Contents lists available atScienceDirect
Earth and Planetary Science Letters
www.elsevier.com/locate/epsl
Deep crustal structure of the Adare and Northern Basins, Ross Sea, Antarctica, from sonobuoy data
M.M. Selvans
a,b,∗, J.M. Stock
a, R.W. Clayton
a, S. Cande
c, R. Granot
daSeismologicalLaboratory,CaliforniaInstituteofTechnology,1200E.CaliforniaBlvd.,MC252-21,Pasadena,CA91125,UnitedStates
bCenterforEarthandPlanetaryStudies,NationalAirandSpaceMuseum,SmithsonianInstitution,4thSt.SWandIndependenceAve.,MRC315,Washington,DC 20013,UnitedStates
cScrippsInstitutionofOceanography,MC0220,LaJolla,CA92093,UnitedStates
dDepartmentofGeologicalandEnvironmentalSciences,Ben-GurionUniversityoftheNegev,BeerSheva,Israel
a r t i c l e i n f o a b s t ra c t
Articlehistory:
Received8March2014
Receivedinrevisedform21August2014 Accepted25August2014
Availableonline20September2014 Editor: P.Shearer
Keywords:
WestAntarcticRiftSystem marineseismicdata crustalstructure sonobuoy AdareBasin NorthernBasin
ExtensionassociatedwithultraslowseafloorspreadingwithintheAdareBasin,inoceaniccrustjustnorth ofthe continentalshelfintheRossSea, Antarctica,extendedsouthintotheNorthern Basin.Magnetic and gravityanomalydatasuggestcontinuityofcrustalstructureacrossthecontinentalshelfbreakthat separatesthe AdareandNorthern Basins.Weuse sonobuoyrefractiondata andmulti-channel seismic (MCS)reflectiondata collectedduringresearchcruiseNBP0701,including71 newsonobuoyrecords,to provideconstraintsoncrustalstructureintheAdareandNorthernBasins.Adjacent1Dsonobuoyprofiles alongseveralMCSlinesrevealdeepcrustal structureinthevicinityofthecontinentalshelfbreak, and agreewithadditional sonobuoydatathatdocument fastcrustal velocities(6000–8000 m/s)atshallow depths(1–6 kmbelow sealevel)fromtheAdareBasintothecontinental shelf,astructureconsistent withthatofotherultraslow-spreadcrust.Ourdeterminationofcrustal structureintheNorthernBasin onlyextendsthroughsedimentaryrocktothebasementrock,andsocannothelptodistinguishbetween differenthypothesesforformationofthebasin.
©2014ElsevierB.V.All rights reserved.
1. Introduction
Deformation processes by which extensional strain at mid- ocean ridges, inrelatively thinand dense oceaniccrust, istrans- ferredtoneighboringcontinentalcrustarepoorlyunderstood.The breakupofGondwanaduring Cretaceoustime producedtheWest AntarcticRiftSystem(WARS)within continentalcrust (e.g.,Elliot, 1992), whichmakes upthemajorityofthecrust intheRossSea.
TheWARSistectonicallycomplexanddifficulttostudyduetobe- inglargelycoveredbyanicesheetandseaice,andbecausemuch oftheriftingtookplacewithincontinentalcrustatextremelyslow rates,making ittheleastwell constrainedlink intheglobalplate circuit.
During mid-Cenozoic time, seafloorspreading occurredfor 17 million years(43–26 Ma)attheAdare Basin(Candeetal., 2000), producingoceaniccrustinthenorthwesternmostRossSea(Fig. 1).
The Adare spreading center lies in the deep water of the Adare
*
Correspondingauthorat:CenterforEarthandPlanetaryStudies,NationalAir andSpace Museum,SmithsonianInstitution, 4thSt.andIndependenceAve. SW, MRC315,Washington,DC20024,UnitedStates.Tel.:+12026332476.E-mailaddress:[email protected](M.M. Selvans).
BasinandtrendssouthwardtowardtheNorthernBasin,whichisa structurally-definedfeatureonthecontinentalshelf(e.g.,Cooperet al.,1995).Magneticanomaliesareusedtoconstrainthetimingand spreadingrateoftheAdarespreadingcenter(Candeetal., 2000);
two of these magnetic anomalies are continuous with similarly strong andnarrowmagnetic features that extendinto theneigh- boringcontinentalshelfoftheNorthernBasin(Granotetal.,2013;
Damaskeetal.,2007; CandeandStock,2006).Multi-channelseis- mic (Granot et al., 2010; Müller et al., 2005; Brancolini et al., 1995),sonobuoy(Cooperetal.,1987; HoutzandDavey,1973),and DeepSeaDrillingProject(Hayesetal.,1975a, 1975b)studieshave characterized the shallow crustal structure at a few locations in thevicinity.
Inthispaperwepresenttheanalysisofsonobuoydatacollected during research cruise NBP0701 on board the R/VIBNathanielB.
Palmer (Adare BasinSonobuoy Data,2007; NBP0701Data Report, 2007), inthe Adare andNorthernBasins (shown withrespect to bathymetry, magneticanomalies,andfree-airgravityanomaliesin Fig. 2a–c).ThisisacompanionpapertotheanalysisbyGranotet al. (2010)ofthemulti-channelseismic(MCS)reflectiondatafrom the same cruise. While MCS data are usefulfor determining the velocity of the shallowest rock layer, sonobuoy data allow us to directlymeasurelayervelocitiesatdepth(seeSelvansetal.,2012).
http://dx.doi.org/10.1016/j.epsl.2014.08.029 0012-821X/©2014ElsevierB.V.All rights reserved.
Fig. 1.BathymetryoftheRossSeaandSouthernOceanshowsacleardelineation ofthecontinentalshelf. Ourstudyarea(blackbox) liesatoneendoftheWest AntarcticRiftSystem,inthenorthwestern-mostRossSea.Totheeastofourstudy area,thecontinentalcrustofIselinBank(IB)jutsoutintooceaniccrust.Sonobuoy datafrompreviousstudies(blackdots)arediscussedinthetext.Keyfeaturesare labeled:AdareBasin(AB),AdareTrough(AT),NorthernBasin(NB),VictoriaLand Basin(VLB),SouthernCentralBasin(SCB),EasternBasin(EB).ShuttleRadarTopogra- phyMissiondataandestimatedseafloorbathymetryare∼1 km resolution(Becker andSandwell,2006).InsetisthelocationofthisfigurewithrespecttoAntarctica.
