JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 104, NO. B9, PAGES 20,431-20,447, SEPTEMBER 10, 1999
Crustal structure beneath the central Oregon convergent
margin from potential-field modeling: Evidence for a
buried basement ridge in local contact with a seaward dipping
backstop
SeanW. Fleming
1andAnneM. Tr•hu
Collegeof OceanicandAtmosphericSciences,
OregonStateUniversity,Corvallis
Abstract. Modelsof magneticandgravityanomaliesalongtwo E-W transects
offshorecentral
Oregon,oneof whichis coincident
with a detailedvelocitymodel,providequantitativelimitson
the structureof the subducting
oceaniccrustandthe crystallinebackstop.The modelsindicate
thatthebackstop-forming
westernedgeof the Siletzterrane,an oceanicplateauthatwasaccretedto
NorthAmerica-50 million yearsago,hasa seawarddip of lessthan60ø3. Seismic,magnetic,
andgravitydataare compatiblewith no morethan2 km of subductedsedimentsbetweenthe
Siletzterraneandthe underlyingcrystallinecrustof the Juande Fucaplate. The dataalsosuggest
thepresence
of a N-S trending,200-km-longbasalticridgeburiedbeneaththe accretionary
complex fromabout43øNto 45øN. Althoughtheheightandwidthof thisridgeprobablyvary along
strike,it may be up to 4 km highandseveralkilometerswide in placesandappearsto be locally
in contactwith the SiletzterranebeneathHecetaBank. Severalmodelsfor the originof thisridge
are discussed. These include: a sliver of Siletz terrane detached from the main Siletz terrane dur-
ing a lateEoceneepisodeof strike-slipfaulting;imbricationandthickeningof subducted
oceanic
crustin place;an aseismicridgeraftedin on the subducting
oceaniccrustduringthe past 1.2 million years;anda seriesof ridgesand/orseamounts
raftedin overa longerperiodof time andtransferredfrom the subductingplateto the overlyingplate. The lastmodel is the mostconsistent
with the complicatedhistoryof localuplift, subsidence,
and slopeinstabilityrecordedin the
ridges,basins,andbanksof this part of the margin. We speculatethat the massiveseawarddipping westernedgeof the Siletzterranein this regioninhibitssubductionof seamountsand sediments,resultingin fomationof buriedridgeasthe accumulated
flotsamandjetsomof subduction.
This processmay alsobe responsible
for thickeningof lower accretionarycomplexmaterial,oversteepening
of slopesleadingto massiveslumping,andnorth-south
extensionthroughstrike-slip
faultingin the accretionarycomplexto the westof the buriedridge. Regardlessof its origin,the
ridgemay currentlybe actingas an asperityinhibitingsubduction.
1.
Introduction
Subduction of the Juan de Fuca plate beneath northern
California, Oregon, Washington, and British Columbia has
produced a geologically complex convergent margin (Figure
l a). The distributionof seismicactivity, however, is unusual
for this tectonic setting. Although great interplate earthquakes have occurred during the Holocene [Nelson et al.,
1995], interplate seismicity is almost absent fromthe historical record,and the distributionof both upper and lower plate
seismicityis very heterogeneous(Figures lb and lc). Earthquake activity is commonin Washington and northern California [e.g., Weaver and Shedlock, 1996] but is infrequent in
central and southernOregon. In Washingtonthe high level of
seismicactivity may result from northwardmotion of a Paleocene to early Eocene accretedbasaltic terrane generally referredto asthe Siletz terrane [Snavelyet al., 1968] relative to
INowatDuncan,
British
Columbia,
Canada.
Copyright1999by the AmericanGeophysical
Union.
Papernumber1999JB900159.
0148-0227/99/1999 JB900159509.00
pre-Tertiary North America [Wells et al., 1998]; here
forearcseismicitygenerally outlines blocks of Siletz terrane
[Trihu et al., 1994; Parsons et al., 1998]. In northern California, seismicityin both the upper and lower plates seemsto
be controlledin large part by the northward motion of the Pacific plate [Wang, 1996]. Understanding the nearly aseismic
behaviorof both the upper and lower plates of the central part
of the subductionzone requires a better understanding of the
crustal structure in this region in order to determinewhether
this is a locked zone with the potential for great earthquakes
or simply a node in the regional stressregime.
To this end, we have modeled marine magneticand gravity
anomalydata recordedalong two E-W profiles. The original
goals of this study were to (1) use potential-field data to regionally extend constraintson crustal structure obtained from
a previousseismicreflection-refractionsurvey(Figure 1d); (2)
constrain the geometry of the backstop-formingwestern edge
of the Siletz terrane; (3) constrainthe structurebeneath Heceta
Bank, a shallow submarinebank on the central Oregon margin; and (4)evaluate the influence of backstop structure on
deformationin the accretionarycomplex. In the course of this
work, we also identified a previously unknown, deeply buried basementridge on the subductingplate, which provides a
possible explanation for the complicatedgeologic history of
20,431
20,432
FLEMINGAND TRI•HU: BURIEDRIDGEBENEATHTHE OREGONMARGIN
forearc that has significantly greater shear strength than
this part of the marginand whichmay currentlybe actingas an
asperitybetweenthe plates.
trenchward
sediments
and which
thus
acts as a bulldozer
blade driving the accretionarywedge [Davis et al., 1983;
Wang and Davis, 1996]. Differing backstopgeometriespro-
1.1. The Siletz Terrane
duce different stress and strain fields within an accretionary
prism[Byrneet al., 1993], and the dynamicsof sedimentsubduction, melangeformation,and prism accretionare in part a
function of the geometry and properties of the backstop
The thick basaltsof the Siletz terrane formthe backstop to
the accretionary complex on the central Oregon continental
margin. A backstop may be defined as a region within a
128
ø
A.
C'_120ø
B.128W
__122W
..-
124
ø
'
. -,,
....................
48ø
. ;•, ....
:....';..'..'
•
•
, •,_.•:•;g•....
::::::::::::½:::;½::
::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::
ß I f.k.'•7•, •.' '-
::::::::::::::::::::::::::::::::::::::::::::::•:ß
•
. :•: • .-:::::::::::::::::::::::::::
•,
46ø
• ::............
•........
•
Figure
2
- ½?'
•
•
43N
Juande Fuca
Plate
':
-.s_,-.... -:,..,:
-•,• •'•... ? t• •::•::•::•::•::•:•::•::•:
"•::•½•;:
....
I'•:J:•:E[•:[[•:::•:.?
...•1
•
•.:7::•::::::•
x
Pacific
ß ['C•_•
•
'
'.•,'•
•.._• x.
42
ø Plate
....
:::.:.....•
••
"-•
Men•
0
200• Transform
I •,
43N
Eocene
Quatem• F•.lt• II ,,•"•.
Vo•c•o.• Accrefion•
com½.• m•r,c
c•t vo•c•ic•oc•
I
I
...........
I'"1
I,
.
x'J•;•;•" ,
, "t,;", "•
0
200 km
i
D
Buried
base-
'
Deformation
Frontment
ridge
W
I
NSProfile
J'
"
0 ...........
Coast
:...;••....'
__
i
Western High
I
' Cascades
.•E
I Cascades
.:.,.,:,;.,..,--:..;.::.::'..•.:.::•::::.•:.-..--..•:.:.`...L.:•..:;•:•:•:•:•:•:•:•:•:•::;::•:•:•;•:•:•:•:•:•:•:}:•:•:•:}:•:•
:' 'x <5.2
10
,6.3
•% 20
•6•
* '••'m•mum
' •;'•
....
:.....
:...... of •letz
, ' -...........
tSmKness
•/u,--,•••:½•__..•
• 40 Accretionary
.............
•:• Eo•
•?••
thickness
ofSiletz
Juan
deFuca ........................................
.....
::•:•:•;•:•;•;•::::•::•
...... Maximum
Complex mafic
crust oceanic
crust
7.•
.......
•::•::•::•::•::•::•::•::•
......
'
50•
'
'
'
I
.....
I
....
I
....
I
....
I
'
'
'
I
....
I
....
I
'
VE2:I
'
'
I
....
I
....
I
....
35 60 85 110 135 •60 185 210 235 260 285
DISTANCE (kin)
I
FLEMING AND TR_I•HU:BURIED RIDGE BENEATH THE OREGONMARGIN
[Shrieveand Cloos, 1986; yon Huene et al., 1996]. The geometryof the backstopthus helps to determinesubduction
zone dynamicsand is an importantconsideration in earthquakehazardassessment.Thermalmodelsof the Cascadia
subductionzone are also sensitiveto the geometryand extent
20,433
al., 1995] (seealso Figure 3). Both the N-S strike of ubiquitous anticlines
and thrust faults and the northwest
strike
of
active left-lateral strike-slipfaults in the accretionarycomplex
offshoreOregon indicate that the maximumhorizontal compressivestressis generallyorientedE-W seawardof the backof the low thermal conductivity basalts of the Siletz terrane stop [Goldfingeret al., 1992, 1997]. In contrast,the orienta(Hyndmanand Wang, 1993; Oleskevich,1994), and these tion of maximumhorizontalcompressivestressin the onshore
modelsare critical for predictingthe size of the locked zone forearcis N-S [Werner et al., 1991; Ludwin et al., 1991].
Moreover, northwest trending, left-lateral strike-slip faults
and thusthe magnitudeof a possiblemegathrustearthquake.
The rocksof the Siletz terrane are tholeiitic submarinepilthat are commonin the accretionaryprism appearto terminate
low lavas and brecciaswhich grade locally into subaerially eastwardnear the seawardedge of the Siletz terrane [Goldfin(3-4 km)
eruptedalkalic basalts[Shayelyet al., 1968; Simpsonand ger et al., 1997]. Finally, a bandof short-wavelength
Cox, 1977;Duncan, 1982], andthey are locally known as the margin-parallelfolds and faults, which contrastssharply with
Siletz River Volcanics in Oregon,the CrescentFormation in longer-wavelength(6-8 km) folds that occur throughout the
Washington,and the MetchosinVolcanicsin British Colum- rest of the accretionarycomplex,overliesthis boundary in the
bia. Seismicdata indicatelarge variations in the thickness of regionof this study. StonewallBank, an actively growing anthis terrane. Its thickness is greatest beneath west central ticline that has recently been studied in detail by Yeats et al.
[1998], may be a particularlywell-developedexampleof one of
Oregon [Trdhu et al., 1994], where it reaches24-32 km (Figthese folds.
ure ld)and the velocity structure resemblesthat of the Ontong Java plateau and other large igneous provinces [Hus- 1.2. Heceta Bank
song et al., 1979; Coffin and Eldholm, 1993]. It thins to ---20
The short-wavelength
folds overlyingthe apparentseaward
km beneathnorthwesternOregon and southwest Washington
[Trdhu et al., 1994; Parsons et al., 1998] and has approxi- edgeof the Siletz terraneare particularlyprominentover He1996), an enigmatelythe samethickness as normal oceanic crust where it is cetaBank(C. Hutto,personalcommunication,
thrust beneath the pre-Tertiary terranes of Vancouver Island maticbathymetrichigh located at approximately44ø00'N to
44ø30'N and extendingwestward on the continentalshelf to
[Hyndman, 1995]. Several investigators [Wells et al., 1984;
Trehu et al., 1994; Parsons et al., 1998; Wells et al., 1998] about 125ø00'W (Figures 2a and 2b). Although the bank is
have suggestedthat these variations have influenced long- associatedwith a large (-50 mGal) free-airgravity high (Figand short-termforearc deformation,with the thickest part of ure 2a), the simple spatial correlation between gravitational
the terrane acting as a coherent rotating block that deflects and bathymetricanomaliesthat would be expectedif the gravseismicand aseismicdeformationto its edges. The gravity ity field reflectedonly topography and large-scaleplate geanomalydata (Figure 2a) reflectthe presenceof dense Siletz ometry is not observed. Magnetic data suggestthat the Siletz
rocksnear the surfacebeneaththe Coast Range data but are of terranedoesnot extend sufficiently far west to entirely underlimited use in mappingthe boundariesof this terranebecause lie Heceta Bank (Figure 2b).
