GEOPHYSICAL RESEARCH LETTERS, VOL. 25, NO.15, PAGES 2769-2772, AUGUST 1, 1998
Bay of Bengal currents during the northeast monsoon
Peter Hacker, Eric Firing, and Julia Hummon
University of Hawaii, Honolulu, Hawaii
Arnold
L. Gordon
Lamont-Doherty Earth Observatory of Columbia University, Palisades,New York
J. C. Kindle
Naval Research Laboratory, Stennis Space Center, Mississippi
Abstract. Velocityand propertyobservations
weremade
An introductionto the richlegacyof numericalmodeling
duringFebruaryandMarch1995aspart ofthe WorldOcean studiesof the Indian Oceancan be foundin McCrearyet
CirculationExperiment(WOCE) Hydrographic
Program al. [1993]and Vinayachandran
andYarnagata
[1998],which
(WHP) expedition
in theIndianOcean.Theobserved
circu- focuses
onthe Bayof Bengalsector.
lation in the upper 300 m of the oceanduring the northeast
monsoon
is compared
to theoutput
ofa high-resolution,
Measurementsand 1ViodelDescription
3-layer, nonlinear model forced by European Center for
MediumRangeWeatherForecasting(ECMWF) winds. The
data identify severalnew featuresin the Bay of Bengal: a
shelf-break coastal current in the northeast corner, an eddy
and subsurfacejet near the South Preparis Channel, and an
eastwardcountercurrentextending from 80øE to the eastern
boundary. The countercurrenttransports high-salinitysurface water from the northwest Indian Ocean into the bay,
and separatesthe historically observedNorth Equatorial
Current into two currents with separate sourcesand water properties. The data also detail the spatial structure
of the South Equatorial Current and Countercurrent, and
showthe Equatorial Undercurrentin the easternhalf of the
Indian Ocean. The model compareswell enough with the
data to suggestthat such realistic modelsmay provide a
usefultemporal context for the WHP snapshot.
The
ADCP
and CTD
data
were taken
on two consec-
utive legsaboardthe R/V Knorr between6 February and
21 March, 1995(Figurela). SeeF•eld[1997] for an overview
of the WOCE Indian Ocean expedition.
The hull-mounted
ADCP
was a 150 kHz unit
manufac-
tured by RD Instruments. Accurate position and heading
data wereprovidedby a Trimble P/Y-code and an Ashtech
3DF (four antennae)GPS receiver,respectively.Errors in
the absolutevelocityestimatesare lessthan 0.02 m s-• for
1-hour averages(about 20 km alongthe track when underway). The velocitywasmeasuredfrom the shallowestdepth
of 21 m to about 300 m. During the leg from the northeast
corner of the Bay of Bengal to Colombo, the ADCP system began to fail, resulting in reduced depth range. After
repair in port and testing at sea, normal ADCP operation
was resumed at 2.8øN on I8N southbound.
The modelresults(Figure lb) are from a high resolution
simulationusingthe NavalResearchLaboratory(NRL) non-
Introduction
The Bay of Bengal is a regionof large freshwaterinput,
high sea-surfacetemperature, and variablemonsoonalforcing. Althoughhistoricaldata are sufficientto identifymajor
featuresof the annual cycleof properties[Conkrightet al.,
linear reduced-gravityIndian Ocean model with three active
layers in the upper 500 m overlying an infinitely deep, quies-
centfourthlayer[Bruceet al., 1998]. The basin-wideaverage
thickness for the upper layer is 115 m. The model domain
1994]and circulation[Cutler and Swallow,1984; Molinari extendsover the entire Indian Ocean north of 30øS,with a
et al., 1990]both within the bay and in the equatorialre- grid resolution of 0.25ø in latitude and 0.37ø in longitude.
gion to the south, the bay itself remainspoorly observed, The model uses a high resolution basin geometry defined
especiallyin the easternregion. Legeckis[1988]discusses by the 200-meter isobath, which resolvesthe three primary
the western boundary current using satellite observations. passagesbetween the Andaman Sea and the Bay of BenRecentfield data [Sheryeet al., 1996]providean improved gal. The simulation was forced by the 12-hourly 1000 mbar
picture in the western bay during the northeastmonsoon, ECMWF winds for the period 1990 through 1995, following
which peaksduring January to March. During this period a 35-yearspinupusingthe Hellermanand Rosenstein[1983]
the anticycloniccirculationof the upperoceanintensifiesin wind stressclimatology.
the northern bay; to the south, the North Equatorial Current (NEC), the SouthEquatorialCountercurrent
(SECC), Results and Discussion
andthe SouthEquatorialCurrent (SEC) arewell developed.
