can gravity waves signiﬁcantly impact psc occurrence in...

Atmos Chem Phys 9 8825ndash8840 2009wwwatmos-chem-physnet988252009copy Author(s) 2009 This work is distributed underthe Creative Commons Attribution 30 License

AtmosphericChemistry

and Physics

Can gravity waves significantly impact PSC occurrence in theAntarctic

A J McDonald S E George and R M Woollands

Department of Physics and Astronomy University of Canterbury Private Bag 4800 Christchurch New Zealand

Received 11 December 2008 ndash Published in Atmos Chem Phys Discuss 3 February 2009Revised 14 October 2009 ndash Accepted 25 October 2009 ndash Published 23 November 2009

Abstract A combination of POAM III aerosol extinctionand CHAMP RO temperature measurements are used to ex-amine the role of atmospheric gravity waves in the formationof Antarctic Polar Stratospheric Clouds (PSCs) POAM IIIaerosol extinction observations and quality flag informationare used to identify Polar Stratospheric Clouds using an un-supervised clustering algorithm

A PSC proxy derived by thresholding Met Office tem-perature analyses with the PSC Type Ia formation tempera-ture (TNAT) shows general agreement with the results of thePOAM III analysis However in June the POAM III obser-vations of PSC are more abundant than expected from tem-perature threshold crossings in five out of the eight years ex-amined In addition September and October PSC identifiedusing temperature thresholding is often significantly higherthan that derived from POAM III this observation probablybeing due to dehydration and denitrification Comparisonof the Met Office temperature analyses with correspondingCHAMP observations also suggests a small warm bias in theMet Office data in June However this bias cannot fully ex-plain the differences observed

Analysis of CHAMP data indicates that temperature per-turbations associated with gravity waves may partially ex-plain the enhanced PSC incidence observed in June (relativeto the Met Office analyses) For this month approximately40 of the temperature threshold crossings observed usingCHAMP RO data are associated with small-scale perturba-tions Examination of the distribution of temperatures rela-tive toTNAT shows a large proportion of June data to be closeto this threshold potentially enhancing the importance ofgravity wave induced temperature perturbations Inspectionof the longitudinal structure of PSC occurrence in June 2005also shows that regions of enhancement are geographically

Correspondence toA J McDonald(adrianmcdonaldcanterburyacnz)

associated with the Antarctic Peninsula a known mountainwave ldquohotspotrdquo The latitudinal variation of POAM III obser-vations means that we only observe this region in JunendashJulyand thus the true pattern of enhanced PSC production maycontinue operating into later months

The analysis has shown that early in the Antarctic win-ter stratospheric background temperatures are close to theTNAT threshold (and PSC formation) and are thus sensitive totemperature perturbations associated with mountain wave ac-tivity near the Antarctic peninsula (40 of PSC formation)Later in the season and at latitudes away from the peninsulatemperature perturbations associated with gravity waves con-tribute to about 15 of the observed PSC (a value which cor-responds well to several previous studies) This lower valueis likely to be due to colder background temperatures alreadyachieving theTNAT threshold unaided Additionally there isa reduction in the magnitude of gravity waves perturbationsobserved as POAM III samples poleward of the peninsula

1 Introduction

The role of PSCs in polar ozone depletion was first identi-fied over twenty years ago (Solomon et al 1986) Hetero-geneous chemical reactions on PSC particles are the centralprocess in chlorine activation and consequent ozone deple-tion In addition to the activation of chlorine from reservoirsof HCl and ClONO2 PSCs can also remove NOy (ie thetotal odd nitrogen) from the lower stratosphere by incorpora-tion of HNO3 from the gas phase and subsequent sedimen-tation This denitrification leads to a slower conversion ofactive chlorine back to the reservoir species ClONO2

In contrast to the Northern Hemisphere very low strato-spheric temperatures in the Antarctic winter generally lead toan abundance of PSCs in the lower stratosphere (Poole andPitts 1994) Three categories of PSCs are generally recog-nized Type Ia composed of solid phase nitric acid trihydrate

Published by Copernicus Publications on behalf of the European Geosciences Union

8826 A J McDonald et al PSC occurrence in the Antarctic

(NAT) Type Ib composed of supercooled ternary solution(STS a liquid mixture of H2O HNO3 and H2SO4) andType II composed primarily of ice (seeLowe and MacKen-zie 2008 and references therein) These three categoriesof PSC form at different temperatures in the stratosphereIn particular the NAT PSC formation temperature (TNAT)is dependent on pressure nitric acid and water vapour con-centrations (Hanson and Mauersberger 1988) and is signifi-cantly warmer than the ice frost point (TICE) The STS for-mation temperature (TSTS) occurs approximately 35 K be-low the value ofTNAT (Carslaw et al 1994) and the ice frostpoint temperature can be derived by using the formulationdescribed inMarti and Mauersberger(1993) Temperaturesare cool enough in the Antarctic stratosphere for all threetypes of clouds to be present during the winter The presentstudy examines the potential contribution of internal grav-ity waves to PSC formation in the Antarctic stratosphere Inthe Arctic the temperature perturbations associated with in-ternal gravity waves particularly mountain waves have beenshown to be responsible for the formation of a significantproportion of PSCs (Carslaw et al 1998 Hopfner et al2001) However the observed relationship between Antarc-tic PSC formation and low background temperatures has ledto less focus on wave induced PSC formation processes inthe Antarctic

Recent studies have indicated that the temperature per-turbations related to these waves may have a more signifi-cant effect on Antarctic PSC occurrence and have a poten-tially important role in ozone depletion (Shibata et al 2003Hopfner et al 2006 Innis and Klekociuk 2006 Eckermannet al 2009) Such effects may be particularly significantin early winter where temperatures hover close to PSC for-mation temperature thresholds Analysis byHopfner et al(2006) has identified a significant role of mountain wavesin the formation of PSC and subsequent denitrification inthe Antarctic Hopfner et al(2006) described spaceborneinfrared measurements of a vortex-wide area of NAT PSCaround Antarctica using the Michelson Interferometer forPassive Atmospheric Sounding (MIPAS) instrument Sim-ulations for the period 10ndash12 June 2003 indicate the obser-vations are likely to be associated with heterogeneous nucle-ation on ice in the cooling phases of large-amplitude strato-spheric mountain waves over the Antarctic Peninsula Theprocess operating here is the same mountain wave NAT for-mation model originally proposed byCarslaw et al(1998)for the Arctic winter stratosphere Recent work byEcker-mann et al(2009) examined the same MIPAS satellite obser-vations over Antarctica with additional Atmospheric InfraredSounder (AIRS) data Their study showed that large moun-tain wave temperature perturbations produced NAT which isadvected to form a circumpolar NAT outbreak which concurswith Hopfner et al(2006)

Other studies have also examined the role of gravity waveson PSC formation For example a ground-based lidar studyby Shibata et al(2003) indicated that Type II PSCs occurred

when stratospheric temperatures dropped belowTICE be-cause of perturbations associated with inertia gravity wavesAnother lidar study byInnis and Klekociuk(2006) also es-timated the influence of gravity waves on PSC occurrenceThey concluded that during the ldquoPSC seasonrdquo when the back-ground temperature was close enough to the NAT formationtemperature (TNAT) derived inHanson and Mauersberger(1988) gravity wave perturbations influence PSC formationapproximately 15 of the time Interestingly this value issimilar to the findings ofFelton et al(2007) who concludedType Ia enhanced PSCs which are likely to be NAT crys-tals with unusually large lidar backscattering ratios which ap-pear down wind of mountain wave-induced ice clouds (Tsiaset al 1999) make up 11 of the clouds observed during theSAGE III Ozone Loss and Validation Experiment (SOLVE)campaign in winter 2000 in the Northern Hemisphere How-ever the warmer mean temperature in the Arctic means thatthis is likely to be coincidental It is also worthy of mentionthat Innis and Klekociuk(2006) found a clear relationshipbetween planetary wave temperature perturbations and PSCoccurrence on the edge of the Antarctic continent Recentanalysis using data from CALIPSO (a space-borne lidar) hasalso mentioned the potential importance of gravity waves onPSC formation (Pitts et al 2007 Noel et al 2008) How-ever enhanced PSC formation has also been suggested to berelated to synoptic-scale motions (Teitelbaum et al 2001Wang et al 2008)

Previous work has shown that NAT PSC formation may beimplicitly linked to gravity wave temperature perturbationsIn particular homogeneous NAT nucleation does not occurdirectly on temperatures crossingTNAT Rather average tem-peratures need to be below theTNAT threshold for a periodgreater than a few days for nucleation to occur (Peter 1997)Carslaw et al(1994) discusses NAT formation and indicatesthat NAT may form by heterogeneous nucleation on pre-existing ice particles which may form because of large tem-perature perturbations produced by mountain waves Workby Fueglistaler et al(2002) suggests that early season PSCproduction associated with Type Ia or Ia enhanced cloudsproduced using the mechanism proposed inCarslaw et al(1994) can potentially be extremely important in denitrifi-cation due to their role in the formation of extremely largeNAT particles if number densities within the clouds are highHowever recent work has additionally identified a need forNAT nucleation mechanisms which are independent of theexistence of ice particles only requiring temperatures belowthe existence temperature for NAT which is significantly eas-ier (approximately 7ndash8 K warmer thanTICE) to meet than thedemand for pre-existing ice particles

Voigt et al(2005) detailed measurements where the con-ditions of particle formation are well enough constrained toconclude that a PSC is observed in air parcels that spent lessthan a day (approximately 18 h) at temperatures no lowerthan 3K belowTNAT Hitchman et al(2003) also suggeststhat a formation mechanism for Type Ia PSC must exist

Atmos Chem Phys 9 8825ndash8840 2009 wwwatmos-chem-physnet988252009

A J McDonald et al PSC occurrence in the Antarctic 8827

without the requirement for temperatures significantly be-low TNAT Their study also highlights the potential impor-tance of non-orographic gravity waves on PSC formation inthe Arctic Pagan et al(2004) conclude that Type Ia PSCscan nucleate in relatively warm synoptic-scale temperaturefields and regions of strong mountain-wave activity with suf-ficient cooling to produce ice nuclei are not required In ad-dition Irie et al(2004) used ILAS trace gas and aerosol ex-tinction measurements AVHRR stratospheric ice cloud mea-surements and a microphysical box model to investigate pro-cesses leading to denitrification in the Arctic vortex in Febru-ary 1997 Their analysis suggest that sedimentation of NADor NAT particles formed through NAD freezing on aerosolsurfaces can cause significant denitrification in the Arcticvortex Their analysis combined with the results inPaganet al(2004) seem to confirm that ice particle surfaces are nota prerequisite for the formation of nitric acid hydrate PSCsReview of these and other studies detailed in theWorld Me-teorological Organization(2007) report indicates increasingevidence that processes occur above the ice frost point How-ever the report indicates that whether these processes are ap-plicable in the Antarctic requires investigation

