with a balloon float altitude of 31.5 km. The
starusedfor the occultation was Rigel, and 100 spectrawere
recorded.The latitude and longitudewere in the 67.7ø-67.0øN
and 27.3ø-18.5øE ranges, respectively, with tangent point
altitudes decreasing from 31.4 to 11.3 km. The LPMA
instrumentperformeda flight fromEsrange(68øN/21øE), just
beforethe flight of AMON. Infrared spectra were recorded at
sunset during occultation l?om 1500 UT to 1540 UT. The
latitude and longitude ranges for the LPMA occultation
185
175
•
10.0
measurements were 67.85ø-66.87øN
E
..9.0
and
26.77ø-18.78øE,
respectively,and so were very close to thoseof AMON.
5.1.1. Aerosol measurementsby AMON. Vertical profiles
of aerosol extinction coefficient obtained every 25 nm are
shown in Plate 1. Almost all the wavelength dependenceof
E
• 1.00
,:
the aerosol extinction
c)
INAT
, ,II
,,I,,,I,,,I
-0.8
-0.6
-0.4
-0.2
at different
altitudes
can be
fitted using a Mie theory algorithm including a log-log size
distribution. The comparison of the aerosols parameters
derived from AMON with those of the OPC obtained the day
before showsthat there is a very good agreementbetween the
-, 0.10
0.01
coefficients
two
measurements
in terms of mean radius.
In
the
altitude
0
rangewhere there is good agreement,the retrieved parameters
time (h)
correspondto background aerosols.
Significant discrepanciesl?omMie theory occur in a layer
Figure4. February25 backtrajectory,
at 500 K. (a) Mesoscale
temperatureperturbation1 hour before the OPC measurements. between21 and 23 km in the AMON profile. In this layer the
Trnelt,
TNAT,
and Ticeare presented(seeFigure 3 caption).(b) wavelength-dependentextinction is maximumfor the larger
Surface
areatimeevolutionof the particlesurface
areaduring wavelengthand cannotbe fitted whatever the refractionindex
the lee wave event shown in Figure 4a.). Liquid particle,
NAT, andPSC II totalsurfaceareaarepresented.
hours before the OPC measurements, NRL/MWFM
calculationsshow that the air parceltemperaturehad been
influenced by lee waves, generatedby flow over the
and
size
distributions
used.
This
result
means
that
the
hypothesisof sphericalparticlesused for Mie scatteringis not
valid. Besides, strong chromatic scintillation was observed
10 4
Scandinavian
mountains.
The temperature
evolutionduring
the last hour of the trajectoryis shownin Figure4a. After a
succession
of cooling-warming
cyclesof 5 K amplitude,the
maximum cooling inducedby the wave is 10 K, which occurs
a few minutes before the measurements,
and Tice-1 K is
reached.Figure 4b shows the evolution of particle surface
areasduringthe lee wave event.When Ti•e-1 K is reached,a
small amountof the largestSTSparticlesfreezesinto PSC II.
10 2
100
Sincethe Tice-3 K thresholdis not reached,someparticles
remainliquid. Ice particlesgrow by condensationof water.
Then, a warming occurswhich makesthe ice evaporate,
leaving NAT particles.A cooling occursagain, but the
10 -2
temperature stays too high to make NAT grow by
condensation
of H20 andHNO3.SAT beginto appearbecause
of NAT evaporationdue to warmtemperatures.
Finally, the
10 -4
distribution computedat the location of the measurement
is
10.00
100.0
0.001
0.010
0.1 O0
1.000
composedof liquid particles,NAT, and low concentrationof
radius (gm)
SAT particles.The overall size distribution resulting fi'om
model calculations(Figure 5) presentsa bimodal structure. Figure 5. February 25, 1997: size distribution measuredby
This bimodal structureis in good agreementwith the the OPC at the 500 K level (stars) and computed by
I
measuredspectrumof OPC. The first mode centeredon a radius
of 0.013 I,trn is mainly composed
of liquid particles(with a
I
I
I
I
II
MiPLaSMO (solid line). Size distribution at initialization
usedfor the simulationis also presented(dashedline).
6850
RIVII•RE ET AL.: LEE WAVES AND FORMATION OF SOLID PSCs
AMON
30
25
'5
Wavelength
February
26,
1997 -----375
nm
•i'•,
'
400 nm
425 nm
.
