Seismic structure of the U.S. Mid-Atlantic continental margin

JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 99, NO. B9, PAGES 17,871-17,891, SEPTEMBER 10, 1994 Seismic structure of the U.S. Mid-Atlantic continental margin W. StevenHolbrook, • G.M. Purdy, • R.E.Sheridan, 2 L. Glover111, 3 M. Talwani, 4 J. Ewing,•,4 andD. Hutchinson 5 Abstract. Multichannelandwide-angleseismicdatacollectedoff Virginia duringthe 1990EDGE Mid-Atlanticseismicexperiment providethemostderailedimageto dateof the continent-ocean transitionon theU.S. Atlanticmargin. Multichanneldatawereacquiredusinga 10,800 in3(177L)airgun array and6-km-long streamer, andcoincident wide-angle data were recorded by tenoceanbottomseismicinstruments. A velocitymodelconstructed by inversionof wide-angleandvertical-incidence traveltimesshowsstronglateralchangesin deep-crustal structure acrossthemargin. Lower-crustal velocitiesare6.8 km/sin riftedcontinental crust, increaseto 7.5 km/s beneaththe outer continentalshelf, and decreaseto 7.0 km/s in oceanic crust.Promine ntseaward-dipping reflections withinbasement liewithinlayers of average velocity6.3-6.5 km/s,consistent with theirinterpretation asbasaltsextrudedduringrifting. Thehigh-velocity lowercrustandseaward-dipping reflections comprise a 100-km-wide,25-kmthickocean-continent transitionzonethatconsists almostentirelyof maficigneousmaterial accretedto the marginduringcontinental breakup.The boundarybetweenriftedcontinentalcrust andthisthickigneouscrustis abrupt,occupyingonlyabout20 km of themargin. Appalachian intracrustal reflectivitylargelydisappears acrossthisboundaryasvelocityincreases from5.9 km/s to >7.0 km/s, implyingthat the reflectivityis disruptedby massiveintrusionand that very little continentalcrustpersistsseawardof thereflectivecrust. The thickigneouscrustis spatiallycorrelated with theEastCoastmagneticanomaly,implyingthatthebasaltsand underlying intrusives causetheanomaly.The detailsof theseismicstructure andlackof independent evidencefor anappropriately locatedhotspotin thecentralAtlanticimplythat nonplumeprocesses are responsible for the igneousmaterial. Introduction A passive, or rifted, continental margin comprises the rifted edge of a continentand the adjoining transitionto normal oceaniccrust. Understandingthe rift-drift evolution of a rifted margin requiresaccurateknowledgeof its crustalstructure, which contains records of such key processesas subsidence,crustal thinning, and volcanism. Although rifted margins are often difficult to probe seismicallydue to their thick sediments and strong lateral heterogeneity, modern seismic data are yielding new insights into their deep structure. The discovery that great thicknesses (up to 25 km) of mafic igneous rocks were emplaced on the margins of the North Atlantic during rifting [Mutter et al., 1984; White et al., 1987] showed that continental breakup cannot always be explained by simple stretching and thermal subsidence [McKenzie, 1978; Sleep, 1971]. This led to a classificationof rifted margins into two types, volcanic and nonvolcanic, and several geodynamic models were put forth to explain this apparent dichotomy [Mutter et al., 1988; White and McKenzie, 1989]. Recent seismic data collected on the U.S. Atlantic margin, including those presentedin this paper, have an important bearingon theorig•nof riftedmarginmagmatism [Holbrook and Kelemen, 1993] and have led to the suggestionthat the volcanic/nonvolcanicclassificationmay be an oversimplification [Mutter, 1993]. The distributionof igneousmaterial accretedduring continentalbreakupis inferredon the basisof two seismiccharacteristics: seaward-dipping reflectionsbeneaththe postriftun1Department of Geologyand Geophysics, WoodsHole conformity interpreted as subaerially emplaced basalt flows Oceanographic Institution,WoodsHole, Massachusetts. and thick, high-velocity (7.2-7.6 km/s) lower crust inter2Department of Geological Sciences, Rutgers University, New preted as the underlying intrusive rocks. Existing seismic Brunswick,New Jersey. 3Department of GeologicalSciences,Virginia Polytechnic transectsare too few in number to thoroughlymap the global distributionof volcanic and nonvolcanicmarginsnor to fully University,Blacksburg,Virginia. assessto what degree intermediate types exist. However, 4Houston Advanced Research Center, TheWoodlands, Texas. 5U.S.Geological Survey, Branch of AtlanticMarineGeology, available data indicate that volcanic margins are found in the Woods Hole, Massachusetts. North Atlantic (Vcring Plateau [Mutter et al., 1984] and Hatton Bank [White et al., 1987]), off Northwest Australia (Cuvier margin [Hopper et al., 1992]) and off the U.S. east Copyright1994by the AmericanGeophysical Union. coast[Holbrook et al., 1994; LASE StudyGroup, 1986; Trdhu et al., 1989], while nonvolcanic margins exist off Iberia Papernumber94JB00729. 0148-0227/94/94JB-0072955.00 [Horsefield et al., 1994; Pinheiro et al., 1992; Whitmarsh et 17,871 17,872 HOLBROOK ET AL.: U.S• MID-ATLANTIC CONTINENTAL MARGIN al., 1990] and Northwest Australia (Exmouth Plateau [flopper et al., 1992]). In this paper we present multichannel seismic reflection (MCS) and wide-angle, ocean bottom seismic data which provide the clearest image to date of deep structureon the U.S. east coast margin. The data were collected in 1990 in the EDGE Mid-Atlantic margin experimentoff Virginia (Figure 1). Interpretationsof the MCS data were presentedby Sheridan et al. [1993] and of wide-angle data on strike line 802 by Holbrook et al. [1992b]. The resultsof this experimentshow that the U.S. Mid-Atlantic margin is strongly volcanic, as it containsrift-related igneous rocks up to 25 km thick in the continent-ocean transition zone. 1973], spreadingwas establishedat the Mid-Atlantic Ridge and has continueduntil the present. This rift-drift history has produceda margin consisting,in the broadest terms, of three structural zones: rifted continental crust of Appalachian origin beneath the continental shelf, oceaniccrustbeneaththe continentalrise, and the intervening 50- to 100-km-widezone of "transitional"or "rift-stage"crust [cf. Klitgord et al., 1988]. The nature of the crust in this ocean-continent transitionzone, betweenthe basementhinge zone and the landward limit of well-defined oceanic crust, has importantimplicationsfor the evolution of the margin and is the subjectof continuingresearch. Critical questionsinclude the amountof rift-related igneousrocks, the thicknessof remnant continental crust, the nature of the contact between continentaland rift-related igneousrocks, the depth, structure, and age of the Moho, and the thicknessof synrift and postrift Geological and GeophysicalSetting sediments. The U.S. Atlantic margin was formed by the rifting of Pangea in the Late Triassic and establishmentof the MidAtlantic Ridge spreadingcenter in the Early Jurassic. During the early stagesof rifting, clastic sedimentsof mainly lacustrine origin were deposited in onshore, "Newark-type" rift basinsfrom about228 to 190 Ma [Manspeizerand Cousininet, 1988]. Sometime in the Early Jurassic,rifting is believed to have steppedeastwardto the eventual location of the basement hinge zone [Harry and Sawyer, 1992]. The earliest spreading history of the central Atlantic is uncertaindue to lack of magnetic anomaliesin seafloor created during the Jurassicquiet zone, but Klitgord and Schouten [1986] estimated that seafloorspreadingbegan at ~175 Ma. Following an inferred ridge jump at 170 Ma, thoughtto be responsiblefor the Blake Spur magnetic anomaly [Klitgord and Schouten,1986; Vogt, Subsidence in the transition zone concentrated sedimentationin several major offshorebasins,including the Blake Plateau Basin, Carolina Trough, Baltimore Canyon Trough, and