aust. met. mag. 51 (2002) 95-106 low ozone concentrations ... · ozonesonde data in mcpeters et al....

Aust. Met. Mag. 51 (2002) 95-106

IntroductionThe year 1997 was notable for the record low ozonecolumn densities observed throughout the southernhemisphere mid-latitude region (Cordero andGrainger (2002), this issue, herein denoted asCG2002). While variability in the large-scale circula-tion is at least partially responsible for inter-annualanomalies (Connor et al. 1999), variations on thetime-scale of a week are also of interest, since ozoneaffects the amount of UV radiation received at theEarth’s surface (WMO 1999).

Macquarie Island (55°S, 159°E) is in a latitudeband where the atmosphere is highly variable, espe-

cially in the troposphere (Hurrell et al. 1998). In addi-tion, there may be passages of low ozone air from thepolar vortex or Antarctic ozone hole in the spring andearly summer (Atkinson 1997). This makes it an ideallocation for studies of short-term ozone changes. Inthe companion paper CG2002, the origins of airparcels from three low ozone events at MacquarieIsland were investigated using trajectory methods. Inthis paper, the large-scale ozone distribution is studiedusing the analysis of satellite ozone data.

The Ozone Analysis System (OASYS) interpo-lates satellite ozone data onto a three-dimensionalgrid to obtain the global ozone mixing ratio distribu-tion every six hours. OASYS consists of an analysisstep and a forecast step, which are described further inthe next section. Next a brief description of the satel-lite data used is given; OASYS is then validated using

95

Low ozone concentrations overMacquarie Island during 1997Part II: satellite ozone analysis

Simon Grainger and Eugene C. CorderoSchool of Mathematical Sciences, Monash University, Melbourne

(Manuscript received October 2000; revised June 2001)The year 1997 was notable for the record low ozone columndensities observed between 20°S and 60°S. The OzoneAnalysis System (OASYS) has been used to investigate lowozone events during 1997 at Macquarie Island (55°S, 159°E).OASYS uses satellite ozone data to generate three-dimen-sional ozone mixing ratio analyses. The analyses were com-pared with data from the Macquarie Island ozonesonde pro-gram. OASYS was able to reproduce the vertical ozone pro-file, although it was on average 10 to 15 per cent lower thanthe ozonesonde profile in the lower stratosphere. This iscaused by a systematic difference between satellite andozonesonde data, and is accentuated when the satellite datais not available. OASYS was used to examine three lowozone events at Macquarie Island in winter, spring and earlysummer. The winter event is dominated by vertical motion inthe lower stratosphere, while the spring and early summerevents are dominated by horizontal advection of low ozoneeither from the polar vortex, or the remnants of theAntarctic ozone hole.

Corresponding author address: Simon Grainger, School ofMathematical Sciences, Monash University, PO Box 28M, Clayton,Vic. 3800, Australia.

data from the Macquarie Island ozonesonde program(Lehmann 2002) and then studies of three low ozoneevents at Macquarie Island in 1997 are presented, fol-lowed by a summary.

The ozone analysis systemThere are two components of OASYS. The analysiscomponent combines all available satellite data with afirst-guess field on a global grid. The forecast step isthen used to generate the first-guess field for the nextanalysis.

The analysis component uses StatisticalInterpolation (SI), which fits a set of observations to agridded first-guess field such that the expected analy-sis error is minimised. The OASYS SI algorithm isbased on the moisture analysis scheme for the GlobalAssimilation and Predication system (GASP)described in Seaman et al. (1995). The methodologyused follows that of Lorenc (1981). Since ozone mix-ing ratio can be vertically integrated to obtain totalozone, both fields are analysed simultaneously. Thehorizontal resolution of OASYS is 2.5x2.5 degrees.Ozone mixing ratio is analysed at 19 pressure levelsbetween 1000 and 0.3 hPa. This includes eight levelsin the lower stratosphere between 100 and 10 hPa.Analyses are six-hourly, using data from three hourseither side of the analysis time. However, most resultsshown here use daily mean ozone mixing ratio, as thisis smoother than the six-hourly analyses.

The forecast step uses Reverse Domain Filling(RDF) with analysed winds to advect the analysedozone mixing ratio. RDF, described by Sutton et al.(1994), finds grid-point tracer values at the currenttime by finding the location of the grid-point air parcelat the previous time step. The wind fields are reversedto advect the air parcel upwind, and the tracer value atthe previous location is found by horizontal linearinterpolation between grid points. For OASYS, ozonemixing ratio is treated as the passive tracer, and theprevious time step is the current analysis.

