Ozone loss from quasi-conservative coordinate mapping during the

1999–2000 SOLVE/THESEO 2000 campaigns

L. R. Lait,1 M. R. Schoeberl,2 P. A. Newman,2 T. McGee,2 J. Burris,2 E. V. Browell,3

E. Richard,4 G. O. Braathen,5 B. R. Bojkov,5 F. Goutail,6 P. von der Gathen,7 E. Kyro,8

G. Vaughan,9 H. Kelder,10 S. Kirkwood,11 P. Woods,12 V. Dorokhov,13 I. Zaitcev,13

Z. Litynska,14 B. Kois,14 A. Benesova,15 P. Skrivankova,15 H. De Backer,16 J. Davies,17

T. Jorgensen,18 and I. S. Mikkelsen18

Received 27 June 2001; revised 27 June 2001; accepted 1 December 2001; published 12 September 2002.

[1] Ozone observations made by the Airborne Raman Ozone, Temperature, and AerosolLidar (AROTEL) and Differential Absorption Lidar (DIAL) on board the NASA DC-8aircraft, the NOAA in situ instrument on board the NASA ER-2 aircraft, and ThirdEuropean Stratospheric Experiment on Ozone 2000 (THESEO 2000) ozonesondes areanalyzed by applying a quasi-conservative coordinate mapping technique. Measurementsfrom the late winter/early spring SAGE III Ozone Loss and Validation Experiment(SOLVE) period (January through March 2000) are incorporated into a time-varyingcomposite field in a potential vorticity-potential temperature coordinate space; ozone lossrates are calculated both with and without diabatic effects. The average loss rate from mid-January to mid-March near the 450 K isentropic surface in the polar vortex is found to beapproximately 0.03 ppmv/d. INDEX TERMS: 1610 Global Change: Atmosphere (0315, 0325); 0340

Atmospheric Composition and Structure: Middle atmosphere—composition and chemistry; 0394 Atmospheric

Composition and Structure: Instruments and techniques

Citation: Lait, L. R., M. R. Schoeberl, P. A. Newman, T. McGee, J. Burris, E. V. Browell, E. Richard, G. O. Braathen, B. R. Bojkov,

F. Goutail, P. von der Gathen, E. Kyro, G. Vaughan, H. Kelder, S. Kirkwood, P. Woods, V. Dorokhov, I. Zaitcev, Z. Litynska, B. Kois,

A. Benesova, P. Skrivankova, H. De Backer, J. Davies, T. Jorgensen, and I. S. Mikkelsen, Ozone loss from quasi-conservative coordinate

mapping during the 1999–2000 SOLVE/THESEO 2000 campaigns, J. Geophys. Res., 107(D20), 8274, doi:10.1029/2001JD000998, 2002.

1. Introduction

[2] Since the discovery of the Antarctic ozone hole byFarman et al. [1985], stratospheric researchers have moni-tored the ozone layer, watching for signs of similar chemical

processes over the Arctic regions. To detect ozone loss,researchers have collected and analyzed measurements fromsatellite instruments such as the Total Ozone MappingSpectrometer (TOMS) [e.g., Newman et al., 1997] andMicrowave Limb Sounder (MLS) [e.g., Manney et al.,1997]. Just as important, however, have been data fromAmerican and European field experiments such as the Air-borne Arctic Stratospheric Expeditions (AASE I and II) andthe European Arctic Stratospheric Ozone Experiment(EASOE), as outlined by Turco et al. [1990], Rodriguez[1993], and Pyle et al. [1994], respectively. Measurementsfrom such experiments typically involve a large number ofspecies from relevant families of atmospheric constituents,and these data—frequently of high spatial and temporalresolution—enable a fairly detailed understanding of thechemical processes associated with stratospheric ozone loss[see, for example, Salawitch et al., 1993].[3] During the winter of 1999–2000, the SAGE III

Ozone Loss and Validation Experiment (SOLVE) wascarried out simultaneously with the Third European Strato-spheric Experiment on Ozone 2000 (THESEO 2000) cam-paign. Meteorological and trace gas constituent data werecollected by many different in situ and remote sensinginstruments. These included measurements of stratosphericozone by airborne lidars, an airborne UV-absorption photo-meter, and ozonesondes.

JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 107, NO. D20, 8274, doi:10.1029/2001JD000998, 2002

1Science Systems and Applications, Inc., Lanham, Maryland, USA.2NASA Goddard Space Flight Center, Greenbelt, Maryland, USA.3NASA Langley Research Center, Hampton, Virginia, USA.4NOAA Aeronomy Laboratory and University of Colorado, Boulder,

Colorado, USA.5Norwegian Institute for Air Research, Norway.6Centre National de la Recherche Scientifique, France.7Alfred Wegener Institute for Polar and Marine Research, Potsdam,

Germany.8Finnish Meteorological Institute, Sodankyla, Finland.9University of Wales, Aberystwyth, UK.10Royal Netherlands Meteorological Institute (KNMI), De Bilt,

Netherlands.11Swedish Institute of Space Physics, Sweden.12National Physical Laboratory, UK.13Central Aerological Observatory, Moscow, Russia.14Institute of Meteorology and Water Management, Legionowo, Poland.15Czech Hydrometeorological Institute (CHMI), Prague, Czech

Republic.16Royal Meteorological Institute of Belgium (KMI/IRM), Belgium.17Meteorological Service of Canada, Ontario, Canada.18Danish Meteorological Institute, Copenhagen, Denmark.

