Estimation of Surface Winds From Upward Looking Acoustic Doppler Current Profilers

advertisement
JOURNAL OF GEOPHYSICAL
RESEARCH,
VOL. 97, NO. Cll, PAGES 17,925-17,930, NOVEMBER
15, 1992
Estimationof SurfaceWinds From Upward Looking
AcousticDoppler CurrentProfilers
J. BROWN
l, E.D. BARTON,
A. TRASVIlqA
2 ANDH.S. VELEZ
Schoolof OceanSciences,Universityof Wales,Bangor
Menai Bridge,Gwynedd Wales
P.M.
KOSRO AND R.L.
SMITH
Collegeof Oceanography,
OregonState University,Corvallis
ThreeupwardlookingacousticDopplercurrentprofilers(ADCP) weredeployedbeneathmeteorological
buoys
in the Gulf of Tehuantepec,Mexico, duringwinter 1988-1989. Hourly averagedwind speeddatafrom the buoys
and from ship when in the vicinity were comparedwith surfaceacousticbackscatterintensity recordedat the
ADCPs. The backscatterwas found to be a significantpredictorof wind speedsfrom both buoy and ship, the
latter when within 50 km of the mootingsite. There wasno apparentsaturationof the backscattersignal at the
maximumwind speeds(<15 m s-t). The resultscastdoubton the abilityof near-surface
Dopplerdirectional
informationto providereliable estimatesof wind direction.
1. INTRODUCTION
Moored acousticDoppler currentprofilers(ADCPs) providea
powerful techniqueby which to routinely measurethe velocity
fields in a diversityof marinesituations[e.g., Schottam:lJohns,
1987; Derecki and Quinn, 1987; Johns, 1988; Schottand Leaman,
1991]. Despitethis versatility,when the air-water interfacefalls
within rangeof the instrument,the near-surfaceinformationis of
limited value.
transientwindsof limited fetchblowingfrom the land. The in situ
meteorologicalbuoy records,absentin the Gulf of Lions, were
curtailedby buoy and instrumentfailure and the ADCP records
offeredthe potentialto constructthe full wind velocitytime series.
While it provedpossibleto determineaccuratelythe wind speed
from the ADCP, on time scalesof the orderof 1 hour, the ability
to determinedirectionalinformationmustbe opento question.
There exists a surface "shadow" zone of vertical
2. THE OBSERVATIONS
extent
D (1 - cos q>)
where D -- depthof the ADCP transducerand q>= angle of the
transmittedpulserelativeto the vertical. Energyreturningto the
transducerfrom the strong sea surface echo coincideswith the
signal from the scattererswithin the shadow zone and swamps
suppressionof the sidelobe against the main beam [RD
Instruments,1989]. Consequently,
within this zonederivedwater
velocitiesare generallydiscarded. Schott[1989], employingthe
well-establisheddependenceof surfacebackscatterenergy on
wind speed [e.g., Apel, 1987; Urick, 1975], has indicatedthat
useful information
can be extracted
from the near-surface
ADCP
signals. Suchdata, recordedat two ADCP mooringsdeployedin
the Gulf of Lions, were foundto be significantlycorrelatedwith
12 hourly averagesof ship boardwind speedrecordedwithin a
radiusof 100 kin. Additionally,directionalinformationobtained
from near-surfaceADCP velocity measurements
in the shadow
zone was closelyrelatedto wind directionover portionsof the
time series.
During December1988 to February1989 a studyof the response
of the Gulf of Tehuantepec,
situatedoff the southwestPacificcoast
of Mexico (Figure 1), to intermittentwind events,known locally
as "Northers"was carried out [Trasvina, 1991; E.D. Barton et al.,
Supersquirt:The Dynamicsof the Gulf of Tehuantepec,Mexico,
submittedto Oceanography,1992]. The winds, first detailedby
Hurd [ 1929], predominantlyoccurduringOctoberto March when
incursionsof high-pressure
polar air into the Gulf of Mexico drive
a seriesof fiercewind jets through the Chivela Pass,a break in
the northwest-southeast
mountainrangeof CentralAmerica,many
hundredsof kilometersinto the lower pressureregime over the
Pacific. The typical durationof a wind pulse is 2-5 days with
velocitiesof 10-30 m s-l. In the centralGulf, strongsurface
circulations,
of order1.2m s-l, andintense
cooling,by asmuchas
15øC with respect to the surrounding Pacific, are induced
[Trasvina, 1991; Roden, 1961].