Heceta Bank and other banks of the central Cascadiamargin
the signal is overprinted by gravity lows associated with
have been interpretedto be an outer arc high formedby undershelf basinsto the west and the Willamette Valley to the east.
The magnetic anomaly data (Figure 2b) and drill holes thrusting of young sedimentsat the base of the accretionary
[Shayely, 1987; Shayely and Wells, 1996], on the other hand, complex [Kulm and Fowler, 1974]. However, it remains
clearly show that the western edge of the Siletz terrane lies poorly understoodwhy the outer arc high exists as a number
beneaththe continentalshelf of central Oregon between 43øN of discretebanksrather than as a continuous bathymetricand
and 45øN. Shayelyet al. [1980] suggestedthat this boundary, structural feature along the entire margin. Recent regional
which they namedthe Fulmar fault, was a right-lateral strike- mappingof the Miocene-Plioceneunconformitysuggeststhat
slip fault duringthe Eocenethat truncatedthe accretedSiletz the outer arc high was once more extensive,extendingnorth of
terrane. Although this boundaryis currentlyaseismic,several Heceta Bank to 46øN beneath what is now the continental
observationssuggestthat it separatesregions with different slope, suggesting post-Miocene collapse of the margin
stressregimesand deformationalstyles and that it has been through gravitational collapse and/or subduction erosion
the site of Neogenedeformation[Shayelyet al., 1980; Trdhu et [McNeill et al., 1998].
Figure 1. (a) Generalized geological and tectonic map of the Cascadiasubduction zone, showing plate
boundaries,Quaternaryvolcanicrocksof the presentarc, Eocenemafic crustof the Siletz terrane in the forearc
(exposedon land,andinferredoffshore),andcrustalfaults. Map is modifiedfrom Wells et al. [1998], with additionaloffshorefaultsfrom Goldfingeret al. [1997]. The locations of the cross-sectionof Figure l d and of
the potentialfield mapsof Figure2 are alsoshown. (b) Earthquakeswith hypocentersshallower than 30 km
from 1980 to September1993 [from Weaverand Shedlock,1996]. Except for near the Mendocino triple junction and in the oceanbasin,theseeventsare in the North Americanplate. (c) Earthquakes with hypocenters
deeperthan30 km from 1980to September1993 [fromWeaverand Shedlock,1996]. Theseevents are in the
subductingJuande FucaYGorda
plate. Note the lack of seismicactivityin eitherplate beneathcentral Oregon.
(d) Velocity modelfor a transectacrossthe Cascadiaforearcin centralOregon. [Reprintedwith permission
fromTrihu et al., 1994. Copyright 1994 AmericanAssociation for the Advancementof Science.]Velocity
contoursare labeled in km/s. This model provided a starting point for the potential-field modeling of the
northernline. The modelwasderivedfrom an offshoremultichannelseismicprofile, which was also recorded
on oceanbottomand onshoreseismometers,
and from severalonshorerefractionprofiles. The seawardextent of
the Siletz ter.rane is indicatedby a sharplateral changein velocity near km 110.
20,434
2.
FLEMING AND TRI•HU' BURIED RIDGE BENEATH THE OREGONMARGIN
Data
fill sediments,as is typical of subduction zones. Between the
"trench"and the Coast Range gravity highs, the data reflecta
complexinterplay between seafloor topography, changing
depthto the Juande Fuca plate, the internal density structure
of the accretionaryprism,the edge of crystalline continental
basement,and the distribution of a number of small, shallow
The prominentgravity highs beneaththe Oregon Coast
Range(Figure2a) are interpretedto be uplifted Siletz River
Volcanicsexposedin the OregonCoastRange[Bromeryand
Shayely,1964;Snavelyet al., 1980; Snavely,1987]. On a regional scale,theseanomaliesexpressthe excessmassof the basins filled with unconsolidated sediments.
Siletz terranewithin the forearc[Wells et al., 1998], consisThe magneticdata have been used to qualitatively locate
tent with its anomalousthickness measuredseismically the seawardedge of crystalline basementoffshorewhere it is
[Trdhuet al., 1994]. The edgesof the anomaly,however,are juxtaposedagainst the accretionarycomplex [Bromery and
affectedby theburialof the Siletzterranebeneathsedimentary Snavely, 1964;Snavelyet al., 1980; Snavely, 1987]. This derocksof continentalshelfbasinsand the WillametteValley, terminationis generallyin agreementwith that determinedon
and theseedgesdo not precisely map the boundariesof the a seismiccrosssection[Trdhuet al., 1994, 1995]. Changesin
terrane.Westof 125ø, the gravitydatareflectthe eastwarddip the gradientat the westernedgeof this anomalysuggestvariaof the subductingJuande Fucaplate and low-densitytrench tion in the shapeof this boundary, assuminga simple magA.
-126 ø
46 ø
B.
_125ø
_124ø
-123 ø
_126ø
-125 ø
_124ø
_123ø
46 ø
45 ø
45 ø
44 ø
44 ø
43 ø
43 ø
42 ø
-126 ø
<-120-100
-125 ø
-80
-60
_124 ø
-40
mGal
-20
0
_123 ø
20 >40
_126 ø
<-155-140-110
-125 ø
-80
-50
_124 ø
-20
10
40
42 ø
_123 ø
70 100 >130
nTesla
Figure2. (a) Gravityanomalies
of thecentralOregon
continental
marginfromthe Geophysics
of NorthAmerica, CD-ROMversion1.0 [1987]. Bouguergravityis shownonshore;
free-airgravityis shownoffshore.Corrections
andgridding
algorithms
usedto construct
thisdatasetaredescribed
by Godson
andScheibe
[1982].
Twoeast-west
trending
lineslabeled
N andS onbothmapsshowthe locationsof the two shipboard
profiles
thatweremodeled
forthispaper.Thecoastline,
deformation
front,andBlancofracture
zone(BFZ)areshownas
blacklines.TheFulmarfaultofSnavely
etal. [1980]is shownasa dashed
line. Thinblacklinesaretopographiccontours
indicatingdepthsof 200, 1000,and2000 m offshoreand elevationsof 200, 600 and 1000 m
onshore.HB, HecetaBank;NB, NehalemBank. (b) Magneticanomalies
of the centralOregoncontinental
margin
fromtheGeophysics
ofNorthAmerica
CD-ROM.Theoriginal
source
forthedatafromtheOregoncontinentalmarginis a series
of aeromagnetic
surveys
flownat an elevationof about500 ft groundclearance
and
1.6-8kmmilelinespacing
[U.S.Geol.Surv.,1970].Datawereregridded
to correspond
to thegravitygridas
described
by Godson[1987]. Overlaysarethe sameasthosefor Figure2a.
FLEMINGAND TRI•HU: BURIEDRIDGEBENEATHTHEOREGONMARGIN
110
115
120
I
I
I
_•_,_,.•ZS..'--:y-.,_;z•-
125
130
135 km
I
I
-7-•_.:
•:-__•_.--:_•
:...•.,•;,,••.••--,:•
.--..*.-_•.•i•'":.•-..
- •.•_..•.•r.-..,_•;__.•.•.'-:•_..--"'"*•
.._:•:--,--_•,.-...
- _•.•••:••.•
:•._ • •
•....
,..,,.::-•--_.-:,..-•'....,•
;•'•.
";z-•5.•.•'•_•_•_
:.'.•ff• ...... '...............................
.::,,----.2
......
-
20,435
..............
..•.:
.•----•.,•..•
,,.,,•
•
'
;;9•
;.,'.
,• .;,'t.:'*',•:•----..
"•'-•,?',a•lk\\..
•.":N..•..
'.-.'_,42
-'-k.•-:--.,-'-.,•:
..;
.•..,::;•,••.•.-"
',n-,',.;..'.
!.:..'.1
;.'?.:.•..?.•
•',2":,,•..',z,,•.•l•.',•..',;'•
:..•,x,,:
,...----':•-'
;;:..'
.... • :,z;.•..'...-•.•:•.--?-...v:-,,..-.
......
i½,. ',.' '-:.'
..':.2_'•?.:,-,-,• '.½ ' 2."'"".-", ",.• ' '-'...,"•".-.'--;-"-'-",'.,
"' ' • ,,.•-"•
-•"' '//" ';•.?'.";:.'.•':,'.•,',•.2•',
•.'..s•.,...?,•"•:...',..'•,.,:.".'*.•
.....'.',, ,,.•.,,
,,a•.,.-,•*_--•
- ß ß.,,...½,;-,,2•.'-'~..
":"'
•; ':':'..,
".,'•;.'
""',7"•'
•'•,
,••'•, .,$•
':?.-,'.3,'.;'•'.
:?:
'-•?__.
•"-".•',
r',, :'•.','q•
.,..x..2:.
:=.,.•;,.'.•
:,:..'.,-,_•.7,:•,;'•._•
:'•?•,','-;i•
•:;-:•;:•,;'•'."",:'r-'
,,,.,,:..
•2'.,•
,..,.:;.
',
,•:;',.,;•._.z...-'•:,•'.i.½.'..•.:•;;',•..;.',.'-..c•.•:•x
.•N
.v..,?.",•,';',z.-".,.;.:
,7•'.,
:.•'-•..,:9:..'2,
;...•.,:•.:..,
:,::,•
,'t..'...,"•2•r,:-.'.•,
•.,,4:;;;,':'
".?,.',-,•_.•,?,
•v,..'' •,,•..:';'•-•.,.,t,.{.,...,;,•,':
•,•.
'.;:•.•%.