Copyright 2002 by the American Geophysical Union.0148-0227/02/2001JD000998

SOL 16 - 1

[4] Given the variety of ozone measurements taken bydifferent instruments at different times and scattered loca-tions, it is desirable to put these data into a consistent wide-scale meteorological framework. As outlined by Schoeberland Lait [1991], constituent mapping using quasi-conservedcoordinates can be a useful technique for this purpose. Acomposite field of trace gas constituent measurements isconstructed in a vortex-relative coordinate system usingpotential vorticity (PV) and potential temperature (q) asabscissa and ordinate, respectively. Then, given griddedmeteorological fields of PV and q, one can map the compo-site into real space, creating a three-dimensional field of thereconstructed constituent. Schoeberl et al. [1989] and Lait etal. [1990] demonstrated examples of this ‘‘PV-q’’ technique.Related techniques have been employed by Manney et al.[1994], Lary et al. [1995], and Kyro et al. [2000].[5] In this paper, we refine the PV-q technique and use it

to investigate ozone depletion observed during the latewinter/early spring of 2000. Section 2 describes the dataused in this analysis, and section 3 describes the enhancedtechnique. Results are presented in section 4, and theconclusions follow.

2. Data

2.1. Meteorological Data

[6] PV-q analysis requires that each measurement beassociated with a value of PV and q, and PV at least mustbe obtained from gridded meteorological analyses. Threesources of such analyses were used for this work. The NASAGoddard Space Flight Center’s Data Assimilation Office(DAO) product for the SOLVE period is obtained from theirGEOS-3 assimilation system for EOS-Terra support. GEOS-3 is the successor to the GEOS-1 system documented byPfaendtner et al. [1995]. These data grids extend from 1000to 0.2 hPa, have a horizontal resolution of 1� longitude by 1�latitude, and are produced four times daily.[7] The United Kingdom Meteorological Office (UKMO)

product generated for the Upper Atmosphere ResearchSatellite (UARS) project is another assimilation effort,described by Swinbank and O’Neil [1994]. The data usedhere extend from 1000 to 0.4 hPa, have a horizontalresolution of 3.75� longitude by 2.5� latitude, and areproduced once per day.[8] The third source of meteorological data used here is

from the long-term assimilation performed by the NationalCenters for Environmental Prediction (NCEP) and theNational Center for Atmospheric Research (NCAR). Theprocedures of the NCEP/NCAR reanalysis system areapplied consistently to over 40 years of raw data, resultingin a data set which is useful for long-term studies (for furtherinformation, see Kalnay et al. [1996]). These data extendfrom 1000 to 10 hPa, have a horizontal resolution of 2.5�longitude by 2.5� latitude, and are produced four times daily.[9] To account for diabatic effects, heating rates were

calculated from the UKMO analyses using the modeldescribed by Rosenfield et al. [1994].

2.2. Ozone Data

[10] SOLVE consisted of three deployments: 2–14December 1999; 10 January to 3 February 2000; and 26February to 16 March 2000. During these periods, the

NASA DC-8 and ER-2 aircraft flew numerous sorties outof Kiruna, Sweden.[11] On board the ER-2, a dual beam UV-absorption

photometer measured ozone volume mixing ratios at flightaltitudes (above 18 km). Proffitt and McLaughlin [1983]describe the instrument. These data are well-calibrated andof good accuracy (around 3%). And at one measurement persecond on an aircraft moving at approximately 200 m/s,they have high spatial and temporal resolution. However,their coverage is limited to the actual position of the aircraft.[12] In contrast, two lidar instruments on board the DC-8

measured profiles of ozone above the aircraft well into thestratosphere. The UV Differential Absorption Lidar (DIAL)measurements have a vertical resolution of approximately750 m and are taken at least 5 min apart (giving a horizontalresolution of up to 70 km). DIAL profiles reach to 25 or 30km. DIAL measures ozone number densities, and the DAOanalyzed temperatures are used to convert these to volumemixing ratios. See Browell et al. [1998] for further descrip-tion of the instrument.[13] The Airborne Raman Ozone, Temperature, and

Aerosol Lidar (AROTEL) uses Rayleigh and Raman scatter-ing to obtain temperature and ozone profiles with a verticalresolution of 0.5 to 1.5 km and which are taken at least 2 minapart (giving a horizontal resolution of up to 24 km).AROTEL profiles reach up to 35 to 40 km. Like DIAL,AROTEL obtains ozone number densities, but AROTELuses its own temperature measurements to convert to vol-ume mixing ratios (J. Burris, personal communication).[14] These three aircraft-based instruments provide strato-

spheric ozone measurements of high resolution along theirflight tracks which cut cross large areas of the Arctic betweenGreenland and Novaya Zemlya. They are limited to flightdays within the three SOLVE deployment periods, however.[15] Ozonesondes launched by the THESEO 2000 cam-

paign, the Meteorological Service of Canada, the WorldMeteorological Organization network, Japan, and Russiameasured profiles at a number of sites in the Arctic. Usingan electrochemical concentration cell, these instruments takemeasurements with a vertical resolution between 60 and 150m and typically reach up to around 30 km. Although thelaunch sites are geographically sparse, these profiles exist fornearly every day of the period examined here.

3. Analysis

[16] PV was obtained from the meteorological analysisclosest in time to each measurement by interpolating thegridded field bilinearly in the horizontal and using cubicspline interpolation in log-pressure for the vertical. The lidarinstruments’ profiles are a function of geometric heightabove the aircraft; the analyzed geopotential heights onpressure surfaces were converted to geometric heights andused to obtain the pressure of each measurement along theflight track. Modified PV [Lait, 1994] was used to avoid thestrong vertical scaling exhibited by regular Ertel’s PV. Forconsistency, the q values were similarly interpolated fromthe meteorological analyses, even for instruments whichmeasured temperature.[17] A time-varying ozone composite was constructed on

a grid in PV-q space. The edges of the grid were chosen tobe well-removed from the values typical of the area of

SOL 16 - 2 LAIT ET AL.: OZONE LOSS DURING THE 1999–2000 SOLVE/THESEO 2000 CAMPAIGNS

interest (the Arctic lower stratosphere): 0 to 50 PV Units(One PVU being one 106 K m2/kg s) and 200 to 1000 K inq. The grid coordinates were nondimensionalized to sim-plify the data weighting calculations.[18] A reasonably long-lived trace gas is expected to be

well-mixed along contours of PV on a surface of constant q[Leovy et al., 1985], and in the absence of diabatic effects andchemical changes a simple composite made by averagingmixing ratios near a grid point should yield an accurate pictureof a time-invariant trace gas distribution in PV-q space. Such acomposite should be constructed only over time periods shortenough that diabatic effects can be neglected.