The 71sonobuoy profiles presented here, deployed along the 19 MCSlines,provideadeeperandmoredetailedlookatthecrustin theAdareandNorthernBasinsthan waspossibleinprior studies ofthisregion.
1.1.TectonichistoryofthenorthwesternRossSea
During the final stages of breakup of Gondwana in mid- Cretaceous time (∼100 Ma), the Ross Sea margin of Antarctica riftedfromthe CampbellPlateauandthe SouthTasmanRise, the formerborderingMarie ByrdLand (West Antarctica)andmostof the Ross Sea continental margin, and the latter adjacent to the westernmost Ross Sea and Cape Adare of East Antarctica (e.g., Weaver etal., 1994). Tectonic reconstructions (e.g., Cande et al., 1995)andcrustalstructure(Cooperetal.,1987) suggestthatIselin Bank is continental material that may have been part of East Antarctica(seeFig. 1).
WhileGondwanawas breakingup, riftingbegan betweenEast andWest Antarcticaalong theWARS.At ∼90 Ma, theSouth Tas- manRisewas closeto EastAntarcticaandtheIselinBank (Gaina etal., 1998). Since the mid-Cretaceous, 500–1000 km ofbroadly distributedtranstensional motion hasoccurred within the WARS (Luyendyketal.,1996).ExtensionbetweentheSouthTasmanRise andEastAntarcticabegan∼60 Ma,riftingslowly,while∼300 km ofextensionbetweenEastandWestAntarctica occurredbetween 80 and 40 Ma (Cande et al., 2000; Molnar et al., 1975). Ma- rine sedimentary rocks from the eastern Ross Sea are Eocene in age (57–35 Ma), consistent with the formation of sediment-
accumulating basinsdue to extension during thistime (Truswell andDrewry,1984).
Additionally, in the western Ross Sea north–south trending basinsandridgesformedduring periodsofextension inlateCre- taceousandearly Cenozoictime (Cooperetal., 1987; Decesariet al., 2007). In particular,at ∼60 Ma the IselinBank was adjacent to Cape Adare, whereas by ∼27 Ma it hadmoved to its current position,accommodatingextensionnorthoftheRossSea,creating the AdareBasinandsignificantly deformingthe continentalcrust at its southern edge, where the Northern Basinnow lies (Cande andStock,2004).
The western Ross Sea has also experienced distributed vol- canic and tectonic deformation since mid-Cenozoic time. Exten- sionandsubsidenceformedtheVictoriaLandandNorthernBasins (CooperandDavey,1985; Daveyetal.,2006).Extensionwas con- centrated in theAdare Basinwhen seafloor spreadingtook place from43to 26 Ma(Candeetal., 2000). Recentminorextensional episodes at ∼24 Ma and ∼17 Ma have resulted in the forma- tion of the Adare Trough (Granot et al., 2010). Finally, ongoing deformation has disrupted the sedimentary rocks that were de- posited on top of the Adare Basin oceanic crust (Granot et al., 2010).YoungPliocenevolcanicknollsonthecontinentalshelfand in the deep water offshore from Cape Adare, which are simi- lar in composition to other volcanoes in the western Ross Sea (PanterandCastillo,2007),postdateformationoftheAdareBasin oceaniccrust(Granotetal.,2010; Panter andCastillo,2007).Fea- turesinmagnetic,gravity,andMCSdataarenot offsetacrossthe boundarybetweentheAdareandNorthernBasins,indicatingthat these basinsare structurallycontinuous (Cande andStock, 2006;
Damaskeetal.,2007).
1.2. Multi-channelseismicdataintheAdareandNorthernBasins
MCSdatacollectedduringresearchcruiseNBP0701(seeFig. 2) haveresolvableprimaryreflectionsupto2.2 s of(two-way)travel time in the continental shelf (Granotet al., 2010). Several sedi- mentaryrockhorizonscanbecontinuously tracedthroughoutthe region.The thicknessofsedimentary covergenerallyranges from 0 s in traveltime alongtheridgesboundingtheAdareTroughto 2.0s atonepointwithintheAdareTrough,andreachtheirthick- est(>2.2 s)intheNorthernBasinandthemarginslope(Granotet al.,2010).
1.3. PreviousseismicrefractionstudiesintheRossSea
Seismicrefractiondatacollected intheRossSeaprovideclues to thedeep crustal structure ofthe basinsandridgestherein. In thecentralandsouthernRossSea(neartheRossIceShelf),ocean bottom seismometers were used to obtain the deepest crustal structurestudiesintheregion.Mantlevelocities(>8.0 km/s)were detectedataminimumdepthof16kmbelowtheseafloor(Treyet al.,1999),andbasementrockwasdetectedasthinas∼5 km,with
∼7 kmof overlyingsediment andan underlying layerof ∼7 km ofinferredmagmaticintrusions(Tréhuetal.,1993).However,only sonobuoyshavebeendeployednearourstudyarea.
Ten sonobuoys deployed in shallow water just south of our studyarea(seelocationsinFig. 1)detectedmaximumsedimentary rockvelocities>4.0 km/s,basementvelocitiesof5.0–5.8 km/s,and assumed basement depths of 1.0–1.6 km subseafloor (Houtz and Davey,1973). Sixsonobuoysin deepwateralong the continental shelfbetweentheNorthernBasinandtheIselinBank,justeastof ourstudyarea,revealedmaximumsedimentaryrockvelocitiesup to4.3 km/s,assumedbasementvelocitiesupto4.7 km/s,andbase- mentdepthsofupto3.0 kmsubseafloor(HoutzandDavey,1973).
