JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 102, NO. C6, PAGES 12,749-12,784,JUNE 15, 1997
Upper ocean thermohaline fields near 2øS, 156øE,
during the Tropical Ocean-Global AtmosphereCoupled Ocean-Atmosphere Response Experiment,
November 1992 to February 1993
Adriana Huyer and P. Michael Kosro
College of Oceanic and Atmospheric Sciences,Oregon State University, Corvallis
Roger Lukas and Peter Hacker
School of Ocean and Earth Science Technology, University of Hawaii, Honolulu
Abstract. Zonal and meridional Seasoarsectionscenteredat 1ø50•S,156ø06•E
were repeated >30 times in three 20-day periods between Nove.mber 13, 1992, and
February 15, 1993. Both sectionswere 130 km long, and sampling depth was 0-
280m, witha verticalresolution
of •-2 dbar(2 x 104Pa) anda horizontal
resolution
of •-3 km. The observed fields are complex and variable and are summarized
graphically in several forms. Characteristicsof the near-surfacelayer varied with
the localwindswhichwerevariableandweak(< 6 m s-1) duringthe first20-day
period,strongandwesterly(>10 m s-1) duringmuchof the second,
andmoderate
andwesterly(4-10 m s-1) duringmostof the third. Near-surface
temperatures
werewarmest(up to 30øC)duringthe first periodand coldest(often <29øC)
during the second.Thermal stratificationin the near-surfacelayer was strongunder
weak winds and weak under strong and moderate winds. Except during and after
heavy rainfall, salinity stratification in the near-surfacelayer was generally weak.
Surface salinity generally decreasedtoward the north. The depth of the surface
isopycnallayer was often but not alwayslimited by salinity stratification; the surface
isohalinelayer was shallowerthan the top of the thermoclinethroughout. Strong
lateral temperature and salinity gradients occurred on a few occasions. Increased
wind stresswas associatedwith lateral homogenizationas well as vertical mixing.
Structure and water properties of the thermocline also varied between cruisesand
within each cruise. The upper thermocline was shallowestin late January after
prolongedeasterlywinds. Isothermsin the upperand midthermocline(20ø-25øC)
slopedgenerallyupwardto the north, while thosein the lowerthermocline(12ø14øC) slopeddown to envelopthe core of the Equatorial Undercurrent,which
shoaled(from 225 to 160 m) and warmed(from 15øto20øC) betweenthe first
and last surveyperiods. Mesoscaleand fine-scalewater massfeatureswere usually
recognizablein sectionsless than a few days apart and migrated eastward at
about0.3 m s-1. Thereis a remarkably
highdegreeof nonstationarity
in these
thermohaline
fields from the Warm
Pool of the western Pacific Ocean.
1. Introduction
periment (COARE), conductedas part of the Tropical Ocean-GlobalAtmosphere(TOGA) program,was
Progressin understandingocean-atmosphere
interacdesignedto fill this vacuum. The program included
tion in the warm pool of the westernequatorial Pacific
Ocean had been hindered by insufficientknowledgeof
the variability of both atmosphereand ocean over suitably small spacescalesand timescales,and a large international Coupled Ocean-AtmosphereResponseExCopyright 1997 by the American GeophysicalUnion.
complementarysampling of both atmosphereand upper ocean over a four-month intensiveobservationpe-
riod (IOP) lastingfrom November1992throughFebru-
ary 1993 [Websterand Lukas,1992]. Both oceanographic and meteorologicalobservationswere concen-
trated in an intensiveflux array (I'FA) centeredat
Paper number 97JC00267.
1ø45•S,156ø00•E[Websterand Lukas, 1992; Parsons
et al., 1994]. The oceanographic
componentincluded
0148-0227/97/9 7JC-0267509.00
studies of ocean mixing, measurementsfrom surface
12,749
12,750
HUYER ET AL' THERMOHALINE FIELDS NEAR 2øS, 156øE
and subsurfacemoorings,and both current and hydro- Table 1. Positionsof Apices and Intersectionof the
graphic observationsfrom severalstationary shipsand Standard Butterly Pattern Used for SeasoarSectionsin
a few surveyships[Moumand Caldwell,1994;Webster the COARE Intensive Flux Array Throughout Most of
the Three COARE SurveyCruises,W9211A (November
and Lukas, 1992]. As part of COARE, we made fre- 8 to December8, 1992), W9211B (December12 to Janquently repeated high-resolutionhydrographicobserva- uary 16, 1993), and W9211C (January22 to February
tions duringthree month-longcruisesof the R/V We- 22, 1993)
comato study the horizontal structure of time-varying
Latitude
Longitude
upper ocean thermohaline and velocity fields at the Way point
center of the IFA; these surveyswere designedto resolve zonal and meridional
SBN
SBS
SBW
SBE
I
scales of 10 to 100 km and
timescalesof a few days to a few months. Results de-
scribedin this papershowthat temperatureand salinity
fields were remarkably nonstationary, both in the sur-
1ø14•S
2ø26•S
1ø50•S
1ø50•S
1o50 •S
156ø06•E
156ø06•E
155ø30•E
156ø42•E
156 ø06•E
face layer and in the main thermocline.
Each of the three Wecomasurveycruises(W9211A,
W9211B, and W9211C) includedmeasurements
of the Figure 2' a rather thick (>30 m) layer of very warm
temperature, salinity, and velocity distribution in the (•28øC) surfacewater extendingmany degreesnorth
upper 300 m of the ocean and continuousmeteorolog- and south of the equator; isothermsin the thermocline
ical measurementsof wind, air temperature, humidity, (•-14ø-23øC)spreadingapart within •-2øofthe equator,
etc. Sampling focusedon frequent repetition of a Stan- away from the core (at •-200-m depth) of maximum
dard Butterfly Pattern consistingof a meridional sec- eastwardflow in the Equatorial Undercurrent(EUC);
tion along 156ø06tWand a zonal sectionalong1ø50tS, and a subsurfacetongue of hi•h-salinity South Pacific
each 130 km long, connectedby diagonalsectionsin water tending northward acrossthe equator at a depth
the southwesternand northeastern quadrants of the of •-150 m, with maximum salinity decreasing from
•35.6 psu at 5øS to •-34.8 psu at •-5øN and interleavCOARE IFA (Figure 1, Table 1).
The centerof our samplingpattern and of the COARE ing layers extending meridionally for •100 km. The
IFA, at • 1ø45tS,lies about 160 km off the equator. typical width of the EUC seemsto be about 200 km,
This distance is less than the equatorial radius of de- typicalcorevelocitiesare 0.3-0.5 m s-1 andthe center
formationfor first-modebaroclinicperturbations[Gill, of the coreusuallyliesbetweenløN and løS [Tsuchiya
1982],and thus our repeatedsectionslie on the south- et al., 1989; Richards and Pollard, 1991; McPhaden
ern flank of equatorial phenomenasuchas the Equato- et al., 1992; Eldin et al., 1994]. The strengthof the
rial Undercurrent
and internal Kelvin waves. Historical
EUC varies from <0.1 to •0.6
m s-1 on timescales
observations
in this region[Tooleet al., 1988;Tsuchiya of months[McPhadenet al., 1990, 1992; McPhaden,
et al., 1989; Richards and Pollard, 1991; McPhaden et 1993]. Our observations
are consistentwith this deal., 1992]andthe averageof 18 cross-equatorial
sections scription
o•theEUC' eastward
velocities
of •0.3 m s-1
madeduringCOARE [Eldin et al., 1994]all showthe in a broad core between IøN and 1øS in early Decemsame basic structure as long sectionsmade by Wecoma ber and •0.6 m s-1 in an apparentlynarrowcorenear
1ON
• Federated States of Micronesia
•
;
[iISS-Vessel
ISS Shore
•1 • .............
•k
I / Chuuk
ß
I/
Pohnpei
Rawinsonde
Site
-
IUETMooring
10øS
SBN
PCM-N
o
ELSl
•
/
o
Kapingamarangi
/ .......
.....
/
1 5ø5
.......
_
IMET
o
Manus
Island
•.• Ka,,ieng••
• .
•
SBE
20øS
_
Papua
NewGuinea •-J
T{
•---.
;•
•
•
,
•
•.
"'
•
•
Solomon
• •
•-7 •
/s/ands
•
H•niara
SBS
2 5øS
10s
,
140E
150E
•),..
Mlslma
I
,
160E
170E
155.OøE
,
I
•
155.5øE
,
I
,
156.0øE
,
I
I
156.5øE
Figure 1. Locationof the StandardButterflyPatternwithinthe COAREoutersounding
array
(dashed
lines,left panel)andthe intensive
fluxarray(solidlines,left panel),andits relationto
the moorings
nearthe centerof the COAREintensive
fluxarray(rightpanel).
157.0øE
HUYER ET AL.' THERMOHALINE FIELDS NEAR 2øS,156øE
12,751
lOO
15o
200
30oO
100
150
200
250
30oO
5O
100
150
200
25O
500
5"N
5øS
2øS
1øS
0ø
1øN
2øN
3"N
4øN
Figure2. (top)Temperature,
(middle)
salinity,
and(bottom)
zonal
velocity
fromlongSeasoar/ADCP
sections
along
156.1øE:
(left)December
1-3and(right)
January
10-12.Shading
indicates
temperatures
of14ø-15øC
and20ø-21øC;
salinities
of>35.2(dark)
and35.1-35.2
(light),
andwestward
flow(light,
0.0to -0.4 ms-1, dark,<0.4ms-1).
1ø45•Sin January(Figure 2). Abovethe EUC, there similarto the one observedduringCOARE (Figure 2)
is usually a broader layer of westwardflow with core [Eldin et al., 1994];theselow salinitiesseemto be asvelocitiesof •0.3 m s-1; this westwardflowsometimes sociated with eastward surface currents and westerly
hasa maximumat a depthof •125 m [Tsuchiyaet al., winds[McPhadenet al., 1992].Equatorialsurfacecur1989;McPhadenet al., 1992]asit did in early Decem- rents are highly correlated with local winds, at least
ber (Figure2), and sometimes
extendsto the surface during periodsof prevailingwesterlywinds;an eastward
[Tsuchiya
etal.,1989;
Richard}
andPollard,
1991]
as jet-likecurrent("Yoshidajet") with a width of •200 km
it did in January (Figure 2). Westwardvelocitiesof and core velocity of •1.0 m s- developsrapidly after
>0.4 m s-1 at the core(Figure2) arewithinthe range the onsetof westerlywinds [McPhadenet al., 1992].
The COARE IOP occurred within one of the longest
of historicalvalues[Tsuchiyaet al., 1989;McPhadenet
E1
Nifio-SouthernOscillation(ENSO) eventson record
al., 1992].
[Halpert
et al., 1994]: the SouthernOscillationIndex
Historical observationsindicate that equatorial surface salinities and surface currents are quite variable.
Sometimesthere is a local salinity maximum with values
(SOI)fellbelow
normal
earlyin 1991andremained
so
through
at leastOctober
1993.TheSOIreached
a deep
>35.0 psuon or nearthe equator[Tsuchiyaet al., 1989; minimum
(indicating
themature
phase
of ENSO)in
Richardsand Pollard, 1991; McPhadenet al., 1992], early
1992andreturned
tonearnormal
forafewmonths
presumablythe resultof equatorialupwelling.On other in mid-1992beforefallingagainin late 1992[Lukaset
occasions,there is a local salinity minimum with values al.,1995b].
Large-scale
monthly
meanwinds
overthe
<34.1 psu [Tooleet al., 1988; Tsuchiyaet al., 1989], COARE domainwerewesterlyduringthe maturephase
12,752
HUYER ET AL.: THERMOHALINE FIELDS NEAR 2øS,156øE
center of the COARE IFA, and our samplingplan was
designedto resolvezonal and meridionalscalesof 10 to
1992,near zeroduringthe first half of the COARE IOP, 100 km, vertical scalesof a few meters,and timescalesof
and westerly again during the secondhalf of the IOP a few days to a few months. The surveyswere designed
[Lukaset al., 1995b]. The 20øCisothermshoaledby to complementtime seriesobservationsfrom moorings
•-20 m in late 1991, during the mature phaseof ENSO, and stationary ships: though fixed platforms provide
and remained shallower than normal through the end better temporal resolution, they can resolveonly a few
of the COARE IOP [Halpertet al., 1994; McPhaden, spatial scales. We chose a butterfly pattern with its
intersection as close as practical to the IMET mooring
1993;Eldin et al., 1994].
The three Wecomasurveycruiseswithin the COARE and nearly stationary R/V Moana Wave at the cenIOP were conducted under quite different weather re- ter of the IFA. The long sectionsof the butterfly were
of ENSO in late 1991 and early 1992, easterlyduring the
relaxation from ENSO between March and September
gimes[Wellerand Anderson,1996]. Justpriorto our oriented in the zonal and meridional directions, parallel and transverse to the predominantly zonal equatorial currents, and connecting diagonal sectionswere in
surveys,there had been a prolongedburst of moderate
westerlies in late October and early November. During our first surveyperiod, winds were generallyweak
with net surface heating and little rainfall; these generally calm conditionswere interruptedonly by a brief
the northeastand southwestquadrants(Figure1). Al-
burst of moderate westerlies on November 24-25.
ional (N2S) line and east alongthe zonal (W2E) sec-
Low
most all of our sampling was done while underway at
7-8 knots(4 m s-1), travelingsouthalongthe merid-
tion. Sampling included standard meteorologicalobwhen there was another burst of moderate westerly servations and underway current profiling, as well as
winds that ended a few days before the beginning of upper ocean temperature and salinity measurements.
our secondsurveyperiod. A seriesof three strongwest- The length of the orthogonal sectionswas limited to
erly wind bursts occurredbetweenDecember20 and •-130 km to allow for completion of the entire pattern
January 2, coincidingwith the first half of our second in •1.5 days, with multiple repetitions per week.
Since Wecomawas based in Guam, there was a long
survey period; these were followedby calm winds and
transit
at the beginning and end of each cruise, and
heavy rain. Between our secondand third surveyperion
some
of these transits we were able to make a
ods,winds were generallyfrom the southeast,i.e., favorlong
cross-equatorial
Seasoarsectionalong156øE(Figable for equatorial upwelling. Westerly winds resumed
ure
2);
the
remaining
time (•-20 days) was spent in
on January 27, at the beginning of our third survey
the
COARE
IFA.
During
each of the three Wecoma
period. Moderate westerly winds with frequentstrong
cruises,
we
sampled
repeatedly
alongthe Standard Butsquallsprevailed from late January through the entire
terfly
pattern
(Figure
1,
Table
1)with a towedundumonth of February.
lating
vehicle,
Seasoar,
equipped
with a conductivityIn this paper we provide a summary of our hydrotemperature-depth
(CTD)
unit.