This study examines the relationship between PSC occur-rence observed using POAM III observations and the occur-rence expected based on Met Office analyses and CHAMPradio occultation measurements The addition of high ver-tical resolution CHAMP temperature observations providesthe possibility of identifying periods and spatial regionswhere PSC formation associated with gravity wave inducedtemperature perturbations may occur Though it should benoted that the CHAMP instrument in a similar manner toall limb viewing instruments can only observe a portion ofthe gravity wave spectrum Comparison of the observationsfrom Met Office analyses and CHAMP observations also al-lows the sensitivity of PSC occurrence to temperature biasesin Met Office analyses to be examined In order to exam-ine the importance of temperature and the impact of gravitywave temperature perturbations we determine the probabilitythat a region is below various threshold temperatures relativeto the frequency of PSC observations made with POAM IIIThis method has previously been used by many authors in-cludingPoole and Pitts(1994) andSaitoh et al(2006)

In this study Section2 describes the POAM III andCHAMP RO observations and also details a new unsuper-vised clustering algorithm which is applied to POAM IIIextinction measurements to identify PSC Comparisons ofthe observed probability of PSC occurrence relative to thosebased on the occurrence of temperatures below theTNAT TSTS andTICE thresholds derived from Met Office analysesand CHAMP RO data are detailed in Sect3 and the interpre-tation of these results is discussed in Sect4

J F M A M J J A S O N D Jminus90

minus85

minus80

minus75

minus70

minus65

minus60

Date

Latit

ude

Fig 1 Seasonal variation of observation latitude measured byPOAM III

2 Dataset and methodology

21 POAM III observations and the PSC identificationalgorithm

The POAM III instrument made observations of the uppertroposphere and stratosphere between its launch on the Satel-lite Pour lrsquoObservation de la Terre (SPOT) 4 satellite inMarch 1998 and the end of 2005 POAM measures watervapor aerosol extinction ozone and nitrogen dioxide usingthe solar occultation limb sounding technique (Lucke et al1999) The high inclination of the SPOT 4 satellite allows oc-cultations to be performed in both the northern and southernpolar regions It should be noted that the POAM III sam-pling volume is about 200 km long 30 km wide and 1 kmthick and thus this data set will not capture very small-scalefeatures The POAM III instrument provides 12 to 14 pro-files per day in each hemisphere around a circle of latitudeThe latitude of the sampled region varies throughout the yearand varies from roughly 63 to 88 in the Southern Hemi-sphere (see Fig1) This latitudinal variation of the solar oc-cultation satellite observations implies an inherent samplingissue when attempting to separate the impact of latitudinaland temporal variations in this dataset For examplePittset al(2007) has shown that sampling CALIPSO data at SAMII latitudes moves the maximum PSC occurrence date fromearly August to September due to the SAM II sampling be-ing at the highest latitude sampled The fact that both SAMII and POAM III have similar sampling patterns means thatissues associated with sampling bias need to be consideredResearch completed using POAM III observations has fo-cused on examinations of the spatial and temporal variabilityof stratospheric ozone aerosols polar stratospheric cloudsand polar mesospheric clouds

wwwatmos-chem-physnet988252009 Atmos Chem Phys 9 8825ndash8840 2009


In this work PSC occurrence is derived from POAM IIIextinction measurements at 1018microm and the ratio of the1018 to the 0603microm extinction channels using a k-meansunsupervised clustering algorithm described in subsequentparagraphsJakob and Tselioudis(2003) have previously uti-lized a simple clustering algorithm for cloud regime identifi-cation in the Tropical Western Pacific region and recent workby Felton et al(2007) has applied a clustering algorithm tolidar observations of PSC We use a similar method to thatindicated inJakob and Tselioudis(2003) to separate differ-ent types of PSC Our algorithm can be used to identify bothType Ia and Ib aerosols using the methodology described inStrawa et al(2002) except that the separation is based on anunsupervised clustering algorithm However in this study wedo not differentiate between Type Ia and Ib PSC The separa-tion into different types will be left to a future study It shouldbe noted that this method provides similar PSC identificationability to algorithms described byFromm et al(2003)

The statistical method of cluster analysis is used to sep-arate background aerosol extinctions from those associatedwith PSCs As its name suggests cluster analysis searchesfor possible ldquoclustersrdquo in a data set by evaluating some mea-sure of distance between individual data points (in this studywe use Euclidean distance in extinction ratioextinction spaceto separate PSC and non-PSC observations and then an an-gular measure can be used to separate PSC Type Ia and Ibbut is not utilized in this study) It should be noted thatto ensure that neither the extinction at 1018microm or the ra-tio of 1018 to the 0603microm extinction channels dominatein the clustering procedure the values are normalized Notethat in this case a ldquodata pointrdquo is an individual set of mea-surements of aerosol extinction from POAM III at a specificaltitude and geographical location The k-means clusteringalgorithm used in this study iteratively searches for a pre-defined number (k) of clusters two in this case using thefollowing scheme

1 k elements of the data set of size N are selected at ran-dom as distinct cluster members

2 each of the remaining N-k elements are assigned to thecluster with the nearest centroid (based on Euclideandistance in extinction ratioextinction) whereby aftereach assignment the centroid of the cluster is recalcu-lated and

3 after all elements have been assigned the centroidsfound in step 2 are used as new seed points and the al-gorithm is iterated

Normally fewer than ten iterations are sufficient for theconvergence of the algorithm Note that the random selectionof data points indicated in the algorithm does not work onobservations from the Arctic because of the small number ofobservations which include PSC This is because it becomesdifficult to identify physically meaningful clusters initially at

random Thus someapriori knowledge to select likely PSCswould be required For example a methodology similar toFromm et al(2003) could be used

It should be noted at this point that other studies (Frommet al 1997 2003) have indicated that the high extinctions as-sociated with Type II PSC can cause the termination or com-mencement of occultation events at anomalously high alti-tudes This is caused by the PSC having large opacity andorsize preventing POAM III from measuring lower Thus theextinction measurements are most useful for identifying TypeI PSCs and will undersample Type II PSCs To reduce thisbias we then utilize quality flags often referred to as ZMINflags to identify profiles with anomalously high commence-ment of occultation events and include these in the statisticsshown later

22 CHAMP RO and Met Office observations

This study also employs data from the radio occultation (RO)experiment onboard the CHAMP satellite available from2002 to 2005 In this experiment GPS radio signal bend-ing due to atmospheric refractivity is measured to determineprofiles of atmospheric parameters The data was processedand provided by the GeoForschungsZentrum (GFZ) Pots-dam Version 005 were used for this study Specific infor-mation on the processing method used is detailed inWick-ert et al(2004) The CHAMP satellite has an inclinationof 87 and therefore the data distribution over the globe isnearly uniform Inversion of bending angle information al-lows the instrument to produce approximately 150ndash200 pro-files of dry temperature each day over the globe The drytemperature is the temperature that results from an inversionprocess that assumes that the radio refractive index variationobserved by radio occultation satellites is related to tempera-ture variations only and this is a very good assumption in thedry stratosphere Dry temperature therefore represents thetrue atmospheric temperature very accurately in this region

From Fresnel diffraction theory it can be shown that theprofiles have a true vertical resolution of approximately14 km in the stratosphere However the data is over-sampledand provided by the GFZ at a vertical spacing of 200 m Thehorizontal resolution of an occultation is 200ndash400 km alongthe line of sight (LOS) and on the order of 1ndash3 km acrossthe LOS A large number of studies have described the qual-ity of these temperature measurements In particular com-parisons of CHAMP RO measurements with analyses andradiosonde observations display excellent agreement (Wick-ert et al 2004 Gobiet et al 2005 Parrondo et al 2007Wang and Lin 2007) For exampleWickert et al(2004)compared CHAMP dry temperature with interpolated datafrom ECMWF meteorological analyses with co-located andnear simultaneous radiosonde observations between 10 and35 km Biases between ECMWF and CHAMP RO profilesare less than 05 K while comparison using radiosonde ob-servations and RO profiles indicates nearly no bias between



approximately 9 and 27 km This lack of bias between theobservational data sets suggests deviations of the analysesfrom the real atmospheric situation Work byGobiet et al(2005) which compares ECMWF operational analyses withCHAMP RO data found close overall agreement worldwideapart from discrepancies in the Antarctic stratosphere wherecold biases up tominus25 K and warm biases up to +35 K wereobserved

These and other previous analyses ledWang and Lin(2007) to use COSMIC RO soundings to examine the distri-bution of the coldest atmospheric temperatures in the Antarc-tic stratosphere Their study suggests that these high qual-ity measurements of vertical temperature structure over theAntarctic region are potentially important inputs for mod-els that predict polar stratospheric clouds occurrence andstratospheric ozone depletion (Huck et al 2005) Very re-cent work byJuarez et al(2009) indicated that CHAMPRO observations are able to detect more PSC-prone temper-ature profiles during winters with disturbed conditions in theArctic than the ECMWF analyses Those authors considerthat this is due to the ability of CHAMP RO to detect short-vertical wavelength features which may represent either lo-calized gravity or global-scale planetary waves

In this study CHAMP observations are utilized to quan-tify the magnitude of the background temperature and per-turbations associated with gravity waves However the res-olution of the CHAMP observations has a direct impact onthe observability of different portions of the gravity wavespectrum and this needs to be considered Work byLangeand Jacobi(2003) shows that while the weighting functionof the RO measurements is 200ndash400 km along the LOS theGPS measurements are sensitive to gravity waves with hor-izontal wavelengths greater than 100 km and vertical wave-lengths between approximately 14 and 10 km This is par-tially associated with the LOS having a parabolic path in theatmosphere and also the fact that the plane waves visible willdepend on the angle between the LOS and the wavefrontsWhen the LOS is perpendicular to the wavefronts then spa-tial averaging will significantly affect the wave activity ob-served In particular work byde la Torre and Alexander(2005) has shown that the observability of mountain waveswhich have horizontal wavelengths of tens to hundreds ofkilometers can be significantly impacted Further work byBaumgaertner and McDonald(2007) suggested that the pre-ponderance of LOS within 30 degrees of the northsouth axisin the Antarctic region means that the viewing geometry overthe Antarctic Peninsula and the Trans-Antarctic Mountainspotentially favoured the detection of the mountain waves byCHAMP in these regions However all this previous worksuggests that CHAMP RO measurements observe a limitedportion of the gravity wave field A study detailed inMcDon-ald and Hertzog(2008) which compared coincident CHAMPand in-situ super-pressure balloon measurements of the grav-ity wave field suggests that overall the wave field is such thatCHAMP underestimates the magnitude of the potential en-