.........
45Ohm
_
20 '
.........
.....
•'
..........
475 nm
- ....
500 nm
...............
550nm _.....
575 nm
60Onto
625nm.
15
_
675 nm
-
700 nm
-
1(
).0
0.0005
0.001
0.0015
ExtinctionCoefficient(km4)
Plate 1. February26, 1997:verticalprofilesof theaerosolextinction
coefficient
obtainedwith AMON in the
nearUV andvisible domain(from 375 nm to 700 nm).
by AMON in the layer centered around 22 kin. These facts
indicate that frozen aerosols are probably present in the
correspondinglayer.
5.1.2. Chemical trace species measurements of LPMA.
At the level of the layer of solid particlesmeasuredby AMEN,
the LPMA infrared remote sensingmeasurements
of HNe3 and
HC1 were 11.3ñ0.6 ppbv and 0.24ñ0.11 ppbv, respectively.
supercooling
threshold,makingall the PSC Ib particlesfreeze
into ice. Then the warming occurs, whereupon the ice
evaporates,and NAT beginsto appear.NAT also evaporates
in this warming,sincethe temperatureis above TNAT.
The air
5.2. Synoptic Situation
until it reaches either the SAT melting temperatureor SAT
The modelwas initialized in the sameway as describedin
section 4.2. We calculate the temperaturehistory of the air
parcelfrom a 20-day back trajectoryat the 520 K level (Figure
6), corresponding to 21.6 km (33 hPa), where the extinction
coefficient for the larger wavelength is maximum. The
temperatureat the locationof the measurementis 197 K, which
is 11 K above Tice and more than 3 K above T•4^T.This is too
warm to associatethis solid layer with the presenceof a PSC.
Furthermore, the coldest temperatureencountered along the
trajectory is more than 4 K above the ice frost point (210
hours before the measurement),so we conclude from the
synoptictemperaturehistory that no PSC should have formed
because of the warm temperatures. Furthermore, no solid
particles are predicted by the model from the synoptic
temperaturehistory.
5.3. MesoscaleInterpretation
Mountain-wave model calculations were performedfor the
two times when the sampled air parcel crossed mountain
ranges.On February 26 over the Scandinavian mountains,the
maximum perturbationamplitudeencounteredby the air parcel
is 5 K. No solid particles can form, since the lowest
temperatures are still 5 K above Tice. However, 92 hours
earlier (February 23), the trajectories indicate that this air
parcel was again over the Scandinavian mountains and the
predictedlee wave activity was much stronger.Figures 7a and
7b show the mesoscaletemperatures(computedas before) and
the evolution of particle surface areas during this period,
respectively, 92 hours prior to sampling by AMEN. A 13 K
strong cooling was first predicted, which exceeds the
coolsagainto 188 K, 2 K abovethefrostpoint,rulingout the
possibility
of ice formationfromheterogeneous
nucleationon
remainingNAT. The remainingNAT particles evaporate,
leavingSAT.The modelassumes
that SAT can remainfrozen
deliquescence
temperature.
Up to the location of the
measurements,
thosethresholds are never reachedat synoptic
scale (Figure 6). Moreover, when the air parcel crosses
Scandinaviajust before the AMeN
measurements,
the
aforementionedmesoscaleprocessesinduce cooling and
warmingevents,but the maximum
cooling inducedis more
than 2 K above the SAT deliquescencetemperature.The
maximum
warmingeventsare 9 K below the SAT melting
220
Tmelt
210
•
Tair
ß
• 200
E
190 •
180
•
-'
...........................
Tice
...................
-500
-400
. .........
-300
• .........
-200
. .........
-100
I
0
time (h)
Figure6. Synoptic
temperature
timeevolutionforthe 20-day
backtrajectory(heavy solid line) at the 520 K level for
February26, 1997. SeeFigure 3 caption for the temperature
labels.
RIVII•RE ET AL' LEE WAVES AND FORMATION OF SOLID PSCs
215
Tmel
t
.......................
a
2O5
_
,Tair
-_-
TNAT
195
(i.e., CIO + 2 C1202)is equal to 0.2 ppbv, indicating a small
chlorineactivationof the air parcel. Around 210 hours before
the measurements,
the synoptic temperaturewas 3-4 K above
the ice frost point (Figure 6), triggering growth of liquid
particles by absorption of nitric acid and water. This
correspondsto an increasein the chlorine activation as seen
by an increaseof the calculatedCIOxmixingratio at night and
a calculated decrease of HCI of 0.36 ppbv. This period
correspondsto a slight increaseof nitric acid of 0.38 ppbv.
