GEOPHYSICAL RESEARCH LETTERS, VOL. 24, NO. 21, PAGES 2573-2576, NOVEMBER 1, 1997
Advection and diffusion of Indonesianthroughflow water within
the Indian Ocean South Equatorial Current
Arnold L. Gordon1, ShubinMa 1, DonaldB. Olson2, PeterHacker3,
Amy Ffieldl, LynneD. Talley4, DouglasWilson5 andMolly Baringer
5
Abstract. Warm, low salinity Pacific water weaves
through the IndonesianSeasinto the easternboundaryof
the Indian Ocean. The Indonesian Throughflow Water
(ITW) addsfreshwaterinto the Indian Oceanas it spreads
by the advectionand diffusion within the Indian Ocean's
South Equatorial Current (SEC). The low salinity
throughflowtrace, centeredalong 12øS,stretchesacross
the Indian Ocean, separating the monsoon dominated
regimeof the northernIndian Oceanfrom the moretypical
subtropical stratification to the south. ITW is well
representedwithin the SEC thermocline,extendingwith
perhapspart of a large interoceannetwork of thermohaline
fluxes [Gordon, 1986]. It may als6 influence SST and
associatedocean atmospherecoupling within the Indian
Ocean, and hence of special interest to monsoonclimate
research.
Using the WOCE meridional WHP CTD sectionsand
hull mountedADCP data [Ffield, 1997] the ITW transport
contribution
to the thermocline
of the SEC is determined as
well as its meridionalisopycnalmixing with Indian Ocean
water.
Indonesian Throughflow Water
concentrations above 80% of initial characteristics from the
sea surface to 300-m within the eastern half of the Indian
The ITW characteristics
are definedby dataobtainedin
Ocean, with 60% concentration reaching well into the
westernIndian Ocean. The ITW transportwithin the SEC
the Timor Sea [Fieux et al., 1994, 1996a; Ilahude and
variationsof the injection rate at the easternboundaryand
to the likelihood of a zonally elongatedrecirculationcell
between the Equatorial Counter Current and the SEC
within the Indian Ocean.Lateral mixing dispersesthe ITW
plume meridionally with an effective isopycnalmixing
profile, a product of freshwaterinput and strongvertical
mixing within the indonesianSeasthat tends to mute the
salinity stratification of the Pacific throughflow water
[Ffield and Gordon, 1992; Hautala et al., 1996]. A major
Gordon, 1996; Gordonand Fine, 1996]. The throughflow
variesfrom 4 to 12 x 106 m3sec-1, partly in response
to thermocline layer is represented by a nearly isohaline
coefficient of 1.1 to 1.6 x 10 4 m2sec-1.
Pacific source is drawn from the North Pacific thermocline
flowing along Makassar Strait, although the lower
thermoclineand deeper waters of the IndonesianSeasare
dominatedby South Pacific water flowing at a slowerrate
throughthe deep channelseast of Sulawesi[Gordonand
Introduction
The SouthEquatorialCurrent(SEC) of the IndianOcean
forms the boundary between the monsoonderived water
masses of the northern Indian Ocean, marked by the
evaporative conditions of the Arabian Sea and excess
freshwater conditions of the Bay of Bengal (as best
exemplified by their sharply different surface salinity,
[Conkrightet al.., 1994]), with the moretypicalsubtropical
Fine, 1996].
The throughflow transport appearsto have a seasonal
signal, with maximum values occurringduring the boreal
summer, with essentially zero values during the boreal
winter [Fieux et al., 1994, 1996b;Ariefand Murray, 1996].
Meyers et al., [1995] XBT basedestimatefor the 1983 to
1989 yields an annual mean of 5 Sv passingthrough the
stratification of the southernIndian Ocean. However, what upper 400 meters across a section between Java and
particularly sets the Indian Ocean SEC apart from its Australia (Figure 4c of Meyers, 1996 using 1983 to 1994
cousins of other oceansis the presenceof low salinity XBT data, indicatesa slightly lower annual mean) with a
Pacific thermoclinewater injectedat the easternboundary, 12 Sv August-Septembermaximum, and essentiallyzero
in what is usually referred to as the Indonesian transportin May-June and October-November.Quadfasel
Throughflow Water (ITW). The ITW is consideredto be et al., [1996] find that of the 22 Sv transportwithin the
an importantelementin the heat and hydrologicalbudget upper470 m of the SouthEquatorialCurrentin the eastern
of the Indian Ocean [Piola and Gordon, 1986], and Indian Ocean in October 1987, approximately9 Sv is
derived from the ITW. Fieux et al., 1994, 1996b find 18
1Lamont-Doherty
EarthObservatory
ofColumbia
University,Sv passing into the Indian Ocean between Java and
Australia in August 1989 and 2.6 Sv headedeastwardin
Palisades, N.Y.