".*,,;.,:,.--.:;,,'•,'_,'..'-',•'½t,.'-.,;.-:.•,-.'.o,:
,;':.•&..-'tS;?7,?"•'_.--2d...,.:",,.
well
Figure 3. Migrated,depth-converted
detailof the seismicprofile of Tr•hu et al. [1995] acrossthe centralOregon continentalmargin. The reflectionidentifiedas the top of Siletz is shown(labeledbasement),as is the
westernextensionof this terraneinto a regionof discontinuous
diffractionsthat is requiredto modelthe magneticanomalydata. S marksthe seawardedge of the Siletz terraneas identified on the seismicdata; M is the
seawardedgebasedon magneticmodeling. M/P is the Miocene/Plioceneunconformity.The position of an
industrytest well, in which two Miocenesills (labeledDB, and CF) were encountered,is also marked. The
short-wavelength
folds and possiblerecentfault overlyingthe regionbetweenkm 115 and 120 are characteristicof a narrow bandof foldsandfaultsthat overliethe seawardedgeof the Siletzterrane.
netization
contrast.Onshore,
themagnetic
dataare quitevari- 3. Modeling' North Line
ableandgenerallyreflectoutcroppatternsof the Siletz volWe used gravity and magneticsmodeling software procanicsandof localizedandvariablymagnetized
upperEocene
duced by Northwest GeophysicalAssociates(Corvallis, OreandMiocene
basalts
[e.g.,Bromery
andSnavely,1964;SimpsonandCox,1977]. Nearthe deformation
frontthe magnetic gon), which is based on the algorithms of Talwani et al.
lineations in this region have been identified as seafloor [1959] and Talwani and Heirtzler [1964] and incorporates
the algorithmof lYebring [1985] for least squaresinversion of
spreading
anomalies
4 and 4a [lyilson, 1993]. Between125ø
selectedmodel parameters.Becausethe large-scalegeometry
andthe edgeof the Siletz terrane,the magneticdata show a
of the Oregon continental margin consistently strikes N-S,
"quiet zone."
perpendicular to the survey lines, and because small-scale,
Although the griddedpotential-fielddata define general
three-dimensionalvariation is poorly constrained,the models
patterns,theyhavebeensmoothed
considerably.We therefore
examinedshipboard data from43ø00'N to 46ø00'N that were
acquiredindependently
by a numberof agenciesand universi-
discussedhere are two-dimensional. We did, however, calculate the effect of the northern
and southern
termination
of He-
ceta Bank on modeled gravity anomaliesand determinedthat
three-dimensionaleffectswere not significant. Becauseof multiple trade-offsamong various parameters,multiple starting
models were considered, and a hybrid forward-inverse approach was adopted in which only a single parameter was
ties [Marine Trackline GeophysicsCD-ROM, version 3.2,
1996]. Data are generally consistentfrom cruise to cruise.
Shipboarddataacquiredby the NationalOceanographic
and
AtmosphericAdministration(NOAA) were selectedfor the
modelingprocedurebecausemagnetic,free-airgravity, and varied at a time.
bathymetrydata were collectedsimultaneously,
the ship
trackswereorientedapproximately
perpendicular
to the pre3.1. Initial Model Parameterization
dominantN-S geologicalstrike,and the surveylines were of
thegreatest
spatialdensity.The datawereresampled
using an
The generalgeometryfor the Siletz terrane, Juan de Fuca
Akima splinewith a 0.5-km samplinginterval prior to model- plate, shelfsediments,and accretionaryprism in the seismic
ing.
model (Figure l d) was used in our initial potential-field
The first profile modeledis approximatelycoincidentwith model. Two important features of this model, the basement
the seismicmodelof Trdhu et al. [1994]. For this profile we the westernedge of the Siletz terrane,were not included in the
focuson modelingthe magneticdataand showthat the grav- initial model, because we wanted to test the existence of these
ity data are compatiblewith this model. A secondprofile features independently.
crossing
HecetaBankwaschosenfor modelingbecauseit repDetailedstructureof the top of the Siletz terranewas interresentsa distinct contrast in margin morphology and the preted from USGS multichannel seismic reflection line
slopeof themagneticanomalymarkingthe seawardedgeof the L376W005 [Snavely, 1987], a proprietary industry seismic
Siletz terrane. For this profile we show that the magnetic reflection line, and the seismic reflection line of Trihu et al.
model is consistentwith conclusionsobtainedfrommodeling [1995] and convertedto depthusinga velocity of 1.8 km/s for
line 1, and we focuson modelingthe gravity data. Models for sedimentsabovethe Pliocene-Mioceneunconformityand 3.0
thesetwo profilesshouldbracketthe rangeof structurealong km/s for sedimentsbelow it. Interpretationsof well log data
this centralsegmentof the Cascadiamargin.
[Petersonet al., 1984] and a syntheticseismogram
computed
20,436
FLEMING AND TRI•HU: BURIED RIDGE BENEATHTHE OREGONMARGIN
from the soniclogs [Cranswickand Piper, 1991] of an industry well generallyconfirm our interpretation of the seismicreflection data and our velocity estimates. The top surfaceof the
Siletz terraneis imaged offshorein these data sets as a blockfaultedreflectorto model km -•115 ("S" in Figure 3). West of
this location, several discontinuous diffractions are observed,
but basementcannot be unambiguouslyidentified.
Two strongreflections,locatedat-•1.25 and 2.4 s in the vicinity of the well, are observedin the data(DB and CF in Figure 3). Drill hole data suggest that these reflections result
from Miocene basaltsassociatedwith the onshore Depoe Bay
and Cape Foulweather basalts [Cranswick and Piper, 1993'
Peterson et al., 1984; Snavely et al., 1980]. The extent of
ity versusdensitydata(from the RAYINVR seismicmodeling
programwritten by C. Zelt and discussedby Zelt and Smith
[1992], extendedto include additional data for crystalline
rocks[e.g.,Godfreyet al., 1997]. This equationwas applied
to the griddedvelocitymodel [Trdhuet al., 1994], which containsboth lateral and vertical velocity gradients. The resulting gridded density model was then simplified to obtain an
initial model parameterizedby isodensityblocks.
3.2. Magnetic Model
AssumingK and M R of 0.003 and +-0.001 for the oceanic
crustof the subductedJuande Fuca plate resultsin a predicted
long-wavelength magnetic high in the magnetic quiet zone
these basalts to the north and south is not well constrained,
(curve
IN on Figure 4a). Increasing K and M R for anomaly 4
but the seismicreflection data suggestthat they terminate at
to
0.0035
and 0.0015 and then gradually decreasingK and
least 10-15 km east of the Siletz terrane'swestern edge, and
M R for anomaly5 to 0.0008 and 0.0001 resultsin a muchbetdrill hole data show that the units are thin [Snavely et al.,
1980; Snavely, 1987; Peterson et al., 1984]. These Miocene ter fit to the observedmagneticanomalieswest of the Siletz
basalts were not explicitly included in the model, because terraneanomaly(curve DEMAG on Figure 4a). We do not attheir impacton the overall anomalyis smalland their magnetic tempt to interpret the variability in magneticproperties for
anomaly4 as it is within the uncertainty of our knowledge of
parametersare poorly constrained.
Startingvaluesfor magneticparameters
of the Juande Fuca the magneticproperties of oceanic crust. The required decreasein magnetizationfor anomaly5, however, is significant.
plate were a remanentmagnetizationintensity MR of +0.001
Thermal
models [Scheidegger, 1984; Hyndman and Wang,
emu/cm
3 (in general
agreement
with Finn [1990]andButler
1992] and a susceptibility K of 0.003 cgs units (in general 1993] suggestthat the subducting slab is not hot enough to
agreementwith Telford et al., 1990). (Note that we cite mag- attribute all of the apparent demagnetizationto thermal denetization propertiesin cgs units to be consistentwith most magnetization, and some chemical demagnetization due to
is thereforealso likely, as discussedin section
prior literature. To convertto SI, multiply susceptibilityval- metamorphism
uesby4It andmultiply
remanent
magnetization
valuesby 103
to convertemu/cm
3 to A/m). No directmeasurements
of MR
5.1.
The next stepin the modeling procedurewas to add the effect
of the Siletz terrane. Becausemagneticmodeling is most
andK are availablefor Juande Fuca plate basalts offshoreOresensitive
to the shorter-wavelength response of shallow
gon, but these initial model values give a reasonableKoestructure
[Webring,
1985], we were able to determinethrough
nigsbergerratio for oceanic basalt of*-I and required little
subsequent
modificationto provide a reasonablematchto the trial and error that the upper surfaceof the Siletz terranemust
magneticdata. InclinationI and declinationD of the remanent extendslightly seawardand upward of the westernmostposition where it is clearly imagedin seismicdata ("M" in Figure
field were set equalto that of the presentgeomagnetic
field (I 68ø andD - 19ø) on the assumption
that no major crustalrota- 3) to include a region of diffractions indicative of rough totions of the young Juan de Fuca plate have occurredin the pography. This seawardextension is only weakly sensitive
study area[Finn, 1990]. This assumptionis justified by the to K. However,the modeledgeometryof the edge of the Siletz
observation that seafloor spreading magneticlineations off- terraneand its modeled susceptibility are theoretically correshorecentralOregon are parallel to the currentorientation of lated [Webring, 1985], and the sensitivity of modeledgeomethe spreadingcenter. Using Wilson's [1993] identificationof try to modeled susceptibilitymust be evaluated when considanomalies4 (-•7.4 Ma) and 4a (-•9.1 Ma), we calculateda half- ering the geometryof the entirebackstop.
Assuming that the western side of the Siletz terrane is a
spreadingrate of 3.65 cm/yr and projectedthe magneticlineastraight
line in crosssection,we created64 models,representtionsof the subductedoceanicplate downdip using the geoing eight values of susceptibility and eight values of dip,
magnetictime-scaleof Candeand Kent [1992].
The internal stratigraphy of the offshoreportion of the holding all other model parametersconstant. The root-meanSiletz terrane is unknown; therefore this block consists of an
squareerror for each model as a function of dip and susceptiunconstrained
vertical and horizontal distribution
of normal
bility is shownin Figure 5. The best fit dip and susceptibiland reversemagnetizationpolarities and variable M R and K. ity for this simplemodel are 55øW and 0.00285 cgs units, reHowever, on the basis of the observedmagneticsignature of spectively. Fit rapidly deterioratesas the dip becomesvertithe Siletz terrane,paleomagneticdata, and other information cal or landward(>90ø). Curve ST in Figure 4a showsthe effect
[Fleming, 1996], we were able to maketwo simplifying as- of addingthe Siletz terraneto the model with theseparameters.
sumptions:(1) Bulk MR for the Siletz terranemay be consid- This value for K is consistent with values measured by
ered negligible and (2) variations in K within the Siletz ter- Brornery and Snavely [1964] for samplesof Siletz rocks from
rane are distributed on a much smaller length scale than the westernOregon. Implicationsof this dip for modelsof the hiscrustal scale of the terrane itself and of our crustal-scale model.
tory of the Siletz terraneare discussedin section5.2.
We assumedthat the western edge of the Siletz terrane is a
As magneticpropertiescommonlyvary by ordersof magnitude
within a single outcrop, this second simplifying assumption straight line in cross section. However, if one relaxesthis
condition,many geometriesare allowablefor the seawardedge
is implicitly made in any magneticmodeling effort [e.g., Finn,
1990], and allows us to assign a single K value for large of the Siletz terrane below 7-km depth, including a slight
crustal blocks.
landwarddip. Figure 6 illustrates the sensitivity of the magDensitiesfor our initial model were determinedby calculat- netic model to the thickness of the Siletz terrane and shows
ing a best fit polynomial to the well-known Nafe-Drake velocthat a landwarddip at depth shallower than 7-km depth leads
FLEMINGAND TRI•HU: BURIEDRIDGEBENEATHTHE OREGONMARGIN
mo
450 -
'.•- Observed,
350
250
,•?•...."''"'•:•
150 •.'-
'I•.
50
•••
20,437
-- = Calculated
• ......................
,..
,v' '\•
..-'
'"'•..... •'
-50
Magnetics
-150
go
sT
40
20
Free-air
Gravity
0
-20
-40
-60
-80
Go
................ .:.,.
..........
5 •• fan4
+{
, 0.0035/•
10
I L2:.
øø•5•
'
accretionaryR'
o•
• 0.003/• • •
• •
20 •vtagnencs
Do
/
............................
:•'
.....................
•-r•-,•;•-Zr
..........
•.,,•,,,,
.,•complex
,/'"..•,3
t' Siletz
Te•ane
......
•
•+• _ .•.•,•....
•7•;•
/
•--JCS} ./
o/oONDEMAG)
-•...•
.......
