VOL. 84, NO. CII
JOURNAL OF GEOPHYSICAL RESEARCH
NOVEMBER 20, 1979
The Spring Transition in Currents Over
The Oregon Continental Shelf
A. HUYER, E. J. C. SOBEY,' AND R. L. SMITH
Schoolof Oceanography,
OregonState University,Corvallis,Oregon97331
Currentmetermooringsmaintainedoverthe Oregoncontinentalshelfin 1973and 1975clearlyshow
the differencebetweenwinterand springoceanographic
regimesand the rapid transitionbetweenthe regimes.In winterthe meanalongshore
currentis northwardat all depthsand strongest
nearshore;thereis
no meanverticalshear,no meanoffshoredensitygradients,and meansealevel is high. In springthe
meanalongshorecurrentis weak nearthe bottomand stronglysouthwardat the surfacewith a maximum
overthe mid-shelf;the strongverticalshearis balancedby strongoffshoredensitygradients,and mean
sealevelis low. In both 1973and 1975the transitionbetweentheseregimesoccurswithin abouta week
duringa strongsouthwardwind event.In manyrespects
the transitioneventdoesnot differsignificantly
from previoussouthwardwind events,but the lateraldensitygradientsestablished
duringthe transition
eventexceedthoseof earlierevents.The transitionappearsto be theresultof a largecumulativeoffshore
Ekman transportcausedby local wind stressratherthan by propagationof effectsgeneratedelsewhere.
The lateraldensitygradientsand the verticalshearestablished
duringthe transitioneventsubsequently
persist,evenduringnorthwardwind eventswith moderateonshoreElanantransport.
INTRODUCTION
middle of the continental shelf about 10 km north of New-
The winter and summercurrent regimesover the Oregon port, Oregon(Figure 1). The mooringwasrecoveredand reincontinentalshelfare quite distinct[Huyer eta!., 1975b,1978]. stalledeight timesat very nearly the sameposition,and with
In winter there is little or no mean shear, the mean flow is the currentmetersat aboutthe samedepths(Table 1). Instal-
by a fisherman,
leavinga
northwardat all depths,and the northwardflow is strongest lationV wasprematurelyrecovered
7-day
gap
before
installation
VI;
all
other
gaps
due
to servvery near shore.In summer the mean surfacecurrent is southward, and there is a strong mean vertical shear such that icing are lessthan a day long. From installation to installadeeper currents are always more northward than shallower
currents;the surfacesouthwardflow is strongestabout 15-30
km from shore.
The distinctionbetweenthe winter and summerregimes
wasclearly observedin a nearly continuousyear-longcurrent
recordobtainedin 1973.This year-longrecordsuggested
that
the springtransitionoccurredquickly, in a few weeksor less.
To studythe winter-to-springtransitionin more detail, we began an observationalprogramat the end of January 1975.The
observationsconsistedof three arraysmooredacrossthe continental sheff at 45øN (Figure 1) from the end of January
through mid-April !975 with the mid-sheff mooring maintained until early September 1975. Supplementaryhydrographic sectionswere made along 45øN at irregular intervals
betweenJanuaryand July 1975.
In this paper we shall presentresultsfrom the long midsheffrecordobtainedbetweenDecember1972and April 1974,
from the secondmid-shelf record obtained betweenJanuary
and September 1975, and from the sectionacrossthe sheff at
45øN betweenJanuary and April 1975.We shall describethe
1975 spring transition in detail, define its essentialcharacteristics, and show that similar transitions have occurred in other
tion, the depths of the current meters remained almost the
same.Current observationsare nearly continuousat 23 m for
12 months, and at 40 m for 17 months;the 80-m data have a
2-weekgap at the end of March 1973and a 6-weekgap in
July.
The current observations were made with Aanderaa current
meterson a taut subsurfacemooring with the shallowestfloat
about 4 m above the top current meter. These Currentmeters
record temperature as well as instantaneousdirection and av-
eragespeed.The raw data were filteredto reducehigh-frequencyvariations,yieldinghourly values.The gapsdue to replacingthe mooringswere filled by artificialdata [U!rychet
al., 1973].The joined time serieswere then low-passfiltered
(half-powerpoint of 40 hours)to suppresstidal and inertial
oscillations.The coordinatesystemwas rotatedclockwise20ø
from the east-northsystemto obtainonshoreand alongshore
componentsof the currents.
Observationsof sealevel, atmosphericpressure,and wind
speedand direction were made at Newport, Oregon, and
hourly data are available. The sea level data have been ad-
justed for the 'invertedbarometer'effect,by which sealevel
falls I cm for every l-mbar increasein atmospheric
pressure.
years.Finally, we shall investigatethe causeof the sp'ring The hourly time seriesof adjustedsea level and wind were
transition and attempt to account for the time of its occurrence.
low-passedin the same manner as the currents to remove
diurnal and shorterperiodvariations.
Daily valuesof the low-passed(LLP) time seriesof the cur-
THE YEAR-LONG RECORD: POINSETTIA, 1973
rentsat Poinsettiaareshownin Figure2, with the adjustedsea
Mooredcurrentmeterobservations
at 'Poinsettia'(44ø45'N, level and the wind. From December 1972 to March 1973 the
124ø17'W) beganin December1972and continueduntil May currentsat all depthsshowfrequentreversalsand are gener1974. Poinsettia is located over the 100-m isobath, about the ally in the samedirectionas the wind. From April through
September1973the currentat 23 m is alwayssouthward,but
• Presentaddress:
Science
Applications,
Boulder,Colorado80302.
Copyright¸ 1979by the AmericanGeophysical
Union.
Papernumber 9C1074.
0148-0227/79/009C-1074501.00
6995
the current at 80 m continues to reverse direction with the
wind. The current at 40 m is continuallysouthwardfrom
April through July; in Augustit beginsto showreversalslike
6996
HUYER ET AL.: SPRING CURRENTS OFF OREGON
45o30 '
i
i
45oZ
44o$0'
I
125'30'
125ø
124' 30'
124•
Fig. 1. Location of current meter arrays:Poinsettia,December 1972to April 1974;Carnation, July-August 1973;and
Wisp array (Wisteria, Sunflower,and Pikake) acrossthe shelf at 45øN, January-April 1975;Sunflowerwas maintained
until September1975.Continuingwind, atmosphericpressure,and sealevel observationsare made at Newport.
thosein the wind and the 80-m current. Beginningin Septem- the maximum southward wind. The difference between the
ber 1973, wind-induced reversalsagain occur at all depths alongshorevelocity at the top and bottom current meters,
through December1973, when the 23-m recordends.The which is a measure of the depth-averaged vertical shear,
data from 40 and 80 m suggestthat thesewind-inducedre- seemsto be almostconstantat 25 cm s-• in springand early
versalscontinueat all depthsthroughMarch 1974.Southward summer,and about zero in winter; the magnitudeof the shear
flow at 40 m is re-established
in April 1974,abouta month be- increasesvery quickly in early springand decreases
slowlyin
fore the end of the record.