H2O

con

cent

ratio

n (p

pmv)

HNO3 concentration (ppbv)

5 6 7 8 9 10 11 12 13 14 152

25

3

35

4

45

5

55

1935

194

1945

195

1955

196

1965

197

1975

198

1985

199

Fig 2 Contour plot of the NAT PSC formation temperature(Kelvin) as a function of nitric acid and water mixing ratios at 18 km(Function based onHanson and Mauersberger(1988))

ergy per unit mass Thus in any analysis CHAMP wouldlikely provide a conservative estimate of the impact of tem-perature perturbations associated with gravity waves Ancil-lary profiles of temperature and pressure from the Met Officeanalyses which have been interpolated to the location andtime of the POAM III measurements have also been used inthis study The details of the Met Office analyses are de-scribed inLorenc et al(2000)

23 PSC temperature thresholds

In our analysis temperatures obtained from CHAMP obser-vations and Met Office analyses are used to determine thefrequency of temperatures below various temperature thresh-olds The NAT formation temperatureTNAT at 18 km asderived using the formulation ofHanson and Mauersberger(1988) which is a function of nitric acid water vapour con-centrations and pressure is shown in Fig2 the ranges se-lected are typical of concentrations in the polar stratosphereFigure2 clearly shows thatTNAT varies much more as a func-tion of water vapour than nitric acid mixing ratios for theranges typical of the Antarctic stratosphere chosen in this di-agram The selection of these values is based on nitric acidconcentrations identified inSantee et al(2007) and typicalwater vapour concentrations at this altitude measured by thePOAM III instrument (see Fig3) The probability of cross-ing the NAT formation temperature threshold is thus deter-mined by comparing Met Office and CHAMP temperatureobservations at a specific altitude with theTNAT thresholdderived using the formulation ofHanson and Mauersberger(1988) for that altitude The H2O variation used is taken fromthe POAM III observations and a constant nitric acid mix-ing ratio of 10 ppbv is assumed unless indicated otherwisein the text As previously indicated this study assumes the



Date

Alti

tude

(km

)

J A J O J A J O J A J O J A J O J12

14

16

18

20

22

24

05

1

15

2

25

3

35

4

45

Fig 3 Time-altitude contour plot of the H2O mixing ratio (ppmv)derived from POAM III observations for the period January 2002 toDecember 2005

STS formation temperature occurs at 35 K below the valueof TNAT (Carslaw et al 1994) and the ice frost point tem-perature is derived using the empirical formulae identified inMarti and Mauersberger(1993)

A time-altitude contour plot of the water vapour mixingratio for the period January 2002 to December 2005 used inthe temperature threshold calculations utilized in the rest ofthis paper is shown in Fig3 Examination of the seasonalvariation of water vapour mixing ratio displays a consistentirreversible dehydration potentially associated with the sed-imentation of ice PSC at altitudes between 12 and 20 kmbetween August and September each yearNedoluha et al(2003) indicates that the higher water vapour mixing ratiosobserved at elevated altitudes are likely to be associated withmoist air moving into the region via diabatic descent

Uncertainties on the probabilities displayed later in thisstudy are derived assuming a binomial distribution in a sim-ilar manner to that utilized inAlfred et al(2007) The stan-dard deviation of a binomial distributionσ can be writtenas

σ =radic

Np(1minusp) (1)

whereN is the number of observations andp is the proba-bility This number indicates the number of observations thatmight occur by chance and thus needs to be divided by thetotal number of observations to determine the uncertainty onthe probability

3 Results

Figure4 displays a time-altitude contour plot of the medianMet Office analysis temperature interpolated to the positions

Date

Alti

tude

(km

)

2000

Jun Jul Aug Sep Oct12

14

16

18

20

22

24

185

190

195

200

205

210

Fig 4 Time-altitude contour plot of the median temperature(Kelvin) derived from Met Office analyses data interpolated toPOAM III measurement locations in 2000 The full green line in-dicates the 25 PSC probability contour and the dotted green lineindicates the 75 PSC probability contour these contours being de-rived from POAM III extinction measurements Note that the tickmarks on the date axis indicate the first day of the month

of POAM III observations for 2000 Overlaid are the 25 and75 PSC probability contours determined from POAM IIIextinction measurements To ensure a large enough statisti-cal sample probabilities are calculated using a sliding 10-daywindow Examination of Fig4 suggests a qualitative corre-spondence between cool temperatures and PSC occurrence

Figure 5 displays the probability of temperatures belowTNAT based on Met Office analyses (red line) the STS PSCformation temperature (green line)TSTS and the ice nucle-ation temperature (blue line)TICE against time of year Fig-ure 5 also displays the probability of PSC detection basedon the k-means detection algorithm applied to POAM III ob-servations (black line) The latter also includes additionalregions identified using the ZMIN quality flag (reducing theunderestimation of Type II PSC) It should be noted that thevalues in Fig5 are the probabilities of PSC occurrence at18 km derived over a 10 day interval This methodologyshould provide accurate values of both PSC occurrence andtemperature threshold crossings but the daily values plottedwill not be independent Statistical uncertainties are calcu-lated as a three-standard-deviation envelope around each ofthe average series (displayed for reference purposes on thefifteenth day of each month) Note that problems with thePOAM III instrument and quality control procedures pro-duce data gaps in several years observations most notice-ably 2004 (see Fig5) Examination of Fig5 shows theexpected increase in PSC incidence from June to Septem-ber and a general reduction in occurrence after this date as-sociated with changes in temperature nitric acid and water



J J A S O N0

50

1001998

Occ

urre

nce

()

J J A S O N0

50

1001999

J J A S O N0

50

1002000

J J A S O N0

50

1002001

J J A S O N0

50

1002002

Occ

urre

nce

()

J J A S O N0

50

1002003

J J A S O N0

50

1002004

DateJ J A S O N

0

50

1002005

Date

Fig 5 The probability of temperatures below theTNAT threshold derived from Met Office analysis and using the H2O variation derivedfrom the POAM III observations and a constant nitric acid mixing ratio of 10 ppbv (red line) the probability of temperatures below theTSTSthreshold (green line)the probability of temperatures below theTICE threshold (blue line) and the probability of PSC derived from POAMIII extinction measurements (black line) against time of year for the 8 years of POAM III observations It should be noted that this is theprobability at 18km derived over a sliding time window of 10 days The vertical lines on the fifteenth of each month display the plus andminus three standard deviations for each probability Note that the tick marks on the date axis indicate the first day of the month

vapour concentrations However comparison of the PSC oc-currence predicted from temperature threshold crossings andthose derived from POAM III extinction data do not corre-spond within the defined uncertainties in several regions

Averaged over the entire period identified theTNAT tem-perature threshold derived from Met Office analyses predictsmore PSC than observed by POAM III This is to be ex-pected because Type Ia PSC formation requires average tem-peratures below theTNAT threshold for a period The exacttemperature value required varies dependent on whether NATonly forms by heterogeneous nucleation on pre-existing iceparticles requiring temperatures belowTICE (Carslaw et al1994) or NAT nucleation mechanisms independent of theexistence of ice particles (Hitchman et al 2003 Pagan et al2004 Svendsen et al 2005) The POAM III observationspredict between 60 and 85 of the occurrence identified bytheTNAT temperature threshold in all the years except 2004where only 45 of the PSC predicted by theTNAT thresholdis observed Comparison of the PSC occurrence observed byPOAM III and identified by theTSTS temperature threshold

displays a better correspondence but this threshold still over-estimates the quantity of PSC It should be noted that this re-sult concurs with recent CALIPSO observations detailed inPitts et al(2007) which suggest that theTSTS temperaturethreshold may be a better measure of PSC existence than theTNAT threshold Inspection of the frequency of PSC occur-rence identified by theTICE threshold derived from analysesand the POAM III frequency suggests theTICE occurrencegenerally significantly underestimates the quantity of PSCobserved by POAM III

Close examination of Fig5 indicates that during certainyears (specifically 1999 and 2002 to 2005 inclusive) the ob-served quantity of PSC in June is greater than that predictedfrom TNAT andTSTS temperature thresholds alone Analysis(not shown here) indicates that this pattern does not changesignificantly even if the nitric acid mixing ratio is increasedto 15 ppbv this change would make the temperature thresh-old warmer and thus easier to cross It is also noticeable thatat all other periods of the year the probability of PSC oc-currence from POAM III measurements is smaller than the



expected value based on the proportion of observations be-low theTNAT threshold The probability of PSC occurrencebased on theTICE temperature threshold does not exceed theprobability of occurrence based on POAM III observations inany of the eight observational years until after mid-AugustComparison of PSC occurrence based on POAM III obser-vations with the quantity of Met Office analysis data belowthe TSTS threshold shows a greater but not perfect corre-spondence which is generally best between the start of Julyand the start of September It should be noted that the tem-perature thresholds determined account for variations in thewater vapour mixing ratio based on POAM III observationsbut that the nitric acid is held constant at 10 ppbv Previousstudies suggest that the difference between the occurrenceprobability based on POAM III observations and those basedon temperature thresholds later in the season is due to den-itrification and dehydration (Nedoluha et al 2003 Alfredet al 2007) Thus the fixed nitric acid concentrations usedin Fig5 are likely to explain some of the over-estimation ob-served after June Note also that while not shown here TypeIa PSCs are identified almost exclusively by the unsupervisedclustering algorithm in June and early July in all years

As previously indicated NAT particles probably normallyform at temperatures several degrees colder thanTNAT(Tabazadeh et al 2001) or at even lower temperatures(Carslaw et al 1998) Therefore a high probability of PSCformation would not be expected in June even if the probabil-ity of temperatures crossing theTNAT temperature thresholdwas relatively large unless a mechanism for Type Ia cloudformation that requires temperatures only belowTNAT ex-ists This study examines whether the enhanced PSC occur-rence observed can be explained by gravity wave motions Itshould be noted that previous work bySaitoh et al(2006)has identified a similar enhanced PSC occurrence in June2003 compared to that identified by temperature thresholdsusing ILAS-II data An enhanced level of PSC occurrence inJune 2003 was also observed and modelled byHopfner et al(2006) Their study showed that only heterogenous nucle-ation on ice formed in mountain waves could reproduce theobserved PSC morphology This analysis has also recentlybeen confirmed by AIRS observations discussed byEcker-mann et al(2009)