The
Seasoarvehicle,
graphicsamplingand an overviewof the spatial strucmanufactured
by
Chelsea
Instruments,
resemblesthe
ture and temporal variability of the temperature, salinvehicle
described
by
Pollard
[1986],
but
we
useda Seaity, and density fields measured along the zonal and
Bird
CTD
with
ducted
temperature
and
conductivity
meridional sections. As this is the first hydrographic
data set obtained from pumped Sea-Bird temperature sensorsrather than a Neil Brown CTD. Seasoartowing
and conductivity sensorsin an undulating Seasoarvehi- speedwas7-8 knots(4 m s-l). Minimumandmaxicle, data acquisitionand processingproceduresare also mum sampling depths were usually •3 m and •270 m,
summarized briefly in section 2. In section3, we show respectively. The cycling rate was 6-8 undulations per
overall time-averagedvertical sectionsalong the merid- hour, yielding a horizontal resolution of •3 km. The
ional and zonal lines and discussthe averageproperties Seasoartows had a typical duration of 2-5 days. Deof the principallayers(surfacelayer,thermocline,ther- tails of the Seasoaroperationshave been presentedin a
winds and net heating continued until December 10,
mostad,salinity maximum). In section4, we compare seriesof data reports[Huyeret al., 1993,1994;O'Malley
cruise-averaged
sectionsfor eachof the three surveype- et al., 1994]. Sincethis wasthe first fielduseof pumped
riods, corresponding
to three rather differentweather Sea-Bird sensors in a Seasoar vehicle and conventional
regimes.In section5, we examinethe temporalvari- CTD processingproceduresproved not to be fully adability on timescales
of daysto weeksby plottingthe equate, sensor configuration and modified processing
propertiesof selectedisobars,isotherms,
andisopycnals proceduresare summarized here.
Seasoar was equipped with dual-ducted Sea-Bird
and discussthe behavior of the principal layers. The
(SBE)
temperature(T) and conductivity(C)sensors
structureof the surfacelayer is further discussed
in secmounted
inside the nose, with intake and outlet ports
tion 6, and the mostimportantresultsare summarized
in section 7.
(•-0.05 m apart) on eachsideof the lowernose.Extra
tubing between the intake and the standard SBE duct
2. Observations
and Data Processing
The objectiveof our surveyswasto measurethe timevarying thermohalinefieldsof the upper oceanat the
was •-0.08 m long. Water was pumped through each
duct by an SBE submersiblepump. Raw 24-Hz CTD
data from Seasoarand unfiltered position and time data
from the Global PositioningSystem (GPS) were ac-
HUYER ET AL' THERMOHALINE FIELDS NEAR 2øS, 156øE
12,753
quiredandarchived
simultaneously.
Forin situconduc- concludethat both the amplitude parameter and the
tivity calibration,conventional
CTD castsweremade time constant of the thermal mass correction are related to flow rate. For the COARE data set, we found
before and after each Seasoartow, and salinity sam-
plesfrom Wecoma's
5-mintakewerecollected
onceper that varying the value of the amplitude constantwas
hour. The sampledata werepairedwith near-surface sufficient to remove most of the systematic difference
(3.0-5.5m) Seasoar
values,andconductivity
corrections between successiveascendingand descendingprofiles,
were determinedfrom regression
of Seasoarconductiv-
with no appreciableimprovementfrom usinga variable
ity onsampleconductivity,
separately
for eachtow and
time
sensor pair.
The entire Seasoardata set was therefore reprocessed
as follows: by calculatingthe T-C lags for consecutive
data segmentsin specifieddepth ranges;usingthe time
seriesof lagsto offsetthe 24-Hz conductivitydata relative to the temperature data within each segment;applying the appropriateconductivitycalibrationequations; applyingLueck's[1990]correctionfor the thermal massof the conductivity cell, with the value of the
amplitudeparameterrelatedto the T-C offsetfor each
data segment;and finally block averagingthe data to
1-Hz values. Since the salinity data from descending
profilessometimesexhibitedhigh-frequency
noisethat
was absent from ascendingprofiles, we use only data
from ascendingprofilesfor all three cruises. We estimate the processedSeasoarCTD data to be accurate
to -F0.01øCin temperature, -F0.01 psu in salinity, and
:El dbar (:El0 kPa) in pressure.
LueckandPicklo[1990]haveshownthat salinities
calculated from Sea-Bird ducted T-C sensorsare subject to
error from two main sources:a finite time lag between
the temperatureandconductivity
measurements
anda
finite transfer of heat from the mantle of the conduc-
tivity cellto the water flowingthroughit. In a conventional SBE CTD, theseerrorsare reducedby pumping
the water throughthe ductedpair at a fixedrate and
by applyinga fixedT-C alignmentoffsetanda thermal
mass correction with fixed vales for the time constant
andamplitudeparameters
to the raw 24-Hzdata [SeaBird Electronics,1992].Eventhoughwe usedthe standard SBE duct with high-speedpumps, the flow rate
throughthe sensors
mountedinsidethe Seasoar
vehicle
wasapparentlynot constant,presumably
because
of dynamicpressuregradientswhichvariedwith vehicleattitude and relativecurrents[Huyeret al., 1993].Cross
constant.
The basic Seasoar data set discussedin this paper
consistsof the 1-Hz temperature, salinity, and density
data from ascendingSeasoarprofiles. A representative ensembleof vertical profilesfor a singlemeridional
correlation between 24-Hz T and C data indicated that
the T-C lags (and flow rate) variedsignificantlyand
that the data should be realigned accordingly. Mori-
sonet al. [1994]havefoundthat flowrate alsoaffects (N2S)sectionfromeachcruiseisshownin Figure3. For
someanalyses,the data wereplacedon a fine-resolution
the rate of heat transfer from the conductivity cell and
20 November
5 January
5 February
20 Nov
5 Jan
5 Feb
5
.• •c
•_. •c
ß
r
300
2.5øS
2.0øS
1.5øS
,•
2 5øS
i
,
2 0øS
I
•
1.5øS
2 5øS
2.0øS
1 5øS
8(
12
16
20
24
28
36
40
44
lOO
15o
......
200
_
300
2.5øS
.
,
I
2.0øS
,
I
1.5øS
,
I ,I
2 5øS
20 Nov
,
20øS
I
1 5øS
•
,
2.5øS
I
2.0øS
,
5 Jan
5 Feb
I
1 5øS
34
35
36
37
38
Figure3. Samples
ofSeasoar
datafromtheCOARE
IFA:(left)vertical
distributions
(plotted
from2-nm,2-dbargridded
data)and(right)ensemble
profiles
(1-Hzdata)of (top)temperature,
and(bottom)
salinity,
foroneN2Slinefromeachcruise.
Profile
axislabels
applytotheNovember
20 data;the January5 profiles
areoffsetby 10øCand1.5psu,andthe February5 profiles
are
offset by 20øC and 3.0 psu.
48
12,754
HUYER ET AL.: THERMOHALINE FIELDS NEAR 2øS, 156øE
Table 2. Times of CompleteMeridional(N2S) Sectionsof the StandardButterfly Pattern
W9211A
W921 lB
0755 to 1938, Nov. 13a
0627 to 1534, Nov. 15
1652, Nov. 17 to 0118, Nov. 18
2325, Nov. 18 to 0754, Nov. 19
0605 to 1508, Nov. 20
1107 to 2032, Nov. 22
1854, Nov. 23 to 0258, Nov. 24
0435 to 1431, Nov. 25
1242 to 2215, Nov. 26
2102, Nov. 27 to 0645, Nov. 28
2023 Nov. 29 to 0550, Nov. 30
0421 to 1330, Dec. 1
1734, Dec. I to 0229, Dec. 2c
1945, Dec.
0553, Dec.
0346, Dec.
1138, Dec.
1238, Dec.
20 to
22 to
24 to
25 to
27 to
W9211C
0516, 21 Dec.
1653, Dec. 22
1323, Dec 24.
2120, Dec. 25
2225, Dec. 27
2209,Dec. 28 to 1721,Dec. 29b
1607, Dec. 30 to 0124, Dec. 31
2358, Dec. 31 to 0922, Jan. 1
1349, Jan. 2 to 2316, Jan. 2
1224, Jan. 3 to 2343, Jan. 3
1336 Jan. 5 to 2311, Jan. 5
2123, Jan. 6 to 0728, Jan. 7
1017, Jan. 8 to 2017, Jan. 8
0917, Jan 27. to 1827, Jan. 27
1930, Jan. 28 to 0430, Jan. 29
0945, Jan. 30 to 1926, Jan. 30
1827, Jan. 31 to 0358, Feb. 1
0320, Feb. 2. to 1221, Feb. 2
0604, Feb. 4 to 1524, Feb. 4
1425, Feb. 5 to 0015, Feb. 6
2200, Feb 7. to 0820, Feb. 8
0803 to 1824, Feb. 9
1819, Feb. 10 to 0452, Feb. 11
0433 to 1456 Feb.
12
0151 to 1139, Feb. 14
Timesaregivenin UT. Almostall sections
weresouthward
along15ø06'EfromSBN(1ø14'S)to SBS(2ø26'S).
aNorthernportion of sectionwas along155ø56•E.
bSection
wasinterrupted
by Seasoar
recovery
anddeployment.
CSection was northward
from SBS to SBN.
grid by block averaginghorizontally over 2 minutes of
latitude or longitude(2 nm) and verticallyovereither
2-dbarpressure
bins(e.g.,Figure3) or over0.1 kg m-3
incrementsof the densityanomaly(at). Althoughdata
tend to be relatively sparsenear the surface,at depths
>230 m, and at the extreme ends of each line, almost
all 2-dbar grid points above 250 dbar were sampledin
> 75% of the sections.
GPS receiverswere used to determineship's heading
and to correct for gyro heading errors. The nominal
depth range is 20-300 m. Vertical resolutionwas nominally 8 m, from data from 8-m bins with 16-m pulse
length (W9211A) or 8-m pulselength (W9211B,C).
The processedvelocity data are accurateto better than
0.02m s-• on horizontalscalesgreaterthan 10 km.
The overall Seasoar data set
from the three cruisesincludesa total of 38 complete 3. Spatial Structure of Time-Averaged
sectionsalongthe meridional(N2S)line (Table2) and Fields
33 completesectionsalongthe zonal (W2E) line (Table 3). Each sectionwas usuallycompletedin about 9
Althoughour sectionsareshort(,,•130km) compared
hours.
to the equatorialradiusof deformation(,,•200km), sigUpper oceancurrentsweremeasuredalongthe ship's
nificant
horizontal
structure
is discernible in the time-
track [Lukaset al., 1995a]with a 150-kHzRD Instru- averagedthermohaline fields. Time averagesand temmentsacousticDopplercurrentprofiler(ADCP). Dual poral standard deviationsfor the entire three-cruisepe-
Table 3. Timesof Complete
andNearlyComplete
Zonal(W2E)Sections
oftheStandard
ButterlyPattern
W9211A
W921 lB
0814 to 1808, Nov. 14
0329 to 1038, Nov. 17•
0723 to 1647, Nov. 18
1422 to 2310, Nov. 19
0242 to 1206, Nov. 23
0925 to 1755, Nov. 24
2120, Nov. 25 to 0525, Nov. 26
0505 to 1353, Nov. 27
2019, Dec. 19 to 0227, Dec. 20•
1217 to 2048, Dec. 21
0000 to 0827, Dec. 23
2026, Dec. 24 to 0455, Dec. 25
2017, Dec. 26 to 0526, Dec. 27
0532 to 1450, Dec. 28
0018 to 0846, Dec. 30
0826 to 1715, Dec. 31
1602, Jan. 1 to 0007, Jan. 2
2230, Jan. 4 to 0639, Jan. 5
0603 to 1443, Jan. 6
1354 to 2234, Jan. 7
0250 to 1139, Jan. 9
1330,Nov. 28 to 1323,Nov. 29b
1225 to 2123, Nov. 30
W9211C
0143 to 1207, Jan. 28
1646, Jan. 29 to 0225, Jan.
0238 to 1112, Jan. 31
1124 to 2015, Feb. 1
1410 to 2304, Feb. 3
2230, Feb. 4 to 0719, Feb.
0747 to 1641, Feb. 6
1524, Feb. 8 to 0114, Feb.
0200 to 1128, Feb. 10
1207 to 2135, Feb. 11
0845 to 1856, Feb. 13
1833, Feb. 14 to 0456, Feb.
30
5
9
15
Timesaregivenin UT. All sections
wereeasthward
along1ø50'S
fromSBW(155ø30'E)
to SBE(156ø42'E).
•Partial section only, completed with CTD stations.
•Section
wasinterrupted
bySeasoar
recovery
anddeployment.
HUYER ET AL.' THERMOHALINE
FIELDS NEAR 2øS, 156øE
12,755
riod (November13, 1992to February15, 1993) were (•34.5 psu), nearlyisothermal,and weaklystratifiedin
calculatedfrom both the 2-dbar pressure-griddeddata salinity[Lukasand Lindstrom,1991];the surfacesalinand the 0.1 kg m-a density-gridded
data. Averages ity was slightly lower than normal and about the same
of pressure-gridded
data (i.e., alongisobars)are more as observed after strong westerly winds in early 1986
appropriate for comparison with data from moored in[Eldin et al., 1994]. The very steepthermocline,with
strumentsand for studying layerswith very weak strat- temperaturedecreasing
•15øC in 200 m [Tooleet al.,
ification. Averagesof density-gridded
data (i.e., along 1988],containsthe broadsubsurface
salinitymaximum
isopycnals)are more appropriatefor studyingwater [Richardsand Pollard, 1991];Eldin et al. [1994]note
mass characteristics. As might be expected from the
closeproximity to the equator, these fields showasymmetry in the meridional plane and lateral homogeneity
in the zonal plane.
Vertical profiles of temperature, salinity, and density anomaly at the four cornersof our sampling pattern Figure 4 show the characteristic structure of the
western equatorial Pacific. The surface layer is rather
that the upper thermocline in this period was 10-40 m
shallower than climatology, consistent with the ongo-
ing E1 Nifio conditions.The deepthermostad(250 to
300 m) is about 1øC colderthan suggestedby its historical name (Thirteen DegreeWater [e.g., Tsuchiya,
1981]), which is more appropriatein the easternPacific. Statistics of the pressure-griddedtemperature,
salinity,and density(Figure 4) showmaximumvarithick (>30 m), very warm (>28øC), relativelyfresh ability at depths with the strongestvertical gradients
Average
Std Dev
Average
50
Std Dev
22
100
'7• 23
15o
T
SBS
200
T
•o•25
250
26
300
,
27
0
,,,3
21
50
22
100
'vf•23
150
S
200
250
•
S
•24
o•25
26
300
...............
34
0
27
34.5
,
i
,
I
35
,
I
35.5
,
I
,
0
0.2
0.4
...............
34
I ,
21
50
34.5
35
35.5
0
,,,,I,,,,I,,,,I,,1,1,,,,I,,,,
I
0.2
0.4
I
I
10
20
22
SBS
SBN
150
ot
200
•
•24
E
•25
250
26
300
27
ß
21
22
23
24
25
26
27
0
0.4
0.8
0
50
100
150
200
250
300
0
Figure 4.
Profilesof the averageand standarddeviationof (top) temperature,(middle)
salinity,and (bottom)sigma-tnearthe four corners:SBN (heavyline), SBS (thin line), SBW,
(thin dottedline), and SBE (heavydottedline). Statisticswerecalculated
fromboththe (left)
pressure-gridded
data and (right) density-gridded
data.