late summerand fall. Linear correlationcoefficientscomputed
To look at the seasonal variations in more detail, we filled among the VLF time series(Table 2) show that the alongthe gapsin the 80-m recordbeforefurtherfilteringto remove shorecurrentsat 23 and 40 m are significantlycorrelatedwith
the wind-induced fluctuations.The first gap in the 80-m rec- sea level at the 99.9% level but are not as well correlated with
ord, a 14-daygap beginningon March 26, 1973,wasfilled by the wind; the 80-m current is not significantlycorrelatedat the
artificial data obtainedby extendingthe previousrecordfor- 95% level with either the sea level or the wind.
ward and the subsequentrecord backward [Ulrych et al.,
In the residual (ILF) time series,which contain signalswith
1973].To fill the second(6-week)gap,we useddata from 80 periodsbetween2 and 50 days, almostall of the wind variam at Carnation [Pillsburyet al., 1974], a mooring over the tions are reflectedin changesin sealevel and currents(Figure
same isobath but about 55 km further north (Figure 1); be- 4). Linear correlation coefficientsbetween the ILF series
causeof the high alongshorecoherenceover the shelf [Huyer (Table 3) showthat the wind, sealevel, and currentsare siget al., 1975a;KunduandAllen,1976],thisdatarecordis prob- nificantly correlatedwith each other at the 99.9% level. This
ably a very goodestimateof the currentsat Poinsettia.A very strongrelationshipbetweenmid-shelf currents,sea level, and
low passfilter (half-powerpointof about50 days)wasapplied wind is consistentwith the resultsof Huyer et al. [1978], who
to the long low-passedtime series.The very low frequency showed seasonal variations in the structure of the low-fre(VLF: 0-0.02 cpd) time seriespassedby this filter are shown quencycurrentfluctuationsbut not in the ratio betweenwind
in Figure 3. The residual (ILF: 0.02-0.6 cpd) time series, and current fluctuations.
which excludeperiodslonger than 50 days and shorterthan 1
day, are shownin Figure 4.
THE 7-MONTH RECORD: SUNFLOWER, 1975
The VLF time series(Figure 3) showa definiteseasonalsignal in the sea level and the alongshorecurrentsat 23 and 40
Measurementsdesignedto observethe 1975 spring transim, with apparent annual rangesof about 35 cm in sea level tion in detail beganat the end of January 1975.The mid-shelf
and 60 cm s-' in currents.Weaker seasonalcyclesare alsoap- mooring, Sunflower (Figure 1), was replaced in April and
parentin the 80-m current,and the wind, with rangesof about again in July and was finally recoveredin September1975
Sunflowerrecordswere
20 cm s-' and 8 ms-', respectively.There is a high degreeof [Gilbert et al., 1976]. The successive
similaritybetweenthe seasonalcyclesof the sealevel and the joined to providecontinuoustime seriesfrom the endof Janualongshorecurrentsat 23 and 40 m. The lowestsealevel and ary to mid-Septemberat 26, 76, and 92 m and to the end of
the maximumsouthwardcurrentoccurseveralmonthsbefore July at 52 m. Althoughtheseobservations
do not spanan en-
HUYER ET AL.: SPRING CURRENTS OFF OREGON
6997
tire seasonalcycle, they do allow us to check the validity of
some of the conclusions based on the Poinsettia
data.
The data seriesfrom the Sunflower current meters (rotated
8ø clockwise)and the wind and adjusted sea level data from
Newport were filtered in the sameway as the Poinsettia data,
yielding LLP (0-0.6 cpd), ILF (0.02-0.6 cpd), and VLF (00.02 cpd) data series.Daily values of the LLP Sunflower current vectors(Figure 5) show variationssimilar to thosein the
Poinsettiadata (Figure 2): in February and March the alongshorecurrentsreversewith the wind; in April, May, and June
the near-surface currents remain generally southward, although the near-bottom current continuesto oscillatewith the
wind. The ILF time series (Figure 5), from which periods
longer than 50 days have been removed, show that current
fluctuationsat all depthsare closelyrelated to fluctuationsin
wind and sealevel. Correlation coefficientsamong the Sunflower ILF time series (Table 4) are very similar to those for
Poinsettia (Table 3). Combining ILF current data from all
depths, the regressioncoefficientsof alongshorecurrent as a
function of sealevel are nearly the samefor the two data sets:
1.7 s-' for Poinsettia and 2.0 s-' for Sunflower.
The VLF Sunflowerseries(Figure 6) are also similar to the
PoinsettiaVLF files:sealevel is high in late winter and low in
spring and summer;the near-surfacecurrentsare nearly zero
in late winter, stronglysouthwardin springand early summer,
and weakly northward in late summer;the differencebetween
the near-surface and near-bottom alongshore currents is
nearlyzeroin late winterbut quitestrong(about30 cm s-') in
spring and early summer. The magnitude of the mean shear
increasesvery quickly in early spring, and its final value
(0.0045s-') at Sunfloweris very nearlyidenticalto its value at
Poinsettia. Thus the Sunflower observationssupport the conclusions based on the Poinsettia data.
The Sunflower observationsprovide some information not
available from the Poinsettia data. Since the Sunflower obser-
vationswere essentiallycontinuous,it is possibleto compute
the low-passedvelocitydifferencebetweenthe top and bottom
current meters(Figure 7); this time seriesshowsthat the persistentmean shear is establishedvery quickly (within I or 2
weeks) in early spring, more quickly than one might guess
from the VLF
time series. Since some of the Sunflower
cur-
rent metersalsorecordedconductivity,it waspossibleto compute time seriesof sigma-t.Figure 7 showsthat the near-bottom sigma-t was highly variable in February and March but
constantwithin about +_0.1from April on. This suggeststhat
the transition
in currents
and sea level is associated
with
a
changein the densityfield over the continentalshefl.
THE OFFSHORE SECTION AT 45 øN, FEBRUARY-APRIL
1975
From late Januaryto mid-April 1975we made intensiveobservations of the currents and hydrography along 45øN.
Three current meter arrays, including Sunflower, were
mooredover the shelf (Figure 1) with a total of 11 current meters [Gilbertet al., 1976];each currentmeter recordedtemperature and pressureas well as speedand direction, and all but
one recordedconductivity,so we were able to compute time
seriesof sigma-t. The duration of theserecordsis not long
enoughto warrant calculatingvery low frequency(VLF) time
series;instead, we shall examine the offshore structure of the
mean winter and spring conditions,i.e., before and after the
springtransitionas definedby the Sunflowerdata.
6998
HUYERET AL.: SPRINGCURRENTS
OFF OREGON
J
F
M
A
M
J
J
F
M
A
M
J
J
A
J
A
'S
0
N
D
J
F
0
N
D
J
F
M
A
M
A
M
6O
50'
23m
,,?, 50
40rn
•
o
• -,50
•
50
-15
S
1973
1974
Fig.2. Dailyvaluesof low-passed
(LLP)timeseries
of thecurrents
of Poinsettia
andof windandadjusted
sealevelat
Newport. The coordinatesystemfor the currentvectorshave beenrotated20ø clockwisefrom north. Sealevel is referredto
its 1973 mean.
