The giant Ruatoria debris avalanche on the northern Hikurangi margin, New Zealand: Result of oblique seamount subduction Jean-Yves Collot, Keith Lewis, Geoffroy Lamarche, Serge Lallemand

To cite this version:

Jean-Yves Collot, Keith Lewis, Geoffroy Lamarche, Serge Lallemand. The giant Ruatoria debris avalanche on the northern Hikurangi margin, New Zealand: Result of oblique seamount subduc- tion. Journal of Geophysical Research. Oceans, Wiley-Blackwell, 2001, 106 (B9), pp.19271-19279. ￿10.1029/2001JB900004￿. ￿hal-01261406￿

HAL Id: hal-01261406 https://hal.archives-ouvertes.fr/hal-01261406 Submitted on 25 Jan 2016

HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non, lished or not. The documents may come from émanant des établissements d’enseignement et de teaching and research institutions in France or recherche français ou étrangers, des laboratoires abroad, or from public or private research centers. publics ou privés. JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 106,NO. B9, PAGES 19,271-19,297,SEPTEMBER 10, 2001

The giant Ruatoria debris avalancheon the northern Hikurangi margin, New Zealand: Result of oblique seamountsubduction

Jean-Yves Collot

UMR G6osciencesAzur, Institut de Recherchepour le D6veloppement,Villefranche sur rner, France

Keith Lewis and Geoffroy Lamarche National Instituteof Water and AtmosphericResearch, Wellington, New-Zealand

Serge Lallemand Laboratoirede G6ophysique,Tectonique et S6dimentologie,Universit6 de MontpellierII, Montpellier,France

Abstract. Despiteconvergent margins being unstablesystems, most reports of huge submarineslope failure havecome from oceanicvolcanoes and passivemargins. Swath bathymetryand seismicprofiles of the northernHikurangi subduction system, New Zealand,show a tapering65-30 km wideby 65 km deepmargin indentation, with a giant,3150+_630 km 3, blocky, debrisavalanche deposit projecting 40 km out acrosshorizontal trench fill, and a debris flow depositprojecting over 100 km. Slide blocksare well-bedded,up to 18 km acrossand 1.2 km high,the largestbeing at the avalanchedeposit's leading edge. Samples dredged from themare mainly Miocene shelfcalc-mudstones similar to thoseoutcropping around the indentation.Cores from coverbeds suggest that failure occurred-170 +_40ka, possiblysynchronously with a major extensioncollapse in the upperindentation. However, the northernpart of the indentationis much older.The steep,straight northern wall is closeto the directionof plate convergenceand probably formedaround 2.0-0.16 Ma as a largeseamount subducted, leaving in its wake a deepgroove obliquelyacross the marginand an unstabletriangle of fracturedrock in the 60ø anglebetween grooveand oversteepenedmargin front. The trianglecollapsed as a blockyavalanche, leaving a scallopedsouthern wall andprobably causing a largetsunami. Tentative calculations of compacted volumessuggest that the indentation is over 600 km 3 largerthan the avalanche, supporting a two- stageorigin that includes subduction erosion. Since failure, convergence has carried the deposits -9 km back towardthe margin,causing internal compression. The eventualsubduction/accretion of the Ruatoriaavalanche explains the scarcityof suchfeatures on activemargins and perhapsthe nature of olistostromes in fold belts.

1. Introduction Large submarineslope failure occursin a variety of forms that can be categorized by what can be regarded as end- Submarine avalanches and debris flows can be enormous. members of a continuum of gravitational processes. Perhaps Thosethat occuron slopesbetween land and deepocean basins the largest submarine landslides are rotational slumps that can be several orders of magnitudelarger than the largest involve the slow or intermittent, downslope movement of landslidesonshore [Hampton et al., 1996]. They can involve largely intact, back-tilting blocks on glide planes as much as the catastrophicmovement of hundredsor even thousandsof 10 km below the seabed [Moore and Normark, 1994]. On the cubic kilometers of broken rock and sediment. They are a other hand, large catastrophic slope failure occurs as threat to offshorestructures, such as cables and platforms, and disaggregateddebris avalanches, with blocks up to many they can devastatecoastal areas both by onshoreretrogression kilometers across and run-out distancesof many tens to more at their head [Coulter and Migliaccio, 1966; Mulder and than a hundredkilometers [Bugge et al., 1987; Moore et al., Cochonat, 1996] and by generation of large tsunamis 1989; Moore and Normark, 1994]. Similar, but generally [Bondevik et al., 1997; Moore and Moore, 1984]. Ancient smaller, thinner, and more disaggregatedsediment slurries, masses of broken blocks have been described as "chaos with fewer and raftedblocks, are generallyreferred to as debris deposits",melanges or olistostromesin fold belts aroundthe flows [Enos, 1977; Masson et al., 1998]. They travel further world lAbbate et al., 1970; Ballance and Sporli, 1979; Hsu, than debris avalanches,perhaps because they travel faster in 1974; Naylor, 1981; Orangeand Underwood,1995], with the same environment [Jacobs, 1995; Weaver, 1995], and part debate often centering on whether particular deposits are of them may incorporatewater and mudto metamorphoseinto gravitationalor tectonicin origin. turbidity currentscapable of travelling a thousandkilometers or more [Garcia and Hull 1994]. Copyright2001 by the AmericanGeophysical Union. Massive margin failure can occurin a variety of geologic Papernumber 2001JB900004. settings. Perhaps the best documentedare on the flanks of 0148-0227/01/2001J B 900004509.00 oceanic "hot spot "volcanoes, wherequenching of lava has

19,271 19,272 COLLOT ET AL.: GIANT RUATORIA DEBRIS AVALANCHE, NEW ZEALAND critically oversteepenedslopes [tlolcomb and Searle, 1991]. [Mooreet al., 1976],the Aleutian Trench [Lewis et al., 1988], At Hawaiian Islands [Jacobs,1995; Moore et al., 1989; Moore Peru[Bourgois et al., 1993; Duperretet al., 1995; von Huene andNormark, 1994], Canary Islands [Masson, 1996; Masson et al., 1989],Costa Rica [Hinz,1996], and Japan [Cadet et al., et al., 1998; Urgeleset al., 1997] and Fournaisevolcano near 1987]. Subductionof oceanic asperities, commonly Reunion Island [Lenat et al., 1989], it has been shown that seamounts,have produced indentations in convergentmargins enormous rotational slumps, debris avalanches, and debris aroundthe world, with only small landslidesin their wake flows, some thousand of cubic kilometers in volume and [Lallemand et al., 1990]. ' extendingabove sea level, have collapsedcatastrophically In this paper, we documentthe massive Ruatoria debris into the surroundingdeep oceanbasin. avalancheand debris flow associatedwith a large-scale, Passivemargins are also the location of large submarine morphologicindentation of the Hikurangisubduction margin slope failure. In some cases, failure is associatedwith excess east of North Island, New Zealand. We interpret the pore pressurein sedimentaryrocks being maintainedby gas, indentationand slopefailure association from geophysical often from unstableclathrates [Bugge et al., 1987; Carpenter, datato suggestthat they result primarily from the Quaternary 1981; Lerche and Bagirov, 1998]. In other cases, failure subductionof a large seamount. We then focus on the results from rapid sedimentoverloading or tectonic stresses dynamicsof avalanchingand massbalance calculations, infer resultingfrom, among other things, isostatic rebound[Bugge that oblique seamountimpact encourageslarger margin et al., 1987]. Notable examplesoccur off Norway [Buggeet collapsecompared with orthogonalconvergence, and finally al., 1987; Jansen, 1987], South Africa [Dingle, 1980], and discussthe apparent scarcity of such features on active Northwest Africa [Massonet al., 1998; Weaver, 1995]. margins. Many of the samecauses of instability occurat convergent margins, where active forearc slopes are maintained at a 2. GeologicalSetting of Ruatoria Indentation critical angle, suggestingthat they should be a privileged and Avalanche location for catastrophicslope failure. Continental collision zones such as the Gibraltar Arc are the location of giant, The Ruatoria indentation and avalanche are located at the submarine,chaotic bodies [Torelli et al., 1997]. Moderate- to northern extremity of the Hikurangi margin, offshore from large-sized landslideshave been reportedfrom the SundaArc EastCape (Figure 1). The Hikurangi margin is at the southern

Figure 1. Location of the Ruatoriaavalanche and margin indentation. Flaggedline is the convergentplate boundarybetween subductingPacific Plate and the edge of the Australian Plate east of the back arc Havre Trough and Taupo Volcanic Zone (TVZ), referred to as the KermadecForearc. The Hikurangi Plateau is thickened,seamount-studded oceanic crust being subductedat the sediment-starvedsouthern Kermadec Trench and sediment-filledHikurangi Trough. EC is EastCape. HC is the HikurangiChannel. AF is Awanui Fault. COLLOT ET AL.: GIANT RUATORIA DEBRIS AVALANCHE, NEW ZEALAND 19,273

PAC-AUS

RuatoriaKnoll .;-' --.• /,

ß ,

--• ?•...

.

/ -k x764 Core Site Gisborne Seamount , e- X765 DredgeSite -- Single-channelseismic reflection profiles

? EM12DMultichannel and MR1seismic multibeamreflection profiles 7,9ø30' I 18iø W177ø30 ' Figure2. Geophysicallines, rock samples and cores used in thisstudy. Bathymetry isat 100-m intervals. Insetshows convergence vectors between Pacific Plate (PAC), Australian Plate (AUS), and Kermadec Forearc (KER)relative to thedeformation front (flagged line). KER-AUS vector was estimated from back arc kinematics.Back arc extension rates decrease from 15-20 mm yr-• in the southernHavre Trough [Wright, 1993]to 8-12 mm yr '• inthe Taupo Volcanic Zone, on the basis of onshore geodetic triangulations [Walcott, 1987]and integration ofGPS measurements [Darby and Meertens, 1995]. The direction of extensionranges fromN124øE +_ 13 to N135øE,on thebasis of GPSdata [Darby and Meertens, 1995] and earthquake T axis azimuths[Anderson etaI., 1990: Pelletier and Louat, 1989]. Averaging these values, we estimate a rate of back arcopening (KER-AUS) of 12.5 mm yr -• in a directionN135øE tbr the latitude of theRuatoria indentation. Usingthese values, the PAC-KER convergence is54 mm yr -• in a directionN277øE.

end of the Tonga-Kermadec-Hikurangisubduction system, volcanic seamounts(over 1 km high) and smaller knolls are whereconvergence between the PacificPlate (PAC) and the elongatedor alignedin ridgestrending N150øE+ 20 ø [Collotet overriding AustralianPlate (AUS) decreasesand becomes al., 1996]. Betweenthe seamountsand ridges, the plateauis progressivelymore oblique toward the south.The relative blanketedby pelagicsediments, < lkm thick, of mainly late PAC-AUSplate motion at the northernextremity of the Cretaceousand Paleogene age [Woodand Davy, 1994], which Hikurangimargin is 45 mmyr -• in a directionof 267øE[De also underlie the subduction trench, both in the sediment- Mets et al., 1994]. Consideringback arc openingin the Taupo floodedHikurangi Trough and in thesediment-starved southern Volcanic Zone and Havre Trough, the speedof convergence KermadecTrench to the north [Lewis et al., 1998]. Within betweenPacific Plate and the KermadecForearc (KER) (Figure -200 km of the Hikurangideformation front, the pelagiclayer 1), includingthe northern Hikurangi margin, is 54 mmyr '• in is coveredby a wedgeof sheetturbidites originating from a a directionN277øE (PAC-KER in Figure 2). channelsystem (Figure 1) that tums out of the Hikurangi Alongthe Hikurangi margin, the oceanicHikurangi Plateau Troughacross the central Hikurangi Plateau [Lewis, 1994]. on the Pacific Plate is subductedbeneath thinning continental The Hikurangimargin changesradically from north to cruston the featheredge of the AustralianPlate [Lewisand south[Lewis and Pettinga, 1993] (Figure1). Its wide central Pettinga,1993; Walcott,1978]. TheHikurangi Plateau is up segmenthas a foundationof imbricatedCretaceous to upper to -600 km wide and is believed to be 12-15 km thick Mioceneshelf slope strata fronted by a 70-km-wide,gently Cretaceousoceanic crust [Davy, 1992; Mortimer and (2ø) slopingaccretionary prism [Davey et al., 1986;Lewis and Parkinson,1996; Woodand Davy, 1994].The northernpart of Pettinga,1993], which grew 50 km seawardwithin the last 0.5 the plateauis heavily studdedwith volcanicedifices of Myr [Barnesand Mercier de Ldpinay, 1997]. In contrast,the probableCretaceous age [Strong, 1994], althoughthere is narrownorthern segment lacks a recentaccretionary wedge and tentative evidence to the south of late Miocene or younger its frontalpart is steep(10ø). It is the siteof tectonicerosion seamounts[Lewis and Bennett, 1985]. Many of the large by a seamount-studdedsubducting plate with limited sediment 19,274 COLLOT ET AL.: GIANT RUATORIA DEBRIS AVALANCHE, NEW ZEALAND