Indeed, the cumulative effectsof large particle size with low
temperatureand the liquid phaseof the particles induce rather
large rate constants of heterogeneousreactions and thus,
Tdeliq
-II•/.............
......................................................
185
': ,...........................................
, Tice
additional
175
.......
'
'
'
i
i
b
•"
6851
?'"':::'i,
Ice
100.
chlorine
CIO concentration
E
activation.
As the first
lee wave
event
occurs over the Scandinavian mountains (at around t =-91
hours),the air parcel is already activated with CIO nighttime
valuesof 0.18 ppbv and CIOx values of 0.68 ppbv. After this
event, differencesof more than a factor of 2 appear in daytime
between
the cases with
and without
lee
• 10.0
15
INAT
•
1.00
' ' ' IV•uht'ain'w'a•e'
'' •
HNO
3
LiquidI•I:[:
SAT
...........................................................................................
..........................................................................
...................................................
_
10
-
ZF
• 0.10
_
_
a REP•US
0.01
,, [ ::•::
t......
-91
,
-90
•
-89
....
LPMA
i
....
i
....
i
....
,
....
i
time (h)
HCI
Figure 7. Sameas in Figure 4 but for February26, 1997 at 52 0
K, 92 hours beforethe AMON measurements:
(a) temperature
evolutionand (b) time evolutionof particlesurfacearea.
1.5
1.0
0.5
temperature,so SAT particles should remainsolid. When the
air parcelis sampledby the AMON instrument,SAT is the
only kind of particlepredictedby the model.
0.0
nitric acidproductionby heterogeneous
reactions(R1), (R2),
(R3), and (R4) on cloud particles.
The model interpretation of AMON solid particle
measurements
is then related to a previous mesoscalePSC
event, which later evaporation led to the release of frozen
sulfuric aerosolparticles.
I
....
i
....
i
....
....
i
....
,
....
,
,
•
•
,
i
....
•
,
i
•
'
ß --I
2.0
1.5
1.0
while the amount measuredby LPMA was between 11.3+0.6
ppbv at this altitude(Figure 8). If a PSC Ia had been observed,
the amount of HNO3 would have been lower, due to
greater than the amount used as initialization. This is due to
1
....
ClONO
2
Regardingthe chemicalcomposition
of the atmosphere,
the
MiPLaSMO microphysicalresults are compatiblewith the
high HNO3 concentrations
measured
by LPMA [Payan et al.,
1999] at this altitude,ruling out the presenceof a PSC Ia. The
nitricacidcomputed
by themodelaboveKirunais 11.3 ppbv,
condensation
of nitric acid and waterfromthe gasphase.The
HNO• amountcomputedover Kiruna is morethan 1 ppbv
• REP•US
0.5
0.0
'•'
2.0
'-'
1.5
o
-.•
E
1.0
ß•
0.5
E
0.0
-250
-200
-150
-100
-50
0
time (h)
5.4. Implications for Chlorine Activation
Figure 8. Case study for February 26, 1997, at 520 K. Time
evolution for HNO•, HC1, C1ONO2, and C10 mixing ratios,
taking into accountmountain-waveperturbationsstarting at
about t =- 91 hours (dotted gray line), and excluding them
evolution of HNO•, HC1, C1ONO2, and CIO are shown in
(solid line). HNO• and HC1 mixing ratios obtained by
Figure 8 for caseswherelee waves were taken into accountor REPROBUS simulations (triangle) and LPMA measurements
excluded.The initial value (REPROBUS simulations)of C1Ox (gray circle) at 21 km are also shown.