.•,,
•
"-.•,..
1.00
2.54
2.65
2.95
2.65
2.95
10
3.30
3.10
15
Gravity
20
0
V.E.- 2
20
40
60
80
100
120
140
160
Distance(km)
Figure 4. Summaryof magneticand gravity anomalymodeling for the north line. Although the gravity and
magneticanomalieswere modeledtogether using modelsspecifiedby the sameboundaries, as is generally
donewith crustalpotentailfield modeling,the results are shown in two separateplots in order to highlight
thosepartsof the modelmostsensitiveto the differentdatasets. (a) Observedand calculatedmagneticanomalies. Stepsin the magneticmodelingare describedin the text. (b) Observedand calculatedgravity anomalies.
(c) Magneticmodel. Exceptfor the seafloor,which is included for reference,only boundariesof bodies with
nonzeromagneticparametersare shown. Magnetic parameters(susceptiblity and remanentmagnetization)of
the subductedJuande Fuca plate are shown; "anX"indicates blocks correspondingto identified seafloor
spreadinganomalies;
+ and- referto the signof the remanentmagnetization.Magnetic parametersof the Siletz
terrane(ST) andthe buriedridge(R andR') are alsoshown.The strategyfor developing
the model,which proceededin stepsfrom initial (IN), to demagnetization
of subductedoceaniccrust (DEMAG), to buried ridge of
oceaniccrust(R), to alternateburiedridgemodel(R'), is discussed
in the text. Magnetizationpropertiesare in
cgsunitsto be consistent
with mostprior literature. To convertto SI, multiply susceptibilityvalues by 4•r
andmultiplyremanent
magnetization
valuesby 103to convert
emu/cm
3 to A/m. (d) Gravitymodel.Only
boundariesseparatingbodiesof different density are shown. Densities were determinedfi'omthe seismicve-
locitymodelof Trdhuet al. [1994]asdescribed
in thetext. Densities
arein unitsof 10'3kg/m3.
20,438
FLEMINGAND TRI•HU: BURIEDRIDGEBENEATHTHE OREGONMARGIN
-43ø00'N) are several circular positive and negative anomalies that may be buried seamountsoffshoreCape Blanco (T.
Brocher,personalcommunication,1996). The velocity model
shows a basementhigh with velocities appropriatefor upper
- 0.0040
oceaniccrustin this region (Figure ld).
The linear anomalycannot be modeled as a seafloormag- 0.0035
netic lineament,becauseits wavelength is too short to be
producedby a source at the depth of the subducting crust.
The data can, however, be matchedby introducing a narrow
- 0.0030
ridgeof materialwith the samemagneticpropertiesthat we assumedin our initial model for positively polarized oceanic
0.0025
crust (K=0.003 and MR=0.001).
The best fitting crosssectionalshapeof a body with these magneticparameters,as- 0.0020
sumingthat it extendsto the subductedplate, is shown in
Figure4c (labeledR), andthe predictedanomalyis shown by
- O.O015
line R in Figure 4a. Assumingthat the magneticparameters
that were derivedfor the Siletz terrane(K=0.00285 and MR=O)
results in a larger,shallower body (labeled R' in Figure 4c)
0.0010
with no perceptiblechangein the predicted magneticanom30
45
60
75
90
105
aly. If the magneticbodyis smallerand deeper,susceptibility
rapidly increasesto unrealistic values. It can, however, be
Figure5. Root-mean-square
misfitof theanomaly
predicted somewhatshallowerif it is not in contactwith the subducting
bythemagnetic
modelasa function
ofsusceptibility
anddip oceanicplate. Severalpossible interpretationsof this feature
Dip (degreesfrom west)
of theSiletzterrane.A dipof 90ø is vertical;seawarddips are
<90ø;andlandward
dipsare>90ø. Thebestfittingsusceptibilityanddipare0.0025and60, respectively.
are discussed in section 5.3.
to a calculated field that does not fit the observed magnetic
data, regardlessof the susceptibilityassumedfor the Siletz ter- •
rane. The seismic data also require a thick Siletz terrane that "•
extends nearly to the subducting oceanic crust, as shown in
Figure 7, which showsray path coverageand travel time fit for •
arrivals refractedthrough the Siletz terrane and underlying
oceanic
crust
and for arrivals
reflected
shelf.
data indicate that these basalts do not extend farther east than
model km 130 (Figure 3). We did not include these sills,
which require a fully three-dimensional modeling approach,
lack of influence
on models of the Siletz
/ Magnetics
150
if'
5o
...!•,•
....
-50
-150
1
•
3
•
7
•
The fit of our calculated magneticanomaly to the observed
data is poor aroundmodel km 140 (Figure 4a). This is probably due to the Miocene sills mentioned above. The reflection
because of their
250
fi'om the base of the
subducted oceanic crust for the model of Trdhu et al. [1994].
These arrivals were recordedon stations deployed onshore to
record offshoreshots, which provided the best experimental
geometryto image the structurebeneaththe coastlineand continental
350
•
9
11
13
15
80
100
120
140
160
Distance
ter-
rane edge. They could account for the high-amplitude, relatively short-wavelength mismatchhere if they have a significant and predominantlyreversedremanentmagnetization.The
limited regional extent of this magnetic low (Figure 2b) supports this interpretation.
The final featureof the magneticdatathatwe modeledis the
local high at km 80-100. This anomaly appearsto be part of a
low-amplitude, linear, margin-parallel anomaly that begins
Figure 6. Exampleof magneticmodelsshowing the sensitivity of the modelto variations in the thicknessand dip of the
seawardpart of the Siletz terrane. Lines 1-3 showthe effects of
varying the configurationof the Siletz terranewhile keeping
the susceptibility fixed at 0.00285. The dashed lines correspond to model 3 with susceptibility varied by +- 0.0005.
Magneticparametersof the subductedJuande Fuca plate crust
and buried ridge are the sameas those in the final model of
Figure4. Magneticparametersof the accretionarycomplex,innear 45ø00'N
and continues to the south to about 43ø00'N.
This magneticanomalywas not recognizedprior to our study. cluding that portion underlying the Siletz terrane, are 0. A
It is a subtlefeaturethat is not visible in maps scaled to span landwarddip of the boundary of the base of the Siletz terrane
(model 1) movesthe main peak of the anomalyto the east. Bethe rangeof magneticanomalyvalues characteristicof seafloor
causethe observedpeak is fixed, such a model requires that
spreadingand the Siletz terrane. However, it is apparent in the Siletz terrane extend fartherwest than is allowed by the
shiptrack magneticdatanorthof HecetaBank. The expression seismicdata. Assuming the selectedmagnetic properties are
of this featurecan alsobe enhanced,as shown in Figure 8, by representative,
the Siletz terraneprobably extendsto a depth
applyinga shaded-relieffilter and by selecting a finer contour of at least 7-9 km, consistent with the seismic and gravity
data.
interval. In line with this feature to the south (-•42ø00'N to
FLEMINGAND TRI•HU: BURIEDRIDGEBENEATHTHE OREGONMARGIN
7O
126
98
154
20,439
182
o
•
5
Rs
B.
o
•
2•
5
10
c•15
c
ß
D.
8
pmpu••onshore
stations
Poc •'=_•...-,
.•" Ill•....
"•"•-•,
coast
anflfnmm
....
• '•=,•__::..?•....
,.........
:.----,
,•
Ps
I
70
I
I
98
I
I
I
126
I
154
I
I
182
Distance(km)
Figure 7. (a-c) Ray diagramsshowingcoverageof seawardedge of Siletz and underlying Juande Fuca plate
usingrefractedarrivalsturningwithin the Siletz terrane(Ps), refractedarrivals turning within the subducted
Juande Fucacrust(Poc), and arrivalsreflectedfromthe base of the subductedJuande Fuca crust (PmP). (d)
The calculatedand observedtravel times for all onshore-offshore
recordings.The model for the calculations is
the sameasthat shown in Figure ld and discussedby Trdhu et al. [1994, 1995]. A small advancefromthe
buried ridge at km 95 is observednear km 84 in phaseslabeled Ps and Poc. The primary evidence for this
phase,however, is in the oceanbottom seismometer(OBS) data, which are not discussedin this paper but
which show a localizedregion of oceaniccrustalvelocitiesat relatively shallow depth.
3.3. Gravity Model
For the gravity model, subdivision of the accretionary
prismand Siletz terraneinto severalbodies(generallyincreasing in densityarcwardand downward)represents,for the most
part, the discretization of continuous density gradients implied by the velocity model. Someminor modification of densities fromthe initial values derived fromthe velocity model
was requiredto modelthe gravity data, reflectinguncertainty
in the velocity model and in the conversionfromseismicvelocity to density. The densitiesin our final gravity modelremain consistentwith those predicted fromthe velocity model
via the Nafe-Drake relationship (Figure 9), showing that the
velocity modelof Trdhu et al. [1994] is generally consistent
with the gravity data. Althoughwe do not showthe sensitiv-
causethese previous models were unconstrained by seismic
data, and overestimatesof the thickness of low-density sedimentsresulted in an underestimateof the depth to the subductingplate beneaththe continentalshelf.
4. Modeling'
4.1.
Initial
Model
South Line
Parameterization
Our startingmodelfor the profile crossingHecetaBank was
adaptedfrom the crosssectionof Snavelyet al. [1985], which
is basedon seismicreflectionlines,regionalgravity and magneticsdata,an industrywell, and onshoresurficialgeology.
This was used to define the shallow basin structure and the
distribution of the upper Eocene-ageYachats basalt, which
ity of the model to the thickness of the Siletz terrane and buried ridge since the thicknessof the Siletz terranebeneaththe
representsbasementin the seismic reflection section. This ba-
shelfhasalreadybeendemonstrated
by the magneticand seismic modeling,we note that a significantdecreasein the thickness or density of the Siletz terraneleads to a poor fit to the
localizedhigh where it outcropsat the coast at -44ø15'N; adjacent, strong,areally limited lows immediatelyto the north
probablyrepresentthe samebasaltoutcropbut reverselymagnetized[Simpsonand Cox, 1977]. The Snavelyet al. estimate
for the geometryof the top of the Siletz terranewas also used
in the model. The density structurewithin the accretionary
prism and the configuration of the subducting Juan de Fuca
data.
The configurationof the subductingplate in this model is
significantlydifferentfrom that in previous crustal modelsderived fromgravity [e.g., Couch and Riddihough, 1989], be-
salt is manifested
in the regionalmagneticsmap as a strong,
20,440
FLEMINGAND TRI•HU: BURIEDRIDGEBENEATHTHE OREGONMARGIN
B.
Ao
_125 ø
46" '
_124 ø
46"
C.
-125 ø
-124 ø
-125 o
,:...-.,•:•...:::.:,. 46"
ß
_124 ø
46"
:
45"
44"
44"
43"
43"
.•?i'iiL.....
:!•?•
':'"•i:;•
42
ø
-125 ø
<-70-60-50-40-30-20-10
-124 ø
0
10 20 30
-125 ø
-124 ø
40 50 60 >70
nTesla
Figure 8. Magnetic anomaliesin the region of the seawardedge of the Siletz terrane. Data are the sameas
thosein Figure2b, but the dataare plottedwith a palettedesignedto show low-amplitudeanomalies. (a) Illuminationfrom the west.(b) No illumination. Contoursat an interval of 60 nT are shown. (c) Illumination
from the east. The peakof the anomalymarkingthe westernedgeof the Siletzterraneis shownas a dashedline
on Figures8a-8c,andthe anomalydueto the buriedmaficridgeis indicatedby whitearrowsin Figures8a and
8c.
plate were taken from our model of the northern line. The regional magneticanomalymap (Figure 2b) showsthat the magnetic lineamentsof the Juan de Fuca plate are uninterrupted
between the two profiles in this study, supporting the assumptionthat the subductingplate is continuousbetween the
two profiles.
netization of the subducted crust determined for the north line
and a seawarddip of 57ø for the Siletz terrane,as determined
for the north line.