A major component of the work was a comprehensive
hydrographicsurveyof the Gulf from January15 to February9
undertaken aboard the R V Wecoma, equipped with a full
meteorological
suite. To assistin definingthe lateralextentof the
wind jet and determinethe shearacrossit, three meteorological
buoys,recordingwind speedand directionapproximately1.3 m
In this notewe considerthe applicationof this techniqueto data
collected in the vicinity of three meteorologicalbuoy/ADCP
mooring pairs deployed in the Gulf of Tehuantepec,Mexico
(Figure 1). This aspectof the studycan be consideredanalogous above the nominal sea surface, were moored in 4000 m of water
to that of Schott [1989] in that both regions experiencestrong perpendicularto the axis of the wind (positions 14ø32.1'N
'Now at M.A.F.F.,Fisheries
Laboratory,
Lowestoft,Suffolk,England.
:Nowat CentrodeInvestigacion
Cientifica
y Educacion
Superior
de
Ensenada,Ensenada, British Colombia, Mexico.
Copyright1992by the AmericanGeophysical
Union.
Papernumber92JC01481.
0148-0227/92/92JC-01481
$05.00
96ø10.8'W (M1), 14ø15.8'N 95ø15.7'W (M2), and 13ø44.5'N
94ø22.3'W (M3)) (Figure 1). M1 was lost; M2 collectedwind
speedand directionfor 74 hours,but sankduringthe strongwind
event;M3 collectedwind speedfor 260 hours,but the direction
sensormalfunctioned
at the startof the experiment.
Situatedbeloweachbuoy on a separatemooringwas an upward
looking 307.2-kHz ADCP configuredfor a 30ø acoustic beam
angle, RDI serial numbers197 (M1), 201 (M2), and 209 (M3).
17,925
17,926
BROWNETAL.:UPWARD
LOOKING
ACOUSTIC
DOPPLER
CURRENT
PROFILERS
100øW
21 Jan
0000
90øW
• Wecoma
1989
UT
i
i
,
I
i
b) Met. M2
15
20ON
100
c) ADCP M2
Gulf
of
,
Tehuantepec
,
,
,
,
,
,
i
,
,
,
M1 e ß
,
I
i
T
,
,
d) Met. M3
15
M2 M3
,
lOON
,
i
,
!
,
,
i
i
,
•
100
e) ADCP M3
)
A
'=
I
i
,
,
,
,
,
,
,
,
,
,
,
,
,
,
Pass
km
A
100
,
Chivela
-
i
98
B
f) ADCP M1
I
96
16
17
18
19
,
,
I
,
!
I
,
,
,
i
•
i
i
i
I
20
21
22
23
24
25
26
27
28
29
30
31
I
2
3
Januaryto February1989 (UT)
Fig. 1. Pressure
regimethatgenerated
theNortherof January
21-23, 1989,
andtherelativepositions
of themeteorological
buoy/ADCP
mooring
pairs Fig.2. Tuneseriesof (a) shipboardwindspeedfromthe Wecoma,
(b)
(Nil, M2, andM3). Section(AB) illustrates
the relief alongthe Sierra meteorological
buoyM2 windspeed,(c)near-surface
E$(dB)signalatM2,
Madre del Sur.
(d)meteorological
buoyM3 windspeed,
(e) near-surface
E$(dB)signalat
M3, and05 near-surface
F.$(dB)signalat M 1.
Therespective
depths
of theinstruments
were135(_1), 124(_2),
and 147 (+1) m, as determinedat Aanderaa RCM4 cut'rentmeters
located12 m beneatheach ADCP. At the depth of the
instruments,
isolatedfrom the vigorouswind-generated
surface
flows(maximum120 cm s-1)by the strongthermocline
(AT •8øC),ambientflowswereof the order10-30cm s'l, inducing
minimalmooringmotion. The ADCP parameters
usedwere a
nominal4-m bin length,135pingsperensemble,
andan ensemble
time averageof 300 seconds.