However, a similar study nearby (Cooper et al., 1987) indicated that many velocities >4.9 km/s are in fact layered sedimentary
Fig. 2.Weinvestigatecrustalstructureusingmulti-channelseismic(MCS)andsonobuoydatafromresearchcruiseNBP0701.(a)TheAdareBasinliesindeepwaternorthof thecontinentalshelfbreak,whereastheNorthernBasin(withinthesinuousblacklines,afterCandeetal.,2000)extendsfromtheshelfbreakto73.5◦S.LocationsofMCS lines(coloredlines,labeledatthebeginningoftheline)(NBP0701DataReport,2007) andsonobuoys(blackdots)aredisplayedontopofmultibeambathymetry(AdareBasin SonobuoyData,2007).SeeSection3.2ofthetextforadditionalclarificationofMCSlinelocationsonthecontinentalshelf;sonobuoysaredepictedsequentiallyfromthestart ofeachMCSline,followingtheexampleoflineL9.DeepSeaDrillingProjectsite(DSDP)274(greendot)iscollocatedwithsonobuoyL9S1.(b)Compiledshipboardmagnetic (Candeetal.,2000; CandeandStock,2006) andaeromagdata(Damaskeetal.,2007) showthatmagneticanomaliesarecontinuousacrossthecontinentalshelf(Granotetal., 2013).(c)Thereisnocorrespondencebetweenbouguergravityanomaliesandthecontinentalshelfbreak;gravitydata(SandwellandSmith,2009) arecorrectedforwater depthandbasement-sedimentdepthofDaveyandBrancolini (1995)(whichisknownonlyasfarnorthas71.0◦S).MCSline,transect,andsonobuoylocationsareoverlainin grayandblackin(b)and(c);1500mwaterdepthisshownasawhitecontour.
rockunits andnot igneous-metamorphicbasement, sothe above basementdepthsmaybeanunderestimate.
Additionalsonobuoysdeployed overtheIselinBankandinthe deep water to its east and west indicate that the Iselin Bank is composed ofcontinental crust, withoceanic crust ∼10 kmthick to its east (Cooper et al., 1987). West of Iselin Bank, just east ofour studyarea,a velocity of5.2 km/swas detected at4.3 km subseafloordepth,andsimilarvelocitieswere detectedat1.8and 2.8 km subseafloordepth in theAdare Basin(at 70.7◦S,175.3◦E, and71.3◦S,175.0◦E,respectively)(Cooperetal.,1987).
InthesouthernRossSea,sonobuoysrevealedarangeofvelocity gradientssimilartothosefoundinsedimentaryrocksfurthernorth (Cochrane etal., 1995; Davey etal., 1983, 1982); RossSea sedi- mentaryrockvelocitiesare systematicallyfasterthanthosefound at similar depths in the Gulf of Alaska (e.g., Bruns andCarlson, 1987) andthe Gulf ofMexico (e.g.,Gardner etal., 1974). Thisis interpretedastheresultofcompactionofRossSeasedimentsdue to past loading by a grounded ice sheet (Cochrane et al., 1995).
Similarly fastsedimentary rocklayer velocitiesat shallowdepths inthenorthernSvalbardmarginarealsoattributedtoiceloading (GeisslerandJokat,2004).
2. Sonobuoyanalysismethods
MCSdata were recorded witha 1.2 km, 48-channel streamer.
Lines1–12,indeeperwater,useda6-gunG/Isourcearraywitha
total capacityof20.6 l.ForseismicLines 13–19, onandnearthe continentalshelf,weuseda6-elementBolt-gunarraywithatotal capacity of 34.8 l. The typical source spacing was approximately 40 m. Sonobuoys were deployed occasionallyin deep water,and with a regular spacing of∼15 km in shallowwater (see Fig. 2).
Most sonobuoys provided data until they were 20–30 km away from theship. We useMCS datato constrain thevelocity of the shallowest rock layer, and sonobuoy data to reveal velocity pro- files deep into the oceanic crust. We assume planar layers for all velocity horizons, consistent withthe linearsegmented travel time plots and the generally flat horizons observed in the MCS data.
We confirm our 1D velocity models for each sonobuoy pro- file through finite difference modeling and depth migration of the data (Selvans et al., 2012). Smoothed andinterpolated adja- cent sonobuoy profiles along MCS Lines 13–19 allow us to also construct 2D velocity models, as do the more widely spaced sonobuoysalongTransect3intheAdareBasin(‘transects’arelines along which we analyze trends in crustal structure,even though they do not coincide with lines of MCS data,shown in black in Fig. 2).WeuseaDelaunaytriangulationalgorithmfortheinterpo- lation,withlocation(depthbelowseaflooranddistancealongthe line)andvelocityasvertices(Selvans et al., 2012).
3. Resultsof1Dand2Dvelocitymodels 3.1.DeepstructureintheAdareBasin
MCSLines 1–13providegoodcoverage oftheAdareBasin;36 sonobuoys deployed along these lines reveal the deep structure of the crust from the southern end of the Adare Trough to the continental shelf at the north end of the Northern Basin. These sonobuoys are generally spaced further apart from each other (>30 km) thanthe offsetsfromthe shipto whichthey recorded data(20–30 km),andsoprovide1Dvelocitymodels.
Sonobuoys in the Adare Basinrecord headwaves from layers atmaximumdepths of0.5–5.4 kmbelow theseafloor;28 ofthe 36 penetrate >2.0 km. These sonobuoy profiles reveal variable- thicknesssedimentarylayers(inagreementwithMCSdata;Granot etal., 2010) on top of basement rock, in four cases withmaxi- mumvelocitiesof7000–7400 m/sat3.0–5.4 kmbelowtheseafloor (for example, see L7S1 in Fig. 5e). In another 15 sonobuoy pro- files,maximumlayervelocitiesof6000–7000 m/sare observedat similardepths(see Table 1,Supplemental Material).Velocitiesare uncertainto ±100 m/s anddepths are uncertain up to ±0.4 km (increasingwithdepth).
To aid in visualizing the trends in these 1D velocity models, we provide a 2D velocity model (see Section 2) along Transect 3 through the southern Adare Basin (Fig. 4). We also provide a comparison of 1D velocity models along Line 9 (Fig. 5), which runsalongtheeastern side oftheAdareTrough,andalong Tran- sects 1 and 2 in the northern and middle Adare Basin (Supple- mentalFigs. S2and S3,respectively).Thesesonobuoysarewidely andirregularlyspaced(23–60 kmapart)andwere collectedalong different azimuths (between the ship and receiver). The 2D ve- locitymodel along Transect 3 (Fig. 4) is accurate within ∼5 km ofeach sonobuoy location (along the transect); furtherfrom the sonobuoy location, velocity contour depth varies from the same contour depth at the sonobuoy location by more than ±0.4 km (thevelocityverticaluncertainty).Apparentlydippingvelocitycon- toursresultfromtheinterpolationprocessratherthanfromdirect measurements of the sonobuoy data, and shallow structure be- tweensonobuoylocations(asobservedinnearbyMCSdata)isnot represented.