The most striking characteristic of the upper ocean velocity stucture is the presenceof intensejet and eddy circulation patterns. The following describesthe major features.
The most northerly station within the Bay of Bengal was
on the continental shelf at a water depth of 107 m. Currents
were weak. Just offshoreof the shelf break, between 19.0øN
Copyright1998by theAmericanGeophysical
Union.
Papernumber98GL52115.
0094-8534/98/98GL-52115505.00
2769
2770
HACKER ET AL.: BAY OF BENGAL CURRENTS, NORTHEAST
C U R RENT
LAYER = I
MONSOON
9 DAY AVERAGE
o.5.,."
70021:3:55.0
DATE = 058 / 1995
>
10'N
ß
EC
.
.•
SECC
lo'$
SEC
15'S
75'E
80'E
85'E
90'E
95'E
100'E
75E
80E
85E
90E
95E
100E
105E
110E
Figure 1. Near-surfacecurrents:(a) ADCP velocityvectorsaveragedbetween25-75m; (b) NRL modelvelocitytrajectories for layer 1, nominally 0-115 meters.
From 9øN to 10øN at 91øE there is an eastward flow
and 19.7øN, the current was to the east (Figure 2) with
near-surface
speedover0.4 rn s-Zanda transportof about of 0.4 rn s-Zwith the high salinitysignatureof northwest
i Sv (1 Sv -- 106mSs
-z) to theeast.Aspointed
out byan Indian Ocean water. It may be entering the Andaman Sea
anonymousreviewer,the current may be part of the annual
or it may be a recirculating eddy on the northern edge of
cyclealongthe easternboundary[McCrearyet al., 1993].
the NEC.
The NRL model run shows the current developing about
The NEC is apparentat three longitudes(Figure1) with
12 February and continuingthrough 5 March, and suggests transport in the upper 100 rn ranging from 8-10 Sv consistent
that the feature is directly forced and deterministic.
with resultsof [$chottet al., 1994]. At 91øEthe variation
A striking feature, the Preparis Eddy, is centered at of near-surfacesalinity with latitude suggestsrecirculation
14.7øN, 91.5øE, just west of the South Preparis Channel eddies on both the north and south sides of the NEC. On
(Figure 3). The eddy is subsurfaceintensifiedwith a maxi- all three sectionsthe NEC has higher salinity (34.5 psu)
mumspeedof 0.8 rn s-• at 140m; hasa diameterof 300km
along its southern edge, representingthe northwest Indian
at 150 m; and is associatedwith a larger surfacegyre in the
top 100 m, with a diameter of nearly 700 kin. The transport in the upper 100 rn associatedwith the larger-scale
gyre is 6 Sv. The transport between100-200 rn depth associated with the subsurfaceeddy is 3 Sv for the northern
half and 6 Sv for the southern half. A possible explanation couldbe a jet flowing out of the Andaman Sea through
the South Preparis Channel, which has a sill depth about
400 m. The south edge of the westward flowing jet has
a salinity anomaly of-0.60 psu and an oxygenanomaly of
Oceansource(Figure 4). South of Sri Lanka the salinity
is reduced(33.0-33.5 psu) alongthe NEC's northern half
representinga mixture of the fresher Bay of Bengal water
with
the northwest
Indian
Ocean water.
Our observations show eastward flow, which we tenta-
tively label the North Equatorial Countercurrent(NECC),
extendingfrom 80øE to 92øE. The current transports a local
surfacesalinitymaximum(34.5 psu)alongits northernedge,
indicatingits northwestIndian Oceansource(Figure4), and
intensifiesto the east. Both the northern and southernparts
-I-15.4/•molkg-• at 17.20øCanddepth140m, suggestingof
the current at 92øE seem to be associated with recircu-
lation eddies as shown by the velocity vectors and the surface salinity distribution. Part of the NECC may enter the
Andaman Sea and part must retrofiect into the NEC east of
a sourceother than the northeast Bay of Bengal. The NRL
model showsa lessintense feature spinning up in layer i on
24 February and continuing until mid-March 1995. For the
5-year period, 1991-1995, the eddy appeared in 1992, 1993
our observations.
and 1995. In 1992 and 1993 it was related to a sub-surface
goesalongthe northernedgeof the NECC. Along 80øE be-
The section from 85øE to 80øE near 3.5øN
jet and was sub-surfaceintensified. However,in 1995 it was tween 2.8øN and 2øN the ADCP shows eastward flow near
surface-trappedand was not associatedwith a sub-surface the surface.This eastwardflow (the NECC) broadenswith
depth and extendsfrom 1.5øNto 4øN at 100 m. Note that
jet. The eddy does not seemto be a simple annual event.