February 1992. Bray et al., [1997] further examinesthe
2RSMAS,
Miami,FL.
Sea circulation using the WHP I10 data. Meyers
3School
of OceanandEarthSciences
andTechnology,Timor
[1996]; Fieux et al., 1996b; Gordon and Fine, 1996, all
University of Hawaii, Honolulu, Hawaii
suggest the throughflow is reduced during E1 Nifio
4Scripps
Institution
ofOceanography,
UCSD,
LaJolla
CA.
episodes([Meyers, 1996] finds a 5 Sv ENSO signalof the
5NOAA/AOML,
Miami,FL.
throughflow), when sea level standslower at the tropical
Pacific entrance to the Indonesian Seas, effectively
Copyright1997by theAmericanGeophysical
Union.
representinga northernwinter analog.
The percentageof ITW within the Indian Ocean SEC is
Papernumber97GL01061.
0094-8534/97/97 GL-01061 $05.00
determined by comparing the potential temperature,
2573
2574
GORDON ET AL.: ADVECTION
AND DIFFUSION OF INDONESIAN THROUGHFLOW
-20
salinity (T/S) data points measuredduring the WOCE
WHP observational period of 1995, with the 'pure'
Indonesia(Timor Sea) sourcewater T/S propertiesand the
principalwater massesof the Indian Ocean(Fig. 1). The
comparisonis made with a simple isopycnalmodel: an
observedT/S point at a WHP station is assumedto be
composed
of a linearadmixtureof theprincipalwatertype
valuesalong the sameisopycnal.This methoddisregards
vertical mixing which is a safe assumptionwithin the
highly stratified •hermocline;it is also noted that the
vertical processeswould tend to reducethe thermocline
salinitywithin the westwardflowing SEC, while isopycnal
mixing would act to increasesalinity,as observed.The
Indian Oceanprincipalwater massesare consideredto be
the southern Indian Ocean Central Water (ICW), and the
sharplycontrasting
surfacelayerwatersof theArabianSea
and Bay of Bengal. The Bay of Bengalsurfacewater may
pass into the throughflow by way of the South Java
Current,which curlsback to the west as it's watersjoin the
SEC. The ICW component is swept into the northern
Indian Ocean within the westernmargins of the tropical
Indian Oceanprovidingthe primary sourceof thermocline
within the northern Indian Ocean [ You and Tomczak, 1993;
You, 1996], thereforethe northernIndian thermoclinemay
be considered as a diluted (with Arabian Sea, Bay of
Bengaland ITW) form of ICW and not as an independent,
-15
-10
WATER
-5
0
0 ' .•I"''
.... •.•...... •'' ....
'"''' '
0
100
........
.-.-.'
'"" :"--•
..'................
:.:.:.'.
....
•:•..
................
•-i-:-i
...........
':, 100
200
300
400
200
300
400
•
100
100
•'
300
300
8
• 400
m
100
100
200
300
400
200
300
400
so8
so8
100
............
'..........
"":":':::
.........
'..............
%......
":":':"'•':'
1
•'
";'•,..•:!.::::•..
•/)
-.
200
300
?a
--,.
_
300
200 ='•
400.t,my
.... -
500
-20
-15
-10
-5
500
400
0
Latitude
Figure 2. Temperature, salinity, oxygen and percentage
IndonesianThroughflow Water along WOCE WHP sectionI9n
(seefig. 1 of Ffield, 1997, this issue).
principalwatermass.As the contrastbetweenthethroughflow and principalwater massesis small for water cooler
The I9n section,nominally along 95øE (Fig. 2) salinity
than 10øC this method cannot detail with acceptable divulge the presenceof the ITW by the low salinitywater
precisionthe throughflowpercentagebelow the thermo- columncenteredat 12øS.The sharpmeridionalgradientin
cline. Although the bulk of throughflow transport is thermoclineoxygen falls immediately southof the plume
believed to occur within the thermocline layer [Wyrtki,
1987; Gordon and Fine, 1996] and the SEC is weak below
the thermocline, the reader is cautionedthat further Pacific
of throughflow water, clearly showing the major frontal
zone associatedwith the plume. The high percentageof
throughflow
waterin theupper50 m southof 15øSis most
watermay enterthe Indian Oceanbelow the thermocline likely a consequenceof southward Ekman transport
characteristic of this region and the doming of the
[Fieux et al., 1996a].
thermocline
between
5 ø and 13øS denotes
the division
between southward and northward Ekman transport
[Trenberth et al., 1990]. Southward flux of freshwater
(southof 15øS)tendsto balancethe net evaporationof the
southIndian Oceansubtropicalregion. North of the plume
axis there is also high concentrationof throughflowwater.