OffshoreStructure
the inner half of the shelf;this changeis more than twice as
large as the greateststandarddeviation in either season.There
Distributions of the means and standard deviations of the is almostno changein densitynear the surfaceover the outer
low-passedalongshorecurrent and the density(Figure 8) half of the shelf or in the deep water at the most offshore
show that there are significant differencesin the offshore mooring.
structure before and after the transition. In winter the mean
The distributionsof the standarddeviationsof alongshore
currentand density(Figure 8) alsochangesignificantlyfrom
winterto spring.In winterthe variabilityin velocitydecreases
the mean and the standarddeviationare greatestnear shore with increasingdepthat eachmooting;in springthe variabiland smallerover the outer shelf.In springthe mean along- ity is almostconstantwith depth over the mid-shelfand outer
shorecurrent is southwardalmost everywhere,and near the shelf. The variability decreaseswith distance from shore,
surfaceit is larger than the standarddeviation,i.e., the surface ratherweaklyin winterbut stronglyin spring.The variability
current over the mid-shelf and outer shelf is persistently in the densityis greatestnear the bottom at mid-shelfin winsouthward.Whereasmean isotachsin winter are generally ter and nearthe surfaceoverthe innershelfin spring.In both
vertical(i.e., the northwardcurrentdecreases
primarilywith casesthe densityvariabilityseemsto be greatestwherethe vedistancefromshore),themeanisotachs
in springaregenerally locityvariabilitydecreases
moststronglywith depth,suggestlateral (i.e., there is significantvertical shear over the inner ing that the densityvariability is an integralpart of the velocshelf and the outer shelf as well as over the mid-shelf).In ity variability. This variability and the difference in its
spring the mean alongshorecurrent at each depth on all characterin winter and springhave beendiscussed
in detail in
alongshorecurrent is northward over the entire shelf, but the
mean is everywheresmaller than the standarddeviation;both
mooringsis southwardrelative to the near bottom current,i.e.,
an earlier paper [Huyer et al., 1978].There, we found that the
the tendencytoward southwardflow penetratesmostof the
variabilityis dueto low-frequency
fluctuations
inducedby the
water column. The changebetweenseasonsin mean currents local wind in both seasons;
the amplitudeof the fluctuationsis
is greatestnear the surfaceand over the mid-shelf;near the a function only of the wind stress;the vertical structureof the
bottom the changeis about the samesizeas the standardde- fluctuationsis more strongly baroclinic in winter and more
viations in either season.
The mean isopycnalsare nearly level with a slightdownward slopetowardshorein winter;in springthey slopeupward towardthe coast,with the steepestslopesnear the bottom over the inner shelf (Figure 8). In winter the lightest
surfacewater is observedat the most inshoremooting;in
springit isobserved
at themostoffshoremooting.The change
in mean densitydistributionis greatestnear the bottom over
nearlybarotropicin spring;the offshorelengthscaleis about
the same as the shelf width in winter and about half the shelf
width in spring;in both seasons
the velocityfluctuationsare
highlycorrelatedwith coastalsealevel,and verticalvelocity
gradientsare balancedby lateraldensitygradients.We speculated that the change in the structure of the wind-induced
fluctuations
wasrelatedto the changein the meanalongshore
currents.
HUYERET AL.: SPRINGCURRENTS
OFFOREGON
I
6999
7000
HUYER ET AL.: SPRING CURRENTS OFF OREGON
6O
i
J
F
M
A
M
J
';'"i
J
A
S
]
I'¾W' ' •"i•
0
!
N
D
N
D
i
J
F
M
A
• _I$•
J
F
M
A
M
J
J
A
S
0
J
F
1973
M
A
1974
Fig. 4. Daily valuesof the'intermediately
low'frequency
(ILF) currents
at Poinsettia,
withILF windandadjusted
sea
level at Newport.
Relationships•4mong Variables
for Newport from the observedhourly data, adding the hourly
atmosphericpressureto adjust for the inverted barometer efThe effectsof the springtransitionare most clearly mani- fect, and adding the annual and semi-annual cyclesto comfestedin the adjustedsea level, the near-surfacealongshore pensatefor their inclusionin the predictedsealevel. (The precurrents,the near-bottom density, and in the differencebe- dicted sea level data were provided by National Ocean
tween near-surface and near-bottom currents over the midSurvey; tidal constantsfor terms with periods of a month or
shefl.In thissectionwe shallusehourlytime seriesto examine less were based on 1971 observations, and the constants for
a few of the relationshipsamongthesevariables.
the annual and semi-annual terms were based on observations
Hourly time seriesof all the currentand densitydata have from 1967 through 1971.) Although this processdoesnot enbeenpresentedby Gilbertet al. [1976].Selectedtime seriesare tirely remove the diurnal and semi-diurnal tides, it does reshownwith hourly northwardwind and adjustedsealevel in duce their amplitude to about 5 cm, and nontidal variationsin
Figure 9. Hourly time seriesof sealevel are stronglydomi- sealevel becomedominant (Figure 9).
natedby the semidiurnaltides,whichhavea muchlargeramCurrentversussealevel. Earlier work [e.g.,Collinsand Patplitudethan the changesin sealevel associated
with changes tullo, 1970; Smith, 1974;Huyer et al., 1978] has shownthat
in wind, currents,and the densityof the water, but it is only
the sealevel 'anomalies'from the regulartidal variationsthat
are of interesthere. We computedthe hourly time seriesof
adjustedsealevel by subtractingthe predictedsealevel data
low-passedvariationsin near-surfacecurrentsand adjusted
sealevelareverywell correlatedandthat theregression
equation between them can be used to estimate the width of the
current fluctuations. In other words, it has been shown that
TABLE 2. LinearCorrelation
Coefficients
(CC) BetweenLinearlyDetrended
VLF Time Seriesof
Alongshore
Currents
at Poinsettia,
Northward
WindandAdjusted
SeaLevel,Using323DailyValues
Depth,m
CC
t,
CR95%
CR99•
CR99.9•
0.80
0.80
0.76
0.90
0.90
0.87
0.83
0.83
0.80
0.92
0.92
0.90
Current Versus Wind
23
40
80
0.79
0.71
0.46
7
7
8
0.67
0.67
0.55
Current Versus Sea Level
23
40
80
0.96
0.93
0.64
6
6
7
0.71
0.71
0.67
Degrees
offreedom
(t,)werecomputed
following
Davis[1976],
andcritical
correlation
coefficients
(CR)
at differentsignificance
levelswereobtainedfrom PearsonandHartley[1970,Table 13].
HUYER ET AL.: SPRING CURRENTS OFF OREGON
TABLE 3. Linear Correlation Coefficients(CC) Between Linearly
Detrended ILF Time Series of AlongshoreCurrents at Poinsettia,
NorthwardWind and AdjustedSeaLevel,Using323 Daily Values
Depth,m
CC
Current
23
40
80
v
low-passedcurrent and densitydata, i.e., the vertical gradient
of the alongshorecurrent is approximately balanced by the
offshore density gradient. Figure 10 showstime seriesof the
hourly velocity differencebetween26 and 53 m at Pikake and
the density differencebetween 53-m Pikake and 52-m Sunflower; thesetime seriesare very well correlated,even during periods of rapid change. The linear correlation coefficient between them is -0.72. Since the thermal wind equation is a
local relationship between local gradients, it is remarkable
that there is suchgood agreementwith finite differenceswhich
are not centeredabout the same point and which were computed from density observationsseparatedby 5 kin. The fact
that such good agreementcan be found for even one pair of
differencetime series(and in particular for sucha shallow-water, near-shorepair) suggestsstrongly that the thermal wind
equation holds for almost all the water over the shelf, except
perhapsthe surfacemixed layer.
