JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 106, NO. D23, PAGES 32,503-32,525,DECEMBER 16, 2001
Ozone in the Pacific tropical troposphere from ozonesonde
observations
Samuel
J.Oltmans,
1BryanJ.Johnson,
1Joyce
M. Harris,
• HolgerV6mel,
•'2AnneM.
Thompson,
3Kanatathu
Koshy,
4Patrick
Simon,
5Richard
J.Bendura,
6Jennifer
A. Logan,
?
FumioHasebe,
8Masato
Shiotani,
9VolkerW.J.H.Kirchhoff,
•øMatakite
Maata,
4Gopal
Sami,
4AnsariSamad,
4JiojiTabuadravu,
4Humberto
Enriquez,
TM
MarioAgama,
•2Jaime
Cornejo,
•: andFrancisco
Paredes
•
Abstract. Ozoneverticalprofile measurements
obtainedfrom ozonesondes
flown at Fiji,
Samoa,Tahiti, andthe Galapagosare usedto characterizeozonein the troposphereover the
tropicalPacific. Thereis a significantseasonalvariationat eachof thesesites. At sitesin both
the easternand westernPacific, ozonemixing ratiosare greatestat almostall levels in the
troposphere
duringthe September-November
seasonand smallestduringMarch-May. The
verticalprofile hasa relative maximumat all of the sitesin the midtroposphere
throughoutthe
year (the largestamountsare usuallyfound near the tropopause).This maximum is particularly
pronouncedduringthe September-November
season.On average,throughoutthe troposphere,
the Galapagoshaslargerozoneamountsthanthe westernPacific sites. A trajectoryclimatology
is usedto identify the major flow regimesthat are associatedwith the characteristicozone
behaviorat variousaltitudesand seasons.The enhancedozoneseenin the midtroposphere
duringSeptember-November
is associatedwith flow from the continents.In the westernPacific
this flow is usuallyfrom southernAfrica (although10-daytrajectoriesdo not alwaysreachthe
continent)but also may come from Australiaand Indonesia. In the Galapagosthe ozonepeak in
the midtroposphere
is seenin flow from the SouthAmericancontinentandparticularlyfrom
northernBrazil. High ozoneconcentrations
within potentialsourceregionsand flow characteristicsassociatedwith the ozone mixing ratio peaksseenin both the westernand easternPacific
suggestthat theseenhancedozone mixing ratiosresult from biomassburning. In the upper
troposphere,low ozone amountsare seenwith flow that originatesin the convectivewestern
Pacific.
1.
Introduction
The tropicalPacific is often consideredto be a regionremote
from major polluting influences becauseof its isolation from
heavilyindustrializedlandmasses.Recentfield campaigns[Hoell
et al., 1999] have emphasizedthat though this is often the case,
•Climate
Monitoring
andDiagnostics
Laboratory,
NOAA,Boulder,
Colorado.
2Cooperative
Institute
forResearch
intheEnvironmental
Sciences,
the signatureof pollution, particularly from biomassburning,
makes a significantimprint on the air chemistryof the region
[Singhet al., 2000]. In a numberof instances,
layersof enhanced
ozone (mixing ratios >80 ppbv) were found in the midtroposphere
in the remotewesternPacific [Stolleret al., 1999].
Theselayersin additionto havingenhancedozonewere replete
with markers of biomass burning such as carbon monoxide,
ethane,ethyne, and notably methyl chloride [Gregory et al.,
1999].
Beginning
in August1995as partof the PacificExploratory
Mission(PEM) TropicsA, ozoneverticalprofilemeasurements
3NASA
Goddard
Space
FlightCenter,
Greenbelt,
Maryland.
4School
ofPureandApplied
Sciences,
University
oftheSouth
Pacific, werebegunat PagoPago,AmericanSamoa(14.3øS,170.6øW),
Suva, Fiji.
andPapeete,
Tahiti(18.0øS,149.0øW).At Samoa,profileswere
5MeteoFrance,
Papeete,
Tahiti,French
Polynesia.
also
obtained
as part of an earliermeasurement
programfrom
6NASA
Langley
Research
Center,
Hampton,
Virginia.
1986to 1989. Theseearlierprofilesprovidean opportunity
for
7Department
ofEarthandPlanetary
Sciences,
Harvard
University,
Cambridge,Massachusetts.
comparison
with measurements
duringthe morerecentperiod.
8Department
ofEnvironmental
Sciences,
Ibaraki
University,
Mito,
Profile measurements were continued at Tahiti and Samoa
Japan.
through
PEM TropicsB withtheprogramat Tahiticompleted
in
9Graduate
School
of Environmental
EarthScience,
Hokkaido
December
1999. At Samoa,weeklysoundings
continue
aspart
University,Sapporo,Japan.
løInstituto
dePequisas
Espaciasis,
SaoJose
dosCampos,
Brazil.
of theSouthern
Hemisphere
AdditionalOzonesondes
(SHADOZ)
•llnstituto
Nacional
deMeteorologia
y Hidroligia,
Quito,Ecuador.
project (A.M. Thompson et al., The 1998-2000 SHADOZ
•2Instituto
Nacional
deMeteorologia
y Hidroligia,
SanCfistobal,
(Southern
Hemisphere
AdditionalOzonesondes)
tropicalozone
Galapagos,Ecuador.
Universityof Colorado,Boulder,Colorado.
Copyright
2001bytheAmerican
Geophysical
Union.
Papernumber2000JD900834.
0148-0227/01/2000JD900834509.00
climatology: Seasonal patterns, the zonal wave-one, and
comparisonwith TOMS and ground-basedmeasurements,
submittedto Journalof GeophysicalResearch,2001, hereinafter
referredto as A.M. Thompson
et al., submitted
manuscript,
2001). Duringmostof themeasurement
period,soundings
were
32,503
32,504
OLTMANS ET AL.' OZONE IN THE PACIFIC TROPICAL TROPOSPHERE
i
!
[
i
t
I
i
i
.d.
dl,
-,
,
,
J
i
i
i
i
i
r
!
4
Sep-Oct-Nov
o
0
100
0
25
50
O0
75
OZONEM•XINGRATIO(PPBV)
Figure1. Theaverage
ozone
profileandvariability
for0.25kmlayers
atTahitifortwoseasons
(a)March-May
and
(b) September-November
fortheperiod1995-1999.
Themedian
is thevertical
lineinside
thebox,andthebox
represents
theinner50thpercentile
ofthedata.The"whiskers"
represent
theinner90thpercentile
ofthedata.The
solid diamond is the mean.
of PEM TropicsB. As part of the Soundings
of
doneweekly. Duringthe aircraftfield campaigns
in September- anticipation
October1996 (PEM TropicsA) and March-April1999 (PEM Ozone and Water Vapor in the EquatorialRegion (SOWER)
were startedon a campaign
Tropics
B) soundings
weredonetwicea week.In January
1997, projectozoneprofile measurements
weeklysoundings
werebegunat Suva,Fiji (18.1øS,178.2øE)
in basisin March 1998 at SanCristobal,Galapagos(0.9øS,89.6øW),
i
I
I
I
•
i
,
,
t
i
i
i
•
, .a.
,
i
..•
,
I
o_
I
•
,,
•
,
.
I
IVlar-Apr-May
o
O0
0
1O0
OZONEMIXING
•TIO(PP•)
Figure2. Theaverage
ozone
profileandvariability
for0.25-km
layers
atGalapagos
fortwoseasons
(a)MarchMay and(b) September-November
for theperiod1998-2000.