Poole and Pitts(1994) have examined the relationship be-tween PSC frequency and temperature previously and shownthat PSC formation in the Southern Hemisphere is less likelylater in the PSC season than earlier Other studies such asMergenthaler et al(1997) have suggested that PSC forma-tion late in the PSC season in the Southern Hemisphere isstrongly affected by the preceding denitrification and dehy-dration which fits with the observations displayed in Fig3and 5 Saitoh et al(2006) also suggests that PSC frequencydepends on the degree of denitrification later in the seasonTheir study also indicates thatTNAT is not a good measure ofPSC occurrence especially above 20 km in late winter

A potential explanation for the differences between thePSC occurrence rates derived from temperature thresholdsusing Met Office analyses and those derived from POAM IIIextinction data could be seasonally varying biases in the MetOffice observations For example a warm bias in the MetOffice temperatures in June would produce lower PSC oc-currence rates than the true values which might explain thediscrepancies observed in some years in this period Figure6therefore compares Met Office analyses and CHAMP ROtemperatures at the POAM III sampling latitudes (see Fig1)It should be noted that studies byParrondo et al(2007) andLuntama et al(2008) have previously shown that CHAMPobservations provide good estimates of the background tem-peratures which can be used to identify issues in reanaly-ses datasets Figure6 displays time-altitude contour plots ofthe median temperature observed by Met Office analyses andCHAMP RO observations in 2005 The black contour line inboth panels of Fig6 displays the 195 K contour line fromthe Met Office analyses for reference purposes In this caseFigure6 is used to identify whether the increasedTNAT oc-currence probability observed in June may be associated withbiases between the temperatures identified in the Met Officeanalyses and the CHAMP RO observations We remove theeffect of gravity wave perturbations on the CHAMP RO ob-servations by using a 4th order polynomial filter Examina-tion shows a good correspondence between CHAMP RO andMet Office analyses in June though detailed examinationsuggests the CHAMP RO observations are slightly coolerIn July to August the temperatures differences are negligibleand in September and October the temperatures of the MetOffice analyses are slightly cooler than the CHAMP obser-vations Over the entire period the mean bias between thetwo sets of data is very small (simminus008 K) Similar analysisfor 2002 2003 and 2004 shows mean differences ofminus02402 and 047 K respectively The similarity of the bias ineach year compared to the change in the difference betweenthe estimated PSC occurrence based on temperature informa-tion and POAM III observations suggests that the differencecan not be solely explained by bias in the Met Office temper-atures

This study now examines the potential of gravity wavesto cause more temperature threshold crossings than expectedfrom the mean represented by Met Office analyses and in par-ticular attempts to identify whether the enhanced PSC fre-quency in June may be associated with temperature varia-tions that are not resolved in the Met Office analyses Toexamine this we use high vertical resolution measurementsmade by the CHAMP RO instrument available between 2002and 2005 It should be noted that the CHAMP observa-tions are not dense enough to directly identify the presence ofgravity waves at POAM III measurement locations and thuswe are limited to examining temperature threshold crossingfrequencies



Jul Aug Sep Oct

14

16

18

20

22

24

Alti

tude

(km

)

190

200

210

220

Jul Aug Sep Oct

14

16

18

20

22

24

Date

Alti

tude

(km

)

190

200

210

220

Fig 6 Time-altitude contour plot of the median temperature (Kelvin) derived from Met Office analysis (top panel) and CHAMP radiooccultation observations (bottom panel) The black contour line in both panels displays the 195K contour line from the Met Office analysisNote that the tick marks on the date axis indicate the first day of the month

We begin by displaying a single CHAMP temperature pro-file observed on 2 June 2006 (see Fig7) Figure7 also dis-plays the mean temperature derived by applying a 4th or-der polynomial fit to the CHAMP observations (red line) andtheTNAT (blue line) andTICE (blue dashed line) temperaturethresholds This temperature profile displays a clear wave-like structure The temperature perturbations associated withthe gravity wave (the difference between the red and greenlines) allow theTNAT threshold to be crossed at around 17 kmand between approximately 20 and 21 km while the meantemperature profile does not cross theTNAT threshold Itshould also be noted that the temperature perturbations typi-cally plusmn2 K do not cross theTICE threshold This is typical inCHAMP observations for this period of the year but periodswhere the temperature perturbations associated with gravitywaves allow a crossing of theTSTS or theTICE threshold dooccur It should be noted that the magnitude of the tempera-ture perturbations observed corresponds well to those identi-fied in Innis and Klekociuk(2006) However the impact ofobservational filtering on CHAMP measurements will meanthat these observations under-represent the true magnitudeof temperature perturbations associated with portions of thegravity wave spectrum Thus it is not possible to guaran-tee that the true minimum temperatures are not significantly

cooler than those observed The analysis inLange and Ja-cobi (2003) suggests that CHAMP observations are likely toprovide a lower limit on the impact of gravity waves on PSCformation It should also be noted that while not perfect therelatively high resolution of CHAMP data in the vertical andhorizontal resolution compared to model resolution does al-low fine-scale structure not observed in analyses to be exam-ined (Juarez et al 2009)

Figure 8 displays temperature distributions relative toTNAT taken from CHAMP RO observations at 18 km for dif-ferent months and latitudinal bands The aim of Fig8 is toshow the temporal and latitudinal structure of temperaturesconducive to PSC formation unhindered by the POAM sam-pling pattern and to be of use for comparisons with Fig5Given the relatively small number of CHAMP profiles ineach month observed poleward of 60S generally less than700 we have used coarse ten degree wide latitudinal bands toensure significant numbers of points in each histogram Theuse of these wide regions make it more difficult to comparedirectly with the Met Office analyses but given the limitedintra-monthly CHAMP data and the more detailed analysisdiscussed later we believe this is not a significant issue Thiscoarse latitudinal sampling also has the benefit of reducingany affects associated with a distorted vortex Examination



180 185 190 195 200 205 21012

14

16

18

20

22

24

Temperature (K)

Alti

tude

(km

)

Fig 7 This diagram shows a CHAMP temperature profile from2nd June 2006 (green line) and a 4th order polynomial fit to theCHAMP data (red line) The polynomial fit approximates the meantemperature observed by CHAMP the difference between the redand green lines represents temperature perturbations due to gravitywaves The blue full dashed and dotted lines indicate the values oftheTNAT TSTSandTICE temperature thresholds respectively

of Fig 8 shows a clear latitudinal variation with colder tem-peratures observed towards the pole A seasonal variationwith coldest temperatures observed in July and August isalso observable in each latitude band Examination of thetemperature relative to theTNAT temperature threshold showsthat in July and August at latitudes poleward of 70 S nearlyall the temperatures observed are belowTNAT Given thattemperature perturbations associated with gravity waves aremost likely to have a significant impact on PSC occurrencewhen the climatological mean is close to a PSC temperaturethreshold it is important to identify the dates and latitudeswhen these conditions occur These conditions preferentiallyoccur between 60 to 70 S in June through August in the70 to 80 S range in August to September and in the 80 to90 range in October Comparison of this variation with thePOAM latitudinal sampling pattern displayed in Fig1 showsa quite good agreement until the end of September Thuswhile the POAM observations are clearly not representativeof the conditions associated with the whole vortex as pre-viously identified byPitts et al(2007) for solar occultationsampling they actually follow the region where temperaturesmay be conducive for gravity wave induced PSC formationquite well at this altitude However it should be noted thatthis correspondence is not perfect and is only relevant to theTNAT threshold and the POAM sampling does not observethe Antarctic peninsula in August and September when thereis still a possibility of significant gravity wave impacts Thegeneral consistency of these patterns from year to year is alsoclear with the exception of 2002 on examination of Fig8

A simple comparison of the occurrence of temperaturesbelow TNAT shows that in June 2002 CHAMP RO obser-vations occur 29 of the time Met Office analyses have amean occurrence of approximately 20 over the month andthe POAM III extinction data suggests an occurrence rate ofroughly 30 A closer correspondence between the CHAMPRO and POAM observations is therefore apparent In 2003CHAMP RO observations occur 30 of the time Met Of-fice analyses have a mean occurrence of approximately 20over the month and the POAM extinction data have an oc-currence rate of roughly 20 In this case the CHAMPRO observations suggest a higher PSC occurrence than thePOAM observations In 2004 CHAMP RO observations sug-gest temperatures belowTNAT occur 23 of the time MetOffice temperatures are below the NAT threshold about 15of the time and POAM III data identifies PSC 19 of thetime Similar results occur in 2005 with a slightly closer cor-respondence between the CHAMP RO observations and thePOAM III values This simple analysis indicates that in somecases CHAMP RO has the potential to provide better esti-mates of temperature than those of the Met Office analysisIt is interesting that in other months and latitudinal bands theprobabilities determined from CHAMP RO data are largerthan those derived from POAM III observations Given thework detailed inJuarez et al(2009) the slightly better re-lationship in each June could be associated with the abilityof CHAMP RO measurements to observe fine-scale struc-ture not included in analyses The enhanced PSC occurrencein the 60 to 70 latitude range in June may be evidence thatthe gravity wave ldquohotspotrdquo of the Antarctic peninsula (Baum-gaertner and McDonald 2007 Hertzog et al 2008) could beparticularly important in this period Basically the combi-nation of mean temperatures close to theTNAT threshold andlarge mountain wave temperature perturbations could be im-portant The over-estimation of PSC by CHAMP RO obser-vations in some years in June notably 2003 could be associ-ated with a number of possible factors The over-estimationcould be associated with the fact that temperatures belowTNAT do not necessarily lead to NAT formation straight away(Peter 1997) The over-estimation could also be associatedwith differences in the spatial and temporal distribution ofCHAMP and POAM III observations Another possibility isthat the limited longitudinal sampling of POAM III obser-vations may underestimate PSC in some years It should benoted that the consistent over-estimation of PSC frequencyby CHAMP RO compared to that identified using POAM IIIis perhaps expected More consistent over-estimation of thetrue PSC occurrence based on temperature thresholds makesphysical sense given that these thresholds are a necessary butnot sufficient condition for PSC formation



minus15minus10 minus5 0 5 10 150

50

100Jun minus60o to minus70o


50



50



50

100

Occ

uren

ce (

)