S37ø40'.•..,

D

o (•

O 20 km I i I A 180 ø

km HikurangiTrough

-1 -20 •.-..._• ":''-'• .'•" •A I Debrisavalanche

-'-'•...... •'•:•':•:' ...... •::"'•

-5 •! ...... I ...... DI ...... I ...... I ...... I ...... I ...... I ...... I ...... I ...... DebrisI ...... flowI ...... 0 10 20 30 40 50 60 70 80 90 100 110 120 km Figure 3. (top) Bathymetryof the Ruatoriaavalanche and associatedmargin indentation with contoursat 25- m intervals.(bottom) Bathymetriccross sections A and D acrossthe margin with base of avalancheshown by dashedline. Cross sectionsare located on bathymetricmap.

cover, the deformationfront being offset landwardby 10-25 nongrowingoversteepened accretionary wedge, dissected by km comparedwith the margins to north and south [Collot et transcurrentfaults and tectonically erodedby the subducting al., 1996; Davey et al., 1997]. Onshore, the northern HikurangiPlateau [Collot and Davy, 1998]. This wedgeis margin's foundation of Cretaceous-Paleogenerocks i s separatedfrom the morestable, upper part of the margin,by overthrust by the vast East Coast Allochthon or nappe the 220-km-long, transcurrent,Awanui Fault, which is cut at obductedfrom the NE during early Miocene time [Rait, 1995; its southernend by the Ruatoriaindentation (Figure 1). Stoneley, 1968]. The allochthon is overlain by Neogeneand The Ruatoria avalanche was first tentatively recognized Quaternary shelf slope sediments, and the whole series has from conventionalbathymetry by [Lewisand Pettinga,1993]. been involved in upper Cainozoic compressional tectonics The preliminaryresults of the first swathmapping survey of [Field et al., 1997]. the area appearedto validate the suggestionthat rugged Immediately north of the Ruatoria indentation, the topographyin thenorthern Hikurangi Trough was a productof Hikurangi Plateauis subductingbeneath the oceanic Kermadec massiveslope failure on the adjacentmargin. However,it Ridge. The southern Kermadec margin consists of a further prompteda suggestionthat the geometryof the COLLOT ET AL.' GIANT RUATORIA DEBRIS AVALANCHE, NEW ZEALAND 19,275

Eastcap...... :•...;..--"•:•e: - , •:i•. :;.,"•,.. Indentation

northern wall

Deep reentrant

HikurangiTrough ...... ,;,,•'.. Blocky .?,.... ß.-::•.'.:..... :...... ??:"'

...... :... ß .

Figure 4. Obliqueterrain model of Ruatoriaavalanche and indentationshowing the linear northernwall, with the deepreentrant in the lower margin, and the scallopedsouthern wall, with the avalanchedeposit at its seawardend. The model also contrasts the block-free upper indentation with the avalanche-coveredlower indentation and Hikurangi Trough. indentationassociated with the slope failure deposit indicates Seismic reflection data were processedusing Globe ClaritasTM not just gravitationaleffects but also the tectoniceffects of seismic processing software to fully migrated sections. seamount subduction [Collot et al., 1996]. This paper Processing included time-domain filtering, predictive describesboth indentation and slope failure deposits in much deconvolution,threefold stack, and 1500 m s'• velocity greater detail using a more comprehensivedata set and migration. We also reinterpreteda series of seismic reflection providesnew evidenceand new interpretationson their nature, profiles archived at National Institute of Water and age, evolution,and relationshipto one another. AtmosphericResearch Ltd (NIWA). In April 1991, April 1995 and May 1999, rocks were 3. Geophysicaland GeologicalData Collection dredgedfrom the indentation walls and from slope toe blocks and Processing (Figure 2 and Table 1) in an eftbrt to determine the relationship between them. Most of the rock sample Multibeamswath bathymetry and backscatter imagery were resembles Cainozoic mudstones on the adjacent land in recordedacross the Ruatoriaavalanche deposit and lower part appearance and degree of induration. Their age and of the indentation during the GeodyNZ cruise of the R/V paleoenvironment were deducedby Stratigraphic Solutions L'Atalante, November 1993 [Collot et al., 1996]. The Ltd. from their nannofossil and, in some cases, their L'Atalante's swath bathymetry and imagery were foraminiferal content. In May 1999, three cores were obtained supplementedwith HAWAII MR1 swath data (Figure 2), from cover beds to try to estimate the age of the indentation collected in September 1994 from the New Zealand vessel and avalanche(Figure 2 and Table 2). Well-defined correlatable Giljanes. The new swath data were integratedwith archived tephrasin two of the cores were identifiedby Auckland bathymetricdata to producea new bathymetrygrid that could UniServicesLtd. using electron microprobeglass shard be contoured at 25-m intervals, revealing the detailed analysis used to calculate sedimentationrates. morphology of the entire indentation and avalanche area (Figures3 and 4). Seismicreflection profiles aligned parallel with the margin 4. Morphostructure of Ruatoria Indentation were obtained during the GeodyNZ cruise. The equipment 4.1. Dimensions and Morphological Divisions of consistedof two 75 cubic inch GI air guns operating in the Ruatoria Indentation harmonic mode and a six-channel seismic streamer. Seismic lines transverseto the margin were collected from the R/V The new bathymetric map and Three-dimensional(3-D) Tangaroain March 1998 with a 75-75 cubic inch GI air gun diagramof the northern Hikurangi margin (Figures 3 and 4) usedin harmonicmode and a 24-channelhydrophone array. showa largeindentation incising a steepcontinental slope for 19,276 COLLOT ET AL.: GIANT RUATORIA DEBRIS AVALANCHE, NEW ZEALAND

Table 1. Rock SamplesFrom Ruatoria Indentationand Avalanchea

Station Latitude Longitude Position Depth,m Lithology Age Paleoenvironment

V469 -38o13' 179o32' avalanche 2600 calc-mudst e Pliocene 1 shelf- u slope V470 -38006' 179009' eastsouthern wall 1540 calc-mudst e-m Miocene u-m slope X280 -37o52' 179o04' top reentrant 1470 calc-mudst m Miocene slope X281 -37056' 179015' head scarp 2200 limest 1Pliocene slope X282 -37005' 179ø37' avalanche 3330 calc-mudst m Miocene shelf- u slope X765a -38020' 179035' RuatoriaKnoll -2500 limest m Oligocene u slope X765b -38020' 179ø35' RuatoriaKnoll -2200 calc-mudst e-m Miocene? shelf- u slope X765c -38020' 179035' RuatoriaKnoll -2200 calc-clayst e Miocene u - m slope X765d -38020' 179035' RuatoriaKnoll -2200 calc-siltst e Miocene shelf- u slope X766a -38005 ' 179010 ' east southern wall 17007 calc-mudst e Miocene ?- e shelf Pliocene X766a -38005 ' 179010 ' east southern wall 17007 calc-mudst e - m Miocene shelf X768 -37041' 179001' westnorthern wall 936 calc-mudst 1 Miocene- e m slope Pliocene X769 -37045' 179034' eastnorthern wall 2260 calc-mudst e Pleistocene 1 shelf- u slope aLithologyabbreviations are mudst, mudstone; limest, limestone; clayst, claystone; siltst, siltstone. Ages are based on nannofossils(e, early;m, mid; 1,late). Paleoenvironmentis basedon nannofossils,foraminifera, and lithology.(1, lower; u, upper).

-65 km landwardof a line joining the deformationfront on steepestparts being inclined at -22 ø. Rocks from both ends of either side (Figure 3). The width of the indentation decreases the wall are calcareous or tuffaceous mudstone of lower landwardfrom-65 km along the interpolateddeformation Miocene to lower Pleistoceneage (Table 1). The northern wall front to -30 km at the continental shelf. The top of the is dividedinto eight 5-10 km long segments,which appear indentation incises the 140-m contour, and its base is the dextrally offset by -1 km, making the wall saw-toothedin 3600-m-deep floor of the Hikurangi Trough. The indentation plan (Figure 5). Seismicreflection profile GNZ-05 (Figure 6) is boundon three sidesby steep walls enclosingan area of suggeststhat on the lower slope, wall segments merge at -3300 km2 (Table3). Onthe basis of bathymetricand seismic depth into steeply south dipping faults that may bottom out reflectiondata we dividedthe indentationin two structurally on the d•collement.All of the steeply dipping fault segments distinct parts, a very hummockylower part, separatedby a of the northern wall togetherform a linear dextral strike-slip seawardconcave, midslope scarp from a more undulatingupper fault system that cuts transversally acrossthe margin (Figure part. In sections4.2-4.4, we describea relatively straight 5). The northern wall on the upper slope is associatedwith northern wall and a highly irregular southernwall, which we clear normal faulting as discussedlater in section 4.5 (Figure considerto be critical to understandingthe formation of both 7). the Ruatoria indentation and avalanche. We then show in sections 4.5-4.6 that the upper part of the indentation 4.3. Irregular, Scalloped, Southern Wall consistsof subsidingsedimentary basins contrastingwith its Compared with the northern wall, the southern wall is lower part that containspart of the -3400 km2 blocky avalanche (Table 3). irregularand is extensivelyscalloped in plan view (Figures3, 4 and 5). On the lower slope its overall trend is -N320øE, 4.2. Straight but Saw-Toothed Northern Wall which is significantly different from any value for plate convergence.On the upper slope it trends N280øE, which is The northern wall is 80 km long. Its overall trend is within a few degreesof the estimatePAC-KER convergence, N276øE within 1ø of our estimated PAC-KER convergence and suggests a genetic link with the northern wall. direction.The wall height ranges from nearly 1400 m on the Conspicuousfeatures of this wall are two largearcuate scarps lower slope to < 200 m on parts of the upper slope, the onthe lower slope as well as smaller ones on theupper slope.

Table 2. CoresFrom RuatoriaIndentation, an EnclosedAvalanche Basin, and the HikurangiTrough Above the Debris Flow Deposiff

Station Latitude Longitude Position Depth, Length, Corelog m m

X764 -38o09 ' 179026 ' avalanche 3095 2.93 Hemipelagicmud, few silt layers,ashes 54-57 cm Taupo tephra (1.8 ka) 98-100 cm containsperalkaline tephra 109-113 cm Waimihia tephra (3.3 ka) 205-207 cm redeposited14.7 ka ash 244-247 cm redepositied14.7 ka ash X767 -37o44' 179o08' top reentrant 1302 3.81 hemipelagicmud, few fine-gradedsilts X770 -38o36' 179o15' Hikurangi 3551 0.97 silt turbidites, hemipelagic mud and ashes Trough 61-64 cm Taupo tephra (1.8 ka) 94-97 cm Waimihia tephra(3.3 ka)

•Ash identificationsare basedon electronmicroprobe analysis of glassshards. COLLOT ET AL.: GIANT RUATORIA DEBRIS AVALANCHE, NEW ZEALz•qD 19,277

Table 3. Areas and Volumes Calculatedfor the Ruatoria Avalanche,Indentation and Debris Flow"

Location Name Area, Volume, Error, Por,% Compacted Error, km2 km3 20%,km 3 Volume,km 3 20%,km 3

Upper indentation V 1 1682 929 186 15 790 158 Lower indentation Material loss V 2 1162 579 116 15 492 98 Remainsin margin V 3 1515 302 15 1287 257 Material gain V4 86 V2+V3 1612 2094 419 15 1780 356 Total indentation V 1+V 2 +V 3 3295 3023 605 15 2570 514 Total avalanche Va 3409 3146 629 37.8 1958 392 Difference: (V I+V2+V3)-Va 612 Debris flow Vd 8000 960 192 60 384 77 aUpperand lower indentations are defined in Plateld. Numbersare rounded up to closestinteger. Por, porosity [Field et al., 1997]'volumes V1, V2, V3, V4 andVa aredefined in Plate1. VolumeV3 andVa arecalculated with P wavevelocity of 2000m s'•. Compactedvolumes, Vol *(100-Por)/ 100.Avalanche (Va) compactedvolume is calculatedin Table4. Vd is calculatedfor an averagethickness of 120m.

Upper marginwith relativehighs lrnbricatedlower marginwith highs& fold-and-thrust belt Upper indentationwith few rafted blocks Debris avalanche with rafted blocks & postavalancheponded basins Anticline axis ,,,.•Structuratlineament. Debds flow & thick, disturbed trench fill Synclineaxis ' Debris flow & thin, disturbed trench fill .Thrust/reverseScarps faultsß Postavalanche turbidites Hikurang!Platea!J. pela9i.c drape w.ith Normalfaults votcanic noges & seamounts Amphitheatre Figure5. Generalizedgeological map showing main structural and gravity-controlled features including (1) saw-toothedsegments of northern wall, (2)scalloped southern wall, (3)almost block-free upper indentation withfolds and faults, some related to imbricatedmargin on eitherside, (4) scarpor amphitheaterseparating upperand lower parts of indentation,(5) lowerindentation with rafted blocks and avalanche flow in two directions,ML, mainlobe, SL, secondary lobe, (6)straight structural lineaments in lowerindentation (see Figure4 and Plate lb) consistentwith shortening along the PAC-KER convergence direction, and (7) seaward extentof avalanche,debris flow, anddisturbed Hikurangi Trough fill, with coveringof turbidites. 19,278 COLLOT ET AL.' GIANT RUATORIA DEBRIS AVALANCHE, NEW ZE••

,o ! ...... GnZ:Os...... I m•'ed :bt'•k Imbricatedlower margin Debs Avalanche

NE o Skr. ,...... i I

Figure 6. (top)GNZ-05 seismic reflection line drawing, with location (inset), showing (1) steepreverse faultsthrough imbricated margin rock at the northernwall and(2) low-anglediscontinuity between the southernwall and the debrisavalanche. (bottom) Detail of a migratedsection of the line showingdebris avalanchewith rafted block and unit F at thetop of thesubducting plate.