Densityand sea level. It is generally believed that geostrophic readjustmentis mostlybarotropicfor the first few daysof
a suddenand major changein flow and that the baroclinic response, including changes in the density field, takes substantially longer [e.g., Reid and Mantyla, 1976]. The hourly
time seriesof sealevel and near-bottomsigma-tover the inner
shelf and over the mid-shelf (Figure 9) suggestthat density
changesover the shelf have essentiallythe same time scaleas
sea level changes,regardlessof season.Comparing hourly
time seriesof verticallyaveragedsigma-tat Pikake (28 and 53
m) and at Sunflower(26, 52, 76, and 92 m) with sea level, we
obtain correlationcoefficientsof-0.54 and -0.61 at zero lag;
the maximum correlation coefficientsare slightly greater
CR99.9
%
Versus Wind
0.53
0.55
0.61
160
160
160
<0.32
<0.32
<0.32
Current Versus Sea Level
23
40
80
0.84
0.84
0.76
80
80
105
0.36
0.36
0.30
Wind Versus Sea Level
0.56
160
<0.32
Degreesof freedom (v) and critical correlationcoefficients(CR)
were obtained as in Table 2.
low-frequency near-surfacealongshorecurrents are approximately in balancewith the offshoresea-surfaceslope.Figure 9
showsthere is alsogoodcorrelationbetweenthe hourly alongshorecurrent at 26-m Sunflowerand sealevel: even very rapid
changesgenerallyoccurin both records.The correlationcoefficient betweenthesehourly time seriesis 0.86, essentiallythe
sameas for the low-passeddata (Table 4).
Thermal wind balance. Earlier work [Huyer et al., 1978]
has also shownthat the 'thermal wind' equation holds for the
60
'
F
'
M '
A
'
M '
J
'
J
'
A
7001
'
•
60
I,,]
'111•
I1"11'
31
•'111'
I,•1
III
,111
I III"1111"11
,,,I,,,,,,
,,,,11,
I11"
I"'111
'111""
IIIIIIIIIII
IIIIIIII1
,IllIt
.... •
"• I'* I•1111
I••1III•1111•
II
.....M
A
M
J
J
' '""'""
'111
....
'11%11•',
'"&•r','
......
"&
"'h""l
•"1h%"
'"'1'
50 •Orm•
o
o
• -50
0
øL
F
M
A
M
1975
J
J
A
M
A
M
J
J
1975
Fig. 5. Daily valuesof LLP (left) Sunflowercurrentswith Newportwind and adjustedsealevel,and of ILF (fight)
Sunflowercurrentswith ILF Newport wind and sealevel. Coordinatesfor currentsare rotated8øclockwisefrom north. Sea
level is referred to its 1975 mean.
A
HUYER
7002
ET AL.: SPRING
CURRENTS
TABLE 4. Linear Correlation Coefficients(CC) Between Linearly
Detrended ILF Time Series of Alongshore Currents at Sunflower,
Northward Wind and AdjustedSeaLevel, Using 134Daily Valuesfor
52 m, and 184 Daily Values for All Other Pairs
Depth, m
CC
v
CR99.9•
Current Versus Wind
26
52
76
92
0.47
0.53
0.58
0.63
Current
26
52
76
92
65
48
70
87
0.40
0.44
0.38
0.34
Versus Sea Level
0.84
0.85
0.85
0.80
42
32
48
56
0.48
0.55
0.44
0.42
OFF OREGON
may alsobe densitychangesin the deeperwater farther from
shore.
Computationsof the stericheightof the seasurfacerelative
to 600 db from hydrographicsections[Gilbert et al., 1976]
showvery goodagreementwith the low-passedsealevel data
(Figure 11). Stericheightat the coastwascalculatedfrom the
hydrographicsectionsusingMontgomery'smethod[Reid and
Mantyla, 1976]of extrapolationfor water shallowerthan 600
db. Similar calculationsusing a shallowerreferencelevel of
400 db show a smaller range (19 cm instead of 29 cm) and
therefore indicate that there are significant density changes
below 400 m. This may explain the poor agreementwhich
Reid and Mantyla [1976] found betweencoastalsealevel and
stericheightvariationsoff Oregonduringthe summerof 1972:
most of the repeated hydrographicsectionsthey used extendedonly deep enoughto directlycomputestericheight to
200 db.
Wind Versus Sea Level
0.54
65
DETAILS
0.39
Degreesof freedom (v) and critical correlationcoefficients(CR)
were obtained as in Table 2.
OF THE
1975
SPRING
TRANSITION
The transition is identified clearly not so much by anything
that occursduring the period of transition but rather by the
dramatic difference between the oceanographicregimes be-
(-0.60 and-0.66), with sealevelleadingdensityby about17
fore and after it. Nevertheless, a close examination of the
hours. What differences there are between the variations in
transitionperiod might reveal one or more distinctivecharacteristicsthat causethis changein the oceanographicregime.
In 1975 the spring transition occursbetween mid-March.
sealevel and densitycan probablybe explainedby the poor'
vertical samplingof the density,and by the fact that there
1
JUN
1
JUL
1
AUG
•t -,15
-50
o
i
i
i
!
i
1
1
1
M•¾
JU•
JUL
-25
-$0
$
-$
Fig. 6. Time seriesof the VLF alongshore
currentsat Sunflower,
with VLF adjustedsealeveland northwardwind at
Newport,andthe differencebetweenthe 26 and 92 m alongshore
current.
HUYERETAL.:SPRING
CURRENTS
OFFOREGON
FEB
MAR
APR
MAY
JUN
7003
JUL
5O
-50
Fig.7. LLPtimeseries
of thedifference
in thealongshore
velocity
between
26and92m, Sunflower,
andof thedensity
anomaly (sigma-t) at 92 m, Sunflower.
andmid-April(Figure6). The LLP timeseries(Figures5 and alsohavevery high valuesduringthis period(Figure7), but
7) indicatethat the transitionoccurswithin a 10-dayperiod they do not exceed,or barely exceed,valuesthat occurredearneartheendof March.In particular,thetransitionapparently lier. Although all of thesechangesare associatedwith the
beginswith the drop in sealevel on March 25; we shall exam- transition,they do not seemto be the distinguishing
characine the period around this date in detail. There is no obvious teristics,sincetheyare alsoassociated
with previousevents.
changein theonshore
component
of thewindduringthispeThe featurethat distinguishes
the springregimefrom the
riod-it is generallyonshorebothbeforeand duringthe tran- winter regime most clearly is the peristent vertical shear,
sitionevent.The southward
windduringthiseventis stronger whichis presentin springbut not in winter(Figures3, 6, and
than in any previousevent,but the differenceis not particu- 8). Sincetheverticalshearis balancedby an offshore
density
larly obviousin the LLP data (Figure 6). Similarly,the ad- gradient,the transitionmustincludea major changein the
justed sea level is lower, and the southward currents are offshoredensitygradientsoverthe shelf.Figure 10 indicates
stronger
thanin anypreviousevent.However,thehourlydata that sucha changeactuallyoccurred:by March 26, within a
(Figure 9), which retain all frequencies
lower than 12 cpd, day of the dropin sealevel,the differencein sigma-tbetween
showthat lower valuesof adjustedsealevel occurredat least 53 m at Pikakeand 52 m at Sunflower
exceeded
the sigma-t
once(about March 15) beforethe transition,and the near-sur- difference
of any previousevent.Althoughthe sigma-tdifferface currentwas more stronglysouthwardon February25 ence betweenthese two instrumentssubsequentlybecame
than duringthe first few daysafter the startof the transition negativeagain(e.g.,on April 26), positivehorizontaldensity
event.The LLP meanverticalshearand the bottomdensity gradientspersistedover the continentalshelf:certainlyat
DISTANCE FROM SHORE (l•rn)
,,,
!