OLTMANS
ET AL.' OZONE IN THE PACIFIC TROPICAL TROPOSPHERE
32,505
96
15
90
84
78
72
66
6O
54
48
42
36
30
24
18
12
6
0 ppb
J
F
M
A
M
J
J
A
S
0
N
D
Month
Plate 1. Time-height crosssectionof ozone mixing ratio at Pago Pago, American Samoa (14.3øE, 170.6øW), for
1996. Soundingsare weekly except during the Pacific Exploratory Mission (PEM) Tropics A intensive
measurement
periodfrom August-September
when soundingswere donetwice weekly.
96
15
9O
84
78
72
66
60
54
48
42
36
3O
24
18
12
o ppb
J
F
M
A
M
J
J
A
S
0
N
D
Month
Plate 2. Time-heightcrosssectionof ozonemixingratio from weeklysoundings
at San Cristobal,Galapagos
(0.9øS, 89.6øW) for 1999.
32,506
OLTMANS ET AL.: OZONE IN THE PACIFIC TROPICAL TROPOSPHERE
i
.
--
I
i
lOO
Figure3. The averageozoneprofileandvariabilityfor 0.25-kmlayersat Samoafor two seasons
(a) March-May
and(b) September-November
for the period1995-2000.
were increasedto biweekly soundingsin September1998 and to throughoutthe troposphereof the tropical Pacific that has not
weeklysoundings
aspartof SHADOZ earlyin 1999. Soundings been available in the past. In particular, new insights on the
continueat Fiji and the Galapagosin 2000 as part of SHADOZ short-termand seasonalvariability of ozonein this regionas well
and SOWER.
as differences
between
the
eastern
and
western
Pacific
are
This set of ozone profiles obtained using balloon-borne revealedin thesedata. In addition, isentropictrajectoriesare used
ozonesondesprovides information on the distributionof ozone to look at the transport associatedwith particular features of
i
II
I
id,
I
i
I
I 4['
I •
t.____•.
i
dl,_dl , i
.I
I
P
i
L
[
i
--•
!
i•lI'
t
I
I
i
iI
, i
i---i..
,
i
I
•-"...,
•'
•
•.
o_
•
i
r'
i
I
i ,
i
,
i
1•
ß
•
•
i
t
i
•'
i
&
,•
i
,
i
i
,
,•
,
i
•
i
,
,•
•'"
I
i
i
•,
•
•
i
I
i
I
ß
•
,,
i
_
,
i
•
i
iI
ar-apr-ay
:
Sep-Oct-Nov
o
0
100
0
'
oZoN[ U•X•NG•0(•)
Figure 4. The averageozoneprofile and variabilityfor 0.25-km layersat Fiji for two seasons(a) March-May and
(b) September-November
for the period 1997-2000.
OLTMANS ET AL.' OZONE IN THE PACIFIC TROPICAL TROPOSPHERE
32,507
individual profiles and also at the influence of climatological ozonesondepump efficiency measurementsand sensitivitytests
transport patterns on the principal features of the ozone of buffered and unbuffered sensor solutions, submitted to
Journal of GeophysicalResearch,2001) at all of the sitesexcept
distributionin the tropicaltroposphere
in this region.
the Galapagos where the new recipe was used /¾om the
beginning. From comparisonflights made at these sites, as well
2. Methods
as laboratorytests,an empirical correctionhas been derived, and
2.1. Ozonesondes
the earlier data have been correctedusing this relationship. This
The ozone vertical profiles were obtained using the correction is largest in the stratosphere.It leaves values for the
electrochemical
concentrationcell (ECC) ozonesonde[Komhyret earlier data at the surface unchangedand reduces amountsby
al., 1995]. This has becomea standardtechniquefor obtaining -2% at 100 hPa. Data placed in the Global Troposphere
ozone profiles with high vertical resolution in both the Experiment (GTE) archive /'or PEM Tropics at NASA Langley
troposphere and stratosphereto altitudes of-35
km. The andin the SHADOZ archivehave this correctionapplied.
measurementsof ozone have an accuracyof _+5%through most
of the tropospherewith somewhatmore uncertainty(_+10%)/br
very low mixing ratios (<10 ppbv) encounteredoccasionallyin
the tropics. The only important interrelent in the measurement
technique,which is basedon the oxidationreactionof ozonewith
potassiumiodide in solution, is sulfur dioxide that is not
2.2. Trajectories
For the purposes of characterizing the tropospheric airflow
patterns influencing transport to the tropical sites, insentropic
trajectorieshave been calculated. The trajectoriesare computed
from the EuropeanCentre for Medium-RangeWeather Forecasts
encountered at these sites at sufficient concentrations
to be of
(ECMWF) analysesusing the model describedby Harris and
significance. One exceptionis the presenceof a layer of zero Kahl [1994]. The limitations in such trajectories must be
ozonemeasuredat the Galapagosin a singleprofile shortlyafter recognizedin interpretingthe resultsthat are obtained,but they
a volcaniceruption in Ecuador. The data were obtainedwith an do provide a useful tool for obtaining a picture of the flow
altitude resolutionof-50 m but for the analysisperformedhere patternsthat may influence ozone behavior at these sites. In
were averaged into 250-m layers. Only at Samoa are total particular for this work the role of convection needs to be
column
ozone
measurements
from
a collocated
Dobson
consideredas a possible influence on the air parcel path. For
spectrophotometer
available for comparisonwith the integrated individual trajectory paths the outgoing longwave radiation
total ozone from the ozonesonde. These comparisonsgive an (OLR) maps available from the NOAA-CIRES Climate
average ratio of i.04 +_0.05 between the Dobson total ozone Diagnostics Center,
Boulder,
Colorado, Web
site
measurementand the integratedtotal ozone from the ozonesonde, (http://www.cdc.noaa.gov/HistData/), that are constructedfrom
giving confidencethat the ozonesondemeasuredozone amounts NCEP operationalOLR data, were inspected/'oreach of the 10
from all of the sitescan be comparedwith each other. During the days to see if low OLR values indicative of convection were
courseof the measurementprogram a change was made in early
found at the air parcel location. The computedtrajectoriesare
1998 in the sensing solution recipe (B.J. Johnsonet al., ECC
used both to investigateindividual cases and by grouping the
•
.
a,
FI -- .......
•r
"•.. '-••..• I [
--,
_77•%• .......
i
_
•••d•=-•
•.e;.,
.....
- 7 -'
,
•_ - _
•' ß 1995-2000
(•as•)
Mar-Apr-May
o
o
1oo
o
1oo
OZONE•IXlNO •TIO(PP•)
Figure5. Comparison
oftheozonemixingratioprofiles
atSamoa
fortheperiods
1986-1989
and1995-2000
for(a)
March-May and (b) September-November.
32,508
OLTMANS
"-
i
ET AL.: OZONE IN THE PACIFIC
'-
TROPOSPHERE
J
_- - 7- *---- •-•
•.-•
TROPICAL
,
=1= _-• - - -,
,&.-Tc•=•--;
•,'•
.
,
....
,
,
i
I
i
- •
.... ,
4• Samoa(solid)
....