12,756
HUYER ET AL.' THERMOHAI,INE FIELDS NEAR 2øS,156øE
and the standard deviation of isopycnal displacement is
depth, and differencesbetween the northern and southern cornersare small at most depths. The surfacelayer
has virtually the sametemperature at all four locations:
10-15 m throughmostof the water column(Figure4).
The isopycnal standard deviations of temperature and
salinity are very small (<0.2øC, •-0.05 psu), exceptin differences
amongprofiles(<0.15øC) are lessthan the
the coreof the salinitymaximum(23.5to 26.3 crt)and standard deviation at each location. Surfacesalinity is
0.05psuhigherin the south(at SBS)and0.05psulower
at the top of the surfacelayer (ort <21.5).
The average profiles for the four cornersare remark- in the north(at SBN) than at eitherthe western(SBW)
ably similar (Figure 4): the profilesfrom the western or eastern(SBE) corners;this latitudinal salinitygraand eastern cornersdo not differ significantlyat any dient might be expected from the usual position of the
AverageT (øC)
AverageS (psu)
Average Sigma-t
29.2
- .--
21.35
- .... --'.,'29.1 -..,--
,'
.....
I
I
-;I,,
- "" ' '-2•.,½5
....
:: ....
5O
'
' ',--21.4-
•'••-21.5
_
_
-
-- •27.•
.-----.--.•26.•
lOO
-•22.0•
__••55.-1
•
- -•------- 23.0 •
- •24..0
••'
1•o
- '"'-'"'"'•"•
-•
2OO
-•
19. •
- •25.5•
17.
•
-_•13.•
25.0'"--'-•-
•::::::***:•:,:..:.•72
i 55.1•
""'---
::1'*::½?
- -•
•-'-'""----26.0
•
--
-_ - •55.0•
250
_
--
_
_
'• 54.9-•••
_
_
_
_
_
_
,
5OO
2.5øS
I
,
I
2.0øS
T
,
1.5øS
2.•øS
2.0øS
1.5øS
2.5øS
I
,
2.0øS
I
I
1.5øS
21.4O
_
_
_
.... _.........%•.•,5
.....,-',,_,•:,,,,,,
_
_
5O
.........
- -----'-----•21.5'•-----•-•_•
28.9 .... ' ...............
_
-'-'---"--•
lOO
28. •
-
-•27.•
-
-•26.•
-
----------•
24-. •
.0-----
-
-
---•23.
150
•
_
_
-
f•::•il::
i•i
;:/;
:•::::•
'•-'•:':'::•::•:•.:
'
..........................
....
•
--•
-
23.0 •
- •----23.5 •
-
,,__ --_.
•
24.5 •
-
i
i
-.----------- 25.0 •
_
-•25.5
-
-'-
_
200
•
17.••
16.•
-
-
•26.0•
..............
•*;•**•**•**•?;:*;:•
............
=====================
*:**•:::'
........................ •
• 55.O•
•54.9•
250
--"-'%
_
26.5 •
_
I
500
155.5øE
,
I
156.0'E
,
I
156.5øE
,
,
155.5'E
I
156.0'E
,
I
156.5'E
155.5øE
156. O"E
156.5øE
Figure 5. Distributionof time-averaged
(left) temperature,(middle)salinity,and(right)sigma-t
in the upper300 m along(top) N2S and (bottom)W2E. Theseaverages
werecalculatedfromthe
2-dbar pressure-gridded
values.Extra (dashed)contoursshowthe structurein the low-gradient
near-surface layer.
_
_
-'-----•-
-
-•18.•
--
-
--
-•19.•
--
-•24.0--
_
-•22.•
_
-•22.0--•22.5•
---'--'-•
-
-
-
-
HUYER ET AL' THERMOHALINE
FIELDS NEAR 2øS, 156øE
12,757
Intertropical Convergence Zone north of the equator, 24øC isotherms slope up toward the equator, and the
though surface salinities on the equator are sometimes 12ø to 13øC isothermsslope down toward the equator,
raisedby equatorialupwelling[Tooleet al., 1988].Wa- while the 15ø to 17øC isotherms are very nearly level.
ters in the upper thermocline are warmer and lighter Isopycnals
generallyparallelthe isotherms(Figure5).
at SBS than at SBN, while waters in the lower thermo- Through geostrophyand the thermal wind relation, the
cline are colder and denser at SBS than at SBN; these average temperature and density sectionssuggestthat
gradients are consistentwith geostrophicshear above the core of the EUC was near the 16øC isotherm, at a
and below the core of the EUC. The salinity maximum depth of about 200 m; this is consistentwith the average
is moreintense(by 0.2 psu) and muchlessvariableat distribution of currents measured by the ship's acous-
SBSthan SBN. The deepthermostad(250to 300 m) is tic Doppler current profiler (Figure 7), which shows
cooler(by 0.6øC) and slightlyfresher(by 0.03 psu) in strongest
eastwardflow (•0.35 m s-1) at a depthof
--175 m at the north end of our section.
the souththan in the north (Figure4).
Verticalsections(Figures5 and 6) showthat meridThe subsurface
salinitymaximum(S •35.3 psu) lies
ional variations are generally monotonic with latitude,
though the sign changeswith depth. Zonal gradientsare
comparatively weak and generally insignificant. Within
the upper 60 m, the average salinity decreasesnorthward toward the equator, and the top of the halocline
within the layer of upwardslopingisothermsand isopycnalsat depthsbetween120 and 200 m (Figures5 and
6). Becauseof the verticalexcursionof isopycnals,
this
feature appears to be lessintense and thicker when sections are averagedon isobars(Figure 5) rather than
(at --34.3 psu) slopesdowntowardthe north. Within isopycnals(Figure 6). The isopycnalaveraging(Figthis surface layer, the average temperature has very ure 6) showsthe 35.5 psuisohalineextendsthroughour
weakstratification
(<0.01øCm-1) and no systematiczonal section, to as far north as 1.5øS. The intensity of
meridional gradient. Within the thermocline, the 19øto the maximum decreases,while its variability increases
AverageT (øC)
AverageS(psu)
Std Dev of Salinity
soL
lOO
150
200
I
250
2.5øS
2.0øS
1.5øS
2.5øS
2.0øS
1.5øS
2.5øS
2.0øS
1.5øS
50
•
'-------lOO
•
28. -•----•--.--.----.--"•
27. •
26. •'••.
' •--•-
•,--,..,-•,-25.•
24.
--
I
'
I
' ttl--o
..,'
-.-'"
I ' -.4._.,.'
4 I.-o,•.,,.,---.,,..-..•
0.05 ------.-.-'-'•'--i
-. .....
-
_
.
-.--.•--•150
]
200
22. •
::i::i::i:;i!::i::i
i::i::iii::iiiiii:i,•:::::::i:i:i:::i:i:i:!:i:•:iiii:2!!!ii..i::i::i:•::i!::!::•
iiii::i::i;iii!i!iii:i:•:•
...........
:::::::::::::::::::::::
'-------•
-•17.•
-
18. •
---- 16.•
_
250 I
155.5•E
,
I
156.0øE
,
I
156.5øE
,
155.5øE
'----tf,:.
I
156.0øE
,
I
156.5øE
,
155.5øE
156.0øE
156.5øE
Figure 6. Distributionof time-averaged
temperatureandsalinitywithinthe thermocline,
along
(top) N2S and (bottom)W2E; the standarddeviationof salinityalongboth linesis alsoshown.
Thesestatistics
werecalulatedfromthe 0.1 kg in-1 density-gridded
data andplottedversusthe
time-averaged pressureof each bin.
12,758
HUYER ET AL.' THERMOHALINE FIELDS NEAR 2øS,156øE
Average U (cm/s)
Average V (cm/s)
10
5O
_
lOO
i!i::•!!'
-'••'••'"•::..•
'"-"--•'- ".:*::•;::iiiiill
I
•lo
•
150
_
_
_
200
_
•o•
250
300
I
2.5
2.0
S
i
S
i
i
i
i
1.5
i
2.5
S
S
2.0
S
1.5
S
Figure 7. Distributionof the averagezonal(U) and meridional(V) currentalongN2S, from
ADCP
on Wecoma. Shaded areas indicate westward or southward flow.
of the
towardthe north (Figure6); the regionsof maximum reflectedin the upperocean,temporalaverages
variability
coincide
roughly
withregions
ofhighvertical thermohalinefieldswere calculatedseparatelyfor each
data and the
shearaboveand belowthe coreof the EUC (Figures6 cruise,usingboth the pressure-gridded
density-gridded
data.
Results
(Figures
8,
9, and10)
and7). Thesalinitymaximum
andtheEUCcorelieat
aboutthe samedepth,but watercarriedby the EUC
are shown for the meridional sectiononly, since cruise-
[Tsuchiya
et al., 1989].
section.
properties
at all locations
onthe zonalsection
is less saline than water farther south becauseof up- averaged
are
similar
to
those
near
the
midpoint
of the meridional
stream dilution with low-salinityNorth Pacificwaters
Cruise-to-cruisedifferencesin surfacelayer properties
(Figures
8 and9) areconsistent
withdifferences
in the
4. Cruise-Averaged Sections
meanwindsandsurfacefluxes(Table4). Althoughthe
The verydifferentlocalforcingregimes
correspond-surfacelayerwasverywarm(>28.5øC)in all threesur-
it wasdistinctly
cooler(by-•0.3•C)during
ingto thethreesurvey
periods
canbecharacterized
by veyperiods,
theiraverage
windstress,surfaceheatfluxandprecipi- December20 to January9 (Figure 9), the surveype-
tationvalues
(Table4): calmwinds,lowrainfallandnet riod with strongestwesterlywinds and net heat loss
stratification
in
heatingduringthe firstsurveyperiod;strongwesterly (Table4). The averagetemperature
the
upper
50
m
was
relatively
strong
(•0.04•C
m
-1)
winds,highrainfall,andnetcooling
duringthesecond;
13to December
2 (Figures
8 and9),
and ]noderatenorthwesterlywinds, moderaterainfall, duringNovember
surface
and n9 net heatingor coolingduringthe third. To see the surveyperiodwith calmwindsandstrongest
whether these cruise-to-cruise weather differenceswere
heating
(Table
4);it wasnearly
absent
(<0.01oC
m-1)
Table 4. AverageandStandardDeviationof Daily AveragedWind Stress(PositiveEastwardor Northward),
Precipitation,and SurfaceHeat Flux (PositiveUpward),for EachSurveyPeriod
< r• >,
Pa
Nov. 13 to Dec. 2
Dec. 20 to Jan. 10
Jan. 27 to Feb. 15
0.006
0.045
0.034
< % >,
Pa
0.001
-0.035
-0.027
r},
r},
< P >,
mm
mm
0.015
0.048
0.022
0.015
0.032
0.013
1.9
12.3
4.5
3.6
17.6
7.8
Pa
Pa
P,
< Qr >,
W m -2
W m -2
-66.3
14.6
-1.3
64.4
77.8
55.9
Wind stressdata (providedby R. Weller)are fromthe WoodsHoleOceanographic
Institutionbuoyat 1ø45'S,
156ø00'E.Rainfall and surfaceheat flux estimates(providedby C. A. Paulson)wereobtainedfrom measurements
aboard Wecoma and do not include the heat flux by rainfall.
HUYER ET AL' THERMOHALINE FIELDS NEAR 2øS,156øE
,•j,l•,,,I,,•lf•,,I,,,,l•,l,,,,I,'''
12,759
21
22
50
23
_
100
SBN
//
-
150
._
E•24
200
25
250
26
3OO
f•
•••r''l
10
15
20
....
I ....
25
30
35
40
Average Temperature
27
I ....
45
34
50
34.5
35
35.5
i
0[IL•'I....
I,],,i....
I,,,,1•]•[I
....
I.... 21
SBS
50
36
36.5
37
37.5
Average Salinity
•
i
J
i
i
.
ß ß
.
.
i
i
i
J
i
.
22
SBI•
100
23
150
m24
200
25
250
26
300
lW9211A B•.........
, ,C/IJ....
j i , ,
, i ,'
,'"
J•
27
, W92•11B
0.2
0
W9211C
i
i
i
j
l
0.2
Std Dev of Salinity
Average Salinity
Figure8. Vertical
profiles
ofthecruise-averaged
temperature
andsalinity
nearSBN(thick
lines)
andSBS(thinlines).
Scales
refertoprofiles
forW9211A;
profiles
areoffset
by10øC
or1.0
psuforW9211B
andby20øC
or2.0psuforW9211C.
Profiles
werecalculated
fromboth(left)
pressure-gridded
dataand(right)density-gridded
data.
scaleof thisavduringDecember
20 to January9, the surveyperiod the third (Figure11). The meridional
spreading
is of O(100km);thedepthof
with strongwesterlywindsandnet surfacecooling.In erageisotherm
in themeridional
temperature
gradient
all threesurveyperiods,
thesurface
layerwasveryfresh thesignreversal
shoaledfrom -•220 to -•160 dbar and warmed from -•14
(<34.3psu)compared
to the underlying
thermocline
andtheirscaleare
waters,andnear-surface
salinities
decreased
towardthe to -•20øC(Figure8). Thesechanges
with the shoaling
of the coreof the
northendof the line (Figures8 and9); bothareconsis- roughlyconsistent
tent with the absenceof equatorialupwelling.The near- EUC from -•190 to -•150 m (Figure10). The average
surfacesalinitystratificationwasstrongerduringthe zonalvelocityat the coreof the EUC alsointensified,
at thenorthern
endofourN2Sline,where
second
surveyperiod(Figure9) whenrainfallwashigh particularly
fromabout0.25ms-• during
thefirstsurvey
(Table4). Thezonalcomponent
ofthenear-surface
cur- it changed
rentswaseastwardduringall threecruises(Figure10) periodto >0.5 m s-• duringthethird(Figure10).
andwasespecially
strong(>0.3m s-•) duringthesec- The structureand intensityof the subsurfacesalinalsochanged
significantly
(Figures
8 and
ondsurveyperiod,whenaneastward
Yoshida
jet devel- ity maximum
thesalinitymaximum
was
opedin response
to thelocalwesterly
winds[Lukaset 11). Duringall threecruises,
al., 1995a].
The thermocline remained at about the same depth
mostintenseat SBS,the southernend of the N2S line.