[19] But over the course of the 10 weeks examined hereone cannot neglect diabatic effects or chemical ozone loss.Both effects appear as a change in ozone mixing ratio at agiven point in PV-q space. One must take into account thistime-varying nature of the field when creating the PV-qcomposite.[20] Over periods of a few weeks (shorter than seasonal

timescales), the ozone mixing ratio c at fixed PV and qvalues can be approximated as a linear change in time t:

c PV; qð Þ ¼ a PV; qð Þ þ b PV; qð Þt ð1Þ

To make the ozone PV-q composite, constants a(PV, q) andb(PV, q) were computed by constructing a time series ateach grid point of all the nearby ozone measurements. Alinear least squares fit in time was applied to themeasurements, which were weighted by

wi ¼ 1þ 1=ffiffiffiffisi

pð Þexp �d2i� �

ð2Þ

where si is the estimated or quoted uncertainty in the ithmeasurement and di is a scaled nondimensional distancefrom the ith measurement to the grid point in PV-qcoordinates. A typical time series and fit for the ozonesondedata is shown in Figure 1.[21] The results of the time fits are two fields a(PV, q) and

b(PV,q), containing intercepts and slopes, respectively. Anozone field in PV-q space can be reconstructed from a and bfor any day. This reconstructed ozone grid can be mappedonto points in real space for which PV and q values areknown in order to obtain nominal values of ozone there.Figure 2, for example, shows ozonesonde data which hasbeen mapped onto vertical profiles above the NASA DC-8for comparison with the AROTEL data. The figure showsthe comparison for the flight of 9 March, when the DC-8flew back and forth across the vortex edge. The remappedsonde data generally compare well with the AROTEL

0 20 40 60 80Days since 00 UTC on 10 January, 2000

0

2

4

6

Ozo

ne (

ppm

v)

Slope = -0.029 +/- 0.007 ppmv/day

Figure 1. Time series of ozonesonde mixing ratiomeasurements near PV = 32 PVU and q = 450 K. Thegray shading of each symbol indicates its weighting in thelinear time fit; the size of each symbol indicates itsproximity to the PV-q grid point. The linear time fit andits uncertainty (± one standard deviation) are overlaid.Dashed vertical lines mark 1 February and 1 March.

Figure 2. Ozone as a function of flight time and altitude for the DC-8 flight of 9 March 2000, (a) asmeasured by the AROTEL instrument on board the DC-8, (b) using THESEO 2000 ozonesondemeasurements mapped into PV-q space and then mapped onto the AROTEL measurement locations.

LAIT ET AL.: OZONE LOSS DURING THE 1999–2000 SOLVE/THESEO 2000 CAMPAIGNS SOL 16 - 3

measurements. Two properties of the reconstructed fieldstand out. First, its spatial resolution is much less than thatof the original data. This is not surprising, considering theaveraging process and the limited resolution of the mete-orological analyses. Nevertheless, the reconstructed fielddoes show the edge structure clearly. Second, the recon-structed ozonesonde data cover a much larger part of theflight than the actual AROTEL measurements, despite thefact that the raw sonde profiles are much more sparselydistributed over the Arctic region than the AROTEL pro-files. The reconstruction process assumes that ozone valuesmeasured under (and averaged within) certain meteorolog-ical conditions will be typical of all ozone values inidentical conditions, and maps them accordingly.[22] In a similar way, one can use gridded three-dimen-

sional fields of PV and q to map the ozone composite ontothe fields’ grid points to obtain an ozone field covering alarge area. After using climatological values to fill inregions above and below the reconstructed profiles, sucha field can be integrated vertically, yielding a total ozonefield that can be compared with TOMS measurements. A

comparison for a typical day is shown in Figure 3. PV-qreconstruction reproduces the morphology of the totalozone field fairly well, and the values themselves arereasonable. For lower total ozone amounts, the values fromthe reconstruction are higher than the TOMS measuredvalues. Inspection of ozonesonde and reconstructed ozoneprofiles indicates that these differences are chiefly causedby uncertainties in the reconstructed values in the lower-most stratosphere/upper troposphere, where small varia-tions in ozone mixing ratio can result in large variationsin total ozone.[23] Reconstructed fields can be used to determine ozone

change rates by differencing the fields at the beginning andend of a time period, or by mapping the slopes from theleast squares fits into real space. However, the changecaused by diabatic descent of air in the polar vortex andthe change caused by chemical destruction of ozone will beindistinguishable. A way is needed to account for thediabatic descent to enable the rate of chemical destructionto be determined. This analysis uses two methods foraccounting for diabatic effects.

Figure 3. (a) Total ozone fields for 10 March 2000 as measured by TOMS. Black areas are those forwhich there are no measurements (polar night). (b) Total ozone fields for 10 March 2000 as reconstructedfrom sonde, lidar, and ER-2 ozone data using the PV-q reconstruction technique. Black areas are those forwhich not enough of the vertical profile could be reconstructed to yield a meaningful total ozone value.(c) Scatterplot of reconstructed values versus the TOMS measurements.