GeorgesBank Basin. The EDGE lines crossthe southernBaltimore Canyon Trough, the thickest(up to 18 km [Grow et al., 1988]) of the offshorebasins,althoughsediment thickness is somewhat less at the location of the EDGE lies, the East Coast, Blake Spur, and Brunswick magnetic anomalies, have been interpreted to mark significant crustal boundaries (Figure 1). The East Coast magnetic anomaly (ECMA) is a strongpositive anomaly that extendsabout2300 km along the margin, from the Blake Spur Fracture Zone to offshore Nova Scotia. The ECMA has been variously interpreted as an edge effect between continentalor rift-stage and 39øN 38øN Ocean 37øN 21 36øN 78øW 76øW line. Several major geophysical lineaments demarcate the U.S. Atlantic margin. Three prominent offshoremagnetic anoma- 74øW 72øW Figure 1. Locationof EDGE Mid-Atlanticseismicexperiment.Resultsfromline 801 arediscussed in thispaper. Positions of WoodsHole oceanbottomhydrophones are shownby solidcircles,U.S.GeologicalSurveyoceanbottomseismometers by open circles, onshoreseismometers by shadedcircles. ECMA is East Coastmagneticanomaly;BSMA is Blake Spur magnetic anomaly;BMA is Brunswickmagneticanomaly. HOLBROOK ET AL.: U.S. MID-ATLANTIC CONTINENTAL MARGIN 17,873 oceaniccrust [Hutchinsonet al., 1983; Keen, 1969], an intra- recorded by a 240-channel, 6-km-long hydrophonestreamer basement ridgeor dike [Drakeet al., 1959;Emeryet al., 1970; Klitgordand Behrendt,1979],a majorbasement fault [Alsop and ten ocean bottom seismic instruments. In addition, shots were recordedby an onshorearray of ten Reftek seismometers that extendedthe apertureof the experimentby about 250 km. The onshoredata are not discussedin this paper. Wide-angle data were recordedon line 802 by one Woods anomaly(BSMA), a positiveanomalylocated150 to 250 km seawardof the ECMA from the Long Island FractureZone to Hole Oceanographic Institution (WHOI) ocean bottom the BahamasFracture Zone, is thought to representa ridge hydrophone(OBH) [Koelschet al., 1982] and on line 801 by jump thatoccurredat 170 Ma [Klitgordand Schouten,1986; five OBH and five U.S. Geological Survey (USGS) ocean Vogt, 1973]. The Brunswickmagneticanomaly(BMA) is a bottom seismometers(OBS) (Figure 1). An interpretationof compositepeak-troughanomaly that extends from Cape wide-angle data from line 802 was presentedby Holbrook et Hatterassouthto the SoutheastGeorgiaembayment,where it al. [1992b]; in this paper we present results from line 801. swingsonshoreinto Georgia. Offshore,the BMA has been The averageinstrumentspacingon line 801 was 13 km on the interpretedas either the result of Mesozoic rift basins continentalshelf and upperslopeand 28 km on the rise, with a [Hutchinson et al., 1983] or as the juxtapositionof highly gap of 38 km on the lower slopedue to the failure of one OBH and Talwani, 1984], or rift-related volcanics [Austin et al., 1990; Talwani et al., 1992]. The Blake Spur magnetic magnetized Mesozoicvolcanicsandlessmagneticcontinental and one OBS to record useful data. crust [Austin et al., 1990]. In addition, an onshoremagnetic anomalyin the EDGE transectarea, the Salisburymagnetic anomaly,has been interpretedas a Paleozoicsuturewithin Seismic Reflection Appalachiancrust [e.g., Pratt et al., 1988; Sheridanet al., 1993]. Offshore free-air gravity anomaliesshow a prominent peak that follows the continentalshelf break. Althoughthis anomaly is partially due to seafloor topographyand sedimentarystructure,a major changein basementdensitybeneath the hingezoneis requiredto matchit [11olbroolc et al., 1994]. Over the past fifteen years, seismic investigationshave provided growing evidencethat rifting of the U.S. Atlantic margin was accompaniedby substantialmagmatic activity. Hinz [1981] first interpreted seaward-dippingreflections on MCS profilesin the BaltimoreCanyonTrough as evidencefor volcanism. Klitgord et al. [1988], in contrast, attributed thosereflectionsto synrift sedimentsbut interpretedsimilar reflectionsin the GeorgesBank Basin as volcanic in origin. The first indicationsof voluminousmagmatismon the margin came from along-strike,two-shipexpandingspreadprofiles in the Baltimore Canyon Trough [Diebold et al., 1988; LASE Study Group, 1986] and along-strikeexplosionrefractionprofiles in the Carolina Trough [Trdhu et al., 1989]. Velocity models constructed from these data show a 6- to 13-km-thick, high-velocity (7.2-7.4 km/s) lower-crustal layer interpreted as mafic igneousrock underplatedbeneaththe margin during rifting. In both of these studies,however, relatively poor constraintswere available on the landward extent of the highvelocity crust and on the structureof the immediately overlying crust. More recently,coincidentMCS and oceanbottom wide-angledata collectedin the CarolinaTrough have imaged seaward-dipping reflections and high-velocity lower crust [Austin et al., 1990; Holbroolc et al., 1994, 1992c; Oh et al., 1991], thus confirming earlier interpretationsthat prodigious volcanismaccompanied the breakupof the margin. Thesedata sets,togetherwith the EDGE data set presentedhere [Sheridan et al., 1993], identify the U.S. Atlantic margin as volcanic Data The EDGE seismicreflectiondata from line 801, originally interpretedby Sheridan et al. [1993], provide a good image of the postrift sedimentsacrossthe entire transectand of the subsedimentary crust beneath the continental shelf (Figure 2). The postrift sediments thicken from about 1.5 s two-way travel time (TWTI') at the landwardend of the profile to about 4.5 s TWTT beneaththe upper continentalrise. A prominent carbonatebank edge is imaged at model km 105-110 at about 4 s TWTr (Figure 2b). Basementis markedby the postriftunconformity,which extendsseawardto the J3 scarp(model km 190) where well-defined oceanic basementis first observed. The reflective character of subsedimentarycrust changes significantlyacrossthe profile. At the landwardend (model km 0-45), subsedimentarycrust consistsof a relatively transparentupper part (2-4 s TWTT) overlying a brightly reflective middle and lower crust (6-12 s TWT'I') that containsnumerous subhorizontal, dipping, and arcuate events that might be diffractions. This zone is interpretedas continental crust of Appalachian origin. Prominent east-dippingreflections that intersectstrike line 802 at 4 s and 7 s TWTT were interpreted by Sheridanet al. [1993] as thrustsequences,the deeperof which was interpretedas the Taconicsuture. Crosscuttingthis reflectivity at its base is a clear, planar reflection that dips landwardfrom 11 s to 13 s TWTT and is interpretedas Moho. Several steeply dipping events are seen beneathMoho; these probablyrepresentdiffractionsfrom the baseof the crust. Seaward of model km 50, bright mid-crustal to lowercrustal reflectivity gives way to prominent seaward-dipping reflections (at 10-11 s TWTT) beneath the postrift unconformiry overlying a largely reflectionlessmiddle crust and, from model km 50-75, bright, subhorizontalreflectionsinterpreted as lowermostcrust and Moho (10-11 s TWTI'). Sheridan et al. [1993] interpreted the seaward-dippingreflections as basalt [Holbrook and Kelemen, 1993]. flows depositedduring rifting and/or early drifting, on the basis of their similarity to seaward-dippingpackagesimaged on other margins [Hinz, 1981; Mutter et al., 1984; White et Data Acquisition al., 1987]. Individual seaward-dippingreflections can be The data presentedhere were shot in September 1990 by traced laterally for up to 20-25 km and extend to at least 9 s the industry vessel Geco Searcher. Three profiles were TWTr. The bright reflectionsat model km 50-75 were interrecorded, two acrossthe margin (lines 801 and 803) and one preted by Sheridan et al. [1993] as mafic/ultramafic sills parallel to the coaston the continentalshelf (line 802). Shots emplacedat the baseof the crustduring rifting. from a 36-element, 10,800 cu. in. (177 L) airgun array were Seawardof model km 75, the only deep crustalreflections fired at constant distance intervals of 50 m (-•22 s) and are the seaward-dippingreflections,which disappearat about 17,874 HOLBROOK ETAL.' U.S.MID-ATLANTIC CONTINENTAL MARGIN SP 1475 1875 2275 2675 3075 3475 20 40 60 80 100 o 5 lO 15 - 0 Distance (kin) Figure 2a.MCSstack oflandward portion ofline801,after Sheridan etal.