ß Galapagos(dashed)
-'
0
Mar-Apr-May
s'o
OZONEM•X•N0RAT•O(PPm')
Sep-Oct-Nov
0
o
' •'•F-•-••1
OZONEMIXINGRATIO(PPBV)
Figure6. Comparison
of theozonemixingratioprofilesat Samoa(1995-2000)andtheGalapagos
(1998-2000)for
(a) March-May and (b) September-November.
trajectoriesusing an objective clustering technique [Moody,
1986; Harris and Kahl, 1990] that gives an indicationof the
primary flow regimesinfluencinga particularlocationat a given
altitude. The trajectoriesare computedtwice daily (0000 and
1200 UT) for 10 days backward in time. During 10 days of
travel,air parcelsmay have undergonediabaticprocesses
that are
percentileof the data. The whiskers representthe inner 90th
percentile of the data. The solid diamond is the mean. The
September-Novemberperiod has greatly enhanced variability
especiallyin the 2-10 km layer when comparedto the MarchMay period. These seasonalprofiles are basedon severalyears
of data and representthe variability that is primarily contributed
not accounted for in the model used here, and these are an
by the short-termfluctuationswithin a season.
In the Galapagosthe time-height crosssectionfor 1999 (Plate
important contribution to the uncertainty in the computed
pathwayof the air parcel [Merrill, 1996]. Transportfrom sources 2), the only completeyear available for this site, showsmany of
or sinksmore than 10 daystravel time from a particularsite could the samefeaturesseen at Samoa.The seasonalprofiles for the
also have a significantinfluence on ozone levels measuredat the Galapagos (Figure 2) show similar behavior to Tahiti with a
site.
maximum in the variability in September-Novemberand a
minimum in the March-May season.The variability is greaterat
3. Day-to-Day Variations
Tahiti in September-Novemberthan it is in the Galapagos.At
Samoa(Figure 3) the variability in September-November
is also
Variationson a timescaleof severaldaysare a largesourceof
the variabilityseenin tropospheric
ozoneat thePacifictropical larger than at the Galapagosbut not quite as large as at Tahiti.
sitesstudiedhereas will be shownin the analysisin this section. Since Fiji (Figure 4) also showsgreater variability than that of
Thesevariationshavebeenstudiedusingsurfaceobservations
at the Galapagosduring this September-November,this feature is
likely a real differencebetweenthe easternand westernPacific.
Samoa [Harris and Oltmans, 1998]. The surfacevariationswere
foundto resultfrom changesin airflow to the sitethat tapped
differentsources
andsinks.Duringall seasons,
air comingfrom 4. Seasonal Variation
higher latitudesand altitudeshas -50% more ozone than air with
a tropicalorigin [Harris and Oltmans,1998]. Althoughthe
As canbe seenin Figures1 - 4, notonlyis thevariability
of
ozonesoundings
are doneon an approximately
weeklyschedule ozonein thetroposphere
greaterin September-November
thanin
compared to the continuous observationsat the surface, the
March-May, but also the mean (and the median)valuesare
variabilityfromsounding-to-sounding
captured
in thetime-height greateras well. Time-heightcrosssections
basedon 14-day
crosssectionof ozonemixing ratio for 1996 at Samoa(Plate 1) is
averagesover the entire recordof measurements
(which varies
also apparent.During 1996 at Samoathe soundingswere done somewhat
foreachsite)areshownin Plate3. Although
thereare
twicea weekduringAugustandSeptember
sothatthevariability unique featuresat each site, a pattern is discernablethat is
is well represented
duringthis time of the year.This can alsobe reflected
at all sites.In thewesternPacific,thereis a prominent
seenin the plot of the averageprofilesat 0.25-km incrementsfor layerof enhanced
ozonein themidtroposphere
in September
and
individualseasons
at Tahiti (Figure1). The medianis the vertical October.
Therealsoseems
to bea distinct
pulsein JuneandJuly
line insideof the box, and the box represents
the inner 50th of somewhatsmallermagnitudewith a relativeminimumin late
OLTMANS ET AL.: OZONE IN THE PACIFIC TROPICAL TROPOSPHERE
32,509
0
0
..,-•
0
0
•
ß
4-o
0
U3
0
U3
0
•Z3
0
,._
.-
U3
....,
0
._
[m•] opn•:•lV
[Ua>l]opnl,llV
E• '""
0
0
•
•
o
o
[tull oPn11•,IV
[m>[] oPn•I!:IIV
•
0 :•
O
32,510
OLTMANS ET AL.: OZONE IN THE PACIFIC TROPICAL TROPOSPHERE
B
0ou
,
0ø
' •-
.-•
'
•",,,
• - • 1•
30os
•_
234X
• 523
%
3
60os -
10øE 40øE
70øE
100øE
130øE 160øE 170øW 140øVV 110øVV 80øW
Longitude
ß
,.
30ON
-•
30os
_
60øS-•
3
•
10øE 40øE
•
70øE
•
•
•
•
100øE 130øE 160øE 170øW
•
140øW 110øW 80øW
Longitude
Figure7. Clusters
of trajectories
arrivingat Tahitiat 1 km for (a) December-February
and(b) SeptemberNovember
for theperiod1990-1999.Theclusters
arenumbered
1-6,andthepercentage
of thetotalnumber
of
trajectories
in eachclusteris alsoshown.
JulyandAugust.Associated
withthisJune-July
enhancement
in 1986-1990periodhas somewhatmoreozonethan in the 1995the midtroposphere,
ozonein the boundarylayeralsoincreases, 2000 periodthat appearsto be drivenby severalprofileswith
and this producesa seasonalmaximumnear the surfacethat greater ozone amounts,particularlyin the low and mid(Figure5a). In September-November,
however,the
occursearlier than the peak in the midtroposphere.During troposphere
amountsand variabilityare very similarduringthe two periods
(Figure5b).
In the Galapagosthe generalpictureis similarto the western
stratosphere. In January-May, ozone at all altitudes is almost
alwayslessthan at a correspondingaltitude during the rest of the Pacificsites(Figure6), but with someimportantdifferences.The
ozonemaximumoccursearlierin the year and
year. Also, at this time of year, ozone in the upper troposphere midtropospheric
australwinter and spring, greaterozone amountsdescendto
-10-12
km.
This seems to be associated with ozone in the lower
has smaller mixing ratios, approachingthose seenat the surface
which are almostalwaysquite small (<15 ppbv)in this season.
An earlier set of profiles obtainedat Samoafrom August 1986
to January1990 (Figure 5) show similar characteristics
to those
seen in the more recent 5-year data set. For March-May the
diminishes earlier as well (see Plate 3). Near the surface the
seasonal variation is smaller. The most noticeable difference is
the lack of very low mixing ratios in the middle and upper
troposphere
especiallyduringtheJanuary-May
time of year.This
is consistent with the fact that the western Pacific is a more
OLTMANS
ET AL.' OZONE IN THE PACIFIC
TROPICAL
60ON•,
o
TROPOSPHERE
. ..,•
32,511
•,.
30ON
- '•
00
--
•
4
30os_
60os -
0o
30OE
60øE
90øE
120øE
150øE
180Wø 150øW 120øW
Longitude
60øN
0ON
0o
3005
60os -
0o
30OE
60øE
90øE
120øE
150øE
180Wø
150øW 120øW
Longitude
Figure8. Clustersof trajectories
arrivingat Tahitiat 6 km for (a) March-Mayand(b) September-November
for
1990-1999.
greatestinfluenceon this site. The following analysisshowsthat
while the airflow patternsat the westernPacific sites are quite
similar, they are much different from those in the Galapagos.
similar to the differences
between
the eastern and western
The differences in airflow direction with seasonare greatestin
Pacific. It is also clear that in the Galapagosozone amountsare the boundary layer. Wind shear at all altitudes is somewhat
greater than in the western Pacific above 4 km in September- greaterin the Galapagosthan in the westernPacific. To carry out
this analysis,a 10-year climatology of 10-day back trajectories
November and above 5 km in March-May.
wascomputedat threelevels(1, 6, and 13 km) for all four of the
sitestbr each of four seasons. These trajectorieswere grouped
5. Flow Characteristics
convectiveregionwhereboundarylayer air can be mixed into the
uppertroposphere.Other than this noticeablelack of low ozone
amounts the differences among the western Pacific sites are
into six clusters at each site for each altitude.