SOL 16 - 4 LAIT ET AL.: OZONE LOSS DURING THE 1999–2000 SOLVE/THESEO 2000 CAMPAIGNS

[24] To determine how much descent took place withinthe vortex over this period, air parcel trajectories weretraced using an isentropic trajectory model with diabaticcorrections (see Schoeberl et al. [1998], but here the netdiabatic heating rate is not balanced, as given by Schoeberlet al. [2002]). Parcels were initialized at approximatelyequally spaced grid points on q surfaces spaced 10 K apartfrom 400 to 600 K, and the trajectories were run bothforward and backward between 10 January and 18 Marchusing the UKMO meteorological data and heating ratesfrom the Rosenfield model.[25] To obtain descent rates within the vortex core, only

those parcel trajectories which started out and ended up inthe core of the vortex were retained. The vortex core wasfound by determining the PV value that marks the highest10% all PV values on the (equidistant) grid points which liepoleward of 50�N on each theta surface. These PV valueswere used to delimit the core of the vortex on the beginningand ending dates; this practice was supported by inspectionof contours of these values on maps of potential vorticity.[26] For each of the starting q surfaces, the parcels’

ending q values were subtracted to get the average descentover this period. Figure 4 shows the result. Parcels startingat 600 K in mid-January had descended by some 40 K, onaverage, by mid-March. Parcels in the lower stratosphere,between 400 and 500 K, descended roughly 15 to 20 Kduring the same time. To get the average diabatic descent asa function of initial potential temperature, two straight lineswere fit to the data, meeting at 475 K.[27] To determine ozone loss, ozone fields were recon-

structed on 10 January and 17 March on pairs of potentialtemperature surfaces within the vortex core, the 17 Marchsurfaces being adjusted upward to compensate for the descentof air. The ozone decline was calculated by differencing thesetwo vertical profiles; this method will be referred to as the‘‘vortex-core average descent’’ method below.

[28] The second method used to remove diabatic effectsinvolved tracing the parcel trajectories of the ozone measure-ments themselves. If the trajectory model were perfect, thenparcels started on 10 January with certain PV and q coor-dinates would end on 18 March with different values of PVand q. If these ‘‘diabatically adjusted’’ PV-q coordinates areused in the linear time fits, then the slopes of those fits shouldreflect chemical loss alone. (Schoeberl et al. [2002] uses asimilar technique.) This method will be termed ‘‘diabaticcoordinate adjustment’’ in the discussion which follows.[29] Of course, trajectories of individual parcels cannot

be traced accurately for such long periods, but Schoeberland Sparling [1995] indicate that reasonable results can beobtained using large ensembles of trajectories whose resultsare interpreted in a statistical sense.

4. Results

4.1. Vortex-Core Average Descent Method

[30] To apply the vortex-core average descent method ofdetermining loss, ozone measurements from the ER-2 in situinstrument, the DC-8 AROTEL and DIAL lidar instru-ments, and the THESEO 2000 ozonesondes were combinedinto a single PV-q composite. Measurements from eachinstrument were thinned out until they were of approxi-mately equal density in PV-q space, so that no one instru-ment would dominate the analysis. The resulting compositewas used to reconstruct ozone on potential temperaturesurfaces for 10 January and 17 March. The surfaces for17 March were adjusted upwards to account for diabaticdescent as determined by the fit in Figure 4. Reconstructedgrid points of ozone in the core of the vortex were selectedand area-averaged to obtain two vertical profiles, one ateach end of the time period being examined.[31] Figure 5 shows these profiles. The difference

between these curves, divided by the time period, is theaverage chemical ozone loss rate. The ozone loss ratecalculated here peaks at 470 K with a value of 0.026ppmv/d and decreased to near 0 at 550 to 650 K. Resultsare shown using the NCAR/NCEP Reanalysis; the UKMOanalyses and DAO data produce similar curves with lossrates of 0.025 and 0.026, respectively, at 470 K.[32] The profiles in Figure 5 extend down to 350 K and

show the loss rates continuing to shrink below the 400 Klevel. The PV-q technique is more suitable for use in thestratosphere, where (modified) PV and q are at least quasi-conserved and quasi-orthogonal. Nevertheless, the techni-que seems to yield reasonable values for loss rates downinto the very lowest part of the stratosphere and into theuppermost troposphere, and possible spurious trends arisingfrom movement of the tropopause and its attendant sharpozone gradient do not appear.

4.2. Diabatic Coordinate Adjustment Method

[33] To apply the diabatic coordinate adjustment method,measurements from the ozonesondes were used, since theywere sparse enough that their trajectories could be tracedeasily. All days’ sonde profiles were run forward to 18March to obtain their diabatically adjusted PV and q valueson that date. They were also run backward to obtain theirdiabatically adjusted PVand q values on 10 January. Parcelswhose PV changed by more than 25% over this period were

400 450 500 550 600 650 700Jan 10 Theta (K)

0

20

40

60

80D

esce

nt (

K)

Figure 4. Vortex descent between 10 January 2000 and 18March 2000 as a function of starting potential temperature.Crosses mark parcels from forward trajectories, anddiamonds mark parcels from backward trajectories. Twolinear least squares fits have been drawn through the points.

LAIT ET AL.: OZONE LOSS DURING THE 1999–2000 SOLVE/THESEO 2000 CAMPAIGNS SOL 16 - 5

discarded to reduce the effects of parcels entering or leavingthe vortex. The remaining measurements, with theiradjusted PV-q coordinates, were then used to construct atime-varying composite, and the slopes from the resultingtime fits were obtained. Uncertainties in the slopes werealso computed from the covariance matrices of the time fitcoefficients (as shown by Meyer [1975]).[34] The loss rates were mapped onto vertical profiles in

the vortex core; Figure 6 shows the area-weighted averageof these profiles. Figure 6a shows the results for the forwardtrajectory run on 10 January. The loss rate peaks at 0.032 ±

0.007 ppmv/d at 460 K. This is larger than that computedfrom the vortex-core average descent method, but it peaksin about the same altitude. (Vortex-core average descent lossrates computed from the ozonesonde data alone had thesame peak location and value as from the composite of allthe instruments.) Figure 6b shows the results for the backtrajectory run on 17 March; the loss rate peaks around 0.030± 0.008 ppmv/d at 440 K. The peak is about the samemagnitude (indicating agreement between the forward andback trajectory runs), but it is located about 20 K lower inpotential temperature.