Air parcels are advected using horizontal winds on

17 isentropic surfaces between 340K and 850K.Persistence, that is, the previous analysis, is usedbelow 340K and above 850K. Chemical and diabaticprocesses, and vertical motion, are assumed to benegligible over six hours in the lower stratosphere.The total ozone first-guess field is found by vertical-ly integrating the ozone mixing ratio.

Winds are obtained from GASP operational analy-ses. They are smoothed to T21 spectral resolution andinterpolated vertically onto the isentropic surfaces.340K is used as the lower surface because it is approx-imately at the tropopause. In the troposphere, diabaticheating becomes important, and isentropes intersect-ing with the surface of the earth need to be accountedfor. The 850K surface is generally around 10 hPa,which is near the top level of the GASP analyses.

Satellite ozone dataThe SI algorithm in OASYS uses all available satel-lite ozone mixing ratio data for input. Data is avail-able from the Solar Backscatter Ultraviolet (SBUV/2)instrument (Frederick et al. 1986), the StratosphericAerosol and Gas Experiment (SAGE II) (Maudlin etal. 1985), the Halogen Occultation Experiment(HALOE) (Russell et al. 1993) and the MicrowaveLimb Sounder (MLS) (Barath et al. 1993). The SIalgorithm also uses total ozone data from the TIROSOperational Vertical Sounder (TOVS) (Smith et al.1979). Table 1 provides a summary of the propertiesof the instruments used. Although total ozone is alsoavailable from the Total Ozone MappingSpectrometer (TOMS) (e.g. McPeters et al. 1996) it isnot used by OASYS. This is because TOMS data areonly publicly available as daily gridded data, which isnot suitable for the SI algorithm, which requires thelocation and time of individual observations.

Because they provide the most observations, theozone mixing ratio analysis is dominated by SBUV/2and total ozone by TOVS. The coverage from MLS isintermittent in the first half of 1997, and there are noMLS data after 15 June.

96 Australian Meteorological Magazine 51:2 June 2002

Table 1. Summary of satellite data used by OASYS. References for each instrument are given in the text. The numberof observations per day and resolution are approximate.

Instrument Type No. obs. per day Horizontal Vertical Accuracy Versionresolution resolution

TOVS IR Nadir 6000 250km N/A ~10% N/ASBUV/2 UV Nadir 2700 200km 5km 5-15% 6.0SAGE II Solar Occultation 30 24° lon 1km 6% 5.96HALOE Solar Occultation 30 24° lon 300m 4.5-16% 19MLS Limb Scanner 1320 400km 5.4km 0.3ppmv 4

Validation at Macquarie IslandIn order to check that the OASYS analyses are correctit is necessary to validate them against an independentset of observations. Surface-based measurements oftotal ozone and in situ ozonesonde data are used forthis, since the relatively sparse network means thatsuch observations are not used by OASYS.

Daily total ozone differences relative to theDobson instrument at Macquarie Island for 1997 areshown in Fig. 1 for OASYS (solid squares) and EarthProbe (EP) TOMS (open squares). Comparison withthe Dobson record (CG2002, Fig. 2) indicates that thedifferences are independent of the total ozoneamount, indicating that OASYS is able to reproducethe day to day variability throughout the year. WhileEP TOMS data are consistently around three per centhigher than the Dobson instrument throughout theyear, the difference between OASYS and the Dobsondata is more variable. In particular, OASYS totalozone is much lower than both EP TOMS and Dobsonduring late March and early November. This is causedby a lack of TOVS data during those times.

A comparison between the OASYS ozone mixingratio profiles and Macquarie Island ozonesondes isshown in Fig. 2. Also shown are differences fromHohenpeissenberg (48°N, 11°E) as an example of theperformance of OASYS at a similar northern hemi-sphere latitude. The ozonesonde profiles are comparedwith the closest OASYS analysis in time, and havebeen grouped into four seasons. In the lower stratos-phere, OASYS is around 10-15 per cent lower than theozonesonde data at Macquarie Island, and 5-10 percent lower at Hohenpeissenberg. These differences areconsistent with differences between SBUV/2 andozonesonde data in McPeters et al. (1994). Theozonesonde profiles are lower than OASYS above 10hPa. As the ozonesonde rises, the efficiency of thepump decreases, reducing the amount ozone that ismeasured. This is accounted for by a pressure-depen-dent correction to the retrieved profile (Kerr et al.1994). However, it is not clear that there is enoughozone being reported in the final profile. Below 100hPa, there are large differences as a result of the lackof satellite data below the tropopause.