The lowestarcuate scarp reaches a heightof over 1500 m with bathymetric and structural trends (Figures 3 and 5). Its a northeastwarddip of 25ø. Seismicprofile GNZ-05(Figure 6) position along the continuation of major continental shelf showsthat this scarpcontinues to a depth of --1.5 km beneath faults suggeststhat its formation may have been structurally the seafloorin the form of a structuralboundary marked by controlled. However, available seismic profile (Figure 9a) sharp reflection terminationsthat strongly diffract seismic across this wall does not allow univocal determination of energy.This structuralboundary terminates at depthagainst a faulting type. stronglyreflective, well-bedded, and generally flat lying layer (unit F, Figure 6), which underliesboth the indentationand the 4.5. Avalanche-Free Upper Indentation lower marginto the south.It is inferredthat thislayer is at the The relatively smooth upper part of the indentationis top of the subductingplate and representspreavalanche pelagic and possiblytrench sedimentsthrust westward beneath generallyless than 1700m deep,although along the northern the margin. Seismic line NZ-47 (Figure 8), which extends margin,it extendsdown to 3300m deep(Figures 3 and5). Its eastwardacross the scarp, shows the northern Hikurangi topographyis depressedby 400-800 m belowthe marginon margin front to consist of imbricate thrust sheets that are eitherside (Figure 3). The upperindentation has only a few sharply cut by the arcuatescarp. Shallow water, calcareous hummocksinterpreted as slideblocks suggesting a largely mudstonesof lower Miocene to lower Pliocene age were avalanche-freetopography. Most of theseblocks appear to havedetached from scarpsinside the upperindentation, but dredgedfrom the main imbricatethrust sheetoutcropping at the scarp(Table 1). We interpretthis scarpas a major scarleft some could have come from the scallopedsouthern wall (Figures 3 and 5). The upperindentation's northern half is by blocks that collapsedin the avalanche. characterizedby ridgesand basins trending subparallel to the 4.4. Western Wall- Indentation Head northernwall. Seismicline 2044 (Figure 7) showsthat the ridges are thinly stratified, southeastwardtilted blocks The westernwall has subduedrelief of only 200-300 m. It boundedby normal faults, which, togetherwith the block's slopesat 4-10 ø and is incisedby gullies and small rotational depressedtopography relative to the adjacentshelf, indicatea slumps. It trendsN27øE, roughly parallel with the regional general subsidencecontrolled by southeastwardextension. COLLOT ET AL.: GIANT RUATORIA DEBRIS AVALANCHE, NEW ZEALAND 19,279

Figure 7. (top) The 2044-3 seismicreflection line drawing,with location (inset) showing rotatedblocks, separatedby normalfaults, with subsequentbasin fill. Central and left basins show sharpdistinction between rotationalblock and basinfill. X767 and X768 are core and samplelocations; see Tables 1 and 2. Right basin shows continuingrotation during basin fill. (bottom) Detail of single-channelseismic data.

The northern flanks of the tilted blocks are unconformably and lower basins, displaces both the seafloor and a 300-m- overlainby horizontallybedded, thinly stratifiedsedimentary thick, seismically transparent layer that outcrops on the wedges,which are -250 m thick and formedmainly after hanging wall of the fault. Rock sample X280 (Table 1) from tilting had ceased.The lower wedgehas, however,recorded this scarp is mid-Miocene calcareousmudstone. Thus our data continuingblock rotationduring fill deposition.We conclude indicate that in the southern half of the upper indentation, that the northernhalf of the upper indentation has subsided early compressionhas been followed by extensionthat is still and that subsidencehas now largely ceased. active. The upperindentation's southern half consistsof two broad benches at 600-1100m deep and 1500-1700m deep, 4.6. Avalanche-Covered Lower Indentation separatedby a seawardconcave scarp centred on 1300 m (Figures3 and5). Both benchesare underlain by sedimentary Most of the lower partof the indentationis depressedmore basins over 1.2 km thick (Figures 9a and, 9b). The basins than 1 km belowthe marginon eitherside and is blanketedby contain strongly reflective and well-stratified sequences a blocky avalanchedeposit describedin section 5. The lower overlain and locally intermingled with a seismically and upper parts of the indentation are separatedby the incoherentlayer of variable thickness(-50-300 m). This layer prominent400-600 m high, 7-8ø eastwarddipping midslope is unconformablyblanketed by 50-100 m of recent reflective scarp which, together with the lowest scarp of the deposits. The basins are deforming now by downslope indentation'ssouthern wall, formsa 30-km-wideamphitheater extension,but earlier compressionis indicated by an anticline (Figure 5). The amphitheateris regardedas the main avalanche near the lower basin's eastern boundary (Figure 9b). The headwall scarp. Seismic reflection line NZ-47 (Figure 8) anticline's crest has been erodedand unconformably overlain supportsthis interpretation, and seismicline GNZ-05 (Figure by a veneerof flat-layingbasin sediments indicating that it is 6) suggestsa total scar height of as much as 2.5 km. Farther no longer active. On the basis of multibeambathymetry, the norththe failuresurface is lesspronounced and appearsmainly anticline's axis trends ENE, which would suggest a westward as an unconformablestratigraphic contact between the margin compressivestress field (Figure 5). Low-amplitudefolds also and avalanchedeposits. Seismic line 3044-37-3 (Figure 9c) occur in basin fill in the upper basin (Figure 9a). Active crosses the midslope scarp and shows a seawarddipping extensionis documentedby eastdipping normal faults N and G boundaryjoining it to the top of the subductingsediments (Figures9a and 9b). Fault N, which boundsthe upperedge of (unit F in Figure 9c inset). Landward dipping, margin the upper basin, cuts a wide antiform, displacing strong reflectorsbelow this boundarycontrast with seawarddipping reflectors. The fault deforms the seafloor, and its recent avalanche reflectors above it. Although not outlined by a activity is recordedon the downthrownside by a rotational major reflector, the boundaryis consideredto be part of the sedimentarybasin. Fault G, which is locatedbetween the upper avalanche failure surface. 19,280 COLLOT ET AL.: GIANT RUATORIA DEBRIS AVALANCHE, NEW ZEALAND

lO km i I LineNZ-471 '-• E VE '4.4

Debris avalanche

o o

Figure 8. (top) Gulfrex NZ-47 seismic reflection line drawing with location (inset) showing imbricate margin lacking a recentaccretionary wedge. The margin possibly includesthe Paleogeneallochthonous mass that outcropsonshore, obducted over well-bedded sedimentscontaining a Miocene microfauna (sample X766, Table 1) and coveredby a basin fill sequencewith thrustingand backthrusting.The margin front is marked by a scar left by blocks that collapsedin the avalanche. Dashed lines are projected structuresfrom line NZ-48 (see inset) suggestingcollapsed blocks. Right side shows debris avalanchedeposit. (bottom) Detail of a stacked section showing imbricated margin, avalanchescar, and debris avalanche.

Northeast of the amphitheater, the lower indentation is a The main lobe of the avalanchedeposit has -70 hummocks 3600-m-deep reentrant of the Hikurangi Trough into the more than 1 km across,including at least 20 hummocks5-18 margin (Figure 4). The reentrantis boundedto the north by the km across.The lobeextends 70 km seawardt¾om the midslope N276øE trendingnorthern wall (Figures3 and 4), landwardby a scarp to -40 km seawardof a line joining the deformation 12ø eastwarddipping slope, and to the south by avalanche fronts on either side. In severalplaces, small enclosedbasins deposits. Seismic lines GNZ-06 and GNZ-11 (Figures 10 and between large blocks have flat-lying and parallel-bedded 11) indicate that the reentrant is underlain by -100 m of sedimentsthat are generally80-120 m thick. In plan view, turbidires above 300-500 m thick avalanche deposits. large hummocks represent -30% of the total area of the avalanche. On the basis of their distribution and size, 5. Debris Avalanche Deposit proximaland distalareas of the avalanchecan be recognized. The proximalarea is mainly landwardof a line joining the The avalanche deposit consists of a main lobe trending deformationfront on either side. It represents-60% of the N155øE, and a northeastern secondarylobe trending N80øE total avalanchearea and is characterizedby hummocksthat are (Figures3 and 5). The secondarylobe has a hummockysurface smallerthan in the distalarea and by tectonic lineaments.The with 30-40 small-sized hummocks (1-5 km across) and a hummocks,ranging from 1 to 3 km acrossand from 50 to 350 topography that steps down toward the northern wall. The m high are predominantly located in the secondarylobe body of the secondarylobe has short, irregularreflectors that (Figures 3, 5 and 10). Others hummocks, located in the main extend beneath the deep water reentrant (Figure 10). It is lobe, are elongated and form ridges, 5-9 km in length, inferred that this northern lobe was prevented from entering trending consistently -N60øE along short structural the KermadecTrench by a basementridge (R1 in Figure 5) that lineaments. Some hummocksalign subtransversallyto the acted as a dam to both avalanche and later turbidity currents. margin, along two remarkablestructural lineaments, trending COLLOTET AL.: GIANT RUATORIA DEBRIS AVALANCHE, NEW ZEALAND 19,281

iii

,- (s) OWLLcq • 19,282 COLLOT ET AL.: GIANT RUATORIA DEBRIS AVALANCHE, NEW ZEALAND

3044'37.2G Rotationalslump 0 5•km SE .,••ß l!t \ ,

•...... ,•.;•;• . :•,.,•:.<-'•:.: .:•:• "[q:...• ' *"•, . - . ':' ;' •'::'*'**.;.,.::.l.,,, '

(b)

3..:.•37.3 0 5km SE ba.'Sal •d.ace Raft ied BloCks

Thrust fau.lt' ?

No data

(c)

.l,I ",' ! ' ' ,, , Knoll 0, s:kin ....- 3044-37.4X765• ...... , ,..,%!-• ...... ':":,i%,i;...... • ...... _

,.' ..,•

(d) Figure 9. (continued) COLLOTET AL.: GIANT RUATORIADEBRIS AVALANCHE, NEW ZEALAND 19,283

...... Debris Avalanche 0I 10 kmi GNZ-11 VE: 3.8

BasementRidge "R1" N E NorthernHikurangi Trough

5

'•'•:.

Figure 10. (top) GNZ-11 seismicreflection line drawing,with location (inset), showingavalanche deposit of the secondarylobe being blockedby a basementridge that preventspassage to the KermadecTrench. X282 sampledata are given in Table 1. Thin turbiditefill overliesthe avalanchedeposit south of the ridge. (bottom) Detail of a migrated sectionof the line.

N125øEon the averageand offsetting the seafloorof the The distal part of the avalanche, which is seawardof the proximal area. The longest of the two lineamentsextends 10 deformationfront on either side, represents40% of the km insidethe upperindentation. A N150øEtrending, 25-km- avalanchesurface. It includesa clusterof megablocks. Five of longlineament, also deforms the proximalavalanche deposit them exceed10 km in their greatestdimension and reach andcuts the upperindentation. These lineaments postdate altitudes> 600 m abovethe surroundingseafloor. The largest boththe indentationand avalanche deposit. Seismic reflection block,which is 18 km long, 1200 m high abovethe adjacent profilesGNZ-05 (Figure6), L-37-3 (Figure9c), and GNZ-06 seabed, is the most distal and has been named Ruatoria Knoll. (Figure11)across the proximalavalanche show generally It has undulatingreflectors, with gentle, landwardapparent incoherentreflections overlying unit F, a stronglyreflective dips(Figure 9d) and[Lewis and Pettinga, 1993], indicating a layerwith distorted,near-parallel bedding. This layer, which deformedsedimentary structure, similar to that of imbricated extends beneath the entire avalanche mass and to seaward in sedimentson the adjacentmargin (Figure 8). Strongreflector the HikurangiTrough, is generally-0.6 s two-waytime (twt) R at a depth of 5.5 stwt beneath the Ruatoria Knoll's thick and can be recognizedat a similar depthbeneath the northwesternflank may be either the base of the block or a marginsouth of the Ruatoriaindentation. We interpretunit F side echo. However, if unit F, which extends beneath its as pelagicdrape perhaps with a coverof preavalanchetrough southeasternflank is extrapolatedbeneath the knoll toward fill [Lewiset al., 1998]. The top of unit F is inferredto mark reflector R, then the block's maximum thickness is the baseof the avalanche.On basisof this interpretation,the -2.5 stwt. Dredged samplesfrom the northeasternface of the maximumavalanche thickness, assuming a soundvelocity of Ruatoria Knoll consist of mid-Oligoceneto mid-Miocene 2000m s-• is -1.7-2.0 km (Figure9c). Despitethe general calcareousmudstone from mainly upperslope environments absenceof coherentbedding in the avalanchedeposit, some (Table 1). Essentiallysimilar rocks and fauna were dredged hummocksreturn well-bedded,high-frequency reflections, fromthe walls of the indentation,including the steepscarp at with clear lateral boundaries,that clearly delineatethinly the easternend of the southernwall (Figure 8 and Table 1). stratified and gently deformedblocks (Figure 11). These Lithologiesand faunas in the blocksare completely different blocksare isolated in theavalanche matrix, and they appearto from those that occur in the pelagic drape of oceanic be rooted at a depth of 0.8-1.2 km beneath seafloor. The seamounts[Lewis and Bennett, 1985], and the samples are internalstructure of someblocks (Figure 11) resemblesthat of proof that even the largest blocks were derived from the the anticline in the upperpart of the indentation(Figure adjacent margin. 9b).We infer that the vast majority of blocks locatedin the The shape and relative position of the blocks give an proximalarea are rafted,having slid from the midslopescarp indicationof the avalanche'sdirection of flow andpattern of or the southern wall. emplacement.Blocks have remarkablemorphologies of two 19,284 COLLOT ET AL.: GIANT RUATORIA DEBRIS AVALANCHE, NEW ZEALAND