[1993]. Data have been processed withcoherency filter. SP 3475 3875 4275 4675 5075 5475 120 140 160 180 200 0 5 10 15 100 Distance (km) Figure2b. Seaward portion ofline801,plotted asin Figure 2a. HOLBROOK ET AL.: U.S.MID-ATLANTICCONTINENTAL MARGIN SP 1475 875 2275 2675 3075 17,875 3475 o _ 15 - 0 20 40 .•..,• ?•.• .. ,"..& ..•; 60 /%. 80 lOO Distance (kin) Figure 2c.Same asFigure 2a,withinterpretive line.drawing ofreflections used infloating reflector analysis. modelkm 110, and a faint, landward-dipping reflectionat arrivals fromthecrystalline crustandMohochanges markedly 10.5-11s TWTY at modelkm 170. As will be shownbelow, from continentalto oceaniccrust. Data from OBH 17, which the wide-angle data confirm that this latter event lies near was located at the edge of the continentalshelf, show this Moho; a similar glimpseof Moho in the ocean-continent change well (Figure3b). Landward, a complexpackage of transition wasobtained onMCSdatain theCarolina Trough reflectionsfrom deep continentalcrust is observed. These [Austinet al., 1990]. Severalprominent seafloor pegleg events are the wide-angle expressionof the mid-crustaland multiples areobserved crosscutting thisreflection.Finally, lower-crustal reflectivityobserved on theMCS stack(Figure seawardof model km 190, hyperbolicreflectionsat 8.5 s 2a). Seaward, a simplerpatternis observed, consisting of a TWTT mark the landwardlimit of well-definedoceanicbase- refraction from the uppercrust,a reflection(or retrograde ment [cf. Klitgordet al., 1988]. divingwave)froma mid-crustal boundary (thelayer2/layer3 transition), a refractionfrom the lowercrust,anda highOcean Bottom Seismic Data Airgun shots were recorded on ten ocean bottom instru- ments,fiveWHOI OBH andfive USGSOBS(Figure3). Data amplitudepostcritical Moho reflection(Figure3b). VelocityModeling quality is excellent on the OBH and fair to excellent on the A velocity model acrossthe entire continent-oceantransi- OBS. Previousshotnoise,generated by the rapid(-20 s) firing rate of the airgunarray [Nakamuraet al., 1987], partiallyobscures arrivalsat 40 km intervals,especially on tion was derivedby ray tracemodelingof wide-angleand deeper-water instruments (e.g., Figure3c), but it did not sub- vertical-incidence traveltimes(Figure4a). Traveltimemodel- ing followedthesesteps: (1) identification of wide-angle reflectionsand refractionson all recordsections;(2) correlastantially hamperphasecorrelation.Therefore, although it is tion of wide-anglereflectionswith vertical-incidencereflecpossibleto suppress thisnoise[Holbrookand Reiter, 1994], tionson MCS datafor sediments andbasement; and(3) iterawe madeno attemptto do so. The processing flow for wide- tive ray trace modelingof both wide-angleand verticalangledataincludedband-pass filter (3-12 Hz), notchfilter on incidencetraveltimesusingthe inversionprocedure of Zelt someOBSto removemonochromatic 5.6-Hznoise,coherency and Smith[1992]. Separation of nodepointswasdetermined filter, and five-trace mix. Reflections and refractions from the entire crust were recorded on the ocean bottom instruments. Reflections and by iterativetesting;nodeswerepositioned as closetogether as possible(-30 km in the lowercrust)withoutgenerating unstableinversionresults. Wide-angle travel times were refractions fromthesediments andpostriftunconformity were invertedsimultaneously for velocity and depth of nodes; recorded across the entire transect. All ten instruments however,only depthsto layerswere allowedto vary when recorded wide-angle Mohoreflections, providinggoodcontrol inverting vertical-incidence data, because the verticalon Mohostructure across theentiretransect.The patternof incidencedata cannot constrainvelocities. Velocities beneath 17,876 HOLBROOK ET AL.' U.S. MID-ATLANTIC OBH 16 were fixed according to the structuredeterminedfor strike line 802 [Holbrook et al., 1992b]. The velocity model shows profound changes in crustal structureacrossthe margin from rifted continentalto oceanic crust (Figure 4a). Postrift sediments increase in thickness from 1.5 km to a maximum of about 8 km beneaththe slope and rise. A striking lateral changeis evident in subsedimentary crust, from normal continental velocity structure at the CONTINENTAL MARGIN landward end of the model to crust consistingof a 6.3-6.5 km/s seaward-dippingwedge overlying a thick, high-velocity (7.5 kin/s) unit beneath the outer continental shelf. Farther seaward, velocities decreaseto 5.3-5.6 km/s in upper basement and 7.0 km/s in lower crust. Subsedimentary crust thins from about 38 km to 7-8 km across the model. It should be noted that important geological boundariesmay not always correlate with layer boundariesin the velocity model: for 10 ' -20 a 20 40 60 OBH 16 -80 b 0 ,' ' ..... 80 , ,'0 • ' :•i•:'!' ;' 100 120 140 160 180 Distance (km) -60 I I I I I I I I -40 -20 0 20 40 60 80 100 OBH 17 Distance(km) Figure3. Recordsections of oceanbottomseismic data,processed witha 4-12 Hz band-pass filter,a coherency filter,anda 500-msautomatic gaincontroloninnertraces.OBSdataareverticalcomponent.. (a) OBH 16,(b) OBH 17,(c) OBSA8, (d) OBS A1, (e) OBH 19, (f) OBH 20. HOLBROOKET AL.: U.S. MID-ATLANTIC CONTINENTALMARGIN -4O c d -20 0 OBS A8 2O 20 40 60 17,877 80 100 Distance (km) - 1o0 OBS -80 -60 60 Distance (kin) Figure 3. (continued) example, theprofound lateralchanges in velocitywithinlayer 19 (e.g.,Figure7). The decrease in lower-crustal velocityto 7, which indicatemajor compositional changes,are not 7.0 km/sin oceanic crustis likewiseconstrained by bothfirst separatedby model interfaces. arrivalsand Moho reflectionson OBS A1 and OBHs 19, 20, Constraints on velocities withinlayersanddepthsto inter- and 21 (e.g., Figure8). facesin the subsedimentary crustcamefrom the following phases:refractions fromlayers5 and7 andwide-angle reflectionsfrominterfaces 5, 6, 7, 8, andM (Figure5). Thesecon- Resolution and Uncertainty straints are summarized in Table 1. The increase in lowerEstimatesof resolutionand uncertaintyin velocity and crustalvelocitybeneaththe seaward-dipping wedge(SDW) is depthare important,thoughdifficult, goalsof deepcrustal constrained by traveltimesof firstarrivalsthatrefractthrough studies.Resolution providesa measure of thedependence of thelowercruston OBH 16 and17 (Figure6), aswell asmove- thetraveltimefit on a givenmodelparameter (i.e., a depthor out on Moho reflections on instrumentsas far seaward as OBH velocitynode). Uncertaintyrefersto the rangeof valuesof a 17,878 HOLBROOK ETAL.' U.S.MID-ATLANTIC CONTINENTAL MARGIN - 140 - 120 - 100 -80 -60 OBH19 f - 120 OBH20 -40 -20 0 20 40 60 Distance(km) - 100 -80 -60 -40 -20 0 20 40 60 Distance(kin) Figure 3. (continued) only by wide-anglereflections. Velocitiesare modelparameterover whichthe traveltime fit remainsaccept- constrained able. Resolutionvalues, which can be calculatedformally dur- least well constrainedwithin the base of the SDW, at the edges ing inversion [Zelt and Smith, 1992], identify generally of the model, and in middle and lower (continental?)crustat whichmodelparameters are well constrained, but theyofferno modelkm 30-50. Bold linesin Figure4b connectdepthnodes valuesgreaterthan0.5; theselinesthusshow insightintouncertainty, whichmustbe estimated by perturba- withresolution