Transect1crossesthesouthernAdare Troughroughlyperpen- dicularto its axis(see Fig. 2fortransect location, andFig. S2a–c for1D velocity profiles);nearbyMCSdatashowssediment layers wedgingoutagainstthetroughflanks(Fig. S2d).Relativelyfastve- locities(4600–6000 m/s)areobserveddeeperinthemiddleofthis transect,wheresonobuoyL3S2(Fig. S2b)detectsthecrustalstruc- turewithin the Adare Trough. All threesonobuoys reveal simple three-layerstructure,withmaximumvelocities(fromwesttoeast) of6500 m/s,5000 m/s,and5400 m/s(Fig. S2a–c).
Transect2 also crossestheAdare Troughroughly perpendicu- larto its axis, andis at an angle to Transect 1 dueto a change inazimuthoftheAdareTrough(seeFig. 2forcontext,andFig. S3 forvelocity profiles). In Transect 2, one sonobuoy profile lies to thewestofthetroughandtwolietoitseast;againsedimentlay- erswedge out against trough flanks (Granot et al., 2010). These sonobuoysshow slow velocities (up to 4000 m/s, likely consoli- datedsediment) atgreater depthon the westside ofthe trough thanon its east side.However, faster velocities(up to6000 m/s, likely basement rock) are detected at the same depth on the eastand westsides. Atthe two ends ofthis transect, sonobuoys L12S2(Fig. S3a)andL7S2(Fig. S3c)reveal maximumvelocitiesof 6500 m/sand7400 m/s,respectively.Thelattervalueisthefastest layervelocitydetected intheAdareBasin, andistypical oflower oceaniccrust(e.g.,Jones,1999).
The 2D velocity model along Transect 3 (Fig. 4a) runs along thebaseofthecontinentalshelf separatingtheAdareandNorth-
ern Basins (see Fig. 2 for transect location). The fastest velocity contours (5000–5600 m/s) dip down slightly to the west, while slower-velocitycontoursdeepeninthemiddle.Thetwosonobuoys closest tothecontinentalshelf(L8S3andL12S6;Figs. 4b and4c) detect the deepestlayers with velocities <6000 m/s observed in theAdareBasin(5700 m/sand5300 m/s,atdepthsof4.1 kmand 4.3 kmbelowtheseafloor,respectively).
SonobuoysalongthetrendofMCSLine9(fouronthatline,and L7S1)spanthelengthoftheAdareTrough,fromDeepSeaDrilling Project (DSDP) site 274 to the north (collocated with sonobuoy L9S1) to the change intrough azimuthto the south (see Fig. 2).
Fromnorthtosouth(Fig. 5a–e),thesesonobuoyprofileshavemax- imumlayervelocitiesof6800 m/s,5900 m/s,5400 m/s,6000 m/s, and 7000 m/s, withrelatively shallow depths for theselayers of 0.7 km, 1.3 km, 2.0 km, 1.4 km, and 3.1 km below the seafloor (corresponding to 3.9, 4.3, 4.0, 3.7, and 5.4 km below sea level, respectively). Sonobuoy data on Line 9 detect only two or three layers, similarly to sonobuoy data along Transect 1. In contrast, sonobuoy L7S1 detects five layers, typical of sonobuoys south of Transect1intheAdareBasinthatdetectvelocities>5000 m/s(of the24sonobuoysthatmatchthesecriteria,21detectfourtoseven layers).
Within theAdare Basin, adjacent sonobuoy profiles were col- lected only along MCSLine 13, while approaching the continen- talshelf fromthenorth. The2D velocitymodelconstructed from thesefivesonobuoyprofiles(Fig. 6)revealsacrustalstructurethat is consistent withthe eastern endof Transect 2, which isbased on a sonobuoy from Line 7. The 2D velocity model for Line 13 indicates that the southern Adare Basin has horizontal velocity contours. Below the seafloor, which is the top of the rock layer with2000 m/svelocity,contoursof3000 m/s,4200 m/s,4900 m/s, 5700 m/s, and 6000 m/s lie at subseafloor depths of ∼0.6 km,
∼1.8 km,∼2.3 km,∼3.1 km,and∼3.3 kmrespectively.
3.2. DeepstructureacrossthecontinentalshelfandintheNorthern Basin
The crustal structure of the continental shelf and the North- ernBasinisexploredindetail,becauseadjacentsonobuoyprofiles alongMCSLines14–19(seeFig. 2)allowforconstructionofthree 2D velocity models. Line 14 trends northeast to southwest, ap- proximatelyperpendiculartothetrendoftheshelfbreakbetween theAdareandNorthernBasins,whileLine 15doublesbackonthe southernhalfofLine14(onthecontinentalshelf).Line 17doubles backonLine 16(blueandgreenlinesinFig. 2a),crossingLines 14 and15atarightanglewithintheNorthernBasin.Line 19(south- ernmostred lineinFig. 2a)runsparallelto Lines16and17,and crossesLines14and15(orangeandyellowlines)another∼20 km furthersouthwest.
The 2Dvelocity modelthat crossesthecontinentalshelf, con- structed from fourteen sonobuoy profiles along Lines 14and 15, shows remarkably horizontal fast (>5000 m/s) velocity contours (within depth uncertainties; Fig. 7), similar to the 2D velocity model along Line 13. Depths of slower velocity contours (up to
∼5000 m/s)agreebetweenthetwomodels;the6000 m/svelocity contour is ∼4.0 km below the seaflooralong Lines 14, ∼0.7 km deeper than the same contour where it is linearly interpolated along Line 13 (a reasonable difference relative to the depth un- certainty of ±0.4 km, since sonobuoys L13S6 and L14S1 are far enough apart to be sampling slightly different subsurface struc- ture).