The depthof maximum
salinityat SBSshoaled
slightly
W9211A
to about145dbardur(70to 230dbar)during
thethreesurvey
periods
(Fig- (from170dbarduring
ures8 and Figure11), exceptthat the 13øCisotherm ingW9211C,Figure8) andlay withinthe layerof up(OT/Oy < 0), just abovethe
at the bottom of the thermocline at SBS was -•10 wardslopingisotherms
dbar shallowerduringthe third period(Figures8 and coreof the EUC duringall three surveyperiods.Dursurvey,whenthe EUC wasapparently
11). The equatorward
spreading
of isotherms
within ingthe second
the thermoclineshoaledsignificantlybetweenNovem- centeredat about1.5ø S (Figure10), the structureof
corewasmorecomplex,
andthe maxber andJanuary:the horizontalisothermchanged
from the high-salinity
15øCat 220 m duringthe first surveyperiodto 17øC imumsalinityat SBN waslower(Figures8 and 11).
at 190m duringthe secondand20øCat 160m during Duringthe third surveyperiod,the EUC corelay far-
12,760
HUYER ET AL.' THERMOHALINE FIELDS NEAR 2øS,156øE
10
20
•
40
EL SO
60
70
8O
2.5øS
2.0øS
1.5øS
2.5øS
2.0øS
1.5øS
2.5os
2.0øS
1.5øS
2.0øS
1.5øS
2.5øS
2.0øS
1.5øS
2.5øS
2.0øS
1.5øS
10
20
•
40
n
5o
60
70
80
2.5øS
1.5 Nov
to
2 Dec
20
Dec to 9 Jan
27
Jan
to
15
Feb
Figure 9. Distributions
of thecruise-averaged
near-surface
(top)temperature
and(bottom)
salinityalongthe N2Sline,calculated
fromthe 2-dbarpressure-gridded
values.Contourintervals
are 0.1øC and 0.05 psu; areaswith temperaturesof 29.0ø- 29.4øCor salinitiesof 34.25-34.30 are
shaded.
ther north (Figure 10), and the intensesalinitymaxi- isothermsas well as in depth. Variationsin the depth
mum layer extended along the entire length of the N2S of the thermoclineand in the strengthand depth of
section(Figure 11).
the EUC have previouslybeen observedin this region
The most surprisingresult of comparingthesecruise- [McPhaden
et al., 1990,1992;McPhaden,
1993]and
averagedsectionsis that the coreof EUC (represented can be explainedin terms of a propagatedresponse
by spreadingisothermsand isopycnals)
migratesacross to remote forcing and large-scalepressuregradients
HUYER ET AL.: THERMOHALINE
50
FIELDS NEAR 2øS, 156øE
12,761
-
100
_
150
-
200
250
300
2.5
2.5
øS
2.0 oS
øS
1.5 •S
2.5 øS
2.0 øS
1.5 øS
o
50
100
150
200
250
300
I
I
2.5 øS
I
I
I
2.0 øS
13 Nov
I
i
I
1.5 øS
to 02 Dec
i
I
2.5 øS
2.0
øS
1.5 øS
19 Dec to 10 Jan
2.5 øS
2.0
øS
1.5 øS
27 Jan to 14 Feb
Figure 10. Distributionsof the average(top) zonaland (bottom)meridionalcomponents
of
ADCP velocityalongthe N2S line, calculatedseparatelyfor eachcruise.Areasof negativevalues
(westwardor southward
flow)are shaded.
[McPhaden
et al., 1990;KesslerandMcPhaden,
1995]. riod, our data can be usedto study local-variationson
We know of no dynamical modelsthat include a mech- timescalesof 2-20 days,aswell asthe spatialdifferences
between cruises. Data from the moored array and staanism for the EUC to migrate acrossisotherms.
tionary shipsare moresuitablefor time seriesanalysis,
but theseplatformsprovidevery limited spatial infor5. Temporal Variability
mation. Resultsfrom the moorings[Weller and AnA central purposeof the Wecomasurveyswasto mea- derson,1996; Eriksen et al., 1995a,b; Kutsuwadaand
surethe temporalvariabilityof the zonalandmeridional Inaba,1995]andstationaryships[Smythet al., 1996a,
and Gregg,1996]showthereis significant
structure on scales of 5-100 km during the COARE b; Wijesekera
(daysto weeks)
IOP. Since each occupationof the butterfly sampling variabilityon boththe longtimescales
pattern lasted about 32 hours, and the averagere- that can be resolvedby the Wecomasamplingand on
peat time was <2 dayswithin each•020-daysurveype- shortertimescales(<1.5 days)that werenot resolved
12,7t52
HUYERET AL- THERMOHALINE
FIELDSNEAR2øS,!56øE
150
200
250
2.5øS
2.0øS
13
1.5øS
Nov to 2 Dec
2.5øS
2.0øS
20
1.5øS
Dec to 9 Jan
2.5øS
2.0øS
1.5øS
27 Jan to 15 Feb
Figure11. Distributions
ofcruise-averaged
(top)temperature
and(bottom)
salinity
along
the
N2Sline,calculated
fromthe2-dbar
pressure-gridded
values
(fortemperature)
andthe0.1kg
m-3 density-gridded
values
(forsalinity).
in our sampling.Much of the high-frequency
variabil- showed
evidence
of surfaceheatingor freshening,
especiallywhenwindswereveryweak(e.g.,November
22,
cyclingin the near-surface
layers[MournandCaldwell, January
8) or following
highrainfall(e.g.,January
8).
ity is due to semidiurnal internal waves and to diurnal
1994];theseandotherhigh-frequency
signals
mightap- Rainfallthat occurred
duringstrongwinds(e.g.,De-
pear as noise in our data. In this presentationof the
temporal variability of the upper oceanin the COARE
cember20-25) causedslightfreshening
throughout
the
upperlayer,to depths>40 m (Figure12, Plate2).
IFA, weincludesampling
time asan independent
variThermalstratification
of the near-surface
layer(Figableand usea varietyof techniques
to reducethe noise
ure 12, Plate 1) wasgenerallystrongduringW9211A
from high-frequencysignals;theseare describedin the
(November
1992),whenweakwindswereprevalent.
appropriate subsections.
Salinitystratification
ofthenear-surface
layerwasgenerally
weak,
under
both
weak
and
strong
winds;
strong
5.1. The Near-Surface Layer
salinitystratificationin the upper 10 m wasobserved
Water abovethe main pycnocline
wasusuallywell ononlya fewoccasions
(notablyJanuary
3-8) immedimixed or very weakly stratified(Figure 12, Plates 1 atelyfollowing
veryheavy
rainfall(Figure
12,Plate2).
and 2) exceptin the upper 5-10 m, whichsometimes The near-surface
layerwaswell-mixed
duringportions
HUYERET AL.' THERMOHALINEFIELDSNEAR2øS,156øE
DEC
NOV
15
0.16
20
25
JAN
JAN
20
25
30
I I I I I I I I I I II III
50
12,763
FEB
30
5
10
I I I I I I I I I I
5
10
15
IIIIIIIIIIIIIIIIIIII
0.08
0.00
'.1
-0.08
40.00
20.00
0.00
100.
x
i
-lOO.
20
25
20
30
JAN
JAN
DEC
NOV
15
50
25
FEB
30
10
5
5
10
15
2O
•::
•
60
8O
ß
lOO
I I I
IIIIII
i
IIII
i
,
,,
IIIII
' i
''•
•
,,'.• •
IIIII
I II II
III
IIII
IIIII
I I
IIIII
II
III
I
IIIIII
IIIII
IIIIIII
III
!'•!!B',iii',!',:!:"!'i
........
•::•::::::4ii•i::
::iii•iiiiii::i:,i'•,•:.::,,::,•,,::.•
...... •...........
•...::•:•:•:•::::::s•.•::•
•'""'••
:::::::::::::::::::::::
::::::::•::
•::•
:•.....
:::::::::::::::::::::::
::::::::::::::::::::::::::::::
.........
.............................
:i.............................
e .......,•.
•
:::::::::•::::
..... i.....
'"':':::
......
.........
.......................
•.?] .•::
ß.......
:..........
:•:::::(
......
•:..•.
:::::::::::::::::::::::::::::::::
.......
;•::•::•::::
•::::•
....
'?::•::
•
,
,'
lOO
I
317
322
327
332
337354
359
364
369
374
,,,%,,,
593
398
403
408
413
W9211A
W921lB
W9211C
Figure12. Time-depth
distributions
of upperlayertemperature
(middle
panel)andsalinity
(bottom
panel)at 1ø50;S,
156ø06;E;
vertical
ticsbetween
panels
indicate
thetimeof Seasoar
profiles
at thislocation.Dailyaveraged
windstress
components
(newtons
persquare
meter,
measured
onIMET buoy),rainfall(millimeters
perday,measured
on Wecoma)
andsurface
heat
flux(wattspersquare
meter,calculated
fromobservations
on liVecoma)
areshown
in thetop
panel.Contour
intervals
are0.1øCat T >28øC,IøCat T <28øC,0.05psuat $ <34.5,and0.1
psu for $ >34.5.
of W9211B(lateDecember)
andW9211C(earlyFebru- nal and tidal variations(Figure12, PlatesI and 2).
ary)whenstrongor moderate
westerly
windsprevailed. We therefore chosethe 20-dbar surface to examine the
Water propertiesat a depthof 20 dbar seemto be time-varyinghorizontalstructureof the near-surface
plotsof the 20typicalof the upperlayerand relativelyfreeof diur- layer.Time-latitudeandtime-longitude
12,764
HUYER ET AL.: THERMOHALINE FIELDS NEAR 2øS,156øE
t
,•
28.5
29
29.5
-1.2
-1.8
a3d '
O/
4O
W9211B
40
o
897
W9211C
400
41•2'4
Plate 1. Time variation of upper layer temperaturealong 156ø06'Eduring the COARE IOP,
represented
by a subsetof N2S sectionsfrom eachcruise(Table2). Contoursof T (28øC are
suppressed.
Time axisis labeledin day number,with day 367 corresponding
to January1, 1993.
The sectionsshown begin on November 13, 17, 22, and 26; on December 1, 20, 25 and 30; on
January 3, 8 and 28; and on February 2, 5, 9 and 14.
HUYER ET AL.' THERMOHALINE FIELDS NEAR 2øS,156øE
•
33.75
34
"
40
12,765
34.25
34.5
I
.•,
80
•J
318 '
-1.2
322
W9211A
40
80
W9211
B a6o a•s
o
40
u'•'
.
•.
,•
-1.8
•
'
ß
397
W9211C
4
400
.,•'•o3
..../
",,"
,'
•
406
•''•,
_,•,.•'•'::'
8
-2.4
Plate 2. Time variationof upperlayersalinityalong156ø06'Eduringthe COARE IOP, repre-
sentedby a subsetof N2Ssections
fromeachcruise.Contours
of S >34.5 psuaresuppressed.
Time axisis labeledin day number,with day 367 corresponding
to January1, 1993. The sections
shownbeginon November
13, 17,.22,and26; on December
1, 20, 25, and 30; on January3, 8
and 28; and on February2, 5, 9 and 14.
12,766
HUYER ET AL.' THERMOHALINE FIELDS NEAR 2øS, 156øE
m temperatureand salinity (Figure 13) showa general welling driven by the strongwesterlywinds. The frontal
warmingand increasein patchiness(at scalesof --10- migration
rate (--0.3m s-1) is consistent
withtheEk30 km) duringthe first cruisewhen windswereweak, mantransport
estimate(Tx/f = 1.5x 104 kgm-1 S-1,
and show general cooling and homogenizationduring where Tx is the zonal wind stress and f is the Coriolis
the strongwesterly winds of the secondcruise;the mod- parameter),assumingthe transportis distributedover
erate winds of the third cruise similarly reduce patchi- the top 60 m. Such a rapid downwellingresponseis also
ness. These trends are more obvious in the time series
consistentwith the rapid formation of the Yoshida jet
of the spatial averagesand standarddeviations(Fig- by eastwardacceleration
of equatorialcurrents[Eldinet
ure 14); the meridionaland zonalaverages
are indistin- al., 1994].This migratingfront seemsto be associated
with large residual cooling in a one-dimensionalheat
guishable,and they are subject to the sametrends.
In general,the large-scale(50-100 km) temperature budgetof surfacelayersat the centerof the IFA [Smyth
gradients were weak in both zonal and meridional di- et al., 1996b].
A secondexception to the generally weak temperarections(Figure 13a), but there weresomenotableexceptions. During December 24-27, while winds were ture gradientswas observedin early January,soonafter
strongandwesterly,a temperaturefront (28.5øto29øC) the endof the strongwesterlywind burst(Figure13a).
migrated 100 km northward. This temperature change A small local temperatureminimum (T <28.5øC, dioccurredsimultaneouslyalong the entire W2E line, sug- ameter--40 km) migrated'50 km northwardalongthe
gestingthat a zonal front moved northward as a result
of equatorward Ekman transport and equatorial down-
a)
NOV
15
0.16
>..
(o 0.08
•
0.00
-0.08
20
southern half of the N2S line during January 5-8 and
crossedthe W2E line on January 9. This local tern-
DEC
25
30
20
JAN
25
30
5
JAN
10
FEB
30
5
10
15
1,,,,,,,,,I,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,
•
.,.,............ ,.
..
•__
NOV
15
2O
;"..,
ß
........
.......
....
:.'":'
DEC
25
50
20
30
5
FEB
•0
10
5
10
15
403
408
415
1.3øS
1.5øS
1.7øS
1.9øS
2.1 øS
2.3øS
1,1,1,1,1,1,1,1
2.5øS
1,1,1,1,1,1,1,1,1,1,1,'1,1,1,1,1,]
,I,1,1•1•1•1•1•1
156.7øE
156.5øE
156.3øE
156.1øE
155.9øE
156.7øE
155.5øE
517
322
327
332
537 354
359
364
369
374
593
398
Figure 13. Latitude-time and longitude-timedistributionsof propertiesof the near-surface
layer
with timeseriesof dailywindstress:(a) temperature
at 20 dbar;(b) salinityat 20 •ibar.Plots
weremadefrom individual1-Hz ascending
profiles(indicatedby dots),interpolatedverticallyto
20 dbar.
120.
-120.
JAN
JAN
25
240.
0.
HUYER ET AL.' THERMOHALINE
b)
NOV
15
20
FIELDS NEAR 2øS, 156øE
DEC
25
50
20
JAN
25
50
5
I I I I I.,I
12,767
JAN
lO
I I I I I I I
FEB
50
5
lO
15
•Z 0.16
>- 0.08
24-
";'•
•......-
•' o.oo
•
o
'
-0.08
NOV
15
20
DEC
25
30
20
JAN
25
30
5
-24-
JAN
10
FEB
30
5
10
15
1.3øS
1.5øS
1.7øS
1.9øS
2.1 øS
2.3os
1,1,1,1,1,1,1•1,1,1,1,'1,1,1,1,1,•
2.5øS
]1
156.7øE
,I,1,1•1,1,1,1,1
i iU.,IJ
i i i 1,1,
i i œi ,Li I i ki I i I i .I
i iJ'..,•J,,.t,..l
i i i i i L.'.J
i ili
11
"iF"'"
.......
'"
::::::::::::::::::::::::::::::::::::::::::
'Dii:
i::•::::::!•::::i::i:'.!:•illiii:i;•i•
::::::::::::::::::::::::::::::::
'i•;',•:,•½:•'"•::'::::':::".