The onsetof northersis often dramatic,as observedat M2 on
February21, whenwind speedincreased
by 10 m s'• in 90 rains
(Figure2b). Additionally,
considerable
spatialvariabilityexisted
in thewindspeedfieldovertheGulf. Forexample,
at thepeakof
theeventlastingfromJanuary21 to 24 the Wecoma
wasforcedto
shelterat the headof the Gulf, resultingin distinctdifferences
in
shipboardspeed(Figure2a) fromthatmeasured
at M3 (Figure
2d). The wind jet fansout over the Gulf as it exitsthe Chivela
Pass, so that mean directions recorded within 50 km of M1 in the
west and M3 in the east differed by 69ø (Table 1). It was
thereforenecessarythat intercomparisons
were made between
instruments
in closeproximity. For the purposes
of thiswork,
meteorological
buoyandADCP pairingsweresufficientlyclose
(<2 kin) to regardthemas a singlemooring,therebyenabling
directcomparisons
of hourlyaveraged
data.Whenbuoydatawere
unavailable,shipboardwind datawithin a 50-km or even20-km
radius were examined.
3. WIND SPEED ESTIMATES
TheADCP backscatter
energy(E) mayberelatedto near-surface
windspeed(U) by a powerlaw dependence
of the formE ~ LP
[$chott,1989], analogous
to that typicallyappliedin satellite
scatterometry
[e.g.,Jones
et aL, 1982]. Urick[1975]alsoprovides
a discussion
of thevariationin scattering
strength
frombeneath
the
seasurface
withangle,frequency,
andwindstrength,
although
the
sound sourcesutilized were primarily nondirectional(i.e.,
explosives).
The echo intensity,or strength(E$(dB)) of the backscattered
signal,withineachbin asmeasured
by theADCP is a byproduct
of the automatic
gaincontrol(AGC) circuitry[RD Instruments,
1989]. When calibrationdata are not available,RD Instruments
suggests
usinga linear scalefactor,0.46 counts/dBat 22øC,to
convertAGC countsto ES(dB). Eachof the ADCPs usedin this
experiment,
however,had previouslybeen calibratedby RD
Instrumentsto determine the variation in this scale factor with the
ambient
temperature
of the electronics
(typically,0.34%per øC
[Heywood
et aL, 1991])andwithsignalstrength.Thesedatawere
fit to an expression
of the form
E$(dB)= c0 (T) + c, (T) * AGC + c2 (T) * AGC
(1)
wherethe coefficients
%, Cl, and c2 were allowedto vary
quadratically
with temperature
[seeFlagg alwlSmith,1989].
BROWNET AL.: UPWARD LOOKINOACOUSTICDOPPLERCURRENTPROFILERS
TABLE
1. Mean Direction of Near-Surface
ADCP
17,927
Data
andShipWinds Within$0 km of the Moorings
a)
Mooring M I
Mooring M2
Mooring M3
0700 I0T Jan 19 to
0100 I0T Jan lg to
0200 I0T Jan 21 to
1300 I0T Feb 4, 1989 1200 I0T Feb 2, 1991 1900 UT Feb 5, 1991
Depth,
m
Angle,
deg
Depth,
m
Angle,
deg
Depth,
m
Angle,
deg.
5O
•Surface
217 ñ 47
Surface
207 ñ 14
Surface
81 ñ 49
12.6
242 ñ 53
13.8
211 ñ 26
16.4
31 ñ 32
20.8
245 ñ 63
22.0
209 + 31
24.6
31 + 21
28.9
243 ñ 66
30.2
206 + 29
32.8
35 + 26
37.1
243ñ62
38.3
205 + 29
40.9
37 ñ 24
45.2
243 ñ 62
46.5
201 + 28
49.1
36 +_ 15
53.4
244 + 60
54.6
199 + 26
57.2
34 ñ 20
61.6
244 ñ 64
62.8
200 ñ 27
65.4
34 ñ 15
69.7
239 + 70
71.0
200 + 27
73.6
33 ñ l g
Ship
230 ñ 95
Ship
191 ñ 52
Ship
161 ñ 42
,4*
4O
0.1
1.0
10.0
100.0
90-
80
•
b)
70
v
CO 60
•
The E$(dB) calculatedin thisway showedonly minordifferences
from what would have been obtainedby using a constantscale
50
40
factor to convert the AGC; effective scale factors for M1 and M2,
30
computedfrom the regressionof AGC andES(dB), were 0.47 and
0.46, respectively. A transducerface on M3 had been replaced
subsequent
to calibration;thereforea value of 0.46 was assumed.