!
o.
- ß
o./•
200
- /
winter
spring
y
wint•;
•/
spr•ng
./•,
/
_ -•.._ )
'spring 'winter
Fig.8. Distributions
ofthemean
andstandard
deviation
ofthealongshore
velocity
andthesigma-t,
forperiods
repre-
sentative
of winter(February
1,to March24,1975)andspring
(March26,to April25,1975),andthedistribution
of the
differences
between
winterandspring
meanvalues
ofthealongshore
current
anddensity.
-lo
12oo
1150
22
1
24
::: ....
1
FEB
HUYER ET AL.: SPRING CURRENTSOFF OREGON
7005
Fig. L0. Hourlytimeseries
of theverticalvelocity
differences
between
28 and53m at Pikake,andthelateraldensity
difference between 53 m at Pikake and 52 m at Sunflower.
about 100m betweenSunflowerand Wisteria (Figure 12), and
probablyalsoat about55 m betweenSunflowerand Wisteria,
where there is a persistentlypositivetemperaturegradientafter the transition [Gilbert et aL, 1976, p. 68]. Horizontal density gradientsalso occurredduringearlier southwardevents,
but they did not reachsucha large,apparentlycritical,value
over the inner shelf. We conclude that the change from the
winter to the spring regimesoccurswhen the horizontal density gradientsover the shelfreachsomecriticalpositivevalue.
At about 50 m, this critical value seemsto be at least about 1-
2 x 10-9 g cm-4 (between0.5 and 1 unit of sigma-tover 5 km,
the distancebetween Pikake and Sunflower).
It seemspossiblethat the criticaldensitygradientis set up
locally by southwardwinds which are sustainedfor at least
severaldays.To obtain a measureof the effectiveness
of each
wind event, we computed the cumulative onshore Ekman
transport(CET) for each event, beginningwhen the alongshore component changed sign and terminating when it
wind changesdirection(Figure 12). At any point in time the
value is proportionalto the Ekman transportsincethe beginningof thisevent.This measureof the wind stressclearlydistinguishesthe transition event from previous southward
events: the cumulative offshore transport is 5 times greater
than during any previousevent (Figure 12). This suggests
the
transitionmay be the resultof a sufficientlystronglocal wind
event.
To determine whether the wind stress alone could account
for the observedchange in the density distributionover the
shelf,we computedsomesimplevolumebudgets.The Ekman
transportcomputedfrom the alongshorecomponentof the
windstress
givesthe massflux normalto thecoast,f-•Y. Mass
must be approximatelyconservedover the continentalshelf,
or largedeeppressuregradientswoulddevelop.Thus an offshore directed Ekman transport will be balanced by an onshoretransportof deeper water which is transportedto the
surface near the coast. The offshoremoving water in the sur-
changed
signagain,i.e.,CET(t) = f,o'f-'•Y(t) dt, wheretois face layer is lessdensethan the onshoreand upwardmoving
the time the alongshorecomponentof the wind stressMy goes deep water, and the net volume(but not the mass)of water
throughzero at the beginningof a wind event.This resultsin over the continental shelf and slopewill be smaller after a pea sawtoothlike curve, which starts from zero whenever the riod of southwardwinds.The rate of changein volume will be
70
-20
t t
t
t
t
t
Fig.11. Steric
height
oftheseasurface
relative
to600db,withLLPadjusted
sealevelatNewport,
January-September
1975.
7006
HUYER ET AL.: SPRING CURRENTS OFF OREGON
I0
FElt
MAR
-5
Fig. 12. Hourly values of lateral densitydifferencesbetween 53 m Pikake and 52 m Sunflower,and between92 m Sunflower and 106m Wisteria, with cumulativeonshoreEkman transportfor eachwind event;the integral is setto zero whenever the low-passedalongshorewind changessign.
proportional to the differencein density betweenthe offshore for the 8-day period beginningon March 24, in order to commovingsurfacewater and the onshoremovingdeeperwater:
pute the change induced by the southwardwind event. The
mean southwardwind stressfor this period is 0.670 dynes
AV
I ?Y
I ?Y
?Y p•-p•
cm-2;the changein volumedue to the Ekmantransportfor
this periodis thus6.95 x 106cm3 per centimeterof coastline.
This value is 55% of the change calculated from the steric
The differencein volume can be estimatedindependently heightprofile.The windsover the watermay well be stronger
from the offshoreprofile of the sea surface.From hydro- than over the Newport jetty: Bakun's coastalupwelling ingraphicsectionson March 19 and April 1-2 we computedthe dicesfor 45øN, which are computedfrom 6-hourly synoptic
offshoreprofiles of the steric height of the surfacerelative to pressurecharts, account for a larger fraction (75%) of the
400 db. The stericheight remained about the samebetween20 changeobservedin stericheightprofiles.The densitydifferand 30 km offshorebut decreasedsignificantlywithin 20 km encesbetweenthe offshoreand onshoreflowingwatersmay
At
Pl f
P: f
f
PIP2
of the coast.The change in volume between thesedates was
also be underestimated.
We conclude that the local wind
1.27 x 107 cm3 per centimeterof coastline.Using the mean stressis probably sufficientto accountfor the changein the
southwardwind stress(computedfrom the Newport wind) for
the 14-dayperiod betweenthesehydrographicsections(0.116
dynescm-2) and a meansigma-tdifferenceof 1.5,we obtaina
volumechangedue to Ekman transportof 2.1 x 106 cm3 per
centimeterof coastline,about one sixth of the changeestimated from the steric height profiles.If, however,the northward wind stressat the beginningof this period had little further effecton decreasingthe densityover the shelf and upper
slope,the stericheightprofile might remain approximatelythe
sameduring March 19-23; indeed, sealevel at the coast(Fig-
densitydistributionover the shelf.