Jul minus60o to minus70o


50

100Jul minus70o to minus80o


50



50

100Aug minus60o to minus70o


50



50



50

100

Occ

uren

ce (

)

Sep minus60o to minus70o


50

100Sep minus70o to minus80o


50



50

100

TminusTNAT

(K)

Oct minus60o to minus70o


50

100

TminusTNAT

(K)



50

100

TminusTNAT

(K)


Fig 8 The set of diagrams displayed show the temperature distribution relative toTNAT observed at 18km by CHAMP RO observationsfor different months and latitudinal regions (indicated on each diagram) The different color bars represent data from different years thesebeing 2002 (dark blue) 2003 (light blue) 2004 (yellow) and 2005 (red)

4 Discussion

As previously indicated temperature biases may explainsome of the differences observed between temperaturethreshold crossing frequencies and the frequency of PSC ob-served by POAM III extinction measurements However fur-ther inspection suggests that some features of the data maypoint to a role for internal gravity wave temperature pertur-bations Figure9 displays a time-longitude contour plot for2005 of the probability of PSC formation based on POAM IIIobservations Examination of Fig9 shows that an enhancedPSC probability which relates to the enhancements in Junedescribed previously in relation to Fig5 occurs around alongitude of 300 E This longitude corresponds to the po-sition of the Antarctic peninsula which has previously beenidentified as a ldquohotspotrdquo for gravity wave activity (Wu andJiang 2002 Baumgaertner and McDonald 2007 Hertzoget al 2008) and therefore might be a potential region of en-hanced PSC occurrence A distinct wavenumber one signa-ture is observable later in the season which is likely to becaused by planetary waves

Thus the expectation that a region of high gravity waveactivity might be associated with rsquoenhancedrsquo PSC occurrenceseems to be confirmed It should be noted that MIPAS mea-surements have identified enhanced PSC occurrence associ-ated with mountain waves in this area earlier in the samemonth in 2003 (Hopfner et al 2006) This suggests thatmountain waves may play a significant role in enhanced PSCformation and that this enhancement may occur frequently inthis region and period This is supported by the fact that thetime-longitude contour plot for 2003 derived from POAM IIIdata (not shown) displays a similar pattern to that observedin Fig 9 However it should be noted that the backgroundtemperatures shown in Fig8 display a large proportion ofobservations close to the NAT formation temperature (TNAT)in this period Thus the observed enhancement is likely tobe a result of both these factors

As previously indicated Figure5 shows that during Junethe number of PSC observed is greater than expected basedon the occurrence of regions where theTNAT andTSTS tem-perature thresholds are crossed in some years most clearly2002 and 2005 but also for a short period in 2003 A previousstudy bySaitoh et al(2006) using ILAS-II observations from



Date

Long

itude

Jul Aug Sep Oct Nov

50

100

150

200

250

300

10

20

30

40

50

60

70

80

90

100

Jul Aug Sep Oct Novminus90minus80minus70minus60

Latit

ude

Fig 9 A plot of the latitude of observations as a function of date(upper panel) and a date-longitude contour plot of the percentageprobability of PSC observation derived from POAM III measure-ments (lower panel) for the year 2005 Note that the tick marks onthe date axis indicate the first day of the month

2003 shows a corresponding enhanced PSC frequency com-pared to that expected based on temperature thresholds Thisis a strong validation of the results presented in this studysinceSaitoh et al(2006) uses the mean plus five standarddeviations from the ILAS-II aerosol extinction data to iden-tify PSC an extremely conservative threshold

To examine the importance of gravity wave temperatureperturbations on temperature threshold crossings we can useCHAMP RO observations with and without these tempera-ture perturbations included Figure10 displays the proba-bility of temperatures below theTNAT temperature thresholdagainst the probability of the threshold crossing being asso-ciated with a gravity wave perturbation based on CHAMPRO observations for 2002 to 2005 The second parame-ter was calculated by identifying all those temperature pro-files where theTNAT threshold was crossed and comparingwhether theTNAT threshold was crossed if only the meantemperature structure identified using a 4th order polyno-mial fit to the CHAMP profile and thereby removing smallvertical scale motions was observed By comparing thenumber of threshold crossings associated with the smoothedprofile and the unfiltered profile we can determine the pro-portion ofTNAT crossings which occur because of tempera-ture perturbations Figure10 suggests that temperature per-turbations associated with gravity waves account for approxi-mately 40 ofTNAT temperature threshold crossings in everyJune observed (2002 to 2005) and roughly 15 in all othermonths and years It should be noted that only CHAMP pro-files in the same latitude region that POAM III observationsexamine have been used in this analysis Figure10 suggeststhat gravity waves may have a significant impact early in the

0 10 20 30 40 50 60 70 80 90 1000

10

20

30

40

50

60

PNAT

()

GW

con

trib

utio

n (

)

June

July

August

September

October

Fig 10 Probability of observations of temperatures below the NATnucleation temperatureTNAT from CHAMP against the contribu-tion from small-scale wave motions (gravity waves) for differentmonths indicated in the legend It should be noted that the individ-ual points for a particular month relate to data from different years

PSC formation season when the mean temperature is close totheTNAT threshold This may explain the very clear identi-fication of mountain waves as the source of PSC detailed inHopfner et al(2006) The value for the months other thanJune compare extremely well with the estimate indicated inInnis and Klekociuk(2006) that gravity wave perturbationsinfluence PSC formation approximately 15 of the time

Examination of the quantity of data close to theTNATthreshold in Fig8 corroborates the estimates of the gravitywave influence derived in Fig10 In particular when POAMobserves between 65 and 70 S the quantity is close to 40in June and between 10ndash20 for July to October when ex-amined at the POAM sampling latitudes Another interestingpoint identified by a reviewer of this paper is that we canthen infer from Fig8 that the gravity wave impact on NATnucleation should be around 20 for August and Septemberabove the Peninsula

Similar analysis to that shown in Fig10 for theTSTS andTICE temperature thresholds shows comparable results thatis an increased frequency of crossing the thresholds associ-ated with gravity wave perturbations However the seasonalpatterns are shifted towards colder months as might be ex-pected In addition no extra threshold crossings associatedwith TSTS andTICE are observed in June Given that NATparticles are usually considered to form at temperatures sev-eral degrees colder thanTNAT (Tabazadeh et al 2001) or ateven lower temperatures (Carslaw et al 1998) the quantity



of PSC observed by POAM III is a little surprising and possi-bly supports studies such asVoigt et al(2005) which suggestthat there is a mechanism for NAT PSC formation that re-quires temperatures no lower than a few degrees belowTNAT But this result is dependent on the level of observational fil-tering in CHAMP measurements and could also be impactedby the constant value of nitric acid assumed in this work

As previously indicated observational filtering means thatCHAMP RO measurements observe only a portion of the ac-tual gravity wave spectrum (Lange and Jacobi 2003) Thefinite spatial resolution of the CHAMP observations meansthat any temperature perturbation observed can be thought ofas a spatial average of the temperature perturbations over thesatellitersquos weighting function This spatial averaging therebyreduces the variance observed The study byLange and Ja-cobi(2003) using a simple assumed gravity wave field whichhas a gravity wave spectrum with constant amplitude in therange 100ndash1000 km horizontal and 1-10 km vertical wave-length indicates that about 80 of the total variance is re-trieved in the worst case when the LOS scans perpendicularto the wave crests and about 88 when the LOS through thegravity waves is arbitrarily oriented Assuming a more re-alistic vertical wavelength spectrum with a spectral gradientof -53 the mean observed variance reduces by an additional2-3 However PSC formation will be impacted only byregions where the temperature is sufficiently cool for forma-tion to occur Thus since the POAM III instrument is also alimb sounder (and effectively produces a spatially averagedextinction) it will observe the integrated affect of a heteroge-neous temperature field on PSC occurrence Given high res-olution lidar observations of heterogeneous structures associ-ated with PSC fields display clear phase structures related togravity waves (Shibata et al 2003) We would suggest thatthe spatial average temperature field may be a useful measureof the PSC formation potential averaged over an area How-ever it should be noted that this measure is by no means per-fect because the nonlinear hysteresis affect inherent in PSCmicrophysics would impact the spatial average

5 Conclusions

Examination of Fig5 shows that PSC occurrence can be rel-atively well predicted by the presence of temperatures be-low either theTNAT or TSTS thresholds This is particularlytrue if varying water vapour and nitric acid concentrationsare taken into account However in some years more PSCthan might be expected from examination of simple temper-ature thresholds applied to Met Office analyses are observedin June The lack of simultaneous nitric acid measurementsmeans that during periods of denitrification the temperaturethresholds significantly overestimate PSC occurrence this isparticularly clear in October in several years

Inspection of the distribution of temperatures relative totheTNAT threshold shown in Fig8 shows that the latitudinal

variation of POAM III measurements means that these valuesare not representative of the whole vortex as previously iden-tified byPitts et al(2007) However the POAM III observa-tions occur in regions where temperatures are conducive forgravity wave induced PSC formation When focusing on theJune results examination of the probability of temperaturesbelow theTNAT threshold based on CHAMP observationsshow a slightly improved predictive capability compared tothe PSC occurrence derived from Met Office data CHAMPRO observations are consistently close or over-estimate thePSC occurrence measured by POAM III This makes somephysical sense given that these thresholds are a necessary butnot sufficient condition for PSC formation This improvedcorrespondence may partially be associated with a warm biasin the Met Office temperature data compared to CHAMP ROmeasurements

Analysis of the CHAMP temperature observations sug-gests that temperature perturbations associated with grav-ity waves are likely to be important in explaining the en-hanced PSC incidence observed in June Examination of thelongitudinal structure in June 2005 also shows that the en-hanced PSC occurrence is close to the Antarctic peninsulaa known mountain wave ldquohotspotrdquo Analysis also suggeststhat the temperature perturbations produced by gravity wavesare most important early in the Antarctic winter but havea smaller effect later in the winter However it should benoted that the rsquoenhancedrsquo PSC occurrence in June is likelyto be associated with both strong temperature perturbationsassociated with mountain waves and temperatures hoveringslightly above theTNAT temperature threshold The lowervalues later in the winter are likely to be affected by thecooler background temperatures making extra temperaturethreshold crossings more unlikely and the reduced magni-tude of gravity waves perturbations observed as POAM IIIbegins to sample poleward of the peninsula