The choice of six
Although the seasonalozone behavior at all of the sites has clusterswas somewhatarbitrary, but it has been found that six
some common featuressuch as the maximum during the austral clusters usually gives sufficient transport detail while not
spring(September-November),this doesnot by itself imply that producing a cluttered figure or redundant transport types.
of variousregimesare discussedhere
the sourcesand sinksinfluencingthe measuredozone at the sites Examplesrepresentative
are the same. For example, the proximity of the Galapagosto (Figures7-12).
Becauseof the similarityin flow patternsat the westernPacific
South America suggeststhat this continentis likely to have the
32,512
OLTMANS ET AL.' OZONE IN THE PACIFIC TROPICAL TROPOSPHERE
60øN
-•
30ON -
00
--
3005 -
60os -
0o
30OE
60øE
90øE
120øE
150øE 180Wø
150øW
Longitude
Figure9. Clusters
oftrajectories
arriving
atTahitiat13kmforallseasons
for1990-1999.
comingfromthe southandwestduringJunesitesthe description
of the behaviorat Tahiti is usedto indicate 65% of trajectories
the westernpatternwith some differencest¾omthe other sites August,while at Tahiti andSamoathe percentage
of southerly
noted. At 1 km, a level in the marine boundarylayer, the largest flow is closerto 30% (Figure7b).
At 6 km above Tahiti airflow (Figure 8) is more uniformly
seasonalcontrastis betweenDecember-February(Figure7a) and
t¾om
the westcomparedwith low altitudeflow, with only-20%
June-August
(Figure7b). Thesearetheseasons
of minimumand
maximum surfaceozone at Samoa [Harris and Oltmans, 1998]. comingfrom the easton an annualbasis. Thereis a southerly
During the australsummer(December-February)
flow at the
1-km level is predominantly
from the tropicalPacificfor all of
the sites in the western Pacific. In the other seasons, 15-65% of
the flow is from the west and more southerlylatitudeswith the
component
to the airflow,butit usuallydoesnotextendsouthof
40øSwithin 10 days,in contrastto the low-levelflow thatreaches
muchhigherlatitudes.Throughout
the australwinterandspring
(June-November),at least 25% of the trajectoriesarriving at
largestpercentage
of flow from the southoccurringin austral Tahiti come from as far west as the mid Indian Ocean and -10%
winter. At Fiji (not shown)this featureis mostprominentwith reachsouthernAfrica (Figure 8b) within 10 days. To Fiji the
40øN
-I
20ON -
,_--T/2
25%\
10o5 -
)
2,•k
4005 l:.IOoW
100øW
70øW
40øW
10øW
Longitude
Figure 10. Clustersof trajectories
arrivingat theGalapagos
at 1 km for all seasons
for 1990-1999.
OLTMANS ET AL.' OZONE IN THE PACIFIC TROPICAL TROPOSPHERE
32,513
60ON -
30ON -
0o -
30o5 -
60o5 -
16001::
170ow
140øW
110øW
80øW
50øW
20W ø
Longitude
30øN
• 0ø
3005
60os
160øE
170øW
140øW
110øW
80øW
50øW
20W ø
Longitude
Figure11. Clusters
of trajectories
arrivingat theGalapagos
at 6 km for (a) December-February
and(b) JuneAugustfor 1990-1999.
flow is even more vigorous from the west and -5% of the
seasons. There is some variation with season(not shown) with
trajectories
comefromSouthAmerica.A numberof trajectories December-Mayshowing10-15% of the trajectoriescomingfrom
also have their 10-day originsover northernAustralia. In the
summer,flow is less vigorousand some trajectorieshave a
northerlycomponent.
In theuppertroposphere
(13 kin), flow is from the west,but a
the Atlantic. In June-Novemberthe flow is exclusivelyfrom the
south,often paralleling the South American coast. The flow at 6
km is in strong contrastto that in the boundarylayer. On an
annualbasis,75% of the trajectoriesarrive at San Cristobalfrom
majorityof the trajectories
also have a northerlycomponent the east and have passedover the South American continentin
(Figure9). This northerlyflow bringsair fromnearor evennorth the previous10 days. In December-February
(Figure 1l a), -10%
of the equatorand is strongest
in September-November
but is of trajectoriesarrive from the tropical northPacific, another10%
presentin otherseasons
as well. As in the midtroposphere,
a from the tropical south Pacific, and the remainder from the east
significantnumber of the trajectoriescome from the Indian off continentalSouthAmerica. During June-August(Figure 1lb)
Ocean and Africa.
and continuingthrough November (not shown), -10-15% of the
The flow patterns
in the Galapagos
differsignificantly
from flow is from the tropical south Pacific with the remainderfrom
thosein thewestern
Pacific.At thelowestlevel(Figure10)the the eastor with little movement(cluster3 in Figure 1lb). At 13
flow is overwhelminglyt¾omthe southand over the oceanin all km (Figure 12) the flow is about equally divided betweenweak
32,514
OLTMANS ET AL.' OZONE IN THE PACIFIC TROPICAL TROPOSPHERE
60oS -
90øE 120øE 150øE 180ø 150øW 120øW 90Wø 60øW 30øW
Longitude
Figure12. Clusters
oftrajectories
arriving
attheGalapagos
at 13kmforallseasons
for1990-1999.
flow from the eastor northeastand vigorousflow from the west. mixingratioof 105 ppbvat 5-6 km hasa trajectory(Figure14)
There is a fairly strong seasonalitywith December-May that reaches back to southern Africa 10 days prior to the
in thetropicalPacific,thelayer
dominated
by trajectories
arrivingfromthe west,but abouthalf sounding.As with all soundings
between
the trajectories
duringJune-November
comefrom the eastand of enhancedozoneis alsomarkedby a large depression
thefrostpointandambientair temperatures,
indicative
of dryair.
northeast.
6.
Discussion
TEMPERATURE & FROST POINT (C)
In the previoussectionsthe short-termand seasonalvariations
-100 -80
-60
-40
-20
0
20
40
have been describedas well as the major air flow characteristics
affectingthe ozonesondesitesin the westernand easterntropical
Pacific. In this section,important featuresof the ozone profile
16
are linked to the trajectoriesto show how important sourceand
sink regions for ozone may contribute to both the shorter-term
14
and seasonalvariations. Since the tropics are known to be a
significantarea of biomassburning (see TRACE-A and SAFARI
special sectionsof the Journal of GeophysicalResearch, 101,
1996), particular attentionis paid to characterizingthe possible
influenceof burning on ozone at thesesitesby linking enhanced
ozone layers with flow from potential sourceregions.It appears
that several burning regions contribute to what is seen at these
sites.The low ozonemixing ratios seenin the uppertroposphere
at the westernPacific sitesalso are discussedbriefly.
It is clearly seen in Plate 3 that there is a persistentlayer of
enhancedozone in the midtropospherein the June-November
season.The climatological trajectory analysis also shows this
seasonto be one of regular occurrencesof flow from potential
burning related sourceregions in southernAfrica and Australia
for the westernlocationsand South America for the Galapagos.
Several individual profiles are examined that contain enhanced
0
20
40
60
80
100
120
140
160
midtropospheric
ozone along with the trajectoriescalculatedfor
OZONE (ppbv)
thesecases.An event of very high ozone seenat Fiji and Samoa
Figure13. Ozonemixingratioprofileat Samoa
onOctober
30,
in November1997 is investigatedbecausethe flow path suggests 1998. The thicker solid line is the ozone mixing ratio. The
an Indonesian
source for the enhanced ozone.