Use of theseconstantscalefactorsgave estimatesof ES(dB) with
90-
0.01
80
a roughness
lengthof z0= 2 * 10-nm.
log,0 U = -1.31(+0.07) + 0.029(+0.001)ES(dB)
(2)
with a correlationcoefficientr of 0.96 (99% significancelevel is
0.76). Similarlyfor the longerrecordsat M3 (Figure3b),
where r = 0.71 (99% significancelevel is 0.37).
10.00
100.00
70
v
CO 60
•
50
40
30
At M2 (Figures 2b and 2c) and M3 (Figures 2d and 2e)
fluctuationsin the time seriesof ES(dB) closely imitated thoseof
wind speedon time scalesas shortas an hour. Consideringfirst
M2, a regression
of thehourlyaveragesof ES(dB)againstlog,0U
(Figure3a) prior to the failureof the meteorological
buoyyields
log•0U = -1.24(+0.09) + 0.029(+0.001)ES(dB)
I•
1.00
c)
standard errors less than 1 dB.
It is importantto notethat absolutecalibrationof power into the
water was not known; therefore values of ES(dB) were defined
relative to an arbitrary level, and intercomparisonsbetween
instruments
would be inappropriate.
No correctionfor propagationeffectshave beenmade,but these
will be minimal as vertical excursions experienced by the
mooringswere minor. Additionally, the presentedwind speeds
have beencorrectedto 10 m, assuminga logarithmicprofile, with
0.10
(3)
Here, the
considerablescatter at U < 1.0 ms-' has been excluded. At low
velocities, near-surface bubble concentrationsare reduced [Wu,
1988], and it may be that surfacewave patternsplay a stronger
role in regulatingthe intensityof the specularlyreflectedsurface
backscatter. In addition, buoy motion may at times induce a
degreeof "pumping"of the anemometer,
producinganomaliesat
speedscloseto the threshold.
Similar analysis at M1 was precluded by the loss of the
meteorologicalbuoy. A less satisfactorycomparisonwith the
0.1
1.0
10.0
100.0
(m/s)
Fig. 3. Hourly values of wind speed (abscissae)versus surface echo
strength(ES(dB)) (ordinates). (a) M2 with the regressionline of equation
(2). (b) M3 with the regression
line of equation(3) (speeds
> 1.0m s'i).
(c) Wecomawindswithin 50 km of M3 represented
by both symbolsand
the solidline of equation(4), with the datawithin 20 km givenby crosses
andthe dashedline of equation(5).
windsrecordedon boardthe RV Wecoma(Uw) can be made. An
inspectionof the records from M2 and M3 (Figures 2b-2e),
separatedby 112 km, revealsthat althoughthe main featuresare
common,there are marked differences. Approachesto mooring
siteswerespasmodic,
rarelyfor morethan12 hoursandseparated
by 2 to 5 days. The decorrelationtime [Davis, 1976] for the
ES(dB) signalwas of order 1 day; thereforeeach approachcan
reasonablybe regardedas 1 degree of freedom. Regressions
appliedto shipdatawithin 50 km and 20 km of M 1, equations(4)
and (5), respectively,yield
log•0 Uw = -0.648(ñ0.143) + 0.019(ñ0.003)E$(dB)
(4)
17,928
BROWN ET AL.: UPWARDLOOKINGACOUSTICDOPPLERCURRENTPROFILERS
r = 0.66 (95 % significancelevel is 0.67) and
log•0 Uw -- -0.869(+0.158) + 0.024(+_0.003)ES(dB)
(5)
r = 0.88 (99% significancelevel is 0.88). Levels of competence
are reducedin comparisonwith equations(2) and (3), particularly
at 50-km separation,emphasizingthe desirability of making
comparisons
as closeto the mooringsaspossible.In addition,the
Wecomawas unableto work in the vicinity of the mooringduring
strongwinds,thereforecomparisons
are fmlher hamperedby the
restricted
rangeof wind speeds(0.0-7.4m s-•). Further,during
periods of light winds the accuracyof velocities,derived by
removalof the ship speedover the ground,becomeslesscertain.
4. WIND
DIRECTION
Here, we have adoptedthe oceanographicconventionthat the
angles,measuredwith respectto truenorth,denotethe directionof
travel,i.e., towardwhich air parcelsmove.