Based on Fig10 temperature perturbations associatedwith gravity waves contribute to the formation of about 15of theTNAT threshold crossings observed a value which cor-responds well to the value derived inInnis and Klekociuk(2006) It is important to note that the values derived fromCHAMP data may be underestimates because of observa-tional filtering and thus should be considered conservativeestimates Studies which focus on ozone depletion in theearly period of the Antarctic winter may need to be awareor parameterize for the effect of gravity waves on PSC for-mation In the future we aim to complete a more conclusivestudy by using measurements from CALIPSO and the COS-MIC satellites

AcknowledgementsWe are grateful to the GFZ Potsdam and theISDC for providing the CHAMP data The POAM III data iskindly provided athttpwvmsnrlnavymilPOAMdatadatahtmlAJM would like to acknowledge the support provided by theBrian Mason Scientific and Technical Trust AJM would alsolike to thank Rebecca Batchelor and Simon Alexander for usefulcomments on an early version of this work We would like to thank



four anonymous referees for providing suggestions for furtheranalysis and improvements to the text

Edited by P Haynes

References

Alfred J Fromm M Bevilacqua R Nedoluha G Strawa APoole L and Wickert J Observations and analysis of polarstratospheric clouds detected by POAM III and SAGE III duringthe SOLVE IIVINTERSOL campaign in the 20022003 North-ern Hemisphere winter Atmos Chem Phys 7 2151ndash21632007httpwwwatmos-chem-physnet721512007

Baumgaertner A J G and McDonald A J A gravity waveclimatology for Antarctica compiled from Challenging Min-isatellite PayloadGlobal Positioning System (CHAMPGPS)radio occultations J Geophys Res-Atmos 112 D05103doi1010292006JD007504 2007

Carslaw K S Luo B P Clegg S L Peter T BrimblecombeP and Crutzen P J Stratospheric Aerosol Growth and Hno3Gas-Phase Depletion from Coupled Hno3 and Water-Uptake byLiquid Particles Geophys Res Lett 21 2479ndash2482 1994

Carslaw K S Wirth M Tsias A Luo B P Dornbrack ALeutbecher M Volkert H Renger W Bacmeister J TReimers E and Peter T H Increased stratospheric ozonedepletion due to mountain-induced atmospheric waves Nature391 675ndash678 1998

de la Torre A and Alexander P Gravity waves above An-des detected from GPS radio occultation temperature pro-files Mountain forcing Geophys Res Lett 32 L17815doi1010292005GL022959 2005

Eckermann S D Hoffmann L Hopfner M Wu D L andAlexander M J Antarctic NAT PSC belt of June 2003 Ob-servational validation of the mountain wave seeding hypothesisGeophys Res Lett 36 doi1010292008GL036629 2009

Felton M A Kovacs T A Omar A H and A H CClassification of Polar Stratospheric Clouds Using LI-DAR Measurements From the SAGE III Ozone Lossand Validation Experiment Tech Rep ARL-TR-4154httpwwwarlarmymilwwwdefaultcfmAction=17ampPage=239ampTopic=TechnicalReportsampYear=2007 2007

Fromm M Alfred J and Pitts M A unified long-term high-latitude stratospheric aerosol and cloud database using SAM IISAGE II and POAM IIIII data Algorithm description databasedefinition and climatology J Geophys Res-Atmos 108(D12)14366 doi1010292002JD002772 2003

Fromm M D Lumpe J D Bevilacqua R M Shettle E PHornstein J Massie S T and Fricke K H Observations ofAntarctic polar stratospheric clouds by POAM II 1994-1996 JGeophys Res-Atmos 102 23659ndash23672 1997

Fueglistaler S and B P L B P C Voigt C K S Carslaw K Sand Peter T NAT-rock formation by mother clouds a micro-physical model study Atmos Chem Phys 2 93ndash98 2002httpwwwatmos-chem-physnet2932002

Gobiet A Foelsche U Steiner A K Borsche M KirchengastG and Wickert J Climatological validation of stratospherictemperatures in ECMWF operational analyses with CHAMPradio occultation data Geophys Res Lett 32 L12806doi1010292005GL022617 2005

Hanson D and Mauersberger K Laboratory Studies of the Nitric-Acid Trihydrate ndash Implications for the South Polar StratosphereGeophys Res Lett 15 855ndash858 1988

Hertzog A Boccara G Vincent R A Vial F and CocquerezP Estimation of gravity wave momentum flux and phase speedsfrom quasi-Lagrangian stratospheric balloon flights Part II Re-sults from the Vorcore campaign in Antarctica J Atmos Sci65 3056ndash3070 2008

Hitchman M H Buker M L Tripoli G J Browell E VGrant W B McGee T J and Burris J F Nonoro-graphic generation of Arctic polar stratospheric clouds dur-ing December 1999 J Geophys Res-Atmos 108(D5) 8325doi10292001JD001034 2003

Hopfner M Blumenstock T Hase F Zimmermann A FlentjeH and Fueglistaler S Mountain polar stratospheric cloud mea-surements by ground based FTIR solar absorption spectroscopyGeophys Res Lett 28 2189ndash2192 2001

Hopfner M Larsen N Spang R Luo B P Ma J SvendsenS H Eckermann S D Knudsen B Massoli P Cairo FStiller G Von Clarmann T and Fischer H MIPAS detectsAntarctic stratospheric belt of NAT PSCs caused by mountainwaves Atmos Chem Phys 6 1221ndash1230 2006httpwwwatmos-chem-physnet612212006

Huck P E McDonald A J Bodeker G E and Struthers HInterannual variability in Antarctic ozone depletion controlled byplanetary waves and polar temperature Geophys Res Lett 32L13819 doi1010292005GL022943 2005

Innis J L and Klekociuk A R Planetary wave and grav-ity wave influence on the occurrence of polar stratosphericclouds over Davis Station Antarctica seen in lidar and ra-diosonde observations J Geophys Res-Atmos 111 D22102doi1010292006JD007629 2006

Irie H Pagan K L Tabazadeh A Legg M J and Sugita TInvestigation of polar stratospheric cloud solid particle formationmechanisms using ILAS and AVHRR observations in the ArcticGeophys Res Lett 31 L15107 doi1010292004GL0202462004

Jakob C and Tselioudis G Objective identification of cloudregimes in the Tropical Western Pacific Geophys Res Lett30(21) 2082 doi1010292003GL018367 2003

Juarez M D Marcus S Dornbrack A Schroder T M KiviR Iijima B A Hajj G A and Mannucci A J Detectionof temperatures conducive to Arctic polar stratospheric cloudsusing CHAMP and SAC-C radio occultation data J GeophysRes-Atmos 114 D07112 doi1010292008JD011261 2009

Lange M and Jacobi C Analysis of gravity waves from radiooccultation measurements in First CHAMP Mission Resultsfor Gravity Magnetic and Atmospheric Studies edited by Reig-ber C Luhr H and Schwintzer P Springer Berlin Germany479ndash484 2003

Lorenc A C Ballard S P Bell R S Ingleby N B Andrews PL F Barker D M Bray J R Clayton A M Dalby T Li DPayne T J and Saunders F W The Met Office global three-dimensional variational data assimilation scheme Q J Roy Me-teor Soc 126 2991ndash3012 2000

Lowe D and MacKenzie A R Polar stratospheric cloud micro-physics and chemistry J Atmos Solar-Terr Phys 70 13ndash402008

Lucke R L Korwan D R Bevilacqua R M Hornstein J S



Shettle E P Chen D T Daehler M Lumpe J D FrommM D Debrestian D Neff B Squire M Konig-Langlo Gand Davies J The Polar Ozone and Aerosol Measurement(POAM) III instrument and early validation results J GeophysRes-Atmos 104 18785ndash18799 1999

Luntama J P Kirchengast G Borsche M Foelsche U SteinerA Healy S von Engeln A OrsquoClerigh E and Marquardt CProspects of the Eps Gras Mission for Operational AtmosphericApplications Bulletin of the American Meteor Soc 89 1863ndash1875 2008

Marti J and Mauersberger K A Survey and New Measurementsof Ice Vapor-Pressure at Temperatures between 170 and 250 kGeophys Res Lett 20 363ndash366 1993

McDonald A J and Hertzog A Comparison of stratosphericmeasurements made by CHAMP radio occultation and Stra-teoleVorcore in situ data Geophys Res Lett 35 L11805doi1010292008GL033338 2008

Mergenthaler J L Kumer J B Roche A E and Massie S TDistribution of Antarctic polar stratospheric clouds as seen bythe CLAES experiment J Geophys Res-Atmos 102 19161ndash19170 1997

Nedoluha G E Bevilacqua R M Fromm M D Hoppel K Wand Allen D R POAM measurements of PSCs and water vaporin the 2002 Antarctic vortex Geophys Res Lett 30(15) 1796doi1010292003GL017577 2003

Noel V Hertzog A Chepfer H and Winker D M Po-lar stratospheric clouds over Antarctica from the CALIPSOspaceborne lidar J Geophys Res-Atmos 113 D02205doi1010292007JD008616 2008

Pagan K L Tabazadeh A Drdla K Hervig M E Ecker-mann S D Browell E V Legg M J and Foschi P GObservational evidence against mountain-wave generation of icenuclei as a prerequisite for the formation of three solid ni-tric acid polar stratospheric clouds observed in the Arctic inearly December 1999 J Geophys Res-Atmos 109 D04312doi1010292003JD003846 2004

Parrondo M C Yela M Gil M von der Gathen P and OchoaH Mid-winter lower stratosphere temperatures in the Antarc-tic vortex comparison between observations and ECMWF andNCEP operational models Atmos Chem Phys 7 435ndash4412007httpwwwatmos-chem-physnet74352007

Peter T Microphysics and heterogeneous chemistry of polarstratospheric clouds Ann Rev Phys Chem 48 785ndash822 1997

Pitts M C Thomason L W Poole L R and Winker D MCharacterization of Polar Stratospheric Clouds with spacebornelidar CALIPSO and the 2006 Antarctic season Atmos ChemPhys 7 5207ndash5228 2007httpwwwatmos-chem-physnet752072007

Poole L R and Pitts M C Polar Stratospheric Cloud Climatol-ogy Based on Stratospheric Aerosol Measurement-Ii Observa-tions from 1978 to 1989 J Geophys Res-Atmos 99 13083ndash13089 1994

Saitoh N Hayashida S Sugita T Nakajima H Yokota T andSasano Y Variation in PSC Occurrence Observed with ILAS-IIover the Antarctic in 2003 SOLA 2 72ndash75 2006