A profile with an enhanced midtropospheric layer
characteristicof those seen during the Septemberand October
periodin the westernPacific is shownin Figure 13 for a sounding
done at Samoaon October30, 1998. The peak in the ozone
dashedline is the air temperature
andthe thinnersolidline the
dew/frostpointtemperature.At temperatures
above0øCthe
moisture
profileis withrespect
to theliquidphaseof water(dew
point)andbelowfreezingprofileis for the solidphase(frost
point).
OLTMANS
ET AL.' OZONE IN THE PACIFIC
0o
30øE
TROPICAL
TROPOSPHERE
32,515
30ON -
00•
30os -
60os -
Ef
UJ
0
60øE
90øE
120øE
150øE
1800
Longitude
1
2
3
4
5
6
7
8
9
10
Days to Samoa
Figure 14. Ten-day back trajectoryarriving at Samoaat 5 km at 0000 GMT on October30, 1998. The numbers
alongthe trajectorypathmark the numberof daysbackin time from the endpointthat the air parcelwaslocated.
The elevationchangeof the air parcelis shownin the bottom.
This suggeststhat air with large ozone amounts is not simply
litied from the surface at the source region and transportedto
Samoa. The changein the temperaturelapse rate near 4 km at the
bottomof the layer with enhancedozone suggeststhat the air has
subsided and warmed in its transit to Samoa. A significant
subsidenceof-3 km can be seen in the air parcel basedon the
calculatedtrajectory(Figure 14). In orderfor a layer with sucha
prominent ozone peak to be transported over the distances
suggested
by the trajectorycalculationit would likely have to
remainrelativelyunmixed,both verticallyand horizontally,with
air from surroundinglayers.
The preservationof the layer as well as the validity of the
trajectorycalculationis dependenton the air parcelnot passing
througha region of strongconvection. Inspectionof the OLR
TEMPERATURE& FROSTPOINT(C)
TEMPERATURE& FROSTPOINT(C)
-100
0
20
-80
40
-60
-40
-20
0
20
60
80
100
120
140
OZONE(ppbv)
-100
40
160
0
20
-80
40
-60
-40
-20
0
20
60
80
100
120
140
40
160
OZONE(ppbv)
Figure15. Ozonemixingratioprofilesat Samoaon (a) October2, 1997at 0000GMT and(b) October30, 1997,at
0000 and 1200 GMT.
32,516
OLTMANSET AL.:OZONEIN THEPACIFICTROPICAL
TROPOSPHERE
•
20os-
50os I
80øE
I
I
110øE
I
170øE
140øE
160øW
Longitude
•
0
1
2
,
,
4
5
6
l:)0øW
...•
7
8
9
10
Daysto Samoa
30ON o•
lOON -
20oS -
50os I
120oE
I
150oE
I
180oE
150oW
,
0
1
2
3
4
5
6
7
,
8
9
10
Days to Samoa
Figure 16. Trajectories
to Samoaat 5 km on (a) October2, 1997,at 0000GMT (solid)and12 GMT (dashed)
and
(b) October30, 1997, at 0000 GMT.
maps for October20-30, 1998, showsthat indeedfor the period.Soundingsfrom Irene, SouthAfrica (25.9øS,28.2øE)were
computed
trajectory
paththeair parcelis notlikelyto havebeen begunlate in November1998 so do not includethe periodof the
stronglyinfluenced
by convection.The otherneededingredient Samoasounding.Profilesfrom Irene in 1999 and 2000 from late
in additionto a transport
pathfor theappearance
of theenhanced Octoberdo showozonelevelsof- 100 ppbv at altitudesbelow 10
ozone layer over Samoa is a sourceof such large ozone km (Thompsonand Witte 11999], and the SHADOZ archive at
http://code916.gsfc.nasa.
gov/Data_services/shadoz/). At
La
Reunion,
Madagascar
(21.1øS,
55.5øE)
there
is
a
profile
earlier
in
Africa (the sourceregionsuggested
by the trajectory)duringthis
concentrations. No ozonesonde data are available from southern
OLTMANS
ET AL.' OZONE IN THE PACIFIC TROPICAL
TEMPERATURE & FROST POINT (C)
-100
-80
-60
-40
-20
0
20
TROPOSPHERE
32,517
derivedfrom TOMS total ozonemeasurements
[Thompson
and
Hudson [1999] and the Web site http://metosrv2.umd.edu/
--tropo/)aroundOctober20, 1997,showhighlyelevatedozonein
the regionto the north(0ø-15øS)
of the areaindicated
by the
calculated
trajectory.This techniquedoesnot producereliable
troposphericdata at more southerlylatitudes. Given the
somewhat
uncertainnatureof the trajectorycalculationafter 10
40
18
days and the indication of enhancedozone amountsover southern
Africa at this time year, it is reasonablethat Africa was the
sourceof the layer of high ozoneconcentrations
over Samoaseen
in theOctober30, 1998, profile.
Ozoneprofileswithpeaksgreaterthan-70 ppbv(Figure15)
do not alwayshavetrajectories(Figure16) that extendbackto
Africain 10 days,buttheyalwayshavepathsthathavea strong
westerlycomponentso that 10-dayoriginsare to the west of
Australiainto the Indian Ocean. On many occasions
the
trajectories
passover Australia,usuallythroughthe middleor
southernpart of the continent.The marked contrastthat can occur
0
20
40
60
80
100
120
140
160
OZONE (ppbv)
Figure 17. Ozonemixingratio profileat Tahiti on October3l,
1995.
duringa month(seeFigure 15) illustratesthe dominantrole that
the airflow (Figure 16) to a site plays in the ozone profile,
particularly during this season. For the profile obtained on
October30, 1997 (Figure 15b), the 10-daytrajectory(Figure 16b)
in the midtroposphere
showslittle air movementfrom the vicinity
of Samoaand mixing ratiosare 20-25 ppbv or lessthroughoutthe
troposphere.This is in contrastto the near 80 ppbv mixing ratio
seen at 5 km on October 2 (Figure 15a) when the flow was
westerly(Figure 16a).
the month with midtroposphericmixing ratios near 100 ppbv.
This is somewhatearlier, however, than when the trajectory for
the Samoa profile indicatesthat the air parcel would have been
over southernAfrica. It is not unusualtbr La Reunion to see very
large ozone enhancementsduring this time of the year that are
often associatedwith biomass burning [Randriambelo et al.,
1999, 2000]. Maps of tropicaltroposphericcolumnozone(TTO)
FromFigure8 it is clearthata largefractionof the air parcels
rcachingTahiti (and also Samoaand Fiji) passover Australia.
From the data that are available from the ozonesondes and
trajectories
it is difficultto determine
if trajectories
thatpassover
Australiaare influencedby burningin Australiaratherthanby
burningin southernAfrica [Olson et al., 1999]. In somecases
thetrajectories
showthattheair movesmoreslowly(for example
30ON
10ON
•
20øS -
50os
•
E
I
140oE
I
110oE
I
140oE
I
160oW
170oE
::': 4
Lu
0
1
2
5
4
5
6
7
8
9
Daysto Tahiti
Figure 18. Trajectoryto Tahiti at 5 km on Novemberl, 1995, at 00 and 12 GMT.