tion of the velocity model. We estimated model resolution in the subsedimentarycrust interfacesthat are well constrainedby wide-anglereflections [Zelt andSmith,1992]. Moho is well constrained acrossmost by interpolatingresolutionvaluesbetweennodesanddisplay- of the model, from km 40 to km 210. Interface 6 is well con- whereit dipsseaward withinthewedgeandagainfrom ing theresultasa contourplot (Figure4b). The appearance of strained Figure 4b is stronglydependenton modelparameterizationmodel km 110 to 220 where it forms the top of lower crust. and shouldbe interpretedsolelyas a qualitativeindicationof The baseof the SDW (interface7) is constrainedonly from relative constraintson velocity in variousparts of the model. about model km 60-70, although additional evidence In rifted continental crust, velocities are well resolved in a supportingthe position of this interface comesfrom reflections on the coincident MCS data, as discussedbelow. Elsewhere in the model, velocities are best constrained in the Intracrustal boundaries and Moho in rifted continental crust are upper part of the seaward-dippingwedge and lower crust. well constrained at the intersection of strike line 802. narrow zone beneath strike line 802 [Holbrook et al., 1992b]. Uncertainty of deep-crustal velocities wasestimated by fixResolutionvaluesdecreasewith depthin thoselayersbecause in a givenportionof a layerandinverting releupper velocity nodesare constrainedby both refractionsand ingvelocities wide-anglereflections,while nodesat the baseof a layer are vant refractionand wide-anglereflectiontravel times for a HOLBROOK ET AL.: U.S. MID-ATLANTIC 16 C9 C4 C3 17 A8 CONTINENTAL A1 19 MARGIN 17,879 20 21 0 5 ::::::::::::::::::::::::.:':::::::::::::::::::::*::::::::: ........................................... i!•.i!•..':..::.':.:..•'::::-"-'•::•::•!• ::::::?::::::::::?::::::. 4 10 15 20 •!:-:....".'•:•'""•':" '"'"" '""•:•q?'"':'" '"':•--'"' •i!]5•i::•:•!i '"' tl •!:•i.• , -"' •-.:•7';: • •,...":'::':"'*:•---::•:?:::'..:-"•...•...:C""" ..•.........:.!•:•.:...:•::,:•:.•:•:..:...:...:..•:•:.....:::•:• •""'"'"" ...,• -:............................. •,:• ,,,.:,. ........ ::** .................... ..:: ........... ß ....................... ............................. ;/½.;½•i':i::-. '•--'-,-........: •:."•i'"'"•.' c'>:••?i:i.•:.•..........•, -':.x*'... ß :.•:ii•!i•: ............. *::•:.-".•....'•i • . ß '."•i',:•i•i :i!•i i I ..... -*............. -'- .... ,•...... •i:-'".:"•!iii:..'•. •i•i::""'-?."•....:...-'•i!::::•!::".:...:•!! -':" ß .... . ::.:i:•ii::::.i!i::ii::::::i!i:i:!i•::i:?•i:•):•':i:..'-:.. ':•.'.'.,•-: ...... •-•.•-•.•:" '2'"'w'"':"':*:.'"::: 555!i:•.•".•:..:• . .•i•!•!!!•':""!:!:•"*?'•' ':::::':':'"':":':'"':<":": "• 5 :'::'::? •'!:: 25 30 35 40 45 SP 14751875 2275 2675 3075 3475 3875 4275 4675 5075 5475 5875':• 6275 a 16 C9 C4 C3 17 A8 A1 19 20 21 0 5 10 15 20 25 30 35 40 45 SP 1475 b 0 I I 1875 2275 2675 3075 3475 3875 4275 4675 5075 5475 5875 6275 20 40 60 80 100 120 140 160 180 200 220 240 Distance (km) Figure4. (a)Velocity modelacross line801for subsedimentary crust, derived frominversion of vertical-incidence andwideangletraveltimes.Shading is proportional to velocity.Locations of OBHsshown by solidcircles, OBSsby opencircles. Interfaces arenumbered 1-9; M is Moho. Interface5 is thepostriftunconformity [Sheridan et al., 1993]. Numberswithin layers arevelocities inkilometers persecond, shown atthepositions ofnodepoints fortheinversion; where onlyonevelocity is shown in themiddleof a layer,thevertical velocitygradient is zero.Smallnumbers onbottom scalereferto shotpoint numbers oftheMCSsection (Figure 2). (b)Resolution values ofvelocity •odeswithin subsedimentary crust, contoured andplotted with shading proportional to resolution value.Thisplotprovides a qualitative basis forevaluating whichportions of themodelare best constrained. zone,theminimum best-fittingreflector structurefor that velocity. After repeat- wedge. In thelowercrustof the transition i,ng thisstepover a suiteof velocityvalues,theresultingRMS RMS travel time misfit (0.13 s) occursfor a velocity of 7.5 misfit values were tabulated to assessthe sensitivity of the km/s (Figure 9). Becausethe misfit becomesunacceptably travel time fit to velocity in that part of the model. The opti- large (> 0.16 s) for velocitiesless than 7.3 km/s or greater mum velocity shouldminimizetravel time misfit while allow- than 7.7 km/s, we estimate an uncertaintyof +0.2 km/s in ing raysto be tracedto a largenumberof observed datapoints. velocity in thatpartof themodel. Similarly,in loweroceanic Thig proceduredoesnot providean exhaustiveexplorationof crust, travel time misfit shows a clear minimum (0.09 s) for a large(> 0.13 the full rangeof possiblevelocitymodels,sinceit merely per- velocityof 7.0 km/s andbecomesunacceptably turbs a single model, which may only be a local rather than s) for deviations greater than +0.2 km/s (Figure 10). globalminimum. However,it doesprovideusefulestimates of Uncertaintyin Moho depth,estimatedfrom its deviationover the range of acceptablevelocities,is +3 km in the ocean.parameteruncertaintyfor the modelin question. We estimateduncertainty in three sectionsof the model: continent transition and +2 km in oceanic crust (Figures 9c lower crust in the ocean-continent transition zone, lower and 10c). Velocities in the seaward-dippingwedge are someoceanic crust, and the upper layer of the seaward-dipping what less well constrainedthan in the underlying lower crust. 17,880 HOLBROOK I , , , I . . ET AL.: U.S. MID-ATLANTIC , I . , , I , . , I . . , I , , , I , , , CONTINENTAL I , , , I , , , I , , , MARGIN I . , , I , . . ......... llll '"•' •-' ".I ,•,,,,• r---n - ..... lll#l "'"""!'-!.'" nil (a) lliii' ......... IIIIIIIIiii . " ..!.'.!!.!?.!!•.'•,•,• Ii I -20 ' ' ' I .... I 0 ' ' ' I 20 , I , ' ' ' I 40 . , I. ' ,, I -40 . I '' I ''' -20 . , , I , I . , ' ' . I ' ' ' 100 I,.. I. I''' I ' ' ' 120 ,, I' 20 I ' I.,, . , I '' 40 , . , I . I''' , I ' ' ' . ' ' ' I 180 ' ' ' I ' ' 200 I, I ' ' ' I 80 , I 160 I.,, 60 . I 140 15rl Iorl '' 0 . ' 80 ., IIIII ' I ' ' ' I' ' 60 . I 100 , . . 7! (c) ii ' I'''1'''1'''1'''1'''1'''1''' -20 0 20 40 60 80 100 120 Distance (km) Figure 5. (a)--(j) Observed(verticalbars)andpredicted (solidlines)traveltimesfor all instruments, basedon thevelocity modelof Figure4a. Observed pickshavebeeninterpolated to 1 km pickspacing for plotclarity;numberof picksusedin inversionexceeds numbershown.PhaselabelsareM, Moho;7, reflection fromboundary 7 of Figure4a;5r, refraction or head wave from beneath interface 5, etc. The RMS misfit curve showsa best velocityof 6.3-6.4 km/s but has a more poorly definedminimum,indicatingan uncertainty of about+0.3 km/s (Figure 11). Uncertaintyon depth to interface6 is about+2 km (Figure 1lb). tions(e.g., Moho) are matchedwell in the synthetics (e.g., Figure 12), althoughthe amplitudematchof refractedarrivals could be substantiallyimprovedby detailed modelingof velocity gradientswithin layers. These observationsindicate that while averagevelocitieswithin layersare fairly accurate Calculation of Synthetic Seismograms We calculated asymptotic ray theory synthetic seismogramsfor representativeinstrumentsto estimatethe amplitude responseof the velocity model. The syntheticsare provided for completeness; no attemptwas made to improvethe match of observedand predictedamplitudesby adjustingthe velocity model. In general,the observedcritical pointsof majorreflec- in our model,velocity gradientsare not. Comparison to Multichannel SeismicData A morecompleteview of the margin'sseismicstructureis obtained by combining in a single presentationthe MCS resultswith the velocity model derived from the wide-angle data. Becauseof the stronglateralheterogeneity