Santee M L Lambert A Read W G Livesey N J CofieldR E Cuddy D T Daffer W H Drouin B J FroidevauxL Fuller R A Jarnot R F Knosp B W Manney G LPerun V S Snyder W V Stek P C Thurstans R P Wagner

P A Waters J W Muscari G de Zafra R L Dibb J EFahey D W Popp P J Marcy T P Jucks K W Toon G CStachnik R A Bernath P F Boone C D Walker K AUrban J and Murtagh D Validation of the Aura MicrowaveLimb Sounder HNO3 measurements J Geophys Res-Atmos112 D24540 doi1010292007JD008721 2007

Shibata T Sato K Kobayashi H Yabuki M and ShiobaraM Antarctic polar stratospheric clouds under temperature per-turbation by nonorographic inertia gravity waves observed bymicropulse lidar at Syowa Station J Geophys Res-Atmos108(D3) 4105 doi1010292002JD002 713 2003

Solomon S Garcia R R Rowland F S and Wuebbles D J Onthe Depletion of Antarctic Ozone Nature 321 755ndash758 1986

Strawa A W Drdla K Fromm M Pueschel R F Hop-pel K W Browell E V Hamill P and Dempsey D PDiscriminating types Ia and Ib polar stratospheric clouds inPOAM satellite data J Geophys Res-Atmos 107(D20) 8291doi1010292001JD000 458 2002

Svendsen S H Larsen N Knudsen B Eckermann S D andBrowell E V Influence of mountain waves and NAT nucle-ation mechanisms on polar stratospheric cloud formation at localand synoptic scales during the 1999-2000 Arctic winter AtmosChem Phys 5 739ndash753 2005httpwwwatmos-chem-physnet57392005

Tabazadeh A Jensen E J Toon O B Drdla K and SchoeberlM R Role of the stratospheric polar freezing belt in denitrifica-tion Science 291 2591ndash2594 2001

Teitelbaum H Moustaoui M and Fromm M Exploring po-lar stratospheric cloud and ozone minihole formation The pri-mary importance of synoptic-scale flow perturbations J Geo-phys Res-Atmos 106 28173ndash28188 2001

Tsias A Wirth M Carslaw K S Biele J Mehrtens H Re-ichardt J Wedekind C Weiss V Renger W Neuber Rvon Zahn U Stein B Santacesaria V Stefanutti L FierliF Bacmeister J and Peter T Aircraft lidar observationsof an enhanced type Ia polar stratospheric clouds during APE-POLECAT J Geophys Res-Atmos 104 23961ndash23969 1999

Voigt C Schlager H Luo B P Dornbrack A D Roiger AStock P Curtius J Vossing H Borrmann S Davies SKonopka P Schiller C Shur G and Peter T Nitric AcidTrihydrate (NAT) formation at low NAT supersaturation in Po-lar Stratospheric Clouds (PSCs) Atmos Chem Phys 5 1371ndash1380 2005httpwwwatmos-chem-physnet513712005

Wang K Y and Lin S C First continuous GPS soundings oftemperature structure over Antarctic winter from FORMOSAT-3COSMIC constellation Geophys Res Lett 34 L12805doi1010292007GL030159 2007

Wang Z Stephens G Deshler T Trepte C Parish TVane D Winker D Liu D and Adhikari L Associa-tion of Antarctic polar stratospheric cloud formation on tro-pospheric cloud systems Geophys Res Lett 35 L13806doi1010292008GL034 209 2008

Wickert J Schmidt T Beyerle G Konig R and Reigber CThe radio occultation experiment aboard CHAMP Operationaldata analysis and validation of vertical atmospheric profiles JMeteor Soc Jpn 82 381ndash395 2004

World Meteorological Organization Scientific Assessment ofOzone Depletion 2006 Global Ozone Research and MonitoringProject ndash Report No 50 572 pp Geneva Switzerland 2007



Wu D L and Jiang J H MLS observations of atmospheric grav-ity waves over Antarctica J Geophys Res-Atmos 107(D24)4773 doi1010292002JD002390 2002



(NAT) Type Ib composed of supercooled ternary solution(STS a liquid mixture of H2O HNO3 and H2SO4) andType II composed primarily of ice (seeLowe and MacKen-zie 2008 and references therein) These three categoriesof PSC form at different temperatures in the stratosphereIn particular the NAT PSC formation temperature (TNAT)is dependent on pressure nitric acid and water vapour con-centrations (Hanson and Mauersberger 1988) and is signifi-cantly warmer than the ice frost point (TICE) The STS for-mation temperature (TSTS) occurs approximately 35 K be-low the value ofTNAT (Carslaw et al 1994) and the ice frostpoint temperature can be derived by using the formulationdescribed inMarti and Mauersberger(1993) Temperaturesare cool enough in the Antarctic stratosphere for all threetypes of clouds to be present during the winter The presentstudy examines the potential contribution of internal grav-ity waves to PSC formation in the Antarctic stratosphere Inthe Arctic the temperature perturbations associated with in-ternal gravity waves particularly mountain waves have beenshown to be responsible for the formation of a significantproportion of PSCs (Carslaw et al 1998 Hopfner et al2001) However the observed relationship between Antarc-tic PSC formation and low background temperatures has ledto less focus on wave induced PSC formation processes inthe Antarctic

Recent studies have indicated that the temperature per-turbations related to these waves may have a more signifi-cant effect on Antarctic PSC occurrence and have a poten-tially important role in ozone depletion (Shibata et al 2003Hopfner et al 2006 Innis and Klekociuk 2006 Eckermannet al 2009) Such effects may be particularly significantin early winter where temperatures hover close to PSC for-mation temperature thresholds Analysis byHopfner et al(2006) has identified a significant role of mountain wavesin the formation of PSC and subsequent denitrification inthe Antarctic Hopfner et al(2006) described spaceborneinfrared measurements of a vortex-wide area of NAT PSCaround Antarctica using the Michelson Interferometer forPassive Atmospheric Sounding (MIPAS) instrument Sim-ulations for the period 10ndash12 June 2003 indicate the obser-vations are likely to be associated with heterogeneous nucle-ation on ice in the cooling phases of large-amplitude strato-spheric mountain waves over the Antarctic Peninsula Theprocess operating here is the same mountain wave NAT for-mation model originally proposed byCarslaw et al(1998)for the Arctic winter stratosphere Recent work byEcker-mann et al(2009) examined the same MIPAS satellite obser-vations over Antarctica with additional Atmospheric InfraredSounder (AIRS) data Their study showed that large moun-tain wave temperature perturbations produced NAT which isadvected to form a circumpolar NAT outbreak which concurswith Hopfner et al(2006)

Other studies have also examined the role of gravity waveson PSC formation For example a ground-based lidar studyby Shibata et al(2003) indicated that Type II PSCs occurred

when stratospheric temperatures dropped belowTICE be-cause of perturbations associated with inertia gravity wavesAnother lidar study byInnis and Klekociuk(2006) also es-timated the influence of gravity waves on PSC occurrenceThey concluded that during the ldquoPSC seasonrdquo when the back-ground temperature was close enough to the NAT formationtemperature (TNAT) derived inHanson and Mauersberger(1988) gravity wave perturbations influence PSC formationapproximately 15 of the time Interestingly this value issimilar to the findings ofFelton et al(2007) who concludedType Ia enhanced PSCs which are likely to be NAT crys-tals with unusually large lidar backscattering ratios which ap-pear down wind of mountain wave-induced ice clouds (Tsiaset al 1999) make up 11 of the clouds observed during theSAGE III Ozone Loss and Validation Experiment (SOLVE)campaign in winter 2000 in the Northern Hemisphere How-ever the warmer mean temperature in the Arctic means thatthis is likely to be coincidental It is also worthy of mentionthat Innis and Klekociuk(2006) found a clear relationshipbetween planetary wave temperature perturbations and PSCoccurrence on the edge of the Antarctic continent Recentanalysis using data from CALIPSO (a space-borne lidar) hasalso mentioned the potential importance of gravity waves onPSC formation (Pitts et al 2007 Noel et al 2008) How-ever enhanced PSC formation has also been suggested to berelated to synoptic-scale motions (Teitelbaum et al 2001Wang et al 2008)

Previous work has shown that NAT PSC formation may beimplicitly linked to gravity wave temperature perturbationsIn particular homogeneous NAT nucleation does not occurdirectly on temperatures crossingTNAT Rather average tem-peratures need to be below theTNAT threshold for a periodgreater than a few days for nucleation to occur (Peter 1997)Carslaw et al(1994) discusses NAT formation and indicatesthat NAT may form by heterogeneous nucleation on pre-existing ice particles which may form because of large tem-perature perturbations produced by mountain waves Workby Fueglistaler et al(2002) suggests that early season PSCproduction associated with Type Ia or Ia enhanced cloudsproduced using the mechanism proposed inCarslaw et al(1994) can potentially be extremely important in denitrifi-cation due to their role in the formation of extremely largeNAT particles if number densities within the clouds are highHowever recent work has additionally identified a need forNAT nucleation mechanisms which are independent of theexistence of ice particles only requiring temperatures belowthe existence temperature for NAT which is significantly eas-ier (approximately 7ndash8 K warmer thanTICE) to meet than thedemand for pre-existing ice particles

Voigt et al(2005) detailed measurements where the con-ditions of particle formation are well enough constrained toconclude that a PSC is observed in air parcels that spent lessthan a day (approximately 18 h) at temperatures no lowerthan 3K belowTNAT Hitchman et al(2003) also suggeststhat a formation mechanism for Type Ia PSC must exist







minus85

minus80

minus75

minus70

minus65

minus60

Date

Latit

ude























H2O

con

cent

ratio

n (p

pmv)


5 6 7 8 9 10 11 12 13 14 152

25

3

35

4

45

5

55

1935

194

1945

195

1955

196

1965

197

1975

198

1985

199







Date

Alti

tude

(km

)


14

16

18

20

22

24

05

1

15

2

25

3

35

4

45





σ =radic

Np(1minusp) (1)


3 Results


Date

Alti

tude

(km

)

2000


14

16

18

20

22

24

185

190

195

200

205

210






J J A S O N0

50

1001998

Occ

urre

nce

()

J J A S O N0

50

1001999

J J A S O N0

50

1002000

J J A S O N0

50

1002001

J J A S O N0

50

1002002

Occ

urre

nce

()

J J A S O N0

50

1002003

J J A S O N0

50

1002004

DateJ J A S O N

0

50

1002005

Date















Jul Aug Sep Oct

14

16

18

20

22

24

Alti

tude

(km

)

190

200

210

220

Jul Aug Sep Oct

14

16

18

20

22

24

Date

Alti

tude

(km

)