z

.,•

(s) em!• COLLOT ET AL.: GIANT RUATORIA DEBRIS AVALANCHE, NEW ZEALAND 19,285 main types. The more distal ones are angular and asymmetric were recently shortened. Shortening appearsto have initiated in map view and cross section. For example, the Ruatoria by backthrustingthese units againstthe toe of the margin, and Knoll, is rhomboidal in plan view with a gently dipping and then a thrust fault propagatedseaward, creating an incipient convex southern flank and three linear and steep sides. We accretionary lobe. infer that the gentle convex flank, now inclined toward N215øE, representsthe original surfaceof the lower Hikurangi 7. Discussion margin. If indeed this is the original margin slope, then the block has rotated 90ø clockwise, from its original dip toward 7.1. Indentation and Slope Failure: N215øE. Significantly, the three steep sides mirror similar, Results From a Two-Stage Process steep,linear flanks of neighboring blocks, implying that all were once joined, perhapsas part of one even larger block. The structural data presented above indicate that the Most obviousare the N150øE trending, parallel walls between Ruatoria indentation and slope failure are genetically linked. the Ruatoria Knoll and another large block to the east, However, they are not synchronousand did not result from the suggestingthat these partedat a very late stage. The N150øE same process. The similarities between geological structures alignment of the block's walls and the longest axis of the within the Ruatoria upper indentation (Figure 9) and those of main lobe reflect the trajectoryof the avalanche. Other nearby the Hikurangi upper margin immediately to the south (Figure large blockshave differentmorphologies. They show multiple 8) support the idea that the depressedupper indentation summits and narrow crests, flanked by locally steep but seafloor reflects local subsidence of the margin and is smooth slopes that are concave in cross section. This generally not the avalanche failure surface. Sedimentary morphology may reflect disintegration of heavily fractured basins within the upper indentation are the northern rocks. Some of these blocks are aligned subparallel and some continuation of the imbricated Neogene and Quaternary shelf near perpendicular to the inferred N150øE direction of basins to the south (Figures 5 and 8). The 500-800 m amount transport. The matching and alignment of peaks and blocks of subsidence together with preservation of coherent could indicate that one massiveblock, possibly as large as 20 geological structuresin the subsidingbasins, inversion from by 35 km, detachedfrom the margin, breaking up, rotating, compressional to extensional tectonics, and ceased and leaving detachedsmaller blocks at the avalanche'strailing extensional faulting in the northern half of the upper edge, as it traveleddown the slope and for 40 km acrossthe indentation (Figure 7) and continuing extension (Figure 9a) in Hikurangi Trough. its southernhalf attest to a relatively slow subsidenceprocess. In contrast, the cleanness of the avalanche scar along the 6. Debris Flow Deposit indentation's southern wall, the avalanche dislocation pattern, and the blocky angular morphology, as well as the A seismically transparentlayer, which extends beneath the fact that the largest block slid farthest away from the margin Hikurangi Trough for up to 100 km in front of the debris on an horizontal seafloor, attest to high energy transport and avalanchedeposit (Figure 5) andcovers -8000 km2, is inferred therefore a catastrophiccollapse. We concludethat most of the to be a debris flow deposit associated with the blocky Ruatoria indentation formed first by subsidence, whereas avalanche. Seismic lines GNZ-14 (Figure 12) and 3044-38 avalancheand debris flow occurredlater on, instantaneously. (Figure 13), which cross the Hikurangi Trough ahead of the avalanche, show five seismic units overlying the Hikurangi 7.2. Causes of Instability on the Northern Plateau acoustic basement. The deepestunit is the strongly Hikurangi Margin reflective unit F recognizedbeneath the avalanchemass. A 0.3 stwt thick weakly reflective, parallel-beddedunit (unit T)is Both the subsidence in the Ruatoria indentation and the overlain by a set of strongly reflective layers (unit X), which catastrophiccollapse reflect large-scalesubmarine instability in turn are irregularly overlain by the 0.1-0.2 stwt thick within consolidatedrocks of the Hikurangi margin. Several transparentlayer (unit DF). Unit X contains numerouslimited regional factors may contributeto this. but well-stratified reflectors that have different dips and The heterogeneousEast Coast Allochthon is a regional locally clear evidencesof disturbance.The base of unit X is factor of instability. Onshore, allochthonous sheets of commonly, but not universally, recognized at a strong fracturedCretaceous to Paleogene sediments and seamount reflector at the top of unit T. Unit DF shows no coherent blocks up to 2 km thick [Field et al., 1997: Rait, 1995; reflections,and both its upper and lower surfacesare irregular. Stoneley,1968] are buried by Neogenesediments up to 4 km Its upper surface is unconformably overlain by the most thick [Field et al., 1997]. To date, there is no clear evidence of recent, 0.2 stwt thick and well-stratified unit PAT interpreted the extent of the allochthon and covering beds offshore. as postavalanche trough turbidites. We interpret the However, we interpret margin strata with discontinuous transparent unit DF as debris flow deposit and unit X as reflectorsand evidenceof thrusting (Figure 8) as a possible preexistingtrough fill disturbedby passageof the debris flow eastern continuation of the allochthon. Overlying, well- or by dewateringafter rapid loadingby the debrisflow deposit. bedded,gently folded and reversefaulted basin sedimentsare Within -20-30 km of the front of the avalanchedeposit unit X Miocene to Recent cover beds. Parts of the allochthon to the increases in thickness from 350 ms to 800 ms twt, in places northwest were subject to Miocene gravitational at a preexisting fault (Figures 5 and 12). We interpret this remobilization [Hayward, 1993], and these parts are so increase to result from extensive disturbance of trough fragmented that they have been variously referred to as sediments by the pressure or bow wave in front of the olistostrome, wild fiysch, megabreccia, and chaos-breccia advancingavalanche. Seismic line 3044-38 (Figure 13) shows [Bradley,1964; Kear and Waterhouse,1967]. We suggestthat that the debris flow and the other units younger than unit T its remobilization can still form "chaotic" deposits. 19,286 COLLOTET AL.: GIANT RUATORIA DEBRIS AVALANCHE, NEW ZE••

II

1 I t I I I ! ! t t t II t t 0 I ! t t 1

t

t I I

I

i ..•

u• (s)eLU!.l. =

(s) eWl.I. COLLOTET AL.: GIANT RUATORIA DEBRISAVALANCHE, NEW ZEALAND 19,287

3044-38

0 I km 2 I I I

Figure 13. (bottom)The 3044-38 seismicreflection line drawing,with location (inset), acrossnorthern HikurangiTrough, showing shortening of the troughsediment including debris flow, againstthe toe of the margin.A backthrust(1)cut by a seawardverging thrust(2) suggestsa two-stageshortening with a total shorteningof near3.2 km; unitsare as definedin Figure12. (top) Detail of a migratedsection of the line.

Throughoutthe eastcoast region, an Eocenebentonite originalenvironment of deposition(Table 1), supporting layeris renownedfor its high fluid pressuresand low shear downwarpingof the slopesince Miocene times. strength[Mazengarb, 1998]. It formsthe main detachment High fluid pressure,particularly from gas, dramatically horizonfor the allochthonand lubricates shearing within it reducesthe shearstrength of susceptiblelayers [Hampton et [Fieldetal., 1997]. This highly mobilemud may be squeezed al., 1996; Papatheodorouet al., 1996], and much of the into overlying sequencesand eruptsat the surfaceas mud Hikurangimargin is percolatedby methane-richfluids derived volcanoes[Ridd, 1970; Stoneley, 1962]. It is inferredto from subductingsediments [Katz, 1981; Lewis and Marshall, lubricatenormal faults downthrowntoward the trench 1996]. In general,the porosity of thin subductingsediments [Mazengarb,1998]. If, as suspected,the bentonirelayer is higherthan the averageporosity of thick subducting underliesthe marginoffshore, then it will continueto sediments,particularly if the top of the thick sequenceis representa significantsource of instability. frontallyaccreted [Lallemand et al., 1994.].Thus high fluid The risk of gravitationalfailure in alreadyfragmented flow might be expectedon the northernmargin, where thin, depositsand at lowstrength horizons may have significantly water-richturbidites are wholly subducted, compared with the increasedbecause of regionalseaward tilting. This tilting southernmargin, where thick turbidites are off-scraped and resultsfrom rapid rise of mountainranges onshore [Walcott, frontallyaccreted [Davey et al., 1986;Lewis and Pettinga, 1987] andsubsidence caused by tectonicerosion by the 1993].In addition,bottom-simulating reflectors attributed to downgoingplate on the lowerslope [Collot et al., 1996]. gashydrate occur widely on the Hikurangi margin [Field et al., Micropaleontologicalevidence indicates that most rock 1997; Katz, 1982]. Gashydrates can becomeunstable and sampleswere dredged from waterdepths much deeper than their dissociateinto waterand 170 times their own volumeof free 19,288 COLLOT ET AL.: GIANT RUATORIA DEBRISAVALANCHE, NEW ZEALAND gas, particularly during the release of hydrostatic pressure equalsthe seamountdiameter, especially if it is large andflat- associatedwith falling sea-level [Lercheand Bagirov, 1998]. topped. Oceanic seamountsclosest to the Ruatoria indentation This drastically reducesthe shear strength of many slope are large,flattished-topped and elongated N155øE. We suggest sediments [Hampton et al., 1996] and is a potentially that a seamountof similar size (-35-40 km long), shape,and destabilizing influence on at least shallower parts of the orientationto Gisborneseamount (Figure 2) partly formedthe avalanche Ruatoriaindentation while subductingbeneath the relatively cohesivebut imbricatedHikurangi margin. 7.3. Origin of the Indentation and Avalanche: 7.3.4. Present-day location and depth of the Results of Oblique Seamount Subduction subducted seamount. Oceanic seamounts buried at shallow