10
32,518
OLTMANS ET AL.' OZONEIN THE PACIFICTROPICALTROPOSPHERE
TEMPERATURE
&FROST
POINT
(C)
-100 -80 -60 -40 -20 0
TEMPERATURE
&FRO•T
POINT
(C)
20 40
-100 -80 -60 -40 -20 0
18
20 40
16
14
12
• •o
6
4
2
0
0
20
40
60
80
100
120
140
160
0
20
40
OZONE(ppbv)
60
80
100
120
140
160
OZONE
(ppbv)
Figure
19.Ozone
mixing
ratio
profile
at(a)FijionNovember
19,1997,
and(b)Samoa
onNovember
20,1997.
30oN -
30ON -
Samoa
10oN -
lOoN -
<t?' /'g
6 km
..• 20øS
-
.• 20øS-
50oS--
Nov.
20,1997
I
80oE
8 km
I
110oE
140oE
• '
Nov.
20,1997
50oS-
I
I
I
170oE
160oW
130oW
I
80OE
I
110oE
Longitude
140oE
I
I
170oE
160oW
Longitude
30ON -
30ON -
ß
S'moa
lOoN -
lOoN
,
•
8km
10 km
20oS -
:m 20os -
50os -
Nov.
23,1997 •
Nov. 21, 1997
I
80OE
110oE
•
.
50oS-
I
140oE
I
170oE
I
160oW
I
130oW
Longitude
800E
110oE
I
140OE
I
I
170oE
160oW
Longitude
Figure20. Trajectories
to Samoa
at6 kmonNovember
20, 1997,andat 10kmonNovember
21,1997,andtoFiji
at 8 km on November 20, 1997, and November 23, 1997.
OLTMANS
ET AL.: OZONE
IN THE PACIFIC
•0
-60
-40
-20
0
20
TROPOSPHERE
32,519
ratherthansouthernAfrica). Althoughburningis knownto take
placeduringthis time of year in northernAustralia[Olsonet al.,
1999] there is not informationfrom ozone profilesfrom this
regionto confirmthathighozoneis produced
in theregionandin
particularwhetherthe high ozoneamountsseenin the layer at
TEMPERATURE & FROST POINT (C)
-100
TROPICAL
40
Tahiti came from this source.
In 1997, extensiveburningtook placein Indonesiaassociated
withdroughtconditions
thatwerea consequence
of thestrongE1
Nifio. Ozoneprofile measurements
[Fujiwara et al., 1999] and
analysis
of satellitedata[Thompson
et al., 2001]foundveryhigh
ozoneamountsin Indonesiaand Malaysiain connectionwith the
burningin theregio_n_
On November19 and20 at Fiji andSamoa
(Figure19) someof the highesttropospheric
ozoneamountswere
seenfor any event recordedat thesesites. These enhancements,
which encompassed
the entire troposphere
abovethe two sites,
increasedthe total tropospheric
columnby -20 DU, or morethan
70% over averagevaluesfor the month. The trajectoriesshow
thatat bothsitesair wascomingfromIndonesia(Figure20a-20d)
0
20
40
60
80
100
120
140
throughout
thedepthof thetropospheric
column.Althoughthese
profiles were the only ones measuredduring the event,
trajectories
showthat a flow patternnearlyidenticalwith the one
duringthislargeozoneenhancement
persisted
for -5 daysaround
thetimethatthe profileswereobtained(Figure20d). Although
Novemberwas not the peak in the burningover the Maritime
160
OZONE (ppbv)
Continent,the presenceof large ozonecolumnamountsover the
regioncanbe seenfrom the TI'O mapsfrom the period(Plate4)
thatthetrajectories
showair movingfromthisregiontowardFiji
and Samoa. Given the large vertical extent of the ozone
enhancement
seenin the profile measurements
andthe coherence
of thetrajectorypathwith altitude,it appearsthata largeplume
Figure21. Ozonemixingratioprofileat SamoaonFebruary
20,
1999.
cluster3 in Figure 8b). An exampleof sucha caseis shownin
the profile of October 31, 1995, at Tahiti (Figure 17). The
correspondingtrajectory just reaches the York Peninsula of
Australia (Figure 18) in 10 days. The indication of relatively
slow transport,the large peak mixing ratio, and the thin vertical
extent of the layer suggestthat the sourceof the elevated ozone
seen at Tahiti may have been relatively nearby (i.e., Australia
20øN
- /
an area of enhanced column ozone amounts of-50
DU near the
datelineandextendingfrom -7øS to15øSthatlasted-10 days.
c•
lOON'
•
of air with high ozone concentrationslid southeastward
from the
MaritimeContinent
regionoverFiji andSamoa.TheTTO maps
at thetimeof theprofilesfromSamoaandFiji showveryclearly
••,•
20os -
50os I
[:::: 80øE
0
I
110øE
1
2
3
I
140øE
4
5
170øE
6
7
8
160øW
9
10
Days to Samoa
Figure22. Trajectory
to Samoaat 13kmonFebruary
20, 1999,at0000GMT (solid)and1002GMT (dashed).
32,520
OLTMANS
ETAL.:OZONE
IN THEPACIFIC
TROPICAL
TROPOSPHERE
showsthatair originated10 daysearlierin northeastern
Australia
in a regionassociated
with extensiveconvection
duringthistime
of year. The OLR map (Plate 5) for the periodprior to the
soundingshowsthat convection(evidencedby the very low
OLR) was widespreadin the region along the South Pacific
Convergence
Zone(SPCZ)throughwhichthe air parceltraveled
basedon thetrajectorycalculation.The validityof thetrajectory
TEMPERATURE & FROST POINT (C)
-100
•0
-60
-40
-20
0
20
40
16
14
calculationunderthesevery convectivecircumstances
is opento
question,
but sincethe calculation
is for an altitudeof 13 km,
thereis somereasonableexpectationthat the air parcelpath is
abovethelayerwith the strongest
verticalmotions.At anyrateit
is clearthat air reachingSamoaduringthis periodis likely to
havepassed
througha regioninfluenced
by convection.
Thelow
ozonemixingratiosare not usuallya resultof mixingdirectly
from the surfacenearthe site,althoughsurfacemixingratiosare
6
low. This can be seen in Plate 3a, 3c, and 3d where the minima
in theuppertroposphere
arenotconnected
to thelow valuesat
the surfaceas illustratedby the individualprofile shownin Figure
21. Instead,there is normally a relative maximum in the mixing
ratio in the midtroposphere.It is more likely that air with very
low ozone is mixed vertically by convection in the region of
o
northernAustraliaand easternIndonesiaand transportedto these
o
20
40
60
80
lOO
12o
14o
1 io
sitesin the uppertroposphere[Kley et al., 1997].
OZONE (ppbv)
In the Galapagosmany of the profiles during August-October
Figure23. Ozonemixingratioprofileat theGalapagos
on alsoshowpronouncedmidtroposphericpeaksas seenon October
October
9, 1999.Thethickersolidlineistheozonemixingratio. 9, 1999 (Figure 23). The trajectory for this event (Figure 24)
Thedashed
lineis theair temperature,
andthethinnersolidline
crossesBrazil in less than 10 days over a region that is heavily
is the dew/frostpoint temperature.
influenced by biomass burning during this time of the year
[Fishman et al., 1996]. Clear evidence for the enhancementof
This is consistentwith the column amountsmeasuredin the ozone over Brazil is seen in two profiles from Natal, Brazil
(5.4øS,35.4øW), on September29 and October6, 1999 (Figure
profiles.