190

200

210

220







180 185 190 195 200 205 21012

14

16

18

20

22

24

Temperature (K)

Alti

tude

(km

)







50



50



50



50

100

Occ

uren

ce (

)



50



50



50



50



50



50

100

Occ

uren

ce (

)



50



50



50

100

TminusTNAT

(K)



50

100

TminusTNAT

(K)



50

100

TminusTNAT

(K)



4 Discussion






Date

Long

itude

Jul Aug Sep Oct Nov

50

100

150

200

250

300

10

20

30

40

50

60

70

80

90

100


Latit

ude




0 10 20 30 40 50 60 70 80 90 1000

10

20

30

40

50

60

PNAT

()

GW

con

trib

utio

n (

)

June

July

August

September

October









5 Conclusions










Edited by P Haynes

References
































































minus85

minus80

minus75

minus70

minus65

minus60

Date

Latit

ude























H2O

con

cent

ratio

n (p

pmv)


5 6 7 8 9 10 11 12 13 14 152

25

3

35

4

45

5

55

1935

194

1945

195

1955

196

1965

197

1975

198

1985

199







Date

Alti

tude

(km

)


14

16

18

20

22

24

05

1

15

2

25

3

35

4

45





σ =radic

Np(1minusp) (1)


3 Results


Date

Alti

tude

(km

)

2000


14

16

18

20

22

24

185

190

195

200

205

210






J J A S O N0

50

1001998

Occ

urre

nce

()

J J A S O N0

50

1001999

J J A S O N0

50

1002000

J J A S O N0

50

1002001

J J A S O N0

50

1002002

Occ

urre

nce

()

J J A S O N0

50

1002003

J J A S O N0

50

1002004

DateJ J A S O N

0

50

1002005

Date















Jul Aug Sep Oct

14

16

18

20

22

24

Alti

tude

(km

)

190

200

210

220

Jul Aug Sep Oct

14

16

18

20

22

24

Date

Alti

tude

(km

)

190

200

210

220







180 185 190 195 200 205 21012

14

16

18

20

22

24

Temperature (K)

Alti

tude

(km

)







50



50



50



50

100

Occ

uren

ce (

)



50



50



50



50



50



50

100

Occ

uren

ce (

)



50



50



50

100

TminusTNAT

(K)



50

100

TminusTNAT

(K)



50

100

TminusTNAT

(K)



4 Discussion






Date

Long

itude

Jul Aug Sep Oct Nov

50

100

150

200

250

300

10

20

30

40

50

60

70

80

90

100


Latit

ude




0 10 20 30 40 50 60 70 80 90 1000

10

20

30

40

50

60

PNAT

()

GW

con

trib

utio

n (

)

June

July

August

September

October









5 Conclusions










Edited by P Haynes

References












































































H2O

con

cent

ratio

n (p

pmv)


5 6 7 8 9 10 11 12 13 14 152

25

3

35

4

45

5

55

1935

194

1945

195

1955

196

1965

197

1975

198

1985

199







Date

Alti

tude

(km

)


14

16

18

20

22

24

05

1

15

2

25

3

35

4

45





σ =radic

Np(1minusp) (1)


3 Results


Date

Alti

tude

(km

)

2000


14

16

18

20

22

24

185

190

195

200

205

210






J J A S O N0

50

1001998

Occ

urre

nce

()

J J A S O N0

50

1001999

J J A S O N0

50

1002000

J J A S O N0

50

1002001

J J A S O N0

50

1002002

Occ

urre

nce

()

J J A S O N0

50

1002003

J J A S O N0

50

1002004

DateJ J A S O N

0

50

1002005

Date















Jul Aug Sep Oct

14

16

18

20

22

24

Alti

tude

(km

)

190

200

210

220

Jul Aug Sep Oct

14

16

18

20

22

24

Date

Alti

tude

(km

)

190

200

210

220







180 185 190 195 200 205 21012

14

16

18

20

22

24

Temperature (K)

Alti

tude

(km

)







50



50



50



50

100

Occ

uren

ce (

)



50



50



50



50



50



50

100

Occ

uren

ce (

)



50



50



50

100

TminusTNAT

(K)



50

100

TminusTNAT

(K)



50

100

TminusTNAT

(K)



4 Discussion






Date

Long

itude

Jul Aug Sep Oct Nov

50

100

150

200

250

300

10

20

30

40

50

60

70

80

90

100


Latit

ude




0 10 20 30 40 50 60 70 80 90 1000

10

20

30

40

50

60

PNAT

()

GW

con

trib

utio

n (

)

June

July

August

September

October









5 Conclusions










Edited by P Haynes

References




























































Date

Alti

tude

(km

)


14

16

18

20

22

24

05

1

15

2

25

3

35

4

45





σ =radic

Np(1minusp) (1)


3 Results


Date

Alti

tude

(km

)

2000


14

16

18

20

22

24

185

190

195

200

205

210






J J A S O N0

50

1001998

Occ

urre

nce

()

J J A S O N0

50

1001999

J J A S O N0

50

1002000

J J A S O N0

50

1002001

J J A S O N0

50

1002002

Occ

urre

nce

()

J J A S O N0

50

1002003

J J A S O N0

50

1002004

DateJ J A S O N

0

50

1002005

Date















Jul Aug Sep Oct

14

16

18

20

22

24

Alti

tude

(km

)

190

200

210

220

Jul Aug Sep Oct

14

16

18

20

22

24

Date

Alti

tude

(km

)

190

200

210

220







180 185 190 195 200 205 21012

14

16

18

20

22

24

Temperature (K)

Alti

tude

(km

)







50



50



50



50

100

Occ

uren

ce (

)



50



50



50



50



50



50

100

Occ

uren

ce (

)



50



50



50

100

TminusTNAT

(K)



50

100

TminusTNAT

(K)



50

100

TminusTNAT

(K)



4 Discussion






Date

Long

itude

Jul Aug Sep Oct Nov

50

100

150

200

250

300

10

20

30

40

50

60

70

80

90

100


Latit

ude




0 10 20 30 40 50 60 70 80 90 1000

10

20

30

40

50

60

PNAT

()

GW

con

trib

utio

n (

)

June

July

August

September

October









5 Conclusions










Edited by P Haynes

References




























































J J A S O N0

50

1001998

Occ

urre

nce

()

J J A S O N0

50

1001999

J J A S O N0

50

1002000

J J A S O N0

50

1002001

J J A S O N0

50

1002002

Occ

urre

nce

()

J J A S O N0

50

1002003

J J A S O N0

50

1002004

DateJ J A S O N

0

50

1002005

Date















Jul Aug Sep Oct

14

16

18

20

22

24

Alti

tude

(km

)

190

200

210

220

Jul Aug Sep Oct

14

16

18

20

22

24

Date

Alti

tude

(km

)

190

200

210

220







180 185 190 195 200 205 21012

14

16

18

20

22

24

Temperature (K)

Alti

tude

(km

)







50



50



50



50

100

Occ

uren

ce (

)



50



50



50



50



50



50

100

Occ

uren

ce (

)



50



50



50

100

TminusTNAT

(K)



50

100

TminusTNAT

(K)



50

100

TminusTNAT

(K)



4 Discussion






Date

Long

itude

Jul Aug Sep Oct Nov

50

100

150

200

250

300

10

20

30

40

50

60

70

80

90

100


Latit

ude




0 10 20 30 40 50 60 70 80 90 1000

10

20

30

40

50

60

PNAT

()

GW

con

trib

utio

n (

)

June

July

August

September

October









5 Conclusions










Edited by P Haynes

References



































































Jul Aug Sep Oct

14

16

18

20

22

24

Alti

tude

(km

)

190

200

210

220

Jul Aug Sep Oct

14

16

18

20

22

24

Date

Alti

tude

(km

)

190

200

210

220







180 185 190 195 200 205 21012

14

16

18

20

22

24

Temperature (K)

Alti

tude

(km

)







50



50



50



50

100

Occ

uren

ce (

)



50



50



50



50



50



50

100

Occ

uren

ce (

)



50



50



50

100

TminusTNAT

(K)



50

100

TminusTNAT

(K)



50

100

TminusTNAT

(K)



4 Discussion






Date

Long

itude

Jul Aug Sep Oct Nov

50

100

150

200

250

300

10

20

30

40

50

60

70

80

90

100


Latit

ude




0 10 20 30 40 50 60 70 80 90 1000

10

20

30

40

50

60

PNAT

()

GW

con

trib

utio

n (

)

June

July

August

September

October









5 Conclusions










Edited by P Haynes

References




























































180 185 190 195 200 205 21012

14

16

18

20

22

24

Temperature (K)

Alti

tude

(km

)







50



50



50



50

100

Occ

uren

ce (

)



50



50



50



50



50



50

100

Occ

uren

ce (

)



50



50



50

100

TminusTNAT

(K)



50

100

TminusTNAT

(K)



50

100

TminusTNAT

(K)



4 Discussion






Date

Long

itude

Jul Aug Sep Oct Nov

50

100

150

200

250

300

10

20

30

40

50

60

70

80

90

100


Latit

ude




0 10 20 30 40 50 60 70 80 90 1000

10

20

30

40

50

60

PNAT

()

GW

con

trib

utio

n (

)

June

July

August

September

October









5 Conclusions










Edited by P Haynes

References





























































50



50



50



50

100

Occ

uren

ce (

)



50



50



50



50



50



50

100

Occ

uren

ce (

)



50



50



50

100

TminusTNAT

(K)



50

100

TminusTNAT

(K)



50

100

TminusTNAT

(K)



4 Discussion






Date

Long

itude

Jul Aug Sep Oct Nov

50

100

150

200

250

300

10

20

30

40

50

60

70

80

90

100


Latit

ude




0 10 20 30 40 50 60 70 80 90 1000

10

20

30

40

50

60

PNAT

()

GW

con

trib

utio

n (

)

June

July

August

September

October









5 Conclusions










Edited by P Haynes

References




























































Date

Long

itude

Jul Aug Sep Oct Nov

50

100

150

200

250

300

10

20

30

40

50

60

70

80

90

100


Latit

ude




0 10 20 30 40 50 60 70 80 90 1000

10

20

30

40

50

60

PNAT

()

GW

con

trib

utio

n (

)

June

July

August

September

October









5 Conclusions










Edited by P Haynes

References






























































5 Conclusions










Edited by P Haynes

References





























































Edited by P Haynes

References



























































can gravity waves signiﬁcantly impact psc occurrence in...

Documents