The Ruatoria avalanche and indentation are not a result of depthbeneath an accretionarywedge have been either imaged from seismic reflection [von Huene et al., 1997] or, more simple slope failure. There is compelling evidencethat they often, indirectly identified by a magnetic anomaly are a responseto oblique subductionof a seamount. [Barckhausen et al., 1998; Lallemand and Chamot-Rooke, 7.3.1. Criteria used to recognize seamount 1986] or by a circular uplifted zone associatedwith the thrust impact. Subductingseamounts have producedindentations in of their summit [Dominguezet al., 1998]. The uplifted zone is convergent margins aroundthe world. Their passage beneath incisedby a divergentnetwork of fine subverticalfractures and margins causescompression with back thrusts ahead of the boundedlandward by back thrusts [Dominguez et al., 1998]. asperity, uplift above it, and "tunneling" and collapse in their These identification criteria are not met in the Ruatoria wake [Dominguez et al., 1998; Lallemand and Le Pichon, 1987; Lallemand et al., 1994; Masson et al., 1990; von Huene indentation,possibly implying that the seamountresponsible for the indentationhas already subductedtoo deepbeneath the and Lallemand, 1990]. Impactingseamounts produce first a U- margin. Experimental modeling shows that normal faulting shapedreentrant in the accretionary wedgedeformation front, deforms the cohesive part of the margin above the trailing then a semicirculardepression in the lower margin, and finally flank of the seamount, but activity ceaseswhen the seamount an elongatedgroove-like indentation of the whole margin. In has subductedfar enough,past the indentation headwall scarp. a non accretionarymargin the groove is generally flanked by Therefore active normal faulting within the upper part of the subparallel scarpsthat trend parallel to the plate convergence Ruatoria indentation (Figure 9a) indicates that the buried direction. Such spectacularimpacts were imaged across the seamount lies somewhere beneath the shelf immediately Costa Rica margin [von Huene et al., 1995]. The indentation landwardof the indentation western wall. A comprehensive may remain long after the seamountthat causedit has passed, seismological study of interplate earthquake distribution and it may accumulateflat-laying sediment and minor slope beneaththe adjacentland [Reyners et al., 1999] suggeststhat failure deposits [Collot and Fisher, 1989]. Experimental a seamount, if located beneath the shelf, would be 10-15 km modeling shows that when the seamount underthrusts the deep, probably too deep to deform the overlying seafloor cohesive part of the margin, subsidenceinitiates above the (Figure 14). seamount trailing flank: former back thrusts are then reactivated into steep normal faults controlling the indentation subsidence[Dominguez et al., 2000]. Seamount 7.4. Timing of Avalanching and Northern Wall Formation subduction is also shown to bulldoze material from the front to beneath the inner part of the margin, thus accountingfor the The avalanche and associated debris flow are shown to be rocks missingin the indentation[Dominguez et al., 2000]. younger than most of the indentation. Their age can be 7.3.2. Impact criteria applied to the Ruatoria estimated from the thickness and sedimentation rates of cover indentation. Such geometrical and morphostructural beds in enclosedbasins surroundedby avalancheblocks, and characteristicscan be recognizedin the Ruatoria indentation trough sedimentoverlying the debrisflow. A 2.93 m core from clearly revealing the structural imprint left by a subducted one of the largest basins (X764 in Figure 2 and Table 2) seamount. First, the northern wall of the indentation, which contains mainly hemipelagic mud with 10 graded silt layers cuts linearly acrossthe entire margin and parallels the PAC- inferred to be turbidites derivedfrom adjacent slopes. It also KER plate convergencedirection, is inferredto be one sideof a containstwo air fall ash layers at 54-57 cm and 109-1 13 cm long, groove-like indentation formed by a subducting below the seabed that are correlated with 1.8 and 3.3 ka seamount.Second, the upper part of the indentation reveals eruptions respectively, on the basis of the geochemistry of the existence of compressional and extensional structures the glass. This indicates a Holocene rate of sedimentation of associatedwith the seamountpassage. Third, the 3600-m-deep -0.33 m kyr'J. An acceleratormass spectrometer(AMS) margin reentrant(Figure 4) at the seawardend of the northern radiocarbon age of 8775+60 years B.P. for planktonic wall denotesremoval of margin material that we believe has foraminifers from the base of the core also gives a rate of been pushedand draggedbeneath the margin by the seamount. 0.33 m kyr't, confirmingthis as a uniformrate for Holocene. Fourth, the compactedvolume of the indentationis larger than Extrapolating the same rate to the base of the 120-m-thick the compactedvolume of the avalancheas discussedin section basin would imply that depositionbegan there -360 ka. South 7.7 (Table 3). of the avalanche deposit, a 0.97-m core was recovered from 7.3.3. Size and shape of subducted seamount. 140-m-thick parallel-beddedtrough fill overlying the debris The shape of the indentation and the margin's deformation flow (X770 in Figure 2 and Table 2). The core consists of four pattern dependupon both the margin and seamountgeometry relatively thick silty turbidireswith thin hemipelagic layers and rock properties [Dominguez et al., 1998]. Experimental and ash layers at 61-64 cm and 94-97 cm below the seabed, modeling indicates that the seamountimpact groove's lateral identified as the same two 1.8 and 3.3 ka ash layers. With scarps are better expressedin a cohesive margin than in an comparatively few but thicker turbidite layers comparedwith accretionarywedge and that the groove's width approximately the enclosed basin on the avalanche, long-term deposition COLLOTET AL.: GIANT RUATORIA DEBRIS AVALANCHE, NEW ZE• 19,289

North Island Ruatoria Indentation Debris avalanche Coast E • km ½ o

30

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 km Figure14. Interpretedcrustal cross section through the Ruatoria indentation and avalanche, showing possibleposition of subductingseamoments. Location shown in inset.Interplate surface beneath North Island eastcoast after [Reyners etal., 1999];HP, HikurangiPlateau thickness after [Davy, 1992]' dashed line is the inferredpreindentation topography.

ratesare difficult to estimatefrom short cores but are broadly 7.5. Scenario for Oblique Margin Indentation and similar to those in the enclosed basins. Slope Failure Age extrapolations basedon Holocenerates over-estimate theavalanche debris flow age as rates of terrigenousinput to Geometricalreconstruction and avalanche dating help to thecontinental slope were several times greater during glacial constrainthe timing of the indentationand slope failure. ages. A piston core from an enclosed basin on the lower Accordingto the 54 km Myr-I PAC-KERplate convergence Hikurangi margin 250 km south of the Ruatoria avalanche rateand N277øE direction and the presumed location of a large suggeststhat glacial age sedimentationrates were -4 times buriedseamount beneath the continental shelf, it is inferred fasterthan Holocene rates, due mainly to windblowndust and that the seamountimpacted the marginbetween 2.0 and 1.3 to an increasedfluvial supplydirectly to the slope from a Ma (Figure15a). The 3600-m-deepmargin reentrant formed as largely deforestedlandscape [Stewart and Neall, 1984]. the seamountdisappeared beneath the marginand probably However,a piston core from the samearea indicates only a actedas a sinkfor small slumpsand debrisflows that occurred modestincrease, perhaps only 50%, from the lastinterglacial in the wakeof the seamount.Experimental modeling shows ageand falling sea level to theheight of thelast glacial age, that whena seamountis beingsubducted, the interplate with onlya slightdecrease into the Holocene(L. Carteret al., d6collementis upliftedabove the seamountroof and remains pets. oral communication, 1999). Farther offshore, a core deflectedupward above the seamounttrailing flank, so that a from the centralpart of the HikurangiPlateau showed a shadowzone develops both in front of andin the wakeof the threefoldincrease in the last glacial age sedimentation seamountallowing frontal margin material and wake slump accumulationrates with respectto that of the Holocene. If the massesand trenchfill to be pushedand draggedinto the threefoldto fourfoldincrease for'glacial ages were correct, subductionalong with the seamount (Figure 14) [Collotet al., thena roughmedian rate for the LateQuaternary based on sea 1992;Dominguez et al., 2000; Lallemandet al., 1994]. We level curves,might be -0.7+_0.1m kyr'•. Consideringa believe that this processaccounts for the formation of the 120+_12m thick enclosedbasin, we derivedan age for the deepwater reentrantof the Ruatoriaindentation. Between 2 avalanche of-180+_40 ka. and 0.16 Ma, the subductingseamount cut a 30-km-wide A 3.81 m core recoveredfrom one of the small basins near groove-likeindentation obliquely through the marginwith the westernend of the indentation'snorthern wall (X767. uplift in front and subsidencebehind the seamount. The Figure2 andTable 3) helpsto constrainits age there.An AMS grooveand northern wall are diachronous. Because of oblique age of 1999+_60years B.P. for planktonic foraminifersfrom convergence,this process left an unstabletriangular wedge in the base of the core gives an exceptionallyhigh theacute angle between impact groove and the steep edge of a sedimentationrate of 1.9 m kyr-•. However,for the last 1000 tectonicallyeroded margin (Figure 15b). This marginwas years or so, shelf deposition rates have increased alreadyrendered unstable by seawardtilting of low strength significantly, following deforestationof the adjacentland Eoceneclays and Miocene slide deposits and by percolating after humanoccupation [Carter et al., 2001]; the only fluids.It requiredonly a largeearthquake, or disruptionby available dates for deforestation on North lsland's east coast subductionof anotherseamount, to causethe triangleto appearto confirm the offshore recorddespite uncertainties collapseat ---170ka (Figure15c). Unusually large earthquakes regardingcontamination [McGlone and Wilmhurst,1999]. We at a site predisposedto failure are commonlycited as the estimatethat an average sedimentationrate for glacial triggeringmechanism for large submarineslope failures (deforested)-interglacial(usually forested) cycles might be [Hamptonet al., 1996].At thenorthern Hikurangi margin the -1.6 +_0.2m kyr-l. Sucha ratewould imply that the 250 +_25m plateinterface is presentlyweakly coupled and accumulating thick basin overlying a tilted block formed-160+_30 ka. strainis generallydissipated in numeroussmall earthquakes Becauseof age uncertainties,we suggest that both the [Reynersand McGinty, 1999; Smith et al., 1989]. However, avalancheand block tilting occurredat -170+_40ka. thesubduction of a largeseamount may locally cause locking 19,290 COLLOTET AL.:GIANT RUATORIA DEBRIS AVALANCHE, NEW ZEALAND

t - ,-,0.17 t ....= 0 M.a

Figure15. Scenario for formation of Ruatoria indentation andavalanche from 1.3 Ma to Present: (a)Initial impactof a largeseamount similar to the Gisborne seamount andformation of the deep-water trough reentrant; (b)formation of the groove-like indentation, the unstable triangular wedge, and beginning of extensional faultingin upperindentation; (c)slope failure that resulted in avalanche (Av)and debris flow (DF) deposits triggeredeither by an earthquake (star?) or/and a small seamount (?)detected asa magneticanomaly; solid arrowis PAC-KERconvergence vector; hatched area is avalanchedeposits that is subsequentlyshortened or subducted;(d)present-day situation with large seamount buried 10-15 km beneath the shelf (Figure 14).

of the plate boundarysufficient to generatelarge earthquakes trailing flank of the largesubducting seamount could have [Cloos, 1992; Scholz and Small, 1997]. The large seamount inducedthe rotational collapse,subsequently triggering the that incised the indentation may have generatedearthquakes failureof the triangleof lowermargin weakened rocks. large enough to cause the wedge failure. In addition, Postindentationand postavalanchetectonic processes have deformation by a small seaTnountidentified from magnetic also affectedthe margin. The saw-toothedpattern of the anomalies [Collot and Davy, 1998, Figure 16] near the northernwall of the indentationis likely to reflect effectsof a indentation'ssouthern wall (Figure 15c) may also have helped postindentationstrike-slip tectonic component. Although the collapse.Reflector D from seismicline 3044-37 acrossthe sucheffect is difficultto quantify,a cumulative5-10 km offset midslope scarp(Figure 9a and 9b), which appearsto truncate can be estimated from our data set. Considering that this the base of an anticline, could be the top of the small motionoccurred since i.5 Ma wouldimply a strike-slip rate of subducting seamount. This scenario implies that the 0.3-0.6mm yr'•. A 0.17-Myr-oldavalanche requires --9 km of indentation's southern wall and avalanche are younger than shorteningto accommodateplate convergence since the slope most of the indentation. However, dating of onlapping, failure occurred(Figure 15c). Such shorteningcould have been horizontal stratason tilted blocks (Figure 7) •ar the western accommodatedby avalanche deposit subduction,internal end of the northern wall indicates that a rotational collapse deformationor thrust motion along its basal plane. Several occurredin the upperindentation at aboutthe sametime as the lines of evidencesupport limited shorteningin the avalanche avalanche, i.e., -170_+40 ka. Thus the overall data set can be deposit.Faint westdipping reflectors on seismicline 3033-37 interpretedto suggestthat the slope failure that causedthe (Figure9c) are tentatively interpretedas incipient thrustsin avalanche, released stress in the margin, thus producing the avalanche incoherent mass. These faults may be part of a headwall collapse in the upper indentation. Alternatively, deformation zone associated with recent motion along the destabilizationof the margin above and in the wake of the interplate d6collement (Figures 5 and 14). Structural COLLOT ET AL.: GIANT RUATORIA DEBRIS AVALANCI-I:E, NEW ZEALAND 19,291 lineaments trending N125øE and N60øE across the avalanche depositionhas buried the leading edge. The leading edge of the depositform a patternof conjugatefaults, which is compatible debris flow associatedwith the avalanche reached 1 10 km with shorteningin responseto the N277øE orientedPAC-KER from the slope. Such run-out distancesare dramatic but convergencevector (Figure 5). Shortening is also evident probably normal in the submarineenvironment. The Nuuanu within the debrisflow. The westernpart of the debrisflow is debrisavalanche, off Hawaii, traveled170 km from the base involved in an incipient accretionarylobe that develops of the slope, including 50 km acrossthe flat floor of the againstthe toe of the margin(Figure 13). An estimated3.2 km HawaiiDeep and then for a further120 km up a slopethat minimumshortening at this site suggestsa 60 ka agefor the rises--300m [Jacobs, 1995]. Blocks several kilometers in debrisflow. However, this presupposesthat the debris flow diameter within the Nuuanu avalanche have traveled over originally touchedthe base of the margin at that point and 100 km from the foot of slope, while the enormous musttherefore be regardedas a minimumage. Tuscaloosablock is >50 km fromthe slope.The Storrega Slide,off Norway,has a totalrun-out distance, including distal 7.6. Dynamics of Avalanching debrisflows, of 800km [Buggeet al., 1987;Jansen, 1987]; 7.6.1. Main avalanche- main and secondary sedimentslabs up to 30 km acrosshaving slid 200 km down lobes. The two lobes of the avalancheare of similar ages a slope of only 0.3ø There are now many examplesof since the cover beds of their small, enclosed basins have submarinedebris deposits that have such extraordin. arily long similar thickness.The splitting of a single avalanchein two run-out distances [Gee et al., 1999]. Theoretical lobes may have beencontrolled by the NW trendinghigh in considerationsimply that effective friction decreaseswith increasing size, so that the enormousdebris avalanches and basement topography that outcrops just seaward of the divisionbetween the lobes(R2 in Figure 5). We infer that the debrisflows that occuroffshore have long run-outdistances main avalanchebegan in the southernpart of the indentation. [Campbelland Grantmackie,1995]. Inevitably too, submarine Its northernedge flowed landwardof the basementhigh. Its flows are fully saturated,and high excesspore pressureis main part was divertedsouth of the high and out acrossthe important,not just at failure, but duringthe whole event Hikurangi Trough. [Noreraet al., 1990]. In manycases, this is achievedby an 7.6.2. Large frontal blocks. The main Ruatoria avalancheor debrisflow overrunningand incorporating saturated fine sediments from the floor of the avalanche avalanchedeposit has the largestblocks at its leading edge, unlike most other large avalanchedeposits, which have the pathwayinto the basallayer [Geeet al., 1999; Sassa,1988]. biggest blocks somewhere in the middle [Jacobs, 1995; The long run-outsalso imply that they have considerable Masson, 1996; Moore et al., 1989]. This may be at least momentumand hence velocity. Onshore debris avalanches partly a function of the original, prefailure profile of the havebeen clocked at 35 m s-• [Jacobs,1995], andalthough margin as well as relative velocities during transport. If the offshoreavalanches may be slower,they arestill likely to be largestblocks form where the failing mass was thickest, then catastrophicevents. Although there may have been low the oversteepened,tectonically eroded, lower slope of the frictiontransport for muchof the long run-out,a narrowzone northern Hikurangi margin would producethe largest blocks of deformed trough sediments in front of the Ruatoria near its leading edge. More typical, concaveupward slopes avalanchedeposit suggestshigh friction "bulldozing," might producethem in the middle. In some situations, large presumably as the avalanchecame to a halt. blocks can "ground" like icebergs with the rest of the flow 7.6.4. Debris flow and pressured turbidites. moving more quickly aroundthem [Massonet al., 1998]. They Unlike debrisavalanches, debris flows are mainly composed can also have the momentum to outrun the rest of an avalanche of unconsolidatedsediment without mega blocks [Urgeleset by trappingfluid or hydroplaningover a fluid substrateat their al., 1997]. Only a smallamount of soft sedimentis likely to leadingedge [Lipman et al., 1988; Mohrig et al., 1998]. The have come from the failed margin, which consists of Ruatoria Knoll and its adjacentblocks moved for over 40 km tectonically eroded, mid-Cainozoic rocks at or near the seabed. over a flat plain by riding over thin, soft, unconsolidated, However,the blocky avalancheadvancing at high speedfor sandymud turbidiresthat rest on Cainozoic pelagic drape. The 50 km on a 30-km-widefront may havegenerated a pressure momentumto achievethis came from their initial fall height. wave capableof mobilising soft sedimentin front of itself. If, as we suspect,the initial massiveblock originated from the The instability causedby an advancingdebris avalanche is lowermost of two large scallopsin the southern wall, then the knownto havetriggered a massivedebris flow in the Canary depth difference between the knoll summit and the upper Islands[Masson, 1996]. Elsewhere,a morefluid basal phase headwall (Figure 8) is -1000 m, suggesting that the block, continuesas the main body"freezes" [Gee et al., 1999]. We do which may have been as large as 20 * 35 km, tull by about this not know what thicknessof trough sedimentwas remobilised height, probably producing a large tsunami. The Ruatoria by the advancingavalanche, but the absenceof units T and X Knoll is certainly enormousto have moved so far acrossa flat (Figure12)beneath the avalanchemass suggests that it could plain.Its volumeis estimatedto be -200 km•. Althoughlarge, reachlocally 400 m and be 250 m on averageover the whole this is lessthan a quarterof the size of the basaltic Tuscaloosa of the troughaffected by the avalanche.We suggestthat the Seamountin the Nuuanudebris avalanche deposit off Hawaii debrisflow is predominantlyredeposited trough turbidires, and [Jacobs, 1995]. However, the Ruatoria Knoll may be the becauseof this, its volumeis irrelevantin latercomparisons of slope failure and indentation volumes. largest sedimentaryavalanche block so far discovered. 7.6.3. Large avalanche, high-speed, long run 7.7. Areas, Volumes, and Mass Balance out. The main lobe of the avalanche traveled on a 30-km-wide Calculation front out across the soft, flat sediments of the Hikurangi Trough. If the avalancheoccurred at least 170 kyr ago, then it Areas and volumes of rocks associated with the Ruatoria traveled at least 50 km acrossthe trench becauseconvergence indentation,avalanche, and debris flow werecalculated using has subsequentlycarried it -9 km back toward the margin and digitalterrain models (DTM) compiledfrom swathbathymetric 19,292 COLLOT ET AL.: GIANT RUATORIA DEBRIS AVALANCHE, NEW ZEALAND