..."••
"::,
'"":'"'"
156.5"E
156.3øE
156.1øE
155.9øE
156.7øE
155.5øE
517
522
527
552
337
354
35g
364-
36g
374
595
598
405
408
415
Figure 13. (continued)
perature minimum was connected to the bottom of the
ciated with westerly winds which induce equatorward
mixedlayer,ratherthan to the seasurface(Figure12); Ekman transport.
it could result from local upwelling or perhaps from
mixing by high-velocity shear at the base of the mixed
layer. Velocity data from a stationary ship adjacent to
There
was little
or no correlation
between
the time-
space fields of near-surfacetemperature and salinity
(Figures13a and 13b). For example,the two mesoscale
the IMET mooring[Smythet al., 1996a,b]andfromthe features that were so obvious in the temperature field
IMET and ADCP moorings[Plueddemann
and Weller, (i.e., the migratingfront seenDecember24-27 and the
1994; Kaneko et al., 1993] indicatethis cold feature cold spot seenJanuary5-10) had no signaturein the
was on the southernflank of the decayingeastwardjet salinity field. Hence, each of the mesoscalefeatures seen
[LukasandHacker,1995].Likethe migratingfront,this in either temperature or salinity is also reflected in the
mesoscalefeature seemsto result in a strong cooling density.
episodeat the centerof the IFA [Smythet al., 1996b]. 5.2. The Thermocline
The near-surface
salinityfield (Figure 13b) was less
patchy than temperature(Figure 13a), and it had a
The near-equatorialthermocline is complex, and its
persistentlarge-scale(50-100 km) meridionalgradient, differentaspects(e.g., depth, thickness)can vary inwith lower valuesfarther north. A rather sharp salin- dependently. In this sectionwe describevariations in
ity front (at --34.2 psu) migratednorth-northeastward isothermdepth, thermoclinethickness,strength of the
during November 13-15 and east-southeastwarddur- vertical gradients,isothermspreading,and temperatureing November 25-30. Northward propagating isoha- salinity characteristics.
lines were also observedduring December 20-25 and
In the COARE IFA, the main thermocline coincides
February 4-10. Northward migration seemsto asso- with a strongpycnocline(Figure 4), and this pycno-
12,768
HUYERET AL.' THERMOHALINE
FIELDSNEAR2øS,156øE
NOV
DEC
10 15 20 25 30
0.16
-
5
JAN
10 15 20 25 30
5
FEB
10 15 20 25 30
5
10 15
•••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••
I
.
>- 0.12 • 0.08
•
>•
0.04
o.oo
• -0.04
-0.08
-
-0.12
••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••
x
29.5
>.
0.4
'•"e•,•e o
©©
I--
**
©
,
o,
eAo
A OAee
ß A'eA AO
AeAeA
AOAeAeAeAeA
eOAAA
ßO
ß
AeAeAe
ß oAOAo•
øA A
27.5 IIII....I....J....]....]....]....]....I....I....III'lIJ....]....]....]....]....I....]....]il•
f•lAI
]....]I1II] 0.0
0.6 >
34.1
0.4
34.3
c•
o.• •
34.5
0.0 m
21.1
-
.
-
21.3
_
-•215
'
-
21.7
21.9
ße
22.1
Aee
eeee ß
e e ß ße
AOAoA
A AAAAA
_
AAAeA
AA
ø A
ß AA
A
ß
A
OAAOAo
O
O
"',,,,0,,,,,,,,
,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,I,t,^•i,,,,,,,,,,,,,
,,,,,,,,,,,,,,,,,,,it,*,
•1th,*,
•,,,
515
525
535
345
555
o N2S Ave
o N2S Std Dev
565
375
585
395
405
415
x W2E Ave
, W2E Std Dev
Figure14.Timeseries
ofthespatial
averages
andstandard
deviations
of20-dbar
temperature,
salinity,
anddensity
anomaly
(sigma-t)
along
theN2SandW2Elines,withtimeseries
ofwind
stress
components
in
the
top
panel.
Statistics
were
calculated
from
the
2-dbar,
2-nm
gridded
data.
cline(with•t of -• 22-26kg m-s) lay in the depth board
observations
fromNoroit[Eldinetal.,1994]
show
they covered
a largeareaand werestronger
on and
Figure15). Temperature
variednearlymonotonically
northoftheequator.
Theresulting
equatorial
upwelling
withdensity,
withonlya fewsmallinversions
(Figure apparentlycauseduplift of the upperthermoclineasfar
15).The28ø,20øand
14øC
isotherms,
laynearthetop, southas 2.5øS,i.e., alongthe entireN2S line. After
middle,andbottomof thethermocline
(Figures
5 and the resumption
of moderate
westerly
windsandequa11) respectively,
andcoincide
approximately
withthe torialdownwelling
(onJanuary26),the 28øCisotherm
22.1,25.2,and26.2kgm-a isopycnals
(Figure
15);the droppedrapidlyto a depthof about80 m andremained
rangeof about80-200 m throughoutthe COARE IOP
23øCisothermlay in the upperthermoclinenear the constant
throughtheremainder
ofthethirdsurveype24.2kg m-a isopycnal.
The depthsof someisothermsriod.
variedwith latitudeaswellasin timeFigure16, and
Within the main thermoclinethere were subtle and
the variousisothermsdo not behavein unison.
complex
variations
in isotherm
depth(Figure16).The
The 28øCisothermnearthe top of the thermocline verticalcoherence
variedfrom one surveyperiodto
wasrelativelydeep(>70 m) duringandimmediatelythe nextandwithinsurveyperiods.Forexample,
the
afterwesterlywinds(Figure16), andit wasshallow- depths
ofthe20øand23øCisotherms
werein phase
durest (-•5 m) afterthe periodof sustained
easterlies
last- ing November
andlate December
but out of phasein
ing fromJanuary7 to 26. Theseeasterlywindswere mid-February
(Figure16).In lateDecember
andearly
quite weak at the centerof the IFA, but the TOGA- January,temporalchanges
in the depthof the 14øC
Tropical
Atmosphere
Ocean(TAO)Atlasmoorings
at isotherm
seemto mirrorthoseof 23øCisotherm,
par2øN,156øE,and 0øN, 157.5øE[McPhaden,
1995;N. ticularlyat the northernendof oursampling
pattern
$oreide,
personal
communication,
1994]andtheship- (Figure16).
HUYER ET AL.' THERMOHALINE
NOV
15
20
FIELDS NEAR 2øS, 156øE
DEC
25
50
20
JAN
25
50
12,769
JAN
5
10
FEB
50
5
o,,,
•26.0 '
'l'½l'l'l'llt•lM'l'l'lH'l'l'l'!'!'l'l'l
I-M-H-H-'f-'-P-I"I'I'I"I'I'I'I'I'I'I'I'I'I'I'I'I'I'I
PH-H'M-H'I'I'I'I'I'I'I
...........
/'1111111
-
22
28.
--------•-----28.
- •27.
•
_
V-•-----26.
25
------2B•'
27. -'•-'•---•-
-----
-
••
--------- -'-----------26.-- -•--------
-•25.-'
E 24
'......
.
_
V
_
23
15
'"'"'"'"'""
200
-
10
_
•
-
-----26.•
•25.-
_
'
-
24.
z__--"--'-'-----24. --'-:--'------•23.--'%•
-- 2•_•• •.•__..•_____________
•
3. :'•
-
•
26
_
. •'•:•_.•
,
__
,
,
....................................
7
322
327
332
337
354
359
364
•
:
_
._
,,,,,•1
369
374
393
398
403
408
4
Figure 15. Time-depthdistributionof (top) densityand (bottom)time-densitydistributionof
temperatureat the intersection
(1ø50•S,156ø06•E),
contoured
fromascending
1-Hzdata;sampling
times are indicatedby the tics between(top) and (bottom)panels. Contourintervalsare 0.5
kg m-3 and 1øC,with heavycontoursat intervalsof 2 kg m-3 and 5øC.
The main thermocline was typically about 140 m
no significant tilt in November but sloped down to the
thick (Figures 15 and 16), but this thicknessvaried, north in December,January,andFebruary(Figure16).
from more than 160 m in November
to less than 120 m
These slow variations
in isothermal
tilts seem to be re-
in early February.
lated to the zonal currents:
A surprisingfeature of the time-depth distribution
of density (Figure 15) is the growth and disappear-
rent (above•28øC and •70 m), the westwardflow at
the top of the thermocline(•23-28øC and •70-120 m),
andthe eastwardEUC (•13-24øC and•120-250 m) all
accelerated
duringthe westerlywindburst[Lukaset al.,
1995a]. The coreof the EUC lay betweenthe 14øand
23øC isotherms(compareFigures10 and 11), and the
ance of weakly stratified layers within the main pycnocline. During much of the first survey period, the
verticaldensitygradientOp/Ozhasa rathersmallrange
the eastward
surface cur-
(Op/Oz= 0.02to 0.05kgm-q, or N = 5 to 10cph).A
morenearlyuniformlayer(with Op/Oz< 0.02kg m-q, thickness of this layer increased with latitude toward
N < 9 cph)occurred
at at of 24.0-24.5kg m-• through the equator(Figure 17).
Water mass characteristics of the thermocline vary
in complexways. Temperature-salinitycharacteristics
tend to be simplest at the southernend of our sampling
from November 20 to 27 and from December 25 to January 4. Strongly stratified layerswere also ephemeral, pattern, where there was almost no temporal variabiland occurredat the top of the pycnocline(i.e., the bot- ity in water characteristics below the 20øC isotherm
tom of the surfacemixedlayer);strongeststratification (Figure 18). Farther north, there is a wide range
of T-S characteristicsin the density range of 23.5 to
(N >20 cph) occurredduringJanuary5-10 .
Latitudinal tilting of isothermswithin the thermo- 25.5kg m-3. Within this range,thereis a generaltenclinevaried with both depth and time. The upper ther- dency for waters to be cooler and fresher farther north
mocline(represented
by 23øC)slopedup to the north than waters on the same isopycnalfarther south.
throughoutthe entireIOP, thoughthe slopevariedfrom
5.3. Salinity Maximum
most of November, December20-30 and January 6-10.
Anotherweakly stratifiedlayer occurredat 25.0-25.5 at
< 2 x 10-4 to > 4 x 10-4 (Figure16). Themidthermo-
cline (20øC) alsoslopedup to the north in November, Water mass variations can be seen most clearly in
December, and early January but had no significant the structure of the salinity maximum embedded in
slopein February. The lowerthermocline(14øC) had the lowerthermocline(Figure 18 and Figure 19). This
12,770
HUYERET AL' THERMOHALINE
FIELDSNEAR2øS,156øE
NOV
DEC
10 15 20 25 30 5
0.16 -!
>- 0.12
•
0.08
JAN
10 15 20 25 30
5
FEB
10 15 20 25 30
5
10 15
•••••••••••••••••••••••••••••••••••••••
t
:><[0.00
v--•,--,•-,
g -0.04
I---0.08
-0.12 IIIIIII
- •
,,
, , '".,,
.-
A•
,,
'
""v••
•"'•-."
III
A
'-- .-., ,-_ .-, ,--,"-,.I-
t-
I IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII
8O
100
160
0
00
0
140
180
o
•
•
•
•
o
o
• • •
•
0 0
•x x
00
0
•
•,
130•
190
/_•
o
• •x
0 0 0•
0 0 0 O00
•
x
xx
x •x x
210
(,,D
190
x
..• 210
x
x
2,.50
xO
u 250
n 270
x
x
O0
•0
x
0
:
0
x
iii I .... I .... I .... I .... I .... I .... illill .... I .... I .... I .... I .... I .... i i i ii I .... I .... I .... I .... I .... i i ii i i
315
325
335
345
355
365
375
585
395
405
415
oSBN
xSBS
Figure
16. Time
series
ofthedepths
ofselected
isotherms:
atthesouthern
(crosses)
and
northern
(circles)
ends
oftheN2Slineandaveraged
along
theN2Sline(shown
aslines
plotted
I standard
error
above
andbelow
themean);
timeseries
ofaverage
isotherm
depths
along
the
W2E
line are
similar.
NOV
15
20
DEC
25
30
' ' 'l' ' ' ' ' ' ' ' ' L , , , ,
20
,
I,
i
25
i ,i
i
i
"•:•:•:•:•:•
: •:•:•:•:•:•E::•:•½:"
: ' 0 : :;
' " ' '•
JAN
30
i
5
't't
......
1o
•
JAN
, i 30 i , ,
5
i
I
i
FEB
, i i , 1o
i I
i
i
1.5
":":•
2.5.S •1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1 1,1,1,1,1,1,1,,I,1,1,1,1,1,1,1, ,I,1,1,1,1,1,1,1 1,1,1,I1,1
, ,I,l,l,l,l,l,l,l,!l,l,l,l,l,i•
317
322
327
332
337 354
359
364
36•
374
393
398
403
408
Figure17.Latitude-time
distributions
ofthethickness
ofthelayer
between
the23ø0and14ø0
isotherms,
smoothed
withhalfpower
at 5 d•ysand20km.
413
HUYER ET AL.' THERMOHALINE
W9211A
FIELDS NEAR 2øS, 156øE
W9211 B
12,771
W9211C
3O
25 -•
SBN
.(o 20
E
: •
30 •
/•
•J•
........
I,,• .......
I
.....
I .........
I .........
20
I
•o•,,•;,......... ,•:• ...../
30
....
J ..........
•' '• .......
20
......
,..
•••
••
- SBS
15
....
,.........
,.........
34
35
Salinity(psu)
36
34
35
36
34
Salini• (psu)
•ia•re zs. T-S aa•a•t•isti•s at (top) t• no• •nd (S•), (middle)•
(bottom)t•e so•thernend(S•S) of t•e •2S line,for ea• s•rveyperiod.
35
36
Salini• (psu)
midpoint(Z), and
12,772
HUYERET AL.' THERMOHALINEFIELDSNEAR2øS,156øE
Nov 15
Nov22
Dec 1-2
I-,-•
-I
21 I ' •-__•
, ,_•
]._, . '---•
_,I '3
•
-
- 54.8....
'
I
'
][•- --•44•••--'--• - •,.34.4
------,-----.---
] I-
-
,34.6•: •
_
2.0øS
Dec
20
Dec
1.5øS
2.5øS
30
I
2.0øS
1.5øS
Jan
21
8
-•34.2 I•
22
34 7
5.
23
•
34._9 ••-•-
F: 24
25
26
•.•
27
2.5øS
2.0øS
Jan
1.5øS
2.5øS
27
I
I
2.0øS
1.5øS
2.5•S
Feb 5
21
-
,
2.0•S
1.5•S
Feb 14
22
•---------•--• 34.6•
_
4.7
23
9
25
26
27
2.5øS
2.0os
I
1.5os
2.5os
I
2.0os
I
1.5os
2.5os
I
2.0os
1.5øS
Figure19. Salinity
distributions
along
theN2Sline,atthebeginning,
middle,
andendofthe
eachcruise,
withdensity
ratherthandepthasthevertical
axisto remove
fluctuations
dueto
waves.
internal
12,773
HUYERET AL.' THERMOHALINEFIELDSNEAR 2øS,156øE
broadtongueof high-salinity
waterextends
northward COARE IOP. A subsetof individualsectionsfrom the
middle,
andendofeachcruise
suggests
that
across
the equator(Figure2) fromthe SouthPacific beginning,
scales>50 km
Ocean[Tsuchiya
et al., 1989;Richards
andPollard, the largerstructure(with meridional
range>0.5kgm-3) tendsto persist
forat
1991],
though
meanmeridional
velocities
at thatdepth anddensity
have
are small(Figures7 and 10). Boththe shapeand least10 daysbut that smalleror thinnerfeatures
Forexample,
the maintongueofthe
the intensityof the high-salinity
corevariedduringthe shortertimescales.