of themargin HOLBROOK ET AL.' U.S. MID-ATLANTIC ,,,I.,. CONTINENTAL I.,,I.,,I,,,I,.,I... I MARGIN 17,881 I, I II ''' I'''1'''1'''1'''1'''1''' -20 -40 I ß , ß I ß ß ß I ß ß I' 0 ß I , , 20 ß 40 60 80 100 I,.,I,.,I,,,I,,, I,,, I, . (e) I'''1'''1''' -80 I'''1'''1'''1'''1''' -60 -40 -20 ' I''' 0 I''' I''' -40 -20 20 I''' 40 I 0 '' 20 60 ' I ' 40 I' 80 ' ' I ' 100 ' ' 60 I ' 80 Distance (krn) Figure 5. (continued) structure, vertical-incidence travel times cannot be converted lower crust beneath the SDW, where velocities increase sea- to depth using a simple vertical stretch;rather, they must be modeledusing the correct two-dimensionalray paths of the reflections. We have accomplishedthis by picking representative eventson the MCS sectionand modeling them as "floating reflectors"in the velocity model. This approachexploits the fact that an MCS stack approximatesa zero-offsetsection ward to 7.0-7.5 km/s, are largely reflectionless. We attribute the seaward disappearance of Appalachian reflectivity to increasing disruption by massive mafic intrusion during Jurassiccontinentalbreakup. The base of the crust is reflec- [Claerbout, 1985], so that vertical-incidence reflection times can be modeled by tracing rays to coincidentsource-receiver pairs. The resulting picture (Figure 13) approximatesa migrated line drawing and allows an effective comparisonof the MCS and wide-angledata sets. Vertical-incidence reflectivity shows a systematiccorrelation with velocity structure. Prominentintracrustalreflectivity is observedin Appalachiancrust with velocities of 6.36.8 km/s and as seaward-dippingreflectionsin the continentocean transition zone (Vp=4.9-6.7 km/s). The middle and tions observed tive at near-vertical incidence beneath continental crust and the landward portion of the SDW. Bright, laminated reflecat 10-11 s TWTT at model km 50-70 were inferred by Sheridan et al. [1993] to be cumulates or sills intruded during rifting, an interpretationsupportedhere by their position at the base of nonreflective crust of velocity 7.1-7.2 km/s. Landward-dippingreflectionsat 10-15 s TWTT beneath continental crust (model km 0-50) fall at or near (within several kilometers) the Moho. Although several steeplydipping eventsin the lower crust are observedseaward of model km 80, Moho is not visible on the MCS data beneath the thickestpart of the SDW or in oceaniccrust. •7,882 HOLBROOK ET AL.' U.S. MID-ATLANTIC CONTINENTAL MARGIN (g) ' I -120 ' ' ' I -100 ' ' ' I -80 ' ' ' I -60 ' ' ' I -40 ' ' ' I -20 ' ' ' I 0 ' ' ' I ' ' ' 20 I ' ' 40 ' I 60 8 (h) (i) Distance (km) (j) 0 ...•...,...•...•[ -80 Figure 5. (continued) -60 -40 -20 0 HOLBROOK ET AL.' U.S. MID-ATLANTIC CONTINENTAL MARGIN 17,883 Table 1. ObservedPhasesthatConstrainDeep-CrustalModel Instrument 16 Layer 5 C9 C4 5r, 6* 6 Interface 6 C3 17 A8 A1 5r, 6 6 6 6 6 6 6 6 Layer6 7 7 7 Interface 7 7 7 7 7r, 8 7r, 8 Interface 8t 8 8 Layer8•' 8 8 M M Layer7 19 20 21 5r, 6 6 7 7 6 6 7r, M 7r, M 7r, M 7r, M 7r, M 7r, M 7r, M 7r, M M M M M M M M M Interface9•' Layer 97 Moho *Phase 5r,refraction frombeneath interface 5;phase 6,reflection frominterface 6;M, Mohoreflection, etc. '•Also constrained bydataonOBH16,line802. 6.0 5.0 4.0 3.0 , I 2.0 , teat , 1.0 i n i I I I I I i I I I I I Range (kin) Figure 6. Detail of recordsectionfrom OBH 17, with travel timesof phasespredictedby the velocitymodel of Figure4a, labeledasin Figure5. 17,884 HOLBROOKET AL.' U.S. MID-ATLANTIC CONTINENTALMARGIN 7.0 OBS 6.0 M 5.0 4.0 3.0 • (31 (• o (• (31 (n o (n (• •J o •J (31 oo o 03 (31 Range (kin) Figure 7. Detailof recordsectionfromOBSAS, withtraveltimesof phases predicted by thevelocitymodelof Figure4a, labeledasin Figure5. 8.0 7.0 6.0 5.0 4.0 o (.n o (.n o (.n o (.n o Range (kin) Figure8. Detailof recordsection fromOBSA1, withtraveltimesof phases predicted by thevelocitymodelof Figure4a, labeledasin Figure5. HOLBROOK ET AL.' U.S. MID-ATLANTIC CONTINENTAL MARGIN I .... I .... I 17,885 The positionof this gravity gradientcoincideswith both the lateralincreasein basement velocity(Figure4a) andthe sharp gradient in magnetic intensity (Figure 14). The density model, which matches the observed data to within +10 mGal, showsthat the gravity anomalyis readily explainedby an 30- accomaoanying increase in basement density, from2650-2760 kg/m-• in continental crustto 2870kg/m 3 in transitional I .... I 50 .... I 100 ' 150 Distance (km) crust. Thus an abruptchangein basementcomposition occurs at modelkm 30-40. Densitieswithin the wedgeare modeled as2650-2920 kg/m 3, consistent withlaboratory-measured densitiesof many basalts[e.g., Christensenand Salisbury, 1973'Hyndman andDrury,1976].Thedensity of3130kg/m 3 in the lower transitional crust is similar to that inferred on othervolcanicmargins[e.g.,Holbrooket al., 1994]. I 11oo . I . I , I . I Discussion lOOO I ' I ' I ' I ' I Investigations of rifted margin structure often seek to locate the so-called "continent-ocean boundary." We do not believe this term accuratelyapplies to the U.S. Atlantic margin, since there is no single boundary that juxtaposescontinental and (normal) oceanic crust. Rather, continental and I . I 0.15 . I . I ............................ •h 0.10 7.0 7.2 7.4 7.6 30 7.8 100 Velocity (km/s) 150 200 Distance (km) Figure 9. Analysisof uncertainty in lower-crustal velocityandMoho depth in transitionalcrust. (a) RMS misfit of observedand modeled Moho reflection and lower-crustal refraction travel times as a function of lower-crustalseismicvelocity. (b) Numberof observations to which rays were successfullytracedfor variouslower-crustalvelocities. Best velocityis chosenas that velocitywhichminimizesRMS misfit while allowing rays to be tracedto a large numberof observations.RMS 450 I, I, I, I. I misfit suggests uncertainty of +0.2 km/sin lower-crustal velocity. (c) Depth to Moho for fixed velocitiesof 7.3, 7.4, 7.5, 7.6, and7.7 km/s in 1.,. .,.i transitional crust;boldline is Mohopositionfor bestvelocity(7.5 km/s). Variationaboutbestpositionsuggests uncertaintyof _+3km in Moho position. 0.15 Gravity Modeling Modeling of gravity data constitutesan importantcheck on the seismic model and enables estimation of crustal densities.We converted theseismicmodelof Figure4a intoa density model by assigningreasonabledensities [e.g., 6.6 6.8 7.0 7.2 7.4 Holbrook et al., 1992a] to layers in the velocity model. Layerswith stronglateralvelocitychanges(e.g., lowercrust) Velocity (km/s) were dividedinto severalbodiesto allow for corresponding changesin density,but no other changesto depthsof inter10. Analysisof uncertainty in lower-crustal velocityandMoho faceswereallowed. Densitieswerethenadjustediteratively Figure depthin oceaniccrust. (a) RMS misfit of observedand modeledMoho until a satisfactory matchbetweenthe predictedandobserved reflection and lower-crustal refraction travel times as a function of gravity data was found. The free air gravity data show a prominentpositive anomalyon the outercontinentalshelf (Figure 14). The seawardgradientof thisanomalyis largelycontrolledby thecontinentalslopeandpaleoshelfedge,but the landwardgradient requiresa relatively sharpdensitycontrastwithin basement. lower-crustalseismicvelocity. (b) Numberof observations to which rays were successfullytracedfor variouslower-crustalvelocities.RMS misfitsuggests uncertainty of +0.2 km/sin lower-crustal velocity. (c) Depthto Moho for fixed velocitiesof 6.8, 6.9, 7.0, 7.1, and7.2 km/s in oceaniccrust;bold line is Moho positionfor bestvelocity(7.0 km/s). Variationaboutbestpositionsuggests uncertainty of -+2km in Moho position. 