As wasnotedearlier,duringtheaustralsummer,
ozonemixing 25), where ozone mixing ratios exceed 100 ppbv. Althoughthe
ratiosare generallysmallthroughout
the troposphere
in the air parceltrajectorypassesalmostdirectly over Natal, it may be
westernPacific. Values less than 10 ppbv are seenas in the that both Natal and the Galapagosare influencedby the area to
profile
forFebruary
20,1999,atSamoa
(Figure
21),where
inthe the west and south of Natal, where the burning actually takes
13-15kmlayerozoneis -5 ppbvsimilarto theverylowvalues place [Kirchhoff et al., 1996]. This is stronglyindicatedby the
seenbyKleyet al. [1996].Thetrajectory
at 13km(Figure22) TTO map for October 4-6, 1999 (Plate 6), that identifies the
20ON -
0 O --
30os -
90ow
120øW
E
60øW
Longitude
8[
,
....w 0
o
30øW
1
2
3
,
,
,
,
4
5
6
7
8
9
Daysto Samoa
Figure
24.Trajectory
totheGalapagos
at6 kmonOctober
9, 1999,at0000GMT.
lO
OLTMANS
ET AL.' OZONE
IN THE PACIFIC
TROPICAL
TROPOSPHERE
32,521
10
•
t
-5
-1o
80
1O0
120
140
160
180
Longitude
!
Modified-Residual
Tropospheric
Ozone
03ohson
Units)
Plate 4. Map of the averagetropical troposphericozone residualcomputedfrom the Total Ozone Mapping
Spectrometer
(TOMS) totalcolumnozonefor the periodNovember10-12, 1997.
20ON --
0 m
!
I
90OE
180 ø
Longitude
180
200
220
240
260
280
300
Plate 5. Map of the averageoutgoinglongwaveradiation(OLR) for the periodFebruary12-20, 1999. Plot from
the NOAA-CIRES Climate DiagnosticsCenter.
32,522
OLTMANS
ET AL.: OZONE IN THE PACIFIC TROPICAL
TROPOSPHERE
lO
l
-5
.1o
-80
-lOO
-60
-20
-40
Longitude
o
oo
,.-,
o
Ig
•
..•
,,.,,1
Modified-Residual
Tropospheric
Ozone
(Dobson
Units)
Plate6. Map of theaveragetropicaltropospheric
ozoneresidualcomputed
fromtheTOMS totalcolumnozonefor
the periodOctober4-6, 1999.
15ON --
90oW
60øW
30ow
Longitude
180
2.00
220
240
260
280
300
Plate 7. Map of the averageOLR for the periodMarch 31, 1998. Plot from the NOAA-CIRES Climate
DiagnosticsCenter.
OLTMANS
ET AL.' OZONE IN THE PACIFIC
TROPICAL
TEMPERATURE& FROSTPOINT(C)
•
- 100
-80
20
40
TROPOSPHERE
32,523
TEMPERATURE& FROSTPOINT(C)
-60
-40
-20
0
20
60
80
100
120
140
40
- 100
-80
-60
-40
-20
0
20
20
40
60
80
100
120
140
40
10
U.I
1:3
::3
!::: 8
0
160
0
OZONE(ppbv)
160
OZONE(ppbv)
Figure 25. Ozonemixingratioprofilesat Natalon (a) September
29, 1999,and(b) October6, 1999.
region of highest troposphericcolumn ozone as being over
centralBrazil. From the Galapagosand Natal soundingsit is
clear that enhancedozone amounts resulting from biomass
burningare transportedboth to the east [Kirchhoff et al., 1996]
wherethey appearat Natal andto the westreachingat leastas far
as the Galapagos. It is possiblebasedon the case studiedhere
that air reachingNatal from the region of burning may be
recirculatedback to the west toward the Galapagosor may be
transported
directlyfrom the burningregionto the Galapagos.
Investigationof eachprofile with an enhancedozonelayer in
the midtroposphere
during the August-Octoberperiod showeda
trajectorythat passedover Brazil. At times of the year when
burningis not occurringin Brazil, profiles at the Galapagosdo
not show pronouncedtroposphericpeaks (Figure 26). Profiles
with mixing ratios near 50 ppbv in the midtroposphereare
occasionallyseen (Figure 26a). These higher amounts are
generally associatedwith flow from continental South America
(Figure 27a). However, flow from the continentmay also have
TEMPERATURE& FROSTPOINT(C)
TEMPERATURE
& FROSTPOINT(C)
-100
•
-80
-60
-40
-20
0
20
100
120
140
- 100
40
-80
-60
-40
-20
0
20
100
120
140
40
10
u.I
1:3
::3
I--
F-- 8
0
20
40
60
80
OZONE
(ppbv)
160
0
20
40
60
80
OZONE
(ppbv)
Figure26. OzonemixingratioprofilesattheGalapagos
on(a) March25, 1999,and(b) April4, 1998.
160
32,524
OLTMANS ET AL.: OZONE IN THE PACIFIC TROPICAL TROPOSPHERE
20ON -.
0 0 --
30os 120øW
90ow
60ow
30ow
Longitude
0
1
2
3
4
5
6
7
8
9
10
Days to Samoa
20ON -
0o -
•
3005 -
I
120oW
iTi
0
0
I
90øW
1
2
3
60øW
4
.5
6
7
30øW
8
9
10
Daysto the Galapagos
Figure27. Trajectories
totheGalapagos
at6 kmon(a)March25,1999,at1200GMTand(b)April4, 1998at120
GMT.
relativelylow ozoneamounts(Figures26b and 27b). It appears
that even thoughtransportto the site is similar,that unlessthis
flow taps a sourceregion with higher ozone,the passageof air
parcelsover a continentis not sufficient(althoughit doesappear
necessary)
to give enhancements
in the ozoneamount.The April
4, 1998, example(Figure26b) wasalsonearthe endof the 19971998 E1 Nifio-southernoscillation(ENSO) warm phasewith well
abovenormalprecipitationboth in the Galapagosand western
equatorialSouthAmerica. The profile of April 4 suggests
that
convectionhasinfluencedthe profile sincehumidityis high up to
-9 km (the depressionbetweenthe air temperatureand frostpointtemperature
shownin Figure26b is small). The convection
thatis responsible
for thecharacteristics
of thisprofileis likelyto
have occurred over equatorial South America where the
trajectory(Figure27b) showsthe air parcelintersecting
an area
of very strongconvection
(Plate7). Convection
nearerto the
OLTMANS ET AL.: OZONE IN THE PACIFIC TROPICAL TROPOSPHERE
Galapagosmay have influencedthe characteristics
of the ozone
profile at lower levels (below -6 km), but this cannotbe
discernedfrom the profile information.
7.
Conclusions
32,525
ExploratoryMission in the tropical Pacific: PEM-Tropics A, AugustSeptember1996,J. Geophys.Res.,104, 5567-5583, 1999.
Kirchhoff, V.W.J.H.,
J.R. Alves, and F.R. da Silva, Observations of
ozone concentrationsin the Brazilian cerrado during the TRACE A
field expedition,J. Geophys.Res., 101, 24,029-24,042, 1996.
Kley, D., P.J. Crutzen, H.G.J. Smit, H. Vomel, S.J. Oltmans,H. Grassel,
and V. Ramanathan, Observations of near-zero ozone concentrations
From an extensiveset of ozone profiles obtainedover the past
several years in the tropical Pacific, the picture emergesof a
regionof both very low troposphericozone amounts(-10 ppbv)
and surprisinglylarge (-100 ppbv) mixing ratiosat all of the sites
studied. The low concentrationsare expected in light of the
strongphotochemicalsink in this region [Kley et al., 1997], and
the importantrole of convection,particularlyduring the austral
summer. On the other hand, the ubiquitous presence of
midtroposphericlayers of enhancedozone demonstratesthat the
influence of biomass burning is widespread throughout the
tropical zone during August-November.In several examplesa
strong connectioncould be made between enhancedozone
amountsin the burning sourceregion and layers of large ozone
concentrationin both the eastern and western tropical Pacific.