Of importance are the differences at MacquarieIsland for October-December 1997 in Fig. 2(d), whereOASYS is over 40 per cent lower than theozonesonde at 30 and 40 hPa. This appears to bebecause the amount of satellite data is limited. Notonly are there no observations from MLS, but there isdata from only one SBUV/2 instrument instead of twoas earlier in 1997. While this does not affect theanalysis in the Northern Hemisphere, the large hori-zontal ozone gradients resulting from the Antarcticozone hole appear to result in large errors in the

OASYS analyses.In summary, differences between OASYS and

independent data shown here are generally consistentwith other studies directly between satellite data (e.g.McPeters et al. 1994; Cunnold et al. 1996;Neuendorffer 1996). However, OASYS does not per-form well when the amount of data is limited. Thelack of data appears to be particularly important whenanalysing regions of high ozone variability.

Low ozone events over MacquarieIslandThree low events at Macquarie Island from the sec-ond half of 1997 are examined in more detail. Theevents were chosen from examination of the totalozone record from the Macquarie Island Dobsoninstrument shown in Fig. 2 of CG2002. The eventsinclude one from winter, one springtime event andone from early summer. Ozone mixing ratio analysesfrom OASYS are used to show the large-scale evolu-tion of each event.

Winter eventThe largest change in total ozone observed atMacquarie Island in 1997 occurred in early August.Total ozone observed by the Dobson instrumentdecreased from 280 Dobson Units (DU) on August2nd to 232DU on 8 August followed by an increase to355DU on 11 August. Figure 3 shows the ozone mix-ing ratio at 40 hPa (approximately 22 km) for six days

Grainger and Cordero: Low ozone over Macquarie Island 1997 - satellite analysis 97

Fig. 1 Differences in daily mean total ozone atMacquarie Island for 1997 relative to theDobson instrument for OASYS (��) and EPTOMS (��), expressed as a percentage of theDobson data. Mean and standard deviationare given in the bottom left corner.

Macquarie Island, Australia (54.50 o S,158.97 o E)

Jan 01 Jan 31 Mar 02 Apr 01 May 01 May 31 Jun 30 Jul 30 Aug 29 Sep 28 Oct 28 Nov 27 Dec 27 -30

-20

-10

0

10

20

30

Diffe

renc

e (%

)

OASYS: 2.31 ±λ 8.76 EP TOMS: 2.96 ±λ 4.19


Fig. 2 Mean OASYS minus ozonesonde difference at Macquarie Island (��) and Hohenpeissenberg (∆∆) for (a) January-March, (b) April-June, (c) July-September and (d) October-December 1997. Differences are shown as a per-centage of ozonesonde data. Bars indicate one standard deviation.

January - March 1997

-40 -20 0 20 40 Difference (%)

1000

100

10

1

Pres

sure

(hPa

)

(a)

Macquarie Island Hohenpeissenberg

April - June 1997

-40 -20 0 20 40 Difference (%)

1000

100

10

1

Pres

sure

(hPa

)

(b)


July - September 1997

-40 -20 0 20 40 Difference (%)

1000

100

10

1

Pres

sure

(hPa

)

(c)


October - December 1997

-40 -20 0 20 40 Difference (%)

1000

100

10

1

Pres

sure

(hPa

)

(d)



Fig. 3 OASYS ozone mixing ratio (ppmv) for the southern hemisphere at 40 hPa for 2, 4, 6, 8, 10 and 12 August 1997.The contour interval is 0.5 ppmv. The ‘X’ symbol indicates the location of Macquarie Island.

2.0 2.0

2.5

2.5

3.0

3.0 3.0

3.0 3.0

3.5

3.5

4.0

02 August 1997

2.0

2.0

2.5

2.5

2.5

3.0

3.0

3.5

3.5

4.0

04 August 1997

2.0

2.5

2.5

3.0

3.0

3.0

3.5

3.5

06 August 1997

2.0

2.5

2.5

3.0

3.0

3.0

3.5

3.5

3.5

3.5

08 August 1997

2.0

2.5

2.5

3.0

3.0

3.5

3.5

3.5

10 August 1997

2.0

2.5

2.5

2.5

3.0

3.0

3.0

3.5

12 August 1997

Southern Hemisphere Ozone Mixing Ratio (ppmv) at 40hPa

OASYS Analysis First time is 02 August 1997 and interval is 2 days.