In step2 (labeledCD, for "changedip," on Figure 10a), the
susceptibilityfor the positively magnetizedblocks responsible for'anomalies 4 and 4a is decreased to 0.002.
As for the
Initial estimatesof K and MR for the Juan de Fuca plate
crust and the Siletz terrane were taken fromthe modeling resultsof the north line. Siletz bulk M R was assumedto be 0 for
north line, we do not attemptto interpretthis minor changein
the magneticpropertiesof the subducting oceanic crust. In
addition,the dip of the westernedge of the Siletz terranewas
decreasedto 30øW, and K is increasedto 0.0033. However, it
is not clear whether these changesare significant given the
uncertainty in the depth and susceptibility of the Siletz ter-
reasons discussed in section 3.2, and K was assumed to be
rane.
4.2. Magnetic Model
A small-amplitude,medium-wavelengthmagneticanomaly
0.00285 cgs units. For the Yachatsbasalt,K and M R were set
to 0.0026 and 0.001 [Bromery and Snavely, 1964], and decli- is clearly superimposedon the main Siletz anomalyat model
nation and inclination were set to 46 ø and 64ø, respectively km 102-112 (Figure 10a). This anomalymay be part of the
[Simpsonand Cox, 1977]. The responseof this initial model buried "ridge" discussedabove for the northernprofile, or it
is labeled IN in Figure 10a. It includes the eastward demag- may reflectirregularity of the seawardedge of the Siletz ter-
FLEMING AND TRI2HU: BURIED RIDGE BENEATH THE OREGON MARGIN
Density
(x10-3kg/m
3)
I
I
-7
-6
-5
-4
-3
-2
I
1.5
2.0
I
2.5
20,441
Figure 10b and shadedregion in Figure 10d). Low densities
in this region are consistentwith the interpretationof Goldfinger et al. [1999] that this regionis a largeslump.
Thereis a significanttrade-offbetweenthe density and the
shapeof the high-densitycore beneathHecetaBank. Assum-
ing a densityof 2.7 x 10'3 kg/m3 resultsin loweringthe
boundary between the upper and lower accretionarycomplex
by-•2 km. A body with this density would most likely be
madeof basalt and would probably representa subducted
seamount. Although we cannot rule out a concealed
seamount,with magneticpropertiessuch that the effectsof inducedand remanentmagnetizationcanceleach other such that
no magneticanomalyis produced,we considerthis interpretation highly unlikely. Our gravity model thereforesuggests
that high-densitymiddle Miocene and older sedimentsof the
deepaccretionarycomplexhave been uplifted to formthe core
of HecetaBank, as originally suggestedby Kulm and Fowler
[1974], andplaceslimits on the volumeof materialuplifted.
I
3.0
5.
Discussion
5.1. Demagnetization of Juan de Fuca Plate Crust
Figure 9. Relationshipbetweendensity in the final modelof
As shown in section 4.3, modeling the seafloorspreading
Figure4d and velocity in Figure ld. The rangeof velocities anomaliesover the Juande Fuca plate suggeststhat the subfor eachdensity results fromthe differentmodel parameteriza- ducted crust is significantly demagnetized. Although the
tions: The velocity model is dominatedby gradients whereas
the density model is constructedfrom a small number of
isodensitybodies. The shadedregion showsthe rangeof experimentaldata. The final velocity-density relationship is
within the allowable range of densities.
rane. These two possibilities are discussed in section 5.3.
AssumingK, MR, D, andI of 0.003, +-0.001, 19ø, and 69ø, respectively,and findingthe crosssectionof a linear body that
fits themagneticdataresultsin the featurelabeledR in Figure
10a and predictedanomalyR in Figure 10a. Assumingmag-
Cascadiasubduction zone is characterizedby high heat flow
dueto the youngageof the Juande Fucaplateand the insulat-
ing effectsof thicksedimentcover,thermalmodelsof Cascadia
[Hyndman and Wang, 1993; Oleskevich,1994] suggestthat
Curie temperaturesof 550ø-600øC are not reached on the
thrust plane until -175-200 km east of the deformationfront.
It is thereforedifficult to attribute the observed demagnetization entirely to thermal demagnetization of the subducted
plate, assuminga reasonablethermal gradient in the oceanic
crust beneath
the thrust zone.
We suggest that hydrothermal alteration of the subducted
crust also contributesto reducing the remanentmagnetization
netic parametersappropriatefor the Siletz terrane (K=0.00285
and MR=O) results in the featurelabeled R' in Figure 10c. of the subductedcrustthroughmetamorphismof the basalt and
While thereis sometrade-off between the magneticparameters recrystallization of iron-titanium oxides [Scheidegger, 1984;
and geometry of this body and while details of the structural Finn, 1990]. Large volumesof fluids releasedat greaterdepth
relationship between the ridge and the Siletz terrane cannot by progrademetamorphismof hydratedbasalt [Peacock, 1987]
may be channeledtrenchward along the shear zone and may
be resolved, if the anomaly is from a buried ridge, then the
contribute to additional hydrothermal alteration and demagridge is in contactwith the Siletz terranein this region.
netization of the subducted crust. Stable oxygen isotopic
4.3. Gravity Model
analysisof carbonatecementsin accretedsedimentsat the central
Oregon deformationfront indicates circulation of highThe primary objective of modeling the gravity anomaly on
temperature
fluids upwardfrom the decollement[Sampleet al.,
this profile was to place constraintson the density structure
1993].
The
petrogeneticgrid compiled by Cloos [1993] sugbeneathHecetaBank. Step 1 (Figure 10b) shows the anomaly
gests that for a young subduction zone, significant metamorpredicted using the density structureobtained for the northphism may take place in the subducting plate at the depths
ern line, with the density of accretionarycomplexmaterial inwhere we requiredemagnetization.The presenceofturbiditecreasinglandward and with depth in three steps. The dashed
line showsthe boundarybetweena densityof 2.28 x 10-3 hosted mesothermalgold deposits within the Alaskan accrekg/m3above
and2.54x 10-3kg/m
3 belowin theinitialmodel; tionary prism also supports the circulation of hot fluids
throughthe forearc [Hauessleret al., 1995].
the densityof the outer part of the accretionary complexis 2.2
x 10-3kg/m3 in the initial model.The anomalypredicted
by
this model indicates that the model density is too high beneath the outer accretionarycomplexand too low beneath Heceta Bank. Lowering the density in the outer accretionary
5.2. Dip of the Western Boundary of Siletz
The geologichistoryof the westernboundary of the Siletz
terraneis enigmatic. Snavely et al. [1980], on the basis of
complex
from2.2to2.0x 10'3kg/m
3 andmovingtheboundary qualitative evaluation of the potential field data and offshore
between
material
witha densityof 2.28 x 10'3 and2.54 x 10'3 drill holes,postulatedthat the western boundary of the Siletz
kg/m3 to neartheseafloor
beneath
HecetaBankimproves
the terraneis an Eocene age strike-slip fault that has had at least
200 km of offset and that juxtaposes accretionary melange
fit of the predictedanomalyto the data (anomaly labeledHB in
20,442
FLEMING AND TR15,HU:BURLEDRIDGE BENEATHTHE OREGONMARGIN
againstthe Siletz terrane. They called this structure the Fulmar fault, and they called the outboard terranethe Fulmar terrane. Onlap by late Eocenesedimentson the continental shelf
near 43øN suggeststhat motion on the Fulmar fault was essentially completeby 43 Ma. The Fulmar fault may thus have
been part of a larger right-lateral fault system that dispersed
terranessuchas the Yakatat terranealong the margin of North
America [Snavelyand Wells, 1996]. The steep seawarddip of
the western edge of the Siletz terrane in our potential-field
models supports the concept that the western edge of the
Ao
375
•._./:.•,,. M
agnetics '"'z2-'••
,.,
....
_
-25
....
"'..............
"""•"•
. ...................
•,,•
.......
CB
..4N
= cu,ated
-225
g.,
Free-air Gravity
"'•
•,
?•'
........
•'. ,•'.•,..
•"'.t,. ßß
g::•f..
........ .... ...
..........
E -40
-80 -
C.
•:•- Observed,
-
alculated
0i
20Do
0
20
V.E.= 2
40
60
80
100
120
140
160
Distance (km)
Figure 10. Summaryof magneticand gravity anomalymodelingfor the south line. Steps in the modeling are
discussedin the text. Gravity and magneticmodeling results are shown in differentfor reasonsdiscussedin
Figure 4. (a) Observedand calculatedmagneticanomalies. (b) Observedand calculatedgravity anomalies. (c)
Magnetic model. Except for the seafloor,which is included for reference,only boundaries of bodies with
nonzeromagneticparametersare shown. The model was developedby modifying an initial model (IN) derived
from the northernprofile. The magneticmodelwas modifiedby changingthe dip of the Siletz terrane(CD), followed by addinga buriedridge of oceaniccrust(R) or Siletz terrane(R'). The gravity model was changedby
increasingthe densitybeneathHeceta Bank (HB) and decreasingthe density to the west. Magnetic parameters (susceptibilityand remanentmagnetization)of the Siletz terraneand of the subductedJuande Fuca plate
are shown ("anX" indicatesblocks correspondingto identified seafloorspreading anomalies;+ and - referto
the sign of the remanentmagnetization). Magnetizationpropertiesare given in cgs units to be consistentwith
most prior literature. To convertto SI, multiply susceptibilityvalues by 4r[ multiply remanentmagnetization
values
by 103to convert
emu/cm
3toA/m. (d) Gravitymodel.Onlyboundaries
separating
bodiesof different
densityare shown. Densities were determinedfromthe seismicvelocity model of Trdhu et al. [1994] as de-
scribed
in thetextandareinunitsof 104kg/m
3. Modification
oftheinitialdensitydistribution
is shownas
the shadedregion, which shows uplift of materialwith densities appropriatefor deeply buried accretionary
complexsediments.Densitiesseawardof this regionare lower than those assumedfor the initial model, supporting the suggestionof Goldfinger et al. [1999] that this material representsa recent slump.
FLEMING AND TRI•HU: BURIED RIDGE BENEATH THE OREGON MARGIN
Siletz terranein this regionwas truncatedby strike-slip faulting. One would expectthat an intact relict subduction
boundarywould dip landward,as has beensuggestedfor the
westernboundaryof the Siletz terranebeneathWashington
[Finn, 1990;Apreaet al., 1998]. The extensionalfaultingrecordedby the Siletz basement
topographyon the north line
(Figure3) mayalsobe a relict of this tectonicepisode,resulting froman extensional
offsetin the Fulmarfault.
5.3. The Buried Mafic Ridge
Severalgeologicinterpretations
are possiblefor the buried
basalticridge that we have identified and modeledon the basis of detailedexamination of the magnetic field. We rule out
an origin as typical abyssalhill topography,becausesuchfeatures are an order of magnitude smaller [Macdonald et al.,
1996].