On average,ozone at the equatorialGalapagossite in the eastern
Pacific is greater than at the three western tropical Pacific sites.
However, for individual profiles the midtropospheric
enhancementsseen in the western Pacific attain larger mixing
ratios than those seen in the Galapagos. This is in spite of the
fact that the Galapagosare located much closer to a potential
sourceof biomassburning (Brazil) than are the westernPacific
sites (southern Africa). The proximity of other sources in
Australia and Indonesia may partially account for this by
providing closer source regions for the western Pacific sites.
Alternatively, the much greater amount of biomassburned in
southernAfrica may producelarger ozone enhancements[Olson
et al., 1999]. The extensiveburningthat took place in Indonesia
and Malaysia during the latter half of 1997 appearsto have been
the sourceof very large ozoneamountsseenin Samoaand Fiji in
November
1997.
over the convectivePacific: Effects on air chemistry, Science,274,
230-233, 1996.
Kley, D., H.G.J. Smit, H. V6mel, H. Grassl,V. Ramanathan,P.J. Crutzen,
S. Williams, J. Meywerk and S.J. Oltmans, Troposphericwatervapour and ozone cross-sectionsin a zonal plane over the central
equatorialPacific, Q. J. R. Meteorol. Soc.,123, 2009-2040, 1997.
Komhyr, W.D., R.A. Barnes, G.B. Brothers, J.A. Lathrop, and D.P.
Opperman, Electrochemical concentration cell ozonesonde
performanceevaluationduring STOIC 1989, J. Geophys.Res., 100,
9231-9244, 1995.
Merrill, J.T., Trajectory results and interpretationfor PEM-West A, J.
Geophys.Res., 101, 1679-1690, 1996.
Moody, J. L., The influenceof meteorologyon precipitationchemistryat
selectedsitesin the easternUnited States,Ph.D. thesis,176 pp., Univ.
of Mich., Ann Arbor, Michigan, 1986.
Olson,J.R., B.A. Baum,D.R. Cahoon,andJ.H. Crawford,Frequencyand
distributionof forest, savannaand crop fires over tropical regions
duringPEM TropicsA, J. Geophys.Res., 104, 5865-5876, 1999.
Randriambelo,T., J.L. Baray, S. Baldy, and P. Bremaud,A casestudyof
extremetroposphericozonecontaminationin the tropicsusingin-situ,
satelliteand meteorologicaldata, Geophys.Res. Lett., 26, 1287-1290,
1999.
Randriambelo,T., J.-L. Baray, and S. Baldy, Effect of biomassburning,
convective venting, and transport on troposphericozone over the
Indian Ocean: Reunion Island field observations,J. Geophys.Res,
105, 11,813-11,832, 2000.
Singh,H.B., et al., Biomassburninginfluenceson the compositionof the
remote South Pacific troposphere:Analysis based on observations
from PEM-TropicsA, Atmos.Environ.,34, 635-644, 2000.
Stoller, P., et al., Measurementof atmosphericlayers from the NASA
DC-8 and P3-B aircraftduringPEM-TropicsA, J. Geophys.Res.,104,
5745-5764, 1999.
Thompson,A.M., and R.D. Hudson,Tropical troposphericozone(TTO)
mapsfrom Nimbus 7 and Earth-ProbeTOMS by the modifiedresidual
method: Evaluation with sondes, ENSO signals and trends from
Atlantic regional time series,J. Geophys.Res., 104, 26,961-26,975,
1999.
Thmnpson, A.M., and J.C. Witte, SHADOZ (Southern Hemisphere
Acknowledgments. The authorsacknowledgethe helpful remarks of
two reviewers whose commentsstrengthenedthe paper and improved
the discussionof the connection between ozone source regions and
featuresin the ozone profiles. The assistanceof R. Rossonin editing
the manuscript and Jacquelyn Witte and Alexander Frolov in
preparingthe TTO mapsis also acknowledged.
Additional
Ozonesondes:
A new ozonesonde
data set for the Earth
sciencecommunity,Earth Obs., 11, 27-30, 1999.
Thompson,A.M., J.C. Witte, R.D. Hudson,H. Guo, J.R. Herman,andM.
Fujiwara, Tropical troposphericozoneand biomassburning,Science,
in press,2001.
M. Agama, J. Cornejo,and F. Paredes,INAMHI, San Cristobal,
Galapagos,Ecuador.
R.J. Bendura,Mail Stop483, NASA LangleyResearchCenter,
Fishman, J., J.M. Hoell Jr., R.D. Bendura, R.J. McNeal, and V.W.J.H.
Kirchhoff, NASA GTE TRACE A Experiment(September-October Hampton, VA 23681.
H. Enriquez, INAMHI, Quito, Ecuador.
1992):Overview,J. Geophys.Res.,101, 23,865-23,879, 1996.
J. M. Harris, B. J. Johnson,S. J. Oltmans,and H. V6mel, Climate
Fujiwara, M., K. Kita, S. Kawakami, T. Ogawa, N. Komala, S.
Saraspriya,and A. Suripto,Troposphericozoneenhancements
during MonitoringandDiagnosticsLaboratory,NOAA, Boulder,CO 80305.
(soltmans
@cmdl.noaa.gov)
the Indonesianforest fire events in 1994 and 1997 as revealed by
F. Hasebe,IbarakiUniversity,Mito, Japan310-0056.
ground-based
observations,
Geophys.Res.Lett., 26, 2417-2420, 1999.
V. W. J. H. Kirchhoff, INPE, Caixa Postal 515, 12201-970 Sao Josedo
Gregory,G.L., et al., Chemicalcharacteristics
of Pacific tropospheric
air
in the regionof the IntertropicalConvergenceZone and SouthPacific Campos,Brazil.
K. Koshy,M. Maata, A. Samad,G. Sami, and T. Tabuadravu,
Convergence
Zone,J. Geophys.Res.,104, 6577-5696, 1999.
Harris, J.M, and J.D. Kahl, A descriptive atmospherictransport Universityof the SouthPacific,P.O. Box 1168,Suva,Fiji.
J.A. Logan,Dept. of Earth andPlanetarySciences,Harvard
climatology for the Mauna Loa Observatory, using clustered
University,Cambridge,MA 02138.
trajectories,J. Geophys.Res.,99, 13,651-13,667, 1990.
M. Shiotani,HokkaidoUniversity,Sapporo,Japan060.
Hams, J.M., and J.D. Kahl, Analysisof 10-day isentropicflow patterns
P. Simon,MeteoFrance,Papeete,Tahiti, FrenchPolynesia.
for Barrow,Alaska: 1985-1992,J. Geophys.Res., 99, 25,845-25,855,
References
1994.
Harris,J.M., andS.J.Oltmans,Variationsin tropospheric
ozonerelatedto
transportat American Samoa, J. Geophys.Res., 102, 8781-8791,
A.M. Thompson,Code916, NASA GoddardSpaceFlightCenter,
Greenbelt, MD 20771.
1998.
Hoell, J.M, D.D. Davis, D.J. Jacob, M.O. Rogers, R.E. Newell, H.E.
Fuelberg, R.J. McNeal, J.L. Raper, and R.J. Bendura, Pacific
(ReceivedSeptember20, 2000; revisedDecember19, 2000;
acceptedDecember21, 2000.)