In order to evaluate the impact of lee wave activity on
chlorine activation, model runs were performedwith and
without lee waves for the caseof February26, 1997. Time
6852
RIVII•RE ET AL.' LEE WAVES AND FORMATION OF SOLID PSCs
waves. HC1 is almost completely consumed (0.1 ppbv
Table 2. ComparisonBetween LPMA (1530 UT) or AMON
remaining) by heterogereousreactions,and C1ONO2decreases (2200 UT)Measurements at 21 and 22 km, and MiPLaSMO
Model Results at 21.6 km for HNO3 and HC1.
by 1.25 ppbv for the same reason. The C1ONO2 loss is in
correlation
with
the nitric
acid increase of the same amount
due to reactions (R1) and (R2). The contribution of reactions
(R3) and (R4) for nitric acid production is negligible because
of the very small amountof N2Os in our conditions.It has been
verified that for 10-days integration period using synoptic
temperaturehistory that REPROBUS and MiPLaSMO results
are in rathergoodagreement(seeFigure 8 and Table 1). When
lee waves are introduced into the model, the MiPLaSMO
values(at 21.6 km at the time of the AMON measurements)
of
HNO3 and HC1 (11.3 ppbv and 0.30 ppbv, respectively)are in
accordancewith the measurementsof LPMA on February 26,
1997, at 21 km (11.3+0.6 ppbv and 0.24+0.11 ppbv,
respectively).ComparisonsbetweenMiPLaSMO results and
LPMA
measurements are summarized
between
LPMA
measurements
in Table 2. Differences
and REPROBUS
simulations
could be explained by the fact that REPROBUS is not
accountingfor mesoscaletemperaturefluctuations.
6.
Conclusion
HNO3 (ppbv)
LPMA (21 km)
MiPLaSMO (21.6 km)
with lee waves
(AMON location)
LPMA (22 km)
HCI (ppbv)
11.3+0.6
0.24+0.11
11.3
0.30
11.0+0.6
0.45+0.20
Here, mesoscaleperturbationsdue to lee wave are introducedin
MiPLaSMO.
This PSC would have been generatedby wave perturbations
above the Scandinavian mountains, several days before the
measurements.
Solid aerosol calculations are qualitatively in
accordancewith the high HNO3 concentrationand the low
valuesof HC1 measuredon February26. For this case, we have
also shown that even if the air parcel was already chlorine
activated before the mesoscaleprocessesoccurred, the lee
wave activity has significantly amplified the activation
processand C10 concentrationis more than 2 times greater
than without lee waves in daytime conditions. This result is
in agreementwith the conclusion of Carslaw et al. [1998b].
The aim of the present study was to interpret two sets of
measurements
performedin February 1997, which cannot be
explained by synoptic-scaletemperaturehistories. For this
purpose, we have used a Lagrangian chemical and
microphysical model (MiPLaSMO) and a mountain wave
model (NRL/MWFM). We have shown that, introducing
mesoscale temperature perturbations from NRL/MWFM,
initialization from the REPROBUS model (chemistry) and
OPC measurements(aerosols), with realistic hypotheses
about the supercoolingthreshold for PSCs freezing(i.e., 3 K
belowthe ice frost point), the MiPLaSMO Lagrangianmodel
could produceresults in agreementwith the layer of large
particlesobservedby the opticalparticlecounteron February
25, 1997. This layer can be interpretedas composedof a small
numberof NAT particles(PSC Ia) mixedwith liquid particles.
The maximum
temperatureamplitudeof the lee waves in this
Acknowledgments. The authorsare grateful to K. Carslaw for
providingthem a parameterizationof SAT deliquescencetemperature
and to the CNES launchingteam at Kiruna. The AMON and LPMA
flights were performed within the French ProgrammeNational de
Chimie Atmosph6riqueand with financial supportfrom CNES and
INSU. The GenevaObservatoryteam is acknowledgedfor the AMON
andLPMA gondolaintegrationandpointing.We would like to thank the
NILU database for access to meteorologicaldata, as well as B.
Knudsen and F. Fierli for the trajectory calculation program. S.
Eckermann acknowledges support from NASA ACMAP grant
simulation
L68786D.
was 10 K.
Chemical results of MiPLaSMO are in accordance with HNO3
and HC1 LPMA
measurements. Furthermore, mesoscale
processescan be an explanationfor the differencesobserved
by Payan et al. [1999] between the REPROBUS model and
LPMA
measurements.