This is Indonesian water embeddedwith the Equatorial
Counter Current (the ADCP reveals strong flow towards
the eastbetween6ø to 8øS), as part of the recirculationcell
discussedbelow. North of 5øS the low salinity water may
reflect local excess precipitation and Bay of Bengal
influence and is not includedin the throughflowtransport
values.
The salinity and percentageof the Indonesianthroughflow on a representativesigma-0surface(24.0, roughlythe
100 m depth, Fig. 3) reveal the westward advection and
dilution of the low salinity throughflow water within the
SEC. The break in the 34.6 isohaline near 95øE may
denote uneven injection of ITW or eddy variability in
meridional mixing process.The plume spreadszonally
across the Indian Ocean, with increased meridionai
Salinity
spreadingon reachingthe roughtopographybetweenthe
Figure 1. Potential TemperatureSalinity relationshipof the
Seychelles and Madagascar. The data suggeststhat the
IndonesianThroughflowWater andprincipalIndian OceanWater
20%
ITW concentrationcontour may pass southward of
masseswhich mix along isopycnalsurfaceswith the Indonesian
20øS
alongthe coastof Madagascar,but suchdetail at the
throughflow water. The IndonesianthroughflowT/S characteristicsare definedby dataobtainedin the Timor Sea (Ilahude lower concentrationlevels requiresa mixing model better
andGordon,1996).
suited to the western Indian Ocean T/S values.
GORDON ET AL.: ADVECTION
(a) 40
50
•
70
80
90
AND DIFFUSION OF INDONESIAN
I O0
:
IiO
•
35.3j ' '
:
..
o
L•I0
20
40
(b) 40
50
60
50
70
80
LONGITUDE
øE
60
7'0
90
I O0
II0
90
i
IO0
I I0
ß
2575
95øE
and 80øE.
The
ADCP
from
19 section
(95øE) doesshow strongsouthwardflow in the upper400
m across5øS [Hacker. personal communication,1997].
The existence of an equatorial recirculation is also
supportedby the 1990 JEXAM driftertracks[Michida and
Yoritaka,
1996] as well as the WOCE
Indian Ocean
drifters, and is the NRL model run forced by ECMWF
winds [Hacker, personal communication, 1997].
Additionally the zonal section12 along 8øSindicatesweak
northward geostrophic flow (upper 400 m) within the
ujIO,-•
;
_•o'----"•
-
i
___ .'•
:
40
WATER
seasonalvariationsin the injecting of Pacific water at the
Indian Oceaneasternboundary,thoughtransportsimilarity
of the 18 one time and repeat sectionsrepresentinga 6
monthdifferenceof injectiontime at the easternboundary,
arguesagainsta seasonalcycle origin. The introductionof
the Indonesianthroughflow into the SEC is expectedto
vary interannual (ENSO) scales, with the throughflow
relaxing during E1 Nifio episodes.However comparisonto
the southernoscillationindex doesnot reveala relationship
to the throughflowtransport.There may be changesin the
throughflowcomponentdue to changesin the subtropical
gyre circulation of the southern Indian Ocean or of a
recirculation cell within the equatorial current system,
which advects throughflow water southward across5øS
between
80
THROUGHFLOW
50
60
70
:
80
LONGITUDE
90
I O0
II0
*E
Figure 3. Distribution of salinity (3a) and percentage of
IndonesianThroughflowWater (3b) on the 24.0 Sigma-0surface,
roughly at 100-m depthß The dots show the positionsof the
WOCE WHP CTD stations used to construct the contour field.
thermocline of the western Indian Ocean, with southward
transportin the east. Thus a zonally elongatedcirculation
cell formed betweenthe SEC and Equatorialcurrentsmay
recirculatethroughflowwater between80øEand 90øE.