NOV
15
20
DEC
25
50
20
25
JAN
JAN
50
211,1,1'1'1'1'1'1'1'1'1'1'1'1'1'1'1'1'1'1'1'1
I•'1'1'1'1'1'1'1'1'1•1'
5
'•'1
10
••'"--'---•
Z
5
10
15
I'l'l'l'l'l'l'l'l'l'l'l'l'l'l'l'l'l'l'l'q
•::
[•34.8
FEB
50
:::: -q,•'•'--.-,---,-'-'-•--- t
3"4.:4"
-•-1
•
•.:•
•'•
.......................................................
SBN
E 24
....
26
- 1,1,1,1•1,1,1,1•1•1,1,1,1•1•1,1,1,1,1,1,1
hl,l,l,l•l•l,l,l,l,l•l•l,l,l,l•l•l,l,l,l,l,l,l,I
27
21
..2_3344.34
22
23
E 24
25
26
I'1'1'1 I I I'1'1'1'1'1'1'1'1'1'1'1
I I I I I I I I'1'•1
•-.-..,-1-._•
iii
i i i i i i i i i i i i iI iI iI iI iI iI iI iI iI iI iI'1iI'1
.
/'
.
23
SBS
E 24
25
26
27
317
1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1
522
527
552
557554
559
564
569
574
595
598
405
408
415
Figure20.Variability
ofthesalinity
profile
at(top)thenorth
end(SBN),
(middle)
themidpoint
(I),and(bottom)
southern
end(SBS)
oftheN2Sline.Density
rather
thandepth
isused
asthe
vertical axis to removefluctuationsdue to internal waves.
12,774
HUYER ET AL' THERMOHALINE
FIELDS NEAR 2øS, 156øE
sent at SBS, the southern apex, where salinity almost
salinity maximum (with S>35.5 psu, centeredat pt -
25 kg m-a) wasnearlyunchanged
fromNovember
13 always decreasedmonotonically away from the maxito 22 and remainedsimilar through December2 (Fig- mum, both aboveand below. At the midpoint(I, at
ure 19). Meanwhile,a smallsecondary
salinityinversion lø50'S,156ø06'E),the verticalstructurewasbothmore
seenat crt= 24.2 kg m-a and2.25øSon November
13 complexand more variable; small secondarysalinity indevelopedinto the intense secondarymaximum at 1.5øS versionsemerged and faded, and the crt range of the
on December 1-2; remnants of this or a similar sec- layerwith S >35.5psuvariedfrom< 0.1to > 1 kgm-3
ondarymaximumwereobserved
alongc4=24.2 kgm-a
(Figure 20). Secondarysalinity inversionswere even
on December
more common at SBN, the northern apex, and the layer
20.
There
is substantial
eastward
flow
at this depth (0.2-0.3 m s-1 in November[Lukaset with S >35.5 psuwastransientratherthanpersistent
al., 1995a]),and thusthe observed
changes
may reflect there. At all threelocations,anomalies
tendto migrate
zonal inhomogeneitiesas well as meridionalvariations. upward through the water column in the crt range of
Figure20 shows
that a veryhighsalinitylayer(with 25.5to 23 kg m-a (Figure20): suchupwardmigration
Smax usually>35.6 and always>35.5) wasalwayspre- is particularlyobviousat SBS and I in early January
a)
NOV
DEC
20
25
JAN
30
5
JAN
10
FEB
,7,0
5
10
15
I I I III I I I I I I I I [ I
1.3øS
1.5øS
1.7øS
25.5
1.9øS
2.1øS
2.3øS
,,,,,,,,,,,,,,,,,,,,,,,,,
2.5øS
,:,
1.3øS
1.5øS
1.7øS
24.5
1.9øS
2.1øS
2.3øS
2.5øS
1.3øS
1.5øS
1.7øS
25.5
1.9øS
2.1øS
2.3øS
2.5øS
317
322
527
352
I,Iil,lll,l,l,l,l,l,l,l,l,l,l,l,t,l,l•l•l•l•l•l•l
557 554
559
364
369
374
393
398
405
408
413
Figure 21. Variation of T-S propertieson three isopycnalsin the main thermocline(Pt = 23.5,
24.5, 25.5). Contoursare labeledin units of salinity;corresponding
valuesof temperatureare
givenin Table 5. (a) The latitude-timedistributions
for the N2S line along156ø06•E.(b) The
longitude-time
distributions
for the W2E sectionalong1ø50•S.
12,775
HUYER ET AL.' THERMOHALINE FIELDS NEAR 2øS,156øE
b)
NOV
30
25
20
15
DEC
25
20
30
JAN
5
JAN
30
10
5
FEB
10
15
'1 ':1
','•!...'.l
'1'
I'I,.L.I.L'
I'.I..'...,....I..I..LI..!..I..'..L,.'
.•::;
l.•..l
I..!..!...Li..!..I...[.I...,...
I!I'
I"::::':i
I:::::::::::::::::::
:::::::::::::::::
156.7øE
156.5øE
:
156.5øE
25.5
156.1øE
155.9øE
155.7øE
155.5øE
156.7øE '1 '.1' I '1 'l'l'l'l'l'l
:.
:
156.5øE
156.5øE
24.5
156.1øE
155.9øE
155.7øE
...........
155.5øEI.
'1
II
I I I I I I I I I I I I I I I I11
156.7øE
•,,•:
?.'::.•
..•'•
::•::>•:.•:>..:>.:<:.'
.........
:::::::::::::::::::::::
156.5øE
•i:•:>.•
:: •::...:.M?•:•.:
•
156.3øE
25.5
156.1øE
155.9øE
155.7øE
I
155.5øE
,317
,322
,327
3,32
,337 ,354
,359
,364
,.369
,374
595
598
405
408
415
Figure 21. (continued)
21). A thinintrusion
ofrelatively
freshwater
(d•ys368-375)
•nd •t I •ndSBNin e•rlyFebruary(Figure
centered
at •t =25.5(Figure18)wasobserved
alongthe
northernportionof the N2SlinefromNovember
19to
26,
but
it
did
not
penetrate
as
far
south
as
W2E
line
tempemture-s•linity
characteristics
onisopycn•ls
sp•n(d•ys395-405).
Time-l•titude •nd time-longitudedistributionsof
21aand21b). A fairlythickisohaline
layer
ningthes•linitym•ximum
•reshown
inFigures
21•nd (Figures
centeredat about •t = 24.5 (Figure 18) was present
22;though
contours
•re l•beled
ass•linity,
e•chisoh•thisperiodalongmuchof the W2E line
line•lsorepresents
•n isotherm
(T•ble5). Thethree throughout
and
at
the
northern
end of the N2S line (Figure 21a
isopycn•ls
shown
in Figure21 sp•nmostof thev•rithisrelatively
fresh
•bilityin T-S characteristics
(Figure18);thoseshown and21b). Thefrontseparating
isohaline intrusion from adjacent waters was narrow
in Figure22 lie nearthecoreof them•ximum
(Figure 19).
(10-20
km)andsharp
(• 2x 10-• psum-l);thisfront
In November,
thes•linity•t •t - 25.2w•snearlyuni- seemedto be stationaryduringNovember15-19, mi-
gratingslowlyeastward
duringNovember
20-24,miform(Figures
18•nd22),though
thereweresignificant
grating
rapidly
northward
(at
>0.1
m
s
-•)
beginning
l•temlgradients
•nd temporal
v•ri•bilityboth•bove
29, andpassing
the W2E lineonNovember
•nd belowthishigh-sMinity
core;w•ters•t •t - 23.5 November
30.
A
much
more
diffuse
but otherwisesimilar front
wereMsorelativelyhomogeneous
•t •,24øC,35.15psu
12,776
HUYERET AL.' THERMOHALINE
FIELDSNEAR2øS,156øE
a)
NOV
15
DEC
20
25
30
20
25
'
1.5øS
JAN
30
5
2.'''
':?'
.......
i:,
....................
.•½:l•.&":
: !i
JAN
10
3O
FEB
•o
15
':
! i :::iii'•i
•
24.8
:It
1.3øS
1.5øS
1.7øS
25.0
1.9øS
2.1øS
2.3øS
2.5øS
4
1.3øS
I
I
!
I
I
I
!
I
!
I
I
I
I
: I•l•-•,t::•J:::•.•J•J•:•J
J ..I•4•,I....
...
:
I
I
I
1.5øS
1.7øS
25.2
1.9øS
' ::::::::•i::iiiiiiii]iiiii
::::::::::::::::::::::::::::::::::::::::::::::
:.::•"•':':•:••••
2.1øS
2.3øS
2.5øS
•1,1,1,1,1,1,1,1,1,1,1,1,1•1•1•1,1,1,1,1•1
3 7
322
327
332
Figure
22.Variation
ofT-Sproperties
onisopycnals
near
thecore
ofthesalinity
maximum
(at
- 24.8,
25.0,
25.2).Contours
arelabeled
inunits
ofsalinity;
corresponding
values
oftemperature
aregiven
inTable
5. (a)Thelatitude-time
distributions
fortheN2Slinealong
156ø06•E.
(b)
Thelongitude-time
distributions
fortheW2Esection
along1ø50•S.
wasobserved
onthe•rt- 24.8and25.0isopycnals
(Fig- andmostintenseat this density;aboveandbelowthe
ure22) atthelower
limitofthisintrusion
(Figure
18). coreoftheintrusion,
thefrontintersected
isopycnals
at
Northward migration of the front coincidedwith net anoblique
angle(Figure19). Thisfrontmigrated
southnorthwardcurrentsin the layerbetween110and 180m wardin lateDecember
andnorthward
in earlyJanuary
[Lukaset al., 1995a].
(Figure22a). Typicalmeridional
migration
rateswere
In lateDecember
andearlyJanuary,
waterproper- about10 kmd-x (0.1m s-i), consistent
withtypitieson•rt - 23.5wereagainnearlyuniform,
though
the calmeridional
ADCPcurrents
at thisdepth[Lukas
et
salinitywasslightly
higher(by0.1psu)thanin Novem- al., 1995a].Thefrontalsomigrated
generally
eastward
ber(Figure21). Relatively
freshwateragainintruded (Figure
22b),withtypical
speeds
of20-30kmd-1 (0.2fromthe north(in late December)
intowatersof uni- 0.3m s-i); these
areconsistent
withtheaverage
eastformlyhighsalinity(35.5-35.6psu)between
the•rt - wardvelocities
measured
at thedepthof thesalinity
24.8andat -- 25.5isopycnals,
andwithdrew
againin maximum(Figures7 and10).
earlyJanuary
(Figures
19,21,and22);thefreshest
waDuringmuchof the third surveyperiod,maximum
ter was centeredat •rt - 25.2. The front was narrowest salinities
were>35.5psuoverlargeportions
ofboththe
HUYER ET AL.' THERMOHALINE FIELDS NEAR 2øS,156øE
b)
NOV
15
20
DEC
25
50
20
' i '.1
' I' I:: I'•I' :1'I 'J'
l' I ',1'I.'
I,'l.!J..',œ•.i•.'
I' ]
.•
•:
: : '•::i::i:::
::iii::::•.
,,,•i.'.•
•.½;
156.7øE
dAN
25
50
5
12,777
JAN
10
FEB
50
5
10
15
........................................
i?,iiiiiiiii',i',i',i',iiiii•',:,',ii::::,:.
• i
!.,:•:,ii??i.:•','?,',•:',
:'.'ii;½ll•l
il•11•-•J•.'.½,•l:';.'•:
,,:;•::::•::'":::::::::::::::::::::::::::::::::::::::::::
156.5øE
156.5øE
156.1 øE
•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:•:*:•:•:•:•::
:i•::•::*•.-"•.•.::•:l•'•**•:
=============================
.....•:•:I:::•::•::*:: •::•]:f•
..........
•l:'.•
155.9øE
24.8
:;::::::•::
.....•I•t¾•..
•t•l'•'•i
•::::•::•::
.....................•.:::M:'i•;•
t
155.7øE
I
155.5øE
•i•:[•*•J•2•.•j•]•.•::•::•::•::•::•..ii/:•::•::•::::•::•::::::•:5::•:•:::;•.
: ===================================================================
½
':!ii::iiii!ii!•ii::i::i!i!iii!iii::i.t..iiil
ii!i::i::il::
::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::
===============================
:::::::::::::::::::::::::::::::::::::::::::::::::::::::::
•1' I:.:.:.':,,•.L!.:.:.:';...•
• '.•'• :::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::
156.7øE
156.5'E
156.5øE
156.1OE
25.0
155.9øE
155.7øE
155.5øE
156.7øE
156.5øE
::?:i:::•i•i::i::•::i::•:i::i•i::i::i::iii::•iiii::•::i::i•`i!i::i::•::i::i::i::.[i•::!•i:•i::i:::•:i•::•::i:•i::i•i::i::•:::.::i::i::ill::i::!::!::iii::i::•::•i::i::ii::::•::i::i::ii::i::•::
156.5OE
25.2
156.1 øE
155.9øE
::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::
..... ::::•.'..•::;•
.... •::i::iii::i::i::11
,,,!•
155.7øE
155.5øE
517
522
527
552
557
554
559
564
569
574
595
598
405
408
415
Figure 22. (continued)
Mostof thesesubtlefeatures
extendfor
N2SandW2E sections
(Figures19,21, and22) andover ary4 section.
a largerangeof densities
(at = 24.5to 25.2). Intrusions >40 km, and most lie on a line or curvethat liesoblique
of relativelyfreshwater fromthe north occurredat the to the isopycnals. Features are observedboth above and
beginningand end of this period(in late Januaryand below the maximum salinity, i.e., in the regimethat is
mid-February),andthe frontalstructureandmigration diffusivelystable as well as in the regime that is favorrates were similar to those observed earlier.
able for salt fingering. The generation,evolution, and
The migrationof theserelativelylargewatermassfea- decay of these small features are very likely affected
tures(extending
>20 km laterallyand>20 m vertically, by the complex equatorial flow field, in which internal
or >0.2 kg m-a in density)seems
to be governed
by the tides and other baroclinic waves are superimposedon
ambientcurrents. Smaller,very subtle featurescan also the highly shearedmean currents,but further study of
be coherentin both spaceand time. For example,the these processesis beyond the scopeof this paper.
"waterfall" plot of salinity profilesin Figure 23 shows
that a very small (<0.04 psu) salinityinversionat at
= 21.8 extendingfor >130 km on February4 (from
1.25øto2.4øS)is still present32 hourslater. Other similarly subtle features are spatially coherentover as few
asthreeor four adjacentSeasoarprofiles(i.e., <12 km),
as seenat at = 24.6 just left of the center of the Febru-
6.