17,886 HOLBROOK ET AL.: U.S. MID-ATLANTIC 10- 20- 50 100 150 CONTINENTAL MARGIN aboutmodel km 160, well landwardof the first appearanceon the MCS sectionof hyperbolic reflections typical of oceanic crust (model km 190, Figure 2b). Little is known about the timing of emplacementof the thick igneous crust. Only indirect inferences can be drawn, because(1) the $DW has not been drilled, and (2) the rift-drift transitionis poorly dateddue to the lack of seafloorspreading magnetic anomalies in the early Jurassic [Klitgord and Schouten,1986]. In the CarolinaTrough, Dillon et al. [1979] and Austin et al. [1990] correlated the J reflector, which marks Distance (kin) the top of the volcanic wedge, with the ClubhouseCrossroads diabase b I , I , I , I , I 0.0 :...i ....... i.... i....... i....... ' 0.05 Newark 6.0 6.2 6.4 6.6 6.8 Velocity (km/s) Figure11. Analysis of uncertainty in velocity withinanddepth tobase of the seaward-dippingwedge. (a) RMS misfit of observedand modeled reflection and refraction travel times as a function of seismic velocityin the wedge. RMS misfit suggests uncertainty of +0.3 km/sin lower-crustal velocity. (b) Depthto baseof wedgefor fixed velocities of 6.1, 6.2, 6.3, 6.4, and6.5 km/sin wedge;boldline is positionfor best velocity(6.3 km/s). Variationaboutbestpositionsuggests uncertainty of +_2km in baseof wedge. oceaniccrustare separatedby a 100-km-widezoneoccupiedby thick igneous crust that bears little resemblanceto normal oceaniccrust (Figure 15). The most pronouncedboundaryon the margin occursbetween continentaland transitionalcrust, where velocity, density, magnetization[Talwani et al., 1992], and reflectivity changedramatically in character. Continental crust is almost completely replaced by rift-related igneous rocks over a lateral distanceof only -20 km (Figure 15). The location of the edge of North America at the basementhinge zone, significantlylandwardof the ECMA, mustbe takeninto accountin predriftreconstructions of North AmericaandAfrica (i.e., the maximum closure pole of Klitgord and Schouten [1986] rather than the minimum closure pole shouldbe used). To our knowledge, the abruptnessof this boundary and the total lack of continental crust beneath the seaward-dipping wedge are not adequatelyaccountedfor in existingmodelsof rifted margin evolution [cf. Mutter, 1993]. Explainingthis profound geological boundary should be a primary goal of future margin studies. The boundarybetweenthick igneouscrustandoceaniccrust drilled in South Carolina and dated at 184 +3 Ma [Lanphere, 1983]. This interpretation,if correct, implies that the transition from thick igneouscrust to normal oceaniccrust generatedby seafloorspreadingoccurredaround184 M a on the Atlantic margin, a date consistentwith the maximum closure pole of Klitgord and Schouten [1986]. Earlier volcanism occurred in the Newark-type rift basins of the Atlantic coast during Hettangianto Sinemuriantime (208-197 Ma [de Boer et al., 1988]), at least 20 m.y. after the onsetof sedimentation [Manspeizer and Cousminer, 1988]. Igneous activity in the Basin was concentrated in a short time interval in earliest Jurassic (-200 Ma [Witte and Kent, 1991]). These dates bracket pre-seafloor spreadingvolcanic activity on the margin from -208-184 Ma, a period of about 24 m.y. However, the temporaloverlap betweenvolcanismin onshore rift basinsand in the offshorethick igneouscrust is unknown; the geodynamicmodel of ttarry and Sawyer [1992] suggests that rifting (and thereforevolcanism)may have migratedfrom the rift basinsto the nascentmargin. High-velocity lower crust on rifted margins is often referred to as "underplate"[e.g., Diebold et al., 1988; Tr•hu et al., 1989], a term that implies intrusionbeneathpreexisting continentalcrust [cf. Furlong and Fountain, 1986; Mooney et al., 1983]. However, on the U.S. Atlantic margin, this term is largely a misnomer, as most of the high-velocity lower crust underliesits extrusivecounterpartin the SDW, not continental crust. Although it is reasonableto supposethat underplating of continentalcrust by rift-related igneousrocks occurredduring breakup,there is no direct evidencein our data to constrain the amountof underplating. Resultsfrom line 802 (Figure 1) suggestthat if a high-velocityunderplateis presentbeneath continental crust, it is limited to about 3 km in thickness [Holbrook et al., 1992b]. There may be indirect evidenceof an underplatedzone, however, in the nonreflectivezone under the seawardtaper of reflective continentalcrust (Figure 15). Our study did not resolve velocity in this region (Figure 4b), so there is no corroboratingevidencefor underplatinghere. The seismicresultspresentedhere offer new insightsinto the long-debatedorigin of the East Coast magnetic anomaly (ECMA). The ECMA is spatially correlated with the thick igneouscrust(Figure 14), implyingthat the basaltsandunderlying intrusives source the anomaly [Talwani et al., 1992]. The correlation is not as simple on previousmargin transects [Austin et al., 1990; Holbrook et al., 1994; LASE Study is not well defined, but at the resolution of our seismic results Group, 1986; Trdbu et al., 1989]: in the CarolinaTrough the it is gradualrather than abrupt. Betweenmodel km 130 and peak of the ECMA is displacedseawardof the thickestbasalts, 160 (Figure 4a), velocity decreasesfrom 7.5 to 7.0 km/s, and beneatha secondset of seaward-dippingreflections[Oh et al. igneouscrustalthicknessdecreasesfrom 15 to 8 km. This 1991]. Despite these poorly understoodcomplexities,it is transitionimplies that a fundamentalchangein magmaticpro- clear that the ECMA residesbetweenthe thick seaward-dipping cesses(e.g., declining temperaturesor cessationof enhanced wedge and normal oceanic crust and is therefore sourcedby convection) occurred prior to establishment of normal mafic igneousrocks emplacedduringrifting. Thus the ECMA seafloorspreading. True oceaniccrust,definedby relatively does not mark an Appalachian suture, as has been suggested normal crustal thickness and velocity structure, begins at elsewhere [Hall, 1990]. HOLBROOK ET AL.' U.S. MID-A]2.,ANTIC CONTINENTAL MARGIN 17,887 8.0 '•, •.o I OBH •7, 6.0 $.0 Z,.O •.0 0.0 a o ! i i ! i i i i 0 0 8 (• .. 0.. _ ,,, • R •ng e (k•) 8.0 : 6.0 ..,' : OB$ A8 . '.,.. $.0 4.0 •.o •.0 •" 0.0 Rcrage 8.0 OBS' A 1 ß : 6.0 $.0 4.0 c 0.0 i i i i i ! i i i Figure 12. Syntheticseismograms, calculated with asymptotic ray theory,for (a) OBH 17, (b) OBS A8, (c) OBS A1, and(d) OBH 20. i 17,888 HOLBROOK ET AL.' U.S. MID-ATLANTIC CONTINENTAL MARGIN &o 7.0 8.0 7'?' 5.0 4.0 &O œ.0 d Figure 12. (continued) crustal velocity fom 7.3 km/s in thick igneous crust to 6.9 km/s in oceanic crust are difficult to reconcile with plume models [e.g., White and McKenzie, 1989]. This suggeststhat nonplume processes,such as secondaryconvection [Mutter et [Skogseidet al., 1992]). From similar crustal structureon the al., 1988], broad, hot zonesin upper manfie [Andersonet al., LASE, USGS line 32, and Carolina Trough BA-6 transects 1992], or supercontinent insulation effects [Gurnis, 1988] [Austin et al., 1990; Holbrook et al., 1994; LASE Study were responsible for the anomalous thickness of igneous Our study clearly demonstratesthat the U.S. Mid-Atlantic m.arginis strongly volcanic, with a thicknessof rift-related igneousrocks (25 km) that rivals that found on marginsof the North Atlantic (e.g., 18.5 km [Morgan et al., 1989]; 23 km Group, 1986; Trdhu et al., 1989] and from the first-order correlation of thick igneous crust with the East Coast magnetic anomaly, Holbrook and Kelemen [1993] inferred that anomalous volcanism extends along the entire U.S. Atlantic margin and therefore comprisesa large igneousprovince. Moreover, the lack of independentevidence for a hotspot in the central rocks [Holbrook and Kelemen, 1993]. Conclusions The EDGE Mid-Atlantic seismic experiment provides the clearest picture to date of the transition from continental to