Fig. 4 GASP isentropic potential vorticity, IPV (in potential vorticity units (PVU)), for the southern hemisphere on 6,8 and 10 August 1997 at (a) 550K, with the -80 PVU contour highlighted, and (b) 450K, with the -20 PVU con-tour highlighted. The ‘X’ symbol indicates the location of Macquarie Island.

-120 -1

00

-80 -60 -40

-20

-20

0

0

0

0

0

550K

06 August 1997

(a)

-50 -40

-30 -20 -15

-15

-10

-10

-10

-5

-5

0

0

0

0

0

0

5

5 450K

06 August 1997

(b)

-120

-100

-80

-60

-40

-40

-20

-20

0

0 0

0

0

08 August 1997

-50

-40

-30

-20

-15

-10

-10

-5

-5

0

0

0

0

0

5

08 August 1997

-140

-140

-120

-100

-80

-60

-40

-40

-20

-20

0

0

0

0

0

10 August 1997

-50

-40

-30

-20

-15

-15

-10

-10

-5

-5

0

0

0

0

0

0

0

5

5

10 August 1997

Southern Hemisphere GASP Potential Vorticity (PVU)

First time is 06 August 1997 and interval is 2 days. θλ = (a) 550K (b) 450K.


in early August 1997. On 2 August the polar vortex,indicated by the 3 ppmv contour, is nearly circular,and Macquarie Island is underneath mid-latitude air.Over the next four days, an outbreak of low ozone airfrom the polar vortex covers Macquarie Island. By 8August, when the total ozone column was lowest, thelow-ozone air also covers New Zealand. Afterwardsthe vortex moves off Macquarie Island, and isreplaced by high-ozone air associated with a stratos-pheric anticyclone.

The synoptic situation can also be examined usingIsentropic Potential Vorticity (IPV, measured in poten-tial vorticity units (PVU)), which is anti-correlatedwith ozone mixing ratio in the lower stratosphere. IPVfor the 550K potential temperature surface from GASPfor 6, 8 and 10 August is shown in Fig. 4(a). The large-scale distribution of IPV generally agrees with the dis-tribution seen in Fig. 3. There is almost no change inthe IPV over Macquarie Island between 6 August and8 August, suggesting there is little difference in ozonemixing ratio at 40 hPa between those dates. However,on the 450K surface, which is approximately 70 hPa(18 km), there is a bulge of high-IPV over MacquarieIsland on 8 August. This can be clearly seen in the -15PVU contour over New Zealand. On August 10th thetransition from high-IPV to low IPV-air can be seen atboth 550K and 450K.

The behaviour of the ozone distribution at differ-ent heights can be clearly seen in the time series inFig. 5, which shows the ozone partial pressure at 100,70, 40 and 20 hPa. At Macquarie Island (Fig. 5(a)),there is a steady decrease in ozone at 100 and 70 hPaover 2 to 8 August, with a sharper decline from 6August. The minimum at 40 hPa is on 7 August, andthere is little difference between 6 August and 8August. There is little change in ozone at 20 hPa over2 to 8 August. All levels show a sharp increase over 8to 11 August. The decrease in total ozone is greatestin the lower stratosphere, with the 60-125 hPa slab,represented by the 70 and 100 hPa levels, contribut-ing 22DU to the overall loss. In contrast, the 25-45hPa slab, represented by the 30 and 40 hPa levels,contributes a maximum loss of 7DU over 2 to 7August.

The fact that the ozone loss is largely confined tothe lowermost stratosphere suggests that it is causedby changes in the vertical structure. This is also indi-cated by the IPV anomaly over Macquarie Island on 8August at 450K (Fig. 4(b)) and by an increase intropopause height (not shown). However, in therecovery phase, horizontal transport plays a muchlarger role. The 60-125 hPa slab contributes 36DU tothe increase over 8 to 11 August, and 25-45 hPa slabcontributes 22DU over 7 to 11 August.