Discussion
6.1. The Surface Layer
Sinceall atmosphere-ocean
interactionsare inevitably
mediated through the surface of the ocean, the ocean
surface layer is the focus of special attention within
12,778
HUYER ET AL.' THERMOHALINE
FIELDS NEAR 2øS, 156øE
Table 5. Values of Temperature Correspondingto Specific Values of Salinity on the Isopycna! Surfaces
with err= 23.5, 24.5, 24.8, 25.0, 25.2, and 25.5 kg m-3
Temperature at Isopycnal,øC
$, psu
err -- 23.5
err = 24.5
err= 24.8
err= 25.0
err----25.2
err = 25.5
35.2
35.3
24.98
25.23
21.54
21.82
20.44
20.72
19.68
19.97
18.91
19.20
17.70
18.02
35.4
35.5
35.6
25.48
25.72
25.96
22.09
22.36
22.62
21.00
21.28
21.56
20.26
20.54
20.83
19.50
19.79
20.08
18.32
18.63
18.93
COARE [WebsterandLukas,1992;Tomczak,
1995;You,
1995; Andersonet al., 1996]. Definitionsof this surface layer abound [Brainerdand Gregg,1995; Lukas
and Lindstrom,1991], and its depth is affectedby a
to surface cooling; stirring by the local wind stressor
by shearinstabilities;restratificationby the addition of
cool but fresh water from rainfall; and differential ad-
variety of processes:diurnal heating due to solar radiation; coolingat the surfaceby evaporation,sensibleheat
flux, and long wave radiation; nocturnal convectiondue
fields of temperature and salinity have been presented
and discussedearlier in this paper, in the form of cruise-
vection from/to the surroundingocean. Near-surface
averagedand individualmeridionalsections(Figure 9,
22
23
• 24
25
26
27
21
la ......
I
I
I
35
36
37
I
I
I
I
38
39
40
41
.
22
23
• 24
25
26
27
42
Salinity(psu)
Figure 23. Salinity profiles showing 1-Hz data from adjacent ascendingSeasoarprofiles from
consecutive
sectionsalong156ø06•E.Sectionswererun from lø14tS (on the right) to 2ø26•S(on
the left), beginning(top) at 0604 UT on February4 and (bottom) at 1425 UT on February5.
Axis labelsapply to the northernmostprofile;successive
profilesare offsetby 0.1 psu for display
and are usually separatedby <2.5 km.
HUYER ET AL.' THERMOHALINE
Plates 1 and 2), time-depthdistributionsfor the intersection(Figure 12), time-distancedistributionsfor the
20-dbarsurface(Figures13aand 13b), and time series
of laterally averagedstatistics(Figure14), but we have
FIELDS NEAR 2øS, 156øE
12,779
nal is nearlycoincidentwith the 28øCisotherm(Figure
15). The second,Ht, estimatesthe depthof the "top
of the thermocline" from a critical gradient criterion
of 0.05øCm-1 but ignoringthe top 20 m [Lukasand
Lindstrom,1991]. For the remaininglayers,definedin
not yet discussedthe structure of the mixed layer per
se. Here we explicitly considerthe vertical structure of
the near-surfacelayer, by estimating the surface layer
depths using several different criteria.
Surface "mixed" layer depth can be defined in terms
terms of differencesfrom the surface, we used the shal-
lowest gridded value above 10 m to representthe surface value; it was usually centered at 0.5 or 2 m. The
"isopycnallayer depth," HD, is defined as the depth
at which the potential density first exceedsthe surface
of a finite changefrom valuesat the surface[e.g.,Andersonet al., 1996; Wijesekeraand Gregg,1996; $myth
et al., 1996b],in termsof a criticalgradient[e.g.,Lukas
and Lindstrom,1991], or even in terms of a specific
isopycnal[e.g., $myth et al., 1996b]. We estimate
"mixed layer" depths from the 2-dbar, 2-nm gridded
Seasoardata using all three of these generalclasses.
The first of these, H22, definedby the depth of the as
densityby Ape ----0.01kg m-3. Thislayeris sometimes
calledthe "diurnalmixed layer" [$mythet al., 1996b].
The "isothermallayer depth", HT, and "isohalinelayer
depth", H$, are definedseparately,by AT - -0.030 ø C
- c•-lA•e and AS -- 0.013psu - -/•-iA•,
respec-
tively, where c• and /• are the thermal and haline expansioncoefficientsat 28øC and 35 psu. FollowingAn-
= 22.0 kg m-3 isopycnal,corresponds
to the "upper dersonet al. [1996],we calculatedHT (and Hs) by
oceanlayer" usedby Smythet al. [1996b]to repre- findingthe maximumtemperature(minimumsalinity)
sentthe regionabovethe main pycnocline;this isopyc- in the top 20 m and starting the differencingfrom that
NOV
15
20
DEC
25
50
20
JAN
25
50
JAN
5
10
FEB
50
5
10
15
,, .
e "'..'. , :-.
', , ,: ? ß e:
I
I
I
I
I
I
1
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'•IIIIIIIIIIIIIIIIt
.....
, ;•..-/,,'../..."./'
20 '"'
,,...
'...
:. ,,..., ..' -•,
..
,,
, , •
,
.:.• ,',:'
,
HD
Isopycnal
... ,,j •
60
8o
100
ß 6 pm to 8 om
1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1•1
1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1
20''•
'"'
•'
''•''•'
'L'
', "...
',...
40
60
'.... •
I
m ' ,' •'
• ' ?
...
•.." ,"
:,
HT
:"'• , ,'
Isothermal
o 8 om to 6 pm
8o ß6,pmto•,I,l,l,l,l,l,l,l,l,l,l,I
om
'b....
100 1,1,1, ,I,l,l,I,
1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,
20
,
-
-',
I,l,l,l,l,l,l,l,l,l,I
/:"'.:,,
,
I
ß
.,.,
Hs
Isohaline
60
80
I
0
ß
20
40
80
ß.
....
'"'",,,--'",
•.',
60
......... . ß
I Ht'
Top
of
ß
100 I
I,I I
I I,
I
•
Therm
0
20
H22
40
60 •
80
....100
317
.-
', ,,,
:.
,'
I•1 I I I I I I I I I I I I I I I I I I,I I
322
•27
332
3•7
•54
359
•64
•.,...,
•
•69
....'.•.%..,.:...•
22.0
kg/m
'"'
•'"f... I I,
374
I
393
,
398
403
408
413
Figure 24. Time seriesof surfacelayer depthsat three locations(SBN (dotted lines), SBS
(dashedlines) and I (solidlines)), estimatedfrom 2-dbargriddedSeasoardata usingcriteria
specified in the text. Symbols are used to distinguish daytime and nighttime estimates of H•
and H•.
12,780
HUYER ET AL.' THERMOHALINE FIELDS NEAR 2øS,156øE
point downward to prevent near-surfaceinversionsfrom
influencing our estimates.
Estimates of these surfacelayer depthswere made for
all of the complete meridional and zonal Seasoarsections. Figure 24 showstime seriesof the values at the
January 4-6, after the cessationof the strong westerly
winds and while the overlying water is restratified by
rainfall(Figures24 and25). The top of the thermocline
is always deeper than the surface "isohaline"layer, H$
(Figure24), but the "barrierlayer" [Lukasand Lind< 10m
intersectionpoint (I) and at the northernand south- strom,1991]that liesbetweenthemissometimes
ern apices(SBN and SBS).In general,eachlayershows thick. By definition,the isopycnallayer, HD, is equalto
similar behavior at all three locations. HD and HT are
or lessthan H$ and HT. The "isothermal"layer, HT,
is often deeper than H$, presumably becausesubsequent precipitation has overlain a remnant mixed layer;
it shouldbe noted that theselayersneednot be strictly
homogenous. The frequent occurrencesof HT > H$
suggestthat at least the upper portion of the barrier
of HD and HT to occur during daytime and deeperval- layer is transient or patchy in nature. Nevertheless,HT
uesto occur at night, but exceptionsare not uncommon approacheswithin 10 m of Ht on only a few brief oc-
subjectto strongdiurnalvariabilityasexpected[Anderson et al., 1996],exceptduringperiodsof weakwinds
combinedwith either sustainedheating(November15
to 22, November29 to December1) or heavyprecipitation (Jan 4-10). There is a tendencyfor smallervalues
(Figure 24). H$ tendsto shoalduringthe November casions(December20-22, December25-28, December
survey period and to deepen during the westerly wind 31, February10-11) of strongwesterlywinds,and HT
burst in late December, before shoalingabruptly in re- and Ht do not coincideon any occasion.Thus the lower
sponse to heavy rainfall in early January; it deepens portion of the barrier layer effectively insulated the sea
rapidlyin unisonwith Ht and H22 (Figure24) during surface from direct exchangewith the cooler waters in
January 27-31, apparently downwellingin responseto the thermoclinethroughout the COARE IOP.
the renewedwesterlywind stress(Figure14). Thereis
a tendencyfor Ht andH22 (andthe 28øCisotherm,Fig6.2. Lateral Homogenization
ure 16) to be deeperin the north, particularlyduring
In section 5.1. we noted that property distribution
our November survey period and during the westerly
wind burst in late December.
on the 20-dbarsurface(Figure13) suggested
an associ-
At all three locations, the top of the thermocline, Ht,
ation betweenthe degreeof patchinessand the strength
liesabovethe 22.0 kg m-3 isopycnal
(Figures24 and of the wind: standard deviations along each section
Figure 25). Their separationis <10 m duringmuchof were relatively large during the weak winds of Novemour secondsurvey period, during and after the strong ber and early January and small during the westerly
westerly wind burst. The two convergecompletelyon wind burst in December and the squallsin early FebruNOV
DEC
20
25
50
20
I
0.16
IIIIIIIIIIIII
0.08
0.00
-0.08
I
JAN
25
I
I
I
I
I
50
I
•1!
I
I
I
I
JAN
5
I
I
I
I
I
I
10
I
I
I
I
I
[
FEB
50
5
10
15
I
/ xx,._......_l/!
• - .,,...__......,•.•x...,,,/____.
....
40.00
20.00
o.oo
lOO.
o.
-lOO.
'
2O
., ," "i :\."i I ".:
.'..' .' """'
,'-',
.'',,'
t/I•..".:
".:
:i '
,'
4O
.:
6O
....?-'/'•,_,,',, ,, ','.,.,.-/,
. ,'
i
"
.,..;•.
,'
/
/
•'--'
':.
•i i!'..'.;;;•
*, !•
•
•/t
;'
, ,'
,,,,,,
',I•'A','/
,•,
;;:;
t :,. :i
.": ;:'.
;.. .'•T
..',,"'
..':
.':' ':: '.,!:.'>.',,
,: , ; :' ':'':, :: ':
•:...., .•........
:,.' ..'
Ht,'
,,'- .....
,,,'-....'
..
•..
. .
: •.
,,.
•._.,
' :'" ","" ....
'
8O
1,1•!,1,1,1,1,1,I,1,1,1,1,1,1,1,1,1,1,1,1
lOO
517
322
527
1,1,1,1,1,1,1,1,1,1,1,
352
357
554
559
1,1,1,1,1,1,1,1,1,1,1•1•1•
564
569
1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,
374
393
398
403
408
Figure 25. Variationof surfacelayerdepthsat the intersection
I (1ø50/S,156ø06/E),estimated
from 2-dbar gridded Seasoardata; time seriesof wind stress,rainfall, and heat flux are repeated
from Figure 12.
413
HUYER ET AL.' THERMOHALINE FIELDS NEAR 2øS, 156øE
12,781
aw (Figure14). Although
thisisobardoesnot always ual patchinessin both zonal and meridional directions
lie within the surfacemixed layer as defined by either
(Figure26). Whetherthis homogenization
resultsfrom
Ho or HT (Figures24 and 25), it is usuallyshallower large-scale accumulation of surface water and broad
than the daily maximum[Andersonet al., 1996, Fig- equatorial downwellingdriven by the westerly winds or
ure 6], and thus it represents,at least approximately, whether it is the result of lateral mixing by small-scale
the nocturnal surface mixed layer. The apparent rela- processessuch as local convergencesdriven by squalls
tion betweenthe degreeof patchinessand the strength is not clear.
of the wind (Figure 14) is verifiedby scatterdiagrams
6.3. The Thermocline,
(Figure26), whichshowlargerspatialvarianceto be Maximum
EUC, and Salinity
associatedwith weak zonal wind stress;similar results
are obtained for the meridional
wind stress and for the
The structure and properties of the main thermocline
wind stressmagnitude,sincewindswerepredominantly were surprisinglyvariable, on timescalesranging from
from the northwestthroughoutthe IOP. Strongwinds days to weeks and months. Some of the variability in
seem to reduce meridional
....
I
trends as well as the resid-
....
I
....
I
.
.
.
,
I
,
,
the upper thermocline was related to variations in the
,
....
+
0.20
I
I
I
I
0.20
-I-_
I_
x
x
x
N2S
x
0.10
x
-I-
I
....
I
+
o
0.00
,
....
N2S
x•x•
+ xß % +
o•o+ ee
x•;ooo o+ ooo o
0.00
x
x
0.10
....
I
....
I
....
....
i
....
i
....
i
....
I
....
0.20
0.20+
+
•
0.10-
W2E
x•+,
>
+
0.10
xXx X X
+
I
....
I
+
o -t-+ +
o
....
....
I
....
i
0.00
....
0.15
"-o
0.15
t+
0.10
W2E
X_FX+
XX
øøøedroø
0.00
x
xx xx
N2S
+
+ ø•1-"•ø•o
o X•
....
I
....
I
....
....
I
....
I
,
+
I
,
,
,
....
I
I
I
N2S
0.10
ß
....
x
x
-o
0.05
+
Xx
o
0.05
X Xf:>øøo
+
+
0.00
0.15
x
0.10-
x
)X
X +
+
x
W2E
0.00
....
0.15
....
X
+
o
o
X•.,,•o
o+ee+
o
I
I
....
....
I
I
....
+
I
I
....
I
,
,
,
0.10
ß
X
x
0.05
+
0.05
x
x +
W2E
x
..•++o+
0.00
....
I
0.0
....
I
....
I
....
0.1
Ave Tau-X (Pa)
I
0.00
..........
0.2
I
0.0
....
I
....
I
0.1
0.2
Ave Tau-X (Pa)
Figure 26. Lateralstandard
deviations
of 20-dbartemperature
andsalinityversus
the average
zonalwind stressfor the preceding24-hourperiod;resultsare shownfor both observedvaluesand
valuesdetrendedby linearregression
versuslatitudeor longitude.Differentsymbolsdistinguish
data fromthe threesurveyperiods'W9211A(crosses),
W9211B(pluses)andW9211C(circles).
12,782
HUYER ET AL.: THERMOHALINE
FIELDS NEAR 2øS, 156øE
zonal wind stress, but we also observedvariations that under calm winds in November, and salinity stratifiwere not related to local winds; this is hardly surpris- cation was strongunder calm winds and heavyrainfall
ing, since long equatorial waves are known to be effi- in early January; both thermal and haline stratificacientconveyors
of signalsfrom the centralPacific[e.g., tion were weak during the westerlywind burst in late
Kesslerand McPhaden,1995]. Someof the variations, December and the moderate winds in February. Une.g., the degree of equatorward spreadingbetweenthe expectedly,we also observedsomemesoscalefeatures
upper and lowerthermocline,are relatedto the strength with strong lateral gradientsin the near-surfacelayer:
and position of the Equatorial Undercurrent, and the mostnotably,a zonaltemperaturefront (with AT of
undercurrentis clearly important in advectingwaters -•0.2øC in --10 km) that migratednorthwardduring
eastward in and along the core of the salinity maxi- the westerlywind burst in late Decemberand a small
cold
mum: we do not yet fully understand the dynamics patchor narrowband(--40 km wide)of unusually
water
(•28.5øC)
that
migrated
slowly
northward
durwhich determine the depth and the temperature of the
core of the undercurrent, nor do we know what governs ing January5-9, the nature and originof thesefeatures
is still not fully understood. Also unexpectedwas the
the properties of its sourcewaters.
observationthat strong winds seemedto causelateral
homogenizationof near-surfacewaters as well as vertical mixing.