Recenttheoreticalwork hassuggested
alternativehypothesesfor the causeof the springtransition:a changein current
and densityregimemightbe a propagatedeffectgeneratedby
windssouthof the region[Allen, 1976;Gill and Clarke, 1974],
or major changesin the current may propagatepoleward
along eastern ocean boundaries as internal Kelvin waves
[McCreary, 1976; Hurlburt et al., 1976]. Since the transition
eventis clearlyassociated
with a changein the currentregime,
it seemsparticularly appropriateto examine the evidencefor
ures 9 and I l) has about the same value on March 19 and 23. alongshorepropagation of this event. Current observations
It is therefore more appropriate to use the mean wind stress were available during the transition event at two other 1oca-
HUYER ET AL.: SPRING CURRENTS OFF OREGON
7007
tionsoverthe continentalshelf,at 46ø09'N, 124ø15'W [Hickey lag betweenSan Franciscoand Newport would be more than
et al., 1979],and at 49ø02'N, 126ø14'W [Huyeret al., 1976].At 2 days; baroclinic coastal-trappedwaves have even slower
both theselocationsthe total water depthwasbetween90 and phase speeds.Along the entire coastthe almost simultaneous
100 m, i.e., about the same as at Sunflower. The currents at
drop in sea level is associatedwith strongsouthwardwinds.
the three locationshave very similar fluctuations[Hickey et Sobey [1977] examined the hypothesisthat the sea level flucal., 1979]. Figure 13 showshourly valuesof the alongshore tuations, including the transition event, are forced by variacomponentof the near-surfacecurrent at each location for the tionsin the wind field southof the region;he concludedthat
March 21-31 period which spansthe transitionat Sunflower; the local wind accountsfor the changesabout as well as the
it showsthe generalsimilarity of the currentsbut also shows Gill and Schumann[1974] model using winds from locations
there are large diurnal variationsat the northern location. The farther south.Thus the transitionis probablygeneratedby the
data from both of the southernlocationsshowa clearchange action of the local wind, though it must be remembered that
in directionof the alongshorecurrenton March 25, which oc- the wind is coherent over scalesapproachingor exceeding
curs earlier
at 46ø09'N
than
at 45ø00'N.
The
data from
1000 kin.
49ø02'Nalsochangesignat aboutthistime, apparentlyat the
sametime as at 46ø09'N. Thus there is no lag between46009'
If the transitionfrom winter to springregimesis truly generatedby the localwind (as suggested
by the roughagreement
and 49ø02', a distance of 300 km, and the currents at both
in volume estimates and the lack of evidence for northward
these locations reverse earlier than the current at 45 øN.
propagation), the event beginning on March 25 is a classical
example of wind-driven coastal upwelling. Unlike the up-
A drop in sea level associatedwith this transition event occursalong the entire coastfrom San Franciscoto Torino, Brit-
wellingeventsobservedduringCUE I and CUE II [Huyeret
aL, 1974;Halpern, 1976],this event occurswith a prior situaoccursvery early on March 25 at San Francisco, and several tion of approximately level isopycnals;it is therefore more
hours later at CrescentCity; the lag betweenSan Francisco suitablefor comparisonwith numericaland analytical models
and Newport (a distanceof about 750 kin) is no more than 6 of time-dependentcoastalupwellingsuchas thoseby O'Brien
hours.Phasespeedsfor first-modebarotropicshelfwavesvary and Hurlburr [1972], Thompsonand O'Brien [1973], and Allen
alongthe coastfrom 95 to 330 km day-' [Clarke, 1977];if the [1973]. Figure 14 showstime seriesof selectedvariableson an
ish Columbia [Osmerand Huyer, 1978].The drop in sealevel
drop in sea level propagatednorthward as a shelf wave, the
expanded time scale for the period March 21-30, 1975. The
lOO
49ø 02'N
22 22 23 2't2••v••• • V• '1•
49 ø 02'N
ß
46ø09' N
ß
•
21
•
22
•
..
.
•""
;'
•:'.
•
I
23
!
28
t
29
t
30
-5O
45 ø O0'N
hours
I
26
2)
ß*.
I
I
I
28
29
•0
:
Fig. 13. Hourlytimeseriesof the alongshore
currentTorino(at 49ø02'N),off the ColumbiaRiver (at 46ø09'N),and at
Sunflower(at 45ø00'N), March 21-31, 1975.
I
10
1200
1175
1150
1125
I
I
I
I
I
I
I
I
I
22
23
24
25
26
27
2•
29
30
•A
•
SeaLevel
I
22
'
"'"'"""• 7hours
I
I
!
23
24
25
;
92m
27
ß
Sunflower
Inverted
See
Level--..•.....:'
:"'"
5hour
;50-hours
PiSk3er•e'•
Sun•
16:•
er
21
22
23
24
25
26
27
2•
29
30
31
InvertedSea Level----T.:": """
ß
ß
ß' '
J•,_,•,,,
.•,,.•Q•.•,I
24' '
22
23
•
-
?
.••-$0, hours
2•
•
24
Fig. 14. Hourlytimeseries
of selected
parameters
for theperiodsurrounding
thetransition,
March21-31, 1975:north-
wardwind;adjusted
sealevel;alongshore
current,
26m Sunflower;
sigma-t
at near-bottom
current
meters;
sigma-t
differencebetweenPikakeand Sunflower;
and the velocitydifference
betweenthe top and bottominstruments
at Sunflower.
Dottedcurvesfor March24-27 arerepeated
fromupperpanelsto showthetimelagsbetweenvariables.
The dashedlines
in the bottom panel were sketchedto showthe gradualincreaseof the verticalshearfrom March 25 to March 30 to its
stable value.
I
HUYER ET AL.: SPRING CURRENTSOFF OREGON
7009
io
•
1976
Fig. 15. The cumulative
onshore
Ekmantransport
for eachwindevent,1973-1976,
andadjusted
sealevelat Newport,
1976.The cumulativeEkmantransportis setto zeroeachtime the LLP alongshore
wind changes
sign.
drop in sea level lags the southwardwind by only 2 hours. is consistentwith the valuespredictedby the numerical modThe changeto southwardcurrentat Sunflowerlagsthe drop els of O'Brien and Hurlburt [1972] and Thompsonand O'Brien
in sealevel by another7 hours(and thereforethe changein [1973],and by Allen's[1973]analyticalmodel.
Vertical velocitiesduring this upwelling event were estiwindby about9 hours).The increase
in densityoccursfirstin
the bottom water over the mid-shelf (where it lagsthe drop in mated from the changein isopycnaldepths.From 12-hourly
sea level by about 15 hours) and later in the bottom water vertical sectionsobtained from the hourly sigma-t time series,
overthe innershelf(whereit lagssealevelby about30 hours). we found that the largestchangein isopycnaldepth seemsto
The suddenincreasein the lateral densitygradient at 52 m be- be about 40 m in a 12-hour period, correspondingto a mean
tween Pikake and Sunfloweralso lags sea level by 30 hours; verticalvelocityof 0.09 cm s-• for this period.This is about
this increasein densitygradientis due almostentirely to the the same order as the maximum vertical velocities estimated
rapid densityincreaseat Pikake. The differencebetweenthe for other upwellingregions[e.g.,Bartonet al., 1977].
near-surfaceand near-bottomcurrentshasstrongtidal oscillaSPRING TRANSITIONS IN OTHER YEARS
tions which make it more difficult to estimatethe lag, but it is
The 1975 data clearly suggestedthat there is a transition
clear that the shear beginsto develop very soon after the
changein surfacecurrent(within 12 hours)but it doesnot from the winter regimewith no mean offshoredensitygradireach its final stable value until about 5 days later, on March entsto the springregime,with strongoffshoredensitygradi30. Already on March 26, the southwardsurfacecurrent is ents whenever the offshore Ekman transport exceeds some
strongerat Pikake and Sunflowerthan at Wisteria; in this minimum value (Figure 12). If suchan event is followedby
sensea southward 'coastaljet' is establishedwithin a few very strongnorthward winds, the winter regimecan perhaps
hours.The 'barocliniccoastaljet' also beginsto develop very
be re-established;otherwise,the spring regime would persist
quickly,but its developmentprobablycannotbe said to be throughspringand into early summer.Figure 15 showstime
completeuntil the mean verticalshearhas reachedits final seriesfor several years of the cumulative onshore Ekman
value 5 daysafter the beginningof the event.This time scale transportfor each wind event:the integrationbeginsagain
7010
HUYER ET AL.'- SPRING CURRENTS OFF OREGON
from zero each time the low-passedalongshorewind stress verticalshear)to the springoceanographic
regime(with isopycnals sloping upward toward the coast, southward surface
Ekman transport resulting from each event; the time series currents,and significantvertical shear) occursvery rapidly,
within a period of severaldays. Before the transition,wind-inthus readily showswhich eventshave the greatesteffect.