The onlypotential detectionof theMoho inthisstudylies at thenorthernendofLine14(sonobuoyL14S1,seeTable 1,Supple- mentaryMaterial),whereavelocityof8000±100 m/s originates at a depth of 5.5±0.4 km below the seafloor(see also Selvans etal., 2012).In orderto verifythat thispossibleMoho detection
Fig. 3.AnalysisandmodelingofonesonobuoyincorporatedinFig. 7(L14S1).The rawdata(a)andfinitedifferencemodel(b)showthedirectwave(betweenthe sourceandthereceiver)asastraightlinethroughtheorigin,curvedreflectionsfrom layerinterfaces(in(a),onlytheseaflooranditsmultiplearedistinct,whereasin(b) othersubsurfaceinterfacesarealsoapparent),andheadwaves.Layervelocitiesare determinedbyapplyingalinearmoveout(lmo)toeachheadwave,shownherefor thefirst(c)andfourth(d)headwavesinthisrecord.SeeSupplementalFig. S1for theresulting1Dvelocityprofile.AdaptedfromFigs. 4and8inSelvansetal. (2012).
isnotawide-anglereflectioncausedbybasementtopography,we usedtopography alonginterfacesimagedby MCSdata(Granotet al.,2010),whichcorrespondtolayersdetectedbysonobuoyL14S1, to runa 2D finite difference model.We found nodetectable dif- ference betweenthe 2D and1D models for sonobuoy L14S1.We dohowevernotethatimagingbasedonMCSdataisonlyavailable downtothebasement(whichcorrespondstotheL14S1sonobuoy layerwithvelocity4400 m/s,seeFig. 3),andthatthe8000 m/sar- rival isthe resultofan interface locatedsubstantially belowthis depth.Energyrefractedoffanup-dipinterfacewitheitherserpen- tenized mantle (with layer velocity ∼7700 m/s;Ritzmann et al., 2004) or lower oceanic crust (velocity ∼7400 m/s; Jones, 1999, p. 342)couldproduceanapparentvelocityof8000±100 m/s.
OftheothersonobuoyprofilesalongLines14and15,onlyfour more detect layers >3.0 km below the seafloor, with the deep- estbeing3.6km(see sonobuoy L14S3inTable 1, Supplementary Material). Maximum velocities detected along Lines 14 and 15 are generally 4400–5000 m/s,with L14S1 (8000 m/s) andL14S3 (5600 m/s) being the only exceptions. Consequently, contours of fastervelocity(>5400 m/s)inthe2D velocitymodelforLines14 and 15 are constrained by two head wave detections along the scarp ofthecontinentalshelf (L14S1 andL14S3),andare uncon- strainedundertheNorthernBasin.
The 2D velocity model based on Lines 16 and 17 (Fig. 8) showsvelocitycontoursthatareapproximatelyflat(3800 m/sand 4400 m/s at ∼1.8 km and ∼2.2 km below the seafloor, respec- tively, or ∼2.3 and 2.7 km below sea level), and a shallow ve-
locity contour that is deflected downward in the middle of the line (2300 m/s, which ranges from 0 kmto ∼0.5 km below the seafloor).Thenorthwesternmostendofthisvelocitymodeliscon- strainedbyarelativelyfastvelocityattheseafloor(2800 m/s)and a relatively slow velocity for a depth of 1.4 km (3000 m/s; see L17S2inTable 1,SupplementaryMaterial);thesedetectionsdeflect velocitycontoursawayfromtheirotherwiserelativelysmoothlat- eralhorizons(Fig. 8).
Velocitycontoursinthe2DvelocitymodelalongLine19(Sup- plementalFig. S4)areflat,depictinghorizonswithvelocitiesupto 4500 m/s (at∼2.5 kmbelowtheseafloor).Thesecontours arein agreementwiththemodelderivedfromLines16and17,andboth are in good agreement with their crossing points on the model derived fromLines 14and15.Individual sonobuoy profilesalong Line19record maximumvelocitiesof4100–4600 m/sat<3.0 km below the seafloor, representative of mostsonobuoy data in the NorthernBasin(seeTable 1,SupplementaryMaterial).
4. Discussion 4.1. AdareBasin
Sonobuoy L12S1 onthe westendof Transect 1(Fig. S2a)and the fouralong Line9 (Fig. 5a–d)demonstratetheoverall charac- teron bothsidesofthe AdareTroughof simplecrustalstructure (i.e., few layers) and shallow fast-velocity layers. In comparison, south of the trough the deep crustal structure is more complex (i.e.,sonobuoysdetectmorelayers),withdeeperfast-velocitylay- ers(e.g.,Figs. 4b–d,S3a,S3c). Thissuggeststhinnerandlesstec- tonicallydeformedcrustalongtheflanksofthetroughthaninthe portion oftheAdareBasinbetweenthe troughandtheNorthern Basin. Previously deployed sonobuoys (Cooper et al., 1987) cor- roboratethisinterpretation, witha velocityof 5200 m/sdetected at 1.8 km subseafloor depth just east of our Transect 1 and at 2.8 km subseafloor depthfurther south,justeast ofLine 13(see Section1.3).
AdjacentsonobuoyprofilesalongLine13,inthesouthernAdare Basin, detect four to six layers, with similar maximum depths (2.7–3.5 km belowtheseafloor), andmaximumvelocitiesindica- tive of intrusiverock (5400–6600 m/s).These fastvelocities sug- gest basementrockandperhaps loweroceaniccrust (i.e.,gabbro, typically6000–7000 m/s;e.g.,Jones,1999,p. 64)atrelativelyshal- low depths below the seafloor. Similar layer velocities at even shallower depthson eitherside ofthe AdareTrough (1.3–2.3 km belowtheseafloor)suggestcontinuityofoceaniccrustalstructure throughoutthe AdareBasin, withlessoverlying sedimentary ma- terialinthenorththaninthesouth.Horizontalvelocity contours interpolated along Line 13 (Fig. 6) suggest that the deep north–
south structure on the east side of the Adare Basin consists of laterallycontinuouslayers,withlittletonodeformation.
However,Transects1–3revealvariationindeepercrustalstruc- turethroughoutthebasin.Basement rockisoverlainby athicker stackofsedimentsalongtheaxisofthetrough(Fig. S2b)thanon its flanks (Figs. S2a andS2c), along thetrend of Transect 1.This transect isnotwithin theisopachcontoursofsediment thickness as determined in the Adare Basin usingNBP0701 MCSdata, but thetrendofgreatersedimentthicknesswithinthetroughthanon itsflanksholdstruesouthofTransect 1(seeFig. 13cinGranotet al.,2010).