HACKER ET AL.: BAY OF BENGAL CURRENTS, NORTHEAST MONSOON
2771
Temperature
100
200
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
Salinity
100
,oo
i
i
i
Zonal Velocity
0
..
...
i
.....
!00•
200
300
-
i
15 øS
i
i
i
i
10 'S
i
5 ON
EQ
10 ON
••OC
,
! 5 ON
20 ON
Figure 2. Propertiesalongthe I9N easternline: temperature,salinity,and u componentof velocity(positiveeastward,
cm/s). Shadedregionsare negative,and the contourintervalis 10 cm/s.
the NECC is very narrow at 80øE comparedwith the eastern side of the basin. At 80øE it may be an occasional
feature pinched off to the south of Sri Lanka between the
NEC and the Equatorial Current (EC). As an alternative
interpretation, the NECC at 92øE could be simply part of
an intense eddy, disconnectedfrom the weak eastwardflow
west of 85øE. The high surfacesalinity would then be attributed to upwelling and diapycnal mixing in the core of
the eddy. While the NRL model does not show the NECC
To the south of the EC we observed the eastward
SECC
and the westward SEC. The transports of both are 50%
higherat 80øE than at 95øE (Figure 4). Our measurements
add spatial resolution to the historical picture developed
by Cutler and Swallow[1984],Molinari et al. [1990],and
Conkrightet al. [1994]for this season. The surfaceve-
at 80øE, it does produce a weak countercurrenttrough with
25m to 75m
100m to 150m
eastward currents between 3øN and 5øN and from 83øE to
the eastern boundary, indicating a tendency for the NEC
20'N
(era/s)
to retroflect to the east south of Sri Lanka, consistent with
modelcirculationin Vinayachandran
and Yamagata[1998].
18'N
To the south of the NECC, we observedthe EC, with a
westwardtransport above100 m of 5 Sv at 80øEand ? Sv
16'N
at 91øE. The surface salinity is consistentlylower in the
east and higherin the west. Cutler and Swallow[1984]and
Molinari et al. [1990]showthe EC, but do not distinguish
14'N
it from the NEC. The NRL model representsthe EC as a
feature associated with the retroflection
of SECC flow across
the equator (Figure lb), a finding also suggestedby the
drifter tracksshownin 3/Iolinariet al. [1990].
/
12'N
Below the EC we observedthe eastward flowing Equato-
rial Undercurrent(EUC) alsonotedin Talley and Baringer
[1997].At 80øE and 93øEthe EUC extendsoverthe depth
/
10'N
I
88'E
90'E
I
92'E
88'E
a
I
90'E
,
I
92'E
,
94'E
range60-140m, haspeakspeedof0.4m s-x anda transport
of i Sv. At 93øE the EUC is shifted north of the Equator,
Figure 3. Layer-averagedvelocity in the easternBay of
extendingfrom 0.7øN to 2.3øN.
Bengal, showingthe Preparis Eddy.
2772
HACKER ET AL.: BAY OF BENGAL CURRENTS, NORTHEAST MONSOON
western leg: ADCP
Acknowledgments.
Thanks to the officers,crew and scientific party aboardthe R/V Knorr for their professionalismand
eastern leg ADCP
help with data collection. CTD data were collected and proceasedby the Scripps Oceanographic Data Facility. This work
wasfundedunderthe followingNSF grants: OCE-9413172 (UH),
OCE-9413171(Columbia),and OCE-9413160(UCSD). NRL par-
NEC
[ NCC
•'
ticipation was provided through the ARI "Forced Upper Ocean
Dynamics", which is funded by the Office of Naval Research. The
numerical simulations were performed through a grant from the
DOD High Performance Computing Initiative. SOEST contribution
number
4651.
References
Bruce, J.G., J.C. Kindle, L.H. Kantha, J.L. Kerling, and J.F. Bailey, Recent observationsand modeling in the Arabian Sea Laccadive High region, J. Geophys.Res., 103, 7593-7600, 1998.
Conkright, M.E., S. Levitus, T.P. Boyer, D.M. Bartolacci,
and M.E. Luther, Atlas of the Northern Indian Ocean,
-5
-10
NOAA/NODC,
-15
(,a) , ,
32 33 34 3'5 -20
psu
-10
0
transport(Sv)
,(c)
-20 -10
0
Washington DC and University of South
Florida, St. Petersburg, Florida, 1994.