In 1973the first major southwardevent beganabout March duced current fluctuations are significantly baroclinic and
23; its total offshoreEkman transportwas 2.7 x 109gcm -I, have an offshore length scale about the same as the shelf
and it was not followed by any major northward wind events. width; after the transition,the fluctuationsare more nearly
The data from Poinsettia(Figure 2) showthat southwardsur- barotropic and confined to about the inner half of the shelf
face currents and low sea level persisted after this event, [Huyer et al., 1978].The transitionseemsto be the result of a
throughspringand early summer1973.Data from the south- single coastal upwelling event (driven by the local wind)
ern Washingtonshelf [Smith et aL, 1976,p. A-88] show that whosecumulative offshoreElanan transporthas some mini-
changessign.The heightof the peaksis a measureof the total
persistentsouthwardsurfacecurrentsthere also began on
March 23, 1973.
In 1974,southwardeventswith morethan 1.3 x 109gcm -I
of offshoreEkman transport began on February 5, March 18,
and April 15. The two earlier eventswere succeededby very
mum value(probablyabout 109 gcm-I) sufficientto establish
stronglateral densitygradients(at least 1-2 x 109 gcm -4)
overtheshelf.The lateraldensitygradients
canbe destroyed
by very strongunfavorablewinds,but they persistthroughperiods of moderate northward wind stress(with total onshore
Ekmantransportsof up to about4 x 109 g cm-I). Thesepersistentlateral densitygradientsare associated
with a persistent
effect of the two earlier southward events was destroyed by southwardsurfacecurrent:this current is initially generated
the very strong northward events but that the effectsof the by a southwardwind event, but its persistencedoes not depend on continued southwardwinds. Although in the very
third southwardevent persisted.
In 1975the first major southwardevent beganon March 25; low frequency data (e.g., Figure 3) the maximum southward
its total offshoteEkman transportexceeded5.4 x 109gcm -I. surfacecurrent leads the maximum southwardwind by sevAlthough it was later followed by two strong northward eral months,this southwardcurrent is apparentlygenerated
J
events,with onshoretransportsexceeding2.5 x 109gcm-I in entirely by the local wind.
The persistenceof the densitygradients,the southwardcurone caseand 3.5 x 109gcm-I in the other,thesewereapparently not strongenoughto re-establishthe winter regime(Fig- rents, and the vertical shear after this first strongupwelling
event is puzzling. There is indirect evidencethat the vertical
ure 5).
From these3 years we estimatethat the minimum offshore shear reduces the effectiveness of events with offshore Elanan
transportto setup the springregimeis only 1.3x 109 gcm-I transportas well as thosewith onshoreElanan transport:the
but that the onshoretransportrequired to destroythe spring southwardsurfacecurrent doesnot get appreciablystronger
regimeand re-establishthe winter regimeis about4 x 109 g with subsequentupwelling events.Why shouldan oceanwith
cm-I. Applyingtheseresultsto the 1976data, we would guess sloped isopycnalsapparently responddifferently to the same
that the southwardeventin March would be offsetby the sub- wind stressthan one with level isopycnals?We know of no
sequent
strong
northward
event.It isnotimmediately
obvious theoreticalstudiesthat directly addressthis problem, though
whether the southward event of early April is strong enough the potentialimportanceof a meanshearin modifyingthe dyby Mysak [1979].
to establishthe springregime,or whetherthis wasdelayed un- namicsof shelfwavesis discussed
Since the rapid transition in the currentsseemsto be a ditil the southward event beginning on May 6, 1976. The sea
level data from 1976 (Figure 15) is somewhatambiguous,but rect responseto the local winds, one might also expect that
persistentlylow sea level did certainly not begin in early there is a rapid transition in the atmosphere.The seasonal
April: it seemsto begin with the rather weak southwardevent variation of the very low frequency wind (Figures 2 and 6)
near the end of April. At first, this seemsinconsistentwith the doesnot sho• any rapid changes.Rather,the wind changes
resultsof previousyears,but the northwardwinds at the end slowly from its northward mean in winter to its southward
of April are so weak that the period from about April 25 mean in summer;in spring the mean wind is very weak. On
through the first week of May shouldprobably be considered the event timescale, however, there seems to be a clearer disas a single southward event. We conclude that southward tinction betweenwinter and spring(Figure 15):in winter there
wind eventswith offshoreEkmantransports
greaterthan 109g is a seriesof strongnorthwardeventswith very strongonshore
Ekman transport;in springthesestrongnorthward eventsare
cm-I are sufficientto setup the springregime.
The asymmetryin the amount of Ekman transportneeded absent.In someyears (e.g., 1974-1976) the strongnorthward
to establishthe spring regime with upward-slopingisopycnals eventsceasequite abruptly in early spring(Figure 15). Bryson
and mean vertical shear, compared to the amount needed to and Lahey [ 1958]attributed the abrupt onsetof springin Wisre-establish the winter conditions with no mean shear, in- consin to the decreasingwave number of the atmosphere:
dicates a difference in the essentialdynamics of the two re- sincethe number of wavesaround the earth is an integer less
gimes.The presenceof upward-slopingisopycnalsseemsto than 10 or so, the changesfrom one seasonalregime to the
add to the stability of the system.Models of coastalupwelling next could be abrupt.
In summary, the transition from winter to spring oceanoeventsin summershouldthereforebegin with upward-sloping
graphic conditionsoccursvery rapidly over the Oregon shelf.
rather than level isopycnals.
The transition seemsto be the result of a wind-generated
coastalupwelling event which resultsin upward-slopingisoDISCUSSION
pycnalsand stronglateral density gradients.For somereason
We have shown that the transition from the winter oceanothat is not understood,theselateral densitygradientsand the
graphic regime (with generallylevel isopycnalsand variable associatedvertical currentsheartend to persistin spiteof subbut predominantly northward currentswith little or no mean sequentmoderatelyunfavorablewind events.The chief result
strongnorthwardeventswhoseonshoreEkman transportexceeded5.4 x 109gcm-I. The Poinsettiadata indicatethat the
HUYER ET AL.: gPRING CURRENTSOFF OREGON
of the transition
is a baroclinic
southward
coastal
current
which persiststhroughspringand summer.