0 2 4 6O3 (ppmv)

350

400

450

500

550

600

650

Pot

entia

l Tem

pera

ture

(K

)

350

400

450

500

550

600

650

Pot

entia

l Tem

pera

ture

(K

)

-0.04 -0.02 0.00 0.02 0.04O3 Rate of Change (ppmv/day)

-2 -1 0 1 2O3 Change (ppmv)

(b)(a)

Figure 5. (a) Reconstructed ozone profiles (area-averaged over the core of the vortex) from 12 Januaryto March 17. The March profile has had its q values adjusted to remove diabatic descent over the period.(b) Ozone average chemical rates of change for the vortex core, calculated by subtracting the two curvesfrom (a).

-0.06-0.06 -0.04 -0.04-0.02 -0.02 0.000.00 0.020.02 0.04 0.04O3 Rate of Change (ppmv/day) O3 Rate of Change (ppmv/day)

350

400

450

500

550

600

650

Pot

entia

l Tem

pera

ture

(K

)

(a) (b)

Figure 6. Ozone chemical rate of change (from the time fit slopes of diabatically adjusted sonde data)as a function of q for the vortex core region for (a) 10 January using diabatic corrections from forwardtrajectories and (b) 17 March using diabatic corrections from back trajectories.

SOL 16 - 6 LAIT ET AL.: OZONE LOSS DURING THE 1999–2000 SOLVE/THESEO 2000 CAMPAIGNS

[35] The back trajectory loss rates mapped to the 10January profile are very similar to the forward trajectoryrates for that date. Likewise, the forward trajectory resultsmapped to 17 March are similar to the back trajectoryresults for that date. That is, both trajectory runs showsimilar peak loss rates which descend in q over the timeperiod examined. Figure 7 shows both rates mapped onto 13February profiles (the middle of the period). One differencebetween the two runs is that the forward trajectory profilesexhibit a small offset near the bottom of the profile (and arevery nearly zero at the top), while the backward trajectory

profiles show a small offset near the top of the profiles (andare very nearly zero at the bottom).[36] A third set of trajectories was calculated in which the

parcels were all traced to 13 February. To reach this date,measurements made earlier were run forward, and measure-ments made later were run backwards. Thus, the averagelength of a trajectory was reduced by approximately one-half. The results are in agreement with the others; at 450 K on13 February, the average loss rate is 0.026 ± 0.005 ppmv/d.[37] Figure 8 shows the time fit slopes (from the back-

wards trajectory run) for the 440 K surface as a function of

-0.06-0.06 -0.04 -0.04 -0.02-0.02 0.00 0.000.02 0.02 0.040.04O3 Rate of Change (ppmv/day)O3 Rate of Change (ppmv/day)

350

400

450

500

550

600

650

Pot

entia

l Tem

pera

ture

(K

)

(a) (b)

Figure 7. Ozone chemical rate of change (from the time fit slopes of diabatically adjusted sonde data)as a function of q for the vortex core region on 13 February using (a) diabatic corrections from forwardtrajectories (b) diabatic corrections from back trajectories.

50 60 70 80 90Equivalent Latitude

-0.06

-0.04

-0.02

0.00

0.02

O3

Rat

e of

Cha

nge

(ppm

v/da

y)

12 UTC on 13 February, 2000 at 440 K

Figure 8. Ozone chemical rate of change (from the time fit slopes of diabatically adjusted sonde data)as a function of equivalent latitude at the 440 K q surface on 13 February 2000. The vertical dotted linemarks the core of the polar vortex.

LAIT ET AL.: OZONE LOSS DURING THE 1999–2000 SOLVE/THESEO 2000 CAMPAIGNS SOL 16 - 7

equivalent latitude [Butchart and Remsberg, 1986] for 13February. The area-weighted average loss rate within theinner vortex region on the 440 K surface is 0.030 ± 0.009,the same as seen in the profile results. Expanding theaverage to include most of the vortex decreases the lossrate slightly to 0.028 ± 0.007 at 440 K.[38] Results from the backwards trajectory run are similar

within the vortex, but the change rates outside the vortexvary greatly between the two runs (from approximately�0.01 to +0.02 ppmv/d at 470 K). The near-zero loss rateshown in Figure 8 outside the vortex must be consideredhighly uncertain.[39] From equation 1, the loss rate of total ozone OT

becomes:

@OT=@t ¼ @ W

Z ps

0

cdp� �

=@t ð3Þ

¼ W

Z ps

0

b PV; qð ÞdpþWc psð Þ@ps=@t ð4Þ

where W = 0.789 m s2/kg and ps is the surface pressure. Wedo not reconstruct ozone values in the lower troposphere,but fixing the (notional) ozone profile to zero at the surface(as is done to derive the total ozone values mapped inFigure 3) makes the second term vanish. (As a dynamicaleffect, the second term has no bearing on chemical ozoneloss anyway.) Thus, integrating the time slopes yields thatpart of the time rate of change of total ozone which iscaused by chemical destruction.[40] Time fit slopes from the diabatically adjusted sonde

data were mapped onto vertical profiles as a function ofpressure. Loss rates from the back trajectory results wereused because they were close to zero below 400 K, a regionmore heavily weighted in the integration. Any ozoneincreases (e.g., from the offsets near the top of the profiles)were zeroed out—only ozone destruction rates are ofinterest. These rate profiles were then integrated and area-averaged within the vortex core for each day between 10January and 17 March.

[41] Figure 9 shows the daily total ozone loss rates for thecore of the vortex. The losses begin at approximately 0.8DU/d and increase to 1.3 DU/d by 17 March. The loss rates’evolution is associated with a decrease in q at the ozone losspeak and with a stronger shift in the PV values associatedwith the vortex core.