Isopycnal Mixing
Attenuation of the low salinity signal of the throughflow
water as it spreadstowardsthe westernIndian Oceanis the
consequenceof lateral mixing. Vertical mixing would not
lead to a reversesignal,as the salinityis low in the surface
water
Geostrophic Transports
The net westwardgeostrophicvelocity field between5ø
to 20øS (6.5øS to 20øS for WHP 17, to avoid the shallow
platform of the Seychelles;Table I) which includesthe
SEC, is determinedby two referencemethods:a 3000 db
zero velocity and the ADCP measuredvelocity perpendicular to the WHP section, averagedbetween stationsat
200 m (below the Ekman layer). The 3000 db zero
referencevaluesmay provide for the more stable,climatic
values;the ADCP representsthe transportfield at the time
of measurements,
includingthe high frequencyand nongeostrophicfeatures. The net westward flow of the SEC
increasestowards the west, presumablyas more water
joins the SEC from the north of 5øS and south of 20øS,
from 11 to 15 Sv across95øE to 23 to 36 Sv across55øE,
in agreementwith the findings of Donguy and Meyers
[1995] basedon XBT analysis.
The contribution of the ITW to the SEC transportis
determinedby combiningthe percentageof throughflow
water with the geostrophictransportvalues. The throughflow transportacross95øE is 4 to 6 Sv, increasingto 11 to
12Svat 80øE.Dropping
to 6 to 10Svat55øEwhichmay
be due to meridional loss of throughflow water on
proceeding westward. The variability may also reflect
and in the intermediate
water
at the base of the
thermocline.Fitting a paraboliccurveto the salinityversus
latitude along isopycnalscharacteristicof the upper400 m
and balancingthe meridionaleddy diffusion againstzonal
advection(using a mean characteristicgeostrophiczonal
flow for the upper 400-m of 10 cm,/sec) requires an
isopycnalmixing coefficientof 1.1 to 1.6 x 104 m2sec
-1
(Table I). Independentestimatesof the diffusivitiesin the
Indonesianplume are derivedfrom the WOCE drifter array
using Taylor's [1921] single particle theory. The WOCE
Indian Ocean surface drifter array over the range of the
plumegivesdiffusivitiesthatrangefrom0.4 to 2.4 X 104
m2sec
4 the meridionaldirectionwith an arealaverageof
1.3 X 104m2sec4.
This value should decrease somewhat
with depthin proportionto thedecrease
in velocitybut will
be augmentedby sheardispersionin the horizontal[Rhines
and Young, 1982; Dewar and Flierl, 1985]. These Ky
valuesmay be considered
to be a ratherlargenumber,but
perhapsnotunreasonable
in the stronghorizontalshearsof
the equatorialcurrentsystem.
,
Acknowledgments. The grant support for the authors is
derived from: Gordon and Ma NSF grant OCE-94-13171,
Lamont-DohertyEarth ObservatoryContributionNumber 5636;
Ffield NOAA grantNA66GP0267; Hacker NSF grantOCE-9413172, Contribution Number 4480; Tally NSF grant OCE-9413160; Olson NSF grantsOCE-94-13165 and OCE-93-14447;
Wilson and Baringer sponsoredby NOAA's Office of Global
Programs.WOCE ContributionNumber 540.
2576
GORDON ET AL.: ADVECTION
AND DIFFUSION
OF INDONESIAN
THROUGHFLOW
WATER
Table1. Transport
in Sv(106 m3/sec)
of theSouth
Equatorial
Current
(SEC)andof theIndonesian
throughflow
component
within the SEC acrossmeridionalsectionsobtainedby the WOCE IndianOceanExpeditionof 1995.The transportvalues
are from 5øSto 20øS(6.5øSto 20øSfor I7 andI7r to avoidSeychelles),
for theupper400-m.
WHP Section Timeof data
TimeatTimor SECtransportSECtransportthroughflow throughflow Ky (m2s
-1)
Number
Sea origin
(SEC=10cm/
sec)
collection
I
I
J
,
3000 db
(upper 400m)
I
200 ADCP
(upper400m)
I
transport
3000 db
(upper 400m)
'
transport
200 ADCP
(upper 400m)
(SEC=
10cm/s)
l
19
I8
I8 repeat
I7
February'95
March'95
October'95
August'95
December
'93
September
'93
March'94
February'93
15
17
27
23
ll
15
30
36
6
12
11
6
4
12
11
9
1.31 x
1.56 x
1.50 x
1.06 x
I7 repeat
April '95
October'92 25
30
8
10
1.06x 104
References
Hautala, S., J. Reid, and N. Bray, The distribution and
mixing of Pacific water massesin the IndonesianSeas,
Arief, D., and S. Murray, Low frequencyfluctuationsin
the IndonesianthroughflowthroughLombok Strait, J.
Geophys.Res., 101(C5), 12455-12464,1996.
Bray, N.A., S.E. Wijffels, J.C. Chong, M. Fieux, S.