On February26, the observationof solid particles can be
interpretedwith MiPLaSMO and the mountain-wavemodel. References
Although the mountain-wave model is considered to
overestimatethe wave amplitude[Carslaw et al., 1999], the Bacmeister, J.T., P.A. Newman, B.L. Gary, and K.R. Chan, An
algorithmfor forecastingmountainwave-relatedturbulencein the
13 K lee wave temperatureamplitudeused in our model is
stratosphere,
WeatherForecasting,9, 241-253, 1994.
neverthelessstill compatible with the observational case
Biermann, U.M., J.N. Crowley, T. Huthwelker, G.K. Moortgat, P.J.
studiedby Carslaw et al. [1998a]. What is observedcould be
Crutzen, and T. Peter, FTIR studieson lifetime prolongationof
solid sulfuric acid aerosol, evidence of a previous PSC event.
stratospheric
ice particlesdue to NAT coating,GeophysRes.Lett.,
25, 3939-3942, 1998.
Table 1. ComparisonBetween REPROBUS at Two Standard
PressureLevels (21 km and 21.9 km) and MiPLaSMO (21.6
km) Model Resultsfor HNO3 and HC1.
REPROBUS(LPMA location)
(21 kin)
MiPLaSMO(21.6 km)
without lee waves
(AMON location)
REPROBUS(LPMA location)
(21.9 kin)
HNO3 (ppbv)
HCI (ppbv)
9.7
1.00
10.3
0.96
9.6
1.30
Temperatureandpressureperturbations
are not takeninto account.
Camy-Peyret,C., Balloon-borneFourier transform spectroscopyfor
measurements
of atmospherictrace gases,Spectrochirn.
Acta, 51A,
1143-1152, 1995.
Carleton,K.L., D.M. Sonnenfroh,W.T. Rawlins, B.E. Wyslouzil, and S.
Arnold,Freezingbehaviorof singlesulfuricacid aerosolssuspended
in a quadrupole
trap,d. Geophys.Res.,102, 6025-6030, 1997.
Carlotti, M., Global fit approach to the analysisof limb-scanning
atmospheric
measurements,
Appl. Opt.,27, 3250-3254, 1988.
Carslaw,K.S. and T. Peter,Uncertaintiesin reactiveuptake coefficients
for solid stratosphericparticles, 2, Effect on ozone depletion,
Geophys.Res.Lett., 24, 1747-1750, 1997.
-Carslaw, K.S., M. Wirth, A. Tsias, B.P. Luo, A. D6rnbrack, M.
Leutbecher,H. Volkert, W. Renger,J.T. Bacmeister,and T. Peter,
Particlesmicrophysics
andchemistryin remotelyobservedmountain
polarstratospheric
clouds,d. Geophys.
Res.,103, 5785-5796,1998a.
Carslaw, K.S., et al., Increased stratosphericozone depletiondue to
mountain-induced
atmospheric
waves,Nature,391,675-678, 1998b.
RIVII•RE ET AL.' LEE WAVES AND FORMATION OF SOLID PSCs
6853
Carslaw, K.S., T. Peter, J.T. Bacmeister, and S.D. Eckermann,
Payan, S., C. Camy-Peyret,P. Jeseck,T. Hawat, G. Durry, and F.
Lefevre, First simultaneousHC1 and C1ONO2 profile measurements
in the Arctic vortex,Geophys.Res.Letr, 25, 2663-2666, 1998.
?ayan,S., C. Camy-Peyret,
P. Jeseck,T. Hawat, M. ?irre, J.-B. Renard,
C. Robert,F. Lefevre, H. Kansawa, and Y. Sasano,Diurnal and
nocturnaldistributionof stratospheric
NO2 from solar and stellar
1997.
occultation measurements:comparison with models and ILAS
De More, W.B., S.P. Sander,D.M. Golden,R.F. Hampson,M.J. Kurylo,
satellitemeasurements,
d. Geophys.Res.,104, 21,585-21,593,1999.
C.J. Howard, A.R. Ravishankara, C.E. Kolb, and M.J. Molina,
Press, W.H., S.A. Teukolsky, W.T. Vetterling, and B.P. Flannery,
Chemicalkineticsand photochemical
data for use in stratospheric
NumericalRectpesin Fortran, 2nd ed., 963 pp., CambridgeUniv.
modeling,Evaluation12,dPL Publ., 97-4, 1997.
Press,New York, 1992.