Further south, Transect 2 reveals structural asymmetry. The low velocity upper crustal layers (2000–4000 m/s) are thicker (by ∼0.3 km) west of the Adare Trough (Fig. S3a) than on its eastern side (Figs. S3b and S3c), with the slowest-velocity lay- ers (2000–3000 m/s) significantly thicker (by ∼0.8 km) to the west.Theseresultsareconsistentwithsedimentthicknessesdeter- mined fromMCSdata(Granotetal.,2010) atthelocationsofthe
sonobuoysalong Transect 2.Velocitiesof∼6000 m/sat∼1.8 km and∼2.3 kmbelowtheseaflooralsoshowthatfastcrustalveloci- tiesarerelativelyshallowtothewestandeastofthesouthernmost extentof the Adare Trough,where it changes azimuth(Figs. S3a andb,respectively).
AtthesouthernmostextentoftheAdareBasin,alongthebase ofthecontinentalshelf,themiddlesonobuoyinTransect3(L12S6) detectsalayervelocityof5300 m/sat4.3 kmbelowtheseafloor, while sonobuoy profiles on either end of the transect detect a slightlyfaster velocity at slightlyshallower depths (5700 m/s,at 4.1 km depth to the west and 2.8 km depth to the east). The greater depth of fast velocity layers along this transect than in the Adare Basin in general could be related to young volcanism clustered along the west side of the Adare Basin, just west of Transect 3(see Fig. 12ain Granotetal., 2010).Alternatively, the greaterdepthoffastvelocity layers alongTransect 3(particularly atsonobuoysL8S2andL12S6,seeFig. 4)maybeduetoathicker sedimentarypackagethanintherestoftheAdareBasin.MCSdata fromNBP0701 show faulting in the shallow crust from ∼17 Ma thatfollowsthetrendoftheAdareTroughaxissouthtotheconti- nentalshelf(Granotetal.,2010).Thiszoneofdeformationroughly coincideswiththethickestAdareBasinsedimentsimagedwiththe MCSdata(Granotetal., 2010),anditsedgesapproximatelycorre- spondto the locationsof sonobuoys L8S2 and L12S6. It may be thatthegeneraltrendofthickercrustundertheaxisoftheAdare Trough and thinner crust in the flanking portions of the Adare Basin(Müller et al., 2005) holds asfar south as the continental shelfedge.
Having variable depths for velocities >5500 m/s, and veloci- tiesof7000–7400 m/s(indicativeofloweroceaniccrust)inthree locations atfairly shallow depths (in sonobuoys L2S2, L7S1, and L7S2; seeFigs. 5e and S3cfor the lattertwo), suggests variation incrustalthicknessthroughouttheAdareBasin.Thedetectionsof 7000+m/svelocitiessuggestthatthiscrustisoceanic,asexpected fromwaterdepth.
4.2.Structureacrossthecontinentalshelfbreak
Three features stand out in the crustal structure across the margin between the Adare and Northern Basins (Fig. 7). One is the horizontal trend of deep velocity contours (5400–7000 m/s) along the first ∼40 km ofLine 14, beneaththe water depths of
∼2000–1500 m that lie adjacent to the continental shelf break.
The second is the overall 1% grade of shallow velocity contours (2600–3700 m/s)fromdeepwatertoshallow,alongthefulllength ofLines 14and15(∼125 km).Thesetwo trendsinvelocitycon- tour depth are in contrast to the greater changein water depth alongthelines,from1960matL14S1to∼500 monthecontinen- talshelf(seeTable 1,SupplementaryMaterial).Thesmallchangein shallowvelocitycontourdepthandlackofchangeindeepvelocity contourdepthsuggestacontinuityofcrustaltypeacrosstheconti- nentalshelf;velocitycontourswouldhavealargerchange,andin theoppositedirection(fromshallowvelocitycontours),ifthecrust were to transition fromoceanic to continental (e.g.,Jones, 1999, p. 391).ContinuitybetweenAdareandNorthernBasincrustisalso suggestedby twowidelinearmagneticanomaliesalongthesides ofthe NorthernBasin, whichappear to be extensionsof anoma- lies16–18oneithersideoftheAdareTrough(Fig. 2b)(Candeand Stock,2006),andbytheabsenceoffeaturesintheBouguergravity anomalydataattheshelfbreak(Fig. 2c)(Daveyetal.,2006).
ExtensionaldeformationfocusedwithintheAdareBasinspread- ing center was more distributed within the adjacent continental crusttothesouth,either(locally)betweenthebrittlecrust ofthe AdareandNorthernBasinsandtheunderlyingmoreductilelitho- sphere(Davey etal.,2006), or(regionally)betweenthe Northern Basinand other basins further east in the Ross Sea (Decesari et
al.,2007; WilsonandLuyendyk,2009).Theformerpredicts>100%
stretching of the crust beneath the Northern Basin, whereas the latter expects the full amount of extension to be spread across three north–south bands of extension (the NorthernBasin being the north end of the westernmost of these bands). Our study cannot resolve which of these is more accurate (i.e., what the stretching factor is for crust beneaththe Northern Basin), but it does support the general idea of it being a transitional crustal type,asarguedbyDaveyetal. (2006).
The third feature of interest in Fig. 7is the maximum veloc- ity of 8000±100 m/s detected by L14S1, a value routinely in- terpreted as uppermost mantle (i.e., the location of the Moho).
Measuring this velocity at 5.5±0.4 km belowthe seafloor may be duetothincrust atthislocation (e.g.,Selvansetal., 2012), or maybeduetoaplanarinterfacewithinthebasementrockthatis significantlytilted(>10◦) uptowardtheNorthernBasin(seeSec- tion 3.2). If the true maximum velocity detected with sonobuoy L14S1 is 7900 m/s (within measurement uncertainty), it is also possible that the interface is in fact the Moho, butis composed ofserpentinizedmantlesuchasisobservedintheSvalbardconti- nentalmargin(Ritzmannetal.,2004).
Although no previous studies of crustal thickness along this portionofthe RossSeacontinentalshelf exist,some estimatesof crustalthicknesshavebeenmadewithin eachbasinusinggravity datatodetermineMohodepth.ModelsofmantleBougueranoma- lies inthe AdareBasin indicatecrustal thicknesses of 9–10.5 km alongtheAdareTroughaxisand5–6 kmeastofthetrough(Müller etal., 2005).This estimate ofcrustal thicknessisconsistent with ourcalculatedvalueof5.5 km,indicatingthatthecrustalstructure attheshelfbreakmaybecontinuouswiththatoftheAdareBasin, againsuggestingatransitional crustalstructureatthecontinental shelfbreak.