One possibility (model A) is that the ridge representsa
sliver of Siletz terrane crust that was sheared off from the main
20,443
note, however, that such a ridge could have formedbeneath
the accretionary complex farther west and then have been
transportedto its presentposition, combiningmodelsB and
C.
A variation of this model (model C') is that the ridge was
formedfrom seamounts
that havebeen partially subductedand
then transferred from the Juan de Fuca to the North
American
plate. In this model,the thick Siletz terranemay prevent subductionof seamounts.The ridge may thus representthe "flotsamandjetsom" of subduction, accumulatedover several million years. This model, however, does not explain why the
ridge appearsto be in contact with the Siletz terrane only beneath Heceta Bank, unless the position where seamountsare
shearedoff the lower plate is not directly determined by the
position of the Siletz backstop(see section5.5).
5.4. PossibleImpact on History and Morphology of the
Accretionary Complex
part of the terrane by a right-lateral strike-slip fault that was
oblique to the seawardedge of the Siletz terrane. Similar but
smallersliversof crusthave been mappedalong the boundary
betweenCrescentterraneand accretionarycomplex material on
the Olympic peninsula [Tabor et al., 1972]. It is, however,
difficult to reconstruct a kinematic history that explains why
the ridge shouldbe in contact with the Siletz terrane beneath
Heceta Bank but separatefrom it both to the north and south
of the bank. Assuming that the strike-slip activity responsible for this terrane distribution occurred during the Eocene,
this model also implies that separation between the Siletz
sliver and the main Siletz terrane has been preservedthrough
nearly 50 million years of compressionaltectonics, which we
consider unlikely.
A secondpossibility(model B) is a ridge of upper oceanic
crustthat is forming in place by imbrication and thickening of
the subductedoceaniccrust in responseto compression. Such
Each of these models predicts a differentpattern of uplift
and subsidenceof the overlying accretionarycomplex. Model
A predictsno post-Eocene uplift or subsidencedirectly related to the ridge, although other processesactive along the
margin (e.g., rotation of the ridge relative to the main Siletz
backstop)might have causeduplift or subsidenceduring this
time period. Model B predicts a stationary center of uplift
overlyingthe ridge as it forms but no pronouncedsubsidence.
Models C and C' predict the greatest uplift and subsidence,
with the margin being first uplifted as the ridge passes
through and then subsiding in its wake [yon Huene and
Scholl, 1991]. Timing of uplift and subsidence,however, is
differentfor the two models. Model C predicts uplift and subsidenceonly duringthe past 1.2 m.y.b.p., with uplift and subsidence being nearly simultaneousalong 200 km, whereas
modelC' predictsa longer, complexhistory of localized uplift
features have been inferred
Considerable Miocene and later uplift and subsidence are
recordedin the morphologyand stratigraphy of the continental margin [e.g., Snavely, 1987; Snavely and Wells, 1996;
Yeatset al., 1998] as indicatedby "textbook"examplesof unconformitiesin the marginal basin stratigraphy (Figure 3).
Regional mapping of the Miocene-Pliocene unconformity
[McNeill et al., 1998] shows that the late Miocene outer arc
high continuesnorthof HecetaBank beneathwhat is now the
continental slope to at least 46øN, where it lies beneath the
outeredgeof NehalemBank. NehalemBank is similar in size
to Heceta Bank but is not associatedwith either a magnetic or
a gravity high [Fleming, 1996], suggestinga differentorigin
(Figure2). Becauseportionsof the outerarc high are now beneath the continental slope, McNeil! et al. (1998) proposed
later slope instability or subsidencedue to subduction erosion. This periodof forearcevolutionmay predateformationof
the buried ridge since the late Miocene forearchigh extends
north of where the buried ridge is observed.
More recently, three massivelandslides have been documentedalongthe Cascadiamarginwest and southwest of Heceta Bank by Goldfinger et al. [1999], who conclude that
subductedseamountsmay have triggered the slides by basal
erosionof the accretionarycomplex as they passedbeneath it.
The locations of the slides are shown in Figure 11 and comparedto the current position of the buried ridge and its track
for the last 2 million years, as predicted by the Juan de FucaNorth Americaplate vector [DeMets et al., 1990] assuming
to exist in the relict
subduction
zone of the central California continental margin [Miller and
Howie, 1993]. This model impliesthat the plate boundary has
steppeddown into the crystalline crust of the Juan de Fuca
plate, transferringmaterialfromthe lower plate to the upper
plate. We expectthat the thrustfault at the coreof sucha ridge
formingin oceaniccrust at these depthswould be markedby
seismicity, which is not observed.
A third possibility(model C) is that the ridge representsa
subductedaseismicridge or seamountchain formedby constructionalvolcanism in the ocean basin along a line parallel
to the spreading center. In this model, the ridge has been
draggedobliquely through the accretionarycomplexfor the
past 1.2 million years(Figure 11). Several examplesof subductedridgesor seamountshave been identified in other subduction
zones and have been shown
to have a considerable
effecton the developmentof the accretionarycomplex along
the path of the subductedseamount[e.g., Lallemand and Le
Pichon, 1987; yon Huene and Scholl, 1991; von Huene et al.,
1996]. Becausethe inferredridge is approximatelyparallel to
magnetic lineations, it probably was not formedat a leaky
transformfault. Moreover, becausethe ridge does not appear
to be offsetwhere it intersects the landward projection of a
pseudofaultgeneratedby a propagatingrift [Wilson, 1993], it
could not have formedat a spreading ridge. Few, if any,
analogousseafloorridges(parallel to the spreadingcenterbut
formedoff axis)are observedin the present oceanbasins. We
and subsidence.
20,444
FLEMING AND TRI•HU' BURIED RIDGE BENEATH THE OREGONMARGIN
_125 ø
-124 ø
45 ø 30'
45 ø 30'
45 ø 00'
45 ø 00'
44 ø 30'
44 ø 30'
44 ø 00'
44 ø 00'
43 ø 30'
43 ø 30'
43 ø 00'
43 ø 00'
42 ø 30'
42 ø 30'
_125 ø
-124 ø
Figure 11. Morphology of the Oregon continental margin,shown as a shadedsurfaceilluminated from the
north. Topographicdata frommultiple onshoreand offshoredata sets were compiled and regridded by C.
Goldfinger. Overlain on this surfaceare the location of the seawardedge of the Siletz terrane (heavy dotted
line), the buriedmafic ridge as inferredfrom the magneticanomalies(heavy solid line), the headwalls of major
Neogeneslumps(dashedlines); [from Goldfingeret at., 1999], and the positionof left-lateral strike-slipfaults
that cut the accretionary
complex(white dashedlines) [fromGoldfinger et at., 1997]. The predictedtrack of
the buried ridge throughthe accretionarycomplexassumingthe NUVEL plate motion vector [DeMets et al.,
1990] is shownby the diagonal heavy dashedlines, and the predicted location of the ridge 1 and 2 million
yearsago, assumingmodelC, is shownby the light dottedlines. For model C any seamountscontributing to
the buried ridge must have been east of theselines 1 and 2 million years ago.
model C. A similar track would be followed by any seamounts
subductedwith the Juande Fuca plate within the past 8 milllion years,the time of the mostrecentmajor changesin relative
motionsof the Pacific and Juan de Fuca plates [Wilson, 1993]
and the Pacific and North American plates [Atwater and
Stock, 1998]. However,the buriedridge extendswell north of
themegaslumps.
Inferredagesof the megaslumps,
which are'
ridge or a seriesof seamounts. It may, however, have led to
conditions favorablefor slumping along this part of the margin by inhibiting subduction, leading to thickening and
oversteepeningof the accretionarycomplexand triggering of
slides in responseto earthquakeswith hypocenters farther
east.
The northern boundary of the predicted track of the sub110, 450, and 1120 kyr, increasing from north to south, are ductedridge does correlatewith a major change in slope moralso generally incompatible with subduction of an aseismic phology and deformationfront structure. North of the buried
FLEMING AND TRI•HU: BURIED RIDGE BENEATH THE OREGON MARGIN
ridge, the margin is markedby a seriesof ridges formedby distributed landward vergent thrusting [Flueh et al., 1998].
Where the subducted ridge and seamountsare predicted to
have passed,the slope morphology is more irregular, and the
deformationfront is generally characterizedby a seawardvergent thrust [MacKay, 1995]. Vergencedirection is obscured
by slump debris from43øN and 44øN, but a seawardvergent
basal thrust is observedcutting through older slump debris
between 42øN to 43øN [Goldfinger et al., 1999]. South of
42øN, where the Siletz terrane is absent and the crystalline
backstop of the Klamath terrane forms a landward dipping
backstop [Beaudoin et al., 1996], vergence appears to be
mixed [Gulick et al., 1998]. These changesin vergence at the
deformation front have been previously attributed to differencesin pore pressureon the decollementresulting from differencesin the thicknessand compositionof sedimentson the
subductingplate. While we find this interpretation well justified, the correlation with the massive Siletz backstop and
buried mafic ridge suggests a feedback effect in which the
along-strikevariation in sedimentproperties on the subducting plate is due, in part, to differencesin the uplift and subsidencehistoryof the adjacentaccretionarycomplex.
The buried ridge also correlateswell with the portionof the
margin characterizedby northwest trending strike-slip faults
(Figure 1) [Goldfingeret al., 1997, 1999]. While a few strikeslip faults have been mappedfarther north, only along this
portion of the margin are they well defined and regularly
spaced.None of thesefaults extendeastof the seawardedge of
the Siletz terrane, and they rapidly die out in the abyssal
plain. Thesefaults imply N-S extensionof the sliver of accretionary complex caught between the deformation front and
Siletz terrane.
Although we cannot definitively rule out any of the four
possible explanationsfor the buried maficridge and none of
the models is consistent with all of the available geologic
data, the apparentcorrelationbetween the predicted track of a
subductedridge through the accretionarycomplex,variations
in slope morphology and deformationfront structure,and the
left-lateral strike-slip faults cutting the accretionary complex
lead us to favor model C'. We suggest that subducted
seamountsrafted in on the subducting plate over roughly the
past 8 million years have accumulatedthrough time to form an
apparentburied ridge. The presenceof this ridge may further
inhibit subduction of accretionary complex material, leading
to east-westshortening and north-south extension, uplift of
lower accretionarycomplexmaterial,oversteepening
of slopes,
and massiveslope collapse.
5.5. SubductionZone Asperities
Regardless
of the optimummodel for formationof the ridge,
it likely results in a changein the materialproperties across
the plate boundary[Dmowskaet al., 1996] and thus may represent an asperity, or region of higher frictional resistance,
along the plate boundary. If this is the case,then the lack of
seismicityin this region may be due to plate locking. Geodetic evidencefor plate locking is ambiguous. This region is
a nodein both long-termand short-termmeasurements
of uplift
[Mitchell et al., 1994], leading to the hypothesisthat the plate
boundary is not locked in this region. However, recent
Global Positioning System (GPS) measurementsindicate
horizontalmotions in the Coast Range and Willamette Valley
indicative of locking [McCafJ•ey et al., 1998]. Within the
next few years, ongoing geodetic work should provide new
20,445
informationabout currentplate interactions. The crustal parametersdeterminedby this study and other studies will provide importantconstraintson geodynamicmodelsto explain
the observedgeodeticdata.