This event was also studied at Lauder, NewZealand (45°S, 170°E) by Brinksma et al. (1998).They found that a combination of the polar vortex atupper levels and mixing of subtropical air in the lowerstratosphere caused the record low ozone observedthere. Figure 5(b) shows the ozone partial pressure atLauder from OASYS. OASYS shows a substantialloss of ozone throughout the lower stratosphere, but itis not possible in this study to indicate the source ofthese losses.

Spring eventThe Antarctic ozone hole develops each year fromearly September, reaching a peak around the end ofthe month. As the polar vortex weakens from Octoberonwards, the outer edge of the Antarctic ozone holemay be drawn off into the middle latitudes.

An example of this occurred in October 1997,

Macquarie Island (54.50oS,158.95oE)

Aug 011997

Aug 161997

0

50

100

150

Parti

al Pr

essu

re (n

bar)

100hPa70hPa40hPa20hPa

(a)

Lauder, NZ (45.04oS,169.68oE)

Aug 011997

Aug 161997

0

50

100

150

Parti

al Pr

essu

re (n

bar)


(b)

Fig. 5 OASYS ozone partial pressure (nbar) for 1 to16 August 1997 at 100hPa (��), 70 hPa (��), 40hPa (∆∆) and 20 hPa (*) at (a) MacquarieIsland, and (b) Lauder, New Zealand.


2.0 2.5

3.0

3.5 4.0

4.0

5.0

5.0 5.0

5.0

5.0

19 October 1997

2.0 2.5 3.0

3.5

4.0

4.0 4.0

5.0 5.0

5.0 5.0

20 October 1997

2.0

2.5

3.0

3.5

4.0

4.0 4.0

5.0

5.0

5.0

5.0

5.0

5.0

21 October 1997

2.0 2.5 3.0

3.5 4.0

4.0 4.0

5.0

5.0

5.0

5.0

5.0

22 October 1997

2.0 2.5

3.0

3.5 4.0

4.0 4.0

4.0

5.0

5.0

5.0 5.0

5.0

23 October 1997

2.0 2.5 3.0

3.5

4.0

4.0

5.0

5.0

5.0

5.0

24 October 1997

Southern Hemisphere Ozone Mixing Ratio (ppmv) at 30hPa

OASYS Analysis First time is 19 October 1997 and interval is 1 day.

Fig. 6 OASYS ozone mixing ratio (ppmv) at 30 hPa for 19 to 24 October 1997. The ‘X’ symbol indicates the locationof Macquarie Island.

where the total ozone at Macquarie Island measuredby the Dobson instrument declined from 399DU on19 October to 281DU on 23 October, followed by asteady recovery. Figure 6 shows the OASYS ozonemixing ratio at 30 hPa for 19 to 24 October 1997. On19 October, Macquarie Island is outside the ozonehole, indicated by the 3ppmv contour. From 21October, the ozone hole is elongated and two tonguesof low ozone air, south of Australia and over theSouth Atlantic ocean, are stripped off the edge of thepolar vortex and into the mid-latitude regions.Macquarie Island is well inside the ozone hole on 22and 23 October, before the ozone hole moves east-ward on 24 October.

The ozone partial pressure at 100, 70, 40 and 20hPa at Macquarie Island for 19 to 24 October 1997 isshown in Fig. 7. All levels show a decline between 19and 22 October, followed by a smaller recovery. Theoverall change at different levels is similar – the 60-125 hPa slab contributes 22DU over 19 to 22 Octoberand the 25-45 hPa contributes 33DU.

That the change in ozone is relatively constantthroughout the stratosphere suggests that this event iscaused by horizontal transport of low-ozone air. Thesource of the low-ozone air is most likely to be fromthe edge of the polar vortex. The process is similar tothe spring event studied in CG2002.

Early summer eventThe final break up of the Antarctic ozone hole typi-cally occurs in late November and early December.During this period, remnants of the ozone hole maybe transported into the middle latitudes (Atkinson1997). The low ozone event at Macquarie Island in

early December 1997 is an example of this. Thisevent was studied in CG2002. There the colocation oflow EP TOMS total ozone and high values of modi-fied potential vorticity indicated that the cause of thisevent was the breakup of the polar vortex. Here theperformance of OASYS is examined.

Figure 8 shows the total ozone for 3 December1997. Both OASYS (Fig. 8(a)) and EP TOMS (Fig.8(b)) show the remnant of the ozone hole to the south-


Macquarie Island (54.50oS,158.95oE)

Oct 191997

Oct 241997

0

50

100

150

Parti

al Pr

essu

re (n

bar)


Fig. 7 As in Fig. 5, except at Macquarie Island for 19to 24 October 1997.