Comparisonsof the near-surfaceADCP directionalinformation
with that from the buoys were hamperedby the loss of M1 and
the malfunctionof the vane at M3; thereforeM2 providedthe sole
mooredvector-winddataset, of duration72 hours. Again, hourly
averagesare considered,which in the case of the buoys were
derivedfrom 6 min spotreadings. Consequently,
at low speeds
contaminationof the data by wave-inducedbuoy motion may
consistent
with the natureof the atmosphericforcingduringthe
Northerseason.Regressions
betweenthe curtailedbuoyrecordat
M2 and the ADCP, or for ship boarddata within 50 km and the
ADCPs at M1 and M2 are of little value, as in each case,
directions
weregenerallyconfinedwithin a sectorof lessthan90ø.
The meandirections
fromthe M2 buoyrecordandthe A,DCP
duringthe sameperiodwere 138 + 26ø and206 + 8%respectively.
This differenceof almost70ø mightbe thoughtattributable
to the
predominance
of low windspeeds
(<5 m s4) in therecord.Mean
wind directions calculated for Wecoma data within 50 km of the
ADCP mooringsgave betterresults. At M1, the meanshipwind
directionof 230 + 95ø agreedwell with the meandirectionof 227
+ 56ø indicatedby the ADCP; at M2 the shipmeanwas 191 + 52ø
comparedwith the ADCP mean of 207 + 19ø.
If it were not for the recordat M3 (Figure 4e), which indicates
an eastwardwind, it mightreasonablybe argued,as postulatedby
Schott[1989], that the near-surfaceADCP directional information
provides an adequate measure of wind direction. At M3
comparisonswere limited to thosewith the Wecomadata. When
within 50 km of the mooringthe meanwind directionwas 161 +
42 ø in contrast to 88 + 24 ø at the ADCP.
consistent with
that for the whole
The ADCP
series of
81
value is
+ 49 ø.
A
comparisonwith the completeWecomarecord(mean 195 + 72ø),
covetingthe entire Gulf, would be inappropriate. The wind jet
occur. Vector wind information was recorded on board Wecoma
fansoutafterexitingthe ChivelaPass,beingsouthwestward
to the
west and southeastwardto the east. Manifestly, there is an
every minute.
Time series of wind direction recorded at M2 and aboard
inconsistency
betweenthe resultsat M3 and thosefrom M1 and
Wecomaare presentedin Figure 4, accompaniedby the surface M2, for which insmn'nentmalfunctionseems an improbable
explanation. Comparisons of direction and speed from
directionalinformationderivedfrom the ADCPs. ExcludingM3,
directionswere predominantlybetweensouthwestand southeast, independentmeasurementsat the Aanderaa meters situated
immediatelybelow each ADCP and ship-mountedADCP data
indicatedexcellentagreement.
5. DISCUSSION
Evidently,near-surface
wind speedandthe acousticbackscattered
energy returnedto ADCPs from the sea surface are strongly
correlated. In the Gulf of Lions, $chott [1989], unable to record
b) Met. M2
c) ADCP M2
wind continuouslyat the ADCP moorings,was constrainedto
considershipboarddatawithin a 100-km radiusandemploya 12houraveragingperiodto smoothshort-periodinconsistencies.
The
close proximity of the instrumentswithin the meteorological
buoy/ADCPpairs duringthe Tehuantepecexperimenthas shown
thatvaluescorrespondaccuratelyon time scalesof an hour.
$chott[1989] citedevidencethattheechoamplitudesignalmight
saturate
at windspeeds
exceeding
10m s-l, a featureabsent
in the
Tehuantepec
datawheremaximumnear-surfacewind speedsat the
moorings
reached15 m s4. Thisdisparity
mightbe attributable
to
the useof 30ø beamanglesin Tehuantepec,
whereasin the former
effective angles of 20ø and 24ø were used. For beam angles
270
relativeto the verticalof between0ø andapproximately
15ø the
dominantbackscattering
processis taken to be specularfrom the
normallyinclinedwave facets[Urick, 1975; $chott, 1989] andthe
360
slopefactorof wind speeddependence
is negative.At roughly15ø
thereexistsa crossoverpoint to a positivedependence
as Bragg
scattering
becomesincreasinglydominant.In this region,changes
in scatteringstrengthbecome progressivelyless sensitiveto
0
increasesin wind speed.