Atlantic, the asymmetricdistributionof igneousrocks along oceanic crust on the U.S. Atlantic margin. The velocity and and acrossthe margin, and the rapid lateral decreasein average reflection images show that a thicknessof up to 25 km of 45 SP 1475•' '1875 • 0 2275 2675 3075 3475 3875 4275 4675 5075 5475 5875 6275 40 60 80 100 120 140 160 180 200 220 240 20 Distance (km) Figure 13. Comparison of velocitymodeland prominentsubsedimentary MCS reflections.Shortwhite and blacklmes represent truepositionsof MCS reflectionshighlightedin Figure2c, determined by ray tracemodeling. Note goodagreement of seaward-dipping reflectorswith layer 5 and 6 of velocity model and restrictionof lower-crustalreflectivity to rifted continental crust and Moho. HOLBROOK ET AL.' U.S. MID-ATLANTIC CONTINENTAL MARGIN 17,889 300 - 600 200observed grav - 400 lOO 0 - 200 I"...:c..al•!..ated grav obj;rv;•t'•;t•t• ...... '...... I I I I I I I I I I I 1800 2650 '• 10 2850 3030 2920 30 3130 328O 40 0 20 40 60 80 100 120 140 160 180 200 220 240 Distance (kin) Figure 14. Gravity model along line801.Numbers aredensities inkilograms percubic meter. •andward gradient of prominentpositiveanomalyis dueto highdensitieswithintransitional crust(2850, 2870, and2920 kg/m-'); seawardgradientis due to continentalslopeandbasementstructure. rifled •' conttnental• crust ., ..... '• [FllCK tgneous crust • •. -. ß oceantc crust ..• .- lO 2o • 25 3o 35 0"'"'•:• cumulates/sills 45 0 20 40 60 80 1• 120 140 160 180 2• 220 Distance (km) Figure 15. Geologicalinterpretation of line 801, basedon vertical-incidence reflections,seismicvelocity,and density. Thick igneouscrust,consisting of basaltin the seaward-dipping wedge.andunderlyingmaficintrusives,occupiesa 100-km-widezone betweenrifted continentalcrustand oceaniccrust. Inferred lateral increasein MgO contentin lower crustreflectsthe lateral increasein velocityfrom 6.7 to 7.5 km/s(Figure4a). Underplate beneathcontinental crustat modelkm 25-40 is inferredfrom lack of continental-typereflections;velocity is not well constrainedin this zone and thereforecannotcorroboratethis interpretation.The boundarybetweenthick igneouscrustand oceaniccrustis obscureand probablygradationalratherthan abrupt. • 240 17,890 HOLBROOKET AL.: U.S. MID-ATLANTIC CONTINENTALMARGIN igneousmaterial was accretedonto the margin during continentalbreakup. The detailsof the seismicstructureand lack of independentevidence for a hotspot in the central Atlantic imply that a long-livedmantleplume was not responsiblefor the igneousmaterial. There is a profoundgeologicalboundarybeneaththe continental shelf between Appalachian crust and thick igneous crust: over a distanceof only 20 km, continentalcrustgives way to basaltsof the seaward-dipping wedgeand their intrusive underpinnings.Thus the mostimportantboundaryon the margin is markednot by the peak of the ECMA but by strong magneticandgravitygradients just landwardof the SDW. The disappearance of intracrustalreflectivity acrossthis boundary, where velocity increasesfrom 5.9 to 7.0+ km/s, suggeststhat the reflectivityis disruptedby massiveintrusionand that very little continentalcrustpersistsseawardof the reflectivecrust. Furlong, K.P., and D.M. Fountain,Continentalcrustalunderplating: Thermal considerationsand seismic-petrologicconsequences,J. Geophys.Res.,91, 8285-8294, 1986. Grow, J.A., K.D. Klitgord, and J.S. Schlee,Structureand evolutionof BaltimoreCanyontrough,in The Geologyof North America,vol. I-2, The Atlantic ContinentalMargin, U.S., editedby R. E. Sheridanand J.A. Grow, pp. 269-290, GeologicalSocietyof America,Boulder, Colo., 1988. Gumis, M., Large-scalemantle convectionand the aggregationand dispersal of supercontinents, Nature,332,695-699, 1988. Hall, D.J., Gulf Coast--East Coast magneticanomaly,I, Root of the main crustal decollementfor the Appalachian-Ouachitaorogen, Geology,18, 862-865, 1990. Harry, D.L., and D.S. Sawyer, A dynamic model of lithospheric extensionin the BaltimoreCanyonTroughregion,Tectonics,11,420436, 1992. Hinz, K., A hypothesis on terrestrialcatastrophes: Wedgesof very thick oceanward dippinglayersbeneathpassivemargins,Geol.Jahrb.,22, 5-28, 1981. Holbrook,W.S., andP.B. Kelemen,Largeigneousprovinceon the U.S. Atlanticmargin and implicationsfor magmatismduringcontinental Acknowledgments. We thankthe captainsandcrewsof theM/V Geco breakup,Nature,364,433-436, 1993. Searcher andR/V Endeavor for the professionalism and cooperation Holbrook,W.S., and E.C. Reiter, Suppression of previousshotnoiseon marinewide-angleseismicdata,Geophysics, in press,1994. thatenabledthe logisticallycomplexcoordination of the OBH andMCS data collection. Recordingof shotinstantswas carriedout by John Holbrook, W.S., W.D. Mooney, and N.I. Christensen,The seismic velocity structureof the lower continental crust, in The Lower Collinsusingsoftwareprovidedby Y. NakamuraandM. Weidersphan ContinentalCrust, editedby D. M. Fountain,R. Arculusand R. Kay, (Universityof Texas)and with the technicalassistance of GECO party pp. 1-43, Elsevier,New York, 1992a. chiefTorsteinSeglem. We thankP.B. Kelemen,K.D. Klitgord,andD. Holbrook,W.S., G.M. Purdy,J.A. Collins,R.E. Sheridan,D.L. Musser,I. Lizarraldefor stimulatingdiscussions, G. Kent and D. Lizarraldefor L. Glover, M. Talwani, J.I. Ewing, R. Hawman, and S. Smithson, software assistance,and D. Sawyer and J. Diebold for reviews. Deepvelocitystructureof rifted continentalcrust,U.S. Mid-Atlantic Acquisitionand analysisof the datapresented hereweresupported by margin,from EDGE wide-anglereflection/refraction data,Geophys. National Science Foundationgrants ES8721194, OCE8917628, and Res.Lett., 19, 1699-1702, 1992b. Holbrook, W.S., E.C. Reiter, G.M. Purdy, and M.N. Toks6z, Image of OCE8917599andby fundsfrom theTexacoOil Company.WoodsHole the Moho across the continent-ocean transition, U.S. east coast, Oceanographic Institutioncontribution 8713. Geology,20, 203-206, 1992c. Holbrook,W.S., E.C. Reiter, G.M. Purdy, D. Sawyer, P.L. Stoffa, J.A. AustinJr., J. Oh, and J. Makris, Deep structureof the U.S. Atlantic continentalmargin,offshoreSouthCarolina,from coincidentoceanReferences bouom and multichannelseismicdata, J. Geophys.Res.,99, 91559178, 1994. Alsop, L.E., and M. Talwani, The East Coast magnetic anomaly, Science, 226, 1189-1191, 1984. Hopper,J.R., J.C. Mutter, R.L. Larson,C.Z. Mutter, and Northwest Anderson,D.L., Y.-S. Zhang,andT. Tanimoto,Plumeheads,continental Australia Study Group, Magmatism and rift margin evolution: lithosphere, flood basalts and tomography, in Magmatism at Evidencefrom northwestAustralia,Geology,20, 853-857, 1992. ContinentalMarginsandthe Causesfor ContinentalBreak-up,edited Horsefield, S.J., R.B. Whitmarsh, R.S. White, and J.