4.3. NorthernBasin
This study indicates that the Northern Basin is filled with approximately horizontal sediment horizons (Figs. 7, 8, and S4), consistent with MCS data (Granot et al., 2010; Brancolini et al., 1995). Sediment layers potentially extendto a depth of >3 km, if maximum velocities of 4100–5000 m/s can be attributed to relatively fast-velocity sedimentary rock (consistent with inter- pretation of similar velocities, and basement depths of up to 3.0 km, to thesouth and east ofour study;Cooper etal., 1987;
HoutzandDavey,1973).Thisissimilartotheuppermost3 kmof crustal structure imagedin thewestern portion ofTransect 3, at the baseof the continentalshelf (Fig. 4). Fast-velocity sediments and shallow bathymetry could be attributed, at least in part,to seawardpropagationofglacio-marinesedimentdepositsthathave filled the basin, out over oceaniccrust, since its formation (e.g., Granotetal.,2013).
Two exceptions to the generally horizontal velocity contours observed in the Northern Basin are a slight down-warping of the 2300 m/s contour along the axis of the basin and close to the continental shelf (see Fig. 7), and a nearly constant velocity of ∼2900 m/s for the entire 1.4 km depth of crust imaged by sonobuoyL17S2(Fig. 8).Additionally,neighboringsonobuoyL16S6 (Fig. 8) detects layer interfaces toan anomalously shallowmaxi- mumdepth(only1.0kmbelowtheseafloor)comparedtotherest of thesonobuoy profiles in the NorthernBasin. Interestingly, the linearmagneticanomalyalongthewestsideoftheNorthernBasin (see Fig. 2b) correspondsto thelocation ofsonobuoysL17S2and L16S6.
CandeandStock (2006)suggestthatthesemagneticanomalies may be the result of massive intrusions that helped accommo- date extension in theNorthern Basin. It isalso possiblethat the crystalline crust of the NorthernBasin is either oceaniccrust or
Fig. 4.2DvelocitymodelalongTransect 3(a)showsvelocitycontoursbasedoninterpolationbetweenthree1Dvelocitymodels(b–d).Notethattheshallowvelocitycontours (2000–4000 m/s)arerelativelyflat,whilethe5000 m/scontourdipsdowntothewestandthe5600 m/scontourisdeflecteddowninthemiddle.SonobuoysL8S3(b)and L12S6(c)detectthedeepestlayerswithvelocities<6000 m/sintheAdareBasin.Depthsofindividuallayersdirectlydetectedinthesonobuoydataareplotted(withan“x”) atthethreesonobuoypositions;arrowspointatthedeepestlayerdetectedbyeachsonobuoy.Thevelocitymodelisaccuratefor±5 kmalongthelineateachsonobuoy location;velocitiesareuncertainto±100 m/s,depthsupto±0.4 km(increasingwithdepth).
Fig. 5.Fromnorthtosouth(a)–(e),sonobuoyprofilesalongtheeastflankoftheAdareTroughhavemaximumlayervelocitiesof6800 m/s,5900 m/s,5400 m/s,6000 m/s, and7000 m/s(respectively),thefirstfouratrelativelyshallowdepthscomparedtoallsonobuoyprofilesintheAdareBasin.SonobuoyL7S1(e)isoneofthreetodetecta layervelocity>7000 m/sintheAdareBasin.
a transitionalcrust (Davey etal., 2006),as alsoinferred between oceanicandcontinentalcrustintheGulfofCalifornia(Oskinetal., 2001). The shallow penetration depths of L17S2 andL16S6 may indicatedisruptedcrust,alikelyoccurrenceeitherduringmassive intrusionsassociatedwithcrustalextension orduring morelocal- izedvolcanicactivity.
Onthecontrary,DaveyandBrancolini (1995)interpretthedeep NorthernBasinstructureascontinentalcruststretchedby>100%.
AnestimateofNorthernBasincrustalthickness,usinggravityand seismicdatatorespectivelydetermineMohoandbasementdepths, indicates ∼12 km ofcrystalline crust underlies sedimentary rock (Davey andBrancolini, 1995). If the NorthernBasin is underlain
Fig. 6.2DvelocitymodelalongLine13showsvelocitycontoursbasedoninterpolationbetweenfiveadjacent1Dvelocitymodels(seeTable1inSupplementaryMaterials).
Notethatvelocitycontoursarehorizontal.DetailsoflabelingandmeasurementuncertaintyareasstatedforFig. 3;inthiscase,themodelisaccuratealongtheentireline.
Fig. 7.2DvelocitymodelalongLines14and15,whichcrossesthecontinentalshelffromtheAdareBasintotheNorthernBasin(waterlayerinwhite).Velocitycontoursare basedoninterpolationbetweenfourteenadjacent1Dvelocitymodels.Notethatvelocitycontoursfrom5400 m/sto7000 m/sareapproximatelyhorizontal,andthosefrom 2600 m/sto3700 m/shaveashallowupwardslope(∼1%gradeoverall)fromdeepoceantoshallowwaterdepths.Contoursareconsistentwiththoseatthesouthernend ofLine13(Fig. 6).Depthsofindividuallayersdirectlydetectedinthesonobuoydata(“x”forLine14,“+”forLine15)indicatethepositionsofthesonobuoys;arrowspoint atthedeepestlayerdetectedbyeachsonobuoy.Velocitiesareuncertainto±100 m/s,depthsupto±0.4 km(increasingwithdepth).
Fig. 8.A2DvelocitymodelalongLines16and17(a)intheNorthernBasinshowsvelocitycontoursbasedoninterpolationbetweenelevenadjacent1Dvelocitymodels.Note thatvelocitycontoursaregenerallyhorizontal,andrevealacrustalstructureconsistentwithLines14and15(Fig. 7)wheretheycross.Detailsoflabelinganduncertainty areasstatedforFig. 7,withtheexceptionthatindividuallayersarelabeled“x”forLine17and“o”forLine16.