•o o

q6ndJ1 !6u•Jml!H

la^al •es MOlaq SaJle• COLLOT ET AL.: GIANT RUATORIA DEBRIS AVALANCHE, NEW ZEALAND 19,293 and seismic reflection data. In addition to the areas of the Table 4. Compaction of Avalanchea indentation,avalanche, and debrisflow deposits(Plate l d and Table 3), three 3-D surfaces were defined and used to calculate Slices,m Vol, km3 P1, P2, Pav, C2, % Vcomp,km 3 % % % volumes. The three surfaces are (S1) the simplified reconstructedtopography of the marginand trench, interpreted 0-300 1208 26 60 46 54 652 as beforesubsidence and avalancheoccurred (Plates I a and 1d), 300-600 968 22 48 37 63 609 (S2) the presentseafloor topography (Plates lb and l d), and 600-900 595 20 38 30 70 416 900-1200 265 19 30 25 75 198 (S3) the base of the debris avalanche derived from seismic 1200-1500 95 18 27 23 77 72 reflection data. 1500-1800 15 17 25 22 78 11 A 3146 km3 grossvolume for the debrisavalanche (Va in >1800 0.6 15 22 19 81 0.5 Table 3; Plate l c) wascalculated by subtractingthe baseof the Total 3146 37.8 62.2 1958 avalanche (S3) from the present topography (S2). The volume's accuracydepends on errorsmade on parametersthat aAvalanchedeposit is divided into 300-m-thick slices. Slice volumes (Vol) are calculatedusing GMT 3.1 package. P1 is porosityfor rafted includeseismic velocities, seismic reflection picking, area blocks;rafted blocks are likely to be fractured and their porosityis contours, and reconstructionof preavalanchetopography. consideredto be -20% on average,slightly higher than the margin rocks Error calculationwas simplified by assumingthat the debris from which they originated.P2 is porosityfor matrix material; avalanche avalancheis a 62-km-long, 39-kin-wide, and 1.3-kin-high matrix has a depth-dependentporosity similar to that of debris flows parallel-sided block on which uncertainties of _+1 km are drilled elsewherein the world. Porositiesas high as 65% were measured over the upper 100 m below seafloor (bsf) of cores from a debris flow estimatedfor length and width and _+0.2 km for height. offshoreBaja California, and a 60% average porosityis reportedover Uncertainties are derived from seismic picking time and 300 m of cores collected in mass flows at the innerwall of the middle velocity.Taking into accountthese errors yields a _+629 km• America trench [Baltuck et al., 1985]. We used a 40% porosityat a uncertainty on the avalanche gross volume (-20%). By depthof 600 mbsf, similar to that obtainedin silt turbiditesat the toe of the Nankai accretionary wedge [Taira et al., 1991]. Porosity per applyingthe same 20% error, a 1508_+ 302 km3 grossvolume avalanche slice calculated for 40% rafted blocks and 60% matrix: was obtained for the indentation (VI+V2 in Plates l d and le Pav=0.4*PI+0.6*P2. Slice compactionis C2 = 100-Pav. Compacted andTable 3) by subtractingthe presenttopography (S2) from volumefor each slice is Vcomp=Vol*C2. Numbersare roundedup to the reconstructedpreindentation topography (S1). Assuminga closestinteger except for vol >1800 to nearestdecimal. seismicvelocity of 1.7 km s'•, the debrisflow generallythins from -170 m thick near the avalanchedeposit to -65 m at its abruptseaward edge, so that, with an areaof 8000 km2, the volumes (Plate ld and Table 3), and C1, C2, and C3 the debrisflow still involves-960 km:• of uncompactedsediment compaction factors for margin rocks, avalanche, and debris (Vd in Table 3). flow deposits, respectively. Volume V3 appears in both Becauseof its size and comprehensivedata set the Ruatoria equationmembers but with different compaction factors since indentation and associatedcatastrophic failure appearto be a it involves margin rocks which remainedin the margin after good natural example to conducta comparison between the slope failure. By considering that most of the debris flow negative volume of the indentation and the positive volumes consists of remobilized trough fill as discussedabove, this of the avalanche and debris flow. Because the avalanche, equationcan be simplified as: debris flow, and margin rockshave different porosities,a valid (V 1+V2+V3) * CI - Va * C2 = 0. (2) comparisonmust take into accountproper compaction factors Porosity and therefore compaction estimates in the for each componentof the system.In a closed system, such as avalanchedeposit and indentationrocks are difficult. Using an a passive margin, mass conservation would imply equality average 15% porosityP1 from adjacentEast Cape rocks [Field betweencompacted volumes of the missing margin rocks and those of the avalanche and debris flow. The mass conservation et al., 1997] that we believe are similar in lithology and age to the rocks of the indentation, we derived a compacted equation is: VI+V2+V3 indentationvolume of 2570 km3 (Table 3). On the (VI+V2+V3) * C1 = Va * C2 + Vd * C3 (1) basis of porosity values indicatedin Table 4 and considering with V1, V2, and V3 the volumesof the different parts of the that the Ruatoria avalanche deposit was emplacedas a single indentation, Va and Vd the avalanche and debris flow gross event, we calculatedthe porosity of the debris avalanche as a

Plate 1. Models usedin mass balance calculations(equal-area Lambert projection). (a) Reconstructedtopography of the marginand trenchbefore subsidence and avalanche,(b) presentseafloor topography, (c) isopachof the avalanche(S2-S3); (d) crosssections before (S1) and after (S2) subsidenceand avalanching;base of avalanche(S3); avalanche(Va=S2-S3) is blue and stripedblue areas; V1 is negativepart of (S2-S1) minusV2; V2 is negative part of (S2-S1) in avalanchecontour; V3 is (S2-S3) in indentationminus V4; V4 is positive part of (S2-S1) in indentationcontour, only usedto calculateV3; inset map showing extentof indentation(green) and avalanche(blue), with black line showingno changein height before and after the avalanche (S2-S1=0); (e) is the differencein height between topographies after (Plate lb) and before (Plate la) subsidenceand avalanching.S1 (stageA)was constructedby replacingindentation topography by interpolatedbathymetric contours between indentation'ssouthern and northernwalls and by removingavalanche and debrisflow on downgoingplate. S3 was obtained by pickingavalanche's base on seismicreflection profiles (Figure 2) and convertingtime to depth using a meanvelocity of 2000 m s-!. Thisvalue is consistentwith that of a debrisavalanche volumetrically composed of 40% of raftedblocks and 60% of matrix as estimatedfrom interpretationof strike and dip seismicreflection profiles. Multichannelseismic reflection data on the continentalshelf provide a 2400 m s'! averageinterval velocity for stratifiedsediments from which rafted blocks are likely derived.An averagevalue of 1700m s'• is assumedfor the matrix by comparison with values measured for debris flows [Curray et al., 1982;Moore et al., 1982;von Huene et al., 1985]or sandy,silty trench turbidites [Taira et al., 1991]. 19,294 COLLOT ET AL.: GIANT RUATORIA DEBRIS AVALANCHE, NEW ZEALAND weightedaverage of the matrix porosity and rafted margin beyondthe toe of theslope on thenorthern Hikurangi margin blocks for a series of 300-m-thick avalanche slices and derived will takeonly •-1 Myr to be carriedback to the margin,where a correspondingcompacted volume for each slice (Table 4). it will beeither shortened and plastered against the marginor The compactedvolume for the debris avalanche is 1958 + 392 cardedbeneath it. Suchdeposits may be the gravitationally km3 assuminga 20% error.This volumeis 612 km3 smaller redepositedand tectonically deformed"chaos breccias" and than the 2570+514 km3 compactedindentation volume, a olistostromes of future landmasses. value that is significant with respect to the error on the compactedvolumes. In an open systemsuch as the Hikurangi 8. Conclusions subductionzone a significant negative differencebetween the compacted indentation and avalanche volumes can be The mainresults of ourstudy are as follows: interpreted as the result of avalanche subduction or tectonic 1. The Ruatoriadebris avalanche deposit covers an areaof erosion of the margin. The lack of evidence of avalanche -3400 km2, is upto 2 km thick,and has a grossvolume of deposit underthrustingthe margin favors tectonic erosion over3100 km 3. It hastwo lobes, the main one originally prior to avalanching,a processthat is alreadysuggested by extendingover 50 km fromthe toe of the margin.It hasat the structuralanalysis of the indentation. One can note that least100 blocks>1 km across,including Ruatoria Knoll that this difference in volumes is close to the 790 km 3 of the is 18 km in maximumdimension. The largest blocks, which compacted upper indentation volume (V1, Plate 1) and maybe over2 km thick,are near its leadingedge. consistentwith removal of margin material by basal erosion. 2. Seawardof the avalanche,a 65-170 m thick debrisflow Our calculations also indicate that the compactedavalanche depositthat extends over 100 km fromthe margin, may have volume (1958 km3) match, within errors, the V2+V3 lower formedlargely from soft trough sediments mobilized by the indentationvolume (1780 km•). However,if our scenariois pressure wave in front of the avalanche. correct, then the northern part of the lower slope had already 3. The Ruatoriadebris avalancheand its associateddebris been removedby seamountsubduction resulting in the deep- flow occurredcatastrophically -170 +40 ka, afterfailure of a water reentrant (Figure 15b). Therefore, in addition to the nonaccretionarymargin, already weakened by localgeologic triangle the collapse may also have involved fracturedrocks processesand by fracturing associatedwith subductionof a from the southern side of the groove-like indentation. This largeseamount. A majorrotational collapse occurred at about may indicatean originally diffuse southernwall producedby a the sametime in the margin'supper indentation. seamount with a long southeastern"tail" like the nearby 4. Theavalanche was so largebecause oblique seamount Gisborne and Mahia seamounts. subductionproduced a groove-likeindentation obliquely acrossthe margin, leaving a triangleof unstablemargin in the 7.8. Scarcity of Avalanches on Active Margins: acute angle with the deformation front. It was failure of the Making Olistostromes trianglethat producedmost of the avalanche. 5. The Ruatoria indentation extends 65 km landwardfrom Comparedwith the 26 km• [Crandellet al., 1984,'Hahcox thetoe of the slopeto within25 km fromthe adjacent land and andPerrin, 1994] and50 km• [Philipand Ritz:, 1999] of the encloses an area of 3300 km 2. largest onshorelandslides and paleolandslides,the 3146+629 6. The Ruatoriaindentation has two partswith different km3 Ruatoriaavalanche deposit is enormous.Even so, it is ages.The northern strip, with its northernwall approximating smallerthan landslides reported from passive margins, [Bugge the direction of plate convergence,is the remnantof the et al., 1987; Dingle, 1977], oceanic volcanoes [Moore and groove cut obliquely from •-2 Ma to 0.16 Ma by the Normark, 1994] and collision zones where volumes up to subductingseamount. The southerntriangle of the indentation several10,000 km• in multipleevents have beenreported and its scalloped southern wall fbrmed the avalanche's [Torelli et al., 1997]. On active margins, slope failure behind headwall at -0.17 Ma. subducting seamounts is generally comparatively minor [Collot and Fisher, 1989; Pautot and al., 1987; von Huene and 7. Volumecalculations suggest that the compactedvolume of the avalanchedeposit is 612 km• smallerthan that of the Lallemand, 1990]. However, moderately large blocky indentation,indicating a lossof marginmaterial other than by avalanche deposits have been reported from the Sunda Arc avalanching and supporting the suggestion that the [Moore et al., 1976] and Peruvian margin (•-250 km•) indentationis primarilya seamountimpact depression. [Duperret et al., 1995], and landslides comparable in size to 8. The Ruatoriaavalanche and debris flow depositshave the Ruatoria avalancheoccur on the southernOregon Cascadia beencarried back toward the marginin the last 170 kyr with margin [Goldfingeret al., 2000]. The latter examplesoccur on indicationsof compressivedeformation in both deposits.The an accretionary margin that is subject to subductionerosion resultanttectonically deformeddebris depositsmay be a and oblique convergence,so that some of the factors causing modem analogue of olistostromes in fold belts around the instability are comparable with those at the Ruatoria world. indentation. Despite the factors that lead to instability on subductionmargins in addition to those on passive margins, Acknowledgments.We are grateful to the New Zealand large slope failures are comparatively rare. The combination Foundationfor ResearchScience and Technology(FRST), the National of oblique subductionand regional instability that occur on Instituteof Water and AtmosphericResearch (NIWA), l'lnstitut de the northernHikurangi margin may not be common. However, Recherchepour le D6veloppement(IRD), andthe FrenchMinistry of the main reason may not be intYequentoccurrence but rather ForeignAffairs for fundingand to IFREMERfor providingship time. equipmentand software to processdata. We thankR. Garlickfor help that slope failure deposits are soon destroyedby continuing in volumecalculations, P. Barnesand A. Kopf for discussions,and G. convergencethat carries them back to the margin they came McMurtry and M. Crozier for helpfulsuggestions. This is UMR 6526 from. For instance, an avalanchedeposit that extends50 k m G6osciences Azur contribution 348. COLLOT ET AL.: GIANT RUATORIA DEBRISAVALANCHE, NEW ZEALAND 19,295