Cutler, A.N. and J.C. Swallow, Surface currents of the Indian
Ocean(to 25øS,100øE),Inst. of Oceanogr.Sci., Wormley,Eng-
transport(Sv)
Figure 4. (a) Averagesalinity(psu) from 10-40 m from
the westernleg (circles,endof I9N andall of I8N) andeastern leg(dotted,easternpart of I9N). ADCP transport(Sv)
integratedfrom the southfor the (b) westernleg and (c)
easternleg, assumingconstantvelocity from 25 m to the
surface,and linearlyinterpolatingacrossmissingdata (small
circles).
land, 1984.
Ffield, A., GRL special section: WOCE Indian Ocean expedition,
Geophys.Res. Letters, ϥ, 2539-2540, 1997.
Gordon, A.L., S. Ma, D.B. Olson, P. Hacker, A. Ffield, L.D. Tal-
ley, D. Wilson, and M. Baringer, Advection and diffusion of
Indonesian throughflow water within the Indian Ocean South
Equatorial Current, Geophys. Res. Letters, ϥ, 2573-2576,
1997.
Hellerman, S. and M. Rosenstein, Normal monthly wind
stress over the world ocean with error estimates, J. Phys.
Oceanogr., 13, 1093-1104, 1983.
Legeckis,R., Satellite observationsof a western boundary current
in the Bay of Bengal, J. Geophys.Res., 9œ,12974-12978, 1988.
McCreary, J.P., P.K. Kundu, and R.L. Molinari, A numerical
investigation of dynamics, thermodynamics and mixed-layer
processesin the Indian Ocean, Prog. Oceanogr., 31, 181-244,
ß
locity and property distributionssuggestrecirculationeddies at the northern and southern boundaries of the SECC
and SEC. In the region south of the Equator, the NRL
1993.
model showslocal eddies and elongated recirculation featureswhichappearconsistentwith our velocities.The effect Molinari, R.L., D. Olson, and G. Reverdin, Surface current distributions in the tropical Indian Ocean derived from compilations
of these features on the IndonesianThroughflow Plume are
discussed
in Gordonet al. [1997].
Maps of the near-surface
temperatureand salinitydistributionswithin the Bay of Bengal appear in Conkrightet al.
[1994].The mapsdonotresolve
the effectsofthe NECC and
of surface buoy trajectories, J. Geophys. Res., 95, 7217-7238,
1990.
Shetye, S.R., A.D. Gouveia, D. Shankar, S.S.C. Shenot,
P.N. Vinayachandran, D. Sundar, G.S. Michael, and G. Nampoothiri, Hydrography and circulation in the western Bay of
Bengal during the northeast monsoon, J. Geophys.Res., 101,
the EC on the property distributionsdue to smoothingand
14011-14025, 1996.
a lack of historical high resolutionobservations.Our data Schott, F., J. Reppin, and J. Fisher, Currents and transports of
providedetailsof the vigorouscirculationpathwayswhich
the Monsoon Current south of Sri Lanka, J. Geophys. Res., 99,
bringhighsalinitywater into closeproximitywith the fresh,
25127-25141, 1994.
surfacewater within the bay, and suggestthat the stirring Talley, L.D. and M.O. Baringer, Preliminary resultsfrom WOCE
processes
occurin veryshallowlayers(Figure2).
In the mean model solutionover the cruiseperiod (not
shown),the NEC and EC are fedby cross-equatorial,
retrofiectedtransport from the SECC. The circulationshownin
Figure lb containsthe essentialelementsof the mean pattern. The model EC's latitude range and the flow south
of Sri Lanka differ somewhat from the observations.
How-
ever, given the quasi-synopticnature of the sampling,the
limitationsof the model (particularlythe vertical and horizontalresolution),the uncertaintyin the modelforcingand
the high variability in this region, the comparisonof the
observations and the model results is encouraging.
hydrographicsectionsat 80øE and 32øS in the Central Indian
Ocean, Geophys. Res. Letters, ϥ, 2789-2792, 1997.
Vinayachandran, P.N. and T. Yamagata, Monsoon responseof
the sea around
Sri Lanka:
Generation
of thermal
domes and
anticyclonic vortices, J. Phys. Oceanogr., in press, 1998.
P. Hacker, E. Firing, and J. Hummon, JIMAR, 1000 Pope Rd.,
Honolulu,HI 96822. (e-mail: [email protected])
A. Gordon, Lamont-Doherty Earth Observatory of Columbia,
Palisades,NY, 10964-8000.(e-mail: [email protected].
edu)
J. C. Kindle, Naval ResearchCenter, Stennis Space Center,
Mississippi,
39529.(e-mail:kindle@polaris.
nrlssc.
navy.mil)
An observationalgoal for the future is to determinethe (ReceivedJuly 16, 1997; revisedMarch 3, 1998;
temporalvariability of the observedfeaturesand pathways. acceptedJune22, 1998.)