Hautala, G. Meyers, and W.M.L. Morawitz,
Characteristicsof the Indo-Pacific throughflow in the
eastern Indian Ocean, GRL WOCE Indian Ocean Issue,
Geophys.Res. Lett., this issue,1997.
Conkright,M., S. Levitus,T. Boyer,D. Bartolacci,andM.
Luther,
104
104
104
104
Atlas
of the northern Indian
Ocean,
J. Geop.Res., 101(C5), 12375-12389,1996.
Ilahude,
A. G.,
and A. L. Gordon,
Thermocline
StratificationWithin the IndonesianSeas,J. Geophys.
Res., 101(C5), 12401-12409, 1996.
Meyers, G., Variations of indonesianthroughflowand E1
Nifio-SouthernOscillation, J. Geophys.Res., 101(C5),
12255-12263, 1996.
Meyers, G., R. Bailey, and A. Worby, Volume transportof
Indonesianthroughflow,Deep-SeaRes., 42(7), 11631174, 1995.
NOAA/NODC,Washintgon,DC and Univ. of South Michida, Y., and H. Yoritaka, Surface currents in the area
Florida, St. Petersburg,FL, 1994
Dewar, W. K., and G. R. Flierl, Particletrajectoriesand
simplemodelsof transportin coherentvortices, Dyn.
of the Indo-Pacific throughflowin the tropicalIndian
Ocean observedwith surfacedrifters, J. Geop. Res.,
101(C5), 12475-12482, 1996.
Piola, A., and A. Gordon, On Oceanic Heat and Fresh-
Atmos. Oceans, 9, 215-252, 1985.
Donguy, J., and G. Meyers, Observationsof geostrophic
transportin the westerntropicalIndian Ocean,Deep-
water
Fluxes
at30øS, J.Phys.
Ocean.,
16(12),21842190, 1986.
SeaRes., 42(6),1007-1028, 1995.
Quadfasel, D., A. Frische, and G. Cresswell, The
Ffield, A.. Introduction to the GRL WOCE Indian Ocean
circulation in the source area of the South Equatorial
Issue,Geophys.Res.Letters,in press,1997.
Current in the easternIndian Ocean, J. Geophys.Res.,
Ffield, A., and A. Gordon, Vertical mixing in the
101(C5), 12483-12488, 1996.
IndonesianThermocline,J. Phys.Oceanog.,22(2), 184- Rhine, P. B., and W. R. Young, Homogenization of
195, 1992.
potentialvorticity in planetarygyres, J. Fluid Mech.,
Fieux, M. R. Molcard, and A. G. Ilahude, Geostrophic
122, 347-367, 1982.
transportof the Pacific-IndianOceansthroughflow,J. Taylor, G.I., Diffusion by continuousmovements, Proc.
Geophys.Res., 101(C5), 12421-12432,1996b.
London Math. Soc., Ser. 2, Vol. 20, 196-212, 1921.
Fieux, M., C. Andrie, E. Charriaud, A. G. Ilahude, N.
Trenberth, K., W Large, and J. Olson, The mean annual
Metzl, R. Molcard, and J. Swallow, Hydrological and
cycle in global oceanwind stress,J. Phys. Oceanog.,
chlorofloromethane
20(11), 1742-1760, 1990
measurements
of the Indonesian
throughflowenteringthe Indian Ocean,J. Geop.Res., Wyrtki, K., Indonesianthroughflow and the associated
101(C5), 12433-12454, 1996a.
pressuregradient, J. Geophys.Res., 92, 12941-12946,
Fieux, M., C. Andri•, P. Delecluse, A. G. Ilahude, A.
Kartavtseff, F. Mantisi, R. Molcard, and J. C, Swallow,
Measurements
within
the
Pacific-Indian
Ocean
1987.
You, Y.,
and M. Tomczak, Thermocline circulation and
ventilation
in the Indian Ocean derived from water mass
ThroughflowRegion, Deep Sea Res., 41(7), 1091-
analysis, Deep-SeaRes..,40(1), 13-56, 1993.
1130, 1994.
You, Y. Dianeutralmixing in the thermoclineof the Indian
Gordon, A. L., and R.A. Fine, Pathwaysof water between
Ocean.Deep-SeaRes. 43(3):291-320, 1996.
the Pacific and Indian Oceans in the Indonesian Seas,
Nature, 379, 146-149, 1996.
Gordon, A. L., Interocean Exchange of Thermocline (ReceivedJanuary28, 1997;revisedApril 7, 1997;
acceptedMarch 24, 1997.)
Water,J. Geophys.Res., 91(C4), 5037-5047,1986.