16
17
18
19
20
21 22
23
24
25
26
27
28 29 30
31
I
2
3
Upward lookingADCPs offer the potentialto determinelocal
January to February1989 (UT)
surfacewind stresswithout additionalexpensiveand vulnerable
surfaceinstrumentation
or recourseto ship board or land-based
Fig. 4. Direction time series,following the conventionthat the angle is
data that often fail to truly representconditionsat surveysites.
relativeto true north and toward the directionof flow. (a) Wecoma, (b)
can be regardedas replacements
for wind
meteorologicalbuoy M2, (e) near-surfaceADCP at M2, (d) near-surface Beforethe instruments
ADCP at M 1, and (e) near-surfaceADCP at M3.
buoys, questionsremain to be answered. For instance;Are
0
,
,
i
,
,
i
i
,
,
,
,
,
,
t
,
i
,
•
,
BROWNET AL.: UPWARD LOOKINGACOUSTICDOPPLERCURRENTPROFILERS
laboratorycalibrations
of transducer
gainsstable?How predictable
aredepthdependencies
dueto spreading,
scattering,
andabsorption
which will affect interpretationfrom separatedeploymentsof an
instrument?Is therea fetchdependence?
Doesthe natureof the
seasurfacelayervarysignificantlywith meteorological
conditions?
Comparisonsof near-surfaceADCP directionalinformationwith
ship board wind directions proved less satisfactory. An
explanation
may be foundfrom an examinationof the near-surface
ADCP data (Table 1). During the observationalperiod the
directionof the flows in the surfacelayer at the western(M 1) and
central(M2) mooringswas predominantly
southwestward.At the
easternmooring (M3) there existed a persistentnortheastward
return flow toward the head of the Gulf. Thus, at M1 and M2,
currentswere essentiallyin the samedirectionas the wind, while
17,929
evidentthat currentdirectionwas similarthroughout
the upper
layersand essentiallyopposedto Wecornawind (separation< 50
km) direction, while the surface bin direction varied between the
two. This seemsreasonable,
asthe Dopplershift measuredat the
ADCP is a power-averaged
meanof thosegenerated
by advancing
and retreatingsurfacewavesand the velocityof scattererswithin
the water column,the latter being additionallyadvectedby the
mean flow lapel, 1987].
Duringthe periodsof consideration,
in both the Gulfs of Lions
[Schott,1989] and Tehuantepee,
circulationwas primarilywind
driven. The experimentsdiffered in that in the latter, an ADCP
(M3) was situatedin a regionforcedindirectlywhere circulation
persistentlyopposedthe wind direction. Indeed, Schott [1989]
discarded
severalperiodsof strongbarotropic
flowsextendinginto
thesurfacezonewhencurrentsclearlyopposed
the winddirection.
It may be that undera more diversecurrentregimehis findings
might have differed.
In conclusion,the presentdata set showsthat the near-surface
backscattered
energy recordedat the ADCP providesa good
measureof wind speed,on time scalesof an hourwith apparently
no saturationof the ES(dB) signalat higherwind speeds. The
considerationto M3, where the ADCP surfacebin direction falls
Dopplerdataagreed
roughlymidway (mean 81ø) betweenthe mean wind direction directionalinformationfrom the near-surface
with
the
observed
wind
direction
only
where
the
underlyingcurrent
derivedfrom the Wecornawhen within 50 km of the mooring
at M3 theywere in oppositionto the dominantwind direction.
Considering
only M1 andM2, onemightconcludefromthe close
agreementin the ship wind and surfacebin directionsthat the
near-surface
ADCP directionalinformationwas representative
of
wind direction. Equally, it could be concludedthat it was
indicative of near-surfacecurrent direction. Extending
(at 20 m) was in nearlythe samedirectionas the wind. At the site
where the directionsof the underlyingcurrentand wind differed
some combination of the current and wind. Consideration of data
by almost 120ø, the near-surfaceDoppler data indicated an
when Wecornaand M3 are in closerproximity(25 km) produces intermediatedirection. This castsdoubt on the ability of the
to providereliable estimatesof the wind
little change,with meansof 159øand87ø, respectively.Excluding Dopplermeasurements
(mean 160ø) and the currentdirection(mean ~ 32ø) (Table 1), it
is reasonable
to concludethat the near-surface
signalis in fact
periodsof weak winds (<3 m s-•), when directionsmight be
expectedto be variable,yields the sameresults.
The ADCP time seriesis illustratedin (Figure 5), where it is
Wecoma
10
direction.