In the caseof models B and C, we must further consider the
effect of topographyon the plate interfaceas it must pass over
the top of the buried ridge. Cloos [1992, 1993] has argued
that large seamountswill be too buoyant to be subducted,
locking the subductionzone until the stressincreasesenough
to shearthemoff andtransferthem to the upper plate. Scholtz
and Small [1997] considera more generalcaseof the effect of a
subducting seamount that includes the effects of flexure.
While they conclude that seamountswill not always be
transferredto the upperplate,their modelpredictsthat a 4-kmhigh seamountwill contribute significantly to the normal
stressand that the effect of subductingseamountsis especially
pronouncedin generally decoupled subduction zone. Either
model predictstemporarylocking and eventual stressrelease
in large earthquakes.
OffshoreOregon,theburiedridgemayhavebeensubducted
to a depthat which the frictional resistanceto subduction on
the plate boundaryexceededthe stressrequiredto shearoff the
subductingseamount. The influence of the Siletz backstop on
this process,which dependson the slope and height of the
subductedseamountand on the flexural strength of the upper
and lower plates,is uncertain. If thickeningand dewateringof
the accretionarycomplex alone are adequate to increasethe
normal stress on the plate boundary enough for the plate
boundaryto jump to the base of a seamount,then model C' is
compatible with the observation that the buried ridge is not
in contact with the Siletz terrane along much of its length.
However, if this is the case, the correlation between the buried
ridge and the Siletz terranemay be coincidental. Geodynamic
models of stressesresulting from interaction between the
Siletz backstop and buried ridge are needed to resolve this
question. Regardlessof whether the seaward dipping western
edge of the Siletz is causing seamountsto be shearedoff the
subducting plate, the process of breaking through oceanic
crust on this scale is likely to be seismogenic. Scholtz and
Small (1997) note that large earthquakes seemto have resulted from this process in the normally decoupled TongaKermadec
6.
and Izu-Bonin
arcs.
Conclusions
Magneticand gravity modelingin conjunctionwith existing geologic and geophysical constraints indicates that the
western edge of the Siletz terrane, which acts as the subduction zonebackstop,has a seawarddip of less than 60ø (measuredfrom horizontal). The Siletz terranemay continue to descendat thesedips to the subductingJuande Fuca plate, but
alternative geometriesfor the lowermostportion of the backstop are also consistentwith the potential-field data. Underthrustingof morethan---2km of accretionarycomplexmaterial,
however, is unlikely. This steep, seaward dip supports the
Snavely et al. [1980] model for Eocene strike-slip truncation
of the Siletz terrane, rather than maintenance of an undeformed
relict subduction boundary since accretion of the Siletz terrane. The magneticdata also require progressive eastward demagnetizationof the Juan de Fuca crust in the magnetic quiet
zone, which is most likely due to a combinationof thermal and
chemicaldemagnetization. Our southern transect reveals that
Heceta Bank is cored by relatively dense sediments(•-2.54
g/cm3),
whichlikely represent
uplift,older,denser,sedimen-
20,446
FLEMING AND TRI•HU: BURIED RIDGE BENEATHTHE OREGONMARGIN
tary rocks that were previously deeply buried within the accretionarycomplex.
Perhapsthe most important result of our investigation is
the identification of a mafic ridge buried beneath the accretionary complexfromabout 45øN to 43øN. From correlation
betweenthe locationof the buried ridge, seafloormorphology,
and the recent geologic history of the margin, we conclude
that the ridge representsseamountsrafted in over the past several million years. We propose that subduction of these
seamountsis impeded by the thick, seawarddipping Siletz
backstop,leading to a complicatedfeedbackbetween accretionary complex uplift, subsidence, collapse, and faulting.
Other possibleorigins includeformationin place through imbrication of the subducting crust and formationas a sliver of
Siletz during the Eocene period of strike-slip tectonics that
formedthe seawarddipping backstop. Regardlessof its origin, the ridge likely represents a contrast in the physical
propertiesof the plate boundaryand may thereforebe acting as
an asperity on the megathrust,although the details of this
processare ambiguous.
The seawarddipping backstop formedby the thick Siletz
basaltsthat we model for the central Oregon margin contrasts
with the landward dipping backstop formedby this terrane
beneath Washington [e.g., Finn, 1990; Aprea et al., 1998]
and by the Klamath terrane beneathnorthernCalifornia [Trdhu
et al., 1995; Beaudoinet al., 1996]. Thesedifferentbackstop
geometriesmay be, in part, responsible for the differentpat. terns of deformationobservedin the accretionarycomplexof
Oregon and Washington.
Cloos,M., Lithospheric
buoyancyandcollisionalorogenesis:
Subduction
of oceanic plateaus, continentalmargins, island arcs, spreading
ridges,and seamounts,Geol. Soc.Am. Bull., 105, 715-737, 1993.
Coffin, M.F., andO. Eldholm,Scratchingthe surface:Estimatingdimensionsof largeigneousprovinces,Geology,21, 515-518, 1993.
Couch,R.W., andR.P. Riddihough,The crustalstructureof the western
continentalmarginof North America, in GeophysicalFrameworkof
the ContinentalUnited States, edited by L.C. Pakiser and W.D.
Mooney, Mem. Geol. Soc.Am., 172, 103-128, 1989.
Cranswick,D.J., andK.A. Piper, Geologicframework of the Washington-Oregoncontinentalshelf: Preliminaryfindings,U.S. Geol. Surv.
Circ. 1092, 146-151, 1991.
Davis, D.M., J. Suppe,and F.A. Dahlen, Mechanics of fold-and-thrust
beltsand accretionarywedges, d. Geophys.Res., 88, 1153-1172,
1983.
DeMets, C., R.G. Gordon, D.F. Argus, and S. Stein,Currentplate motions,Geophys.d. Int., 101,423-478, 1990.
Dmowska,R., G. Zheng, and J.R. Rice, Seismicityand deformationat
convergentmarginsdue to heterogeneouscoupling,d. Geophys.
Res., 101, 3015-3029, 1996.
Duncan,R.A., A capturedislandchain in the CoastRangeof Oregon
andWashington,d. Geophys.Res.,87, 10,827-10,837,1982.
Finn, C., Geophysicalconstraintson Washingtonconvergentmargin
structure,d. Geophys.Res.,95, 19,533-19,546, 1990.
Fleming, S.W., Bulldozer blades and colliding submarinemountain
chains: Constraintson central Oregon convergentmargin tectonics
from magneticsand gravity, M.S. thesis,84 pp., Oreg. State Univ.,
Corvallis, 1996.
Flueh,E.R., M.A. Fisher,J. Bialas, J.R. Childs,D. Klaeschen,N. Kukowski, T. Parsons,D.W. Scholl,U. ten Brink, A.M. Tr•hu, and N.
Vidal, New seismicimagesof the Cascadia subductionzone from
cruiseSO108 - Orwell, Tectonophysics,
293, 69-84, 1998.
Geophysics of North America [CD-ROM set, version 1.0],
http://www.ngdc.noaa.gov,
NationalGeophysicalData Center,Boulder, CO, 1987.
Godfrey,N.J., B.C. Beaudoin,S.L. Klemperer, and MTJSE Working
Group,OphioliticBasement
to the Great Valley forearc basin,CaliAcknowledgments.We thank NorthwestGeophysicalAssociates
fornia,from seismicandgravitydata:Implicationsfor crustalgrowth
(Corvallis,
OR) foruseof theirpotential-field
modeling
software.Vern
at the North Americancontinentalmargin,Geol. Soc.Am. Bull., 108,
KulmandBobYeatsprovidedmanyhelpfulcomments
overthe course
1536-1562, 1997.
of thiswork. LisaMcNeill, Tom Hildebrand,
CarolFinn,andRay Wells Godson,R.H., Descriptionof magnetic tape containingconterminous
provided
carefulreviews.ChristofLendlprovided
technical
assistance
U.S. magneticdatain a griddedformat, U.S. Geol. Surv. Open File
withdataprocessing
anddisplay.The location
mapin Figurel a was
Rep., PB87-126017/,YAB,1987.
modifiedfroma figureprepared
by RayWells. Manyof thefiguresin Godson,R.H., and D.M. Scheibe, Descriptionof magnetictape con-
thispaperwere createdusingGenericMappingTools[kVessel
and
tainingconterminous
U.S. gravitydata in griddedformat, U.S. Geol.
Surv. OpenFile Rep. PB82-254798, 1982.
(USGS),Department
of theInterior,underUSGSaward1434-95-G- Goldfinger,C., L.D. Kulm, R.S. Yeats, B. Appelgate,M.E. MacKay,
2617to Oregon
StateUniversity.The viewsandconclusions
contained
and G.F. Moore, Transversestructuraltrendsalong the Oregon conin this documentare thoseof the authorsand shouldnot be interpreted
vergentmargin: Implicationsfor Cascadiaearthquakepotentialand
asnecessarily
representing
theofficialpolicies,eitherexpressed
or imcrustalrotations,Geology,20, 141-144, 1992.
plied, of the U.S. Government.
Goldfinger, C., L.D. Kulm, R.S. Yeats, L. McNeill, and C. Hummon,
Oblique strike-slipfaulting of the central Cascadia submarine
forearc,d. Geophys.Res., 102, 8217-8243, 1997.
References
Goldfinger,C., L.D. Kulm, L.C. McNeill, and P. Watts, Super-scale
Aprea, C., M. Unsworth, and J. Booker, Resistivitystructureof the
failure of the southernOregon Cascadiamargin, Pure Appl. GeoOlympic mountainsand Pugetlowland, Geophys.Res. Lett., 25, ! 09phys., in press,1999.
112, 1998.
Gulick, S.P.S.,A.M. Meltzer, and S.H. Clarke Jr., Seismicstructureof
Atwater, T., and J. Stock,Implicationsof recent refinementsin Pacificthe southernCascadiasubductionzone and accretionaryprism north
North America plate tectonic reconstructions,Eos Trans. AGU,
of the Mendocino triple junction, d. Geophys.Res., 103, 27,20727,222, 1998.
79(45), Fall Meet. Suppl.,F206, 1998.
Beaudoin,B.C., et al., Transition from slab to slabless:Resultsfrom the
Hauessler,P.J., D. Bradley, R. Goldfard, L. Snee, and C. Taylor, Link
1993 Mendocino triple junction seismicexperiment,Geology, 24,
between ridge subduction and gold mineralization in southern
195-199, 1996.
Alaska, Geology,23, 995-998, 1995.
Bromery,R.W., andP.D. Snavely Jr., Geologic interpretationof reconHussong,D.M., L.K. Wipperman, and L.W. Kroenke, The crustal
naissancegravity and aeromagnetic surveys in northwesternOrestructureof the OntongJavaand Manihiki oceanicplateaus,d. Geogon, U.S. Geol. Surv.Bull. 1181-N, 1964.
phys.Res.,84, 6003-6010, 1979.
Butler,R.F., Paleomagnetism:
MagneticDomainsto Geologic Terranes, Hyndman, R.D., The lithoprobecorridor acrossthe Vancouver Island
319 pp., Blackwell Sci., Cambridge,Mass., 1992.
continentalmargin: The structuraland tectonicsconsequencesof
Byrne,D.E., W. Wang, and D.M. Davis, Mechanicalrole of backstops
subduction,Can. d. Earth Sci., 32, 1777-1802, 1995.
in the growthof forearcs,Tectonics,12, 123-144, 1993.