Fig. 9.Summaryofresultsfromsonobuoyanalysis,shownwithregionalbathymetry(perspectiveisfromthenortheast).1DvelocityprofileseastoftheAdareTrough(in theAdareBasin)revealvelocitiesindicativeoflowercrustatrelativelyshallowdepths,whereasthetroughitselfhasfastvelocitieslocateddeeperinthecrust.Adjacent sonobuoysneartheshelfbreakandintheNorthernBasinshowthickersedimentandflatvelocitycontoursatdepth,consistentwithacontinuouscrustaltypeacrossthe shelf.ApotentialMohodetectionislocatedinneartheshelfbreak.
by continentalcrust,it maybethat thetransitionincrustal type betweenthe AdareandNorthernBasinsoccurssouthoftheshelf break.
Theshallowsubsurface structureoftheNorthernBasinasde- terminedbythisstudy(to∼3 kmdepth)isconsistentwithprevi- ousseismicstudiesintheRossSea,ascompiled andinterpolated byCooperetal. (1995).Thosestudiesalsorevealedtheuppermost of the sediment horizons to be nearly flat inthe portion ofthe basinstudiedduringresearchcruiseNBP0701.Cooperetal. (1995) determinedthattheacousticbasementisalmostentirely>4.0 km belowtheseafloorinthisregion(consistentwithourdeepvelocity contoursacrossthecontinentalshelf;seeFig. 7),withamaximum of∼5.5 kmandaminimumof2.0–3.0 km(thelatteratthenorth- west extremeof thestudy area, approximatelywhere sonobuoys L17S2andL16S6 are located). However,our velocity contoursdo not directly correspond to the interfaces mapped by Cooper et al. (1995),because theRSU4a and RSU5 unconformitiescross up intotheeasternportionsofour2DvelocitymodelsforLine17/16 (Fig. 8)andLine19(Fig. S4).Itcouldbethatsedimentvelocitiesin theNorthernBasinare determined moreby overburdenpressure (i.e.,depth)thanbyage,inwhichcasepreviousseismicresultsand thoseofNBP0701wouldbefullyconsistent.
4.4. Comparisontosimilarcrustaroundtheworld
UltraslowspreadingzonessuchastheAdareBasin(∼12 mm/yr full-spreadingrate;Candeetal.,2000)aretypicallyassociatedwith thinoceaniccrust.Thetraditionalviewofoceaniccrustalstructure was determinedfromstudiesofophiolite suitesonlandandma- rine seismic data, both of which indicated subhorizonal layering beneaththesediment–basementcontactof1)basalticpillowlavas (vp∼5000 m/s), 2) sheeted dikes, 3) gabbro (vp∼7000 m/s), and 4) peridotite mantle (vp∼8000 m/s), with mean oceanic crust (composed of basalt andgabbro) being ∼7 km thick (e.g., Christensen,1978). However,insitustudiesoftheselayers reveal thattheoceaniccrustalstructureiscomplexandvariable(Karson, 1998; Dicketal., 2003).Fault scarpson theseafloorreveal devi- ations from mean oceanic crustal structure on the scale of tens ofmetersto tensofkilometers, particularlyalongslowspreading ridges andin magma-poor locations, where stretching and thin- ning of the lithosphere often results in crust <7 km thick and thepresenceofoceaniccorecomplexes(exposedalonglow-angle detachmentfaults,whererockunitsaremissing,andcontactsbe- tween units that are neither horizontal or continuous) (Karson, 1998).
Peridotites are found along ultraslow (≤12 mm/yr) spreading ridges(Dick etal.,2003),indicatingthat evenmantlerockcanbe exhumedinthesesituations.Modelingofthedetachmentfaulting process andresulting crustalthicknessatspreading ridges, based ondifferingamountsoftectonicandmagmaticcrustalformationin boththeupper(basalt)andlower (gabbro)oceaniccrust (Oliveet al.,2010),indicatesthat ourmeasuredcrustalthicknessof5.5 km atthecontinentalmargin(2.1 kmsedimentand3.4 kmbasement rock)isconsistentwiththeformationofanoceaniccorecomplex along a detachment fault at a slow-spreading ridge, in the case wherehalfoftheextensionisaccommodatedthroughcrustalthin- ningandhalfthrough magmatism(see summary Fig. 9). Thefast velocities we measure at shallow depths along the eastern side of the Adare Trough (6000–7400 m/s,originating at1–3 km be- lowtheseafloor)arealsoconsistentwithscenariosofoceaniccore complexformationasmodeledbyOliveetal. (2010).
The crust around the arctic Gakkel Ridge (6–13 mm/yr full- spreading rate) reduces to thicknesses of 1.9–3.3 km along the ridge(Jokatetal.,2003),whilethicknessesof2.5–5.4 km(Minshull etal.,2006; Mülleretal.,2000) areobservedalongtheSouthwest IndianRidge(11–18 mm/yr;ChuandGordon,1999).Crustalthick- nesswithintheAdareBasinisnotdirectlymeasuredinthisstudy, butthinoceaniccrustwouldbeconsistentwiththecrustalstruc- ture detected by sonobuoy profiles along Lines 9 and13 (Figs. 5 and 6), aswell as sonobuoys L7S1 and L7S2 (Figs. 5e and S3c), wherefastcrustalvelocities(upto7400 m/s)aredetectedatrela- tivelyshallowdepthsinthecrust.AdareBasincrustappearstobe similartothatofoceaniccorecomplexesobservedandmodeledat ultraslowspreadingridges(e.g.,Karson,1998).
5. Conclusions
Sonobuoy data in the Adare Basin suggest particularly thin oceanic crust along the east flank of the Adare Trough, consis- tent with estimates of 5–6 km thickness based on gravity data (Müller etal., 2005). Sonobuoy profiles also suggest the crust is thickerwithin theAdareTroughandalong thebaseoftheconti- nental shelf,againconsistent withMülleretal. (2005).Detection ofavelocityof8000±100 m/s,whichmayhaveoriginatedfrom the Mohoat5.5±0.4 km belowtheseafloor, suggeststhat typi- cal AdareBasincrustal thicknessesof5–6 kmcould extendasfar southasthecontinentalshelf.
Adjacent sonobuoy profiles from the Adare Basin to within 12 kmofthecontinentalshelfbreakrevealdeepvelocitystructure with horizontal contours, suggestingthat the deep crustal struc- ture of the Northern Basin may be continuous with that of the