References deformationacross the Taupoback-arc and Hikurangitore arc regionsin New Zealand,J. Geophys.Res., 100, 8221-8232,1995. Abbate,E., V. Bortolotti,and P. Passerini,Olistostromes and olistoliths, Davey, F.J.,M. Hampton,J. Childs,M.A. Fisher,K.B. Lewis, and J.R. Sediment.Geol, 4, 521 - 557, 1970. Pettinga,Structure of a growingaccretionary prism, Hikurangi Anderson,H., E. Smith,and R. Robinson,Normal faulting in a backarc margin,New Zealand,Geology, 14, 663-666, 1986. basin: Seismologicalcharacteristics of the March 2, 1987, Davey, F.J., S. Henry, and E. Lodolo, A seismiccrustal section across Edgecumbe,New Zealand,eartliquake, J. Geophys.Res., 95, 4709- theEast Cape convergent margin, New Zealand,Tectonophysics, 4723, 1990. 269, 199-217, 1997. Ballance,P.F., and K.B. Sporli, Northland Allochthon, J. R. Soc.N.Z., 9, Davy,B.W., The influence of subductingplate buoyancy on subduction 259-275, 1979. of the Hikurangi-ChathamPlateau beneath North Island, New Baltuck,M., E. Taylorand K. McDougall,Mass movement along the Zealand,in Geologyand. Geophysics of ContinentalMargins, edited innerwall of theMiddl• America Trench, Costa Rica., Initial Rep. byJ.S. Watkins, F. Zhiqiang,and K.J. McMillen, pp. 75-91, Am. Deep SeaDrill. Proj., 84, 551-565, 1985. Assoc.of Petr.Geol., Tulsa, Okla., 1992. Barckhausen,U., H. A. Roeser,and R. vonHuene, Magnetic signature De Mets,C., R.G. Gordon,D.F. Argus,and S. Stein,Effect of recent of upperplate structures and subducting seamounts at the convergent revisionsto the geomagneticreversal time scale on estimatesof marginoff CostaRica, J. Geophys.Res., 103, 7079-7093, 1998. currentplate motions, Geophys. Res. Lett., 21, 2191-2194,1994. Barnes,P.M., and B. Mercierde L6pinay, Rates and mechanics of rapid Dingle,R.V., The anatomyof a largesubmarine slump on a sheared frontalaccretion along the very obliquelyconvergent southern continentalmargin (SE Africa),J. Geol.Soc. London, 134, 293-310, Hikurangimargin, New Zealand,J. Geophys.Res., 102, 24,931- 1977. 24,952, 1997. Dingle,R.V., Largeallochthonous sediment masses and their role in the Bondevik,S., J.l. Svendsen,ahd J. Mangerud,Tsunami sedimentary constructionof the continentalslope and rise off southwestern faciesdeposited by theStoregg• tsunami in shallowmarine basins Africa, Mar. Geol., 37, 333-354, 1980. andcoastal lakes, western Norway, Sedimentology, 44, 1115-1131, Dominguez,S., S. Lallemand, J. Malavieille,and R. vonHuene, Upper 1997. plate deformation associated with seamount subductiom Bourgois,J., Y. Lagabrielle,P. De Wever,E. Suess,and the Nautiperc Tectonophysics,293, 207-224, 1998. Team,Tectonic history of the northernPeru convergentmargin Dominguez, S., J. Malavieille, and S. Lallemand, Deformation of duringthe past400 ka, Geology,21,531-534, 1993. marginsin responseto seamountsubduction insights from sandbox Bradley,J., Megabreccias,in SyntaphralTectonics and Diagenesis, experiments,Tectonics, 19, 182-'196,2000. editedby S.W. Carey,pp. KI-K5, Univ. of Tasmania,Hobart, 1964. Duperret,A., J. Bourgois,Y. Labagrielle,and E. Suess,Slope Bugge,T., S. Befring, R.H. Belderson,T. Eidvin, E. JanKen,N.H. instabilitiesat an activecontinental margin: Large-scale polyphase Kenyon,H. Holtedahl,and H.P. Sejrup,A giant three-stage submarineslides along the northernPeruvian mai'gin, between 5øS submarineslide off Norway,Geo Mar. Lett., 7, 191-198, 1987. and 6øS, Mar. Geol., 122, 303-328, 1995. Cadet,J.P., et al., The Japantrench and its juncturewith the Kurile Enos,P., Flow regimesin debrisflow, Sedimentology,24, 133-142, trench:Cruise results of theKaiko project,Leg 3, EarthPlanet. Sci. 1977. Lett., 83, 267-284, 1987. Field,B.D., et al., Cretaceous-CenozoicGeology and Petroleum Systems Campbell,H.J., and J.A. Grantmackie,Jurassic Pholadomyidae of theEast Coast Region, New Zealand, Geol. Monogr., 19, 301 pp., (Bivalvia) from New Zealand and New Caledonia,N. Z J. Geol. Inst. of Geol. and Nucl. Sci., Lower Hutt, New Zealand, 1997. Geophys.,38(1), 47-59, 1995. Garcia, M.O., and D.M. Hull, Turbiditesfrom giant Hawaiian Carpenter,G., Coincidentsediment slump/clathrate complexes on the landslides:Results from OceanDrilling Program Site 842, Geology, U.S. Atlanticcontinental slope, Geo Mar. Lett.,1, 29-32, 1981. 22, 159-162,1994. Carter,L., B. Manighetti,M. Elliot, N. Trustramand B. Gomez,Source, Gee, M.J.R.,D.G. Masson,A.B. Watts,and P.A. Allen, The Saharan sea-leveland circulation effects on the sedimentflux to the deep debrisflow: An insightinto the mechanicsof longrunout submarine oceanover the past 15 ka off easternNew Zealand, Global Planet. debrisflows, Sedimentology, 46, 317-335, 1999. Change,in press,2001. Goldfinger,C., L.D. Kuim, L.C. McNei!l, and P. Watts, Super-scale Cloos,M., Thrust-typesubduction-zone earthquakes and seamount failure of the southernOregon Cascadiamargin, Pure Appl. asperities:A physicalmodel for seismicrupture, Geology, 20, 601- Geophys.,157, 1189-1226, 2000. 604, 1992. Hampton, M.A., H.J. Lee, and J. Locat, Submarine Landslides, Rev. Collot,J.-Y., and B. Davy,Forearc structures and the tectonicregimes Geophys.,34, 33-59, 1996. at the obliquesubduction zone between the HikurangiPlateau and Hancox,G.T., and N.D. Perrin,Green Lake landslide:A very large the southernKermadec margin,J. Geophys.Res., 103, 623-650, ancient rock slide in Fiordland, New Zealand, in Seventh 1998. International CongressInternational Association of Engineering Collot,J.-Y., and M.A. Fisher,Formation of Forearcbasins by collision Geology,edited by R. Oliveira, L.F. Rodrigues,A.G. Coelho,and betweenseamounts and accretionary wedges: An examplefrom the A.P. Cunha,pp. 1677-1689,A.A. Baikema,Brookfield, Vt., 1994. New Hebridessubduction zone, Geology, 17, 930-933,1989. Hayward, B.W., The tempestuous10 millionyear life of a doublearc Collot,J.-Y., S. Lallemand,B. Pelletier,J.-P. Eissen, G. Glaqon,M.A. and intra-arc basin-NewZealand's Northland Basin in the early Fisher,H.G. Greene,J. Boulin,J. Daniel,and M. Monzier,Geology Miocene,in Sedimentary.Basins of the World,vol. 2, SouthPacific of the d'Entr•cas•eaux-New Hebrides Arc collision zone: Results Sedimentary.Basins, edited by P.F. Ballance,pp. 113-142, Elsevier from a deep submersiblesurvey, Tectonophysics,212, 213-241, Sci., New York, 1993. 1992. Hinz,K., Tectonicstructure of theconvergent Pacific margin offshore Collot, J.-Y., et al., From oblique subductionto intra-continental CostaRica from multichannel seismic reflection data, Tectonics,15, transpression:structures of the southernKermadec-Hikurangi 54-66, 1996. margin from multibeambathymetry, side scan sonar and seismic Holcomb,R.T., and R.C. Searle, Large landslidesfrom oceanic reflection,Mar. Geophys.Res., 18, 357-381, 1996. volcanoes,Mar. Geotechnol.,10, 19-32, 1991. Coulter,H.W., andR.R. Migliaccio,Effects of the earthquakeof March Hsu,K.J., Melangesand their distinctionfrom olistostromes,in Modern 27, 1964,at Valdez,Alaska, U.S. Geol. Surv.Prof. Pap., 542-C, 36 andAncient Geosynclinal Sedimentation, Spec. Publ., 19, editedby pp., i966. R.H. Dott andR.H. Shaver,pp. 321-333,Soc. ofEcon. Paleontol.and Crandeli, D.R., C.D. Miller, R.L. Christiansen,H.X. Glicken, and C.G. Mineral., Tulsa, Okla., 1974. Newhall, Catastrophicdebris avalanche from an ancestral Mount Jacobs,C.L., Mass-wastingalong the Hawaiian Ridge: giant debris Shastavolcano, California, in Geologyof the Upper Cretaceous avalanches,in Atlas of Deep Water Environments:Architectural HornibrookFormation, Oregon and Cal•brnia, edited by T.H. Stylein TurbiditeSystems, edited by K.T. Pickeringet al., pp. 26-28, Nilsen, pp. 197-201, Soc. of Econ. Paleontol.and Mineral., Tulsa, Chapman_ and Hall, New York, 1995. Okla., 1984. Jansen, E., S. Befring, T. Bugge, T. Eidvin, H. Holtedahl and H.P. Curray,J.R., et al., Baja Californiapassive margin transect: Sites 474, Sejrup, Large submarine slides on the Norwegian continental 475 and476, Initial Rep.Deep Sea Drill. Proj., 64, 35-69, 1982. margin, sediments,transport and timing, Mar. Geol., 78, 77-107, Darby, D.J., and C.M. Meertens,Terrestrial and GPS measurementsof 1987. 19,296 COLLOTET AL.: GIANT RUATORIA DEBRIS AVALANCHE, NEW' ZEALAND