5. Discussion and Conclusions

5.1. Analysis Uncertainties

[42] The PV-q analysis has several potential pitfalls. Itdepends on its constituent being long-lived enough to bewell-mixed along contours of isentropic potential vorticity.In the middle stratosphere and above, where photochemistrydominates, the method is of limited applicability to ozone.This analysis has been confined to the lower stratosphere.[43] When an ozone measurement is made near the

vortex edge in the sunlight, where chlorine-driven photo-chemical destruction is likely to be enhanced, this measure-ment may not be typical of ozone values around a PVcontour on a q surface that intersects the measurementlocation. With enough additional measurements around thatcontour, PV-q analysis will yield an averaged view of theozone field which the ozone-depleted measurements willinfluence but not necessarily dominate. Poor sampling,however, may lead to the sunlit region being underrepre-sented or overrepresented in the average until the ozone-depleted air is mixed along the contour. This analysis relieson the time series fit to average out the sudden drops inozone that might appear in such cases.[44] To distinguish between diabatic effects and chemical

destruction of ozone, this analysis uses diabatic trajectoriesof large numbers of parcels traced over about 10 weeks.Averaging the effects of large numbers of parcels might wellcompensate for the inaccuracy of individual parcel trajecto-ries, but there remains the possibility of systematic artifactsof the meteorological analysis used or of the radiationmodel used to generate the heating rates.[45] The diabatic heating rates have been verified in

previous work [Rosenfield and Schoeberl, 2001] for theUKMO meteorological products, and they appear to bereasonable.[46] Another approach would be to analyze a tracer such

as N2O and test whether the diabatic trajectories reduce theapparent rate of change, which should be a purely diabaticeffect. This has been attempted, and the results suggest thatusing diabatically adjusted trajectories tends to reduce therates of change in the vortex core at altitudes higher than500 K. However, the N2O measurements from SOLVE weresparse enough, and the uncertainties in the rates are con-sequently large enough to make these results inconclusive.That is, the differences between the uncorrected rate profilesand the diabatically corrected rate profiles are smaller thanthe uncertainties in either.[47] Another effect which would complicate matters is

mixing across the vortex boundary. Although the polar vortexis generally recognized as being isolated from the midlati-tudes [Bowman, 1993; Schoeberl et al., 2002], it is notcompletely isolated—slow or weak mixing may occur overthe course of a season, in addition to occasional majorintrusion events such as one described by Plumb et al.[1994]. The PV-q technique may still reconstruct ozone fields

Vortex Core Total Ozone Loss Rate

0.6

0.8

1.0

1.2

1.4

1.6

Loss

Rat

e (D

U/d

ay)

1-10 1-20 2-01 2-10 2-20 3-01 3-10Date

Figure 9. Vertically integrated chemical ozone loss ratesfor 10 January through 17 March, within the core of thepolar vortex.

SOL 16 - 8 LAIT ET AL.: OZONE LOSS DURING THE 1999–2000 SOLVE/THESEO 2000 CAMPAIGNS

with some degree of fidelity in these circumstances, butuntangling chemical destruction from mixing becomes prob-lematic. In this analysis, the issue is addressed by excludingparcels whose diabatic trajectories involved a change of PVofmore than 25%, but it is unclear whether this is sufficientgiven the limitations of tracing individual parcels. In otherwords, it is difficult to say whether a parcel has been excludedbecause it moved into or out of the vortex, or because itstrajectory is simply not accurately traced.[48] Inspection of the grid point trajectories used to

determine descent rates shows only a few cases wheretrajectories which originated outside the vortex in Januaryended up inside the vortex in March: about 1.3% of the mid-March vortex parcels at 420–430 K for the back-trajectoriesonly. For other theta levels and for the forward trajectoryrun, no such parcels were found. Schoeberl et al. [2002],who used a more comprehensive trajectory modeling run,support the assertion that mixing into the vortex wasnegligible over most of this period.[49] Toward the middle of March, as the vortex was

breaking up, the effects of mixing are expected to increase.A sharper drop-off in ozone at the end of the time series canbe seen in Figure 1, and it is difficult to separate the effectsof mixing from those of chemical destruction in theincreased sunlight. This analysis partially side-steps theissue by fitting a linear trend over the 10-week period, sothat the sudden drop-off at the end does not influence theoverall trend too strongly.

5.2. Other Techniques

[50] PV-q analysis of diabatically adjusted sonde datagives an average ozone loss rate of around 0.03 ± 0.008ppmv/d near 440 to 470 K. Strictly speaking, cumulativeozone loss should be estimated from a Lagrangian frame-work, following air parcels. But because the rates arederived from parcels whose coordinates have been correctedfor diabatic effects, and because the vortex is fairly isolated,a rough estimate for the total loss can be obtained bymultiplying the rate by the length of the time period:approximately 2.0 ppmv.[51] Other techniques have been used to categorize ozone

loss. Instead of placing all the data into a composite time-dependent field, the Match program [Rex et al., 1999] usespaired individual ozonesonde profiles to determine ozoneloss. Match selects the pairs by tracing the trajectories ofsonde-sampled air parcels until they intersect other sondelaunches days later. By using small clusters of trajectories foreach measurement, and by averaging differenced pairstogether, Match tends to reduce the uncertainties associatedwith the trajectory calculations. Because Match uses changesin PV as a criterion in matching pairs of sonde launches, itsozone values are effectively categorized by PV in a waysimilar to PV-q analysis. Instead of dealing with pairwisedifferences at discrete q levels, though, the PV-q techniquehas the effect of averaging all ozone measurements of similarPVand q values. More data will tend to enter such an averageat the cost of a higher dependence upon the assumption thanozone is well-mixed along an isentropic PV contour. Matchresults for SOLVE [Rex et al., 2002] are similar to thosepresented here, about 2.6 ppmv cumulative loss near 460 K.[52] The vortex-averaged analysis of Schoeberl et al.