Acknowledgements.UCNW funding was provided by the Natural
EnvironmentalResearchCouncil(grantGR3/6719) and for OSU by the
Office of Naval Research(contractN00014-87-K-0009;grantN00014-90J-1177). Thanks are also due to all those who contributed to the
Tehuantepecproject.
wind
REFERENCES
E
Apel, J.R., Principlesof OceanPitysics,lnt. Geophys.Set'..vol. 38, 534
pp., Academic, San Diego, Calif., 1987.
Davis, R.E., Predictabilityof sea surfacetemperatureand sea level
pressure
anomaliesoverthe NorthPacificOcean,J. Phys.Oceanogr.,6,
-10
Surface
120
249-266, 1976.
Derecki, J.A., and F.H. Quinn, Use of current meteta for continuous
measurementsof flows in large rivers, Water Resour. Res., 23,
1751-1756, 1987.
o
Flagg,C.N. and S.L. Smith, On the useof the acousticDopplercun'ent
profilerto measurezooplankton
abundance,
DeepSeaRes.,36, 455-474,
-120
16.4
120
1989.
m
Heywood, K.J., S. Scrope-Howe, and E.D. Barton, Estimation of
zooplanktonabundancefrom shipborneADCP backscatter,Deep Sea
Res., 38, 677-691, 1991.
-121
73.6
120
Hurd, W.E., Northernof the Gulf of Tehuantepec,Mon. WeatherRev.,57,
192-194, 1929.
Johns,W.E., Near-surface
currentmeasurements
in theGulf Streamusing
an upward-lookingacousticDopplercurrentprofiler,d. Atmos.Oceanic
m
Technol., 5, 602-613, 1988.
Jones,W.L., L.C. Schroeder,D.H. Boggs,E.M. Bracalente,R.A. Brown,
G.J. Dome, W.J. Pierson, and F.J. Wentz,
o
-120
i
i
i
i
i
i
15
20
25
30
35
40
Time (UT) Julian Days 1989
Fig. 5. Horizontalcurrentvector time seriesat M3 with Wecomawinds
when within 50 km of the mooring.
The Seasat A satellite
scatterometer:The geophysicalevaluation of remotely sensedwind
vectorsover the ocean,J. Geophys.Res.,87, 3297-3317, 1982.
RD Instruments, Acoustic Doppler Current Profilers principles of
operation:A practicalprimer, 36 pp., San Diego, Calif., 1989.
Roden,G.I., On the wind drivencirculationin the Gulf of Tehuantepec
and its effectuponsurfacetemperatures,
Geofis.Int., 1, 55-80, 1961.
Schott,F., Measuringwinds from undemeaththe oceansurfaceby
upwardlookingacousticDopplercurrentprofilers,J. Geophys.Res.,94,
8313-8321, 1989.
17,930
BROWN
ETAL.:UPWARD
LOOKING
ACOUSTIC
DOPPLER
CURRENT
PROFILERS
Schott,F., andW. Johns,Haft-yearlongmeasurements
witha buoymounted
AcousticDopplerCurrentProfilerin the SomaliCurrent,J.
Geophys.
Res.,92,5169-5176,
1987.
Schott,
F.,andK.D. Leaman,Observati
• withmoored
acoustic
Doppler
currentprofilersin the convectionregimeof the Guff of Lions,J. Phys.
Oceanogr.,21,558-574, 1991.
Tm.veina,A., Offshorewind forcingin a coastalocean:observations
and
modelingof the Guff of Tehuantepec,Mexico, Ph.D. Thesis,88 pp.,
Universityof Wales, 1991.
Urick, P,.J., Principlesof Ut•lerwater Sound,2nd ed., 384 pp., McGrawHill, New York, 1975.
Wu, J., Bubbles in the near-surface ocean: A general description,J.
Geophys.Res., 93, 587-590, 1988.
E.D.BartOn
'and
H.S.Velez,
School
ofOcean
Sciences,
University
of
Wales,Bangor,Menai Bridge,Owynedd,LI.59//EY, Wales.
J. Brown,M.A.F.F., FisheriesLaboratory,Lowestoft,Suffolk,NR33 0HT,
England.
P.M. Kosro and R.L. Smith, College of Oceanography,
OregonState
Univemity, Corvallis, Oregon 97331.
A. Trasvina,Centrode InvestigacionCienti•lcay EducacionSuperiorde
œnsenada,Ensenada,British Colombia, Mexico.
(ReceivedJanuary27, 1992;
acceptedMarch 30, 1992.)
Download