7. Summary and Conclusions
Within the main thermocline, isotherm depths varThe COARE survey cruiseshave provideda large set ied from section to section and from cruise to cruise.
of repeated 130-km meridional and zonal sections at High-frequencyinternalwavesand tides,thoughnot rethe center of the COARE intensive flux array. These solvedby our samplingpattern, were manifest as noise
sectionshave fine vertical (--2 dbar) and horizontal with vertical displacementsof --10 m. Cruise-averaged
(--2 nm) resolutionandextendto a depthof •280 dbar. sectionsshow that both the equatorward spreadingof
Observationswere repeated at intervalsof --2 days dur- isothermsand the Equatorial Undercurrentshoaledsiging three 20-day periods,spanningmostof the 4-month nificantly during the 100-dayobservationperiod;while
COARE intensive observation period from November the core layer shoaled from --220 to --160 m, its tem1992 to February 1993.
perature increasedfrom -•14øto -•20øC. The separation
Our time-averaged fields show a number of features between the 23øC and 14øC isotherms in the upper
that are typical of the westernequatorialPacificOcean: and lower thermocline spreading seemed to be greata very warm and relatively freshsurfacelayer,with tem- est (• 120 m) duringperiodswhenthe EUC xvasstrong
peratures >28.5øC and salinity decreasingnorthward; (•0.4 m s-i). Watermasscharacteristics
of the salina very sharp thermocline with temperature decreasing ity maximum within the thermocline were surprisir•gly
from •-28øC at --50 m to •-12øC at •-250 m and invariable: maximum salinity values ranged from --35.2
termediate isotherms diverging northward toward the to --35.7 psu. The front between high-salinity southequator to envelopthe core of the Equatorial Undercur- ern and low-salinity northern waters was often narrow
rent at •-175 m; a subsurfacesalinity maximum, con- (--30 km), and it migratedmeridionallybetweenand
sisting of a northward tending tongue of high-salinity within cruises;this front was also observedto migrate
South Pacific water; and a deep vertically mixed layer, eastward at a speed consistentwith advectionby the
--50 m thick, with a temperature of --12øC and salinity Equatorial Undercurrent. Secondarysalinity inversions
of 34.8 psu, correspondingto the equatorialthermostad. with lateral scales of 10-50 km were observed both
The duration of the COARE
Intensive Observation
aboveand below the high-salinitycore;thesewere easPeriod was exceptionally long for an intensive study ily recognizablein sectionsa few days apart but did not
of the upper ocean, but this was clearly appropriate persist from cruise to cruise.
This large, high-resolutionSeasoardata set shows
given the lack of stationarity in both atmosphericand
oceanicregimes. In general, near-surfacetemperature that the thermohaline fields in the COARE intensive
respondedas expected from the variations in the lo- flux array are both heterogeneousand nonstationary.
cal winds: general heating during the calm winds in Even within the broadly uniform Warm Pool of the
November, rapid coolingduring the westerlywind burst western equatorial Pacific, there are instancesof mesoin late December, warming during weak easterly winds scale and submesoscalefronts in the surface layer. We
in January, and gradual coolingduring moderatewest- look forward to combiningthis data set with observaerlies in February. Variations in the depth of the 28øC tions from other survey shipsand with time seriesfrom
isothermandthe 22.0kgm-3 isopycnal
at thetopofthe stationary platforms. Work to estimate advectiveheat
thermocline also varied with the wind: deepeningdur- and freshwaterfluxes and three-dimensionalbudgetsof
ing the westerlywind burst, shoalingduring the easter- heat, freshwater, and momentum and to study the evolies in January, and deepeningrapidly with the resump- lution of particular features and phenomenais undertion of westerly winds in early February. Vertical strat- way. This large data set will provide many opportuniification within the near-surfacelayer varied with wind ties to improve communityunderstandingof the upper
stress and rainfall: thermal stratification was strong ocean dynamics in this region.
HUYER ET AL.: THERMOHALINE
Acknowledgments.
We are deeply indebted to
COARE Surveys co-investigatorsClayton Paulson and Eric
Firing, who participated fully in the cruise planning and
data collection.
We are also indebted
to Wecoma's
Marine
FIELDS NEAR 2øS, 156øE
12,783
pp. 36-55, Ocean ResearchInstitute, University of Tokyo,
Tokyo, 1993.
Kessler, W. S., and M. J. McPhaden, Oceanic equatorial
waves and the 1991-3 E1 Nifio, J. Clim., 8, 1757-1774,
1995.
Technicians(Marc Willis, Brian Wendler, Mike Hill, and
Tim Holt) for the successful
operationof the Seasoarsystem Kutsuwada, K., and H. Inaba, Year-long measurementsof
upper-ocean currents in the western equatorial Pacific by
acoustic Doppler current profilers, J. Meteorol. $oc. Jpn,
73, 399-409, 1995.
Lueck, R., Thermal inertia of conductivity cells: Theory, J.
Atmos. Oceanic Techol., 7, 741-755, 1990.
Weller and Melora Park (WHOI) and by Clayton Paulson Lueck, R., and J. J. Picklo, Thermal inertia of conductivity cells: Observations with a Sea-Bird cell, J. Atmos.
and Lynne deWitt (OSU). We are gratefulfor frequentdisOceanic Technol., 7, 756-768, 1990.
cussionswith Eric Antonissen,Bill Smyth, Clayton Paulson,
and Hemantha Wijesekera. This work was supported by the Lukas, R., and P. Hacker, Spinup of a submesoscaleeddy in
Ocean Sciences Division of the National Science Foundation
the TOGA COARE intensive flux array during the spindown of an eastward jet, paper presentedat the TOGA 95,
and by the NOAA Office under TOGA through NSF grants
OCE-9113510
and OCE-9113948.
International Science Conference, World Meteorol. Organ., Melbourne, Australia, April 1995.
Lukas, R., and E. Lindstrom, The mixed layer of the western
equatorial Pacific Ocean, J. Geophys. Res., 96, suppl.,
References
3343-3357, 1991.
Anderson, S. P., R. A. Weller, and R. B. Lukas, Surface Lukas, R., P. Hacker, M. Mao, E. Firing, A. Huyer, and
buoyancy forcing and the mixed layer of the western PaP.M. Kosro, Upper ocean currents in the inter•sive flux
cific Warm Pool: Observations and 1-D model results, J.
array during TOGA COARE, in Proceedingsof TOGA 95
Clim., 9, 3056-3085, 1996.
Int'l Sci. Conf., April 1995, WRCP-91-WMO/TD 717,
Brainerd, K. E., and M. C. Gregg, Surfacemixed and mixing
vol.l, pp. 542-546, World Meteorol. Organ., Geneva,
layer depths, Deep Sea Res., Part I, •2, 1521-1543, 1995.
Switzerland, 1995a.
Eldin, G, T. Delcroix, C. Henin, K. Richards, Y. Du Pen- Lukas, R., P. J. Webster, M. Ji, and A. Leetmaa, The largehoat, J. Picaut, and P. Rual, Large-scale current and
scale context for the TOGA Coupled Ocean-Atmosphere
thermohalinestructuresalong 156øE during the COARE
ResponseExperiment, Meteorok Atmos. Phys., 56, 3-16,
1995b.
intensive observation period, Geophys. Res. Left., 21,
2681-2684, 1994.
McPhaden, M. J., TOGA-TAO and the 1991-93 E1 NifioEriksen, C. C., A. J. Plueddemann, and R. A. Weller, AdSouthern Oscillation event, Oceanography,6, 36-44, 1993.
vection and local changeof temperature and salinity in the McPhaden, M. J., Ocean-atmosphere variability observed
COARE Intensive Flux Array estimated from a 3-mooring
from the TOGA-TAO array, in Proceedingsof TOGA 95
array, in Summary Report of the TOGA COARE InterInt'l Sci. Conf., April 1995, WRCP-91 - WMO/TD 717,
national Data Workshop,Toulouse, France, 2-11 August
vol. 1, pp. 181-185, World Meteorol. Organ, Geneva,
1994, pp. 100, TOGA COARE International Project OfSwitzerland, 1995.
rice, Univ. Corp. for Atmos. Res., Boulder, Colo., 1995a. McPhaden, M. J., S. P. Hayes, L. J. Mangum, and J. M.
Eriksen, C. C., A. J. Plueddemann, and R. A. Weller, Tidal
Toole, Variability in the western equatorial Pacific Ocean
prediction of current and vertical displacement in the
during the 1986-87 E1 Nifio-Southern Oscillation event, J.
COARE IFA estimated from a 3-mooring array, in SumPhys. Oceanogr., 20, 172-190, 1990.
mary Report of the TOGA COARE International Data McPhaden, M. J., F. Bahr, Y. du Penhoat, E. Firing, S. P.
Workshop,Toulouse, France, 2-11 August 1994, pp. 101,
Hayes, P. P. Niiler, P. L. Richardson, and J. M. Toole,
TOGA COARE International Project Office, Univ. Corp.
The response of the western equatorial Pacific Ocean to
for Atmos. Res., Boulder, Colo., 1995b.
westerly wind bursts during November 1989 to January
Gill, A. E., Atmosphere-OceanDynamics, Academic, San
1990, J. Geophys. Res., 97, 14,289-14,303, 1992.
Diego, Calif., 1982.
Morison, J., R. Andersen, N. Larson, E. D'Asaro, and T.
Halpert, M. S., G. D. Bell, V. E. Kousky, and C. F. RoBoyd, The correction for thermal lag effects in Sea-Bird
pelewski(Eds.), Fifth Annual ClimateAssessment
1993,
CTD data, J. Atmos. Oceanic Technol., 11, 1151-1164,
111 pp., Natl. Oceanic and Atmos. Admin., U.S. Dep. of
1994.
Commer., Clim. Anal. Cent., Washington, D.C., 1994.
Moum,
J. N., and D. R. Caldwell, Experiment explores the
Huyer, A., P.M. Kosro, R. O'Malley, and J. Fleischbein,
dynamics
of oceanmixing, Eos Trans.A GU, 75(•2), 489,
SEASOAR and CTD observationsduring a COARE Sur490, 495, 1994.
veys cruise, W9211C, 22 January to 22 February 1993,
Data Rep. 15•, Reft 93-2, 325 pp., Coll. of Oceanic and O'Malley, R., P.M. Kosro, R. Lukas, and A. Huyer, Seasoar
observationsduring a COARE Surveys cruise, W9211B,
Atmos. Sci., Oreg. State Univ., Corvallis, 1993.
12 December 1992 to 16 January 199.3, Data Rep. 156,
Huyer, A., P. Hacker, P.M. Kosro, J. Fleischbein, E. AnReft 9•-2, 426 pp., Coll. of Oceanic and Atmos. Sci.,
tonissen,and R. O'Malley, SEASOAR and CTD observaOreg. State Univ., 1994.
tions during a COARE Surveyscruise,W9211A, 8 November to 8 December 1992, Data Report 155, Ref. 9d-1, 364 Parsons, D., et al., The integrated sounding system: Descriptions and preliminary observations from TOGApp., Coll.of Oceanic and Atmos. Sci., Oreg. State Univ.,
COARE, Bull. Am. Meteorok $oc., 75, 553-567, 1994.
Corvallis, 1994.
Kaneko, A., N. Gohda, H. Nakajima, T. Itoh, K. Kawatate, Plueddemann, A. J., and R. A. Weller, Observationsof the
upper-ocean velocity structure from the central mooring
T. Nagahama, and H. Furukawa, Profiling measurement
of the TOGA COARE IFA (abstract), Eos Trans. A GU,
of the equatorial currents by an ADCP on the mooring
75(3), OceanSci. Meet. Suppl., 185, 1994.
lines, in Preliminary report of the Hakuho Maru Cruise
KH-92-5 (TOGA/COARE Cruise), edited by T. Asai, Pollard, R., Frontal surveys with a towed profiling con-
and to Robert O'Malley and Jane Fleischbein for archiving
and processingthe data and assistancein analysis. We are
also grateful to every cruise participant and to the officers
and crew of Wecoma. ProcessedADCP data were provided
by Eric Firing. Meteorologicaldata were provided by Robert
12,784
HUYER ET AL.: THERMOHALINE FIELDS NEAR 2øS, 156øE
ductivity/temperature/depthmeasurement
package(SeaSoar), Nature, 232, 433-435, 1986.
Richards, K. J., and R. T. Pollard, Structure of the upper
ocean in the western equatorial Pacific, Nature, 350, 4850, 1991.
Sea-Bird Electronics, Inc., Bellevue, Wash., CTD Data Acquisition Software, SEASOFT Version 4.015, 1992.
Smyth, W. D., D. Hebert, and J. N. Moum, Local ocean
responseto a multiphase westerly windburst, 1, Dynamic
response, J. Geophys. Res., 101, 22,495-22,512, 1996a.
Smyth, W. D., D. Hebert, and J. N. Moum, Local ocean
response to a multiphase westerly windburst, 2, Thermal
and fleshwater responses, J. Geophys. Res., 101, 22,51322,533, 1996b.
Tomczak, M., Salinity variability in the surface layer of the
tropical western Pacific Ocean, J. Geophys. Res., 100,
20,499-20,512, 1995.
Toole, J. M., E. Zou, and R. C. Millard, On the circulation
of upper waters in the western equatorial Pacific Ocean,
Deep Sea Res., Part A, 35, 1451-1482, 1988.
Tsuchiya,M., The origin of the Pacificequatorial13øCwater, J. Phys. Oceanogr., 11, 794-812, 1981.
Tsuchiya, M., R. Lukas, R. A. Fine, E. Firing, and E. Lindstrom, Source waters of the Pacific Equatorial Undercurrent, Progr. Oceanogr., 23, 101-147, 1989.
Webster, P. J., and R. Lukas, TOGA COARE: The Coupled
Ocean-AtmosphereResponseExperiment, Bull. Amer.
Meteorol. Soc., 73, 1378-1416, 1992.
Weller, R. A., and S. P. Anderson, Surfacemeteorologyand
air-sea fluxes in the western equatorial Pacific warm pool
during the TOGA Coupled Ocean-AtmosphereResponse
Experiment, J. Clim., 9, 1959-1990, 1996.
Wijesekera, H. W., and M. G. Gregg, Surface layer response
to weak winds, westerly bursts, and rain squalls in the
western Pacific Warm Pool, J. Geophys. Res., 101, 977997, 1996.
You, Y., Salinity variability and its role in the barrier layer
formation during TOGA-COARE, J. Phys. Oceanogr.,
25, 2778-2807, 1995.
A. Huyer and P.M. Kosro, Collegeof Oceanicand AtmosphericSciences,OregonState University,104 Ocean Administration Building, Corvallis, OR 97331-5503. (email:
ahuyer@oce.orst.edu;
kosro@oce.orst.edu)
R. Lukas and P. Hacker, School of Ocean and Earth
Science and Technology, University of Hawaii, Hon-
olulu, HI 96822. (email: phacker@iniki.soest.hawaii.edu;
rlukas@iniki.so
est.hawaii.edu)
(ReceivedJune 28, 1996; revisedNovember21, 1996;
acceptedDecember5, 1996.)