Deshler, T., B.J. Johnson,and W.R. Rozier, Balloonborne measurements Ramaroson,R.A., M. Pirre, and D. Cariolle, A box model for on-line
of Pinatuboaerosolduring1991and 1992 at 41øN:Vertical profiles,
computationof diurnal variation in a 1-D model: Potentialfor
size distribution,and volatility,Geophys.Res. Lett., 20, 1435-1438,
application
in multi-dimentionnal
cases,
Ann. Geophys.,10, 416-428,
Widespreadsolidparticleformationby mountainwavesin the Arctic
stratosphere,
d. Geophys.Res., 104, 1827-1836, 1999.
Clapp, M.L., R.F. Niedziela, L.J. Richwine, T. Dransfield, R.E. Miller,
and D.R. Worsnop, Infrared spectroscopyof sulfuric acid/water
aerosols:Freezingcharacteristics,
d. Geophys.Res., 102, 8899-8907,
1993.
1992.
Eckermann,S.D., and J.T. Bacmeister,Global parameterizationof
Ravishankara,
A. R., and D. R. Hanson,Differencesin the reactivityof
gravity wave temperature perturbations for chemical and
type I polar stratosphericcloudsdependingon their phase, d.
microphysicalmodels,paper presentedat the MEPS workshop,
Geophys.Res., 101, 3885-3890, 1996.
European CommissionEnvironnementand Climate Programme, Renard, J.B., M. ?irre, C. Robert, G. Moreau, D. Huguenin, and J.M.
BMBF POLECAT, DLR, andSPARC, Bad T61z,Nov. 8-11, 1998.
RussellIII, Nocturnal vertical distributionof stratospheric03, NO2
Eckermann,S.D., D.E. Gibson-Wilde,and J.T. Bacmeister,Gravity
and NO3 from balloon measurements,
d. Geophys.Res., 101, 28,793wave perturbationsof minor constituents:A parcel advection
28,804, 1996.
methodology,
ar.Atmos.Sci., 55, 3521-3539, 1998.
Rosen,J. M., The vertical distributionof dustto 30 km, d. Geophys.Res.,
Holmann,D.J., andT. Deshler,Stratospheric
cloudobservations
during
69, 4673-4676, 1964.
formationof the Antarcticozoneholein 1989, d. Geophys.Res., 96,
Sasano,Y., M. Susuki,T. Yokota, and H. Kanzawa, Improved Limb
2879-2912, 1991.
AtmosphericSpectrometer(ILAS) for stratospheric
ozone layer
Huret,N., E.D. Rivi•re,F. Taupin-Goffinont,
and M. Pirre, Lagrangian
measurements
by solar occultationtechnique,Geophys.Res. Lett.,
studyof polar stratospheric
clouds:The role of lee wave eventsfor
26, 197-200, 1999.
cloudsformationandchlorineactivationat the endof February 1997 Solomon,S., R.R. Garcia, F.S. Rowland, and D.J. Wuebbles, On the
closeto Kiruna, paperpresentedat the MEPSworkshop,European
depletionof theAntarcticozone,Nature,321, 755-758,1986.
CommissionEnvironnement and Climate Programme, BMBF Tabazadeh, A., and O. B. Toon, The presence of metastable
POLECAT, DLR, and SPARC, Bad TOlz,Nov. 8-11, 1998.
HNO3/H20 solidphasesin the stratosphere
inferredfrom ER 2 data,
Iraci, L.T., T.J. Fortin, and M.A. Tolbet, Dissolutionof sulfuric acid
d. Geophys.Res., 101,9071-9078, 1996.
tetrahydrateat low temperaturesand subsequent
growth of nitric Tabazadeh,A., R.P. Turco, K. Drdla, M.Z. Jacobson,and O.B. Toon, A
acidtrihydrate,d. Geophys.Res.,103, 8491-8498, 1998•
studyof Type I polar stratospheric
cloudformation,Geophys.Res.
Knudsen,B.M., and G.D. Carver, Accuracy of the isentropic
Lett., 21, 1619-1622, 1994.
trajectoriescalculatedfor the EASOE campaign,Geophys.Res. Tolbert, M. A., Polar clouds and sulfate aerosols, Science, 272,
Lett., 21, 1199-1202, 1994.