225

225

250

250

250

250

250

275

275

275

300

300

300

300

325

(a)

250

275 300

300

350

(b)

Southern Hemisphere Total Ozone (DU)

Analysis time is 03 December 1997. (a) OASYS Total Ozone Analysis. (b) Earth Probe TOMS.

Fig. 8 Total ozone (DU) for 3 December 1997 from (a)OASYS, and (b) EP TOMS. The contour inter-val is 25 DU, with low values shaded darkly.White regions indicate no data. The ‘X’ sym-bol indicates the location of Macquarie Island.

east of Macquarie Island, with the OASYS analysisindicating slightly lower values than EP TOMS.However, OASYS underestimates the size and depthof the other remnant, centred on 0° longitude, andincludes a region of high total ozone.

The ozone mixing ratio for 2300 UTC 3 December

1997 at 70, 50, 30 and 20 hPa is shown in Fig. 9. Alsoplotted are the locations of the satellite ozone mixingratio observations for that time. The ozone hole rem-nants can be seen in the same locations in Fig. 8, butthe dominant feature is the very low ozone between45°S and 60°S south of Australia at 50 and 30 hPa


Fig. 9 OASYS ozone mixing ratio (ppmv) for 2300 UTC 3 December 1997 at (a) 70 hPa, (b) 50 hPa, (c) 30 hPa and (d)20 hPa. Also shown are the locations of observations used by OASYS from SBUV/2 (+), HALOE (∆∆) and SAGEII (��). The ‘X’ symbol indicates the location of Macquarie Island.

0.5

0.5

0.5

0.5 1.0

1.0

1.5

1.5

2.0

70hPa

(a)

1.0

1.0 1.0

1.0 1.5

1.5

2.0

2.0

2.0

2.5 3.0

50hPa

(b)

3.0

3.0 3.5

3.5 3.5

3.5

3.5

3.5

3.5

3.5

3.5

4.0

4.0

4.0

4.0

4.0 4.0

30hPa

(c)

3

4

4

4

5

5

5

5

6

6

6

6

6

6

6

6

7

7

7

7

7

7

7

20hPa

(d)

Southern Hemisphere Ozone Mixing Ratio (ppmv)

OASYS OMR Analysis Analysis time is 23Z 03 Dec 1997.

Pressure = (a) 70hPa (b) 50hPa (c) 30hPa (d) 20hPa.

(Figs 9(b) and 9(c)). This feature is less prominent at70 and 20hPa (Figs 9(a) and 9(d)).

One of the causes of this appears to be systematicdifferences between the satellite instruments. Byexamining the contours near the observation loca-tions, SBUV/2 (pluses) act to increase the ozone mix-ing ratio at 70 and 50 hPa, and decrease it at 30 and20 hPa. SAGE II data (squares) has the oppositeimpact. HALOE observations (triangles) generallyincrease the ozone mixing ratio. An example can beseen at 30 hPa (Fig. 9(c)) between 0° and 60°E at60°S, where it corrects the very low ozone, at least ina small region. Systematic differences alone cannotexplain the problem, however, since it does not occurin the northern hemisphere (not shown), or earlier inthe year (e.g. Fig. 2). Two factors that may accentuatethe problem are the low ozone values and large hori-zontal gradients associated with the Antarctic ozonehole, and the limited amount of satellite data availablecompared with earlier in 1997.

To check that the feature in Fig. 9 is caused by

OASYS rather than the satellite data, the MacquarieIsland ozonesonde profile from 2300 UTC 3December 1997 is compared with the OASYS profileand those from the closest HALOE and SBUV/2observations in Fig. 10. Given the limited resolution ofHALOE, it agrees well with the ozonesonde profile.HALOE may be lower than the ozonesonde around 40hPa because it is southeast of Macquarie Island, nearthe ozone hole remnant in Fig. 8. The SBUV/2 profiledoes not agree as well with the ozonesonde, with moreozone below the partial pressure maximum at 50 hPa,and less between 40 and 20 hPa. OASYS bears littleresemblance to either the ozonesonde or the satellitedata in the lower stratosphere, with a large deficitbetween 50 and 15 hPa. It is interesting to note thelarge ozone partial pressure in OASYS at 300 hPa.When integrating the column to find total ozone, thisanomaly almost cancels out the ozone deficit in thelower stratosphere. The cause of this is not known,however it is clear that further investigation is requiredin circumstances where OASYS has limited satellitedata available for analysis.