Acknowledgments. We are grateful to John Allen and Dale Pillsbury for many valuable discussions,
and to Barbara Hickey for providing the current data from her mooring near the Columbia River.
Bill Gilbert, Henry Pittock, and Joseph Bottero assistedin the data
analysis.This study was supportedby the National ScienceFoundation through grants DES 74-22290 and OCE 77-07932 and by its Office for the International Decade of Ocean Exploration throughgrant
OCE
78-03381.
REFERENCES
Allen, J. S., Upwelling and coastaljets in a continuouslystratified
ocean, J. Phys. Oceanogr.,3, 245-257, 1973.
Allen, J. S., Some aspectsof the forcedwave responseof stratified
coastalregions,J. Phys. Oceanogr.,6, 113-199, 1976.
Barton, E. D., A. Huyer, and R. L. Smith, Temporal variation observed in the hydrographic regime near Cabo Corveiro in the
northwestAfrican upwellingregion,Februaryto April 1974,Deep
Sea Res., 24, 7-23, 1977.
Bryson,R. A., and J. F. Lahey, The march of the seasons,Rep.
ASTIA A0-1525 00, 41 pp., Dep. of Meteorol., Univ. of Wisc.,
Madison, 1958.
Clarke, A. J., Observational and numerical evidence for wind-forced
coastaltrappedlong waves,J. Phys.Oceanogr.,7, 231-247, 1977.
Collins, C. A., and J. G. Pattullo, Ocean currents above the continen-
tal shelfoff Oregonas measuredwith a singlearray of current meters, J. Mar. Res., 28, 51-68, 1970.
701 1
overthecontinental
shelfoff OregonandWashington,
J. Geophys.
Res., 80, 3495-3505, 1975a.
Huyer, A., R. D. Pillsbury,and R. L. Smith, Seasonalvariation of the
alongshore
velocityfieldoverthecontinental
shelfoffOregon,Limnol. Oceanogr.,20, 90-95, 1975b.
Huyer, A., J. Gagnon, and S. Huggett,Observationsfrom currentme-
tersmooredoverthe continentalshelfoff VancouverIsland,28 No-
vember1974and8 April 1975,andrelatedoceanographic
andmeteorological
data,Tech.Rep.4, Environ.Can.,Fish.andMar. Serv.,
Mar. Environ. Data Serv.,Ottawa, Ont., 1976.
Huyer, A., R. L. Smith, and E. J. C. Sobey,Seasonaldifferencesin
low-frequencycurrent fluctuationsover the Oregoncontinental
shelf,J. Geophys.Res.,83, 5077-5089, 1978.
Kundu, P. K., and J. S. Allen, Some three-dimensional characteristics
of low-frequencycurrentfluctuationsnear the Oregoncoast,J.
Phys. Oceanogr.,6, 181-199, 1976.
McCreary,J., Easterntropicaloceanresponse
to changing
windsystems:With application
to El Nino, J. Phys.Oceanogr.,
6, 632-645,
1976.
Mysak,L. A., Recentadvances
in shelfwavedynamics,
Rev.Geophys.
SpacePhys.,17, in press,1979.
O'Brien,J. J., andH. E. Hurlburr,A numerical
modelof coastalupwelling, J. Phys.Oceanogr.,2, 14-26, 1972.
Osmer,S. R., and A. Huyer,Variationsin the alongshore
correlation
of sealevelalongthe WestCoastof North America,J. Geophys.
Res., 83, 1921-1927, 1978.
Pearson,
E. S.,andH. O. Hartley(eds.),BiometrikaTables
for Statisticians,vol. 1,270 pp., CambridgeUniversityPress,New York, 1970.
Pillsbury,R. D., J. S. Bottero,R. E. Still, and W. E. Gilbert, A compilationof observations
frommooredcurrentmeters,vol. VII, OregonContinentalShelf,July-August1973,Data Rep.58, Ref 74-7,
87 pp., Sch.of Oceanogr.,OregonStateUniv., Corvallis,1974.
Davis, R. E., Predictabilityof sea surfacetemperatureand sea level
flowupon
pressureanomaliesover the North PacificOcean,J. Phys.Ocean- Reid,J. L., andA. W. Mantyla,Theeffectof geostrophic
coastalsea elevationsin the northern North Pacific Ocean, J.
ogr., 6, 249-266, 1976.
Geophys.Res., 81, 3100-3110, 1976.
Gilbert, W. E., A. Huyer, E. D. Barton,and R. L. Smith, Physical
oceanographic
observations
off the Oregoncoast,1975,Wisp and Smith, J. D., B. Hickey, and J. Beck,Observationsfor mooredcurrent
meterson the WashingtonContinentalShelffrom February1971
UP-75 data,Data Rep.64, Ref 76-4, 189pp., Sch.of Oceanogr.,Oregon State Univ., Corvallis, 1976.
to February1974,Spec.Rep.65, Ref M76-44, 347 pp., Dep. of
andgeneration
offluctuations
incurrents
andsealevelonthePa-
vailis,
1977.
toJ.Phys.
Oceanogr.,
1979.
welling,
J.Phys.
Oceanogr.,
3,33-46,1973.
Oceanogr.,Univ. of Wash., Seattle, 1976.
Gill, A. E., and A. J. Clarke, Wind-inducedupwelling,coastalcurSmith, R. L., A descriptionof current, wind, and sea level variations
rentsand sea-levelchanges,DeepSea Res.,21, 325-345, 1974.
duringcoastalupwellingofftheOregoncoast,July-August1972,J.
Gill, A. E., and E. H. Schumann,The generationof longshelfwaves
Geophys.Res., 79, 435-443, 1974.
by the wind, J. Phys.Oceanogr.,4, 83-90, 1974.
of Oregonshelfwatersto wind fluctuaHalpern, D., Structureof a coastalupwellingeventobservedoff Ore- Sobey,E. J. C., The response
tions:Differences
and the transitionbetweenwinterand summer,
gon during July 1973,Deep Sea Res.,23, 495-508, 1976.
Ph.D. thesis,153pp., Sch.of Oceanogr.,OregonStateUniv., CorHickey, B. M., A. Huyer, and R. L. Smith,The alongshorecoherence
cific
Northwest
continentalshelf,
winter
andspring
1975,
submitted
Thompson,
J. D., andJ. J. O'Brien,
Time-dependent
coastal
upandG. K. C. Clarke,PreHurlburr, H. E., J. C. Kindle, and J. J. O'Brien, A numerical simula- Ulrych,T. J., D. E. Smylie,O. G. Jensen,
dictivefilteringandsmoothing
of shortrecords
by usingmaximum
tion of the onsetof E1Nino, J. Phys.Oceanogr.,6, 621-631, 1976.
entrophy,J. Geophys.
Res., 78,4959-4964, 1973.
Huyer, A., R. L. Smith, and R. D. Pillsbury,Observtionsin a coastal
upwellingregionduring a periodof variablewinds(Oregoncoast,
July 1972), Tethys,6, 391-404, 1974.
Huyer, A., B. M. Hickey, J. D. Smith, R. L. Smith, and R. D. Pillsbury, Alongshorecoherenceat low frequenciesin currentsobserved
(ReceivedApril 30, 1979;
revisedJuly 13, 1979;
acceptedJuly 16, 1979.)