Koop, T. and K. Carslaw, Melting of H2SO4.4H20 particlesupon
cooling:Implicationsfor polar stratosphericclouds,Science, 272,
1638-1641, 1996.
Koop, T., U. M. Biermann, W. Raber, B. P. Luo, P. J. Crutzen and T.
Peter, Do stratosphericaerosoldropletsfreeze above the ice frost
point?,Geophys.Res.Lett., 22, 917-920, 1995.
Koop, T., K.S. Carslaw and T. Peter, Thermodynamicstability and
phasetransitionsof ?SC particles, Geophys.Res. Lett., 24, 2199-
1597,1996.
Tsias, A., A. J. Prenni, K. S. Carslaw, T. P. Onasch, B. P. Luo, M. A.
Tolbert, and T. Peter, Freezing of polar stratosphericclouds in
orographically
inducedstrongwarmingevents,Geophys.Res. Lett.,
24, 2303-2306, 1997.
Zhang,R., M.T. Leu, andM.J. Molina,Formationof polar stratospheric
cloudson preactivated
background
aerosols,Geophys.Res. Lett.,23,
1669-1672, 1996.
C. Camy-?eyretand S. ?ayan, Laboratoire de PhysiqueMoleculaire
et Applications,CNRS, 4 Place Jussieu,Bte 76, 75252 Paris cedex 05,
France.(camy•ccr.jussieu.fr;payan•ccr.jussieu.fr)
T. Deshler, Department of Atmospheric Science, University of
Wyoming, 16th and Gibbon, Laramie, WY 82071. (deshler•
grizzly.uwyo.edu)
S. D. Eckermann,E. O. Hulburt Center for Space Research,Naval
Research Laboratory, Mail code 7641, Washington, DC 20375.
(eckerman•ismap4.nrl.navy.mil)
N. Huret, M. Pirre, J.-B. Renard, E. D. Riviere, and F. G.-Taupin,
Res., 100, 11275-11288, 1995.
Meilinger, S.K., T. Koop, B.P. Luo, T. Huthwelker, K.S. Carslaw, U. Laboratoire de Physiqueet Chimie de l'Environnement,CNRS, 3A,
Krieger, P.J.Crutzen and Th. Peter, Size dependentstratospheric Avenue de la RechercheScientifique,45071 Orleans cedex 2, France.
fr,
mpirre•cnrs-orleans.
fr,
renard•cnrsdropletcompositionin lee wave temperaturefluctuationsand their (nhuret•cnrs-orleans.
potentialrole in PSC freezing, Geophys.Res. Lett., 22,. 3031-3034, orleans.fr,riviere•cnrs-orleans.fr,goffinon•cnrs-orleans.fr)
1995.
N. Larsen,Departmentof Research,DanishMeteorologicalInstitute,
Denmark.(nl•dmi.dk)
Middlebrook,A.M., et al., Formationof modelpolar stratospheric
cloud Lyngbyvej100, DK-2100, Copenhagen,
F. Lefevre, Service d'Aeronomie du CNRS, Universite Pierre et
film, Geophys.Res.Letr, 19, 2417-2420, 1992.
Naudet, J.P., C. Robert, and D. Huguenin,Balloon measurementsof Marie Curie, BP 102, 4 place Jussieu,75252 Paris cedex 05, France.
stratospheric trace species using a multichanel UV-visible (franck.1efevre• aero.jussieu.fr)
spectrometer,
proceedings
of the 14thESA symposium
on European
rocket and balloon programsresearched,Montreux, European (ReceivedMay 4, 1999; revisedJuly 19, 1999;
acceptedAugust23, 1999.)
SpaceAgencySpec.Publ. 355, 165-168, 1994.
2202, 1997.
Larsen, N., B. Knudsen,J. Rosen,N. Kjome, R. Neuber and E. Kyr6,
Temperaturehistoriesin liquid and solid polar stratosphericcloud
formation,d. Geophys.Res.,102, 23505-23517, 1997.
Lef•vre, F., F. Figarol,K.S. Carslaw, and T. Peter, The 1997 Arctic
ozone depletion quantified from three-dimensional model
simulations,Geophys.Res.Lett., 25, 2425-2428, 1998.
MacKenzie, A.R., M. Kulmala, A. Laaksonen,and T. Vesala, On the
theoriesof type 1 polar stratospheric
cloudformation,d. Geophys.