[2002] involves diabatic trajectory calculations initialized

from massive numbers of ozone measurement locations, sothat the inaccuracies associated with individual parcels areaveraged out. The data which remain inside the vortex are fitto a quadratic curve at various theta levels to obtain ozoneloss rates from December 1999 through mid-March 2000(around 0.025 ppmv/d). Solar exposure is also used todiscriminate between parcels in the core of the vortex andparcels from the edge region; excluding parcels near thesunlit edge is important for separating out parcels with higherloss rates which might be seen in January. Their analysisyields an ozone loss rate of approximately 0.025 ppmv/d.[53] Another analysis that uses trajectory calculations to

correct for diabatic effects has been performed by Swartz etal. [2002]. From stellar occultation measurements analyzedby applying a vortex-averaged descent correction and bytracing individual trajectories, they found an average dailyloss rate of about 0.024 ppmv/d from 23 January to 4March.[54] The PV-q technique with diabatic corrections may be

thought of as a way of enhancing such trajectory analyses toproduce a continuously evolving ozone field. The peakozone loss rates produced here are somewhat less thanMatch’s, but somewhat higher than that obtained by theSchoeberl trajectory analysis. The peak loss rates seem to bea little lower and broader in altitude in the latter results, aswell, but this may be an artifact of the heavy smoothingdone here.

5.3. Summary

[55] Ozone measurements from AROTEL, DIAL, theNOAA ER-2 instrument, and the THESEO 2000 ozone-sondes have been combined using a PV-q analysis techniquewhich allows the ozone field to evolve in time. Realisticozone fields have been obtained by mapping the ozone fieldfrom PV-q coordinates back into real space. Diabatic effectshave been estimated using modeled trajectories of largenumbers of parcels. Estimates of chemical ozone loss in thelower stratosphere are found to peak at 0.030 ± 0.009 ppmv/d near 460 K in January, with this peak descending to 440 Kin March. The loss rates decrease to near 0 at 550–600 Kand below 400 K. The cumulative loss near 450 K is thusabout 2 ppmv over the period. These results appear to be inbroad agreement with other analyses.

[56] Acknowledgments. The author wishes to acknowledge the DC-8and ER-2 pilots and the flight and ground crews who made the aircraftmeasurements possible under sometimes difficult conditions; the sondelaunch personnel; the SOLVE logistics staff and the personnel of the ArenaArctica in Kiruna, Sweden; and SOLVE and THESEO 2000 management.This research was supported by the NASA’s Atmospheric ChemistryModeling and Analysis Program and the Upper Atmosphere ResearchProgram.

ReferencesBowman, K. P., Large-scale isentropic mixing properties of the Antarcticpolar vortex from analyzed winds, J. Geophys. Res., 98, 23,013–23,027,1993.

Browell, E. V., S. Ismail, and W. B. Grant, Differential absorption lidar(DIAL) measurements from air and space, Appl. Phys. B, 67, 399–410,1998.

Butchart, N., and E. E. Remsberg, The area of the stratospheric polar vortexas a diagnostic for tracer transport on an isentropic surface, J. Atmos. Sci.,43, 1319–1339, 1986.

Farman, J. C., B. G. Gardiner, and J. D. Shanklin, Large losses of totalozone in Antarctica reveal seasonal ClOX/NOX interaction, Nature, 315,207–210, 1985.

LAIT ET AL.: OZONE LOSS DURING THE 1999–2000 SOLVE/THESEO 2000 CAMPAIGNS SOL 16 - 9

Kalnay, E., et al., The NCEP/NCAR 40-year reanalysis project, Bull. Am.Meteorol. Soc., 77, 437–471, 1996.

Kyro, E., R. Kivi, T. Turunen, H. Aulamo, V. V. Rudakov, V. Khattatov,A. R. MacKenzie, M. P. Chipperfield, A. M. Lee, L. Stefanutti, and F.Ravegnani, Ozone measurements during the Airborne Polar Experiment:Aircraft instrument validation, isentropic trends, and hemispheric fieldsprior to the 1997 Arctic ozone depletion, J. Geophys. Res., 105,14,599–14,611, 2000.

Lait, L., An alternative form for potential vorticity, J. Atmos. Sci., 51,1754–1759, 1994.

Lait, L. R., et al., Reconstruction of O3 and N2O fields from ER-2, DC-8,and balloon observations, Geophys. Res. Lett., 17, 521–524, 1990.

Lary, D. J., M. P. Chipperfield, J. A. Pyle, W. A. Norton, and L. P. Riishøj-gaard, Three-dimensional tracer initialization and general diagnosticsusing equivalent PV latitude-potential-temperature coordinates, Q. J. R.Meteorol. Soc., 121, 187–210, 1995.

Leovy, C. B., C.-R. Sun, M. H. Hitchman, E. E. Remsberg, J. M. RussellIII, L. L. Gordley, J. C. Gille, and L. V. Lyjak, Transport of ozone in themiddle stratosphere: Evidence for wave breaking, J. Atmos. Sci., 42,230–244, 1985.

Manney, G. L., et al., Chemical depletion of ozone in the Arctic lowerstratosphere during winter 1992–93, Nature, 370, 429–434, 1994.

Manney, G. L., L. Froidevaux, M. L. Santee, R. W. Zurek, and J. W. Waters,MLS observations of Arctic ozone loss in 1996–97, Geophys. Res. Lett.,24, 2697–2700, 1997.

Meyer, Stewart L., Data Analysis for Scientists and Engineers, 513 pp.,John Wiley, New York, 1975.

Newman, P. A., J. F. Gleason, R. D. McPeters, and R. S. Stolarski, Anom-alously low ozone over the Arctic, Geophys. Res. Lett., 24, 2689–2692,1997.