SummaryTo examine the record low total ozone observedbetween 20°S and 60°S in 1997, several low ozoneevents at Macquarie Island have been studied.Macquarie Island is an important location for moni-toring such events, as they can be caused by transportbetween polar and middle latitude regions, or bychanges in the vertical structure of the stratospherecaused by synoptic-scale tropospheric systems.

The study has been done in two parts. In the firstpart, CG2002, trajectory methods were used to deter-mine the origin of air parcels over Macquarie Island.In this paper, analyses of satellite data by OASYSwere used to examine the large-scale ozone distribu-tion. OASYS is suitable for this purpose, as compar-isons showed agreement with Macquarie Islandozonesonde data, within known instrument differ-ences. However, this is provided that there was suffi-cient data to analyse, otherwise OASYS tended todrift to unrealistically low values.

Three low ozone events from the second half of1997 were studied. The first case looked at a winterevent in early August. The large ozone decrease main-ly in the lowermost stratosphere (60-125 hPa) sug-gested a change in the vertical structure. However, theozone recovery in this case was more strongly influ-enced by horizontal advection. The second caseexamined a spring event from late October. This wascaused by large-scale horizontal advection of lowozone air from the edge of the Antarctic ozone hole.


Macquarie Is. (54.50oS,158.95oE)

0 20 40 60 80 100 120 140 160Ozone Partial Pressure (nbar)

1000.0

100.0

10.0

1.0

0.1

Pres

sure

(hPa

)

OzonesondeOASYSSBUV/2 04/0437Z 52.8oS, 159.7oEHALOE 03/1533Z 60.6oS, 167.8oE

Time: 23Z 03 Dec 1997

Fig. 10 Ozone partial pressure (nbar) profiles atMacquarie Island for 2300 UTC 3 December1997 from OASYS (��) and the ozonesonde(thick line). Also shown are profiles from thenearest SBUV/2 (+) and HALOE (∆∆) observa-tions.


The third case was an early summer event in the firstweek of December. The total ozone analysis was inagreement with EP TOMS, which suggests thatOASYS was analysing the ozone hole remnants indi-cated by CG2002. However, the limited data resultedin a large belt of unrealistically low ozone south ofAustralia between 50 and 30 hPa.

The results presented here are in broad agreementwith the trajectory methods used in the companionpaper. Together they enable the construction of a com-plementary picture of the short-term evolution of theglobal ozone distribution.

AcknowledgmentsThe authors would like to thank J. Easson and theBureau of Meteorology’s Global Atmosphere Watchsection for providing the Macquarie Islandozonesonde data. Satellite ozone data was made avail-able by NASA (TOMS, MLS, HALOE and SAGE II)and NOAA (TOVS and SBUV/2). GASP data wasprovided by the Australian Bureau of Meteorology.The RDF code was originally developed by D. Waughand R. Atkinson. Comments from an anonymousreviewer led to the strengthening of this paper.

ReferencesAtkinson, R.J. 1997. Ozone variability over the southern hemisphere.

Aust. Met. Mag., 46, 195-201.Barath, F.T., Chavez, M.C., Cofield, R.E., Flower, D.A., Frerking,

M.A., Gram, M.B., Harris, W.M., Holden, J.R., Jarnot, R.F.,Kloezeman, W.G., Klose, G.J., Lau, G.K., Loo, M.S., Maddison,B.J., Mattauch, R.J., McKinney, R.P., Peckham, G.E., Pickett,H.M., Siebes, G., Soltis, F.S., Suttie, R.A., Tarsala, J.A., Waters,J.W. and Wilson, W.J. 1993. The Upper Atmosphere ResearchSatellite Microwave Limb Sounder instrument. J. geophys. Res.,98, 10751-62.

Brinksma, E.J., Meijer, Y.J., Connor, B.J., Manney, G.L., Bergwerff,J.B., Bodeker, G.E., Boyd, I.S., Liley, J.B., Hogervorst, W.,Hovenier, J.W., Livesey, N.J. and Swart, D.P.J. 1998. Analysis ofrecord-low ozone values during the 1997 winter over Lauder,New Zealand. Geophys. Res. Lett., 25, 2785-8.