Katz, H.R., Probablegas hydrate in continentalslope east of the North Moore, J.G., and G.W. Moore, Depositfrom a giant wave on the island Island, New Zealand, J. Pet. Geol., 3, 315-324, 1981. of Lanai, Hawaii, Science, 226, 1312-1315, 1984. Katz, H.R., Plate margin transition from oceanic arc-trench to Moore, J.G., and W.R. Normark,Giant Hawaiianlandslides, Annu. Rev. continental system: The Kermadec-New Zealand example, Earth Planet. Sci., 22, 119-144, 1994. Tectonophysics,87, 49-64, 1982. Moore,J.G., D.A. Clague,R.T. Holcomb,P.W. Lipman,W.R. Normak, Kear, D., and B.C. Waterhouse, Onerahi chaos-breccia of Northland, andM.E. Torresan,Prodigious submarine landslides on the Hawaiian N. Z. J. Geol. Geophys.,10, 629-646, 1967. Ridge,J. Geophys.Res., 94, 17,465-17484,1989. Lallemand, S.E., and N. Chamot-Rooke, Sur la cause du d6crochement Mortimer, N., and D.L. Parkinson,Hikurangi Plateau:A Cretaceous s6nestre entre les fosses du Japon et des Kouriles: Subduction- largeigneous province in the southwestPacific Ocean, J. Geophys. collision d'un ancien volcan sous-marin, C. R. Acad. Sci., 303, 1433- Res., 101,687-696, 1996. 1448, 1986. Mulder, T., and P. Cochonat, Classification of offshore mass Lallemand, S.E., and X. Le Pichon,Coulomb wedge model applied to movements,J. Sediment.Res., 66, 43-5, 1996. subductionof seamountsin the Japan Trench, Geology. 15, 1065- Naylor, M.A., Debrisflow (olistostromes)and slumpingon a distal 1069, 1987. passivecontinental margin: the Palombinilimestone-shale sequence Lallemand, S., J.-Y. Collot, B. Pelletier, C. Rangin, and J.-P. Cadet, of the northernApennines, Sedimentology, 28, 113-123,1981. Impact of oceanic asperities on the tectonogenesisof modern Norera,H., J. Locat,and B. Scheildrop,An approachto the physicsand convergentmargins, Oceanol. Acta, 10, 17-30, 1990. modellingof submarineflowslides, Mar. Geotechnol.,9, 93-111, Lallemand, S.E., P.S. Schnurle, and J. Malavieille, Coulomb theory 1990. applied to accretionary and non accretionary wedges: Possible Orange,D.L., andM.B. Underwood,Patterns of thermalmaturity as causes for tectonic erosion and/or frontal accretion, J. Geophys. diagnosticcriteria for interpretationof melanges,Geology, 23, 1, Res., 99, 12,033-12,055, 1994. 144-1,148, 1995. Lenat, J.-F., P. Vincent, and P. Bachelery,The off-shore continuationof Papatheodorou,G., T. Hasiotis,and G. Ferentinos,Gas-charged an active basaltic volcano: Piton de la Fournaise (Reunion Island, sedimentsin theAegean and IonianSeas, Greece, Mar. Geol., 112, Indian Ocean); structuraland geomorphologicalinterpretation from 171-184, 1993. Sea Beam mapping,J. Volcanol. Geotherm.Res., 36, 1-36, 1989. Pautot,G., et al., Deep-seasubmersible survey in the Sugura,Sagami Lerche, I., and E. Bagirov, Guide to gas hydrate stability in various andJapan trenches: Preliminary results of the 1985 Kaiko cruise, geologicalsettings, Mar. Pet. Geol., 15, 427-437, 1998. Leg 2, EarthPlanet. Sci. Lett., 83, 300-312,1987. Lewis, K.B., The 1500-kin-long Hikurangi Channel: Trench-axis Pelletier,B., and R. Louat, Seismotectonicsand present-dayrelative channel that escapesits trench, crossesa plateau,and feeds a fan plate motionsin the Tonga-Lauand Kermadec-Havreregion, drift, Geo Mar. Lett., 14, 19-28, 1994. Tectonophysics,165, 237-250, 1989. Lewis, K.B., and D.J. Bennett, Structural patterns on the Hikurangi Philip,H., andJ.-F. Ritz, Gigantic paleolandslide associated with active margin: An interpretationof new seismic data, in New Seismic faultingalong the Bogdfault (Gobi-Altay, Mongolia), Geology, 27, Profiles, Cores and Dated Rocks From the Hikurangi Margin, New 211-214, 1999. Zealand, edited by K.B. Lewis, pp. 3-25, N. Z. Oceanogr. Inst., Rait, G.J.,Tectonics of theEast Coast Allochthon, 164 pp., Geol. Soc. of Wellington, 1985. N. Z., 1995. Lewis, K.B., and B.A. Marshall, Seep faunas and other indicatorsof Reyners,M., and P. McGinty,Shallow subduction tectonics in the methane-richdewatering on New Zealand convergentmargins, N. Z RaukumaraPeninsula, New Zealand, as illuminatedby earthquake J. Geol. Geophys.,39, 181-200, 1996. focalmechanisms, J. Geophys.Res., 104, 3025-3034,1999. Lewis, K.B., and J.R. Pettinga,The emerging,imbricate frontal wedge Reyners,M., D. Ebemart-Phillips,and G. Stuart,A three-dimensional of the Hikurangi margin, in SedimentaryBasins of the World, vol.2, imageof shallow:,ubduction: crustal structure of the Raukumara SouthPact.'l•c Sedimentary Basins, edited by P.F. Ballance,pp. 225- Peninsula,New Zealand,Geophys. J. Int. 137, 873-890, 1999. 250, Elsevier Sci., New York, 1993. Ridd, M.F., Mud volcanoesin New Zealand,Am. Assoc. Pet. Geol. Bull., Lewis, S.D., J.W. Ladd, and T.R. Bruns, Structuraldevelopment of an 54,601-616, 1970. accretionary prism by thrust and strike-slip faulting: Shumagin Sassa, K., Geotechnicalmodel for the motion of landslides,in region,Aleutian Trench, Geol. Soc. Am. Bull., 100, 767-782, 1988. Landslides-Glissementsde Terrains, edited by C. Bonnard,pp. 37-55, Lewis, K.B., J.-Y. Collot, and S. Lallemand, The dammed Hikurangi A. A. Balkema, Brookfield, Vt., 1988. Trough:A channel-fedtrench blocked by subductingseamounts and Scholz,C.H., and C. Small,The effect of seamountsubduction on their wake avalanches (New Zealand-France GeodyNZ Project), seismiccoupling, Geology, 25, 487-490,1997. Basin Res., I0, 44 !-468, 1998. Smith,E.G.C., T.A. Stern,and M. Reyners,Subduction and back-arc Lipman, P.W., W.R. Normark, J.G. Moore, J.B. Wilson, and C.E. activityat theHikurangi Convergent Margin, New Zealand,Pure Gutmacher, The giant submarineAlika Deris Slide, Mauna Loa, Appl.Geophys., 129, 203-231, 1989. Hawaii, J. Geophys.Res., 93, 4279-4299, 1988. Stewart,R.B., and V.E. Neall, Chronologyof palaeoclimatechange at Masson,D.G., Catastrophiccollapse of the volcanicisland of Hiefro 15 theend of thelast glaciation, Nature, 311, 47-48, 1984. ka agoand the historyof landslidesin the Canary Islands,Geology, Stoneley,R., Marl diapirism near Gisborne, New Zealand, N. Z. J. Geol. 24, 231-234, 1996. Geophys.,5, 630-641, 1962. Masson, D.G., L.M. Parson, J. Milsom, G. Nichols, N. Sikumbang, B. Stoneley,R., A lowerTertiary decollement on the eastcoast, North Dwiyanto,and H. Kallagher,Subduction of seamountsat the Java Island,New Zealand,N. Z. J. Geol.Geophys., 11, 128-156,1968. Trench: A view with long-rangesidescan sonar, Tectonophysics, Strong,C.P., Late cretaceousforaminifera from HikurangiPlateau, 185, 51-65, 1990. New Zealand, Mar. Geol., 119, 1-5, 1994. Masson, D.G., M. Canals, B. Alonso, R. Urgeles, and V. Huhnerbach, Taira,A., et al., Site808, Proc. OceanDrill. ProgramInitial Rep.,131, The Canary debris flow: Source area morphologyand failure 71-285, 1991. mechanisms,Sedimentology, 45, 411 - 432, 1998. Torelli,L., R. Sartori,and N. Zitellini,The giantchaotic body in the Mazengarb, C. Late Neogene structuraldevelopment of eastern Atlantic Ocean off Gibraltar: New results from a deep seismic Raukumara Peninsula:implications for oil exploration, in New survey,Mar. Pet.Geol., 14, 125-138,1997. ZealandPetroleum Co•½krence Proceedings, pp. 201-205, Publicity Urgeles,R., M. Canals,J. Baraza,B. Alonso,and D. Masson,The most Unit, Crown Miner., Minist. of Cornruer.,Queenstown, 1998. recentmegaslides of the Canary Islands: E1 Golfo debris avalanche McGlone, M.S., and J.M. Wilmhurst,Dating initial Maori environmental andCanary debris flow, west El Hierro Island,J. Geophys.Res., impactin New Zealand,Quat. Int., 59, 5-16, 1999.• 102, 20,305-20,323, 1997. Mohrig, D., K.X. Whipple, M. Hondzo, C. Ellis, and G. Parker, yonHuene, R., andS. Lallemand,Tectonic erosion along the Japanand Hydroplaningof subaqueousdebris flows, Geol. Soc. Am. Bull., 110, Peruconvergent margins, Geol. Soc. Am. Bull., 102, 704-720, 1990. 387-394, 1998. vonHuene, R., et al., Site565, Initial Rep. Deep Sea Drill. Proj.,84, 21- Moore, D.G., J.P. Curray, and F.J. Emmcl, Large submarineslide 78, 1985. (olistostrome) associated with Sunda Arc subduction zone, yon Huene, R., J. Bourgois,J. Miller, and G. Pautot. A large, northeasternIndian Ocean, Mar. Geol.. 21,211-226, 1976. tsunamogeniclandslide and debrisflow alongthe Peru Trench,J. Moore,D.G., J.R.Curray and G. Einsele,Salado-Vinorama submarine Geophys.Res., 94, 1703-1714,1989. slideand turbidity current off the southeasttip of Baja California, von Huene, R., et al., Morphotectonicfeatures of the Costa Rican Initial Rep.Deep Sea Drill. Proj.,64, 1071-1082,1982. Pacificmargin surveyed during the Sonne 76 cruise,in Geologicand COLLOTET AL.: GIANT RUATORIADEBRIS AVALANCHE, NEW ZEALAND 19,297

TectonicDevelopment of the CadbeanPlate Boundaryin Southern Wright,I.C., Pre-spreadrifting and heterogeneous volcanism in the CentralAmerica, edited by P. Mann,Spec. Pap. Geol. Soc. Am., 295, southernHavre Through back-arc basin, Mar. Geol.,I13, 179-200, 291-308, 1995. 1993. vonHuene, R., T. Reston,N. Kukowski,G. Dehghani,W. Weinrebe,X. Le Pichon,and S. Lallemant,A subductingseamount beneath the MediterraneanRidge, Tectonophysics, 271,249-261, 1997. Walcott, R.I., Geodetic strainsand large earthquakesin the axial J.-Y. Collot, UMR G•osciences Azur, IRD, B.P. 48, F-06235 tectonic belt of North Island, New Zealand, J. Geophys. Res., 83, Villefranche sur met, France. ([email protected]) 4419-4429, 1978. S. Lallemand, Laboratoire de G•ophysique, Tectonique et Walcott,R.I., Geodeticstrain and the deformationhistory of the North S•dimentologie,Universit• de Montpellier II, F-34000 Montpellier, Islandof New Zealandduring the late Cainozoic,Philos. Trans. R. France. (lallem @dstu.univ-montp2.fr) Soc. London, Ser. A, 321, 163-181, 1987. G. Lamarche and K. Lewis, NIWA, P.O. Box 14901, Wellington, Weaver,P.P.E., Sediment mass wasting in the CanaryBasin., in Atlasof New Zealand. ([email protected];[email protected]) DeepWater Environments: Architectural Style in TurbiditeSystems, editedby K.T. Pickeringet al.,pp. 289-297, Chapman and Hall, New York, 1995. Wood,R., andB. Davy, The HikurangiPlateau, Mar. Geol.,118, 153- (ReceivedApril 13,2000; revised December 8, 2000; 173, 1994. acceptedDecember 22, 2000.)