-C. Sibuet, Crustal by B. Storey, T. Alabaster,and R. J. Pankhurst,Geol. Soc. Spec. structure of the GobanSpurpassivecontinental margin,NE AtlanticPubl., London, 68, 99-124, 1992. Resultsof a detailed seismicrefraction survey, Geophys.J. lnt., in Austin, J.A., Jr., P.L. Stoffa, J.D. Phillips, J. Oh, D.S. Sawyer,G.M. press,1994. Purdy, E. Reiter, and J. Makris, Crustal structureof the Southeast Hutchinson,D.R., J.A. Grow, K.D. Klitgord, and B.A. Swift, Deep Georgia embayment-Carolina Trough: Preliminary results of a structure and evolution of the Carolina Trough, in Studies in compositeseismicimageof a continentalsuture(?) and a volcanic Continental MarginGeology,editedby J. S. WatkinsandC. L. Drake, AAPG Mem., 34, 129-152, 1983. passivemargin,Geology,18, 1023-1027, 1990. Christensen,N.I., and M.H. Salisbury, Velocities, elastic moduli and Hyndman,R.D., and M.J. Drury, The physicalpropertiesof oceanic weathering-age relationsfor Pacificlayer 2 basalts,Earth Planet. Sci. basementrocksfrom deep drilling on the Mid-Atlantic ridge,J. Lett., 19, 461-470, 1973. Geophys.Res.,81, 4042-4052, 1976. Claerbout, J.F., Imaging the Earth's interior, 398 pp., Blackwell Keen, M.J., Possibleedgeeffectto explainmagneticanomaliesoff the Scientific, Boston, Mass., 1985. de Boer, J.Z., J.G. McHone, J.H. Puffer, P.C. Ragland, and D. Whiuington,MesozoicandCenozoicmagmatism,in The Geologyof North America,vol. I-2, The Atlantic ContinentalMargin, U.S., edited by R. E. SheridanandJ.A. Grow, pp. 217-241, GeologicalSocietyof eastemseaboardof the U.S., Nature, 222, 72-74, 1969. Klitgord,K.D., and J.C. Behrendt,Basinstructureof the U.S. Atlantic margin,in Geologicaland GeophysicalInvestigationsof Continental Margins,AAPGMere.,29, 85-112, 1979. Klitgord,K.D., andH. Schouten, Platekinematicsof the centralAtlantic, America, Boulder, Colo., 1988. in The Geologyof North America,vol. M, The WesternNorthAtlantic Diebold, J.B., P.L. Stoffa, and LASE Study Group, A large aperture Region, edited by P. R. Vogt and B.E. Tucholke, pp. 351-378, seismicexperimentin the BaltimoreCanyonTrough,in The Geology GeologicalSocietyof America,Boulder,Colo., 1986. of North America,vol. I-2, The AtlanticContinentalMargin, U.S., Klitgord, K.D., D.R. Hutchinson, and H. Schouten, U.S. Atlantic continental margin; Structural and tectonic framework, in The editedby R. E. Sheridanand J. A. Grow, pp. 387-398, Geological Societyof America,Boulder,Colo., 1988. Geologyof North America,vol. I-2, TheAtlanticContinentalMargin, Dillon, W.P., C.K. Paull, R.T. Bufflet, and J.-P. Fail, Structure and U.S., editedby R. E. SheridanandJ.A. Grow, pp. 19-55, Geological Societyof America,Boulder,Colo., 1988. development of the southeast Georgiaembayment andnorthernBlake Plateau: Preliminary analysis, in Geological and Geophysical Koelsch, D.E., K.R. Peal, and G.M. Purdy, The Woods Hole Investigationsof Continental Margins, edited by J. S. Watkins L. OceanographicInstitution digital ocean bottom hydrophone Montadert, and P.W. Dickerson,AAPG Memoir 29, 27--41, 1979. instrument,Rep. WH01-82-30, Woods Hole Oceanogr.Inst., Woods Drake, C.L., M. Ewing, and G.H. Sutton, Continental margins and Hole, Mass., 194802. 39 At/ Ar agesof basaltfrom ClubhouseCrossroads geosynclines: The east coast of North America north of Cape Lanphere,M.A., test hole 2, near Charleston, South Carolina, in Studies related to the Hatteras,Phys.Chem.Earth, 3, 110-198, 1959. Charleston,South Carolina earthquakeof 1886 -- Tectonicsand Emery,K.O., E. Uchupi,J.D. Phillips,C.O. Brown,E.T. Bunce,andS.T. Knott, Continental rise of eastern North America, Am. Assoc. Petr. seismicity,editedby G.S. Gohn, U.S. Geol. Surv.Prof. Pap., 1313, Geol. Bull., 54, 44-108, 1970. B1-B8, 1983. HOLBROOK ET AL.: U.S. MID-ATLANTIC LASE Study Group, Deep structureof the U.S. East Coast passive margin from large apertureseismicexperiments(LASE), Mar. Pet. CONTINENTAL MARGIN 17,891 Continental Break-up,editedby B. Storey,T. Alabasterand R. $. Pankhurst,Geol. Soc.Spec.Publ. London,68, 305-320, 1992. Sleep, N.H., Thermal effects of the formationof Atlantic continental Manspeizer,W., and H.L. Cousminer,Late Triassic-EarlyJurassic marginsby continentalbreakup, Geophys.J. R. Astron.Soc.,24, 324, 1971. synrift basinsof the U.S. Atlantic margin,in The Geologyof North America,vol. I-2, The Atlantic ContinentalMargin, U.S., editedby R. Talwani, M., J. Ewing, R.E. Sheridan, D.L. Musser, L. Glover III, S. E. Sheridan and $.A. Grow, pp. 197-216, Geological Society of Holbrook, and M. Purdy, EDGE lines off the U.S. Mid-Atlantic America, Boulder, Colo., 1988. marginand the EastCoastMagnetic Anomaly (abstract),EOS, Trans. McKenzie, D., Some remarks on the developmentof sedimentary AGU, 73(14), Fall Meetingsuppl.,490-491, 1992. basins,Earth Planet. Sci. Lett., 40, 25-32, 1978. Trthu, A.M., A. Ballard, L.M. Dorman,J.F. Gettrust,K.D. Klitgord, and A. Schreiner, Structure of the lower crust beneath the Carolina Mooney, W.D., M.C. Andrews, A. Ginzburg, D.A. Peters,and R.M. Hamilton, Crustal structureof the northernMississippiembayment Trough, U.S. Atlantic margin,J. Geophys.Res.,94, 10,585-10,600, 1989. and a comparisonwith othercontinentalrift zones,Tectonophysics, 94, 327-348, 1983. Vogt, P.R., Early events in the opening of the North Atlantic, in Morgan,$.V., P.S.Barton,and R.S. White, The Hatton Bank continental Implicationsof ContinentalDrift to the Earth Sciences,editedby D. H. margin,IlI, Structurefrom wide-angleOBS andmultichannelseismic Tarling and S. K. Runcorn, pp. 693-712, Academic, San Diego, Cald., 1973. refractionprofiles,Geophys. J. lnt., 98, 367-384, 1989. Mutter, J.C., Marginsdeclassflied, Nature, 364, 393-394, 1993. White, R., and D. McKenzie, Magmatismat rift zones: The generation Mutter, J.C., M. Talwani, and P.L. Stoffa, Evidence for a thick oceanic of volcaniccontinentalmarginsand flood basalts,J. Geophys.Res., 94, 7685-7729, 1989. crustadjacentto the Norwegian margin,J. Geophys.Res., 89, 483502, 1984. White, R.S., G.D. Spence, S.R. Fowler, D.P. McKenzie, G.K. Westbrook, and A.N. Bowen, Magmatism at rifted continental Mutter,J.C., W.R. Buck,andC.M. Zehnder,Convectivepartialmelting, margins,Nature, 330, 439-444, 1987. 1, A model for the formationof thick basalticsequencesduring the Whitmarsh, R.B., P.R. Miles, and A. Mauffret, The ocean-continent initiationof spreading, J. Geophys.Res.,93, 1031-1048, 1988. boundaryoff the western continentalmargin of Iberia, I, Crustal Nakamura,Y., P.L. Donoho,P.H. Roper, and P.M. McPherson,Largestructureat 40ø30'N,Geophys.J. lnt., 103, 509-531, 1990. offset seismic surveyingusing ocean-bottomseismographsand air Witte, W.K., and D.V. Kent, Tectonicimplicationsof a remagnetization guns: Instrumentationand field technique,Geophysics,52, 16011611, 1987. event in the Newark Basin, J. Geophys.Res., 96, 19,569-19,582, 1991. Oh, J., J.D. Phillips,J.A. AustinJr., and P.L. Stoffa, Deep-penetration seismicreflectionimagesacrossthe United Statescontinentalmargin, Zelt, C.A., and R.B. Smith, Seismic traveltime inversion for 2-D crustal velocitystructure,Geophys.J. lnt., 108, 16-34, 1992. in Continental lithosphere: Deep seismicreflections,Geodyn.Ser., vol. 22, edited by R. Meissner,L. Brown, H.-J. D•irbaum,W. Franke, K. Fuchs,andF. Seffert,pp. 225-240, AGU, Washington, D.C., 1991. J. Ewing, W.S. Holbrook, and G.M. Purdy,Departmentof Geology Pinheiro, L.M., R.B. Whitmarsh, and P.R. Miles, The ocean-continent WoodsHole Oceanographic Institution,WoodsHole, boundaryoff the westerncontinentalmarginof Iberia, Part II, Crustal andGeophysics, structurein the TagusAbyssalPlain, Geophys.J. lnt., 109, 106-124, MA 02543. 1992. L. Glover III, Department of Geological Sciences, Virginia Pratt,T.L., J.K. Costain,C. Coruh,andI. L. Glover,A geophysical study PolytechnicUniversity,Blacksburg,VA 24061-0420. of the Earth'scrustin centralVirginia: Implicationsfor Appalachian D. Hutchinson,U.S. Geological Survey, Branch of Atlantic Marine crustalstructure,J. Geophys.Res.,93, 6649-6667, 1988. Geology,WoodsHole, MA 02543. Sheridan,R.E., D.L. Musser,L. Glover IIl, M. Talwani, J. Ewing, W.S. R.E. Sheridan, Department of Geological Sciences, Rutgers Holbrook,G.M. Purdy,R. Hawman,and S. Smithson,Deep seismic University, New Brunswick,NJ 08903. reflection data of EDGE U.S. Mid-Atlantic continental margin Geol., 3, 234-242, 1986. experiment: Implications for Appalachian suturesand Mesozoic riftingandmagmaticunderplating, Geology,21,563-567, 1993. Skogseid,J., T. Pedersen,O. Eldholm,and B.T. Larsen,Tectonismand magmatismduring NE Atlantic continentalbreak-up: The Voring Margin, in Magmatism at ContinentalMargins and the Causesfor M. Talwani, Houston Advanced Research Center, 4802 Research ForestDrive, The Woodlands,TX 77381. (Received October 4, 1993; revised March 14, 1994; acceptedMarch 17, 1994.)