Pfaendtner, J., S. Bloom, D. Lamich, M. Seablom, M. Sienkiewicz, andA. da Silva, Documentation of the Goddard Earth Observing System(GEOS) data assimilation system—Version 1, NASA Tech. Memo.10,4606, vol. 4, 1995.

Plumb, R. A., D. W. Waugh, R. J. Atkinson, P. A. Newman, L. R. Lait,M. R. Schoeberl, E. V. Browell, A. J. Simmons, and M. Lowenstein,Intrusions into the lower stratosphere Arctic vortex during the winter of1991–1992, J. Geophys. Res., 99, 1089–1105, 1994.

Proffitt, M. H., and R. J. McLaughlin, Fast-response dual-beam UV-absorp-tion ozone photometer suitable for use on stratospheric balloons, Rev. Sci.Instrum., 54, 1719–1728, 1983.

Pyle, J. A., et al., An overview of the EASOE campaign, Geophys. Res.Lett., 21, 1191–1194, 1994.

Rex, M., et al., Chemical ozone loss in the Arctic winter 1994/1995 asdetermined by the Match technique, J. Atmos. Chem., 32, 35–59, 1999.

Rex, M., et al., Chemical depletion of Arctic ozone in winter 1999/2000,J. Geophys. Res., 10.1029/2001JD000533, in press, 2002.

Rodriguez, J., Probing stratospheric ozone, Science, 261, 1128–1129,1993.

Rosenfield, J. E., and M. R. Schoeberl, On the origin of polar vortex air,J. Geophys. Res., 106, 33,485–33,497, 2001.

Rosenfield, J. E., P. A. Newman, and M. R. Schoeberl, Computations ofdiabatic descent in the stratospheric polar vortex, J. Geophys. Res., 99,16,677–16,689, 1994.

Salawitch, R. J., et al., Chemical loss of ozone in the Arctic polar vortex inthe winter of 1991–1992, Science, 261, 1146–1149, 1993.

Schoeberl, M. R., and L. R. Lait, Conservative coordinate transformations

for atmospheric measurements, in The Use of EOS for Studies in Atmo-spheric Physics, edited by J. C. Gille and G. Visconti, Proc. Int. School ofPhys. ‘‘Enrico Fermi,’’ CXV Course, pp. 419–430, 1991.

Schoeberl, M. R., and L. C. Sparling, Trajectory modelling, in DiagnosticTools in Atmospheric Physics, edited by J. C. Gille and G. Visconti, Proc.Int. School of Phys. ‘‘Enrico Fermi,’’ vol. 124, pp. 419–430, 1995.

Schoeberl, M. R., et al., Reconstruction of the constituent distribution andtrends in the Antarctic polar vortex from ER-2 flight observations,J. Geophys. Res., 94, 16,815–16,845, 1989.

Schoeberl, M. R., C. H. Jackman, and J. E. Rosenfield, A Lagrangianestimate of aircraft effluent lifetime, J. Geophys. Res., 103, 10,817–10,825, 1998.

Schoeberl, M. R., et al., An assessment of the ozone loss during the 1999–2000 SOLVE/THESEO 2000 Arctic campaign, J. Geophys. Res.,10.1029/2001JD000412, this issue, 2002.

Swartz, W. H., J.-H. Yee, R. J. Vervack Jr., S. A. Lloyd, and P. A. Newman,Photochemical ozone loss in the Arctic as determined by MSX/UVISIstellar occulatation observations during the 1999–2000 winter, J. Geo-phys. Res., 10.1029/2001JD000412, in press, 2002.

Swinbank, R., and A. O’Neil, A stratosphere-troposphere data assimilationsystem, Mon. Weather Rev., 122, 686–702, 1994.

Turco, R., A. Plumb, and E. Condon, The Airborne Arctic StratosphericExpedition—Prologue, Geophys. Res. Lett., 17, 313–316, 1990.

�����������L. Lait, Code 916, NASA Goddard Space Flight Center, Greenbelt, MD

20071, USA. (lrlait@ertel.gsfc.nasa.gov)J. Burris, T. McGee, P. A. Newman, and M. R. Schoeberl, NASA

Goddard Space Flight Center, Greenbelt, MD, USA.E. V. Browell, NASA Goddard Space Flight Center, Greenbelt, MD,

USA.E. Richard, NOAA Aeronomy Laboratory and University of Colorado,

Boulder, CO, USA.B. R. Bojkov and G. O. Braathen, Norwegian Institute for Air Research,

Norway.F. Goutail, Centre National de la Recherche Scientifique, France.P. von der Gathen, Alfred Wegener Institute for Polar and Marine

Research, Potsdam, Germany.E. Kyro, Finnish Meteorological Institute, Sodankyla, Finland.G. Vaughan, University of Wales, Aberystwyth, UK.H. Helder, Royal Netherlands Meteorological Institute (KNMI), De Bilt,

Netherlands.S. Kirkwood, Swedish Institute of Space Physics, Sweden.P. Woods, National Physical Laboratory, UK.V. Dorkhov and I. Zaitcev, Central Aerological Observatory, Moscow,

Russia.B. Kois and Z. Litynska, Institute of Meteorology and Water Manage-

ment, Legionowo, Poland.A. Benesova and P. Skrivankova, Czech Hydrometeorological Institute

(CHMI), Prague, Czech Republic.H. De Backer, Royal Meteorological Institute of Belgium (KMI/IRM),

Belgium.J. Davies, Meteorological Service of Canada, Ontario, Canada.I. S. Mikkelsen and T. Jorgensen, Danish Meteorological Institute,

Copenhagen, Denmark.

SOL 16 - 10 LAIT ET AL.: OZONE LOSS DURING THE 1999–2000 SOLVE/THESEO 2000 CAMPAIGNS