Connor, B.J., Bodeker, G.E., McKenzie, R.L. and Boyd, I.S. 1999.The total ozone anomaly at Lauder, NZ in 1997. Geophys. Res.Lett., 26, 189-92.

Cordero, E.C. and Grainger, S. 2002. Low ozone concentrations overMacquarie Island during 1997. Part I: trajectory analysis. Aust.Met. Mag., 52, 85-94.

Cunnold, D.M., Froidevaux, L., Russell, J.M., Connor, B. and Roche,A. 1996. Overview of UARS ozone validation based primarily on

intercomparisons among UARS and Stratospheric Aerosol andGas Experiment II measurements. J. geophys. Res., 101, 10335-50.

Frederick, J.E., Cebula, R.P. and Heath, D.F. 1986. Instrument char-acterization for the detection of long term changes in stratos-pheric ozone: An analysis of the SBUV/2 radiometer. J. Atmos.Oceanic Technol., 3, 472-80.

Hurrell, J.W., van Loon, H. and Shea, D.J. 1998. The mean state ofthe troposphere, in Meteorology of the Southern Hemisphere, edsD.J. Karoly and D.G. Vincent, American Meteorological Society,1-46.

Kerr, J.B., Fast, H., McElroy, C.T., Oltmans, S.J., Lathrop, J.A.,Kyro, E., Pakkunen, A., Claude, H., Kohler, U., Sreedharan,C.R., Takao, T. and Tsukagoshi, Y. 1994. The 1991 WMO inter-national ozonesonde intercomparison at Vanscoy, Canada.Atmos.-Ocean, 32, 685-716.

Lehmann, P. 2002. A climatological study of the Macquarie Islandozonesonde record 1994-2000. Aust. Met. Mag.,52, 79-83.

Lorenc, A.C. 1981. A global three-dimensional multivariate statisti-cal interpolation scheme. Mon. Weath. Rev., 109, 701-21.

Maudlin, L.E. III, Zuan, N.H., McCormick, M.P., Guy, J.H. andVaughn, W.R. 1985. Stratospheric Aerosol and Gas Experiment IIinstrument: A functional description. Opt. Eng., 24, 307-12.

McPeters, R.D., Miles, T., Flynn, L.E., Wellemeyer, C.G. andZawodny, J.M. 1994. Comparison of SBUV and SAGE II ozoneprofiles: Implications for ozone trends. J. geophys. Res., 99,20513-24.

McPeters, R.D., Bhartia, P.K., Krueger, A.J., Herman, J.R.,Schlesinger, B.M., Wellemeyer, C.G., Seftor, C.J., Jaross, G.,Taylor, S.L., Swissler, T., Torres, O., Labow, G., Byerly, W. andCebula, R.P. 1996. Nimbus-7 Total Ozone MappingSpectrometer (TOMS) data products user’s guide. NASAReference Publication 1384, 67pp.

Neuendorffer, A.C. 1996. Ozone monitoring with TIROS-N opera-tional vertical sounders. J. geophys. Res., 101, 18807-28.

Russell, J.M. III, Gordley, L.L., Park, J.H., Drayson, S.R., Hesketh,W.D., Cicerone, R.J., Tuck, A.F., Frederick, J.E., Harries, J.E.and Crutzen, P.J. 1993. The Halogen Occultation Experiment. J.geophys. Res., 98, 10777-97.

Seaman, R., Bourke, W., Steinle, P., Hart, T., Embery, G., Naughton,M. and Rikus, L. 1995. Evolution of the Bureau of Meteorology’sGlobal Assimilation and Prediction system. Part I: analysis andinitialisation. Aust. Met. Mag., 44, 1-18.

Smith, W.L., Woolf, H.M., Hayden, C.M., Wark, D.Q. and McMillin,L.M. 1979. The TIROS-N Operational Vertical Sounder. Bull.Am. Met. Soc., 60, 1177-87.

Sutton, R.T., Maclean, H., Swinbank, R., O’Neill, A. and Taylor, F.W.1994. High-resolution stratospheric tracer fields estimated fromsatellite observations using Lagrangian trajectory calculations. J.Atmos. Sci., 51, 2995-3005.

World Meteorological Organisation (WMO). 1999. Scientific assess-ment of ozone depletion: 1998, WMO Global Ozone Researchand Monitoring project. Report No. 44, WMO, Geneva.

aust. met. mag. 51 (2002) 95-106 low ozone concentrations ... · ozonesonde data in mcpeters et al....

Documents