Hyndman,R.D., and K. Wang, Thermal constraintson the zone of maCande,S.C., andD.V. Kent, A new geomagneticpolarity time scale for
jor thrustearthquakefailure: The Cascadiasubductionzone, d. Geothe late Cretaceousand Cenozoic, d. Geophys.Res., 97, 13,917phys.Res.,98, 2039-2060, 1993.
13,951, 1992.
Kulm, L.D., andG.A. Fowler, Oregon continentalmargin structureand
Cloos,M., Thrust-typesubduction-zone
earthquakesand seamountasstratigraphy:
A testof the imbricatethrustmodel, in The Geology of
perities:A physicalmodelfor seismicrupture,Geology,20, 601-604,
ContinentalMargins, editedby C.A. Burke and C.L. Drake, pp. 2611992.
283, Springer-Verlag,New York, 1974.
Smith,1991]. Research
was supported
by theU.S. Geological
Survey
FLEMINGAND TRI•HU: BURIEDRIDGEBENEATHTHE OREGONMARGIN
Lallemand,S., andX. Le Pichon,Coulombwedge model appliedto the
subductionof seamountsin the Japan trench, Geology, 15, 10651069, 1987.
Ludwin,R.S., C.S. Weaver,andR.S. Crosson,Seismicityof Oregon and
Washington,
editedby D.B. Slemmonset al., in Decade Map, vol. 1,
Neotectonics
of North America, pp. 77-98, Geol. Soc. of Am., Boulder, Colo., 1991.
MacDonald,K.C., P.J. Fox, R.T. Alexander,R. Pockalny, and P. Gente,
Volcanic growthfaults and the origin of Pacific abyssalhills, Nature,
380, 125-129, 1996.
MacKay, M.E., Structuralvariation and landwardvergenceat the toe of
the Oregonaccretionary
prism,Tectonics,14, 1309-1320, 1995.
Marine Trackline Geophysics [CD-ROM set, version 3.2],
http://www.ngdc.noaa.
gov,NationalGeophysicalData Center, Boulder, CO, 1996.
McCaffrey, R., C. Goldfinger,M.H. Murray, P. Zwick, J.L. Nabelek,
and C.K. Johnson,GPS constraintson forearc sliver motion, plate
coupling,and strainpartitioningin northwesternOregon (abstract),
Eos Trans.AGU, 79(45), Fall Meet. Suppl.,F875, 1998.
McNeill, L.C., L.D. Kulm, C. Goldfinger, and R.S. Yeats, Evolutionof
the late NeogenecentralCascadiaforearc basin:investigations
of a
late Miocene unconformity, Ph.D. thesis, Oreg. State Univ.,
Corvallis, 1998.
Miller, K.C., andJ. Howie, Integratedcrustalstructureacrossthe south
centralCaliforniamargin:SantaLucia Escarpmentto San Andreas
fault, d. Geophys.Res.,98, 8173-8196, 1993.
Mitchell, C.E., P. Vincent,R.J. Weldon II, and M.A. Richards,Presentdayverticaldeformation
of the Cascadiamargin,PacificNorthwest,
USA, d. Geophys.Res.,99, 12,257-12,277,1994.
Nelson,A.R., et al., Radiocarbonevidencefor extensiveplate-boundary
rupture about300 years ago at the Cascadiasubductionzone, Nature, 378, 371-374, 1995.
Oleskevich,D.A., Comparisonof thermal constraintmodellingon selectedprofiles along the Cascadiasubductionzone, Work team report,Univ. of Victoria Sch.of Earth and Ocean Sci., Victoria, B.C.,
20,447
Snavely,P.D., Jr., H.C. Wagner,D.L. Lander,Geologiccrosssectionof
the centralOregoncontinentalmargin,report,Plate Tecton. Group,
U.S. Geodyn.Comm., Geol. Soc. of Am., Boulder,Colo., 1980.
Snavely,P.D., Jr., H.C. Wagner,D.L. Lander, Land-seageologiccross
sectionof the southernOregon continentalmargin, U.S. Geol. Surv.
Misc. Invest. Ser. Map, 1-1463, 1985.
Tabor, R.W., R.S. Yeats, and M.L. Sorensen,Geologic map of the
Mount Angelesquadrangle,Clallam and Jeffersoncounties,WA, U.
S. Geol. Surv.Map, GQ-958, 1972.
Talwani, M., and J.R. Heirtzler, Computationof magnetic anomalies
causedby two-dimensionalbodiesof arbitrary shape,in Computers
in the Mineral Industries,vol. 9, part 1, edited by G.A. Parks, pp.
464-480, StanfordUniv. Press,Stanford,Calif., 1964.
Talwani, M., J.L. Worzel, and M. Landisman,Rapid gravity computationsfor two-dimensionalbodieswith applicationto the Mendocino
submarinefracturezone,J. Geophys.Res., 64, 49-59, 1959.
Telford,W.M., L.P. Geldart,R.E. Sheriff, AppliedGeophysics,770 pp.,
CambridgeUniv. Press,New York, 1990.
Tr•hu, A.M., I. Asudeh, T.M. Brocher, J.H. Luetgert, W.D. Mooney,
J.L. Nabelek, and Y. Nakamura, Crustal architecture of the Cascadia forearc, Science, 265, 237-243, 1994.
Tr•hu, A.M., G. Lin, E. Maxwell, C. Goldfinger,A seismicreflection
profile acrossthe Cascadiasubductionzone offshore central Oregon:New constraintson methanedistributionand crustalstructure,J.
Geophys.Res., 100, 15,10!-15,116, 1995.
U.S. GeologicalSurvey, Aeromagneticsurvey compositemap, Oregon
coast,U.S. Geol. Surv. OpenFile Rep., 70-341, 1970.
yon Huene,R., andD. Scholl,Observationsat convergentmarginsconcerningsedimentsubduction,subductionerosion,and the growth of
continentalcrust,Rev. Geophys.,29, 279-316, 1991.
yon Huene, R., I.A. Pecher, M.-A. Gutscher,Development of the accretionaryprism along Peru and material flux after subductionof
Nazca Ridge, Tectonics,15, 19-33, 1996.
Wang, K., Simplified analysisof horizontal stressesin a buttressed
forearcsliver at an obliquesubductionzone, Geophys.Res. Lett., 23,
2021-2024, 1996.
Wang, W.-H., and D.M. Davis, Sandboxmodel simulationof forearc
evolutionand noncritical wedges, d. Geophys. Res., 101, 11,329-
Canada, 1994.
Parsons,T., et al., A new view into the Cascadia subductionzone and
volcanicarc: Implicationsfor earthquakehazards along the Wash11,339, 1996.
ingtonmargin,Geology,26, 199-202, 1998.
Peacock,S.M., Thermal effects of metamorphicfluids in subduction Weaver, C.S., and K.M. Shedlock,Estimatesof seismicsource regions
from the earthquakedistributionandregionaltectonicsin the Pacific
zones,Geology,15, 1057-1060, 1987.
northwest,in Assessing
EarthquakeHazards and ReducingRisk in
Peterson,C.P., P.W. Loubere,L.D. Lulm, andJ.S. Peper, Stratigraphyof
thePacificNorthwest,editedby A.M. Rogers, U.S. Geol. Surv. Prof
continental shelf and coastal region, in Western North American
Pap., 1560, 161-182, 1996.
ContinentalMargin and Adjacent Ocean Floor off Oregon and
Webring,M., SAKI: A fortranprogramfor generalizedlinear inversion
Washington,
OceanMargin Drill. Prog. Reg. Atlas Set., atlas 1, edof gravity and magneticprofiles, U.S. Geol. Surv. Open File Rep.
ited by L.D. Kulm, sheet8, Mar. Sci. Int., WoodsHole, Mass., 1984.
85-122, 1985.
Sample,J.C.,M.R. Reid,H.J. Tobin, andJ.C. Moore, Carbonatecements
Wells, R.E., D.C. Engebretson,
P.D. SnavelyJr., andR.S. Coe, Cenozoic
indicate channeledfluid flow along a zone of vertical faults at the
platemotionsandthe volcano-tectonicevolutionof westernOregon
deformationfront of the Cascadiaaccretionarywedge, Geology, 21,
andWashington,Tectonics,3, 275-294, 1984.
507-510, 1993.
Wells, R.E., C.S. Weaver, and R.J. Blakely, Fore-arc migration in CasScheidegger,K.F., Thermal evolutionof the Juan de Fuca plate, in
cadiaand its neotectonicsignificance,Geology,26, 759-762, 1998.
WesternNorth American ContinentalMargin and Adjacent Ocean
Werner, K.S., E.P. Graven, T.A. Berkman, M.J. Parker, Direction of
Floor off Oregon and Washington,Ocean Margin Drill. Prog. Reg.
maximumhorizontalcompressionin western Oregon determinedby
Atlas Set., atlas 1, edited by L.D. Kulm, sheet 8, Mar. Sci. Int.,
borehole breakouts,Tectonics,10, 948-958, 1991.
Woods Hole, Mass., 1984.
Wessel, P., and W.H.F. Smith, Free software helps map and display
Scholtz,C.H., and C. Small, The effect of seamountsubductionon seisdata,Eos Trans.AGU, 72, 441, 1991.
mic coupling,Geology,25, 487-490, 1997.
Shrieve,R.L., andM. Cloos,Dynamicsof sedimentsubduction,melange Wilson, D.S., Confidence intervalsfor motion and deformation of the
formation,andprismaccretion,d. Geophys.Res., 91, 10,229-10,245,
1986.
Simpson,R.W., and A. Cox, Paleomagneticevidence for tectonicrotation of the OregonCoastRange,Geology,5, 585-589, 1977.
Snavely,P.D., Jr., Tertiary geologicflamework, neot½ctonics,
and petroleumpotentialof the Oregon-Washington
continentalmargin, in
Geologyand ResourcePotentialof the ContinentalMargin of Western North America and Adjacent Ocean Basins: Beaufort Sea to
Baja California,Earth Sci. Set., vol. 6, edited by D.W. Scholl,A.
Grantz,andJ.G. Vedder,J.G., pp. 305-335, Circum-Pac.Counc.for
Energy and Miner. Resour.,Houston,Tex., 1987.
Shayely,P.D., and R.E. Wells, Cenozoic evolutionof the continental
marginof OregonandWashington,in Assessing
EarthquakeHazards
andReducingRiskin the Pacific Northwest,edited by A.M. Rogers,
U.S. Geol.Surv.Prof Pap., 1560, 161-182, 1996.
Snavely,P.D., Jr., N.S. MacLeod, and H.C. Wagner, Tholeiitic and alkalic basaltsof the Eocene Siletz River volcanics, Oregon Coast
Range,Am. d. Sci., 266, 454-481, 1968.
Juande Fucaplate,J. Geophys.Res.,98, 16,053-16,071,1993.
Yeats, R. S., L.D. Kulm, C. Goldfinger, L.C. McNeill, Stonewall anticline:An activefold on the Oregoncontinentalshelf,GSA Bull., 110,
572-587, 1998.
Zelt, C.A., andR.S. Smith,Seismictravel-time inversionfor 2-D crustal
velocitystructure,Geophys.J. lnt., 108, 16-34, 1992.
S. W. Fleming, 204-620 Dobson Road, Duncan, British Columbia,
Canada V9L
4R8.
A.M. Tr•hu, College of OceanicandAtmosphericSciences,Oregon
StateUniversity,104 OceanAdministrationBldg., Corvallis,OR 973315503. (trehu•oce.orst.edu)
(ReceivedJuly 10, 1998; revisedApril 16, 1999;
acceptedApril 27, 1999.)